Springer Handbook of Advanced Catalyst Characterization 3031071247, 9783031071249

Citation preview

Springer

Handbook

oƒ

Advanced Catalyst Characterization Wachs Bañares Editors

123

Springer Handbooks

Springer Handbooks maintain the highest standards of references in key areas of the physical and applied sciences for practitioners in industry and academia, as well as graduate students. Designed to be useful and readable desk reference books, but also prepared in various electronic formats, these titles allow fast yet comprehensive review and easy retrieval of essential reliable key information. Springer Handbooks cover methods, general principles, functional relationships and fundamental data and review established applications. All Springer Handbooks are edited and prepared with great care by editors committed to harmonizing the content. All chapters are written by international experts in their field. Indexed by SCOPUS. The books of the series are submitted for indexing to Web of Science.

Israel E. Wachs • Miguel A. Ban˜ares Editors

Springer Handbook of Advanced Catalyst Characterization With 747 Figures and 45 Tables

Editors Israel E. Wachs Operando Molecular Spectroscopy & Catalysis Laboratory Department of Chemical & Biomolecular Engineering Lehigh University Bethlehem, PA, USA

Miguel A. Bañares Catalytic Spectroscopy Laboratory Spectroscopy and Industrial Catalysis, SpeiCat Institute for Catalysis and Petroleum Chemistry, CSIC Madrid, Spain

ISSN 2522-8692 ISSN 2522-8706 (electronic) Springer Handbooks ISBN 978-3-031-07124-9 ISBN 978-3-031-07125-6 (eBook) https://doi.org/10.1007/978-3-031-07125-6 © Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

This Springer Handbook focuses on Advanced Catalyst Characterization and covers the cutting-edge catalyst characterization methods for heterogeneous catalysts in powder form. Special emphasis is given to characterization under relevant reaction conditions (in situ and operando spectroscopy) and catalysis dynamics. Each modern catalyst characterization method is covered by a leading expert that is intimately familiar with the specific method. Additionally, case studies are provided to showcase how the specific advanced characterization methods are now incorporated in the investigation of different types of heterogeneous catalytic materials in powder form. The chapters are arranged in eight parts: Part I: Vibrational Spectroscopy Part II: Electron and Photoelectron Spectroscopy Part III: Electron Microscopy Part IV: Particle Scattering Part V: X-Ray Methods Part VI: Magnetic Resonances Part VII: Transient and Thermal Methods Part VIII: Soft Operando This handbook brings the knowledge of leading experts and practitioners of characterization techniques for heterogeneous catalytic materials at the atomic and molecular levels, with emphasis on techniques that can operate under reaction conditions, thus providing an advanced characterization, closer to actual working conditions. The chapters in this handbook provide insights on emerging methodologies and advanced in situ and operando spectroscopy characterization methods for the characterization of the bulk and surface of solid catalytic materials. The application of in situ and operando spectroscopy to catalysis studies is poised to advance catalysis science by establishing fundamental structure-activity relationships that can potentially guide the molecular design of advanced catalysts. We would like to especially thank the contributing authors since without their excellent contributions this handbook would not be possible. Madrid, Spain, and Bethlehem, PA May 2023

Israel E. Wachs Miguel A. Bañares

v

Editors’ Introduction

“in the beginning, . . . there was light . . .” when light met matter, . . . there was spectroscopy.

Catalysis is a phenomenon that was initially not understood as first described by Berzelius in 1835 (“I shall. . ... call [the catalytic] bodies [i.e., substances] the catalytic force and the decomposition of [other] bodies by this force catalysis. . .”) [1]. (After reviewing Eilhard Mitscherlich’s research on the formation of ether, Berzelius coins the word katalys (catalysis) on p. 245.) The nature of this catalytic force was not at all understood and required about a century for its understanding to begin since methods were not available to determine the state of the catalysts (surface and bulk) and the nature of the “catalytic force” and activated molecules (“[other] bodies”) during catalysis. Without knowledge of the catalyst structure and chemical state and their effects on reacting molecules, the catalyst was viewed as a black box and the fundamentals of catalysis couldn’t be established. In subsequent years, methods were developed based on the interaction of multiple radiation energies with matter that provided information about catalyst structure, chemical state, and activated molecules, which is the basis for spectroscopy. Illuminated by radiation, spectroscopy allows us to finally “see” the catalyst structures and chemical states (surface and bulk), catalytic force (active sites), activated molecules (reaction intermediates), and their dynamics as a function of environmental conditions and reaction time. In practice, it is difficult to understand catalysis without spectroscopy, especially at the molecular level. The initial spectroscopic catalyst characterization studies appeared in the literature approximately in 1950, and the total number of reported catalyst characterization studies increased exponentially with time reflecting the desire of catalysis researchers to better understand how their catalysts function. The fraction of catalysis papers using “characterization” is constantly increasing over the years (Fig. 1a). Today, most of the heterogeneous catalyst publications include characterization studies. Of the reported characterization studies for heterogeneous catalysis, more than 50% use in situ methodologies and nearly 10% use operando methodologies (Fig. 1b). A few years after the early characterization studies were reported, in situ catalyst characterization studies began to appear. The term in situ is Latin and means “on site” and is used to refer to catalyst characterization studies under controlled environmental conditions (e.g., ultrahigh vacuum, reducing, oxidizing, dehydrated, molecular adsorption or reaction). Although the number of in situ catalyst characterization studies lags those of all catalyst characterization publications, more than 50% of all current catalyst characterization publications have incorporated in situ studies (Fig. 1b). The in situ characterization studies revealed that heterogeneous catalysts, especially the surface phases, are dynamic with respect to reaction environments and reaction time. This realization has recently prompted simultaneous catalyst characterization and performance measurements under reaction conditions. Such catalyst characterization studies are termed operando, from the Latin meaning “operating” [2, 3]. Having both catalyst characterization and performance data at the same time allows for directly establishing structure-performance relationships from the same location and at the same time, which minimizes variability when compared to independent measurements of structure and performance in different instruments and possibly reaction conditions. Although the term vii

viii

Editors’ Introduction

a)

b) 8%

6000 5000

60% 50%

6% 4000

40%

3000

4%

2000

30% 20%

2% 1000

10%

0 1945

1965

1985

charac

in situ

operando

2005

0% 2025 %CHARAC

0% 1945

1965 in situ ratio

1985

2005

2025

operando ratio

Fig. 1 (a) Number of papers with the term catalyst and “characterization” (orange); among these, those using “in situ” (green); and among those, those using operando (yellow); right axis, percentage of catalysis papers in all catalysis fields using the term “characterization.” (b) Percentage of papers using the term “in situ” in papers using “catalyst” and “characterization” (orange) and using the term operando in papers using “catalyst” and “characterization” (green). (Source: Scopus (July 9, 2022))

operando was just introduced into the catalysis literature in 2003, the number of operando studies has increased exponentially with time and now represents 10% of total characterization studies (Fig. 1b). The importance of in situ and operando characterization studies in heterogeneous catalysis research is reflected by the fact that they now account for ~60% of all catalyst characterization studies, and such studies will only increase with time. We would like to point out that the term in operando has crept into the heterogeneous catalysis literature by analogy with the Latin term “in situ,” but this is an incorrect Latin expression. In operando means “on working (conditions),” which assumes that the catalyst is working, but since the catalyst is already working, the addition of the term “in” is incorrect. Operando is the correct term to use when performing simultaneous characterization and performance measurements of working catalysts. Analysis of Application of Characterization Techniques in the Heterogeneous Catalysis Literature Catalysts are dynamic materials that can change with environmental conditions (temperature, pressure, gas/liquid composition, and reaction time). It is, thus, critical that the environmental conditions employed for measurements be specified to avoid confusion between the different states of the catalysts. Unfortunately, many literature studies can be found that do not indicate the measurement conditions. This is problematic since surfaces are usually hydrated and oxidized under ambient conditions and change upon heating in different environments. Under ambient conditions, the surfaces are solvated and not representative of the state of the surface under reaction conditions where solvation may be absent and the surface is populated by adsorbed reactants, reaction intermediates, and products. Furthermore, the incident radiation can also cause changes in the catalysts under certain conditions and corrupt the data. To avoid such confusion, the section below will examine the influence of environmental conditions and radiation sources on the state of the catalysts. Another confusion that exists in the catalysis literature is the assignments of the collected spectral information. Where possible, references will be provided that will assist in the spectral assignments and, thus, lead to reliable catalyst structure-activity/selectivity relationships.

Editors’ Introduction

ix

Infrared (IR) and Sum-Frequency Generation (SFG) Spectroscopies Infrared spectroscopy (IR) provides the vibrations of the catalyst skeletal region and is primarily employed for examination of surfaces of solid catalysts to obtain molecular level information about surface hydroxyls, surface M¼O bonds, adsorbed molecules, and reaction intermediates. As indicated above, the surface functionalities monitored by IR are strongly dependent on the environmental conditions. Experimental conditions do have an effect on the samples that infrared spectroscopy readily detects. Infrared can clearly identify the nature of surface species formed during reactions and it will deliver relevant information if the experiments properly reflect relevant reaction conditions. During in situ and operando IR spectroscopy studies under reaction conditions, it is important to distinguish between the vibrations of both adsorbed species and gas phase molecules because IR can give rise to strong vibrations of gas phase molecules. This is a great challenge, and several techniques may help discriminate between the gaseous and adsorbed molecules. For instance, reflection absorption infrared spectroscopy (RAIRS) can probe the structure and surface chemistry on model catalysts, taking advantage that only the polarization component of IR light that is parallel to the plane of incidence interacts with the molecules on the surface. These aspects and more are covered in the chapters by Busca, Negri et al., and Trenani and Ranjan. Furthermore, modulation-excitation spectroscopy methodologies, as described in the chapters by Urakawa et al., Ferri et al., and Collins et al., can be applied with polarized excitation to discriminate between gaseous and adsorbed species and their transient states during reaction. A step forward to investigate vibrational spectra of molecules located at surfaces and interfaces is infrared-visible sum frequency generation (SFG), described in the chapter by Pramhaas and Rupprechter, which is a nonlinear optical method that can identify such molecular structures with sub-picosecond time resolution and is applicable at nearly any interface that is optically accessible. Raman Spectroscopy Raman spectroscopy is a very versatile optical spectroscopy to understand the working catalyst structure and surface chemistry in heterogeneous catalysis. Its extreme experimental versatility teams up with a growing number of instruments that continuously increases the characterization information. Raman experiments can be carried out under wide reaction conditions without any limitations of the sample phase. Raman is not only sensitive to crystalline phases, but it is a molecular spectroscopy with sensitivity to detect vibrations from amorphous surface metal oxide sites on oxide supports and small crystalline nanoparticles ( E > 0.0496 eV, medium IR) and 400–10 cm
1 (0.0496 > E > 1.24 10
3 eV, far-infrared, FIR), respectively. The simplest way to cause vibrational excitation is to allow the chemical species to absorb a quantum of energy from electromagnetic radiation of an appropriate energy, which gives rise to a vibrational transition from the ground state to an excited vibrational state. When electromagnetic radiation with intensity I0 is incident on a sample, the light may be absorbed, reflected, or transmitted. From the conservation of energy, it follows: I0 ¼ IR þ IT þ IA

ð1:1Þ

where I0, IR, IT, and IA are the intensities of the incident, reflected, transmitted, and absorbed radiation, respectively. By dividing Eq. (1.1) by I0, we obtain RþTþA¼1

ð1:2Þ

where R ¼ IR/I0, T ¼ IT/I0, and A ¼ IA/I0 are the apparent reflectance, transmittance, and absorptance, respectively, while

A ¼
log 10 I T =I 0

ð1:3Þ

is the absorbance. All these quantities depend, obviously, on the particular radiation used, specifically on its energy E, its frequency v, its wavelength λ, or its wave number v : E ¼ hv ¼ hc=λ ¼ hc v

ð1:4Þ

where h is Plank’s constant and c is the velocity of light. Analysis of the quanta that are absorbed by a polyatomic chemical species and those that are not absorbed (primarily being transmitted) gives information on the vibrational structure of these species and, consequently, on the chemical and geometric structure. This is the transmission/absorption IR technique. However, selection rules apply to this phenomenon. They are simplified as follows: Δv ¼ 1   @μ 6¼ 0 @Q 0

ð1:5Þ ð1:6Þ

The absorbance at a particular wavelength also depends upon the population of the corresponding vibrational states. According to Boltzmann’s law: N i =N 0 ¼ gi =g0 e


ΔE=kT

ð1:7Þ

where Ni’s are the state populations, gi’s are the state multiplicities, ΔE is the energy difference between the states 0 and i, and k is the Boltzmann constant. ΔEs for fundamental vibrational transitions fall in the IR region, so that the population of the first excited states is very small at room temperature and much lower than the population of the ground state even at 1000 K. This means that only transitions originating from the ground state can be ordinarily excited at room or relatively low temperature. Equation (1.5), where v is the vibrational quantum number, means that only transitions between nearest vibrational states can directly occur in the case of the harmonic oscillator. This means that the IR spectrum is generally composed of fundamental transitions, i.e., those associated with the excitation from the fundamental state to the first excited state. This condition (1.5), however, is relaxed in the case of anharmonic oscillators, so that not only fundamentals but also overtone and combination modes (also called the “harmonics,” those associated with the excitation from the fundamental state to a second or third excited state) can sometimes be observed, although they are usually weak. Equation (1.6) indicates that only vibrational modes that are associated with a change in the dipole moment μ of the molecule between non-zero extremes (the “polar” modes) of the atomic displacements (Q is the normal coordinate) can be directly excited.

1

Infrared (IR) Spectroscopy

5

In the solid state, the polar phonons (those that are IR active) split into two components, the transverse mode (TO) and the longitudinal mode (LO). This TO/LO splitting occurs because the electric field associated with the transverse wave is ¼ 0 while that associated with the longitudinal wave is 6¼ 0. When these modes couple with the electric fields associated with the vibration, vLO > vTO. This factor is relevant in relation to the shape and interpretation of the IR spectra of solid materials and will be further considered below. According to the Lambert-Beer law: A ¼ εcl

ð1:8Þ

the absorbance linearly depends on the molar concentration of the absorbing species c, the molar absorption coefficient ε, and the sample path-length l. This relation (rigorously valid only in non-scattering media) is the basis for quantitative analysis performed with IR spectroscopy. As is well known, the selection rules for the Raman scattering spectroscopy, which is the most common alternative vibrational technique, may be simplified as follows: Δv ¼ 1   @μ 6¼ 0 @Q 0

ð1:9Þ ð1:10Þ

This means that in Raman scattering spectroscopy, only transitions between nearest levels and associated with changes in the polarizability α upon motion are allowed for a harmonic oscillator. The first selection rule is relaxed for anharmonic oscillators. Group theory shows that for centrosymmetric chemical species, Raman-active modes are infrared inactive and viceversa. This is the so-called mutual exclusion rule. Other modes can be both IR and Raman inactive. For non-centrosymmetric molecular species, modes also occur that are both IR and Raman active. Thus, in most cases, Raman spectroscopy is complementary with respect to infrared spectroscopy. It may occur that chemical species are vibrationally silent, such as monoatomic noble gases, or IR inactive such as homo-diatomic molecules (H2, N2, O2, F2, Cl2,. . .). The energy and the frequency of the molecular vibrational modes depend on the weight of the atoms directly involved as well as on the bond energies. Consequently, the position of the absorption peaks is shifted when isotopically labeled molecules are used. This is a useful property to confirm the assignments of vibrational modes when needed.

1.3

Experimental Techniques

Infrared spectroscopy in the heterogeneous catalysis field is mostly performed today with either the transmission/ absorption technique or with the diffuse reflectance

technique. In the transmission/absorption technique, the IR beam is focused on a sample part of the radiation (the “transmitted” light) passing through the sample, and some of the photons are absorbed by the solid and excite its vibrational modes. In the diffuse reflectance technique, part of the light is scattered from the powder, while the rest is absorbed by the solid causing excitation of its vibrational modes. With modern instruments, the spectra obtained by the two techniques are fundamentally the same. Other experimental methods have been demonstrated, such as emission, reflection, multiple attenuated reflection, photothermal, and photoacoustic techniques. However, these techniques are rarely applied to catalysts [12, 13]. In this review, we will not describe these different experimental techniques. Additional information can be found in more specialized publications as well as in refs [11, 12] and the papers cited therein.

1.4

The Bulk Characterization of Solid Catalysts by Infrared Spectroscopy

Infrared spectroscopy is a very useful technique for the characterization of solids. Nonconducting solids usually absorb photons in the IR range with the corresponding excitation of vibrational modes. Conducting solids usually absorb the totality of the infrared radiation with the corresponding excitation of electronic energy states, while semiconducting materials can absorb either the entire range of IR radiation or single wavelengths with the excitation of electronic states either in bands or isolated states. The most common practical technique to obtain a good spectrum of the fundamental vibrations of a powdered insulating material with the transmission/absorption IR technique is to prepare a thin disc of appropriately diluted material. To do this, the most common technique is that of KBr-pressed discs [14]. KBr is an easily available powdered material which does not absorb in the medium IR region (MIR, down to near 400 cm
1, it only cuts out the far IR, FIR). It can be easily mixed homogeneously with the powder to be investigated and pressed, thus obtaining “diluted” self-supporting discs that are very useful for IR transmission. It does not usually react with the powders under study. Other materials (such as CsI or polyethylene for far IR studies) allow the production of similar pressed discs with cut-off limits at even lower frequencies in the far IR region. Note that IR absorptions of solids usually occur in both MIR and FIR regions. Alternatively, the powders can be simply deposited in the form of a thin layer on a disc of a transmitting medium. Skeletal IR spectra can also be obtained in the diffuse reflectance mode, usually by diluting the powder with the powder of a nonabsorbing material.

1

6

G. Busca

1.4.1

IR Absorption Spectra of Crystalline Nonconducting Solids

The application of skeletal IR spectroscopy to the structural characterization of crystalline nonconducting solids (such as heterogeneous catalytic materials) [15, 16], if coupled with the application of Raman scattering spectroscopy, may allow complete definition of the crystal structure. The analysis of the symmetry elements of the crystal structure (the space group and factor group [17] of the smallest Bravais cell) according to the site symmetry of every atom allows the determination of the irreducible representation of the total modes. After the subtraction of the “hindered translational modes” (i.e., the “acoustic modes”), the “irreducible representation” of the vibrational (or “optical”) modes can be obtained. This means that the number of the vibrational modes belonging to the symmetry species associated with the molecular or crystal symmetry can be counted. Consequently, the number of the active vibrational modes can be counted, according to the symmetry selection rules. Vibrational modes are active in infrared spectroscopy (i.e., they correspond to absorption of the corresponding photon) only when the excitation of the vibrational mode results in modification of the dipolar moment of the molecule/species. Otherwise, if the excitation of the vibrational mode does not result in modification of the dipolar moment of the species but results in a change of its polarizability, the mode is active in Raman scattering spectroscopy but inactive in IR absorption spectroscopy. Finally, fully inactive modes can occur

a)

(neither polarity nor polarizability are modified during the vibrational transition), or for asymmetric molecules, modes active in both techniques can occur. Additionally, the position (in terms of frequency or wave number of the absorbed photon) of the vibrational mode can be forecast from the nature of the elements and the bond with which they are held. Thus, the theoretical vibrational spectra can be forecast and calculated [18, 19] for a crystalline solid structure and compared with the recorded experimental infrared absorption and Raman scattering spectra. In Fig. 1.1, the FTIR and the FT Raman spectra of the solid compounds NiTiO3 and CoTiO3 (nickel and cobalt metatitanate) are reported. Several structures may correspond to the ABO3 stoichiometry (see Table 1.1) and can be distinguished by vibrational spectroscopies. The spectra clearly show that the two solids are isostructural [20]. The number of peaks observed in both infrared and Raman spectra in Fig. 1.1 excludes the cubic perovskite structure and agree with the ilmenite structure, that was also determined by XRD spectrometry. The recording of skeletal IR spectra is a reliable and fast technique to confirm the dominant crystal phase of a crystalline sample. This can be applied, for example, in the case of aluminas, which are important catalytic materials [21, 22]. During catalyst preparation and use, alumina can either rehydrate to form the hydroxide or thermally convert to different phases. In Fig. 1.2, the skeletal IR spectra of alumina after different treatments are reported. The presence of phases like boehmite (the most common oxyhydroxide form γ-AlOOH),

b)

b)

Intensity

Absorbance

b)

a)

a)

1200

1000

800

600

Wavenumber

400

200

800

(cm–1)

Fig. 1.1 Infrared spectra (a) and Raman spectra (b) of NiTiO3 (a) and CoTiO3 (b)

700

600

500

400

Wavenumber (cm–1)

300

200

1

Infrared (IR) Spectroscopy

7

Table 1.1 Crystal structures of some solid ternary-oxide structures with the stoichiometry ABO3 and the irreducible representations of the optical modes Structure name Ilmenite

Example FeTiO3

LiNbO3 Cubic perovskite Tetragonal perovskite Orthorombic perovskite Rhombohedral perovskite

3305

C3i2

LiNbO3 SrTiO3 BaTiO3 LaFeO3

R3 R3c Pm3m P4mm Pnma

Z 6

C3v6 Oh1 C4v1 D2h16

6 1 1 4

LaMnO3

R 3c

D3d6

2

S. G.

4A1 (IR) þ 5A2 (R) +9E(IR,R) 3F1u(IR) þ F2u(in) 3A1(IR,R) þ 4 E(IR,R) þ B1(R) 7Ag(R) þ 7B1g(R) þ 5B2g(R) þ 5B3g(R) þ 8Au(in) þ 7B1u (IR) þ 9B2u(IR) þ 9B3u(IR) A1g(R) þ 3A2g(in) þ 4Eg(R) þ 2A1u(in) þ 2A2u(IR) þ 4Eu(IR)

3090 630

3600

Irreducible representation 5Ag (R) þ 5Eg(R) þ 4Au(IR) þ 4Eu(IR)

731

3200

490

50% NiO/Al2O3 0.1

1074

G-Al2O3

Absorbance

650 °C J-Al2O3

800 °C J/G-Al2O3

Absorbance

100 °C J-AlOOH

1170

NiAl2O4 NiO

950 °C T-Al2O3 1000 °C D/T–Al2O3 1100 °C D-Al2O3

1300

1100

900

700

1400

1200

1000 800 600 Wavenumbers (cm–1)

500

Wavenumbers (cm–1)

Fig. 1.2 Infrared spectra of boehmite and aluminas

of different transitional forms such as γ-Al2O3, δ-Al2O3, and θ-Al2O3, of the thermodynamically most stable form α-Al2O3 (corundum), and of their mixtures can be obtained by IR. Specifically, in the case of the oxyhydroxide boehmite, the bands at 3305 and 3090 cm
1 are due to the stretching modes of structural O-H bonds, while those at 1170, 1074 cm
1 are due to the corresponding in plane deformation modes. Obviously, these features are not present in the spectra of Al2O3 polymorphs. The spectra of oxide structures in the MIR region are dominated by the vibrations of oxide anions. Bands in the 1000–700 cm
1 region are due to Al-O bonds involving mainly tetrahedral AlO4 environments in spinel-type structures, while the features at lower frequencies mainly involve AlO6 structures present both in spineland corundum-type lattices. IR spectroscopy also allows the analysis of multiphasic crystalline materials. Figure 1.3 shows the spectrum of a 50% NiO/Al2O3 catalyst prepared by impregnation of δ-Al2O3

Fig. 1.3 Infrared skeletal spectra of catalytic materials in the NiO-Al2O3 system

containing 5% SiO2 using an aqueous solution of Ni nitrate, followed by drying and calcination. The spectrum shows that the NiO phase is present together with the δ-Al2O3 support, while the bulk NiAl2O4 spinel phase was not formed. In the solid state, the polar phonons (those that are IR-active) split into two components, the transverse optical mode (TO) and the longitudinal optical mode (LO) [23]. This TO/LO splitting occurs because the electric field associated with the transverse wave is ¼ 0, while that associated with the longitudinal wave is 6¼ 0. Coupling of these modes with the electric fields associated with the vibration gives rise to vLO > vTO. This consideration is particularly relevant when the IR reflection spectra of monocrystals are considered and provides information on particle shape distribution of nanoparticles. This is evident in the spectrum of NiO in Fig. 1.3. NiO (bunsenite) belongs to the rock salt family (Fm3m space group, Oh5, n. 225) with Z ¼ 4. With only one formula unit in the smallest Bravais cell, the irreducible representation of the vibrational modes is

1

8

G. Busca

Γopt ¼ F1u ðIRÞ i.e., only a single, triply degenerate, fundamental vibrational mode exists, which is active in IR. Thus, only one IR band is expected. The observed spectrum (Fig. 1.3) shows two vTO components at 467 and 425 cm
1, likely due to two dominant particle morphologies and vLO at ca 565 cm
1.

1.4.2

3800 3600 Kaolin

According to the microscopic nature of the absorption of the IR photons, amorphous solids (also like liquids and gases, i.e., fully disordered phases) also absorb infrared radiation. This is particularly interesting in the field of catalysis, where amorphous materials may be used, as well as in the field of glasses, because in these cases, diffraction techniques are not applicable. Thus, vibrational spectroscopies, and in particular infrared spectroscopy, play an important role in the characterization of these solids [24]. The vibrational patterns of amorphous solids are associated with the elemental vibrations of the fundamental units of the structure and are frequently simpler than those of crystalline structures where multiple couplings between repeated units in the smallest Bravais cell may occur and produce band splitting. Peaks in IR spectra of amorphous solids are frequently broader than those of crystalline materials. This observation results from the larger dispersion of slightly different environments of the vibrating units in the amorphous disordered solids than in the crystalline ones, in which they are virtually all identical. In Fig. 1.4, the IR spectrum of amorphous silica is compared with that of α-quartz. The main bands in the spectra of any silica belong to the vibrational modes of the oxygen atoms in the structural units of the SiO4 tetrahedra and the Si-O-Si bent bridges. These bands are the Si-O-Si asymmetric stretching (expected at 1300–1000 cm
1), Si-O-Si symmetric stretching/ in plane bending (around 800 cm
1), and out of plane bending or rocking mode (around 450 cm
1) [25]. These modes may be multiplets due to the couplings of the movements of the four different oxygen atoms. As for the spectrum of α-quartz, S.G. P3221 with Z ¼ 3, the irreducible representation is Γopt ¼ 4A1 ðRÞ þ 4A2 ðIRÞ þ 8EðR, IRÞ Thus, 12 IR modes are expected and are evident in the spectrum. In Fig. 1.4, the spectra of kaolin and metakaolin are also reported. Kaolin is composed of kaolinite, Al2Si2O5(OH)4, which has a triclinic structure, space group P1 with Z ¼ 2. The IR spectrum is quite complex, with the appearance of many bands corresponding to the 47 IR active modes expected [26].

G,T-Al2O3

Metakaolin

IR Absorption Spectra of Amorphous Solids

D-quartz Amorphous silica (aerosil ) 1600 1400

1200

1000

800

Wavenumbers

600

400

200

(cm–1)

Fig. 1.4 Infrared spectra of materials in the silica-alumina system

Four of these modes are associated with (SiO)-H stretching modes and are observed in the region 3800–3600 cm
1. After thermal treatment at 500–800 K, kaolinite decomposes to metakaolin, an amorphous porous material. The IR spectrum of metakaolin shows the features of amorphous silica and of a mixed transitional alumina phase, with the corresponding disappearance of the bands due to the hydroxy groups. Figure 1.5 presents the spectrum of a catalyst obtained by impregnating amorphous silica with cobalt acetate followed by calcination. The spectrum is clearly the superposition of that of the amorphous silica support with that of crystalline Co3O4 nanoparticles [27]. However, the differences in the region 1000–900 cm
1 show that some cobalt ions have entered the silica framework.

1.4.3

IR Characterization of Spent Catalysts

Infrared spectroscopy is also frequently applied to spent catalysts after they have been used in laboratory plants or even in real industrial reactors. This is very useful to investigate the deactivation phenomena frequently occurring under operating conditions. Fluid catalytic cracking is a refinery process that produces high-octane gasoline and light olefinic and paraffinic gases from heavy oils. The catalyst is a complex mixture of several powders extruded in small balls which are fluidized and transported cyclically from the riser reactor, through a disengager to the regenerator, where the coke formed during the reaction is burned off. The active catalytic phase in this mixture is a zeolitic solid acid powder normally based on low-aluminum faujasite (Ultrastable Y

1

Infrared (IR) Spectroscopy

9

668

1097

584

zeolite), with added rare earth oxides, usually denoted as REUSY zeolite. Metakaolin is an additional significant component of this complex material, together with some alumina and binders based on silica [28]. The IR spectrum of the fresh catalyst (Fig. 1.6) is dominated by that of the faujasite zeolite but is clearly broader and less defined due to the additional presence of amorphous metakaolin and silica. The latter are responsible for the broad bands in the spectrum. Upon time on stream, the catalyst progressively loses activity. As it is evident from Fig. 1.6, one of the reasons for the loss of activity is the amorphization of the zeolite phase, as

1097 1200

470

585

669

810 1000

Wavenumbers

805

977

468 1400

Infrared Detection of Impurities in Catalysts

IR spectroscopy is also very useful to reveal the presence of impurities in catalysts. Typical examples are carbonates and sulfates which are common impurities in some oxide materials. Also, catalysts prepared by sol-gel techniques can contain significant amounts of organic impurities, which are responsible for characteristic infrared bands. As an example, catalysts based on Zn-Mn oxides that were obtained by recycling spent alkaline and zinc-carbon batteries were found to contain significant residues of sulfate species that are responsible for a strong band around 1200 cm
1 [29].

1.4.5

1094

Absorbance

SiO2

1.4.4

Co3O4

0.2 Co3O4/SiO2

evidenced by the almost complete disappearance of the sharp peaks in the spectrum.

800

600

(cm–1)

Fig. 1.5 Infrared spectra of silica, cobalt oxide, and a Co3O4/SiO2 catalyst

Absorbance

E-cat (used catalyst)

Fresh catalyst

Application of Skeletal IR Spectroscopy in the Characterization of Unsupported and Supported Metal Nanoparticles

As previously discussed, IR spectroscopy cannot be applied to bulk metals due to their absorption of the entire range of IR radiation. However, IR spectroscopy can be applied to checking the purity of unsupported metal nanoparticles. In this case, the absorption of IR radiation by the metallic particles is incomplete, and the absorption due to nonconducting impurities can be observed. This is shown in the right side of Fig. 1.7, where the spectra of three differently prepared samples of unsupported metallic cobalt nanoparticles are compared. The absorption at 1378 cm
1 is due to trigonal borate species while that around 1000 cm
1 is associated with tetrahedral borates, both arising from sodium borohydride used as the reducing agent in the preparation procedures [30]. The pair of bands at 667 and 567 cm
1 is associated with traces of Co3O4, while the single band around 510 cm
1 is associated with the presence of CoO. On the other hand, in the case of supported metal nanoparticles, the IR spectrum on the left side of Fig. 1.7 shows the features of the support, amorphous silica. The weak feature at 668 cm
1 is attributed to cobalt silicate species [31] showing that some reaction occurred between the supported species and the support.

REUSY Zeolite with binder 1400

1200

1000

800

600

Wavenumbers (cm–1)

Fig. 1.6 IR skeletal spectra of a fresh industrial FCC catalyst (from Grace Gmbh), the original zeolite component, and of the same catalyst after use in industrial refinery FCC reactor

1.4.6

Revealing the State of Oxidation of Catalysts by Skeletal IR Spectroscopy

The Mars-van Krevelen or redox mechanism [32] occurs for many oxidation reactions over metal/metal oxide catalysts. This implies that the oxidized catalyst surface oxidizes the

1

10

G. Busca

a)

b) 1093

BO33–

Co3O4 667

Co/SiO2

1000

465

1370

BO45–

567

Co NP C Co NP B

1376

668

797

0.05

CoO 510

SiO2

Co NP A

1400

1200

1000

800

600

1800

1600

Wavenumbers (cm–1)

1400

1200

1000

800

600

Wavenumbers (cm–1)

Fig. 1.7 Infrared spectra of reduced Co/SiO2 catalyst, compared with that of the SiO2 support (a) and of three different samples of unsupported Co metal nanoparticles (CoNP) (b)

reactant and is reoxidized by gas-phase O2 in the following step. Bulk or subsurface atomic oxide species may be active in this mechanism. However, the real working oxidation couple is not always well established. As previously discussed, skeletal IR spectroscopy provides detailed information on the bulk metal oxide bonds. During bulk oxidation and reduction, the skeletal spectrum is modified. For example, the skeletal features of Co3O4 in the spectrum of a Co3O4/SiO2 catalyst (see Fig. 1.5) disappear upon total oxidation of 1,2-dichloropropane [26], showing that it is probably not the active phase in the catalytic act. They also disappear upon the use of the same catalyst in CO2 hydrogenation to methane (compare Figs. 1.5 and 1.7a) showing that in this case, the active phase is metallic cobalt.

1.5

Surface Characterization of Catalysts by IR Spectroscopy

IR spectroscopy can be applied to the surface chemistry of catalytic materials using the pure powders, frequently after activation by outgassing or under controlled atmospheres. In the transmission/absorption technique, self-supporting pressed discs of the pure oxide powders are prepared and put into the IR beam in an appropriate cell that allows heating, cooling, and gas/vapor manipulation. In the case of DRIFT experiments, the catalyst is deposited in the sample holder, and activation is typically performed by flowing inert dry gas.

1.5.1

The Infrared Spectra of Pure Catalyst Powders

Line-Base Slope and Light Scattering The transmission/absorption IR spectrum of a disc of a nonconducting fine powder has a baseline slope that is flat or increasing toward higher frequency (Fig. 1.8). This observation reflects the elastic scattering of the beam, which increases with increasing frequency (wave number). When radiation is incident on a nonabsorbing layer consisting of N particles per unit volume, its intensity decreases exponentially with the sample thickness x following the law: IT ¼ I0 e


φx

¼ I0 e


N σx

ð1:11Þ

so that an apparent absorbance A′ can be defined for an absorbing and scattering medium: 0

A ¼
log 10 I T =I 0 ¼ 0:43 N σx   0 0 ¼ 0:43 N ε þ s x ¼ A þ 0:43 Nsx ¼ A þ S ð1:12Þ where φ is the linear attenuation coefficient and σ is the linear attenuation coefficient of one particle. σ contains a component due to absorption, ε′, and a component due to scattering, s. Consequently, the apparent absorbance is composed of the true absorbance A superimposed upon the component due to the light scattering S′. This simplified approach is similar to

1

Infrared (IR) Spectroscopy

11

a)

b)

1

TiO2 (from sulphate process)

Overtones

Absorbance

Amorphous SiO2

J-Al2O3 mZrO2 4000

3500

3000

MgO-Al2O3 (calcined hydrotalcite) 2500

2000

1500

1000

4000

3500

3000

2500

2000

1500

1000

Wavenumbers (cm–1)

Fig. 1.8 Infrared spectra of pure powder-pressed discs of catalytic materials after outgassing at 773 K

that given by Henry [33] who proposed the following expression for the component dependent on scattering: 0

2

S ¼ k  d ðn
1Þ =λ

2

ð1:13Þ

where k is a constant independent of λ but dependent on the particle arrangement, d is the particle diameter and n is the refractive index of the material. On the other hand, according to the Rayleigh theory, S′ depends on the third power of the particle diameter (and so linearly on the particle volume) and on the fourth power of the radiation frequency v. The effect of scattering on the IR transmission spectra of powder samples has been studied by Duyckaerts [13, 34] who showed that if the particle diameter is sufficiently small (considerably smaller than the IR wavelength), the scattering of the IR radiation is very small. In this instance, the apparent absorbance of a layer of a powder depends on its own absorptivity, on the disc thickness and density, on the particle size, and on the wavelength. The slope depends on the particle size and may be very steep for powders which consist of quite large particles (near 1 μm). Thus, transmission for large particle-size powders may be nil at 4000 cm
1 or even lower frequencies. In the case of DRIFT, the reflectance actually increases as scattering increases, and the baseline is flat or even decreasing with increasing wave number.

The Cutoff At low frequencies, below 1300–600 cm
1 depending on the skeletal spectrum of the solid, the absorptions due to bulk skeletal fundamental vibrations cut off the spectra of the pure powder discs. The position of the cutoff in the spectra of pure powder-pressed discs depends on the position of the higher frequency side of the bulk spectrum. Samples containing covalent bonds, such as silica and silicates (strongly absorbing in the region 1200–1000 cm
1) have a cutoff limit at high frequencies (e.g., 1300 cm
1) but may have transmittance windows at lower frequencies. Typical metal oxides have a cutoff limit depending on the atomic weight of the metal atom, e.g., near 1300 cm
1 for silica (with windows in the lower frequency range, see left side of Fig. 1.8), near 1000 cm
1 for alumina and titania, near 800 cm
1 for zirconia (Fig. 1.8) and ceria, and 600 cm
1 for thoria. Thus, the position of the cutoff limit can also be indicative of the composition of the catalyst. The Bulk Vibration Overtones In the mid-infrared region, a series of moderately weak bands can be observed which are not perturbed by outgassing and/or adsorption of molecular probes. These absorptions are associated with overtone or combination modes of lattice adsorption. In particular, the spectra of highly siliceous materials, such as those based on amorphous silica or crystalline

12

silicates, including zeolites, present a triplet of broad bands around 2000 cm
1, 1880 cm
1, and 1640 cm
1, which are combinations of bulk Si-O-Si modes (Fig. 1.8a, top). A number of other solids that contain covalent structural units also give rise to characteristic absorptions in the region 2500–1500 cm
1 due to overtones of bulk vibrational modes. This is the case of covalent oxides, such as V2O5 [35], MoO3, and WO3 [36], and of metal oxosalts, such as metal phosphates (such as (VO)2P2O7, the active phase of catalysts for n-butane oxidation to maleic anhydride [37]), metal vanadates (such as Mg vanadates for oxidative dehydrogenation reactions [38]), molybdates, tungstates, and heteropolyacid. The presence of these absorptions can be used as diagnostic tools to reveal the presence of phases, such as bulk MoO3 in the case of heavily loaded MoO3/ TiO2 supported oxides [39].

Spectra of Surface or Bulk Impurities The IR spectra of some samples may contain bands in the 2200–1200 cm
1 region which can be associated with the presence of impurities. As an example, titania (anatase) produced from ilmenite with the “sulfate” process usually contains surface sulfate impurities that, after surface dehydration by outgassing, give rise to a sharp band at ca. 1370 cm
1 due to the S¼O band of sulfates (Fig. 1.8b, top). Basic oxides are frequently contaminated, at the surface or even in the bulk, by carbonate species. In Fig. 1.8b lower , the spectrum of a mixed oxide containing MgO and Al2O3 (obtained by thermal decomposition of hydrotalcite Mg6Al2(OH)16(CO3).4H2O, a layered double hydroxide-carbonate), is reported and shows a strong absorption in the region 1700–1200 cm
1 due to residual carbonates. The IR Spectra of the Surface Hydroxyl Groups The fragments arising from the dissociative adsorption of water on the surface of metal oxides give rise to hydroxyl groups that are potentially active Brønsted acid or base sites. The vibrations of hydroxyl groups are typically composed of O-H stretching (3800–2000 cm
1, depending on the extent of H bonding), in-plane bending (1200–800 cm
1), and out-ofplane deformation (generally below 1000 cm
1). In most cases, only stretching modes can be observed in the spectra of pure catalyst-pressed discs. They can be detected in the 3800–3000 cm
1 region by recording the IR spectra of the oxide catalyst powders after desorption of molecularly adsorbed water (Fig. 1.8). Confirmation of the assignment of these bands to O-H stretching can be obtained by deuteration, i.e., isotopic exchange with gaseous deuterium 2H2 [40] or deuterated water 2H2O [41], hydrogen-deuterium exchange results in the formation of surface deuteroxyls O-D, with well-shifted stretching modes (vOH/vOD ~1.355) and usually retaining the same absorption shapes.

G. Busca

The position and shape of the vOH bands of the surface hydroxyl groups are informative about their coordination (Table 1.2). Covalent oxide components [42, 43] usually give rise to very typical strong sharp peaks due to covalently bonded terminal OHs. Conversely, bridging and triply bridging surface hydroxyl groups are also formed on ionic oxides. Thus, the spectrum of surface hydroxyl groups is frequently more complex for ionic than for covalent oxides. It is generally agreed that the OH stretching frequency is the highest for terminal OHs, intermediate for bridging OHs, and the lowest for triply bridging OHs. This approach was first systematically proposed by Tsyganenko and Filimonov [44]. The Surface Hydroxyl Groups of Aluminas Transitional aluminas, in particular the defective spinel phase γ-Al2O3, find wide application as a Lewis acid catalyst, as a catalyst support and as an adsorbent [20, 21]. Many studies have been devoted to the multiplicity of the surface hydroxyl groups of aluminas. At least five components are usually present in the IR spectrum of the hydroxyl groups of aluminas, at ca 3790, 3770, 3740–3720, 3700–3690, and 3580 cm
1. In many cases, multiplets may be observed. A number of different assignments have been proposed for these bands. Attempts have been made by authors to assign the different bands to structures located on different planes exposed at the surface of alumina nanocrystals. In Fig. 1.9a, the IR spectra of two samples obtained from the same precursors with a similar preparation, but having significantly different porous structures, are reported. Specifically, the sample Puralox SBA (SBET 198 m2/g, VP 0.45 cm3/g, dP 90 Å) has far smaller pore volume and larger pore sizes than PuraloxTH (SBET 166 m2/g, VP 1.20, cm3/g, dP 188 Å). Both samples are proposed to be composed of laminar/ prismatic nanoparticles that differ in their thickness. This means that the two samples should have different ratios between the extents of the basal plane (101 spinel plane) and of the “vertical” planes (100 spinel plane arising from the basal plane of boehmite platelets). Interestingly, the two spectra are identical that suggests the corresponding OH groups are located on defects or on edges more than on the different extended planes. The analysis of the OH groups of aluminas can also be indicative of the purity of the alumina surface. Impurity levels of silica have been shown to be responsible for increasing intensity of the band around 3725 cm
1 which is attributed at least in part to surface silanol groups on alumina [45], Fig. 1.9b, bottom. On the other hand, the presence of sodium impurities (quite common in samples coming from sodium aluminate and Bayer alumina) generally results in a decrease or the disappearance of the bands at higher frequencies (3790 and 3770 cm
1), a shift of the medium band (from 3726 to

1

Infrared (IR) Spectroscopy

13

3747 cm
1) and an increase in intensity of the band at 3680 cm
1 [46] (Fig. 1.9b, top). The presence of chlorine atoms in samples produced by flame hydrolysis of AlCl3 can be responsible for a decrease in intensity of the bands above 3700 cm
1 [47].

As it will be discussed later on, surface hydroxyl groups may display Brønsted acidity (the ability to protonate sufficiently strong bases), hydrogen-bonding activity allowing adsorption of even weakly basic molecules, or also basic or nucleophilic behavior, functioning as hydroxide anions to

Table 1.2 Position of typical OH stretching bands on dehydrated surfaces of metal oxides Position cm
1 and tentative assignment Terminal 3790, 3770, 3740–3720 3680 3710–3690 3685–3670 3700–3600

3710 3735–3710 3725 3730–3680 3645 3700 3690

3726

3747

Puralox TH

3505

Home made Na rich

3687

3718 3786

3726

3685 3670

3771

0.5

3790

3580

3670 3690–3660 3640 3662–3640 3650

b)

1.0

Triply bridging

3688, 3670 3680 3640 3675, 3640 3620

3738 3680, 3660 3740, 3725 3650 3745–3 3682 3680 3680 3770

a)

Absorbance

Bridging 3748–3730 3700–3690

3681

Silica (SiO2)/silicated oxides Al2O3 Phosphated oxides Borated oxides Germania (GeO2) Sulfated oxides α-Ga2O3 β-Ga2O3 α-Fe2O3 γ-Fe2O3 α-Cr2O3 MgO MnO CoO NiO Monoclinic zirconia Tetragonal zirconia Ceria CeO2 TiO2 anatase TiO2 rutile Lanthana and La-doped oxides V2O5 Nb2O5  n H2O WO3

3675

Material

Degussa (Cl– rich) Puralox TH

Puralox SBA 4000

3800

3600

3400

3200

3900

3800

3700

Wavenumbers (cm–1)

Fig. 1.9 Infrared spectra of alumina samples in the OH stretching region, all activated at 773 K under outgassing

3600

3500

1

3655

G. Busca

3600

14

3743

3662

Niobium phosphate

Phosphated silica

Amorphous silica-alumina

3561

USY zeolite

3602

3743

3626

3740

Absorbance

3665

Tungstated zirconia

H-MOR zeolite 4000

3800

3600 3400 Wavenumbers (cm–1)

3200

Fig. 1.10 Infrared spectra of Brønsted acidic catalyst samples in the OH stretching region, all activated at 773 K under outgassing

interact with nucleophilic/acidic species. In Fig. 1.10, the spectra of catalysts that display Brønsted acidity (they are able to protonate the basic probe molecule pyridine, see below) and, correspondingly, high catalytic activity in Brønsted acid catalysis are shown. Although surface hydroxyl groups characterized by strong Brønsted acidity tend to absorb at lower frequency than those that are not acidic, the OH stretching frequency is not simply correlated with the acidic/ basic character of the hydroxyl groups. Thus, the adsorption of molecular probes is needed to characterize the chemical behavior of hydroxyl groups, as will be discussed below.

Absorptions Due to Surface Metal-Oxygen “Double” Bonds Metals and other elements in very high oxidation states can give rise to element-oxygen double bonds in their oxides, very short M¼O bonds that are characterized by very high stretching frequencies and also detected with Raman spectroscopy [48]. Usually these bonds are also responsible for overtone bands in the IR spectra [49]. This is the case in vanadyl, niobyl, molybdenyl, chromyl, and wolframyl groups as well as of P¼O and S¼O bonds present in sulfates and phosphates. Several catalysts or catalyst precursors on common oxide carriers contain such species on the surface. 16 18 O/ O exchange on the surface of oxides can be performed with 18O2 or with H218O. The position of M¼16O stretching is different from that of M¼18O stretching, and this isotopic shift confirms the presence of surface metal-oxygen double bonds [50]. In Fig. 1.11, the spectrum of a sample of zirconia containing both sulfates and wolframates is reported after

activation under vacuum at 773 K. The sharp band at 1385 cm
1 is due to surface sulfate species (S¼O stretching), while the band at 1008 cm
1 with a second component at 1014 cm
1 is due to surface wolframyl complexes (W¼O stretching). The species responsible for these bands are certainly located at the surface, because the frequencies observed are far higher than those of bulk sulfate and wolframate species, and also because the bands are perturbed by adsorption of different molecules such as water and pyridine. In Fig. 1.11, the spectrum of the same sample after adsorption of pyridine and following outgassing at 473 K is also reported. The band of the sulfate species is now broader and centered at 1335 cm
1, while that of wolframyl species is centered now at 990–980 cm
1. Wolframyl species are also responsible for the weaker absorptions observed at 2025 and 2012 cm
1 after activation in vacuum. These bands are the first overtone of the W¼O stretching vibrations and shift to 1994 cm
1 after adsorption of pyridine. It is not infrequent that the region below 1200–1000 cm
1 is not available in the IR spectra of oxides of light metals owing to the absorbance of bulk metal-oxygen stretching. Consequently, careful inspection of the region around 2000 cm
1 to look for first overtones of these M¼O bonds is often useful.

Absorptions by Surface Metal-Oxygen-Metal Bridges The adsorption of probe molecules may provide evidence of the existence of metal-oxygen bonds on the surface of metal oxides, whose stretching modes are located just at the limit of the cutoff. Upon adsorption of molecular probes, a broad “negative” band is observed in the subtraction spectrum that is attributed to the relaxation of surface metal-oxygen bonds. This is shown in Fig. 1.12 for the system methanol/iron oxide. After adsorption of methanol over α-Fe2O3, intense bands of adsorbed methanol appear at 2929, 2815 cm
1 (CH3 stretching) and at 1068 cm
1 with a shoulder at 1030 cm
1, (CO stretching of adsorbed methoxy groups and undissociated methanol, respectively). Simultaneously, the OH stretching bands of surface hydroxyl groups disappear because they are involved in H bonding with undissociated methanol species. Additionally, an absorption at 774 cm
1 also disappears, as shown in the subtraction spectrum. This absorption is completely restored when methanol is desorbed. It is attributed to surface Fe-O-Fe bond stretching. A similar absorption is found in several other cases, such as on aluminas [51], aluminates [52], gallia, and ferrite spinels.

1.5.2

The IR Spectra of Adsorbed Probe Molecules

Additional and exhaustive information on the surface chemistry of solids can be obtained by examining the adsorption of

Infrared (IR) Spectroscopy

15

Absorbance

2.6

2025

2036

2.8

981

990

990

2012

3.0

1

1008

1014 1008

3.2

1994

2012

1

2.4 1611 2.2

1046

0.5

1068 2.0

1486 1489

1.8

1385

1541 1.6

1226

1638 1335

1.4

2200

2000

1800

1600

1400

1200

1000

Wavenumbers (cm–1)

1030

2815

774

Subtraction result 3408

Absorbance

3670 3640

2929

1068

Fig. 1.11 Infrared spectra of a ZrO2 sample containing surface sulfate and tungstate species activated at 773 K under outgassing (solid line) and after adsorption of pyridine (broken line)

1.5

0.86

0.66

0.4 3800

3600

3400

3200

3000

2800

1300

1200

1100

1000

900

800

Wavenumbers (cm–1)

Fig. 1.12 Infrared spectra of α-Fe2O3 activated at 673 K under outgassing (solid line) and after adsorption of methanol at r.t and brief outgassing (broken line)

16

G. Busca

probe molecules on such surfaces. Surface probes test the features of the surface, such as acidity, basicity, oxidizing ability, and redox properties. A large number of molecules have been proposed for such studies; however, only a few of them find general application and are useful for standard catalyst characterization studies. In Table 1.3, some of the most common and most useful molecular probes are reported.

IR Spectra of Basic Probes for Surface Acidity Characterization Surface Acidity and Catalysis Acid catalysis occurs via Brønsted and Lewis acidity [53]. By definition, Brønsted acidity is the property of surface protonic sites to protonate bases and produce the corresponding conjugated acids (protonated bases). The extent of protonation depends on the basic strength of the basic probe and on the Brønsted acid strength of the surface site. Very strong Brønsted acids are able to protonate not only typical bases (n-type bases, those Table 1.3 Most used molecular probes for the characterization of the surface chemistry of catalytic materials Acidity Pyridine Hindered pyridines Nitriles Hindered nitriles Carbon monoxide Basicity/nucleophilicity Carbon dioxide Oxidizing ability Carbon monoxide Alcohols Metallic sites Carbon monoxide

Coordination, protonation, hydrogen bonding Location of acid sites in porous materials Coordination, hydrogen bonding Location of acid sites in porous materials “Polarization,” hydrogen bonding Exposed basic/nucleophilic sites Oxidizing cations/oxide couples Oxidizing cations/oxide couples Clusters, extended surfaces, defects

that use nonbonding electron pairs for acid-base interaction) but also very weak bases such as hydrocarbons, which act as π-type bases (olefins or aromatics in which C¼C π bonding orbitals are converted to carbenium or arenium ions)) or σ-type bases (paraffins, in which C-H or C-C σ bonding orbitals are converted to carbonium ions). The protonation of “basic” reactants is a way to initiate and promote catalytic reactions. Conversely, Lewis acidity is the property of a chemical species having empty orbitals, such as coordinatively unsaturated cations. For metal cations, the strength of the Lewis acid sites may be correlated to the “polarizing power” of the cations which can be roughly measured by the ratio of charge/diameter [49]. Lewis acids usually interact and “activate” n-bases, while Lewis acids typically do not interact with π or σ bases. Thus, Brønsted acid catalysts play a very significant role in the field of organic chemistry, including hydrocarbon chemistry, while Lewis acidity is mostly relevant in the chemistry of organic molecules containing heteroatoms (oxygen-, nitrogen-, sulfur-, and halide-containing organic compounds). Table 1.4 lists a number of solid acid catalysts and the industrial processes carried out with them. Characterization of Brønsted and Lewis Acidity Using Adsorbed Pyridine Starting with the pioneering work of Parry [5], pyridine is frequently used as a probe to characterize both Brønsted and Lewis acidity [54]. In Fig. 1.13, the spectra of pyridine adsorbed on a purely H-bonding solid (amorphous silica), on a purely Lewis acidic solid (γ-Al2O3), on a purely Brønsted acidic material (low-Al content ZSM-5 zeolite), and on a catalyst containing both Lewis and Brønsted acid sites (USY Faujasite) are shown. The bands of pyridine species bonded to Lewis sites, H-bonded, or protonated on Brønsted sites are denoted as L, H, and B. Pyridine is used for several reasons: (i) its liquid state under normal conditions; (ii) its chemical stability; (iii) its

Table 1.4 Families of industrial solid acid catalysts Solid catalysts Protonic zeolites

H-ZSM5 H-Ferrierite H-beta H-Faujasite Rare earth ultrastable Y faujasite Omega (Mazzite)

Silicoaluminophosphates Sulfated zirconia Tungstated zirconia Chlorided alumina Pure alumina Amorphous silica-alumina Solid phosphoric acid

Notation H-MFI H-FER H-BEA H-FAU REUSY H-MAZ SAPO

ASA SPA

Typical reactions Aromatic chemistry Methanol to hydrocarbons Light olefins skeletal isomerization Benzene alkylation Benzene alkylation Fluid catalytic cracking Light paraffin isomerization Methanol to olefins Light paraffin isomerization Light paraffin isomerization Light paraffin isomerization Light olefins skeletal isomerization Ether syntheses from alcohols Mild cracking Olefin oligomerization

Infrared (IR) Spectroscopy

17

medium basicity (pKa 5.2, proton affinity 928 kJ/mol) that is useful for characterizing weak to strong solid acids, which also progressively desorbs by outgassing in a useful temperature range (300–800 K); (iv) the marked and clear difference between the spectra of the protonated form (pyridinium ion) and that of the molecular form (Table 1.5); and (v) the sensitivity of the IR spectrum of its unprotonated form to the strength of the electron-withdrawing interaction (H bonding or Lewis acid-base interaction). The last point is shown in Table 1.5 by the shift of the 8a and 19b bands when comparing the spectrum of liquid pyridine with that of pyridine adsorbed on the strongest Lewis acid sites of alumina. This is also evident from the data reported in Fig. 1.14, where the spectra of pyridine species Lewis bonded on acid sites of different oxides are displayed. The shift to higher wave numbers of the 8a (1583–1624 cm
1) and 19b (1436–1456 cm
1) modes with the increasing Lewis acid

B+L(+H)

1546

1596

1622

L H

Al-poor USY outg. 323 K

1577,2 1547

J-Al2O3 outg. 473 K

1595

1445

1634

1623 1618 1637 1626

Absorbance

B

1491

B B+L H

Characterization of Lewis Acid Sites of Aluminas Alumina is a typical and widely used solid Lewis acid. The Lewis acid

1454 1446

Fig. 1.13 Infrared spectra of pyridine adsorbed at room temperature on different solids previously activated at 773 K under outgassing and outgassed after adsorption at the given temperature. B ¼ band due to protonated species on Brønsted acid sites; H ¼ band due to species hydrogen bonded on surface OHs; L ¼ band due to coordinated species on Lewis acid sites

strength of the adsorbing ion is evident. The use of pyridine as a reference adsorbate is so widespread that a Brønsted solid acid is frequently defined as a solid whose surface sites are able to protonate pyridine. Drawbacks of the use of pyridine are its toxicity and its strong interaction with glass and greases, that in turn gives rise to persistent contamination of the experimental equipment. Evaluation of the absorption coefficients of pyridinium ions and pyridine molecules in principle also allows the evaluation of the number of Lewis and Brønsted acid sites [55–57]. The strength of the sites can indirectly be determined by the temperature of pyridine desorption from Lewis and Brønsted acid sites during outgassing at increasing temperature.

1494 1491

1

1650

1600

1487

1578

Al-poor ZSM5 outg. 423 K

1550 1500 Wavenumbers (cm–1)

Silica outg. 303 K 1450

1400

Table 1.5 Position (v, cm
1) of the vibrational bands of pyridine and pyridinium ion as compared to those of benzene Point group C2v

D6h Benzene C6H6 Sym A1g E2g

Not. 1 8

Activity R R

ν 992 1596

E1u

19

IR

1479

Sym A1 A1 B1 A1 B1

Not. 1 8a 8b 19a 19b

Activity IR,R IR,R IR,R IR,R IR,R

Pyridine C5H5N liquid ν 991 1583 1577 1481 1436

Pyridine on alumina C5H5N ! Al3+

Pyridinium C5H5NH+ (Cl
)

ν 1020 1624 1577 1495 1456

ν 1009 1638 1608 1535 1485

νCC νCC νCC νCC νCC, δCC νCC, δCC

1

18

G. Busca

J-Al2O3

1590

Absorbance

1601

1441

1447 1445

1453

1494

1577

1620 1615 1606 1608

1624 1618

1453

Fig. 1.14 Infrared spectra of pyridine adsorbed at room temperature on different solids previously activated at 773 K under outgassing and outgassed after adsorption at 373–573 K

E-Ga2O3 TiO2-anatase ZrO2-monoclinic 1573

1488

MgO-Al2O3

NaX-zeolite 1650

Fig. 1.15 Infrared spectra of pyridine adsorbed at room temperature on γ-Al2O3 previously activated at 773 K under outgassing and outgassed at the given temperature

1600

1550 1500 Wavenumbers (cm–1)

1450

1445

I II III IV 3.0 2.8

1.6

1455

1490

1.8

1578

2.0

1613 -5 a 1610

2.2 1624

Absorbance

2.4

1593 -8

2.6

1 torr 1.4 1.2

Ev 473 K Ev 673 K

1.0 1700

1650

1600

Ev 300 K Ev 373 K

1550

1500

1450

Wavenumbers (cm–1)

sites of alumina have been well characterized by adsorption of basic probes such as pyridine, carbon monoxide, and several other bases followed by IR, as well as with a variety of other molecules and techniques [20]. They are the strongest Lewis acids among binary metal oxides. Although it is clear that surface Lewis acid sites on alumina are due to

coordinatively unsaturated Al3+ ions, the coordination of such surface ions is not well understood. The spectra recorded during a typical experiment of pyridine adsorption on γ-Al2O3 are reported in Fig. 1.15. At least three different 8a absorption band maxima are observed, one shifting from 1593 to 1598 cm
1, another shifting from

Infrared (IR) Spectroscopy

19

SBA

I II

(110)al (110)al

(111)al

(100)al

(110)al

TH

(110)al (111)al

1455

0.5

1614

g alumina

(100)al

1493

1575

1577

1495

g alumina

Absorbance

Fig. 1.16 Infrared spectra of pyridine adsorbed at room temperature on two different samples of γ-Al2O3 previously activated at 773 K under outgassing and outgassed at the given temperature. The sample Puralox SBA (SBET 198 m2/g, VP 0.45 cm3/g, dP 90 Å) has far smaller pore volume and higher pore sizes than PuraloxTH (SBET 166 m2/g, VP 1.20, cm3/g, dP 188 Å)

have different ratios between the extents of the basal plane (101 spinel plane, corresponding to the basal plane of the original boehmite crystals) and of the “vertical” planes (100 spinel plane). Interestingly, the two spectra are identical, suggesting that sites that adsorb pyridine are primarily located on defects or on edges rather than on different extended planes. On the other hand, theoreticians simulated the adsorption of pyridine on aluminum oxide clusters and found that the calculated shifts of the vibrational modes of pyridine adsorbed on tricoordinated Al3+ ions (giving a tetrahedral complex) agree with those measured experimentally for pyridine adsorbed on the strongest Lewis sites [61, 62] (8a mode at 1624 cm
1 and 19b mode at 1456 cm
1). We recently proposed [63] that a contribution to the spectra can arise from polypyridine species. The species characterized by the 8a mode at 1624 cm
1 and the 19b mode at 1456 cm
1 is not initially observed but starts to form after outgassing at 473 K, when other species (8a at 1615–1610 and 1591 cm
1 and 19b at 1447–1441 cm
1) are disappearing. It is possible that the lower frequency bands are due in part to di- or tripyridine species that decompose upon heating into monopyridine species. This would parallel the behavior of AlCl3 that can coordinate a single pyridine molecule to form the tetrahedral complex PyAlCl3 as well as three pyridine molecules to form the octahedral complex Cl3AlPy3 [64]. The likely formation at high pyridine vapor pressure of polypyridine complexes provides further support for the identification of the strongest Lewis sites of alumina as tricoordinated Al3+. An additional consideration is that surface reconstruction of the structure can occur depending on the conditions. As an example, surface tricoordinated Al species can be formed from

1449

1613- to 1615 cm
1, and the third at 1624 cm
1. In addition, a fourth component may also be present around 1610 cm
1. While at the beginning of the experiment, in contact with pyridine vapor or after short outgassing at ambient temperature, the lower frequency band is the strongest, it progressively disappears. After outgassing at higher temperature, the intermediate frequency component initially is stable or even increases a little bit in intensity but later progressively disappears. The highest frequency band, at 1624 cm
1, increases in intensity up to outgassing at 773 K. The three main components of the spectra have been assigned to pyridine adsorbed on three different types of Lewis acid sites [58], weak, medium, and strong. These Lewis acid sites arise from the divalent or trivalent coordination states of the aluminum ions in the bulk spinel structure, specifically octahedral and tetrahedral (normal spinel positions) and trigonal. Liu and Truitt [59] emphasized the close proximity of Lewis acid sites with surface OHs, while Lundie et al. [60] identified four different Lewis acid sites arising from coordinatively unsaturated octahedral (the weakest) and tetrahedral sites (the three strongest). Three of these sites are considered to be associated with three different types of hydroxyl groups. A possibility for the presence of different Lewis acid sites may be their location on different surface planes, where Al3+ ions may be differently exposed. In Fig. 1.16, the IR spectra of pyridine adsorbed on two samples made from the same precursors with similar preparations (both produced by Sasol), but that differ significantly in pore structure, are reported. Specifically, the sample Puralox SBA has far smaller pore volume and higher pore sizes than PuraloxTH. The two samples are believed to be constituted of laminar/prismatic nanoparticles, which differ in their thickness. This means that the two samples should

1624

1

TH ev 373 K TH ev 573 K SBA ev 373 K SBA ev 573 K 1750

1700

1650

1600

1550 –1

Wavenumbers (cm )

1500

1450

1

20

G. Busca

groups is evident too. In Fig. 1.17, the spectrum of amorphous silica and H-ZSM5 zeolite after adsorption of pyridine is reported. The band of H-zeolite at 3605 cm
1 has disappeared in conjunction with the formation of pyridinium ion. Instead, for silica, strong H bonding occurs between silanol groups and pyridine. The position of the IR band due to bridging OHs is dependent on the size of the zeolite cavities. Generally, the lower the vOH is, the smaller the cavity. In particular, the OH stretching band position and width can be influenced by weak H bonding through the cavities [69]. In the case of zeolites with more than one type of different cavities, splitting of the band of the bridging hydroxyl groups can be observed. Parallel 1H NMR and IR studies show that the IR extinction coefficient of the zeolite’s bridging OHs is far higher than that of the silanol groups Based on this observation, Kazansky et al. [70] proposed the use of the intensity of the IR band to determine the surface acid strength.

Al ions in tetrahedral sites exposed on the 110 and 111 surface of the 110–100 edges of the spinel-type structure. In the bulk spinel structure, the tetrahedra share a face with an empty octahedron of a non-spinel site (an interstice which is not occupied in the spinel structure). Surface reconstruction certainly occurs for dehydroxylated surfaces [65]. Thus, the exposed tricoordinated Al ion may reduce its free energy by slightly shifting down, below the surface, entering this octahedron environment to increase its coordination to four, five, or six. This shift can be reversed in the presence of gas-phase bases, thus producing tetrahedral Al species. A similar hypothesis was proposed by Busco et al. [66] to occur on zeolites. Brønsted Acid Sites in Protonic Zeolites: The Bridging Hydroxyl Group The strongly acidic hydroxyl groups of zeolites are characterized by the presence in the IR spectrum of moderately sharp and strong bands in the region between 3650 and 3500 cm
1 (Fig. 1.10) as well as by 1H MAS NMR peaks in the region 3.6–8.0 ppm [67, 68]. Both techniques demonstrate the acidity of these groups. These spectroscopic signals disappear upon contact with bases like ammonia, pyridines, amines, and phosphines, in parallel with the appearance of the features of the corresponding protonated bases. In the presence of weak basic probes (see below), a significant perturbation of the spectral characteristics of these

a)

Characterization of the Location of Adsorbing Sites in Protonic Zeolites Interestingly, bridging OHs are only detected in the interior of the zeolitic cavities. The corresponding spectroscopic features (both IR and NMR) are absent for nonzeolitic materials based on silica and alumina [71, 72] and also on the external surfaces of zeolites.

b)

1445

1595

Absorbance

1.5

0.4

1487

1578

Silica outg. 30 °C

1491

1.5

0.3 4000

3500

3000

2500 Wavenumbers

1650

1547

1637 1626

Al-poor ZSM5 outg. 150 °C

1600

1550

1500

1450

1400

(cm–1)

Fig. 1.17 (b) Infrared spectra of H-ZSM-5 (very low Al content) and of amorphous silica after activation at 773 K and after pyridine adsorption at r.t. and outgassing at the given temperature. (a) Subtraction spectrum of adsorbed pyridine in the same conditions

1

Infrared (IR) Spectroscopy

21

stabilization of the bridging OH sites as well as in the strengthening of their acidity [58, 73]. This has been confirmed using sterically hindered probes [74] like those reported in Fig. 1.18. For example, monoaromatics like

Thus, the existence of the bridging hydroxyl groups Al-(OH)-Si should imply the existence of interior cavities. In other words, the cavities (or the microporous zeolitic framework) are possibly involved in the generation and/or Fig. 1.18 Molecular probes found useful for the determination of the location of acid sites on and in microporous materials

Pyridines

CH3 CH3

H3C H3C

N

CH3

N

H3C

Quinolines

CH3

CH3

H3C

2,4,6-Collidine

2,6-Lutidine

pyridine

N

N CH3 H3C CH3 2,6-Di-tert-butyl-pyridine

CH3

N 4,6-Dimethylquinoline

N Quinoline

Aromatic hydrocarbons

CH3

CH3

H3C

H3C

CH3

CH3 Paraxylene

Orthoxylene

Metaxylene

Amines H3C

H3C

H3C NH2

H3C H3C Si H3C

N

CH3 CH2

N CH3 NH H3C CH2 H3C H3C Methylamine Dimethylamine Trimethylamine Trimethylsilyldiethylamine

Paraffins

H3C CH2 CH2 CH3

H3C

H

H3C

C

H3C

CH3

-

n butane Isobutane

C

H3C

CH2 CH3 CH3

2,2-Dimethylbutane

Nitriles H3C H3C C N

H3C CH2

C N

H3C CH C N

Propionitrile

Acetonitrile

Isobutyronitrile

H3C H3C C C N H3C Pivalonitrile

H3C C C N

H3C C N Benzonitrile

Cycic siloxanes CH3 CH3 CH 3 Si H3C Si O CH3 Si O O H3C Hexamethylcyclotrisyloxane

C N Ortho-toluonitrile

2,2-Diphenylpropionitrile

1

22

3747 'A = 0.1 3676

Silica

3590 3748

Wavenumbers (cm–1) 3304

Absorbance

pyridine and para xylene do not enter the cavities of Ferrierite easily or rapidly at room temperature [75], while smaller molecules such as acetonitrile enter easily [67]. Para xylene rapidly enters the cavities of ZSM-5 zeolite, while ortho xylene does not [76]. These studies distinguished the location of linear silanols (outer surface) from bridging OHs (internal) and also the location of different hydroxyl groups in different cavities such as those of mordenite [77, 78]. The Lewis acidity of protonic zeolites has also been studied. While it is frequently reported in the literature that Lewis acidity is due to extraframework alumina in defective zeolites, it has been pointed out that Al3+ ions in framework positions can very likely enlarge their coordination from four to five, thus acting as Lewis acid sites. This will occur when the position of the Lewis adsorption site can be spatially decoupled from the Brønsted acidic proton, i.e., when they are exposed to different cavities [79].

G. Busca

3615

'A = 0.1 3415

3748

3480 H-MFI A 3675

Application of CO as a Probe for the Evaluation of the Acid Strength of Surface Brønsted Sites The strength of the acidbase interaction depends on the acid strength of the acid as well as on the basic strength of the base. When the acid or the base is too weak to allow proton transfer, the interaction results in hydrogen bonding. In intermediate cases, the proton transfer can be only partial so that a “symmetrical strong hydrogen bonding” occurs. These two situations can be detected and distinguished spectroscopically by protonation and interaction with Lewis sites. In both of these cases, the spectrum of the adsorbed base is only weakly perturbed with respect to that of the liquid. However, hydrogen bonding causes the shift to lower wave numbers and the broadening of the OH stretching bands of the surface hydroxyl groups. Simple hydrogen bonding gives rise to the formation of well-defined, but broad, OH stretching bands, The shift of the maximum of the vOH band upon interaction can be taken as a measure of the hydrogen bonding interaction. If the same base is used on different surfaces, the shifts measure the acid strengths of the corresponding protonic centers, according to the Bellamy-Hallam-Williams relation [80]. Very weak bases that are not readily protonated, like nitriles or hydrocarbons (such as ethylene or benzene), can be used for this purpose. Low-temperature adsorption of even weaker “bases,” like hydrogen or carbon monoxide (CO), is also a useful experiment in this respect. In particular, the low temperature adsorption of CO is particularly popular for this application. In Fig. 1.19, the spectra of low-temperature CO adsorption on a fumed pure silica sample and on two different samples of H-ZSM-5 zeolite are reported. The adsorption of CO on the silica’s silanols results in the formation of H-bonded Si-O: CO complexes with vCO at 2155 cm
1 (shifted to nearly 18 cm
1 higher wave number with respect to liquid CO) and

3470 3670 3600

3747

3290 'A = 0.1

H-MFI B

3622 3780 3800

3600

3400

Wavenumbers

3200

3000

(cm–1)

Fig. 1.19 Infrared spectra of amorphous silica and two samples of H-MFI zeolite (A large Si/Al atomic ratio, B small Si/Al atomic ratio) after activation at 773 K (solid line) and after CO adsorption at 130 K (broken line)

vOH at 3676 and 3590 cm
1 (shifted to lower wave numbers of ΔvOH ~70 cm
1 and ~155 cm
1). The spectrum of a typical low Al content protonic zeolite, H-MFI (sample A), is reported. As always it shows two bands. The band at 3748 cm
1 is assigned to terminal silanol groups, while that observed at 3615 cm
1 is assigned to bridging Al-(OH)-Si groups which constitute the strong acidic sites of protonic zeolites. The very weak feature observed for this sample at 3675 cm
1 is assigned to OHs on extraframework material. Upon adsorption of CO, the most evident interaction is the shift of the bridging hydroxyls’ stretching band from 3615 to 3304 cm
1 (ΔvOH ~310 cm
1). Additional OH stretching shoulders near 3480 and 3415 cm
1 are likely due to perturbation by adsorbed CO of the weak band near 3675 cm
1 that is due to OHs on extraframework material. In the case of the higher Al-content H-MFI sample B, the presence of extraframework material gives rise to the presence of additional bands in the OH stretching spectrum at 3780 and a strong

1

Infrared (IR) Spectroscopy

band at 3670 cm
1. The low-temperature CO adsorption experiment confirms the strong acidity of the bridging OHs of this zeolite that absorb at 3622 cm
1 and that are shifted upon interaction with CO to 3290 cm
1 (ΔvOH ~330 cm
1). Additionally, the significant acidity of the OHs on extraframework material is evident (see the shift from 3670 to 3470 cm
1, ΔvOH ~200 cm
1). The shift of the terminal silanols from 3747 to 3600 cm
1 (ΔvOH ~150 cm
1) is similar to that of the silanols of pure silica, confirming their quite low acidity. The Olefin Polymerization Method for the Characterization of the Acid Strength of Bronsted Acid Sites Olefins are very reactive toward the electrophilic attack of a Brønsted acid and can undergo proton-catalyzed cationic polymerization at low temperature. This phenomenon occurs more rapidly as the strength of the Brønsted acid and the electron density of the olefinic double bond increase. The experiment must be performed in a medium to low temperature range (e.g., room temperature) with olefin pressures of the order of 20–200 Torr in order to favor rapid oligomerization from the point of view of thermodynamics. The observed polymerization rate in our experimental conditions follows the order: 1,3-butadiene > isobutene > propene > ethylene [50, 81, 82]. Ethylene polymerization (under the conditions used to evaluate the surface acidity of oxide catalysts) is only observed with very strong Brønsted acids, while some weak Brønsted acids only polymerize butadiene. Additionally, very strong Brønsted acids cause the formation of branched polymeric chains from linear olefins, such as polyisobutene formation from both isobutene and 1-butene and also possibly from ethylene. Alumina hydroxyls, although unable to initiate polymerization of the four butene isomers, initiate polymerization of 1,3-butadiene [50].

IR Spectra of Acidic Probes for Surface Basicity Characterization Heterogeneous basic catalysis is primarily applied in fine chemistry and in the chemistry of fine chemical intermediates and very infrequently in refinery and petrochemistry [83]. Basic solid catalysts are mainly alkali and alkaline earth oxides. Some examples of them are summarized in Table 1.6. In contrast to the characterization of catalyst acidity with basic probes, the use of “acidic” molecules to probe surface basicity is far less satisfactory. In fact, all “acidic” (or electrophilic) molecules also contain accessible nucleophilic (acidic) atoms. It seems impossible to find a molecule that only interacts specifically with basic sites. On the other hand, metal oxides that display significant surface basicity always are very ionic and also possess weak Lewis acidity [39, 84]. So, in this case, the acidity-basicity more than pure basicity properties are relevant.

23 Table 1.6 Catalysts and conditions of industrial solid base or acid-base catalyzed reactions Catalyst composition MgO-Al2O3 (calcined hydrotalcite) MgO-Al2O3 (calcined hydrotalcite) with Pt MgO-SiO2 NaOH-SiO2 Cs2O/SiO2

ThO2 or ZrO22

Process/reaction Epoxide ring-opening/ ethoxylation alcohol to polyethoxylates Acetone to methyl isobutyl ketone

Butadiene from ethanol Acrolein from acetaldehyde + formaldehyde Methyl methacrylate from methyl propionate + formaldehyde Crotonic condensation of 2-butanone 2,4-dimethyl-3-pentanone from isobutyric acid

Conditions 400–450 K 4–5 bar 400–500 K

650–690 K 570–600 K 600–700 K 600–700 K 700 K

IR Spectra of Adsorbed Carbon Dioxide for Surface Basicity/Nucleophilicity Characterization Carbon dioxide is the most frequently used probe for surface basicity characterization. Spectroscopic studies [85, 86] show that carbon dioxide may adsorb both as an acid on basic sites and as a base on acid sites. When behaving as a base, CO2 uses one of the lone pairs of its oxygen atoms to coordinate with Lewis acid sites of ionic surfaces. The OCO asymmetric stretching band shifts to higher wave numbers from the gas-phase value of 2340 cm
1 [87–89]. These weakly bonded species disappear quite easily upon outgassing at room or slightly higher temperatures. When acting as an acid (or better as an electrophile), CO2 reacts with surface oxygen or hydroxide species to form carbonate or bicarbonate species. The free carbonate ion, as a result of its trigonal D3h symmetry, has one characteristic strong infrared vibration (v3; doubly degenerate asymmetric CO stretching) found near 1415 cm
1 for bulk metal carbonates, together with two lower frequency IR active deformation modes. The lowering of the ion symmetry, due to coordination, causes the splitting of this doubly degenerate v3 vibration. The Raman active symmetric deformation mode v1 is active in the IR as well. As deduced by IR spectra, different types of carbonate ions have been considered as possible surface species: (i) planar trigonal carbonate species that may form on the surface of strongly basic materials; (ii) monodentate carbonates; (iii) bridging and/or chelating (bidentate) carbonates; and (iv) polydentate carbonates. These structures may be distinguished by consideration of their stability (polydentate > bidentate > chelating > monodentate) and the extent of the splitting of v3, the band present at 1415 cm
1 in the symmetric carbonate species. The extent of splitting is bidentate  chelating > monodentate  polydentate > trigonal. Bicarbonate ions

1

24

G. Busca

[HCO3]
can also be formed by adsorption of/contamination by CO2. These species are characterized by vOH modes near 3620 cm
1, δOH modes near 1300–1200 cm
1, as well as by two vC¼O modes (~1600 and ~1450 cm
1). Although surface cations are also involved in carbonate formation, the basicity of the oxide is the dominating factor in stabilizing surface carbonates. Carbon dioxide desorption may be followed by monitoring the disappearance of the features of these species using IR spectroscopy: the higher the desorption temperature, the more basic the adsorption site. For very basic materials, complete desorption of surface carbonates needs treatment in the 800–1100 K range. For example, Fig. 1.20 shows the IR spectra of the adsorbed species arising from the adsorption of CO2 on samples of the La2O3/γ-Al2O3 system. The bands at 1643 cm
1, 1484 cm
1 (with 1450 cm
1 as a shoulder), and 1236 cm
1 observed upon adsorption of CO2 on pure γ-Al2O3 are due to surface bicarbonate species (C¼O asymmetric and symmetric stretching, and OH deformation, respectively) [81]. These bands show that few nucleophilic (basic) hydroxyl groups are present on hydroxylated alumina to convert CO2 into bicarbonate species. The spectrum of adsorbed species is strongly reduced in intensity by the addition of small amounts of lanthanum (5% wt/wt La2O3), while addition of 20% La2O3 further modifies the spectrum. In the latter case, the bands due to bicarbonate species broaden and shift to ca. 1630, 1428, and 1229 cm
1, while bands due to carbonate species at ca 1530 and 1350 cm
1 (asymmetric and symmetric C¼O stretching, possibly with further components in the region near

1643

1485

A 0.5

1450

1235

1703 1773 1815

1460

Absorbance

1570 1650

1438 1228

5LaA

1630

1550

20LaA

1376 1428 1340 1229

1520 1900

1800

1700

1600

1500

1380 1400

1300

1200

Wavenumbers (cm–1)

Fig. 1.20 Infrared spectra of CO2 adsorption at r.t. (solid line) and after evacuation at 423 K (dashed line) on pure alumina and on samples containing 5% and 20% of La2O3

1630 cm
1) are formed. These results show that addition of La first inhibits the basic sites of alumina that produce bicarbonate species by CO2 adsorption then, at higher loadings, provides new basic and nucleophilic sites for CO2 adsorption [90]. These data are in good agreement with those reported for K2O/γ-Al2O3 [91]. Samples containing almost a full monolayer of the supported basic phase are very basic, while small amounts of supported basic species only neutralize the acidity of the γ-Al2O3 carrier.

IR Spectra of Adsorbed Carbon Monoxide for Metallic/Cationic Sites Characterization Carbon monoxide is a very weak base primarily used for the surface characterization of cationic centers on metal oxide surfaces [92]. Low-temperature CO adsorption experiments allow the observation of weakly acidic but active adsorption sites, including surface hydroxyl groups (see above) and alkali cations, whose interaction with CO is not observed in room or higher temperature experiments. A summary of metal sites revealed by low-temperature CO adsorption on a variety of alumina-supported catalysts is reported in Table 1.7. Low-temperature adsorption experiments enable more precise photography of the state of the surface as they avoid misunderstandings due to the reactivity of CO, which may reduce the surface species and produce carbonate species and CO2, and sometimes react to form undetectable surface carbide species. However, room temperature experiments may provide evidence for “activated” adsorption phenomena, revealing adsorption modes that require energy to be allowed. As discussed by Hadjiivanov and Vayssilov [91], the use of isotopically labeled 13CO, whose stretching modes are shifted with respect to those of 12CO, or of 13CO-12CO mixtures, also helps in identifying the nature of the adsorbed species. As is well known [93], CO bonds preferentially end-on (terminally) through the lone pair on the C atom (σ-bonding) on oxidized cations [94], although very weak O-bonded species can also sometimes form [95]. The position of the CO band is sensitive to the cation Lewis acid strength that enhances the C-O stretching frequency as a result of electron withdrawal through the σ-bond on C, although d- π* electron backdonation can have the reverse effect. In any case, the position of the CO stretching band is indicative of the cation oxidation state. Additionally, when the cation is strongly oxidizing and oxide species are available, oxidation of CO gives rise to CO2 at room temperature or even at temperatures as low as 140 K, as shown in the IR spectra. This is the case for Ru4+, Pd4+, Cu2+, Pt2+, and Co3+ centers (see Table 1.7). Thus, CO is an efficient probe to reveal the reducibility of oxide surfaces. In the case of zeolitic internal surfaces, where several cations may be located in close proximity, CO provides evidence for the possibility of complex interactions. In such interactions, both C and O lone pairs and even the π-orbital

1

Infrared (IR) Spectroscopy

25

Table 1.7 Position of CO stretching of surface carbonyls on catalytic materials Material SiO2

Position 2155

Adsorbing species Si-OH

Al2O3 TiO2 ZrO2 H-zeolites

2240–2180 2210–2170 2200–2165 2170–2180

Al3+ Ti4+ Zr4+ Si-OH-Al

2220–2190 2160–2190 2160–2180 2150–2100 2170–2160 2150–2100 2160–2140 2150–2100 2150–2140 2190

Al3+ Li+ Na+ More Na+ ions K+ More K+ ions Cs+ More Cs + ions K+ Co3+

CO species Terminal H-C bonded Terminal Terminal Terminal Terminal H-C bonded Terminal Terminal Terminal C- and O multiply bonded Terminal C- and O multiply bonded Terminal C- and O multiply bonded Terminal Terminal

2170 2050 2030–1990

Co2+ Co0 Co0

Terminal Terminal Terminal

2167–2173 2150–2000

Ni2+ Nin+

2080–2020 1930–1870 2160–2140 2120–2100 2080–2050 2135 2142, 2075, 2004 2040–1980

Ni Ni Cu2+ Cu+ Cu Ru4+ Rux+ Ru

Terminal Terminal Polycarbonyls Terminal Bridging Terminal Terminal Terminal Terminal Terminal Terminal

1960 ca 2090, ca. 2020 2080–2050 2160–2170 2150–2140

Ru Rh+ ions

Bridging Dicarbonyls Terminal Terminal Terminal

2130 2120–2080

Rh Pd2+ ions Pd2+ on PdO particles Pdn+ Partially oxidized part. Pd

1990–1980 1820 and 1778 2186 2135 2100–2030 1815–1780

Pd Pd Pt4+ ions Pt2+ Pt Pt

Bridging Triply bridging Terminal Terminal Terminal Bridging

Li-zeolites Na-zeolites K-zeolites Cs-zeolites K2O/Al2O3 Co/Al2O3

Ni/Al2O3

Cu/Al2O3

Ru/Al2O3

Rh/Al2O3

Pd/Al2O3

Pt/Al2O3

may be involved [96, 97]. An interaction of CO with oxide ions has also been suggested [98]. The use of IR spectroscopy of adsorbed CO for the characterization of supported metal catalysts takes advantage of

Terminal Terminal

Notes

Different sites depending on activation T

Framework or extraframework ions

More resistant t outgassing More resistant t outgassing More resistant t outgassing On Co3O4 particles Oxidation CO to CO2 at rt On defects On faces Shift due to vibration couplings Several sharp bands

Oxidation CO to CO2 at rt Strong, resistant Weak very labile Oxidation CO to CO2 at rt Polycarbonyls On faces Shift due to vibrat. Couplings On faces Strong doublet

Oxidation CO to CO2 at rt

On faces Shift due to vibrat. Couplings

Oxidation CO to CO2 at rt Shift due to vibrat. Couplings Weak

the many surface science studies performed on single crystal metal faces [99]. The technique of low-temperature CO adsorption followed by vibrational spectroscopies (such as IRAS, high-resolution EELS (HREELS), or, more recently,

1

26

G. Busca

sum frequency generation spectroscopy (SFG)) is a widely used technique in metal surface science studies. These techniques give very precise reference data for catalyst characterization and allow good comparison among the respective results. Coupling the data obtained from FTIR spectra of real catalysts and those from single crystals with the results of computational studies may provide a picture of the morphology and activity of metal nanoparticles in real catalysis. Figure 1.21 presents the spectra of CO on a Ru/Al2O3 sample recorded upon warming from 140 K at room temperature [100]. In the spectrum obtained with a slightly pre-reduced sample, the bands due CO interacting with Ru metal particles are weak, with a maximum at 2027 cm
1 and a clear shoulder near 2000 cm
1 that are resistant to outgassing at r.t.. Under these conditions, a strong band is also formed at 2156 cm
1 with a shoulder at 2135 cm
1, which is due at least in part to carbonyls interacting with Rux+ ions together with H-bonded CO. During outgassing while warming, these features progressively disappear, while the band of linearly adsorbed CO2 initially grows and subsequently disappears. These results show that unreduced Rux+ centers oxidize CO to CO2 that later desorbs. In contrast, for highly reduced samples, this phenomenon does not occur. Instead, the bands due to CO adsorbed on metallic ruthenium

1.6

Application of IR Spectroscopy to the Study of the Mechanisms of Heterogeneous Catalysis

From the first applications of the IR technique for surface studies and heterogeneous catalysis, it was recognized that this technique can be useful, not only for the characterization of the catalyst surface structure by detection of the spectra of adsorbed probe molecules but also to give direct information on the mechanisms of heterogeneous catalysis. Beginning with the work of Robert P. Eischens and co-workers [3], in which the adsorption of reactants of catalytic reactions of industrial interest over the relevant catalysts

a)

2027

0.38

2000 2156

1.5 2344 Absorbance

Fig. 1.21 Infrared spectra of the surface species after CO adsorption on Ru/Al2O3 samples, slightly pre-reduced (a) and more hardly pre-reduced (b). Spectra recorded at 140 K, in contact with 10 torr of CO and upon outgassing upon warming until RT (a) and also after heating to 673 K (b). Inset: magnification of the low frequency region for the experiment with the weakly reduced sample

are more intense. They disappear during outgassing and heating to 673 K and progressively shift from 2036 to 1990 cm
1 due to vibrational coupling. Weak bands 2142, 2075, and 2004 cm
1 due to polycarbonyl species are also observed, as is interaction of CO with the alumina support at 2190–2220 cm
1. The behavior of the bands in the range 250–1950 cm
1 is similar to that found by surface science techniques (Electron Energy Loss Spectroscopy and Reflection-Absorption IR) on the Ru (0001) single crystal [101] plane.

0.26 2080

2040

2000

1960

2027 0.3 2400

2300

2200

2100

2000

1900

b) 0.66

2160 2192 2036 2073

1990

2140

0.40 2400

2300

2200 2100 Wavenumbers (cm–1)

2000

1900

1920

1

Infrared (IR) Spectroscopy

was investigated, a possible indication of the mechanism of the catalytic steps was obtained. This approach was later developed by Richard Kokes and co-workers [102] in the 1960s. The concept of experiments under dynamic conditions was also developed in the 1960s by Kenzi Tamaru [103] and applied to IR spectroscopy and other techniques. During his experiments, Tamaru measured the conversion of reactants in parallel with the modification of the spectra of the surface species for several reactions such as (i) olefin isomerization and hydrogenation over ZnO, (ii) decomposition of formic acid, (iii) water-gas shift reaction over ZnO and MgO, (iv) decomposition of methanol over ZnO, and (v) oxidation of CO on supported Pd. Starting in the 1970s, appropriate IR cells for experiments performed under flow conditions – with the simultaneous detection of the reaction product – while also measuring the conversion and selectivities have been developed. In those times, these experiments were denoted as in situ studies. In an extensive review published in 2001, J. Ryczkowski described and discussed most of the designs of IR cells used for these studies [11]. More recently, this approach has evolved into the concept of operando spectroscopy [104] applied to different spectroscopic techniques including IR [105, 106], with the possibility of applying several spectroscopic techniques simultaneously to the same catalyst sample under working conditions [107]. A number of cells have been developed to evaluate kinetics of catalytic reactions [108]. A number of problems, however, still exist, and different attempts to resolve them are currently under investigation [109, 110]. In any case, operando techniques, including IR, are very useful today for monitoring the state of the catalyst during reaction. The real problem for mechanistic studies with operando techniques, in particular for operando infrared, is that reaction intermediates are frequently not detectable under operando conditions. Although IR spectroscopy is a very powerful technique for detecting surface intermediates, most of them are always not detectable. Additionally, under reaction conditions, most intermediates are present in extremely low concentrations, especially all those formed subsequent to the rate determining step. For the past 40 years in our lab, we have approached the study of reaction mechanisms in heterogeneous catalysis using in situ IR experiments similar to those of Kokes and Tamaru and other earlier authors [111], in conjunction with simultaneous analysis of the gas phase coupled with independent flow reactor experiments. With these experiments, some IR detectable intermediates trapped on the surface at lower temperatures, when they still do not react, and their evolution with temperature can be followed. In several cases, intermediates have been found to form the reaction products

27

under the reaction conditions when they disappear, confirming that they can be identified when they are not reacting but are not detectable in reaction conditions. As an example, many studies are devoted to the hydrogenation of carbon oxides. As is well known, both CO and CO2 and their mixtures can be hydrogenated to methane (methanation) over Ni- or Ru-based catalysts, linear hydrocarbons (Fischer Tropsch synthesis; FTS) over Co- or Ru-based catalysts, as well as to alcohols over Cu-based catalysts. For all these reactions, reaction mechanisms via carbide intermediates are proposed by analogy with the reactivity found for metallic single crystals that form surface carbides. However, mechanisms via oxygenates, including the formation of formate species as key intermediates on the support surface, have also been suggested, mainly based on IR spectroscopic studies [112]. Such oxygenated intermediates are usually evident in the spectra recorded in situ or under operando conditions when the catalyst support is an ionic oxide, like alumina, titania, and zirconia, and are believed to be formed on the uncovered support surface. However, these species do not form when the support is silica. Figure 1.22 shows the spectra of surface and gas-phase species formed by the interaction of CO and H2 at room temperature and 523 K for 3 h over a Ru/TiO2 catalyst which is active for FTS [113]. The gas-phase spectra show that CO hydrogenation is already occurring under these conditions (negative CO band is due to its consumption) to produce methane (the main product of FT synthesis on a molar basis, 3015 and 1305 cm
1), ethane, ethylene, and C3 and C4 hydrocarbons (bands at 1050–850 cm
1). The spectrum of the surface shows the presence of surface carbonyls on metallic ruthenium (2011, 1960 cm
1, see above), bands due to oxygenated compounds such as formates (1620, 190, 1365 cm
1), acetates (1550, 1460 cm
1), and linear paraffin chains (2954, 2853, 1420 cm
1). The presence of oxygenated compounds together with reaction products could suggest that they are intermediates in the reaction. However, ruthenium carbide species which are believed to be the active species in the reaction [114] cannot be detected by infrared spectroscopy. A similar result is obtained with Co/Al2O3 catalysts [115]. A similar situation also occurs for CO2 and CO methanation over supported metal catalysts, in which formate species adsorbed on ionic oxide supports appear to be present during the reaction [96, 108] and are believed to be the active species [116]. Again in this case, however, carbide species could be also present but are not detectable. Additionally, unsupported metals as well as metals supported on silica (where formates do not form) are also active in these reactions. Formate species are always easily formed in the presence of CO or CO2 þ H2 mixtures over any ionic oxide support, but this is likely the result of their poor chemical activity [117]. Thus, it seems that the

1

28

G. Busca

formation of such species is not necessary for the reaction. In the opinion of the present author, studies ignoring these points are unreliable. Another example is related to the conversion of ethanol over alumina to produce diethyl ether and ethylene [86, 118, 119]. The IR in situ study (Fig. 1.23) shows that at up to 423 K ethanol (CO stretching at 1066 cm
1) is the only Fig. 1.22 Infrared spectra of the surface species (subtracted spectrum, A) and gas phase (B) after contact of Ru/TiO2 catalyst with CO þ H2 mixture for 3 h at 523 K

compound found in the gas phase except for a small amount of water vapor. This agrees with the low conversion (0.2%) found at 423 K in the flow reactor study. Under these conditions, the surface is covered by ethoxy groups (1449, 1389, 1167, 1121, 1074 cm
1, CH2 deformation, CH3 stretching, CH3 rocking, coupled C-C and C-O stretching) with weak bands due to surface acetate species (1578, 1462 cm
1,

3015 949 B expanded

2954

912

990

889

0.50 1200

1100

1000

900

800 1305

Absorbance

2853

B 3000

2500

2000

2011

1500 1620 1460, 1420 1550

1960

1000

1390 1365

A 2600

b)

1200

423 K

949

423 K 523 K

523 K

573 K

573 K

0.2 1600

1400

473 K

473 K

1700

1600

1167

1389

1462 1449

1578

Absorbance

1058

1.9

2400 2200 2000 1800 Wavenumbers (cm–1)

1066

1121

a)

2800

1141

3000

1074

–0.05 3200

1500

1400

1300

1200

1100

1300

Wavenumbers

1200

1100

1000

900

(cm–1)

Fig. 1.23 Infrared spectra of the surface species (subtracted spectrum, a) and gas phase (b) after contact of γ-Al2O3 catalyst with ethanol for 30 min at the given temperatures

1

Infrared (IR) Spectroscopy

asymmetric and symmetric stretching of the carboxylate group -COO
) and a weak feature due to weakly bonded undissociated ethanol (1056 cm
1, CO stretching). At 473 K, ethanol is still the predominant gas-phase component, with a low concentration of diethyl ether (1141 cm
1, C-O-C asymmetric stretching). This agrees with the low conversion found with the flow reactor (20.8%) to produce diethyl ether. The spectrum of the surface species is almost unchanged: only the weak CO stretching band of hydrogen-bonded ethanol has disappeared. At 523 K, the intensity of the bands of gaseous ethanol has decreased to half of the original intensity, while a strong band for diethyl ether is present. A trace of ethylene is also present (CH2 wagging mode at 949 cm
1). This agrees with the substantial conversion of ethanol (78.6%) to diethyl ether (selectivity 79.7%) and small amounts of ethylene (20.2% selectivity) found with the flow reactor. Under these conditions, the intensity of the bands of the ethoxy groups is considerably lower. Finally, at 573 K, the gas-phase spectra of both ethanol and diethyl ether are absent, and ethylene is the only product. Again, this agrees with the results of flow reactor data at 573 K: ethanol conversion, 97.7%; ethylene selectivity, 99.7%. Only traces of ethoxy groups are present on the catalyst surface, while acetate species remain. These data, coupled with data arising from diethyl ether cracking experiments, clearly suggest that cracking of ethoxy groups (monomolecular mechanism) is the main reaction producing ethylene from ethanol at high temperature and conversion under conditions of practical interest. Under these conditions, the concentration of the intermediate ethoxy groups is near zero, while the acetate groups, which are clearly “spectator species,” appear more intense. Thus, under operando conditions, this mechanism probably cannot be observed. Our study also shows that gaseous ethanol (i.e., uncomplete ethanol conversion) and ethoxy groups are needed to produce diethyl ether, which suggests that alkoxides react with gaseous or H-bonded ethanol to produce the ether. These experiments show that carefully performing parallel flow reactor and IR experiments can even be more informative than trying to conduct the two experiments together.

1.7

Conclusions

In this chapter, we summarized the work done using IR spectroscopy for the characterization of real catalysts in our laboratory. The wide availability of this technique today in most laboratories demonstrates how useful it is 80 years after the pioneering work of the academician Alexander Terenin and his coworkers and 70 years after the pioneering studies by Robert Eischens at Texaco. The wide availability of this technique, however, does not always correspond to the same high level of knowledge of spectroscopic details in these studies. In any case, IR spectroscopy will continue to play

29

an important role in the development of catalysis for application to the new industrial chemistry that will be based on renewable raw materials.

References 1. Buswell, A., Krebs, K., Rodebush, W.: Infrared studies. III. Absorption bands of hydrogels between 2.5 and 3.5 micrometers. J. Am. Chem. Soc. 59, 2603–2605 (1937) 2. Terenin, A., Roev, L.: Infra-red spectra of nitric oxide adsorbed on transition metals, their salts and oxides. Spectrochim. Acta. 11, 946–957 (1959) 3. Eischens, R., Pliskin, W.: The infrared spectra of adsorbed molecules. Adv. Catal. Relat. Subj. 10, 1–56 (1958) 4. Peri, J.B., Hannan, R.B.: Surface hydroxyl groups on γ-alumina. J. Phys. Chem. 1960(64), 1526–1530 (1960) 5. Parry, E.P.: An infrared study of pyridine adsorbed on acidic solids. Characterization of surface acidity. J. Catal. 2, 371–379 (1963) 6. Stavitski, E., Weckhuysen, B.M.: Infrared and Raman imaging of heterogeneous catalysts. Chem. Soc. Rev. 39, 4615–4625 (2010) 7. Zaera, F.: New advances in the use of infrared absorption spectroscopy for the characterization of heterogeneous catalytic reactions. Chem. Soc. Rev. 43, 7624–7663 (2014) 8. Busca, G.: Infrared spectroscopy in oxidation catalysis. In: Duprez, D., Cavani, F. (eds.) Handbook of Advanced Methods and Processes in Oxidation Catalysis, pp. 447–495. Imperial College Press (2014) 9. Bordiga, S., Lamberti, C., Bonino, F., Travert, A., ThibaultStarzyk, F.: Probing zeolites by vibrational spectroscopies. Chem. Soc. Rev. 44, 7262–7341 (2015) 10. Concepción, P.: Application of infrared spectroscopy in catalysis: impacts on catalysts’ selectivity. In: El-Azazy, M. (ed.) Infrared Spectroscopy - Principles, Advances, and Applications. Intechopen (2019)., available on-line. London 11. Nakamoto, K.: Infrared and Raman Spectra of Inorganic and Coordination Compounds: Part A: Theory and Applications in Inorganic Chemistry, Sixth edn. Wiley: Hoboken, New Jersey, USA (2009) 12. Busca, G.: The use of vibrational spectroscopies in studies of heterogeneous catalysis by metal oxides: an introduction. Catal. Today. 7, 323–352 (1996) 13. Ryczkowski, J.: IR spectroscopy in catalysis. Catal. Today. 68, 263–381 (2001) 14. Duykaerts, G.: The infra-red analysis of solid substances. A review. Analyst. 84, 201–214 (1959) 15. Durig, J.R.: Vibrational spectra of solids. In: Infrared and Raman Spectroscopy of Biological Molecules Theo M. Theophanides red., pp. 69–80. D. Reidel Publishing Company, Dordrecht (1979) 16. Busca, G., Resini, C.: Vibrational spectroscopy for the analysis of geological and inorganic materials. In: Meyers, R.A. (ed.) Encyclopedia of Analytical Chemistry, pp. 10984–11102. Wiley, Chichester (2000) 17. Fateley, W., Dollish, F., McDevitt, N., Bentley, F.F.: Infarared and Raman Selection Rules for Molecular and Lattice Vibrations. Wiley, New York (1972) 18. Pascale, F., Zicovich-Wilson, C.M., Lopez, F., Civalleri, B., Orlando, R., Dovesi, R.: The calculation of the vibration frequencies of crystalline compounds and its implementation in the CRYSTAL code. J. Comput. Chem. 25, 888–897 (2004) 19. Pandey, S.C., Xu, X., Williamson, I., Nelson, E.B., Li, L.: Electronic and vibrational properties of transition metal-oxides: comparison of GGA, GGA + U, and hybrid approaches. Chem. Phys. Lett. 669, 1–8 (2017) 20. Baraton, M.I., Busca, G., Prieto, M.C., Ricchiardi, G., Sanchez Escribano, V.: On the vibrational spectra and structure of FeCrO3

1

30 and of the ilmenite-type compounds CoTiO3 and NiTiO3. J. Solid State Chem. 112, 9–14 (1994) 21. Busca, G.: Structural, surface, and catalytic properties of aluminas. Adv. Catal. Relat. Subj. 57, 319–404 (2014) 22. Busca, G.: The surface of transitional aluminas: a critical review. Catal. Today. 226, 2–13 (2014) 23. Born, M., Huang, K.: Dynamic Theory of Crystal Lattices. Clarendon Press, Oxford (1956) 24. Tarte, P., Rulmont, A., Liégeois-Duyckaerts, M., CahayJ, R., Winand, M.: Vibrational spectroscopy and solid state chemistry. Solid State Ionics. 42, 177–196 (1990) 25. Astorino, E., Peri, J.B., Willey, R.J., Busca, G.: Spectroscopic characterization of silicalite-1 and titanium silicalite-1. J. Catal. 157, 482–500 (1995) 26. Balan, E., Saitta, A.M., Mauri, F., Calas, G.: First-principles modeling of the infrared spectrum of kaolinite. Am. Miner. 86, 1321–1330 (2001) 27. Finocchio, E., Gonzalez-Prior, J., Gutierrez-Ortiz, J.I., LopezFonseca, R., Busca, G., de Rivas, B.: Surface characterization of mesoporous CoOx/SBA-15 catalyst upon 1,2-dichloropropane oxidation. Materials. 11, 912 (2018) 28. Busca, G., Riani, P., Garbarino, G., Ziemacki, G., Gambino, L., Montanari, E., Millini, R.: The state of nickel in spent fluid catalytic cracking catalysts. Appl. Catal. A Gen. 486, 176–186 (2014) 29. Gallegos, M.V., Peluso, M.A., Finocchio, E., Thomas, H.J., Busca, G., Sambeth, J.: Removal of VOCs by catalytic process. A study of MnZnO composites synthesized from waste alkaline and Zn/C batteries. Chem. Eng. J. 313, 1099–1111 (2017) 30. Riani, P., Garbarino, G., Cavattoni, T., Canepa, F., Busca, G.: Unsupported cobalt nanoparticles as catalysts: effect of preparation method on catalytic activity in CO2 methanation and ethanol steam reforming. Int. J. Hydrcarb. Eng. 44, 27319–27328 (2019) 31. Riani, P., Garbarino, G., Cavattoni, T., Busca, G.: CO2 hydrogenation and ethanol steam reforming over CO/SiO2 catalysts: deactivation and selectivity switches. Catal. Today. (in press) 32. Doornkamp, C., Ponec, V.: The universal character of the Mars and Van Krevelen mechanism. J. Mol. Catal. A Chem. 162, 19–32 (2000) 33. Henry, R.L.: The transmission of powder films in the infra-red. J. Opt. Soc. Am. 38, 775–789 (1948) 34. Lejeune, R., Duykaerts, G.: Note concernant les spectres infrarouges de poudres. Spectrochim. Acta. 6, 194–197 (1954) 35. Busca, G., Ramis, G., Lorenzelli, V.: FT-IR study of the surface properties of polycrystalline vanadia. J. Mol. Catal. 50, 231–240 (1989) 36. Ramis, G., Cristiani, C., Elmi, A.S., Villa, P., Busca, G.: Characterization of the surface properties of polycrystalline WO3. J. Mol. Catal. 61, 319–331 (1990) 37. Busca, G., Cavani, F., Centi, G., Trifirò, F.: Nature and mechanism of formation of vanadyl pyrophosphate:active phase in n-butane selective oxidation. J. Catal. 90, 400–414 (1986) 38. Busca, G., Ricchiardi, G., Siew Hew Sam, D., Volta, J.-C.: Spectroscopic characterization of magnesium vanadate catalysts part 1 .-vibrational characterization of Mg3(VO4)2, Mg2V2O7 and MgV2O6 powders. J. Chem. Soc. Faraday Trans. I. 90, 1161–1170 (1994) 39. Ramis, G., Busca, G., Lorenzelli, V.: FT-IR study of the dispersion of the supported phase on MoO3-TiO2 catalysts. Z. Phys. Chem. Neue Folge. 153, 189–200 (1987) 40. Carter, J.L., Lucchesi, P.J., Corneil, P., Yates, D.J.C., Sinfelt, J.H.: Exchange of deuterium with the hydroxyl groups of alumina. J. Phys. Chem. 69, 3070–3074 (1965) 41. Contescu, C., Popa, V.T., Schwarz, J.A.: Heterogeneity of hydroxyl and deuteroxyl groups on the surface of TiO2 polymorphs. J. Colloid Interface Sci. 180, 149–161 (1996)

G. Busca 42. Busca, G.: The surface acidity of solid oxides and its characterization by IR spectroscopic methods. An attempt at systematization. Phys. Chem. Chem. Phys. 1, 723–736 (1999) 43. Busca, G.: The surface acidity and basicity of solid oxides and zeolites. In: Fierro, J.L.G. (ed.) Metal Oxides: Chemistry and Applications, pp. 247–318. CRC Press, Boca Raton (2005) 44. Tsyganenko, A., Filimonov, V.N.: Infrared spectra of surface hydroxyl groups and crystalline structure of oxides. Spectrosc. Lett. 5, 477–487 (1972) 45. Garbarino, G., Travi, I., Pani, M., Carnasciali, M.M., Busca, G.: Pure vs ultra-pure γ-alumina: a spectroscopic study and catalysis of ethanol conversion. Catal. Commun. 70, 77–81 (2015) 46. Phung, T.K., Herrera, C., Larrubia, M.Á., García-Diéguez, M., Finocchio, E., Alemany, L.J., Busca, G.: Surface and catalytic properties of some gamma-Al2O3 powders. Appl. Catal. A Gen. 483, 41–51 (2014) 47. Lavalley, J.C., Benaissa, M., Busca, G., Lorenzelli, V.: FT-IR study of the effect of pretreatment on the surface properties of alumina produced by flame hydrolysis of AlCl3. Appl. Catal. 24, 249–255 (1986) 48. Busca, G.: Differentiation of mono-oxo and poly-oxo and of monomeric and polymeric vanadate, molybdate and wolframate species in metal oxide catalysts by IR and Raman spectroscopies. J. Raman Spectrosc. 33, 348–358 (2002) 49. Busca, G., Lavalley, J.C.: Use of overtone bands to monitor the state of catalyst active phases during infrared studies of adsorption and catalytic reaction. Spectrochim. Acta. 42 A, 443–445 (1986) 50. Saur, O., Bensitel, M., Mohammed Saad, A.B., Lavalley, J.C., Tripp, C.P., Morrow, B.A.: The structure and stability of sulfated alumina and titania. J. Catal. 99, 104–110 (1986) 51. Lavalley, J., Benaissa, M.: IR spectroscopic evidence for surface vibrational modes upon dehydroxylation of alumina. J. Chem. Soc. Chem. Commun. 14, 908–909 (1984) 52. Busca, G., Lorenzelli, V., Sanchez Escribano, V., Guidetti, R.: FTIR study of the surface properties of the spinels NiAl2O4 and CoAl2O4 in relation to those of transitional aluminas. J. Catal. 131, 167–177 (1991) 53. Busca, G.: Acid catalysts in the industrial hydrocarbon chemistry. Chem. Rev. 107, 5366–5410 (2007) 54. Busca, G.: Spectroscopic characterization of the acid properties of metal oxide catalysts. Catal. Today. 41, 191–206 (1998) 55. Emeis, C.A.: Determination of integrated molar extinction coefficients for infrared absorption bands of pyridine adsorbed on solid acid catalysts. J. Catal. 141, 347–353 (1993) 56. Massa, M., Andersson, A., Finocchio, E., Busca, G.: Gas-phase dehydration of glycerol to acrolein over Al2O3-, SiO2-, and TiO2supported Nb- and W-oxide catalysts. J. Catal. 307, 170–184 (2013) 57. Phung, T.K., Carnasciali, M.M., Finocchio, E., Busca, G.: Catalytic conversion of ethyl acetate over faujasite zeolites. Appl. Catal. A Gen. 470, 72–80 (2014) 58. Morterra, C., Magnacca, G.: A case study: surface chemistry and surface structure of catalytic aluminas, as studied by vibrational spectroscopy of adsorbed species. Catal. Today. 27, 497–532 (1996) 59. Liu, X., Truitt R.E.: DRFT-IR studies of the surface of gammaalumina. J. Amer. Chem. Soc. 119, 9856–9860 (1997) 60. Lundie, D.T., McInroy, A.R., Marshall, R., Winfield, J.M., Jones, P., Dudman, C.C., Parker, S.F., Mitchell, C., Lennon, D.: Improved description of the surface acidity of η-alumina. J. Phys. Chem. B. 109, 11592–11601 (2005) 61. Nastova, I., Skapin, T., Pejov, L.: Effect of partial fluorination on the Lewis sites of microcrystalline γ-alumina studied by adsorption of pyridine as a probe molecule: a quantum chemical cluster model study. Surf. Sci. 605, 1525–1536 (2011)

1

Infrared (IR) Spectroscopy

62. Kassab, E., Castellà-Ventura, M.: Theoretical study of pyridine and 4,4‘-Bipyridine adsorption on the Lewis acid sites of alumina surfaces based on ab initio and density functional cluster calculations. J. Phys. Chem. B. 109, 13716–13728 (2005) 63. Busca, G.: Catalytic materials based on silica and alumina: structural features and generation of surface acidity. Prog. Mater. Sci. 104, 215–249 (2019) 64. Krahl, T., Stösser, R., Kemnitz, E., Scholz, G., Feist, M., Silly, G., Buzaré, J.Y.: Structural insight into aluminum chlorofluoride (ACF). Inorg. Chem. 42, 6474–6483 (2003) 65. Sohlberg, K., Pantelides, S.T., Pennycook, S.J.: Surface reconstruction and the difference in surface acidity between γ- and η-alumina. J. Am. Chem. Soc. 123, 26–29 (2001) 66. Busco, C., Bolis, V., Ugliengo, P.: Masked Lewis sites in proton exchanged zeolites: a computational and microcalorimetric investigation. J. Phys. Chem. 111, 5561–5567 (2007) 67. Hunger, M.: Bronsted acid sites in zeolites characterized by multinuclear solid-state NMR spectroscopy. Catal. Rev. Sci. Eng. 39, 345–393 (1997) 68. Stepanov, A.G.: Basics of solid-state NMR for application in zeolite science: material and reaction characterization. In: Sels, B., Kustov, L. (eds.) Zeolites and Zeolite-Like Materials, pp. 137–188. Elsevier, Amsterdam (2016) 69. Eichler, U., Brändle, M., Sauer, J.: Predicting absolute and site specific acidities for zeolite catalysts by a combined quantum mechanics/interatomic potential function approach. J. Phys. Chem. B. 101(48), 10035–10050 (1997) 70. Kazansky, V.B., Seryk, A.I., Semmer-Herledan, V., Fraissard, J.: Intensities of OH IR stretching bands as a measure of the intrinsic acidity of bridging hydroxyl groups in zeolites. Phys. Chem. Chem. Phys. 5, 966–969 (2003) 71. Montanari, T., Bevilacqua, M., Busca, G.: Use of nitriles as probe molecules for the accessibility of the active sites and the detection of complex interactions in zeolites through IR spectroscopy. Appl. Catal. A Gen. 307, 21–29 (2006) 72. Freitas, C., Barrow, N.S., Zholobenko, V.: Accessibility and location of acid sites in zeolites as probed by Fourier transform infrared spectroscopy and magic angle spinning nuclear magnetic resonance. Johnson Matthey Technol. Rev. 62, 279–290 (2018) 73. Busca, G.: Acidity and basicity of zeolites: a fundamental approach. Microporous Mesoporous Mater. 254, 3–16 (2017) 74. Busca, G.: Location and accessibility of hydroxy groups in silicoaluminate porous materials as studied by IR spectroscopy. Curr. Phys. Chem. 2, 136–150 (2012) 75. Trombetta, M., Busca, G.: On the characterization of the external acid sites of ferrierite and other zeolites. J. Catal. 187, 521–523 (1999) 76. Armaroli, T., Bevilacqua, M., Trombetta, M., Gutièrrez Alejandre, A., Ramirez, J., Busca, G.: A study of the adsorption of aromatic hydrocarbons and 2,6-lutidine on H-FER and H-ZSM5 zeolites. Appl. Catal. A Gen. 220, 181–190 (2001) 77. Bevilacqua, M., Gutièrrez Alejandre, A., Resini, C., Casagrande, M., Ramirez, J., Busca, G.: An FT-IR study of the accessibility of the protonic sites of H-mordenites. Phys. Chem. Chem. Phys. 4, 4575–4583 (2002) 78. Bevilacqua, M., Busca, G.: A study of the localization and accessibility of Brønsted and Lewis acid sites of H-mordenite through the FT-IR spectroscopy of adsorbed branched nitriles. Catal. Commun. 3, 497–502 (2002) 79. Phung, T.K., Busca, G.: On the Lewis acidity of protonic zeolites. Appl. Catal. A Gen. 504, 151–157 (2015) 80. Bellamy, L.J., Hallam, H.E., Williams, R.E.: Infra-red spectra and solvent effects. Part 1.—X—H stretching frequencies. Trans. Faraday Soc. 54, 1120–1127 (1958) 81. Ramis, G., Busca, G., Lorenzelli, V.: Surface acidity of solid acids and superacids: a FT-IR study of the behaviour of titania doped

31 with phosphoric, sulphuric, tungstic and molybdic acids. In: Zecchina, A., Costa, G., Morterra, C. (eds.) Structure and Reactivity of Surfaces, pp. 777–786. Elsevier, Amsterdam (1989) 82. Trombetta, M., Busca, G., Rossini, S.A., Piccoli, V., Cornaro, U.: FT-IR studies on light olefin skeletal isomerization catalysis. Part II: the interaction of C4 olefins with H-ZSM5 zeolite. J. Catal. 168, 349–363 (1997) 83. Busca, G.: Bases and basic materials in industrial and environmental chemistry. A review of commercial processes. Ind. Eng. Chem. Res. 48, 6486–6511 (2009) 84. Busca, G.: Bases and basic materials in industrial and environmental chemistry. Liquid versus solid basicity. Chem. Rev. 110, 2217–2249 (2010) 85. Busca, G., Lorenzelli, V.: Infrared spectroscopic identification of species arising from reactive adsorption of carbon oxides on metal oxide surfaces. Mater. Chem. 7, 89–126 (1982) 86. Ramis, G., Busca, G., Lorenzelli, V.: Low-temperature CO2 adsorption on metal oxides: spectroscopic characterization of some weakly adsorbed species. Mater. Chem. Phys. 29, 425–435 (1991) 87. Morterra, C., Zecchina, A., Coluccia, S., Chiorino, A.: I.r. spectroscopic study of CO2 adsorption onto η-Al2O3. J. Chem. Soc., Faraday Trans. 1. 73(73), 1544–1560 (1977) 88. Montanari, T., Busca, G.: On the mechanism of adsorption and separation of CO2 on LTA zeolites: an IR investigation. Vibr. Spectrosc. 46, 45–51 (2008) 89. Villarreal, A., Garbarino, G., Riani, P., Finocchio, E., Bosio, B., Ramírez, J., Busca, G.: Adsorption and separation of CO2 from N2-rich gas on zeolites: Na-X faujasite vs Na-mordenite. J. CO2 Util. 19, 266–227 (2017) 90. Garbarino, G., Wang, C., Valsamakis, I., Chitsazan, S., Riani, P., Finocchio, E., Flytzani-Stephanopoulos, M., Busca, G.: Acidobasicity of lanthana/alumina catalysts and their activity in ethanol conversion. Appl. Catal. B Environ. 200, 458–468 (2017) 91. Montanari, T., Castoldi, L., Lietti, L., Busca, G.: Basic catalysis and catalysis assisted by basicity: FT-IR and TPD characterization of potassium-doped alumina. Appl. Catal. A Gen. 400, 61–69 (2011) 92. Hadjiivanov, K., Vayssilov, G.: Characterization of oxide surfaces and zeolites by carbon monoxide as an IR probe molecule. Adv. Catal. 47, 307–511 (2002) 93. Blyholder, G.: Molecular orbital view of chemisorbed carbon monoxide. J. Phys. Chem. 68, 2772–2777 (1964) 94. Lupinetti, A.J., FauGernot Frenking, S., Strauss, S.H.: Theoretical analysis of the bonding between CO and positively charged atoms. J. Phys. Chem. A. 101(49), 9551–9559 (1997) 95. Sthorozhev, P.Y., Yanko, V.S., Tsyganenko, A.A., RodriguezDelgado, G., Areán, C.O.: Isomeric states of polar molecules on ionic surfaces: electrostatic model and FTIR studies. Appl. Surf. Sci. 238, 390 (2004) 96. Salla, I., Montanari, T., Salagre, P., Cesteros, Y., Busca, G.: Fourier transform infrared spectroscopic study of the adsorption of CO and nitriles on Na-mordenite: evidence of a new interaction. J. Phys. Chem. B. 109, 915–922 (2005) 97. Kozyra, P., Salla, I., Montanari, T., Datka, J., Salagre, P., Busca, G.: FT-IR study of the adsorption of carbon monoxide and of some nitriles on Na-faujasites: additional insight on the formation of complex interactions. Catal. Today. 114, 188–196 (2006) 98. Tsyganenko, A.A., Kondratieva, E.V., Yanko, V.S., Storozhev, P.Y.: FTIR study of CO adsorption on basic zeolites. J. Mater. Chem. 16, 2358–2363 (2006) 99. Yost, A.J., Evans, P.E., Tanabe, I., Hao, G., Gilbert, S., Komesu, T.: Integrated experimental methods for the investigation of the electronic structure of molecules on surfaces. In: Rocca, M.A., Rahman, T.S., Vattuone, L. (eds.) Spriger Handbook of Surface Science, pp. 331–350. Springer, Cham (2020)

1

32 100. Garbarino, G., Bellotti, D., Finocchio, E., Magistri, L., Busca, G.: Methanation of carbon dioxide on Ru/Al2O3: catalytic activity and infrared study. Catal. Today. 277, 21–28 (2016) 101. Payne, S.H., McEwen, J.S., Kreuzer, H.J., Menzel, D.: Adsorption and desorption of CO on Ru(0001): a comprehensive analysis. Surf. Sci. 594, 240–262 (2005) 102. Kokes, R.: Characterization of adsorbed intermediates on zinc oxide by infrared. Acc. Chem. Res. 6, 226–233 (1973) 103. Tamaru, K.: Dynamic Heterogeneous Catalysis. Academic Press, New York (1978) 104. Bañares, M.: Operando methodology: combination of in situ spectroscopy and simultaneous activity measurements under catalytic reaction conditions. Catal. Today. 100, 71–77 (2005) 105. Vimont, A., Thibault-Starzyk, F., Daturi, M.: Analysing and understanding the active site by IR spectroscopy. Chem. Soc. Rev. 39, 4928–4950 (2010) 106. Ferri, D.: Toward operando infrared spectroscopy of heterogeneous catalysts. In: Teoh, W.Y., Urakawa, A., Ng, Y.H., Sit, P. (eds.) Heterogeneous Catalysts: Advanced Design, Characterization and Applications, pp. 311–338. Wiley-VCH, Weinheim (2021) 107. Tinnemans, S., Mesu, J., Kervinen, K., et al.: Combining operando techniques in W. Y. Teoh, Atsushi Urakawa Yun Hau Ng Patrick SitOne spectroscopic-reaction cell: new opportunities for elucidating the active site and related reaction mechanism in catalysis. Catal. Today. 113, 3–15 (2006) 108. Meunier, F.C.: The design and testing of kinetically-appropriate operando spectroscopic cells for investigating heterogeneous catalytic reactions. Chem. Soc. Rev. 39, 4602–4614 (2010) 109. Aguirre, A., Collins, S.E.: Selective detection of reaction intermediates using concentration-modulation excitation DRIFT spectroscopy. Catal. Today. 205, 34–4035 (2013) 110. Aguirre, A., Collins, S.E.: Design of an optimized DRIFT cell/ microreactor for spectrokinetic investigations of surface reaction mechanisms. Mol. Catal. 481, 100628 (2020) 111. Busca, G.: Infrared studies of the reactive adsorption of organic molecules over metal oxides and of the mechanisms of their heterogeneously-catalyzed oxidation. Catal. Today. 27, 457–496 (1996) 112. Sanchez-Escribano, V., Larrubia Vargas, M.A., Finocchio, E., Busca, G.: On the mechanisms and the selectivity determining steps in syngas conversion over supported metal catalysts: an IR study. Appl. Catal. A Gen. 316, 68–74 (2007) 113. González Carballo, J.M., Finocchio, E., García-Rodriguez, S., Ojeda, M., García Fierro, J.L., Busca, G., Rojas, S.: Insights into

G. Busca the deactivation and reactivation of Ru/TiO2 during Fischer–Tropsch synthesis. Catal. Today. 214, 2–11 (2013) 114. Filot, I.A.W., van Santen, R.A., Hensen, E.J.M.: Quantum chemistry of the Fischer–Tropsch reaction catalysed by a stepped ruthenium surface. Cat. Sci. Technol. 4, 3129–3140 (2014) 115. Cornaro, U., Rossini, S., Montanari, T., Finocchio, E., Busca, G.: K-doping of Co/Al2O3 low temperature Fischer–Tropsch catalysts. Catal. Today. 197, 101–108 (2012) 116. Wang, X., Shi, H., Szanyi, J.: Controlling selectivities in CO2 reduction through mechanistic understanding. Nat. Commun. 8, 513 (2017) 117. Meunier, F., Reid, D., Goguet, A., Shekhtman, S., Hardacre, C., Burch, R., Deng, W., Flytzani-Stephanopoulos, M.: Quantitative analysis of the reactivity of formate species seen by DRIFTS over a Au/Ce(La)O2 water–gas shift catalyst: first unambiguous evidence of the minority role of formates as reaction intermediates. J. Catal. 247, 277–287 (2007) 118. Phung, T.K., Lagazzo, A., Rivero Crespo, M.A., Sanchez Escribano, V., Busca, G.: A study of commercial transition aluminas and of their catalytic activity in dehydration of ethanol. J. Catal. 311, 102–114 (2014) 119. Garbarino, G., Vijayakumar, R.P.P., Riani, P., Finocchio, E., Busca, G.: Ethanol and diethyl ether catalytic conversion over commercial alumina and lanthanum-doped alumina: reaction paths, catalyst structure and coking. Appl. Catal. B Environ. 236, 490–500 (2018)

Guido Busca is a Full Professor of Chemical Engineering at the Genova University. It is the author of a scientific book and of 480 full papers with more than 36000 citations (h index 97) in the fields of applied spectroscopy, chemical engineering, heterogeneous catalysis and adsorption technologies.

2

Case Studies: Infrared (IR) Spectroscopy Chiara Negri , Michele Carosso Silvia Bordiga

, Eleonora Vottero, Elena Groppo

, and

To Carlo, with love and gratitude

Contents 2.1

The FT-IR Experimental Setups . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2.2

Case Study 1: Formation of Cu Nitrates on Cu-CHA Catalyst by Operando FT-IR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of Temperature on the NO/O2 Reactivity on Cu-CHA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpretation of the CuII NO
3 Bands . . . . . . . . . . . . . . . . . . . . . . In Situ FT-IR Spectroscopy Monitoring Nitrates Formation with Isotopic Labelled 15NO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2.1 2.2.2 2.2.3 2.2.4 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.4

Case Study 2: Dynamic Behavior of Pt-Hydrides on a Pt/Al2O3 Catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pt-Hydride Species as a Function of the H2 Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Explaining the Dynamic Behavior of the Pt-H Species . . . . The Behavior of the Pt-Hydrides During a Hydrogenation Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35 36 37 39 41

Keywords 41 42

FT-IR spectroscopy · Cu-chabazite · Pt/Al2O3 · In situ · Operando

43 45 46

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Abstract

The chapter illustrates the role of FT-IR spectroscopy in clarifying the nature of the surface species formed on a catalyst in reaction conditions and their behavior as a function of the reaction variables (such as the temperature and the reactants’ concentration). In particular, two case studies have been selected, belonging to two different categories of heterogeneous catalysts: a Cu exchanged chabazite zeolite (Cu-CHA), which can be classified as a C. Negri · M. Carosso · E. Vottero · E. Groppo · S. Bordiga (*) Department of Chemistry, NIS Centre and INSTM, University of Torino, Torino, Italy e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]

single-site catalyst; and a Pt/Al2O3 catalyst, where the active Pt phase is in the form of nanoparticles and hence is an example of multisite catalyst. Taken together, the two case studies demonstrate the potentials of FT-IR spectroscopy in catalysis, provided that the experiments (either in situ or operando) are properly designed by tuning the reaction conditions.

This chapter illustrates the potentials of FT-IR spectroscopy in characterizing the reactive species formed at the catalyst surface in the presence of the reaction mixture, under different experimental conditions. In particular, two case studies have been selected, belonging to two different categories of heterogeneous catalysts: a Cu exchanged chabazite zeolite (Cu-CHA), which can be classified as a single-site catalyst; and a Pt/Al2O3 catalyst, where the active Pt phase is in the form of nanoparticles and hence is an example of multisite catalyst. For both systems, in situ and operando FT-IR experiments were performed, by adopting the same experimental setups, which are shortly described in Sect. 2.1. By comparing the results collected on the same catalyst using a different approach, it is hence possible to evaluate advantages and disadvantages of in situ and operando FT-IR measurements. The first case study, the Cu-CHA catalyst, is described in Sect. 2.2. In this case, FT-IR spectroscopy was used to investigate the formation, the nature, and the thermal stability of the Cu-nitrate species formed in the presence of the NO/O2 reaction mixture, which is an open debate in the recent literature. Initially, a series of operando FT-IR experiments

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_2

33

34

C. Negri et al.

performed in the presence of the NO/O2 mixture at fixed temperature (200  C or 50  C) are discussed (Sect. 2.2.2). These experiments allowed detecting the presence of intermediate species (such as the nitrosonium ion NO+) and understanding that the reaction temperature affects the overall amount of adsorbed nitrates and influences their speciation. A second series of operando FT-IR experiments were then conducted at variable temperature (from 200 to 50  C or the opposite, Sect. 2.2.2). The obtained results confirmed that the behavior observed at constant temperature is reproducible and dependent on the temperature at which the catalyst is exposed to the reactive NO/O2 gas mixture. Finally, in situ FT-IR measurements in the presence of 15NO/O2 gas mixture were performed to definitely identify the structure of the adsorbed nitrates, clarifying the occurrence of a Fermi resonance effect. Section 2.3 is dedicated to the second case study, the Pt/Al2O3 catalyst. In this case, FT-IR spectroscopy was adopted to investigate the formation of Pt-hydrides in the presence of H2 and their dynamic behavior as a function of the hydrogen coverage, which is a topic of great relevance in the field of hydrogenation catalysis. Initially, a series of in situ FT-IR experiments were performed, which allowed identifying the presence of four different types of Pt-hydrides and suggested that their relative concentration is a function of the hydrogen coverage. To demonstrate this intuition, operando FT-IR experiments were designed ad hoc. Operando FT-IR spectra were collected in both transmission and DRIFT modes, by adopting different experimental setups, and the results were perfectly comparable, demonstrating their reproducibility. Finally, operando FT-IR measurements were also performed in the presence of the H2/toluene mixture, allowing to discriminate between active and spectator Pt-hydride species. The two approaches substantiate and support each other, making FT-IR spectroscopy, alone, capable to unravel the nature of the reactive species and their behavior in different experimental conditions.

2.1

Fig. 2.1 The three experimental setups employed for performing the FT-IR experiments are described in the following. Part (a) The homemade quartz cell employed for in situ FT-IR measurements in transmission mode. Part (b) The AABSPEC cell used for operando FT-IR

experiments in transmission mode. Part (c) The DRIFT cell employed at the BM23 beamline (ESRF, Grenoble) for operando DRIFT/XAS/MS experiments

The FT-IR Experimental Setups

Figure 2.1 shows the cells and the experimental setups adopted for the in situ and the operando FT-IR measurements performed on the Cu-chabazite and Pt/Al2O3 catalysts described in the following sections. The in situ FT-IR experiments were carried out by adopting a homemade quartz cell equipped with KBr windows (Fig. 2.1a), which allows the sample activation in the 25–700  C temperature range and the successive measurements at room temperature, either in vacuum or in the presence of gases (at pressures from 10
4 to 760 Torr). The sample was pressed in the form of thin selfsupporting pellet (ca. 10 mg) without any dilution and placed inside a gold envelope. Although the total cell volume is of ca. 100 mL, the optical path of the cell is of ca. 0.5 cm, hence optimized for observing phenomena occurring at the catalyst surface and for minimizing the contribution of the gas phase. For the operando FT-IR measurements, a commercial FT-IR reactor cell (AABSPEC, no. 2000-A multimode) was employed, which allows recording the FT-IR spectra under controlled temperature (in the 25–700  C range) and in the presence of a gas flow (Fig. 2.1b). Also in this case, the sample was measured in the form of a thin self-supported pellet (ca. 10 mg). The cell volume is of. ca. 45 mL, and the cell is designed in such a way that the gas flow crosses the catalyst pellet. It is worth noticing that the huge internal volume of the cell should be taken into account for results interpretation under the point of view of gas flow dynamics. For these reasons, the cell should be better presented as a next-to-operando or quasi-operando cell rather than as an operando reactor. Finally, the operando DRIFT experiments on the Pt/Al2O3 catalyst were carried out at the BM23 beamline [1] of the European Synchrotron Radiation Facility (ESRF, Grenoble, France), in synchronous with XAS and MS. The experiments were performed using a low-volume cell (ca. 2 mL)

2

Case Studies: Infrared (IR) Spectroscopy

35

developed at BM23/ID24 for transient experiments (Fig. 2.1c) [2], equipped with a CaF2 IR-transparent window. The cell, which can be resistively heated in the 25–500  C range and can work in the presence of a gas flow, was mounted into a dedicated infrared optical system designed for DRIFT measurements. The Pt/Al2O3 catalyst (ca. 5 mg) was loaded in powder form into a cylindrical channel in the sample holder (ca. 1 mm in diameter). Two thin glassy carbon windows were employed to ensure that the X-ray beam passes through the sample. The dimension of the channel holder was selected to maximize the quality of the XAS spectra.

2.2

Case Study 1: Formation of Cu Nitrates on Cu-CHA Catalyst by Operando FT-IR

Copper ion-exchanged chabazite zeolites (Cu-CHA) have been studied intensively over the last decade, due to their outstanding performance in the selective catalytic reduction (SCR) of NOx with NH3 [3]. The NH3-SCR reaction is nowadays one of the most important technologies to abate NOx emissions from diesel engines and involves the reduction of NO with NH3 in the presence of O2 to release N2 and H2O. In this context, the catalytic activity of Cu-CHA is associated with the capability of Cu ions to reversibly change their oxidation state, depending on the reactants and conditions. CuII is reduced to CuI while releasing N2 and H2O, and the cycle is closed by re-oxidizing CuI with O2 and NO. Notwithstanding many in-depth studies about the nature and location of CuII (present as a mixture of Z[CuII(OH)]I/Z Fig. 2.2 Part (a) Schematic representation of the possible structures of surface NO
3 species. Part (b) Pictorial representation of the structure of different CuII NO
3 complexes involving CuII ions in the chabazite 8-member ring: chelating bidentate (left), bridging bidentate (middle), and monodentate (right). Cu: green, Si: white, Al: yellow, O: red and N: blue. The labels νNO2 and νN¼O indicate the corresponding vibrational modes observed and discussed in this work

a)

[CuII(O2)]I and Z2CuII, where Z represents a framework negative charge) and CuI sites [4], some aspects of the NH3-SCR mechanism over Cu-CHA are still debated. In particular, the formation of different CuII–(N,O) species (Cu nitrate and Cu nitrite) by reaction with NO/O2 and/or NO2 is still an intriguing research playground [5–8]. To obtain more insights in the formation and structural characterization of the CuII-(N,O) species that is formed by oxidation in a mixture of NO and O2, the technique that is most commonly applied is FT-IR spectroscopy. In fact, at variance with other characterization methods such as adsorption methods (TPD or TPSR) [9] or spectroscopic methods as X-ray absorption spectroscopy or EPR spectroscopy [10, 11], FT-IR provides direct information on the identity of the surface species. In spite of that, there is no common agreement on the type of nitrogen-oxo species that are formed when exposing the Cu-CHA catalyst to NO/O2 or NO2 [6, 7, 12–14]. The difficulties in assigning unambiguously the observed spectroscopic features to a specific CuII-(N,O) species are related to many reasons, among which: [15, 16] 1. A large variety of (N,O) compounds can be formed and coexist under the same reaction conditions. 2. Different nitrogen-oxo species and/or different configuration of the same species give absorption bands in the same spectral region (1650–1500 cm
1). 3. The bands rising below 1300–1200 cm
1 are overshadowed by the zeolites skeleton vibrational modes. 4. The same nitrogen-oxo compound can bond to the metal cation in different configuration (see Fig. 2.2a), resulting in a superimposition of vibrational bands.

O

O

O

O

N

N

N O M

O

O

O

M

M

M

b)

nN=O

nNO2 nN=O

5Å

O

2

36

Nevertheless, some (N,O) compounds can be ruled out, considering: (I) the reaction conditions, in terms of temperature, pressure, gas compositions (e.g., the excess of oxygen in the reaction environment favors the formation of nitrates on nitrites); (II) the relative stabilities to changes in gas composition and/or temperature (e.g., charged species usually cannot be removed at room temperature); (III) the reactivity toward specific reactants (e.g., nitrates can react with NO to give nitrites, lowering the intensity of their bands); and (IV) exchanges with isotopic labelled molecules (e.g., can define which atoms are involved in the observed vibrations) [15–17]. All these elements should be examined, when assigning the absorption bands observed after the exposure of the catalyst to the selected gas mixture. Considering the reaction conditions, exposing the Cu-CHA catalyst to NO in excess of O2 results in the formation of nitrates. Generally, surface nitrates have a C2v symmetry, resulting in three IR active stretching modes regardless of the assumed configuration (Fig. 2.2a) [16]. The monodentate is characterized by three single N-O bonds, one connected to the cation (N-O) and two equivalent ones (O-N-O) (right structure in Fig. 2.2a). The absorption bands are reported in the 1530–1480 and 1290–1250 cm
1 range, respectively for the asymmetric and symmetric stretching modes of monodentate nitrates (νasNO2 and νsymNO2) [15, 16, 18]. The νN-O stretching falls in the lower energy region, which is covered by the zeolite framework vibrations and therefore cannot be detected [19]. Conversely, bidentate nitrates can form chelating or bridging structures, linked to one or two metal ions, respectively (left and middle structures in Fig. 2.2a). These structures present an identical symmetry for the NO
3 ligand, resulting in one stretching vibration for the N¼O group (νN¼O), and symmetric and asymmetric stretching modes for the O-N-O group (νsymNO2 and νasNO2) binding one or two Mn+ ions. Both the νsymNO2 and νasNO2 modes fall in the frequency region of the intense zeolite framework vibrations (below 1300 cm
1). It is worth remembering that the split of the free D3h NO
3 ion stretching mode once adsorbed on the catalyst increases progressively from monodentate to bridging bidentate nitrates [16]. The study of nitrates on zeolites can thus only focus on the νsymNO2 and νasNO2 vibrations for monodentate nitrates and on the νN¼O vibrations of the bidentate configuration. Generally, the modes ascribed to bridging structures are reported to be in the 1650–1600 cm
1 range and those of the chelating ones at 1585–1500 cm
1 [15, 16, 20, 21]. Recently, these assignments were rediscussed on the basis of theoretical calculations, proposing a higher frequency for the chelating structures with respect to bridging ones, in relation to a more strained O-N-O angle and shorter N¼O distance in the former [22]. This might imply that FTIR spectroscopy and the νN¼O frequency cannot be used alone with the aim of discriminating between the two bidentate

C. Negri et al.

structures involving one or two CuII ions, eventually formed on Cu-CHA in this specific case study. Support to the assignment can come from other techniques, as X-ray absorption spectroscopy, more sensitive to the structural coordination of copper ions. Recently with this technique, on the basis of EXAFS evidences, CuII-(N,O) species formed upon interaction of NO/O2 (or NO2 alone) at 200  C on a Cu-CHA zeolite were assigned to bidentate chelating nitrates [6, 14]. Nonetheless, in the following, we will demonstrate that actually FT-IR spectroscopy can be sufficient to determine the type of nitrates formed on the catalyst. With this aim, FT-IR in operando conditions was used to investigate nitrates formation on a Cu-CHA zeolite as a function of the reaction temperature (at 200  C and at 50  C). Furthermore, the same reaction has been studied at room temperature by in situ FT-IR spectroscopy, exposing the catalyst to 15NO and O2, to get closer insight on the nature of some of the observed experimental spectral features.

2.2.1

The Catalyst

The case study analyzed in this chapter reports the measurements performed on a Cu-CHA catalyst with Si/Al ¼ 15, Cu/Al ¼ 0.5 (CHA15_05). The chabazite with Si/Al of 15 is prepared tuning suitably the synthesis gel composition (1.0 SiO2/0.033 Al2O3/0.5 TMAdaOH/0.5 HF/3 H2O). The gel is prepared by dissolving aluminum isopropoxide (98%, SigmaAldrich) in tetraethyl orthosilicate (98% Aldrich) and adding N,N,N-trimethyladamantammonium hydroxide (TMAdaOH, 25 wt %, Sachem) to the solution. This mixture is then stirred to homogenize, and hydrofluoric acid (48 wt %, 99.99% tracemetal basis, Sigma-Aldrich) is added; the mixture is then dried at 60  C and stirred by hand until the desired water content is obtained. Subsequently, the gel is heated to 150  C for 3 days under rotation in a Teflon-lined autoclave. The CHA product is recovered by filtration, washed several times with water, and calcined at 580  C for 3 h to remove the TMAdaOH. To prepare the corresponding Cu-CHA zeolites, the required amount of copper (II) acetate monohydrate (Sigma-Aldrich, 99.99%) is dissolved in water, and the proton form of the zeolite is added to the solution (250 ml per zeolite gram). The solution is stirred at room temperature for 24 h, and then the obtained copper zeolite is filtered, washed, and dried at 100  C overnight; then it is calcined at 500  C for 3 h to remove the residual acetate. In all the reported experiments, the catalyst was first treated in O2 at 400  C for 60 min (heating rate 5  C min
1), leading to a mixture of roughly 70/80% of Z[CuII(OH)]I/Z [CuII(O2)]I and 30/20% of Z2-CuII ions [23]. Subsequently, it has been cooled to 200  C or to 50  C (cooling rate 3  C min
1) in O2 and exposed to a mixture of 1000 ppm NO and 10% O2 in N2.

2

Case Studies: Infrared (IR) Spectroscopy

2.2.2

37

Effect of Temperature on the NO/O2 Reactivity on Cu-CHA

To investigate nitrate formation on the selected copper chabazite catalyst, we applied operando FT-IR spectroscopy, measuring the spectra in transmission mode. The use of an operando setup allowed investigating the catalyst reactivity in real time, monitoring the species adsorbed on the surface while they are forming. Furthermore, we were capable of working under controlled conditions in terms of temperature and gas flow concentration and rate.

Operando FT-IR Spectroscopy at Fixed Temperature Figure 2.3a shows the evolution of the FT-IR spectra in the range of 3900–1300 cm
1 for CHA15_05 in a NO/O2 flow at 200  C. After the activation treatment, the sample was cooled to 200  C in pure O2 (grey) to preserve the CuII oxidation state before dosing the NO/O2 mixture. The initial spectrum

a) c)

1625

3800 3600 3400

#

3800 3600 3400 C

A B 0.1

#

0.05

1500

Absorbance (a.u.)

1625

1610

d) f)

0.1

3544

e)

3670

QOH

*

* C

#

1310

0.1

0.05

1575 1610

0.1

3655

b)

1575

Absorbance (a.u.)

QOH

3600 3200 2800 2400 2000 1600 1800 1650 1500 1350 Wavenumbers (cm–1)

Fig. 2.3 Operando FT-IR spectra measured during exposure of Cu-CHA15_05 to 1000 ppm NO/10% O2 after pretreatment in O2 (grey curve), at 200  C, from dark to light blue curves (part a), and at 50  C, from dark to light blue curves (part d). Parts (b) and (e) Magnifications of the ν(OH) region. Parts (c) and (f) Magnifications and of the CuII-(N,O) regions. The spectra for the CuII-(N,O) region were corrected for the background by subtraction of the spectrum for the zeolite before NO/O2 admission. See text for the meaning of the * and # symbols

shows the typical signals of surface OH species in the ν(OH) stretching region (3800–3300 cm
1). The four main bands observed in ν(OH) region are at 3732 (isolated Si-OH), 3604 and 3582 (CHA Brønsted sites), and 3655 cm
1, assigned to the ν(OH) mode of the Z-[Cu(OH)]I site [24]. The presence of the intense signals due to Si(OH)Al Brønsted sites points to a not complete exchange of the available Brønsted sites with Cu, in agreement with the Cu/Al ratio of about 0.5. The 2100 and 1300 cm
1 region is characterized by the typical overtone and combination modes of the TO4 zeolite framework vibrations (T representing a Si or Al framework atom), at 1998, 1868, 1682, and 1542 cm
1. The lower-frequency region (not shown) is dominated by the intense zeolite framework modes. Dosing the 1000 ppm NO/10% O2 mixture at 200  C (from dark to light blue curves), results in changes in the entire spectral region. In the ν(OH) and overtone/combination regions, the band at 3655 cm
1 grows and reaches a steady state after 3 min in NO/O2 (see Fig. 2.3b), while a very weak band at 2159 cm
1 (labelled with * in Fig. 2.3a) grows progressively, indicating the formation of nitrosonium NO+ ion exchanging Brønsted sites (see below) [25–28]. Furthermore, a slight vertical shift of the zeolite spectrum is observed. In the 1800–1300 cm
1 range, three peaks at 1625, 1610 (shoulder), and 1575 cm
1 appear, which can be assigned to CuII NO
species [6, 14, 28, 29]. The 3 extremely weak signal at 1380 cm
1 (labelled with # in Fig. 2.3c) is due to the formation of a very small amount of water molecules, hydrogen bonded to the Brønsted sites. These traces of H2O are also responsible for the growth of the band at 3655 cm
1 (see below). Lowering the NO/O2 exposure temperature to 50  C promotes major changes in the FT-IR spectra (Fig. 2.3d). Still, the spectrum measured for the Cu-CHA pretreated in O2 at 400  C and subsequently cooled to 50  C in O2 (grey) is comparable to the one measured at 200  C, with no observable traces of water contamination. However, more evident changes can be seen when dosing 1000 ppm NO/10% O2 at 50  C, if compared to those observed at 200  C. The main differences between the experiments carried out at 200 and 50  C can be summarized as follows: first, we observed changes in the ν(OH) region, where the bands at 3670 cm
1 and 3544 cm
1 grow constantly, with the parallel consumption of Brønsted signals (Fig. 2.3e). Furthermore, we can note the development of broad signals centered at 2884, 2416, and 1670 cm
1 (labelled in Fig. 2.3d as A, B, and C, respectively) and of the 2159 cm
1 band (labelled with *), which reaches its maximum after the first minutes (mauve curve), to progressively decrease in time. The three bands at 1625, 1610, and 1575 cm
1 (magnified in Fig. 2.3f) assigned to CuII NO
3 species increase with time, reaching a final intensity significantly higher when compared to the

2

38

C. Negri et al.

reaction performed at 200  C. Finally, some weaker bands are also visible at 1500, 1380 (labelled with #), and 1310 cm
1. The A, B, and C triad of bands and the band at 1380 cm
1 (labelled with #) are fingerprints of highly perturbed OH groups involved in strong hydrogen bond. The appearance of these bands in the spectrum points to the presence of water molecules, interacting with the zeolites Si(OH)Al Brønsted sites [30, 31]. These H2O molecules are formed when NO reacts in the presence of O2 and H+ Brønsted sites to form NO+ ions (band at 2159 cm
1, labelled with *), which are consequently replacing the reacted Brønsted sites [25–28, 32]. According to the mechanisms proposed by Hadjiivanov and by Henriques et al., the reaction involves a transfer of the NO unpaired electron, located in an antibonding π* orbital, with the formation of water (2NO + 1=2 O2 þ 2Olattice–H+ Ð H2O þ 2Olattice–NO+) [25, 26, 32]. Thus, the ABC triplet arises from a Fermi resonance phenomenon, involving the ν(OH) mode in [zeoliteO-H∙∙∙Owater] complexes (labelled as (a) in Fig. 2.4), with the overtone of the in-plane bending mode (2δ[zeoliteO-H∙∙∙Owater]) giving the AB diad and the overtone of the out-of-plane bending mode (2δ[zeoliteO-H∙∙∙Owater]) giving the C band. Contributing to the B band is also the second overtone of the out-of plane bending, coupled with the stretching. The weak band at 1380 cm
1 (labelled with # in Fig. 2.3c, f) is assigned to the δ[zeoliteO-H∙∙∙Owater], labelled as (a) in Fig. 2.4. The bands around 3670 and 3544 cm
1are assigned to ν(OH) stretching modes of slightly perturbed H2O molecules (labelled as (b) and (c) in Fig. 2.4). This indicates that in the tested conditions, (H2O)n/H+ adducts are formed, with n  1. It is worth noticing that the same features (though much weaker, in agreement with the lower intensity of the NO+ band) are observed at 200  C (see Fig. 2.3c, where the 1380 cm
1 peak is visible and marked with #). Comparing the changes observed for the 1625, 1610, and 1575 cm
1 bands when dosing NO/O2 at 50  C or at 200  C

H (b) O

(c) H

H (a) O Si

O Al

Si

Fig. 2.4 Schematic representation of a zeolite Brønsted site involved in a strong hydrogen bond with an H2O molecule

allows a better understanding of the properties and structure of the formed CuII NO
3 species. On the basis of the wave number positions, we would be tempted to assign the higherfrequency band to bridging bidentate nitrates and the two at lower frequencies to chelating bidentate nitrates. However, as reported in Sect. 2.2.1, the three bands at 1625, 1610, and 1575 cm
1 have been previously assigned to chelating bidentate structures, involving one CuII ion (Fig. 2.2b, left) on the basis of X-ray absorption spectroscopy data analysis [6, 14]. Furthermore, the data in Fig. 2.3f show that the intensity ratio 1625/1610 cm
1 is high at the beginning but gradually decreases, leading to a final state where the relative intensity of the two bands is inverted with respect to what is observed at 200  C. The well-defined peak at 1625 cm
1 observed in the experiment performed at 200  C appears as a defined shoulder of the 1610 cm
1 band in the experiment performed at 50  C. This observation seems to indicate a dependence on the temperature of the two observed bands, possibly pointing to the presence of more than one chelating nitrate. Moreover, a third species virtually absent at 200  C is formed at 50  C, which bands at 1500 and 1310 cm
1 are assigned to monodentate CuII NO
3 complexes (right structure in Fig. 2.2b), on the basis of their position and their low thermal stability (see below). As a preliminary conclusion, lowering the reaction temperature affects the overall amount of adsorbed nitrates (compare the bands intensity in Fig. 2.3c, f) and influences their speciation. This observation is in agreement with the relatively low thermal stability of nitrates, which follows the order monodentate < chelating bidentate < bridging bidentate [16].

Operando FT-IR Spectroscopy at Variable Temperature To further demonstrate the dependency on temperature of the stability of the diverse CuII-(N,O) species observed in the experiments conducted at fixed temperature (200 vs 50  C), FT-IR spectra were measured continuously, both during cooling from 200 to 50  C (Fig. 2.5a) and during heating from 50 to 200  C (Fig. 2.5b), maintaining the atmosphere of 1000 ppm NO/10% O2. When decreasing the temperature (Fig. 2.5a), a general broadening of the spectra is observed, as a consequence of the formation of strongly perturbed hydroxyls groups (Fig. 2.4), with observable fingerprint around 1380 cm
1 (labelled as #). The gradual growing of these features is accompanied in the whole range (not reported) by the evolution of the ν(OH) absorption bands with formation of the AB diad, and by the appearance of the NO+ feature at 2159 cm
1. Furthermore, we can observe the progressive formation of monodentate nitrates with bands at 1500 and 1310 cm
1, once the temperature is sufficiently low (below 150  C). The final state (orange in Fig. 2.5a) is very similar to that

2

Case Studies: Infrared (IR) Spectroscopy

a)

39

b)

1610 1625

1575

2

1625 1500

1610

# 1310 50

200

§ 200 1800

1700

1600

1400

1500

1300

1575

1610

50 1800

1700

–1)

§

1500

#

1600

1500

1400

Wavenumbers

Wavenumbers (cm

1310 1300

(cm–1)

Fig. 2.5 Operando FT-IR spectra measured during exposure of Cu-CHA15_05 to 1000 ppm NO/10% O2 after pretreatment in O2, while decreasing temperature, from blue to orange curves (part a), and

while increasing temperature, from orange to blue curves (part b). The spectra were corrected for the background by subtraction of the spectrum for the zeolite at 200  C before NO/O2 admission

obtained after dosing NO/O2 at 50  C for 30 min (light blue in Fig. 2.3f). Furthermore, we can see that the intensity of the band at 1625 cm
1 reaches a maximum, and then it slightly decreases, while the bands at 1610 and 1575 cm
1 grow continuously, resulting in a lower I(1625)/I(1610) ratio. The same trend, reversed, is observed when increasing the temperature (Fig. 2.5b). In this case, it is possible to follow the progressive disappearance of the bands present only at low temperature. Specifically, the bands associated to monodentate nitrates diminish until the species is no longer adsorbed on Cu ions, confirming their thermal stability below 150  C, while the 1380 cm
1 band, fingerprint of perturbed OH groups, disappears constantly when increasing the temperature. In parallel, we can observe that the nitrate band at 1625 cm
1 is progressively restored, as well as the relative intensity ratio with the 1610 cm
1 band. As in the previous case, the final state at 200  C (blue in Fig. 2.5b) closely resembles the spectrum obtained after 30 min in NO/O2 flow at 200  C (light blue in Fig. 2.3c). The overall intensity of the bands observed in the spectrum is lowered with respect to lower temperature, as previously observed. These findings are in agreement with the reported thermal stability of the involved species. The data collected in these experiments clearly show that the observed behavior is reproducible and dependent on the temperature at which the catalyst is exposed to the reactive NO/O2 gas mixture. Furthermore, recording the spectra in operando condition is of a fundamental importance to unravel the behavior of the species adsorbed on the catalyst, aiming to rationalize the changes observed. Following the evolution with temperature points to a no correlation between the decrease of the band at 1625 cm
1 and the growth of

monodentate copper nitrates, as they start to be formed (or consumed) independently from the decrease (or growth) of the 1625 cm
1 peak. These findings confirm that the three bands at 1625, 1610, and 1575 cm
1 are related to the νN¼O vibration of different families of bidentate chelating nitrates (left structure in Fig. 2.2b).

2.2.3

Interpretation of the CuII NO
3 Bands

On the basis of the considerations reported, we can try to interpret the bands observed in the presence of NO/O2 in the 1650–1550 cm
1 range in terms of two distinct chelating nitrates, structurally similar to the left one of Fig. 2.2b. As highlighted previously, CuII ions can be present in Cu-CHA as Z-[Cu(OH)]I (requiring only one vicinal framework Al for charge compensation, denoted as Z) and/or as Z2-CuII, stabilized by two framework negative charges. The investigated sample is estimated to have roughly 70/80% of Z-[Cu(OH)]I/ Z-[CuII(O2)]I and the remaining Cu ions as Z2-CuII. Based on this, the observed behavior for the bidentate bands could be at first interpreted in terms of Z-[Cu(OH)]I and Z2-CuII sites. However, it has been recently reported that nitrates are hardly formed on a copper chabazite with Si/Al ¼ 5, hosting most of CuII ions in the Z2-CuII sites, under similar conditions to those applied in the reported experiments [33]. This allows us to exclude a contribution of Z2-CuII sites to the spectroscopic multiplicity of chelating nitrates. The speciation of copper sites can also be discriminated on the basis of spatial distribution in the chabazite cage. Recent studies have shown that CuI sites can be located in the 6- and 8-member rings (6r and 8r) of the zeolite framework, with a

40

parallel with the main CuII NO
3 peaks. The band is well defined and supports the absence of bridging bidentate nitrates structures in the zeolite. Thus, the three bands at 1625, 1610, and 1575 cm
1 are all related to the νN¼O vibration of bidentate chelating nitrates (left structure in Fig. 2.2b) differently affected by interaction with the chabazite environment. This assignment has been done based on a careful analysis of the intensity “minimum” between the bands at 1610 and 1575 cm
1 (1585 cm
1, indicated by § in Figs. 2.5 and 2.7). This is typically a

vasNO2 + vN=O

O

Absorbance (a.u.)

N O

O CuII

0.01

2620

2600 2580 Wavenumbers (cm–1)

2560

Fig. 2.6 Operando FT-IR spectra in the combination mode region of nitrates, measured during exposure of Cu-CHA15_05 to 1000 ppm NO/10% O2 at 200  C after initial heating in O2 at 400  C – from dark to light blue curves

a)

b)

1570

0.2

0.1 1535

Absorbance (a.u.)

relative fraction dependent on the Cu/Al and Si/Al ratio of the Cu-CHA zeolites [34, 35]. The different νN¼O bands might then be assigned to chelating nitrates formed during the reaction of NO and O2 with Z-[Cu(OH)]I sites, located in 6r or 8r windows. However, DFT calculations demonstrated almost identical geometries (including O-N-O bond angles and N¼O distance) for chelating nitrates involving CuII ions in the two chabazite rings [14]. Moreover, similar set of bands are observed in zeolites with different topologies, such as Cu-BEA [28, 36] and Fe-MFI [27, 37, 38]. Thus, our propose is that the observed band multiplicity is due to a perturbation of the nitrate stretching mode (νN¼O) by the zeolite cage or by the presence of vicinal Brønsted sites. This could explain the relative decrease of the band at 1625 cm
1 when the temperature is lowered (Fig. 2.5a) and its reappearance when increasing the temperature (Fig. 2.5b). Furthermore, since reducing the temperature favors the formation of NO+, with parallel water evolution and 1:1 H2O/H+ adduct formation, a nitrate species interacting with vicinal Brønsted sites (which induces a shortening of the N¼O bond and/or straining of the O-N-O angle) [22] would be perturbed when H+ are consumed to produce and exchange NO+ ions and are involved in strong hydrogen bonds. These is confirmed by the observation of the evolution of the band at 1380 cm
1 (used as fingerprint of the presence of H2O/H+ adducts) and at 1625 cm
1 during cooling and heating. Looking at the spectra reported in Fig. 2.5, the correlation between the disappearance of the 1380 cm
1 band and the restoration of the 1625 cm
1 one is visible, once the Brønsted sites are no longer overshadowed by the presence of H2O/ NO+ moieties. In the same conditions, monodentate nitrates are also observed, showing the coexistence of different structures (all involving single CuII sites) at low temperature. No correlation could be observed between the decrease of the band at 1625 cm
1 and the growth of NO+ or monodentate nitrates, suggesting that these are parallel reactions, which are mainly affected by changes in the reaction temperature. Further support to this interpretation can be found in the works by Hadjiivanov [15] and Venkov [39], who showed the possibility of analyzing the nitrates overtone and combination modes region to discriminate the very similar bidentate structures. This is an experimental approach that should not be underestimated, as the bands that are superimposed in the 1700–1200 cm
1 region are characterized by separate signals in the overtones and combination mode region. This can provide useful help in assigning bands whose interpretation is not straightforward. In the examined case, the authors assigned a band around 2580 cm
1 to the combination mode of chelating nitrates (νasNO2 þ νN¼O) and one around 2620 cm
1 to that of bridged ones. These bands lie in a region with no other signals, so they can be easily detected. In all the experiments discussed so far, only a single weak band at 2590 cm
1 (Fig. 2.6) is observed, developing in

C. Negri et al.

1800

§ 1615

1700 1600 1500 1400 1300

1280

1470 #

1700

1600

1500

1400

1300

Wavenumbers (cm–1)

Fig. 2.7 Part (a) In situ FT-IR spectra measured during exposure of Cu-CHA15_05 to 15NO/O2 at RT after pre-treatment in O2 – from dark to light blue curves. The spectra were corrected for the background by subtraction of the spectrum for the zeolite before 15NO/O2 admission. Part (b) Comparison between the Cu-NO
3 spectrum after 30 min NO/O2 at 50  C and the Cu-15 NO
3

2

Case Studies: Infrared (IR) Spectroscopy

41

derivative signal, usually deriving from resonance phenomena, between vibrational modes of the same molecular groups. In the data presented in this book chapter, this phenomenon is observed for the AB diad of the ν(OH) (zeoliteOHOwater) mode previously described [30, 31]. Interestingly, this peculiar spectral shape seems to be a characteristic of nitrate species formed on zeolites [27–29, 36–38, 40], while it is not reported on other catalysts, as oxides [41–43]. To exclude a resonance effect involving δ(NO
3 Þ/δ(OH) modes, we performed isotopic exchange experiments with 15NO.

“minimum” in nitrate bands, which is still present, shifted, between the 1570 and 1535 cm
1 peaks. This is well depicted in Fig. 2.7b, where we can distinctively see that the overall shape of the spectrum has not been affected by the use of 15 NO. Thus, the reported isotopic exchange experiment with 15 NO/O
2 allowed us to exclude a resonance effect involving δ NO
3 /δ(OH) modes. However, we have had a further confirmation of our previous assignments, and we thus put forward the hypothesis that the two bands centered at 1610 and 1575 cm
1 are related to a single chelating nitrate.

2.2.4

2.3

In Situ FT-IR Spectroscopy Monitoring Nitrates Formation with Isotopic Labelled 15NO

The use of isotopic marked molecules in FT-IR spectroscopy is a powerful tool to interpret the absorption bands observed in an experimental spectrum. Isotopic exchange can help in resolving the structure of the species responsible for a given band and can be used to resolve some spectra phenomena as the Fermi resonance [16, 44]. It is worth remembering that, since this effect is related to the superimposition of a stretching mode and the in and/or out-of-plane bending mode of a given adduct adsorbed on the catalyst, the isotopic substitution would shift the involved bands, re-separating their vibrational modes in the spectrum. In the present case, the aim is to understand if the observed “minimum” between the nitrate bands at 1610 and 1575 cm
1, which seems surprisingly unaffected by any changes the system undergoes, might be related to a Fermi resonance effect. Figure 2.7 shows the spectra obtained when contacting the Cu-CHA15_05 zeolite in situ with a mixture of 15NO and O2 at room temperature. Upon dosing the reactive gas mixture, we can observe the appearance of the Cu nitrates bands between 1750 and 1250 cm
1. In the reported spectra are clearly distinguishable all the features previously assigned, only shifted in frequencies because of the use of the marked isotope: this effect is visible from the direct comparison between the final states reported in Fig. 2.7b. Thus, we can observe the bands of chelating bidentate nitrates that are now at 1570 and 1535 cm
1, while the bands of the monodentate nitrates are now found at 1470 and 1280 cm
1. The calculated isotopic shift ði ¼ ν=ν0 Þ for the N14O16 ! N15O16 is found at 1.025–1.020, respectively, for the two nitrates species, in perfect agreement with the values reported in literature [16, 44, 45]. The band at 1615 cm
1 can be assigned to 14 NO
3 nitrate species that can be present due to traces of 14 NO in the reactive gas mixture. We can also observe the presence of the band at 1380 cm
1, associated to the formation of H2O/H+ adducts. Nonetheless, from the reported spectra is evident that the isotopic exchange has had no effect on the observed

Case Study 2: Dynamic Behavior of Pt-Hydrides on a Pt/Al2O3 Catalyst

Pt-based heterogeneous catalysts are employed in a wide range of industrial processes involving hydrogen, ensuring high selectivity and conversion at relatively mild operating conditions [46–49]. The key step of all these processes is the homolytic dissociation of molecular hydrogen at the Pt surface, with the consequent formation of different Pt-hydride species, including surface, subsurface, and bulk hydrides, depending on the system and on the reaction conditions [50–53]. These atomic hydrogen species are directly involved in the hydrogenation (or hydrognolysis) of several organic substrates. For example, Pt-based catalysts play a fundamental role in the production of bulk chemicals, where hydrogenation reactions are used either for purification (e.g., in the production of purified terephthalic acid or for alkyne removal from alkenes) or for synthetic purposes, as well as in the production of fine and specialty chemicals and in environmental processes (e.g., hydrogenolysis of organohalogens or nitrate reduction from water). Although the first experimental evidences of hydrogen splitting over Pt date back to the early 1800 [54, 55], and despite the intense research activity by means of both experimental and theoretical approaches, both the nature of the Pt-hydrides and their relative concentration under different operating conditions are still a matter of discussion [56, 57]. Pt-hydrides can be characterized by means of adsorption methods (such as H2 chemisorption and H-TPD), structural methods (such as XRD and XAS), and vibrational methods (such as FT-IR and INS spectroscopies). Starting from the adsorption methods, H2 chemisorption measurements provide an estimation of the H/Pt ratio, allowing the quantification of the amount of adsorbed H2 [47]. However, the stoichiometry between the surface adsorption sites and the adsorbate must be known (the assumption of a H:Pt ¼ 1:1 stoichiometry is usually wrong) and phenomena other than H2 dissociation on the metal phase (such as spillover on the support) might lead to overestimations of the H/Pt ratio. Hydrogen TPD measurements provide information on the Pt-H bonding energies. Up to six different features have

2

42

been observed in H-TPD curves for polycrystalline Pt [58–60], suggesting the presence of at least six different adsorption sites characterized by different Pt-H bonding energies. The formation of Pt-hydrides can be indirectly proved by structural methods, such as XRD and XAS, which detect the elongation of the Pt-Pt bond due to the insertion of H [61–71]. However, these techniques do not differentiate the identity of the Pt-hydrides and are really effective only when bulk hydrides are formed, which is usually not the case for Pt particles. Finally, vibrational spectroscopies have the potential to discriminate between different Pt-hydride species, since the Pt-H vibrational frequencies depend on how hydrogen is coordinated to the metal, independently on the nature of the support. For example, according to surface science works conducted on well-defined platinum surfaces, the ν(Pt-H) of terminal Pt-H species falls in the 1800–2200 cm
1 range, while multifold coordinated Pt-H species absorb in the range from 500 to 1720 cm
1 [72–76]. Although FT-IR spectroscopy was adopted since the early 1960s [77–82] to investigate the behavior of Pt-supported catalysts in the presence of hydrogen, so far it has been largely overlooked, especially in comparison to the huge amount of works based on XAS. There are three main reasons behind this fact. 1. First of all, as a consequence of the relatively low dipole moment involved in the Pt-H bond, the ν(Pt-H) absorption bands are usually extremely weak. This fact complicates the detection of Pt-hydrides for those catalysts having a low metal dispersion (i.e., a low metal surface area) and/or when the FT-IR spectrum is dominated by the absorption features of the support. 2. In addition, only linear Pt-H species can be detected by FT-IR spectroscopy, since the vibrational features of multi-coordinated Pt-H species fall in the region usually obscured by the framework modes of the majority of support materials [77–80]. 3. Finally, several works in the past pointed out the possibility that at least a fraction of the observed bands in the FT-IR spectrum of a Pt-based catalyst in the presence of H2 could be due to CO impurities [79, 82, 83]. As a matter of fact, H2 is commonly produced by the water-gas shift reaction, where CO is one of the reagents, and Pt-carbonyls absorb in the same range of frequencies as Pt-hydrides, but are characterized by a much higher extinction coefficient. At least part of the limitations of FT-IR spectroscopy in detecting Pt-hydrides on Pt-based heterogeneous catalysts can be overcome by incoherent inelastic neutron scattering (INS) spectroscopy. The transparency of the supports to the neutron beam combined with the selectivity to hydrogen

C. Negri et al.

allows the determination of the hydrogen-related properties on macroscopic amounts of samples, still keeping surface sensitivity due to the large surface area of the support [84]. For these reasons, INS has been used since 1990s for investigating Pt-based hydrogenation catalysts in the presence of H2 [85–88], and recently, it demonstrated to be sensitive enough to provide information also on low-loaded Pt-based catalysts [89, 90]. Nevertheless, with respect to FT-IR spectroscopy, INS has an important limitation; that is the inability to follow the dynamic of Pt-hydrides in reaction conditions, since the measurements are usually conducted at very low temperature (20 K) to increase the signal/noise ratio. In the last decades, a large number of experimental and theoretical evidences demonstrated that the Pt clusters may undergo electronic and morphological reconstructions in hydrogenation conditions, as a function of the H-coverage and of the reaction temperature [63–65, 91–97]. Hence, if the purpose of the measurements is to monitor the dynamic changes undergone by the catalysts in reaction conditions, FT-IR spectroscopy becomes the technique of choice. In the following, we will demonstrate the potential of FT-IR spectroscopy in the detection and discrimination of Pt-hydrides on a highly dispersed 5 wt% Pt/Al2O3 catalyst [98]. Such catalyst was investigated by means of in situ FT-IR spectroscopy in transmission mode, operando FT-IR spectroscopy in transmission mode, and operando DRIFT spectroscopy, as a function of the hydrogen coverage. By comparing the obtained results, we will highlight analogies and differences of the in situ and operando approaches. The same catalyst was then investigated by means of operando DRIFT spectroscopy during the hydrogenation of toluene, allowing to discriminate between the active and the spectators Pt-hydride species.

2.3.1

The Catalyst

The case study discussed in this chapter involves a 5 wt% Pt/Al2O3 catalyst prepared in the Chimet S.p.A. laboratories. The catalyst was prepared following a depositionprecipitation method [99], using a high-surface area alumina as support (SSA ¼ 121 m2 g
1; pore volume ¼ 0.43 cm3 g
1). After preparation, the sample was water washed and dried at 120  C overnight. The estimated average Pt particle size is 1.4  0.4 nm, as determined by HR-TEM analyzing more than 700 particles, which is in good agreement with the nominal dispersion D ¼ 63% determined by means of H2/O2 titration [100]. For all the experiments, the catalyst activation and reduction were accomplished following two subsequent steps: (1) the catalyst was heated up to 120  C (heating rate 5  C/ min), either under vacuum (for in situ experiments) or under an inert flow (N2 or He, for operando experiments) and left at

2

Case Studies: Infrared (IR) Spectroscopy

43

this temperature for ca. 30 min, in order to eliminate most of the physisorbed water; (2) the catalyst was treated at 120  C in H2 atmosphere (100 mbar dosed twice, separated by evacuation, for in situ experiments) or in a H2/inert flow (for operando experiments). During this step, the platinum oxide phase is reduced forming water as a by-product, followed by formation of Pt-hydrides. As far as the operando measurements are concerned, several experiments were performed by changing the total flow rate (in the range of 10–50 mL/min) and the H2 concentration in the gas feed (2, 5, 10 mol%) [98].

2.3.2

Pt-Hydride Species as a Function of the H2 Concentration

In Situ FT-IR Spectroscopy in Transmission Mode Figure 2.8a shows the typical FT-IR spectrum of the Pt/Al2O3 catalyst after the in situ H2-reduction at 120  C (spectrum 1, black). The spectrum shows an intense and very broad absorption band at ca. 3500 cm
1, which is due to the ν(OH) of the hydroxyl groups at the alumina surface. Additional weak bands are observed at 2030, ca. 1975, and ca. 1740 cm
1, which are not present in the spectrum of a pure Al2O3 subjected to the same treatment. These bands, which are magnified in Fig. 2.8a’, are ascribed to linearly adsorbed Pt-hydride species [72–82]. Their persistence after evacuation at 120  C suggests that they are associated to strongly adsorbed Pt-hydrides. Curiously, these bands are still observed even after a prolonged treatment in dynamic vacuum at 120  C, although with a different relative intensity. In particular, the band at 2030 cm
1 almost doubles in

a) 2.0

1. On this highly dispersed Pt/Al2O3 catalyst and under the adopted experimental conditions, at least four types of linearly adsorbed Pt-hydrides are detected by in situ FT-IR spectroscopy, labelled as species I, II, III and IV in Fig. 2.8b. Species I is observed only in excess of H2, while all the others are strongly adsorbed at the Pt surface and are observed also in the absence of molecular H2. Similar bands were already reported in the literature [77, 78, 80, 81, 83], although the difference among the three strongly adsorbed Pt-H species was never clarified. 2. The relative abundance and the ν(Pt-H) of the four linear Pt-hydride species is a function of the hydrogen coverage. It is worth noticing that bands I–IV are quite weak, because the stretching mode of linear Pt-H species has a weak dipole moment, as anticipated at the beginning of this chapter. This is the reason why Pt-hydrides have been rarely observed by FT-IR spectroscopy, and their assignment has been often questioned [79, 82]. In particular, one of the most frequent criticisms is that these bands might be due to chemisorbed CO [83], which is a common impurity in the H2 feed. To confirm our assignment, an experiment was performed by

a') 1

Absorba nce (a .u.)

intensity (spectrum 2, blue, in Fig. 2.8a’), which is counterintuitive, since during the vacuum treatment, the amount of adsorbed hydride species should decrease. Upon dosing H2 at room temperature on the latter sample (Fig. 2.8b, from spectrum 2 to spectrum 3), a new band appears at 2115 cm
1, those at 2030 and 1975 cm
1 upward shift to 2041 and 1990 cm
1 and slightly decrease in intensity, and that at 1740 cm
1 remains almost unaltered. These observations lead to the following preliminary conclusions:

b)

0.2

b')

II

2

2

1.5

0.2

0.2 2

1.0

1

0.0

0.0 0.1

2100 2000 1900 1800 1700

III

1500 1400 1300 1200

I

0.5

IV

3

0.0 0.0 4000

3500

3000

2500

2000

1500

Wavenumbers (cm–1)

Fig. 2.8 Part (a) In situ FT-IR spectrum of the Pt/Al2O3 catalyst after the reduction step (spectrum 1, black) and the successive degassing at 120  C overnight (spectrum 2, blue). Part (a’) shows a magnification the ν(Pt-H) region. Part (b) The effect of dosing H2 at room temperature on

1000

2160

2080

2000

1920

Wavenumbers

1840

1760

1680

(cm–1)

the sample degassed at 120  C overnight up to the maximum coverage (spectrum 3, red). Part (b’) shows the spectrum obtained by using D2 instead of H2

2

44

C. Negri et al.

dosing deuterium after the reduction step in H2. This time, four bands are observed at 1515, 1465, 1422, and 1220 cm
1 (Fig. 2.8b’). The isotopic shift of ca. 1.39 with respect to bands I, II, III, and IV in Fig. 2.8b confirms the assignment to hydride species [78].

Operando FT-IR Spectroscopy in Transmission Mode The formation of the Pt-hydrides during the H2-reduction of the Pt/Al2O3 catalyst at 120  C was successively monitored by means of operando FT-IR spectroscopy in transmission mode [98]. The adoption of an operando experimental setup allowed us to monitor the reduction process and hence to observe the Pt-hydride species as soon as they are formed. Figure 2.9a (from black to blue spectra) shows, as an example, the FT-IR spectra collected during reduction of Pt/Al2O3 in a 10 mol% H2/N2 flow of 50 mL/min. The same four absorption bands at 2115, 2041, ca. 1990, and ca. 1740 cm
1 already observed during the in situ FT-IR experiments (Fig. 2.8a) rapidly grow in intensity and then reach a steady-state situation. Stimulated by the observation done during the in situ experiments that the relative proportion of the four Pt-hydride species changes as a function of the hydrogen coverage, we decided to follow also the dehydrogenation process. In order to do so, we switched the gas feed composition to pure N2, keeping constant the total flow rate (50 mL/min) and the temperature (120  C). The sequence of FT-IR spectra collected in these conditions is shown in Fig. 2.9a (from blue to yellow spectra), while Fig. 2.9b shows the change in intensity of the four Pt-H bands as a function of

a) a)

time. The band at 2115 cm
1 (species I ) rapidly disappears, confirming that it is associated with a very weakly adsorbed Pt-H species, which is stable only in the presence of H2. The other three bands, instead, evolve in a counterintuitive way as already observed in the in situ experiment (Fig. 2.8a’). In particular, the two bands at 2041 and 1990 cm
1 (species II and III) reach approximately the double of their original intensity after around 150 min, while the band at ca. 1740 cm
1 (species IV) remains almost unchanged. After reaching the maximum, all the three bands rapidly disappear. This differentiates the operando experiments from the in situ ones, where the Pt-H species were still observed even after degassing at 120  C overnight, and indicates that a gas flow is more effective than a vacuum in removing the adsorbed species by keeping the temperature constant. We performed similar experiments by varying either the H2 concentration during the reduction step (10 mol%, 5 mol%, and 2 mol% H2/N2) or the total flow rate (50 mL/min and 20 mL/min). The FT-IR spectra evolved qualitatively in the same way but with a different kinetic. In general, the higher the H2 concentration in the flow during the reduction step, the longer the dynamic, while the higher the total flow rate, the shorter the dynamic. As an example, Fig. 2.10 shows the evolution of the intensity of the Pt-H bands corresponding to species II and IV as a function of time for three different experiments. By keeping the H2 concentration constant during the reduction step (10 mol%) but increasing the total flow rate (from 20 mL/min to 50 mL/min), the evolution is the same but occurs much faster (Fig. 2.10a, b). A further increase in the dynamic of the spectral evolution is observed by

b)

Pt-H formation: 10%H2 Flow: 50 mL/min

0

50

Time (min) 100 150

200

250

0.2 H2 Intensity (a.u.)

III

IV II I

N2

I: 2115 cm–1 II: 2041 cm–1 III: 1990 cm–1 IV: 1740 cm–1

0.1

N2

H2 2200

2100

2000

1900

1800

in) (m e Tim

1700

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Number of spectra

Wavenumbers (cm–1)

Fig. 2.9 Part (a) Evolution of the operando FT-IR spectra (collected in transmission mode) as a function of time for the Pt/Al2O3 catalyst during the reduction step in a 10 mol% H2/N2 flow at 120  C (from black to blue) and the successive dehydrogenation step in N2 flow at 120  C (from blue

to yellow). The spectra are shown in the ν(Pt-H) region after subtraction of the spectrum of the catalyst at 120  C before reduction. Part (b) Evolution of the intensity of the four bands due to Pt-hydrides (species I–IV) as a function of time, for the FT-IR spectra shown in part (a)

2

Case Studies: Infrared (IR) Spectroscopy

45

a) Pt-H formation: 10%H2 Flow: 20 mL/min

0.1

I n t e n sit y (a .u.)

0.0

b)

Pt-H formation: 10%H2 Flow: 50 mL/min

0.1 0.0

c) Pt-H formation: 5%H2 Flow: 50 mL/min

0.1

II IV

0.0 0

40

80

120 160 200 Time (min)

240

280

Fig. 2.10 Evolution of the intensity of the two bands due to Pt-hydrides of species II and IV as a function of time during the dehydrogenation step, for three operando experiments performed in different conditions. Part (a) H2 concentration during the reduction step ¼ 10 mol% H2/N2, total flow rate ¼ 20 mL/min; part (b) H2 concentration during the reduction step ¼ 10 mol% H2/N2, total flow rate ¼ 50 mL/min; part (c) H2 concentration during the reduction step ¼ 5 mol% H2/N2, total flow rate ¼ 50 mL/min

keeping the total flow rate (50 mL/min) constant but decreasing the H2 concentration during the reduction step (from 10 to 5 mol%, compare Fig. 2.10b, c). The data summarized in Fig. 2.10 allow also to appreciate that in all the cases, the intensity of the band ascribed to species IV starts to decrease as soon as the band ascribed to species II increases in intensity. The reason for the different kinetics in the evolution of the Pt-H species during the dehydrogenation step is ascribable to the experimental setup. Indeed, the operando FT-IR cell (Fig. 2.1b) has a large dead volume, and the complete removal of hydrogen by the N2 flow requires time. With our experimental setup, we were able to decrease the hydrogen partial pressure in the cell very slowly, and this permitted us to follow the dynamic of the platinum hydrides as a function of the hydrogen coverage. In this respect, what might be considered as a defect of the experimental setup turned out to be an advantage, since it allowed us to realize that the Pt-H species dynamically evolve as a function of the hydrogen coverage [98].

Operando FT-IR Spectroscopy in DRIFT Mode As a confirmation of the hypothesis formulated above, when we repeated the experiment shown in Fig. 2.9 at the BM23 beamline (ESRF) by using the DRIFT cell illustrated in Fig. 2.1c, we were not able to observe the same evolution of the Pt-H bands: during the dehydrogenation step, all the four bands simply decreased in intensity up to vanish. We explained this observation by considering that the

dehydrogenation rate was too fast, due to the low volume of the cell, which corresponded to a too fast complete removal of H2 from the gas flow. Hence, we tried to prolong the duration of the dehydrogenation step, which was also the desired condition to collect in synchronous a significant number of good-quality XAS spectra. To this end, we added to the setup a dead volume filled with the same hydrogen amount used during the hydride formation step (20 mol %), which was left open during the dehydrogenation step at 120  C. In this way, a very small and declining amount of H2 was constantly stripped by He. As revealed by MS at the outcome of the cell (Fig. 2.11b), the H2 concentration in the feed suddenly dropped, followed by a very slow decrease, and fell down to zero only after closure of the dead volume. Figure 2.11a shows the evolution of the DRIFT spectra collected in these conditions, whereas Fig. 2.11c shows the intensity of the four bands due to Pt-hydrides (species I–IV) as a function of time. The evolution of the absorption bands attributed to the linear Pt-H species is even more complex than that reported in Fig. 2.9, as a consequence of the much slower dehydrogenation, but the overall behavior is the same. Band I rapidly disappears, whereas bands II and III increase gradually and finally double their original intensity. The relative intensity of bands II and III changes with time, and both of them rapidly disappear when closing the dead volume. Band IV does not change until the end when it disappears quickly together with bands II and III.

2.3.3

Explaining the Dynamic Behavior of the Pt-H Species

The experiments discussed above demonstrate that four types of linear Pt-hydrides are rapidly formed (and are detectable by FT-IR spectroscopy) when our Pt/Al2O3 catalyst is exposed to hydrogen at 120  C. These species differ in terms of their interaction strength. Their relative concentration changes as a function of time during the dehydrogenation step, but the phenomenon is observable only when the hydrogen concentration decreases very slowly. The evolution of our FT-IR spectra [98] can be explained with the theoretical model proposed by Mager-Maury et al. [92]. According to this model, a decrease in the hydrogen coverage causes a reconstruction of the Pt particles from a cuboctahedral morphology, covered mainly by n-fold coordinated hydrides, to a biplanar one, mostly covered by linear hydrides. During the reconstruction, the relative concentration of the linear hydrides (which are those visible by FT-IR spectroscopy) is predicted to increase at the expenses of the n-fold coordinated hydrides (invisible by FT-IR spectroscopy). This explains why bands II and III, which are due to strongly chemisorbed linear Pt-hydrides, always increase in intensity

2

46

C. Negri et al.

a)

b)

III

m/Z = 2

Counts

II

2E-12

1E-12

IV

I

in)

m e(

Intensity (K.M.)

c)

I II III IV

2

1

m

Ti 2200

2100

2000 1900 1800 Wavenumbers (cm–1)

0

1700

90

190

290 390 Time (min)

490

590

Fig. 2.11 Part (a) Evolution of the operando DRIFT spectra as a function of time for the Pt/Al2O3 catalyst during dehydrogenation step in He flow (20 mL/min) at 120  C. The reduction step preceding the dehydrogenation was accomplished in the presence of 20 mol% H2 in He at 120  C. The spectra are shown in the ν(Pt-H) region after subtraction of the spectrum of

the catalyst at 120  C before reduction. Part (b) Evolution of the signal corresponding to H2 (m/Z ¼ 2) as detected by the mass spectrometer placed at the outcome of the reaction cell. Part (c) Evolution of the intensity of the four bands due to Pt-hydrides (species I–IV) as a function of time, for the FT-IR spectra shown in part a

before disappearing, with a kinetic that depends on the speed with which hydrogen is removed from the gas feed. The abovementioned model also predicts the presence of hydrogen at the interface between the Pt particles and the support. In this respect, we assign the absorption band at ca. 1740 cm
1 (band IV) to Pt-hydrides in interaction with alumina. Species IV survives for prolonged time (i.e., also at low hydrogen coverage), suggesting that the hydrides at the Pt/Al2O3 interface are quite stable irrespective of the particle morphology. Our FT-IR data are also in good agreement with the threesite model proposed by the group of Koningsberger [66]. At the maximum hydrogen coverage, the Pt particles are solvated by weakly adsorbed “on-top” hydrides (species I) and n-fold coordinated hydrides (not visible by IR). When the H2 coverage is decreased, the “on-top” hydrides desorb and the n-fold coordinated hydrides are converted into strongly adsorbed “atop” species. On these bases, we can explain our spectroscopic evidences with the following desorption model:

2. Upon dehydrogenation at 120  C, the weakly adsorbed hydrides are immediately desorbed, and the overall amount of chemisorbed hydrogen decreases. A conversion of n-fold coordinated hydrides into strongly adsorbed linear hydrides is responsible for the enhancement in intensity of species II and III. Being the strongest adsorbed species, interfacial hydrides are not yet desorbed, and band IV is unaffected. 3. After having reached the maximum intensity, both linear and multi-coordinated hydrides are quickly desorbed, leading to a complete dehydrogenated Pt phase. Species IV also disappears, meaning that the hydride species at the particle-support interface are also removed. Concomitantly, the spherically shaped PtHx NPs evolve toward Pt NPs with a droplike morphology strongly anchored to the support.

1. At the maximum H-coverage, strongly adsorbed linear hydrides (species II and III), n-fold coordinated hydrides (not detectable by FT-IR spectroscopy), and weakly adsorbed linear hydrides (species I ) are co-present. The presence of atomic hydrogen species at the interface between the Pt nanoparticles and the support (species IV) indicates a weak interaction between the spherical PtHx particles and the Al2O3 surface.

This hydrogen-induced reconstruction model was validated by the analysis of the XAS data collected synchronously to the DRIFT spectra shown in Fig. 2.11 [98].

2.3.4

The Behavior of the Pt-Hydrides During a Hydrogenation Reaction

As a final step, we investigated the surface dynamics of the Pt nanoparticles in Pt/Al2O3 applying the same DRIFT/XAS/ MS approach described above, during the hydrogenation of

2

Case Studies: Infrared (IR) Spectroscopy

47

toluene to methylcyclohexane [98]. After the reduction step, toluene vapors from a saturator were stripped by the 20 mol% H2 in He flow (20 mL/min), and DRIFT/XAS/MS data were collected for ca. 1 h. In the adopted conditions, we were far from the stoichiometric toluene hydrogenation. Figure 2.12a shows the DRIFT spectra of the Pt/Al2O3 catalyst collected in the presence of H2 at 120  C and their evolution during the hydrogenation of toluene at the same temperature. Toluene conversion starts immediately (MS data in Fig. 2.12b) and proceeds at a constant rate for the whole time. Correspondingly, important changes are visible in the DRIFT spectra, which are better visualized in Fig. 2.12c: bands I and II disappear and band III fast increases in intensity. In contrast, band IV does not change. Then, the DRIFT spectra do not change anymore during the whole reaction. A reverse behavior is observed when toluene is removed from the reaction feed. These data provide important information on the types of Pt-hydrides directly involved in the hydrogenation reaction. Hydrides of type I are rapidly consumed and no more observed during the reaction, meaning that they directly participate to the reaction. This observation is perfectly in agreement with the weak nature of these hydrides, which are hence also the most available for hydrogenating an organic substrate. Apparently, also the strongly adsorbed hydrides of

type II are fast consumed in the presence of toluene. However, the observation of an isosbestic point at 2006 cm
1 in the sequence of DRIFT spectra suggests that band II is converted into band III, as already observed during the dehydrogenation (Fig. 2.11), substantiating the assignment of bands II and III to similar strongly adsorbed “atop” platinum hydrides. The difference between the two species might be searched in their surface environment: species III is more isolated than species II, and this explains why it is favored in the presence of toluene, which is a competitor for the adsorbing sites at the platinum surface. All in all, the only linear Pt-hydrides involved in the hydrogenation reaction appear to be those of type I. Nevertheless, it is important to notice that the XAS data collected synchronously to the DRIFT spectra indicate that the Pt nanoparticles are structurally and electronically stable during the whole reaction time (i.e., no changes are observed neither in the XANES nor in the EXAFS spectra). This is made possible only because the Pt nanoparticles remain solvated by hydrogen during the reaction. In other words, the greatest fraction of the Pt-hydrides, being not directly involved in the reaction, can be basically considered spectator species, whose role is that of retaining the Pt nanoparticles morphology, avoiding worsening in the catalytic performances.

b)

a)

2

III

Counts

1

IV

H2 MeCHex

0.01

Tol

1E-3 2.5

I II III IV

c) Intensity (K.M.)

2.0 H2/Tol

Time

(min)

II

1.5 1.0 0.5

H2 I 2200

2100

0.0 2000

1900

1800

Wavenumbers (cm–1)

Fig. 2.12 Part (a) Evolution of the operando DRIFT spectra as a function of time for the Pt/Al2O3 catalyst during hydrogenation of toluene to methylcyclohexane at 120  C. The reduction step preceding the reaction was accomplished in the presence of 20 mol% H2 in He. The spectra are shown in the ν(Pt-H) region after subtraction of the spectrum of the catalyst at 120  C before reduction. Part (b)

1700

0

10

20

30 40 50 Time (min)

60

70

Evolution of the signals corresponding to H2 (m/Z ¼ 2), toluene (m/Z ¼ 91) and methylcyclohexane (m/Z ¼ 55), as detected by the mass spectrometer placed at the outcome of the reaction cell. Part (c) Evolution of the intensity of the four bands due to Pt-hydrides (species I–IV) as a function of time, for the FT-IR spectra shown in part a

2

48

2.4

C. Negri et al.

Conclusions

This chapter illustrated the role of FT-IR spectroscopy in clarifying the nature of the surface species formed on a catalyst in reaction conditions, and their behavior as a function of the reaction variables (such as the temperature and the reactants concentration). As far as the Cu-CHA catalyst is concerned, we demonstrated that upon exposure of Cu-CHA to a NO/O2 gas mixture, two types of nitrates are formed, whose relative concentration is a function of the reaction temperature. Monodentate copper nitrates are stable at a temperature between room temperature and 150  C, while chelating bidentate copper nitrates are the main species above 150  C. The former are characterized by two well-defined bands at 1500 and 1310 cm
1, while the latter are characterized by a triad of bands at 1625, 1610, and 1575 cm
1. The reason for the presence of three bands is associated to the perturbation of nitrate signals by the zeolite structure, as clarified by performing an in situ FT-IR experiment with isotopic labelled 15NO. For the Pt/Al2O3 catalyst, four types of linear Pt-hydrides were detected by FT-IR spectroscopy, differing in terms of interaction strength. Their relative concentration changes with the hydrogen coverage as a consequence of a morphological reconstruction of the Pt particles. Operando FT-IR experiments performed in the presence of the toluene/H2 reaction mixture allowed understanding that only one out of the four hydride species are actually involved in the toluene hydrogenation, while the others behave as fundamental spectators, which allows maintaining the Pt nanoparticles hydrogen-solvated and protect them from deactivation phenomena. Taken together, the two case studies demonstrate the potentials of FT-IR spectroscopy in catalysis, provided that the experiments (either in situ or operando) are properly designed by tuning the reaction conditions. It is worth noticing that all the conclusions summarized above could be safely derived from the FT-IR results alone, although they have been supported by additional data obtained with complementary techniques. Acknowledgments M.C., E.V., and E.G. would like to thank their colleagues and friends who took part in the story of Pt/Al2O3, and in particular Riccardo Pellegrini (Chimet S.p.A.), Andrea Piovano and Monica Jimenez Ruiz (Institut Laue Langevin, ILL), Kirill Lomachenko (European Synchrotron Radiation Facility, ESRF), Sara Morandi and Maela Manzoli (University of Turin), and Andrea Lazzarini (University of Oslo). C.N. and S.B. are gratefully indebt to Gloria Berlier, Elisa Recchia, Michele Cutini (University of Turin) and to Ton V.W. Janssens and Peter S. Hamartin (Umicore Denmark ApS) for the fruitful contribution given in the study of Cu-NO
3 on Cu-CHA.

References 1. Mathon, O., Beteva, A., Borrel, J., Bugnazet, D., Gatla, S., Hino, R., Kantor, I., Mairs, T., Munoz, M., Pasternak, S., Perrin, F., Pascarelli, S.: The time-resolved and extreme conditions XAS (Texas) facility at the European Synchrotron Radiation Facility: the general-purpose EXAFS bending-magnet beamline BM23. J. Synchrotron Radiat. 22, 1548–1554 (2015) 2. Agostini, G., Meira, D., Monte, M., Vitoux, H., Iglesias-Juez, A., Fernandez-Garcia, M., Mathon, O., Meunier, F., Berruyer, G., Perrin, F., Pasternak, S., Mairs, T., Pascarelli, S., Gorges, B.: XAS/DRIFTS/MS spectroscopy for time-resolved operando investigations at high temperature. J. Synchrotron Radiat. 25(6), 1745–1752 (2018) 3. Beale, A.M., Gao, F., Lezcano-Gonzalez, I., Peden, C.H.F., Szanyi, J.: Recent advances in automotive catalysis for NOx emission control by small-pore microporous materials. Chem. Soc. Rev. 44(20), 7371–7405 (2015) 4. Borfecchia, E., Beato, P., Svelle, S., Olsbye, U., Lamberti, C., Bordiga, S.: Cu-CHA – a model system for applied selective redox catalysis. Chem. Soc. Rev. 47, 8097–8133 (2018) 5. Ruggeri, M.P., Luo, J.Y., Nova, I., Tronconi, E., Kamasamudram, K., Yezerets, A.: Novel method of ammonium nitrate quantification in SCR catalysts. Catal. Today. 307, 48–54 (2018) 6. Janssens, T.V.W., Falsig, H., Lundegaard, L.F., Vennestrom, P.N. R., Rasmussen, S.B., Moses, P.G., Giordanino, F., Borfecchia, E., Lomachenko, K.A., Lamberti, C., Bordiga, S., Godiksen, A., Mossin, S., Beato, P.: A consistent reaction scheme for the selective catalytic reduction of nitrogen oxides with ammonia. ACS Catal. 5(5), 2832–2845 (2015) 7. Colombo, M., Nova, I., Tronconi, E.: Detailed kinetic modeling of the NH3-NO/NO2 SCR reactions over a commercial Cu-zeolite catalyst for diesel exhausts after treatment. Catal. Today. 197(1), 243–255 (2012) 8. Marberger, A., Petrov, A.W., Steiger, P., Elsener, M., Krocher, O., Nachtegaal, M., Ferri, D.: Time-resolved copper speciation during selective catalytic reduction of NO on Cu-SSZ-13. Nat. Catal. 1(3), 221–227 (2018) 9. Ruggeri, M.P., Selleri, T., Colombo, M., Nova, I., Tronconi, E.: Investigation of NO2 and NO interaction with an Fe-ZSM-5 catalyst by transient response methods and chemical trapping techniques. J. Catal. 328, 258–269 (2015) 10. Godiksen, A., Vennestrom, P., Rasmussen, S., Mossin, S.: Identification and quantification of copper sites in zeolites by electron paramagnetic resonance spectroscopy. Top. Catal. 60(1–2), 13–29 (2017) 11. Mossin, S., Godiksen, A., Janssens, T., Falsig, H., Lundegaard, L., Vennestrom, P., Rasmussen, S., Moses, P.G., Giordanino, F., Borfecchia, E., Lomachenko, K., Lamberti, C., Bordiga, S., Beato, P.: Consistent reaction mechanism for the selective catalytic reduction of NO with NH3. Abstr. Pap. Am. Chem. Soc. 249 (2015) 12. Selleri, T., Ruggeri, M.P., Nova, I., Tronconi, E.: The low temperature interaction of NO + O-2 with a commercial Cu-CHA catalyst: a chemical trapping study. Top. Catal. 59(8–9), 678–685 (2016) 13. Tronconi, E., Nova, I., Novakova, J., Bernauer, M., Sazama, P., Sobalik, Z.: Mechanistic study of the NO+NH4NO3 reaction on Hand Fe/H-BEA zeolites using N-15 and O-18 labeled species. Top. Catal. 61(18–19), 1967–1973 (2018) 14. Tyrsted, C., Borfecchia, E., Berlier, G., Lomachenko, K.A., Lamberti, C., Bordiga, S., Vennestrom, P.N.R., Janssens, T.V.W., Falsig, H., Beato, P., Puig-Molina, A.: Nitrate-nitrite equilibrium in the reaction of NO with a Cu-CHA catalyst for NH3-SCR. Cat. Sci. Technol. 6(23), 8314–8324 (2016)

2

Case Studies: Infrared (IR) Spectroscopy

15. Hadjiivanov, K.: Use of overtones and combination modes for the identification of surface NOx anionic species by IR spectroscopy. Catal. Lett. 68(3–4), 157–161 (2000) 16. Hadjiivanov, K.I.: Identification of neutral and charged NxOy surface species by IR spectroscopy. Catal. Rev. 42(1–2), 71–144 (2000) 17. Djonev, B., Tsyntsarski, B., Klissurski, D., Hadjiivanov, K.: IR spectroscopic study of NOx adsorption and NOx-O-2 coadsorption on Co2+/SiO2 catalysts. J. Chem. Soc. Faraday Trans. 93(22), 4055–4063 (1997) 18. Pozdnyakov, D.V., Filimonov, V.N.: Use of IR spectroscopy to investigate chemisorption of nitric oxide and nitrogen dioxide on metallic oxides. Adv. Mol. Relax. Process. 5, 55–63 (1973) 19. Field, B.O., Hardy, C.J.: Volatile and anhydrous nitrato-complexes of metals: preparation by the use of dinitrogen pentroxide, and measurement of infrared spectra, J. Chem. Soc., 4428–4434 (1964). https://doi.org/10.1039/JR9640004428 20. Szanyi, J., Kwak, J.H., Chimentao, R.J., Peden, C.H.F.: Effect of H2O on the adsorption of NO2 on gamma-Al2O3: an in situ FTIR/ MS study. J. Phys. Chem. C. 111(6), 2661–2669 (2007) 21. Borensen, C., Kirchner, U., Scheer, V., Vogt, R., Zellner, R.: Mechanism and kinetics of the reactions of NO2 or HNO3 with alumina as a mineral dust model compound. J. Phys. Chem. A. 104(21), 5036–5045 (2000) 22. Tribe, L., Hinrichs, R., Kubicki, J.D.: Adsorption of nitrate on kaolinite surfaces: a theoretical study. J. Phys. Chem. B. 116(36), 11266–11273 (2012) 23. Borfecchia, E., Lomachenko, K.A., Giordanino, F., Falsig, H., Beato, P., Soldatov, A.V., Bordiga, S., Lamberti, C.: Revisiting the nature of Cu sites in the activated Cu-SSZ-13 catalyst for SCR reaction. Chem. Sci. 6(1), 548–563 (2015) 24. Giordanino, F., Vennestrom, P.N.R., Lundegaard, L.F., Stappen, F.N., Mossin, S., Beato, P., Bordiga, S., Lamberti, C.: Characterization of Cu-exchanged SSZ-13: a comparative FTIR, UV-Vis, and EPR study with Cu-ZSM-5 and Cu-beta with similar Si/Al and Cu/Al ratios. Dalton Trans. 42(35), 12741–12761 (2013) 25. Hadjiivanov, K., Saussey, J., Freysz, J.L., Lavalley, J.C.: FT-IR study of NO + O2 co-adsorption on H-ZSM-5: re-assignment of the 2133 cm
1 band to NO+ species. Catal. Lett. 52(1–2), 103–108 (1998) 26. Henriques, C., Marie, O., Thibault-Starzyk, F., Lavalley, J.C.: NO+ ions as IR probes for the location of OH groups and Na+ ions in main channels and side pockets of mordenite. Microporous Mesoporous Mater. 50(2–3), 167–171 (2001) 27. Rivallan, M., Ricchiardi, G., Bordiga, S., Zecchina, A.: Adsorption and reactivity of nitrogen oxides (NO2, NO, N2O) on Fe-zeolites. J. Catal. 264(2), 104–116 (2009) 28. Chen, H.-Y., Kollar, M., Wei, Z., Gao, F., Wang, Y., Szanyi, J., Peden, C.H.F.: Formation of NO+ and its possible roles during the selective catalytic reduction of NOx with NH3 on Cu-CHA catalysts, Catal. Today 320, 61–71 (2019) 29. Ruggeri, M.P., Nova, I., Tronconi, E., Pihl, J.A., Toops, T.J., Partridge, W.P.: In-situ DRIFTS measurements for the mechanistic study of NO oxidation over a commercial Cu-CHA catalyst. Appl. Catal. B-Environ. 166, 181–192 (2015) 30. Bordiga, S., Regli, L., Lamberti, C., Zecchina, A., Bjorgen, M., Lillerud, K.P.: FTIR adsorption studies of H2O and CH3OH in the isostructural H-SSZ-13 and H-SAPO-34: formation of H-bonded adducts and protonated clusters. J. Phys. Chem. B. 109(16), 7724–7732 (2005) 31. Paze, C., Bordiga, S., Lamberti, C., Salvalaggio, M., Zecchina, A., Bellussi, G.: Acidic properties of H-beta zeolite as probed by bases with proton affinity in the 118-204 kcal mol(
1) range: a FTIR investigation. J. Phys. Chem. B. 101(24), 4740–4751 (1997)

49 32. Aylor, A.W., Lobree, L.J., Reimer, J.A., Bell, A.T.: NO Adsorption, Desorption, and Reduction by CH4 over Mn-ZSM-5. J. Catal. 170, 390–401 (1997) 33. Negri, C., Hammershoi, P.S., Janssens, T.V.W., Beato, P., Berlier, G., Bordiga, S.: Investigating the low temperature formation of Cu-II-(N,O) species on Cu-CHA zeolites for the selective catalytic reduction of NOx. Chem.-Eur. J. 24(46), 12044–12053 (2018) 34. Martini, A., Borfecchia, E., Lomachenko, K.A., Pankin, I.A., Negri, C., Berlier, G., Beato, P., Falsig, H., Bordiga, S., Lamberti, C.: Composition-driven Cu-speciation and reducibility in Cu-CHA zeolite catalysts: a multivariate XAS/FTIR approach to complexity. Chem. Sci. 8(10), 6836–6851 (2017) 35. Andersen, C.W., Borfecchia, E., Bremholm, M., Jorgensen, M.R. V., Vennestrom, P.N.R., Lamberti, C., Lundegaard, L.F., Iversen, B.B.: Redox-driven migration of copper ions in the Cu-CHA zeolite as shown by the in situ PXRD/XANES technique. Angew. Chem. Int. Edit. 56(35), 10367–10372 (2017) 36. Chen, H.Y., Wei, Z.H., Kollar, M., Gao, F., Wang, Y.L., Szanyi, J., Peden, C.H.F.: NO oxidation on zeolite supported Cu catalysts: formation and reactivity of surface nitrates. Catal. Today. 267, 17–27 (2016) 37. Devadas, M., Krocher, O., Elsener, M., Wokaun, A., Mitrikas, G., Soger, N., Pfeifer, M., Demel, Y., Mussmann, L.: Characterization and catalytic investigation of Fe-ZSM5 for urea-SCR. Catal. Today. 119(1–4), 137–144 (2007) 38. Sun, Q., Gao, Z.X., Chen, H.Y., Sachtler, W.M.H.: Reduction of NOx with ammonia over Fe/MFI: reaction mechanism based on isotopic labeling. J. Catal. 201(1), 89–99 (2001) 39. Venkov, T., Hadjiivanov, K., Klissurski, D.: IR spectroscopy study of NO adsorption and NO+O-2 co- adsorption on Al2O3. Phys. Chem. Chem. Phys. 4(11), 2443–2448 (2002) 40. Szanyi, J., Kwak, J.H., Zhu, H., Peden, C.H.F.: Characterization of Cu-SSZ-13 NH3 SCR catalysts: an in situ FTIR study. Phys. Chem. Chem. Phys. 15(7), 2368–2380 (2013) 41. Morandi, S., Prinetto, F., Castoldi, L., Lietti, L., Forzatti, P., Ghiotti, G.: Effect of water and ammonia on surface species formed during NOx storage-reduction cycles over Pt-K/Al2O3 and Pt-Ba/Al2O3 catalysts. Phys. Chem. Chem. Phys. 15(32), 13409–13417 (2013) 42. Morandi, S., Prinetto, F., Ghiotti, G., Livi, M., Vaccari, A.: FT-IR investigation of NOx storage properties of Pt-Mg(Al)O and Pt/CuMg(Al)O catalysts obtained from hydrotalcite compounds. Microporous Mesoporous Mater. 107(1–2), 31–38 (2008) 43. Cao, L., Chen, L., Wu, X.D., Ran, R., Xu, T.F., Chen, Z., Weng, D.: TRA and DRIFTS studies of the fast SCR reaction over CeO2/ TiO2 catalyst at low temperatures. Appl. Catal. A-Gen. 557, 46–54 (2018) 44. Lagunov, O., Drenchev, N., Chakarova, K., Panayotov, D., Hadjiivanov, K.: Isotopic labelling in vibrational spectroscopy: a technique to decipher the structure of surface species. Top. Catal. 60, 1486–1495 (2017) 45. Weingand, T., Kuba, S., Hadjiivanov, K., Knoezinger, H.: Nature and reactivity of the surface species formed after NO adsorption and NO+O2 coadsorption on a WO3-ZrO2 catalyst. J. Catal. 209, 539–546 (2002) 46. Blaser, H.-U., Schnyder, A., Steiner, H., Roessler, F., Baumeister, P.: Selective hydrogenation of functionalized hydrocarbons. In: Ertl, G., Knoezinger, H., Schuth, F., Weitkamp, J. (eds.) Handbook of Heterogeneous Catalysis, vol. 1, 2nd edn, pp. 3284–3308. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany (2008) 47. Ertl, G., Knoezinger, H. (eds.): Handbook of Heterogeneous Catalysis, vol. 5. Wiley-VCH, Weinheim, Germany (1997) 2800 pp 48. Nishimura, S.: Handbook of Heterogeneous Catalytic Hydrogenation for Organic Synthesis. Wiley, New York (2001) 784 pp.

2

50 49. Rylander, P.: Catalytic Hydrogenation in Organic Syntheses. Academic Press, New York (1979) 325 pp 50. Armbruester, M., Behrens, M., Cinquini, F., Foettinger, K., Grin, Y., Haghofer, A., Kloetzer, B., Knop-Gericke, A., Lorenz, H., Ota, A., Penner, S., Prinz, J., Rameshan, C., Revay, Z., Rosenthal, D., Rupprechter, G., Sautet, P., Schloegl, R., Shao, L., Szentmiklosi, L., Teschner, D., Torres, D., Wagner, R., Widmer, R., Wowsnick, G.: How to control the selectivity of palladium-based catalysts in hydrogenation reactions: the role of subsurface chemistry. ChemCatChem. 4(8), 1048–1063 (2012) 51. Lewis, F.A.: The Palladium-Hydrogen System. Academic Press, London (1967) 178 pp 52. Teschner, D., Borsodi, J., Wootsch, A., Revay, Z., Haevecker, M., Knop-Gericke, A., Jackson, S.D., Schloegl, R.: The roles of subsurface carbon and hydrogen in palladium-catalyzed alkyne hydrogenation. Science. 320(5872), 86–89 (2008) 53. Wicke, E., Brodowsky, H., Zuechner, H.: Hydrogen in palladium and palladium alloys. Top. Appl. Phys. 29(Hydrogen Met., Vol. 2), 73–155 (1978) 54. Che, M.: Nobel Prize in chemistry 1912 to Sabatier: organic chemistry or catalysis? Catal. Today. 218–219, 162–171 (2013) 55. Zecchina, A., Califano, S.: Chapter 1: From the onset to the first large-scale industrial processes. In: The Development of Catalysis: A History of Key Processes and Personas in Catalytic Science and Technology, I edn, pp. 1–57. Wiley (2017) 56. Andrews, L., Wang, X., Manceron, L.: Infrared spectra and density functional calculations of platinum hydrides. J. Chem. Phys. 114(4), 1559–1566 (2001) 57. Mueller, W.M., Blackledge, J.P., Libowitz, G.G.: Metal Hydrides. Academic Press, New York (1968) 791 pp 58. Tsuchiya, S., Amenomiya, Y., Cvetanovic, R.J.: Study of metal catalysts by temperature programmed desorption. II. Chemisorption of hydrogen on platinum. J. Catal. 19(3), 245–255 (1970) 59. Dixon, L.T., Barth, R., Kokes, R.J., Gryder, J.W.: Hydrogen adsorption by alumina-supported platinum. J. Catal. 37(2), 376–382 (1975) 60. Miller, J.T., Meyers, B.L., Modica, F.S., Lane, G.S., Vaarkamp, M., Koningsberger, D.C.: Hydrogen temperature-programmed desorption (H2 TPD) of supported platinum catalysts. J. Catal. 143(2), 395–408 (1993) 61. Tew, M.W., Janousch, M., Huthwelker, T., van Bokhoven, J.A.: The roles of carbide and hydride in oxide-supported palladium nanoparticles for alkyne hydrogenation. J. Catal. 283(1), 45–54 (2011) 62. Gorczyca, A., Moizan, V., Chizallet, C., Proux, O., Del Net, W., Lahera, E., Hazemann, J.-L., Raybaud, P., Joly, Y.: Monitoring morphology and hydrogen coverage of nanometric Pt/γ-Al2O3 particles by in situ HERFD–XANES and quantum simulations. Angew. Chem. Int. Ed. 53(46), 12426–12429 (2014) 63. Alexeev, O.S., Li, F., Amiridis, M.D., Gates, B.C.: Effects of adsorbates on supported platinum and iridium clusters: characterization in reactive atmospheres and during alkene hydrogenation catalysis by X-ray absorption spectroscopy. J. Phys. Chem. B. 109(6), 2338–2349 (2005) 64. Bus, E., Miller, J.T., Kropf, A.J., Prins, R., van Bokhoven, J.A.: Analysis of in situ EXAFS data of supported metal catalysts using the third and fourth cumulant. Phys. Chem. Chem. Phys. 8(27), 3248–3258 (2006) 65. Kang, J.H., Menard, L.D., Nuzzo, R.G., Frenkel, A.I.: Unusual non-bulk properties in nanoscale materials: thermal metal-metal bond contraction of γ-alumina-supported Pt catalysts. J. Am. Chem. Soc. 128(37), 12068–12069 (2006) 66. Oudenhuijzen, M.K., Van Bokhoven, J.A., Miller, J.T., Ramaker, D.E., Koningsberger, D.C.: Three-site model for hydrogen adsorption on supported platinum particles: influence of support ionicity

C. Negri et al. and particle size on the hydrogen coverage. J. Am. Chem. Soc. 127(5), 1530–1540 (2005) 67. Frenkel, A.I., Small, M.W., Smith, J.G., Nuzzo, R.G., Kvashnina, K.O., Tromp, M.: An in situ study of bond strains in 1 nm Pt catalysts and their sensitivities to cluster-support and clusteradsorbate interactions. J. Phys. Chem. C. 117(44), 23286–23294 (2013) 68. Jung, U., Elsen, A., Li, Y., Smith, J.G., Small, M.W., Stach, E.A., Frenkel, A.I., Nuzzo, R.G.: Comparative in operando studies in heterogeneous catalysis: atomic and electronic structural features in the hydrogenation of ethylene over supported Pd and Pt catalysts. ACS Catal. 5(3), 1539–1551 (2015) 69. Bugaev, A., Guda, A., Lazzarini, A., Lomachenko, K., Groppo, E., Pellegrini, R., Piovano, A., Emerich, H., Soldatov, A., Bugaev, L., Dmitriev, V.P., Bokhoven, J., Lamberti, C.: In situ formation of hydrides and carbides in palladium catalyst: when XANES is better than EXAFS and XRD. Catal. Today. 283, 119–126 (2017) 70. Bugaev, A.L., Guda, A.A., Lomachenko, K.A., Shapovalov, V.V., Lazzarini, A., Vitillo, J.G., Bugaev, L.A., Groppo, E., Pellegrini, R., Soldatov, A.V., van Bokhoven, J.A., Lamberti, C.: Core–shell structure of palladium hydride nanoparticles revealed by combined X-ray absorption spectroscopy and X-ray diffraction. J. Phys. Chem. C. 121(33), 18202–18213 (2017) 71. Bugaev, A.L., Guda, A.A., Lomachenko, K.A., Srabionyan, V.V., Bugaev, L.A., Soldatov, A.V., Lamberti, C., Dmitriev, V.P., van Bokhoven, J.A.: Temperature- and pressure-dependent hydrogen concentration in supported PdHx nanoparticles by Pd K-edge X-ray absorption spectroscopy. J. Phys. Chem. C. 118(19), 10416–10423 (2014) 72. Dong, Y., Hu, G., Hu, X., Xie, G., Lu, J., Luo, M.: Hydrogen adsorption and oxidation on Pt film: an in situ real-time attenuated total reflection infrared (ATR-IR) spectroscopic study. J. Phys. Chem. C. 117(24), 12537–12543 (2013) 73. Feng, X., Yao, M., Hu, X., Hu, G., Jia, A., Xie, G., Lu, J., Luo, M.: Co-adsorption of hydrogen and CO on Pt film: an in-situ ATR-IR study combined with DFT calculations. Int. J. Hydrog. Energy. 38(31), 13673–13679 (2013) 74. Gorodetskii, V.V., Sametova, A.A., Matveev, A.V., Tapilin, V.M.: From single crystals to supported nanoparticles in experimental and theoretical studies of H2 oxidation over platinum metals (Pt, Pd): intermediates, surface waves and spillover. Catal. Today. 144(3), 219–234 (2009) 75. Pennemann, B., Oster, K., Wandelt, K.: Hydrogen adsorption on platinum (100) at low temperatures: work function and thermal desorption data. Surf. Sci. 249(1–3), 35–43 (1991) 76. Peremans, A., Tadjeddine, A.: Vibrational spectroscopy of electrochemically deposited hydrogen on platinum. Phys. Rev. Lett. 73(22), 3010–3013 (1994) 77. Pliskin, W.A., Eischens, R.P.: Infrared spectra of hydrogen and deuterium chemisorbed on platinum. Z. Phys. Chem. (Muenchen., Ger.). 24, 11–23 (1960) 78. Eley, D.D., Moran, D.M., Rochester, C.H.: Infrared study of interaction between hydrogen and supported platinum catalysts. Trans. Faraday Soc. 64(8), 2168–2180 (1968) 79. Primet, M., Basset, J.M., Mathieu, M.V., Prettre, M.: Infrared investigation of hydrogen adsorption on alumina-supported platinum. J. Catal. 28(3), 368–375 (1973) 80. Dixon, L.T., Barth, R., Gryder, J.W.: Infrared active species of hydrogen adsorbed by alumina-supported platinum. J. Catal. 37(2), 368–375 (1975) 81. Candy, J.P., Fouilloux, P., Primet, M.: Hydrogen adsorption between 300 and 873 K on a platinum/magnesium oxide catalyst. Surf. Sci. 72(1), 167–176 (1978) 82. Szilagyi, T.: Fourier-transform infrared study of weak adsorption of hydrogen on platinum/silica. J. Catal. 121(2), 223–227 (1990)

2

Case Studies: Infrared (IR) Spectroscopy

83. Paleček, D., Tek, G., Lan, J., Iannuzzi, M., Hamm, P.: Characterization of the platinum–hydrogen bond by surface-sensitive timeresolved infrared spectroscopy. J. Phys. Chem. Lett. 9(6), 1254–1259 (2018) 84. Mitchell, P.C.H., Parker, S.F., Ramirez-Cuesta, A.J., Tomkinson, J.: Vibrational Spectroscopy with Neutrons, with Applications in Chemistry, Biology, Materials Science and Catalysis. World Scientific, Singapore (2005) 85. Albers, P., Auer, E., Ruth, K., Parker, S.F.: Inelastic neutron scattering investigation of the nature of surface sites occupied by hydrogen on highly dispersed platinum on commercial carbon black supports. J. Catal. 196(1), 174–179 (2000) 86. Albers, P.W., Krauter, J.G.E., Ross, D.K., Heidenreich, R.G., Koehler, K., Parker, S.F.: Identification of surface states on finely divided supported palladium catalysts by means of inelastic incoherent neutron scattering. Langmuir. 20(19), 8254–8260 (2004) 87. Albers, P.W., Lopez, M., Sextl, G., Jeske, G., Parker, S.F.: Inelastic neutron scattering investigation on the site occupation of atomic hydrogen on platinum particles of different size. J. Catal. 223(1), 44–53 (2004) 88. Parker, S.F., Frost, C.D., Telling, M., Albers, P., Lopez, M., Seitz, K.: Characterisation of the adsorption sites of hydrogen on Pt/C fuel cell catalysts. Catal. Today. 114(4), 418–421 (2006) 89. Carosso, M., Lazzarini, A., Piovano, A., Pellegrini, R., Morandi, S., Manzoli, M., Vitillo, J.G., Ruiz, M.J., Lamberti, C., Groppo, E.: Looking for the active hydrogen species in a 5 wt% Pt/C catalyst: a challenge for inelastic neutron scattering. Faraday Discuss. 208(Designing Nanoparticle Systems for Catalysis), 227–242 (2018) 90. Piovano, A., Agostini, G., Carosso, M., Groppo, E., Jimenez Ruiz, M., Lamberti, C., Lazzarini, A., Manzoli, M., Morandi, S., Pellegrini, R., Vottero, E.: Study of the Pt-hydride formation and spillover effect on Pt/Al2O3 and Pt/C catalysts (2016) 91. Wang, L.-L., Johnson, D.D.: Shear instabilities in metallic nanoparticles: hydrogen-stabilized structure of Pt37 on carbon. J. Am. Chem. Soc. 129(12), 3658–3664 (2007) 92. Mager-Maury, C., Bonnard, G., Chizallet, C., Sautet, P., Raybaud, P.: H2-induced reconstruction of supported Pt clusters: metalsupport interaction versus surface hydride. ChemCatChem. 3(1), 200–207 (2011) 93. Small, M.W., Sanchez, S.I., Marinkovic, N.S., Frenkel, A.I., Nuzzo, R.G.: Influence of adsorbates on the electronic structure, bond strain, and thermal properties of an alumina-supported Pt catalyst. ACS Nano. 6(6), 5583–5595 (2012) 94. Bus, E., van Bokhoven, J.A.: Hydrogen chemisorption on supported platinum, gold, and platinum-gold-alloy catalysts. Phys. Chem. Chem. Phys. 9(22), 2894–2902 (2007) 95. Bus, E., Miller, J.T., van Bokhoven, J.A.: Hydrogen chemisorption on Al2O3-supported gold catalysts. J. Phys. Chem. B. 109(30), 14581–14587 (2005) 96. Li, L., Wang, L.-L., Johnson, D.D., Zhang, Z., Sanchez, S.I., Kang, J.H., Nuzzo, R.G., Wang, Q., Frenkel, A.I., Li, J., Ciston, J., Stach, E.A., Yang, J.C.: Noncrystalline-to-crystalline transformations in Pt nanoparticles. J. Am. Chem. Soc. 135, 13062–13072 (2013) 97. Sanchez, S.I., Menard, L.D., Bram, A., Kang, J.H., Small, M.W., Nuzzo, R.G., Frenkel, A.I.: The emergence of nonbulk properties in supported metal clusters: negative thermal expansion and atomic disorder in Pt nanoclusters supported on γ-Al2O3. J. Am. Chem. Soc. 131, 7040–7054 (2009) 98. Carosso, M., Vottero, E., Lazzarini, A., Morandi, S., Manzoli, M., Lomachenko, K.A., Ruiz, M.J., Pellegrini, R., Lamberti, C., Piovano, A., Groppo, E.: Dynamics of reactive species and reactant-induced reconstruction of Pt clusters in Pt/Al2O3 catalysts. ACS Catal. 9(8), 7124–7136 (2019) 99. Kaprielova, K.M., Yakovina, O.A., Ovchinnikov, I.I., Koscheev, S.V., Lisitsyn, A.S.: Preparation of platinum-on-carbon catalysts

51 via hydrolytic deposition: factors influencing the deposition and catalytic properties. Appl. Catal., A. 449, 203–214 (2012) 100. Benson, J.E., Boudart, M.: Hydrogen-oxygen titration method for the measurement of supported platinum surface areas. J. Catal. 4(6), 704–710 (1965)

Chiara Negri received a PhD in Chemical and Materials Sciences in 2018 at the Chemistry Department of the University of Torino, in Italy. Currently, she is working as a research fellow at the Chemistry Department of Oslo University. Her scientific researches deal with the study and characterization of heterogeneous catalysts of industrial interest (e.g., oxides/metal-exchanged zeolites), applying complementary in situ/operando spectroscopy techniques.

Michele Carosso received a M.Sc. degree in Chemistry in 2016 and a PhD in Chemical and Materials Sciences in 2019, both at the Chemistry Department of the University of Torino, in Italy. Currently, he is working in the same institution within a fellowship. His main research interest resides in the characterization of industrial catalysts based on supported noble metal nanoparticles by means of in situ/operando spectroscopies.

Eleonora Vottero received a M.Sc. degree in Chemistry in 2017 at the Chemistry Department of the University of Torino. Currently she is a PhD student in a co-tutorship position between the University of Torino and the Institut Laue-Langevin in Grenoble. Her studies are mostly focused on the characterization of supported metal nanoparticles catalysts by means of FT-IR and INS spectroscopy, complemented by DFT simulations.

2

52

Elena Groppo is an associate Professor in Physical Chemistry at the Department of Chemistry of the University of Torino. Her research interests are mainly focused on the understanding the physical-chemical working principles of heterogeneous catalysts by applying a multitechnique approach, comprising in situ and operando spectroscopic techniques. Most investigated systems are heterogeneous catalysts for olefin polymerization and supported metal nanoparticles for selective hydrogenation and oxidation reactions.

C. Negri et al.

Silvia Bordiga, full Prof. in Physical Chemistry at the Department of Chemistry of the University of Torino, since 2012, is also Prof. II at the Department of Chemistry of the University of Oslo. Her scientific activity is mainly devoted to the characterization of the physicalchemical properties of high-surface-area nanostructured materials used as heterogeneous catalysts, materials for adsorption, separation, and storage, through in situ spectroscopic studies.

3

Reflection Absorption Infrared Spectroscopy Ravi Ranjan

and Michael Trenary

Contents 3.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.2 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Hydrocarbons on Metallic and Single-Atom Alloy Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Adsorption of Methanol on Palladium . . . . . . . . . . . . . . . . . . . . . . 3.2.3 CO2 Activation on a ZrO2 Film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 CO on Metallic, Bimetallic, and Single-Atom Alloy (SAA) Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3

55 55 59 60 61

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Abstract

Recent case studies are presented that illustrate the capabilities of the technique of reflection absorption infrared spectroscopy (RAIRS) for probing the structure and surface chemistry of model metal catalysts. A key feature of RAIRS is the polarization dependence of the interaction of infrared radiation with surface vibrations. Only the component of polarization parallel to the plane of incidence ( p-polarization) interacts with molecules on the metal surface, whereas surface vibrations are invisible to light polarized perpendicular to the plane of incidence (s-polarization). Since both polarizations interact with gas-phase molecules, the polarization dependence of the spectra allows surface vibrations to be distinguished from those of gas-phase species. This is the basis for using RAIRS to study surfaces under ambient pressure conditions, in addition to studies in ultrahigh vacuum. The case studies presented include the use of RAIRS of CO to probe the structure of surfaces, the hydrogenation of acetylene, the activation of CO2 by H2O on a ZrO2 thin-film grown on a R. Ranjan · M. Trenary (*) Department of Chemistry, University of Illinois Chicago, Chicago, IL, USA e-mail: [email protected]; [email protected]

Pt3Zr(0001) surface, and the formation of hydrogenbonded clusters of methanol on a Pd(111) surface. Keywords

Polarization-dependent reflection absorption infrared spectroscopy · Reflection absorption infrared spectroscopy · Polarization modulation reflection absorption infrared spectroscopy · Ambient pressure studies · Model catalysts

3.1

Introduction

Infrared spectroscopy has long been used to characterize catalysts [1]. For supported metal catalysts, the IR spectra are typically obtained either by transmission through a thin layer or pressed disk of catalyst material or by diffuse reflectance from the catalyst powder. As the vibrational frequencies of molecular adsorbates depend sensitively on the identity of the adsorption site, the heterogeneous distribution of different types of adsorption sites can lead to broad peaks, which can mask important information. In addition, the support materials, which are typically oxides, possess their own IR active vibrations, and it can be difficult to identify peaks attributable solely to molecules on the metal sites. As an alternative to studying actual catalysts, metal single crystals that can be well characterized by a variety of surface science techniques can serve as model catalysts. By studying a single crystal under ultrahigh vacuum (UHV) conditions, the surface composition and structure can be well defined and controlled. Another advantage of using UHV conditions is that molecular adsorption at low temperatures can be studied as a way to establish adsorbate structure before reaction takes place as the temperature is raised. Infrared spectra are obtained after a single reflection from a small single crystal of a model metal catalyst in UHV using the technique of reflection absorption infrared spectroscopy (RAIRS), which

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_3

53

54

R. Ranjan and M. Trenary

is also called infrared reflection absorption spectroscopy (IRAS or IRRAS). However, the method can also be used in the presence of an ambient pressure of gases during catalytic reactions. In this way, RAIRS can bridge the pressure gap between UHV studies and more realistic catalytic conditions. Here, we present some recent examples that illustrate the capabilities of RAIRS in studies of the structure and surface chemistry of model catalysts. Although our focus is on the gas-solid interface, RAIRS is also widely used for studies at the liquid-solid interface, which are of relevance to electrocatalysis [2]. The theoretical foundation for RAIRS as currently practiced is generally credited to Greenler [3]. He showed that the absorption factor, A, defined as in Fig. 3.1, reaches a sharp maximum for most metals at angles of incidence in the range of 85–89
from the surface normal [4]. Furthermore, he showed that the phase change upon reflection was quite different for light polarized parallel ( p-polarized) and perpendicular (s-polarized) to the plane of incidence. The phase change is such that there is essentially no electric field at the surface for s-polarized light, while for p-polarized light, there is an enhanced electric field component along the surface normal, with a negligible component parallel to the surface. This leads to the surface dipole selection rule that only molecular vibrations with a non-zero component of the dynamic dipole moment along the surface normal will be allowed in RAIRS. This rule is quite rigorous and is best applied by considering the point group symmetry of an adsorbate [5]. The different behavior of s- and p-polarized infrared radiation upon reflection from metal surfaces is the basis for discriminating between absorption features due to molecules on the surface from molecules in the surrounding gas phase or liquid phase. In a UHV RAIRS experiment, a background spectrum is obtained of the reflectivity (R
) from the clean surface and then at a later time, the reflectivity (R) from the

Reflectance

1.0

adsorbate-covered surface. The absorption factor, as defined in Fig. 3.1, A¼

R0  R ΔR ΔR ¼ 0 ffi , R R0 R

is a positive quantity since the reflectivity at frequencies where the adsorbates absorb radiation is lower than the reflectivity from the clean surface. Since the change in reflectivity is generally quite small, R ffi R0. A major shortcoming of measuring spectra in this way is that during the time lag between the sample and background spectra, which is seldom shorter than a few minutes, the overall optical performance of the system can drift, leading to miscancellation of various extraneous features present in both the background and sample spectra. Such miscancellations are a major source of spurious signals in the spectra. Some authors use the definition ΔR ¼ R  R0, in which case RAIRS peaks are negative. It is also common to present spectra in absorbance, which like ΔR/R, is a unitless quantity. For low values of absorbance, R/R0 is proportional to absorbance, and there is little practical significance in choosing one over the other. In most spectroscopic measurements, absorbance is preferred as it is proportional to concentration, according to Beer’s law. However, as the derivation of Beer’s law is based on a well-defined path length and a concentration based on amount of substance per unit volume, the applicability of Beer’s law to reflection from a surface with a submonolayer coverage of an adsorbate is unclear. Since the s-polarized component does not contribute to ΔR/R, peak intensities obtained with unpolarized radiation are approximately half of the values obtained with p-polarized radiation. Gas-phase (or liquid phase) molecules anywhere in the optical path will absorb the s- and p-polarized components equally, so that the difference between the two should remove all features not due to molecules adsorbed on the metal surface. There are several ways to take practical advantage of this polarization dependence. The technique of polarization modulation RAIRS (PM-RAIRS) has been well described in the literature [6–9]. The method involves sending linearly polarized light through a photoelastic modulator (PEM) to obtain a signal proportional to Rp  Rs at the PEM modulation frequency, and a separate signal proportional to Rp + Rs, to yield:

0.5 R°

'R = R° – R

ΔR ¼

Rp  Rs : Rp þ Rs

R A= 0.0

Oo

R° – R R°

Wavelength

Fig. 3.1 Definition of the absorption factor A. (From Ref. [3]. Reproduced with permission from AIP Publishing)

This method has several appealing features. (1) It directly yields spectra sensitive only to surface features in real time, eliminating problems associated with slow drift in overall reflectivity. (2) There is no need to measure separate clean and adsorbate-covered surfaces. For example, spectra could be obtained, in principle, in air for a coated metal surface for

3

Reflection Absorption Infrared Spectroscopy

which a clean surface reference spectrum is unobtainable. (3) There is no need, in principle, to purge the spectrometer and the optical path external to the sample cell (assuming there is a sample cell) with dry air as absorption due to H2O (g) and CO2(g) should cancel out. In practice, the full advantages of PM-RAIRS are difficult to realize. One factor is that the polarization efficiency is wavelength dependent, which generates a strongly varying background signal that does not always cancel out completely in the ΔR/R spectra. The overall throughput for a PM-RAIRS setup is somewhat less than for a conventional one, which can lower the signal-to-noise ratio. At ambient pressures of even only a few Torr, some molecules absorb so much IR radiation that essentially no light is transmitted in certain frequency ranges, making these ranges inaccessible to the method. In addition, at elevated temperatures, the infrared radiation emitted from the sample can interfere with the operation of some IR detectors. Thus, it is usually impractical to carry out RAIRS studies under truly realistic catalytic reaction conditions, such as those for ammonia synthesis where pressures of 150–300 bar and temperatures of around 400
C are typically used [10]. Given these shortcomings, the extra expense and complexity of PM-RAIRS may not be not justified for many applications of RAIRS. Schennach et al. constructed a PM-RAIRS setup similar to the conventional one, but instead of a PEM, they placed one polarizer in front of the sample oriented at 45
to equalize the intensities of both s- and p-polarized light. The rotatable polarizer was placed after the sample and rotated between 0
and 90
to collect s- and p-polarized spectra sequentially [11]. This method is both simpler to implement than polarization modulation using a PEM and is less restrictive on the accessible spectral range. We have used an even simpler method that uses only one polarizer in which we separately record ΔR/R for s- and p-polarized spectra. The ΔRp/Rp spectra are sensitive to both surface and gas-phase species, while the ΔRs/Rs spectra show only peaks due to gas-phase species. The difference, ΔRp/Rp  ΔRs/Rs, yields spectra of the surface species only. Conversely, the ΔRs/Rs spectra can be used to monitor the changes only in the gas-phase composition as gas-phase reactants are converted to gas-phase products. The method relies on being able to take background spectra of the clean surface and sample spectra for both s and p polarized light in separate experiments. Since the ΔRp/Rp and ΔRs/Rs spectra are separated by considerable time intervals of as much as a day or two, the method requires a highly stable system and a high degree of reproducibility in the experimental conditions. Fortunately, such stability and reproducibility are possible with RAIRS. We present here case studies that use PM-RAIRS as well as polarizationdependent RAIRS (PD-RAIRS), showing that both methods provide spectra of adsorbates on model catalyst surfaces in the absence and presence of ambient gas.

55

A typical experimental apparatus used for ambient RAIRS studies on well-characterized single crystals is illustrated in Fig. 3.2. The apparatus consist of an upper analysis chamber equipped with a cylindrical mirror analyzer for Auger electron spectroscopy (AES), an apparatus for lower-energy electron diffraction (LEED), and a quadrupole mass spectrometer for performing temperature-programmed desorption (TPD) experiments. The elemental composition of the surface is measured with AES, the structure of the clean surface, and the presence of ordered overlayer structures are established with LEED, and TPD can be used to explore reaction chemistry and to analyze the gas-phase products evolving from the surface after ambient pressure experiments are performed in the IR cell. The single-crystal sample is attached to the bottom of a stainless steel tube via a vacuum-tight ceramic feedthrough containing leads for sample heating and a thermocouple for temperature measurements. An XYZ manipulator is used to translate the sample from the analysis chamber into the cell through three stacked spring-loaded fluorocarbon polymer seals. The gaps between the seals are differentially pumped. When the sample tube is translated through the seals, the high-pressure cell is isolated from the analysis chamber. The basic idea follows a design of Goodman’s group [12, 13]. The cell is equipped with differentially pumped KBr windows following a design by Hollins and Pritchard [14]. A gate valve is located between the sliding seals and the analysis chamber so that once the cell is evacuated after a measurement and the sample retracted back into the analysis chamber, the cell can be isolated from the chamber. Otherwise, slow degassing from the cell would place an additional gas load on the analysis chamber pumping system and raise the base pressure. This apparatus was used for some of the case studies described below [15–17].

3.2

Case Studies

The next sections cover studies of hydrocarbons, carbon dioxide, and carbon monoxide as a probe molecule on metallic, bimetallic, and single-atom alloy (SAA) surfaces at pressures ranging from ultrahigh vacuum (UHV) to several hundred Torr.

3.2.1

Hydrocarbons on Metallic and SingleAtom Alloy Surfaces

The reactions of hydrocarbons on transition metal surfaces are of enormous importance in heterogeneous catalysis. Consequently, there have been many RAIRS studies of their adsorption and reactions [18–20]. In particular, the hydrogenation of unsaturated hydrocarbons over metal catalysts has been extensively investigated since its discovery by Sabatier

3

56

R. Ranjan and M. Trenary

Q D

S N

C

K

R

B G

H C

I D

F

E

L

P

A B

K

J G

P

N Q

M

O

Fig. 3.2 Ambient RAIRS system. (A) FTIR, (B) CMA detector for AES, (C) LEED optics, (D) leak valve, (E) ion gauge, (F) main chamber, (G) metal evaporator, (H) titanium sublimation pump, (I) ion pump, (J) viewport, (K) mass spectrometer, (L) manual ion pump gate valve,

(M) MCT detector, (N) purge box for IR optics, (O) turbo pump, (P) interlocked gate valve, (Q) IR cell, (R) 4-axis manipulator, (S) crystal attached to manipulator transfer rod

and Senderens in 1897 [21–23]. Because they are the simplest unsaturated hydrocarbons and because of the many catalytic processes involving them, acetylene and ethylene have been thoroughly studied, including with RAIRS. In one of the first such studies, Malik et al. reported the RAIR spectrum of ethylidyne (CCH3) on Pt(111), which is a stable intermediate formed from ethylene [24]. An advantage of

RAIRS is that the spectra can be directly compared to transmission infrared spectra obtained on supported metal catalysts. Malik et al. compared their RAIR spectra of ethylidyne with transmission IR spectra of ethylidyne on a Pt/Al2O3 supported metal catalyst [24]. The spectra have remarkably similar peak positions and relative intensities but are different in two important ways. First, the peak widths are larger for

3

Reflection Absorption Infrared Spectroscopy

the Pt/Al2O3 sample. For example, the symmetric CH3 deformation mode has a peak position and full width at half maximum (FWHM) at 300 K of 1339 and 2.1 cm1, while on Pt/Al2O3, these values are 1342 and 5.7 cm1. While still unusually sharp, the greater width on the supported metal can be attributed to the greater heterogeneity of the adsorption sites available. Second, the asymmetric CH3 deformation and asymmetric C-H stretch modes are observed at 1410 and 2940 cm1 on Pt/Al2O3 but are absent on Pt(111) as they are not allowed by the surface selection rules. Similar transmission IR spectra have been obtained in studies of ethylene conversion to ethylidyne for Pt, Pd, Rh, Ir, and a PtRh alloy on Al2O3 supports [25–27]. Ethylidyne can also form from acetylene with many of the same intermediates proposed. On Pt(111), RAIR spectra have been used to identify an ethylidene (CHCH3) intermediate following ethylene adsorption [28], while a vinyl intermediate was identified following the adsorption of acetylene [29]. Zaera et al. had earlier also proposed that ethylidyne forms from ethylene via an ethylidene intermediate on Pt(111) [22]. Tysoe and coworkers used RAIRS to show that ethylene adsorbs on Pd(111) in a di-σ configuration, but in the presence of pre-adsorbed hydrogen, it changes to a π-bonded species [30–32]. The hydrogenation of acetylene to form ethylidyne on Pd(111) was found to occur via a vinyl intermediate [33]. In a PD-RAIRS study by Krooswyk et al., acetylene hydrogenation over Pt(111) was observed at ambient pressure, as shown in Fig. 3.3 [15]. The p-polarized spectra in Fig. 3.3a show peaks due to both gas-phase and surface species. The s-polarized spectra, which detect gas-phase molecules only, are shown in Fig. 3.3b. Peaks due to three different gases are present. C2H2(g) peaks are at 1304 and 1343 cm1 and are assigned to the P- and R-rotational branches of the combination band of the symmetric and asymmetric bending modes, while the peaks at 3269 and 3309 cm1 are the P and R branches of the asymmetric C-H stretch. The peak at 949 cm1 seen at 350 K is assigned to the ρw(CH2) mode of ethylene, which has B1u symmetry for the D2h point group. The peak at 2988 cm1 is the B3u CH stretch. At 370 K, the peaks between 2881 and 2987 cm1 are due to CH stretches of gas-phase ethane. Figure 3.3b reveals that acetylene is hydrogenated to ethylene, and then the ethylene is hydrogenated to ethane. By subtracting the spectra of Fig. 3.3b from those of Fig. 3.3a, the spectra in Fig. 3.3c reveal features due to surface species only. The weak peaks at 902, 2976, and 3025 are assigned to adsorbed acetylene. After addition of H2(g), ethylidyne peaks at 1340 and 2881 cm1 appear. After annealing the crystal to higher temperatures, weaker peaks of ethylidyne at 1118 and 2793 cm1 are seen. Other peaks in the spectra relate to C2Hx species such as vinylidene (CCH2), vinyl (CHCH2),

57

ethylene (both di-σ- and π-bonded), ethyl (CH2CH3), and ethynyl (CCH) [15]. In a different study, selective hydrogenation of acetylene to ethylene has been observed on the Pd/Cu(111) single-atom alloy (SAA) at ambient pressure (Fig. 3.4) [17]. The gas-phase acetylene peaks appeared at 1300, 1352, 3265, and 3312 cm1 after adding 1102 Torr of C2H2 and 1 Torr of H2 to the IR cell at 300 K with 0.08 monolayer (ML) of Pd on a Cu(111) surface. The peak at 950 cm1 is due to gas-phase ethylene formed upon heating the surface to 340 K (Fig. 3.4a). No conversion of ethylene to ethane is detectable, in contrast to the results in Fig. 3.3. This study also shows formation of oligomers on the surface as indicated by the peaks in Fig. 3.4b that develop after annealing to 320 K and above [17]. The formation of oligomers on Pd catalysts during acetylene hydrogenation is well known, and the spectral features observed with RAIRS in Fig. 3.4b and as described in Ref. [17] are a good match to transmission IR spectra of the oligomers formed on Pd catalysts. These oligomers are an example of what is often referred to as a carbonaceous deposit. The results shown in Fig. 3.4 show that hydrogenation of acetylene to ethylene proceeds despite the formation of the carbonaceous deposit. Simonovis et al. used RAIRS to investigate ethylene hydrogenation over Pt(111) surfaces in the presence of well-defined carbonaceous deposits [34]. They did this by first saturating the clean Pt(111) surface at room temperature in UHV with ethylene, propylene, butene, and toluene. The first three compounds produce a saturation coverage of the corresponding alkylidyne. They then inserted the crystal into a high-pressure cell to which they added a C2H4/H2 mixture and monitored the ethane formation as a function of time with a mass spectrometer. Figure 3.5 shows the ethane turn over number (TON) versus time after admitting to 2 Torr of C2H4 and 10 Torr of H2 to the cell for the propylidyne covered Pt (111) surface. For the first run, the reaction is completed after about 220 s. After about 470 s, the gases are pumped out, then readmitted for the second run. The reaction was completed about 20 s later than in the first run. The process is repeated for the third run, which yields essentially the same results as the second run. The surface species present while the data of Fig. 3.5 were obtained are identified through the RAIRS results shown in Fig. 3.6. The initial presence of propylidyne is revealed by the asymmetric C-H stretch peak at 2961 cm1. Concurrent with hydrogenation of ethylene to ethane is the replacement of propylidyne by ethylidyne, as indicated by growth of the peaks at 1120 (C-C stretch), 1341 (symmetric CH3 deformation), and 2882 (symmetric C-H stretch) cm1. Figure 3.6 also reveals a common experimental problem with RAIRS; the presence of the sharp rovibrational lines of H2O in the range of 1400–1700 cm1. These arise from water in the part of the optical path external to the vacuum chamber and can be

3

58

R. Ranjan and M. Trenary

a)

b) 1.0 × 10–2 torr C2H2 at 300 K

1.0 × 10–2 torr C2H2 at 300 K 902

2976

+ 1.0 torr H2

1304 1343

2994

320 K 910

320 K

3025

330 K 'RP

3269 3309

1462

'RS

2963

350 K

RP

+ 1.0 torr H2

330 K

RS

2793

970 1378

949

350 K

3269

3309

2929 2879 370 K

2988 370 K

949 1420

1118

0.002 5 × 10–4 1340 900

1100

1300

1500

2800

3000

3200

900

3400

1100

1300

Wavenumbers (cm–1)

1500

2881 2895 2987 2978 2800

3000

3200

3400

Wavenumbers (cm–1)

c)

1.0 × 10–2 torr C2H2 at 300 K + 1.0 torr H2

902

2976

2994

320 K 910

3025

330 K

'R

1462 350 K

R 970

2963

1378

2929 370 K 2793

1118

1420 2881 1340

900

1100

1300

1500

2800

3000

3200

–1)

Wavenumbers (cm

Fig. 3.3 Acetylene hydrogenation over Pt(111). p-polarized (a), s-polarized (b), and spectra obtained by subtracting the s-polarized spectra from the p-polarized spectra (c). 1.0102 Torr of C2H2 was added to the cell at 300 K. 1.0 Torr of H2 was then added, and the

crystal was annealed at 320, 330, 350, and 370 K. (Reprinted with permission from Ref. [15]. Copyright (2015) American Chemical Society)

3

Reflection Absorption Infrared Spectroscopy

59

a)

b) 1300 1352

1 × 10–2 torr C2H2 + 1 Torr H2 at 300 K

3265

927 1288

3312

1 × 10–2 torr C2H2 + 1 Torr H2 at 300 K

320 K 'Rs

340 K

950

1008 1380 1465

360 K 380 K

0.00025

Evacuate to UHV

900 1100 1300 1500 1700

2700 2900 3100 3300 3500

3

340 K

R

Rs

3011

320 K 'R

2915

360 K

380 K 2855 0.0025

900 1100 1300 1500 1700

Wavenumber (cm–1)

2960

2924 2700 2900 3100 3300 3500

Wavenumber (cm–1)

Fig. 3.4 Acetylene hydrogenation over a Pd/Cu(111) single-atom alloy. (a) s-polarized spectra after 1102 Torr of C2H2 and 1 Torr of H2 were added to the IR cell with the 0.08 ML Pd/Cu(111) crystal at 300 K followed by annealing to the indicated temperatures for 30 s. (b)

Subtraction of the s-polarized spectra from the p-polarized spectra yielded spectra of the surface species. (Reprinted with permission from Ref. [17]. Copyright (2017) American Chemical Society)

eliminated by purging the optical path with dry air or by evacuating the entire optical path. Purging with dry air has the advantage of lower cost but is often inadequate because the very low signal characteristics of many adsorbates are often comparable to peaks due to even trace amounts of water. In principle, this should not be a problem with PM-RAIRS as the gas-phase water peaks should cancel out. In practice, it is best to remove as much water vapor as possible from the optical path. Comparison of the results of Figs. 3.5 and 3.6 shows that while ethylene hydrogenation proceeds in the presence of both alkylidynes, the turnover frequency (the slope of the TON lines during hydrogenation) is higher when the surface is covered with propylidyne than with ethylidyne. Comparing the TON plots for clean Pt(111), with the surface covered with ethylidyne, propylidyne, butylidyne, and benzl (formed from toluene exposures), reveals that the highest reactivity is for the propylidyne-covered surface, followed by the clean surface, with the other surfaces all having about the same but slightly lower reactivity. Hydrogenation over the carbonaceous layer formed by annealing the propylidyne-covered surface to increasingly higher temperatures was also studied. A layer designated CnH(ad) had reduced reactivity, but annealing to 750 K and above to produce a hydrogen-free graphitic layer partially restored the reactivity. This was attributed to graphite having a smaller footprint, thereby making more bare Pt sites available. The study of Simonovis et al. makes it clear that the identity of the carbonaceous layer affects catalytic activity in subtle ways and that research on

the nature of such layers is needed [34]. Their study illustrates how in situ RAIRS can be used to provide crucial information on the species present during a reaction.

3.2.2

Adsorption of Methanol on Palladium

Although PM-RAIRS is especially useful for surface studies conducted in the presence of a background of gas-phase or even liquid-phase IR active molecules, it can also be useful for studies under UHV conditions. In a recent example, Kaichev et al. used PM-RAIRS to study methanol adsorption on Pd (111) [35]. As shown in Fig. 3.7, spectra were obtained after a 50 Langmuir (1 L ¼ 106 Torr sec) exposure of methanol at 80, 90, and 100 K. The 100 K spectrum consists of sharp peaks at 3309, 2959, 2832, 1473, 1144, and 1029 cm1, assigned to the O-H stretch, asymmetric and symmetric CH3 stretches, CH3 symmetric bend, CH3 rock, and CO stretch, respectively. The spectra indicate that methanol adsorbs molecularly at these temperatures. It desorbs from the surface above 120 K and no dehydration product was found. For adsorption at 80 and 90 K, the peaks are broader than for adsorption at 100 K. This is most prominent for the O-H stretch, which is resolved into two components at 3300–3303 and 3210–3216 cm1. This broadening and redshift of the O-H stretch is a clear indication of hydrogen bonding. The simple and obvious interpretation of the results is that methanol adsorbs as hydrogen-bonded clusters at 80 and 90 K, but as isolated molecules at 100 K. The ν(CO) peak also shows two

60

R. Ranjan and M. Trenary

Pt3{CCH2CH3/Pt(111) C2H4 + H2 Conversion, RAIRS

Pt3{CCH2CH3/Pt(111) C2H4 + H2 conversion, kinetics

0.0005 Ethylidyne 1341

40 L C3H6 P(C2H4) = 2 Torr P(H2) = 50 Torr T = 300 K

Absorbance

1st run 2nd run 3rd run

Ethylidyne 2882

After 1st run

2

40 L C3H6 in UHV

2

1

3

0 100

After 3rd run

4

0 Run: 0

T = 300 K P(C2H4) = 2 Torr P(H2) = 50 Torr

After 2nd run

Ethylidyne 1120

500

TOF (s–1)

TON (molec Pt-atom–1)

1000

200

300

400

1000

1200

1400

1600

Wavenumber

Propylidyne 2961

2800

3000

3200

(cm–1)

Time (s)

Fig. 3.5 Main frame: Accumulation of ethane (in turnover numbers, TON ¼ ethane molecules per Pt surface atoms) versus time during the hydrogenation of 2 Torr of ethylene with 50 Torr of H2 catalyzed by a Pt (111) single-crystal surface at 300 K. Three curves are reported, for three consecutive kinetic runs starting with a propylidyne-presaturated Pt (111) surface (prepared in the ultrahigh vacuum chamber beforehand by exposing the clean surface to 40 L of propylene at 300 K). Inset: Turnover frequencies (TOF, in units of TON/s) calculated by numerical derivatization of the TON versus time data. A slight decrease in conversion rate is seen between the first and second runs, as the initial surface species are replaced by ethylidyne moieties. (Reprinted with permission from Ref. [34]. Copyright (2017) American Chemical Society)

components for the hydrogen-bonded clusters, but for this mode, hydrogen bonding leads to a blue shift.

3.2.3

CO2 Activation on a ZrO2 Film

The conversion of CO2 into fuels is desirable, as it would lead to no net CO2 emissions upon combustion. Consequently, various catalytic schemes have been explored for its conversion into compounds that can be used, among other things, to generate energy by burning. For example, CO2 activation or adsorption on Cu/Cu2O or ZnO surfaces can be used for methanol synthesis [36–38]. Because of the interest in the catalytic activation of CO2, the surface chemistry of CO2 has long been of interest, as described in a series of review articles [39–41]. In a recent example, Li et al. have used PM-IRAS in a study of CO2 interaction, in the presence and absence of water, with an ultrathin ZrO2 film grown on Pt3Zr(0001) [43]. Without water, no reaction was observed for CO2 pressures up to 3102 mbar at room temperature. In the presence of H2O, CO2 was activated

Fig. 3.6 Reflection-absorption infrared spectra (RAIRS) for the initial propylidyne-dosed Pt(111) surface (bottom) and after each of the three kinetic runs reported in Fig. 3.5. A stepwise change is observed as the initial propylidyne moieties, identified mainly by the 2961 cm1 peak due to the CH asymmetric stretching mode, is replaced by the features characteristic of ethylidyne, at 1120, 1341, and 2882 cm1. The additional peaks seen in the 1400–1800 cm1 range originate from water vapor due to incomplete purge of the path of the IR beam and are not relevant to the chemistry studied here. (Reprinted with permission from Ref. [34]. Copyright (2017) American Chemical Society)

on ZrO2 at ambient pressure and led to the formation of several oxygenated adsorbates. The PM-RAIR spectra (Fig. 3.8) were used along with near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) to identify the surface species. Upon exposure to CO2 and H2O, each at 1106 mbar, a peak at 1265 cm1 is seen and is assigned to a ρ(CH2) mode of either formaldehyde or dioxymethylene. At 3104 mbar of H2O, the intensity of the 1265 cm1 peak increased, and new peaks appeared at 1383 and 1455 cm1 and were assigned to ω(CH2) and δ(CH2) modes. The peak at 1383 cm1 can also be due to the νs(OCO) mode of formate. By increasing the CO2 pressure to 7104 mbar (3104 mbar H2O), the spectrum does not change much. When the CO2 and H2O pressures were increased to 3102 mbar and 3103 mbar, respectively, peaks were observed at 1735 and 1290 cm1 and assigned to the ν(C¼O) and ν(C-OH) modes of HCOOH. By comparing this with PM-RAIRS results obtained in the presence of HCOOH(g) and HCHO(g), each at 106 mbar, Li et al. concluded that the adsorbed surface species are formate, dioxymethylene, and formaldehyde. Also present in the spectra are sharp peaks in the 1450–1800 cm1 range due to gas-phase H2O. Their presence reveals the imperfect discrimination against gas-phase species with PM-RAIRS [42].

3

Reflection Absorption Infrared Spectroscopy

61

1473 G(CH3) U(CH3) 1144

Qas(CH3) 2959 2832 Qs(CH3)

100 K

Q(CO)

90 K

80 K

1029 1045

PM IRRAS intensity (arb.un.)

PM IRRAS intensity (arb.un.)

Q(OH) 3309

100 K

3

90 K

80 K

3400

3200

3000

2800 2200 2000 Wavenumber (cm–1)

1800

1600

1400

1200

1080 1060 1040 1020 1000 980 960 Wavenumber (cm–1)

Fig. 3.7 PM-IRRAS spectra obtained after dosing 50 L of methanol at 80, 90, and 100 K on Pd(111). (Reprinted with permission from Ref. [35]. Copyright (2019) American Chemical Society)

3.2.4 343 K

Q(CO) 1735 Q(CO) 1600

3×10–2 CO2 3×10–3 H2O

Zs(CH2) 1383 G(CH2) 1455

mbar

Intensity (a.u.)

1×10–3

Co-adsorption

7×10–4 CO2 3×10–4 H2O

Q(COH) 1290

U(CH2) 1265

1×10–6 CO2 3×10–4 H2O

1×10–6 CO2 1×10–6 H2O

2000

1800

CO on Metallic, Bimetallic, and SingleAtom Alloy (SAA) Surfaces

1600

1400

1200

1000

Wavenumber (cm–1)

Fig. 3.8 In situ PM-RAIRS spectra at 343 K of different CO2 and H2O mixtures exposed to an ultrathin ZrO2 film grown on a Pt3Zr(0001) surface. (Reprinted from Ref. [42]. Copyright (2019), with permission from Elsevier)

Vibrational spectroscopy of adsorbed CO is often used to probe surface properties as the C-O stretch frequency is a sensitive indicator of the adsorption site and the balance between electron donation from the 5σ orbital and backdonation into the 2π* orbital according to the Blyholder model [43, 44]. The CO molecule is also of great importance in catalysis and environmental chemistry. It is a strong IR absorber and the C-O stretch occurs in a frequency range generally devoid of other peaks. Consequently, early RAIRS studies were primarily of adsorbed CO [45]. A more recent example is reported by Anic et al., who studied CO adsorption on reconstructed Ir(100) surfaces using RAIRS and PM-RAIRS [46]. Three different surface structures were probed: the (11) phase, the (51) reconstruction, and the oxygen terminated (21)–O surface. After exposing 1 L of CO to the (11) surface, two bands appeared, at 2061 and ~2040 cm1 (Fig. 3.9a). Both peaks are assigned to on-top CO species, the one at 2061 cm1 to CO on a normal terrace and the one at ~2040 cm1 to CO on step sites. The peak at 2028 cm1 corresponds to background CO adsorbed on the surface and then displaced upon deliberate CO exposure. This feature is also seen in the background spectrum in Fig. 3.9b. The spectra for 10 L and 50 L CO exposures have similar features to those of the 1 L dose, suggesting that a nearly saturated CO coverage is reached for relatively low exposures. The effect of gas pressure on the surface was studied through PM-RAIRS in the presence of 106 to 100 mbar CO (Fig. 3.9b). The spectra

62

R. Ranjan and M. Trenary

a)

IRAS

300 K

2063

b)

700 K

2073

300 K

PM-IRAS

300 K 2028

2028

Evacuated

50 L CO 100 mbar 10 mbar

Intensity (a.u.)

2062

10 L CO

1 mbar 10–4 mbar 10–6 mbar 2065

1 L CO 2061

Background

~2040 2000

2050 Wavenumber

2100

2150

2000

(cm–1)

2050 Wavenumber

2100

2150

(cm–1)

Fig. 3.9 (a) RAIR (IRAS) spectra of Ir(100) (11) exposed to the indicated amounts of CO at room temperature; (b) pressure-dependent PM-RAIR (PM-IRAS) spectra. (Adapted from Ref. [46] with permission1,*)

1

show a band that shifts in frequency from 2065 to 2073 cm1 with increasing pressure. Comparison of the UHV RAIRS and PM-RAIRS results reveals that slightly higher CO coverages, as indicated by the higher C-O stretch frequency, are obtained in an ambient of 104 mbar of CO but that little change is observed as the pressure is increased by six orders of magnitude to 100 mbar. Blueshifts with increasing coverage such as this are usually attributed to a combination of dipole coupling and a reduction in electron backdonation [45]. The most interesting observation is that upon evacuation to UHV, the peak frequency of 2073 cm1 is still higher than the value seen after a 50 L exposure in UHV, which compared to the 1 and 10 L exposures appeared to reach a limiting value of 2063 cm1. This implies that a higher coverage structure of CO that is stable at 300 K can be reached from ambient pressures than would be inferred from the trend with increasing exposure in UHV. This then implies that the sticking probability of CO onto the structure achieved after a 50 L exposure is so low that much higher exposures are needed to reach the true saturation coverage at 300 K. The authors also show that the high coverage achieved at 300 K in an ambient pressure of 104 mbar and above can be reached by cooling the surface to 160 K in the presence of 106 mbar of CO. A similar relationship between the RAIR spectra at 300 K under ambient pressures and the spectra

obtained by cooling in an ambient pressure of 106 mbar of CO was observed for the (51) and (21)–O surfaces [46]. Bimetallic catalysts have long been of interest as a way to achieve properties not available with the relatively small number of catalytically active metals. RAIRS and PM-RAIRS of adsorbed CO can provide valuable new insights into the surface structure and properties of model bimetallic catalysts. Yao and Goodman studied the interaction of CO with Cu-Ni supported on SiO2 thin films through PM-RAIRS under UHV and ambient pressure conditions [48]. The Cu-Ni surface was prepared on a thin film of SiO2 grown on Mo(110). The CO was adsorbed at ambient pressure on 5 MLE Cu-5 MLE Ni/SiO2 (1 monolayer equivalent (MLE) ¼ 1.431015 atom per cm2). The in situ PM-RAIR spectra of CO at pressures from 1106 to 5 Torr are shown in Fig. 3.10. In UHV; there is only one peak at about 2000 cm1, which is attributed to atop CO on Cu sites. At 0.05 Torr of CO, two peaks at 2117 and 2078 cm1 are observed that are assigned to CO adsorption on Cu that is directly attached to the Ni surface and CO adsorption on a Cu cluster or multilayer Cu, respectively. In the presence of 0.5 Torr of CO, adsorption on Cu sites was replaced by adsorption on Ni sites as revealed by the peak at 2043 cm1. This peak was unchanged as the pressure was increased to 5 Torr. The disappearance of CO from Cu sites

https://pubs.acs.org/doi/10.1021/acs.jpcc.5b12494 *Further permissions related to the material excerption should be taken from ACS

Reflection Absorption Infrared Spectroscopy

63

indicates that Cu-Ni/SiO2 bimetallic nanoparticles are covered by a Ni layer at high CO pressures. It also indicates that there are few monometallic Cu clusters on the bimetallic surface. After pumping down to UHV, the on-top CO peaks shift to 2031 cm1, and a broad peak at 1900 cm1 appears

2031

Cu-Ni/SiO2

2022 2043

UHV overnight PM-IRRAS intensity (a.u.)

1900 2117 Pump down to UHV 5.00 torr 1.00 torr 0.50 torr 2078

0.10 torr 0.05 torr 2000

2200

2100

1×10–6 torr

2000 Wavenumber (cm–1)

1900

1800

Fig. 3.10 PM-IRRAS of CO adsorption on a 5.0 MLE Cu-5 MLE Ni/SiO2 surface at the indicated CO pressures varying from 1  106 Torr to 5.0 Torr, and subsequent pumping down to UHV, and pumping in UHV overnight. (Adapted from Ref. [47]. Reproduced by permission of The Royal Society of Chemistry)

a) 200 K

b)

PM-IRAS

PM-IRAS

c)

XPS C1s 283.9

2075 Redosed 1E–4 mbar

10 mbar CO 1 mbar CO

CO pressure

100 mbar CO

1E–4 mbar CO

PM-IRAS intensity (a.u.)

PM-IRAS intensity (a.u.)

2075

that is assigned to CO at Ni bridging sites. The bridging site CO disappears after pumping overnight, but the on-top CO peak remains and shifts to 2022 cm1. This result contrasts with the finding under UHV conditions, where the surface of the bimetallic Cu-Ni/SiO2 catalyst is enriched in Cu. As the RAIRS results indicate, the CO can alter the surface composition, a fact that must be considered when using CO as a probe [47]. In another study, Anic et al. employed PM-RAIRS and XPS to investigate a model Ni-ZrO2 catalyst prepared by evaporating Ni onto a thin ZrO2 film grown on a Pd3Zr (0001) crystal in UHV [48]. This was motivated by the importance of Ni-ZrO2 in various catalytic methane reforming reactions. In addition to studying the model catalyst, they also studied methane reforming over a technological catalyst consisting of Ni nanoparticles on a ZrO2 support. In the latter studies, they employed FTIR for operando measurements of the catalyst and of the surface species present during the reactions. This permits direct comparison of PM-RAIRS data for a model catalyst with conventional FTIR measurements of the technological catalyst under similar reaction conditions. The spectra for the structurally welldefined model catalyst can be quite useful for interpreting the spectra of the more complex supported technological catalyst. They studied CO interaction with the model reforming catalysts through PM-RAIRS and XPS as shown in Fig. 3.11 [49]. CO is a product of methane dry reforming and of the reverse water-gas shift reaction. CO PM-RAIRS spectra were collected at 200 K from 106 mbar to 100 mbar (Fig. 3.11a). The peak at 2075 cm1 gradually increases

Annealed to 550K

CO at 200K Annealed to 450K Annealed to 300K

Intensity (a.u.)

3

Annealed to 350K Annealed to 300K

Evacuated at 200K

Annealed to 200K

1E–6 mbar CO

CO saturated 90K 284.9

2200 2100 2000 1900 1800 1700 Wavenumber (cm–1)

2200 2100 2000 1900 1800 1700 Wavenumber (cm–1)

Fig. 3.11 CO adsorption on the Ni-ZrO2/Pd3Zr(0001) model catalyst monitored by PM-RAIRS and XPS. (a) CO adsorption at 200 K at increasing pressure; (b) PM-RAIR spectra after adsorption and evacuation, followed by heating to 300 K. Upon a second CO exposure at

294

292

290

288 286 284 282 Binding energy (eV)

280

278

276

200 K, the CO band at 2075 cm1 did not reappear; (c) XPS spectra of the C1s region after stepwise annealing of the CO saturated surface. (Reprinted by permission from Ref. [48]. Copyright (2016), with permission from Springer Nature)

3

64

R. Ranjan and M. Trenary

with exposure and is attributed to atop CO chemisorbed on individual nickel atoms. CO adsorption on the ZrO2 trilayer is not detected as its desorption temperature is 155 K [49]. After evacuation of the reaction chamber, the atop CO peak was present at 200 K (Fig. 3.11b) as CO desorbs from Ni (111) and Ni(100) at ~425 K [50, 51]. The signal of atop CO vanished at 300 K, which is below the expected CO desorption temperature. The peak at 2075 cm1 is not seen again after recooling and redosing 104 mbar CO at 200 K (Fig. 3.11b). To understand the reason for the disappearance of this peak upon annealing, an XPS study was performed at 90 K following a 50 L CO exposure. The C1s and Ni2p regions were scanned. The spectrum gave peaks at 284.9 eV (Fig. 3.11c) and 852.7 eV (not shown), which are characteristic of molecular CO and metallic Ni, respectively. Upon annealing, the C1s peak changes from 284.9 eV at 90 K (molecular CO) to 283.9 eV at 550 K (sp2-bound carbon), while the Ni2p peak is not affected. From this observation, the authors inferred that CO dissociates on nickel and forms a carbon layer on the surface that hinders CO re-adsorption. There is no Ni sintering as the Ni2p peak does not change. This study again demonstrates the power of the combined use of PM-RAIRS and XPS in the investigation of the surface chemistry of a model catalyst. Weng et al. used RAIRS to probe CO oxidation over partially oxidized Pd surfaces in the presence of ambient pressures of CO and O2 [52]. Three O/Pd(100) model surfaces were investigated: p(22) chemisorbed O, (√5√5)

b) 0.01Torr O2

90 K

0.01Torr O2 0.01Torr CO

120 K

2128

90 K 100 K

0.03Torr CO

130 K

230 K 300 K

130 K

IR intensity (a.u.)

IR intensity (a.u.)

190 K

140 K 150 K 160 K 190 K 230 K 300 K

400 K 2110

0.5%

2500

2300

2100

90 K 150 K 190 K 230 K 300 K 350 K 375 K 400 K 425 K 450 K 475 K

After evacuating

300 K 2108

400 K

2107 2001

0.005Torr CO 0.04Torr CO

110 K 120 K

140 K 150 K

c)

1943

0.005Torr CO

IR intensity (a.u.)

a)

R27
surface oxide, and bulk-like surface oxide (PdO). These surfaces were characterized by low-energy electron diffraction (LEED) and Auger electron spectroscopy (AES). When the p(22) O/Pd(100) surface was exposed to 0.01 Torr of O2 at 90 K, no peaks were observed, nor expected, in the range of 1700–2500 cm1 (Fig. 3.12a). In the presence of 0.01 Torr of CO, peaks at 1986 cm1 and 2100 cm1 were observed and were assigned to CO on twofold bridging sites of Pd(100) and atop CO on the Pd atom bonded directly to oxygen, respectively (Fig. 3.13). The 1986 cm1 peak intensity increased after annealing, while the 2100 cm1 peak vanished after heating to 300 K. At 400 K, a new peak at 2001 cm1 appeared which corresponds to CO saturation adsorption on the Pd(100) surface. On the (√5√5) R27
-O surface, peaks at 2128 cm1, 1994 cm1, and 1943 cm1 were seen at 90 K (Fig. 3.12b), which were assigned to CO adsorbed on the partially oxidized Pd surface (Fig. 3.13). As the surface temperature increased to 130 K, the peaks at 2128 cm1 and 1943 cm1 disappeared, and new peaks at 1994 and 2110 cm1 formed. Similar features were seen for the p(22) O/Pd(100) surface. This suggests that in a background of a CO and O2 gas mixture, the (√5√5)R27
-O surface is reduced to the p(22) O/Pd(100) surface in the temperature range of 100–130 K. On the bulk-like surface oxide, no peak is seen until 300 K. The IR band at 2108 cm1 appeared at 350 K followed by a 1994 cm1 peak at 375 K (Fig. 3.12c). These features are similar to CO adsorption on the metallic surface. It indicates that the bulk-like PdO

0.5%

1986 1900

Wavenumber (cm–1)

0.5%

1994

1995 2005

2001 1700 2500

2300

2100

1900

Wavenumber (cm–1)

Fig. 3.12 RAIRS spectra for CO adsorption as a function of surface temperature on three O/Pd(100) model surfaces. (a) p(22) chemisorbed O; (b) (√5√ 5)R27
surface oxide; (c) bulk-like surface

1700 2500

2300

2100

1900

1700

Wavenumber (cm–1)

oxide. (Reprinted by permission from Ref. [52]. Copyright (2014), with permission from Springer Nature)

3

Reflection Absorption Infrared Spectroscopy

65

O

O

C

C

Pd0

PdG+

2107 cm–1

2128 cm–1

Fig. 3.13 Schematics for IR bands of 2107 cm1 and 2128 cm1. (Reprinted by permission from Ref. [52]. Copyright (2014), with permission from Springer Nature)

surface reduces to metallic Pd at 350 K. This study shows that all three O/Pd surfaces are reduced to a metallic surface by the CO/O2 mixture. In a UHV study, Thuening et al. used RAIRS and temperature-programmed desorption (TPD) to study the low temperature reaction of CO with O on Au(111) surfaces [53]. The O-covered Au(111) surfaces were prepared through exposure to ozone. The TPD results in Figure 3.14a reveal that CO desorbs from Au(111) covered by 0.2 ML of O in two distinct states at 132 and 174 K with the intensity increasing with CO exposure. This leads to CO2 formation in the low temperature region, with the first peak appearing at 168 K that intensifies with increasing CO exposure and later another peak appears at 142 K (Fig. 3.14b). The 168 K CO2 peak is associated with the more stably adsorbed CO state

a)

b) 142

CO on Au(111) + 0.2 ML Oads

168

132

28 amu signal

200

250

300

350

0 L CO 0.05 L CO 0.1 L CO 0.5 L CO 1 L CO 3 L CO

44 amu signal

0 L CO 0.05 L CO 0.1 L CO 0.5 L CO 1 L CO 3 L CO

174

150

CO on Au(111) + 0.2 ML Oads

400

150

200

T/K

250

300

350

400

T/K

c) CO on Au(111) + 0.2 ML Oads 500

32 amu signal

0 L CO 0.05 L CO 0.1 L CO 0.5 L CO 1 L CO 3 L CO

470

350

400

450

500

550

600

T/K

Fig. 3.14 TPD profiles [(a) 28 (CO), (b) 44 (CO2), and (c) 32 (O2) amu] as a function of CO exposures at 80 K to a Au(111) surface precovered by 0.2 ML of atomic oxygen using a heating rate of 3.4 K/s. (Reprinted from Ref. [53]. Copyright (2015), with permission from Elsevier)

3

66

R. Ranjan and M. Trenary

that desorbs at 174 K. The decrease in oxygen (32 amu) intensity confirms that there is a reaction between adsorbed CO and adsorbed oxygen (Fig. 3.14c). Figure 3.15 shows RAIR spectra for CO on clean and oxygen-covered Au(111). Following a 10 L CO exposure to the clean surface at 87 K, a peak is seen at 2110 cm1 that is assigned to CO bonded to gold in the zero oxidation state, Au0 (Fig. 3.15a). The intensity of this peak fell as the surface was annealed to the indicated temperatures for ~10 s and then cooled back to 87 K, where the spectra were acquired. Two peaks at 2135 and 2110 cm1 are observed when CO is exposed to Au(111) covered with 0.2 ML of oxygen

(Fig. 3.15b). Both peaks rise together as CO exposure is increased, consistent with the growth of the two desorption peaks seen with TPD (Fig. 3.14a). Comparing these spectra with the spectra in Fig. 3.15a indicates that the infrared band at 2135 cm1 is induced by co-adsorbed atomic oxygen, whereas 2110 cm1 is due to unaffected sites. The peak at 2135 cm1 is related to CO adsorption on Auδ+ sites, which is generally higher in frequency than CO on Au0 sites. A 10 L CO exposure leads to a shift in frequency from 2135 to 2141 cm1 as shown in Fig. 3.15c. The variation in integrated peak area as a function of annealing temperature indicates that the low-temperature 132 K desorption peak (Fig. 3.14a)

a) 2110

87 K 93 K 106 K 118 K 130 K 141 K 153 K 168 K 174 K 178 K

2200

2150

I/I (87 K)

Absorbance

CO/Au(111)

1.0 0.8 0.6 0.4 0.2 0.0 100 120 140 160 180 T/K

2100

2050

2000

1950

Frequency (cm–1)

c)

0 L CO 0.3 L CO 0.6 L CO 0.9 L CO 1.2 L CO

2135

2200

2100

2000

Frequency

1900

1800

(cm–1)

Fig. 3.15 (a) RAIR spectra of CO on Au(111) at 87 K and heated to various temperatures. Inset shows the variation in integrated absorbance with temperature. (b) CO adsorbed on Au(111) with 0.2 ML of atomic oxygen as a function of CO exposure. (c) RAIR spectra of a saturated overlayer of CO (10 L exposure) on Au(111) with 0.2 ML of atomic

1.0 0.8 0.6 0.4 0.2 0.0

174

132

2141 cm–1 2115 cm–1

100 120 140 160 T/K

Absorbance

Absorbance

2110

2110

2141

CO on Au(111) + 0.2 ML Oads

I/I (96 K)

b)

2200

2100

2000

1900

96 K 100 K 106 K 110 K 115 K 120 K 133 K 137 K 144 K 155 K 167 K 1800

Frequency (cm–1)

oxygen as a function of annealing temperature. Inset shows the variation in integrated absorbance of the 2141 and 2110 cm1 peaks as a function of annealing temperature. (Reprinted from Ref. [53]. Copyright (2015), with permission from Elsevier)

3

Reflection Absorption Infrared Spectroscopy

67

is associated with CO attached to Auδ+ sites and the 174 K desorption peak is associated with CO adsorbed on Au0 sites. This study illustrates the valuable insights that can be gained through the combined use of the complementary techniques of RAIRS and TPD. In contrast to the study of Thuening et al. [53] who used RAIRS to study CO on Au(111) in UHV at low temperature, Piccolo et al. obtained RAIR spectra of CO on Au(111) in the pressure range of 103–103 Torr at room temperature [54]. They describe their spectra as polarized, by which they mean they displayed (Ip  Is)/(Ip þ Is), thus eliminating signal from gas-phase CO. They do not state if they used polarization modulation or merely took the difference between p- and s-polarized spectra from different experiments [54]. Their RAIR spectra shows one peak at 2060 cm1 that is attributed to CO on top of gold atoms (Fig. 3.16). The peak appears for pressures higher than ~1 Torr and has a negligible shift in frequency with increase in pressure up to 100 Torr, which is attributed to a lack of CO-CO interactions due to a low CO coverage. They note that 2060 cm1 is a low wavenumber for CO on Au surfaces. It is also low compared to the value of 2110 cm1 seen in Fig. 3.15a by Thuening et al. [53]. Piccolo et al. [54] used DFT to calculate frequencies for CO on several Au surfaces and found a good match between their experimental value and CO adsorbed at the step sites of Au (331) and the kink sites of Au(874) (Fig. 3.17). The calculated binding energies were found to be in the order: terraces < steps < kinks < adatoms. They also used STM to show

2060 cm–1

0.024

25 cm–1

Polarized RAIRS intensity

0.020

100 Torr

0.016

Au(111)

Au(331)

Au(874)

Au(111)-ad

30 Torr

0.012

10 Torr

0.008

3 Torr

0.004 0.000 2200

that elevated CO pressures removed the herringbone reconstruction of Au(111) to produce a 11 surface (Au(111)(22√3!11). From these results, they conclude that the peak they observe at 2060 cm1 for CO pressures 1 Torr is due to CO adsorbed on top of Au atoms at defect sites of the Au(111) surface. This would also explain the large difference in frequency from the value reported by Thuening et al. [53] at low temperature in UHV, where the CO was bound on the top sites of the (111) terraces. In a recent PD-RAIRS study by Kruppe et al., CO was used as a probe molecule to characterize the structure and to quantify the Pd coverage on a Pd/Cu(111) single-atom-alloy (SAA) surface [16]. It was found that at room temperature, CO only adsorbs on the surface Pd atoms and not on Cu sites. Figure 3.18a shows that low CO exposures under UHV conditions to a Cu(111) surface with 0.14 monolayer (ML) of Pd leads to a peak at 2064 cm1 that increases in intensity as the exposure increases. However, the peak frequency does not change significantly with exposure, indicating that the CO molecules are well separated from each other. This is consistent with the isolated nature of the Pd atoms on this SAA [55–58]. This peak is assigned to atop CO on Pd sites. The fact that no peak is observed for bridge site CO further indicates that the Pd atoms exist as isolated atoms, rather than as dimers or other aggregates. The maximum CO coverage achievable under UHV conditions is quite low, consistent with a weak heat of adsorption for CO on the Pd atoms of the Pd/Cu(111) SAA. In contrast, on Pd(111), CO adsorbs at room temperature in UHV with a saturation coverage of 0.6 ML [59, 60]. Higher CO coverages can be achieved on the Pd/Cu(111) SAA under an ambient pressure

1 Torr

2150

2100

2050

2000

1950

1900

Wavenumber (cm–1)

Fig. 3.16 RAIR spectra of CO adsorbed on Au(111) at RT for various CO pressures. The primary single-beam spectra were averaged over 1024 scans with a total recording time of ~10 min for each polarization. (Reprinted from Ref. [54]. Copyright (2004), with permission from Elsevier)

Fig. 3.17 Top views of the optimized geometries for adsorption of CO on various gold surfaces. Au(331) and (874) are vicinal surfaces of Au (111). “ad” means Au adatom. The lighter metallic planes indicate the upper terraces, and the darker planes, the lower terraces. C and O are represented by small balls, C being darker. (Reprinted from Ref. [54]. Copyright (2004), with permission from Elsevier)

3

68

R. Ranjan and M. Trenary

of CO. Figure 3.18b shows p-polarized spectra of Pd-free Cu (111) (red) and of 0.14 ML Pd/Cu(111) (blue) under an ambient pressure of 1102 Torr of CO. In both the red and blue spectra in Fig. 3.18b, the P- and R-branch envelopes of gas-phase CO are seen as weak broad peaks at 2115 and 2176 cm1. A strong peak due to CO adsorbed on Pd atoms is seen at 2068 cm1, which is only slightly shifted from the position for the much weaker peak in UHV in Fig. 3.18a. The C-O stretch peak area was then used to quantify the Pd coverage. Because of the high signal-to-noise ratio observed for the CO peak, a Pd coverage as low as 0.004 ML of Pd was detectable by this method. The 2063 cm1 peak decreases in intensity after evacuation to UHV (Fig. 3.18c) as CO slowly desorbs from the Pd sites. A Pd coverage up to 0.2 ML can be Fig. 3.18 RAIR spectra for CO exposed to 0.14 ML Pd/Cu(111) at 300 K. In (a) the CO exposure increased from 0.5 L to a total of 10 L. (b) Spectra from the sample exposed to a static pressure of 1 102 Torr of CO. The red spectrum overlaid in (b) is from a Pd-free Cu(111) surface. The IR cell was then evacuated to UHV, where the last spectrum was taken. (c) Repeated scans (scan time ¼ 4 min, scans taken immediately after each other) after the cell had been evacuated. The last spectrum was obtained after annealing the sample to 500 K for 2 min. (Reprinted with permission from Ref. [16]. Copyright (2017) American Chemical Society)

quantified by this method. However, at higher coverages, the Pd atoms form aggregates on the surface as confirmed by the appearance of bridge site CO [16]. As another example of the use of RAIRS of adsorbed CO to probe surface structure, Noei et al. studied a novel system consisting of Ir nanoclusters grown on a graphene monolayer on the Ir(111) surface [61]. Due to the lattice mismatch between graphene and the Ir(111) surface, a periodic moiré pattern is observed with a lattice constant of ~25 Å. The moiré pattern can serve as a template for the growth of ordered arrays of metal nanoclusters [62, 63]. For comparison to the results for CO on the Ir nanoclusters, Noei et al. also obtained spectra versus CO exposure for the clean Ir(111) surface, as shown in Fig. 3.19. These results agree with an

a)

b) 0.5 L CO

0.0001

1 × 10–2 Torr CO

+ 0.5 L CO 'RP RP

'RP RP

+ 1 L CO

2068

+ 3 L CO

Evacuate to UHV

0.0025 + 5 L CO 2064 1900

2000

2100

2200

Wavenumber (cm–1)

1900

2000

c) After evacuating to UHV 2063 'RP

0.0002

Repeated scan

RP

Repeated scan

After 500 K anneal 1900

2100

2200

Wavenumber (cm–1)

2000 Wavenumber

2100 (cm–1)

2200

2300

Reflection Absorption Infrared Spectroscopy

a)

Background

A B

Clean Ir(111) 0.05 L

C D E F G

Intensity

69

b)

Clean Ir/ G/Ir(111)

2050

A

353 K

F

0.1 L

0.05 L

0.5 L 1L 3L

H

6L

I J K

10 L 50 L 100 L

L

500 L

M

1000 L

2062

0.1 L

Intensity

3

B C

2050 343 K

E D

0.5 L

D

333 K 323 K

3L

E

313 K

B

10 L

F

195 K

A

50 L

G

60 L

H

C

2043

0.1 % 0.2 %

2053 2066

2150 0.1%

20702054

2057 2070

2100 2050 2000 2100 2050 2000 2150 Wavenumber (cm–1) Wavenumber (cm–1)

2082 2100

2050 Wavenumber

2000 (cm–1)

Fig. 3.19 Adsorption of CO on a clean Ir(111) surface. RAIRS spectra obtained after exposing the clean sample to CO at 195 K. (A) Background, (B) clean sample with back pressure of CO, and dosing CO: (C) 0.05 L, (D) 0.1 L, (E) 0.5 L, (F) 1 L(G) 3 L, (H) 6 L, (I) 10 L, (J) 50 L, (K) 100 L, (L) 500 L, and (M) 1000 L. (Reprinted with permission from Ref. [61]. Copyright (2018) American Chemical Society)

earlier RAIRS study of CO on Ir(111) [64, 65]. Due to the presence of background CO, the lowest coverage peak is observed at 2043 cm1 on the clean Ir surface (Fig. 3.19b), whereas at their lowest coverages, Lauterbach et al. reported a peak as low as 2030 cm1 [64, 65]. After exposing CO at 195 K, the C-O stretch shifts to higher wavenumbers, reaching a final value of 2082 cm1. The 2082 cm1 peak is associated with the formation of a (3√ 33√ 3)R30
superstructure at a CO coverage of 0.7 ML [66]. The CO occupies on-top sites for all coverages. No IR band of CO appeared on the graphene/Ir(111) surface after dosing CO at 195 K, indicating both that CO does not adsorb on the graphene layer and that this layer completely covers the Ir (111) surface. Figure 3.20a shows CO adsorption on 0.1 ML Ir on the graphene/Ir(111) surface, which produces one peak at 2050 cm1 at low CO coverage (0.1 L) that shifts to 2054 cm1 at higher coverages. Additionally, another peak appeared at 2062 cm1 and moved to 2070 cm1 as the coverage increased. Both IR bands at 2054 and 2070 cm1

Fig. 3.20 (a) Adsorption of CO on clean 0.1 ML Ir/graphene/Ir(111). RAIRS spectra obtained after exposing the clean sample to CO at 195 K. (A) Before CO dosing, and dosing CO: (B) 0.05 L, (C) 0.1 L, (D) 0.5 L, (E) 3 L, (F) 10 L, (G) 50 L, and (H) 60 L. (b) Annealing the sample after dosing 60 L CO. (A) 60 L CO at 195 K; (B) 313 K, (C) 323 K, (D) 333 K, (E) 343 K, and (F) 353 K. (Reprinted with permission from Ref. [61]. Copyright (2018) American Chemical Society)

Fig. 3.21 Respective view of the graphene moiré unit cell with (a) 0.05 ML and (b) 0.15 ML Ir clusters on graphene/Ir(111). The Ir surface is depicted in red and the graphene layer in black. (Reprinted with permission from Ref. [61]. Copyright (2018) American Chemical Society)

remain unchanged at higher CO exposures at 195 K. The peaks at 2050 and 2070 cm1 are assigned to CO on one-layer high clusters and two- or several-layer high

3

70

R. Ranjan and M. Trenary

b)

0.1 mbar CO / Cu(100) 2093

2103

After pump-down

2074

PM-IRAS intensity (arb. units)

0.1 mbar CO / Cu(100) @ 300 K

T = 300 K

2082 T = 225 K

2112

PM-IRAS intensity (arb. units)

a)

100 80 60 40 t = 20 min

T = 200 K 2093

2000

2050

2100 Wavenumber

2150

2200

(cm–1)

Fig. 3.22 (a) PM-RAIR spectra on Cu(100) recorded in 0.1 mbar CO at 200, 225, and 300 K. The features at 2082 and 2112 cm1 are assigned to CO adsorbed on terraces and CO/Cu1 complexes, respectively. (b) Time evolution of PM-RAIR spectra recorded in 0.1 mbar CO at 300 K. The features at 2093 and 2103 cm1 are assigned to CO

clusters, respectively. As the surface is annealed to higher temperatures, Fig. 3.20b shows that the intensity of the peak at 2054 cm1 decreases, while the intensity of the one at 2070 cm1 increases. This is attributed to sintering of the one-layer high clusters into clusters of two or more layers in height. This agrees with observations with scanning tunneling microscopy (STM). The decrease in peak widths indicates that the arrangement of the CO molecules on the clusters becomes more ordered with increasing annealing temperature. Figure 3.21 depicts the arrangements of the clusters on the graphene/Ir(111) surface at Ir coverages of 0.05 and 0.15 ML as observed with STM. Although vibrational spectroscopy of adsorbed CO is often used to probe the properties of metal surfaces, a recent ambient pressure STM study showed that elevated pressures of CO can lead to the formation of Cu nanoclusters on a Cu (111) surface [67]. Earlier STM work had shown a different type of restructuring of the Cu(100) surface [68]. Recently, Roiaz et al. used PM-RAIRS to investigate structural changes to a Cu(100) surface induced by ambient pressures of CO [69]. Figure 3.22a shows spectra of CO on Cu(100) in the presence of 0.1 mbar of CO at 200, 225, and 300 K. At 200 K, two peaks at 2082 and 2112 cm1 are observed.

2000

2050

2103

2100 Wavenumber

2150

2200

(cm–1)

adsorbed at Cu clusters and adatoms, respectively. (Adapted from Ref. [69] with permission1,*) 1 https://pubs.acs.org/doi/10.1021/acs.jpcc.8b07668) *Further permissions related to the material excerption should be taken from ACS

Fig. 3.23 Calculated structures of CO adsorbed on various Cu sites: (a) (100) terraces, (b) step, (c) kink atom, (d) Cu1/Cu(100), and (e) Cu5/Cu (100). (Adapted from Ref. [69] with permission1,*) 1 https://pubs.acs.org/doi/10.1021/acs.jpcc.8b07668 *Further permissions related to the material excerption should be taken from ACS

The peak at 2082 cm1 is due to the c(22)CO overlayer on the Cu(100) terraces and the second peak at 2112 cm1 is assigned to CO adsorbed on Cu1 adatoms on the basis of

3

Reflection Absorption Infrared Spectroscopy

density functional theory (DFT) results (Fig. 3.23). Both peaks are red shifted at 300 K. Additionally, a new peak at 2093 cm1 appeared that is assigned to CO adsorbed on Cu5 clusters. These clusters are formed by coalescence of Cu1 adatoms via Ostwald ripening [70, 71]. This is also confirmed by time-dependent PM-RAIR (Fig. 3.22b) spectra as the peak at 2093 cm1 (assigned to Cu clusters) increases while the 2103 cm1 peak (assigned to CO/Cu1) decreases in intensity. This study reveals that CO exposure leads to roughening of the surface through the formation of Cu adatoms and clusters on the (100) terraces.

3.3

Conclusion

The experimental determination of the identity and properties of molecular species, including intermediates, on metal surfaces is essential for establishing reaction mechanisms in heterogeneous catalysis. Surface vibrational spectroscopies are a primary means of obtaining such information. The technique of RAIRS offers high resolution and a broad spectral range and has sufficient sensitivity to detect a variety of surface species. In many cases, definitive assignment of spectral features to a particular adsorbate can be challenging. Nevertheless, with sufficient care in the design and execution of experiments, RAIRS can now provide unique insights into surface chemical reactions. The technique is particularly attractive for studies related to catalysis because it can be used both under ultrahigh vacuum and ambient pressure conditions. Ambient pressure studies have the advantage that high coverages of reactants can be achieved at the elevated temperatures where reactions occur. High coverages can also be achieved by lowering the temperature but at the expense of suppressing reactions of interest. A key lesson gained over the many decades of surface science studies is that the use of multiple complementary techniques can provide a more complete understanding than is possible from a single method. For example, the use of RAIRS and XPS is proving to be a particularly powerful combination for ambient pressure studies. Acknowledgments We gratefully acknowledge support by a grant from the National Science Foundation, CHE-2102622. We thank Mr. Arephin Islam for preparing the drawing in Fig. 3.2.

References 1. Madey, T.E., Yates Jr., J.T.: Vibrational Spectroscopy of Molecules on Surfaces, 1. Springer Science & Business Media, New York (2013) 2. Kas, R., Ayemoba, O., Firet, N.J., Middelkoop, J., Smith, W.A., Cuesta, A.: In-situ infrared spectroscopy applied to the study of the

71 electrocatalytic reduction of CO2: theory, practice and challenges. ChemPhysChem. 20, 2904–2925 (2019) 3. Greenler, R.G.: Infrared study of adsorbed molecules on metal surfaces by reflection techniques. J. Chem. Phys. 44, 310–315 (1966) 4. Greenler, R.G.: Design of a reflection–absorption experiment for studying the ir spectrum of molecules adsorbed on a metal surface. J. Vac. Sci. Technol. 12, 1410–1417 (1975) 5. Fan, J., Trenary, M.: Symmetry and the surface infrared selection rule for the determination of the structure of molecules on metal surfaces. Langmuir. 10, 3649–3657 (1994) 6. Griffiths, P.R., De Haseth, J.A.: Fourier Transform Infrared Spectrometry, vol. 171. Wiley, New York (2007) 7. Rupprechter, G., Weilach, C.: Spectroscopic studies of surface–gas interactions and catalyst restructuring at ambient pressure: mind the gap! J. Phys. Condens. Matter. 20, 184019 (2008) 8. Föttinger, K., Weilach, C., Rupprechter, G.: Sum frequency generation and infrared reflection absorption spectroscopy. In: Che, M., Védrine, J.C. (eds.) Characterization of Solid Materials and Heterogeneous Catalysts: From Structure to Surface Reactivity, pp. 211–253. Wiley-VCH, Weinheim (2012) 9. Méthivier, C., Pradier, C.-M.: Infra red reflection absorption spectroscopy and polarisation modulation-IRRAS. In: Che, M., Védrine, J.C. (eds.) Characterization of Solid Materials and Heterogeneous Catalysts: From Structure to Surface Reactivity, pp. 255–287. Wiley-VCH, Weinheim (2012) 10. Somorjai, G.A.: Introduction to Surface Chemistry and Catalysis. Wiley, New York (1994) 11. Schennach, R., Hirschmugl, C., Gilli, E., Tysoe, W.T.: A new method for performing polarization modulation infrared reflection-adsorption spectroscopy of surfaces. Appl. Spectrosc. 63, 369–372 (2009) 12. Szanyi, J., Goodman, D.W.: Combined elevated pressure reactor and ultrahigh vacuum surface analysis system. Rev. Sci. Instrum. 64, 2350–2352 (1993) 13. Campbell, R.A., Goodman, D.W.: A new design for a multitechnique ultrahigh vacuum surface analysis chamber with high pressure capabilities. Rev. Sci. Instrum. 63, 172–174 (1992) 14. Hollins, P., Pritchard, J.: Versatile infrared window mounting for ultrahigh vacuum applications. J. Vac. Sci. Technol. 17, 665–665 (1980) 15. Krooswyk, J.D., Waluyo, I., Trenary, M.: Simultaneous monitoring of surface and gas phase species during hydrogenation of acetylene over Pt(111) by polarization-dependent infrared spectroscopy. ACS Catal. 5, 4725–4733 (2015) 16. Kruppe, C.M., Krooswyk, J.D., Trenary, M.: Polarization-dependent infrared spectroscopy of adsorbed carbon monoxide to probe the surface of a Pd/Cu(111) single-atom alloy. J. Phys. Chem. C. 121, 9361–9369 (2017) 17. Kruppe, C.M., Krooswyk, J.D., Trenary, M.: Selective hydrogenation of acetylene to ethylene in the presence of a carbonaceous surface layer on a Pd/Cu(111) single-atom alloy. ACS Catal. 7, 8042–8049 (2017) 18. Zaera, F.: An organometallic guide to the chemistry of hydrocarbon moieties on transition metal surfaces. Chem. Rev. 95, 2651–2693 (1995) 19. Zaera, F.: Probing catalytic reactions at surfaces. Prog. Surf. Sci. 69, 1–98 (2001) 20. Bent, B.E.: Mimicking aspects of heterogeneous catalysis. Generating, isolating, and reacting proposed surface intermediates on single crystals in vacuum. Chem. Rev. 96, 1361–1390 (1996) 21. Horiuti, J., Miyahara, K.: Hydrogenation of Ethylene on Metallic Catalysts. National Standard Reference Data Series 13, National Bureau of Standards (1968) 22. Zaera, F., Janssens, T.V.W., Öfner, H.: Reflection absorption infrared spectroscopy and kinetic studies of the reactivity of ethylene on Pt (111) surfaces. Surf. Sci. 368, 371–376 (1996)

3

72 23. Sabatier, P., Senderens, J.: Comptes Rendus Des Seances De L’Academie Des Sciences Section VI—Chimie. Paris, Imprimerie Gauthier-Villars (1902) 24. Malik, I.J., Brubaker, M.E., Mohsin, S.B., Trenary, M.: Infrared vibration-rotation selection rules for chemisorbed molecules with free internal rotation: results for ethylidyne on Pt(111). J. Chem. Phys. 87(5554), 5554–5561 (1987) 25. Mohsin, S.B., Trenary, M., Robota, H.J.: Infrared identification of the low-temperature forms of ethylene adsorbed on Pt/Al2O3. J. Phys. Chem. 92, 5229–5233 (1988) 26. Mohsin, S.B., Trenary, M., Robota, H.J.: Identification of ethylenederived species on Al2O3-supported Rh, Ir, Pd, and Pt catalysts by infrared-spectroscopy. J. Phys. Chem. 95, 6657–6661 (1991) 27. Celio, H., Trenary, M., Robota, H.: A comparative IR study of ethylene adsorption on a PtRh alloy and monometallic Pt and Rh catalysts supported on Al2O3: identification of alloy-specific binding sites. J. Phys. Chem. B. 99, 6024–6028 (1995) 28. Deng, R.P., Herceg, E., Trenary, M.: Formation and hydrogenation of ethylidene on the Pt(111) surface. Surf. Sci. 560, L195–L201 (2004) 29. Deng, R.P., Jones, J., Trenary, M.: Spectroscopic characterization of vinyl formed from acetylene on Pt(111). J. Phys. Chem. C. 111, 1459–1466 (2007) 30. Kaltchev, M., Stacchiola, D., Molero, H., Wu, G., Blumenfeld, A., Tysoe, W.T.: On the reaction pathway for the formation of benzene from acetylene catalyzed by palladium. Catal. Lett. 60, 11–14 (1999) 31. Stacchiola, D., Azad, S., Burkholder, L., Tysoe, W.T.: An investigation of the reaction pathway for ethylene hydrogenation on Pd (111). J. Phys. Chem. B. 105, 11233–11239 (2001) 32. Stacchiola, D., Calaza, F., Zheng, T., Tysoe, W.T.: Hydrocarbon conversion on palladium catalysts. J. Mol. Catal. A Chem. 228, 35–45 (2005) 33. Azad, S., Kaltchev, M., Stacchiola, D., Wu, G., Tysoe, W.T.: On the reaction pathway for the hydrogenation of acetylene and vinylidene on Pd(111). J. Phys. Chem. B. 104, 3107–3115 (2000) 34. Simonovis, J., Tillekaratne, A., Zaera, F.: The role of carbonaceous deposits in hydrogenation catalysis revisited. J. Phys. Chem. C. 121, 2285–2293 (2017) 35. Kaichev, V.V., Selivanova, A.V., Tsapina, A.M., Saraev, A.A., Bukhtiyarov, V.I.: Nonclassical adsorption of methanol on palladium: the competition between adsorption of single molecules and clusters. J. Phys. Chem. C. 123, 7259–7265 (2019) 36. Arena, F., Italiano, G., Barbera, K., Bordiga, S., Bonura, G., Spadaro, L., Frusteri, F.: Solid-state interactions, adsorption sites and functionality of Cu-ZnO/ZrO2 catalysts in the CO2 hydrogenation to CH3OH. Appl. Catal. A Gen. 350, 16–23 (2008) 37. Gouvêa, D., Ushakov, S.V., Navrotsky, A.: Energetics of CO2 and H2O adsorption on zinc oxide. Langmuir. 30, 9091–9097 (2014) 38. Bukhtiyarova, M., Lunkenbein, T., Kähler, K., Schlögl, R.: Methanol synthesis from industrial CO2 sources: a contribution to chemical energy conversion. Catal. Lett. 147, 416–427 (2017) 39. Solymosi, F.: The bonding, structure and reactions of CO2 adsorbed on clean and promoted metal surfaces. J. Mol. Catal. 65, 337–358 (1991) 40. Freund, H.J., Roberts, M.W.: Surface chemistry of carbon dioxide. Surf. Sci. Rep. 25, 225–273 (1996) 41. Taifan, W., Boily, J.-F., Baltrusaitis, J.: Surface chemistry of carbon dioxide revisited. Surf. Sci. Rep. 71, 595–671 (2016) 42. Li, H., Rameshan, C., Bukhtiyarov, A.V., Prosvirin, I.P., Bukhtiyarov, V.I., Rupprechter, G.: CO2 activation on ultrathin ZrO2 film by H2O CO-adsorption: in situ NAP-XPS and IRAS studies. Surf. Sci. 679, 139–146 (2019) 43. Blyholder, G.: Molecular orbital view of chemisorbed carbon monoxide. J. Phys. Chem. 68, 2772–2777 (1964) 44. Blyholder, G.: CNDO model of carbon monoxide chemisorbed on nickel. J. Phys. Chem. 79, 756–761 (1975)

R. Ranjan and M. Trenary 45. Hoffmann, F.M.: Infrared reflection-absorption spectroscopy of adsorbed molecules. Surf. Sci. Rep. 3, 107–107 (1983) 46. Anic, K., Bukhtiyarov, A.V., Li, H., Rameshan, C., Rupprechter, G.: CO adsorption on reconstructed Ir(100) surfaces from UHV to mbar pressure: a LEED, TPD, and PM-IRAS study. J. Phys. Chem. C. 120, 10838–10848 (2016) 47. Yao, Y., Goodman, D.W.: In situ IR spectroscopic studies of Ni surface segregation induced by CO adsorption on Cu–Ni/SiO2 bimetallic catalysts. Phys. Chem. Chem. Phys. 16, 3823–3829 (2014) 48. Anic, K., Wolfbeisser, A., Li, H., Rameshan, C., Föttinger, K., Bernardi, J., Rupprechter, G.: Surface spectroscopy on UHV-grown and technological Ni–ZrO2 reforming catalysts: from UHV to operando conditions. Top. Catal. 59, 1614–1627 (2016) 49. Li, H., Choi, J.-I.J., Mayr-Schmölzer, W., Weilach, C., Rameshan, C., Mittendorfer, F., Redinger, J., Schmid, M., Rupprechter, G.: Growth of an ultrathin zirconia film on Pt3Zr examined by highresolution X-ray photoelectron spectroscopy, temperatureprogrammed desorption, scanning tunneling microscopy, and density functional theory. J. Phys. Chem. C. 119, 2462–2470 (2015) 50. Rupprechter, G., Dellwig, T., Unterhalt, H., Freund, H.J.: CO adsorption on Ni(100) and Pt(111) studied by infrared–visible sum frequency generation spectroscopy: design and application of an SFG-compatible UHV–high-pressure reaction cell. Top. Catal. 15, 19–26 (2001) 51. Miller, J.B., Siddiqui, H.R., Gates, S.M., Russell, J.N., Yates, J.T., Tully, J.C., Cardillo, M.J.: Extraction of kinetic parameters in temperature programmed desorption: a comparison of methods. J. Chem. Phys. 87, 6725–6732 (1987) 52. Weng, X., Yuan, X., Li, H., Li, X., Chen, M., Wan, H.: The study of the active surface for CO oxidation over supported Pd catalysts. Sci China Chem. 58, 174–179 (2015) 53. Thuening, T., Walker, J., Adams, H., Furlong, O., Tysoe, W.T.: Kinetics of low-temperature CO oxidation on Au(111). Surf. Sci. 648, 236–241 (2016) 54. Piccolo, L., Loffreda, D., Cadete Santos Aires, F.J., Deranlot, C., Jugnet, Y., Sautet, P., Bertolini, J.C.: The adsorption of CO on Au (111) at elevated pressures studied by STM, RAIRS and DFT calculations. Surf. Sci. 566-568, 995–1000 (2004) 55. Liu, J., Lucci, F.R., Yang, M., Lee, S., Marcinkowski, M.D., Therrien, A.J., Williams, C.T., Sykes, E.C.H., FlytzaniStephanopoulos, M.: Tackling CO poisoning with single-atom alloy catalysts. J. Am. Chem. Soc. 138, 6396–6399 (2016) 56. Darby, M.T., Stamatakis, M., Michaelides, A., Sykes, E.C.H.: Lonely atoms with special gifts: breaking linear scaling relationships in heterogeneous catalysis with single-atom alloys. J. Phys. Chem. Lett. 9, 5636–5646 (2018) 57. Darby, M.T., Sykes, E.C.H., Michaelides, A., Stamatakis, M.: Carbon monoxide poisoning resistance and structural stability of single atom alloys. Top. Catal. 61, 428–438 (2018) 58. Therrien, A.J., Hensley, A.J.R., Marcinkowski, M.D., Zhang, R., Lucci, F.R., Coughlin, B., Schilling, A.C., McEwen, J.-S., Sykes, E.C.H.: An atomic-scale view of single-site Pt catalysis for low-temperature CO oxidation. Nat. Catal. 1, 192–198 (2018) 59. Bradshaw, A.M., Hoffmann, F.M.: The chemisorption of carbon monoxide on palladium single crystal surfaces: IR spectroscopic evidence for localised site adsorption. Surf. Sci. 72, 513–535 (1978) 60. Butorac, J., Wilson, E.L., Fielding, H.H., Brown, W.A., Minns, R.S.: A RAIRS, TPD and femtosecond laser-induced desorption study of CO, NO and coadsorbed CO + NO on Pd(111). RSC Adv. 6, 66346–66359 (2016) 61. Noei, H., Franz, D., Creutzburg, M., Müller, P., Krausert, K., Grånäs, E., Taube, R., Mittendorfer, F., Stierle, A.: Monitoring the interaction of CO with graphene supported Ir clusters by vibrational spectroscopy and density functional theory calculations. J. Phys. Chem. C. 122, 4281–4289 (2018)

3

Reflection Absorption Infrared Spectroscopy

62. N’Diaye, A.T., Bleikamp, S., Feibelman, P.J., Michely, T.: Two-dimensional Ir cluster lattice on a graphene Moiré on Ir(111). Phys. Rev. Lett. 97, 215501 (2006) 63. N’Diaye, A.T., Gerber, T., Busse, C., Myslivecek, J., Coraux, J., Michely, T.: A versatile fabrication method for cluster superlattices. New J. Phys. 11, 103045 (2009) 64. Lauterbach, J., Boyle, R.W., Schick, M., Mitchell, W.J., Meng, B., Weinberg, W.H.: The adsorption of CO on Ir(111) investigated with FT-IRAS. Surf. Sci. 350, 32–44 (1996) 65. Lauterbach, J., Boyle, R.W., Schick, M., Mitchell, W.J., Meng, B., Weinberg, W.H.: Erratum to “The adsorption of CO on Ir(111) investigated with FT-IRAS” [Surface Science 350 (1996) 32]. Surf. Sci. 366, 228 (1996) 66. Grånäs, E., Andersen, M., Arman, M.A., Gerber, T., Hammer, B., Schnadt, J., Andersen, J.N., Michely, T., Knudsen, J.: CO intercalation of graphene on Ir(111) in the millibar regime. J. Phys. Chem. C. 117, 16438–16447 (2013) 67. Eren, B., Zherebetskyy, D., Patera, L.L., Wu, C.H., Bluhm, H., Africh, C., Wang, L.W., Somorjai, G.A., Salmeron, M.: Activation of Cu(111) surface by decomposition into nanoclusters driven by CO adsorption. Science. 351, 475–478 (2016) 68. Eren, B., Zherebetskyy, D., Hao, Y.B., Patera, L.L., Wang, L.W., Somorjai, G.A., Salmeron, M.: One-dimensional nanoclustering of the Cu(100) surface under CO gas in the mbar pressure range. Surf. Sci. 651, 210–214 (2016) 69. Roiaz, M., Falivene, L., Rameshan, C., Cavallo, L., Kozlov, S.M., Rupprechter, G.: Roughening of copper (100) at elevated CO pressure: Cu Adatom and cluster formation enable CO dissociation. J. Phys. Chem. C. 123, 8112–8121 (2019) 70. Ostwald, W.: Studien über die Bildung und Umwandlung fester Körper. Z. Phys. Chem. 22, 289–330 (1897) 71. Baldan, A.: Review Progress in Ostwald ripening theories and their applications to nickel-base superalloys part I: Ostwald ripening theories. J. Mater. Sci. 37, 2171–2202 (2002)

73

3 Ravi Ranjan is currently a Ph.D. candidate in the Chemistry Department at the University of Illinois Chicago. He completed his integrated bachelor’s and master’s dual degree from the Indian Institute of Science Education and Research Mohali (India) in Chemistry in 2017. His current research work focuses on the adsorption and hydrogenation of hydrocarbons on the Pd(111) surface.

Michael Trenary is a Professor of Chemistry at the University of Illinois Chicago. A major theme of his research program is the use of reflection absorption infrared spectroscopy for the identification and characterization of surface intermediates that form during chemical reactions related to problems in heterogeneous catalysis. Thirty students have received their PhD degrees under his supervision.

4

Raman Spectroscopy Jisue Moon

, Meijun Li

, Anibal J. Ramirez-Cuesta

, and Zili Wu

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76

4.4.3 Increase Spatial Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.2 Description of Raman Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1 Theory of Raman Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 Benefits of Raman Spectroscopy for Characterization of Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.3 Limitations of Raman Spectroscopy for Characterization of Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.4 Comparison of Method to Other Techniques: Pros and Cons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76 76

4.5

4.1

4.3

Description of Reaction Cells for In Situ and Operando Raman Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

84

78

Chronology of Application of Raman Spectroscopy to Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.1 Early Ambient Conditions and In Situ Condition . . . . . . . . . 4.6.2 Reaction Conditions: In Situ/Operando Measurement . . . .

86 86 89

79

4.7

Time-Resolved Raman Spectroscopy . . . . . . . . . . . . . . . . . . . .

91

4.8

Spatial-Resolved Raman Spectroscopy: Microscopy . . .

92

4.9

Modulation Excitation Raman Spectroscopy . . . . . . . . . . .

94

4.10

Applications of Raman Spectroscopy to Catalyst Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

96

Applications of Raman Spectroscopy Study to Catalyst Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

4.6 77

Description of General Raman System to Conduct Characterization of Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . Excitation Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Illumination and Light Collection System . . . . . . . . Sample Holder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Detection System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

80 80 81 82 82

4.4 New Instrumental Advances in Raman Spectroscopy . . . 4.4.1 Avoidance of Fluorescence Effect . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Increase Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

82 82 83

4.3.1 4.3.2 4.3.3 4.3.4

J. Moon Radioisotope Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA e-mail: [email protected] M. Li Manufacturing Science Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA e-mail: [email protected] A. J. Ramirez-Cuesta Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA Neutron Technologies Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA e-mail: [email protected] Z. Wu (*) Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, TN, USA Radioisotope Science Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA e-mail: [email protected]

84

4.11 4.12

Applications of Raman Spectroscopy to Catalyst Structure-Activity Relationships . . . . . . . . . . . . . . . . . . . . . . . . . 100

4.13

Combining Raman Spectroscopy with Other Techniques (Multimodal) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

4.14

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Abstract

Raman spectroscopy is one of the mostly utilized optical spectroscopic tools for revealing both the catalyst structure and surface chemistry in heterogeneous catalysis. It has recently seen increasing role in catalysis research, thanks to the development of new Raman instrumentations, reactors, and combination with other techniques, leading to in situ and operando studies with significant temporal and spatial resolutions. This chapter aimed to provide a general overview of the applications of Raman spectroscopy in heterogeneous catalysis. It starts with an introduction to the fundamentals of Raman scattering including theory and pros and cons for catalysis research; followed by a

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_4

75

76

J. Moon et al.

The principle of the Raman effect is based on the inelastic light-scattering process between incident light and an irradiated substance. Raman spectra are acquired by irradiating a sample with a laser source that is available to do visible or near-infrared monochromatic radiation, and the scattered light is usually observed in the perpendicular direction to the incident beam. Figure 4.1 shows a section of Raman spectrum that was obtained by irradiating a sample of carbon tetrachloride with a laser source that has the wavelength of 488 nm. The scattered light includes three types of emitted irradiation: Stokes scattering, anti-Stokes scattering, and Rayleigh scattering. Rayleigh scattering shows more intense than either of the other two types of scattering. This is due to the wavelength that exactly matches with that of the excitation source, and there is no energy transfer occurring between the incident light and the scattered light. However, when the photons fall into a new energy level, which is different from the initial energy level, energy transfer happens and results in the laser photon energy being shifted down or up, which gives information about the vibrational mode in the system. In a typical Raman spectrum, the x-axis shows the difference in wave numbers between the observed irradiation and that of the source, which is called as the wave number shift, Δv. Identical Raman peaks are found on both sides of the Rayleigh peak with different intensity. In the case of CCl4 (Fig. 4.1), Stokes lines are found at wave

Rayleigh Stokes

Anti-stokes

–459

–314

Raman spectroscopy, a branch of vibrational Spectroscopy, was first observed experimentally by Raman and Krishnan in 1928 [1]. It is specialized in measuring a frequency shift of inelastically scattered light from materials when incident photons of light strike a molecule and produce a scattered photon. The difference in frequency between incident and scattered light provides information of lattice vibrations. The earliest application of Raman spectroscopy in catalysis is back to 1970s. Early Raman experiments were conducted to study the structure of supported molybdenum oxide catalysts for hydrodesulfurization of petroleum products and coals [2, 3]. However, for many years, Raman spectroscopy had limited applications because of very low efficiency of the normal Raman scattering and the limitation of instrumentation. In the past decades, significant improvements of lasers and spectrometers had happened to Raman spectroscopy. This allows to overcome limitations and changes capabilities of the spectroscopy fundamentally. Today, Raman spectroscopy is one of the most powerful and efficient techniques in catalysis for characterizing the structure of catalysts and investigating the bulk and surface chemistry occurring during reaction conditions. The high resolution of Raman spectroscopy (about 1 cm
1) and its wide spectral range (50–5000 cm
1) make it possible to study the nature of vibration features of molecular species (e.g., NH3, peroxide ions) [4, 5], identification of the crystalline oxide phase, and determination of the structure of noncrystalline surface species. Due to these advantages, the application of Raman spectroscopy to heterogeneous catalysis has been extensively discussed in excellent reviews [6–15] and books [16–20]. In this chapter, we will present a general overview of Raman spectroscopy including theoretical

Theory of Raman Scattering

–218

Introduction

4.2.1

218

4.1

Description of Raman Method

314

Raman · Catalysis · In situ · Operando · Instrumentation · UV Raman · Time-resolved Raman · Spatial-resolved Raman · Raman imaging · Modulation excitation Raman · Structure-activity relationships · Reaction mechanism

4.2

459

Keywords

background, instrumentations, and the applications of Raman spectroscopy in the field of heterogeneous catalysis.

Intensity

description of the typical setup of a Raman system and recent advances in Raman instrumentations; then a chronology of the applications of Raman spectroscopy for ex situ, in situ, and operando studies of catalysis; elucidations of the advances in improving the temporal and spatial resolution of Raman spectroscopy of catalysis; Raman application case studies related to catalyst synthesis, treatments, and function under reaction conditions; illustrations of the power of multimodal approach including Raman spectroscopy in catalysis research; and ended with a brief summary and a future outlook.

Raman shift (cm–1)

Fig. 4.1 Raman spectrum of CCl4 excited by laser radiation of 488 nm. (Adapted from ref. [23], Copyright 2003, with permission from Elsevier)

4

Raman Spectroscopy

77

numbers that are located at 218, 314, and 459 cm
1, which is smaller in energy than the Rayleigh peaks, while the anti-Stokes lines existed at
218,
314, and
459 cm
1, which is greater in energy than the Rayleigh peaks. The Stokes scattering is generally used due to its intense peak intensity compared with anti-Strokes lines. However, it is noteworthy that fluorescence may interfere with the observation of Stokes shifts, but this is not affected by anti-Stokes li In this manner, for fluorescing samples, anti-Stokes signals may be more useful [21, 22]. It is also important to understand that the frequency of Raman shifts is independent of the wavelength of excitation, and thus, pattern shifts can be obtained with different excitation source and wave numbers. According to the classical theory, Raman scattering can be explained in terms of polarizability. Let us assume that the irradiation beam has a frequency of v0, and the sample is exposed to this beam. The strength of electric field E of this irradiation can be described by the following equation: E ¼ E0 cosð2πv0 tÞ where E0 is the amplitude of the wave and t is time. When the electric field of the radiation interacts with an electron from the chemical bonding from the molecules, it induces an electric dipole moment, P. P ¼ αE ¼ αE0 cosð2πv0 tÞ α is called a polarizability, which is a proportionality constant in an electric field. The polarizability is an important factor in determining whether the vibration is active in Raman spectra or not. In order to be Raman active, the polarizability, α, should vary as a function of the distance between nuclei through the following equation:  α ¼ α0 þ

@α @r

 r
r eq




where α0 is the polarizability at the equilibrium internuclear distance (req) and r is the internuclear separation at any distance. The intermolecular separation distance can be varied with the frequency of the vibration vv through the following equation: r
r eq ¼ r m cosð2πvv tÞ where rm is the maximum internuclear separation relative to the equilibrium position. By applying this intermolecular separation distance into the polarization equation, we can then obtain an expression for the induced dipole moment P.

P ¼ αE ¼ αE0 cosð2πv0 tÞ     @α ¼ α0 þ r m cosð2πvv tÞ ½E0 cosð2πv0 tÞ @r   @α ¼ α0 E0 cosð2πv0 tÞ þ r E cosð2πvv tÞ cosð2πv0 tÞ @r m 0 By using trigonometry, the equation can be simplified as the following equation:

4

  Er @α P ¼ α0 E0 cosð2πv0 tÞ þ 0 m cos½2π ðv0
vv Þt @r 2   E0 r m @α þ cos½2π ðv0 þ vv Þt 2 @r The first term represents oscillating dipole that radiates with the frequency of beam, v0, which corresponds to Rayleigh scattering. The second and third terms correlate to the Raman scattering of frequency v0 þ vv (anti-Stokes) and v0
vv (Stokes). It is noteworthy that Raman scattering requires that the polarizability term varies as a function of distance. In other   @α words, must be greater than zero to be Raman active. @r

In the case of small molecules, it is easy to see whether the polarizability changes during the vibration. Let us consider the diatomic molecule such as CO2. The electron clouds have an elongated shape with circular cross-section. If the size, shape, or orientation of polarizability ellipsoid changes during normal vibration, then the vibration is Raman active. Although the vibrations of v1 and v3 show size change of ellipsoid, v1 is Raman active and v3 is not. This is due to the difference between polarizability term of v1 and v3 vibration.   @α The difference can be shown in Fig. 4.2, which shows @r

of v1 is nonzero, but zero for v3 vibration.

4.2.2

Benefits of Raman Spectroscopy for Characterization of Catalysts

Catalysis is a key technology offering effective routes for many chemical processes. Catalysts must possess several properties to do the job in catalytic process such as activity, selectivity, durability, and regenerability. To optimize chemical reactions and design effective catalytic materials, it is important to understand the nature of catalyst active site and reaction mechanism underlying catalytic process, which yet remains challenge. Over the last decades, a wide variety of characterization techniques have been developed, and among them, Raman spectroscopy is one of the most effective tools to examine over various catalysts owing to following advantages. First, Raman spectroscopy does not have any special requirement for sample preparations and can be extensively

78

J. Moon et al.

n1

n3

a0

a0

+q

q=0 (da/dq)0 ≠ 0

–q

+q

–q

q=0 (da/dq)0

0

Fig. 4.2 Difference between v1 and v3 vibrations in CO2 molecules. (Adapted from ref. [23], Copyright 2003, with permission from Elsevier)

used on various type of materials: liquid, gas, solid (single crystals, glass, powder, amorphous materials, nanocrystalline), surface adsorbates, and their mixture. Thus, it is potentially adapted in wide range of the catalytic materials to address important catalytic problems that cannot be examined by other spectroscopies. For examples, Raman spectroscopy could directly monitor the molecular event that occur in aqueous solution during catalyst synthesis and reactions because of extremely low Raman scattering from water [24–27]. Second, Raman spectroscopy can be employed to directly provide molecular level information of catalytic active sites at a very wide range of temperature (sub-ambient up to 1000  C) [28, 29] and pressure (UHV to 100 atm) [30, 31], without any special interference from ambient air. The interference of gas phase often works as limitation of techniques, but its weak scattering of the gas phase in Raman spectroscopy makes it suitable to carry the experiment under gas flowing condition [32–35]. Third, significant technical development and enhancement allow it to operate at wide scan region (from 50 to 4000 cm
1), time resolution measurement with femtosecond time and operando condition with quartz fiber access. The combination of modern Raman spectroscopy instrumentation (e.g., Fourier transform (FT) Raman [36, 37], surface-enhanced Raman (SERS) [9, 10, 38, 39], confocal Raman, ultraviolet (UV) Raman [15], tip-enhanced Raman scatterings (TERS) [38, 40, 41]), and the operando Raman spectroscopy methodology [17, 42–46] (e.g., Raman-MS and Raman-GC) can measure the state of catalysts (chemical composition, hydration degree, and redox state) and chemical structure of the reaction intermediates/products, which allows the establishment of direct structure-activity/selectivity relationships that will have a significant impact on catalysis science. In addition, Raman spectroscopy with different laser sources make it possible to detect phases in different depths and different area of interest [42, 47–49]. Laser wavelengths ranging from ultraviolet through visible to near infrared can

be used for Raman spectroscopy. In general, most samples absorb strongly in the UV while less in the visible to near IR; the longer the excitation wavelength the deeper light penetration into the sample. The shorter laser wavelength gives information closer to the surface. For a Raman spectroscopy instrument, the laser spot size (spatial resolution) determines the size of analyzed point, which also strongly depends on choice of laser wavelength. Therefore, with the appropriate laser wavelength, Raman spectroscopy should be able to deliver dynamic information on the different areas of catalyst in terms of their structure performance, providing a more complete picture of the overall catalytic process.

4.2.3

Limitations of Raman Spectroscopy for Characterization of Catalysts

Although Raman spectroscopy shows tremendous advantages for catalyst characterization, there are some shortcomings that limit its usability. First, laser irradiation may cause the degradation of catalyst, which makes it hard to apply to catalytic materials that have photochemical properties and heat sensitivity [50–52]. For instance, once the sensitive sample is exposed to the laser, the heating effect accomplished by conduction could lead to the dehydration of catalyst, the phase transition, reduction, or even complete decomposition. It is reported by Xie and co-workers [53] that the local temperature at laser focus importantly depends on the color of the sample together with the power and wavelength of the excitation source (UV  vis > IR). Due to the advanced technology, there are several ways to reduce the laser heating effect: using lower laser power (400  C). SiO2supported (Fig. 4.6a) and Al2O3-supported (Fig. 4.6b) molybdenum oxide presented different phases under the ambient and controlled environment with different loadings of Mo [150]. For the SiO2- and Al2O3-supported MoO3 catalysts, the typical band at ~820 cm
1 due to crystalline MoO3 nanoparticle is essentially absent, and the supported molybdena phase is 100% dispersed as a two-dimensional surface MoOx overlayer on these supports. Under ambient conditions, the MoOx phase is hydrated by moisture, and the surface molybdenum oxide species have been shown to be present in the same molecular structures as found in aqueous molybdena solution as [Mo8O26]4
, [Mo7O24]6
, and [MoO4]2
anions. For MoO3/SiO2 catalysts (Fig. 4.6a), [Mo7O24]6
cluster dominated the hydrated surface, which gives rise to Raman bands at 950 and 880 cm
1 for 1% MoO3 loading and 950, 880, 695, 381, and 232 for 5% MoO3 loading. For MoO3/Al2O3 catalysts, both hydrated [MoO4]2
and [Mo7O24]6
clusters are present in the Al2O3 surface

a)

under ambient conditions, and their relative concentration depends on the surface molybdena coverage. At low coverage (1% MoO3/Al2O3), the isolated [MoO4]2
, exhibiting band at 912, 846, and 320 cm
1, is the main surface species, as well as minor amount of [Mo7O24]6
anions. At high coverage, very different Raman features are present as shown in Fig. 4.6b. The Raman bands for the hydrated 20% MoO3/Al2O3 are present at 950, 846, 501, 367, and 210 cm
1, which most closely match the vibrations of the aqueous [Mo7O24]6
clusters, and may reflect the presence of a minor amount of [MoO4]2
anion. Upon dehydration, the adsorbed moisture evaporates, and the molybdena anions decompose, and the molybdena are anchored to the oxide support by titrating the support surface hydroxyls. For the dehydrated SiO2-supported MoO3, isolated surface MoOx species are detected, of which Raman spectra exhibit a sharp band at 980 cm
1 assigned to the isolated dioxo surface (O¼)2MoO2 species, and the additional weaker band at 357 cm
1 for the higher loaded sample corresponds to the associated surface MoO4 bending mode. The absence of the bridging Mo-O-Mo deformation at 220 cm
1 indicates the presence in isolated surface MoOx species on SiO2 support under dehydrated condition. Unlike the dehydrated supported MoO3/SiO2, the Al2O3-supported MoOx contains isolated surface MoOx species at low loading, and it evolves into polymeric species at higher loadings. Afterward, Raman investigations were conducted over various supported metal oxides including Re2O7 [101], V2O5 [151], MoO3 [151, 152], WO3 [151, 152], CrO3 [153], and Nb2O5 [154] under

b) 980

1006 20% MoO3/Al2O3, dehydrated 210

940 870

950 880 950 880

810

490

600

1% MoO3/SiO2, dehydrated

695 605

232 5% MoO3/SiO2, ambient

976

605

976

601

381

1% MoO3/SiO2, ambient

490 490

1200

1000

810 800

600

Raman shift (cm–1)

377 Intensity (a.u.)

Intensity (a.u.)

597

5% MoO3/SiO2, dehydrated

357

490

590

990 838 950 846 912 846

452 345

1% MoO3/Al2O3, dehydrated

20% MoO3/Al2O3, ambient 210 367 1% MoO3/Al2O3, 561 320 ambient

SiO2 400

200

Al2O3 1200

1000

800 600 400 Raman shift (cm–1)

200

Fig. 4.6 Raman spectra of supported (a) MoO3/SiO2 catalyst and (b) MoO3/Al2O3 catalyst under ambient and dehydrated conditions as a function of surface molybdena coverage. (Reprinted from Tian et al. [150], copyright (2010) with permission from American Chemical Society)

4

88

J. Moon et al.

different environments, which concluded that the dehydrated surface metal oxides have unique molecular structure compared with their hydrated forms. These studies demonstrated that Raman spectroscopy is a useful tool to determine surface molecular structures of metal oxide catalysts as a function of environmental conditions. Utilization of different laser excitations has been shown as a powerful approach to characterize supported oxide catalysts on the structure of not only the surface dispersed oxide species but also sometimes the support oxides. It is known that the oxide support plays an essential role in catalysis by supported oxides. In this regard, the support shape is an important factor that can affect various reactions. To understand the changes of the surface species related to the shape of the support, the synthesized vanadium (VOx) supported by ceria (CeO2) nanocrystals with defined surface planes was studied by Wu et al. using in situ UV and Vis Raman spectroscopy [155]. In the study, different shapes of ceria (e.g., rod, cube, and octahedra) were applied, and the structure of loaded vanadia has been studied as a function of vanadia loading and calcination temperature. Depending on the wavelength of the laser source, either the nature of surface VOx species can be examined by using 632.8 nm excitation (Fig. 4.7a) or the defect sites in ceria can be studied by using 325 nm laser (Fig. 4.7b). For visible Raman, the spectra typically have two regions including V¼O mode between 960 and 1050 cm
1 and bridging modes (V-O-V and V-O-Ce)

b)

10V

930

F2g 1040

2V

lex = 325 nm

V2O5 995

0.1V 1008

Intensity (a.u.)

5V

1.6 1.2

D 0.8

0 2 4 6 8 10 Surface VOx density (V/nm2)

VOx 10 V 5V 2V

1003

0.5V

2.0

VOx/r-CeO2

Intensity (a.u.)

864

lex = 632.8 nm

ID/IF2g

CeVO4

1020 1028

a)

around at 960 cm
1. For low vanadium loading, it showed characteristic peaks of monovanadate with increasing V¼O band intensity. As the loading increases to 2 V nm
2, the peak splits into two bands at 1021 and 1031 cm
1, which indicates the formation of dimeric and trimeric VOx species. Eventually, crystalline V2O5 and CeVO4 are formed at higher loading. This phenomenon occurred in both cube and rod shape of ceria. For UV Raman laser (Fig. 4.7b), the 2LO mode at 1179 cm
1 and defect-associated D band at 592 cm
1 are greatly enhanced in intensity compared to the F2g mode of ceria as in the visible Raman spectra. The ratio of the D band to F2g band (ID/IF2g) represents the relative amount of the defect site in ceria. This ratio decreased with increasing VOx loadings, indicating that the surface VOx interacts closely with the defect sites in ceria. Such trend was not obvious for VOx/cube samples, i.e., UV Raman is less sensitivity in detecting the defect changes as a function of surface VOx density on ceria cube, which is due to the much large particle size of the cube than the rod. Additionally, the solid-state reaction between VOx and ceria to form CeVO4 was also studied over the three ceria shapes upon calcination to high temperatures (900  C) at similar surface vanadia density. The result again suggests that the both surface structure and the amount of defect sites on the different ceria shapes play major roles in the easiness of the formation of CeVO4.

2LO r-CeO2

700

800

900 Raman shift

1000 (cm–1)

1100

400

600

800

1000

Raman shift

1200

0.5 V 0.1 V r-CeO2 1400

1600

(cm–1)

Fig. 4.7 Visible (a) and UV (b) Raman spectra of dehydrated VOx/r-CeO2 sample as a function of surface VOx density. (Reprinted from Wu et al. [155], copyright (2012) with permission from Wiley-VCH Verlag GmBH & Co. KGaA, Weinheim)

4

Raman Spectroscopy

4.6.2

89

Reaction Conditions: In Situ/Operando Measurement

The active site or the structure of catalysts may go through significant changes under reaction conditions, which necessitates characterization approaches ideally under reaction conditions. Taking its advantages, Raman spectroscopy has been widely employed to investigate metal oxide catalysts under various reaction conditions such as temperature-programmed reactions; reaction-induced dispersion studies for supported metal oxide [20]; oxidative coupling of methane and for NOx storage [29, 156]; the temperature effect on surface peroxide, dioxide, and superoxide species on ceria [157, 158]; oxidation of ethanol by ozone on MgO catalyst [34]; etc. Raman spectroscopy is often applied on the structural determination and the number of terminal M¼O bonds in surface metal oxide species that are present in supported metal oxide catalysts. Taking supported chromia catalysts as example, combination of in situ Raman and IR discovered that only mono-oxo surface chromate species is present in supported chromia catalysts. Later on, in situ Raman studies on butane dehydrogenation over the supported chromia catalyst revealed that isolated and polymerized surface monooxo chromate species coexisted on the supported and the

a)

polymerized surface species, which was preferentially reduced under reducing condition [159]. In this work, by using Raman spectroscopy, the molecular structures of surface metal oxide were directly determined. Not only the nature of surface reaction intermediate was identified, but also the reducibility of the polymerized surface species was discovered by in situ Raman investigations. Although in situ Raman measurement gives various information under reaction condition, it is limited to provide useful information under genuine reaction condition due to the shortcoming of in situ cell. Operando Raman spectroscopy offers an opportunity to understand the dynamic states of catalytic sites and the reaction intermediates during catalytic reaction, with both structure and catalytic activity being simultaneously investigated at molecular-level stage. The first time use of an “operando” term for Raman was from Bañares et al. in 2002 (Fig. 4.8) [160]. This work investigated the structure transformation of the supported V-Sb-O catalyst during propane ammoxidation by in situ Raman and online GC measurements so that both the structure and activity/selectivity information were simultaneously obtained, of which the methodology is donated as operando Raman spectroscopy. The conversion and selectivity of the reaction did not show any significant differences in comparison to the reaction recorded with convention

SbVO4

b)

Sb2O4 VO-R

V-OR 29.2 4.6

d

480 °C

22.0

V=O

18.1 10.9

c 200 °C

b

16.3 2.2 480 °C

4.2

Acrylonitrile

15.0

Acetonitrile

2.6

Propylene

a

CO2 200 °C

CO

Sb2O3 30

20

10

0

1100

Yield (%)

Fig. 4.8 Operando Raman spectra (a) obtained during propane ammoxidation and yield of the different products during Raman acquisition (b) on 2Sb5V/Al catalysts (a, b) and 1SbV/Al catalyst. (a) and (c) correspond to fresh dehydrated samples, and (b) and (d) show the

900

700 Raman shift

500

300

100

(cm–1)

spectrum from the propane ammoxidation at 480  C. (Reprinted from Bañares et al. [160], copyright (2002) with permission from The Royal Society of Chemistry)

4

90

J. Moon et al.

e)

1580 1320

c)

1320

1580

a)

accuracy of catalytic data collected over operando cells [161]. Thus, it is crucial to compare the reaction rate between operando Raman cells and laboratory catalytic reactors. The activity of the catalyst based on vanadia and molybdena support on alumina using two different reactors (traditional fixedbed reactor and operando reaction cell) has been analyzed by Bañares and Khatib [162]. For both reactors, the activity was measured with an online GC equipped with a thermal conductivity detector. The activity agrees well between Raman cell and the traditional reactor within the experimental error. Therefore, more reliable information under genuine reaction conditions with their operando Raman cell. More recently, operando Raman spectroscopy has been used to monitor the structural evolution of iron-based catalyst in Fischer-Tropsch synthesis (FTS) at a pressure of 3.0 MPa and a temperature of 260  C, which is a typical industrial reaction at harsh conditions. Han’s group [46] has successfully determined the active phase of iron-based catalyst for FTS using operando Raman spectroscopy (Fig. 4.9). The

1580 1320

fixed-bed microreactor. The fresh dehydrated 2Sb5/V showed Raman bands at 716, 451, 372, 255, and 190 cm
1, characteristic of Sb2O3, and a broad band of SbVO4 phase appeared at 800 cm
1. Upon catalytic reaction with increased temperature at 480  C, Sb2O3 bands are replaced by bands at 459, 399, 261, and 190 cm
1, corresponding to α-Sb2O4 phase. Meanwhile, the broad band at 800 cm
1 becomes more intense. The low Sb-loaded 1Sb1V/Al showed Raman bands at 1024 and 900 cm
1 due to polymeric mono-oxo vanadium oxide species. However, the Raman spectrum significantly changed during the reaction and showed new Raman bands at 1060 and 670 cm
1 corresponding to surface alkoxy (V-O-R) species. A significant benefit of operando Raman spectroscopy is that it provides direct evidence on the oxidation states of V and Sb oxide species in the Sb-V-O phases under reaction condition. Although operando Raman spectroscopy has been successfully adapted to obtain both catalyst structural and reaction kinetic information, it is worthwhile to mention that heat and mass transport limitation could dramatically affect the

Intensity (a.u.)

420 min 390 min 360 min 330 min 300 min 270 min 240 min 210 min 180 min 150 min 120 min 90 min 60 min 30 min 0 min

Conversion (%)

1.2 1.0

100

70

70

60 0.8

50

0.6

40 30

0.4

20 0.2 0.0

10 0

0 60 120 180 240 300 360 420 TOS (min)

1.0

2500 2000 1500 1000 500 Raman shift (cm–1)

80 60 50

0.8

40 30

0.6

20 10

0.4 0

0 60 120 180 240 300 360 420 TOS (min)

Fig. 4.9 Operando Raman spectra of a-Fe2O3 during the reaction processes under FTS condition (3.0 MPa, 260  C) recorded in the time interval of 30 min for 7 h. (a, b) 10% H2/Ar, (c, d) 10% CO/Ar, (e, f) 5% CO/5%H2/Ar. (a, c, e) present the structural evolution of catalyst, and

f)

Conversion C2+C4= selectivity

1.4

90 C2+C4= selectivity (%)

Conversion C2+C4= selectivity

1.4

Conversion (%)

b)

2500 2000 1500 1000 500 Raman shift (cm–1) 100 d) 1.4 Conversion 90 C2+C4= selectivity 1.2 80 C2+C4= selectivity (%)

2500 2000 1500 1000 500 Raman shift (cm–1)

100 90

1.2

80

1.0

70 60

0.8

50

0.6

40 30

0.4

20 0.2 0.0

10 0

0 60 120 180 240 300 360 420 TOS (min)

(b, d, f) represent the corresponding catalytic performance. Solid square is CO conversion, and the empty triangle represents selectivity of C2-C4 olefins. (Reprinted from Fu et al. [46], copyright (2015) with permission from Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim)

Raman Spectroscopy

91

structure change of α-Fe2O3 catalyst was monitored under extensive activation and reaction processes. α-Fe2O3 could only be partially transformed into γ-Fe2O3 (1580 cm
1) upon hydrogen treatment, and the Raman spectra (Fig. 4.9a) indicated that the structure of working catalyst remained unchanged during the FTS. Upon CO and syngas (H2/CO) activation, the formation of γ-Fe2O3 (1580 cm
1) was observed at lower treatment temperature (below 300  C) and subsequently transformed to Fe3O4 (300, 540, and 660 cm
1) at 320  C. The mixture phase mainly consisting of Fe3O4 with a certain amount of Fe5C2 was detected with prolonged treatment of 320  C in both CO and syngas (H2/ CO) treatment. Carbonaceous species appeared with Raman bands at 1320 and 1570 cm
1, which was assumed as reaction intermediates and was found to be responsible for relative higher catalytic activity over CO- and syngas (H2/CO)activated catalysts. Fe3O4 was found to be the main species when the surface carbonaceous of CO- and H2/CO-activated catalysts was consumed and thus transformed to lower olefins, which is absent from H2-activated catalyst. Under the same reaction conditions, CO-activated catalyst exhibited the highest catalytic activity, followed by H2/CO- and H2-activated catalysts. This work opens a new strategy to investigate the structure-selectivity relationship of industrial catalysts under harsh conditions by operando Raman spectroscopy.

Time-resolved in situ Raman spectroscopic studies on working catalysts have hardly been reported. The main obstacle to the time-resolved capabilities is low sensitivity of Raman on working catalytic process. In other words, high sensitivity requires relatively long data accumulation time and, thus, making it hard to obtain high-resolution data within a short period of time under reaction condition. The initial timeresolved Raman study was conducted over vanadium phosphorous oxide catalysts to study active species of selective oxidation of n-butane [163]. However, the time resolution of this study is in the range of hours due to instrument limitations. After the development of highly sensitive chargecoupled device (CCD) detectors and the advancement of notch filters were achieved, the accumulation time in the range of minutes or even less can be accomplished. Kuba and Knӧzinger [81] obtained spectra of time-dependent in situ Raman experiments on the sulfated-zirconia (SZ) and tungstated-zirconia (WZ) catalyst during isomerization reaction of n-pentane (Fig. 4.10). A fixed-bed microflow reactor was employed as an in situ Raman cell combining with onstream gas chromatography to simultaneously monitor the activity of the catalysts at a time resolution of a few minutes. For fresh sulfated ZrO2 in a helium flow (0 min, SZ), two types of surface sulfated species were present: strong Zr-O-S stretching bands at 1015 and 1045 cm
1 and

c) 1590

WZ

WZ *

1190

1022

SZ 840

1170 1143

1409 1390

50,000

Time-Resolved Raman Spectroscopy

b) 1045 1015

a)

4.7

1590

4

0 min

156 min

8 min

156 min

40,000

32 min 40 min

124 min *10

1590

24 min

Intensity (a.u.)

Intensity (a.u.)

16 min

92 min 60 min

0 min

48 min

28 min 4 min

56 min

30,000

64 min

4 min

72 min

28 min 60 min 92 124 156 min

80 min

1600

1400

1200

1000

800

1600

Wavenumbers (cm–1)

Fig. 4.10 Time-dependent in situ Raman spectra (0–80 min TOS) of (a) SZ and (b), (c) WZ catalyst during isomerization reaction of n-pentane (b) show the spectrum before the intensity correction, and (c) shows the spectra after intensity correction for darkening. The asterisks

1400

1200

0 min 1000 1600 Wavenumbers (cm–1)

1400

1200

* 1000

on the peaks in (b) correspond to the laser plasma line at 1057 cm
1. (Reprinted from Kuba and Knӧzinger [81], copyright (2002) with permission from John Wiley & Sons, Ltd.)

4

92

S¼O stretching bands at 1390 and 1490 cm
1. Upon the catalytic n-pentane conversion, one of these species (1045 and 1390 cm
1) remained unchanged, whereas another one (1015 and 1490 cm
1) disappeared with increasing reaction time, which is possibly due to the reduction to H2S without any coke formation. For the working tungstated-zirconia, the Raman spectrum of fresh catalyst in He displayed two bands at 910 cm
1 (W-O-Zr vibration) and 1022 cm
1 (W¼O stretching vibration). With the catalyst on stream, the bands of tungsten oxide species vanished and pre-graphitic particles (coke) was formed, indicating by the band at 1590 cm
1 and a broad shoulder between 1300 and 1550 cm
1. This coke formation caused the continuously darkening of the sample, which strongly affected the intensity of the Raman scattering and limits the time-resolved availability. However, by using the intensity of laser plasma lines, the change of the diffuse reflectance of the catalyst can be determined, and thus, the intensity correction of the spectra allowed to observe that most of the coke was formed in the first few minutes of the reaction followed by a constant formation rate with time of stream. Very recently, as an effort to improve time resolution, operando time-resolved Raman spectroscopy was applied to study the structural dynamics of the supported vanadia catalyst under oxidative dehydrogenation (ODH) of ethanol reaction. Waleska et al. [42] studied the dynamics of supported vanadia species under reaction conditions using multiwavelength and time-resolved Raman spectroscopy combined with in situ FTIR. Structural dynamics of catalysts upon switching the reaction cycles from oxidative to the reactive condition were studied using different excitation wavelengths for Raman spectroscopy. Three targeted excitation wavelengths have been selected: 256.7 nm for full resonance enhancement, 385.1 nm for pre-resonance conditions, and 514.5 nm for negligible resonance enhancement. The time-dependent correlation over several reaction cycles allows the identification of active vanadia surface structures through temporal behaviors of related VOx species (Fig. 4.11). Prior to operando Raman experiments, Raman spectra of the dehydrated catalysts were recorded, and it showed six fundamental vibrations of the dispersed VOx structure denoted as νa
vf at 461, 548, 695, 914, 1031, and 1064 cm
1. The bands at 461, 548, and 695 cm
1 are associated with oligomerized surface structure of vanadia, and the two bands at 914 and 1064 cm
1 are assigned to the in-phase and out-of-phase stretching vibration of the V-OSi linkage, respectively. The main signal at 1031 cm
1 represents the vanadyl (V¼O) stretching vibration. This six bands were observed carefully under reaction conditions, and temporal behavior of each related peaks was obtained with the reaction time (Fig. 4.11b). The bands va-c and ve showed significant intensity changes, and the temporal

J. Moon et al.

behavior of this species is compared to that of the acetaldehyde concentration in the gas phase. During the ethanol reaction, all three peaks go through a redshift, and the intensities and full widths at half-maximum of all three peaks decrease. The intensity of these three peak decreases linearly, while the concentration of acetaldehyde increases. With the combination of IR and the temporal behavior, it showed that the first contact between the catalyst and the ethanolcontaining gas phase leads to an irreversible change in the VOx structure, and thus, the vanadyl vibrations of small oligomeric structures appear at lower frequencies in the Raman spectrum. Through periodically and reversibly occurring spectral changes in the spectra, this time-resolved study gives insights into the nature of each species and a better understanding of the reaction mechanism.

4.8

Spatial-Resolved Raman Spectroscopy: Microscopy

Catalysts might exhibit a gradient of their properties under working conditions due to the changes of reactant concentrations, temperature, and structural and oxidation states on different length scale of the particles [164]. However, conventional Raman spectroscopy is limited on the structure information from the focused area of surface/bulk. Raman mapping/imaging can be a solution to observe the gradient of various effects. Raman spectroscopy and microscopy have been successfully merged into a one unique technique and provide further understandings of reaction pathways over various spots/areas of the catalyst surface as well as catalyst synthesis. Moulijin et al. showed that Raman imaging is an effective tool to measure carbon deposits along the cross section of the surface of the spent hydroprocessing catalyst pellet through the line scan procedure with Raman microscopy [165]. Results from this study suggest that activity is significantly reduced by exposing to heavy molecules forming coke. Spatially resolved Raman microscopy has been also applied in the catalyst synthesis procedure to obtain insights to determine the active phase over the support pallets during the catalyst preparation. Bergwerff et al. [79] studied the formation of Mo(Hcitrate)2O11 using various Mo precursor over Al2O3 pallets using Raman microscopy. Figure 4.12 shows one of the examples in this study, which represents the distribution of the support surface impregnated with a ratio of 1:1 with Mo-citrate. The distribution illustrated in a three-dimensional plot at 15, 60, and 180 min with the explanation of the pellet location. The cross-sectional scans were used as a faster way to monitor the distribution of Mo complex after impregnation and gave validated concentration variations over the support bodies. The results confirmed that a high concentration of Mo in combination with long contact

4

Raman Spectroscopy

a)

93

256.7 nm Prior to reaction During reaction

nr: 1056 nd: 922

ne: 662

(a)

ng: 1095

Raman intensity (absorption-corrected)

3080 2941 385.1 nm Prior to reaction During reaction

(b)

4

809

1460

514.5 nm Prior to reaction During reaction

(c)

3743 600

400

800

1000

1400

1200

3000

3500

Raman shift (cm–1)

Raman intensity (absorption-corrected)

Raman shift (cm–1)

CAA (ppm)

b) 200

200

200

100

100

100

0

0

0

535

700

1030

680

1025

530 525

660

150

40 30 20 10

50 40 30 20 10

120 100 80 60 40

50

100 50

150 FWHM (cm–1)

1020

520

125 100 0

50 100 150 Time (min)

200

40 30 0

Raman peak na,b

Fig. 4.11 (a) Comparison of the Raman spectra before (black) and during the reaction (red) using multiwavelength Raman spectroscopy. (a) 256.7 nm shows the full resonance enhancement, (b) 385.1 shows the pre-resonance conditions, and (c) 514.5 nm showed the negligible resonance enhancement. (b) Temporal behavior of the three Raman

50 100 150 Time (min) Raman peak nc

200

0

50 100 150 Time (min)

200

Raman peak ne (V=O)

peaks b (black), νc (red), and νe (blue) of dispersed VOx structures compared to the acetaldehyde concentration (green, top panels) under oxidative and reactive conditions. Marked position (*) are artifacts resulting from cosmic radiation. (Reprinted from Waleska et al. [42], copyright (2018) with permission from American Chemical Society)

94

J. Moon et al.

Pellet t = 15 min

50

50

0

0 0

–50

0 –50

1

0

1

2 X-axis (mm)

3

0

2 Y-axis (mm)

3

1 1

2 X-axis (mm)

2 Y-axis (mm)

3

3

25 20 15

t = 60 min

I (n MoO2t) I (NO3–)

10

t = 180 min

5 0

50

50 0

0 1

–50 0

1

2 2 X-axis (mm)

3

3

Y-axis (mm)

0 0 –50

1 0

1

2 2 X-axis (mm)

3

3

Y-axis (mm)

Fig. 4.12 Three-dimensional plots illustrating the spatial distribution of the Mo(Hcitrate)2O114
complex in 3 mm Al2O3 pellets 15, 60, 180 min after impregnation with a Mo-citrate solution. The intensity

plots are obtained by referencing to the NO3
Raman band. (Reprinted from Bergwerff et al. [79] copyright (2004) with permission from the American Chemical Society)

times can result in a homogeneous distribution of the complex. The spatial resolution of Raman spectroscopy can be significantly improved when a metal-coated tip is implemented. Tip-enhanced Raman spectroscopy (TERS) improves the diffraction-limited spatial resolution of Raman spectroscopy to the nanoscale, which allow mapping even single molecules [41, 166]. A TERS study of photocatalytic reaction with nanoscale spatial resolution was demonstrated for the first time by Kumar et al. [129] (Fig. 4.13). An Ag-coated TERS tip in an AFM-TERS configuration was adapted in this experiment to observe the nanoscale Raman mapping of the plasmon-assisted photocatalytic reaction of p-mercaptoaniline (pMA) occurring at an isolated catalytic point. It is noted that the reaction from pMA to p,p′dimercaptoazobenzene (DMAB) occurs only at a few locations on the Ag substrate because Ag substrate contains different size of Ag nanoparticles. TER experiments were conducted over the area marked with dotted rectangle in Fig. 4.13a, and the high intensity of the 1142 cm
1 Raman peak of DMBA appeared. TERS spectra from positions 1, 2, and 3 showed intense DMAB signature peaks, while these peaks are very weak at positions 4, 5, and 6. This showed the heterogeneous nature of the surface and the spatial resolution

within 20 nm, which is a reasonable range to differentiate the active and inactive particles and ultimately spectroscopic mapping of individual catalytic sites. Raman microscopy can be greatly improved by combination with other imaging spectroscopy including scanning near-field optical microscopy (SNOM) that can provide further insight within molecular spectroscopy, atomic spectroscopy, or topographical information. The detailed information will be explained in the following Raman microscopy chapter.

4.9

Modulation Excitation Raman Spectroscopy

Modulation excitation spectroscopy (MES) was introduced in early 2001 by Baurecht and Fringeli as a novel methodology [37] which provides an additional sensitive and selective method to the toolbox of in situ/operando techniques. MES uses a periodic perturbation of a system by external parameter called stimulation. A stimulation is chosen to influence the concentration of the target species. Common stimulation parameters are concentration, pH, temperature, light flux, electric field, etc. After periodic perturbation, the

4

Raman Spectroscopy

95

a)

40

b)

40 6

5 1

3 2

4 4 50 nm

200 nm nm

c)

0

a.u.

d)

0

30

Raman intensity (a.u.)

25

2 3

1000

1400 1200 Raman shift (cm–1)

Intensity (a.u.)

1

15

4

10

5 6

5

1600

20 nm

20

0

20

40

60

80

Distance along the line (nm)

Fig. 4.13 (a) AFM topography image of an Ag substrate with Ag nanoparticle on a glass substrate. (b) TER map from the dashed rectangle mark in (a), (c) near-field spectra from the positions mark in (b), and (d) intensity profile along the dotted line marked in (b) showing the

spatial resolution of the TERS map. (Reprinted from Kumar et al. [129], copyright (2015) with permission from The Royal Society of Chemistry)

concentration of targeted species reaches a quasi-steady state, largely at the same frequency as that of the stimulation. The amplitude and the phase delay of the response are stimulation-frequency dependent and contain the information on the kinetics of the active species [167]. As a result, this makes it possible to selectively pick up on these small changes and in which sequence they occur. Initially, attenuated total reflection infrared spectroscopy (ATR-IR) was the choice for MES spectroscopy. However, recently, any time-resolved spectroscopic techniques can be combined with MES (e.g., polarization-modulation infrared reflection-absorption spectroscopy (PM-IRRAS), diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), X-ray absorption spectroscopy (XAS), X-ray diffraction (XRD), and Raman spectroscopy) [168] . In the combined MES-Raman spectroscopic study, the structure of a spin-crossover material was investigated using

temperature as a stimulus [169]. However, it is technically challenging to perform the temperature modulation since it requires equal heating and cooling rate in order to obtain a symmetric stimulation. Nuguid et al. reported the modulated excitation Raman spectroscopy of V2O5/TiO2 under selective catalytic reduction condition (SCR) (Fig. 4.14) [170]. In this study, reactants were varied periodically, while spectra were obtained with sufficient time resolution simultaneously. The collected spectra in the time domain are converted into the corresponding spectra in the phase domain, and this contains the signals only from the species that responded to the modulation sequence. As a result, signals from unresponsive or static species are effectively eliminated, and signals from the active sites and perturbed species are emphasized. For example, time-resolved Raman spectra (Fig. 4.14a) and the corresponding phase-resolved spectra (Fig. 4.14c) of the 2 wt. % V2O5/TiO2 catalyst were compared. This spectrum

96

J. Moon et al.

a)

b)

1025 5 a.u.

Intensity (a.u.)

100 a.u.

c)

d) 5 a.u.

1031 1022

Intensity (a.u.)

1 a.u.

1100 1000

900

800

700

600

500

400

300

1050

Raman shift (cm–1)

Fig. 4.14 (a, b) Time-resolved Raman spectra and (c, d) corresponding phase-resolved spectra of 2 wt% V2O5/TiO2 at 250  C during 30 s pulses of 500 ppm NH3 in a gas feed of 500 ppm NO, 2 vol% H2O, and 5 vol %

showed three main peaks dominated from TiO2 anatase, which include intense peaks at 395 and 636 cm
1 assigned to B1g and Eg modes and a peak at 516 cm
1 assigned to a convolution of the two Raman active mode B1g and A2g. Additionally, the weak peak feature of the VOx showed in the region of 1010–1040 cm
1, due to the stretching mode of V¼O. In the time-resolved spectra, there was no clear difference with time, but it showed clear changes after phasesensitive detection (PSD) that the signals of both VOx and TiO2 responded to the changes in the gas-phase composition. This result can be concluded that both the VOx phase and the TiO2 support are somewhat involved in the SCR process and that only a defined portion of VOx species are active in the SCR reaction by analyzing the peak at 1031 cm
1. ME Raman spectroscopy is the technique that can easily be adapted to other catalytic studies, and it opens up new possibilities to understanding structure-reactivity relationship of catalyst systems under operating conditions.

1025

1000

975

Raman shift (cm–1)

O2 balanced in Ar. (Reprinted from Nuguid et al. [170], copyright (2019) with permission from the American Chemical Society)

4.10

Applications of Raman Spectroscopy to Catalyst Synthesis

Raman spectroscopy has been applied for the investigation of catalyst preparation process to reveal information about molecular structures and the formation of crystalline phases under synthesis conditions. As an early example, Hill and Wilson reported the detailed Raman studies over iron molybdate at each synthesis conditions related to coprecipitation process [24] (Fig. 4.15). In this study, the authors studied not only the effect of pH but also in various conditions including the temperature, precursor concentration, mixing, aging, filtration, washing, drying time/temperature, and calcination time/temperature. Ferric molybdate includes several unique Raman active bands: 780 (Mo-O-Mo vibration of Fe2(MoO4)3 phase), 820 and 990 (molybdenum octahedrally coordinated with oxygen), 860 (Mo¼O stretching mode),

4

Raman Spectroscopy

97

c)

f)

Finished Fe-MoO catalyst calcined 1 h at 500 °C

Fe-Mo-O catalyst

b)

e) Dry catalyst precursor

a)

1200

Wet catalyst precursor

800

400

0

Raman shift (cm–1)

Precursor after 12 days at 105 °C

d)

1200

Precursor derided at room temperature

800

400

0

Raman shift (cm–1)

Fig. 4.15 Raman spectra of catalyst from different procedures of the synthesis. (a) Wet catalyst precursor, (b) dry catalyst precursor, (c) FeMo-O catalyst, (d) precursor dried at room temperature (24  C), (e)

precursor after 12 days at 105  C, and (f) finished Fe-Mo-O catalyst, calcined 1 h at 500  C. (Reprinted from Hill and Wilson [24], copyright (1990) with permission from Elsevier)

and 530 cm
1 (Mo-O-Mo bending mode). Figure 4.15a–c showed Raman spectra of the wet precipitate, the precipitate after being dried at room temperature, and the final Fe-Mo-O catalyst after calcination at 500  C for 1 h to reveal the effect of the precursor compared with the final catalyst. Figure 4.15d–f showed the effect of drying time and temperature by comparing the drying temperature between room temperature and over 100  C. It exhibited that the catalyst precursor begins to transform into an Fe/Mo oxide at temperatures in excess of 100  C which possesses similar features to the final Fe-Mo-O catalysts. Figure 4.15 just presents two variables studied in this work where several variables have been studied to give insights to understand the phase changes during the synthesis process.

programmed Raman spectroscopy can provide useful structural information of catalysts at different temperatures under reductive or oxidative condition. For example, temperature-programmed reduction (TPR) and oxidation (TPO) profiles of dispersed VOx on silica were obtained by Bañares et al. [171]. In situ temperatureprogrammed Raman spectra provide information of the changes of vanadia species during TPO and TPR treatments (Fig. 4.16). The fresh catalyst includes several broad bands (495, 606, 802, and 907 cm
1) representing silica support and a sharp peak at 1037 cm
1 assigned to dehydrated surface isolated VOx species. At 473 K during TPO, new Raman bands at 146, 284, 702, and 994 cm
1 appeared, which represent crystalline V2O5. It is noted that during the reduction process, this vanadium oxide bands of crystalline vanadia became more evident as the temperature increased and then disappeared at 923 K, but the crystalline vanadia restored through TPO treatment. The TPR- and TPO-Raman profiles revealed the interaction among surface vanadia species through the sharing of oxygen sites. Additionally, the effect of loadings of vanadia on the silica surface was examined in this study, and it was shown that the interaction of vanadia and silica is directly related to the surface vanadia

4.11

Applications of Raman Spectroscopy Study to Catalyst Treatments

Catalytic reactions often occur at high temperatures with different gas environments, and thus, it is important to understand the most efficient reduction condition or oxidation condition for each catalyst. Consequently, temperature-

4

98

J. Moon et al.

a)

b) 923 K

923 K

823 K

873 K

673 K

773 K

573 K

723 K

473 K

573 K Fresh catalyst

100

300

500

700

Raman shift

900

Reduced catalyst

1100

(cm–1)

100

300

500

700

Raman shift

900

1100

(cm–1)

Fig. 4.16 Obtained Raman spectra during TPR over (a) fresh sample) (b) during the TPO Raman of the reduced sample. (Reprinted from Bañares et al. [171], copyright (2000) with permission from J. C. Baltzer AG, Science Publishers)

density and does not alter the total reducibility as shown by the O/V ratios in TPR results. The removal of oxide ions causes the rearrangement and aggregation of VOx into nanocrystalline V2O5 over the silica. Understanding the reductivity of different surface oxide species is important to reveal their potential role in catalytic reactions. A combined UV and visible Raman study of the reduction of supported VOx provides a clear picture of the reducibility of different VOx [83]. The UV- and visibleexcited Raman studies of the V/θ-Al2O3 samples reduced in hydrogen show that polyvanadate and V2O5 are more easily reduced than monovanadate species. UV Raman is more capable in obtaining information on reduced vanadia species than visible Raman, mainly because of a decrease in selfabsorption and resonance enhancement in the UV region. Comparison of the UV Raman spectra from reduced V/θ-Al2O3 with a series of model bulk vanadium oxide compounds suggests that reduced VOx species can assume a V2O3-like form. The reduced VOx species can re-disperse on the support surface upon reoxidation in O2. Reduction of oxides generates various oxygen vacancies on the surface and in the bulk. Raman spectroscopy coupled with O2 adsorption is a powerful to probe the surface

vacancies. Defect sites of ceria have been studied using O2 adsorption with Raman spectroscopy by Wu et al. [158]. O2 adsorption was conducted at room temperature on both calcined and reduced ceria samples of three different shapes (Fig. 4.17). Generally, Raman bands from adsorbed O2 species are not shown on calcined samples. On reduced ceria nanorod surface, adsorbed O2 species were observed with Raman bands at 1139 (superoxide), 862 as a shoulder, and 830 cm
1 (both as peroxides), which suggests that oxygen vacancies generated from the H2 reduction play an important role to adsorbed O2 species on the surface. On nanocubes, O2 adsorption shows in an asymmetric band at 833 cm
1 with tailing at the high Raman shift. This indicates that one type of oxygen, peroxide, is dominant on reduced nanocube surface. Unlike with nanorods and nanocubes, very weak Raman features are observed from O2 adsorption on H2-treated nano-octahedra, due to the stable surface facets (111) present on this ceria shape. This experiment shows that the type of defect sites and consequently the nature of surface-adsorbed oxygen species depend strongly on the surface structure of ceria. Hydrocarbon reactions can often result in carbonaceous species (coke) deposits on catalyst surfaces. Insights into the

4

Raman Spectroscopy

99

a)

b)

lex = 532 nm

- 673 K reduced 2

* 673 K reduced

*

773 K reduced

**

833

*

860

* 18O

Intensity (a.u.)

1139

862

813 830

Intensity (a.u.)

782

lex = 532 nm

4

673 K reduced * 773 K reduced 873 K reduced

873 K reduced

800

900 1000 1100 Raman shift (cm–1)

1200

1300

800

900 1000 1100 Raman shift (cm–1)

1200

1300

lex = 532 nm

673 K reduced * * *

832 851

Intensity (a.u.)

c)

*

*

773 K reduced 873 K reduced

800

900 1000 1100 Raman shift (cm–1)

1200

1300

Fig. 4.17 Raman spectra of O2 adsorption at room temperature on different temperature-reduced ceria (a) nanorods, (b) nanocubes, and (c) nano-octahedra. Asterisks (*) represent Raman bands from trace

nitrate/phosphate species in ceria. (Reprinted from Wu et al. [158], copyright (2010) with permission from the American Chemical Society)

coke formation process during the reaction are important to understand the deactivation mechanism of catalysts, an area Raman spectroscopy has proved powerful [81, 125, 165, 172–175]. Wu and Stair studied the coke formation in butane dehydrogenation on alumina-supported vanadium catalysts through UV Raman spectroscopy [173]. Figure 4.18 shows Raman spectra from butane dehydrogenation on 1.2 V/Al2O3 catalyst at 673 K with controlled amounts of dosed butane. As the amounts of dosed butane increase, Raman bands at 851, 1005, 1186, 1432, 1500, and 1600 cm
1 grow intensely, which are corresponding to a single surface coke species,

while the bands at 915 and 1021 cm
1 due to surface VOx vanish gradually because the coke species covered VOx species on the surface. Additionally, the structural effect of the surface vanadate species has been studied, and it showed that the amount of coke formed in butane dehydrogenation follows the sequence: polymeric VOx > monomeric VOx > V2O5, Al2O3. This showed that the deactivation of the catalyst in butane dehydrogenation is mainly due to the formation of coke species over the surface, and both the nature and amount of formed coke are related to the structure of VOx species.

J. Moon et al.

1432 1500

1186

Intensity

851

1005

1600

100

600

900

1021

915

Increasing amount Fresh 1200

1500

1800

Raman shift (cm–1)

Fig. 4.18 Raman spectra of butane dehydrogenation on oxidized 1.2 V at 673 K with different dosage of butane. (Reprinted from Wu and Stair [173], copyright (2006) with permission from Elsevier Inc.)

4.12

Applications of Raman Spectroscopy to Catalyst Structure-Activity Relationships

With the development and design of new Raman cells, the difficulty of real-time experiments under reaction conditions has been somewhat resolved for Raman spectroscopy. Consequently, GC or QMS can be combined with Raman spectroscopy to analyze the catalytic activity, in addition to the structural transformation of the surface species at the same time, allowing possible correlation of structure-catalysis relationship. Taking bismuth molybdate catalyst as an example, there have been various studies related to the stability of different bismuth molybdate phases, and the phase diagram has been formed based on individual experiments [176–180]. To clarify the stability of each phase, in situ Raman studies of bismuth molybdate catalysts have been conducted in the presence of reactive gases at elevated temperature by Snyder and Hill [181]. Raman spectra of a bismuth molybdate catalyst were measured under various conditions including calcination in stagnant air with different temperatures, heated in air flow, and propylene oxidation reaction conditions (400  C, 500  C) with the early design of Raman cell in addition to online GC analysis. This Raman study clears out the debates about the stability issues of each phases and elucidates that α, β, and γ bismuth molybdate can be prepared in air between 400  C and 500  C, and the thermal stabilities of these species are not dependent on their purity but more on temperatures.

With the adoption of fixed-bed reactor, better insights of the structure-activity relationship can be obtained over different types of catalysts at genuine reaction condition. Mestl adapted in situ confocal Raman spectroscopy and in situ Raman with split-bed reactor which allowed to acquire more accurate structure-activity information of MoO3 and (MoVW)5O14 catalysts during propene oxidation [182]. Raman spectra and catalytic data of MoO3
x (Fig. 4.19a, b) showed the increasing intensity of 820 cm
1 band corresponding to M¼O stretching mode with decreasing selectivity to the partial oxidation product (acetic acid). In other words, the initial material that includes more oxygen defects showed higher selectivity for partial oxidation products. This indicates that the activity and selectivity of MoO3
x catalysts are a function of their degree of reduction, and oxygen defects are required for the higher catalytic activity. In comparison with MoO3
x, Raman spectra and catalytic data of (MoVW)5O14 catalysts were shown in Fig. 4.19c, d. The Raman spectra collected at 523 K showed two broad bands at 840 and 880 cm
1 due to the formation of nanocrystalline Mo5O14-type mixed oxide and a sharp peak at 995 cm
1 attributed to vanadium oxide. During propene oxidation at higher temperatures, the nanocrystalline Mo5O14 phase became dominant and was very stable. Meanwhile, the selectivity to partial oxidation products including acetic acid and acrylic acid went through a maximum of around 650 K. This indicates that Mo5O14 mixed oxide is considerably a better catalyst than MoO3
x for the formation of partial oxidation product even at high temperature of general propene oxidation condition. The in situ Raman results combined with the different catalytic properties of MoO3
x and the Mo5O14-type oxide led to the development of a structureactivity relationship which explains the behavior of industrial catalysts based on MoOx for partial oxidation. Catalyst deactivation through coke formation is one of the well-known drawbacks for many catalysts. The coking mechanism under working conditions is still not very well understood and the types and roles of carbon-containing species are also required to be clarified. Raman spectroscopy, especially in operando mode, is extremely powerful in studying coke formation chemistry and can study structure-activity relationship. Mutz et al. [183] studied the methanation of CO2 under various industrially relevant feed composition with a focus on the formation of carbonaceous species. Operando Raman spectra were recorded on 5 wt% Ni/Al2O3 with fixed-bed microreactor under varying feed concentration (Fig. 4.20). The experiment showed that gas atmospheres like CO2/H2 or CO2/H2/CH4 for the methanation or in extreme scenarios, CO2 or CO2/CH4 were tolerated by the Ni-based catalysts and no carbon depositions were observed. Only in pure CH4 atmosphere after 250 min, two intensive carbon deposition

Raman Spectroscopy

b) 100

3.0

80

10 h

2.5

60 2.0 40 1.5

5h 20 1h 1000

800 Wavenumbers

Raman intensity ratio I823/I0823

a)

Selectivity (%)

1.0 0

600

0

(cm–1)

2

4

6

8

10

Time on stream (h)

840

d) 100

716 K

80

Selectivity (%)

Intensity (arbitrary units)

880

c)

623 K

50

40

60 30 40

20

523 K

995

1000

Acrolein CO2 Acetic acid Acrylic acid Conversion

Conversion (%)

Fig. 4.19 In situ Raman spectra of MoO3
x (a, b) and (Mo6V3W1)5O14 (c, d) recorded during propene oxidation with catalytic data selectivity to acetic acid and the integrated intensity of prominent Raman band of MoO3 at 823 cm
1 referenced to its initial integrated intensity I0. (Reprinted from Mestl [182], copyright (2002) with permission from John Wiley & Sons, Ltd.)

101

Intensity (arbitrary units)

4

20

800

600

Wavenumbers (cm–1)

bands with graphitic structure have been confirmed by operando Raman spectroscopy. The two bands include an ideal graphitic G band at 1574 cm
1 and a disordered graphitic lattice mode D1 band at 1324 cm
1. The intensity of these two bands increased over time and did not stabilized after 30 min under CH4 gas environment. Simultaneously, the deactivation of the catalyst has been observed through MS data and the selectivity shifted toward CO. This work clearly demonstrates that operando Raman spectroscopy is ideally suited to investigate the formation of carbon deposits and its nature on a working catalyst upon varying gas atmospheres. This makes it a powerful and valuable method for further and more detailed studies under reaction conditions and

0 550

10

0 600

650

700

Temperature (K)

modulations of the reaction gas atmosphere as well as longterm experiments under harsh reaction conditions.

4.13

Combining Raman Spectroscopy with Other Techniques (Multimodal)

One of the major challenges for the in situ/operando characterization of catalysts is the fact that only small amounts of sites present over the catalyst surface are the true active sites. To obtain the most exact information from the reaction, it is important to have a highly sensitive analyzing tool. Moreover, catalysis involves many different phenomena such as

4

102

J. Moon et al.

b)

G

A

CO 2

O2

H2

+C

+

+C

+

0

O2

A

H2

+C

0

100

1

200

Ti

m

0

CH 4 0 140

160

180

D

E

0 000

2

–1 )

m ft (c

i n sh

1.0 0.8

D1/G ratio

0.6 0.4 0.2

D1 band intensity

a

Ram

C

G band intensity

50

CH 4

O2

H2

0 10

1E–9 Ion current (A)

H4

+C

0 15

m e(

c iti ph ra l g re t ea tu Id u c fec de st r tic hi ap ure Gr uct str

B

CH 4

) in

0 20

B

Intensity (a.u.)

0 25

D1

5 wt.% Ni/Al2O3

D1/G ratio

a)

0.0

1E–10 1E–11 H2 (m/z 2) CO2 (m/z 44) CO (m/z 28) CH4 (m/z 15)

1E–12 1E–13 0

50

100 150 200 250 Time (min)

400

450

Fig. 4.20 (a) Raman spectra recorded continuously with various gas mixture during CO2 methanation using the 5 wt. % Ni/Al2O3 catalyst and the spectroscopic fixed-bed microreactor. (b) Intensities of the characteristic G and D1 Raman band and the corresponding mass spectrometer data for the reaction cycle monitored (370  C, 1 bar,

6000 mLCO2 gcat
1 h
1): Phases A, C, and E are methanation under CO2/H2/N2, phase B is C deposition under CH4/N2, and phase D is reactivation under H2/N2. (Reprinted from Mutsz et al. [183], copyright (2015) with permission from Elsevier B.V.)

adsorption, product or byproduct formation, desorption, and coke deposition over different scales in time and space. The interpretation of these phenomena, which is spectroscopically superimposed and happening simultaneously, makes spectra interpretation highly challenging. Additionally, there has been inconsistency between different techniques due to its different operation conditions. As a solution, multitechnique approaches are desirable and have been developed for a single measurement. The most prominent advantage of multimodal analysis lies in the enhanced, correlated information of dynamic chemical and physical processes when simultaneous and synchronized data acquisition is ensured. A good number of techniques have been combined with Raman spectroscopy in a single measurement, including IR [13, 95, 184], UV-vis [185, 186], X-ray absorption spectroscopy (XAS) [187, 188], EPR [185], XRD [46, 169, 186, 188–190], etc. Among the multimodal measurements, Raman spectroscopy has been mostly combined with IR and UV-vis spectroscopies. IR and Raman are the most common and experimentally easy to combine, and they are complementary to each other due to the different selection rules. For example, Raman spectroscopy can detect carbon deposits due to the sensitivity of C-C bond vibration among sp2 hybridized carbons including polyolefin and graphitic carbon species, but it is not sensitive to C-O vibrations, which can be successfully detected by IR [32]. Payen et al. [95] studied Pd/γ-Al2O3 catalyst under the DeNOx reaction by a combination of Raman and IR spectroscopy in a single benchtop instrument. IR spectroscopy was able to detect different adsorbed NOx

species (nitrito, nitrato, nitro, and nitrate compounds) on the surface of alumina-supported palladium during NO exposure at 473 K at low concentration. Raman spectroscopy allowed the characterization of these surface species in the low spectral range regarding their bonds with the support. The ability to monitor the very same reaction via both techniques in a quasi-simultaneous approach provided full and complementary vibrational information and thus a better insight into the catalytic reaction mechanism. With the fast development of synchrotron-based X-ray techniques and their application in heterogeneous catalysis, multimodal experiments involving Raman scattering and synchrotron X-ray have shown the power of the combination. Due to its brightness from synchrotron compared with lab-based X-rays, a variety of techniques can be obtained in different beamlines. Several highly specialized beamlines such as diffraction [188, 190–193], X-ray adsorption spectroscopy (XAS) [188, 194], small-angle X-Ray scattering (SAXS) [195, 196], and high-pressure research [192, 193] have been equipped with an online Raman spectrometer in addition to the IR spectrometer. Beale et al. [187] designed a cell to combine UV-Vis, Raman, and energy-dispersed (ED) XAS spectroscopy to understand the mechanism of the dehydrogenation of propane on alumina- or silica-supported molybdenum oxide (Fig. 4.21). Initial measurements of the fresh Mo/SiO2 catalyst include a distinct 1s-4d pre-edge feature of Mo K-edge at 20002 eV in the ED-XANES (X-ray absorption near-edge structure), a strong ligand-to-metal charge transfer (LMCT) band at ca. 350 nm in the UV-Vis, and Raman bands at

4

Raman Spectroscopy

103

a) Synchrotron radiation

Absorbance

b) 0.6

Gas outlet to MS

0.4 Time

Spectrometer

0.3

300

400

UV-Vis probe

UV-Vis

Raman

PSD EDXAS Curved crystal polychromator

Catalyst

Raman intensity

c) 2000

UV-Vis source

1500

Quartz reactor tube Raman laser Gas inlet

Baseline

1000

900

4 Time

500

d) 1.2

Furnace

500 600 700 800 Wavelength (nm) 820 cm–1 1350 cm–1

0

Normalized absorbance

Thermocouple

0.5

500

1000 1500 Raman shift (cm–1)

2000

1s-5p transition

0.8

White line 0.4 1s-4d transition 0.0

20,000

20,050 Energy (eV)

Fig. 4.21 (a) Scheme of the multi-technique operando setup for UV-Vis, Raman, and ED-XAFS and time-resolved (b) UV-Vis, (c) Raman, and (d) ED-XANES spectra of Mo/SiO2 sample. (Reprinted

from Beale et al. [187], copyright (2005) with permission from The Royal Society of Chemistry)

992 (Mo¼O stretching mode), 820 and 664 (Mo-O-Mo stretching mode), and 335 cm
1 (Mo-O-Mo bending mode) for surface MoOx species in +6 oxidation state. These features, however, were observed to quickly diminish during the subsequent propane dehydrogenation step. While the ED-XANES showed the reduction from Mo6+ to Mo4+ during the reaction, there were no obvious features in the UV-Vis absorption for such reduced Mo species, and the Raman bands from MoOx decreased greatly. The contrast was explained by the occurring of both Mo reduction and coke formation that caused the increase in UV-Vis absorbance at the beginning of the reaction. Similar combined study was also carried out on Mo/Al2O3. Overall, the research showed that both catalysts deactivate over time in propane dehydrogenation via two processes: an initial deactivation due to coke formation and a more permanent one as a result of bulk changes (reduction). Coke formation, a common problem affecting the activity of propane dehydrogenation, was found to occur more quickly on Mo/Al2O3 than on Mo/SiO2. From this multimodal experiment, it showed that this operando setup is capable of providing a lot of congruous information in one experiment, and thus, it could demonstrate the potential of a new and powerful tool for studying

heterogeneous catalytic processes as well as other scientific disciplines. Although synchrotron-based X-ray experiment combined with Raman gives a high resolution, it is limited by experiment duration because it is not common to be granted more than a couple of days of beamtime at a time. Due to this limitation, it is hard to study long time duration such as the deactivation of FTS catalysts which is typically quite slow (days to weeks). For this reason, the effort to combine lab-scale XRD and Raman has been conducted. For example, Cats and Weckhuysen [189] designed and constructed the operando setup for combining XRD and Raman spectroscopy. It is equipped with Raman probe near the X-ray detector and a GC at the end of the vent. Through this setup, the authors were able to measure the product stream simultaneously in addition to obtaining deactivation information over 100 h of a cycle over Co/TiO2 Fischer-Tropsch synthesis (FTS) catalysts. Based on various types of combination examples, it is clear that the combination of multi-techniques gives more complete picture of how catalysts work and thus enable the establishment of structure-activity relationship of heterogeneous catalysis. Furthermore, the technical evolution makes

104

J. Moon et al.

it possible to design more compact and simple equipment so that it will be possible to combine more techniques into one single instrument and obtain more information in a single measurement.

4.14

Summary

Raman spectroscopy has become a powerful tool in the field of catalysis. Raman experiments can be carried out under wide reaction conditions without any limitations of the sample phase. It is thus possible to investigate structure evolution of catalysts from birth (synthesis) to death (reaction and deactivation). It can provide substantial information about bulk transition metal oxides by elucidating the phase structure, transformation, and reaction mechanism. Insightful information can also be revealed on the molecular structure and the distribution of surface dispersed metal oxide species and the role of these oxide species in catalytic reactions. Due to the sensitivity for carbon materials, Raman spectroscopy has been also applied to analyze coke formation chemistry, and thus, it could be used to elucidate the deactivation mechanism of certain catalysts under reaction conditions. Developments of operando Raman techniques allow to probe the structure of catalysts under real reaction conditions and obtain both structural and kinetic information at the same time, permitting the powerful correlation of structure-catalysis relationship. Currently, operando Raman experiments have greatly improved our understanding of different types of catalytic systems and reactions, but with limitations on temporal-spatial resolution. The combination of operando Raman spectroscopy and imaging techniques such as surface-enhanced Raman or tip-enhanced Raman will become a very important approach to deliver valuable insights with high resolution in both time and space into complex catalytic reactions, thus enabling the design of more efficient catalysts. Notably, the development of proper reaction cells for such combined Raman and imaging techniques will be a key. Raman spectroscopy has seen its extensive utilization in catalysis research in the past few decades, and thus, large Raman data sets have been generated in the literature. Such data has not been well capitalized for catalysis research yet. Owing to the recent developments of quantum chemistry and computer science including high-accuracy methodologies, numerical treatment of complex systems, and machine learning, we believe the coupling of theory, modeling, and data science with Raman (and other spectroscopy) experiments will be timely and could revolutionize the field of in situ/ operando characterization of catalysts and catalysis, thus playing important roles in promoting the understanding and advancement of catalysis in the near future.

Acknowledgments This work is sponsored by the US Department of Energy, Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division, Catalysis Science program. ZW is partly supported by the Center for Nanophase Materials (CNMS), which is a US Department of Energy, Office of Science User Facility at Oak Ridge National Laboratory. Notice: This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the US Department of Energy. The US Government retains, and the publisher, by accepting the article for publication, acknowledges that the US Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript or allow others to do so, for US Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/ downloads/doe-public-access-plan).

References 1. Raman, C.V., Krishnan, K.S.: A new type of secondary radiation. Nature. 121, 501–502 (1928) 2. Brown, F.R., Makovsky, L.E.: Raman spectra of a cobalt oxidemolybdenum oxide supported catalyst. Appl. Spectrosc. 31, 44–46 (1977) 3. Brown, F.R., Makovsky, L.E., Rhee, K.H.: Raman spectra of supported molybdena catalysts. II. Sulfided, used, and regenerated catalysts. J. Catal. 50, 385–389 (1977) 4. Lunsford, J.H., Yang, X., Haller, K., Laane, J., Mestl, G., Knoezinger, H.: In situ Raman spectroscopy of peroxide ions on Ba/Mgo catalysts. J. Phys. Chem. 97, 13810–13813 (1993) 5. Pittman, R.M., Bell, A.T.: Raman investigations of Nh3 adsorption on TiO2, Nb2O5, and Nb2O5/TiO2. Catal. Lett. 24, 1–13 (1994) 6. Wachs, I.E., Roberts, C.A.: Monitoring surface metal oxide catalytic active sites with Raman spectroscopy. Chem. Soc. Rev. 39, 5002–5017 (2010) 7. Chua, Y.T., Stair, P.C., Wachs, I.E.: A comparison of ultraviolet and visible Raman spectra of supported metal oxide catalysts. J. Phys. Chem. B. 105, 8600–8606 (2001) 8. Wachs, I.E.: In situ Raman spectroscopy studies of catalysts. Top. Catal. 8, 57–63 (1999) 9. Campion, A., Kambhampati, P.: Surface-enhanced Raman scattering. Chem. Soc. Rev. 27, 241–250 (1998) 10. Arya, K., Zeyher, R.: Theory of surface-enhanced Raman scattering from molecules adsorbed at metal surfaces. Phys. Rev. B. 24, 1852–1865 (1981) 11. Zhang, Y., Fu, D., Xu, X., Sheng, Y., Xu, J., Han, Y.-F.: Application of operando spectroscopy on catalytic reactions. Curr. Opin. Chem. Eng. 12, 1–7 (2016) 12. Kim, H., Kosuda, K.M., Van Duyne, R.P., Stair, P.C.: Resonance Raman and surface- and tip-enhanced Raman spectroscopy methods to study solid catalysts and heterogeneous catalytic reactions. Chem. Soc. Rev. 39, 4820–4844 (2010) 13. Stavitski, E., Weckhuysen, B.M.: Infrared and Raman imaging of heterogeneous catalysts. Chem. Soc. Rev. 39, 4615–4625 (2010) 14. Stair, P.C.: Advances in Raman spectroscopy methods for catalysis research. Curr. Opinion Solid State Mater. Sci. 5, 365–369 (2001) 15. Fan, F., Feng, Z., Li, C.: UV Raman spectroscopic study on the synthesis mechanism and assembly of molecular sieves. Chem. Soc. Rev. 39, 4794–4801 (2010) 16. Wachs, I.E., Hardcastle, F.D.: Applications of raman spectroscopy to heterogeneous catalysis. In: Spivey, J.J., Agarwal, S.K. (eds.) Catalysis: Volume 10, vol. 10, pp. 102–153. The Royal Society of Chemistry, London (1993)

4

Raman Spectroscopy

17. Wachs, I.E., BaÑAres, M.: In situ and operando Raman spectroscopy of oxidation catalysts. In: Handbook of Advanced Methods and Processes in Oxidation Catalysis, pp. 420–446. Imperial College Press (2011). https://doi.org/10.1142/9781848167513_0017 18. Grasselli, J.G., Snavely, M.K., Bulkin, B.J.: Applications of Raman spectroscopy. Physics Reports. 65, 231–344 (1980) 19. Bañares, M.A., Mestl, G.: Chapter 2. Structural characterization of operating catalysts by Raman spectroscopy. In: Advances in Catalysis, vol. 52, pp. 43–128. Academic Press, Amsterdam (2009) 20. Cai, Y., Wang, C.B., Wachs, I.E.: Reaction induced spreading of metal oxides: in situ Raman spectroscopic studies during oxidation reactions. In: Grasselli, R.K., Oyama, S.T., Gaffney, A.M., Lyons, J.E. (eds.) Studies in Surface Science and Catalysis, vol. 110, pp. 255–264. Elsevier, Amsterdam (1997) 21. Carreira, L.A., Goss, L.P., Malloy, T.G.: Preresonance enhancement of the coherent anti-stokes Raman spectra of fluorescent compounds. J. Chem. Phys. 68, 280–284 (1978) 22. Tolles, W.M., Nibler, J.W., McDonald, J.R., Harvey, A.B.: A review of the theory and application of coherent anti-stokes Raman spectroscopy (CARS). Appl. Spectrosc. 31, 253–271 (1977) 23. Ferraro, J.R., Nakamoto, K., Brown, C.W.: Chapter 1 – Basic theory. In: Ferraro, J.R., Nakamoto, K., Brown, C.W. (eds.) Introductory Raman Spectroscopy (Second Edition), pp. 1–94. Academic Press, San Diego (2003). https://doi.org/10.1016/B978012254105-6/50004-4 24. Hill, C.G., Wilson, J.H.: Raman spectroscopy of iron molybdate catalyst systems: part I. Preparation of unsupported catalysts. J. Mol. Catal. 63, 65–94 (1990) 25. Griffith, W.P., Wickins, T.D.: Raman studies on species in aqueous solutions. Part II. Oxy-species of metals of groups VIA, VA, and IVA. J. Chem. Soc. A, 675 (1967). https://doi.org/10.1039/ J19670000675675-679 26. Griffith, W.P., Lesniak, P.J.B.: Raman studies on species in aqueous solutions. Part III. Vanadates, molybdates, and tungstates. J. Chem. Soc. A, 1066 (1969). https://doi.org/10.1039/ J196900010661066-1071 27. Kervinen, K., Korpi, H., Gerbrand Mesu, J., Soulimani, F., Repo, T., Rieger, B., Leskelä, M., Weckhuysen, B.M.: Mechanistic insights into the oxidation of veratryl alcohol with Co(Salen) and oxygen in aqueous media: an in-situ spectroscopic study. Eur. J. Inorg. Chem. 2005, 2591–2599 (2005) 28. Mestl, G., Knözingers, H., Lunsford, J.H.: High temperature in situ Raman spectroscopy of working oxidative coupling catalysts. Ber. Bunsenges. Phys. Chem. 97, 319–321 (1993) 29. Haller, K., Lunsford, J.H., Laane, J.: Temperature dependence of the Raman spectrum of barium peroxide. J. Phys. Chem. 100, 551–555 (1996) 30. Urakawa, A., Trachsel, F., von Rohr, P.R., Baiker, A.: On-chip Raman analysis of heterogeneous catalytic reaction in supercritical CO2: phase behaviour monitoring and activity profiling. Analyst. 133, 1352–1354 (2008) 31. da Silva, A.N., Pinto, R.C.F., Freire, P.T.C., Junior, J.A.L., Oliveira, A.C., Filho, J.M.: Temperature and high pressure effects on the structural features of catalytic nanocomposites oxides by Raman spectroscopy. Spectrochim. Acta A Mol. Biomol. Spectrosc. 138, 763–773 (2015) 32. Korhonen, S.T., Bañares, M.A., Fierro, J.L.G., Krause, A.O.I.: Adsorption of methanol as a probe for surface characteristics of zirconia-, alumina-, and zirconia/alumina-supported chromia catalysts. Catal. Today. 126, 235–247 (2007) 33. Hu, H., Wachs, I.E.: Catalytic properties of supported molybdenum oxide catalysts: in situ Raman and methanol oxidation studies. J. Phys. Chem. 99, 10911–10922 (1995) 34. Li, W., Oyama, S.T.: Ethanol oxidation using ozone over supported manganese oxide catalysts: an in situ laser Raman study. Proc.

105 Stud. Surf. Sci. Catal., 873–882 (1997) https://www.scopus.com/ inward/record.uri?eid¼2-s2.0-33750852821&partnerID¼40& md5¼36673441bbe31674fc4ff388e4ec776a 35. Gao, X., Bañares, M.A., I. E. Wachs: Ethane and N-butane oxidation over supported vanadium oxide catalysts: an in situ UV-visible diffuse reflectance spectroscopic investigation. J. Catal. 188, 325–331 (1999). https://doi.org/10.1006/jcat.1999.2647 36. Chase, D.B.: Fourier transform Raman spectroscopy. J. Am. Chem. Soc. 108, 7485–7488 (1986) 37. Baurecht, D., Fringeli, U.P.: Quantitative modulated excitation Fourier transform infrared spectroscopy. Rev. Sci. Instrum. 72, 3782–3792 (2001) 38. Hartman, T., Wondergem, C.S., Kumar, N., van den Berg, A., Weckhuysen, B.M.: Surface- and tip-enhanced Raman spectroscopy in catalysis. J. Phys. Chem. Lett. 7, 1570–1584 (2016) 39. Lombardi, J.R., Birke, R.L.: The theory of surface-enhanced Raman scattering. J. Chem. Phys. 136, 144704 (2012) 40. Harvey, C.E., Weckhuysen, B.M.: Surface- and tip-enhanced Raman spectroscopy as operando probes for monitoring and understanding heterogeneous catalysis. Catal. Lett. 145, 40–57 (2015) 41. Kumar, N., Mignuzzi, S., Su, W., Roy, D.: Tip-enhanced Raman spectroscopy: principles and applications. EPJ Tech. Instrum. 2, 9 (2015) 42. Waleska, P., Rupp, S., Hess, C.: Operando multiwavelength and time-resolved Raman spectroscopy: structural dynamics of a supported vanadia catalyst at work. J. Phys. Chem. C. 122, 3386–3400 (2018) 43. Guerrero-Pérez, M.O., Banares, M.A.: Operando Raman study of alumina-supported Sb-V-O catalyst during propane ammoxidation to acrylonitrile with on-line activity measurement. Chem. Commun. 2, 1292–1293 (2002) 44. Guerrero-Pérez, M.O., Bañares, M.A.: Operando Raman-Gc studies of alumina-supported Sb-V-O catalysts and role of the preparation method. Catal. Today. 96, 265–272 (2004) 45. Martínez-Huerta, M.V., Deo, G., Fierro, J.L.G., Bañares, M.A.: Operando Raman-Gc study on the structure
activity relationships in V 5+/Ceo 2 catalyst for ethane oxidative dehydrogenation: the formation of Cevo 4. J. Phys. Chem. C. 112, 11441–11447 (2008) 46. Fu, D., Dai, W., Xu, X., Mao, W., Su, J., Zhang, Z., Shi, B., Smith, J., Li, P., Xu, J., Han, Y.-F.: Probing the structure evolution of ironbased Fischer–Tropsch to produce olefins by operando Raman spectroscopy. ChemCatChem. 7, 752–756 (2015) 47. Wu, Z., Dai, S., Overbury, S.H.: Multiwavelength Raman spectroscopic study of silica-supported vanadium oxide catalysts. J. Phys. Chem. C. 114, 412–422 (2010) 48. Wu, Z., Rondinone, A.J., Ivanov, I.N., Overbury, S.H.: Structure of vanadium oxide supported on ceria by multiwavelength Raman spectroscopy. J. Phys. Chem. C. 115, 25368–25378 (2011) 49. Wu, Z.: Multi-wavelength Raman spectroscopy study of supported vanadia catalysts: structure identification and quantification. Chin. J. Catal. 35, 1591–1608 (2014) 50. Payen, E., Kasztelan, S., Grimblot, J., Bonnelle, J.P.: Study of the reduction of Moo3/gamma-Al2o3 and Wo3/gamma-Al2o3 catalysts by laser Raman spectroscopy. J. Mol. Struct. 143, 259–262 (1986) 51. Payen, E., Grimblot, J., Kasztelan, S.: Study of oxidic and reduced alumina-supported molybdate and heptamolybdate species by in situ laser Raman spectroscopy. J. Phys. Chem. 91, 6642–6648 (1987) 52. Liu, H., Gaigneaux, E.M., Imoto, H., Shido, T., Iwasawa, Y.: Performance and characterization of novel re – Sb – O catalysts active for the selective oxidation of isobutylene to methacrolein. J. Phys. Chem. B. 104, 2033–2043 (2000) 53. Xie, S., Rosynek, M.P., Lunsford, J.H.: Effect of laser heating on the local temperature and composition in Raman spectroscopy: a

4

106 study of Ba(No3)2 and Bao2 decomposition. Appl. Spectrosc. 53, 1183–1187 (1999) 54. Cheng, C.P., Ludowise, J.D., Schrader, G.L.: Controlledatmosphere rotating cell for in situ studies of catalysts using laser Raman spectroscopy. Appl. Spectrosc. 34, 146–150 (1980) 55. Yilbas, B.S.: Chapter 2 – Conduction-limited laser pulsed laser heating: Fourier heating model. In: Yilbas, B.S. (ed.) Laser Heating Applications, pp. 7–51. Elsevier, Boston (2012). https://doi.org/10. 1016/B978-0-12-415782-8.00002-4 56. Sheik-Bahae, M., Epstein, R.I.: Laser cooling of solids. Laser Photonics Rev. 3, 67–84 (2009) 57. Chua, Y.T., Stair, P.C.: A novel fluidized bed technique for measuring UV Raman spectra of catalysts and adsorbates. J. Catal. 196, 66–72 (2000) 58. Puppels, G.J., Huizinga, A., Krabbe, H.W., de Boer, H.A., Gijsbers, G., de Mul, F.F.M.: A high-throughput Raman notch filter set. Rev. Sci. Instrum. 61, 3709–3712 (1990) 59. Barron, L.D., Hecht, L., Hug, W., MacIntosh, M.J.: Backscattered Raman optical activity with a CCD detector. J. Am. Chem. Soc. 111, 8731–8732 (1989) 60. Richet, P., Gillet, P., Pierre, A., Bouhifd, M.A., Daniel, I., Fiquet, G.: Raman spectroscopy, X-ray diffraction, and phase relationship determinations with a versatile heating cell for measurements up to 3600 K (or 2700 K in air). J. Appl. Phys. 74, 5451–5456 (1993) 61. Zouboulis, E.S., Grimsditch, M.: Raman scattering in diamond up to 1900 K. Phys. Rev. B. 43, 12490–12493 (1991) 62. Herchen, H., Cappelli, M.A.: First-order Raman spectrum of diamond at high temperatures. Phys. Rev. B. 43, 11740–11744 (1991) 63. Alvarenga, A.D., Grimsditch, M., Polian, A.: Raman scattering from cubic boron nitride up to 1600 K. J. Appl. Phys. 72, 1955–1956 (1992) 64. Brookes, A., Dyke, J.M., Hendra, P.J., Meehan, S.: The Ft-Raman spectroscopic study of polymers at temperatures in excess of 200 C. Spectrochim. Acta A Mol. Biomol. Spectrosc. 53, 2313–2321 (1997) 65. Bravo-Suárez, J.J., Srinivasan, P.D.: Design characteristics of in situ and operando ultraviolet-visible and vibrational spectroscopic reaction cells for heterogeneous catalysis. Catal. Rev. 59, 295–445 (2017) 66. Zobeiri, H., Xu, S., Yue, Y., Zhang, Q., Xie, Y., Wang, X.: Effect of temperature on Raman intensity of Nm-thick Ws2: combined effects of resonance Raman, optical properties, and Interface optical interference. Nanoscale. 12, 6064–6078 (2020) 67. Zouboulis, E., Renusch, D., Grimsditch, M.: Advantages of ultraviolet Raman scattering for high temperature investigations. Appl. Phys. Lett. 72, 1–3 (1998) 68. Kato, D.: Improvement of Raman spectroscopy by quenching the fluorescence. J. Appl. Phys. 45, 2281–2282 (1974) 69. Jeziorowski, H., Knözinger, H.: Laser induced electronic excitation of surface hydroxide ions and scattering background in laser Raman spectra of oxide surfaces. Chem. Phys. Lett. 51, 519–522 (1977) 70. Jeziorowski, H., Knoezinger, H.: Raman and ultraviolet spectroscopic characterization of molybdena on alumina catalysts. J. Phys. Chem. 83, 1166–1173 (1979) 71. Chase, B.: Fourier transform Raman spectroscopy. Anal. Chem. 59, 881A–889A (1987) 72. Stair, P.C.: The application of UV Raman spectroscopy for the characterization of catalysts and catalytic reactions. In: Gates, B.C., Knözinger, H. (eds.) Advances in Catalysis, vol. 51, pp. 75–98. Academic Press, Amsterdam (2007) 73. Bruckner, S., Jeziorowski, H., Knözinger, H.: Frequency modulation Raman spectroscopy. Chem. Phys. Lett. 105, 218–222 (1984) 74. Giles, J.H., Gilmore, D.A., Denton, M.B.: Quantitative analysis using Raman spectroscopy without spectral standardization. J. Raman Spectrosc. 30, 767–771 (1999)

J. Moon et al. 75. Wu, Z., Zhang, C., Stair, P.C.: Influence of absorption on quantitative analysis in Raman spectroscopy. Catal. Today. 113, 40–47 (2006) 76. Tinnemans, S.J., Kox, M.H.F., Nijhuis, T.A., Visser, T., Weckhuysen, B.M.: Real time quantitative Raman spectroscopy of supported metal oxide catalysts without the need of an internal standard. Phys. Chem. Chem. Phys. 7, 211–216 (2005) 77. Chan, S.S., Wachs, I.E., Murrell, L.L.: Relative Raman crosssections of tungsten oxides: [Wo3, Al2(Wo4)3 and Wo3 Al2o3]. J. Catal. 90, 150–155 (1984) 78. Lin, L.-T., Mann, C.K., Vickers, T.J.: Feasibility of quantitative UV resonance Raman spectroscopy with a KRF excimer laser. Appl. Spectrosc. 41, 422–427 (1987) 79. Bergwerff, J.A., Visser, T., Leliveld, G., Rossenaar, B.D., de Jong, K.P., Weckhuysen, B.M.: Envisaging the physicochemical processes during the preparation of supported catalysts: Raman microscopy on the impregnation of Mo onto Al2o3 extrudates. J. Am. Chem. Soc. 126, 14548–14556 (2004) 80. Aarnoutse, P.J., Westerhuis, J.A.: Quantitative Raman reaction monitoring using the solvent as internal standard. Anal. Chem. 77, 1228–1236 (2005) 81. Kuba, S., Knözinger, H.: Time-resolved in situ Raman spectroscopy of working catalysts: sulfated and tungstated zirconia. J. Raman Spectrosc. 33, 325–332 (2002) 82. Dou, X., Takama, T., Yamaguchi, Y., Yamamoto, H., Ozaki, Y.: Enzyme immunoassay utilizing surface-enhanced Raman scattering of the enzyme reaction product. Anal. Chem. 69, 1492–1495 (1997) 83. Wu, Z., Stair, P.C., Rugmini, S., Jackson, S.D.: Raman spectroscopic study of V/Θ-Al2o3 catalysts: quantification of surface vanadia species and their structure reduced by hydrogen. J. Phys. Chem. C. 111, 16460–16469 (2007) 84. Seki, H.: The effect of Co on the UHV Sers of pyridine on silver. J. Vac. Sci. Technol. 20, 584–587 (1982) 85. Kudelski, A., Pettinger, B.: Fluctuations of surface-enhanced Raman spectra of co adsorbed on gold substrates. Chem. Phys. Lett. 383, 76–79 (2004) 86. Cai, W.-B., She, C.-X., Ren, B., Yao, J.-L., Tian, Z.-W., Tian, Z.Q.: Surface Raman spectroscopic investigation of pyridine adsorption at platinum electrodes—effects of potential and electrolyte. J. Chem. Soc. Faraday Trans. 94, 3127–3133 (1998) 87. Góra-Marek, K., Datka, J.: IR studies of OH groups in mesoporous aluminosilicates. Appl. Catal. A Gen. 302, 104–109 (2006) 88. Lamberti, C., Zecchina, A., Groppo, E., Bordiga, S.: Probing the surfaces of heterogeneous catalysts by in situ IR spectroscopy. Chem. Soc. Rev. 39, 4951–5001 (2010) 89. Ferraro, J.R.: Introductory Raman Spectroscopy. Elsevier, Amsterdam (2003) 90. Wiercigroch, E., Szafraniec, E., Czamara, K., Pacia, M.Z., Majzner, K., Kochan, K., Kaczor, A., Baranska, M., Malek, K.: Raman and infrared spectroscopy of carbohydrates: a review. Spectrochim. Acta A Mol. Biomol. Spectrosc. 185, 317–335 (2017) 91. Fan, F., Feng, Z., Li, G., Sun, K., Ying, P., Li, C.: In situ UV Raman spectroscopic studies on the synthesis mechanism of zeolite X. Chem. Eur. J. 14, 5125–5129 (2008) 92. Mougin, J., Le Bihan, T., Lucazeau, G.: High-pressure study of Cr2o3 obtained by high-temperature oxidation by X-ray diffraction and Raman spectroscopy. J. Phys. Chem. Solids. 62, 553–563 (2001) 93. Vlasko-Vlasov, V., Joshi-Imre, A., Bahns, J.T., Chen, L., Ocola, L., Welp, U.: Liquid cell with plasmon lenses for surface enhanced Raman spectroscopy. Appl. Phys. Lett. 96, 203103 (2010) 94. Silverwood, I.P., Hamilton, N.G., McFarlane, A.R., Kapitán, J., Hecht, L., Norris, E.L., Mark Ormerod, R., Frost, C.D., Parker, S.F., Lennon, D.: Application of inelastic neutron scattering to

4

Raman Spectroscopy

studies of CO2 reforming of methane over alumina-supported nickel and gold-doped nickel catalysts. Phys. Chem. Chem. Phys. 14, 15214–15225 (2012) 95. Le Bourdon, G., Adar, F., Moreau, M., Morel, S., Reffner, J., Mamede, A.S., Dujardin, C., Payen, E.: In situ characterization by Raman and IR vibrational spectroscopies on a single instrument: denox reaction over a Pd/Γ-Al2O3 catalyst. Phys. Chem. Chem. Phys. 5, 4441–4444 (2003) 96. Okamoto, Y., Imanaka, T.: Interaction chemistry between molybdena and alumina: infrared studies of surface hydroxyl groups and adsorbed carbon dioxide on Aluminas modified with molybdate, sulfate, or fluorine anions. J. Phys. Chem. 92, 7102–7112 (1988) 97. Kerkhof, F.P.J.M., Moulijn, J.A., Thomas, R.: Laser-Raman spectroscopy of the alumina-supported rhenium oxide metathesis catalyst. J. Catal. 56, 279–283 (1979) 98. Wang, L., Hall, W.: On the genesis of molybdena-alumina catalyst. J. Catal. 66, 251–255 (1980) 99. Wu, Z., Kim, H.-S., Stair, P.C., Rugmini, S., Jackson, S.D.: On the structure of vanadium oxide supported on aluminas: UV and visible Raman spectroscopy, UV
visible diffuse reflectance spectroscopy, and temperature-programmed reduction studies. J. Phys. Chem. B. 109, 2793–2800 (2005) 100. Lewandowska, A.E., Bañares, M.A., Khabibulin, D.F., Lapina, O.B.: Precursor effect on the molecular structure, reactivity, and stability of alumina-supported vanadia. J. Phys. Chem. C. 113, 20648–20656 (2009) 101. Wang, L., Hall, W.K.: The preparation and properties of rheniaalumina catalysts. J. Catal. 82, 177–184 (1983) 102. Williams, C.C., Ekerdt, J.G., Jehng, J.M., Hardcastle, F.D., Wachs, I.E.: A Raman and ultraviolet diffuse reflectance spectroscopic investigation of alumina-supported molybdenum oxide. J. Phys. Chem. 95, 8791–8797 (1991) 103. Knözinger, H., Jezlorowski, H.: Raman spectra of molybdenum oxide supported on the surface of aluminas. J. Phys. Chem. 82, 2002–2005 (1978) 104. Agudo, A.L., Gil, F.J., Calleja, J.M., Fernández, V.: Raman spectroscopic study of MoO3/Γ-Al2O3 and Coo(or NiO)MoO3/Γ-Al2O3 oxide catalysts prepared by different methods. J. Raman Spectrosc. 11, 454–458 (1981) 105. Leyrer, J., Mey, D., Knözinger, H.: Spreading behavior of molybdenum trioxide on alumina and silica: a Raman microscopy study. J. Catal. 124, 349–356 (1990) 106. Hashimoto, K., Badarla, V.R., Kawai, A., Ideguchi, T.: Complementary vibrational spectroscopy. Nat. Commun. 10, 4411 (2019) 107. Brown, F.R., Makovsky, L.E., Rhee, K.H.: Rotating vacuum cell for Raman spectroscopy. Appl. Spectrosc. 31, 563–565 (1977) 108. Li, M., Feng, Z., Xiong, G., Ying, P., Xin, Q., Li, C.: Phase transformation in the surface region of zirconia detected by UV Raman spectroscopy. J. Phys. Chem. B. 105, 8107–8111 (2001) 109. Schrader, B., Hoffmann, A., Keller, S.: Near-infrared Fourier transform Raman spectroscopy: facing absorption and background. Spectrochim. Acta A: Mol. Spectrosc. 47, 1135–1148 (1991) 110. Kubelka, P., Munk, F.J.Z.T.P.: An article on optics of paint layers. Fuer Tekn. Physik. 12, 593–609 (1931) 111. Yassin, O.A., Alamri, S.N., Joraid, A.A.: Effect of particle size and laser power on the Raman spectra of CuAlO2 delafossite nanoparticles. J. Phys. D. Appl. Phys. 46, 235301 (2013) 112. Graham, G.W., Weber, W.H., Peters, C.R., Usmen, R.: Empirical method for determining CeO2-particle size in catalysts by Raman spectroscopy. J. Catal. 130, 310–313 (1991) 113. Schlotter, N.: Analytical Raman Spectroscopy. Chemical Analysis Series. Grasselli, J.G., Bulkin, B.J. (eds.) vol. 114, Winefordner, J.D., Kolthoff, I.M. (Series editors). Wiley, New York, (1991) (ISBN 0-471-51955-3). 462 pp. (1992)

107 114. Schwab, S.D., McCreery, R.L.: Versatile, efficient Raman sampling with fiber optics. Anal. Chem. 56, 2199–2204 (1984) 115. Newman, C.D., Bret, G.G., McCreery, R.L.: Fiber-optic sampling combined with an imaging spectrograph for routine Raman spectroscopy. Appl. Spectrosc. 46, 262–265 (1992) 116. Borio, V.G., Vinha, R., Nicolau, R.A., de Oliveira, H.P.M., de Lima, C.J., Silveira, L.: Quantitative evaluation of acetaminophen in oral solutions by dispersive Raman spectroscopy for quality control. Spectr.-Int. J. 27, 215–228 (2012) 117. Santos, L.F., Wolthuis, R., Koljenović, S., Almeida, R.M., Puppels, G.J.: Fiber-optic probes for in vivo Raman spectroscopy in the high-wavenumber region. Anal. Chem. 77, 6747–6752 (2005) 118. Woodward, L.A., Waters, D.N.: A water-cooled mercury arc lamp for the excitation of Raman spectra. J. Sci. Instrum. 34, 222–224 (1957) 119. Williamson, J.M., Bowling, R.J., McCreery, R.L.: Near-infrared Raman spectroscopy with a 783-NM diode laser and CCD array detector. Appl. Spectrosc. 43, 372–375 (1989) 120. Fujiwara, M., Hamaguchi, H., Tasumi, M.: Measurements of spontaneous Raman scattering with Nd:Yag 1064-nm laser light. Appl. Spectrosc. 40, 137–139 (1986) 121. The High-Power Ion Laser from Coherent. Inc. https://www. coherent.com/lasers/main/ion-lasers 122. Wang, Y., McCreery, R.L.: Evaluation of a diode laser/charge coupled device spectrometer for near-infrared Raman spectroscopy. Anal. Chem. 61, 2647–2651 (1989) 123. Robert, B.: Resonance Raman spectroscopy. Photosynth. Res. 101, 147–155 (2009) 124. Kagan, M.R., McCreery, R.L.: Reduction of fluorescence interference in Raman spectroscopy via analyte adsorption on graphitic carbon. Anal. Chem. 66, 4159–4165 (1994) 125. Li, C., Stair, P.C.: Ultraviolet Raman spectroscopy characterization of coke formation in zeolites. Catal. Today. 33, 353–360 (1997) 126. Li, C., Stair, P.C.: An advance in Raman studies of catalysts: ultraviolet resonance Raman spectroscopy. In: Hightower, J.W., Nicholas Delgass, W., Iglesia, E., Bell, A.T. (eds.) Studies in Surface Science and Catalysis, vol. 101, pp. 881–890, Elsevier, Amsterdam (1996) 127. Moskovits, M.: Surface-enhanced spectroscopy. Rev. Mod. Phys. 57, 783–826 (1985) 128. van Schrojenstein Lantman, E.M., Deckert-Gaudig, T., Mank, A.J. G., Deckert, V., B. M.: Weckhuysen: catalytic processes monitored at the nanoscale with tip-enhanced Raman spectroscopy. Nat. Nanotechnol. 7, 583–586 (2012) 129. Kumar, N., Stephanidis, B., Zenobi, R., Wain, A.J., Roy, D.: Nanoscale mapping of catalytic activity using tip-enhanced Raman spectroscopy. Nanoscale. 7, 7133–7137 (2015) 130. Avouris, P., Demuth, J.E.: Electronic excitations of benzene, pyridine, and pyrazine adsorbed on Ag(111). J. Chem. Phys. 75, 4783–4794 (1981) 131. Albrecht, M.G., Creighton, J.A.: Anomalously intense Raman spectra of pyridine at a silver electrode. J. Am. Chem. Soc. 99, 5215–5217 (1977) 132. Jeanmaire, D.L., Van Duyne, R.P.: Surface Raman spectroelectrochemistry: part I. heterocyclic, aromatic, and aliphatic amines adsorbed on the anodized silver electrode. J. Electroanal. Chem. Interfacial Electrochem. 84, 1–20 (1977) 133. Harvey, C.E., van Schrojenstein Lantman, E.M., Mank, A.J.G., Weckhuysen, B.M.: An integrated AFM-Raman instrument for studying heterogeneous catalytic systems: a first showcase. Chem. Commun. 48, 1742–1744 (2012) 134. Łojewska, J., Knapik, A., Kołodziej, A., Jodłowski, P.: Far field combined AFM and micro-Raman imaging for characterisation of surface of structured catalysts: example of Pd doped CoOx catalysts on precalcined kanthal steel. Top. Catal. 56, 1088–1095 (2013)

4

108 135. Zhang, R., Zhang, Y., Dong, Z.C., Jiang, S., Zhang, C., Chen, L.G., Zhang, L., Liao, Y., Aizpurua, J., Luo, Y., Yang, J.L., Hou, J.G.: Chemical mapping of a single molecule by plasmon-enhanced Raman scattering. Nature. 498, 82 (2013) 136. Zumbusch, A., Holtom, G.R., Xie, X.S.: Three-dimensional vibrational imaging by coherent anti-stokes Raman scattering. Phys. Rev. Lett. 82, 4142–4145 (1999) 137. Min, W., Lu, S., Holtom, G.R., Xie, X.S.: Triple-resonance coherent anti-stokes Raman scattering microspectroscopy. ChemPhysChem. 10, 344–347 (2009) 138. Linkam Scientific Instruments – CCR1000 Stage. http://www. linkam.co.uk/ccr1000-features 139. Beato, P., Schachtl, E., Barbera, K., Bonino, F., Bordiga, S.: Operando Raman spectroscopy applying novel fluidized bed micro-reactor technology. Catal. Today. 205, 128–133 (2013) 140. Vannice, M.A., Joyce, W.H.: Kinetics of Catalytic Reactions, vol. 134. Springer, New York (2005) 141. Kapteijn, F., Moulijn, J.: Laboratory Catalytic Reactors: Aspects of Catalyst Testing, 2019–2045 (2008) 142. Berty, J.J.A.I.C.: Laboratory reactors for catalytic studies. Appl. Ind. Catal. 1, 41–67 (1983) 143. Knitter, R., Liauw, M.A.: Ceramic microreactors for heterogeneously catalysed gas-phase reactions. Lab Chip. 4, 378–383 (2004) 144. Maghsoumi, A., Ravanelli, A., Consonni, F., Nanni, F., Lucotti, A., Tommasini, M., Donazzi, A., Maestri, M.: Design and testing of an operando-Raman annular reactor for kinetic studies in heterogeneous catalysis. React. Chem. Eng. 2, 908–918 (2017) 145. Villa, P.L., Trifirò, F., Pasquon, I.: Study of the interaction between CoMoO4 and 341-1341-1341-1-Al2O3 by Raman spectroscopy. React. Kinet. Catal. Lett. 1, 341–344 (1974) 146. Chan, S.S., Bell, A.T.: Characterization of the preparation of Pd SiO2 and Pd La2O3 by laser Raman spectroscopy. J. Catal. 89, 433–441 (1984) 147. Roozeboom, F., Gellings, P.J., Medema, J.: Vanadium oxide monolayer catalysts. Z. Phys. Chem. 111, 215–224 (1978) 148. Schrader Jr., G.L., Hill Jr., C.G.: Variable temperature sample cell and optical mount for laser Raman studies of adsorbed species. Rev. Sci. Instrum. 46, 1335–1337 (1975) 149. Krasser, W., Ranade, A., Koglin, E.: Low temperature Raman measurements of carbon monoxide adsorbed on nickel at high pressures. J. Raman Spectrosc. 6, 209–212 (1977) 150. Tian, H., Roberts, C.A., Wachs, I.E.: Molecular structural determination of molybdena in different environments: aqueous solutions, bulk mixed oxides, and supported MoO3 catalysts. J. Phys. Chem. C. 114, 14110–14120 (2010) 151. Chan, S.S., Wachs, I.E., Murrell, L.L., Wang, L., Hall, W.K.: In situ laser Raman spectroscopy of supported metal oxides. J. Phys. Chem. 88, 5831–5835 (1984) 152. Stencel, J.M., Makovsky, L.E., Sarkus, T.A., de Vries, J., Thomas, R., Moulijn, J.A.: Raman spectroscopic investigation of the effect of H2O on the molybdenum surface species in MoO3 Al2O3 catalysts. J. Catal. 90, 314–322 (1984) 153. Hardcastle, F.D., Wachs, I.E.: Raman spectroscopy of chromium oxide supported on Al2o3, Tio2 and Sio2: a comparative study. J. Mol. Catal. 46, 173–186 (1988) 154. Jehng, J.M., Wachs, I.E.: Niobium oxide solution chemistry. J. Raman Spectrosc. 22, 83–89 (1991) 155. Wu, Z., Li, M., Overbury, S.H.: A Raman spectroscopic study of the speciation of vanadia supported on ceria nanocrystals with defined surface planes. ChemCatChem. 4, 1653–1661 (2012) 156. Mestl, G., Rosynek, M.P., Lunsford, J.H.: Decomposition of nitric oxide over barium oxide supported on magnesium oxide. 2. In situ Raman characterization of phases present during the catalytic reaction. J. Phys. Chem. B. 101, 9321–9328 (1997)

J. Moon et al. 157. Pushkarev, V.V., Kovalchuk, V.I., D'Itri, J.L.: Probing defect sites on the CeO2 surface with dioxygen. J. Phys. Chem. B. 108, 5341–5348 (2004) 158. Wu, Z., Li, M., Howe, J., Meyer, H.M., Overbury, S.H.: Probing defect sites on CeO2 nanocrystals with well-defined surface planes by Raman spectroscopy and O2 adsorption. Langmuir. 26, 16595–16606 (2010) 159. Weckhuysen, B.M., Wachs, I.E.: In situ Raman spectroscopy of supported chromium oxide catalysts: reactivity studies with methanol and butane. J. Phys. Chem. 100, 14437–14442 (1996) 160. Bañares, M.A., Guerrero-Pérez, M.O., Fierro, J.L.G., Cortez, G.G.: Raman spectroscopy during catalytic operations with on-line activity measurement (operando spectroscopy): a method for understanding the active centres of cations supported on porous materials. J. Mater. Chem. 12, 3337–3342 (2002) 161. Dumesic, J.A., Topsoe, N.Y., Slabiak, T., Morsing, P., Clausen, B.S., Törqvist, E., Topsoe, H.: Microiunetic analysis of the selective catalytic reduction (SCR) of nitric oxide over vanadia/titaniabased catalysts. In: Guczi, L., Solymosi, F., TÉTÉNyi, P. (eds.) Studies in Surface Science and Catalysis, vol. 75, pp. 1325–1337. Elsevier, Amsterdam (1993) 162. Bañares, M.A., Khatib, S.J.: Structure–activity relationships in alumina-supported molybdena–vanadia catalysts for propane oxidative dehydrogenation. Catal. Today. 96, 251–257 (2004) 163. Hutchings, G.J., Desmartin-Chomel, A., Olier, R., Volta, J.-C.: Role of the product in the transformation of a catalyst to its active state. Nature. 368, 41–45 (1994) 164. Cavani, F., Caldarelli, A., Luciani, S., Cortelli, C., Cruzzolin, F.: Selective oxidation of O-xylene to phthalic anhydride: from conventional catalysts and technologies toward innovative approaches. In: Catalysis: Volume 24, vol. 24, pp. 204–222. The Royal Society of Chemistry, London (2012) 165. Vogelaar, B.M., van Langeveld, A.D., Eijsbouts, S., Moulijn, J.A.: Analysis of coke deposition profiles in commercial spent hydroprocessing catalysts using Raman spectroscopy. Fuel. 86, 1122–1129 (2007) 166. Latorre, F., Kupfer, S., Bocklitz, T., Kinzel, D., Trautmann, S., Gräfe, S., Deckert, V.: Spatial resolution of tip-enhanced Raman spectroscopy – DFT assessment of the chemical effect. Nanoscale. 8, 10229–10239 (2016) 167. Urakawa, A., Bürgi, T., Baiker, A.: Sensitivity enhancement and dynamic behavior analysis by modulation excitation spectroscopy: principle and application in heterogeneous catalysis. Chem. Eng. Sci. 63, 4902–4909 (2008) 168. Müller, P., Hermans, I.: Applications of modulation excitation spectroscopy in heterogeneous catalysis. Ind. Eng. Chem. Res. 56, 1123–1136 (2017) 169. Urakawa, A., Van Beek, W., Monrabal-Capilla, M., GalánMascarós, J.R., Palin, L., Milanesio, M.: Combined, modulation enhanced X-ray powder diffraction and Raman spectroscopic study of structural transitions in the spin crossover material [Fe(Htrz)2 (Trz)](Bf4). J. Phys. Chem. C. 115, 1323–1329 (2011) 170. Nuguid, R.J.G., Ferri, D., Marberger, A., Nachtegaal, M., Kröcher, O.: Modulated excitation Raman spectroscopy of V2O5/ TiO2: mechanistic insights into the selective catalytic reduction of NO with NH3. ACS Catal. 9, 6814–6820 (2019) 171. Bañares, M.A., Cardoso, J.H., Agulló-Rueda, F., Correa-Bueno, J.M., Fierro, J.L.G.: Dynamic states of V-oxide species: reducibility and performance for methane oxidation on V2O5/SiO2 catalysts as a function of coverage. Catal. Lett. 64, 191–196 (2000) 172. Hou, X., Zhao, L., Diao, Z.: Roles of alkenes and coke formation in the deactivation of ZSM-5 zeolites during N-pentane catalytic cracking. Catal. Lett. 150, 2716 (2020). https://doi.org/10.1007/ s10562-020-03173-4 173. Wu, Z., Stair, P.C.: UV Raman spectroscopic studies of V/Θ-Al2o3 catalysts in butane dehydrogenation. J. Catal. 237, 220–229 (2006)

4

Raman Spectroscopy

174. Li, C., Stair, P.C.: Ultraviolet Raman spectroscopy characterization of sulfated zirconia catalysts: fresh, deactivated and regenerated. Catal. Lett. 36, 119–123 (1996) 175. Guerra, P., Zaker, A., Duan, P., Maag, A.R., Tompsett, G.A., Brown, A.B., Schmidt-Rohr, K., Timko, M.T.: Analysis of coke formed during zeolite-catalyzed supercritical dodecane cracking: effect of supercritical water. Appl. Catal. A Gen. 590, 117330 (2020) 176. Egashira, M., Matsuo, K., Kagawa, S., Seiyama, T.: Phase diagram of the system Bi2O3–MoO3. J. Catal. 58, 409–418 (1979) 177. Theobald, F., Laarif, A., Tachez, M.: Bismuth molybdate: the Γ′-Bi2O3  MoO3 polymorph. J. Catal. 73, 357–360 (1982) 178. Carson, D., Coudurier, G., Forissier, M., Vedrine, J.C., Laarif, A., Theobald, F.: Synergy effects in the catalytic properties of bismuth molybdates. J. Chem. Soc. Faraday Trans. 1: Phys. Chem. Condensed Phases. 79, 1921–1929 (1983) 179. Kumar, J., Ruckenstein, E.: Thermal decomposition of the 1:1 bismuth molybdate and its implications for catalytic oxidation. J. Solid State Chem. 31, 41–46 (1980) 180. Batist, P.A., van de Moesdijk, C.G.M., Matsuura, I., Schuit, G.C. A.: The catalytic oxidation of 1-butene over bismuth molybdates: promoters for the Bi2O33MoO3 catalyst. J. Catal. 20, 40–57 (1971) 181. Snyder, T.P., Hill, C.G.: Stability of bismuth molybdate catalysts at elevated temperatures in air and under reaction conditions. J. Catal. 132, 536–555 (1991) 182. Mestl, G.: In situ Raman spectroscopy for the characterization of MoVW mixed oxide catalysts. J. Raman Spectrosc. 33, 333–347 (2002) 183. Mutz, B., Sprenger, P., Wang, W., Wang, D., Kleist, W., Grunwaldt, J.-D.: Operando Raman spectroscopy on CO2 Methanation over alumina-supported Ni, Ni3Fe and NiRh0.1 catalysts: role of carbon formation as possible deactivation pathway. Appl. Catal. A Gen. 556, 160–171 (2018) 184. Stavitski, E., Kox, M.H.F., Swart, I., de Groot, F.M.F., Weckhuysen, B.M.: In situ synchrotron-based IR microspectroscopy to study catalytic reactions in zeolite crystals. Angew. Chem. 120, 3599–3603 (2008) 185. Brückner, A.: Killing three birds with one stone—simultaneous operando EPR/UV-Vis/Raman spectroscopy for monitoring catalytic reactions. Chem. Commun., 1761 (2005). https://doi.org/10. 1039/B418790C1761-1763 186. Dieterle, M., Weinberg, G., Mestl, G.: Raman spectroscopy of molybdenum oxides part I. structural characterization of oxygen defects in MoO3
X by DR UV/VIS, Raman spectroscopy and X-ray diffraction. Phys. Chem. Chem. Phys. 4, 812–821 (2002) 187. Beale, A.M., van der Eerden, A.M.J., Kervinen, K., Newton, M.A., Weckhuysen, B.M.: Adding a third dimension to operando spectroscopy: a combined UV-VIS, Raman and XAFS setup to study heterogeneous catalysts under working conditions. Chem. Commun., 3015 (2005). https://doi.org/10.1039/ B504027B3015-3017 188. van Beek, W., Safonova, O.V., Wiker, G., Emerich, H.: SNBL, a dedicated beamline for combined in situ X-ray diffraction, X-ray absorption and Raman scattering experiments. Phase Transit. 84, 726–732 (2011) 189. Cats, K.H., Weckhuysen, B.M.: Combined operando X-ray diffraction/Raman spectroscopy of catalytic solids in the laboratory: the Co/Tio2 Fischer–Tropsch synthesis catalyst showcase. ChemCatChem. 8, 1531–1542 (2016) 190. Boccaleri, E., Carniato, F., Croce, G., Viterbo, D., van Beek, W., Emerich, H., Milanesio, M.: In situ simultaneous Raman/highresolution X-ray powder diffraction study of transformations occurring in materials at non-ambient conditions. J. Appl. Crystallogr. 40, 684–693 (2007)

109 191. Tardif, S., Pavlenko, E., Quazuguel, L., Boniface, M., Maréchal, M., Micha, J.-S., Gonon, L., Mareau, V., Gebel, G., BayleGuillemaud, P., Rieutord, F., Lyonnard, S.: Operando Raman spectroscopy and synchrotron X-ray diffraction of lithiation/delithiation in silicon nanoparticle anodes. ACS Nano. 11, 11306–11316 (2017) 192. Yao, H., Cheng, H., Gao, X.J., Li, C., Cui, R., Huang, H., Guo, X., Li, X., Zhang, L., Liu, B., Xu, B., Dong, J., Gao, X., Li, Y., Sun, B.: In situ synchrotron X-ray diffraction and Raman spectroscopy studies of Gd@C82–S8 under high pressure. J. Phys. Chem. C. 122, 10992–10998 (2018) 193. Jiang, J., Wu, X., Li, D., Ma, B., Liu, R., Wang, X., Zhang, J., Zhu, H., Cui, Q.: High pressure Raman scattering and synchrotron X-ray diffraction studies of benzyl azide. J. Phys. Chem. B. 119, 513–518 (2015) 194. Briois, V., Vantelon, D., Villain, F., Couzinet, B., Flank, A.-M., Lagarde, P.: Combining two structural techniques on the micrometer scale: micro-XAS and micro-Raman spectroscopy. J. Synchrotron Radiat. 14, 403–408 (2007) 195. Rastogi, S., Goossens, J.G.P., Lemstra, P.J.: An in situ SAXS/ WAXS/Raman spectroscopy study on the phase behavior of syndiotactic polystyrene (SPS)/solvent systems: compound formation and solvent (dis)ordering. Macromolecules. 31, 2983–2998 (1998) 196. Martel, A., Burghammer, M., Davies, R.J., Di Cola, E., Vendrely, C., Riekel, C.: Silk fiber assembly studied by synchrotron radiation SAXS/WAXS and Raman spectroscopy. J. Am. Chem. Soc. 130, 17070–17074 (2008)

Jisue Moon received his PhD from the University of California, Irvine, in 2018. Between 2018 and 2019, she was a postdoc at Surface Chemistry and Catalysis Group, Oak Ridge National Laboratory. She is currently a staff of Isotope Applications Research Group at ORNL. Her research interests lie in the mechanism studies of catalysts by neutron scattering and in situ spectroscopy.

Meijun Li received her PhD from Dalian Institute of Chemical Physics in 2002. She worked at Northwestern University and at the University of Tennessee. She is currently a staff of Chemical Process Scale Up Group in Manufacturing Science Division at Oak Ridge National Laboratory. Her research focus includes heterogeneous catalysis, nanomaterials, carbon capture, and in situ/operando spectroscopy.

4

110

Anibal J. Ramirez-Cuesta studied Physics in Argentina and joined the group of Dr. Philip C.H. Mitchell at the University of Reading, UK. Later, he joined Rutherford Appleton Laboratory as a senior instrument scientist at the TOSCA spectrometer. He joined Oak Ridge National Laboratory in 2013 as a leader of chemical spectroscopy group. He is currently the leader of Neutron Instrument Development Group. His research interests include neutron scattering, computer modelling, and surface science.

J. Moon et al.

Zili Wu received his PhD from Dalian Institute of Chemical Physics in 2001. He worked as a postdoc at the Catalysis Center at Northwestern University before he joined Oak Ridge National Laboratory in 2006. He is currently the leader of Surface Chemistry and Catalysis Group with research focus on heterogeneous catalysis, nanomaterials, in situ/operando spectroscopy, neutron scattering, and reaction mechanisms.

5

Case Studies: Raman Spectroscopy Ragamaye Tigiripalli

, Vishal Agarwal

, and Goutam Deo

Contents 5.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

5.2

In Situ Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

5.3

Case Studies on the Application of Operando Raman Spectroscopy to Heterogeneous Catalysts . . . . . . . . . . . . . . 113 5.3.1 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 5.4

Present Challenges and Future Recourse . . . . . . . . . . . . . . . 124

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

Abstract

The application of Raman spectroscopy to heterogeneous catalysts has undergone several changes over the years. Based on the information collected during characterization of the catalysts, these changes can be categorized into the three generations of applications. In the firstgeneration of applications, information from Raman spectroscopy is used to characterize the catalysts when the environment surrounding the catalyst is not well controlled, for example, ambient conditions. This is also referred to as ex situ characterization. The secondgeneration applications are those where the vibrational information from the catalysts under controlled environments or in situ studies is examined, and the nature of the solid catalyst and gas-solid catalyst interactions are deciphered. Finally, the third-generation applications involve the simultaneous use of in situ Raman spectroscopy and reaction data to directly correlate the structure and reactivity information of the working heterogeneous catalysts. Such studies are referred to as operando Raman spectroscopy. More recently, Raman spectroscopy has been coupled with other characterization techniques to obtain additional information simultaneously, which are R. Tigiripalli · V. Agarwal (*) · G. Deo (*) Department of Chemical Engineering, Indian Institute of Technology Kanpur, Kanpur, India e-mail: [email protected]; [email protected]; [email protected]

then correlated with the reactivity data. Furthermore, computational techniques coupled with operando Raman and other techniques are becoming in vogue for a more comprehensive understanding of the catalytic system. In this chapter, we primarily take up two case studies. In one, we highlight the three generations of application of Raman spectroscopy, and in the other, we show the importance of using density functional theory coupled with Raman spectroscopy to identify the molecular nature of the catalyst. The proper applications of Raman spectroscopy to heterogeneous catalysis is faced with several challenges. However, the future looks bright, and it is expected that we will continue to see a significant increase in research in this exciting area. Keywords

Raman · In situ · Operando · DFT

5.1

Introduction

During the process of obtaining the Raman spectra of the catalyst, a beam of monochromatic photons, usually in the visible light region, is focused on the catalyst surface and the scattered photons are collected, cleaned, and then processed. The photons from the incident beam are scattered with or without a change in the frequency. The phenomena of scattering with the change in the incident frequency is called inelastic scattering. Raman spectroscopy is based on inelastic scattering of light, which is extensively used by experimentalists to obtain vibrational information about the catalyst. From the vibrational information, the structure of the vibrating species is deduced. The Raman phenomena [1] is named after its discoverer C.V. Raman, an Indian physicist who received the Nobel prize in 1930 for this remarkable discovery.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_5

111

112

R. Tigiripalli et al.

The Raman spectra provides information about the vibrational and rotational excitations, and their presence or absence are given by certain selection rules. The selection rules would depend on the polarizability of the vibrating entities and are discussed in ▶ Chap. 8. Different types of detectors, filters, and other excitation sources have also been used and improved upon over the years. These have also been elaborated in ▶ Chap. 8. Information from the molecular vibration of species is of primary importance in the application of Raman spectroscopy to heterogeneous catalysts. For a gas-solid system, the Raman spectra can provide information about the nature of solid catalyst and the adsorbed/reacting species. Raman spectroscopy has also been applied to liquid-solid systems [2, 3] and more recently to a gas-liquid-solid system [4]. Furthermore, “unconventional” but related resonance Raman techniques, such as surface-enhanced, tip-enhanced Raman spectroscopy, and other complex Raman techniques have also been used to study heterogeneous catalysts [5–7]. However, such “unconventional” Raman spectroscopy is not dealt with in this Chapter, which is restricted to the advances of “conventional” Raman spectroscopy applications to heterogeneous catalysts. In this chapter, we begin by a section on ex situ, in situ, and operando Raman spectroscopy. We then proceed to provide two case studies where the application of Raman spectroscopy is highlighted, and the improvement of knowledge about the catalytic system is discussed. Finally, we conclude the chapter by providing the present challenges and future recourse that we envision in the application of Raman spectroscopy to heterogeneous catalysis.

5.2

In Situ Raman Spectroscopy

During the initial days, Raman spectroscopy was applied to a few heterogeneous catalysts so as to obtain information about vibrating species in a solid catalyst, which were difficult to detect with commonly available techniques at those times, such as X-ray diffraction. Sometimes these vibrating species are the active site or form part of the catalytic active sites. With the recognition that Raman spectroscopy is well suited for carrying out catalyst characterization under controlled environments, several in situ characterization studies of the catalysts were then carried out [8]. With the use of advanced laser sources and detector, a lot of emphasis was placed on the development of proper in situ cells that enabled the collection of Raman spectra under controlled environments. These controlled environments were chosen such that the Raman spectra at different partial pressures of reactants and at various temperatures could be analyzed. There are several types of in situ cells that have been used to obtain the Raman spectra of a number of

heterogeneous catalysts under controlled environments [9, 10]. One of the earliest in situ cells used for Raman spectroscopy to study adsorbed species was developed in 1975 by Schrader and Hill [11]. The Raman cell was a batch reactor, with the ability to vary the temperature up to 450
C. Such type of characterization can be classified as the second-generation application of Raman spectroscopy to heterogeneous catalysts. The first-generation applications are those where Raman spectroscopy was applied for catalyst characterization under ambient conditions or under conditions where the exposure to the catalyst was not well defined. The second-generation characterization studies can be further classified into “in situ spectra,” “variable programmed in situ spectra,” and “reaction in situ spectra” [12]. Furthermore, it was known that to obtain meaningful in situ Raman spectra, showing the changes in the catalysts during reaction, it was necessary to supplement the spectral information with online reactivity data. Thus, the information about the vibrating species giving rise to the Raman spectra and their reactivity could be simultaneously obtained. This simultaneous collection of information from structure and reactivity was coined as operando spectroscopy [8, 13–15], which can be referred to as the third-generation application of Raman spectroscopy to heterogeneous catalysts. Without collection of online reactivity information, we still categorize these cases belonging to the second-generation characterization of heterogeneous catalysts as defined above. It must be mentioned here that the three generations of applications do not reflect the chronological development of the application of Raman spectroscopy, since it was recognized early on that characterization under appropriate conditions was important for the understanding of the catalytic science. A schematic representation of the three generations of application of Raman spectroscopy to heterogeneous catalysts is given in Fig. 5.1. It is worth noting that prior to the coinage of the word “operando,” there were some studies that have taken into consideration the simultaneous collection of spectral and reactivity information. For example, by using surfaceenhanced Raman spectroscopy with online mass spectroscopy for an electrochemically deposited Rh catalyst, the oxidation of CO by NO was studied at ambient pressures and temperatures up to 623 K [6]. Based on the Raman spectra of the adsorbed species and reaction intermediates and changes in the gas-phase composition, Tolia et al. proposed that the reaction of CO with NO proceeded via a dissociative mechanism under reaction conditions. The historical development of operando spectroscopy in general, and operando Raman in particular, has been reported by Wachs and co-workers [16]. Wachs and co-workers have elucidated how operando spectroscopy has successfully been used to improve our understanding of the catalytic phenomena. Several applications of operando Raman and other spectroscopies to heterogeneously catalyzed systems have been

5

Case Studies: Raman Spectroscopy

113

a)

b) Laser source

Laser source

Raman detector

Raman detector

Vent gases

Reactant gases In situ

Ex situ

5 c) Laser source

Raman detector

GC/MS/IR to detect outlet composition

Reactant gases In situ

Fig. 5.1 The uses of Raman spectroscopy in heterogeneous catalysis can be classified into three generation of applications. This classification does not reflect the chronological development of the applications of

Raman spectroscopy to heterogeneous catalysis but is based on how the spectral information was obtained. (a) First generation application, (b) Second generation application, (c) Third generation application

discussed. In what follows, we discuss a few case studies showing the applications of Raman spectroscopy to heterogenous catalysis.

“working catalyst” was postulated from the Raman spectra, while the reactivity information was simultaneously provided by online gas chromatography (GC) analysis. This study systematically carried out the three generations of application of Raman spectroscopy to improve the understanding of an important catalytic system. Other methods of online analysis of reactants and products are also possible, such as the use of mass spectrometry and/or a Fourier-transform infrared (FTIR) gas cell. Before we start our case studies on the application of Raman spectroscopy to heterogeneous catalysts, it is worthwhile to mention some of the requirements of an ideal operando system. Preferably, the relationship between the structure and reactivity can be obtained for well-defined and industrially relevant systems. What is crucial for the implementation of operando spectroscopy is the proper design of the reactor. The reactor should be a properly defined system, with minimum dead volume and bypass, and should facilitate capturing spectroscopic information [9, 10, 17]. Due consideration for possible mass and heat transfer effects is required. For example, proper temperature estimates in an operando cell is vital since the temperature and time evolution of reactions are required in the study of catalytic mechanisms

5.3

Case Studies on the Application of Operando Raman Spectroscopy to Heterogeneous Catalysts

As mentioned above, operando Raman spectroscopy is the characterization of a working catalyst with simultaneous online analysis of reactants and products. By analyzing the Raman spectrum, the structure arising due to the interaction of the solid-fluid interaction can be inferred. Furthermore, from the online analysis of the reactants and products, the important reactivity parameters, such as conversion, yield, and selectivity, can be determined. Thus, the primary goal of operando spectroscopy, which is to examine the relation between the structure and the reactivity for a catalyst during operation, can be achieved. For example, an operando Raman-GC study was carried out over a nanocrystalline vanadium-antimonate catalyst for the propane ammoxidation to acrylonitrile [13], where the structural information of a

114

R. Tigiripalli et al.

and determination of apparent activation energies. Furthermore, the reactor should be able to operate under temperatures, partial pressures, and contact times of the reactants that are relevant to industrial applications. Using computational fluid dynamics (CFD), it may be possible to analyze the effect of operating variables for a specific reaction and the temporal and spatial changes that may occur. Mass and heat transfer effects are inherently addressed. Alternatively, an isothermal reactor with a specially designed optical fiber to provide the excitation source and collection of the scattered light to obtain the Raman spectra would be well suited for operando studies [18]. Such a setup has been attempted in the simultaneous collection of Raman spectra and online reactivity measurements [19]. Furthermore, the spectral requirements of the operando Raman study should be such that they are [20]: (a) Time resolved: The Raman spectra and reactivity information should be measured at the timescales of breaking and making of the chemical bonds. However, only timescales of microseconds to several seconds are possible. (b) Spatially resolved: The Raman spectra should be able to identify different reaction species and distinguish the difference between different active sites and spectator species. The catalytic phenomena usually occur at nanometer scales in heterogeneous catalysts, and the spatial resolution ranging from micrometers to nanometers will be able to assist in distinguishing different reaction species.

5.3.1

Case Studies

Operando Raman spectroscopy has been applied to various heterogeneous catalysts, which include (i) supported and unsupported metal oxides, (ii) mesoporous silica and zeolites, (iii) polyoxometalates, (iv) sulfides, (v) adsorbed molecules, (vi) reaction intermediates, and (v) carbon formation during reactions. Previous compilation of studies dealing with operando Raman spectroscopy has been carried out till 2017 [16]. We examine two classes of catalysts where the application of Raman spectroscopy has significantly improved the general understanding about the heterogeneous catalyst. These two case studies are meant to be representative since the number of research publications dealing with the application of Raman spectroscopy and the respective impact may be varying.

Case Study 1: The Three Generations of Characterization of Zirconia-Supported Vanadium Oxide Catalysts Supported metal oxide catalysts are an important class of catalysts that find numerous applications in the petrochemical and environmental industry [21–25]. In this class of catalyst,

a metal oxide phase, usually in the form of molecularly dispersed species, is deposited on the surface of a robust and high-surface-area carrier or support [26]. The supported metal oxide phase is often a transition metal oxide, such as vanadium oxide, molybdenum oxide, tungsten oxide, or chromium oxide. The robust supports typically considered are silica (SiO2), alumina (Al2O3), titania (TiO2), zirconia (ZrO2), ceria (CeO2), and others. We begin with this class of catalytic systems since we can show the progressive improvement of knowledge over the three generations of application of Raman spectroscopy. Most transitional metal oxides are amenable to being detected and analyzed by Raman spectroscopy, shown by the studies by Wachs, Banares, and co-workers [15]. These studies have provided conclusive evidence about the vibrational structure of the supported metal oxide phase under different environments. Furthermore, the authors have shown that these systems are ideally suited for in situ and operando studies. The presence of molecularly dispersed species below monolayer coverage and the formation of bulk metal oxide species (e.g., V2O5, MoO3, WO3, or Cr2O3) above monolayer coverage were established from the ambient Raman spectra [21] or the first-generation application as classified above. For example, the monolayer coverage of vanadium oxide supported on zirconia support was determined to be about 8 V atoms/nm2 [27]. Furthermore, the first generation of Raman characterization also established that the supported metal oxide phase possessed a structure that was largely determined by the net pH at the point-of-zero charge of the supported metal oxide catalyst [28]. These structures of the supported metal oxide phase closely resembled those present in aqueous solutions of various pHs and concentration of the metal oxide phase in the solution. However, upon heating the surface, moisture was removed, and the structure of the surface metal oxide phase changed as inferred from the Raman spectra of the samples [29]. Supported vanadium oxide catalysts form an important subclass of the supported metal oxide catalysts that find specific use in the chemical industry [25]. With a carefully planned experiment using in situ Raman spectroscopy, Jehng et al. showed that the dehydrated structure of a supported vanadium oxide catalyst possessing molecularly dispersed metal oxide species was different from the structure present under ambient environments [30]. At temperatures of about 230
C and above the interaction of gas-phase moisture with the vibrational features of the surface, vanadium oxide species was minimal. Upon exposure of the supported vanadium oxide catalysts to humid air at room temperatures, the spectrum reverted to the one obtained under ambient conditions. An experimental and theoretical study showing the effect of hydration of supported vanadium oxide catalysts has also been performed [31]. We show a simplified schematic of this shuttling between two states, the hydrated state and the dehydrated state, in Fig. 5.2.

5

Case Studies: Raman Spectroscopy

[V2O7]–4

[VO4]–3 [V10O28]–6 Oxide support

115

[V3O9]–3

O

O

O

O

V

V

V

V

O

OH O O

+ nH2O OH V O

x y

OH

VxOy OH V O OH x y

– nH2O

OH

O O

O

O

O O

O

O

O

V

V O OH

O O

O

O OH

Oxide support VxOy

VxOy OH V O OH x y OH OH

Oxide support

Fig. 5.2 The structure of the molecularly dispersed species shuttling between the ambient and dehydrated structures depending on the presence or absence of moisture. Typical ambient structures, [VO4]3, [V2O7]4, [V3O9]3, and [V10O28]6, are shown to be present in the film of water. The dehydrated structures are shown as VxOy without detail. Representative dehydrated structures are shown below. The reversibility of the ambient and dehydrated vanadium oxide structures has also been used to confirm the surface nature of the molecularly dispersed species

Under ambient conditions, the molecularly dispersed species was shown to be present in several layers of moisture, and under dehydrated conditions, the molecularly dispersed species anchor themselves to the support. It was also shown that the shuttling between these two classes of structures was common to several supported metal oxide catalysts [29]. The effect of moisture was also used to confirm the surface nature of the supported metal oxide phases. Furthermore, bulk metal oxide species, if present, were found to be not affected by the presence or absence of moisture. The presence of bulk metal oxide species indicates that the monolayer coverage had been exceeded or the surface metal oxide species were not well dispersed, perhaps due to the ill-synthesized catalysts. Thus, the initial second-generation application of Raman spectroscopy revealed that the structure of the molecularly dispersed vanadium oxide phase reversibly shuttled between these two environments. Based on the various vibrating groups detected and using multiple characterization techniques [25], supplemented with 16 O2-18O2 exchange experiments [32], some representative surface vanadium oxide structures under dehydrated conditions were proposed using information gathered from secondgeneration characterization of Raman spectroscopy of supported vanadium oxide catalysts. Representative structures are shown in Fig. 5.3. Supported vanadium oxide catalysts have been extensively studied for the oxidative dehydrogenation of propane to propylene [33, 34]. Propylene is a key petrochemical intermediate that can be used to produce a variety of chemicals, which include polypropylene, propylene oxide, acrylonitrile, cumene, and others [35, 36]. A relatively recent critical review of the propane oxidative dehydrogenation (P-ODH) reaction has been reported, where a significant

Fig. 5.3 Representative surface vanadium oxide structures present on oxide supports under dehydrated conditions. The three structures shown are as follows: one isolated and two polymeric. Hydroxyl (-OH) species on the surface of the support are also indicated. The surface vanadium oxide species are primarily present in the +5 oxidation state and participate in the Mars-van Krevelen reaction mechanism. Above monolayer coverage, microcrystalline V2O5 are also present. There are three different types of oxygen species associated with the surface vanadium oxide species: one terminal (V¼O) and two bridging (V-O-V or V-O-S, where S is the support cation)

number of results from different research groups have been compared in terms of their turnover frequency and apparent activation energy [34]. Other comparisons have also been provided. The P-ODH reaction has been tested with vanadium oxide supported on silica (SiO2), alumina (Al2O3), titania (TiO2), and zirconia (ZrO2). The turnover frequency over these supported vanadium oxide catalysts (V/support) revealed that V/ZrO2 ~ V/TiO2 > V/Al2O3 > V/SiO2. Thus, vanadium oxide supported on ZrO2 is an active catalyst for the P-ODH reaction, and the V/ZrO2 catalyst is a worthy candidate to track over the three generations of applications of Raman spectroscopy to heterogeneous catalysts. Previous in situ Raman studies on V/ZrO2 catalysts showed that the oxygen-18 exchanged surface vanadium oxide species anchored to the zirconia support possessed a Raman band at about 990 cm1 under dehydrated conditions. In contrast, the V/ZrO2 catalyst with 16O2 possessed a band at about 1030 cm1 [30]. Unfortunately, analysis of the vibrations below 800 cm1 was not possible due to the dominant Raman bands of the ZrO2 support. The Raman band positions for the vanadium oxide species with 18O2 and for the vanadium oxide species with oxygen-16 clearly indicate the presence of a single V¼O terminal bond and confirm the nature of the vanadium species proposed above in Fig. 5.3. In situ Raman characterization of a V/ZrO2 catalyst, in slight excess of a monolayer loading, in the presence of O2 þ He or a mixture of C3H8 þ O2 þ He, with different partial pressures of the reactants, provided results that were quite revealing [37]. Such a study also qualifies for our classification of second-generation application of Raman spectroscopy since the spectra were obtained in a controlled environment and the simultaneous product analysis was not carried out. As expected from the discussion above, molecularly dispersed surface vanadium oxide species and bulk V2O5 are detected under dehydrated conditions based on the Raman bands observed. The presence of bulk V2O5 was evident from the

5

116

R. Tigiripalli et al.

presence of typical sharp Raman bands at 994, 700, 402, 280, and 146 cm1. The molecularly dispersed vanadium oxide species showed bands at ~1034 cm1 and ~930 cm1 due to the V¼O and V-O-M vibrating bonds, respectively. In the V-O-M bond, M is either another V atom or Zr atom of the support [32, 38–41]. One may refer to Fig. 5.3, where the three different types of oxygen species are illustrated. The molecularly dispersed vanadium oxide species was suggested to be of two kinds: an isolated vanadium oxide species possessing a single terminal V¼O bond and three V-O-Zr bonds, and a polymeric vanadium oxide species possessing a terminal V¼O bonds, V-O-V, and V-OZr bridging bonds. The terminal V¼O bond gave rise to a Raman species at ~1030 cm1 for the isolated species (seen at ~1034 cm1 by Gao et al. [37]) and at ~1020 cm1 for the polymeric species [42]. Gao et al. [37] further claimed that the V¼O terminal bond due to the polymeric species was not readily detected due to the relatively intense bands at ~1034 and 994 cm1. To account for the V¼O terminal bond due to the polymeric species the authors deconvoluted the Raman spectra obtained in the 800–1035 cm1 region into four Raman bands, as shown in Fig. 5.4. The four Raman bands, at ~935, ~994,

a) PO2 = 0.2 994

1020

935

930

c)

1016

b) PO2 = 0.2, PC3H8 = 0.04 1029

1016

Raman intensity (arbitrary units)

1034

1029

PO2 = 0.02, PC3H8 = 0.18 930

850

900

950

Wavenumber

1000

1050

(cm–1)

Fig. 5.4 The deconvoluted in situ Raman spectra obtained at different partial pressures of the reactants for the V/ZrO2 catalyst at 300
C used by Gao et al. [37] are shown. The partial pressure of the reactants, PO2 and PC3H8, are shown in the figure along with the positions of the Raman bands due to the different vanadium oxide species. (Reprinted from Gao et al. [37], Copyright © (2002), with permission from Elsevier)

~1020, and ~1034 cm1, were tracked in the deconvoluted spectra obtained under different conditions. The change in the intensity of the Raman bands at ~935, ~994, ~1020, and ~1034 cm1 of the deconvoluted spectra was tabulated as in Table 5.1. The decrease in the intensity of the in situ Raman bands at ~935 and ~1020 cm1 bands appeared to be similar. These two Raman bands were assigned to the polymeric surface vanadium oxide species. The decrease in intensity of the in situ Raman bands due to the polymeric species were more than the decrease observed for the isolated surface vanadium oxide species, which possessed a terminal V¼O Raman band at ~1034 cm1. Furthermore, the isolated surface species and microcrystalline V2O5 appeared to reduce to a similar extent. Thus, the polymeric species was suggested to be more reduced than the isolated species under steady-state reaction. Information from the characterization of the vanadium oxide species in supported catalysts from second-generation studies, such as that done by Gao et al. [37] for the zirconiasupported vanadium oxide catalysts under dehydrated conditions, and in the presence of reactant molecules provides an excellent platform for the understanding of the changes these structures undergo during catalytic reactions using the thirdgeneration operando Raman studies. For example, the reduction of the molecularly dispersed vanadium oxide species on an alumina support was pursued with a “MultiTRACK” Operando Raman system by Banares and co-workers [43]. Using this “Multiple Transient Analysis of Catalytic Kinetics” approach, the authors were able to reveal the participation of lattice oxygen during the formation of propene which followed the Mars-van Krevelen mechanism [44]. Very recently, a multicharacterization operando study has been undertaken over a V/ZrO2 catalyst for the P-ODH reaction [45]. In this study, transmission FTIR and Raman spectroscopy was carried out during the P-ODH reaction in the temperature range of 200–400
C, and the exhaust gases were analyzed by an IR gas cell and a mass spectrometer (MS). In situ and operando FTIR spectroscopy have also been used successfully for the characterization of heterogeneous catalysts [46]. Like Raman, IR spectroscopy also looks at the molecular vibrations of the species, and it is often used as a complimentary technique for the characterization of supported metal-oxide catalysts [46]. The selection rules of IR and Raman are different, and FTIR spectroscopy has been often used to identify the adsorption of relevant gas-phase species on catalyst surfaces during P-ODH [47, 48]. Thus, the combination of FTIR and Raman spectra used in the above study provides simultaneous information about the supported catalyst and the adsorbed species. In the abovementioned study, the transmission FTIR and Raman spectra were collected over three V/ZrO2 catalysts under ambient and in situ dehydrated conditions and during operando mode of the P-ODH reaction. The tetragonal phase

5

Case Studies: Raman Spectroscopy

117

Table 5.1 Change in the in situ Raman band intensity for the V/ZrO2 sample in O2 þ He and different mixtures of C3H8 þ O2 þ He at a temperature of 300
C [37] In situ environment in partial pressures of the gas phase PC3H8 PHe PO2 0.20 0.00 0.80 0.20 0.04 0.76 0.02 0.18 0.80

of ZrO2 was identified by Raman spectroscopy and was used to normalize the intensity of the Raman spectra. However, the Raman bands were uncharacteristically broad, perhaps due to the method of preparation of the V/ZrO2 catalyst [49]. The FTIR spectra revealed the presence of adsorbed moisture under ambient conditions, as expected from characterization under ambient conditions, and the intensity of the adsorbed moisture bands decreased during heating to P-ODH reaction conditions of 340
C. Upon heating the catalyst wafer up to 340
C in an oxidized atmosphere, IR and Raman bands due to the surface vanadium oxide species were detected at about 1020 cm1. Additional Raman bands were detected from 700 to 900 cm1. However, the IR bands in this region (lesser than 900 cm1) were dominated by the strong absorption of the zirconia support. The overlapping IR and Raman at ~1020 cm1 is consistent with the presence of a single terminal V¼O bond of the surface vanadium oxide species, which has been previously proposed by matching the independently obtained Raman and IR spectra of the same sample [42, 47]. With increase in vanadium oxide loading in the V/ZrO2 samples, the V¼O gradually blueshifts in the operando FTIR and Raman spectra. The relatively broad Raman bands at ~880 and 770 cm1, due to the V-O-V and V-O-Zr vibrating bonds, also become more intense with an increase in vanadium oxide loading. FTIR peaks due to occluded CO2 or carbonaceous material were also detected. To detect changes in the IR and Raman bands during reaction, operando spectroscopic characterization experiments were then carried out. The reactivity data revealed that the conversions and propene yield increase with vanadium oxide content, and this trend was like those achieved in a fixed bed micro-reactor. The reactivity data was then correlated with the IR and Raman operando spectra. To improve the understanding of the reactivity of the V/ZrO2 samples, the spectra were collected using the following gas mixtures on the three catalysts, referred to as conditions 1, 2, and 3. 1. Flow of 10% O2 þ Ar at 340
C. This was referred to as “activation” flow. These spectra were collected before and after reaction and were similar to the dehydrated conditions mentioned above. Activation after reaction was important to detect if any irreversible deactivation takes place.

Relative band intensity of the Raman bands at ~935 cm1 994 cm1 ~1020 cm1 1.00 1.00 1.00 0.71 0.77 0.69 0.57 0.64 0.60

~1430

~1034 cm1 1.00 0.81 0.65

~1600 340 °C 300 °C

Zr-V6.4

340 °C 300 °C

5

Zr-V5.0 20% C3H8 + He

Zr-V2.5

340 °C 300 °C

1300

1400 1500 Raman shift (cm–1)

1600

Fig. 5.5 Raman bands showing carbon formation (peaks at ~1430 and ~1600 cm1) for the 5 V/ZrO2 and 2.5 V/ZrO2 samples during P-DH reaction at 340
C. In contrast, no formation is seen at 300
C and for the 6.4 V/ZrO2 sample. (Reprinted with permission from Ternero-Hidalgo et al. [45]. Copyright © 2020 American Chemical Society)

2. Flow of 20% C3H8 þ 10% O2 þ Ar at 340
C. These were the conditions for the P-ODH reaction and corresponded to operando conditions. Some runs were carried out with the reactant gases at lower temperatures, where the conversions were not significant. 3. Flow of 20% C3H8 þ He at 340
C. These were the conditions for propane dehydrogenation (P-DH) reaction. Activation flow of the catalysts in O2 þ Ar, condition 1, was necessary at the start of each experiment. During the flow of 20% C3H8 þ He, condition 3 above, the Raman band due to the surface vanadium oxide species was absent. However, the intensity of the IR bands of the V¼O species somewhat decreased. Simultaneously, the growth of carbon was detected in the Raman spectra as peaks at 1430 and 1600 cm1 for two V/ZrO2 catalysts, as shown in Fig. 5.5 and as a shoulder at 1600 cm1 in the FTIR spectra shown

118

R. Tigiripalli et al.

later. Interestingly, peaks due to carbon were not detected in the V/ZrO2 sample containing the high vanadium oxide loading. The absence of the Raman bands due to the surface vanadium oxide species may be related to the dark coloration of the sample due to carbon being deposited on the catalyst surface, and the authors aim to address this observation in the future. Furthermore, catalyst deactivation was evident from the evolution of the P-DH reaction products analyzed by IR spectroscopy. The evolution of the reaction products revealed that propylene and carbon oxides (CO and CO2) were initially formed and their concentration decreased with time. The decrease in these products with time was attributed

to the occurrence of the Mars-van Krevelen mechanism [44], which required the participation of lattice oxygen. Once the lattice oxygen was depleted, the reaction ceased to occur. The FTIR spectra during P-DH revealed that the presence of acetates, with bands at 1540, 1448, and 1355 cm1, dominate the spectra. The formation of formate was excluded since lattice oxygen was not required for the conversion of formate to CO. During the flow of the P-ODH reaction mixture, as given in condition 2 above, the IR and Raman bands due to the surface vanadium oxide species are detected as shown in Fig. 5.6. A small shift and a slight decrease in intensity of IR band of V¼O was detected for the three catalysts.

Color coding of spectra is done as follows: Activation flow: 10% O2 + Ar at 340 °C, condition 1 P-ODH flow: 20% C3H8 + 10% O2 + Ar balance at 340 °C, condition 2 P-DH flow: 20% C3H8 + He at 340 °C, condition 3 Activation flow after reaction

a)

b)

V-O-Zr

V=O V-O-V 2.5V/ZrO2

V=O

V=O

V-O-Zr

V-O-V

5V/ZrO2

V=O

V=O V-O-Zr V-O-V

6.4V/ZrO2

1100

1000 Wavenumbers (cm–1)

900

Fig. 5.6 FTIR and Raman spectra of the three V/ZrO2 catalysts, 6.4 V/ ZrO2, 5 V/ZrO2, and 2.5 V/ZrO2, during the P-ODH and P-DH reaction conditions showing the vanadium-oxygen vibrations. The panel (a) shows the FTIR spectra and the panel (b) the corresponding Raman

V=O

1000

900 800 Raman shift (cm–1)

700

600

spectra under different flow conditions as mentioned. The spectra have been color coded as per the corresponding flow conditions in the operando reactor. (Reprinted with permission from Ternero-Hidalgo et al. [45]. Copyright © 2020 American Chemical Society)

5

Case Studies: Raman Spectroscopy

119

However, no change in the Raman band intensity of the V¼O band at ~1025 cm1 was mentioned. This is somewhat in contrast to the study of Gao et al. [Ref. 39], where a decrease in the Raman intensity of the V¼O was shown. This minor discrepancy may be related to the different temperatures used in both the studies. Furthermore, the Raman intensity of the V-O-V and V-O-Zr bonds significantly decreased. The original Raman spectra of the surface vanadium oxide species was completely restored upon oxidation using activation flow, condition 1. It is evident that the V¼O, V-O-V, and V-O-Zr

a)

bonds were participating in the P-ODH reaction. Moreover, the significant change in the V-O-V and V-O-Zr bonds relative to the V¼O bond was attributed to the higher reactivity of the bridging oxygen species by the authors. In the FTIR spectra of the three catalysts, bands were detected in the 1100–1750 cm1 region. Specifically, IR bands due to oxygenated species were detected, as shown in Fig. 5.7. At lower reaction temperatures (200
C), where the catalysts are essentially inactive, IR bands due to formates dominate the spectra for the 5 V/ZrO2 and 6.4 V/ZrO2

b)

Reaction temperature = 200 °C

1540 1570 1650

5

Reaction temperature = 340 °C

1520

1448 1371

1600

1355

2.5V/ZrO2

1570

1540 1371

1355

1448 1350

1650

5V/ZrO2

6.4V/ZrO2

1600

1400 Wavenumbers

1200

(cm–1)

Fig. 5.7 FTIR spectra showing the carbon-related vibrations for the three V/ZrO2 during P-ODH and P-DH reactions conditions at the following: (a) 200
C and (b) 340
C. The acetate bands are at 1540, 1448, and 1355 cm1; formate bands are at 1570, 1371, and 1350 cm1;

1600

1400 Wavenumbers

1200

(cm–1)

the acrylate bands are at 1650, 1540, 1440, and 1355 cm1; and the carbon band is at 1600 cm1. Each spectrum is color coded as in Fig. 5.6. (Reprinted with permission from Ternero-Hidalgo et al. [45]. Copyright © 2020 American Chemical Society)

120

samples in the 1100 to 1700 cm1 region, as shown in Fig. 5.7a. Thus, it was proposed that the formate species do not have a major role to play in the P-ODH reaction. Under P-ODH reaction conditions, IR bands due to bidentate acetates, acrylates, and bidentate formates were observed, as shown in Fig. 5.7b. These oxygenated species were completely removed in the presence of 20% O2 þ Ar. Furthermore, the authors proposed that these oxygenated species were responsible for the formation of carbon oxides and water since these total oxidation products were detected when the reactant mixture was switched to a 20% O2 þ Ar flow. Furthermore, the proportion of acetates and acrylates increased relative to the formate species when the vanadium oxide loading was increased. It was suggested that the acetates and acrylates were preferentially formed on the polymerized vanadium oxide species. No vibrations due to carbon was detected during the P-ODH reaction, indicating that carbon deposition that may occur during P-DH can be removed during P-ODH. In summary, the application of the three generations of Raman spectroscopy to propane oxidative dehydrogenation over zirconia-supported vanadium oxide catalyst has been successful. However, there are several uncertainties that still need to be addressed, such as the effect of temperature, contact time (conversion), and partial pressure of the reactant gases. Other recent and interesting applications of the use of operando Raman spectroscopy for supported metal-oxide catalysts have also been carried out. For example, using a “spectrokinetic” approach, Carrero and co-workers attempted to determine the redox rates of binary V/SiO2, Nb/SiO2, and ternary V/Nb/SiO2 catalysts and evaluate the effect of niobium (Nb) [50]. By combining operando Raman and transient reaction kinetics, the redox rates of the catalysts were estimated in real time. Furthermore, based on their approach, the site-specific rate (based on changes in the V¼O bond intensity) and overall rate were measured. Future analysis of using other reactant molecules and different vanadium oxygen vibrations is awaited. Supported vanadium oxide catalysts were also analyzed by operando Raman spectroscopy during the selective catalytic reduction (SCR) of NO using NH3 [51]. The SCR reaction, 4NO þ 4NH3 þ O2 ¼ 4 N2 þ 6H2O, was carried out over a V/TiO2 catalyst and followed by operando Raman spectroscopy. Previous studies revealed that the ambient and dehydrated structures of the surface vanadium oxide species on TiO2 were similar to those discussed above for V/ZrO2 catalysts [41]. The presence of “slightly perturbed” V¼O species, with vanadium in the +4 oxidation state and having a smaller Raman cross-section, was suggested to be present during the SCR reaction. The presence of a V¼O terminal bond for the vanadium in the +4 and +5 oxidation state was associated with the exceptional stability of the V¼O unit. In contrast, when vanadium oxide is reduced to the +3 oxidation

R. Tigiripalli et al.

state, the V¼O terminal bond is absent. Operando multiwavelength Raman spectroscopy with a FTIR analyzer has also been applied to ceria-supported vanadium oxide catalysts, V/CeO2, for the ethanol ODH reaction [52]. However, this study falls under the domain of the unconventional Raman spectroscopy applications and is not elaborated further in this chapter.

Case Study 2: Operando Raman Spectroscopy and Density Functional Theory-Based Vibrational Assignment The ability to study catalytic material during reaction using operando Raman spectroscopy, while simultaneously being able to assign peaks using computational techniques, such as density functional theory (DFT), shows the increased curiosity of the researchers to improve their knowledge about the changes in the different catalytic sites during reaction. We take up one such case involving molybdena-zeolite catalysts, MoOx/ZSM-5, used for the de-hydroaromatization of methane [53]. Zeolites are an important class of nanoporous catalysts made of TO4 tetrahedra [54], where “T” in all-silica zeolites are Si atoms. All-silica zeolites are itself not active; however, replacing Si with Al introduces a negative charge in the framework, which is commonly compensated by positive ions such as H+, Na+, or others. Often these ions are responsible for the activity of the zeolites. Several other strategies have been used to make all-silica framework of zeolite active. For example, nitrogen-doping introduces Lewis basic sites inside the zeolite framework [55, 56]; CuOx dispersed inside zeolites selectively oxidizes methane to methanol [57]; and dispersed MoOx moiety catalyzes methane de-hydroaromatization reaction [53]. We refer the readers elsewhere for more details [54, 58–60]. The intricate pore channels of the zeolite provide huge surface area for the reaction and nanosize pores offer shape selectivity. Because of these properties, zeolites are one of the preferred catalysts for several reactions in the petrochemical and biofuel industry [61]. For example, zeolites are used for alkylation of benzene [62], methanol to olefins [63], isomerization of glucose into fructose [64], and aldol condensation of furfural [65]. Over several decades, vibrational spectroscopies such as infrared, inelastic neutron scattering, energy loss spectroscopy, and Raman have played a crucial role in characterizing zeolites. After infrared spectroscopy, Raman is one of the most used spectroscopies for understanding zeolite science [66]. Previously, Raman spectroscopy has been applied to identify zeolite framework vibrations [67, 68], mechanism of zeolite synthesis [69], adsorbed molecules inside zeolite framework [70, 71] and metal/metal-oxide complexes inside zeolites [71]. Recently, Auerbach and co-workers [72] correlated spectral information and density functional theory (DFT) calculations to show that Raman-

5

Case Studies: Raman Spectroscopy

121

active vibrational modes of tricyclic bridges inside zeolites provide a definite chemical fingerprint. Such a fingerprint analysis is shown for the LTA-type zeolite in Fig. 5.8. The figure shows that vibrations associated with 4-4-6 and 4-4-8 tricyclic bridges possess a Raman fingerprint at 476 cm1, and vibrations associated with 6-6-8 tricyclic bridge has a Raman fingerprint at 318 cm1, though the experimentally observed Raman bands positions were not exactly the same. Thus, computational techniques, such as DFT, can play an important role in analyzing the Raman spectra. Here, we focus on the use of Raman spectroscopy on one selected application of zeolite catalysis which underscores the utility of operando studies and the application of DFT calculations to assist in the understanding of the catalytic system. Methane is the main component of natural gas, which is one of the most abundant fossil-fuel reserves present on earth. Natural gas can be harnessed to meet our current energy requirements. However, the efficient utilization of natural gas poses several challenges – much of which is because it is a gas and needs extremely high pressures to be converted to liquids. It is, therefore, desirable to convert methane to liquid fuels. One of the promising technologies hinges on the use of dispersed metal-oxide species in acidic zeolites as shape-selective catalysts to convert methane to liquid aromatics [73]. The distinct advantage comes from the nonoxidative environment in which this reaction is performed. Among the various metal-oxide catalysts tested, MoOx/ZSM-5 was found to be the most effective catalyst mainly because of the very high selectivity toward benzene and near equilibrium conversions [73]. However, rapid

506

476 cm–1 318 cm–1

Intensity

333

432 Experiment 318

403

476 Simulated

404 200

408

400

600 800 1000 Wavenumber (cm–1)

1200

Fig. 5.8 Experimental and simulated Raman spectra of LTA-type zeolite. Also shown are normal modes of tricyclic bridges in the upper right of the figure. (Adapted with permission from Wang et al. [72]. Copyright © 2019 American Chemical Society)

deactivation of the catalyst hampers the prospects of commercialization of this process. The engineering of such a catalyst can significantly benefit from understanding the molecular nature of catalyst under different conditions. It is well established that catalysts have a dynamic character [16] and characterization under ambient or ex situ conditions may not be representative of characterization under “real” reaction conditions. Although these studies are informative, metal oxides, especially those that do not have a long-range order, tend to hydrate at these conditions. These hydrated structures, like the ones mentioned above for supported metal oxide catalysts, obfuscates the true nature of the active sites under experimental conditions. Here also, operando Raman-MS allows for simultaneous analysis of the catalyst surface as well as the reaction product via an online mass spectrometer. The experimental Raman spectra can then be combined with powerful predictive tools like DFT calculations to elucidate the molecular nature of the dynamic catalyst during reactions. In what follows, we highlight operando Raman experiments coupled with DFT calculations performed by Podkolzin and co-workers to investigate the molecular nature of molybdenum oxide species inside ZSM-5 [53]. Particularly interesting was that the authors recognized that the Raman spectra were to be obtained using a UV laser since the samples were dark [53]. Previous to this study, there was general consensus that Brønsted acid sites were responsible to anchor the surface MoOx moieties in the ZSM-5 framework [74]. There were indications that the active sites could be anchored on two adjacent Brønsted acid sites [74]; however, direct supporting evidence to corroborate this fact was missing. The authors examined the nature of MoOx moieties with in situ Raman spectra of MoOx/ZSM-5 with different Mo loadings (0, 0.7, 1.3, 2.0, 2.7, and 3.1 wt%) and different Si/Al ratios (Si/Al ¼ 15, 25, 40, and 140), which are shown in Fig. 5.9 [53]. The figure is zoomed in to the 800–1200 cm1 region where MoOx moieties are known to have Ramanactive peaks due to Mo-oxygen terminal and Mo-O-M bridging bonds. As expected, the Raman spectra is featureless at 0 wt% Mo loading and appearance of peaks (c.a. 1000 cm1) at higher loading are indicative of dispersed MoOx species on H-ZSM5 catalyst. The obtained Raman spectra has the following features: (a) At Si/Al ¼ 15, there is a peak at 993 cm1 for all Mo loadings. (b) At Si/Al ¼ 15, a shoulder appears at 950 cm1 as Mo loadings are increased from 0.7 wt% to 1.3 wt%. (c) At 1.3 wt% Mo loading, a band at 1026 cm1 appears as Al content in the sample is decreased. (d) A new band appears at 975 cm1 as Si/Al is increased from 15 to 25, and an additional band appears at 984 cm1 at the highest ratio of Si/Al ¼ 140. The independent nature of the peaks at 993 cm1 and 975 cm1 suggested that they originated from two distinct

5

122

R. Tigiripalli et al.

Fig. 5.9 In situ Raman spectrum of MoOx moieties in ZSM5 at different loadings of Mo and Si to Al ratio in the zeolite framework. On top we also show snippets of MoOx moieties in ZSM-5 corresponding to the dominant peaks in the Raman spectrum. For clarity, the zeolite framework has been removed. (From Gao et al. [53]. Adapted with permission from AAAS)

a)

Double Al-site in 10T ring O

O

Single Al-site in 10T ring OH O O

External double Si-site

Mo

Mo

Mo O

O

Al

Al

O

b)

O Al

n = 993 cm1

n = 975 cm1

O

O

O

O

Si

Si

n = 984 cm1

c)

993

984 975 1026 950

1.3 wt% Mo Si/Al Al/Mo 140

0.8

40

2.8

25

4.3

15

7.0

442 nm laser

Raman intensity (a.u.)

Si/Al=15 Mo Al/Mo wt %

800

3.3

2.8

2.7

3.5

993 950

1000 Raman shift (cm–1)

MoOx species. To investigate this suggestion further, the authors performed DFT calculations by considering several permutations of MoOx in ZSM-5 [53]. The systematic procedure used by authors for generating MoOx moieties anchored on Al-doped ZSM5 is shown in Fig. 5.10. Figure 5.10 includes (a) single Al site, (b) double Al site with double oxygen framework bonding, and (c) double Al site with multiple oxygen framework bonding. Apart from these, the authors also considered isolated MoOx moieties on all-silica and T-vacancy site. The vibrational frequencies were computed using quantum-mechanical calculations with GGA-corrected RPBE [75] density functional. The computed vibrational frequencies were compared to the experimental Raman spectra, and the authors were able to assign the bands as follows: The 993 cm1 and 975 cm1 Raman bands were assigned to the symmetric stretches of isolated >Mo(¼O)2 anchored on

2.0

4.7

1.3

7.0

0.7

14

0

N/A

1200

800

1000

1200

Raman shift (cm–1)

double Al sites and isolated >Mo(¼O)2OH anchored on single Al sites, respectively. The assignments are also consistent with the fact that the band at 993 cm1 is only observed at higher Al doping, where double Al sites are expected to be present in more abundance, and a band at 975 cm1 appears when Al doping is decreased, and single Al sites are expected to be dominant. The shoulder at 950 cm1 was attributed to MoOx moieties in ZSM-5 with vacancy defects, and 1026 cm1 was attributed to MoOx moieties on extra-framework alumina nanoparticles. The band at 984 cm1, which appeared for Si/Al ¼ 140, was assigned to MoOx moieties on the silanol groups of zeolite external surface based on DFT calculations. The authors further evaluated the dynamic changes in MoOx moieties under reaction and regeneration conditions using the operando Raman-MS setup. The setup allowed them to simultaneously capture the Raman spectrum and

5

Case Studies: Raman Spectroscopy

123

Al sites

Single Al site

Al-O-(Si-O)2-Al in the same channel but not in the same plain Distance, nm d > 0.7

Double Al site d < 0.7

Al-O-(Si-O)2-Al

2 Al atoms are in the same 10-member ring or in two adjacent 10member rings

with multi (3 or 4) framework oxygen bonding

2 Al atoms are in 6or 8- member rings with d > 0.6 nm. Double Al site

Si

O O Al Si O

with double framework oxygen bonding

Si

O

O O Si Si O

OH

O

Mo Si O Al

Si O O

Single Al site

Si O

Si Si

O

O O Al Si O

O Mo

Si O Si

Si O

O

Al O

O

Si Si

Double Al site with double framework oxygen bonding

Si

O Mo

Si O Si

O

O O Al O

Si O Al O Si O

O Mo

Si O Si

Si O

O

Al O

Si

Double Al site with multi framework oxygen bonding

Fig. 5.10 Permutations of various MoOx moieties anchored on Al doping in ZSM5. (From Gao et al. [53]. Adapted with permission from AAAS)

online mass spectrometry measurements. The catalyst was studied first in methane flow at 953 to 1053 K and then under oxygen flow at 773 K. The Raman spectra under these conditions are shown in Fig. 5.11. The measurements were done for a ZSM-5 zeolite containing Si/Al ¼ 15. For such a sample, the MoOx moieties shows a sharp Raman-active peak at 993 cm1. Initially, when methane is introduced, CO2 is the only carbon-containing product. The peak at 993 cm1 gradually disappears till CO2 no longer evolved. The authors attributed this to formation of MoOxCy or MoCx moieties based on previous experimental studies [52, 76]. The dehydroaromatization of methane started just after disappearance of the 993 cm1 band, clearly indicating that either molybdenum carbide or molybdenum oxycarbide are the catalytic sites for this reaction – underscoring the utility of operando studies in providing direct evidence for the nature of catalytic active sites under true reaction conditions. The catalyst regeneration was further studied in operando Raman setup. As seen from Fig. 5.12, the original Raman-active band at 993 cm1 is regenerated suggesting the reversal of carburization process and regeneration of the original oxide catalyst.

Case Study 3: Recent Studies on the Operando Raman Spectroscopy Identification of Coke During Heterogeneously Catalyzed Reactions In addition to the two case studies above, carbon or coke formation during catalytic reactions has also been explored by operando Raman spectroscopy. Carbon or coke can be an intermediate in the reaction mechanism or formed as a by-product. Carbon formation is often associated with catalyst deactivation [77]. Recently, the formation of carbon during a few important reactions have also been followed by operando Raman spectroscopy. During the CO2 methanation reaction, the role of carbon formation, which might lead to deactivation, was monitored by operando Raman spectroscopy over Ni-based catalysts [78]. However, no carbon formation was detected by Raman spectroscopy during the CO2 methanation reaction at 370
C. Carbon formation was also tracked during the dehydration of methanol to dimethyl ether over a H-ZSM5 catalyst using operando UV Raman spectroscopy for obtaining spectral information and GC þ MS for reactivity data measurements [79]. UV Raman is particularly useful for obtaining the Raman spectra of dark samples. Methyl benzenium carbenium ions, “soft coke,” were identified by a Raman band at 1605 cm1 and proposed to be the

5

124

R. Tigiripalli et al.

1583

a)

b)

1372

953 K

1003 K amu 44 CO2

1169 1033 K, 354 min 1024 K, 333 min

325

1013 K, 308 min 993

Mass spec signal (a.u.)

Raman intensity (a.u.)

442 nm laser 1003 K

amu 28 C2H4 amu 16 CH4

953 K 1003 K

amu 78 C6H6 1033 K 1013 K

953 K, 297 min

0

500

1000

1500

2000

100

Raman shift (cm–1)

1.3 wt% Mo/ZSM-5, Si/Al=15 a.

993

Raman spectra

442 nm laser Regeneration time, min Raman intensity (a.u.)

140

250

300

350

400

times. (b) Evolution of various molecules determined using mass spectrometer as a function of time. (From Gao et al. [53]. Adapted with permission from AAAS)

key precursor for coke formation. This precursor converted to “hard coke” (Raman band at 1390 cm1) at 473 K. However, the conversion of “soft coke” to “hard coke” could be suppressed with water, which was formed during the methanol dehydration reaction. The bands due to carbon formation were also observed above during the P-DH reaction at ~1430 and ~1600 cm1. Additional operando Raman studies identifying the importance of coke formation during reactions are expected in the future.

120 100

5.4

80 60 40

20

800

200

Reaction time (min)

Fig. 5.11 The results of operando Raman mass spectroscopy. The measurements are done for CH4 dehydroaromatization on a 2 wt% Mo/ZSM-5 (Si/Al ¼ 15) catalyst in a flow of 1.5 mol% CH4/He gas. (a) Raman spectrum of the MoOx/ZSM5 sample shown at different

975

150

1000 Raman shift (cm–1)

1200

Fig. 5.12 Operando Raman spectra of regeneration of Mo/ZSM5 carburized catalyst under the flow of oxygen at 773 K. (From Gao et al. [53]. Adapted with permission from AAAS)

Present Challenges and Future Recourse

One of the important goals of petrochemical, environmental, or biochemical industry is to understand the existing catalytic process in detail and to develop a “reaction-specific” catalyst. Such goals assist in the design and operation of industrial reactions. A bottleneck in this direction is to decipher what happens on the surface of the catalyst at industrial operating conditions. Operando spectroscopies have pushed this envelope closer to reality. However, a thorough analysis and assignment of spectral data poses its own challenges. As shown by Wachs and Roberts in 2010, the application of in situ and operando Raman spectroscopy to catalysis, in term of number of publications, has been increasing over the years [80] and was predicted to explode in the future. The updated figure, as shown in Fig. 5.13, shows that this prediction is indeed true, and the application of in situ and operando Raman to catalysis continues to grow further as we search for

5

Case Studies: Raman Spectroscopy

125

350 In-situ Raman

Operando Raman

Number of publications

300 250 200 150 100 50 0 1980

1985

1990

1995

2000 Year

2005

2010

2015

Fig. 5.13 Number of publications per year published using in situ and operando Raman spectroscopy in the area of catalysis. A similar plot till 2010 was reported by Wachs and Roberts [80]. We obtained this data from Web of Science database with search keywords: “catalysis + catalyst + in situ Raman” or “catalysis + catalyst + operando Raman”

new applications and more detailed understanding of the catalytic phenomena. Starting from the application of a single characterization technique, researchers have proceeded to simultaneously use multiple characterization techniques in the operando mode to obtain additional information about the catalytic system. This is in addition to obtaining complementary reactivity data. For example, with a judicious choice of multiple characterization techniques operating in the operando mode, along with Raman, a more comprehensive explanation of the progress of the reaction can be proposed. Such hyphenated Raman characterization techniques are becoming more common in the study of heterogeneous catalysts. For example, Bruckner applied operando EPR-Raman-UV–vis-MS to investigate the propane ODH reaction for a V/TiO2 catalyst that possessed surface vanadium oxide species and bulk V2O5 [81]. This study revealed that VO+2 species are present in the isolated and polymeric species and on the surface of microcrystalline V2O5. These species dominate the surface of the catalyst during reaction and are possibly the active sites. Similarly, Weckhuysen and co-workers [82] applied time-resolved operando UV-vis and Raman spectroscopy with online mass spectroscopy for the propane dehydrogenation reaction over a supported chromium oxide catalyst. Different oxidation states of the chromium oxide catalysts and the growth of coke was evaluated as a function of time-on-stream and different cycles. These results illustrate that additional information about the catalytic system can be obtained by the simultaneous application of two characterization techniques with online analysis of the gas phase. Subsequently, Weckhuysen and co-workers used three spectroscopic

techniques to study alumina- and silica-supported molybdenum oxide for the propane dehydrogenation reaction. The three complementary techniques used were energy dispersive X-ray absorption fine structure (ED-XAFS), UV-vis, and Raman spectroscopy. They proposed that molybdenum, in the +4 oxidation state, may be necessary for the dehydrogenation reaction. Furthermore, deactivation occurs by coke formation (reversible) and solid-state reaction (irreversible). However, information about the adsorbed species was limited, and a complete picture of the catalytic phenomena is unclear. Furthermore, the case studies discussed above needs to be analyzed with the mass and heat transfer considerations. In other words, a proper CFD study along with the operando hyphenated Raman analysis appears essential. As highlighted in the case study above, the computational models based on quantum chemical calculations can be used as a powerful tool for assigning Raman peaks and elucidating the molecular nature of the catalysts. These calculations are, however, done on an idealized in silico models. A highthroughput calculation assigning the dynamic changes in the spectra or high-throughput calculation screening several catalysts is far from reality. Although, we have matured with respect to the theory, computing Raman spectra of a complex working catalyst under true experimental conditions remains a daunting task. First, the computational expensive nature of quantummechanical calculations limits its feasibility to only hundreds of atoms. The periodic repetition of the simulation cell allows modelling of an infinite solid. However, for larger simulation cells, one has to estimate the potential using classical force fields. One is then left to the mercy of the “goodness” of the force-field parameters to correctly predict the normal mode vibrations – which for an unknown catalytic system is highly improbable. The DFT calculations scale as O(Ne3), where Ne is the total number of electrons in the system. The field can significantly benefit from developing more tractable DFT methods or semi-empirical methods [83, 84]. Another line of research would be to develop accurate classical force fields. Machine learning approaches to develop interatomic potentials supervised from quantum mechanical calculations is gaining popularity in this direction [85–88]. The development of faster and accurate computational methods will allow simulations of a much larger and wider spectrum of catalytic systems. Second, computation of Raman spectra requires generation of representative structural models of the working catalyst. Given the multiplex structural and compositional changes which a catalyst surface undergoes during a reaction [89–91], creating a library of possible surface morphology is a challenging task. Often, this library is manually created using chemical intuition and experimental understanding, rendering this process cumbersome and tedious. The field

5

126

would gain by the development of fast and easy-to-use automated approaches aimed at finding the global minima of the catalyst structures [92]. Third, the state-of-the-art ab initio calculations are structural optimization calculations which do not consider the effect of temperature and pressure, i.e., they are performed at 0 K and zero pressure, the ubiquitous temperature and pressure gap from actual reaction conditions. Recently, there has been a rapid increase in approaches which gauge the stability of potential catalyst structures by computing chemical potentials at realistic temperatures and pressures [92–94]. One such approach is ab initio thermodynamics [95–98], which uses statistical mechanics and energetics from ab initio calculations to compute surface free energy of the candidate catalyst. At realistic temperatures and pressures, the candidate which minimizes the surface energy is deemed to be the most probable structure. Thus, the application of Raman spectroscopy to heterogeneous catalysis holds a bright future. In the foreseeable future, the applications would monotonically increase, and the upcoming applications are expected to be more detailed in terms of multiscale analysis. The molecular-level understanding provided by Raman needs to be translated to optimization, operation, and design of industrial reactors. An ideal operando system including the above aspects remains a challenge, and future improvements along these lines are necessary. It is inevitable that this would be possible by further development of operando experimental techniques coupled with multiscale computational methods. Acknowledgments R. T. acknowledges the financial support from PMRF (MHRD, GOI) for her doctoral research. V. A. acknowledges the funding support provided by SERB (Ramanujan Fellowship No. SB/S2/RJN-013/2017).

References 1. Raman, C.V., Krishnan, K.S.: A New Type of Secondary Radiation, Nature. 121, 501 (1928) 2. Heidary, N., Kornienko, N.: Operando vibrational spectroscopy for electrochemical biomass valorization, Chem. Commun. 56, 8727 (2020) 3. Lu, Y., Zhou, P., Lei, K., Zhao, Q., Tao, Z., Chen, J.: Selenium Phosphide (Se4P4) as a New and Promising Anode Material for Sodium-Ion Batteries, Adv. Energy Mater. 7, 1601973 (2017) 4. De Sousa Duarte, M., Rolland, M., Sagnard, C., Suire, D., Flacher, F., Delpoux, O., Lienemann, C.P.: High-Pressure High-Temperature Transparent Fixed-Bed Reactor for Operando Gas-Liquid Reaction Follow-up, Chem. Eng. Technol. 42, 655 (2019) 5. Hartman, T., Wondergem, C.S., Kumar, N., Van Den Berg, A., Weckhuysen, B.M.: Surface- and Tip-Enhanced Raman Spectroscopy in Catalysis, J. Phys. Chem. Lett. 7, 1570 (2016) 6. Tolia, A.A., Weaver, M.J., Takoudis, C.G.: In situ surface‐enhanced Raman spectroscopic study of CO oxidation by NO and O2 over rhodium‐coated gold surfaces, J. Vac. Sci. Technol. A Vacuum, Surfaces, Film. 11, 2013 (1993)

R. Tigiripalli et al. 7. Kim, H., Kosuda, K.M., Van Duyne, R.R., Stair, P.C.: Resonance Raman and surface- and tip-enhanced Raman spectroscopy methods to study solid catalysts and heterogeneous catalytic reactions, Chem. Soc. Rev. 39, 4820 (2010) 8. Stavitski, E., Weckhuysen, B.M.: Infrared and Raman imaging of heterogeneous catalysts, Chem. Soc. Rev. 39, 4615 (2010) 9. Bravo-Suárez, J.J., Srinivasan, P.D.: Design characteristics of in situ and operando ultraviolet-visible and vibrational spectroscopic reaction cells for heterogeneous catalysis, Catal. Rev. – Sci. Eng. 59, 295 (2017) 10. Meunier, F.C.: The design and testing of kinetically-appropriate operando spectroscopic cells for investigating heterogeneous catalytic reactions, Chem. Soc. Rev. 39, 4602 (2010) 11. Schrader, G.L., Hill, C.G.: Variable temperature sample cell and optical mount for laser Raman studies of adsorbed species, Rev. Sci. Instrum. 46, 1335 (1975) 12. Bañares, M.A.: Operando methodology: combination of in situ spectroscopy and simultaneous activity measurements under catalytic reaction conditions, Catal. Today. 100, 71–77 (2005) 13. Guerrero-Pérez, M.O., Bañares, M.A., Banares, M.A.: Operando Raman study of alumina-supported Sb–V–O catalyst during propane ammoxidation to acrylonitrile with on-line activity measurement, Chem. Commun. 2, 1292 (2002) 14. Bañares, M.A., Guerrero-Pérez, M.O., Fierro, J.L.G., Cortez, G.G.: Raman spectroscopy during catalytic operations with on-line activity measurement (operando spectroscopy): a method for understanding the active centres of cations supported on porous materials, J. Mater. Chem. 12, 3337 (2002) 15. Bañares, M.A., Wachs, I.E.: Molecular structures of supported metal oxide catalysts under different environments, J. Raman Spectrosc. 33, 359 (2002) 16. Chakrabarti, A., Ford, M.E., Gregory, D., Hu, R., Keturakis, C.J., Lwin, S., Tang, Y., Yang, Z., Zhu, M., Bañares, M.A., Wachs, I.E.: A decade+ of operando spectroscopy studies, Catal. Today. 283, 27 (2017) 17. Bañares, M.A.: In Situ to Operando Spectroscopy: From Proof of Concept to Industrial Application, Top. Catal. 52, 1301 (2009) 18. Mestl, G.: In situ Raman spectroscopy — a valuable tool to understand operating catalysts, J. Mol. Catal. A Chem. 158, 45 (2000) 19. Geske, M., Korup, O., Horn, R.: Resolving kinetics and dynamics of a catalytic reaction inside a fixed bed reactor by combined kinetic and spectroscopic profiling, Cat. Sci. Technol. 3, 169 (2013) 20. Tinnemans, S.J., Mesu, J.G., Kervinen, K., Visser, T., Nijhuis, T.A., Beale, A.M., Keller, D.E., van der Eerden, A.M.J., Weckhuysen, B.M.: Combining operando techniques in one spectroscopic-reaction cell: New opportunities for elucidating the active site and related reaction mechanism in catalysis, Catal. Today. 113, 3 (2006) 21. Thomas, C.L.M.: Catalytic Processes and Proven Catalysts. Academic Press, New York (1970) 22. Bañares, M.A.: Supported metal oxide and other catalysts for ethane conversion: a review, Catal. Today. 51, 319 (1999) 23. Fu, M., Li, C., Lu, P., Qu, L., Zhang, M., Zhou, Y., Yu, M., Fang, Y.: A review on selective catalytic reduction of NOx by supported catalysts at 100–300
C—catalysts, mechanism, kinetics, Cat. Sci. Technol. 4, 14 (2014) 24. Einaga, H., Maeda, N., Nagai, Y.: Comparison of catalytic properties of supported metal oxides for benzene oxidation using ozone, Cat. Sci. Technol. 5, 3147 (2015) 25. Wachs, I.E.: Catalysis science of supported vanadium oxide catalysts, Dalton Trans. 42, 11762 (2013) 26. Wachs, I.E.: Raman and IR studies of surface metal oxide species on oxide supports: Supported metal oxide catalysts, Catal. Today. 27, 437 (1996) 27. Gao, X., Wachs, I.E.: Investigation of Surface Structures of Supported Vanadium Oxide Catalysts by UV–vis–NIR Diffuse Reflectance Spectroscopy, J. Phys. Chem. B. 104, 1261 (2000)

5

Case Studies: Raman Spectroscopy

28. Deo, G., Wachs, I.E.: Predicting molecular structures of surface metal oxide species on oxide supports under ambient conditions, J. Phys. Chem. 95, 5889 (1991) 29. Vuurman, M.A., Wachs, I.E.: In situ Raman spectroscopy of alumina-supported metal oxide catalysts, J. Phys. Chem. 96, 5008 (1992) 30. Jehng, J.M., Deo, G., Weckhuysen, B.M., Wachs, I.E.: Effect of water vapor on the molecular structures of supported vanadium oxide catalysts at elevated temperatures, J. Mol. Catal. A Chem. 110, 41 (1996) 31. Lewandowska, A.E., Calatayud, M., Tielens, F., Bañares, M.A.: Dynamics of Hydration in Vanadia–Titania Catalysts at Low Loading: A Theoretical and Experimental Study, J. Phys. Chem. C. 115, 24133 (2011) 32. Weckhuysen, B.M., Jehng, J.-M., Wachs, I.E.: In Situ Raman Spectroscopy of Supported Transition Metal Oxide Catalysts: 18O2–16O2 Isotopic Labeling Studies, J. Phys. Chem. B. 104, 7382 (2000) 33. Cavani, F., Ballarini, N., Cericola, A.: Oxidative dehydrogenation of ethane and propane: How far from commercial implementation? Catal. Today. 127, 113 (2007) 34. Carrero, C.A., Schloegl, R., Wachs, I.E., Schomaecker, R.: Critical Literature Review of the Kinetics for the Oxidative Dehydrogenation of Propane over Well-Defined Supported Vanadium Oxide Catalysts, ACS Catal. 4, 3357 (2014) 35. Hocking, M.B.: Handbook of Chemical Technology and Pollution Control, 3rd edn. Elsevier Inc., Amsterdam, Netherlands (2005) 36. Zimmermann, H.: Ullmann’s Encyclopedia of Industrial Chemistry. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim (2013) 37. Gao, X., Jehng, J.M., Wachs, I.E.: In Situ UV–vis–NIR Diffuse Reflectance and Raman Spectroscopic Studies of Propane Oxidation over ZrO2-Supported Vanadium Oxide Catalysts, J. Catal. 209, 43 (2002) 38. Went, G.T., Ted Oyama, S., Bell, A.T.: Laser Raman spectroscopy of supported vanadium oxide catalysts, J. Phys. Chem. 94, 4240 (1990) 39. Cristiani, C., Forzatti, P., Busca, G.: On the surface structure of vanadia-titania catalysts: Combined laser-Raman and fourier transform-infrared investigation, J. Catal. 116, 586 (1989) 40. Wachs, I.E., Weckhuysen, B.M.: Structure and reactivity of surface vanadium oxide species on oxide supports, Appl. Catal. A Gen. 157, 67 (1997) 41. Christodoulakis, A., Machli, M., Lemonidou, A.A., Boghosian, S.: Molecular structure and reactivity of vanadia-based catalysts for propane oxidative dehydrogenation studied by in situ Raman spectroscopy and catalytic activity measurements, J. Catal. 222, 293 (2004) 42. Burcham, L.J., Deo, G., Gao, X., Wachs, I.E.: In situ IR, Raman, and UV-Vis DRS spectroscopy of supported vanadium oxide catalysts during methanol oxidation, Top. Catal. 11–12, 85 (2000) 43. Mul, G., Bañares, M.A., Cortéz, G.G., Van Der Linden, B., Khatib, S.J., Moulijn, J.A.: MultiTRACK and operando Raman-GC study of oxidative dehydrogenation of propane over alumina-supported vanadium oxide catalysts, Phys. Chem. Chem. Phys. 5, 4378 (2003) 44. Mars, P., van Krevelen, D.W.: Oxidations carried out by means of vanadium oxide catalysts, Chem. Eng. Sci. 3, 41 (1954) 45. Ternero-Hidalgo, J.J., Guerrero-Pérez, M.O., Rodríguez-Mirasol, J., Cordero, T., Bañares, M.A., Portela, R., Bazin, P., Clet, G., Daturi, M.: Operando Reactor-Cell with Simultaneous Transmission FTIR and Raman Characterization (IR Raman) for the Study of Gas-Phase Reactions with Solid Catalysts, Anal. Chem. 92, 5100 (2020) 46. Vimont, A., Thibault-Starzyk, F., Daturi, M.: Analysing and understanding the active site by IR spectroscopy, Chem. Soc. Rev. 39, 4928 (2010) 47. Sanchez Escribano, V., Busca, G., Lorenzelli, V.: Fourier transform infrared spectroscopic studies of the reactivity of vanadia-titania

127 catalysts toward olefins. 1. Propylene, J. Phys. Chem. 94, 8939 (1990) 48. Shee, D., Deo, G.: In situ DRIFT studies of alkane adsorption on vanadia supported titania-doped catalysts, Catal. Today. 325, 25 (2019) 49. Ternero-Hidalgo, J.J., Torres-Liñán, J., Guerrero-Pérez, M.O., Rodríguez-Mirasol, J., Cordero, T.: Electrospun vanadium oxide based submicron diameter fiber catalysts. Part I: Preparation procedure and propane ODH application, Catal. Today. 325, 131 (2019) 50. Moncada, J., Adams, W.R., Thakur, R., Julin, M., Carrero, C.A.: Developing a Raman Spectrokinetic Approach To Gain Insights into the Structure–Reactivity Relationship of Supported Metal Oxide Catalysts, ACS Catal. 8, 8976 (2018) 51. Godiksen, A.L., Rasmussen, S.B.: Identifying the presence of [V=O]2+ during SCR using in-situ Raman and UV Vis spectroscopy, Catal. Today. 336, 45 (2019) 52. Ober, P., Rogg, S., Hess, C.: Direct Evidence for Active Support Participation in Oxide Catalysis: Multiple Operando Spectroscopy of VOx/Ceria, ACS Catal. 10, 2999 (2020) 53. Gao, J., Zheng, Y., Jehng, J.M., Tang, Y., Wachs, I.E., Podkolzin, S.G.: Identification of molybdenum oxide nanostructures on zeolites for natural gas conversion, Science (80-.). 348, 686 (2015) 54. Auerbach, S.M., Carrado, K.A., Dutta, P.K.: Handbook of Zeolite Science and Technology. Marcel Dekker, Inc., New York (2003) 55. Agarwal, V., Huber, G.W., Conner, W.C., Auerbach, S.M.: DFT study of nitrided zeolites: Mechanism of nitrogen substitution in HY and silicalite, J. Catal. 269, 53 (2010) 56. Agarwal, V., Huber, G.W., Curtis Conner, W., Auerbach, S.M.: Kinetic stability of nitrogen-substituted sites in HY and silicalite from first principles, J. Catal. 270, 249 (2010) 57. Grundner, S., Markovits, M.A.C., Li, G., Tromp, M., Pidko, E.A., Hensen, E.J.M., Jentys, A., Sanchez-Sanchez, M., Lercher, J.A.: Single-site trinuclear copper oxygen clusters in mordenite for selective conversion of methane to methanol, Nat. Commun. 6, 1 (2015) 58. Kulprathipanja, S.: Zeolites in Industrial Separation and Catalysis. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim (2010) 59. Weitkamp, J., Puppe, L.: Catalysis and Zeolites: Fundamentals and Applications. Springer Science & Business Media, Berlin, Germany (1999) 60. Čejka, J., Corma, A., Zones, S.: Zeolites and Catalysis: Synthesis, Reactions and Applications. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim (2010) 61. Taarning, E., Osmundsen, C.M., Yang, X., Voss, B., Andersen, S.I., Christensen, C.H.: Zeolite-catalyzed biomass conversion to fuels and chemicals, Energy Environ. Sci. 4, 793 (2011) 62. Christensen, C.H., Johannsen, K., Schmidt, I., Christensen, C.H.: Catalytic Benzene Alkylation over Mesoporous Zeolite Single Crystals: Improving Activity and Selectivity with a New Family of Porous Materials, J. Am. Chem. Soc. 125, 13370 (2003) 63. Liang, T., Chen, J., Qin, Z., Li, J., Wang, P., Wang, S., Wang, G., Dong, M., Fan, W., Wang, J.: Conversion of Methanol to Olefins over H-ZSM-5 Zeolite: Reaction Pathway Is Related to the Framework Aluminum Siting, ACS Catal. 6, 7311 (2016) 64. Moreau, C., Durand, R., Roux, A., Tichit, D.: Isomerization of glucose into fructose in the presence of cation-exchanged zeolites and hydrotalcites, Appl. Catal. A Gen. 193, 257 (2000) 65. Kikhtyanin, O., Kelbichová, V., Vitvarová, D., Kubů, M., Kubička, D.: Aldol condensation of furfural and acetone on zeolites, Catal. Today. 227, 154 (2014) 66. Bordiga, S., Lamberti, C., Bonino, F., Travert, A., Thibault-Starzyk, F.: Probing zeolites by vibrational spectroscopies, Chem. Soc. Rev. 44, 7262 (2015) 67. Dutta, P.K., Shieh, D.C., Puri, M.: Correlation of framework Raman bands of zeolites with structure, Zeolites. 8, 306 (1988)

5

128 68. Dutta, P.K., Rao, K.M., Park, J.Y.: Correlation of Raman spectra of zeolites with framework architecture, J. Phys. Chem. 95, 6654 (1991) 69. Dutta, P.K., Puri, M.: Synthesis and structure of zeolite ZSM-5: a Raman spectroscopic study, J. Phys. Chem. 91, 4329 (1987) 70. Tam, N.T., Cooney, R.P., Curthoys, G.: Vibrational spectra of molecules on zeolites. Part 1.—Acetylene on A-type zeolites, J. Chem. Soc. Faraday Trans. 1 Phys. Chem. Condens. Phases. 72, 2577 (1976) 71. Knops-Gerrits, P.P., De Vos, D.E., Feijen, E.J.P., Jacobs, P.A.: Raman spectroscopy on zeolites, Microporous Mater. 8, 3 (1997) 72. Wang, T., Luo, S., Tompsett, G.A., Timko, M.T., Fan, W., Auerbach, S.M.: Critical Role of Tricyclic Bridges Including Neighboring Rings for Understanding Raman Spectra of Zeolites, J. Am. Chem. Soc. 141, 20318 (2019) 73. Zhang, Z.-G.: Process, reactor and catalyst design: Towards application of direct conversion of methane to aromatics under nonoxidative conditions, Carbon Resour. Convers. 2, 157 (2019) 74. Borry, R.W., Kim, Y.H., Huffsmith, A., Reimer, J.A., Iglesia, E.: Structure and Density of Mo and Acid Sites in Mo-Exchanged HZSM5 Catalysts for Nonoxidative Methane Conversion, J. Phys. Chem. B. 103, 5787 (1999) 75. Hammer, B., Hansen, L.B., Nørskov, J.K.: Improved adsorption energetics within density-functional theory using revised PerdewBurke-Ernzerhof functionals, Phys, Phys. Rev. B - Condens. Matter Mater. Phys. 59, 7413 (1999) 76. Zheng, H., Ma, D., Bao, X., Jian, Z.H., Ja, H.K., Wang, Y., Peden, C.H.F.: Direct Observation of the Active Center for Methane Dehydroaromatization Using an Ultrahigh Field 95Mo NMR Spectroscopy, J. Am. Chem. Soc. 130, 3722 (2008) 77. Argyle, M., Bartholomew, C.: Heterogeneous Catalyst Deactivation and Regeneration: A Review, Catalysts. 5, 145 (2015) 78. Mutz, B., Sprenger, P., Wang, W., Wang, D., Kleist, W., Grunwaldt, J.D.: Operando Raman spectroscopy on CO2 methanation over alumina-supported Ni, Ni3Fe and NiRh0.1 catalysts: Role of carbon formation as possible deactivation pathway, Appl. Catal. A Gen. 556, 160 (2018) 79. An, H., Zhang, F., Guan, Z., Liu, X., Fan, F., Li, C.: Investigating the Coke Formation Mechanism of H-ZSM-5 during Methanol Dehydration Using Operando UV–Raman Spectroscopy, ACS Catal. 8, 9207 (2018) 80. Wachs, I.E., Roberts, C.A.: Monitoring surface metal oxide catalytic active sites with Raman spectroscopy, Chem. Soc. Rev. 39, 5002 (2010) 81. Brückner, A.: Killing three birds with one stone—simultaneous operando EPR/UV-vis/Raman spectroscopy for monitoring catalytic reactions, Chem. Commun. 1761 (2005) 82. (Xander) Nijhuis, T.A., Tinnemans, S.J., Visser, T., Weckhuysen, B.M.: Operando spectroscopic investigation of supported metal oxide catalysts by combined time-resolved UV-VIS/Raman/on-line mass spectrometry, Phys. Chem. Chem. Phys. 5, 4361 (2003) 83. Goringe, C.M., Hernández, E., Gillan, M.J., Bush, I.J.: Linear-scaling DFT-pseudopotential calculations on parallel computers, Comput. Phys. Commun. 102, 1 (1997) 84. Prentice, J.C.A., Aarons, J., Womack, J.C., Allen, A.E.A., Andrinopoulos, L., Anton, L., Bell, R.A., Bhandari, A., Bramley, G.A., Charlton, R.J., Clements, R.J., Cole, D.J., Constantinescu, G., Corsetti, F., Dubois, S.M.M., Duff, K.K.B., Escartín, J.M., Greco, A., Hill, Q., Lee, L.P., Linscott, E., O’Regan, D.D., Phipps, M.J.S., Ratcliff, L.E., Serrano, Á.R., Tait, E.W., Teobaldi, G., Vitale, V., Yeung, N., Zuehlsdorff, T.J., Dziedzic, J., Haynes, P.D., Hine, N.D.M., Mostofi, A.A., Payne, M.C., Skylaris, C.K.: The ONETEP linear-scaling density functional theory program, J. Chem. Phys. 152, 174111 (2020)

R. Tigiripalli et al. 85. Li, Z., Wang, S., Xin, H.: Toward artificial intelligence in catalysis, Nat. Catal. 1, 641 (2018) 86. Yang, W., Fidelis, T.T., Sun, W.H.: Machine Learning in Catalysis, From Proposal to Practicing, ACS Omega. 5, 83 (2020) 87. Kitchin, J.R.: Machine learning in catalysis, Nat. Catal. 1, 230 (2018) 88. Van Santen, R.A., Sautet, P.: Computational Methods in Catalysis and Materials Science: An Introduction for Scientists and Engineers. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim (2015) 89. Schlögl, R.: Heterogeneous Catalysis, Angew. Chem. Int. Ed. 54, 3465 (2015) 90. Newton, M.A.: Dynamic adsorbate/reaction induced structural change of supported metal nanoparticles: heterogeneous catalysis and beyond, Chem. Soc. Rev. 37, 2644 (2008) 91. Kalz, K.F., Kraehnert, R., Dvoyashkin, M., Dittmeyer, R., Gläser, R., Krewer, U., Reuter, K., Grunwaldt, J.D.: Future Challenges in Heterogeneous Catalysis: Understanding Catalysts under Dynamic Reaction Conditions, ChemCatChem. 9, 17 (2017) 92. Grajciar, L., Heard, C.J., Bondarenko, A.A., Polynski, M.V., Meeprasert, J., Pidko, E.A., Nachtigall, P.: Towards operando computational modeling in heterogeneous catalysis, Chem. Soc. Rev. 47, 8307 (2018) 93. Kirchhoff, B., Braunwarth, L., Jung, C., Jónsson, H., Fantauzzi, D., Jacob, T.: Simulations of the Oxidation and Degradation of Platinum Electrocatalysts, Small. 16, 1905159 (2020) 94. Asthagiri, A., Janik, M.: Computational Catalysis. Royal Society of Chemistry, Cambridge, UK (2014) 95. Reuter, K., Metiu, H.: Handbook of Materials Modeling, pp. 1309–1319. Springer International Publishing, Cham, Switzerland (2020) 96. Rogal, J.; Reuter, K.: Ab Initio Atomistic Thermodynamics for Surfaces: A Primer, In: Experiment, Modeling and Simulation of Gas-Surface Interactions for Reactive Flows in Hypersonic Flights, NATO Research and Technology Organization: Neuilly-sur-Seine, France (2007) pp 2-1–2-18. 97. Reuter, K.: Ab Initio Thermodynamics and First-Principles Microkinetics for Surface Catalysis, Catal. Lett. 146, 541 (2016) 98. Reuter, K., Stampf, C., Scheffler, M.: Ab Initio Atomistic Thermodynamics and Statistical Mechanics of Surface Properties and Functions. In: Yip, S. (eds) Handbook of Materials Modeling, pp. 149–194. Springer, Dordrecht, Netherlands (2005)

Ragamaye Tigiripalli is a PhD student in the chemical engineering department at IIT Kanpur under the Prime Minister’s Research Fellowship Program. Her supervisors are Prof. Goutam Deo and Prof. Vishal Agarwal. She graduated from NIT Warangal in 2018 and is a gold medalist. Her current focus is on developing an efficient catalyst for hydrogenation of CO2 to methanol using experiments and computations.

5

Case Studies: Raman Spectroscopy

Vishal Agarwal is presently working as an assistant professor in the department of chemical engineering at the Indian Institute of Technology Kanpur, India. Previously, he worked as a postdoctoral scholar at the University of California, Santa Barbara. He obtained his PhD in Chemical Engineering from the University of Massachusetts, Amherst. He now works on computational catalysis in the areas of CO2 and biomass conversion to fuels.

129

Goutam Deo has been a faculty member at the Indian Institute of Technology Kanpur since 1996. He received his PhD from Lehigh University, Masters from the University of Iowa, and Bachelors from the Indian Institute of Technology Kharagpur, all in chemical engineering. His main areas of research are in catalysis and reaction engineering, with specific interest in supported metals and metal oxides.

5

6

Ultraviolet (UV) Raman Spectroscopy Peter C. Stair

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

6.1

6.2 Description of Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . 132 6.2.1 Raman Scattering Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 6.2.2 Benefits and Limitations for Catalyst Characterization . . . 134

measurements with a special focus on sample handling and in situ reaction cells. The remainder of the chapter summarizes studies of catalyst synthesis, catalyst deactivation by coke formation, and catalytic metal oxide speciation. The identification and appearance of resonanceenhanced Raman scattering and how it can be exploited are emphasized.

6.3

UV Raman Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

6.4

New Instrument Advances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

6.5

Reaction Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

6.6

Chronology of Application to Catalysis . . . . . . . . . . . . . . . . . 138

Keywords

6.7

Time Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

6.8

Spatial Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

UV Raman · Resonance Raman · Catalyst deactivation · Supported oxide catalysts

6.9 6.9.1 6.9.2 6.9.3 6.9.4 6.9.5 6.9.6

Applications of UV Raman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Silica/Zeolite Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal, Oxidation, and Reduction Treatments . . . . . . . . . . . Catalyst Deactivation by Coke Formation . . . . . . . . . . . . . . . . . Speciation of Titania . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supported Vanadium Oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ceria Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6.10

Multimodal Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

6.11

Conclusions and Future Outlook . . . . . . . . . . . . . . . . . . . . . . . . 146

140 140 140 141 143 145 146

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

Abstract

Using an ultraviolet laser to excite Raman scattering has the benefit of avoiding fluorescence interference and increasing Raman intensity via resonance enhancement. The high photon energy inherent at ultraviolet wavelengths requires special precautions for sample handling to minimize transformations caused by laser-induced heating and photochemistry. This chapter begins by covering the basic theory of resonance-enhanced Raman spectroscopy and the instrumentation for making P. C. Stair (*) Department of Chemistry, Northwestern University, Evanston, IL, USA e-mail: [email protected]

6.1

Introduction

When a sample of interest is irradiated by a narrow-band, intense laser, the light is transmitted, scattered, or absorbed. A small fraction of the scattered light appears at longer or shorter wavelengths from the incident laser (Stokes and antiStokes scattering, respectively) as a consequence of energy transfer to (or from) bond vibrations in the sample. When spectrally dispersed, this inelastically scattered light comprises the vibrational Raman spectrum of the sample. As indicated elsewhere in this Handbook, Raman spectroscopy has proven to be one of the most powerful techniques for the characterization of catalytic systems, via their vibrational spectrum, under ex situ, in situ, and operando conditions. As a light scattering technique it can be applied to a wide range of high-surface-area, solid catalytic materials, including bulk metal oxides, zeolites, metal- or oxide-supported catalysts. Both gas and liquid phase reaction environments that are transparent to the laser beam are accessible to Raman spectroscopy. It is one of the few techniques that can provide information about both the solid catalyst and the molecular reagents in a single measurement. UV and UV Resonance Raman spectroscopy are specific variants of Raman spectroscopy that make use of one or more

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_6

131

132

P. C. Stair

ultraviolet laser wavelengths to excite Raman scattering. Using UV laser excitation it is possible to avoid interference from fluorescence or blackbody emission that, depending on the sample, can otherwise totally obscure the Raman spectrum under a broad, featureless background [1]. With a cross section that can be 106 times larger than the Raman cross section, fluorescence from a trace impurity can make the Raman measurement impossible [2]. A common source of fluorescence in catalytic systems is coke deposited during reaction [2–7]. In 1984 Asher and Johnson demonstrated that by exciting the Raman scattering using a UV laser at 260 nm and below, the spectrum could be measured even from a highly fluorescent coal liquid [2]. The success of this approach is not due to a suppression of the fluorescence but because the fluorescence appears at longer wavelengths and does not overlap the Raman spectrum. The application of UV Raman to catalytic materials was first published in 1996 [8]. Avoidance of fluorescence or blackbody radiation extends the range of catalyst materials and reaction conditions beyond what can be measured using conventional, visible laser excitation or infrared spectroscopy. In addition, by tuning the excitation laser to specific wavelengths within the optical absorption region of the sample, it is possible to enhance the Raman spectrum from only select components in a mixture of species via resonance Raman scattering (RRS), even if that species is a small percentage of the mixture. While UV and UV Resonance Raman spectroscopies can extend the capabilities of conventional Raman spectroscopy, they also present specific challenges to successfully making the measurements and interpreting the results that are somewhat different from conventional Raman spectroscopy. After a brief recap of Raman and Resonance Raman spectroscopy theory, this chapter will address the practical experimental and interpretational considerations associated with these two measurements.

6.2

Description of Raman Spectroscopy

Here we provide a brief, practical description of the Raman scattering process sufficient to understand its application to catalyst characterization. For a complete description of Raman and resonance Raman from molecules see the monograph by D. A. Long [9].

6.2.1

Raman Scattering Theory

Like infrared spectroscopy (FTIR), Raman spectroscopy measures the vibrational spectrum of the sample system. By comparison to FTIR, Raman spectroscopy can cover the entire vibrational spectrum from low frequency (4000 cm
1), whereas FTIR is often limited to vibrations above ~1000 cm
1 due to absorption by the metal oxide materials that are common in heterogeneous catalysis. Raman spectroscopy can often be successfully applied to samples at high reaction temperatures (>800  C) where the background arising from blackbody radiation obscures the IR spectrum. On the other hand Raman scattering is inherently weaker than IR absorption. The fraction of photons absorbed in a strong IR band (depending on the concentration of the absorbing species) is 0.01–0.1, whereas, the fraction of Raman scattered photons from a strongly scattering vibration is 10
7. With these differences it might seem that the measurement of Raman spectra is hopeless. Indeed, before the advent of intense laser sources, Raman spectra were difficult to measure. A typical incident laser power of 1 mW, corresponding to ~1015 photons per second, will produce 106–108 counts per second of Raman scattered light at a commercial spectrometer detector, a readily detectable signal. The intensity of classical Stokes Raman scattering (inelastic energy loss) can be written:


I ωL
ωfi




4 X 0 2 4π ωL
ωfi I ð ω ÞN αρσ ¼ 0 L 3c4 ρσ

ð6:1Þ

I(ωL
ωfi) is the intensity of the Raman band in power per steradian; ωL & ωfi are the frequencies of the laser and the vibrational transition from state i to state f; c is the speed of light; I0(ωL) is the laser power per steradian; N is the number of scatterers; and α0ρσ is the ρσ-component tensor of polarizability derivatives with respect to the normal coordinate of vibration. From this equation we see that Raman scattering is produced when the polarizability changes as a bond vibrates. In a typical experiment, using a visible or ultraviolet laser to excite Raman scattering, ωL  ωfi and ωL
ωfi ≈ ωL so that the Raman intensity increases as  ω4L , i.e., the Raman intensity increases as the laser excitation moves from visible to ultraviolet. However, the Raman intensity can increase much more dramatically when the laser photon energy is sufficient to excite an electronic transition in the sample, giving rise to resonance Raman scattering (RRS). Figure 6.1 illustrates the processes of elastic (Rayleigh) scattering, Stokes and anti-Stokes Raman scattering for (a) normal and (b) resonance Raman on a Jablonski diagram. Experimentally, resonance enhancement can be identified by Raman bands whose intensities deviate strongly from the normal ω4-dependence or by the observation of overtone progressions. Physical insight into the RRS process can be obtained from the Kramer-Heisenberg-Dirac (KHD) or sum-overstates formalism for α0ρσ in second-order time-dependent perturbation theory [9]:

Ultraviolet (UV) Raman Spectroscopy

133

a)

v = 5

v = 0

v = 1

v = 2v = 3

Energy

v = 0

Stokes v=4 v=2 v=2 v=1 Anti-stokes

v=0 Rayleigh

v = 5

v = 4

E1 E0

Virtual states

b)

Normal scattering

Internuclear separation

Energy

6

v = 1

v = 2v = 3

v = 4

E1 E0

Stokes Fluorescence

v=4 v=3 v=2 v=1 Resonance scattering Anti-stokes v=0 Rayleigh Internuclear separation

Fig. 6.1 Transitions for (a) normal and (b) resonance scattering on a Jablonski diagram. Note: The energies are not to scale 0 αρσ

¼

X r6¼i,f

"

pσ jii p jii pρ jrihr jb pσ jrihr jb hf jb hf jb ρ þ
ℏðωri
ωL Þ
iΓr ℏ ωrf þ ωL þ iΓr

#

ð6:2Þ pσ jii pbσ is the dipole operator in the σ-direction, hence hr jb is σth component of the electric dipole transition moment between initial state |ii and electronic excited state hr|; ωri is the frequency of that transition; and iΓr is a damping factor related to the lifetime of the excited electronic state. The remaining terms have the analogous meaning for the indicated states. From this equation we see that if the laser frequency matches the electronic transition frequency, then the first term in brackets is large and so is the polarizability derivative, α0ρσ . This can lead to enhancements of the Raman intensity as large as 108 [10]. The degree of resonance enhancement depends on the details of the vibrational and electronic states involved. Consequently, a resonance Raman spectrum may have a very different peak intensity pattern than the normal Raman spectrum, and this pattern can change with the excitation wavelength. In some cases resonance Raman spectra exhibit strong overtone progressions, unlike normal Raman spectroscopy. Particularly useful for characterization of catalyst materials is the ability to enhance the Raman spectrum for a subset of species in a mixture by taking advantage of differences in electronic absorption between species. It is even possible to obtain a measurement that is dominated by the Raman spectrum of a minority species that makes up only a few percent of the mixture. Reference [9] provides both rigorous mathematical and qualitative descriptions of the mechanisms for resonance enhancement. In general the resonance enhanced intensity

of a spectral feature depends on (1) the oscillator strength of the electronic absorption at the excitation wavelength, (2) whether the electronic transition is dipolar or vibronically allowed, (3) the magnitude of any change in the equilibrium normal coordinate, Q0, that accompanies the electronic excitation (e.g., bond length or structural distortion), and (4) any vibronic coupling between electronic states, both ground$excited and excited$excited. As originally formulated, three mechanisms for resonance enhancement were described designated A-term, B-term, and C-term [11, 12]. The characteristics of these mechanisms are 1. Dipole-allowed electronic transitions (ground to excited), e.g., π ! π* or charge transfer. A-term: ω(ground) 6¼ ω(excited) (vibrational frequency changes), Q0(ground) 6¼ Q0(excited) (structural change); totally symmetric modes enhanced, overtones enhanced when ΔQ is large. B-term: Excited state is vibronically coupled to another excited state through normal coordinate Q. Only important for non-totally symmetric modes and excited electronic states that are close in energy. 2. Vibronically coupled electronic transitions (ground to excited). C-term: Since the strength of vibronic coupling depends inversely on the energy separation between electronic states, this mechanism can be ignored because the energy separation between ground and excited states is typically large.

Reference [9] also designates a D-term mechanism that involves vibronic coupling of the excited electronic state to two additional excited states. This term is normally small and

6

134

P. C. Stair

species that represent a small minority of the catalyst material, even when there is significant overlap between the spectral features of minority and majority species. As an example, Fig. 6.2 shows Raman spectra measured for a catalyst composed of molybdenum oxide supported on silica [14]. The spectra on the left were measured using 488 nm excitation, while those on the right used 244 nm excitation. Under visible excitation the only detectable Mo species at high loadings of 20.1 and 30.2 wt% is crystalline MoO3. Under UV excitation, which is in resonance with a UV absorption band of dispersed Mo oxide, only the dispersed Mo species are detectable at high loadings. This result demonstrates that even when 99% of Mo is in MoO3 crystals, there is still dispersed Mo present, and it dominates the measured UV Raman spectrum. UV Raman is less susceptible to interference from thermal emission produced by samples at elevated temperatures. Thermal emission generally follows the wavelength dependence of a black-body radiator with negligible intensity at the ultraviolet wavelengths corresponding to the region occupied by UV Raman spectra. Finally, practitioners of UV Raman have noticed stronger intensities measured for bands at large Raman shifts, such as from OH and CH bonds, as compared to lower Raman shifts, such as from aromatic ring vibrations, in comparison to similar spectra obtained using visible wavelength excitation. It is not clear exactly why this is the case, but it is possibly due to the wavelength-dependent efficiency of the CCD detectors typically used in Raman spectrometers. Because of the elevated energy of UV photons, a full UV Raman spectrum covers a smaller range of wavelengths

can be ignored. From a practical point of view, the A-term resonance is likely to be the most important in the characterization of catalytic systems. In fact, the observation of resonance Raman in a catalytic system for which some species may be unknown can be used as a diagnostic tool for species identification.

6.2.2

Benefits and Limitations for Catalyst Characterization

Raman spectroscopy in any form provides a vibrational spectrum that spans the entire range of molecular and lattice vibrations. In principle, Raman “sees” everything, both the solid catalyst material and the molecular reagents in a single measurement, but the information it provides is limited by what can be deduced from a vibrational spectrum. For example, information about atomic structure is indirect (cf. EXAFS or XRD) and must rely on comparisons to well-characterized materials or molecules or on computational modeling. UV Raman, in particular, offers the benefit of avoiding interference from fluorescence, which can otherwise obscure the Raman spectrum [13]. This benefit can be crucial in the study of coke formation, as described later in this chapter. Ultraviolet excitation also increases Raman cross sections via the ω4-dependence of normal Raman scattering and/or by resonance enhancement, since many, if not most, materials absorb at the typical laser wavelengths used (16 times, and resonance enhancement of 1-methylnaphthalene was sufficient to make it the strongest contributor to the spectrum. The lower panel illustrates the analogous situation for vanadium oxide supported on alumina. Compared to the spectrum from pure alumina, the spectrum from a material containing only 0.03 mol% vanadium oxide on the surface exhibits dramatically attenuated alumina peaks and a peak due to the V¼O vibration at ca. 1010 cm
1 that is of comparable intensity due to resonance enhancement. Avoiding the problem of sample decomposition will be discussed in the section describing in situ and operando reaction cells.

6 6.3

UV Raman Instrumentation

The instrumentation components required to measure UV Raman spectra are specialized versions of the same components used for conventional Raman spectroscopy. Integrated instrument packages are commercially available with the capability to perform UV Raman spectroscopy at a few specific excitation wavelengths. Here we describe some of the variations that are specific to UV and resonance Raman measurements, with reference to Fig. 6.4. The excitation lasers most often used for catalysis studies are commercially available, continuous-wave (CW) Ar+ lasers that produce laser radiation at 514.5 nm or 488.0 nm and ultraviolet light at ca. 257 or 244 nm via frequency-doubling by a BBO (β-BaB2O4) nonlinear optical crystal mounted in the laser cavity. The helium-cadmium (HeCd) laser, operating at 325 nm, is also commonly used. The most versatile (and expensive) UV laser systems are based on second-, third-, and fourth-harmonic generation of a wavelength-tunable, Ti: Sapphire quasi-CW, pulsed laser system [16]. Using combinations of these lasers, excitation wavelengths in the range 210–900 nm are accessible to excite Raman scattering. The methods for steering and focusing of the laser beam on the sample and collecting the Raman scattered light depend on the nature of the sample cell used for in situ and operando conditions. If the cell and working distance are relatively large, with characteristic dimensions of 5–10 cm, then the focusing optic can be a simple, UV-transparent lens. If the cell is small (ca. 1 cm) then both focusing and light collection, are performed by a microscope objective. If excitation by multiple wavelengths is planned, then a reflecting microscope objective, which maintains a constant focal length, independent of the laser wavelength, is preferred. The collection optic for UV Raman scattered light in the 180 backscattering geometry can be the same optic used to focus the laser on the sample, e.g., a microscope objective. Separation of the incident and scattered light is performed by

136

P. C. Stair

user-friendly, possess capabilities not found in single-stage instruments. In addition to being able to tune the prefilter stages to block Rayleigh scattering at any excitation wavelength, the triple spectrometers can be operated so as to measure both Stokes and anti-Stokes Raman scattering at the same time in order to measure the sample temperature or so that the overall dispersion of the Raman spectrum is additive over the three stages to achieve a high resolution measurement.

Triple-stage spectrometer and CCD detector

Wavelength tunable (210–900 nm) laser

6.4

Ellipsoidal reflector

Heater in situ Raman cell Gas in

Gas out

Fig. 6.4 UV resonance Raman system consisting of wavelengthtunable excitation lasers, a fluidized bed reactor, an ellipsoidal reflector for light collection, and a triple-grating spectrometer. (After reference [16])

a narrow-band, reflecting edge filter, which also serves to remove Rayleigh scattering from the collected light before input to the spectrometer. For a long working distance setup, the collection optic is typically a separate, ellipsoidal mirror designed to match the scattered light cone to the numerical aperture of the spectrometer (Fig. 6.4) or to input to a fiber bundle that conveys the light to the spectrometer [17]. Removal of Rayleigh scattering from the collected light is essential for scattering from solid samples and can be performed by an edge filter when only a few, discrete laser wavelengths (e.g., 325, 257, 244 nm) are employed. If a wavelength tunable measurement is planned, then a “triple” spectrometer, such as that shown in Fig. 6.4, is required. The first two stages in the “triple” serve to remove the Rayleigh scattered light, while the third state presents the dispersed Raman scattering to a CCD detector. UV Raman spectra are dispersed by an aberrationcorrected spectrometer using a high groove density (>2400/ mm), holographic grating, and detected by a UV-sensitive CCD detector. As mentioned above, Rayleigh scattered light, at the wavelength of the excitation laser, is attenuated by a factor ~10
6 using either a long-pass edge filter when using a single-stage spectrometer or the first two stages of triple-stage spectrometers operating in subtractive mode. Overall efficiencies of ca. 15% are achievable with currently available commercial instruments, even with the triple-stage instruments. Triple-stage spectrometers, while often not as

New Instrument Advances

The ability to excite Raman scattering at multiple wavelengths makes possible the selective detection of multiple species in a mixture via resonance enhancement. While tunable laser systems offer the most flexibility, they are large, complex instruments that require significant skill to operate. Advances in hollow cathode technology offer reliable, high power deep UV lasers at discrete wavelengths (e.g., 224.3 or 248.6 nm) in a compact package. Edge filters corresponding to these excitation wavelengths have also recently been developed. Advances in the technology of in situ and operando cells as well as fiber optic coupling of both the laser and the Raman scattered light can be anticipated. The combination of these technologies has the potential to make UV Raman applicable to more realistic catalytic reaction conditions of high pressure and temperature as well as industrial installations for process monitoring.

6.5

Reaction Cells

Because of the high energy nature of the ultraviolet photons used for UV Raman spectroscopy and the high probability for decomposition of molecular samples, special handling during measurements has been a requirement since the early days of the technique. The issues are two-fold. First, many materials absorb strongly at the wavelengths employed for UV Raman spectroscopy (e.g., 244 nm), which leads to localized sample heating even at the low laser powers used in typical measurements (~1 mW). Second, many materials will rapidly undergo photochemical degradation under the UV irradiation. Only thermally and photochemically robust samples can be measured without special precautions. A common method used to avoid sample transformations due to heating is referred to as the “spinning disk.” Solid samples are pressed into the shape of a disk and mounted on a rotating platform with the axis of rotation perpendicular to the plane of the disk. With the excitation laser incident on the face of the disk at some distance off the rotation axis, the irradiated spot traces a ring whose length and thickness are

6

Ultraviolet (UV) Raman Spectroscopy

137

determined by the radial distance from the rotation axis and the diameter of the focused laser spot. This geometry mitigates against sample heating by spreading the absorbed energy over the area of the ring and by allowing the sample surface to cool between passes under the beam. The impact on heating can be appreciated by comparing the time average incident power density for a 0.1 mm diameter beam of 10 mW power. For the stationary spot the power density is ~130 W/cm2, while for the spinning disk with the laser 3 mm off the rotation axis the power density is ~0.5 W/cm2. The power density on a spinning sample can be further reduced by translating the sample in a direction perpendicular to the rotation axis so that the laser spot moves from the center to the edge of the disk and back, impinging over a spiral area on the sample disk. The effectiveness of these methods was compared in an early study of alumina-supported metal oxide materials [18]. Obviously, if the speed of measurement is not an issue, it is always possible to reduce the laser power in order to avoid heating. One limitation of the spinning sample approach has been the complexity associated with implementing the spinning motion when the sample is in a reactor for in situ and operando measurements. An innovation that avoids problematic mechanical coupling to the sample by implementation of magnetic coupling was recently reported in Reference [19]. A schematic diagram of the sample holder is depicted in Fig. 6.5 [19]. The success of this approach was demonstrated by spectra taken of anthracene using 244 nm excitation, with and without spinning. Under spinning there was no obvious buildup of decomposition products, and a spectrum of intact anthracene was obtained. When a sample is sensitive to photochemical damage, spinning may not be sufficient. The problem stems from the fact that there is no recovery from photochemical damage as there is from transient heating. The sample transforms via reactions driven by single photon absorption events so that even at very low laser powers degradation products

accumulate. In this case degradation cannot be avoided, but its effect on the measured spectrum can be minimized by introducing a continuous supply of fresh sample during the measurement. For biological or liquid samples this can be done using a flow geometry [20]. In principle the same approach could be used for solid catalyst materials; however, what amounts to the same thing can be accomplished using the fluidized bed Raman cell shown schematically in Fig. 6.6 [21]. The cell consists of a double-wall quartz reactor with reagent gases flowing up the center tube and down the space between tubes and an optical window fused to the top. The tubes are sealed against atmosphere to metal flanges at the bottom using O-rings. The flanges are bored to allow connections to inlet gases and outlet analytical instrumentation. The catalyst takes the form of a loose powder supported by a porous disk near the top of the center tube. The sample in Fig. 6.6 was H-USY zeolite that had been pressed into a pellet and then lightly ground and sieved to a particle size of ~0.1 mm to facilitate powder movement. Using a combination of gas flow control and horizontal vibration, driven by an electromagnetic shaker, the catalyst power moves continuously. A 1 mm deep bed provides a sufficient amount of catalyst to ensure that photochemical damage is mitigated by dilution with undamaged material. Implementation of heating and selection of desired reagent gases makes the fluidized bed cell suitable for in situ and operando experimentation. One important disadvantage of the fluidized bed is movement of the catalyst material in and out of the collection optic focal plane, causing a decrease in light collection efficiency. The severity of this effect depends on the numerical aperture of the light collector because numerical aperture and depth of field are anti-correlated. When there are no concerns about degradation of robust samples, the cell can be used without fluidization with a corresponding improvement in signal gathering efficiency.

Scattered photons

cⴕ)

244 nm laser Zeolite bed

e) Furnace

c)

Porous disc

d)

Electromagnetic shaker

b)

a) 6 mm

Fig. 6.5 Magnetically coupled spinning sample holder: (a) stainless steel cylinder; (b) magnet; (c) gold foil for holding sample pellet; (d) mounting strip; and (e) sample pellet. (Reprinted by permission from Springer Nature Customer Service Centre GmbH: Springer Nature, Signorile et al. [19], Copyright 2018)

Gas out Gas in

Fig. 6.6 Schematic diagram of a fluidized bed reaction/Raman cell. (Reprinted from Chua and Stair [21], Copyright 2000, with permission from Elsevier)

6

138

P. C. Stair

The success of the fluidized bed approach can be seen from the data presented in Fig. 6.7 [21]. Spectrum (a), recorded using the spinning disk, shows the UV Raman spectrum of intact naphthalene. Spectrum (b) was obtained from a stationary sample of H-USY zeolite that had been impregnated with naphthalene. While there are some features detectable in spectrum (b), they are clearly not from molecular naphthalene. Spectrum (c) shows the result of spinning the naphthalene/H-USY sample. The spectrum is of rather poor quality, and the broadening of the peak at 1620 cm
1 to higher values is characteristic of polycyclic aromatic hydrocarbons that form as a consequence of naphthalene decomposition. Spectrum (d), taken using the fluidized bed cell, is clearly that of intact naphthalene, by comparison to spectrum (a). It should be noted that the systematic shifts of the naphthalene peaks to lower wavenumber for naphthalene/H-USY compared to solid naphthalene are also seen for naphthalene dissolved in a hydrocarbon solvent. We have used the fluidized bed Raman cell successfully for nearly 20 years. However, it is homemade and not easily

1373

1623

d)

0.087 cts/s.mW 760

1020

6.6

1620

c)

Intensity

1.9 cts/s.mW 1390

6.8 cts/s.mW 1625

b)

1626

a)

1.6 cts/s.mW 1384 768

800

1029

interfaced to current Raman instrumentation due to the long working distance. Recently, the clever modification of a commercial in situ and operando cell was described, which makes possible the fluidized bed approach in a commercial Raman instrument [22, 23]. A schematic diagram of the cell is shown in Fig. 6.8 [22]. The reactor is a catalyst cell available from Linkam Scientific Instruments, model CCR1000. The overall flow direction is down, through the catalyst bed, but fluidization is obtained “via a micro-device [membrane pump], placed downstream from the reactor that provides pressure oscillations in such a way that the flow direction is reversed in discrete pulses of about 40–100 Hz [22]. A very nice video showing the particle motion is included in the Supporting Information to reference [22]. Another strategy is to mount the sample cell on an X-Y stage and move the sample continuously. This approach has been demonstrated by Christian Hess and results in a combination of reaction cell and collection optics with outstanding sensitivity (Fig. 6.9) [24, 25]. Whether sample translation is sufficient to mitigate interference from photochemical decomposition will depend on the system. A best case would be for an oxidation reaction, such as oxidative dehydrogenation, where the exposure to oxygen at elevated temperatures is expected to clean the catalyst of organic photochemical decomposition products.

1155 1200

1580

1600

Raman shift (cm–1)

Fig. 6.7 (a) Spectrum of naphthalene recorded on a spinning disc. (b) Spectrum of naphthalene/H-USY recorded on a stationary disc. (c) Spectrum of naphthalene/H-USY recorded on a spinning disc; (d) Spectrum of naphthalene/ZH-USY recorded using the fluidized bed. (Reprinted from Chua and Stair [20], Copyright 2000, with permission from Elsevier)

Chronology of Application to Catalysis

The first measurements of deep UV Raman spectra from catalyst materials were performed in the spring of 1995, and the first reports were published by Can Li and Peter Stair in 1996 [8, 26]. Almost immediately the problem of photochemical sample degradation was recognized as a significant barrier to the application of UV Raman to catalyst systems containing molecular reagents. Interestingly, the formation of degradation products was observed even for alkanes absorbed in zeolites despite the lack of optical absorption by the sample. In mid-1999 the idea for the fluidized bed system was developed and tested, with the first publication submitted in spring 2000 [20]. While there was certainly a scientific driver, the desire to avoid photochemical degradation required the coupling of UV Raman to in situ reaction cells, which could then be readily adapted for operando operation, i.e., under reaction conditions, starting around the year 2000. The development of small volume, commercial reaction cells has enabled UV Raman measurements with commercial instruments. However, the problem of photochemical degradation has been a barrier to the application of UV Raman in these commercial cells under reaction conditions where molecular transformations can be driven by the laser. Another barrier has been the time required to obtain a spectrum, which is at best ~60 s and sometimes hours. As a consequence UV Raman has been

6

Ultraviolet (UV) Raman Spectroscopy

139

a)

b) Circular ceramic heating element

Optical window

Sample holder

In gasflow (educts)

“Back puls” (40–100 Hz)

Catalyst particles (150–300 μm)

gas Out

In gas

In gasflow (educts)

6

Circular channel for cooling liquid Electric connections for heating and thermocouple

Out gasflow (products)

Fig. 6.8 (a) Sketch of the Linkam CCR1000 reactor. (b) Schematic drawing of the sample holder with gas flow directions indicated. (Reprinted from Beato et al. [22], Copyright 2013, with permission from Elsevier)

Spectrometer entrance slit Spherical mirror Laser beam Parabolic mirror A

Parabolic mirror B Sample

Tunable laser system

Customized mirror system

Triple-stage spectrometer

CCD camera

Reaction cell

Fig. 6.9 Optical beam path for sample cell mounted on a movable stage. (Reprinted from Waleska et al. [24] © 2016 American Chemical Society)

primarily applied to catalytic systems under steady-state conditions or by interrupting the reagent feed.

6.7

Time Resolution

As alluded to above, the typical time required to obtain a UV Raman spectrum with sufficient signal-to-noise is of order 2 min for a strong Raman scattering species at high concentration [19]. Collection times are often not reported in the

literature because the measurement conditions are selected so that the catalyst material is not changing or changing very slowly. In our laboratory, collection times are typically 5–60 min for measurements on an aromatic hydrocarbon adsorbed in a zeolite, using a fluidized bed reaction cell coupled to a single-stage spectrometer [27, 28]. In our experience there is ca. a 5X loss of intensity that accompanies the fluidization process. More recent reports on analogous samples involving aromatic hydrocarbons cite similar collection times [23]. Raman measurements on supported metal oxide catalyst materials using the home-built reactor require collection times of 10 min to 6 h [14, 29]. In what represents the current state of the art, a time-resolved UV resonance Raman measurement with collection times of 1 min monitored changes in vanadium oxide species that correlated with gas composition during ethanol oxidative dehydrogenation [25].

6.8

Spatial Resolution

In principle the spatial resolution of UV Raman can be diffraction limited. The smallest reported beam size is 1 μm [30]. More typical beam sizes are reported to be 10–100 μm [31]. It should be clear from the previous comments about the danger of thermal and photochemical sample decomposition that sample motion is required to avoid artifacts when there are organic molecules in the sample system. Consequently, the ability to map species or material phases by scanning UV Raman microscopy is limited to robust materials.

140

6.9

P. C. Stair

Applications of UV Raman

The bulk of UV Raman studies in catalysis have been performed at five laboratories: Northwestern University/ Argonne National Laboratory (Peter Stair), Dalian Institute of Chemical Physics (Can Li), Oak Ridge National Laboratory (Zili Wu), Technical University of Darmstadt (Christian Hess), and University of Torino (Sylvia Bordiga). The applications described here are drawn from publications over the last 10 years from these laboratories covering the subjects of zeolite synthesis, catalyst deactivation, and the speciation and catalytic roles of metal oxides. Additional applications and descriptions of UV Raman studies can be found in a number of reviews published over the past 20 years [13, 16, 32–39].

6.9.1

while at 325 nm the spectra emphasized the silica species. In this way it was possible to study different aspects of the synthesis system, i.e., the coordination of Fe and the silica structural units, during the condensation and crystallization process. The results led to a proposed mechanism [42]. The ability to excite resonance Raman scattering from FeO-M structures is one of the important applications of UV Raman spectroscopy in catalytic materials [43–47]. Fe-containing zeolites often contain only small concentrations of iron, making their spectroscopic investigation very difficult. Resonance enhancement not only increases the intensity from iron species sufficiently to make them detectable, but the fact that other vibrations in the material are not enhanced makes the bands due to iron standout against the background of majority species.

Silica/Zeolite Synthesis 6.9.2

UV Raman spectroscopy in combination with NMR and computational modeling was used to monitor the silica species formed, in real time, during sol gel formation from tetraethylorthosilicate (TEOS) in acidic aqueous media [40]. The application of UV Raman was selected to achieve lower backgrounds and increased signal-to-noise. Experiments were performed at three water/TEOS ratios (0.2, 0.7, 1.2) revealing the formation of linear chains of 3–6 Si atoms, branched structures, and rings with 4–6 Si atoms. A gel formation mechanism involving oligomer condensation, rather than oligomer growth, was deduced. For the most part UV Raman played a supporting role to the NMR characterization, confirming the species present from NMR analysis; however, the synergistic combination of the two methods strengthened the species assignment of NMR shifts and Raman bands. Immediately following reference [40] was a report about gel formation from tetramethylorthosilicate (TMOS) as the Si precursor, using the same combination of NMR, UV Raman, and computational modeling [41]. The gels produced from the two precursors exhibit different porosity with microporosity derived from TEOS and mesoporosity derived from TMOS. The studies found differences in the condensing species with TEOS providing oligomers with ring structures and TMOS more branched and fewer ring structures. The iron-containing zeolite Fe-ZSM-35 is an interesting catalyst for oxidation and isomerization reactions; however, it is difficult to prepare so as to avoid extra-framework Fe. A new method, which promises to reduce or eliminate the extraframework Fe, involves two steps: (1) Hydrothermally synthesize partially crystallized Fe-containing silicate species and partially crystallized ZSM-35. (2) Combine the two partially crystallized gels and complete the crystallization process. UV Raman, combined with XRD and UV-Vis, measurements was excited at 266 nm and 325 nm to follow the speciation during each of the two steps [42]. At 266 nm resonance Raman spectra from the Fe species were obtained,

Thermal, Oxidation, and Reduction Treatments

UV Raman spectroscopy has been used to study the influence of various catalyst treatments typically employed for activation. Examples of its application to treatments such as calcination, oxidation, and reduction will be given here. Perhaps the most well-known example on thermal treatment is the dehydration of supported metal oxide catalysts by calcination. The observation that supported oxide Raman spectra change under dehydration goes back to classic work by Israel Wachs using conventional, visible Raman spectroscopy [49]. In later work the spectra from visible and UV excitation were compared from a variety of supported metal oxides to determine the extent of dehydration produced by strong absorption of the UV light and formulate experimental strategies to mitigate laser heating [18]. These early studies showed how to control laser-induced thermal effects and clarified the experimental procedures required for meaningful measurements at temperature. The influence of hydrogen reduction and reoxidation on the nature of vanadia species supported on alumina was studied by UV and visible Raman spectroscopy [48]. The sample contained 1.2 V/nm2 loaded onto θ-alumina by incipient wetness impregnation. After drying and calcination at 550  C in air for 4 h the sample was placed in the fluidized bed Raman cell and further oxidized prior to the hydrogen reduction. The sample was reduced at progressively higher temperature in 5% H2/N2 with UV (244 nm) and visible (488 nm) Raman spectra measured after 1 h at each temperature. The resulting spectra are shown in Fig. 6.10. The position of the V¼O band in UV (1021 cm
1) and visible (1034 cm
1) differs as a consequence of species-specific resonance enhancement of monomers and polymers, respectively. With increasing hydrogen reduction temperature the V¼O intensity decreases in both sets of spectra. In the visible-excited spectra the band red-shifts progressively to

Ultraviolet (UV) Raman Spectroscopy

a)

141

b)

l = 488 nm 910

1021

620

865

910

l = 244 nm

1023 1034

6

Red-873 K

Red-873 K

Intensity

Intensity

Red-973 K Red-773 K Red-673 K Red-773 K Red-573 K Red-673 K

Red-473 K Red-373 K

Red-573 K Cal-823 K Cal-823 K 900 400

600

800

1000

1200

Raman shift (cm–1)

975 Raman shift

1050

(cm–1)

Fig. 6.10 (a) UV Raman and (b) visible Raman spectra of vanadia/alumina catalysts with 1.2 V/nm2 reduced H2 at various temperatures. (Reprinted from Wu et al. [48] © 2007 American Chemical Society)

1023 cm
1 at 400  C and then disappears. The V¼O persists in the UV-excited spectra up to the highest temperature studied, 700  C. The differing behavior at the two excitation wavelengths reflects the differing species detected. The polymeric species detected under 488 nm excitation are more easily reduced than the monomers detected under 244 nm excitation. Moreover, the shift in the band seen in panel (B) is consistent with a population of various polymer sizes with the largest polymers being the easiest to reduce. Under 244 nm excitation, a new broad band at 870 cm
1 appears at 500  C, which is assigned to a V-O-V vibration based on a comparison of UV Raman spectra from bulk VO2 and V2O3 [48]. The supported vanadia Raman spectra are restored to their original form by reoxidation in 5% O2/N2. The spectra, measured following a similar protocol to the samples under hydrogen, are shown in Fig. 6.11. These experiments demonstrate the reversibility of the supported vanadia structure after cycling between oxidation and reduction and that the redox properties depend somewhat on the nature of the vanadia species.

6.9.3

Catalyst Deactivation by Coke Formation

One of the original motivations for development of UV Raman was to avoid fluorescence interference from carbonaceous deposits (coke) that can accumulate on a catalyst during hydrocarbon transformations [33, 51]. The ability to

perform Raman measurements with coke on the catalyst opened the door to vibrational spectroscopy studies of their formation and speciation. Early work in our laboratory focused on the methanol to hydrocarbons (MTH) reaction catalyzed by H-ZSM5 [18]. While detection and identification of cyclopentadiene as a key intermediate in the formation of coke was accomplished, due to the complexity of the coke formation reactions and the high rates at which they occur UV Raman was unable to follow the progress of the chemistry in real time, under reaction conditions. Instead, in a later study of butane dehydrogenation by vanadium supported on alumina, the catalysts were exposed to the hydrocarbon reagent at reaction temperature for 30 min intervals, cooled to room temperature, and measured. Following this heatsoak-quench procedure at successively higher temperatures, the formation and evolution of carbonaceous deposits was observed [50]. Anticipating that product olefins from butane dehydrogenation are the starting point for coke formation, the adsorption and reaction vs. temperature of 1-butene, cis/ trans-2-butene, and 1,3-butadiene on a supported vanadia/ alumina catalyst were examined, with the results shown in Fig. 6.12a–c, respectively. Each of these olefins produced a set of bands at 850, 1006, 1181, 1375, 1440, 1501, and 1604 cm
1 after heating to 400–500  C, and, with the exception of the band at 1375 cm-1, the same bands are produced by butane dehydrogenation at 400  C (Fig. 6.12d-(b)). This similarity points to the common chemistry performed by butane and the C4 olefins. From the observation that these

6

142

P. C. Stair

b)

l = 488 nm

Reox-873 K

Intensity

Intensity

Cal-823 K

1032

910

1024

l = 244 nm 905

a)

Cal-823 K Reox-873 K Reox-773 K

Reox-773 K

Reox-673 K

Reox-673 K

Reox-573 K

Reox-573 K

Reox-473 K Reox-300 K Red-873 K

Red-873 K 400

600 800 Raman shift (cm–1)

1000

1200

850

900

950

1000

1050

Raman shift (cm–1)

Fig. 6.11 UV Raman (a) and visible Raman (b) spectra of 1.2 V/nm2 on alumina reduced at 600  C, then reoxidized at various temperatures. (Reprinted from Wu et al. [48] © 2007 American Chemical Society)

bands were produced at room temperature on adsorption of 1,3-butadiene, the possibility that they could be due to polystyrene was considered. Figure 6.12d-(a) and (c) show the Raman spectra from styrene and polystyrene formed by heating the catalyst in the presence of styrene. The nearly identical band positions and intensities confirm the formation of polystyrene on the vanadia/alumina catalyst whether the starting reagent is butane or a C4 olefin, pointing to polystyrene as a key intermediate in coke formation in this system. In the laboratory of Silvia Bordiga there have recently been renewed efforts to examine the formation of carbonaceous species in zeolites during the MTH reaction using novel in situ and operando reaction cells [23, 52–54]. First, in reference [52], Raman spectra were obtained from a set of polycyclic aromatic hydrocarbons (PAH) as solids, in solution, and adsorbed on silica or carbon, using excitation lasers in the visible and UV. The spectra demonstrated that the characteristic peaks from PAHs were essentially independent of their environment and could be reliably used as reference spectra for the same molecules located in or on a zeolite catalyst. UV Raman was then used to examine the carbonaceous deposits produced during MTH catalyzed by Mordenite, ZSM-5, Beta, ZSM-22, and SAPO-34 [53]. The carbons were classified according to whether their Raman spectra exhibited the broad, characteristic features of extended, 2D graphitic structures or sharper, well-defined peaks characteristic of smaller, more molecular species.

Interestingly, there was a clear correlation between the classification of the carbons and the topology of the zeolite. Mordenite, ZSM-5, and Beta, with 3D, medium-large pores produced the graphitic carbon. ZSM-22 and SAPO-34, with smaller, 1D pores, produced the more molecular carbons. Since graphitic carbon species are too large to fit in even the large pore zeolites, the authors deduced that it must be located on the external surface of the zeolite crystals. They attributed the absence of graphitic structures on the smaller pore ZSM-22 and SAPO-34 materials to restricted diffusion of the molecular precursors to form graphitic carbon [53]. Later, authentic operando experiments were performed on the same set of five zeolites for the first 60 min on stream using the fluidized bed cell described in reference [22] and employing a mass spectrometer to monitor product formation [54]. The results confirmed the classification of deposited carbons in reference [53] but with some additional detail about the presence of molecular species formed early in the reaction. Finally, in a very recent publication the Torino group has investigated the influence of reactor geometry to obtain new information on the variation in carbon deposition that mimic changes in the process between fixed bed and fluidized bed reactors [23]. It is likely that further experimentation along this direction will make it possible to better understand the deactivation behavior of particular reactor configurations.

143

873 K 773 K 673 K 573 K 473 K 373 K 298 K c-2-butene (g)

573 K 473 K 373 K 1-butene (g)

1640

1640

773 K 673 K

1180

852

1385 1440 1502

1173

1003

851 Intensity

873 K

1003

b)

1603

a)

1380 1447 1502 1604

Ultraviolet (UV) Raman Spectroscopy

Intensity

6

Fresh

Fresh

6

1603 1437 1501

c) 1.2V + styrene heated at 393K b) Butane DH673K 1320 1415

1636

1183

773 K 673 K 573 K 473 K 373 K

1298

1800

1636

850

1181

1006

1375 1446 1502

873 K

Intensity

850

900 1200 1500 Raman shift (cm–1)

d)

1605

c)

600

1800

1500

1002

900 1200 Raman shift (cm–1)

Intensity

600

t-2-butene (g)

298 K 1.3-butadiene (g)

a) Styrene

Fresh 600

900 1200 1500 Raman shift (cm–1)

900

1800

1200 1500 Raman shift (cm–1)

1800

Fig. 6.12 Raman spectra of 1-butene (a), cis/trans-2-butene (b), 1.3butadiene (c) subjected to dehydrogenation (DH) on oxidized 1.2 V/nm2 vanadia/alumina at different temperatures. (d) Comparison of Raman

spectra from styrene (a), butane DH at 400  C (b) and a mixture of the vanadia/alumina and styrene at 150  C. (Reprinted from Wu and Stair [50], Copyright 2006, with permission from Elsevier)

Additional reports have studied the formation of coke during methane dehydroaromatization catalyzed by Mo/ HZSM-5 [55], during side chain alkylation of toluene catalyzed by Cs- and Na-modified zeolite X [56, 57], and during HZSM-5 catalyzed methanol dehydration [58]. Notable here is reference [58] where measurements were made at both the inlet and the outlet of a fixed bed reactor. At a low reaction temperature, 150  C, the Raman spectra at the beginning and end of the bed are essentially the same and interpreted as “soft coke.” When the reaction temperature is raised to 200  C, the Raman spectra at the beginning of the bed suggest the formation of more extended PAH structures, interpreted as “hard coke.” At the end of the bed the nature of the coke remains the same as at the lower temperature. It was concluded that the key coke precursor is the methylbenzenium cation, whose formation at the end of the

catalyst bed is suppressed by the presence of water produced by the dehydration reaction.

6.9.4

Speciation of Titania

One of the powerful features of UV Raman spectroscopy is the possibility for resonance enhancement of the Raman signal from key, minority species in a catalytic system. This was demonstrated for the first time by the group of Can Li with titanium silicalite, TS-1, [59] the prominent catalyst used for selective oxidation reactions using hydrogen peroxide [60]. Clear signatures of framework titanium: Bands at 490, 530, and 1125 cm
1 were obtained from TS-1, with 244 nm excitation, which overlaps a prominent UV charge transfer absorption in this material (see Fig. 6.13). The bands

144

P. C. Stair

220

a)

244

were not observed from pure silicalite or by excitation of TS-1 at 325 and 488 nm. Spectra using the latter wavelengths only exhibited bands from silica and anatase TiO2. Recent UV Raman studies of TS-1 have focused on the identification of the active site for propylene epoxidation [61, 62] and on optimizing the synthesis of TS-1 [63]. Reference [63] is notable for its focus on identifying multiple Ti-species via a multi-wavelength excitation approach. UV Raman spectra at 325, 266, and 244 nm were obtained from solids formed at various times during the synthesis of TS-1 in order to identify the TiOx species that lead to the deleterious amorphous TiOx that decomposes HOOH under propene epoxidation conditions [64]. Measurements at 244 and

A 244

TS-1

Silicalite-1

200

300

400

500

l (nm)

O Si

O Ti

Si

a)

b)

H

960

815

290

380 490 530

1125

b)

325 nm from 2 h onward did not show changes during synthesis that elucidated the species of interest. Measurements at earlier times, when combined with measurements after 2 h, reveal the appearance of a 695 cm
1 band appearing between 110 and 120 min. This band is associated with the TiO6 species that leads to amorphous titania. Based on this evidence, a model was suggested where the octahedral titania species is squeezed out of the bulk as the MFI crystalizes. By slowing down the crystallization, the formation of the octahedral species was diminished. A material related to TS-1 is simply titania supported on amorphous silica. Because the support structure is ill-defined, compared to silicalite, the potential range of titania species and structures is much broader than for TS-1. Consequently, there must be more emphasis on computational modeling of measurements. An exemplar of this approach is the work from the group of Christian Hess [65, 66]. In reference [65] two models for the titania/silica structure, based on Ti(OH)4 grafted to a polyhedral oligomeric silsequioxane (POSS) structure, were used to perform normal coordinate calculations for the assignment of Raman bands (see Fig. 6.14). Raman bands at 945 and 1045 cm
1 were identified as interphase Ti-O-Si vibrations with force constants dominated by (Si-O)Si-O-Si and 1085 cm
1 as a sym/asym Ti-O-Si vibration with a force constant having substantial contributions from (Ti-O)Ti-O-Si and (Si-O)Ti-O-Si, where the subscripts designate the structural unit containing the M-O bond. Regrettably, the vibrational frequencies produced by the two model structures are too similar to be distinguished, suggesting that Raman spectra cannot be used to determine the structure in an experimental system. The Hess laboratory has gone on to study the oxidative dehydrogenation of ethanol by the silicasupported titania system [66]. UV Raman data show the effect of both dehydration and ethanol reaction on the population of Ti-O-Ti, Ti-O-Si, and Ti-OH structures that is

O

H

O H

O

TS-1

I

Ti

H

O

Ti O Si

O Si

O Si O

Si

Si

1250 ~ Δn (cm–1)

750

1750

Fig. 6.13 UV-Vis diffuse reflectance (a) and UV Raman spectra (b) of titanium silicalite (TS-1) and pure silicalite-1. (Reprinted from Li et al. [59] © 1999)

Si

Si

O

O O

O

250

O O

Si

O

Silicalite-1

O

O O

Si

Si

O

O

Si Si

O

Si

O

OO O

Si

O

O

Si

Si

i

O

O

Fig. 6.14 Models for monomeric titania species on silica with 3 support bonds (a) and 2 support bonds (b). (Reprinted from Nitsche and Hess [65], Copyright 2014, with permission from Elsevier)

6

Ultraviolet (UV) Raman Spectroscopy

consistent with a mechanism whereby breaking a Ti-O bond in either Ti-O-Ti or Ti-O-Si occurs upon adsorption of ethanol leading to formation of ethoxy bound to Ti.

6.9.5

Supported Vanadium Oxide

Supported vanadium oxide catalyst materials have been studied intensively for many years because of their importance in selective oxidation and pollution prevention applications [67]. These materials have been characterized extensively by normal and UV resonance Raman spectroscopy, in part, because the supported species often include the V¼O bond, which appears as a narrow band at Raman shifts >990 cm
1 where there is little or no interference from other metaloxygen vibrations. In addition, supported vanadia catalysts are often used for oxidations where the reaction conditions keep the catalyst clean of coke and adsorbed hydrocarbons that might otherwise obscure the catalyst Raman spectrum. The recent literature on vanadium oxide supported by silica, alumina, and ceria provides excellent examples of how UV and resonance Raman spectroscopy can provide detail about supported vanadia structures and their role in catalysis. Many of the Raman results for VOx/Al2O3 are discussed in a review published in 2010 [16]. At loadings below a saturated monolayer, fully oxidized, supported vanadium oxide adopts a distorted tetrahedral configuration, with one vanadyl oxygen and three bridging oxygens having the chemical structure O¼V(-O-)3. At higher loadings crystalline V2O5 forms, which is easily identified by a sharp band at 993 cm
1. For the chemical structure O¼V(-O-)3 there are a number of atomic structures, e.g., O¼V(-O-Al)3, O¼V(-OAl)2(-OH), O¼V(-O-Al)2(-O-V), etc. By taking advantage of resonance enhancement effects it was possible to distinguish, and even quantify, monomeric vanadia species, O¼V(-OAl)3, from oligomeric or polymeric species, O¼V(-OAl)2(-O-V) or O¼V(-O-V-)2(-O-Al) [29, 48]. Using multiple wavelength resonance Raman measurements in combination with computational modeling it has also been possible to distinguish different monomer structures, e.g., O¼V(-OAl)3 and O¼V(-O-Al)2(-OH) [68], and show that the different structures exhibit differences in their temperature for hydrogen reduction [69]. Interestingly, despite differences in reducibility by hydrogen, there is at most a modest impact of the structure on the activity and selectivity for catalytic oxidative dehydrogenation of alkanes to olefins [70, 71]. Dehydrated VOx/SiO2 has also seen extensive investigation by UV Raman as a model supported catalyst where the support is expected to be relatively inert [24, 25, 72–76]. For the vanadia/silica system there has been significant controversy over the vanadia Raman band assignments [72–74, 77]; however, a recent normal coordinate analysis resolved the controversy using a model built upon the well-understood

145

POSS structure as a model for the silica support and showing that an earlier model was not sufficiently large to properly include the silica structure in the normal coordinate analysis [75]. Still unresolved are the VOx structures. Two monomeric VOx structures were proposed in reference [72], corresponding to the chemical formulas O¼V(-O-Si)3 and O¼V(-O-Si)2(-OH), having tripodal and bipodal coordination to the silica support. The evidence for two structures was the observation of two bands, with majority V¼O character, at 1032 and 1041 cm
1. The 1032 cm
1 band intensity dominated when Raman scattering was excited at 244 or 325 nm and was identified as resonance enhanced from the appearance of strong overtone bands. Under 442 or 532 nm excitation the 1041 cm
1 band dominated; the 1032 cm
1 band was undetectable; and the spectra were excited by normal Raman scattering. The lower and higher frequency bands were proposed to be from bipodal and tripodal structures, respectively. Fitting of Raman spectra from VOx/ SiO2(SBA-15) excited at 217.5 nm also produced two bands in the relevant frequency region but at 1020 and 1037 cm
1 [75]. Moreover, the normal coordinate analysis of reference [75] produced the opposite structural assignment, i.e., the low frequency band comes from the tripodal species and the high frequency band from the bipodal species. Finally, by combining detailed UV-Vis [76] or XPS [24] with UV Raman for VOx supported on SBA-15 or planar SiO2/Si(100), respectively, the Hess laboratory examined the case for formation of dimers and oligomers. They concluded that monomers and multimers of VOx on silica coexist, even at extremely low loadings (Fig. 6.15) [76]. How and to what extent the various structural forms participate in catalytic turnovers remains uncertain, however the hydrogen reduction data presented in reference [72] would seem to suggest that all structural forms undergo reduction/oxidation and can participate. The final system we consider is vanadium oxide supported on ceria. References [78] and [79] from the group at Oak

Ia O V O O O

Ib O V O

O

IIb O

O V

O

V O V O O O

IIa OH

O

O

O

O

V

V O

O

IIc O

O V O V O O

SBA-15

Fig. 6.15 Proposed structures for highly dispersed vanadia species on silica under dehydrated conditions. (Reprinted from Nitsche and Hess [76] © 2016)

6

146

P. C. Stair

Ridge National Laboratory examined the structure of vanadium oxide supported on ceria polycrystalline and nanoshaped supports, respectively, using a combination of IR, Raman, and UV-Vis spectroscopies. The supported VOx species were only detectable using visible wavelength excitations (442, 532, and 633 nm) due to strong absorption by the ceria support in the UV. Raman spectra excited at 325 nm were useful for the monitoring of ceria defect sites and demonstrating the inverse correlation between the population of defect sites and the loading of vanadia. On the polycrystalline ceria support the VOx Raman spectra showed the presence of monomers and multimers at all loadings, even as small as 0.05 V/nm2. On the crystalline nanoshapes, the ceria surfaces stem from terminations along only (100), (110), and (111) orientations. The impact on vanadia speciation is the absence of multimers at low vanadia loading, 0.1 and 0.5 V/nm2. Only monomers are detected. This shows a direct relationship between the surface orientation of the support and the vanadia speciation.

6.9.6

Ceria Support

Comparison of the Raman spectra excited at 532 and 325 nm demonstrate an enhanced sensitivity for defects marked by the so-called D band at 592 cm
1 under UV excitation as a consequence of resonance Raman scattering [80]. The intensity of this band, relative to the bulk F2g band, provides the means to monitor the presence and population of ceria structural defects. The results show a dependence of the defect population on the shape of the ceria crystals: rods > cubes > octahedra. The sensitivity to hydrogen reduction temperature followed the same trend. Spectral changes accompanying adsorption of O2 at room temperature indicate that the defects derived from hydrogen reduction are associated with surface oxygen vacancies. Reference [80] laid the foundation for subsequent investigations of the role ceria defects play in adsorption and reaction of methanol [81] and in oxidation reactions catalyzed by supported Pt [82–84] and Au [85, 86].

6.10

Multimodal Operation

The only sample and environmental requirements for UV Raman measurements are an optical window that is transparent to ultraviolet wavelengths, such as fused silica or sapphire, and the means to move the sample in order to minimize laser-induced thermal and photochemical sample changes. It is certainly possible to observe reaction products via the usual gas monitoring mass spectrometer or gas chromatograph detection. Depending on the reaction cell configuration, one can imagine the incorporation of IR and UV-Vis measurements in addition to UV Raman. By making use of UV-transparent fiber optic bundles, it should also be possible

to perform UV Raman measurements at hard-to-reach sample locations, such as at a synchrotron beamline.

6.11

Conclusions and Future Outlook

It should be evident from this short narrative that UV Raman spectroscopy has advantages in terms of resonance enhancement and avoidance of fluorescence over normal Raman spectroscopy using visible lasers for sample excitation. Resonance-enhanced Raman spectroscopy is one of the few techniques that can detect catalytic active sites when they are a small minority of the catalyst material. An outstanding example of this feature is the detection of the monomeric dioxo molybdenum species, (O¼)2Mo(-O-Si)2, that are the precursor to the active sites in Mo/SiO2 olefin metathesis catalysts [14] even when present at only 1% of the Mo-content (Fig. 6.2). The ability to avoid sample fluorescence makes possible measurements on complex or “dirty” materials that often comprise commercial heterogeneous catalysts. It must be said that a UV Raman measurement is only useful when the spectrum can be interpreted. It is only a vibrational spectrum, which is indirectly related to atomic structure or a fingerprint for a material species or phase. The technique certainly benefits from robust computational tools that provide information on the relationship between band positions and the species identity or structure. When resonance enhancement is involved, at the current state of the art, the computation of band intensities is generally limited to relatively simple systems, but significant progress has been made using cluster models [87, 88]. For systems with a complex mixture of organic molecules, it would be helpful to implement chemometric analysis techniques, such as Principle Components Analysis (PCA) or SIMPLISMA (SIMPLe-to-use Interactive Self-modeling Mixture Analysis), in order to identify and quantify the components in the mixture. Of course the utility of advanced spectral data treatments will always improve when the experiments are well controlled and the measured data is of high quality. Acknowledgments PCS acknowledges support by the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy under Contract DE-AC0206CH11357 and Grant DE-FG02-03ER15457. Thanks to Dr. Dingdi Wang for drafting Fig. 6.1.

References 1. Angell, C.L.: Raman spectroscopic investigation of zeolites and adsorbed molecules. J. Phys. Chem. 77, 222–227 (1973) 2. Asher, S.A., Johnson, C.R.: Raman spectroscopy of a coal liquid shows that fluorescence interference is minimized with ultraviolet

6

Ultraviolet (UV) Raman Spectroscopy

excitation. Science (Washington, D. C., 1883-). 225, 311–313 (1984) 3. Careri, G., Signorel, G., Sampoli, M., Mazzacur, V.: Raman scattering from water adsorbed on oxide surfaces. J. Catal. 26, 494-& (1972) 4. Jeziorowski, H., Knoezinger, H.: Scattering background in laser Raman spectra of oxide surfaces. Chem. Phys. Lett. 42, 162–165 (1976) 5. Jeziorowski, H., Knoezinger, H.: Laser induced electronic excitation of surface hydroxide ions and scattering background in laser Raman spectra of oxide surfaces. Chem. Phys. Lett. 51, 519–522 (1977) 6. Turner, I.D.M., Paul, S.O., Reid, E., Hendra, P.J.: Laser-Raman study of isomerization of olefins over alumina. J. Chem. Soc. Faraday Trans. I. 72, 2829–2835 (1976) 7. Egerton, T.A., Hardin, A.H., Kozirovs, Y., Sheppard, N.: Reduction of fluorescence from high-area oxides of silica, gamma-alumina, silica-alumina, and Y-zeolite types and Raman-spectra for a series of molecules adsorbed on these surfaces. J. Catal. 32, 343–361 (1974) 8. Li, C., Stair, P.C.: Ultraviolet Raman spectroscopy characterization of sulfated zirconia catalysts: fresh, deactivated and regenerated. Catal. Lett. 36, 119–123 (1996) 9. Long, D.A.: The Raman Effect: A Unified Treatment of the Theory of Raman Scattering by Molecules. Wiley, Chichester; New York (2001) 10. Hagen, G., Glavaski, B.S., Yeager, E.: The Raman Spectrum of an adsorbed species on electrode surfaces. J. Electroanal. Chem. Interfacial Electrochem. 88, 269–275 (1978) 11. Asher, S.A.: UV resonance Raman studies of molecular structure and dynamics: applications in physical and biophysical chemistry. Annu. Rev. Phys. Chem. 39, 537–588 (1988) 12. Hudson, B., Kelly, P.B., Ziegler, L.D., Desiderio, R.A., Gerrity, D.P., Hess, W., Bates, R.: Far ultraviolet laser resonance Raman studies of electronic excitations. In: Garetz, B.A., Lombardi, J.R. (eds.) Advances in Laser Spectroscopy, vol. 3, pp. 1–32. Wiley, New York (1986) 13. Stair, P.C., Li, C.: Ultraviolet Raman spectroscopy of catalysts and other solids. J. Vac. Sci. Technol. A. 15, 1679–1684 (1997) 14. Ding, K., Gulec, A., Johnson, A.M., Drake, T.L., Wu, W., Lin, Y., Weitz, E., Marks, L.D., Stair, P.C.: Highly efficient activation, regeneration, and active site identification of oxide-based olefin metathesis catalysts. ACS Catal. 6, 5740–5746 (2016) 15. Wu, Z., Zhang, C., Stair, P.C.: Influence of absorption on quantitative analysis in Raman spectroscopy. Catal. Today. 113, 40–47 (2006) 16. Kim, H., Kosuda, K.M., Van Duyne, R.P., Stair, P.C.: Resonance Raman and surface- and tip-enhanced Raman spectroscopy methods to study solid catalysts and heterogeneous catalytic reactions. Chem. Soc. Rev. 39, 4820–4844 (2010) 17. A dimensional line drawing of the off-axis ellipsoidal mirror is available from the author 18. Chua, Y.T., Stair, P.C., Wachs, I.E.: A comparison of ultraviolet and visible Raman spectra of supported metal oxide catalysts. J. Phys. Chem. B. 105, 8600–8606 (2001) 19. Signorile, M., Bonino, F., Damin, A., Bordiga, S.: A novel Raman setup based on magnetic-driven rotation of sample. Top. Catal. 61, 1491–1498 (2018) 20. Asher, S.A.: UV resonance Raman spectroscopy for analytical, physical, and biophysical chemistry. Part 1. Anal. Chem. 65, 59A–66A (1993) 21. Chua, Y.T., Stair, P.C.: A novel fluidized bed technique for measuring Uv Raman spectra of catalysts and adsorbates. J. Catal. 196, 66–72 (2000) 22. Beato, P., Schachtl, E., Barbera, K., Bonino, F., Bordiga, S.: Operando Raman spectroscopy applying novel fluidized bed micro-reactor technology. Catal. Today. 205, 128–133 (2013)

147 23. Signorile, M., Gama, D.R., Bonino, F., Svelle, S., Beato, P., Bordiga, S.: Operando Uv-Raman study of the methanol to olefins reaction over Sapo-34: spatiotemporal evolution monitored by different reactor approaches. Catal. Today. 336, 203–209 (2019) 24. Waleska, P.S., Hess, C.: Oligomerization of supported vanadia: structural insight using surface-science models with chemical complexity. J. Phys. Chem. C. 120, 18510–18519 (2016) 25. Waleska, P., Rupp, S., Hess, C.: Operando multiwavelength and time-resolved Raman spectroscopy: structural dynamics of a supported vanadia catalyst at work. J. Phys. Chem. C. 122, 3386–3400 (2018) 26. Li, C., Stair, P.C.: An advance in Raman studies of catalysts: ultraviolet resonance Raman spectroscopy. Stud. Surf. Sci. Catal. 101, 881–890 (1996) 27. Allotta, P.M., Stair, P.C.: Time-resolved studies of ethylene and propylene reactions in zeolite H-Mfi by in-situ fast Ir heating and Uv Raman spectroscopy. ACS Catal. 2, 2424–2432 (2012) 28. Zhang, C., Allotta, P.M., Xiong, G., Stair, P.C.: Uv-Raman and fluorescence spectroscopy of benzene adsorbed inside zeolite pores. J. Phys. Chem. C. 112, 14501–14507 (2008) 29. Wu, Z., Kim, H.-S., Stair, P.C., Rugmini, S., Jackson, S.D.: On the structure of vanadium oxide supported on Aluminas: Uv and visible Raman spectroscopy, Uv-visible diffuse reflectance spectroscopy, and temperature-programmed reduction studies. J. Phys. Chem. 109, 2793–2800 (2005) 30. Nitsche, D., Hess, C.: In situ diagnostics of catalytic materials using tunable confocal Raman spectroscopy. J. Raman Spectrosc. 44, 1733–1738 (2013) 31. Cheong, C.U.A., Stair, P.C.: In situ studies of the lubricant chemistry and frictional properties of perfluoropolyalkyl ethers at a sliding contact. Tribol. Lett. 10, 117–126 (2001) 32. Stair, P.C.: Advances in Raman spectroscopy methods for catalysis research. Curr. Opin. Solid State Mater. Sci. 5, 365–369 (2001) 33. Stair, P.C.: The application of Uv Raman spectroscopy for the characterization of catalysts and catalytic reactions. Adv. Catal. 51, 75–98 (2007) 34. Hess, C.: In situ Raman spectroscopy of catalysts: examples from current research. Top. Catal. 56, 1593–1600 (2013) 35. Wu, Z.L.: Multi-wavelength Raman spectroscopy study of supported vanadia catalysts: structure identification and quantification. Chinese J. Catal. 35, 1591–1608 (2014) 36. Bordiga, S., Lamberti, C., Bonino, F., Travert, A., Thibault-Starzyk, F.: Probing zeolites by vibrational spectroscopies. Chem. Soc. Rev. 44, 7262–7341 (2015) 37. Jin, S., Feng, Z., Fan, F., Li, C.: Uv Raman spectroscopic characterization of catalysts and catalytic active sites. Catal. Lett. 145, 468–481 (2015) 38. Merlen, A., Buijnsters, J.G., Pardanaud, C.: A guide to and review of the use of multiwavelength Raman spectroscopy for characterizing defective aromatic carbon solids: from graphene to amorphous carbons. Coatings. 7, 153/151–153/155 (2017) 39. Wu, Q., Meng, X., Gao, X., Xiao, F.-S.: Solvent-free synthesis of zeolites: mechanism and utility. Acc. Chem. Res. 51, 1396–1403 (2018) 40. Depla, A., Lesthaeghe, D., van Erp, T.S., Aerts, A., Houthoofd, K., Fan, F., Li, C., Van Speybroeck, V., Waroquier, M., Kirschhock, C.E.A., Martens, J.A.: 29si Nmr and Uv-Raman investigation of initial oligomerization reaction pathways in acid-catalyzed silica sol-gel chemistry. J. Phys. Chem. C. 115, 3562–3571 (2011) 41. Depla, A., Verheyen, E., Veyfeyken, A., Van Houteghem, M., Houthoofd, K., Van Speybroeck, V., Waroquier, M., Kirschhock, C.E.A., Martens, J.A.: Uv-Raman and 29si Nmr spectroscopy investigation of the nature of silicate oligomers formed by acid catalyzed hydrolysis and polycondensation of tetramethylorthosilicate. J. Phys. Chem. C. 115, 11077–11088 (2011)

6

148 42. Ju, X., Tian, F., Wang, Y., Fan, F., Feng, Z., Li, C.: A novel synthetic strategy of Fe-Zsm-35 with pure framework Fe species and its formation mechanism. Inorg. Chem. Front. 5, 2031–2037 (2018) 43. Yu, Y., Xiong, G., Li, C., Xiao, F.-S.: Characterization of iron atoms in the framework of Mfi-type zeolites by Uv resonance Raman spectroscopy. J. Catal. 194, 487–490 (2000) 44. Li, C.: Identifying the isolated transition metal ions/oxides in molecular sieves and on oxide supports by Uv resonance Raman spectroscopy. J. Catal. 216, 203–212 (2003) 45. Fan, F.T., Sun, K.J., Feng, Z.C., Xia, H.A., Han, B., Lian, Y.X., Ying, P.L., Li, C.: From molecular fragments to crystals: a Uv Raman spectroscopic study on the mechanism of Fe-Zsm-5 synthesis. Chem Eur J. 15, 3268–3276 (2009) 46. Koekkoek, A.J.J., Kim, W., Degirmenci, V., Xin, H., Ryoo, R., Hensen, E.J.M.: Catalytic performance of sheet-like Fe/Zsm-5 zeolites for the selective oxidation of benzene with nitrous oxide. J. Catal. 299, 81–89 (2013) 47. Wang, J., Xia, H., Ju, X., Feng, Z., Wang, J., Fan, F., Li, C.: Influence of extra-framework Al on the structure of the active iron sites in Fe/Zsm-35. J. Catal. 300, 251–259 (2013) 48. Wu, Z., Stair, P.C., Rugmini, S., Jackson, S.D.: Raman spectroscopic study of V/Θ-Al2o3 catalysts: quantification of surface vanadia species and their structure reduced by hydrogen. J. Phys. Chem. C. 111, 16460–16469 (2007) 49. Chan, S.S., Wachs, I.E., Murrell, L.L., Wang, L., Hall, W.K.: In situ laser Raman spectroscopy of supported metal oxides. J. Phys. Chem. 88, 5831–5835 (1984) 50. Wu, Z., Stair, P.C.: Uv Raman spectroscopic studies of V/Θ-Al2o3 catalysts in butane dehydrogenation. J. Catal. 237, 220–229 (2006) 51. Li, C., Stair, P.C.: Coke formation in zeolites studied by a new technique: ultraviolet resonance Raman spectroscopy. Stud. Surf. Sci. Catal. 105A, 599–606 (1997) 52. Signorile, M., Bonino, F., Damin, A., Bordiga, S.: In situ resonant Uv-Raman spectroscopy of polycyclic aromatic hydrocarbons. J. Phys. Chem. C. 119, 11694–11698 (2015) 53. Rojo-Gama, D., Signorile, M., Bonino, F., Bordiga, S., Olsbye, U., Lillerud, K.P., Beato, P., Svelle, S.: Structure-deactivation relationships in zeolites during the methanol-to-hydrocarbons reaction: complementary assessments of the coke content. J. Catal. 351, 33–48 (2017) 54. Signorile, M., Rojo-Gama, D., Bonino, F., Beato, P., Svelle, S., Bordiga, S.: Topology-dependent hydrocarbon transformations in the methanol-to-hydrocarbons reaction studied by operando Uv-Raman spectroscopy. Phys. Chem. Chem. Phys. 20, 26580–26590 (2018) 55. Tempelman, C.H.L., Hensen, E.J.M.: On the deactivation of Mo/ Hzsm-5 in the methane dehydroaromatization reaction. Appl Catal B. 176-177, 731–739 (2015) 56. Han, H., Liu, M., Ding, F., Wang, Y., Guo, X., Song, C.: Effects of cesium ions and cesium oxide in side-chain alkylation of toluene with methanol over cesium-modified zeolite X. Ind. Eng. Chem. Res. 55, 1849–1858 (2016) 57. Han, H., Liu, M., Nie, X., Ding, F., Wang, Y., Li, J., Guo, X., Song, C.: The promoting effects of alkali metal oxide in side-chain alkylation of toluene with methanol over basic zeolite X. Microporous Mesoporous Mater. 234, 61–72 (2016) 58. An, H., Zhang, F., Guan, Z., Liu, X., Fan, F., Li, C.: Investigating the coke formation mechanism of H-Zsm-5 during methanol dehydration using operando Uv-Raman spectroscopy. ACS Catal. 8, 9207–9215 (2018) 59. Li, C., Xiong, G., Xin, Q., Liu, J.K., Ying, P.L., Feng, Z.C., Li, J., Yang, W.B., Wang, Y.Z., Wang, G.R., Liu, X.Y., Lin, M., Wang, X.Q., Min, E.Z.: Uv resonance Raman spectroscopic identification of titanium atoms in the framework of Ts-1 zeolite. Angew. Chem. Int. Ed. Engl. 38, 2220–2222 (1999)

P. C. Stair 60. Notari, B.: Synthesis and catalytic properties of titanium-containing zeolites. Stud. Surf. Sci. Catal. 37, 413–425 (1988) 61. Wang, L., Xiong, G., Su, J., Li, P., Guo, H.: In situ Uv Raman spectroscopic study on the reaction intermediates for propylene epoxidation on Ts-1. J. Phys. Chem. C. 116, 9122–9131 (2012) 62. Xiong, G., Cao, Y., Guo, Z., Jia, Q., Tian, F., Liu, L.: The roles of different titanium species in Ts-1 zeolite in propylene epoxidation studied by in situ Uv Raman spectroscopy. Phys. Chem. Chem. Phys. 18, 190–196 (2016) 63. Guo, Q., Feng, Z., Li, G., Fan, F., Li, C.: Finding the “missing components” during the synthesis of Ts-1 zeolite by Uv resonance Raman spectroscopy. J. Phys. Chem. C. 117, 2844–2848 (2013) 64. Su, J., Xiong, G., Zhou, J., Liu, W., Zhou, D., Wang, G., Wang, X., Guo, H.: Amorphous Ti species in titanium silicalite-1: structural features, chemical properties, and inactivation with sulfosalt. J. Catal. 288, 1–7 (2012) 65. Nitsche, D., Hess, C.: New insight into the structure of dispersed titania by combining normal-mode analysis with experiment. Chem. Phys. Lett. 616, 115–119 (2014) 66. Waleska, P., Hess, C.: Structural dynamics of dispersed titania during dehydration and oxidative dehydrogenation studied by in situ Uv Raman spectroscopy. Catal. Lett. 148, 2537–2547 (2018) 67. Weckhuysen, B.M., Keller, D.E.: Chemistry, spectroscopy and the role of supported vanadium oxides in heterogeneous catalysis. Catal. Today. 78, 25–46 (2003) 68. Kim, H.-S., Zygmunt, S.A., Stair, P.C., Zapol, P., Curtiss, L.A.: Monomeric vanadium oxide on a Θ-Al2o3 support: a combined experimental/theoretical study. J. Phys. Chem. C. 113, 8836–8843 (2009) 69. Kim, H., Ferguson, G.A., Cheng, L., Zygmunt, S.A., Stair, P.C., Curtiss, L.A.: Structure-specific reactivity of alumina-supported monomeric vanadium oxide species. J. Phys. Chem. C. 116, 2927–2932 (2012) 70. Wegener, S.L.: Supported vanadium oxides prepared by organometallic grafting and sulfur as a “soft” oxidant for catalytic methane conversion. Thesis (2011) 71. Samek, I.A., Bobbitt, N.S., Snurr, R.Q., Stair, P.C.: Interactions of Vox species with amorphous Tio2 domains on Ald-derived aluminasupported materials. J. Phys. Chem. C. 123, 7988–7999 (2019) 72. Wu, Z., Dai, S., Overbury, S.H.: Multiwavelength Raman spectroscopic study of silica-supported vanadium oxide catalysts. J. Phys. Chem. C. 114, 412–422 (2010) 73. Stiegman, A.E.: Comment on “Multiwavelength Raman spectroscopic study of silica-supported vanadium oxide catalysts”. J. Phys. Chem. C. 115, 10917–10924 (2011) 74. Wu, Z.L., Dai, S., Overbury, S.H.: Reply to comment on “Multiwavelength Raman spectroscopic study of silica-supported vanadium oxide catalysts”. J. Phys. Chem. C. 115, 10925–10928 (2011) 75. Nitsche, D., Hess, C.: Normal mode analysis of silica-supported vanadium oxide catalysts: comparison of theory with experiment. Catal. Commun. 52, 40–44 (2014) 76. Nitsche, D., Hess, C.: Structure of isolated vanadia and titania: a deep Uv Raman, Uv-Vis, and Ir spectroscopic study. J. Phys. Chem. C. 120, 1025–1037 (2016) 77. Moisii, C., van de Burgt, L.J., Stiegman, A.E.: Resonance Raman spectroscopy of discrete silica-supported vanadium oxide. Chem. Mater. 20, 3927–3935 (2008) 78. Wu, Z.L., Rondinone, A.J., Ivanov, I.N., Overbury, S.H.: Structure of vanadium oxide supported on ceria by multiwavelength Raman spectroscopy. J. Phys. Chem. C. 115, 25368–25378 (2011) 79. Wu, Z., Li, M., Overbury, S.H.: A Raman spectroscopic study of the speciation of vanadia supported on ceria nanocrystals with defined surface planes. ChemCatChem. 4, 1653–1661 (2012) 80. Wu, Z.L., Li, M.J., Howe, J., Meyer, H.M., Overbury, S.H.: Probing defect sites on Ceo2 nanocrystals with well-defined surface planes

6

Ultraviolet (UV) Raman Spectroscopy

by Raman spectroscopy and O-2 adsorption. Langmuir. 26, 16595–16606 (2010) 81. Wu, Z.L., Li, M.J., Mullins, D.R., Overbury, S.H.: Probing the surface sites of Ceo2 nanocrystals with well-defined surface planes via methanol adsorption and desorption. ACS Catal. 2, 2224–2234 (2012) 82. Peng, R., Sun, X., Li, S., Chen, L., Fu, M., Wu, J., Ye, D.: Shape effect of Pt/Ceo2 catalysts on the catalytic oxidation of toluene. Chem. Eng. J. (Amsterdam, Neth.). 306, 1234–1246 (2016) 83. Wang, B., Chen, B., Sun, Y., Xiao, H., Xu, X., Fu, M., Wu, J., Chen, L., Ye, D.: Effects of dielectric barrier discharge plasma on the catalytic activity of Pt/Ceo2 catalysts. Appl Catal B. 238, 328–338 (2018) 84. Xiao, H., Wu, J., Wang, X., Wang, J., Mo, S., Fu, M., Chen, L., Ye, D.: Ozone-enhanced deep catalytic oxidation of toluene over a platinum-ceria-supported Bea zeolite catalyst. Mol. Catal. 460, 7–15 (2018) 85. Schilling, C., Hess, C.: Real-time observation of the defect dynamics in working Au/Ceo2 catalysts by combined operando Raman/UvVis spectroscopy. J. Phys. Chem. C. 122, 2909–2917 (2018) 86. Schilling, C., Hess, C.: Elucidating the role of support oxygen in the water-gas shift reaction over ceria-supported gold catalysts using operando spectroscopy. ACS Catal. 9, 1159–1171 (2019) 87. Maganas, D., Trunschke, A., Schloegl, R., Neese, F.: A unified view on heterogeneous and homogeneous catalysts through a combination of spectroscopy and quantum chemistry. Faraday Discuss. 188, 181–197 (2016)

149 88. Kubas, A., Noak, J., Trunschke, A., Schloegl, R., Neese, F., Maganas, D.: A combined experimental and theoretical spectroscopic protocol for determination of the structure of heterogeneous catalysts: developing the information content of the resonance Raman spectra of M1 Movox. Chem. Sci. 8, 6338–6353 (2017)

6 Peter C. Stair received a Ph.D. from University of California, Berkeley in 1977. He joined the chemistry faculty at Northwestern University. He has served as Professor, Director of the Center for Catalysis and Surface Science, Director of the Institute for Catalysis in Energy Processes, and Department Chair. He works to develop fundamental understanding in catalysis science.

7

Surface Enhanced Raman Spectroscopy (SERS) Ramo´n A. Alvarez-Puebla

Contents 7.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

7.2

Raman Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

7.3 Surface-Enhanced Raman Scattering (SERS) . . . . . . . . . . 152 7.3.1 The Electromagnetic Effect in SERS . . . . . . . . . . . . . . . . . . . . . . 152 7.3.2 Chemical Mechanism (CT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 7.4

Surface-Enhanced Resonance Raman Scattering (SERRS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

7.5

Surface Selection Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

7.6 7.6.1 7.6.2 7.6.3

SERS Active Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metallic Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Highly Ordered Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hybrid Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7.7

Tip-Enhanced Raman Scattering . . . . . . . . . . . . . . . . . . . . . . . . 161

7.8

SERS Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

7.9

SERS Applications in Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . 163

7.10

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

156 157 158 160

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

Abstract

SERS is a spectroscopic technique that combines modern laser spectroscopy with the exciting optical properties (localized surface plasmon resonances, LSPRs) of metallic nanostructures that result in strongly increased Raman signals when molecules are attached to nanometer-sized metallic structures. Such enhancement can be over 12 orders of magnitude with respect to the signal obtained by normal Raman scattering. Because the registered signal is essentially a vibrational spectrum and, therefore, it contains all the information of the molecular system R. A. Alvarez-Puebla (*) Department of Physical and Inorganic Chemistry, Universitat Rovira i Virgili, Tarragona, Spain Experimental Sciences & Mathematics, ICREA, Barcelona, Spain e-mail: [email protected]

studied, SERS spectroscopy is considered as a powerful analytical technique. Since its discovery in the 1970s, SERS has attracted much interest in various fields of physics, chemistry, biology, and medicine. Keywords

SERS · Localized surface plasmon resonances · Plasmonic materials · Catalysis

7.1

Introduction

The Raman scattering (RS) is the inelastic part of the light scattered by a molecule when illuminated. The energy difference between this inelastically scattered light and the incident light is due to the interaction of the photons with the vibrational states of the molecule under study. The effect was theoretically postulated in 1923 by Smekal [1] and experimentally demonstrated in 1928 by Raman [2] using the sun as the excitation source and some basic experimental setup. Raman scattering spectroscopy provides the unique vibrational fingerprint, containing rich variety of structural and chemical information, of the molecular system under study. Unfortunately, the Raman effect is very weak with cross sections around 10
29 cm2/molecule. These cross sections, however, can be enhanced up to 5 orders of magnitude under resonant conditions, i.e., when the frequency of incident light match an electronic transition of the molecular system [3]. This phenomenon known as resonance Raman scattering [4]. Fifty years after the discovery of the Raman effect, in 1974, Fleischmann and his collaborators informed of a surprisingly strong Raman signal from a pyridine monolayer adsorbed onto an electrochemically rough silver electrode [5]. The unusual signal strength was explained as the result of the increase of the surface area due to the several oxidation-reduction cycles and the subsequent increase of the analyte concentration retained on the electrode. This

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_7

151

152

R. A. Alvarez-Puebla

misinterpretation was simultaneously, and independently, resolved in 1977 by Van Duyne and Jeanmaire [6] and Albrecht and Creighton [7]. These manuscripts confirmed that an intensity enhancement of about 106, compared with the signal from pyridine molecules in the absence of metal, could not be explained by just a local concentration at the surface and thus, the effect, was electromagnetic in nature. Later observations pointed toward the excitation of surface plasmon resonances in the metal as the most likely cause of the increase in the intensity of the Raman signals. During the last 40 years, surface-enhanced Raman scattering (SERS) has been a mature ultrasensitive technique that combines the interaction of light, molecules and plasmonic nanostructures to increase Raman signals of molecular systems enough to resolve their chemical structure, dynamics, and concentration even at the single-molecule regime [8, 9].

7.2

Raman Scattering

When a propagating oscillating dipole (light) interacts with a molecule, it distorts the electronic cloud around the nuclei. This energy is released in the form of scattered radiation. As the oscillating dipole is much larger than the molecule, the electrons are polarized and move to a higher energy state. At that moment, the energy present in the light is transferred to the molecule. This interaction can be viewed as the formation of a very short-lived “complex” between the light energy and the electrons of the molecule in which the nuclei do not have time to move appreciably. This “complex” between the light and the molecule is not stable and the light is released immediately as scattered radiation. This is often called the virtual state of the molecule. Two types of dispersion are easily identified. The most intense (Rayleigh scattering) occurs when the electron cloud relaxes without any nuclear motion. Thus, the process is essentially an elastic and there is no appreciable change in the energy of the photon. The other (Raman scattering, RS), is a much atypical event and involves only one millionth of the total scattered photons. This process occurs when the light and the electrons interact resulting in a movement of the nuclei. Since the nuclei are much heavier than the electrons, there is an appreciable change in energy of the molecule to either lower or higher energy depending on whether the process starts with a molecule in the ground state (Stokes scattering) or from a molecule in a vibrationally excited state (anti-Stokes scattering). Figure 7.1 shows a simple diagram illustrating Rayleigh and Raman scattering. Stokes scattering has an energy of hνS ¼ hνL
hνm, while anti-Stokes scattering has an energy of hνaS ¼ hνL þ hνm, where hνL is the energy of the excitation light and hνm the molecule vibrational energy. Thus, the Raman shift in a Raman spectrum corresponds to the frequency shift between excitation photon and scattered photon.

Consequently, the Raman shift of a band remains constant independently of the energy of the excitation light applied. At room temperature most molecules are in the ground vibrational state. Thus, most of the Raman scattering is Stokes scattering. The power of the Stokes scattering, P(νS), depends on the intensity of the excitation I(νL), the number (N ) of molecules in the probed volume, and the Raman cross section σ R, which is determined by the polarizability of the molecule under study. Intense RS occurs from vibrations, which cause a change in the polarizability of the electron cloud round the molecule. Usually, symmetric vibrations cause the largest changes yielding the largest scattering. Experimentally, the total Raman scattered light, IRS, is averaged over all the random molecular orientations and is proportional to the incoming flux of photons, I0(νL), IRS ¼ σ R I0. The proportionality constant σ R has the dimensions of area and is a function of the energy of the excitation light. The Raman cross section is proportional to the square of the polarizability derivative for the m ! n vibrational transition and the fourth power of the scattering frequency. The efficiencies of the absorption and scattering processes are determined by the function cross section, which is the meeting point of experiments with theory. The cross section is a characteristic value for each atomic or molecular species and for each analytical technique.

7.3

Surface-Enhanced Raman Scattering (SERS)

Surface-enhanced Raman scattering (SERS) can be defined as conventional Raman scattering but carried out on a plasmonic substrate. The fact of the necessity of the plasmonic substrate adds another degree of complexity to this vibrational spectroscopy. SERS effect can be briefly described by the excitation of the molecule under study by the strong electromagnetic field (localized surface plasmon resonance, LSPR) generated by the plasmonic platform when illuminated by the appropriate light. This process is called electromagnetic effect (EM). Further, the electronic interaction between the molecule and the plasmonic material can modify the scattering process of the complex molecule-metal and give rise to an effectively larger cross section than would occur by light scattering from the molecule alone, an effect known as chemical enhancement (CT). The total enhancement in SERS is usually a product of the two mechanisms but, rigorously, to obtain SERS the EM is always required.

7.3.1

The Electromagnetic Effect in SERS

The main contribution to the increase of the Raman signal in SERS is due to the efficient coupling between the localized

7

Surface Enhanced Raman Spectroscopy (SERS)

153

Fig. 7.1 Diagram of the Rayleigh and Raman scattering processes

hnAS Virtual state

Virtual state

hnL hnS

hnS

hnaS

nAS = nL + nm hnL

hnL

hnL

nS = nL – nm Vibrational states hnm

Ground state Rayleigh

Stokes

Anti-stokes

7 surface plasmon resonance (LSPR) of a plasmonic nanostructure and the incident light [10]. Thus, before describing this mechanism, it is necessary to briefly explain the concept of the localized surface plasmon resonances. Notably, in this context, all LSPR-related effects can be understood as electromagnetic effects and the relation to the free electrons of the plasmonic material is only secondary. Thus, all characteristics of LSPR are contained within the dielectric function (and its wavelength dependence) and the geometry of a specific problem. A given material (usually a metal) would be a suitable plasmonic platform if its refractive index has large negative real part and a small imaginary part of the dielectric function (preferably large and negative). The materials that better fulfill these conditions are alkaline and the noble metals, especially silver and gold. The LSPR not only depends on the material but also on the size and strongly the shape. For example, given isotropic particles of the same material, the intensity of the LSPR slightly increases with size although its position does not modify substantially. The shape of the particles has, however, a deeper effect on the LSPR as may significantly modify the intensity, position, and number of plasmons on the nanostructure. For instance, when a spherical metal nanostructure is illuminated with light, the oscillating electric field generates a coherent motion of the conduction electrons (see Fig. 7.2). If the metallic sphere is small (compared to the radiation wavelength), the collective oscillation of the electrons is dipolar. Therefore, the resonance condition for the coherent electronic motion with the electric field is unique and, at this

resonance frequency, metals absorb and scatter light most efficiently. Experimentally, this is observed as a single plasmon band with a UV-Vis-NIR spectrophotometer (Fig. 7.2a). On the other hand, in the case of anisotropic particles, the in-phase motion of the electron cloud can occur along any of the symmetry axes of the particle (Fig. 7.2). Thus, in the case of a nanorod, the plasmon band splits into two modes, either along (longitudinal) or across (transverse) the particle (Fig. 7.2b). The simplest description of the electromagnetic mechanism is based on models of a small metallic sphere (Fig. 7.3). When the light is resonant with the LSPR, the metallic nanosphere radiates its own dipolar field (ESP). The intensity of this radiation is a function of the size of the sphere (r), the distance (d) sphere-molecule, the dielectric constants of the metal ε and the surrounding medium ε0, and the strength of the incident field (E0). Thus, the molecule feels an enhanced local field (EM), which includes both electric-field magnitudes: E0 þ ESP. The light field enhancement A(ν) is determined by the ratio between the field at the position of the molecule and the incoming field. Að ν Þ ¼

  EM ð ν Þ ε
ε0 r 3 E0  rþd E0 ðνÞ ε þ 2ε0

ð7:1Þ

where A(ν) is particularly strong when the real part of ε(ν) is equal to
2ε0. For a intense electromagnetic enhancement, the imaginary part of the dielectric constant requires to be small. These conditions define the resonant excitation of

154

R. A. Alvarez-Puebla

Electric field

Metal sphere

Metal rod

Electron cloud

b) 1.0

1.0

0.8

0.8 Absorbance

Absorbance

a)

0.6

0.6

0.4

0.4

0.2

0.2

0.0 400

500

600

700

Longitudinal

Transversal

0.0 400

800

600

800

1000

Wavelength (nm)

Wavelength (nm)

Fig. 7.2 Representative scheme depicting the induction of a localized surface plasmon resonance in a metallic sphere and an elongated (ellipsoidal) particle upon excitation with light. An electric dipole is created in the nanosphere, whereas the oscillation of the electron cloud can occur

along any of the two symmetry axes of the metal rod. Experimentally, this is reflected as one or two (transversal and longitudinal) LSPR bands in the UV-Vis-NIR spectra of Au nanospheres (a) or nanorods (b)

for the laser and the Stokes field, the electromagnetic SERS enhancement factor, GSERS(νS), can be expressed as: Molecule SERS

d

e = e⬘+ie⬙

G

¼j

EM = E0 + Esp r Esp = r3 Metallic sphere r l

e – e0 e + 2e 0

2

ðνS Þ ¼ jAðνL Þ j jAðνs Þ j

E0

1 (r + d)3

≤ 0.05

Fig. 7.3 Diagram of electromagnetic mechanism for SERS enhancement

surface plasmons for a metal nanosphere [11]. Consequently, the scattered field will be enhanced if it is in resonance with the particle LSPR. Thus, considering the enhancing effects

2

εðνL Þ
ε0 2 εðνS Þ
ε0 2  r 12 ð7:2Þ j j j εðνL Þ þ 2ε0 εðνS Þ þ 2ε0 r þ d

This equation shows that the enhancement scales as the fourth power (|E|4) of the local field of the metallic nanostructure and that it is particularly strong when excitation and scattered fields are in resonance with the surface plasmons. On the other hand, the EM predicts that SERS does not require direct contact between the molecule and the metal but, as shown in Eqs. 7.1 and 7.2, the electromagnetic enhancement factor strongly decreases with the distance to the surface due to the decay of the dipolar field [1/d]3 to the fourth power, resulting in [1/d]12. Maximum values for electromagnetic enhancement for spherical isolated nanoparticles are on the order of 106 [12].

Surface Enhanced Raman Spectroscopy (SERS)

Theory predicts stronger electromagnetic enhancement for sharp features and large curvature regions, which may exist on silver and gold nanostructures. For example, the EM SERS enhancement factor can be increased up to 1011 when the particle presents sharp tips or edges [13], which localizes the electric field within that confined regions allowing to obtain extremely high enhancement factors [13]. Also, closely spaced interacting particles can provide extra field enhancement. Electromagnetic enhancement factors up to 1011 have been estimated for dimers of silver particles [14]. These large enhancements are normally attributed to highly concentrated electromagnetic fields associated with strong LSPRs at intergaps (i.e., electromagnetic hot spots) in structures consisting of two or more coupled nanoparticles with closely spaced features. The size of these hot areas is small and the field concentration on them depends on the geometry of the site where the molecule is retained, the wavelength and the polarization of the incident light. When optical excitation is localized in those electromagnetic hot spots, extremely large SERS enhancement, up to 1012, can be expected [13].

7.3.2

Chemical Mechanism (CT)

Although the electromagnetic mechanism is the essential contribution to the SERS signal, early observations revealed a dependence of the scattering signal on the electrode potential [15]. This suggested an electronic coupling between the molecule and the plasmonic material. Also, the metallic nature of the absorbant may alter the polarizability of adsorbed molecules increasing the Raman-scattering efficiency. Observations such as dependence of the effect on the chemical structure of the sampled analyte [16] and a strong molecular selectivity [17] clearly indicates the existence of an additional chemical enhancement. Further, comparison between the best theoretical SERS enhancement factors (1012) differs about two or three orders of magnitude to the best experimentally observed nonresonant SERS enhancement factors (1015) [18], suggesting the existence of additional enhancement mechanism(s) responsible of this difference. Different mechanisms had been proposed to explain the chemical SERS effect. The formation of a surface complex between the molecule (adsorbate) and metal (adsorbent) may lead to a charge transfer (CT) from the metal to the molecule or vice versa resulting in an increased Raman signal. Figure 7.4 shows a typical energy level diagram for a molecule-metal system, where the energies of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) are approximately symmetric relative to the Fermi level of the metal, together with the possible CT processes involving molecular states (path (a))

155

Metal

Molecule

c LUMO

Energy

7

Fermi level

a b

HOMO

Fig. 7.4 Energy level diagram for a “molecule-metal system” showing a possible charge transfer involving molecular states (path (a)) and molecular and metallic states (paths (b), (c))

and molecular and metallic states (paths (b), (c)). In general, the chemical SERS enhancement factor is considered to contribute enhancement factors on the order of 10–103 and co-adds to the EM enhancement.

7.4

Surface-Enhanced Resonance Raman Scattering (SERRS)

Surface-enhanced resonance Raman scattering is expected when the analyte is in resonance with the frequency of the light used to excite the LSPR. Thus, both enhancement from the plasmon resonance (SERS) and molecular resonance from the analyte (R) contribute to give very intense signal. The wavelength dependence of the intensity obtained from SERRS is different than in SERS because the greatest enhancement is found with excitation close to the molecular resonance, which means that molecular resonance process has a larger effect than plasmon resonance. An advantage of SERRS in practical applications is that fluorescence is quenched on the active surface. This is also true for SERS and can be used to reduce fluorescence interference, but for SERRS it means that excellent Raman scattering can be obtained from fluorescent molecules. The additional sensitivity of SERRS can be an advantage where the electronic structure of the analyte is suitable. The most effective SERRS applications are related with the structural study of some proteins with heme chromophores [19], detection of labeled DNA fragments or antibodies [20]. However, the number of common analytes is much more limited.

7

156

7.5

R. A. Alvarez-Puebla 2

Surface Selection Rules

The surface selection rules, formulated by Moskovits in the 80s [21], studies the modification of the relative intensity of the bands of a given SERS spectrum as a consequence of the orientation of the molecule under study with respect to the plasmonic surface. For example, considering a pfunctionalized mercaptobenzene (Fig. 7.5) [22], and a slightly simplified C2v symmetry for the mercaptobenzene group, its vibrational modes can be classified into in-plane (ip) a1 and b2 modes and out-of-plane (oop) a2 and b1 modes. On the other hand, considering that the surface electric field, E, effectively has only a normal component (Z direction in Fig. 7.5) [23], the intensity of a vibrational mode is proportional to the square of the scalar product of the electric field ! and the dipole moment derivative of the mode, dμ =dQ [24]. !

!

!

Z(E)

x(b1)

Analyte

z(a1) q

Functional group

c

s y(b2) Y X

Fig. 7.5 Model for the estimation of the molecular orientation. Absolute orientation of the molecule on the surface and relative orientation of the ring over the surface are represented by XYZ and xyz axes, respectively

0

ð7:5Þ

2

2

0

ð7:6Þ

where I0 represents the intrinsic intensity of the corresponding mode without the surface effects (i.e., normal Raman spectrum). Thus θ þ 90 , which describes the angle metal surface-S-C, and χ, which describes the orientation of the phenyl ring against the silver surface, can be calculated as follows: 2

I ð b2 Þ I 0 ð b1 Þ I ð b1 Þ I 0 ð b2 Þ

ð7:7Þ

I ð b1 Þ I 0 ð a1 Þ 1 I ða1 Þ I 0 ðb1 Þ cos 2 χ

ð7:8Þ

tan χ ¼

ð7:3Þ

where α is the angle between E and dμ =dQ. Defining θ as the tilt angle of the z axis of the mercaptobenzene unit with the surface normal (Z ), and χ as the twist angle of the molecular plane around the z axis (which is 0 when y is parallel to the surface). Then, by considering that the ip a1 and b2 modes have dipole moment derivatives along z and y axes, respectively, and the oop b1 modes have dipole moment derivatives perpendicular to the phenyl ring (along x axis), the molecularfixed axis system xyz can be correlated with the experimental axis system XYZ by the two Eulerian angles, θ and χ. The intensities of a1, b1, and b2 can be then represented as follows from the above equation [25]:

2

I ðb2 Þ / sin θ sin χI ðb2 Þ

tan θ ¼

dμ ! 2 ⅆμ 2 ! 2 2 I / j E j ¼ j j jE j cos α ⅆQ ⅆQ

ð7:4Þ

2

I ðb1 Þ / sin θ con χI ðb1 Þ

2

!

0

I ða1 Þ / cos θI ða1 Þ

Thus, by assigning the a1, b1, and b2 vibrational modes in the SERS spectra, before and after the coupling, it is possible to know the tilt and twist angles. The deformation of these angles is, however, restricted by the fact that the chemoreceptor is chemically bounded to the surface and requires of very large analytes to be effective.

7.6

SERS Active Substrates

Since the first observation of SERS on silver electrodes, the methods for the preparation of plasmonic substrates for SERS have gone through the following stages: (i) random and nonuniform substrates prepared by electrochemical oxidation and reduction cycles(s) (EC-ORC) [5], physical vapor deposition methods [26], or colloidal casting [27]; (ii) large size distribution nanoparticle colloids prepared by chemical or a laser ablation reduction [28]; (iii) nanoparticles with controlled size and shape prepared by controlled chemical reduction method [29]; and, (iv) large-area surface nanostructures with defined size, shape, and interparticle spacing prepared by self-assembly of colloidal particles [30], or lithography methods [31]. Combining the concluding remark made by Natan [32], we can outline the features of an ideal SERS substrate should: (i) have a high SERS activity and therefore provide high sensitivity; (ii) have good stability and reproducibility; (iii) be uniform so that the deviation in enhancement over the whole surface can be less than 20%, which requires a relatively ordered arrangement of the nanoparticles on the substrate; and, (iv) should be clean enough so that it can be applied to study not only strong adsorbates but also some weak adsorbates or even unknown samples.

7

Surface Enhanced Raman Spectroscopy (SERS)

157

Unfortunately, at present, it is still difficult to obtain SERS substrates that can simultaneously fulfill all the above requirements. Thus, nowadays, SERS many platforms are prepared considering the final application. For instance, for quantitative analysis, a uniform and reproducible substrate is extremely important; however, for the ultratrace analysis of some species, where their mere presence indicates contamination, a maximized enhancement is preferred. Bio-detection requires of platforms that can be easily separated and cleaned from the biological fluid of interest.

7.6.1

Metallic Nanoparticles

In recent years, the chemical [29, 33] methods for the fabrication of metallic nanocolloids permit the fine tuning of nanoparticle size and shape. The most widely used strategy to obtain metallic colloids relies in the chemical reduction of

a)

a metallic precursor with reducing agent (organic acids, borohydride, etc.) This process usually requires the use of surfactants (i.e., cetyltrimethylammonium bromide, CTAB, or poly (vinylpyrrolidone), PVP) firs to direct the size and shape of the nanoparticles but also to prevent the aggregation and flocculation of the colloids. In this scenario, all variables in the reduction reaction (metal precursor, the reducing agent and their relative concentrations, pH, temperature and surfactants) have an impact in the size, shape, and aggregation degree of the colloids prepared. For instance, the variation of the relative concentration between the reducing agent and the metallic seeds allows for the preparation of nanoparticles with different morphologies and different size (Fig. 7.6) [34]. Since the first repot of the use of metallic colloids as SERS platforms in 1979 [35], these substrates, specially of gold, silver, and their alloys, had been widely used as optical enhancers due to their remarkable SERS activity and their reasonable chemical stability. Nowadays, with the fast

MTP

b)

SC

100 nm

100 nm

Seeds

d)

c)

50 nm

20 nm

50 nm

e)

50 nm

f)

AuIII / Ag

g)

100 nm 100 nm

AuIII 250 nm

250 nm

Fig. 7.6 TEM and SEM images of Au nanoparticles with different shapes, all derived of the DMF/polyol reduction in the presence of PVP. (Reproduced with permission from Ref. [34]). Examples derived

150 nm 200 nm

50 nm

from (a-d) seeded-growth mechanisms and (e-f) direct preparation of different shapes without the need of seeds

7

158

R. A. Alvarez-Puebla

b)

100 nm

c)

f)

200 nm

e)

100 nm

100 nm

Au NPs Ag NPs Au NSt Au NRs

Norm. absorbance

a)

300

400

500

600 700 800 900 1000 1100 Wavelength (nm)

Fig. 7.7 Example of the localized surface plasmon resonance response of nanoparticles of different materials and shape: (a) gold and (c) silver nanospheres, (b) gold nanostars, and (d) gold nanorods and (f) their corresponding LSPR spectra

development in nanofabrication, the process to obtain gold and silver sols has become quite sophisticated. In general, it can be stated that silver is a more efficient optical material than gold, giving rise to SERS signals 10- to 100-fold higher than similar gold nanostructures [36]. As mentioned above, the dielectric constant is divided into parts, real and imaginary, scattering being associated with the real part. In the case of silver, the scattering contribution is larger, providing higher enhancement than gold. Additionally, silver can be excited from the UV to the IR while gold is restricted to the red or IR due to damping by the interband transitions [37]. Therefore, silver nanoparticles are preferred when dealing with practical applications. On the other hand, for applications involving living organisms the chemical inertia of gold is a plus. This, together with the biocompatibility of gold colloids [38] and the possibility of fine tuning their size and shape, and thus their optical properties (Fig. 7.7) [39], make particles of this metal very appropriate in biological applications such as in vivo or even intracellular monitoring. Further, due to the so-called biological window, where the tissues are transparent to light SERS community, among others, had been specially interested in preparing SERS substrates with LSPRs in the near-infrared region [40].

7.6.2

Highly Ordered Substrates

As experimentally reported many times, corroborated by theoretical calculations, interactions of two or more nanoparticles result in a much larger optical enhancement than that

expected by the same number of nanoparticles specially isolated. In many experiments, the SERS platform compromises a collection of randomly organized nanoparticles formed by aggregation. Figure 7.8 shows colloidal silver particles in different aggregation stages. In such colloidal cluster structures, the individual dipole of the isolated particles couple generating normal modes of plasmon excitation that cover the cluster with a wide frequency region from the visible to the near infrared (NIR). This figure demonstrates the localized surface plasmon broadening as a consequence of the intercoupling of the different electromagnetic fields provided by each of the nanoparticles as excited with light. Notably, single isolated silver particles that present a narrow spectrum at 420 nm become broader when aggregated in a window over 400–1000 nm. As, in this case, the film is not homogeneous, the interaction of the nanoparticles between themselves is neither homogeneous, giving rise to such broad optical response. Therefore, the surface of a randomly aggregated colloidal cluster structure shows a very heterogeneous electromagnetic field distribution [41]. The size of the hot areas is as small as a few nanometers. Their locations depend strongly on the geometry of the aggregate and on the excitation wavelength and polarization of the optical fields. When optical excitation is localized in such small hot spots, extremely large electromagnetic SERS enhancement up to 1012 was predicted for these areas, as we have mentioned above. The heterogeneous electromagnetic field distribution in aggregated clusters of nanoparticles leads to irreproducible signal intensities from point to point of the same substrate

7

Surface Enhanced Raman Spectroscopy (SERS)

159

c) a)

2.4

a

0.2 μm

Absorbance (a.u.)

d) 1.6

a c

b

c

d

e

d

b)

0.8 e 0.2 μm b

7

0.0

e) 400

500

600

700

800

Wavelength (nm)

0.1 μm

0.2 μm

Fig. 7.8 SERS active colloidal silver particles (a) as prepared and (b-e) after different aggregation stages. The UV-V is and the optical image shows the effect of the aggregation in the optical properties of the colloids

with the same sample. In these circumstances, although these substrates are commonly used in ultrasensitive applications, they should be avoided when quantification of the analyte is required. An alternative to prepare highly efficient optical platforms with homogeneous SERS response involves the control of the number of hot spots and their geometry (nanoparticle interdistance). To prepare highly ordered SERS substrates some strategies include the use of electron or ion lithography to cut the surface of a silver of gold smooth thin film [42], nanosphere lithography [31], or oblique angle vapor deposition [43]. These methods give rise to highly uniform nanostructured surfaces with a wellcontrolled number. However, although appropriate for academic research, for applicability to real life they are time consuming, expensive, as they require clean rooms and special nanofabrication tools and often cannot be sufficiently upscaled. Other most common methods to immobilize metal nanoparticles onto a planar surface are self-assembled monolayers (SAM) by covalent bonding of metal nanoparticles to solid substrates [30], or chemically grown on a solid support synthesized in a polymer or blockcopolymer film [44]. These platforms usually are more appropriate as they can be

easy prepared in common chemical labs with almost no restriction in the size of the surface to be coated. Alternatively two-dimensional ordered assemblies of nanoparticles into periodic arrays represent a highly promising type of nanostructures for SERS because they provide large enhancements and reproducible signals (Fig. 7.9) [45]. Recent developments in nanophotonics [46] have led to new ways of producing highly efficient photonic materials. In fact, it has been theoretically demonstrated that the localization of plasmons occurring in nanoantennas can be propagated between coupled nanorods. Analogously, ordered colloidal crystals (also referred to as nanoparticle supercrystals) are materials that are characterized by three-dimensional periodicity with or without preferential orientation of individual nanoparticles (Fig. 7.10) [47]. In the case of periodic assemblies, the positions of particles and, as a result, packing density and chemical composition are uniform throughout the entire structure. This uniformity makes periodic nanoparticle arrays (films, 3D crystals) especially attractive for practical applications and for fundamental studies. Additionally, these materials exploit not only the generation of LSPRs but also produce other electromagnetic effects such as whispering gallery modes yielding

160

R. A. Alvarez-Puebla

a)

c)

1 μm

b)

1 μm

950 1000 1050 1100 Raman shift (cm–1)

Fig. 7.9 SEM, Dark field, and SERS images of colloidal imprinting to form organized lines (either (a) single, SL, or (b) doubled, DL) upon the deposition of (c) benzenethiol

extraordinarily strong electric fields and, consequently, SERS intensity [45].

7.6.3

Hybrid Materials

In general, a hybrid material is a mixture of materials that add different functionalities, for example, optical properties in order to obtain a good SERS response and on the other hand mechanical properties to provide stability or any other particular property to the material. Although pure metal nanostructures can provide strong SERS signals with homogeneous intensities, still some shortcomings require to be resolved in the preparation of SERS platforms. For instance, some molecular families show very low affinity toward gold or silver surfaces and, therefore, cannot be studied with SERS as they are not retained onto the sensor surface. Thus, the integration of nanoparticles into advanced hybrid platforms, which may offer flexibility and functionality, is key in sensor engineering. These strategies rely on the rational design of materials where properties such as electrostatic charge, chemical affinity, and mechanical response can be tuned to facilitate or even force the retention of the desired analyte onto the plasmonic substrate. Some methods have been proposed through the design of hybrid materials that can unspecifically trap molecules from a solution. Mechanical retention and

identification of analytes have been reported using polymers that are responsive to temperature [48] or pH [49]. Another typical example is the functionalization of plasmonic materials with molecular compounds such as cyclodextrins [50], calixarenes [17], or cucurbiturils [51]. These host molecules present selective affinities, depending on their functionalization, which can trap target analytes that, otherwise, would not be retained onto the metallic surfaces. In fact, the use of host–guest interaction can be so selective to permit the ultra-quantification of chiral enantiomers within a mixture [50]. These materials allow the detection of compounds that would otherwise be elusive. As SERS is essentially an ultrasensitive technique, the direct analysis of complex solutions (i.e., biological fluid or environmental liquid sample), in many cases, the selective sieving of a minimal part of the molecules contained in the sample maybe necessary. To this end, plasmonic materials can be coated with porous silica or metal-organic frameworks. Nowadays, growth of porous silica on a metallic surface can be controlled to the nm in thickness thus offering an additional parameter to control the sieving capacity of the material (Fig. 7.11) [52]. Silica coating can be applied to single particles [53, 54], aggregates [55, 56], or even thin films [57]. In single particles, also known as shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS) [53, 54], the coating also provides colloidal stability to the

7

Surface Enhanced Raman Spectroscopy (SERS)

a)

161

b)

4 μm

4 μm

d)

c)

7

20 μm

500 nm

Fig. 7.10 Example of nanoparticle supercrystals of spherical gold nanospheres organized in (a) grids, (b) ellipsoids, and (c) pyramids. (d) Shows the high-resolution image of a single pyramid

particles while avoiding undesired plasmonic couplings, with the subsequent increase in the quantitative intensity of the signal, but at the cost of a dramatic decrease in its absolute intensity. Thus, to fulfill this drawback, controlled aggregates can be also coated with silica. In this case, the coating still stabilizes the particles, but those, dense aggregates in the interior, promote the formation of highly active electromagnetic hot spots and, consequently, give rise to gigantic SERS intensities. Unfortunately, the porosity of silica cannot be controlled. However, during the last years the coating of metallic nanosurfaces with metal- organic frameworks (MOFs) has been successfully achieved. These polymeric organometallic compounds allow for the control of the pore size, defined by the size of the ligands that coordinate the metallic cation [58], providing an excellent material for size-selective SERS analysis. However, thickness cannot be controlled, and they are sensitive to water. Other strategies involve the incorporation of colloidal silver nanoparticles into bulk materials, generating highly active optical materials that allow direct and ultrasensitive

analysis of molecules that cannot be retained by other methods [50]. Similarly fabrication of optical accumulators to monitor large sample volumes in continuous flows is possible. The concept of optical accumulators is based on the analyte adsorption during a continuous flow of the sample, until the sufficient concentration is reached to allow the detection of SERS [59, 60].

7.7

Tip-Enhanced Raman Scattering

As a derived technique from SERS, tip-enhanced Raman scattering (TERS) spectroscopy is also assisted by the LSPR response of a nanoscaled plasmonic tip [61]. The highly confined electromagnetic field at the tip apex not only enhances the Raman signal of species in the vicinity of the tip (usually less than 5 nm) but also provides high spatial resolution below 5 nm, which allows the capability to ultimately resolve submolecular details [62]. Rather than colloidal of or plasmonic materials, TERS enhancement is derived from the plasmonic tip mounted in an atomic probe

162

R. A. Alvarez-Puebla

BT

*

c) 1

NAT

*

0 1

a)

*

CV

b)

*

3S

Peak area/peak areamax

a)

500 nm

3S 0

500

1000

1600

CV

NAT

microscope (SPM). This tip should not only be sharp enough to obtain high resolution topological images of the sample, but also support the LSPR at the right wavelength to enhance the Raman signal of the sample (Fig. 7.12). TERS study in air is usually hindered by sample damage due to plasmon-induced hot carriers, heating, and oxygen activation. Thus, ultrahigh vacuum-TERS (UHV-TERS) is preferred as it is free of oxygen and impurities and can reduce photo-induced reactions. Also, UHV-SPM has ultrahigh stability and, if combined with a low temperature facility, can achieve TERS imaging at extremely high spatial resolution. This high resolution allows the spatial distinction of two adjacent molecules within van der Waals interactions [63]. Another important direction in TERS is electrochemical TERS (EC-TERS), which involves working in an electrolyte and under potential control. In this way, the surface state and the interaction and reaction of molecules with the substrate can be controlled in a flexible manner. In EC-TERS, the molecular system to be probed is under complex interaction with its environment close to the operando condition.

2500

0.6

3S

CV

0.4 0.2

3000 6

4 NAT BT

2

R2 = 0.9114 R2 = 0.9993

0.0 0.0005

Fig. 7.11 Examples of (a) single particle and (b) silica capsule of gold aggregates. (c) Example of the characterization of the porosity of a silica capsule of gold aggregates coated with silica (a) SERS spectra and (b) molecular structure of the SERS probes used for the exploration of the porous structure of the capsules. (c) Sorption kinetics of the different

2000 1500 Time (s)

0.8

d) 3S

BT

CV

0 1

Molecular volume (nm3/molecule)

b)

100 nm

1200 800 Raman shift (cm–1)

NAT

0 1

0 400

BT

0.001 0.0015 Rate constant

0 0.002

Cross-sectional molecular area (nm3/molecule)

c)

probes onto the inner gold films of the capsules. (d) Comparison of the sorption rate constant of the different probes on the capsules with their molecular volume ad their molecular cross-sectional area (* indicates the band used for representation)

Laser

Tip

TERS

Fig. 7.12 Scheme of a TERS experiment. A plasmonic tip, usually of gold or silver, mounted in a SPM, is illuminated with the appropriate light to excite its LSPR. This electromagnetic field induces an intensification of the signal of the sampled surface providing a vibrational (chemical) image at high resolution

7

Surface Enhanced Raman Spectroscopy (SERS)

163

Although EC-TERS is still at an early stage, it shows impressive potential in addressing more interesting and challenging issues in the electrochemical interface, including electro(photo)catalysis, corrosion, plasmon-driven electron transfer, and energy storage. With an extremely high spatial resolution down to several nm, it is possible to achieve nanoscale spectral imaging of an electrochemical interface and reveal the structure–function correlation at the electrochemical interfaces by correlating the sub-nanometer resolution SPM image with a simultaneous nm resolution TERS image [64].

7.8

SERS Imaging

State-of-the-art Raman spectrometers may acquire spectra of complex plasmonic nanostructures within 100 ms in very optimized substrates, being more reliable acquisition times in the order from 1 to 10 s. While these times may be more than enough for applications requiring single spectra acquisitions (i.e., detection and/or quantification of biomolecules, toxic pollutants, explosives, etc.), it hinders the applicability to CRSAs. Unlike widefield fluorescence microscopy, SERS imaging is still predominantly performed using scanning microscopes. Thus, contrary to fluorescence where a snapshot image can be acquired in milliseconds to seconds, SERS imaging is registered point-by-point or line-by-line, increasing the acquisition times from minutes to hours [65]. In fact, this procedure not only restricts the area to be mapped but also dramatically increases the damage of the mapped surface due to an excessive exposure to the excitation light. As an alternative, holographic measurements of spontaneous Raman images and, simultaneously, measured spectral content of all individual SERS emitters can be achieved [66]. For Raman holographic reader (Fig. 7.13), the output of a λ CW laser (λ wavenumber fitting the hot spots to maximize optical efficiency) is collimated and transmitted through a bandpass filter prior to being focused into the back focal plane of

780 nm

immersion objective by reflecting off a λ nm dichroic beam splitter thus widefield illuminating the sample. The Raman signal is then collected by the same objective and passes the λ nm dichroic beam splitter as well as an additional long pass filter to efficiently isolate the Raman scattering from residual laser light. A Michelson interferometer placed between the microscope objective and the imaging lens allows acquiring Fourier transform Raman spectra as outlined previously. After passing the interferometer, the lens forms a conjugate image of the sample plane slightly in front of the 2D 0-π transmission phase grating. The first four (+1+1;+1
1;
1+1;
1
1) diffraction orders off the grating pass the imaging system, which is arranged such as to form a conjugate image of the sample on the CMOS camera’s chip. Importantly, all other diffraction orders, including the zeroth order, are blocked by a hardaperture mask placed in the Fourier plane of L2/3. This configuration allows for the acquisition of threedimensional SERS images in seconds rather than in hours avoiding sample damage and other changes derived from the environment.

7.9

SERS Applications in Catalysis

As previously stated, SERS requires a plasmonic material, which provides the electromagnetic field for the vibrational excitation of the analyte. This plasmonic material is commonly gold, silver, or copper as their LSPRs are in the visible-NIR regions. As well, these two metals are eminently catalytic. Thus, it is possible to use the optical enhancer (the gold, silver, or copper nanostructure) also as a catalyzer to monitor the reaction. Plasmonic nanostructures illuminated with the appropriate light produce two simultaneous, but linked, effects. First, the collective oscillation of the free electron gas (i.e., plasmon polariton) and, with that, the production of heat. Due to the small size of the nanostructures and the lack of instrumental techniques with resolution enough to resolve small parts of such nanostructures, usually

L0 DBS

BP

L2/3

LP L1

Δt

BS

2DG

Fig. 7.13 Detailed schematic of the experimental setup. L0: widefield lens, DBS: dichroic beam splitter, BP: bandpass filter, LP: long pass filter, L1 lens, BS: 50:50 NIR beamsplitter cube, 2DG: 2D 0-π phase

CAM

grating, L2/3: relay imaging system with aperture stop in its Fourier plane, CAM: imaging camera

7

164

R. A. Alvarez-Puebla

exploiting the ultrasensitivity of SERS it has been possible to monitor the kinetics of a single enzyme [67]. Conversely, this strategy has also been well employed for the study of kinetics of organic reactions (Fig. 7.14a) [56] and, by appropriately engineering of the material, to monitor the levels of important metabolites in the interior of living cells (Fig. 7.14b) [55]. As a second level of complexity, the optical enhancers can be used as the support of the catalyst. Typical examples of

these two mechanisms, electronic and thermal, are difficult to separate. Additionally, the illumination of the reactants with high coherent light should also be considered. Thus, SERS can be used in catalysis in different scenarios. First, the plasmonic nanostructure can be used as a solid support to one of the reactants or products of the reaction. The main goal of this strategy is to gain information of the reaction upon illumination with light. For example, by

a)

b) Laser

a)

785 nm

Reactants

Products

NO I

785 nm T Mesoporous shell Plasmonic NPs

1200

1300

b)

1600

1700

b) 70 Intensity (kcps)

a)

1400 1500 Raman shift (cm–1)

Time

60 50 40 30 20 10 0

1650 1750 1850 Raman shift (cm–1) c)

10 μm

0

900 1800 2700 3600 Time (s)

60

10 μm

10 μm

With H2O2 t = 48000s t = 16000s

Diene + Dienophile + Capsules + light Diene + Dienophile + Capsules Diene + Dienophile 0

100

200

300

400

Normalized yield

c(NO) (nM)

Diene + Dienophile + light 40 t=0s 1500

20

1540 1580 1620 Raman shift (cm–1) Without H2O2

0

10000

20000

30000

40000

t (s)

Fig. 7.14 Schematic cross-sectional view of the plasmonic nanoreactor where reactants and products diffuse through the mesoporous silica shell and NIR laser irradiation promotes the chemical reaction, allowing simultaneous in situ SERS monitoring of the process. (a, b) SERS monitoring of the reaction occurring inside the nanocapsule over the time. (c) Comparison of the reaction yields after 30 min determined by 1 H NMR for our system and controls including the diene and dienophile in the presence and absence of capsules and light. Reproduced with permission from Ref. [56]. (b) Spectra of p-aminobenzenethiol (ABT, blue) and p-hidroxybenzethiol (HBT, red) are shown for comparison.

Thin traces in between represent the resulting spectra with growing concentrations of NO. (b) Optical images of 3T3 cells and intracellular NO formation over time for three different samples upon NO induction with oxygen peroxide (H2O2). A control sample without the presence of H2O2 is also shown for comparison. Representative normalized SERS spectra obtained at different times are shown. The SERS dashed (blue) and dotted (red) spectra represent the reference vibrational pattern for ABT and HBT, respectively. (Reproduced with permission from Ref. [55])

Surface Enhanced Raman Spectroscopy (SERS)

165

20 4

0.0 10

20 t (sec)

1600 cm–1 0.5 1580 cm–1 0.0

h)

5 t (sec)

0.0

1095 cm–1 1600 cm–1 Pt-Au-Ag CM Pd-Au-Ag CM 150 100

Ag-Au CM

50

10 t (sec)

20

NH

CC

vCC00

+d

0v

16

0.5 0.0 1580 cm–1

i)

2 t (sec)

1095 cm–1

4

1600 cm–1

80

40

40

20

0

0

Au NP

0 0

1.0

0

Intensity (a.u.)

Intensity (a.u.)

Pt-Au-Ag CM Pd-Au-Ag CM Ag-Au CM

7 1600 cm–1

10

200

0.5

158

1300 1400 1500 1600 Raman shift (cm–1)

f) Pt-Au-Ag CM

1.0

0

1.0

4 6 t (sec)

0 8 t (sec)

1300 1400 1500 1600 Raman shift (cm–1)

30

g)

0 2

500

1500 2000 1000 Raman shift (cm–1)

Fig. 7.15 (a–c) In situ SERS monitoring of the NBT reduction photocatalyzed by (a) bimetallic Ag–Au NPs, (b) trimetallic Pd–Au–Ag NPs, and (c) Pt–Au–Ag NPs. Note that the characteristic band of NBT (1580 cm–1) shifted to that of ABT (1600 cm–1) during the reduction process. (d–f) Changes in the normalized SERS spectra intensity at the characteristic bands of NBT (red) and ABT (blue) as a function of the reaction time depending on the type of NPs. (g) Ratio-metric changes in the characteristic peaks of ABT and NBT as a function of time. The red, green, and blue colors indicate the Ag–Au, Pd–Au–Ag, and Pt–Au–Ag

Pd Ag- Au -A A NP Pt u-Au C -A g M u- C A M g CM

1580 cm–1

40

A Pd Ag- u N -A Au P Pt u- C -A A M u- g C A M g CM

0.5

80

0

Normalized intensity (a.u.)

1600 cm–1

135

16

0

e) Pd-Au-Ag CM

1.0

0 v N

NH

+d

vCC00

158

40

t (sec)

d) Ag-Au CM Normalized intensity (a.u.)

60

Normalized intensity (a.u.)

0

10 20

Normalized relative intensity (In(I1600/I1580))

CC

O2

0

Intensity (a.u.)

100

0

0 v N

16

80

200

1300 1400 1500 1600 Raman shift (cm–1)

135

vCC00

+d

NH

CC

158

0v

0 v N 135 Intensity (a.u.)

300

c) Pt-Au-Ag CM O2

b) Pd-Au-Ag CM O2

a) Ag-Au CM

p-nitrobenzenethiol (NBT) on silver/gold and silver/gold decorated with Pd or Pt nanoparticles [68]. Finally, the last scenario is the use of the optical enhancer as catalyst. This

Intensity (a.u.)

this are plasmonic nanoparticles decorated with other non-plasmonic but catalytic metals such us Pd, Pt, etc. Figure 7.15 shows the classical plasmonic reduction of

0v

7

NPs, respectively. The logistic fitting curves are shown as light-colored lines accordingly. (h) Averaged single-nanoparticle SERS profiles of ABT obtained from five individual single NPs and spherical Au NPs. The characteristic peaks of ABT (1095 and 1600 cm–1) are marked by yellow lines. (i) Comparison of the single-nanoparticle SERS intensities of the characteristic peaks at 1095 (left) and 1600 cm–1 (right) depending on the type of nanomaterials. (Reproduced with permission from Ref. [68])

166

R. A. Alvarez-Puebla

n1CO32–

b)

m

nasCO2– ù

ù

–1.2 nCu-C 1602

1680

1211

ù

m

Raman intensity (cnts)

1375

–1.1

n1CO32– –0.3 V

–1.2 V

1050

–0.9

1100

nasCO2–

–0.7

1330

nsCO2–

703

535 608

dCO2–

1540

350

–1.0

250 300 350 400

Raman intensity (AU)

2060

1295

1059 1119

298

m

1429

m 20,000

nCu-C

nCO

ù

E, V (Ag/AgCl)

a)

–1.0 V

–0.5 –0.3

CuOx

–0.1 200

400

600

800

1000

1200

1400

1600

1800

2000

2200

–1.2 V

–0.9 V

–1.1 V 1500

1600

Raman shift (cm–1)

Fig. 7.16 (a) Operando SERS of rough Cu surface in CO2-saturated in 0.1 M NaHCO3 (pH 6.8). Spectra were measured from
0.1 V toward the cathodic direction. Panel (b) shows the enlarged and scaled by intensity peaks of adsorbed *CO32
and *CO2
. The intensity of the

spectrum measured at
1.2 V is underestimated because the SERS signal dropped due to the hydrogen evolution reaction. (Reproduced with permission from Ref. [69])

strategy has been successfully applied to the study of the CO2 electroreduction to formate and CO on copper electrode (Fig. 7.16) [69].

and realizing the effective transfer of SERS as a standard analytical method in the near future.

7.10

Conclusions

Surface-enhanced Raman scattering is a powerful method to analyze molecular systems closely spaced from a plasmonic substrate. For analytical purposes, high sensitivity, efficiency, and intensity reproducibility of SERS substrates are recurrently discussed as key factors. In this context, specific platforms have been designed and fabricated toward a myriad of sensing strategies including direct SERS, indirect SERS with chemosensors, SERS encoded particles, chiral-selective systems, or remote SERS. These strategies are, nowadays, routinely applied in a widefield including material science, analytical chemistry, biochemistry, or medicine, to name a few. However, optimization of the physical, optical, and chemical properties of SERS platforms will be indispensable for obtaining a deeper understanding of all relevant aspects

Acknowledgments This work was funded by the Spanish Ministerio de Economia y Competitividad (CTQ2017-88648R), the Generalitat de Cataluña (2017SGR883) the Universitat Rovira i Virgili (2018PFR-URV-B2-02) and the European Union Horizon 2020 (No.713679).

References 1. Smekal, A.: Zur Quantentheorie der Dispersion. Naturwissenschaften. 11(43), 873–875 (1923) 2. Raman, C.V., Krishnan, K.S.: A new type of secondary radiation. Nature. 121(3048), 501–502 (1928) 3. Shim, S., Stuart, C.M., Mathies, R.A.: Resonance Raman crosssections and vibronic analysis of rhodamine 6G from broadband stimulated Raman spectroscopy. ChemPhysChem. 9(5), 697–699 (2008) 4. Apanasevich, P.A., Ordabaev, D.N.: Stimulated resonance Raman scattering. J. Appl. Spectrosc. 4(2), 102–107 (1966) 5. Fleischmann, M., Hendra, P.J., McQuillan, A.J.: Raman spectra of pyridine adsorbed at a silver electrode. Chem. Phys. Lett. 26(2), 163–166 (1974)

7

Surface Enhanced Raman Spectroscopy (SERS)

6. Jeanmaire, D.L., Van Duyne, R.P.: Surface raman spectroelectrochemistry: part I. Heterocyclic, aromatic, and aliphatic amines adsorbed on the anodized silver electrode. J. Electroanal. Chem. Interfacial Electrochem. 84(1), 1–20 (1977) 7. Albrecht, M.G., Creighton, J.A.: Anomalously intense Raman spectra of pyridine at a silver electrode. J. Am. Chem. Soc. 99(15), 5215–5217 (1977) 8. Nie, S., Emory, S.R.: Probing single molecules and single nanoparticles by surface-enhanced Raman scattering. Science. 275(5303), 1102–1106 (1997) 9. Kneipp, K., Wang, Y., Kneipp, H., Perelman, L.T., Itzkan, I., Dasari, R.R., Feld, M.S.: Single molecule detection using surface-enhanced Raman scattering (SERS). Phys. Rev. Lett. 78(9), 1667–1670 (1997) 10. Moskovits, M.: Surface-enhanced spectroscopy. Rev. Mod. Phys. 57(3), 783–826 (1985) 11. Kelly, K.L., Coronado, E., Zhao, L.L., Schatz, G.C.: The optical properties of metal nanoparticles: the influence of size, shape, and dielectric environment. J. Phys. Chem. B. 107(3), 668–677 (2003) 12. Langer, J., Jimenez de Aberasturi, D., Aizpurua, J., Alvarez-Puebla, R.A., Auguié, B., Baumberg, J.J., Bazan, G.C., Bell, S.E.J., Boisen, A., Brolo, A.G., Choo, J., Cialla-May, D., Deckert, V., Fabris, L., Faulds, K., García de Abajo, F.J., Goodacre, R., Graham, D., Haes, A.J., Haynes, C.L., Huck, C., Itoh, T., Käll, M., Kneipp, J., Kotov, N.A., Kuang, H., Le Ru, E.C., Lee, H.K., Li, J.-F., Ling, X.Y., Maier, S.A., Mayerhöfer, T., Moskovits, M., Murakoshi, K., Nam, J.-M., Nie, S., Ozaki, Y., Pastoriza-Santos, I., Perez-Juste, J., Popp, J., Pucci, A., Reich, S., Ren, B., Schatz, G.C., Shegai, T., Schlücker, S., Tay, L.-L., Thomas, K.G., Tian, Z.-Q., Van Duyne, R.P., Vo-Dinh, T., Wang, Y., Willets, K.A., Xu, C., Xu, H., Xu, Y., Yamamoto, Y.S., Zhao, B., Liz-Marzán, L.M.: Present and future of surface-enhanced Raman scattering. ACS Nano. 14(1), 28–117 (2020) 13. Alvarez-Puebla, R., Liz-Marzan, L.M., Garcia de Abajo, F.J.: Light concentration at the nanometer scale. J. Phys. Chem. Lett. 1(16), 2428–2434 (2010) 14. Hao, E., Schatz, G.C.: Electromagnetic fields around silver nanoparticles and dimers. J. Chem. Phys. 120(1), 357–366 (2003) 15. Kudelski, A., Bukowska, J.: The chemical effect in surface enhanced Raman scattering (SERS) for piperidine adsorbed on a silver electrode. Surf. Sci. 368(1), 396–400 (1996) 16. Corio, P., Temperini, M.L.A., Santos, P.S., Rubim, J.C.: Contribution of the charge transfer mechanism to the surface-enhanced Raman scattering of the binuclear ion complex [Fe2((Bpe)(CN)10] 6- adsorbed on a silver electrode in different solvents. Langmuir. 15(7), 2500–2507 (1999) 17. Guerrini, L., Garcia-Ramos, J.V., Domingo, C., Sanchez-Cortes, S.: Functionalization of Ag nanoparticles with dithiocarbamate calix[4] arene as an effective supramolecular host for the surface-enhanced Raman scattering detection of polycyclic aromatic hydrocarbons. Langmuir. 22(26), 10924–10926 (2006) 18. Kneipp, K., Kneipp, H., Kartha, V.B., Manoharan, R., Deinum, G., Itzkan, I., Dasari, R.R., Feld, M.S.: Detection and identification of a single DNA base molecule using surface-enhanced Raman scattering (SERS). Phys. Rev. E. 57(6), R6281–R6284 (1998) 19. Murgida, D.H., Hildebrandt, P.: Heterogeneous electron transfer of cytochrome c on coated silver electrodes. Electric field effects on structure and redox potential. J. Phys. Chem. B. 105(8), 1578–1586 (2001) 20. Faulds, K., Smith, W.E., Graham, D.: Evaluation of surfaceenhanced resonance Raman scattering for quantitative DNA analysis. Anal. Chem. 76(2), 412–417 (2004) 21. Moskovits, M., Suh, J.S.: Surface selection rules for surfaceenhanced Raman spectroscopy: calculations and application to the surface-enhanced Raman spectrum of phthalazine on silver. J. Phys. Chem. 88(23), 5526–5530 (1984) 22. Pazos, E., Garcia-Algar, M., Penas, C., Nazarenus, M., Torruella, A., Pazos-Perez, N., Guerrini, L., Vázquez, M.E., Garcia-Rico, E.,

167 Mascareñas, J.L., Alvarez-Puebla, R.A.: Surface-enhanced Raman scattering surface selection rules for the proteomic liquid biopsy in real samples: efficient detection of the oncoprotein c-MYC. J. Am. Chem. Soc. 138(43), 14206–14209 (2016) 23. Ataka, K.-i., Yotsuyanagi, T., Osawa, M.: Potential-dependent reorientation of water molecules at an electrode/electrolyte Interface studied by surface-enhanced infrared absorption spectroscopy. J. Phys. Chem. 100(25), 10664–10672 (1996) 24. Allara, D.L., Nuzzo, R.G.: Spontaneously organized molecular assemblies. 2. Quantitative infrared spectroscopic determination of equilibrium structures of solution-adsorbed n-alkanoic acids on an oxidized aluminum surface. Langmuir. 1(1), 52–66 (1985) 25. Chalmers, J.M., Griffiths, P.R.: Handbook of Vibrational Spectroscopy. Wiley, Chichester (2002) 26. Rowe, J.E., Shank, C.V., Zwemer, D.A., Murray, C.A.: Ultrahighvacuum studies of enhanced Raman scattering from pyridine on Ag surfaces. Phys. Rev. Lett. 44(26), 1770–1773 (1980) 27. Alvarez-Puebla, R.A., Aroca, R.F.: Synthesis of silver nanoparticles with controllable surface charge and their application to surfaceenhanced Raman scattering. Anal. Chem. 81(6), 2280–2285 (2009) 28. Lee, P.C., Meisel, D.: Adsorption and surface-enhanced Raman of dyes on silver and gold sols. J. Phys. Chem. 86(17), 3391–3395 (1982) 29. Litti, L., Reguera, J., García de Abajo, F.J., Meneghetti, M., Liz-Marzán, L.M.: Manipulating chemistry through nanoparticle morphology. Nanoscale Horiz. 5(1), 102–108 (2020) 30. Freeman, R.G., Grabar, K.C., Allison, K.J., Bright, R.M., Davis, J.A., Guthrie, A.P., Hommer, M.B., Jackson, M.A., Smith, P.C., Walter, D.G., Natan, M.J.: Self-assembled metal colloid monolayers: an approach to SERS substrates. Science. 267(5204), 1629–1632 (1995) 31. Haynes, C.L., Van Duyne, R.P.: Nanosphere lithography: a versatile nanofabrication tool for studies of size-dependent nanoparticle optics. J. Phys. Chem. B. 105(24), 5599–5611 (2001) 32. Natan, M.J.: Concluding remarks surface enhanced Raman scattering. Faraday Discuss. 132(0), 321–328 (2006) 33. Maillard, M., Huang, P., Brus, L.: Silver nanodisk growth by surface plasmon enhanced photoreduction of adsorbed [Ag+]. Nano Lett. 3(11), 1611–1615 (2003) 34. Grzelczak, M., Pérez-Juste, J., Mulvaney, P., Liz-Marzán, L.M.: Shape control in gold nanoparticle synthesis. Chem. Soc. Rev. 37(9), 1783–1791 (2008) 35. Creighton, J.A., Blatchford, C.G., Albrecht, M.G.: Plasma resonance enhancement of Raman scattering by pyridine adsorbed on silver or gold sol particles of size comparable to the excitation wavelength. J. Chem. Soc., Faraday Trans. 2. 75(0), 790–798 (1979) 36. García de Abajo, F.J.: Colloquium: light scattering by particle and hole arrays. Rev. Mod. Phys. 79(4), 1267–1290 (2007) 37. Zhao, J., Pinchuk, A.O., McMahon, J.M., Li, S., Ausman, L.K., Atkinson, A.L., Schatz, G.C.: Methods for describing the electromagnetic properties of silver and gold nanoparticles. Acc. Chem. Res. 41(12), 1710–1720 (2008) 38. Murphy, C.J., Gole, A.M., Stone, J.W., Sisco, P.N., Alkilany, A.M., Goldsmith, E.C., Baxter, S.C.: Gold nanoparticles in biology: beyond toxicity to cellular imaging. Acc. Chem. Res. 41(12), 1721–1730 (2008) 39. Alvarez-Puebla, R.A., Liz-Marzán, L.M.: SERS-based diagnosis and biodetection. Small. 6(5), 604–610 (2010) 40. Huang, X., El-Sayed, I.H., Qian, W., El-Sayed, M.A.: Cancer cells assemble and align gold nanorods conjugated to antibodies to produce highly enhanced, sharp, and polarized surface Raman spectra: a potential cancer diagnostic marker. Nano Lett. 7(6), 1591–1597 (2007) 41. Kneipp, K., Kneipp, H., Itzkan, I., Dasari, R.R., Feld, M.S.: Surfaceenhanced Raman scattering and biophysics. J. Phys. Condens. Matter. 14(18), R597–R624 (2002) 42. Abu Hatab, N.A., Oran, J.M., Sepaniak, M.J.: Surface-enhanced Raman spectroscopy substrates created via electron beam lithography and nanotransfer printing. ACS Nano. 2(2), 377–385 (2008)

7

168 43. Driskell, J.D., Shanmukh, S., Liu, Y., Chaney, S.B., Tang, X.J., Zhao, Y.P., Dluhy, R.A.: The use of aligned silver nanorod arrays prepared by oblique angle deposition as surface enhanced Raman scattering substrates. J. Phys. Chem. C. 112(4), 895–901 (2008) 44. Sánchez-Iglesias, A., Aldeanueva-Potel, P., Ni, W., Pérez-Juste, J., Pastoriza-Santos, I., Alvarez-Puebla, R.A., Mbenkum, B.N., Liz-Marzán, L.M.: Chemical seeded growth of Ag nanoparticle arrays and their application as reproducible SERS substrates. Nano Today. 5(1), 21–27 (2010) 45. Blanco-Formoso, M., Pazos-Perez, N., Alvarez-Puebla, R.A.: Fabrication of plasmonic supercrystals and their SERS enhancing properties. ACS Omega. 5(40), 25485–25492 (2020) 46. Bryant, G.W., García de Abajo, F.J., Aizpurua, J.: Mapping the plasmon resonances of metallic nanoantennas. Nano Lett. 8(2), 631–636 (2008) 47. Murray, C.B., Kagan, C.R., Bawendi, M.G.: Synthesis and characterization of monodisperse nanocrystals and close-packed nanocrystal assemblies. Annu. Rev. Mater. Sci. 30(1), 545–610 (2000) 48. Álvarez-Puebla, R.A., Contreras-Cáceres, R., Pastoriza-Santos, I., Pérez-Juste, J., Liz-Marzán, L.M.: Au@pNIPAM colloids as molecular traps for surface-enhanced, spectroscopic, ultra-sensitive analysis. Angew. Chem. Int. Ed. 48(1), 138–143 (2009) 49. Qian, X., Li, J., Nie, S.: Stimuli-responsive SERS nanoparticles: conformational control of plasmonic coupling and surface Raman enhancement. J. Am. Chem. Soc. 131(22), 7540–7541 (2009) 50. Abalde-Cela, S., Hermida-Ramón, J.M., Contreras-Carballada, P., De Cola, L., Guerrero-Martínez, A., Alvarez-Puebla, R.A., Liz-Marzán, L.M.: SERS chiral recognition and quantification of enantiomers through cyclodextrin supramolecular complexation. ChemPhysChem. 12(8), 1529–1535 (2011) 51. Tan, L.-L., Wei, M., Shang, L., Yang, Y.-W.: Cucurbiturils-mediated noble metal nanoparticles for applications in sensing, SERS, theranostics, and catalysis. Adv. Funct. Mater. 31(1), 2007277 (2021) 52. Serrano, I.C., Vazquez-Vazquez, C., Adams, A.M., Stoica, G., Correa-Duarte, M.A., Palomares, E., Alvarez-Puebla, R.A.: The effect of the silica thickness on the enhanced emission in single particle quantum dots coated with gold nanoparticles. RSC Adv. 3(27), 10691–10695 (2013) 53. Li, J.F., Huang, Y.F., Ding, Y., Yang, Z.L., Li, S.B., Zhou, X.S., Fan, F.R., Zhang, W., Zhou, Z.Y., Wu, D.Y., Ren, B., Wang, Z.L., Tian, Z.Q.: Shell-isolated nanoparticle-enhanced Raman spectroscopy. Nature. 464(7287), 392–395 (2010) 54. Li, J.F., Tian, X.D., Li, S.B., Anema, J.R., Yang, Z.L., Ding, Y., Wu, Y.F., Zeng, Y.M., Chen, Q.Z., Ren, B., Wang, Z.L., Tian, Z.Q.: Surface analysis using shell-isolated nanoparticle-enhanced Raman spectroscopy. Nat. Protoc. 8(1), 52–65 (2013) 55. Riveragil, P., Vazquez-Vazquez, C., Giannini, V., Callao, M.P., Parak, W.J., Correa-Duarte, M.A., Alvarez-Puebla, R.A.: Plasmonic nanoprobes for real-time optical monitoring of nitric oxide inside living cells. Angew. Chem. Int. Ed. 52(51), 13694–13698 (2013) 56. Vázquez-Vázquez, C., Vaz, B., Giannini, V., Pérez-Lorenzo, M., Alvarez-Puebla, R.A., Correa-Duarte, M.A.: Nanoreactors for simultaneous remote thermal activation and optical monitoring of chemical reactions. J. Am. Chem. Soc. 135(37), 13616–13619 (2013) 57. López-Puente, V., Abalde-Cela, S., Angelomé, P.C., AlvarezPuebla, R.A., Liz-Marzán, L.M.: Plasmonic mesoporous composites as molecular sieves for SERS detection. J. Phys. Chem. Lett. 4(16), 2715–2720 (2013) 58. Carrillo-Carrión, C., Martínez, R., Navarro Poupard, M.F., Pelaz, B., Polo, E., Arenas-Vivo, A., Olgiati, A., Taboada, P., Soliman, M.G., Catalán, Ú., Fernández-Castillejo, S., Solà, R., Parak, W.J., Horcajada, P., Alvarez-Puebla, R.A., del Pino, P.: Aqueous stable gold nanostar/ ZIF-8 nanocomposites for light-triggered release of active cargo inside living cells. Angew. Chem. Int. Ed. 58(21), 7078–7082 (2019) 59. Ko, H., Chang, S., Tsukruk, V.V.: Porous substrates for label-free molecular level detection of nonresonant organic molecules. ACS Nano. 3(1), 181–188 (2009)

R. A. Alvarez-Puebla 60. Blanco-Formoso, M., Pazos-Perez, N., Alvarez-Puebla, R.A.: Fabrication of colloidal platforms for surface-enhanced Raman spectroscopy on optically inert templates. J. Raman Spectrosc. n/a(n/a), 554 (2020) 61. Anderson, M.S.: Locally enhanced Raman spectroscopy with an atomic force microscope. Appl. Phys. Lett. 76(21), 3130–3132 (2000) 62. Zhang, R., Zhang, Y., Dong, Z.C., Jiang, S., Zhang, C., Chen, L.G., Zhang, L., Liao, Y., Aizpurua, J., Luo, Y., Yang, J.L., Hou, J.G.: Chemical mapping of a single molecule by plasmon-enhanced Raman scattering. Nature. 498(7452), 82–86 (2013) 63. Nguyen, D., Kang, G., Chiang, N., Chen, X., Seideman, T., Hersam, M.C., Schatz, G.C., Van Duyne, R.P.: Probing molecular-scale catalytic interactions between oxygen and cobalt phthalocyanine using tip-enhanced Raman spectroscopy. J. Am. Chem. Soc. 140(18), 5948–5954 (2018) 64. Sonntag, M.D., Pozzi, E.A., Jiang, N., Hersam, M.C., Van Duyne, R.P.: Recent advances in tip-enhanced Raman spectroscopy. J. Phys. Chem. Lett. 5(18), 3125–3130 (2014) 65. Palonpon, A.F., Ando, J., Yamakoshi, H., Dodo, K., Sodeoka, M., Kawata, S., Fujita, K.: Raman and SERS microscopy for molecular imaging of live cells. Nat. Protoc. 8(4), 677–692 (2013) 66. Liebel, M., Pazos-Perez, N., van Hulst, N., Alvarez-Puebla, R.A.: Surface-enhanced Raman scattering holography. Nat. Nanotech. 15, 1005–1011 (2020) 67. Moore, B.D., Stevenson, L., Watt, A., Flitsch, S., Turner, N.J., Cassidy, C., Graham, D.: Rapid and ultra-sensitive determination of enzyme activities using surface-enhanced resonance Raman scattering. Nat. Biotechnol. 22(9), 1133–1138 (2004) 68. Ryu, H.-J., Shin, H., Oh, S., Joo, J.H., Choi, Y., Lee, J.-S.: Wrapping AgCl nanostructures with trimetallic nanomeshes for plasmonenhanced catalysis and in situ SERS monitoring of chemical reactions. ACS Appl. Mater. Interfaces. 12(2), 2842–2853 (2020) 69. Chernyshova, I.V., Somasundaran, P., Ponnurangam, S.: On the origin of the elusive first intermediate of CO2 electroreduction. Proc. Natl. Acad. Sci. 115(40), E9261–E9270 (2018)

Prof. Ramón A. Alvarez-Puebla is an expert in surface science, nanoscience, and spectroscopy with emphasis on the manufacturing and characterization of plasmonic materials and their integration into advanced detection devices especially with applications in nano- biomedicine. He has a degree in chemistry (Universidad de Navarra, 2000) and a doctorate in surface science (Universidad Publica de Navarra, 2003, summa cum laude). He completed his postdoctorate at the University of Windsor (Windsor, ON, Canada) and General Motors Corporation (Warren, Mi, USA) in nanofabrication and surface spectroscopy. He worked as a Research Officer and Principal Investigator at the National Institute for Nanotechnology of the National Research Council of Canada (NINTNRC, Edmonton, AB, Canada). In 2008 he returned to Spain as an associate Professor at the Universidade de Vigo. Since 2012, he is the ICREA Professor at the Universitat Rovira i Virgili in Tarragona, where he leads the Plasmonics and Ultradetection Group (Zeptonic).

8

Nanoscale Raman Spectroscopy Tanja Deckert-Gaudig, Marie Richard-Lacroix, and Volker Deckert

Contents 8.1

Abstract

Short Introduction to TERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

8.2

Theoretical Background of Plasmon-Induced Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 8.2.1 Thermal Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 8.3

Nanoscale Spectroscopic Investigation of Catalyzed Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

8.4

Nanoscale Catalytic Reactions with Plasmon Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . pNTP and pATP Dimerization to DMAB and Other Azo Bridge Containing Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Triple Bond Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (De)Protonation of Pyridine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Miscellaneous: Bond Cleavages . . . . . . . . . . . . . . . . . . . . . . . . . . .

8.4.1 8.4.2 8.4.3 8.4.4 8.5 8.5.1 8.5.2 8.5.3 8.5.4

Electrochemical Processes Using EC-AFM-TERS and EC-STM-TERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reversible Redox Reaction of Nile Blue . . . . . . . . . . . . . . . . . . Protonation Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cleavage of Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manipulating Phthalocyanine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8.6 Catalytic Reactions Without Plasmon Contribution . . . 8.6.1 Porphyrin and Phthalocyanine: NO, CO, O, O2 Complexation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.6.2 Bimetallic Substrates: Oxidation on Au/Pd and Au/Pt Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.6.3 Cis-Trans Isomerization Around an Azo Bridge . . . . . . . . . . 8.7

174 174 177 177 177 178 178 179 179 180 180 180 181 183

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

T. Deckert-Gaudig · V. Deckert (*) Leibniz Institute of Photonic Technology, Jena, Germany e-mail: [email protected]; [email protected] M. Richard-Lacroix Department of Nanoscopy, Leibniz Institute of Photonic Technology, Jena, Germany e-mail: [email protected]

Heterogeneous metal-catalysis plays a significant role in organic synthesis as it lowers the activation barrier necessary to initiate a chemical reaction. In this field, mainly thermally activated catalysts are established whereby unwanted by-products are often an issue. Photocatalysts can be a promising alternative, especially, if they can be activated with visible light under ambient conditions. To promote the application and development of this rather young class of catalysts, a detailed understanding of their surface structure, processes at the interface, reaction kinetics and mechanisms on the atomic and molecular level is of great importance. Photocatalysts based on noble metal nanoparticles like gold and silver are highly attractive candidates in this field. When irradiated with the appropriate wavelength, resonant light absorption of the nanoparticles can lead to the excitation of localized surface plasmons (LSPs) and the generation of a highly amplified electromagnetic field in their immediate vicinity. This can be exploited to initiate charge-driven reactions with an energy transfer between metal and adsorbate as the so-called “hot electrons” or “hot holes.” (Zhang et al., Acc Chem Res 52:2506–2515, 2019; Wang et al., Appl Mat Today 15:305–314, 2019; Ren et al., RSC Adv 7: 31189–31203, 2017; Linic et al., Nat Mater 14:567–576, 2015). Still, there are many fundamental questions unresolved, such as the catalyst-molecule electronic interactions and transfer and the structural (re)-organization of the adsorbates on the catalytic surface. Since the general aspects of plasmon-activated catalysis are discussed in another chapter of this book, in this chapter we focus mainly on the high lateral resolution aspects. The consequences for spectroscopy when using single plasmon particles as analytical probes will be considered specifically in the respective topics. If necessary, specific properties of the probe composition and its influence on the spectra and on the overall results of the reactions are also discussed.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_8

169

170

T. Deckert-Gaudig et al.

Keywords

Localized surface plasmon (LSP) · Nanoscale · Tipenhanced Raman scattering · Optical near-field

8.1

Short Introduction to TERS

Since a catalytic process in general involves several reaction steps, the aim is to detect intermediates highly specific – preferably in situ, in real time and with high spatial resolution. Raman spectroscopy can contribute significantly to this field as it detects components highly specifically – without the need for labeling. The detection sensitivity can be enormously increased by the use of plasmonic nanostructures in surface- and tip-enhanced Raman scattering (SERS and TERS) [1–4]. Both methods are based on the excitation of localized surface plasmons (LSPs) on rough metal nanostructures when irradiated with the appropriate laser wavelength. Under resonant conditions the sample molecules interacting with such plasmonic nanoparticles experience an enormous electromagnetic field. Thus, the resulting Raman signals are detected with an intensity that is enhanced by several orders of magnitude. Typically, nanoparticle films, arrays, electrodes, or colloids are used as enhancing devices. Although SERS can be operated on the single molecule level, the spatial resolution is limited by the diffraction limit and in the experiments often large spectral and temporal fluctuations are observed [5]. In SERS, the experimental reproducibility strongly depends on the substrate homogeneity, which is difficult to control – albeit, SERS has been successfully applied to study plasmon-induced reactions under ambient or under electrochemical conditions [4, 6–10]. It could be demonstrated that the inherent random distribution of the so-called “hot spots,” their temporal reorganization, diffusion processes, and conformational variations of the probed molecules are mainly responsible for these effects [2, 5, 11–13]. It is noteworthy that the spectral variability in single molecule SERS and also TERS experiments, can obstruct the clear band assignment and thus the identification of the compound. Therefore, a uniform arrangement of the sample molecules on the substrate is advantageous. Presently, the commonly accepted way to achieve this is via generation of self-assembled monolayers, i.e., by gold-sulfur or silversulfur bonding, although diffusion processes can play a role at ambient conditions [14]. As mentioned, SERS generally cannot be used for high spatial resolution detection. This changes when only a single nanoparticle is used as a sensor, particularly if only a specific orientation of such particle is applied. This approach is realized in TERS, which combines the sensitivity of SERS with the nanoscale spatial resolution of scanning probe

microscopy [3, 4, 15, 16]. In TERS a single plasmonic nanoparticle at the apex of an atomic force microscope (AFM) cantilever tip boosts the Raman signal in such a way that only a few or even single molecules (even without a Raman resonance contribution) can be identified. An etched/ sharpened wire can be used instead, as is mainly the case in scanning tunneling microscopy (STM)-TERS. In the experiments, the topography and the chemical structure of the probed sample are characterized simultaneously with (sub-) nanometer spatial resolution. In Fig. 8.1, three commonly used TERS configurations are sketched. In transmission mode the scattered light is collected in back reflection and the illumination and detection path utilize the same optical components. Ideally, setups are equipped with a high numerical aperture objective (i.e., oil immersion) to optimize the signal collection efficiency. It should be mentioned that in transmission mode only transparent samples can be analyzed by default, but a modified tip holder and an adapted sampling geometry can be used to also access opaque samples [17]. More commonly, top and side illumination setups equipped with a long working distance objective are used for nontransparent samples. For more details on TERS setups and experimental procedures it is referred to refs. [3, 15, 18] and literature cited therein. Depending on the scanning probe microscope incorporated in the setup, either etched gold or silver wires (mainly for STM-TERS) or nanoparticle coated AFM cantilever tips (for AFM-TERS) are commonly used. The excitation wavelength is chosen to match the absorption maximum of the nanoparticles at the tip (i.e., their plasmon resonance).

a)

b)

TERS tip Opaque sample Transparent sample Excitation

c) Scattering

Fig. 8.1 Schemes of commonly used TERS setups. (a) Transmission mode for transparent samples (b) Side illumination and (c) top illumination for transparent and opaque samples. TERS tip: etched silver or gold wire or nanoparticle coated AFM cantilever probe; excitation: laser wavelength in the visible range matching the absorption maximum of the nanoparticles at the tip

8

Nanoscale Raman Spectroscopy

171

A crucial step in each TERS experiment is the precise positioning of the tip in the laser focus. Then, the region of interest on the sample is moved under the tip and spectra are recorded pixel-wise or on predefined positions on the sample. Since the Raman signal-enhancing field at the tip apex is confined to a few nanometers or even less, only a few (down to single) molecules contribute to each spectrum. Thus, by correlating spectroscopic and topographic data, a molecular structure analysis with highest precision can be achieved. Interestingly, plasmonic nanoparticles cannot only be used as sensors but can also catalyze reactions when irradiated with the appropriate laser wavelength. The catalytic effect is based on the formation of LSPs around the tip, which can lead to the generation of the so-called “hot electrons”or “hot holes.” [8, 13, 19–21] An energy transfer to the probed molecules in close vicinity can trigger various reactions such as reduction, oxidation, isomerization, (de-)protonation, and bond cleavage. So far, numerous works demonstrated that TERS is very useful to track heterogeneous catalytic reactions, electrochemical processes, isomerization, and oxygen reduction reactions at metal organic complexes as illustrated in Fig. 8.2. This chapter gives an overview of the achievements obtained with TERS in the field of catalytic reactions. Most importantly TERS provides a direct experimental route to unravel reaction mechanisms and identify potential intermediates and products with a lateral resolution at or close to the single molecule level. It is important to emphasize again that in some experiments the tip acts as sensor only, while in others it influences or even induces the reaction simultaneously. Some research focuses on the chemical structural analysis of the reactants at varying experimental conditions. In other studies, the time-dependent tracking of dynamic Fig. 8.2 TERS as a tool for tracking catalytic reactions. If hot electrons or holes are involved a TERS tip can act as sensor and catalyst, simultaneously. (a) Oxidation and reduction reactions (e.g., coupling and cleavage of molecules) and (b) reversible configuration changes (e.g., cis-trans isomerization) can be initiated; a catalytically inactive TERS tip can be used to follow electrochemical reactions (c) or to characterize bimetallic catalytic surfaces (d)

processes is the main objective. Finally, there are also reports regarding the characterization of the catalyst itself (i.e., the metallic particles) to understand the effects of morphological defects and edges on its activity. All approaches pursue the same goal: to deepen the understanding of catalytic reactions on the smallest lateral scale possible. Initially, the theoretical background of plasmon-driven reactions is discussed as this is an important aspect to keep in mind when using TERS to investigate chemically active sites. After this section applications of TERS in the field of catalysis will be presented, emphasizing the aspects of the tip as an active or passive part of the experiment.

8.2

Theoretical Background of Plasmon-Induced Catalysis

Plasmon-induced catalysis is based on the transfer of electrons (or holes) from the sample surface to the tip (or vice versa) involving LSPs. The LSP is a coherent electron charge density oscillation in the conduction band of a metal that acts as a standing (or quasi-static) [22] wave at the dielectric interface. Among other phenomena, it creates a strong and localized enhancement of the electric field at the surface of plasmonically active materials. This principle is the basis of SERS in general and of specific interest naturally also for TERS since the enhancement of incoming and scattered radiation causes a significant increase of the intensity of the Raman signal and consequently of the technique’s sensitivity. Since the same physical phenomenon is the origin of the field enhancement and generation of highly energetic electrons (or holes), the simultaneous injection of electrons and probing of the changes induced in the molecular structure via Raman scattering are possible.

a)

d)

Coupling and cleavage

Bimetallic catalyst

stat

ntio

Pote

c) Electrochemistry

Hot e –

b) Cis-trans isomerization

Hot h+

– Ef +

8

172

The detailed process of hot carrier generation has been the topic of several reviews [6, 23–26]. Here, we summarize the main concepts leading to high energetic or “hot electron” transfer from metal to molecules and indicate optimization pathways. “Hot electrons” are created directly from LSP dephasing, while the word “hot” typically refers to electronhole pairs whose energy is higher than “just” thermal excitation, with the electron energy above the Fermi level (Fig. 8.3a, b) [27]. The production of hot electrons and thus their potential transfer to a surrounding or underlying substrate depends on plasmon damping mechanisms, their relative life-time, the nature of the metal, the size and geometry of the particles, to just name a few factors [28, 29]. In general, plasmonic dephasing can be divided into two mechanisms: radiative and non-radiative decay. Their respective importance depends on the nature of the metal and the size of the plasmonic particles. Radiative plasmon decay (i.e., leading to the re-emission of a photon) through fluorescence is considered as an inefficient mechanism, although several strategies have been developed recently to understand the phenomena [30–32]. Fluorescence from the molecules in contact with the metal (metallic surface or the plasmonic object itself) is also usually expected to be quenched [33]. However, plasmon-enhanced fluorescence [30] where the molecules (or particles) at the surface of nanoscale metallic structure can also be enhanced in the strong coupling regime [34–36]. These recent developments open the possibility for faster and higher sensitivity detection [30] that may be critical for resolving mechanistic details of catalytic activities. Importantly, radiative emission also implicitly includes elastic and inelastic photon scattering [37, 38], which is important for the detection in SERS and TERS. For hot-electron-based catalysis, the radiative pathways must be avoided. The competing dominant non-radiative mechanism involves the formation of a highly energetic electron-hole pair with a typical life time of a few to a few hundred femtoseconds [39]. The number of hot carriers produced and their spatial distribution are therefore approximately proportional to the LSP absorption [29]. By definition, the electron-hole pair energy distribution is nonthermal (out-of-equilibrium, Fig. 8.3b). Then, the hot carriers thermalize (energy redistribution, Fig. 8.3c) by creating a “hot” Fermi-Dirac distribution (an energy distribution similar to that found at high temperatures) [40] on a time scale of ~100 fs to picoseconds (depending on the hot carriers energy) [26]. This relaxation can happen in different ways, the main contribution is generally electron-electron scattering [41]. Thermalized electrons then relax through electron-phonon scattering (energy transfer into nuclei vibrational motion [26]), which happens on a time scale of up to a few ps. The latter implies heat generation (also often referred to as Joule effect [42]), which then dissipates “slowly” in the medium, i.e., generally on a much longer time scale of nanoseconds,

T. Deckert-Gaudig et al.

depending on the system geometry [23]. The impact of this local temperature raise on catalytic activity can be significant and will be treated in the next section. The energy of the produced hot electrons determines their transfer probability to a molecule. However, once such energetic electron-hole-pairs are formed, rapid relaxation pathways mentioned above render their transfer to an adjacent molecule inefficient. In general, electron transfer can be classified into two mechanisms: indirect and direct transfer processes. “Indirect transfer” refers to the formation of a hot electron, which is consecutively transferred to the lowest unoccupied molecular orbital (LUMO) of the adjacent molecules (Fig. 8.3d) [43]. This process suggests an overlap between the hot electrons energy level from the plasmonic particles and the electron accepting energy level from the substrate. This is particularly prominent for a transfer toward TiO2 substrates [44] or TiO2/nanoparticle interfaces [45], due to the d-orbitals’ availability. The hot carriers can also be transferred during the thermalization process if the available states of the molecules are close enough to the Fermi level [24, 25, 46, 47]. Another mechanism called chemical-interface damping (CID) implies the “direct transfer” of an electron from the plasmon excitation to a LUMO of the molecule, i.e., prior and without the formation of an hot electron-hole pair (Fig. 8.3e)

a)

b)

E

E

c)

ee–Ef

e–

Ef e–

d)

Ef

e)

E

LUMO e– Ef h+

h+

h+

E

f)

E

e–

e– HOMO

E

Ef

e– Ef

h+

ee–-

e–

Metal Adsorbate

Fig. 8.3 Schematic representation of plasmonic damping and the possibilities of an electron transfer into an adsorbate. (a) Neat metallic particle and energy level distribution prior to light absorption (thermal equilibrium); (b) irradiation and formation of highly energetic electronhole pairs in the metal; (c) thermalization and relaxation of the electronhole pair in the metal; (d) irradiation and plasmon generation where the indirect electron transfer leads to the formation of an electron-hot pair where electrons are transferred to a LUMO of the adsorbed molecule (e) chemical-interface damping (CID) where an electron in the metal is directly transferred to a LUMO of the adsorbed molecule; (f) direct electron transfer from a HOMO to a LUMO in the molecule. E energy, e
electron, h+ hole, Ef Fermi energy level, LUMO lowest occupied molecular orbital, HOMO highest occupied molecular orbital. The green arrow indicates irradiation with a wavelength matching the plasmon resonance of the metal

8

Nanoscale Raman Spectroscopy

[48, 49]. This process is known to be particularly efficient, especially at the interface between plasmonic particles and semiconductors that enables an efficient coupling of the states [50, 51] but is also applicable to organic molecules adsorbed on the metal surface [25, 49]. It has been recently suggested that CID mechanisms may be more bond selective and thus, enable a better control of catalytic activity [48]. It was proposed that plasmon-induced catalytic activity can happen via another type of “direct transfer” where LSP photon absorption leads to the direct excitation of an electron from an occupied to an unoccupied orbital of the molecule–nanoparticle system (Fig. 8.3f) [47, 52]. Similarly, it is also worth mentioning the so-called “plasmon-pump adsorbate excitation,” where a reaction that can be photo-activated at a particular wavelength in the absence of plasmonic activity is strongly promoted by local field enhancement in resonant conditions with the plasmonic activity [53–55]. For optimizing catalytic efficiency, improving both the quantity and the energy of the hot carrier generated and eventually transferred to a molecule is a main target [23]. This in turn implies a profound understanding of the reaction as for different synthesis a specific optimization of those “plasmonic” parameters is required. In general, smaller particles show less scattering (radiative decay) and higher absorption, which is the basic parameter for hot electron-hole formation. Plasmon life time can be increased, in part, through hybridization of plasmonic modes [56], especially for out-of-phase modes [57]. On the other hand, sharper plasmon modes generate more electronsholes pairs with higher energy, which has been shown to be more likely for silver than for gold particles [28]. The rapid oxidation of silver particles is however detrimental to their practical and more commercial use. Nowadays, the use of combinations or assemblies of metallic and/or semiconductor structures has become a popular approach, partly because it enables a better control of the reaction specificity [6, 23]. For instance, the Hallas group developed recently the concept of antenna-reactor complexes where the plasmonic activity of a particle is transferred to an adjacent non-plasmonic active particle enabling a better control of catalytic activity [58, 59]. Similarly, the use of core-shell plasmonic particles where the shell is composed of a semiconductor such as TiO2 has been shown as a promising avenue for optimizing hot carrier collection and for controlling the resulting reaction [60–62].

8.2.1

Thermal Effects

The Arrhenius equation demonstrates that, in the simplest scenario, reaction rates exponentially depend on the local temperature. In the context of plasmonic probes, highly localized plasmonic activity at the origin of hot carrier

173

formation simultaneously creates highly localized heating sources. As for hot carrier generation, heat release is a non-radiative process and is thus theoretically proportional to the amplitude of the plasmonic absorption [63–65], which is nm to Å scale localization dependent. Consequently, strong inhomogeneities of locally dependent thermal gradients may be found, even on a single particle, suggesting a spatial distribution of reaction rates. Thermal effects can act synergistically with hot carriers to increase reaction rates, which is ultimately a desired outcome. But it suggests the preservation of the chemical integrity of the molecules adsorbed or chemically linked to the plasmonic surface at high local temperatures. One must also keep in mind potentially competitive reaction pathways involved in both mechanisms, that are by definition, detrimental to generating the desired product with the highest efficiency [66]. Disentangling hot carrier transfer, near-field temperature rise and catalytic activity is a great challenge and a topic of active discussion in the field of plamonics [43]. In fact, differentiating qualitatively and quantitatively whether catalytic activity arises from hot carrier transfer or from near-field heating (or a mixture of two processes) is a topic of current research. Sivan et al. [67–70] recently compared numerically the hot electron energies with reaction rates observed in several experimental contexts and argued that catalytic activity naturally attributed to hot carrier transfer would actually arise mainly from thermal effects on the surface. Their novel theory based solely on Boltzmann-heat equations, energy conservation, and thermodynamic arguments show that most of the power density must go through thermal pathways (Fig. 8.4a) [69]. One of the main difficulties at addressing these questions experimentally is related to the difficulties of quantifying the temperature in such local and heterogeneous environment. We recently proposed a method to extract in situ and on-site near-field temperature information based on simultaneous measurements of the Stokes and anti-Stokes spectral ranges [72]. To date, most experimental studies have used techniques that provide insight into the role of heat in some hot electroncatalyzed reactions but were not able to fully address the local subtleties. In particular, the field of electrochemistry is currently especially active at investigating this aspect thanks to the time resolution and wealth of possibilities offered by current measurements, especially when used in combination with an AFM. Yu et al. [73], for instance, used scanning electrochemical microscopies and mass transport measurements at the tip to expose the often underestimated impact of heat on plasmon-driven catalysis. On the other hand, Rodio et al. [74] concluded from their wavelength-dependent photocurrent measurements that near-field heating had little impact on their reaction, which is believed to be hot-carrier driven. Using a combination of electrochemistry

8

174

T. Deckert-Gaudig et al.

7 × 101

7 × 102

1012

7 × 103

/phonon

Electron

108

heating

s

ctron

) ele

m

-ther

Nom

104 106

hot” al (“

7 × 104

107

108 |E|2

109

(V/m)2

b) 2 1

IRC′ ISC′ Itotal′

5.0 2.5 0.9 0.6

0

0.3

–1

ABS (IRC /ISC′)

Power (W/m3)

a)

Current (μA)

related to plasmonic-induced heat process studies in the field of plasmonic catalysis, but also the disagreement regarding interpretation of main mechanistic paths [68, 77, 79, 80]. Nevertheless, the active discussions imply that in the future a further deepening of the understanding of the process can be expected that will greatly help to control catalytic reaction conditions. Many of the items discussed also play an important role in plasmonic properties of ensembles of particles. It is important to note that the unique capabilities of the combination of scanning probe microscopy and near-field optical spectroscopy allows not only to investigate and even control chemical reactions with nanometer resolution, due to its controllability it also provides a valuable tool to test existing theories.

Iinc (W/m2)

0.0 0.39 0.33 0.30 0.27 0.21 0.18 0.15 0.12 0.09 0.06 E vs SHE (V)

Fig. 8.4 (a) Comparison of power densities going to thermal pathways (temperature increase by electron-electron scattering- green diamond – or by electron-phonon scattering – yellow triangles) vs. nonthermal pathway as hot carriers (blue squares) as a function the square of the incident electric field (or intensity). (Reproduced with permission from Ref. [69]) (b) Experimental measurement of rapid response current (RC and RC`) vs. slow response current (SC) as a function of the voltage applied at a silver electrode in a KOH electrolyte solution and under illumination with a 785 nm laser. Negative ratios of IRC/ISC (red) expose the dominance of the thermal heating at all voltages except at values close to bias voltage (0.15 V). (Reproduced with permission from Ref. [71])

measurements and finite element simulations Maley et al. [75] showed that a significant enhancement of current occurs related to near-field heat generation. Ou et al. [71] used a similar approach approximating the hot carrier transfer as a short-range effect while heat transfer is a long-range effect. By measuring the slow vs. rapid current response of Ag rough electrodes as a function of the applied voltage, they observed that slow current response was almost systematically dominating the total current and thus concluded that heat must be the dominating parameter (Fig. 8.4b). Zhan et al. [76] pushed a similar concept further by illustrating the difference in wavelength dependence of the two processes. Other methods such as in situ measurements of photocatalyst surfaces with a thermal imaging camera [77] and bath temperature gradients [78] as a function of illumination wavelengths were also used to differentiating/quantifying relative contributions to the mechanism. The few examples mentioned above illustrate already the research activity

8.3

Nanoscale Spectroscopic Investigation of Catalyzed Reactions

This following survey of practical applications of near-field spectroscopic investigations of surface reactions will be divided into two sections. The first and major part will discuss applications with a strong influence of the utilized nearfield probe itself. This is due to the fact that, as discussed in the previous section, a plasmon-induced electron or hole transfer will often influence the reactivity. Presently this active part of the probe is particularly utilized for systematic studies of otherwise difficult accessible plasmon catalyzed reactions on rough substrates. Particularly the distinct control of the localization and the particle-sample distance is an extremely valuable tool. In the second part we discuss applications that utilize mostly the analytical properties of the near-field optical probes without an active participation of the tip in the reactions observed.

8.4

Nanoscale Catalytic Reactions with Plasmon Contribution

8.4.1

pNTP and pATP Dimerization to DMAB and Other Azo Bridge Containing Molecules

Among plasmon-catalyzed reactions the dimerization of p-nitrothiophenol (pNTP) or p-aminothiophenol (pATP) to p,p’-dimercaptoazobenzene (DMAB) are studied particularly well. The processes had been known under SERS conditions for several years [9, 10, 81–84] when it was first reported in 2012, when pNTP was found to dimerize to DMAB under TERS conditions with 532 nm irradiation [85]. Evidently, a TERS tip can act as detecting and catalytic tool, simultaneously. In the experiments researchers took advantage of the wavelength-dependent behavior of the reaction, which is

Nanoscale Raman Spectroscopy

175

c)

nNN nNN nNO2

b)

nNN nNN nNO2

a)

bCH

1335 cm
1 and DMAB (C-H) at 1140 cm
1 and (-N¼N-) at 1440 cm
1 do not overlap, the reactant and the product could be directly identified from the TERS spectra (see Fig. 8.5c). Any fluctuations in the spectra in Fig. 8.5b were attributed to a disturbance of the monolayer, indicating the high spatial resolution of the measurements where only a few molecules contributed to a spectrum. The experiments using a silver substrate and a gold tip were confirmed at λ ¼ 633 nm under high vacuum (HV) STM-TERS conditions [86] and subsequently by further experiments [87–93]. Recently, a systematic approach to understand the contribution of the metal substrate was provided for the reaction of pATP and pNTP under plasmonic conditions on gold and silver (111) surfaces [92]. The irradiation of the configuration Au/pATP/Au tip with 633 nm yielded DMAB. In contrast, no reaction was observed at 532 nm because the required plasmon resonance on the Au tip could not be excited at this wavelength. Exchanging the Au tip for a Ag tip initiated the transformation to DMAB again and the authors proposed that this was due to the higher energy of the generated hot electrons at the silver surface. If a Ag substrate was used, the reaction did not proceed under any condition. Based on DFT (density functional theory) calculations, the authors considered the orientation of pATP on the different metal substrates. It was found that pATP on Ag was almost vertically oriented, which rendered the coupling of two molecules sterically unfavorable. In contrast, pATP was strongly tilted on Au, which explained the reaction of two adjacent molecules. It was concluded that in general not only the energy of the hot electrons generated in the substrate-tip gap plays a role but the orientation of molecules to each other. Finally, the results were compared with those from pNTP. Using a Ag tip, pNTP

generally not initiated when 633 nm is used. Consequently, the photo-catalytic reduction (oxidation state of the nitrogen atom changes from +III to
I) was initiated with green light while a 633 nm laser line was used for detection. The TERS results have demonstrated that a single Ag nanoparticle was sufficient for triggering the reaction and monitoring the dynamic process in situ. For a successful tracking of a chemical reaction with TERS, however, some issues have to be considered. Due to the high spatial resolution of TERS, only a few molecules under the tip are probed and the molecules orientation influences the spectra. In other words, different vibrational modes can be detected or are differently enhanced, depending on the functional group pointing toward the tip. In order to obtain the most uniform spectra of a compound, a homogeneous molecular arrangement on the substrate is advantageous. This way, spectral fluctuations can be minimized and dynamic reactions can be tracked more efficiently. If TERS spectra are acquired time-dependently, possible intermediates of a chemical reaction can thus be identified more easily. In the experiments a uniform sample adsorption was realized by immobilizing pNTP on a smooth gold nanoplate as sketched in Fig. 8.5a [85]. The obtained self-assembled monolayer (SAM) permitted the recording of stable TERS spectra of the starting material. When the sample was excited with 633 nm a constant TERS signal was collected within the first 90 s as shown in Fig. 8.5b. Switching the light source to λ ¼ 532 nm for 30 s (indicated by the white bar in Fig. 8.5b) initiated the reaction and the spectra changed drastically when spectra acquisition was resumed at λ ¼ 633 nm excitation. Since the characteristic bands of the functional groups in pNTP (-NO2) at

bCH

8

0.8

S

pNTP 0.6

200 0.4

S

Intensity

S

S

100

NN

NO2

Time (s)

NO2

DMAB 300

0.2

1600

1400

1200

1000

Wavenumber (cm–1)

Fig. 8.5 (a) Sketch of the TERS setup used to monitor the catalytic reaction of a pNTP monolayer to DMAB on a single gold nanoplate; (b) Time-dependent TERS spectra intensity plot (λ ¼ 633 nm, P ¼ 380 μW, tacq ¼ 5 s/spectrum), the gap indicates a period of 90 s when the sample

1600

1400

1200

1000

Wavenumber (cm–1)

was irradiated for 30 s at λ ¼ 532 nm to initiate the reaction; (c) Selected spectra from (b) indicate pNTP (black) and DMAB (red), the dotted spectrum shows the reference spectrum on glass. ((b) and (c) are modified from Ref. [85] with permission from Nature Springer)

8

176

on Au substrates reacted to DMAB at 532 nm but not at 633 nm. Interestingly, neither pNTP nor DMAB were detected when flat Ag substrates were used. According to theoretical considerations, pNTP adopted a tilted orientation on Au but was almost flatly aligned on silver due to involved Ag-NO2 interactions. Thus, the nitro groups were perpendicular to the tip and experienced only a small part of the electromagnetic field enhancement, whose main part in TERS is known to be parallel to the main axis of the tip. The effect of the molecular orientation on the band intensity was also observed. A similar reason for changing band intensities in pNTP spectra was reported from numerical simulations [93]. However, the focus of this AFM-TERS-based work was on explaining why in a minute part ( 0.6 V the deprotonated species was present. In order to obtain an insight into the orientation of adenine on the gold surface DFT calculations were performed. In agreement with the experimental results it was suggested that protonated molecules were oriented perpendicular/tilted to the substrate and reversibly changed to a flat geometry after proton release. The latter was enabled by strong Au-N interactions. The detection of subtle differences in EC-TERS was only possible due to the high specificity and spatial resolution of TERS. It was assumed that only 50–70 adenine molecules contributed to a spectrum.

8.5.3

Cleavage of Water

An interesting study reported on the electrochemical cleavage of water, which was investigated with EC-STM-TERS in aqueous sulfuric acid. At E ¼ 1.45 V the generated oxygen oxidized spatially confined edges on the Au(111) electrode. Using 633 nm for excitation and a gold tip, TERS-cyclic voltammograms were recorded and the reactivity changes on the Au substrate were discriminated with 10 nm spatial resolution [115]. The EC-TERS data indicated the formation of two different AuOx species and the redox reaction was found to be reversible indicated by the absence and presence of the characteristic Au-Ox at band at ~560–580 cm
1. It is noteworthy that the potential at the gold TERS tip was kept constant at 1.0 V to prevent its oxidation. The correlation of sample topography and TERS map revealed the heterogeneity of the Au surface. Interestingly, not all detected 2–4 nm

8

180

T. Deckert-Gaudig et al.

high surface defects were catalytically active and in the most effective areas the oxidation process stopped after the formation of six monolayers of AuOx (¼ 3 nm). From variations in the band position between 560 and 580 cm
1 the presence of two species was concluded, i.e., Au2O3 on flatter defects and Au2O on sharper protrusions. From the overall results it was estimated that approximately 4.5% of the surface contained defects.

8.5.4

Manipulating Phthalocyanine

In EC-STM-TERS experiments under UHV conditions, a silver tip was used to manipulate a cobalt phthalocyanine (CoPc) monolayer adsorbed on a gold surface. As already mentioned, metal phthalocyanines are promising candidates to replace Pt in oxygen reduction reactions. Therefore, it is not only important to analyze the complexation of gaseous molecules but to track associated dynamic configuration changes of the phthalocyanine molecule in situ. A possible approach to assess and control those processes is the variation of the substrate potential. It was demonstrated that by sweeping the applied sample potential (Esample) at a constant TERS tip potential (Etip) the transfer of CoPc from the substrate to the probe was enabled [116]. After the tip was lifted from the sample, CoPc signals could be detected at the tip indicating that CoPc was attached to the tip. It was postulated that a bias potential threshold above Esample
Etip ¼ 0.15 Vbias was necessary to induce a strong dipole in CoPc. Thus, the attraction forces between the oxidized molecule and the Ag tip overcame the adhesion forces to the Au substrate. If the potential sweep was reversed CoPc desorbed and a clean TERS tip was regained. The transfer of CoPc from the substrate to the tip could be avoided by working at a constant Vbias ¼
0.05 V. In the experiments using 633 nm for excitation a shift of Esample ¼
0.06 to
0.6 V resulted in a notable signal intensity decrease. This observation was explained by the reduction of CoPc, which resulted in the loss of the resonance Raman effect. A detailed insight into the arrangement of CoPc molecules on the Au substrate was presented in a follow-up study. In the EC-STM-TERS experiments the ordered molecule arrangement could be disturbed by driving Esample down to 0.1 V, which resulted in the formation of a (non-ordered) diffusion phase [117]. At this point, it should be mentioned once again that during switching to negative potentials CoPc is reduced. EC-TERS spectra were acquired substrate potential dependent at different excitation wavelengths. Interestingly, at λ ¼ 633 nm a signal intensity decrease with increasingly negative potential was observed. At λ ¼ 700 nm the opposite behavior was reported while at λ ¼ 720 nm no change was detected. Complementary cyclic voltammetry (CV) and UV-vis measurements showed that the absorption band of CoPc was

red-shifted when the molecule was reduced. In other words, the resonance effect Raman effect was no longer observed at 633 nm but at 700 nm. A definite explanation why no changes occurred at 720 nm could not be provided. It was merely postulated that the substrate-molecule interactions must have played a role. A further model catalyst for oxygen reduction reactions is iron phthalocyanine (FePc) and two oxygen-catalyzed processes were followed time-dependently using ECSTM-TERS. While the observed flat and tilted molecular orientations on the Au(111) substrate were reversible, the loss of the central Fe atom was found to irreversibly damage the molecule [118]. When the initially applied potential was quickly switched between 0.7 and 0.4 V, a reversible intensity decrease and increase of some bands in the TERS was observed. The spectral changes could be correlated with topographic changes based on irregularities of the height profile in the hn0

Ground state E0

Rayleigh filters

(i)

Focusing mirror

Laser

Beam expander

Grating (iii) Microscope

(iv) (ii) CCD detector

Focusing mirror

Sample

192

transition energy of the sample and hence produce a signal on its own. If higher laser energies (closer to UV region) are utilized, then the probability for fluorescence is significantly increased. Lasers in the visible light region, typically exhibiting a wavelength in the range from 400 nm to 900 nm are commonly utilized for the acquisition of Raman spectra to counterbalance fluorescence effects. The use of near infrared (NIR) or near UV lasers has recently been reported, bearing specific advantages such as diminished damaging of the sample and reduction of background fluorescence in the case of NIR, or increased Raman scattering intensities under UV illumination [19]. The laser power can vary from few μW to hundreds of mW. Even though, the Raman signal is proportional to it, it is advisable to use a rather low excitation power in order to avoid thermal decomposition of the samples [19]. Microscope objectives have a high impact on the spatial resolution of Raman signals. A high magnification leads to higher axial resolution at the cost of a shorter working distance and vice versa. Low magnification is therefore recommended for the analysis of liquids and bulky homogeneous samples [15]. After irradiating the sample with the light source, collection of scattered photons is conducted either in a 90
(conventional) or in a 180
(backscattering) scattering geometry. In the former, the light beam irradiates the sample and the scattered light is collected at 90
by means of a lens, whereas in the case of the 180
geometry, the scattered light is received through the same lens where the laser was emitted using a mirror [18]. Commercial Raman spectrometers are generally available in two versions, the dispersive and the nondispersive configuration [20] depending on the means by which Rayleigh scattering is secerned. In the case of dispersive devices, single or multistage monochromators and/or special filters, such as the holographic notch and dielectric edge filters are used to achieve spatial separation of wavelengths, whereas nondispersive instruments integrate an interferometer and use Fourier transformation to obtain Raman spectra by frequency modulation. Single monochromators involve a diffraction grating, typically in the range from 150 to 4000 grooves per mm, to disperse Rayleigh scattered photons. Even though a high grating frequency grants better spectral resolution, the spectral intensity and range are reduced compared to low grating frequencies [15]. Multistage monochromators operates as a separator for, firstly, Raman scattering from other radiation, and secondly, for the individual Raman peaks [18]. Nondispersive Raman instruments usually employ NIR lasers, such as the neodymium-doped yttrium aluminum garnet (Nd:YAG, 1064 nm), which is particularly interesting for samples requiring that fluorescence is almost completely avoided. Yet, since the probability of Raman scattering events is also reduced, relatively high laser power and highly sensitive interferometer detectors are required [18]. Raman spectra are typically presented as the Raman scattering intensity

W. Schuhmann et al.

(number of photons) as a function of the frequency shift (reciprocal centimeters) over a range from about 10 up to 4000 cm1. The count of photons can be normalized by the exposure time (counts per second) or/and the laser power (counts per mW). Due to the low probability of inelastic scattering, Raman spectroscopy requires highly sensitive detectors. CCDs are typical detectors used in Raman spectrometers and consist of multichannel arrays capable of collecting charge from scattered photons. They exhibit comparatively high quantum efficiencies and low signal-to-noise ratios as compared to other detectors, including photomultiplier tubes or photodiode arrays [15], and are capable of monitoring several Raman shifts concurrently [21]. The suitability of the different components and types of Raman spectrometers depends on the specific application. For instance, for biological and fluorescent samples it is preferable to use NIR lasers to avoid sample damage and background interference, respectively. For samples with high micro-heterogeneity, achieving high spatial resolution could be more important than it is for homogeneous samples. Other criteria to consider include the sample mode, data analysis, and operating costs. For a detailed discussion on those criteria, we would like to refer the reader to specialized literature [21].

9.3

Surface-Enhanced Raman Spectroscopy

Due to the scarcity of inelastically scattered photons, the signal intensity in conventional Raman spectroscopic measurements is inherently low and, hence, the collection of spectra with a reasonable signal-to-noise ratio often requires long exposure times and/or high-intensity laser illumination. Additionally, the typically small cross sections for Raman scattering are about 1030–1025 cm2 per molecule, thus requiring a high number of analyte molecules to be detected [22]. Consequently, the applicability of Raman spectroscopy for the detection of small quantities of molecules at a short time scale seems to be precluded. However, observations made in the 1970s paved the way to overcome that limitation. Extraordinarily intense Raman spectra of pyridine adsorbed on electrochemically roughened silver electrodes were initially reported by Fleischmann et al. [23, 24] and later explained by Van Duyne et al. [25] and by Albrecht and Creighton [26]. That effect, nowadays commonly referred to as surface-enhanced Raman spectroscopy (SERS), increased the signal intensity in these initial measurements by a factor of up to 106 with respect to conventional Raman measurements [18]. Henceforth, Raman spectroscopy got renewed attention by the scientific community since it was no longer necessarily limited by low signal intensities, and the development of surface-sensitive Raman analysis advanced rapidly. In 1997, Kneipp et al. [27] concomitantly

9

Operando Electrochemical Raman Spectroscopy

193

with Nie and Emory [28] reported the ultimate sensitivity goal: the first detection of single molecules using SERS with enhancement factors (EF) of up to 1014–1015. With this, a new stimulus for the use of Raman spectroscopy as state-ofthe-art analytical technique was provided. A vital role in the process of developing SERS-related applications was the tremendous progress in the fabrication of nanomaterials, which are the main contributor for high signal enhancements observed in Raman spectroscopy [29]. For instance, it has been shown that nanostructuring materials of sizes below 100 nm gives rise to extraordinarily high Raman scattering intensities [30]. The exact mechanism of SERS is still strongly debated. Nevertheless, researchers have come to the conclusion that more than one phenomenon is responsible for the large signal intensity enhancements. In the following section, we provide a brief summary of the basic physics underlying SERS, which should be sufficient to grasp the fundamentals of the phenomenon necessary to understand case studies discussed later. Fabrication and possible designs of Raman substrates will be discussed in Sect. 9.3.2.

9.3.1

Surface-Enhanced Raman Spectroscopy

SERS relies not only on the interaction between the incident light and the probed molecules, but also on the interaction of the light with electron clouds in metal nanostructures [31]. Light, as an electromagnetic wave, can excite the conduction electrons in a metal and cause them to oscillate collectively thus forming surface plasmons. Ag and Au, for example, oscillate at frequencies in the visible light region. Since in Raman spectroscopy the most commonly used lasers emit visible light, Ag and Au turn out to be excellent substrates for Raman experiments. During resonant excitation, the oscillating electric field of the incident light, e.g., a laser beam, causes oscillation of the electrons in the metal nanostructure and thus induces a charge separation, namely, a localized surface plasmon resonance (LSPR). Fig. 9.3 Electromagnetic field enhancements of a NP upon excitation induced by an electromagnetic wave. The enhancement occurs as a consequence of LSPRs in the respective NPs acting as nanoantennas. (Adapted from ref. [32, 33])

For illustration, a nanoparticle (NP) smaller than the wavelength of the incident light is represented in Fig. 9.3, in which an electromagnetic wave induces a dipole within the structure with its magnitude being defined by both the polarizability of the metal and the electric field strength of the light. Furthermore, a Hertzian dipole, depicted by the field lines in Fig. 9.3, is evolved forming a “nanoantenna.” [32] Elastic light scattering of the nanostructure leads then to an increased local electrical field around it as compared to that of the incident light. Next, the influence of the local electrical field on molecules attached to the nanostructured metal surface has to be considered, since, similarly to what occurs at the substrate, a dipole is induced at each of these molecules. The induced dipole of a molecule depends on its electronic polarizability, and the incident electric field can be modulated by the vibrating molecule at the substrate surface. If it is not modulated by the vibrational frequency of the molecule, then Rayleigh scattering occurs, whereas if it is modulated, Stokes and anti-Stokes shifted Raman scattering can take place. The intensity of the scattering in SERS depends on both the incoming and the outgoing fields. Thus, optimization of the SERS enhancement requires that the incident radiation and the Stokes Raman radiation are resonant with the plasmon peak of the metal substrate. Plasmonic excitation also generates “hot electrons,” which constitute the core of plasmonic catalysis and the foundation for solar-chemical energy shift reactions. The generation of hot electrons can be transferred to molecules and take part in (electro)chemical reaction at the surface of the plasmonic substrate. This can potentially render plasmonic substrates as taking a cocatalytic role in the conversion of chemicals by activating molecules such as H2 or O2 [34, 35]. It is important to mention that extremely high-field enhancements are observed in the so-called “hot spots.” As an example, the gaps between NP can be mentioned [36], since they offer an extraordinary, yet localized electromagnetic enhancement enabling even single molecule detection [27, 28]. As schematically depicted in Fig. 9.4 hot spots can

Electromagnetic wave Hertzian dipole

–––

+++

Nanoparticle

Nanoparticle +++

––– Electron cloud

9

194

W. Schuhmann et al.

High EF

Nanoparticle

Nanoparticle

Low EF

Fig. 9.4 Electromagnetic surface EF reaches its maximum between closely spaced nanostructures with overlapping plasmon resonances

be induced by coupled dipolar plasmon resonances between two nanostructures. The hot spot theory suggests that these spots make up the largest fraction of the overall surface enhancement in a given area, whereas the majority of the area remains SER inactive. The elucidation of the exact nature of the enhanced intensities remains an active field of research. Initial thoughts expressed by Fleischmann et al. [23] implied, yet not stated, that the increased surface on roughened electrodes accommodates a higher amount of analyte molecules leading to a higher signal intensity [37]. However, that theory was disproved by Van Duyne et al. [25, 38, 39], by Albrecht and Creighton [26], and later by means of theoretical calculations by Gersten, who additionally predicted EFs as high as 1011 [40–42]. Noteworthy, the enhancements described so far take place in the absence of adsorbed molecules, since they are an intrinsic property of the substrate. It is generally assumed that in addition to electromagnetic field-related effects, there are other effects that may account for a small fraction of the higher signal intensities, as it is the case of chemical enhancement. For example, chemical signal boosts can be induced by means of charge transfer resonance [43] or via addition of anions such as chlorine [44]. However, the chemical enhancement may only account for a factor of ~10 as compared to 103–1010 from the electromagnetic contribution [45]. Nevertheless, the difficulty of the definition of EFs coupled to the variety of methodologies for their estimation results in significant fluctuations of EFs reported in the literature [46]. Therefore, to obtain profound knowledge about EFs we would like to redirect the reader to further in-depth studies [47].

9.3.2

SERS Substrates and Fabrication

On the forefront of the increasing popularity as well as the increasing sensitivity of SERS measurements is the progress in the fabrication of nanostructured substrates by means of a multitude of different techniques, enabling optimized and reproducible fabrication of ordered highly active substrates

with large surface areas (cm2) [48]. Typical SERS substrates consist of colloidal NP [49, 50], NP dimers/trimers [51], or ordered arrays of voids [48, 52–54], domes [55, 56], or nanowires [57]. Presently, advanced methods for their preparation are available, including inkjet/screen printing [58, 59], direct deposition of plasmonic metal nanostructures [60], or nanosphere lithography [61]. The latter involves templating of a substrate with polystyrene nanospheres with subsequent vapor [62]- or electrodeposition [48] of the metal on the modified substrate. Even though advanced preparatory methods are available, the most commonly used systems for SERS experiments are colloidal NP solutions or self-assembled monolayers thereof [63, 64], which can enable single molecule detection. It turned out that SERS substrates should comprise coinage metals such as Ag, Au, and Cu since they offer the highest signal enhancements [65]. Several alternative materials have been tested and it has been demonstrated that SERS can also be observed using other transition metals, such as Pt or Pd [66], though most research relies on those metals supporting plasmons in the visible region [67].

9.3.3

Extensions of SERS

The applicability of SERS is severely limited by the prerequisite of using nanostructured coinage metals, firstly because fundamental single crystal experiments are excluded, and secondly because typically used materials, including Cu and Ag, are not stable in oxidative environments, hampering the application of SERS at low pH values as well as at highly anodic conditions. Different approaches have been developed to overcome these limitations and to extend the field of use of SERS. One of these approaches is known as tip-enhanced Raman spectroscopy (TERS), developed in the early 2000s [68–70], which overcomes the need for nanometric rough surfaces as well as the substrate/adsorbate specificity, while greatly enhancing the spatial resolution. The working principle emulates that of scanning probe microscopies, such as scanning tunneling microscopy (STM) and atomic force microscopy (AFM), in which a tip is closely approached to any sample substrate. From the “hot spot hypothesis” it is known that the strongest enhancements are generated in the region between nanostructures in close vicinity to each other. Similar effects can be induced if a sharp coinage metal tip, fabricated typically from Ag or Au (e.g., a silvered AFM tip or a thin wire), is moved closely to a sample under concomitant laser illumination of the tip. The illumination excites the localized surface plasmons at the tip apex, and, acting as antenna, an enhanced EM field is generated in its close proximity [38]. Thus, as soon as the tip is in tunneling distance, significant Raman enhancements of molecules attached to non-enhancing substrates are observed

9

Operando Electrochemical Raman Spectroscopy

a)

Ag or

195

b)

c)

High EF

Au

Ag

Non SERS-active substrate

Au

Ag

Non SERS-active substrate

Au

Non SERS-active substrate

Low EF

Fig. 9.5 Different strategies for inducing SERS activity on otherwise non SERS-active substrates. Molecules under study are represented by black bars. The color scale bar at the right displays the intensity of the EFs, with red and blue depicting the high and low intensities, respectively. (a) TERS setup in which a metal tip (Ag or Au) is approached

closely to the surface under study. (b) Direct deposition of SERS-active surfaces, such as NP or discontinuous films onto the substrate. (c) A non SERS-active shell consisting of the material of interest coated onto AuNP

(Fig. 9.5a). The small dimension of the tip enables unprecedented spatial resolution since the enhancement is localized under the tip. Advances in TERS even include the utilization of liquid cells, rendering it an excellent technique for the simultaneous acquisition of chemical and structural data during changes of dynamic interfaces, as it is the case of electrochemical processes. However, the comparably low EF and the need for sophisticated and often expensive instrumentation, such as position controlling elements or electrochemical TERS cells, still hamper its popularity. Another approach to allow for substrate generality is the so-called “strategy of borrowing SERS activity” [71], which was proposed by Van Duyne et al. in the early 1980s [72], in which a discontinuous film of Ag islands was deposited onto the substrate as schematically depicted in Fig. 9.5b. It was found that the overlayer is able to enhance the recorded resonance Raman scattering intensity by exploitation of the long-range effect of the electromagnetic enhancement. However, limitations arise when adsorption, taking place specifically at the non-SERS-active surface is the subject of the study since it turned out that most molecules prefer to adsorb at the SERS-active sites. To overcome this, a reversed approach was later introduced by Fleischmann et al. [73, 74] in which a thin layer of the material of interest was coated onto a SERS-active plasmonic substrate, for instance a AuNP, to spectroscopically study the molecules adsorbed on the non-SERS-active layer (Fig. 9.5c). However, the enhancing electromagnetic field attenuates rapidly within a few nanometers, as it was shown in a report where a thickness of 6.8 nm (24 atomic layers) of Pd on Au offered a considerably diminished SERS activity as compared to 0.7 nm (2.5 atomic layers) [71]. Consequently, the films used for this purpose should consist of only a few monolayers. Nonetheless, depositing ultra-thin layers requires considering several phenomena concerning surface-specific experiments. Firstly, the layer has to be deposited without the presence of

pinholes, so that the recorded signal is substrate-specific for the non-SERS-active material. Secondly, the electronic properties of metals may change, since electron transfer can occur due to the differences in the Fermi level. In other words, the electronic properties of the shell may be different to those of its bulk counterpart. Several strategies for the preparation of such core-shell NP have been developed, including Cu underpotential deposition and subsequent redox replacement [75, 76], electrodeposition [77], and chemical vapor deposition [78]. With the ability of coating AuNP with different metals, the substrate generality was greatly improved. However, the requirement of coating NP with the desired material still hampers some applications, since fabrication protocols for the specific application have to be developed for different substrates. The utilization of coatings or films of the SERS active materials may be the key in some stationary studies, in which chemical components are the subject of study. That technique is, however, not suited for electrochemical studies since reactions at the SERS substrate could be induced due to the electrical contact of SERS substrate and the probed material. Moreover, the electronic properties of the coated metal could be changed, which can further complicate the spectral analysis. A breakthrough that greatly enhanced the applicability of SERS substrates was reported in 2010 by Tian et al. [79], in which AuNP were coated with inert ultrathin films of SiO2/ Al2O3/MnO2 The coating prevented, on the one hand, direct (electrical) contact of the substrates with the AuNPs, and on the other hand, their agglomeration. These engineered plasmonic substrates, called shell-isolated nanoparticleenhanced Raman spectroscopy (SHINERS), allow for monitoring surface processes via SERS at biological samples or single crystal surfaces [80]. Furthermore, a significant amount of time is saved during sample preparation since the NP powder can simply be spread on the probed surface. For electrochemical measurements SHINERS is extraordinarily

9

196

W. Schuhmann et al.

suitable because electrical contact of the isolated NP and the probed substrate is avoided by the nonconductive coating. Thus, the detection of reaction products and intermediates during electrocatalytic and heterogeneous thermal reactions [81] on any surface with a high sensitivity and a high spectral resolution is enabled [82, 83].

9.4

Operando Electrochemical Raman Spectroscopy in Electrocatalysis

Up until this point, all necessary tools for the understanding of Raman spectroscopy and its different varieties have been described. In the following sections, a general introduction of electrochemistry and in particular its coupling with Raman spectroscopy is provided. Selected examples from literature are highlighted, which, in our opinion, represent key aspects of Operando Electrochemical Raman spectroscopy (OERS).

9.4.1

Coupling Raman Spectroscopy and Electrochemistry

Electrochemistry, as a branch of physical chemistry, is an interdisciplinary science dealing with the interaction of electrical energy and chemical bonds. An externally supplied current can drive a chemical reaction, as in the case of electrolysis devices, or, for instance, reactions can be exploited to generate a flow of electrons, as it occurs in batteries. Electrochemistry spans its applications through different fields of science, including (but not limited to) chemistry, biology, and engineering, for example in sensing or energy conversion devices even on the industrial scale. At the base of these technologies stands the fundamental understanding of the electrochemical processes occurring at the interface between the surface of a solid electronic conductor (electrode) and an ion-conducting phase (electrolyte). These processes take place in an electrochemical cell, which consists of at least two electrodes immersed in an electrolyte. The presence of ions at the interfacial region between the electrolyte and each of the electrodes induces charges and an equilibrium is established as depicted in Fig. 9.6a. If a source of electrons, such as a power supply, is connected and electrons are supplied to one of the electrodes the equilibrium formed between the metal and the ions in solution is disrupted. In case of excess electrons supplied to a metal plate the electrons will flow to the metal-solution interface and can be transferred to species in solution to invoke an electrochemical reduction reaction. At the same time, an oxidation reaction as counter reaction takes place at the other electrode/electrolyte interface. The electron flow is equal on both electrodes, guaranteeing charge neutrality within the system. Hence,

a) Positive net charge e–

Negative net charge

b) Power source

e– 4 e–

e–

e–

e–

e–

4 e–

4 e– e– e– e–

+ + +

+ e–– + e– + e

O2 + 4 H+ 2 H2O

4 e–

4 H+ 2 H2

Fig. 9.6 (a) Isolated electronic conductors immersed in an electrolyte exhibit surface charge. (b) Schematic representation of a water electrolysis cell, in which two electrodes are connected by a power supply driving water oxidation at the anode and proton reduction at the cathode

each electrochemical reaction consists of an oxidation and a reduction reaction taking place at a different electrode, and the individual reactions are called half-cell reactions. An example of an electrochemical reaction is water electrolysis (Fig. 9.6b) [84], in which water is split at the anode forming O2 and protons while releasing four electrons, which are later transferred to protons at the cathode, reducing them and yielding 2 H2 molecules. In general, electrochemical reactions are heterogeneous processes, which involve electron transfer from an electrolyte to an electrode material or vice versa. A complete understanding of the processes taking place at electrode/electrolyte interfaces requires structural characterization of the electrode surface under electrochemical reaction conditions, i.e., operando conditions. The interest in detecting very low concentrations operando in dynamic processes grew rapidly in recent years. One of the main challenges encountered by OERS is the necessity of integrating components that allow technical coupling of both, the electrochemical and the spectroscopic analysis. OERS was initially reported by Fleischmann et al. in 1974 [23]. Adsorption of pyridine at silver electrodes under potential cycling conditions was investigated using a three-electrode configuration electrochemical cell designed with a polished optical glass window through which the sample was irradiated (Fig. 9.7a). Similar and more sophisticated cell designs have been reported [85, 86], some integrating, for instance, submersible objectives to eliminate background signals and reduce laser power [87]. Kornienko et al. [88] designed a quartz crystal microbalance (QCM) spectroelectrochemical flow-cell, which enabled concurrent electrochemical, spectroscopic, and gravimetric analysis (Fig. 9.7b). QCM and Raman spectroscopy were used to

9

Operando Electrochemical Raman Spectroscopy

a)

197

b)

To reference electrode

Optical window

Subsidiary electrode Working electrode

Pt C.E.

O-ring Ag/AgCl R.E.

Radiation scattered at 90°

Flow in

Optical flat window

Flow out

Quartz/Au W.E. Laser

c)

d) Computer

Microscope objective Spectrometer

Laser

Stepper motors (x,y,z)

Su per hea d

AD/DA Potentiostat

z

WE

x

a

z

Furnace

SOFC MEA

Coarse manipulator (z, a)

y

9

Vibration damping table RE

WE CE

Cooling coil

Quartz enclosure Silicone stopper

Alumina tube Fuel Fuel exhaust feed Air feed Air exhaust

Fig. 9.7 Electrochemical cells for the collection of Raman spectra under operando conditions. (a) Electrochemical cell integrating an optical glass window to allow sample irradiation [23]. (Reproduced with permission of Elsevier). (b) Electrochemical flow-cell coupled to a QCM and Raman spectroscopy. (Reproduced from [88] with permission from the Royal Society of Chemistry). (c) Scanning

electrochemical microscopy setup for the precise positioning of the working electrode under laser irradiation [90]. (Reproduced with permission of John Wiley and Sons) (d) Solid oxide fuel cell design for the acquisition of Raman spectra at 715
C [91]. (Copyright (2007) American Chemical Society)

monitor structural changes and the nature of metastable phases, respectively, during the electrocatalytic evolution of oxygen and hydrogen on CoPX catalysts with continuous removal of the formed gases. Integration of high-precision positioning scanning electrochemical microscopy (SECM) setups with Raman spectroscopy [89] and SERS [90] analysis have been reported for the investigation of local electrochemical surface processes (Fig. 9.7c). High-temperature OERS was demonstrated by Pomfret et al. [91] for the investigation of temperature-dependent structural changes of YSZ electrolytes and redox processes of nickel using a solid oxide fuel cell operating at temperatures above 700
C (Fig. 9.7d). Although technical solutions have been already developed for OERS, there are still limitations regarding operating

conditions including highly corrosive environments and high temperatures. Additionally, typical SERS substrates (Cu, Ag), are not stable in oxidative environment and hence they can only be used for cathodic reactions such as the oxygen reduction reaction (ORR) [92], carbon dioxide reduction (CO2RR) [93] or the hydrogen evolution reaction (HER) [94].

9.4.2

Central Topics in Electrocatalysis

One important advantage of electrochemical reactions is that the rate of electron supply can be easily controlled by means of a power source. By enhancing the electron flow,

198

reaction rates can be altered and even different reactions or reaction steps can be triggered. If a catalyst is used on the electrode surface the approach is known as electrocatalysis, which constitutes the very core of many sustainabilityfocused inventions including electrosynthesis of chemicals, conversion and storage of energy, carbon dioxide recycling, among others. However, scarcity and cost of the required catalyst materials, which are often based on noble metals, impede their grid-scale implementation. Hence, the rational design of either more efficient or cheaper catalyst materials is critical for the development of many sustainable technologies and processes. In order to do so, it is crucial to obtain an in-depth knowledge of the reactions in question, including understanding the adsorption/desorption processes of chemical species at the interface, as well as identifying active sites and formed intermediates while a certain potential is applied to the electrode. OERS can be used for such purposes. Operando, in this context, refers to monitoring the catalyst in its working state with concurrent characterization of activity and, in some cases, selectivity [95]. It is hence necessary to probe the catalyst state in dependence on the applied potential, temperature, or pressure, while allowing for the determination of intermediates and/or reaction products [96]. Probing the electrode/electrolyte interface under reaction conditions may pave the way for catalyst optimization as well as for gaining fundamental understanding of interfacial processes. For example, identification of active sites [97, 98] could enable rational tailoring of catalysts in a way that the active atoms/motifs are enriched on the surface, ultimately improving the catalytic performance of the material [88]. Identifying reaction intermediates and elucidating the role of adsorbed intermediates in the mechanism of cascade reactions, such as oxidation of alcohols [99] or reduction of CO2 [100], remains a daunting challenge, which could be the key for tuning the selectivity of those and other relevant reactions for electrosynthesis of high-value products [101]. A multitude of techniques is nowadays utilized to solve related questions, such as nuclear magnetic resonance spectroscopy (NMR) [102], X-ray photoelectron spectroscopy (XPS), or X-ray absorption spectroscopy (XAS) [103]. However, both sample preparation, and technical and experimental requirements hamper the feasibility of these experiments. Raman spectroscopy, as a tool utilized under ambient conditions and especially in liquids, seems to be an excellent candidate for an operando electrochemical technique, with the additional advantage that unlike in IR spectroscopy aqueous solutions do not lead to interferences with the surface signal [104]. Furthermore, it does not require extensive sample preparation and the method is only minimally invasive if the material under study does not undergo photo-conversion and if moderate laser powers are employed.

W. Schuhmann et al.

9.4.3

Case Studies on Electrocatalytic Energy Conversion

In the following section, we focus on key studies involving energy conversion and storage technologies, such as water electrolyzers, fuel-cells, and batteries. We show how the application of Raman spectroscopy and SERS under operando conditions allowed monitoring catalyst states as well as reaction intermediates/products formed during different energy conversion-related electrocatalytic processes.

Water Electrolysis: OERS and SECM As depicted in Fig. 9.6b, water electrolysis consists of two half-reactions: the oxygen evolution reaction (OER) taking place at the anode (Eq. 9.1a and b), and the HER, occurring at the cathode (Eq. 9.2a and b) 

þ

2H2 O ! 4 e þ 4 H þ O2 , acidic OER 



4OH ! 2 H2 O þ 4 e þ O2 , þ



2H þ 2 e ! H2 , 



2H2 O þ 2 e ! H2 þ 2OH ,

alkaline OER acidic HER alkaline HER

ð9:1aÞ ð9:1bÞ ð9:2aÞ ð9:2bÞ

The theoretical thermodynamic overpotential for water electrolysis is 1.23 V vs. reversible hydrogen electrode (RHE) [105]. The possibility to split water to form oxygen and hydrogen offers a promising route for sustainable storage of intermittently produced energy, for instance by sunlight or wind, in the form of chemical bonds. However, it is primarily hampered by the sluggish kinetics of the OER, because of which much higher cell voltages have to be applied in practice [106]. Moreover, the high anodic potentials may not exclusively oxidize the reactant, but also the catalyst material, so that the true active state may be different from its pristine structure. To reduce the high-energy input requirements, highly active electrocatalytic materials are necessary. The most active electrocatalysts involved in water electrolysis, namely RuO2 or IrO2 for the OER, and Pt for the HER, are scarce and expensive, which still hinders the economic viability of these energy conversion and storage devices. A massive body of literature has been generated in the endeavor to identify active and stable catalysts consisting of earthabundant elements, such as Ni, Co, and Fe [107]. The active catalyst states are usually the oxide or the oxyhydroxide forms of the used metal(s) [108]. Especially, composite materials of transition metals (i.e., Ni-Fe), exploiting synergistic metal–metal interactions have evoked considerable research efforts, since they offer a promising alternative to noble metal-based catalysts and comparable activities are already achievable [109]. Louie et al. investigated the structure and composition-related OER activity of Ni-Fe thin film

Operando Electrochemical Raman Spectroscopy

199

b)

12

c) 494 449

8 4 0 67

100% Fe

41

0.3

0.4

0.5

0.2 0.3 0.4 0.43 0.47 0.5 0.6

474

0.6

0.7

Potential vs. Hg/HgO (1 M) (V)

200

554

0% Fe 8 19 32 51 65 78 88

715

100

555

0% Fe 8 19 32 51

494

65 78 88

573 715

100

0.7 V

400 600 800 Raman shift (cm–1)

Fig. 9.8 (a) Cyclic voltammograms demonstrating the effect of Fe incorporation on the Ni(OH)2/NiOOH redox couple and OER activity. (b) OERS at low wavenumbers for 100% Ni films on a roughened Au substrate as a function of the applied potential (vs. Hg/HgO, 1 M KOH) in 0.1 M KOH. (c) OERS of Ni-Fe with different compositions

475

450

0.1

578

15 0% Fe

–4

d) 494

0V

Intensity (arb. unit)

Apparent current density (mA cm–2)

a)

through application of an ORR potential to the closely positioned tip (Fig. 9.9a). The potential at the Ni-Fe substrate was then stepped to higher anodic potentials until the OER started to proceed (Fig. 9.9b). As the oxygen was formed at the substrate and was diffusing toward the tip, a reductive current was generated at the tip due to ORR. A higher OER rate led to an increased ORR current and can be visualized as a function of the potential applied to the Ni-Fe substrate (Fig. 9.9c). In this particular study, the authors found that a Fe content of 15% seems to be optimal to enhance the OER activity. The OERS spectra of different composites (Fig. 9.9d (6% Fe) and Fig. 9.9e (22% Fe)) show the impact of the Fe incorporation on the formation of NiOOH and the resulting degree of disorder, which is increased with the introduction of a higher Fe fraction and indicated by peak ratios closer to 1. The overpotentials as determined by means of SECM are highlighted in red and show that a lower overpotential is linked to an increased Fe fraction. By coupling OERS with SECM as a further measure of catalyst activity, a more sophisticated and more sensitive distinction between different catalyst materials was introduced. Some other recent studies employ Raman spectroscopy in combination with OER catalysts to study NiFe layered double hydroxides [112], pure NiOx [113], cobalt oxides [114], manganese oxides [115], or to identify the active state of Cu-based materials during OER catalysis [97]. The latter reference is especially intriguing since Cu-based catalysts can exhibit a multitude of oxidation states under OER conditions that may enable electrocatalysis. It was found that different copper oxides unequivocally transform into CuO and Cu(OH)2 under working conditions, which both offer a similar OER activity. Furthermore, SERS has been used to identify the key intermediates on Au electrodes during gas

Intensity (arb. unit)

electrodes (Fig. 9.8a) as well as the influence of the Fe incorporation (up to 95% Fe) on the Ni oxidation states under working conditions in a very detailed study [110]. One of the key findings was that the Ni in Ni-Fe-based catalysts is present as NiOOH, whereas this phase is dramatically modulated by the presence of Fe (Fig. 9.8b–d). Changes in the NiOOH phase (and hence changes in the corresponding Raman peaks) correlate with an increased OER activity and they suggest that a reduction of the average Ni oxidation state (e.g., by Fe addition) increases the OER activity. The changes in the NiOOH phase (β/γ transition) are most likely an increase in the degree of disorder. Apparently, Ni-Fe films with an Fe-content between 15% and 41% exhibited the highest activities among the investigated samples. The spectral findings are furthermore underlined by voltammetric measurements, showing that the potential of the Ni(OH)2/NiOOH redox couple shifts to higher potentials until it coincides with the OER current for samples with Fe-contents equal or higher than 41%. A different approach for the investigation of NiFe-based OER catalysts was demonstrated by Steimecke et al., in which SECM was coupled to Raman spectroscopy in order to obtain electrochemical and spectral data with a high spatial resolution [111]. They employed a confocal Raman microscope and utilized transparent thin films of Ni and Ni-Fe to record the spectra in transmission mode. The setup allowed for the parallel observation of electrochemical processes by means of a positioned Pt microelectrode above the sample, while collecting Raman spectra at the same position from below. Ni-Fe films with a Fe content of up 30% have been studied. For determination of the OER activity, the SECM technique came into play. By using a sample generation/tip collection (SG/TC) mode, the evolved oxygen was detected

Intensity (arb. unit)

9

200

400 600 800 Raman shift (cm–1)

200

400 600 800 Raman shift (cm–1)

(Fe fraction given) in 0.1 M KOH at 0.2 V. (d) OERS of Ni-Fe with different compositions (Fe fraction given) in 0.1 M KOH at 0.6 V (OER proceeding). (Adapted with permission from reference [110]. Copyright (2013) American Chemical Society)

9

200

a)

W. Schuhmann et al.

b)

Pt UME

c) 0.000

O2

NiOOH ITO

jUM E (mA cm–2)

H2O

jWE 1 (mA cm–2)

0.6 0.4 0.2

–0.005 –0.010 –0.015

0.0

Confocal Raman microscopy

0.50

0.55 0.60 0.65 0.70 EWE 1 (V vs. Hg/HgO/1 M KOH)

d)

–0.020 0.50

0% Fe 6% Fe 7% Fe 15% Fe 22% Fe 30% Fe

0.60 0.65 0.55 EWE 1(V vs. Hg/HgO/1 M KOH)

0.70

e) Eonset = 0.63 V

Eonset = 0.62 V

I

400

500 600 700 a Q (cm–1)

800

)

300

400

500 600 700 a Q (cm–1)

800

0 vs .45 .H –0 g/ .66 H gO V( /1 'E M = KO 10 H mV

300

0 vs .45 .H –0 g/ .70 H gO V( /1 'E M = KO 10 H mV

)

I

Fig. 9.9 (a) Scheme of SG/TC SECM mode with the opportunity to record Raman spectra simultaneously with the electrochemical measurements. (b) Current densities at different potentials as measured at different Ni-Fe thin film electrodes in 0.1 M KOH and (c) corresponding Pt microelectrode response, biased to 0.6 V vs. Hg/HgO/1 M KOH to

detect evolved O2. OERS of Ni electrode containing (d) 6% Fe and (e) 22% Fe. Red highlights correspond to overpotentials determined by means of SECM. (Adapted with permission from ref. [111]. Copyright (2017) American Chemical Society)

evolution [116]. It was found that Au-OOH species as a precursor for O2 are forming [87]. As mentioned earlier, electrolytic water splitting is crucial for advanced energy conversion technologies. While before, the attention was directed to the anodic half reaction, the focus is now shifted to the cathodic reaction, i.e., the HER. An interesting class of noble metal-free HER catalysts are sulfides, such as CoS2 [117] or MoS2 NPs, discovered as especially suited for acidic HER by Nørskov [118] and Chorkendorff [119] et al. MoS2 catalysts were inspired by active centers of enzymes such as the FeMo cofactor in nitrogenases. Especially, amorphous MoSx materials have attracted attention since these easy to prepare catalysts offer a cheap and comparably active alternative to noble metals. Together with CoSe2 even catalytic hydrogen evolution activities close to commercial Pt/C are achievable [120]. Yet, these materials are poorly understood and their multielement nature complicates the analysis of the origin of their remarkable catalytic activity. The identification of active sites is challenging

especially when multiple sites are involved. For MoSx it is generally agreed that the structure consists of MoIV3 clusters with several types of S22 such as bridging, terminal, unsaturated, and apical. The amorphous type in its native state usually possesses a composition close to MoS3, whereas post-catalysis XPS analysis suggests that MoS2 is irreversibly formed during cathodic electrocatalysis [121, 122]. Further studies aimed at the elucidation of the nature of the extraordinary HER capability by identifying the active structural motifs by means of STM and it was found that the MoS2 edge sites are the origin of the catalytic activity [119]. The next stage of active site identification addresses the question, which atoms are involved in the catalytic process – Mo, S or both? In 2016, Yeo et al. employed OERS on MoSx materials to identify the active atoms involved in the HER and to characterize the elemental composition of the electrodeposits. They subjected cathodically and anodically electrodeposited films of MoSx to high cathodic overpotentials and a band appeared at around 2530 cm1 during gas evolution [98]. The band was first ascribed to the S-H stretch vibration based on

9

Operando Electrochemical Raman Spectroscopy

201

a)

320 415 445

earlier studies. A small peak was observed after applying 0.18 V vs. Hg/Hg2SO4/sat. K2SO4, which grew as a function of a further increased cathodic potential and disappeared upon lowering of the cathodic bias (Fig. 9.10a–d). These results were then verified by density functional theory (DFT) calculations and isotope substitution by performing similar experiments in DClO4 instead of HClO4. The high frequency of the observed band indicates that the H atom is bonded to a single S suggesting that it is not bridgecoordinated, implicating that Mo can be ruled out as active center. Thus, it could be demonstrated that the sulfur atoms are the catalytically active site of MoSx. Other OERS studies dealing with the HER include, for example, the identification of the active state of an amorphous cobalt sulfide catalyst to be CoS2 as demonstrated by Kornienko et al. [123], as well as further studies concerning HER on MoS2 [124]. Zhu et al. recently reported on the active state identification of CoSe2 during alkaline HER and

OER, concluding that Co0 is the active species in the case of HER, while CoOOH catalyzes the OER [125].

The Oxygen Reduction Reaction (ORR): OERS and SHINERS Electrocatalysis can be utilized not only for the generation of fuels but also for the sustainable consumption of fuels and generation of electric energy using fuel cells [126]. The electrochemical processes shown in a water electrolysis cell (Fig. 9.5b) are reversed in a fuel cell. The fuel hydrogen is oxidized at the anode while O2 is reduced at the cathode. The ORR is a multistep reaction involving the net transfer of four protons and four electrons (Eq. 9.3a and b), and it requires large overpotentials, which impedes attaining high energy efficiencies. The ORR takes place following two reaction mechanisms resulting either in the formation of hydroxide or hydrogen peroxide (Eq. 9.4c–d) [127, 128].

b) 320

MoSx-AE cathodic scan

415

630 934 –0.16 V –0.12 V –0.07 V

400

800

H2O

1200 2000 2500 3000 3500 Raman shift (cm–1) MoSx-CE anodic scan

320 415

415 630 *

934 0.40 V 0.28 V 0.18 V –0.07 V –0.12 V

400

–0.16 V –0.12 V –0.07 V 0.18 V 0.28 V 0.41 V

d)

MoSx-AE anodic scan

320

934

H2O

1200 2000 2500 3000 3500 Raman shift (cm–1)

800

9

2530

630

Intensity (A.U)

Intensity (A.U)

0.18 V 0.28 V 0.40 V

550

400

c)

Intensity (A.U)

Intensity (A.U)

2530

520

MoSx-CE cathodic scan

800

630

934 0.41 V 0.28 V 0.18 V

2530

–0.07 V

H2O

1200 2000 2500 3000 3500 Raman shift (cm–1)

Fig. 9.10 Operando Raman spectra of anodic (AE) and cathodic (CE) electrodeposits of MoSx in the low and high wavenumber region during CVs at 0.5 mV s1 in 1 M HClO4 and as deposited materials vary in their elemental stoichiometry (Mo:S 1:3.1 for AE and 1:1.8 for CE).

2530

H2O

–0.12 V

400

800

1200 2000 2500 3000 3500 Raman shift (cm–1)

(a/c) correspond to the anodic deposit, and (b/d) to the cathodic deposit, respectively. (Adapted with permission from ref. [98]. Copyright (2016) American Chemical Society)

202

W. Schuhmann et al. þ



O2 þ 4 H þ 4 e ! H2 O,

acidic 4 electron ORR ð9:3aÞ

þ



O2 þ 2 H þ 2 e ! H2 O2 ,

acidic 2 electron ORR ð9:4aÞ

þ



H2 O2 þ 2 H þ 2e ! 2H2 O,

acidic ½2 þ 2 electron ORR

ð9:4bÞ 



O2 þ 2 H2 O þ 4 e ! 4 OH ,

alkaline 4 electron ORR ð9:3bÞ







O2 þ H2 O þ 2 e ! OOH þ OH , alkaline 2 electron ORR

ð9:4cÞ 





OOH þ H2 O þ 2 e ! 3 OH , alkaline ½2 þ 2 electron ORR

ð9:4dÞ Oxygen electrocatalysis is of paramount importance not only for fuel cell applications, but also in modern chlor-alkali industry [129] and metal-air battery applications [130]. The state-of-the-art catalyst material for the ORR is Pt/C, exhibiting the highest catalytic activity [131]. Even though there is a large body of literature on the ORR, key questions regarding the origin of the high overpotentials and the identification of rate-determining steps remain unclear. That is because the ORR, as a multistep reaction, involves a variety of intermediates including OH22, O2, and radicals [132]. A short lifetime of the involved intermediates complicates their operando detection. It is hence important to identify the reaction intermediates and their adsorbed configuration on a given surface in order to tailor next generation catalysts adequately. To bridge that gap, Dong et al. conducted OERS measurements to investigate intermediately formed products during the ORR on several Pt single-crystal surfaces [133]. For that, the SHINERS method (see Sect. 9.3.3) was used and in particular Pt(111), Pt(100), and Pt(110) were investigated. Their corresponding polarization curves are displayed in Fig. 9.11a. In a first experiment utilizing Pt(111), the most active of these surfaces, the authors observed a pronounced peak growth at 732 cm1 upon application of potentials below 0.8 V vs. RHE. As the potential increased toward more negative values, the same peak grew in intensity (Fig. 9.11b). Based on deuterium substitution experiments, they deduced that this peak is correlated to a hydrogen atom and was thus assigned to the O-O stretching vibration of HO2*. Furthermore, the assignment was corroborated by DFT calculations, and it was suggested that this intermediate adsorbs on a bridge site (Fig. 9.11c). They were able to propose a mechanism for Pt(111), stating that the adsorbed

O2 transforms to HO2* after proton and electron transfer, and then dissociates to form OH* and O*, whereas the OH* reacts with hydrogen and forms H2O. According to their calculations, the activation energy for the HO2* dissociation was highest for all examined surfaces. In a different set of experiments, Pt(100) and Pt(110) were subjected to SHINERS and it was found that both behave similarly, yet different from Pt(111) (Fig. 9.11d, e). Peaks at around 1030 and 1080 cm1 appear as a consequence of increasing negative potentials starting at around 0.9 V. The former is assigned to the solvent-related ClO3 species, and the latter can be attributed to the Pt-OH bending mode of OH*. By means of DFT, it was possible to identify the presence of an OH* next to an adsorbed O* (Fig. 9.11f), which bended the H atom, so that the Pt-OH bending vibration increased from 875 to 1080 cm1. Furthermore, HO2*, being less stable at Pt(110) and Pt(100) as compared to Pt(111), is easily dissociated to Pt-O and Pt-OH making only the latter vibration visible. The diminished catalytic activity was caused by OH* blocking active sites on Pt surfaces. In addition, the ORR in alkaline solution was studied and a similar behavior of all the three Pt facets was observed. The experiments were conducted in a similar manner as those displayed in Fig. 9.12 but with the difference that a pH value of 10.3 was used. Considering pH changes during ORR and its effect on the reaction itself is crucial for its in-depth understanding, since, in alkaline conditions, the reaction consumes protons in the form of water and generates hydroxide ions, leading to an increased interfacial pH, which could influence the formation/consumption of intermediates as well as the electrode surface coverage. The authors found that a broad peak centered at around 1150 cm1 appeared upon application of 0.65 V (vs. RHE). The intensity of the peak rises as the potential increases to 0.35 V and decreases at more negative potentials (Fig. 9.12a–c). Isotopic substitution measurements conducted in D2O revealed that this species is not coupled to hydrogen the peak arises due to the O-O stretch in the O2 superoxide molecule bound in an on top configuration. Furthermore, it was shown that the CV shape does not alter upon decoration of the substrate with SHINs, thus demonstrating the capability of operando electrochemical SHINERS to allow for substrate generality without interfering with the electrochemical reaction. Further studies dealing with the ORR include the utilization of Au surfaces for SERS measurements and the identification of the superoxide intermediate [92]. Additionally, Radjenovic and Hardwick published a study concerning time-resolved SERS of the ORR in ionic liquids in the frame of nonaqueous lithium-oxygen cells [134].

9

Operando Electrochemical Raman Spectroscopy

203

a)

b) 0

1.0 V 0.9 V

–2

0.85 V –3

732

933

0.8 V

–4

0.7 V

–5

0.6 V

–6 –7

0.5 V 0.6

0.4

0.8

1.0

400

600

E versus RHE (V)

d)

ns (ClO4–)

1000

800

1200

Raman shift (cm–1)

e)

ns (ClO3) G (Pt-OH)

933

ns (ClO4–)

G (Pt-OH) 1.1 V 1.0 V

1,030 1.0 V

1,080

0.7 V

0.9 V 0.85 V 0.8 V 0.75 V 0.7 V

0.65 V

0.65 V

0.6 V

0.6 V

0.9 V 1,081

0.85 V 0.8 V

400

600

1000

800

1200

400

600

Raman shift (cm–1)

933

ns (ClO4–)

600

800

n (O2–)

1000

Raman shift (cm–1)

1000

800

9

1200

Raman shift (cm–1)

Fig. 9.11 (a) Polarization curve corresponding to the ORR proceeding at Pt(111), Pt(100), and Pt(110) surfaces recorded in O2-saturated 0.1 M HClO4 solution at a rotation rate of 1600 rpm and a scan rate of 50 mV s1. (b) Electrochemical SHINERS spectra during ORR at Pt (111) surface. (c) Illustration of HO2* adsorbed on the bridge site of a Pt (111) surface derived from DFT calculations. (d/e) Electrochemical

a)

f)

ns (ClO3) 933

1.1 V

1,030

400

Bridge site

1.1 V

Pt(111) Pt(100) Pt(110)

–1 j (mA cm–2)

c)

ns (ClO4–)

n (HO2*)

b)

1.05 V 0.95 V 0.85 V 0.75 V 0.65 V 1,151 0.55 V 0.45 V 0.35 V 0.25 V 0.15 V 1200 400

ns (ClO4–)

600

800

SHINERS spectra of Pt(100) and Pt(110) collected during ORR. (f) Side-view of OH* and O* adsorbed on a Pt(110) surface. Gray, red, and white colors in the illustrations represent Pt, O, and H atoms, respectively. The arrows in the spectra represent the potential scanning direction. The applied potential is referenced vs. RHE. (Reprinted by permission from Springer Nature, Nature Energy from ref. [133])

933

n (O2–)

1.05 V 0.95 V 0.85 V 0.75 V 0.65 V 1,149 0.55 V 0.45 V 0.35 V 0.25 V 0.15 V 400 1000 1200

Raman shift (cm–1)

Fig. 9.12 Electrochemical SHINERS study of ORR on (a) Pt(110), (b) Pt(111), and (c) Pt(100). Experiments were conducted in O2-saturated 0.1 M NaClO4 solution (pH 10.3). Arrows indicate the order of the

c)

ns (ClO4–)

933

n (O2–) 1.05 V 0.95 V 0.85 V 0.75 V 0.65 V 0.55 V 1,150 0.45 V

0.35 V 0.25 V 0.15 V 600

800

1000

1200

Raman shift (cm–1)

applied potential steps (vs. RHE). (Reprinted by permission from Springer Nature, Nature Energy from ref. [133])

204

W. Schuhmann et al.

The Carbon Dioxide Reduction Reaction (CO2RR): OERS and SERS A new frontier in modern electrochemistry with a recent spike of attention is the CO2RR. During the 1980s, CO2 was already exploited as a potential abundant carbon source by Hori and several others [135, 136]. CO2RR enables the production of commodity chemicals utilizing the greenhouse gas under consumption of electrical energy, which in a sustainable scenario should be supplied by a renewable energy source. For example, it could be used to store intermittently produced energy while providing a means to add economic value to a chemical compound that negatively affects the climate. The CO2RR is a multistep and multi-electron transfer process, involving up to 18 electrons and leading to 16 different products [137]. The products depend on both the applied potential and catalyst material. The multitude of pathways coupled to these processes poses a great challenge to the elucidation of the underlying reaction mechanisms. The highest variety of products and, especially, C2 products such as ethylene, are usually observed when Cu is utilized as catalyst material [138]. Other metals such as Ag and Au produce predominantly CO and H2 at high cathodic overpotentials [48, 139–141]. By reduction of CO2 to CO, while producing H2 in parallel, a direct formation of syngas may be conceivable, which can be utilized in relevant industrial processes including the Fischer-Tropsch synthesis to produce higher hydrocarbons. Since these metals are extraordinarily suited for SERS applications (see Sect. 9.3.1) the detection of even extremely small concentrations of intermediates and products is possible by means of OERS. In particular, Ag is a promising material for the conversion of CO2, since it is comparably cheap and offers a high selectivity for CO, among HCOOH as a minor byproduct.

b)

1000 500

40,000 20,000 0

0 0 5 10 15 20 25 Raman position (mm)

)

(mm

60,000

Count

Polymer size 200 nm 300 nm 500 nm

8 7 6 5 4 3 2 1

ion

1500

400

osit

Maximum peak intensity of νs(N-O)

d)

200

an p

2000

c)

Ram

Raman counts (mW–1 s–1)

2500

Raman counts (mW–1 s–1)

a)

To enable the investigation of reactions with high Raman signal intensities, sophisticated nanostructures are necessary, and their development and identification of highly active structures can be tedious. An optimal morphology of Ag nanovoid structures to maximize the signal enhancement with subsequent spectroscopic CO2RR investigation was shown recently [48]. In order to identify the optimal void morphology, differently sized nanospheres (200, 300, and 500 nm) were deposited onto a Ag-coated Si wafer and a gradient of Ag was electrodeposited between the nanosphere templates using a bipolar electrochemical setup, thus generating a substrate with different void opening sizes and wall thicknesses. These were then subjected to line scans utilizing the Raman spectrometer (20 objective) and lateral displacement increments of 100 μm detecting a Raman probe molecule (4-nitrothiophenol, 4-NTP) to differentiate which template displayed the highest signal enhancement. Out of the three measured samples, the substrate that was decorated with 300 nm nanospheres exhibited the highest signal intensity, as shown in Fig. 9.13a, in which the detected Raman signal is plotted as a function of the position on the substrate. The 300 nm sample was then scanned with a 60 objective, which offers a smaller laser spot and a smaller lateral displacement increment (10 μm), to increase the resolution, and thus, to determine the exact position of the highest signal enhancement (Fig. 9.13b). Using scanning electron microscopy (SEM), the exact void morphology at that particular spot was identified (Fig. 9.13c, d). Afterward, a sample for Raman spectroscopic investigation was fabricated using the determined optimal void morphology, and it was used as electrode for operando SERS investigation of the CO2RR aiming at the identification of the formed products. Potential steps from 0.1 V to 1.3 V vs. Ag/AgCl/3 M KCl were applied for 120 s (Fig. 9.14a) and

1000 1200 1400 1600 Wavenumber (cm–1)

Fig. 9.13 (a) Change in intensity of the N-O stretching band at 1343 cm1 across a nanovoid-modified Ag gradient. (b) Raman line scans across nanostructured (300 nm template) sample with 10 μm lateral displacement for a total length of 8 cm. (c) SEM image nanovoid structures at the point of highest Raman signal enhancement as

2 μm 0

150 200 250 Void diameter (nm)

determined in (b). (d) Void diameter distribution of the substrate shown in (b) and (c) revealing an average void opening size of 190 nm (N ¼ 4000, group size 5 nm). (Adapted with permission from ref. [48]. Copyright (2018) American Chemical Society)

2724 2826 2860 2929

2090

1308 14101440 1536 1612

487

b) 709

a)

OCP after –1.3 V –1.25 V

0.0

–1.2 V

–0.2

–1.15 V

750 Raman counts (mW–1 s–1)

Fig. 9.14 (a) Potential step chronoamperometry on a nanovoid Ag surface during Raman experiments. Each potential was held for 120 s. (b) Raman spectra recorded (60 s after potential application). (c) Magnification of the potentialdependent peak intensity of the CO-associated signal. The spectra were recorded during a cathodic potential sweep from 0.8 V to 1.2 V vs. Ag/AgCl/3 M KCl. (d) Potential-dependent peak growth in the C-H region during cathodic potential steps. (Adapted with permission from ref. [48]. Copyright (2018) American Chemical Society)

205

934 1003 1076

Operando Electrochemical Raman Spectroscopy

i (mA)

9

–0.4

–0.6

–1.1 V –1.05 V –1 V –0.95 V –0.9 V –0.85 V –0.8 V –0.75 V –0.7 V –0.6 V –0.5 V –0.4 V –0.3 V –0.2 V –0.1 V OCP before

–0.8

–1.0 –1.0

–0.5

0.0

1000

c)

2074

2000

3000

Wavenumber (cm–1)

E vs. Ag/AgCl/3M KCl (V)

d)

2097

2724

2827 2860 2929

9

OCP OCP

–1.2 V

–1.2 V –1.15 V –1.1 V –1.15 V –1.1 V –1.05 V

2000

2050

2100

2150

Wavenumber (cm–1)

Raman spectra were recorded after 60 s of each potential bias (Fig. 9.14b). From the overall spectrum (Fig. 9.14b) it can be deduced that a peak growth predominantly occurs centered at 2100 cm1 (Fig. 9.14c) and at around 2800 cm1 (Fig. 9.14d). It was concluded that the increase in the peak intensity in the region around 2100 cm1 corresponds to the carbon oxygen vibration of on-top bound CO. The observed frequency red shift may indicate a conformational change to bridge bound CO, which would arise at 1950 cm1. In

–1.05 V

–1 V

–1 V –0.95 V

–0.9 V –0.8 V OCP before

–0.9 V

2200

–0.8 V OCP before

2700

2800 2900 Wavenumber (cm–1)

3000

contrast, the peaks emerging and growing at higher wavenumbers have been reported to arise due to the formation of methylene-containing species adsorbed on the surface. It was not possible to gain spectral evidence for the emergence of formate intermediate species, which reportedly is a product of the CO2RR on silver. The preparation and optimization of SERS substrates for a certain application and its utilization for operando characterization of the CO2RR was demonstrated.

206

W. Schuhmann et al.

SERS-active catalysts can, once identified, be used to identify possible reaction pathways especially in combination with modeling. Ultimately, by tuning catalysts in a way that certain intermediate species and reaction paths are favored, the selectivity for the final product could be modified. Hence, a crucial research task is monitoring intermediate formation and studying its impact on the final products. In the case of the CO2RR, it is generally accepted that the formation of CO proceeds via a *COOH intermediate and the formation of HCOO has a bidentate, O-bound *OCHO precursor [142]. However, the activation barrier (on Ag(110) surfaces) for the bidentate intermediate is considerably lower than that for the *COOH precursor, yet a significantly higher selectivity for CO is observed as compared to that for HCOO [93]. The role of the bidentate intermediate was investigated by Wilson et al. in a study comprising DFT coupled operando SERS measurements [93] aiming at the elucidation of mechanistic differences in the formation of both intermediates and their impact on the H2 formation, shedding light on the role of parameters that could influence the CO2RR selectivity on a Ag(110) model by means of DFT calculations. These results were confirmed by spectroscopy on polycrystalline Ag surfaces, which in fact showed preliminary evidence for the formation of the bidentate intermediate. They hypothesized that formation of *OCHO on the catalyst surface promotes CO formation and inhibits HCOO production. They ascribed the lateral adsorbate interactions of *OCHO to be a crucial factor that influences the coverage of *H, which in turn inhibits HCOO and H2 formation and thus induces a higher selectivity towards CO. In another interesting study involving SERS conducted by Gewirth et al. the influence of an additive (3,5-diamino1,2,4-triazole, DAT) on the CO2RR to CO was investigated. It was observed that the overpotential is slightly lower and that CO preferably adsorbs on sites with less surface coordination in the presence of the AgDAT complex. That means that adsorption of CO is weaker with respect to a bare Ag surface, it dissociates faster and impedes the reaction to continue, which also may account for an increase in CO selectivity. They showed using Raman spectroscopy that hydrocarbon formation is suppressed and occurs at 0.2 V more negative potentials with respect to the metallic Ag surface [143].

9.5

Experiences from a Practical Point of View

One of the shortcomings in SERS experiments is the extreme sensitivity of the EF to the nanostructure of the electrode material, which can lead to very heterogeneous results associated with slightly different yet supposedly similar SERSactive surfaces. That means, that different signal intensities

can be recorded at different positions of the substrate, which complicates reproducibility and comparability. That can be overcome, for instance, by recording spectra at different positions of the substrate followed by averaging. Another mathematical operation could involve the multiplication of spectra to normalize them to a certain constant reference peak, which is not influenced by the potential. Alternatively, the laser (or the sample) can be moved around until the spot offering the highest signal intensity is identified. Furthermore, surfaces may change during the course of the experiment leading to spectral intensities, which differ as a function of the time, and may be induced by thermal, potential, or chemical processes. This phenomenon, commonly referred to as “irreversible loss,” is known for more than a decade and could be due to different effects such as the metastable nature of atomic roughness, which could be a signature for the chemical enhancement mechanism, or to the morphological instability of roughness on the nm scale, which would be evidence for the EM mechanism [144]. For instance, a temperature rise increases adatom mobility, which could allow for the migration of atoms into the metal lattices, disrupting the delicate surface structure. Similar effects hold true for morphological changes, for example, induced by oxidationreduction cycles (ORC) of Ag electrodes, where signal intensity losses of up to 75% were reported upon application of ORC. As a consequence of the potential sweeps, the morphology of the SERS substrate alters, for instance the particle diameter may change as probed by means of SEM [145]. Similar signal loss effects have been observed in a TERS environment triggered by thermal effects [146]. Hence, the possible destructive impact of various sources should be considered beforehand. In particular, when working with Ag surfaces, oxidation or tarnishing processes may occur. It is advisable to introduce a conditioning step for the electrodes prior to any highly sensitive spectroscopic measurement to ensure similar surfaces for each measurement. It is important to keep in mind that the substrates should be stored under Ar and possibly under exclusion of light to hamper oxidative processes. A recent work performed by Dau et al. [100] highlights the problems that can be encountered during data treatment of Raman spectra, such as baseline correction or background subtraction. They report that a strong SERS background is often observed, which depends on the electrochemical potential complicating the analysis of the SERS bands. They found that the background intensities decrease as a function of an increased negative potential on Cu foam electrodes utilized in the CO2RR. Furthermore, the potential dependence is also dictated by the history of the electrode exposure to cathodic potentials. The broad background was separated from the actual SERS peaks by subtraction of polynomial functions and then normalizing the Raman intensities to the background intensity. The normalization was carried out by averaging of the

9

Operando Electrochemical Raman Spectroscopy

background curve, whereas the integrated background intensity was then used to divide the Raman spectra. They found that a background subtraction and a subsequent normalization reveals previously not observable spectral data, and it was suggested that in order to adequately interpret SERS data these steps should be undertaken before evaluating the signals. The problems associated with a high intensity SERS background were already acknowledged and tackled in a study authored by Bin Ren et al. [147], in which a general method was developed to correct for spectral deviations induced by the SERS background. They pointed out that the local field induced by excitation of LSPRs dominates the overall enhancement of SERS thus distorting the spectral features that contain the desired information. The distortion of the relative intensity in the recorded SERS spectra is commonly called plasmonic spectral shaping effect (PSSE). They claimed that photoluminescence (PL) of metal nanostructures contributes to the background in SERS spectra, which is usually either neglected or subtracted to clean the acquired spectrum and they presented a method to subtract the SERS background in order to retrieve the molecular fingerprint that reflects the chemical interaction between molecule and substrate (Au nanorods therein). It was proposed to measure the PL of the bare substrate in question, since it is an intrinsic response of the nanostructure and the measured PL closely resembles the SERS background in the presence of analyte molecules. The conclusion was drawn that the PL is a dominating factor dictating the SERS background, and on that basis a methodology is proposed to filter out the contribution of the PL exhibited by the substrate. Briefly, the background-subtracted SERS spectra are divided by the fitted backgrounds and multiplied by the substrate PL. For further information and an in-depth explanation of the procedure, see reference [147].

9.6

Conclusion

Raman spectroscopy has been developed significantly in the past years, enabling a wide range of research applications and spanning its influence through various areas of science and industry. We focused on Raman spectroscopy in combination with electrochemistry, or more specifically, electrocatalysis as a tool for investigating phenomena occurring at electrolyte/electrode interfaces operando during electrochemical reactions. A general description of the basic principles and instrumentation of Raman spectroscopy was provided, as well as an introduction to next generation surface-sensitive Raman spectroscopy techniques, including SERS, SHINERS, and TERS. We discussed the main challenges related to conducting spectroscopic analysis under reaction conditions (operando conditions) and different technical strategies for overcoming

207

them. Three relevant topics in electrocatalysis were chosen as study cases: water electrolysis, reduction of oxygen for fuelcell applications, and reduction of CO2 for converting it to valuable chemicals. We showed that operando electrochemical Raman spectroscopy has been effectively utilized for aiding the identification of active sites, reaction products, and intermediates, material stability, among other aspects of electrocatalytic processes, undoubtedly rendering this technique a powerful analytical tool with the ability to address state-of-the-art scientific questions while consuming both a reasonable amount of time and money.

References 1. Smekal, A.: Zur Quantentheorie der Dispersion. Naturwissenschaften. 11, 873–875 (1923) 2. Raman, C.V., Krishnan, K.S.: Polarisation of scattered lightquanta. Nature. 122, 169 (1928) 3. Landsberg, G., Mandelstam, L.: Eine neue Erscheinung bei der Lichtzerstreuung in Krystallen. Naturwissenschaften. 16, 557–558 (1928) 4. Komai, T., Kawamura, T., Kang, S., Nagashima, K., Yamamoto, Y.: In situ observation of gas hydrate behaviour under high pressure by Raman spectroscopy. J. Phys. Condens. Matter. 14, 11395–11400 (2002) 5. Welsh, H.L., Stansbury, E.J., Romanko, J., Feldman, T.: Raman spectroscopy of gases. J. Opt. Soc. Am. 45, 338 (1955) 6. Yu, Y., Lin, K., Zhou, X., Wang, H., Liu, S., Ma, X.: New CH stretching vibrational spectral features in the Raman spectra of gaseous and liquid ethanol†. J. Phys. Chem. C. 111, 8971–8978 (2007) 7. Iwata, K., Okajima, H., Saha, S., Hamaguchi, H.-O.: Local structure formation in alkyl-imidazolium-based ionic liquids as revealed by linear and nonlinear Raman spectroscopy. Acc. Chem. Res. 40, 1174–1181 (2007) 8. Boyaci, I.H., Genis, H.E., Guven, B., Tamer, U., Alper, N.: A novel method for quantification of ethanol and methanol in distilled alcoholic beverages using Raman spectroscopy. J. Raman Spectrosc. 43, 1171–1176 (2012) 9. Carey, D.M., Korenowski, G.M.: Measurement of the Raman spectrum of liquid water. J. Chem. Phys. 108, 2669–2675 (1998) 10. Ferrari, A.C., Meyer, J.C., Scardaci, V., Casiraghi, C., Lazzeri, M., Mauri, F., Piscanec, S., Jiang, D., Novoselov, K.S., Roth, S., Geim, A.K.: Raman spectrum of graphene and graphene layers. Phys. Rev. Lett. 97, 187401 (2006) 11. Mammone, J.F., Sharma, S.K., Nicol, M.: Raman spectra of methanol and ethanol at pressures up to 100 kbar. J. Phys. Chem. 84, 3130–3134 (1980) 12. Frost, R.L., Henry, D.A., Weier, M.L., Martens, W.: Raman spectroscopy of three polymorphs of BiVO4: Clinobisvanite, dreyerite and pucherite, with comparisons to (VO4)3-bearing minerals: namibite, pottsite and schumacherite. J. Raman Spectrosc. 37, 722–732 (2006) 13. Griffith, W.P.: Raman spectroscopy of minerals. Nature. 224, 264–266 (1969) 14. Rygula, A., Majzner, K., Marzec, K.M., Kaczor, A., Pilarczyk, M., Baranska, M.: Raman spectroscopy of proteins: a review. J. Raman Spectrosc. 44, 1061–1076 (2013) 15. Butler, H.J., Ashton, L., Bird, B., Cinque, G., Curtis, K., Dorney, J., Esmonde-White, K., Fullwood, N.J., Gardner, B., MartinHirsch, P.L., Walsh, M.J., McAinsh, M.R., Stone, N., Martin,

9

208 F.L.: Using Raman spectroscopy to characterize biological materials. Nat. Protoc. 11, 664–687 (2016) 16. Puppels, G.J., de Mul, F.F., Otto, C., Greve, J., Robert-Nicoud, M., Arndt-Jovin, D.J., Jovin, T.M.: Studying single living cells and chromosomes by confocal Raman microspectroscopy. Nature. 347, 301–303 (1990) 17. Das, R.S., Agrawal, Y.K.: Raman spectroscopy: recent advancements, techniques and applications. Vib. Spectrosc. 57, 163–176 (2011) 18. Smith, E., Dent, G.: Modern Raman Spectroscopy – A Practical Approach. John Wiley & Sons, Ltd, Chichester (2004) 19. Toporski, J., Dieing, T., Hollricher, O.: Confocal Raman Microscopy. Springer International Publishing, Cham (2018) 20. Kellner, R., et al. (eds.): Analytical Chemistry: The Approved Text to the FECS Curriculum Analytical Chemistry. Wiley-VCH, Weinheim (1998) 21. McCreery, R.L.: Raman Spectroscopy for Chemical Analysis. Wiley, Hoboken (2000) 22. Kneipp, K., Kneipp, H., Itzkan, I., Dasari, R.R., Feld, M.S.: Ultrasensitive chemical analysis by Raman spectroscopy. Chem. Rev. 99, 2957–2976 (1999) 23. Fleischmann, M., Hendra, P.J., McQuillan, A.J.: Raman spectra of pyridine adsorbed at a silver electrode. Chem. Phys. Lett. 26, 163–166 (1974) 24. McQuillan, A., Hendra, P.J., Fleischmann, M.: Raman spectroscopic investigation of silver electrodes. J. Electroanal. Chem. 65, 933–944 (1975) 25. Jeanmaire, D.L., van Duyne, R.P.: Surface Raman Spectroelectrochemistry. J. Electroanal. Chem. Interfacial Electrochem. 84, 1–20 (1977) 26. Albrecht, M.G., Creighton, J.A.: Anomalously intense Raman spectra of pyridine at a silver electrode. J. Am. Chem. Soc. 99, 5215–5217 (1977) 27. Kneipp, K., Wang, Y., Kneipp, H., Perelman, L.T., Itzkan, I., Dasari, R.R., Feld, M.S.: Single molecule detection using surface-enhanced Raman scattering (SERS). Phys. Rev. Lett. 78, 1667–1670 (1997) 28. Nie, S., Emory, R.: Probing single molecules and single nanoparticles by surface-enhanced Raman scattering. Science. 275, 1102–1106 (1997) 29. Schlögl, R.: Heterogeneous catalysis. Angew. Chem. Int. Ed. 54, 3465–3520 (2015) 30. Wang, Y., Becker, M., Wang, L., Liu, J., Scholz, R., Peng, J., Gösele, U., Christiansen, S., Kim, D.H., Steinhart, M.: Nanostructured gold films for SERS by block copolymer-templated galvanic displacement reactions. Nano Lett. 9, 2384–2389 (2009) 31. Moskovits, M.: Surface-enhanced Raman spectroscopy: a brief retrospective. J. Raman Spectrosc. 36, 485–496 (2005) 32. Schlücker, S.: Surface-enhanced Raman spectroscopy: concepts and chemical applications. Angew. Chem. Int. Ed. 53, 4756–4795 (2014) 33. Haynes, C.L., McFarland, A.D., Van Duyne, R.P., Haynes, C.L., McFarland, A.D., van Duyne, R.P.: Surface-enhanced Raman spectroscopy. Anal. Chem. 77, 338A–346A (2005) 34. Zhang, H., Wei, J., Zhang, X.-G., Zhang, Y.-J., Radjenovica, P.M., Wu, D.-Y., Pan, F., Tian, Z.-Q., Li, J.-F.: Plasmon-induced interfacial hot-electron transfer directly probed by Raman spectroscopy. Chem. 6, 689–702 (2020) 35. Liu, R., He, Z., Sun, J., Liu, J., Jiang, G.: Tracking the fate of surface plasmon resonance-generated hot electrons by in situ SERS surveying of catalyzed reaction. Small. 12, 6378–6387 (2016) 36. Xu, H., Aizpurua, J., Käll, M., Apell, P.: Electromagnetic contributions to single-molecule sensitivity in surface-enhanced Raman scattering. Phys. Rev. E. 62, 4318–4324 (2000) 37. Campion, A., Kambhampati, P.: Surface-enhanced Raman scattering. Chem. Soc. Rev. 27, 241 (1998)

W. Schuhmann et al. 38. Kneipp, K., Moskovits, M., Kneipp, H.: Surface-Enhanced Raman Scattering. Springer, Berlin Heidelberg (2006) 39. Haynes, C.L., Yonzon, C.R., Zhang, X., van Duyne, R.P.: Surfaceenhanced Raman sensors: early history and the development of sensors for quantitative biowarfare agent and glucose detection. J. Raman Spectrosc. 36, 471–484 (2005) 40. Gersten, J., Nitzan, A.: Electromagnetic theory of enhanced Raman scattering by molecules adsorbed on rough surfaces. J. Phys. Chem. 73, 3023–3037 (1980) 41. Gersten, J., Nitzan, A.: Spectroscopic properties of molecules interacting with small dielectric particles. J. Phys. Chem. 75, 1139–1152 (1981) 42. Gersten, J., Rayleigh, I.: Mie, and Raman scattering by molecules adsorbed on rough surfaces. J. Phys. Chem. 72, 5780–5781 (1980) 43. Wu, D.-Y., Duan, S., Ren, B., Tian, Z.-Q.: Density functional theory study of surface-enhanced Raman scattering spectra of pyridine adsorbed on noble and transition metal surfaces. J. Raman Spectrosc. 36, 533–540 (2005) 44. Otto, A., Bruckbauer, A., Chen, Y.X.: On the chloride activation in SERS and single molecule SERS. J. Mol. Struct. 661-662, 501–514 (2003) 45. Schlücker, S. (ed.): Surface Enhanced Raman Spectroscopy: Analytical, Biophysical and Life Science Applications. Weinheim, Wiley-VCH (2011) 46. Natan, M.J.: Concluding remarks: surface enhanced Raman scattering. Faraday Discuss. 132, 321 (2006) 47. Le Ru, E.C., Blackie, E., Meyer, M., Etchegoin, P.G.: Surface enhanced Raman scattering enhancement factors: a comprehensive study. J. Phys. Chem. C. 111, 13794–13803 (2007) 48. Öhl, D., Kayran, Y.U., Junqueira, J.R.C., Eßmann, V., Bobrowski, T., Schuhmann, W.: Optimized Ag nanovoid structures for probing electrocatalytic carbon dioxide reduction using operando surfaceenhanced Raman spectroscopy. Langmuir. 34, 12293–12301 (2018) 49. Kneipp, K., Kneipp, H., Kneipp, J.: Surface-enhanced Raman scattering in local optical fields of silver and gold nanoaggregates-from single-molecule Raman spectroscopy to ultrasensitive probing in live cells. Acc. Chem. Res. 39, 443–450 (2006) 50. Creighton, J.A., Blatchford, C.G., Albrecht, M.G.: Plasma resonance enhancement of Raman scattering by pyridine adsorbed on silver or gold sol particles of size comparable to the excitation wavelength. J. Chem. Soc., Faraday Trans. 2. 75, 790–798 (1979) 51. Wustholz, K.L., Henry, A.-I., McMahon, J.M., Freeman, R.G., Valley, N., Piotti, M.E., Natan, M.J., Schatz, G.C., van Duyne, R.P.: Structure-activity relationships in gold nanoparticle dimers and trimers for surface-enhanced Raman spectroscopy. J. Am. Chem. Soc. 132, 10903–10910 (2010) 52. Kayran, Y.U., Eßmann, V., Grützke, S., Schuhmann, W.: Selection of highly SERS-active nanostructures from a size gradient of Au nanovoids on a single bipolar electrode. ChemElectroChem. 3, 399–403 (2016) 53. Lang, X., Qiu, T., Yin, Y., Kong, F., Si, L., Hao, Q., Chu, P.K.: Silver nanovoid arrays for surface-enhanced Raman scattering. Langmuir. 28, 8799–8803 (2012) 54. Cintra, S., Abdelsalam, M.E., Bartlett, P.N., Baumberg, J.J., Kelf, T.A., Sugawara, Y., Russell, A.E.: Sculpted substrates for SERS. Faraday Discuss. 132, 191–199 (2006) 55. Yüksel, H., Özbay, A., Solmaz, R., Kahraman, M.: Fabrication and characterization of three-dimensional silver nanodomes: application for alkaline water electrolysis. Int. J. Hydrog. Energy. 42, 2476–2484 (2017) 56. Fang, Y., Seong, N.-H., Dlott, D.D.: Measurement of the distribution of site enhancements in surface-enhanced Raman scattering. Science. 321, 385–388 (2008) 57. Lee, S.J., Morrill, A.R., Moskovits, M.: Hot spots in silver nanowire bundles for surface-enhanced Raman spectroscopy. J. Am. Chem. Soc. 128, 2200–2201 (2006)

9

Operando Electrochemical Raman Spectroscopy

58. Yu, W.W., White, I.M.: Inkjet printed surface enhanced Raman spectroscopy array on cellulose paper. Anal. Chem. 82, 9626–9630 (2010) 59. Miccichè, C., Arrabito, G., Amato, F., Buscarino, G., Agnello, S., Pignataro, B.: Inkjet printing Ag nanoparticles for SERS hot spots. Anal. Methods. 10, 3215–3223 (2018) 60. Tao, A., Kim, F., Hess, C., Goldberger, J., He, R., Sun, Y., Xia, Y., Yang, P.: LangmuirBlodgett silver nanowire monolayers for molecular sensing using surface-enhanced Raman spectroscopy. Nano Lett. 3, 1229–1233 (2003) 61. Haynes, C.L., van Duyne, R.P.: Nanosphere lithography: a versatile nanofabrication tool for studies of size-dependent nanoparticle optics. J. Phys. Chem. B. 105, 5599–5611 (2001) 62. Alvarez-Puebla, R., Cui, B., Bravo-Vasquez, J.-P., Veres, T., Fenniri, H.: Nanoimprinted SERS-active substrates with tunable surface plasmon resonances. J. Phys. Chem. C. 111, 6720–6723 (2007) 63. Freeman, R.G., Grabar, K.C., Allison, K.J., Bright, R.M., Davis, J.A., Guthrie, A.P., Hommer, M.B., Jackson, M.A., Smith, P.C., Walter, D.G., Natan, M.J.: Self-assembled metal colloid monolayers: an approach to SERS substrates. Science. 267, 1629–1632 (1995) 64. Grabar, K.C., Freeman, R.G., Hommer, M.B., Natan, M.J.: Preparation and characterization of Au colloid monolayers. Anal. Chem. 67, 735–743 (1995) 65. Tian, Z.-Q., Ren, B., Wu, D.-Y.: Surface-enhanced Raman scattering: from Noble to transition metals and from rough surfaces to ordered nanostructures. J. Phys. Chem. B. 106, 9463–9483 (2002) 66. Abdelsalam, M.E., Mahajan, S., Bartlett, P.N., Baumberg, J.J., Russell, A.E.: SERS at structured palladium and platinum surfaces. J. Am. Chem. Soc. 129, 7399–7406 (2007) 67. Stiles, P.L., Dieringer, J.A., Shah, N.C., van Duyne, R.P.: Surfaceenhanced Raman spectroscopy. Annu. Rev. Anal. Chem. 1, 601–626 (2008) 68. Stöckle, R.M., Suh, Y.D., Deckert, V., Zenobi, R.: Nanoscale chemical analysis by tip-enhanced Raman spectroscopy. Chem. Phys. Lett. 318, 131–136 (2000) 69. Pettinger, B., Ren, B., Picardi, G., Schuster, R., Ertl, G.: Nanoscale probing of adsorbed species by tip-enhanced Raman spectroscopy. Phys. Rev. Lett. 92, 96101 (2004) 70. Hayazawa, N., Inouye, Y., Sekkat, Z., Kawata, S.: Metallized tip amplification of near-field Raman scattering. Opt. Commun. 183, 333–336 (2000) 71. Tian, Z.-Q., Ren, B., Li, J.-F., Yang, Z.-L.: Expanding generality of surface-enhanced Raman spectroscopy with borrowing SERS activity strategy. Chem. Commun. 121, 3514 (2007) 72. van Duyne, R.P., Haushalter, J.P.: Surface-enhanced Raman spectroscopy of adsorbates on semiconductor electrode surfaces: Tris (bipyridine)ruthenium(II) adsorbed on silver-modified n-gallium arsenide(100). J. Phys. Chem. 87, 2999–3003 (1983) 73. Fleischmann, M., Tian, Z.Q., Li, L.J.: Raman spectroscopy of adsorbates on thin film electrodes deposited on silver substrates. J. Electroanal. Chem. Interfacial Electrochem. 217, 397–410 (1987) 74. Mengoli, G., Musiani, M.M., Fleischman, M., Mao, B., Tian, Z.Q.: Enhanced Raman scattering from iron electrodes. Electrochim. Acta. 32, 1239–1245 (1987) 75. Park, S., Yang, P., Corredor, P., Weaver, M.J.: Transition metalcoated nanoparticle films: vibrational characterization with surfaceenhanced Raman scattering. J. Am. Chem. Soc. 124, 2428–2429 (2002) 76. Mrozek, M.F., Xie, Y., Weaver, M.J.: Surface-enhanced Raman scattering on uniform platinum-group overlayers: preparation by redox replacement of Underpotential-deposited metals on gold. Anal. Chem. 73, 5953–5960 (2001) 77. Zou, S., Weaver, M.J.: Surface-enhanced Raman scattering on uniform transition-metal films: toward a versatile adsorbate

209 vibrational strategy for solid-nonvacuum interfaces? Anal. Chem. 70, 2387–2395 (1998) 78. Yang, C., Zhang, C., Huo, Y., Jiang, S., Qiu, H., Xu, Y., Li, X., Man, B.: Shell-isolated graphene@Cu nanoparticles on graphene@Cu substrates for the application in SERS. Carbon. 98, 526–533 (2016) 79. Li, J.F., Huang, Y.F., Ding, Y., Yang, Z.L., Li, S.B., Zhou, X.S., Fan, F.R., Zhang, W., Zhou, Z.Y., Wu, D.Y., Ren, B., Wang, Z.L., Tian, Z.Q.: Shell-isolated nanoparticle-enhanced Raman spectroscopy. Nature. 464, 392–395 (2010) 80. Huang, Y.-F., Kooyman, P.J., Koper, M.T.M.: Intermediate stages of electrochemical oxidation of single-crystalline platinum revealed by in situ Raman spectroscopy. Nat. Commun. 7, 12440 (2016) 81. Hartman, T., Wondergem, C.S., Weckhuysen, B.M.: Practical guidelines for Shell-isolated nanoparticle-enhanced Raman spectroscopy of heterogeneous catalysts. ChemPhysChem. 19, 2461–2467 (2018) 82. Wang, Y.-H., Wei, J., Radjenovic, P., Tian, Z., Li, J.-F.: In situ analysis of surface catalytic reactions using SHINERS. Anal. Chem. 91(3), 1675–1685 (2019) 83. Lin, X.-D., Uzayisenga, V., Li, J.-F., Fang, P.-P., Wu, D.-Y., Ren, B., Tian, Z.-Q.: Synthesis of ultrathin and compact Au@MnO2 nanoparticles for shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS). J. Raman Spectrosc. 43, 40–45 (2012) 84. Bockris, J.O.’.M., Reddy, A.K.N., Gamboa-Aldeco, M.E.: Modern Electrochemistry, 2nd edn. Plenum Press, New York (1998–2000) 85. Chen, D., Ding, D., Li, X., Waller, G.H., Xiong, X., El-Sayed, M.A., Liu, M.: Probing the charge storage mechanism of a pseudocapacitive MnO2 electrode using in operando Raman spectroscopy. Chem. Mater. 27, 6608–6619 (2015) 86. Hannauer, J., Scheers, J., Fullenwarth, J., Fraisse, B., Stievano, L., Johansson, P.: The quest for polysulfides in lithium-sulfur battery electrolytes: An operando confocal Raman spectroscopy study. ChemPhysChem. 16, 2755–2759 (2015) 87. Yeo, B.S., Klaus, S.L., Ross, P.N., Mathies, R.A., Bell, A.T.: Identification of hydroperoxy species as reaction intermediates in the electrochemical evolution of oxygen on gold. ChemPhysChem. 11, 1854–1857 (2010) 88. Kornienko, N., Heidary, N., Cibin, G., Reisner, E.: Catalysis by design: development of a bifunctional water splitting catalyst through an operando measurement directed optimization cycle. Chem. Sci. 9, 5322–5333 (2018) 89. Etienne, M., Dossot, M., Grausem, J., Herzog, G.: Combined Raman microspectrometer and shearforce regulated SECM for corrosion and self-healing analysis. Anal. Chem. 86, 11203–11210 (2014) 90. Clausmeyer, J., Nebel, M., Grützke, S., Kayran, Y.U., Schuhmann, W.: Local surface modifications investigated by combining scanning electrochemical microscopy and surface-enhanced Raman scattering. ChemPlusChem. 83, 414–417 (2018) 91. Pomfret, M.B., Owrutsky, J.C., Walker, R.A.: In situ studies of fuel oxidation in solid oxide fuel cells. Anal. Chem. 79, 2367–2372 (2007) 92. Li, X., Gewirth, A.A.: Oxygen electroreduction through a superoxide intermediate on bi-modified Au surfaces. J. Am. Chem. Soc. 127, 5252–5260 (2005) 93. Bohra, D., Ledezma-Yanez, I., Li, G., de Jong, W., Pidko, E., Smith, W.: Lateral adsorbate interactions inhibit HCOO while promoting CO selectivity for CO2 electrocatalysis on Ag. Angew. Chem. Int. Ed. (2018) 94. Chen, Y.X., Tian, Z.Q.: Dependence of surface enhanced Raman scattering of water on the hydrogen evolution reaction. Chem. Phys. Lett. 281, 379–383 (1997) 95. Bañares, M.A.: Operando methodology: combination of in situ spectroscopy and simultaneous activity measurements under catalytic reaction conditions. Catal. Today. 100, 71–77 (2005)

9

210 96. Deng, Y., Yeo, B.S.: Characterization of electrocatalytic water splitting and CO2 reduction reactions using in situ/operando Raman spectroscopy. ACS Catal. 7, 7873–7889 (2017) 97. Deng, Y., Handoko, A.D., Du, Y., Xi, S., Yeo, B.S.: In situ Raman spectroscopy of copper and copper oxide surfaces during electrochemical oxygen evolution reaction: identification of Cu III oxides as catalytically active species. ACS Catal. 6, 2473–2481 (2016) 98. Deng, Y., Ting, L.R.L., Neo, P.H.L., Zhang, Y.-J., Peterson, A.A., Yeo, B.S.: Operando Raman spectroscopy of amorphous molybdenum sulfide (MoSx) during the electrochemical hydrogen evolution reaction: identification of sulfur atoms as catalytically active sites for H+ reduction. ACS Catal. 6, 7790–7798 (2016) 99. James, O.O., Sauter, W., Schröder, U.: Towards selective electrochemical conversion of glycerol to 1,3-propanediol. RSC Adv. 8, 10818–10827 (2018) 100. Jiang, S., Klingan, K., Pasquini, C., Dau, H.: New aspects of operando Raman spectroscopy applied to electrochemical CO2 reduction on Cu foams. J. Chem. Phys. 150, 41718 (2019) 101. Barwe, S., Weidner, J., Cychy, S., Morales, D.M., Dieckhöfer, S., Hiltrop, D., Masa, J., Muhler, M., Schuhmann, W.: Electrocatalytic oxidation of 5-(Hydroxymethyl)furfural using high-surface-area nickel boride. Angew. Chem. Int. Ed. 57, 11460–11464 (2018) 102. Wiaderek, K.M., Borkiewicz, O.J., Castillo-Martínez, E., Robert, R., Pereira, N., Amatucci, G.G., Grey, C.P., Chupas, P.J., Chapman, K.W.: Comprehensive insights into the structural and chemical changes in mixed-anion FeOF electrodes by using operando PDF and NMR spectroscopy. J. Am. Chem. Soc. 135, 4070–4078 (2013) 103. Li, X., Yang, X., Zhang, J., Huang, Y., Liu, B.: In situ/operando techniques for characterization of single-atom catalysts. ACS Catal. 9, 2521–2531 (2019) 104. Wu, D.-Y., Li, J.-F., Ren, B., Tian, Z.-Q.: Electrochemical surfaceenhanced Raman spectroscopy of nanostructures. Chem. Soc. Rev. 37, 1025–1041 (2008) 105. Anantharaj, S., Ede, S.R., Sakthikumar, K., Karthick, K., Mishra, S., Kundu, S.: Recent trends and perspectives in electrochemical water splitting with an emphasis on sulfide, selenide, and phosphide catalysts of Fe, Co, and Ni: a review. ACS Catal. 6, 8069–8097 (2016) 106. Wang, J., Cui, W., Liu, Q., Xing, Z., Asiri, A.M., Sun, X.: Recent progress in cobalt-based heterogeneous catalysts for electrochemical water splitting. Adv. Mater. 28, 215–230 (2016) 107. Tahir, M., Pan, L., Idrees, F., Zhang, X., Wang, L., Zou, J.-J., Wang, Z.L.: Electrocatalytic oxygen evolution reaction for energy conversion and storage: a comprehensive review. Nano Energy. 37, 136–157 (2017) 108. Andronescu, C., Barwe, S., Ventosa, E., Masa, J., Vasile, E., Konkena, B., Möller, S., Schuhmann, W.: Powder catalyst fixation for post-electrolysis structural characterization of NiFe layered double hydroxide based oxygen evolution reaction electrocatalysts. Angew. Chem. Int. Ed. 56, 11258–11262 (2017) 109. Lu, X., Zhao, C.: Electrodeposition of hierarchically structured three-dimensional nickel-iron electrodes for efficient oxygen evolution at high current densities. Nat. Commun. 6, 6616 (2015) 110. Louie, M.W., Bell, A.T.: An investigation of thin-film Ni-Fe oxide catalysts for the electrochemical evolution of oxygen. J. Am. Chem. Soc. 135, 12329–12337 (2013) 111. Steimecke, M., Seiffarth, G., Bron, M.: In situ characterization of Ni and Ni/Fe thin film electrodes for oxygen evolution in alkaline media by a Raman-coupled scanning electrochemical microscope setup. Anal. Chem. 89, 10679–10686 (2017) 112. Lu, Z., Xu, W., Zhu, W., Yang, Q., Lei, X., Liu, J., Li, Y., Sun, X., Duan, X.: Three-dimensional NiFe layered double hydroxide film for high-efficiency oxygen evolution reaction. Chem. Commun. 50, 6479–6482 (2014) 113. Yeo, B.S., Bell, A.T.: In situ Raman study of nickel oxide and goldsupported nickel oxide catalysts for the electrochemical evolution of oxygen. J. Phys. Chem. C. 116, 8394–8400 (2012)

W. Schuhmann et al. 114. Yeo, B.S., Bell, A.T.: Enhanced activity of gold-supported cobalt oxide for the electrochemical evolution of oxygen. J. Am. Chem. Soc. 133, 5587–5593 (2011) 115. Rabe, M., Toparli, C., Chen, Y.-H., Kasian, O., Mayrhofer, K.J.J., Erbe, A.: Alkaline manganese electrochemistry studied by in situ and operando spectroscopic methods – metal dissolution, oxide formation and oxygen evolution. Phys. Chem. Chem. Phys. 21, 10457 (2019) 116. Diaz-Morales, O., Calle-Vallejo, F., de Munck, C., Koper, M.T.M.: Electrochemical water splitting by gold: evidence for an oxide decomposition mechanism. Chem. Sci. 4, 2334 (2013) 117. Faber, M.S., Dziedzic, R., Lukowski, M.A., Kaiser, N.S., Ding, Q., Jin, S.: High-performance electrocatalysis using metallic cobalt pyrite (CoS2) micro- and nanostructures. J. Am. Chem. Soc. 136, 10053–10061 (2014) 118. Hinnemann, B., Moses, P.G., Bonde, J., Jørgensen, K.P., Nielsen, J.H., Horch, S., Chorkendorff, I., Nørskov, J.K.: Biomimetic hydrogen evolution: MoS2 nanoparticles as catalyst for hydrogen evolution. J. Am. Chem. Soc. 127, 5308–5309 (2005) 119. Jaramillo, T.F., Jørgensen, K.P., Bonde, J., Nielsen, J.H., Horch, S., Chorkendorff, I.: Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts. Science. 317, 100–102 (2007) 120. Gao, M.-R., Liang, J.-X., Zheng, Y.-R., Xu, Y.-F., Jiang, J., Gao, Q., Li, J., Yu, S.-H.: An efficient molybdenum disulfide/cobalt diselenide hybrid catalyst for electrochemical hydrogen generation. Nat. Commun. 6, 5982 (2015) 121. Benck, J.D., Chen, Z., Kuritzky, L.Y., Forman, A.J., Jaramillo, T.F.: Amorphous molybdenum sulfide catalysts for electrochemical hydrogen production: insights into the origin of their catalytic activity. ACS Catal. 2, 1916–1923 (2012) 122. Benck, J.D., Hellstern, T.R., Kibsgaard, J., Chakthranont, P., Jaramillo, T.F.: Catalyzing the hydrogen evolution reaction (HER) with molybdenum sulfide nanomaterials. ACS Catal. 4, 3957–3971 (2014) 123. Kornienko, N., Resasco, J., Becknell, N., Jiang, C.-M., Liu, Y.-S., Nie, K., Sun, X., Guo, J., Leone, S.R., Yang, P.: Operando spectroscopic analysis of an amorphous cobalt sulfide hydrogen evolution electrocatalyst. J. Am. Chem. Soc. 137, 7448–7455 (2015) 124. Tran, P.D., Tran, T.V., Orio, M., Torelli, S., Truong, Q.D., Nayuki, K., Sasaki, Y., Chiam, S.Y., Yi, R., Honma, I., Barber, J., Artero, V.: Coordination polymer structure and revisited hydrogen evolution catalytic mechanism for amorphous molybdenum sulfide. Nat. Mater. 15, 640–646 (2016) 125. Zhu, Y., Chen, H.-C., Hsu, C.-S., Lin, T.-S., Chang, C.-J., Chang, S.-C., Tsai, L.-D., Chen, H.M.: Operando unraveling of the structural and chemical stability of P-substituted CoSe2 electrocatalysts toward Hhydrogen and oxygen evolution reactions in alkaline alectrolyte. ACS Energy Lett. 4, 987–994 (2019) 126. Bing, Y., Liu, H., Zhang, L., Ghosh, D., Zhang, J.: Nanostructured Pt-alloy electrocatalysts for PEM fuel cell oxygen reduction reaction. Chem. Soc. Rev. 39, 2184–2202 (2010) 127. Ge, X., Sumboja, A., Wuu, D., An, T., Li, B., Goh, F.W.T., Hor, T.S. A., Zong, Y., Liu, Z.: Oxygen reduction in alkaline media: from mechanisms to recent advances of catalysts. ACS Catal. 5, 4643–4667 (2015) 128. Jiao, Y., Zheng, Y., Jaroniec, M., Qiao, S.Z.: Design of electrocatalysts for oxygen- and hydrogen-involving energy conversion reactions. Chem. Soc. Rev. 44, 2060–2086 (2015) 129. Moussallem, I., Jörissen, J., Kunz, U., Pinnow, S., Turek, T.: Chloralkali electrolysis with oxygen depolarized cathodes: history, present status and future prospects. J. Appl. Electrochem. 38, 1177–1194 (2008) 130. Cheng, F., Chen, J.: Metal-air batteries: from oxygen reduction electrochemistry to cathode catalysts. Chem. Soc. Rev. 41, 2172–2192 (2012)

9

Operando Electrochemical Raman Spectroscopy

131. Kulkarni, A., Siahrostami, S., Patel, A., Nørskov, J.K.: Understanding catalytic activity trends in the oxygen reduction reaction. Chem. Rev. 118, 2302–2312 (2018) 132. Nørskov, J.K., Rossmeisl, J., Logadottir, A., Lindqvist, L., Kitchin, J.R., Bligaard, T., Jónsson, H.: Origin of the overpotential for oxygen reduction at a fuel-cell cathode. J. Phys. Chem. B. 108, 17886–17892 (2004) 133. Dong, J.-C., Zhang, X.-G., Briega-Martos, V., Jin, X., Yang, J., Chen, S., Yang, Z.-L., Wu, D.-Y., Feliu, J.M., Williams, C.T., Tian, Z.-Q., Li, J.-F.: In situ Raman spectroscopic evidence for oxygen reduction reaction intermediates at platinum single-crystal surfaces. Nat. Energy. 4, 60–67 (2019) 134. Radjenovic, P.M., Hardwick, L.J.: Time-resolved SERS study of the oxygen reduction reaction in ionic liquid electrolytes for non-aqueous lithium-oxygen cells. Faraday Discuss. 206, 379–392 (2018) 135. Hori, Y., Kikuchi, K., Suzuki, S.: Production of CO and CH4 in electrochemical reduction of CO2 at metal electrodes in aqueous hydrogencarbonate solution. Chem. Lett. 14, 1695–1698 (1985) 136. Jitaru, M., Lowy, D.A., Toma, M., Toma, B.C., Oniciu, L.: Electrochemical reduction of carbon dioxide on flat metallic electrodes. J. Appl. Electrochem. 27, 875–889 (1997) 137. Kuhl, K.P., Cave, E.R., Abram, D.N., Jaramillo, T.F.: New insights into the electrochemical reduction of carbon dioxide on metallic copper surfaces. Energy Environ. Sci. 5, 7050 (2012) 138. Kortlever, R., Shen, J., Schouten, K.J.P., Calle-Vallejo, F., Koper, M.T.M.: Catalysts and reaction pathways for the electrochemical reduction of carbon dioxide. J. Phys. Chem. Lett. 6, 4073–4082 (2015) 139. Dunwell, M., Lu, Q., Heyes, J.M., Rosen, J., Chen, J.G., Yan, Y., Jiao, F., Xu, B.: The central role of bicarbonate in the electrochemical reduction of carbon dioxide on gold. J. Am. Chem. Soc. 139, 3774–3783 (2017) 140. Wuttig, A., Yaguchi, M., Motobayashi, K., Osawa, M., Surendranath, Y.: Inhibited proton transfer enhances Au-catalyzed CO2-to-fuels selectivity. Proc. Natl. Acad. Sci. U. S. A. 113, E4585–E4593 (2016) 141. Kim, C., Jeon, H.S., Eom, T., Jee, M.S., Kim, H., Friend, C.M., Min, B.K., Hwang, Y.J.: Achieving selective and efficient electrocatalytic activity for CO2 reduction using immobilized silver nanoparticles. J. Am. Chem. Soc. 137, 13844–13850 (2015) 142. Hatsukade, T., Kuhl, K.P., Cave, E.R., Abram, D.N., Jaramillo, T.F.: Insights into the electrocatalytic reduction of CO2 on metallic silver surfaces. Phys. Chem. Chem. Phys. 16, 13814–13819 (2014) 143. Schmitt, K.G., Gewirth, A.A.: In situ surface-enhanced Raman spectroscopy of the electrochemical reduction of carbon dioxide on silver with 3,5-diamino-1,2,4-triazole. J. Phys. Chem. C. 118, 17567–17576 (2014) 144. Dick, L.A., McFarland, A.D., Haynes, C.L., van Duyne, R.P.: Metal film over nanosphere (MFON) electrodes for surfaceenhanced Raman spectroscopy (SERS): improvements in surface nanostructure stability and suppression of irreversible loss. J. Phys. Chem. B. 106, 853–860 (2002) 145. Cross, N.A., Pemberton, J.E.: Surface enhanced Raman scattering of pyridine on Ag electrodes formed with controlled-rate oxidation-reduction cycles. J. Electroanal. Chem. Interfacial Electrochem. 217, 93–100 (1987) 146. Zhang, W., Schmid, T., Yeo, B.-S., Zenobi, R.: Near-field heating, annealing, and signal loss in tip-enhanced Raman spectroscopy. J. Phys. Chem. C. 112, 2104–2108 (2008) 147. Lin, K.-Q., Yi, J., Zhong, J.-H., Hu, S., Liu, B.-J., Liu, J.-Y., Zong, C., Lei, Z.-C., Wang, X., Aizpurua, J., Esteban, R., Ren, B.: Plasmonic photoluminescence for recovering native chemical information from surface-enhanced Raman scattering. Nat. Commun. 8, 14891 (2017)

211

Wolfgang Schuhmann studied chemistry at the University of Karlsruhe, and completed his Ph.D. in 1986 at the Technical University of Munich. He was appointed professor for Analytical Chemistry at the Ruhr University Bochum in 1996. His research interests cover a broad spectrum of different fields of electrochemistry from bioelectrochemistry to nanoelectrochemsitry and electrocatalysis.

9 Denis Öhl after finishing his master’s degree and Ph.D. under supervision of Prof. Dr. Wolfgang Schuhmann, is now a post-doctoral researcher in Prof. Schuhmann’s group at the Ruhr University Bochum, Germany. His work focuses on electrocatalysis and particularly on the utilization of analytical tools such as nano –and spectroelectrochemistry for advanced characterization of metal catalysts under operando conditions.

Dulce M. Morales has a background in Industrial Chemical Engineering from Instituto Politécnico Nacional (México, 2012), and received her Masters (2015) and Dr. rer. nat. (2019) degrees from RuhrUniversität Bochum (Germany). She is presently an assistant professor at University of Groningen (the Netherlands). Her research focuses on the electrochemical conversion of renewable resources to fuels and value-added chemicals.

Sum Frequency Generation (SFG) Spectroscopy

10

Verena Pramhaas and Gu¨nther Rupprechter

Contents 10.1

Introduction to Sum Frequency Generation (SFG) Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

10.2 10.2.1

SFG Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 SFG Signal Intensity and Lineshape . . . . . . . . . . . . . . . . . . . . . 215

10.3

SFG Instrumentation and Operation Modes . . . . . . . . . . 216

10.4

10.4.4

Applications of SFG Spectroscopy and Selected Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SFG Spectroscopy on Metal Surfaces . . . . . . . . . . . . . . . . . . . . SFG Spectroscopy on Oxide Surfaces . . . . . . . . . . . . . . . . . . . SFG Spectroscopy on Polymer and Biomaterial Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SFG Spectroscopy of Water and Ice Layers . . . . . . . . . . . . .

10.5

Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

10.4.1 10.4.2 10.4.3

218 219 224 226 227

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

Abstract

Infrared-visible sum frequency generation (SFG) laser spectroscopy is an interface-specific nonlinear optical method that provides vibrational spectra of molecules located at surfaces or interfaces. Provided the interface is accessible to light, SFG allows identifying chemical species, molecular densities, and molecular orientations, with sub-picosecond time resolution. Resulting from its selection rules, SFG is typically not sensitive to bulk or isotropic phases, making the method highly interface-specific. SFG can thus be utilized to investigate buried interfaces in gaseous or liquid environments. Keywords

SFG · In situ spectroscopy · Surface sensitivity · Adsorbates · Ambient pressure

V. Pramhaas · G. Rupprechter (*) Institute of Materials Chemistry, TU Wien, Vienna, Austria e-mail: [email protected]

10.1

Introduction to Sum Frequency Generation (SFG) Spectroscopy

Molecules located at surfaces and interfaces exhibit important functions in various fields of chemistry, physics, materials science, biology, and medicine. Specific applications include, among numerous others, anticorrosive, antireflecting, self-cleaning, antibacterial coatings, optical devices and quantum dots, dyes of solar cells, magnetic devices, microelectronics and data storage, sensors, selfassembled monolayers, polymers, membranes, etc. Among these technologies, heterogeneous catalysis is a classical surface phenomenon that is widely used for gross and fine chemical synthesis, syngas production, CO2 utilization, waste remediation, and exhaust gas treatment. In the context of “green chemistry,” the further development of atom- and energy-efficient and environmentally benign processes holds large promises toward a sustainable society [1, 2]. It is thus easy to understand that studies of molecules at surfaces and interfaces are key to future technological advances. Among the many physical methods characterizing molecules, vibrational spectroscopy is one of the most versatile ones [3, 4]. Vibrations can be excited by (far-, mid-, and near-) infrared as well as visible light and, indeed, infrared and Raman spectroscopy are the most frequently applied vibrational spectroscopies. One of their downsides is, however, that they are not surface-specific, i.e., the vibrational signatures of molecules at surfaces/interfaces may be obscured by surrounding gaseous or liquid bulk phases. Clearly, there are ways to overcome such obstacles, e.g., by specific measurement geometries, polarization, and concentration modulation [5, 6], thin films and nanoparticles, etc. Nevertheless, in some way or the other the vibrational background of bulk phases must be subtracted before the surface/ interface-specific information can be deduced.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_10

213

214

V. Pramhaas and G. Rupprechter

This is when one calls for sum frequency generation (SFG), as this laser spectroscopic method is inherently surface-specific, i.e., it only records vibrational signatures of molecules at surfaces and interfaces [7, 8]. The differentiation between surfaces/interfaces and bulk phases is based on symmetry considerations and selection rules, as described in detail below. Consequently, bulk phases do not show up in SFG, which makes it a prime method for interfacial studies but also for the increasingly popular in situ and operando studies. As a laser spectroscopic method, SFG also holds advantages with respect to time resolution (ps to fs) [9–11] and molecular orientation (polarization dependence) [12–15]. As illustrated below, SFG can be applied to various solid–vacuum, solid–gas, solid–liquid, liquid–gas, and liquid–liquid interfaces. In the following, the basic principles and operation modes of SFG spectroscopy will be explained, followed by a series of carefully selected case studies, intended to illustrate how SFG can be successfully applied to solve various problems in interface science.

10.2

SFG Theory

SFG spectroscopy makes use of the second-order nonlinear optical process of sum frequency generation, i.e., two light waves at different frequencies interact in a medium characterized by a second-order nonlinear susceptibility tensor χ (2), and generate a wave at the sum of their frequencies (ωSFG ¼ ω1 þ ω2) [8]. Because second-order susceptibility is typically Fig. 10.1 Illustration of vibrational IR-vis sum frequency generation (SFG). In case of a vibrational resonance of an adsorbed molecule, visible light is generated at a frequency that is the sum of the frequencies of two incident optical fields (ωSFG ¼ ωIR þ ωvis). Light beams can be p (parallel to plane of incidence) or s (“senkrecht”, orthogonal to plane of incidence) polarized. The angle of the resulting SFG beam lies in between the angles of the IR and visible light. A schematic energy level diagram shows the steps of this nonlinear process

several magnitudes smaller than the first order, this process yields only a small signal, so high incident light intensities, i.e., pulsed lasers, are required. To acquire an SFG vibrational spectrum of molecules adsorbed on a solid surface, two (e.g., picosecond) laser pulses are spatially and temporally overlapped on the sample (Fig. 10.1). One input beam is in the visible range at fixed frequency (ωvis), and the second one is tunable in the mid-IR region (ωIR) to probe the vibrational modes of the surface species. In a simplified picture, when the IR beam is tuned through a vibrational resonance of the adsorbate, it induces a vibrational transition from the ground state to an excited state, and simultaneously the visible beam induces a transition to a higher-energy virtual state through an anti-Stokes Raman process (cf. energy scheme in Fig. 10.1). When the high-energy virtual state relaxes, light is generated at a frequency that is the sum of the frequencies of two incident optical fields (ωSFG ¼ ωIR þ ωvis), resulting in a signal in the visible region. By tuning of the wavelength of the IR beam and monitoring of the intensity of the SFG output, an adsorbate vibrational spectrum is obtained as a plot of the SFG intensity against the IR wavenumber. As mentioned, the power of SFG lies in its surface-specificity. According to the selection rules discussed below, only vibrational modes that simultaneously satisfy both IR and Raman selection rules are SFG-active. Therefore, SFG is not allowed in media with inversion symmetry (such as in the centrosymmetric bulk of a solid or in an isotropic gas or liquid phase), but has a finite value at an interface, where the inversion symmetry is broken. The underlying physical principles are discussed in the following.

w

IR

p-polarization

q IR

q SFG

s-polarization

wV

w SFG

q VIS

IS

90°

E

Virtual level

Vibrational level

10

Sum Frequency Generation (SFG) Spectroscopy

215

The interaction between light and matter induces a polarization p. In the electric–dipole approximation the microscopic polarization p and the electric field E are related through molecular (hyper)polarizability tensors α, β, and γ, as well as other higher order terms (Eq. 10.1) [8, 16, 17]. p ¼ α E þ β EE þ γ EEE þ . . . :

ð10:1Þ

The polarization produced by conventional light sources scales linearly with the electric field E (coefficient α; polarizability), whereas irradiation with intense laser light induces significant nonlinear polarization, characterized by the higher-order hyperpolarizability tensors β and γ. However, macroscopic amounts of substances are typically investigated (involving large numbers of molecules). Macroscopic polarizability tensors χ (n) are obtained by multiplying the ensemble/orientation averages (denoted “< >”) of the corresponding molecular (hyper)polarizabilities with the number of molecules N (Eq. 10.2)

χ

ð1Þ

¼ N < α >, χ

ð2Þ

¼ N < β >, χ

ð3Þ

¼ N < γ > ð10:2Þ

and the macroscopic polarization P can thus be written as P¼χ

ð1Þ

Eþχ

ð2Þ

EE þ χ

ð3Þ

EEE þ : . . .

ð10:3Þ

The magnitude of the local electric field E is described as

ð2Þ

¼χ

ð2Þ

EE:

ð10:4Þ

ð10:5Þ

The electric field at a position r irradiated by two light waves with different frequencies ω1 and ω2 can be expressed by the vector sum of the two electric fields, Eðr, tÞ ¼ Eðω1 ; r, tÞ þ Eðω2 ; r, tÞ:

ð2Þ

EðωIR Þ Eðωvis Þ

ð10:6Þ

Using the electric field from Eq. (10.6) for E in Eq. (10.5) yields a second-order polarization P(2), comprising polarizations oscillating at ω1 + ω2 (sum frequency generation, SFG)

ð10:7Þ

As shown below, when the frequency/energy of the IR beam is in resonance with a vibrational transition of a surface/interface species, the SFG intensity is resonantly enhanced giving rise to a signal. IR-vis SFG has thus become a versatile tool for vibrational spectroscopy of molecules located at surfaces and interfaces, utilizing frequency-tunable or broadband infrared and visible lasers [7, 18–23]. As mentioned, not all vibrational modes are SFG active, as discussed below.

10.2.1 SFG Signal Intensity and Lineshape The intensity of the generated sum frequency output is proportional to the square of P(2), i.e., to the absolute square of the effective second-order susceptibility of the surface/interface, denoted χ s(2), and to the intensities of the two incident IR and vis beams. 2

ð2Þ

with r being the position vector, t as the time, E0(ω) the amplitude vector, k the wave vector, and c.c. as the complex conjugate. With increasing electric field strength, the second-order term gets more relevant in optical processes, while third- and higher-order terms are still nearly negligible due to small prefactors. In the following, we will thus focus on this second-order (nonlinear) polarization P(2), which is responsible for SFG, with P

ð2Þ

P ðωSFG ¼ ωIR þ ωvis Þ ¼ χ

I SFG ðωIR Þ / jχ s ðωIR , ωvis Þj Ι ðωIR Þ Ι ðωvis Þ

iðωtkrÞ

Eðω; r, tÞ ¼ E0 ðωÞ e þ c:c: ¼ 2E0 ðωÞ cos ðω t  krÞ

and ω1  ω2 (difference frequency generation, DFG), in addition to polarizations oscillating at zero frequency (optical rectification), 2ω1 (second harmonic generation, SHG of ω1), and 2ω2 (SHG of ω2). For the specific case of sum frequency generation by infrared and visible light (IR-visible SFG), the nonlinear polarization P(2) is

ð10:8Þ

Whereas Ι(ωIR) and Ι(ωvis) “simply” depend on the sources (lasers) of the IR and vis radiation, the surface/interface properties are characterized by χ s(2), which has two components: The first part is the “resonant” nonlinear susceptibility χ r(2) near vibrational resonances ð2Þ

χ r ðωIR Þ ¼

X q

Ar ð qÞ ωIR  ωq þ iΓq

ð10:9Þ

where AR(q), ωq, and Γq are the resonance amplitude, resonance frequency, and damping constant (homogeneous linewidth 2Γq ¼ FWHM) of the qth vibrationally resonant mode, and ωIR is the IR laser frequency. The term χ r(2) incorporates the resonance condition (ωIR  ωq), and as the IR beam is tuned through vibrational resonances of surface species, χ r(2) and thus ISFG reach a maximum. The amplitude of the vibrationally resonant susceptibility Ar(q) is determined by the adsorbate concentration (number density N ) and the product of the IR and Raman transition moments of the vibration (Tq, Mq; δρ is the population difference between the vibrational ground and excited state).

10

216

V. Pramhaas and G. Rupprechter

ArðqÞ / N T q Mq δρ

ð10:10Þ

Equation (10.10) represents the selection rule for the SFG process. In order to generate a sum frequency emission, the excited vibrational mode must be both IR and Raman active. Therefore, SFG is not allowed in media with inversion symmetry (due to the “Rule of Mutual Exclusion”). Sum frequency generation, or any other second-order nonlinear optical process, therefore does not occur in (isotropic) gases, liquids, or centrosymmetric solids. The SFG signal is hence dominated by the vibrational modes of the molecules at the surface/interface where the inversion symmetry is broken, making SFG inherently surface/interface specific. Nevertheless, as most other spectroscopic methods, SFG has to cope with a signal background. In addition to the resonantly enhanced nonlinear susceptibility χ r(2) (related to surface/adsorbate vibrations), the surface/interface may generate an SFG signal that is not connected to vibrational resonances, characterized by χ nr(2). A “nonresonant” SFG signal can be produced when a surface electronic state is excited by either the visible or SFG light. However, in most cases the visible or SFG light frequencies are far from resonances of the (bulk-)surface and the nonresonant signal arises from oscillations of localized surface electrons. The nonresonant response is not changed by different IR frequencies and can thus be approximated by a frequency-independent nonresonant susceptibility χ nr(2) (cf. Eq. 10.11; assuming that this also incorporates nonresonant contributions of higherorder components). Consequently, χ s(2)(ωIR) can be expressed as

ð2Þ

ð2Þ

ð10:11Þ

where Anr is the amplitude of the vibrationally nonresonant susceptibility, and ϕ represents its phase relative to the resonant term. Interference between the resonant and the nonresonant parts governs the SFG lineshape [24]. In case χ nr(2)  χ r(2), nearly symmetric lineshapes are obtained (see Fig. 10.2 green curve), with an increase of the total SFG intensity at the resonance frequency. However, as apparent from Eq. (10.11), when χ nr(2) ≈ χ r(2), the lineshape of an SFG-resonance is primarily determined by the phasedifference ϕ between the resonant and nonresonant signal. The phase-difference is a result of the different physical origin of vibrationally resonant and nonresonant SFG fields/ signals. Depending on ϕ, the sum of χ r(2) þ χ nr(2) (Eq. (10.11)) may give rise to an asymmetry or even a signal decrease at the resonance frequency, as seen for the blue curve with 1:1 amplitude ratio in Fig. 10.2.

10.3

SFG Instrumentation and Operation Modes

The main components of an SFG spectrometer include (i) a laser system to generate the incident visible and IR beams (ps or fs pulses), (ii) a sample preparation/manipulation stage (e.g., a spectroscopy cell with controlled environment), and

f=0 10000

8000

6000

|c (2)|2

Fig. 10.2 Simulated lineshape of the SFG signal at resonance with a phase difference of ϕ ¼ 0. Different amplitude ratios between resonant and nonresonant contributions are displayed: Ar/Anr ¼ 20, i.e., with almost negligible background, Ar/Anr ¼ 5, with obvious asymmetry, and Ar/Anr ¼ 1, with a similar amount of both contributions leading to regions of decreased and increased intensity throughout the resonance

ð2Þ

χ s ðωIR Þ ¼ χ r ðωIR Þ þ χ nr X A r ð qÞ iϕ ¼ þ Anr e q ωIR  ωq þ iΓq

Ar /Anr = 1

4000

2000 Ar /Anr = 5 Ar /Anr = 20

0 –6

–4

–2

0 Δw

2

4

6

10

Sum Frequency Generation (SFG) Spectroscopy

217

(iii) a detection system for the generated SF signal as shown in Fig. 10.3. For scanning SFG systems, most frequently, neodymium yttrium-aluminum-garnet (Nd:YAG) and titanium sapphire (Ti:Sa) lasers are used [13, 25–29]. The output of a Nd:YAG picosecond laser (1064 nm, ~30 mJ/pulse, ~25 ps, 10–50 Hz) is partly converted to 532-nm light by second-harmonic generation (SHG) using, e.g., a KD*P (Potassium Dihydrogen Phosphate) crystal. About 200 μJ/pulse 532-nm light is used as a visible (green) beam in the SFG experiment; the 1064- and remaining 532-nm beams are mixed in an optical parametric generator/amplifier (OPG/OPA) to generate tunable IR light (3–6 μm, ca. 200 μJ/pulse, resolution ~3 cm1) in a difference frequency generation (DFG) stage [13]. A delay line allows optimizing the temporal overlap of IR and visible pulses and there are also devices (half-plates, polarizers (Glan-Taylor prisms)) to adjust the intensity and the polarization of the visible and IR beam. Using different DFG crystals (such as AgGaS2, AgGaSe2) or even synchrotron radiation [30] allows carrying out spectroscopy in a range of ~1000–4000 cm1. In order to differentiate the (weak) SFG signal from the (strong) reflected vis beam a dedicated detection system is required, which combines spatial and spectral filtering [31]. In co-propagating geometry, the IR and visible beams have different incidence angles, e.g., ~55 and 50 respectively, producing an SFG signal that lies in between them (compare to Fig. 10.1)

and can be (spatially) separated by an aperture (as shown in Fig. 10.3). The SFG signal is then further (spectrally) filtered by an edge filter and a monochromator (both remove the remaining 532 nm light and allow only the SFG light to pass) before it reaches the detector, which in case of a scanning IR beam is a photomultiplier, the signal of which is processed by a gated boxcar integrator (triggered by the laser for temporal filtering), and sent to a PC via an analogue-to-digital interface. SFG spectroscopic measurements can be carried out in different modes, such as scanning SFG (as described above), broadband SFG, and time-resolved (pump-probe) SFG, and can be altered by polarization changes. In case of scanning SFG, the IR energy is tuned stepwise over the range of interest (which makes this mode rather slow). To speed up spectral acquisition, “broadband” SFG takes advantage of ultrashort and thus spectrally broad IR laser pulses (e.g., 150 fs; width ~150 cm1), covering an IR region of interest, e.g., centered around C-O or C-H stretching vibrations ([7, 27, 32, 33] and references therein). The broadband IR pulse is overlapped with a narrowband visible pulse (e.g., 7 ps; width 2 cm1). Dispersing the spectrum allows measuring all resonances within the IR range at once, with resolution limited by the bandwidth of the visible pulse. The broadband approach thus allows capturing an entire SFG vibrational spectrum in a few single laser shots, without tuning of the IR wavelength.

Sample preparation e.g. vacuum chamber

IR laser pulse scanning ps/broadband fs

Sample stage w

IR

Delay

wV

VIS laser pulse narrowband

Photon detector multiplier tube/CCD screen

w SFG 90°

IS

Iris

Polarizer

Co-propagating

Counter-propagating

IR Vis

SFG

Vis

SF

Time resolved

Total internal reflection IR

IR

G IR

Fig. 10.3 Schematic illustration of an SFG spectrometer. The IR laser pulse is generated either by a scanning picosecond laser system or a broadband femtosecond laser, while the visible pulse is a narrowband pulse. Polarizers can be used to set a specific beam polarization. The sample stage is often coupled with corresponding sample preparation

Vis

Pump

SFG

Vis

SFG

possibilities, i.e., for catalysis tests the sample stage could be in a gas cell, connected to a vacuum chamber for sample preparation. The resulting SFG beam is spatially separated from the incident beams by an iris aperture before further spectral filtering and detection. The most common beam geometries are shown in the lower panels

10

218

An increasingly popular form of detection is heterodyne SFG spectroscopy. In comparison to the straightforward homodyne detection, a second surface is used to generate a reference SFG signal, functioning as a local oscillator. Using the interference between local oscillator and measured SFG signal, the measurement sensitivity can be improved and the phase of the nonlinear susceptibility can be deduced directly. However, the additional surface and different signal detection require a more complex optical design and the phase must be calibrated for different surfaces and geometries. This technique is used in the second case study presented in Sect. 10.4.4, showing the imaginary part of the nonlinear susceptibility, which is directly dependent on the vibrational resonances. It therefore does not sensitively depend on signal fitting, which may lead to significant errors for some systems with homodyne detection. For time-resolved “pump-probe” SFG spectroscopy, the surface species are first excited by a “pump” laser pulse, e.g., intense near-IR (ps or fs) for thermal excitation, which pushes the system out of its equilibrium, followed by a time-delayed SFG (IR and vis) “probe,” which monitors the changes in the vibrational properties of the adsorbate–substrate complex. Varying the delay time between pump and probe, the effect of the excitation can be studied. For example, excitation (pump) may allow detection of surface reaction intermediates that are too short-lived for steady-state scanning SFG. By taking “snapshots” of the transient vibrational spectrum at different delay times, the lifetime of transient species can be determined. A prerequisite for such measurements is, of course, that the surface species relax in between the probe and next pump pulses (e.g., for measurements at 1 kHz, the adsorbate-substrate system must relax within 1 millisecond). In the same way vibrational coupling, etc., can be studied. Time-resolved measurements with different pump frequencies can be used to investigate temperature/vibrational relaxation (shown in the second case study in section “SFG Spectroscopy on Metal Single Crystals”). The so-called 2D SFG uses a pump beam scanning through different frequencies to excite vibrations to show cross coupling between different states (this is illustrated in the second case study in Sect. 10.4.2). Using lasers also allows one to take advantage of the light polarization. SFG is frequently carried out in ppp geometry on solid surfaces (i.e., by detecting a p (parallel)-polarized SFG signal produced by a p-polarized visible and a p-polarized IR beam). This combination is typically used for adsorption/reaction investigations of metal surfaces, because it produces the most intense adsorbate SFG signal [14, 34]. However, other polarization combinations can also be used, for example, ssp (s-polarized SFG, s-polarized visible and p-polarized IR) is used preferably on water interfaces. For metals, the IR beam is always p-polarized because

V. Pramhaas and G. Rupprechter

the light field of an s-polarized IR beam is screened by the conduction electrons of metal surfaces. Comparison of the signal intensities for different polarization combinations (e.g., Ippp vs. Issp) can then be utilized to obtain information about the molecular orientation of surface species [13, 14, 35–37]. The Ippp/Issp ratio is used in case studies in section “SFG Spectroscopy on Metal Single Crystals” and Sect. 10.4.3. It is often specifically interesting for longer organic molecules combined with studies of overall changes in intensity. Applications of SFG microscopy have also been reported, imaging the spatial distribution of selected SFG vibrational bands, thus combining the analytical ability of IR with the spatial resolution of visible light microscopy [38–44]. Figure 10.4 shows a recently developed SFG setup that combines sample preparation and characterization in ultrahigh vacuum (UHV) with SFG vibrational spectroscopy from UHV to atmospheric-pressure (sample temperature from 90 to 1273 K). This versatile setup allows performing surface science, SFG spectroscopy, catalysis, and electrochemical investigations on model systems, including single crystals, thin films, and deposited metal nanoparticles, under wellcontrolled conditions of gas composition, pressure, temperature, and potential. The UHV section enables preparing model surfaces/catalysts and analyzing their surface structure and composition by low energy electron diffraction (LEED) and Auger electron spectroscopy (AES), respectively. Thereafter, a sample transfer mechanism moves samples under UHV to the spectroscopic cell, avoiding air exposure. In the catalytic cell, SFG spectroscopy and catalytic tests (reactant/ product analysis by mass spectrometry or gas chromatography) are performed simultaneously, representing a true operando approach. For post-reaction analysis, the SFG cell is rapidly evacuated and samples are transferred back to the UHV chamber. Results obtained by this versatile setup are shown in sections “SFG Spectroscopy on Metal Single Crystals” (polarization dependent SFG) and “SFG Spectroscopy on Supported Metal Nanoparticles” (catalytic measurements).

10.4

Applications of SFG Spectroscopy and Selected Case Studies

Applications of SFG include solid–gas, solid–liquid, solid–solid, liquid–gas (air), and liquid–liquid interfaces, and each combination has been reviewed in detail [7, 18, 19, 21, 22, 45–47]. SFG can be used for a large range of different materials, as long as the surface allows guiding a decent amount of SFG light toward the detector. In the following sections we discuss selected recent case studies for different material interfaces to illustrate the versatility of

10

Sum Frequency Generation (SFG) Spectroscopy

219

10 Fig. 10.4 SFG setup for catalysis measurements. The whole setup consists of a load lock (orange), a UHV preparation chamber (blue) and a UHV-to-high-pressure chamber for SFG spectroscopy during catalytic processes (purple). A front view of the SFG chamber is shown in the left inset showing the beam paths. The sample is mounted

on a sample stage that allows cooling to LN2 temperature and heating to 1273 K, shown in the right inset. Sample receivers in both chambers are connected to manipulators with linear translation in three dimensions and rotation along the manipulator axis

SFG. The case studies also feature various SFG setups as discussed above, like using polarization, time dependence, and other features taking advantage of specific SFG properties.

SFG Spectroscopy on Metal Single Crystals The early surface science studies employing SFG examined the adsorption of small molecules such as CO, NO, or C2H4 on metal single crystal surfaces [23, 51–56]. The SFG vibrational spectra provided insight into the adsorption configuration and adsorption sites, for example, the relative population of hollow/bridge/on-top bonded CO or di-σ/π-bonded ethylene. Coverage-dependent and temperature-dependent measurements yielded information on adsorbate–adsorbate interactions and adsorbate-binding energies. As an example, Fig. 10.5a shows SFG spectra of CO adsorption on Pt(111) and Pd(111), measured in ppp polarization. This polarization combination, with all beams (SFG, vis, IR) being p (parallel) polarized, typically yields the strongest signal on metal surfaces, but using additional polarization combinations provides information on the molecular orientation of the adsorbed molecules [14]. As mentioned, on metals the IR laser beam must be p-polarized, as the surface electric

10.4.1 SFG Spectroscopy on Metal Surfaces Early applications of SFG on solid surfaces were dedicated to examine the adsorption of small molecules in a UHV environment ([7, 48, 49] and references therein). Soon after, timeresolved and polarization-dependent spectroscopy were carried out [37, 50] and the surface-specificity of SFG was utilized to monitor surface-adsorbed molecules at elevated (ambient) gas pressure, during catalytic reactions and in liquid environments during electrochemistry [23, 34]. To introduce the method we start with the prototype system of CO/Pt(111).

V. Pramhaas and G. Rupprechter

50×10–6

a) PPP

8×10–3

2092

Pt 40

6

mbar

2073

Simulated Ippp

Normalized SFG intensity (a.u.)

220

Pd

400 4

b) PPP

30 20

4

Pt

10

2

10

6 1800

2000

1900

0

2100

0.4

Wavenumber (cm–1)

4

0.8

1.2

1.6

2.0

R = baac /bccc 100

c) SSP

d) Ippp /Issp 80

Issp(Pd)/Issp(Pt) |2

Pd

Simulated Ippp /Issp

5×10–6

Simulated Issp

Pd

Ippp(Pd)/Ippp(Pt) |3

3 2

Pt

R = 0.49

60

Pd: 48 40 Pt: 27 20

1 0

0 0.4

0.8

1.2

1.6

2.0

R = baac /bccc

0.4

0.8

1.2

1.6

2.0

R = baac /bccc

Fig. 10.5 Comparison of the SFG intensity of on-top CO on Pt(111) and Pd(111). (a) Experimental SFG spectra in ppp polarization at 300 K for Pt in blue and Pd in purple. (b)Simulated Ippp versus R. (c) Simulated Issp versus R. (d) Simulated Ippp/Issp ratio versus R. Parameters used in the simulation are αIR ¼ 55 , αVIS ¼ 58.5 , ωVIS ¼ 532 nm, N ¼ 1,

βccc ¼ 1, and surface tilt θ ¼ 0 . (Adapted from Ref. [15], distributed under CC-BY, http://creativecommons.org/licenses/by/4.0/, from Ref. [57], copyright (2002) American Chemical Society, and from Ref. [58] copyright (2003) with permission from Elsevier)

field of an s-polarized IR laser beam is screened by the conduction band electrons of the metal. Contrary, the visible beam can either be s- or p-polarized, because the surface electric field is not as efficiently screened due to the lower dielectric constants of metals in the visible region. The polarization of the SFG output can be either s- or p-polarized, leading to the ppp and ssp polarization combinations (ssp refers to a s-polarized SFG, s-polarized visible, and p-polarized IR; psp and spp yield little to no signal on metals [14]). In a simplified picture, one can state that an SFG signal is generated only when the electric fields of the visible and infrared light have components parallel to the molecular bond axis. Consequently, ppp detects predominantly molecules with molecular axes parallel or slightly inclined to the surface normal, whereas ssp spectra are more sensitive to tilted molecules. Thus, by comparing signal intensities Ippp and Issp

of ppp and ssp spectra, respectively, the orientation (tilt angles) of molecular bonds can be determined. The quantitative analysis of polarization-dependent SFG spectra is quite complex, as described in the literature [14, 56, 59, 60]. On the other hand, if the tilt angle is already known from independent measurements (other than SFG), the Ippp to Issp ratio can be used to calculate molecular hyperpolarizabilities. Previous studies comparing ppp and ssp spectra of CO adsorbed on Pt(111) clearly illustrated the upright (“perpendicular”) adsorption geometry, leading to on-top (linear) CO on individual Pt atoms. Simulations of ppp and ssp intensities (Fig. 10.5b, c, respectively) can then be used to calculate molecular hyperpolarizability ratios R as shown in Fig. 10.5d. If this value is known, adsorbates on more complex surface structures can be interpreted or differences for special overlayer structures can be understood. For CO on Pd(111) [14]

10

Sum Frequency Generation (SFG) Spectroscopy

221

and Ni(111) [13] an upright orientation of on-top and bridged CO was reported as well. Having managed small molecules on single crystal surfaces, systems of higher complexity can be examined by SFG, such as an electro-catalytic Re complex on a gold single crystal electrode (Fig. 10.6) upon light-induced excitation by an additional IR pump pulse. Such studies may reveal new routes to influence catalytic processes. After combining SFG spectroscopy with theory simulations to determine the molecular orientation, time-dependent studies were performed to investigate vibrational relaxation dynamics and their possible influence on reaction pathways. The pump-probe-spectra in Fig. 10.6a show peaks of different C-O vibrations at varying delay times after an initial excitation. The measurements were performed in ppp polarization combination, accounting for the metal surface and orientation of the adsorbate. Three peaks were observed featuring strong changes, at 2021 cm1, a ground state bleach by stimulated emission of the a’(1) mode, at 1991 cm1, an excited state absorption of the a’(1) mode, and at 1920 cm1, a ground state bleach by stimulated emission of low frequency modes. The intensity changes of the three resonances

a) 0.05

are plotted versus time in Fig. 10.6b, displaying the temporal evolution of energy transfer after excitation. The decay consists of a fast and a slow component. Using theoretical calculations it was shown that the decay time was consistent with energy transfer through electron-hole pair excitation. As electron hole pairs can strongly influence electronic states at the surface, this observation may open up new ways to influence catalytic processes.

SFG Spectroscopy on Supported Metal Nanoparticles Apart from single crystal measurements, SFG can also be used for studies of supported metal nanoparticles, which are more relevant to technological catalysis. Clearly, the macroscopic surface must still be flat enough to allow sufficient signal collection, why model systems were mostly used. The first SFG spectra of CO adsorbed on oxide supported Pd nanoparticles (mean size 3–6 nm) were reported by Dellwig et al. [63], those of CO on Pt aggreagates (mean size around 40 nm) by Baldelli et al. [64]. Figure 10.7 shows SFG spectra of CO adsorption on UHV grown ~6 nm Pd nanoparticles supported on a thin Al2O3 film

b)

'S(t,w)

'S(t,w) –0.10 1800

2021 cm–1 1991 cm–1 1920 cm–1

0.00

0.00

–0.05

0.05

0~0.3 ps 0.3~1 ps 1~5 ps 5~10 ps 10~20 ps 20~30 ps 30~50 ps 1850

1900

0.05 –0.05

0.00 –0.05 –0.10

–0.10 1950

2000

2050

2100

–1 0 0

20

40 60 Delay time (ps)

IR frequency (cm–1)

c)

d)

ac(2): 1883 cm–1

1

2

3 80

4

5 100

e)

as: 1977 cm–1

Fig. 10.6 Time-dependent spectral changes of RediCN complex on Au. (a) Time resolved SFG difference spectra at different delay times. (b) Kinetics at 2021 cm1, 1991 cm1, and 1920 cm1 with kinetic fits. (c–e) CO stretching modes of the complex on Au. The green frame

ac(1): 2028 cm–1

highlights the axial CO. Ab initio frequencies of the resonances are given underneath. (Adapted with permission from Ref. [61], copyright (2019) American Chemical Society, and Ref. [62], copyright (2018) American Chemical Society)

10

222

V. Pramhaas and G. Rupprechter

Pd-Al2O3 600 mbar

CO pressure

1990

SFG intensity (a.u.)

1895

High

2100

Low

10–6 mbar 1800

1900

190 K 2000

2100

2200

Wavenumber (cm–1)

Fig. 10.7 SFG spectra of CO on Pd nanoparticles supported on alumina. The lower spectrum (black) was taken at 106 mbar CO, the upper spectrum (red) at 600 mbar CO. Three peaks were observed characterizing hollow (1895 cm1), bridge (1990 cm1), and linearly bonded CO (2100 cm1). While at high pressure all peaks were well visible, for lower pressure the bridge bonded peak was significantly larger and a different peak asymmetry was observed. (Adapted with permission from Ref. [65]. © Materials Research Society 2007)

with spectra were acquired at 190 K in a pressure range from 106 up to 1000 mbar. Based on SFG reference measurements on smooth and rough single crystals, IR spectroscopy and density functional theory, the observed peaks were assigned to CO adsorbed on threefold hollow sites (1895 cm1), bridge bonded CO at particle edges (1990 cm1) and particle terraces (1950 cm1), and to linearly (on-top) bonded CO (~2100 cm1). No high-pressure species were observed, but there were marked differences between Pd nanoparticles and Pd single crystals, as described in more detail in Refs. [5, 65]. In addition to the adsorbate geometries discussed before, SFG of nanoparticles may also yield information on morphology. Alyabyeva et al. [66] studied CO adsorption on Al2O3/ Ni3Al(111)-supported Pd nanoparticles of various size, controlled by the amount of deposited metal (from 0.0025 to 2 ML). Measurements were performed in ppp polarization at 108 and 1 mbar CO. Based on their previous studies using Pd nanoparticles on MgO [67] and theoretical calculations [68], the observed resonances seen in Fig. 10.8a–d were attributed to bridge (1900–1980 cm1) and linear bonded CO (2020–2100 cm1).

Furthermore, one can differentiate linear CO and bridge bound CO on terraces from edge bound linear CO. The relative amount of edge to terrace sites deduced from SFG spectra is shown in Fig. 10.8e for various particle sizes. After fitting the molecular hyperpolarizability, calculations for edge/terrace ratios are given for varying Pd amounts/geometries. Four geometries were investigated, that is half spheres and 1-, 2-, and 3-layer flat pyramids. Based on these measurements, a switch in the nucleation and growth regime can be observed. The edge ratio indicated that flat multilayer particles converted to half-spheres when the deposited Pd amount exceeded 1 ML. Pd nanoparticles have also been used for catalytic studies. As significant differences were observed between spectra on (low-index) single crystal and (multifaceted) nanoparticles [57, 63] it is advisable that catalytic model studies should be performed on nanoparticle model catalysts. For example, SFG was employed to monitor the co-adsorption of CO and hydrogen under UHV, and atmospheric pressure CO hydrogenation, both on oxide supported Pd nanoparticles and Pd (111). For the high pressure/high temperature reaction the relative populations of hollow and on-top sites were different from those observed under UHV and indicated a partly disordered CO overlayer and/or surface roughening [7, 69, 70]. Because the coverage of the CO layer under reaction conditions was quite high (around 0.5 ML), preventing dissociative hydrogen adsorption, the disorder/roughening seems essential to allow for CO-H reaction under the conditions of technical catalysis. Furthermore, Pd nanoparticles supported on Al2O3/NiAl (110) and Pd(111) were employed in studies to examine methanol oxidation (partial oxidation to formaldehyde CH2O or full oxidation to CO2) [7, 71–73]. Under the applied experimental conditions, adsorbed CO was the only surface species observed, whereas other reaction intermediates were too short-lived for SFG detection. Nevertheless, the in situ studies revealed a marked difference between nanoscale and macroscopic catalyst surfaces: whereas Pd nanoparticles were partially oxidized during the reaction, Pd(111) remained metallic, again pointing to the need to perform fundamental studies on realistic nanoscale models [7, 73–75]. With the setup described in Sect. 10.3, combined studies of adsorbates (SFG) and catalytic processes (monitoring of the gas phase) could be performed. Figure 10.9 shows the response of a Pt/ZrO2 system during CO oxidation with a starting mixture of 1:2 CO and O2. In Fig. 10.9a spectra of on-top CO at 250 and 350  C in a 1:2 CO and O2 mixture are shown. It is obvious that the wavenumber is red-shifted and the intensity is decreased at higher temperature. A comparison of amplitude and position of the peak is shown in Fig. 10.9b for heating to 350  C and subsequent cooling back to 250  C. The reduced peak height upon heating is due to a lower amount of adsorbed CO on the surface. High

Sum Frequency Generation (SFG) Spectroscopy

a)

1500 1000

223

0.0025 ML Pd Pd4 Pd128

500 0 1500

b)

c)

SFG intensity (cpm)

1000

×3

e)

×3

1900 2000 2100 0.05 ML Pd Pd19

1000

d)

0.8 ×3

×3

1900 2000 2100 0.5 ML Pd Pd145

1900 2000 2100

1000

0.4

×3

×3

500

1900 2000 2100

SFG data No fit Fitted ECO

0.6

0.2 1900 2000 2100 2 ML Pd Pd525

Pd525

Pd145

500 0 1500

Pd19 Pd31

Pd10

1.0

500 0 1500

Pd4

Pd58

1900 2000 2100

Edge site ratio

10

NP model: 1-L 3-L

2-L Half-sphere Nucleation

0.0

0.1

0.01

Data Fit NR CO/alumina Bridge CO/(111) Bridge CO/(111) and (100) Linear CO/edges Linear CO/(111)

Growth regime 1

Pd deposit (ML)

×3

0

1900 2000 2100 1900 2000 2100 IR wavenumber (cm–1)

10

105

350 °C

2085

Amplitude Position

2095 2090

100

2085

95

2080

90

250 °C

2075

85

375

c)

O2 350 CO 325 300 275 CO2

2070 2040

2080

2120

Wavenumber (cm–1)

250 300 350 300 250 Temperature (deg)

Temperature (deg)

2100

b) 110

2077

Peak position (cm–1)

Data Fit

Amplitude (%)

SFG intensity (arb. units)

a)

sites. (e) shows the edge site ratio for theoretical models of Pd nanoparticles, as lines, and the experimentally obtained ratios as open circles. (Adapted with permission from Ref. [66]. Copyright (2019) American Chemical Society)

Mass spectrometer (arb. units)

Fig. 10.8 SFG spectra of CO for different Pd particles with a nominal metal amount of (a) 0.0025 ML, (b) 0.05 ML, (c) 0.5 ML, and (d) 2 ML, with spectra at 108 mbar CO on the left and 1 mbar CO on the right. Orange and pink peaks correspond to bridge CO, blue corresponds to on-top CO on terraces, and green corresponds to on-op CO on edge

0

10

20

30

40

50

250 60

Time (min)

Fig. 10.9 CO oxidation in 1:2 CO to O2 atmosphere on Pt/ZrO2 catalyst between 250  C and 350  C. (a) SFG spectra of on-op CO at starting temperature of 250  C and after heating to 350  C. (b) Evolution of peak amplitude and position during temperature increase to 350  C

and following decrease to 250  C. (c) MS data showing decrease of CO and O2 as well as increase of CO2 depending on measurement temperature. (Adapted from Ref. [76] with permission of AIP Publishing)

CO coverage causes “CO-poisoning”, i.e., blocks the surface sites and hinders dissociative oxygen adsorption, thus a lower CO coverage enables catalytic CO oxidation. In Fig. 10.9c

the gas phase change measured by mass spectrometry shows in agreement with the SFG spectra, that at higher temperatures, with less CO on the surface, the CO oxidation process

224

V. Pramhaas and G. Rupprechter

is much faster than for lower temperatures. SFG allows using this combined approach, so more detailed information on catalytic processes can be gained, observing reaction intermediates.

10.4.2 SFG Spectroscopy on Oxide Surfaces Oxides are a frequently used catalyst support and thus contribute to measurements on metal nanoparticles to some extent. For example, it was found that the oxide electronic surface structure may influence the nonresonant SFG background signal [24]. However, the pristine oxide surfaces exhibit interesting properties as well, for example in photocatalysis, and have been increasingly investigated by SFG spectroscopy. Al2O3 is widely used for various applications and can serve as model for environmentally abundant aluminosilicate surfaces. It has a very specific interaction with water and has therefore been widely studied, particularly its surface

b)

0.14 ML

IVSF /IIR (a.u.)

2.0 1.5 1.0

|(Xr)(2)| 2

2762

2812 ×3

0.4

0.05 ML 2.0 1.5 1.0 2762

2839

|(Xr)(2)| 2

IVSF /IIR (a.u.)

a)

dynamics. The study by Heiden et al. [77] reported SFG spectra of D2O on alumina surfaces and DFT calculations of favorable adsorption sites and dissociative adsorption. Two surfaces with different pretreatment (heating to 300 or 400 K) retained different amounts of adsorbed D2O. SFG measurements of both samples at 130 K, as shown in Fig. 10.10, were fitted under the assumption that both have the same resonances and resonance positions, independent of surface coverage. This yielded three resonances at 2762 cm1, 2812 cm1, and 2839 cm1, respectively. The peaks were then assigned by comparing (relative) peak intensities with DFT calculations. The alumina surface used in calculations is shown in Fig. 10.10c with a topmost layer of oxygen coordinated to two Al atoms (O-μ2, golden) and a second layer of oxygen atoms coordinated to 3 Al atoms (o-μ3, red). Below is an Al layer with coordinatively undersaturated (CUS) Al atoms (gray “a,” bound to one O-μ2 and one O-μ3, black “b,” bound to two O-μ3). While the calculated absolute resonance positions did not match the measured frequencies, the relative shifts between the resonances

×20

0.2

0.2

2812 ×10

0.1

2839 ×30

0.0

0.0 2700

2800

2900

2700

IR frequency (cm–1)

c)

d)

b

2800

2900

IR frequency (cm–1)

e)

O-P2 CUSa O-P CUSb 3

CUSa

CUSb

O-P3 a

Fig. 10.10 SFG spectra after exposing α-Al2O3(112¯0) to D2O at 130 K, reaching either (a) 0.14 ML or (b) 0.05 ML coverage, depending on pretreatment of the sample. Both spectra were fitted with three equally positioned peaks at 2762 cm1, 2812 cm1, and 2839 cm1. DFT calculations were used to assign these resonances with (c) showing

Inter-CUSa||O-m2

CUSb||O-m2

the surface slab used for modeling. (d) and (e) show the two assigned surface adsorption geometries based on the modeling. Color code: O red and gold at the surface, Al black and gray at the surface, H white. (Adapted with permission from [77]. Copyright (2018) American Chemical Society)

Sum Frequency Generation (SFG) Spectroscopy

225

and stability calculations led to an assignment of two adsorption geometries responsible for the observed spectra. The assigned oscillations of the experimental data were two inter-CUSa||O-μ2 at 2762 cm1 and 2812 cm1 and one CUSb||O-μ2 at 2839 cm1. Another important oxide surface, known for its superior performance in photocatalysis, is TiO2. As common for catalytic processes, adsorbate–surface interactions strongly influence reaction pathways and thus activity/selectivity. Vanselous et al. [78] performed 2D SFG spectroscopy on TiO2, which is a dedicated pump-probe technique to detect vibrational crosscoupling, comparing the pump frequency with measured resonance frequencies. A Re(CO)3 complex was deposited on differently structured TiO2 surfaces. Figure 10.11 shows the 2D spectra for nanocrystalline TiO2, Rutile (110) and (001), and Anatase (101). The observed peaks are shown in red (positive signal) and blue (negative signal). Two regions, a symmetric and asymmetric stretch, are visible when

t1

a)

t2 LO t3=0

10

Time

b)

c)

Nanocrystalline

Rutile (110)

Probe frequency (cm–1)

0.2 2050

0.04

2050

2000

0.02

2000

0.00

1950

0.1 0.0

1950

–0.02

–0.1

1900

1900 –0.04

–0.2 1900 1950 2000 2050

1900 1950 2000 2050

d)

e)

Rutile (001)

Anatase (101)

0.2 Probe frequency (cm–1)

Fig. 10.11 Measurement scheme and spectra of the C-O stretching vibrations. (a) shows the pulse sequence for 2D SFG spectroscopy and the combination with the local oscillator (LO). (b–e) show 2D spectra of the Re (CO)3 complex on several TiO2 surfaces for τ2 ¼ 0. Off-diagonal peaks show cross coupling between different vibrational modes. A clear difference in coupling is observed between the nanocrystalline surface (b), where energy is coupled from the symmetric peak (pump with lower wavenumbers) to the asymmetric stretch, and single crystal surfaces (c–e), where the coupling is vice versa. (Adapted with permission from Ref. [78]. Copyright (2018) American Chemical Society)

considering the diagonal for which pump and probe frequencies are the same. Each feature on the diagonal is accompanied by a second resonance with inverse sign due to induced absorbance of the ν ¼ 1 level. Most interesting are the off-diagonal peaks showing the coupling between symmetric and asymmetric stretches. For nanocrystalline material a signal could be observed for the asymmetric stretch after exciting the symmetric stretch, but not vice versa. However, for the Re (CO)3 complex on single crystalline surfaces the opposite trend was observed, with a symmetric response after an asymmetric excitation. This difference hints to differences in the molecular structure resulting from the different adsorption properties of the different surfaces. Theoretical calculations attributed the differences to the specific binding angle of the molecule to the respective surface. The observed binding angle on single crystalline TiO2 matched that reported in previous studies [79, 80].

Intensity

10

2050

1.0 2050

0.1

2000

0.5 2000

0.0

1950

–0.1

1900

0.0

1950

–0.5

1900

–0.2 1900 1950 2000 2050 Pump frequency (cm–1)

–1.0 1900 1950 2000 2050 Pump frequency (cm–1)

226

V. Pramhaas and G. Rupprechter

10.4.3 SFG Spectroscopy on Polymer and Biomaterial Interfaces

hydrogel surfaces [90–92] indicated changes in the molecular orientation of poly(2-hydroxyethyl methacrylate) (pHEMA) in various environments. The complexity was further increased by studying the adsorption of biomolecules on polymers, for example the coverage-dependent surface structures of fibrinogen, lysozyme, and bovine serum albumin (BSA) on deuterated polystyrene, partially deuterated polystyrene, and silica surfaces [93–95]. Olenick et al. [96] investigated lipid bilayers with a phosphatidylcholine (PC) headgroup (e.g., DMPC) due to their importance in various biological processes (e.g., due to their abundance in eukaryotic membranes [97]). Several samples were investigated at different temperatures, and additionally DMPC bilayers supported on fused silica were examined (formed from DMPC/DMPG 1:9 mixtures in either H2O or D2O as buffer solution). For this bilayer (grown from 9:1 mixture of DMPC/DMPG in D2O) they conducted polarization-dependent measurements with orientation analysis as shown in Fig. 10.12. Two peaks can be observed in the SFG spectra in Fig. 10.12a, centered at 2875 cm1 and around 2980 cm1. The peak at 2875 cm1, characteristic of the terminal methyl group, was significantly smaller in the ppp spectrum (blue) than in the ssp. spectrum (orange). By fitting the peaks it was possible to obtain a ppp/ssp amplitude ratio of 0.44  0.01. As shown in Fig. 10.12b, this corresponds to a tilt angle of the methyl group of 15–35 , taking into account the standard deviation and a 1 Gaussian

SFG has been applied to study self-assembled monolayers and polymers on metal, semiconductor, and insulator (solid) surfaces, in gaseous (air) or liquid environments [81–85]. IR or Raman vibrational spectra of polymer films of micrometer thickness just provide bulk spectra whereas SFG selectively detects the conformation of the hydrocarbon chains in the surface layer. Because materials properties (e.g., adhesion or hydrophilicity/hydrophobicity) are often governed by the surface, which may restructure when changing the environment, only SFG provides the relevant information. Zhang et al. [82, 86] and Gracias et al. [87] studied the surface structure of polyethylene and polypropylene. The surface glass transition temperature of polypropylene was found equivalent to the bulk glass transition temperature. Gautam et al. [88] reported studies of the structure and melting transition temperatures of alkyl-side chain acrylate polymers at air and solid interfaces. The SFG polarization dependence suggested a structural rearrangement at the polymer–air but not at the polymer–solid interface. SFG studies were also performed on polymeric biomaterials that are commonly used in bioimplants. Zhang et al. [89] showed that polyurethane with poly(dimethylsiloxane) (PDMS) end groups changed its interfacial structure depending on the environment (e.g., in air and liquids). The examination of

a)

b)

0

15

0.5

80

1 5 10 20 30 40 0.43 0.44 0.45

CH

30

3

0.4

le

45

ng

ta

til

60

c ppp /c ssp

ISFG (a.u.)

0.3

s)

20

ree

eg

(d

40

60 0.2 75

0

0.1

–20

90

0 3200

3100

3000

2900

2800

2700

Wavenumber (cm–1)

Fig. 10.12 Polarization-dependent SFG spectra enabling tilt angle estimation of lipid bi-layer formed from a 9:1 mixture of DMPC/ DMPG in D2O buffer solution. (a) ssp (red/yellow) and ppp (blue) polarized spectra with two resonances. The peak at 2875 cm1 shows a significant difference between the two polarizations and is assigned to the symmetric stretch of CH3. (b) shows the ratio of χ ppp/χ ssp as a

0.1

0.2

0.3 c ppp /c ssp

0.4

0.5

function of the CH3 tilt angle in black. The dashed red line is the estimate from the measured SFG spectra with an error margin of 0.01. The red cones show the overlap of the computed ratio to the measured ratio and estimates a tilt angle around 30 for CH3. (Adapted with permission from the PCCP Owner Societies from Ref. [96]; permission conveyed through Copyright Clearance Center, Inc.)

10

Sum Frequency Generation (SFG) Spectroscopy

monomodal angular distribution. This angle is in good agreement with earlier studies [98]. Self-assembled monolayers (SAMs) were also studied by SFG. Chen et al. [99] studied the “even-odd” effect of SAMs on different surfaces. This effect arises from the different tilt angles of the terminal group, which depends on the odd or even number of carbons. Figure 10.13a, b show spectra of different SAMs on template striped gold (AuTS) with a very flat surface, and as-deposited gold samples (AuAD) with a rougher surface. By comparing the different spectra it is obvious that on the rough surface all resonances have similar intensity and the spectra look almost the same, whereas on the smooth gold surface the SAMs produced significantly different resonances (between 2900 cm1 and 3000 cm1), both for odd and even amounts of carbon in the molecule chains. The respective resonance intensities with respect to chain length are plotted in Fig. 10.13c–e to emphazise the difference. While the evenodd effect on the asymmetric CH3 stretch is striking for the SAMs on the AuTS surface, suggesting a strong variation of the terminal group orientation, this effect was not observed for SAMs on the AuAD. The lack of intensity change for SAMs on the rough surface was explained by a disorder of the molecular chains, as the signal is generated statistically from the terminal group, tilted with respect to the surface normal. For a disordered system the average tilt angle is the same for SAMs of all lengths, as observed on the AuAD surface. In comparison, on AuTS the interface is ordered, causing a significant zigzag in the SFG spectra.

10.4.4 SFG Spectroscopy of Water and Ice Layers Water (containing) surfaces and interfaces play a key role in many physical, chemical, and biological processes [22], why SFG was initially employed to examine O-H stretch vibrations at the water–vapor interface [100, 101]. Polarizationdependent and time-resolved SFG spectra demonstrated that the orientation of the O-H bond varied on a ~1 ps time scale [102–107]. Due to the strong dependence of the SFG intensity on molecular ordering, this technique was soon recognized for its ability to detect the transition of liquid water ordering to ice-like structures. Along these lines, Pandey et al. [108] studied the ice-nucleation activity of different biological samples by investigating structural changes in D2O. The SFG spectra in Fig. 10.14 show the D-O stretching region (2300–2600 cm1) as well as several C-H resonances (above 2800 cm1). For P. syringae bacteria (shown in Fig. 10.14a), which are known to contain ice-nucleation active InaZ proteins, the heavy water signal increased with decreasing temperature. This was especially pronounced in the lower wavenumber region around 2390 cm1, i.e., for

227

strongly hydrogen-bonded water molecules. The increasing signal is directly connected to increasing molecular ordering and alignment of water, which promotes the high temperature ice nucleation. For comparison, reference substances were investigated as shown in Fig. 10.14b–e, such as synthetic misfolded acid of the protein, extract containing denaturated protein, a model DPPG lipid, and lysozyme as a model protein. There is no increase in the water signal for any of these samples, which means that the changing water signal is not simply due to the temperature change or the overall presence of a biological surface. Clearly, it is directly related to the studied bacteria interface. Calculations of water aligned at the interface also matched the observed spectral changes. Thus, SFG spectroscopy supported the hypothesis of ice-nucleation active sites on the P. syringae. Another SFG ice study was reported by Smit and Bakker [109], who investigated the molecular ordering on the ice surface, which has been reported to exhibit high disorder even at temperatures far below the freezing point. A heterodyne-detected SFG setup was used, allowing the direct measurement of the SFG signal’s phase and, thus distinction of the real and imaginary part of the second-order susceptibility. In Fig. 10.15a the imaginary χ (2) is displayed for super-cooled water at 270 K, as well as for ice at 270 K and 254 K, which allows direct identification of the vibrations at the surface. Figure 10.15b shows an equivalent bulk response obtained by combining infrared and Raman spectra. The surface spectra of liquid water and ice at 270 K are very similar, whereas the ice surface spectrum at 245 K is significantly different. For the constructed bulk responses both ice spectra were quite similar, while the liquid water spectrum was significantly different. Thus, for the ice at 270 K the bulk is obviously ice-like, while the surface behaves like super-cooled water, as has been observed previously for multilayer water on Pt(111) [110]. Further analysis of the surface response analyzing the three major components indicates that there is one component that can be ascribed to super-cooled water. Even for the 245 K ice spectrum a significant amount of ordering comparable to super-cooled water can be observed, which can possibly be explained by a disordered surface bilayer.

10.5

Synopsis

Vibrational sum frequency generation spectroscopy, being an inherently interface-specific method, is able to significantly contribute to our understanding of the chemistry and physics at solid–gas, solid–liquid, liquid–gas, and liquid–liquid interfaces. Such studies elucidate surface adsorption and surface reactions relevant for heterogeneous catalysis, electrochemistry, polymer science, biotechnology, etc. SFG can be operated in various gaseous, liquid, and solid environments, and is specifically powerful in detecting species at “buried”

10

228

V. Pramhaas and G. Rupprechter

a) 1.2

b) 0.6

Evens

SFG intensity (a.u.)

0.8 0.6

AuTS

C9 C10 C11 C12 C14 C15 C16

0.5

SFG intensity (a.u.)

C9 C10 C11 C12 C14 C15 C16

1.0

Odds

0.4

0.4 0.3 0.2 0.1

0.2 0.0 2800

2850

2900

2950

3000

0.0 2800

3050

2850

Wavenumber (cm–1)

c)

CH3 Asy CH2 Asy CH3 Sym

AuTS

6

2900

2950

3000

7

CH3 Asy CH2 Asy CH3 Sym

6

5

AuAD

Amplitude

5

4 3

4 3

2

2

1

1

0

0 9

10

11 12 13 14 Number of carbons

15

16

e)

9

f)

10

11 12 13 14 Number of carbons

15

16

6.0 AuTS

4

CH3-asy peak width (cm–1)

5

IAsy /ISym (CH3)

3050

Wavenumber (cm–1)

d)

7

Amplitude

AuAD

AuTS AuAD

3 2

AuAD

5.5

5.0

4.5

1 4.0 0 9

10

11 12 13 14 Number of carbons

15

16

Fig. 10.13 SFG spectra of SAMS with different chain length on smooth (AuTS) and as-deposited rough (AuAD) gold surfaces. (a) and (b) show the ppp polarized spectra of the C-H region on AuTs and AuAD, respectively. While the spectra on AuTS show strong differences for the odd and even length SAM spectra between 2950 and 3000 cm1, the signals on the AuAD surface are all similar. (c) and (d) show the peak

9

10

11 12 13 14 Number of carbons

15

16

amplitude for different chain lengths on AuTS and AuAD, highlighting this trend. (e) shows the comparison of symmetric to asymmetric peak intensity for both samples. (f) shows the different peak width of the asymmetric CH3 stretch for both samples. (Adapted with permission from the PCCP Owner Societies from Ref. [99]; permission conveyed through Copyright Clearance Center, Inc.)

10

Sum Frequency Generation (SFG) Spectroscopy

a) P. syringae Normalized SFG intensity (a.u.)

22 °C 15 °C 10 °C 5 °C

2400

2600

229

b) Fragment

c) Extract

d) DPPG lipid

e) Lysozyme

2800

2400 2600 2800 IR frequency (cm–1)

Fig. 10.14 SFG spectra of D2O on ice-nucleation active bacteria (a) and control substances (b–e) at different temperatures. For the P. syringae bacteria the D2O signal increases at lower temperatures, especially in the low wavenumber region, which is due to increased molecular ordering. For the control substances, the changes in D2O spectra are negligible, showing that the increased ordering of water

Im c (2) (a.u.)

a)

0.5

Surface response

0 Liquid water 270 K Ice 270 K

–0.5

2400

2600

2800

molecules is not simply due to the decreasing temperature, but depends on the specific surface. (Adapted with permission of AAAS from Ref. [108] © The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC) http://creativecommons.org/licenses/by-nc/4.0/)

scanning SFG spectroscopy to polarization-dependent, timeresolved (pump-probe broadband), phase-resolved operation and even locally resolved microscopic modes will certainly lead to fascinating new insights into inorganic, organic, and biochemical interface materials on the nanoscale.

10

–1

IIR · IRaman (a.u.)

b)

References

Ice 245 K

–1.5 Bulk response 1

Ice 245 K Ice 270 K Liquid water 270 K

0.5

0

3000

3200 3400 3600 IR frequency (cm–1)

3800

Fig. 10.15 Vibrational resonances of H2O in the form of a super cooled liquid at 270 K, ice at 270 K, and ice at 245 K. (a) Imaginary part of χ deduced from a heterodyne detected SFG spectrum. As SFG is interface limited, this is a surface response, showing that the surface of both samples at 270 K is similar, while the surface of the ice at 245 K is significantly different. (b) Constructed bulk response from IR and Raman spectroscopy measurements. In the bulk, both ice samples seem similar while the liquid water is distinctly different. (Adapted from Ref. [109] with permission. Copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)

interfaces. The range of materials that can be studied by SFG is meanwhile quite wide, spanning from single crystals via nanostructured surfaces to complex biological surfaces like proteins to technological materials. Extending conventional

1. Schlögl, R.: Sustainable energy systems: the strategic role of chemical energy conversion. Top. Catal. 59(8), 772–786 (2016) 2. Lewis, N.S., Nocera, D.G.: Powering the planet: chemical challenges in solar energy utilization. Proc. Natl. Acad. Sci. 103(43), 15729 (2006) 3. Somorjai, G.A.: Introduction to Surface Chemistry and Catalysis. Wiley, New York (1994) 4. Somorjai, G.A., Rupprechter, G.: The flexible surface: molecular studies explain the extraordinary diversity of surface chemical properties. J. Chem. Educ. 75(2), 161 (1998) 5. Rupprechter, G., Weilach, C.: Spectroscopic studies of surface-gas interactions and catalyst restructuring at ambient pressure: mind the gap! J. Phys. Condens. Matter. 20, 184019 (2008) 6. Haghofer, A., Ferri, D., Föttinger, K., Rupprechter, G.: Who is doing the job? Unraveling the role of Ga2O3 in methanol steam reforming on Pd2Ga/Ga2O3. ACS Catal. 2, 2305–2315 (2012) 7. Rupprechter, G.: Sum frequency generation and polarizationmodulation infrared reflection absorption spectroscopy of functioning model catalysts from ultrahigh vacuum to ambient pressure. Adv. Catal. 51, 133–263 (2007) 8. Shen, Y.R.: Surface properties by second-harmonic and sum-frequency generation. Nature. 337, 519 (1989) 9. Backus, E.H.G., Eichler, A., Kleyn, A.W., Bonn, M.: Real-time observation of molecular motion on a surface. Science. 310(5755), 1790 (2005) 10. Hess, C., Funk, S., Bonn, M., Denzler, D.N., Wolf, M., Ertl, G.: Femtosecond dynamics of chemical reactions at surfaces. Appl. Phys. A Mater. Sci. Process. 71, 477 (2000) 11. Arnolds, H.: Vibrational dynamics of adsorbates – Quo vadis? Prog. Surf. Sci. 86(1), 1–40 (2011)

230 12. Hirose, C., Yamamoto, H., Akamatsu, N., Domen, K.: Orientation analysis by simulation of vibrational sum frequency generation spectrum: CH stretching bands of the methyl group. J. Chem. Phys. 97, 10064 (1993) 13. Bandara, A., Dobashi, S., Kubota, J., Onda, K., Wada, A., Domen, K., Hirose, C., Kano, S.: Adsorption of CO and NO on NiO (111) Ni (111) surface studied by infrared-visible sum frequency generation spectroscopy. Surf. Sci. 387, 312 (1997) 14. Galletto, P., Unterhalt, H., Rupprechter, G.: The molecular orientation of CO on Pd (1 1 1): a polarization-dependent SFG study. Chem. Phys. Lett. 367, 785 (2003) 15. Li, X., Roiaz, M., Pramhaas, V., Rameshan, C., Rupprechter, G.: Polarization-dependent SFG spectroscopy of near ambient pressure CO adsorption on Pt(111) and Pd(111) revisited. Top. Catal. 61, 751–762 (2018) 16. Bain, C.D.: Sum-frequency vibrational spectroscopy of the solid/ liquid interface. J. Chem. Soc. Faraday Trans. 91, 1281 (1995) 17. Hirose, C., Akamatsu, N., Domen, K.: Formulas for the analysis of surface sum-frequency generation spectrum by CH stretching modes of methyl and methylene groups. J. Chem. Phys. 96, 997 (1992) 18. Shen, Y.R.: Surfaces probed by nonlinear optics. Surf. Sci. 299–300, 551 (1994) 19. Buck, M., Himmelhaus, M.: Vibrational spectroscopy of interfaces by infrared–visible sum frequency generation. J. Vac. Sci. Technol. A. 19, 2717 (2001) 20. Williams, C.T., Beattie, D.A.: Probing buried interfaces with non-linear optical spectroscopy. Surf. Sci. 500(1), 545–576 (2002) 21. Richmond, G.L.: Structure and bonding of molecules at aqueous surfaces. Annu. Rev. Phys. Chem. 52, 357 (2001) 22. Eisenthal, K.B.: Liquid interfaces probed by second-harmonic and sum-frequency spectroscopy. Chem. Rev. 96, 1343 (1996) 23. Somorjai, G.A., Rupprechter, G.: Molecular studies of catalytic reactions on crystal surfaces at high pressures and high temperatures by infraredvisible sum frequency generation (SFG) surface vibrational spectroscopy. J. Phys. Chem. B. 103, 1623 (1999) 24. Morkel, M., Unterhalt, H., Klüner, T., Rupprechter, G., Freund, H.J.: Interpreting intensities in vibrational sum frequency generation (SFG) spectroscopy: CO adsorption on Pd surfaces. Surf. Sci. 586, 146 (2005) 25. Peremans, A., Tadjeddine, A., Zheng, W.Q., Le Rille, A., GuyotSionnest, P., Thiry, P.A.: Vibrational dynamics of CO at singlecrystal platinum electrodes in aqueous and non-aqueous electrolytes. Surf. Sci. 368(1), 384–388 (1996) 26. van der Ham, E.W.F., Vrehen, Q.H.F., Eliel, E.R.: Self-dispersive sum-frequency generation at interfaces. Opt. Lett. 21, 1448 (1996) 27. Richter, L.T., Petrallimallow, T.P., Stephenson, J.C.: Vibrationally resolved sum-frequency generation with broad-bandwidth infrared pulses. Opt. Lett. 23, 1594 (1998) 28. Rupprechter, G., Dellwig, T., Unterhalt, H., Freund, H.-J.: CO adsorption on Ni(100) and Pt(111) studied by infrared–visible sum frequency generation spectroscopy: design and application of an SFG-compatible UHV–high-pressure reaction cell. Top. Catal. 15, 19 (2001) 29. Rupprechter, G.: Surface vibrational spectroscopy from ultrahigh vacuum to atmospheric pressure: adsorption and reactions on single crystals and nanoparticle model catalysts monitored by sum frequency generation spectroscopy. Phys. Chem. Chem. Phys. 3, 4621 (2001) 30. Williams, C.T., Yang, Y., Bain, C.D.: Prospects for detecting metal–adsorbate vibrations by sum-frequency spectroscopy. Catal. Lett. 61, 7 (1999) 31. Rupprechter, G., Dellwig, T., Unterhalt, H., Freund, H.-J.: Highpressure carbon monoxide adsorption on Pt (111) revisited: a sum frequency generation study. J. Phys. Chem. B. 105, 3797 (2001)

V. Pramhaas and G. Rupprechter 32. Bonn, M., Hess, C., Funk, S., Miners, J., Persson, B.N.J., Wolf, M., Ertl, G.: Femtosecond surface vibrational spectroscopy of CO adsorbed on Ru (001) during desorption. Phys. Rev. Lett. 84, 4653 (2000) 33. Roeterdink, W.G., Aarts, J.F.M., Kleyn, A.W., Bonn, M.: Broadband sum frequency generation spectroscopy to study surface reaction kinetics: a temperature-programmed study of CO oxidation on Pt (111). J. Phys. Chem. B. 108, 14491 (2004) 34. Cremer, P.S., Su, X., Shen, Y.R., Somorjai, G.A.: High pressure catalytic processes studied by infrared-visible sum frequency generation. J. Mol. Catal. 131, 225 (1998) 35. Guyot-Sionnest, P., Hunt, J.H., Shen, Y.R.: Sum-frequency vibrational spectroscopy of a Langmuir film: study of molecular orientation of a two-dimensional system. Phys. Rev. Lett. 59, 1597 (1987) 36. Baldelli, S., Markovic, N., Ross, P., Shen, Y.R., Somorjai, G.A.: Sum frequency generation of CO on (111) and polycrystalline platinum electrode surfaces: evidence for SFG invisible surface CO. J. Phys. Chem. B. 103, 8920 (1999) 37. Bandara, A., Kubota, J., Onda, K., Wada, A., Kano, S., Domen, K., Hirose, C.: Short-lived reactive intermediate in the decomposition of formate on NiO (111) surface observed by picosecond temperature jump. J. Phys. Chem. B. 102, 5951–5954 (1998) 38. Flörsheimer, M.: Second-harmonic microscopy—new tool for the remote sensing of interfaces. Phys. Status Solidi A. 173, 15 (1999) 39. Flörsheimer, M., Brillert, C., Fuchs, H.: Chemical imaging of interfaces by sum frequency microscopy. Mater. Sci. Eng. C. 8-9, 335 (1999) 40. Kuhnke, K., Hoffmann, D.M.P., Wu, X.C., Bittner, A.M., Kern, K.: Chemical imaging of interfaces by sum-frequency generation microscopy: application to patterned self-assembled monolayers. Appl. Phys. Lett. 83, 3830 (2003) 41. Schaller, R.D., Johnson, J.C., Wilson, K.R., Lee, L.F., Haber, L.H., Saykally, R.J.: Nonlinear chemical imaging nanomicroscopy: from second and third harmonic generation to multiplex (broadbandwidth) sum frequency generation near-field scanning optical microscopy. J. Phys. Chem. B. 106, 5143 (2002) 42. Humbert, B., Grausem, J., Burneau, A., Spajer, M., Tadjeddine, A.: Step towards sum frequency generation spectromicroscopy at a submicronic spatial resolution. Appl. Phys. Lett. 78, 135–137 (2001) 43. Hanninen, A., Shu, M.W., Potma, E.O.: Hyperspectral imaging with laser-scanning sum-frequency generation microscopy. Biomed. Opt. Express. 8(9), 4230–4242 (2017) 44. Fang, M., Baldelli, S.: Surface-induced heterogeneity analysis of an alkanethiol monolayer on microcrystalline copper surface using sum frequency generation imaging microscopy. J. Phys. Chem. C. 121(3), 1591–1601 (2017) 45. Shen, Y.R.: Frontiers in laser spectroscopy. Proc. Int. School Phys. 120, 139–156 (1994) 46. Somorjai, G.A.: Modern surface science and surface technologies: an introduction. Chem. Rev. 96, 1223 (1996) 47. Tadjeddine, A., Peremans, A.: Advances in Spectroscopy, p. 159. Wiley, NewYork (1998) 48. Hall, R.B., Russell, J.N., Miragliotta, J., Rabinowitz, P.R.: In: Vanselow, R., Howe, R. (eds.) Chemistry and Physics of Solid Surfaces Springer Series in Surface Science, p. 87. Springer, Berlin (1990) 49. Somorjai, G.A., McCrea, K.R.: Sum frequency generation: surface vibrational spectroscopy studies of catalytic reactions on metal single-crystal surfaces. Adv. Catal. 45, 385 (2000) 50. Hirose, C., Bandara, A., Katano, S., Kubota, J., Wada, A., Domen, K.: Sum-frequency generation study of reacting species at surfaces. Appl. Phys. B Lasers Opt. 68, 559–565 (1999) 51. Su, X., Cremer, P.S., Shen, Y.R., Somorjai, G.A.: The pressure dependence of the vibrational spectra of adsorbed CO on Pt(111) studied by SFG. Phys. Rev. Lett. 77(18), 3858 (1996)

10

Sum Frequency Generation (SFG) Spectroscopy

52. Cremer, P.S., McIntyre, B.J., Salmeron, M., Shen, Y.R., Somorjai, G.A.: Monitoring surfaces on the molecular level during catalytic reactions at high pressure by sum frequency generation vibrational spectroscopy and scanning tunneling microscopy. Catal. Lett. 34(1/2), 11 (1995) 53. Cremer, P.S., Su, X., Shen, Y.R., Somorjai, G.A.: Ethylene hydrogenation on Pt (111) monitored in situ at high pressures using sum frequency generation. J. Am. Chem. Soc. 118(12), 2942 (1996) 54. Cremer, P.S., Stanners, C., Niemantsverdriet, J.W., Shen, Y.R., Somorjai, G.A.: The conversion of di-σ bonded ethylene to ethylidyne on Pt (111) monitored with sum frequency generation: evidence for an ethylidene (or ethyl) intermediate. Surf. Sci. 328, 111 (1995) 55. Somorjai, G.A.: Molecular level studies of solid-gas and solidliquid interfaces. Surf. Sci. 335(1–3), 10 (1995) 56. Metka, U., Schweitzer, M.G., Volpp, H.R., Wolfrum, J., Warnatz, J.: In-situ detection of NO chemisorbed on platinum using infraredvisible sum-frequency generation (SFG). Z. Phys. Chem. 214, 865 (2000) 57. Unterhalt, H., Rupprechter, G., Freund, H.J.: Vibrational sum frequency spectroscopy on Pd (111) and supported Pd nanoparticles: CO adsorption from ultrahigh vacuum to atmospheric pressure. J. Phys. Chem. B. 106, 356 (2002) 58. Morkel, M., Unterhalt, H., Salmeron, M., Rupprechter, G., Freund, H.-J.: SFG spectroscopy from 108 to 1000 mbar: less-ordered CO structures and coadsorption on Pd(111). Surf. Sci. 532–535, 103–107 (2003) 59. Chen, Z., Gracias, D.H., Somorjai, G.A.: Sum frequency generation (SFG) – surface vibrational spectroscopy studies of buried interfaces: catalytic reaction intermediates on transition metal crystal surfaces at high reactant pressures; polymer surface structures at the solid–gas and solid–liquid interfaces. Appl. Phys. B Lasers Opt. 68, 549 (1999) 60. Hunt, J.H., Guyot-Sionnest, P., Shen, Y.R.: Observation of CH stretch vibrations of monolayers of molecules optical sum-frequency generation. Chem. Phys. Lett. 133, 189 (1987) 61. Ge, A., Rudshteyn, B., Videla, P.E., Miller, C.J., Kubiak, C.P., Batista, V.S., Lian, T.: Heterogenized molecular catalysts: vibrational sum-frequency spectroscopic, electrochemical, and theoretical investigations. Acc. Chem. Res. 52(5), 1289–1300 (2019) 62. Ge, A., Rudshteyn, B., Zhu, J., Maurer, R.J., Batista, V.S., Lian, T.: Electron–hole-pair-induced vibrational energy relaxation of rhenium catalysts on gold surfaces. J. Phys. Chem. Lett. 9(2), 406–412 (2018) 63. Dellwig, T., Rupprechter, G., Unterhalt, H., Freund, H.-J.: Bridging the pressure and materials gaps: high pressure sum frequency generation study on supported Pd nanoparticles. Phys. Rev. Lett. 85, 776 (2000) 64. Baldelli, S., Eppler, A.S., Anderson, E., Shen, Y.R., Somorjai, G.A.: Surface enhanced sum frequency generation of carbon monoxide adsorbed on platinum nanoparticle arrays. J. Chem. Phys. 113, 5432 (2000) 65. Rupprechter, G.: Sum frequency laser spectroscopy during chemical reactions on surfaces. MRS Bull. 32(12), 1031–1037 (2007) 66. Alyabyeva, N., Ouvrard, A., Zakaria, A.-M., Bourguignon, B.: Probing nanoparticle geometry down to subnanometer size: the benefits of vibrational spectroscopy. J. Phys. Chem. Lett. 10(3), 624–629 (2019) 67. Ouvrard, A., Ghalgaoui, A., Michel, C., Barth, C., Wang, J., Carrez, S., Zheng, W., Henry, C.R., Bourguignon, B.: CO chemisorption on ultrathin MgO-supported palladium nanoparticles. J. Phys. Chem. C. 121(10), 5551–5564 (2017) 68. Yudanov, I.V., Sahnoun, R., Neyman, K.M., Rösch, N., Hoffmann, J., Schauermann, S., Johánek, V., Unterhalt, H., Rupprechter, G., Libuda, L., Freund, H.-J.: CO adsorption on Pd nanoparticles:

231 density functional and vibrational spectroscopy studies. J. Phys. Chem. B. 107, 255 (2003) 69. Morkel, M., Rupprechter, G., Freund, H.J.: Ultrahigh vacuum and high-pressure coadsorption of CO and H2 on Pd(111): a combined SFG, TDS, and LEED study. J. Chem. Phys. 119, 10853 (2003) 70. Morkel, M., Rupprechter, G., Freund, H.-J.: Finite size effects on supported Pd nanoparticles: interaction of hydrogen with CO and C2H4. Surf. Sci. 588, 209 (2005) 71. Borasio, M., Rodríguez de la Fuente, O., Rupprechter, G., Freund, H.-J.: In situ studies of methanol decomposition and oxidation on Pd(111) by PM-IRAS and XPS spectroscopy. J. Phys. Chem. B Lett. 109, 17791 (2005) 72. Borasio, M.: Polarization modulation infrared reflection absorption spectroscopy on Pd model catalysts at elevated pressure, PhD thesis, Free University Berlin (2006) 73. Bäumer, M., Libuda, J., Neyman, K.M., Rösch, N., Rupprechter, G., Freund, H.-J.: Adsorption and reaction of methanol on supported palladium catalysts: microscopic-level studies from ultrahigh vacuum to ambient pressure conditions. Phys. Chem. Chem. Phys. 9, 3541–3558 (2007) 74. Rupprechter, G.: A surface science approach to ambient pressure catalytic reactions. Catal. Today. 126, 3 (2007) 75. Rupprechter, G., Weilach, C.: Vibrational studies of surface-gas interactions at ambient pressure. J. Phys. Condens. Matter. 20, 184020 (2008) 76. Roiaz, M., Pramhaas, V., Li, X., Rameshan, C., Rupprechter, G.: Atmospheric pressure reaction cell for operando sum frequency generation spectroscopy of ultrahigh vacuum grown model catalysts. Rev. Sci. Instrum. 89, 045104-(045101-045110) (2018) 77. Heiden, S., Yue, Y., Kirsch, H., Wirth, J., Saalfrank, P., Campen, R.K.: Water dissociative adsorption on α-Al2O3(112¯0) is controlled by surface site undercoordination, density, and topology. J. Phys. Chem. C. 122(12), 6573–6584 (2018) 78. Vanselous, H., Videla, P.E., Batista, V.S., Petersen, P.B.: Distinct binding of rhenium catalysts on nanostructured and singlecrystalline TiO2 surfaces revealed by two-dimensional sum frequency generation spectroscopy. J. Phys. Chem. C. 122(45), 26018–26031 (2018) 79. Anfuso, C.L., Snoeberger, R.C., Ricks, A.M., Liu, W., Xiao, D., Batista, V.S., Lian, T.: Covalent attachment of a rhenium bipyridyl CO2 reduction catalyst to rutile TiO2. J. Am. Chem. Soc. 133(18), 6922–6925 (2011) 80. Ge, A., Rudshteyn, B., Psciuk, B.T., Xiao, D., Song, J., Anfuso, C.L., Ricks, A.M., Batista, V.S., Lian, T.: Surface-induced anisotropic binding of a rhenium CO2-reduction catalyst on rutile surfaces. J. Phys. Chem. C. 120(37), TiO2(110) 20970–20977 (2016) 81. Chen, P., Westerberg, S., Kung, K.Y., Zhu, J., Grunes, J., Somorjai, G.A.: CO poisoning of catalytic ethylene hydrogenation on the Pt (1 1 1) surface studied by surface sum frequency generation. Appl. Catal. A Gen. 229, 147 (2002) 82. Zhang, D., Ward, R.S., Shen, Y.R., Somorjai, G.A.: Environmentinduced surface structural changes of a polymer - an in situ IR VIS SFG study. J. Phys. Chem. B. 101, 9060 (1997) 83. Wang, J., Woodcock, S.E., Buck, S.M., Chen, C.Y., Chen, Z.: Different surface-restructuring behaviors of poly (methacrylate) s detected by SFG in water. J. Am. Chem. Soc. 123, 9470 (2001) 84. Wang, J., Chen, C.Y., Buck, S.M., Chen, Z.: Molecular chemical structure on poly (methyl methacrylate)(PMMA) surface studied by sum frequency generation (SFG) vibrational spectroscopy. J. Phys. Chem. B. 105, 12118 (2001) 85. Kim, J., Koffas, T.S., Lawrence, C.C., Somorjai, G.A.: Surface structural characterization of protein- and polymer-modified polystyrene microspheres by infraredvisible sum frequency

10

232 generation vibrational spectroscopy and scanning force microscopy. Langmuir. 20, 4640 (2004) 86. Zhang, D., Shen, Y.R., Somorjai, G.A.: Studies of surface structures and compositions of polyethylene and polypropylene by IR plus visible SFG vibrational spectroscopy. Chem. Phys. Lett. 281, 394 (1997) 87. Gracias, D.H., Zhang, D., Lianos, L., Ibach, W., Shen, Y.R., Somorjai, G.A.: A study of the glass transition of polypropylene surfaces by sum-frequency vibrational spectroscopy and scanning force microscopy. Chem. Phys. Lett. 245, 277 (1999) 88. Gautam, K.S., Dhinojwala, A.: Melting at alkyl side chain comb polymer interfaces. Phys. Rev. Lett. 88, 145501 (2002) 89. Zhang, D., Gracias, D.H., Ward, R., Gauckler, M., Tian, Y., Shen, Y.R., Somorjai, G.A.: Surface studies of polymer blends by sum frequency vibrational spectroscopy, atomic force microscopy, and contact angle goniometry. J. Phys. Chem. B. 102, 6225 (1998) 90. Chen, Q., Zhang, D., Somorjai, G.A., Bertozzi, C.: Probing the surface structural rearrangement of hydrogels by sum-frequency generation spectroscopy. J. Am. Chem. Soc. 121, 446 (1999) 91. Kweskin, S.J., Komvopoulos, K., Somorjai, G.A.: Entropically mediated polyolefin blend segregation at buried sapphire and air interfaces investigated by infraredvisible sum frequency generation vibrational spectroscopy. J. Phys. Chem. B. 109, 23415 (2005) 92. Kweskin, S.J., Komvopoulos, K., Somorjai, G.A.: Conformational changes at polymer gel interfaces upon saturation with various liquids studied by infrared-visible sum frequency generation vibrational spectroscopy. Appl. Phys. Lett. 88, 134105 (2006) 93. Mermut, O., Phillips, D.C., York, R.L., McCrea, K.R., Ward, R.S., Somorjai, G.A.: In situ adsorption studies of a 14-amino acid leucine-lysine peptide onto hydrophobic polystyrene and hydrophilic silica surfaces using quartz crystal microbalance, atomic force microscopy, and sum frequency generation vibrational spectroscopy. J. Am. Chem. Soc. 128, 3598 (2006) 94. Kim, J., Somorjai, G.A.: Molecular packing of lysozyme, fibrinogen, and bovine serum albumin on hydrophilic and hydrophobic surfaces studied by infraredvisible sum frequency generation and fluorescence microscopy. J. Am. Chem. Soc. 125, 3150 (2003) 95. Amitay-Sadovsky, E., Komvopoulos, K., Tian, Y., Somorjai, G.A.: Correlation of surface molecular composition to nanoscale elastic behavior and topography of stretched polyurethane films. Appl. Phys. Lett. 80, 1829 (2002) 96. Olenick, L.L., Chase, H.M., Fu, L., Zhang, Y., McGeachy, A.C., Dogangun, M., Walter, S.R., Wang, H.-f., Geiger, F.M.: Singlecomponent supported lipid bilayers probed using broadband nonlinear optics. Phys. Chem. Chem. Phys. 20(5), 3063–3072 (2018)

V. Pramhaas and G. Rupprechter 97. van Meer, G., Voelker, D.R., Feigenson, G.W.: Membrane lipids: where they are and how they behave. Nat. Rev. Mol. Cell Biol. 9(2), 112–124 (2008) 98. Liu, J., Conboy, J.C.: Structure of a gel phase lipid bilayer prepared by the LangmuirBlodgett/Langmuir-Schaefer method characterized by sum-frequency vibrational spectroscopy. Langmuir. 21(20), 9091–9097 (2005) 99. Chen, J., Liu, J., Tevis, I.D., Andino, R.S., Miller, C.M., Ziegler, L.D., Chen, X., Thuo, M.M.: Spectroscopic evidence for the origin of odd–even effects in self-assembled monolayers and effects of substrate roughness. Phys. Chem. Chem. Phys. 19(10), 6989–6995 (2017) 100. Du, Q., Superfine, R., Freysz, E., Shen, Y.R.: Vibrational spectroscopy of water at the vapor/water interface. Phys. Rev. Lett. 70, 2313 (1993) 101. Du, Q., Freysz, E., Shen, Y.R.: Surface vibrational spectroscopic studies of hydrogen bonding and hydrophobicity. Science. 264, 826 (1994) 102. Wei, X., Shen, Y.R.: Motional effect in surface sum-frequency vibrational spectroscopy. Phys. Rev. Lett. 86, 4799 (2001) 103. Simonelli, D., Baldelli, S., Schultz, M.J.: Ammonia–water complexes on the surface of aqueous solutions observed with sum frequency generation. Chem. Phys. Lett. 298, 400 (1998) 104. Schnitzer, C., Baldelli, S., Shultz, M.J.: Sum frequency generation by water on supercooled H2SO4/H2O liquid solutions at stratospheric temperature. Chem. Phys. Lett. 313, 416 (1999) 105. Schnitzer, C., Baldelli, S., Shultz, M.J.: Sum frequency generation of water on NaCl, NaNO3, KHSO4, HCl, HNO3, and H2SO4 aqueous solutions. J. Phys. Chem. B. 104, 585 (2000) 106. Brown, M.G., Raymond, E.A., Allen, H.C., Scatena, L.F., Richmond, G.L.: The analysis of interference effects in the sum frequency spectra of water interfaces. J. Phys. Chem. A. 104, 10220 (2000) 107. Allen, H.C., Raymond, E.A., Richmond, G.L.: Surface structural studies of methanesulfonic acid at air/aqueous solution interfaces using vibrational sum frequency spectroscopy. J. Phys. Chem. A. 105, 1649 (2001) 108. Pandey, R., Usui, K., Livingstone, R.A., Fischer, S.A., Pfaendtner, J., Backus, E.H.G., Nagata, Y., Fröhlich-Nowoisky, J., Schmüser, L., Mauri, S., Scheel, J.F., Knopf, D.A., Pöschl, U., Bonn, M., Weidner, T.: Ice-nucleating bacteria control the order and dynamics of interfacial water. Sci. Adv. 2(4), e1501630 (2016) 109. Smit, W.J., Bakker, H.J.: The surface of ice is like supercooled liquid water. Angew. Chem. Int. Ed. 56(49), 15540–15544 (2017) 110. Starke, U., Materer, N., Barbieri, A., Döll, R., Heinz, K., Van Hove, M.A., Somorjai, G.A.: A low-energy electron diffraction study of oxygen, water and ice adsorption on Pt(111). Surf. Sci. 287–288, 432–437 (1993)

10

Sum Frequency Generation (SFG) Spectroscopy

Verena Pramhaas received her doctoral degree in physical chemistry at TU Wien working on SFG spectroscopy. She was associated student of the interdisciplinary doctoral school “Solids for Function” funded by the Austrian Science Fund (FWF). Her master’s degree in quantum optics at the University of Innsbruck gave her deep insight into laser physics and the specifics of complex spectroscopy. She has been postdoctoral fellow at TU Wien from 2019–2021, focusing on research in model catalysis as well as teaching, before moving to industry.

233

Günther Rupprechter received his Ph.D. in Physical Chemistry in 1996 from the University Innsbruck, Austria. After being a postdoctoral fellow at the University of California at Berkeley and the Lawrence Berkeley National Lab (LBNL) till 1998 (with Gabor A. Somorjai), he was group leader for Laser Spectroscopy and Catalysis at the Fritz Haber Institute of the Max Planck Society in Berlin (Germany) from 1999 to 2005 (with Hajo Freund). In 2005 he accepted a Full Professorship in Surface and Interface Chemistry at Technische Universität Wien. His research focuses on heterogeneous catalysis, particularly in situ (operando) spectroscopy/microscopy on model and technological catalysts (about 260 articles). In 2005 he received the Jochen Block Award of the German Catalysis Society for “the application of surface science methods to heterogeneous catalysis” and became corresponding member of the Austrian Academy of Sciences in 2012. He was also appointed “Renowned Overseas Professor” of Shanghai University of Engineering Science.

10

Part II Electron and Photoelectron Spectroscopy

Ultraviolet-Visible (UV-Vis) Spectroscopy Charlotte Vogt , Caterina Suzanna Wondergem Bert M. Weckhuysen

11

, and

Contents 11.1

Basic Principles of Ultraviolet-Visible Spectroscopy . . . 238

11.2 The Spectrometer and Related Accessories . . . . . . . . . . . 240 11.2.1 The UV-Vis Spectrometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240 11.2.2 Sample Preparation, Mode of Measuring, and Catalytic Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 11.3

Probe Molecule UV-Vis Spectroscopy . . . . . . . . . . . . . . . . . . 246

11.4

Coupling UV-Vis Spectroscopy with Other Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

11.5

Complementing Data Interpretation with Density Functional Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

11.6

Application of Chemometrics and Multivariate Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

11.7 11.7.1 11.7.2 11.7.3 11.7.4

Selected Applications of UV-Vis Spectroscopy in the Field of Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Heterogeneous Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Homogeneous Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11.8

Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

254 254 256 259 259

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

Abstract

Ultraviolet-Visible (UV-Vis) spectroscopy is a versatile and powerful analytical method, which allows to C. Vogt Schulich Faculty of Chemistry, Technion, Israel Institute of Technology, Haifa, Israel e-mail: [email protected] C. S. Wondergem Research Center for Materials Science, Graduate School of Science, Nagoya University, Nagoya, Japan e-mail: [email protected] B. M. Weckhuysen (*) Inorganic Chemistry and Catalysis, Debye Institute for Nanomaterials Science, Utrecht University, CG Utrecht, The Netherlands e-mail: [email protected]

investigate a wide variety of catalysts in both the liquidphase and solid-state and their interfaces at elevated temperatures and pressures. In the case of solid catalysts, they can be studied in the form of powders (e.g., in diffuse reflectance mode) and as thin wafers (in transmission mode), and when combined with a microscope even in the form of catalyst bodies (e.g., extrudates) and single crystals. In the past two decades, UV-Vis spectroscopy has been increasingly used under in situ and operando conditions to shed light on/gain insight in the working principles of heterogeneous catalysts, homogeneous catalysts, electrocatalysts, as well as photocatalysts. One of the advantages of this method is that it can simultaneously measure, e.g., the electronic transitions of organic molecules (mainly via their n ! π* and π ! π* transitions) and transition metal oxides or ions (via their d-d and charge transfer transitions). Unfortunately, absorption bands in the UV-Vis range are often broad and overlapping and hence their interpretations are not always trivial. Advanced theoretical calculations are required to provide a proper foundation of their interpretation, while, e.g., chemometrics can help prevent biased analysis when many (time-resolved) spectra are collected. Finally, UV-Vis spectroscopy is often combined with other analytical methods to provide complementary information. Examples include X-ray absorption spectroscopy and diffraction, next to vibrational spectroscopy (i.e., infrared and Raman) and magnetic resonance (i.e., electron spin resonance and nuclear magnetic resonance) methods. The above-described scientific and instrumental developments will be illustrated by using a selection of showcase examples, covering the different areas of catalysis. The chapter concludes with some main observations as well as some future developments on what might become possible in the not-too-distant future.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_11

237

238

C. Vogt et al.

Keywords

UV-Vis spectroscopy · Catalysis · Operando characterization · In-situ spectroscopy

11.1

Basic Principles of Ultraviolet-Visible Spectroscopy

Ultraviolet-Visible spectroscopy makes use of radiation in the Ultraviolet (UV) and Visible (Vis) radiation regions of the electromagnetic spectrum from approximately 200–800 nm (50,000–12,500 cm1) [1]. Often the spectroscopic method extends into the near-infrared (NIR) region and is then referred to as UV-Vis-NIR spectroscopy, with a spectral range up to, e.g., 2000 nm (i.e., down to 5000 cm1). The UV range constitutes the first 200–400 nm of the spectrum, followed by the Vis region (400–800 nm) and finally the NIR region (>800 nm). In UV-Vis-NIR spectroscopy, the interaction of UV, Vis, and NIR radiation photons with the sample is measured. Since this is partly the same region as where human eyes detect color, the absorption, scattering, or reflectance of UV-Vis-NIR radiation and thus the suitability of the characterization technique for a particular sample can (roughly) be qualitatively predicted by its color. Completely black samples, as is the case for solid catalysts containing coke deposits, are often not suitable for UV-Vis-NIR spectroscopy as they may absorb all the incident photons, whereas completely white or transparent samples are not likely to be suitable either, as they may reflect, scatter, or transmit (respectively) all incident radiation depending on, e.g., the dimensions of the grains or crystals constituting the catalyst material. The latter can be the case when investigating (highvalency) transition metal oxides (e.g., Re2O7) supported on a silica support. Samples of which color changes occur during reaction (e.g., CrO3 ! Cr2O3) or those that are vibrantly colored (e.g., a solid containing an organic dye molecule as can be formed/synthesized/produced within the channels of a zeolite-based catalyst) are generally particularly suitable for UV-Vis-NIR spectroscopy. UV-Vis-NIR spectroscopy is often called electronic spectroscopy because electrons are transferred from low-energy into high-energy atomic or molecular orbitals when the material is irradiated with electromagnetic radiation [2]. The σ ➔ σ* transitions in organic molecules (e.g., originating from a C-H or C-C chromophore) are generally high-energy transitions with corresponding short wavelength (i.e., λ often below 200 nm). These transitions are generally not observed with UV-Vis-NIR spectroscopy unless vacuum is applied and/or synchrotron-based UV-radiation (and related detection) is employed because such transitions will also be probed in the molecules that make up air (e.g., O2), and vacuum or higher photon flux from synchrotron radiation are ways to

eliminate this issue. A similar situation exists for n ➔ σ* transitions (e.g., NH2 chromophore), which occur at ~150–250 nm. On the other hand, n ➔ π* (e.g., C¼O chromophore) and π ➔ π* (e.g., C¼C chromophore) transitions occur with incident radiation of approximately 200–700 nm and show weak and intense absorption responses, respectively. As such, many organic molecules containing (conjugated) π systems, as is the case for aromatic compounds, as well as metal-organic species and transition metal complexes (which possess characteristic d-d transitions or charge transfer (CT)), are species excellent for detection with UV-Vis-NIR spectroscopy. The NIR addition to UV-Vis spectrometers is especially useful as it allows to detect the combination and overtone bands of characteristic vibrational bands of, e.g., adsorbed organic molecules (e.g., aromatics) or the terminal groups of metal oxides (e.g., silanol groups of a SiO2 support). Furthermore, the more delocalized the electrons are in a system (i.e., the more conjugated the bonds of a molecule are in, e.g., an aromatic species), the lower the energy required to excite an electron from its ground state. Consequently, the absorption peak for similar molecules with increasing degree of conjugation will be found at increasingly higher wavelengths in the UV-Vis absorption spectrum. Additional background on the principles, theory, and applications of UV-Vis-NIR spectroscopy in the field of catalysis can be found in several book chapters and review articles. As becomes clear from Table 11.1, the application of UVVis-NIR spectroscopy in catalysis is particularly relevant to study transition metal ion (TMI) centers (in particular 3d-ions), rare earth metal ions (especially lanthanides), adsorbed molecules, molecular ions and (organic) radicals, and catalyst supports (which can be amorphous metal oxides with a high surface area or porous crystalline materials, such as zeolites, metal organic frameworks (MOFs) and mesoporous materials). In the case of 3d-ions, detailed UV-VisNIR characterization studies can be performed on the low-valent transition metal cations (i.e., Cu2+, Cr3+, Co2+, and Ni2+) revealing their coordination environment in the Table 11.1 Overview of the possible transitions that can be probed with UV-Vis-NIR spectroscopy and are relevant to the characterization of catalysts Species Transition metal ion

Rare earth ions Molecules, radicals, molecular ions Supports (oxide) Supports, inorganic and organic species Multinuclear mixedvalence species

Relevant transitions d-d transitions, metal-to-ligand charge transfer (MLCT) transitions, ligand-tometal charge transfer (LMCT) transitions f-f and f-d transitions n ➔ π* and π ➔ π* transitions Band gap, impurities, defects Overtone and combination bands of vibrations in the infrared region Intervalence charge transfer (IVCT)

11

Ultraviolet-Visible (UV-Vis) Spectroscopy

239

presence and absence of reactants, while for high-valent transition metal anions (i.e., Cr6+, V5+, Mo6+, and Re7+) the redox behavior and nuclearity (e.g., chromate, dichromate, and polychromate) can be investigated as a function of, e.g., support oxide and reduction conditions. Examples of the former can be found in the work of the groups of, e.g., Schoonheydt [3], Che [4], Zecchina [5], and Wichterlova [6], while the latter has been performed by a.o. the groups of Wachs [7], and Bell/Iglesia [8]. The earlier works of Che, Wichterlova, and Schoonheydt on first row transition metal ions, such as Cu2+/+ and Ni2+/+, already illustrated the strong complementarity of UV-Vis-NIR spectroscopy with another powerful analytical method, namely, Electron Spin Resonance (ESR) which also measures ground state properties of transition metal ions (e.g., V4+ (d1) and Cu2+ (d9). This has been nicely summarized by Sojka et al. [9]. Two theories that are highly relevant to explain the effects observable by UV-Vis-NIR spectroscopy are Crystal Field Theory (CFT) and Molecular Orbital Theory (MOT) [10]. CFT describes the breaking of degeneracies of electron orbital states, while MOT uses quantum mechanics to describe the electronic structure of molecules. These theories describe the origins and the consequences of interactions of the surroundings on the orbital energy levels of transition metal ions and conjugated π-system, among others. Such interactions are considered electronic even if there is no net charge, as lone pairs of electrons are interchanged. When considering a transition metal ion, molecular ion, or radical, and its interaction with an adsorbate, such adsorbates are considered point negative charges surrounding the ionic radius of the ion or radical under consideration. In this case the corresponding energy is the crystal field potential (Eq. 11.1), where Rj is the distance of the ion-adsorbate j, ri is the distance between the ion-electron i, and n is the number of adsorbates. V ðr i Þ ¼

n X j¼1

e2 j Rj  r i j

ð11:1Þ

For m electrons the crystal field potential is given as Eq. 11.2: V CF ¼

m X

V ðr i Þ

ð11:2Þ

i¼1

The crystal field potential causes destabilization of d-orbitals (to differing extents, based on their orientation). Depending on the applied crystal field geometry (e.g., octahedral coordination) and the (non-)axial nature of the orbitals, splitting of d-orbital energies into relatively lower and higher energies can occur. This separation is then given as Δ0 and suggests that the electrons that occupy the

d-orbitals can rearrange. An electron can be excited from the low energy d-orbitals to the high energy ones by absorption of a photon with the corresponding energy. This energy is represented by ΔE ¼ hν ¼ Δ0. Such rearrangements form the basis of d-d spectra (crystal field spectra), and a large part of electronic absorption spectra. However, electronic absorption spectra also include charge transfer complexes, and interelectronic transitions. Charge transfer complexes involve the association of two or more molecules in which a fraction of electronic charge is transferred between the molecular entities. The resulting electrostatic attraction provides a stabilizing force for the molecular complex. The combination of CFT and MOT led to the more realistic and complex ligand field theory (LFT), which can give insight into the chemical bonding of (supported) transition metal ion complexes. A recent literature example can be found in the elucidation of the redox properties of Mn-promoted sulfated ZrO2, a highly active alkane isomerization catalyst, by in situ UV-Vis spectroscopy and CFT [11]. More specifically, a cluster model that considers the Mn center as a complex with the adjacent ions of the lattice as ligands was developed and used to determine the Mn3+/Mn4+ ratio in the catalyst based on temperature-dependent UV-Vis spectroscopic data. The combination of such theories with density functional theory (DFT) has allowed for accurate modeling of UV-Vis-NIR spectra of model complexes, a tool now often used in literature to build, e.g., libraries providing the ability to relate specific species to convoluted spectra or real system, thereby increasingly approaching realistic reaction environments (i.e., the operando approach). This aspect will be later on illustrated with some showcases in the field of zeolite-based catalysis and olefin polymerization catalysis. There are several excellent reference works describing the principles and applications of UV-Vis-NIR spectroscopy in the broad field of catalysis. To the best of our knowledge, the last review paper has been written by Jentoft in 2012, which appeared in a textbook edited by Che and Védrine [12], but also noteworthy are the 2010 Chemical Society Reviews paper by Schoonheydt [13], the 2009 Advances in Catalysis paper by Jentoft [14] and the older reviews by Schoonheydt [15], Kellerman [16], Klier [17], as well as the textbook of Kortüm [18]. Our own review articles date back to one 1999 Catalysis Today paper [19] and two book chapters in 2000 [20] and 2004 [21], in which the state-of-the-art of in situ and operando UV-Vis-NIR diffuse reflectance spectroscopy of solid catalysts has been summarized. Hence, this chapter approximately summarizes the tremendous progress made in this field of research the past 20 years. In this time, the instrumental capabilities have increased dramatically, offering more possibilities currently than we could have imagined 20 years ago. Among these new possibilities is the branching out of UV-VIS-NIR spectroscopy to other application

11

240

C. Vogt et al.

domains, including photocatalysis and electrocatalysis. The search for complementarity with other analytical methods is now well-established as no single technique can provide the ultimate answer to a fundamental catalytic problem. Furthermore, UV-Vis-NIR spectroscopy is increasingly used under relevant reaction conditions, including the verification that the measurement cell indeed performs (to a certain degree) as a catalytic reactor (i.e., the operando measurement approach [22]).

11.2

The Spectrometer and Related Accessories

11.2.1 The UV-Vis Spectrometer A UV-Vis spectrometer can be a relatively inexpensive piece of laboratory equipment and generally has the following basic components; a source of UV-Vis radiation, a monochromator which ensures the correct wavelength of radiation illuminates the sample, a sample holder, and a detector. In a typical spectrometer for liquid-phase measurements, these components are placed on a single axis, as shown in Fig. 11.1. The monochromator gradually changes its energy Fig. 11.1 (a) Schematic representation of the basic components of a UV-Vis spectrometer, including a proper radiation source, a wavelength selector, sample holder, and detector; and (b) The difference between specular reflected radiation (radiation reflected in a mirror-like fashion, at a definite angle) and diffuse reflected radiation (the reflection of radiation from a surface such that an incident ray is reflected at many angles), which forms the basis of UV-Vis spectroscopy when working with, solid catalysts

along a given set of energies in the region of interest, and the photons that reach the detector are counted. Because the amount of photons output by the source (I0) is known, the transmission of the sample can be calculated (the measured intensity through the sample (I ) divided by the intensity of the source, I/I0). In a classical UV-Vis spectroscopy measurement, often a cuvette with a precise path length is filled with the sample, which can be a solution containing solvent, reactant, and a homogeneous catalyst. The amount of radiation that is absorbed by the sample depends on its concentration, the path length of the radiation through the cuvette, and how much radiation the sample absorbs at a certain wavelength. The absorbance can be calculated by taking the natural logarithm of the transmittance (T ) (Eq. 11.3), and this is in turn related to the concentration of the sample via LambertBeer’s law (Eq. 11.4). A ¼  log 10ðT Þ

ð11:3Þ

A ¼ εlc

ð11:4Þ

where A is absorbance (unitless); ε is the molar absorption coefficient, or molar absorption constant of the sample for a

a)

l I Exit slit

Detector

I0

Dispersion Entrance slit

Sample

Light source Monochromator

b) Incident light

Specular reflected light

Diffuse reflected light

11

Ultraviolet-Visible (UV-Vis) Spectroscopy

certain wavelength (Lmol1cm1); l is path length (cm) through the cuvette; and c is the concentration of the sample (molL1). In other words, for simple solution-based experiments quantitative UV-Vis experiments can be performed, allowing to elucidate the change of a coordination complex of a transition metal ion as a function of reaction time. UV-Vis spectroscopy is a fast and simple method with which one can determine the concentration of the sample or molecule, being it either a reactant, reaction, intermediate, or reaction product under investigation. It is a versatile method that can be applied to gases, liquids, and solids. UV-Vis spectroscopy is extremely sensitive to concentrations and is also a noninvasive technique, which make it possible to perform in situ measurements for, e.g., chemical reactor control. Downsides of UV-Vis spectroscopy are the fact that the technique is diffraction-limited and hence the spatial resolution is limited to ~0.5 μm when combined with a microscope, and that the signals that are obtained are generally broad and convoluted, meaning that several components have overlapping features. Furthermore, for characterization of solid catalysts, it is often difficult to obtain UV-Vis transparent samples (i.e., thin-films of powdered samples are difficult to obtain), which limits the application of, e.g., transmission UV-Vis spectroscopy. There are only a few examples of zeolites being pressed into self-supported catalyst wafers allowing for the UV-Vis transmission study of solid catalysts (in use) [23–25]. The limiting factor is that much of the radiation does not reach the detector due to scattering. To that end, diffuse reflectance spectroscopy (DRS) can be applied, where the radiation that is diffusely reflected off the sample is collected by convergent mirrors or an integration sphere. In what follows, we will discuss the different modes of measuring UV-Vis-NIR spectroscopy, including fiber optics UV-Vis spectroscopy, but we refer to the earlier cited book chapters and review articles for a more comprehensive overview of the different cell designs and measurement options available [12–21].

11.2.2 Sample Preparation, Mode of Measuring, and Catalytic Reactors The typical undergraduate degree laboratory introduction to UV-Vis spectroscopy is to measure calibration lines in a cuvette for a solution with a specific concentration of an analyte [1, 2]. While surely interesting as an analytical technique in that sense, the cuvette has little relevance for the application of UV-Vis spectroscopy for studying chemical reactions taking place in the presence of, e.g., a homogeneous or heterogeneous catalyst. For most catalytic applications, requirements of an experimental setup for in situ or operando UV-Vis spectroscopy are – next to a suitable spectrometer – a methodology apt to measure with sufficient signal-to-noise

241

ratio and operable at reaction conditions of elevated temperatures and/or pressures. By doing so, a setup is used that mimics to a certain extent the reactor type (e.g., plug flow) of the catalytic reactor when there is no spectroscopy involved. The latter aspect is often evaluated by attaching a residual gas analysis device to the operando setup, such as a mass spectrometer (MS) or a gas chromatograph (GC), in order to ensure that product formation is as expected, and true structure-reactivity correlations can be made [22]. The operando UV-Vis spectroscopy approach (employing online GC analysis) in the field of heterogeneous catalysis was already practiced at the end of the 1990s [26], and at the start of the 2000s [27, 28], although it was not yet coined that way, and operando UV-Vis spectroscopy formally entered the scientific literature in 2003 when two combined setups (UV-Vis/Raman spectroscopy [29] and ESR/UV-Vis spectroscopy [30]) were first discussed at the first Congress on Operando Spectroscopy (Lunteren, the Netherlands). As mentioned above, it can be difficult to prepare heterogenous catalytic samples in particular, samples that are UV-Vis transparent, and while it has been proven to be advantageous to measure in transmission in some cases [31], more often than not high absorption and scattering limit the application of transmission-mode UV-Vis spectroscopy for heterogeneous catalysis. As such, there are three main modes of operation for UV-Vis spectroscopy that have become highly relevant to catalysis research. 1. Diffuse reflectance UV-Vis (DRS, or DR-UV-VIS) spectroscopy, either by use of a Praying Mantis™-type accessory or (to a lesser extent) by use of a classical diffuse reflectance integrating sphere, which is typically coated with materials that reflect more than 95–99% of radiation in the relevant spectral region 2. UV-Vis spectroscopy by fiber optics (a variation of diffusion and specular reflectance), and 3. UV-Vis micro-spectroscopy (a variation of diffusion and specular reflectance, or transmission depending on the mode of operation).

Transmission Spectroscopy The application of transmission UV-Vis spectroscopy is the preferred mode of measurement, as this is the simplest mode, requiring the least amount of additional contributing factors. For typical homogeneous catalytic experiments, provided that the solution has sufficient transmission of UV-Vis radiation, the setup need not be different from that described in Fig. 11.1; a simple cuvette will then often suffice, although this cuvette can be adapted to make it possible to apply both temperature and pressure. To perform, e.g. (homogeneous) electrocatalytic experiments, one can include a working electrode, counter electrode, and reference electrode in the cuvette provided they are small enough and are not in the optical path of the

11

242

C. Vogt et al.

UV-Vis beam. For these types of reactions, standard spectroscopic equipment can generally be used. Performing heterogeneous electrocatalytic experiments in transmission mode should in principle be possible if the sample can be prepared as a thin-film, while still providing sufficient transmission, although it might be more relevant to measure such heterogeneous systems (including photocatalysis) using the experimental methodologies discussed below. Nevertheless, there are only limited examples of heterogeneous catalysts successfully measured in transmission mode (e.g., zeolites [9–11]), as these samples have to be sufficiently thin and non-absorbing to allow such measurements. A self-supported catalyst wafer of a support material or catalyst can sometimes fulfill these requirements. These

Fig. 11.2 (a) Picture of different catalyst wafers for measuring transmission UV-Vis spectroscopy under operando conditions; (b) An example of a spectroscopic setup that can be used for operando UV-Vis spectroscopy studies of selfsupported catalyst wafers in transmission (or reflection) mode, as used for, e.g., the methanol-toolefins (MTO) reaction over zeolite-based catalysts; and (c) Picture of the particular cell design for transmission operando UV-Vis spectroscopy measurements. (Reproduced from [32]; with permission from Wiley-VCH)

a)

wafers are prepared by grinding and sieving catalyst or support particles (optional), and later applying an appropriate amount of pressure in a pellet press with a dye. Around a few tons of pressure is typically sufficient to create appropriate wafers. Care should be taken not to apply too much pressure, as the material will become very brittle and crystal structures may be damaged. In addition, a certain degree of gaseous transmission is required for in situ measurements. As a showcase, Fig. 11.2a shows a self-supported catalyst wafer of H-SAPO-34, used as a catalyst in the Methanol-toOlefins (MTO) reaction, of 16 mm diameter and around 0.1 mm thickness, weighing 5–7 mg, by using a Laboratory Pellet Press and around 4 tons of pressure under vacuum [32]. These prepared wafers can be placed in commercial

1 cm KBr windows Gas outlet UV-Vis probe

b)

Heating stage Catalyst pellet

Mass spectrometer Mass flow controllers N2 and O2

Operando UV-Vis cell

Gas saturator with methanol feedstock Custom-built hydrocarbon trap

c)

11

Ultraviolet-Visible (UV-Vis) Spectroscopy

transmission cells with appropriate windows, such as those designed by Linkam (Fig. 11.2b, c). These cells contain a heating stage, liquid cooling, and gas connections, so that a catalytic experiment may be performed in the gas-solid or liquid-solid range. Depending on the modes of operation, the specific requirements of the catalytic reaction and the sample, and the potential coupling with other spectroscopic techniques, one may choose to design spectroscopic cells specifically for the experiment in mind. One of the important aspects to verify is that the dead volume of the cell is low to avoid contributions of the gas-phase, as well as to ensure a timely response on the online GC or MS device by decreasing the length of the gas lines between cell and GC or MS device when structure-performance relationships have to be established. This can be done by performing a pulse experiment and following a time-on-stream GC or MS profile. The operando transmission UV-Vis spectroscopy setup with online MS in Fig. 11.2 has been used to investigate the effect of different types of impurities (i.e., feed and internal impurities) on the hydrocarbon pool species [32]. It was found that feedstock impurities strongly influence the location of coke deposits, whereas organic impurities retained after regeneration function as “seeds” for coke formation.

Diffuse Reflectance Spectroscopy Diffuse reflectance measurement cells for performing in situ or operando UV-Vis spectroscopy measurements, making use of the Praying Mantis™-type accessories as schematically shown in Fig. 11.3a, are now widely available for powdered samples, offering a plug-flow type reactor (Fig. 11.3b). These cells make use of convergent mirrors to collect diffuse-reflected radiation off of the loaded catalyst powder in the reactor cell, allowing it to be studied at elevated temperatures and under gaseous atmosphere, which fulfills all of the requirements listed above. This Praying Mantis™ setup was also the first one used for performing operando UV-Vis-NIR spectroscopy of supported metal oxide catalysts [26, 27]. A classical integration sphere is shown in Fig. 11.3c and is usually acquired as an accessory with a UV-Vis-NIR spectrometer when someone wants to conduct diffuse reflectance spectroscopy (DRS) measurements of powdered materials. This cell can then be used, shown in Fig. 11.3d, to measure solid catalysts under hydrated/dehydrated or oxidized/ reduced conditions at room temperature [19–21]. However, the classical integration sphere is more problematic to use when performing in situ or operando measurements of catalysts. Due to this, there are no commercial cells available. Nevertheless, in situ cells have been developed in academic labs, as the one shown in Fig. 11.3e [33]. The sample is then positioned outside the integration sphere in a quartz container surrounded by a heating mantle. This cell can be operated at

243

elevated temperatures and pressures (e.g., temperatures n!π*. The electronic transition in an irradiated solid sample under investigation can occur via metal-centered (MC) transitions (e.g., d-d transition in transition metals) or charge transfer transitions, such as ligand-to-metal (LMCT), metal-to-ligand (MLCT), and metal-to-metal (MMCT). UV-Vis Diffuse Reflectance spectroscopy (UV-Vis DRS) is used to analyze: (1) the chemical structure, electronic state, coordination state of the active metal, (2) the interaction between the active metal and support during catalyst preparation, pretreatment, and reaction, (3) physical/chemical adsorption behaviors of the reacting molecules on catalyst surface, and (4) the carbocations and unsaturated hydrocarbon species formed during reaction. The spectra can be interpreted with equal vigor in terms of absorption intensity or Kubelka-Munk function (F(ρ)) depending on the instrumental mode of data collection. The analysis of UV-Vis spectra under in situ conditions (situated in the original, natural, or existing position) provides essential information for understanding the reaction mechanism. In the subsequent subsections of this chapter, we will discuss the versatility of UV-Vis spectroscopy in the context of catalyst materials utilized in treatments or chemical reactions, including other relevant spectroscopic techniques and example-driven product analysis.

0.15 0.10

NH3-SCR Over Supported Vanadium/ Copper Catalysts

Nitrogen oxide (NOx), primarily emitted from the combustion of fossil fuels, has serious detrimental impacts on the environment [1, 2]. Among the various efforts that have been devoted to reducing the NOx emission, selective catalytic reduction (SCR) technique is the most widely used owing to its low cost and high efficiency [3, 4]. Several types of catalysts have been found to be active, including those containing vanadium [5], copper [6], manganese [7], chromium [8], etc. Vanadium-based catalysts are usually implemented for stationary sources and diesel vehicles to catalyze the “Standard SCR” reaction (4NH3 þ 4NO þ O2 ! 4N2 þ 6H2O) as well as “Fast SCR” reaction (4NH3 þ 2NO þ 2NO2 ! 4 N2 þ 6H2O) [9]. UV-Vis spectroscopy is a powerful tool to elucidate the nature and function of active sites in vanadium-based

0.05 0.00

F (Rf)

12.2

0.8 I-VHMS-9.5 A B C D 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 2 3 4 5 Energy (eV)

S-VHMS-11.7 E

6

A B

2

C

D

3 4 5 Energy (eV)

E

6

Fig. 12.1 UV-Vis spectroscopy of vanadium-based catalysts with five bands identified by peak deconvolution. X-axis: energy of the incident light; Y-axis: reflectance intensity given by Kubelka-Munk function. (Reproduced with permission from ref. [10]. Copyright 2011 The American Chemical Society)

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

paper, 500 ppm NO/O2/N2 gas mixture was fed to the reactor after the adsorption of NH3, and DRIFTS and UV-Vis spectra were collected simultaneously. The adsorption of NH3 on the catalyst surface results in both BNH3 and LNH3 being titrated on the Brønsted acid and Lewis acid sites, respectively. The responses of BNH3 and LNH3 intermediates were studied by DRIFTS, while the oxidation state of vanadium was determined by UV-Vis spectroscopy. It was found that the reduction of V5+ to V4+ occurred simultaneously with the rapid consumption of LNH3 species, while BNH3 remained unchanged. This indicated that NO preferentially reacted with NH3 species adsorbed on Lewis acid sites, while NH3 species on Brønsted acid sites were less reactive. The standard SCR reaction mechanism was further found to proceed through a nitrosamide intermediate with the aid of additional pulse experiments. The oxidation state of vanadium during the different stages of catalysis was also studied by Zhu et al. under varying conditions using in situ UV-Vis spectroscopy [12]. These studies concluded that the interaction between NH3 and NO contributed to the reduction of V5+ to V4+ state, followed by the initiation of re-oxidation of V4+ by O2. The extent of V5+ reduction was also quantified according to the d-d transition band of V4+ at ~700 nm. The time-dependent evolution trends of V5+/V4+ demonstrated that re-oxidation at low temperature ( 0.2, prominent bands at 6200, 9200, 22,700, and 30,000 cm1 were observed in their UV-Vis spectra (Fig. 12.2). They had compared the fingerprint spectrum to a series of model complexes with various Cu/O2 ratio and found that dinuclear copper complexes were the best fit. Based on the intense absorption band at 22700 cm1 and a Cu-Cu distance of 2.79–2.92 Å according to in situ EXAFS results, the bis(μ-oxo)dicopper sites (i.e., [Cu2(μ-O)2]2+) were originally considered to be the active sites of Cu-ZSM-5 catalyst. However, this theory was challenged by Woertink et al. based on their 18O isotopesensitive resonance Raman spectroscopic studies [14]. Their results excluded all the previously known copper-oxygen sites and the research team proposed a bent mono-(μ-oxo) dicupric site as the active site. The geometric and electronic

267

CZ-31-0.16 CZ-31-0.34 CZ-31-0.58

12

Absorption (a.u.)

12

10,000

20,000

30,000

40,000

50,000

Wavenumber (cm–1)

Fig. 12.2 UV-Vis spectra of CZ-31-0.16, CZ-31-0.34 and CZ-31-0.58 after being treated by O2 at 623 K for 3 h. X-axis: wavenumber; Y-axis: absorbance intensity. (Reprinted with permission from Ref. [13]. Copyright 2003 The American Chemical Society)

structure of this site was determined through normal coordinate analysis (NCA) and density functional theory (DFT) calculation [14, 15]. The copper precursors can also be directly characterized by UV-Vis spectroscopy. For the hydrated forms of Cu-SSZ-13 catalysts with a Cu:Altot ratio < 0.2, a narrow feature at ~42,000 cm1, was identified and attributed to the O ! Cu charge transfer in isolated Cu(II) and a d-d transition signal from hydrated Cu(II) species at 12,500 cm1 [16]. The UV-Vis signal intensities of hydrated Cu(II) species in Cu-SSZ-13 with various Cu:Altot ratios appeared to be linearly related to the standard SCR reaction rate, suggesting a direct correlation between the hydrated Cu(II) species and the active sites for SCR. This was then revealed by operando (in real time and under operating conditions with simultaneous activity measurements) X-ray absorption spectroscopy (XAS) analysis that the precursor or hydrated Cu(II) transformed to the catalytically active, isolated Cu(II) species in the six-membered ring under dehydrating condition.

12.3

Dehydrogenation of Propane Over Supported Catalysts

Propylene is widely used for the production of polymers, oxygen containing compounds, and other valuable commodity chemicals [17]. Oxidative dehydrogenation of propane (ODH) is a promising supplement to the conventional petroleum-based technology [18]. Vanadium [19], platinum [20], chromium [21], and other metal-based heterogeneous catalysts have been reported to be active in catalyzing the

268

Z. Yang and M. Zhu

ODH reaction. Among them, vanadium-based catalysts have received maximum attention due to their lower cost, higher catalytic activity, and better selectivity toward propylene production [22–24]. This section describes some examples of determination of vanadium oxidation states using UV-Vis spectroscopy. To investigate the effect of propane/O2 ratio and catalyst loading on the molecular structure of the active sites, Gao et al. performed in situ UV-Vis spectroscopy analyses on 4% V2O5/ZrO2 catalysts at 300  C in the propane/O2 mixtures with different ratios (Fig. 12.3). With an increased propane/ O2 ratio, the oxygen-to-vanadium LMCT band at above 20,000 cm1 decreased, while the V3+/V4+ signal below 20,000 cm1 increased accordingly [25]. The authors also quantified the relative degree of reduction of V5+ species at different propane/O2 ratios and investigated catalysts with higher metal loadings (1% V2O5/ZrO2 and 2% V2O5/ZrO2) [26]. It was concluded that the reduction of V5+ to V3+/V4+ was more pronounced with higher propane/O2 ratio and vanadium loading under steady-state ODH reaction conditions. The effect of the two oxidants, N2O and O2, on the extent of catalysis was investigated by Ovsitser et al. using in situ UV-Vis spectroscopic analysis (Fig. 12.4) [27]. Under steady-state ODH reaction conditions with O2 as the oxidant, the spectrum was nearly identical to that in the O2 atmosphere. However, when the oxidant was switched to N2O, the spectrum became similar to that in H2 environment, with pronounced UV-Vis features of V3+/V4+ species at 400–700 nm. The authors then performed an in situ UV-Vis analysis in a transient mode to compare the reoxidation capability of O2 and N2O. The oxidized catalyst was sequentially treated under conditions of reduction (C3H8/Ne ¼ 40/

a b V5+ c d e

F (R∞)

3

2

1 V3+/V4+ 0 45,000

35,000 25,000 Wavenumber (cm –1)

15,000

Fig. 12.3 In situ UV–vis spectra of 4% V2O5/ZrO2 at 300  C in (a) O2/ He; (b) 0.8C3H8/4O2/45.2He; (c) 9C3H8/9O2/32He; (d) 9C3H8/1.5O2/ 39.5He; and (e) 9C3H8/41He (cm3/min). X-axis: wavelength; Y-axis: reflectance intensity given by Kubelka-Munk function. (Reprinted with permission from Ref. [25]. Copyright 2002 Elsevier)

60, 773 K, 10 min), oxidation (O2/Ne ¼ 20/80), reduction (C3H8/Ne ¼ 40/60, 773 K, 10 min), and re-oxidation (N2O/ Ne ¼ 40/60, 773 K, 10 min). After exposing the reduced catalyst to O2 for 5 s, a complete re-oxidation of V3+/V4+ to V5+ species was observed. In contrast, the re-oxidation process took place after 25 s for N2O. This directly proved that the oxidation ability of N2O is weaker than that of O2 and the average oxidation states of VOx in N2O þ C3H8 flow was lower than that in O2 þ C3H8 flow. The authors concluded that the lower oxidation state of VOx induced by N2O is the key factor for the improved selectivity toward propylene over undesired CO2. Recently, extensive efforts have been focused on the change of vanadium on the catalyst surface by UV-Vis technology. Hess et al. [28] observed the reduction degree of vanadium and CeO2 on the surface of VOx/CeO2 catalysts using operando UV-Vis spectroscopy. The UV-Vis spectra was analyzed using Tauc plots [29]. The results showed that the band gap of bare CeO2 shifted by 0.05 eV after switching from the oxidation to the reaction condition, and even less for the sample spiked with VOx. The shift in band gap decreased as VOx loading increased. Besides, Ce3+ oxide has a smaller band gap than Ce4+ oxide [29, 30], which indicates that an increase in the vanadium loading can lead to a decrease in CeO2 reduction under the reaction conditions. Figure 12.5 shows the increase in absorption at higher wavelengths upon transition from the oxidized to the reaction state. This is attributed to the d-d transition of the reduced vanadium species (e.g., V3+ or V4+), which increases with the increase in vanadium loading. However, intensity of the d-d transition did not correlate with conversion and was much weaker for VOx/CeO2 than for other catalysts in which vanadium participated in the process. [25, 31]. The above results illustrate that vanadium hinders the reduction of CeO2 and the reduced vanadium species do not affect the ODH reaction of propane. Zhang et al. [32] also used UV-Vis to investigate the transition mechanism of vanadium (V) into various species on the catalyst surface. It was found that for the catalyst with 1.7 wt.% V, the peaks at 216, 240, and 255 nm are considered as charge transfer (CT) transition between O2 ligands and V5+ ions in isolated VO4 species. This peak is considered to be a CT transition of the tetrahedrally coordinated V5+ moiety in Mg3V2O8 [33]. This indicates that the catalyst surface changes from VO4 tetrahedra to Mg3V2O8 nanoclusters and eventually to Mg3V2O8 crystallites with the gradual addition of V. Kharlamova et al. [34] used the UV-Vis Diffusive reflectance (DR) method to study the change in V on the catalyst surface after adding MgO. The addition of MgO caused a blue shift in the spectrum, corresponding to LMCT transitions of V5+ cation. The blue shift became more obvious with the increase of MgO amount. The edge energy of the LMCT transition of V5+ cations was shown to correlate with the coordination environment of V5+ cation

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

a)

269

d)

1.5

1.5

1.0

H2

0.5

C3H8+N2O

Kubelka-Munk

Fig. 12.4 In situ UV-Vis spectra of (a) VOx(2.7)/MCM-41, (b) VOx(2.8)/SiO2, and (c) VOx(11.2)/MCM-41 under different reaction conditions at 773 K as well as VOx(2.7)/MCM41 during (d) O2-C3H8-O2 and (e) O2-C3H8-N2O cycles. X-axis: wavelength; Y-axis: reflectance intensity given by Kubelka-Munk function. (Adjusted with permission from Ref. [27]. Copyright 2007 The American Chemical Society)

Kubelka-Munk

12

C3H8

0.5 5s in O2 after C3H8

0.0

Oxidized O2 C3H8+O2

200

12

1.0

400

0.0

600

200

800

400

600

800

l (nm)

l (nm)

e)

b)

1.5

6

4 C3H8+N2O C3H8 2

Kubelka-Munk

Kubelka-Munk

C3H8 1.0

0.5 5s in N2O after C3H8 0.0

O2

0 200

400

C3H8+O2 600

800

l (nm)

Kubelka-Munk

c)

Oxidized 25 s in N2O after C3H8

200

400

600

800

l (nm)

0.40

C3H8+N2O

0.35

C3H8

0.30 C3H8+O2

0.25 0.20

O2

0.15 200

400

600

800

l (nm)

and the number of covalent V–O–V bonds around the central V5+ cation (CVB) [35]. According to this correlation, the number of V-O-V bonds changes with the addition of MgO. Vanadium-based catalysts can also be used to catalyze the non-oxidative dehydrogenation of propane (DH). Sokolov et al. investigated the deactivation mechanism of CrOx/ MCM-41 and VOx/MCM-41 catalysts during reaction by comparing the UV-Vis spectra of H2-pretreated catalysts and deuterium-labeled DH stream-treated catalysts [36]. In a single DH cycle, the formation of carbon deposits was

found to be the main cause of deactivation, as reflected by the appearance of low-condensed aromatic species at 325 and 450 nm, as well as polyaromatic graphitic species at 570–800 nm after exposure to DH stream for 30 min [37]. It was further revealed that the loss of activity in CrOx/MCM41 after each cycle was caused by structural changes in the catalytically active metal/metal oxide species. On the contrary, the structure of the active VOx species on MCM-41 remained unchanged during the DH/regeneration cycles, which led to a stable catalytic performance.

270

Z. Yang and M. Zhu

Intensity (K.-M. units)

0.9

514

3850.6

60

532

0.3

40

0

480

560

640

720

20

b)

Before reaction During reaction 1.36 V/nm2 275 °C

1.8

70

1.5

60

1.2

514

385 0.9

50

532

0.6

40

d-d transition (V3+/V4+)

0.3 0

30

480

560

640

20

720

c) 0.3

400 500 600 700 Wavelength (nm)

800

0.04 0.15 0.03 0.1

0.02 0.01

0 300

400 500 600 700 Wavelength (nm)

Fig. 12.5 Operando UV-Vis spectra (red) of bare ceria (a) and VOx/ CeO2 (1.36 V/nm2) (b) at 275  C compared to spectra recorded under oxidative conditions prior to reaction (blue). The laser excitation wavelengths (385, 514, and 532 nm) are mentioned in the graph. Insets highlight the region between 470 and 760 nm. Band gap shifts were

12.4

0.05

0.2

0.05

0 300

Electroreduction of CO2 Over Molecular Catalysts

Excess CO2 emission from anthropogenic activities has broken the global carbon balance and led to a series of environmental issues, such as global warming and ocean acidification [38, 39]. The electroreduction of CO2 (CO2RR) into fuels and value-added chemicals using renewable electricity as the energy input is a promising technique for energy storage and CO2 mitigation [40, 41]. Among the various catalysts, porphyrins and their derivatives have showed extraordinary efficiency in converting CO2 to CO, which can further be utilized as a feedstock in the FischerTropsch process to produce downstream chemicals [42–45]. Porphyrin molecules typically feature a Q band and a Soret band in their UV-Vis spectra. The Q band is the a1u(π)-eg(π*) transition from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO), and the Soret band is the deeper a2u(π) to LUMO transition. Due to their relatively high sensitivity, this technique has been employed to understand the nature of active sites of porphyrin-based catalysts during the catalytic reaction [47–49]. For example, cobalt porphyrin-based covalent organic frameworks (COFs) were synthesized and found to exhibit high turnover numbers (TON) of around 290,000 for CO production at an overpotential of 550 mV (Fig. 12.6a) [46]. In order to determine the valence state of the Co center during this reaction, the catalysts were deposited on fluorine-doped tin oxide (FTO) glass and analyzed by in situ UV-Vis analysis in 0.5 M KHCO3 electrolyte saturated

800

0.07 0.06

0.25

10 0

d-d transition intensity Bandgap energy shift

Bandgap shift (eV)

80

Before reaction During reaction CeO2 275 °C

d-d intensity (K.M. units)

1.2

Intensity (K.-M. units)

a)

0 0

0.5 1 1.5 2 2.5 Vanadia loading (Vnm–2)

determined using Tauc plots and the d-d transition intensity was determined based on intensities measured at 700 nm (c). X-axis: wavelength or V loading; Y-axis: reflectance intensity given by Kubelka-Munk function. (Reprinted with permission from Ref. [28] Copyright 2021 Elsevier)

with CO2 under different applied potentials. As shown in Fig. 12.6b, the intensities of the bands at 525 and 640 nm increased with higher applied potentials, indicating that Co (I) was favored at reaction conditions and may be responsible for CO2RR. The time-dependence of UV-Vis signal at 640 nm could be further tracked at 0.57 V vs. RHE, which was close to the redox potential of Co1+/2+. An apparent diffusion coefficient of 2  1012 cm2s1 could be obtained by fitting the data using a modified Cottrell equation (Fig. 12.6c). Similarly, Kornienko et al. synthesized metal-organic frameworks (MOF) with cobalt porphyrin as the repeating unit. The obtained catalysts exhibited a CO selectivity of 76% over 7 h time-on-stream with a TON of 1400 [50]. Meanwhile, in situ UV-Vis analysis was also conducted to gain insights into the oxidation state of Co during the reaction. In the open-circuit state, the cobalt porphyrin-based catalyst (Al2(OH)2TCPPCo) showed a Soret band at 422 nm and a Q band at 530 nm. When the potential varied from 0.2 to 0.7 V vs. RHE, the Soret band corresponding to Co(I) species could be found at 408 nm. Plotting the difference spectra clearly illustrates the UV-Vis signal responses under reductive conditions, which could be used to determine the redox potential of the cobalt species. The first derivative of the time-dependent UV-Vis signal was calculated with a maximum at 0.4 V vs. RHE. This is consistent with the cyclic voltammogram results showing a cathodic wave at ca. 0.4 V vs. RHE. Interestingly, the UV-Vis spectra of cobalt porphyrins dissolved in organic solvents exhibited different trends compared to those collected in aqueous electrolytes. Behar

12

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

271

a)

N

O O

BDA

12

N N N M N N

N

N

.5

Å

26

.5

23

NH2

H2N

COF-366-M

COF-366-Co

M = Co

N N M N N H2N

M(TAP)

NH2

N

N N N M N N

O BPDA

O N

N

COF-367-M

b)

COF-367-Co COF-367-Co(10%) COF-367-Co(1%)

c)

0.10 0.08

M = Co M = Co/Cu (10/90) M = Co/Cu (1/99)

Å

COF-366-Co Applied potential:

COF-366-Co at –0.57 V, 640 nm

0.08

Δ Absorption (a.u.)

Δ Absorption (a.u.)

–0.57 V 0.06

0.33 V

0.04 0.02

0.06

0.04

0.02

0.00 0.02

0.00 400

500

600

700

800

Time1/2 (s1/2)

Fig. 12.6 (a) Design and synthesis of metalloporphyrin-derived 2D covalent organic frameworks. (b) In situ UV-Vis spectrum at various applied potentials (0.23 to 0.57 V vs. RHE) with reference to that under 0.33 V vs. RHE. (c) Time dependence of the relative UV-Vis

0

2

4

6

Time1/2 (s1/2)

absorbance at 640 nm at 0.57 V vs. RHE. X-axis: wavelength; Y-axis: absorbance intensity. (Adjusted with permission from Ref. [46]. Copyright 2015 The American Association for the Advancement of Science)

272

Z. Yang and M. Zhu

et al. performed cyclic voltammetry analysis on a cobalt tetraphenylporphyrin (CoTPP) solution in butyronitrile saturated with Argon and observed two successive redox features: CoIITPP to CoITPP at 0.86 V vs. SCE and CoITPP to Co0TPP at 2.02 V vs. SCE [51]. Then, they purged the electrolyte with CO2 and collected the cyclic voltammograms again. While the CoIITPP ! CoITPP reduction was unaffected, the current density of CoITPP ! Co0TPP reductive wave was increased by a factor of 8, indicating that the reduction of CO2 might be catalyzed by Co0TPP species (Fig. 12.7a). To validate this hypothesis, the authors dissolved cobalt porphyrins in tetrahydrofuran (THF) and introduced sodium metal as the reducing agent. CoIITPP was gradually reduced to CoITPP as evidenced by the appearance of two new UV-Vis bands at 364 and 426 nm (Fig. 12.7b). Later, further reduction of CoITPP led to the formation of new species with bifurcated Soret bands at 314 and 420 nm, possibly from Co0TPP (Fig. 12.7c, d). Interestingly, purging CO2 into the THF solution containing CoITPP did not result in any spectral change. In contrast, addition of CO2 to the Co0TPP/THF solution immediately oxidized the porphyrin back to CoITPP, and then CoIITPP after several hours. This directly confirmed that Co0TPP is the active phase for electroreduction of CO2 in presence of CoTPP solution.

12.5

Methanol to Olefin (MTO) Process Over Zeolite Catalysts

The methanol-to-olefins (MTO) reaction, which is usually catalyzed by protonated zeolites or zeotype materials, is a promising process for upgrading C1 feedstock (methanol) to petrochemicals [52–54]. To date, a consensus has been reached that the MTO reaction occurs at the Brønsted acid site together with an organic reaction center, and follows a “hydrocarbon pool (HP) hypothesis” [55, 56]. Active organic compounds, acting as reactive intermediates, are formed during the induction period and can either bind with methanol or directly dissociate to form olefins [57–59]. Due to the complexity of the MTO reaction, it still remains a challenge to fully understand the intermediates and reaction pathways. Among the various characterization techniques, UV-Vis spectroscopy is an ideal one for its high sensitivity to the charged aromatic compounds, which are known as the key intermediates [60–62]. To evaluate whether the organic intermediates remained in neutral or cationic form, samples that were pretreated with methanol vapor can be exposed to an ammonia stream, which is believed to diffuse through the nanopores and react with the cationic species. As shown in Fig. 12.8a, b, Hemelsoet et al. found that the intensities of the UV-Vis band at ~400 nm for H-SAPO-34 and H-SSZ-13 started to decrease immediately upon exposure to NH3 [63].

150

Current

Molar absorptivity (L mmol–1 cm–1)

a)

b)

100

50

0 60

c)

50 40 30 20

0.0

0.5

1.0

1.5

2.0

Potential (V)

Fig. 12.7 (a) Cyclic voltammograms of cobalt porphyrins in butyronitrile solutions saturated with Ar (dotted line) and CO2 (solid line). (b) Spectral changes upon reduction of CoIITPP by sodium in THF. (c) Spectral changes reduction of CoITPP by sodium in THF. (d)

2.5

Δε

10 0 –10 –20 –30

d) 300

400 500 600 Wavelength (nm)

700

The difference spectrum of CoITPP/Co0TPP in THF. (Adjusted with permission from Ref. [51]. X-axis: wavelength; Y-axis: molar absorptivity/Lmmol1cm1. Copyright 1998 The American Chemical Society)

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

a)

b)

H-SAPO-34 0.25

750 s (α) 900 s (β) 1250 s (γ) 3450 s (δ)

0.20 0.15 A

0.15

H-SSZ-13 0.25

1050 s (α) 1150 s (β) 1400 s (γ) 3150 s (δ)

0.20

0.10

0.10

0.05

0.05

0.00

0.00 400

c)

500 600 Wavelength (nm) αβ γ

0.25

400

700

d)

δ

0.20

0.20

0.15

0.15

0.10

0.10

0.05

0.05

0.00

0.00 0

1000

2000 TOS (s)

The UV-Vis spectra were then deconvoluted into a few Gaussian bands for further analysis of the underlying reaction intermediates, which appeared to behave differently. For H-SAPO-34, the Gaussian bands at around 400 and 450 nm dropped faster than the one at 505 nm. In addition, the intensities of all three bands for H-SSZ-13 decreased much faster and tended to stabilize after introducing NH3 for 2000 s (Fig. 12.8c, d). These differences were explained as being due to distinct acidity and/or pore diffusion features. Establishment of the structure–performance relationship for the MTO process catalyzed by zeolite catalysts can also be achieved through operando UV-Vis spectroscopy. For instance, Goetze et al. evaluated the MTO process for three small-pore zeolite catalysts with different framework consisting of large cages interconnected by small eight-ring windows (Chabazite: CHA, Deca-Dodecasil 3 Rhombohedral: DDR, and Levynite: LEV). The generation of hydrocarbon pool species in these catalysts was quite different as identified by operando UV-Vis spectroscopy and subsequent GC/MS analysis (Fig. 12.9) [64, 65]. In order to get more specific information on the band positions from the broad and often convoluted bands in operando UV-Vis spectra, the spectra collected at different stages of the reaction (i.e., one during the induction period, two during the active period, and two during deactivation) can be deconvoluted by fitting the

3000

4000

500 600 Wavelength (nm)

αβ γ

0.25

A

A

Fig. 12.8 Time-resolved UV-Vis spectra during MTO reaction at 509 K and the corresponding time evolution of Gaussian bands over (a, c) H-SAPO-34 and (b, d) H-SSZ-13 crystals. (black squares, 400 nm; red circles, 449 nm; green triangles, 505 nm). At time β (1125 s for H-SAPO-34 and 900 s for H-SSZ-13) the methanol feed was stopped and at time γ (1400 s for H-SAPO-34 and 1250 s for H-SSZ-13) the ammonia feed was started. X-axis: wavelength or time-on-stream; Yaxis: absorbance intensity. (Reprinted with permission from Ref. [63]. Copyright 2013 WILEY-VCH)

273

A

12

0

1000

700

δ

2000 TOS (s)

3000

4000

spectra with Gaussian curves using a peak-fitting programFityk [66]. As a result, the species formed inside the zeolites during MTO can be identified to be methylated naphthalene and pyrene in CHA, 1- methylnaphthalene and phenalene in DDR and methylated benzene and naphthalene in LEV, leading to a lattice expansion of all three frameworks. In a nutshell, operando UV-Vis spectroscopy is an efficient approach to gain insight into the effect of catalytic structure in MTO processes. Previous studies have suggested that the rate-determining step of the MTO reaction is the methylation reactions in which the methanol molecule is protonated and transferred to a ring carbon atom to form cationic aromatics [67–70]. The in situ UV-Vis spectrum of the experimental solution recorded during the reaction can therefore be utilized for kinetic analysis. Speybroeck et al. monitored the evolution of Gaussian bands at 400, 450, 505, and 580 nm as a function of time during MTO reactions over H-SAPO-34 and H-SSZ-13 catalysts (Fig. 12.10a, b) [71]. They fitted the rapidly developing regions that follow first-order kinetics to the equation ln(Amax  A) ¼ lnAmax  kabst (A: absorbance intensity/a.u.; kabs: rate constant/s1; t: time/s). To avoid deviations due to induction period at the beginning and secondary reactions near the end, regions between 0.2Amax and 0.5Amax for H-SAPO-34 and between 0.3Amax and

12

274

Z. Yang and M. Zhu

a)

25,000 cm–1 10.9 Å

17,000 cm–1 10.9 Å

b)

24,800 cm–1 9.4 Å

c

a

c

b

a 6.7 Å Wavelength (nm) 300 400

b

6.7 Å

500 600 700

7.1 Å 7.1 Å Wavelength (nm) 300 400 500 600 700

1.4

1.2

1.2

1.0

1.0 Absorbance

Absorbance

1.4

20,000 cm–1 9.4 Å

0.8 0.6 0.4

0.8 0.6 0.4 0.2

0.2 0.0

0.0 40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

c)

25,000 cm–1 7.5 Å

7.5 Å

c

a

b

1.4

6.5 Å 6.5 Å Wavelength (nm) 300 400 500 600700

1.2 Absorbance

1.0 0.8 0.6 0.4 0.2 0.0 40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

Fig. 12.9 Operando UV-Vis spectra during methanol conversion and corresponding hydrocarbon species over the (a) CHA (b) DDR, and (c) LEV catalysts. X-axis: wavelength; Y-axis: absorbance intensity.

(Reprinted with permission from Ref. [64]. Copyright 2018 The American Chemical Society)

0.7Amax for HSSZ-13 were selected. The absorption rate constant kabs could then be determined and plotted as a function of 1/T, from which the activation energy for forming the charged (poly)aromatic species was calculated

(Fig. 12.10c, d). Using this method, the activation energy for the species at 400 nm could be determined to be 98 and 81 kJmol1 for H-SAPO-34 and H-SSZ-13, respectively. These numbers agreed well with the theoretically predicted

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

a)

275

b)

H-SAPO-34 0.4

403 nm 448 nm 505 nm 580 nm

0.3

0.2

0.1

H-SSZ-13 400 nm 448 nm 505 nm 575 nm

0.3 Absorbance (a.u.)

Fig. 12.10 (a, b) Time development of the Gaussian functions during the MTO reaction at 553 K. X-axis: time/s; Y-axis: absorbance intensity. (c, d) Arrhenius plots derived from the temporal evolution of various absorption bands in the temperature range 500–600 K. (Reprinted with permission from Ref. [71]. Copyright 2012 WILEY-VCH)

Absorbance (a.u.)

12

0.2

0.1

0.0

0.0 200

600

1000

100

1400

200

t (s)

c)

0

In(kabs) /In(s–1)

In(kabs) /In(s–1)

500

0

Ea = 99 kJ/mol

–2 –3

–5

400

d) Ea = 98 kJ/mol

–1

–4

300 t (s)

403 nm 448 nm 0.0017

–2

T–1 (K–1)

methylation activation energies of the same species, i.e., 105 and 77 kJmol1 in H-SAPO-34 and H-SSZ-13, respectively. Furthermore, the activation energies of forming the species at 450 nm were similar to those species at 400 nm, indicating that the governing reaction steps for the methylation reaction involves the cationic, highly methylated benzenic compounds with an UV-Vis absorption band at ~400 nm. To study the kinetic evolution of the retained hydrocarbons inside the zeolites in more detail, the time evolution of the operando UV-Vis spectra can be analyzed using Multivariate Curve Resolution (MCR). This method provides a decomposition that maximizes the explained variance in the data while allowing for the definition of constraints for the obtained components. Goetze et al. performed a chemometric analysis using the MCR-ALS toolbox, which performs a deconvolution of the time series of UV-Vis spectra into contributions from pure components [72]. Noticeably, these pure components are not necessarily pure chemical phases or species but a set of UV-Vis spectral features that show the same time behavior. Using this method, the obtained time series of the operando UV-Vis spectra can be split into two parts: (i) components, i.e., UV-Vis spectral features that follow the same kinetics during the reaction, and (ii) the contributions of these components versus time. An initial estimate

0.0019

0.002

Ea = 77 kJ/mol

–3 –4 –5

0.0018

Ea = 81 kJ/mol

–1

407 nm 448 nm 0.0017

0.0018

0.0019

0.002

T–1 (K–1)

of the contribution was made using Evolving Factor Analysis, a completely model-free resolution of overlapping peaks into concentration profiles and absorption spectra [72]. As a unimodality and non-negativity constraint was imposed on the contributions of components over time, a reconstruction of the original operando UV-Vis spectra is obtained [73]. To obtain an accurate reconstruction with a cumulative variance (CVE) > 99.9%, three groups of absorbing species formed with different spectral features at the induction, active, and deactivation period of the MTO reaction are presented by the blue, black, and red components, respectively (Fig. 12.11) [65]. This is well exemplified in the case of CHA, where the blue component, which has two bands around 35,000 cm1and 25,000 cm1 is dominant at the beginning of the reaction. These two bands are assigned to monoenyl carbocations and alkylated benzene carbocations, respectively. At a reaction temperature of 450  C, only two components were enough to reconstruct the original spectrum, which behaved similarly in time to the black and red components at lower temperatures. This indicates that the induction period species rapidly evolve into the active species at this reaction temperature. The broad feature around 25,000 cm1 is visible in the black component at a reaction temperature of 450  C and is dominant during the active period. On the other

12

276

Z. Yang and M. Zhu

300

Wavelength (nm) 400 500 600 700

0.8

1.2 Contribution

0.8 0.6 0.4

1.0

0.6

0.8 0.4

0.6 0.4

0.2

0.0 0

40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

300

Wavelength (nm) 400 500 600 700

100

200 300 Time (min)

400

0.8

1.2 Contribution

0.6 0.4

1.0

0.6

0.8 0.4

0.6 0.4

0.2

0.2

0.2 0.0

0.0

0

40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

300

Wavelength (nm) 400 500 600 700

100

200 300 Time (min)

400

0.0 500

1.0

1.6 1.4

1.0

0.8

1.2 Contribution

0.8 Absorbance

Methanol conversion

1.4

0.8

c)

0.0 500

1.0

1.6

1.0

Absorbance

0.2

0.2

0.0

b)

Methanol conversion

1.4

1.0

Absorbance

1.0

1.6

0.6 0.4

1.0

0.6

0.8 0.4

0.6 0.4

0.2

Methanol conversion

a)

0.2

0.2 0.0

0.0 40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

0

100

200 300 Time (min)

400

0.0 500

Fig. 12.11 Left panel: MCR-ALS components. Right panel: their respective contributions to the overall operando UV-Vis spectra measured as absorbance intensity vs. time for the conversion of methanol over CHA zeolite at a reaction temperature of (a) 350  C, (b) 400  C,

and (c) 450  C. The colors of the contribution plots correspond to the colors of the components. (Reprinted with permission from Ref. [65]. Copyright 2017 The American Chemical Society)

hand, at reaction temperatures of 350 and 400  C this feature is part of the red component, which is correlated with deactivation. Using the kinetics of the formation of different species described by the chemometric analysis, it is possible to evaluate the evolution of the retained hydrocarbon pool species using operando UV-Vis spectroscopy.

Although the zeolite materials exhibit outstanding activities and selectivities toward olefin production, they suffer from the coke formation where methyl substitution of the aromatic intermediates ends up as the polycyclic aromatic structures, leading to pore blocking and severely restricted mass transportation [74, 75]. Extensive studies have been

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

performed to investigate the nature of coke species during the deactivation process [76, 77]. Mores et al. investigated the coke formation on the H-ZSM-5 catalyst with different Si/Al molar ratio of 11, 12, 17, 37, and 44 (A–E) using in situ UV-Vis spectroscopy [78]. As shown in Fig. 12.12, the evolution of absorption bands associated with more extended

aromatic species (475 and 500 nm) was delayed with decreasing Brønsted acid site density (represented by the Si/Al molar ratio). Additionally, the formation of the 500 nm absorption band did not occur at the expense of 400 nm absorption band, indicating that the formation of important HP compounds and their extended variants is not impeded by morphological

in )

500 600 Wavelength (nm)

0 700

m

0.0 400

0 700

500 600 Wavelength (nm)

in

)

50 500 600 Wavelength (nm)

0 700

m

0 700

m

0.0 400

in

)

50

0.8 0.6 0.4 0.2 0.0 400

t(

0.2

Absorption (a.u.)

0.4

t(

500 600 Wavelength (nm)

m

0 700

t(

0.0 400

50 )

0.2 0.0 400

500 600 Wavelength (nm)

0 700

h)

0.0 400

500 600 Wavelength (nm)

0 700

t( m

0.0 400

500 600 Wavelength (nm)

0 700

j)

50 in

)

0.2 0.0

400

500 600 Wavelength (nm)

0 700

0.6 0.4 50

0.2 0.0 400

in )

0.4

500 600 Wavelength (nm)

0 700

m

0.6

t(

i)

50

0.2

in )

in

)

50

0.4

m

0.2

0.6

t(

0.4

Absorption (a.u.)

g)

0.4 in

in

)

50

m

0.2

0.6

t(

0.4

Absorption (a.u.)

f)

Absorption (a.u.)

Absorption (a.u.)

500 600 Wavelength (nm)

d)

e) Absorption (a.u.)

30

0.0 400

c)

Absorption (a.u.)

0.2

in )

50

0.2

0.4

m

0.4

0.6

t(

0.6

Absorption (a.u.)

b)

t(

Absorption (a.u.)

a)

Absorption (a.u.)

Fig. 12.12 UV-Vis spectra of the H-ZSM-5 catalysts with Si/Al molar ratio of 11, 12, 17, 37, 44 (A–E) taken during the MTO reaction. (a) A at 623 K, (b) A at 773 K, (c) B at 623 K, (d) B at 773 K, (e) C at 623 K, (f) C at 773 K, (g) D at 623 K, (h) D at 773 K, (i) E at 623 K, (j) E at 773 K. X-axis: wavelength; Y-axis: absorbance intensity. (Reprinted with permission from Ref. [78]. Copyright 2011 WILEY-VCH)

277

t( m

12

12

278

Z. Yang and M. Zhu

1.2 0.2 a.u.

AE3

0.8 0.6 0.4

b) 2.0 0.0

Absorbance (a.u.)

1.0

0.2 a.u.

Z2

0.8 0.6 0.4

100 80

1.6 Primary coke

60

1.2 0.8

40

Secondary coke

0.4 0.0 0

3

6

0.2

9 Time (h)

12

Top 20 Middle Bottom 0 15

Methanol conversion (%)

0.2

Absorbance at 10,000 cm–1

Absorbance (a.u.)

1.0

of the amount of Brønsted acid site. Moreover, high density of Brønsted acid site solely leads to rapid deactivation of the catalyst [79]. To solve this issue, Brønsted acid sites were isolated by the incorporation of alkaline-earth metals such as Ca2+ to form [Ca(μ-OH)2Ca]2+— a Lewis acid sites species. This species has a tendency to suppress aromatic growth by methylation and destabilize crucial carbenium ions for the aromatic cycle, which blocks the reaction pathway and efficiently impedes the formation of polyaromatic species with

C m har po ethyged lya lat rom ed ati cs

C m har be ethyged nz la en ted es

N aro eutra ma l tic s

a)

M d on catienylo- an ion ic d s

constraints under mild reaction conditions. At temperatures higher than 773 K, more extended aromatic species with adsorption bands at 600 nm could be observed for all catalysts. Higher Brønsted acid site density facilitates the formation of methyl-substituted aromatic species and its subsequent growth toward larger amount of coke species. However, the nature of those coke species remains unaffected. According to the UV-Vis DRS results, Yarulina et al. found that the nature of hydrocarbon species is independent

0.2 a.u.

Z1

c) Absorbance at 10,000 cm–1

Absorbance (a.u.)

1.0 0.8 0.6 0.4 0.2 0.0

40,000 35,000 30,000 25,000 20,000 15,000 Wavenumber (cm–1)

Fig. 12.13 (a) Operando UV-Vis spectra (measured as absorbance intensity) collected during methanol conversion over Z1 and Z2 zeolites with different Si/Al ratios and Ca-modified AE3 at 500  C indicates that aromatic species do not accumulate in Ca-modified samples. UV-Vis absorbance at 10,000 cm1 for (b) H-ZSM-5 and (c) Mg-ZSM-5 at the three positions along the reactor bed vs. time-on-stream. Secondary coke caused by conversion of MTO products is indicated in blue while the

2.0

100

1.6

80

1.2

60

0.8

Primary coke

0.4 Secondary coke 0.0 0

10

20

30 40 Time (h)

40 Top 20 Middle Bottom 0 50 60

Methanol conversion (%)

0.0

primary coke caused by conversion of methanol is indicated in gray. Photo of the reactor bed and corresponding operando UV-Vis spectra at three positions along the reactor bed of (d) H-ZSM-5 and (e) Mg-ZSM-5 after 2 h of catalyzing methanol-to-olefins (MTO) process. (Reprinted with permission from Ref. [79, 80]. Copyright 2018 Springer Nature, 2018 Royal Society of Chemistry)

12

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

d)

Methanol

279

e)

Methanol

1.4

1.0

12

1.2

Absorbance

Absorbance

0.8 1.0 0.8 0.6 0.4

0.6 0.4 0.2

0.2 0.0

0.0

40,000 35,000 30,000 25,000 20,000 15,000 10,000

40,000 35,000 30,000 25,000 20,000 15,000 10,000

1.4

Wavenumber (cm–1)

1.0

Wavenumber (cm–1)

0.8

1.0

Absorbance

Absorbance

1.2

0.8 0.6 0.4

0.6 0.4 0.2

0.2 0.0

0.0

40,000 35,000 30,000 25,000 20,000 15,000 10,000

40,000 35,000 30,000 25,000 20,000 15,000 10,000

1.4

Wavenumber (cm–1)

1.0

Wavenumber (cm–1)

1.0 0.8 0.6

5 mm

0.4

Absorbance

5 mm

Absorbance

1.2 0.8 0.6 0.4 0.2

0.2 0.0

0.0 40,000 35,000 30,000 25,000 20,000 15,000 10,000

Wavenumber (cm–1)

40,000 35,000 30,000 25,000 20,000 15,000 10,000

Wavenumber (cm–1)

Fig. 12.13 (continued)

the absorption bands below 15,000 cm1, a characteristic of catalyst deactivation (Fig. 12.13a). This results in a longer lifetime of the catalysts. Goetze et al. prepared Mg2+-modified ZSM-5 (Mg-ZSM-5) via spray impregnation of the aforementioned sample with a magnesium nitrate solution. The resulting catalysts exhibited a 3.4 times longer lifetime than ZSM-5 in its protonic form (H-ZSM-5) [80]. In order to investigate the progression of the coke front through the catalyst bed over time, operando UV-Vis spectra were measured at three positions along the reactor bed during the MTO process for both H-ZSM-5 and Mg-ZSM-5. There was an evolution of UV-Vis absorbance at 10,000 cm1, which was assigned to light absorbing poly-aromatic species directly leading to catalyst deactivation. This was used to evaluate the coke process of catalyst bed (as shown in Fig. 12.13b, c). For H-ZSM-5, there was rapid coke formation along the entire catalyst bed at the beginning of the process, while for Mg-ZSM-5, the increase in UV-Vis absorbance was much slower in the middle and bottom of the catalyst bed. As reaction proceeded, the absorbance at the top, middle, and bottom of the reactor reached its maximum absorbance value successively for both catalysts, with Mg-ZSM-5 showing a

much longer lifetime. It is worthy to mention hereby that the coke formation in the middle and bottom of the reactor for H-ZSM-5 occurred rapidly, although the methanol conversion zone (a small part of the catalyst bed to achieve 100% methanol conversion) was still at the top of the catalyst bed. This effect was observed and discussed by other researchers, with the coke formation being caused by secondary reaction of species formed in the methanol conversion zone [55, 81]. A Snapshot of the H-ZSM-5 catalyst bed and the corresponding operando UV-Vis spectra for the first 2 h of reaction are shown in Fig. 12.13d, e. For both H-ZSM-5 and Mg-ZSM-5, the color of top catalysts bed was darker than the other two positions, but the Mg-ZSM-5 showed a much lighter color and weaker intensity of absorption bands compared to H-ZSM-5, indicating much less formation of coke species. Therefore, the authors confirmed that the amount of strong Brønsted acid sites responsible for MTO activity can be reduced by the magnesium modification, limiting the formation of secondary coke caused by olefin aromatization, resulting in a higher lifetime to our benefit. This work highlights the potential of UV-Vis spectroscopy in investigating the MTO process catalyzed by these types of materials.

280

12.6

Z. Yang and M. Zhu

Conclusions and Remarks

This chapter has demonstrated through several case studies that UV-Vis spectroscopy is one of the fastest and most sensitive tools to study the reaction mechanism of heterogeneously catalyzed reactions. The application of UV-Vis spectroscopy in thoroughly understanding the reaction kinetics, role of intermediates, and the transportation of reactants, products, and intermediates at its transient or dynamic state is available in different gas-solid, liquid-solid, and gas-solidliquid systems. Like other methods such as FTIR, Raman, and X-ray photoelectron, spectroscopic studies with UV-Vis can be performed in both transmission or diffuse reflectance modes where the relation between spectral intensity and absorbance or reflectance can be quantitatively described using Lambert-Beer’s law or Schuster-Kubelka-Munk function. It also allows the acquisition of simultaneous kinetic and spectroscopic information about the active sites and the adsorbed species on catalyst surface via in situ or operando spectrokinetic investigations under realistic reaction conditions. Although UV-Vis spectroscopy is a sensitive tool to investigate the catalytic reactions, changes in the spectral features may be caused by multiple intertwined factors during the reaction process. Some examples of these factors are variations in particle size, electronic states of the supported metal, and the formation of intermediates. Complementary approaches like combinatorial analytical techniques and theoretical studies are needed for precise spectra interpretation. The modern catalysis research is devoted to elevating our understanding of catalytic reactions at the very molecular or atomic level. The development of in situ spectroscopy with high spaciotemporal resolution has undoubtedly provided a solid basis for mechanistic models of heterogeneous catalysis. The application of UV-Vis spectroscopy to study working heterogeneous catalysts is only limited to restricted reaction conditions due to limitations in equipment and reaction cells. For spectroscopic studies under reaction conditions, the highest workable temperature and pressure are very important to meet multiple kinetic demands of specific catalytic processes. Moreover, gradients in temperature and concentrations of reactants in the catalytic bed should be avoided. Thus, overcoming the challenges related to the experimental evaluation of mass and heat transfer in UV-Vis spectroscopy is crucial to obtain rigorous and reliable spectrokinetics of heterogeneous catalysts for industrial applications. Comparison of the kinetic parameters from spectroscopic studies with the observation from laboratory-scale experimental studies is also critical to validate the interpretation of the results – for example, investigations on the time variation of the percentage of an important intermediate.

References 1. Chen, C., Cao, Y., Liu, S., Chen, J., Jia, W.: Review on the latest developments in modified vanadium-titanium-based SCR catalysts. Chin. J. Catal. 39(8), 1347–1365 (2018) 2. Forzatti, P.: Present status and perspectives in De-NOX SCR catalysis. Appl. Catal. A Gen. 222(1–2), 221–236 (2001) 3. Segura, Y., Chmielarz, L., Kustrowski, P., Cool, P., Dziembaj, R., Vansant, E.F.: Characterisation and reactivity of vanadia–titania supported SBA-15 in the SCR of NO with ammonia. Appl. Catal. B Environ. 61(1–2), 69–78 (2005) 4. Arfaoui, J., Boudali, L.K., Ghorbel, A., Delahay, G.: Effect of vanadium on the behaviour of unsulfated and sulfated Ti-pillared clay catalysts in the SCR of NO by NH3. Catal. Today. 142(3–4), 234–238 (2009) 5. Lai, J.-K., Wachs, I.E.: A perspective on the selective catalytic reduction (SCR) of NO with NH3 by supported V2O5–WO3/TiO2 catalysts. ACS Catal. 8(7), 6537–6551 (2018) 6. Pereda-Ayo, B., De La Torre, U., Illán-Gómez, M.J., Bueno-López, A., González-Velasco, J.R.: Role of the different copper species on the activity of Cu/zeolite catalysts for SCR of NOX with NH3. Appl. Catal. B Environ. 147, 420–428 (2014) 7. Ettireddy, P.R., Ettireddy, N., Mamedov, S., Boolchand, P., Smirniotis, P.G.: Surface characterization studies of TiO2 supported manganese oxide catalysts for low temperature SCR of NO with NH3. Appl. Catal. B Environ. 76(1–2), 123–134 (2007) 8. Duffy, B.L., Curryhyde, H.E., Cant, N.W., Nelson, P.F.: 15N labeling studies of the reduction of nitric oxide by ammonia over amorphous and crystalline chromia in the presence and absence of oxygen. J. Catal. 149(1), 11–22 (1994) 9. Arnarson, L., Falsig, H., Rasmussen, S.B., Lauritsen, J.V., Moses, P.G.: A complete reaction mechanism for standard and fast selective catalytic reduction of nitrogen oxides on low coverage VOx/ TiO2(001) catalysts. J. Catal. 346, 188–197 (2017) 10. Bulánek, R., Čapek, L., Setnička, M., Čičmanec, P.: DR UV–vis study of the supported vanadium oxide catalysts. J. Phys. Chem. C. 115(25), 12430–12438 (2011) 11. Marberger, A., Ferri, D., Elsener, M., Krocher, O.: The significance of Lewis acid sites for the selective catalytic reduction of nitric oxide on vanadium-based catalysts. Angew. Chem. Int. Ed. 55(39), 11989–11994 (2016) 12. Zhu, M., Lai, J.-K., Tumuluri, U., Ford, M.E., Wu, Z., Wachs, I.E.: Reaction pathways and kinetics for selective catalytic reduction (SCR) of acidic NOx emissions from power plants with NH3. ACS Catal. 7, 8358–8361 (2017) 13. Groothaert, M.H., van Bokhoven, J.A., Battiston, A.A., Weckhuysen, B.M., Schoonheydt, R.A.: Bis(μ-oxo)dicopper in Cu-ZSM-5 and its role in the decomposition of NO: a combined in situ XAFS, UVVisnear-IR, and kinetic study. J. Am. Chem. Soc. 125(25), 7629–7640 (2003) 14. Woertink, J.S., Smeets, P.J., Groothaert, M.H., Vance, M.A., Sels, B.F., Schoonheydt, R.A., Solomon, E.I.: A [Cu 2 O] 2+ core in Cu-ZSM-5, the active site in the oxidation of methane to methanol. Proc. Natl. Acad. Sci. 106(45), 18908–18913 (2009) 15. Snyder, B.E.R., Bols, M.L., Schoonheydt, R.A., Sels, B.F., Solomon, E.I.: Iron and copper active sites in zeolites and their correlation to metalloenzymes. Chem. Rev. 118(5), 2718–2768 (2018) 16. Bates, S.A., Verma, A.A., Paolucci, C., Parekh, A.A., Anggara, T., Yezerets, A., Schneider, W.F., Miller, J.T., Delgass, W.N., Ribeiro, F.H.: Identification of the active Cu site in standard selective catalytic reduction with ammonia on Cu-SSZ-13. J. Catal. 312, 87–97 (2014) 17. Grant, J.T., Carrero, C.A., Goeltl, F., Venegas, J., Mueller, P., Burt, S.P., Specht, S.E., McDermott, W.P., Chieregato, A., Hermans, I.:

12

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

Selective oxidative dehydrogenation of propane to propene using boron nitride catalysts. Science. 354(6319), 1570–1573 (2016) 18. Chaar, M.A., Patel, D., Kung, H.H.: Selective oxidative dehydrogenation of propane over V-Mg-O catalysts. J. Catal. 109(2), 463–467 (1988) 19. Sattler, J.J.H.B., Ruiz-Martinez, J., Santillan-Jimenez, E., Weckhuysen, B.M.: Catalytic dehydrogenation of light alkanes on metals and metal oxides. Chem. Rev. 114(20), 10613–10653 (2014) 20. Pham, H.N., Sattler, J.J.H.B., Weckhuysen, B.M., Datye, A.K.: Role of Sn in the regeneration of Pt/γ-Al2O3 light alkane dehydrogenation catalysts. ACS Catal. 6(4), 2257–2264 (2016) 21. Gascón, J., Téllez, C., Herguido, J., Menéndez, M.: Propane dehydrogenation over a Cr2O3/Al2O3 catalyst: transient kinetic modeling of propene and coke formation. Appl. Catal. A Gen. 248(1–2), 105–116 (2003) 22. Carrero, C.A., Schloegl, R., Wachs, I.E., Schomaecker, R.: Critical literature review of the kinetics for the oxidative dehydrogenation of propane over well-defined supported vanadium oxide catalysts. ACS Catal. 4(10), 3357–3380 (2014) 23. Bai, P., Ma, Z., Li, T., Tian, Y., Zhang, Z., Zhong, Z., Xing, W., Wu, P., Liu, X., Yan, Z.: Relationship between surface chemistry and catalytic performance of mesoporous γ-Al2O3 supported VOX catalyst in catalytic dehydrogenation of propane. ACS Appl. Mater. Interfaces. 8(39), 25979–25990 (2016) 24. Liu, G., Zhao, Z.-J., Wu, T., Zeng, L., Gong, J.: Nature of the active sites of VOx/Al2O3 catalysts for propane dehydrogenation. ACS Catal. 6(8), 5207–5214 (2016) 25. Gao, X., Jehng, J.-M., Wachs, I.E.: In situ UV–vis–NIR diffuse reflectance and Raman spectroscopic studies of propane oxidation over ZrO2-supported vanadium oxide catalysts. J. Catal. 209(1), 43–50 (2002) 26. Gao, X., Bañares, M.A., Wachs, I.E.: Ethane and N-butane oxidation over supported vanadium oxide catalysts: an in situ UV–visible diffuse reflectance spectroscopic investigation. J. Catal. 188(2), 325–331 (1999) 27. Ovsitser, O., Cherian, M., Kondratenko, E.V.: In-situ UV/Vis and transient isotopic analysis of the role of oxidizing agent in the oxidative dehydrogenation of propane over silica-supported vanadia catalysts. J. Phys. Chem. C. 111(24), 8594–8602 (2007) 28. Schumacher, L., Hess, C.: The active role of the support in propane ODH over VOx/CeO2 catalysts studied using multiple operando spectroscopies. J. Catal. 398, 29–43 (2021) 29. Filtschew, A., Hofmann, K., Hess, C.: Ceria and its defect structure: new insights from a combined spectroscopic approach. J. Phys. Chem. C. 120(12), 6694–6703 (2016) 30. Skorodumova, N.V., Ahuja, R., Simak, S.I., Abrikosov, I.A., Johansson, B., Lundqvist, B.I.: Electronic, bonding, and optical properties of CeO2 and Ce2O3 from first principles. Phys. Rev. B. 64(11), 115108 (2001) 31. Zhao, Z., Yamada, Y., Ueda, A., Sakurai, H., Kobayashi, T.: The roles of redox and acid–base properties of silica-supported vanadia catalysts in the selective oxidation of ethane. Catal. Today. 93–95, 163–171 (2004) 32. Zhang, S., Liu, H.: Oxidative dehydrogenation of propane over MgV-O oxides supported on MgO-coated silica: structural evolution and catalytic consequence. Appl. Catal. A Gen. 573, 41–48 (2019) 33. Rybarczyk, P., Berndt, H., Radnik, J., Pohl, M.-M., Buyevskaya, O., Baerns, M., Brückner, A.: The structure of active sites in Me–V–O catalysts (Me ¼ Mg, Zn, Pb) and its influence on the catalytic performance in the oxidative dehydrogenation (ODH) of propane. J. Catal. 202(1), 45–58 (2001) 34. Kharlamova, T.S., Timofeev, K.L., Salaev, M.A., Svetlichnyi, V.A., Vodyankina, O.V.: Monolayer MgVOx/Al2O3 catalysts for propane oxidative dehydrogenation: insights into a role of structural, redox, and Acid-Base properties in catalytic performance. Appl. Catal. A Gen. 598, 117574 (2020)

281 35. Gao, X., Wachs, I.E.: Investigation of surface structures of supported vanadium oxide catalysts by UVvisNIR diffuse reflectance spectroscopy. J. Phys. Chem. B. 104(6), 1261–1268 (2000) 36. Sokolov, S., Stoyanova, M., Rodemerck, U., Linke, D., Kondratenko, E.V.: Comparative study of propane dehydrogenation over V-, Cr-, and Pt-based catalysts: time on-stream behavior and origins of deactivation. J. Catal. 293, 67–75 (2012) 37. Nijhuis, T.A., Tinnemans, S.J., Visser, T., Weckhuysen, B.M.: Towards real-time spectroscopic process control for the dehydrogenation of propane over supported chromium oxide catalysts. Chem. Eng. Sci. 59(22–23), 5487–5492 (2004) 38. Solomon, S., Plattner, G.-K., Knutti, R., Friedlingstein, P.: Irreversible climate change due to carbon dioxide emissions. Proc. Natl. Acad. Sci. 106(6), 1704–1709 (2009) 39. Shakun, J.D., Clark, P.U., He, F., Marcott, S.A., Mix, A.C., Liu, Z., Otto-Bliesner, B., Schmittner, A., Bard, E.: Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation. Nature. 484(7392), 49–54 (2012) 40. Gattrell, M., Gupta, N., Co, A.: Electrochemical reduction of CO2 to hydrocarbons to store renewable electrical energy and upgrade biogas. Energy Convers. Manag. 48(4), 1255–1265 (2007) 41. Sakakura, T., Choi, J.-C., Yasuda, H.: Transformation of carbon dioxide. Chem. Rev. 107(6), 2365–2387 (2007) 42. Zhang, X., Wu, Z., Zhang, X., Li, L., Li, Y., Xu, H., Li, X., Yu, X., Zhang, Z., Liang, Y., et al.: Highly selective and active CO2 reduction electrocatalysts based on cobalt phthalocyanine/carbon nanotube hybrid structures. Nat. Commun. 8, 14675 (2017) 43. Chen, J., Li, J., Liu, W., Ma, X., Xu, J., Zhu, M., Han, Y.-F.: Facile synthesis of polymerized cobalt phthalocyanines for highly efficient CO2 reduction. Green Chem. 21(22), 6056–6061 (2019) 44. Li, C.W., Ciston, J., Kanan, M.W.: Electroreduction of carbon monoxide to liquid fuel on oxide-derived nanocrystalline copper. Nature. 508(7497), 504–507 (2014) 45. van de Loosdrecht, J., Botes, F.G., Ciobica, I.M., Ferreira, A., Gibson, P., Moodley, D.J., Saib, A.M., Visagie, J.L., Weststrate, C.J., Niemantsverdriet, J.W.: Fischer–Tropsch synthesis: catalysts and chemistry. In: Comprehensive Inorganic Chemistry II, pp. 525–557. Elsevier (2013) 46. Lin, S., Diercks, C.S., Zhang, Y.-B., Kornienko, N., Nichols, E.M., Zhao, Y., Paris, A.R., Kim, D., Yang, P., Yaghi, O.M., et al.: Covalent organic frameworks comprising cobalt porphyrins for catalytic CO2 reduction in water. Science. 349(6253), 1208–1213 (2015) 47. Ahrenholtz, S.R., Epley, C.C., Morris, A.J.: Solvothermal preparation of an electrocatalytic metalloporphyrin MOF thin film and its redox hopping charge-transfer mechanism. J. Am. Chem. Soc. 136(6), 2464–2472 (2014) 48. Kalyuzhny, G., Vaskevich, A., Ashkenasy, G., Shanzer, A., Rubinstein, I.: UV/Vis spectroscopy of metalloporphyrin and metallophthalocyanine monolayers self-assembled on ultrathin gold films. J. Phys. Chem. B. 104(34), 8238–8244 (2000) 49. Han, N., Wang, Y., Ma, L., Wen, J., Li, J., Zheng, H., Nie, K., Wang, X., Zhao, F., Li, Y., et al.: Supported cobalt polyphthalocyanine for high-performance electrocatalytic CO2 reduction. Chem. 3(4), 652–664 (2017) 50. Kornienko, N., Zhao, Y., Kley, C.S., Zhu, C., Kim, D., Lin, S., Chang, C.J., Yaghi, O.M., Yang, P.: Metal–organic frameworks for electrocatalytic reduction of carbon dioxide. J. Am. Chem. Soc. 137(44), 14129–14135 (2015) 51. Behar, D., Dhanasekaran, T., Neta, P., Hosten, C.M., Ejeh, D., Hambright, P., Fujita, E.: Cobalt porphyrin catalyzed reduction of CO2. Radiation chemical, photochemical, and electrochemical studies. J. Phys. Chem. A. 102(17), 2870–2877 (1998) 52. Stöcker, M.: Methanol-to-hydrocarbons: catalytic materials and their behavior. Microporous Mesoporous Mater. 29(1–2), 3–48 (1999) 53. Bjorgen, M., Svelle, S., Joensen, F., Nerlov, J., Kolboe, S., Bonino, F., Palumbo, L., Bordiga, S., Olsbye, U.: Conversion of methanol to

12

282 hydrocarbons over zeolite H-ZSM-5: on the origin of the olefinic species. J. Catal. 249(2), 195–207 (2007) 54. Tian, P., Wei, Y., Ye, M., Liu, Z.: Methanol to olefins (MTO): from fundamentals to commercialization. ACS Catal. 5(3), 1922–1938 (2015) 55. Haw, J.F., Marcus, D.M.: Well-defined (supra)molecular structures in zeolite methanol-to-olefin catalysis. Top. Catal. 34(1), 41–48 (2005) 56. Haw, J.F., Song, W., Marcus, D.M., Nicholas, J.B.: The mechanism of methanol to hydrocarbon catalysis. Acc. Chem. Res. 36(5), 317–326 (2003) 57. Svelle, S., Joensen, F., Nerlov, J., Olsbye, U., Lillerud, K.-P., Kolboe, S., Bjørgen, M.: Conversion of methanol into hydrocarbons over zeolite H-ZSM-5: ethene formation is mechanistically separated from the formation of higher alkenes. J. Am. Chem. Soc. 128(46), 14770–14771 (2006) 58. Cui, Z., Liu, Q., Ma, Z., Bian, S., Song, W.: Direct observation of olefin homologations on zeolite ZSM-22 and its implications to methanol to olefin conversion. J. Catal. 258(1), 83–86 (2008) 59. Cui, Z.-M., Liu, Q., Song, W.-G., Wan, L.-J.: Insights into the mechanism of methanol-to-olefin conversion at zeolites with systematically selected framework structures. Angew. Chem. Int. Ed. 45(39), 6512–6515 (2006) 60. Song, W., Fu, H., Haw, J.F.: Selective synthesis of methylnaphthalenes in HSAPO-34 cages and their function as reaction centers in methanol-to-olefin catalysis. J. Phys. Chem. B. 105(51), 12839–12843 (2001) 61. Hereijgers, B.P.C., Bleken, F., Nilsen, M.H., Svelle, S., Lillerud, K.P., Bjørgen, M., Weckhuysen, B.M., Olsbye, U.: Product shape selectivity dominates the methanol-to-olefins (MTO) reaction over H-SAPO-34 catalysts. J. Catal. 264(1), 77–87 (2009) 62. Schoonheydt, R.A.: UV-VIS-NIR spectroscopy and microscopy of heterogeneous catalysts. Chem. Soc. Rev. 39(12), 5051 (2010) 63. Hemelsoet, K., Qian, Q., De Meyer, T., De Wispelaere, K., De Sterck, B., Weckhuysen, B.M., Waroquier, M., Van Speybroeck, V.: Identification of intermediates in zeolite-catalyzed reactions by in situ UV/Vis microspectroscopy and a complementary set of molecular simulations. Chem. A Eur. J. 19(49), 16595–16606 (2013) 64. Goetze, J., Yarulina, I., Gascon, J., Kapteijn, F., Weckhuysen, B.M.: Revealing lattice expansion of small-pore zeolite catalysts during the methanol-to-olefins process using combined operando X-ray diffraction and UV–vis spectroscopy. ACS Catal. 8(3), 2060–2070 (2018) 65. Goetze, J., Meirer, F., Yarulina, I., Gascon, J., Kapteijn, F., RuizMartínez, J., Weckhuysen, B.M.: Insights into the activity and deactivation of the methanol-to-olefins process over different small-pore zeolites as studied with operando UV–vis spectroscopy. ACS Catal. 7(6), 4033–4046 (2017) 66. Wojdyr, M.: Fityk: a general-purpose peak fitting program. J. Appl. Crystallogr. 43(5), 1126–1128 (2010) 67. Hemelsoet, K., Nollet, A., Vandichel, M., Lesthaeghe, D., Van Speybroeck, V., Waroquier, M.: The effect of confined space on the growth of naphthalenic species in a chabazite-type catalyst: a molecular modeling study. ChemCatChem. 1(3), 373–378 (2009)

Z. Yang and M. Zhu 68. Lesthaeghe, D., De Sterck, B., Van Speybroeck, V., Marin, G.B., Waroquier, M.: Zeolite shape-selectivity in thegem-methylation of aromatic hydrocarbons. Angew. Chem. Int. Ed. 46(8), 1311–1314 (2007) 69. McCann, D.M., Lesthaeghe, D., Kletnieks, P.W., Guenther, D.R., Hayman, M.J., Van Speybroeck, V., Waroquier, M., Haw, J.F.: A complete catalytic cycle for supramolecular methanol-to-olefins conversion by linking theory with experiment. Angew. Chem. Int. Ed. 47(28), 5179–5182 (2008) 70. Maihom, T., Boekfa, B., Sirijaraensre, J., Nanok, T., Probst, M., Limtrakul, J.: Reaction mechanisms of the methylation of ethene with methanol and dimethyl ether over H-ZSM-5: an ONIOM study. J. Phys. Chem. C. 113(16), 6654–6662 (2009) 71. Van Speybroeck, V., Hemelsoet, K., De Wispelaere, K., Qian, Q., Van der Mynsbrugge, J., De Sterck, B., Weckhuysen, B.M., Waroquier, M.: Mechanistic studies on chabazite-type methanol-toolefin catalysts: insights from time-resolved UV/Vis microspectroscopy combined with theoretical simulations. ChemCatChem. 5(1), 173–184 (2013) 72. Jaumot, J., de Juan, A., Tauler, R.: MCR-ALS GUI 2.0: new features and applications. Chemom. Intell. Lab. Syst. 140, 1–12 (2015) 73. Maeder, M., Zilian, A.: Evolving factor analysis, a new multivariate technique in chromatography. Chemom. Intell. Lab. Syst. 3(3), 205–213 (1988) 74. Lesthaeghe, D., Horré, A., Waroquier, M., Marin, G.B., Van Speybroeck, V.: Theoretical insights on methylbenzene side-chain growth in ZSM-5 zeolites for methanol-to-olefin conversion. Chem. A Eur. J. 15(41), 10803–10808 (2009) 75. Li, J., Xiong, G., Feng, Z., Liu, Z., Xin, Q., Li, C.: Coke formation during the methanol conversion to olefins in zeolites studied by UV Raman spectroscopy. Microporous Mesoporous Mater. 39(1–2), 275–280 (2000) 76. Dehertog, W.J.H., Froment, G.F.: Production of light alkenes from methanol on ZSM-5 catalysts. Appl. Catal. 71(1), 153–165 (1991) 77. Mores, D., Stavitski, E., Kox, M.H.F., Kornatowski, J., Olsbye, U., Weckhuysen, B.M.: Space- and time-resolved in-situ spectroscopy on the coke formation in molecular sieves: methanol-to-olefin conversion over H-ZSM-5 and H-SAPO-34. Chem. A Eur. J. 14(36), 11320–11327 (2008) 78. Mores, D., Kornatowski, J., Olsbye, U., Weckhuysen, B.M.: Coke formation during the methanol-to-olefin conversion: in situ microspectroscopy on individual H-ZSM-5 crystals with different brønsted acidity. Chem. A Eur. J. 17(10), 2874–2884 (2011) 79. Yarulina, I., De Wispelaere, K., Bailleul, S., Goetze, J., Radersma, M., Abou-Hamad, E., Vollmer, I., Goesten, M., Mezari, B., Hensen, E.J.M., et al.: Structure–performance descriptors and the role of Lewis acidity in the methanol-to-propylene process. Nat. Chem. 10(8), 804–812 (2018) 80. Goetze, J., Weckhuysen, B.M.: Spatiotemporal coke formation over zeolite ZSM-5 during the methanol-to-olefins process as studied with operando UV-vis spectroscopy: a comparison between H-ZSM-5 and Mg-ZSM-5. Cat. Sci. Technol. 8(6), 1632–1644 (2018) 81. Schulz, H.: “Coking” of zeolites during methanol conversion: basic reactions of the MTO-, MTP- and MTG processes. Catal. Today. 154(3), 183–194 (2010)

12

Case Studies: Ultraviolet-Visible (UV-Vis) Spectroscopy

283

12

Zixu Yang obtained his BS and MS degrees from Huazhong University of Science and Technology, China in 2009 and 2012, and Ph.D. degree from Oklahoma State University, USA in 2016. He worked as a postdoctoral researcher at Washington State University, USA from 2016 to 2017. He works now as a lecturer at East China University of Science and Technology in China, focused on heterogeneous catalysis and reaction engineering to address the challenges in converting fossil and biomass feedstocks to fuels and chemicals.

Minghui Zhu obtained his BS degree from Zhejiang University, China in 2011 and Ph.D. degree from Lehigh University, USA in 2016. He worked as a postdoctoral researcher at Massachusetts Institute of Technology, USA. He works now as a Professor at East China University of Science and Technology in China, focused on heterogeneous catalysis.

Fluorescence Microscopy

13

Xianwen Mao, Rong Ye, and Peng Chen

Contents 13.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

13.2

Reactivity and Heterogeneity of Individual Particles . . . 286

13.3

Restructuring and Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

13.4

Super-resolution Mapping of Catalytic Activities at the Single to Subparticle Level . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

13.5

Scalable Parallel Screening of Catalyst Activities . . . . . . 290

13.6

Spatial and Temporal Catalysis Cooperativity Within and Between Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292

13.7

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293

Abstract

Heterogeneous catalysts exhibit intrinsic heterogeneities, both structurally and compositionally. For example, the size and shape of nanoparticle catalysts often show dispersions, and could also change over time during reactions. Therefore, it is important to study heterogeneous catalysts with experimental tools that allow in situ, realtime, spatially resolved characterization of catalytic activities. Single-molecule fluorescence microscopy has recently emerged as a powerful tool with the abovementioned capabilities. In this chapter, we discuss the development and application of single-molecule fluorescence microscopy for characterizations of heterogeneous catalysts at the single-particle to subparticle level, covering topics ranging from the static/dynamic activity heterogeneities of individual catalyst particles and subparticle regions, to the scale-up ability of catalyst screening, and to the catalysis cooperativity between spatially distinct locations. X. Mao · R. Ye · P. Chen (*) Department of Chemistry and Chemical Biology, Cornell University, Ithaca, NY, USA e-mail: [email protected]; [email protected]; [email protected]

Keywords

Catalysis · Nanoparticle · Single molecule · Optical microscopy

13.1

Introduction

A myriad of developments in the catalysis science stem from research on model catalysts that have uniform catalytic sites, e.g., single crystal catalysts or ordered oxides, via modern surface-science approaches in ultrahigh vacuum [1–3]. However, the composition and structure of real-world heterogeneous catalysts, most of which are nanoparticles (NPs), have inherent heterogeneity, so their properties might not be accurately represented by the model catalysts in the ideal system. Moreover, the structures of NP catalysts are dependent on the synthesis and activation conditions (static disorder) and often change under reaction conditions (dynamic disorder), making it challenging to study the properties of individual NPs. Recent advances in single-molecule fluorescence microscopy promise long-awaited progress in the understanding of heterogeneous NP catalysts [4–9]. Single-molecule fluorescence microscopy detects fluorescence signals involved in catalytic fluorogenic reactions, i.e., reactions that generate fluorescent product molecules (Fig. 13.1a), for example, in a wide-field imaging format (Fig. 13.1b) [10]. This technique determines the time and location of single catalytic turnovers on single NPs (Fig. 13.1b, c) with nanometer precision in a high throughput manner in real time while the catalysts are working, i.e., in operando. It can, in a straightforward manner, elucidate the vital roles of both static and dynamic heterogeneity of catalysts, offering a way to approach the complexity of realistic working catalysts [11]. The image of an infinitely small bright object, even in focus, is not infinitely small. Instead, it is a circular Airy diffraction image – a central bright disk and progressively weaker

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_13

285

286

X. Mao et al.

–O

Resazurin O O N O

Resorufin –O O

b)

c)

O

Counts

a)

N

Au nanoparticle

ton

60k 40k toff 20k 0 1930

Glass slide

1940

1950

1960

Time (s)

5 µm

532 nm

e)

d)

f)

Prism

10

Objective Filter

No. of NPs

532 nm

5

0 0.4

0.2

1000 nm EMCCD

Mirror

geff (s–1)

Fig. 13.1 Single-turnover detection of single Au NP catalyzing the transformation of nonfluorescent resazurin to fluorescent resorufin. (a) Schematics of the experimental setup using total internal reflection fluorescence microscopy. (b) A representative image of fluorescence products generated on 6-nm Au NPs in operando. (c) A segment of the fluorescence trajectory

from the fluorescence spot indicated by the arrow in b. (d) A zoomed-in view of a typical fluorescence spot. The center position of this molecule can be determined with nanometer accuracy as labeled by the red cross mark. (e) Schematics of a typical imaging experimental setup. (f) Distribution of γeff from many gold NPs. (Adapted with permission from Xu et al. [10])

concentric dark and bright rings. The radius R of the first dark ring around the central disk of the Airy diffraction image defines the resolution of an optical microscope: R ¼ 0.61λ/ NA, where λ is the wavelength of the illuminating light and NA is the numerical aperture of the microscope objective. The resolution is typically 200–400 nm for visible light, and resolution below this theoretical limit is called super-resolution. Single-molecule fluorescence microscopy enables the localization of a single fluorescent product molecule in a snapshot at nanometer precision. This is done through fitting the point spread function (the intensity profile of the molecule emitting fluorescence in the image) with a 2D Gaussian function, and therefore achieves super-resolution imaging – improving the imaging resolution down to a few nanometers (Fig. 13.1b). Such super-resolution localization of fluorescent molecules requires that these molecules are more than ~200 nm away from each other in a given image frame, which is often achievable for studying heterogeneous catalysis due to the slow and stochastic nature of product formation on the catalyst surface and relatively fast desorption. The exact resolution is affected mainly by signal/noise ratio and can be improved by increasing the fluorescence intensity and decreasing the background noise. The single-molecule fluorescence approach provides a myriad of advantages in studying nanoparticle catalysis: (I) real-time sampling with single-reaction temporal

resolution – the best meaningful sensitivity, (II) singleparticle (and occasionally single-reactive-site) resolution, (III) tens of nanometers down to a few nanometers in spatial precision from super-resolution imaging, (IV) chemical specificity based on the selection of fluorescence wavelength, (V) parallel sampling of many particles simultaneously from wide-field imaging, (VI) in situ characterizations under ambient solution conditions without sample destruction from the optical detection, and (VII) continuous, steadystate reaction conditions via a continuous supply of reactant solutions into a flow cell [4–9]. This chapter discusses the application of single-molecule fluorescence microscopy in the characterization of heterogeneous NP catalysts using works in our group as examples. The instrumentation setup using total internal reflection fluorescence microscopy is briefly illustrated in Fig. 13.1e and covered in detail elsewhere [12–14]. Interested readers are referred to recent reviews [4–7, 15, 16] for other relevant works.

13.2

Reactivity and Heterogeneity of Individual Particles

The real-time single-turnover analysis of single-particle catalysis can be achieved via the trajectory of the emission intensity of fluorescent products on a single particle

13

Fluorescence Microscopy

287

(Fig. 13.1c), e.g., resorufin produced from the gold-NPcatalyzed reduction of resazurin. The waiting time between or during the stochastic fluorescence bursts, τoff or τon, respectively, contain information of reaction kinetics of individual NPs. The product formation rate hτoffi1 of a single NP, obtained by averaging over many events, characterizes the mean activity of that NP. hτoffi1 is a function of the substrate concentration [S], following the singlemolecule Langmuir– Hinshelwood equation: hτoff i

1

¼

γ eff K ½S 1 þ K ½ S

ð13:1Þ

where γeff can be understood as a rate constant that represents the combined reactivity of all surface catalytic sites on one NP and has the unit of inverse time, and K is the reactant adsorption equilibrium constant and has the unit of inverse concentration. Taking advantage of Eq. (13.1), one can measure the activity of many individual NPs at various substrate concentrations, compute γeff of each NP, and obtain the distribution among many NPs in a batch of a sample. The differences of γeff among different NPs in the same batch (Fig. 13.1f) quantifies the heterogeneity (static disorder) of the catalysts [17].

13.3

Restructuring and Switching

In addition to the heterogeneity among different NP catalysts, single-molecule fluorescence microscopy can readily identify and quantify the restructuring of the catalysts under reaction conditions. Restructuring, which leads to the dynamic disorder of catalytic activity, is the temporal fluctuations of the catalytic activity of a single NP. Even though in situ TEM or scanning probe microscopy [18–21] can visualize the morphology change of NP catalysts, the quantification of the restructuring timescale or activation energy is very difficult to be characterized via these techniques. An intuitive way to evaluate the restructuring of a single NP is to monitor its activity (time-binned rate of turnover) over a period (Fig. 13.2a inset) [10]. However, the stochastic nature of single turnover behaviors and the insufficient statistics of activity contained in each binned data point compromise the robustness of such evaluation. Instead, a reliable and quantitative analysis for the temporal fluctuations of activity is the autocorrelation function Cτ-off (m) of the waiting time τoff of catalytic turnovers, in other words, the correlation of a turnover event (i.e., waiting time τoff) with a later event as a function of delay time in the units of turnover index m along the single-molecule turnover trajectories (Fig. 13.2a) [10]. The exponential decay of the

autocorrelation function clearly indicates the restructuringinduced activity fluctuations of the NP catalysts, and the exponential time constant of this function represents the timescale of the underlying restructuring. Such temporal fluctuations of activity originate from the dynamic surface restructuring, which is related to the catalytic turnovers, as the adsorbate-surface interactions during catalysis can induce surface restructuring. Relatedly, the dynamic surface restructuring can occur spontaneously when a particle becomes small enough so that the surface energy is high and, concurrently, the activation energy of restructuring is lower to be comparable to the thermal energy (~2.5 kJ/mol at room temperature). As the rate of activity fluctuation indicates the surface restructuring rate, it is not surprising that such fluctuation slows down when the catalytic turnover declines (e.g., from lower substrate concentrations), or when the NP is bigger in size (Fig. 13.2b) [22]. The rate of fluctuation can be correlated with the catalytic turnover rate and the diameter of the NP based on a thermodynamic model, from which one can calculate the activation energy and spontaneous surface restructuring rate. Predictably, the activation energy of restructuring increases as the size of the NP is bigger, while the restructuring rate shows the opposite trend (Fig. 13.2c) [22]. In addition, the fluctuation rate is an intrinsic property of a given material, i.e., independent of the probe catalytic reactions, but varies among different materials. Consequently, the fluctuation rates of single ~5 nm Pt NPs are the same when catalyzing an N-deoxygenation or an N-deacetylation reaction, but is slower than single Au NPs of the same size when catalyzing the same Ndeoxygenation reaction (Fig. 13.2d) [23]. Single-molecule fluorescence microcopy can further reveal the [S]-dependent surface switching, between a low catalytic activity and a high activity state at a specific substrate concentration, a phenomenon caused by the adsorbate-surface interactions and likely adsorbateadsorbate interactions on the NP surface. Such [S]-dependent dynamic surface switching is demonstrated by the 1 statistics of variances of τ1 off and τon from trajectories of single NPs (Fig. 13.2e–g) [17], as they quantify the amplitudes of the temporal fluctuations of the corresponding NPs. When [S] is small, the two-dimensional histogram of 1 the variances of τ1 off and τon among many Au NPs in catalyzing the resazurin reduction reaction has one population (type-a, Fig. 13.2e). A portion of the type-a population switch to a different population at larger Var(τ1 off ) and Var 1 (τon ) (type-b, Fig. 13.2f) as [S] increases. When [S] reaches a high-enough value, all type-a NPs switch to type-b (Fig. 13.2g). Comparing the properties, e.g., γeff and K defined in Eq. 13.1, of NPs in these two populations, type-b has larger γeff (higher reactivity) and smaller K (weaker substrate binding) compared with type-a.

13

288

X. Mao et al.

Ct-off (m)

0.6

1.2 0.8

0.12

0.4 0.0 0

0.3

1000 2000 Time (s)

3000

0.06

0.00 0.2 s–

0.0

0.1

1)

10

0.0

40

d)

6

0.03 4

e)

20

30

f)

2

a [S] = 0.05 PM 2 1 Var (toff–1)

Super-resolution Mapping of Catalytic Activities at the Single to Subparticle Level

As discussed earlier in the Introduction, single-molecule fluorescence microscopy allows the localization of a single fluorescent product molecule at nanometer precision (typically about 5–40 nm depending on the signal-to-noise ratio) [4, 15, 24], thus enabling super-resolution mapping of catalytic activities within a single nanostructure, ranging from

0.02 0.04 Turnover rate (s–1 particle–1)

g) 3

Var (ton–1)

3

38 32 26 20 14 8 2

6

18

14 d (nm)

0.03

0.00 0.00

0.00

10

Au, deoxygenation Pt, deoxygenation Pt, deacetylation

0.06

2

33 29 25 21 17 13 9 5 1

3 b a

1

Var (ton–1)

10 d (nm)

5

Var (ton–1)

rsp (s–1)

'Esp (kcal mol–1)

0.06

Fluctuation rate (s–1)

c)

20 30 m, turnover index

1

13.4

toff reaction

b)

r (s–1)

Turnover rate (s–1 particle–1)

a)

v(

Fig. 13.2 Spontaneous and catalysis-induced restructuring and switching of single Au NPs. (a) Autocorrelation function of waiting time τoff during the reduction of resazurin at a saturating concentration catalyzed by a single 6 nm Au NP. Solid line: exponential fit. Inset: the corresponding time-binned turnover rate vs. time of this NP; each data point presents the mean of 10 turnover events. (b) Surface restructuring rate (r, equivalent to the activity fluctuation rate) vs. the single-NP turnover rate (v) and the NP diameter (d ) for pseudospherical Au NPs. Meshed surface: a fit based on a thermodynamic model. (c) Activation energy (ΔEsp) and the rate (rsp) of spontaneous dynamic surface restructuring vs. the diameter of pseudospherical Au NPs. Gray regions: the approximate errors. (d) Activity fluctuation rate vs. single-NP turnover rate of 4.6 nm Pt and Au NPs in catalyzing an Ndeoxygenation and/or an Ndeacetylation reaction. (e–g) Two-dimensional histograms of 1 the variances of τ1 off and τ on of individual 6 nm Au NPs at various concentrations of resazurin [S]. (Adapted with permission from Zhou et al [22] and from Xu et al. [17])

2

20 17 13 9 5 1

b

1

[S] = 1.2 PM

[S] = 0.07 PM 1

2 Var (toff–1)

1

2 Var (toff–1)

simple nanoparticles such as pseudospherical particles, nanorods, and nanoplates, to more complicated structures such as linked nanoparticle-nanorod or nanorod-nanorod bimetallic nanostructures. These spatially resolved activity patterns obtained under operando conditions, when correlated with particle sizes/shapes determined accurately from electron microscopic techniques, yield insights into catalytic performance at a single particle or subparticle level. Such information is unattainable from conventional ensemble measurements wherein a mixture of catalyst particles with inevitably varying sizes and shapes are employed.

13

Fluorescence Microscopy

a)

289

d)

100 nm

v (×10–6 s–1nm–2)

f)

4 3 2 1 Surface

Specific turnover rate (×10–6 s–1nm–2)

200 nm

c)

200 nm

e)

Edge

Corner

2.4 2.0 1.6 0

b R (×10–11 s–1nm–4)

b)

3 r2 (×104 nm2)

6

3 R (×102 nm)

4

10 5 0 2

Fig. 13.3 (a) TEM image of a Au nanoplate encapsulated in a mesoporous silica shell. (b) Spatial mapping of product molecules overlaid with the SEM image of a single Au@mSiO2 nanoplate. mSiO2: mesoporous silica shell. Every dot represents a single fluorescent product molecule, and is color-coded based on the region where it is generated. (c) Specific turnover rate v for the three different types of regions of the Au@mSiO2 nanoplate shown in b. (d) Spatial mapping of product molecules at different radial segments of the top facet, represented by different colors, overlaid with the SEM image of the single Au@mSiO2 nanoplate shown in b. (e) Specific turnover rate of different radial segments vs. r2, where r is defined as the distance along the center to corner vector of the nanoplate. (f) Radial activity gradient βR vs. the radius of a nanoplate R, defined along the center to corner vector. (Adapted with permission from Andoy et al. [25])

For example, single-molecule super-resolution catalysis imaging revealed site-specific activities within a single Au nanoplate (a transmission electron microscopy (TEM) image is shown Fig. 13.3a) in catalyzing an N-deoxygenation reaction: the corner regions are the most catalytically active, followed by the edge regions and then the flat facets (Fig. 13.3b, c) [25]. The mesoporous silica shell stabilizes Au nanoplates in solution and improves detection efficiency for fluorescent product molecules by temporarily trapping them at locations away from the Au surface. Such site-specific activity patterns likely result from the more active, low-coordination surface sites, the population of which increases from the flat facets to the corner region and to the edge regions. Furthermore, this catalysis imaging technique with high spatial resolutions enables the examination of the activity variation within a single facet (i.e., at a subparticle level). As shown in

Fig. 13.3d (a scanning electron microscopy (SEM) image of a Au nanoplate) and Fig. 13.3e, the specific activity within the flat facet of the Au nanoplate exhibits a pronounced radial gradient for catalyzing the N-deoxygenation reaction: the specific turnover rate decreases from the center to the periphery [25]. Such observed activity gradients within the same surface facets of a single nanocrystal catalyst could be explained by the density gradients of the low-coordination defect sites within a single facet, which arise possibly due to the decreasing growth rate during the solution-phase synthesis of the Au nanoplates via a seeded growth mechanism. Additionally, these activity gradients also show strong particle size dependence (Fig. 13.3f): larger nanoplates exhibit smaller specific activity gradients. Such observations suggest that the surface defect site densities of larger particles have a more gradual variation from the center region to the periphery. Single-molecule super-resolution catalysis imaging can also be exploited for direct quantification of catalytic activity for bimetallic nanoparticles, which are of increasing research interest because they usually exhibit enhanced catalytic performance in terms of the activity, stability, and selectivity compared with their monometallic counterparts [26]. The first quantitative visualization of enhanced bimetallic activity was demonstrated in a study by Chen et al. [27], in which heteronuclear bimetallic PdAu nanoparticles with well-defined Pd-Au interfaces (Fig. 13.4a, b) were used as a model bimetallic catalyst system. Enhanced catalytic activity toward catalyzing the disproportionation of resazurin to generate the fluorescent resorufin as one of the products was directly observed near the Pd-Au interface using super-resolution single-molecule catalysis imaging (Fig. 13.4b, upper panel), demonstrating a direct, first-ofits-kind visualization of bimetallic effect for catalytic enhancement. Individual bimetallic particles can be dissected into four regions (Fig. 13.4b, lower panel): monometallic Pd (Pd), monometallic Au (Au), Pd-doped Au near the Pd-Au interface (AuPd), and Au-doped Pd near the Pd-Au interface (PdAu). The specific catalytic rates of all four regions display saturation kinetics with increasing substrate concentration (Fig. 13.4c). The catalytic rate constant (which is equivalent to the rate at the saturation regime) exhibits the following order for the four subparticle regions: PdAu > Pd > AuPd > Au. Additionally, catalytic hotspots on plasmonic nanostructures, such as linked Au-Au nanorod structures (Fig. 13.5b) and linked Au-Ag nanorod-nanoparticle structures (Fig. 13.5d), can be visualized and quantitatively examined by super-resolution single-molecule catalysis imaging [28]. Surface-plasmon-induced catalytic enhancement toward the resazurin reduction reaction was clearly observed at the Au-Au or Au-Ag gap regions compared with the nongap regions (Fig. 13.5a, c).

13

290

X. Mao et al.

a)

b)

c)

Super-resolution catalysis mapping

Electron microscopy

Sub-particle kinetic titration

Specific TOR

PdAu 200 nm Pd

PdAu

Pd AuPd Au

200 nm AuPd

Au

[Substrate]

No. of products 0

20

Fig. 13.4 (a) SEM image of a single PdAu nanoparticle (Pd nanorod with a pseudospherical Au particle grown at the tip) coated with a mesoporous silica shell. (b) Upper panel: Two-dimensional histogram of the positions of product molecules on the nanoparticle shown in a (the outer and inner white lines, determined based on the SEM image in a, represent the structural contours of the entire PdAu@mSiO2 particle and the Au nanoparticle,

b) 30 20 10

No. of products

a)

0

c)

d)

25

No. of products

50

0

Fig. 13.5 (a) Mapping of catalytic product molecules on a linked Au-Au nanorod nanostructure. White line: structural contour from b. (b) SEM image of the nanostructure in a. (c) Mapping of catalytic product molecules on a linked Au-Ag nanorod-nanoparticle nanostructure. White line: structural contour form d. (d) SEM image of the nanostructure in c. The red and black circles of 70 nm radii in (a) – (d) define the gap and nongap regions, respectively. All scale bars: 200 nm. (Adapted with permission from Zou et al. [28])

40

respectively). Lower panel: dissection of a typical PdAu nanoparticle into segments of monometallic Pd (Pd), monometallic Au (Au), Pd-doped Au near the Pd-Au interface (AuPd), and Au-doped Pd near the Pd-Au interface (PdAu). (c) Specific turnover rate (TOR) vs. [Rz] (i.e., the substrate) for the four dissected regions (Pd, Au, AuPd, PdAu) of the PdAu nanoparticle in a. (Adapted with permission from Chen et al. [27])

13.5

Scalable Parallel Screening of Catalyst Activities

Because of their ability to implement wide-field imaging, single-molecule catalysis imaging techniques can be used to screen the catalytic performance of a large quantity of nanoparticles simultaneously, enabling, for example, immediate identification of high-performance catalyst particles and collection of important statistical information that might provide useful guidelines for catalyst design and preparation [29]. A single-molecule super-resolution catalysis image for about 1000 Au nanoparticles in catalyzing the resazurin reduction reaction is shown in Fig. 13.6a, which is directly correlated with the corresponding SEM image (Fig. 13.6b). Single particles with quantifiable activities can be unambiguously identified (see Fig. 13.6a, b, insets). Statistical analysis of the entire set of ~1000 particles provides information on the particle size dependence of the catalytic activities (Fig. 13.6c): generally larger particles exhibit stronger activities on a per-particle basis, and two subpopulations can be clearly resolved (Fig. 13.6d). Such a screening approach could potentially be scaled up to be applicable to simultaneous examination of many thousands of catalyst particles using automated microscopes and large camera formats, which in turn could accelerate the discovery and development of novel catalytic systems.

13

Fluorescence Microscopy

291

a)

c) 0 2

15

3

4

Log (v)

No. of molecules

1 10

5

–1

13

3 Pm –2

0

10 Pm

100

5 200

6

7 300

SEM diameter (nm)

d)

0 No. of particles

b) 2

Log (v)

1

3 Pm

–1

40

20 0

–2

10 Pm

100

200

300

SEM diameter (nm)

Fig. 13.6 (a) Image of catalytic product molecules for about 1000 Au@mSiO2 particles in catalyzing the reductive N-deoxygenation reaction of resazurin. Inset: zoom-in of the red rectangle. (b) SEM image of the same catalyst particles in a (c) Scatter plot of the turnover rate on a

per-particle basis (v, in log scale) of individual particles vs. their diameters determined by SEM. Each dot denotes one particle. Insets: SEM images of selected particles. (d) Contour plot of the two-dimensional histogram in c. (Adapted with permission from Zhou et al. [29])

While to date the single-molecule catalysis imaging can only be applied to fluorogenic reactions, it is possible to obtain useful information on the performance of the catalyst particles in catalyzing nonfluorogenic reactions using this imaging technique, if an activity correlation between the fluorogenic and nonfluorogenic reaction can be established from conventional ensemble measurements. For instance, for Au and Au@mSiO2 particles, a linear activity correlation was observed between the fluorogenic resazurin reduction reaction and the nonfluorogenic 4-nitrophenol reduction reaction (Fig. 13.7a), and was also observed between the fluorogenic amplex red oxidation reaction and the nonfluorogenic hydroquinone oxidation reaction (Fig. 13.7b) [29]. Such strong correlations indicate that for the nonfluorogenic reactions, one could use the corresponding, activity-correlatable fluorogenic probe reaction to acquire equivalent information for assessing the catalyst performance. Notably, a very recent study has introduced a “COMPEITS” (competition-enabled imaging technique with super-resolution) technique [30] that enables

super-resolution imaging of nonfluorescent reactions via competition. This technique has overcome a long-standing limitation in single-molecule catalysis imaging, and more generally in super-resolution optical imaging, of being only applicable to fluorescent processes. The COMPEITS technique is based on the incorporation of competitive chemistry into a single-molecule fluorescence detection scheme, and could, in principle, be exploited to visualize a variety of nonfluorescent processes at nanometer resolution, in situ and in operando. For example, COMPEITS can be used to image a nanoparticle-catalyzed nonfluorescent reaction, via using a fluorogenic auxiliary reaction that competes for the same surface site on the nanoparticle catalyst where the nonfluorescent reaction occurs. The fluorogenic reaction can be imaged at super-resolution, and introduction of the competing nonfluorescent reaction leads to suppression of fluorogenic reaction rate. The degree of suppression can, therefore, be imaged at the same nanometer resolution, offering super-resolution spatial information of the target nonfluorescent reaction.

292

X. Mao et al.

Spatial and Temporal Catalysis Cooperativity Within and Between Nanoparticles

Allosteric effects, wherein the occurrence of a reaction at one site of a macromolecule influences the reaction at a remote site of some distance away, are commonly observed for biological systems and processes (e.g., enzymatic catalysis). A recent study by Zou et al. [31] suggests that such allosteric effects are also present for nonbiological nanoparticle catalysts. These authors found that, using the single-molecule catalysis imaging approach, the catalytic reactions on a single metal-based nanoparticle are positively correlated with each other, with a temporal memory of ~10 –100 s and over a distance of hundreds of nanometers. Dissecting of a singlemolecule catalysis image for a Pd nanorod, one of the model catalytic systems investigated in their study, into different, Fig. 13.7 (a) Catalytic activity correlation between the reductive N-deoxygenation reaction of resazurin (i.e., RZ-rdx) and the reduction of 4-nitrophenol to 4-aminophenol (i.e., NIP-rdx). (b) Catalytic activity correlation between the oxidative N-deacetylation reaction of amplex red (i.e., AR-ox) and the oxidation of hydroquinone to quinone (i.e., HQ-ox). The catalytic activities were quantified by the turnover rates v. (Adapted with permission from Zhou et al. [29])

spatially distinct regions (Fig. 13.8a) allows for establishing spatial correlation between reactions that occur at different regions within the same catalyst particle. Additionally, the time sequence of each reaction can also be obtained from every subparticle region. An important discovery is that the intraparticle catalytic communication strength, quantified through a correlation analysis, starts with a positive value first, followed by an exponential decay with increasing intraparticle distance or longer time separation between temporally subsequent reactions (Fig. 13.8b). Based on these trends, the physical picture is that a fast reaction at one subparticle region tends to be followed by another fast reaction nearby, giving rise to a positive cooperativity between these temporally neighboring reactions at different locations. Such intraparticle catalytic communication has been identified to result from the surface migration of positively charged messenger species, likely a surface localized hole that, for

a)

b) 10M 5c

vNIP-rdx (s–1 particle–1)

1G

5c

3c

1M

5 6 4

2c 1c

3

1k 2 1

1

1m

5 6 4

10k 3

10 1

10m 1µ

a)

1m

1k

1

b) Communication strength

20 No. of products

2c 1c

2

1µ

1M

1m

vRZ-rdx (s–1 particle–1)

'xij 15 10

4c 3c

4c vHQ-ox (s–1 particle–1)

13.6

j i

5

Time separation (s) 0

1

1k

vAR-ox (s–1 particle–1)

30

c)

60

0.08

u

Pd

0.04

200 nm 0

Fig. 13.8 (a) Dissection of a super-resolution catalysis image of a mSiO2coated Pd nanorod into several subparticle regions with widths of ~100 nm. Green line: structural contour from SEM. (b) Catalytic communication strength within a single Pd nanorod as a function of the intraparticle

0

or A

600 300 Distance separation (nm)

distance separation or the time separation between temporally subsequent product-generation events. (c) Schematic illustration of the intraparticle catalytic communication mechanism due to the presence of positively charged messenger species. (Adapted with permission from Zou et al. [31])

13

Fluorescence Microscopy

example, could reside on a surface metal oxide cluster (Fig. 13.8c). This study also demonstrates similar communication between individual catalyst particles, while it originates from a molecular diffusion mechanism involving negatively charged product molecules and its communication distance is about several micrometers, much larger compared with that for intraparticle communication.

13.7

Conclusion

In this chapter, we have provided a pedagogical monograph on the development and application of single-molecule fluorescence microscopy for characterizations of heterogeneous catalysts at the single-particle to subparticle level, covering important topics ranging from static/dynamic heterogeneities of individual catalyst particles, to the scale-up ability of catalyst screening, and to catalysis cooperativity between spatially distinct locations. Acknowledgments We thank the following agencies that supported our research in the general area of single-molecule single-particle catalysis: Army Research Office (grant no. W911NF-17-1-0590 and. W911NF-18-1-0217), and the US Department of Energy, Office of Science, Basic Energy Sciences, Catalysis Science Program (award no. DE-SC0004911, and as part of the Center for Alkaline-based Energy Solutions (CABES), an Energy Frontier Research Center, award no. DE-SC0019445).

References 1. Ye, R., Hurlburt, T.J., Sabyrov, K., Alayoglu, S., Somorjai, G.A.: Molecular catalysis science: perspective on unifying the fields of catalysis. Proc. Natl. Acad. Sci. 113, 5159–5166 (2016) 2. Ertl, G.: Reactions at surfaces: from atoms to complexity (nobel lecture). Angew. Chem. Int. Ed. 47, 3524–3535 (2008) 3. Somorjai, G.A., Li, Y.: Introduction to Surface Chemistry and Catalysis. Wiley, Hoboken, New Jersey, U.S (2010) 4. Chen, P., Zhou, X., Andoy, N.M., Han, K.-S., Choudhary, E., Zou, N., Chen, G., Shen, H.: Spatiotemporal catalytic dynamics within single nanocatalysts revealed by single-molecule microscopy. Chem. Soc. Rev. 43, 1107–1117 (2014) 5. De Cremer, G., Sels, B.F., De Vos, D.E., Hofkens, J., Roeffaers, M.B.J.: Fluorescence micro(spectro)scopy as a tool to study catalytic materials in action. Chem. Soc. Rev. 39, 4703–4717 (2010) 6. Chen, T., Dong, B., Chen, K., Zhao, F., Cheng, X., Ma, C., Lee, S., Zhang, P., Kang, S.H., Ha, J.W., Xu, W., Fang, N.: Optical superresolution imaging of surface reactions. Chem. Rev. 117, 7510–7537 (2017) 7. Tachikawa, T., Majima, T.: Single-molecule, single-particle fluorescence imaging of TiO2-based photocatalytic reactions. Chem. Soc. Rev. 39, 4802–4819 (2010) 8. Roeffaers, M.B.J., Sels, B.F., Uji-i, H., De Schryver, F.C., Jacobs, P.A., De Vos, D.E., Hofkens, J.: Spatially resolved observation of

293 crystal-face-dependent catalysis by single turnover counting. Nature. 439, 572–575 (2006) 9. Hendriks, F.C., Mohammadian, S., Ristanović, Z., Kalirai, S., Meirer, F., Vogt, E.T.C., Bruijnincx, P.C.A., Gerritsen, H.C., Weckhuysen, B.M.: Integrated transmission electron and singlemolecule fluorescence microscopy correlates reactivity with ultrastructure in a single catalyst particle. Angew. Chem. Int. Ed. 57, 257–261 (2018) 10. Xu, W., Kong, J.S., Yeh, Y.-T.E., Chen, P.: Single-molecule nanocatalysis reveals heterogeneous reaction pathways and catalytic dynamics. Nat. Mater. 7, 992 (2008) 11. Goldsmith, B.R., Peters, B., Johnson, J.K., Gates, B.C., Scott, S.L.: Beyond ordered materials: understanding catalytic sites on amorphous solids. ACS Catal. 7, 7543–7557 (2017) 12. Moerner, W.E., Fromm, D.P.: Methods of single-molecule fluorescence spectroscopy and microscopy. Rev. Sci. Instrum. 74, 3597–3619 (2003) 13. Pawley, J.: Handbook of Biological Confocal Microscopy. Springer Science & Business Media, New York, NY, U.S (2010) 14. Lakowicz, J.R.: Instrumentation for fluorescence spectroscopy. In: Principles of Fluorescence Spectroscopy, pp. 25–61. Springer, New York, NY, U.S (1999) 15. Hesari, M., Mao, X., Chen, P.: Charge carrier activity on singleparticle photo(electro)catalysts: toward function in solar energy conversion. J. Am. Chem. Soc. 140, 6729–6740 (2018) 16. Ye, R., Mao, X., Sun, X., Chen, P.: Analogy between enzyme and nanoparticle catalysis: a single-molecule perspective. ACS Catal. 9, 1985–1992 (2019) 17. Xu, W., Kong, J.S., Chen, P.: Probing the catalytic activity and heterogeneity of Au-nanoparticles at the single-molecule level. Phys. Chem. Chem. Phys. 11, 2767–2778 (2009) 18. Avanesian, T., Dai, S., Kale, M.J., Graham, G.W., Pan, X., Christopher, P.: Quantitative and atomic-scale view of CO-induced Pt nanoparticle surface reconstruction at saturation coverage via DFT calculations coupled with in situ TEM and IR. J. Am. Chem. Soc. 139, 4551–4558 (2017) 19. Hansen, P.L., Wagner, J.B., Helveg, S., Rostrup-Nielsen, J.R., Clausen, B.S., Topsøe, H.: Atom-resolved imaging of dynamic shape changes in supported copper nanocrystals. Science. 295, 2053–2055 (2002) 20. Nolte, P., Stierle, A., Jin-Phillipp, N.Y., Kasper, N., Schulli, T.U., Dosch, H.: Shape changes of supported Rh nanoparticles during oxidation and reduction cycles. Science. 321, 1654–1658 (2008) 21. Dou, J., Sun, Z., Opalade, A.A., Wang, N., Fu, W., Tao, F.: Operando chemistry of catalyst surfaces during catalysis. Chem. Soc. Rev. 46, 2001–2027 (2017) 22. Zhou, X., Xu, W., Liu, G., Panda, D., Chen, P.: Size-dependent catalytic activity and dynamics of gold nanoparticles at the singlemolecule level. J. Am. Chem. Soc. 132, 138–146 (2010) 23. Han, K.S., Liu, G., Zhou, X., Medina, R.E., Chen, P.: How does a single Pt nanocatalyst behave in two different reactions? A singlemolecule study. Nano Lett. 12, 1253–1259 (2012) 24. Sambur, J.B., Chen, P.: Approaches to single-nanoparticle catalysis. Annu. Rev. Phys. Chem. 65(65), 395–422 (2014) 25. Andoy, N.M., Zhou, X.C., Choudhary, E., Shen, H., Liu, G.K., Chen, P.: Single-molecule catalysis mapping quantifies site-specific activity and uncovers radial activity gradient on single 2D nanocrystals. J. Am. Chem. Soc. 135, 1845–1852 (2013) 26. Gilroy, K.D., Ruditskiy, A., Peng, H.C., Qin, D., Xia, Y.N.: Bimetallic nanocrystals: syntheses, properties, and applications. Chem. Rev. 116, 10414–10472 (2016)

13

294 27. Chen, G.Q., Zou, N.M., Chen, B., Sambur, J.B., Choudhary, E., Chen, P.: Bimetallic effect of single nanocatalysts visualized by super-resolution catalysis imaging. ACS Cent. Sci. 3, 1189–1197 (2017) 28. Zou, N.M., Chen, G.Q., Mao, X.W., Shen, H., Choudhary, E., Zhou, X.C., Chen, P.: Imaging catalytic hotspots on single plasmonic nanostructures via correlated super-resolution and electron microscopy. ACS Nano. 12, 5570–5579 (2018) 29. Zhou, X.C., Choudhary, E., Andoy, N.M., Zou, N.M., Chen, P.: Scalable parallel screening of catalyst activity at the single-particle level and subdiffraction resolution. ACS Catal. 3, 1448–1453 (2013) 30. Mao, X., Liu, C., Hesari, M., Zou, N., Chen, P.: Super-resolution imaging of nonfluorescent reactions via competition. Nat. Chem. 11, 687 (2019) 31. Zou, N.M., Zhou, X.C., Chen, G.Q., Andoy, N.M., Jung, W., Liu, G.K., Chen, P.: Cooperative communication within and between single nanocatalysts. Nat. Chem. 10, 607–614 (2018)

Xianwen Mao obtained his BS degree in Polymer Materials and Engineering from Tsinghua University, China. He then came to Massachusetts Institute of Technology to pursue his graduate study, and obtained his Ph.D. degree in Chemical Engineering with T. Alan Hatton and Gregory C. Rutledge. His Ph.D. work is focused on molecular engineering of soft electrochemical materials for healthcare and sustainability. He is currently a postdoctoral associate with Peng Chen in the Department of Chemistry and Chemical Biology at Cornell University, working on operando functional imaging of (photo)electrocatalysts for water decontamination and artificial photosynthesis.

X. Mao et al.

Rong Ye was born in Guangdong, China. He received his B.E. degree from Sun Yat-sen University in China and M.Sc. degree in Chemistry from University of California, Los Angeles. He obtained his Ph.D. degree with Profs Gabor A. Somorjai and F. Dean Toste at University of California, Berkeley, where he investigated strategies to combine the advantages of heterogeneous and homogeneous catalysis. Currently, he is a Presidential Postdoctoral Fellow at Cornell University and working with Prof. Peng Chen on ligand adsorption on single nanoparticles and on polymerization kinetics.

Peng Chen is the Peter J.W. Debye Professor of Chemistry at Cornell University. He received his B.S. in Chemistry from Nanjing University, China in 1997, and his Ph.D. in bioinorganic/physical inorganic chemistry (with Ed Solomon) from Stanford University in 2004. After postdoctoral training in single-molecule biophysics (with Sunney Xie) at Harvard University, he started his faculty appointment at Cornell University in 2005. His current research focuses on single-molecule imaging of heterogeneous and homogeneous catalysis as well as of metal homeostasis machineries in vitro and in living cells.

Photoluminescence (PL) Spectroscopy

14

Qinghe Li, Masakazu Anpo, Jinmao You, Tingjiang Yan, and Xinchen Wang

Contents 14.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296

14.2 14.2.1

Basic Principles of Photoluminescence . . . . . . . . . . . . . . . . 296 Absorption Spectrum, Franck-Condon Principle, and Vibration Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 The Fate of Electronic Excitation Energy . . . . . . . . . . . . . . . . 297

14.2.2 14.3 14.3.1 14.3.2 14.3.3 14.3.4 14.3.5 14.3.6 14.3.7 14.3.8 14.3.9 14.4 14.4.1 14.4.2 14.4.3 14.4.4

Practical Aspects of Photoluminescence . . . . . . . . . . . . . . . Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spectral Parameters to Identify Photoluminescence Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wavelength and Spectral Shape . . . . . . . . . . . . . . . . . . . . . . . . . . Quantum Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lifetimes and the Stern-Volmer Expression . . . . . . . . . . . . . . Energy Transfer and Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . Ultrafast Time-Resolved PL Spectroscopy . . . . . . . . . . . . . . . Relevance of Photoluminescence to Surface Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

301 302 302

305

Characterization of Catalytically Active Sites by In Situ Photoluminescence Spectroscopy . . . . . . . . . . . . . . . . . Ti-Oxide Single-Site Containing Samples . . . . . . . . . . . . . . . V-Oxide Single-Site Containing Samples . . . . . . . . . . . . . . . . Mo-Oxide Single-Site Containing Samples . . . . . . . . . . . . . . Carbon Containing Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

306 306 307 308 309

303 303 304 304 305 305

14.5

Characterization of Acidic and Basic Sites by Means of Luminescent Probe Molecules and In Situ Photoluminescence Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 310

14.6

In Situ Photoluminescence Studies of Photocatalytic Processes Involving Inorganic and Organic Semiconductor Photocatalytic Systems . . . . . . . . . . . . . . . . 311

Q. Li · X. Wang Fuzhou University, Fuzhou, China e-mail: [email protected]; [email protected] M. Anpo (*) Osaka P. U., Osaka, Japan e-mail: [email protected] J. You Qufu Normal University, Qufu, China Shaoxing University, Shaoxing, China T. Yan Qufu Normal University, Qufu, China

14.7

Effect of Temperature on Photoluminescence Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

14.8

Effect of Magnetic Fields on Photoluminescence Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317

14.9

Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318

Abstract

This chapter deals with the fundamental of the photoluminescence (PL) spectroscopy and its applications to study the chemical reaction of molecules on solid surfaces and the reactivity of various heterogeneous solid catalysts in relation to their properties in adsorption, catalysis, and photocatalysis. After a short introduction, the basic principles of PL spectroscopy are explained in relation to the definitions of fluorescence and phosphorescence. And, the PL features of the semiconducting catalysts are discussed in relation to the surface band structures. Next, the practical aspects of static and dynamic PL with the spectral parameters including wavelength and spectral shape, lifetime and the Stern-Volmer expression, energy transfer and migration, and ultrafast time-resolved PL spectroscopy are discussed. In Sect. 14.4, which is one of the cores of this chapter, the characterization of the catalytically active sites by applying in situ PL spectroscopy are discussed with various single-sites heterogeneous catalysts such as Ti-oxide, V-oxide, and Mo-oxide single-site containing catalysts, and carbon containing catalysts such as polymeric carbon nitride. In Sect. 14.5, characterization of acidic and basic surface sites is discussed by means of luminescence probe molecules and in situ PL spectroscopy. In Sect. 14.6, in situ PL studies are discussed in relation to the photocatalytic reaction processes on inorganic and organic semiconducting catalysts. Especially, photo-generation of electron and holes, their lifetimes,

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_14

295

296

Q. Li et al.

and reaction dynamics are discussed. In Sect. 14.7, effects of temperature on PL spectra of their intensity and wavelength are discussed. In Sect. 14.8, effect of magnetic fields on PL spectra are discussed. Section 14.9 is the conclusion and outlook of the PL spectroscopy in catalysis and photocatalysis. Keywords

Photoluminescence · Adsorption · Catalysis · Photocatalysis · Single-site catalyst · Semiconducting catalyst · Dynamic of photocatalysis · In situ study

14.1

Introduction

What is “Photoluminescence”? Electronically excited states of molecules and substances such as catalysts are at the heart of all physical and chemical processes, which occur after their absorption of light. Some thermal pathways are also known to produce electronically excited states. Physical and chemical reactions occur at these electronically excited states, and the physical processes involve both radiative and radiationless processes. “Photoluminescence” is associated with the radiative process, thus, studies of photoluminescence in its static and dynamic measurements have played an important role in understanding the photochemistry of compounds and substances at their molecular level. There have been remarkable developments in photochemistry through studies of the transition states of chemical reactions using femtosecond photoluminescence spectroscopy by Prof. Ahmed H. Zewail who was awarded the Nobel Prize in Chemistry in 1999. Investigations into the photochemistry of molecules within certain constrained environments such as their adsorption on solid surfaces have led to remarkable advances in molecular photochemistry. Photochemical studies in such heterogeneous systems have also led to the application of photoluminescence spectroscopy to understand the solid surface itself. At the same time, many studies of photochemical reactions on solids such as semiconducting materials in relation to the conversion of light energy into chemical energy through the excitation of such solid materials have also been carried out. Solid catalysts are usually utilized in a high dispersion state on a support in order to obtain efficiency and effectiveness in the catalytic reaction. Especially, highly dispersed heterogeneous solid catalysts have a higher surface area and large numbers of active sites, leading to enhanced catalytic activity. Newly synthesized zeolites allow the incorporation of various metal cations in their framework structures and/or the walls for the construction of highly dispersed single-site metal oxide heterogeneous catalysts. Interestingly, they are

not only considered solid surface models for bulk metal oxide catalysts but also exhibit very unique photocatalytic performance, which cannot be observed with such bulk metal oxide catalysts alone. Such highly dispersed transition metal oxide photocatalysts were found to exhibit unique photoluminescence under UV light irradiation, its intensity depending on the concentration of reactant molecules. These results showed that measurement of the photoluminescence provides important information on the reaction mechanism of photocatalysis on highly dispersed single-site oxide heterogeneous catalysts [1–4]. These pioneering works along with various studies in catalysis now commonly use photoluminescence spectroscopy to understand the catalysts themselves. Thus, a summary of its fundamentals and applications is vital to make sense of such data in understanding the reaction mechanisms for the further development of catalysis and photocatalysis. (Previous chapters and research in this field): Adv. Catal. Vol. 44, 1999, “Application of photoluminescence techniques to the characterization of solid surfaces in relation to adsorption, catalysis, and photocatalysis” by M. Anpo & M. Che [3], Adv. Catal., Vol. 52, 2009, “Applications of photoluminescence spectroscopy to the investigation of oxidecontaining catalysts in the working state” by M. Anpo, S. Dzwigai, & M. Che [4], and Res. Chem. Intermediat., Vol. 46, 2020, “Application of photoluminescence spectroscopy to elucidate photocatalytic reactions at the molecular level” by Q. H. Li, M. Anpo, & X. C. Wang [5]. Here, we have detailed the basic principles and practical aspects of photoluminescence, the characterization of catalytically active sites by in situ photoluminescence spectroscopy, characterization of acidic and basic sites by luminescent probe molecules and in situ photoluminescence spectroscopy, in situ photoluminescence studies of photocatalytic processes involving inorganic and organic semiconductor photocatalytic materials, the effect of temperature on photoluminescence spectra, and the effect of the magnetic field on photoluminescence spectra. We sincerely hope this chapter on “Photoluminescence Spectroscopy” will contribute to an understanding of catalysts and photocatalysis that will benefit our world.

14.2

Basic Principles of Photoluminescence

Photoluminescence, which could be defined as the radiation emitted from a molecule or a solid when it has absorbed the energy from an external source and transfer into an electronic excited state returning to the ground electronic state subsequently, is composed of fluorescence and phosphorescence. Presented here is the context of photoluminescence as an analytical tool applied in material analysis science involving the solid surface represented by alkaline earth oxides, carbon

14

Photoluminescence (PL) Spectroscopy

297

material, and organic polymers, and its basic principles could also be found in physical chemistry and photochemistry books [6–14].

14.2.1 Absorption Spectrum, Franck-Condon Principle, and Vibration Structure Taking a solid oxide surface as an example, the photoluminescence principles could be easily elucidated by assuming some simplifications. As shown in Fig. 14.1, the harmonic oscillator represents an alkaline earth oxide in the Morse potential energy curves related to an ion pair such as M2+-O2
. Absorbing the light similar to the fundamental absorption edge of the oxide, the electron will be excited concurrently, followed by a charge-transfer process to create an exciton (the electron-hole pair), which is essentially free to migrate through the lattice: M

2þ

 ns

0

 O

2


 2p

6



2þ

"# ðS0 Þ ! M

 ns

0

 O

2




5

2p 3s

1



"#

ð14:1Þ M

2þ

 ns

0

 O

2




5

2p 3s

1



"#! M

þ



1

ns






O



5

0

2p 3s



"# ðS1 Þ

ð14:2Þ The horizontal lines in the potential wells represent the possible vibrational states of the oscillator, either in the ground electronic singlet state (S0) in which all the electrons are paired or in the electronic excited singlet state (S1) in which the charge-transfer shown in Eq. (14.2) occurs and the unpaired electrons remain with opposite spins. The vibrational states (ν) are labeled by increasing energy, starting with the vibrational ground state ν ¼ 0, of the electronic ground state, followed by the vibrational excited states ν1, ν2, etc. Labeling is performed for the electronic excited state, starting again with the vibrational ground state ν′ ¼ 0, followed by the vibrational excited states ν′1, ν′2, etc. It is readily understood that the potential energy curve of the S1 is located at higher energy than that of the S0, with its potential well located at larger internuclear distance because, electrostatically, the charge transfer leads to a M+-O
bond, i.e., a weakening of the M2+-O2
bond. The observed spectral transitions, which lead to the absorption spectrum, are related to the energy diagram such as that of Fig. 14.1 on the basis of the Franck-Condon principle. This principle is based on the fact that electrons move and rearrange at a much faster speed than that of the vibrational movement of the nuclei of a molecule. For example, the time for an electron to circle a hydrogen nucleus can be calculated from Bohr’s model to be about 10
16 s. A typical vibration period for a molecule is only about 10
13 s.

Potential energy

Continuum convergence limit

s1 4 23 1 0

5 Q s0

Line spectrum

3 12 0 Internuclear distance

Fig. 14.1 Potential energy curves for the ground (S0) and excited states (S1), which account for the absorption spectrum with a vibrational fine structure shown on the right-hand vertical side [3]

The comparison of these times suggests that an electronic configuration changes within a time so short that the nuclei do not change their positions during absorption. The same reasoning state for the molecule also applies to the M2+-O2
ion pair. Spectral transitions with the highest probability then occur at constant internuclear distance and, therefore, should be drawn vertically and not, as might be expected, from the potential minimum of the lower curve to that of the upper curve. The transition probability decreases as the transition departs from the vertical and is highest when starting near the midpoint of the lowest vibrational level of the ground electronic state. In absorption, an electronic transition may show a series of lines, called the vibrational or fine structure, corresponding to the different vibrational states reached and in which the most intense transition refers to the most probable vertical transition.

14.2.2 The Fate of Electronic Excitation Energy In absorbing light, the system acquires the associated excitation energy and moves to an upper electronic excited state. There are many ways by which the system can return to the lower energy state. The excess energy can be lost to the vibration, rotation, and translation of the surrounding molecules or ions. This loss of energy is most efficient for gases and liquids and less so for solids, in which vibrations essentially are predominant since rotations and translations are hindered. This thermal degradation transforms the excitation energy into thermal motion of the environment, i.e., heat. Another, more interesting, possibility is that the excess

14

298

Q. Li et al.

energy becomes involved in a chemical reaction, and the result is photochemistry [10–14]. Other possibility is radiative decay (as opposed to nonradiative or radiationless decay), which occurs when the molecule or ion pair loses its excitation energy as a photon. Two different radiative decay mechanisms are possible: fluorescence and phosphorescence. The radiation emitted in a transition between states of the same spin multiplicity (i.e., singlet-singlet or triplettriplet transitions) is called fluorescence, and the radiation emitted in a transition between states of different spin multiplicity (i.e., triplet-singlet transitions) is called phosphorescence. The fluorescence lifetimes are usually very short (e.g., about 10
9 s for organic molecules), whereas phosphorescence lifetimes are about 10
3 s to minutes or longer because transitions between states of different spin multiplicity are forbidden and thus have very low probabilities.

Excitation, Emission (Fluorescence), and Stokes Shift The sequence of steps involved in fluorescence is described in Fig. 14.2. The initial absorption takes the molecule or ion pair from the S0 to the S1, and the resulting absorption (excitation) spectrum should resemble that shown in Fig. 14.3.

Potential energy

Radiationless decay

3 Absorption

1

4

5

6 (S1)

2

(S0)

0

hn

2

5 Radiation 4 (fluorescence) 3 hn ′

1 0 Internuclear distance

Fig. 14.2 The sequence of steps leading to fluorescence. After the initial absorption, the upper vibrational states in the excited state (S1) undergo radiationless decay (vibrational relaxation) by giving up energy to the lattice environment and/or surrounding gas-phase molecules, when the ion pair is located at the solid–gas interface, a radiative transition then occurs from the vibrational ground state (ν’0) of the electronically excited state (S1) to the ground state (S0) [3]

In the excited state, the ion pair is subjected to the influence of its lattice environment, and as it gives up energy (by radiationless decay or a vibrational relaxation process) it steps down the ladder of vibrational levels. The lattice environment, however, may be unable to accept the larger energy difference needed to lower the ion pair to the ground electronic state; therefore, the ion pair may survive long enough to undergo spontaneous emission, releasing the remaining excess energy as radiation. The downward electronic transition is vertical in accordance with the FranckCondon principle, and the fluorescence spectrum shown has a vibrational structure characteristic of the lower electronic state. This mechanism accounts for the observation that fluorescence occurs at a frequency lower than that of the incident light. The difference between the two frequencies is referred to as the Stokes shift. The emission occurs after some vibrational energy has been dissipated into the surroundings. The mechanism also suggests that the intensity of the fluorescence should depend on the ability of the lattice environment (or the surrounding gas phase molecules if the ion pair is located at the solid–gas interface) to accept the electronic and vibrational quanta. It has been observed that molecules with widely spaced vibrational levels, such as molecular oxygen, may be able to accept the large quantum of electronic energy and extinguish or quench the fluorescence. Also, the environment is less likely to accept the excitation energy as the temperature is lowered since lattice vibrations are less favored at low temperatures [10–14].

Excited Triplet State, Intersystem Crossing, Phosphorescence, and Selection Rules When a spin flip occurs for an electron in the S1 state, an excited triplet state (T1) is created: þ

M

 ns

1



"O




"O






5

0



2p 3s 

5 0

2ps s



# ð S1 Þ ! M " ðT 1 Þ

þ

  1 ns ð14:3Þ

The triplet state lies lower than the singlet state since electron-electron repulsions are less in the triplet state. The sequence of steps involved in phosphorescence can be understood as described in Fig. 14.4. The first steps are the same as the fluorescence, but the presence of the T1 plays a decisive role. At the point at which the potential energy curves intersect, the two states share a common geometry. If there is a mechanism for changing the spin state of the S1 state, then the ion pair M+-O
may cross into the triplet state by the intersystem crossing. This singlet-triplet transition may take place in the presence of spin orbit coupling. Such intersystem crossing is expected to occur more efficiently in a molecule with a heavy atom (the so-called heavy atom effect) or in an

14

Photoluminescence (PL) Spectroscopy

Fig. 14.3 Schematic diagram showing the origin of absorption (a), fluorescence (b), and phosphorescence (c) spectra. The absorption spectrum (a) shows a vibrational structure characteristic of the upper excited state (S1). The fluorescence spectrum (b) shows a structure characteristic of the lower ground state (S0); it is also displaced to lower frequencies (the 0-0 transitions show a coincidence) and resembles a mirror image of the absorption [3]

299

v′ = 4 v′ = 3 v′ = 2 v′ = 1 v′ = 0

S1

T1 v′ = 0

Vibrational level v=4 v=3 v=2 v=1 v=0

S0

Absorption or emission intensity (a.u.)

Absorption

a)

Fluorescence

Phosphorescence

b) c)

Wavelength (cm−1)

Radiationless decay

Potential energy

Intersystem crossing (S→T)

Singlet

(S1)

Triplet

(T 1)

Singlet (S0)

Absorption hn Phosphorescence hn ′

Internuclear distance

Fig. 14.4 The sequence of steps leading to phosphorescence. The important step is the intersystem crossing, the passage from the excited singlet (S1) to triplet state (T1) induced by spin orbit coupling. The triplet state acts as a slowly radiating reservoir since return to the ground state is spin forbidden [3]

ion pair in an oxide solid since the corresponding spin orbit coupling is large and electrons can acquire the same spin, leading to the excited triplet state [11–13]. The ion pair in the excited triplet state continues to release energy into the surroundings, moving down the vibrational ladder, until it is trapped in the lowest vibrational state. Its surroundings cannot accept the large quantum of electronic excitation energy, and the ion pair cannot radiate its energy because the return to the electronic ground state involves a forbidden triplet-singlet transition. This transition is not totally forbidden, however, because the spin-orbit coupling that was responsible for the intersystem crossing also breaks the selection rule. The ion pairs are able to emit phosphorescence weakly, and the emission may continue long after the original excited state is formed. This mechanism is in agreement with the experimental observation that the excitation energy appears trapped in a reservoir that slowly leaks. Generally, the phosphorescence is most intense when catalysts have highly dispersed metal ions and solid materials have highly localized emitting sites because the energy transfer becomes less efficient and there

14

300

Q. Li et al.







O þ CH4 ! OH þ CH3

ð14:4Þ

The transitions involved in the photochemical processes discussed so far are electronic in nature and thus associated with electric dipole changes during either absorption or emission. From a theoretical standpoint, the dipole moment change, which occurs during the transition of a molecule moving from a state of energy Ei (wave function: φi) to a state of higher energy Ef (wave function: φf), is represented by the transition-moment integral, Rif ¼ ˂φi jRjφf>. The probability of the transition (absorption or emission) is proportional to the magnitude of Rif2. If Rif ¼ 0, the corresponding transition is forbidden, whereas if Rif differs from zero it is allowed [10–14]. From a more practical point of view, electronic transitions follow two types of selection rules because of the orbital and spin nature of the electronic wave function. The first requires that Δl ¼  1 for the orbitals involved in the transition. It predicts, for instance, that electronic transitions for transition metal ions in Td symmetry (involving orbitals with d-p character) should be more intense than d-d transitions in Oh symmetry involving orbitals of the same character thus leading to Δl ¼ 0. By contrast, charge-transfer transitions are essentially. In the case of centrosymmetric complexes, a change of parity: u ! u and g ! g transitions (as, for instance, t2g ! eg transitions in octahedral complexes) are forbidden, whereas transitions such as g ! u are allowed. The second wave function referred to as the spin rule, states that the spin multiplicity 2S + 1 of the levels involved must be the same, i.e., ΔS ¼ 0. Thus, singletsinglet transitions such as those involved in fluorescence are spin-allowed, whereas singlet-triplet transitions such as those involved in phosphoresce are spin-forbidden. These selection rules may be partially or even totally relaxed by various mechanisms, such as vibronic coupling, spin orbit coupling, and exchange interaction in the case of polymetallic systems [22–25].

Vibrational Deactivation and Internal Conversion As stated previously, the loss of the excitation energy of the system can occur by a nonradiative process. Figure 14.5 shows a representative arrangement of potential energy curves, which lead to such deactivation. The excess energy is first lost by vibrational deactivation, internal conversion; note that this process refers to electronic excited states with the same spin multiplicity within the highest potential energy curve of the excited singlet state S1. At the crossover point (C) with the excited triplet state, the geometry and potential energies of the two electronic states are equal so that the system can transfer to the lower potential energy curve T1 by intersystem crossing (this process involves states of different spin multiplicities). Other crossover points (C′ and C00 ) allow the systems to reach their original ground states. This case is commonly encountered with polyatomic molecules for which potential curves often exist with suitable relative positions so that combinations of internal conversion and intersystem crossing return the molecule to its ground state before another emission process has had a chance to occur. The previous points are schematically represented in Fig. 14.6, which summarizes the processes described so far between various energy levels [10–14]. Radiative Processes on Semiconducting Catalyst: Effect of the Adsorption of Various Reactant Molecules Upon the Radiative Process (PL Behaviors) on TiO2 The photoluminescence (PL) properties of semiconducting materials related to the photocatalytic reactions on their

(S1) Excited singlet state Potential energy

is enough time for intersystem crossing to occur, as the singlet excited state steps slowly past the intersection point. The mechanism also suggests that the phosphorescence efficiency should depend on the presence of heavy atoms (with strong spin-orbit coupling) in molecules and on the bonding strength of the ion pair with its surroundings. Finally, it also predicts that, due to the unpaired spins in the triplet, magnetic properties should be involved either as a S ¼ 1 system, if the two unpaired electrons are present, or as a S ¼ 1/2 system, if only one unpaired electron remains after reaction (e.g., with gas-phase molecules) [15–21]. A good illustration of the latter case is the activation of CH4 by surface O
ions, produced by processes shown in Eqs. (14.1)–(14.3), following the reaction

(T1) Excited triplet state

C

C⬘

(S0) Ground state

C⬙

Internuclear distance

Fig. 14.5 Energy dissipation (dashed arrow) by vibrational deactivation and internal conversion [3]

14

Photoluminescence (PL) Spectroscopy

Fig. 14.6 Schematic representation of changes in electronically energy levels, which may occur upon absorption of radiation. Nonradiative processes are represented by wave arrows and radiative processes by straight line arrows [3]. (1) Absorption, (2) Internal conversion, (3) Fluorescence, (4) Intersystem crossing, (5) Phosphorescence

301

S3 (2) Internal conversion Excited singlet S2 state (2) Internal conversion S1

14 (4) Intersystem crossing (3) Fluorescence Excited triplet stste

T1 (1) Absorption

(5) Phosphorescence Intersystem crossing

Ground singlet S0 state

surfaces are remarkably changed by the addition of reactant molecules such as O2 and H2O. One of the most famous inorganic semiconducting photocatalysts, TiO2 catalyst, exhibits the PL spectrum at around 450–550 nm when excited with UV light having energies greater than the band gap (3.2 eV) of the catalyst. With an increase in the adsorption of O2 onto the TiO2, the PL spectrum decreases remarkably in intensity, leading to the quenching of the PL. The addition of N2O also leads to the quenching of the PL. On the other hand, the addition of 1-C4H4 onto the TiO2 caused the PL to increase in intensity. Similarly, the addition of various unsaturated hydrocarbons such as ethylene, propylene, acetylenes, as well as H2O and H2 enhanced the PL in intensity. It is known that the extent of PL enhancement depends strongly on the ionization potential of the added molecules (i.e., the lower the ionization potential of the added molecule, the larger the PL enhancement in intensity). The addition of N2 molecule scarcely changes in the PL in intensity [3, 5]. These results are well understood by the fact that the formation of negatively charged adduct species through electron donation by the TiO2 surfaces causes the quenching of the PL Case 3 in Fig. 14.7), whereas the formation of positive adduct species through hole trapping results in an enhancement of the PL in intensity (case 1 in Fig. 14.7). Thus, the PL enhancement by the adsorption of H2, H2O, or unsaturated hydrocarbons is attributed to the reduction in work function of TiO2 semiconductor due to a decrease in the extent of structural changes in the bending of the surface band

structure, which is explained in the dead-layer model shown in Fig. 14.7. The reduction in the band bending at the surfaces results in an increase in the radiative recombination of the photo-generated electrons and holes and subsequently an enhancement of the PL. On the other hand, the addition of O2 or N2O to these semiconducting catalysts causes a suppression of the radiative recombination of the photogenerated electrons and holes, leading to an efficient charge separation and subsequently a quenching of the PL. Thus, the PL behaviors strongly depend on the nature of the surface electronic structures of the catalysts in the reaction atmosphere, their extents depending on the electronegativity or electron affinity of the reactant molecules. Not surprisingly, such PL behaviors are found to be closely related to the photocatalytic reactivity of these semiconducting catalysts [3, 5].

14.3

Practical Aspects of Photoluminescence

Photoluminescence analysis generally involves various molecular spectroscopic techniques used to observe the sensitivity and selectivity of the emitting species of molecules and/or materials. The high sensitivity of photoluminescence analysis has allowed the detection and determination of extremely low concentrations of emitting materials. Lower limits of detection are in the parts per million to parts per billion range. The selectivity of photoluminescence analysis

302

2

Electron CB

O2 (N2O)

CB

Adsorption VB

VB

Radiative recombination

1

3

H2O (H2)

Adsorption VB

o Hole TiO2

CB

Interface

Fig. 14.7 Schematic descriptions of the surface band bending structures of the semiconducting TiO2 catalyst. (1) Just after degassing in a high vacuum environment, (2) after adsorption of H2O (or H2) (in this case, positive charged species are formed on the surface), (3) after adsorption of O2 (or N2O) (in this case, negative charged species are formed on the surface)

Q. Li et al.

Bulk inside

has become possible using variations in excitation and/or analytical wavelengths, allowing the simultaneous determination of emitting components in many mixtures. Potentially, phosphorescence can be observed with a higher degree of sensitivity and selectivity than fluorescence due to its longer lifetime, which offers the possibility of investigations with differing spectroscopic parameters. For a better understanding and more efficient application of photoluminescence spectroscopy, it is essential to gain insight into the nature of the excited states of the target species and of the various photochemical processes [12, 13]. A molecule in an excited state can be considered a new entity, only remotely related to the parent molecule in the ground state. In the excited state, the molecule will have a completely different electron distribution from that of the molecule in the ground state as well as a different geometry. Also, it usually undergoes chemical reactions that are quite different from those of the molecule in the ground state. These photochemical and photophysical processes can be used to explain the basic principles of the photocatalytic reactivity of catalysts.

14.3.1 Instrumentation A variety of commercially available autocompensating spectrophotofluorometers (photoluminescence instrumentation) have been employed to measure and analyze the photoluminescence spectra of catalysts. The three principal components of all spectrophotofluorometers are the excitation light source, the chamber to set the sample, and the emitted photons detector (Fig. 14.8). The light source is usually either a mercury or a xenon arc lamp. The light passes through a monochromator (equipped with a grating or a

prism) to select the excitation radiation to be focused on the sample cell. Interference filters can also be used to provide greater selectivity and resolution for analysis of the photoluminescence spectra. The emission leaving the sample cell usually passes directly into the emission monochromator and then into a high gain photomultiplier tube. Although the photoluminescence output can be displayed in a variety of ways, the spectrum is recorded with the correct compensation to cover small variations in the excitation light power [10–14].

14.3.2 Sample Preparation Although the design of the sample cell holder may determine the amount of scattered and stray light from the samples, glass filters can be used to minimize scattered light. Some kinds of quartz may fluoresce under ultraviolet (UV) excitation causing errors in the photoluminescence measurements. Therefore, secondary filters may be used, and the sample cell should be made from high-quality nonfluorescent quartz. It is not possible to eliminate the Raman emission from the sample, which is sometimes used to interrelate the fluorescence spectra, but the fluorescence spectrum of the blank, which comes from the sample cell, can be used to detect the background signals due to scattering. The presence of oxygen is critical to measure the photoluminescence because the interaction of oxygen with the sample can cause serious errors since oxygen shows a strong ability to quench not only the fluorescence but also phosphorescence. Therefore, to measure the meaningful photoluminescence it is mandatory to degas the sample. Also, the photoluminescence intensity usually decreases when the temperature of the sample is increased; this is attributed to the

14

Photoluminescence (PL) Spectroscopy

303

a)

CCD detector

b)

Spectrograph grating

Laser and line filter

Notch filter Microscope Sample lens

14

Mirrors

Adjustable entrance slit

Fig. 14.8 Basic setup to measure the photoluminescence spectrum and its lifetime. (a) photonbeam path diagram and (b) appearance of physical equipment

higher probability for other nonradiative deactivations in the excited state of the molecule. Therefore, to minimize such temperature effects, the photoluminescence spectra are often measured at liquid-nitrogen temperature and even liquidhelium temperature [10–14].

14.3.3 Spectral Parameters to Identify Photoluminescence Sites Measuring the fluorescence intensities of fluorescent reference compounds for the purpose of identification and standardization is essential in monitoring the emission yields of the sample and calibrating the number of photons emitted from the excitation source. The ideal is a series of reference compounds that cover the visible to near-UV regions of the spectrum. The reference compound should also have its absorption in the same regions that of the sample compound, and its emission spectrum should be in the same general region as the emission of the compound under investigation. The most important parameters of the photoluminescence spectrum, which are necessary to identify the emitting species, are its spectral shape, from which the emission intensity Ιe is obtained as a function of wavelength λ; its quantum efficiency Фe, by which the intensity is measured relative to the absorption intensity Ιa; and its lifetime τ, determined from the decay curves of variation of the intensity measured as a function of time [10–14].

14.3.4 Wavelength and Spectral Shape In a photoluminescence spectrum, the emission intensity Ιe, is plotted as a function of the frequency μ (or wavelength λ) of

the emitted light I, for which the excitation frequency (or wavelength) and the intensity of the exciting light are fixed at constant values. The photoluminescence spectrum may be well defined but it is also often partially resolved or even unresolved. This can be explained by the molecular electronic transition, which does not always correspond to a well-defined quantum of energy because different nuclear geometries are associated with the initial or final electronic states, leading to partially resolved or unresolved spectra. Such a situation is most likely found for molecules in solution or for solids. In certain cases, however, some vibrational fine structure is observed in the band corresponding to the electronic transition, subsequently leading to the prominent vibrational progression of the emission band of the vibrations associated with this emission, the nuclear equilibrium positions are most dramatically changed by the radiative electronic transition. As shown in Fig. 14.9, the vibrational bands are clearly resolved in the photoluminescence spectrum of vanadium oxide single-site supported on Vycor glass [22]. The vibrational separation between the (0 ! 0) and (0 ! 1) of about l040 cm
1 is in good agreement with the energy of the V¼O stretching vibration of the ground state of the vanadyl group (V¼O) of the oxide, which was measured by infrared (IR) or Raman spectroscopies. Under the light radiation, the tetrahedrally coordinated (V5+-O2
) single-site species will be changed to the charge transfer excited triplet state (V4+-O
). On the other hand, in the excitation spectrum, the emission intensity Ie, at the monitored emission band, is plotted as a function of the wavelength (λ) of the excitation light, which varies as the extinction coefficient εa of the absorbing molecules. Therefore, the excitation spectrum exhibits the same spectral appearance as that of the absorption spectrum. The advantage of measuring the excitation spectrum in addition to the emission spectrum is the greater sensitivity even for low

304

Q. Li et al.

a)

Wavenumber (cm–1) 20,000 17,000 Relative intensity

25,000

0-3 0-2 0-4 0-1 6 0-5 4 0-0 0-6 2

10 8

0-0 0-1 0-2 0-3 0-4 0-5 0-6

Vibrational transitions Vibrational progression

Intensity (arb. unit)

Fig. 14.9 (a) The photoluminescence spectrum of vanadium oxide single site heterogeneous catalyst supported on porous Vycor glass at 77 K with a beautiful vibrational fine structure. Insert shows its vibrational progression. (b) The charge transfer processes of the tetrahedrally coordinated V5+ oxide single-site species [22]

b) O2– V5+

0 1 Band

O O O

O– hv hv⬘

V4+ O O O

0 0 Band 1040 cm–1

400

500

600

Wavelength (nm)

concentrations of photoluminescent material compared to standard absorption measurements.

77 or 4 K, the quantum yields are generally less than 1.00 since radiationless processes occur even at such low temperatures [10–14].

14.3.5 Quantum Efficiency The quantum efficiency or yield of the photoluminescence, Фe, is defined as the ratio of the number of photons emitted from the excited molecules to the number of photons absorbed by the ground state molecules, being expressed as 

Фe ¼ ke =ðke þ kIC Þ ¼ Ф ke τ

ð14:5Þ

where ke and kIC are the radiative rate constant and the sum of the unimolecular rate constants of the nonradiative deactivation processes from the radiative excited state, respectively, and Ф* and τ are the concentrations of entities in the excited state and the measured experimental lifetime of the excited state, respectively. All excited states emit a certain number of photons, but it has been found that quantum yields of 20 1.86 480 0.52 35

O2–

0.41 ms 0.27 ms

(c) (H2O)

(d) (H2O) 400

500 600 Wavelength (nm)

O–

*

Ti3+ O

O O

Absorbance (arb. unit)

(b) (CO2)

hv

210 – 260 nm

Ti4+ O O O Relative intensity (a.u.)

Fig. 14.11 (A) The PL spectrum of tetrahedrally coordinated Ti-oxide single-site species constructed within porous silica thin film (Ti-PS (h, 50) and the effect of the addition of CO2 and H2O on the PL spectrum. Excitation at 260 nm, measured at 77 K. amount of CO2: (b) 0.15 and H2O: (c) 0.15, (d) 1.5 mmol/g-cat. (B) The excitation (absorption) spectrum of tetrahedrally coordinated Ti-oxide single-site species constructed within porous silica thin film (Ti-PS (h, 50) (charge transfer absorption). Measured at 77 K [29]

Ti-βeta(OH) Tetra 7.66.4 1.84 480 0.42 21

200 700

300

400

Wavelength (nm)

500

14

308

Q. Li et al.

Charge transfer Excited state O2–

250~300 nm

Discrete energy level of single-site species

Ti4+ O2–

O2–

O2–

Ground state

40

80

30

60

20

40

10

20

0 3.5

4

4.5

5

5.5

Selectivity for N2 formation (%)

Isolated Ti-oxide single-site species

Selectivity for CH3OH formation (%)

100

0 6.5

6

O2– O2– Ti4+

Coordination number Conduction band

O2–

O2– O2–

O2– 2NO

hv

CO2 + H2O

N2 + O2 (quantum yield for N2 (and O2) formation: 17.5%) hv

CH3OH + CH4 (quantum yield: 0.3%)

380 nm Valence band

Unit structure of bulk TiO2 semiconductor

Fig. 14.12 Relationship between the coordination numbers of Ti-oxide species of various Ti-oxide-based catalysts and the selectivity of photocatalytic reduction of CO2 and H2O to form CH3OH and photocatalytic direct decomposition of NO into N2 and O2 at 275 K [33]

vibrational energy level of the CT excited triplet state (V4+O
)* of tetrahedral O¼VO3 units to the various vibrational energy levels of its ground singlet state (V5+-O2
). In order to ease the analysis of very complex spectra due to the superimposition of different vibrational fine structures, the secondderivative photoluminescence spectra were measured. As can be seen in Fig. 14.14, C-Hyd-VSiβ exhibits three different types of vibrational fine structure, α, β, and γ due to the presence of different kinds of tetrahedral V5+-oxide species. The spectrum of C-Hyd-VSiβ is less resolved than that of C-VSiβ and exhibits mainly the α type species. The energy separation between the (0 ! 0) and (0 ! 1) vibrational transitions can thus be determined from the second derivative spectrum, and the vibrational energies are found to be 1018 cm
1 for α species, 1054 cm
1 for β species, and 1036 cm
1 for γ species, in good agreement with the vibrational energy for V¼O bond obtained by IR and Raman measurements for gaseous O¼VF3, liquid O¼VCl3, and O¼VBr3 molecular compounds and various supported vanadium oxide catalysts (V2O5/SiO2, V2O5/PVG, and V2O5/γ-Al2O3, where PVG stands for porous Vycor glass) [50, 52, 54].

14.4.3 Mo-Oxide Single-Site Containing Samples So far the selective photo-oxidation of miner CO in the presence of H2 by O2 is concerned an important reaction in fuel cell technology since it is essential to selectively remove CO impurities from H2 without any consumption of H2 [55], otherwise, electrode performance is poisoned. Anpo et al. found that Mo-oxide single-site heterogeneous catalysts worked as a unique photocatalyst to oxide CO into CO2 without any oxidation of rich H2 into H2O, leading to the selective oxidation of CO in rich H2 [56]. They have proposed a plausible photocatalytic reaction mechanism shown in Fig. 14.15 based on the known structures determined by XAFS of the tetrahedrally coordinated Mo-oxide single-site species. The CT excited triplet state of the Mo-oxide species is implicated to play an important role in the reaction and it was clearly observed. The good quantum yields of this reaction were also obtained. Such Mo-oxide single-site photocatalytic reaction to purify H2 from its unwanted CO impurity is an alternative to the thermally activated heterogeneous

14

Photoluminescence (PL) Spectroscopy

309

J

(b) x 1 J

SiOH

D

Intensity (a.u.)

D

(c) x 1.5 E DE D D E

D D

(a) x 2.5

400

450

500

(b) (a) (c) 0

20 40 60 Time (ms)

J

D

0-1 (E)

Intensity (a.u.)

D

J

Intensity (a. u.)

J

J

E

D

E

0-1 (J)

550

600

D E

J

D

D E

a) E

J

J

b)

J

650

Wavelength (nm)

c)

0-1 (D)

Fig. 14.13 Photoluminescence spectra of VSiβ (curve a)‚ C-VSiβ (curve b), and C-Hyd-VSiβ (curve c) recorded at 77 K. All the catalysts were evacuated at 353 K for 2 h [49]. C: calcination at 773 K, Hyd: further rehydration at 298 K

catalytic reaction to achieve the same end using a Pt-Fe and other noble-metal catalysts supported on alumina. In addition to these catalysts, currently various kinds of other single-site catalysts constructed in MOFs and on MCFs have been reported by Yamashita et al. [57–62] and Zhang et al. [63].

14.4.4 Carbon Containing Samples Polymeric carbon nitride (CN), the most stable allotrope of covalent carbon nitride solids at ambient conditions, have been successfully introduced as a new metal-free visible light responsive photocatalyst for the splitting or decomposition of water to produce H2 since the pioneering work by Wang et al. in 2009 [64], owing to their unique electronic band structure. Since then, many strategic studies have been adopted to modify its physical and chemical properties to enhance their activity by doping, sensitization, nanostructured design, and hybridization, as well as copolymerization. In addition to those treatments, the use of different precursors that contain CN core structures for its synthesis has also been proven to be an effective means of enhancing the photocatalytic performance of CN. Wang and coworkers have advanced the strategy by employing cheap and easily available elemental sulfur as the external sulfur species instead of sulfur-containing precursors for the sulfur-mediated synthesis of polymeric carbon

400

D 450

D

D 500

D 550

600

Wavelength (nm)

Fig. 14.14 Second derivative photoluminescence spectra of VSiβ (curve a)‚ C-VSiβ (curve b), and C-Hyd-VSiβ (curve c). All the catalysts were evacuated at 353 K for 2 h. The decay curves (insets a, b, and c, logarithmic scale) were monitored at 77 K [49]

nitride photocatalysts [65]. They found that a strong PL peak is detected for all CN samples at room temperature, as shown in Fig. 14.16, which indicates the generation and recombination of photo-formed electrons and holes by following of the absorption of light at around 400 nm by carbon nitride semiconducting materials. The quenching of fluorescence was also observed for CN-Sx, indicating that the radiative recombination of electron and hole in the polymer CN is efficiently suppressed after the treatment with sulfur, S8-mediated synthesis. These results clearly indicated that the optimized textural structure and electronic properties of polymeric CN easily facilitates and controls the fate of the photo-formed electrons and holes, in other words, the charge separation can be controlled by the treatment of CN by various elements. As mentioned above, intrinsic radiationless transition causes serious quenching of fluorescence, which is an unwanted process in the photocatalysis. Furthermore, with the treatment by sulfur, a slight red-shift of fluorescence emission peak of CN was found to change from 460 to 493 nm, due to its bandgap narrowing effect associated with the S8-mediated synthesis. The modification of semiconductor materials by doping of metal cluster was also applied to enhance the photocatalytic

14

310

Q. Li et al.

CO + ½ O2 (in H2)

CO2

O2

4

CO2

199,60

0

2 4 Distance (Å)

Tetrahedrally coordinated Mo6+-oxide single-site CO

190 10 20 30 40 Reaction time (min) Reoxidation of Mo4+ Quantum yield: reduced-species with O2 12.6% 0

λ

O

O

Mo4+

2150

O2

O O

O

d)

H2

510

610

Stern-Volmer plots Φ0/Φ=1+τ0kq[Q] CO

2.0

H2O

UV+CO

410

Mo5+

Φ0/Φ

Absorbance (a.u.)

2127

310

Formation of the charge transfer excited triplet state ∗ O–

O2–

O 2077

210

Wavelength (nm)

(CO)n

O2– Adsorption of CO on Mo4+ reduced-species

Emission from excited state

1. 68

Mo6+

1/2 O2

e)

hv

O2–

O2–

1. 90

CO

6

λ

2 0

200,40 Energy (eV)

Absorption

Mo-O

Intensity (a.u.)

6

~ ~

Amounts of gasses (μmol)

H2

23 Dark Light on ~ ~ 8

c) Mo=O

Magnitude (a.u.)

Normalized absorbance (a.u.)

b) a)

O2 1.5

2 CO

H2

2043 1.0 CO2

2100

2050

2000

Reduction of Mo-oxide single-site with CO

0

0.2

0.4

0.6

Concentration of quencher (μmol/l)

Wavenumber (cm–1)

Fig. 14.15 Photocatalytic selective elimination of CO with O2 in H2-rich fuel cells on Mo-oxide single-site heterogeneous catalyst. Insert figures: (a) photocatalytic reaction time profiles, (b) XAFS spectra, (c) UV-vis absorption and photoluminescence of Mo-oxide single-site

species, (d) Stern-Volmer plots for the quenching of photoluminescence by CO, O2, and H2, and (e) FT-IR spectra of the absorption of CO on Mo-oxide single-site species [33]

efficiency. The overall water splitting using Pt/CN photocatalysts without using any sacrificial agents has been reported by Wang group [66]. The creation of metal/polymer surface junctions promotes the interfacial redox reaction, which can be confirmed by a rapidly decreasing PL intensity by increased Pt loading (Fig. 14.17). The optimum activity was achieved at a Pt loading of 3 wt.%. When Pt or PtOx was singly deposited on the CN nanosheets, the sample exhibited very poor activity in both cases, which clearly suggested that the charge separation to form electron and hole are important to enhance the photocatalytic activity of such semiconductor photocatalysts to achieve a high overall splitting of water. Here it is shown an example to study the very fast movement of photo-formed electrons and holes in CN semiconducting material by the ultrafast transient absorption spectra (TA). Schlenker and coworkers presented the TA spectra of a bulk CN at select time points from 1 ps to 5 ns (Fig. 14.18), exhibiting broad positive-induced absorption features in the wavelength between 600 and 1350 nm, being attributable to

overlapping of TA of electrons and holes in the heptazine network [67].

14.5

Characterization of Acidic and Basic Sites by Means of Luminescent Probe Molecules and In Situ Photoluminescence Spectroscopy

The measurement of phosphorescence of benzophenone adsorbed on Ti/Al binary oxides clearly suggested the presence of its protonated form in addition to benzophenone hydrogen-bonded to surface OH groups [68, 69]. As can be seen in Fig. 14.19, a triplet-triplet (T-T) absorption spectrum is observed at around 570 nm, which is due to benzophenone ketyl radicals. Direct detection of timeresolved T-T transient absorption spectra of the adsorbed benzophenone ketyl radicals indicated that those radicals are formed on the surface of Al2O3 via a hydrogen abstraction

14

Photoluminescence (PL) Spectroscopy

311

460 nm CN

I (a.u.)

S1.0

S2.0 S3.0 S5.0 CN-S10.0

S10 493 nm 450

500

550

600

650

l (nm)

Fig. 14.16 PL spectra of CN and the effect of the Sx doping into CN on in situ PL. PL were measured under the excitation at 400 nm and at 295 K [65]

were found to increase with UV-irradiation time, while the intensity of the phosphorescence of benzophenone decreased. Furthermore, the yield of benzhydrol was also found to strongly depend on the Ti/Al ratio. Benzpinacol was the only product observed in the photolysis of benzophenone adsorbed on SiO2 and porous Vycor glass [15–17, 68, 69]. These results clearly showed that surface acidic sites of Ti/Al binary oxides play a key role in the photolysis of benzophenone not only on adsorption surfaces but also on reactant species. As a matter of fact, protonation was found to be favored when benzophenone was adsorbed, thus leading to the formation of benzhydrol. The Bronsted basicity of a surface was found to relate with the deprotonation ability of molecules that can be probed by investigating the dissociative adsorption of protic molecular. The OLC2
$ OLCH
transformation thus induced can be followed by PL, which is one of the few techniques able to simultaneously characterize oxide ions and their protonated form. The same kind of equilibrium is also involved when a hydroxylated surface is undergoing thermal pretreatment. PL is thus an interesting tool to evaluate the surface basic properties of alkaline earth oxides.

14.6 0.0 wt% Pt

g-C3N4

In Situ Photoluminescence Studies of Photocatalytic Processes Involving Inorganic and Organic Semiconductor Photocatalytic Systems

Intensity (arb. units)

0.2

1.0 3.0

5 wt. % Pt-g-C3N4

5.0

450

500

550

600

l (nm)

Fig. 14.17 PL spectra of the CN and the effect of Pt deposition on in situ PL. PL were measured under the excitation at 400 nm and at 295 K [66]

from acidic surface OH groups by the excited triplet state of benzophenone, in addition to the T-T absorption spectrum at around 520–540 nm, due to adsorbed benzophenone [68]. After adsorption of benzophenone on Ti/Al binary oxides, the systems were UV-irradiated, and benzhydrol and benzpinacol were detected as major products of photolysis. Their yields

Anpo et al. [70] have investigated the photoluminescence behavior of TiO2 photocatalysts in the presence of various kinds of reactant molecules. The dependence of the photoluminescence intensity on the nature of the atmosphere was explained in terms of surface band bending of TiO2 particles, its extent depending on the electronegativity or electron affinity of the reactant molecules. Furthermore, such effect in photoluminescence intensity was found to be reversible after elimination of the reactant molecules by flowing N2, allowing to propose the idea to use such effect for remotecontrolled gas sensors. Murabayashi et al. have studied the photoluminescence properties of rutile and anatase TiO2 powders at room temperature in air in the absence or presence of reactants such as methanol or ethanol, in order to understand the mechanism of gas-phase photocatalytic reactions [18]. For rutile, they observed that the photoluminescence intensity in the presence of methanol or ethanol linearly increases with the square root of UV-irradiation time. They also found a linear time dependence for the integrated amount of photoinduceddesorbed O2, as can be seen on Fig. 14.20. The time dependence of the photoluminescence intensity and the effect of O2 photoinduced-desorption was explained as mentioned above in terms of band bending of the surfaces on the powders.

14

312

Q. Li et al.

b)

h+ 1 ps 10 ps 50 ps 100 ps 500 ps 1 ns 5 ns

3 2 1 0

0.5 0.0 –0.5

–1 600

600

800 1000 1200 Wavelength (nm)

1.0 0.8 0.6 0.4 0.2 0.0

4.89 ns 130 ps 0

1

800 1000 Wavelength (nm)

2 3 4 Time (ns)

5

1200

50

25

20

T-T absorption of benzophenoneketyl radical (60 Ps)

1

15

Methanol Ethanol 1-propanol 2-propanol

40 PL intensity (a.u.)

T-T absorption of benzophenone (24 Ps)

% absorption

1.0

EADS (norm ΔOD)

e–

SE

EADS (ΔmOD)

a)

ΔmOD

Fig. 14.18 (a) Transient absorption spectra of graphitic carbon nitride in visible to nearinfrared regions at selected time points from 1 picosecond to 5 nanosecond. (b) Evolutionassociated decay spectra and kinetics from global analysis. (Reproduced with permission from Schlenker [67])

30 20

0 Ps

10

1 Ps

0

2 10

0

3

2

6

4

8

Square root of time (s ) ½

10 Ps 5 4

0

100 Ps

Fig. 14.20 PL intensities of rutile TiO2 (Kanto-R) in vacuum and in air vs. the square root of UV irradiation time from 1 s to 60 s [4]

5 24,000 Ps 400

500

600

700

Wavelength (nm)

Fig. 14.19 Time-resolved Triplet-Triplet transient absorption spectra of BP adsorbed on Al2O3 following excitation at 355 nm, (pulse duration) 1.6 μs, pulse intensity: 10 mJ). After laser flash, 1; 0 μs, 2; 1 μs, 3; 10 μs, 4; 100 μs, 5; 24,000 μs [69]

These results, however, were not observed for anatase powders. The authors explained that such difference in the photoluminescence behaviors was due to the difference of photocatalytic activity toward these alcohols. It is important to understand the dynamics of photogenerated electrons and holes as well as their role in the redox reactions on the surfaces in order to design and develop efficient photocatalysts [3, 42, 43]. As reported earlier [3], the presence of a small amount of Pt particles on TiO2 photocatalyst is known to dramatically increase the efficiency of

photocatalytic reactions. This is attributed to an easy migration of photo-formed electrons from the conduction band of TiO2 to the Pt particles while the holes remain in the valence band, leading to an efficient charge separation of the photogenerated carriers, in addition to the catalytic effect of Pt to easily reduce H+ to form H2. However, the direct observation of the photo-formed electron-hole charge separation process in TiO2 photocatalyst is not easy. The time-resolved pumpprobe spectroscopy is a powerful method to investigate such dynamics of charge carriers with high time resolution. Furube et al. investigated the dynamics of charge separation of electrons and holes photogenerated in TiO2 and Pt-loaded TiO2 (Pt/TiO2) photocatalysts by transient absorption measurements in the whole visible region using femtosecond diffuse reflectance spectroscopy [19–21]. Transient absorption spectra and their time profiles of the intensity for TiO2 and Pt/TiO2 are shown in Fig. 14.21, with an excitation wavelength of 390 nm. A spectral shape with

14

Photoluminescence (PL) Spectroscopy

313

Fig. 14.21 Transient absorption spectra and their temporal profiles at 600 nm for non-loaded TiO2 and 0.2 wt.% Pt-loaded TiO2 power under 390 nm excitation [21]

Pt-loaded TiO2 (0.2 wt%)

TiO2

14

2 ps

0.5 ps

20

12

1.3 ps 4 ps

8

%absorption (%)

%absorption (%)

10

100 ps

6

15

100 ps

14 10

1 ns

1 ns 4 5 2

6 ns

0 400

500 600 Wavelength (nm)

0 400

700

500 600 Wavelength (nm)

700

%absorption (%) [TiO2]

12

20

10 600 nm

8

t = 5 ps

TiO2

15

Pt/TiO2

6

10

4 5

2 0

%absorption (%) [Pt/TiO2]

6 ns

0 0

10 20 Delay time (ps)

larger %-absorption at longer wavelength regions is observed for Pt/TiO2 for any delayed time, especially at shorter delay times. A fast decay at 4 picosecond (at around 4 or 5 picosecond) is observed at the %-absorption at around 600 nm wavelength regions for Pt/TiO2 while there is almost no absorption and decay for TiO2 [21]. Figure 14.22 shows that the decay of the transient absorption of Pt-loaded TiO2 is changed by the amount of Pt-loading, more rapid decay being observed for TiO2 with increasing Pt-loading. It was found that the inverse of the decay time is approximately proportional to Pt loading. These results clearly indicated that an electron migration from TiO2 moiety to Pt occurs, its extent depending on the amount of Pt. Thus, the charge separation in Pt/TiO2 photocatalysts was directly measured for the first time with a process taking place within a few ps [21]. On the other hand, slower electron-hole recombination was observed in the presence of air, which was explained in terms of upward band bending on/near the surface due to the presence of adsorbed O2.

0

1000

2000 3000 Delay time (ps)

4000

5000

Furthermore, femtosecond diffuse reflectance spectroscopy with a white continuum probe pulse has been applied to detect the dynamics of hole transfer from photoexcited TiO2 to adsorbed reactant molecules. As shown in Fig. 14.23, at lower pH than pH 7 of the TiO2 aqueous suspension with KSCN, ultrafast hole transfer was found to take place in less than 1 ps. Subsequent structure stabilization of dimer anion radicals, (SCN)2
, within a few picosecond and slow hole transfer with a time constant of a few hundred picosecond were clearly observed. Fast hole transfer was also found to be due to a surface-trapped state interacting strongly with adsorbed molecules. On the other hand, slow hole transfer observed at higher pH than pH 7 was found to be due to deep trapped states with a Boltzmann distribution of energy levels. Thus, the time-resolved spectroscopic techniques can be applied to the heterogeneous photocatalytic reaction systems to understand its dynamics. Such direct analysis of the interfacial CT for various kinds of photocatalysts such as surfacemodified, size-controlled, ion-implanted and doped TiO2, and hybrid multilayer systems such as SiO2-TiO2 binary

314

Q. Li et al.

20

600 nm

5 ps

%absorption (%)

in acidic (pH ¼ 1.2) than in alkaline (pH ¼ 12.8) solution. These results were attributed to the photo-formation of a hole at a surface lattice O2
site, which upon nucleophilic attack of a H2O molecule lead to oxygen evolution. Fornasiero and Murray et al. performed the ultrafast TA spectroscopy of the TiO2 nanorod films to quantify lifetimes of the photoexcited carriers [78]. The interesting feature is the decay of the trapped holes showing a recombination with electrons or reaction to oxidize adsorbates. The detailed dynamics will be described and discussed in a case study by Martra et al.

15 Pt-loaded TiO2 0.2 wt% 1.0 wt% 2.0 wt%

1.4 ps 10 0.6 ps 5

0 –5

0

5 10 Delay time (ps)

15

20

Fig. 14.22 Dependence of transient absorption decay at 600 nm for Pt-loaded TiO2 powder with different wt.% of Pt loading by the excitation at 390 nm [21]

Hole transfer rate (109 s–1)

10 7 6 5 4 3 2

1 7 6 5 4 0

1

2

3

4

5

6

7

pH

Fig. 14.23 The rates of hole transfer from photoexcited TiO2 to SCN
are plotted against pH of the solution [77]

oxides is expected to contribute to designing of an efficient TiO2-based photocatalysts [71–76]. Primary intermediates of O2 evolution in the photodecomposition of H2O at the water-rutile TiO2 interface was investigated by in situ multiple internal reflection IR absorption and photoluminescence. From the spectroscopic data, Nakato et al. [75] proposed a reaction mechanism in which O2 photoevolution is initiated by nucleophilic attack of H2O on a photogenerated hole at a surface lattice O2
site, and not by oxidation of a surface OH group by a hole. For a photoetched rutile TiO2 electrode, Nakato et al. observed a photoluminescence band at around 820 nm with an intensity higher

14.7

Effect of Temperature on Photoluminescence Spectra

The effect of temperature on photoluminescence spectra is important as regards to the internal conversion between the lowest excited triplet state and its ground state since the lifetime of the excited triplet states is long. As a result, the radiationless deactivation process in the excited triplet states is efficient. Therefore, phosphorescence spectra can be observed only when the system is frozen at liquid nitrogen temperature to reduce the efficiency of radiationless processes. For example, pyridine molecules show neither fluorescence nor phosphorescence at room temperature, however, these can be observed in rigid glass state at liquid nitrogen temperature. In general, the intensity of emission decreases with increasing temperature due to higher probabilities of the radiationless deactivation at the excited state. Furthermore, much better resolution of the vibrational fine structure of the emission (fluorescence and phosphorescence) can be observed at low temperature such as rigid glass state at liquid nitrogen temperature. The phosphorescence spectrum of highly dispersed tetrahedrally coordinated V-oxide single oxide species is known to exhibit a beautiful vibrational fine structure due to the presence of the surface V¼O double bond. The measurement at low temperature is indispensable to observe such well-resolved beautiful phosphorescence spectrum with a vibrational fine structure. Thus, the internal conversion is a strongly temperature-dependent process, decreasing the temperature of the system causes a sharp increase in the emission, both for fluorescence and phosphorescence. In solid materials research field, the mapping technique has been applied for defect profiling in SiC wafers, including photoluminescence, spectroscopic optical transmission, capacitive contactless resistivity, Raman imaging, and thermally stimulated luminescence (TSL) [79–81]. When the temperature was increased from 4.2 to about 150 K, the photoluminescence intensity was reduced to a level below the sensitivity of the photoluminescence setup; however, above 165 K the photoluminescence spectrum was found to

Photoluminescence (PL) Spectroscopy

315

show a new broad “orange” luminescence with the spectral maximum at 1.82 eV, assigned hereafter as the O-band. The O-band persisted also above room temperature, as shown in Fig. 14.24. Thus, photoluminescence spectra could be analyzed for Al- and B-doped p-type silicon carbide single crystals showing broad spectra between 0.7 eV and 3.25 eV at temperatures up to 300 C. The O-band was found to compose at least two strongly overlapped sub-bands with individual maximum at 1.86 (O1) and 1.50 eV (O2).

30

PL intensity (arb. un.)

PL

25 20

1.2

1.6

15

2.0

2.4

(eV)

4 3

10

2 5

1

0 1.2

1.4

1.6

1.8 2.0 Energy (eV)

2.2

2.4

2.6

Fig. 14.24 Photoluminescence spectra in Al-doped (4H) SiC at temperatures (1) 295 K, (2) 335 K, (3) 343 K, and (4) 403 K. Inset shows computer deconvolution of the 1.82 eV PL band on two individual Gaussian components [4]

a)

Recently, inorganic halide perovskite quantum dot (IPQD) system CsPbX3 (X ¼ Cl, Br, I) has been reported by Kovalenko et al. [82] as well as Li et al. [83] These systems attracted tremendous interest in solar cell research using different spectrum technology, especially, the temperaturedependent PL technology. As shown in Fig. 14.25 [81], the exciton-binding energy was determined by the temperature-dependent PL spectra, the PL peak at 15 K is 8 nm, presenting a typical strong excitonic PL character with few defect states according to the PL emission-excitation intensity relationship. A power law dependence (I / Lk) has been observed for both CPB-C, CsPbBr3 cubic phase and CPB-M QDs, CsPbBr3 monoclinic phase quantum dots with k values as 1.36 and 1.14, respectively. Such 1 < k < 2 coefficients revealed that the emissions are mainly from free and bound-excitons, which are associated with the long- and short-lived PL lifetimes. With temperature increasing, the PL peaks in Fig. 14.25a show a blue-shift and become weaker, which was previously observed from CH3NH3PbX3 and PbS QDs [85, 86]. At the same time, the PL intensity monotonously decreases with increasing temperature as plotted in Fig. 14.25b, from which the corresponding exciton-binding energy is derived as 40 meV. Such experimentally obtained Eb values are found to well match with the theoretical values of CsPbBr3 QDs, but slightly higher than that of bulk CsPbBr3 (35 meV) due to quantum confinement effect [74–89]. Polymeric semiconducting photocatalysts are widely used in the solar energy conversion and environmental pollutants degradation benefiting from the perfect physicochemical

b)

PL intensity (a.u.)

15 K 30 K 50 K 70 K 100 K 130 K 150 K 170 K 200 K 230 K 250 K 273 K 500

520

540

Wavelength (nm)

560

Integrated intensity (a.u.)

14

R2=0.98 Eb=40 meV >26 meV

0.00

0.01

0.02

0.03

1/T (K–1)

Fig. 14.25 Excitonic PL and its decay characteristics of IPQDs. (a) Temperature-dependent PL spectra. (b) Integrated PL emission intensity as a function of temperature from 30 to 273 K. (Reproduced with permission from Zeng [84])

14

316

Q. Li et al.

property. The fast photo-formed charge carrier separation and transfer to the surfaces is the key factors for improving the efficiency of these materials for solar energy utilization. Exciton-binding energy (Eb) has been regarded as a crucial parameter for mediating the charge separation. Minimizing the exciton-binding energy of the polymers leads to the high yield of the charge carrier generation, thus improving the photocatalytic activities. Wang et al. [90] reported a series of linear donor-acceptor conjugated polymers with low Eb by modulating the charge transfer pathway. The results clearly revealed that reducing the energy loss at the charge transfer state facilitates electron transfer from donor to acceptor, as a result, more electrons are ready to lead subsequent reduction reactions. The exciton-binding energy, Eb of these polymers could be estimated from the temperature-dependent PL spectra. As shown

b)

Eb = 104 meV

400 450 500 550 600 650 700 Wavelength (nm) 0.00

0.01

0.02

0.03

0.04

0.05

0.06

Eb = 91 meV

FSO-F

400 450 500 550 600 650 700 Wavelength (nm) 0.07

0.00

0.01

0.02

1/T (K–1)

d)

0.02

0.03 1/T

0.04 (K–1)

0.05

0.06

0.07

0.05

0.06

Eb = 105 meV

15 K 25 K 50 K 100 K 150 K 200 K 250 K 300 K

Intensity (a.u.)

Integrated intensity (a.u.)

15 K 25 K 50 K 100 K 150 K 200 K 250 K 300 K

Intensity (a.u.) 0.01

0.04

FSO-FSZ

400 450 500 550 600 650 700 Wavelength (nm) 0.00

0.03

1/T (K–1) Eb = 88 meV

FSO-FS

Integrated intensity (a.u.)

c)

15 K 25 K 50 K 100 K 150 K 200 K 250 K 300 K

Intensity (a.u.)

15 K 25 K 50 K 100 K 150 K 200 K 250 K 300 K

Intensity (a.u.)

Integrated intensity (a.u.)

FSO-BP

Integrated intensity (a.u.)

a)

in Fig. 14.26, upon a decrease in the temperature, the integrated PL peak intensity increases monotonically, and the corresponding Eb could be experimentally obtained by fitting these data using the Arrhenius equation, I(T ) ¼ I0/(1 þ A exp (
Eb/kBT)). Accordingly, the Eb of FSO-BP conjugated polymer is derived at 104 meV, and is due to its more planar structure leading to a more delocalized charge transfer pathway, while FSO-F conjugated polymer has a much lower Eb (91 meV). Based on the computed results, a planar structure and extended conjugation are beneficial for minimizing Eb through the construction of a more delocalized pathway for charge transfer. Thus, minimizing the Eb of the linear DonorAcceptor conjugated polymers with high planarity and extended conjugation led to exhibit an outstanding photocatalytic performance with visible light.

400 450 500 550 600 650 700 Wavelength (nm) 0.07

0.00

0.01

0.02

0.03 1/T

0.04

0.05

0.06

0.07

(K–1)

Fig. 14.26 Integrated PL emission intensity as a function of temperature (Inset: temperature-dependent PL spectra, λexcitation ¼ 375 nm) of polymers [90]. FSO: dibenzothiophene-S,S-dioxide, BP: biphenyl, F: fluorine, FS: dibenzothiophene

Photoluminescence (PL) Spectroscopy

14.8

317

Effect of Magnetic Fields on Photoluminescence Spectra

14.9

Polimeni et al. [91] investigated the zincblende phase GaAs nanowires with wurtzite (WZ) lattice using the magnetophotoluminescence under high magnetic fields (B ¼ 0–28 T ) at T ¼ 4.2 and 77 K. The PL spectra exhibited an extremely narrow ( 1 meV) emission line at 1.522 eV, originating from the band gap exciton recombination in WZ GaAs. Magnetic fields up to 28 T were applied along and perpendicular to the WZ ĉ axis, in order to disclose possible band structure anisotropies related to the lattice hexagonal symmetry. Figure 14.27 shows the PL spectra of WZ GaAs in the !

!

Voigt configuration (B ⊥k // ĉ) recorded at T ¼ 77 K under different magnetic fields. This PL was attributed to the band gap free exciton (FE) recombination. With increasing magnetic field, B, the energy of the PL peak max exhibits blue shifts, while, as shown in panel (b), a narrowing of the PL line is observed. Such narrowing can be attributed to a consequence of the magnetic field-induced modification of the density of states from 3D- to 1D-like, its extent being apparent at higher temperature. In addition, panel (c) shows that the PL spectra are identical upon circular polarization filtering.

The basic principles and practical aspects of photoluminescence techniques and their applications to the adsorption, catalysis, and photocatalysis phenomena were introduced, which exhibited from various solid catalysts and photocatalytic materials in their working states and summarized for the complete understanding of the surface active sites and catalytic processes at the molecular level with a high sensitivity and nondestructive nature. Compared with other spectroscopic techniques, photoluminescence opens up the possibility of directly observing surface lattice O2
ions in various positions with different coordination numbers and offer insights into the reactivity, dynamics of excited states, and determination of the absolute reaction rate constants as well as the dynamic aspects of intermediate species in catalytic and photocatalytic reactions. Also, the photoluminescence is applied to study oxide-supported sulfides as well as unsupported or oxide-supported (oxi)carbides or (oxi)nitrides with better defined characteristics. In order to have more comprehensive views of the working states of catalysts and photocatalysts, the combined utilization of other physical techniques and spectroscopies including the diffuse reflectance, electron paramagnetic resonance, and fast timescale IR and UV as well as Raman techniques are

a)

b) →

→

T = 77 K

T = 77 K

FE

Γ5/6

∧

B ⊥ k // c

Line narrowing

B = 28 T B=0T

26 T

PL intensity (arb. units)

Fig. 14.27 (a) Photoluminescence spectra of WZ GaAs NWs in Voigt configuration at T ¼ 77 K under different magnetic fields. (b) Comparison between normalized PL spectra at B ¼ 0 and 28 T, highlighting the absence of a sizable Zeeman splitting and the presence of a line narrowing induced by the magnetic field. The 28 T spectrum has been red-shifted for ease of comparison with the 0 T spectrum. Relative multiplication factors are provided. (c) Comparison between PL spectra at B ¼ 27 T recorded under opposite circular light polarizations, showing the absence of circular dichroism in the Voigt configuration. (Reproduced with permission from Polimeni [91])

Conclusions and Outlook

x 1.3

24 T

Γ5/6

22 T

B = 28 T (shifted)

20 T 1.48

18 T

1.50

x1

1.52

1.54

Energy (eV)

16 T

c)

14 T

T = 77 K B = 27 T

FE

12 T 10 T 8T

No circular dichroism

V+ V–

6T 4T

Γ5/6

2T 0T 1.50

FE 1.52 Energy (eV)

1.54

1.50

1.52

1.54

Energy (eV)

1.56

PL intensity (arb. units)

14

14

318

suggested. Such complete understanding will provide us very useful information to rational design and develop the highly functional catalytic and photocatalytic systems, which will benefit the environment and mankind. In the twenty-first century, energy depletion and environmental pollution on a global scale are the most serious and urgent issues facing mankind. It is, thus, of vital importance to develop effective catalytic and photocatalytic systems for novel and clean energy production such as the conversion of CO2 released in air and the production of clean H2 from water. Such developments will allow sustainable economic growth and development without environmental destruction or pollution. Various approaches have been intensively investigated to address these issues and trials have also been successfully carried out. We hope and believe that photoluminescence techniques will play an important role in the developments of such efficient and effective catalytic and photocatalytic materials.

References 1. Anpo, M., Tanahashi, I., Kubokawa, Y.: Photoluminescence and photoreduction of vanadium pentoxide supported on porous Vycor glass. J. Phys. Chem. 84, 3440 (1980) 2. Anpo, M., Kondo, M., Coluccia, S., Louis, C., Che, M., Coluccia, S.: Application of dynamic photoluminescence spectroscopy to the study of the active surface sites on supported molybdenum/silica catalysts: features of anchored and impregnated catalysts. J. Am. Chem. Soc. 111, 8791 (1989) 3. Anpo, M., Che, M.: Applications of photoluminescence techniques to the characterization of solid surfaces in relation to adsorption, catalysis, and photocatalysis. Adv. Catal. 44, 119 (1999) 4. Anpo, M., Dzwigai, S., Che, M.: Applications of photoluminescence spectroscopy to the investigation of oxide containing catalysts in the working state. Adv. Catal. 52, 1 (2009) 5. Li, Q.H., Anpo, M., Wang, X.C.: Application of photoluminescence spectroscopy to elucidate photocatalytic reactions at the molecular level. Res. Chem. Intermed. 46, 4325 (2020) 6. Barrow, G.M.: Physical Chemistry. McGraw-Hill, New York (1988) 7. Atkins, P.W.: Physical Chemistry. Oxford University Press, Oxford (1986) 8. Daniels, F., Alberty, R.A.: Physical Chemistry. Wiley, New York (1975) 9. Olmsted, J., Williams, G.M.: Chemistry. Mosby, St. Louis (1994) 10. Calvert, J.G., Pitts Jr., J.N.: Photochemistry. Wiley, New York (1966) 11. Simons, J.P.: Photochemistry and Spectroscopy. Wiley Interscience, London (1971) 12. Turro, N.J.: Molecular Photochemistry. Benjamin, New York (1967) 13. Turro, N.J.: Modern Molecular Photochemistry. Benjamin/Cummings, Menlo Park (1978) 14. Demas, J.N.: Excited State Lifetime Measurements. Academic, New York (1984) 15. Anpo, M., Kubokawa, Y.: Photoinduced and photocatalytic reactions on supported metal oxide catalysts. Excited states of oxides and reaction intermediates. Res. Chem. Intermed. 8, 105 (1987) 16. Anpo, M.: Photocatalysis on small particle TiO2 catalysts. Reaction intermediates and reaction mechanisms. Res. Chem. Intermed. 9, 67 (1989)

Q. Li et al. 17. Anpo, M., Yamada, Y., Coluccia, S., Zeccina, A., Che, M.: Photocatalysed isomerization of butenes on MgO powders with coordinatively unsaturated surface ions. J. Chem. Soc. Faraday Trans. I. 85, 609 (1989) 18. Nakajima, H., Itoh, K., Murabayashi, M.: Influences of C1–C3 alcohols and purities of TiO2 powders on their photoluminescence properties at room temperature. Bull. Chem. Soc. Jpn. 75, 601 (2002) 19. Garrone, E., Stone, F.: In: Che, M., Bond, G.C. (eds.) Adsorption and Catalysis on Oxide Surface, p. 97. Elsevier, Amsterdam (1985) 20. Che, M., Giamello, E.: In: Fierro, J.L.G. (ed.) Spectroscopy Characterization of Heterogeneous Catalysts, vol. B, p. 265. Elsevier, Amsterdam (1990) 21. Furube, A., Asahi, T., Masuhara, H., Yamashita, H., Anpo, M.: Direct observation of a picosecond charge separation process in photoexcited platinum-loaded TiO2 particles by femtosecond diffuse reflectance spectroscopy. Chem. Phys. Lett. 336, 424 (2001) 22. Kubokawa, Y., Anpo, M.: In: Che, M., Bond, G.C. (eds.) Adsorption and Catalysis on Oxide Surface, p. 127. Elsevier, Amsterdam (1985) 23. Tang, J., Durrant, J.R., Klug, D.R.: Mechanism of photocatalytic water splitting in TiO2. Reaction of water with photoholes, importance of charge carrier dynamics, and evidence for four-hole chemistry. J. Am. Chem. Soc. 130, 13885 (2008) 24. Kubacka, A., Fernandez-Garcia, M., Colon, G.: Advanced nanoarchitectures for solar photocatalytic applications. Chem. Rev. 112, 1555 (2012)., and references therein 25. Pesci, F., Wang, G., Klug, D.R., Li, Y., Cowan, A.J.: Efficient suppression of electron-hole recombination in oxygen-deficient hydrogen-treated TiO2 nanowires for photoelectrochemical water splitting. J. Phys. Chem. C. 117, 25837 (2013) 26. Anpo, M., Yamashita, H., Ikeue, K., Fujii, Y., Zhang, S.G., Ichihashi, Y., Park, D.R., Suzuki, Y., Koyanao, K., Tatsumi, T.: Photocatalytic reduction of CO2 with H2O on Ti-MCM-41 and TiMCM-48 mesoporous zeolite catalysts. Catal. Today. 44, 327 (1998) 27. Ikeue, K., Yamashita, H., Anpo, M.: Photocatalytic reduction of CO2 with H2O on titanium oxides prepared within the FSM-16 mesoporous zeolite. Chem. Lett. 28, 1135 (1999) 28. Zhang, J.L., Minagawa, M., Ayuzawa, T., Natarajan, S., Yamashita, H., Matsuoka, M., Anpo, M.: In situ investigation of the photocatalytic decomposition of NO on the Ti-HMS under flow and closed reaction systems. J. Phys. Chem. B. 104, 11501 (2000) 29. Anpo, M., Higashimoto, S., Shioya, Y., Ikeue, K., Harada, M., Watanabe, M.: Characteristics in the photocatalytic reactivity of the tetrahedrally coordinated Ti-oxide species designed within various types of zeolites and on support surfaces. Stud. Surf. Sci. Catal. 140, 27 (2001) 30. Ikeue, K., Yamashita, H., Anpo, M., Takewaki, T.: Photocatalytic reduction of CO2 with H2O on Ti-β zeolite photocatalysts: effect of the hydrophobic and hydrophilic properties. J. Phys. Chem. B. 105, 8350 (2001) 31. Zhang, J.L., Hu, Y., Matsuoka, M., Yamashita, H., Minagawa, M., Hidaka, H., Anpo, M.: Relationship between the local structures of titanium oxide photocatalysts and their reactivities in the fecomposition of NO. J. Phys. Chem. B. 105, 8395 (2001) 32. Ogawa, M., Ikeue, K., Anpo, M.: Transparent self-standing films of titanium-containing nanoporous silica. Chem. Mater. 13, 2900 (2001) 33. Anpo, M., Thomas, J.M.: Single-site photocatalytic solids for the decomposition of undesirable molecules. Chem. Commun. 31, 3273 (2006) 34. Anpo, M., Takeuchi, M., Ikeue, K., Dohshi, S.: Design and development of titanium oxide photocatalysts operating under visible and UV light irradiation: the applications of metal ion-implantation techniques to semiconducting TiO2 and Ti/zeolite catalysts. Curr. Opin. Solid State Mater. Sci. 6, 381 (2002) 35. Matsuoka, M., Anpo, M.: Local structures, excited states, and photocatalytic reactivities of highly dispersed catalysts constructed within zeolites. J Photochem Photobiol C: Photochem Rev. 3, 225 (2003)

14

Photoluminescence (PL) Spectroscopy

36. Hu, Y., Higashimoto, S., Martra, G., Zhang, J.L., Matsuoka, M., Coluccia, S., Anpo, M.: Local structures of active sites on Ti-MCM41 and their photocatalytic reactivity for the decomposition of NO. Catal. Lett. 90, 161 (2003) 37. Anpo, M. (ed.): Photofunctional Zeolite-Synthesis, Characterization, Photocatalytic Reactions, Light Harvesting. Nova Science Publishers, New York (2000) 38. Yamashita, H., Ikeue, K., Takewai, T., Anpo, M.: In situ XAFS studies on the effects of the hydrophobic-hydrophilic properties of Ti-Beta zeolites in the photocatalytic reduction of CO2 with H2O. Top. Catal. 18, 95 (2002) 39. Hu, Y., Martra, G., Higashimoto, S., Zhang, J.L., Matsuoka, M., Coluccia, S., Anpo, M.: The relationship between the local structures and photocatalytic reactivity of Ti-MCM-41 catalysts. Stud. Surf. Sci. Catal. 146, 593 (2003) 40. Yamashita, H., Anpo, M.: Local structures and photocatalytic reactivities of the titanium oxide and chromium oxide species incorporated within micro- and mesoporous zeolite materials: XAFS and photoluminescence studies. Curr. Opin. Solid State Mater. Sci. 7, 471 (2003) 41. Jung, K.Y., Park, S.B.: Effect of calcination temperature and addition of silica, zirconia, alumina on the photocatalytic activity of titania. Korean J. Chem. Eng. 18, 879 (2001) 42. Curr. Opin. Solid State Mater. Sci. 6, (2002) (Special issue on Photoluminescence, M. Anpo, ed.) 43. Matsuoka, M., Anpo, M.: Measurements of the time-resolved photoluminescence and reaction dynamics. Shokubai Catal. Catal. 47, 328 (2005) 44. Dzwigaj, S., Matsuoka, M., Franck, R., Anpo, M., Che, M.: Probing different kinds of vanadium species in the VSiβ zeolite by diffuse reflectance UV-visible and photoluminescence spectroscopies. J. Phys. Chem. B. 102, 6309 (1998) 45. Anpo, M., Zhang, S.G., Higashimoto, S., Matsuoka, M., Yamashita, H.: Characterization of the local structure of the vanadium silicalite (VS-2) catalyst and its photocatalytic reactivity for the decomposition of NO into N2 and O2. J. Phys. Chem. B. 103, 9295 (1999) 46. Higashimoto, S., Matsuoka, M., Yamashita, H., Anpo, M.: Investigation of the local structure of vanadium silicalite catalyst (VS-1) using XAFS, FT-IR and photoluminescence spectroscopic methods. Jpn. J. Appl. Phys. 38, 47 (1999) 47. Dzwigaj, S., Matsuoka, M., Anpo, M., Che, M.: Evidence of three kinds of tetrahedral vanadium (V) species in VSiβ zeolite by diffuse reflectance UV
visible and photoluminescence spectroscopies. J. Phys. Chem. B. 104, 6012 (2000) 48. Higashimoto, S., Matsuoka, M., Yamashita, H., Anpo, M., Kitao, O., Hidaka, H., Che, M., Giamello, E.: Effect of the Si/Al ratio on the local structure of V oxide/ZSM-5 catalysts prepared by solid-state reaction and their photocatalytic reactivity for the decomposition of NO in the absence and presence of propane. J. Phys. Chem. B. 104, 10288 (2000) 49. Dzwigaj, S., Matsuoka, M., Anpo, M., Che, M.: A comparative study of V species in ß zeolite by photoluminescence, diffuse reflectance UV-visible and 51V NMR spectroscopies. Catal. Lett. 72, 211 (2001) 50. Higashimoto, S., Matsuoka, M., Zhang, S.G., Yamashita, H., Kitao, O., Hidaka, H., Anpo, M.: Characterization of the VS-1 catalyst using various spectroscopic techniques and its unique photocatalytic reactivity for the decomposition of NO in the absence and presence of C3H8. Microporous Mesoporous Mater. 48, 329 (2001) 51. Anpo, M., Higashimoto, S., Matsuoka, M., Zhanpeisov, N., Shioya, Y., Dzwigaj, S., Che, M.: The effect of the framework structure on the chemical properties of the vanadium oxide species incorporated within zeolites. Catal. Today. 78, 211 (2003) 52. Hu, Y., Wada, N., Matsuoka, M., Anpo, M.: Preparation and characterization of highly dispersed gold nanoparticles within channels of mesoporous silica. Catal. Lett. 97, 49 (2004) 53. Matsuoka, M., Higashimoto, S., Yamashita, H., Anpo, M.: In-situ investigations of the photocatalytic reaction of no with propane on the vanadium silicalite-1 catalyst. Res. Chem. Intermed. 26, 85 (2000)

319 54. Dzwigaj, S.: Recent advances in the incorporation and identification of vanadium species in microporous materials. Curr. Opin. Solid State Mater. Sci. 7, 461 (2003) 55. Siani, A., Captain, D.K., Alexeev, O.S., Stafyla, E., Hungria, A.B., Midgley, P.A., Thomas, J.M., Adams, R.D., Amiridis, M.D.: Improved CO oxidation activity in the presence and absence of hydrogen over cluster-derived PtFe/SiO2 catalysts. Langmuir. 22, 5160 (2006) 56. Kamegawa, T., Matsuoka, M., Takeuchi, R., Anpo, M.: Photocatalytic oxidation of CO with various oxidants by Mo oxide species highly dispersed on SiO2 at 293 K. Catal. Today. 111, 248 (2006) 57. Horiuchi, Y., Toyao, T., Takeuchi, M., Matsuoka, M., Anpo, M.: Recent advances in visible-light-responsive photocatalysts for hydrogen production and solar energy conversion from semiconducting TiO2 to MOF/PCP photocatalysts. Phys. Chem. Chem. Phys. 15, 13243 (2013) 58. Mori, K., Kawashima, M., Che, M., Yamashita, H.: Enhancement of the photoinduced oxidation activity of a Ruthenium(II) complex anchored on silica-coated silver nanoparticles by localized surface plasmon resonance. Angew. Chem. Int. Ed. 49, 8598 (2010) 59. Mori, K., Tottori, M., Watanabe, K., Che, M., Yamashita, H.: Photoinduced aerobic oxidation driven by phosphorescence Ir(III) complex anchored to mesoporous silica. J. Phys. Chem. C. 115, 21358 (2011) 60. Mori, K., Watanabe, K., Che, M., Yamashita, H.: Anchoring of Pt(II) pyridyl complex to mesoporous silica materials: enhanced photoluminescence emission at room temperature and photooxidation activity using molecular oxygen. J. Phys. Chem. C. 115, 1044 (2011) 61. Che, M., Mori, K., Yamashita, H.: Elaboration, characterization and properties of silica-based single-site heterogeneous photocatalysts. Proc. R. Soc. A. 468, 2113 (2012) 62. Yamashita, H., Mori, K., Kuwahara, Y., Kamegawa, T., Wen, M., Verma, P., Che, M.: Single-site and nano-confined photocatalysts designed in porous materials for environmental uses and solar fuels. Chem. Soc. Rev. 47, 8072 (2018) 63. Xing, M.Y., Zhang, J.L., Qiu, J.B.C., Tian, B.Z., Anpo, M., Che, M.: A brown mesoporous TiO2-x/MCF composite with an extremely high quantum yield of solar energy photocatalysis for H2 evolution. Small. 11, 1920 (2015) 64. Wang, X.C., Maeda, K., Thomas, A., Takanabe, K., Xin, G., Carlsson, J.M., Domen, K., Antonietti, M.: A metal-free polymeric photocatalyst for hydrogen production from water under visible light. Nat. Mater. 8, 76 (2009) 65. Zhang, J.S., Zhang, M.W., Zhang, G.G., Wang, X.C.: Synthesis of carbon nitride semiconductors in sulfur flux for water photoredox catalysis. ACS Catal. 2, 940 (2012) 66. Zhang, G.G., Lan, Z.A., Lin, L.H., Lin, S., Wang, X.C.: Surface engineering of graphitic carbon nitride polymers with cocatalysts for photocatalytic overall water splitting. Chem. Sci. 7, 3062 (2017) 67. Corp, K.L., Schlenker, C.W.: Ultrafast spectroscopy reveals electron-transfer cascade that improves hydrogen evolution with carbon nitride photocatalysts. J. Am. Chem. Soc. 139, 7904 (2017) 68. Lewis, G.N., Kasha, M.: Phosphorescence and the triplet state. J. Am. Chem. Soc. 66, 2100 (1944).; Phosphorescence in fluid media and the reverse process of singlet-triplet absorption. 67, 994 (1945); Lewis, G.N., Calvin, M., and Kasha, M., Photomagnetism. Determination of the paramagnetic susceptibility of a dye in its phosphorescent state. J. Chem. Phys. 17, 804 (1947) 69. Terenin, A.N.: J. Phys. Chem. SSSR 14, 1362 (1940); Nishiguchi, H., Zhang, J.L., Anpo, M., and Masuhara, H.: Photochemical properties of benzophenone adsorbed on Ti-Al binary oxides: the effects of the surface acidity. J. Phys. Chem. B 105, 3218 (2001) 70. Anpo, M.: Preparation, characterization, and reactivities of highly functional titanium oxide-based photocatalysts able to operate under UV-Visible light irradiation: approaches in realizing high efficiency in the use of visible light. Bull. Chem. Soc. Jpn. 77, 1427 (2004)

14

320 71. Fujihara, K., Izumi, S., Ohno, T., Matsumura, M.: Time-resolved photoluminescence of particulate TiO2 photocatalysts suspended in aqueous solutions. J. Photochem. Photobiol. A. 132, 99 (2000) 72. Miyashita, K., Kuroda, S., Tajima, S., Takehira, K., Tobita, S., Kubota, H.: Photoluminescence study of electron-hole recombination dynamics in the vacuum-deposited SiO2/TiO2 multilayer film with photo-catalytic activity. Chem. Phys. Lett. 369, 225 (2003) 73. Yu, J.G., Yu, H.G., Cheng, B., Zhao, X.J., Yu, J.C., Ho, W.K.: The effect of calcination temperature on the surface microstructure and photocatalytic activity of TiO2 thin films prepared by liquid phase deposition. J. Phys. Chem. B. 107, 13871 (2003) 74. Furube, A., Asahi, T., Masuhara, H., Yamashita, H., Anpo, M.: Direct observation of interfacial hole transfer from a photoexcited TiO2 particle to an adsorbed molecule SCN-by femtosecond diffuse reflectance spectroscopy. Res. Chem. Intermed. 27, 177 (2001) 75. Nakamura, R., Nakato, Y.: Primary intermediates of oxygen photoevolution reaction on TiO2 (Rutile) particles, revealed by in situ FTIR absorption and photoluminescence measurements. J. Am. Chem. Soc. 126, 1290 (2004) 76. Chen, H.J., Matsuoka, M., Zhang, J.L., Anpo, M.: The effect of the Si/Al ratio on the photoluminescence properties of Cu(I)SAPO-5 catalysts. Chem. Lett. 33, 1254 (2004) 77. Chizallet, C., Bailly, M.L., Costentin, G., Lauron-Pernot, H., Krafft, J.M., Bazin, P., Saussey, J., Che, M.: Thermodynamic brønsted basicity of clean MgO surfaces determined by their deprotonation ability: role of Mg2+-O2
pairs. Catal. Today. 116, 196 (2006) 78. Cargnello, M., Montini, T., Smolin, S., Priebe, J., Jaén, J., DoanNguyen, V., McKay, I., Schwalbe, J., Pohl, M., Gordon, T., Lu, Y., Baxter, J., Brückner, A., Fornasiero, P., Murray, C.: Engineering titania nanostructure to tune and improve its photocatalytic activity. Proc. Natl. Acad. Sci. 113, 3966 (2016) 79. Korsunska, N.E., Tarasov, I., Kushnirenko, V., Ostapenko, S.: Hightemperature photoluminescence spectroscopy in p-type SiC. Semicond. Sci. Technol. 19, 833 (2004) 80. St. Amour, A., Stum, J.C., Lacroix, Y., Thewalt, M.L.W.: Enhancement of high temperature photoluminescence in strained Si1-xGex/ Si heterostructures by surface passivation. Appl. Phys. Lett. 65, 3344 (1994) 81. Bissiri, M., Gaspari, V., Polimeni, A., Baldassarri, G., von Hogersthal, H., Capizzi, M., Frova, A., Fischer, M., Reinhardt, M., Forchel, A.: High temperature photoluminescence efficiency and thermal stability of (InGa)(AsN)/GaAs quantum wells. Appl. Phys. Lett. 79, 2585 (2001) 82. Protesescu, L., Yakunin, S., Bodnarchuk, M.I., Krieg, F., Caputo, R., Hendon, C.H., Yang, R.X., Walsh, A., Kovalenko, M.V.: Nanocrystals of cesium lead halide Perovskites (CsPbX3, X = Cl, Br, and I): novel optoelectronic materials showing bright emission with wide color gamut. Nano Lett. 6, 3692 (2015) 83. Song, J., Li, J., Li, X., Xu, L., Dong, Y., Zeng, H.: Quantum dot light-emitting diodes based on inorganic Perovskite cesium lead halides (CsPbX3). Adv. Mater. 27, 7162 (2015) 84. Li, X.M., Wu, Y., Zhang, S.L., Cai, B., Gu, Y., Song, J.Z., Zeng, H. B.: CsPbX3 Quantum dots for lighting and displays: room-temperature synthesis, photoluminescence superiorities, underlying origins and white light-emitting diodes. Adv. Funct. Mater. 26, 2435 (2016) 85. Zhang, F., Zhong, H., Chen, C., Wu, X.G., Hu, X., Huang, H., Han, J., Zou, B., Dong, Y.: Brightly luminescent and color-tunable colloidal CH3NH3PbX3 (X-Br, I, Cl) quantum dots: potential alternatives for display technology. ACS Nano. 9, 4533 (2015) 86. Dey, P., Paul, J., Bylsma, J., Karaiskaj, D., Luther, J.M., Beard, M.C., Romero, A.H.: Origin of the temperature dependence of the band gap of PbS and PbSe quantum dots. Solid State Commun. 165, 49 (2013) 87. Kondo, S., Takahashi, K., Nakanish, T., Saito, T., Asada, H., Nakagawa, H.: High intensity photoluminescence of microcrystalline CsPbBr3 films: evidence for enhanced stimulated emission at room temperature. Curr. Appl. Phys. 7, 1 (2007)

Q. Li et al. 88. Fröhlich, D., Heidrich, K., Künzel, H., Trendel, G., Treusch, J.: Cesium-trihalogen-plumbates a new class of ionic semiconductors. J. Lumin. 18, 385 (1979) 89. Burschka, J., Pellet, N., Moon, S.J., Humphry-Baker, R., Gao, P., Nazeeruddin, M.K., Gratzel, M.: Sequential deposition as a route to high-performance perovskite-sensitized solar cells. Nature. 499, 316 (2013) 90. Lan, Z.A., Zhang, G.G., Chen, X., Zhang, Y.F., Zhang, K.A.I., Wang, X.C.: Reducing the exciton binding energy of donor-acceptor-based conjugated polymers to promote charge-induced reactions. Angew. Chem. Int. Ed. 58, 10236 (2019) 91. Luca, M.D., Rubini, S., Felici, M., Meaney, A., Christianen, P.C.M., Martelli, F., Polimeni, A.: Addressing the fundamental electronic properties of wurtzite GaAs nanowires by high-field magnetophotoluminescence spectroscopy. Nano Lett. 17, 6540 (2017)

Dr. Qinghe Li is now a postdoctor in Dalian Institute of Chemical Physics, Chinese Academy of Sciences. The co-advisors are Prof. Tao Zhang and Prof. Botao Qiao. He received his PhD in Physical Chemistry from Fuzhou University in 2020 with Prof. Xinchen Wang. His current research interest mainly focuses on tuning the product selectivity and mechanism study of CO2 hydrogenation on single-atom catalysts.

Prof. Masakazu Anpo is presently a Special Honor Professor and International Advisor of the State Key Laboratory of Photocatalysis on Energy and Environment, Fuzhou University in China. He worked for 40 years at Osaka Prefecture University (renamed as Osaka Metropolitan University) in Japan and served as Dean, Vice President, and Executive Director for the last 10 years and is now a Professor Emeritus. He is a pioneer in the research of photochemical reactions on solid surfaces, design of visible-light-responsive TiO2 photocatalysts, and single-site heterogeneous photocatalysts constructed within zeolites. He is the editor-in-chief of Res. Chem. Intermed. (Springer). He has published more than 110 books and 550 original peer-reviewed papers. He is a member of Academia Europaea and Science Council of Japan, and Honorary Fellow of Chinese Chemical Society.

14

Photoluminescence (PL) Spectroscopy

Prof. Jinmao You received his PhD degree in analytical chemistry from Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences in 1999. Then he joined Dalian Institute of Chemical Physics, Chinese Academy of Sciences for postdoctoral research for two years. In 2001, he joined Qufu Normal University as a Professor and served as the Dean of College of Chemistry and Chemical Engineering since 2010. He has also served as Guest Scientist in Western-north Institute of Plateau Biology, Chinese Academy of Sciences since 2003 and Shandong University since 2010. Now he works in Shaoxing University. His current research interests mainly focus on the development of new reagent, new material, and new method for chromatography-mass spectrometry in food, environment, medicine, life, etc. He is the author or coauthor of more than 400 peer-reviewed scientific publications with Hindex of 36. Professor You has won many important awards including the Second Prize of National Scientific and Technological Progress Award in 2007, Hundred-Talent Award Program from Chinese Academy of Science in 2008, and Qinghai Province Scientific and Technological Progress Award in 2012

Prof. Tingjiang Yan received his PhD degree from the Research Institute of Photocatalysis at Fuzhou University in 2009. After a brief research at Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, he joined the College of Chemistry and Chemical Engineering, Qufu Normal University in 2011. His current research interests primarily focus on the design and synthesis of highly efficient heterogeneous photocatalysts for solar energy conversion and environmental purification applications.

321

Prof. Xinchen Wang is the Director of the State Key Laboratory of Photocatalysis on Energy and Environment, and the Dean of the College of Chemistry at Fuzhou University. He received his PhD from The Chinese University of Hong Kong in 2005 with Prof. Jimmy Yu, and moved to Tokyo University as a JSPS post-doctor in 2006 with Prof. Domen. In 2007, he worked at the Max Planck Institute of Colloids and Interfaces, Germany, as an Alexander von Humboldt fellow, and was appointed as a group leader between 2008 and 2012. He was as the Fellow of The Royal Society of Chemistry (UK) in 2015, and Cheung Kong Scholar Program in 2016 (China). He is one of the pioneers in the field of carbon nitride photocatalysis. He developed polymeric carbon nitride photocatalysts for water splitting, CO2 reduction, water purification, air-pollution degradation, and organo-synthesis. He developed a series of synthesis approaches to promote the performance of carbon nitride-based photocatalysts and other conjugated polymer photocatalysts. He also attempts to address the fundamental issues in polymer photocatalysts, including visible light absorption, charge separation and transport, and kinetics of the redox reaction. To date, he has achieved more than 250 peer-reviewed publications, including Nat. Mater., Nat. Rev. Mater., J. Am. Chem. Soc., Angew. Chem. Int. Ed., Nat. Commu., and more. His total citations are over 43,000 times with H index of 101.

14

Case Studies: Photoluminescence (PL) Spectroscopy

Lorenzo Mino

15

, Masaya Matsuoka, and Gianmario Martra

Contents 15.1 15.1.1 15.1.2 15.1.3 15.2 15.2.1 15.2.2

Investigation of Charge Carrier Dynamics in Photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Time-Resolved PL Studies of Pure and Doped TiO2 Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PL Studies of Reactant Interactions with TiO2 Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PL Investigation of Charge Carrier Separation in g-C3N4 Heterojunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

323 324 324 326

Investigation of Photocatalytic Reactions Promoted by Supported Transition Metal Ions . . . . . . . . . . . . . . . . . . . 327 NO Photo-Reduction by CO on Mo6+/SiO2 . . . . . . . . . . . . . 328 Photo-PROX Reaction on Visible Light Responsive Cr6+-MCM-41 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332

on silica supports, discussing in particular PL studies of NO photo-reduction by CO on Mo6+/SiO2 and preferential photo-oxidation of CO by O2 on Cr6+/MCM-41. Keywords

TiO2 · g-C3N4 · Charge carrier dynamics · Silicasupported transition metal ions · Kinetics modeling · Photoreaction mechanisms · In situ photoluminescence spectroscopy · Time-resolved photoluminescence spectroscopy

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334

Abstract

The chapter presents two representative case studies, which highlight the potentialities of photoluminescence (PL) spectroscopy for the study of photocatalytic systems at the molecular level in their working state. The first part of the chapter is devoted to the investigation of charge carrier dynamics in semiconductor photocatalysts (TiO2 and g-C3N4) by steady-state and ultrafast time-resolved PL, focusing also on the interaction with the semiconductor surface of key reactants in photocatalytic processes. The second part, on the other hand, describes single-site photocatalytic systems based on transition metal ions dispersed

Gianmario Martra: deceased. L. Mino (*) · G. Martra Department of Chemistry and Interdepartmental NIS Centre, University of Torino, Torino, Italy e-mail: [email protected]; [email protected] M. Matsuoka Department of Applied Chemistry, Graduate School of Engineering, Osaka Prefecture University, Osaka, Japan e-mail: [email protected]

15.1

Investigation of Charge Carrier Dynamics in Photocatalysts

Photoluminescence spectroscopy (PL) has emerged as a powerful tool to investigate the electronic structure, the photochemical properties, and the nature of catalytically active surface sites of photocatalysts under working conditions in order to better clarify their role in photocatalytic processes [1–3]. Recent developments of ultrafast time-resolved PL spectroscopy in the picosecond/nanosecond-time domain allow investigating the charge carriers photodynamics and discriminating between processes sensitive to surface terminations and bulk processes [4, 5]. Moreover, the study of the variations in the PL intensities and lifetimes in presence of specific molecules can provide deeper insights into the interaction mechanism of the different reactants during photocatalytic reactions [6]. In particular, this technique has been successfully applied to semiconductor oxide photocatalysts, like ZnO and TiO2, to better elucidate the catalytic properties of surface centers. For such systems, PL spectroscopy can also provide detailed information about oxygen vacancies, defective sites, and related efficiency of charge carrier trapping and transfer [7]. In the first part of this case study, the

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_15

323

324

main findings obtained by PL spectroscopy concerning pure and doped TiO2 nanomaterials are reported and commented on. We will mainly focus our attention on the TiO2 anatase crystal phase, which is the most widely employed in photocatalysis and it is the most stable at the nanoscale [8, 9]. We will also show the application of PL spectroscopy to organic semiconductors and in particular to g-C3N4.

15.1.1 Time-Resolved PL Studies of Pure and Doped TiO2 Nanoparticles It is well known that the PL signals of semiconductor oxides originate from the recombination of photo-excited charge carriers (see also Sect. ▶ 14.2 in ▶ Chap. 14) and can be classified into [2]: (i) band–band PL processes resulting from electron transitions from the conduction band (CB) bottom to the valence band (VB) top; (ii) excitonic PL processes in which the photo-excited electrons undergo a non-radiative decay from the CB bottom to different sub-bands (or surface states) followed by radiative transitions from the sub-band to the VB top. In general, a lower PL intensity is related to a slower electron–hole recombination, which results in a higher photocatalytic activity of the material. However, the inherent relationships between PL signals and photocatalytic performance are very complex, especially when dopants are present, therefore several factors can modify this general rule. Dozzi et al. [10] investigated F-doped TiO2 nanoparticles (NPs) by time-resolved PL spectroscopy highlighting that fluorine induces the formation of long-living (tens of nanoseconds or more) luminescent surface trapping sites. Their relative quantity and lifetime with respect to other shorter living PL components grows at higher calcination temperatures. This behavior is ascribed to an increase in the crystallinity of the material. The presence of these long-living surface traps results in a higher photocatalytic activity. Indeed, the trapping of the photogenerated electron and holes in these surface sites reduces the detrimental charge carrier recombination, increasing the probability of interaction with adsorbed target molecules. In a more recent study [11] they also clarified how noble metal NPs can interact with the photogenerated electrons in N,F co-doped TiO2. They observed an evident modification in the PL signal intensity (Fig. 15.1a) and decay (Fig. 15.1b) upon noble metal deposition. In particular, Au NPs seem to interact more efficiently than Pt NPs with the photogenerated electrons trapped at the defect sites presenting doped TiO2, which give rise to the long-lasting PL emission. Based on these evidences, they proposed a possible scheme (Fig. 15.1c) for the different electron transfer and recombination paths: (i) direct electron recombination with VB holes on the subnanosecond timescale; (ii) direct electron scavenging by noble metal NPs (path a in Fig. 15.1c); (iii) electron scavenging by noble

L. Mino et al.

metal NPs after being trapped at the defect sites introduced by doping (path b, c in Fig. 15.1c); (iv) recombination of the electrons at the trap sites with VB holes resulting in the typical long-living PL emission (path d in Fig. 15.1c). The decreased PL intensity observed in noble metal functionalized TiO2 NPs was then correlated to the measured higher rate constants for photocatalytic H2 evolution (Fig. 15.1d). Very recently Brüninghoff et al. [5] investigated the temporal evolution of the PL spectra of anatase TiO2 thin films in aqueous media, mimicking real photocatalytic conditions. In particular, they reported that the PL spectrum is time-dependent, with a pH-dependent broadening, showing a progressive increase of the intensity of the low energy component relative to the higher energy one with time. Moreover, they demonstrated that the dynamics and spectral changes in the picosecond/nanosecond time range can be ascribed to electrons moving from the nanoparticle bulk to the depletion layer/surface in about 1 ns (see Fig. 15.2). Therefore, it is inferred that the effect of surface states on the PL increases with time. This directionality in electron diffusion is crucial for photocatalytic processes and is likely to critically influence the photocatalytic performance of TiO2 nanoparticles. They concluded that tailoring charge recombination by surface engineering could be a promising way to improve the photocatalytic activity.

15.1.2 PL Studies of Reactant Interactions with TiO2 Nanoparticles The PL study of the interaction of selected reactant molecules with TiO2 anatase NPs in the gas-solid regime generally requires a specific pretreatment of the samples in vacuum at suitable temperatures (usually up to 873 K since at higher temperatures the anatase to rutile phase transition is occurring) to clean the NPs surface from adventitious contaminants and to tune the degree of surface hydration/hydroxylation [12–14]. This treatment is usually followed by an oxidation step to effectively remove the organic contaminants and to restore the stoichiometry of the oxide before adsorption of the target molecule [15]. This approach allows studying by PL the effect of the interaction of key reactants in photocatalytic processes (e.g., H2O, O2) [16–18] on the electronic properties and charge carrier dynamics of the semiconductor oxide. A similar methodology is applied also for the characterization of the TiO2 active surface sites with other in situ spectroscopies [19–21]. For instance, IR spectroscopy employing CO as probe molecule has been widely applied to investigate the Lewis acid properties of the Ti4+ sites and to correlate their spectroscopic features with the NPs morphology and their photocatalytic activity [22–24].

15

Case Studies: Photoluminescence (PL) Spectroscopy

a)

b)

6 D_5

Pt/D_5

100

Au/D_5

D_5

4

Normalized PL

PL intensity (a.u.)

325

2

10

–1

10

–2

Pt/D_5 Au/D_5

10–3 10–4

15 450

550 O(nm)

650

E°/V (NHE) pH 0 –0.16

200

300

400

Time (ns) a)

c)

100

0

750

d)

80 kAu/k

58.2 CB

60

e– e–

b)

c)

hQ–355 nm

PL

Au/Pt

kNM (k)

0 350

kPt/k

40 23.1

d)

21.0 7.1

4.5 3.0 VB

19.1

20

0

h+ TiO2

NM

D_0

D_5

D_12

Fig. 15.1 (a) Photoluminescence spectra, recorded upon excitation at 355 nm, of N,F co-doped TiO2 bare and functionalized with Pt or Au NPs. (b) PL decay measured for the same set of samples reported in part (a). (c) Schematic representation of the different electron transfer paths taking place in N,F co-doped TiO2 functionalized with noble metal NPs

(see the main text for the details). (d) Ratios between the zero-order rate constants for H2 evolution obtained with Au or Pt-functionalized TiO2 NPs (kAu or kPt) and those obtained with the corresponding bare TiO2 sample (k). (Reprinted with permission from Ref. [11]. Copyright 2018 American Chemical Society)

Anpo and coworkers performed systematic investigations of the effect of different reactants on the PL emission of TiO2 [25–27]. The dependence of the PL intensity on the nature of the atmosphere was discussed considering the TiO2 surface band bending, whose extent was correlated to the electronegativity or electroaffinity of the adsorbed molecules. In particular, they showed that H2O adsorption resulted in downward band bending, favoring the recombination of photogenerated charge carriers [26]. On the contrary, O2 adsorption caused an upward band bending, increasing the lifetime of electron-hole pairs. The upward band bending process and related decrease of the PL intensity induced by oxygen was confirmed also by Yates and coworkers [28] (see Fig. 15.3a–c). They also investigated the adsorption of donor molecules, like NH3 and CO, on TiO2 under irradiation observing a decrease of the PL signal (Fig. 15.3b). This phenomenon was explained

by a downward band bending with a consequent expansion of the depletion region, providing a lower concentration of photon-accessible defect recombination centers (Fig. 15.3d). They also highlighted that the reduction of the PL emission is completely reversible upon CO and NH3 desorption, while this is not the case for O2. This observation suggests that O2 is also irreversibly dissociated at the TiO2 surface, healing some oxygen vacancies naturally present in the sample. Recently, Ma et al. [29] studied the interaction between hydrogen species and TiO2 NPs owing to its implications for photocatalytic hydrogen production [30]. They employed PL spectroscopy to show that, during the photoreduction of hydrogen cations at the TiO2 surface, the intrinsic upward band bending of TiO2 is decreased owing to charge transfer. In particular, hydrogen cations impinging the oxide surface are reduced by the photogenerated electrons. The

326

L. Mino et al.

Time

+ –

Mobile electrons – –

– PL sensitive to surface/ environment

Intermediate behavior

Bulk PL

4.64 eV, 300 fs pulse

–

– +

STE

–

CB

– + +

+

+

+

+

VB Polarons

Potential depletion layer

Fig. 15.2 Scheme highlighting the different timescales of PL due to bulk or surface charge carrier recombination. The intensity of PL signals arising from surface terminations increases with time owing to the diffusion of electrons from the bulk to the potential surface depletion

layer followed by recombination with immobile hole polarons. Also self-trapped excitons (STE) can contribute to this process after phononassisted detrapping. (Reprinted from Ref. [5])

photogenerated holes left in the VB are driven toward the surface by the electric field in the band bending region (depletion layer). The accumulation of holes at the oxide surface tend to compensate the negative charges initially present in the material (TiO2 is a n-type semiconductor) decreasing the upward band bending. This effect reduces the volume of the depletion region, consequently increasing the electronhole recombination rate as monitored by the higher PL intensity. These results provide deeper insights into the mechanism of reduction of protons by photoexcited electrons to form hydrogen molecules, a key step in the water photocatalytic splitting reaction.

PL spectroscopy has been successfully applied to unravel the electron dynamics and the effectiveness of charge carrier separation in these heterojunction systems containing g-C3N4. For instance, Ong et al. [33] compared pure and Pt-loaded (up to 10%) g-C3N4 samples in the photoreduction under visible light of CO2 to CH4 in the presence of water vapor at ambient temperature and atmospheric pressure (Fig. 15.4). They highlighted that 2% Pt-loaded samples showed the best photocatalytic activity (Fig. 15.4b) and correlated these results with the intensity of the corresponding PL signals (Fig. 15.4a). In particular, they concluded that an interfacial transfer of photogenerated electrons occurs from g-C3N4 sheets to Pt nanoparticles owing to the lower Fermi level of Pt in the Pt/CN hybrid heterojunctions. This results in a slower recombination of the charge carriers, which leads to a decreased intensity of the PL signal. A similar approach was applied also to a carbon nanodot (CND)-protonated g-C3N4 (pCN) heterojunction photocatalyst [34] to correlate the catalytic activity in CO2 photoreduction process under visible light to the carrier dynamics. For pure pCN, the PL spectrum showed an intense and broad emission band, centered at about 450 nm, indicating rapid recombination of electron–hole pairs, through band–band transitions. Conversely, the CND/pCN hybrid photocatalysts showed significantly reduced PL intensities and a modification of the PL decay lifetime. These variations were ascribed to an effective interfacial charge separation in the CND/pCN nano-hybrids arising from well-formed 0D/2D heterojunction interfaces, which hinder the charge recombination.

15.1.3 PL Investigation of Charge Carrier Separation in g-C3N4 Heterojunctions In the last years, graphitic carbon nitride (g-C3N4) has attracted increasing attention as one of the most appealing next generation photocatalyst, owing to its facile synthesis, promising electronic band structure, high photochemical stability, and “earth-abundant” chemical composition [31]. However, pure g-C3N4 still suffers from its low efficiency in the separation of the photogenerated charge carriers. Therefore, increasing research efforts are devoted in the development of g-C3N4-based heterostructures to achieve enhanced charge carrier separation efficiency and, thus, photocatalytic performance [32].

15

Case Studies: Photoluminescence (PL) Spectroscopy

327

18

a)

PL intensity at 529.5 nm

14

(a)

(c) In vacuum Ta=200K Add NH3

12

c)

T=120 K

hQ on in vacuum

16

NH3 desorption at 120K

CB

CB

(b)

In vacuum at 120K hQ off NH3 adsorption

10 8

+ donor molecule

G+

at 200K hQ off

6

VB

Ta

4 120 K

2

120 K

VB

L0

Inactive region

Active PL region

0

50

100

150 200 250 Time (min)

300

350

15

In dark Add O2 2 1.5u10 Torr

O2 increases

PL intensity at 529.5 nm

(a)

10

(b)

PL intensity at 529.5 nm

hQ on in vacuum

20

d)

14 12 10 8 6 4

(e) In dark

CB

0.0 0.5 1.0 1.5 2.0 2.5 3.0 PO2 (Torr)

+ acceptor molecule

(d)

hQ on

(c) 3.00 Torr

In vacuum

0

VB

L0

Inactive region

G

Active PL region

VB

L2 L2 < L0

Depletion region

50 100 150 200 250 300 350 400 450 500 Time (min)

Fig. 15.3 Effect of NH3 (a) and O2 (b) adsorption and desorption on the PL intensity of TiO2 NPs. Schematic representation of the effect of donor (c) and acceptor (d) molecules on the thickness of the active PL region. In both cases, upon adsorption, the active PL region narrows

15.2

G

CB

In dark

5

0

15

400

25 T=120 K

L1 L1 < L0 Depletion layer

0

b)

G+

Investigation of Photocatalytic Reactions Promoted by Supported Transition Metal Ions

The dispersion of photoactive sites, typically transition metal ions (TMI), on/in a (porous) support is a key strategy to overcome rapid recombination of photoexcited charge carriers and limited number of active sites negatively affecting the performances of bulk photocatalysts. Highly dispersed TMI, to the limit of single sites, and their surrounding ligands can be considered as isolated semiconductor quantum dots. In the case of metal organic frameworks (MOF), they can be in interaction with light antenna systems, i.e., the organic linkers [35]. The huge number of reviews appeared in the last 5 years highlight the increasing interest in MOF as potential photocatalysts [36–41] but in general their performances must still be significantly improved to impact on actual applications [35]. Definitely well settled and

(L1, L2 < L0) resulting in a decrease of the PL intensity. (Reprinted with permission from Ref. [28]. Copyright 2011 American Chemical Society)

rationalized is the behavior of TMI photocatalytic sites anchored on oxides, and among them SiO2, both in the non-porous and porous forms, are still playing a prominent role [41–46]. In this domain, relevant insights on the reaction mechanism of the reduction of NO by CO on silica-supported molybdena (Mo6+/SiO2) photocatalysts were attained, by combining in situ investigations of the PL of supported TMI in the presence of reactants or products [47] and kinetics studies [48]. In addition to the silica-supported molybdena (Mo6+/ SiO2) photocatalysts, mesoporous silica-supported chromia (Cr6+-MCM-41) was reported to act as visible light responsive photocatalysts. Photoluminescence investigations revealed that Cr6+-MCM-41 react in its photo-excited state more efficiently with CO than H2, indicating that preferential photo-oxidation of CO (photo-PROX) by O2 can proceed on Cr6+-MCM-41 even in the presence of excess amount of H2. The detailed mechanism will also be shown in the following section [49].

328

a) (a)

(a) Pure CN (b) 0.5Pt/CN (c) 1Pt/CN (d) 2Pt/CN

Intensity (a.u.)

Fig. 15.4 (a) Photoluminescence spectra of the pure and Pt-loaded g-C3N4 samples. (b) Total yield of CH4 photoproduction over the different photocatalysts under visible light irradiation for 10 h. (Adapted from Ref. [33])

L. Mino et al.

(b)

Total yield of CH4 (Pmol gcatalyst1)

(c)

(f) 10Pt/CN

(e) (d)

440

390

b)

(f)

(e) 5Pt/CN

490

540 Wavelength (nm)

14.0

590

640

690

13.02

12.0

10.76 10.24

10.0

8.85

8.29 8.0 6.0 4.0

2.55

2.0 0.0 Pure CN

0.5Pt/CN

15.2.1 NO Photo-Reduction by CO on Mo6+/SiO2 Mo6+/SiO2 was found to act as an efficient UV-photocatalyst for converting NO to N2O and N2 in the presence of CO (converted to CO2) [48, 50], a reaction of actual interest for the abatement of air pollutants. A redox reaction mechanism was proposed [50] resulting from two steps: (i) a photostimulated reduction ~Mo6+ ¼ O2 þ CO þ hν ! Mo4+ þ CO2 (~Mo6+ ¼ O2denoting one of the two molybdenil bonds of a tetrahedrally coordinated Mo6+ ion on the SiO2 surface) followed by (ii) dark oxidation of ~Mo4+ by NO, producing N2O and N2, and restoring initial molybdenil bonds ~Mo6+ ¼ O2. The first step of the reaction has been studied by several researchers [51–55] and has been recently analyzed in detail combining in situ UV-Vis and IR spectroscopies [56]. In particular, the FT-IR investigation under simultaneous UV irradiation (Fig. 15.5a, b) clarified the structure of the Mo active sites and the mechanism of the photoreduction process. As schematized

1Pt/CN

2Pt/CN Samples

5Pt/CN

10Pt/CN

in Fig. 15.5c, the initially fully oxidized Mo6+/SiO2 system first gives rise to meridionally coordinated molybdenum tricarbonyls, Mo4+(CO)3 [52], which gradually convert into dicarbonyl species, Mo4+(CO)2. A prolonged photoreduction in CO leads, then, also to the appearance of highly reduced molybdenum hexacarbonyls, Mo0(CO)6. Finally, CO evacuation at room temperature mainly promotes the formation of monocarbonyl species, Mo4+(CO) [56]. A key intermediate species in this process is the short-lived excited state (~Mo5+ – O)*, resulting from the charge transfer excitation of a surface ~Mo6+ ¼ O2 by UV light absorption. In particular, in Mo6+/SiO2 photocatalysts with Mo content in the 0.23–2.5 wt% range considered in these studies, this ligand-to-metal charge transfer transition can be excited by irradiation at λ ¼ 265 nm and λ ¼ 330 nm [47]. In addition to the chemical quenching by CO, such excited species can be quenched by the interaction with the other reactant, NO, and, among the products, by N2O. Moreover, radiative

15

Case Studies: Photoluminescence (PL) Spectroscopy

329

a)

b) D

a) UV ON t=0 b) UV ON t=30' c) UV ON t=150'

Absorbance

2140 2107

H

D

M

+  T 0.1 2177

2200

T

2150

2100

2050

2000

1950

1900

0.1

T

2150

2050

2100

2000

1950

Wavenumber (cm1)

1900

1850

15

CO OC

0

Mo

CO CO

OC C

c) O

O Mo+6

Isolated O dioxo species

CO CO

OC hn

Mo+4

O O CO

O

O hn

T

hn CO

O CO

Mo+4

CO

O

CO

H

O

O

CO

Q(CO)=1990 cm1

D

O

hn Q(CO)=2177, 2140, 2107 cm1

Q(CO)=2127, 2080 cm1

CO Mo+4

O

O

M Q(CO)=2044 cm1

Fig. 15.5 (a) FT-IR spectra during in situ photoreduction up to 150 min of Mo6+/SiO2 in the presence of CO at 50 mbar. The spectrum of the oxidized material in the presence of CO gas, in dark conditions, has been subtracted from all spectra. (b) Difference spectrum (b–c of part a) to underline the spectral evolution from 30 min to 150 min of UV

(photoluminescence) and thermal radiationless processes can contribute to the decay of such excited species. The actual chemical character of the quenching by CO was proved by the strong inhibition occurring when decreasing the interaction temperature from 300 K down to 77 K [47]. The whole set of processes involving (~Mo5+ – O)* is as follows:  Mo

6þ

2

¼ O þ hν  ðαI UV Þ
 5þ  !  Mo  O 

ðexcitation Þ

ð15:1Þ


   5þ   Mo  O  þCO  kCQ ! 4þ

 Mo þ CO2

ðreaction, chemical quenchingÞ ð15:2Þ

irradiation. (c) Scheme of the photoreduction pathways of silicasupported Mo in CO at room temperature. T, D, M, and H indicate tricarbonyl, dicarbonyl, monocarbonyl, and hexacarbonyl species, respectively. The vibrational frequencies ascribed to the different Mo-CO surface complexes are also reported. (Adapted from ref. [56])


   5þ  6þ  Mo  O  þX  kPQx ! Mo 2



¼ O þ X ðphysical quenching Þ
   5þ  6þ  Mo  O   kph ! Mo ¼O

2

0

þ hν ðphotoluminescenceÞ
 5þ  6þ  Mo  O  ðkT Þ ! Mo 2

¼O

þ hϖ ðthermal radiationless decayÞ

ð15:3Þ

ð15:4Þ

ð15:5Þ

where IUV is the intensity of the UV light, α is a proportionality coefficient, kRQ is the chemical quenching rate constant by quencher molecule CO, kPQx is the physical quenching rate constant by quencher molecule X ¼ NO or N2O, and kph and kT are the rate constants for deactivation of (~Mo5+ – O) via emission of a photon (hν′) or phonon, (hϖ), respectively.

330

L. Mino et al.

The competition between the quenching by a given molecules and radiative and thermal radiationless decays can be evaluated by rationing the photoluminescence intensity in the absence and in the presence of a quencher molecule (I0, [Q] ¼ 0 and I, [Q] 6¼ 0, respectively). On the basis of the above set of processes, and under the assumption that (~Mo5+ – O) is at steady state, i.e., d[~Mo5+ – O]  / dt ¼ 0, the photoluminescence intensity in the absence (I0) and in the presence (I ) of a quencher molecule can be expressed as:
h i  6þ 2 I 0 ¼ αI UV  Mo ¼ O kph =kD

Noteworthy, a linear trend is obtained for N2O, also showing a very weak quenching effect, whereas data plots are nonlinear, not only for CO, as expected, but also for NO. Moreover, NO acts as a quencher definitely stronger than CO. The nonlinearity indicates that the principal mechanism of photoluminescence quenching in the presence of CO or NO involves the formation of complexes between excited (~Mo5+ – O) species and adsorbed quencher molecules. Thus, the interaction with by CO and NO must be expressed as:


 Mo

ð15:6Þ




where kD is kph þ kT and
h i  6þ 2 I ¼ αI UV  Mo ¼ O kph =ðkD þ kOX ½YÞ ð15:7Þ

5þ

O

 Mo

5þ

 Mo

  4þ  þCOads  kRQ ! Mo þ CO2 ð15:9Þ 

¼O



2

   þNOads  kPQ,NO ! þ NO



ð15:10Þ

and the I0/I equation becomes:   I 0 =I ¼ 1 þ kQz Θz =kD

ð15:8Þ

That, in the case of a pure collisional quenching, is the well-known Stern-Volmer equation, predicting a linear dependence of I0/I on [Y]. Nevertheless, I0/I measurements for Mo6+/SiO2 photocatalysts with Mo content of 0.23 or 2.5 wt% contacted with NO, CO, and N2O, separately, gave origin to the results depicted in Fig. 15.6a [47].

a)



O

6þ

where Y ¼ CO, NO, N2O Hence, the resulting photoluminescence intensity is   I 0 =I ¼ 1 þ kQX ½Y=kD



ð15:11Þ

where Z ¼ CO(kQz ¼ kRQ), NO(kQz ¼ kPQ, NO), and Θ is the fraction of quenchable surface sites occupied by adsorbed molecules. The Θz value can be found from the classical equation for the Langmuir-type adsorption isotherm Θz ¼ bz Pz =ð1 þ bz Pz Þ

ð15:12Þ

where Z ¼ NO, CO, and P is the pressure of a quenching gas. The combination of the last two equations results in

b) 250

140 NO CO N2O

200 Partial pressure (arb. units)

120

I0 (I)

100 80 60 40

NO 150

100 CO 50

20

CO2 N2O

0

0 0

5 Pressure (Torr)

10

Fig. 15.6 (a) Relative photoluminescence intensity (λex ¼ 330 nm; λem ¼ 430 nm) of Mo6+/SiO2 photocatalysts when in contact with NO, CO (Mo content in the photocatalyst: 0.23 wt%), and N2O (Mo content in the photocatalyst: 2.50 wt%). Symbols represent experimental data, while solid lines are the best fits resulting from the equation reported in

0

5

10

15

20

25

30

35

40

45

50

55

UV irradiation time (min)

the text. (Adapted from ref. [47]). (b) Kinetics of the CO þ 2NO ! CO2 þ N2O reaction over a 2.50 wt% Mo6+/SiO2 photocatalysts. Relevant amounts are: initial 13CO pressure: 2.6 Torr; initial NO pressure: 5.2 Torr; catalyst mass: 0.3 g; (NO+CO)gas/Mocat ¼ 4.9. (Adapted from ref. [48])

15

Case Studies: Photoluminescence (PL) Spectroscopy

  I 0 =I ¼ 1 þ kQz bz Pz =ðkD ð1 þ bz Pz Þ

331

ð15:13Þ

A best fit procedure carried out varying parameters kQz/kD and bz (Table 15.1) produced the solid curves in Fig. 15.6a, with a good agreement with experimental points. For the sake of completeness, it can be noticed that in the low pressure range, i.e., when bzPz  1, the I0/I reduces again to the SternVolmer equation   I 0 =I ¼ 1 þ kQz bz Pz =kD

ð15:14Þ

From data in Table 15.1, it is evident that kD is much smaller than both quenching rate constant kPQ for NO and kCQ for CO, and that NO is a stronger quencher molecule than CO for (~Mo5+ – O), being kPQ ≈ 2kCQ. These results are then used as a “feed” for the investigation of the kinetics of the actual photocatalytic reaction. Subbotina et al. [48], monitored the evolution of the gas-phase composition during UV irradiation (by a high-pressure mercury lamp) of a 2.5 wt% Mo6+/SiO2 photocatalyst in a circulating CO:NO ¼ 1:2 mixture, i.e., a stoichiometric mixture for the CO þ 2NO ! CO2 þ N2O reaction (Fig. 15.6b), which can be considered as the first part of a full photocatalytic process including the subsequent reduction of N2O to N2. Noteworthy, the detection of CO with respect N2 by mass-spectroscopic analyses (by a monopole-type MKh-7304 mass-spectrometer) was made possible by using isotopically labeled 13CO. It can be observed that kinetic curves are linear (constant rates regime), and no N2 is formed even under prolonged UV-irradiation. The occurrence of an actual photocatalytic process was proved by determining a turnover number (TON), i.e., the number of CO and NO molecules converted over 1 Mo atom of the photocatalyst sample, of 2.3, definitely larger than 1. Kinetic data can be rationalized by completing the previous set of interactions/reaction with the reactions yielding to the formation of N2O, which do not require UV irradiation  Mo

6þ

2

¼O


 5þ  þ hν  ðαI UV Þ !  Mo  O  ð15:15Þ

Table 15.1 Fitting parameters kQz/kD and bz for the I0/I equation Molecule (Z) NO CO a

kQz/kD 126  25a 67  10b

Average from fitting of three data sets Average from fitting of two data sets

b

1

bx (torr ) 6.8  0.9a 4.2  0.4b




 Mo

5þ









O

  4þ  þCO  kCQ ! Mo þ CO2 ð15:16Þ




 Mo ¼O

5þ

2

O

  6þ  þNO  kPQ ! Mo



þ NO ð15:17Þ
 5þ  6þ  Mo  O  ðkD Þ ! Mo
 2 0 ¼O þhν or þ hω where kD ¼ kph þ kT ð15:18Þ  Mo

4þ


 6þ 2 þ NO  ðk1 Þ !  Mo   NO

ð15:19Þ


 6þ 2 6þ  Mo   NO þ NO  ðk2 Þ ! Mo ¼O

2

þ N2 O

ð15:20Þ

Under the assumption that (~Mo5+ – O)  , ~Mo4+ and (~Mo6+  NO2) being intermediate species are at steadystate concentration under UV irradiation, the following set of differential equations can be written on the basis of the above set of interactions/reactions: d½CO=dt ¼ d½CO2 =dt ¼ d½N2 O=dt

h i 6þ 2 ¼ αI UV kCQ  Mo ¼ O   ½CO= kD þ kCQ ½CO þ kPQ ½NO h i 6þ 2 d½NO=dt ¼ 2αI UV kCQ  Mo ¼ O   ½CO= kD þ kCQ ½CO þ kPQ ½NO

ð15:21Þ

ð15:22Þ

Nevertheless, the quantitative analysis of the photoluminescence intensity reported above allowed to conclude that kD  kCQ,kPQ, thus the two differential equations can be simplified as follows: d½CO=dt ¼ d½CO2 =dt ¼ d½N2 O=dt h i 6þ 2 ¼ αI UV kCQ  Mo ¼ O   ½CO= kCQ ½CO þ kPQ ½NO

ð15:23Þ

h i 6þ 2  d½NO=dt ¼ 2αI UV kCQ  Mo ¼ O   ½CO= kCQ ½CO þ kPQ ½NO

ð15:24Þ

It is worth noting that: (a) the rate of NO consumption is as twice as high as those of CO consumption and CO2 and N2O formation; (b) neither k1 nor k2 enters in the differential equations. Subbotina et al. [48] provided the results of the integration of these differential equations

15

332

L. Mino et al.

ð1 þ 2K Þð1  xÞ  K ðγ  2Þ ln x h i 6þ 2 ¼ αI UV  Mo ¼ O t=½CO0

ð15:25Þ

ð0:5 þ K Þðγ  x1 Þ  K ðγ  2Þ ln f0:5ðx1  γ Þ þ 1g h i 6þ 2 ¼ αI UV  Mo ¼ O t=½CO0 ð15:26Þ ð1 þ 2K Þx2  K ðγ  2Þ ln ð1  x2 Þ h i 6þ 2 ¼ αI UV  Mo ¼ O t=½CO0

ð15:27Þ

where γ ¼ [NO]0/[CO]0, x ¼ [CO]/[CO]0, x1 ¼ [NO]/[CO]0, and x2 ¼ [N2O]/[CO]0 are dimensionless concentrations and K ¼ kPQ/kCQ. Index “0” refers to the initial concentrations. Another result of the quantitative analysis of photoluminescence quenching was kPQ kCQ, thus 2K 1. Hence, the last set of three equations can be further reduced in: h i 6þ 2 2ð1  xÞ  ðγ  2Þ ln x ¼ αI UV  Mo ¼ O t=K ½CO0 ð15:28Þ ðγ  x1 Þ  ðγ  2Þ ln f0:5ðx1  γ Þ þ 1g h i 6þ 2 ¼ αI UV  Mo ¼ O t=K ½CO0

ð15:29Þ

h i 6þ 2 2x2  ðγ  2Þ ln ð1  x2 Þ ¼ αI UV  Mo ¼ O t=K ½CO0 ð15:30Þ which allow to find the dependence of dimensionless concentrations of reactants and products on irradiation time t. Moreover, it must be considered that for a stoichiometric CO: NO ¼ 1:2, γ ¼ 2, thus dimensionless concentrations are linearly dependent on the irradiation time, in agreement with the linear trends of the evolution of the gas-phase composition observed experimentally (Fig. 15.6b). The authors [48] provided also some examples of the fitting of the last set of equations to experimental data derived from Fig. 15.6, using a computer best fit procedure. Results are show in Fig. 15.7 the agreement between the fitting and the experimental values is satisfactorily good, demonstrating the useful synergy between the quantitative analysis of the photoluminescence quenching and the modeling of the kinetics of the photocatalytic reaction (Fig. 15.7). A further proof of the usefulness and correctness of this synergy is represented by the results obtained by Subbotina et al. [48] for nonstoichiometric CO-NO mixtures (γ 6¼ 2),

namely, CO:NO ¼ 1:1 and CO:NO ¼ 2:1, resulting in nonlinear kinetics curves, with corresponding nonlinear evolution of the dimensionless concentrations vs. irradiation time well fitted by the above set of equations.

15.2.2 Photo-PROX Reaction on Visible Light Responsive Cr6+-MCM-41 As already mentioned, transitions metal ions supported on mesoporous silicas, like MCM-41 [57], are promising systems for the photocatalytic preferential oxidation of CO even in the presence of H2 (photo-PROX reaction). In this respect, Cr appears to be particularly interesting to exploit visible light for this reaction. The structure of the active sites in Cr6+MCM-41 has been elucidated by X-ray absorption spectroscopy [49]. Indeed, the XANES spectra of the sample show an intense pre-edge peak, confirming the tetrahedral nature of the Cr surface sites. Moreover, curve fitting of the Cr – O peaks in the EXAFS Fourier transform (Fig. 15.8a) highlights that the Cr6+ species exist in a highly distorted tetrahedral coordination with two shorter Cr ¼ O double bonds (bond length ¼ 1.59 Å) and two longer Cr - O single bonds (bond length ¼ 1.85 Å) [49]. The photocatalytic activity of the system has been assessed employing also PL spectroscopy since Cr6+-MCM41 exhibits a photoluminescence spectrum at around 550–800 nm at 298 K upon excitation of its typical absorption bands at around 240, 350, and 460 nm, as shown in Fig. 15.8b. The absorption and emission spectra are attributed to the following charge-transfer processes on the Cr¼O moieties of the tetrahedral monochromate species (CrO42) involving an electron transfer from O2 to Cr6+ ions and a reverse radiative decay from the charge-transfer excited triplet state [49].  Cr

6þ

2

¼ O þ hν ðabsorption Þ
 5þ  6þ !  Cr  O  ! Cr 2

¼O

0

þ hν

ðphotoluminescenceÞ

ð15:31Þ

The photoluminescence is quenched by the addition of CO, O2, and H2, showing that the Cr6+-oxide species, in its chargetransfer excited triplet state, interacts with CO, O2, and H2. The absolute quenching rate constants (kq(l/mols)) for each gas were determined by the SternVolmer plots as follows: H2 (8.63 105)  CO (5.91 109) < O2 (1.12 1010), showing that CO interacts more efficiently with the photoexcited Cr6+oxide species than H2. This result shows that preferential photooxidation of CO (photo-PROX) with O2 can proceed on Cr6+MCM-41 even in the presence of excess amount of H2. In this

15

Case Studies: Photoluminescence (PL) Spectroscopy

333

2.6 0.5 2.4 0.4 [N2O]/[CO]0

[NO]/[CO]0

2.2 2.0

0.3

1.8

0.2

1.6

0.1

15 0.0

1.4 0

10

20

30

40

50

0

10

20

30

40

50

0

10

20

30

40

50

0.7 1.0

0.6 0.5 [CO2]/[CO]0

[CO]/[CO]0

0.9

0.8

0.4 0.3 0.2

0.7 0.1 0.0

0.6 0

10

20 30 40 UV irradiation time (min)

50

UV irradiation time (min)

Fig. 15.7 Analysis of the experimental data for the 13CO : NO ¼ 1 : 2 mixture. Points are from experimental data; solid lines are calculated by the equations reported above. (Adapted from ref. [48])

a)

b)

(a)

Cr=O (c)

Cr-O Intensity (a.u.)

Magnitude (a.u.)

Fig. 15.8 (a) Fourier transform modulus of the EXAFS spectrum of Cr6+-MCM-41. (b) Photoluminescence spectra of Cr6+-MCM-41 measured: (a) in vacuum, (b) after visible light irradiation in the presence of CO for 0.5 h and its subsequent evacuation, and (c) after the addition of O2 on (b) at 298 K and its subsequent evacuation. (Measurement temperature: 298 K; Excitation: λex ¼ 500 nm). (Reprinted with permission from ref. [49]. Copyright 2007 American Chemical Society)

0

2

4 Distance (Å)

6

(b)

550

600

650 700 Wavelength (nm)

750

800

334

L. Mino et al.

Amounts of gasses (Pmol)

25 Dark

23

Light on

Tetrahedrally coordinated Cr6+ oxide species in Cr6+-MCM-41

H2

8 6

O2

4

CO2

O2–

O2–

CO

Cr6+

Reoxidation of Cr4+ reduced-species by O2

O

hv

O

1/2 O2

2

O2–

CO 0 0

30

60 90 120 Reaction time (min)

150

180

Cr5+

CO O

Cr4+ O

Charge transfer excited triplet state * O– O2– O

H2

O H2O

Fig. 15.9 Reaction time profiles, monitored by gas chromatography, of the photocatalytic oxidation of CO with O2 in the presence of H2 on Cr6+-MCM-41 under visible light irradiation (λ > 420 nm). (Initial amount of gasses for CO: 3.8 μmol, O2: 7.5 μmol, and for H2: 24.6 μmol). (Reprinted with permission from ref. [49]. Copyright 2007 American Chemical Society)

reaction, highly required is the high oxidation selectivity of CO, that is, only CO should be oxidized in CO2 by O2, while H2 oxidation into H2O by O2 should be minimized. In fact, as shown in Fig. 15.9 [49], the photo-PROX proceeds under visible light irradiation on Cr6+-MCM-41 and CO conversion and CO selectivity reached 100% and 97%, respectively, after visible light irradiation for 150 min. The reaction mechanism of photo-PROX on Cr6+-MCM41 is proposed as shown in Fig. 15.10 [49]. Initially, the Cr6+oxide species is photoexcited to its charge-transfer excited triplet state and preferentially reacts with CO (not with H2) to form CO2 and photoreduced Cr4+-oxide species, which can be detected by in situ FT-IR measurements [49, 58]. Then the Cr4+-oxide species are oxidized by O2 and the original Cr6+-oxide species are generated. In fact, as shown in Fig. 15.8b, photoluminescence intensity of Cr6+-oxide species decreases after the visible light irradiation in the presence of CO, suggesting the reduction of Cr6+-oxide species into Cr4+-oxide species. Moreover, the addition of O2 on this system at 298 K led to the complete recovery of the original intensity of the photoluminescence, indicating the regeneration of the original Cr6+-oxide species from Cr4+-oxide species. The high CO selectivity observed for the Cr6+-MCM-41 can be attributed to the high and selective reactivity of the photoexcited Cr6+-oxide species with CO, as indicated by the high quenching efficiency of CO as compared to H2. Thus, in situ photoluminescence study combined with FT-IR investigations revealed that the efficient redox cycles of Cr-oxide species under visible light irradiation play significant roles in the selective oxidation reaction of CO in the presence of H2.

2 CO CO2

Reduction of Cr-oxide species by CO

Fig. 15.10 Complete reaction cycle in the photocatalytic oxidation of CO with O2 in the presence of H2 on Cr6+-MCM-41. (Reprinted with permission from ref. [49]. Copyright 2007 American Chemical Society)

References 1. Anpo, M., Che, M.: Applications of Photoluminescence Techniques to the Characterization of Solid Surfaces in Relation to Adsorption, Catalysis, and Photocatalysis. In: Haag, W.O., Gates, B.C., Knozinger, H. (eds.) Advances in Catalysis, Vol 44 Advances in Catalysis, vol. 44, pp. 119–257. Elsevier Academic Press Inc, San Diego (1999) 2. Jing, L.Q., Qu, Y.C., Wang, B.Q., Li, S.D., Jiang, B.J., Yang, L.B., Fu, W., Fu, H.G., Sun, J.Z.: Review of photoluminescence performance of nano-sized semiconductor materials and its relationships with photocatalytic activity. Sol. Energy Mater. Sol. Cells. 90, 1773–1787 (2006) 3. Pallotti, D.K., Passoni, L., Maddalena, P., Di Fonzo, F., Lettieri, S.: Photoluminescence mechanisms in anatase and rutile TiO2. J. Phys. Chem. C. 121, 9011–9021 (2017) 4. Furube, A., Asahi, T., Masuhara, H., Yamashita, H., Anpo, M.: Direct observation of a picosecond charge separation process in photoexcited platinum-loaded Tio2 particles by femtosecond diffuse reflectance spectroscopy. Chem. Phys. Lett. 336, 424–430 (2001) 5. Bruninghoff, R., Wenderich, K., Korterik, J.P., Mei, B.T., Mul, G., Huijser, A.: Time-dependent photoluminescence of nanostructured anatase TiO2 and the role of bulk and surface processes. J. Phys. Chem. C. 123, 26653–26661 (2019) 6. Anpo, M., Dzwigaj, S., Che, M.: Applications of photoluminescence spectroscopy to the investigation of oxide-containing catalysts in the working state. In: Gates, B.C., Knozinger, H. (eds.) Advances in Catalysis, Vol 52 Advances in Catalysis, vol. 52, pp. 1–42. Elsevier Academic Press Inc, San Diego (2009) 7. Mascaretti, L., Russo, V., Zoppellaro, G., Lucotto, A., Casari, C.S., Kment, S., Naldoni, A., Li Bassi, A.: Excitation wavelength- and medium-dependent photoluminescence of reduced nanostructured TiO2 films. J. Phys. Chem. C. 123, 11292–11303 (2019) 8. Hummer, D.R., Kubicki, J.D., Kent, P.R.C., Post, J.E., Heaney, P.J.: Origin of nanoscale phase stability reversals in titanium oxide polymorphs. J. Phys. Chem. C. 113, 4240–4245 (2009)

15

Case Studies: Photoluminescence (PL) Spectroscopy

9. Mino, L., Cesano, F., Scarano, D., Spoto, G., Martra, G.: Molecules and heterostructures at TiO2 surface: the cases of H2O, CO2, and organic and inorganic sensitizers. Res. Chem. Intermed. 45, 5801–5829 (2019) 10. Dozzi, M.V., D’Andrea, C., Ohtani, B., Valentini, G., Selli, E.: Fluorine-doped TiO2 materials: photocatalytic activity vs timeresolved photoluminescence. J. Phys. Chem. C. 117, 25586–25595 (2013) 11. Dozzi, M.V., Candeo, A., Marra, G., D’Andrea, C., Valentini, G., Selli, E.: Effects of photodeposited gold vs platinum nanoparticles on N,F-doped TiO2 photoactivity: a time-resolved photoluminescence investigation. J. Phys. Chem. C. 122, 14326–14335 (2018) 12. Mino, L., Negri, C., Santalucia, R., Cerrato, G., Spoto, G., Martra, G.: Morphology, surface structure and water adsorption properties of TiO2 nanoparticles: a comparison of different commercial samples. Molecules. 25, 4605 (2020) 13. Mandavi-Shakib, A., Arce-Ramos, J.M., Austin, R.N., Schwartz, T.J., Grabow, L.C., Frederick, B.G.: Frequencies and thermal stability of isolated surface hydroxyls on pyrogenic TiO2 nanoparticles. J. Phys. Chem. C. 123, 24533–24548 (2019) 14. Mino, L., Morales-Garcia, A., Bromley, S.T., Illas, F.: Understanding the nature and location of hydroxyl groups on hydrated titania nanoparticles. Nanoscale. 13, 6577–6585 (2021) 15. Mino, L., Mandrile, L., Iannarelli, L., Portesi, C., Martra, G., Rossi, A.M.: Vibrational spectroscopy. In: Characterization of Nanoparticles: Measurement Processes for Nanoparticles, pp. 457–480. Elsevier (2019). https://doi.org/10.1016/B978-0-12-814182-3. 00024-9 16. Henderson, M.A.: A surface science perspective on TiO2 photocatalysis. Surf. Sci. Rep. 66, 185–297 (2011) 17. Mino, L., Zecchina, A., Martra, G., Rossi, A.M., Spoto, G.: A surface science approach to TiO2 P25 photocatalysis: an in situ Ftir study of phenol photodegradation at controlled water coverages from sub-monolayer to multilayer. Appl. Catal. B-Environ. 196, 135–141 (2016) 18. Mino, L., Negri, C., Zecchina, A., Spoto, G.: Photodegradation of organic pollutants on TiO2 P25 surfaces investigated by transmission Ftir spectroscopy under in situ UV-vis irradiation. Z. Phys. Chemie-Int. J. Res. Phys. Chem. Chem. Phys. 230, 1441–1451 (2016) 19. Gao, X.T., Bare, S.R., Fierro, J.L.G., Banares, M.A., Wachs, I.E.: Preparation and in-situ spectroscopic characterization of molecularly dispersed titanium oxide on silica. J. Phys. Chem. B. 102, 5653–5666 (1998) 20. Mihaylov, M., Chakarova, K., Hadjiivanov, K.: Formation of carbonyl and nitrosyl complexes on titania- and zirconia-supported nickel: Ftir spectroscopy study. J. Catal. 228, 273–281 (2004) 21. Pellegrino, F., Morra, E., Mino, L., Martra, G., Chiesa, M., Maurino, V.: Surface and bulk distribution of fluorides and Ti3+ species in TiO2 nanosheets: implications on charge carrier dynamics and photocatalysis. J. Phys. Chem. C. 124, 3141–3149 (2020) 22. Xu, M.C., Gao, Y.K., Moreno, E.M., Kunst, M., Muhler, M., Wang, Y.M., Idriss, H., Woll, C.: Photocatalytic activity of bulk TiO2 anatase and rutile single crystals using infrared absorption spectroscopy. Phys. Rev. Lett. 106, 4 (2011) 23. Mino, L., Spoto, G., Bordiga, S., Zecchina, A.: Rutile surface properties beyond the single crystal approach: new insights from the experimental investigation of different polycrystalline samples and periodic DFT calculations. J. Phys. Chem. C. 117, 11186–11196 (2013) 24. Mino, L., Pellegrino, F., Rades, S., Radnik, J., Hodoroaba, V.-D., Spoto, G., Maurino, V., Martra, G.: Beyond shape engineering of TiO2 nanoparticles: post-synthesis treatment dependence of surface hydration, hydroxylation, Lewis acidity and photocatalytic activity of TiO2 anatase nanoparticles with dominant {001} or {101} facets. ACS Appl. Nano Mater. 1, 5355–5365 (2018)

335 25. Anpo, M., Tomonari, M., Fox, M.A.: In situ photoluminescence of TiO2 as a probe of photocatalytic reactions. J. Phys. Chem. 93, 7300–7302 (1989) 26. Anpo, M., Chiba, K., Tomonari, M., Coluccia, S., Che, M., Fox, M.A.: Photocatalysis on native and platinum-loaded TiO2 and Zno catalysts-origin of different reactivities on wet and dry metal-oxides. Bull. Chem. Soc. Jpn. 64, 543–551 (1991) 27. Jung, K.Y., Park, S.B., Anpo, M.: Photoluminescence and photoactivity of titania particles prepared by the sol-gel technique: effect of calcination temperature. J. Photochem. Photobiol. A-Chem. 170, 247–252 (2005) 28. Stevanovic, A., Buttner, M., Zhang, Z., Yates, J.T.: Photoluminescence of TiO2: effect of UV light and adsorbed molecules on surface band structure. J. Am. Chem. Soc. 134, 324–332 (2012) 29. Ma, S.L., Zhang, Z., Harrison, I.: Photoreduction of hydrogen cations on TiO2 and its impact on surface band bending and the charge carrier recombination rate: a photoluminescence study under high vacuum conditions. J. Phys. Chem. C. 122, 8288–8294 (2018) 30. Pellegrino, F., Sordello, F., Mino, L., Minero, C., Hodoroaba, V.D., Martra, G., Maurino, V.: Formic acid photoreforming for hydrogen production on shape-controlled anatase TiO2 nanoparticles: assessment of the role of fluorides, {101}/{001} surfaces ratio, and platinization. ACS Catal. 9, 6692–6697 (2019) 31. Ong, W.J., Tan, L.L., Ng, Y.H., Yong, S.T., Chai, S.P.: Graphitic carbon nitride (g - C3N4)-based photocatalysts for artificial photosynthesis and environmental remediation: are we a step closer to achieving sustainability? Chem. Rev. 116, 7159–7329 (2016) 32. Fu, J.W., Yu, J.G., Jiang, C.J., Cheng, B.: G-C3n4-based heterostructured photocatalysts. Adv. Energy Mater. 8(31), 1701503 (2018) 33. Ong, W.J., Tan, L.L., Chai, S.P., Yong, S.T.: Heterojunction engineering of graphitic carbon nitride (g - C3N4) via Pt loading with improved daylight-induced photocatalytic reduction of carbon dioxide to methane. Dalton Trans. 44, 1249–1257 (2015) 34. Ong, W.J., Putri, L.K., Tan, Y.C., Tan, L.L., Li, N., Ng, Y.H., Wen, X.M., Chai, S.P.: Unravelling charge carrier dynamics in protonated g - C3N4 interfaced with carbon nanodots as co-catalysts toward enhanced photocatalytic CO2 reduction: a combined experimental and first-principles Dft study. Nano Res. 10, 1673–1696 (2017) 35. Dhakshinamoorthy, A., Li, Z.H., Garcia, H.: Catalysis and photocatalysis by metal organic frameworks. Chem. Soc. Rev. 47, 8134–8172 (2018) 36. Meyer, K., Ranocchiari, M., van Bokhoven, J.A.: Metal organic frameworks for photo-catalytic water splitting. Energy Environ. Sci. 8, 1923–1937 (2015) 37. Dhakshinamoorthy, A., Asiri, A.M., Garcia, H.: Metal-organic framework (Mof) compounds: photocatalysts for redox reactions and solar fuel production. Angew. Chem.-Int. Edit. 55, 5414–5445 (2016) 38. Dias, E.M., Petit, C.: Towards the use of metal-organic frameworks for water reuse: a review of the recent advances in the field of organic pollutants removal and degradation and the next steps in the field. J. Mater. Chem. A. 3, 22484–22506 (2015) 39. Zeng, L., Guo, X.Y., He, C., Duan, C.Y.: Metal-organic frameworks: versatile materials for heterogeneous photocatalysis. ACS Catal. 6, 7935–7947 (2016) 40. Zhang, H.B., Liu, G.G., Shi, L., Liu, H.M., Wang, T., Ye, J.H.: Engineering coordination polymers for photocatalysis. Nano Energy. 22, 149–168 (2016) 41. Liang, J., Liang, Z.B., Zou, R.Q., Zhao, Y.L.: Heterogeneous catalysis in zeolites, mesoporous silica, and metal-organic frameworks. Adv. Mater. 29(21), 1701139 (2017) 42. Yamashita, H., Mori, K., Kuwahara, Y., Kamegawa, T., Wen, M.C., Verma, P., Che, M.: Single-site and nano-confined photocatalysts designed in porous materials for environmental uses and solar fuels. Chem. Soc. Rev. 47, 8072–8096 (2018)

15

336 43. Yoo, B.K., Su, Z.X., Thomas, J.M., Zewail, A.H.: On the dynamical nature of the active center in a single-site photocatalyst visualized by 4D ultrafast electron microscopy. Proc. Natl. Acad. Sci. U. S. A. 113, 503–508 (2016) 44. Anpo, M., Thomas, J.M.: Single-site photocatalytic solids for the decomposition of undesirable molecules. Chem. Commun., (31), 3273–3278 (2006) 45. Mino, L., Barzan, C., Martino, G.A., Piovano, A., Spoto, G., Zecchina, A., Groppo, E.: Photoinduced ethylene polymerization on the Crvi/SiO2 Phillips catalyst. J. Phys. Chem. C. 123, 8145–8152 (2019) 46. Peek, N.M., Jeffcoat, D.B., Moisii, C., van de Burgt, L., Profeta, S., Scott, S.L., Stiegman, A.E.: Reassessment of the electronic structure of Cr(Vi) sites supported on amorphous silica and implications for Cr coordination number. J. Phys. Chem. C. 122, 4349–4358 (2018) 47. Shelimov, B., Dellarocca, V., Martra, G., Coluccia, S., Che, M.: Quantitative analysis of photoluminescence quenching of silicasupported molybdena catalysts. Relation to photocatalytic reduction of nitric oxide by carbon monoxide. Catal. Lett. 87, 73–79 (2003) 48. Subbotina, I.R., Shelimov, B.N., Che, M., Coluccia, S.: Kinetics of no photocatalytic reduction by CO over MoO3/SiO2 catalysts. In: Gamba, A., Colella, C., Coluccia, S. (eds.) Oxide-Based Systems at the Crossroads of Chemistry Studies in Surface Science and Catalysis, vol. 140, pp. 421–430. Elsevier Science BV, Amsterdam (2001) 49. Kamegawa, T., Morishima, J., Matsuoka, M., Thomas, J.M., Anpo, M.: Eliminating traces of carbon monoxide photocatalytically from hydrogen with a single-site, non-noble metal catalyst. J. Phys. Chem. C. 111, 1076–1078 (2007) 50. Subbotina, I.R., Shelimov, B.N., Kazansky, V.B., Lisachenko, A.A., Che, M., Coluccia, S.: Selective photocatalytic reduction of nitric oxide by carbon monoxide over silica-supported molybdenum oxide catalysts. J. Catal. 184, 390–395 (1999) 51. Gerasimov, S.F.: CO and NO adsorption on photoreduced Mo/SiO2 catalysts. React. Kinet. Catal. Lett. 32, 275–280 (1986) 52. Williams, C.C., Ekerdt, J.G.: Infrared spectroscopic characterization of molybdenum carbonyl species formed by ultraviolet photoreduction of silica-supported Mo(Vi) in carbon-monoxide. J. Phys. Chem. 97, 6843–6852 (1993) 53. Aigler, J.M., Houalla, M., Hercules, D.M.: Surface structure and metathesis activity of photoreduced allyl-based Mo/SiO2 catalysts. Top. Catal. 10, 123–126 (2000) 54. Kamegawa, T., Takeuchi, R., Matsuoka, M., Anpo, M.: Photocatalytic oxidation of Co with various oxidants by Mo oxide species highly dispersed on SiO2 at 293 K. Catal. Today. 111(248), 248–253 (2006) 55. Lee, E.L., Wachs, I.E.: In situ spectroscopic investigation of the molecular and electronic structures of SiO2 supported surface metal oxides. J. Phys. Chem. C. 111(14410), 14410–14425 (2007) 56. Santalucia, R., Spoto, G., Mino, L.: Probing molybdenum active sites during in situ photoreduction of the Mo6+/SiO2 catalyst. Molecules. 26, 1700 (2021) 57. Corma, A.: From microporous to mesoporous molecular sieve materials and their use in catalysis. Chem. Rev. 97, 2373–2419 (1997) 58. Toyao, T., Morishima, J., Saito, M., Horiuchi, Y., Kamegawa, T., Martra, G., Coluccia, S., Matsuoka, M., Anpo, M.: FT-IR study of the reaction mechanisms for photocatalytic reduction of no with co promoted by various single-site photocatalysts. J. Catal. 299, 232–239 (2013)

L. Mino et al.

Lorenzo Mino received his Ph.D. in Materials Science from the University of Torino in 2012. He is now Assistant Professor in Physical Chemistry at the Chemistry Department of the same University. His research activity is currently focused on the investigation of (photo) catalytic reactions on oxide nanoparticles coupling spectroscopic techniques and ab initio modeling.

Masaya Matsuoka is Full Professor of Physical Chemistry at the Osaka Prefecture University, Japan, where he obtained his Ph.D. in 1997, then he spent a postdoctoral period at the Université Pierre et Marie Curie, Paris. His current research interests include the development of the visible light-responsive photocatalysts, PCP/MOF photocatalysts, and their applications for environmental applications and H2 production from water.

Gianmario Martra is Full Professor in Physical Chemistry at the University of Torino, Italy, where he received his Ph.D. in 1994. He worked also at the Laboratoire de Réactivité de Surface of the Université Paris VI, France, and at the Osaka Prefecture University, Japan. He works on molecular events occurring at the surface of nanomaterials.

Near Ambient Pressure (NAP) X-Ray Photoelectron Spectroscopy (XPS) Zongyuan Liu

, Sanjaya Senanayake

16

, and Jose´ A. Rodriguez

Contents 16.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

16.2

Technical Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

16.3 16.3.1 16.3.2 16.3.3

Applications of NAP-XPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CO Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CO2 Hydrogenation and Methanol Synthesis . . . . . . . . . . . . Methane Activation and Conversion . . . . . . . . . . . . . . . . . . . . .

16.4

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

339 339 340 341

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

Abstract

Near atmospheric pressure (NAP)-XPS is a technique frequently used for the characterization of catalysts under working conditions and in the study of the surface chemistry associated with catalytic process at solid-gas interfaces. In the last 10 years, NAP-XPS has been used to study the adsorption/desorption of many common molecules (CO, CO2, NO, H2O, CH3OH, CH4, C2H4, C6H6, etc.) and a wide set of catalytic reactions (CO oxidation, NO reduction, forward and reverse water-gas shift reaction, CO and CO2 hydrogenation to methanol or methane, methane dry reforming, methane conversion to methanol, hydrogenation of olefins, desulfurization, etc.). It has provided a fundamental understanding of chemical phenomena associated with the manipulation of C-H, C-O and C-S bonds on model or powder catalysts. Keywords

XPS · In situ characterization · Surface reactions · Mechanistic studies · C1 catalysis · Methane · Carbon dioxide · Methanol synthesis · CO oxidation · Water-gas shift reaction Z. Liu · S. Senanayake · J. A. Rodriguez (*) Chemistry Division, Brookhaven National Laboratory, Upton, NY, USA e-mail: [email protected]; [email protected]; [email protected]

16.1

Introduction

The standard X-ray photoelectron spectroscopy (XPS) is a well-known surface-sensitive technique that is useful for detecting elements, changes in their oxidation state, and their concentration in a given sample [1, 2]. The surface sensitivity of XPS is a consequence of the limited mean free path of the photoelectrons ejected from the sample as a result of the photoelectric effect. The standard XPS instruments are operated under ultrahigh vacuum (UHV) conditions to prevent scattering of the photoelectrons by gases or liquids. In recent years, the development of near atmospheric pressure (NAP)XPS allowed the application of this powerful technique to in situ measurements under a background of gas (up to a few torr) and henceforth helped to bridge the pressure gap between UHV and industrial or technical pressure conditions (
atmospheric pressure) [3–5]. Both X-ray lab sources (Al and Mg K-α) and synchrotron-generated radiation have been utilized in the implementation of NAP-XPS (see Fig. 16.1). With lab source X-rays, NAP-XPS experiments can be performed routinely on a daily basis. In general, the technique is able to provide sufficient signal-to-noise ratio for obtaining elemental compositions and oxidation states of the most common transition metals used in catalytic processes. Synchrotron-based NAP-XPS systems require the radiation generated by large synchrotron facilities, but it has clear advantages in the studies of surface chemistry (C 1s, N 1s, O 1s regions with high signal) as well as depth profiling owing to the intense, focused, and tunable X-rays. In the field of heterogeneous catalysis where reactions mainly occur at the gas-solid interface, NAP-XPS has demonstrated its ability to identify catalytic active phases, variations in the concentration of surface species, and associated reaction mechanisms, giving in situ and surface-sensitive information which cannot be obtained by other “bulk” characterization techniques such as X-ray diffraction or X-ray absorption spectroscopy.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_16

337

338

Z. Liu et al.

Fig. 16.1 (a) Photograph of the IOS beamline at NSLS-II (BNL) equipped with chambers for AP-XPS. (b) Photograph of the lab-based AP-XPS at BNL-Chemistry with dual Mg and Al K-α sources, the inset image shows the aperture of the analyzer entrance cone. (c) A schematic of a typical AP-XPS setup, demonstrating the differential pumping system and analyzer aperture

a)

IOS –NSLS II

b)

c)

Lens 1 pumping Lens 2 pumping

ple

Main chamber 10–10–1 torr

Sam

Preparation chamber

e–

Al coated hv SiNx window

Analyzer 10–10–10 –7 torr

Analyzer pumping

16.2

Technical Issues

The general working principle of NAP-XPS relies on the separation of ambient pressure (up to few Torr) in the main chamber from the UHV of the analyzer through an aperture

(few hundred μm in diameter) and a differential pumped lens system (2–3 stages). A SiNx window is commonly used to allow the pass-through of X-rays while maintain the UHV of the beamline or X-ray source. Typical NAP-XPS systems based on synchrotron or lab source X-rays are shown in Fig. 16.1a and b, respectively. Figure 16.1c depicts the

Near Ambient Pressure (NAP) X-Ray Photoelectron Spectroscopy (XPS)

schematic of the setup illustrating the basic components of NAP-XPS. At the top, in the right side of Fig. 16.1b, is shown an inset with a typical analyzer entrance cone. In the image, a 300 μm diameter aperture at the inlet of the first lens and the differential pumping setup can be seen. In many instruments, a residual gas analyzer (or mass spectrometer) is installed within the differential pumping stages of the analyzer or operates via a capillary located close to the front of the sample. This allows the sampling of reaction products coming from the surface of the catalysts [5, 6]. Thus, using this approach it is possible to study the composition and reactivity of active phases in a catalytic material.

16.3

Applications of NAP-XPS

In recent years, the technique has been used to study the adsorption of many common molecules (CO, CO2, NO, H2O, CH3OH, CH4, C2H4, etc.) and a wide set of catalytic a)

b)

O 1s

hv= 650eV Pd 3p

Pd 3d5/2

hv= 435eV Bulk

c)

16.3.1 CO Oxidation The oxidation of CO is a prototypical reaction carried out on model catalysts [17]. NAP-XPS has been used to elucidate more accurate reaction mechanisms and surface electronic structures during ambient pressure exposure to CO and O2 [18, 19]. For example, NAP-XPS spectra displayed in Fig. 16.2 show the evolution of surface adsorbates and surface electronic structure for the CO þ O2 reaction on Pd(100) [20].

C1s

hv= 435eV 435 °C

O2 CO2

390 °C

Intensity (arb. units)

335 °C

Ads.CO 310 °C

CO

290 °C

Activation temperature (°C)

350 °C CO2

339

reactions (CO oxidation, NO reduction, forward and reverse water-gas shift reaction, CO and CO2 hydrogenation to methanol or methane, methane dry reforming, methane conversion to methanol, etc.) [7–12]. It has provided a fundamental understanding of chemical phenomena associated with C1 processes on model or powder catalysts [13–17]. Several examples which highlight the main results of these studies are discussed below.

Oxygen coverage (ML)

16

1 0.9 d) 0.8 0.7 0.6 0.5 0.4 p(2×2) 0.3 0.2 0.1

(—5ח5)

CO:O2 1:4

c(2×2)

e) 320 280 CO:O2 1:4

240 200

260 °C

250 °C

130 °C 540

535

530

337 335 295 Binding energy (eV)

290

285

Fig. 16.2 CO oxidation in a gas mixture of 0.25 Torr of CO and 0.25 Torr O2 showing the binding energy regions of (a) O 1s, (b) Pd 3d, and (c) C 1s. (d) Derived oxygen coverage and (e) activation temperature for a CO:O2 ratio of 1:1 and increasing total pressure. One measuring point

Activation temperature (°C)

275 °C 300

f)

250 200 150 100 10–6

10–5

10–4 10–3 10–2 Total pressure (Torr)

10–1

100

from a CO:O2 ratio of 1:4 is also included. (f) Calculated 1p-kMC activation temperature for 1:1 CO:O2 ratio and increasing total pressure. (Reproduced with permission from Ref. [20])

16

340

Z. Liu et al.

The O 1s, Pd 3d, and C 1s regions were monitored during a temperature ramp from 130 to 435  C under a reactant pressure of 500 mTorr of O2 þ CO (1:1 feed ratio). At relatively low temperatures ( 1.2 Å
1, it revealed an intermediate rate (k2,3,4 ≈ 0.35 h
1). The first diffraction peak is related to the change in contrast between the mesoporous substrate and absorbed material, but the second process also contributes to this change on a different timescale. Therefore, the first process is attributed to the dissociative adsorption of D2, and this is consistent with the rapid and relatively large increase in temperature over the first 30 minutes following introduction of the D2. In Fig. 24.12c, the slow component k1 does not reflect a chemical change and is likely to be related to the mass transport of the products within the pore. The change at Q ¼ 1.22 Å
1 reflects nearest-neighbor molecular interactions, while those at 3.1 and 4.3 Å
1 are associated with atomistic chemical changes within the system. Additionally, these features evolve with similar time constants and thus are correlated with the reduction of benzene. This information gives the kinetic information of the hydrogenation of benzene reaction and elucidates the process scheme for the reaction at conditions of room temperature with 250 mBar of D2 (Fig. 24.12e). This study provides a method of examining structure-reactivity correlations for these complex systems in detail, thus allowing the effects of mass transport within the

J. Moon et al.

catalyst and surface reaction to be decoupled. Additionally, the technique takes a step to allow both structure and spatial data to be tied effectively with the kinetics of the underlying processes.

24.7

Other Neutron Scattering Techniques for Heterogeneous Catalysis

In addition to INS, QENS, and ND, SANS and neutron imaging have also been used in catalysis. SANS is an elastic neutron scattering at a small angle to investigate the structure covering mainly the nanometer (10
9 m) to micrometer scale (10
6 m) [46]. Scattering measurements are performed in the Fourier space (also called reciprocal space) but not in real space like microscopy. Obtained scattering data needs to be converted to real space or fitted to various models to describe structures in reciprocal space. Normally, SANS is complementary to small-angle X-ray scattering (SAXS) which uses X-ray instead of the neutron. Due to the high penetration of thermal neutrons, the bulk structure of the catalyst may be studied, and sample environments are easily varied over a wide range of pressure and temperature. Additional advantages of SANS over SAXS include minimum beam damage and contrast variation achieved through isotopic labeling. The isotope labeling is particularly useful for the study of hydrogen-rich organic and biological materials. However, the low flux of neutron sources compared to the X-ray source means it requires longer data acquisition time. One application of SANS in catalysis is to study fuel cell electrode, focused on electrode layers of high-temperature PEM fuel cells. Holderer et al. utilized neutron scattering experiment where simultaneously data in the small-angle and diffraction regime are taken on an electrode layer with different Pt loading with and without phosphoric acid (Fig. 24.13) [140]. It probes the ensemble average of the sample and is in the sense complementary to the local real space information obtained with microscopy techniques. Both SANS and wideangle neutron scattering (WANS) have been applied in this study, and thus it can cover the Q range from 1 to 1000 Å
1 including diffraction part of the scattering curve with q > 1 Å
1. The characteristic length scale with the two instances of the Beaucage model with Rg,0 of about 10 Å from Pt particles and a large length scale Rg,1 ≈ 500 Å of the supporting structure indicating Porod scattering of a flat interface such as no surface fractal structure. This work showed that scattering experiments over a large range of length scales provide an insight into the individual components such as catalyst particles and carbon structure and investigate at the same time on larger scale to understand the fractal structure of the electrode material and its evolution upon filling with electrolyte. SANS analysis covering a large range in reciprocal space provides the potential to reveal structural properties on length scale from sub-nm to micrometers with unique contrast

Neutron Scattering (NS) Spectroscopy

509

9 8 7 6 5 4 3 2 1 0

a) D D

D

D

D

3D2

D

D D D D

D

D D D D D D

D Log10(F(Q)) (barns sr −1 atom−1)

1.5 1

1.25

2

Q (Å−1)

3

4

t (h)

24

24

5 9

1

8

0.75

7

6 5

0.5

4

t (h) 3

0.25

2

0

1

−0.25

0 1

3

Q (Å−1)

4

5

c) 300

0.6 Q = 0.19

+10

T (K)

0.5

Å−1

0.4 0.3

Q = 1.22

0.2 Q = 3.1 Å−1

0.1 0

Q = 4.3 Å−1

−0.1 −0.2 0

2

4

6

Bottom 295

Top

Å−1

d) P(D2) (mBar)

ΔF(Q) (barns sr −1 atom−1)

b)

2

250

200 0

8

1

2

t (h)

e)

3

4

5

6

7

8

9

t (h) D2(g) → 2D(ads), k0 = 2.146 h−1

C6D6(1) + 6D(ads)

C6D12(1), k2,3,4 ≈ 0.35 h−1

C6D12(1) pore diffusion, k1 = 0.136 h−1

Fig. 24.12 (a) F(Q)/time domain data over the course of the reaction with the benzene hydrogenation reaction scheme on the left panel. (b) Slices taken through (a) at specific Q values and corresponding exponential fits to the data. (c) Sample temperature as measured at the top and

bottom extremes of the TiZr cell (d) pressure of D2 gas present in 4 L supply reservoir. (e) Reaction scheme elucidated from the experiment. (Adapted from Ref. [139], Copyright 2013, with permission from The Royal Society of Chemistry)

variation properties useful for samples with light elements and heterogeneous multicomponent environment. Neutron image is an imaging process based on the neutron attenuation properties of the objects. It is a complementary tool to neutron scattering; it can directly probe the samples in real space on macroscopic length scales within the range from micrometers to decimeters. It is however rarely used to characterize catalysts due to the length scale it probes.

Presently, neutron imaging in catalysis is mainly used to understand the flow pattern in the reactor or porous media. For example, Borgschulte et al. [141] studied the CO2 methanation process in a sorption fixed bed reactor using time-resolved neutron imaging. Thanks to the high neutron attenuation coefficient of hydrogen, the adsorbed water in the sorption catalyst, such as LTA zeolite, gives a high contrast allowing researchers to follow the formation and distribution

510

J. Moon et al.

typical neutron imaging result of the relative neutron absorbance of the reactor during the reaction and regeneration. Dark regions normally occurred when the neutron absorbance increased, and thus it represents the higher water contents in the sorption reactor. The water forms from the Sabatier reaction (the hydrogenation of CO2), which is

map. LTA zeolite is known to absorb large quantities of water. The removal of water from the reaction centers is the key to enhance the conversion yield while suppressing side products such as CO. Normally, sorption reactor includes significantly large amount of catalyst to improve the sorption to make water remains in the bed. Figure 24.14 shows a Fig. 24.13 (a) SANS-WANS diffraction data for the empty 20% platinum coating electrode layer, (b) only the high-q region on a linear scale, (c) the same sample filled with phosphoric acid, (d) high-q region of the phosphoric acid-doped electrode. The hydrogen content is responsible for the higher incoherent background. (Adapted from Ref. [140], Copyright 2020, with permission from MDPI, Basel, Switzerland)

a)

b) 3e–06 2e–06 I(q) (cm−1)

I(q) (cm−1)

10−1 10−3 10−5

2e–06 2e–06 1e–06

10−7

5e–07 10−2

10−1

100

1

2 q (Å−1)

3

4

1

2 q (Å−1)

3

4

q (Å−1)

c)

d) 7e–06

10−1

I(q) (cm−1)

I(q) (cm−1)

6e–06 10−3 10−5

5e–06 4e–06 3e–06

10−7

2e–06 10−2

a)

10−1 q (Å−1)

100

b)

Gas inlet Camera

Reaction

Regeneration

ys

t

Side view

ca

To collimator

To IRGas cell

Front view

1.5

0.99 0.98

1.0

0.97 0.96

0.5

0.95 0.94

0.0

0.93 Reaction

Time

Regeneration

Fig. 24.14 (a) Top: simplified experimental setup used to measure the water content in sorption reactors at the NEUTRA beamline at SINQ, PSI Switzerland. (a) Bottom left: drawing of the reactor with dimension and gas fluxes and gas analysis. (b) Hydrogen flux is kept constant, and

CH4 , CO2 -IR absorbance (1)

tio n

Neutrons

Integrated camera signal (1 @time =0)

CO2/H2 (reaction); H2 (regeneration)

So rp

Light

Scintillator

tal

1.90

0

10

20

30

40 50 60 Time (min)

70

80

90

during the reaction and regeneration phase, the CO2 flux is switched on and off, respectively. (Adapted from Ref. [141], Copyright 2016, with permission from the Owner Societies)

24

Neutron Scattering (NS) Spectroscopy

adsorbed by the zeolite while the methane leaves the reactor. The neutron images show the formation of a reaction front running through the reactor. The gas product from the outline was also continuously monitored, and the water contents of the sorbent and catalytic activity can be correlated. There is no CO2 leaving from the reactor as long as the total water contents increase linearly. The neutron imaging may be applied to a similar process with sorption-enhanced reactions such as reversed water-gas shift reaction and purification of gases by adsorption under static and dynamic conditions.

24.8

Summary

The applications of neutron scattering to catalysis have been mainly focused on the systems which are difficult to study by other available techniques. It has a major recognized role in structural studies including light atoms and made a significant contribution to understanding the role of hydrogen or hydrogen-containing molecules over the catalysts. Throughout this chapter, theories and the applications for catalyst researches of main neutron scattering techniques, INS, QENS, and ND, along with other neutron techniques have been reviewed. INS has provided a different perspective of adsorbates on catalysts and sometimes complementary information on surface reactions over catalysts to other techniques such as IR and Raman. QENS has been used to understand transport motions of hydrogen-containing species on catalyst surfaces and in catalyst pores. This technique deserves more attention as it reveals the significance of transport processes in controlling catalytic environments. ND allows one to locate the light element in the lattice structure of the catalyst and gives better insights at a higher diffraction angle with stronger intensity. Overall, the unique power of neutron scattering in probing the vibration modes of adsorbed atoms and dynamics of molecular processes in catalytic systems has been demonstrated from the case studies with these neutron techniques. Neutron scattering provides observable characteristic results that can be straightforward calculated, and thus neutron scattering combined with modeling allows one to obtain better understating of the observed neutron scattering results. Due to the recent development of computer science and quantum chemistry, we believe that the coupling of modeling and data science with neutron scattering experiments could revolutionize the field of the catalyst research. With further developments in neutron sources, instrumentation, detectors, and more techniques and methodologies, neutron scattering is becoming available for the advanced characterization of catalysis materials. Although catalysis research is a well-established area, the fundamental properties and the nature of the catalytically active site of heterogeneous catalysts under realistic reaction conditions have remained largely unknown. Having access to

511

such information is of utmost importance to understand the rate-determining processes and steps of many heterogeneous reactions and identify important structure-activity/ selectivity relationships. In situ and operando methods have become available to identify the structural and morphological properties of the catalyst under reaction conditions, but only limited techniques of neutron scattering such as ND can be combined with in situ/operando methods. It is challenging to improve spatial and time resolution of neutron scattering; however, with further development of instrumentation and detection methods under working condition, we believe neutron scattering could revolutionize the field of in situ/operando characterization of catalyst and thus plays an important role in understanding the catalysis in the future.

References 1. Schimmel, H.G., Johnson, M.R., Kearley, G.J., Ramirez-Cuesta, A.J., Huot, J., Mulder, F.M.: The vibrational spectrum of magnesium hydride from inelastic neutron scattering and density functional theory. Mater. Sci. Eng. B. 108, 38–41 (2004) 2. Bridges, C.A., Fernandez-Alonso, F., Goff, J.P., Rosseinsky, M.J.: Observation of hydride mobility in the transition-metal oxide hydride Lasrcoo3h0.7. Adv. Mater. 18, 3304–3308 (2006) 3. Polo-Garzon, F., Luo, S., Cheng, Y., Page, K.L., Ramirez-Cuesta, A.J., Britt, P.F., Wu, Z.: Neutron scattering investigations of hydride species in heterogeneous catalysis. ChemSusChem. 12, 93–103 (2019) 4. Fu, Q.J., Ramirez-Cuesta, A.J., Tsang, S.C.: Molecular aluminum hydrides identified by inelastic neutron scattering during H2 regeneration of catalyst-doped Naalh4. J. Phys. Chem. B. 110, 711–715 (2006) 5. Tang, Y., Kobayashi, Y., Tassel, C., Yamamoto, T., Kageyama, H.: Hydride-enhanced Co2 methanation: water-stable Batio2.4h0.6 as a new support. Adv. Energy Mater. 8, 1800800 (2018) 6. Sato, T., Ramirez-Cuesta, A.J., Daemen, L.L., Cheng, Y., Orimo, S.-i.: Evidence of intermediate hydrogen states in the formation of a complex hydride. Inorg. Chem. 57, 867–872 (2018) 7. Noritake, T., Aoki, M., Towata, S., Ninomiya, A., Nakamori, Y., Orimo, S.: Crystal structure analysis of novel complex hydrides formed by the combination of Libh4 and Linh2. Appl. Phys. A. 83, 277–279 (2006) 8. Masakuni, O., Chun-Keung, L.: Neutron studies of nanostructured Cuo□Al2o3 Nox removal catalysts. Phys. B Condens. Matter. 241243, 269–275 (1997) 9. Huq, A., Chen, W.-R.: Neutron scattering for in-situ characterization of heterogeneous catalysis. In: In-Situ Characterization of Heterogeneous Catalysts, pp. 169–189 (2013). https://doi.org/10. 1002/9781118355923.ch6 10. Omojola, T., Silverwood, I.P., O’Malley, A.J.: Molecular behaviour of methanol and dimethyl ether in H-Zsm-5 catalysts as a function of Si/Al ratio: a Quasielastic neutron scattering study. Catal. Sci. Technol. 10, 4305 (2020). https://doi.org/10.1039/ D0CY00670J 11. Wu, Z., Cheng, Y., Tao, F., Daemen, L., Foo, G.S., Nguyen, L., Zhang, X., Beste, A., Ramirez-Cuesta, A.J.: Direct neutron spectroscopy observation of cerium hydride species on a cerium oxide catalyst. J. Am. Chem. Soc. 139, 9721–9727 (2017) 12. Davidson, A.L., Webb, P.B., Silverwood, I.P., Lennon, D.: The application of quasi-elastic neutron scattering to investigate hydrogen diffusion in an iron-based Fischer–Tropsch synthesis catalyst.

24

512 Top. Catal. 63, 378 (2020). https://doi.org/10.1007/s11244-02001259-2 13. Warringham, R., McFarlane, A.R., MacLaren, D.A., Webb, P.B., Tooze, R.P., Taylor, J., Ewings, R.A., Parker, S.F., Lennon, D.: The application of inelastic neutron scattering to explore the significance of a magnetic transition in an iron based Fischer-Tropsch catalyst that is active for the hydrogenation of Co. J. Chem. Phys. 143, 174703 (2015) 14. Terreni, J., Sambalova, O., Borgschulte, A., Rudić, S., Parker, S.F., Ramirez-Cuesta, A.J.: Volatile hydrogen intermediates of Co2 methanation by inelastic neutron scattering. Catalysts. 10 (2020) 15. Mitchell, P.C.H.: Tomkinson, J., Parker, S.F., Ramirez-Cuesta, A.J.: Vibrational Spectroscopy with Neutrons: With Applications in Chemistry, Biology, Materials Science and Catalysis, vol. 3. World Scientific, Singapore (2005) 16. Parker, S.F., Lennon, D., Albers, P.W.: Vibrational spectroscopy with neutrons: a review of new directions. Appl. Spectrosc. 65, 1325–1341 (2011) 17. Mitchell, P.C.H., Ramirez-Cuesta, A.J., Parker, S.F., Tomkinson, J., Thompsett, D.: Hydrogen spillover on carbon-supported metal catalysts studied by inelastic neutron scattering. Surface vibrational states and hydrogen riding modes. J. Phys. Chem. B. 107, 6838–6845 (2003) 18. Jobic, H., Renouprez, A., Bee, M., Poinsignon, C.: Quasi-elastic neutron scattering study of the molecular motions of methanol adsorbed on H-Zsm-5. J. Phys. Chem. 90, 1059–1065 (1986) 19. Jobic, H., Bee, M., Renouprez, A.: Quasi-elastic neutron scattering of benzene in Na-Mordenite. Surf. Sci. 140, 307–320 (1984) 20. Silverwood, I.P., Sakai, V.G.: Propane diffusion in Zsm-5 pores measured by quasielastic neutron scattering under macroscopic flow. Chem. Eng. Sci. 186, 116–121 (2018) 21. Eklöf-Österberg, C., Nedumkandathil, R., Häussermann, U., Jaworski, A., Pell, A.J., Tyagi, M., Jalarvo, N.H., Frick, B., Faraone, A., Karlsson, M.: Dynamics of hydride ions in metal hydride-reduced Batio3 samples investigated with quasielastic neutron scattering. J. Phys. Chem. C. 123, 2019–2030 (2019) 22. Turner, J.F.C., Benmore, C.J., Barker, C.M., Kaltsoyannis, N., Thomas, J.M., David, W.I.F., Catlow, C.R.A.: Probing the nature of acetylene bound to the active site of a Nina
zeolite Y catalyst by in situ neutron scattering. J. Phys. Chem. B. 104, 7570–7573 (2000) 23. Thomas, J.M.: High-resolution electron microscopy, neutron diffraction with isotopic substitution and X-ray absorption fine structure for the characterisation of active sites in oxide catalysts. In: Gamba, A., Colella, C., Coluccia, S. (eds.) Studies in Surface Science and Catalysis, 140, pp. 1–12. Elsevier, Amsterdam (2001) 24. Kammert, J., Moon, J., Cheng, Y., Daemen, L., Irle, S., Fung, V., Liu, J., Page, K., Ma, X., Phaneuf, V., Tong, J., Ramirez-Cuesta, A.J., Wu, Z.: Nature of reactive hydrogen for ammonia synthesis over a Ru/C12a7 electride catalyst. J. Am. Chem. Soc. 142, 7655–7667 (2020) 25. Stockmeyer, R., Conrad, H.M., Renouprez, A., Fouilloux, P.: Study of hydrogen chemisorbed on Raney-nickel by neutron inelastic spectroscopy. Surf. Sci. 49, 549–566 (1975) 26. Wright, P.A., Thomas, J.M., Cheetham, A.K., Nowak, A.K.: Localizing active sites in zeolitic catalysts: neutron powder profile analysis and computer simulation of deuteropyridine bound to gallozeolite-L. Nature. 318, 611–614 (1985) 27. Goyal, R., Fitch, A.N., Jobic, H.: Powder neutron and X-ray diffraction studies of benzene adsorbed in zeolite Zsm-5. J. Phys. Chem. B. 104, 2878–2884 (2000) 28. Falkowska, M., Chansai, S., Manyar, H.G., Gladden, L.F., Bowron, D.T., Youngs, T.G.A., Hardacre, C.: Determination of toluene hydrogenation kinetics with neutron diffraction. Phys. Chem. Chem. Phys. 18, 17237–17243 (2016)

J. Moon et al. 29. Neumann, D.A.: Neutron scattering and hydrogenous materials. Mater. Today. 9, 34–41 (2006) 30. Parker, S.F., Lennon, D.: Applications of neutron scattering to heterogeneous catalysis. J. Phys. Conf. Ser. 746, 012066 (2016) 31. Lennon, D., Parker, S.F.: Inelastic neutron scattering studies of methyl chloride synthesis over alumina. Acc. Chem. Res. 47, 1220–1227 (2014) 32. Ramirez-Cuesta, A.J., Jones, M.O., David, W.I.F.: Neutron scattering and hydrogen storage. Mater. Today. 12, 54–61 (2009) 33. O’Malley, A.J., Parker, S.F., Catlow, C.R.A.: Neutron spectroscopy as a tool in catalytic science. Chem. Commun. 53, 12164–12176 (2017) 34. Xue, Z.-L., Ramirez-Cuesta, A.J., Brown, C.M., Calder, S., Cao, H., Chakoumakos, B.C., Daemen, L.L., Huq, A., Kolesnikov, A.I., Mamontov, E., Podlesnyak, A.A., Wang, X.: Neutron instruments for research in coordination chemistry. Eur. J. Inorg. Chem. 2019, 1065–1089 (2019) 35. Albers, P.W., Parker, S.F.: Inelastic incoherent neutron scattering in catalysis research. In: Gates, B.C., Knözinger, H. (eds.) Advances in Catalysis, vol. 51, pp. 99–132. Academic Press, Amsterdam (2007) 36. Albers, P.W., Lennon, D., Parker, S.F.: Chapter 5: catalysis. In: Fernandez-Alonso, F., Price, D.L. (eds.) Experimental Methods in the Physical Sciences, vol. 49, pp. 279–348. Academic Press, Cambridge (2017) 37. Mittal, R., Chaplot, S.L., Choudhury, N.: Inelastic neutron scattering, lattice dynamics, computer simulation and thermodynamic properties. In: Thermodynamic Properties of Solids, pp. 75–121 (2010). https://doi.org/10.1002/9783527630417.ch3 38. Ramirez-Cuesta, A.J., Mitchell, P.C.H.: Neutrons and neutron spectroscopy. In: Local Structural Characterisation, pp. 173–224. Wiley, Chichester (2013). https://doi.org/10.1002/9781118681909.ch3 39. Squires, G.L.: Introduction to the Theory of Thermal Neutron Scattering. Courier Corporation (1996) 40. Andersen, K.H., Argyriou, D.N., Jackson, A.J., Houston, J., Henry, P.F., Deen, P.P., Toft-Petersen, R., Beran, P., Strobl, M., Arnold, T., Wacklin-Knecht, H., Tsapatsaris, N., Oksanen, E., Woracek, R., Schweika, W., Mannix, D., Hiess, A., Kennedy, S., Kirstein, O., Petersson Årsköld, S., Taylor, J., Hagen, M.E., Laszlo, G., Kanaki, K., Piscitelli, F., Khaplanov, A., Stefanescu, I., Kittelmann, T., Pfeiffer, D., Hall-Wilton, R., Lopez, C.I., Aprigliano, G., Whitelegg, L., Moreira, F.Y., Olsson, M., Bordallo, H.N., MartínRodríguez, D., Schneider, H., Sharp, M., Hartl, M., Nagy, G., Ansell, S., Pullen, S., Vickery, A., Fedrigo, A., Mezei, F., Arai, M., Heenan, R.K., Halcrow, W., Turner, D., Raspino, D., Orszulik, A., Cooper, J., Webb, N., Galsworthy, P., Nightingale, J., Langridge, S., Elmer, J., Frielinghaus, H., Hanslik, R., Gussen, A., Jaksch, S., Engels, R., Kozielewski, T., Butterweck, S., Feygenson, M., Harbott, P., Poqué, A., Schwaab, A., Lieutenant, K., Violini, N., Voigt, J., Brückel, T., Koenen, M., Kämmerling, H., Babcock, E., Salhi, Z., Wischnewski, A., Heynen, A., Désert, S., Jestin, J., Porcher, F., Fabrèges, X., Fabrèges, G., Annighöfer, B., Klimko, S., Dupont, T., Robillard, T., Goukassov, A., Longeville, S., Alba-Simionesco, C., Bourges, P., Le Bouffy, J.G., Lavie, P., Rodrigues, S., Calzada, E., Lerche, M., Schillinger, B., Schmakat, P., Schulz, M., Seifert, M., Lohstroh, W., Petry, W., Neuhaus, J., Loaiza, L., Tartaglione, A., Glavic, A., Schütz, S., Stahn, J., Lehmann, E., Morgano, M., Schefer, J., Filges, U., Klauser, C., Niedermayer, C., Fenske, J., Nowak, G., Rouijaa, M., Siemers, D.J., Kiehn, R., Müller, M., Carlsen, H., Udby, L., Lefmann, K., Birk, J.O., Holm-Dahlin, S., Bertelsen, M., Hansen, U.B., Olsen, M.A., Christensen, M., Iversen, K., Christensen, N.B., Rønnow, H.M., Freeman, P.G., Hauback, B.C., Kolevatov, R., Llamas-Jansa, I., Orecchini, A., Sacchetti, F., Petrillo, C., Paciaroni, A., Tozzi, P., Zanatta, M., Luna, P., Herranz, I., del Moral, O.G., Huerta, M., Magán, M., Mosconi, M., Abad, E., Aguilar, J., Stepanyan, S.,

24

Neutron Scattering (NS) Spectroscopy

Bakedano, G., Vivanco, R., Bustinduy, I., Sordo, F., Martínez, J.L., Lechner, R.E., Villacorta, F.J., Šaroun, J., Lukáš, P., Markó, M., Zanetti, M., Bellissima, S., del Rosso, L., Masi, F., Bovo, C., Chowdhury, M., De Bonis, A., Di Fresco, L., Scatigno, C., Parker, S.F., Fernandez-Alonso, F., Colognesi, D., Senesi, R., Andreani, C., Gorini, G., Scionti, G., Schreyer, A.: The instrument suite of the European spallation source. Nucl. Instrum. Methods Phys. Res., Sect. A. 957, 163402 (2020) 41. The Advanced Photon Source - a National Synchrotron Radiation Research Facility at Argonne National Laboratory, Technical Bulletins Series Tb-25. (1997) 42. Mason, T.E.: Pulsed neutron scattering for the 21st century. Phys. Today. 59, 44–49 (2006) 43. Sears, V.F.: Neutron scattering lengths and cross sections. Neutron News. 3, 26–37 (1992) 44. Neutron Scattering Lengths and Cross Sections. https://www.ncnr. nist.gov/resources/n-lengths/ 45. Colognesi, D., Celli, M., Cilloco, F., Newport, R.J., Parker, S.F., Rossi-Albertini, V., Sacchetti, F., Tomkinson, J., Zoppi, M.: Tosca neutron spectrometer: the final configuration. Appl. Phys. A. 74, s64–s66 (2002) 46. Feigin, L., Svergun, D.I.: Structure Analysis by Small-Angle X-Ray and Neutron Scattering, vol. 1. Plenum Press, New York (1987) 47. Zhou, X.-L., Chen, S.-H.: Theoretical foundation of X-ray and neutron reflectometry. Phys. Rep. 257, 223–348 (1995) 48. Kohn, W., Sham, L.J.: Self-consistent equations including exchange and correlation effects. Phys. Rev. 140, A1133–A1138 (1965) 49. Hedin, L., Lundqvist, B.I.: Explicit local exchange-correlation potentials. J. Phys. C Solid State Phys. 4, 2064 (1971) 50. Senftle, T.P., Hong, S., Islam, M.M., Kylasa, S.B., Zheng, Y., Shin, Y.K., Junkermeier, C., Engel-Herbert, R., Janik, M.J., Aktulga, H.M.: The Reaxff reactive force-field: development, applications and future directions. npj Comput. Mater. 2, 1–14 (2016) 51. McGreevy, R.L., Pusztai, L.: Reverse Monte Carlo simulation: a new technique for the determination of disordered structures. Mol. Simul. 1, 359–367 (1988) 52. Cheng, Y.Q., Daemen, L.L., Kolesnikov, A.I., Ramirez-Cuesta, A.J.: Simulation of inelastic neutron scattering spectra using Oclimax. J. Chem. Theory Comput. 15, 1974–1982 (2019) 53. Ramirez-Cuesta, A.J.: aCLIMAX 4.0.1, The new version of the software for analyzing and interpreting INS spectra. Comput. Phys. Commun. 157(3), 226–238 (2004) 54. Dymkowski, K., et al.: AbINS: The modern software for INS interpretation. Phys. B Condens. Matter. 551, 443–448 (2018) 55. Ramirez-Cuesta, A.J., Mitchell, P.C.H.: Hydrogen adsorption in a copper Zsm5 zeolite: an inelastic neutron scattering study. Catal. Today. 120, 368–373 (2007) 56. Mojet, B.L., Eckert, J., van Santen, R.A., Albinati, A., Lechner, R.E.: Evidence for chemisorbed molecular hydrogen in Fe-Zsm5 from inelastic neutron scattering. J. Am. Chem. Soc. 123, 8147–8148 (2001) 57. Juárez, R., Parker, S.F., Concepción, P., Corma, A., García, H.: Heterolytic and heterotopic dissociation of hydrogen on ceriasupported gold nanoparticles. Combined inelastic neutron scattering and Ft-Ir spectroscopic study on the nature and reactivity of surface hydrogen species. Chem. Sci. 1, 731–738 (2010) 58. Jobic, H., Clugnet, G., Lacroix, M., Yuan, S., Mirodatos, C., Breysse, M.: Identification of new hydrogen species adsorbed on ruthenium sulfide by neutron spectroscopy. J. Am. Chem. Soc. 115, 3654–3657 (1993) 59. Lacroix, M., Jobic, H., Dumonteil, C., Afanasiev, P., Breysse, M., Kasztelan, S.: Role of adsorbed hydrogen species on ruthenium and molybdenum sulfides. Characterization by inelastic neutron scattering, thermoanalysis methods and model reactions. In:

513 Hightower, J.W., Nicholas Delgass, W., Iglesia, E., Bell, A.T. (eds.) Studies in Surface Science and Catalysis, vol. 101, pp. 117–126. Elsevier, Amsterdam (1996) 60. Cavallari, C., Pontiroli, D., Jiménez-Ruiz, M., Ivanov, A., Mazzani, M., Gaboardi, M., Aramini, M., Brunelli, M., Riccò, M., Rols, S.: Hydrogen on graphene investigated by inelastic neutron scattering. J. Phys. Conf. Ser. 554, 012009 (2014) 61. Tsao, C.-S., Liu, Y., Chuang, H.-Y., Tseng, H.-H., Chen, T.-Y., Chen, C.-H., Yu, M.-S., Li, Q., Lueking, A., Chen, S.-H.: Hydrogen spillover effect of Pt-doped activated carbon studied by inelastic neutron scattering. J Phys. Chem. Lett. 2, 2322–2325 (2011) 62. Itoh, K., Miyahara, Y., Orimo, S.-i., Fujii, H., Kamiyama, T., Fukunaga, T.: The local structure of hydrogen storage nanocrystalline graphite by neutron scattering. J. Alloys Compd. 356–357, 608–611 (2003) 63. Colognesi, D., Ulivi, L., Zoppi, M., Ramirez-Cuesta, A.J., Orecchini, A., Karkamkar, A.J., Fichtner, M., Gil Bardají, E., Zhao-Karger, Z.: Hydrogen-storage materials dispersed into nanoporous substrates studied through incoherent inelastic neutron scattering. J. Alloys Compd. 538, 91–99 (2012) 64. Rowsell, J.L.C., Eckert, J., Yaghi, O.M.: Characterization of H2 binding sites in prototypical metal
organic frameworks by inelastic neutron scattering. J. Am. Chem. Soc. 127, 14904–14910 (2005) 65. Moon, J., Cheng, Y., Daemen, L.L., Li, M., Polo-Garzon, F., Ramirez-Cuesta, A.J., Wu, Z.: Discriminating the role of surface hydride and hydroxyl for acetylene semihydrogenation over ceria through in situ neutron and infrared spectroscopy. ACS Catal. 10, 5278–5287 (2020) 66. Hochard, F., Jobic, H., Massardier, J., Renouprez, A.J.: Gas phase hydrogenation of acetonitrile on Raney nickel catalysts: reactive hydrogen. J. Mol. Catal. A Chem. 95, 165–172 (1995) 67. Hawkins, A.P., Zachariou, A., Parker, S.F., Collier, P., Barrow, N., Silverwood, I.P., Howe, R.F., Lennon, D.: Effect of steam de-alumination on the interactions of propene with H-Zsm-5 zeolites. RSC Adv. 10, 23136–23147 (2020) 68. Warringham, R., Davidson, A.L., Webb, P.B., Tooze, R.P., Ewings, R.A., Parker, S.F., Lennon, D.: Examining the temporal behavior of the hydrocarbonaceous overlayer on an iron based Fischer–Tropsch catalyst. RSC Adv. 9, 2608–2617 (2019) 69. Hawkins, A.P., Zachariou, A., Collier, P., Ewings, R.A., Howe, R.F., Parker, S.F., Lennon, D.: Low-temperature studies of propene oligomerization in Zsm-5 by inelastic neutron scattering spectroscopy. RSC Adv. 9, 18785–18790 (2019) 70. Hall, P.G., Pidduck, A., Wright, C.J.: Inelastic neutron scattering by water adsorbed on silica. J. Colloid Interface Sci. 79, 339–349 (1981) 71. Beta, I.A., Böhlig, H., Hunger, B.: Structure of adsorption complexes of water in zeolites of different types studied by infrared spectroscopy and inelastic neutron scattering. Phys. Chem. Chem. Phys. 6, 1975–1981 (2004) 72. Jobic, H., Tuel, A., Krossner, M., Sauer, J.: Water in interaction with acid sites in H-Zsm-5 zeolite does not form hydroxonium ions. A comparison between neutron scattering results and ab Initio calculations. J. Phys. Chem. 100, 19545–19550 (1996) 73. Copéret, C., Estes, D.P., Larmier, K., Searles, K.: Isolated surface hydrides: formation, structure, and reactivity. Chem. Rev. 116, 8463–8505 (2016) 74. Norton, J.R., Sowa, J.R.: Introduction: metal hydrides. Chem. Rev. 116, 8315–8317 (2016) 75. Nicol, J.M., Rush, J.J., Kelley, R.D.: Inelastic neutron scattering studies of the interaction of hydrogen with palladium black. Surf. Sci. 197, 67–80 (1988) 76. Albers, P.W., Krauter, J.G.E., Ross, D.K., Heidenreich, R.G., Köhler, K., Parker, S.F.: Identification of surface states on finely divided supported palladium catalysts by means of inelastic incoherent neutron scattering. Langmuir. 20, 8254–8260 (2004)

24

514 77. Albers, P., Burmeister, R., Seibold, K., Prescher, G., Parker, S.F., Ross, D.K.: Investigations of palladium catalysts on different carbon supports. J. Catal. 181, 145–154 (1999) 78. Parker, S.F., Walker, H.C., Callear, S.K., Grünewald, E., Petzold, T., Wolf, D., Möbus, K., Adam, J., Wieland, S.D., Jiménez-Ruiz, M., Albers, P.W.: The effect of particle size, morphology and support on the formation of palladium hydride in commercial catalysts. Chem. Sci. 10, 480–489 (2019) 79. Parker, S.F., Adroja, D., Jiménez-Ruiz, M., Tischer, M., Möbus, K., Wieland, S.D., Albers, P.: Characterisation of the surface of freshly prepared precious metal catalysts. Phys. Chem. Chem. Phys. 18, 17196–17201 (2016) 80. Albers, P., Auer, E., Ruth, K., Parker, S.F.: Inelastic neutron scattering investigation of the nature of surface sites occupied by hydrogen on highly dispersed platinum on commercial carbon black supports. J. Catal. 196, 174–179 (2000) 81. Albers, P.W., Lopez, M., Sextl, G., Jeske, G., Parker, S.F.: Inelastic neutron scattering investigation on the site occupation of atomic hydrogen on platinum particles of different size. J. Catal. 223, 44–53 (2004) 82. Rush, J.J., Cavanagh, R.R., Kelley, R.D.: Summary abstract: neutron scattering from adsorbates on platinum black. J. Vac. Sci. Technol. A. 1, 1245–1246 (1983) 83. Wright, C.J.: Alternative explanation of the inelastic neutron scattering from hydrogen adsorbed by Raney nickel. J. Chem. Soc., Faraday Trans. 2. 73, 1497–1500 (1977) 84. Jobic, H., Candy, J.-P., Perrichon, V., Renouprez, A.: Neutronscattering and volumetric study of hydrogen adsorbed and absorbed on Raney palladium. J. Chem. Soc., Faraday Trans. 1. 81, 1955–1961 (1985) 85. Barrera, G.D., Colognesi, D., Mitchell, P.C.H., Ramirez-Cuesta, A.J.: Lda or Gga? A combined experimental inelastic neutron scattering and ab initio lattice dynamics study of alkali metal hydrides. Chem. Phys. 317, 119–129 (2005) 86. Sivadinarayana, C., Choudhary, T.V., Daemen, L.L., Eckert, J., Goodman, D.W.: The nature of the surface species formed on Au/Tio2 during the reaction of H2 and O2: an inelastic neutron scattering study. J. Am. Chem. Soc. 126, 38–39 (2004) 87. Breck, D.W.: Zeolite molecular sieves: structure, chemistry, and use D. W. Breck (Union Carbide Corporation, Tarrytown, New York) John Wiley and Sons, New York, London, Sydney, and Toronto. 1974. 771. J. Chromatogr. Sci. 13, 18A (1975) 88. Weitkamp, J., Fritz, M., Ernst, S.: Zeolites as media for hydrogen storage. Int. J. Hydrog. Energy. 20, 967–970 (1995) 89. Yang, R.T.: Adsorbents: Fundamentals and Applications. Wiley, New Jersey (2003) 90. Ramirez-Cuesta, A.J., Mitchell, P.C.H., Ross, D.K., Georgiev, P.A., Anderson, P.A., Langmi, H.W., Book, D.: Dihydrogen in cation-substituted zeolites X—an inelastic neutron scattering study. J. Mater. Chem. 17, 2533–2539 (2007) 91. Rosi, N.L., Eckert, J., Eddaoudi, M., Vodak, D.T., Kim, J., O’Keeffe, M., Yaghi, O.M.J.S.: Hydrogen storage in microporous metal-organic frameworks. Science. 300, 1127–1129 (2003) 92. Rowsell, J.L.C., Millward, A.R., Park, K.S., Yaghi, O.M.: Hydrogen sorption in functionalized metal
organic frameworks. J. Am. Chem. Soc. 126, 5666–5667 (2004) 93. Kroto, H.W., Fischer, J.E., Cox, D.: The Fullerenes. Newnes, Oxford (2012) 94. Schimmel, H.G., Kearley, G.J., Nijkamp, M.G., Visser, C.T., de Jong, K.P., Mulder, F.M.: Hydrogen adsorption in carbon nanostructures: comparison of nanotubes, fibers, and coals. Chem. Eur. J. 9, 4764–4770 (2003) 95. Georgiev, P.A., Ross, D.K., Monte, A.D., Montaretto-Marullo, U., Edwards, R.A.H., Ramirez-Cuesta, A.J., Colognesi, D.: Hydrogen site occupancies in single-walled carbon nanotubes studied by

J. Moon et al. inelastic neutron scattering. J. Phys. Condens. Matter. 16, L73–L78 (2004) 96. Horsewill, A.J., Panesar, K.S., Rols, S., Johnson, M.R., Murata, Y., Komatsu, K., Mamone, S., Danquigny, A., Cuda, F., Maltsev, S., Grossel, M.C., Carravetta, M., Levitt, M.H.: Quantum translatorrotator: inelastic neutron scattering of dihydrogen molecules trapped inside anisotropic fullerene cages. Phys. Rev. Lett. 102, 013001 (2009) 97. Horsewill, A.J., Panesar, K.S., Rols, S., Ollivier, J., Johnson, M.R., Carravetta, M., Mamone, S., Levitt, M.H., Murata, Y., Komatsu, K., Chen, J.Y.C., Johnson, J.A., Lei, X., Turro, N.J.: Inelastic neutron scattering investigations of the quantum molecular dynamics of a H${}_{2}$ molecule entrapped inside a fullerene cage. Phys. Rev. B. 85, 205440 (2012) 98. Mao, W.L., Mao, H.-k., Goncharov, A.F., Struzhkin, V.V., Guo, Q., Hu, J., Shu, J., Hemley, R.J., Somayazulu, M., Zhao, Y.: Hydrogen clusters in clathrate hydrate. Science. 297, 2247 (2002) 99. Lee, H., Lee, J.-w., Kim, D.Y., Park, J., Seo, Y.-T., Zeng, H., Moudrakovski, I.L., Ratcliffe, C.I., Ripmeester, J.A.: Tuning clathrate hydrates for hydrogen storage. Nature. 434, 743–746 (2005) 100. Bée, M.: Localized and Long-range diffusion in condensed matter: state of the art of Qens studies and future prospects. Chem. Phys. 292, 121–141 (2003) 101. Mamontov, E., Herwig, K.W.: A time-of-flight backscattering spectrometer at the spallation neutron source, BASIS. Rev. Sci. Instrum. 82, 085109 (2011) 102. Shibata, K., Takahashi, N., Kawakita, Y., Matsuura, M., Yamada, T., Tominaga, T., Kambara, W., Kobayashi, M., Inamura, Y., Nakatani, T., Nakajima, K., Arai, M.: The performance of Tof near backscattering spectrometer DNA in Mlf, J-Parc. In: Proceedings of the 2nd International Symposium on Science at J-Parc? Unlocking the Mysteries of Life, Matter and the Universe ? Jps Conference Proceedings vol 8, vol. 8. Journal of the Physical Society of Japan (2015) 103. Bernhard, F., Eugene, M., van Lambert, E., Tilo, S.: Recent backscattering instrument developments at the Ill and Sns. Z. Phys. Chem. 224, 33–60 (2010) 104. Tsapatsaris, N., Lechner, R.E., Markó, M., Bordallo, H.N.: Conceptual design of the time-of-flight backscattering spectrometer, miracles, at the European spallation source. Rev. Sci. Instrum. 87, 085118 (2016) 105. Azuah, R.T., Kneller, L.R., Qiu, Y., Tregenna-Piggott, P.L., Brown, C.M., Copley, J.R., Dimeo, R.M.: Dave: a comprehensive software suite for the reduction, visualization, and analysis of low energy neutron spectroscopic data. J. Res. Natl. Inst. Stand. Technol. 114, 341–358 (2009) 106. Mamontov, E., Smith, R.W., Billings, J.J., Ramirez-Cuesta, A.J.: Simple analytical model for fitting Qens data from liquids. Phys. B Condens. Matter. 566, 50–54 (2019) 107. Gautam, S., Le, T.T.B., Rother, G., Jalarvo, N., Liu, T., Mamontov, E., Dai, S., Qiao, Z.-A., Striolo, A., Cole, D.: Effects of water on the stochastic motions of propane confined in Mcm-41-S pores. Phys. Chem. Chem. Phys. 21, 25035–25046 (2019) 108. Ruthven, D.M., Eic, M., Richard, E.: Diffusion of C8 aromatic hydrocarbons in silicalite. Zeolites. 11, 647–653 (1991) 109. Teixeira, A.R., Chang, C.-C., Coogan, T., Kendall, R., Fan, W., Dauenhauer, P.J.: Dominance of surface barriers in molecular transport through silicalite-1. J. Phys. Chem. C. 117, 25545–25555 (2013) 110. Song, L., Sun, Z.-L., Rees, L.V.C.: Experimental and molecular simulation studies of adsorption and diffusion of cyclic hydrocarbons in silicalite-1. Microporous Mesoporous Mater. 55, 31–49 (2002) 111. Fried, J.R., Weaver, S.: Atomistic simulation of hydrocarbon diffusion in silicalite. Comput. Mater. Sci. 11, 277–293 (1998)

24

Neutron Scattering (NS) Spectroscopy

112. Koriabkina, A.O., de Jong, A.M., Schuring, D., van Grondelle, J., van Santen, R.A.: Influence of the acid sites on the intracrystalline diffusion of hexanes and their mixtures within Mfi-zeolites. J. Phys. Chem. B. 106, 9559–9566 (2002) 113. Krishna, R., van Baten, J.M.: Diffusion of hydrocarbon mixtures in Mfi zeolite: influence of intersection blocking. Chem. Eng. J. 140, 614–620 (2008) 114. Leroy, F., Rousseau, B., Fuchs, A.H.: Self-diffusion of N-alkanes in silicalite using molecular dynamics simulation: a comparison between rigid and flexible frameworks. Phys. Chem. Chem. Phys. 6, 775–783 (2004) 115. Jiang, M., Eic, M., Miachon, S., Dalmon, J.-A., Kocirik, M.: Diffusion of N-butane, isobutane and ethane in a Mfi-zeolite membrane investigated by gas permeation and Zlc measurements. Sep. Purif. Technol. 25, 287–295 (2001) 116. Matam, S.K., O’Malley, A.J., Catlow, C.R.A., Suwardiyanto, Collier, P., Hawkins, A.P., Zachariou, A., Lennon, D., Silverwood, I., Parker, S.F., Howe, R.F.: The effects of Mtg catalysis on methanol mobility in Zsm-5. Cat. Sci. Technol. 8, 3304–3312 (2018) 117. Tang, Y., Kobayashi, Y., Masuda, N., Uchida, Y., Okamoto, H., Kageyama, T., Hosokawa, S., Loyer, F., Mitsuhara, K., Yamanaka, K., Tamenori, Y., Tassel, C., Yamamoto, T., Tanaka, T., Kageyama, H.: Metal-dependent support effects of oxyhydride-supported Ru, Fe, Co catalysts for ammonia synthesis. Adv. Energy Mater. 8, 1801772 (2018) 118. Shull, C.G., Wollan, E.O.: X-ray, electron, and neutron diffraction. Science. 108, 69 (1948) 119. Shoemaker, D.P., Li, J., Seshadri, R.: Unraveling atomic positions in an oxide spinel with two Jahn
Teller ions: local structure investigation of Cumn2o4. J. Am. Chem. Soc. 131, 11450–11457 (2009) 120. Page, K., Stoltzfus, M.W., Kim, Y.-I., Proffen, T., Woodward, P.M., Cheetham, A.K., Seshadri, R.: Local atomic ordering in Batao2n studied by neutron pair distribution function analysis and density functional theory. Chem. Mater. 19, 4037–4042 (2007) 121. Page, K., Hood, T.C., Proffen, T., Neder, R.B.: Building and refining complete nanoparticle structures with total scattering data. J. Appl. Crystallogr. 44, 327–336 (2011) 122. Liu, J., Olds, D., Peng, R., Yu, L., Foo, G.S., Qian, S., Keum, J., Guiton, B.S., Wu, Z., Page, K.: Quantitative analysis of the morphology of {101} and {001} faceted anatase Tio2 nanocrystals and its implication on photocatalytic activity. Chem. Mater. 29, 5591–5604 (2017) 123. Mawdsley, J.R., Vaughey, J.T., Krause, T.R.: Neutron diffraction studies of nickel-containing perovskite oxide catalysts exposed to autothermal reforming environments. Chem. Mater. 21, 4830–4838 (2009) 124. Rauch, H., Waschkowski, W., Neutron scattering lengths. In: Dianoux, A.-J., Lander, G. (eds.) Neutron Data Booklet, 2nd ed, pp. 11–15. Institut Laue-Langevin, Grenoble (2003) 125. Kandemir, T., Girgsdies, F., Hansen, T.C., Liss, K.-D., Kasatkin, I., Kunkes, E.L., Wowsnick, G., Jacobsen, N., Schlögl, R., Behrens, M.: In situ study of catalytic processes: neutron diffraction of a methanol synthesis catalyst at industrially relevant pressure. Angew. Chem. Int. Ed. 52, 5166–5170 (2013) 126. Otomo, T., Ikeda, K., Honda, T.: Structural studies of hydrogen storage materials with neutron diffraction: a review. J. Phys. Soc. Jpn. 89, 051001 (2020) 127. Orimo, S.-i., Nakamori, Y., Eliseo, J.R., Züttel, A., Jensen, C.M.: Complex hydrides for hydrogen storage. Chem. Rev. 107, 4111–4132 (2007) 128. Chater, P.A., David, W.I., Johnson, S.R., Edwards, P.P., Anderson, P.A.: Synthesis and crystal structure of Li 4 Bh 4 (Nh 2). Chem. Commun. 3, 2439–2441 (2006)

515 129. Weidenthaler, C.: Crystal structure evolution of complex metal aluminum hydrides upon hydrogen release. J. Energy Chem. 42, 133–143 (2020) 130. Brinks, H., Langley, W., Jensen, C., Graetz, J., Reilly, J., Hauback, B.: Synthesis and crystal structure of Β-Ald3. J. Alloys Compd. 433(433), 180–183 (2007) 131. Brinks, H.W., Istad-Lem, A., Hauback, B.C.: Mechanochemical synthesis and crystal structure of Α ‘-Ald3 and Α-Ald3. J. Phys. Chem. B. 110(110), 25833–25837 (2006) 132. Orimo, S., Majer, G., Fukunaga, T., Züttel, A., Schlapbach, L., Fujii, H.: Hydrogen in the mechanically prepared nanostructured graphite. Appl. Phys. Lett. 75, 3093–3095 (1999) 133. Yildirim, T., Hartman, M.R.: Direct observation of hydrogen adsorption sites and nanocage formation in metal-organic frameworks. Phys. Rev. Lett. 95, 215504 (2005) 134. Lee, H., Choi, Y.N., Choi, S.B., Kim, J., Kim, D., Jung, D.H., Park, Y.S., Yoon, K.B.: Liquid-like hydrogen stored in nanoporous materials at 50 K observed by in situ neutron diffraction experiments. J. Phys. Chem. C. 117, 3177–3184 (2013) 135. Zhao, Y., Xu, H., Daemen, L.L., Lokshin, K., Tait, K.T., Mao, W.L., Luo, J., Currier, R.P., Hickmott, D.D.: High-pressure/lowtemperature neutron scattering of gas inclusion compounds: Progress and prospects. Proc. Natl. Acad. Sci. 104, 5727 (2007) 136. Dincǎ, M., Dailly, A., Liu, Y., Brown, C.M., Neumann, D.A., Long, J.R.: Hydrogen storage in a microporous metal
organic framework with exposed Mn2+ coordination sites. J. Am. Chem. Soc. 128, 16876–16883 (2006) 137. Parvanov, V.M., Schenter, G.K., Hess, N.J., Daemen, L.L., Hartl, M., Stowe, A.C., Camaioni, D.M., Autrey, T.: Materials for hydrogen storage: structure and dynamics of borane ammonia complex. Dalton Trans., 4514 (2008). https://doi.org/10.1039/B718138H4514-4522 138. Kandemir, T., Schuster, M.E., Senyshyn, A., Behrens, M., Schlögl, R.: The Haber–Bosch process revisited: on the real structure and stability of “ammonia iron” under working conditions. Angew. Chem. Int. Ed. 52, 12723–12726 (2013) 139. Youngs, T.G.A., Manyar, H., Bowron, D.T., Gladden, L.F., Hardacre, C.: Probing chemistry and kinetics of reactions in heterogeneous catalysts. Chem. Sci. 4, 3484–3489 (2013) 140. Holderer, O., Carmo, M., Shviro, M., Lehnert, W., Noda, Y., Koizumi, S., Appavou, M.-S., Appel, M., Frielinghaus, H.: Fuel cell electrode characterization using neutron scattering. Materials (Basel). 13, 1474 (2020) 141. Borgschulte, A., Delmelle, R., Duarte, R.B., Heel, A., Boillat, P., Lehmann, E.: Water distribution in a sorption enhanced methanation reactor by time resolved neutron imaging. Phys. Chem. Chem. Phys. 18, 17217–17223 (2016)

Jisue Moon received her PhD from the University of California, Irvine, in 2018. She joined the Oak Ridge Laboratory in 2018 as a postdoctoral researcher at the Surface Chemistry and Catalysis Group. Presently, she is an R&D associate in the Isotope Application Group with research focus on molten salt chemistry and isotope target production using electrodeposition.

24

516

Yongqiang Cheng received his PhD from Johns Hopkins University in 2010. He joined the Oak Ridge National Laboratory in 2011 as a Shull Fellow. He is currently a computational instrument scientist focusing on applying atomistic modeling and advanced analytics to interpret neutron scattering data.

Zili Wu received his PhD from Dalian Institute of Chemical Physics in 2001. He worked as a postdoc at the Catalysis Center at Northwestern University before he joined the Oak Ridge National Laboratory in 2006. He is currently the leader of the Surface Chemistry and Catalysis Group with research focus on heterogeneous catalysis, nanomaterials, in situ/ operando spectroscopy, neutron scattering, and reaction mechanisms.

J. Moon et al.

Anibal J. Ramirez-Cuesta studied Physics in Argentina. He worked at the University of Reading, and the Rutherford Appleton Laboratory in the UK. He joined Oak Ridge National Laboratory (ORNL) in 2013 as Leader of Chemical Spectroscopy group and has leads the project Integrated Computational Environment for Modeling and Analysis of Neutron data ICEMAN.

Part V X-Ray Methods

25

X-Ray Diffraction (XRD) Daniyal Kiani

Contents 25.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519

25.2 25.2.1 25.2.2

Physics of XRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521 Sources of X-Rays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521 XRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521

25.3

Crystalline Solids and XRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 522

25.4

Understanding X-Ray Diffractograms . . . . . . . . . . . . . . . . . 525

25.5

Toward In Situ and Operando XRD Characterization of Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527

25.6

Case Studies Highlighting In Situ/Operando XRD in Catalyst Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Tailored Multifunctional Catalyst for Ultraefficient Styrene Production Under a Cyclic Redox Scheme . . . . . In Situ Studies of the Active Sites for the Water-Gas Shift (WGS) Reaction Over Cu
CeO2 Catalysts: Complex Interaction Between Metallic Copper and Oxygen Vacancies of Ceria . . . . . . . . . . . . . . . . . . . . . . . . . . Combined In Situ X-Ray Powder Diffractometry/Raman Spectroscopy of Iron Carbide and Carbon Species Evolution in Fe(
Na–S)/α-Al2O3 Catalysts During the Fischer-Tropsch Synthesis (FTS) . . . . . . . . . . . . . . . . . . . . .

25.6.1 25.6.2

25.6.3

529 529

529

532

25.7

Limitations of XRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532

25.8

Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 537

Abstract

This chapter provides a brief overview of the historic significance, basic theory/fundamentals, various types of reaction cells, and the application of X-ray diffraction (XRD) for catalyst characterization. Since this chapter is written primarily for a catalysis science audience, special emphasis is placed on the in situ and operando XRD characterization of catalysts. Finally, present limitations of XRD and recent developments that are helping in D. Kiani (*) Cummins Emission Solutions, Cummins Inc., Stoughton, WI, USA e-mail: [email protected]

overcoming those limitations of XRD for application toward catalysis science research are also highlighted. Keywords

In situ · Operando · X-ray scattering · Catalyst · Phases · Size · Bulk · Characterization · Crystalline

25.1

Introduction

X-rays were discovered in 1885 by the German physicist Wilhelm Conrad Röntgen, who called them X-rays because their nature was not known at the time [1]. Soon after the discovery of X-rays, in 1912, Max von Laue [2] discovered that crystalline materials scattered X-rays, which earned him the 1914 Nobel Prize in physics – more on this discovery can be found in references [3–5]. Since Laue’s discovery, X-ray diffraction (XRD) has become one of the most useful and impactful techniques to reveal the structure of crystalline materials. The understanding generated by XRD has paved way in various fields including polymer science, metallurgy [1], structural biology and pharmaceutical science [6–10], glass and ceramics science [11], geology/mineralogy [12, 13], and catalysis [2, 14–16]. The unique ability of X-rays to provide structural information is a direct consequence of the wavelengths of X-rays, which are typically about ~0.5–2.5 Å [1]. Testament to its significance and impact, three physicists were recognized with the Nobel Prize in physics for the development of XRD as a technique, and subsequently 22 researchers have been awarded the Nobel Prize for application of XRD to elucidate structure of biological molecules. A schematic chronological summary is shown in Fig. 25.1, highlighting the Nobel laureates and a brief description of their work’s relevance to XRD. Fundamentally, XRD is an experimental, bulk characterization technique that can determine the atomic and molecular structure of crystalline solids, in which the crystalline solid

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_25

519

520

D. Kiani

1914 M. Laue Physics: for his discovery of the diffraction of X-rays by crystals

1915 W. L. Bragg W. H. Bragg Physics: for their services in the analysis of crystal structure by means of X-rays

1962

M. F. Perutz J. C. Kendrew Chemistry: for their studies of the structures of globular proteins

1964

D. Hodgkin Chemistry: for her determinations by X-ray techniques of the structures of important biochemical substances

J. D. Watson F. H. C. Crick M. H. F. Wilkins Medicine: for their discoveries concerning the molecular structure of nucleic acids and its significance for information transfer in living material 1972

S. Moore W. H. Stein Chemistry: for their contribution to the understanding of the connection between chemical structure and catalytic activity of the active centre of the ribonuclease molecule. Remaining 1/2 of the prize was awarded to Christian B. Anfinsen for his work on ribonuclease, especially concerning the connection between the amino acid sequence and the biologically active conformation.

1976

1985

J. Karle H. A. Hauptman Chemistry: for their outstanding achievements in developing direct methods for the determination of crystal structures

W. N. Lipscomb Chemistry: for his studies on the structure of boranes illuminating problems of chemical bonding

1988 R. Huber J. Deisenhofer H. Michel Chemistry: for their determination of the three-dimensional structure of a photosynthetic reaction centre 1997

P. D. Boyer J. E. Walker Chemistry: for their elucidation of the enzymatic mechanism underlying the synthesis of adenosine triphosphate (ATP). Remainder 1/2 of the prize was awarded to Jens C. Skou for the first discovery of an ion-transporting enzyme, Na+, K+ -ATPase. R. D. Kornberg Chemistry: for his studies of the molecular basis of eukaryotic transcription

2003 2006

P. Agre R. MacKinnon Chemistry: for discoveries concerning channels in cell membranes;

2009

2012 Robert J. Lefkowitz B. Kobilka Chemistry: for studies of G-proteincoupled receptors

A. E. Yonath T. A. Steitz V. Ramakrishnan Chemistry: for studies of the structure and function of the ribosome

Fig. 25.1 Schematic summary of Nobel Prizes awarded between 1900 and 2020 for the development and application of XRD. The list of Nobel laureates and prize details are available publicly on wikipedia.org and nobelprize.org

causes a beam of incident X-rays to diffract into many directions. By measuring the angles and intensities of these diffracted beams, a three-dimensional picture of the density of electrons within the crystalline sample can be drawn. In

turn, from this electron density map, the mean positions of the atoms in the crystalline solid can be also determined. Therefore, XRD serves as a cornerstone characterization tool for the design and development of catalysts, since it

25

X-Ray Diffraction (XRD)

can provide critical information regarding the bulk structure and composition of solid catalysts such as metal oxides, perovskites, zeolites, etc. In the coming sections, the fundamentals of XRD will be discussed briefly. Curious readers are directed to reference [16] for a list of pertinent texts introducing the physics and data analysis of XRD.

25.2

Physics of XRD

25.2.1 Sources of X-Rays The simplest source of X-rays is an X-ray tube, which is comprised of a vacuum sealed cathode and anode. X-rays are produced whenever high-speed electrons from the cathode (e.g., W filament) collide with a metal anode (e.g., Cu, Mo, Al, etc.), generating X-rays and heat. An electron is ejected from one of the inner electron shells (i.e., K) of the metal atom as the high-speed electron hits the metal anode. To occupy the generated hole, an electron from a higher shell (i.e., L or M) drops to the vacant level with the emission of an X-ray photon, corresponding to the difference in energy between the two levels. When the electronic transition occurs between two adjacent shells (i.e., L ! K; 2p ! 1s), the radiation is termed Kα, and when the transition occurs between nonadjacent shells (i.e., M ! K; 3p ! 1s), the radiation is termed Kβ. Naturally, corresponding to the larger energy difference in the case of β transition, λβ is smaller than λα. Historically, X-ray filters were used to eliminate the unwanted radiation. However, filters cannot prevent the high background radiation and transmit radiation of a broad wavelength range. Therefore, most modern instruments utilize monochromators (single-crystal monochromators to be more precise), typically made of pyrolytic graphite (less selective) and silicon (more selective), to select the desired λ range (λβ or λα) of the radiation. Depending on the anode composition, that is, Cr, Fe, Co, Cu, Mo, Ag, and Mg, the wavelength of the generated X-rays can be chosen, although discretely, due to the sparse choices of suitable options (high melting point and good thermal conductance are among some of the requirements for choosing the anodic metal). According to Moseley’s law, as the atomic number of the anode metal increases, the wavelength of Kα, i.e., λα, decreases. For instance, the Kα wavelength for commonly used anode metals is as follows: Cr, 2.29 Å; Fe, 1.94 Å; Co, 1.79 Å; Cu, 1.54 Å; Mo, 0.71 Å; and Ag, 0.56 Å. Note that Cu and Mo are by far the most used anodes in typical laboratory XRD units. The choice of anode metal and ultimately the λα is critical toward successful employment of XRD characterization as λ determines the energy of the X-rays, which in turn affect the penetration (mean free path length) of the incident X-rays within the material being characterized. It suffices to note that the choice of anode

521

(and hence the λα) will affect some of the features of the final diffractogram including peak intensity, number of peaks observed, and the background in the measured XRD. Moreover, it is worth noting that recent advancements in X-ray sources have now enabled generation of extremely high-flux X-rays in comparison to conventional X-ray tubes [17]. Although still widely used, X-ray tubes are not the brightest sources of X-rays, as clearly seen in Fig. 25.2, with the highest intensity X-rays being produced by latest generation synchrotrons around the world [17]. In comparison to a standalone XRD instrument using an X-ray tube source – the kind most laboratories in academic institutions utilize – a synchrotron X-ray source produces spectral brightness 9–24 orders of magnitudes higher, providing unparalleled temporal, spatial, and spectral resolution. The brightness of X-rays produced in synchrotrons is one of the major reasons that most temporally and spatially resolved XRD studies in the field of catalysis science are only possible at such synchrotron sources and would not furnish clean, revealing results if conducted at standalone units utilizing conventional X-ray tubes.

25.2.2 XRD Diffraction is the elastic scattering experienced by X-rays when they interact with the atoms in a crystalline solid. Elastic scattering occurs when a plane wave interacts with an obstacle or a slit whose size is approximately that of the wavelength (0.5–2.5 Å) [1]. In the most simplified X-ray scattering experiment, a beam of X-rays hits the sample, and the intensity of the scattered rays is measured as a function of the scattering angle. As multiple X-rays are scattered in various directions as they interact with multiple planes within the crystalline solid, constructive and destructive interference in the scattered X-rays occurs due to interference. Where constructive interference occurs, the intensity of the scattered rays is higher than the incident rays, and where destructive interference occurs, the intensity is lower (often negligible). Constructive interference only occurs when a specific condition is met, which was outlined by the English physicists Sir W.H. Bragg and his son Sir W.L. Bragg in 1913 [18–20], resulting in a shared Nobel Prize in physics in 1915. According to Bragg’s law [4, 18], constructive interference occurs when: nλ ¼ 2d sin θ where n is an integer describing the order of reflection, λ is the wavelength of the incident X-rays, d is the spacing between the planes scattering the X-rays, and θ is the angle of incidence.

25

522

D. Kiani

Wavelength l (Å)

a) 1240

1020

12.4

1.24

b) 18

Undulators Wigglers

1016

1014

Bends 7 GeV

1012

1.9 GeV Liquid metal-jet PT Mo K

1010

Cu K Microfocus XRTs

108

Standard XRTs

106 Bremsstrahlung 104

European XFEL

1033

Peak brightness, photon s−1 mrad−2 mm−2 (Δl/l = 0.1%)

Spectral brightness, photon s−1 mrad−2 mm−2 (Δl/l = 0.1%)

10

1035 SACLA PAL-XFEL SwissFEL 1031

LCLS FLASH FERMI @Elettra

1029

1027

1025

PETRA III

SPring-8

ESRF 1023 APS SLS 1021

XRT

BESSY 1019

102 1

10

2

3

4

10 10 10 Photon energy e (eV)

10

5

Fig. 25.2 A comparison of the brightness of modern sources of X-rays where (a) shows the spectral brightness of some third-generation synchrotron sources and laboratory-used X-ray tube sources. The vertical dashes denote the brightness of the characteristic radiation lines of standard rotating-anode tubes (Cu Kα and Mo Kα) of various powers; the squares indicate the brightness of the characteristic lines of a sealed tube with a fixed anode (standard XRTs), a microfocus tube with a fixed anode (microfocus XRTs), and a microfocus tube with a liquid metal

25.3

Crystalline Solids and XRD

A crystalline solid is a material in which particles (atoms, molecules, or ions) occupy fixed positions, forming a periodic arrangement called a lattice [21]. Conceptually, a lattice (point lattice) is defined as an array of points arranged in space such that each point has identical surroundings [1]. Depending on the identity of the particles (atoms, molecules, or ions) and the nature of the attractive force that is holding those particles in fixed positions, crystalline solids can be ionic (as in NaCl), covalent (as in diamond), molecular (as in ice), or metallic (as in Au), with each type of crystalline

101

102

103

104

105

106

Photon energy e (eV)

anode (liquid metal-jet XRT) GaKα 9.25 keV (1.34 Å); bremsstrahlung is indicated as a shaded area where the lower boundary corresponds to fixed-anode tubes, the intermediate area to a standard rotating-anode XRT, and the upper boundary to a microfocus rotating-anode XRT. (b) The maximum peak brightness of the most powerful operating sources of third-generation synchrotrons and X-ray free-electron lasers (XFELs) – fourth-generation synchrotrons sources. (Figure reprinted with permission from Ref. [17])

material exhibiting unique structure and physicochemical properties. Regardless of the type of the crystalline solid, in each case, the smallest repeating unit that generates the crystalline lattice upon translation in three dimensions is called a unit cell [21]. Note that while a lattice is only composed of fixed points, a unit cell includes the empty space between the points as well as those points. Fourteen major types of lattices, called Bravais lattices, are encountered in crystalline solids, generated by combining the seven major types of lattice systems [22, 23]. A Bravais lattice – named after August Bravais (1850) [24] – is an infinite array of points, generated by translating them in three-dimensional space to describe the overall crystalline

25

X-Ray Diffraction (XRD)

523

structure and boundaries. Each of the 14 types of Bravais lattices has unique properties based on six parameters (Fig. 25.3) that describe the unit cell in each case, namely: I. The relative lengths of the unit cell in three-dimensional space (a, along x-axis; b, along y-axis; c, along z-axis) II. The angles between those relative lengths (α, between b, c; β, between a, c; γ, between a, b) Depending on the lattice type, values for these lattice constants/parameters (a, b, c; α, β, γ) can differ [23]. For example, in an orthorhombic unit cell, a 6¼ b 6¼ c and α ¼ β ¼ γ ¼ 90 [23]. On the other hand, in a tetragonal unit cell, a ¼ b 6¼ c, and α ¼ β ¼ γ ¼ 90 [23]. Moreover, in a triclinic unit cell, a 6¼ b 6¼ c, and α 6¼ β 6¼ γ. Note that the congruence of a, b, and c with rectangular coordinate axes x, y, and z, respectively, is true only for a simple cubic, tetragonal, and orthorhombic unit cells for which α ¼ β ¼ γ ¼ 90 . However, it is not the case generally. The rectangular coordinate axes x, y, and z should be therefore interpreted as referencing the threedimensionality rather than general orthogonality. Broadly speaking, the 14 Bravais lattices are grouped into seven crystal systems, which are further characterized into 32 crystallographic point groups. A crystallographic point group describes all the symmetry operations (e.g., rotation) that can be performed on a crystal lattice that, when performed, would leave one point in the crystal structure unchanged or fixed. Importantly, a crystallographic point group must maintain the three-dimensional translational symmetry that defines crystalline nature of the solid in the first place. A point group in addition to other symmetry operations collectively then forms a space group, where the space group

is defined as the group of all transformations that would leave the crystal lattice unchanged. In three-dimensional space, there are a total of 230 space groups. With rare few exceptions, all crystalline solids can be classified according to the 14 Bravais lattices, the 32 point groups, and the 230 space groups briefly introduced in this section. Within the crystalline lattice, the positions of its constituent particles leads to distinctive flat surfaces called planes, described by the bounds or intercepts on the three axes. When the planes in a crystalline lattice are described in the reciprocal three-dimensional space, this system of description is referred to as Miller indices. Miller indices are shorthand notations, well understood throughout the crystallography community that unambiguously describe a plane within a crystalline lattice using the h, k, l fractional notation. Specifically, Miller indices are defined as the reciprocals of the fractional intercepts which the plane makes with the crystallographic axes [1]. A plane running parallel to an axis is assumed to intersect that axis at infinity. For instance: Coordinates: x, y, z. x y z Fractional coordinates: , , a b c:   x y z
1 Miller indices (h, k, l): , , a

∴ (h, k, l):

a

b

c

x

y

z

,

,

c

 where a, b, and c are lattice

parameters and x, y, and z are coordinates bounding the plane. The following example (Fig. 25.4) illustrates Miller indexing of a crystalline lattice plane that intersects the x-axis at a but does not intersect the y- and z-axes:

z

z

c

c

a

b



b

b

b

y

y

g a

a

x

Fig. 25.3 An orthorhombic unit cell representation in threedimensional space, with relative lengths labeled as a (x), b ( y), and c (z) and angles labeled as α, β, and γ

x

Fig. 25.4 An orthorhombic unit cell representation in threedimensional space, with relative lengths labeled as a (x), b ( y), and c (z). The shaded region represents the (100) plane

25

524

D. Kiani a 1 1

Fractional coordinates: , , a

∴ (h, k, l):



a a

,

b

,

1

b

c

c

 : (1, 0, 0)

1

Lastly, the following example (Fig. 25.5) illustrates the Miller indices of a crystalline lattice plane that intersects the x, y, and z axes at ½ (a, b, c): Fractional coordinates:

0:5a 0:5b 0:5c a

,

b

,

c

∴ (h, k, l): ( 2, 2, 2) Having briefly introduced the concept of Miller indices and planes, we now progress to the next step in understanding how Miller indices and planes are related to the observed XRD peaks. As shown in Figs. 25.3, 25.4, and 25.5, atomic planes are described by the Miller indices, e.g., (222). In practical terms, a (222) plane will not exist by itself; there will be a whole set of parallel planes to the (222) plane, written as , and the distance between the two parallel planes is termed as interplanar spacing or d – the same d as in the Bragg’s equation above. In common notation, d has Miller indices mentioned in the subscript to highlight the set of parallel planes being described. In the case of , interplanar spacing would be noted as d222. A list of formulas is readily available that mathematically correlates interplanar spacing with the lattice constants/parameters a, b, and c for various types of unit cell systems, e.g., cubic, tetragonal, hexagonal, etc. For instance, for a simple cubic unit cell where a ¼ b ¼ c,

1 d2

¼

h2 þk2 þl2 a2

orthorhombic unit cell where a 6¼ b 6¼ c,

1

d2

, while for an

¼

h2 a2

þ

k2 b2

l2

þ 2. c

z

c

b

y

a

x

Fig. 25.5 An orthorhombic unit cell representation in threedimensional space, with relative lengths labeled as a (x), b ( y), and c (z) and angles labeled as α, β, and γ. The shaded region represents the (222) plane

However, the derivation or the mathematical context for these formulas is beyond the scope of this chapter. We direct the readers to readily available formula lists correlating d and the lattice constants for all crystal systems, like those found in appendix of reference [1]. It suffices to note at this point that the interplanar spacing, Miller indices, and lattice constants/ parameters (a, b, c; α, β, γ) are critical in understanding the physical explanation behind XRD peaks and using XRD data to understand the structure of a solid sample. For a more rigorous explanation and mathematical description of the relevant background, curious readers are directed to seminal works such as reference [1]. Unique planes and interplanar spacing can help identify the phase of the crystalline material being studied, where different phases of a crystalline material are chemically identical but structurally different due to the differences in constituent unit cells. For example, as shown in Fig. 25.6, a mixture of Fe2O3 polymorphs can be generated by oxidizing Fe3O4 at various temperatures [25]. The α-Fe2O3 (hematite) crystallizes in a corundum structure characterized by a slightly distorted hexagonal close-packed Bravais lattice, exhibiting distinct XRD peaks below 80 corresponding to (012), (104), (113), and (024) planes present in the α-Fe2O3 phase. On the other hand, γ-Fe2O3 (maghemite) crystallizes in the cubic tetragonal lattice, exhibiting unique XRD peaks corresponding to (111), (220), (311), and (400) below 80 . Summarily, as shown in Fig. 25.6, stronger oxidation at 600  C produces α-Fe2O3, while a milder oxidation at 350  C produces γ-Fe2O3 [25]. Therefore, using this information, one can analyze an oxidized Fe3O4 sample via XRD to determine the identity of the Fe2O3 polymorph (hematite vs maghemite) present using the unique peaks corresponding to different planes from each polymorph. Likewise, similar XRD analysis can also help determine the polymorph population if a mixture of various polymorphs is present, although careful calibration is required before XRD data can be used to quantify phase population in a mixed sample. Summarily, noting the aforementioned key concepts, bulk structure of a crystalline solid can be determined via XRD in the following three steps [1], where step c is the most difficult and not always achievable: (a) Shape and size of the unit cell are deduced from the angular positions of the XRD lines (reflections) by first assuming the lattice type and assigning miller indices to each reflection from various planes. (b) Atoms per unit cell volume (i.e., atomic density) are calculated from the pertinent unit cell parameters mentioned above. (c) Atomic positions within the unit cell are determined by analyzing the XRD peak intensities. Beyond steps a to b, structure refinement becomes critical, especially to accomplish c. The Rietveld refinement method

X-Ray Diffraction (XRD)

γ-Fe2O3

300

440

511

422

018

214

116

α-Fe2O3

400

220

350 °C

311

600 °C

024

113

012

110

104

25

111

Fig. 25.6 XRD pattern of uncoated Fe3O4 nanoparticles oxidized at different oxidation temperatures of 25, 250, 350, and 600  C in air exhibiting diffraction peaks assigned to various Miller indices of α-Fe2O3 and γ-Fe2O3. (Figure reprinted with permission from reference [25]. Copyright 2008 American Chemical Society)

525

Intensity (a.u.)

25

250 °C 25 °C 20

Fe3O4 40

60

80

100

2q (deg)

[26], developed in the 1960s, is the most used method for crystal structure “refinement” from XRD data. It is important to remember that the Rietveld refinement requires a crystal structure model as the starting point and offers no way to deduce such a model on its own from the XRD data. However, it can be used to find structural details missing from a partial structure or infer new structural details from the experimental data. The Rietveld method involves nonlinear least squares method of fitting a calculated profile based on an assumed structural model to the acquired experimental data, necessitating tuning, approximating, and even guessing all structural and instrumental parameters. To ensure successful refinement (i.e., good fit between model and experimental data), reasonable initial approximation of many parameters including peak height, unit cell dimensions, and atomic coordinates in the crystal structure is necessary. The successful outcome of the refinement is directly related to the quality of the experimental XRD data, the accuracy of initial approximations of the structural parameters and the structural model, and the experience of the user.

25.4

Understanding X-Ray Diffractograms

An X-ray diffractogram is the “fingerprint” of a crystalline solid, for it reveals critical information including unit cell, space group, distribution of electron density, crystallite size, and strain within the crystalline solid. Ideally, a peak on the diffractogram (i.e., where Bragg’s law is satisfied and constructive interference of scattered X-rays occurs) should be a

sharp line. However, in practice, XRD peaks are seldom narrow and exhibit appreciable width. It is fairly common to have peaks with a finite width, i.e., broadening, which can reveal yet another critical piece of information regarding the crystalline solid: its coherent scattering domain size. The shape of the peak (position, full width at half maximum intensity – FWHM) is related to the sub-micrometer coherent domain size via Scherrer’s equation [27, 28] as follows: D¼

Kλ βcosθ

where D is the calculated crystallite size; K is a dimensionless constant depending on the shape of the crystallite and its size distribution, usually close to 1; [28] λ is the wavelength of incident X-rays; β is the FWHM of the XRD peak in radians; and θ is the peak position in radians. It should be noted that Scherrer’s equation is not applicable to particles above 100–200 nm in dimension and largely remains applicable to smaller nanoscale dimensions. Moreover, Scherrer’s equation is used to calculate the coherent scattering domain in a crystalline solid, i.e., its crystallite size, which may be smaller than the actual particle size [27]. Coherent scattering domain size is defined as the size within which three-dimensional periodicity in an atomic structure is retained, which may not be the case if a particle is composed of multiple crystallites where the crystallite size will always be smaller than the particle size. An example, taken from reference [29], where the crystallite size (of CoPt3 nanocrystals) impacts the width of the observed reflections in a diffractogram is shown in

526

D. Kiani

Fig. 25.7. HR-TEM images clearly show that sample A (left panel) contains smaller sized nanocrystals of CoPt3 than sample B (right panel). Quantitative analysis of the

a)

HR-TEM data shows that indeed the mean particle size is 4.86 nm in sample A, while it is 8.50 nm in sample B. Both samples A and B were also studied via XRD, followed by

b)

Sample A

Sample B

HR-TEM 20 nm

20 nm

350 a) Sample A Particles Mean Std. deviation Skewness Excess kurtosis

Frequency

250 200 150

b) Sample B

150 : 2824 : 4.86 nm : 0.45 nm : −0.36 : 1.62

Frequency

300

: 1156 : 8.50 nm : 0.59 nm : 3.60 : 38.06

Counted particles Mean Standard deviation Skewness Excess kurtosis

100

50

100 50

0

0 1

2

3 4 5 Diameter (nm)

6

7

6

8

10 12 14 Diameter (nm)

16

XRD sample A, diameter = 5.1 nm

(111)

3000

(200)

(311)

(220)

2000

6000 Counts

Counts

4000

(111)

sample B - diameter = 8.4 nm (200)

4000

(220) 2000

1000 30

45

60 2q / °

75

35

45

(200)

40 2q / °

45

68 72 2q / °

52

35

40 45 2q / °

90

(200)

80 84 2q / °

44

d = 8.52 nm 48 52 2q / °

(220)

d = 4.98 nm 76

75

(111)

(311)

d = 5.33 nm

60

d = 8.40 nm

d = 4.92 nm 44 48 2q / °

(220)

64

30

2q / °

(111) d = 5.09 nm

30

90

d = 8.21 nm 88

60

64

68 72 2q / °

76

Fig. 25.7 HR-TEM and XRD for (a) ~5 nm and (b) 8.5 nm nanoparticles. (Figure adapted with permission from reference [29]. Copyright 2005 American Chemical Society)

25

X-Ray Diffraction (XRD)

Scherrer’s analysis. Visually, it is apparent in the diffractogram in each case that the smaller CoPt3 nanocrystals in sample A exhibit broader XRD peaks than the larger nanocrystals in sample B. The crystallite sizes calculated using Scherrer’s equation for nano-CoPt3 in sample A applied to various reflections in the diffractograms, i.e., (111), (200), (220), and (311), all yield similar sizes of 4.92–5.33 nm. Scherrer’s analysis of various reflections in sample B also leads to a narrow range of crystallite sizes between 8.2 and 8.5 nm. In both cases, Scherrer’s analysis leads to crystallite sizes in good agreement with the size estimated from quantitative analysis of HR-TEM data. At this point, a word of caution is necessary regarding the applicability of Scherrer’s equation to XRD results to calculate size. Crystallite size is not the only factor that can lead to peak broadening in a diffractogram, and a variety of other factors may also potentially contribute to the broadening of the diffraction peak. For example, secondary factors can also cause peak broadening, for example, inhomogeneous strain (compressive or tensile) and stress within a lattice, lattice imperfections such as dislocations, stacking faults, twinning, vacancies, chemical heterogeneities, and anisotropic crystallite shape where one dimension is significantly smaller than the other dimension as in nano-rods. If secondary factors are also contributing to the observed broadening in the XRD peaks, then the crystallite size will be larger than that predicted by Scherrer’s equation, with the additional peak width coming from the secondary factors. Therefore, secondary factors that also cause peak broadening in conjunction with the primary factor i.e finite size effects, need to be differentiated accurately to estimate the crystallite size from a diffractogram [16]. Another source of broadening comes from the instrument itself. Instrumental broadening is dependent on various factors like the energy range of the X-ray source (i.e., monochromator efficiency), macroscopic topology of the specimen surface (i.e., how flat the sample is), and axial divergence of the X-ray beam (typically corrected by employing Soller slits) [30]. Uniform strain can shift the diffraction peak position to higher or lower diffraction 2θ angles without changing the intensity of those peaks [16]. On the other hand, nonuniform strain will not change the position of the diffraction peaks but will lead to a higher FWHM (i.e., broadening) and a decrease in the peak intensities [16, 31]. Therefore, one key qualitative difference between peak broadening in XRD originating in finite size and other factors (discussed above) is that when finite size effects are the predominant reason for peak broadening, all XRD peaks will be broadened similarly, and applying Scherrer’s equation to all peaks should give very similar (ideally same) crystallite size, as in Fig. 25.7. On the other hand, if secondary factors are also contributing to peak broadening, all peaks will not give the same or very similar crystallite size estimates owing to varying FWHM of the

527

different peaks. In such a case, Scherrer’s equation is typically applied to the first peak, as it is the least affected by secondary factors.

25.5

Toward In Situ and Operando XRD Characterization of Catalysts

Since its humble beginnings in the early twentieth century, XRD has come a long way in terms of its capabilities for structural analysis. The advancements in XRD were brought about thanks to the development of energy-resolved, highflux X-ray in conjunction with extremely sensitive detectors. Such advancements have made XRD quite a useful characterization tool for the catalysis science community to study catalysts at working conditions. A lengthy discussion on the origin, explanation, and effects of each of the major advancement in XRD is beyond the scope of this chapter, but we believe it will be beneficial for the readership to briefly appreciate the basics of various XRD cells used for characterization of catalysts at or near reaction conditions via in situ and operando XRD. There are two basic types of XRD cells encountered in the application of XRD for catalyst characterization, based on the diffractometer geometry being used in the instrument: (a) Bragg-Brentano (reflection) geometry: In this geometry, schematically shown in Fig. 25.8a, an infinitely thick, flattened layer of the powdered catalyst is deposited on the sample holder, while the X-ray source and the detectors are rotated during XRD. Bragg-Brentano geometry is widely used in standalone XRD units such as those in academic research labs and universities. Practically, if ex situ XRD is being conducted on a powdered sample, a glass slide or any flat sample holder cup can be used to hold the flat, infinitely thick film of the powder. Here, infinitely thick implies a sample thickness that is greater than the X-ray penetration depth into the sample. X-ray penetration depth in a sample in turn depends on the mass absorption coefficient of that sample and the incident angle of the X-ray beam. Moreover, if a thin sample film needs to be characterized, a zero
/low-background holder is required. On the other hand, if in situ or operando XRD is being conducted, where the catalyst powder has to be exposed to a certain temperature and gaseous environment, commercial cells are available that enable such operation. For example, a commercially available Anton Paar’s DHS-1100 cell is shown in Fig. 25.8b. Note that any in situ/operando XRD cell (commercial or homemade) is still following the BraggBrentano reflection geometry and hence means that the catalyst powder is not exposed to the gaseous environment in the same way as a catalyst operating in a plug

25

528

D. Kiani

flow reactor would be. Therefore, researchers should keep in mind that spectro-kinetics and structural dynamics attained through such in situ/operando reflection cells should only be correlated to reactor data cautiously as in one case gas flows through the catalyst bed and in the other it flows over the catalyst bed. Furthermore, such a setup primarily analyzes the interface between the surrounding gas and the catalyst’s outer layers, where substantial temperature gradients are possible since the heater is usually only in contact with the bottom layer of the catalyst film depending on the location of the heater [16].

(b) Debye-Scherrer (transmission) geometry: In this geometry, schematically shown in Fig. 25.9, a near-parallel incident beam of X-rays with sufficient cross section to bathe the whole powder sample is irradiated over the catalyst bed. In this case, the powdered sample is held inside a capillary, with gas flowing from one end to the other through the catalyst bed, just like in a plug flow reactor. Broadly speaking, this type of geometry and such capillary reaction cells are mostly used in synchrotronbased XRD facilities. Examples of capillary cells used in transmission geometry can be found in the next section: case study b-c. It suffices to note that the Debye-Scherrer

a)

b) 7

4

5

Detector x-ray tube

2q

Powder film

1

Sample holder

6

Fig. 25.8 (a) A simplified schematic of the Bragg-Brentano (reflection) geometry used in most standalone XRD units. (b) An image of a commercial in situ/operando XRD cell used with the Bragg-Brentano geometry. Major parts of the cell are as follows: (1) AlN base plate,

2

3

(2) sample, (3) TiAl holders, (4) graphite dome, (5) ring with cooling air exhausts, (6) heat sink, and (7) feed pipes. ((b) is reprinted from reference [32] under the Creative Commons license)

Fig. 25.9 A simplified schematic of the Debye-Scherrer (transmission) geometry, often used in synchrotron-based XRD facilities. (Figure reprinted with permission from reference [33])

Detector Scan

Parallel beam

2q

λ λ

Synchrotron white beam

Monochromator

Sample in capillary

25

X-Ray Diffraction (XRD)

geometry utilizing capillary cells is the preferable one for catalyst characterization studies, owing to better quality diffraction data from high-flux synchrotron sources than laboratory-based instruments and more representative structural dynamics and spectro-kinetics due to plug flow and due to the ability to incorporate other beams and detectors into this setup to create multimodal techniques like Raman-XRD, IR-XRD, etc.

25.6

Case Studies Highlighting In Situ/ Operando XRD in Catalyst Characterization

25.6.1 A Tailored Multifunctional Catalyst for Ultraefficient Styrene Production Under a Cyclic Redox Scheme Styrene is an important commodity chemical that is highly energy intensive to produce, generating a large CO2 footprint. In this study, a multifunctional (Ca/Mn)1
xO@KFeO2 coreshell redox catalyst was used to convert ethylbenzene to styrene with up to 97% single-pass conversion and >94% selectivity in a stepwise redox-ODH autothermal operation, as shown in Fig. 25.10a [34]. This catalyst represents a 72% yield increase compared to commercial dehydrogenation on a relative basis, leading to 82% energy savings and 79% CO2 emission reduction [34]. In situ XRD was used to study the structural dynamics of this novel catalyst during styrene production. In situ XRD experiments were conducted on an Empyrean X-ray diffractometer equipped with an Anton-Paar XRK-900 reactor chamber. Figure 25.10c shows the dynamic phase change observed with in situ XRD under cyclic ethylbenzene ODH to styrene, followed by reoxidation in air at 600  C. Under the styrene-selective region (denoted as operating regime), the primary phases observed include a CaO–MnO solid solution phase and a potassium ferrite (KFeO2) phase. The CaO–MnO solid solution, with cation defects, continues to release lattice oxygen during the ODH step, as indicated by the continuous peak shift (from 40.5 to 34.4 ). As the catalyst reduces (under over-reduction regime), metallic iron Fe0 formed, which was found to reversibly incorporate back into the redox catalyst upon reoxidation. During the reoxidation regime, full reoxidation of (Ca/Mn)1
xO@KFeO2 catalyst led to the formation of two new phases, namely, Ca2Fe2O5 and K0.296Mn0.926O2. KFeO2 was also observed in the fully reoxidized catalyst. Based on critical insights generated from in situ XRD in conjunction with other detailed characterizations (not discussed herein), this study showed that under the working conditions, the (Ca/Mn)1
xO@KFeO2 catalyst had a KFeO2-rich surface, which was identified as the catalytically active phase for

529

ethylbenzene to styrene conversion [34]. On the other hand, the core was found to be composed of a cation-deficient CaO–MnO solid solution, which was suggested to be involved in reversible lattice oxygen donation (in the ODH step) and uptake (in the reoxidation step) [34].

25.6.2 In Situ Studies of the Active Sites for the Water-Gas Shift (WGS) Reaction Over Cu
CeO2 Catalysts: Complex Interaction Between Metallic Copper and Oxygen Vacancies of Ceria Cu-doped CeO2 catalysts (written as Ce(1
x)CuxO2) are known to be active for WGS reaction, which converts a mixture of CO and H2O to a mixture of H2 and CO2. The respective role of CeO2 and Cu dopant in the WGS over the Ce(1
x)CuxO2 catalysts had been a point of contention in the catalysis literature. Regarding Cu, both metallic Cu0 and reduced Cu1+ had been proposed as active sites for the WGS reaction. On the other hand, CeO2 (i.e., Ce4+) had been proposed to be a spectator. These speculative hypotheses were put to rest when an in situ XRD study categorically showed that metallic Cu0 and oxygen vacancies in reduced CeO2 (i.e., Ce3+) were both involved in the WGS reaction and that both needed to be present in the catalyst for a high WGS activity [35]. In the aforementioned study, in situ, timeresolved X-ray diffraction (TR-XRD) experiments were carried out at beamline X-7B of the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory, using a flow-through capillary cell depicted in Fig. 25.11a [35]. The WGS reaction was carried out isothermally at 200, 300, 400, and 500  C, during which in situ XRD was used to characterize the catalyst during reaction. By comparing the XRD results in Fig. 25.11b and mass spec product curves in Fig. 25.11c, it was found that the signals for the H2 and CO2 products increased concomitantly with the appearance of metallic copper, i.e., Cu0, indicating the important role of reduced Cu0 during WGS reaction [35]. Moreover, since Ce3+ has a larger ionic radius than Ce4+, the lattice of crystalline CeO2 expands as Ce4+ in CeO2 reduces to Ce3+ along with the formation of oxygen vacancy. Likewise, if Ce4+ is replaced by a smaller radius Cu2+, the lattice of the mixed oxide Ce(1
x)CuxO2 shrinks. Therefore, changes in the lattice parameter calculated from the in situ XRD data can be directly correlated to the concentration of oxygen vacancies, the degree of Ce4+ reduction, and the Cu2+ ions in the lattice of Ce(1
x)CuxO2. In the present study, a composition of Ce(1
x)CuxO2 was studied such that x ¼ 0.2, i.e., Ce0.8Cu0.2O2. As shown in Fig. 25.11d, e, the lattice parameters for CeO2 and metallic Cu, which were determined from the (111) diffraction peak of Ce0.8 Cu0.2O2 in different gases at 300  C, vary as the gaseous environment changes [35]. The

25

530

D. Kiani

CeO2 lattice parameter increased significantly after exposure to CO (Fig. 25.11d) and decreased partially with exposure to H2O, indicating that CO reduced Ce4+ to Ce3+ where Ce3+ has a larger ionic radius, while H2O oxidized reduced Ce3+ back to the smaller Ce4+ [35]. This can be further interpreted as CO creating oxygen vacancies and H2O eliminating those

a)

Styrene, H2O

vacancies [35]. After exposure to O2 at the end of the experiment, the CeO2 lattice parameter decreases back to its initial value observed at the start of the experiment prior to WGS reaction, signifying that the reduction and oxidation of Ce0.8Cu0.2O2 was fully reversible.

b)

O2-lean air

Redox catalyst

MeMe’Ox-1

MeMe’Ox

Ethylbenzene

Air

KFeO2

c)

Fe appearance with over-reduction

Ca2Fe2O5 support

1 120 Over-reduction

0.9 0.8

Time (min)

0.7 Operating regime

80

0.6 0.5

60

Oxygen release from CaO-MnO solid solution

K0.296Mn0.926O2 appearance

Re-oxidation

0.4

Relative peak intensity

100

0.3

40 Over-reduction

0.2 0.1

20

Operating regime 0 15

20

25

30

35

40

45

2q (q)

Fig. 25.10 (a) A schematic illustration of redox-ODH of ethylbenzene. MeMe′Ox represents a generic redox catalyst. (b) An image of the commercial Anton-Paar XRK-900 reactor cell, similar to the in situ XRD cell used in reference [34]. (Figure reprinted with permission

from reference [16]. Copyright 2009 Elsevier). (c) In situ XRD under cyclic ethylbenzene ODH and air reoxidation steps at 600  C. (Figures adapted from reference [34] under the Creative Commons license)

25

X-Ray Diffraction (XRD)

531

a)

25

Catalyst sample

Gas flow

Thermocouple

Sapphire cappilary

To goniometer head

To G.C.

b)

d)

5.50 CO

H2O

CO

Cu 5.48

400 Ceria lattice (Å)

300 150 118 16

18

20

22 24 2q (degree)

25 qC

25q 30

28

CO/H2O 5.44 O2 300 qC

5.42

400 qC

He

H2 CO2

0

300 qC

200 qC

5

10

15

20

25

Time (h)

e)

150 qC

Ce0.8Cu0.2O2

5.40

3.66 Cu lattice (Å)

Relative intensity (a.u.)

c)

26

5.46

3.65

CO CO/H2O

H2O

CO

O2

3.64 300 qC

He

3.63

Ce0.8Cu0.2O2

3.62 0 0

2

4

6

8 Time (h)

10

12

14

5

10

15

20

25

Time (h)

Fig. 25.11 (a) Simplified schematic of the flow cell used for in situ XRD of catalysts. (b) TR-XRD patterns for Ce0.8Cu0.2O2 catalyst during the WGS reaction at different temperatures with (λ) 0.992 Å. (c) The relative concentration of H2 and CO2 products as a function of time at different temperatures during the WGS reaction. Lattice parameters

during the gas-switch experiments at 300  C for (d) ceria and (e) metallic copper. (Figure (a) adapted with permission from reference [36]. Copyright 2001 American Chemical Society. Figures (b–e) adapted with permission from reference [35]. Copyright 2006 American Chemical Society)

On the other hand, the behavior of the Cu0 formed during WGS reaction in Ce0.8Cu0.2O2 catalyst was much different [35]. Firstly, as the sample was heated to 300  C in He, no metallic Cu0 was observed (Fig. 25.11e), and oxidized Cu2+ persisted in the sample. Metallic Cu0, however, formed in the Ce0.8Cu0.2O2 sample as the gas was switched to 5% CO/He (i.e., reducing conditions) and persisted during the WGS reaction where CO and H2O are present simultaneously. Even during H2O-only exposure, the metallic Cu0 formed during WGS persisted and only disappeared after the gaseous environment was switched to O2 (i.e., strongly oxidizing)

[35]. The disappearance of metallic Cu0 in O2 environment confirms that the reduction and oxidation of Cu in Cu0.2Ce0.8O2 was completely reversible, although not under H2O [35]. Lastly, the lattice parameter of the formed metallic Cu0 was found to be invariant toward different gas environments and remained relatively constant (Fig. 25.11e) [35]. Summarily, these results confirmed that both metallic Cu0 and reduced Ce3+ species were critically involved in the WGS reaction [35]. Further experiments (not shown herein) revealed that neither metallic Cu0 nor the O vacancies and reduced Ce3+ exhibit high WGS activity when present

532

independently. However, when both types of sites are co-populated, as in Cu0.2Ce0.8O2, high WGS activity is observed [35].

25.6.3 Combined In Situ X-Ray Powder Diffractometry/Raman Spectroscopy of Iron Carbide and Carbon Species Evolution in Fe(
Na–S)/α-Al2O3 Catalysts During the Fischer-Tropsch Synthesis (FTS) FTS is an industrially viable chemical route to convert syngas (a mixture of CO and H2) into value-added hydrocarbons. Fe-based catalysts like the unpromoted and promoted alphaAl2O3 supported Fe (written as Fe/α-Al2O3) are known to be excellent catalysts for high-temperature FTS. However, the exact nature and identity of the active phase and the type of carbon formed that leads to deactivation of the catalyst remained unclear until recently, when combined in situ XRD and Raman (not discussed herein) were utilized to study the evolution of carbon species under in situ conditions [37]. In the aforementioned study, the catalysts were characterized under FTS reaction conditions of 340  C at 10 bar in a H2/CO 2:1 gas mixture using a modified iKey in situ XRD plug flow capillary reactor cell from Cape Catalytix, as shown in Fig. 25.12a [37, 38]. Both unpromoted Fe/α-Al2O3 and Na- and S-promoted Fe/α-Al2O3 catalysts (Na-S-Fe/α-Al2O3) were studied via in situ XRD and Raman spectroscopy to elucidate the role of Na and S promoters in FTS [37]. The Rietveld quantitative phase analysis (R-QPA) was used to quantify from the collected in situ powder XRD (pXRD) data the Fe phase content changes and the Fe phases’ crystallite diameters [37]. Three independent Lorentzian peaks were used for fitting the amorphous content scattering and/or diffraction contribution within the range of 8–12 2θ [37]. In situ XRD was used to monitor the phase transformations and changes in the crystallite sizes during FTS over promoted and unpromoted Fe/α-Al2O3 catalysts. Shown in Fig. 25.12b–i, in situ pXRD evidences most intense and sharp diffraction peaks in the individual diffraction patterns from the support α-Al2O3 (e.g., at 19.58 and 27.13 2θ) [37]. These intense XRD peaks are excluded for convenience as they are invariant and not pertinent to the chemistry. With respect to the phase changes in the Fe/α-Al2O3 vs Na–SFe/α-Al2O3 catalysts during FTS reaction, only minor changes were observed in contrast to the initial carburized catalysts [37]. A critical difference observed during the phase evolution of each catalyst was the small amount of Fe3O4 (7%) is present in the Na–S-Fe/α-Al2O3 catalyst during the first 10 h ToS of the FTS reaction [37]. However, no Fe oxides (in particular Fe3O4) were observed in the Fe/α-Al2O3 (Fig. 25.12b–d vs f–h) [37]. The presence of Fe3O4 was hypothesized to be due to the higher CO conversion over

D. Kiani

the Na–S-Fe/α-Al2O3 catalyst, which produces higher amounts of H2O and CO2, hence increasing the likelihood of Fe species getting reoxidized during the FTS reaction [37]. Moreover, another important difference noted between the structural evolution of Fe/α-Al2O3 and Na–S-Fe/α-Al2O3 catalysts during FTS reaction was the formation of larger Fe5C2 crystallites in the case of Na–S-Fe/α-Al2O3 catalyst during the first 10–15 hr ToS (Fig. 25.12e vs i) [37]. Hence, one role of the Na- and S-promoters can be to stabilize certain carbidic phases like Fe5C2 over other carbidic phases and to control the carbide crystallite size. Lastly, via in situ Raman analysis (not shown herein) of the carbonous species formed on the two catalysts during FTS, it was concluded that Na–S promotion also affected the type of carbon species formed during FTS by increasing the initial C sp2 content in chainlike carbon structures and by increasing the sixfold cyclic carbon species in catalyst materials with the initial CO carburization [37].

25.7

Limitations of XRD

Up till now, we have discussed the fundamentals and applications of XRD, with a focus on its use for catalysis science research. XRD as a technique also has its limitations, which the reader should keep in mind. The following list briefly describes the limitations, especially as pertaining to XRD’s application in catalyst characterization: (a) Fundamentally, XRD elucidates the average atomic positions of the repeating molecule in the three-dimensional crystalline lattice, making it a bulk characterization technique [16]. Being a bulk characterization technique, XRD’s suitability for catalysis science research is entirely dependent on the nature of the catalyst being used (i.e., crystalline vs amorphous, bulk vs supported, oxide vs metallic, etc.) and the information being sought. Otherwise, heterogeneous catalysis is primarily a surface phenomenon, clearly related to the surface structure and composition of the catalyst and not the bulk structure and composition [16]. That is not to say that XRD should not be used to study catalysts; the point is to keep in mind that XRD is providing bulk structural information, which might not be correlated to the catalyst’s surface structure and composition that ultimately controls its catalytic activity and selectivity. Often in catalysis literature, bulk structure from XRD (mostly ex situ) is used to propose structure-function relationships, which is a problematic practice, and we caution the readers against it. More on bulk vs surface structure and composition of crystalline catalysts can be found in references [16, 39–43]. Curious readers are also directed to reference [44] for a profound commentary on the importance and power of operando catalyst characterization.

25

X-Ray Diffraction (XRD)

533

a)

Z-height adjustment

Insulating kapton cover

25

Mounting plate to commercial XRDs

Gas in X-rays Gas out

ing Slid ate pl e s ba

Thermocouple

Capillary reactor

Infrared heaters

70

0.18

N/A

Normalized counts (a.u.) FF F F c) 3 2 1 Normalized counts (a.u.)

Time-on-stream (h)

60 50 40 30 20 10

Image

0 4

2 1

2 1

F

Measured Calculated d) 100 70 h FF FF F 80

FF FF F

FF FF F

5h

3 h

e)

60 100%

100%

40 20

0

Average crystalline size (nm)

UP 340

b)

0.09

Crystalline Fe fraction (%)

0.00

3D-rendering

Fe3O4 F-Fe5C2

30 25 20 15 10 5 0

0

Na-S 340

30 20 10 0 4

2 1

2 1

30

FF FF F

FF FF F

5h

3 h

80 60

91%

93%

40 20

0 14 16 18 20 22 24 26 28

2q (q) - Mo Kg

2q (q) - Mo Kg

Fig. 25.12 (a) 3D rendering and actual image of an iKey in situ XRD plug flow capillary reactor cell from Cape Catalytix taken with permission from reference. In situ pXRD results of the UP (Fe/α-Al2O3) (b–e) and Na–S 340 (Na–S-Fe/α-Al2O3) (f–i) catalysts during a 72 h Fischer–Tropsch synthesis (FTS). (b, f) Heatmaps of aligned and background-corrected, in situ pXRD patterns, (c, g) measured and the Rietveld method calculated pXRD patterns at
3, 5, and 70 h time-onstream (ToS), (d, h) the crystalline Fe phase composition of the catalyst,

20 15 10 5 0

0

14 16 18 20 22 24 26 28

25

0 10 20 30 40 50 60 70 Time-on-stream (h)

0 15 30 45 6070 Time-on-stream (h)

Q-FTS

40

FF FF F

70 h

Average crystalline size (nm)

50

i)

100

Q-FTS

Normalized counts (a.u.)

Time-on-stream (h)

60

h) 3 2 1

Carb.

g)

Carb.

FF

70

Crystalline Fe fraction (%)

FF

f)

and (e, i) Fe phases’ average crystallite sizes during the 72 h FTS reaction. The measurements before (Carb.) and quasi in situ (Q-FTS) after the FTS reaction run are included for comparison. FTS conditions: 340  C, 10 bar, H2/CO/He ¼ 2.0:1.0:0.33 v/v, 7200 h
1. (Figure (a) adapted with permission from reference [38], Copyright 2016 Elsevier, while figures (b–i) adapted with permission from reference [37] under the Creative Commons license.) The readers are directed to the fullsized, high-resolution figures in reference [37] for visual clarity

534

D. Kiani

(b) If a catalyst sample contains multiple polymorphs and/or compounds, deconvoluting the overlapping peaks can be extremely challenging, especially if the peaks are also broadened due to the factors discussed in previous sections. The overlap makes it extremely difficult to get accurate intensities. Although several analytical approaches have been developed to estimate intensities of overlapping peaks from intensities of non-overlapping peaks, such methods are not reliable for a high degree of overlap [45]. Therefore, complementary structural information from other techniques like Raman, X-ray absorption spectroscopy (XAS), etc., should be included to aid XRD [45]. As discussed in a later section, the modern approach to overcome this issue is to conduct combined complementary characterization with XRD. (c) Catalysts, especially nano-powders, are rarely monodispersed in terms of size and more commonly are polydisperse, exhibiting a range of particle sizes [46]. XRD analysis, especially one utilizing Scherrer’s equation to determine the size, does not account for such polydispersity in sizes [46]. Moreover, other factors, as discussed above, which can also affect the XRD peak broadness, intensities, and positions, should also be kept in mind when analyzing XRD results via Scherrer’s equation to avoid calculating incorrect size values. (d) XRD cannot be used to characterize a sample where crystallite size is below ~3 nm, as it will not be detected by XRD due to its fundamental limit. Conversely, just because XRD does not detect a crystalline phase in a sample does not mean that phase is not present; it can either be absent or be present in coherent scattering domains of smaller than 3 nm. Therefore, intrinsically, XRD analysis is not unambiguous and should be paired with more robust characterization techniques like Raman spectroscopy, which can detect isolated single sites, oligomeric sites, nanoparticles, and large crystalline domains all in one measurement [47]. Recently, a new method, called the pair distribution function (PDF) [48–51], has alleviated this limitation to certain extent. By using extremely short wavelength X-rays and analyzing a wider range of reflection angles, PDF enables quantitative structural analysis of nanomaterials that would appear amorphous under conventional XRD [48].

25.8

Outlook

Unfortunately, it is often seen in the heterogenous catalysis literature that speculative, handwavy structure-function proposals are much more common than pertinent, operando structural characterization of the catalysts that can provide direct structure-function relationships. Therefore, in situ/operando XRD can serve as a critical tool to address this knowledge

gap. Noting that XRD is a bulk characterization technique, it can still be utilized in catalytic applications where the catalytic material is a crystalline solid like bulk metal and mixed metal oxides, micro- and mesoporous crystalline frameworks, etc. Likewise, XRD can be especially useful when the catalytic reaction occurs with a change in the bulk structure of the catalyst, as in a Mars-van Krevelen oxidation mechanism or during a chemical looping (cyclic redox) operation. Recent advancements in XRD have been notable in making variants of this technique that are even more informative and hence useful for catalysis science research, albeit resource demanding. Here, three notable recent developments are highlighted as they overcome some of the limitations of XRD technique summarized in the previous sections: (a) As demonstrated in numerous recent studies by multiple groups, multidimensional temporally and spatially resolved operando XRD mapping can now be conducted at synchrotrons around the world to study the whole catalyst bed or a single catalyst pellet under operation and observe phase evolution and structural dynamics as a function of temperature, gaseous environment, and reaction [50, 52–57]. Synchrotron light sources can produce micro- or nano-focused hard X-rays with high brilliance and tunable energy, which has enabled this advance [58]. By further combining X-ray imaging with computed tomography (CT), a three-dimensional analysis of working catalysts is possible, revealing structural features which may be otherwise undetectable by bulk methods such as XAS, XRD, etc [58]. For example, a very neat study on core@shell structured Cu/ZnO/Al2O3@ZSM-5 catalysts for dimethyl ether (DME) synthesis from syngas utilized in situ μ-X-ray fluorescence (XRF)-CT, μ-XRD-CT, and scanning transmission X-ray microscopy (STXM-CT) simultaneously under model DME synthesis conditions to render 3D distribution of various phases in the working catalyst. As shown in Fig. 25.13, the 3D tomographic map of the catalyst clearly evidences the core@shell structure of catalyst where the CuO/ZnO/ Al2O3 forms the core, while ZSM-5 forms the shell. Special attention was given to study the fate of CuO in the core, as numerous speculative proposals have been put forth in the prior literature regarding the role of this component. It was found that under oxidizing conditions, the Cu species were found to be primarily in CuO phase (i.e., Cu2+) with a minor fraction of reduced phases Cu2O (i.e., Cu1+) present. However, as this catalyst is activated in 5%H2 at 250  C for 6–8 h, the CuO core reduced to a mixture of Cu2O and metallic Cu0, where the Cu2O phase was localized at the upper 40% of the core volume. This Cu2O phase was found to be irreducible and persisted throughout the reaction [58]. Upon switching from activation to reaction conditions, the catalyst core evolved further, where the majority of the core still comprised of

25

X-Ray Diffraction (XRD)

535

25 Cu2+

core (2q 18.1°)

Zeolite shell (2q 14.1°) 25 μm

Other Cu phases (2q 19.8−20.1°)

Cu+ layers (2q 17°) Cu2+ core (2q 18.1°)

Cu0 body (2q 23°)

25 μm

25 μm

Under oxidizing condition

Activated under reducing condition

Fig. 25.13 Core@shell particle under oxidizing conditions, with μ-XRD-CT rendering of core and zeolite shell occupying distinct 3D space. The inset below shows the chemical state and spatial distribution of the CuO core under oxidizing conditions where the CuO core constitutes 95 vol %, i.e., Cu2+, and reduced fractions constitute 5 vol % of the core. When this catalyst was activated for ~6–8 h at 250  C in a flow

of 5% H2, the catalyst core changed drastically. CuO was seen to reduce to metallic Cu0, except for some small Cu2O clusters (Cu1+) within the catalyst core which proved irreducible. Such Cu2O clusters were found to be localized at the top 40% of the catalyst core volume [58]. (Figure adapted with permission from reference [58]. Copyright 2017 American Chemical Society)

Cu0, but a minor population of poorly crystalline nanoparticles of partially oxidized phases like Cu2O and an unidentified CuOx formed during reaction. Summarily, the spatially resolved studies uncovered the presence of metastable Cu2O during activation through interaction with the zeolite shell and a significant increase in the presence of mixed Cu oxidation states under model reaction conditions, indicating the possible role of Cu2O in DME synthesis [58]. (b) Another critical advancement is the development and application of grazing incidence angle XRD or GIXD [59–63], also sometimes called surface-XRD or SXRD [64–66]. SXRD/GIXD differs from regular XRD in that it can be used to determine the average positions of atoms at well-defined surfaces, whereas XRD determines the average position of atoms throughout the bulk of a crystalline solid [65]. However, the use of SXRD/GIXD is limited to adsorbates and ultrathin layer samples of single-crystal quality – hence not applicable to polycrystalline powdered catalysts [16]. Note that SXRD/GIXD requires atomically flat samples, a near-perfect alignment of sample and X-ray beam, and, in practice, a synchrotron source when a cell

containing a catalyst in a reactive atmosphere is to be used [16]. For example, a simple study utilizing GIXD to characterize Ti-oxide film on metallic Ti substrate formed via anodization of titanium metal enabled composition profiling of the oxide layer [67]. The depicted GIXD in Fig. 25.14 yields information about the proportions of the oxide phases at various depths in the thin film. The left panel clearly shows that after mild anodization at 5 V for 15 min, a very thin layer of oxide is expected to form on the Ti substrate. Conventional XRD of the resulting film shows no peak besides the two intense peaks from the Ti substrate. However, a GIXD at 6 angle of incidence clearly shows the formation of Ti2O and Ti2O3 on the surface of the Ti substrate that was not detected via conventional XRD. Likewise, after harsher anodization (100 V for 2 h) when a thicker oxide layer is expected, GIXD of the strongly anodized film at various incidence angles (0.2–10 ) is shown in the right panel [67]. The GIXD as a function of incidence angle enables compositional depth profiling of the resulting film. By contrast, conventional XRD (not shown in figure but similar to the spectrum at α ¼ 10 ) only indicated the simultaneous presence of rutile and

536

D. Kiani

cannot provide all necessary chemical and structural information about the working catalyst [44]. However, a carefully designed combination of two or more spectroscopic techniques, which provides complementary information on the molecular structure and composition of a single sample of catalyst at the same time and under optimal spectroscopic and catalytic conditions, can overcome this inherent limitation [44]. For example, as noted in the example discussed earlier, a core@shell structed Cu/ZnO/Al2O3@ZSM-5 catalyst for DME synthesis from syngas was studied via combined in situ μ-XRF-CT, μ-XRD-CT, and STXM-CT under model DME synthesis conditions (schematically depicted in Fig. 25.15) [58]. The simultaneous use of three

b)

Low angle (6°) incidence X-ray diffraction

Intensity (a.u.)

Intensity (a.u.)

Ti/ TiOx Eanod = 100V tanod = 120min

Ti Conventional XRD (Bragg-Brentano geometry)

Ti

TiO0.5

TiO0.5

TiO1.5

Ti

TiO1.5 0 30

GIXD as a function of incidence angle

40

35

(101) A (110) R

Ti

Ti A R

Ti

R

A - anatase R - rutile

A

10 4.0 2.5 1.5 0.8 0.5 0.2

45

2q (degrees)

0 20

Fig. 25.14 (a) Conventional and GIXD patterns of Ti anodized at 5 V for 15 min. (b) GIXD characterization of an oxide layer obtained by anodization of Ti at 100 V for 2 h. Anatase, rutile, and Ti substrate reflections are indicated. Note that the diffraction signals from the Fig. 25.15 A schematic overview of the combined structural characterization conducted on the core@shell catalysts for DME synthesis. (Figure adapted with permission from reference [58]. Copyright 2017 American Chemical Society)

A Ti R

25

30

35 40 45 2q (degrees)

50

55

gl e (d of in eg c re ide es n ) ce

a)

An

anatase phases of TiO2 [67]. Hence, this example serves to show that GIXD can be a powerful tool in compositional characterization of suitable catalyst samples not possible with conventional XRD. (c) Lastly, another key development has been the combinatorial approach that incorporates multiple complementary spectroscopies with XRD to yield multimodal characterization tools. Examples of such multimodal approaches are Raman-XRD [37], XAS-XRD [68–70], IR-XRD [70–73], etc. As recognized in the literature [44], combination of multiple techniques into a single operando experiment is especially beneficial because every spectroscopic technique has its own sensitivity and limitations. Thus, one single characterization technique

substrate get weaker and disappear by lowering α. Shallow angles give information about layers closer to the sample surface. (Figures adapted from reference [67] with permission under the SCIRP open access license)

Multimodal characterization techniques

Working catalyst

Complementary structural information XRF spectrum (Cu, Zn)

X-ray

Per pixel

Absorption image XRD pattern (12−38° 2q)

z q x

x: 81 steps z: 8 steps 101 rotations

Tomogram

25

X-Ray Diffraction (XRD)

complimentary techniques in this multimodal approach enabled the authors of the study to map the crystalline structure and elemental distribution of the core@shell catalyst grain noninvasively, with effective spatial resolution of 2 μm. These complementary techniques allowed them to unravel structural features of both the core and zeolite shell in a single measurement, which could not have been possible otherwise. Importantly, the same catalyst volume was measured throughout in these structural characterization experiments, which is crucial to allow direct and accurate comparison of the catalyst structure under different chemical environments like oxidation, activation, and reaction [58]. Summarily, the key advantage of using XRD for catalysis science research is the technique’s maturity, given its centurylong history and widespread use in the broader scientific community. This means that the possibilities and limitations and the advantages and disadvantages of this technique are very well understood, especially in comparison to some more advanced but recent characterization techniques developed in the past two decades. Therefore, as the need and urgency to study catalysts under operating conditions continues to grow, the capabilities and use of operando XRD for catalysis research is expected to grow as well, and XRD is expected to become ever more important to structural evaluation of catalysts under reaction conditions. Acknowledgments D.K. would like to acknowledge the use of several open sources including Wikipedia and Chemistry LibreTexts for the preparation of this chapter. Teaching notes from the course “Characterization of Catalysts and Surfaces” by Professor Jeroen A. van Bokhoven, available online [74], were also used for referencing fundamental topics pertinent to XRD. Disclaimer: the opinions expressed within this chapter are solely the author’s and do not reflect the opinions, beliefs, or policies of the author’s affiliation.

References 1. Cullity, B.D.: Elements of X-Ray Diffraction, p. 547. AddisonWesley Publishing Company, Inc., USA (1956) 2. Coppens, P., Penner-Hahn, J.: Introduction: X-rays in chemistry. Chem. Rev. 101(6), 1567–1568 (2001) 3. Authier, A.: Early Days of X-Ray Crystallography, 1st edn. Oxford University Press, Oxford (2013) 4. Brooks-Bartlett, J.C., Garman, E.F.: The nobel science: one hundred years of crystallography. Interdiscip. Sci. Rev. 40(3), 244–264 (2015) 5. Eckert, M.: Disputed discovery: the beginnings of X-ray diffraction in crystals in 1912 and its repercussions. Acta Crystallogr. A. 68(1), 30–39 (2012) 6. Meisburger, S.P., Thomas, W.C., Watkins, M.B., Ando, N.: X-ray scattering studies of protein structural dynamics. Chem. Rev. 117(12), 7615–7672 (2017) 7. Egli, M.: Diffraction techniques in structural biology. Curr. Protoc. Nucleic Acid Chem. 41(1), 7.13.1–7.13.35 (2010)

537 8. Parker, M.W.: Protein structure from x-ray diffraction. J. Biol. Phys. 29(4), 341–362 (2003) 9. Shi, Y.: A glimpse of structural biology through X-ray crystallography. Cell. 159(5), 995–1014 (2014) 10. Garman, E.F.: Developments in X-ray crystallographic structure determination of biological macromolecules. Science. 343(6175), 1102–1108 (2014) 11. Warren, B.E.: X-ray diffraction study of the structure of glass. Chem. Rev. 26(2), 237–255 (1940) 12. Kahle, M., Kleber, M., Jahn, R.: Review of XRD-based quantitative analyses of clay minerals in soils: the suitability of mineral intensity factors. Geoderma. 109(3), 191–205 (2002) 13. Bish, D.L., Blake, D.F., Vaniman, D.T., Chipera, S.J., Morris, R.V., Ming, D.W., Treiman, A.H., Sarrazin, P., Morrison, S.M., Downs, R.T., Achilles, C.N., Yen, A.S., Bristow, T.F., Crisp, J.A., Morookian, J.M., Farmer, J.D., Rampe, E.B., Stolper, E.M., Spanovich, N.: X-ray diffraction results from mars science laboratory: mineralogy of rocknest at gale crater. Science. 341(6153), 1238932 (2013) 14. Hargreaves*, J.S.J.: Powder X-ray diffraction and heterogeneous catalysis. Crystallogr. Rev. 11(1), 21–34 (2005) 15. Newton, M.A., van Beek, W.: Combining synchrotron-based X-ray techniques with vibrational spectroscopies for the in situ study of heterogeneous catalysts: a view from a bridge. Chem. Soc. Rev. 39(12), 4845–4863 (2010) 16. Schlögl, R.: Chapter 5: X-ray diffraction: a basic tool for characterization of solid catalysts in the working state. In: Advances in Catalysis, vol. 52, pp. 273–338. Academic Press, Amsterdam, Netherland (2009) 17. Fetisov, G.V.: X-ray diffraction methods for structural diagnostics of materials: progress and achievements. Physics-Uspekhi. 63(1), 2–32 (2020) 18. Jauncey, G.E.M.: Atomic structure as revealed by crystal reflection of X-rays. Chem. Rev. 5(4), 437–449 (1929) 19. Bragg, W.H., Bragg, W.L.: The reflection of X-rays by crystals. Proc. R. Soc. London, Ser. A, Containing Papers of a Mathematical and Physical Character. 88(605), 428–438 (1913) 20. Bragg, W.L., Bragg, W.H.: The structure of some crystals as indicated by their diffraction of X-rays. Proc. R. Soc. London, Ser. A, Containing Papers of a Mathematical and Physical Character. 89(610), 248–277 (1913) 21. Strinati, G.C.: Hartree and Hartree–Fock methods in electronic structure. In: Bassani, F., Liedl, G.L., Wyder, P. (eds.) Encyclopedia of Condensed Matter Physics, pp. 311–318. Elsevier, Oxford (2005) 22. Souvignier, B.: International tables for crystallography. In: Aroyo, M.I. (ed.) 1.3.4. Classification of Space Groups [Online]. International Union of Crystallography, pp. 31–41. John Wiley and Sons Inc. & IUCR (2016). https://onlinelibrary.wiley.com/doi/10.1107/ 97809553602060000921/full 23. Kittel, C.: Introduction to Solid State Physics, 8th edn. Wiley, CA, USA (2005) 24. Aroyo, M.I., Müller, U., Wondratschek, H.: International Tables for Crystallography, vol. A1. Springer & IUCR, Dordrecht (2006). https://doi.org/10.1107/97809553602060000101 25. Ayyappan, S., Gnanaprakash, G., Panneerselvam, G., Antony, M.P., Philip, J.: Effect of surfactant monolayer on reduction of Fe3O4 nanoparticles under vacuum. J. Phys. Chem. C. 112(47), 18376–18383 (2008) 26. Rietveld, H.: A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 2(2), 65–71 (1969) 27. Hargreaves, J.S.J.: Some considerations related to the use of the Scherrer equation in powder X-ray diffraction as applied to heterogeneous catalysts. Catal. Struct. React. 2(1–4), 33–37 (2016) 28. Langford, J.I., Wilson, A.J.C.: Scherrer after sixty years: a survey and some new results in the determination of crystallite size. J. Appl. Crystallogr. 11(2), 102–113 (1978)

25

538 29. Borchert, H., Shevchenko, E.V., Robert, A., Mekis, I., Kornowski, A., Grübel, G., Weller, H.: Determination of nanocrystal sizes: a comparison of TEM, SAXS, and XRD studies of highly monodisperse CoPt3 particles. Langmuir. 21(5), 1931–1936 (2005) 30. Shiraishi, M., Inagaki, M.: Chapter 10: X-ray diffraction methods to study crystallite size and lattice constants of carbon materials. In: Yasuda, E.-i., Inagaki, M., Kaneko, K., Endo, M., Oya, A., Tanabe, Y. (eds.) Carbon Alloys, pp. 161–173. Elsevier Science, Oxford (2003) 31. Chattot, R., Asset, T., Bordet, P., Drnec, J., Dubau, L., Maillard, F.: Beyond strain and ligand effects: microstrain-induced enhancement of the oxygen reduction reaction kinetics on various PtNi/C nanostructures. ACS Catal. 7(1), 398–408 (2017) 32. Hegele, P., von Kobylinski, J., Hitzler, L., Krempaszky, C., Werner, E.: In-situ XRD study of phase transformation kinetics in a co-Cr-Walloy manufactured by laser powder-bed fusion. Crystals. 11(2), 176 (2021) 33. Barnes, P., Jacques, S., Vickers, M.: Mode (3): Debye-Scherrer. http://pd.chem.ucl.ac.uk/pdnn/inst2/debyes.htm 34. Zhu, X., Gao, Y., Wang, X., Haribal, V., Liu, J., Neal, L.M., Bao, Z., Wu, Z., Wang, H., Li, F.: A tailored multi-functional catalyst for ultra-efficient styrene production under a cyclic redox scheme. Nat. Commun. 12(1), 1329 (2021) 35. Wang, X., Rodriguez, J.A., Hanson, J.C., Gamarra, D., MartínezArias, A., Fernández-García, M.: In situ studies of the active sites for the water gas shift reaction over Cu
CeO2 catalysts: complex interaction between metallic copper and oxygen vacancies of ceria. J. Phys. Chem. B. 110(1), 428–434 (2006) 36. Chupas, P.J., Ciraolo, M.F., Hanson, J.C., Grey, C.P.: In situ X-ray diffraction and solid-state NMR study of the fluorination of γ-Al2O3 with HCF2Cl. J. Am. Chem. Soc. 123(8), 1694–1702 (2001) 37. Paalanen, P.P., van Vreeswijk, S.H., Weckhuysen, B.M.: Combined in situ X-ray powder diffractometry/Raman spectroscopy of iron carbide and carbon species evolution in Fe(
Na–S)/α-Al2O3 catalysts during Fischer–Tropsch synthesis. ACS Catal. 10(17), 9837–9855 (2020) 38. Fischer, N., Claeys, M.: Phase changes studied under in situ conditions—a novel cell. Catal. Today. 275, 149–154 (2016) 39. Wachs, I.E., Routray, K.: Catalysis science of bulk mixed oxides. ACS Catal. 2(6), 1235–1246 (2012) 40. Wachs, I.E., Jehng, J.-M., Ueda, W.: Determination of the chemical nature of active surface sites present on bulk mixed metal oxide catalysts. J. Phys. Chem. B. 109(6), 2275–2284 (2005) 41. Phivilay, S.P., Roberts, C.A., Puretzky, A.A., Domen, K., Wachs, I.E.: Fundamental bulk/surface structure
photoactivity relationships of supported {Rh(2
y)Cr(y)O3}/GaN photocatalysts. J. Phys. Chem. Lett. 4, 3719–3724 (2013) 42. Reece, C., Madix, R.J.: Moving from fundamental knowledge of kinetics and mechanisms on surfaces to prediction of catalyst performance in reactors. ACS Catal. 11, 3048–3066 (2021) 43. Uskokovic, V.: X-ray photoelectron and ion scattering spectroscopic surface analyses of amorphous and crystalline calcium phosphate nanoparticles with different chemical histories. Phys. Chem. Chem. Phys. 22(10), 5531–5547 (2020) 44. Weckhuysen, B.M.: Determining the active site in a catalytic process: operando spectroscopy is more than a buzzword. Phys. Chem. Chem. Phys. 5(20), 4351–4360 (2003) 45. Li, J., Sun, J.: Application of X-ray diffraction and electron crystallography for solving complex structure problems. Acc. Chem. Res. 50(11), 2737–2745 (2017) 46. Jensen, H., Pedersen, J.H., J⊘rgensen, J.E., Pedersen, J.S., Joensen, K.D., Iversen, S.B., S⊘gaard, E.G.: Determination of size distributions in nanosized powders by TEM, XRD, and SAXS. J. Exp. Nanosci. 1(3), 355–373 (2006) 47. Kiani, D., Sourav, S., Wachs, I.E., Baltrusaitis, J.: Synthesis and molecular structure of model silica-supported tungsten oxide

D. Kiani catalysts for oxidative coupling of methane (OCM). Cat. Sci. Technol. 10, 3334–3345 (2020) 48. Billinge, S.J.L.: The rise of the X-ray atomic pair distribution function method: a series of fortunate events. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 377(2147), 20180413 (2019) 49. Du, P., Kokhan, O., Chapman, K.W., Chupas, P.J., Tiede, D.M.: Elucidating the domain structure of the cobalt oxide water splitting catalyst by X-ray pair distribution function analysis. J. Am. Chem. Soc. 134(27), 11096–11099 (2012) 50. Sottmann, J., Di Michiel, M., Fjellvåg, H., Malavasi, L., Margadonna, S., Vajeeston, P., Vaughan, G.B.M., Wragg, D.S.: Chemical structures of specific sodium ion battery components determined by operando pair distribution function and X-ray diffraction computed tomography. Angew. Chem. Int. Ed. 56(38), 11385–11389 (2017) 51. Jacques, S.D.M., Di Michiel, M., Kimber, S.A.J., Yang, X., Cernik, R.J., Beale, A.M., Billinge, S.J.L.: Pair distribution function computed tomography. Nat. Commun. 4(1), 2536 (2013) 52. Vamvakeros, A., Jacques, S.D.M., Di Michiel, M., Matras, D., Middelkoop, V., Ismagilov, I.Z., Matus, E.V., Kuznetsov, V.V., Drnec, J., Senecal, P., Beale, A.M.: 5D operando tomographic diffraction imaging of a catalyst bed. Nat. Commun. 9(1), 4751 (2018) 53. Vamvakeros, A.: Operando Chemical Tomography of Packed Bed and Membrane Reactors for Methane Processing (2017) 54. Vamvakeros, A., Matras, D., Jacques, S.D.M., di Michiel, M., Price, S.W.T., Senecal, P., Aran, M.A., Middelkoop, V., Stenning, G.B.G., Mosselmans, J.F.W., Ismagilov, I.Z., Beale, A.M.: Real-time multilength scale chemical tomography of fixed bed reactors during the oxidative coupling of methane reaction. J. Catal. 386, 39–52 (2020) 55. Becher, J., Sanchez, D.F., Doronkin, D.E., Zengel, D., Meira, D.M., Pascarelli, S., Grunwaldt, J.-D., Sheppard, T.L.: Chemical gradients in automotive Cu-SSZ-13 catalysts for NOx removal revealed by operando X-ray spectrotomography. Nat. Catal. 4(1), 46–53 (2021) 56. Matras, D., Jacques, S.D.M., Poulston, S., Grosjean, N., Estruch Bosch, C., Rollins, B., Wright, J., Di Michiel, M., Vamvakeros, A., Cernik, R.J., Beale, A.M.: Operando and postreaction diffraction imaging of the La–Sr/CaO catalyst in the oxidative coupling of methane reaction. J. Phys. Chem. C. 123(3), 1751–1760 (2019) 57. Matras, D., Vamvakeros, A., Jacques, S.D.M., di Michiel, M., Middelkoop, V., Ismagilov, I.Z., Matus, E.V., Kuznetsov, V.V., Cernik, R.J., Beale, A.M.: Multi-length scale 5D diffraction imaging of Ni–Pd/CeO2–ZrO2/Al2O3 catalyst during partial oxidation of methane. J. Mater. Chem. A. 9(18), 11331–11346 (2021) 58. Sheppard, T.L., Price, S.W.T., Benzi, F., Baier, S., Klumpp, M., Dittmeyer, R., Schwieger, W., Grunwaldt, J.-D.: In situ multimodal 3D chemical imaging of a hierarchically structured Core@Shell catalyst. J. Am. Chem. Soc. 139(23), 7855–7863 (2017) 59. Landree, E., Marks, L.D., Zschack, P., Gilmore, C.J.: Structure of the TiO2
x(100)-13 surface by direct methods. Surf. Sci. 408(1), 300–309 (1998) 60. Singh, M.A., Groves, M.N.: Depth profiling of polymer films with grazing-incidence small-angle X-ray scattering. Acta Crystallogr. A. 65(Pt 3), 190–201 (2009) 61. Scherzer, M., Girgsdies, F., Stotz, E., Willinger, M.-G., Frei, E., Schlögl, R., Pietsch, U., Lunkenbein, T.: Electrochemical surface oxidation of copper studied by in situ grazing incidence X-ray diffraction. J. Phys. Chem. C. 123(21), 13253–13262 (2019) 62. Bergmann, A., Roldan Cuenya, B.: Operando insights into nanoparticle transformations during catalysis. ACS Catal. 9(11), 10020–10043 (2019) 63. Kim, T., Kumar, R.E., Brock, J.A., Fullerton, E.E., Fenning, D.P.: How strain alters CO2 electroreduction on model Cu(001) surfaces. ACS Catal. 11, 6662–6671 (2021)

25

X-Ray Diffraction (XRD)

64. Vlieg, E., Van Der Veen, J.F., Gurman, S.J., Norris, C., Macdonald, J.E.: X-ray diffraction from rough, relaxed and reconstructed surfaces. Surf. Sci. 210(3), 301–321 (1989) 65. Harlow, G.S., Lundgren, E., Escudero-Escribano, M.: Recent advances in surface x-ray diffraction and the potential for determining structure-sensitivity relations in single-crystal electrocatalysis. Curr. Opin. Electrochem. 23, 162–173 (2020) 66. Robinson, I.K., Tweet, D.J.: Surface X-ray diffraction. Rep. Prog. Phys. 55(5), 599–651 (1992) 67. Bouroushian, M., Kosanovic, T.: Characterization of thin films by low incidence X-ray diffraction. Crystal Structure Theory and Applications. 1(3), 35–39 (2012) 68. Ehrlich, S.N., Hanson, J.C., Lopez Camara, A., Barrio, L., Estrella, M., Zhou, G., Si, R., Khalid, S., Wang, Q.: Combined XRD and XAS. Nucl. Instrum. Methods Phys. Res., Sect. A. 649(1), 213–215 (2011) 69. Beale, A.M., Sankar, G.: In situ characterization of iron phosphate and bismuth molybdate catalysts prepared by hydrothermal methods: an EDXRD and combined XRD/XAS study. Nucl. Instrum. Methods Phys. Res., Sect. B. 199, 504–508 (2003) 70. Marchionni, V., Nachtegaal, M., Ferri, D.: Influence of CO on dry CH4 oxidation on Pd/Al2O3 by operando spectroscopy: a multitechnique modulated excitation study. ACS Catal. 10(8), 4791–4804 (2020) 71. Hinokuma, S., Wiker, G., Suganuma, T., Bansode, A., Stoian, D., Huertas, S.C., Molina, S., Shafir, A., Rønning, M., van Beek, W., Urakawa, A.: Versatile IR spectroscopy combined with synchrotron XAS–XRD: chemical, electronic, and structural insights during thermal treatment of MOF materials. Eur. J. Inorg. Chem. 2018(17), 1847–1853 (2018) 72. Marchionni, V., Kambolis, A., Nachtegaal, M., Kröcher, O., Ferri, D.: High energy X-ray diffraction and IR spectroscopy of Pt/Al2O3 during CO oxidation in a novel catalytic reactor cell. Catal. Struct. React. 3(1–2), 71–78 (2017)

539 73. Newton, M.A., Di Michiel, M., Kubacka, A., Fernández-García, M.: Combining time-resolved hard X-ray diffraction and diffuse reflectance infrared spectroscopy to illuminate CO dissociation and transient carbon storage by supported Pd nanoparticles during CO/NO cycling. J. Am. Chem. Soc. 132(13), 4540–4541 (2010) 74. Pinar, A.B., Bokhoven, J.A.: v. Fundamentals and applications of X-ray diffraction. Applications in catalysts characterization. https:// ethz.ch/content/dam/ethz/special-interest/chab/icb/van-bokhovengroup-dam/coursework/Characterization-Techniques/2019/XRD_ lecture_AnaBPinar_2019_for_students1.pdf

Daniyal Kiani received B.S. chemistry from Georgetown University and Ph.D. chemical engineering from Lehigh University. He has worked on numerous catalysis applications including upgradation of methane (OCM, MDA), dry methane reformation via DBD-plasma activation, ethylene dimerization, ethanol coupling, methanol oxidation, and selective catalytic reduction of NOx. Daniyal now works as a technical specialist in the chemistry and kinetics group at Cummins Emission Solutions.

25

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

26

Pablo Beato, Lars Fahl Lundegaard, Stian Svelle, and David Stephen Wragg

Contents Starting from the adsorption of simple molecules, over postmortem catalyst analysis, to the development of advanced in situ and operando diffraction techniques, the chapter will provide an overview of the state of the art.

26.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 541

26.2

MTH Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542

26.3

Diffraction for Zeolite Characterization . . . . . . . . . . . . . . . . . 542

26.4

Organic Molecules Adsorbed in Zeolites . . . . . . . . . . . . . . . . 544

26.5

Catalysts “Postmortem” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547

Keywords

26.6

In Situ Studies on Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549

26.7

Operando Catalytic Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551

Diffraction · Zeolites · MTH · Catalysis · In situ · Operando · Time and space resolved

26.8

Time- and Space-Resolved Operando Studies . . . . . . . . . . 552

26.9

Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 556

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557

Abstract

Heterogeneous catalytic reactions typically occur at the solid-gas interface, which represents a challenge for many characterization methods and often limits structural analysis to surface-sensitive techniques. However, in the case of zeolites, the large microporous surface area is intimately linked to the three-dimensional bulk structure, which leads to the phenomenon called shape selectivity. The crystallography of zeolites can therefore be exploited to obtain a wealth of information on the nature and precise location of the interaction between reactant molecules and the catalyst surface. This chapter focuses on how crystallography can be used as a unique tool for studying the conversion of methanol to hydrocarbons in zeolites. P. Beato (*) · L. F. Lundegaard Topsoe A/S, Lyngby, Denmark e-mail: [email protected]; lafl@topsoe.com S. Svelle · D. S. Wragg Center for Materials Science and Nanotechnology (SMN), Department of Chemistry, University of Oslo, Oslo, Norway e-mail: [email protected]; [email protected]; [email protected]

26.1

Introduction

The tool of choice for basic zeolite characterization is powder X-ray diffraction (PXRD). The powder pattern is the fingerprint of the crystal structure, a simple and unambiguous proof of identity. While this kind of basic identification remains the biggest single application of diffraction in zeolite science, the technique is not limited by it. The diffraction pattern from a single crystal or crystalline powder contains not only the full details of the three-dimensional configuration of atoms in the material but also microstructural information like degree of crystallinity, crystallite size, and strain [1, 2]. We can also use the diffraction pattern to study defects and disorder in the lattice. For a family of crystalline materials with properties so intimately linked to their three-dimensional structures as the zeolites, diffraction is an exceptionally rich source of data. This chapter will concentrate mainly on how diffractionbased methods have been used to study methanol-to-hydrocarbon (MTH) conversion in zeolites but also gives a general introduction to the use of crystallography in zeolite catalysis beyond structure identification. Furthermore, although we make a few references to studies using neutron and electron diffraction, we will concentrate principally on X-ray diffraction.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_26

541

542

26.2

P. Beato et al.

MTH Conversion

MTH conversion is a key technology for the production of synthetic fuels, light alkenes for polymer production, or aromatics for chemicals [3]. Methanol can in principle be produced from virtually any carbon-rich feedstock: natural gas, liquid petroleum, coal, biomass, or even CO2. Hydrogenation of captured CO2 using renewable hydrogen to methanol [4], which can be further processed with MTH technology, is envisaged to play an important role in carbon capture and utilization. The development of small-scale gas conversion and MTH technology has recently received increased attention due to the World Bank’s “Zero Routine Flaring by 2030” initiative [5]. The commercial operational history of MTH technologies starts with the methanol-to-gasoline (MTG) plant started in New Zealand in 1985. This plant was based on Mobil’s MTG process using the ZSM-5 zeolite catalyst. Topose A/S has developed the Topose integrated gasoline synthesis (TIGAS), in which methanol, dimethyl ether (DME), and gasoline are synthesized in a single loop, eliminating the intermittent storage of methanol. TIGAS-based technology was operated in a 25 barrels per day “Wood-2-Gasoline” pilot at the Gas Technology Institute in Chicago from 2013 to 2014 [6]. At present, the world’s only natural gas-to-gasoline plant uses TIGAS technology and is in operation in Turkmenistan. The full capacity of this plant is 15,500 barrels of gasoline per day [7]. The conversion of methanol to light alkenes (ethene and propene) relies primarily on SAPO-34-based catalysts. The UOP (now Honeywell UOP) and Norsk Hydro (now Ineos) methanol-to-olefin (MTO) technology is operated in several plants worldwide. For example, the Jiangsu Sailboat Petrochemical Company operates an MTO unit capable of producing 833,000 metric tons annually, based on Honeywell UOP’s “Advanced MTO Process,” which combines the UOP/Hydro MTO process and the Total/UOP Olefin Cracking Process [8]. The Dalian MTO (DMTO) technology is also in widespread commercial operation, and several generations of the DMTO technology exist [9]. Finally, Lurgi’s MTP ® process (now offered by Air Liquide) is also based on ZSM-5, but the catalyst and operating conditions are optimized toward propene [10]. Several plants are in operation in China. A common challenge for all these technologies is (reversible) deactivation by coke formation. Commercial SAPO-34based MTO reactors use complex fluid bed and continuous regeneration technology to deal with coke formation. ZSM-5based plants have substantially longer cycle lifetimes before they are deactivated by coke, allowing the use of several fixed bed reactors arranged in parallel. Even so, the catalyst beds require periodic regeneration. Both catalysts also suffer from longer-term, permanent deactivation due to structural

changes. The motivation for most studies of zeolite catalyst structure during and after operation (especially the in situ and operando investigations) is a desire to understand these deactivation phenomena and model the reactor performance under industrial conditions in order to bring about process improvements and cost savings. In Sects. 26.5, 26.7, and 26.8, we will illustrate how diffraction-based studies have recently made a significant contribution to this effort.

26.3

Diffraction for Zeolite Characterization

PXRD is the first choice tool for zeolite structure identification, but a diffraction pattern also provides a wealth of other information on the crystal structure [11]. We can measure three simple things from a diffraction pattern; each of them provides different details about the sample: • Peak positions: unit cell dimensions, d-spacings of crystal lattice planes, lattice plane identities (Miller indices), and systematic absences (for lattice centering and space group determination) • Peak intensities: quantitative analysis, atom types, occupancies, positions and thermal motion, and crystallite orientation and texture • Peak widths: Crystallite size, strain, shape, and stacking defects From this information, we can solve and refine the structures of novel materials and understand the structure of known materials in great detail. A range of tools have been developed as the science of crystallography has advanced. In structure solution, model building to simulate observed diffraction data was the first method applied. Mathematical and statistical methods were then developed to solve the “phase problem,” i.e., that although it is easy to measure the intensities of the waves diffracted from the crystal(s) in reciprocal (diffraction) space, their phases cannot be measured. Patterson maps [12] and atom substitution [13] offered initial but unreliable mathematical routes to the phases (Patterson maps work well with structures containing heavy atoms but are less effective for organic structures and zeolites). With the advent of modern computing, “direct methods,” in which many sets of randomly generated phases are tested until one which can reasonably reproduce the electron density map is found, became practical [14]. Direct methods have been used to solve the structures of many zeolites from both singlecrystal and powder diffraction data [15]. Direct methods use mathematical relationships between certain classes of reflections in reciprocal space to help solve the phase problem, but the structure in real space can also be used to help solve structures. The bond lengths and angles in zeolites can be predicted within fairly narrow ranges [16, 17], and this makes

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

them ideal cases for structure solution with computational structure prediction based on energy minimization by simulated annealing or a combination of real and reciprocal space fitting, with programs like DLS-76/FOCUS [18], ZEFSAII [19], and XLENS [20]. In the last 20 years, with greater computing power available, dual-space methods such as charge flipping [21] and intrinsic phasing [22] have been developed and proved highly effective for solving zeolite structures from powder diffraction data, especially when combined with electron diffraction data [23]. An unusual approach to solving the phase problem, which was tested on zeolites, is using modulation-enhanced in situ X-ray diffraction data [24]. The method used data collected during adsorption and desorption of xenon from TS-1 (MFI framework). The data were processed using the variation in Xe content to obtain the phases of the framework, the structure of which does not vary significantly with the changes in channel occupancy. Once the basic framework structure is determined, structural parameters like lattice parameters, atom positions, thermal factors, and occupancies are refined by the method of least squares [25, 26], and methods such as difference Fourier mapping and simulated annealing are applied to discover and place missing elements of the structure like cations or organic templates (see below) [27]. Most of the data used in structure solution and refinement comes from the diffraction peak positions and intensities, and the peak shapes also contain structural information. Scherrer’s method for obtaining crystallite size from diffraction peak widths is almost as old as the technique of X-ray diffraction [28], and strain analysis is also well established [29]. A key point in this analysis is that the instrumental peak shape must be corrected for, so that only sample broadening contributions are used in the size and strain calculations. With modern Rietveld software, it is quite easy to extract size and strain information from powder diffraction data using the fundamental parameters approach to model the underlying instrumental peak shape [30–32]. Models for obtaining 3D crystallite shape from PXRD data have also been established [33–35], and the effect of crystal structure defects on peak shapes has also been extensively studied. The study of stacking faults using diffraction data has been particularly significant in zeolite science as several industrially important zeolite frameworks notably zeolite-beta (BEA) [36] and chabazite (CHA) [37, 38]) have stacking faulted structures. The recursive probabilistic approach to modeling diffraction from this kind of disorder was first developed in the DIFFAX program by Treacy et al. [39]. A database of stacking faulted zeolite intergrowths can be found in the IZA zeolite database [40]. The first use of X-ray diffraction to identify zeolites was reported by Leonard in 1927 [41], and the determination of crystal structures followed soon after in work by Taylor [42],

543

Bragg [43], and Pauling [44, 45] using the Laue and oscillation methods to study the crystals. As zeolites developed into one of the most important groups of industrial catalysts throughout the second half of the twentieth century, diffraction was used extensively in identification and structure solution. Powder diffraction has proved especially important in zeolite structure solution due to the difficulty of obtaining large, high-quality single crystals from hydrothermal synthesis [46]. Two of the most important zeolite frameworks for methanol conversion were discovered during this period, and their structures determined from X-ray diffraction data. Materials with the ZSM-5 framework were first reported in 1978, and a structure was proposed the same year by Kokotailo and co-workers who combined powder and single-crystal diffraction with model building [47]. The structure was confirmed by a single-crystal study in 1981 [48], which also begins to detail the exciting, shape-related properties of ZSM-5 in sorption and catalysis. The structure of SAPO-34 was determined in 1984 [49] by comparison of the PXRD pattern of the new material to the structure of its natural analogue chabazite (solved by Dent and Smith in 1958 [50]). When we study zeolites in catalysis, we are interested in both the framework and the material inside the pores and channels. The first examples of non-framework material in zeolite crystal structures come from the metal cations found inside the cages of natural aluminosilicates, balancing the negative charge created when Si4+ is replaced by Al3+ in the framework. The crystal structure of zeolite-A including the position of the Na+ cation was reported by Reed and Breck as early as 1956 [51], immediately after the first report of its synthesis [52]. They created a model of the framework and applied this to help them determine the positions of the cations in Fourier projections of the diffraction data, a method that (with slight modernization) has since been applied to many similar problems in zeolite structure determination. In 1971 a single-crystal study by Grammlich and Meier showed that the water molecules in the large cage of zeolite-A form highly ordered clusters [53]. The development of the Rietveld method [26] for fitting crystal structure models to powder diffraction data (from 1969 onward) accelerated the development of zeolite structure studies. As more exotic methods came into use in zeolite synthesis, crystallographers became interested in the positions of organic template molecules inside the new frameworks, with the aim of understanding the template-framework interactions and creating targeted routes to specific frameworks. The location of the tetrapropylammonium ion in ZSM-5 by Baerlocher with PXRD data in 1984 [54] and Van Koningsveld et al. [55] (using difference Fourier maps from single-crystal diffraction data) in 1987 is an early example, and a recent review (focused on powder diffraction) by Smeets and McCusker shows that this is still an active area of research [27]. The review describes several tools used to

26

544

P. Beato et al.

determine template positions, e.g., difference Fourier mapping, molecular docking calculations, rigid bodies, and simulated annealing. Their technique of combining soft constraints on the framework (bond lengths and bond angles) with the charge flipping method has proven powerful for direct observation of cations in ion exchange positions and molecular species in zeolite cages and channels. Probe atoms are then placed at the positions of the observed residual electron density and the occupancies later refined in a Rietveld analysis. Here we begin to close in on catalytic applications.

26.4

Organic Molecules Adsorbed in Zeolites

The study of adsorbed organics in zeolites with diffraction began in 1984 with Fyfe et al. using XRD to confirm solidstate MAS-NMR results suggesting that adsorption of molecules including acetylacetonate caused distortions of the framework of ZSM-5 [56]. In 1985, the first two reports of adsorbed organic molecule locations from diffraction data were reported. Both were based on difference Fourier maps from neutron diffraction: Wright reported the structure of pyridine adsorbed in gallosilicate zeolite-L [57] (also using energy minimization calculations based on atom-atom potentials), and Fitch et al. determined the structure of benzene adsorbed in zeolite-Y [58]. In the same year, Mentzen began a series of papers describing powder diffraction studies of adsorbed organics in zeolites, mainly using laboratory powder diffraction data, which continued into the 2000s [59–70]. These include a notable early contribution on time-resolved in situ PXRD during desorption of several aromatics [71, 72]. In a similar time frame, Van Koningsveld and co-workers released a number of papers on the adsorption of organic molecules in zeolites, especially ZSM-5, using single-crystal XRD [73–78]. The groups of both Mentzen and Van Koningsveld also made extensive use of computer simulations (molecular mechanics energy minimizations [62, 65, 69], atom-atom potentials [68], force fields [75, 78]) to help locate the adsorbed molecules and compare to the diffraction data. The computational information is used both to help with initial placement of the molecules for refinement and to help understand the energetics of adsorption and interaction between sorbate and framework, although even with the latest computational methods disagreements between theoretical calculations and experimental observations can be found. Structures from diffraction data help to give a much more detailed understanding of the properties of internal zeolite surfaces – allowing features from thermal and kinetic measurements to be linked to real structural features and specific sites [64, 79, 80]. The structure of benzene adsorbed in ZSM-5 has also been studied by Taylor [81, 82] using neutron diffraction, Goyal et al. with a combination of

neutron and high-resolution synchrotron PXRD [83], and Kamiya et al. using single-crystal XRD [84]. Gameson and co-workers employed laboratory PXRD to locate and study the temperature-dependent occupancy of methyl chloride molecules in zeolite rho [80], while Parise used synchrotron PXRD to study temperature dependence in the adsorption of stilbene in ZSM-5 [85]. Nair and Tsapatsis determined the location of both o- and m-xylene in silicalite (the pure silica analogue of ZSM-5) using laboratory PXRD data [86, 87]. Fyfe, Gies, and co-workers combined single-crystal and powder diffraction with solid-state MAS-NMR spectroscopy to determine the monoclinic distortion of silicalite under high hexane loading, both at room temperature and 180 K [88]. The same authors also determined the structure of p-dichlorobenzene adsorbed in ZSM-5, along with two other zeolite structures, by similar methods [89]. NMR is highly sensitive to local structure distortions making it an exceptionally useful complement to diffraction, which reveals the long-range structure, especially for spotting the subtle changes that occur in zeolite geometry when molecules are adsorbed. NMR was again paired with PXRD by Grey and co-workers to resolve the structures of various loadings of hydrofluorocarbon-134 in zeolite-Y [90]. Kaszkur et al. also studied halogenated hydrocarbons, using both PXRD and molecular dynamics, revealing framework distortions, changes in cation occupancy, and specific sorbate-framework interactions [91, 92]. Most recently, Rodeghero and co-workers published a series of studies of organic molecules and mixtures desorbing from heated zeolites combining in situ synchrotron PXRD with various other techniques to provide a detailed picture of the desorption process and how it affects the structure of the framework [79, 93–95]. A summary of the works on organic adsorption in zeolites is presented in Table 26.1 (see also Table 1 in Wragg et al. [96] for a summary of crystallographic data for organic adsorption on ZSM-5, including unit cell parameters). In 1996, Smith and co-workers used a combination of IR spectroscopy, PXRD, and neutron powder diffraction to study the location of water and acid sites in SAPO-34 [106] and later its aluminosilicate analogue SSZ-13 [107]. The charge balancing protons in these frameworks are catalytically active and a key part of the MTO conversion mechanism. Another study of significant catalytic relevance is the location of methylamines in zeolite rho, a catalyst used for their synthesis, with a combination of PXRD and neutron diffraction by Weidenthaler et al. [100–102] More recently, in situ methods have been applied to the adsorption of methanol and water in SAPO-34 by Wragg and co-workers. [97] This study first determined the adsorption sites of methanol using high-resolution synchrotron PXRD and then used the model to fit the increasing population of the site to in situ PXRD data collected on a calcined SAPO-34 sample during

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

545

Table 26.1 A summary of the organic molecules which have been adsorbed in zeolite frameworks and located using crystallographic methods, along with the year of the work and other methods used to determine and validate the structures Year 1984 1985 1985 1985 1989 1993 1997

Organic molecule Acetylacetonate Pyridine Benzene n-Hexane, p-xylene Pyridine d-Benzene Benzene

Framework MFI LTL FAU MFI B-MFI MFI MFI

1994 1995 1992

MFI MFI MFI

1998

Methanol p-Xylene p-Xylene, p-chlorotoluene, p-dibromobenzene, p-dichlorobenzene, p-Nitroaniline

MFI

1993 1993

p-Dichlorobenzene Naphthalene

MFI MFI

2002

Tetrachloroethylene

MFI

1998 1989 1996 1996 1996 1997 1998 1987 1987 2000 2011 1988 1995 2000 2002 1991

p-Xylene p-Xylene Naphthalene p-Dichlorobenzene (low loading) p-Dichlorobenzene (high loading) p-Nitroaniline p-Dichlorobenzene and p-xylene d-Benzene d6-Benzene Benzene Benzene Methyl chloride Stilbene o-, m-Xylene (second reference correction) n-Hexane p-Chlorobenzene (in MFI)

1997 1991 1993 2015 2017 2016 2017

1,1,2,2-Tetrafluoroethane Chloroform, m-dichlorobenzene 1,4-Dibromobutane 1,2-Dichloroethane Methyl tert-butyl ether Toluene Methyl tert-butyl ether/toluene mixture, MTBE/1,2dichloroethane mixture Methanol o-Xylene DME Methylamine Dimethylamine Trimethylamine Pyridine, methanol, ammonia Pyridine d/l-Lysine

Cs-MFI MFI MFI MFI MFI MFI MFI MFI MFI MFI MFI RHO MFI MFI MFI MEL, MTW, MFI Na-FAU Na-FAU Na-FAU MFI MFI MFI MFI

2010 2008 2014 1997 1997 1997 2016 2017 2020

CHA MFI MFI RHO RHO RHO MFI FAU MFI

Data type used X-ray, 29Si MAS-NMR Neutron, atom-atom potentials Neutron X-ray X-ray Neutron X-ray, 29Si MAS-NMR, molecular mechanics X-ray X-ray, adsorption isotherms X-ray, molecular mechanics

Ref. [56] [57] [58] [59] [60] [61] [62]

X-ray, 29Si MAS-NMR, molecular mechanics X-ray X-ray, 29Si MAS-NMR, atom-atom potentials X-ray, 29Si MAS-NMR, molecular mechanics X-ray, adsorption isotherms Single-crystal X-ray Single-crystal X-ray, 29Si MAS-NMR, Single-crystal X-ray, force fields Single-crystal X-ray Single-crystal X-ray Single-crystal X-ray, force fields Neutron Neutron X-ray and neutron Single-crystal X-ray X-ray X-ray, TGA, 2H NMR X-ray, force fields X-ray, 29Si MAS
NMR, modeling X-ray, 29Si MAS -NMR

[66]

[70] [71] [74] [75] [76] [77] [78] [81] [82] [83] [84] [80] [85] [86, 87] [88] [89]

X-ray, 23Na MAS-NMR X-ray X-ray, molecular dynamics X-ray, TGA, adsorption isotherms X-ray, sorption data X-ray, adsorption isotherms X-ray, adsorption isotherms, TGA

[90] [91] [92] [93] [94] [79] [95]

X-ray Neutron Single-crystal X-ray X-ray, neutron, TGA-DTA X-ray, neutron, TGA-DTA X-ray, neutron, TGA-DTA X-ray X-ray, 27Al/31P NMR X-ray, TGA

[97] [98] [99] [100] [101] [102] [103] [104] [105]

[63] [64] [65]

[67] [68] [69]

26

546

P. Beato et al.

exposure to a flow of methanol-saturated helium. A similar in situ study on water adsorption (using the Smith et al. model [106] as a starting point) was used for comparison (Fig. 26.1). The in situ methanol adsorption results show a clear correlation between increasing occupancy of the two methanol locations and an increase in unit cell volume, in addition to which there appears to be some transfer of methanol from position 2, in the eight-ring window to position 1, inside the chabazite-type cage as the unit cell swells. Water adsorption, by contrast, reduces the unit cell volume suggesting a stronger interaction with the framework compared to methanol. The study of adsorbed molecules remains relevant. In 2008, Fyfe and co-workers determined the structure o-xylene in ZSM-5 with neutron diffraction [98]. Fujiyama et al. have used single-crystal XRD to determine the structure a)

O1 3.001 Å 3.250 Å

O4

3.060 Å

O7

O2m C2m

2.922 Å

O6

O1m 3.524 Å C1m

C2 3.281 Å O3

b) 1.19 2340 1.02

0.68

2334

0.51

2331

0.34 Frac. methanol 1

2328

Frac. methanol 2

0.17

Volume

2325

0.00 100

110

120

130 Time (min)

140

150

Volume (Å3)

2337

0.85 Fractional occupancy

Fig. 26.1 (a) Structure of methanol adsorbed in SAPO-34 showing significant contact distances between the molecules and the framework. (b) Plots of the unit cell volume and fractional occupancy of the two methanol sites vs time under flowing methanol-saturated helium. Increases in occupancy of the methanol sites are clearly correlated to an increase in unit cell volume. (Reproduced from Wragg et al. [97])

of DME (an intermediate in the MTG process [3]) in silicalite and determined the entrance and diffusion pathways of DME and carbon dioxide in the same framework via an interesting methodology involving sealing large, faceted crystals after controlled exposure to gas using a silicone resin [99, 108]. In situ studies of these processes with single-crystal diffraction remain a challenge for the future. Finally, the group of Tsang has, alongside the operando studies discussed below, determined several zeolite-organic structures using highresolution synchrotron PXRD, in the process transferring the idea of using probe molecules to reveal and quantify Brønsted sites neatly from spectroscopy to diffraction [103–105]. Determination of the structures of organic molecules adsorbed in the nano-spaces of zeolitic materials has

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

contributed significantly to our present understanding of the shape selectivity of zeolites in technologically important processes like sorption and catalysis and paved the way for the more recent developments in in situ and operando diffraction studies.

26.5

Catalysts “Postmortem”

Despite the large number of studies on adsorption of molecules into zeolites and the ubiquity of PXRD in the characterization of fresh zeolite catalysts, relatively few studies have used diffraction to examine zeolites postmortem, i.e., after catalytic deactivation. The earliest such study we have identified revealed a relationship that has proved extremely useful in further studies of coking of ZSM-5 in the MTG process. Alvarez and co-workers used PXRD to examine ZSM-5 samples before and after deactivation in a MTG reactor and noted from careful analysis of peak splitting that rather than transforming to a lower symmetry when coked, the ZSM-5 unit cell appeared to increase its symmetry to tetragonal [109]. The fitted unit cell changes from orthorhombic (space group Pnma, No. 62) with a ¼ 20.104, b ¼ 19.897, and c ¼ 13.392 Å to tetragonal (P42212, No. 94) with a ¼ b ¼ 19.965 and c ¼ 13.380 Å. Calcination reverses the change. This apparent shift to higher symmetry (apparent, as in fact the symmetry never truly becomes tetragonal) is contrary to what is observed in crystal structures of ZSM-5 with adsorbed p-xylene (high loading) [73], methanol [63],

a)

b)

and hexane [59, 88] and as-synthesized ZSM-5 containing the tetrapropylammonium template [110]. In all of these cases, the natural orthorhombic symmetry of the empty MFI-type framework is reduced to monoclinic. Recent studies (see below) have shown that coked ZSM-5 does not, in fact, increase its symmetry when coked: it remains orthorhombic but with a approximately equal to b. Rojo-Gama et al. have developed the transition from orthorhombic to tetragonal-like lattice parameters noticed by Alvarez et al. into a very useful descriptor for coke levels in ZSM-5 MTG catalysts [111]. It turns out that the progress of the difference in the a and b lattice parameters from significantly different (orthorhombic) to approximately equal (metrically tetragonal) correlates in a linear manner with the degree of coking (Fig. 26.2). By simply measuring a PXRD pattern, fitting the lattice parameters, and calculating the a-b parameter, they determine coke content in a range of samples from different positions in a MTG test reactor and fit the data to a model of the deactivation processes. They also obtained the a-b parameter from computational simulations. The organics retained in the catalyst at various times and at three different positions in the reactor bed were analyzed using gas chromatography to find the coke molecules most likely to be present at different reaction stages. Using density functional theory (DFT) methods, coke molecules were docked into the framework, and the entire model was relaxed to give new unit cell dimensions. The fully relaxed DFT models were verified with the Rietveld fits. The DFT calculations also revealed that fixing specific types of coke

c)

10

X-ray diffraction

Coke content Coke content (% wt)

547

Top

8

Top 6 Middle 4 Bottom 2 0 0.20

Middle Bottom Fresh 0.15 (a-b) (Å)

44.0 0.10

Fig. 26.2 Relationship between coke content from TGA and the transition from orthorhombic to (near) tetragonal symmetry in coked ZSM-5 MTG catalysts. (From ref [111]). (b) The panel shows how the catalyst bed was subdivided into three sections. The coloration (and extensive other characterization not shown here) demonstrates that the coking is most severe at the reactor inlet and less so at the outlet. (a) The panel

44.5

45.0

45.5

46.0

46.5

47.0

Angle (2-θ)

shows the total coke as a function of the a minus b parameter and position in the reactor. (c) The panel shows a very diagnostic peak splitting feature in the diffractogram. For fresh MFI, a distinct doublet is seen at 2θ ¼ 45 (Cu Kα1 radiation). As deactivation progresses and the a and b unit cell dimensions become very similar, the two peaks merge into one

26

548

2

1

Empty cages: low strain

3

More cages filledstrain is released (though lattice is still expanded)

One cage filled with large poly-aromatic coke: high strain

0.150 FWHM (101)

molecule at specific positions (straight channels, sinusoidal channels, and intersections) led to different levels of variation in the a-b parameter. Since changes in the experimental a-b parameter were observed for all levels of coking, this helps to define both the likeliest types and positions of coke molecules. This methodology is closely related to that of some of the molecular adsorption studies discussed above but cannot provide a perfect fit, since multiple coke species are present in the real samples. The deactivation by coking of SAPO-34 in the MTO process can also be tracked with crystallographic descriptors. Zokaie [112] and co-workers studied coked samples of SAPO-34 using a combination of laboratory PXRD and molecular mechanics calculations. Taking the operando study of Wragg et al. [113] (see below) as a starting point, they used the c-axis of the rhombohedral unit cell (space group R-3, hexagonal axes) as a descriptor for coke buildup. As in the Rojo-Gama et al. study [111], catalyst samples taken from test reactors at different times were analyzed to find the likeliest coke molecules. Molecular mechanics calculations were used to dock the molecules and relax the framework, giving theoretical unit cell parameters, which they compared to PXRD data fitted with the Rietveld method. In addition, they noted from calculations on a supercell comprising 8 hexagonal unit cells that the maximum expansion of the framework was reached before all the chabazite cages in the framework were filled with coke molecules. This was related to the strain values obtained from operando PXRD peak widths, which reach a maximum after the initial c-axis expansion. The authors suggest based on this that the formation of the initial, isolated, large coke molecules creates a highly strained lattice and that the strain is released as the type of molecules filling the cages becomes more uniform (Fig. 26.3). The approach of postmortem analysis of partially spent catalyst samples employed in this study has particular limitations for MTH conversion, where deactivation progresses gradually from reactor inlet to outlet. PXRD measurements of material taken from the entire bed and ground up will thus represent an average of the different states of deactivation at different positions in the catalyst bed. Spatially resolved studies (see Fig. 26.2 and Sect. 26.8) provide much more information in these cases. Wragg and co-workers used a combination of highresolution synchrotron PXRD and UV Raman spectroscopy to study SAPO-34 samples coked during conversion of methanol and propene feeds to mixed olefins [114]. A tapered element oscillating microbalance system was used to prepare samples with specific masses of coke per 100 g of catalyst. The HRPXRD data revealed peak splitting across the entire diffraction pattern for methanol feeds, corresponding not to a lowering of symmetry but to the formation of two rhombohedral unit cells with different coke contents and lattice parameters. These are attributed to different regions of the catalyst crystallites containing monocyclic (potentially active) or polycyclic aromatics (coke), which cause different degrees of lattice

P. Beato et al.

2 3

0.148 0.146 0.144

1

0.142 0

100 Time (min)

200

Fig. 26.3 The relationship between filling of chabazite cages in SAPO34 and microstrain measured by PXRD (from ref [112]). Simple models of cage expansion caused by large, polyaromatic coke molecules are linked to microstrain measurements from PXRD (increase in the full width at half maximum (FWHM) height of the SAPO-34 (101) peak). The peak strain is associated with a low coke occupancy state (2) where there are huge differences in volume between the empty and filled cages. As more cages fill with coke (3), the strain is somewhat reduced due to more cages having equivalent volume

expansion. For propene feeds, no splitting was observed, and expansion of the lattice in an in situ PXRD experiment was minimal compared to methanol [113]. UV Raman indicates that both feeds produce high levels of polycyclic aromatics at high coke levels but that methanol produces more of monocyclic species. They conclude that the propene feed was converted very quickly to coke on the crystallite surfaces, blocking the pores and preventing further propene from penetrating into the crystallites. For methanol, the feed can penetrate deeper into the crystallites before the pores are blocked. This leads to inactive polycyclic species on the periphery of the crystallites and potentially reactive monocyclic aromatics inside. The monocyclic species could react further but are insulated from the methanol feed by pore blocking. The authors suggest that by providing larger pathways into the crystallites, the feed could be allowed to reach deeper and utilize areas where the monocyclic species remain unreacted (an example of a catalyst subjected to this kind of treatment is found in the time- and space-resolved operando study of del Campo and co-workers [115], described in Sect. 26.8). PXRD has also been used alongside NMR and DRIFTS to study the long-term deactivation of SAPO-34 due to migration of isolated, catalytically active silicon into so-called silicon islands [116]. The authors used variations in the unit cell volume determined by structureless (Pawley method [117]) fits to help characterize the amount of silicon substitution and

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

how it is affected by steaming (carried out to simulate the hydrothermal conditions experienced by SAPO-34 in industrial MTO reactors). Such effects are closely related to dealumination of aluminosilicate zeolites [118] (see the study by Agostini et al. [119] described in Sect. 26.6). Finally, a pair of papers by Smith and co-workers describes the structure and catalytic properties (for MTO conversion) of SAPO-34/18 intergrowth materials in which the intergrowth levels were carefully controlled and characterized [120, 121]. This is to date the only detailed study which has attempted to understand the role of stacking faults in MTH catalysis using diffraction. The number of detailed, postmortem crystallographic studies of zeolitic methanol conversion catalysts is surprisingly small, considering the amount of useful information obtained from the studies carried out. We believe there is significant potential for gathering useful process information from quick and easy measurements for researchers willing to dive a little deeper into data analysis. However, as famously advocated by Somorjai, there are limits to what one can achieve when characterizing catalysts before and after use (“prenatal and postmortem” characterization) [122]. The incentive to carry out in situ and operando studies is strong.

26.6

In Situ Studies on Zeolites

Most of the in situ PXRD studies on zeolites have been motivated by the desire to understand how these immensely useful and strikingly beautiful structures emerge from solutions or mixtures of a limited number of reagents. A huge range of in situ work on zeolite synthesis has been published, and much of it has been reviewed elsewhere [123–126]. Although no clear single mechanism of zeolite synthesis has yet emerged, the experiments have provided excellent kinetic information [127] and often revealed stable intermediate phases [128]. Combining diffraction with other techniques has been informative [129]. Total scattering analysis uses the Fourier transform of the diffraction pattern to create a pair distribution function (PDF) representing the distances between atom pairs in real space [130]. The power of this method lies in the fact that it uses all of the scattering from the sample, not just the Bragg diffraction peaks from the crystal lattice. This means that the PDF describes the local structure (even for noncrystalline materials) as well as the long-range (crystalline) structure of the material. The analysis of total scattering data on in situ synthesis has potential to reveal noncrystalline precursors and intermediates. Removal of organic templates is another part of the zeolite preparation process that has been studied in situ. Synchrotron radiation combined with the area detectors (CCDs and image plates) that became available in the 1990s gives the time resolution (from 1–2 min down to under a second) needed

549

to follow the processes occurring during calcination. Most of the studies mentioned here were carried out at synchrotrons. The first reported in situ PXRD study of calcination (or template burning) was by Milanesio et al. in 2003 [131]. They were able to follow both lattice parameter variations and site occupancy changes for TS-1 and Fe-MFI zeolites using the Rietveld refinements and fit the results with standard kinetic models. In 2006, Lassinantti Gualtieri and co-workers studied the removal of the tetrapropylammonium template from silicalite, using a laboratory diffractometer and similar analysis to Milanesio et al. [132]. Bhange et al. studied the same system with a combination of PXRD and thermogravimetric analysis [133]. Martucci et al. studied thermal effects on CoAPO-34 [134] in a detailed study which had sufficient data quality to reveal small variations in the framework geometry during heating, as well as showing the loss of both the morpholine template and extraframework fluorine. More recently, Kalantzopoulos et al. made use of the high time resolution (10 s per diffractogram in this study, but the latest version of the detector used is capable of a frame rate of 500 Hz [135]) afforded by the latest photon counting pixel area detectors to study the calcination of SAPO-37 in detail with simultaneous synchrotron PXRD and mass spectrometry and separate TGA-DSC studies [136]. Photon counting detectors represent a major advance over earlier technology, combining the high dynamic range and sensitivity to highenergy photons of image plates with speed that can match and exceed that of CCD systems. Different behavior was observed in the two different cage types (SOD and FAU) of the structure. Careful analysis of variations in the refined framework bond lengths and angles showed that the template in the smaller SOD cage does not leave the structure until the FAU cage is completely clear. The same team also used in situ PXRD to study the effect of water on the SAPO-37 framework, this time combining the data with DFT calculations, in situ DRIFTS, and adsorption measurements. SAPO37 is known to have poor stability in air below 345 K [137]. The results clearly indicate a mechanism in which water clusters in the smaller cages of the structure (SOD cages and double-six-ring units). Here, the clusters distort the framework (Fig. 26.4) and take on an acidic character which can break down the framework in a violent manner as soon as the temperature drops to a point where water interacts strongly with the framework [138]. The studies of Kalantzopoulos et al. utilized parametric Rietveld refinement [139], a powerful method for treating large PXRD datasets in which the model is refined against all diffraction data simultaneously, with some model parameters unique for each diffractogram and others linked across the whole data surface. This allows rapid fitting of large datasets with a minimized number of least squares parameters. It also gives precise values of parameters like zero error

26

550

P. Beato et al.

a)

b)

O2

O2

O4

O4 H2O

Fig. 26.4 The structures of dry SAPO-37 (left) and the same material with adsorbed water (right), showing the significant distortion of the framework as water fills the smaller cages. (Taken from ref [138]). As

soon as water fills these cages, this distortion makes the structure susceptible to hydrothermal damage

which must be equal for all diffractograms and permits structural parameters to be linked to known, observed, or expected physical behavior (note that the latter should be avoided as it risks forcing the model to fit our own ideas of how it should behave!). An excellent example of parametric Rietveld refinement is found in the method’s first application to in situ zeolite data (combined synchrotron PXRD and XAS) [119]. Agostini et al. studied the dealumination of zeolite-Y. Initial analysis of the PXRD data using sequential Rietveld analysis showed interesting trends in the occupancy of the ammonium and water sites in the framework as the sample was heated; however, the data had high levels of uncertainty (Fig. 26.5a). Analysis of the refined parameters showed large oscillations in the framework atom positions. Analysis of selected datasets across the temperature range showed that these actually varied in a linear manner with temperature and so could be parameterized with the equation of a straight line:

and P0 is the starting value of the parameter. Now only a and P0 are needed to be refined for each set of structural parameters, reducing the number of least squares parameters in the refinement of all diffractograms from 62 to 2 for each structural parameter linked to a straight line. Variation in framework atom positions and occupancies observed in the ion exchange (I, I′, II, and II′) and water positions (A, B, C) affects the relative intensities of the Bragg peaks. If many of these parameters are refined freely, their values become highly correlated. Parameterization of the framework atom positions significantly reduces their correlation with the refined parameters, resulting in lower errors in the ion exchange site occupancies and revealing slightly different trends as shown in Fig. 26.5b. Only after minimizing these correlations was it possible to resolve the true details of the dealumination process. The parametric results revealed that the aluminum site occupancy in the framework increased when the temperature was held at 875 K, explaining the shrinking unit cell parameter observed during this stage of heating in both the sequential and parametric refinements. Note that the lattice parameter immediately reveals the trend later linked to the occupancies. Lattice parameters are



PðiÞ ¼ a i þ P0 where i is the number of the diffractogram, P(i) is the parameter value in diffractogram i, a is the slope of the fitting line,

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

24.78

24.78 b)

a)

80 24.76

60

24.74 H2O (A) H2O (B) H2O (C)

40

24.72

NH4+ (I') NH4+ (II)

20

24.70 24.68

Cell parameter (Å) Number of atoms per unit cell

80 Number of atoms per unit cell

551

24.76 24.74

60

24.76 4

Isotherm

40

24.74

2

24.73

Ramp up

20

24.75

300

400

500 600 700 Temperature (K)

800

24.66 900

260 280 300 320 340 Time (min)

300

400

H2O (A)

H2O (B)

500 600 700 Temperature (K) H2O (C)

+

NH4 (I')

24.70 24.68 24.66 24.64

0

24.62

0 0

24.72

Cell parameter (Å)

26

24.60 900

800 +

NH4 (II)

Al3+ (I')

c) Im

IIm

II A B C

Fig. 26.5 Plots of extra framework site occupancy factors vs temperature from in situ PXRD data on dealumination of zeolite-Y, fitted with (a) sequential Rietveld refinement and (b) parametric Rietveld refinement. Diagram (c) shows the locations of the extra-framework sites. Both sets of results show clear correlations between unit cell volume changes and loss of water and ammonium; however, the lower noise

levels in the parametric refinement make the patterns much clearer and allow us to see the link between aluminum site occupancy and unit cell parameter (b inset) during the 875 K isotherm. Note the significantly smaller error bars and the absence of population of H2O site (B) above 500 K in the parametric refinement results. (Taken from reference [119])

generally much better determined (a few parameters are refined against many independent observations, i.e., the peak positions) in diffraction data than the parameters which influence the peak intensities (where a large number of parameters influence a few small variations in the observed data). Least squares refinement in general relies on having more data than parameters [25]. Finally, Kalantzopoulos et al. have recently used total scattering analysis of synchrotron PXRD data to study the response of SAPO-34 to hydration. This method, sensitive to both local structure and long-range order [130], was combined with in situ flow-NMR spectroscopy to reveal the subtle structural changes taking place [140]. Total scattering studies at the latest generation of high-energy synchrotron beamlines (e.g., beamline ID15A [141] of the European synchrotron, with its Extremely Brilliant Source upgrade

[142]) open up new possibilities for in situ and operando studies of crystalline, amorphous, and pre-crystalline catalysts which are only just beginning to be developed. In situ XRD studies have already made a huge contribution to our understanding of zeolite catalysts, their synthesis, and properties. The further application of methods like total scattering and the development of faster experimental facilities will undoubtedly lead to new and exciting discoveries in the future.

26.7

Operando Catalytic Studies

In an operando experiment, structure and activity data are collected simultaneously [143]. This is one of the most informative ways of studying a functional material as we can often see the actual active structures at specific points in the cycle

26

552

of reaction. The term operando was first coined by Bañares in the early 2000s to describe spectroscopic experiments [144–148]. In fact, as early as 1991, Clausen and co-workers [149] reported a capillary-based cell which allowed XRD data to be collected on a solid catalyst under catalytic reaction conditions (water-gas shift), while activity was monitored, i.e., an operando experiment. The simple capillary cell has been further developed for a range of applications [150–152], and variants for gas and liquid flow were used in most of the operando studies described here. Some recent reviews have covered the activity in this predominantly synchrotron-based area in a general way [124, 126, 153, 154], and a laboratory-based operando PXRD setup has been reported [155]. Here we have selected some specific examples from the field of methanol conversion to hydrocarbons (MTH). A prerequisite for the success of such experiments is to find a practical experimental setup that provides excellent diffraction data, proper control of the catalyst sample environment (reactor geometry, flow characteristics, temperature control), and adequate analysis of reaction products. The capillary setup originally described by Clausen et al. [156] has proved to be very useful. A version of this cell, allowing gas flow over a catalyst bed with analysis of the product stream by mass spectrometry, is shown in Fig. 26.6. This system was used in several studies described in this chapter. The first operando study on methanol conversion was reported by Wragg and co-workers in 2009 [113]. They studied the deactivation of SAPO-34 by coke during the MTO process with a combination of PXRD and Raman spectroscopy while measuring catalyst activity with MS. The work shows a clear relationship between the

Fig. 26.6 A Clausen-type Swagelok cell for PXRD studies on a catalyst bed. The capillary is glued into a metal bracket, which can be aligned on the goniometer and rotated through a small angular range to improve powder averaging. The product gas stream passes out through the tube on the left for MS analysis. The capillary is heated from below by a hot air blower. The Pilatus 2 M area detector is on the left, and the X-ray beam passes through from the pointed gray collimator tube on the right

P. Beato et al.

c-axis of the SAPO-34 unit cell and the production of propene, with the maximum propene production at roughly 1/3 of the c-axis expansion, dropping to zero as the c-axis expansion levels off (Fig. 26.7a). The flow of unreacted methanol through the bed rises roughly in step with the c-axis expansion. Difference Fourier maps revealed the buildup of coke in the SAPO-34 CHA cages, and quantification of the electron density showed a trend perfectly correlated with c-axis expansion (Fig. 26.7 right). The Raman data capture the change from simple aromatic intermediates inside the zeolite to polyaromatic coke. The c-axis was later used as a simple descriptor for coke development in SAPO-34 in the postmortem work of Zokaie et al. (see Sect. 26.5) [112]. A similar operando study by the same team on SAPO-18 as an MTO catalyst revealed related but different behavior compared to SAPO-34 [157]. Once again, they observed a relationship between electron density in the zeolite cages and unit cell expansion, but in the case of SAPO-18, higher relative levels of electron density led to a smaller relative level of unit cell expansion than in SAPO-34. This was rationalized by comparing the different cage shapes in the two zeolites with the large coke molecules found in deactivation studies by Marcus and co-workers [158]. SAPO-18 can accommodate the largest coke molecules without significant distortion due to larger cages and a rigid structure of double-six-ring units. Lo et al. recently published a study on the behavior of ZSM-5 in the MTG process which reveals the interactions between methanol and the zeolite framework at various stages of reaction with a combination of synchrotron PXRD and molecular dynamics calculations [159]. The study provides structures of methanol, dimethyl ether, and carbon adsorbed on the acid sites of ZSM-5; however, the PXRD data were collected after cooling of the sample from reaction temperature at each stage, and the maximum reaction temperature of 250  C bears little relation to industrial MTG reaction conditions. Moreover, the success of this study relies on very specific insights about aluminum distribution and acid site location to ensure ordering of the adsorbates. Operando studies are a well-established part of the development and understanding of catalyst systems both for academia and industry. Crystallography is playing an increasingly important part in these studies, especially in the world of zeolite science, and we expect that it will continue to do so.

26.8

Time- and Space-Resolved Operando Studies

Real MTH reactors are large and contain extruded catalyst bodies composed not only of active material but also supports, binders, and sometimes promoters [160]. True

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

a)

b)

15.35

Methanol (M/e = 31) Propene (M/e = 42) c-axis

15.2

300 15.30 15.25

15.1

2.00E–010

c-axis (Å) Void 2 electron count

200

15.20 15.15

0.00E+000

15.0

250

150

Void electron count

4.00E–010

Mass spectrum intensity

553

c-axis (Å) c-axis (Å)

26

15.10 100 15.05

0

20 Time (min)

40

0

20

40

60 80 Time (min)

100

120

140

Fig. 26.7 Results from operando study of the MTO process at 450  C. (a), correlations between the c-axis of SAPO-34 and propene/methanol MS signals; (b), correlation between the SAPO-34 c-axis and the electron density in the CHA cages. (Taken from ref [113]). Note the different

timescales in the panels. Clearly, a buildup of material retained in the catalyst pores (as seen from the electron density, “Void 2 electron count”), an extension of the c-axis length (from refinement), and catalyst deactivation (from MS data) occur in parallel

operando studies, at real working conditions, seek to recreate these conditions and really understand the industrial catalyst systems. To do this requires moving beyond single-point measurements on model catalyst powders and into two- and three-dimensional studies of something closer to the real industrial catalyst bodies. At the forefront of this movement have been Beale and co-workers, who were the first to report spatially resolved studies of dense oxide catalyst beds with the fast z-scanning method [161] and operando studies of catalysts in three dimensions using both X-ray diffraction and PDF computed tomography [162–165]. They have also applied these methods to study battery materials [166], fuel cells [167], and teeth [168]. The present state of the art here is “five-dimensional imaging” where the fifth dimension (in addition to the three spatial dimensions and time) is chemical structure information [165, 169]. Wragg and co-workers have applied similar methods for zeolitic MTH catalysts. The first study in 2012 [170] used the SAPO-34 c-axis descriptor for coking in the MTO process described in Sect. 26.7 [113]. The experiment was carried out at the ESRF on beamline ID15B. A 4 mm internal diameter quartz reactor filled with a bed of sieved SAPO-34 particles was set up on Swagelok mount which allowed gas to be flowed through the catalyst bed. This setup was mounted on a z-axis scanning stage so that it could be translated up and down through the X-ray beam (a 0.5 mm pencil beam). Diffractograms were collected with a 1-second exposure time on a Pixium detector, allowing the authors to track changes in the catalyst structure as methanol flowed up through the bed and coke developed. MS data were collected at the reactor outlet to monitor the products. Catalysts containing 4% and 8% silicon were studied under flow rates of 30 and 50 mL

min
1 of methanol-saturated helium. By extracting the c-axis values with parametric Rietveld methods, the authors produced maps of the progress of coking in time and space. This showed the initial development of active hydrocarbon intermediates at a level in the reactor that depends on the flow rate (and thus the time which the feed spends in contact with the catalyst) and silicon content (Fig. 26.8d). The intermediates then spread back toward the reactor inlet against the reagent flow before spreading toward the outlet, occupying the whole reactor. Development of heavier, inactive coke molecules, which cause greater c-axis extension, is observed later in the process (Fig. 26.8c). A kinetic model of the formation and spread of active intermediates and coke is also described. This reproduces much of the behavior observed in the XRD, providing a clear link between the structural changes and the chemistry of the MTO process. The same team applied the fast z-scanning method to a ZSM-5 sample with the same setup [96], showing small structural changes during MTG conversion. The results provided an interesting demonstration of how a very large dataset can be carefully analyzed to capture very subtle trends in catalyst behavior, and the range of structural parameters which can be extracted from data of this kind with parametric Rietveld refinement, but did not reveal significant detail of the deactivation processes. Likely reasons for this are that the experiment was run under conditions where the catalyst did not deactivate significantly in the timescale of the experiment and peak broadening due to the sample thickness. Peak broadening effects from large tubular samples are described in this paper and elsewhere [161]. Del Campo and co-workers carried out a similar study on ZSM-22 catalysts with the linear channel-based TON

26

554

a)

15.05 15.04

3

15.02

10

15.00 14.98

8

14.96

2

4

1

MS intensity

14.94 6 Bed height (mm)

Fig. 26.8 Color map plots of c-axis variation in time and space in a SAPO-34 reactor bed during the MTO process at 340  C. Map (b) shows the first 15 min of the same reaction shown in map (a). Mass spectrometry data are plotted as lines over the maps. (Figure from reference [170]). The c-axis expansion caused by formation of active intermediates at 4 mm up the bed after 3 min on stream, followed by their spread back toward the reactor inlet, is clearly visible in (b). The larger expansion caused by inactive coke is observed in (a) from 4 mm and up after 80 min

P. Beato et al.

2

14.92 c-axis length (Å)

Methanol DME Propene 0

0 0

20

40

60

80

100

120

140

160

180

TOS (min)

b)

15.05 15.04

3

15.02

10

15.00 14.98

8

14.96

2

Bed height (mm)

4

1

2

MS intensity

14.94 6

14.92 c-axis length (Å)

Methanol DME Propene

0

0 0

2

4

6

8

10

12

14

TOS (min)

framework [115]. The paper compares methanol conversion over a pristine ZSM-22 and a sample of the same material post-treated with base and acid to produce additional porosity and improve the penetration of the feed gas into the catalyst. Finally, the effect of feeding a mixture of methanol and propanol over a sample of the parent catalyst under MTH conditions was studied. Data were collected at the SwissNorwegian beamline (BM01) of the ESRF using a Pilatus 2 M detector [171]. This two-dimensional, solid-state photon counting detector represents the present state of the art in X-ray detection [172] and allows very fast data collection

even at a relatively low-flux source like BM01. The study shows a clear link between the volume of ZSM-22 unit cell and the degree of coking, and despite pronounced peak broadening from the 4 mm reactor tube, significant differences are shown between the progress of the reaction in time and space through the different catalyst beds and with different feeds. The changes are related to different organic species present at different positions in the beds at different times (Fig. 26.9) which control the rates of reaction, coking, and (consequently) unit cell expansion.

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

a) 0.5

555

H-ZSM-22C co-fed with MeOH + iPrOH

26

Unit cell volume (Å3)

IV III Positions at the catalytic bed from 1 (start) to 8 (end)

1 4 7

II

2 5 8

3 6

I 0

100

50

150

Time on stream (min)

I

MeOH + iPrOH II

Alkenes

Rojo Gama et al. used the a-b descriptor for MTG coking of ZSM-5 described in Sect. 26.5 [111] in a time- and spaceresolved operando experiment [173]. To obtain data with sharper PXRD peaks, the 4 mm reactor tube was replaced with a 0.5 mm internal diameter capillary packed with a long bed of sieved ZSM-5 powder. The capillary reactor was scanned through the X-ray beam at ESRF BM01, while high-quality diffraction patterns were collected with 10-s time resolution in ten positions for each scan. The data presented improve significantly on the ZSM-5 fast z-scanning experiment described above [96] and show clear evidence of the “burning cigar” model of autocatalytic deactivation [174]. Rietveld refinements allowed the authors to extract a range of structural parameters including direct determination of the coke content from the residual electron density in the cages as described for single-point operando experiments in Sect. 26.7. A strong correlation exists between the coke content (i.e., residual electron density on the cages, obtained from variations in the peak intensities) and the a-b parameter (obtained from the peak positions), and the structural variations observed through the bed in time and space could be related to process parameters by a kinetic model based on autocatalytic deactivation (Fig. 26.10).

MeOH + iPrOH

Alkenes

III

MeOH + iPrOH

Active (slow) Alkenes

IV

Deactivated

MeOH + iPrOH

Deactivated

d

b) Active (slow) Dea zo cti ne va te

Fig. 26.9 (a) Unit cell expansion of a ZSM-22 catalyst vs time and bed position under MTH conditions with a flow of methanol and isopropanol. (b) Sketches showing the authors’ hypotheses for the intermediate and coke molecules present at different positions in the bed at the time points marked I, II, III, and IV in the plot of unit cell volume vs time above. On initial exposure to reagent flow, the whole bed expands as it is filled with active linear and branched hydrocarbons. By time point (II), the inlet end of the bed has been deactivated by aromatic coke. By time point (III), more than half of the bed is filled with polyaromatic coke with a small portion near the outlet still active but expanding more slowly due to limited penetration of reagents through deactivated portion of the bed. By time point (IV), the entire bed is filled with inactive, polyaromatic coke

Active zone

26

MeOH/DME + iPrOH

Despite the ease with which crystallographic information can be extracted from zeolite PXRD data, 3D XRDCT studies are still in their infancy. Wragg et al. reported a postmortem, three-dimensional study of an MTO reactor bed (in fact one of the samples from the fast z-scan study described above [170]) in 2015 [175], at the same time introducing parametric Rietveld methods as an efficient tool for extracting structural data from large XRDCT datasets. More recently, we have used the upgraded ESRF beamline ID15, now equipped with superior optics, sample environment, and a Pilatus 2 M pixel area detector adapted for high-energy applications with a cadmium telluride sensor, to carry out 5D operando studies on ZSM-5 catalyst extrudates under MTG conditions. The experiment allows us to observe coking patterns in three dimensions through the catalyst pellets with a time resolution of around 10 min for a complete three-dimensional scan of the pellet (each 3D scan consists of 3 XRDCT slices; see Fig. 26.11) with 20 μm spatial resolution within the slices. The series of almost 100 3D XRDCT tomograms over 8 h of MTG reaction reveals both axial and radial gradients in the coking of the extrudate vs time. The experiment also shows, significantly, that much of the catalyst pellet remains unreacted after the apparent activity measured by online MS drops to zero. The changes can be tracked both from

556

P. Beato et al.

a)

MeOH (31)

DME (46)

Alkene (41)

Toluene (91)

c)

0 5

1 6

2 7

3 8

4 9

6 5

Intensity (a.u.)

Occupancies (a.u.)

4 3 2 1 0 0

b) 0.16

40 80 TOS (min) 0 5

120 1 6

160 2 7

3 8

200

240

0

d)

4 9

40 80 TOS (min)

120 0 5

1.0

1 6

160 2 7

200 3 8

240

4 9

0.14 0.8

0.12 0.10

0.6 Coke coverage (a.u.)

(a − b) (Å)

0.08 0.06 0.04 0.02 0.00

0.4 0.2 0.0

0

40 80 TOS (min)

120

160

200

240

0

40 80 TOS (min)

120

160

200

240

Fig. 26.10 Operando activity and PXRD data for the MTG process over ZSM-5. (a) MS signals for MeOH, DME, alkenes, and aromatics (multiplied by a factor of 10). (b) The (a – b) parameter for the different layers of the catalytic reactor at increasing TOS. (c) The total coke occupancies at increasing reaction times for the ten reactor slices. (d)

Simulated coke coverage derived from the autocatalytic deactivation model. (Figure from reference [173]). The reaction and deactivation by coke formation clearly proceeds in a cascade fashion from inlet to outlet. Once coke is built up in the final reactor segment, methanol breakthrough is observed in the MS data

the a-b parameter and the coke content as for the spatially resolved capillary experiment of Rojo Gama et al. above [173]. Mapping of the quantitative distributions of catalyst and alumina binder and variations in crystallite size and strain is also possible.

conversion but also applications of metal-loaded zeolites in process like selective catalytic reduction [176–178, 186]. PDF analysis of total scattering data will also play an important role in future developments. For heterogeneous catalysis more generally, the long-range ordered structures that can be detected by X-ray diffraction often represent only one part of the catalyst structure, and amorphous or sub-nanometer domains can play a significant role for the catalytic performance. In addition, information on the interaction of molecules with the solid surface, i.e., the solid-gas interface is highly desirable for a better understanding of catalytic reactions. Therefore, the further development of methods which provide direct information on different length scales is necessary to obtain the complete picture. In this respect, the Swiss-Norwegian beamlines (SNBL) at the ESRF have been pioneering the field by combining XRD,

26.9

Perspective

Crystallography is an increasingly important technique in catalysis, particularly for zeolites. Its applications now extend well beyond the basic characterization of solid catalysts after synthesis. The highly crystalline nature of zeolite catalysts makes them ideal for studies of this kind, especially operando experiments. Recent examples have shown the suitability of PXRD methods for studying not only methanol

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

a)

b)

0.150

557

c)

0.145

a − b (Å)

0.135 0.130 0.125 0.120 0.115

Coke occupancies

0.140

0.110 0.105

5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

0.100

d)

e)

0.150

f)

0.145

a − b (Å)

0.135 0.130 0.125 0.120 0.115 0.110 0.105

Coke occupancies

0.140

5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

0.100

Fig. 26.11 XRDCT reconstructions of slices through a ZSM-5 catalyst extrudate, fresh (a) and after 17 h (d) on stream in the MTG process. The variation across the diameter of the extrudate is also plotted for the a-b parameter (b) and (e), and the total coke content (c) and (f) is determined by the dummy atom method. The data come from an operando XRDCT experiment. The extent of coking on the outer circumference of the pellet and the relatively coke-free interior are clearly visible in the data collected after 17 h on stream (d, e, f). Time- and space-resolved

diffraction experiments offer a powerful route to connect the fundamental science of structure-property relationships in catalyst materials to real-world applications in understanding the behavior and kinetics of industrial catalyst beds and extrudates. As such, they represent a massively valuable source of information for both industry and academia which is as yet relatively untapped. We expect these methods to find many further applications in the study of methanol conversion catalysts and beyond

XAS, Raman spectroscopy [113], and now PDF [179]. Further combination of techniques for long- and short-range structure will also feature in future studies. With high-intensity X-ray rotating anodes and liquid jet source technology [180] becoming more commonplace in the home laboratory, ultrafast detectors providing data of unprecedented resolution from ever shorter exposures, improvements in experimental methodology, and a new generation of extremely brilliant synchrotron sources coming online [142, 181–185], the future for operando crystallography of zeolite catalysts is bright.

2. Dinnebier, R.E., Billinge, S.J.L.: Powder Diffraction: Theory and Practice, 1st edn, p. 582. Royal Society of Chemistry, London (2008) 3. Olsbye, U., Svelle, S., Bjørgen, M., Beato, P., Janssens, T.V.W., Joensen, F., Bordiga, S., Lillerud, K.P.: Conversion of methanol to hydrocarbons: how zeolite cavity and pore size controls product selectivity. Angew. Chem. Int. Ed. 51(24), 5810–5831 (2012) 4. Pérez-Fortes, M., Schöneberger, J.C., Boulamanti, A., Tzimas, E.: Methanol synthesis using captured CO2 as raw material: technoeconomic and environmental assessment. Appl. Energy. 161, 718–732 (2016) 5. Murphy-McGreevey, C.; Whitaker, G.: Zero routine flaring by 2030. https://www.worldbank.org/en/programs/zero-routine-flar ing-by-2030 6. Making ‘green’ gasoline in an integrated biorefinery. https://www. gti.energy/making-green-gasoline-in-an-integrated-biorefinery/ 7. Ravn, S.: World’s only natural gas-to-gasoline plant in operation in Turkmenistan. https://blog.topsoe.com/worlds-only-natural-gas-togasoline-plant-in-operation-in-turkmenistan 8. UOP. https://www.uop.com/?press_release¼jiangsu-sailboataccepts-worlds-largest-single-train-coal-to-chemicals-plant-fromhoneywell-uop

References 1. Giacovazzo, C., Monaco, H.L., Artioli, G., Viterbo, D., Milanesio, M., Gilli, G., Gilli, P., Zanotti, G., Ferraris, G., Catti, M.: Fundamental of Crystallography, p. 864. Oxford University Press, Oxford (2011)

26

558 9. Tian, P., Wei, Y., Ye, M., Liu, Z.: Methanol to olefins (MTO): from fundamentals to commercialization. ACS Catal. 5(3), 1922–1938 (2015) 10. Koempel, H., Liebner, W.: Lurgi’s Methanol To Propylene (MTP ®) Report on a successful commercialisation. In: Bellot Noronha, F., Schmal, M., Falabella Sousa-Aguiar, E. (eds.) Studies in Surface Science and Catalysis, vol. 167, pp. 261–267. Elsevier, Amsterdam (2007) 11. Burton, A.W.: Powder diffraction in zeolite science. In: Chester, A.W., Derouane, E.G. (eds.) Zeolite Characterization and Catalysis. Springer, Dordrecht (2009) 12. Rossmann, M.G., Arnold, E.: Patterson and molecular replacement techniques, and the use of noncrystallographic symmetry in phasing. In: Shmueli, U. (ed.) International Tables for Crystallography Volume B: Reciprocal Space, vol. B, pp. 235–238. International Union of Crystallography, Chester (2006) 13. Vijayan, M., Ramaseshan, S.: Isomorphous replacement and anomalous scattering. In: Shmueli, U. (ed.) International Tables for Crystallography Volume B: Reciprocal Space, vol. B, pp. 264–275. International Union of Crystallography, Chester (2006) 14. Giacovazzo, C.: Direct methods. In: Shmueli, U. (ed.) International Tables for Crystallography Volume B: Reciprocal Space, vol. B, pp. 210–234. Springer Nature (2010) 15. Burton, A.W.: Structure solution of zeolites from powder diffraction data. Zeitschrift für Kristallographie – Crystalline Materials. 219(12), 866–880 (2004) 16. Wragg, D.S., Morris, R.E., Burton, A.W.: Pure silica zeolite-type frameworks: a structural analysis. Chem. Mater. 20(4), 1561–1570 (2008) 17. Dawson, C.J., Kapko, V., Thorpe, M.F., Foster, M.D., Treacy, M.M.J.: Flexibility as an indicator of feasibility of zeolite frameworks. J. Phys. Chem. C. 116(30), 16175–16181 (2012) 18. Grosse-Kunstleve, R.W., McCusker, L.B., Baerlocher, C.: Powder diffraction data and crystal chemical information combined in an automated structure determination procedure for zeolites. J. Appl. Crystallogr. 30(6), 985–995 (1997) 19. Deem, M.W., Newsam, J.M.: Determination of 4-connected framework crystal structures by simulated annealing. Nature. 342(6247), 260–262 (1989) 20. Rius, J.: XLENS, a direct methods program based on the modulus sum function: its application to powder data. Powder Diffract. 14(4), 267–273 (1999) 21. Palatinus, L.: The charge-flipping algorithm in crystallography. Acta Crystallogr. B. 69(1), 1–16 (2013) 22. Sheldrick, G.: SHELXT - integrated space-group and crystalstructure determination. Acta Crystallogr. A. 71(1), 3–8 (2015) 23. O’Keeffe, M.: Flipping marvelous: new zeolites by new methods. Angew. Chem. Int. Ed. 48(44), 8182–8184 (2009) 24. Caliandro, R., Chernyshov, D., Emerich, H., Milanesio, M., Palin, L., Urakawa, A., van Beek, W., Viterbo, D.: Patterson selectivity by modulation-enhanced diffraction. J. Appl. Crystallogr. 45(3), 458–470 (2012) 25. Prince, E., Boggs, P.T.: Least squares. In: Prince, E. (ed.) International Tables for Crystallography, vol. C, pp. 678–688. International Union of Crystallography, Chester (2006) 26. Rietveld, H.: A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 2(2), 65–71 (1969) 27. Smeets, S., McCusker, L.B.: Location of organic structuredirecting agents in zeolites using diffraction techniques. In: Gómez-Hortigüela, L. (ed.) Structure and Bonding: New Insights into the Chemistry of Organic Templating in the Synthesis of Zeolitic Materials, vol. 175, pp. 1–31. Springer, Cham, Switzerland (2017) 28. Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-Physikalische Klasse (1918)

P. Beato et al. 29. Le Bail, A.: Chapter 5. The profile of a Bragg reflection for extracting intensities. In: Powder Diffraction: Theory and Practice, pp. 134–165. The Royal Society of Chemistry. Cambridge, UK (2008) 30. Coelho, A.: TOPAS and TOPAS-Academic: an optimization program integrating computer algebra and crystallographic objects written in C++. J. Appl. Crystallogr. 51(1), 210–218 (2018) 31. Bergmann, J., Friedel, P., Kleeberg, R.: Handling of unusual instrumental profiles by the BGMN Rietveld program. Mater. Sci. Forum. 321–324, 192–197 (2000) 32. Cheary, R.W., Coelho, A.: A fundamental parameters approach to X-ray line-profile fitting. J. Appl. Crystallogr. 25(2), 109–121 (1992) 33. Katerinopoulou, A., Balic-Zunic, T., Lundegaard, L.F.: Application of the ellipsoid modeling of the average shape of nanosized crystallites in powder diffraction. J. Appl. Crystallogr. 45(1), 22–27 (2012) 34. Balić Žunić, T., Dohrup, J.: Use of an ellipsoid model for the determination of average crystallite shape and size in polycrystalline samples. Powder Diffract. 14(3), 203–207 (1999) 35. Zuev, A.: Chapter 6. Instrumental contributions to the line profile in X-ray powder diffraction. Example of the diffractometer with Bragg–Brentano geometry. In: Powder Diffraction: Theory and Practice, pp. 166–205. The Royal Society of Chemistry. Cambridge, UK (2008) 36. Newsam, J.M., Treacy, M.M.J., Koetsier, W.T., De Gruyter, C.B.: Structural Characterization of Zeolite Beta. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences. 420(1859), 375–405 (1988) 37. Sławiński, W.A., Wragg, D.S., Akporiaye, D., Fjellvåg, H.: Intergrowth structure modelling in silicoaluminophosphate SAPO-18/ 34 family. Microporous Mesoporous Mater. 195, 311–318 (2014) 38. Lillerud, K.P., Akporiaye, D.: Systematic relationships between the structures of CHA, AEI and KFI. In: Weitkamp, J., Karge, H.G., Pfeifer, H., Hölderich, W. (eds.) Studies in Surface Science and Catalysis, vol. 84, pp. 543–550. Elsevier, Amsterdam, Netherlands (1994) 39. Treacy, M.M.J., Newsam, J.M., Deem, M.W.: A general recursion method for calculating diffracted intensities from crystals containing planar faults. Proceedings of the Royal Society of London. Series a, Mathematical and Physical Sciences. 433, 499–520 (1991) 40. Baerlocher, C.; McCusker, L. http://www.iza-structure.org 41. Leonard, R.J.: The hydrothermal alteration of certain silicate minerals. Econ. Geol. 22(1), 18–43 (1927) 42. Taylor, W.H.: I. The structure of analcite (NaAlSi2O6 H2O). Zeitschrift für Kristallographie – Crystalline Materials. 74, 1 (1930) 43. Bragg, W.L.: The structure of silicates*. Nature. 125(3152), 510–511 (1930) 44. Pauling, L.: XXII. The Structure of Sodalite and Helvite. Zeitschrift für Kristallographie – Crystalline Materials. 74, 213 (1930) 45. Pauling, L.: The structure of some sodium and calcium aluminosilicates. Proc. Natl. Acad. Sci. U. S. A. 16(7), 453–459 (1930) 46. McCusker, L.: Zeolite crystallography. Structure determination in the absence of conventional single-crystal data. Acta Crystallogr. A. 47(4), 297–313 (1991) 47. Kokotailo, G.T., Lawton, S.L., Olson, D.H., Meier, W.M.: Structure of synthetic zeolite ZSM-5. Nature. 272(5652), 437–438 (1978) 48. Olson, D.H., Kokotailo, G.T., Lawton, S.L., Meier, W.M.: Crystal structure and structure-related properties of ZSM-5. J. Phys. Chem. 85(15), 2238–2243 (1981) 49. Lok, B.M., Messina, C.A., Patton, R.L., Gajek, R.T., Cannan, T.R., Flanigen, E.M.: Silicoaluminophosphate molecular sieves: another

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

new class of microporous crystalline inorganic solids. J. Am. Chem. Soc. 106(20), 6092–6093 (1984) 50. Dent, L.S., Smith, J.V.: Crystal structure of chabazite, a molecular sieve. Nature. 181(4626), 1794–1796 (1958) 51. Reed, T.B., Breck, D.W.: Crystalline zeolites. II. Crystal structure of synthetic zeolite, type A. J. Am. Chem. Soc. 78(23), 5972–5977 (1956) 52. Breck, D.W., Eversole, W.G., Milton, R.M., Reed, T.B., Thomas, T.L.: Crystalline zeolites. I. The properties of a new synthetic zeolite, type A. J. Am. Chem. Soc. 78(23), 5963–5972 (1956) 53. Gramlich, V., Meier, W.M.: The crystal structure of hydrated NaA: a detailed refinement of a pseudosymmetric zeolite structure. Zeitschrift für Kristallographie – Crystalline Materials. 133, 134 (1971) 54. Baerlocher, C.: Proceedings of the 6th International Zeolite Conference. Butterworths, Guildford (1984) 55. van Koningsveld, H., van Bekkum, H., Jansen, J.C.: On the location and disorder of the tetrapropylammonium (TPA) ion in zeolite ZSM-5 with improved framework accuracy. Acta Crystallogr. B. 43(2), 127–132 (1987) 56. Fyfe, C.A., Kennedy, G.J., De Schutter, C.T., Kokotailo, G.T.: Sorbate-induced structural changes in ZSM-5 (silicalite). Journal of the Chemical Society, Chemical Communications. (8), 541–542 (1984) 57. Wright, P.A., Thomas, J.M., Cheetham, A.K., Nowak, A.K.: Localizing active sites in zeolitic catalysts: neutron powder profile analysis and computer simulation of deuteropyridine bound to gallozeolite-L. Nature. 318(6047), 611–614 (1985) 58. Fitch, A.N., Jobic, H., Renouprez, A.: The localization of benzene in a Y-zeolite. Journal of the Chemical Society, Chemical Communications. (5), 284–286 (1985) 59. Mentzen, B.F., Vedrine, J.C.: Crystalline phases of pentasil zeolites formed upon normal-hexane and para-xylene adsorption. Comptes Rendus De L Academie Des Sciences Serie Ii. 301(14), 1017–1020 (1985) 60. Mentzen, B.: Localizing adsorption sites in zeolitic materials by X-ray powder diffraction: pyridine sorbed in B.ZSM-5. J. Appl. Crystallogr. 22(2), 100–104 (1989) 61. Sacerdote-peronnet, M., Mentzen, B.F.: Location of perdeuterated benzene sorbed at low pore-filling in a HMFI material: a neutron powder diffraction study. Mater. Res. Bull. 28(8), 767–774 (1993) 62. Mentzen, B.F., Lefebvre, F.: Flexibility of the MFI SILICALITE framework upon benzene adsorption at higher pore-fillings: a study by x-ray powder diffraction, NMR and molecular mechanics. Mater. Res. Bull. 32(7), 813–820 (1997) 63. Mentzen, B.F., Favre, R., Cauwet, F.: Characterization of guest molecules sorbed in zeolites of known structure. Part VI — localization of methanol at several pore fillings in a highly siliceous MFI silicalite material. Mater. Res. Bull. 29(6), 587–594 (1994) 64. Mentzen, B.F., Gelin, P.: The silicalite/p-xylene system: Part I – flexibility of the MFI framework and sorption mechanism observed during p-xylene pore-filling by X-ray powder diffraction at room temperature. Mater. Res. Bull. 30(3), 373–380 (1995) 65. Mentzen, B.F.: Characterization of guest molecules sorbed in zeolites of known structure. Part IV – localization of para-disubstituted benzene molecules at low sorbate-fillings in a zsm-5 material (SiAl ¼ 159). Mater. Res. Bull. 27(8), 953–960 (1992) 66. Mentzen, B.F., Lefèbvre, F.: The silicalite(MFI)/p-nitroaniline system. I. Location of the sorbed molecules at medium and high porefillings. A study by X-ray powder diffraction, 29Si mas NMR and energy minimization calculations. J. Chim. Phys. 95(5), 1052–1067 (1998) 67. Mentzen, B.F., Sacerdote-peronnet, M.: Characterization of guest molecules sorbed in zeolites of known structure. Part V – localization of para-dichlorobenzene in MFI materials at medium and higher loadings. Mater. Res. Bull. 28(11), 1161–1168 (1993)

559

68. Mentzen, B.F., Sacerdote-Peronnet, M., Bérar, J.-F., Lefebvre, F.: Structural study of an MFI/naphthalene system by X-ray powder diffraction, 29Si MAS n.m.r, and energy-minimization calculations. Zeolites. 13(6), 485–492 (1993) 69. Mentzen, B.F., Lefebvre, F.: Evolution of host/guest interactions during adsorption of tetrachlorethylene (TCET) by silicalite (MFI): a study by X-ray powder diffraction, NMR and molecular mechanics. Mater. Res. Bull. 37(5), 957–970 (2002) 70. Mentzen, B.F., Gélin, P.: Location of p-xylene and cesium cations in ZSM-5 and Cs-ZSM-5: structural evidence for the formation of a π-complex. Mater. Res. Bull. 33(1), 109–116 (1998) 71. Mentzen, B.F.: Modeling diffusion pathways in MFI materials by time-resolved powder diffraction techniques. In: Karge, H.G., Weitkamp, J. (eds.) Studies in Surface Science and Catalysis, vol. 46, pp. 477–484. Elsevier, Amsterdam, Netherlands (1989) 72. Mentzen, B.: Time-resolved powder diffraction as an analytical tool for diffusion studies in microporous topologies. J. Appl. Crystallogr. 21(3), 266–271 (1988) 73. van Koningsveld, H., Tuinstra, F., van Bekkum, H., Jansen, J.C.: The location of p-xylene in a single crystal of zeolite H-ZSM-5 with a new, sorbate-induced, orthorhombic framework symmetry. Acta Crystallogr. B. 45(4), 423–431 (1989) 74. van Koningsveld, H., Jansen, J.C.: Single crystal structure analysis of zeolite H-ZSM-5 loaded with naphthalene. Microporous Mater. 6(3), 159–167 (1996) 75. van Koningsveld, H., Jansen, J.C., Man, A.J.M.D.: Single-crystal structure analysis and energy minimizations of a MFI-type zeolite at low p-dichlorobenzene sorbate loading. Acta Crystallogr. B. 52(1), 131–139 (1996) 76. van Koningsveld, H., Jansen, J.C., van Bekkum, H.: The location of p-dichlorobenzene in a single crystal of zeolite H-ZSM-5 at high sorbate loading. Acta Crystallogr. B. 52(1), 140–144 (1996) 77. van Koningsveld, H., Koegler, J.H.: Preparation and structure of crystals of zeolite H-ZSM-5 loaded with p-nitroaniline. Microporous Mater. 9(1), 71–81 (1997) 78. van Koningsveld, H.: Localization of non-framework species in MFI. J. Mol. Catal. A Chem. 134(1), 89–95 (1998) 79. Rodeghero, E., Martucci, A., Cruciani, G., Bagatin, R., Sarti, E., Bosi, V., Pasti, L.: Kinetics and dynamic behaviour of toluene desorption from ZSM-5 using in situ high-temperature synchrotron powder X-ray diffraction and chromatographic techniques. Catal. Today. 277, 118–125 (2016) 80. Gameson, I., Rayment, T., Thomas, J.M., Wright, P.A.: Studies of model zeolitic catalysts: determination of the site of adsorption of methyl chloride sorbed in zeolite rho by powder x-ray diffraction. J. Phys. Chem. 92(4), 988–991 (1988) 81. Taylor, J.C.: Locations of deuterobenzene molecules in the synthetic zeolite ZSM-5 at 77 K by powder neutron diffraction. Zeolites. 7(4), 311–318 (1987) 82. Taylor, J.C.: Hexadeuteriobenzene locations in the cavities of ZSM-5 zeolite by powder neutron diffraction. J. Chem. Soc. Chem. Commun. (15), 1186–1187 (1987) 83. Goyal, R., Fitch, A.N., Jobic, H.: Powder neutron and X-ray diffraction studies of benzene adsorbed in zeolite ZSM-5. J. Phys. Chem. B. 104(13), 2878–2884 (2000) 84. Kamiya, N., Iwama, W., Kudo, T., Nasuno, T., Fujiyama, S., Nishi, K., Yokomori, Y.: Determining the structure of a benzene7.2silicalite-1 zeolite using a single-crystal X-ray method. Acta Crystallogr. B. 67(6), 508–515 (2011) 85. Parise, J.B.: Structural case-studies of inclusion phenomena in zeolites – XE in rho and stilbene in ZSM-5. J. Incl. Phenom. Mol. Recogn. Chem. 21(1–4), 79–112 (1995) 86. Nair, S., Tsapatsis, M.: The location of o- and m-xylene in silicalite by powder X-ray diffraction. J. Phys. Chem. B. 104(38), 8982–8988 (2000)

26

560 87. Nair, S., Tsapatsis, M.: The location of o- and m-xylene in silicalite by powder X-ray diffraction. J. Phys. Chem. B. 105(6), 1276–1276 (2001) 88. Morell, H., Angermund, K., Lewis, A.R., Brouwer, D.H., Fyfe, C.A., Gies, H.: Structural investigation of silicalite-I loaded with n-hexane by X-ray diffraction, 29Si MAS NMR, and molecular modeling. Chem. Mater. 14(5), 2192–2198 (2002) 89. Gies, H., Marler, B., Fyfe, C., Kokotailo, G., Feng, Y., Cox, D.E.: The combination of synchrotron powder diffraction and highresolution solid-state NMR experiments. J. Phys. Chem. Solids. 52(10), 1235–1241 (1991) 90. Grey, C.P., Poshni, F.I., Gualtieri, A.F., Norby, P., Hanson, J.C., Corbin, D.R.: Combined MAS NMR and X-ray powder diffraction structural characterization of hydrofluorocarbon-134 adsorbed on zeolite NaY: observation of cation migration and strong sorbate
cation interactions. J. Am. Chem. Soc. 119(8), 1981–1989 (1997) 91. Kaszkur, Z.A., Jones, R.H., Couves, J.W., Waller, D., Catlow, R.A., Thomas, J.M.: Locating the sites of sorbed chloroform and dichlorobenzene in a zeolite solid: a synchrotron-based diffraction study of zeolite Y at room temperature. J. Phys. Chem. Solids. 52(10), 1219–1227 (1991) 92. Kaszkur, Z.A., Jones, R.H., Waller, D., Catlow, C.R.A., Thomas, J.M.: Combined Rietveld-molecular dynamics powder diffraction approach to the location of molecules in porous solids: application to 1,4-dibromobutane in zeolite Y. J. Phys. Chem. 97(2), 426–431 (1993) 93. Martucci, A., Rodeghero, E., Pasti, L., Bosi, V., Cruciani, G.: Adsorption of 1,2-dichloroethane on ZSM-5 and desorption dynamics by in situ synchrotron powder X-ray diffraction. Microporous Mesoporous Mater. 215, 175–182 (2015) 94. Rodeghero, E., Pasti, L., Sarti, E., Cruciani, G., Bagatin, R., Martucci, A.: Temperature-induced desorption of methyl tertbutyl ether confined on ZSM-5: an in situ synchrotron XRD powder diffraction study. Minerals. 7(3), 34 (2017) 95. Rodeghero, E., Martucci, A., Cruciani, G., Sarti, E., Cavazzini, A., Costa, V., Bagatin, R., Pasti, L.: Detailed investigation of thermal regeneration of high-silica ZSM-5 zeolite through in situ synchrotron X-ray powder diffraction and adsorption studies. J. Phys. Chem. C. 121(33), 17958–17968 (2017) 96. Wragg, D.S., Bleken, F.L., O'Brien, M.G., Di Michiel, M., Fjellvåg, H., Olsbye, U.: The fast Z-scan method for studying working catalytic reactors with high energy X-ray diffraction: ZSM-5 in the methanol to gasoline process. Phys. Chem. Chem. Phys. 15(22), 8662–8671 (2013) 97. Wragg, D.S., Johnsen, R.E., Norby, P., Fjellvåg, H.: The adsorption of methanol and water on SAPO-34: in situ and ex situ X-ray diffraction studies. Microporous Mesoporous Mater. 134(1), 210–215 (2010) 98. Fyfe, C.A., Joseph Lee, J.S., Cranswick, L.M.D., Swainson, I.: Powder neutron diffraction determination of the structure of the o-xylene/zeolite ZSM-5 complex. Microporous Mesoporous Mater. 112(1), 299–307 (2008) 99. Fujiyama, S., Seino, S., Kamiya, N., Nishi, K., Yokomori, Y.: Adsorption structure of dimethyl ether on silicalite-1 zeolite determined using single-crystal X-ray diffraction. Acta Crystallogr. B. 70(5), 856–863 (2014) 100. Weidenthaler, C., Fischer, R.X., Abrams, L., Hewat, A.: Zeolite rho loaded with methylamines. I. Monomethylamine loadings. Acta Crystallogr. B. 53(3), 429–439 (1997) 101. Weidenthaler, C., Fischer, R.X., Abrams, L.: Zeolite rho loaded with methylamines. II. Dimethylamine loadings. Acta Crystallogr. B. 53(3), 440–443 (1997) 102. Weidenthaler, C., Fischer, R.X., Abrams, L., Hewat, A.: Zeolite rho loaded with methylamines. III. Trimethylamine loadings. Acta Crystallogr. B. 53(3), 444–450 (1997)

P. Beato et al. 103. Lo, B.T.W., Ye, L., Qu, J., Sun, J., Zheng, J., Kong, D., Murray, C.A., Tang, C.C., Tsang, S.C.E.: Elucidation of adsorbate structures and interactions on Brønsted acid sites in H-ZSM-5 by synchrotron X-ray powder diffraction. Angew. Chem. Int. Ed. 55(20), 5981–5984 (2016) 104. Ye, L., Teixeira, I., Lo, B.T.W., Zhao, P., Edman Tsang, S.C.: Spatial differentiation of Brønsted acid sites by probe molecule in zeolite USY using synchrotron X-ray powder diffraction. Chem. Commun. 53(70), 9725–9728 (2017) 105. Chen, T., Huang, B., Day, S., Tang, C.C., Tsang, S.C.E., Wong, K.Y., Lo, T.W.B.: Differential adsorption of l- and d-lysine on achiral MFI zeolites as determined by synchrotron X-Ray powder diffraction and thermogravimetric analysis. Angew. Chem. Int. Ed. 59(3), 1093–1097 (2020) 106. Smith, L., Cheetham, A.K., Morris, R.E., Marchese, L., Thomas, J.M., Wright, P.A., Chen, J.: On the nature of water bound to a solid acid catalyst. Science. 271(5250), 799–802 (1996) 107. Smith, L.J., Davidson, A., Cheetham, A.K.: A neutron diffraction and infrared spectroscopy study of the acid form of the aluminosilicate zeolite, chabazite (H- SSZ-13). Catal. Lett. 49(3), 143–146 (1997) 108. Fujiyama, S., Yoza, K., Kamiya, N., Nishi, K., Yokomori, Y.: Entrance and diffusion pathway of CO2 and dimethyl ether in silicalite-1 zeolite channels as determined by single-crystal XRD structural analysis. Acta Crystallogr. B. 71(1), 112–118 (2015) 109. Alvarez, A.G., Viturro, H., Bonetto, R.D.: Structural changes on deactivation of ZSM-5. A study by X-ray powder diffraction. Mater. Chem. Phys. 32(2), 135–140 (1992) 110. van Koningsveld, H., Jansen, J.C., van Bekkum, H.: The orthorhombic/monoclinic transition in single crystals of zeolite ZSM-5. Zeolites. 7(6), 564–568 (1987) 111. Rojo-Gama, D., Nielsen, M., Wragg, D.S., Dyballa, M., Holzinger, J., Falsig, H., Lundegaard, L.F., Beato, P., Brogaard, R.Y., Lillerud, K.P., Olsbye, U., Svelle, S.: A straightforward descriptor for the deactivation of zeolite catalyst H-ZSM-5. ACS Catal. 7(12), 8235–8246 (2017) 112. Zokaie, M., Wragg, D.S., Grønvold, A., Fuglerud, T., Cavka, J.H., Lillerud, K.P., Swang, O.: Unit cell expansion upon coke formation in a SAPO-34 catalyst: a combined experimental and computational study. Microporous Mesoporous Mater. 165, 1–5 (2013) 113. Wragg, D.S., Johnsen, R.E., Balasundaram, M., Norby, P., Fjellvåg, H., Grønvold, A., Fuglerud, T., Hafizovic, J., Vistad, Ø. B., Akporiaye, D.: SAPO-34 methanol-to-olefin catalysts under working conditions: a combined in situ powder X-ray diffraction, mass spectrometry and Raman study. J. Catal. 268(2), 290–296 (2009) 114. Wragg, D.S., Grønvold, A., Voronov, A., Norby, P., Fjellvåg, H.: Combined XRD and Raman studies of coke types found in SAPO-34 after methanol and propene conversion. Microporous Mesoporous Mater. 173, 166–174 (2013) 115. del Campo, P., Slawinski, W.A., Henry, R., Erichsen, M.W., Svelle, S., Beato, P., Wragg, D., Olsbye, U.: Time- and space-resolved high energy operando X-ray diffraction for monitoring the methanol to hydrocarbons reaction over H-ZSM-22 zeolite catalyst in different conditions. Surf. Sci. 648, 141–149 (2016) 116. Arstad, B., Lind, A., Cavka, J.H., Thorshaug, K., Akporiaye, D., Wragg, D., Fjellvåg, H., Grønvold, A., Fuglerud, T.: Structural changes in SAPO-34 due to hydrothermal treatment. A NMR, XRD, and DRIFTS study. Microporous Mesoporous Mater. 225, 421–431 (2016) 117. Pawley, G.: Unit-cell refinement from powder diffraction scans. J. Appl. Crystallogr. 14(6), 357–361 (1981) 118. Hriljac, J.A., Eddy, M.M., Cheetham, A.K., Donohue, J.A., Ray, G.J.: Powder neutron diffraction and 29Si MAS NMR studies of siliceous zeolite-Y. J. Solid State Chem. 106(1), 66–72 (1993)

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

119. Agostini, G., Lamberti, C., Palin, L., Milanesio, M., Danilina, N., Xu, B., Janousch, M., van Bokhoven, J.A.: In situ XAS and XRPD parametric Rietveld refinement to understand dealumination of Y zeolite catalyst. J. Am. Chem. Soc. 132(2), 667–678 (2010) 120. Smith, R.L., Sławiński, W.A., Lind, A., Wragg, D.S., Cavka, J.H., Arstad, B., Fjellvåg, H., Attfield, M.P., Akporiaye, D., Anderson, M.W.: Nanoporous intergrowths: how crystal growth dictates phase composition and hierarchical structure in the CHA/AEI system. Chem. Mater. 27(12), 4205–4215 (2015) 121. Smith, R.L., Svelle, S., del Campo, P., Fuglerud, T., Arstad, B., Lind, A., Chavan, S., Attfield, M.P., Akporiaye, D., Anderson, M.W.: CHA/AEI intergrowth materials as catalysts for the methanol-to-olefins process. Appl. Catal. A Gen. 505, 1–7 (2015) 122. Somorjai, G.A., Aliaga, C.: Molecular studies of model surfaces of metals from single crystals to nanoparticles under catalytic reaction conditions. Evolution from prenatal and postmortem studies of catalysts. Langmuir. 26(21), 16190–16203 (2010) 123. Norby, P.: Hydrothermal conversion of zeolites: an in situ synchrotron X-ray powder diffraction study. J. Am. Chem. Soc. 119(22), 5215–5221 (1997) 124. Norby, P., Hanson, J.C.: In-situ powder X-ray diffraction in heterogeneous catalysis. In: Rodriguez, J.A., Hanson, J.C., Chupas, P.J. (eds.) In-Situ Characterization of Heterogeneous Catalysts, pp. 121–146. Wiley, Hoboken (2013) 125. Muncaster, G., Davies, A.T., Sankar, G., Catlow, C.R.A., Thomas, J.M., Colston, S.L., Barnes, P., Walton, R.I., O'Hare, D.: On the advantages of the use of the three-element detector system for measuring EDXRD patterns to follow the crystallisation of openframework structures. Phys. Chem. Chem. Phys. 2(15), 3523–3527 (2000) 126. Lo, B.T.W., Ye, L., Tsang, S.C.E.: The contribution of synchrotron X-ray powder diffraction to modern zeolite applications: a minireview and prospects. Chem. 4(8), 1778–1808 (2018) 127. O’Brien, M.G., Beale, A.M., Kuipers, B.W.M., Erné, B.H., Lewis, D.W., Catlow, C.R.A.: Role of germanium on the nucleation and growth of zeolite a from clear solutions as studied by in situ smallangle X-ray scattering, wide-angle X-ray scattering, and dynamic light scattering. J. Phys. Chem. C. 113(43), 18614–18622 (2009) 128. Vistad, Ø.B., Akporiaye, D.E., Lillerud, K.P.: Identification of a key precursor phase for synthesis of SAPO-34 and kinetics of formation investigated by in situ X-ray diffraction. J. Phys. Chem. B. 105(50), 12437–12447 (2001) 129. Grandjean, D., Beale, A.M., Petukhov, A.V., Weckhuysen, B.M.: Unraveling the crystallization mechanism of CoAPO-5 molecular sieves under hydrothermal conditions. J. Am. Chem. Soc. 127(41), 14454–14465 (2005) 130. Egami, T., Billinge, S.J.L.: Underneath the Bragg Peaks: Structural Analysis of Complex Materials, vol. 6, p. 57. Pergamon, Kiddington, Oxford (2003) 131. Milanesio, M., Artioli, G., Gualtieri, A.F., Palin, L., Lamberti, C.: Template burning inside TS-1 and Fe-MFI molecular sieves: an in situ XRPD study. J. Am. Chem. Soc. 125(47), 14549–14558 (2003) 132. Lassinantti Gualtieri, M., Gualtieri, A.F., Hedlund, J.: The influence of heating rate on template removal in silicalite-1: an in situ HT-XRPD study. Microporous Mesoporous Mater. 89(1), 1–8 (2006) 133. Bhange, D.S., Pandya, N.A., Jha, R.K., Ramaswamy, V.: Non-isothermal kinetic studies of the template decomposition from silicalite-1 framework-high temperature X-ray diffraction and thermogravimetric analysis. Microporous Mesoporous Mater. 113(1), 64–71 (2008) 134. Martucci, A., Alberti, A., Cruciani, G., Frache, A., Marchese, L., Pastore, H.O.: Temperature-induced transformations in CoAPO-34 molecular sieve: a combined in situ X-ray diffraction and FTIR study. J. Phys. Chem. B. 109(28), 13483–13492 (2005)

561

135. Loeliger, T.; Brönnimann, C.; Donath, T.; Schneebeli, M.; Schnyder, R.; Trüb, P.: The new PILATUS3 ASIC with instant retrigger capability. In: 2012 IEEE Nuclear Science Symposium and Medical Imaging Conference Record (NSS/MIC), 27 Oct.–3 Nov. 2012; pp 610–615 (2012) 136. Kalantzopoulos, G.N., Lundvall, F., Lind, A., Arstad, B., Chernyshov, D., Fjellvåg, H., Wragg, D.S.: SAPO-37 microporous catalysts: revealing the structural transformations during template removal. Catal. Struct. React. 3(1–2), 79–88 (2017) 137. Briend, M., Shikholeslami, A., Peltre, M.-J., Delafosse, D., Barthomeuf, D.: Thermal and hydrothermal stability of SAPO-5 and SAPO-37 molecular sieves. Journal of the Chemical Society, Dalton Transactions. (7), 1361–1362 (1989) 138. Kalantzopoulos, G.N., Lundvall, F., Thorshaug, K., Lind, A., Vajeeston, P., Dovgaliuk, I., Arstad, B., Wragg, D.S., Fjellvåg, H.: Factors determining microporous material stability in water: the curious case of SAPO-37. Chem. Mater. 32(4), 1495–1505 (2020) 139. Stinton, G.W., Evans, J.S.O.: Parametric Rietveld refinement. J. Appl. Crystallogr. 40(1), 87–95 (2007) 140. Kalantzopoulos, G.N., Lundvall, F., Checchia, S., Lind, A., Wragg, D.S., Fjellvåg, H., Arstad, B.: In situ flow MAS NMR spectroscopy and synchrotron PDF analyses of the local response of the Brønsted acidic site in SAPO-34 during hydration at elevated temperatures. ChemPhysChem. 19(4), 519–528 (2018) 141. Vaughan, G.B.M., Baker, R., Barret, R., Bonnefoy, J., Buslaps, T., Checchia, S., Duran, D., Fihman, F., Got, P., Kieffer, J., Kimber, S.A.J., Martel, K., Morawe, C., Mottin, D., Papillon, E., Petitdemange, S., Vamvakeros, A., Vieux, J.-P., Di Michiel, M.: ID15A at the ESRF – a beamline for high speed operando X-ray diffraction, diffraction tomography and total scattering. J. Synchrotron Radiat. 27(2), 515–528 (2020) 142. Pacchioni, G.: An upgrade to a bright future. Nat. Rev. Phys. 1(2), 100–101 (2019) 143. Weckhuysen, B.M.: Determining the active site in a catalytic process: operando spectroscopy is more than a buzzword. Phys. Chem. Chem. Phys. 5(20), 4351–4360 (2003) 144. Bañares, M.A., Guerrero-Pérez, M.O., Fierro, J.L.G., Cortez, G.G.: Raman spectroscopy during catalytic operations with on-line activity measurement (operando spectroscopy): a method for understanding the active centres of cations supported on porous materials. J. Mater. Chem. 12(11), 3337–3342 (2002) 145. Bañares, M.A., Wachs, I.E.: Molecular structures of supported metal oxide catalysts under different environments. J. Raman Spectrosc. 33(5), 359–380 (2002) 146. Guerrero-Pérez, M.O., Bañares, M.A.: Operando Raman study of alumina-supported Sb–V–O catalyst during propane ammoxidation to acrylonitrile with on-line activity measurement. Chem. Commun. 12, 1292–1293 (2002) 147. Weckhuysen, B.M.: Snapshots of a working catalyst: possibilities and limitations of in situ spectroscopy in the field of heterogeneous catalysis. Chem. Commun. 2, 97–110 (2002) 148. Cortez, G.G., Bañares, M.A.: A Raman spectroscopy study of alumina-supported vanadium oxide catalyst during propane oxidative dehydrogenation with online activity measurement. J. Catal. 209(1), 197–201 (2002) 149. Clausen, S.B.: Combined (Q)EXAFS/XRD: Technique and applications. Catalysis Today. 39(4), 293–300 (1998), ISSN 0920-5861. https://doi.org/10.1016/S0920-5861(97)00115-6 150. Chupas, P.J., Chapman, K.W., Kurtz, C., Hanson, J.C., Lee, P.L., Grey, C.P.: A versatile sample-environment cell for non-ambient X-ray scattering experiments. J. Appl. Crystallogr. 41(4), 822–824 (2008) 151. Johnsen, R.E., Norby, P.: Capillary-based micro-battery cell for in situ X-ray powder diffraction studies of working batteries: a study

26

562 of the initial intercalation and deintercalation of lithium into graphite. J. Appl. Crystallogr. 46(6), 1537–1543 (2013) 152. Norby, P.: In-situ time resolved synchrotron powder diffraction studies of syntheses and chemical reactions. Mater. Sci. Forum. 228-231, 147–152 (1996) 153. Newton, M.A., van Beek, W.: Combining synchrotron-based X-ray techniques with vibrational spectroscopies for the in situ study of heterogeneous catalysts: a view from a bridge. Chem. Soc. Rev. 39(12), 4845–4863 (2010) 154. Topsøe, H.: Developments in operando studies and in situ characterization of heterogeneous catalysts. J. Catal. 216(1), 155–164 (2003) 155. Cats, K.H., Weckhuysen, B.M.: Combined operando X-ray diffraction/Raman spectroscopy of catalytic solids in the laboratory: the Co/TiO2 Fischer–Tropsch synthesis catalyst showcase. ChemCatChem. 8(8), 1531–1542 (2016) 156. Clausen, B.S., Steffensen, G., Fabius, B., Villadsen, J., Feidenhans’l, R., Topsøe, H.: In situ cell for combined XRD and on-line catalysis tests: studies of Cu-based water gas shift and methanol catalysts. J. Catal. 132(2), 524–535 (1991) 157. Wragg, D.S., Akporiaye, D., Fjellvåg, H.: Direct observation of catalyst behaviour under real working conditions with X-ray diffraction: comparing SAPO-18 and SAPO-34 methanol to olefin catalysts. J. Catal. 279(2), 397–402 (2011) 158. Marcus, D.M., Song, W., Ng, L.L., Haw, J.F.: Aromatic hydrocarbon formation in HSAPO-18 catalysts: cage topology and acid site density. Langmuir. 18(22), 8386–8391 (2002) 159. Lo, B.T.W., Ye, L., Murray, C.A., Tang, C.C., Mei, D., Tsang, S.C. E.: Monitoring the methanol conversion process in H-ZSM-5 using synchrotron X-ray powder diffraction-mass spectrometry. J. Catal. 365, 145–152 (2018) 160. Stöcker, M.: Methanol-to-hydrocarbons: catalytic materials and their behavior1 Dedicated to my wife Wencke Ophaug.1. Microporous Mesoporous Mater. 29(1), 3–48 (1999) 161. O’Brien, M.G., Beale, A.M., Jacques, S.D.M., Di Michiel, M., Weckhuysen, B.M.: Closing the operando gap: the application of high energy photons for studying catalytic solids at work. Appl. Catal. A Gen. 391(1), 468–476 (2011) 162. Jacques, S.D.M., Di Michiel, M., Kimber, S.A.J., Yang, X., Cernik, R.J., Beale, A.M., Billinge, S.J.L.: Pair distribution function computed tomography. Nat. Commun. 4(1), 2536 (2013) 163. O'Brien, M.G., Jacques, S.D.M., Di Michiel, M., Barnes, P., Weckhuysen, B.M., Beale, A.M.: Active phase evolution in single Ni/Al2O3 methanation catalyst bodies studied in real time using combined μ-XRD-CT and μ-absorption-CT. Chem. Sci. 3(2), 509–523 (2012) 164. Jacques, S.D.M., Di Michiel, M., Beale, A.M., Sochi, T., O'Brien, M.G., Espinosa-Alonso, L., Weckhuysen, B.M., Barnes, P.: Dynamic X-ray diffraction computed tomography reveals realTime insight into catalyst active phase evolution. Angew. Chem. Int. Ed. 50(43), 10148–10152 (2011) 165. Beale, A.M., Jacques, S.D.M., Gibson, E.K., Di Michiel, M.: Progress towards five dimensional diffraction imaging of functional materials under process conditions. Coord. Chem. Rev. 277-278, 208–223 (2014) 166. Finegan, D.P., Vamvakeros, A., Tan, C., Heenan, T.M.M., Daemi, S.R., Seitzman, N., Di Michiel, M., Jacques, S., Beale, A.M., Brett, D.J.L., Shearing, P.R., Smith, K.: Spatial quantification of dynamic inter and intra particle crystallographic heterogeneities within lithium ion electrodes. Nat. Commun. 11(1), 631 (2020) 167. Li, T., Heenan, T.M.M., Rabuni, M.F., Wang, B., Farandos, N.M., Kelsall, G.H., Matras, D., Tan, C., Lu, X., Jacques, S.D.M., Brett, D.J.L., Shearing, P.R., Di Michiel, M., Beale, A.M., Vamvakeros,

P. Beato et al. A., Li, K.: Design of next-generation ceramic fuel cells and realtime characterization with synchrotron X-ray diffraction computed tomography. Nat. Commun. 10(1), 1497 (2019) 168. Egan, C.K., Jacques, S.D.M., Di Michiel, M., Cai, B., Zandbergen, M.W., Lee, P.D., Beale, A.M., Cernik, R.J.: Non-invasive imaging of the crystalline structure within a human tooth. Acta Biomater. 9(9), 8337–8345 (2013) 169. Vamvakeros, A., Jacques, S.D.M., Di Michiel, M., Matras, D., Middelkoop, V., Ismagilov, I.Z., Matus, E.V., Kuznetsov, V.V., Drnec, J., Senecal, P., Beale, A.M.: 5D operando tomographic diffraction imaging of a catalyst bed. Nat. Commun. 9(1), 4751 (2018) 170. Wragg, D.S., O'Brien, M.G., Bleken, F.L., Di Michiel, M., Olsbye, U., Fjellvåg, H.: Watching the methanol-to-olefin process with time- and space-resolved high-energy operando X-ray diffraction. Angew. Chem. 124(32), 8080–8083 (2012) 171. Dyadkin, V., Pattison, P., Dmitriev, V., Chernyshov, D.: A new multipurpose diffractometer PILATUS@SNBL. J. Synchrotron Radiat. 23(3), 825–829 (2016) 172. Broennimann, C.: The PILATUS detectors: hybrid pixel detectors for synchrotron and industrial applications. Acta Crystallogr. A. 64(a1), C162 (2008) 173. Rojo-Gama, D., Mentel, L., Kalantzopoulos, G.N., Pappas, D.K., Dovgaliuk, I., Olsbye, U., Lillerud, K.P., Beato, P., Lundegaard, L.F., Wragg, D.S., Svelle, S.: Deactivation of zeolite catalyst H-ZSM-5 during conversion of methanol to gasoline: operando time- and space-resolved X-ray diffraction. J. Phys. Chem. Lett. 9(6), 1324–1328 (2018) 174. Haw, J.F., Marcus, D.M.: Well-defined (supra)molecular structures in zeolite methanol-to-olefin catalysis. Top. Catal. 34(1), 41–48 (2005) 175. Wragg, D.S., O'Brien, M.G., Di Michiel, M., Lonstad-Bleken, F.: Rietveld analysis of computed tomography and its application to methanol to olefin reactor beds. J. Appl. Crystallogr. 48(6), 1719–1728 (2015) 176. Agote-Arán, M., Kroner, A.B., Islam, H.U., Sławiński, W.A., Wragg, D.S., Lezcano-González, I., Beale, A.M.: Determination of molybdenum species evolution during non-oxidative dehydroaromatization of methane and its implications for catalytic performance. ChemCatChem. 11(1), 473–480 (2019) 177. Beale, A.M., Lezcano-Gonzalez, I., Slawinksi, W.A., Wragg, D.S.: Correlation between Cu ion migration behaviour and deNOx activity in Cu-SSZ-13 for the standard NH3-SCR reaction. Chem. Commun. 52(36), 6170–6173 (2016) 178. Lezcano-Gonzalez, I., Wragg, D.S., Slawinski, W.A., Hemelsoet, K., Van Yperen-De Deyne, A., Waroquier, M., Van Speybroeck, V., Beale, A.M.: Determination of the nature of the Cu coordination complexes formed in the presence of NO and NH3 within SSZ-13. J. Phys. Chem. C. 119(43), 24393–24403 (2015) 179. Abdala, P.M., Mauroy, H., van Beek, W.: A large-area CMOS detector for high-energy synchrotron powder diffraction and total scattering experiments. J. Appl. Crystallogr. 47(1), 449–457 (2014) 180. AB, E.: https://www.excillum.com/products/metaljet/ 181. Banks, M.: Advanced photon source set for $815m upgrade. https://physicsworld.com/a/advanced-photon-source-set-for815m-upgrade/ 182. Jr, G.R.: Milestone in advanced light source upgrade project will bring in a new ring. 2020 183. Streun, A., Garvey, T., Rivkin, L., Schlott, V., Schmidt, T., Willmott, P., Wrulich, A.: SLS-2 - the upgrade of the Swiss light source. J. Synchrotron Radiat. 25(Pt 3), 631–641 (2018) 184. Falk, T.J.: MAX IV Is Ready to Make the Invisible Visible (2016)

26

Case Studies: Crystallography as a Tool for Studying Methanol Conversion in Zeolites

563

185. Tanaka, H.: Current status of the SPring-8 upgrade project. Synchrotron Radiat. News. 27, 23 (2014) 186. Andersen, C.W., Borfecchia, E., Bremholm, M., Jørgensen, M.R.V., Vennestrøm, P.N.R., Lamberti, C., Lundegaard, L.F., Iversen, BB.: Redox-driven migration of copper ions in the Cu-CHA Zeolite as Shown by the In Situ PXRD/XANES Technique. Angew Chem Int Ed Engl. 56(35), 10367–10372 (2017). https://doi.org/10.1002/anie. 201703808. Epub 2017 Jul 24. PMID: 28670829

Pablo Beato obtained the degree in chemistry from the PhilippsUniversität Marburg and accomplished his PhD in physical inorganic chemistry at the Fritz Haber Institute in Berlin. Since 2005, Pablo has been working at the chemical engineering company Topsoe A/S in Copenhagen, Denmark, where he is currently a lead scientist, in charge of the optical spectroscopy labs, and project manager for catalyst and catalytic process development. His research interests are dedicated to the development of spectroscopic tools to obtain fundamental understanding of the synthesis and the working principles of heterogeneous catalysts.

26

Stian Svelle (born 1975) has been professor at the University of Oslo since 2013. He received his PhD from the same institution in 2004. His main research focus is on fundamental studies of zeolite catalysis, especially the conversion of methanol to hydrocarbons, and more recently the direct conversion of methane to methanol using metalloaded zeolites. He combines kinetic, quantum chemical, and operando studies of individual reaction steps and deactivation phenomena occurring within zeolite catalysts. Another activity is related to the influence of zeolite catalyst crystal morphology and the development of nanostructured zeolites, in both cases aiming for improved performance.

David Stephen Wragg is the manager of the Norwegian National Resource Centre for X-ray Diffraction and Scattering and has specialized in the field of in situ and operando diffraction for over 15 years. His PhD research at the University of St Andrews was on zeolite chemistry and crystallography, under the supervision of Prof. Russell Morris FRS.

Lars Fahl Lundegaard is a senior scientist in R&D at Topsoe A/S where he has been employed since 2009. He is a team leader with responsibility for the X-ray diffraction laboratory and synchrotronrelated activities. He is involved in many different R&D projects, where he primarily contributes with his expertise in characterization, crystal chemistry, and solid-state reactions.

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

27

Mark A. Newton, Patric Zimmermann, and Jeroen A. van Bokhoven

Contents 27.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565

27.2

Recent Technical Developments for Operando XAS Studies of Catalysts and Catalysis: The Induction of Highly Time-Resolved, Single-Shot, FluorescenceYield XAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567 Fast, Single-Shot, Fluorescence-Yield XAS Using a Passivated Implanted Planar Silicon (PIPS) Diode Detector [36] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568

27.2.1

27.3 27.3.1

27.3.2 27.3.3 27.3.4

27.3.5 27.3.6 27.3.7 27.4

Recent Selected Examples of Advanced Operando XAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advances in Spatially Resolved Operando XAS: From Single Metal Nanoparticles to Reactors and from Two- to Three- to Four-Dimensional Experimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operando XAS in Two and Three Dimensions . . . . . . . . . . XAS on the Microseconds Timescale: Investigating Mechanisms of Photocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemistry/Catalysis: Novel Approaches to Combining XAFS with Electrochemical Techniques and Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Combined XAS and Photo-Electro-Catalysis . . . . . . . . . . . . Operando XAS Goes Soft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operando Studies Using Laboratory XAS Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

569

569 573 575

578 583 584 587

Checks, Balances, and Outlook . . . . . . . . . . . . . . . . . . . . . . . . 590

M. A. Newton (*) Department of Chemistry and Applied Biosciences, Institute for Chemical and Bioengineering, ETH, Zürich, Switzerland Laboratory of Nanochemistry for Energy (LNCE), EPFL Valais Wallis, Switzerland e-mail: mark.newton@epfl.ch P. Zimmermann Laboratory for Catalysis and Sustainable Chemistry, Paul Scherrer Institute (PSI), Villigen, Switzerland e-mail: [email protected] J. A. van Bokhoven Department of Chemistry and Applied Biosciences, Institute for Chemical and Bioengineering, ETH, Zürich, Switzerland Laboratory for Catalysis and Sustainable Chemistry, Paul Scherrer Institute (PSI), Villigen, Switzerland e-mail: [email protected]

27.5

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 594

Abstract

In this chapter, we review recent developments in the application of operando X-ray absorption spectroscopy (XAS) and look forward to how the outcomes of present synchrotron upgrades might extend, enhance, or otherwise affect such studies. We will consider some recent developments in beamline, detector technology, cell designs, and experimental arrangements, which are bringing new capacity to operando XAS studies and extending them into new areas of catalysis. Specific emphasis will be placed on developments in operando XAS as applied to electrocatalysis and photocatalysis, along with the different challenges that they present, and we will, in some cases, also consider methodological developments in related fields. We shall also investigate how operando research is venturing into the spectroscopically rich “soft X-ray” (sub ca. 2 keV) regime. Lastly, we shall consider the rise of laboratory XAS spectrometers and assess their role in making new types of operando study possible and how they may fit into the wider picture of operando XAS resources. Keywords

Catalysis · Photocatalysis · Electro-catalysis · Operando · X-ray absorption spectroscopy EXAFS · XANES · Fuel cells · Fluorescence

27.1

Introduction

Just over 20 years ago, the term “operando” was coined by Guerrero-Perez and Banares [1]. This term describes an approach to experimentation in catalysis that goes beyond

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_27

565

566

the concept of “in situ.” Axiomatic to operando study is that the material being considered is not just “in place” but is also “working” under conditions that are relevant to the catalytic process of interest. Since its introduction, this concept has become a guiding principle within catalysis that has driven much development in terms of how catalysts and catalytic processes are investigated. Within the panoply of resources that can be applied in an operando manner, synchrotron sources have been at the forefront of many developments. The ability of X-rays to penetrate matter makes them very well suited to the study of catalysts from wide-ranging points of view and circumstances. These points of view cover a range of length and timescales, alongside the experimental conditions required to satisfy the axioms of operando study; and, within the numerous approaches to study that X-rays permit, X-ray absorption spectroscopy (XAS) remains at the forefront of catalysis research [2–4]. XAS is founded upon the photoelectric effect, wherein electromagnetic radiation of sufficient energy excites electrons into unoccupied electronic states or into the continuum (ionization). The tunable nature of synchrotron X-rays means that XAS can be used to probe virtually all the elements in a selective manner, through excitation of the core levels of the element in question. The information that XAS provides depends upon the energy of excitation relative to the intrinsic energy of the core level that is excited, as this determines what may subsequently happen to the excited electron. If the energy of excitation is insufficient to for the element to be ionized completely, the electron interacts with the local density of unoccupied states and may do so in a spin-selective manner. As a result, these electrons, within the bounds permitted to them by their inelastic mean-free path, sample the local electronic structure. In turn, therefore, these electrons report upon valence and geometric aspects of the bonding of the element under study, the most common of which is the oxidation state of the element. This spectral region, which in general extends to no more than ca. 50 keV from the excited edge, is referred to as XANES (X-ray absorption near-edge structure) or NEXAFS (near-edge X-ray absorption fine structure). If the energy of the applied X-rays is sufficient to result in ejection of the photo-electron into the surrounding continuum then, the further one goes in energy from the elemental edge, and the photoelectron carries with it more kinetic energy, the more the resulting X-ray absorption is dominated by the physics of electron-atom scattering. In this region of kinetic energy, the electron inelastic mean-free path goes through a minimum of a few angstroms, at ca. 100 keV, before rising monotonically the farther away from the edge one probes in energy. As a result, the electron-atom scattering that dominates the behavior of the electrons is less subject to multiple scattering events (though they still play a role which, in certain cases, cannot be neglected), and, as a result of this relative simplicity, analytical solutions which can accurately

M. A. Newton et al.

describe this behavior have long since been arrived at. This spectroscopic regime reports upon the local structure (number and types of atoms in the immediate neighborhood of the element under study) and is known as EXAFS (extended X-ray absorption fine structure). The details of the fundamental physics of the processes occurring in each of these regions, and how they are analyzed, are beyond the scope of this work, and the interested reader is therefore referred elsewhere [2–4]. Suffice it to say, however, that together these two regions of the XAS spectrum provide a great deal of elementally specific information regarding physical and electronic structure, and can be used to specify the nature of chemical speciation and how it may change as result of a catalytic process. Hard X-ray XAS measurements are typically made in two distinct ways, transmission and fluorescence, for which the optimal geometries are different and their use is subject to, historically, rather different limitations. We note that XAS can also be measured using the direct detection of electrons (total or partial electron yields). As we shall see in a later section, these latter methods come into their own as the energy of the X-rays is decreased into the so-called “tender” (ca. 2–4 keV) and then “soft” X-ray ( ca. 20 seconds) timescales. The PIPS detector therefore provides a cheap and versatile solution to the problem of measuring fast XAFS from otherwise intractable systems. As ever, however, there are some caveats. Principal among them is that the PIPS comes with no intrinsic energy resolution, and in systems wherein multiple fluorescence lines can overlap, this could be problematic. However, and to date, the implementation of this sort of detector for fast fluorescence XAS might be regarded as the most general solution thus far arrived at. It is also worthy of note that very fast, photon counting, 2D detectors, such as the Pilatus [37], could also be used to the same ends and, indeed, are central to the HEROS approach [32–34]. These sorts of detectors are extremely fast and large (vis-à-vis solid angle of detection) and have, to a certain degree, the ability to “window out” fluorescence X-rays below the energy of those of interest. That said, these are also very expensive detectors, which may impede their widespread adoption as “simple” fluorescence detectors. A further potential detection scheme, which is a technique in itself, is that of X-ray excited optical luminescence (XEOL) [38–42]. XEOL reports on core-level excitation events, via multistep cascade processes, in the optical part of the electromagnetic spectrum. This method is highly amenable to both time-resolved and imaging studies and often used in tandem with more conventional XANES. To date, and as far as we are aware, however, this approach has not seen application in the field of catalysis.

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

b)

Ni K-edge XANES

c)

k3·F(k) (a.u.)

Absorption (a.u.)

a)

569

100 s 60 s 30 s 10 s

100 s

60 s

60 s

30 s

30 s

10 s 5s

5s 3s

1s

1s

5 Magnitude Imaginary part

Untreated 8320

8340

8360

8380

2

8400

27

10 s

3s

Untreated single spectrum (0.5 s)

5s 3s 1s

100 s

| F(R)| (Å–4)

27

4

6

10

8

12

0

k (Å–1)

Energy (eV)

d) 100

1

2

3

4

Radial distance (Å)

e)

References

O2 Absorption (a.u.)

Fraction (%)

80 60 Ni0 40

NiO Nin+oct

20

1 Ni0 LaFe0.8Ni0.2O3 n+ (Nioct ) 0

0 0

1

10 Time (min)

Fig. 27.1 Summary of results obtained from a fast, fluorescence yield XAS study of an LaFe0.8Ni0.2O3 catalyst using a PIPS detector. (a) Dependence of signal to noise for Ni K-edge XANES as a function of time per spectrum. (b) Quality of Ni k3-weighted Ni K-edge EXAFS as a function of spectral exposure time. (c) As (b), but using the Fourier transform representation of the K3-weighted EXAFS. (d) Time-resolved

27.3

NiO

Recent Selected Examples of Advanced Operando XAS

In this section, we shall look at some of the varied ways in which operando XAS has developed in the last few years in relation to the study of catalysts. Where necessary, a detour into other areas of research will be taken to illustrate a method that could well be applicable to the study of working catalytic materials but which has yet to be applied within this arena. Furthermore, though not exclusively, we shall focus somewhat on approaches that cannot make use of transmission XAS and for which different solutions to enable study have had to be arrived at.

100

8320

8340

8360

8380

Energy (eV)

in situ speciation of the nickel from linear combination fitting of the Ni K-edge XANES during re-oxidation of a pre-reduced LaFe0.8Ni0.2O3 at 650
C in 20 vol% O2/Ar. (e) Ni K-edge XANES spectra of reference components used for linear combination fitting. (Reproduced, with permission from IUCR, from Ref. [36])

27.3.1 Advances in Spatially Resolved Operando XAS: From Single Metal Nanoparticles to Reactors and from Two- to Three- to Four-Dimensional Experimentation In addition to multiple viewpoints, catalysis is a discipline that presents a very wide range of length and timescales, over which one would like to have a detailed insight into. In the case of length scales, these range from the molecular, wherein the intrinsically local nature of XAS can provide fundamental information regarding structure and speciation, through to the dimensions of laboratory-scale reactors, wherein XAS probes can also be used to derive a great deal of process-relevant

570

information. Furthermore, catalyst bodies or reactors are three-dimensional in nature, and, therefore, along with oneor two-dimensional assessments of the behavior of a sample within a reactor, obtaining three-dimensional information is highly sought after. The next three examples therefore explore aspects of this sort of XAS-based research.

One-Dimensional Spatial Investigation of Reactor Beds One-dimensional spatial interrogation of catalyst beds and reactors has been demonstrated for a considerable time [42–53], and numerous examples now exist where, for instance, the nature of reaction light-off and the propagation of catalytic wave fronts have been investigated using XAS. There are also some examples of the use of spatially resolved XAS to investigate a time line [35] or that have utilized chemistry that is then probed by XAS to assess whether or not temperature gradients exist within a given reactor [53]. Figure 27.2a–d gives a salient example of the diverse information that time- and space-resolved XAS can yield when applied to studying a chemical reactor. This example is not per se catalytic but aims to understand the effects of individual components that are typical of, and, in the case of a base, obligatory components of, carbon-carbon coupling catalysis, within the paradigm of a single-pass, plug-flow reactor [48–52]. Figure 27.2a illustrates how the spatiotemporal mapping is achieved, along a catalyst bed held within a reactor system previously established as being isothermal and yielding true plug-flow conditions [48]. Once under a solvent (aqueous ethanol) flow, the sample is heated at 1 Kmin1, while the evolution of the palladium is continuously monitored using Pd K-edge EXAFS by repeated stepping through the different positions within the bed from the inlet to the outlet Figure 27.2b shows a summary of some of the information restored from analysis of the spatiotemporal XAS as a function of temperature (red line in each panel), time, and space, for four palladium catalysts supported upon two different aluminas, carbon and titanium dioxide. N1PdPd refers to the EXAFS Pd-Pd coordination number at around 2.75–2.8 Å that is indicative for the formation of nano-particulate fcc Pd metal or Pd hydrides depending on the distance [54–56]. What emerges from these studies is that the different samples behave in different ways under the influence of the solvent flow as they are heated, yet the end points in each case (in terms of N1PdPd) are very similar, save for the gradients in the palladium phase that appear under the influence of the solvent mixture alone. These gradients manifest (Fig. 27.2b) as sharp changes in the value of N1PdPd that are very pronounced in three of the four samples. These low values of N1PdPd result from the evolution and maintenance of a phase/ speciation gradient, wherein PdII persists at the inlet of the

M. A. Newton et al.

bed, while Pd0 dominates at the outlet. Analysis of the PdPd bond distances associated with the Pd0 phases further indicates that the reduced palladium formed from interactions with the solvent is of PdHx nature, rather than hydrogenfree, metallic palladium An apparent exception is the Pd/Al2O3 (JM325) case (Fig. 27.2b, panel (c)). EXAFS shows that, uniquely among the catalysts tested, this material is inherently comprised of very highly dispersed metallic palladium nanoparticles, rather than nano-particulate PdO. In this case, what the gradual rise in N1PdPd is reflecting is agglomeration of these small entities into significantly larger PdHx particles at the behest of the solvent molecules with concurrent loss of the very high initial dispersion of the palladium [51]. As this reactor has been verified to be isothermal [48], these gradients result from reaction of the solvent with the palladium. This solvent mixture, which is consistently highly rated in solvent selection guides produced by the pharmaceutical industry [57–59], is therefore not a reactively neutral element in respect to the supported palladium. Lastly, panels (C) and (D) show the effects upon the levels of palladium (from the magnitude of the Pd K-edge jump) seen to be present at both the reactor inlet and outlet as the sample is heated. Under the aqueous-ethanolic solvent alone, the edge jump itself does not to change throughout the experiment. However, with the addition of 0.1 M K2CO3, a base commonly used in carbon-carbon coupling catalysis, the levels of Pd present at the inlet of the bed are seen to decrease, whereas at the outlet less change is observed. For one of the catalysts studied (Pd/Al2O3 (JM325)), panel (D) summarizes how, at different positions in the bed, the palladium edge jump varies. This shows that the addition of the base elicits leaching and mass transport of the palladium in the direction of flow and, as XAS is quantitative in this respect, permits an assessment of the rates of leaching/redeposition occurring along the catalyst bed [51]. This is important as, in general, the pernicious problem of palladium leaching in carbon-carbon coupling catalysis has always been thought of as being due to the presence of halogens (in the reactant molecules) [60–62]. In this case, however, as these are not present, leaching can only be due to the actions of the base on the reduced metal nanoparticles as and when they are present, either innately (as in JM325) or as a result of the reductive action of the solvent. Subsequent investigations have further shown that the degree to which a system leaches, and indeed the fate of the support material itself, under the influence of the basified feed, can vary very considerably [52]. Experiments of this type, which braid the operando philosophy with space- and time- resolved XAS, can be extremely informative as to how any given catalyst may develop within a given process situation. In this case, the catalytic process involves both solvents and adjunct materials that are either

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

571

27

Fig. 27.2 (a) Schematic illustration of a plug-flow approach to spatiotemporal monitoring of palladium speciation along a catalyst bed. (b) Spatiotemporal development of four different supported palladium catalysts (supported upon Al2O3, carbon, or TiO2, as indicated) during heating under a flow of 50:50 ethanol/water and as determined from analysis of Pd K-edge EXAFS (Pd-Pd coordination indicative of fcc Pd nanoparticles (N1PdPd)) recorded as a function of axial position within the catalyst bed and temperature. The temperature of the system is given by the red curve. (c) Development of Pd K-edge EXAFS (not

normalized) for a Pd/Al2O3 (JM325) sample in the absence and presence of 0.1 M K2CO3 and for the inlet to the catalyst bed and at the outlet as a function of temperature. (d) Palladium leaching behavior for PdA (JM325 as received): changes in Pd K-edge edge jump (Δedge). In each case, three positions within the bed are indicated: green, inlet; red, middle; and blue, outlet of the bed. (Reproduced with permission from Cat. Sci. Tech., 2016, 6, 8525–8531 and Cat. Sci. Technol., 2020, 10, 466–474, The Royal Society of Chemistry [50, 51])

highly preferred on the basis of the metrics used to determine the “green” credentials of the solvent, [57–59] or required to be present to facilitate the desired chemistry. Other notable examples of this sort of approach, also dealing with palladium-based catalysis, have recently appeared in respect of the catalytic production of hydrogen peroxide at pressures up to 80 bar, wherein equally a great deal of process-related and fundamental information regarding how the catalysis proceeds has also been elucidated [63, 64].

“Nano-focus” XAS: Toward Interrogating Single Supported Nanoparticles At the other end of the spectrum of length scales, X-ray beam foci of a few tens of nanometers have recently become achievable. This leads to the possibility of the selective study of individual, nano-sized elements within a catalyst. By definition, such a situation exists at the very extreme of “dilution,” but, nonetheless, experiments that aim toward the study of single adsorbed nanoparticles have been undertaken. Figure 27.3

572

M. A. Newton et al.

a)

(i)

Gas inlet

(ii) (1)

10 mm 25 mm

(2)

Si frame SiNx window (4)

60 mm

X-ray fluorescence out

(3) (5)

(6) X-ray in

Gas outlet

b) (ii)

(i)

200 nm

50 nm

c)

24300

Pd foil reference A-1 under H2 atmosphere A-2 under H2 atmosphere 24325

24350

24375 24400 Energy (eV)

24425

(2)

Normalized absorption (a.u.)

Normalized absorption (a.u.)

(1)

(3)

Pd foil H2 O2 CO PdO

24450

Fig. 27.3 (a) Overall setup for nano XAS measurements: (i) schematic of measurement cell; (ii) set up as mounted on ID16B at the ESRF, Grenoble, France. (b) TEM characterization of the palladium samples

24300

24325

24350 24375 24400 Energy (eV)

24425

24450

measured with average particle sizes (dp) and inter-particle separations (rnm) of: (i) dp ¼ 2.3  0.6 nm, rnm ¼ 18.6  5.1 nm) and (ii) dp ¼ 5.6  1.1 nm, rnm ¼ 92.1  18.1 nm). The red square given in (ii) shows the

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

shows some results from an experiment designed to demonstrate the feasibility of observing working catalysts using nano-sized X-ray beams, such that either very small ensembles or even individual catalysis particles can be reported upon [65]. In this study, two different palladium samples (supported upon planar (0.5  0.5 mm) silicon nitride) were synthesized to yield either very small, but densely deposited, palladium nanoparticles or larger, but more dispersed, palladium particles, which were then probed using a ca. 400  400 nm2 X-ray beam. For comparison, a further sample was made using silicon nitride powder and measured using a standard transmission EXAFS. While very much a feasibility study, this work shows that XANES can be measured under reactive environments with very high spatial resolution (ca. 400  400 nm2) that, in the systems studied, contain no more than ca. 20 individual palladium nanoparticles. Moreover, the XANES obtained is of a sufficient quality that distinct differences between different catalysts and different gaseous treatments (Fig. 27.3c) can be observed.

27.3.2 Operando XAS in Two and Three Dimensions While the previous examples (sections “One-Dimensional Spatial Investigation of Reactor Beds” and “Nano-focus” XAS: Toward Interrogating Single Supported Nanoparticles”) showed, firstly, how XAS can be used at a macroscopic level to interrogate a catalytic system and, secondly, that spectroscopically addressing individual supported nanoparticles is no longer an implausible notion, the next examples illustrate possibilities for detailed mapping, in two and three dimensions, of working materials. μ-XAS probes, which have spatial resolutions of a few microns or less, have led to the possibility of specific assessment of element distributions and speciation within catalyst bodies and even single catalyst particles. Such measurements have been achieved using both energy-dispersive [66–69] and scanning XAS [70–74] variants and in both 2D and 3D (tomographic) senses. For a more detailed elucidation of the status of this field, the reader is referred to the subsequent chapters within this book dedicated to this subject [75, 76]. Here we shall consider two operando examples of these types

573

of approach, and the information that they may yield, for which in one case we look beyond mainstream catalysis and to the world of battery research. The first example considers the two-dimensional speciation of iron within a battery that results from repeated charging and discharging of the device. The second investigates the distribution and speciation of platinum within a polymer electrolyte membrane (PEM) fuel cell catalyst using a three-dimensional, computed-tomographic (CT) approach to XANES.

Quantifying Changes in Iron Speciation in LiFePO4 Battery Materials Using XAS Imaging The study of spatial changes in the speciation of iron as a function of charging and discharging due to Katayama et al. [77] is nicely illustrative of how two-dimensional XAS imaging on the tens of microns scale can yield significant insight into the distribution and speciation of an element within a functional material. Moreover, this experiment represents a rare example of a full-field approach to X-ray imaging, as opposed to the μ-XAS approaches alluded to above, which has some potentially significant benefits compared to the μ-XAS method. In this example, the spatial resolution of the experiment is not determined by the intrinsic focal spot size of the X-ray beam but by the pixel size of the detector employed (10  10 μm2) [77–79]. In this full-field approach, each pixel, after various corrections, yields a single XAS spectrum, and therefore image collection depends only upon the time required to accumulate sufficient spectral statistics. This is to contrast with the far more common use of micron-sized X-ray beams [66–76], which may also be used to build up an image in a step-by-step manner. In these latter types of imaging, the rapidity of image retrieval is a function of the time required to obtain a suitable spectrum but also the time required to physically move the sample from point to point to build up the XAS speciation map. In the full-field case, no movement of the sample is required, and therefore the speed of image acquisition depends only upon the time to obtain suitable statistics. Moreover, the use of a relatively large, and therefore not so brilliant, X-ray beam to illuminate the sample [77–79] also minimizes the possibility that the X-rays themselves might influence the oxidation states that are to be investigated and mapped. The case in point is founded upon the reliable detection of the partitioning of iron into FeII and FeIII states, and, as FeIII could all too easily be subject to X-ray-induced

ä Fig. 27.3 (continued) approximate footprint (ca. 400  400 nm2) of the X-ray beam obtained at ID16B. (c) (i): normalized XANES spectra of sample a-1 and a-2 against a palladium foil reference. (ii): XANES spectra of sample a-1 under different atmospheres, including Pd foil and

PdO standard; arrows point out the Pd XANES features at: (1) 24.36 keV; (2) 24.38 keV; (3) 24.42 keV. (Reproduced from Catal. Struct. & React., 2017, 3, 63 – 70, [65], copyright Taylor Francis, 2017)

27

574

M. A. Newton et al.

reduction if the source is too powerful, the validity of the measurements depends strongly on this possibility being avoided. As such, this low-brilliance, full-field approach to imaging could be regarded as preferred in this instance. The oxidation of FeII to FeIII is, within the operation of the battery material, indicative both of the levels and location of the charging/discharging events occurring in the cathode but also, in an indirect way, the degree of lithiation/delithiation present as these processes proceed. As can be seen from Fig. 27.4, at rest (0) % charging (I(b), II(a)), the sample is essentially comprised of entirely FeII, whereas when fully charged (II (b)) it is 100% FeIII. The change from FeII to FeIII, and therefore the distribution of charge, revealed by the mapping experiment is, however, decidedly inhomogeneous, despite the apparent uniformity of iron speciation achieved in both fully discharged and fully charged states. One potential source of the inhomogeneous charging of the material – that of significant nonuniformity in the iron component – can be ruled out by the application of this twodimensional XAS imaging. As this method allows a parallel assessment of both the net absorption of X-rays and the oxidation state of the iron at any given point within the field of view, it can be shown [77] that the localization of charging/discharging has no relation to any spatial variations in the concentration of the iron. This method further shows

I.

Operando Computed Tomography (CT) XANES of Structure and Speciation in Pt Cathodes in a PEM Fuel Cell Having moved from one-dimensional operando assessment of working materials through to a two-dimensional XAS imaging approach, we move into the rendering of structure and speciation of an active phase using tomographic methods based upon XAS. In recent years, operando tomographic

II. Fe speciation during successive charging discharging cycles

Fe speciation as a function of charging capacity

b) 0%

that the charging and discharging processes follow the same reactive channels, the iron speciation observed in one process being, to a high degree, the negative of the other. In spite of the high degree of uniformity achieved in the distribution of iron within this device, there exist routes or channels that are of variable resistance, and it is these, rather than any other phenomenon (such as differences in pressure across the device), which result in the observed nonuniformity of charge distribution during operation. On the basis of these data, therefore, the authors argue that the inhomogeneity, and the radial manner in which charge propagates from specific areas, depends on how particles of LiFePO4 – that have a variable conductivity – are distributed within the carbon matrix used to enhance conductivity. As these particles themselves are stationary, these low-resistance pathways are maintained from cycle to cycle.

d) 1st discharge

b) 100%

c) 25%

f) 2nd discharge

7715 7719 E0 (eV)

State of charge

500 μm

d) 50% Normalized absorbance

1

a)

0.5 0 0 0 0

7100 7120 7140 7160 E (eV) e) In-situ experiment 0% 10% 20%

a) 0%

30%

40%

b)

100% c)

50% 0%

d)

e)

f)

a) 500 μm Cycle c) 1st charge

e) 2nd charge

50%

1 mm

Fig. 27.4 Fe K-edge XAS imaging of a LiFePO4 cathode material as a function of charging and discharging cycles. (a) Changes in distributions of iron in differing oxidation states (FeII (blue)/FeIII (red)) as a function of the degree (as indicated) of charging capacity. (b) Iron

chemical state maps during successive charge/discharge cycles. (Reproduced with permissions from J. Power Sources, 2014, 269, 994–999, [77], copyright Elsevier, (2014))

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

methods have developed greatly [67–76]. The principal methods used in catalysis have, however, been founded upon X-ray scattering [72, 80–84]. This is not to say, however, that XAS has been absent from this development into three-dimensional imaging, but to date there exist relatively few operando examples of this very powerful method. Figure 27.5 gives a picture of how such measurements are accomplished, in this case to permit the operando study of Pt speciation, and how it changes in three dimensions, within the structure of a PEM fuel cell subject to accelerated degradation tests (ADT) [85, 86]. In CT, a three-dimensional representation of the catalyst body is built up through rotating the sample (Panel I (A and B)) and at each angle collecting a XAS slice. These slices (Panel I (C)) may then be used as the basis for reconstructing the three-dimensional nature of the sample according to whatever information may be extracted from the method applied. For XAS, this may be the overall texture, from absorption measurements made below the edge of the element of interest (experiment A in panel II); the distribution of the element being addressed, from a difference made between measurements made above and below the edge of interest (experiment B in panel II); the oxidation state of that element, from the position of the edge (experiment C in panel II); and even (not shown) the coordination number and bond distance associated with the Pt from collection of complete EXAFS spectra and their analysis. As can be seen from Fig. 27.5, this approach yields a wealth of information that shows how long-term operation affects all the aspects listed above and how both lateral (2D) and depth-profiling (3D) information may be restored from CT-XAS. Operation leads to profound changes in the morphology of the cathode that manifest as apparent widening and deepening (in terms of absorption contrast below the Pt L3 edge, Panel II, A1 and A2) of preexisting, nonhomogeneous areas of the cathode which extend into the Nafion membrane separating the cathode from the anode. These are accompanied by (Panel II, B1 and B2) significant changes in the lateral distribution of Pt that again extend into the Nafion membrane (B1 and B2 Z′ ¼ 45 μm). It is interesting to note that the largest change in the distribution of platinum resulting from the ADT appears to occur within the center of the catalyst layer (B1, B2, Z′ ¼ 10 μm). This suggests a loss of platinum from this region, which is most evidently associated with the cracks and fissures already apparent from the pre-edge absorption measurement. What is more, whereas prior to ADT there was precious little evidence of the presence of platinum in the Nafion membrane, the post-ADT measurement shows that, at certain lateral positions in the membrane, measurably enhanced levels of platinum can now be found. Moreover, these areas appear to correlate with preexisting areas of high platinum density at the cathode membrane

575

boundary (B1, B2, Z′ ¼ 20 μm). In other words, this experiment is able to pinpoint, with considerable accuracy, how the distribution of platinum has changed during operation and that the Pt itself may be far more mobile than would otherwise be desired. Lastly, experiments C1 and C2 address the platinum oxidation state and how this has changed within different parts of the device as a result of ADT. Before ADT, and within both the catalyst itself and at the interface with the Nafion, the platinum is found to be of predominantly mixed valence, save to a few areas of high density that are more metallic in nature. However, post-ADT the map of platinum valence state shows that it is almost uniformly metallic, with only remnants of higher oxidation states remaining clustered around the edges of the areas of low X-ray absorption that appears as “cracks” in the structure of the material. This example nicely illustrates the power of approaches that can result in a vast amount of information regarding the three-dimensional behavior of materials such as catalysts and how it may be related to aspects of their function. More specifically, these measurements can reveal a variety of behavior within, in this case, the platinum phase and are able to determine whether such behavior should be promoted or avoided. Lastly, in this section, a further imaging method, that of X-ray spectro-ptychography, is also worthy of mention, though, to our knowledge, it too has yet to be applied in an operando sense. X-ray ptychography is a high-spatial-resolution (to sub ca. 5 nm depending on the energy of the X-rays) method based upon coherent diffraction [87]. However, as X-rays from a synchrotron are continuously tunable, it can be used to retrieve highly spatially resolved XANES maps of chemical speciation. As such, it has been used to map, in a very precise manner, variations in the concentration and chemical speciation occurring within lanthanum and iron phases in fluid catalytic cracking catalysts, as a function of the size of the individual catalyst particles [88]. It has also been used to study the heterogeneity of redox speciation of cerium in Pt/Ce2Zr2Ox systems to understand the factors that underpin the high oxygen storage capacity (OSC) of this important support material [89]. The upgrades in synchrotron technology, which are now occurring widely among the world’s X-ray laboratories, mean that this type of approach will only become more powerful and widespread and offer considerable opportunities for more advanced and operando studies of this nature.

27.3.3 XAS on the Microseconds Timescale: Investigating Mechanisms of Photocatalysis The ability to follow the progress of chemical reactions, report quantitatively upon changes in structure and

27

576

M. A. Newton et al.

CT-XAS set up and method Rotation axis: q

(A)

(B)

(C)

Rotation axis: q (r80 degree)

I (E) image

Z axis

I

X-rays Flat object

Y, Zc

Angle (r80°, 0.1° step)

PEFC cell

I0(E) image

X-rays

X, Xc

X axis

Z, Yc Energy scan at Pt Ln-edge (11.451-11.631 keV; 184 pts)

Idark image

II

Operando CT-XAS before and after ADT (A1) Before ADT

(A2) After ADT

(B1) Before ADT

(B2) After ADT

(C1) Before ADT

(C2) After ADT

Yc 100 Pm Xc Zc Zc= 10 Pm

Zc= 20 Pm

Zc= 45 Pm

Pt0+

Pt2+

(D)

Intensity

Pt density

Fig. 27.5 Panel I: CT-XAS setup and method: (A) Coordinates for CT measurement and reconstruction (XYZ and X′Y′Z′) axes indicating the coordinates in the experimental setup and reconstructed data, respectively. (B) The PEM fuel cell for operando CT-XAFS measurements. (C) Illustration of the CT-XAFS measurement protocol, in this case comprised of 184 datasets recorded at X-ray energies around PtLIII-

1.0 0.9 0.8 0.7 0.6 OCV

0

ADT /cycle number 5000 10000 15000 20000

1 cycle

80 70 60 50 40 30

ECSA (m2 g-Pt –1)

Pt valence Cell voltage (V)

Center of Nafion membrane

Interface with Nafion membrane

Center of cathode catalyst layer

0

0 6 12 18 24 30 Time (s)

edge. Panel II: Operando CT-XAS before and after ADT; threedimensional and X′-Y′-cross-sectional images at Z′ ¼ 10 μm (center of cathode catalyst layer), 20 μm (interface of cathode catalyst layer and Nafion membrane), and 45 μm depths (center of Nafion membrane) of (A) morphology reconstructed by XCT data measured 11.497 keV, (B) the distribution of the Pt catalyst, and (C) the Pt oxidation state in

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

coordination around a reactive element in real time, and accurately determine reaction kinetics is one of the many features of XAS that makes it so attractive as a means to understand how catalysts work. Both modern quick-scanning [6, 7, 11] and dispersive [5, 8–10] sources are capable (in transmission) of reporting upon reactive chemistry on the timescales of a few tens of milliseconds in a single shot. However, there is a great deal of catalytic chemistry that occurs on timescales which are much faster than this, and therefore the question arises as to how to access the details of such structural-reactive chemistry. The answer is to adopt the long-known step-scanning approach for cases where a reactive “pump” can be repeatedly and rapidly applied (for instance, [12–14]). This method is extremely well suited to investigating the very fast events (from the pico- to the microsecond range) that follow photoexcitation and which are the basis of photocatalysis. In recent years, a number of examples of the application of time-resolved XAS studies of photochemical or photocatalytic processes have appeared [90–98]. Here, we restrict ourselves to the microsecond regime [94–97] and light-driven water splitting catalysis based upon cobalt coordination compounds. Though much faster experimentation of this sort has long been possible, the microsecond regime is also rich territory for investigation and one that lies in between the ultrafast (pico- and nanoseconds) and the “kinetic” (milliseconds and upward) regimes, the latter of which can (in transmission XAS) be addressed in a single shot without recourse to a step-scan methodology. Aside from the requirement for high time resolution, these sorts of investigations are made particularly challenging by the intrinsic dilution of the species of interest (in the given example, [Co] ≈ 1 mM), that redox-capable organometallics in solution can be rather sensitive to the presence of the X-rays, and that photo-stimulation generally only leads to a very small fraction of the photo-activated species being formed. As such, these measurements must be conducted using fluorescence detection and continuous flow, and often rely upon the interpretation of extremely small differences in the XAS that result from the photoexcitation event. Nonetheless, this is an area of catalysis that has become the focus of increased levels of investigation. Figure 27.6 summarizes the results of an experiment that seeks to understand the kinetics and changes in structure that occur within a [CoII(DPA-Bpy)Cl]Cl species, where Bpy ¼ bipyridine, which works in tandem with a RuII-based

577

sensitizer (Fig. 27.6a), to effect the splitting of water and release hydrogen when illuminated using UV irradiation [95]. The central element of the reaction cycle (Fig. 27.6b) is an electron transfer from the reduced ruthenium complex to the CoII complex. This results in a reduction of the CoII to CoI, which is subsequently perceived to undergo oxidation to CoIII, which then affects the splitting of water, before both ruthenium and cobalt centers are finally returned to their +2 oxidation states to complete the reactive cycle. The significant challenges that this investigation faces are illustrated in Fig. 27.6c panels (I) and (II) where the Co K-edge XANES derived from a 1.5 mMol solution of this complex (in the presence of 3 mMol of the ruthenium sensitizer) is shown in dark and illuminated conditions, whilst Fig. 27.6c shows the difference between these two cases. As can be observed, the fraction of cobalt that is excited is extremely small, though from Fig. 27.6c panel (II), that changes induced by photoexcitation can be resolved in various parts of the spectrum. Moreover, Fig. 27.6c panel (III) shows that the passage of the photocatalytic reaction can be modeled on the μsec timescale through monitoring of different energies (corresponding to features (E) and (A) in panels (II) and (III)). The experiment is even sensitive to changes in intensity in the 1s ! 3d (quadrupole transition) pre-edge feature (P, panel (II)) which contains within it information regarding the symmetry of the Co center. These changes, though small, are evidence of reduction of the CoII starting state. Moreover, the change (increase) in intensity of this feature under irradiation is indicative either of an augmentation of the degree of distortion, which already exists in the cobalt complex, away from a perfect octahedral symmetry, or a discrete change to a C4v symmetry. Panel (IV) then summarizes the result of fitting the difference EXAFS also obtained from this study, in an effort to separate out the contribution from the starting CoII complex and access aspects of the nature of the CoI species derived from photo-irradiation. Fitting of difference EXAFS is a nontrivial exercise that requires an extremely high level of precision. The retrieved EXAFS cannot specify whether the reduced CoI species retains a sixfold coordination or enters into a fivefold coordination state (though the latter possibility is preferred from DFT), but it does resolve a very significant contraction in the first shell Co-N coordination of ca. 0.09 Å

ä Fig. 27.5 (continued) the MEA (1) before and (2) after the ADT cycles. (D) Accelerated degradation test (ADT) procedure (left axis) and electrochemical surface area (ECSA) versus the number of the ADT cycles (right axis). Square and circle plots indicate ECSAs for before

and after the CT-XAFS measurement. (Reproduced from Chem. Rec., 2019, 19, 1380–1392 [86], with permission from The Chemical Society of Japan & Wiley-VCH Verlag GmbH)

27

578

M. A. Newton et al.

a)

2⊕

HO O

O

H C

C

N N

CH2OH

RuII

N

H

N

N N

N

N

CoII N N

N HO

⊕

Cl

O-(H+)

Ascorbic acid /Ascorbate H2A/HA-

[Ru(bpy)3]Cl2-6H2O

[CoII(DPA-Bpy)Cl]Cl

Ru(II) photosensitizer

Co(II) pre-catalyst

⊕

b)

From Ru(I) e–

N

N CoI

N

[CoI-VS]

N N H+

H2O

Cl N N

⊕H O 2

OH2

N CoII N N

N N

2⊕ 0.09 ± 0.03 Å bond contraction and penta-coordinated

N CoII

N

N Cl

[CoII-Cl]+

[CoII-OH2]2+

H2

H N N

H2O, H+ and

e– From Ru(I)

2⊕

N CoIII N N

?

[CoIII-H]2+

Fig. 27.6 (a) Molecular components of the Co/Ru-based system for the photocatalytic splitting of water. (b) Schematic of the proposed mechanism. (c): (I) Co K-edge XANES of the cobalt orgamometallic in both dark and illuminated states; (II) experimental (black) and calculated (red) difference XANES (light–dark); (III) time-dependent behaviour of the two difference XANES features (E and A in (II)) together with

kinetic fits (blue); (IV) extracted EXAFS due to the majority CoII state (blue) and the reconstructed, minority, reduced, CoI state (black) derived from photo excitation. Fits to the two datasets are also given by the red and pink lines. (Reprinted with permission from J. Phys. Chem. Lett., 2016, 7, 5253–5258). Copyright (2016) American Chemical Society)

which has accompanied the photo-reduction of the cobalt center.

some examples of recent operando studies of “conventional” PEM fuel cells, before moving on to consider other types of electro-catalytic materials.

27.3.4 Electrochemistry/Catalysis: Novel Approaches to Combining XAFS with Electrochemical Techniques and Cells

The Importance of Pd Hydride Phases and Pd-Based PEMFC Fuel Cells [99] Figure 27.7 nicely illustrates how operando study can be employed to understand basic aspects of reactive chemistry within PEM cells but also how different types of fuel cell configurations may affect what is observed. The essence of the problem being attacked in this investigation [99] is the resolution of the source of discrepancies that had appeared in the literature regarding the presence or absence during electrochemical cycling of a feature in voltammetry

Among the areas where operando X-ray spectroscopy has flourished in recent years, electrochemistry and electro-catalysis have been both very prominent [99–129] and have evolved and diversified greatly within this generic categorization. For the purposes of this chapter, we shall begin with

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

c)

579

1.6

0.3

Laser off Laser on

(I)

Integrated feature intensity (a.u.)

27

Normalized P (E)

1.2

0.8

0.4

0.0

7708

E

7725

(II)

0.01 Difference F(E)

7712

(III) 0.0

Feature E Feature A Multiple expotenial fit

–0.6

7740

Fit Exp

27

–0.3

0

10

20 30 Delay (Ps)

0.6 (IV) 0.4

40

Reconstructed Co(I) Co(I) fit Co(II) exp Co(II) fit

0.00 0.0 –0.01 –0.4 –0.02 A 7720

7740

7760

7780

X-ray energy (eV)

7800 –0.8

2

4 6 8 Wavenumber k (Å-1)

10

Fig. 27.6 (continued)

related to hydrogen desorption from a palladium hydride phase. Figure 27.7a describes some of the properties of the cell employed for these studies. Figure 27.7b then shows, firstly, (a) the Pd K-edge XANES and near-edge EXAFS obtained at two different applied voltages, (b) the k2-weighted EXAFS, and (c) the corresponding Fourier transforms of these EXAFS data. The high quality of the EXAFS (Fig. 27.7b (b)) is plain to see, as is, in all three panels, the distinct phase shift indicative of the formation of an expanded (relative to hydrogen-free palladium) fcc lattice. This phase shift results from the dissolution of hydrogen into the palladium nanoparticles (in this case 80 wt% Pd/C; average particle diameter ¼ 9.8 nm) to form palladium hydride phases [54–56]. Figure 27.7c then constructs isotherms that plot the stoichiometry of the PdHx as a function of operating temperature and applied voltage, from (a) purely electrochemical measurements and (b) the Pd K-edge XAS, using the first shell Pd-Pd bond distance as extracted from the EXAFS. Through comparison to a variety of similar and preexisting measurements [54–56], these data show that the degree to which PdHx is formed at high potentials is very much a function of the Pd dispersion. Low Pd dispersion

results in ¼ low (x < 0.05) levels of adsorbed H and the α-phase of PdHx, whereas the production of PdHx in more highly dispersed systems is much higher for comparable applied voltages (with x up to ca. 0.25). Further measurements (not shown) establish that under operando PEMFC conditions (80
C and 1 bar H2), the PdHx phase is maintained across the potential range (to ca. 90 mV) at a high level (x ≈ 0.6). This final observation is in stark contrast to the behavior of palladium-based cells which use liquid electrolytes. This difference is then resolved as the result of the very different mass transport characteristics of these different PEM fuel cells. One concludes, therefore, that great care must be taken when choosing the conditions of characterization applied to a working fuel cell, specifically the use of appropriate mass transfer regimes, especially for any electrochemical systems wherein the reactant molecule is supplied to the working cell in a gaseous form.

The Role of Iron Dopants in Cobalt-Based Perovskite Catalysts in the Oxygen Evolution Reaction (OER) The OER is one of two key components to the electrocatalytic splitting of water, the other being the corresponding

580

M. A. Newton et al.

Pnorm. (a.u.) BACK

a

40 °C

1.0

a 24.34

5

10

15

20

25

30

k2X (k)

−1.0 0

|F(R)| (Å–3)

Ifluo

WE N2 or H2 GDL

4

8 12 k (Å–1)

c

90

0.6 2.82

1.5 1.0

0.5

2.80

0.4

2.78

0.3

XAS

0.2

2.76

0.5

c 0

1

2

3 R (Å)

4

5

0.1 b

2.74

GDL

CE / RE H2 @ 1 bar

30 60 E (mVRHE) p(H2) (mbar) 100 10

16

2.0

6

b Pd-H

Region (I) a Pd-H + Had

a

1000

b

Region (II)

EC

0

Ex-ray (keV) I0

0.2

0.0

RPd-Pd (Å)

0

0.3

1.0 b

0.01

24.38 24.42 Ex-ray (keV)

0.4

0.1

Pd K-edge

0.1

20 °C 40 °C 60° C 80 °C 100 °C

0.5

15mV

0.0 24.30

Reactive correlations p(H2) (mbar) 100 10 1000 0.6

90mV

0.5

1 0.5m m 1.5m m 4m m

Transmission

FRONT

c)

Exemplary EXAFS

x in Pd-Hx

b)

Cell configuration

x in Pd-Hx

a)

0

0.0 30 60 E (mVRHE)

90

Fig. 27.7 (a) (a) Photograph of the fluorescence-yield-compatible XAS operando PEMFC cell; (b) calculated X-ray transmission of the graphite windows of the cell at the incident/detection angle of 45
as a function of X-ray energy and window thickness. The energy of the Pd K-edge (24.35 keV) is also given; (c) cross section of the cell indicating the XAS geometry used and the gaseous conditions applied to the working electrode (WE) and the counter/reference (CE/RE) electrodes. (b) Exemplary Pd K-edge EXAFS data obtained at an operating temperature of 40
C. (a) Example Pd K-edge XANES for two applied voltages and wherein the phase shift due to hydride formation is already evident above ca. 24.38 keV; (b) K2-weighted Pd K-edge operando

EXAFS at the two voltages given in (a) and again showing very clearly the substantial phase shift in the Pd-Pd scattering. (c) The corresponding Fourier transforms of the K2-weighted EXAFS data. (c) (a): Relationship between the applied voltage and the stoichiometric hydrogen content of the palladium as a function of the temperature of fuel cell and the partial pressure of hydrogen, as determined from purely electrochemical measurement. (b): Relationship between the first shell fcc Pd-Pd bond length (derived from analysis of the EXAFS) and the stoichiometry of the palladium phase, the temperature of operation, and the partial pressure of hydrogen used. (Reprinted with permission from (ACS Catal., 2016, 6, 7326–7334). Copyright (2016) American Chemical Society)

evolution of hydrogen (HER). Among the diverse materials that catalyze this reaction are perovskites founded upon cobalt [130]. Moreover, the efficiency of such materials can be greatly enhanced if a portion of the cobalt, which initially occupies an octahedral site in these complex oxides, is substituted by iron. A question that therefore arises is: By what mechanism does the iron promote this process? To gain insight into this promotional effect, Kim et al. [121] used operando XAS, made within a purpose-built electrochemical cell that permitted transmission EXAFS (and indeed small-angle X-ray scattering (SAXS) [102, 107]) to be made during electro-catalytic operation, to establish the nature of the speciation of both iron and cobalt and how they change with time during operation. The application of operando Co K-edge XAS had already established that a correlation existed between activity at OER potentials in these materials and an increased level of oxidation in the

cobalt, [104] and that the activity of this material itself was derived from the formation, at the catalyst surface, of an extruded metal oxy(hydroxide) phase. In other words, the high activity of these materials was not derived directly from their innate structure but rather from the formation of a surface oxy(hydroxide) layer that formed under the basic reaction conditions. However, precision as to the structural foundation of this promotional effect remained elusive. Figure 27.8 summarizes how these measurements were made and the significant structural insight that operando Co K-edge XAFS leads to in terms of the active structures present in these materials. What these data show, beyond the previously established segregation of an oxidic (Fe/Co) layer to the surface of the perovskite as a result of the basic nature of the reaction medium and the applied potential [104], is that the OER active phase is comprised of edgesharing, rather than corner-sharing, polyhedra that are most

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

a)

Holes for mounting on beamline sample stage

581

b) a

Counter electrode contact

a /í 2

Working electrode contact

27

Reference electrode

Corner-sharing octahedra

Edge-sharing octahedra

c) 6

Edge-sharing polyhedra from flame spray Ba0.5Sr0.5Co0.8Fe0.2O2+δ synthesis

|x(R)| Å4

Co/Fe – O(OH) construction 4

1.20 VRHE

1.54 VRHE

2

OH−

Self-assembled oxy(hydroxide) layer Partial A-site catlons dissolution

OER/ LOER Sr O Ba O Redeposition Fe O

0 1

2

3

4

5

Radial distance (Å)

Fig. 27.8 (a) Experimental arrangement of the electrochemical cell developed for combined XAS/SAXS studies of electrochemical materials and processes. [c,d] (b) Illustration of the Co-Co scattering interactions expected from cobalt oxides comprised of corner-sharing and edge-sharing polyhedral, as indicated. (c) The evolution of structure of the cobalt oxide phase as a function of operating voltage and the

commencement of OER activity, together with schematic illustration of the formation of a metal oxide shell about the body of the perovskite under reaction conditions, and its eventual arrangement into an OER active Co/Fe oxy(hydroxide) surface layer. (Reprinted and adapted with permission from (ACS Catal., 2018, 8, 7000–7015). Copyright (2016) American Chemical Society [121])

likely two-dimensional in nature. The promotional role of the iron dopant, for which parallel assessment at the Fe K-edge elicits no evidence for a change in terms of its local electronic state, therefore lies in directing the formation of the extruded metal layer into this specific configuration, rather than any other possible Fe/Co oxide phases under the conditions of operation.

detection systems, must be employed. These situations therefore require different solutions in terms of the environments within which the sample must be presented. In this example [112], the nature of iron-based cathodes, and how they corrode under conditions required for the electrochemical production of NaClO3, a precursor to ClO2 extensively used as a bleaching agent, is studied. This route to NaClO3 is, however, blighted by high corrosion rates that lead to diminished performance in the steel-based electrodes used to facilitate it. Ex situ studies [131] have highlighted the formation of thin layers of two types of iron oxy(hydroxides): α-FeOOH (goethite) and γ-FeOOH (lepidocrocite) at the cathode of the working system. Consequently, considerable interest in understanding how these two different polymorphs

The Nature, Stability, and Reversibility of Active Iron Phases Under Conditions of Hydrogen Evolution (HER) There are many situations in catalysis wherein transmission measurements cannot be made, and fluorescence-yield XAS, together with its different geometric requirements and

582

M. A. Newton et al.

behave under conditions of electrochemical reduction has arisen, though significant debate has persisted as to the nature of the reduced phases [132–134]. Figure 27.9 delineates various aspects of operando XAS which this study addressed through the design and implementation of a specific cell configuration [101] that allows electrochemical measurements to be made in tandem with fluorescence-yield XAS. Given the intrinsic limitations that have held back timeresolved, single-shot fluorescence-yield XAS (outlined in Sect. 27.2.1), for which solutions had not, until very recently

a)

c)

0.4 0.76 V

0.41 V

[36], been available, a different approach was used to gain insight into the dynamic behavior of the system. Here, in order to follow the redox character of the iron phases during cyclic voltammetry (Panel I, (a) and (b)), fixed energy X-ray absorption voltammetry (FEXRAV; panel I (c) and (d) [a]) was used to induce sufficient time resolution in the measurement to see into processes to hand. However, FEXRAV, while enabling investigation, does so at the expense of much of the information that could be obtained if the fast, fluorescence-yield XANES/EXAFS of the type demonstrated in Sect. 27.2.1 was to be utilized [36].

0.2

0.0

I (mA)

I (mA)

0.2

–0.2 –0.4

b)

–0.2

0.32 V –0.55 V

–0.4

–0.65 V

–0.7 V

d) 0.06

0.06

0.05 m (arb. un.)

m (arb. un.)

0.0

0.32 V

–0.4 V

–0.6

0.7 V

0.05 0.04

0.04 0.03

0.03 0.02

0.02 –0.9 –0.6 –0.3 0.0

0.3

0.6

0.9

–0.9 –0.6 –0.3 0.0

Potential vs. RHE (V)

e)

0.4 0.0 7110 7120 7130 7140 7150 7160

1.2 γ-Pristine electrode γ −0.4 V γ −0.65 V γ 0.41 V γ 0.76 V (first acquired) γ 0.76 V (second acquired)

0.8 0.4 7120

7130

7140

7150

0.9

2.0 1.5 1.0

0.8

1.6

0.0 7110

2.8 2.4

m (arb. un.)

m (arb. un.)

f)

1.2

2.0

0.6

Potential vs. RHE (V)

1.6

2.4

0.3

2.0

0.5 0.0 7110 7120 7130 7140 7150 7160

1.6 1.2 0.8

α-Pristine electrode α −0.55 V α −0.7 V α 0.7 V

0.4 7160

E (eV)

Fig. 27.9 Panel I: Top cyclic voltammetry for (a) γ-FeOOH and (b) αFeOOH cathodes under conditions of appropriate to electrochemical NaClO3 productions (0.2M Na2SO4 (aq.), NaOH, pH = 11). Operando FEXRAV @ 7.125 keV for (c) γ-FeOOH and (d) α-FeOOH cathodes. Voltage-dependent Fe K-edge XANES for (e) γ-FeOOH and (f) αFeOOH cathodes. Panel II: (a) Fourier transforms calculated L.H.S pristine γ-FeOOH (red line) and γ-FeOOH at 0.76 V (blue line), and R.H.S pristine α-FeOOH (green line) and α-FeOOH at 0.7 V (pink line). Clusters of (L.H.S, γ-FeOOH) and (α-FeOOH, R.H.S) are also shown wherein the central iron is gray, oxygens are red, and all the others are iron atoms. Bottom panels: (b) XANES simulation of γ-FeOOH at 0.4

0.0 7110

7120

7130

7140

7150

7160

E (eV)

V and α-FeOOH at 0.7 V. The black and blue dots represent the experimental signals for γ- and α-FeOOH, respectively, while the colored lines are the simulations with the model of reduced green rust (red and pink lines) and with Fe(OH)2 (green line). The bond angles corresponding to the two structures are also shown. (c) Cluster of reduced green rust obtained after XANES simulation. (d) EXAFS signal and (e) corresponding Fourier transform of γ-FeOOH at 0.4 V. The black line represents the experimental curve, while the red line is the fit with the reduced green rust model. (Reprinted and adapted with permission from Appl. Energy Mater., 2018, 1, 1716 – 1725. Copyright (2018) American Chemical Society. [112])

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

583

m (arb. un.)

a)

b) 3

81°

2

99°

1

27

76°

0 104°

–1 7120

7130

7140

7150

7160

E (eV)

c)

d)

3

FT | k2 c (k) (arb. un.)|

k2 c (k) (arb. un.)

2 1 0 –1 –2

Exp Fit

–3 –4

4

6 k (Å–1)

8

10

20

Exp Fit

15 10 5 0

0

2

6

4

8

10

r (Å)

Fig. 27.9 (continued)

Nonetheless, it yields a very informative window into the similarities and differences that exist in the redox behavior of the two FeOOH phases studied. Most notably, while in the first half of the voltammetric experiment, the two FeOOH cathodes behave in a similar manner (orange curves), in the second half (oxidative, red curve), pronounced differences appear. This result shows that the electrochemically induced reduction shows a very similar overall character in both cases. However, whereas the re-oxidation of reduced γ-FeOOH is facile, the re-oxidation of reduced α-FeOOH is subject to significant kinetic barriers. Further, detailed analysis of the XANES collected at different potentials establishes that in both cases FeIII is reduced to FeII and that the spectra obtained are very similar to both FeII(OH)2 and “sulfate green rust.” From the relatively high intensities of the Fe K-edge white line, and the multiplicity and relative intensity of the pre-edge features observed, it is further concluded that these reduced phases are comprised of FeII coordinated in an essentially undistorted, octahedral, and high-spin coordination [112]. Lastly, both fluorescence-yield XANES and EXAFS were then subjected to extensive modeling to further specify the nature of the reduced FeII, as both FeII(OH)2 and “sulfate green rust” candidates conform to the overall characteristics of the XANES. In Fig. 27.9, panel II, top, the result of full curved-wave and multiple scattering fitting of the EXAFS for the two starting FeIII(OOH) compounds is shown, along with indications of the origins of specific scattering

pathways that pertain to each polymorph (left, γ-FeOOH; right, α-FeOOH). Below this, panels (a)–(d) show (a) fitting of the XANES of the reduced iron using models pertaining to Fe(OH)2 (top curve, green fit) and “sulfate green rust” (bottom two curves, red and purple fits). This analysis, which makes use of the sensitivity of the XANES coordination and bonding geometry and bond angles, clearly favors the “sulfate green rust” model. Panel (b) then shows the structural model, based upon the consideration of the known structure of “sulfate green rust,” but comprised only of FeIII centers without the presence of the sulfate anions. Refinement of this structure leads to the fits to the EXAFS of the reduced material given in panel II (c) and (d). The conclusion of this study is that under conditions (potentials < 0.4 V) of hydrogen evolution (HER) both αand γ-FeOOH cathode are reduced to a common FeII species that has the character “green rust.” However, despite this commonality, the facility with which this phase is re-oxidized is subject to very different kinetic limitations.

27.3.5 Combined XAS and Photo-ElectroCatalysis In some areas of research into catalytic materials, there exist important processes that require both photonic and electrochemical components to be combined. As such, for these to be adequately studied in an operando sense, the sample

584

presentation must permit this combination in a manner that is as close to optimal as can be achieved. The photoelectrochemical splitting of water is such a case, and one that has recently been addressed in a manner that complies with the above conditions using XAS [103]. The 3D printed cell developed for these studies [101] is arranged such that the distances between the required working, counter, and reference electrodes are minimized and that a substantial reservoir of electrolyte is present. Moreover, this arrangement permits the back illumination, as required, of the sample electrode using a suitable diode while, at the same time, conducting electrochemical (voltammetry) measurement and fluorescence-yield XAS. The right-hand panel delineates some of the results achieved using this novel arrangement. The top diagram shows illustrative Cu K-edge XANES derived from two bulk copper oxide samples measured within the cell (CuO (a), green, and Cu2O (d), blue). The pink spectrum (b) is that obtained from the copper photocathode as synthesized. The black spectrum (c) is derived from illumination (at 0.3 V relative to the reference electrode) at λ ¼ 400 nm for 8 h, while the red curve is a linear combination fitting of this spectrum using the reference CuO and Cu2O spectra. This yields a phase/oxidation state composition of the photocathode after 8 h of operation of ca. 80% CuI and 20% CuII. Continuous operation has therefore led to a significant reduction of the photocathode. As in the previous example, given that single-shot, timeresolved fluorescence was not available, to gain dynamical insight into the behavior of this material, FEXRAV (righthand side, middle panels, in Fig. 27.10) was employed. In this case, the X-ray energy was tuned to the CuI pre-edge feature at 8.891 keV, and the potential across the cell scanned in both dark (middle panel (A)) and then illuminated conditions (B), while measurement of the resulting currents (mA cm2) was also made. In the “dark” experiment, the FEXRAV shows that initial reduction (increased values of the absorption coefficient (μ)) of the photocathode occurs only at potentials below 0.2 V, and that as the voltage is cycled the level of reduction systematically increases. Illumination of the cathode significantly promotes reduction, which is then seen to commence at voltages of ca. ca. 5 keV. There are many reasons for this, not least of which is that the majority of XAS beamlines active in the last 20 or so years have low-energy cutoffs as a result of their optical arrangements. As such, studies in the “soft” (ca. < 2 keV) and “tender” (ca. 2–5 keV) X-ray regimes have been far less commonplace. Moreover, it is also the case that below ca. 2 KeV new optical configurations, such as different mechanisms for the monochromation of the X-rays, are required. The soft X-ray region is particularly rich in information. Not only are the K-edges of elements lighter than phosphorous (i.e., up to silicon) to be found in this region of energy but also the L and M edges of many heavier catalytic elements. Importantly, therefore, soft X-rays permit one to directly address, from a spectroscopic point of the view, the lighter elements (such as carbon, oxygen, and nitrogen) that are often core components of catalysts and/or form the reactive species adsorbed at the catalyst surface or in the gas phase. These elements, and the speciation associated with them, are not at all easily reported upon via XAS measurements founded upon the higher energy edges, and therefore there is considerable motivation to be able to use operando XAS in this energy regime. However, moving into this energy regime also poses significant challenges to operando measurement, as the

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

Electrolyte reservoir

1.0 0.5 0.0

a) 170

27 8980

8990 9000 E (eV)

0.00 –0.01 –0.02 –0.03 –0.04 –0.05 –0.06

FEXRAV signal CV

160 150 140 130 120 110 100

m (a.u.)

b)

d c b a

1.5

b)

9010

Exp. Fit 4

e) 2.0 k2 c(k) (arb. un.)

d)

100 μm

1.5 1.0 0.5 0.0 –0.5 –1.0 –1.5 –2.0 2

Potential vs RHE (V)

d)

6 8 k (Å–1)

10

12

f)

Exp. Fit 4

6 8 k (Å–1)

10

12

FT | k2 c(k) (arb. un.) |

50 mm

4 3 2 1 0 –1 –2 –3 –4 2

40

FT | k2 c(k) (arb. un.) |

k2 c(k) (arb. un.)

12.44 mm

c)

–0.10.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Potential vs RHE (V)

c)

0.03 0.00 –0.03 –0.06 –0.09 –0.12

FEXRAV signal CV

800 700 600 500 400 300 200

–0.10.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Working electrode

Current (mA cm–2)

m (arb. un.)

a)

585

Current (mA cm–2) m (a.u.)

27

18 15 12 9 6 3 0

Exp. Fit

30 20 10 0

0

2

6 4 r (Å)

8

10

Exp. Fit

0

2

6 4 r (Å)

8

10

Fig. 27.10 Panel I: (a–d) schematic of the cell developed by Achilli et al. [101] for investigation of photo-spectro-electrochemical devices using fluorescence yield XAS. The red circular hole indicated in (a) permits sample illumination, as required, using a suitable source of UVvisible light. Right hand panel, top: Cu K-edge XANES measured in fluorescence for: (a) bulk CuO (green) and (d) bulk Cu2O standards (blue). (b) is the CuOx photo cathode measured in its original as made state, while (c) (black spectrum) is the operando spectrum from the photocathode at a potential of 0.3V (relative to the reference electrode) and after 8 hours of illumination @ λ ¼ 400 nm. The red line is a linear combination fit to the data based upon a combination of (0.2) CuO and

(0.8) Cu2O components. Middle panels: FEXRAV of CuxO electrodes at 8981 eV in the dark (a) and under LED illumination (b); 0.1 M K2HPO4 + KH2PO4 (pH 7) was used as an electrolyte. Shadow zones indicate the instability potentials. Black curves: μ; blue or brown curves: j (current density). Starting points are indicated by red arrows. Bottom panels c) to f): (c) and (e) Cu–K edge EXAFS spectra of CuxO-E and CuxO-E kept at 0.3 V vs RHE for 8 h upon irradiation with λ ¼ 400 nm and (d, f) their corresponding Fourier transforms, respectively. Black line is experimental data, and red line the fit in each case. (Reprinted and adapted with permission from Appl. Mater. & Interfaces, 2016, 8, 21250 – 21260. Copyright (2016) American Chemical Society. [103])

penetration of X-rays into matter reduces dramatically below ca. 4 keV. This comes with two considerable knock-on effects: firstly, transmission measurements become much more difficult, and, secondly, the variety of materials (and their thickness) available for windowing of reactor systems dwindles considerably. As such, from both reactor design and methodological points of view, moving into this energetic regime requires a number of challenges to be overcome. Nonetheless, various solutions have been arrived at [101, 135–146]. Moreover, numerous hard X-ray resources have

also extended their lower operational limits into the tender X-ray regime, and many more beamlines can, or will be able to, make X-ray spectroscopic measurements down to ca. 2 KeV than were previously available. Of the growing number of catalysis investigations that have ventured into the soft ( 2 keV) range, we shall consider two examples that illustrate how such studies can be achieved for both gas-solid and electrochemical situations. Figure 27.11 outlines (a) the design of the soft X-raycompatible operando cell [101]. The key element to this

586

M. A. Newton et al.

rt

design is the use of an extremely thin (ca. 10 nm) Si3N4 window, held within a conically cut flange. This arrangement yields both a large solid angle of detection for any X-ray fluorescence desired to be detected and also a high transmission of the incident X-ray beam (i.e., ca. 60% @ 500 eV [147]). In this case, however, total electron yield (drain current) is the preferred method of detection. This method samples all the electrons (photoelectrons, Auger electrons, and secondary electrons) emitted from the sample in a manner that is inherently surface sensitive, the escape depth of electrons of these kinetic energies being limited to but a few nanometers into the surface of the material. Figure 27.11b then illustrates the sensitivity of O k-edge soft X-ray XAS to gas phase speciation for a number of relevant species. Panels (c) and (d) summarize the results obtained from a SnO2 sample previously dried through thermal treatment to

b)

Top part

e

O K-edge

Bo

tto

m

pa

a)

573 K [126]. After this treatment, the SnO2 sample retains a degree of hydration, corresponding to chemisorbed water and an –OH-terminated surface. Figure 27.11c gives the spectra derived at the O K-edge, while Fig. 27.11d gives those derived from the Sn M-edge together with standard spectra derived from SnO2 and SnO. These results nicely illustrate the richness of the spectroscopy in this soft regime and how sensitive it is to both surface and gas phase speciation in a manner that high-energy XAS is not. What arises from these studies is an in-depth view of how SnO2 interacts with hydrogen (Fig. 27.11), methane, and propene (not shown) [126]. The degree to which reduction of the SnO2 occurs, to yield an oxygen-deficient SnO layer, the thickness of which is a function of the reductant, which, even under hydrogen, extends only ca. 3–4 layers into the sample, results from this soft X-ray approach.

f

CO2

d CO+O2

Heater 1×1 cm2

Cer

am

ic s

Membrane

TEY (arb. un.)

epa

rato

r

Gas inlet O

Air b c

Gas inlet a

H2O

528 530 532 534 536 538 540 542 544 546 Photon energy (eV)

c) O K-edge

Sn M4,5 edges

SnO

230 °C

In O2

360 °C

In H2

240 °C 144 °C

In H2 In H2

TEY (arb. un.)

TEY (arb. un.)

RT

0.1 arb. un.

300 °C

250 °C

In H2 flux

d)

200 °C

150 °C

120 °C

SnO2

100 °C

490

495 500 Photon energy (eV)

505

488

490

Fig. 27.11 (a) Schematic illustration of the soft X-ray compatible gas-solid reactor system, (b) O1s XAS spectra due to varied gas phase species (as indicated), (c) temperature dependence of the O1s spectral region during heating of SnO2 under flowing H2, and (d) temperature depedence of the Sn M-edge region of the SnO2 samples under the same

528 530 532 534 536 538 540 542 544 546 Photon energy (eV)

conditions. Spectra derived from SnO2 (bottom, black) and SnO (top, red) are also given. The gray box highlights the evolution of an SnO phase as the temperature is increased. (Reprinted and adapted with permission from J. Phys Chem, 2020. Copyright (2020) American Chemical Society [126])

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

Concurrent with this reduction, the evolution of water to the gas phase is clearly signaled (Fig. 27.1c, vertical, black, dashed lines). Moreover (though not shown here), when considering reduction using hydrocarbons, such as methane and propene, no evidence is forthcoming for the production of carbonyl-type species, such as carbon monoxide and carbon dioxide. This in turn suggests that the interaction of the hydrocarbons with SnO2 within this temperature range is founded upon a “selective” two-electron transfer process that leads to the formation of surface methoxy/propoxy, with the outer layers of the SnO2 acting as an “electron sponge” to accommodate the required electrons. A similar soft X-ray approach has also recently been used in the operando study of the behavior of nickel oxide-based (NiMnO3) electro-catalysts engaged in the OER [122], which yields our second example of this type of approach to operando spectroscopy. Figure 27.12 panel I, (A–E) illustrates the overall design of the reactor system used to achieve operando electro-catalytic study under potential control and with provision for the sample to be saturated with a suitable electrolyte (in this case 0.1 M KOH). As can be seen, through comparison to Fig. 27.11, the overall design of the cell is very similar to that used for the previous soft X-ray example, and again the central element is a Si3N4 window (150 nm thick). In this example, however, the whole system is mounted in vacuum to permit detection of the total X-ray fluorescence yield obtained at the Ni K-edge for the XAS measurements. More than that, and though not shown here, this study goes a step further, by also measuring resonant inelastic X-ray spectroscopy (RIXS) at the Ni 2p edge to gain additional information regarding the nature of the intermediates formed during the OER. Importantly, possible effects of exposure to the extremely powerful and focused X-ray beam used in this study were explicitly looked for and found to be absent, save for an X-ray-induced corrosion of the Si3N4 window, which eventually compromises its integrity and leads to failure after ca. 4-min X-ray exposure at any single point. As a result, spectra were collected in a “point-wise” fashion across the Si3N4 membrane and using a transient exposure to the X-rays of significantly less than the time required to induce failure of the window. The voltage-dependent XAS from the Ni 2p edge given in Fig. 27.12 shows a significant sensitivity to the changes in nickel speciation. Even the act of simply wetting the sample in the 0.1 M KOH electrolyte solution is seen to elicit measurable changes in the overall XAS envelope. Addition of KOH results in a 2.5 eV shift (to lower binding energy), which is accompanied by a significant increase in intensity, in a feature at 854.6 eV. This is interpreted as being the result of a partial conversion of the NiO present to β-Ni(OH)2 which, as can be seen from the standard spectra given in Fig. 27.12 panel II, C, also has a prominent feature at this energy. Importantly, β-Ni(OH)2 is clearly

587

distinguishable from β-Ni(OOH) when the whole Ni 2p envelope is considered (vis-à-vis the region between 867 and 875 eV). Indeed, the “wet” spectrum can be reproduced to a high level using linear combination of the spectra obtained using a 50:50 mixture of NiO and β-Ni (OH)2. A change in the applied potential to 1.25 V, just below the voltage required to elicit OER, causes only very slight further changes in the spectra before more significant changes occur as a result of a further increment in the voltage to 1.47 V and then to 1.55 V. These potentials are greater than the redox potential of NiO, and therefore oxidation of NiII to NiIV might be expected. In the L3-edge spectrum, an energetic shift of almost 2.5 eV is observed, though two components to this feature may still be clearly observed. A similar shift is also noted in the L2-edge but one that also results in the collapse of the two-peak nature of this part of the XAS into a much broader and less well-defined feature. These changes are consistent with the formation of NiIII and/or a mixture of NiIII and NiIV, and comparison to the standard spectra given in the bottom panel strongly suggests a conversion of most of the nickel sampled into the γ-NiOOH phase. However, and within the sampling depth of the experiment, further analysis indicates that this conversion is not complete. Both linear combinations of the relevant reference spectra, and a consideration of the Ni L3 RIXS (not shown), indicate that a proportion of the wet catalyst precursor (estimated at ca. 20%) remains even at 1.55 V. Nonetheless, the Ni L-edge XAS data very clearly indicate that the dominant phase present under conditions required for OER is γ-NiOOH.

27.3.7 Operando Studies Using Laboratory XAS Instruments In recent years, laboratory-based X-ray spectrometers have appeared and become commercially available [20–27]. Potentially, these sorts of spectrometers may have an important role to play within the overall palette of X-ray resources that are available to researchers, especially for experimentation that lies outside of the remit of dedicated central facilities such as synchrotrons. Once such operando case is that of long timescale experiments that consider the behavior of catalytic materials on timescales well beyond those generally available at central facilities (ca. one week at most), and which are extremely relevant to application in terms of time-dependent activation of catalysts, or the long-term degradation of their performance. While most of these sorts of spectrometers are not currently used for operando study, there has been at least one published example [148] that shows that these instruments can, and should, have a place within the pantheon of operando XAS resources.

27

588 Fig. 27.12 Panel I: Photographs and a schematic illustration of a soft X-ray-compatible electrochemical cell arrangement. Panell II: (A) the overall catalytic sequence of events, and changes in Ni(OOH) phases, proposed to be occurring within the nickel oxide-based catalysts during electrochemical recycling. (B): Voltage-dependent spectra obtained from the working catalyst; (C), corresponding Ni M-edge spectra derived from different (α, β, γ, as indicated) phases of Ni(OOH) and NiO. (Reprinted and adapted with permission from Appl. Mater. & Interfaces, 2019, 11, 42, 38595–38605, Copyright (2019), American Chemical Society [122])

M. A. Newton et al.

I A

B

Inlet

C

Outlet

Inlet

O ring Electrodes Electrodes

Outlet

Membrane loading area Au-coated Si3N4 membrane thickness~ 150 nm

In-situ holder for liquid phase Fully assembled in-situ holder experiments Vacuum Atmospheric E side pressure side

D JIS90 Flange

Si3N4 membrane/WE

O-ring Si3N4 membrane/WE Gasket (PTFE) Sample Emitted Electrodes X-rays

Emitted X-rays Incident X-rays

Electrode Liquid inlet/outlet Electrode

Incident X-rays Echem cell Echem cell JIS90 flange II

A

Charge

Discharge

J-NiOOH

Normalized intensity

B

Overcharge

E-Ni(OH)2 Charge

Discharge

E-NiOOH

846 849 852 855 858 861 864 867 870 873 876 879 0.81

Ni3MnO4 dry Ni3MnO4 wet

Ni3MnO4 1.25 V Ni3MnO4 1.47 V Ni3MnO4 1.55 V

0.54 0.27 C

Normalized intensity

Aging

D-Ni(OH)2

0.00 0.81

NiO E-Ni(OH)2 E-NiOOH J-NiOOH

0.54 0.27 0.00 846 849 852 855 858 861 864 867 870 873 876 879 Incident energy (eV)

Figure 27.13a shows the overall configuration of the reactor system, constructed at the University of Helsinki, Finland [26, 27]. Figure 27.13b then shows exemplary results obtained from a Co/TiO2 catalyst, mounted within

a plug-flow reactor, and submitted to a feed typical of the Fischer-Tropsch catalysts in the production of olefins from carbon monoxide and hydrogen (523 K, 1 and 5 bar) [148].

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

589

a) X-ray tube Slits Mass flow controllers (MFC)

Reactor Back set-up pressure regulator (BPR)

27

Detector Si (533) crystal I

II

0.8 0.6 0.4

R: 0.0098757 Co: 0.67 ± 0.02 CoTiO3: 0.33 ± 0.02

0.2 0.0 7680

7700

7720

7740 7760 Energy (eV)

7780

1.0

0.6 0.4

R: 0.0097597 Co: 0.66 ± 0.03 CoTiO3: 0.34 ± 0.02

0.2

7700

7720

7760 7740 Energy (eV)

7780

Normalized absorption (a.u.)

Normalized absorption (a.u.)

c

0.0

c)

0.4

R: 0.010622 Co: 0.65 ± 0.03 CoTiO3: 0.35 ± 0.02

0.2

7700

7720

7740 7760 Energy (eV)

7780

7800

0.8 0.6 0.4

R: 0.0014699 Co: 1 ± 0.01 CoTiO3: 0 ± 0.01

0.2

7680

7700

d

7740 7760 Energy (eV)

7780

0.6 0.4

R: 0.0065 Co: 0.66 ± 0.01 CoTiO3: 0.34 ± 0.01

0.2 0.0 7700

7720

7740 7760 Energy (eV)

7780

7800

b

0.8 0.6 0.4

R: 0.0044708 Co: 1 ± 0.01 CoTiO3: 0 ± 0.01

0.2 0.0

7800

7680

7700

7720

7740 7760 Energy (eV)

7780

7800

1.2

c

1.0

105

1.0

0.8 0.6 0.4

R: 0.0019226 Co: 1 ± 0.01 CoTiO3: 0 ± 0.01

0.2 0.0 7700

7680

7720

7740 7760 Energy (eV)

7780

7800

d

0.8 0.6 0.4

R: 0.0019695 Co: 1 ± 0.01 CoTiO3: 0 ± 0.01

0.2 0.0 7680

7700

7720

7740 7760 Energy (eV)

7780

7800

105 I

100

II

100

Composition (%)

Composition (%)

7720

1.2

0.8

7680

1.0

0.0

7800

1.2

0.8

7680

0.8 0.6

7680

1.2

a

1.0

0.0

7800

1.2 1.0

1.2

b

Normalized absorption (a.u.)

1.0

Normalized absorption (a.u.)

a Normalized absorption (a.u.)

Normalized absorption (a.u.)

1.0

1.2

Normalized absorption (a.u.)

1.2

Normalized absorption (a.u.)

b)

Rowland circle

95

5

0

95

5

0

−5

−5 20 40 60 80 100 120 140 160 180 200 Time (h)

Fig. 27.13 (a): Schematic of the laboratory XAS spectrometer arrangement [26, 27] and the plug-flow reactor system employed for the transmission XANES study of a cobalt-based Fisher-Tropsch catalyst.

20 40 60 80 100 120 140 160 180 200 Time (h)

(b): (I) Co K-edge XANES derived from the laboratory XAS setup at 523 K and under an H2/CO feed of ratio 0.5 and at ambient pressure at different operational durations: (a) 6 h, (b) 17.5 h, (c) 25 h, and (d) 35 h.

590

M. A. Newton et al.

This example demonstrates that laboratory-based XAS spectrometers have evolved to the point where longtimescale, operando experiments can be undertaken and used to good effect. In this case, the data obtained with the laboratory source is used to assess the phase composition of the catalyst into active (cobalt metal) and inactive phases (CoTiO3), to determine how these phases change under reaction conditions over extended periods of operation. One conclusion that may be drawn from these measurements is that oxidation of the catalyst into inactive phases, such as CoTiO3, is not the source of the deactivation observed in these catalysts on these long timescales. While it is also very clear that sources of this type are still some considerable way off even bending magnet sources at present third-generation synchrotrons, in terms of the quality and overall data length achievable in a given time, useful operando catalysis experiments can be achieved using them. Moreover, the time-on-line study, which one could not realistically hope to obtain time for at a synchrotron (over 200 h on line, >8 days), is unique and showcases a new type of operando capacity. As with synchrotrons themselves, it seems inevitable that, as these types of source become more common and powerful, then their spectroscopic capacity will also increase, and such systems will become a viable option for a wider range of experiments.

27.4

Checks, Balances, and Outlook

Amidst all of the above, regarding how various technological challenges are being met and how the diverse and highly informative field operando XAS is venturing into new territory, there are caveats that need to be clearly delineated and considered. These may be broken down into three broad areas: firstly, whether the X-rays being used are really an experimentally neutral probe of the chemistry to hand; secondly, whether the presentation of the sample for operando investigation is fit for purpose; and, lastly, how does one ensure that measurements of reactivity, which generally integrate over the entirety of a sample, are accurately reflected by XAS measurements that, in general, sample only a fractional volume of any sample bed?

That X-rays have the capacity to interact strongly with matter is well known. However, there exists a considerable gulf between communities as to how seriously the potential of the X-rays to influence the result of any given experiment is considered. In medical physics (e.g., [149–152]), this is a matter that has long been at the core of the field. In the field of macromolecular crystallography as well, the issue of the induction of spurious results through the various ways that X-rays can alter the nature of biological matter is constantly, and critically, assessed [152–162]. In the world of catalysis – among other fields – however, systematic testing and dissemination (there are, of course, exceptions [163–171]) of this potentially problematic possibility are much less common. On the face of it, it appears largely assumed that the X-rays can be regarded as a neutral probe that do not change the behavior of the material under study. However, while this is, no doubt, a valid assumption in the majority of cases, it is to ignore the developments in synchrotron science that, along with making many new things possible, has systematically increased the power densities that are being applied to catalysts; and with the best will in the world, there is only so much power any given material can withstand before it is altered by the probe itself. Moreover, operando operation presupposes that the materials are being studied in the presence of reactive components (be they gases or liquids) or other media (for instance, solvents) that may also be affected by the presence of ionizing radiation. Figure 27.14 gives a salient example of the sorts of deleterious effects that can manifest within an inorganic system, a copper-exchanged zeolite (mordenite, Cu/MOR), while under study for its ability to activate methane [171]. Spectroscopically, this reaction can be followed using time-resolved Cu K-edge XANES, as this is appropriately sensitive to the relevant oxidation states of copper (I and II), as well as to levels of hydration (for CuII). In this experiment, the reactive chemistry is probed as a function of the brilliance of the X-rays and the total applied flux. What arises from this investigation, is that the apparent behavior of the materials, in terms of the starting state of the experiment, observed rates of reaction, and the overall extent of reaction as a function of time, is a significant function of the applied X-ray dose [171].

ä Fig. 27.13 (continued) The red lines, and the fitting parameters, correspond to least squares fitting of the XANES on the basis of previously obtained spectra pertaining to cobalt in the metallic state and CoTiO3. (II) Co K-edge XANES derived from an equivalent experiment conducted at the DUBBLE beamline at the European synchrotron radiation facility (ESRF) after (a) 6 h, (b) 10 h, (c) 12 h, and (d) 15 h of operation. As before, the red lines and the fitted parameters correspond

to least squares fitting of the Co K-edge XANES. (c) Cobalt phase fractions (metallic cobalt, black; CoTiO3, red) as a function of timeon-line during the Fisher-Tropsch synthesis (523 K, 5 bar) derived for two different feedstock ratios, (I) H2/CO ¼ 0.5 and (II) H2/CO ¼ 2. (Reproduced with permission from ChemCatChem, 2019, 11, 1039–1044 Wiley-VCH Verlag GmbH [148])

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

a)

Activated CUII/MOR + methane @ 413 K 1.4 1.2

b)

591

Fractional CUI formation versus applied X-ray dose

0.7

Beam dimensions _attenuation

0.6

80 × 80 mm

Normalized absorbance

CUI 1

0.5

0.8

0.4

0.6

0.3

0.4

0.2

0.2

0.1

0 8970 8980 8990 9000 9010 9020 Energy (eV)

Increasing X-ray brilliance

750 × 50 mm 2500 × 150 mm 2500 × 150 mm _2C 2500 × 150 mm _5C

0 0

500 1000 1500 2000 2500 3000 Time (seconds)

Fig. 27.14 (a) Exemplary time-resolved Cu K-edge XANES spectra obtained during the exposure of Cu/MOR to methane at 413 K. The red curve is obtained under oxygen at 413 K, the black after ca. 3000 seconds of exposure to methane, and the blue intermediate spectra. (b) The apparent rates of CuI formation derived from successive activation-

reaction cycles over Cu/MOR as a function of X-ray brilliance (beam focal dimensions and additional front end attenuation using different carbon (2 mm ¼ 2C and 5 mm ¼ 5C) filters). (Adapted and reproduced with permission from Phys. Chem. Chem. Phys., (2020), 22, 6826–6837, Royal Society of Chemistry [171])

Moreover, it can be shown that a bending magnet beamline, which does nothing to focus its X-rays, yields kinetic data for this system that is quantitatively comparable to those derived from other techniques (i.e., UV-vis [172]). To reach the point where the results from a bending magnet beamline that focuses its X-rays match those from a non-focusing line in all respects, the instantaneous X-ray dose must be reduced by  ca. 1/1500th of the optimal throughput and brilliance of the higher brilliance line (Fig. 27.14 (b), 80  80 μm2, black curves). Of note here is by measuring on two very different XAS lines, we understand that: (a) a good degree of caution needs to be exercised at modern XAS beamlines when making operando studies using even inorganic materials; and (b) that the situation is retirveable, but requires considerable reductions to be made in term of the levels of X-ray throughput and brilliance applied. Furthermore, it should also be borne in mind that most third-generation synchrotrons have either upgraded or will, in the near future, upgrade their machines. The result of these upgrades will be the production of X-ray sources that are even more brilliant than those that currently exist. These upgrades will have far-reaching ramifications, in terms of the new sorts of experiment they will permit. However, their increased power will also demand a much higher level of circumspection, in respect of their use, to ensure that the

results obtained from any given study do in fact reflect the true behavior of the materials in question and not processes and chemistry that result from the application of the X-rays themselves. In regard to sample environments and their suitability for purpose, in the examples given here, the performance of the cells developed and used has been quantified, in the absence and presence of the X-rays. As such, any results derived from their use can be considered as having a high level of confidence associated with them. Moreover, in conventional gassolid catalysis, many have gone to significant lengths to assess, in a critical manner, how their cells perform in a number of ways. If structure-reactivity relationships are to be considered quantitative in all respects, these types of confirmatory studies must be entered into. From a catalytic point of view, there are three very important considerations, when using any cell developed for operando X-ray measurements, that must be established in order to have complete confidence in any result derived from such cells. The first is the degree to which they are inherently isothermal. That temperature gradients might develop, as a result of catalytic operation, is entirely possible, but they should certainly not exist prior to the onset of any chemistry if an accurate picture of the behavior (be it kinetic or structural) is to be arrived at. Moreover, this is a criterion that

27

592

applies to any form of sample presentation and must been seen a key foundation to any catalysis experiment, irrespective of applied technique. A case in point is the use of capillary micro-reactors that are heated in a unidirectional manner by a hot air stream. In their original application [53], which was the study of Cu/ZnO/Al2O3 catalysts for methanol synthesis, a significant amount of confirmatory work was engaged in prior to any synchrotron usage, to make sure that, from the point of view of reactivity, they produced results that were consistent with much larger reactor systems. More than that, the very nature of the materials under study meant that capillary diameters of  1 mm diameter had to be used in order to be able to make measurements upon them (for XAS and XRD) in transmission, and, even then, actions to guard against non-isothermality were put in place. Since this early work, these types of systems have been widely adopted at synchrotrons, as they permit a great many different types of measurement to be made. However, they have often been used to study catalyst beds of much larger dimensions without any action taken to ensure that the heating of the sample remained uniform or acceptably close to it. Indeed, a recent study [173] has shown how easy it can be to enter into a situation, using a unidirectional heating of capillaries of only 2 mm in diameter, where both axial and radial temperature gradients exist across the sample that make it unfit for purpose in any quantitative sense. The solution to that particular issue was to move to another (bidirectionally heated [174]) design and verify that this produced an isothermal zone of twice the axial extension than the sample beds to be measured. A second critical aspect of the study of catalysis, one which is more difficult to assess in operando reactors, is that of how the reactants contact the sample as they pass over (for instance, in a pressed disk system) or through (as in a plug-flow reactor) it. Furthermore, a distinction should be made between these two types of sample presentation. The former, at least from the point of view of EXAFS, is preferred, as it guarantees a high level of spatial uniformity within the sample. However, in guaranteeing this aspect, control over other important parameters (for instance, porosity, penetration of the reactant into the disk, and the concept of a contact time) is lost. As such, while a pressed sample, even if uniformly heated across its dimensions, may yield quantitative structural data in terms of the phases and speciation present, it is unlikely that reaction kinetics will be quantitatively accessible using such an arrangement, a weakness that has been previously demonstrated [175]. By contrast, the increasingly common approach to the operando study of catalysts, that of plug flow, retains the possibility of defining a contact time, and permits the restoration if quantitative kinetics alongside information

M. A. Newton et al.

regarding structure and speciation. This configuration also permits facile spatial exploration of the catalyst bed (see section “One-Dimensioanl Spatial Investigation of Reactor Beds”). However, for these positive aspects to be properly realized, the contact pattern of the reactants should be as uniform as possible within the sample bed, and both the formation of significant pressure drops across the sample be, or bypassing of the sample itself along the walls of the containing vessel, must be avoided. While the presence of significant pressure drops across a bed may be relatively easily established and attended to, the issue of bypassing is far more difficult to spot and/or quantify. By and large in this case, only guidelines – that can be found in various engineering textbooks – exist to relate the average particle size of the packed sample to the overall radial dimension of the container. A reasonable rule of thumb in this respect would be to ensure that the sieve fraction employed should be both narrow and of the order at least 10–20 times less than the internal dimensions of the reactor. Lastly, we come to the issue of how to ensure that one can quantitatively, and accurately, correlate what is being measured by the X-ray probe and what is being measured by whatever other probes of reactivity might be being applied. Given the strong trend toward the use of smaller and smaller X-ray probes, there exists a growing disconnect between the volume of the catalyst bed sampled by the X-rays and that detected by downstream methods of reactivity assessment (e.g., mass spectrometry or gas chromatography), which integrate over the entirety of the sample. This issue also applies to any other simultaneously applied method, such as infrared or Raman, each of which will sample different volumes according to the nature of the sample. This last aspect is a particularly thorny issue as, quite legitimately, not all parts of the bed may show the same level of reactivity, as a result of, for instance, reactive gradients developing as a natural result of the chemistry under study (see section “One-Dimensional Spatial Investigation of Reactor Beds”), or the formation of localized hotspots resulting from exothermic processes. Equally, however, the use of very small and powerful beams makes the determination of the integral reactivity/selectivity of a given system in the absence of X-rays a poor guide as to whether what is being measured under illumination results from the chemistry at work or from unwanted effects that arise from the application of the X-rays themselves (vide supra). At present, and in this respect, there appears no obvious fail-safe manner to ensure that what is observed using the X-ray probe is an accurate reflection of the intrinsic reactivity characteristics of the sample under study. The best that may be suggested is either direct comparison to other, larger, catalytic reactors, as was undertaken many years ago by Clausen and co-workers [173], or the cross-referencing of the behavior of a system to aspects of behavior (be they

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

kinetics or selectivity) that are already known from other methods [53]. However, even when doing this, the possibility that a small and focused X-ray probe is altering what is being observed in terms of the behavior of the material under reaction conditions, while having no readily discernable impact on the measured integral reactivity of the system, exists. This being the case, we might reasonably consider where laboratory-based reactor systems, which yield X-ray fluxes and brilliance orders of magnitude less powerful than even a bending magnet at currently existing synchrotrons, might fit into the operando landscape. These sources offer by far the most reduced possibility that the application of X-rays may lead to reactive behavior that is not intrinsic to the materials themselves. The corollary of this is that they are currently extremely limited in terms of the rapidity with which they may sample a working system with sufficient data quality to permit in-depth analysis. As such, and at present, they are therefore best suited to simpler, ex situ, measurements of samples using XAS or indeed XES. In terms of operando spectroscopy, their most obvious present application (as with the example given above, [148]) is in performing experiments on timescales (e.g., several days to more than a week) that would not be possible within the extremely competitive and time-restricted world of synchrotron-based study. A final question, regards how these sources might evolve within the near future, and to what degree the significant gap in their performance, even compared to bending magnets ay synchrotron sources, might be closed? At present, commercial laboratory sources operate exclusively through using dispersive optics coupled to relatively low-power (100–300 watts) X-ray tubes. As such, their intrinsic photon throughput massively reduced compared to a bending magnet, and the total flux is spread across an energetic bandwidth that is determined by the nature of energy-dispersing crystal used. More than that, as a result of their optical construction, to build up a XAS spectrum, the sample and detector (generally a single-element silicon diode) must be physically rotated over the angular range required for the desired spectrum to be acquired. Given this, there appear to be only two routes to enhance their performance: the implementation of much more powerful sources and/or the use of significantly more effective means of detection.

27.5

Conclusions

Operando XAS spectroscopy continues to evolve into an extremely diverse number of areas and applications in catalysis. The importance of XAS, in all its forms, for obtaining highly specific information regarding all aspects of catalysis – from fundamental speciation to the investigation of more general aspects of the behavior of catalysts within reactor

593

situations – cannot be understated. This is especially the case where XAS may be combined with other probes (for instance, X-ray scattering, vibrational spectroscopies, or, for instance, voltammetry, or mass spectrometry) that can inform us regarding aspects of speciation and behavior that XAS either cannot or can only do so with a reduced specificity or in particular circumstances (for instance, soft X-ray XAS). The continued evolution of synchrotrons, and technologies allied to them (for instance, detectors and the counting chains which lie behind them), means that XAS will only become more powerful in its ability to address, in both space and time, those fundamental aspects of material behavior that are the foundation of quantitative structure function relationships. That said, the immense quantitative power of modern XAS – and indeed many other methods – may only be realized if it is conducted in full respect of the processes and chemistry under study and with the knowledge of what it can and cannot do. Moreover, this extends to the manner in which it is achieved (i.e., sample environments verified to be suitable for the process and chemistry to be studied) and, for instance, that the applied X- rays might inadvertently and unwantedly alter the behavior of the system under study simply through their application. By and large, the former is attended to very well. Virtually all the studies presented in this chapter have been preceded by extensive development and testing of the sample environment applied to ensure that it is fit for the intended purpose. That said, practitioners of operando study, be it with or without the use of X-rays, must remain vigilant in this respect as things can go easily awry and sometimes for the most unexpected of reasons [53]. That a probe (such as X-rays) should not alter or interfere in any way with the materials/process under study is a founding axiom to any form of meaningful experimentation, and it is in this respect that present and future operando XAS investigations face their most challenging examination. The present and foreseen upgrades of third-generation sources will produce the most powerful synchrotron X-ray resources yet contrived, and both sides of this decidedly double-edge sword are required to be considered. On the one hand, the very considerable increases, in terms of X-ray coherence, accessible probe dimensions, and throughput (especially at high X-ray energies), will open up entirely new vistas for the study of catalysts and the processes they facilitate. On the other hand, however, the increased power densities that may result from these developments must necessarily place a proportionately higher burden of circumspection regarding the possible effects of these increased loads upon the materials and processes desired to be studied. Moreover, and as recent results have demonstrated [171], this is not a matter that only concerns medical physicists or those who study so-called soft matter. Nor is it,

27

594

these days, only a concern to those using powerful insertion device-based resources; bending magnets at third-generation synchrotrons are also more than capable of unwantedly, and chronically, interfering with the restoration of meaningful results from inorganic materials engaged in reactive chemistry [171]. Being vigilant to these possibilities, and then actively reporting them as and when they might be encountered, is an area of operando research, and indeed X-ray science in general, that could certainly be improved to the benefit of all. In this respect, we might look to the continuing development for operando study of laboratory-based spectrometers as providing some sort of alternative [20–27]: firstly, as they can provide a new means to make different types of operando experiments, especially those that are concerned with longterm process application, and, secondly, as these sources are the least likely to result in unwanted effects that may arise from the application of X-rays. However, whether these types of sources can and will be developed further, to bridge what is a very big gap between their present operational capacity and those of even bending magnet resources that do not focus their X-rays, remains to be seen. What we do know is that modern-day synchrotron resources, and their upgraded counterparts, provide the finest, and most powerful, palette of methods and opportunities across the whole spectrum of X-ray-based techniques. However, a fine palette is no guarantee of a great picture, and power is nothing without control. As such, the realization of the insight that could be achieved using these advanced resources depends very much upon how and why the powerful paint is applied; and that is an ecumenical matter. Acknowledgments MAN would like to acknowledge Shell Global Solutions for part finding of his present position. He would also like to thank all those with whom he has worked with over the years, both within and without central facilities, from whom he has learned so much.

References 1. Gerrrero-Perez, M.O., Banares, M.: Operando Raman study of alumina-supported Sb–V–O catalyst during propane ammoxidation to acrylonitrile with on-line activity measurement. Chem. Commun., 1292–1293 (2002) 2. Bunker, G.: Introduction to XAFS: A Practical Guide to X-Ray Absorption Fine Structure Spectroscopy. Cambridge University Press (2010) 3. van Bokhoven, J.A., Lamberti, C. (eds.): X-Ray Absorption and X-Ray Emission Spectroscopy: Theory and Applications, 1st edn. Wiley (2016) 4. Evans, J.: X-Ray Absorption Spectroscopy for the Chemical and Materials Sciences. Wiley (2017) 5. Newton, M.A., Dent, A.J.: Chapter 3. In: Chupas, P.J., Hanson, J., Rodriguez, J.A. (eds.) In-Situ Characterization of Heterogeneous Catalysts, pp. 75–118. Wiley (2013)

M. A. Newton et al. 6. Mueller, O., Nachtegaal, M., Just, J., Lutzenkirchen-Hecht, D., Frahm, R.: Quick-EXAFS setup at the SuperXAS beamline for in situ X-ray absorption spectroscopy with 10 ms time resolution. J. Synchrotron Radiat. 23, 260–266 (2016) 7. Stotzel, J., Lutzenkirchen-Hecht, D., Fonda, E., De Oliveira, N., Briois, V., Frahm, R.: Quick scanning monochromator for millisecond in situ and in operando X-ray absorption spectroscopy. Rev. Sci. Instrum. 79, 083107 (2008) 8. Labiche, J.-C., Mathon, O., Pascarelli, S., Newton, M.A., Ferre, G.G., Curfs, C., Vaughan, G., Homs, A., Carreiras, D.F.: The fast readout low noise camera as a versatile x-ray detector for time resolved dispersive extended x-ray absorption fine structure and diffraction studies of dynamic problems in materials science, chemistry, and catalysis. Rev. Sci. Instrum. 78, 091301 (2007) 9. Headspith, J., Groves, G., Luke, P.N., Kogimtzis, M., Salvini, G., Thomas, S.L., Farrow, R.C., Evans, J., Rayment, T., Lee, J.S., Goward, W.D., Amman, A., Mathon, O., Diaz-Moreno, S.: The fast readout low noise camera as a versatile x-ray detector for time resolved dispersive extended x-ray absorption fine structure and diffraction studies of dynamic problems in materials science, chemistry, and catalysis. In: IEEE Nuclear Science Symposium and Medical Imaging Conference, pp. 2421–2428 (2007) 10. Headspith, J., Salvini, G., Thomas, S.L., Derbyshire, G., Dent, A., Rayment, T., Evans, J., Farrow, R., Anderson, C., Cliche, J., Dobson, B.R.: XSTRIP-a silicon microstrip-based X-ray detector for ultra-fast X-ray spectroscopy studies. Nucl. Instrum. Methods Phys. Res., Sect. A. 512, 239–244 (2003) 11. Muller, O., Stotzel, J., Lutzenkirchen-Hecht, D., Frahm, R.: Gridded ionization chambers for time resolved X-ray absorption spectroscopy. JPCS. 425, 092010 (2013) 12. Bressler, C., Chergui, M.: Ultrafast X-ray absorption spectroscopy. Chem. Rev. 104, 1781–1812 (2004) 13. Bressler, C., Chergui, M.: Molecular structural dynamics probed by ultrafast X-ray absorption spectroscopy. Annu. Rev. Phys. Chem. 61, 263–282 (2010) 14. Chen, L.X., Zhang, X., Shelby, M.L.: Recent advances in ultrafast spectroscopy in the chemical sciences. Chem. Sci. 5, 4136–4152 (2014) 15. Kurian, R., Kunnus, K., Wernet, P., Butorin, S.M., Glatzel, P., de Groot, F.M.F.: Intrinsic deviations in fluorescence yield detected x-ray absorption spectroscopy: the case of the transition metal L 2,3-edges. J. Phys. C. 24, 452201 (2012) 16. Achkar, A.J., Regier, T.Z., Wadati, H., Kim, Y.-J., Zhang, H., Hawthorn, D.G.: Bulk sensitive X-ray absorption spectroscopy free of self-absorption effects. Phys. Rev. B. 83, 081106 (2011) 17. Frentrup, W., Schroeder, D., Manzke, R.: Correction of selfabsorption on XAS measurements in fluorescence mode. J. Phys. IV. 7(C2), 509–510 (1997) 18. Bauer, M.: HERFD-XAS and valence to core XES: new tools to push the limits in research with hard X-rays? Phys. Chem. Chem. Phys. 16, 13827–13837 (2014) 19. Pollock, C.J., DeBeer, S.: Insights into the geometric and electronic structure of transition metal centers from valence-to-core X-ray emission spectroscopy. Acc. Chem. Res. 48, 2967–2975 (2015) 20. Anklamm, L., Schlesiger, C., Malzer, W., Grotzsch, D., Neitzel, M., Kanngiesser, B.: A novel von Hamos spectrometer for efficient X-ray emission spectroscopy in the laboratory. Rev. Sci. Instrum. 85, 053110 (2014) 21. Seidler, G.T., Mortensen, D.R., Remesnik, A.J., Pacold, J.I., Ball, N.A., Barry, N., Styczinski, M., Hoidn, O.R.: A laboratory-based hard x-ray monochromator for high-resolution x-ray emission spectroscopy and x-ray absorption near edge structure measurements. Rev. Sci. Instrum. 85, 113906–111391 (2014) 22. Legall, H., Stiel, H., Schnurer, M., Pagels, M., Kanngiesser, B., Mueller, M., Beckhoff, B., Grigorieva, I., Antonov, A., Arkadiev, V., Bjeoumikhov, A.: An efficient X-ray spectrometer based on thin

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

mosaic crystal films and its application in various fields of X-ray spectroscopy. J. Appl. Crystallogr. 42, 572–579 (2009) 23. Schlesiger, C., Anklamm, L., Stiel, H., Malzer, W., Kanngiesser, B.: XAFS spectroscopy by an X-ray tube based spectrometer using a novel type of HOPG mosaic crystal and optimized image processing. J. Anal. At. Spectrom. 30, 1080–1085 (2015) 24. Nemeth, Z., Szlachetko, J., Bajnoczi, E.G., Vanko, G.: Laboratory von Hamos X-ray spectroscopy for routine sample characterization. Rev. Sci. Instrum. 87, 103105 (2016) 25. Mortensen, D.R., Seidler, G.T.: Determination of hexavalent chromium fractions in plastics using laboratory-based, high-resolution X-ray emission spectroscopy. J. Electron Spectrosc. Relat. Phenom. 215, 8–15 (2017) 26. Bes, R., Ahopelto, T., Honkanen, A.P., Huotari, S., Leinders, G., Pakarinen, J., Kvashnina, K.: Laboratory-scale X-ray absorption spectroscopy approach for actinide research, experiment at the uranium L3-edge. J. Nucl. Mater. 507, 50–53 (2018) 27. Honkanen, A.-P., Ollikkala, S., Ahopelto, T., Kallio, A.-J., Blomberg, M., Huotari, S.: Johann-type laboratory-scale X-ray absorption spectrometer with versatile detection modes, http:// arxiv.org/abs/1812.01075 28. Kutzenkirchen-Hecht, D., Grundmann, S., Frahm, R.: PiezoQEXAFS with fluorescence detection: fast time-resolved investigations of dilute specimens. J. Synchrotron Radiat. 8, 6–9 (2001) 29. Pascarelli, S., Neisius, T., De Panfilis, S.: Turbo-XAS: dispersive XAS using sequential acquisition. J. Synchrotron Radiat. 6, 1044–1050 (1999) 30. Nagai, Y., Dohmae, K., Ikeda, Y., Takagi, N., Tanabe, T., Hara, N., Guilera, G., Pascarelli, S., Newton, M.A., Kuno, O., Jiang, H.Y., Shinjoh, H., Matsumoto, S.: In situ redispersion of platinum autoexhaust catalysts: an on-line approach to increasing catalyst lifetimes? Angew. Chem. Int. Ed. 47, 9303–9306 (2008) 31. Nagai, Y., Dohmae, K., Ikeda, Y., Takagi, N., Hara, N., Tanabe, T., Guilera, G., Pascarelli, S., Newton, M.A., Takahashi, N., Shinjoh, H., Matsumoto, S.: In situ observation of platinum sintering on ceria-based oxide for autoexhaust catalysts using turbo-XAS. Catal. Today. 175, 135–140 (2011) 32. Błachucki, W., Szlachetko, J., Hoszowska, J., Dousse, J.-C., Kayser, Y., Nachtegaal, M., Sá, J.: High energy resolution off-resonant spectroscopy for X-ray absorption spectra free of self-absorption effects. Phys. Rev. Lett. 112, 173003 (2014) 33. Błachucki, W., Hoszowska, J., Douse, J.-C., Kayser, Y., Stachura, R., Tyra, K., Wojtaszek, K., Sá, J., Szlachetko, J.: High energy resolution off-resonant spectroscopy: a review. Spectrochim. Acta B. 136, 23–33 (2017) 34. Szlachetko, J., Ferri, D., Machionni, V., Kambolis, A., Safonova, O.V., Milne, C.J., Kröcher, O., Naachtegaal, M., Sá, J.: Subsecond and in situ chemical speciation of Pt/Al2O3 during oxidation reduction cycles monitored by high-energy resolution off-resonant X-ray spectroscopy. J. Am. Chem. Soc. 135, 19071–19074 (2013) 35. Chang, E.M., Marcus, M.A., Fakra, S., El-Naggar, M., Mathies, R.A., Alisviatos, A.P.: Millisecond kinetics of nanocrystal cation exchange using microfluidic X-ray absorption spectroscopy. J. Phys. Chem. A. 111, 12210–12215 (2007) 36. Clark, A.H., Steiger, P., Bornmann, B., Hitz, S., Frahm, R., Ferri, D., Nachtegaal, M.: Fluorescence-detected quick-scanning X-ray absorption spectroscopy. J. Synchrotron Radiat. 27, 681–688 (2020) 37. Broennimann, C., Eikenberry, E.F., Heinrich, B., Horisberger, R., Huelsen, G., Pohl, E., Schmitt, B., Schulze-Briese, C., Suzuki, M., Tomizaki, T., Toyokawa, H., Wagner, A.: The PILATUS 1M detector. J. Synchrotron Radiat. 13, 120–130 (2006) 38. Dowsett, M.G., Adriaens, A., Jones, G.K.C., Poolton, N., Fiddy, S., Nikitenko, S.: Optically detected X-ray absorption spectroscopies measurements as a means of monitoring corrosion layers on copper. Anal. Chem. 80, 8717–7724 (2008)

595 39. Dowsett, M., Adriaens, A., Martin, C., Bouchenoire, L.: Real time observation of X-ray induced surface modification using simultaneous XANES and XEOL-XANES. Anal. Chem. 84, 4866–4872 (2012) 40. Dowsett, M., Hand, M., Sabbe, P.J., Thompson, P., Adriaens, A.: XEOM 1 – a novel microscopy system for the chemical imaging of heritage metal surfaces. Herit. Sci. 3, 14 (2015) 41. Li, J., Liu, L., Sham, T.K.: 2D XANES-XEOL spectroscopy studies on morphology dependent phase transformation and corresponding luminescence from hierarchical TiO2 nanostructures. Chem. Mater. 27, 3021–3029 (2015) 42. Wang, Z., Guo, X., Sham, T.K.: 2D XANES-XEOL mapping: observation of enhanced band gap emission from ZnO nanowire arrays. Nanoscale. 6, 6531–6536 (2014) 43. Grunwaldt, J.-D., Baiker, A.: Axial variation of the oxidation state of Pt-Rh/Al2O3 during partial methane oxidation in a fixed-bed reactor: an in situ X-ray absorption spectroscopy study. Catal. Lett. 99, 5–12 (2005) 44. Newton, M.A., Jyoti, B., Dent, A.J., Diaz-Moreno, S., Fiddy, S.G., Evans, J.: Rapid monitoring of the nature and interconversion of supported catalyst phases and of their influence upon performance: CO oxidation to CO2 by gamma-Al2O3 supported Rh catalysts. Chem. Eur. J. 12, 1975–1985 (2006) 45. Hannemann, S., Grunwaldt, J.D., Vegten, N., Baiker, A., Boye, P., Schroer, C.G.: Distinct spatial changes of the catalyst structure inside a fixed-bed microreactor during the partial oxidation of methane over Rh/Al2O3. Catal. Today. 126, 54–63 (2007) 46. Hanneman, S., Grunwaldt, J.D., Kimmerle, B., Baiker, A., Boye, P., Schroer, C.: Axial changes of catalyst structure and temperature in a fixed-bed microreactor during noble metal catalysed partial oxidation of methane. Top. Catal. 52, 1360–1370 (2009) 47. Grunwaldt, J.D., Beier, M., Kimmerle, B., Bauker, A., Nachtegaal, M., Griesbock, B., Lützenkirchen-Hecht, D., Stötzel, J., Frahm, R.: Structural changes of noble metal catalysts during ignition and extinction of the partial oxidation of methane studied by advanced QEXAFS techniques. Phys. Chem. Chem. Phys. 11, 8779–8789 (2009) 48. Brazier, J.B., Nguyen, B.N., Adrio, L.A., Barreiro, E.M., Leong, W.P., Newton, M.A., Figueroa, S.J.A., Hellgardt, K., Hii, K.K.: Catalysis in flow: Operando study of Pd catalyst speciation and leaching. Catal. Today. 229, 95–103 (2014) 49. Newton, M.A., Brazier, J.B., Barreiro, E.M., Nguyen, B.N., Parry, S., Emerich, H., Adrio, L.A., Hellgardt, K., Hii, K.K.: Operando XAFS of supported Pd nanoparticles in flowing ethanol/water mixtures: implications for greener catalysis. Green Chem. 18, 406–441 (2016) 50. Newton, M.A., Brazier, J.B., Barreiro, E.M., Emerich, H., Adrio, L.A., Mulligan, C.J., Hellgardt, K., Hii, K.K.: Restructuring of supported Pd by green solvents: an operando quick EXAFS (QEXAFS) study. Catal. Sci. Technol. 6, 8525–8531 (2016) 51. Newton, M.A., Ferri, D., Mulligan, C.J., Alxneit, I., Emerich, H., Thompson, P.B.J., Hii, K.K.: In situ study of metal leaching from Pd/Al2O3 induced by K2CO3. Catal. Sci. Technol. 10, 466–474 (2020) 52. Checchia, S., Mulligan, C.J., Emerich, H., Alxneit, I., Krumeich, F., Di Michiel, M., Thompson, P.B.J., Hii, K.K., Ferri, D., Newton, M.A.: Pd-LaFeO3 catalysts in aqueous ethanol: Pd reduction, leaching, and structural transformations in the presence of a base. ACS Catal. 10, 3933–3944 (2020) 53. Newton, M.A., Checchia, S., Knorpp, A.J., Stoian, D., van Beek, W., Emerich, H., Longo, A., van Bokhoven, J.A.: On isothermality in some commonly used plug flow reactors for X-ray based investigations of catalysts. Catal. Sci. Technol. 9, 3081–3089 (2019) 54. McCaulley, J.A.: In-situ X-ray absorption spectroscopy studies of hydride and carbide formation in supported palladium catalysts. J. Phys. Chem. 97, 10372–10379 (1993)

27

596 55. Rose, A., South, O., Harvey, I., Diaz-Moreno, S., Owen, J.R., Russell, A.E.: In situ time resolved studies of hydride and deuteride formation in Pd/C electrodes via energy dispersive X-ray absorption spectroscopy. Phys. Chem. Chem. Phys. 7, 366–372 (2005) 56. Rose, A., Maniguet, S., Mathew, R.J., Slater, C., Yao, J., Russell, A.E.: Hydride phase formation in carbon supported palladium nanoparticle electrodes investigated using in situ EXAFS and XRD. Phys. Chem. Chem. Phys. 5, 3220–3225 (2003) 57. Alfonis, K., Colberg, J., Dunn, P.J., Fevig, T., Jennings, S., Johnson, T.A., Kleine, H.P., Knight, C., Nagy, M.A., Perry, D.A., Stefaniak, M.: Green chemistry tools to influence a medicinal chemistry and research chemistry based organization. Green Chem. 10, 31–36 (2008) 58. Dunn, P.J.: The importance of green chemistry in process research and development. Chem. Soc. Rev. 41, 1452–1461 (2012) 59. Prat, D., Hayler, J., Wells, A.: CHEM21 selection guide of classical- and less classical-solvents. Green Chem. 18, 288–296 (2016) 60. Pachon, L.D., Rothenburg, G.: Tansition-metal nanoparticles: synthesis, stability, and the leaching issue. Appl. Organomet. Chem. 22, 288–299 (2008) 61. Cantillo, D., Kappe, O.C.: Immobilized transition metal catalysts for cross-couplings in a continuous flow – a critical assessment of the reaction mechanism and metal leaching. ChemCatChem. 6, 3286–3305 (2014) 62. Reimann, S., Stotzel, J., Frahm, R., Kleist, W., Grunwaldt, J.D., Baiker, A.: Identification of the active species generated from supported Pd catalysts in Heck reactions: an in situ quick scanning EXAFS investigation. J. Am. Chem. Soc. 133, 3921–3930 (2011) 63. Selinsek, M., Benedikt, J., Doronkin, D.E., Sheppard, T.L., Grunwaldt, J.-D., Dittmayer, R.: Revealing the structure and mechanism of palladium during direct synthesis of hydrogen peroxide in continuous flow using operando spectroscopy. ACS Catal. 8, 2456–2557 (2018) 64. Doronkin, D.E., Wang, S., Sharapa, D., Deschner, B., Sheppard, T., Zimina, A., Studt, F., Dittmeyer, R., Behrens, S., Grunwaldt, J.D.: Dynamic structural changes of supported Pd, PdSn, and PdIn nanoparticles during continuous flow high pressure direct H2O2 synthesis. Catal. Sci. Technol. 10, 4746–4742 (2020) 65. Martin, D.J., Decarolis, D., Tucoulou, R., Martínez-Criado, G., Beale, A.M.: Towards the interrogation of the behaviour of a single nanoparticle under realistic catalytic reaction conditions. Catal. Struct. React. 3, 63–70 (2017) 66. Ouvrard, G., Zerrouki, M., Soudan, P., Lestriez, B., Masquelier, C., Morcrette, M., Hamelet, S., Belin, S., Flank, A.M., Baudelet, F.: Heterogeneous behavior of the lithium battery composite. J. Power Sources. 229, 16–21 (2013) 67. Sanchez, D.F., Simionovici, A.S., Lemelle, L., Cuatero, V., Mathon, O., Pascarelli, S., Bonnin, A., Shapiro, K., Konhauser, K., Grolimund, D., Bleuet, P.: Microanalysis by energy dispersive X-ray absorption spectroscopy tomography. Sci. Rep. 7, 16453 (2017) 68. Ihli, J., Sanchez, D.F., Jacob, R.R., Cuatero, V., Mathon, O., Krumreich, F., Borca, C., Huthwelker, T., Cheng, W.-C., Shu, Y., Pascarelli, S., Grolimund, D., Menzel, A., van Bokhoven, J.A.: Localization and speciation of iron impurities within a fluid catalytic cracking catalyst. Angew. Chem. Int. Ed. 56, 14031–14035 (2017) 69. Qi, P., Samadi, N., Martinson, M., Ponomerenko, O., Bassey, B., Gomz, A., George, G.N., Pickering, I.J., Chapman, L.D.: Wide field imaging energy dispersive X-ray absorption spectroscopy. Sci. Rep. 9, 17734 (2019) 70. Grunwaldt, J.D., Kimmerle, B., Baiker, A., Boye, P., Schroer, C.G., Galtzel, P., Borca, C.N., Beckmann, F.: Catalysts at work: from integral to spatially resolved X-ray absorption spectroscopy. Catal. Today. 145, 67–278 (2008)

M. A. Newton et al. 71. Hoffman, G., Rochet, A., Baier, S., Casapu, M., Ritter, S., Wilde, F., Orgurreck, M., Beckmann, F., Grunwaldt, J.D.: Ageing effects on exhaust gas catalysts: microscopic changes captured by X-ray tomography. JPCS. 499, 012017 (2014) 72. Grunwaldt, J.D., Wagner, J.B., Dunin-Borkowski, R.E.: Imaging catalysts at work: a hierarchical approach from the macro- to the meso- and nano-scale. ChemCatChem. 5, 62–80 (2013) 73. Portela, R., Perez-Ferreras, S., Serrano-Lotina, A., Bañares, M.A.: Engineering operando methodology: understanding catalysis in time and space. Front. Chem. Sci. Eng. 12, 509–536 (2018) 74. Gambino, M., Vesely, M., Filez, M., Oord, R., Sanchez, D.F., Grolimund, D., Nesterenko, N., Monoux, D., Maquet, M., Meirer, F., Weckhuysen, B.M.: Nickel poisoning of a cracking catalyst unraveled by single-particle X-ray fluorescence-diffraction-absorption tomography. Angew. Chem. Int. Ed. 59, 3922–3927 (2020) 75. Chapter on imaging and tomography in this book 76. Matras, D., Vamvakeros, A., Jacques, S.D.M., Price, S.W.T., Pritchard, J., Beale, A.M.: Chapter X. In: Springer Handbook of Advanced Catalyst Characterization, this volume Advances in Chemical Tomography, pp. A–B. Springer (2020) 77. Katayama, M., Sumiwaka, K., Miyhara, R., Yamashige, H., Arai, H., Uchimoto, Y., Ohta, T., Inada, Y., Z.: Ogumi: X-ray absorption fine structure imaging of inhomogeneous electrode reaction in LiFePO4 lithium-ion battery cathode. J. Power Sources. 269, 994–999 (2014) 78. Katayama, M., Sumiwaka, K., Hayashi, K., Ozutsumi, K., Ohta, T., Inada, Y.: Development of a two dimensional imaging system of X-ray absorption fine structure. J. Synchrotron Radiat. 19, 717–721 (2012) 79. Orikasa, Y., Gogyo, Y., Yamashige, H., Katayama, M., Chen, K.Z., Mori, T., Yamamoto, K., Masese, T., Iadna, Y., Ohta, T., Siroma, Z., Kato, S., Kinoshita, H., Arai, H., Ogumi, Z., Uchimoto, Y.: Ionic conduction in lithium ion battery composite electrode governs cross-sectional reaction distribution. Sci. Rep. 6, 26382 (2016) 80. Grunwaldt, J.D., Schroer, C.G.: Hard and soft X-ray microscopy and tomography in catalysis: bridging the different time and length scales. Chem. Soc. Rev. 39, 4741–4753 (2010) 81. Jacques, S.D.M., Di Michiel, M., Beale, A.M., Sochi, T., O’Brien, M.G., Espinosa-Alonso, L., Weckhuysen, B.M., Barnes, P.: Dynamic X-ray diffraction computed tomography reveals realtime insight into catalysts active phase evolution. Angew. Chem. Int. Ed. 43, 10148–10152 (2011) 82. O’Brien, M.G., Jacques, S.D.M., Di Michiel, M., Barnes, P., Weckhuysen, B.M., Beale, A.M.: Active phase evolution in single Ni/Al2O3 methanation catalyst bodies studied in real time using combined μ-XRD-CT and μ-absorption-CT. Chem. Sci. 3, 509–523 (2012) 83. Jacques, S.D.M., Di Michiel, M., Kimber, S.A.J., Yang, X.H., Cernik, R.J., Beale, A.M., Billinge, S.J.L.: Pair distribution function computed tomography. Nat. Commun. 4, 2536 (2013) 84. Beale, A.M., Jacques, S.D.M., Gibson, E.K., Di Michiel, M.: Progress towards five dimensional imaging of functional materials under process conditions. Coord. Chem. Rev. 227, 208–223 (2014) 85. Matsui, H., Ishiguro, N., Uruga, T., Sekizawa, O., Higashi, K., Maejima, N., Tada, M.: Operando 3D visualization of migration and degradation of a platinum cathode catalyst in a polymer electrolyte fuel cell. Angew. Chem. Int. Ed. 56, 9371–9375 (2017) 86. Matsui, H., Maejima, N., Ishiguro, N., Tan, Y., Uruga, T., Sekizawa, O., Sakata, T., Tada, M.: Operando XAFS imaging of distribution of Pt cathode catalysts in PEFC MEA. Chem. Rec. 19, 1380–1392 (2019) 87. Thibault, P., Dierolf, M., Menze, A., Bunk, O., David, C., Pfeiffer, F.: High-resolution X-ray diffraction microscopy. Science. 321, 379–382 (2008) 88. Wise, A.M., Weker, J.N., Kalirai, S., Farmand, M., Shapiro, D.A., Meirer, F., Weckhuysen, B.M.W.: Nanoscale chemical imaging of

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

an individual catalyst particle with soft X-ray ptychography. ACS Catal. 6, 2178–2181 (2016) 89. Hirose, M., Ishiguro, M., Shimomura, K., Burdet, N., Matsui, H., Tada, M., Takahashi, Y.: Visualization of heterogeneous oxygen storage behavior in platinum-supported cerium-zirconium oxide three-way catalyst particles by hard X-ray spectro-ptychography. Angew. Chem. Int. Ed. 57, 1474–1479 (2018) 90. Tromp, M., Dent, A.J., Headspith, J., Easun, T.L., Sun, X.Z., George, M.W., Mathon, O., Smolentsev, G., Hamilton, M.L., Evans, J.: Energy dispersive XAFS: characterization of electronically excited states of copper(I). J. Phys. Chem. B. 117, 7381–7387 (2013) 91. Stickrath, A.B., Mara, M.W., Lockhard, J.V., Harpham, M.R., Huang, J., Zhang, X.Y., Attenkofer, K., Chen, L.X.: Detailed transient heme structures of Mb-CO in solution after CO dissociation: an X-ray transient absorption spectroscopic study. J. Phys. Chem. B. 117, 4705–4712 (2013) 92. Smolentsev, G., Syndstrom, V.: Time-resolved X-ray absorption spectroscopy for the study of molecular systems relevant for artificial photosynthesis. Coord. Chem. Rev. 304, 117–132 (2015) 93. Zhang, X.Y., Canton, S.E., Smolentsev, G., Wallentin, C.J., Liu, Y.Z., Kong, Q.Y., Attenkofer, K., Stickrath, A.B., Mara, M.W., Chen, L.X., Warnmark, K., Sundstrom, V.: Highly accurate excited-state structure of [Os(bpy)(2)dcbpy](2+) determined by X-ray transient absorption spectroscopy. J. Am. Chem. Soc. 136, 8804–8809 (2014) 94. Smolentsev, G., Cecconi, B., Guda, A., Chavanot-Kerlidon, M., van Bokhoven, J.A., Nachtegaal, M., Artero, V.: Microsecond X-ray absorption spectroscopy identification of CoI intermediates in cobaloxime catalyzed hydrogen evolution. Chem. Eur. J. 21, 15158–11516 (2015) 95. Li, Z.J., Zhan, F., Xiao, H., Zhang, X., Kong, Q.Y., Fan, X.B., Liu, W.-Q., Huang, M.-Y., Huang, C., Gao, Y.-J., Li, X.-B., Meng, Q.Y., Feng, K., Chen, B., Tung, C.-H., Zhao, H.-F., Tao, Y., Wu, L.Z.: Tracking Co(I) intermediate in operando in photocatalytic hydrogen evolution by X-ray transient absorption spectroscopy and DFT calculation. J. Phys. Chem. Lett. 7, 5253–5258 (2016) 96. Song, F.Y., More, T., Schilling, M., Smolentsev, G., Azzaroli, N., Fox, T., Luber, S., Patzke, G.R.: {Co4O4} and Co, Ni4O4 cubane water oxidation catalysts as surface cut-outs of cobalt oxides. J. Am. Chem. Soc. 139, 14198–14208 (2017) 97. Huijser, A., Pan, Q., van Duinen, D., Laursen, M.G., Al Nahhas, A., Chabera, P., Feitag, L., Gonzalez, L., Kong, Q.Y., Zhang, X.Y., Haldrup, K., Browne, W.R., Smolentsev, G., Uhlig, J.: Shedding light on the nature of photo-induced states formed in hydrogengenerating supramolecular RuPt, photocatalysts by ultrafast spectroscopy. J. Phys. Chem. A. 122, 6396–6406 (2018) 98. Smolentsev, G., Soldatov, M.A., Probst, B., Bachmann, C., Azzaroli, N., Alberto, R., Nachtegaal, M., van Bokhoven, J.A.: Structure of the CoI intermediate of a cobalt pentapyridyl catalyst for hydrogen evolution revealed by time-resolved X-ray spectroscopy. ChemSusChem. 11, 3087–3091 (2018) 99. Siebel, A., Gorlin, Y., Durst, J., Proux, O., Hasché, F., Tromp, M., Gasteiger, H.A.: Identification of catalyst structure during the hydrogen oxidation reaction in an operating PEM fuel cell. ACS Catal. 6, 7326–7334 (2016) 100. Siebel, A., Gorlin, Y., Durst, J., Proux, O., Hasché, F., Tromp, M., Gasteiger, H.A.: Surface adsorption affects the performance of alkaline anion-exchange membrane fuel cells. ACS Catal. 6, 7326–7334 (2016) 101. Achilli, E., Minguzzi, A., Visibile, A., Locatelli, C., Vertova, A., Naldoni, A., Rondinini, S., Auricchio, F., Marconi, S., Fracchia, M., Ghigna, P.: 3D-printed photo-spectro-electrochemical devices for in situ and in operando X-ray absorption spectroscopy investigation. J. Synchrotron Radiat. 23, 622–628 (2016)

597 102. Binninger, T., Fabbri, E., Patru, A., Garganourakis, M., Han, J., Abbott, D.F., Sereda, O., Koetz, R., Menzel, A., Nachtegaal, M., Schmidt, T.J.: Electrochemical flow-cell setup for in situ X-ray investigations I. Cell for SAXS and XAS at synchrotron facilities. J. Electrochem. Soc. 163, H906–H912 (2016) 103. Baran, T., Wojtyla, S., Lenardi, C., Vertova, A., Ghigna, P., Achilli, E., Fracchia, M., Rondanini, S., Minguzzi, A.: An efficient CuxO photocathode for hydrogen production at neutral pH: new insights from combined spectroscopy and electrochemistry. Appl. Mater. Interfaces. 8, 21250–21260 (2016) 104. Fabbri, E., Nachtegaal, M., Binninger, T., Cheng, X., Kim, B.-J., Durst, J., Bozza, F., Graule, T., Schäublin, R., Wiles, L., Pertoso, M., Danilovic, N., Ayers, K.E., Schmidt, T.J.: Dynamic surface self-reconstruction is the key of highly active perovskite nanoelectrocatalysts for water splitting. Nat. Mater. 16, 925–932 (2017) 105. Kim, B.J., Abott, D.F., Cheng, X., Fabbri, E., Nachtegaal, M., Bozza, F., Castelli, I.E., Lebedev, D., Schaublin, R., Coperet, C., Gaule, T., Marzari, N., Schmidt, T.J.: Unravelling thermodynamics, stability, and oxygen evolution activity of strontium ruthenium perovskite oxide. ACS Catal. 7, 3245–3256 (2017) 106. Zitolo, A., Ranjbar-Sahraie, N., Mineva, T., Li, J., Jia, Q., Stamatin, S., Harrington, G.F., Lyth, S.M., Krtil, P., Mukerjee, S., Fonda, E., Jaouen, F.: Identification of catalytic sites in cobalt-nitrogen carbon materials for the oxygen reduction reaction. Nat. Commun. 8, 957 (2017) 107. Povia, M., Herranz, J., Binninger, T., Nachtegaal, M., Diaz, A., Kohlbrecher, J., Abbott, D.F., Kim, B.-J., Schmidt, T.J.: Combining SAXS and XAS to study the operando degradation of carbonsupported Pt-nanoparticle fuel cell catalysts. ACS Catal. 8, 7000–7015 (2018) 108. Miller, T.E., Davies, V., Li, J., Ghoshal, S., Stavitski, E., Attenkofer, K., Mukerjee, S., Jia, Q.: Actualizing in situ X-ray absorption spectroscopy characterization of PEMFC-cycled Pt-electrodes. J. Electrochem. Soc. 165, F597–F603 (2018) 109. Jeon, H.S., Sinev, I., Scholten, F., Divins, N.J., Zegkinoglou, I., Pielsticker, L., Roldan Cuenya, B.: Operando evolution of the structure and oxidation state of size controlled Zn nanoparticles during CO2 electroreduction. J. Am. Chem. Soc. 140, 9383–9386 (2018) 110. Enman, L.J., Burke Stevens, M., Dahan, M.H., Nellist, M.R., Toroker, M.C., Boettcher, S.W.: Operando X-ray absorption spectroscopy shows iron oxidation is concurrent with oxygen evolution in cobalt–iron (oxy)hydroxide electrocatalysts. Angew. Chem. Int. Ed. 57, 12840–12844 (2018) 111. Müller, R.J., Lan, J., Lienau, K., Moré, R., Triana, C.A., Iannuzzi, M., Patzke, G.R.: Monitoring surface transformations of metal carbodiimide water oxidation catalysts by operando XAS and Raman spectroscopy. Dalton Trans. 47, 10759–10766 (2018) 112. Fracchia, M., Visibile, A., Ahlberg, E., Vertova, A., Minguzzi, A., Ghigna, P., Rondinini, S.: α- and γ-FeOOH: stability, reversibility, and nature of the active phase under hydrogen evolution. ACS Appl. Energy Mater. 1, 1716–1725 (2018) 113. Cavillo, L., Carraro, F., Vozniuk, O., Celorrio, V., Nodari, L., Russell, A.E., Debellis, D., Fermin, D., Cavani, F., Agnoli, S., Yang, G.H.B., Hung, S.-F., Liu, S., Yuan, K., Miao, S., Zhang, L., Huang, X., Wang, H.-Y., Cai, W., Chen, R., Gao, J., Yang, X., Chen, W., Huang, Y., Chen, H.M., Li, C.M., Zhang, T., Liu, B.: Atomically dispersed Ni(I) as the active site for electrochemical CO2 reduction. Nat. Energy. 3, 140–147 (2018) 114. Yang, H.B., Hung, S.-F., Liu, S., Yuan, K., Miao, S., Zhang, L., Huang, X., Wang, H.-Y., Cai, W., Chen, R., Gao, J., Yang, X., Chen, W., Huang, Y., Chen, H.M., Li, C.M., Zhang, T., Liu, B.: Atomically dispersed Ni(I) as the active site for electrochemical CO2 reduction. Nat. Energy. 3, 140–147 (2018) 115. Song, S., Zhou, J., Su, X., Wang, Y., Li, J., Zhang, L., Xiao, G., Guan, C., Liu, R., Chen, S., Lin, H.-J., Zhang, S., Wang, J.-Q.:

27

598 Operando X-ray spectroscopic tracking of self-reconstruction for anchored nanoparticles as high-performance electrocatalysts towards oxygen evolution. Energy Environ. Sci. 11, 2945–2953 (2018) 116. Handoko, A.D., Wei, F., Jenndy, Yeo, B.S., Seh, Z.W.: Understanding heterogeneous electrocatalytic carbon dioxide reduction through operando techniques. Nat. Catal. 1, 922–924 (2018) 117. Weng, Z., Wu, Y., Wang, M., Jiang, J., Yang, K., Huo, S., Wang, X.-F., Ma, Q., Brudvig, G.W., Batista, V.S., Liang, Y., Feng, Z., Wang, H.: Active sites of copper-complex catalytic materials for electrochemical carbon dioxide reduction. Nat. Commun. 9, 415 (2018) 118. Kornienko, N., Heidary, N., Cibin, G., Reisner, E.: Catalysis by design: development of a bifunctional water splitting catalyst through an operando measurement directed optimization cycle. Chem. Sci. 9, 5322–5333 (2018) 119. Yang, X., Nash, J., Anibal, J., Dunwell, M., Kattel, S., Stavitski, E., Attenkofer, K., Chen, J.G., Yan, Y., Xu, B.: Mechanistic insights into electrochemical nitrogen reduction reaction on vanadium nitride nanoparticles. J. Am. Chem. Soc. 140, 13387–13391 (2018) 120. Fracchia, M., Cristino, V., Vertova, A., Rondininic, S., Caramori, S., Ghigna, P., Minguzzi, A.: Operando X-ray absorption spectroscopy of WO3 photoanodes. Electrochim. Acta. 320, 134561 (2019) 121. Kim, B.-J., Fabbri, E., Abbott, D.F., Cheng, X., Clark, A.H., Nachtegaal, M., Borlaf, M., Castelli, I.E., Graule, T., Schmidt, T.J.: Functional role of Fe-doping in Co-based perovskite oxide catalysts for oxygen evolution reaction. J. Am. Chem. Soc. 141, 5231–5240 (2019) 122. Al Samarai, M., Hahn, A.W., Askari, A.B., Cui, Y.-T., Yamazoe, K., Miyawaki, J., Harada, Y., Rüdiger, O., DeBeer, S.: Elucidation of structureactivity correlations in a nickel manganese oxide oxygen evolution reaction catalyst by operando Ni L-edge X-ray absorption spectroscopy and 2p3d resonant inelastic X-ray scattering. ACS Appl. Mater. Interfaces. 11, 38595–38605 (2019) 123. Cao, L., Luo, Q., Liu, W., Lin, Y., Liu, X., Cao, Y., Zhang, W., Wu, Y., Yang, J., Yao, T., Wei, S.: Identification of single-atom active sites in carbon-based cobalt catalysts during electrocatalytic hydrogen evolution. Nat. Catal. 2, 134–141 (2019) 124. Zhu, Y., Chen, H.-C., Hsu, C.-S., Lin, T.-S., Chang, C.-J., Chang, S.-C., Tsai, L.-D., Ming Chen, H.: Operando unraveling of the structural and chemical stability of P-substituted CoSe2 electrocatalysts toward hydrogen and oxygen evolution reactions in alkaline electrolyte. ACS Energy Lett. 4, 987–994 (2019) 125. Fracchia, M., Ghigna, P., Pozzi, T., Tamburini, U.A., Colombo, V., Braglia, L.: Stabilization by configurational entropy of the Cu (II) active site during CO oxidation on Mg0.2CO0.2,Ni0.2Cu0.2Zn0.2. J. Phys. Chem. Lett. 11, 3589–3593 (2020) 126. Braglia, L., Fracchia, M., Ghigna, P., Minguzzi, A., Meroni, D., Edla, R., Vandichel, M., Ahlberg, E., Cerrato, G., Torelli, P.: Understanding solid gas reaction mechanisms by operando soft X-ray absorption spectroscopy at ambient pressure. J. Phys. Chem. C. 124, 14202–14212 (2020) 127. Varsha, M.V., Nageswaran, G.: Operando X-ray techniques: a focus on hydrogen and oxygen evolution reactions. Front. Chem. 8, 23 (2020) 128. Carbonio, E.A., Velasco-Velez, J.J., Schlögl, R., Knop-Gericke, A.: Perspective-outlook on operando photoelectron and absorption spectroscopy to probe catalysts at the solid-liquid electrochemical interface. J. Electrochem. Soc. 167, 054509 (2020) 129. Qi, J., Lin, Y.P., Chen, D.D., Zhou, T.H., Zhang, W., Cao, R.: Autologous cobalt phosphates with modulated coordination sites for electrocatalytic water oxidation. Angew. Chem. Int. Ed. 259, 8917–8921 (2020)

M. A. Newton et al. 130. Suntivich, J., May, K.J., Gasteiger, H.A., Goodenough, J.B., Horn, Y.: A perovskite oxide optimized for oxygen evolution catalysis from molecular orbital principles. Science. 334, 1383–1385 (2011) 131. Hedenstedt, K., Gomes, A.S.O., Busch, M., Ahlberg, E.: Study of hypochlorite reduction related to the sodium chlorate process. Electrocatalysis. 7, 326–335 (2016) 132. Stratmann, M., Hoffmann, K.: In situ Mossbauer spectroscopic study of reactions within rust layers. Corros. Sci. 29, 1329–1352 (1989) 133. Monnier, J., Réguer, S., Foy, E., Testemale, D., Mirambet, F., Saheb, M., Dillmann, P., Guillot, I.: XAS and XRD in situ characterisation of reduction and re-oxidation processes of iron corrosion products involved in atmospheric corrosion. Corros. Sci. 78, 293–303 (2014) 134. Hedenstedt, K., Bäckström, J., Ahlberg, E.: In-situ Raman spectroscopy of α- and γ-FeOOH during cathodic load. J. Electrochem. Soc. 164, H621–H627 (2017) 135. Niwa, H., Kiuchi, H., Miyawaki, I., Harada, Y., Oshima, M., Nabae, Y., Aoki, T.: Operando soft X-ray emission spectroscopy of iron phthalocyanine-based oxygen reduction catalysts. Electrochem. Commun. 35, 57–60 (2013) 136. Nagasaka, M., Yuzawa, H., Horigome, T., Kosugi, N.: In operando observation system for electrochemical reaction by soft X-ray absorption spectroscopy with potential modulation method. Rev. Sci. Instrum. 85, 104105 (2014) 137. Nagasaka, M., Yuzawa, H.: Development and application of in situ/ operando soft X-ray transmission cells to aqueous solutions and catalytic electrochemical reactions. J. Elecron. Spectrosc. Relat. Phenom. 200, 293–310 (2015) 138. Toyoshima, R., Kondoh, H.: In-situ observations of catalytic surface reactions with soft X-rays under working conditions. J. Phys. C. 27, 083003 (2015) 139. Nguyen, L., Tao, F.: Development of a reaction cell for in-situ/ operando studies of a catalyst under a reaction condition and during catalysis. Rev. Sci. Instrum. 87, 064101 (2016) 140. Schwanke, C., Xi, L., Lange, M.: A soft XAS transmission cell for operando studies. J. Synchrotron Radiat. 23, 1390–1394 (2016) 141. Castan-Guerrero, C., Krizmancic, D., Bonanni, V., Edla, R., Deluisa, A., Salvador, F., Rossi, G., Panaccione, G., Torelli, P.: A reaction cell for ambient pressure soft x-ray absorption spectroscopy. Rev. Sci. Instrum. 89, 054191 (2018) 142. Nakamura, T., Oike, R., Kimura, Y., Tamenori, Y., Kawada, Y., Amezawa, K.: Operando soft X-ray absorption spectroscopic study on a solid oxide fuel cell cathode during electrochemical oxygen reduction. ChemSusChem. 10, 2008–2014 (2017) 143. Chang, S.-Y., Kathyola, A., Willneff, E.A., Willis, C.J., Wilson, P., Dowding, P.J., Cibin, G., Kroner, A.B., Shotton, E.J., Schroeder, S.L.M.: A versatile liquid-jet/sessile droplet system for operando studies of reactions in liquid dispersions and solutions by X-ray absorption spectroscopy. React. Chem. Eng. 4, 679–687 (2019) 144. Liu, Y.-S., Jeong, S., White, J.L., Feng, X., Cho, E.S., Stavila, V., Allendorf, M.D., Urban, J.J., Guo, J.: In-situ/operando X-ray characterization of metal hydrides. ChemPhysChem. 20, 1261–1271 (2019) 145. Masanari, N., Havato, Y., Nobuhiro, K.: Soft X-ray absorption spectroscopy of liquids for understanding chemical processes in solution. Anal. Sci. 36, 95–105 (2020) 146. Ye, Y.F., Wu, C.H., Zhang, L., Liu, Y.S., Glans-Suzuki, P.A., Guo, J.H.: Using soft X-ray spectroscopy to characterize electro/electrolyte interfaces in-situ and in-operando. J. Electron Spectrosc. Relat. Phenom. 221, 2–9 (2017) 147. See, for instance, https://henke.lbl.gov/optical_constants/filter2. html

27

X-Ray Absorption Spectroscopy (XAS): XANES and EXAFS

148. Moya-Cancino, J.G., Honkanen, A.-P., Van der Eerden, A.M.J., Schaink, H., Folkmertsma, L., Ghiasi, M., Longo, A., De Groot, F.M.F., Meifer, F., Huotari, S., Weckhuysen, B.M.: In situ X-ray absorption near edge structure spectroscopy of a solid catalyst using a laboratory setup. ChemCatChem. 11, 1039–1044 (2019) 149. Bartzsch, S., Corde, S., Crosbie, J.C., Day, L., Donzelli, M., Krisch, M., Lerch, M., Pellicioli, P., Smyth, L.M.L., Tehei, T.: Technical advances in X-ray microbeam radiation therapy. Phys. Med. Biol. 65, 02TR01 (2020) 150. Hertault, A., Manuel, B., Midulla, M., Bordier, C., Desponds, L., Kilani, M.S., Sobocinski, J., Haulon, S.: Minimizing radiation exposure during endovascular procedures: basic knowledge, literature review, and reporting standards. Eur. J. Vasc. Endovasc. Surg. 50, 21–36 (2015) 151. Das, I.J., Cheng, C.W., Watts, R.J., Ahnesjo, A., Gibbons, J., Li, X.A., Lowenstein, J., Mitra, R.K., Simon, W.E., Zhu, T.C.: Accelerator beam data commissioning equipment and procedures: report of the TG-106 of the therapy physics committee of the AAPM. Med. Phys. 35, 4186–4215 (2008) 152. Huda, W., Nickoloff, E.L., Boone, J.M.: Overview of patient dosimetry in diagnostic radiology in the USA fort the past 50 years. Med. Phys. 35, 5713–5728 (2008) 153. Garman, E.F.: Radiation damage in macromolecular crystallography: what is it and why should we care? Acta Cryst. D66, 339–351 (2010) 154. Owen, R.L., Rudino-Pinera, E., Garman, E.F.: Experimental determination of the radiation dose limit for cryocooled protein crystals. PNAS. 103, 4912–4917 (2006) 155. Holton, J.M.: A beginner’s guide to radiation damage. J. Synchrotron Radiat. 16, 133–142 (2009) 156. Paithankar, K.S., Owen, R.L., Garman, E.F.: Absorbed dose calculations for macromolecular crystals: improvements to RADDOSE. J. Synchrotron Radiat. 16, 152–162 (2009) 157. Meents, A., Gutmann, S., Wagner, A., Schulze-Briese, C.: Origin and temperature dependence of radiation damage in biological samples at cryogenic temperatures. PNAS. 107, 1094–1099 (2010) 158. De la Mora, E., Lovett, J.E., Blanford, C.F., Garman, E.F., Valderrama, B., Rudino-Pinera, E.: Structural changes caused by radiation-induced reduction and radiolysis: the effect of X-ray absorbed dose in a fungal multicopper oxidase. Acta Crystallogr. D Biol. Crystallogr. D68, 564–577 (2012) 159. Serrano-Posada, H., Centeno-Leija, S., Rojas-Trjo, S.P., Rodriguez-Almazan, C., Stojanoff, V., Rudino-Pinera, E.: X-rayinduced catalytic active-site reduction of a multicopper oxidase: structural insights into the proton-relay mechanism and O-2-reduction states. Acta Cryst. D71, 2396–2411 (2015) 160. Horrell, S., Antonyuk, S.V., Eady, R.R., Hasnain, S.S., Hough, M.A., Strange, R.W.: Serial crystallography captures enzyme catalysis in copper nitrite reductase at atomic resolution from one crystal. IUCrJ. 3, 271–281 (2016) 161. Grundahl Frankaer, C., Mossin, S., Stahl, K., Harris, P.: Towards accurate structural characterization of metal centres in protein crystals: the structures of Ni and Cu T6bovine insulin derivatives. Acta Cryst. D70, 110–122 (2014) 162. George, G.N., Pickering, I.J., Pushie, M.J., Nienaber, K., Hackett, M.J., Ascone, I., Hedman, B., Hodgson, K.O., Aitken, J.B., Levina, A., Glover, C., Lay, P.A.: X-ray-induced photo-chemistry and X-ray absorption spectroscopy of biological samples. J. Synchrotron Radiat. 19, 875–886 (2012)

599 163. Butenko, Y.V., Alves, L., Brieva, A.C., Yang, J., Krishnamurthy, S., Siller, L.: X-ray induced decomposition of gold nitride. Chem. Phys. Lett. 430, 89–92 (2006) 164. Mesu, G.J., Beale, A.M., de Groot, F.M.F., Weckhuysen, B.M.: Probing the influence of X-rays on aqueous copper solutions using time-resolved in situ combined video/X-ray absorption near-edge/ ultraviolet-visible spectroscopy. J. Phys. Chem. B. 110, 17671–17677 (2006) 165. Plech, A., Kotaidis, V., Siems, A., Sztucki, M.: Kinetics of the X-ray induced gold nanoparticle synthesis. Phys. Chem. Chem. Phys. 10, 3888–3894 (2008) 166. Chang, S.H., Kim, J., Phatak, C., D’Aquila, K., Kim, S.K., Kim, J., Song, S.J., Hwang, C.S., Eastman, J.A., Freeland, J.W.: X-ray irradiation induced reversible resistance change in Pt/TiO2/Pt cells. ACS Nano. 8, 1584–1589 (2014) 167. Jiang, P., Porsgaard, S., Borondics, F., Kober, M., Caballero, A., Bluhm, H., Besenbacher, F., Salmeron, M.: Room-temperature reaction of oxygen with gold: an in situ ambient-pressure X-ray photoelectron spectroscopy investigation. J. Am. Chem. Soc. 132, 2858–2859 (2010) 168. Martis, V., Nikitenko, S., Sen, S., Sankar, G., van Beek, W., Flinichuk, Y., Snigireva, I., Bras, W.: Effects of X-rays on crystal nucleation in lithium disilicate. Cryst. Growth Des. 11, 2858–2865 (2011) 169. Stanley, H.B., Banerjee, D., van Breeman, L., Ciston, J., Liebscher, C.H., Martis, V., Merino, D.H., Longo, A., Pattison, P., Peters, G.W.M.: X-ray irradiation induced reduction and nanoclustering of lead in borosilicate glass. CrystEngComm. 16, 9331–9339 (2014) 170. Feldman, V.I., Zezin, A.A., Abramchuk, S.S., Zezina, E.A.: X-ray induced formation of metal nanoparticles from interpolyelectrolyte complexes with copper and silver ions: the radiation-chemical contrast. J. Phys. Chem. C. 117, 7286–7293 (2013) 171. Newton, M.A., Knorpp, A.J., Meyet, J., Stoian, D., Nachtegaal, M., Clark, A.H., Safonova, O.V., Emerich, H., van Beek, W., Sushkevich, V.L., van Bokhoven, J.A.: Unwanted effects of X-rays in surface grafted copper (II) organometallics and copper exchanged zeolites, how they manifest, and what can be done about them? Phys. Chem. Chem. Phys. 22, 6826–6837 (2020) 172. Vanelderen, P., Snyder, B.E.R., Tsai, M.-L., Hadt, R.G., Vancauwenbergh, J., Coussens, O., Schoonhedyt, R.A., Sels, B.F., Solomon, E.I.: Spectroscopic definition of the copper active sites in mordenite: selective methane oxidation. J. Am. Chem. Soc. 137, 6383–6392 (2015) 173. Clausen, B.S., Steffensen, G., Fabius, B., Villadsen, L., Feidenhansl, R., Topsoe, H.: In-situ cell for combined XRD and online catalyst tests: studies of Cu based water-gas shift and methanol catalysts. J. Catal. 132, 524–535 (1991) 174. Chupas, P.J., Chapman, K.W., Kurtz, C., Hanson, J.C., Lee, P.L., Grey, C.P.: A versatile sample environment cell for non ambient X-ray scattering measurements. J. Appl. Crystallogr. 41, 822–824 (2008) 175. Grunwaldt, J.D., Caravati, M., Hannemann, S., Baiker, A.: X-ray absorption spectroscopy under reaction conditions: suitability of different reaction cells for combined catalyst characterization and time-resolved study. Phys. Chem. Chem. Phys. 6, 3037–3047 (2004)

27

600

Mark Newton is a graduate of the University of Liverpool, UK, where he also obtained his PhD in surface science and catalysis, under the supervision of Prof. M. Bowker. He has worked at the University of Washington (Prof. C. T. Campbell), Seattle, USA, the University of Southampton, UK, (Profs. J. Evans and B. E. Hayden), and the European Synchrotron Radiation Facility (ESRF, beamline scientist ID24/BM23), Grenoble, France. In 2017 he moved to the department of Chemistry and applied biosciences, ETH, Zurich, under the auspices of Profs. J. A. van Bokhoven and C. Coperet before, in late 2022, moving to EPFL and the laboratory (LNCE) of Prof. R. Buonsanti.

Dr. Patric Zimmermann, born 15th of February 1978 in Germany, obtained his BSc and Msc in X-ray physics in the group of Prof. Birgit Kanngiesser at the TU Berlin in 2013. After 1 year in the group of Prof. Majed Chergui at the EPFL in Switzerland, he moved to Utrecht in the Netherlands and obtained his PhD in the group of Prof. Frank de Groot in 2018. After that, he joined the Laboratory for Catalysis and Sustainable Chemistry (LSK) at the Paul Scherrer Institute in Switzerland where he is responsible for the group’s laboratory XAS/XES spectrometer.

M. A. Newton et al.

Jeroen A. van Bokhoven completed a degree in chemistry at Utrecht University (Netherlands) in 1995 and went on to obtain a PhD in inorganic chemistry and catalysis from the same university in 2000 (with honors). From 1999 to 2002, he was head of the XAS (X-ray absorption spectroscopy) users support group at Utrecht University. In 2002, he moved to the ETH, where he worked as a postdoc in the group of Professor Prins. In 2006, he obtained an SNF assistant professorship in the Department of Chemistry and Applied Biology. He was the 2008 recipient of the Swiss Chemical Society Werner Prize. Since 2010, Jeroen A. van Bokhoven has a chair in heterogeneous catalysis at the Institute for Chemical and Bioengineering at ETH Zurich and is head of Laboratory for Catalysis and Sustainable Chemistry at Paul Scherrer Institute.

Time-Resolved X-Ray Absorption Spectroscopy (XAS) , Caterina Suzanna Wondergem

Bert M. Weckhuysen Charlotte Vogt

28

, and

Contents 28.1 28.1.1 28.1.2

Basic Concepts of X-Ray Spectroscopy . . . . . . . . . . . . . . . . 602 Interaction of X-Rays with Matter . . . . . . . . . . . . . . . . . . . . . . . . 602 The EXAFS Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 604

28.2 28.2.1 28.2.2 28.2.3 28.2.4

The X-Ray Absorption Experiment . . . . . . . . . . . . . . . . . . . . The X-Ray Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spectral Versus Time Resolution . . . . . . . . . . . . . . . . . . . . . . . . . Spatially Resolved X-Ray Absorption Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Choosing the Correct Experimental Mode . . . . . . . . . . . . . . .

28.3

X-Ray Absorption Spectroscopy in Catalysis . . . . . . . . . 613

28.4 28.4.1 28.4.2 28.4.3 28.4.4

Showcases from the Field of Heterogeneous Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Automotive Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hydrogenation Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28.5

Toward Ultrafast X-Ray Spectroscopy of Catalysts . . . 618

28.6

Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 620

605 605 608 608 610

614 614 616 618 618

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621

Abstract

X-ray absorption spectroscopy (XAS) plays a crucial role in the characterization of catalysts as it can in principle B. M. Weckhuysen (*) Inorganic Chemistry and Catalysis, Debye Institute for Nanomaterials Science, Utrecht University, CG Utrecht, The Netherlands e-mail: [email protected] C. S. Wondergem Research Center for Materials Science, Graduate School of Science, Nagoya University, Nagoya, Japan e-mail: [email protected]

characterize any chemical element under well-defined conditions (i.e., in the liquid, gas, and solid phase) as well as under reaction conditions (i.e., at elevated temperatures and pressures, in the so-called in situ or operando mode). The past decades have seen an increase in XAS capabilities in part due to the higher brilliance of X-ray sources at synchrotrons, which in combination with more powerful detectors and optics allows to conduct timeresolved measurements, such as sub-second X-ray absorption near-edge spectroscopy (XANES). Such measurements allow investigation of solid catalysts at different stages of existence, i.e., their birth, life, and death. Furthermore, by the combination of XANES with microscopic capabilities by use of Fresnel zone plates to focus, X-rays allow for 2D and 3D imaging of a catalyst material as a function of reaction time, while the further development of lab-based X-ray source has made it possible to bring the X-ray experiment into both academic and industrial labs, comparable to what we currently do for measuring, e.g., X-ray diffraction (XRD) and Raman spectroscopy. Finally, as no single analytical technique can offer the ultimate answer to a scientific question, often (time-resolved) XAS is combined with optical, diffraction, and/or scattering methods, thereby allowing to distinguish between local and structural (bulk) properties of catalyst materials. The above-described developments will be illustrated by using a selection of showcases. The chapter concludes with some general observations as well as with an outlook. Keywords

X-ray absorption spectroscopy · Operando spectroscopy · Synchrotron · Core electron spectroscopy · X-rays with matter · Time-resolved X-ray absorption spectroscopy

C. Vogt Schulich Faculty of Chemistry, Technion, Israel Institute of Technology, Haifa, Israel e-mail: [email protected] © Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_28

601

602

28.1

B. M. Weckhuysen et al.

Basic Concepts of X-Ray Spectroscopy

28.1.1 Interaction of X-Rays with Matter Electromagnetic radiation interacts with matter in ways that are characteristic for the type of matter, depending on the energy of the light. X-rays are generally defined as photons with energies between 0.1 keV and 200 keV (approximately 6 pm–12 nm, Fig. 28.1a). Such photons have the ability to interact with electrons that surround the cores of atoms, namely, core electrons [1–5]. Unlike valence electrons, which make up the outer shells of an atom, core electrons are tightly bound to the nucleus and do not participate in chemical bonding. Because of their tight binding to the nucleus, the energy required to excite them to an unoccupied orbital is relatively high. The closer the electrons are located to the positively charged core, the higher this energy requirement becomes [6]. The core electron shells are termed K, L, M, and N, depending on their vicinity to the core (Fig. 28.1b). Their nomenclature is based on quantum numbers, i.e., 1s corresponds to K, 2s and 2p to L, and so on. This means that within one element, electrons in the K shell require the highest amount of energy for excitation as this corresponds to the excitation of a 1s electron. The energy at which the excitation of a 1s electron occurs is called the “Kedge” (>1 keV for transition metals). The required energy for excitation decreases moving to the L- and M-edges (generally ~0.1–1 keV for (transition) metals) [7]. This is illustrated in Fig. 28.1c, where the absorption edges of the K, L, and M shells and their corresponding orbitals are shown as a function of energy. X-ray absorption spectroscopy (XAS) is used to measure these edge energies and to analyze the properties of the core electrons of an element under investigation. XAS is a bulk technique, and as such the spectrum that results from the acquired XAS data is that of the average atom in the sample. To obtain an X-ray absorption spectrum, the sample under study is exposed to an X-ray source, which scans along a range of relevant energies close to a known absorption edge of the element under study [6, 7]. An intensity jump in the spectrum can be observed at a specific energy, which relates to the absorption of photons by the sample as a function of energy. Three distinct energy-dependent events are of importance while scanning X-rays across an absorption edge. These three events give rise to distinct features in X-ray absorption spectra [2, 5]: 1. The pre-edge, the threshold of absorption which is distinct for excitations into the lowest unoccupied states. 2. The edge, where core transitions to quasi-bound states occur. 3. Scattering of electrons with neighboring atoms. In the high-kinetic-energy range, the constructive and

destructive interference of the resonance of excited electrons at fixed positions of neighboring atoms gives rise to clearly distinguishable absorption features (oscillations). A typical example of an X-ray absorption spectrum featuring the abovementioned events is shown in Fig. 28.1d, including the division of the spectrum in two separate regions. These two regions are referred to as the X-ray absorption near-edge structure (XANES), also known as near-edge X-ray absorption fine structure (NEXAFS), and the extended X-ray absorption fine structure (EXAFS). In XANES, events 1 and 2 can be observed. This technique focuses on the absorption edge, approximately 200 eV before the edge and approximately 1000 eV after it for K-edge spectra of transition metals. The scattering of electrons with neighboring atoms, event 3, is what is considered in EXAFS, which is found in the higher-kinetic-energy region of the X-ray absorption spectra [5–7]. The areas of a spectrum corresponding to these separate events hold different information on the material under study. First, the energy of the measured edge gives information on the characteristic absorption edge of the material and thus yields element-specific information [8]. For the K-edge, this involves excitation from the 1s ➔ p shells. Since there is only one electron in an s-orbital, there is only one (main) K-edge. There are however three L-edges, since these involve both an s and two p-orbitals. The L1-edge corresponds to s ➔ p transitions, while the L2- and L3-edges correspond to p ➔ d transitions. The former is excited at higher energy than the latter two, which generally occur closer to each other in energy (Fig. 28.1c), and can even overlap or be difficult to distinguish in a spectrum. This also complicates the scattering spectrum in the high-kinetic-energy range, requiring different modes of measurement. These different edges not only give information on which element is being studied, but also provide information on the average oxidation state. When observing the profile of a K-edge spectrum in more detail, often a small peak can be observed in the lowerenergy region of the spectrum, right before the intense increase in absorption. This is the pre-edge (point 1 described above), corresponding to the excitation of a 1s electron to an unoccupied d-orbital (this is a dipoleforbidden, quadrupole-allowed transition and therefore a much weaker feature than the electric dipole allowed 1s to p-orbital transitions). This pre-edge can already give information on the oxidation state of the average atom by revealing the occupancy of the d-orbitals [2, 9]. This includes the observation of ligand (geometry) and local distortions, e.g., Chen et al. reported strong geometry-dependent 3d molecular orbital energies, resulting in a broad 1s ➔ 3dx2-y2 transition and elongated M-R bonds in their studies of

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

603

a) 10–12 10–11 10–10

106

105

γ-rays

b)

10–9

104

Hard X-rays

103

10–8

10–7

10–6

101

100

102

Soft X-rays

UV Vis

10–5

10–1

10–4

10–3

10–2

10–3

10–2

10–4

μ-Waves

IR

100 l (m)

10–1

10–5

10–6 E (eV)

Radio-waves

c)

hn

K

28

L1 Absorption

L2

K LMN

M M5 3 M1 M4 M2

L3

3d 3p 3s

2p 2p

1s

2s r

Energy (eV)

d)

e)

Absorption

Continuum

EXAFS

XANES Edge 1s to continuum Rising edge 1s to (n+1)p Pre-edge 1s to nd

M L

m0 K Energy (eV)

f)

g) Continuum

Continuum M

M

L

L

K

K

Fig. 28.1 (a) Wavelengths in m (upper number) and corresponding energies in eV (lower numbers) of electromagnetic radiation or light. The electromagnetic radiation can be divided into subgroups as indicated in the figure. X-rays constitute the range of 0.1–200 keV and in turn can be divided into hard and soft X-rays, where the first has higher energy than the latter; (b) schematic representation of core electron excitation from the K, L, and M shell by incoming light with sufficient energy (i.e., X-rays); (c) schematic representation of absorption edges. The highest energy is required to excite a core electron from the 1s or K shell, followed by the L (2s, 2p) and M (3s, 3p, 3d) shells; (d) typical example of an X-ray absorption spectrum, depicting the X-ray absorption

near-edge structure (XANES) and the extended X-ray absorption fine structure (EXAFS); (e) schematic representation of X-ray absorption, where an incident X-ray photon excites a core electron, ejecting it from the atom (into the continuum). The excited atom can fill the core hole through several mechanisms to return to the ground state; (f) schematic representation of X-ray fluorescence. Electrons from higher shells (i.e., for K shell core holes, electrons from the L and M shells) fill the core hole, releasing X-ray photons with characteristic energies depending on the element and shell; (g) schematic representation of Auger electron emission. The core hole is filled by an electron from a higher shell, while emitting an electron of the same shell into the continuum

axial ligation of Ni metalloporphyrins [10, 11]. As mentioned, the pre-edge and the edge together are often referred to as the XANES or NEXAFS regions. The XANES region of an X-ray absorption spectrum is generally used to

determine electronic (e.g., oxidation state) and (to a lesser extent) geometric information about the sample. Examples are estimations of ligand fields, spin states, or the charge on a metal carrier (oxidation state).

604

B. M. Weckhuysen et al.

Above the edge, at high kinetic energies, we can observe the EXAFS as post-edge oscillations [5, 12]. These oscillations originate from scattering interactions of the emitted electrons with neighboring atoms. The resulting constructive and destructive interference results in the oscillations observed in the EXAFS and thereby provides information on the local surroundings of an atom [13]. The EXAFS is generally used to determine precise local geometric information on the average absorbing atom. First, the spacing between the oscillations or their frequency is inversely proportional to the distance between the atom that absorbed the X-ray photon and a neighboring atom that the emitted electron scattered off of, allowing determination of the average local atomic surroundings. For example, if a solid being measured is in its metallic form, the distance between two atoms will be larger than when the solid is oxidized. In the latter case, the scattering atoms (oxygen) would be found at a position that is closer to the atom than a metallic neighbor would be (i.e., at bond lengths of below 2 Å, whereas metalmetal bonds are found at around or above 2 Å). Additionally, scattering probability generally increases with atomic mass and is element-dependent, meaning that we can extract information not only about the distance to neighboring atoms but also about their nature. Finally, the amplitude of the EXAFS oscillations reflects the number of scattering neighbors and thus can be used as a measure for the average coordination number of the element under study. If the element of interest is present in low concentrations, insufficient signal might be obtained following XAS measurement configuration. In this case, X-ray emission spectroscopy, XES, techniques may be employed to collect spectra with good signal-to-noise ratios [2, 6, 7]. XES is closely related and complementary to XAS. Figure 28.1e illustrates how a core hole is created upon X-ray absorption. This core hole can be filled by an electron from a higher shell through the mechanisms shown in Fig. 28.1f, g. The emitted photons or electrons can be detected, leading to X-ray fluorescence spectroscopy (XRF) and Auger electron spectroscopy, respectively.

28.1.2 The EXAFS Equation The description given so far has been purely qualitative, but it can be derived using quantum mechanics. This derivation results in what is often referred to as the EXAFS equation in which the EXAFS signal χ(k) is expressed as the sum of the different contributions. Note first that the EXAFS equation is given as a function of k, wave number, and not of E, energy, by the following conversion: k ¼ 1=ℏ√ð2me ðE E0 ÞÞ, and with that:

2

2 2D=λðkÞ 2k2 σ 2

χ ðkÞ ¼ S0 ΣðN i Þ f ðkÞ=kDi e

e

sinð2kDi þ δi ðkÞÞ,

where S02 represents the amplitude reduction factor, which is an element-dependent (and sample-specific) constant that accounts for the increased influence that the electrons experience from the positively charged nucleus after the emission of a core electron. This results in the orbitals adapting to the decrease in shielding from the core, or in other words, it results in relaxation of all the other electrons in the absorbing atom. Because there is now a slight symmetry mismatch in orbitals compared to the initial atom, the amplitude of EXAFS oscillations is reduced by a factor of S02. This term is obtained for EXAFS analysis by measuring a standard for the specific experimental or beamline setup and generally lies (or is kept) between 0.7 and 1.0. Ni in the EXAFS equation represents the degeneracy, or the number of neighbors of the absorbing atom. Upon inspection of the expression for the EXAFS absorption as a function of the wave number, χ(k), we see that if we know fi(k) and δi(k) we can determine the coordination number of the neighboring atom. fi(k) and δi(k) are both dependent on atomic number and represent the scattering amplitude and the phase shift, respectively. These factors can be calculated, given a valid approximation of the periodic structure of the sample, by software that was built to model EXAFS. Aside from the degeneracy Ni, the EXAFS equation can also be used to determine Di (the distance from the absorber to the scatterer and back, often represented a R) and σ, the measure of disorder in neighboring atoms. These factors are the quantitative descriptors to what was described in the last paragraph of the previous section. As the EXAFS equation implies, data is often represented in k-space to show dependence of momentum instead of energy. Additionally, data can be weighted (often by k2 or k3) to reveal weak signals at high kinetic energies, even though this approach generally also amplifies noise. Subsequently, the EXAFS data in the form of (k-weighted) χ(k) will often undergo a Fourier transform, resulting in the expression of χ in R-space, which is related to a radial distribution function. This can in turn be fitted using specialized EXAFS software to obtain information on bond distances. However, the trained eye can already gain preliminary understanding based on distinguishable features. Ultimately, proper analysis of XAS results cannot be performed without the measurement of appropriate standards and reference materials. This requires some level of understanding of the system under study – often obtained through standard laboratory techniques – before conducting X-ray experiments. For example, transmission electron microscopy can be used to determine nanoparticle size and dispersion, and when combined with energy-dispersive X-ray spectroscopy, the correct edges to consider can be determined; UV-Vis spectroscopy can be used to get an idea of charge

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

605

schematic overview of a synchrotron and its workings is given in Fig. 28.2. After generation of electrons in an electron gun, they are accelerated to (nearly) the speed of light in a linear accelerator. They are then transferred to the booster ring, with the aim of increasing their energy to the desired level. This level differs per synchrotron radiation facility. For example, the Super Photon Ring-8 (SPring-8) in Japan operates at 8 GeV, while the Swiss Light Source (SLS) in Switzerland operates at 2.4 GeV, and the Advanced Light Source (ALS) in the United States operates at 1.9 GeV. Finally, the electrons are moved to the storage ring, where they can be used to generate electromagnetic radiation with high photon flux [15]. The trajectory of the electrons in the storage ring is determined by bending magnets, steering them in a circle instead of a straight line. This comes with the release of energy, i.e., light, with a broad range of wavelengths and high photon flux. The generated light can be extracted from the storage ring in beamlines, which are placed at specific locations in the storage ring and as such can be used for experiments. Furthermore, beamlines can be built behind undulators and wigglers, which are different types of equipment in the storage ring that induce movement in the electrons and thus produce X-rays. Light generated at

transfer mechanisms, and probe molecule (infrared) spectroscopy can reveal the type of sites present in a sample. This knowledge should be used to determine appropriate standards. Along with possible modeling and calculation of EXAFS using dedicated software (e.g., Artemis, Larch), such appropriate standards are essential to a successful XAS experiment and its in-depth analysis. For more information on data treatment and dedicated EXAFS software, we refer the reader to a number of textbooks [2, 5–7].

28.2

The X-Ray Absorption Experiment

28.2.1 The X-Ray Source Synchrotron Radiation Sources Because a range of energies at high photon flux in the X-ray energy range is required to measure XAS data, standard laboratory equipment could not be used for a long time. Instead, a tunable X-ray source is employed, often in the form of a synchrotron or particle accelerator, which produces light in a broad range of wavelengths and with a high photon flux by the circulation of electrons in a storage ring [14]. A a)

Booster

Electrons

Beamline

Undulators

Bending magnets

X-rays

X-rays

Accelerator

b)

Slit

Monochromator

Storage ring Sample

I0

I1

Reference

I2

Incoming radiation

c)

If Fluorescence detector

Fig. 28.2 (a) Schematic illustration of a synchrotron light source. After generation, electrons are accelerated to nearly the speed of light in the accelerator and when entering the booster ring undergo an increase in their energy to the desired level. They are then transferred to the storage ring where they pass by bending magnets and insertion devices (i.e., undulators and wigglers) to generate the desired X-ray beam. Bending magnets change the trajectory of the electron bunch, while insertion devices like undulators consist of magnets with alternating poles; (b) schematic setup of an X-ray absorption experiment. After extraction from the storage ring, the X-rays pass through a slit to tune the beam

size. Subsequently, a monochromator in combination with another slit is used to select the desired energy of the X-ray beam. The X-ray beam then passes through the first ionization chamber, I0, which is a gas-filled radiation integration detector, before passing through the sample. After the sample, the beam passes through another ionization chamber, I1. If a reference sample is measured alongside the sample, it will be placed after I1, and another ionization chamber, I2, will be employed; (c) X-ray fluorescence experiment. The setup is virtually the same as described in a, but a fluorescence detector (If) is placed at an angle θ with respect to the sample and usually contains more than one detection channel

28

606

bending magnets, undulators, and wigglers all have their own photon flux (with the latter two resulting in a more brilliant beam than bending magnets) and energy range which in turn all come with their own advantages and drawbacks. Electrons travel together in small groups called bunches, and only when these bunches pass by the bending magnets, undulators, or wigglers, light is generated. The bunches are present in pulses in the GHz range, which means that for most practical purposes, the generated light can be deemed continuous [2, 7].

Lab-Based X-Ray Absorption Spectroscopy Recently, lab-based X-ray spectroscopy methods have been developed that can meet the requirements for XAS, i.e., high photon flux over a range of X-ray energies. Lab-based X-ray sources generally comprise X-ray tubes, which are vacuum tubes consisting of a cathode and an anode side, as can be seen in Fig. 28.3. Electrons are generated at the cathode by applying a voltage, and X-rays are generated at the anode. Most of the energy is converted into heat; only about ~1% is converted into X-rays and can be used as light source for a spectroscopy experiment. To obtain sufficiently high photon flux, the voltage can be increased, but due to the high amount of heat generation, the melting point of the anode (often a metal like W) is a limiting factor of X-ray tubes. Therefore, with the aim of applying XAS in the lab, studies have focused on overcoming these limitations by implementing cooling techniques, alternative anode materials, and using techniques that lower the heat dissipation. An example comprises rotation of the anode. As most of the heat is generated at the focal spot, which is typically very small, rotating the anode means a more homogeneous irradiation and thus lower average temperature generated. Alternatively, by focusing the electron beam as a line instead of a spot, heat dissipation becomes much more efficient. Such a source is called a line focus X-ray tube, LFXT, and is regarded a promising new development in the lab-based XAS community. Apart from the developments in the X-ray source area of the setup, significant advances in detection equipment also contribute to the ongoing increase of implementation of lab-based XAS. Techniques like wavelength-dispersive X-ray (WDX) detection that originated at synchrotron beamlines can be used in a lab setup as well, increasing the energy resolution. WDX detectors exploit Bragg’s law to achieve spatial dispersion of photons, and they can be used both for soft and hard X-ray experiments, although with slightly different setups. The advances in lab-based XAS methodologies have made it possible to obtain spectra in the lab that are similar to those obtained at synchrotrons. This has been shown by, e.g., Moya-Cancino et al., who demonstrate that despite a lower signal-to-noise ratio, the observed spectral features in

B. M. Weckhuysen et al.

their XAS spectra of the Co K-edge for a Co/TiO2 FischerTropsch synthesis (FTS) catalyst closely resemble spectra obtained on the same system measured at a synchr a synchrotron radiation facility, which can be seen in Fig. 28.3c [16]. For a more detailed and comprehensive overview of various lab-based setups and their development, we refer the interested reader to the excellent review article recently published by Zimmerman et al. [17] The authors expect that lab-based XAS will become more routine and eventually will be used in a similar manner as we are currently doing with lab-based X-ray diffraction (XRD).

Free Electron Lasers XAS experiments can also be carried out at facilities where free electron lasers (FELs) are available [18, 19]. In FELs, electrons are accelerated close to the speed of light and led by a number of undulators, producing light through the same physical principles as in storage rings. Through interaction of the produced radiation with the oscillation of the electrons, micro-bunches producing coherent radiation are formed. This results in intense X-ray pulses that can be used for ultrahigh time resolution XAS experiments [20]. The emission wavelength of FELs is tunable by adjusting the magnetic field strength of the undulators. The authors expect that a lot of exciting new time-resolved studies will become feasible with FELs, especially where repetition-type experiments can be conducted. The latter is possible when studying photocatalytic processes. Soft Versus Hard X-Rays X-rays are roughly divided into two types: soft (i.e., below ~5 keV) and hard X-rays (i.e., above 5 keV), although the region between 2 and 5 keV is sometimes also referred to as tender X-rays. Hence, soft X-rays have lower energy than hard X-rays and interact to a greater extent with absorbing materials they come across. On the one hand, the lower amount of energy they carry makes it easier for them to interact with light materials, while on the other hand the attenuation length in heavy materials is much lower than that for hard X-rays. These two together potentially make the collection of sufficient signal challenging in soft X-ray experiments. In that sense, it seems preferable to use hard X-rays that have a larger penetration depth, but the drawback is that these cannot detect some light materials with absorption edges at relatively low energies, like carbon. Generally speaking, hard X-rays are often applied to measure transition metals. The main reasons for this are ease of use, or increased practicality in experimental setups, and better signal. For these reasons, spectral acquisition times can currently be as low as tens of milliseconds at especially equipped beamlines (i.e., quick-XAS, QXAS). The large majority of XAS beamlines comprises hard X-ray beamlines for measurement of transition metal-based catalysts. Figure 28.4 shows two reactor designs, one that is typically

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

607

a)

b)

X-ray tube

28 Gas chromatograph Crystals Rowland circle

Sample

Slit

Product analysis

c)

Detector

e)

Normalized absorption (a.u.)

Normalized absorption (a.u.)

d)

d c

d c

b

b a

a 7680

7700

7720

7740 7760 Energy (eV)

7780

7800 7680

7700

7720

7740 7760 Energy (eV)

7780

7800

Fig. 28.3 (a) Schematic and (b, c) photographs of a lab-based X-ray absorption spectroscopy setup to perform time-resolved long-duration operando spectroscopy of solid catalysts. (d, e) Comparison between lab-based X-ray absorption spectra (d) and synchrotron-based X-ray

absorption spectra (e) [64] for cobalt reference materials measured: (a) Co, (b) CoTiO3, (c) Co3O4, and (d) CoO, thereby illustrating the overall quality of the performance of current lab-based setups. (Reproduced from [64], with permission from Wiley-VCH)

used for soft X-ray techniques called a microelectromechanical system (MEMS) reactor. MEMS reactors can be used both for liquid and solid studies (provided that the attenuation length of the components to study does not lead to total absorption). The other is a capillary reactor that is – in

different variations – often the choice for operando studies of heterogeneous catalyst powders reacting mainly with gases. It can be used, for example, in combination with hard X-ray microscopy and provides 3D information on, e.g., metals in a catalyst particle.

608

B. M. Weckhuysen et al.

a) MEMS reactor Fresnel zone plate

b) Capillary reactor Heater connection

Gas outlet

CO/H2

CxHy

Fresnel zone plate

Monocapillary Gas inlet

X-ray detector

2 mm

X-ray detector

Fig. 28.4 (a) A microelectromechanical system (MEMS) reactor, which can be used for time-resolved long-duration in situ or operando soft X-ray spectroscopy experiments, as illustrated in Fig. 28.6. (b) A capillary reactor, which can be used for time-resolved long-duration

in situ or operando hard X-ray absorption spectroscopy of solid catalysts, as illustrated in Fig. 28.7a. Both reactor designs have been used to study, for example, the Fischer-Tropsch synthesis (FTS) reaction over a Co/TiO2 catalyst system under realistic reaction conditions

28.2.2 Spectral Versus Time Resolution

etc. However, if we take these elements away, the factor that ultimately determines the maximum time resolution is the number of electrons (the bunches) in the storage ring. As mentioned earlier, light is generated in pulses in the GHz range, so for experiments with a timescale of seconds or longer, in all practicality, the X-ray light source can be deemed continuous. Still, the pulsed nature of the incoming X-ray radiation can be exploited to carry out transient experiments [2]. In the end, the time resolution of an experiment will be technically limited by the amount of electron bunches in the storage ring, the photon flux they generate (which differs per end-station/beamline type), the time required for mechanical movements of the equipment, and the readout speed of detectors and sensors. For more in-depth discussion of XAS theory, how synchrotrons work, and practical matters like how to prepare for X-ray spectroscopy experiments, we refer the reader to several excellent handbooks that are dedicated to XAS [1, 3, 4].

An important parameter for all spectroscopic experiments is resolution, which is defined as the capability of the used equipment to distinguish between different unit scales (e.g., time or length). Ideally, a measurement has both good spectral resolution (e.g., the amount of points that is measured between two wave numbers) so that one can identify and separate all important features in a spectrum, as well as a good time resolution so that you can conduct measurements quickly. Quick measurements allow the ability to maximize the amount of samples measured during your limited beamtime and/or to distinguish important transitions during a time-resolved experiment. However, spectral and time resolution are a weigh-off, as generally the first increases with increasing measurement time (either by a longer acquisition time or by increasing the amount of accumulations that make up one spectrum). Therefore, a balance needs to be found between the two, based on the aim of the experiment. Note that although advancement in equipment and methods plays an important part in optimizing the spectral resolution, time resolution is ultimately the limiting factor. With this in mind, it is important to realize that nowadays in XAS, the spectral resolution can be increased greatly without sacrificing time resolution. This is because of the implementation of energy-dispersive X-ray or QXAS equipment, both of which severely reduce the acquisition time (as well as the risk of sample damage) [21]. An example of a QXAS beamline is the X10DA at the Swiss Light Source (SLS) with a special monochromator developed by Wuppertal University, that can oscillate at several tens of Hz frequency, and is shown in Fig. 28.5. Just like spectral resolution, time resolution is largely determined by equipment, since machinery needs to move, detectors need to be read out, data needs to be saved,

28.2.3 Spatially Resolved X-Ray Absorption Spectroscopy An in-depth discussion of spatially resolved XAS methods is not within the scope of this chapter, but we feel any work on XAS in catalysis would be lacking if this possibility was not mentioned as more and more we are moving toward the spatiotemporal characterization of catalyst systems, i.e., simultaneously measuring the system as function of space and time and in essence taking snapshots of a catalysts at work. In short, there are several techniques that can combine imaging with time-resolved XAS that can be roughly divided into microscopy and computational techniques [22]. Not all spatially resolved techniques are suitable for all types of

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

609

a) Experimental hutch

Optics hutch QEXAFS monochromator Double crystal monochromator

Toroidal mirror

Front-end

White beam slit High order harmonic rejection mirror

Slit Bremsstrahlung stop

24.0 m

Super bend

Collimating mirror

15.6 m

12.9 m

11.9 m

7.7 m

Graphite filter, optional 0m

b)

QEXAFS monochromator

Beam Goniometer Torque motor Toroidal mirror

c)

2.5 Cu

50 Hz 40 Hz 30 Hz 20 Hz 10 Hz

50 Hz (10 ms) 40 Hz (12.5 ms)

2.0

Norm. absorption (a.u.)

30 Hz (16.7 ms) 1.5

20 Hz (25 ms) 10 Hz (50 ms)

1.0

0.5

0.0 8950

9000

9050

Energy (eV)

Fig. 28.5 (a) Schematic overview of the X10DA beamline at the Swiss Light Source (SLS, Switzerland), one of the most advanced quick-X-ray absorption spectroscopy (QXAS) beamlines currently in operation, and

9100

9150

8980

8990

9000

Energy (eV)

which is used for the investigation of solid catalysts, including the example described in Fig. 28.9. (b) Photograph of the QEXAFS monochromator made by Wuppertal University, which enables the acquisition

28

610

in situ and operando measurements. Especially the computational measurements of “classic” heterogeneous catalysis processes with high spatial resolution are still very challenging. However, some beamlines are dedicated to the investigation of a niche form of catalysis, like the ATMOS beamline at the ALS for the investigation of batteries with X-ray ptychography or the BL36XU at SPring-8 for polymer electrolyte fuel cells (PEFCs) for computed tomography-XAS (CT-XAS) measurements [23]. For computed tomography techniques, the field of view and pixel sizes are (still) generally quite large for operando experiments (in the order of microns). Therefore, statistically relevant sampling sizes are generally obtained. However, for X-ray microscopy techniques, the obtained resolution is often much higher, and single particle approaches can be used, even under in situ and operando conditions. It is important to stress that one has to be aware that while X-ray microscopy can yield valuable insights, before drawing conclusions, statistically relevant amounts of samples or particles need to be analyzed. Conventional operando XAS is in principle a bulk technique and therefore does not generally suffer from this drawback. Using both soft and hard X-rays, several advances in in situ (and ex situ) spatially resolved X-ray microscopy have been published over the last decade. For example, using scanning transmission X-ray microscopy and soft X-rays, van Ravenhorst et al. were able to visually capture the genesis of an active Co/TiO2 Fischer-Tropsch synthesis (FTS) catalyst particle using a MEMS reactor, shown in Fig. 28.4a, during a study lasting approximately 2 days. This study showed that during an induction period, the TiO2 pores slowly fill with long-chain hydrocarbons, which takes several tens of hours and shows intraparticle hydrocarbon heterogeneities (Fig. 28.6) [24]. Such spatiotemporal differences in a single catalyst particle had previously not been brought to light and could only be captured by performing time-resolved long-duration operando XAS experiments. Cats et al. studied the same catalyst system with transmission X-ray microscopy (TXM) using hard X-rays, making use of the reactor cell shown in Fig. 28.4b. They were able to note similar intraparticle effects for Co as shown in the corresponding 2D TXM images (Fig. 28.7a). Furthermore, time-resolved long-duration XANES in 3D could be measured for an individual catalyst particle, and the study indicates that the catalyst material is under operando conditions mainly in its Co

ä Fig. 28.5 (continued) of an X-ray absorption spectrum in the up and down movements of the monochromator with several tens of Hz frequency. (c) X-ray absorption spectra of the Cu K-edge recorded at

B. M. Weckhuysen et al.

metallic state. This X-ray microscopy technique has also shown to be particularly relevant for a fluid catalytic cracking (FCC) particle (shown in Fig. 28.7b, c) where 3D reconstructions with voxels of down to 20–30 nm showed detailed information of metal (i.e., Fe and Ni) speciation throughout the FCC particle (Meirer et al. [25]) as a function of reaction time, and more recently, a 3D reconstruction of the carbon speciation throughout the FCC catalyst particle (Vesely et al. [26]), allowing a detailed view of the flow resistance of the crude oil feedstock through the pore network, the ~50-μmsized FCC particle in the absence and presence of metal deposits (i.e., a long-term catalyst deactivation phenomenon) and carbon deposits (i.e., a short-term catalyst deactivation phenomenon).

28.2.4 Choosing the Correct Experimental Mode In order to choose the right mode of measurement (i.e., transmission or fluorescence), it is important to have a clear idea of the kind of information that should be contained in the acquired data. This includes the element of interest, more specifically its absorption edge energies, its concentration and distribution, and the physical state of the sample. Transmission experiments involve simply measuring the X-rays before and after passing through a sample (that should be uniform) in a straight line. Samples should not be too thick or too thin, with a rule of thumb of ~4 absorption lengths. Fluorescence mode can be employed either when the concentration of the element of interest is low in thick samples or high in thin samples. In fluorescence experiments not absorption but emission is detected, and the detector is placed at an angle. This angle is often chosen to be 90 of the incident beam, while the sample is placed at 45 to maximize emission yield and minimize noise from scattering. The sample angle can also be decreased to a grazing incidence-type setup. In this mode of measurement, the angle is chosen such that the penetration depth of the X-rays is limited to a few nanometers. This makes it possible to selectively investigate the surface of flat samples by detecting the X-ray fluorescence at 90 .

different frequencies, 10, 20, . . . 50 Hz [65]. (Reproduced from [65], with permission from Wiley-VCH)

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

a)

611

b)

C Co

65 h

Ti

50.5 h

280 285 290 295 Energy (eV)

43 h 770 780 790 800 810 Energy (eV)

0.3 μm

Heating spiral 450 460 470 Energy (eV)

Fresnel zone plate

Monochromatic X-rays

1.2 mm

50 h 49 h

Normalized absorption

500 × 500 μm

51.5 h

47 h 26 h 45 h 22 h 17 h 2h

Electrical connections

Back view Gas intlet

Gas outlet

Front view

0.5 h

275 X-ray detector

280

285 290 Energy (eV)

295

300

15 nm – 0.1 μm thickness

c)

0.5 h

47 h

17 h

26 h

50.5 h

65 h 1 μm

Fig. 28.6 Schematic overview of the operando microelectromechanical system (MEMS) reactor and related data acquisition method by scanning transmission X-ray microscopy (STXM) for measuring a single Co/TiO2 Fischer-Tropsch synthesis (FTS) catalyst particle as a function of both space and time. (a) Schematic representation of the MEMS reactor. Bottom-left, optical microscopy photograph of heating spiral with 0.1 μm overall thickness, with SiNx windows of 15 nm thickness containing the single Co/TiO2 catalyst particles. Bottom, data acquisition method, the focused soft X-ray beam is changed in energy by the monochromator to measure an X-ray absorption spectrum

for each pixel as a function of increasing time-on-stream [24]. (b) Timeresolved long-duration X-ray absorption spectra show the C-edge spectra of the catalyst particle, thereby capturing the organic part of an active Co/TiO2 catalyst particle [56]. (c) Time-resolved long-duration spatial maps of an active Co/TiO2 catalyst particle [56]. The different spectra represent different chemometrically resolved C-edge spectra, thereby gradually showing the different carbon species formed, including aromatic species, with increasing time-on-stream. (Reproduced with permission from [56])

28

612

a)

CoTiO3 Co

CoTiO3

CoO

Co

Cobalt-K-edge spectra

1.5 1.0

e) d) c) b)

0.5 0.0

Reduction

Fischer-Tropsch synthesis

100 80 60 40 20 0

2.5 2.0

CoO

Composition (%)

Normalized absorption (a.u.)

CoO CoTiO3 Co

a)

7640 7660 7680 7700 7720 7740 7760 7780

0

2

4

Energy (eV)

b)

6 8 10 12 14 Time (h)

X-ray nanomography Fe

Ni

Catalyst particle age Metals affecting pore accesibility

c) Fe

Ni

16 Pm High Flow Low 105

Intensity (cts)

Fig. 28.7 (a) Two time-resolved long-duration 2D X-ray operando nanotomography reconstructions of a single Fischer-Tropsch synthesis (FTS) Co/TiO2 catalyst particle [62]. The in situ X-ray absorption near-edge spectroscopy (XANES) data of the cobalt K-edge are an example of the data quality of K-edges studied with this type of technique and measured as a function of reaction time. These results indicate that cobalt mainly was found in the metallic state under FTS conditions with minor contribution of CoTiO3 [66]. (b) 3D X-ray nanotomography reconstruction of a fluid catalytic cracking (FCC) catalyst particle of approximately 50 μm in diameter, visualizing the distribution of Fe and Ni in the FCC catalyst particle as a function of its age (i.e., days). (c) 3D X-ray nanotomography reconstruction of an FCC catalyst particle, with both Fe and Ni distribution and the pore network shown on its right-hand side the pore network includes streamlines, which illustrate the connectivity between the different macropores within the catalyst particle [25]. The same system was examined with X-ray holotomography allowing the 3D imaging of different carbon deposits formed during the FCC process. The carbon K-edge spectra are shown to the right [26]. (Reproduced from [25, 62, 66], with permission from Wiley-VCH and AAAS)

B. M. Weckhuysen et al.

Surface and non-surface carbon deposits

100 0

10 5 Energy (keV) Carbon K-edge spectra

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

28.3

X-Ray Absorption Spectroscopy in Catalysis

In the previous section, we have described both qualitatively and quantitatively that synchrotron-based XAS is an extremely powerful characterization technique and capable of studying catalysts in a spatiotemporal manner. By selecting the proper incident energy, important characteristics of the material of interest can be studied, and even very small distortions in crystal structure can be detected. This not only holds for static samples but also for dynamic samples or systems when time resolution is high enough. XAS therefore enables the elucidation of how introduced species change the (average) state of the catalyst [27, 28]. This is exactly what we are interested in when we want to study catalysis. In catalysis research, important goals include the elucidation of active sites and (subsequent) structure-performance relationships [29]. For example, an in-depth understanding of reaction and deactivation mechanisms can lead to the design of new and improved catalytic processes in terms of activity, selectivity, stability, and as such can have major effects on sustainability and economic factors. Taking into account that 80% of all manufactured goods have encountered at least one catalyst in their production process [30], small improvements in such processes can be greatly advantageous both from an economical and a societal point of view. In situ and operando spectroscopy offers the possibility to investigate various aspects of such catalytic processes in great detail as it is aimed at characterization of catalysts in their active state. Catalysts may also change their structure upon heating or when subjected to increased pressures, for example, metal nanoparticles are known to be dynamic in structure particularly at the elevated temperatures and pressures relevant to industrially applied catalysis [31–34]. Therefore, ex situ analysis will never be sufficiently capable of capturing the full complexity of a catalyst at work, and in situ or operando analysis is essential to generate relevant structure-performance relationships [35, 36]. The term operando spectroscopy is used when simultaneously analyzing reaction products made by the catalyst at work (usually by gas chromatography, GC, and/or mass spectrometry (MS)). More specifically, when using operando spectroscopy, one makes sure that the catalyst of interest is actually representative of the catalytic process aimed to study in terms of activity and (side) products. For example, in situ or operando UV-Vis spectroscopy can monitor the buildup of organics in a catalyst during reaction (e.g., polyaromatics or coke) [37], while in situ or operando infrared spectroscopy can provide a fingerprint-like molecular map of a reaction [38]. Such lab-based techniques are relatively easy to apply

613

even in operando mode – assuming a suitable reactor is available – but cannot always provide a complete picture. Sometimes, to unravel relevant important details of a (catalytic) process, information on the system is required that can only be obtained by using (advanced) X-ray-based characterization methods, for example, to yield highly specific and precise information on the local electronic environment of the catalyst materials (under reaction conditions) [39]. It is important to note that the design or use of a suitable operando reactor or cell is not trivial and requires serious consideration regarding your reaction, research goal, and the type of X-rays you will use [40]. Apart from the possibility to bring your own reactor, some beamlines also offer the possibility of using theirs [39, 41, 42]. As discussed above, XAS allows for the identification of parameters like coordination number, oxidation state, and local bond distances. Aside from these, the features in the spectra containing information on the different constituents all show up at different energies, both with regard to different elements as well as the different phases a specific element may boast in a sample. That means that XAS is elementspecific (i.e., an element shows up in a XAS spectrum within a predetermined set range of energies) and that it can distinguish between different valences or coordination numbers of an element within this specific range of energies. This is especially useful for catalysis as a heterogeneous catalyst typically consists of an active material in the form of transition metal (oxide) nanoparticles29 carried on a porous, metal oxide (e.g., SiO2, Al2O3, and TiO2) or carbonaceous support (e.g., active carbon and carbon nanofiber), most of which are materials suitable for (hard) X-ray operando studies [43] (recall that soft X-rays easily interact with light materials, which include gases, products, and even the operando cell, making operando soft X-ray studies more challenging). Therefore, in principle, the entire catalyst (i.e., both active material and support) can be characterized in depth, both in ex situ and in situ/operando modes. When introduced species, such as the aforementioned gaseous reactants, bind to the surface of the catalyst, corresponding changes in the average electronic structure can be observed in the spectra, and this information can be used to identify the oxidation state of the active phase. This holds true for many relevant processes like catalyst activation (e.g., reduction), catalytic conversion, and deactivation of the catalyst material. As such, after thorough analysis of the observed changes, a clear picture of the entire catalytic reaction can be obtained [44]. The events that a catalytic reaction encompasses take place simultaneously albeit on different timescales, ranging from picoseconds for bond breaking to (typically, depending

28

614

B. M. Weckhuysen et al.

on the process) weeks-months-years for deactivation [34, 45–47]. The different timescales can all be investigated using XAS but require different approaches. For example, when trying to figure out the equilibrium state of the catalyst at work, often steady-state operation is used as the system then is most representative for industrial operation [46]. In principle, due to the equilibrium, the timescales involved are such that we do not need to sacrifice spectral resolution for time resolution. This is especially true due to the advent of QXAS. In contrast, when we want to investigate the exact moment a reactant first comes into contact with the catalyst, kinetics, or reaction intermediates, we want to look at shorter timescales which may also require different equipment or a different beamline altogether. Such alternate approaches may, for example, require introducing reactants in a pulsed fashion. It is important to realize that under such conditions, even when the reaction products are being monitored by GC/MS, the system is inherently not representative for its industrial counterpart (should there be one) even though such experiments are typically still referred to as operando. Although technically you could measure steady-state or deactivation processes with sub-second or sub-microsecond resolution, the amount of data generated and that would require analysis would be very large (in the order of terabytes). If the scope is to learn more about deactivation, obtaining such an amount of data would not increase the chances of achieving your goal. Similarly, if you want to study a process taking place at a sub-second timescale, it would not make sense to use a time resolution of seconds or more since the information will be averaged out or not even recorded at all. Hence, a careful choice of the proper beamtime setup is important to realize the goals of a particular characterization study, including the type of catalysis (e.g., photocatalysis), and both short-duration and longduration time-resolved studies have their merits to capture the transient behavior of catalysts under relevant reaction conditions.

28.4

Showcases from the Field of Heterogeneous Catalysis

28.4.1 Automotive Catalysis Zeolites are an archetypal class of heterogeneous catalysts, combining shape selectivity originating from their ordered crystalline and microporous structure with catalytic activity originating from single sites, like Brønsted acid or metal ion sites. SSZ-13, a zeolite with the chabazite (CHA) topology meaning it consists of cages connected by 6- and 8-membered rings (Fig. 28.8a), is such a material, and its copper-exchanged derivative known as Cu-SSZ-13 has received a great deal of attention over its activity in selective

catalytic reduction (SCR) of NOx with NH3 as reducing agent. NH3-SCR is currently already being applied in trucks and other vehicles that run on diesel. Cu-SSZ-13-catalyzed NH3-SCR has been investigated intensively with numerous analytical methods, including time-resolved XAS. Figure 28.8 summarizes some important aspects of this catalytic process. First, in Fig. 28.8c, we see that this process involves two operational regimes: one at low temperature (standard), for example, when the engine is just being turned on, and one at high temperature (fast) [48, 49]. The distinct regimes favor different mechanisms [50–52] and therefore different reaction products. The different sites involved in catalysis comprise the different oxidation states (+1 and +2) Cu ions can adopt and where they are located within the zeolite framework (Fig. 28.8b), which can be identified using operando XAS. The (type of) species or sites these distinct Cu ions prefer to bind or coordinate to can be identified in the EXAFS, also revealing catalyst deactivation due to binding to poisons in the gas feed [53]. All these mechanisms were investigated with operando XAS on different timescales and conditions accommodating to the aim of the experiment. The different temperature regimes for NH3-SCR operation over zeolite Cu-SSZ-13 have been observed in time-resolved operando Cu K-edge XAS experiments at a number of different beamlines [49–52, 54]. During the low-temperature regime run at steady conversion, typically below 200  C, lower NO conversion and N2 production are observed in MS measurements compared to the high-temperature regime. Inspection of the XANES spectra at this low T shows both a pre-edge peak originating from Cu(II) (1s ➔ 3d) and an edge-rising peak attributed to Cu(I) (1s ➔ 4p). Comparison with reference spectra and application of linear combination fit (LCF) reveals that these species are present in around a 1:1 ratio, albeit that Cu(I) is only present as NH3-solvated mobile species, whereas Cu(II) is present as both mobile NH3-solvated species and framework-bound species. This distinction can also be observed in the k2-weighted FT-EXAFS spectra, where part of the Cu(II) is observed as frameworkcoordinated species by an established, well-defined second shell peak in Fig. 28.8d. The other species predominantly show a first shell coordination number of 2 for Cu(I) and 4 for Cu(II) and induce perturbation in the second shell peak through their mobile nature (recall that more crystalline environments result in well-defined peaks, and vice versa). Upon moving to the high-T regime, these perturbations decrease through the formation of more framework-coordinated Cu (II) species, as is also observed in the XANES. During the reaction in the low-T regime, XANES measurements revealed that the mobile Cu(II) species are reduced to Cu(I) by ammonia, before the latter species are oxidized back to Cu(II) to complete the redox cycle [50]. Transient operando XAS experiments in the form of LCF analysis of time-resolved QEXAFS spectra were conducted to track the

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

615

b)

c) Im {FT[2X(k)]} (Å–3) | FT[2X(k)] (Å–3)

Standard 4 NO + 4 NH3 + O2

4 N2 + 6 H2O

Fast 2 NO2 + 2 NO + 4 NH3 + O2

Z-(Cu(II)OH–) Z-(Cu(II)NO3–) (Cu(II)(NH3)4)2+ (Cu(II)(H2O)6)2+ (Cu(I)(NH3)2)+

0.5 Å–3

1

2

3

0.5 Å–3

0.5

2

3

2 R (Å)

3

f)

2 R (Å)

1

e) 1.4 1.2 1.0 0.8 0.6 O2 cutoff Steady-state NO cutoff NH3 cutoff

0.4

0.0 8.96 8.97 8.98 8.99 9.00 9.01 9.02 9.03 Photon energy (keV)

3

g)

Cu(I)(NH3)2 1s 4p

SCR 400 °C SCR 300 °C SCR 250 °C SCR 150 °C

Cu(II) 1s

3d

600

80 60

400

40 200 20 0

H-SSZ-13 Cu-SSZ-13 0 4 vol% O2/N2 4 vol% N2 Cu(I)-(NH3)2

8.975 8.98

8.979

9.00 9.01 8.99 Photon energy (keV)

dynamic Cu speciation induced by the addition of NO to the gas feed. The results confirmed that in this operational regime, the reoxidation of twofold coordinated NH3-solvated Cu(I) species to form the active fourfold coordinated Cu(II)(NH3)4 species is the rate-limiting step of the reaction [52].

28

Å–3

0.2

1

Normalized absorption (a.u.)

1

4 N2 + 6 H2O

Normalized absorption (a.u.)

d)

SCR 400 °C SCR 300 °C SCR 250 °C SCR 100 °C

0.5 Å–3

NH3 (ppm)

a)

No conversion (%)

Fig. 28.8 Selective catalytic reduction (SCR) of NOx over Cu-loaded zeolite Cu-SSZ-13 catalysts. (a) Schematic of the chabazite (CHA) topology, which is the framework structure of the SSZ-13 zeolite; (b) zoom-in of the CHA framework showing the possible positions of the Cu ions in SSZ-13, namely, in the 8-membered ring (red) and the 6-membered ring (blue) [49, 51]; (c) the two typical NOx reduction reactions during SCR operation; (d) k2-weighted extended X-ray absorption fine structure (EXAFS) spectra [54]; (e) operando X-ray absorption near-edge spectroscopy (XANES) spectra of Cu-SSZ-13 during low-temperature SCR of NOx under various conditions [67]; (f) operando XANES of the Cu-SSZ-13-catalyzed SCR recorded at various reaction temperatures [54]; (g) temperature-dependent evolution of Cu species related to NO conversion and NH3 temperatureprogrammed desorption [52]. (Reproduced from [49, 51, 52, 54, 67], with permission from Wiley-VCH and Springer Nature Publications)

Im {FT[2X(k)]} (Å–3) | FT[2X(k)] (Å–3)

28

Cu(II)-(NH3)4

200

Cu(I)-Z

Cu(II)-Z

300 400 Temperature (°C)

500

At temperatures above 350  C, most Cu(II) is present as coordinated to zeolite framework, indicating a different mechanism is at play [48, 52]. In the high-temperature regime, Cu(I) species are still observed, even though Cu(II) species are postulated to be

616

the active site. The presence of stable Cu(I) species is therefore not desired, and insight into their continued presence is required in order to increase NH3-SCR activity. Cu ions dispersed over and coordinated to well-defined materials like zeolites may seem similar activity-wise, but the fact is that the chemical environment in a 6- and in an 8-membered ring leads to differences in reducibility of Cu(II) in the presence of NH3 [49]. Through fundamental alteration of the catalyst ensuring the presence of Cu ions only in 6-membered rings, operando XANES measurements revealed that the active Cu(II) species did not suffer from reduction into stable Cu(I) with subsequent loss of high-T regime SCR activity in this position. This indicates that merely Cu(II) present in 8-membered rings is sufficiently reducible to be converted into the stable Cu(I) under influence of NH3 and high T. Finally, catalyst deactivation can be caused by several factors, for instance, the presence of steam or impurities like SO2 in the feedstock [53]. The first causes deactivation of zeolite-based materials by extracting Al from the framework, with a subsequent loss of Brønsted acid sites. However, as Ye et al. showed using XANES-based STXM, even when extra-framework Al species were generated upon treatment with steam, the Cu(II) sites proved to be very stable and remained active [55]. Whereas XANES shows that its presence can cause the formation of Cu bisulfate species which is detrimental for the activity of the catalyst, EXAFS also indicates that at low T the presence of SO2 inhibits the formation of active sites. Such a combination of operando XAS studies of the different aspects of Cu-SSZ-13-catalyzed NH3-SCR showcases that time-resolved XANES and EXAFS, especially when combined with other spectroscopic techniques, can provide fundamental understanding of the active sites, reaction mechanism, and deactivation of heterogeneous catalysts [49, 54].

28.4.2 Hydrogenation Catalysis Heterogeneous catalysts often consist of metal nanoparticles supported on oxidic or carbonaceous material. The support materials can take on different crystal phases (e.g., titania can be present in the rutile, anatase, and brookite phase) or be amorphous. Porosity often varies throughout the support material, as large-scale preparation of such materials is not conducted with structure-inducing templates. Deposition of catalytic material on these supports is not always easy to control, leading to nanoparticles showing a size distribution and/or an inhomogeneous dispersion. Furthermore, metal nanoparticles are facetted, consequently supporting many kinds of sites available for interaction with reactants. The activity of certain sites can be enhanced by the addition of promotors or, oppositely, blocked to prevent deactivation or poisoning. Additionally, for many catalysts supported on

B. M. Weckhuysen et al.

metal oxide supports, strong metal support interactions (SMSI) can take place. All of this complicates the analysis of such industrially relevant materials, especially under operando conditions where the catalyst material is in a dynamic state and changes continuously. In conglomeration with (operando) characterization data from other analytical techniques and advanced data analysis, invaluable insights can be obtained on such systems through the use of time-resolved operando XAS. For instance, Vogt et al. studied Ni/SiO2 catalysts with different particle sizes prepared by an industrial partner for CO2 reduction with operando QXAS [38]. In their operando FT-IR spectra, they observed different reaction mechanisms occurring over the various Ni particle sizes, indicating structure sensitivity. They then conducted operando QXAS experiments probing the catalysts with alternating pulses of the reactant, CO2, and a reducing gas, H2. By recording spectra with a very high time resolution (100 ms), it was possible to study the dynamics of changes in nanoparticles at relevant timescales. Nevertheless, such experiments yielded enormous datasets (>70.000 spectra), and the authors set out to study very subtle changes (1–2% of absorption). To this end, the XANES spectra were analyzed by principal component analysis (PCA), which significantly reduced the noise level. Subsequent least square linear combination fitting with Ni and NiO references revealed subtle changes (1–2%) in oxidation state of the Ni catalysts as a result of the alternating pulses, as shown in Fig. 28.9. Among the different particle sizes, the subtle changes varied in accordance with FT-IR spectroscopy results that had shown the presence of different reaction mechanisms taking place depending on catalyst size. By relating these results, these/such distinct reaction pathways could be related to specific sites on the Ni nanoparticles, thus explaining the observed structure sensitivity. This study is an excellent example of high time resolution XAS and its relevance to millisecond dynamics of solid catalysts. On the other hand, as mentioned before, catalytic processes have many relevant timescales. Recently, van Ravenhorst et al. published a time-resolved long-duration (i.e., 47 h) operando hard X-ray XAS study of FischerTropsch synthesis (FTS) catalysis. A selection of the corresponding XANES data is shown in Fig. 28.10 [56]. Here the authors also used PCA and cluster analysis (CA) to reduce the 47 spectra into 5 clusters representing the largest variation in spectral changes. These five clusters happened to align chronologically, indicating a gradual change in the Co spectra during the FTS reaction which the authors linked to the formation of Co2C, a commonly proposed deactivation mechanism. Interestingly, the formation of this cobalt carbide did not correlate to a decrease in activity, which does not support a common theory that the formation of Co2C is linked to the deactivation mechanism of this catalyst material.

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

a)

617

Normalized absorption

Kinetic (100 ms) XAS study

1.0

0.5

28 0.0 8300

8400

8500 Energy (eV)

8600

8700

Principal component analysis and clustering of r70.000 spectra per experiment

Gaussian mixture model clustering, least square linear combination fitting of clusters to Ni/NiO references, and subsequent reconstruction of spectra with GMM contribution

b)

Catalyst C 1.2 r 0.5 nm

Catalyst B 1.4 r 0.4 nm

Catalyst A 1.8 r 0.8 nm

Catalyst F 2.1 r 1.1 nm

Catalyst H 6.0 r 1.9 nm

5 nm

5 nm

2 nm

5 nm

2 nm

c)

H c d 99 b F

Metallic Ni (%)

98 N2 CO2 H2

97

96

A B

95 a

C

94 0

10

20 Time-on-stream (min)

Fig. 28.9 A time-resolved short-duration operando X-ray absorption spectroscopy of a set of Ni/SiO2 methanation catalysts. (a) Spectrum recorded at 100 ms time resolution in a study where kinetic details of the CO2 methanation over Ni/SiO2 catalyst materials with increasing Ni

30

40

metal nanoparticle size were investigated [38]. (b) A selection of transmission electron microscopy (TEM) images of these catalyst. In several Ni/SiO2 catalysts with differing mean particle size (1–6 nm), were studied with operando quick X-ray absorption spectroscopy (QXAS)

618

28.4.3 Electrocatalysis Electrocatalytic processes are also a major point of interest for time-resolved XAS studies. Polymer electrolyte fuel cells (PEFCs) aimed at H2 conversion for energy production, for instance, have been studied by a number of groups. The PEFCs often consist of a Pt catalyst supported on carbon as both the anode (H2 dissociation) and the cathode (oxygen reduction reaction, ORR) catalyst. Especially the cathode side presents an interesting topic of study since the ORR over Pt catalysts suffers from slow kinetics and agglomeration of the Pt NPs that results in deactivation [57]. Due to dedicated operando PEFC setups representative for largescale application that have been built at various beamlines, these materials are uniquely suited for advanced operando XAS studies. As such, the ORR over Pt/C under working conditions has been extensively investigated. For example, a study by the Tada group used time-resolved operando XAS to reveal oxidation kinetics at various potentials relevant for PEFC operation (i.e., 0.4–1.0 V) [58]. Additionally, they compared the pure Pt/C catalysts to bimetallic Pt3Co/C catalysts. TEM and EXAFS confirmed that the latter consisted of an alloy Pt-Co core and a Pt-rich shell as the characteristics of Pt-Co bonding did not change upon oxidation, whereas the contribution of Pt-Pt decreased and a characteristic Pt-O shell was observed. Furthermore, they observed that upon alloying with Co (and Ni [59]), kinetic control of the Pt oxidation rate could be achieved. However, in a study involving accelerated degradation tests, they found that the long-term stability of Pt3Co/C as cathode catalysts was not suitable for long-term successful PEFC operation [60]. This degradation was further investigated with computed tomography XAS, also denoted as CT-XAS. By rotating the sample 161 and recording XANES spectra at all angles, the catalyst could be reconstructed. Both Pt and Co were imaged, and this revealed differences between their degradation mechanisms that seemed to be largely influenced by cracks in the carbon support [61].

28.4.4 Photocatalysis The field of photocatalysis is particularly interesting for timeresolved XAS studies with high resolutions, since photocatalytic processes occur on the sub-microsecond timescale.

B. M. Weckhuysen et al.

For example, charge transfer processes taking place on the photocatalytically active metal center can be investigated using pump-probe techniques. This approach involves excitation of the catalyst under pulsed UV-Vis light typically in the order of ns, while the storage ring of the beamline operates in single or few bunches mode, resulting in X-ray pulses in the order of ps, separated (and limited) by the bunch frequency. Such an exploitation of the electron bunches when appropriately equilibrated with the probe can result in ns time resolution. This is demonstrated by Baran et al., who studied water splitting over a model α-Fe2O3/IrOx catalyst using Ir-LIII XANES spectra, pulse-activated by 400 nm LED light in synchronization with the incoming X-rays. Comparison between illuminated and dark samples revealed that more holes were present in IrOx during irradiation, inducing a cascade of charge transfers from the Fe2O3 layer to the IrOx layer, resulting in more photo-generated holes. These holes in the IrOx layer were observed through analyzing the density of states obtained from the XANES spectra, while the catalyst was operating under a realistic potential. A change in potential showed a change in the charge transfer process, whereby minor amounts of Ir(IV) were reduced to Ir(III), while transiently (in the order of ns) maintaining Ir(IV) coordination distances. In the XANES spectra, this resulted in a shift of the spectrum to lower energies [2]. Such a study showcases that time resolutions suitable for the detection of transient species, in this case 600 ns, can be achieved on catalysts when the capabilities of the equipment are properly exploited and the experiment is designed appropriately.

28.5

Toward Ultrafast X-Ray Spectroscopy of Catalysts

To push the time resolution down even further and into the direction of ps-fs, a different X-ray source can be employed, namely, FELs. This is demonstrated by the investigation of an elementary step in many heterogeneously catalyzed reactions: the dynamics of CO adsorption and desorption on metal surfaces in the absence and presence of, for example O2. This exciting field of frontier research has been summarized in a recent article by Nilsson and co-workers [62]. A model Ru(0001) surface was employed, and CO was probed with soft X-rays to inspect the C-K-edge with XAS, and particularly the 1s ➔ π* transition [63]. As CO adsorbs

ä Fig. 28.9 (continued) by pulsing CO2 and H2, and comparing the QXAS data of the different particle size responses to those reactant pulses. The large amount of QXAS data recorded during each experiment was refined by several (multivariate) data analysis steps allowing the discrimination of several interesting particle-size-dependent phenomena of the CO2 methanation reaction. (c) Differences between the

different time-resolved metallic Ni contribution as a function of the pulse duration when switching between H2 and CO2. Inserts a and b as well as c and d show the dynamics of the Ni metal surface under different reaction environments and its dependency of the metal nanoparticle size. (Reproduced from [38], with permission from Nature Springer Publishers)

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

619

Long term (47 h) activation/deactivation XAS study

a) 1h

Normalized absorption (a.u.)

1

28

47 h

7713 7716

7730

7740

7750

7760

0 7720

7700

7740

7760

7780

7800

Energy (eV) Principal component analysis and clustering, least square linear combination fitting and Cauchy wavelet transform

e) # h time-on-stream

c)

Normalized absorption (a.u.)

Norm. abs. (a.u.)

b)

7750 7780 Energy (eV)

7700

d)

f)

1 2

7 14 21 28 35 42 Co - ref 8 15 22 29 36 43 Cl. - 1 9 16 23 30 37 44 Cl. - 2

3 10 17 24 31 38 45 Cl. - 3 4 11 18 25 32 39 46 Cl. - 4 5 12 19 26 33 40 47 Cl. - 5 6 13 20 27 34 41

Co-hcp

Co-fcc CO2C

2.0E–3

100

Cluster - 2 Cluster - 1 Co - Ref

7710

7712

Energy (eV)

Fig. 28.10 (a) A time-resolved long-duration operando X-ray absorption spectroscopy study of a Co/TiO2 Fischer-Tropsch synthesis (FTS) catalyst at 16 bar and 220  C. (b) By performing principal component analysis (PCA) and cluster analysis (CA), the large number of X-ray

7714

7716

Energy (eV)

7718

0 Cl. - 5

0 8400

Cl. - 4

8200

Cl. - 3

8000

Cl. - 2

7800

5.0E–4

Cl. - 1

7600

1.0E–3

R

50

Co-hcp (ref)

Cluster - 3

%

Cluster - 4 Time

1.5E–3

Norm. abs. (a.u.)

Cluster - 5

absorption spectra obtained during 2 days of operation could be downsized to five different clusters representing the spectral changes observed, where the formation of cobalt carbide (a commonly proposed deactivation mechanism in FTS) could not be linked to a decrease in

620

B. M. Weckhuysen et al.

linearly through the C atom, preferably on a top site, a change in this transition is indicative of a change in orientation of the CO molecule on the surface. Additionally, both in-plane and out-of-plane polarized beams were used to investigate the behavior of Ru(0001)-CO. Such changes in behavior of CO were brought about by pulsing the Ru(0001) surface with 400 nm light, heating it up. As such, XAS spectra revealed a redshift in the peak position upon of the surface. This redshift is attributed to an increase in population of the π* orbital. This indicates CO moving from a top to a bridge or hollow site, a phenomenon taking place on the ps timescale. In the ps that follow, the polarization-dependent XAS spectra show peaks closely resembling gas-phase CO, but still interacting with the surface in several configurations. Upon subsequent cooling down of the Ru(0001) substrate, only some CO desorbs completely, while most readsorbs to the surface. This indicates that the short-lived, transient phase is in fact a precursor state. The entire process described here takes place over a period of 25 ps, illustrating the potential of time-resolved XAS measurements conducted at FELs, an experimental approach which will become more frequently used in the upcoming decade.

28.6

Conclusions and Outlook

Time-resolved in situ and operando X-ray absorption spectroscopy (XAS) is a powerful analytical tool for heterogeneous catalysis research to shed new insights in their (local) structure and functioning, including their reaction and deactivation mechanism. XANES and EXAFS spectra allow in-depth characterization of solid catalysts under various conditions (hydrated, dehydrated, and reduced state) and, most importantly, in their working state (operando characterization approach) on different timescales, very short (milliseconds and faster) as well as very long (hours, days, and longer). Both timescales are of importance in the field of catalysis; especially the latter is sometimes forgotten as solid catalysts evolve toward their active state (“genesis” of the catalyst) and has to be evaluated as well during its deactivation phase to shed insight in deactivation phenomena in order to extend the lifetime of solid catalysts. The XAS methodology can be applied to a plethora of catalytic reactions ranging

from photocatalysis over electrocatalysis to thermocatalysis. Processes on different timescales can be investigated by smart exploitation of equipment and design of experiments. While a true operando experiment investigates the catalyst working at industrially relevant conditions in terms of gas feed composition, temperature, pressure, and reaction products formed, not all timescales relevant to catalysis can be investigated under the same conditions. Kinetics, intermediate species, and transient states are generally investigated on shorter timescales (ns-ps), using, for example, pulsed feed or excitation, lower conversion rates, or alternate catalyst compositions. Such measurements provide time-resolved XAS with the capability to investigate fundamental and shortlived phenomena occurring during catalytic reactions. However, as XAS is element-specific and can therefore only look at a single material at a time, it is important to combine timeresolved XAS with results from other spectroscopic and analytical methods, including theoretical modelling and chemometrics. Only through combing the insights obtained from a multi-technique characterization approach can we truly get to understand a catalytic process in depth. In the coming decade, we expect that technological advancements in various fields will push down the time resolution for (Q)XAS even further without sacrificing spectral resolution or sample integrity. Next to that, we expect that developments in imaging methods, like operando X-ray microscopy and ptychography, will play a key role in revealing structure-performance relationships in all fields of catalysis with both high spatial and temporal resolution, especially combined with techniques from data science like unsupervised learning and data mining. It is expected that long-duration experiments at synchrotron radiation facilities, which can last for months or years, will become feasible by building setups which can be temporarily moved in and out of the X-ray beam, while still fully functional. Finally, the aforementioned technological advancements will not only play an important role at synchrotron radiation facilities and FELs, but they will also make lab-based XAS more viable as a standard laboratory technique, similar to lab-based [XRD] studies. This will make XAS so accessible that in the future we will be able to simply measure our XAS experiments whenever we want and under almost identical reaction conditions as we are currently doing in lab-based catalysis

ä Fig. 28.10 (continued) observed catalyst activity [56]. (b, c) A zoomin of the areas with the highest amount of spectral changes. (d) The time frame of the different clusters with increasing time-on-stream. The colored boxes represent the cluster and the number represents the time in hours. The different clusters were fitted with the spectra of post-

reduction, post-H2 treatment (post-H2) and the Co2C reference, and the linear combination result with the residual is shown in (e). (f) With increasing time-on-stream, the amount of Co2C phase increased, while the amount of metallic Co decreased. (Reproduced from [56], with permission from Wiley-VCH)

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

experiments, thereby allowing for time-resolved (e.g., an entire year or even longer) long-duration operando XAS measurements to fully capture the genesis, life, and death of a solid catalyst. In this manner, the operando concept can be fully captured as the experimental conditions of an industrial reactor can be best mimicked in a lab environment with the ultimate goal to make a molecular movie of the working catalyst. Acknowledgments This work is part of the Advanced Research Center for Chemical Building Blocks, ARC CBBC, which is co-founded and co-financed by the Netherlands Organisation for Scientific Research (NWO) and the Netherlands Ministry of Economic Affairs and Climate Policy. This work was supported by the Netherlands Center for Multiscale Catalytic Energy Conversion (MCEC), an NWO Gravitation program funded by the Ministry of Education, Culture and Science of the government of the Netherlands, and the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement no. 801359. The authors thank T. Hartman (Utrecht University) for the graphical illustrations.

References 1. Newton, M.A., Dent, A.J., Evans, J.: Bringing time resolution to EXAFS: recent developments and application to chemical systems. Chem. Soc. Rev. 31, 83–95 (2002) 2. van Bokhoven, J., Lamberti, C. (eds.): X-Ray Absorption and X-Ray Emission Spectroscopy. Wiley, New York (2016) 3. Koningsberger, D.C., Prins, R. (eds.): X-Ray Absorption: Principles, Applications, Techniques of EXAFS, SEXAFS and XANES. Wiley, New York (1988) 4. Bordiga, S., Groppo, E., Agostini, G., Van Bokhoven, J.A., Lamberti, C.: Reactivity of surface species in heterogeneous catalysts probed by in situ X-ray absorption techniques. Chem. Rev. 113, 1736–1850 (2013) 5. Iwasawa, Y., Asakura, K., Tada, M.: XAFS Techniques for Catalysts, Nanomaterials, and Surfaces. Springer, New York (2017) 6. Calvin, S.: XAFS for Everyone. CRC Press, Boca Raton (2013) 7. Bunker, G.: Introduction to XAFS. Cambridge University Press, Cambridge (2010) 8. Bearden, J.A., Burr, A.F.: Re-evaluation of X-ray atomic energy levels. Rev. Mod. Phys. 39, 125–142 (1967) 9. Wong, J., Lytle, F.W., Messmer, R.P., Maylotte, D.H.: K-edge absorption spectra of selected vanadium compounds. Phys. Rev. B. 30, 5596–5610 (1984) 10. Chen, L.X., et al.: X-ray snapshots for metalloporphyrin axial ligation. Chem. Sci. 1, 642 (2010) 11. Chen, L.X., et al.: Tracking electrons and atoms in a Photoexcited Metalloporphyrin by X-ray transient absorption spectroscopy. J. Am. Chem. Soc. 129, 9616–9618 (2007) 12. Sayers, D.E., Stern, E.A., Lytle, F.W.: New technique for investigating noncrystalline structures: Fourier analysis of the extended X-rayabsorption fine structure. Phys. Rev. Lett. 27, 1204–1207 (1971) 13. Yano, J., Yachandra, V.K.: X-ray absorption spectroscopy. Photosynth. Res. 102, 241–254 (2009) 14. (a) Meirer, F., Weckhuysen, B.M.: Spatial and temporal exploration of heterogeneous catalysts with synchrotron radiation. Nat. Rev. Mater. 3, 324–340 (2018); (b) Beale, A.M., Jacques, S.D.M., Weckhuysen, B.M.: Chemical imaging of catalytic solids with synchrotron radiation. Chem. Soc. Rev. 39, 4656–4672 (2010); (c) Beale, A.M., Jacques, S.D.M., Gibson, E.K., Di Michiel, M.:

621 Progress towards five dimensional diffraction imaging of functional materials under process conditions. Coord. Chem. Rev. 277, 208–223 (2014); (d) Grunwaldt, J.D., Schroer, C.G.: Hard and soft X-ray microscopy and tomography in catalysis: Bridging the different time and length scales. Chem. Soc. Rev. 39, 4741–4753 (2010) 15. Mobilio, S., Boscherini, F., Meneghini, C.: Synchrotron Radiation. Springer, Berlin (2015) 16. Moya-Cancino, J.G., et al.: Elucidating the K-edge X-ray absorption near-edge structure of cobalt carbide. ChemCatChem. 11, 3042–3045 (2019) 17. Zimmermann, P., et al.: Modern X-ray spectroscopy: XAS and XES in the laboratory. Coord. Chem. Rev. 423, 213466 (2020) 18. McNeil, B.W.J., Thompson, N.R.: X-ray free-electron lasers. Nat. Photonics. 4, 814–821 (2010) 19. Dell’Angela, M., Parmigiani, F., Malvestuto, M.: Time resolved X-ray absorption spectroscopy in condensed matter: a road map to the future. J. Electron Spectros. Relat. Phenomena. 200, 22–30 (2015) 20. Bressler, C., Chergui, M.: Ultrafast X-ray absorption spectroscopy. Chem. Rev. 104, 1781–1812 (2004) 21. Prestipino, C., Mathon, O., Hino, R., Beteva, A., Pascarelli, S.: Quick-EXAFS implementation on the general purpose EXAFS beamline at ESRF. J. Synchrotron Radiat. 18, 176–182 (2011) 22. Matsui, H., et al.: Operando XAFS imaging of distribution of Pt cathode catalysts in PEFC MEA. Chem. Rec. 19, 1380–1392 (2019) 23. Uruga, T., et al.: SPring-8 BL36XU: synchrotron radiation X-raybased multi-analytical beamline for polymer electrolyte fuel cells under operating conditions. Chem. Rec. 19, 1444–1456 (2019) 24. van Ravenhorst, I.K., et al.: Capturing the genesis of an active Fischer–Tropsch synthesis catalyst with operando X-ray nanospectroscopy. Angew. Chem. Int. Ed. 57, 11957–11962 (2018) 25. Meirer, F., et al.: Life and death of a single catalytic cracking particle. Sci. Adv. 1, e1400199 (2015) 26. Vesely, A.M., et al.: 3-D X-ray nanotomography reveals different carbon deposition mechanisms in a single catalyst particle. ChemCatChem. 13, 2494–2507 (2021) 27. Fabbri, E., Abbott, D.F., Nachtegaal, M., Schmidt, T.J.: Operando X-ray absorption spectroscopy: a powerful tool toward water splitting catalyst development. Curr. Opin. Electrochem. 5, 20–26 (2017) 28. Bell, A.T.: The impact of nanoscience in heterogeneous catalysis. Science. 229, 1688–1692 (2013) 29. Nørskov, J.K., et al.: The nature of the active site in heterogeneous metal catalysis. Chem. Soc. Rev. 37, 2163–2171 (2008) 30. Hagen, J.: Industrial Catalysis: A Practical Approach. Wiley-VCH, Weinheim (2015) 31. Erlandsson, R.: Gas-induced restructuring of palladium model catalysts studied with atomic force microscopy. J. Vac. Sci. Technol. B Microelectron. Nanom. Struct. 9, 825 (1991) 32. Gai, P.L.: Environmental high resolution electron microscopy of gas-catalyst reactions. Top. Catal. 8, 97–113 (1999) 33. Rupprechter, G., Weilach, C.: Spectroscopic studies of surface-gas interactions and catalyst restructuring at ambient pressure: mind the gap! J. Phys. Condens. Matter. 20, 184019 (2008) 34. Kalz, K.F., et al.: Future challenges in heterogeneous catalysis: understanding catalysts under dynamic reaction conditions. ChemCatChem. 9, 17–29 (2017) 35. Weckhuysen, B.M.: Determining the active site in a catalytic process: operando spectroscopy is more than a buzzword. Phys. Chem. Chem. Phys. 5, 4351–4360 (2003) 36. Bañares, M.A.: Operando spectroscopy: the knowledge bridge to assessing structure-performance relationships in catalyst nanoparticles. Adv. Mater. 23, 5293–5301 (2011) 37. De Wispelaere, K., et al.: Insight into the effect of water on the methanol-to-olefins conversion in H-SAPO-34 from molecular simulations and in situ microspectroscopy. ACS Catal. 6, 1991–2002 (2016)

28

622 38. Vogt, C., et al.: Unravelling structure sensitivity in CO2 hydrogenation over nickel. Nat. Catal. 1, 127–134 (2018) 39. Agostini, G., et al.: XAS/DRIFTS/MS spectroscopy for timeresolved operando investigations at high temperature. J. Synchrotron Radiat. 25, 1745–1752 (2018) 40. Kondrat, S.A., van Bokhoven, J.A.: A perspective on counting catalytic active sites and rates of reaction using X-ray spectroscopy. Top. Catal. 62, 1218–1227 (2018) 41. Bare, S.R., Boubnov, A., Hong, J., Hoffman, A.S.: The consortium for operando and advanced catalyst characterization via electronic spectroscopy and structure (Co-ACCESS) at Stanford synchrotron radiation lightsource (SSRL). Synchrotron Radiat. News. 33, 15–19 (2020) 42. Marinkovic, N.S., Ehrlich, S.N., Northrup, P., Chu, Y., Frenkel, A.I.: Synchrotron catalysis consortium (SCC) at NSLS-II: dedicated beamline facilities for in situ and operando characterization of catalysts. Synchrotron Radiat. News. 33, 4–9 (2020) 43. Frenkel, A.I., Hills, C.W., Nuzzo, R.G.: A view from the inside: complexity in the atomic scale ordering of supported metal nanoparticles. J. Phys. Chem. B. 105, 12689–12703 (2001) 44. Timoshenko, J., Roldan Cuenya, B.: In situ/operando electrocatalyst characterization by X-ray absorption spectroscopy. Chem. Rev. 121, 882–961 (2020) 45. Rothenberg, G.: Catalysis: Concepts and Green Applications. Wiley-VCH, Weinheim (2017) 46. Chorkendorff, I., Niemantsverdriet, J.W.: Concepts of Modern Catalysis and Kinetics, 3rd edn. Wiley-VCH, Weinheim (2013) 47. Calderone, V.R., et al.: De Novo Design of Nanostructured IronCobalt Fischer-Tropsch Catalysts. Angew. Chem. Int. Ed. 52, 4397–4401 (2013) 48. Krishna, S.H., Jones, C.B., Miller, J.T., Ribeiro, F.H., Gounder, R.: Combining kinetics and operando spectroscopy to interrogate the mechanism and active site requirements of NOx selective catalytic reduction with NH3 on Cu-zeolites. J. Phys. Chem. Lett. 11, 5029–5036 (2020) 49. Greenaway, A.G., et al.: Operando spectroscopic studies of Cu–SSZ-13 for NH3–SCR deNOx investigates the role of NH3 in observed Cu(II) reduction at high NO conversions. Top. Catal. 61, 175–182 (2018) 50. Liu, C., et al.: In situ spectroscopic studies on the redox cycle of NH3SCR over CuCHA zeolites. ChemCatChem. 12, 3050–3059 (2020) 51. Paolucci, C., et al.: Dynamic multinuclear sites formed by mobilized copper ions in NOx by selective catalytic reduction. Science. 357, 898–903 (2017) 52. Marberger, A., et al.: Time-resolved copper speciation during selective catalytic reduction of NO on Cu-SSZ-13. Nat. Catal. 1, 221–227 (2018) 53. Bergman, S.L., et al.: In-situ studies of oxidation/reduction of copper in Cu-CHA SCR catalysts: comparison of fresh and SO2-poisoned catalysts. Appl. Catal. B Environ. 269, 118722 (2020) 54. Lomachenko, K.A., et al.: The Cu-CHA deNOx catalyst in action: temperature-dependent NH3-assisted selective catalytic reduction monitored by operando XAS and XES. J. Am. Chem. Soc. 138, 12025–12028 (2016) 55. Ye, X., et al.: Deactivation of Cu-exchanged automotive-emission NH3-SCR catalysts elucidated with nanoscale resolution using scanning transmission X-ray microscopy. Angew. Chem. Int. Ed. 59, 15610–15617 (2020) 56. van Ravenhorst, I.K., et al.: On the cobalt carbide formation in a Co/TiO2 Fischer-Tropsch synthesis catalyst as studied by highpressure, long-term operando X-ray absorption and diffraction. ACS Catal. 11, 2956–2967 (2021)

B. M. Weckhuysen et al. 57. Karan, K.: PEFC catalyst layer: recent advances in materials, microstructural characterization, and modeling. Curr. Opin. Electrochem. 5, 27–35 (2017) 58. Ozawa, S., et al.: Operando time-resolved X-ray absorption fine structure study for Pt oxidation kinetics on Pt/C and Pt3Co/C cathode catalysts by polymer electrolyte fuel cell voltage operation synchronized with rapid O2 exposure. J. Phys. Chem. C. 122, 14511–14517 (2018) 59. Ishiguro, N., et al.: Rate enhancements in structural transformations of Pt-Co and Pt-Ni bimetallic cathode catalysts in polymer electrolyte fuel cells studied by in situ time-resolved X-ray absorption fine structure. J. Phys. Chem. C. 118, 15874–15883 (2014) 60. Tan, Y., Matsui, H., Ishiguro, N., Tada, M.: Three-dimensional XAFS imaging of polymer electrolyte fuel cell cathode catalysts in membrane electrode assembly. Acc. Mater. Surf. Res. 3, 165–171 (2018) 61. Tan, Y., et al.: Pt-Co/C cathode catalyst degradation in a polymer electrolyte fuel cell investigated by an infographic approach combining three-dimensional spectro-imaging and unsupervised learning. J. Phys. Chem. C. 123, 18844–18853 (2019) 62. Nilsson, A., et al.: Catalysis in real time using X-ray lasers. Chem. Phys. Lett. 675, 145–173 (2017) 63. Wang, H.Y., et al.: Time-resolved observation of transient precursor state of CO on Ru(0001) using carbon K-edge spectroscopy. Phys. Chem. Chem. Phys. 22, 2677–2684 (2020) 64. Moya-Cancino, J.G., et al.: In-situ X-ray absorption near edge structure spectroscopy of a solid catalyst using a laboratory-based set-up. ChemCatChem. 11, 1039–1044 (2019) 65. Müller, O.: Hard X-Ray Synchrotron Beamline Instrumentation for Millisecond Quick Extended X-Ray Absorption Spectroscopy. Universitätsbibliothek Wuppertal (2016) 66. Cats, K.H., et al.: X-ray nanoscopy of cobalt Fischer-Tropsch catalysts at work. Chem. Commun. 49, 4622–4624 (2013) 67. Paolucci, C., et al.: Isolation of the copper redox steps in the standard selective catalytic reduction on Cu-SSZ-13. Angew. Chem. Int. Ed. 53, 11828–11833 (2014)

Bert M. Weckhuysen is Distinguished University Professor at Utrecht University. His group focuses on the development of in situ and operando spectroscopy and microscopy to investigate solid catalysts at work. For his work, he has received national and international awards, including the International Catalysis Award and the Spinoza Prize. He is an elected (foreign) member of a.o. KNAW, KVAB, and Academia Europaea.

28

Time-Resolved X-Ray Absorption Spectroscopy (XAS)

Caterina Suzanna Wondergem received her PhD from Utrecht University, the Netherlands, in 2019. During this time, she worked on the development of in situ techniques for heterogeneous catalysis with professor Bert Weckhuysen. In 2020, she received a JSPS postdoctoral fellowship from the Japanese Society for the Promotion of Science to work on advanced X-ray imaging of electrocatalysts at Nagoya University, and in 2021 was promoted to designated assistant professor at the same institute.

623

Charlotte Vogt received her PhD with highest distinctions from Utrecht University in 2020 working with professor Bert Weckhuysen. In 2021, she was named one of “Forbes 30 under 30 Europe” and started her own research group as assistant professor at the Technion Institute for Technology in Israel, which focuses on a deep fundamental understanding of catalytic processes.

28

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS) , Camille La Fontaine

Aline Ribeiro Passos Vale´rie Briois

, Ame´lie Rochet

29

, and

Contents 29.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625

29.2

Multivariate Curve Resolution with Alternating Least Square (MCR-ALS) Analysis . . . . . . . . . . . . . . . . . . . . Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rank Determination of Matrix D . . . . . . . . . . . . . . . . . . . . . . . . . Initial Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations: Deviation of the Bilinearity Model for Evolutionary Data Set Recorded in Temperature . . . . Limitations: Rank Deficiency by Existing Correlated Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29.2.1 29.2.2 29.2.3 29.2.4 29.2.5 29.3 29.3.1 29.3.2 29.4

626 626 628 628 630 633

How Time Resolution Can Give Insights on “Birth, Life, and Death” of Solid Catalysts . . . . . . . . . . . . . . . . . . . . 636 Preparation of Catalysts: From Solution Processes to Solid-State Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636 Catalysts in Operation: From Active Phases to Spent Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654

Abstract

Intermediate and active site identification and quantification are crucial for understanding the structure-activity relationships and designing efficient catalysts. We overview how time-resolved XAS (TR-XAS) with the emerg-

ing use of multivariate data analysis allows the isolation of the pure species and the description of their evolution over time. Special emphasis is placed on the concepts of multivariate curve resolution with alternating least square (MCR-ALS) fitting. Investigation of catalyst preparation steps highlights the strength of TR-XAS to solve fast chemical transformations in liquid and solid states, including nanoparticle nucleation and growth, metal dispersion and redistribution, formation of undesirable phases, and metal-support interaction. Moreover, recent operando monitoring of the direct methane to methanol conversion, CO oxidation, methanol synthesis by CO2 hydrogenation, and ethanol steam reforming are presented illustrating the great potential of TR-XAS to quantify active sites and reaction intermediates. The main achievements to identify deactivation mechanisms including particle sintering, atom redistribution in bimetallic particles, and oxidation-state change are discussed. The combination of TR-XAS with complementary techniques is outstanding to fully describe deactivation by carbon deposition and regeneration. It is expected that TR-XAS combined with MCR-ALS analysis will be groundbreaking in research in many catalysis science areas. Keywords

A. R. Passos Cateretê Group, Brazilian Synchrotron Light Laboratory, Campinas, Brazil e-mail: [email protected] C. La Fontaine ROCK Beamline, SOLEIL Synchrotron, Gif-sur-Yvette, France

Time-resolved X-ray absorption spectroscopy · QuickEXAFS · Multivariate analysis · MCR-ALS · Catalysis · Nucleation and growth · Active species · Reaction · Deactivation · Regeneration

A. Rochet LNLS, CNPEM, Campinas, Brazil e-mail: [email protected]

29.1

V. Briois (*) ROCK Beamline, SOLEIL Synchrotron UR1-CNRS, Gif-sur-Yvette, France e-mail: [email protected]

In the last decade, several new time-resolved X-ray absorption spectroscopy (TR-XAS) beamlines [1–4], totally or partly dedicated to operando characterization of catalysts,

Introduction

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_29

625

626

have offered the time resolution to unravel complex mechanisms involved in catalyst synthesis, activation, reaction, or deactivation. The time resolution improvement was strongly related to the access to third-generation synchrotron radiation facilities offering the mandatory monochromatic flux of 1011 (or more) photons per second on the sample, to record high-quality sub-second absorption spectra [5]. Moreover, as beneficial “side effect” of the high brightness of new-generation synchrotron sources, more sophisticated sample environments can be used for in situ characterizations. In this framework, all the aspects of the catalyst life span can be scrutinized with deeper details by XAS. New strategies can be proposed for the design of more efficient catalysts which encompass the development of challenging synthesis routes and optimization of activation treatments. The study of the structure-performance relationships using operando characterizations under realistic conditions allows the identification of active species and finally the development of new regeneration strategies of deactivated catalysts through the in-depth understanding of their performance loss. The larger volume of XAS data obtained with the current sub-second time resolution gives rise to more complex and precise information on all steps of the catalyst life span but also to an apparent greater workload to analyze in depth the huge number of spectra. To overcome this difficulty, multivariate data analysis has been explored by several groups to reduce the data complexity of time-resolved studies. Sometimes qualified as an unsupervised machine learning method, the multivariate data analysis allows reducing the deep analysis of thousands of measured spectra to one of a few spectra representative of the evolving pure species. Therefore, the data analysis has been largely simplified thanks to fast-handling tools for automatic energy calibration/normalization [6–10] and thanks to those methods isolating the chemical species and describing their evolution upon a reaction parameter (temperature, time, and so on) [11–17]. Further, using conventional EXAFS data treatment and/or comparison with spectra of known phases, the chemical nature of those pure species can be identified. The chapter is organized as follows: the first part will present the basic concepts underlying the multivariate data analysis of TR-XAS including a discussion of their limitations and strategies to overcome them. In the second part, selected TR-XAS studies illustrating the state of the art of catalyst characterizations at different stages of their life span will be discussed. We will focus on the use of the technique for unravelling dynamic processes which are of paramount importance in the preparation, use, and regeneration of heterogeneous catalysts.

A. R. Passos et al.

29.2

Multivariate Curve Resolution with Alternating Least Square (MCR-ALS) Analysis

29.2.1 Basic Concepts The catalytic steps are hardly a two-phase transformation process. The temperature programmed reduction (TPR) of supported Cu(II) species reveals, for instance, a two-stage process with the formation of Cu(I) and Cu(0) species [11]. Supported cobalt catalysts, which find wide applications in many catalytic processes such as Fischer-Tropsch [18–22] or ethanol steam reforming reactions [23–25], must be firstly reduced into metallic active species. This activation, simple at first glance, often involves a mixture of oxidic species. Alumina-supported calcined cobalt catalysts presenting Co3O4 and CoAl2O4 phases go through the formation of CoO before Co(0) species but with a well-known different reducibility. Those systems stress out that XAS investigations of the activation face a problem of phase mixtures. The knowledge gained on the identification and quantification of those phases evolving under activation is of prime importance for improving catalyst performances. Multivariate curve regression methods are highly relevant to solve the problem of mixture of phases observed by analytical techniques in which the measure is the sum of the signal contributions coming from each one of the mixture components [26]. XAS is governed by the Beer-Lambert’s law and thus totally fulfils the aforementioned condition for the use of the MCR-ALS method. The rows of the experimental matrix D, gathering the q-XAS spectra collected during a reaction with k energy points, can be described by a bilinear model according to the following relation: T

D5C :S þE

ð29:1Þ

where the columns of matrix C contain the relative proportions of the n pure species in the mixture, the rows of the matrix ST are the spectra of those n pure species (ST meaning the transpose of matrix S), and E is the matrix expressing the error or variance unexplained by the C.ST product which should be in principle not larger than the experimental noise (Fig. 29.1). The data presented herein correspond to the Cu K-edge quickXAS monitoring of the self-reduction of an alumina-supported Cu catalyst [27]. In this example, the minimization considering three components leads to the isolation of spectra which are characteristic of pure Cu(II)-Cp1, Cu(I)-Cp2, and Cu(0)-Cp3 species and to their related concentration profiles. The principle of the MCR-ALS analysis is to proceed to a least square minimization of E based on the alternative optimization of C and S at each iterative cycle. For this scope,

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

q : number of spectra

k : number of energy points

n

a1 a2 .......ai ai+1 ai+2.........ak a1 a2 .......ai ai+1 ai+2.........ak

627

k

k

ST Pure spectrum =

q

n

C

D

+

E = experimental noise

a1 a2 .......ai ai+1 ai+2.........ak

Composition (%)

1.2 1 0.8 0.6 0.4 0.2 0 200 250 T (°C)

Cp1 Cp2 Cp3

80 60 40

29

1.4

100 Normalized absorbance

q

n 2.6 Å1)

rather than the XANES spectrum (k  2.6 Å1). Consequently, MCR-ALS analysis of a data set built from temperature-evolving EXAFS spectra can potentially suffer from an overestimation of the number of components, some of them being chemically identical but affected by the DebyeWaller damping. This is, for instance, illustrated in Fig. 29.4 with the MCR-ALS analysis of the matrix composed of the

632

A. R. Passos et al.

a) 8 6

k3 chi (k)

4

500 °C

2 0 –2 –4 –6

RT

–8 0

2

4

6

8

10

k scale (Å–1)

b)

PCA1 evolution

PCA2 evolution

0.055

PCA3 evolution 0.08

0.12 PCA1 eig. = 904.013

PCA2 eig. = 226.7852

0.1

0.045

0.08

0.04

0.06

0.02

0.04 0.035 0.03

0.04

0

0.02

–0.02

0

–0.04

0.025

–0.02

–0.06

0.02

–0.04

0

100

200 300 400 Spectrum index

500

0

100

PCA4 evolution

200 300 400 Spectrum index

500

–0.08

0

100

PCA5 evolution

0.1

200 300 400 Spectrum index

500

PCA6 evolution

0.15

0.1

PCA4 eig. = 24.5268

PCA5 eig. = 16.4895

0.1

PCA6 eig. = 13.8251

0.05

0.05 0.05

0

0

0

–0.05

–0.05 –0.05

–0.1

–0.1 –0.1

PCA3 eig. = 61.3491

0.06

0.05

0

100

200 300 400 Spectrum index

–0.15

500

Eigenvalue

c)

0

100

904.0

200 300 400 Spectrum index

500

–0.15

0

100

200 300 400 Spectrum index

500

Scree plot in log scale

103 226.8 102

61.3 24.5

101 0

2

8.0

4 6 Component number

Fig. 29.4 MCR-ALS analysis of Ni K-edge quick-EXAFS data recorded during the TPR of a calcined Ni oxidic precursor catalyst. (a) Quick-EXAFS normalized data. Rank analysis using (b) the 1D score trajectory plot and (c) the scree plot. (d) Outcome of the MCR-ALS optimization considering 3 components. It is noteworthy that the

8

10

MCR-ALS resolution has been performed with the addition of EXAFS spectra of fully reduced Ni(0) supported species recorded at 500  C (data with indexes higher than 599 in (d)) and corresponding to the yellow-colored EXAFS spectrum in (d)

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

d)

ALS opitimization

Copy

Print

Concentration profiles 1

cp1 cp3

0.5 cp2 0

100 200 300 400 500 600 700 800 900 1000 Spectra

5 0 –5 20

40

60

Plots are optimum in the iteration Nr. 7

Information

80

100 120 140 160 180 200 220 Convergence is achieved!!! Std. dev of residuals vs. exp. data = 0.18245

Fitting error (lack of fit, lot) in % (PCA) = 4.3042 Fitting error (lack of fit, lot) in % (exp) = 7.4713

MCR bands

Percent of variance explained (r2) at the optimum is = 99.4418 Back 2 const

Back 2 param

Fig. 29.4 (continued)

EXAFS spectra recorded during the TPR of a Ni catalyst [31]. Chemical rank analysis concludes to the presence of three components (Fig. 29.4b, c). The outcome of the MCR-ALS analysis (Fig. 29.4d) reveals that the first two components display EXAFS oscillations in phase, both characteristic of the same NiO-based species but with amplitude for the second component suffering from damping when the temperature is increased. The transformation of the NiO species into Ni(0) species occurs only above 300  C, which corresponds to a spectrum index greater than 420 (Fig. 29.4d). Considering the elegant demonstration done by Martini et al. [15] on the negligible temperature dependence of the XANES spectra, a first strategy to overcome the bilinearity deviation associated to the temperature increase can be the MCR-ALS minimization considering the XANES region only. This strategy, presented in Fig. 29.5 (same data set as Fig. 29.4), allows obtaining accurate determination of the number of “pure” chemical species involved in the reaction process and their associated concentration profiles. Only two components are significant in the scree plot. Identification of the chemical species involved in the reaction is therefore obtained from the analysis of the XANES spectra of pure species, using comparison with spectra of already recorded known references or the use of ab initio XANES spectrum calculations [32]. The minimization based on XANES data set has the drawback that no pure EXAFS spectra can be

633

isolated and further used for identifying the pure chemical species by least square fittings. However the knowledge of concentrations determined by MCR-ALS at a given process temperature can be used for a better simulation of the experimental EXAFS data, thus without temperature limitation [25], taking into account the right contributions of the different species in the EXAFS signal. A second strategy for temperature-dependent XAS data set is to consider that most of the variance related to structure modifications (by phase transformation and not by temperature increase) is contained in the XANES part and that a minimization of a normalized data set containing both the XANES and the EXAFS energy regions is a good compromise to overcome the temperature-damping effect, while keeping the possibility to isolate the EXAFS spectra of pure chemical species. This approach has been successfully used for a wide number of catalyst activation and reaction processes and has demonstrated its efficiency to further identify the chemical species [6, 12, 24, 29, 33–37]. However, it is noteworthy that for a given species, the EXAFS spectrum extracted by MCR-ALS and rebuilt from the sum of the contributions of this species over the full process is affected by a damping contribution related to the temperature window of existence of this component. Therefore, this strategy allows for EXAFS fitting and chemical identification but in no case for any fine discussion on Debye-Waller modifications occurring during the chemical process.

29.2.5 Limitations: Rank Deficiency by Existing Correlated Data One of the main issues of the multivariate analysis is related to the so-called rank deficiency of the data, an issue which occurs when large similarity among the spectra of pure species or coevolving concentration profiles exist [26]. In that case, the number of components detected by PCA-SVD is lower than the number of chemical species present in the mixture. This is, for example, illustrated in Fig. 29.6a with the Zn K-edge quick-XAS monitoring of the thermal decomposition of the same Zn-Cu-Al layered double hydroxide (LDH) sample with a heating ramp of 5  C/min (D1 matrix) and 10  C/min (D2 matrix) in static air. Both data sets are described individually with a chemical rank of 2, a value which is underestimated considering the extra information gained by wide-angle X-ray scattering (WAXS) and the final XANES spectra. Indeed, WAXS monitoring of the thermal decomposition of the same sample points out the formation of nanocrystalline ZnO phase while a careful examination of the XANES spectra obtained at the end of both decompositions [38] does not reveal characteristic XANES fingerprint of nanocrystalline ZnO. The outcome of the decompositions is

29

634

A. R. Passos et al.

a) 1.6

c)

1

1.4 0.8

Fraction of species

Normalized absorbance

500 °C 1.2 1 0.8 0.6 0.4

Comp1-NiO Comp2-Ni(0) 0.6

0.4

0.2

RT 0.2 0 8300

0 8320

8340

b)

8360 8380 Energy (eV)

8400

0

8420

d) 1.6

Normalized absorbance

431.3

Eigenvalue

102 16.7 1.47

0.66

0.05 0

2

4

6

200

300 400 Spectrum index

500

600

1.4

Scree plot in log scale

100

100

8

10

Component number

1.2 1 0.8 0.6

Comp1-NiO Comp2-Ni(0)

0.4 0.2 0 8300

8320

8340

8360 8380 Energy (eV)

8400

8420

Fig. 29.5 MCR-ALS analysis of Ni K-edge quick-XANES data recorded during the TPR of a calcined Ni oxidic precursor. (a) QuickXANES normalized data, (b) rank analysis using the scree plot, (c and d) outcome of the MCR-ALS optimization considering two

components. It is noteworthy that the MCR-ALS resolution has been performed with the addition of XANES spectra of fully reduced Ni (0) supported species recorded at 500  C. LoF ¼ 0.46% and 99.9979% of variance explained

actually a mixture of species, one of them being ZnO. In order to overcome the rank deficiency characterizing each individual data set, the so-called column-wise augmented (CWA) data strategy can be used. It consists in analyzing simultaneously both data sets with the assumption that irrespective of the heating ramp, D1 and D2 matrices share common chemical components, expressed in the single ST matrix. C1 and C2 matrices contain the concentration profiles related to each data set. SVD analysis of the augmented data set detected three components. Figure 29.6b, c displays the concentration profiles and the spectra obtained within the CWA strategy. The thermal decomposition of the pristine LDH gives rise to the formation of nanocrystalline ZnO and ZnAl2O4 phases which have been identified by comparing their spectra isolated by MCR-ALS with those of synthesized references. From the comparison of the concentration

profiles, the reason for rank deficiency of each data set is clearly related to coevolving concentrations of ZnO and ZnAl2O4 nanophases. The simultaneous resolution of both data sets allows breaking this rank deficiency because the relative concentrations of both phases obtained by thermal decomposition vary sufficiently from D1 to D2. This example emphasizes how powerful is the multi-data set analysis to overcome the rank deficiency issue of individual data sets. It points out also that the users must take a critical look at the PCA-SVD analysis regarding the estimation of the number of components for evolutionary processlike data. MCR-ALS is often described as a “blind-source separation” method [13]. “Blind” is only justified by the fact that MCR-ALS does not require a priori the use of any standards to isolate the chemical species involved in the reaction. In that sense, the method is really unique. In

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

D1 5 °C/min

Normalized absorbance

2 1.8

1.8

1.6

1.6

1.4 1.2 1 0.8 0.6

1 0.8 0.6

0.2

0.2 9660

9680

9700 9720 Energy (eV)

9740

9760

RT

[S]T

= D2

E2

110 LDH

100 90

LDH

D1 5 °C/min

80

D2 10 °C/min

60 50 40

Nano-ZnO

Nano-ZnAl2O4

20

1 MCR-ALS Spectrum index 1

0.5

9650

60 40 Spectrum index

Plots are optimum in the iteration Nr. 73

80

100

120 140 160 Spectrum index

Convergence is achieved!!! Std. dev of residuals vs. exp. data = 0.0054723 Fitting error (lack of fit, lof) in % (PCA) = 0.18726 Fitting error (lack of fit, lof) in % (exp) = 0.79967 Percent of variance explained (r2) at the optimum is = 99.9936

1 MCR-ALS Nano-ZnAl2O4 Bulk-ZnAl2O4

0.5

9650

9700

9750

2 Nano-ZnO 1.5 1 MCR-ALS 4 nm-ZnO Bulk-ZnO

0.5 0

Fig. 29.6 (a) Zn K-edge quick-XAS monitoring of the thermal decomposition of Zn-Cu-Al LDH samples at 5  C/min (D1 matrix) and 10  C/ min (D2 matrix). (b) Concentration profiles of each data set and (c)

9750 Nano-ZnAl2O4

180 Normalized absorbance

20

9700

1.5

0 0

9760

Pristine-LDH

1.5

10 0

9740

2

Nano-ZnAl2O4

30

9700 9720 Energy (eV)

MCR-components identification

0

Nano-ZnO

70

9680

2

+

C2

9660

c)

E1

C1

29

0 9640

Normalized absorbance

D1

% Zn species

1.2

0.4

Column-wise augmented data matrix in the C-direction: a solution for system with rank deficiency (e.g. 2 species with ª formation rate)

b)

1.4

0.4

0 9640

D2 10 °C/min

450 °C 2

Normalized absorbance

a)

635

Normalized absorbance

29

9650

9750 9700 Energy (eV)

9800

MCR-ALS isolated spectra of pure species obtained in the CWA data resolution, considering a rank for D1 þ D2 equal to 3, and identified by comparison with spectra of known references

636

A. R. Passos et al.

particular in catalysis, it allows the identification of chemical species from process-like data set, for which references are not available either due to the nanocrystalline nature of the species leading to spectra different from their bulk counterparts [11] or due to the difficulty to synthesize them as, for instance, for metallic species in interaction with zeolite framework [16]. However, the way to properly perform the MCR-ALS optimization often requires other information obtained by complementary techniques [14, 35, 38–40]. For instance, additional equality constraints associated with the knowledge of the presence of pure species can be introduced in the minimization, either as the spectra of those species or as their presence windows in the concentration profiles [26]. Finally, it is noteworthy that the CWA method has been successfully used in numerous recent publications related to MCR-ALS analysis of time-resolved data [29, 34, 37–41]. This clearly points out that any rank deficiency of a data set should be detected at the beamline during the data collection to allow the user to immediately perform additional measurements by changing a parameter of evolution (heating ramp, gaseous atmosphere, and so on) to break the rank deficiency. This justifies the strong effort carried out nowadays by beamline scientists to offer more automated data analysis strategies. Until recently, linear combination fitting (LCF) of XAS spectra was the common methodology for the determination of species concentration provided that the spectra of appropriate reference compounds were available. Currently, the more sophisticated and powerful MCR-ALS analysis is widely used with the advantage that not only the concentration profiles are determined but also the XAS spectra of those species allowing their further identification by EXAFS fitting.

29.3

How Time Resolution Can Give Insights on “Birth, Life, and Death” of Solid Catalysts

As pointed out in a perspective article [42], the developments of time-resolved operando and in situ methods are crucial to study the “birth, life, and death” of catalysts operating under harsh conditions.

29.3.1 Preparation of Catalysts: From Solution Processes to Solid-State Reactions Among the different stages of the catalyst life span, the “birth” stage, which refers to the catalyst preparation, is probably the one which greatly took advantages of the last decade of development of TR-XAS beamlines, at least for the genesis of multicomponent nanoparticles from solution processes [14]. Taking into account the incisive influence of the

morphology and size of metallic nanoparticles on the catalytic activity [43], the understanding of the mechanisms and kinetics of precursor transformation is highly demanded to achieve a comprehensive control of the morphological parameters. Besides those bottom-up preparation routes of colloidal particles in liquid media, the solid-state preparation route, encompassing support impregnation and subsequent thermal treatment, is also crucial to control and rationalize. This section will be focused on a few selected results gained by TR-XAS: first on nucleation and growth of colloidal nanoparticles and then on thermal decomposition routes for better controlling the dispersion and nature of active phases.

Solution Preparation of Colloidal Particles The study of bottom-up particle preparation routes gives rise to tremendous efforts to develop appropriate in situ characterizations due to the short time span of nucleation processes [44] but also to explore new preparation strategies. The nucleation and growth mechanisms are not unique and, depending on the chemical systems, can be supported either by the classical nucleation theory (CNT), where nucleation and growth are conceptually separated in time, or the nonclassical nucleation theory, where growth occurs partially simultaneously with the nucleation [45]. Further, various growth laws can be highlighted for ripening or oriented attachment growth models. Irrespective of the nucleation pathways, in situ TR-XAS has been recognized as being unique for shedding light on the nucleation stage. For instance, considering the LaMer model for nanoparticle generation which is the basis of the CNT, nucleation spontaneously occurs when the concentration of free monomers formed by the chemical transformation of pristine precursor reactants reaches a level of supersaturation (Fig. 29.7). During the nucleation encompassing stage I and stage II of Fig. 29.7, chemical reactions of precursors with solvent or reducing species can induce ligand exchange around the metal ions and/or change of formal oxidation state of metal ions. Both can be dynamically well studied by TR-XAS, revealing the formation of the free monomers and when nucleation occurs. Due to their relevant applications in catalysis, noble metal nanoparticles have been considerably studied using various preparation routes in solution with, as quite common characteristic, a fast nucleation rate in the second or millisecond timescale when using strong reducing chemical agents [46–49]. For instance, the formation of gold nanoparticles by reduction of AuCl3 by BH4 ions required the use of sub-second timescale available at dispersive XAS [47] or quick-EXAFS beamlines [46]. Comparison of Au L3-edge XANES spectra recorded with 100 ms time resolution upon nanoparticle formation with references of Au(III), Au(I), and Au(0) species reveals that after 104 ms, the pristine Au(III) ions have been reduced all into the Au(I) form (Fig. 29.8a).

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

Atomic concentration (C)

29

nu Cmax

nu Cmin

Cs I

II

III Time (t)

M+

M

Generation of free monomers

Selfnucleation

Formation of seed

Growth

Fig. 29.7 Schematic illustration of the LaMer model describing the three stages of colloidal nanoparticle formation in solution according to the CNT. The generation of precursors encompasses the transformation of precursor reactant (M+ in orange) into nanoparticle precursors also called free monomers (M(0) in blue). (Reprinted with permission from Ref. [45]. Copyright (2016) Elsevier)

Then gradually Au(I) is transformed into Au(0), and the complete reduction is achieved in less than 20 s. Figure 29.8b compares the concentration profiles of Au(0) after the reactant mixing (red curve) with the amount of Au(0) specifically in nanoparticles derived from the analysis of time-resolved small-angle X-ray scattering (SAXS) data (blue curve). It is noteworthy that the red curve determined by XAS is equal to all the Au(0) species in solution, i.e., nanoparticles but also free monomers formed upon reduction of Au(I) species. The subtraction of both curves is consequently equal to the amount of Au(0) free monomers, and their variation upon time is a perfect illustration of a nucleation governed by the CNT [45] as presented in Fig. 29.7. CNT model assumes that the self-formation of seeds occurs through the successive accretion of free monomers and through dissolution of unstable pre-nucleation clusters characterized by a subcritical size. Smart experiments were recently carried out for the study of the onset of nucleation of palladium nanoparticles [49], in which the concept of reaction at the interface of two immiscible liquids has been used in order to confine the nucleation in space but also to control the onset of nucleation through electrochemical trigger. TR-XAS measurements revealed that at peculiar PdCl42 concentrations, spontaneous reduction of aqueous PdCl42 occurred at the interface with an organic solution of ferrocene acting as a reducing agent, according to a stochastically fluctuating equilibrium. The Pd(0) nuclei concentration fluctuated over a number of hours suggesting random oscillations between PdCl42 reduction (leading to the formation of pre-nuclei with subcritical size) and their subsequent

637

dissolution (Fig. 29.9b). At the opposite, under potential control, the oscillatory growth is eliminated, and a continuous increase of the XAS edge jump is observed attesting to the increase of concentration of stable Pd(0) nuclei (Fig. 29.9a). XAS spectroscopy can be used beyond the only oxidationstate characterization of the multivalent precursors before nanoparticle nucleation. For some systems, XAS provides also important information about the chemical nature of the ligands and intermediates before nucleation, which enlightens their crucial role in tailoring the final shape of the nanoparticles [50, 51]. The versatile use of oleylamine (OLAM) reagent has been recently explored to tailor the composition, size, shape, and crystalline structure of various nanoparticle systems. Using TR-XAS measurements, the role of oleylamine in combination with trioctylphosphine (TOP) or trioctylphosphine oxide (TOPO) has been comprehensively discussed rationalizing the shape control of copper nanoparticles prepared from CuBr under high temperature [51]. Thanks to the combination of XAS, nuclear magnetic resonance, and mass spectrometry (MS), it has been shown that the choice of reagents first impacts the structure of the Cu (I) precursor (Fig. 29.10a). Then, the subsequent heating and aging at 260  C lead to the formation of nanospheres for the CuBr-OLAM-TOP system and nanocubes for the CuBrOLAM-TOPO as measured by transmission electron microscopy (TEM) (Fig. 29.10a). The in situ TR-XAS monitoring reveals very different kinetics of Cu nanoparticle nucleation driven by the structural differences of the pristine complexes. Actually, under temperature, a mixture of equal proportions of Cu(I), Cu(II), and Cu(0), resulting from a disproportionation equilibrium (2 Cu(I) ! Cu(II) þ Cu(0)), is evidenced from the presence of characteristic XAS features related to each oxidation state (Fig. 29.10b, e). For the CuBr-OLAMTOPO system, the three species evolve with the same kinetics upon heating to 260  C (Fig. 29.10d), in agreement with the disproportionation. Under the same conditions, a deviation from the equilibrium is observed for the CuBr-OLAM-TOP system with the faster appearance of Cu(0) species (Fig. 29.10c), indicating that Cu(0) nuclei are formed and no longer participate to the disproportionation equilibrium. The nucleation occurs continuously during the heating ramp for the CuBr-OLAM-TOP system, whereas it appears abruptly after 10 minutes at 260  C for the CuBr-OLAMTOPO system, as evidenced with the rising of the signal characteristic of Cu(0) at the detriment of Cu(I) and Cu (II) signals (Fig. 29.10e). For the CuBr-OLAM-TOP system, the gradual flux of Cu(0) monomers favors the thermodynamically controlled shape, i.e., nanospheres, while for the CuBr-OLAM-TOPO, the sudden increase of Cu(0) monomers leads to the formation of particle shapes dictated by the kinetic regime, i.e., nanocubes.

29

638

A. R. Passos et al.

a) 4

b) 1 Ref Au(III)

3

Normalized gold concentration

Normalized absorbance

3.5

t=104 ms

2.5

Ref Au(I)

2

t=20s

1.5

Ref Au(0)

1

)

tal To

0.8

(0 Au

tnucl Particles

0.6

0.4 Monomers

0.2

0.5 0 11.9

11.95

12

12.05

0

0

5

10

Energy (keV)

15

Time (s)

Fig. 29.8 (a) Au L3-edge XANES spectra recorded at different times of the gold nanoparticle synthesis and (b) concentrations of the different species during the reaction: Au(0) as determined by LCF of the timeresolved XANES data recorded upon synthesis (red symbols); Au (0) contained in the nanoparticles measured by SAXS (blue symbols);

b)

1.55

1.45 1.40

m (E )

1.50

0

30

0

0.12 0.10 0.08

6

0.06 5

V)

(e gy ner

E

Ti m

e(

h)

2

0

0. 24 04 60 24 550 0 5 24 00 45 24 0 40 V) 24 0 y (e 35 g r 24 0 Ene 30 0 24

0

0

24

0

0

35

0

0

45

40

60

55

50

0.14

1

24

24

2

h)

4

e(

24

24

6

Ti m

24

24

0.16

4

8

10

12

1.35

0.03350 0.04665 0.05980 0.07295 0.08610 0.09925 0.1124 0.1255 0.1387 0.1518 0.1650

3

1.303 1.331 1.359 1.387 1.415 1.443 1.471 1.499 1.527 1.555 1.583

m (E )

a)

Au(0) free monomers (green symbols) deduced by subtraction of the red and blue curves, responsible beyond the supersaturation point to spontaneous aggregation to form nuclei. (Reprinted with permission from Ref. [47]. Copyright (2010) American Chemical Society)

Fig. 29.9 Pd K-edge XAS monitoring of the nucleation of Pd nanoparticles at the liquid-liquid interface (a) under electrochemical triggering and (b) in the absence of applied potential. For (a), [PdCl42] ¼

5 mM ¼ [reducing agent], whereas for (b), [PdCl42] ¼ 10 mM and [reducing agent] ¼ 20 mM. (Reprinted with permission from Ref. [49]. Copyright (2017) Elsevier)

The bottom-up preparation of colloidal particles is specially used for the preparation of photocatalysts, electrocatalysts, or heterogeneous catalysts for selective hydrogenation of organics in autoclaves. Such catalysts do not require severe pretreatments before their use, except the deposition of the preformed particles on the adequate

support [22]. The removal of capping ligands remains nevertheless a challenge for obtaining active and stable heterogeneous catalysts, which is probably insufficiently addressed by in situ and/or TR-XAS characterizations. The landscape is very different regarding conventional preparation of heterogeneous catalysts, for which XAS

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

a)

639

Br

TOPO Cu

Br N

80 ºC 15 min

Cu N

N

N CuBr(OLAM)3

Heating to 260 °C 260 °C, 60 min

P

Cu

Br Br

Cu {CuBr(TOP)2}2

P

100 nm

P

b)

CuBr+OLAM +TOPO

O N

P

TOP

CuBr+OLAM +TOP

CuBr(OLAM)2(TOPO)

c)

100 nm

e)

Time

CuOAc CuO Cu2O Cu

0.6 0.4 Cu (II) Cu (I) Cu (0)

0.2

20 40 Time (min)

60

CuBr + OLAM + TOP

d) 1.0

0.1 CuBr + OLAM + TOPO

0.2

CuBr+OLAM +TOPO 0.8

8990 Energy (eV)

9000

0.6

0.5

0.4

CuBr+ OLAM+ TOPO

0.6 0.3 0.4 Cu (II) Cu (I) Cu (0)

0.2 0.0

8980

CuBr+ OLAM+ TOP

0.7

0

0.0

60 min

0.8

0.0

0.1

Normalized intensity

Normalized intensity

0.2

0 min Cu(0) Cu(I) Cu(II)

Normalized intensity

Cu(0) Cu(I) Cu(II)

Normalized intensity

0.3

1.0 CuBr+OLAM +TOP 0.8

0

20 40 Time (min)

0.2 60

8980 8982 8984 8986 8988 Energy (eV)

Fig. 29.10 (a) Proposed structures for the complexes formed from reactions of CuBr(OLAM)3 with TOPO and TOP (TOP, TOPO, and OLAM ligands are, respectively, represented by their donor atoms P, O, and N) and representative shapes of nanoparticles after heating and aging at 260  C. (b) displays some reference spectra and stack plots of Cu K-edge spectra recorded upon heating from 80 to 260  C for both systems and obtained by subtracting the XANES spectrum at 80  C

from each subsequent spectrum. (c) and (d) display the intensities of the Cu(0), Cu(I), and Cu(II) pre-edge structures normalized by the highest value observed during the heating ramp. (e) is the stack plots of Cu K-edge spectra recorded upon aging at 260  C for 60 min. (Reprinted with permission from Ref. [51]. Copyright (2019) American Chemical Society)

has merely contributed to better understand the impact of the catalyst pretreatment on the structure-activity relationship.

calcination, and activation [22]. Irrespective of the catalytic reaction, synthesis parameters and support characteristics are of crucial importance. They directly alter the catalyst activity by affecting the metal dispersion but also its distribution. This section will examine some elementary preparation steps of heterogeneous catalysts impacting the catalyst properties and for which TR-XAS provides real-time knowledge of phase transformations.

Synthesis of Supported Catalysts The industrial preparations of supported catalysts often use solid-state routes based on the simple and inexpensive impregnation of metal salts on supports followed by drying,

29

640

Active-Phase Dispersion: Particle Size and Particle Density Dispersion herein encompasses two different concepts: the first one is the ratio between the number of atoms exposed at the surface of the particle and the total number of atoms in the particle. The smaller the particles, the larger their surface contributions, the greater the dispersion values. The second is the density of particles, which is related to the concept of distance between adjacent particles on the support. Both concepts have a direct bearing on catalytic activity and stability. The recent progress in in situ TEM under reactive environments leads to unique opportunities for accessing the sizes, shapes, particle size distribution, and density [52]. XAS offers complementary advantages, such as its high time resolution for in situ and operando characterizations, to highlight a specific key parameter of the preparation protocol, influencing the particle dispersion. For instance, in-depth investigations of preparation parameters of cobalt-based Fischer-Tropsch (FT) catalysts were carried out to optimize the dispersion of the active metallic Co(0) particles. Among them, the calcination atmosphere for the thermal activation of supported cobalt hydrated species has been scrutinized by several groups [18–22]. In particular, it was reported that the calcination under NO-containing atmosphere enabled the formation of smaller Co3O4 particles on porous silica supports [53] than the one carried out under air. The Co K-edge quick-XAS monitoring of the calcination revealed different decomposition pathways depending on the reactive atmosphere (Fig. 29.11) [20]. Starting from the Co(H2O)6(NO3)2 species, the calcination under air first led to the formation of the anhydrous Co(NO3)2 species, which was further gradually transformed into Co3O4, as evidenced by the characteristic energy shift of the rising edge associated to the cobalt oxidation and by the second neighbor contributions above 160  C in the Fourier transform moduli (Fig. 29.11a1 to c1). NO-assisted decomposition gave rise to the formation of a supplementary intermediate species before the formation of Co3O4, identified as a cobalt hydroxynitrate Co(OH)x(NO3)y phase (Fig. 29.11a2 to c2). It was underlined that the brutal O2 release for the transformation under air of Co(NO3)2 into Co3O4 (3 Co(NO3)2 $ Co3O4 þ 6NO2 þ O2) is detrimental to oxide particle dispersion and that similar O2 release is not in the equilibrium balance during the transformation of Co(OH)x(NO3)y into Co3O4. De Jong et al. [21] further explained the role of the hydroxynitrate as an immobile phase preventing redistribution of the pristine Co(H2O)6(NO3)2 species and yielding small cobalt oxide nanoparticles by its further decomposition at higher temperature. Unfortunately for the catalytic activity, smaller Co3O4 particles (8.4 nm) obtained upon NO-assisted calcination are more difficult to reduce than larger (10.7 nm) ones prepared by calcination under air [20]. This behavior, observed in many studies involving NO-assisted calcination [54], was ascribed to a stronger interaction of smaller cobalt

A. R. Passos et al.

oxide particles with the support leading to the formation of cobalt silicate-like species. The use of noble metal promoters (Ru, Pt, Re, and so on) for preparing cobalt FT catalysts has been also the subject of intense researches aiming to relate the noble metal reduction to cobalt reducibility improvement [19]. The element selectivity of XAS can be used to determine not only the local environment of cobalt but also of the promoter. In a comprehensive study combining Co and Ru K-edge quick-XAS monitoring with MCR-ALS analysis, the impact of Ru promoter on the active Co(0) site density has been evaluated for CoRu/SiO2 catalysts prepared in the presence of sorbitol [33]. The transformation of Co3O4 into CoO under H2 is characterized by an increase of the white line intensity of the XANES spectra associated to the conversion of the tetrahedral cobalt sites in the spinel-like Co3O4 into octahedral ones in CoO. This transformation occurs at significantly lower temperatures for CoRu-sorbitol/SiO2 oxidic precursors (Fig. 29.12b) compared to non-promoted ones (Fig. 29.12a) in total agreement with the TPR profiles (Fig. 29.12d) [18, 20]. But more remarkably is the quasi absence of Co (0) species detected at 500  C for the non-promoted sample compared to the promoted one (Fig. 29.12a, b). This is ascribed to the formation of cobalt silicate species which are reduced at high temperatures, as displayed in Fig. 29.12d with the peak at 855  C, whereas the strong peak of the TPR profile around 369  C for the CoRu-sorbitol/SiO2 catalysts is ascribed to the complete reduction of CoO into Co(0) observed by XAS. The MCR-ALS analysis of the Ru K-edge data reveals also a two-stage mechanism (Fig. 29.12e). First, the initial oxidic species obtained after calcination, identified by EXAFS fitting as characteristic of a Ru(IV) ion substituting Co(III) ion in an octahedral cationic site of the spinel-like Co3O4 species, is transformed into a Ru (III) ion in substitution of Co(II) in the NaCl structure of CoO. In a second step, the reduction to the metallic state takes place from 150 to 500  C and yields Ru embedded in a bimetallic Co-Ru nanoparticle, as revealed by EXAFS fitting. The size of the metallic nanoparticles determined by TEM emphasizes the beneficial impact of sorbitol for enhancing the dispersion of the nanoparticles with sizes lower than 5 nm compared to the CoRu/SiO2 catalysts with 25% of the nanoparticles of sizes up to 12 nm. The TPR and Co K-edge quick-XAS monitoring highlight also the beneficial role of the promoter for improving the reduction of cobalt and thus leading to a higher density of active particles. The Ru K-edge quick-XAS characterization of the promoted catalysts rationalizes the synergetic effect between the reduction steps of both metals since they occur at approximately the same temperatures due to the close vicinity of Ru inside the cobalt oxidic network, first at the end of the calcination stage into Co3O4 then as belonging to the CoO network and finally through an autocatalytic process forming bimetallic CoRu particles.

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

2.0 1.5 1.0

160 °C

0.5

b1)

11 9 7 5 160 °C

1 240 210 180 150 120 90 60 30

7720

7730

7740

0

1

2

Photon energy (eV)

3

4

R, Angstrom

b2) 2.0 1.5

110 °C

1.0 200 °C

0.5

11

Norm. absorbance (a.u.)

a2) Calcination under 5% NO/He

9 7 200 °C

1

era

per

atu

ture

re (

(°C

)

°C)

240 210 180 150 120 90 60 30

7720

7730

7740

0

Photon energy (eV)

c1)

Co(H2O)6(NO3)2

Tem p

Tem

7710

1

2

3

4

R, Angstrom Co(NO3)2

Co3O4

c2) Co(H2O)6(NO3)2

100

Co(OH)x(NO3)y

Co3O4

100

Percentage of cobalt species

Percentage of cobalt species

3

110 °C

0.0 30 60 90 120 150 180 210 240 7700

5

k3-weighted FTmoduli

7710

Tem per atu re (

Tem per atu re (

°C)

°C)

0.0 30 60 90 120 150 180 210 240 7700

3

k3-weighted FTmoduli

a1) Calcination under air

641

Norm. absorbance (a.u.)

29

80 60 40 20 0

80 60

Co(NO3)2 40 20 0

40

80

120 160 Temperature (°C)

200

Fig. 29.11 Co K-edge quick-XAS monitoring of the thermal activation of a Co/SiO2 FT catalyst with heating from RT to 240  C: (a1 to c1) under air and (a2 to c2) under 5% NO/He with (a) XANES spectra, (b)

40

80

120 160 Temperature (°C)

200

EXAFS k3-weighted Fourier transform moduli, and (c) cobalt-phase speciation determined by LCF of the TR-XAS spectra. (Reprinted with permission from Ref. [20]. Copyright (2012) Wiley)

29

642

A. R. Passos et al.

a)

d) 244

1.5

1

0.5

0 100

5 0

200

400

600

300 400 500

T (°C)

190

CoRu-sorb/SiO2

7800 7820 7760 7780 7740 7700 7720 Energy (eV)

b)

0

400

200

600

800

T (°C)

1

CoRu-sorbitol/SiO2 Co K edge

1.5

800

369

5

200

CoO

Co3O4

Co

0.8

Molar fraction of Co

Normalized absorbance

Co-sorb/SiO2 855

TPR H2 consumption (Pmol H2/(°C.g cat))

Normalized absorbance

Co-sorbitol/SiO2 Co K edge

1

0.5

0.6

0.4

0.2

0

Isothermal

100 0

200 300 400 500

T (°C)

7780 7800

0

100

200

300 400 T (°C)

500

Ru K edge Ru/Co

Ru(IV)Co3O4 1 1 0.8

Molar fraction of Ru

Normalized absorbance

c)

7700

7760 7720 7740 Energy (eV)

7820

0.5

0 100

Ru(III)/CoO 0.6

0.4

0.2

200 300 400 T (°C)

500

22200 22220 22160 22180 22140 22100 22120 Energy (eV)

Fig. 29.12 Quick-XAS monitoring of the reduction (a) at the Co K-edge of Co-sorbitol/SiO2 and (b) at the Co and (c) Ru K-edges of CoRu-sorbitol/SiO2 and related MCR-ALS speciations. (d) TPR profiles of Co-sorbitol/SiO2 and CoRu-sorbitol/SiO2. (e) Representation of

0 0

100

200 300 T (°C)

400

500

the Ru species involved during the reduction of CoRu-sorbitol/SiO2 identified by EXAFS fitting of the different Ru MCR-ALS components: red, ruthenium; black, cobalt; gray, oxygen. (Reprinted with permission from Ref. [33]. Copyright (2015) American Chemical Society)

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

643

e)

After oxidative activation

After 1st step of reduction

After 2nd step of reduction

Fig. 29.12 (continued)

Metal Distribution: Formation of Undesirable Phases and Interaction with the Support The formation of phases involving one of the chemical elements of the support at the impregnation step could have strong impact in the further catalyst preparation stages, such as calcination or activation. For instance, the Anderson-like heteropolyanion (HPA) Al(OH)6Mo6O183 is a well-known species reported during the preparation of alumina-supported Mo catalysts used in hydrodesulfurization (HDS) process for the removal of sulfur in crude oils [34]. The formation of this HPA observed for high Mo loading solutions results from the support lixiviation due to the solution acidity and further reactions with soluble Mo species during the maturation in solution and drying stages. The impact of its formation has been described in the literature as leading to MoO3 upon calcination, which is detrimental for the dispersion of the further active MoS2 phase formed under sulfiding atmosphere. The formation of the Anderson-like HPA species can be also observed upon further exposure under moisture conditions of calcined catalysts, as illustrated in Fig. 29.13 [55, 56]. Metal Distribution Within Bimetallic Particles Besides the addition of promoters for improving the formation of active species, the use of bimetallic systems is also frequently considered. Taking advantage of an interplay of electronic and geometric characteristics of both metals, the catalytic properties are improved. The knowledge and mastering of the atomic distribution of the two metals inside those systems are of paramount importance for the design of efficient catalysts. With selected examples, we will examine herein the use of TR-XAS for highlighting processes leading to the formation of bimetallic particles and to composition adjustments passing from core-shell structures to alloys and vice versa. The thermal decomposition of LDH-like compounds is an effective synthesis route for the preparation of highly dispersed

multimetallic catalysts. Quick-XAS monitoring combined with MCR-ALS analysis has been used for unravelling and quantifying the intermediate species involved during the transformation of NiCuAl LDH-based oxidic precursors into NiCu/ Al2O3 catalyst [37]. The use of a direct activation route, starting from the as-prepared NiCuAl LDH-based oxidic gel and involving an oxidative treatment up to 210  C followed by reduction under H2 up to 500  C, displays very different speciation patterns for nickel and copper compared to conventional activation [31], as illustrated in Fig. 29.14. During the reductive step of the direct activation, the nickel and copper phases are completely transformed into metallic species before reaching 500  C at the opposite of the conventional route where only 75% of Ni(0) is formed after 3 h of H2 treatment at 500  C. As the onset of reduction of the oxidic nickel phases formed upon LDH oxidative decomposition is coincident with the completion of Cu reduction, it was assumed that Cu(0) catalyzes the decomposition of adsorbed H2, supplying active H species for promoting the reduction of nickel. The comparison with the monometallic Ni LDH gel evidences a decrease of the temperature for complete nickel reduction, from 495  C in the monometallic sample to 450  C for the bimetallic one. This was explained by a heterogeneous nucleation of Ni(0) particles over the Cu(0) ones early formed, leading to the formation of NiCu alloyed particles [31, 37]. The emergence of environmental TEM cells provides unique opportunities for imaging at the atomic level the particle composition and the dynamic restructuration of bimetallic nanoparticles under reaction conditions [52]. However, information afforded by TR-XAS still remains very complementary to the ones obtained by TEM. The element sensitivity of XAS offers the clear advantage of local-order information around each of the elements composing the materials. It provides a complete description of degree of element mixing and oxidation state of the atoms. Furthermore, the so-obtained information are averaged over a larger field of view than TEM and consequently over a large

29

644

A. R. Passos et al.

a) 6

b)

5 AIMo6

3

AIMo6

Intensity (a.u.)

FFT (A–4)

4

40 h 18h 9h

2

40 h

4h

18 h

0h

9h

1

4h *

*

0h *

0 0

1

2

3

4

5

6

R (A)

200

300

400

500

600

*

* 700

800

900

1000

Raman shift (cm–1)

Fig. 29.13 Simultaneous monitoring by Mo K-edge XAS (a) and Raman spectroscopy (b) of the RT rehydration of a calcined 5 wt. % Mo catalyst supported on amorphous silica-alumina. The spectra of the

Al(OH)6Mo6O183 ion, labeled AlMo6, are presented for comparison. Stars in (b) indicate the position of characteristic MoO3 lines. (Reprinted with permission from Ref. [55]. Copyright (2005) IOP Publishing)

number of multimetallic nanoparticles, providing some advantages such as a higher statistical relevance of the observation but also mitigation to a certain extent of artifacts induced by electron beams for sensitive materials. Quick-XAS was applied to unravel, with one-second time resolution, the reaction mechanism of the dynamic restructuring of Pt13In9 nanocrystals during high-temperature O2-H2 redox cycling [57]. Similar redox cycling treatments are used for reactivating dehydrogenation Pt-In nanocatalysts after coking in high-temperature alkane flows. In situ Pt L3-edge quick-XAS measurements were used to probe the Pt oxidation state during nanoalloy segregation dynamics induced by the repetitive H2-O2 redox cycling. The reversible change of the white line (WL) intensity and of the energy position of the WL maximum upon cycling (Fig. 29.15a) is interpreted, by comparison with the XANES spectra of references (Fig. 29.15b), as being related to the reversible transformation between a reduced state, in which Pt is alloyed with In, and an oxidized state where PtOx is formed. This interpretation is clearly deduced from the analysis of the wavelet-transformed quick-EXAFS magnitude measured at specific points of a cycle (Fig. 29.15c) where at the end of the H2 pulse, Pt is surrounded by In and Pt first neighbors in a Pt-In alloy. Upon O2 exposure, the Pt-In alloy has decomposed by segregation into a PtOx phase in which Pt is surrounded by O in the first coordination shell and Pt as

second nearest neighbors. The segregation is fully confirmed by microscopy (Fig. 29.15e, f) showing a metallic Pt core particle surrounded by an In2O3 crust. It is noteworthy that the WL intensity characteristic of the PtOx species significantly decreases upon successive cycles as the result of the Pt nanoparticle sintering. Namely, electron microscopy highlights the particle growth from ca. 2 nm at the end of the first cycle (Fig. 29.15d) to ca. 10 nm after 60 cycles (Fig. 29.15g). The in-depth analysis of the results performed at different temperatures upon cycling allowed the identification of the reaction mechanism describing the dynamic restructuring of Pt-In nanocatalysts and the calculation of activation energy of alloying and dealloying processes.

29.3.2 Catalysts in Operation: From Active Phases to Spent Catalysts The ability to visualize a catalytic transformation in real time under real conditions using TR-XAS has motivated studies of many catalytic reactions due to the potential for unprecedented understanding of reaction mechanisms. The structure-performance relationship can be either indirectly established by comparison of in situ XAS measurements, i.e., under gas, temperature, and pressure, with catalytic activity measured independently or directly via TR-XAS

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

Conventional activation 1.6

1.6

1.4

1.4 1.2 1 0.8 0.6 0.4 Ni K edge 0 8320 8340 8360 8380 8400 8420 8440 Energy (eV)

1.2 1 0.8 0.6 0.4

80 70 60

NiO Ni(0) CuO Cu2O Cu(0)

50 40 30 20 10

Cu K edge 0 8960 8980 9000 9020 9040 9060 9080 Energy (eV)

0

0

1.6

1.6

1.4

90

1.2

80

1.4 1.2 1 0.8 0.6 0.4 0.2

Ni K edge 0 8320 8340 8360 8380 8400 8420 8440 Energy (eV)

100

200 300 400 Temperature (°C)

500

Direct activation

1.8

Normalized absorbance

Normalized absorbance

90

0.2

0.2

b)

100

Species percentage (%)

1.8

Normalized absorbance

Normalized absorbance

a)

645

1 0.8 0.6 0.4

100

Species percentage (%)

29

0.2 Cu K edge 0 8960 8980 9000 9020 9040 9060 9080 Energy (eV)

70

Ni-LDH NiOx Ni(0) Cu-LDH CuCl2 Atacamite Cu(0)

60 50 40 30 20 10 0 100

200 300 400 Temperature (°C)

500

Fig. 29.14 Simultaneous Ni and Cu K-edge quick-XAS monitoring of the activation of a NiCu/Al2O3 oxidic precursor derived from LDH. (a) Conventional activation, carried out after stagnant air calcination, under H2 from RT to 500  C. The MCR-ALS speciation revealed the transformation of NiO into Ni(0) and a two-stage transformation for CuO into Cu2O and Cu(0) [31]. (b) Direct activation route. Due to the presence of chloride anions as charge compensating of the positively charged LDH

layers, CuCl2- and atacamite-like species, first formed at the early stage of the oxidative treatment, are promptly reduced into Cu(0) species at 300  C, whereas Cu-depleted LDH is transformed into a first oxidic nickel species, which is a mixture of NiO and NiAl2O4 evolving with a similar formation rate before being reduced into Ni(0) species at the completion of Cu reduction. (Reprinted with permission from Ref. [37]. Copyright (2020) The Royal Society of Chemistry)

and simultaneous monitoring of the products of the catalytic reaction by an analytic method, through the operando approach. This section is focused on a few recent studies which illustrate the insights gained by TR-XAS into the modification of catalyst structure and composition in relation to the catalytic performances. Studies carried out in order to optimize the process conditions regarding catalytic activity will be distinguished to those specifically performed to better understand the origin of deactivation.

to methanol (DMTM) conversion. DMTM is considered as a potential alternative route to the conventional and costly indirect ones, involving steam reforming of methane to produce syngas further converted into methanol by hydrogenation. As reviewed in [58] and [41], considerable debates about the basic mechanisms and structure of active sites in those systems still exist. However, those reviews agree on the importance of TR-XAS to provide dynamical information on the structural and electronic changes impacting copper atoms in Cu-based zeolites during the DMTM conversion cycling process. For instance, Lomachenko et al. [59] tested on Cu mordenite (Cu-MOR) catalysts different O2 activation and CH4 loading duration before performing the last step of cycling dealing with the steam-assisted CH3OH extraction. For the 0.18Cu-HMOR(7) composition presented in Fig. 29.16a–d, it was observed that the abundance of the Cu (I) species is strongly dependent of the O2 protocol duration

Insights of Catalyst Structure and Composition vs Activity: A Lever for Process Optimization Because of major concerns in relation to climate change, the mitigation of greenhouse gases in the atmosphere has been the subject of intense researches in the last decade. In particular, an abundant literature can be found about the nature of active sites in Cu-based zeolites used for the direct methane

29

646

A. R. Passos et al.

H2 O2 0

10

20 30 40 50 # Redox cycles

d)

60

2 PtO2

Pt

0 s in H2

60 s in O2 3

4

4 s in H2 5

Pt

In

O

k (Å–1) 11 0 k (Å–1) 11 0 k (Å–1) 11 0 k (Å–1) 11 0 k (Å–1) 11

0 11560 11570 11580 X-ray energy (eV)

2 nm Pt [011]

10 nm

2

Pt

Alloying

f) 10 nm

In

Pt13In9

1

0 s in O2

1

1 0

g)

e)

c) 3.5 116 s in H2

Flow

O2 H2

XANES reference spectra

R (Å)

b)

773 K 723 K

WL energy (eV) WL height (–) 11 11 1. 1. 56 57 30 15 8 0

823 K

Oxidation

WL energy (eV) WL height (–) 11 11 1. 1. 56 57 15 55 0 7

923 K 873 K

Normalized absorption (–)

a)

10 nm

H2

O2

H2

60 sec

120 sec

60 sec

Time

Fig. 29.15 Pt L3-edge quick-XAS monitoring of the H2-O2 redox cycling performed at variable temperatures to induce structural changes in the Pt-In nanocatalysts: (a) WL intensity and energy position of the WL maximum as a function of the number of H2-O2 redox cycles, (b) XANES reference spectra, (c) zoom of the WL intensity and WL energy position within the 21st redox cycle. On the top, Pt L3-edge wavelet-

transformed quick-EXAFS magnitude maps at specific points of the 21st redox cycle. The plots reveal the element type (k-axis) and position (R-axis) of neighbors around Pt, namely, In, O, or Pt. (d to g) Electron microscopy images of Pt-In nanoparticles after the 1st (d and e) and the 60th (f and g) redox cycles. (Reprinted with permission from Ref. [57]. Copyright (2018) Wiley VCH)

with more than 60% at the onset of methanol extraction for long protocol (LP). Rapidly, after CH4 introduction, the fraction of Cu(I) species in interaction with the zeolite is stabilized. This behavior indicates a saturation by CH4 of the Cu (II) active species interacting with the zeolite framework formed during the O2 activation at a level which is nearly two times larger for LP than for SP (short protocol). Adding the results obtained for the 0.36Cu-HMOR(11), a clear linear correlation between the methanol productivity per Cu versus the fraction of Cu(I) determined at the end of the CH4 loading step was established (Fig. 29.16e): the higher the Cu (I) fraction, the better the performance. This study highlights the correlation between Cu speciation and performances thanks to a spectroscopic descriptor of the CH3OH productivity per Cu represented by the Cu(I) fraction formed after CH4 exposure. It is noteworthy that the full description of Cu speciation during activation required to go beyond the description gained by the common LCF method from XANES of standard compounds using the CWA-MCR-ALS strategy detailed in Sect. 29.2.5, with the analysis of a data set built from the matrix concatenation of several high-energy resolution fluorescence detected (HERFD) XANES data sets collected on the same zeolite systems activated under O2 and He but with a higher-energy resolution [41]. This strategy revealed that the Cu(II)-MOR reference used for LCF is actually a mixture of three different Cu(II) intermediates, which are characterized in each individual data set by close

HERFD-XANES spectra and nearly correlated concentration profiles during activation. The understanding of the dynamic evolution under reaction conditions of promising single-atom catalysts (SACs), considering the relationship between ultimate active site dispersion and catalytic efficiency, is of prime importance for the design of industrially relevant catalysts. Furthermore, when catalysts involve precious metals, like in many industrial applications, e.g., petrochemistry, the decrease of the catalyst metal loading is a parameter of importance due to its impact on cost production. This quest of highly efficient lowly loaded noble metal catalysts motivates many studies involving operando/in situ characterization of the archetype catalytic CO oxidation reaction [60, 61]. Recently, the operando monitoring of the thermal treatments on Pt/Al2O3 catalysts used for CO oxidation (Fig. 29.17) was discussed at the light of TR-XAS and DRIFTS results obtained under cycling reaction conditions with CO conversion measured simultaneously by MS [60]. From TR-XAS, it was concluded that the oxidized single Ptm+ atoms (m > 2) resulting from impregnation-calcination procedure were stable under O2 heating and presented a moderate activity for CO oxidation with a half-conversion temperature around 230  C (Fig. 29.17c). Upon cooling under reaction conditions, a clustering of SACs occurred. It is revealed by both TR-XAS (appearance of Pt-Pt contribution at the same time than Pt-O ones and decrease of the WL intensity) and DRIFTS (signature

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

647

O2 activation at 500 °C

CH4 loading at 200 °C

CH3OH extraction at 200 °C

Short protocol

90 min

120 min

60 min

Long protocol

480 min

360 min

60 min

He flush before CH4 > CH4 int. > CH4 end > He flush after CH4 > H2O int. > H2O end

a)

c)

0.2

e) 1.0

0

mol CH3OH/mol Cu

0.8

0

(m

(m

e Tim

e Tim

~ 120

0.18CuHMOR(7), LP

in)

in)

0.18CuHMOR(7), SP

~ 360

~ 240

R-factor

0.4 0.2

0.0 0

30

60

90 120 150 180 210 240 Time (min)

– 9x

y=

0.4

29

0.1

0.

C A

B

10

00

0.2

90

90

0.0

0.18Cu-HMOR(7), LP

1.0

CH4

He H2O

0.4 0.6 0.8 (mol Cu(I))max/mol Cu

1.0

6.0×10–4

0.8

Cu(II)-hyd

0.6

3.0×10–4 R-factor

H2O

3.0×10–4

0.0

90

80

d) 6.0×10–4

0.8 0.6

D

Energy (eV)

Fraction Cu-species

He

89

89

10

00

90

90

90

80

89

89 Fraction Cu-species

CH4

x

0.0

0.18Cu-HMOR(7), SP 1.0

y=

4

0.6

0.2

~ 480 Energy (eV)

b)

A. 0.36Cu-HMOR(11) SP B. 0.18Cu-HMOR(7) SP C. 0.36Cu-HMOR(11) LP D. 0.18Cu-HMOR(7) LP

0.4 0.2

0.0

0.0 0

Cu(II)-MOR Cu(I)-MOR LCF R-factor

60 120 180 240 300 360 420 480 Time (min)

Fig. 29.16 Cu K-edge time-resolved operando XANES recorded for Cu-MOR zeolite (Cu/Al ¼ 0.18 and Si/Al ¼ 7) during CH4 loading and subsequent CH3OH extraction (a, b) for the SP and (c, d) for the LP processes. (b) and (d) display the fraction of Cu species during the DMTM reaction obtained by LCF using as references Cu(II)-hydrated,

Cu(II)-MOR, and Cu(I)-MOR. (e) Correlation of the normalized productivity to the total fraction of Cu(I) per Cu after methane loading for the experiments presented for Cu/Al ¼ 0.18 and Si/Al ¼ 7 and for Cu/Al ¼ 0.36 and Si/Al ¼ 11. (Adapted with permission from Ref. [59]. Copyright (2019) Elsevier)

of CO adsorption on oxidized clusters together with remaining fingerprint of CO linearly adsorbed on Ptm+ atoms). The formation of partially oxidized Ptδ+ clusters (δ < 2), with size ≈ 1 nm, was observed in a high CO conversion temperature window suggesting outperformance of Ptδ+ clusters for CO oxidation than single Ptm+ atoms (m > 2). The hypothesis was fully confirmed by the second part of the experiments dealing with heating under H2, during which reduced Pt clusters were mostly formed (Fig. 29.17a, d). Upon further exposure to CO/O2 atmosphere, those reduced clusters were found stable until reaching the light-off temperature around 150  C (Fig. 29.17c) at which a sudden disappearance of the

2070–2050 cm1 DRIFTS band ascribed to Pt(0) clusters and a Pt-O contribution and increase of WL intensity characteristic of oxidation were observed. Actually, the 100% CO conversion observed at this stage led to a highly oxidative atmosphere in the catalytic bed inducing the formation of amorphous PtOx clusters. This set of results, in particular with the significant decrease of the light-off temperature after reductive treatment favoring cluster formation (Fig. 29.17c), clearly highlights that clusters are more active than SACs for CO oxidation. Those examples illustrate how powerful is TR-XAS operando characterization to assess the structure-activity relationships. They also emphasize that its combination

648

a) Pt – 0

RT 100

EXAFS Pt – Pt

100 200 100% 300 conv. 200 100

200 100

H2

RT 2200

200 100

2150 2100 2050 2000 Wavenumber (cm–1)

1950

RT 100 200 100% 300 conv. 200 100

200 300 200 100 RT 300

RT 2200

200

2150 2100 2050 2000 Wavenumber (cm–1)

c) 100

100 RT 0.5

0

1.0

1.5 2.0 2.5 3.0 Radial distance (Å)

3.5

4.0

Pt(0) Pt(II) Pt(IV) 2.5 White line intensity (absorbance)

CO2 yield (%)

CO+O2

300

RT 100

CO+O2

DRIFTS 0 Pt1m+ Ptd+ n Ptn

RT

RT 300

80 60 40 20 0 50

XANES white line intensity (absorbance) EXAFS Pt-Pt coordination number

2.5

O2

CO+O2

H2

100 150 200 250 Temperature (°C)

CO+O2

a

2.0

1.5 0.3Pt COOX10 1.0Pt COOX10 0.3Pt COOX2 b 6 4 2

0 300 °C

RT

100% CO conversion

100% CO conversion Time

1950

After calcination After reduction

d)

Temp.

O2

b)

XANES

200

Time

Fig. 29.17 Operando Pt L3 TR-XAS and DRIFTS monitoring of CO oxidation on a Pt/γ-Al2O3 catalyst, throughout the calcination/reaction/reduction/ reaction procedure. (a) EXAFS color map showing the temporal evolution of the phase-corrected FT with two FT signals shown in classical view. The white points represent the XANES WL intensity. (b) DRIFTS color maps showing the υ(CO) absorption band evolution and post-reduction reaction steps. Three spectra corresponding to the horizontal lines (at 150  C during heating and 100  C during cooling) on the DRIFTS maps are plotted in classical view. (c) CO oxidation light-off and extinction curves. (d) Evolution of XANES WL intensities and Pt-Pt coordination numbers derived from EXAFS fitting for the catalyst monitored in (a) and for other Pt loadings and CO:O2 gas composition [60]. (Copyright (2019) American Chemical Society)

A. R. Passos et al.

300

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

with state-of-the-art multivariate analysis or complementary techniques is mandatory for unravelling the complexity of the dynamic transformations of working catalysts. Similar strategy involving operando XAS-XRD monitoring assisted by MCR-ALS was recently used to understand the driving force leading to catalyst stabilization and further deactivation during methanol synthesis by CO2 hydrogenation under pressure (Fig. 29.18) [17]. In agreement with the In-O and In-In coordination numbers derived from EXAFS fitting and XANES rising-edge energy position, the MCR-ALS analysis isolates during the reaction process three phases involving the pristine In2O3 nanoparticles, an In2O3-x phase resulting from the formation of oxygen vacancy in In2O3 and a molten In (0) species. The comparison of the catalytic activity with the MCR-ALS profiles associated to those species highlights that the active species for CO2 hydrogenation into CH3OH is the

649

oxygen-vacant In2O3-x phase, accounting for about 60% of the indium speciation at the reaction stage where outperformance is observed. The onset of increase of In(0) proportion corresponding to the reductive amorphization of In2O3 observed by XRD is linked to the gradual deactivation observed upon time on stream (TOS). The higher the proportion of In(0), the lower the crystallinity of the bcc-In2O3 crystallites, the lower the catalytic activity. It is noteworthy that upon activation or deactivation stages, no sintering of the In2O3 phase has been revealed by XRD. As the crystallinity and particle sizes remain nearly constant during the activation stage and only a slight crystallite size decrease is observed during the deactivation stage, it can be concluded that the change of indium speciation is mainly responsible for the catalyst stabilization and further deactivation.

a)

b) g(MeOH) g(In2O3)–1 h–1

1.2

In-O 1.2 27954

0.4

le ab ge D n a ch XR Ex AS/ X

0.0

27990

6

In2O3 catalyst

0.80 0.85 0.90 0.95 1.00 1.05 1.10 Time (min)

It

60

80

100

120

140 200

0.2 0.0 In-O In-In

3.6

*

* BN bcc-In2O3 * (440)

(400) (411)

* (200)

9 10 11 12 13 14 15 16 2 q (°)

Fig. 29.18 (a) Setup used for operando In K-edge TR-XAS and XRD experiments performed for a model In2O3 catalyst under CO2 hydrogenation conditions (300  C, 20 bar) with presentation of selected TR-XANES and Fourier transforms of EXAFS signals, selected XRD patterns (λ ¼ 0.506 Å) with indexed reflection peaks for bcc-In2O3, and selected chromatographs showing the methanol amounts with TOS. The arrows indicate the direction of changes with TOS. (b) Summary of the

In2O3

0.0

InCl2

–1.5

In0

–3.0 1.2

XAS detector

(222) Intensity (a.u.)

MeOH

40

0.0

XRD

GC FID+TCD

20

Deactivation

1.8

Capillary reactor

Gas outlet

0

Stable performance

5.4 2 4 Radial distance (Å)

0

H2, CO2

Monochromators Ion chamber GC

0.4

27945

300 °C 20 bar I0

0.6

CN

27940

27960 Energy (eV)

0.4

'E0 (eV)

27935

27930

In-In

In-out

0.0

0.8

Activation

MCR fraction

27948

XRD detector

0.8

|F(R)| (Å–3)

Normalized P (a.u.)

TR-XAS

C-1: In2O3

C-2: In2O3-x

C-3: In0

0.8 0.4 0.0

Crystallinity (%)

29

100 80 60 40 20

7.5 7.0 6.5 0

20

40

60 80 100 TOS (min)

120

140 200

results gained upon TOS by GC for methanol production; by EXAFS with the fitted In-O and In-In coordination numbers (CN), XANES edge energy shift (ΔE0) with respect to In2O3 reference, and MCR-ALS concentration profiles; and by XRD with the average crystallite size and fraction of crystalline bcc-In2O3. (Reprinted with permission from Ref. [17]. Copyright (2019) American Chemical Society)

29

650

A. R. Passos et al.

Catalyst Deactivation and Regeneration Besides the in-depth identification of active sites leading to the best performances, operando TR-XAS provides also unique opportunity to address the deactivation topic since, at the same time, the technique provides the monitoring of the evolution of the active site structure and of catalytic activity. Bartholomew et al. [62] reviewed the mechanisms of catalyst deactivation. Some of them can be readily characterized by XAS insofar as they affect the same catalyst characteristics than those already evaluated by XAS during activation studies detailed in section “Synthesis of Supported Catalysts”, such as active-phase dispersion, metal distribution, or catalyst composition. In this last section, we will examine some deactivation mechanisms with the emphasis to propose regeneration strategies. Particle sintering, which reduces the active site dispersion, is reported using TR-XAS as the main deactivation mechanism for supported Pt catalysts used in the harsh temperature conditions of automotive converters with high temperature and fluctuation of exhaust gas composition between oxidative and reductive conditions [63]. Redistribution of atoms in bimetallic particles, which changes the active site nature, has been also identified by TR-XAS for Pd-Pt/Al2O3 catalyst deactivation during total oxidation of methane, with segregation in oxidized state of Pt and Pd in core-shell bimetallic particles [64]. Another deactivation route well characterized by TR-XAS is the change in the oxidation state of the active phase induced by the presence of reactive gases in the feed, some of them being present as contaminants [65], whereas others as result of the catalytic reaction [19, 66]. For the latter case, many publications aimed to

a)

study the fast oscillatory behaviors induced by peculiar reaction conditions, leading to bistability of the catalyst structure, passing from active to inactive states driven by changes in the oxidation of the metal [66, 67]. Understanding catalyst evolution under oscillatory conditions is of prime importance for the development of diesel oxidation catalysts, which suffer similar ignition and extinction conditions. For instance, TR-XAS monitoring of the catalytic partial oxidation of methane over Pd/Al2O3 catalysts is presented in Fig. 29.19a [66]. The in-depth characterization of the Pd oxidation state along the catalyst bed together with the reactant conversion measured by MS at the reactor outlet allowed for proposing the scheme shown in Fig. 29.19b. It explains the cycling behavior of the catalyst upon reaction conditions, based on four stages: (1) PdOx first promotes the complete CH4 oxidation; (2) upon total O2 consumption, catalytic partial oxidation occurs leading to the reduction of PdOx into Pd (0) from the reactor outlet; (3) propagation of the reduction occurs toward the inlet driven by temperature hot spots; (4) when Pd is fully reduced, deactivation occurs dropping the CH4 oxidation activity leading to oxidative conditions and reoxidation of Pd(0). The poisoning by contaminants leads to active site occupation by extraneous species blocking the reactant access. Deactivation and regeneration processes are often addressed by operando steady-state XAS characterization, e.g., with sulfur poisoning of noble metal catalysts used for methane oxidation under typical conditions of lean burn gas engines [68] or for methanation from dry biomass [69]. However, experiments carried out under high frequency of gaseous pulses of contaminants or regenerative reactants can require TR-XAS, as illustrated with the use of modulation excitation

b) 1

1.15

Pd Al2O3

CO2, H2O

CH4, O2

1.10 Absorption

PdO

Extinction

Ignition

1.05 4 Al2O3

Al2O3

CH4, O2

0.90

s)

e( m Ti

600 500 400 300 200 100

PdO

Pd

Pd

0.95

2

PdCx

PdO

1.00

Al2O3

CO2, H2O

Pd

Pd

Al2O3

Al2O3

CH4, O2

H2, CO 24600 24550 24500 24450 24400 ) (eV 24350 0 rgy e 24300 n E

3

Deactivation

Fig. 29.19 (a) Pd K-edge TR-XAS monitoring of oscillations with a Pd/Al2O3 catalyst loaded in a fixed bed catalyst and heated at 410  C under 6% CH4 : 3% O2:He gas mixture. (b) Suggested scheme for the

CH4, O2

CO2, H2O H2, CO

Pd Al2O3

CO2, H2O H2, C`O

Formation of hot spot & autoreduction

oscillatory behavior of the catalytic partial oxidation (CPO) of CH4. (Reprinted with permission from Ref. [66]. Copyright (2012) American Chemical Society)

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

spectroscopy (MES) [70] for studying the stability of catalysts for hydrogenation of ethylene glycol. As already mentioned, MoS2-based catalysts find applications for HDS treatment. Their optimization makes a great use of TR-XAS for understanding the impact of the sulfiding activation processes [6, 12, 29, 36] on Mo and promotor (Ni, Co) speciation. The use of Mo-based catalysts for biomass conversion into upgraded bio-oil by catalytic hydrodeoxygenation (HDO) requires to optimize their tolerance against water and to develop strategies for their stabilization in oxygenrich feedstock (20–50 wt. %). The catalytic active sites being located at the corners and edges of MoS2 slabs only constitute a small fraction of all the Mo atoms in the slabs letting the bulk XAS technique mainly dominated by the contribution of the Mo spectator atoms which are not involved in the reaction. To overcome the poor sensitivity of XAS toward minority species, MES can be used to separate the spectroscopic answer of active species from spectator ones [71]. The principle is to periodically (and reversibly) perturbate the catalytic system by external stimulation, such as the alternate use of different gaseous pulses, which only influences the active species. The further mathematical treatment, called demodulation or phase-sensitive detection, allows for isolating only signals of the active species, which respond to the external simulation with the same frequency, and their further identification. MES allows also a better differentiation of signals with different kinetics. Readers are invited to refer to the dedicated literature for a deep understanding of the mathematical treatment [71, 72]. In order to understand the influence of H2O and H2S partial pressures on the stability of active MoS2 slab edges used for HDO, MES experiments were carried out by cycling the catalysts to pulses of H2O/H2 and H2S/H2 [70]. Figure 29.20a shows that no modification of the Mo K-edge XANES data can be observed during pulse cycling. The match of the demodulated spectra with the difference between the XANES spectra of MoS2 and MoO3 references (Fig. 29.20b) demonstrates a reversible oxidationsulfidation process in which 1% of Mo(IV) atoms undergo oxidation to Mo(VI) with the replacement of S atoms at the edge of MoS2 by O atoms. Similar behaviors are reported for MES of Co- and Ni-promoted MoS2 catalysts (Fig. 29.20c), however with a reduced amplitude of the demodulated spectra at the energy position P centered on the shoulder H characteristic of MoS2, suggesting a decrease in S-O exchange upon promotion. Further information on the kinetics of oxidation-sulfidation can be obtained by plotting the variation of the shoulder H intensity with time (Fig. 29.20d). The decrease of absorption during the H2O pulse and its increase during the H2S pulse are delayed by about 30 s for CoMo compared to Mo (not shown) or NiMo. Mo and CoMo catalysts suffer from oxidation-sulfidation at the first period of pulses, whereas nothing is happening for the NiMo one

651

before the fifth period. All these results emphasize a different response of the promoted catalysts compared to the non-promoted one with as general conclusion an improvement of the stability of the promoted catalyst against H2O exposure explaining their better activity for ethylene glycol HDO and the benefit of small quantity of H2S to regenerate the catalysts at each cycle. Besides, the fruitful combination of TR-XAS with Raman spectroscopy can address the impact on catalyst deactivation of carbonaceous deposits. Raman spectroscopy enables the specific characterization of the carbon species at the surface of the catalysts under reaction conditions and highlights the onset of coke formation inside the catalytic bed [31, 37]. Recently, the reaction, deactivation, and regeneration of a Co/Al2O3 catalyst used for ethanol steam reforming (ESR) have been studied through operando monitoring by TR-XAS, MS, and Raman spectroscopy. Figure 29.21 displays the main results gained during the monitoring of successive ESR and oxidative conditions aiming to regenerate the catalyst after coking. Figure 29.21c evidences that the decrease of ethanol conversion during the first ESR reaction is coincident with the increase of coke deposits at the surface of the catalyst, for which the abundance is measured by the Raman G-line intensity. MCR-ALS analysis of the TR-XAS data recorded simultaneously with Raman shows nearly invariant proportions of Co(0) (≈ 89%) and CoO (≈ 11%) after passing the first 10 minutes of reaction, and EXAFS fitting reveals that no sintering occurs. These results indicate that the fouling of the catalyst surface by coke is responsible for the observed deactivation after 90 min of reaction. In order to regenerate the catalyst, 5% O2/He is then added to the reactant flow aiming to oxidize the coke deposits. The duration of the oxidative ESR treatment (OX) is determined from the careful in-line examination of the recorded Raman spectra. As soon as a total vanishing of the Raman G-line is observed (Fig. 29.21a), the O2 flow is stopped, and the catalyst is exposed back to the original ESR conditions for 60 min. Due to the removal of coke deposits, the metallic cobalt species in contact to O2 at the end of OX conditions have been partially oxidized (Fig. 29.21b). However, ESR restarts immediately at the contact of ethanol steam with regenerated catalyst, which produces H2 and gives rise to a self-reactivation of the catalyst with 74% of Co(0) after 60 min of the second ESR. Again, Raman indicates that the deactivation occurring in the second ESR is related to the coke formation. Finally at the end of the third ESR, 67% of Co(0) are obtained. It is noteworthy that despite a decrease of the concentration of active species upon the successive regenerative treatments, the ethanol conversion is still observed at the same level around 77–80%, when it reaches its maximum. This behavior is explained as the result of the fragmentation of the coke-encapsulated Co

29

652

A. R. Passos et al.

1.2

0.9

c)

H

0.6 0.9

H

0.6

0.3

20000 20008 20016

0° 60° 120° 180° 240° 300°

0.2 0.1

0.1 0.0

–0.1

0.0 –0.2 19950

–0.1

20000

–0.2

20050 Energy (eV)

20100

20150

–0.3 19950

20000

20050 Energy (eV)

20100

20150

d)

b)

H2O/H2

0.776

H2S/H2

1.0

1.0 0.775

0.5

0.774

0.0

0.6

MoO3 MoO2

0.4

Exp (MES) Std diff

0.2 0.0 0.45

0.003

P

0.002

0.30 0.15

0.001

0.00

0.000

–0.15

–0.001

Normalized absorption at 20008 eV

0.8 0.773

–0.5

0.772 –1.0

CoMo

0.771 0.758 0.757

1.0

0.756

0.5

0.755

0.0

0.754

–0.5

0.753 19950

20000

20050 Energy (eV)

20100

20150

sin ((360°/360 s) t + 90°)

0.777 1.2

Absorbance (a.u.)

Normalized absorbance (a.u.)

0.2

–1.0

NiMo

sin ((360°/360 s) t + 120°)

Absorbance (a.u.) × 102

0.0 0.3

Absorbance (a.u.)

Mo (120°) CoMo (90°) NiMo (120°)

0.3 Absorbance (a.u.) × 102

Normalized absorption (a.u.)

a)

0.752 0

60

120

180 Time (s)

240

300

360

Fig. 29.20 (a) Mo K-edge TR-XAS monitoring of unpromoted Mo catalyst during 10 periods of 180 s pulses of 3% H2O/H2 and 0.1% H2S/ H2 and demodulated spectra at selected values of phase angle. (b) Experimental XANES spectra for MoS2 and MoO3 references and comparison of their difference with the maximum amplitude signal obtained after demodulation for the Mo catalyst. (c) Comparison of

the maximum amplitude demodulated signal obtained for the Mo, CoMo, and NiMo catalysts and (d) kinetics of the phase transition of Mo observed for the promoted catalysts considering the intensity of the shoulder H versus time. The dashed line is the sine curve with the phase shift revealed by demodulation. (Reprinted with permission from Ref. [70]. Copyright (2019) American Chemical Society)

(0) particles upon O2 exposure and induced by the so-called Kirkendall effect. Due to the encapsulation of particles by carbon shells, the lattice stress associated with the volume expansion during particle oxidation cannot be released,

except by the rupture of the particles into smaller fragments at the Kirkendall nanovoids (Fig. 29.21e). The fragmentation of the Co(0) nanoparticles has been fully confirmed by microscopy, showing a decrease of the mean particle size

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

a)

c)

OX condition

653

1st ESR

2nd ESR

OX

3rd ESR

OX

100

G line intensity (103 counts)

3.0

Time under O2

600

70 2.0

60 50

1.5 40 1.0

30 20

30 min

0.5

800 1000 1200 1400 1600 1800 2000

0.0 100

10

Raman shift (cm–1)

0 10

90 Percentage of cobalt species

b)

Normalized absorbance

80

2.5

1 0.8 0.6 0.4 0.2

9

80

Co(0)

8

70 7 60 6

50

5

40

4

CoO

30

3

20 10 15 5 )

in

m

e(

m Ti

25

d)

2

Co3O4

0 7780 7700 7720 7740 7760 Energy (eV) End of ESR

Rescaled N(Co-Co)

0 min

D

90

Ethanol conversion (%)

G

1 0

50

100

200

150

250

Time (min)

e) Kirkendall effect

End of OX

O2

Co2+ O2 Co0 O2

50 nm

50 nm 30 20 15 10

5.4 nm

Count

Count

25

5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Diameter (nm)

35 30 25 20 15 10 5 0

4.0 nm

1 2 3 4 5 6 7 8 9 10 11 12 13 Diameter (nm)

Fig. 29.21 Monitoring of successive ESR and oxidative (OX) conditions over a Co/Al2O3 catalyst by (a) Raman spectroscopy

and (b) Co K-edge quick-XAS. (c, top) Evolution of ethanol conversion and intensity of the Raman G-line (diamond symbols). (c, bottom)

29

654

A. R. Passos et al.

(Fig. 29.21d), and by EXAFS fitting (Fig. 29.21c), which displays a decrease of the coordination number NCo-Co. This study clearly demonstrates how powerful is the combination of TR-XAS/MS with Raman spectroscopy for understanding the deactivation mechanisms related to coke deposits but also to propose strategies to regenerate catalysts.

29.4

Acknowledgments The development of MCR-ALS analysis at the ROCK beamline (SOLEIL) has been financially supported by a public grant overseen by the French National Research Agency (Project ANR-10-EQPX 10-45). We thank Olga Roudenko (SOLEIL) for providing fast handling calibration and normalization skills for the TR-XAS data and Ludovic Duponchel (LASIR) for allowing us the discovery of the power of the multivariate analysis applied to process-like data. The authors also acknowledge Celso V. Santilli and Sandra H. Pulcinelli (UNESP) for the collaboration and discussions about ESR catalysts.

Conclusion

This chapter clearly demonstrates how TR-XAS, available with state-of-the-art quick-EXAFS or energy-dispersive beamlines, is essential for an in-depth description of the structural and electronic modifications suffered by catalysts. The technique can be used in very versatile experimental conditions to solve the nucleation and growth of nanoparticles in liquid media, eventually at high temperature, or to monitor liquid or gas-solid reactions leading to the activation of catalysts and its further evolution in reaction conditions encompassing high pressure and temperature. The timeresolved data collected with sub-second time resolution are successive snapshots of the mixture of species representative of the evolution of the reaction. To describe the precise movie of the catalyst life span, multivariate data analysis methods such as MCR-ALS are nowadays emerging. They lead to the isolation of the species involved in the reaction together with the description of their evolution as function of the reaction coordinates. Beyond the speciation provided by MCR-ALS or LCF from known references, X-ray absorption spectroscopy, by EXAFS fittings, offers also unique opportunity to identify the species participating to the reaction. Not yet fully used, the speciation available by MCR-ALS analysis of TR-XAS data offers also solid bases to investigate the kinetic laws driving the phase formation with the determination of growth exponent and the effective activation energy. Many of the examples presented herein also use complementary techniques, such as simultaneous Raman spectroscopy or SAXS to cite only a few, in order to obtain a full and detailed description of the complexity of the reactions in which catalysts are involved and eventually guide the multivariate data analysis toward a more complete description of the chemical transformation. Finally, the next step for the precise description of the catalyst life span will be to routinely combine sub-second time resolution of XAS data with spatial resolution at the micro- or nanometer scales relevant for the investigated processes.

References 1. Müller, O., Nachtegaal, M., Just, J., Lützenkirchen-Hecht, D., Frahm, R.: Quick-EXAFS setup at the SuperXAS beamline for in situ X-ray absorption spectroscopy with 10ms time resolution. J. Synchrotron Radiat. 23, 260–266 (2016) 2. Briois, V., La Fontaine, C., Belin, S., Barthe, L., Moreno, T., Pinty, V., Carcy, A., Girardot, R., Fonda, E.: ROCK: the new quickEXAFS beamline at SOLEIL. J. Phys. Conf. Ser. 712, 012149–012155 (2016) 3. Palomino, R.M., Stavitski, E., Waluyo, I., Chen-Wiegart, Y.c.K., Abeykoon, M., Sadowski, J.T., Rodriguez, J.A., Frenkel, A.I., Senanayake, S.D.: New in-situ and operando facilities for catalysis science at NSLS-II: the deployment of real-time, chemical, and structure-sensitive X-ray probes. Synchrotron Radiat. News. 30, 30–37 (2017) 4. Caliebe, W.A., Murzin, V., Kalinko, A., Görlitz, M.: High-flux XAFS-beamline P64 at PETRA III. AIP Conf. Proc. 2054, 060031 (2019) 5. Van Bokhoven, J.A., Lamberti, C. (eds.): X-Ray Absorption and X-Ray Emission Spectroscopy. Wiley, Chichester, UK (2016) 6. Lesage, C., Devers, E., Legens, C., Fernandes, G., Roudenko, O., Briois, V.: High pressure cell for edge jumping X-ray absorption spectroscopy: applications to industrial liquid sulfidation of hydrotreatment catalysts. Catal. Today. 336, 63–73 (2019) 7. Clark, A.H., Imbao, J., Frahm, R., Nachtegaal, M.: ProQEXAFS: a highly optimized parallelized rapid processing software for QEXAFS data. J. Synchrotron Radiat. 27, 551–557 (2020) 8. Figueroa, S.J.A., Prestipino, C.: PrestoPronto: a code devoted to handling large data sets. J. Phys. Conf. Ser. 712, 12012 (2016) 9. Newville, M.: Larch: an analysis package for XAFS and related spectroscopies. J. Phys. Conf. Ser. 430, 12007 (2013) 10. Stötzel, J., Lützenkirchen-Hecht, D., Grunwaldt, J.-D., Frahm, R.: T-REX: new software for advanced QEXAFS data analysis. J. Synchrotron Radiat. 19, 920–929 (2012) 11. Cassinelli, W.H., Martins, L., Passos, A.R., Pulcinelli, S.H., Santilli, C.V., Rochet, A., Briois, V.: Multivariate curve resolution analysis applied to time-resolved synchrotron X-ray Absorption Spectroscopy monitoring of the activation of copper alumina catalyst. Catal. Today. 229, 114–122 (2014) 12. Rochet, A., Baubet, B., Moizan, V., Pichon, C., Briois, V.: Co-K and Mo-K edges Quick-XAS study of the sulphidation properties of Mo/Al2O3 and CoMo/Al2O3 catalysts. C. R. Chim. 19, 1337–1351 (2016)

ä Fig. 29.21 (continued) MCR-ALS quick-XAS speciation of Co phases and NCo–Co coordination numbers rescaled by the percentage of Co(0) species during ESR-OX cycles. (d) STEM images of the catalyst after the first ESR and the first OX, together with the respective

size distributions. (e) Scheme for the oxidative regeneration of a cokeencapsulated Co(0) particle. (Reprinted with permission from Ref. [25]. Copyright (2018) The Royal Society of Chemistry)

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

13. Voronov, A., Urakawa, A., van Beek, W., Tsakoumis, N.E., Emerich, H., Rønning, M.: Multivariate curve resolution applied to in situ X-ray absorption spectroscopy data: an efficient tool for data processing and analysis. Anal. Chim. Acta. 840, 20–27 (2014) 14. Staniuk, M., Hirsch, O., Kränzlin, N., Böhlen, R., van Beek, W., Abdala, P.M., Koziej, D.: Puzzling mechanism behind a simple synthesis of cobalt and cobalt oxide nanoparticles: in situ synchrotron X-ray absorption and diffraction studies. Chem. Mater. 26, 2086–2094 (2014) 15. Martini, A., Borfecchia, E., Lomachenko, K.A., Pankin, I.A., Negri, C., Berlier, G., Beato, P., Falsig, H., Bordiga, S., Lamberti, C.: Composition-driven Cu-speciation and reducibility in cu-CHA zeolite catalysts: a multivariate XAS/FTIR approach to complexity. Chem. Sci. 8, 6836–6851 (2017) 16. Clark, A.H., Nuguid, R.J.G., Steiger, P., Marberger, A., Petrov, A.W., Ferri, D., Nachtegaal, M., Kröcher, O.: Selective catalytic reduction of NO with NH3 on CuSSZ-13: deciphering the low and high-temperature rate-limiting steps by transient XAS experiments. ChemCatChem. 12, 1429–1435 (2020) 17. Tsoukalou, A., Abdala, P.M., Stoian, D., Huang, X., Willinger, M.G., Fedorov, A., Müller, C.R.: Structural evolution and dynamics of an In2O3 catalyst for CO2 hydrogenation to methanol: an operando XAS-XRD and in situ TEM study. J. Am. Chem. Soc. 141, 13497–13505 (2019) 18. Rochet, A., Moizan, V., Diehl, F., Pichon, C., Briois, V.: Quick-XAS and Raman operando characterisation of a cobalt alumina-supported catalyst under realistic Fischer–Tropsch reaction conditions. Catal. Today. 205, 94–100 (2013) 19. Jacobs, G., Ma, W., Gao, P., Todic, B., Bhatelia, T., Bukur, D.B., Davis, B.H.: The application of synchrotron methods in characterizing iron and cobalt Fischer–Tropsch synthesis catalysts. Catal. Today. 214, 100–139 (2013) 20. Hong, J., Marceau, E., Khodakov, A.Y., Griboval-Constant, A., La Fontaine, C., Briois, V.: A time-resolved in situ quick-XAS investigation of thermal activation of Fischer–Tropsch silica-supported cobalt catalysts. Chem. – A Eur. J. 18, 2802–2805 (2012) 21. Munnik, P., Krans, N.A., de Jongh, P.E., de Jong, K.P.: Effects of drying conditions on the synthesis of Co/SiO2 and Co/Al2O3 Fischer–Tropsch catalysts. ACS Catal. 4, 3219–3226 (2014) 22. Munnik, P., de Jongh, P.E., de Jong, K.P.: Recent developments in the synthesis of supported catalysts. Chem. Rev. 115, 6687–6718 (2015) 23. Passos, A.R., Martins, L., Pulcinelli, S.H., Santilli, C.V., Briois, V.: Effect of the balance between Co(II) and Co(0) oxidation states on the catalytic activity of cobalt catalysts for Ethanol Steam Reforming. Catal. Today. 229, 88–94 (2014) 24. Passos, A.R., Martins, L., Pulcinelli, S.H., Santilli, C.V., Briois, V.: Correlation of sol-gel alumina-supported cobalt catalyst processing to cobalt speciation, ethanol steam reforming activity, and stability. ChemCatChem. 9, 3918–3929 (2017) 25. Passos, A.R., La Fontaine, C., Martins, L., Pulcinelli, S.H., Santilli, C.V., Briois, V.: Operando XAS/Raman/MS monitoring of ethanol steam reforming reaction–regeneration cycles. Cat. Sci. Technol. 8, 6297–6301 (2018) 26. de Juan, A., Jaumot, J., Tauler, R.A.: Multivariate Curve Resolution (MCR). Solving the mixture analysis problem. Anal. Methods. 6, 4964–4976 (2014) 27. Cassinelli, W.H., Martins, L., Magnani, M., Pulcinelli, S.H., Briois, V., Santilli, C.V.: Time-resolved XAS/MS/Raman monitoring of mutual copper self-reduction and ethanol dehydrogenation reactions. RSC Adv. 6, 20453–20457 (2016) 28. Jaumot, J., De Juan, A., Tauler, R.: Chemometrics and intelligent laboratory systems MCR-ALS GUI 2.0: new features and applications. Chemom. Intell. Lab. Syst. 140, 1–12 (2015) 29. Rochet, A., Baubet, B., Moizan, V., Devers, E., Hugon, A., Pichon, C., Payen, E., Briois, V.: Intermediate species revealed during

655 sulfidation of bimetallic hydrotreating catalyst: a multivariate analysis of combined time-resolved spectroscopies. J. Phys. Chem. C. 121, 18544–18556 (2017) 30. Fornasini, P., Grisenti, R.: On EXAFS Debye-Waller factor and recent advances. J. Synchrotron Radiat. 22, 1242–1257 (2015) 31. Passos, A.R., Pulcinelli, S.H., Santilli, C.V., Briois, V.: Operando monitoring of metal sites and coke evolution during non-oxidative and oxidative ethanol steam reforming over Ni and NiCu ex-hydrotalcite catalysts. Catal. Today. 336, 122–130 (2019) 32. Bugaev, A.L., Skorynina, A.A., Braglia, L., Lomachenko, K.A., Guda, A., Lazzarini, A., Bordiga, S., Olsbye, U., Lillerud, K.P., Soldatov, A.V., Lamberti, C.: Evolution of Pt and Pd species in functionalized UiO-67 metal-organic frameworks. Catal. Today. 336, 33–39 (2019) 33. Hong, J.P., Marceau, E., Khodakov, A.Y., Gaberova, L., GribovalConstant, A., Girardon, J.S., La Fontaine, C., Briois, V., Gaberová, L., Griboval-Constant, A., Girardon, J.S., La Fontaine, C., Briois, V.: Speciation of ruthenium as a reduction promoter of silicasupported Co catalysts: a time-resolved in situ XAS investigation. ACS Catal. 5, 1273–1282 (2015) 34. Rochet, A., Baubet, B., Moizan, V., Devers, E., Hugon, A., Pichon, C., Payen, E., Briois, V.: Influence of the preparation conditions of oxidic NiMo/Al2O3 catalysts on the sulfidation ability: a quick-XAS and Raman spectroscopic study. J. Phys. Chem. C. 119, 23928–23942 (2015) 35. Rochet, A.A., Passos, A.R., Legens, C.C., Briois, V.V., Ribeiro Passos, A., Legens, C.C., Briois, V.V.: Sulphidation study of a dried Ni/Al2O3 catalyst by time-resolved XAS-MS combined with in situ Raman spectroscopy and multivariate Quick-XAS data analysis. Catal. Struct. React. 3, 33–42 (2017) 36. Nikulshina, M., Blanchard, P., Lancelot, C., Griboval-Constant, A., Marinova, M., Briois, V., Nikulshin, P., Lamonier, C.: Genesis of active phase in MoW/Al2O3 hydrotreating catalysts monitored by HAADF and in situ QEXAFS combined to MCR-ALS analysis. Appl. Catal. B Environ. 269, 118766 (2020) 37. Passos, A.R., La Fontaine, C., Pulcinelli, S.H., Santilli, C.V., Briois, V.: Quick-EXAFS and Raman monitoring of activation, reaction and deactivation of NiCu catalysts obtained from hydrotalcite-like precursors. Phys. Chem. Chem. Phys. 22, 18835 (2020) 38. Carvalho, H.W.P., Pulcinelli, S.H., Santilli, C.V., Leroux, F., Meneau, F., Briois, V.: XAS/WAXS time-resolved phase speciation of chlorine LDH thermal transformation: emerging roles of isovalent metal substitution. Chem. Mater. 25, 2855–2867 (2013) 39. Girard, V., Chiche, D., Baudot, A., Bazer-Bachi, D., Lemaitre, L., Moizan-Basle, V., Rochet, A., Briois, V., Geantet, C.: In situ QXAS study of sulfidation/oxidative regeneration reactions of zinc molybdate (ZnMoO4) and ZnO-MoO3 materials. Phys. Chem. Chem. Phys. 21, 8569–8579 (2019) 40. Rabeah, J., Briois, V., Adomeit, S., La Fontaine, C., Bentrup, U., Brückner, A.: Multivariate analysis of coupled operando EPR/XANES/ EXAFS/UV–vis/ATR–IR spectroscopy: a new dimension for mechanistic studies of catalytic gas-liquid phase reactions. Chem. – A Eur. J. 26, 7395–7404 (2020) 41. Kvande, K., Pappas, D.K., Borfecchia, E., Lomachenko, K.A.: Advanced X-ray absorption spectroscopy analysis to determine structure-activity relationships for Cu-zeolites in the direct conversion of methane to methanol. ChemCatChem. 12, 2385–2405 (2020) 42. Weckhuysen, B.M.: Studying birth, life and death of catalytic solids with operando spectroscopy. Natl. Sci. Rev. 2, 147–149 (2015) 43. Nørskov, J.K., Bligaard, T., Rossmeisl, J., Christensen, C.H.: Towards the computational design of solid catalysts. Nat. Chem. 1, 37–46 (2009) 44. Sun, Y., Ren, Y.: In situ synchrotron X-ray techniques for real-time probing of colloidal nanoparticle synthesis. Part. Part. Syst. Charact. 30, 399–419 (2013)

29

656 45. You, H., Fang, J.: Particle-mediated nucleation and growth of solution-synthesized metal nanocrystals: a new story beyond the LaMer curve. Nano Today. 11, 145–167 (2016) 46. Tanaka, T., Ohyama, J., Teramura, K., Hitomi, Y.: Formation mechanism of metal nanoparticles studied by XAFS spectroscopy and effective synthesis of small metal nanoparticles. Catal. Today. 183, 108–118 (2012) 47. Abécassis, B., Testard, F., Kong, Q., Baudelet, F., Spalla, O.: Influence of monomer feeding on a fast gold nanoparticles synthesis: time-resolved XANES and SAXS experiments. Langmuir. 26, 13847–13854 (2010) 48. Harada, M., Kamigaito, Y.: Nucleation and aggregative growth process of platinum nanoparticles studied by in situ quick XAFS spectroscopy. Langmuir. 28, 2415–2428 (2012) 49. Booth, S.G., Chang, S.-Y., Uehara, A., La Fontaine, C., Cibin, G., Schroeder, S.L.M., Dryfe, R.A.W.: In situ XAFS study of palladium electrodeposition at the liquid/liquid interface. Electrochim. Acta. 235, 251–261 (2017) 50. Yao, T., Liu, S., Sun, Z., Li, Y., He, S., Cheng, H., Xie, Y., Liu, Q., Jiang, Y., Wu, Z., Pan, Z., Yan, W., Wei, S.: Probing nucleation pathways for morphological manipulation of platinum nanocrystals. J. Am. Chem. Soc. 134, 9410–9416 (2012) 51. Strach, M., Mantella, V., Pankhurst, J.R., Iyengar, P., Loiudice, A., Das, S., Corminboeuf, C., van Beek, W., Buonsanti, R.: Insights into reaction intermediates to predict synthetic pathways for shapecontrolled metal nanocrystals. J. Am. Chem. Soc. 141, 16312–16322 (2019) 52. Pan, Y.-T., Yang, H.: Design of bimetallic catalysts and electrocatalysts through the control of reactive environments. Nano Today. 31, 100832 (2020) 53. Wolters, M., Munnik, P., Bitter, J.H., de Jongh, P.E., de Jong, K.P.: How NO affects nickel and cobalt nitrates at low temperatures to arrive at highly dispersed silica-supported nickel and cobalt catalysts. J. Phys. Chem. C. 115, 3332–3339 (2011) 54. Jacobs, G., Ma, W., Davis, B.H., Cronauer, D.C., Kropf, J.A., Marshall, C.L.: Fischer–Tropsch synthesis: TPR-XAFS analysis of Co/silica and Co/alumina catalysts comparing a novel NO calcination method with conventional air calcination. Catal. Lett. 140, 106–115 (2010) 55. Briois, V., Belin, S., Villain, F., Bouamrane, F., Lucas, H., Lescouzec, R., Julve, M., Verdaguer, M., Tokumoto, M.S., Santilli, C.V., Pulcinelli, S.H., Carrier, X., Krafft, J.M., Jubin, C., Che, M.: New insights for materials science characterisation using different complementary techniques combined with X-ray absorption spectroscopy. Phys. Scr. T115, 38–44 (2005) 56. Carrier, X., Marceau, E., Che, M.: Physical techniques and catalyst preparation: determining the interactions of transition-metal complexes with oxide surfaces. Pure Appl. Chem. 78, 1039–1055 (2006) 57. Filez, M., Poelman, H., Redekop, E.A., Galvita, V.V., Alexopoulos, K., Meledina, M., Ramachandran, R.K., Dendooven, J., Detavernier, C., Van Tendeloo, G., Safonova, O.V., Nachtegaal, M., Weckhuysen, B.M., Marin, G.B.: Kinetics of lifetime changes in bimetallic nanocatalysts revealed by quick X-ray absorption spectroscopy. Angew. Chem. Int. Ed. 57, 12430–12434 (2018) 58. Newton, M.A., Knorpp, A.J., Sushkevich, V.L., Palagin, D., van Bokhoven, J.A.: Active sites and mechanisms in the direct conversion of methane to methanol using Cu in zeolitic hosts: a critical examination. Chem. Soc. Rev. 49, 1449–1486 (2020) 59. Lomachenko, K.A., Martini, A., Pappas, D.K., Negri, C., Dyballa, M., Berlier, G., Bordiga, S., Lamberti, C., Olsbye, U., Svelle, S., Beato, P., Borfecchia, E.: The impact of reaction conditions and material composition on the stepwise methane to methanol conversion over Cu-MOR: an operando XAS study. Catal. Today. 336, 99–108 (2019) 60. Dessal, C., Len, T., Morfin, F., Rousset, J.-L., Aouine, M., Afanasiev, P., Piccolo, L.: Dynamics of single Pt atoms on alumina

A. R. Passos et al. during CO oxidation monitored by operando X-ray and infrared spectroscopies. ACS Catal. 9, 5752–5759 (2019) 61. Li, X., Yang, X., Zhang, J., Huang, Y., Liu, B.: In situ/operando techniques for characterization of single-atom catalysts. ACS Catal. 9, 2521–2531 (2019) 62. Argyle, D.M., Bartholomew, H.C.: Heterogeneous catalyst deactivation and regeneration: a review. Catalysts. 5, 145–269 (2015) 63. Nagai, Y., Dohmae, K., Ikeda, Y., Takagi, N., Hara, N., Tanabe, T., Guilera, G., Pascarelli, S., Newton, M.A., Takahashi, N., Shinjoh, H., Matsumoto, S.: In situ observation of platinum sintering on ceria-based oxide for autoexhaust catalysts using Turbo-XAS. Catal. Today. 175, 133–140 (2011) 64. Gremminger, A.T., Pereira de Carvalho, H.W., Popescu, R., Grunwaldt, J.-D., Deutschmann, O.: Influence of gas composition on activity and durability of bimetallic Pd-Pt/Al2O3 catalysts for total oxidation of methane. Catal. Today. 258, 470–480 (2015) 65. Mutz, B., Gänzler, A.M., Nachtegaal, M., Müller, O., Frahm, R., Kleist, W., Grunwaldt, J.: Surface oxidation of supported Ni particles and its impact on the catalytic performance during dynamically operated methanation of CO2. Catalysts. 7, 279 (2017) 66. Stötzel, J., Frahm, R., Kimmerle, B., Nachtegaal, M., Grunwaldt, J.D.: Oscillatory behavior during the catalytic partial oxidation of methane: following dynamic structural changes of palladium using the QEXAFS technique. J. Phys. Chem. C. 116, 599–609 (2012) 67. Boubnov, A., Gänzler, A., Conrad, S., Casapu, M., Grunwaldt, J.-D.: Oscillatory CO oxidation over Pt/Al2O3 catalysts studied by in situ XAS and DRIFTS. Top. Catal. 56, 333–338 (2013) 68. Lott, P., Doronkin, D.E., Zimina, A., Tischer, R., Popescu, S., Belin, S., Briois, V., Casapu, M., Grunwaldt, J.D., Deutschmann, O.: Understanding sulfur poisonning of bimettalic Pd-Pt methane oxidation catalysts and their regeneration. Appl. Catal. B Environ. 278, 119244 (2020) 69. Kuzmenko, D., Nachtegaal, M., Copéret, C., Schildhauer, T.J.: Molecular-level understanding of support effects on the regenerability of Ru-based catalysts in the sulfur-poisoned methanation reaction. J. Catal. 375, 74–80 (2019) 70. Gaur, A., Hartmann Dabros, T.M., Høj, M., Boubnov, A., Prüssmann, T., Jelic, J., Studt, F., Jensen, A.D., Grunwaldt, J.-D.: Probing the active sites of MoS2 based hydrotreating catalysts using modulation excitation spectroscopy. ACS Catal. 9, 2568–2579 (2019) 71. Urakawa, A., Bürgi, T., Baiker, A.: Sensitivity enhancement and dynamic behavior analysis by modulation excitation spectroscopy: principle and application in heterogeneous catalysis. Chem. Eng. Sci. 63, 4902–4909 (2008) 72. Ferri, D., Newton, M.A., Nachtegaal, M.: Modulation excitation X-ray absorption spectroscopy to probe surface species on heterogeneous catalysts. Top. Catal. 54, 1070 (2011)

Aline Ribeiro Passos received her cotutelle PhD from São Paulo State University (UNESP), Araraquara, and Paris-Sud University, Orsay, in 2015. She was a postdoctoral researcher in X-ray absorption at ROCK beamline, Synchrotron SOLEIL (2016–2017). She joined the Brazilian Synchrotron Light Laboratory in 2017; she works now on coherent X-ray scattering, focused on materials science and catalysis.

29

Case Studies: Time-Resolved X-Ray Absorption Spectroscopy (XAS)

Camille La Fontaine received his PhD from the Poitiers University (France) in 2007. He worked at the Laboratoire de Réactivité de Surface (Paris) and SOLEIL, the French Synchrotron Radiation facility, where he was beamline scientist on the ROCK beamline from 2015 to 2021. His activities were focused on operando quick-EXAFS characterizations of catalysts and on the development of hyperspectral imaging.

Amélie Rochet received her PhD in physico-chemistry in 2011 (ParisSud, France) in a joint collaboration between SOLEIL and IFP Energies nouvelles. She worked at the Institute for Chemical Technology and Polymer Chemistry in Karlsruhe (Germany). In 2015, she joined the Brazilian synchrotron light source (CNPEM). Her work focuses on the investigation of catalytic materials using synchrotron radiation methods.

657

Valérie Briois received her PhD from the Paris VI University (Paris) in 1991. She worked at LURE and now at SOLEIL, the French synchrotron radiation facilities. CNRS research director, she is the head of the ROCK quick-EXAFS beamline at SOLEIL. She works on the operando quickEXAFS characterizations of catalysts and contributes to the spread of MCR-ALS applied to quick-EXAFS analysis.

29

X-Ray Absorption Spectroscopy (XAS): Surface Structural Determination of Alloy Nanoparticles Guanghui Zhang , Nicole LiBretto, Stephen Purdy Jeffrey Miller

, Laryssa Cesar

30 , and

Contents 30.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660

30.2

Surface XAS of Alloy Metal Nanoparticle Catalysts: Basic Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660

30.3

Case Study 1: Incomplete Formation of a Pt3Cr Surface Alloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660

30.4

Case Study 2: Identification of the Evolution of the Core-Shell Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663

G. Zhang Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN, USA State Key Laboratory of Fine Chemicals, PSU-DUT Joint Center for Energy Research, School of Chemical Engineering, Dalian University of Technology, Dalian, Liaoning, People’s Republic of China e-mail: [email protected] N. LiBretto Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN, USA Manufacturing Product Technology, Research & Development, Honeywell UOP, McCook, IL, USA e-mail: [email protected]; [email protected] S. Purdy Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN, USA Manufacturing Science Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA e-mail: [email protected]

30.5

Case Study 3: Identification of Bimetallic Alloy Compositions Suitable for Determination of Electronic Changes by XANES or RIXS Spectroscopy . . . . . . . . . . . . . 664

30.6

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668

Abstract

In catalysis since reactions occur on the surface of nanoparticles (NP), it is essential to determine the composition of this structure, rather than that of the nanoparticles, since the two may be, and often are, different. Conventional techniques including X-ray absorption spectroscopy (XAS) and X-ray diffraction (XRD) are powerful techniques, but these data reflect the average composition of the entire particle. In this chapter, we introduce the method of EXAFS analysis, which isolates the surface atoms of nanoparticles based on its sensitivity to chemical reactions, specifically surface oxidation. As shown in our case studies, if the surface of a Pt-based nanoparticle is contacted by air at room temperature, the surface will selectively oxidize, resulting in the loss of Pt-Pt and Pt-M bonds to the formation of Pt-O bonds. The difference between the completely reduced and surface-oxidized nanoparticle allows for the isolation of signal from the catalytic surface. Although these examples highlight Pt alloys, similar analysis is also possible for other group 8 and IB bimetallic catalysts.

L. Cesar Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN, USA

Keywords

Dow Performance Silicones, Dow Chemical Company, Midland, MI, USA e-mail: [email protected]

Surface structure · Surface-sensitive X-ray absorption · Intermetallic alloy · Resonant inelastic X-ray scattering · Propane dehydrogenation · Pt alloy catalysts

J. Miller (*) Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN, USA e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_30

659

660

30.1

G. Zhang et al.

Introduction

The identification of the geometric and electronic structures of heterogeneous catalysts and their evolution under realistic reaction conditions is essential for developing structurefunction relations that motivates catalyst development. With the availability of a new generation of synchrotron radiation facilities, scientists are now able to characterize catalytic nanomaterials under working conditions with control of specific reaction conditions (e.g., temperature, pressure, reactor configuration, and chemical feedstock). In particular, X-ray absorption spectroscopy (XAS) has been extensively used by the catalysis community to identify the oxidation state and geometric arrangement of atoms. This has further been enabled through the development of fast scan modes, allowing for the capture of information about key reaction intermediates and mechanistic steps in the catalytic cycle [1–5]. Many elements of interest in catalytic materials have absorption edges in the hard X-ray regime. This allows for a variety of in situ and in operando sample environments since hard X-rays can penetrate cell windows, gas or liquid reactants, support materials, etc. Hence, one can make measurements under realistic reaction conditions and detect all atoms of interest in the sample. While XAS measures every atom, which is important for understanding the nanoparticle structure, chemical reactions occur at the surface of heterogeneous catalysts, where the structure of the active sites may differ from the average composition [6–14]. A precise understanding of the surface structure in alloy nanoparticles, therefore, is critical for relating the active site structure to catalytic performance. There are many excellent books and reviews on the theory and practice of XAS [15–24], and in this chapter, we will not introduce the basic principles but, instead, present a method for determination of the surface composition and structure in alloy nanoparticle and show its potential with three examples. In the first, the average nanoparticle composition of two Pt-Cr alloy catalysts is similar, but their catalytic performance is not, and this difference is related to the surface composition. The second is an example of PtCo alloys in which the surface structure is the same but the structure of the nanoparticle interior changes with varying metal loading. The third example describes the formation of bimetallic Pt-V catalysts with differing composition and average structure. While the catalytic surface structure and reaction performance are identical, to assess the changes in the energy of the valence orbitals responsible for catalysis, for example, by resonant inelastic X-ray scattering (RIXS), requires that the surface and bulk nanostructures are identical since the latter method is not surface-sensitive.

30.2

Surface XAS of Alloy Metal Nanoparticle Catalysts: Basic Approach

In monometallic Pt nanoparticle catalysts, CO or H2 chemisorption is often used to determine the number of surface atoms, i.e., the dispersion. Alternatively, the catalytic surface can be selectively oxidized by O2 at room temperature, and the dispersion can be determined by re-reduction of the oxidized surface by H2 titration. For many (reduced) nanoparticle catalysts, especially for noble metals, ambient oxidation leads to selective oxidation of the (monolayer) catalytic surface. The surface EXAFS method takes advantage of this selective oxidation. In the fully reduced nanoparticle, all atoms are reduced, while, in the oxidized nanoparticle, the surface atoms are oxidized, but the interior atoms remain reduced. For surface EXAFS analysis, the spectra of both the reduced and oxidized catalyst are determined. Subtraction of the oxidized from the reduced EXAFS gives a difference spectrum that allows for determination of the surface structure. See references [6, 9, 10, 25], and for more details on the procedure for this analysis. Since the interior atoms in both the reduced and oxidized nanoparticle are unchanged, these interior atoms subtract and are not present in the difference spectrum leaving only EXAFS from surface atoms. Specifically, there are fewer metallic atoms in the oxidized sample, while oxidation also leads to new M-O bonds. The difference spectrum (reduced – oxidized) has metal-metal scattering with phases identical to those in the reduced catalyst, while the M-O bonds are π-radians out of phase (due to the subtraction process) from a normal scattering pair [6, 9, 10, 25]. As will be demonstrated in the examples, this method works well when the fraction of surface atoms is large enough to get a reliable EXAFS in the difference spectrum, for example, nanoparticle less than about 5 nm in size where there are greater than about 20% surface atoms. For nanoparticle larger than about 8 nm, the error in the difference EXAFS fit becomes larger due to the small CNs, although qualitative characterization is still possible. The potential of the surface EXAFS technique will be demonstrated in the following three examples.

30.3

Case Study 1: Incomplete Formation of a Pt3Cr Surface Alloy

Pt-Cr bimetallic catalysts show promising performance for propane dehydrogenation with propylene selectivity higher than 95% compared to monometallic Pt nanoparticles [10]. 2%Pt/SiO2 (denoted 2Pt), 2%Pt-1%Cr/SiO2 (denoted 2Pt-1Cr), and 2%Pt-3%Cr/SiO2 (denoted 2Pt-3Cr) were

30

X-Ray Absorption Spectroscopy (XAS): Surface Structural Determination of Alloy Nanoparticles

prepared by incipient wetness impregnation, and all catalysts had a similar particle size of about 2 nm measured by scanning transmission electron microscopy (STEM). Detailed structural characterization and catalytic performance have been previously reported [10]. Therefore, only a brief summary will be given here to demonstrate why it is necessary to determine the surface structure and how this can be determined using the surface EXAFS analysis. The propane dehydrogenation selectivity of 2Pt shows lower propylene selectivity (75%) than 2Pt-1Cr (95%) and 2Pt-3Cr (98%) at ~15% conversion. Additionally, the selectivity of 2Pt and 2Pt-1Cr decreases with increasing propane conversion, while that of 2Pt-3Cr remains little changed at all conversions, Fig. 30.1. The crystalline phases (long-range order) and coordination environment (local structure) of 2Pt and 2Pt-3Cr were determined using in situ X-ray diffraction (XRD) and XAS, respectively. The XRD pattern of 2Pt-3Cr after reduction at 800
C showed that the bimetallic phase was a Pt3Cr intermetallic alloy. Under these conditions, the XAS indicated that a Pt has about 8 Pt neighbors and 4 Cr neighbors, with a Pt-Cr to Pt-Pt coordination number ratio (CNPt-Cr:Pt-Pt) of 0.5, consistent with the Pt3Cr phase (Cu3Au structure type). However, when 2Pt-3Cr was reduced at 550
C under the conditions of the catalytic reaction, XRD indicates that both Pt and Pt3Cr phases were present. The XAS was also consistent with a Pt-rich morphology with the CNPt-Cr:Pt-Pt ratio of about 0.25. For 2Pt-1Cr reduced at 550
C, the XRD pattern and EXAFS CNPt-Cr:Pt-Pt ratio and bond distances were nearly identical to that of 2Pt-3Cr indicating that both catalysts have

661

very similar average structures (Fig. 30.2). The catalytic performances, however, suggest some difference in the surface compositions. As discussed above, the XAS difference spectrum of the reduced minus oxidized catalyst, depicted in Fig. 30.3, provides surface-sensitive structural information for the two Pt-Cr catalysts. Figure 30.4a shows the two Fourier transform of the k2-weighted EXAFS spectra (reduced and oxidized) for 2Pt-3Cr. Upon room temperature oxidation, there is a small loss of surface Pt-Pt and Pt-Cr bonds and the formation of Pt-O bonds. Cr-O bonds formed from the oxidation of surface chromium in the alloy are not observed at the Pt edge due to the element specificity of XAS. While the oxidized spectrum could be fit directly to determine the core composition, the small changes and overlapping features make quantification of these changes in coordination numbers and bond distances less accurate for determination of the surface composition. By subtracting the oxidized from reduced EXAFS spectrum, Fig. 30.4b, changes in the two catalysts are more readily resolved and can be more accurately fit, Table 30.1. The Pt-Cr to Pt-Pt coordination numbers, bond distances, and CNPt-Cr:Pt-Pt ratio can be used to determine the Cr incorporation into these nanoparticles for the fully reduced alloy, metallic core (in the oxidized nanoparticle), and surface monolayer (from the difference EXAFS), Table 30.2. In the average EXAFS fit of 2Pt-3Cr, i.e., the reduced catalyst, the coordination ratio, CNPt-Cr:Pt-Pt, was 0.30, while that of the surface EXAFS is near 0.5. The latter is consistent with a stoichiometric Pt3Cr intermetallic alloy

100

0.02 2Pt-3Cr

90

2Pt-1Cr

80

FT (k 2 × c(k))

Selectivity (%)

0.01

70 60 2Pt-3Cr

−0.01

2Pt-1Cr

50

0.00

2Pt −0.02

40 0

10

20

30

40

50

Conversion (%)

Fig. 30.1 Propylene selectivity as a function of propane dehydrogenation conversion at 550
C, 2.5% C3H8 and 2.5% H2 (balanced with N2)

0

1

2 R (Å)

3

4

Fig. 30.2 EXAFS of Pt-Cr bimetallic nanoparticles containing 1 wt% and 3 wt% Cr after reduction at 550
C in 3.5% H2

30

-

Pt

=

Pt

Pt

PtO Pt x 3 Cr

Photons hv

Pt

Fig. 30.3 Approach for difference analysis, where reduced Pt-Cr nanoparticles are subsequently oxidized and the EXAFS (chi) data is subtracted to isolate the surface atoms [10]

3 Cr

G. Zhang et al.

3 Cr

662

In situ XAS Reduced: Average composition of whole nanoparticle -

Pt-Pt, Pt-Cr

a)

b)

0.020 Reduced

Oxidized: Composition of the subsurface + oxidized surface Pt-O, Cr-O

=

Pt-Pt, Pt-Cr, Pt-O

0.003

Oxidized

Pt-Cr & Pt-Pt

0.010

Pt-O

FT (k 2 × c(k))

Pt-Cr & Pt-Pt

0.015 FT (k 2 × c (k))

Difference: Monolayer Surface

Pt-Cr & Pt-Pt

0.002 Pt-O

0.001

0.005

0.000

0.000 1

2

3

4

R (Å)

1

2

3

4

R (Å)

Fig. 30.4 The k2-weighted Fourier transform of chi of 2Pt-3Cr after reduction at 550
C; (a) the reduced, oxidized, and (b) difference EXAFS

Table 30.1 Difference EXAFS fits for 2Pt-1Cr and 2Pt-3Cr Catalyst 2Pt-1Cr

Scattering pair Pt-Pt Pt-Cr Pt-O Pt-Pt Pt-Cr Pt-O

2Pt-3Cr

CN 1.5 0.6 0.4 0.9 0.5 0.3

R (Å) 2.75 2.71 2.05 2.73 2.73 2.05

Table 30.2 The CNPt-Cr/CNPt-Pt ratios of 2Pt-1Cr and 2Pt-3Cr catalysts Sample 2Pt-1Cr

CNPt-Cr/CNPt-Pt ratio Average Interior 0.28 0.25

Surface 0.40

2Pt-3Cr

0.30

0.56

0.22

Catalyst structure

surface. The bond distances of the surface EXAFS are also consistent with the Pt3Cr intermetallic alloy. The oxidized EXAS fit also represents the structure of the non-surface atoms in the nanoparticle, i.e., the nanoparticle core. Although the surface is a Pt3Cr phase, the interior is much more Pt-rich than the surface. The analysis of the surface and particle interior also suggests that Cr alloy formation during synthesis starts at the surface of a reduced Pt nanoparticle with metallic Cr atoms diffusing to the nanoparticle interior. A similar difference surface EXAFS analysis of the reduced and oxidized 2Pt-1Cr shows that the surface CNPtCr:Pt-Pt ratio is 0.40 compared to 0.56 for the 2Pt-3Cr catalyst. In other words, the surface of 2Pt-1Cr is Pt-rich suggesting incomplete formation of a surface Pt3Cr monolayer (Table 30.2), which leads to lower catalytic selectivity. By identifying the surface structures and compositions, small changes in the catalytic performance can be explained. For

30

X-Ray Absorption Spectroscopy (XAS): Surface Structural Determination of Alloy Nanoparticles

these two catalysts, the 2Pt-3Cr had a full monolayer Pt3Cr intermetallic alloy composition, while the surface of the 2Pt-1Cr was an incomplete monolayer; see schematic in Table 30.2. For the 2Pt-1Cr catalyst, even reduction at higher temperatures did not lead to a complete surface alloy monolayer [10].

30.4

Case Study 2: Identification of the Evolution of the Core-Shell Structures

In this example, using the difference surface EXAFS analysis, we show that the average Pt-Co NP composition changes with increasing Co loading, while the surface composition remains similar [9]. The average EXAFS of bimetallic Pt-Co nanoparticles containing 2 wt% Pt with 0.6 (2Pt-0.6Co), 1 (2Pt-1Co), 2 (2Pt-2Co), and 4 wt% Co (2Pt-4Co) is compared to a monometallic Pt nanoparticle (3Pt). The k2weighted magnitude of the Fourier transform (FT) at the Pt L3 edge shows different shapes between the four bimetallic particles and Pt (Fig. 30.5 and Table 30.3). The shape of the

0.03 Pt 2Pt-0.6Co

FT (k 2 × c(k))

2Pt-1Co 0.02

2Pt-2Co 2Pt-4Co

0.01

0.00 1

2

3

4

R (Å)

Fig. 30.5 Comparison of EXAFS magnitudes of Pt and Pt-Co catalysts

663

EXAFS spectrum of 2Pt-0.6Co particles is distorted compared to monometallic Pt. The EXAFS fitting results suggest 7.8 Pt-Pt bonds at 2.73 Å and 2.5 Pt-Co bonds at 2.56 Å. The larger Pt-Pt coordination number is consistent with Pt-rich nanoparticles. As the nominal Co content increases, the number of Pt-Pt bonds decreases, and the number of Pt-Co bonds increases. For instance, 2Pt-1Co has 4.5 Pt-Pt bonds at 2.73 Å and 2.2 Pt-Co bonds at 2.56 Å (Table 30.3). The Pt-Co to Pt-Pt coordination number ratio (CNPt-Co:Pt-Pt) is about 0.5, which matches the Pt3Co phase. This structure was also confirmed by in situ synchrotron XRD. In 2Pt-2Co, the number of Pt-Pt bonds decreases to 3.0, and the number of Pt-Co bonds increases to 2.9, i.e., a CNPt-Co:PtPt of about 1.0, or about an equal number of Pt-Pt and Pt-Co bonds. With a further increase in Co, in 2Pt-4Co, the CNPt-Co: Pt-Pt ratio was approximately 2.0. Thus, with increasing Co loading, the average nanoparticle composition changes from Pt-rich to Co-rich from low to high loading, respectively. Since these nanoparticles are small (1–2 nm), there is a sufficiently large fraction of surface atoms, and the difference EXAFS analysis can be performed to identify the surface composition. Upon oxidation, there are loss of surface Pt-Pt and Pt-Co metallic bonds. The remaining metallic bonds in the spectra (Pt-Co and Pt-Pt), therefore, are due to metallic atoms from the nanoparticle interior, i.e., the nanoparticle core. From the difference EXAFS, it is possible to evaluate whether the composition is homogeneous throughout the particle or if the surface and particle interior have different compositions. In addition, the ratio of Pt-Co to Pt-Pt neighbors can be useful to identify the ordered surface structure or, at least, rule out other compositions and structures. The reduced, oxidized, and difference spectra of 2Pt-1Co are shown in Fig. 30.6. In Fig. 30.6a, the large peaks (solid line between about 2–3 Å) of the reduced catalysts represent both Pt-Pt and Pt-Co bonds. The red spectrum in Fig. 30.6a shows the oxidized spectrum with loss of metallic neighbors and addition of a Pt-O peak at about 1.5 Å (phase uncorrected distance). In these bimetallic Pt-Co catalysts, since the fraction of surface atoms is high, there is a significant Pt-O peak, which can be fit. In the difference spectrum, any atoms that are unchanged are not present in the difference spectrum.

Table 30.3 Summary of EXAFS analysis for the reduced bimetallic and monometallic Pt and Co samples Sample 3Pt 2Pt-0.6Co 2Pt-1Co 2Pt-2Co 2Pt-4Co

Scattering pair Pt-Pt Pt-Pt Pt-Co Pt-Pt Pt-Co Pt-Pt Pt-Co Pt-Pt Pt-Co

Bond length (Å) 2.75 2.73 2.56 2.73 2.56 2.73 2.56 2.73 2.56

CN 9.3 7.8 2.5 4.5 2.2 3.0 2.9 2.5 5.2

CNPt-Co:Pt-Pt 0 0.32

Phase Pt Pt þ Pt3Co

0.49

Pt3Co

0.97

Pt3Co þ PtCo

2.08

Pt3Co þ PtCo

30

664

G. Zhang et al.

a)

b)

0.020 Reduced

Pt-Co/Pt

Pt-Co/Pt

0.010

FT (k 2 × c(k))

FT (k 2 × c(k))

0.015

Difference EXAFS

0.009

Surface oxidized

Pt-O Pt-Co/Pt

Pt-O

0.006

Pt-Co/Pt

0.003

0.005

0.000

0.000 1

2

3

4

1

R (Å)

2

3

4

R (Å)

Fig. 30.6 FT magnitude of the EXAFS for 2Pt-1Co at the Pt L3 edge Table 30.4 Fitting results for the difference spectra (reduced minus oxidized catalyst) at the Pt L3 edge Sample 3Pt 2Pt-0.6Co

2Pt-1Co

2Pt-2Co

2Pt-4Co

Bond Pt-O Pt-Pt Pt-O Pt-Pt Pt-Co Pt-O Pt-Pt Pt-Co Pt-O Pt-Pt Pt-Co Pt-O Pt-Pt Pt-Co

Bond length (Å) 2.05 2.77 2.05 2.73 2.56 2.05 2.73 2.56 2.05 2.73 2.56 2.05 2.73 2.56

CN 0.9 2.5 0.2 0.6 0.3 0.9 2.1 1.0 0.9 2.4 1.2 0.7 1.9 0.9

CNPt-Co:Pt-Pt – – – 0.5 – 0.5

dispersion can be estimated from the Pt-O coordination number. For example, Pt(II) compounds have CNs of 4, thus the Pt dispersion is the ratio of Pt-O CN/4 in the difference spectrum. The accuracy of the Pt-O coordination number is more accurate in the difference EXAFS than in the partially oxidized sample, since there is less overlap of the Pt-O and metallic Pt scatters; see Figs. 30.4b and 30.6b, for example. This energy specificity of the surface Pt-O allows for determination of the number of catalytic sites for determination of the turnover rates and identification of the active site, for example, Pt or Co [9].

– 0.5 – 0.5

Thus, the Pt-Pt, Pt-Co, and Pt-O are more easily resolved and fit in the difference EXAFS spectrum (Fig. 30.6b). The fits for the difference EXAFS are shown in Table 30.4. An unexpected result in Table 30.4 is that the CNPt-Co:Pt-Pt is near 0.5 for all catalysts, despite the clear difference in their average and core compositions. This suggests that all catalysts likely have the same surface structure, i.e., Pt3Co (structure type AuCu3) with different core compositions from Pt-rich to Pt3Co to PtCo. As also shown in Fig. 30.6, the difference EXAFS gives a Pt-O scattering peak, which can be related to the number of surface Pt atoms in the alloy. For alloy nanoparticles, the STEM particle size gives the fraction of surface atoms, but not the number of catalytic surface atoms. In addition, as in the case of Pt-Co alloys, since both Pt and Co may adsorb standard adsorbates like H2 and CO, the Pt and Co dispersions are not possible to determine. However, the Pt

30.5

Case Study 3: Identification of Bimetallic Alloy Compositions Suitable for Determination of Electronic Changes by XANES or RIXS Spectroscopy

While the surface alloy composition and structure are important, there are also important electronic changes in the energy of the catalytic, i.e., surface, valence orbitals, which control the metal-reactant bond energies, surface coverage, and performance. For the row 5 catalytic elements, e.g., Pt, the L2 and L3 edge XANES measure the energy of the unfilled 5d orbitals, while resonant inelastic X-ray scattering (RIXS) allows for determination of the energy of the filled 5d orbitals. Both XANES and RIXS measure all atoms in the sample and, thus, are not surface-sensitive. The absorption and emission processes for RIXS are shown in Fig. 30.7 [26–30]. It is the energy of the filled 5d orbitals of the surface-active atoms that is responsible for the catalytic performance. As shown in the case studies above, the composition of bimetallic catalysts is often not uniform, and the surface can have a different composition from that of the average nanoparticle. Since hard X-rays sample all atoms in

30

X-Ray Absorption Spectroscopy (XAS): Surface Structural Determination of Alloy Nanoparticles

Fig. 30.7 RIXS spectroscopy: excitation of the 2p electron to the empty 5d orbitals (XANES absorption spectra) and decay to the core hole from the filled 5d state (emission spectrum). The difference in energy of the absorbed (Ω) and emitted (ω) photon gives the energy difference between filled and unfilled 5d valence orbitals

6s

665

hv 2

6s 5d5/2

Ω = hv1

5d5/2 5d3/2

w = hv2

5d3/2

hv1 2p3/2

2p3/2 L3 -O4,5 X-ray emission (filled 5d states)

L3 X-ray absorption (unfilled 5d states)

RIXS process: ΔE = Ω − w

30 Table 30.5 XANES energies and EXAFS fitting results for 3Pt and two Pt-V catalysts reduced at 550
C

0.02 2Pt-5V

FT (k 2 × c (k))

3Pt 0.01

Sample 3Pt 5Pt-5V

Edge energy (keV) 11.5640 11.5642

2Pt-5V

11.5644

Scattering pair Pt-Pt Pt-Pt Pt-V Pt-Pt Pt-V

CN 8.8 6.5 2.0 6.2 2.9

R (Å) 2.74 2.73 2.71 2.72 2.72

0.00

−0.01 0

1

2 R (Å)

3

4

Fig. 30.8 R-space EXAFS magnitude (solid lines) and imaginary components (dashed lines) for 2Pt-5V/SiO2 and 3Pt/SiO2 collected after reduction in hydrogen at 550
C

the catalyst, in order to accurately determine the energy shifts due to catalytic surface, one must measure the XANES, and RIXS, on catalysts where the nanoparticle and surface compositions are identical, i.e., a fully alloyed nanoparticle with the same surface and interior structure. The following example shows two Pt-V bimetallic alkane dehydrogenation catalysts with similar catalytic selectivity and rates but differ in their structure. Determination of the nanoparticle and surface structures allows for identification of the catalyst in which all atoms have identical geometry and electronic properties allowing for accurate determination of the changes in energy of the Pt valence orbitals due to alloy formation with V [25]. A 3%Pt/SiO2 catalyst (denoted 3Pt) and two Pt-V/SiO2 catalysts with different Pt loadings, 2%Pt-5%V/SiO2 (2Pt-5V) and 5%Pt-5% V/SiO2 (5Pt-5V), were synthesized, and full characterization of the structures and catalytic performance have been previously reported [25]. Thus, only a

brief summary is given. All three catalysts had metal particle sizes of approximately 2.5 nm. Both Pt-V catalysts had propylene selectivity above 95% at 20% propane conversion and comparable propylene turnover rates of 0.3  0.1 s1. The bimetallic structure of the catalyst was verified using in situ EXAS, Fig. 30.8. The 3Pt catalyst showed scattering from Pt neighbors, with three characteristic peaks between 2 and 3 Å. The ratio of the three peaks in the 2Pt-5V catalyst is modified due to Pt-V scattering. First shell fits of the two spectra are given in Table 30.5. The coordination number ratio of Pt-V to Pt-Pt (CNPt-V:Pt-Pt) for 5Pt-5V was 0.31, while the ratio for 2Pt-5V was 0.47, demonstrating that 5Pt-5V was Pt-rich compared to 2Pt-5V. The coordination number ratio of Pt3V intermetallic alloy is 0.5, suggesting that the 2Pt-5V nanoparticles are a full alloy, while those in 5Pt-5V are a phase mixture, i.e., Pt þ Pt3V. The nanoparticle surface compositions of 2Pt-5V and 5Pt-5V were determined using the difference EXAFS spectra (reduced – oxidized) and are shown in Fig. 30.9, and fits are given in Table 30.6. Both spectra show Pt-O scattering between 1 and 2 Å from the surface oxidation process and Pt-Pt and Pt-V scattering between 2 and 3 Å. Fitting the difference spectra and taking the ratio of CNPt-V to CNPt-Pt gave a ratio close to 0.5 for both catalysts, which is consistent with their similar catalytic performance and suggests a Pt3V surface structure. While the surface and bulk ratio for 2Pt-5V were similar, suggesting a pure-phase Pt3V nanoparticle, the

666

G. Zhang et al.

a)

b)

0.004

0.004

Pt-V/Pt

Pt-V/Pt Pt-O

FT (k 2 × c (k))

Pt-O 0.002

0.002

0.000

0.000

0.002

0.002

2Pt-5V

5Pt-5V Fit

Fit 0

1

2

3

4

0

R (Å)

1

2

3

4

R (Å)

Fig. 30.9 Pt EXAFS difference spectra for 2Pt-5V (a) and 5Pt-5V (b) and difference spectra fit Table 30.6 EXAFS fit for Pt-V difference spectra Sample 5Pt-5V

2Pt-5V

Scattering pair Pt-O Pt-Pt Pt-V Pt-O Pt-Pt Pt-V

CN 0.6 1.2 0.7 0.5 0.9 0.6

R (Å) 2.03 2.74 2.74 2.05 2.70 2.72

CNPt-V/CNPt-Pt ratio 0.58

0.66

bulk coordination ratio of 5Pt-5V was lower than that of the surface suggesting a structure with a Pt-rich core and a Pt3V shell consistent with the in situ synchrotron XRD. Consistent with changes in average coordination geometry in the bimetallic Pt-V catalyst, the XANES spectra and edge energies shift slightly to higher energy with increasing V content in the nanoparticles suggesting changes in the energy of the unfilled Pt 5d orbitals; see Fig. 30.10 and Table 30.5. For the 5Pt-5V and 2Pt-5V bimetallic catalysts, there is a shift to higher energy, 11.5642 and 11.5644 keV, respectively, compared to Pt (11.5640 keV); however, edge energy of 5Pt-5V is the average of atoms in the NP, Pt3V þ Pt, rather than the catalytic surface, i.e., Pt3V. Only in the 2Pt-5V is the surface composition and structure the same as the NP. Thus, the true shift in the energy of the empty Pt 5d orbitals between Pt and Pt3V phases is an increase of 0.4 eV. Thus, to obtain accurate determination of the electronic properties of the catalytic surface, i.e., the Pt3V phase, one not only has to determine the nanoparticle composition but also confirm that the surface and bulk structures are the same. Since the average and surface EXAFS of 2Pt-5V indicate a full Pt3V intermetallic alloy, the RIXS spectrum was obtained and compared to that of Pt, Fig. 30.11. The energy transfer is the difference in energy between the absorbed (XANES) and

emitted photons. When the energy transfer is combined with the XANES energy, the energy of the filled 5d orbitals, which are responsible for adsorbate bonding and surface chemistry, can be determined. For Pt, the energy transfer value is 2.7 eV. For Pt3V, the energy transfer value is larger, 3.5 eV. The energy separation between the filled and unfilled states in Pt3V is 0.8 eV larger than Pt, which can be separated into a 0.4 eV increase in the energy of the unfilled states and a 0.4 eV decrease in the energy of the filled 5d states relative to Pt. These values in Pt 5d orbital energies reflect the true nature of the electronic changes due to the Pt3V alloy and, thus, are comparable with density functional theory (DFT) calculations and essential for understanding of the catalytic properties [25].

30.6

Summary

Bimetallic nanoparticles often have complex morphologies where the average and surface structures differ significantly. In the examples highlighted in this chapter, a method is presented which allows for determination of the surface structure even though hard X-rays sample all atoms in the nanoparticle. The difference EXAFS approach relies on the selective oxidation of surface atoms. While these examples were for Pt alloys, the method is not limited to noble metals. For example, the surface atoms of Cu(0) nanoparticles can be selectively oxidized to Cu(I) forming a surface Cu2O layer. The quantity of Cu(I) sites generated during N2O oxidation could be used to estimate the number of active Cu(0) surface sites [31]. Since in the Cu2O surface layer each Cu(I) site is surrounded by two oxygen atoms, the increase of the Cu-O coordination number could also be used to determine the

30

X-Ray Absorption Spectroscopy (XAS): Surface Structural Determination of Alloy Nanoparticles

a)

1.5

b)

667

1.0

3Pt 0.8

2Pt-5V

Normalized absorption

Normalized absorption

5Pt-5V

1.0

0.5

0.6

3Pt

0.4

5Pt-5V 2Pt-5V 0.0 11.55

0.2 11.56

11.57

11.58

11.560

11.59

11.561

11.562

11.563

11.564

11.565

30

Energy (keV)

Energy (keV)

Fig. 30.10 (a) In situ Pt L3 edge XANES of 3Pt/SiO2, 2Pt-5V/SiO2, and 5Pt-5V/SiO2; (b) zoomed-in region of the XANES spectra. Spectra were collected at room temperature in He after reduction at 550
C in 3.5% H2

8

8 Energy transfer (eV)

b) 10

Energy transfer (eV)

a) 10

6

4

6

4

3.50 eV

2.70 eV

Pt3V

Pt

2

2

0 11.555

0 11.560

11.565

11.570

11.575

Photon energy (keV)

11.555

11.560

11.565

11.570

11.575

Photon energy (keV)

Fig. 30.11 (a) Pt L3-Lβ5 RIXS maps of 3Pt and (b) 2Pt-5V (Pt3V). Spectra were collected after a reduction treatment at 550
C in 3.5% H2. Horizontal dashed lines denote the maximum of energy transfer (difference in the absorbed and emitted photons) for each catalyst

surface composition of Cu catalysts. For particle sizes of less than about 5 nm, there is a sufficient fraction of surface atoms to provide accurate surface analysis. For example, in the study of the bimetallic Pt-V catalysts [25], there are small differences in XRD patterns of Pt and Pt3V, which makes identification of the intermetallic alloy structure or solid solutions difficult, especially for small nanoparticles where the peaks are small and broad and can overlap. The surface

EXAFS coordination ratio can also be useful for elimination of several possible alloy structures, for example, that are inconsistent with the fits. In addition, comparison of the NP and surface EXAFS can also rule out the formation of solid solutions. For the latter, the surface and average composition are the same, while for intermetallic alloys core-shell structure are common, i.e., regions of different composition [25]. Finally after elimination of these structures, the NP (for

668

2Pt-5V) and surface CN ratios (and XRD) were consistent with formation of the Pt3V intermetallic alloy. While these examples focused on differences in surface and average structures resulting from different compositions, this method is equally applicable to changes in the surface composition under reaction including changes due to the reacting gases, deactivation, or poisoning. The basis of this technique relies on changes in the surface structure during a chemical reaction. Here the reaction was surface oxidation but could also be applied to changes in the catalyst under reaction conditions. For example, deactivation may lead to M-C surface bonds, sintering, or surface reconstruction. For coke formation one would expect to determine new M-C scatting peaks, which could be fit to determine the CN, i.e., C/Msurface and M-C bond distance. Sintering would lead to increases in the average NP CN, while the surface CN would decrease. Finally, surface reconstruction would lead to changes in surface composition, and the difference EXAFS before and after reaction would show which Ms have lower and higher surface concentrations after reaction. The full potential of the method is to determine these changes in the surface during reaction, deactivation, etc. Acknowledgments This chapter is based on work supported by the National Science Foundation under cooperative agreement no. EEC-1647722. GZ would like to acknowledge the National Natural Science Foundation of China (21902019) and Fundamental Research Funds for the Central Universities (DUT18RC(3)057 and DUT20RC(5)002). Use of the Advanced Photon Source was supported by the US Department of Energy, Office of Basic Energy Sciences, under contract no. DE-AC02-06CH11357. MRCAT operations, beamline 10-BM, are supported by the Department of Energy and the MRCAT member institutions. The authors also acknowledge the use of beamline 11-IDC at APS.

References 1. Zhang, G., Yi, H., Zhang, G., Deng, Y., Bai, R., Zhang, H., Miller, J.T., Kropf, A.J., Bunel, E.E., Lei, A.: Direct observation of reduction of Cu(II) to Cu(I) by terminal alkynes. J. Am. Chem. Soc. 136, 924–926 (2014) 2. Macmillan, S.N., Lancaster, K.M.: X-ray spectroscopic interrogation of transition-metal-mediated homogeneous catalysis: primer and case studies. ACS Catal. 7, 1776–1791 (2017) 3. Paolucci, C., Khurana, I., Parekh, A.A., Li, S., Shih, A.J., Li, H., Di Iorio, J.R., Albarracin-Caballero, J.D., Yezerets, A., Miller, J.T., Delgass, W.N., Ribeiro, F.H., Schneider, W.F., Gounder, R.: Dynamic multinuclear sites formed by mobilized copper ions in NOx selective catalytic reduction. Science. 357, 898–903 (2017) 4. Zhang, G., Yi, H., Xin, J., Deng, Y., Bai, R., Huang, Z., Miller, J.T., Kropf, A.J., Bunel, E.E., Qi, X., Lan, Y., Lei, A.: Aromatic C–H bond cleavage by using a Cu(i) ate-complex. Org. Chem. Front. 3, 975–978 (2016) 5. Wan, G., Zhang, G., Lin, X.-M.: Toward efficient carbon and water cycles: emerging opportunities with single-site catalysts made of 3d transition metals. Adv. Mater. 32, 1905548 (2020)

G. Zhang et al. 6. Wu, Z., Bukowski, B.C., Li, Z., Milligan, C., Zhou, L., Ma, T., Wu, Y., Ren, Y., Ribeiro, F.H., Delgass, W.N., Greeley, J., Zhang, G., Miller, J.T.: Changes in catalytic and adsorptive properties of 2 nm Pt3Mn nanoparticles by subsurface atoms. J. Am. Chem. Soc. 140, 14870–14877 (2018) 7. Ye, C., Wu, Z., Liu, W., Ren, Y., Zhang, G., Miller, J.T.: Structure determination of a surface tetragonal Pt1Sb1 phase on Pt nanoparticles. Chem. Mater. 30, 4503–4507 (2018) 8. Zhang, G., Ye, C., Liu, W., Zhang, X., Su, D., Yang, X., Chen, J.Z., Wu, Z., Miller, J.T.: Diffusion-limited formation of nonequilibrium intermetallic nanophase for selective dehydrogenation. Nano Lett. 19, 4380–4383 (2019) 9. Cesar, L.G., Yang, C., Lu, Z., Ren, Y., Zhang, G., Miller, J.T.: Identification of a Pt3Co surface intermetallic alloy in Pt–Co propane dehydrogenation catalysts. ACS Catal. 9, 5231–5244 (2019) 10. LiBretto, N.J., Yang, C., Ren, Y., Zhang, G., Miller, J.T.: Identification of surface structures in Pt3Cr intermetallic nanocatalysts. Chem. Mater. 31, 1597–1609 (2019) 11. Zhu Chen, J., Wu, Z., Zhang, X., Choi, S., Xiao, Y., Varma, A., Liu, W., Zhang, G., Miller, J.T.: Identification of the structure of the bi promoted Pt non-oxidative coupling of methane catalyst: a nanoscale Pt3Bi intermetallic alloy. Cat. Sci. Technol. 9, 1349–1356 (2019) 12. Wu, Z., Wegener, E.C., Tseng, H.-T., Gallagher, J.R., Harris, J.W., Diaz, R.E., Ren, Y., Ribeiro, F.H., Miller, J.T.: Pd–In intermetallic alloy nanoparticles: highly selective ethane dehydrogenation catalysts. Cat. Sci. Technol. 6, 6965–6976 (2016) 13. Wegener, E.C., Wu, Z., Tseng, H.-T., Gallagher, J.R., Ren, Y., Diaz, R.E., Ribeiro, F.H., Miller, J.T.: Structure and reactivity of Pt–in intermetallic alloy nanoparticles: highly selective catalysts for ethane dehydrogenation. Catal. Today. 299, 146–153 (2018) 14. Shen, X., Zhang, C., Zhang, S., Dai, S., Zhang, G., Ge, M., Pan, Y., Sharkey, S.M., Graham, G.W., Hunt, A., Waluyo, I., Miller, J.T., Pan, X., Peng, Z.: Deconvolution of octahedral Pt3Ni nanoparticle growth pathway from in situ characterizations. Nat. Commun. 9, 4485 (2018) 15. Frenkel, A.I.: Applications of extended X-ray absorption finestructure spectroscopy to studies of bimetallic nanoparticle catalysts. Chem. Soc. Rev. 41, 8163–8178 (2012) 16. Newville, M.: Fundamentals of XAFS. Rev. Mineral. Geochem. 78, 33–74 (2014) 17. Silvia, B., Elena, G., Giovanni, A., Van Bokhoven, J.A., Carlo, L.: Reactivity of surface species in heterogeneous catalysts probed by in situ X-ray absorption techniques. Chem. Rev. 113, 1736–1850 (2013) 18. Koningsberger, D.C., Prins, R.: X-Ray Absorption Principles, Applications, Techniques of EXAFS, SEXAFS and XANES. Wiley (1988) 19. Bunker, G.: Introduction to XAFS. Cambridge University Press (2010) 20. Calvin, S.: XAFS for Everyone. CRC Press (2013) 21. Rehr, J.J., Albers, R.C.: Theoretical approaches to X-ray absorption fine structure. Rev. Mod. Phys. 72, 621–654 (2000) 22. Bokhoven, J.A.V., Lamberti, C.: X-Ray Absorption and X-Ray Emission Spectroscopy: Theory and Applications. Wiley (2016) 23. Nelson, R.C., Miller, J.T.: An introduction to X-ray absorption spectroscopy and its in situ application to organometallic compounds and homogeneous catalysts. Cat. Sci. Technol. 2, 461–470 (2012) 24. Frenkel, A.I., Rodriguez, J.A., Chen, J.G.: Synchrotron techniques for in situ catalytic studies: capabilities, challenges, and opportunities. ACS Catal. 2, 2269–2280 (2012) 25. Purdy, S.C., Ghanekar, P., Mitchell, G., Kropf, A.J., Zemlyanov, D.Y., Ren, Y., Ribeiro, F., Delgass, W.N., Greeley, J., Miller, J.T.: The origin of electronic modification of platinum in a Pt3V alloy and

30

X-Ray Absorption Spectroscopy (XAS): Surface Structural Determination of Alloy Nanoparticles

their consequences for propane dehydrogenation catalysis. ACS Appl. Energy Mater. 3, 1410–1422 (2020) 26. Ravel, B., Kropf, A.J., Yang, D., Wang, M., Topsakal, M., Lu, D., Stennett, M.C., Hyatt, N.C.: Nonresonant valence-to-core X-ray emission spectroscopy of niobium. Phys. Rev. B. 97 (2018). https://doi.org/10.1103/PhysRevB.97.125139 27. Cybulskis, V.J., Bukowski, B.C., Tseng, H.-T., Gallagher, J.R., Wu, Z., Wegener, E., Kropf, A.J., Ravel, B., Ribeiro, F.H., Greeley, J., Miller, J.T.: Zinc promotion of platinum for catalytic light alkane dehydrogenation: insights into geometric and electronic effects. ACS Catal. 7, 4173–4181 (2017) 28. Glatzel, P., Sikora, M., Smolentsev, G., Fernández-García, M.: Hard X-ray photon-in photon-out spectroscopy. Catal. Today. 145, 294–299 (2009) 29. Pieter, G., Jagdeep, S., Kvashnina, K.O., Van Bokhoven, J.A.: In situ characterization of the 5d density of states of Pt nanoparticles upon adsorption of CO. J. Am. Chem. Soc. 132, 2555–2557 (2010) 30. Wegener, E.C., Bukowski, B.C., Yang, D., Wu, Z., Kropf, A.J., Delgass, W.N., Greeley, J., Zhang, G., Miller, J.T.: Intermetallic compounds as an alternative to single-atom alloy catalysts: geometric and electronic structures from advanced X-ray spectroscopies and computational studies. ChemCatChem. 12, 1325–1333 (2020) 31. Hanukovich, S., Dang, A., Christopher, P.: Influence of metal oxide support acid sites on cu-catalyzed nonoxidative dehydrogenation of ethanol to acetaldehyde. ACS Catal. 9, 3537–3550 (2019)

669

Stephen Purdy received his PhD in chemical engineering from Purdue University in 2019, where he studied platinum and palladium intermetallic compounds for light alkane dehydrogenation by synchrotron X-ray techniques. He is currently a postdoc in the diffraction group at the spallation neutron source, Oak Ridge National Laboratory studying catalyst structure by X-ray and neutron total scattering.

30

Laryssa Cesar received her PhD at Purdue University in 2019. She joined Dow Chemical Company shortly after in 2019. She currently works as a Senior Research Specialist in the Dow Performance Silicones Business. Her current research interests include silicone-related heterogeneous and homogeneous catalysis as well as development and scale up of new processes.

Guanghui Zhang received his PhD at Wuhan University in 2014. He spent several years as a postdoc researcher at Illinois Institute of Technology and Purdue University. He joined Dalian University of Technology in 2018. His current research interests include energy and environment-related catalysis and synchrotron-based in situ X-ray techniques.

Jeffrey Miller spent 25 years working in industry R&D for AmocoBP. In 2008 he became the heterogeneous catalysis group leader at Argonne National Laboratory and has been a professor in chemical engineering at Purdue University since 2015. His research interests include energy and environment-related catalysis, especially the characterization of catalysts using in situ synchrotron-based X-ray characterizations. Nicole LiBretto is a PhD candidate at the Davidson School of Chemical Engineering at Purdue University. She graduated with a Bachelor of Engineering from Stony Brook University in 2016. Her research is with Dr. Jeffrey T. Miller in developing new catalyst compositions for shale gas conversion.

Case Studies: Mapping Using X-Ray Absorption Spectroscopy (XAS) and Scattering Methods

31

Dorota Matras, Antonis Vamvakeros, S. D. M. Jacques, and Andrew M. Beale

Contents 31.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 671

31.2

X-Ray Scattering-Based Imaging . . . . . . . . . . . . . . . . . . . . . . . . 672

31.3

X-Ray Absorption Spectroscopy-Based Imaging . . . . . . . 676

31.4

X-Ray Coherent Diffraction Imaging . . . . . . . . . . . . . . . . . . . . 681

31.5

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684

Abstract

Understanding of complex structure-function relationships is crucial in designing catalytic materials with optimized properties. The past 20 years have seen significant progress in the development of imaging techniques (i.e., acquisition methodology, sample environment, data handling) for performing experiments under industrially relevant operating conditions (i.e., temperature, pressure, D. Matras School of Materials, University of Manchester, Manchester, Lancashire, UK Research Complex at Harwell, Harwell Science and Innovation Campus, Rutherford Appleton Laboratory, Didcot, Oxon, UK e-mail: [email protected] A. Vamvakeros Department of Chemistry, University College London, London, UK Finden Limited, Abingdon, UK e-mail: antony@finden.co.uk S. D. M. Jacques Finden Limited, Abingdon, UK e-mail: simon@finden.co.uk A. M. Beale (*) Research Complex at Harwell, Harwell Science and Innovation Campus, Rutherford Appleton Laboratory, Didcot, Oxon, UK Department of Chemistry, University College London, London, UK Finden Limited, Abingdon, UK e-mail: [email protected]

chemical environment). These can now provide invaluable insight into the nature and structure of active catalyst components. In this chapter a variety of chemical and structural imaging techniques are discussed, using exemplar recent studies where it has been investigated how the catalyst activity and stability can be affected by the interplay of micro
/macrostructure, distribution, and nature of active components/sites. Keywords

Tomography · Imaging · Diffraction · Scattering · Fluorescence · Spectroscopy · Absorption · X-rays · Coherence · Structure · Composition · Operando · Synchrotron

31.1

Introduction

Much of modern society has come to depend on the performance of solid catalysts. The rarely uniform 3D structure of solid catalysts requires multi-length-scale characterization (i.e., from atomic to millimeter scale and over the entire catalyst volume). The various heterogeneities present in the catalyst influence the overall performance, and in order to gain an insight into the complex structure-function relationships, it is crucial to study catalysts in situ/operando with spatially resolved techniques [1–7]. X-rays with energies typically ranging from 1 to 100 keV are employed to investigate the internal structure of bulk samples. X-rays can penetrate materials in a noninvasive way providing chemical information for both crystalline and amorphous components including elemental composition, local coordination, and electronic state as well as sample physical properties such as porosity, density, and tortuosity. The spatially resolved chemical imaging techniques can be divided into two main groups based on the data collection and subsequent analysis: (1) 2D scanning which results in a two-dimensional grid

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_31

671

672

(map) where every pixel corresponds to a projected/summed chemical signal along the X-ray beam path [8–13] and (2) computed tomography (CT) which represents a two-dimensional cross section of the sample where every voxel corresponds to a mathematically reconstructed chemical signal [14–18]. In parallel beam X-ray computed tomography, the sample is illuminated with an X-ray beam (full-field illumination) at different tomographic angles, typically covering a range from 0 to 180 . This type of geometry is used for conventional imaging (i.e., absorption and phase contrast) and absorption spectroscopy in transmission mode at synchrotron facilities; the voxel size in the reconstructed image is determined by the detector pixel size and/or additional optics applied in front of the detector. In the case of chemical imaging, a narrow and focused X-ray beam, also known as “pencil” beam, is typically used. In this approach the sample needs to be additionally translated across the X-ray beam at each tomographic angle; the combination of the translation step size and the size of the focused beam determines the voxel size in the reconstructed images. In both types of tomography, the realspace images are obtained through the application of reconstruction algorithms such as the filtered back projection, direct Fourier reconstruction, or iterative algebraic algorithms [19, 20]. There are various data collection strategies aiming to improve the temporal resolution of the tomographic measurements; for more detailed information, the reader is referred to other publications [21–24]. Tomographic imaging is considered a superior characterization technique over the conventional single point bulk measurements; the spatially resolved signals provide information regarding the distribution of the various components present in the catalyst cross section/volume. It also allows for the extraction of local signals and effectively removing the undesired reactor vessel signal, which can significantly complicate the data analysis process. With the use of hard X-rays generated at synchrotron facilities, the tomographic measurements can be applied to study materials in their industrial shape and size, without the additional step of sample preparation. Additionally, state-of-the-art detectors and highly advanced optic systems available at the synchrotron beamlines allow to perform the imaging experiments with high temporal (~ 10 ms per point) and spatial (< μm) resolution. This chapter aims to present some recent examples of chemical tomography studies applied to catalysis, based on scattering and spectroscopic techniques. The last section of this chapter discusses the application of imaging techniques using coherent X-rays to study the internal structure of catalytic materials and changes in the shape and strain of the nanomaterials induced during a chemical reaction.

D. Matras et al.

31.2

X-Ray Scattering-Based Imaging

Catalysis research often employs Bragg diffraction (X-ray diffraction) and more recently total scattering (pair distribution function) techniques to obtain structural information from long-range and short-range ordered materials (crystalline and amorphous/nanocrystalline, respectively). The X-ray diffraction (XRD) technique is suitable for characterization of crystalline materials; the diffraction patterns provide information regarding the sample composition (i.e., phase fraction) as well as physicochemical information related to the structure, crystallinity, strain, texture, and defects all of which are typically calculated through full-profile analysis using least squares minimization of a structure model (e.g., Rietveld analysis). The pair distribution function (PDF) may be considered as a complimentary technique to XRD, as it allows the investigation of the local structure of nanocrystalline (< 3 nm) and amorphous materials. In contrast to XRD, the PDF analysis works in real space (rather than reciprocal space) after calculating the Fourier transform of the total scattering measured up to a high Q range (typically up to ~30 Å
1). It is a measure of the probability of finding an atom at a specific distance from another atom which makes the technique appropriate for studying materials at the nanoscale (e.g., calculating nanoparticle size) yielding also atomic-scale information (bond distances). It is therefore of no wonder that both XRD-CT and PDF-CT have been proven to be invaluable tools for characterization of solid catalysts, functional materials, and devices under operating conditions [25–34]. As an example, in the study of Li et al. [35], the XRD-CT technique was applied to investigate the thermal stability of a novel type of micro-monolithic solid oxide fuel cell. The authors performed in situ several thermal cycling (from room temperature to 800  C) of the solid oxide fuel cell, collecting an XRD-CT cross section after each cycle to assess the structural integrity and distribution of the crystalline phases within the device (Fig. 31.1). These measurements revealed that although the device was exposed to harsh temperature conditions, the distribution of Ni, its lattice parameter, and thermal strain remained homogenous for the duration of the experiment (six thermal cycles), suggesting great mechanical robustness and resistance to thermal shock. Vamvakeros et al. [36] performed a multimodal chemical tomography experiment through simultaneous acquisition of combining XRD-CT and X-ray fluorescence-CT (XRF-CT) data from operating catalysts used for the oxidative coupling of methane reaction. More specifically, the authors studied the effect of La promoter on the activity and stability of a Na-Mn-W/SiO2 catalyst both at the reactor and at the single catalyst particle level. Under the operating conditions

31

Case Studies: Mapping Using X-Ray Absorption Spectroscopy (XAS) and Scattering Methods

a)

b)

Phase distribution maps NiO

YSZ

673

LSM

Red : YSZ Green: LSM Blue : NiO

SrO 1

0.5

0

c)

Weight percentages NiO

(%)

YSZ

(%)

LSM

(%)

SrO

(%) 7

50

50

80

40

40

60

5

30

30

40

4

20

20

20

6

3

Fig. 31.1 X-ray diffraction computed tomography of the fresh solid oxide fuel cell. (a) The phase distribution maps of NiO, Yttria-Stabilised Zirconia (YSZ), Lanthanum Strontium Manganese (LSM), and SrO as derived from the Rietveld analysis of the fresh Solid Oxide Fuel Cell (SOFC) XRD-CT datasets (color bar indicates intensity in arbitrary

units); (b) red green blue (RGB) image showing the distribution of YSZ (red), LSM (green), and NiO (blue); (c) weight % of all crystalline phases present in the SOFC. The scale bar corresponds to 0.5 mm. (Source: reproduced with permission from Li et al. [35], Copyright 2019, Nature Publishing Group)

(i.e., high temperature of 780  C and reaction mixture containing CH4 and O2), the Na2WO4 species were found to be volatile and migrate from the catalyst particles, accumulating at the catalyst surface and reactor vessel. The loss of active species during the OCM reaction was shown to be directly responsible for the lower catalyst activity [33] and hence the need for the effective stabilization of Na2WO4 species through a chemical promotion. The combination of diffraction and spectroscopic measurements showed that the La promoter effectively stabilized the volatile species, forming a crystalline phase of NaLa(WO4)2. In addition, the La species improved the dispersion and interaction of metallic species with the SiO2 support as well as suppressed the crystallization and evolution of SiO2 support during the hightemperature operation (Fig. 31.2). The main drawback of pencil beam tomography compared to parallel beam (full-field) tomography experiments is considered to be the lower temporal resolution. Currently, with the use of X-rays generated in the third-generation synchrotrons (i.e., high intensity and brilliance) and state-of-the-art detectors (e.g., Pilatus and Eiger series from Dectris), the limiting step is associated with the motors’ movements, as the sample needs to be both translated and rotated across the

illuminating X-ray beam. A significant development in the data collection strategy has been recently reported by Vamvakeros et al. [37]. In their study, a multicomponent Pd-Ni/CeO2-ZrO2/Al2O3 catalyst was studied during a reduction-oxidation (redox) experiment, using a new data collection strategy where both the translation and rotation axes were moving simultaneously. This allowed to minimize the “dead time” of the measurement and probe the entire volume of the catalyst bed (i.e., multiple XRD-CT cross sections) in a reasonable timescale by acquiring each XRD-CT dataset (7200 patterns) in less than 2 min. The evolution of the solid-state chemistry was tracked for the first time in 5D (i.e., three spatial dimensions, one diffraction dimension, and one temporal dimension/imposed chemical environment) during the reduction and oxidation stages, revealing significant heterogeneities present both across the catalyst particles and along the catalyst bed (Fig. 31.3). These are expected to have an important effect on the catalyst performance; an improved catalyst design can only be achieved by understanding the impact of these complex chemical gradients and structure-activity relationships in 3D. The first demonstration of the PDF-CT technique was reported on the in situ study of an industrial Pd/γ-Al2O3

31

674 Fig. 31.2 Phase distribution maps created based on the intensities of the scale factors of each crystalline phase. Right: relative changes of each component during the OCM experiment with the Mn-Na-W/ SiO2 catalyst (operating conditions 1–5). Scale bar corresponds to 0.75 mm. (Reproduced with permission from Vamvakeros et al. [36], Copyright 2020, Elsevier)

D. Matras et al.

Temperature

RT

Gases

800 °C

800 °C

800 °C

He

Air

CH4/air

(2)

(3)

10/1 (4)

Ratios (1)

RT

(5) 1

Cristobalite-high

0.8 0.6 1

Cristobalite-low 0.9 Tridymite-high

1 0.8 0.6 1

Tridymite-low 0.5 Quartz-high

1 0.8 0.6

Quartz-low

1 0.8 0.6 1

Mn2O3

0.5 1

Na2WO4

0.5 0 1

MnWO4

0.5 0 12345

catalyst pellet during calcination and reduction processes [26]. According to the XRD-CT measurements, the active catalyst component, present either as PdO or metallic Pd, demonstrated an egg-shell distribution. However, the PDFCT measurements revealed that in addition to the larger Pd particles (4.5 nm), the core of the catalyst pellet was rich in small Pd nanoparticles (1.4 nm). The results of this study illustrated that the choice of the characterization techniques is crucial for catalyst characterization; the distribution of the small nanoparticles could not be observed with XRD-CT, and without this information, a misleading conclusion could have been made regarding the structure-activity relationships. Senecal et al. [25] applied both XRD-CT and PDF-CT techniques to study the behavior of a Co/γ-Al2O3, industrial

catalyst pellet used for Fischer-Tropsch synthesis (FTS). Both techniques were able to follow the evolution of Co speciation during reduction (catalyst activation) and early stage of the FTS reaction and determine their spatial distribution within the catalyst pellet. More specifically, two types of Co species were observed in the catalyst sample: (1) small nanoparticles (< 6.5 nm) that strongly interacted with Al2O3 support and (2) larger nanoparticles that agglomerated and did not have strong interaction with the Al2O3 support. Furthermore, these Co species behaved in a different way during the reduction process, with the large nanoparticles being readily reduced to metallic Co and the smaller ones being present in the oxidized form as CoO. The distribution of CoO and metallic fcc Co, obtained with both XRD-CT and

31

Case Studies: Mapping Using X-Ray Absorption Spectroscopy (XAS) and Scattering Methods

675

5D evolution of the Ni species

NiO

NiO

b)

Top of the bed

Bottom of the bed 1

1

0.5

0.5

0

0 2

x = 0.9

x = 0.15

x = 0.2

x = 0.3

2

O2

Top of the bed

NiAl2O4

Bottom of the bed 1

0.5

0.5

0

0 2

RT H e H 2 O

NiAl2O4

1

x = 0.8

x = 0.6

2

H2

RT H e H 2 O

He

RT H e H 2 O

RT

RT H e H 2 O

a)

x = 0.8

x Ni

x=1

0

Fig. 31.3 5D chemical evolution of the Ni-containing species in the catalyst bed. (a) Phase distribution volumes of NiO, NiAl2O4, and Ni as obtained from the Rietveld analysis of the 3D XRD-CT data collected at the four different operating conditions. The values in the color bar axes have been chosen to achieve the best possible contrast. (b) Solid-state

evolution of the NiO and NiAl2O4 phases at the bottom and top of the catalyst bed during the redox experiment (i.e., XRD-CT scan 1 and 30). The results presented in this figure correspond to the Rietveld analysis of ca. 1.2  106 diffraction patterns. (Reproduced with permission from Ref. [37], Copyright Nature Publishing Group, 2018)

PDF-CT measurements during the three representative stages, is shown in Fig. 31.4. It can be clearly seen that during the reduction of the catalyst there are significant differences in the distribution of CoO and Co obtained with XRD-CT and PDF-CT; at the beginning of reduction, the PDF-CT suggested that CoO was more uniformly dispersed in the catalyst pellet, and this could be explained by the presence of smaller nanoparticles not detected with the XRD-CT technique. Similarly, at the end of the reduction stage, a strong signal of metallic Co at the periphery of pellet was again observed with PDF-CT, suggesting formation of a large number of small Co nanoparticles. Finally, the spatial distribution maps of both CoO and Co during the FTS reaction were accurately described by the XRD-CT and PDF-CT measurements, both showing the reoxidation of metallic Co to CoO at the sample periphery. The oxidation of small nanoparticles of metallic Co had a detrimental effect on the catalyst performance, as the catalyst became active for the undesired water-gas shift reaction and lowered the selectivity for the C2+ products.

It is important to note that there are some challenges associated with the PDF-CT technique compared to XRD-CT: (1) the analysis of PDF-CT data is not trivial and can become complicated when a multicomponent catalytic system is studied, and (2) significantly longer acquisition time is necessary to obtain good-quality patterns when compared to XRD patterns (i.e., with good signal-to-noise ratio). The high Q range necessary for the total scattering analysis can be achieved with high-energy X-rays. In scattering-based measurements, there are also some extra factors one has to take into account when designing such an experiment. For example, while very hard X-rays (e.g., >70–80 keV) are able to probe even relatively large samples (i.e., several cm in diameter), several artifacts can form in the acquired data/reconstructed images such as selfabsorption and parallax. The former can be accounted for by collecting conventional X-ray absorption-contrast CT data of the sample before the in situ/operando experiment and applying them in self-absorption correction algorithms [31]. The latter, the parallax artifact, is more complex and results in loss

31

676

D. Matras et al.

Fig. 31.4 Reconstructed 2D integrated Fourier transform intensity maps based on the intensity of the Co–Co scattering features at 3 Å (CoO) and 2.5 Å fcc Co observed under reduction and during FTS conditions as a function of temperature and time. (Modified and reproduced with permission from Ref. [25], Copyright American Chemical Society, 2017)

XRD CoO

Co fcc

CoO

Co fcc

x4

Reduction

450 °C

FT

310 °C

250 °C

of physicochemical information (e.g., accuracy in lattice parameter and crystallite size determination) when X-rays scattered at a scattering angle 2θ arrive at multiple detector elements. The observed diffraction peaks can become broader and can shift toward different scattering angles. A solution to this problem, employing a new reconstruction algorithm, was proposed by Vamvakeros et al. [38]. It is also important to take into consideration the relationship between the illuminating X-ray beam and the size of crystallites present in the sample. Samples consisting of large crystallites (i.e., large single crystals/grains) yield spotty single-crystal-like diffraction patterns rather than powder diffraction rings. Filters have been developed to suppress this spottiness in the diffraction data [39], but in severe cases, other types of analysis are needed alternative to XRD-CT such as scanning 3D XRD [40, 41] which can reconstruct the shape, orientation, position, and strain of the large grains in the sample cross section.

31.3

PDF

X-Ray Absorption Spectroscopy-Based Imaging

In X-ray absorption spectroscopy, the illuminating X-ray beam primarily interacts with the electrons from the atom core levels, and if the energy of X-ray photons is above a certain value, these electrons are ejected from the atom subshells. This phenomenon leads to the formation of a

discontinuity in the absorption spectrum, known as the absorption edge, and it occurs at characteristic X-ray energies which are element specific. The analysis of the X-ray absorption near-edge spectrum (XANES) provides important information regarding the oxidation state and local environment of the absorbing/target element [42–46]. At the same time, the ejection of an electron from the core levels sets the atom in an excited state of higher energy. The atom can return to the ground state by filling the created electron hole with an electron from higher levels. This is accompanied with an emission of X-ray radiation, known as X-ray fluorescence, in which energy can be used to identify the sample elemental composition [47–53]. The data collection for the XANES-CT technique can be performed with X-rays of different characteristics which lead to the following variations of the technique: 1. Scanning approach with a monochromatic X-ray pencil [54] – in this approach, the tomographic cross-section data need to be collected for multiple energy points along the XANES edge. 2. Scanning approach with a polychromatic X-ray pencil beam [55] – full XANES spectra are obtained per acquisition point, and this can significantly reduce the measurement time. To the authors’ knowledge, this data collection strategy is yet to be applied in the catalysis studies. 3. Full-field 3D XANES-CT [56] – it is an approach similar to the pencil beam approach where multiple energy points

31

Case Studies: Mapping Using X-Ray Absorption Spectroscopy (XAS) and Scattering Methods

need to be scanned across the XANES edge, but here information regarding the whole sample volume is acquired with projections. The resulting resolution in the reconstructed sample volume depends on the detector properties and optics system applied. 4. Absorption-contrast [45, 57–59] – in this approach, the absorption image is being collected for two different energy values, before and after the absorption edge of chosen element. This approach is believed to offer and improve contrast for the chosen element in samples suffering from weak absorption contract (i.e., consisting of weekly absorbing elements or similar features/elements). The first example of a XANES-CT study on a catalyst, using the pencil beam approach was reported by Schroer et al. [54]. The authors acquired a cross section of a CuO/ZnO catalyst oxidized at room temperature to investigate the oxidation state of Cu species. Through the fitting of reference spectra of Cu species to the spectra present in each reconstructed voxel, it was possible to identify and map the distribution of metallic Cu and Cu2O within the catalyst sample. Another approach was demonstrated in the work of Nelson et al. [56]; here XANES spectroscopy was combined

with X-ray nanotomography employing the full-field mode and the zone plane objective. The 3D mapping of Ni oxidation state was obtained using a composite Ni/NiO sample, and the measurements were performed at the Ni K-edge energy. With this approach it was possible to visualize the distribution and oxidation state of Ni species with high spatial resolution (due to the application of Fresnel zone plate objective lens, field of view between 5 and 30 μm) which is very important when considering applications in heterogeneous catalysis and electrocatalysis. Matsui et al. [60] performed the first operando XANESCT experiment of a practical membrane electrode assembly (MEA) with a Pt cathode. The aim of this study was to map the distribution and oxidation state of the Pt cathode catalyst in the fuel cell under operating conditions of accelerated degradation. The results of this study revealed a homogenous degradation of Pt from the cathode catalyst layer, which dissolved and migrated toward the Nafion polymer membrane (Fig. 31.5). Since there is no versatile imaging technique able to provide all the desired chemical and physical information when investigating a material system, it has been realized that a practical approach is to employ multimodal imaging.

Interface of cathode catalyst layer and nafion membrane

Center of cathode catalyst layer X

Z = 10 μm

Center of nafion membrane

Z = 20 μm

Z = 45 μm

Z = 20 μm

Z = 45 μm

Pt density

(ii-CS) 600 μm

(ii)

677

a)

Y Z = 10 μm

b) c)

Fig. 31.5 (ii) 3D images of the Pt density and (ii-CS) their crosssectional images at Z ¼ 10, 20, and 45 mm reconstructed by in situ CT-XANES before and after the 20,000 Accelerated Durability Tests (ADT) cycles. (a) Observed line-like structures, (b) areas that the Pt

Pt density

600 μm

catalyst aggregated, and (c) areas that the Pt catalyst was lost (dashed circles). (Modified and reproduced with permission from Ref. [60], Copyright Wiley-VCH Verlag GmbH & Co., 2017)

31

678

D. Matras et al.

Here, the scattering and spectroscopic imaging data are collected simultaneously (or in some cases sequentially). In the work of Price et al. [61] an industrial FTS catalyst 5% Ti/10% Co/1% Re on SiO2 was studied operando using multimodal imaging by employing the XRD-CT, XRF-CT, and absorption-contrast CT techniques. The results of this study demonstrated the importance of the catalyst preparation protocol, more specifically focusing on the impregnation sequence of the different compounds. Although different impregnation sequences yielded materials with the same chemical compositions, the size and the distribution of crystalline species were significantly altered. For instance, it was shown that impregnating the silica support with the Ti precursor before the Co and Re precursors suppressed the diffusion/migration of Re and Co species to the core of catalyst particles. At the same time, the reduced Co nanoparticles were smaller and more strained; this phenomenon attributed to the higher degree of intergrowth between hexagonal and cubic Co particles (Fig. 31.6). Although during the FTS reaction the nanoparticles of Co underwent partial reoxidation to CoO, the catalyst activity remained stable which demonstrated the importance of the intergrowth in the Co nanoparticles. On the other hand, when the Co and Re precursors were deposited before the Ti precursors, the Co and Re species freely diffused through the support material. In this case, larger but less strained Co nanoparticles were formed, and the catalyst exhibited higher stability and

Fig. 31.6 Conventional catalyst structure during FTS. (a) XRF-CT reconstructions showing elemental distributions for the conventional catalyst during FTS at 2-bar pressure. Green, Ti; blue, Co; orange, Re. Absorption-CT reconstruction (gray) also shows the capillary wall surrounding the particles. (b) XRD-CT reconstructions of the conventional catalyst revealing the phases present. (c) Average crystallite size per pixel for each phase identified. Each pixel is 5 mm  5 mm. Gas flow was 4 ml min
1 5% H2/He and 2 ml min
1 5% CO/He at 200  C (H2/CO 2:1 and 95% inert). (Reproduced with permission from Ref. [61], Copyright, AAAS, 2017)

a)

selectivity for longer-chain hydrocarbons. These results clearly showed that the catalyst activity and selectivity for specific products can be tuned by altering the physiochemical properties of the catalytic material; it was shown that the spatial distribution of both crystalline and noncrystalline components plays an import role in the properties of the obtained catalyst. Sheppard et al. [62] applied in situ multimodal imaging of a core@shell Cu/ZnO/Al2O3@ZSM5 catalyst with XRD-CT, XRF-CT, and absorption-contrast CT techniques during the calcination, activation (reduction), and dimethyl ether (DME) synthesis reaction using synthesis gas. The results obtained during the activation stage showed that the formation of Cu2O species (Cu+) at the core-shell interface was a result of an interaction between metastable Cu+ species and the zeolite shell (Fig. 31.7). This rather unexpected behavior led to the formation of a catalyst with the core being significantly reduced to metallic Cu and clusters of Cu2O phase being present at the core-shell interface. When the reaction mixture was introduced to the reactor, the catalyst structure changed significantly, yielding a range of species with mixed Cu oxidation states and the aforementioned Cu2O, this time being present in the form of small nanoclusters in the core of catalyst particle. This example of a core@shell catalyst clearly showed the importance of performing the measurements in the spatially resolved manner, as otherwise such heterogeneities would have been lost in bulk measurements.

Ti

Co

Re

Absorption-CT

XRF & absorption

200 μm

b) XRD

c)

CoO

Anatase TiO2

Co cubic

Co intergrown

Fitted XRD

2.0 nm

3.5 nm

3.0 nm

10.0 nm

7.0 nm

12.0 nm 2.0 nm

10.0 nm

31

Case Studies: Mapping Using X-Ray Absorption Spectroscopy (XAS) and Scattering Methods

a) Other Cu phases (2q 19.8–20.1°)

Cu2+ core (2q 18.1°)

25 μm

b) Cu2+ core (2q 18.1°)

Zeolite shell (2q 14.1°) 25 μm

Fig. 31.7 Core@shell particle under oxidizing conditions, with μ-XRD-CT rendering of (a) CuO core (95 vol %) and reduced fractions (5 vol %); (b) core and zeolite shell occupying distinct 3D space. (Modified and reproduced with permission from Ref. [63], Copyright, American Chemical Society, 2017)

In another work by Price et al. [64], XANES-CT and XRF-CT were applied to study an industrial single Pt-Mo/C catalyst particle for the selective hydrogenation of nonfunctionalized nitrobenzene compounds. Both imaging techniques showed a good dispersion of the Pt and Mo species on the surface as well as in the internal pore structure of the carbon support. Surprisingly, the promotion of Mo with Pt was found to occur as a result of close proximity between these two species rather than through the direct interaction and formation of a bimetallic Mo-Pt solid solution. XRF-CT combined with X-ray ptychography-CT was applied to study the fragmentation in a single particle from an industrial Ziegler-Natta catalyst [65]. This complex catalyst material, consisting of MgCl2 support, TiCl4 active site precursor, and trialkylaluminum cocatalyst, is used in the production of polyethylene and isostatic polypropylene. During the polymerization process, the fragmentation of the catalyst framework is necessary to sustain the catalyst activity as the formed polymer products tend to adsorb on the catalyst surface and prevent monomers from diffusing toward the catalyst active sites. The fragmentation of the catalyst framework is possible to occur following two different models: (1) shrinking core/layer by layer where the particles

679

from the surface are fragmenting until reaching the core and (2) continuous bisection where a particle breaks internally into smaller clusters. The XRF-CT measurements revealed the formation of isolated clusters of Ti and Cl, encapsulated by the formed polymer product. The separation of the large clusters was proposed to follow the fragmentation model of continuous bisection. At the same time, the presence of small clusters (i.e., with a volume smaller than 0.135 μm3) in close proximity to larger Ti clusters was also observed, suggesting the subsequent layer-by-layer fragmentation type occurring in the large clusters (Fig. 31.8). Gambino et al. [66] investigated the poisoning effect of Ni and Fe in the catalyst for the fluid catalytic cracking (FCC) process with a combined XRD, XRF, and XANES-CT (Fig. 31.9). The FCC catalyst, composed of an alumina-silica matrix of mixed porosity, clay binder, and zeolite, is prone to irreversible deactivation due to accumulation of poisoning metallic species and to hydrothermal degradation of the active zeolite phase. In this study, two different catalyst particles (i.e., with respect to their Ni content) were extracted from an industrial reactor after deactivation, and the relationship between the metallic species and the catalyst structure was investigated ex situ. The results of the XRF and XRD-CT measurements showed that the catalyst shell consisted of Ni, Fe, and Al2O3, while the core of catalysts consisted mainly of Ni and Al2O3 in the form of highly concentrated regions, suggesting a high correlation between the Ni and Al2O3. The XANES measurements suggested that the Ni species are present in the form of metallic spinel (NiAl2O4) most probably also containing Fe species. In addition, the analysis of the XRD patterns revealed higher dealumination of the zeolite phase when the catalyst particle contained Ni species as well as structure rearrangement and reduced mesoporosity of the alumina matrix, whereas the Fe and Ni species are deposited in the core of catalyst particle. The spectroscopic measurements, especially the XANES measurements, can provide important information regarding both crystalline and nanocrystalline materials. However, it is important to note that the analysis of XANES data is mainly based on the fitting of reference compound spectra to the experimental data (linear combination fitting), and to authors’ knowledge, in contrast to the XRD/crystallography data, there are no large and standardized databases for X-ray absorption fine structure (XAFS) spectra, and one has to measure multiple reference compounds prior to the main experiment. In addition, the XANES measurements of multicomponent materials can be very time-consuming as each element has its own characteristic absorption edge that needs to be scanned. Sample self-absorption problems often impose restrictions on the reactor size (i.e., single catalyst particle), and at the same time, the implementation of self-absorption correction algorithms in the XRF-CT data is not a trivial task [67]. Also, the interference in the absorption spectra of

31

680

D. Matras et al.

Fig. 31.8 Reconstructed 3D volume rendering of a propylenepolymerized Ziegler-Natta catalyst particle, 41 μm in diameter. (a) Electron density reconstruction (gray scale) with a voxel size of 43.2  43.2  43.2 nm3. (b) Extracted volume showing the position of the Ti species (in red) within the polymer-catalyst composite particle. (c) Elemental distribution of the Ti species (red color map) with a voxel size of 150  150  150 nm3. (d) Elemental distribution of Ti and Cl (in green) within the composite particle (only external surface rendered). (Reproduced with permission from Ref. [65], Copyright American Chemical Society, 2020)

a)

b)

50 μm

μm

45 μm

28

53 μ m

c)

d)

Ti Cl High I

Low I

a)

b)

Poisoning metals ECAT1-F1

ECAT2

Crystallographic components ECAT2

ECAT1-F1

923

20

0.043

0.05

FAU [111]

Ni

20 μm

10 μm

10 μm

20 μm

0

0

0

0

296

284

0.027

0.031

γ-Al2O3 [400]

Fe

20 μm 0

10 μm

0

20 μm 0

10 μm

0

Fig. 31.9 Obtained information about (a) poisoning metals, in particular Fe and Ni; (c) crystallographic phases of the different FCC catalyst components. (Modified and reproduced with permission from Ref. [66], Copyright Wiley-VCH Verlag GmbH &Co., 2020)

31

Case Studies: Mapping Using X-Ray Absorption Spectroscopy (XAS) and Scattering Methods

elements with similar absorption edges can significantly complicate the analysis of the acquired spectra. However, it is important to point out that sometimes the XAFS measurements are the only viable way to identify the structure or contents of a sample, especially when the sample is present in a form close to molecular species. The current trend points toward the development of multimodal imaging capabilities at synchrotron beamlines where several techniques can be simultaneously employed to collect different types of chemical information.

31.4

compressive strain occurring at the crystal edges. This was considered as an indication of the position of the catalyst active sites at the nanocrystal. After the introduction of methane, the deformation of the crystal continued with the contraction propagating further into the crystal interior due to the bonding of methane molecules to adsorbed oxygen atoms (Fig. 31.10). The in situ BCDI results were in good agreement with reactive molecular dynamic simulations performed on this system.

X-Ray Coherent Diffraction Imaging

The previous sections of this chapter focused on chemical imaging, where each pixel/voxel contains a chemical scattering or spectroscopic signal. However, physical properties such as surface area, porosity, shape, and mechanical stability play also an important role in the performance of catalytic materials. To yield this type of information at the nanoscale level, high-resolution imaging techniques such as electron microscopy have been previously employed [68, 69]. However, the main drawback of electron microscopy is its short penetration depth, and the obtained information is limited to surface features from thin samples (4 keV or wavelength 0 nuclei that are magnetically coupled to the paramagnetic center. At times, the substitution can narrow the spectrum by removing a hyperfine interaction, such as the deuteration of organics reducing broadening from unresolved hyperfine interaction with protons. Conversely, interactions can also be detected when isotopic substitution causes additional broadening, as in the case of H217O [52, 53]. In cases where the natural abundance distribution of an element is among several isotopes, the spectrum can be simplified by using an isotopically pure source. This has been employed in studies of Cu (II) centers; using 63Cu instead of the natural abundance mixture of 63Cu/65Cu reduces inhomogeneous broadening and can allow for detection of superhyperfine interactions [52]. However, the most common implementation of isotopic substitution is to track substrates as they are converted to products, or to positively identify binding. Figure 38.6 shows such an example, where the binding of hydrogen to a Fe

a)

H

Ph N

P P N

H

Fe I



(I) complex is confirmed by repeating the experiment with deuterium [54]. The EPR spectrum of the hydrogen complex was least squares fit using Easyspin [19], then the same parameters were used to simulate a spectrum with the only change being the protons changed to deuterium and the hyperfine interaction energy being divided by the ratio of the proton and deuterium gyromagnetic ratios. All other parameters such as g values and hyperfine coupling to other nuclei were left at the values determined by the least squares fit. The simulation fits the experimental spectrum with D2 in the headspace extremely well, confirming that the hyperfine detected is from a bound H2 (or D2), Fig. 38.7. The catalytic system used for preferential CO oxidation (PROX) was investigated by Wang et al. using both multifrequency ex situ EPR and operando EPR [53]. The question under review was whether the CuO/CeO2 catalyst operated using a synergistic mechanism involving the redox cycling of Cu(0/I)/Cu(II) and Ce(III)/Ce(IV) or if only copper was involved (the direct mechanism), with copper being reoxidized directly by oxygen after being reduced by CO. Preliminary ex situ EPR spectra showed a variety of Cu(II) species: monomers, dimers (S ¼ 1), and tetramers (S ¼ 2). The dimers were identified through splitting associated with strong dipolar coupling and by the semi-forbidden



b)

Ph P Ph Ph P

N

Ph

3400

3200 Field (G)

A1H = 35 Hz, A31P = 5572 Hz

P P N

Ph

Fe-H2, 2-MeTHF, 80 K

3000

D

Ph

Fe I

D

Ph P Ph P Ph Ph

Ph

Fe-D2, 2-MeTHF, 90 K

Exp

Exp

Sim

Sim

3600

3000

3100

3200

3300

3400

3500

3600

Field (G) A2D = 5.4 Hz, A31P = 5573 Hz

Fig. 38.7 The effect of isotopic substitution. (a) EPR spectrum of an Fe(I) compound with hydrogen bound. (b) The spectrum of the same compound but with deuterium bound

38

880

half-field signal [55]. The tetramers were identified using Q-band EPR and a technique called nutation spectroscopy. This method can distinguish overlapping spectra from species with different total spin angular momenta. The rate of nutation of the magnetization of a spin system depends not only on the impinging RF power, but is also dependent on S. In situ experiments were performed using a quartz plug flow reactor to contain the 20% wt. CuO/CeO2 sample in the EPR resonator, with the output directed into a gas analyzer. Introduction of either N2 or O2 did not change the EPR spectrum of the catalyst in the temperature range of 293–473 K. However, 1% CO in N2 immediately caused the spectrum of all Cu species to decrease at a temperature of 453 K. After 30 minutes, only a portion of the spectrum of aggregated copper species remained incalcitrant to reduction. While the spectrum was reducing in amplitude, the downstream gas analyzer showed only CO2. Upon switching the gas stream to 1% O2 the copper spectrum returned to full intensity. When the gas stream was instead switched to N2, the copper spectrum also recovered to a small extent presumably from the reduction of Ce(IV) by a reduced copper species. Together these data were consistent with both the synergistic and the direct mechanism playing a role. The experiments were repeated, but with 17O2 replacing natural abundance O2. The hyperfine coupling to 17O is usually too small to resolve individual peaks, but it can result in measurable broadening when bound directly to copper [52]. With repeated cycles of CO followed by 17O2 the proportion of reaction cycles for each mechanism could be calculated. For example, with the first cycle of the synergistic mechanism, the copper reoxidized by Ce16O2 will be Cu16O while those directly reoxidized by 17O2 will be Cu17O. Statistical analysis of the change in Cu17O content with each cycle allowed the authors to calculate that the synergistic mechanism was responsible for 80% of the catalytic activity. Ex situ electron nuclear double resonance (ENDOR) performed at W-Band (~95 GHz) confirmed that 17O had entered CeO2 framework, validating the role of the synergistic mechanism.

38.2.3 Metals and Free Radicals At times a reaction can involve several oxidation states of a metal, some paramagnetic and some diamagnetic. Grauke et al. (2019) used operando EPR in conjunction with in situ UV-vis, IR, and XANES/EXAFS spectroscopies to monitor the performance of the homogeneous catalysis of some chromium catalysts toward the oligomerization of ethylene [56]. These catalysts are often modified by an aluminum containing cocatalyst. The authors explored the effect of the R-group on cocatalysts of the for AlR3. Operando experiments were performed in a heavy-wall quartz reactor connected to a gas circulation system that bubbled ethylene through the reaction solution. The EPR spectra revealed

E. Walter

varying amounts of the activated Cr(III) precursor and an inactive Cr(I) species. An inverse correlation of the Cr (I) species and the activity of the catalyst displayed the utility of the method for evaluating cocatalysts. In situ measurements by UV-vis and EXAFS supplemented this by added data for the EPR-silent Cr(II) active state. This Cr(II) is part of a redox cycle with Cr(IV), giving the system four redox states in operando conditions, displaying the need for a multimodal in situ approach for these experiments. Another group also used in situ EPR to study chromiumbased catalysts for the oligomerization of ethylene. Morra et al. used EPR at both the most common frequency, X-band or ~9.5 GHz, and higher-frequency Q-band, ~35 GHz, to study a Cr/SiO2 Philips catalyst [57]. They found that the temperature at which the ethylene was introduced to the catalyst strongly affected the distribution of chromium oxidation states. The use of multiple EPR frequencies aided in sorting out the mixture of high-spin and low-spin states, as well as variations in the zero-field splitting of the high-spin states. High-frequency EPR is often advantageous when dealing with large zero-field splittings; when the Zeeman energy term is dominated, the effects of the zero-field splitting can be mathematically modeled as a perturbation. In a multimodal operando study of a homogeneous catalytic system, Yin et al. combined photoluminescence with EPR to monitor the TEMPO-mediated oxidation of benzyl alcohol to benzaldehyde [58]. In these systems, TEMPO redox cycles with its reduced form, TEMPOH. The authors designed and synthesized a bifunctional molecule that contains a paramagnetic TEMPO moiety, that when reduced to the diamagnetic TEMPOH then becomes fluorescent. In this way, both redox forms of the catalyst were detectable in situ. Operando experiments were performed using this new TEMPO derivative and BAIB (Bis(acetoxy)iodobenzene or phenyliodine(III) diacetate) to oxidize benzyl alcohol. Able to detect all TEMPO species, they were able to discern the three processes involved. First the disproportionation of TEMPO to TEMPOH and its oxammonium salt. Second was the oxidation of the alcohol by the oxamonium salt of TEMPO yielding a ketone or aldehyde and TEMPOH. And lastly, the rejuvenation of TEMPO from TEMPOH and BAIB. Qi et al. examined the reaction a β-O-4 model compound over Ni/γ-Al2O3 using a combination of in situ MAS NMR and EPR, and isotopic labeling [59]. Benzyl phenyl ether (BPE) was labeled with 13C at both the benzylic and phenolate-1 positions, providing a convenient way of following the reaction using in situ high-temperature/high-pressure MAS NMR. Samples containing BPE, Ni/γ-Al2O3 catalyst, and isopropanol were loaded in WHiMS MAS rotors [60] and then pressurized with either H2 or D2. The MAS spectra were acquired at temperatures ranging from 150–175 C, and showed the formation of products including phenol, toluene, and cyclohexanol. However, the reaction showed a temperature-dependent induction period with no reaction

38

Electron Paramagnetic Resonance (EPR)

881

taking place for 43 minutes at 150 C. Reduction of the catalyst was suspected, so EPR experiments performed to examine the oxidation state. The results of variable temperature experiments were consistent with Ni(0) nanoparticles showing superparamagnetic behavior – as the temperature was lowered the spectra shifted to lower fields and broadened. A sample of the catalyst was then placed in an EPR flow cell, and in situ reduction with flowing hydrogen was performed. Integration of the spectra showed an increase in signal intensity after reduction consistent with the conversion of a surface monolayer of Ni(II)O to Ni(0). These results were collaborated with temperature-programed reduction (TPR) and HAADF-STEM imaging. Operando X-ray absorption near-edge spectroscopy (XANES) and EPR were used by Lu et al. to observe the single-electron reduction of a copper catalyst by benzenesulfinic acid (PhSO2H) [61]. The reduction of copper bromide (CuBr2) by PhSO2H was monitored by EPR, with the EPR spectrum of Cu(II) rapidly disappearing at room temperature. XANES spectra confirmed that the Cu(II) had been reduced to Cu(I). The effect of anions on the copper catalyst was explored, with the result that copper compounds with coordinative saturation did not show reduction by EPR. Since the copper was undergoing single-electron reduction, sulfonyl radicals should be generated and be available for reaction. A number of arylacrylic compounds were tested, with β -keto sulfones being produced. e

Gas

b) 3 2.75

Potential (V)

a)

Radicals can be generated electrochemically, and in situ EPR cells for electrochemical experiments are available commercially. In many cases the experiment may necessitate the construction of custom cells, particularly for studies involving detection of radicals at the electrode rather than in the solution. In addition to the sample cell, a potentiostat or galvanostat is required, these can be of a conventional design, however all hardware for making connections near the spectrometer need to be nonmagnetic. One powerful method of generating data is to collect in situ EPR spectra while performing a simultaneous electrochemical experiment, such as cyclic voltammetry. Then features in a voltammogram can be correlated with the appearance or disappearance of an EPR spectrum. Wang et al. examined the just such a correlation using an in situ electrochemical cell to study the chemical processes involved in the charge/discharge of a lithium sulfur cell [62]. Simultaneous cyclic voltammetry and in situ EPR revealed that the signal from sulfur anion radicals (S3) peaked during the cathodic scan and did not return during the anodic scan, Fig. 38.8. Two different reaction pathways are used depending on the direction of the scan, this asymmetry is caused by differences in the solubility of different species in the series of products as elemental sulfur is converted to Li2S.

2.5 2.25 2 1.75 1.5 1.25 1

Amplitude S3 EPR spectrum

Fig. 38.8 In situ electrochemistry. (a) An example of a homemade in situ electrochemical cell. The glass flat cell has an internal width of 8 mm and an internal thickness of 0.8 mm. This cell has access for both electrical inputs as well as gas. (b) Amplitude of an EPR spectrum during a cyclic voltammetry experiment. Trace in red shows the voltage ramps, while the trace in black shows the amplitude of the EPR spectrum of a sulfur anion radical at those voltages

38.2.4 Electrochemistry and EPR

400

500

600 Time (minutes)

700

800

38

882

E. Walter

An example from Ali et al. shows the power of this method in their study of the presumed electrocatalytic oxidation of NADH [63]. The authors functionalized flexible carbon fiber electrodes with quinones using two different methods: one pot and electrochemical grafting. Comparing unfunctionalized electrodes with the two functionalized electrodes shows that the amount of radicals detected correlates with NADH oxidation, suggesting that the quinone groups on these electrodes act electrocatalytically rather than as redox mediators. Neukermans et al. reported on the development of continuous flow in situ electrochemical reactor that is suitable for the screening of electrocatalysts [64]. This is an interesting design, as it combines elements of an in situ electrochemical cell with a plug flow reactor. It is a three-electrode electrochemical cell supplemented with a channel for substrate flow that can be fitted to a standard EPR resonator. During use, electrochemical control was provided by an external potentiostat. The flow capability allows for continuous renewal of reactants, which can then be used to explore the effect of mass action on the results. This setup was used to explore systems with both organic solvents (benzoquinone [BQ] in acetonitrile) and with aqueous solutions (methyl viologen [MV] in water). Both BQ and MV radicals were detected at the expected potentials. The Nocera catalyst [65] is a cobalt/phosphate derivative formed on electrodes in situ prior to its use in water oxidation. Kutin et al. described a new spectroelectrochemical cell designed to determine the cobalt oxidation state at various potentials [34]. Since the EPR spectra needed to be obtained at cryogenic temperatures, while the cell had to be electrochemically cycled at room temperature, the cell was brought to the desired electrical potential and then flash frozen in liquid nitrogen. In this way the EPR spectrum of cobalt could be collected at a series of applied potentials, on the same electrode. Previous studies had made separate electrodes for each potential. After bringing the electrode to the desired potential, some of the material was removed by mechanically scraping it off, then placing it in an EPR tube. The advantage of the in situ method includes no variation in preparation (i.e., accidental exposure to oxygen), having the same amount of sample each time to allow direct comparison for quantitation, and having no difference in sample history (i.e., no difference in the dimensions of the electrode, or differences in electrical contact, etc.).

38.3

The Future of In Situ EPR

Many future advances in the study of catalysis using in situ EPR are currently being used in other fields, and are only awaiting the adaption to catalytic systems. One example is EPR imaging, which has mostly been applied to biological

questions, but with a few ex situ examples in the field of catalysis [66, 67]. With the availability of commercial instruments, more applications should soon follow. For instance, Spitzbarth et al. used a combination of time-, spatial-, and spectrally resolved EPR to study the diffusion of molecules in the nanopores of aerogel [68]. They used the spin labels TEMPO and trityl in ethanol solutions. Analysis of the line shapes allowed the measurement of diffusion rates and of spatial distribution of two different molecules simultaneously. Pulsed EPR is already widely used in the study of catalysis, however the vast majority of experiments require cryogenic temperatures. Two advances may make some in situ pulsed EPR experiments possible. First is the development of spin labels with long T2 times at ambient temperature [69–71]. While the current driver for their development is the study of protein structure, but just like more traditional spin labels, they may also find applications in catalysis. The other advance is the development of high-field instruments with large (for frequency) nonresonant sample holders and very short dead times [72, 73]. For the same amount of data collection time after a pulse (limited by T2) an instrument that is operating at 10–30 times higher frequency will see 10–30 times more nuclear modulations than a traditional X-band instrument. In addition to high-frequency EPR, multifrequency EPR also has unexploited possibilities for in situ studies. Several papers reviewed here used Q-band (~35 GHz) or W-band (95 GHz), but as supporting ex situ techniques for in situ studies at X-band. Xu et al. used the combination S-, X-, Q-, and W-band ex situ EPR to study silica-supported Ni/Pt catalysts. These different frequencies are sensitive to different dynamic ranges and to different couplings with other magnetic species. Lastly, while we have seen several multimodal in situ experiments combining simultaneous optical or electrochemical analysis with EPR, there are other techniques that are not only compatible with EPR but provide complimentary information. These include X-ray or other beam-line techniques, mass spectrometry (for plug flow cells), and potentially NMR/DNP. Nuclear magnetic resonance (NMR) is a natural partner with EPR, in that it is sensitive to diamagnetic compounds, giving the combination of the two the potential to detect all relevant species. However, it typically uses magnets >10 time higher in field than EPR, making multimodal instruments rare. However, dynamic nuclear polarization (DNP) enhances NMR sensitivity by transferring polarization from paramagnetic species that have been irradiated at their EPR frequency. That makes these instruments, which are now commercially available, essentially combined EPR/NMR instruments and therefore have the potential to be modified for multimodal in situ experiments. Of particular interest is the possibility of using endogenous metal ions as the source of the polarizing radicals [74–76]. In addition to hardware, an important component of multimodal studies is many times labels or tags that are

38

Electron Paramagnetic Resonance (EPR)

sensitive to multiple techniques [58]. It is likely that new examples of these multimodal agents may be imported from biochemical studies [77, 78], much in the same way as have spin traps and spin labels been adapted to material science and catalytic studies.

References 1. Lunsford, J.: EPR methods in heterogeneous catalysis. In: Catalysis, pp. 227–256. Springer (1987) 2. Louis, C., Lepetit, C., Che, M.: EPR characterization of oxide supported transition metal ions: relevance to catalysis. In: Radicals on Surfaces, pp. 3–38. Springer (1995) 3. Sojka, Z.: Molecular aspects of catalytic reactivity. Application of EPR spectroscopy to studies of the mechanism of heterogeneous catalytic reactions. Catal. Rev. 37, 461–512 (1995) 4. Werst, D., Han, P., Trifunac, A.: Radiation chemical studies in zeolites: radical cations and zeolite catalysis. Radiat. Phys. Chem. 51 (1998) 5. Sojka, Z., Che, M.: Catalytic chemistry of transition metal ions on oxide surfaces. A molecular approach using EPR techniques. C. R. Acad. Sci. 3, 163–174 (2000) 6. Labanowska, M.: EPR monitoring of redox processes in transition metal oxide catalysts. ChemPhysChem. 2, 712–731 (2001) 7. Van Doorslaer, S., Murphy, D.M.: EPR spectroscopy in catalysis. In: EPR Spectroscopy, pp. 1–39. Springer (2011) 8. Drescher, M.: EPR Spectroscopy: Applications in Chemistry and Biology. Springer (2012) 9. Goswami, M., Chirila, A., Rebreyend, C., de Bruin, B.: EPR spectroscopy as a tool in homogeneous catalysis research. Top. Catal. 58, 719–750 (2015) 10. Hao, Z., Fen, L., Lu, G., Liu, J., An, L., Wang, H.: In situ electron paramagnetic resonance (EPR) study of surface oxygen species on Au/Zno catalyst for low-temperature carbon monoxide oxidation. Appl. Catal. A Gen. 213, 173–177 (2001) 11. Hunger, M., Weitkamp, J.: In situ IR, NMR, EPR, and UV/Vis spectroscopy: tools for new insight into the mechanisms of heterogeneous catalysis. Angew. Chem. Int. Ed. 40, 2954–2971 (2001) 12. Brückner, A.: Monitoring transition metal ions (TMI) in oxide catalysts during (re) action: the power of operando EPR. Phys. Chem. Chem. Phys. 5, 4461–4472 (2003) 13. Zykwinska, A., Domagala, W., Czardybon, A., Pilawa, B., Lapkowski, M.: In situ EPR spectroelectrochemical studies of paramagnetic centres in poly (3, 4-ethylenedioxythiophene)(PEDOT) and poly (3, 4-butylenedioxythiophene)(PBuDOT) films. Chem. Phys. 292, 31–45 (2003) 14. Panchenko, A., Dilger, H., Kerres, J., Hein, M., Ullrich, A., Kaz, T., Roduner, E.: In-situ spin trap electron paramagnetic resonance study of fuel cell processes. Phys. Chem. Chem. Phys. 6, 2891–2894 (2004) 15. Weckhuysen, B. M., In-Situ Spectroscopy of Catalysts. American Scientific Publishers, Valencia, California (2004) 16. Brückner, A.: In situ electron paramagnetic resonance: a unique tool for analyzing structure–reactivity relationships in heterogeneous catalysis. Chem. Soc. Rev. 39, 4673–4684 (2010) 17. Risse, T., Hollmann, D., Brückner, A.: In Situ Electron Paramagnetic Resonance (EPR)–a Unique Tool for Analysing Structure and Reaction Behaviour of Paramagnetic Sites in Model and Real Catalysts, vol. 27. The Royal Society of Chemistry London (2015) 18. Prokopchuk, D.E., Wiedner, E.S., Walter, E.D., Popescu, C.V., Piro, N.A., Kassel, W.S., Bullock, R.M., Mock, M.T.: Catalytic N2 reduction to silylamines and thermodynamics of N2 binding at square planar Fe. J. Am. Chem. Soc. 139, 9291–9301 (2017)

883 19. Stoll, S., Schweiger, A.: Easyspin, a comprehensive software package for spectral simulation and analysis in EPR. J. Magn. Reson. 178, 42–55 (2006) 20. Kim, K.H., Dutta, T., Walter, E.D., Isern, N.G., Cort, J.R., Simmons, B.A., Singh, S.: Chemoselective methylation of phenolic hydroxyl group prevents quinone methide formation and repolymerization during lignin depolymerization. ACS Sustain. Chem. Eng. 5, 3913–3919 (2017) 21. Bromberg, S.E., Chan, I.: Enhanced sensitivity for high-pressure EPR using dielectric resonators. Rev. Sci. Instrum. 63, 3670–3673 (1992) 22. Jaworski, M., Checiński, K., Bujnowski, W., Porowski, S.: Highpressure EPR cavity. Rev. Sci. Instrum. 49, 383–384 (1978) 23. Batchelor, S., Henningsen, B., Fischer, H.: EPR of transient free radicals during photochemical reactions in high temperature and pressure gases. Chem. A Eur. J. 101, 2969–2972 (1997) 24. Mendt, M., Vervoorts, P., Schneemann, A., Fischer, R.A., Pöppl, A.: Probing local structural changes at Cu2+ in a flexible mixed-metal metal-organic framework by in situ electron paramagnetic resonance during Co2 ad-and desorption. J. Phys. Chem. C. 123, 2940–2952 (2019) 25. Hogben, H., Krzystyniak, M., Charnock, G., Hore, P., Kuprov, I.: Spinach–a software library for simulation of spin dynamics in large spin systems. J. Magn. Reson. 208, 179–194 (2011) 26. Hanson, G.R., Gates, K.E., Noble, C.J., Griffin, M., Mitchell, A., Benson, S., Xsophe-Sophe-Xeprview ®: A computer simulation software suite (V. 1.1. 3) for the analysis of continuous wave EPR spectra. J. Inorg. Biochem. 98, 903–916 (2004) 27. Bou-Abdallah, F., Chasteen, N.D.: Spin concentration measurements of high-spin (G′¼ 4.3) rhombic iron (iii) ions in biological samples: theory and application. J. Biol. Inorg. Chem. 13, 15–24 (2008) 28. Dyrek, K., Rokosz, A., Madej, A.: Spin dosimetry in catalysis research. Appl. Magn. Reson. 6, 309–332 (1994) 29. Eaton, G.R., Eaton, S.S., Barr, D.P., Weber, R.T.: Quantitative EPR. Springer Science & Business Media (2010) 30. Eaton, S.S., Eaton, G.R.: Signal area measurements in EPR. Bull. Magn. Reson. 1, 130–138 (1980) 31. Madej, A., Dyrek, K., Mattusch, J.: Preparation and evaluation of the quality of standards for quantitative EPR measurements of spin concentration. Fresenius J. Anal. Chem. 341, 707–708 (1991) 32. Zhang, Y., Peng, Y., Li, J., Groden, K., McEwen, J.-S., Walter, E.D., Chen, Y., Wang, Y., Gao, F.: Probing active-site relocation in Cu/ SSZ-13 SCR catalysts during hydrothermal aging by in situ EPR spectroscopy, kinetics studies, and DFT calculations. ACS Catal. 10, 9410–9419 (2020) 33. Eaton, S.S., More, K.M., Sawant, B.M., Eaton, G.R.: Use of the ESR half-field transition to determine the interspin distance and the orientation of the interspin vector in systems with two unpaired electrons. J. Am. Chem. Soc. 105, 6560–6567 (1983) 34. Kutin, Y., Cox, N., Lubitz, W., Schnegg, A., Rüdiger, O.: In situ EPR characterization of a cobalt oxide water oxidation catalyst at neutral pH. Catalysts. 9, 926 (2019) 35. Hug, G.L., Camaioni, D.M., Carmichael, I.: EPR detection of Hno2•-in the radiolysis of aqueous nitrite and quantum chemical calculation of its stability and hyperfine parameters. Chem. A Eur. J. 108, 6599–6604 (2004) 36. Ohno, K.: Reaction kinetics with rapid mixing. EPR Imag. In Vivo EPR, 175–180 (2018) 37. Augusto, O., Goldstein, S., Hurst, J.K., Lind, J., Lymar, S.V., Merenyi, G., Radi, R.: Carbon dioxide-catalyzed peroxynitrite reactivity–the resilience of the radical mechanism after two decades of research. J Free Radic. Biol. Med. 135, 210 (2019) 38. Huang, X., Chen, Y., Walter, E., Zong, M., Wang, Y., Zhang, X., Qafoku, O., Wang, Z., Rosso, K.M.: Facet-specific photocatalytic degradation of organics by heterogeneous fenton chemistry on

38

884 hematite nanoparticles. Environ. Sci. Technol. 53, 10197–10207 (2019) 39. Frejaville, C., Karoui, H., Tuccio, B., le Moigne, F., Culcasi, M., Pietri, S., Lauricella, R., Tordo, P.: 5-Diethoxyphosphoryl-5-methyl1-pyrroline N-oxide (DEPMPO): a new phosphorylated nitrone for the efficient in vitro and in vivo spin trapping of oxygen-centred radicals. J. Chem. Soc. Chem. Commun., 1793–1794 (1994) 40. Yang, X., Cao, Y., Yu, H., Huang, H., Wang, H., Peng, F.: Unravelling the radical transition during the carbon-catalyzed oxidation of cyclohexane by in situ electron paramagnetic resonance in the liquid phase. Cat. Sci. Technol. 7, 4431–4436 (2017) 41. Chou, P.-W., Song, J.-M., Xie, Z.-Y., Akaike, M., Suga, T., Fujino, M., Lin, J.-Y.: Low temperature de-oxidation for copper surface by catalyzed formic acid vapor. Appl. Surf. Sci. 456, 890–898 (2018) 42. Duan, X., Su, C., Miao, J., Zhong, Y., Shao, Z., Wang, S., Sun, H.: Insights into perovskite-catalyzed peroxymonosulfate activation: maneuverable cobalt sites for promoted evolution of sulfate radicals. Appl. Catal. B Environ. 220, 626–634 (2018) 43. Duan, X., Li, W., Ao, Z., Kang, J., Tian, W., Zhang, H., Ho, S.-H., Sun, H., Wang, S.: Origins of boron catalysis in peroxymonosulfate activation and advanced oxidation. J. Mater. Chem. A. 7, 23904–23913 (2019) 44. Li, D., Duan, X., Sun, H., Kang, J., Zhang, H., Tade, M.O., Wang, S.: Facile synthesis of nitrogen-doped graphene via low-temperature pyrolysis: the effects of precursors and annealing ambience on metal-free catalytic oxidation. Carbon. 115, 649–658 (2017) 45. Wei, Z., Villamena, F.A., Weavers, L.K.: Kinetics and mechanism of ultrasonic activation of persulfate: An in situ EPR spin trapping study. Environ. Sci. Technol. 51, 3410–3417 (2017) 46. Tan, C., Dong, Y., Shi, L., Chen, Q., Yang, S., Liu, X., Ling, J., He, X., Fu, D.: Degradation of Orange II in ferrous activated peroxymonosulfate system: efficiency, situ EPR spin trapping and degradation pathway study. J. Taiwan Inst. Chem. Eng. 83, 74–81 (2018) 47. Mendoza, C., Désert, A., Khrouz, L., Páez, C.A., Parola, S., Heinrichs, B.: Heterogeneous singlet oxygen generation: inoperando visible light EPR spectroscopy. Environmental Science and Pollution Research. 28(20), 25124–25129 (2019) 48. Janzen, E.G., Wang, Y.Y.: Spin trapping with immobilized spin traps. poly (P [. alpha.-(N-tert-butylnitronyl)] styrene). J. Phys. Chem. 83, 894–896 (1979) 49. Kuno, N., Sakakibara, K., Hirota, M., Kogane, T.: A new polymerincorporated spin-trapping reagent aimed at environmental use. Reactions with organic free radicals. React. Funct. Polym. 43, 43–51 (2000) 50. Earla, A., Walter, E.D., Braslau, R.: Synthesis and spin trapping properties of polystyrene supported trifluoromethylated cyclic nitrones. Free Radic. Res. 53, 1084–1100 (2019) 51. Khramtsov, V.V., Reznikov, V.A., Berliner, L.J., Litkin, A.K., Grigor’ev, I.A., Clanton, T.L.: NMR spin trapping: detection of free radical reactions with a new fluorinated DMPO analog. Free Radic. Biol. Med. 30, 1099–1107 (2001) 52. Chattopadhyay, M., Walter, E.D., Newell, D.J., Jackson, P.J., Aronoff-Spencer, E., Peisach, J., Gerfen, G.J., Bennett, B., Antholine, W.E., Millhauser, G.L.: The octarepeat domain of the prion protein binds Cu (II) with three distinct coordination modes at pH 7.4. J. Am. Chem. Soc. 127, 12647–12656 (2005) 53. Wang, F., Büchel, R., Savitsky, A., Zalibera, M., Widmann, D., Pratsinis, S.E., Lubitz, W., Schüth, F.: In situ EPR study of the redox properties of CuO–CeO2 catalysts for preferential CO oxidation (PROX). ACS Catal. 6, 3520–3530 (2016) 54. Prokopchuk, D.E., Chambers, G.M., Walter, E.D., Mock, M.T., Bullock, R.M.: H2 binding, splitting, and net hydrogen atom transfer at a paramagnetic iron complex. J. Am. Chem. Soc. 141, 1871–1876 (2019) 55. Aboukaïs, A., Bennani, A., Aïssi, C.F., Wrobel, G., Guelton, M., Vedrine, J.C.: Highly resolved electron paramagnetic resonance

E. Walter spectrum of copper (II) ion pairs in CuCe oxide. J. Chem. Soc. Faraday Trans. 88, 615–620 (1992) 56. Grauke, R., Schepper, R., Rabeah, J., Schoch, R., Bentrup, U., Bauer, M., Brückner, A.: Impact of Al activators on structure and catalytic performance of Cr catalysts in homogeneous ethylene oligomerization–a multitechnique in situ/operando study. ChemCatChem. 12(4), 1025–1035 (2020) 57. Morra, E., Martino, G.A., Piovano, A., Barzan, C., Groppo, E., Chiesa, M.: In situ X-and Q-band EPR investigation of ethylene polymerization on Cr/SiO2 Phillips catalyst. J. Phys. Chem. C. 122, 21531–21536 (2018) 58. Yin, L., Zhang, J., Yao, J., Li, H.: A designed tempo-derivate catalyst with switchable signals of EPR and photoluminescence: application in the mechanism of alcohol oxidation. ChemCatChem. 10, 3513–3519 (2018) 59. Qi, L., Chamas, A., Jones, Z.R., Walter, E.D., Hoyt, D.W., Washton, N.M., Scott, S.L.: Unraveling the dynamic network in the reactions of an alkyl aryl ether catalyzed by Ni/Γ-Al2O3 in 2-propanol. J. Am. Chem. Soc. 141, 17370–17381 (2019) 60. Walter, E.D., Qi, L., Chamas, A., Mehta, H.S., Sears, J.A., Scott, S.L., Hoyt, D.W.: Operando MAS NMR reaction studies at high temperatures and pressures. J. Phys. Chem. C. 122, 8209–8215 (2018) 61. Lu, Q., Zhang, J., Peng, P., Zhang, G., Huang, Z., Yi, H., Miller, J.T., Lei, A.: Operando X-ray absorption and EPR evidence for a single electron redox process in copper catalysis. Chem. Sci. 6, 4851–4854 (2015) 62. Wang, Q., Zheng, J., Walter, E., Pan, H., Lv, D., Zuo, P., Chen, H., Deng, Z.D., Liaw, B.Y., Yu, X.: Direct observation of sulfur radicals as reaction media in lithium sulfur batteries. J. Electrochem. Soc. 162, A474–A478 (2015) 63. Ali, M.A., Hassan, A., Sedenho, G.C., Gonçalves, R.V., Cardoso, D.R., Crespilho, F.N.: Operando electron paramagnetic resonance for elucidating the electron transfer mechanism of coenzymes. J. Phys. Chem. C. 123, 16058–16064 (2019) 64. Neukermans, S., Hereijgers, J., Ching, H.V., Samanipour, M., Van Doorslaer, S., Hubin, A., Breugelmans, T.: A continuous in-situ EPR electrochemical reactor as a rapid in-depth mechanistic screening tool for electrocatalysis. Electrochem. Commun. 97, 42–45 (2018) 65. Kanan, M.W., Nocera, D.G.: In situ formation of an oxygenevolving catalyst in neutral water containing phosphate and Co2+. Science. 321, 1072–1075 (2008) 66. Nefed’ev, E., Musin, K., Mirakova, T.Y., Kadirov, M., Aminov, K., Salikhov, K., Silaev, V.: EPR imaging study of paramagnetic centre distribution in thiokol-epoxy hermetics. Appl. Magn. Reson. 11, 115–123 (1996) 67. Ulbricht, K., Ewert, U., Herrling, T., Thiessenhusen, K., Aebli, G., Völter, J., Schneider, W.: EPR imaging on zeolites and zeolite catalysts. EPR Imag. In Vivo EPR, 241–250 (2018) 68. Spitzbarth, M., Scherer, A., Schachtschneider, A., Imming, P., Polarz, S., Drescher, M.: Time-, spectral-and spatially resolved EPR spectroscopy enables simultaneous monitoring of diffusion of different guest molecules in nano-pores. J. Magn. Reson. 283, 45–51 (2017) 69. Shevelev, G.Y., Krumkacheva, O.A., Lomzov, A.A., Kuzhelev, A.A., Rogozhnikova, O.Y., Trukhin, D.V., Troitskaya, T.I., Tormyshev, V.M., Fedin, M.V., Pyshnyi, D.V.: Physiologicaltemperature distance measurement in nucleic acid using triarylmethyl-based spin labels and pulsed dipolar EPR spectroscopy. J. Am. Chem. Soc. 136, 9874–9877 (2014) 70. Meyer, V., Swanson, M.A., Clouston, L.J., Boratyński, P.J., Stein, R.A., Mchaourab, H.S., Rajca, A., Eaton, S.S., Eaton, G.R.: Roomtemperature distance measurements of immobilized spin-labeled protein by DEER/PELDOR. Biophys. J. 108, 1213–1219 (2015) 71. Kuzhelev, A.A., Strizhakov, R.K., Krumkacheva, O.A., Polienko, Y.F., Morozov, D.A., Shevelev, G.Y., Pyshnyi, D.V., Kirilyuk, I.A.,

38

Electron Paramagnetic Resonance (EPR)

Fedin, M.V., Bagryanskaya, E.G.: Room-temperature electron spin relaxation of nitroxides immobilized in trehalose: effect of substituents adjacent to no-group. J. Magn. Reson. 266, 1–7 (2016) 72. Cruickshank, P.A., Bolton, D.R., Robertson, D.A., Hunter, R.I., Wylde, R.J., Smith, G.M.: A kilowatt pulsed 94 Ghz electron paramagnetic resonance spectrometer with high concentration sensitivity, high instantaneous bandwidth, and low dead time. Rev. Sci. Instrum. 80, 103102 (2009) 73. Raitsimring, A., Astashkin, A., Enemark, J., Kaminker, I., Goldfarb, D., Walter, E., Song, Y., Meade, T.J.: Optimization of pulsed-DEER measurements for Gd-based labels: choice of operational frequencies, pulse durations and positions, and temperature. Appl. Magn. Reson. 44, 649–670 (2013) 74. Harchol, A., Reuveni, G., Ri, V., Thomas, B., Carmieli, R., Herber, R.H., Kim, C., Leskes, M.: Endogenous dynamic nuclear polarization for sensitivity enhancement in solid-state NMR of electrode materials. J. Phys. Chem. C. 124, 7082–7090 (2020) 75. Wenk, P., Kaushik, M., Richter, D., Vogel, M., Suess, B., Corzilius, B.: Dynamic nuclear polarization of nucleic acid with endogenously bound manganese. J. Biomol. NMR. 63, 97–109 (2015) 76. Wolf, T., Kumar, S., Singh, H., Chakrabarty, T., Aussenac, F., Frenkel, A.I., Major, D.T., Leskes, M.: Endogenous dynamic nuclear polarization for natural abundance 17O and lithium NMR in the bulk of inorganic solids. J. Am. Chem. Soc. 141, 451–462 (2018) 77. Mitchell, N., Kalber, T.L., Cooper, M.S., Sunassee, K., Chalker, S.L., Shaw, K.P., Ordidge, K.L., Badar, A., Janes, S.M., Blower, P.J.: Incorporation of paramagnetic, fluorescent and PET/SPECT contrast agents into liposomes for multimodal imaging. Biomaterials. 34, 1179–1192 (2013)

885 78. Boś-Liedke, A., Walawender, M., Woźniak, A., Flak, D., Gapiński, J., Jurga, S., Kucińska, M., Plewiński, A., Murias, M., Elewa, M.: EPR oximetry sensor—developing a tam derivative for in vivo studies. Cell Biochem. Biophys. 76, 19–28 (2018)

38 Eric Walter attended the University of Delaware and received his Bachelor of Science in chemistry. After a short stint in manufacturing, he became a graduate fellow at Montana State University, receiving his Ph.D. in chemistry for work under mentor David Singel on high-field EPR and many-spin systems. After this he did postdoctoral research in the lab of Glen Millhauser at the University of California Santa Cruz, employing EPR to examine the metal binding sites in the prion protein. He then became a staff scientist at the Environmental Molecular Sciences Laboratory, a national user facility located at the Pacific Northwest National Laboratory. There he specialized in operando techniques for magnetic resonance.

Case Studies: Time-Resolved Electron Paramagnetic Resonance (EPR) Susanne Mossin

39

and David Nielsen

Contents 39.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887

39.2

Opportunities and Challenges in In Situ EPR of Transition Metal Centers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 888

39.3

Information from Analysis of EPR Spectra . . . . . . . . . . . 889

39.4

Information from Time-Resolved EPR Spectral Intensity During In Situ Measurements . . . . . . . . . . . . . . . 891 Cu-Zeolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892 Vanadium in a Keggin-Type Polyoxometalate on Titania . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893

39.4.1 39.4.2 39.5

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 894

39.6

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895

Abstract

Electron paramagnetic resonance (EPR) can be collected both under ex situ, in situ, and operando conditions on solid catalyst materials exposed to relevant gas mixtures. The reactions can be followed with a time resolution of less than a minute and the EPR response can be used to quantify the content of the EPR active component. This is exemplified by the reaction between NO, NH3, and O2 to form N2 and water, the selective catalytic reduction (SCR) reaction. Reactive SCR catalysts containing redox active copper and vanadium find use in industrial applications. Both metals alternate between two oxidation states, one diamagnetic and one paramagnetic having an informative EPR spectrum. The spectral resolution at the relevant temperatures (150–250
C) in in situ or operando experiments is often compromised and the main information obtained is the change in spectrum intensity with time and reaction conditions. Here we S. Mossin (*) · D. Nielsen Centre for Catalysis and Sustainable Chemistry, Department of Chemistry, Technical University of Denmark, Lyngby, Denmark e-mail: [email protected]; [email protected]

present and compare the information obtained on a copper-exchanged zeolite material and on a vanadiumsubstituted polyoxometalate immobilized on titania during in situ cycling between different gas mixtures relevant for the SCR reaction. Keywords

EPR · SCR · SSZ-13 · Keggin polyoxometalates · Catalyst · In situ

39.1

Introduction

In situ EPR is a very diverse field with a large number of paramagnetic centers being amenable to investigation. EPR can be used to follow or identify species both in solution, in the solid state, and even in the gas phase. The paramagnetic centers investigated can be due to one or more unpaired electrons in d-orbitals in transition metals, but also to radicals associated with lighter atoms or point defects in solid materials. EPR is a very sensitive method, and it is often possible to detect paramagnetic species in exceedingly low concentrations or even transient intermediates. In solution, if the concentration is too low for traditional detection, it is still possible to use EPR to detect by using spin-trapping methods. Diamagnetic materials can be investigated by introducing defects by irradiation or electrolysis. The EPR spectra of many types of paramagnetic centers are quite characteristic due to the interaction with the nuclear spin of several elemental isotopes, which gives characteristic splitting patterns and the possibility to assign the spectra. The magnetic field position of the features in the spectra reveals information about the molecular orbital of the unpaired

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_39

887

888

S. Mossin and D. Nielsen

electron(s) and the Zeeman splitting pattern of the electronic states. Here we will focus on a single case study; solid-state catalyst materials consisting of transition metals on oxide supports. The materials are very active catalysts for the selective catalytic reduction (SCR): 4NO þ 4NH3 þ O2 ! 4N2 þ 6H2 O

ð39:1Þ

The SCR reaction is the reaction between a dilute stream of nitrogen oxides, NOx, a stoichiometric amount of ammonia, and oxygen in excess from the air stream. The products of the reaction are dinitrogen and water. The reaction is highly favorable and is catalyzed by several redox active transition metals such as copper, vanadium, iron, and manganese on different supports. Of these, especially vanadium as V2O5 supported on titania, TiO2 (in the anatase morphology) and copper as ion-exchanged Cu2+ in aluminum-containing zeolite materials such as chabazite (CHA) find use in industrial applications. The temperature is typically between 150 and 350
C, with increasing research efforts focused on materials that are active at low temperatures, such as between 150 and 200
C. For both of these types of catalysts, the metals alternate between two oxidation states in the catalytic cycle, one diamagnetic and one with one unpaired electron having an informative EPR spectrum. For copper, the oxidized state, Cu2+, is paramagnetic and can be observed by EPR, whereas the reduced state, Cu+ is diamagnetic and EPR silent. For vanadium, the EPR active state is the reduced state of the metal, V4+, whereas the oxidized state, V5+, is EPR silent. Typically, the spectral resolution is compromised in in situ or operando experiments due to the high temperatures and the high concentration of paramagnetic centers. Another source of information is the change in spectrum intensity with time and the change of reactant gases. The EPR signal intensity (obtained as the double integral of the measured spectra) can be related to the amount of paramagnetic metal, which then can be compared to the total amount of metal present in the sample. The analysis of the spectra themselves and the change in intensity with time reveals how individual EPR active species interconvert during the reaction, and it is possible to extract timeresolved reaction profiles of EPR active species present in the material. The quantification of the number of potentially catalytically active sites in a given material is a piece of very valuable information about the catalyst material. If the response of the material to the change in the gas flow is followed at temperatures where the changes take effect within the timeframe of a few minutes, it is also possible to

obtain reaction profiles and to extract information about the initial reaction rates of EPR active species present.

39.2

Opportunities and Challenges in In Situ EPR of Transition Metal Centers

Electron paramagnetic resonance (EPR) has been used to identify paramagnetic transition metal centers in a range of heterogeneous catalyst materials [1]. It is very attractive due to the low energy of the radiation that does not interfere with the catalytic reaction in any way and the flexibility of the experimental setup. The magnetic field necessary to perform routine X-band EPR is modest and can be reached with a normal electromagnet. The probed volume is convenient for the typical investigation of solid catalyst materials. For 1-electron redox catalysts, one of the relevant oxidation states will be paramagnetic. Most support materials are diamagnetic metal oxides and therefore invisible in an EPR experiment, whereas the redox active catalytic site of interest can be probed. However, there are some challenges in using EPR at room temperature and above. The intensity scales inversely with temperature, so that the signal intensity becomes lower at the high temperatures relevant for heterogeneous catalysis. For systems with integer spin, the allowed transitions in EPR may be outside the accessible frequency or field interval. Of even greater concern is that a short lifetime of the excited state will broaden the signal or even make it disappear completely. If an electronic excited state is close in energy and populated at the relevant temperature and if it is able to communicate with the ground state, then there is usually an efficient and fast pathway to relax the excited spin state, resulting in the excitation being unobservable by EPR. Fortunately, several systems exist where the ground state is well separated in energy from the excited states and informative EPR signals can be observed even at high temperatures. This is true for Cu2+ (d9 electron configuration) in a coordination environment with four equatorial donor atoms (tetragonal coordination) and for vanadium(IV) (d1 electron configuration) in the form of oxidovanadium(IV) species VO2+ (also known as vanadyl). The approximate d orbital energy diagrams of these systems are given in Fig. 39.1. The electronic ground states for square planar Cu2+ and square pyramidal VO2+ have no orbital degeneracy and are far from other electronic states (>10,000 cm1). In addition, the vibrations around the equilibrium bond distances do not influence the orbital energies to any large extent. As a result, observation of the EPR signal of the unpaired electron is possible even at high temperatures. For comparison, the energy level diagram of an approximately tetrahedral V4+ ion is also shown in Fig. 39.1. This system is hypothetical since the system will distort away from the high symmetry due to the Jahn-Teller

39

Case Studies: Time-Resolved Electron Paramagnetic Resonance (EPR)

889

E (cm–1) d(z2)

20000 d(x2–y2)

10000 d(x2–y2) 0

d(x2–y2), d(z2)

d(z2) CuII(OF)4 D4h

VIVO(OF)4 C4v

VIV(OF)4 Td

d(z2) d(x2–y2) VIV(OF)4 JT distorted to D2d

Fig. 39.1 Approximate energy level diagrams of selected metal centers surrounded by framework oxygens, OF. From left to right: Cu2+ in square planar coordination environment (approximate D4h symmetry) [3]. VO2+ with vanadium in a square pyramidal coordination environment (approximate C4v symmetry) [4]. V4+ in a tetrahedral environment

(approximate Td symmetry) and an indication of how the energy levels would split after a decrease in symmetry from Td to D2d. Breaking of symmetry is expected since V4+ in perfect tetrahedral symmetry would be unstable due to the Jahn-Teller effect

effect. A distortion along the C2 axis (chosen as an elongation) will result in D2d symmetry and the new ground state will be nondegenerate. The result is shown in Fig. 39.1, right. A distortion along the C3 axis, on the other hand, will not split the ground state and is therefore not considered relevant here. A small distortion towards D2d will result in only a small energy gap between the ground state and the first excited state. The close-lying excited state interacts with the ground state, and consequently, there is an efficient pathway for fast relaxation of the excited spin state. In this case, the EPR spectrum cannot be observed at non-cryogenic conditions. A similar argument were presented to explain why EPR could not detect 3-coordinate Cu2+ in [Cu(OF)2OH]+, which was proposed to be present in not too diluted Cu-zeolites after dehydration [2]. As a consequence of the many criteria that have to be fulfilled in order to observe an EPR spectrum at a high temperature, the data from EPR have to be interpreted carefully. The absence of an EPR spectrum does not constitute proof that no paramagnetic species are present, but only that none of the types of EPR active centers are present. On the other hand, the observation of an EPR signal is absolute proof of the presence of paramagnetic species. Cu2+ and V4+ are both relevant for the industrially important SCR reaction since the redox couples Cu2+/Cu+ and V5+/ V4+ are part of the reaction in copper-exchanged zeolites and in vanadia on titania, respectively [5, 6]. EPR can probe the oxidized state of copper, whereas the reduced state is EPR silent. For vanadium, it is the reduced oxidation state which is observable, whereas the oxidized state is EPR silent. Copperexchanged zeolite materials are used in automotive deNOx in diesel engine exhaust gas treatment systems [7], whereas

vanadia on titania is used in stationary deNOx in flue gases, e.g., power plants [8], and for use in heavy-duty trucks. In the following, representatives of these two types of materials have been chosen as examples of redox active catalyst materials. EPR gives two separate types of information for both samples: (1) the appearance of the spectrum itself, giving information about the coordination environment of the paramagnetic metal ion and the distribution between different sites; and (2) the EPR signal intensity.

39.3

Information from Analysis of EPR Spectra

The information obtainable from well-resolved EPR spectra of Cu2+ and V4+ (as oxidovanadium(IV), VO2+) is well described in the literature. For this study, a chabazite zeolite, SSZ-13, was synthesized according to a literature procedure [9]. It contained Si/Al ¼ 11.3 according to analysis by X-ray fluorescence (XRF). Sodium was not present in any of the precursors [10, 11]. The zeolite was ion exchanged with aqueous Cu(NO3)2 (5 mM) for 16 h; filtered, dried at 90
C; and calcined at 580
C for 3 h. The material was fractioned into a particle size distribution of 150–300 μm. According to inductively coupled plasma elemental analysis (ICP analysis), the sample contained 1.14 wt. % Cu. A Keggin-type heteropolyacid, H3PMo12O40, also commonly referred to as a polyoxometalate (POM), was modified by exchanging 3 of the Mo present for V. The resulting POM was supported on TiO2 (anatase). The synthesis of the material is described in Ref. [12]. The material contains 1.6 wt. %

39

890

S. Mossin and D. Nielsen

of V by ICP. Both the Cu-zeolite and the supported V-containing POM were checked and were found to be active catalysts for the SCR reaction [9, 12]. Examples of spectra obtained for the copper-exchanged zeolite and the vanadia on titania samples are shown in Fig. 39.2. The spectra were measured with a continuous wave Bruker X-band EPR spectrometer (model EMX) with a ST4102 cavity, a microwave frequency of 9.5 GHz, a power of 6.3 mW, a modulation frequency of 100 kHz, and a modulation amplitude of 0.52 mT using a Bruker variable temperature insert controlled by a separate thermocouple. The literature is rich in EPR spectra of Cu2+ and V4+ in different coordination environments. For copper, the immediate coordination environment induced by four oxygen ligands gives an EPR spectrum, which is significantly different from the spectrum of copper surrounded by four other atoms, such as nitrogen [13]. The particular shape of a typical EPR spectrum can be reproduced by fitting the spectrum calculated using a spin Hamiltonian model of the energy levels to the experimental spectrum. The simplest form only includes the interaction between the unpaired electron and the magnetic field, and treats interactions between the electron and any nuclear spins as a fieldindependent perturbation with coupling parameters for each nucleus.

a) 2.8

2.4

2.2

2.0

1.8

g||(1)

Cu-Zeolite

2.6

g-value 2.3

V-POM on titania

A||(1)

ð39:2Þ

Here μB is the Bohr magneton, g and A are the diagonal g and A tensors (gx, gy, gz) and (Ax, Ay, Az), B is the magnetic field, I is the nuclear spin, and S is the electron spin. When applied to explain the EPR transitions, this model works excellently for systems where the magnetic moment is dominated by the electron spin and other contributions, such as spin-orbit coupling, can be treated as perturbations. The spin Hamiltonian is routinely applied to fit EPR spectra by deriving the spin Hamiltonian tensors, g and A. It is usually assumed that the axes that provide diagonal g and A tensors coincides. An EPR spectrum simulation program such as EasySpin [14] uses this spin Hamiltonian model to predict the field dependence of the energy levels of the different electronic and nuclear spin states. The program searches for all transitions between energy levels for a given experimental microwave frequency. Allowed transitions are weighted by their transition probability, and spectra are evaluated over all orientations of the molecular coordination system with respect to the laboratory frame since we are investigating randomly oriented microcrystalline powder materials. Finally, a line shape function, Lorentzian or combined Lorentzian and Gaussian, is applied to obtain a final simulated spectrum. For both of the investigated paramagnetic metal centers, the approximate local symmetry of the metal centers is

b)

g-value 2.6

H ¼ gμB SB þ ASI

2.0

1.8

1.6

gA(1) g||(1)

gA(1,2) A||(1)

A||(2) g||(2) A||(2) g||(2)

gA(2) 240

260

280

300 320 Field (mT)

340

360

380

Fig. 39.2 (a) EPR spectrum of Cu-zeolite recorded at 200
C in a flow of 10% O2 þ 1000 ppm NO and balance He. For each Cu2+ species, a quartet signal arise due to the hyperfine splitting as illustrated by the black and blue lines. The axial part of the hyperfine interaction is not

260

300

340 Field (mT)

380

420

resolved. (b) EPR spectrum of a vanadium-substituted Mo polyoxometalate deposited on TiO2 recorded at 250
C in a flow of 550 ppm NH3 þ 500 ppm NO in He

39

Case Studies: Time-Resolved Electron Paramagnetic Resonance (EPR)

axial. This implies that we should use an axial spin Hamiltonian. Taking into account that both nuclei have nonzero nuclear spin, the spin Hamiltonian thus has four parameters. Two g-values, gz ¼ g|| and gx ¼ gy ¼ g⊥, are necessary to describe the field position of the features of the spectra, and two coupling parameters, A-values, Az ¼ A||, Ax ¼ Ay,¼ A⊥, are necessary to describe the fine structure of the spectra due to the interaction between the electron and the nearby nuclei. The nuclear spin quantum number is I ¼ 3/2 for both naturally abundant isotopes of copper, giving rise to a hyperfine quartet, and the difference between the isotopes only gives rise to a modest 7% deviation in coupling parameters, which is typically not resolved in the experimental spectra of Cu-zeolite materials. For X-band EPR of typical tetragonal copper systems, the splitting of spectral features due to the g anisotropy is significant and larger than the splitting due to the interaction with the copper nuclei, which is typically only significant in the parallel direction. For vanadium, the 51V isotope (100% abundance) has I ¼ 7/2, giving rise to a hyperfine octet. For X-band EPR of typical tetragonal oxidovanadium (IV) species, the splitting pattern due to the hyperfine octet dominates over the small splitting due to the g anisotropy, see Fig. 39.2a, b. Peisach and Blumberg collected EPR spectral information of Cu2+ from a literature study of a large number of biomolecules and conceived a plot of A|| versus g|| [13]. These parameters are easily discerned in a typical EPR spectrum of a Cu2+ compound. It is found empirically to describe the coordination environment of copper and to reveal well which type of donor atoms coordinate directly with Cu in the tetragonal plane. Carl and Larson applied this plot to data obtained on a range of copper-zeolite materials [15]. Davidson and Che devised a similar plot for V4+, using the average g- and A-values, which were found empirically to be more descriptive of the coordination environment in this case [16]. The spectrum of the Cu-zeolite in Fig. 39.1 shows spin Hamiltonian parameters that correspond to two similar, yet distinguishable, Cu-species coordinated to four oxygen atoms. The two separate sites were assigned to Cu coordination to the zeolite framework in two different sites that are subtly different due to a different Al distribution in a double 6 ring in the framework [2, 17]. The spectrum for the vanadium-based catalyst shows at least two different species as well as a broad unresolved peak. With two sets of overlapping octets with pronounced anisotropy in the hyperfine coupling tensor for each species, the spectrum becomes more difficult to interpret. An analysis of the two resolved species reveals that both correspond to square pyramidal 5-coordinate or 6-coordinate oxidovanadium (IV) species. The isotropic g-value of the underlying broad

891

species is also compatible with this type of coordination [18]. The assignment of each of these species to particular oxidovanadium(IV) sites on anatase TiO2 has been discussed recently [19, 20].

39.4

Information from Time-Resolved EPR Spectral Intensity During In Situ Measurements

The intensity of the EPR spectra is another source of valuable information. It can be obtained from the background-corrected first-derivative spectra by simple double integration. For concentrated samples, the individual features of the EPR spectrum are often not discernible, and the overall line position and the total intensity are the only information that can be extracted. The magnitude of the magnetic moment of an unpaired electron is always the same and, therefore, in the absence of strong exchange coupling between paramagnetic metal centers with single unpaired electrons, the intensity of the EPR spectrum at a given temperature is directly proportional to the number of paramagnetic centers. This is one of the most powerful aspects of the method. For heterogeneous catalyst materials at high metal concentration, the spectra are less informative due to the high temperatures and relatively high concentration of paramagnetic centers, but the relative intensity during an experiment can be obtained even during changes in reactant gases and temperature as long as the resonant cavity of the instrument can be maintained at a constant ambient temperature. Using a commercial double-walled quartz insert with a vacuum between the glass walls in a ST4102 standard cavity (Bruker) and a variable temperature unit using a flow of air or nitrogen gas, the sample can be maintained at another temperature than the cavity and reliable and quantifiable results can be obtained in a broad temperature interval of 100–573 K. EPR spectra are almost exclusively measured as firstderivative spectra, and therefore, the spectral intensities are obtained by double integration over field. If care is taken to compare the measured intensity to reference samples with well-known spin concentration and to perform careful background corrections as well as corrections due to changes in temperature, it is possible to find the percentage of EPR active centers and EPR inactive centers in the sample. A little care must be taken before translating the number into a percentage of oxidized and reduced metal ions since EPR signal intensity, in some cases, can be lost due to subtle changes in the coordination environment and not only due to a change of oxidation state. As an example of this, even the loss of one ligand in a Cu2+ ligand sphere (going from square planar Cu2+ to trigonal Cu2+) may result in the loss of the observable EPR spectrum [2]. Another example is the tetrahedral V4+ system illustrated in Fig. 39.1.

39

892

S. Mossin and D. Nielsen

Here we will show how much information can be extracted from a single descriptor, the total intensity of the spectra, as a function of time and reaction conditions. The method will be exemplified by measurements upon a copperexchanged zeolite sample and vanadium in polyoxometalate supported upon titania. We investigate the active SCR catalyst materials in situ during exposure to a gas stream of NO + NH3, which will effectively reduce the metal species according to the reaction NO þ NH3 þ Cu

2þ

þ

! N2 þ H2 O þ Cu þ H

þ

ð39:3Þ

for the Cu-based material and NO þ NH3 þ VO

3þ

! N2 þ H2 O þ VO

2þ

þ

þH

ð39:4Þ

for the vanadium-based material. The paramagnetic EPR active species are underlined. We also expose the catalysts to a gas stream comprised of NO + excess O2, which will effectively oxidize the metal centers back to Cu2+ and VO3+, but the precise reaction scheme is in both cases less well understood. We also investigate the response of the Cu-zeolite material to a gas mixture of NO þ NH3 þ O2 in the same ratio as in a realistic SCR test (1000 ppm for NO and NH3, 10% for O2). Careful initial ex situ experiments are performed at room temperature in both cases to quantify the EPR signal intensity of the fresh materials relative to reference samples. An explanation of how this quantification is performed can be found in Ref. [2] for Cu and in Refs. [20, 21] for V. In all cases, the response of the catalyst material is followed just after changing the content of the gas stream until steady-state conditions are reached. The amount of active metal that is able to change its oxidation state when the content of the gas phase is adjusted can be quantified by EPR and compared to the total amount of metal or the total mass of catalyst material in the sample. This number is a descriptor of how much of the active metal that is actually able to enter into the catalytic reaction. The amount of metal that does not alternate between the two oxidation states but stays in one oxidation state despite the gas mixture changing can be assumed to be irrelevant or inaccessible to the catalytic reaction. There can be several reasons for the redox active metal to be caught in one oxidation state; e.g., inaccessibility (blocking of micropores or slow diffusion in the surface of the material), strong coordination of the active metal to catalyst poisons (e.g., sulfates for Cu-zeolites), or a strong interaction with the support material.

39.4.1 Cu-Zeolite 20 mg of Cu-zeolite was immobilized in a 4 mm inner diameter quartz tube as described in Ref. [17] and a time-resolved experiment was started recording an EPR spectrum every 90 s. After acquiring a few spectra, a flow of 200 mL/min of 10% O2 in He was applied. The temperature was ramped to 250
C at 10 degrees/min and left at 250
C for 30 min. Subsequently, the sample was exposed to reaction gases at 200
C. The results are shown in Fig. 39.3 as the double integral of the EPR spectra. Experimental spectra are background corrected by subtracting the spectra of an empty tube and by subtracting a linear slope generated from the start and end level (averaged over 100 measurement points) of each individual spectrum. For comparison with the ex situ quantification performed at room temperature, the intensity of each spectrum is multiplied with 295 K/ T, where T is the temperature of the particular spectrum. This is done in order to correct for the decrease in intensity due to the different Boltzmann distribution of the spin states at the temperature T of the measured spectrum and that of the reference spectrum (295 K). For this Cu-zeolite material, >90% of the Cu is EPR active after heating at 250
C in air and ~ 100% is EPR active after oxidation with NO + O2. Samples with higher Cu-loading have previously been shown to have a decreased EPR signal of the activated catalyst compared to the fresh catalyst despite all copper being Cu2+. This is due to the change in copper coordination during the dehydration procedure and the inability of EPR to see some types of Cu2+ sites. For intermediate Cu loading samples, a considerable part of the invisible sites were assigned to be CuII(OF)2OH close to one Al in the zeolite framework [2]. For higher loading of Cu, dimerization of the Cu species is also expected, and since the dominant coupling between Cu2+ bridged by oxygen is antiferromagnetic, the total paramagnetic response decreases when Cu starts to cluster. Since we observe here that all Cu is still observable after the dehydration treatment, we conclude that this particular sample has all Cu as CuII(OF)4 close to two Al in the framework [22]. Figure 39.3a shows that the EPR signal decreases very fast when Cu2+ is reduced to Cu+ by exposure to NO + NH3. The oxidation of Cu+ to Cu2+ by NO + O2 occurs at a slower rate than the reduction and appears to occur in two steps with different rates. The Cu-zeolite is microcrystalline and has a completely uniform porosity, and we observe changes in the gas flow very fast in the EPR response during reduction. This sets the limit for how fast we can possibly observe changes during oxidation. The slower rate after a few minutes during the oxidation step must be due to a slower reaction and not due to diffusion limitation. Investigation of the intermediate spectrum during the first sharp increase in intensity shows that only one Cu2+ species giving sharp features is present, namely the species (2) marked in blue in Fig. 39.2a [23]. After a longer exposure to NO + O2, species (1) also appear.

39

Case Studies: Time-Resolved Electron Paramagnetic Resonance (EPR)

a)

NO+NH3

NO+O2

Cu-Zeolite

b)

NO+NH3

0.8

1.0

NO+O2

V-POM on titania

0.7 EPR active V4+ (wt. %)

EPR active Cu2+ (wt. %)

893

0.8

0.6

0.4

0.6 0.5 0.4 0.3 0.2

0.2 SCR 0.0

0.1 0.0

0

60

120

180

240

300

Time (min)

0

60

120 Time (min)

Fig. 39.3 Quantification of EPR active metal centers as a function of time during exposure to different reaction gasses: 1000 ppm NO, 1000 ppm NH3, 10% O2 in He for Cu-zeolite at 200
C and 500 ppm NO, 550 ppm NH3, and 10% O2 for vanadia on titania at 250
C. The white areas indicate a reaction mixture of NO + NH3, and the green areas indicate NO + O2. The region marked SCR represents a flow of 1000 ppm NO +1000 ppm NH3 þ 10% O2 in He. In the left graph,

1.14 wt. % Cu corresponds to 100% of the copper content observed by EPR in the fresh sample (verified by ex situ comparison to reference samples). In the right graph, 0.8 wt. % V corresponds to 50% of the total vanadium content in the sample by comparison between the EPR intensity of a fresh sample and a calibration curve based upon EPR signal intensity of five reference samples

Approximately 9% of the total amount of Cu2+ is still present in the EPR spectrum even after prolonged reduction with NO + NH3. This indicates that 0.1 wt. % Cu in this sample is not involved in the catalytic reaction at 200
C. Apparently, the gas phases NO and NH3 can reach and reduce 91% of all Cu and only 9% is not accessible for reduction. Comparing different samples in this way can give insight into the fraction of metal that is available for catalysis and the fraction of transition metal that is wasted in inactive or inaccessible sites. After around 220 minutes and three completely reproducible cycles, suddenly the relative Cu signal only reaches 95% after oxidation instead of 100% as before. There was no change in the procedure or in the features of the spectra, and we also did not note a change in the sample such as shifting of the fractioned powder in the reactor due to a shift in the catalyst bed in the gas flow, although this explanation cannot be excluded completely. This illustrates the inherent uncertainty in performing quantification in an in situ setup. We recommend that all reliable quantification results should be based upon several replica of experiments. Indeed, this unexpected development was not observed the next time the sample was investigated. Furthermore, it is observed that the intensity of the Cu-zeolite sample when exposed to SCR gases (marked in Fig. 39.3a) is close to the reduced state of the catalyst. This is in accordance with mechanistic studies that suggest that the ratelimiting step of the catalytic cycle should be found in the part of the catalytic cycle that involves oxidation of Cu+ to Cu2+ [5].

39.4.2 Vanadium in a Keggin-Type Polyoxometalate on Titania 20 mg of the sample was mounted in the same way as the Cu-zeolite and the experiment proceeded in almost the same way, including the initial activation in air at 250
C. The differences were that he concentration of NO and NH3 was lower (500 and 550 ppm, respectively), and the temperature during reaction was higher (250
C). In this case, the EPR signal intensity increases when the sample is exposed to NO + NH3 and decreases when it is exposed to NO + O2. This is the opposite of what was observed for Cu since EPR is sensitive to the reduced state of vanadium. For this sample, the EPR signal calibration to the total metal content is more challenging since the analysis relies on comparison between the ex situ quantification experiment and the beginning of the in situ experiment, before the temperature is increased. The procedure for how this comparison can be made is described in detail in [20]. The EPR signal from fresh vanadium samples has low intensity since most of the vanadium in the fresh catalyst is in the oxidized state. Due to the low intensity of the EPR spectrum of the reference state of the material, the result of the quantification has a larger uncertainty. We alleviate this uncertainty by performing several replicates of the experiments. For this sample, we use the quantification procedure to determine that the signal after exposure to NO + O2 corresponds to 3% of the V present. In other words, at least 97% of

39

894

S. Mossin and D. Nielsen

the vanadium in the sample is in the oxidized state after this treatment. During exposure to NO + NH3, 19% of the total amount of vanadium is EPR active. Information from EPR alone cannot prove that the other 81% of the vanadium is all V5+ as there might be another reason for the lack of signal. This could be fast relaxation of the exited state for some types of V4+ species (such as the approximately tetrahedral V4+ site shown in Fig. 39.1 [16]). On the other hand, 19% of vanadium is identified as oxidovanadium(IV) (VO2+) by the EPR analysis. Interestingly, the magnitude of the change in EPR signal intensity between the fully oxidized catalyst (in NO + O2 flow) and the fully reduced catalyst (in NO + NH3 flow) was observed to be a relevant descriptor for the SCR activity on a range of the investigated materials [12]. This measure is observable in Fig. 39.3 as the difference in the upper and lower plateaus and corresponds to 19 – 3 ¼ 16% of the total vanadium content or 0.26 wt. % V in the sample. For the vanadium sample, the reduction and oxidation rates are comparable. The intensity profile suggests that both reduction and oxidation of vanadium takes place in two different regimes. For the oxidation, there is first a fast rate of response and then it stabilizes at a slower rate. For the reduction, the profile reveals first a slow rate of response and then a faster rate. In other words, there is a kink in the slope of both profiles, but they have opposite signs. We inspected the individual EPR spectra of the vanadium sample to search for a clue to these observations, but we did not find any additional information here to help us understand the difference between the two regimes. We hypothesize that the availability of ammonia on the surfaces is the reason for the changes in the rates of response in the oxidation and reduction of vanadium species. The reason we cannot detect this in the spectra themselves is that the EPR spectrum observed is not sensitive to the coverage of NH3. Ammonia is not coordinated directly with the paramagnetic vanadium center but further away from the vanadium. Fig. 39.4 Zoom in on sections of the graphs in Fig. 39.3 showing the EPR signal intensity of the catalyst materials throughout the in situ experiment at 200
C (Cu-zeolite) and 250
C (V-POM on titania). The gas flow alternated between two different gas mixtures: NO + NH3 (white sections) and NO + O2 (green sections). Lines are drawn as an aid in assessing the relative slope of different parts of the graph. Numbers 1–7 have been added to aid the discussion below

Another observation is that the first reduction (at 42–58 minutes in the experiment shown in Fig. 39.3b) has a different EPR intensity profile than the following reductions. This could indicate a reordering of surface species during the first cycle. In the following cycles, the evolution during the shift in gas composition is reproducible between cycles.

39.5

Discussion

The information obtainable by quantification of the EPR signal shows that both materials respond fast and efficiently to the change in gas composition of the flow. In general, the well-defined and crystalline Cu-zeolite materials give correspondingly detailed and sharp EPR spectra and signal intensity profiles, and almost all the copper in the sample (91% or 1.0 wt. % Cu) changes oxidation state during the procedure. The supported vanadium samples have a higher degree of disorder on the molecular level and the EPR spectra themselves are complicated by a higher degree of magnetic interactions between paramagnetic species, giving a large contribution from species with a signal broadened by exchange interactions of a range of directions and magnitudes. About 16% of the total amount of vanadium was found to change the EPR response after the changes in gas composition. For 84% of the vanadium, we did not detect any EPR response after exposing it to a gas mixture that should result in a reduction of vanadium(V) to vanadium(IV). Indeed, the vanadium-based catalyst was also found to be less active for the SCR reaction than the Cu sample on both a weight basis and a molar basis [9, 12]. The quantification of the EPR signal proceeded equally well on both samples, and the EPR signal intensity was reproduced within 1% for the vanadium sample and within

Cu-Zeolite

V-POM on titania NO+O2

NO+NH3

NO+NH3 6

5 1

NO+O2

3

7 4

2

150

180 Time (min)

150 Time (min)

180

39

Case Studies: Time-Resolved Electron Paramagnetic Resonance (EPR)

1–5% for the Cu sample in subsequent oxidation-reduction cycles despite the complexity of the in situ setup and quantification procedure. The reduction of copper occurs fast with a constant slope all the way down to complete reduction; see the red line marked 1 in Fig. 39.4. The reduction is the fastest we can observe, and the rate is calculated to be ~1/3 of the rate of how fast one molecule of NH3 comes into contact with one Cu atom in the sample at the used space velocity. The oxidation of copper is significantly slower and can be followed on the time scale of the experiment. It has first a fast rate of response (2), then almost stabilizes for a short period, and finally it follows a slower rate (3). The reduction of vanadium follows a different type of profile: It has first a slow (4) and then a significantly faster rate of response (5). The oxidation of vanadium is initially fast (6), and then, after a small transition period, the rate is slightly slower (7). The overall features of both reduction and oxidation of both samples are reproduced several times in subsequent cycles, see Fig. 39.3. This includes reproduction of the short nonlinear periods in the oxidation profile of both samples: the interval between the slopes belonging to (2) and (3) in the Cu-zeolite and the interval between the slopes belonging to (6) and (7) in the V-POM. A similar Cu-zeolite sample was previously followed by EPR during reduction with NO + NH3 at a lower reaction temperature (100
C) in order to slow the reaction down [23]. At this temperature, first a fast and then a slower rate were observed, and the change in reaction rate was assigned to a change in the ammonia coverage on Cu; excess ammonia has been suggested to inhibit Cu slightly towards reduction due to the stability of the coordination complex formed: [Cu (NH3)4]2+. For the V-POM material during the reduction, we observe the opposite: first the rate is slow (4) and then fast (5). For this material, in contrast to the Cu-zeolite, it is reasonable to suggest that the first ammonia entering the sample is used to saturate the very acidic sites on the V-POM materials. Since most of the ammonia is making strong bonds to the surface, it is not available for reaction with NO and the reduction progresses only slowly (4). When the acidic sites are saturated, NH3 is available for reaction and the rate increases (5). Ammonia is intimately involved in the redox reactions of the metals and we hypothesize that the ammonia coverage is the underlying reason for the changes in reaction rates we observe, both during reduction of V-POM and during reduction of Cu-zeolite (which is so fast that the rate has to be investigated at lower temperatures), but also during oxidation of both materials. During oxidation, the argument is as follows: As long as ammonia is available on the surface, one rate of reaction is observed. This corresponds to the catalyst running the SCR reaction since NO, NH3, and O2 are all present. When running the SCR reaction, the distribution

895

between reduced and oxidized metal is moving fast towards the steady state value for the catalyst at these conditions. Once the steady-state value is reached, the rate of change in the EPR response slows down or stops since, as the surface is depleted of ammonia, the SCR reaction cannot run anymore. After that, new slower reaction rates are relevant for both materials, corresponding to oxidation by NO + O2 in the absence of ammonia: This is seen in the linear rate region (3) for the Cu-zeolite and (7) for the V-POM on titania.

39.6

Conclusion

It has been shown through two examples how the observation of EPR signal intensity during redox cycling can provide valuable information. It is therefore possible to quantify the amount of the redox active metals taking part in the reaction and ensure that they are in good contact with the reactants in the gas phase. This is useful information to guide the design of new and more efficient catalysts for the SCR reaction and to characterize the response of the material to gas mixtures relevant to the SCR reaction. The method is equally valuable for other catalytic reactions with 1-electron redox cycling if at least one of the oxidation states gives a robust quantifiable EPR signal. In practice, this implies that the paramagnetic species should possess an uneven number of unpaired electrons and an electronic ground state that is well separated from excited electronic states. By taking a closer look at the profiles for the EPR signal intensity during the cycling between a NO + NH3 gas mixture and a NO + O2 gas mixture, the influence of ammonia coverage on the rate of change of metal oxidation states was revealed for two different types of NH3-SCR catalyst materials. Acknowledgments Anna Bukowski and Jakob Albert, Friedrich-Alexander-Universität Erlangen-Nuremberg, are thanked for supplying the vanadium polyoxometalate material. Søren B. Rasmussen and Anita L. Godiksen, Haldor Topsøe A/S, are thanked for their contributions to the operando and in situ EPR setup at DTU Chemistry. The Carlsberg Foundation and The Independent Research Fund Denmark (DFF 1335-00175) are thanked for their support of the EPR instrument at the Department of Chemistry.

References 1. Brückner, A.: In situ electron paramagnetic resonance: a unique tool for analyzing structure-reactivity relationships in heterogeneous catalysis. Chem. Soc. Rev. 39(12), 4673–4684 (2010). https://doi. org/10.1039/b919541f 2. Godiksen, A., Stappen, F.N., Vennestrøm, P.N.R., Giordanino, F., Rasmussen, S.B., Lundegaard, L.F., Mossin, S.: Coordination environment of copper sites in Cu-CHA zeolite investigated by electron paramagnetic resonance. J. Phys. Chem. C. 118, 23126–23138 (2014). https://doi.org/10.1021/jp5065616

39

896 3. Vanelderen, P., Vancauwenbergh, J., Tsai, M.-L., Hadt, R.G., Solomon, E.I., Schoonheydt, R.A., Sels, B.F.: Spectroscopy and redox chemistry of copper in mordenite. ChemPhysChem. 15(1), 91–99 (2014). https://doi.org/10.1002/cphc.201300730 4. Ballhausen, C.J., Gray, H.B.: The electronic structure of the vanadyl ion. Inorg. Chem. 1(1), 111–122 (1961). https://doi.org/10.1021/ ic50001a022 5. Janssens, T.V.W., Falsig, H., Lundegaard, L.F., Vennestrøm, P.N.R., Rasmussen, S.B., Moses, P.G., Giordanino, F., Borfecchia, E., Lomachenko, K.A., Lamberti, C., et al.: A consistent reaction scheme for the selective catalytic reduction of nitrogen oxides with ammonia. ACS Catal. 5(5), 2832–2845 (2015). https://doi.org/10. 1021/cs501673g 6. Arnarson, L., Falsig, H., Rasmussen, S.B., Lauritsen, J.V., Moses, P.G.: A complete reaction mechanism for standard and fast selective catalytic reduction of nitrogen oxides on low coverage VOx / TiO2 (001) catalysts. J. Catal. 346, 188–197 (2017). https://doi. org/10.1016/j.jcat.2016.12.017 7. Deka, U., Lezcano-Gonzalez, I., Weckhuysen, B.M., Beale, A.M.: Local environment and nature of Cu active sites in zeolite-based catalysts for the selective catalytic reduction of NOx. ACS Catal. 3, 413–427 (2013) 8. Forzatti, P.: Present status and perspectives in De-NOx SCR catalysis. Appl. Catal. A Gen. 222, 221–236 (2001) 9. Nielsen, D., Gao, Q., Mossin, S.: Manuscript in Preparation 10. Di Iorio, J.R., Gounder, R.: Controlling the isolation and pairing of aluminum in chabazite zeolites using mixtures of organic and inorganic structure-directing agents. Chem. Mater. 28(7), 2236–2247 (2016). https://doi.org/10.1021/acs.chemmater.6b00181 11. Zones, S.I.: Zeolite SSZ-13 and its method of preparation. US4544538 A, 1985 12. Bukowski, A., Schill, L., Nielsen, D., Mossin, S., Riisager, A., Albert, J.: NH3-SCR of NO with novel active, supported vanadium-containing Keggin-type heteropolyacid catalysts. React. Chem. Eng. 5, 935–948 (2020). https://doi.org/10.1039/ d0re00033g 13. Peisach, J., Blumberg, W.E.: Structural implications derived from the analysis of electron paramagnetic resonance spectra of natural and artificial copper proteins. Arch. Biochem. Biophys. 165, 691–708 (1974) 14. Stoll, S., Schweiger, A.: EasySpin, a comprehensive software package for spectral simulation and analysis in EPR. J. Magn. Reson. 178, 42–55 (2006). https://doi.org/10.1016/j.jmr.2005.08.013 15. Carl, P.J., Larsen, S.C.: EPR study of copper-exchanged zeolites: effects of correlated g - and A -Strain , Si/Al ratio , and parent zeolite. J. Phys. Chem. B. 104, 6568–6575 (2000) 16. Davidson, A., Che, M.: Temperature-induced diffusion of probe vanadium(IV) ions into the matrix of titanium dioxide as investigated by ESR techniques. J. Phys. Chem. 96, 9909–9915 (1992). https://doi.org/10.1021/j100203a061 17. Godiksen, A., Vennestrøm, P.N.R., Rasmussen, S.B., Mossin, S.: Identification and quantification of copper sites in zeolites by electron paramagnetic resonance spectroscopy. Top. Catal. 60, 13–29 (2017). https://doi.org/10.1007/s11244-016-0731-7 18. Kompio, P.G.W.A., Brückner, A., Hipler, F., Auer, G., Löffler, E., Grünert, W.: A new view on the relations between tungsten and vanadium in V2O5-WO3/TiO2 catalysts for the selective reduction of NO with NH3. J. Catal. 286, 237–247 (2012). https://doi.org/10. 1016/j.jcat.2011.11.008 19. Vuong, T.H., Radnik, J., Rabeah, J., Bentrup, U., Schneider, M., Atia, H., Armbruster, U., Grünert, W., Brückner, A.: Efficient VOx/Ce1xTixO2 catalysts for low-temperature NH3-SCR: reaction mechanism and active sites assessed by in situ/operando

S. Mossin and D. Nielsen spectroscopy. ACS Catal. 7, 1693–1705 (2017). https://doi.org/10. 1021/acscatal.6b03223 20. Godiksen, A.L., Funk, M.H., Rasmussen, S.B., Mossin, S.: Assessing the importance of V(IV) during NH3-SCR using operando EPR spectroscopy. ChemCatChem. 12, 4893–4903 (2020). https://doi.org/10.1002/cctc.202000802 21. Putluru, S.S.R., Schill, L., Godiksen, A., Poreddy, R., Mossin, S., Jensen, A.D., Fehrmann, R.: Promoted V2O5/TiO2 catalysts for selective catalytic reduction of NO with NH3 at low temperatures. Appl. Catal. B Environ. 183, 282–290 (2016). https://doi.org/10. 1016/j.apcatb.2015.10.044 22. Paolucci, C., Parekh, A.A., Khurana, I., Di Iorio, J.R., Li, H., Albarracin Caballero, J.D., Shih, A.J., Anggara, T., Delgass, W.N., Miller, J.T., et al.: Catalysis in a cage: condition-dependent speciation and dynamics of exchanged Cu cations in SSZ-13 zeolites. J. Am. Chem. Soc. 138(18), 6028–6048 (2016). https://doi.org/10. 1021/jacs.6b02651 23. Godiksen, A., Isaksen, O.L., Rasmussen, S.B., Vennestrøm, P.N.R., Mossin, S.: Site-specific reactivity of copper chabazite zeolites with nitric oxide, ammonia, and oxygen. ChemCatChem. 10(2), 366–370 (2018). https://doi.org/10.1002/cctc.201701357

Susanne Mossin received her PhD from the University of Copenhagen in 2006. She worked at the Friedrich-Alexander University of ErlangenNuremberg and joined the Technical University of Denmark in 2009. She is working on in situ and operando electron paramagnetic resonance spectroscopy of heterogeneous catalysts and spectroscopic, electronic, and magnetic properties of inorganic materials.

David Nielsen graduated as an MSc in chemical engineering in 2017 at the Technical University of Denmark and received his PhD in 2021. In his PhD project, he has been working with in situ and operando electron paramagnetic spectroscopy of environmental catalysts, especially on catalyst materials relevant for the selective catalytic reduction of nitrogen oxides.

Part VII Transient and Thermal Methods

Temporal Analysis of Product (TAP)

40

Rebecca Fushimi

Contents 40.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 899

40.2 40.2.1 40.2.2

Basic Experimental Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . 900 Instrument Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 900 Experimental Concepts, Distinctions . . . . . . . . . . . . . . . . . . . . . 902

40.3

Theoretical Tools for Extracting Kinetic Information from the Pulse Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Collisions in a Diffusion Reactor . . . . . . . . . . . . . . . . . . . . . . . . . Reactor Transport: The Standard Diffusion Curve . . . . . . . Numerical Solution to Diffusion/Reaction Systems . . . . . Model-Free Analysis of Pulse Response Data . . . . . . . . . . .

40.3.1 40.3.2 40.3.3 40.3.4 40.4

904 904 905 908 910

40.4.5

Experimental Studies of TAP Catalyst Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Coverage-Dependent Sticking Coefficients . . . . . . . . . . . . . . Active Site Titration at Working Temperatures . . . . . . . . . . Mechanistic Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Role of Dynamic Surface Species in Reaction Mechanism, Lifetime of Fast Surface Intermediates . . . . . The Pressure Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40.5

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 931

40.4.1 40.4.2 40.4.3 40.4.4

915 916 917 921 927 928

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 931

Abstract

The Temporal Analysis of Products (TAP) pulse response method for characterization of catalyst kinetic properties and mechanistic features is presented. Beginning with details of the instrument configuration and different experimental formats, the basic experimental concepts that distinguish the TAP method from kinetic tools such as molecular beam scattering and pulse response in advecting reactors are discussed. Three guiding principles that define the TAP experiment are (i) insignificant perturbation of the catalyst state, (ii) spatial uniformity of the gas within the active zone of the reactor, and (iii) wellR. Fushimi (*) Catalysis and Transient Kinetics Group, Idaho National Laboratory, Idaho Falls, ID, USA e-mail: [email protected]

defined diffusion in the Knudsen transport regime. In the Knudsen regime, only gas/solid collisions are significant and an intrinsic measurement of catalyst kinetic properties becomes possible. Theoretical tools for extracting kinetic information from TAP pulse response data are described, including numerical solutions, moment-based analysis, and time-dependent analysis of rate and concentration data. Experimental studies are presented that demonstrate the TAP approach to characterization of coveragedependent sticking coefficients, active site titration at working temperatures for both irreversibly and reversibly adsorbing molecules (heat of adsorption), multiple active sites working within mixtures, mechanistic features including adsorption processes and microkinetic network discrimination, and lifetime of surface intermediates as well as connection of kinetic features across the pressure gap. Keywords

TAP Reactor · Temporal analysis of products · Transient kinetics · Pulse response · Catalysis

40.1

Introduction

Basic kinetic evaluation of a catalyst typically takes place in an advecting device operating at steady-state, e.g., the plug flow reactor, PFR, or the continuous stirred tank reactor, CSTR, which offers a global snapshot of performance at conditions similar to the industrial operating environment. Transient experiments, whether used to observe a change in temperature, concentration, or pressure, are more challenging than steady-state, both from an experimental and analysis perspective, but can offer much greater insight into how a catalytic surface controls a chemical reaction. The temporal analysis of products, TAP reactor, uses a pulse transient in a packed bed at low pressure to reveal details of chemical

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_40

899

900

R. Fushimi

kinetic steps that comprise the global reaction [1–3]. Whereas the steady-state device is useful for screening materials to determine which compositions perform better (conversion and yield), the transient device can be used to derive greater understanding of microkinetic features that describe why one catalyst performs better than another. In this chapter, the TAP method for precise characterization of catalyst kinetic properties is presented. The experimental apparatus is described along with key features that distinguish the TAP diffusion reactor from pulse response experiments in advecting devices. Theoretical tools used to extract kinetic features from pulse response data are described including a numerical approach, moment-based techniques, and a more recently developed inverse diffusion method that preserves rate and concentration timedependence. Calling on different published works, experimental methodology and analysis will be described for obtaining the following catalyst characterization information: • Coverage-dependent sticking coefficients • Active site titration at working temperatures – For irreversibly adsorbing probe molecules including discrimination of sites from a working mixture – For reversibly adsorbing probe molecules, equilibrium constants, and heat of adsorption • Mechanistic features – Adsorption mechanisms including competitive and inhibited adsorption – Microkinetic network discrimination • Lifetime of fast surface intermediates and their role in reaction mechanism • Pressure-sensitive reaction kinetics – Distinction of gas phase and surface reactions

40.2

Basic Experimental Concepts

The TAP concept was originally conceived by John Gleaves while working at the Monsanto Company in the late 1970s. Inspired by molecular beam-scattering (MBS) experiments that could capture detailed kinetic information on welldefined single crystal surfaces, the TAP method was developed to provide the same level of detail on real catalysts, without any minimization of their physical complexity. In other words, a system that could address the study of ill-defined, multicomponent, multiphasic, polycrystalline, inhomogeneous materials with abundant defect structures directly from industrial reactors. The key components of a TAP reactor system consist of a high-speed pulse valve, an evacuated packed bed reactor and a mass spectrometer immediately at the exit of the packed bed. The packed bed reactor is primarily composed of inert particles with the exception of

a thin layer of active material at the center. A TAP experiment uses an ultralow pulse intensity (10 nmols/pulse) such that the number of reacting molecules is significantly less than the number of active sites available on the catalyst sample. The low intensity pulse into an evacuated bed results in reactor transport that is purely Knudsen diffusion with negligible gas/gas collisions and first-order gas/solid interactions. These features give rise to three guiding principles that constitute the TAP approach to chemical kinetics: • Insignificant perturbation of the catalyst state • Spatial uniformity of the gas within the active zone of the reactor • Well-defined diffusion in the Knudsen transport regime These principles are what distinguish the TAP experiment from a pulse response in an advecting device where transport is complex and rate expressions are often nonlinear, apparent functions of multiple component concentrations. Much of the theoretical foundation used for deriving kinetic and mechanistic information from TAP pulse response curves was led by Gregory Yablonsky. Gleaves and Yablonsky began collaborating in 1990, and during this early period the guiding principles were established. The TAP principles create an environment where the intrinsic features of the catalyst can be directly assessed, that is, the capacity for, and mechanism, by which the solid chemically transforms molecules from the gas phase. Although the kinetics can be readily addressed due to the high vacuum conditions within the TAP reactor, the high vacuum, particularly in combination with elevated temperatures, may substantially alter reactions, e.g., promotion of auto-reduction or prevention of favorable adsorbate-induced structural changes, as compared to industrial operating conditions. In addition, the Knudsen diffusion regime prevents intramolecular collisions; consequently, gas-phase reactions do not occur. If needed, these phenomena can be investigated further at elevated pressure within the same TAP system.

40.2.1 Instrument Configuration This section describes the TAP-3 system, other TAP systems have been developed, and the main distinguishing feature is their use of diffusion pump or a turbo-molecular pump to generate the high vacuum [4–6]. The basic components of a modern TAP reactor and vacuum system are illustrated in Fig. 40.1a. In the original TAP-1 design, the components were arranged horizontally with differential pumping between the reactor and mass spectrometer chamber; the reader can find more information in references [7, 8]. In the TAP-2 and TAP-3 design, the components are arranged vertically; these include a pulse valve assembly for generating

40

Temporal Analysis of Product (TAP)

901

a)

b) O2, CO, CH4, C2H4, C2H6 C3H8, C3H6 C4H10, Ar, etc.

Pulsevalve assembly Reactant source

Mass spec. chamber Liquid nitrogen chamber

Microreactor Catalyst

Metals, supported metals, mixed metal oxides, zeolites

Mass spectrometer Detector

Diffusion pump

High speed vacuum system

Fig. 40.1 (a) Basic components of a TAP reactor system (Illustration courtesy of Mithra Technologies, Inc.); (b) photograph of a TAP-3 reactor system built by Mithra Technologies, Inc. in 2015. (Photo courtesy of Idaho National Laboratory)

reactant/probe molecule pulses, a microreactor which holds the catalyst sample and a mass spectrometer detection system mounted in a high-speed vacuum chamber driven by an oil diffusion pump and supported by a liquid nitrogen trap. Figure 40.1b shows an example of a modern TAP reactor system, and Fig. 40.2 shows greater detail of the key components. A valve manifold generally consists of two or four pulse valves designed to deliver fast, 100–300 μs, low-intensity pulses, ca. 10 nmols. Reactant sources, oxygenates, hydrocarbons, etc. are mixed with inert gases such as helium or argon that serve to quantify conversion and yield as well as to validate Knudsen flow characteristics. Both flow through and static heated valve feed systems have been used for injecting vapors of low boiling point liquids such as water, methanol, benzene, etc. In advanced designs, the pulse valve manifold eliminates dead volume between the valve aperture and the entrance to the microreactor packed bed. The microreactor supports a quartz, stainless-steel, or similar tube, typically around 6 mm ID
25 mm length, that is resistively heated from the exterior. In some designs, a thermocouple is located at the radial/axial center of the reactor (for direct measurement of the catalyst bed); in others, the thermocouple is fixed to the exterior and the thermal gradient through the reactor wall is measured separately. The reactor is removed from the system and inverted for loading. As shown in Fig. 40.2, the reactor primarily consists of two layers of inert particles and a thin zone of active material (catalyst

powders, particles, wire, gauze, etc.) in between two inert zones; a stainless-steel screen, and snap ring to hold the packing in place. A slide valve assembly is used to isolate the vacuum system while the reactor is removed for loading. The slide valve assembly also enables the reactor to be pressurized and operated as a conventional flow through packed bed reactor. The pulse valve manifold will typically contain a continuous flow valve for admitting oxidant or reductant gases for pretreatment as well as gas mixtures for reaction conditions. The majority of the gas flow exits the slide valve while a small needle valve within the assembly can be used to admit a portion of the gas flow to the mass spectrometer. In a TAP pulse response experiment, the slide valve is retracted, and in some cases the reactor/pulse valve assembly is lowered into the vacuum system to minimize the distance from the reactor exit to the ionizer of the mass spectrometer. Many different mass spectrometers (residual gas analyzers) have been used including the original UTI 100C (MKS Instruments), HAL 3F RC (Hiden Analytical), SRS RGA300 (Stanford Research Systems), and 150 QC (Extrel QMS), among others. Electron impact ionization is utilized, and a mass range of 200 AMU is typically sufficient for catalysis studies. An electron multiplier and signal amplification enable high signal to noise ratio (1000:1) and reliable observation of low-intensity reaction products three orders of magnitude less abundant than the reactants. With the slide valve retracted, the reactor exit is open to the high-throughput vacuum system. Typically, pressures of

40

R. Fushimi

Flow valve

Inert zone Active zone Inert zone

Slide-valve open TAP vacuum pulse response experiments Slide-valve closed Atmospheric pressure steadystate experiments Microreactor

Pulse valve Heater wire

Normalized intensity

902

Mass spectrometer 10–8 torr

High-throughput vacuum pump

Normalized intensity

Slide valve (open)

FWHH = 360 ms Pulse intensity ≈ 10–10 moles

0.01

Thermocouple Ionizer

Input pulse

0.0

0.05 Time (s)

0.1

Outlet response

0.2

0.5 Time (s)

1.0

Fig. 40.2 Key components supporting a TAP pulse response experiment. (Illustration courtesy of Mithra Technologies, Inc.)

108 torr are achieved at the mass spectrometer ionizer. A relatively large vacuum chamber should be used to minimize detection of molecules backscattered from the chamber walls. The pulse intensity should neither significantly influence the chamber pressure nor diminish the flux of ionizing electrons which may lead to splitting of the peak into two maxima. Most TAP reactor systems utilize a large diameter oil diffusion pump (e.g., Varian VHS-400) since this type of pump offers the highest throughput at a reasonable cost. Turbomolecular pumps have been used in some examples, but the cost of achieving sufficient throughput, to avoid an undesirable tailing signal after the pulse, can be significant. In the pulse response experiment, only one atomic mass unit (AMU) is monitored in each pulse. Thus, from one pulse to the next, the mass spectrometer will parse through a series of AMUs of interest, often switching to higher amplification settings for product masses. The raw data obtained such as this must undergo a number of preprocessing steps, baseline correction, filtering, deconvolution, normalization, calibration, etc., prior to interpretation [9].

40.2.2 Experimental Concepts, Distinctions High throughput catalyst screening tools have been designed to rapidly screen large libraries of catalyst compositions for basic kinetic properties, namely, activity and selectivity. This is useful for identifying trends in composition that may lead to superior performance (both activity and selectivity). Although reactor carousels have been designed for the TAP method, their implementation for sample screening has not been emphasized to date. Using current methods, the TAP

technique might successfully screen 10 catalyst samples in a day, but more often researchers spend more time on detailed kinetic characterization of fewer samples. Instead of focusing on composition, the TAP method can be thought of as providing high throughput kinetic data on a smaller set of materials. The substantial quantity of kinetic data comes from two features: (i) the sum of many infinitesimal nanomole pulses that add up to a finite change, a sort of “chemical calculus” [10] that yields a kinetic snapshot at each step (see section “Screening Multiple Active Site Mixtures”), and (ii) the timedependent changes within a single pulse where the reaction rate can be sampled as a function of changing gas surface concentrations (see section “General Reaction Model Analysis,” Kinetic Petals). Two important experimental concepts have been distinguished in the TAP literature: the state-defining and state-altering experiments [11]. In the state-defining experiment, the catalyst composition is not significantly perturbed by the pulse, e.g., the number of active sites or active surface intermediates is sufficiently large compared to the number of probe molecules injected. Thus, while the physicochemical properties of the material do not significantly change during the measurement, a kinetic snapshot of the catalyst state may be captured from the measurement. The state-defining regime can be identified by quantifying the extent to which a pulse shapes changes from one pulse to the next or by examining the independence of the pulse shape on the pulse intensity [11]. From the shape of the pulse response, the kinetic parameters defining the present state of the catalyst are derived (Sect. 40.3.4). For example, in the case of an irreversible hydrocarbon conversion over an oxide catalyst,

40

Temporal Analysis of Product (TAP)

the rate constant is directly related to the oxidation state or oxygen content of the material at the time of measurement. A long series of many such infinitesimal pulses will slowly change the oxygen content, and the changes in the pulse response shape will record the evolution of kinetic properties. This is an example of a state-altering experiment where the experiment changes the catalyst composition in a precisely controlled manner. The state-altering experiment can be used to capture the landscape of kinetic features across a wide domain, e.g., from oxidized to reduced (chemically or thermally), from fresh to deactivated, from promoted to poisoned, etc. The state-altering experiment can also be conducted to evaluate the impact of reaction products on the forward process. TAP operating conditions are far from equilibrium, so products generated from the reactant pulse are not likely to have significant backward reactions, but these common reaction products can be intentionally pulsed together (or in short time delay) with the reactant. Such an experiment could be used to understand the effect of changing the oxygen content by reaction with different reducing agents such as CO or H2 that are common by-products of the overall reaction process. The role of a water by-product on catalyst selectivity could be similarly investigated. When the evolution of kinetic features show inflection, then optimal regimes can be identified. Chemical titrations are not uncommon, but what makes the TAP titration advantageous is the precise kinetic characterization encoded in the pulse response shape for each state-defining pulse. The state-altering experiment can be conducted with a long series of state-defining pulses or with a single large pulse where the number of molecules pulsed is similar to the number of active sites. The pulse response shape is often summarized through the use of integral quantities. This analysis method is convenient but also greatly reduces the quantity and quality of the kinetic data. Exploitation of the time-dependent features was significantly enhanced with the introduction of the Y-Procedure inverse-diffusion method in 2007 [12] (section “TimeDependent Analysis of Rate and Concentration”). This method translates the experimentally observed exit flow into time-dependent rate and concentration in the catalyst bed. Thus, the complex interplay of the gas and surface concentration on the reaction rate can be directly examined. The rate, gas, and surface concentration dependencies, as they evolve over the course of one pulse, yield a significant quantity of kinetic data. These dependences, referred to as kinetic petals, can be used to understand the kinetic mechanism of surface regulation processes [13]. Moreover, they provide a unique identifier or fingerprint for comparison of one catalyst to another. In addition to examination of the singular pulse response, and the evolution of pulses in a long sequence, the pump/ probe experimental format is a powerful tool for examining adsorption and reaction dynamics between two different gases. Figure 40.3 illustrates the pump/probe format in comparison to the single pulse and multipulse modes. In the

903

pump/probe experiment, two pulses are separated by a small time-delay. The surface state can be pumped or primed with one reactant, such as oxygen, and a time-delay imposed before the probe pulse, e.g., a hydrocarbon, is injected. The short delay between the pulses allows for relaxation of the perturbing pulse, perhaps the conversion of oxygen to different active forms, surface to bulk transport, or other surface processes. The probe pulse then samples the catalyst state during the relaxation; by varying the time-delay, the timing of the relaxation processes can be studied. The precision of the kinetic characterization using these different experimental formats comes from the diffusionreactor concept. In the TAP reactor, diffusion provides uniform mixing: an efficient impeller. This also creates the possibility for uniform kinetic measurements at higher conversions (section “Uniformity of the Gas and Solid in the Kinetic Measurement”). In an advecting kinetic device such as the plug flow reactor, diffusion is a disturbing influence and is minimized to the extent possible. The TAP experiment eliminates viscous flow, and the transport mechanism is solely comprised of well-defined Knudsen diffusion, in which the complexity and diffusional resistance arising from intermolecular collisions is avoided. This Knudsen diffusion regime implies that concentration gradients in the interparticle space are negligible; in other words, external diffusion limitations are avoided under TAP conditions. Internal diffusion limitations may still exist, although the distinction between meso- and microporous diffusion disappears when all inter- and intraparticle diffusion is within the Knudsen regime. The small pulse size also eliminates thermal effects (e.g., hot spots, surface restructuring) even for highly exothermic or endothermic reactions. For example, for an N2 exothermic reaction such as ammonia oxidation (ΔH1073 K¼ 635 kJ/mol), 100% conversion of 10 nmols of NH3 to N2 results in a temperature rise of only 2.8 K [15]. The small pulse into an evacuated bed also renders the gas/solid interaction to first-order processes. The reactor transport is precisely modeled by Fick’s second law of diffusion and is hence readily separated from kinetic effects. Operation in high vacuum enables the use of a closely coupled mass spectrometer detection device which offers high time resolution (millisecond) suitable for resolving the time characteristics of most catalytic adsorption and desorption events. In a number of examples, the intrinsic kinetic parameters extracted from low-pressure TAP are corroborated by apparent kinetics captured in high-pressure steady-state laboratory reactors (Sect. 40.4.5). However, higher intensity pulses can be used to depart from Knudsen flow conditions to investigate the effects of gas phase reactions. The TAP pulse response concept has been primarily used as a kinetic characterization tool, to elucidate reaction mechanisms and to distinguish materials based on how they control probe reactions. The methodology can be considered as a

40

904

R. Fushimi

Fig. 40.3 Illustration of the different experimental formats in the TAP reactor. In the pump/ probe experimental format, input pulses of reactants A and B are injected with a time-delay, Δt. Output responses indicating reactant conversion and product formation, P and R, are recorded over the entire duration. (Reproduced with permission from Ref. [14]). (Reprinted from Catalysis Today, 121, PerezRamirez, J., Kondratenko, E., Evolution, achievements, and perspectives of the TAP technique, 160–169, (2007), with permission from Elsevier)

Single pulse experiments Input pulses

Output responses

A Size ~ 1013-1017 molec. Width ~ 100 Ps Pulse rate ~ 0.1-50 s–1

A P

t (s)

t (s)

Multipulse experiments Input pulses

Output responses

A

P

t (s)

t (s)

Pump/probe experiments

Input pulse

A

Output responses

B 't

A

B

P

R

t (s)

t (s)

combination of molecular beam-scattering methods developed by Herschbach [16] and Lee [17] with the relaxation methods developed by Eigen [18]. The concepts of insignificant perturbation and diffusional flow primarily distinguish the TAP reactor from other kinetic tools.

the mean free path, λ (m), calculated according to the kinetic theory of gases, to the diameter of the interstitial voids in the packed bed, dvoid (m):

40.3

where kB is the Boltzmann constant (J/K), T is the absolute temperature (K), P is the pressure (Pa), σ is the collision diameter of the molecule (m), εb is the bed voidage, and dp is the diameter of the bed particles (m). In the Knudsen flow regime, the Knudsen diffusion coefficient of a gas A, DA (m2/s), does not depend on the concentration of the gas mixture and is calculated from the kinetic theory of gases:

Theoretical Tools for Extracting Kinetic Information from the Pulse Response

This section describes the theoretical tools that can be used to extract kinetic information from TAP pulse response data and as such is fairly theory oriented. A reader with a primary interest in the application of TAP can skip ahead to Sect. 40.4, where several examples are presented of how TAP can be used in catalysis research.

40.3.1 Collisions in a Diffusion Reactor According to the three principles that define a TAP experiment (Sect. 40.2), a low-intensity pulse into an evacuated packed bed results in Knudsen diffusion as the dominant mass transport mode for gas molecules. Operation in the Knudsen diffusion regime can be estimated by comparing

k T 2 εb  dvoid ¼ λ ¼ pffiffiffiB d 3 ð 1  εb Þ p 2Pσ m

d DA ¼ void 3

rffiffiffiffiffiffiffiffiffiffi 8RT πMA

ð40:1Þ

ð40:2Þ

where Rg is the ideal gas constant (J/(mol K)) and MA is the molecular weight of the gas A (kg/mol). The square root term represents the mean velocity of the gas. In a packed bed, the effective Knudsen diffusivity, DeA, (m2/s), can be used in place of the Knudsen diffusion coefficient:

40

Temporal Analysis of Product (TAP)

DeA ¼

εb D τb A

905

ð40:3Þ

where τb is the tortuosity of the bed (dimensionless). The collision frequency, Z (s1), is related to the mean free path, λ (m), and the mean molecular velocity, υ (m/s): υ Z¼ λ

ð40:4Þ

If the diameter of the interstitial bed voids is used in place of the mean free path and the velocity is expressed using the effective Knudsen diffusivity, Eq. (40.4) is then: Z¼

υ 3 DeA τb ¼ 2 di d void εb

εb L2 3τ L2 ¼ b 2 2DeA 2 d void

0  x  L, t ¼ 0, CA ¼ δx

ð40:6Þ

where L is the length of the catalyst bed (m). Here, it can be seen that the average number of collisions only depends on parameters related to the reactor packing, reactor length, and diameter of particles. It is often mistaken that the number of collisions depends on the temperature, but when increasing temperature, the velocity of the molecules increases as does the mean free path. As described by Schuurman [19], the average number of collisions in a typical nonporous particlepacked TAP reactor is approximately 2
105. For a pulse size of 2
1014 molecules, every active site is hit on average approximately 1000 times. For measurement of the sticking coefficient, in comparison to a molecular beam-scattering experiment, the random diffusive flow of the TAP reactor offers a significant advantage by increasing the number of collisions with the active site.

@CA ¼0 @x

x ¼ 0,

x ¼ L, CA ¼ 0

@CA @ 2 CA ¼ DeA @t @x2

ð40:9Þ ð40:10Þ

CA ¼

1 X

cosððn þ 0:5ÞπξÞ

n¼0

  2 2
exp ðn þ 0:5Þ π τ x L

ð40:12Þ

tDeA εb L2

ð40:13Þ

ξ¼ τ¼

ð40:11Þ

where CA is a dimensionless form of concentration, ξ is the dimensionless axial coordinate, and τ is dimensionless time. The negative gradient of concentration is the flow, FA, (mol/s): @CA @ξ

ð40:14Þ

and at the reactor exit, ξ ¼ 1, the dimensionless flow is then:

In the TAP experiment, the concentration of gas A, CA (mol/m3), changes over length and time according to Fick’s second law of diffusion: εb

ð40:8Þ

Applying the method of separation of variables, the solution to Eq. (40.7), according to initial condition (40.8) and boundary conditions (40.9) and (40.10) has been described by Gleaves et al. [20] as:

FA ¼ 

40.3.2 Reactor Transport: The Standard Diffusion Curve

N pA εb AL

where NpA is the number of molecules of gas A in the inlet pulse (mol) and A is the cross-sectional areas of the reactor (m2). The boundary conditions specify that there is no flux at the reactor entrance when the pulse valve is closed, and at the reactor exit, the concentration of gas at the reactor exit is nearly zero due to the high pumping speed of the vacuum system:

ð40:5Þ

The average number of collisions, ncoll (dimensionless), is then obtained by multiplying the collision frequency by the average residence time in the reactor, tres(s): ncoll ¼ Z  tres ¼ Z 

gas concentration can be represented by a delta function, δx, at the reactor inlet;

ð40:7Þ

Here, DeA is the effective Knudsen diffusivity as described in Eq. (40.3), t is time (s), and x represents the reactor axial coordinate (m). The initial condition specifies that at t ¼ 0 the

FA ¼ π

1 X

n

ð1Þ ð2n þ 1Þ

n¼0

  2 2
exp ðn þ 0:5Þ π τ

ð40:15Þ

Equation (40.15) is referred to as the standard diffusion curve and describes the dimensionless exit flow as a function of dimensionless time as plotted in Fig. 40.4 along with a number of characteristic features. This curve represents the time dependence of molecules exiting the reactor and is

40

R. Fushimi

Dimensionless flow (FAebL2/NpADeA)

906

2.4 1.9

tp = 0.17

1.4

FA,p = 1.85 t = 0.35 (maximum average flow) t = 0.50 (mean residence time)

0.9 Criterion for Knudsen diffusion FA,ptp=0.31

0.4

–0.1 0.0

0.5

1.0 1.5 2.0 2.5 Dimensionless time (tDeA/ebL2)

3.0

Fig. 40.4 The standard diffusion curve in dimensionless form, representing the exit flow when molecules are governed by Knudsen diffusion only (absence of any kinetic process). (Reproduced with permission from Ref. [20]). (Reprinted from Applied Catalysis A: General, 160, Gleaves, J.T., Yablonskii, G., Phanawadee, P., Schuurman, Y., TAP-2: an interrogative kinetics approach, (1997), with permission from Elsevier)

constant for any given gas, bed length, particle size, or reactor temperature. It is the signature of the TAP pulse response experiment and provides a “measuring stick” against which a pulse response of reactant and product molecules can be compared. The spatiotemporal profile inside the TAP reactor when an inert gas pulsed in the Knudsen regime as presented by Schuurman [19] is shown in Fig. 40.5a. From this data, the unique operation of the TAP reactor can be made apparent. The change in pressure over the length of the reactor is shown in Fig. 40.5b. Over a short period of time near the reactor entrance, the Knudsen criterion will be violated. The pressure gradient, however, is rapidly diminished with no more than 1 standard deviation after 5 milliseconds. In a TAP experiment, the pulse quickly spreads out over the length of the empty reactor and then, much more slowly, the total concentration of the reactor is diminished as molecules escape the exit. Thus, the center of the reactor, where the catalyst is typically placed, is well-mixed and will have a flat concentration profile for noninteracting molecules. The physics of a pulse into an advecting device are very different; in this case, the concentration peak travels through the reactor [21], Fig. 40.5c, forced by the steady flow of the carrier gas (see also simulations by Rothaemel and Baerns [22]). In the diffusion reactor, the axial gradient is always negative by comparing Fig. 40.5a, c. In the TAP reactor, diffusion can be viewed as a “measuring stick” against which the timecharacteristics of kinetic processes are judged. In an advecting reactor, however, diffusion is a disturbing factor which causes axial dispersion and broadening of the peak shape.

Active Zone Configurations The solution to Eq. (40.7) is valid over the entire reactor. In practice however, the majority of the TAP reactor is simply inert material, e.g., quartz or silicon carbide particles that give rise to random gas molecule collisions creating well-mixed diffusional flow. The active material can be loaded in different configurations, e.g., a layer of particles, a single particle [25], gauzes [15, 26], and monoliths [27, 28]. Almost predominantly, a thin zone configuration is used in TAP reactor studies; a layer of catalyst particles is located at the center of the reactor between two zones of inert particles. As a guide, the thickness of the thin zone should not exceed 10% of the total reactor length; this enables a physical decoupling of the transport and kinetic processes, and concentration gradients across the catalyst bed can be neglected [29]. In the case of a thin zone configuration [30], Eq. (40.7) applies in the inert zones: εin

@CA @ 2 CA ¼ Din @t @x2

ð40:16Þ

where εin and Din refer to the bed voidage and diffusivity in the inert zones, respectively, and in the catalyst zone: εcat

@CA @ 2 CA  RðCA , Cθ , N m Þ ¼ Dcat @t @x2

ð40:17Þ

where εcat and Dcat refer to the bed voidage and diffusivity in the catalyst zone, respectively, and R(CA, Cθ, Nm) describes the gas reaction rate (mol/(m3s)), which is a complex function of the gas concentration, CA, surface concentration (mol/m3), Cθ (mol/m2), and fixed properties of the catalyst, Nm, such as the number of active sites. Mathematically, the boundaries between the catalyst and inert zones are accounted for by appropriate matching conditions between the zones. These are the continuity of gas concentration given in Eq. (40.18) and a matching condition for the diffusional flows that includes a term accounting for reaction at the thin zone boundary given Eq. (40.19): CI ðx, tÞjx¼LTZ ¼ CII ðx, tÞjx¼LTZ ¼ CTZ ðtÞ

@CI ðx, tÞ @C ðx, tÞ þ ADin II j jx¼LTZ @x x¼LTZ @x ¼ FlowI ðx, tÞjx¼LTZ  FlowII ðx, tÞjx¼LTZ

ð40:18Þ

ADin

ð40:19Þ

¼ ALcat RTZ ðCA , Cθ , N m Þ

where CI and CII are solutions for the gas concentrations in the first and second inert zones, LTZ is the length of the thin zone (m), and CTZ(t) is the spatial average concentration in the thin zone; Lcat is the length of the catalyst zone (mol/m3),

40

Temporal Analysis of Product (TAP)

907

40

a)

100

b) 1.00

60 40

0

0.2

0.6 0.8

0.4

0.6 0.8 t/(HL 2 /D )

1

Normalized concentration

0.2

20

Time (s) 1e–05 0.00025 0.0025 0.005

0

0.4 Z/L

Pressure (Pa)

80

1

c)

0.75

0.50

0.25

Response

0.00 0.00 0.06 0.12

0.20

0.32

0.48

Dimensionless reactor length 1 1 Time Position 0

Fig. 40.5 (a) Spatiotemporal profile inside the TAP reactor for a pulse of 1 nmol of argon into a reactor (4 mm diameter, 25.4 mm length, 0.25 mm particle size, 0.4 bed voidage, and 5.5 bed tortuosity) at 298 K. (Reproduced with permission from Ref. [19] Reprinted from Catalysis Today, 121, Schuurman, Y., Assessment of kinetic modeling procedures of TAP experiments, 187–196, (2007), with permission from Elsevier); (b) simulated concentration as a function of reactor length taken at early time snaptshots following a pulsed injection of 1
1015 molecules of argon at room temperature into a 4 mm
60 mm evacuated reactor

containing 210–250 μm spheres. Derived from the model described in Ref. [24]; and (c) simulated data for a pulse into flow through a packed bed with superficial velocity ¼ 2 m/s and Peclet number ¼ 100. (Reproduced with permission from Ref. [21]) (Reprinted from Applied Catalysis A: General, 151, van der Linde, S.C., Nijhuis, T., Dekker, F., Kapteijn, F., Moulinj, J.A., Mathematical treatment of transient kinetic data: combination of parameter estimation with solving the related partial differential equations, 27–57 (1997), with permission from Elsevier.)

and RTZ(CA, Cθ, Nm) is the spatial average reaction rate (mol/(m3 s)).

configuration; there the reactor profile can be approximated by a linear function, and the reaction rate becomes a function of the spatial concentration average; and R is the average of Cin and Cout. However, in the TAP reactor, being diffusiondriven, the rate is instead determined by the difference in flux or flow in/out of the active zone, that is, the concentration gradients, as in Eq. (40.19). Shekhtman and Yablonsky analyzed the nonuniformity of the differential PFR and the thin zone TAP reactor [31]. In terms of conversion, they approximate the nonuniformity in the thin zone TAP according to:

Uniformity of the Gas and Solid in the Kinetic Measurement In any kinetic characterization, a reliable rate measurement requires uniformity of the gas phase and catalyst composition. This is the primary reason why kinetic experiments in the convection-driven plug flow reactor (PFR) are restricted to conversions less than 20%. The axial concentration gradient can be minimized by adopting a differential PFR

908

R. Fushimi

Cin  Cout L X 2 cþ Cin Lr 1 þ ð1  XÞLr=Lc

ð40:20Þ

where Lc is the length of the catalyst zone (m) and Lr is the length of the reactor (m). Figure 40.6a compares the nonuniformity for both devices as a function of conversion. In the TAP reactor, nonuniformity can remain low over a much wider range of conversion; nonuniformity only exceeds 20% at conversions >20%. Figure 40.6b provides greater insight by examining the axial concentration profile in the thin zone TAP. It can be seen that the difference in concentration at the boundaries of the thin zone are not as dramatic as the difference in gradients. As a result, the thin zone TAP reactor can sample uniform rate measurements at much higher conversion. Equation (40.20) is an estimate, and Wang et al. investigated the impact of high conversion measurements in a thin zone TAP reactor using ammonia decomposition [32]. At high temperatures, conversion was near 90% (gas concentration nonuniformity is 24% according to Eq. (40.20)); however, they found that the first-order Arrhenius plot showed a high correlation coefficient across a wide domain of conversion (from 10% to 90%). Furthermore, they did not observe deviation in the shape of the pulse response across this range or with 2000 pulses at isothermal sampling. Thus, even at 90% conversion, they concluded that gas concentration and catalyst composition were not significantly impacted by nonuniformity. This finding that a TAP thin zone can be considered uniform has been corroborated by

In the catalyst zone, the mass balance for gas phase species includes a set of terms that comprise the complex gas transformation as it depends on the gas and surface concentrations, along with the fixed properties of the catalyst, phenomenologically R (CA, Cθ, Nm) in Eq. (40.17). A source term for the rate along with a mass balance for surface species can be accounted for explicitly by a kinetic model: Rj ¼ k j

Y n Y nm Ck θ

ð40:21Þ

m

k

Then we have for Eq. (40.17):

b) 2.00

1.0

1.75

Differential PFR

0.8 Non-uniformity

40.3.3 Numerical Solution to Diffusion/Reaction Systems

Dimensionless gas concentration

a)

Phanawadee et al. [30]. The thin zone TAP reactor is thus a very near approximation to a perfectly mixed reactor which arises from diffusion playing the role of an “efficient impeller.” There is a trade-off however for samples with low activity, the thickness of the catalyst zone should not exceed 10% of the reactor length to be approximated as a thin zone, and the quantity of active sites should be sufficient to maintain an insignificant change during the pulse response. In such scenarios, it may be beneficial to decrease the particle size of the material to increase the active surface area available for testing.

0.6 TZTR, Lr/Lc=20 0.4 TZTR, Lr/Lc=30 TZTR, Lr/Lc=50

0.2

Lr /2

1.50 Lc 1.25 1.00

Inert zone I

.75

0.0

0.2

0.4

0.6

0.8

1.0

X

Fig. 40.6 (a) Nonuniformity of gas concentration as a function of conversion for the differential plug flow reactor (PFR) and thin-zone TAP reactor at different ratios of reactor length, Lr, to catalyst zone length, Lc; (b) axial concentration profile in the thin zone TAP reactor. Dashed lines indicate nonuniformity in the catalyst zone. (Reprinted

0.00 0.0

'Ccat,R

'Ccat,G

.50 .25

0.0

Inert zone II

'Ccat Cin

.1

.2

CTZ

.3 .4 .5 .6 .7 .8 Dimensionless axial coordinate

.9

1.0

with permission from Shekhtman, S.O., Yablonsky, G.S.: Thin-zone TAP reactor versus differential PFR: analysis of concentration nonuniformity for gas-solid systems. Ind. Eng. Chem. Res. 44, 6518–6522 (2005). Copyright 2005 American Chemical Society)

40

Temporal Analysis of Product (TAP)

εcat

909

Xn @CA @ 2 CA ¼ Dcat  ð1 εcat Þσaυ j¼1 υij Rj 2 @t @x

ð40:22Þ

and for the surface: @θA Xn ¼ υ R j¼1 ij j @t

ð40:23Þ

where υA is the fractional coverage of component A (dimensionless), σ is the surface site density (mol/m2), aυ is the specific surface area (m2/m3), and υij are stoichiometric coefficients. This model applies to nonporous catalyst pellets, and the case of microporous materials is not included here; the reader is referred to the excellent references in the literature that deal with this subject [28, 33–41]. The use of Eqs. (40.22) and (40.23) can be demonstrated with the reaction of CO oxidation via the LangmuirHinshelwood-competitive adsorption mechanism along with the corresponding rate equations: kþ 1

COðgÞ þ Z

!

COZ

k 1 þ

ð40:24Þ



RCO ¼ k1 CCO θZ  k1 θCO kþ 2

O2 ðgÞ þ 2Z

!

2OZ

k 2 þ

2

ð40:25Þ

 2

RO2 ¼ k2 CO2 θZ  k2 θO k3

COZ þ OZ ! CO2 ðgÞ þ 2Z RCO2 ¼ k3 θCO θO

ð40:26Þ

where θI is the fractional coverage of the ith species or empty sites. The following mass balance equations in the catalyst zone must be solved simultaneously: εcat

@CCO @ 2 CCO ¼ Dcat @t @x2

εcat

  þ   ð1 εcat Þσaυ k1 CCO θZ  k1 θCO ð40:27Þ

@CO2 @ 2 CO2 ¼ Dcat @t @x2

  þ 2  2  ð1 εcat Þσaυ k2 CO2 θZ  k2 θO

εcat

ð40:28Þ

@CCO2 @ 2 CCO2 ¼ Dcat @t @x2  ð1 εcat Þσaυ ðk3 θCO θO Þ

ð40:29Þ

@θCO þ  ¼ k1 CCO θZ  k1 θCO  k3 θCO θO @t

ð40:30Þ

@θO þ 2  2 ¼ 2k2 CO2 θZ  k2 θO  k3 θCO θO @t

ð40:31Þ

1 ¼ θZ þ θCO þ θO

ð40:32Þ

In state-defining experiment, where the number of active sites is large compared to the number of adsorbing molecules, the dependence on θZ can be neglected. Nonetheless, these governing equations cannot be solved explicitly, and numerical integration must be used. The partial differential equations (PDEs) must first be converted to ordinary differential equations (ODEs), and two approaches have been used: one using the Laplace transformation and one using numerical integration. Following the Laplace transform, a numerical Fast Fourier Transform (FFT) algorithm is used to calculate the exit flow. This method however is limited to linear equations, i.e., first-order kinetics and low surface coverages [19, 42–45]. For numerical integration of the PDEs, first a spatial approximation, for example, the method of lines [46], must be used to discretize the axial coordinate [21]. The axial length of the reactor is divided into a grid, and the derivatives at each grid point are approximated by finite difference. Next, an ODE solver that can accommodate stiff ODEs is applied [47–51]. Several different robust numerical methods for solving coupled ODEs have been discussed by van der Linde et al. [21]. Simulation tools such as these are convenient for exploring subtle differences in reaction mechanism. For example, Kondratenko et al. explored six different microkinetic mechanisms for N2O decomposition. They found only one model where simulated pulse response curves for N2O, N2, and O2 showed strong correlation with experimental data [47]; more discussion can be found in section “General Reaction Model Analysis.” Estimation of rate constants from experimental data requires a nonlinear regression routine that varies the model parameters in order to minimize the difference between the simulated and experimental data. These methods rely on good initial values in order to converge on a global minimum. For example, the paper by Kumar et al. solves the TAP equations numerically and demonstrates how the use of pump-probe timing can alter the selectivity to NH3/N2, therefore provide insight into the reaction pathways [52]. Furthermore, the reliability of regressed parameters should also be statistically assessed, e.g., F-test for variation of estimates across pulses or sampling time, confidence limits, joint confidence regions, Durbin-Watson test for autocorrelation in the residuals, etc.; see examples in reference [19]. Sensitivity analysis should be conducted at a number of distinct experimental conditions to verify if the parameters in question can even be correctly estimated from the

40

910

R. Fushimi

experimental data, or perhaps to identify the desired experimental regime a priori. Similar to the degree of rate control used in computational catalysis [53], a sensitivity analysis on the kinetic parameters can be used to simplify models by identifying rate-limiting steps. Parameter sensitivity is dependent on derivatives and requires the Jacobian matrix of the system. Methods for derivative calculation should be chosen that do not significantly create noise. An example can be found in Logg et al. [54]. In the TAP experiment, only gas phase measurements are possible; information about the accumulation of surface species is inferred from this data. One advantage of the modelfitting approach is that confident estimates in the regression of gas phase exit flow will also yield time-dependent information of changes in surface species, “accumulation dynamics.” This information can be directly examined for understanding storage properties and surface regulation of different species.

40.3.4 Model-Free Analysis of Pulse Response Data One drawback of starting data analysis with an explicit kinetic model, as described in the previous section, is that oftentimes validating the kinetic model is the goal of the investigation. The strong coupling of reaction and transport complicates parameter estimation, and as the complexity of the model grows, so too does uncertainty and the challenge in discriminating multiple solutions when fitting data. To compare one catalyst to another, one needs an intrinsic set of kinetic quantities that are not biased by the experimental device (e.g., heat and mass transport limitations, complex transport models) nor the choice of a kinetic model. Experimental reactor data is a mixture of both transport and kinetic phenomena that can be difficult to decouple, and a primary kinetic characterization should be derived directly from pure kinetic data. Using a steady-state plug flow reactor (PFR), the reaction rate can be determined in a kinetic model-free manner by taking the derivatives of experimental conversion versus space time Fig. 40.7 Analysis workflow for deriving moment quantities from exit flow data. Moments and Shekhtman reactivities can be used for kinetic characterization without assumption of a model. From these quantities, kinetic and transport parameters can be determined ex post facto

data. In a nonsteady-state PFR, an accumulation term must be included and requires assumptions about the details of the kinetic mechanism. The thin zone TAP experiment enables both physical and mathematical decoupling of transport from kinetic phenomena. As in Fig. 40.7, this section describes a primary kinetic analysis based on simple integral quantities, moments of the exit flow data, or more advanced features, Shekhtman reactivities, that are derived from moment data. These quantities, described in section “Shekhtman Reactivities”, can form the basis for comparison of kinetic function among a set of materials with measures that describe the ability for a solid to transform the gas phase. From these integral quantities, if desired, a series of different kinetic models can be scrutinized ex post facto. Integral analysis is convenient and well-developed but reduces the transient data to a singular feature and a great deal of information lost. The Y-Procedure was introduced as a method to preserve the transient [12], section “TimeDependent Analysis of Rate and Concentration.” The transient rate and concentration represent pure kinetic data from which changes in reaction rate constants can be observed during the pulse response. By examining how the concentration and rate evolve together, one can then propose plausible kinetic models.

Primary Analysis: Moment-Based Quantities The primary analysis of pulse response data based on moments, or time-weighted integral quantities, from a mathematical perspective is fairly straightforward and has been widely used to address dynamic reactors [55]. The moments, Mn, of the exit flow are defined by a set of integral equations: Mn ¼

ð þ1 0

n

t Fexit ðtÞdt

ð40:33Þ

The zeroth moment, M0, is simply the area under the pulse response and represents the total number of molecules exiting the reactor. When calibration is performed by pulsing known quantities of gas into an inert reactor, M0 can be used to

Derived characteristics

Primary data +f

tn Fexit (F)dt

Mn =

Exit flow (mol/s)

Moments

0

Time (s)

{r0, r1, r2}

Shekhtman reactivities

{kads/des, kr, tp}

Kinetic and transport parameters

Kinetic model-free Kinetic model-specific

40

Temporal Analysis of Product (TAP)

911

calculate conversion, X (dimensionless), and yield in reacting systems, for example, X ¼ 1  M0

ð40:34Þ

In a process with no conversion, e.g., porous transport or reversible adsorption, M0 will not change with temperature. The first moment, M1, is the integral when the pulse is weighted by time and can be used to calculate the average diffusional reactor residence time (s), τres, when normalized to the number of molecules in the pulse [20]: τres ¼

εb L2 M1 ¼ 2De M0

ð40:35Þ

Note that the diffusivity of the gas may be determined directly from Eq. (40.35). Comparison of the reactor residence time for adsorbing molecules and inert molecules can be used to gauge the average delay in transport through the reactor void space as a result of the kinetic process. The second moment, M2, with dimension of molecules multiplied by time squared, represents the dispersion of residence times for different molecules and compares the residence time of a molecule to the average residence time. The second moment is related to the time-characteristics of confined processes, e.g., desorption time or residence time in catalyst pores. As the order of the moment increases, the contribution of information from the tail portion of the curve increases. Only three moments, zeroth, first, and second, can be reliably extracted from TAP pulse response data; higher order moments are generally unreliable due to experimental noise. Thus, these three moments have physicokinetic meaning and can be used in a primary analysis of pulse response data for comparison of different materials. The derivation of model-specific kinetic parameters from moment data has been described for simple irreversible and reversible adsorption as well as a two-step catalytic mechanism [29, 56]. For example, for an irreversible reaction in a thin zone TAP reactor, the apparent rate constant can be estimated from the conversion and residence time according to: kapp τres,cat ¼

X 1X

ð40:36Þ

where τres,cat is the residence time for diffusion through the catalyst zone, τres,cat ¼

εb Lcat LII M1 Lcat ffi De M0 L

ð40:37Þ

Lcat is the length of the catalyst thin zone, LII is the length of the second symmetric inert zone; see derivations in references [30] and [29]. Equation (40.36) is analogous to the

expression found for a first-order reaction in a CSTR and emphasizes the designation of TAP as a perfectly mixed reactor. For reversible adsorption, although a true equilibrium is not obtained, the relative rates of adsorption and desorption can be captured in an apparent equilibrium constant, Keq, calculated from the first moment using the following Eq. (40.57): K eq ¼ ðM1 1=2Þ

τres L ¼ ðM1 1=2Þ Lcat τres,cat

ð40:38Þ

when the catalyst thin zone is at the center of the reactor. From the temperature dependence of the equilibrium constant, the enthalpy and entropy of adsorption can then be predicted using the Van’t Hoff equation:   ΔH ΔS ln K eq ¼  þ RT R

ð40:39Þ

Shekhtman Reactivities Instead of an explicit kinetic model, as described in Sect. 40.3.3, Shekhtman and Yablonsky [57] describe a phenomenological model that provides a basis for developing a detailed reaction mechanism, or allows for screening of catalysts on a desired characteristic, e.g., promoted catalysts that facilitate desorption of an intermediate. In state-defining experiments, where there is insignificant chemical perturbation, any elementary reaction can be considered a linear function of the gas or adsorbed intermediate concentrations [57, 58]. These considerations allow for a polynomial approximation of the gas-phase component consumption/production rates: RA ðt, xÞ ¼ r 0 CA ðt, xÞ þ r 1 þ r2

dCA ðt, xÞ dt

d2 CA ðt, xÞ þ ... dt2

ð40:40Þ

The coefficients, rin(Ccs), are referred to as Shekhtman reactivities (While they were originally referred to as Laplace reactivities, they were colloquially renamed as Shekhtman reactivities in memory of their developer, Sergiy Oskavovich Shekhtman (1970–2018)), and represent basic kinetic coefficients for a specific probe molecule and are only a function of the catalyst state; they do not depend on time, the concentration of reacting surface intermediates, Knudsen diffusivity, reactor packing, etc. As such, they can be used as an intrinsic catalyst “fingerprint” for direct comparison of kinetic properties among a set of different catalysts. The Eq. (40.40) describes the nonsteady-state reaction rate of gas component A as a function of the concentration of A and its derivatives. At steady-state, the rate is determined

40

912

R. Fushimi

by the first term only which is linear with respect to concentration with r0 as the proportionality constant. The Shekhtman reactivities can be described as follows: • r0 has the dimension of reciprocal seconds and represents an apparent kinetic constant for linear dependence of the rate and gas concentration. • r1 is dimensionless and represents an apparent constant for the balance of consumption and release of the gas on the surface; qualitatively, it can be thought of as a “reversibility” parameter [59]. • r2 has the dimension of seconds and represents an apparent time delay caused by processes on or in the solid material (e.g., pore residence time in the case of microporous diffusion). The physicochemical meaning becomes clearer when the reactivities are expressed in terms of a detailed kinetic mechanism (Table 1 in reference [57]). For example, for simple irreversible adsorption and reaction: ka

A ! AZ kr

AZ ! B r0 ¼ ka; for the reactant |r1| ¼ 0 while for the product jr 1 j ¼ ka=kr and jr2=r1 j ¼ 1=kr : For reversible adsorption: ka

A

!

AZ

kd

r 0 ¼ 0; jr 1 j ¼ ka=kd and jr2=r1 j ¼ 1=kd Hence a single reversible reaction, a process with no conversion, can be quickly identified where r0 ¼ 0 and an irreversible reaction where r1 ¼ 0. A number of additional mechanisms with greater complexity (e.g., adsorption and microporous diffusion) have been analyzed and the reactivities expressed in terms of kinetic parameters [57]. The Shekhtman reactivities can be determined directly from the integral quantities of the experimentally observed exit flow data. For the case of the thin zone TAP reactor, the reactivities are expressed in terms of exit flow moments: r 0,r ¼ r 1,r ¼ 

1 1 þ τr τr M0,r

M1,r 2 τin 1 τin  þ 3 τr 3 τr M0,r M20,r

ð40:41Þ ð40:42Þ

r 2,r ¼

4 τ2in 7 τ2in 1 τin M1,r 1 M2,r þ   45 τr 90 τr M0,r 3 τr M20,r 2 τr M20,r þ

M21,r

ð40:43Þ

τr M30,r τr ¼

Lcat Lin 2Dr

ð40:44Þ

τin ¼

εb,in L2in 2Dr

ð40:45Þ

where r and in indicate the reactant and inert, respectively, τr and τin are the residence time of the reactant in the catalyst zone and the residence time in both inert zones, respectively. Similar expressions have been derived for the products [59, 60] and a convenient worksheet tool is available that will directly calculate reactivities from exit flow moments [59]. The Shekhtman reactivities (r0, r1, and r2) are useful quantities for comparing multiple catalyst samples in that they: • Represent basic kinetic coefficients that relate the rate of chemical transformation to concentration and concentration derivatives. • Are functions of the catalyst composition and temperature and not dependent on the dynamic gas concentration (i.e., an intrinsic measurement of the material). • Are directly calculated from the observed moment of exit flow pulse response data hence can be viewed as experimentally measured characteristics. • Contain no assumptions regarding a kinetic model. • Have been applied to a number of different kinetic models where convenient solutions for rate constants may be determined. While the moment-based analysis approach is straightforward, convenient, and well-developed, it also reduces the rich time-dependent information of the transient experiment to a singular data point. More advanced analysis techniques are available that enable analysis of the pulse response with the time-dependence intact.

Time-Dependent Analysis of Rate and Concentration The Y-Procedure analysis method is a more powerful method for decoupling transport and kinetic data while keeping the time-dependence intact [12, 61]. This inverse-diffusion method enables one to reconstruct the exit flux as a time-dependent rate and concentration in the catalyst thin zone. By preserving the time-dependence, one can examine the interplay of the rate as it evolves in response to the gas concentration transient. This is

40

Temporal Analysis of Product (TAP)

913

significant because in the transient regime the gas and surface concentrations are decoupled; at steady-state, they are fixed. As a result, the evolution of the gas and surface concentration, together with the rate, enables one to sample a dynamic range of kinetic states as well as derive mechanistic understanding for how the surface transforms the surrounding gas and regulates different species. This quantitative rate, gas, and surface concentration dependence creates a detailed “fingerprint” for characterizing a catalytic material and for distinguishing one from another at a very fundamental level. The experimental observable in the TAP experiment is the exit flux or flow, which contains both transport and kinetic information. Transport information can be decoupled, while leaving the time-dependence intact, by subtracting the response of the same molecule through an inert bed, or by scaling an inert gas pulse, FInert, according to Graham’s law: rffiffiffiffiffiffiffiffiffiffiffi MInert FA ¼ FInert MA

ð40:46Þ

where MInert is the molecular weight of the inert gas and MA is the molecular weight of the probe molecule. Equation (40.46) is based on a proportionality of Eq. (40.2) with all other reactor parameters being equal. Equation (40.46) can only be applied to the exit flow of an inert molecule: While transport scales according to Graham’s law, any kinetic features of the response curve will not.

Inert zones I

Mathematically, the Y-Procedure, as described in detail in reference [12], first moves the transport Eq. (40.7) into the Laplace domain: @ 2 LCðx, sÞ , for x0 < x @x2 < x0 þ Δx

εb sLCðx, sÞ ¼ De

where an exact solution for concentration can be found:  rffiffiffiffiffiffi sεb LCðx, sÞ ¼ AðsÞ cosh x De  rffiffiffiffiffiffi sεb þ BðsÞ sinh x De

∂CI ∂t

= Din

Catalyst zones

Inert zones II 2 Continuity matching conditions

∂ 2CI ∂x2

ecat

3-zones model ∂ 2C =D Rg(C,X) ∂t ∂x2

ein

∂C

CI(x, t) –D

∂CI(x, t) ∂x

x=LTZ

= CII(x, t)

+ D x=LTZ

Fig. 40.8 Conceptual demonstration of the Y-Procedure for “dragging” the inlet and exit pulse to the reactor catalyst zone. The schematic is rotated 90 of the physical configuration with the pulse entering on the left and exiting from the right. (Reprinted from Chemical Engineering

∂CII(x, t) ∂x

∂CII ∂t

= Din

∂ 2CII ∂x2

Vacuum BC at the exit:

TZTP model Inlet pulse BC: ∂C ArDin = P(t) ∂x x=0

ð40:48Þ

where A(s) and B(s) are functions that are determined by the boundary conditions (inlet pulse and exit flux) that are described by the experiment being analyzed. The derivative of (40.48) with respect to x can be used to similarly describe the flux. Equation (40.48) essentially instructs how the concentration, and hence flux, translates along the reactor position. Conceptually, this equation and its derivative are used to “feed forward” the inlet concentration to the left boundary of the catalyst zone and “pull backward” the exit concentration and flux to right boundary of the catalyst zone as depicted in Fig. 40.8; note that the depiction of the reactor is rotated 90 of the physical configuration with the pulse

2 Continuity matching conditions ein

ð40:47Þ

C

=0 x=Lr=2Lin+Lcat

x=LTZ

= LcatRTZ

Observed exit flow:

x=LTZ

Science, 62, Yablonsky, G.S., Constales, D., Shekhtman, S.O., Gleaves, J.T., The Y-procedure: how to extract the chemical transformation rate from reaction-diffusion data with no assumption on the kinetic model, 6754–6767, (2007), with permission from Elsevier)

40

914

R. Fushimi

LR

active

 ðsÞ ¼

pffiffiffiffiffiffi sinh sτ3 outlet pffiffiffiffiffiffi LF ðsÞ γ 3 sτ3

ð40:49Þ

sffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi! τ1 γ 21 sinh sτ1 sinh sτ3 pffiffiffiffiffiffi ðsÞ ¼ cosh sτ1 τ3 γ 23   outlet outlet
LFdiffusion ðsÞ  LF ðsÞ pffiffiffiffiffiffi cosh sτ3 þ

ð40:50Þ ðΔxÞ2 De

ð40:51Þ

De ðΔxÞ

ð40:52Þ

τ ¼ εb γ¼

where τ and γ summarize reactor parameters for convenience, subscripts 1 and 3 are used to designate reactor properties in the two inert zones, and Foutlet diffusion represents the outlet flux for the reactant in an inert reactor (or the inert flux scaled according to Eq. (40.46) to adjust for molecular weight); this term is zero when calculating the rate of product formation. It can be seen that the first terms on the right hand side of Eqs. (40.49) and (40.50) represent trigonometric operations that scale the outlet flux to physical reactor parameters; they account for transport and accomplish the inverse diffusion through the reactor packing according to length, bed voidage, and diffusivity. Equations (40.49) and (40.50) form the basis of the Y-Procedure and are used to translate the experimentally observed exit flux (in the Laplace domain) to the catalyst zone concentration and reaction rate, as a function of time. Equations (40.49) and (40.50) are more efficiently utilized if switched to the Fourier domain first; there they become: FC

active

 ðωÞ ¼

pffiffiffiffiffiffiffiffiffi sinh iωτ3 outlet pffiffiffiffiffiffiffiffiffi F F ðωÞ γ 3 iωτ3

ð40:53Þ

1. Scale the exit flux of the inert molecule according to molecular weight of the probe molecule, Eq. (40.46). 2. Apply the Fourier transform to the outlet flux and scaled inert flux. 3. Using the diffusivity calculated from Eq. (40.35) and reactor physical parameters, εb, and position of the catalyst bed (typically reactor length/2), calculate the concentration and rate dependence using Eqs. (40.53) and (40.54). 4. Apply smoothing to reduce high-frequency contributions to noise. Apply the inverse Fourier transform to determine Cactive(t) and Ractive(t). The time-dependence of the transformation rate for reactants and products can be directly compared to evaluate sequential product formation. However, it can be more informative to examine the rate and concentration examined together. Similar to rate/concentration (RC) dependencies captured in steady-state experiments, Fig. 40.9a, in the transient experiment, the RC dependency in the transient pulse response experiment forms a petal shape, Fig. 40.9b. As described by Wang et al., a petal shape arises when the dynamic experiment forces the concentration and rate to return to zero [13]. The petal width, branches, and

a)

High pressure limit r = f(k(T), Nm)

b)

(I)

A

R

LC

active

noise. Filtering can be achieved by multiplication of F Foutlet(ω) with the function exp(ω2h2σ s2/2) where σ s > 0 is a smoothing parameter. One must exercise caution, however, in the application of σ, as oversmoothing can lead to distortion of experimental results [9, 12]. Analysis of experimental noise and filtering strategies have been additionally discussed by other researchers [62, 63]. From the experimentally observed exit flux as a function of time, the steps for implementing the Y-Procedure method consist of the following:

R

entering from the left and exiting from the right. Continuitymatching conditions apply at the boundaries of the catalyst and inert zones. The exact mathematical solutions for the active zone concentration and rate are determined from the exit flux according to:

Low pressure limit r = f(k(T), C, Nm)

B (II)

sffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffi! C C τ1 γ 21 sinh iωτ1 sinh iωτ3 active pffiffiffiffiffiffiffiffiffi FR ðωÞ ¼ 2 τ3 γ 3 cosh iωτ1   outlet outlet Fig. 40.9 Typical rate/concentration dependence encountered in (a)
F Fdiffusion ðωÞ  F F ðωÞ pffiffiffiffiffiffiffiffiffi cosh iωτ3 þ

ð40:54Þ One drawback of analyzing experimental data into the Fourier domain is the amplification of high-frequency

steady-state CSTR experiments, (b) transient pulse response experiments. (Reprinted with permission from Wang, Y., Kunz, M.R., Constales, D., Yablonsky, G., Fushimi, R.: Rate/concentration kinetic petals: a transient method to examine the interplay of surface reaction processes. J. Phys. Chem. A. 123, 8717–8725 (2019). Copyright (2019) American Chemical Society)

40

Temporal Analysis of Product (TAP)

915

directionality, along with the location of the rate and concentration maxima, are features of the petal that can be used to distinguish how different catalysts control chemical reactions. The petal width is a qualitative indication of the degree to which the catalyst changes during the kinetic measurement. A narrow petal, approximately a line, indicates a statedefining experiment where the change in available active sites or active species is insignificant. A wide petal shape is an indication of a significant change in the catalyst during the measurement, a state-altering experiment. Two branches can be identified from the petal shape; in Fig. 40.9b, the ascending branch (I) and descending branch (II) are characterized by an increase and decrease of the gas concentration during the pulse. In the gas-solid interaction of an adsorbing reactant, the slope of the ascending branch represents the apparent rate constant which includes the number of free sites, @R=@C

g

¼ kapp ¼ kf ðθÞ

Cs ðtÞ ¼

0

  X   þ 0  0 0 ν R t ν R t dt  i i i i i i

40.4

Experimental Studies of TAP Catalyst Characterization

ð40:55Þ

where θ indicates the nonsteady state surface coverage. As adsorption proceeds and the surface sites are occupied by the reactant, the number of free sites decreases while the gas concentration from the pulse injection continues to rise; the reaction rate exhibits a plateau (a Langmuir-type dependence). If additional consequences impact the rate at high surface coverage, e.g., inhibiting lateral interactions, then the rate may experience a maximum, as opposed to a plateau while the gas concentration continues to increase. Following the concentration maximum, the rate may decline or extinguish (indicating an exhaustion of free active sites) on the descending branch. The directional (clockwise versus counterclockwise) of the petal shape can be used to understand the adsorption and product formation mechanism (more discussion in section “General Reaction Model Analysis,” Kinetic Petals). In any state-altering petal, there will be two values of the reaction rate at any given gas concentration. This is attributed to a change in the surface concentration over the duration of the pulse response and a rate corresponding to different surface coverage. In the TAP experiment, only gas phase species can be measured; however, changes in the instantaneous surface concentration, Cs(t), can be calculated from a linear combination of gas phase transformation rates: ð t X

on the surface, are related to the population of surface intermediates. Whereas a surface intermediate is described by a distinct molecular arrangement, the surface storage refers to the collected atoms and molecular components that may exist in multiple arrangements. Structural details can only be inferred based on stoichiometry and the connectivity of the reaction mechanism. With time-dependent information describing changes in the rate, gas, and surface concentration, a three-dimensional kinetic space can be presented that describes how surface regulates different species and controls the transformation rate [9, 61]. This detailed kinetic characterization over a dynamic range of gas and surface concentrations creates a unique “fingerprint” for identifying and distinguishing how different catalyst compositions control chemical reactions.

ð40:56Þ

where νi is the stoichiometric coefficient of the intermediate ith step and the plus/minus signs correspond to the reaction steps where the intermediate is produced/consumed. Both the instantaneous and cumulative storage, or uptake and release

With approximately 20 TAP systems in use, the number of published experimental results has recently seen accelerated growth. This section draws examples from the global community to demonstrate how the tool can be used to measure critical quantities needed in the characterization and development of industrial catalysts. These include (i) coveragedependent sticking coefficients, (ii) active site titration at working temperatures for irreversibly adsorbing probe molecules as well as for mixtures of active sites, (iii) active site titration for reversible adsorption: equilibrium constants and the heat of adsorption, (iv) mechanistic features including adsorption and reaction network distinction, (v) the lifetime of fast surface intermediates and their role in reaction mechanism, and finally (vi) pressure-sensitive reaction kinetics and the distinction between gas phase and surface reactions. As these examples do not fully illustrate the unique capabilities of TAP, the reader is referred to the following kinetic network studies for additional detail: • N2O reduction over ex-framework FeZSM-5 [64]: it was demonstrated that the release of oxygen from the catalyst surface during direct N2O decomposition is the ratedetermining step, due to slow oxygen recombination, which is favored at high temperatures. • C2-C4 alkene reaction [65]: the TAP methodology was employed to isolate a well-defined population of surface species, consistent with surface methoxy species on Bronsted acid sites that are reactive in alkene methylation on a ZSM-22 (TON) zeolite. Their coverage was determined by TAP titration to be ca. 5% of the total amount of Brønsted acid sties, which was also indirectly suggested by FTIR data. This study highlights the potential benefit of low-pressure kinetic measurements for fundamental

40

916

R. Fushimi

investigations of zeolite-mediated reactions, which is a promising research avenue toward reconciliation of experimental data with ab initio calculations and rigorous single-parameter variation studies of structureperformance relationships. • Methanol to olefins over HZSM-5 [66]: the overall reaction scheme was shown to proceed through an initial methanol dehydration to dimethyl ether, which subsequently reacts to form olefins, where ethylene and propylene are the primary olefin products. • Ethylene oxychlorination over ceria [67]: simultaneous HCl, C2H4, and O2 pulse experiments suggest that vinyl chloride formation proceeds through an ethylene dichloride intermediate followed by dehydrochlorination. • Isotopic labeling: the small pulse intensity allows for efficient use of expensive isotopically labeled gases, which can be used to resolve the details of surface or lattice species. For example, the role of N-species during NO reduction using 15NO with 14NO [68, 69]; the role of adsorbed and lattice oxygen species using 18O2 and 16O2 [68, 70].

40.4.1 Coverage-Dependent Sticking Coefficients The sticking coefficient, S0, is the ratio of adsorbing molecules that successfully adsorb to the total number of atoms impinging on the surface during the same period of time. This coefficient is a function of temperature, surface coverage θ, and structural characteristics of the material as well as the

a)

kinetic energy of impinging molecules. In the state-defining TAP experiment, where the change in the surface concentration during one pulse is negligible, the sticking coefficient is directly related to conversion, XA, when adsorption is irreversible. After a number of collisions, ncoll, the fraction that does not adsorb (1  XA) is the product of unsuccessful collisions: ncoll

1  X A ¼ ð 1  S0 Þ

ð40:57Þ

Equation (40.57) is only valid for very small, constant values of S0. As described by Schuurman [19], if a thin zone of catalyst is used, the sticking coefficient must be less than 102 in order to observe an exit flux; if any higher, conversion will be complete. The sticking coefficient is a function of the adsorption rate constant and the rate of impinging molecules: ka S0 ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   8RT σ 2πM

ð40:58Þ

where σ is the surface site density (mol/m2). The adsorption rate constant can be derived from the TAP experiment accordingly using any of the methods described in Sects. 40.3.3 and 40.3.4. In one example, Schuurman presented a multipulse titration of a 1 wt% Pd/SiO2 sample using oxygen pulses at 100  C [19]. The pulse responses are shown in Fig. 40.10a; at the beginning of the experiment, oxygen conversion is complete and no exit flow is detected; in this case the pulse is state-defining as adsorption sites are in excess. As

b) 1.0

1–Xo2

0.8

0.6

0.4

0.2 100 0.4 1

0.0

Pulse #

Time (s)

Fig. 40.10 (a) Pulse response series when oxygen (2
1015 molecules/pulse) is pulsed over 1 wt% Pd/SiO2, (b) oxygen breakthrough curve as a function of pulse number. (Reprinted from Catalysis Today,

0.0 0

20

40 60 Pulse number

80

100

121, Schuurman, Y., Assessment of kinetic modeling procedures of TAP experiments, 187–196, (2007), with permission from Elsevier)

40

Temporal Analysis of Product (TAP)

917

irreversible oxygen uptake proceeds, the number of available surface sites decreases, the conversion decreases, and the pulses become state-altering. Figure 40.10b shows the oxygen breakthrough curve, unconverted oxygen, 1 – XO2, as a function of pulse number. Assuming irreversible adsorption requiring a number of surface sites, α, the change in the sticking coefficient with respect to coverage can be modeled according to: 0

Sð θ o Þ ¼ S ð 1  θ o Þ

α

ð40:59Þ

where θo is the coverage of oxygen (equivalent to the converted oxygen). The model fit shown in Fig. 40.10b shows how the sticking coefficient changes as a function of coverage. An initial sticking coefficient, S0, was estimated to be 0.24 with α ¼ 2.5, and the number of adsorption sites was estimated to be Nsites ¼ 3
105 mol∙gcat from the total accumulation. A similar titration approach was used to distinguish materials with trace impurities based on the sticking coefficient measured at low coverage conditions and with different oxygen exposure levels [71]; see also reference [72]. Compared to the molecular beam-scattering method on single crystals, the diffusive flux of the TAP measurement significantly increases the number of collision events (see Sect. 40.3.1), but importantly, it is amenable to measurement on complex industrial catalysts.

40.4.2 Active Site Titration at Working Temperatures The turnover frequency (TOF), defined as the number of revolutions of the catalyst cycle per unit time, and the turnover number (TON), the number of molecules converted per catalytic site per unit time, are essential measurements for benchmarking catalyst performance in terms of activity and productivity, respectively [73, 74]. These quantities rely on the measurement of rate and the number of active sites, and the challenge is certainly greater with the latter. The number of active sites is generally found using chemisorption studies [74] which are conducted at low temperatures. This is a key drawback since the sites measured at low temperature may be different from those working at reaction temperatures. Steady-state isotopic transient kinetic analysis (SSITKA), which can determine most active reaction intermediates, MARI [75] under steady-state reaction conditions by using a step change from an unlabeled reaction mixture to a labeled one [76], is not plagued by this key drawback as realistic conditions can be used. The number of active sites determined by SSITKA, however, is related only to the number of active sites occupied by intermediates with isotope and, therefore, is prone to underestimate the involved number of

active sites. One advantage of the TAP method is that it can be used to count active sites at higher temperatures using precise titration due to the nanomole pulse intensity. Different strategies exist depending on whether the molecule adsorbs irreversibly or reversibly. In the case of the irreversibly adsorbing species, a long series of titrating pulses can also be used to quantitatively distinguish two working sites from a mixture. First, irreversible and reversible adsorption are readily distinguished by examination of the time-dependence of the exit flux of the adsorbate molecule in a single pulse experiment. For example, Fig. 40.11a compares the exit flux predicted for the case of irreversible adsorption with dimensionless rate constants of ka ¼ 3 and 10 to that of the standard diffusion curve; simulation conditions are described in more detail by Gleaves et al. [20]. Where irreversible adsorption takes place, the normalized exit flux will always fit within the standard diffusion curve. On the other hand, where reversible adsorption takes place, the exit flux of the probe molecule will be observed to cross over the standard diffusion curve. As shown in Fig. 40.11b, pulse responses modeled with ka ¼ 20, kd ¼ 20, ka ¼ 20, and kd ¼ 5 are observed to intersect the standard diffusion curve.

Titrating Sites with Irreversibly Adsorbing Species Using a long series of state-defining pulses, the total capacity of a catalyst can be measured either from the perspective of occupying empty active sites or consuming preadsorbed active species. Figure 40.10 is an example of titrating empty sites. Similarly, Widman et al. [75] determined the total oxygen storage capacity (OSC) of a supported gold catalyst using CO titration at 120  C. Alternatively, on a nanoporous (Ag)Au catalyst, preadsorbed oxygen was titrated using methanol pulses by measuring the formaldehyde production at 150  C [72]. There, the amount of adsorbed atomic oxygen was also quantified by measuring the total uptake during oxygen pulses, and a 1:1 ratio of active atomic oxygen to formaldehyde production was observed. Since the oxygen conversion was low, the pulse exiting the reactor was large and small changes can be difficult to detect on top of a large signal. Observation of the formaldehyde product provided a more sensitive measurement since the product signal is compared to the baseline. These examples assume the number of active sites covered by a gas molecule is either 1 or 2. The summation of conversion data over a long series of pulses is a measure of capacity. An advanced analysis method of the same type of data was demonstrated by Constales et al. that distinguishes adsorption capacity from the number of active sites as it exists in the rate relationship [76]. This method takes into account the interdependence of the conversion, reaction rate, and number of active sites as they change throughout the titration and includes information about the stoichiometry of adsorption as

40

918

R. Fushimi

b) 2.4

tp = 0.17

1.9

Dimensionless flow (FAebL2/NpADeA)

Dimensionless flow (FAebL2/FpADeA)

a) 2.4

FA,p = 1.85

1.4

0.9

A B

0.4 C

–0.1 0

0.5

1 1.5 2 Dimensionless time (tDeA/ebL2)

2.5

3

1.9 A 1.4

0.9 B 0.4

C

–0.1 0

0.5

1

1.5 2 2.5 3 3.5 4 Dimensionless time (tDeA/ebL2)

4.5

5

Fig. 40.11 Simulated dimensionless exit flux for (I) irreversible adsorption/reaction with (a) ka ¼ 0 (standard diffusion), (b) ka ¼ 3, and (c) ka ¼ 10; and (II) reversible adsorption with (a) ka ¼ 0 (standard diffusion), (b) ka ¼ 20, kd ¼ 20, (c) ka ¼ 20, and kd ¼ 5: (Reprinted

from Applied Catalysis A: General, 160, Gleaves, J.T., Yablonskii, G., Phanawadee, P., Schuurman, Y., TAP-2: an interrogative kinetics approach, (1997), with permission from Elsevier)

well as the kinetic order of the adsorption reaction (molecular vs. dissociative). Here, the conversion, X, of a gas causes the surface coverage to increase, qY. Pulse-by-pulse changes in the coverage bring about a change in the reaction rate,

concentration of adsorbed species, Cads (mol/m2), in a single pulse from the integral of the rate:

R ¼ ka Cg ðN a  qY Þ

α

ð40:60Þ

where Cg is the gas concentration, q is stoichiometric coefficient which is number of active sites occupied by one gas molecule, Y is the number of adsorbed molecules, Na is the total number of active sites, α is the kinetic order, and ka is the adsorption constant. With changes in the reaction rate, the apparent constant is changed and the Damköhler number, Da, changes as well; τ is the reactive residence time in the reactor. The change in the Damköhler number then leads to a change of conversion according to; Da ¼ kapp τres ¼ X=ð1 XÞ

ð40:61Þ

This interdependence of conversion, surface concentration, and rate was analyzed using both differential and integral methods. For a long series of small oxygen pulses over a Pt/SiO2 catalyst, the analysis of pulse-by-pulse conversion data using a linear rectification method indicated secondorder oxygen chemisorption with 1:1 Pt to O stoichiometry and a minor increase in the number of active sites over the temperature range of 125–175  C [76]. As opposed to measurements based on conversion, the time-dependence of the rate (calculated using the Y-Procedure as described in section “Time-Dependent Analysis of Rate and Concentration”) can be used to determine the total

Cads ¼

qV Scat

ðt 0

RðtÞdt

ð40:62Þ

Here, V and Scat are the catalyst volume and surface area (m3) and (m2) and Rg is the rate of gas consumption (mol/(m3s)). For example, from a series of pulses, the total accumulation of oxygen (6
107 mol/gcat) was measured on polycrystalline platinum using oxygen pulses at 150  C. Using a multipulse series, measurement based on conversion or the rate are both integral quantities, but from the timedependence of the rate one can also estimate the number of active sites from a single pulse [13]. For irreversible adsorption, the conventional rate expression is: R ¼ kapp C kapp ¼ k0 θfree sites ¼ k0 ðN a total uptakeÞ

ð40:63Þ ð40:64Þ

where k0 is the rate constant at low coverage. The apparent rate constant, kapp, will be observed to decrease over the course of the experiment. By rearranging Eqs. (40.63) and (40.64):   ðt R ¼ k0 N RðtÞdt C 0 then taking the derivative:

ð40:65Þ

40

Temporal Analysis of Product (TAP)

919

 0 R R0 RC0 ¼  2 ¼ k0 R C C C

ð40:66Þ

From the regression of Eq. (40.66), one can determine k0 and N. At Rmax , k0 ¼

C0 C2

ð40:67Þ

from different oxidants (O2, CO2, and H2O) [77]; the data in Fig. 40.12 is recorded at three different temperatures. Early in the titration, the rate is controlled by a highly active site which is quickly extinguished. The material is still active following the initial rapid decrease, but the rate constant then assumes a new linear decline. The overall activity can be modeled as a weighted average of the rate constant for two distinct sites:

and at t ¼ 0,

kapp ¼ k1 ð1 θ1 Þ þ k2 ð1 θ2 Þ R ¼ k0 N C

ð40:68Þ

Therefore, if k0 is estimated from the initial slope of R over C′, then the number of active sites, N, can be determined. This method was demonstrated on a polycrystalline iron catalyst and a CoFe bimetallic catalyst at 550  C using ammonia pulses [13].

Screening Multiple Active Site Mixtures More than the measure of capacity, another valuable aspect of the multipulse titration is the ability to observe the evolution of kinetic properties as the catalyst is changed over the longer course of the state-altering experiment, e.g., from oxidized to reduced, as in the case of butene titration of active oxygen on vanadyl pyrophosphate [57]. Generally, the rate constant is proportional to the number of active sites and should decrease linearly as sites are consumed. In the case of CO titration of 2% Pt/CeO2, the rate constant for CO conversion presented two distinct linear regimes as CO reacted with oxygen added

θ1 ¼ N CO,1 =N 1

ð40:70Þ

θ2 ¼ N CO,2 =N 2

ð40:71Þ

where θ1 and θ2 are the coverage of CO over more active sites (N1) and less active sites (N2), respectively. k1 and k2 are the rate constants associated activity on the different sites. The fit of Eq. (40.69) is shown in Fig. 40.12 and can be used to distinguish two active sites, with very different activity, working independently. From the two distinct linear regions characterized at different temperatures, the values of fitting parameters were used to quantify 1.33
107 mol/mgcat (~9% of the total) and operate with an activation energy of 23 kJ/mol and 1.28
106 mol/mgcat with an activation energy of 44 kJ/mol. By virtue of the Arrhenius dependence, even a small fraction of more highly active sites (here associated with the platinum) will dominate the rate at low temperatures. At high temperatures, the lower activity site will be dominant, which can result in a switch to a different reaction

100 90 80 CO apparent rate constant (s–1)

Fig. 40.12 CO apparent rate constant as a function of CO consumed in the pulse titration of 2% Pt/CeO2 at different temperatures; green triangle 250  C, blue diamond 300  C, red circle 350  C, and black line model fit from Eq. (40.69). (Reprinted from Journal of Catalysis, 253, Shekhtman, S., Goguet, A., Burch, R., Hardacre, C., Maguire, N., CO multipulse TAP studies of 2% Pt/CeO2 catalysts: influence of catalyst pretreatment and temperature on the number of active sites observed, 303–311, (2008), with permission from Elsevier)

ð40:69Þ

70 60 50 40 30 20 10 0 0.0E+00

2.5E–07

5.0E–07

7.5E–07

1.0E–06

1.3E–06

CO consumed (mol. mgcat–1)

1.5E–06

1.8E–06

2.0E–06

40

920

R. Fushimi

mechanism as a function of temperature, a feature which is particularly relevant to the water-gas shift reaction. A similar result was found in a study on Au/CuMnOx catalysts [78].

1.0 Ar

COðgÞ þ Z

F eL2

CO, 188 C CO, 127 C

0.6

CO, 99 C 0.4

ka

!

Np D

0.8

Heat of Adsorption, Counting Sites Where Adsorption Is Reversible For reversibly adsorbing species,

COZ

ð40:72Þ

CO, 72 C CO, 50 C

0.2

kd

0.0

the adsorption equilibrium Keq,

0

K eq ¼ ka=kd

4

ð40:73Þ

can be determined by fitting model parameters to pulse response data (see Sect. 40.3.3). In this case, for the gas phase: εcat

2

@CCO @ 2 CCO ¼ Dcat @t @x2  ð1 εcat Þσaυ ðka CCO θZ  kd θCO Þ

ð40:74Þ

Such an approach has been used in a number of examples: Dewaele and Froment characterized the sorption of CO and CO2 on dehydroxylated γ-Al2O3 [79]; Kapteijn, Moulijn, and coworkers studied both sorption and diffusion in activated carbon [80]; Schuurman et al. characterized sorption and diffusion of hydrocarbons in FCC catalysts [81]; Breitkopf examined C4-hydrocarbons in sulfated zirconias [82]; and Farusseng et al. characterized noble gases, hydrocarbons, and oxygenates in metal-organic frameworks (MOFs) [83]. Compared to other measurement methods, the absence of external mass-transfer limitations, negligible temperature change from 10 nmol adsorption, and the ability to screen small samples ( 80  C the dominant mechanism involved removal of surface lattice oxygen as opposed to adsorbed oxygen, but at lower temperatures the Mars van Krevelen pathway became inhibited due to activated removal of CO [94]. Oxygen isotope pulsing is a straightforward method to quickly distinguish the role of lattice oxygen in the formation of different oxygenate products. When the reducing species is pulsed simultaneously, or in pump/probe time delay with 18 O2, if lattice oxygen is the principle oxidizing species then only 16O will be detected in products. Incrementally, as lattice oxygen is depleted and resupplied from the 18O2 pulse, the formation of 18O containing products will gradually emerge. If the adsorption of oxygen is reversible, then 16 18 O O will be detected in the pulse response. Morgan et al. used such isotopic experiments to distinguish the Langmuir-Hinshelwood and Mars van Krevelen mechanisms for CO oxidation on CuMnOx with and without the addition of Au [78]. When CO was pulsed in varying delay time following 18O2, the yield of different CO2 isotopes was examined as a function of the delay time, Fig. 40.17. It is clear that the addition of Au increases the activity which is primarily observed in the 13C16O2 (AMU 45) product containing only unlabeled oxygen atoms; thus, gold doping promotes the Mars van Krevelen mechanism. However, on

923

40

0.4

0.3 CO2 yield

40

0.2

0.1

0.0

0

1

2 Time delay (s)

3

4

Fig. 40.17 CO2 yield when CO is pulsed over CuMnOx and Au/CuMnOx catalysts at varied time delays after an O2 pulse: (◊) CO2 (40.45) undoped, (+) CO2 (40.45) Au doped, (□) CO2 (40.47) undoped, ( ) CO2 (40.47) Au doped, (Δ) CO2 (40.49) undoped, and (○) CO2 (40.49) Au doped. (Reprinted from Journal of Catalysis, 276, Morgan, K., Cole, K.J., Goguet, A., Hardacre, C., Hutchings, G.J., Maguire, N., Shekhtman, S.O., Taylor, S.H.: TAP studies of CO oxidation over CuMnOX and Au/CuMnOX catalysts. 38–48 (2010), with permission from Elsevier)

both materials, there is a significant drop in the 13C18O16O (AMU 47) production from simultaneous pulsing to the 0.5 s time delay experiment. Therefore, adsorbed oxygen from the 18 O2 pulse also reacts according to the LangmuirHinshelwood mechanism, but the addition of gold has little effect to enhance this mechanism. While both mechanisms are at play, isotopic pump/probe experiments can be used to distinguish the contribution of each.

General Reaction Model Analysis Compared to steady-state pressure dependencies, transient experiments provide much greater detail in a single pulse response for discriminating the elementary reaction steps of a microkinetic mechanism. In either case, the most convenient microkinetic models will have at least one gas phase species either generated or consumed in each elementary reaction step. In this manner, the entire mechanism is observable from gas phase measurement and is more readily validated. In the case where individual steps consist entirely of surface reactions, they might be lumped together with steps observable from the gas phase or resolved explicitly when fitting data numerically. Numerical Fitting With data from only three TAP response curves, when N2O was pulsed over Fe-silicalite at 798 K, Kondratenko and Pérez-Ramírez demonstrated how six different plausible reaction schemes for direct N2O decomposition could be reliably discriminated [47]. Figure 40.18 shows

924

R. Fushimi

N2O

Flux (1015 molecules s–1)

a) 4.0

Experiment Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

3.0

2.0

1.0

0.0 0.0

0.1

0.2

0.3

0.4

0.5

t (s) N2

O2

c) 4.0 Experiment Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

6.0

4.0

Flux (1014 molecules s–1)

Flux (1015 molecules s–1)

b) 8.0

2.0

Experiment Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

3.0

2.0

1.0

0.0

0.0 0.0

0.2

0.4

0.6

0.8

1.0

t (s)

0

1

2

3

t (s)

Fig. 40.18 Experimental data and model fit for (a) N2O, (b) N2, and (c) O2 pulse response when N2O is pulsed over Fe-silicalite at 798 K. Models described in Table 40.1. (Reprinted from Catalysis Today, 121, Kondratenko, E.V., Pérez-Ramírez, J.: Micro-kinetic analysis of

direct N2O decomposition over steam-activated Fe-silicalite from transient experiments in the TAP reactor. 197–203 (2007), with permission from Elsevier)

their experimental data and model fit for the elementary reaction schemes shown in Table 40.1. The first reaction step is the same for each, and the models increase in complexity from 2 to 5 steps. Models 3 and 5 both include one surface reaction step with no gas phase reactants or products. Based on DFT studies in the literature, Model 5 distinguishes two types of biatomic oxygen species: The O-Z-O structure has no chemical bond between the oxygen atoms and must be transformed to ZO2 before oxygen desorption can occur. It is readily apparent from Fig. 40.18 that Model 5 yielded the best optimization for all three experimental pulse responses; the rate constants and activation energies are included in

Table 40.1; and note that the rate constants for the adsorption steps are the product of the intrinsic rate constant and the total number of active sites. The validity of the model was confirmed over a broader temperature range under both transient and steady-state conditions. The numerical fitting approach for transient pulse response data has been widely used in the TAP community to discriminate and validate different microkinetic models [21, 42, 47, 50, 62, 95–103]. As discussed previously in Sect. 40.3.3, regressed parameters should be carefully assessed statistically, and sensitivity analysis should be

40

Temporal Analysis of Product (TAP)

925

Table 40.1 Reaction schemes for direct N2O decomposition used for model fitting of N2O, N2 and O2 pulse response data by Kondratenko and Pérez-Ramírez [95] 1 2 3

4

5

6

N2O þ Z ! N2 þ ZO ZO þ ZO ! O2 þ 2Z N2O þ Z ! N2 þ ZO N2O þ ZO ! N2 þ O2 þ Z N2O þ Z ! N2 þ ZO ZO þ ZO ! ZO2 þ Z ZO2 ! O2 þ Z N2O þ Z ! N2 þ ZO N2O þ ZO ! N2 þ ZO2 ZO2 ! O2 þ Z N2O þ Z ! N2 þ ZO N2O þ ZO ! O-Z-O þ N2 O-Z-O ! ZO2 ZO2 ! O2 þ Z N2O þ Z ! N2 þ ZO N2O þ ZO ! ZO2 þ N2 N2O þ ZO2 ! ZO3 þ N2 ZO2 ! O2 þ Z ZO3 ! O2 þ ZO

slow

fast

Z ! AZ ! BZ þ Cðor Z þ CÞ: A counterclockwise trend in the product petal trajectory indicates that product generation is indirect and transformation through some surface species is involved. In this case, the product is formed from some intermediate that is not equilibrated with the gas phase reactant or it is offset by fast consumption. The transformation of the intermediate to fast

k798 K (1/s) 1170  68 167  9.5 26  5.5 2.7  0.1

Ea (kJ/mol) 120  8 81  8 298  44 129  8

Optimized kinetic parameters are listed for Model 5 which demonstrated the best fit to experimental data

conducted at a number of different experimental conditions. The process becomes less reliable however, as the complexity of the model grows and in cases where there are multiple surface reaction steps in sequence that cannot be observed from gas phase measurements. Kinetic Petals As a complement or alternative to numerical model fitting, examination of the rate and gas concentration time-dependence can be used to derive microkinetic mechanistic features. As introduced in section “Time-Dependent Analysis of Rate and Concentration,” the RC kinetic petal of reactants can be used to characterize the adsorption process. Interpretation of the product RC petal shape and directionality can be used to understand the complex interplay of reactant adsorption, surface transformation, and product desorption [13]. Three cases may be considered for clockwise product petals: (a) Direct formation. The product is generated directly through interaction with the catalyst; A þ Z ! AZ þ C, where A and C are reactant and product, respectively. Z and AZ are the active center and surface intermediate, respectively. (b) Quasi-equilibrium. A two-step reaction with fast, reversible conversion of the reactant to a surface intermediate; fast

than conversion of the intermediate to product; A þ

A þ Z $ AZ ! BZ þ C ðor Z þ CÞ, where AZ and BZ are corresponding intermediates. (c) Quasi-steady-state. A two-step reaction where conversion of the reactant to the surface intermediate is slower

slow

product must therefore be the slow step: A þ Z ! AZ ! BZ þ Cðor C þ Z Þ: Thus, the counterclockwise petal shape indicates the product formation depends on the concentration of surface intermediates. These features of reactant/ product petal shape and direction and rate/concentration maximum can be used to derive physicochemical insight into how the surface regulates different species and how reaction processes are orchestrated. For example, Wang et al. used the case of ammonia decomposition to illustrate key features of RC kinetic petals [13]. In Fig. 40.19, the hydrogen and nitrogen product rateconcentration dependencies are compared when ammonia is pulsed over an iron, CoFe bimetallic, and cobalt catalyst at 550  C. For hydrogen, the counterclockwise petal shape observed for iron and CoFe, Fig. 40.19a, b, indicates hydrogen formation on these materials is limited by slow surface reaction steps. For cobalt, Fig. 40.19c, fast hydrogen production, coherent with ammonia conversion, was observed. Moreover, the lower branch of the petal following the concentration maximum returns linearly to zero which indicates an additional slow hydrogen formation step that is dependent on the surface concentration of remaining H species. In these experiments, nitrogen formation was observed but the signal was too low for analysis using the Y-Procedure algorithm. For both CoFe and cobalt, the nitrogen petal shapes were counterclockwise, indicating an anticipated dependence on surface transformations, i.e., dehydrogenation steps. Interestingly, the nitrogen petal shape for CoFe indicated a negative reaction order which represents a self-inhibition process where N species block the adsorption of ammonia required to accelerate N2 formation. A convenient feature of the rate-concentration presentation is direct extraction of the apparent rate constant from the slope; Wang et al. reported the kinetic data in Table 40.2 [13]. In particular, for state-altering petal shapes such as that observed for hydrogen formation on cobalt, Fig. 40.19e, different constants can be distinguished in regions of low and high surface coverage. The rate-concentration petal shapes are used to derive different microkinetic features directly from the experimental data. This information can be used independently to discriminate the behavior of different materials. It can serve as the

40

926

R. Fushimi

a)

b)

Fe

1.5×1011

RH2/(molecules/s/cm2)

1.0×1010

1.0×1011

5.0×109

1.0×1011

5.0×1010

5.0×1010

0.0 1×1015 2×1015 CNH3/(molecules/cm3)

3×1015

d)

0

1×1014 CNH3/(molecules/cm3)

1.0×1010

RN2/(molecules/s/cm2)

0

e)

5.0×109

0.0

0

1×1014 CNH3/(molecules/cm3)

0.0

2×1014

2×1014

0

1×1014 CNH3/(molecules/cm3)

2×1014

0

1×1014 CNH3/(molecules/cm3)

2×1014

1.0×1010

RN2/(molecules/s/cm2)

0.0

Co 1.5×1011

RH2/(molecules/s/cm2)

RH2/(molecules/s/cm2)

1.5×1010

c)

CoFe

5.0×109

0.0

Fig. 40.19 Rate-Concentration (RC) kinetic petals for hydrogen on (a) iron, (b) CoFe, and (c) cobalt; nitrogen for (d) CoFe, and (e) cobalt when ammonia is pulsed at 550  C. (Reprinted with permission from Wang, Y., Kunz, M.R., Constales, D., Yablonsky, G., Fushimi, R.: Rate/

concentration kinetic petals: a transient method to examine the interplay of surface reaction processes. J. Phys. Chem. A. 123, 8717–8725 (2019). Copyright (2019) American Chemical Society)

Table 40.2 Apparent rate constants for ammonia transformation, hydrogen and nitrogen production as reported by Wang et al. [13]

Table 40.3 Elementary steps comprising different mechanism of CO oxidation

kNH3 ð1=sÞ

Iron 4.8
107

CoFe 4.2
109

Cobalt 3.5
1010

kH2 ð1=sÞ

1.4
106

5.7
108

kN2 ð1=sÞ

N/A

N/A

1.7
1010 (low coverage) 2.3
108 (high coverage) 1.5
107

starting point to propose different microkinetic models and conveniently provides an initial guess of kinetic parameters for fitting algorithms. Kinetic Coherency An alternative approach for more detailed model validation is based on nonsteady-state coherency of kinetic parameters and was proposed by Yablonsky and coworkers [104]. This approach is enabled by the calculation of the rate, gas, and surface concentration timedependence from exit flux data via the Y-Procedure method (section “Time-Dependent Analysis of Rate and Concentration”). From these quantities, two types of temporal coherency can provide a fingerprint for distinguishing different reaction networks: rate-rate coherency and rate-concentration coherency. Based on the way each network is constructed, different coherency features can be identified. Starting with a

1 2 3 4 5

Elementary step O2 þ 2Z ! 2ZO CO þ Z $ ZCO CO þ ZO ! Z þ CO2 ZCO þ ZO ! 2Z þ CO2 ZO þ Zb ! Z þ ZbO

comprehensive reaction network of all possible reaction steps, a sequential examination of experimental rateconcentration and rate-rate transients is used to reject mechanistic steps where inconsistencies are found. This process was described based on different combinations of the elementary steps, Table 40.3, that comprise variations of different CO oxidation mechanisms, Table 40.4. The typical Langmuir-Hinshelwood (LH) mechanism was distinguished from the Eley-Rideal (ER) mechanism in addition to combinations with different buffer steps (BS) and surface processes involving oxygen (referred to as Additional Oxygen Process, AOP). For example, based on numerical simulation in a thin zone TAP configuration where kinetic parameters give similar conversion, the CO and CO2 time-dependent rate data can

40

Temporal Analysis of Product (TAP)

927

Table 40.4 Combinations of elementary reaction steps involved in possible mechanisms of CO oxidation Abbreviation ER LH ER þ LH ER þ BS ER þ AOP ER þ BS þ AOP LH þ AOP ER þ LH þ AOP

Sequence of steps 1, 3 1,2,4 1,2,3,4 1,2,3 1,3,5 1,2,3,4 1,2,4,5 1,2,3,4,5

ER Eley-Rideal, LH Langmuir-Hinshelwood, BS buffer step, AOP additional oxygen process. As reported in [104]

be distinguished for the ER impact and LH adsorption mechanisms. In the ER mechanism, CO is consumed at a rate which equals to CO2 production and hence demonstrates rate-rate coherency. However, for the LH mechanism, there is a time-delay between CO consumption and CO2 release that is dictated by the surface lifetime of the ZCO intermediate. Hence these two mechanisms could be distinguished solely on CO and CO2 rate-rate coherency. In a similar manner, Redekop et al. parse through different scenarios for rate-concentration coherency and constructed a decision tree, Fig. 40.20, for distinguishing the six different reaction networks listed in Table 40.4 [104]. From experimental exit flux data, the time-dependence of the rate, gas, and surface concentration for reactants and products are first calculated. The first decision comes with examination of the CO and CO2 rate-rate coherency. Following this, different rate-concentration quantities are examined which are simply statements of the mass action law for each mechanism: þ Y nr Cr

Ri ¼ k i

 Y np Cp

 ki

ð40:79Þ

where the subscript r refers to reactants and p refers to products. For example, for the pure LH mechanism, Fig. 40.21, the rate of CO2 production should be linear with respect to the product of the instantaneous oxygen storage, CΣ O, and the instantaneous carbon storage, CΣ C: þ

þ

RCO2 ðtÞ ¼ k4 CZO ðtÞCZCO ðtÞ ¼ k4 CΣO ðtÞCΣC ðtÞ

ð40:80Þ

Recall that the instantaneous storage of species is calculated using integral quantities of linear combinations of gas phase rate (uptake and release) according to stoichiometry, Eq. (40.56). Other discrimination steps are described in greater detail in the original work [104]. It should be noted that in certain cases, especially where elementary steps are confined to the surface without gas phase reactants or products, it may not be

possible to discriminate networks based on experimental data. Likewise, there are limitations associated with the kinetic parameter space where steps may be too fast or too slow for observation on the timescale of the TAP experiment. Lastly, high-frequency experimental noise amplified by the Y-Procedure analytic method is another limitation of the coherency-based hypothesis discrimination.

40.4.4 Role of Dynamic Surface Species in Reaction Mechanism, Lifetime of Fast Surface Intermediates As described in sections “Titrating Sites with Irreversibly Adsorbing Species” and “Screening Multiple Active Site Mixtures,” the multipulse experimental format can be used to titrate empty active sites as well as long-lived active surface species added through different pretreatment methods; see Fig. 40.12. Many surface species, however, may have a short surface lifetime due to processes such as reversible adsorption, conversion to more/less active forms, diffusion into the bulk, etc. Short-lived surface species, and their role in the reaction mechanism, can be determined using pump/probe dynamic experiments. As described in Sect. 40.2.2 and illustrated in Fig. 40.3, the introduction of two reactants (typically a hydrocarbon and oxidant) are separated by a short time-delay (generally 0.01–4 seconds). The first pulse (the pump or priming pulse) creates a dynamic surface population of active species which are then sampled by the second pulse (the probe pulse). Changes in conversion and product distribution are then investigated as a function of pump/probe delay timing. This method was demonstrated early in the development of the TAP technique by Gleaves et al. with O2/C2D4 pump/ probe experiments in the investigation of perdeuterated ethylene epoxidation on silver powder [105]. This method was recently applied to understand ammonia decomposition/synthesis by Wang et al. [106]. The isotopic pump/probe experiments between NH3 and D2 were used to distinguishing the lifetime of surface species of H* and NHx* during the ammonia decomposition. After a NH3 pulse, a subsequent D2 pulse was introduced using a time delay of 0.2, 0.4 or 1.0 s (Fig. 40.22a). The relative intensities for NH2D, NHD2, and HD isotopic products are presented in Fig. 40.22b–d. The primary ammonia pulse sets forth the decomposition reaction and formation of NH2*, NH*, H*, and N* species on the surface with which the D2 probe pulse can then interact. Changing the time delay for introduction of D2 changes the isotopic product distribution according to the “snapshot” of NHx* surface species that have evolved from the initial NH3 pulse. As shown in Fig. 40.22b–d, at 550  C, the intensity of all products decreases with a delay time on all three materials but with a different trend, which can be explained by the

40

928

R. Fushimi

{RCO2, RCO, CCO, CSO, RSC}

1 Yes

No

RCO2 = RCO

2 Yes

RCO2

3 No

=k

RCO2

No

CS OCCO

Yes =k

CS OCS C ER+AOP 4

LH

ER RCO2

No

=k

Yes

CS OCCO

No

Slope > 0

ER+LH+AOP LH+AOP

RCO2 vs CS OCCO vs CS OCS C

5

RCO2 vs CS OCCO vs CS OCS C

Slope < 0

Yes ER+BS

ER+LH

ER+BS+AOP

Fig. 40.20 Decision tree for discriminating different mechanisms of CO oxidation based on rate-rate and rate-concentration coherency data. (Reprinted from Chemical Engineering Science, 110, Redekop, E.A.,

Yablonsky, G.S., Constales, D., Ramachandran, P.A., Gleaves, J.T., Marin, G.B.: Elucidating complex catalytic mechanisms based on transient pulse-response kinetic data. 20–30 (2014), with permission from Elsevier)

different surface lifetime of different species on different catalysts. This method provides a new approach for understanding how the network of surface reaction steps unfold on complex industrial catalysts as well as for deriving quantitative data on lifetimes of surface species and kinetic constants of individual surface reactions. In another example, based on the selective catalytic reduction (SCR) of N2O, two different surface oxygen species are believed to originate from N2O decomposition on Fe-silicalite catalysts: one highly reactive (nascent) and one nonreactive (thermally accommodated) [107]. Kondratenko and Pérez-Ramírez conducted pump/probe experiments with N2O and different reducing agents CO, C3H8, 13CH4, to demonstrate the importance of the lifetime of oxygen species in alkane activation and changes in the reaction mechanism with different reductants [108]. The pump/probe delay spacing was increased between N2O and methane or propane, and

the activity for hydrocarbon oxidation decreased. Since gas phase oxygen was not observed, it can be concluded that the change in product formation does not result from a decrease in the concentration on the surface but rather from a transformation of highly reactive (short-lived) oxygen species to less reactive oxygen. On the contrary, when CO was used as a reductant, there was no significant influence of the pump/ probe delay timing. Thus, CO and hydrocarbons clearly have different reaction pathways.

40.4.5 The Pressure Gap The TAP experiment is conducted under high vacuum conditions (typically 108 torr), and as such, it is similar to surface science experiments with the distinction that complex industrial catalysts may be probed, not just well-defined

40

Temporal Analysis of Product (TAP)

929

0.8

RCO2 (mmol/kgcat/s)

0.6

0.4 Time 0.2

0.0 0.0

0.2

0.4

0.6

0.8

1.0

C S C C S O, (mmol S C·mmol S O/kgcat/m3)

Fig. 40.21 Rate-concentration temporal coherency prescribed by the Langmuir-Hinshelwood reaction mechanism for CO oxidation. Simulated noise-free data is shown as the solid line. Arrows indicate the direction of time during the transient experiment. (Reprinted from Chemical Engineering Science, 110, Redekop, E.A., Yablonsky, G.S., Constales, D., Ramachandran, P.A., Gleaves, J.T., Marin, G.B.: Elucidating complex catalytic mechanisms based on transient pulse-response kinetic data. 20–30 (2014), with permission from Elsevier)

model catalysts. However, at the peak of the pulse, the pressure will reach 103 – 102 torr in the catalyst thin zone; this pressure is similar to that used in many near ambient pressure operando spectroscopic measurements [109]. These peak pressures, however, may be insufficient to induce certain catalytically active phases or surface species which can form only at elevated pressure [110]. The TAP low-pressure condition and dynamic response simplify mechanistic studies, and backward reactions from evolving products can be neglected; however they can certainly be probed directly by pulsing product molecules. A key distinction between high- and low-pressure regimes is surface coverage. However, by implementing different pulsing formats (such as with pump/probe dynamics) surface coverage can be manipulated and the implication for mechanism and kinetics can be studied. The tool has often been used to distinguish properties of different catalyst formulations, but reaction mechanism and intrinsic kinetics can also be used to inform reactor models. A number of experimental examples have corroborated intrinsic kinetics assessed under TAP conditions with global kinetics collected from conventional steady-state flow experiments. CO oxidation on precious metals is a common example where pressure does not show an influence on TAP or

steady-state kinetics, e.g., supported ruthenium [111, 112], platinum [25, 56], and gold [71]. More than corroborating coverage-dependent regimes and the apparent rate constant of the slowest reaction step, the low-pressure-transient experiments can characterize many more steps in the reaction network. This information yields much greater information as to why one obtains certain global behaviors at higher pressure. For example, Reece et al. used TAP pulse response characteristics of methanol on nanoporous Ag0.03Au0.97 to bridge the kinetic behavior observed on single crystal Au(111) under UHV conditions to flow reactor experiments at atmospheric pressure [103]. While selectivity to methyl formate was very different under TAP and atmospheric flow, the microkinetic model parameters derived from fitting of the TAP pulse response data were able to predict the observed performance across nine orders of magnitude in pressure, five orders of oxygen surface coverage as well as a 200 K range of reaction temperature. Their microkinetic model indicates that the key distinction in performance is due to changes in surface methoxy coverage as a function of pressure and temperature. While microkinetic models have traditionally been restricted to well-defined systems, either single crystal surface science measurements or DFT (density functional theory) calculations, the TAP measurement provides microkinetic model measurements for real materials, with all unique surface complexity intact. These previous examples demonstrate catalysts/reaction systems where kinetic characterizations found agreement across the pressure gap. Still, some reaction pathways remain pressure sensitive and the role of gas-phase kinetic processes cannot be directly addressed when TAP is used in the Knudsen flow regime. Furthermore, some catalytically active phases or surface species might only be formed under pressurized conditions, e.g., the Zn species in the standard methanol catalyst Cu/ZnO/Al2O3 are not metallic but have a positively charged nature under industrial high-pressure conditions [110]. The Knudsen flow regime is a defining characteristic of a TAP experiment, and it essentially removes pressure-dependent effects by eliminating gas phase collisions, allowing the experiment to exclusively probe firstorder gas/solid adsorption kinetics. When transport is welldefined, it can be easily separated from kinetics. The TAP apparatus can, however, accommodate experiments outside the Knudsen flow regime, at the onset of viscid flow, by injecting larger pulses. In these experiments, caution should be used in order to avoid detector saturation (insufficient ionizing electrons) that can lead to split-peak artifacts that might be erroneously interpreted as intriguing transport/ kinetic phenomena. The unfortunate fact is that while higher pressure pulse response experiments can be conducted in the TAP apparatus, even pulsing into a gas flow, the complexity of density-

40

R. Fushimi

a)

Pulse intensity (a.u.)

930

NH3/Ar D2/Kr

NH2D detected during D2 pulse NHD2 detected during D2 pulse

Dt

HD detected during D2 pulse t (s)

b) 0.05

c)

d)

Fe

0.8

CoFe

Co

Intensity (a.u.)

0.04

0.6

0.03 0.4 0.02 0.2

0.01 0.00

0.2

0.4

0.6 Dt (s)

0.8

1.0

0.2

0.4

0.6 Dt (s)

0.8

1.0

0.2

0.4

0.6 Dt (s)

0.8

1.0

0.0

Fig. 40.22 (a) Definition of pump-probe experiments with two injection valves at 550  C, pump molecules: NH3, probe molecules: D2. The pulse size normalized intensity of NH2D, NHD2, and HD intensity versus the pump/probe delay time (Δt) over; (b) Fe; (c) CoFe; and (d) Co [106]. (Reprinted from Wang, Y., Qian, J., Fang, Z., Kunz, M.R.,

Yablonsky, G., Fortunelli, A., Goddard III, W.A., Fushimi, R.R.: Understanding reaction networks through controlled approach to equilibrium experiments using transient methods. J. Am. Chem. Soc. 143, 10998– 11006 (2021). Copyright (2021) American Chemical Society)

dependent viscous flow of these experiments renders the separation of transport and kinetic a significant challenge, although a number of modeling efforts have been undertaken [21, 22, 24, 50, 113, 114]. Without a well-defined transport model, the kinetic information is not much more sophisticated than that from a plug flow reactor (conversion and selectivity). Nonetheless, excellent qualitative observations regarding reaction mechanism and intermediate species have been made with non-Knudsen pulsing [14, 115]. Moreover, the experimental apparatus is convenient for bridging the pressure regimes across 9 orders of magnitude: conducting Knudsen pulse response, increasing intensity into the non-Knudsen regime, then all the way to conventional plug flow reaction (see description of slide valve operation in Sect. 40.2.1) on the same catalyst without disturbing the packing. The study of ammonia oxidation, a fast, highly exothermic reaction, is an illustrative example where the kinetic steps of N2O formation had not been resolved. N2O is a reaction product under industrial operating conditions; however, its formation is not observed in single crystal surface science experiments (a significant pressure and materials gap). Indeed, in Knudsen TAP pulse response experiments of NH3 and O2, Pérez-Ramírez and Kondratenko did not observe N2O formation [14]. However, with higher pressure pulses, just beyond the Knudsen regime N2O formation was observed. Moreover, by using an isotopic experiment with sequential pulsing of 15NH3 and 14NO, N2O was only

observed on the nitric oxide pulse with 15N14NO being the major isotopic product. From this result, it was clear that nitrous oxide is not a primary product of ammonia oxidation, but rather, it is formed as a result of a secondary reaction between NH3 and NO. A similar need for beyond-Knudsen experiments was encountered in studies of catalysts for the oxidative coupling of methane (OCM). In this reaction, on sodium tungstatebased systems, the catalyst is described as responsible for homolytic cleavage of a C-H bond in methane. Then, ethane formation results from a gas-phase coupling of methyl radicals. In a study comparing MgO and Mn/Na2WO4/SiO2 catalysts, Beck et al. pulsed methane and oxygen under TAP Knudsen conditions with only CO2 formation [115]. Higher intensity pulsing was needed for the observation of C2 products. From the pulse shape characteristics in pump/ probe sequencing of oxygen and methane with different time delays, they surmised that two distinct oxygen species form on the surface: a weakly adsorbed but highly reactive species that leads to CO2 formation and a strongly adsorbed species that is responsible for making the methyl radicals that form ethane in the gas phase. While it would not be possible to resolve these features from steady-state experiments, the transient experiments measured different relative amounts of the two types of oxygen which were used to explain why these materials perform so differently under steady-state conditions.

40

Temporal Analysis of Product (TAP)

40.5

Conclusion

The TAP pulse response technique is a unique tool for characterizing intrinsic kinetic properties and discriminating mechanistic detail of reactions controlled by complex industrial catalytic materials. The low-intensity pulse in the low-pressure packed bed creates well-defined transport, uniform mixing, isothermal operation, incremental titration, and sampling of kinetic properties without disturbing the kinetic state of the system. In the 40 years since its conception, the tool has significantly advanced both theoretically and experimentally and is still evolving. With regard to the theoretical foundations, the recent development of a Python-based TAP reactor simulation and processing program [116] and the ratereactivity model [117] are both data driven methods, which can be extended to capture more accurate descriptions of microkinetic models for complex reaction mechanism in the future. A machine-learned kinetic model is currently under development that aims to predict kinetic performance over a wide range of chemical space. Experimentally, the incorporation of in situ time-resolved transient spectrokinetic measurement, e.g., Raman, IR, and UV-Vis, within TAP is expected on the short term and will provide for a much better understanding of the mechanistic details behind the kinetic measurement and structural changes of the catalysts. TAP has demonstrated unique insight into a wide variety of different catalytic applications. In the future, as accessibility to this distinct tool increases, the impact of this and other transient methods will grow in heterogeneous catalysis and broaden into new research applications beyond catalysis, for example, characterization of the aging, poisoning and degradation process of anode/cathode materials, fuel cells, batteries [118], electrochemical cycles, and surface-reactive properties of sensors and membranes. Acknowledgments The author would like to acknowledge the contributions of J.T. Gleaves, G. Yablonsky, M.R. Kunz, C. Reece, R. Batchu, Y. Chen, J. Posthuma de Boer and Y. Wang in the editing and discussion of this chapter. J.M. Yoda is acknowledged for many fruitful discussions. This chapter was supported by US Department of Energy (USDOE), Office of Energy Efficiency and Renewable Energy (EERE), and Advanced Manufacturing Office Next Generation R&D Projects under contract no. DE-AC0705ID14517.

References 1. Morgan, K., Maguire, N., Fushimi, R., Gleaves, J., Goguet, A., Harold, M., Kondratenko, E.V., Menon, U., Schuurman, Y., Yablonsky, G.: Forty years of temporal analysis of products. Catal. Sci. Technol. 7, 2416–2439 (2017) 2. Fushimi, R., Gleaves, J.: Recent advances in dynamic chemical characterization using Temporal Analysis of Products. Curr. Opin. Chem. Eng. 21, 10–21 (2018)

931 3. Pérez-Ramírez, J., Kondratenko, E.V.: Evolution, achievements, and perspectives of the TAP technique. Catal. Today. 121, 160–169 (2007) 4. Leppelt, R., Hansgen, D., Widmann, D., Häring, T., Bräth, G., Behm, R.: Design and characterization of a temporal analysis of products reactor. Rev. Sci. Instrum. 78, 104103 (2007) 5. Nijhuis, T., Makkee, M., Van Langeveld, A., Moulijn, J.: New insight in the platinum-catalyzed CO oxidation kinetic mechanism by using an advanced TAP reactor system. Appl. Catal. A Gen. 164, 237–249 (1997) 6. Mul, G., Banares, M., Cortez, G.G., Van der Linden, B., Khatib, S., Moulijn, J.: MultiTRACK and operando Raman-GC study of oxidative dehydrogenation of propane over alumina-supported vanadium oxide catalysts. Phys. Chem. Chem. Phys. 5, 4378–4383 (2003) 7. Ebner, J.R., Gleaves, J.T.: US 5264183. Method and apparatus for carrying out catalyzed chemical reactions and for studying catalysis, 1993 8. Gleaves, J.T., Ebner, J.R., Kuechler, T.C.: Temporal analysis of products (TAP)—a unique catalyst evaluation system with submillisecond time resolution. Catal. Rev. Sci. Eng. 30, 49–116 (1988) 9. Kunz, R., Redekop, E.A., Borders, T., Wang, L., Yablonsky, G.S., Fushimi, R.: Pulse response analysis using the Y-procedure computational method. Chem. Eng. Sci. 192, 46–60 (2018) 10. Yablonsky, G.S., Marin, G.B.: Kinetics of Chemical Reactions : Decoding Complexity. Wiley-VCH, Weinheim (2011) 11. Gleaves, J., Yablonskii, G., Phanawadee, P., Schuurman, Y.: “StateDefining” TAP pulse response experiments. In: Studies in Surface Science and Catalysis, pp. 333–340. Elsevier, Amsterdam, the Netherlands (1997) 12. Yablonsky, G.S., Constales, D., Shekhtman, S.O., Gleaves, J.T.: The Y-procedure: how to extract the chemical transformation rate from reaction–diffusion data with no assumption on the kinetic model. Chem. Eng. Sci. 62, 6754–6767 (2007) 13. Wang, Y., Kunz, M.R., Constales, D., Yablonsky, G., Fushimi, R.: Rate/concentration kinetic petals: a transient method to examine the interplay of surface reaction processes. J. Phys. Chem. A 123, 8717–8725 (2019) 14. Perez-Ramirez, J., Kondratenko, E.V.: Evolution, achievements, and perspectives of the TAP technique. Catal. Today. 121, 160–169 (2007) 15. Perez-Ramirez, J., Kondratenko, E.V., Kondratenko, V.A., Baerns, M.: Selectivity-directing factors of ammonia oxidation over PGM gauzes in the Temporal Analysis of Products reactor: primary interactions of NH3 and O2. J. Catal. 227, 90–100 (2004) 16. Herschbach, D.R.: Molecular dynamics of elementary chemical reactions (Nobel lecture). Angew. Chem. Int. Ed. Engl. 26, 1221–1243 (1987) 17. Lee, Y.T.: Molecular beam studies of elementary chemical processes (Nobel lecture). Angew. Chem. Int. Ed. Engl. 26, 939–951 (1987) 18. Eigen, M.: Immeasurably fast reactions. Nobel Lecture. 11, 1963–1979 (1967) 19. Schuurman, Y.: Assessment of kinetic modeling procedures of TAP experiments. Catal. Today. 121, 187–196 (2007) 20. Gleaves, J.T., Yablonskii, G.S., Phanawadee, P., Schuurman, Y.: TAP-2: an interrogative kinetics approach. Appl. Catal. A Gen. 160, 55–88 (1997) 21. van der Linde, S.C., Nijhuis, T., Dekker, F., Kapteijn, F., Moulijn, J.A.: Mathematical treatment of transient kinetic data: combination of parameter estimation with solving the related partial differential equations. Appl. Catal. A Gen. 151, 27–57 (1997) 22. Rothaemel, M., Baerns, M.: Modeling and simulation of transient adsorption and reaction in vacuum using the temporal analysis of products reactor. Ind. Eng. Chem. Res. 35, 1556–1565 (1996)

40

932 23. Van, V.A.C., Farrusseng, D., Rebeilleau, M., Decamp, T., Holzwarth, A., Schuurman, Y., Mirodatos, C.: Acceleration in catalyst development by fast transient kinetic investigation. J. Catal. 216, 135–143 (2003) 24. Gering, K., Baroi, C., Fushimi, R.: Transport modeling and mapping of pulsed reactor dynamics near and beyond the onset of viscid flow. Chem. Eng. Sci. 192, 576–590 (2018) 25. Zheng, X., Gleaves, J.T., Yablonsky, G.S., Brownscombe, T., Gaffney, A., Clark, M., Han, S.: Needle in a haystack catalysis. Appl. Catal. A Gen. 341, 86–92 (2008) 26. Perez-Ramirez, J., Kondratenko, E.V., Kondratenko, V.A., Baerns, M.: Selectivity-directing factors of ammonia oxidation over PGM gauzes in the Temporal Analysis of Products reactor: secondary interactions of NH3 and NO. J. Catal. 229, 303–313 (2005) 27. Kumar, A., Medhekar, V., Harold, M.P., Balakotaiah, V.: NO decomposition and reduction on Pt/Al2O3 powder and monolith catalysts using the TAP reactor. Appl Catal B. 90, 642–651 (2009) 28. Maguire, N., Sasegbon, A., Abdelkader, A., Goguet, A., Hardacre, C., Hellgardt, K., Morgan, K., Shekhtman, S.O.: Using temporal analysis of products and flux response technology to determine diffusion coefficients in catalytic monoliths. Chem. Eng. Sci. 87, 224–233 (2013) 29. Shekhtman, S.O., Yablonsky, G.S., Chen, S., Gleaves, J.T.: Thinzone TAP-reactor - theory and application. Chem. Eng. Sci. 54, 4371–4378 (1999) 30. Phanawadee, P., Shekhtman, S.O., Jarungmanorom, C., Yablonsky, G.S., Gleaves, J.T.: Uniformity in a thin-zone multi-pulse TAP experiment: numerical analysis. Chem. Eng. Sci. 58, 2215–2227 (2003) 31. Shekhtman, S.O., Yablonsky, G.S.: Thin-zone TAP reactor versus differential PFR: analysis of concentration nonuniformity for gas-solid systems. Ind. Eng. Chem. Res. 44, 6518–6522 (2005) 32. Wang, Y., Kunz, M.R., Siebers, S., Rollins, H., Gleaves, J., Yablonsky, G., Fushimi, R.: Transient kinetic experiments within the high conversion domain: the case of ammonia decomposition. Catalysts. 9, 104 (2019) 33. Keipert, O.P., Baerns, M.: Determination of the intracrystalline diffusion coefficients of alkanes in H-ZSM-5 zeolite by a transient technique using the temporal-analysis-of-products (TAP) reactor. Chem. Eng. Sci. 53, 3623–3634 (1998) 34. Delgado, J.A., Nijhuis, T.A., Kapteijn, F., Moulijn, J.A.: Determination of adsorption and diffusion parameters in zeolites through a structured approach. Chem. Eng. Sci. 59, 2477–2487 (2004) 35. Phanawadee, P., Phongaksorn, M., Chaimongkol, N., Jaree, A., Limtrakul, J.: Mathematical analysis of TAP models for porous catalysts. Chem. Eng. J. (Amsterdam, Neth.). 115, 51–62 (2005) 36. Colaris, A.H.J., Hoebink, J.H.B.J., De, C.M.H.J.M., Schouten, J.C.: Intrapellet diffusivities from TAP pulse responses via moment-based analysis. AICHE J. 48, 2587–2596 (2002) 37. Berger, R.J., Kapteijn, F., Moulijn, J.A., Marin, G.B., De Wilde, J., Olea, M., Chen, D., Holmen, A., Lietti, L., Tronconi, E.: Dynamic methods for catalytic kinetics. Appl. Catal. A Gen. 342, 3–28 (2008) 38. Li, W., Xie, L., Gao, L., Zhao, X., Hu, R., Cheng, Y., Wang, D.: The measurement of gas diffusivity in porous materials by temporal analysis of products (TAP). Catal. Today. 121, 246–254 (2007) 39. Wang, D., Li, F., Zhao, X.: Diffusion limitation in fast transient experiments. Chem. Eng. Sci. 59, 5615–5622 (2004) 40. Nayak, S.V., Morali, M., Ramachandran, P.A., Dudukovic, M.P.: Transport and sorption studies in beta and USY zeolites via temporal analysis of products (TAP). J. Catal. 266, 169–181 (2009) 41. Breitkopf, C., Galinsky, M., Lutecki, M.: Diffusion and reaction in nanostructured sulfated zirconias: a transient TAP study. Int. J. Transp. Phenom. 12, 85–92 (2011) 42. Lobera, M.P., Biausque, G., Tellez, C., Herguido, J., Menendez, M., Schuurman, Y.: Comparison of different methods for quantitative analysis of TAP pulse-response data for propane

R. Fushimi dehydrogenation over Pt-Sn-K/γ-Al2O3. Chem. Eng. Sci. 65, 2417–2424 (2010) 43. Mills, P., Duduković, M.: Convolution and deconvolution of nonideal tracer response data with application to three-phase packedbeds. Comput. Chem. Eng. 13, 881–898 (1989) 44. Svoboda, G.D., Gleaves, J.T., Mills, P.L.: New method for studying the pyrolysis of VPE/CVD precursors under vacuum conditions. Application to trimethylantimony and tetramethyltin. Ind. Eng. Chem. Res. 31, 19–29 (1992) 45. Svoboda, G.D., Gleaves, J.T., Mills, P.L.: Detailed modeling of transport-kinetics interactions of ethylene epoxidation at high vacuum and atmospheric pressures. In: Corberán, V.C., Bellón, S.V. (eds.) Studies in Surface Science and Catalysis, pp. 481–493. Elsevier, Amsterdam, the Netherlands (1994) 46. Schiesser, W.E.: The Numerical Method of Lines: Integration of Partial Differential Equations. Elsevier, Amsterdam, the Netherlands (2012) 47. Kondratenko, E.V., Pérez-Ramírez, J.: Micro-kinetic analysis of direct N 2 O decomposition over steam-activated Fe-silicalite from transient experiments in the TAP reactor. Catal. Today. 121, 197–203 (2007) 48. Soick, M., Wolf, D., Baerns, M.: Determination of kinetic parameters for complex heterogeneous catalytic reactions by numerical evaluation of TAP experiments. Chem. Eng. Sci. 55, 2875–2882 (2000) 49. Sincovec, R.F., Madsen, N.K.: Software for nonlinear partial differential equations. ACM Trans. Math. Softw. 1, 232–260 (1975) 50. Huinink, J.P.: A Quantitative Analysis of Transient Kinetic Experiments: The Oxidation of CO by O2/NO on Pt. Eindhoven University of Technology, Eindhoven, the Netherlands (1995) 51. Kushekbayev, N.: Kinetic modeling of complex heterogeneously catalyzed reactions using temporal analysis of product method (2017) 52. Kumar, A., Zheng, X., Harold, M.P., Balakotaiah, V.: Microkinetic modeling of the NO+ H2 system on Pt/Al2O3 catalyst using temporal analysis of products. J. Catal. 279, 12–26 (2011) 53. Campbell, C.T.: The degree of rate control: a powerful tool for catalysis research. ACS Catal. 7, 2770–2779 (2017) 54. Logg, A., Mardal, K.-A., Wells, G.: Automated Solution of Differential Equations by the Finite Element Method: The FEniCS Book. Springer Science & Business Media, New York, USA (2012) 55. Doğu, G., Doğu, T.: Fifty years of moment technique for dynamic analysis of chemical reactor parameters. Int. J. Chem. React. Eng. 17 (2019) 56. Yablonskii, G.S., Shekhtman, S.O., Chen, S., Gleaves, J.T.: Moment-based analysis of transient response catalytic studies (TAP experiment). Ind. Eng. Chem. Res. 37, 2193–2202 (1998) 57. Shekhtman, S.O., Yablonsky, G.S., Gleaves, J.T., Fushimi, R.: “State defining” experiment in chemical kinetics—primary characterization of catalyst activity in a TAP experiment. Chem. Eng. Sci. 58, 4843–4859 (2003) 58. Yablonskii, G.V., Bykov, V., Elokhin, V., Gorban, A.: Kinetic Models of Catalytic Reactions. Elsevier, Amsterdam, the Netherlands (1991) 59. Constales, D., Yablonsky, G.S., Wang, L., Diao, W., Galvita, V.V., Fushimi, R.: Precise non-steady-state characterization of solid active materials with no preliminary mechanistic assumptions. Catal. Today. 298, 203 (2017) 60. Shekhtman, S., Maguire, N., Goguet, A., Burch, R., Hardacre, C.: Three primary kinetic characteristics observed in a pulse-response TAP experiment. Catal. Today. 121, 255–260 (2007) 61. Redekop, E.A., Yablonsky, G.S., Constales, D., Ramachandran, P.A., Pherigo, C., Gleaves, J.T.: The Y-procedure methodology for the interpretation of transient kinetic data: analysis of irreversible adsorption. Chem. Eng. Sci. 66, 6441–6452 (2011) 62. Reece, C.: Kinetic Analysis and Modelling in Heterogeneous Catalysis. Cardiff University, Cardiff, Wales (2017)

40

Temporal Analysis of Product (TAP)

63. Roelant, R., Constales, D., Yablonsky, G.S., Van Keer, R., Rude, M.A., Marin, G.B.: Noise in temporal analysis of products (TAP) pulse responses. Catal. Today. 121, 269–281 (2007) 64. Pérez-Ramırez, J., Kapteijn, F., Mul, G., Moulijn, J.A.: NO-assisted N2O decomposition over Fe-based catalysts: effects of gas-phase composition and catalyst constitution. J. Catal. 208, 211–223 (2002) 65. Redekop, E.A., Lazzarini, A., Bordiga, S., Olsbye, U.: A temporal analysis of products (TAP) study of C2-C4 alkene reactions with a well-defined pool of methylating species on ZSM-22 zeolite. J. Catal. 385, 300–312 (2020) 66. Dewaele, O., Geers, V.L., Froment, G.F., Marin, G.B.: The conversion of methanol to olefins: a transient kinetic study. Chem. Eng. Sci. 54, 4385–4395 (1999) 67. Scharfe, M., Capdevila-Cortada, M., Kondratenko, V.A., Kondratenko, E.V., Colussi, S., Trovarelli, A., López, N., PérezRamírez, J.: Mechanism of ethylene oxychlorination on ceria. ACS Catal. 8, 2651–2663 (2018) 68. Wang, Y., Kapteijn, F., Makkee, M.: NOx reduction in the Di-Air system over noble metal promoted ceria. Appl. Catal. B Environ. 231, 200–212 (2018) 69. Kumar, A., Harold, M.P., Balakotaiah, V.: Isotopic studies of NOX storage and reduction on Pt/BaO/Al2O3 catalyst using temporal analysis of products. J. Catal. 270, 214–223 (2010) 70. Bueno-López, A., Krishna, K., Van der Linden, B., Mul, G., Moulijn, J., Makkee, M.: On the mechanism of model diesel soot-O2 reaction catalysed by Pt-containing La3+-doped CeO2: a TAP study with isotopic O2. Catal. Today. 121, 237–245 (2007) 71. Wang, L.-C., Zhong, Y., Widmann, D., Weissmüller, J., Behm, R.J.: Oxygen adsorption and low-temperature CO oxidation on a nanoporous Au catalyst: reaction mechanism and foreign metal effects. Top. Catal. 61, 446–461 (2018) 72. Wang, L.-C., Personick, M.L., Karakalos, S., Fushimi, R., Friend, C.M., Madix, R.J.: Active sites for methanol partial oxidation on nanoporous gold catalysts. J. Catal. 344, 778–783 (2016) 73. Boudart, M.: Turnover rates in heterogeneous catalysis. Chem. Rev. 95, 661–666 (1995) 74. Boudart, M., Djéga-Mariadassou, G.: Kinetics of Heterogeneous Catalytic Reactions. Princeton University Press, Princeton, USA (1984) 75. Widmann, D., Liu, Y., Schueth, F., Behm, R.J.: Support effects in the Au-catalyzed CO oxidation - correlation between activity, oxygen storage capacity, and support reducibility. J. Catal. 276, 292–305 (2010) 76. Constales, D., Fang, Z., Kunz, M.R., Yablonsky, G., Fushimi, R.: Methods for determining the intrinsic kinetic characteristics of irreversible adsorption processes. Chem. Eng. Sci. 207, 344–351 (2019) 77. Shekhtman, S., Goguet, A., Burch, R., Hardacre, C., Maguire, N.: CO multipulse TAP studies of 2% Pt/CeO 2 catalyst: influence of catalyst pretreatment and temperature on the number of active sites observed. J. Catal. 253, 303–311 (2008) 78. Morgan, K., Cole, K.J., Goguet, A., Hardacre, C., Hutchings, G.J., Maguire, N., Shekhtman, S.O., Taylor, S.H.: TAP studies of CO oxidation over CuMnOX and Au/CuMnOX catalysts. J. Catal. 276, 38–48 (2010) 79. Dewaele, O., Froment, G.F.: TAP study of the sorption of CO and CO2 on γAl2O3. Appl. Catal. A Gen. 185, 203–210 (1999) 80. Linders, M., Van Den Broeke, L., Nijhuis, T., Kapteijn, F., Moulijn, J.: Modelling sorption and diffusion in activated carbon: a novel low pressure pulse-response technique. Carbon. 39, 2113–2130 (2001) 81. Schuurman, Y., Pantazidis, A., Mirodatos, C.: The TAP-2 reactor as an alternative tool for investigating FCC catalysts. Chem. Eng. Sci. 54, 3619–3625 (1999)

933 82. Breitkopf, C.: A transient TAP study of the adsorption of C4-hydrocarbons on sulfated zirconias. J. Mol. Catal. A Chem. 226, 269–278 (2005) 83. Farrusseng, D., Daniel, C., Gaudillere, C., Ravon, U., Schuurman, Y., Mirodatos, C., Dubbeldam, D., Frost, H., Snurr, R.Q.: Heats of adsorption for seven gases in three metal organic frameworks: systematic comparison of experiment and simulation. Langmuir. 25, 7383–7388 (2009) 84. Zheng, X., Veith, G.M., Redekop, E., Lo, C.S., Yablonsky, G.S., Gleaves, J.T.: Oxygen and CO adsorption on Au/SiO2 catalysts prepared by magnetron sputtering: the role of oxygen storage. Ind. Eng. Chem. Res. 49, 10428–10437 (2010) 85. Redekop, E.A., Yablonsky, G.S., Galvita, V.V., Constales, D., Fushimi, R., Gleaves, J.T., Marin, G.B.: Momentary Equilibrium (ME) in transient kinetics and its application for estimating the concentration of catalytic sites. Ind. Eng. Chem. Res. 52, 15417–15427 (2013) 86. Hong, J., Pietrzyk, S., Khodakov, A.Y., Chu, W., Olea, M., Balcaen, V., Marin, G.B.: TAP investigation of hydrogen and carbon monoxide adsorption on a silica-supported cobalt catalyst. Appl. Catal. A Gen. 375, 116–123 (2010) 87. Kondratenko, E., Buyevskaya, O., Soick, M., Baerns, M.: Transient kinetics and mechanism of oxygen adsorption over oxide catalysts from the TAP-reactor system. Catal. Lett. 63, 153–159 (1999) 88. Olea, M., Kunitake, M., Shido, T., Iwasawa, Y.: TAP study on CO oxidation on a highly active au/Ti(OH)4* catalyst. Phys. Chem. Chem. Phys. 3, 627–631 (2001) 89. Buyevskaya, O., Rothaemel, M., Zanthoff, H., Baerns, M.: Transient studies on the role of oxygen activation in the oxidative coupling of methane over Sm 2 O 3, Sm 2 O 3/MgO, and MgO catalytic surfaces. J. Catal. 150, 71–80 (1994) 90. Schwartz, S., Schmidt, L., Fisher, G.B.: Carbon monoxide+ oxygen reaction on rhodium (III): steady-state rates and adsorbate coverages. J. Phys. Chem. 90, 6194–6200 (1986) 91. Morgan, K., Inceesungvorn, B., Goguet, A., Hardacre, C., Meunier, F.C., Shekhtman, S.O.: TAP studies on 2% Ag/γ–Al 2 O 3 catalyst for selective reduction of oxygen in a H 2-rich ethylene feed. Catal. Sci. Technol. 2, 2128–2133 (2012) 92. Kondratenko, E.V., Sakamoto, Y., Okumura, K., Shinjoh, H.: Transient analysis of oxygen storage capacity of Pt/CeO2-ZrO2 materials with millisecond- and second-time resolution. Appl Catal B. 89, 476–483 (2009) 93. Burch, R., Millington, P.J., Walker, A.: Mechanism of the selective reduction of nitrogen monoxide on platinum-based catalysts in the presence of excess oxygen. Appl. Catal. B Environ. 4, 65–94 (1994) 94. Widmann, D., Krautsieder, A., Walter, P., Brückner, A., Behm, R.J.: How temperature affects the mechanism of CO oxidation on Au/TiO2: a combined EPR and TAP reactor study of the reactive removal of TiO2 surface lattice oxygen in Au/TiO2 by CO. ACS Catal. 6, 5005–5011 (2016) 95. Baerns, M., Imbihl, R., Kondratenko, V.A., Kraehnert, R., Offermans, W.K., Van, S.R.A., Scheibe, A.: Bridging the pressure and material gap in the catalytic ammonia oxidation: structural and catalytic properties of different platinum catalysts. J. Catal. 232, 226–238 (2005) 96. Granger, P., Renème, Y., Dhainaut, F., Schuurman, Y., Mirodatos, C.: NO adsorption and reaction on aged Pd–Rh natural gas vehicle catalysts: a combined TAP and steady-state kinetic approach. Top. Catal. 60, 289–294 (2017) 97. Renème, Y., Pietrzyk, S., Dhainaut, F., Chaar, M., van Veen, A.C., Granger, P.: Reaction pathways involved in CH(4) conversion on Pd/Al(2)O(3) catalysts: TAP as a powerful tool for the elucidation of the effective role of the metal/support interface. Front. Chem. 4, 7 (2016)

40

934 98. Mallens, E., Hoebink, J., Marin, G.: The reaction mechanism of the partial oxidation of methane to synthesis gas: a transient kinetic study over rhodium and a comparison with platinum. J. Catal. 167, 43–56 (1997) 99. Creten, G., Lafyatis, D.S., Froment, G.F.: Transient kinetics from the TAP reactor system: application to the oxidation of propylene to acrolein. J. Catal. 154, 151–162 (1995) 100. Hu, R., Xie, L., Ding, S., Hou, J., Cheng, Y., Wang, D.: CO oxidation and oxygen-assisted CO adsorption/desorption on Ag/MnOx catalysts. Catal. Today. 131, 513–519 (2008) 101. Pantazidis, A., Bucholz, S., Zanthoff, H., Schuurman, Y., Mirodatos, C.: A TAP reactor investigation of the oxidative dehydrogenation of propane over a V–Mg–O catalyst. Catal. Today. 40, 207–214 (1998) 102. Kushekbayev, N.: Kinetic Modeling of Complex Heterogeneously Catalyzed Reactions Using Temporal Analysis of Product Method. Nazarbayev University School of Engineering, Astana, Kazakhstan (2017) 103. Reece, C., Redekop, E.A., Karakalos, S., Friend, C.M., Madix, R.J.: Crossing the great divide between single-crystal reactivity and actual catalyst selectivity with pressure transients. Nat. Catal. 1, 1 (2018) 104. Redekop, E.A., Yablonsky, G.S., Constales, D., Ramachandran, P.A., Gleaves, J.T., Marin, G.B.: Elucidating complex catalytic mechanisms based on transient pulse-response kinetic data. Chem. Eng. Sci. 110, 20–30 (2014) 105. Gleaves, J.T., Sault, A.G., Madix, R.J., Ebner, J.R.: Ethylene oxidation on silver powder: a TAP reactor study. J. Catal. 121, 202–218 (1990) 106. Wang, Y., Qian, J., Fang, Z., Kunz, M.R., Yablonsky, G., Fortunelli, A., Goddard III, W.A., Fushimi, R.R.: Understanding reaction networks through controlled approach to equilibrium experiments using transient methods. J. Am. Chem. Soc. 143, 10998–11006 (2021) 107. Nobukawa, T., Yoshida, M., Kameoka, S., Ito, S.-i., Tomishige, K., Kunimori, K.: Role of nascent oxygen transients in N2O decomposition and selective catalytic reduction of N2O. Catal. Today. 93, 791–796 (2004) 108. Kondratenko, E.V., Perez-Ramirez, J.: Importance of the lifetime of oxygen species generated by N2O decomposition for hydrocarbon activation over Fe-silicalite. Appl Catal B. 64, 35–41 (2006) 109. Chakrabarti, A., Ford, M.E., Gregory, D., Hu, R., Keturakis, C.J., Lwin, S., Tang, Y., Yang, Z., Zhu, M., Banares, M.A.: A decade+ of operando spectroscopy studies. Catal. Today. 283, 27–53 (2017) 110. Laudenschleger, D., Ruland, H., Muhler, M.: Identifying the nature of the active sites in methanol synthesis over Cu/ZnO/Al2O3 catalysts. Nat. Commun. 11, 3898 (2020) 111. Assmann, J., Loffler, E., Birkner, A., Muhler, M.: Ruthenium as oxidation catalyst: bridging the pressure and material gaps between ideal and real systems in heterogeneous catalysis by applying DRIFT spectroscopy and the TAP reactor. Catal. Today. 85, 235–249 (2003) 112. Assmann, J., Narkhede, V., Breuer, N., Muhler, M., Seitsonen, A., Knapp, M., Crihan, D., Farkas, A., Mellau, G., Over, H.:

R. Fushimi Heterogeneous oxidation catalysis on ruthenium: bridging the pressure and materials gaps and beyond. J. Phys. Condens. Matter. 20, 184017 (2008) 113. Delgado, J.A., Nijhuis, T.A., Kapteijn, F., Moulijn, J.A.: Modeling of fast pulse responses in the Multitrack: an advanced TAP reactor. Chem. Eng. Sci. 57, 1835–1847 (2002) 114. Svoboda, G.: Fundamental Transport-Kinetics Models for Interpretation of Temporal Analysis of Products (TAP) Reactor Transient Response Data with Application to Reactive Systems. Washington University Saint Louis, Saint Louis, USA (1993) 115. Beck, B., Fleischer, V., Arndt, S., Hevia, M.G., Urakawa, A., Hugo, P., Schomäcker, R.: Oxidative coupling of methane—a complex surface/gas phase mechanism with strong impact on the reaction engineering. Catal. Today. 228, 212–218 (2014) 116. Yonge, A., Kunz, M.R., Batchu, R., Fang, Z., Issac, T., Fushimi, R. Medford, A.J.: TAPsolver: A Python package for the simulation and analysis of TAP reactor experiments. Chem. Eng. J., 420, 129377 (2021) 117. Kunz, M.R., Yonge, A., Fang, Z., Batchu, R., Medford, A.J., Constales, D., Yablonsky, G., Fushimi, R.: Data driven reaction mechanism estimation via transient kinetics and machine learning. Chem. Eng. J., 420, 129610 (2021) 118. Fang, Z., Confer, M.P., Wang, Y., Wang, Q., Kunz, M.R., Dufek, E. J., Liaw, B., Klein, T. M., Dixon, D.A., Fushimi, R.: Formation of surface impurities on Lithium–Nickel–Manganese–Cobalt oxides in the presence of CO2 and H2O. J. Am. Chem. Soc. 143(27), 10261–10274 (2021)

Dr. Fushimi is a research scientist working in the emerging areas of dynamic catalyst science and flexible chemical manufacturing. Her research is focused on using/developing transient kinetic tools where dynamics in chemical systems can reveal reaction networks and mechanism. Using TAP and other dynamic techniques, her research group investigates selective oxidation, dehydrogenation, selective hydrogenation, ammonia synthesis, and other reactions on supported metals and mixed metal oxide catalysts.

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

41

Anders Holmen, Jia Yang, and De Chen

Contents 41.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 936

41.11.3

41.2

SSITKA Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 937

41.11.4

41.3 41.3.1 41.3.2

SSITKA Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Single Pool First-Order Irreversible Reaction . . . . . . . . . . Multiple Pools in Series for a First-Order Irreversible Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multiple Pools in Parallel for a First-Order Irreversible Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reversible Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41.3.3 41.3.4

937 937 939 940 940

41.4 41.4.1

Complicating Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 941 Gas Phase Hold-Up, Readsorption, and Chromatographic Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . 941

41.5 41.5.1 41.5.2 41.5.3

Reactivity Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fitting to Exponential Functions (Parametric Approach) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inverse Laplace Transform (ILT) Method . . . . . . . . . . . . . . Tikhonov-Fredholm (T-F) Method . . . . . . . . . . . . . . . . . . . . . .

41.6

Reactors and Isotope Effects . . . . . . . . . . . . . . . . . . . . . . . . . . 943

41.7

Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 944

41.8 41.8.1 41.8.2

Combination of SSITKA with Spectroscopic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 944 SSITKA-FTIR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 944 SSITKA-Neutron Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 946

41.9

Other Considerations and Recent Developments . . .

41.10 41.10.1

Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 947 Discriminate Between Different Reaction Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 947

41.11 41.11.1

Examples of the Use of SSITKA . . . . . . . . . . . . . . . . . . . . . . 948 CO Hydrogenation on Al2O3-Supported Co Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 948 The Use of Multicomponent SSITKA to Obtain Kinetic Parameters for Higher Hydrocarbons in CO Hydrogenation [8] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 951

41.11.2

943 943 943 943

946

A. Holmen (*) · J. Yang · D. Chen Department of Chemical Engineering, Norwegian University of Science & Technology, Trondheim, Norway e-mail: [email protected]

Surface Species and Mechanistic Studies by Combination of SSITKA and Kinetic Isotope Effect . . . 951 Combing DFT and Transient and Steady-State Modeling to Study Reaction Mechanism of CO Hydrogenation . . . 952

41.12

Examples of Combining SSITKA-DRIFT for WGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954

41.13

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 958

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 959

Abstract

The principles and practical use of steady-state isotopic transient kinetic analysis (SSITKA) for studying heterogeneous catalytic reactions are presented. SSITKA combines the advantages of steady-state and transient techniques. A basic requirement for SSITKA is that steady-state conditions are maintained during the isotope switch, and for 12C/13C, 14N/15N, and 16O/18O, the isotope effect is small and these isotopes are all used. The use of H2/D2 is not straightforward since a H2/D2 switch may induce kinetic isotopic effects. Hydrogen spillover to the support and exchange with surface hydroxyl groups may also take place during a H2/D2 switch. The SSITKA principle is outlined based on a simple reaction with adsorbed intermediates. The presentation of the technique is then extended to include complex multi-pool SSITKA and to reactivity distributions. A standard experimental set-up is described and several examples of reactions studied with SSITKA are included. The technique makes it possible to identify the abundance of intermediates and their kinetic parameters. It is not possible to determine the composition of the adsorbed intermediates, but recent developments combining SSITKA with spectroscopic techniques are promising developments. Examples of the combination of SSITKA and spectroscopic techniques are included.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_41

935

936

A. Holmen et al.

Keywords

Steady state · Transient experiments · Isotopic labeling · Surface intermediates · Mean residence time · First-order rate constant · Site coverage · Turnover frequency (TOF)

41.1

Introduction

Steady-state isotopic transient kinetic analysis (SSITKA) is an in situ technique suitable for addressing the intrinsic kinetics of catalytic reactions by analysis of surface species and their residence times. The advantage of this technique is that it can decouple the effect of the concentration of intermediates and site reactivity on the catalytic activity. SSITKA can provide information at, or near to, molecular level under realistic, steady-state conditions such as coverage of surface intermediates, surface residence time, intrinsic turnover frequency, surface heterogeneity, catalyst dispersion, and reactivity distribution. SSITKA is usually carried out at close to atmospheric pressure, but high-pressure experiments have also been carried out [1]. The SSITKA method was initially developed by Happel [2], Bennett [3], and Biloen [4]. The methodology has already been reviewed in 1995 by Shannon and Goodwin [5], and more recently by Ledesma, Yang, Chen, and Holmen [6]. The review from 2014 [6] contains 259 references and includes references to most of the groups that have been working with SSITKA. However, several articles showing the broad area of applications of SSITKA have appeared since 2014 and a brief summary of recent articles is given at the end of the introduction. A SSITKA experiment involves an abrupt switch to the corresponding isotopically labelled feed stream with identical

flow and pressure. The steady state is maintained during the whole experiment, provided that no isotope effect occurs. The transient response of the product is recorded by a mass spectrometer, and it contains the kinetic information. An illustration of a setup used for SSITKA studies is shown in Fig. 41.1 together with experimental results obtained by a switch from 12 CO/H2/Ar to 13CO/H2/Kr for Ru supported on Al2O3 [7]. Since the reactor operates under isothermal and isobaric conditions, the catalytic surface does not suffer any change during the isotopic switch and mechanistic studies can be carried out. SSITKA including the combination with spectroscopic techniques has been applied to several catalytic systems. The reaction systems studied by SSITKA since 1995 are summarized in Table 41.1. Several recent studies using SSITKA are concentrated on CO hydrogenation. Studies based on Co catalysts are carried out by Vasiliades et al. [11], Rebmann et al. [15], Ledesma et al. [8], den Otter et al. [18], and Carvalho et al. [12]. Studies based on Fe catalysts include Yang et al. [13] and Xie et al. [19]. CO2 hydrogenation has also recently become a popular subject and the reaction has been studied by Zabilskiy et al. [119] using a large number of different techniques in addition to SSITKA, by Vrijburg et al. [61] and also by Gutterød et al. [59]. Examples of other reactions recently studied by SSITKA are DeNOx (Jablonska et al. [120], Pinaeva et al. [121], Kalamaras et al. [89]), hydrogenation of acetylene (Gulyaeva et al. [114]), alcohol conversion (Polo-Garzon et al. [122], Castellanos et al. [103], Hanspal et al. [123]), ammonia synthesis (Kammert et al. [62]), CH4 combustion (Rotko et al. [98, 99]), reforming (Takise et al. [92], Vasiliades et al. [94], Polo-Garzon et al. [93]), and water–gas shift

1.0

0.8 Kr Ar 13CO 12CO 13CH 4

F(t)

0.6 12CO/H /Kr 2

0.4

GC Vent

Reactor

MS

12CH 4

0.2

0.0 12CO/H /Ar 2

0

5

10

15

20

25

30

35

40

Time (s)

Fig. 41.1 Flow sheet of a setup for carrying out SSITKA experiments. Normalized transients obtained for a switch from H2/12CO/Ar to H2/13CO/Kr obtained at 523 K, 5.5 kPa CO, 55 kPa H2, and

124.5 kPa inert on Ru/Al2O3 (Reprinted with permission from Ref. [7]. Copyright © 2011 Elsevier Inc.)

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

937

Table 41.1 Reaction systems studied by SSITKA since 1995 Reaction systems CO hydrogenation/Fischer Tropsch CO2 hydrogenation Ammonia synthesis Water–gas shift Oxidative dehydrogenation Butane isomerization Selective NO reduction Methanol decomposition/steam reforming Toluene steam reforming Methane dry reforming Phenol steam reforming Propene epoxidation CH4 oxidation CO oxidation Methanol oxidation Additional oxidation reactions Selective hydrogenation of acetylene Surface acidity N2 adsorption O2 adsorption/desorption and diffusion

References [1, 7–57] [58–61] [62–66] [67, 68] [69–72] [73–76] [77–89] [90, 91] [92] [93, 94] [95, 96] [97] [98, 99] [100–102] [103–109] [110–113] [114] [115] [116] [117, 118]

(Polo-Garzon et al. [67]). Several studies focusing on the method including experimental details have also appeared in recent literature: van Bellegham et al. [10], Terreni et al. [124], Olds et al. [116, 125], Thomas et al. [126], Brush et al. [127], and Rochoux et al. [117]. SSITKA has been used extensively in our laboratory in Trondheim. Experimental details can be found in several PhD theses [128–131]. The experimental setup has recently been rebuilt and extended to include the use of IR.

41.2

SSITKA Principle

For heterogeneous catalytic reactions at steady state, the amount of reactant (R), intermediates (X), and products (P) are constant: R$X$P

ð41:1Þ

However, irrespective of the underlying mechanism and kinetic model, a SSITKA experiment permits the determination of the amount of adsorbed intermediates (NP) converted to products and the overall mean surface residence time of the intermediates (τp). Following the isotopic switch: R˟ ! X˟ ! P˟

ð41:2Þ

the total amount of old isotopic label in the product species P from t ¼ 0 to t ¼ 1 is equivalent to the total steady-state amounts of intermediate species, Np, present on the catalyst surface that leads to product species P.

NP ¼

ð1 0

r P ðtÞdt ¼

ð1



0

 r p  r˟p ðtÞ dt

ð41:3Þ

where rP(t) is the reaction rate of unlabelled products. Following the isotopic switch, the reaction rate of unlabelled products can be expressed as: r P ðtÞ ¼ r p  r˟p ðtÞ

ð41:4Þ

where r p is the steady-state reaction rate and r˟p(t) is the reaction rate of labelled products. F(t) in Fig. 41.1 is simply the normalized transient response, defined as: Fp ðtÞ ¼

r p ðtÞ rp

ð41:5Þ

where Fp(t) ¼ 1  Fp˟ (t). The overall mean surface residence time is obtained by integration of the normalized transient response: τp ¼

ð1 0

Fp ðtÞdt ¼

ð1



0

 1  Fp˟ ðtÞ dt

ð41:6Þ

Rearranging of Eqs. (41.3), (41.5), and (41.6) yields Eq. (41.7): N P ¼ r P  τP

ð41:7Þ

The amount of adsorbed intermediates converted to products (Np) and the overall mean surface residence time (τp) are the most general parameters obtainable from SSITKA. The Np is typically in the range of 10–500 umol/g, τp 0.1–200 s for most of the reactions studied [94, 132]. As shown above, τp and Np can be obtained without any assumption of the reaction mechanism. However, additional kinetic parameters may be calculated but are dependent on assumptions and specific models for the reaction. In SSITKA, the catalyst surface is considered to be composed of a system of interconnected pools where each pool represents a homogeneous or well-mixed subsystem within the reaction pathway. The pool can be treated as a continuously stirred tank reactor (CSTR) meaning that there are no changes in composition on the catalyst surface. A separate pool is assumed to exist for each unique adsorbed reaction species or type of catalytic active sites.

41.3

SSITKA Modeling

41.3.1 Single Pool First-Order Irreversible Reaction Consider the first-order irreversible reaction R ! P that takes place on a heterogeneous catalyst:

41

938

A. Holmen et al.

R!X!P

ð41:8Þ

where X is the adsorbed intermediate for the reaction R ! P. (Intermediates can be located in different well-mixed pools or subsystems on the surface as will be discussed later.) A SSITKA experiment consists of a switch from R to the corresponding isotopically labelled Rx at t  t0: x

x

R !X !P

x

ð41:9Þ

It is assumed that the concentration of X is constant over the catalytic surface and that the reservoir of X on the surface can be modeled as a CSTR. It means that the mixing of adsorbed intermediates on the surface is fast compared to the change of composition of the isotopes and that the composition of isotopes on the surface is equal to the composition in the product gas as illustrated in Fig. 41.2. It is also assumed that no isotopic effect occurs meaning that the rate of reaction is constant and not dependent on whether the reactant is R or Rx. Since all intermediates on the catalytic surface have the same reactivity, the composition of isotopes on the surface reservoir at t  t0 is: FX ð t Þ ¼

NX N ¼ X N X þ N Xx N X

ð41:10Þ

For the adsorbed intermediate, Xx, the symmetrical function prevails since, at any time, the sum of adsorbed intermediates is constant and equal to N X ¼ N X þ N Xx . Rearranging Eq. (41.10) gives: 1  FX ð t Þ ¼

N Xx NX

d N X ðtÞ r R ðtÞ r p ðtÞ ¼  dt N X NX NX N X ðtÞ NX

the intermediate X, FX(t). Since rR(t) ¼ 0 for t  0 and combining with Eq. (41.5), we have: dFX ðtÞ 1 1 ¼ r p ðtÞ ¼ r p  Fp ðtÞ  dt NX NX

For t > 0, no more R reacts, since R is switched to R˟, Eq. (41.9). It is assumed that no isotopic effect occurs meaning that the reaction rate is constant and independent of whether the reactant is R or R˟. Dividing both sides of

ð41:14Þ

According to the definition, the mean surface residence time of the intermediates is: τX ¼

NX rp

ð41:15Þ

If the pool can be modeled as a CSTR, the isotopic composition of the pool equals that of the product, hence FX(t) ¼ Fp(t). Substitute FX(t) and Fp(t) and

1 τX

NX rp

in Eq. (41.14) with

yield dFp ðtÞ Fp ðtÞ ¼ dt τX

ð41:16Þ

Integration of Eq. (41.16) with boundary conditions Fp(t) ¼ 1 at t ¼ 0 yields: Fp ðtÞ ¼ e

t=τX

ð41:17Þ

The overall mean surface residence time is then given by: τp ¼

ð41:12Þ

ð41:13Þ

is equivalent to the normalized transient response for

ð41:11Þ

The mass balance for unlabelled intermediates can be expressed as: dN X ðtÞ ¼ r R ðtÞ  r p ðtÞ dt

Eq. (41.12) with the total steady-state amount of the intermediate species, N X , yields:

ð1 0

Fp ðtÞdt ¼

ð1 e 0

t=τX

dt ¼ τX

ð41:18Þ

The first-order rate expression is given as r p ¼ k  N X , where k is the rate constant. It follows then from Eq. (41.18): k¼

rp 1 1 ¼ ¼ τ τ NX X p

ð41:19Þ

The turnover frequency (TOF) is defined as the rate based on the number of active sites on the surface:

RX

TOFp ¼

X

XX

P

PX

Fig. 41.2 Composition of intermediates (X and Xx) on the surface after an isotope switch

rp k  N X 1 ¼ ¼  θX τp Ns Ns

ð41:20Þ

Ns is the total number of active sites, typically determined by H2 chemisorption, and θX is surface coverage of intermediates at steady state. N X , the concentration of adsorbed intermediates determined by SSITKA, was sometimes used to represent “working” active sites, which will be distinct from

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

the total number of active sites determined by chemisorption. Being able to quantify the number of “working” active sites is a clear advantage of the method. It should be noted that for a first-order reaction, k is pseudo-first-order rate constant, but for general cases, k represents a “reactivity.” SSITKA makes it possible to deconvolute the contributions to the turnover frequency, which at steady state for a first-order reaction is the product of the intrinsic rate constant and the surface coverage (θ). This is one of the main advantages of SSITKA. As shown in Fig. 41.3, the isotope composition of adsorbed intermediates in the reservoir on the catalyst surface follows an exponential curve, i.e., the curve is characterized by one single parameter τx ¼ τp. As shown by Eq. (41.20), it is possible to calculate θx, the surface coverage of intermediates based on τp, provided that TOF is known from independent measurements such as H2 chemisorption.

41.3.2 Multiple Pools in Series for a First-Order Irreversible Reaction Consider the reaction R ! P that involves two adsorbed intermediates: R!X!Y!P

939

with the time necessary for the labelled components to go through the two pools. After the switch at t  t0, the amount adsorbed X will decrease exponentially as shown above. At the reactor outlet, however, the reduction in product P is determined by the reduction in intermediate Y. The reduction of Y does not follow an exponential curve since it takes time for the labelled component to reach pool two. The study of the initial response after an isotope switch may therefore give important information about adsorbed intermediates as indicated in Fig. 41.5. The first pool is independent of the second pool and the isotopic composition after the isotope switch given by: F1 ð t Þ ¼ e

ð41:22Þ

A mass balance for the second pool (CSTR) gives: F2 ðtÞ ¼

t t   τ1 τ2  e τ1   e τ2 τ1  τ2 τ1  τ2

ð41:23Þ

The composition in pool 2, (F2(t)), is the composition observed with the MS at the reactor outlet, and τ1 and τ2 are the time constants for pool 1 and 2, respectively. In general, the procedure for analyzing experimental SSITKA-data is to consider n CSTR in series. A certain

ð41:21Þ

X and Y are the adsorbed intermediates located in two pools of CSTR in a series as illustrated in Fig. 41.4. For two pools of adsorbed intermediates in a series, the reduction in product P after the isotope switch will be delayed

t=τ1

R

X

Y

P

Fig. 41.4 Two pools of CSTR in a series for the reaction (Eq. 41.21): R!X!Y!P

1

1

Fp(t)

Ro X oYoP

F(t)

tp = tx = k –1

e–t/t p

Ro X oP tp

t=0

0 Time

Fig. 41.3 Illustration of a SSITKA experiment

Time

Fig. 41.5 Comparison of a single pool model and two pools in series

41

940

A. Holmen et al.

coverage of adsorbed intermediates on the catalyst surface, θ, is divided into n pools of adsorbed intermediates. This corresponds to a CSTR with a given volume that is divided into n smaller CSTR. It should be mentioned that for n ! 1, the transient response can be modeled as a PFR reactor. The transient response for 2 CSTR in series (Eq. 41.21) is shown in Fig. 41.5 and compared with 1 CSTR, Eq. (41.8).

41.3.3 Multiple Pools in Parallel for a First-Order Irreversible Reaction Pools in parallel are typically the case for a nonhomogeneous catalyst surface with multiple sites with varying activity or multiple pathways to the same product. For irreversible first-order reactions, each single reaction pathway can be treated as in Eq. (41.17), and for n pathways in parallel, we simply obtain: Fp ðtÞ ¼

n X

Fp ð t Þ ¼

t t   N1 N2  e τ1 þ  e τ2 N1 þ N2 N1 þ N2

ð41:26Þ

The differences between the different kinetic models can easily be seen in a plot of F(t) or (ln F(t)) as shown in Fig. 41.6. The possibility by experiments to distinguish between different reservoirs of adsorbed intermediates is another important advantage of transient kinetics compared with stationary kinetics. A detailed discussion of the use of the different models can be found in the review of Shannon and Goodwin [5], and their results can be summarized in Table 41.2 reproduced from Shannon and Goodwin [5, 7] and Ledesma et al. [6]. In Table 41.2, reaction models, transient responses, and kinetic parameters obtained by SSITKA are shown.

41.3.4 Reversible Reactions Equation (41.17) is based on a first-order irreversible reac-

yi e

t=τi

ð41:24Þ

i¼1

where yi is the steady-state fractional amount of the surface intermediates (Ni) in the ith pool: Ni yi ¼ P n Ni



t

tion: Fp ðtÞ ¼ e τX . If a reversible reaction is considered: R $ X ! R and the corresponding isotopic labelled species: Rx $ Xx ! Px, the mass balance equations is:

ð41:25Þ

Rate of formation of X : r !  FX ðtÞ

ð41:27Þ

Rate of consumption of X : r  FX ðtÞ þ r p  FX ðtÞ ¼ r !  FX ðtÞ

ð41:28Þ

i¼1

For two pools in parallel, Eqs. (41.24) and (41.25) yields:

where r! and r are the rates for the reversible reaction step. Equations (41.27) and (41.28) then give:

a)

b) Single pool 1

2 pools in parallel

0

F(t)

Ln(F(t))

2 pools in series

Single pool 2 pools in parallel 2 pools in series

t=0

Time

t=0

Time

Fig. 41.6 The transient response (a) and the logarithmic transient response (b) of a single pool, two pools in series and two pools in parallel (Reproduced from Refs. [5, 6]. Copyright © 1995, 2014 American Chemical Society)

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

941

Table 41.2 Mechanistic models, transient responses, and kinetic parameters obtained by SSITKA 1

Reaction Irreversible

Adsorption Irreversible

Pool system Single

Model rR

Transient response and kinetic parameter FpðtÞ ¼ et=τX ; r R ¼ r P N τX ¼ x τP; k ¼ τ1 i

rP

R!X!P

41

rP

TOFi ¼ 2

3

Irreversible

Irreversible

Reversible

Single

Irreversible

rdR

¼ τ1 i

Ni

¼ kθi

Ns

FpðtÞ ¼ et=τX ; r P ¼ r R  r dR

rP

R,X!P

Multiple in series

ri Ns

τX ¼ r

R ! Xi ! P

Nx d P þr R

FP ðtÞ ¼ n P

τP ¼

¼ τP

n P

Qn

τn1 i

Qn

τn1 i

t

τ τ j¼1; j6¼i ð i j Þ

i¼1

τ

e

i

τi

i¼1

4

Irreversible

Reversible

Multiple in series

rdR

R , X1 ! Xi ! P

FP ðtÞ ¼

n P

j¼1; j6¼i

i¼1

τ1 ¼ τP ¼

N1 rP þrdR n P

; τi,i>1 ¼

ðτi τl Þ

et=τi

Ni rP

τi

i¼1

5

Irreversible

Irreversible

Multiple in parallel

R ! Xi ! P R!⋮!P R ! Xn ! P

FP ðtÞ ¼

n P

xi ¼ Pn

i¼1

6

Irreversible

Reversible

Multiple in parallel

rdR

rP

R,X!P R!⋮!P R ! Xn ! P

xi et=τi

i¼1 Ni

FP ðtÞ ¼

Nj

n P

N i¼1 j

τP ¼

n P

n P

xi τ i

i¼1

xi et=τi

i¼1 Ni

xi ¼ Pn

; τP ¼

; τi ¼

Ni r i,P þr dR,i

xi τi

i¼1

Reproduced with permission from Refs. [5] and [6]. Copyright © 1995, 2014 American Chemical Society

τX ¼

NX ¼ r!

N  X  r rp  1 þ r

ð41:29Þ

p

The equivalent to Eq. (41.20) will then be for a reversible reaction:  r θX ¼ TOFP  τX  1 þ r

ð41:30Þ

As shown by Eq. (41.29), reversible reactions tend to delay the transient response of X and therefore also the product P.

41.4

Complicating Factors

41.4.1 Gas Phase Hold-Up, Readsorption, and Chromatographic Effect An experimental setup, as shown in Fig. 41.11, typically consists of three parts: feed, reaction zone with the catalytic bed, and detection of products. Mass flow controllers have been used to control the same total flow with the labelled and

unlabelled gas lines. The pressures of the two parallel lines are controlled with digital back pressure controllers. The surface residence time obtained from the integration of the measured transient responses is the sum of all of these sections. The reactions take place in the catalytic bed, and therefore, it is important to account for the gas hold-up in the system. The gas phase will also immediately after a switch slightly mix with the gas phase prior to the switch. This is observed as a non-frontal movement through the reactor, as shown in Fig. 41.7. The gas phase hold-up is minimized by using 1/1600 gas lines after the switching valve and MS. The reactor volume can be minimized by filling the void with a quartz rod on top of the catalyst bed. An easy way to account for the gas phase hold-up and the nonideal flow is to include an inert tracer in one of the feeds or an inert switch simultaneously as the isotopic switch. The residence time of the inert tracer is only due to the gas phase hold-up: τp,corrected ¼ τp,measured  τinert

ð41:31Þ

Another important issue is the readsorption of products after reaction; a typical example is the readsorption of

942

A. Holmen et al.

methanol in the methanol synthesis reaction [132]. Readsorption is also part of the measured transient response. The effect of intraparticle readsorption on the measured residence time can be reduced by performing SSITKA at different space times and extrapolate to zero space-time as shown

1.0 Product Inert

F(t)

0.8 0.6

in Fig. 41.8a, b. The intrapore readsorption can be corrected by co-feeding unlabelled methanol for competitive adsorption, as shown in Fig. 41.8c [132]. In addition to the readsorption of products, the adsorption of reactants should also be accounted for. The reactants can adsorb not only on the catalyst surface but also throughout the whole reactor system, leading to a delay in the change of isotope composition with respect to an inert tracer. The chromatographic effects should be kept small to avoid interference with the intrinsic kinetics. An example of chromatographic effect is shown in Fig. 41.9. It has been suggested that this can be corrected for by Eq. (41.32):

0.4 0.2

τp,corrected ¼ τp,measured  τinert  x  τreactant t gas

0.0 Time

Fig. 41.7 A schematic representation of gas phase hold-up as indicated by the residence time of the inert gas

Fig. 41.8 (a) Experimental calculation of the mean surfaceresidence time in the presence of interparticle readsorption; (b) τMeOH vs. space time during steady-state MeOH synthesis at 493 K; (c) τMeOH vs. space time at constant average PMeOH (~ ¼ 26.5 Pa, ● ¼ 38.5 Pa, and ■ ¼ 50.5 Pa) during steady-state MeOH synthesis at 493 K on Pd (Reproduced from Ref. [133]. Copyright © 2006 World Scientific Publishing Co.)

ð41:32Þ

The value of x in Eq. (41.32) varies among different users. It is typically set to 0.5 as proposed by Biloen et al. [134]. Here, 0.5 is considered an average of two boundary situations, where the product formed at the inlet is not delayed by the readsorption of the reactant while the product formed at the end of the bed will be fully delayed by the readsorption of the reactant.

a)

t pobs

t psurf Lb or 1/WHSV

b)

c)

200

160

t MeOH (s)

t MeOH (s)

140

240

t 0MeOH 60

40

t 00MeOH

20 0

0.1

0.3 Space time (s)

0.5

0

0.1

0.3 Space time (s)

0.5

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

irreversible first-order reaction on all sites, the transient product formation after a switch is given by:

1.0 0.8

F(t)

943

R R' Inert

0.6 0.4 0.2 0.0 t=0

Time



r ðtÞ ¼ θ0

ð1 0

k  f ðk Þ  e

kt

dk

ð41:33Þ

where θ0 is the initial surface coverage of intermediates that react to the detected product with a distribution f(k). The ILT method is based on the fact that r(t) represents the Laplace transform of θ0 k f(k), hence an inverse Laplace transformation of the measured r(t) or F(t) transient can be applied to extract the f(k) distribution. The numerical procedure requires a fit of the experimental transient to a sum of exponential terms according to Eq. (41.34) below in order to determine the coefficients an.

Fig. 41.9 The delay of the reactant R as compared to inert without reaction is due to chromatographic effect



r ðt Þ ¼

N X

an  expðkn tÞ

ð41:34Þ

n¼1

41.5

Reactivity Distributions

The surface of a heterogeneous catalyst is characterized by the intrinsic reactivity of active sites. The basic interpretation of the transient experimental results is done by the integration and corrections as shown above. The multi-pool model in the previous section describes discrete site reactivity. The reactivity distribution function f(k) describes a continuous reactivity distribution of the catalyst surface. The reactivity distribution function f(k) can be calculated by numerical deconvolution methods, such as parametric and nonparametric methods as described by Rothaemel et al. [135] and Ledesma et al. [6]. The parameter k is a measure of site activity.

41.5.1 Fitting to Exponential Functions (Parametric Approach) The reactivity of the active sites can be obtained by calculation of the reactivity distribution function ( f(k)) from the isotopic transient curve using parametric methods. However, an assumption of its functional form is needed. If the number of pools exceeds two, this method may be difficult to use due to the complexity of the exponential function. Nonparametric methods are preferred since no assumption of a specific form of the function is required. Two nonparametric methods are mentioned below:

41.5.2 Inverse Laplace Transform (ILT) Method The ILT method was developed to derive a continuous reactivity distribution of the catalyst surface, instead of using distinct values for the intrinsic reactivity. Assuming

Although the number of exponential terms, N, has to be infinity, for practical reasons, a limited number (N < 20) resulting in a sufficient large upper limit of the reaction rates has been used to fit the data. For a more detailed description, refer to Rothaemel et al. [135] and Ledesma et al. [6]. Regarding the determined rate constants, it should be noted that if the initial surface coverage, θ0, is unknown, only the effective, pseudo-first-order rate constants k′i ¼ k1 θ0 can be determined.

41.5.3 Tikhonov-Fredholm (T-F) Method In the T-F method, the reactivity distribution can be extracted from the measured r(t) by a Tikhonov regularization of Eq. (41.34), which can be regarded as a Fredholm integral of the first kind. Detailed description of the method has been given by Shannon and Goodwin [5], Ledesma et al. [6], Rothaemel et al. [135], and Hoost and Goodwin [136].

41.6

Reactors and Isotope Effects

The simplest reactor to make a mathematical model for is the continuous stirred tank reactor (CSTR) whereas the plug flow (PFR) requires a more complicated reactor model. The CSTR is a complex reactor setup which requires a high recycle flow rate to maximize the superficial linear velocity through the catalyst bed. The high flow increases the consumption of expensive isotopes and decreases their detectability. The PFR, however, can be made with a low reactor volume and a simple operation. As a result, it has been common practice to use micro-PFR, which can be modeled with gradientless

41

944

A. Holmen et al.

conditions in the catalyst bed, and hence it will be analogous to a CSTR. If the reaction mechanism contains reversible reaction steps of reactants or products, the CSTR approach is no longer valid and a PFR model should be applied. The transient curve using PFR modelling applied to CO hydrogenation is discussed in detail by Ledesma et al. [6]. State-state conditions are of course essential for the correct use of SSITKA and mostly the use of isotopes maintains steady state. For 12C/13C, 14N/15N, and 16O/18O, the isotopic effect is small, and the mentioned isotopes are all used in SSITKA studies. However, for H2/D2, the use is not straightforward and a H2/D2 switch may induce kinetic isotopic effects. If spillover of hydrogen to the support or/and exchange of H2/D2 with surface hydroxyl groups takes place, an additional reservoir has to be accounted for in the hydrogen trajectory [137, 138]. For further reading about H2/ D2 exchange, see Basallote et al. [138]. The kinetic isotope effects (KIE) is defined as the ratio of the rates with CO/H2 and CO/D2 as reactants (rH/rD) in CO hydrogenation reaction. If the ratio is greater than unity, the isotope effect is normal; if less than unity, it is inverse. Typically, an inverse KIE was found on cobalt-based catalyst for CO hydrogenation. This means that the reaction rate with D2/CO reactants was higher than with H2/CO, and an inverse KIE is shown in Fig. 41.10. The magnitude of KIE can give us information on the involvement of H up to the rate-limiting step [140]. The approach of combining experimental KIE and DFT calculations was proven as an efficient method to study reaction mechanisms for complex reaction, such as CO activation, methane formation, and C-C coupling mechanism in the Fischer-Tropsch synthesis [141–143]. Basically, the overall KIE of all steps up to the rate-determining step of each reaction scenario will be calculated and compared to the

41.7

Experimental Setup

A flow sheet of a typical experimental equipment is shown in Fig. 41.11. There are two parallel lines for the feed, one for unlabelled gas mixture (ul) and one for the labelled mixture (l). The flow and pressure are regulated by mass flow controllers (MFC) and back pressure controllers (PC). The switching valve has a response time of 20 ms. A carbonyl trap is installed after the CO bottles to remove iron carbonyl. The reactor is a U-shaped quartz microreactor with an inner diameter of 4 mm. The dead volume of the reactor is minimized by filling the voids with quartz wool and a quartz rod with a diameter of 3 mm. The temperature is controlled with an Eurotherm temperature controller. The reactants and the products were analyzed with an online gas chromatograph (GC) equipped with TCD as well as FID. The transient responses were monitored with a quadrupole mass spectrometer (MS). The experimental error is typically in the range of 2–5%.

41.8

Combination of SSITKA with Spectroscopic Methods

41.8.1 SSITKA-FTIR SSITKA is a very powerful technique to determine the amount and reactivity of surface intermediates, but unable to determine the composition/nature of intermediate species. However, coupling of SSITKA with spectroscopic techniques

a) 4.0

b) 1.0 H2 D2

3.5

0.8

0.6

3.0 rH /rD

rCO*10–3 (mol/gcat/h)

Fig. 41.10 (a) Reaction rate of CO at different H2 or D2 partial pressures; (b) an example of inverse kinetic isotope effect (Reproduced from Ref. [139]. Copyright © 2013 Springer Science Business Media New York)

experiments to differentiate reaction pathways. Details can be found in the cited papers. An example will be given later.

2.5

0.4

2.0

0.2

1.5 120

160 200 PH2 or PD2 (psi)

240

0.0 80

120

160 200 PH2 or PD2 (psi)

240

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

945

Vent

41

1

2

1

MFC

Feed

PC 1

12CO

PI

2

MFC

Unlabeled

Ar

Switch 3

MFC

PC 2 Reaction

GC TCD/FID

4

H2

MFC PI

5

MFC

Labeled

MS QMG 422

Kr 6

MFC 13CO

7

MFC Reactor Furnace

Air

Unlabeled Inject Vent

Reactor

Labeled

Vent

Unlabeled Load Reactor

1

2

ºC

Labeled

Fig. 41.11 A standard experimental setup for carrying out SSITKA studies at or close to atmospheric pressure

could allow for the direct observation of adsorbed species under reaction conditions. Infrared spectroscopy is very sensitive to detect surface species based on vibration frequencies. The main advantage of combining SSITKA and FTIR technique is the ability to differentiate active intermediates from species that do not participate in the reaction, so-called “spectators.” The discrimination can be achieved by comparing the isotopic exchange rate of surface species measured by IR to the exchange rate of the species leading to gas-phase products measured by MS. Only surface species with the same/higher exchange rate can contribute to the main reaction pathway. Minor reaction intermediates and spectator species typically have slower isotopic exchange rates or no exchange, respectively [144]. The main challenge for combining SSITKA with spectroscopic techniques such as FTIR is the construction of a proper reaction cell that allows the acquisition of both the kinetic and spectroscopy data. FTIR can be performed both in transmission mode and reflectance mode. Earlier studies used the transmission mode, which allows quantitative measurements of all samples. However, the sample has to be pressed

into a wafer, which can cause diffusion limitation in the wafer. A DRIFTS reactor is particularly suitable to carry out catalytic gas-solid heterogeneous reactions since the catalyst powder can be used as it is, therefore it is becoming the major technique. Commercial DRIFTS cells are available for performing operando DRIFTS experiments combined with SSITKA at high temperature and high pressure. Care has to be taken to ensure that representative IR spectra are measured in order to correlate to kinetic data since the IR data are measured in the catalyst bed while the kinetic data are measured at the effluent of the reactor. Sometimes a modification of the commercial cell is needed in order to optimize the flow pattern and temperature control in the cell [145]. It is always wise to calibrate the activity obtained in the IR cell against normal fixed-bed reactor results to ensure a similar reaction environment. In order to measure the exchange rate of surface species, one has to trace active surface species with IR. When the gas molecule is replaced with a heavier isotope, it is expected to have a red shift in IR vibration frequency, as shown in Fig. 41.12, and the intensity change during the isotope

946

A. Holmen et al. Table 41.3 Reaction systems studied by SSITKA-FTIR 0.5 wt% Pt/Ce0.8La0.2O2-G T = 250 °C

(3)

Absorbance

(1)

(2) (3*)

12CO/H O 2 13CO/H O 2

3050

0.02

3000

(1*)

2950

2900

Wavenumber

(2*)

2850

Reaction system CO/CO2 hydrogenation Oxidative catalysis Water–gas shift (WGS) Selective catalytic reduction Methane reforming Ethylene hydroformylation n-butane isomerization Hydrogenation of nitrates in water

References [11, 58, 59, 119, 153–161] [103, 162–168] [68, 146, 169–187] [188–197] [198–200] [201–205] [206] [207]

41.8.2 SSITKA-Neutron Scattering 2800

2750

(cm–1)

Fig. 41.12 Formate species have a red isotope shift due to the switch from 12CO/H2O to 13CO/H2O in water–gas shift reaction (Reproduced from Ref. [146]. Copyright © 2013 Elsevier B.V.)

exchange can be measured. Care has to be taken when quantifying the active reaction intermediates with DRIFTS [147]: (1) the surface concentrations of active intermediates should be within the detection limit of the IR instrument; (2) appropriate procedures for background subtraction and baseline correction have to be followed [148]; (3) the theoretically expected red isotopic shift of the isotope labelled intermediates [149] should be larger than the resolution of the FTIR; (4) the red isotopic shift is only due to isotope exchange and not to any other surface chemical reactions; and (5) appropriate deconvolution procedures should be applied [149] for overlapping IR bands which might come from active and inactive (spectator) intermediate species. The combination of SSITKA-FTIR has been applied to various catalytic reactions. Following the pioneering work of Dalziel et al. [150] and Tamaru et al. [151, 152], the combination of SSITKA with spectroscopic techniques has become an important part of the use of SSITKA. A list of reaction systems studied by SSITKA-FTIR is presented in Table 41.3. A comprehensive discussion for the most used systems can be found in the review article by Ledesma et al. [6]. Recent studies have been on CO2 hydrogenation [58, 119], WGS [169], etc. Efstathiou [169] has outlined in a great detail the application of the SSITKA-operando (MS-FTIR) methodology on the water–gas shift reaction on Pt catalysts. Efstathiou [169] also includes DFT calculations and experimental studies in the discussion of advantages and limitations. Other recent studies on the combination of SSITKA and FTIR include the study of Zabilskiy et al. [119] on CO2 hydrogenation to methanol using Cu/ZnO catalysts. The study of Zabilskiy et al. [119] is also an example of combining a large number of techniques.

A high-precision gas flow cell has been developed for performing time-of-flight neutron scattering at both constant and pulsed gas flow. The neutron scattering can provide detailed structural information of solid materials. The combination of neutron scattering and SSITKA experiments has been demonstrated for N2 adsorption on Ca-exchanged zeolite-X. The detailed design of the gas cell and reaction system can be found in the relevant reference [125]. The combined system has been applied to study gas-sorbent interaction under operational conditions, i.e., N2 adsorption on Ca-exchanged zeolite-X at 1 atm and 300 K. By performing in situ neutron scattering and SSITKA simultaneously, the crystal structure of the sorbent can be linked to the macroscopic kinetics involved in the gas adsorption process [116]. The neutron scattering with static N2 gas gives lattice information relative to the vacuum condition. The isotope exchange allows for the identification of structures involved in the gas adsorption process under steady-state conditions through isotopic contrasted neutron signals [116]. The technique is useful to probe the kinetics of detailed surface structure species in steady-state processes through the switch of isotopic labelled gases. For example, utilizing the contrast of 14N versus 15N in neutron scattering to study processes involving N atoms, such as NH3 synthesis or NO reduction processes, or utilizing the contrast of 12C versus 13 C to study reactions such as CH4 oxidation, CO oxidation, propene epoxidation, or CO2 reduction/fixation.

41.9

Other Considerations and Recent Developments

SSITKA can and has been used at high pressures, a work that has been pioneered by Bertole, Mims, and Kiss [1]. However, due to isotope costs, the pressure and gas flows are usually restricted. It is also preferred to keep the product composition as simple as possible. Multicomponent SSITKA is used to study the kinetics of complex reactions such as hydrogenation of CO [8,

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

947

208, 209]. Different samples of the multiple products formed during CO hydrogenation are collected, separated by gas chromatography, and then the C2+ products can be converted to CH4 by hydrogenation in a separate reactor before MS measurements [209] or GC-MS can be used to analyze C2+ directly. With the development of advanced instrumentation, multicomponent analysis can be performed using GC-MS without additional hydrogenation reactor [8, 208]. Basically, GC will separate C2–C5 hydrocarbons, and the isotopic signal for each hydrocarbon can be detected by the subsequent MS. The MS signal is used to calculate isotopic composition of isotopologues of each hydrocarbon. The observed MS signal is a linear combination of the fragmentation patterns of isotopologues. A detailed example will be given later. The main advantage of the method is that it is possible to construct the isotopic transients for the various products and determine their surface parameters without overlapping the signals of the different products due to the MS fragmentation. However, due to the sensitivity of GC and overlap of fragmentation of hydrocarbons in the MS, this approach is usually limited to analyze products up to C5. Recently, in order to deal with complex reaction networks, especially those

A o [θ1] o B

II)

A o [θ1] o [θ2] o B

III) A

o

IV )

A o [θ1] o B l

I)

[θ2]

[θ1] o B [θ2]

V)

A o [θ1] o [θ3] o B

o

l

o

[θ2]

Fig. 41.13 Different reaction pathways from A to B (Reproduced from Ref. [210]. Copyright © 2009 Elsevier B.V.)

41.10 Applications SSITKA has been used to study many catalytic systems. The straightforward information is the mean residence time and the number of surface intermediates, as well as intrinsic kinetic rate constants. This information is very useful to discriminate the contribution of quantity or reactivity change in the active site in terms of the effect of the support, promoters, alloys, and the deactivation mechanism.

41.10.1 Discriminate Between Different Reaction Mechanisms In addition to obtaining information on the number and reactivity of active sites under reaction conditions, SSITKA can also be used to discriminate between different reaction mechanisms, as shown by Angelos Efstathiou and Xenophon Verykios [104] and Bal’zhinimaev et al. [210]. By modeling the transient curve, it is possible to discriminate between various reaction mechanisms towards product formation. Take the simplest reaction A ! B through one or two intermediates (θi) as an example. There are several pathways to form B from A as shown in Fig. 41.13. The reaction pathways differ in the number of intermediates, one pool or two pools, and in the way the pools are connected, parallel or consecutive. The shape of the isotopic transient curve differs due to the difference in the relaxation time of the product B detected by MS, as indicated by Fig. 41.14.

a) 1.0

b) 0 III, IV

ZB

Fig. 41.14 Time dependencies of the linear (a) and logarithmic (b) isotope fractions in the reaction product for different types of reaction mechanisms (Reproduced from Ref. [210]. Copyright © 2009 Elsevier B.V.)

involving isotope labelled gases, a computer code has been developed by E. Rebmann et al. [14] to automatically generate reaction networks. This approach makes the modeling of both steady-state and transient responses of SSITKA easier [14].

In (1 – ZB)

41

0.5 I

III, IV –1 I V

–2

II

0.0

II

–3 0

10 Time (s)

20

0

10 Time (s)

20

41

948

A. Holmen et al.

At steady state, the rate of elementary steps for isotope transfer also has constant rates and can be considered as a first-order reaction with respect to the isotope fraction. Therefore, for a single mechanism (Scheme I), the relaxation curve is a single exponential. For a reaction with consecutive pools (Scheme II), a different exponential is obtained. For Schemes III and IV, the relaxation curve is a sum of two exponentials. It is easier to distinguish features from different reaction mechanisms on a logarithmic scale, as shown in Fig. 41.14b. Scheme I is a straight line; Scheme II will convex upward; Schemes III and IV will convex downward; while Scheme V will have an S-shaped isotope response. The detailed modelling can be found in the paper of B.S. Bal’zhinimaev et al. [210]. SSITKA transient modeling has been used by several groups to identify and discriminate between different mechanistic models for the Fischer-Tropsch synthesis for both cobalt- and iron-based catalysts [211]. For mechanisms involving reversible interaction of reactants or products with catalyst surfaces, a plug flow reactor model should be used. According to Van Dijk et al. [212], the general continuity equations for the labeled gaseous component X′ are represented by Eq. (41.35) and for the labeled surface component Y′ by Eq. (41.36).

Using CO hydrogenation as an example, modeling of CO transient curve provides information on the kinetics of CO adsorption, namely CO adsorption, desorption rate constant, and the conversion of surface-adsorbed CO. Modeling of the CH4 transient curve helps to differentiate different reaction pathways for methane formation and provide the kinetic rate constants involved. An example of a model result is shown in Fig. 41.15. Based on the model fit, different methane formation pathways can be discriminated, as shown in Fig. 41.16. It was found that model 5, two parallel carbon pools Cα and Cβ were involved in methane formation. SSITKA transient modeling can be used to estimate rate coefficients for chain initiation, propagation, and termination by fitting theoretically generated model curves to the observed transient responses. The rate constants (ka, kb, kc, kd, ke, kf, and kg) involved in the reaction step can be estimated from modeling results. Transient modeling has been applied to study C2+ reaction mechanisms for CO hydrogenation [208]. The prerequisite is that possible reaction pathways have to be proposed first. Once the reaction pathways are identified, the kinetic parameters along the reaction pathways can be extracted as in methane formation example.

41.11 Examples of the Use of SSITKA @CX0 1 @CX0 ρb þ ¼ Rw,X0 τ @x @t εb

ð41:35Þ

@LY 0 ¼ Rw,Y0 @t

ð41:36Þ

The initial condition for labeled compound based on isotopic switch from R to R′:

The catalysts shown in Table 41.4 have been characterized under methanation conditions (483 K, 1.85 bar, H2/CO ¼ 10) by SSITKA [213].

t ¼ 0, CX0 ¼ CY0 ¼ 0 t > 0, x ¼ 0, Creactant0 ¼ input function (t) t > 0, x ¼ 0, Cproduct0 ¼ 0

1.2 1

3

where CX′ is gas phase concentration mol/m gas; LY′ is surface concentration mol/kgcat; τ is the residence time in s; t is time in s; εb is the porosity of the bed m3 gas/m3 bed; ρb is the density of the catalysts bed in kg/m3 bed; Rw0 is the steady-state reaction rate of different species in mol/kgcat; and x is a dimensionless axial position in the ε catalyst bed. The residence time τ is calculated as τ ¼ b Fv , VR

where VR is the volume of the catalyst bed in m3 and Fv is the total volumetric flow rate in m3/s at reaction temperature and pressure. The input function is represented by the transient of the inert tracer Ar, corrected for the residence time.

Normalized response F(t)

Initial conditions Boundary conditions

41.11.1 CO Hydrogenation on Al2O3-Supported Co Catalysts

0.8

Ar 13CO 13CO calculated 13CH 4 13CH calculated 4

0.6 0.4 0.2 0

–0.2

0

10

20

30 Time (s)

40

50

60

Fig. 41.15 Example of simulated transient curve for 13CO and 13CH4 (Reproduced from Ref. [6]. Copyright © 2014 American Chemical Society)

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

949

Model 1

Model 2

COg

kd*

ka*

kd*

ka*

kc*

41 kb*

CO*

CO*

ke*

C1*

C1*

k d*

ka* kf*

Cα*

CH4,g

COg

ke*

kd*

CO*

kb*

kg* Cβ* Model 5

Cα*

kf* ke* k g*

kc*

Cβ*

C1+*

C1+* Cα*/Cβ*

Cα*/Cβ*

Model 6 COg

CH4,g

COg

CH4,g

ke*

kd* CO*

C1+* Cα*/Cβ*

CH4,g

kc*

ka*

kg* Cβ*

Model 4

COg

kb*

ke*

C1+*

Model 3

CO*

Cα*

kf*

kc*

kb*

ka*

CH4,g

COg

CH4,g

kb*

Cα*

kd*

ka*

kf*

Cβ*

kc*

kg*

kf*

ke*

CO* kc*

Cβ*

k b*

Cα*/Cβ*

kg* Cα*

C1+* Cα*/Cβ*

Fig. 41.16 Possible reaction routes for methane formation considered in SSITKA transient modeling. (Reproduced from Ref. [212]. Copyright © 2003 Plenum Publishing Corporation) Table 41.4 The supported Co catalysts involved in the study with characterization results Catalyst 12Co/γ-Al2O3 12Co0.5Re/γ-Al2O3 20Co0.5Re/γ-Al2O3 20Co0.5Re/α-Al2O3 12Co/TiO2f 12Co0.5Re/TiO2f 12Co/SiO2

Surface areaa (m2/g) 161 155 150 23 8 12 297

Dispersionb (%) 6.6 10.2 7.8 3.7 2.3 2.4 5.3

vads COd (μmol/gcat) 39 87 124 61 31 22 37

vads H2c (μmol/gcat) 67 104 132 63 23 24 54

vads COe (μmol/gcat) 40 76 117 69 29 22 39

Reprinted with permission from Ref. [213]. Copyright © 2005 Elsevier B.V. BET surface area calculated from N2 adsorption measurements b Cobalt metal dispersion from H2 chemisorption at 313 K, assuming adsorption on Co atoms only and H2:Co ¼ 1:2 c Amount of H2 chemisorbed at 313 K d Amount of CO adsorbed at 373 K, calculated from SSITKA (S1), CO/inert ¼ 1.5/33.5 Nml/min e Amount of CO adsorbed at 373 K, calculated from SSITKA (S2), H2/CO/inert ¼ 15/1.5/33.5 Nml/min f Support pretreated at 973 K for 10 h a

It is assumed a first-order reaction with no readsorption. The rate constant is given by Eqs. (41.19) and (41.20):

k¼τ

1

¼ TOF  θ

1

ð41:37Þ

The surface residence time is given by the area under the normalized transient curve:

950

A. Holmen et al.

τi ¼

ð1 0

Fi ðtÞ  dt

ð41:38Þ

As shown in Fig. 41.1, the surface residence times are corrected for gas phase hold-up by using Ar and Kr, respectively, as inert during the isotope switching. The surface residence time of intermediates leading to methane is corrected for the chromatographic effect of CO as follows: τCH4,corrected ¼ τCH4,measured  0:5  τCO

ð41:39Þ

The number of adsorbed species is calculated from the mean residence time and the exit flow (EFi) of species i (i ¼ CH4, CO): N i ¼ τi  EFi

ð41:40Þ

The turnover frequency (TOF) in Table 41.5 is calculated based on the total number of active sites measured by H2 chemisorption at 313 K. The procedure for the SSITKA experiments involved 100 mg catalyst mixed with 200 mg of SiC with the same particle size fraction (53–90 μm). The catalysts were reduced in flowing H2 (10 ml/min) for 16 h at 623 K using a heating rate of 1 K/min from ambient temperature to 623 K followed by cooling to 373 K. The feed was then switched to 12CO (1.5 Nm/min) and Ar (33.5 Nml/min) at 1.85 bar. When no H2 was detected, the feed was switched to 13CO and Kr without changing the flow rates in order to obtain the amount

of reversible adsorbed CO on the catalysts. When steady state was obtained (after about 3 min), the feed was switched back to the 12CO/Ar feed, and after 5 min, H2 (15 Nml/min) was added to the feed. After a short period, the feed was switched from H2/12CO/ Ar to H2/13CO/Kr in order to obtain the amount of reversible adsorbed CO when H2 is present. The feed was switched back to H2/12CO/Ar after about 3 min, and the temperature was increased at 5 K/min to 443 K and finally by 1 K/min to 483 K. After having obtained steady state after about 5 h, the final switch to H2/13CO/Kr was performed to obtain steadystate reaction kinetic parameters at actual temperature and pressure. The SSITKA results are given in Table 41.5. Table 41.5 shows that the increase in activity (RCO) is accompanied by a similar increase in NCH4 (surface concentration of active intermediates leading to CH4). It follows from this that τCH4 is approximately constant and independent of Co loading, promoter, or support. Table 41.5 also shows that a value for τCH4 (the mean surface residence time of intermediates leading to methane) of 11 s was obtained. α-Al2O3 has a much lower surface area compared with γ-Al2O3 (Table 41.4) and the overall reaction rate is also much lower. However, TOFCH4 and the coverage of methane precursors are almost unchanged, indicating that the change in the reaction rate is due to a change in the number of active sites on the surface. The increase in active sites can be explained by the more dispersed metal and consequently smaller Co particles on the high surface alumina.

Table 41.5 SSITKA results obtained at 483 K, 1.85 bar, and H2/CO/inert ¼ 15/1.5/33.5 Nml/mina Catalyst 12Co/γ-Al2O3 12Co0.5Re/γ-Al2O3 20Co0.5Re/γ-Al2O3 20Co0.5Re/α-Al2O3 12Co/TiO2

c

12Co0.5Re/Ti2O2 12Co/SiO2 12Co0.5Re/SiO2

c

12Co/γ-Al2O3 12Co0.5Re/γ-Al2O3 20Co0.5Re/γ-Al2O3 20Co0.5Re/α-Al2O3 12Co/TiO2

c

12Co0.5Re/TiO2 12Co/SiO2 12Co0.5Re/SiO2

c

TOFCH4 (103 s1) b

RCO (μmol/(gcat s)) 1.6 2.7 4.1 1.6 1.0

7.1 8.6 11.9 10.2 13.9

1.0

15.4

1.0 1.3

7.5 8.7

τCH4 (s) 10 11 9 12 12

τCO (s) 3 8 11 7 2

NCH4 (μmol/gcat) 9 20 30 15 8

SCH4 (%) 59 67 76 80 67

SC2 (%) 9 9 9 8 9

SC3+ (%) 32 24 15 12 24

74

9

18

78 78

9 9

13 13

NCO (μmol/gcat) 32 67 79 63 22

b

b

θCH4

θCO

0.07 0.10 0.11 0.12 0.16

0.24 0.32 0.3 0.5 0.47

11

2

8

20

0.16

0.41

13 11

3 4

10 11

31 38

0.10 0.10

0.29 0.32

Reprinted with permission from Ref. [213]. Copyright © 2005 Elsevier B.V. CH(4) indicates surface intermediates leading to CH4. τi: residence time of surface intermediates, Ni: concentration of surface intermediates, and θi: coverage of surface intermediates (i ¼ CO, CH(4)) b Based on H2 chemisorption, assuming adsorption only on Co-atoms c Support pretreated at 973 K for 10 h a

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

41.11.2 The Use of Multicomponent SSITKA to Obtain Kinetic Parameters for Higher Hydrocarbons in CO Hydrogenation [8] The residence time of higher hydrocarbon is determined by analyzing the residence time of non-labelled, partially labelled, and fully labelled species with GC-MS. The fragmentation pattern of each hydrocarbon is considered as a linear combination of fragmentation patterns of all isotopologues (non-labelled, partially labelled, and fully labelled) as described by Schouten J.C. and coworkers [16]. The contribution of each isotopologue to the final MS pattern can be calculated by minimizing the objective function, as shown in Eq. (41.41):

G ðyÞ ¼

m n X X i¼1

!2 ref yij f ij



obs fi

ð41:41Þ

j

where G( y) is the objective function, m is the total number of different m/e values for each hydrocarbon products, n is the total number of isotopologues for each hydrocarbon, yij is the contribution of a given isotopologue to the mixture at its m/e value, f ref ij is the intensity of the fragmentation pattern of a given isotopologue at its m/e value, and f obs i is the intensity of the fragmentation pattern observed in the GC-MS for each m/e value. Once the contribution of isotopologues is calculated, the normalized responses can be obtained over time on stream as shown in Fig. 41.17 as an example. Using ethane as an example, based on the transience curve, the residence time of non-labelled, partially labelled, and fully labelled isotopologues (e.g., 12C2H6, 12C13CH6, 13C2H6) can be determined. The overall residence time of ethane can be calculated by the overall 12C content of ethane, according to Eq. (41.42): τ C2 ¼

 1 1  τ12 C1 13 C1 H6 þ 2  τ12 C2 H6 2

ð41:42Þ

A generic equation of hydrocarbon with carbon number n can be derived as Eq. (41.43):

τ Cn ¼

n 1X iτ12 Ci 13 Cni n i¼1

ð41:43Þ

Thereafter, the total number of intermediates Ni leading to the higher hydrocarbon products can be calculated similarly to CH4 as described by Eq. (41.38). The site coverage of intermediates leading to different products can therefore be calculated by dividing Ni with the total number of active site Ns as typically determined by H2 chemisorption. The kinetic data is very useful for analyzing the contribution of site coverage of intermediates and site reactivity to the overall

951

turnover frequency. It was found that the site reactivity towards C2–C5 olefin and paraffin formation are similar, and the TOF differences are attributed to the change in site coverage of surface intermediates.

41.11.3 Surface Species and Mechanistic Studies by Combination of SSITKA and Kinetic Isotope Effect Supported Co catalysts with different particle size were studied by isothermal hydrogenation (IH), temperatureprogrammed hydrogenation (TPH), and SSITKA. Kinetic isotope effect experiments were used to probe possible mechanisms on Co/γ-Al2O3 with particle size in the range of 4–15 nm. The performance of the catalysts in FischerTropsch synthesis (FTS) depends on properties such as metal loading, particle size, and support. Decreased catalytic activity and decreased selectivity towards higher hydrocarbons are found on highly dispersed Co catalysts. In the present work, IH and TPH are used to distinguish different types of surface species and their distributions for Co catalysts at working conditions. SSITKA gives the intrinsic activity and the number of active intermediates in situ, and H2–D2 experiments are carried out to study the dependence of the rate-determining step on Co particle size. This study is also an example of the use of other isotopes than 12CO/13CO and that several different experiments can be carried out in the same experimental setup. The main purpose of using SSITKA in this case was to study the dependence of the intrinsic activity on the cobalt particle size. The SSITKA experiments were carried out as a switch between Ar/12CO/H2 and Kr/13CO/H2 at 210  C, 1.85 bar, and H2/CO ¼ 10. The amount of CO and surface intermediates leading to methane and the intrinsic reactivity toward methane formation were measured and summarized in Table 41.6. It must be emphasized that the TOF reported in the Table 41.6 is based on the steady-state reaction rate and number of active sites determined by ex-situ H2 chemisorption. The actual number of active sites may change during the reaction. In situ CO adsorption measured by SSITKA at 100  C may be a better approximation to the total number of active sites than ex situ H2 adsorption. Dispersion based on in situ CO adsorption has a better correlation with characteristic surface parameters on a large variety of cobalt catalysts, as shown in Fig. 41.18. Another alternative is to use surface intermediates (NCHx) as the total number of active sites. The intrinsic activity k is then the inverse of surface residence time of surface intermediates for a pseudo-first-order reaction (1/τ) and also called intrinsic turnover frequency (TOFITK). Figure 41.19 clearly shows that TOF increases with increasing particle size, while the intrinsic rate constant/activity/TOFITK remains the same

41

952

A. Holmen et al.

Fig. 41.17 Isotopic response of C2H6 (a), C3H8 (b), C3H6 (c), C4H10 (d), C4H8 (e), C5H12 (f), and C5H10 (g) as a function time on stream at 483 K, Ptot ¼ 1.85 bar, PCO ¼ 0.17 bar, and PH2 ¼ 0.56 bar (Reprinted with permission from Ref. [8]. Copyright © 2016 American Chemical Society)

a) 1.2 Normalized F(t)

1.0 0.8 0.6

12C

13C

2H6

0.4

2H6

12C13CH 6

0.2 0.0

−0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

1.0

1.0

0.8 0.6 0.4

12C H 3

12C 13CH 2

0.2

13C H 3

12C13C

2H8

Normalized F(t)

c) 1.2

Normalized F(t)

b) 1.2

0.0

0.8 0.6 0.4

12C H 4 10

13C H 4 10

0.6 12C13C H 3 10

0.4 12C 13CH 3 10

12C 13C H 2 2 10

Normalized F(t)

1.0

Normalized F(t)

1.0

0.0

0.8

0.4 0.2

12C

4H8

13C

4H8

12C13C H 3 8 12C 13CH 3 8

12C 13C H 2 2 8

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

1.0

1.0 12C H 5 12

13C

5H12

0.6 12C 13C H 3 2 12 12C 13CH 4 12

12C13C H 4 12 12C 13C H 2 3 12

0.0

0.8 0.6 0.4 0.2

12C

5H10

12C 13C H 3 2 10 12C 13CH 4 10

13C 12C13C

5H10

4H10

12C 13C H 2 3 10

0.0

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

for catalysts with different cobalt particle sizes. The increase of TOF with increasing particle size is ascribed to higher number of active sites on large particles instead of higher intrinsic activity (k ¼ 1/τ). The reason that no leveling off of TOF was found for larger particle sizes (>10 nm) as have been reported previously is probably due to the larger pore size of LP Al2O3.

Normalized F(t)

g) 1.2

Normalized F(t)

2H6

0.6

f) 1.2

0.2

12C13C

3H6

0.0

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

0.4

13C

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

e) 1.2

0.8

12C 13CH 2 6

0.2

d) 1.2

0.2

3H6

0.0

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

0.8

12C

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Time (min)

41.11.4 Combing DFT and Transient and Steady-State Modeling to Study Reaction Mechanism of CO Hydrogenation A new approach combining DFT, transient, and steady-state kinetic modeling to elucidate the CO activation mechanism

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

953

Table 41.6 SSITKA on Co/γ-Al2O3 catalyst with different cobalt particle sizes

Catalyst samples MP_4 MP_8 MP_11 LP_15

Rate of CO conversion [μmol/(gcat s)] 1.2 2.4 2.5 2.6

Rate of CH4 formation [μmol/(gcat s)] 0.7 1.5 1.7 1.8

NCO [μmol/ gcat]a 56 53 63 60

τ12CO (s)b 5.7 6 7.2 7

NCHx [μmol/ gcat]c 6 12 13 15

τ12CHx (s)b 8.2 8 7.6 8.5

k (s1)d 0.12 0.13 0.13 0.12

TOF (103 s1)e 3.3 10 12.1 14.8

Reproduced from Ref. [48]. Copyright © 2010 American Chemical Society Surface number of reversibly adsorbed CO b Mean surface residence time c Number of surface intermediates leading to methane d Reaction rate constant for methanation assuming pseudo-first-order reaction (k ¼ 1/τCHx) e TOF is calculated based on the CO reaction rate and H2 chemisorption results a

300 J-Al2O3 TiO2 SiO2 D-Al2O3 Re D-Al2O3 Re D-Al2O3 (modified) P-CNF FB-CNF

NCHx (Pmol/gCo)

250 200 150

a)

b)

c)

d)

100 50 0 1000

NCO (Pmol/gCo)

800 600 400 200 0 0

200

400

600

800

1000

1200 0

NH2 (Pmol/gCo)

200

400

600

800

1000

1200

1400

NCO (Pmol/gCo)

Co/TiO2;◆, Co/SiO2; ⋆, Co/α-Al2O3; ●, CoRe/α-Al2O3; ○, CoRe/α-Al2O3 (modified); □, Co/P-CNF; and ■, Co/FB-CNF (Reproduced from Ref. [214]. Copyright © 2015 Elsevier B.V.)

Fig. 41.18 The number of surface intermediates leading to methane and number of reversibly adsorbed CO (1.85 bar 483 K H2/CO/Ar ¼ 15/ 1.5/33.5 Nml/min) as a function of the amount of H2 adsorption (a, c) and in situ CO adsorption at 373 K (b, d), respectively. △, Co/γ-Al2O3;

~,

has been demonstrated for CO hydrogenation reaction. First, conventional steady-state Langmuir-Hinshelwood models were established based on the assumption of different reaction pathways and rate-determining steps. The main

reaction pathways can be divided into direct CO dissociation pathways with steps 3, 4, and 5 as rate-determining steps and H-assisted CO dissociation with steps 3, 4, and 5 as RDSs, respectively, as shown in the scheme below.

41

954

A. Holmen et al.

Initial screening was made based on the reaction order with respect to CO and H2. All other mechanisms apart from reaction mechanism 1.2, 2.2, and 2.3 are excluded. The TOF for CO conversion was simulated by LangmuirHinshelwood models at various reaction conditions. The model does not fit very well especially for data at different CO partial pressure. It was observed that there is a need to correct CO equilibrium constant at different CO partial pressures to take into account lateral interaction of co-adsorbed CO on the surface. Mechanism 1. Direct dissociation followed by stepwise hydrogenation H2 þ 2*$ 2H* CO þ *$ CO* CO* þ * $ C* þ O * C* þ H* $ CH* þ * CH* þ H* ! CH2* þ * . . .. . . O* þ H* $ OH* þ * OH* þ H* $ H2O þ 2*

Case 1.1 Case 1.2 Case 1.3

[1] [2] [3] [4] [5] [6] [7]

Mechanism 2. Hydrogen-assisted CO dissociation H2 þ 2*$ 2H* CO þ *$ CO* CO* þ H* $ HCO* þ * Or CO* þ H* $ COH* þ * HCO* þ H* ! HCOH* þ * Or COH* þ H* ! HCOH* þ * HCOH* þ * ! CH* þ OH*

Case 2.1 Case 2.2 Case 2.3

[1′] [2′] [3′] [300 ] [4′] [400 ] [5′]

SSITKA studies can directly measure the site coverage and residence time of surface-adsorbed CO species. The

16 TOF (*103 s–1)

0.12

12

0.10

10

0.08

8

0.06

6

0.04

4

0.02

2 2

4

6

8 10 12 Particle size (nm)

14

k = 1/tCHx (s–1)

0.14

14

0 16

Fig. 41.19 TOF and pseudo-first-order rate constant k (1/τCH4) (Reprinted with permission from Ref. [48]. Copyright © 2010 American Chemical Society)

transient response of CO contains information on the adsorption and desorption kinetic information. The transient response of CO was then modeled in a plug flow reactor model to extract kinetic parameter with respect to CO adsorption and desorption. The model adequately simulates the CO transient curve (Fig. 41.20) and the pseudo-adsorption rate constant ka (containing contribution from site vacancy) was obtained from model fitting. However, SSITKA cannot measure the site coverage of H2 due to the presence of KIE and possible exchange with OH groups on the alumina surface. The combination of the transient modeling and conventional Langmuir-Hinshelwood steady-state modeling can predict very well the experimental TOF by taking into account the effect of CO coverage on the CO equilibrium (KCO), Fig. 41.21. The models also allow the determination of H2 equilibrium constant and site coverage. The surface mapping of adsorbed species can be obtained at different reaction conditions as shown in Fig. 41.22. Based on transient and steady-state modeling, two models fit equally well the experimental data. It seems difficult to exclusively elucidate the reaction mechanism based solely on kinetic modeling. To further distinguish reaction mechanism, e.g., direct CO dissociation with C þ H ! CH+ * as RDS and hydrogen-assisted CO dissociation with either the hydrogenation of HCO* species or decomposition of the HCOH* species as RDSs, density functional theory (DFT) was used. DFT calculations can give energetic information of the reaction, such as activation energy, and it can give kinetic and equilibrium isotopic effects of the elementary steps through calculating the vibration frequencies of the ISs (initial states), TSs (transition states), and the FSs (final states). The overall KIE can be obtained by summing up all the elementary steps. The DFT calculated KIE for four possible kinetic routes in H-assisted CO dissociation are shown in Table 41.7. The experimental KIE can be determined by measuring the rate at CO/H2 and CO/D2. By comparing experimental KIE and calculated KIE, hydrogen-assisted CO dissociation with HCO* þ H* ! HCOH* as the kinetically relevant step could be a favorable pathway for CO activation.

41.12 Examples of Combining SSITKA-DRIFT for WGS SSITKA-DRIFTS has been used to elucidate the reaction mechanism for water–gas shift reaction (WGS) [147]. The reaction mechanism is heavily debated as to whether it goes through “Redox mechanism” or “Associated mechanism/ formate” or “Carbonate mechanism.” Jacobs and Davis [176] investigated low-temperature WGS mechanisms on Pt/Ceria and related catalysts. SSITKA-DRIFTS has been

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

955

f(t) -13CO

a) 1.2

b)

1.2

1.0

1.0

0.8

0.8

0.6 PCO = 0.037 0.4

0.0555

41

0.6 PH2 = 0.555 0.4

0.4625 0.37

0.111

0.2

f(t) -13CO

41

0.2 0.2775

0.1665 0.0

0.0

0.1665

−0.2

−0.2 0

2

4

6

8

10

12

14

Time (s)

2

4

6

8

10

12

Time (s)

Fig. 41.20 Modeling of transient response of 13CO upon a switch form 12 CO/H2/Ar to 13CO/H2/Kr at different (a) CO pressure (PH2 ¼ 0.555 bar) and different (b) H2 pressure (PCO ¼ 0.0555 bar) at 483 K

and 1.85 bar. Points are experimental results; lines are model results. f(t) is the normalized transient response (Reproduced from Ref. [141]. Copyright © 2013 Elsevier Inc.)

shown useful to identify the role of Pt-CO and formate as a potential intermediate in the catalytic mechanism. The experiments were performed in a Nicolet Nexus 870 spectrometer equipped with a DTGS-TEC detector. An in situ cell fitted with ZnSe windows was used as the reactor for the operando measurements at close to atmospheric pressure. Scans were taken at a resolution of 4 to give a data spacing of 1.928 cm1. For steady-state initial (i.e., only 12CO) and final (i.e., only 13CO) conditions, 256 scans were taken to improve the signal-to-noise ratio. During the transient isotopic switch experiment (i.e., where 12CO was switched to 13 CO) under steady-state water–gas shift conditions, however, 32 scans were taken between points to provide an adequate level of signal to noise but at the same time allow the gathering of enough data points. The point representing the time where 50% of fractional isotopic exchange in the CO2 product was identified. FTIR was used to quantify the fractional isotopic exchange of 13C in CO2 instead of using MS to quantify CO2 in the effluent. The feed consists of 3.75 ccm CO, 62.5 ccm H2O, and 67.5 ccm H2 [176]. Figure 41.23a shows a typically DRIFTS signal during an isotopic switch. Figure 41.23b presents the spectrum before the switch, the spectrum at approximately the time for 50% exchange in formate/CO2, and at a point close to complete reactive exchange. It was found that at the time of 50%

fraction isotopic exchange in CO2, the v(CO) bands for CO adsorbed on Pt have virtually been completely exchanged. This means that the exchange rate of product CO2 is faster than the exchange rate of Pt-CO species. This result suggests that the “Redox” mechanism, where adsorbed CO directly reacts with O adatoms to form gas phase CO2, is less likely. The coincident time to achieve 50% fractional isotopic exchange for both CO2 and formate indicates that CO2 can be produced from formate decomposition. The authors also measured the half-life when the feed was switched to CO, D2O, and D2. A similar half-life increase was observed for both formate and adsorbed CO2, which further supports a formate mechanism. Combining with the normal isotope effect, the authors concluded that C-H bond breaking of the adsorbed formate species is the rate-limiting step for CO conversion during WGS. Meunier et al. developed a quantitative measurement of DRIFTS signals for adsorbed species that enables the comparison of fractional isotope exchange as a function of time for both surface intermediates and products CO2, as shown in Fig. 41.24 [177, 178]. The analysis of the formate exchange curves at 493 K showed two formate pools. The “Fast formates” showed a similar exchange rate as with that of CO2. The “Slow formates” displayed an exchange rate constant 10to 20-fold slower than that of the reaction product CO2. In addition, the concentration (mol/g) of formates on the surface

956

A. Holmen et al.

was determined based on calibration against formate standards and a specific rate of formate decomposition was calculated. It was reported that the rate of CO2 formation was more than an order of magnitude higher than the rate of

a)

decomposition of formates (slow + fast species). Based on that, the authors concluded that not all of the formates detected by DRIFTS could be the main reaction intermediates in the production of CO2.

b)

50

50

40 Calculated TOFCO (*103 s–1)

Calculated TOFCO (*103 s–1)

40

30

20

30

20

10

10

0

0 0

10

20

30

40

0

50

10

30

Experimental TOFCO

Experimental TOFCO (*103 s–1)

c)

20

40

50

(*103 s–1)

40

TOFCO (*103 s–1)

30

20

10

0

Experimental TOF Calculated TOF Calculated TOF (correct for Kco) 0.00

0.04

0.08

0.12

0.16

0.20

PCO (bar)

Fig. 41.21 Calculated TOF as a function of experimental TOF at varying syngas pressure: PH2 ¼ 0.17–0.56 bar, PCO ¼ 0.01–0.167 bar at 1.85 bar and 483 K for direct CO dissociation (a) and H-assisted CO

dissociation (b). (c) Measured and calculated TOF at different CO pressure at PH2 ¼ 0.56 bar and 483 K for H-assisted mechanism (Reproduced from Ref. [141]. Copyright © 2013 Elsevier Inc.)

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

957

b) 0.6

0.5

0.5

0.4

0.4 Site coverage

a) 0.6

Site coverage

41

0.3

0.3

0.2

0.2

0.1

0.1

0.0 0.1

0.2

0.3

0.4

0.5

H2 partial pressure (bar)

0.6

41

0.0 0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

CO partial pressure (bar)

Fig. 41.22 Site coverage of different species as a function of the partial pressure of H2 and CO, respectively, at 483 K and 1.85 bar. (a) PCO ¼ 0.056 bar and (b) PH2 ¼ 0.56 bar. Symbols: ●, CO*; ○, C1*; ▾, H*; and Δ, * (Reproduced from Ref. [141]. Copyright © 2013 Elsevier Inc.)

Table 41.7 The isotope effect at 483 K for four possible kinetic routes in H-assisted CO dissociation Elementary step ZPE contribution Entropic contribution Case 2.1 The isotope effect at 483 K assuming HCO* þ H* ! HCOH* as the rate-determining step 1 H2 þ 2* ! 2H* 0.35 1.75 2 CO* þ H* ! HCO* 0.57 1.09 3 HCO* þ H* ! HCOH* 1.98 1.00 Overall KIE Case 2.2 The isotope effect assuming HCOH* ! HC* þ OH* as the rate-determining step via HCO* 1 H2 þ 2* ! 2H* 0.35 1.75 2 CO* þ H* ! HCO* 0.57 1.09 3 HCO* þ H* ! HCOH* 1.09 0.53 4 HCOH* ! HC* þ OH* 1.23 0.83 Overall KIE Case 2.3 The isotope effect assuming COH* þ H* ! HCOH* as the rate-determining step 1 H2 þ 2* ! 2H* 0.35 1.75 2 CO* þ H* ! COH* 0.60 1.06 3 COH* þ H* ! HCOH* 0.56 1.23 Overall KIE Case 2.4 The isotope effect assuming HCOH* ! HC* þ OH* as the rate-determining step via COH* 1 H2 þ 2* ! 2H* 0.35 1.75 2 CO* þ H* ! COH* 0.6 1.06 3 COH* þ H* ! HCOH* 0.41 1.21 4 HCOH* ! HC* þ OH* 1.23 0.83 Overall KIE Reproduced from Ref. [141]. Copyright © 2013 Elsevier Inc. a Equilibrium isotope effect b Kinetic isotope effect

Overall KIE or IE 0.61a 0.62a 1.98b 0.75 0.61a 0.62a 0.58a 1.02b 0.22 0.61a 0.63a 0.69b 0.27 0.61a 0.63a 0.49a 1.02b 0.19

958

A. Holmen et al.

a)

225 °C, 1%Pt/ceria WGS with co-fed H2

Formate C-H band region

CO2

1

Pt-CO

CO2

15

Absorbance

10 0.25 0.20 0.15 0.10 0.05 0.00

b)

Absorbance

0.3

0.2

Time (min)

20

Relative intensity

0.8

0.6

0.4 Carbonates 0.2 Formates

5

0 0 3000 2800 2600 2400 2200 2000 1800 Wavenumber (cm–1)

1%Pt/ceria 225 °C with H2O and co-fed H2 Dash

Initial Bold 7.9 min Normal 23.2 min

Formate C-H band region

10 20 Time after isotopic switch (min)

0

Pt-CO

30

Fig. 41.24 Relative evolution of the intensity of 12C-containing carbonate band (●), 12C-containing formate band (), and 13CO2 (○) signal with time on stream at 493 K under 2% 13CO þ 7% H2O, following steady-state under 2% 12CO þ 7% H2O (Reprinted with permission from Ref. [178]. Copyright © 2007 Elsevier Inc.)

CO2 0.1 10%H2O/Ar

600

0.0

0.6 wt% Pt/CeO2 T = 300 °C

Fig. 41.23 (a) Sample-modified SSITKA–DRIFTS experiment for 1% Pt/ceria in co-fed H2 at 225  C. (b) DRIFTS spectra at the beginning, half-life point, and the end of a SSITKA switching experiment during water–gas shift using H2O as reactant under co-fed H2 (Reprinted with permission from Ref. [176]. Copyright © 2007 Elsevier B.V.)

Kalamaras et al. [184] designed a novel transient isotopic experiment to quantify the initial transient rates of reactions of adsorbed formate(–COOH) and CO by water, as shown in Fig. 41.25. Based on the fact that CO reacts faster with H2O compared to -COOH, the authors proposed that the WGS reaction on ceria-supported Pt at 300  C occurs largely via the “redox” mechanism, and to a lesser extent via the “formate” mechanism. It appears that the reactivity of surface formate species is very much dependent on the reaction temperature [215], reactant (co-fed H2 or H2O) [174, 216], the presence/loading of noble metal [217], Pt particle size [184], and promoters [146], which makes an unequivocal determination of reaction mechanism difficult.

41.13 Further Reading In the text above, the basic principles of SSITKA are presented together with a description of a standard experimental setup. Three examples from the studies of the hydrogenation of CO

Concentration (ppm)

500 H2 400 13CO 2

300 200

12CO 2

100 0 0

50

100

150

200

250

Time (s)

Fig. 41.25 Transient response curves of H2, 12CO2, and 13CO2 obtained on the 0.6 wt.% Pt/CeO2 catalyst at 300  C according to the following gas delivery: 3 vol.% 12CO/10 vol.% H2O/Ar (300  C, 30 min) ! 3 vol.% 13CO/Ar (300  C, 3 min) ! 10 vol.% H2O/Ar (300  C, t) (Reprinted with permission from Ref. [184]. Copyright © 2011 Elsevier Inc.)

using supported Co catalyst are also enclosed. Three examples of using SSITKA-FTIR for studying WGS are also included. The examples are selected to show how SSITKA combined with spectroscopy can be used to obtain useful information about the kinetics and mechanism. A detailed overview of the SSITKA literature is presented in the review by Ledesma et al. [6]. An update of recent literature has been included above. The technique

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

has been used to study many different and important chemical reactions. The review also contains details about advanced modelling and recent developments, in particular combinations of SSITKA with spectroscopic methods. Combination with other techniques is necessary in order to determine the identity and nature of reaction intermediates.

References 1. Bertole, C.J., Mims, C.A., Kiss, G.: Support and rhenium effects on the intrinsic site activity and methane selectivity of cobalt Fischer–Tropsch catalysts. J. Catal. 221, 191–203 (2004) 2. Happel, J.: Transient tracing. Chem. Eng. Sci. 33, 1567 (1978) 3. Bennett, C.O.: Understanding heterogeneous catalysis through the transient method. In: Catalysis Under Transient Conditions, vol. 178, pp. 1–32. American Chemical Society, Washington, DC (1982) 4. Biloen, P.: Transient kinetic methods. J. Mol. Catal. 21, 17–24 (1983) 5. Shannon, S.L., Goodwin, J.G.: Characterization of catalytic surfaces by isotopic-transient kinetics during steady-state reaction. Chem. Rev. 95, 677–695 (1995) 6. Ledesma, C., Yang, J., Chen, D., Holmen, A.: Recent approaches in mechanistic and kinetic studies of catalytic reactions using SSITKA technique. ACS Catal. 4, 4527–4547 (2014) 7. Carballo, J.M.G., Yang, J., Holmen, A., García-Rodríguez, S., Rojas, S., Ojeda, M., Fierro, J.L.G.: Catalytic effects of ruthenium particle size on the Fischer–Tropsch synthesis. J. Catal. 284, 102–108 (2011) 8. Ledesma, C., Yang, J., Blekkan, E.A., Holmen, A., Chen, D.: Carbon number dependence of reaction mechanism and kinetics in Co hydrogenation on a co-based catalyst. ACS Catal. 6, 6674–6686 (2016) 9. Athariboroujeny, M., Raub, A., Iablokov, V., Chenakin, S., Kovarik, L., Kruse, N.: Competing mechanisms in Co hydrogenation over Co-Mnox catalysts. ACS Catal. 9, 5603–5612 (2019) 10. Van Belleghem, J., Constales, D., Thybaut, J.W., Marin, G.B.: Complex reaction network generation for steady state isotopic transient kinetic analysis: Fischer-Tropsch synthesis. Comput. Chem. Eng. 125, 594–605 (2019) 11. Vasiliades, M.A., Kyprianou, K.K., Govender, N.S., Govender, A., Crous, R., Moodley, D., Efstathiou, A.M.: The effect of CO partial pressure on important kinetic parameters of methanation reaction on Co- based FTS catalyst studied by SSITKA-MS and operando DRIFTS-MS techniques. Catalysts. 10, 583 (2020) 12. Carvalho, A., Ordomsky, V.V., Luo, Y., Marinova, M., Muniz, A.R., Marcilio, N.R., Khodakov, A.Y.: Elucidation of deactivation phenomena in cobalt catalyst for Fischer-Tropsch synthesis using SSITKA. J. Catal. 344, 669–679 (2016) 13. Yang, X., Wang, R., Yang, J., Qian, W., Zhang, Y., Li, X., Huang, Y., Zhang, T., Chen, D.: Exploring the reaction paths in the consecutive Fe-based Ft catalyst-zeolite process for syngas conversion. ACS Catal. 10, 3797–3806 (2020) 14. Rebmann, E., Fongarland, P., Lecocq, V., Diehl, F., Schuurman, Y.: Kinetic modeling of transient Fischer–Tropsch experiments over Co/Al2o3 catalysts with different microstructures. Catal. Today. 275, 20–26 (2016) 15. Rebmann, E., Fongarland, P., Lecocq, V., Diehl, F., Schuurman, Y.: Measurement of the influence of the microstructure of aluminasupported cobalt catalysts on their activity and selectivity in Fischer–Tropsch synthesis by using steady-state and transient kinetics. ChemCatChem. 9, 2344–2351 (2017)

959 16. Chen, W., Pestman, R., Zijlstra, B., Filot, I.A.W., Hensen, E.J.M.: Mechanism of cobalt-catalyzed co hydrogenation: 1. Methanation. ACS Catal. 7, 8050–8060 (2017) 17. Chen, W., Filot, I.A.W., Pestman, R., Hensen, E.J.M.: Mechanism of cobalt-catalyzed Co hydrogenation: 2. Fischer-Tropsch synthesis. ACS Catal. 7, 8061–8071 (2017) 18. Den Otter, J.H., Yoshida, H., Ledesma, C., Chen, D., De Jong, K.P.: On the superior activity and selectivity of PtCo/Nb2O5 Fischer Tropsch catalysts. J. Catal. 340, 270–275 (2016) 19. Xie, J., Yang, J., Dugulan, A.I., Holmen, A., Chen, D., De Jong, K.P., Louwerse, M.J.: Size and promoter effects in supported iron Fischer-Tropsch catalysts: insights from experiment and theory. ACS Catal. 6, 3147–3157 (2016) 20. Schanke, D., Vada, S., Blekkan, E.A., Hilmen, A.M., Hoff, A., Holmen, A.: Study of Pt-promoted cobalt co hydrogenation catalysts. J. Catal. 156, 85–95 (1995) 21. Vada, S., Chen, B., Goodwin, J.G.: Isotopic transient study of La promotion of Co/Al(2)O(3)or CO hydrogenation. J. Catal. 153, 224–231 (1995) 22. Vada, S., Goodwin, J.G.: Analysis of Li promotion of methanol synthesis on Pd/SiO2 using isotopic transient kinetics. J. Phys. Chem. 99, 9479–9484 (1995) 23. Vada, S., Hoff, A., Adnanes, E., Schanke, D., Holmen, A.: FischerTropsch synthesis on supported cobalt catalysts promoted by platinum and rhenium. Top. Catal. 2, 155–162 (1995) 24. Haddad, G.J., Chen, B., Goodwin, J.G.: Effect of La3+ promotion of Co/SiO2 on CO hydrogenation. J. Catal. 161, 274–281 (1996) 25. Kogelbauer, A., Goodwin, J.G., Oukaci, R.: Ruthenium promotion of Co/Al2O3 Fischer-Tropsch catalysts. J. Catal. 160, 125–133 (1996) 26. Bajusz, I.G., Goodwin, J.G.: Hydrogen and temperature effects on the coverages and activities of surface intermediates during methanation on Ru/SiO2. J. Catal. 169, 157–165 (1997) 27. Bajusz, I.G., Kwik, D.J., Goodwin, J.G.: Methanation on K+modified Pt/SiO2: the impact of reaction conditions on the effective role of the promotor. Catal. Lett. 48, 151–157 (1997) 28. Hanssen, K.F., Blekkan, E.A., Schanke, D., Holmen, A.: Transient and steady-state studies of the effect of water on cobalt FischerTropsch catalysts. In: Froment, G.F., Waugh, K.C. (eds.) Studies in Surface Science and Catalysis, vol. 109, pp. 193–202. Elsevier, Amsterdam (1997) 29. Rothaemel, M., Hanssen, K.F., Blekkan, E.A., Schanke, D., Holmen, A.: The effect of water on cobalt Fischer-Tropsch catalysts studied by steady-state isotopic transient kinetic analysis (SSITKA). Catal. Today. 38, 79–84 (1997) 30. Ali, S.H., Goodwin, J.G.: SSITKA investigation of palladium precursor and support effects on CO hydrogenation over supported Pd catalysts. J. Catal. 176, 3–13 (1998) 31. Rohr, F., Holmen, A., Barbo, K.K., Warloe, P., Blekkan, E.A.: Modified alumina supports for cobalt Fischer-Tropsch catalysts. In: Parmaliana, A., Sanfilippo, D., Frusteri, F., Vaccari, A., Arena, F. (eds.) Natural Gas Conversion V, vol. 119, pp. 107–112. Elsevier, Amsterdam (1998) 32. Rohr, F., Lindvag, O.A., Holmen, A., Blekkan, E.A.: FischerTropsch synthesis over cobalt catalysts supported on zirconiamodified alumina. Catal. Today. 58, 247–254 (2000) 33. Bertole, C.J., Mims, C.A., Kiss, G.: The effect of water on the cobalt-catalyzed Fischer-Tropsch synthesis. J. Catal. 210, 84–96 (2002) 34. Panpranot, J., Goodwin, J.G., Sayari, A.: CO hydrogenation on Ru-promoted Co/MCM-41 catalysts. J. Catal. 211, 530–539 (2002) 35. Jongsomjit, B., Panpranot, J., Goodwin, J.G.: Effect of zirconiamodified alumina on the properties of Co/Gamma-Al2O3 catalysts. J. Catal. 215, 66–77 (2003) 36. Panpranot, J., Goodwin, J.G., Sayari, A.: Effect of H-2 partial pressure on surface reaction parameters during Co hydrogenation

41

960 on Ru-promoted silica-supported Co catalysts. J. Catal. 213, 78–85 (2003) 37. Sudsakorn, K., Goodwin, J.G., Adeyiga, A.A.: Effect of activation method on Fe FTS catalysts: investigation at the site level using SSITKA. J. Catal. 213, 204–210 (2003) 38. Bertole, C.J., Kiss, G., Mims, C.A.: The effect of surface-active carbon on hydrocarbon selectivity in the cobalt-catalyzed FischerTropsch synthesis. J. Catal. 223, 309–318 (2004) 39. Froseth, V., Storsaeter, S., Borg, O., Blekkan, E.A., Ronning, M., Holmen, A.: Steady state isotopic transient kinetic analysis (SSITKA) of CO hydrogenation on different Co catalysts. Appl. Catal. A Gen. 289, 10–15 (2005) 40. Lohitharn, N., Goodwin Jr., J.G.: Effect of K promotion of Fe and FEMN Fischer-Tropsch synthesis catalysts: analysis at the site level using SSITKA. J. Catal. 260, 7–16 (2008) 41. Lohitharn, N., Goodwin Jr., J.G.: Impact of Cr, Mn and Zr addition on Fe Fischer-Tropsch synthesis catalysis: investigation at the active site level using SSITKA. J. Catal. 257, 142–151 (2008) 42. Lohitharn, N., Goodwin Jr., J.G., Lotero, E.: Fe-based FischerTropsch synthesis catalysts containing carbide-forming transition metal promoters. J. Catal. 255, 104–113 (2008) 43. den Breejen, J.P., Radstake, P.B., Bezemer, G.L., Bitter, J.H., Froseth, V., Holmen, A., de Jong, K.P.: On the origin of the cobalt particle size effects in Fischer-Tropsch catalysis. J. Am. Chem. Soc. 131, 7197–7203 (2009) 44. Lohitharn, N., Goodwin Jr., J.G.: An investigation using SSITKA of chain growth on Fe and FEMN k Fischer-Tropsch synthesis catalysts. Catal. Commun. 10, 758–762 (2009) 45. Govender, N.S., de Croon, M.H.J.M., Schouten, J.C.: Reactivity of surface carbonaceous intermediates on an iron-based FischerTropsch catalyst. Appl. Catal. A Gen. 373, 81–89 (2010) 46. Graf, B., Schulte, H., Muhler, M.: The formation of methane over iron catalysts applied in Fischer-Tropsch synthesis a transient and steady state kinetic study. J. Catal. 276, 66–75 (2010) 47. Rane, S., Borg, O., Yang, J., Rytter, E., Holmen, A.: Effect of alumina phases on hydrocarbon selectivity in Fischer-Tropsch synthesis. Appl. Catal. A Gen. 388, 160–167 (2010) 48. Yang, J., Tveten, E.Z., Chen, D., Holmen, A.: Understanding the effect of cobalt particle size on Fischer-Tropsch synthesis: surface species and mechanistic studies by SSITKA and kinetic isotope effect. Langmuir. 26, 16558–16567 (2010) 49. den Breejen, J.P., Frey, A.M., Yang, J., Holmen, A., van Schooneveld, M.M., de Groot, F.M.F., Stephan, O., Bitter, J.H., de Jong, K.P.: A highly active and selective manganese oxide promoted cobalt-on-silica Fischer-Tropsch catalyst. Top. Catal. 54, 768–777 (2011) 50. Phan, X.K., Yang, J., Bakhtiary-Davijnay, H., Myrstad, R., Venvik, H.J., Holmen, A.: Studies of macroporous structured alumina based cobalt catalysts for Fischer-Tropsch synthesis. Catal. Lett. 141, 1739–1745 (2011) 51. Tsai, Y.-T., Goodwin Jr., J.G.: Comparison of chemisorption close to ambient vs. under reaction conditions for group VIII metal catalysts. J. Catal. 281, 128–136 (2011) 52. Mo, X., Tsai, Y.-T., Gao, J., Mao, D., Goodwin Jr., J.G.: Effect of component interaction on the activity of Co/CUZnO for CO hydrogenation. J. Catal. 285, 208–215 (2012) 53. Tsai, Y.-T., Mo, X., Goodwin Jr., J.G.: Synergy of components in CuZnO and CuZnO/Al2O3 on methanol synthesis: analysis at the site level by SSITKA. Top. Catal. 55, 757–770 (2012) 54. Yang, J., Chen, D., Holmen, A.: Understanding the kinetics and re promotion of carbon nanotube supported cobalt catalysts by SSITKA. Catal. Today. 186, 99–108 (2012) 55. Enger, B.C., Froseth, V., Yang, J., Rytter, E., Holmen, A.: Ssitka analysis of CO hydrogenation on Zn modified cobalt catalysts. J. Catal. 297, 187–192 (2013)

A. Holmen et al. 56. Lillebo, A.H., Patanou, E., Yang, J., Blekkan, E.A., Holmen, A.: The effect of alkali and alkaline earth elements on cobalt based Fischer-Tropsch catalysts. Catal. Today. 215, 60–66 (2013) 57. Patanou, E., Lillebø, A.H., Yang, J., Chen, D., Holmen, A., Blekkan, E.A.: Microcalorimetric studies on Co–Re/Γ-Al2O3 catalysts with Na impurities for Fischer–Tropsch synthesis. Ind. Eng. Chem. Res. 53, 1787–1793 (2014) 58. Wang, F., He, S., Chen, H., Wang, B., Zheng, L., Wei, M., Evans, D.G., Duan, X.: Active site dependent reaction mechanism over Ru/CeO2 catalyst toward CO2 Methanation. J. Am. Chem. Soc. 138, 6298–6305 (2016) 59. Gutterød, E.S., et al.: Hydrogenation of CO2 to methanol by Pt nanoparticles encapsulated in Uio-67: deciphering the role of the metal-organic framework. J. Am. Chem. Soc. 142, 999–1009 (2020) 60. Wang, X., Szanyi, J.: In The Mechanism of CO2 Hydrogenation over Pd/Al2o3: A Steady State Isotopic Transient Kinetic Analysis (Ssitka)/Operando-Ftir Spectroscopy Investigation, Catalysis and Reaction Engineering Division 2016 - Core Programming Area at the 2016 AIChE Annual Meeting, 2016; pp. 704–706 61. Vrijburg, W.L., Garbarino, G., Chen, W., Parastaev, A., Longo, A., Pidko, E.A., Hensen, E.J.M.: Ni-Mn catalysts on silica-modified alumina for CO2 Methanation. J. Catal. 382, 358–371 (2020) 62. Kammert, J., et al.: Nature of reactive hydrogen for ammonia synthesis over a Ru/C12A7 electride catalyst. J. Am. Chem. Soc. 142, 7655–7667 (2020) 63. Nwalor, J.U., Goodwin Jr., J.G.: Isotopic tracing study of K promotion of NH3 synthesis on Ru. Top. Catal. 1, 285–293 (1994) 64. McClaine, B.C., Davis, R.J.: Importance of product readsorption during isotopic transient analysis of ammonia synthesis on Ba-promoted Ru/BaX catalyst. J. Catal. 211, 379–386 (2002) 65. McClaine, B.C., Davis, R.J.: Isotopic transient kinetic analysis of Cs-promoted Ru/MgO during ammonia synthesis. J. Catal. 210, 387–396 (2002) 66. Siporin, S.E., Davis, R.J.: Isotopic transient analysis of ammonia synthesis over Ru/MgO catalysts promoted by cesium, barium, or lanthanum. J. Catal. 222, 315–322 (2004) 67. Polo-Garzon, F., et al.: Elucidation of the reaction mechanism for high-temperature water gas shift over an industrial-type copperchromium-iron oxide catalyst. J. Am. Chem. Soc. 141, 7990–7999 (2019) 68. Petallidou, K.C., Kalamaras, C.M., Efstathiou, A.M.: The effect of La3+, Ti4+ and Zr4+ dopants on the mechanism of WGS on ceriadoped supported Pt catalysts. Catal. Today. 228, 183–193 (2014) 69. Brundage, M.A., Chuang, S.S.C.: Experimental and modeling study of hydrogenation using deuterium step transient response during ethylene hydroformylation. J. Catal. 164, 94–108 (1996) 70. Liu, C., Ozkan, U.S.: Spectroscopic and structural characterization of chlorine loading effects on Mo/Si: Ti catalysts in oxidative dehydrogenation of ethane. J. Phys. Chem. A. 109, 1260–1268 (2005) 71. Kondratenko, E.V., Ovsitser, O., Radnik, J., Schneider, M., Kraehnert, R., Dingerdissen, U.: Influence of reaction conditions on catalyst composition and selective/non-selective reaction pathways of the ODP reaction over V2O3, VO2 and V2O5 with O-2 and N2O. Appl. Catal. A Gen. 319, 98–110 (2007) 72. Ovsitser, O., Kondratenko, E.V.: Similarity and differences in the oxidative dehydrogenation of C-2-C-4 alkanes over nano-sized Vox species using N2O and O-2. Catal. Today. 142, 138–142 (2009) 73. Kim, S.Y., Goodwin, J.G., Galloway, D.: N-butane isomerization on sulfated zirconia: isotopic transient kinetic analysis of reaction at the site level. Catal. Lett. 64, 1–8 (2000) 74. Kim, S.Y., Goodwin, J.G., Galloway, D.: N-butane isomerization on sulfated zirconia: active site heterogeneity and deactivation. Catal. Today. 63, 21–32 (2000)

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

75. Kim, S.Y., Goodwin, J.G., Hammache, S., Auroux, A., Galloway, D.: The impact of Pt and H-2 on N-butane isomerization over sulfated zirconia: changes in intermediates coverage and reactivity. J. Catal. 201, 1–12 (2001) 76. Kim, S.Y., Lohitharn, N., Goodwin, J.G., Olindo, R., Pinna, F., Canton, P.: The effect of Al2O3-promotion of sulfated zirconia on N-butane isomerization: an isotopic transient kinetic analysis. Catal. Commun. 7, 209–213 (2006) 77. Ocal, M., Oukaci, R., Marcelin, G., Agarwal, S.K.: Steady-state isotopic transient kinetic-analysis (SSITKA) investigation of no reduction with CO over perovskite catalysts. Ind. Eng. Chem. Res. 33, 2930–2934 (1994) 78. Burch, R., Shestov, A.A., Sullivan, J.A.: A steady-state isotopic transient kinetic analysis of the No/O-2/H-2 reaction over Pt/SiO2 catalysts. J. Catal. 188, 69–82 (1999) 79. Burch, R., Shestov, A.A., Sullivan, J.A.: A transient kinetic study of the mechanism of the No+H-2 reaction over Pt/SiO2 catalysts 1. Isotopic transient kinetics and temperature programmed analysis. J. Catal. 186, 353–361 (1999) 80. Burch, R., Shestov, A.A., Sullivan, J.A.: A transient kinetic study of the mechanism of the NO/C3H6/O-2 reaction over Pt-SiO2 catalysts - part 2: steady-state isotopic transient kinetic analysis. J. Catal. 182, 497–506 (1999) 81. Sadovskaya, E.M., Suknev, A.P., Pinaeva, L.G., Goncharov, V.B., Mirodatos, C., Balzhinimaev, B.S.: On the role of oxygen in the reaction of selective no reduction with methane over Co/ZSM-5 catalyst. In: Corma, A., Melo, F.V., Mendioroz, S., Fierro, J.L.G. (eds.) Studies in Surface Science and Catalysis, vol. 130, pp. 1505–1510. Elsevier, Amsterdam (2000) 82. Sullivan, J.A., Burch, R., Shestov, A.A.: Transient techniques in the study of lean-NOx reduction over supported Pt catalysts. Chem. Eng. Res. Des. 78, 947–953 (2000) 83. Sadovskaya, E.M., Suknev, A.P., Pinaeva, L.G., Goncharov, V.B., Bal’zhinimaev, B.S., Chupin, C., Mirodatos, C.: Mechanism and kinetics of the selective NO reduction over CO-ZSM-5 studied by the SSITKA technique part 1: no Chi adsorbed species formation. J. Catal. 201, 159–168 (2001) 84. Burch, R., Coleman, M.D.: An investigation of promoter effects in the reduction of NO by H-2 under lean-burn conditions. J. Catal. 208, 435–447 (2002) 85. Burch, R.: The investigation of mechanisms in environmental catalysis using time-resolved methods. Top. Catal. 24, 97–102 (2003) 86. Sadovskaya, E.M., Suknev, A.P., Goncharov, V.B., Bal’zhinimaev, B.S., Mirodatos, C.: Reaction kinetics and mechanism of selective no reduction on a CoZSM-5 catalyst as studied by SSITKA. Kinet. Catal. 45, 436–445 (2004) 87. Sadovskaya, E.M., Suknev, A.P., Pinaeva, L.G., Goncharov, V.B., Bal’zhinimaev, B.S., Chupin, C., Perez-Ramirez, J., Mirodatos, C.: Mechanism and kinetics of the selective NO reduction over Co-ZSM-5 studied by the SSITKA technique 2. Reactivity of NOX-adsorbed species with methane. J. Catal. 225, 179–189 (2004) 88. Schuurman, Y., vanVeen, A.C., Chupin, C., Mirodatos, C.: Can isotopic transient and formal steady-state kinetics lead to a converging surface chemistry analysis? Case study on the selective reduction of NO by CH4 over Co-ZSM-5 catalysts. Top. Catal. 39, 45–52 (2006) 89. Kalamaras, C.M., Olympiou, G.G., Pârvulescu, V.I., Cojocaru, B., Efstathiou, A.M.: Selective catalytic reduction of NO by H2/C3H6 over Pt/Ce1-xZrxO2-δ: the synergy effect studied by transient techniques. Appl. Catal. B Environ. 206, 308–318 (2017) 90. Avgouropoulos, G.: Isotopic transient study of methanol decomposition over noble metal/ceria catalysts. Catal. Commun. 10, 682–686 (2009) 91. Papavasiliou, J., Avgouropoulos, G., Ioannides, T.: Steady-state isotopic transient kinetic analysis of steam reforming of methanol

961 over Cu-based catalysts. Appl. Catal. B Environ. 88, 490–496 (2009) 92. Takise, K., Manabe, S., Muraguchi, K., Higo, T., Ogo, S., Sekine, Y.: Anchoring effect and oxygen redox property of Co/La0.7Sr0.3AlO3-δ perovskite catalyst on toluene steam reforming reaction. Appl. Catal. A Gen. 538, 181–189 (2017) 93. Polo-Garzon, F., Pakhare, D., Spivey, J.J., Bruce, D.A.: Dry reforming of methane on Rh-doped pyrochlore catalysts: a steady-state isotopic transient kinetic study. ACS Catal. 6, 3826–3833 (2016) 94. Vasiliades, M.A., Makri, M.M., Djinović, P., Erjavec, B., Pintar, A., Efstathiou, A.M.: Dry reforming of methane over 5 wt% Ni/Ce1-xPrxO2-δ catalysts: performance and characterisation of active and inactive carbon by transient isotopic techniques. Appl. Catal. B Environ. 197, 168–183 (2016) 95. Polychronopoulou, K., Costa, C.N., Efstathiou, A.M.: The role of oxygen and hydroxyl support species on the mechanism of H-2 production in the steam reforming of phenol over metal oxidesupported-Rh and -Fe catalysts. Catal. Today. 112, 89–93 (2006) 96. Polychronopoulou, K., Efstathiou, A.M.: Spillover of labile OH, H, and 0 species in the H-2 production by steam reforming of phenol over supported-Rh catalysts. Catal. Today. 116, 341–347 (2006) 97. Nijhuis, T.A., Sacaliuc-Parvulescu, E., Govender, N.S., Schouten, J.C., Weckhuysen, B.M.: The role of support oxygen in the epoxidation of propene over gold-titania catalysts investigated by isotopic transient kinetics. J. Catal. 265, 161–169 (2009) 98. Rotko, M., Machocki, A., Słowik, G.: The mechanism of the CH4/ O2 reaction on the Pd-Pt/Γ-Al2O3 catalyst: a SSITKA study. Appl. Catal. B Environ. 160-161, 298–306 (2014) 99. Rotko, M., Zawadzki, W., Redko, V.: Steady state isotopic transient kinetic analysis of the combustion of CH4 over Co-Mn-O catalysts. Catal. Commun. 125, 32–36 (2019) 100. Calla, J.T., Davis, R.J.: Investigation of alumina-supported an catalyst for CO oxidation by isotopic transient analysis and X-ray absorption spectroscopy. J. Phys. Chem. B. 109, 2307–2314 (2005) 101. Calla, J.T., Bore, M.T., Datye, A.K., Davis, R.J.: Effect of alumina and titania on the oxidation of CO over Au nanoparticles evaluated by C-13 isotopic transient analysis. J. Catal. 238, 458–467 (2006) 102. Calla, J.T., Davis, R.J.: Oxygen-exchange reactions during CO oxidation over Titania- and alumina-supported Au nanoparticles. J. Catal. 241, 407–416 (2006) 103. Castellanos, I., Bazin, P., Thomas, S., Marie, O., Daturi, M.: A relevant estimation of the TOF for methanol oxidation over Au/CeO2: a combined SSITKA and FTIR operando contribution. Top. Catal. 59, 337–346 (2016) 104. Efstathiou, A.M., Verykios, X.E.: Transient methods in heterogeneous catalysis: experimental features and application to study mechanistic aspects of the CH4/O2 (OCM), NH3/O2 and NO/He reactions. Appl. Catal. A Gen. 151, 109–166 (1997) 105. Ocal, M., Oukaci, R., Marcelin, G., Jang, B.W.L., Spivey, J.J.: Steady-state isotopic transient kinetic analysis on Pd-supported hexaaluminates used for methane combustion in the presence and absence of NO. Catal. Today. 59, 205–217 (2000) 106. Sadykov, V.A., et al.: Fuel-rich methane combustion: role of the Pt dispersion and oxygen mobility in a fluorite-like complex oxide support. Catal. Today. 117, 475–483 (2006) 107. Machocki, A., Rotko, M., Stasinska, B.: SSITKA studies of the catalytic flameless combustion of methane. Catal. Today. 137, 312–317 (2008) 108. Machocki, A., Rotko, M., Gac, W.: Steady state isotopic transient kinetic analysis of flameless methane combustion over Pd/Al2O3 and Pt/Al2O3 catalysts. Top. Catal. 52, 1085–1097 (2009) 109. Rotko, M., Machocki, A., Stasinska, B.: Studies of catalytic process of complete oxidation of methane by SSITKA method. Appl. Surf. Sci. 256, 5585–5589 (2010) 110. Heracleous, E., Lemonidou, A.A.: Ni-Nb-O mixed oxides as highly active and selective catalysts for ethene production via

41

962 ethane oxidative dehydrogenation. Part II: mechanistic aspects and kinetic modeling. J. Catal. 237, 175–189 (2006) 111. Frank, B., Rinaldi, A., Blume, R., Schloegl, R., Su, D.S.: Oxidation stability of multiwalled carbon nanotubes for catalytic applications. Chem. Mater. 22, 4462–4470 (2010) 112. Ouyang, X., Susannah, L.S.: Co oxidation catalyzed by Pd-doped BaCeO3: coexistence of Langmuir-Hinshelwood and BaCeO3mediated mechanisms. MRS Online Proceedings Library. 1217, 205 (2010) 113. Zhang, R., Haddadin, T., Rubiano, D.P., Nair, H., Polster, C.S., Baertsch, C.D.: Quantification of reactive CO and H-2 on CuOxCeO2 during CO preferential oxidation by reactive titration and steady state isotopic transient kinetic analysis. ACS Catal. 1, 519–525 (2011) 114. Gulyaeva, Y.K., Kaichev, V.V., Zaikovskii, V.I., Suknev, A.P., Bal’zhinimaev, B.S.: Selective hydrogenation of acetylene over Pd/fiberglass catalysts: kinetic and isotopic studies. Appl. Catal. A Gen. 506, 197–205 (2015) 115. Glazneva, T.S., Kotsarenko, N.S., Paukshtis, E.A.: Surface acidity and basicity of oxide catalysts: from aqueous suspensions to in situ measurements. Kinet. Catal. 49, 859–867 (2008) 116. Olds, D., Lawler, K.V., Paecklar, A.A., Liu, J., Page, K., Peterson, P.F., Forster, P.M., Neilson, J.R.: Capturing the details of N2 adsorption in zeolite X using stroboscopic isotope contrasted neutron total scattering. Chem. Mater. 30, 296–302 (2018) 117. Rochoux, M., Guo, Y., Schuurman, Y., Farrusseng, D.: Determination of oxygen adsorption-desorption rates and diffusion rate coefficients in perovskites at different oxygen partial pressures by a microkinetic approach. Phys. Chem. Chem. Phys. 17, 1469–1481 (2015) 118. Gu, X.K., Carneiro, J.S.A., Samira, S., Das, A., Ariyasingha, N.M., Nikolla, E.: Efficient oxygen electrocatalysis by nanostructured mixed-metal oxides. J. Am. Chem. Soc. 140, 8128–8137 (2018) 119. Zabilskiy, M., Sushkevich, V.L., Palagin, D., Newton, M.A., Krumeich, F., van Bokhoven, J.A.: The unique interplay between copper and zinc during catalytic carbon dioxide hydrogenation to methanol. Nat. Commun. 11, 2409 (2020) 120. Jabłońska, M.: An application of steady-state isotopic-transient kinetic analysis (SSITKA) in DeNOx process. ChemCatChem. 13, 818–827 (2021) 121. Pinaeva, L.G., Prosvirin, I.P., Dovlitova, L.S., Danilova, I.G., Sadovskaya, E.M., Isupova, L.A.: MeOx/Al2O3 and MeOx/CeO2 (Me ¼ Fe, Co, Ni) catalysts for high temperature N2O decomposition and NH3 oxidation. Catal. Sci. Technol. 6, 2150–2161 (2016) 122. Polo-Garzon, F., Blum, T.F., Bao, Z., Wang, K., Fung, V., Huang, Z., Bickel, E.E., Jiang, D.E., Chi, M., Wu, Z.: In situ strong metalsupport interaction (SMSI) affects catalytic alcohol conversion. ACS Catal. 11, 1938–1945 (2021) 123. Hanspal, S., Young, Z.D., Shou, H., Davis, R.J.: Multiproduct steady-state isotopic transient kinetic analysis of the ethanol coupling reaction over hydroxyapatite and magnesia. ACS Catal. 5, 1737–1746 (2015) 124. Terreni, J., Trottmann, M., Delmelle, R., Heel, A., Trtik, P., Lehmann, E.H., Borgschulte, A.: Observing chemical reactions by time-resolved high-resolution neutron imaging. J. Phys. Chem. C. 122, 23574–23581 (2018) 125. Olds, D., et al.: A high precision gas flow cell for performing in situ neutron studies of local atomic structure in catalytic materials. Rev. Sci. Instrum. 88, 034101 (2017) 126. Thomas, S., Marie, O., Bazin, P., Lietti, L., Visconti, C.G., Corbetta, M., Manenti, F., Daturi, M.: Modelling a reactor cell for operando IR studies: from qualitative to fully quantitative kinetic investigations. Catal. Today. 283, 176–184 (2017) 127. Brush, A., McDonald, S., Dupré, R., Kota, S., Mullen, G.M., Buddie Mullins, C.: Apparatus for efficient utilization of isotopically-labeled gases in pulse transient studies of

A. Holmen et al. heterogeneously catalyzed gas phase reactions. React. Chem. Eng. 2, 512–520 (2017) 128. Vada, S.: Isotopic transient kinetic investigations of catalytic reactions. Trondheim, Norway: Doctoral Theses at NTH 1994: 79 (1994) 129. Frøseth, V.: A steady-state isotopic transient kinetic study of Co catalysts on different supports. Trondheim, Norway: Doctoral Theses at NTNU, 2006:102 (2006) 130. Yang, J.: A steady-state Isotpic transient kinetic study of cobalt catalysts: mechanistic insights and effect of cobalt particle size, supports and promotes. Trondheim, Norway: Doctoral Theses at NTNU, 2001:280. (2011) 131. Bariås, O.A.: Transient kinetic investigation of the catalytic dehydrogenation of propane. Trondheim, Norway: Doctoral Theses at NTH, 1993:104. (1993) 132. Ali, S.H., Goodwin, J.G.: Impact of readsorption effects and their removal from surface reaction parameters obtained by isotopic transient kinetic analysis: methanol synthesis on Pd/SiO2. J. Catal. 171, 339–344 (1997) 133. Pansare, S., Sirijaruphan, A., James, G., Goodwin, J.: Investigation of reaction at the site level using SSITKA. In: Isotopes in Heterogeneous Catalysis, pp. 183–211, Imperial College Press, Lodon 134. Biloen, P., Helle, J.N., Vandenberg, F.G.A., Sachtler, W.M.H.: On the activity of Fischer-Tropsch and methanation catalysts - a study utilizing isotopic transients. J. Catal. 81, 450–463 (1983) 135. Rothaemel, M., Hanssen, K.F., Blekkan, E.A., Schanke, D., Holmen, A.: Determination of reactivity distributions from steady-state isotopic transient data: a model study. Catal. Today. 40, 171–179 (1998) 136. Eddy Hoost, T., Goodwin, J.G.: Nonparametric determination of reactivity distributions from isotopic transient kinetic data. J. Catal. 134, 678–690 (1992) 137. Hedrick, S.A., Chuang, S.S.C., Brundage, M.A.: Deuterium pulse transient analysis for determination of heterogeneous ethylene hydroformylation mechanistic parameters. J. Catal. 185, 73–90 (1999) 138. Basallote, M.G., Bernal, S.n., Gatica, J.M., Pozo, M.: Steady-state isotopic transient kinetic analysis of the H2/D2 exchange reaction as a tool for characterising the metal phase in supported platinum catalysts. Appl. Catal. A Gen. 232, 39–50 (2002) 139. Yang, J., Shafer, W.D., Pendyala, V.R.R., Jacobs, G., Ma, W., Chen, D., Holmen, A., Davis, B.H.: Fischer–Tropsch synthesis: deuterium kinetic isotopic effect for a 2.5% Ru/NaY catalyst. Top. Catal. 57, 508–517 (2014) 140. Ojeda, M., Li, A., Nabar, R., Nilekar, A.U., Mavrikakis, M., Iglesia, E.: Kinetically relevant steps and H2/D2 isotope effects in FischerTropsch synthesis on Fe and Co catalysts. J. Phys. Chem. C. 114, 19761–19770 (2010) 141. Yang, J., Qi, Y., Zhu, J., Zhu, Y.-A., Chen, D., Holmen, A.: Reaction mechanism of CO activation and methane formation on Co Fischer–Tropsch catalyst: a combined DFT, transient, and steady-state kinetic modeling. J. Catal. 308, 37–49 (2013) 142. Qi, Y., Yang, J., Duan, X., Zhu, Y.-A., Chen, D., Holmen, A.: Discrimination of the mechanism of CH4 formation in Fischer–Tropsch synthesis on Co catalysts: a combined approach of DFT, kinetic isotope effects and kinetic analysis. Cat. Sci. Technol. 4, 3534–3543 (2014) 143. Qi, Y., Yang, J., Holmen, A., Chen, D.: Investigation of C1 + C1 coupling reactions in cobalt-catalyzed Fischer-Tropsch synthesis by a combined DFT and kinetic isotope study. Catalysts. 9, 551 (2019) 144. Meunier, F.C.: The power of quantitative kinetic studies of adsorbate reactivity by operando FTIR spectroscopy carried out at chemical potential steady-state. Catal. Today. 155, 164–171 (2010) 145. Meunier, F.C., Goguet, A., Shekhtman, S., Rooney, D., Daly, H.: A modified commercial DRIFTS cell for kinetically relevant

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

operando studies of heterogeneous catalytic reactions. Appl. Catal. A Gen. 340, 196–202 (2008) 146. Kalamaras, C.M., Petallidou, K.C., Efstathiou, A.M.: The effect of La3+-doping of CeO2 support on the water-gas shift reaction mechanism and kinetics over Pt/Ce1xLaxO2δ. Appl. Catal. B Environ. 136-137, 225–238 (2013) 147. Efstathiou, A.M.: Elucidation of mechanistic and kinetic aspects of water–gas shift reaction on supported Pt and Au catalysts via transient isotopic techniques. In: Catalysis, vol. 28, pp. 175–236. The Royal Society of Chemistry, Lodon (2016) 148. Paredes-Nunez, A., Jbir, I., Bianchi, D., Meunier, F.C.: Spectrum baseline artefacts and correction of gas-phase species signal during diffuse reflectance Ft-IR analyses of catalysts at variable temperatures. Appl. Catal. A Gen. 495, 17–22 (2015) 149. Hadjiivanov, K., Mihaylov, M., Panayotov, D., Ivanova, E., Chakarova, K.: Isotopes in the FTIR investigations of solid surfaces. Spectrosc. Prop. Inorg. Organomet. Compd. 45, 43 (2014) 150. Dalziel, K., Obrien, J.R.P.: Spectrokinetic studies of the reaction of hydrogen peroxide with haemoglobin in dithionite solutions. Biochem. J. 67, 124–136 (1957) 151. Tamaru, K.: Adsorption measurements during surface catalysis. Adv. Catal. 15, 65–96 (1964) 152. Ueno, A., Onishi, T., Tamaru, K.: Dynamic technique to elucidate reaction intermediate in surface catalysis - water-gas shift reaction. Trans. Faraday Soc. 66, 756 (1970) -& 153. Stockwell, D.M., Chung, J.S., Bennett, C.O.: A transient infrared and isotopic study of methanation over Ni/Al2O3. J. Catal. 112, 135–144 (1988) 154. Balakos, M.W., Chuang, S.S.C., Srinivas, G., Brundage, M.A.: Infrared study of the dynamics of adsorbed species during Co hydrogenation. J. Catal. 157, 51–65 (1995) 155. Agnelli, M., Swaan, H.M., Marquez-Alvarez, C., Martin, G.A., Mirodatos, C., Co Hydrogenation on a Nickel Catalyst - II: A mechanistic study by transient kinetics and infrared spectroscopy. J. Catal. 175, 117–128 (1998) 156. Eckle, S., Denkwitz, Y., Behm, R.J.: Activity, selectivity, and adsorbed reaction intermediates/reaction side products in the selective methanation of Co in reformate gases on supported Ru catalysts. J. Catal. 269, 255–268 (2010) 157. Schweicher, J., Bundhoo, A., Frennet, A., Kruse, N., Daly, H., Meunier, F.C.: DRIFTS/MS studies during chemical transients and SSITKA of the Co/H-2 reaction over Co-MgO catalysts. J. Phys. Chem. C. 114, 2248–2255 (2010) 158. Eckle, S., Anfang, H.-G., Behm, R.J.: Reaction intermediates and side products in the methanation of CO and CO2 over supported Ru catalysts in H-2-rich reformate gases. J. Phys. Chem. C. 115, 1361–1367 (2011) 159. Yang, Y., Mims, C.A., Disselkamp, R.S., Peden, C.H.F., Campbell, C.T.: Simultaneous MS-IR studies of surface formate reactivity under methanol synthesis conditions on Cu/SiO2. Top. Catal. 52, 1440–1447 (2009) 160. Rezvani, A., Abdel-Mageed, A.M., Ishida, T., Murayama, T., Parlinska-Wojtan, M., Behm, R.J.: CO2 reduction to methanol on Au/CeO2 catalysts: mechanistic insights from activation/deactivation and SSITKA measurements. ACS Catal. 10, 3580–3594 (2020) 161. Wang, X., Shi, H., Szanyi, J.: Controlling selectivities in CO2 reduction through mechanistic understanding. Nat. Commun. 8, 513 (2017) 162. Matyshak, V.A., Krylov, O.V.: In-situ IR spectroscopy of intermediates in heterogeneous oxidative catalysis. Catal. Today. 25, 1–87 (1995) 163. van Neer, F., Bliek, A.: The feedback mechanism in selfoscillations for Co oxidation over Europt-3. Chem. Eng. Sci. 54, 4483–4499 (1999)

963 164. Cheung, P., Liu, H.C., Iglesia, E.: Kinetics and mechanism of dimethyl ether oxidation to formaldehyde on supported molybdenum oxide domains. J. Phys. Chem. B. 108, 18650–18658 (2004) 165. Rajasree, R., Hoebink, J., Schouten, J.C.: Transient kinetics of carbon monoxide oxidation by oxygen over supported palladium/ ceria/zirconia three-way catalysts in the absence and presence of water and carbon dioxide. J. Catal. 223, 36–43 (2004) 166. Bazin, P., Thomas, S., Marie, O., Daturi, M.: New insights into the methanol oxidation mechanism over Au/CeO2 catalyst through complementary kinetic and FTIR operando SSITKA approaches. Catal. Today. 182, 3–11 (2012) 167. El-Roz, M., Bazin, P., Daturi, M., Thibault-Starzyk, F.: Operando infrared (IR) coupled to steady-state isotopic transient kinetic analysis (SSITKA) for photocatalysis: reactivity and mechanistic studies. ACS Catal. 3, 2790–2798 (2013) 168. El-Roz, M., Bazin, P., Daturi, M., Thibault-Starzyk, F.: On the mechanism of methanol photooxidation to methylformate and carbon dioxide on TiO2: an operando-FTIR study. Phys. Chem. Chem. Phys. 17, 11277–11283 (2015) 169. Efstathiou, A.M.: Elucidation of mechanistic and kinetic aspects of water-gas shift reaction on supported Pt and Au catalysts via transient isotopic techniques. In: Catalysis, vol. 28, pp. 175–236, The Royal Society of Chemistry, Lodon (2016) 170. Cámara, A.L., Chansai, S., Hardacre, C., Martínez-Arias, A.: The water-gas shift reaction over CeO2/CuO: operando SSITKADRIFTS-mass spectrometry study of low temperature mechanism. Int. J. Hydrog. Energy. 39, 4095–4101 (2014) 171. Goguet, A., Meunier, F.C., Tibiletti, D., Breen, J.P., Burch, R.: Spectrokinetic investigation of reverse water-gas-shift reaction intermediates over a Pt/CeO2 catalyst. J. Phys. Chem. B. 108, 20240–20246 (2004) 172. Meunier, F.C., Tibiletti, D., Goguet, A., Reid, D., Burch, R.: On the reactivity of carbonate species on a Pt/CeO2 catalyst under various reaction atmospheres: application of the isotopic exchange technique. Appl. Catal. A Gen. 289, 104–112 (2005) 173. Meunier, F.C., Tibiletti, D., Goguet, A., Burch, R.: DRIFTS-MSSSITKA study of the reverse water-gas shift reaction. Oil Gas Sci. Technol.-Revue D Ifp Energies Nouvelles. 61, 497–502 (2006) 174. Tibiletti, D., Goguet, A., Reid, D., Meunier, F.C., Burch, R.: On the need to use steady-state or operando techniques to investigate reaction mechanisms: an in situ DRIFTS and SSITKA-based study example. Catal. Today. 113, 94–101 (2006) 175. Tibiletti, D., Meunier, F.C., Goguet, A., Reid, D., Burch, R., Boaro, M., Vicario, M., Trovarelli, A.: An investigation of possible mechanisms of the water-gas shift reaction over a ZrO2-supported Pt catalyst. J. Catal. 244, 183–191 (2006) 176. Jacobs, G., Davis, B.H.: Low temperature water-gas shift: applications of a modified SSITKA-DRIFTS method under conditions of H2 co-feeding over metal/ceria and related oxides. Appl. Catal. A Gen. 333, 192–201 (2007) 177. Meunier, F.C., Goguet, A., Hardacre, C., Burch, R., Thompsett, D.: Quantitative DRIFTS investigation of possible reaction mechanisms for the water-gas shift reaction on high-activity Pt- and Au-based catalysts. J. Catal. 252, 18–22 (2007) 178. Meunier, F.C., Reid, D., Goguet, A., Shekhtman, S., Hardacre, C., Burch, R., Deng, W., Flytzani-Stephanopoulos, M.: Quantitative analysis of the reactivity of formate species seen by DRIFTS over a Au/Ce(La)O-2 water-gas shift catalyst: first unambiguous evidence of the minority role of formates as reaction intermediates. J. Catal. 247, 277–287 (2007) 179. Meunier, F.C., Tibiletti, D., Goguet, A., Shekhtman, S., Hardacre, C., Burch, R.: On the complexity of the water-gas shift reaction mechanism over a Pt/CeO2 catalyst: effect of the temperature on the reactivity of formate surface species studied by operando DRIFT during isotopic transient at chemical steady-state. Catal. Today. 126, 143–147 (2007)

41

964 180. Olympiou, G.G., Kalamaras, C.M., Zeinalipour-Yazdi, C.D., Efstathiou, A.M.: Mechanistic aspects of the water-gas shift reaction on alumina-supported noble metal catalysts : in situ DRIFTS and SSITKA-mass Spectrometry Studies, Catal. Today. 127, 304–318 (2007) 181. Daly, H., Ni, J., Thompsett, D., Meunier, F.C.: On the usefulness of carbon isotopic exchange for the operando analysis of metalcarbonyl bands by IR over ceria-containing catalysts. J. Catal. 254, 238–243 (2008) 182. Kalamaras, C.M., Olympiou, G.G., Efstathiou, A.M.: The watergas shift reaction on Pt/gamma-Al2O3 catalyst: operando SSITKADRIFTS-mass spectroscopy studies. Catal. Today. 138, 228–234 (2008) 183. Kalamaras, C.M., Panagiotopoulou, P., Kondarides, D.I., Efstathiou, A.M.: Kinetic and mechanistic studies of the watergas shift reaction on Pt/TiO2 catalyst. J. Catal. 264, 117–129 (2009) 184. Kalamaras, C.M., Americanou, S., Efstathiou, A.M.: “Redox” vs “associative formate with -OH group regeneration” WGS reaction mechanism on Pt/CeO2: effect of platinum particle size. J. Catal. 279, 287–300 (2011) 185. Kalamaras, C.M., Dionysiou, D.D., Efstathiou, A.M.: Mechanistic studies of the water-gas shift reaction over Pt/CexZr1–xO2 catalysts: the effect of Pt particle size and Zr dopant. ACS Catal. 2, 2729–2742 (2012) 186. Shekhar, M., Wang, J., Lee, W.-S., Akatay, M.C., Stach, E.A., Delgass, W.N., Ribeiro, F.H.: Counting Au catalytic sites for the water-gas shift reaction. J. Catal. 293, 94–102 (2012) 187. Wang, J., Kispersky, V.F., Delgass, W.N., Ribeiro, F.H.: Determination of the Au active site and surface active species via operando transmission FTIR and isotopic transient experiments on 2.3 Wt.% Au/TiO2 for the WGS reaction. J. Catal. 289, 171–178 (2012) 188. Costa, C.N., Efstathiou, A.M.: Transient isotopic kinetic study of the NO/H-2/O-2 (lean de-NOx) reaction on Pt/SiO2 and Pt/La-CeMn-O catalysts. J. Phys. Chem. B. 108, 2620–2630 (2004) 189. Anastasiadou, T., Loukatzikou, L.A., Costa, C.N., Efstathiou, A.M.: Understanding the synergistic catalytic effect between La2O3 and CaO for the CH4 lean de-NOx reaction: kinetic and mechanistic studies. J. Phys. Chem. B. 109, 13693–13703 (2005) 190. Haneda, M., Kintaichi, Y., Hamada, H.: Reaction mechanism of no decomposition over alkali metal-doped cobalt oxide catalysts. Appl. Catal. B Environ. 55, 169–175 (2005) 191. Haneda, M., Nakamura, I., Fujitani, T., Hamada, H.: Catalytic active site for NO decomposition elucidated by surface science and real catalyst. Catal. Surv. Jpn. 9, 207–215 (2005) 192. Lambrou, P.S., Efstathiou, A.M.: The effects of Fe on the oxygen storage and release properties of model Pd-Rh/CeO2-Al2O3 threeway catalyst. J. Catal. 240, 182–193 (2006) 193. Cheung, P., Bhan, A., Sunley, G.J., Law, D.J., Iglesia, E.: Site requirements and elementary steps in dimethyl ether carbonylation catalyzed by acidic zeolites. J. Catal. 245, 110–123 (2007) 194. Savva, P.G., Costa, C.N., Efstathiou, A.M.: The mechanism of reduction of NO with H-2 in strongly oxidizing conditions (H-2Scr) on a novel Pt/Mgo-CeO2 catalyst: effects of reaction temperature. Kinet. Catal. 49, 743–747 (2008) 195. Savva, P.G., Efstathiou, A.M.: The influence of reaction temperature on the chemical structure and surface concentration of active NOx in H-2-SCR over Pt/MgO-CeO2: SSITKA-DRIFTS and transient mass spectrometry studies. J. Catal. 257, 324–333 (2008) 196. Klukowski, D., Balle, P., Geiger, B., Wagloehner, S., Kureti, S., Kimmerle, B., Baiker, A., Grunwaldt, J.D.: On the mechanism of the SCR reaction on Fe/HBEA zeolite. Appl. Catal. B Environ. 93, 185–193 (2009) 197. Chansai, S., Burch, R., Hardacre, C., Breen, J., Meunier, F.: The use of short time-on-stream in situ spectroscopic transient kinetic isotope techniques to investigate the mechanism of hydrocarbon selective catalytic reduction (HC-SCR) of NOx at low temperatures. J. Catal. 281, 98–105 (2011)

A. Holmen et al. 198. Verykios, X.E.: Mechanistic aspects of the reaction of CO2 reforming of methane over Rh/Al2O3 catalyst. Appl. Catal. A Gen. 255, 101–111 (2003) 199. Stevens, R.W., Chuang, S.S.C.: In situ IR study of transient CO2 reforming of CH4 over Rh/Al2O3. J. Phys. Chem. B. 108, 696–703 (2004) 200. Bobin, A.S., et al.: Mechanism of CH4 dry reforming on nanocrystalline doped ceria-zirconia with supported Pt, Ru, Ni, and Ni-Ru. Top. Catal. 56, 958–968 (2013) 201. Balakos, M.W., Chuang, S.S.C., Srinivas, G.: Transient infrared study of methanation and ethylene hydroformylation over Rh/SiO2 catalysts. J. Catal. 140, 281–285 (1993) 202. Srinivas, G., Chuang, S.S.C., Balakos, M.W.: An insitu IR study coupled with transient kinetic-analysis of hydroformylation. AICHE J. 39, 530–532 (1993) 203. Balakos, M.W., Chuang, S.S.C.: Dynamic and LHHW kineticanalysis of heterogeneous catalytic hydroformylation. J. Catal. 151, 266–278 (1995) 204. Chuang, S.S.C., Brundage, M.A., Balakos, M.W.: Mechanistic study in catalysis using dynamic and isotopic transient infrared spectroscopy: Co/H-2/C2H4 reaction on Mn-Rh/SiO2. Appl. Catal. A Gen. 151, 333–354 (1997) 205. Brundage, M.A., Balakos, M.W., Chuang, S.S.C.: LHHW and PSSA kinetic analysis of rates and adsorbate coverages in Co/H2/C2H4 reactions on Mn-Rh/SiO2. J. Catal. 173, 122–133 (1998) 206. Hammache, S., Goodwin, J.G.: Characteristics of the active sites on sulfated zirconia for N-butane isomerization. J. Catal. 218, 258–266 (2003) 207. Theologides, C.P., Olympiou, G.G., Savva, P.G., Pantelidou, N.A., Constantinou, B.K., Chatziiona, V.K., Valanidou, L.Y., Piskopianou, C.T., Costa, C.N.: Novel catalytic and mechanistic studies on wastewater denitrification with hydrogen. Water Sci. Technol. 69, 680–686 (2014) 208. van Dijk, H.A.J., Hoebink, J.H.B.J., Schouten, J.C.: Steady-state isotopic transient kinetic analysis of the Fischer–Tropsch synthesis reaction over cobalt-based catalysts. Chem. Eng. Sci. 56, 1211–1219 (2001) 209. Gao, J., Mo, X., Goodwin, J.G.: Relationships between oxygenate and hydrocarbon formation during CO hydrogenation on Rh/SiO2: use of multiproduct SSITKA. J. Catal. 275, 211–217 (2010) 210. Bal’zhinimaev, B.S., Sadovskaya, E.M., Suknev, A.P.: Transient isotopic kinetics study to investigate reaction mechanisms. Chem. Eng. J. 154, 2–8 (2009) 211. Govender, N.S., Botes, F.G., de Croon, M.H.J.M., Schouten, J.C.: Mechanistic pathway for methane formation over an iron-based catalyst. J. Catal. 260, 254–261 (2008) 212. van Dijk, H.A.J., Hoebink, J., Schouten, J.C.: A mechanistic study of the Fischer-Tropsch synthesis using transient isotopic tracing. Part-1: model identification and discrimination. Top. Catal. 26, 111–119 (2003) 213. Frøseth, V., Storsæter, S., Borg, Ø., Blekkan, E.A., Rønning, M., Holmen, A.: Steady state isotopic transient kinetic analysis (SSITKA) of CO hydrogenation on different Co catalysts. Appl. Catal. A Gen. 289, 10–15 (2005) 214. Yang, J., Frøseth, V., Chen, D., Holmen, A.: Particle size effect for cobalt Fischer–Tropsch catalysts based on in situ CO chemisorption. Surf. Sci. 648, 67–73 (2016) 215. Kalamaras, C.M., Gonzalez, I.D., Navarro, R.M., Fierro, J.L.G., Efstathiou, A.M.: Effects of reaction temperature and support composition on the mechanism of water-gas shift reaction over supported-Pt catalysts. J. Phys. Chem. C. 115, 11595–11610 (2011) 216. Jacobs, G., Davis, B.H.: Reverse water-gas shift reaction: steady state isotope switching study of the reverse water-gas shift reaction using in situ DRIFTS and a Pt/ceria catalyst. Appl. Catal. A Gen. 284, 31–38 (2005)

41

Steady-State Isotopic Transient Kinetic Analysis (SSITKA)

965

217. Jacobs, G., Graham, U.M., Chenu, E., Patterson, P.M., Dozier, A., Davis, B.H.: Low-temperature water-gas shift: impact of Pt promoter loading on the partial reduction of ceria and consequences for catalyst design. J. Catal. 229, 499–512 (2005)

41

Anders Holmen is Professor Emeritus of Chemical Engineering at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway. He worked as a research scientist and a group leader at SINTEF, Applied Chemistry before he joined the Faculty of Chemical Technology, Norwegian Institute of Technology in 1981. He was appointed Professor of Industrial Chemistry in 1985. He has been a visiting scientist/professor at Stanford University, USA, the Institute de Recherché sur la Catalyse in Villeurbanne, France, the University of Bologna, Italy, and the Research Centres of Norsk Hydro, Porsgrunn and Statoil, Trondheim. He received Statoil’s Research Price in 1994 and the Award for Excellence in Natural Gas Conversion in Lyon in 2010. He has been the main supervisor for 50 Ph.D. students and 179 M. Sc. students. He has published about 300 articles in international journals.The main subjects of his teaching courses have been Heterogeneous Catalysis and Chemical Reaction Engineering.

Jia Yang has been an Associate Professor at the Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU), since 2016. She received her Ph.D. in Heterogeneous Catalysts from NTNU in 2011. She carried out postdoctoral research in the same field and then worked as a research scientist at SINTEF for three years before joining NTNU in 2016. Her main research interests include Fischer–Tropsch synthesis, biomass conversion to fuels and chemicals, electrochemical CO2 reduction, fuel cell catalysts, and photocatalytic H2 generation. She has strong expertise in mechanistic investigation with Steady-State Isotopic Transient Kinetic Analysis (SSITKA) and isotopically labeled gases.

Dr. De Chen is Professor of catalysis at the Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU) since 2001 (associate professor 1998–2001). He earned his Ph.D. in industrial catalysis at NTNU, Norway, in 1998. He was a visiting professor at the University of California at Berkeley (2009– 2010) and East China University of Science and Technology (2017– 2018). His research is mainly on a multiscale approach at the interface between catalysis science and industrial chemical processes. His work on combined theoretic and experimental heterogeneous catalysis has in several instances led to the development of new catalysts for gas to liquids, monomer production for polyvinyl chloride (PVC), biomass to liquids, natural gas to olefins, hydrogen production and fuels, as well as materials for CO2 capture technologies, plastic waste recycling, and energy storages. He is a member of the Norwegian Academy of Technological Science and Royal Norwegian Academy of Sciences and Letters. He is the director of innovation hub of upcycling of wastes, a member of the leader group at the national innovation center (iCSI), and FME center of biomass for fuels (Bio4Fuels). He well-published more than 450 scientific papers in peer-reviewed journals (H-index:79) and more than 10 patents.

Modulation Excitation Spectroscopy (MES) Atsushi Urakawa

, Davide Ferri

42

, and Rob Jeremiah G. Nuguid

Contents 42.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 967

42.2

Modulation and Phase-Sensitive Detection . . . . . . . . . . . . . . 968

42.3

Use and Interpretation of Phase-Resolved Data . . . . . . . . 971

42.4

Summary and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 975

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 975

Abstract

In spectroscopy and diffraction methods, the signatures of catalytically active sites are often submerged by the contribution of spectator species. In some cases, the signals may also superimpose with each other, hindering proper peak identification. Rationalizing a reaction pathway becomes very challenging, if not impossible, under these circumstances. Accordingly, the implementation of transient, dynamic methods such as modulation/modulated excitation spectroscopy (MES) can improve the analysis of these signals. In MES, the catalyst sample is subjected to periodic changes in the environment (e.g., reactant concentration) that stimulate the active species periodically while spectra or diffractograms are recorded with sufficient time resolution. Combined with phase-sensitive A. Urakawa (*) Catalysis Engineering, Department of Chemical Engineering, Delft University of Technology, Delft, The Netherlands e-mail: [email protected] D. Ferri Energy and Environment Research Division, Paul Scherrer Institut, Villigen, Switzerland e-mail: [email protected] R. J. G. Nuguid Energy and Environment Research Division, Paul Scherrer Institut, Villigen, Switzerland École polytechnique fédérale de Lausanne (EPFL), Institute for Chemical Sciences and Engineering, Lausanne, Switzerland e-mail: [email protected]

detection (PSD) analysis, this approach selectively enhances the signals of the responsive (and possibly active) species and at the same time attenuates the contribution from the spectator and static species. Overall, this results in increased analytical capabilities irrespective of the type of spectroscopy or diffraction technique that is used for the experiment. In this chapter, we introduce the basic concepts of MES, discuss the theory of PSD, and provide general guidelines that are useful for whoever encounters PSD data for the first time. Keywords

Modulated/modulation excitation spectroscopy · Phasesensitive detection · In situ · Operando · Spectroscopy · Diffraction

42.1

Introduction

In recent years, operando spectroscopy contributed to shedding light on the “black box” enclosing the mechanism of catalytic processes [1]. Through our knowledge of lightmatter interaction, we have gained access to the molecular events occurring on the catalyst surface that was not possible a few decades ago. As a result, the rational design of catalytic materials is starting to become more realistic than ever, and the possibility to perform it is increasingly in our hands. Despite its great successes, however, modern spectroscopy still faces a number of limitations when applied to heterogeneous catalysis research due to the inherent complexity of solid catalytic systems. The prevailing difficulties stem from the following: 1. Catalysis results from simultaneous and often superimposed chemical phenomena [2]. These include reactant adsorption, surface reaction, and product desorption. As a consequence, the arising spectral signal will be a

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_42

967

968

convolution of all of these processes. In some cases, one of the processes, thus its spectral features, will predominate over the others in a steady-state process. Furthermore, the spectroscopic signatures of different chemical species may have similar energy values and, in practice, appear as overlapping signals. For instance, Lewis-bound NH3, water, and nitrate species possess vibrational signature in the infrared (IR) spectrum at around 1620 cm
1. The resulting observed peak is then the result of the overlap of the signals of these three species, with no straightforward way of deconvolution [3]. 2. In commercial catalyst formulations, the active phase is present in a much lower concentration in comparison with the other catalyst components (e.g., support, promoter, and binder) in order to maximize the atom efficiency of the active phase, which is usually more expensive and not Earth-abundant. Conventional spectral features are therefore dominated by the signal originating from species coordinated to the support, while the active phase, which is the catalytically relevant component, can only contribute weakly because of its low concentration and the occurring turnover. In addition, the active phase itself can be further divided into catalytically relevant and spectator/ unresponsive species. Normally, the spectator species predominate in number, and it remains a challenge to extract the signal contribution from the real species responsible for the activity [4]. 3. Reactive intermediates are short-lived species owing to their thermodynamic instability. In rare instances, intermediate species are sufficiently stable that they can be isolated and characterized extensively. This is especially true for some organic reactions [5, 6], but it is not the norm for inorganic reactions and those that occur via a radical mechanism. The transient presence of intermediates in the reaction is further complicated by the fact that, compared to the reactants and products, only a very small fraction of the intermediates is present at any given time and thus will most likely be obscured by the signal contributed by the reactants and products. Difference spectroscopy may partially solve the problem of separating the signal of the active phase from that of the support [7], but it fails to address the other issues. Chemometric methods can be employed to resolve complex envelopes of superimposed signals and are increasingly gaining popularity, but they normally require large datasets and computing power. Alternatively, a different experimental approach may be implemented. First described in a seminal work more than two decades ago [8], modulated excitation spectroscopy (MES, as it is also often called, modulation excitation spectroscopy) holds the potential of addressing all of the aforementioned difficulties.

A. Urakawa et al.

42.2

Modulation and Phase-Sensitive Detection

Selective and sensitive detection of signals containing relevant information on catalytic reactions, such as catalytic active site/material and active chemical species, is of great challenge as described above in the limitations of general spectroscopic methods. MES is a transient response technique. It makes use of periodic perturbation of a reaction system with external parameter(s) called stimulus (or stimulation) [2, 8]. The stimulus is chosen in a way that the population of what we wish to monitor, e.g., the active catalyst structure or active chemical species, is affected most exclusively by the perturbation. Through the exclusive disturbance of particular species/structure and their subsequent detection, one can add selectivity to the detection. In principle, there is no limitation on the types of stimulus, and there are many examples such as fluid (gas/liquid) concentration [9, 10], temperature [11], electric field [12], light flux [13], and X-ray energy [14]. One can choose the most appropriate one depending on what one wishes to see and also on experimental convenience and restrictions. By far the most common stimulus type is fluid concentration, because the extent of a chemical reaction and the concentrations of active structure/species can be more easily modulated using a concentration stimulus [15]. The experimental protocol and data processing procedures are shown in Fig. 42.1. Let us assume that we use a periodic sinusoidal perturbation of a catalytic system by an external parameter as the stimulus (e.g., reactant concentration, Fig. 42.1a). At the start of such periodic perturbation experiments, typically there is an initial period when the chemical system changes irreversibly to some extent and reaches a stably oscillating period where the mean value is constant (quasi steady state, Fig. 42.1a). It should be noted that there is always some noise in the signal and the effects of noise are more significant when the time resolution of our measurements is increased (i.e., fast spectral acquisition). Figure 42.1 shows the stimulus and signal responses with lines, but in a typical spectroscopic measurement, we record signal responses as a function of a range of energy. The signal intensity shown in Fig. 42.1 represents a response only at one energy point, while in practice, there are distinct signal responses at every energy point recorded in a set of spectra (e.g., if IR spectroscopy is used and the spectra are recorded in the range of 1000–4000 cm
1 at the spectral resolution of 1 cm
1, there are 3001 signal intensity responses in the recorded time-resolved data). For simplicity, here only one signal response at one energy point is described. The first data processing generally performed in MES is a simple averaging of the signal over the cycles into one period after reaching a so-called quasi steady state after the initial equilibration period (Fig. 42.1b). According to Poisson

42

Modulation Excitation Spectroscopy (MES)

a)

969

b)

Active species/structure response (A)

c)

Astatic +

Signal intensity

Aave =

Stimulus (S)

Areal

j

Equilibration Quasi steady-state (s.s.)

= Aamplitude sin(w t + j)

+

Averaging over quasi s.s.

Anoise S = Samplitude sin(w t) 0

Time

e)

1/2T T Time (T)

d)

A (f PSD) = Aamplitude cos(j – f PSD)

Phase sensitive detection (PSD) Ak (f kPSD ) =

S (f PSD) = Samplitude cos(f PSD) 0

180

at k = 1

T

2 T

Aave sin(kw t + f kPSD )dt 0

360

f PSD (°)

Fig. 42.1 (a) Sinusoidal stimulus of an external parameter and a representative response of active species/structures affected by the stimulus; (b) the one-period response obtained after averaging over the multiple modulation cycles after reaching quasi steady state; (c) the three signal elements consisting of the actual response (Aave), namely

static signal (Astatic), actual signal of interest (Areal), and noise (Anoise); (d) the mathematical engine of MES, phase-sensitive detection (PSD); and (e) the phase-domain response after PSD at the fundamental stimulus frequency demodulation (k ¼ 1)

statistics, this reduces the data size and improves pffiffiffiffi sensitivity as well as the signal-to-noise ratio by about N . (N is the number of averaged cycles.) Actively responding species mainly respond at the same frequency as that of the stimulus, and the signal intensity oscillates in a sinusoidal manner but with some delay (φ) with respect to the stimulus (Fig. 42.1b). The magnitude of the delay is dependent on the kinetics of the underlying physicochemical processes (e.g., sorption, reaction, convection, and diffusion) and consequently the delay changes with the modulation frequency of the stimulus (ω). In other words, the frequency of the stimulus affects the magnitude of the delay, and kinetic studies can be performed by varying the stimulus frequency. One can study the averaged response and its delay (Fig. 42.1b) to understand catalytic processes. However, the relative intensity of the signal actively changing in response to the stimulus is often extremely small and can be even at the level of noise. One can improve the signal-to-noise ratio by increasing the number of cycles (N ), but this increases the burden in the experiments and data processing. In principle, the averaged signal (Aave) consists of three signal elements (Fig. 42.1c): (i) static signal (Astatic), (ii) actual signal of interest (Areal), and (iii) noise (Anoise). Astatic may arise from catalyst support material, solvent, and inactive spectator

species, often dominating the detected signal. They are all important for the reaction one way or the other, but to understand the catalytic reactions, their dominant signal intensity should be reduced or eliminated completely. Areal is obviously the one we wish to extract to understand its chemical origin. Anoise is always present for any detection method, and typically it contains high-frequency elements with respect to the stimulus frequency. What we wish to do is to extract only Areal out of Aave, and this is exactly what is done in MES by its mathematical processing, called phase-sensitive detection (PSD, Fig. 42.1d) [8]. PSD is the mathematical treatment often implemented in the hardware in signal processing such as lock-in amplifiers. On the other hand, the lock-in (i.e., extracting the signal at the same frequency as that of the stimulus, k ¼ 1 in Fig. 42.1d) is performed numerically in MES, and this has major advantages over the hardware lock-in due to the richer information contained (i.e., less oversimplification of the data). What PSD does mathematically is presented by the equation in Fig. 42.1d. In a nutshell, the averaged signal (Aave) is multiplied by another sine function at the frequency of kω. The sine function also contains a new variable ϕPSD which is k called phase angle. Generally, the multiplied sine function has the fundamental frequency of the stimulus (ω), meaning

42

970

that k ¼ 1 (if k ¼ 1, k is often not written). The thus obtained response is integrated over the period (T ) and then normalized by multiplying with 2/T (Fig. 42.1d). Interestingly, the resulting response after PSD is not a function of time any longer but of ϕPSD(Fig. 42.1e). PSD is also called demodulation, and k is sometimes called the demodulation index. Let us evaluate what PSD does more closely. Regarding the stimulus, Samplitude sin (ωt) is transformed by PSD to a cosine function of the same amplitude (Samplitude cos (ϕPSD)) where the domain changes from time (t) to phase angle (ϕPSD). For the actively responding signal, the domain also changes and Aamplitude cos (φ
ϕPSD) is obtained after PSD. This phase-domain signal is a very similar sinusoidal function containing the delay term (φ) with the same amplitude as that of the signal we wish to extract (Areal) which is Aamplitude sin (ωt þ φ). Most importantly, in the phase domain, the static signal (Astatic) is completely vanished and so is the noise (Anoise). Precisely, only the noise signal varying at the same frequency as kω will appear in the phase domain, but this is negligible because high-frequency noise is generally dominating. Hence, what we can achieve with MES and especially by its mathematical engine PSD is that we can add selectivity to the measurement and boost its detection sensitivity by removing the static signals and reducing the noise while retaining the key information of interest. After PSD and the initial averaging process (Fig. 42.1b), in MES, one can often achieve 2–3 orders of magnitude improvement in signal-to-noise. This drastic improvement in the sensitivity is extremely useful or even necessary to study the weak and transiently present signals in catalytic reactions. The mathematical expression for the active sites/species (Aamplitude cos (φ
ϕPSD), Fig. 42.1e), shows that the amplitude of the response (Aamplitude) is at its maximum when φ
ϕPSD ¼ 0, i.e., ϕPSD is equal to φ, which is called the “inphase” condition. This “in-phase” condition has interesting and useful implications. From the phase-domain response, one can easily check when (at which phase angle) the signal becomes maximum by varying ϕPSD, and this value is directly related to the delay of the active species/structure. If we have multiple species and thus multiple bands in the time-resolved spectra with different kinetic responses, these bands will show maximum value in the phase domain at different ϕPSD. By converting in-phase ϕPSD to φ for different bands (e.g., different chemical species/structures), one can understand what is happening first and gain insights into transformation pathways. This analysis is called in-phase angle analysis [2, 16]. So far, a sinusoidal shape of stimulus has been assumed. In practice, this shape is often not the most convenient one to generate and, instead, a square-wave shaped stimulus is often used because such shapes can be more easily generated, for example, by valve switching for concentration modulation and by light on-off or by means of a chopper for light flux

A. Urakawa et al.

modulation. One may wonder about the influence of the stimulus shape on the data analysis procedure and on the quantitative nature of such analysis. More precisely, the meaning of the amplitudes and delays of phase-domain signals when a square-wave stimulus is used can be questioned, compared to the case of the unaltered amplitude of the active species/structure signals after PSD when a sinusoidal stimulus is used (Fig. 42.1). Within the linear response approximation, the analysis procedure has been well established and shows even advantages of a square-wave stimulus compared to a sine-wave one [2, 17]. Mathematically, A square-wave (SW(t)) with an amplitude of ASW can be written as the sum of odd-frequency sine functions with their amplitude scaled by the factor of 4/π and with more reducing amplitudes at higher frequencies.   4 sin 3ωt sin 5ωt SWðtÞ ¼ ASW sin ωt þ þ þ ... π 3 5 1 X sin ½ð2n
1Þωt  4 ¼ ASW 2n
1 π n¼1 ð42:1Þ Clearly, a square-wave contains the sine wave of the fundamental frequency besides all other sine waves of odd frequencies. In other words, with a square-wave stimulus, the system is perturbed not only by the fundamental frequency (ω) but also by the many higher frequencies (odd-frequency harmonics: 3ω, 5ω¸ 7ω¸. . .). The beauty of PSD is that by changing the demodulation index k in the PSD equation (Fig. 42.1d), we can selectively extract the response varying at kω frequency, which can also be performed conveniently by Fourier transform [18]. It has been shown that the signal delay φk (in the phase domain, it is also called phase delay) obtained by the high-frequency demodulation (PSD with k > 1) is identical to that obtained with the MES experiment at the higher stimulus frequency as the fundamental frequency (e.g., MES experiment at (2n
1)ω, n defined in Eq. 42.1). The same also holds for the response amplitude, but to convert the amplitude to a comparable scale, one needs to multiply the demodulated signal amplitude by 2n
1 (Eq. 42.1) after high-frequency demodulation since the amplitudes of the higher frequency terms are smaller by 2n
1 (Eq. 42.1). Importantly, this means that one square-wave stimulus experiment is equivalent to multiple sine-wave stimulus experiments, facilitating to understand the kinetic responses of the amplitude and delay of active signals. It should be remembered that this is precise and exact only when the responses are linear. In practice, one can obtain a more linear response by reducing the stimulus amplitude [17]. In this case, however, the amplitude of the higher frequency terms is also reduced, and the signal-to-noise ratio may not be high enough for reliable analysis.

42

Modulation Excitation Spectroscopy (MES)

42.3

971

Use and Interpretation of Phase-Resolved Data

The increasing number of reports on the use of MES and PSD to elucidate mechanisms of catalyzed reactions calls for an introduction into the use, analysis, and utility of the phaseresolved data. In order to keep the discussion broad enough, we consider two general signals close to each other (Fig. 42.2a). The abscissa is generic and could be mute, because the idea is to discuss any type of spectroscopy and diffraction experiment performed under the paradigm of the MES methodology. Moreover, it is clear from the PSD equation (Fig. 42.1d) that the formula is valid irrespective of the method that is used to interrogate the sample. This is

a)

A

demonstrated by the increasing variety of analytical methods that have been coupled to the modulation approach in recent years (ATR-IR, DRIFTS, PM-IRRAS, Raman, XAS, XRD, XES, HEROS, and PDF) [15]. For the sake of clarity, we will refer to the spectra as the data in Fig. 42.2a–d. Figure 42.2a, e represents a common graph that is encountered in most works making use of MES and PSD: timeresolved spectra (Fig. 42.2a) are presented together with their corresponding phase-resolved data (Fig. 42.2e) obtained at the fundamental demodulation frequency k ¼ 1. In Fig. 42.2e, only half of the phase-resolved data upon PSD is presented. The other half of the data is a mirror image of these ones, as it follows from the sinusoidal phase angular dependence of signals. While the information delivered by

b)

c)

d)

f)

g)

h)

1.0 Intensity (–)

0.8 0.6 B 0.4 0.2 0.0

e) 0.6 Intensity (–)

0.4 0.2 0.0 –0.2 –0.4 –0.6 0

i)

10

20 30 40 Data point (-)

50

0

j)

10

20 30 40 Data point (-)

50

0

k)

10 20 30 40 50 60 Data point (-)

0

l)

10 20 30 40 50 60 Data point (-)

Intensity (–)

0.8 0.4 0.0 –0.4 –0.8

0

60 120 180 240 300 360 Phase angle (°)

0

60 120 180 240 300 360 Phase angle (°)

Fig. 42.2 (a–d) Simulated data sets of two generic signals A and B in modulated excitation experiments in which (a) A does not vary, (b) A and B vary with identical kinetics, (c) A is retarded compared to B, and (d) A is present when B is absent (i.e., opposite kinetics). (e–h) Corresponding phase-resolved data obtained at the fundamental

0

60 120 180 240 300 360 Phase angle (°)

0

60 120 180 240 300 360 Phase angle (°)

frequency of stimulus (k ¼ 1). Only half of the full set of phase-resolved data is shown for simplicity. Green data are used to guide the eye and show relevant traces. (i–l) Phase angular dependence of A and B at the maximum peak position. Red data in (a–d) represents the first halfperiod, blue data the second half-period

42

972

a) Intensity (–)

1.0

b) A

0.8 0.6 B

0.4 0.2 0.0

c)

d)

1.0 Intensity (–)

comparison of Fig. 42.2a, e is inherently sufficient to claim that the signals appearing in the phase-resolved data belong to species responding to the external stimulus, there are a number of further actions that can be taken to analyze the phase-resolved data and take full advantage of MES. Let us discuss the appearance of the phase-resolved data. Several sub-cases can be encountered that depend on how two signals behave with respect to each other along the time axis. We will discuss the following sub-cases: (a) one signal that we label as A remains unchanged, and the second (B) varies following a given kinetics of consumption and formation; (b) both signals vary following identical kinetics (synchronous); (c) both signals vary following different kinetics; and finally, (d), when A is present, B is not and vice versa (the signals vary with opposite phase; asynchronous). Figure 42.2a–d displays the sets of time-resolved spectra during one modulation period in which either B or both signals vary as indicated just above. Figure 42.3a–d presents the corresponding simulated temporal dependence of the two signals A and B in that modulation period. Ideally, the dataset of Fig. 42.2a–d is the result of averaging of e.g., n time-resolved spectra (in Fig. 42.3, n ¼ 40) obtained in m modulation periods to a single modulation period, which we have shown above contributes to increase the signal-tonoise ratio of the measurement significantly [19–21]. Figure 42.2e–h show the corresponding phase-resolved dataset for each case obtained from application of the PSD equation using k ¼ 1 and a half-period length T/2 (i.e., a half range of the whole ϕPSD). This PSD at k ¼ 1 is very common for most characterization techniques enabling such measurements; however, for X-ray diffraction, it was demonstrated that active responses appear at k ¼ 2 due to the fact that the detected intensity in XRD is the square of the scattered energy, leading to frequency doubling [22]. It should be noted that the maximum intensity of the signals in the phase-resolved data is typically scaled compared to the intensity of the changes in the time-resolved data (in the case of Fig. 42.2, the scaling factor is ca. 1.6; this scaling factor can be explained more precisely by considering that (i) the amplitude in the time domain is ca. 0.2 (Fig. 42.2a, b) and not ca. 0.4 since the amplitude change from the mean value is considered as in a sinusoidal wave); (ii) the amplitude obtained after PSD at k ¼ 1 needs to be scaled by 4/π because of the square-wave stimulus and consequent scaling factor (Eq. 42.1); and (iii) 0.4/(0.2∙4/π) yields roughly 1.6 [23, 20]. In Fig. 42.2a, only signal B varies while signal A remains constant throughout the modulation period. The corresponding phase-resolved data shows only the contribution of signal B, because signal A does not change in intensity along the experiment and such signals are fully vanished by PSD (Fig. 42.1). The intensity of signal B changes from a maximum value to a minimum negative value following a sinusoidal function (Fig. 42.2i). The dataset of Fig. 42.2e is

A. Urakawa et al.

0.8 0.6 0.4 0.2 0.0 0

10

30 20 Time (–)

40 0

10

20 30 Time (–)

40

Fig. 42.3 Simulated temporal dependences of generic signals A and B. PSD of the corresponding time-resolved delivers the data shown in Fig. 42.2

composed of 18 spectra in the range of phase angles ϕPSD of 0–180 . In the 190–360 range, the same set of spectra is obtained but of the opposite sign, following the sinusoidal phase angular dependence of signal intensity shown in Fig. 42.2i. This data set is shown with intervals of ϕPSD ¼ 10 , but more precise intervals of 5, 2, or 1 can be obtained that might be required to better resolve kinetic behaviors of signals in phase-resolved data [24]. Consider then that also signal A varies during the modulation experiment and that the intensity of both signals exhibits the same kinetics of decay and evolution as shown in Fig. 42.3b. The two signals are set to exhibit different intensities. Figure 42.2f shows the phase-resolved spectra of this experiment. As in the previous case, the intensity of both signals varies along a sinusoidal function (Fig. 42.2j) and the two signals are in-phase as the maximum intensity reached by signal A corresponds to the maximum intensity obtained for signal B, but the two maxima do not correspond in absolute value. The phase angular dependence of the intensity of the two signals coincides (Fig. 42.2j); both signals cross the x-axis at the same points (phase angle, ϕPSD) and rise to the maximum and minimum values at the same points. This case is, for example, that very simple of an adsorbed species whose vibrational spectrum is composed of two signals A and B of different intensities: when the species appears in the experiment, both signals grow simultaneously when the species vanishes, both signals disappear with the same kinetics. This case could also be that of two species, one assigned to signal A and one to signal B, that display the same kinetic behavior.

42

Modulation Excitation Spectroscopy (MES)

A more complex case is that obtained when the two signals exhibit different kinetics but they still disappear/ appear in the same half-period. Keeping the situation of two signals of different intensities, Fig. 42.2c shows very similar data to those of Fig. 42.2b. However, the kinetics of decay and growth clearly are not equal, implying that the phaseresolved data intersect. The phase-resolved data are also very similar to those shown in Fig. 42.2f. Here, the details are important. To explain, let us consider the phase angular dependence (Fig. 42.2k). Contrary to the previous case, the two sinusoids do not overlap; rather their maxima are shifted indicating a delay in the evolution of the two signals. This shift also reveals that, for example, at ϕPSD ¼ 60 , signal A is absent in the phase-resolved spectrum, because its intensity is nil. Similarly, at ϕPSD ¼ 80 , signal B intersects the x-axis and will not be present in the corresponding phase-resolved spectrum. This demonstrates that the spectra at these two ϕPSD values will display only the species corresponding to signal B and the species assigned to signal A, respectively. Therefore, spectra containing only the signal(s) of one species can be obtained by PSD and can be selected [2]. This analysis is the kinetic analysis of phase-resolved data. This may deliver the same results as the analysis of the timeresolved spectra as shown in Fig. 42.3 but could allow improving the differentiation of signals of different species that is sometimes not possible in the time-dependent dataset. The other characteristic situation is shown in Fig. 42.2d, where the two signals are retarded by a half-period: when A is present, B is not, and vice versa in the subsequent halfperiod. This case may correspond to the situation when one species transforms into the other one, e.g., A transforms into B reversibly during the experiment. The corresponding phase-resolved data (Fig. 42.2h) always show lobes of the opposite sign. The phase angular dependence of the signals in Fig. 42.2l shows that signal A is maximum when signal B is minimum (and vice versa) and that both signals cross the x-axis at the same intersection points. Clearly, the cases in Fig. 42.2a, c, and d represent two signals belonging to two different species. The above analysis suggests that rather than all phase-resolved data, a careful selection of the phase-resolved data can improve our knowledge about the behavior of the system. While the analysis presented in Fig. 42.2 is relatively straightforward in the case of vibrational spectra [17, 25], it is particularly delicate in the case of X-ray absorption spectroscopy data. Given the nature of the experiment, there are a number of points within the sequence of transformations that can be performed on XAS data, from the raw near-edge data to the Fourier transformed data, that can provide different types of information. The raw XAS data can be treated by PSD as demonstrated in Refs. [19, 20, 26]. PSD at this point provides spectra in the whole energy range selected for the measurement that are very similar to difference spectra

973

(though with significantly enhanced signal-to-noise ratio). This allows the visualization of subtle changes in oxidation state by the absorbing element, even for low levels of the absorbing element, e.g., 0.3 wt% Rh/Al2O3, hence likely for very small aggregated entities of the active phase [19]. Comparison with a difference spectrum obtained from the metal foil and a corresponding spectrum of an oxidized phase of the element (for example, PdO in the case of Pd) can support qualitative identification of additional phases. This has been, for instance, the case for the presence of Pd-C species in modulation experiments consisting of alternate pulses of CO and NO on Pd/Al2O3 [19]. Large deviations of the phase-resolved spectra in the region starting 5 eV above the Pd K-edge (E0 ¼ 24,350 eV) from the Pd-PdO difference spectrum provided qualitative information on the formation of carbide species that were reflected in the quantitative analysis of the time-resolved EXAFS spectra, suggesting expansion of the Pd-Pd bond length in the CO pulse from 2.73 to ca. 2.77 Å. The PSD of the normalized raw data cancels the contribution of the edge jump that hampers identification of small changes around the edge energy and the whiteline as well as of the predominant phase. Continuing with the example of reduced Pd/Al2O3, the contribution of metallic Pd representing the bulk of Pd nano-particles after reduction in this material is cancelled by PSD. For quantitative purposes, Chiarello et al. [23] showed that for XAS data, PSD can be carried out on the data in the k-space followed by or after Fourier transform. While the information contained in the PSD data should ideally be identical, demodulation in k-space followed by Fourier transform of the PSD data is preferable to visualize only the changes occurring on the radial distribution function. Hence, changes in the distances from neighboring atoms and in the coordination number relative to a specific coordination shell are made visible by PSD. Because PSD eliminates the contribution of species to the spectra that do not respond to the modulation stimulus, the PSD of the k-space data eliminates the contribution of the bulk phase that may dominate in the radial distribution function. This approach allowed the determination of small coordination number values from fitting the data after sequential PSD and Fourier transform, suggesting the growth of thin oxidic layers on metal particles. For example, on nano-particles of typical sizes of 2–6 nm, values for Pd-O and Pd-O-Pd were 2.0  0.2 and 1.8  0.6, respectively, for Pd/Al2O3 [23]. On larger particles (>10 nm), where the contribution of the bulk to the XAS is even larger, an average Ru-O coordination number of 0.25  0.09, a bond distance of 1.945  0.016 Å, and a pseudo-Debye-Waller factor of 0.0035  0.0035 Å2 were determined for a reduced Ru/Al2O3 catalyst subject to periodic oxidation-reduction indicating the presence of a very thin oxide layer [20]. On the contrary, the PSD data after Fourier transform contain also the information on the bulk

42

974

A. Urakawa et al.

Fig. 42.4 Sets of phase-resolved data are obtained from PSD of the time-resolved data of Fig. 42.2c using (a) k ¼ 1 (identical to Fig. 42.2g), (b) 3, and (c) 5. Bold traces are used to guide the eye and to highlight data where the two signals A and B have the opposite sign. For simplicity, only half of the full set of phaseresolved data is shown

Intensity (–)

phases [25], which overlaps strongly with the changes induced by the modulation typically overwhelming them. König et al. [24] have shown that the modulation-PSD approach provides sufficient sensitivity to detect Ru-S bonds in Al2O3-supported Ru that are believed to be located at the surface of the large Ru particles (10–30 nm) with a Ru-S coordination number of 0.07  0.02 at R ¼ 2.299  0.012. This analysis of enhanced precision was possible only upon the selection of suitable phase-resolved spectra among the obtained set, as we have described in Fig. 42.2. Given the surface nature of the heterogeneous catalytic processes, the studies using XAS combined with modulated excitation and PSD demonstrate that surface sensitivity can be bestowed on XAS through this experimental approach. We believe that this is also the case for XRD, another typical bulk characterization method. High energy XRD modulation experiments on Pd supported on ceria-zirconia demonstrated the potential of XRD in determining the redox behavior of small metal nano-particles that are otherwise not resolved by XRD [21], especially because of the contribution of the background scattering. This is completely removed by the PSD thus allowing the observation of small diffraction peaks. Marchionni et al. [27, 28] showed that reoxidation of reduced Pd/Al2O3 occurs in two stages using high-energy XRD. The phase-resolved XRD data exhibited different degrees of combination of two features of different widths in the range of the (101) reflection of PdO (2θ ¼ 32.5 ). The kinetic analysis demonstrated above, upon adequate selection of the phase-resolved data, allowed to isolate the contribution

of an amorphous phase (ϕPSD ¼ 210 ), characterized by a very broad signal, from that of a more crystalline phase (ϕPSD ¼ 70 ), characterized by an overlapping sharp signal [27]. Alternate CO and O2 pulses allowed demonstrating that Pt nano-particles in Pt/Al2O3 oscillate between fully reduced and an amorphous Pt-O layer [29]. The corresponding phaseresolved XRD data (belonging to the case shown in Fig. 42.2d) displayed sharp peaks in correspondence of Pt reflections and very broad features of the opposite sign in correspondence of the Pt oxide reflections. While most phase-resolved data are typically obtained and shown at k ¼ 1, further analysis can be performed upon demodulation using higher harmonics, typically k ¼ 3, 5, etc. [17, 25, 27]. This additional analysis is useful to distinguish between species of different nature when this differentiation is ambiguous at k ¼ 1. A detailed mathematical analysis is provided in Ref. [27] in combination with kinetic considerations. Figure 42.4 shows the same data as in Fig. 42.2c after PSD at k ¼ 1, 3 and 5. The major effect of the use of higher harmonics on the sets of phase resolved data is to decrease the overall intensity, which is more significant at k ¼ 3 (Fig. 42.4b) and attenuates above this value. This is the case because the contribution of higher harmonics to the approximation of a square-wave stimulus decreases with increasing k [2]. However, Fig. 42.4 also shows that the decrease in intensity with increasing k occurs faster for signal A, the A:B ratio decreasing in the order of 2.45 (k ¼ 1) > 1.60 (3) > 1.17 (5). Hence, in this process, a simultaneous enhancement in the response of signals whose

a)

b)

c)

0.6

0.15

0.15

0.4

0.10

0.10

0.2

0.05

0.05

0.0

0.00

0.00

–0.2

–0.05

–0.05

–0.4

–0.10

–0.10

–0.15

–0.15

–0.6 0

10 20 30 40 50 Data point (-)

0

10 20 30 40 50 Data point (-)

0

10 20 30 40 50 Data point (-)

42

Modulation Excitation Spectroscopy (MES)

temporal dependence is faster is obtained relative to signals evolving more slowly. In the case considered here, the decreased A:B ratio with increasing k agrees with the temporal profiles of A and B signals in Fig. 42.3c, showing that signal B appears and disappears faster than signal A in both half-periods. It is also evident that with increasing k, it becomes easier to see differences between the kinetics of the two signals, as an increasing number of phase-resolved data exhibits signals of opposite sign with increasing k.

42.4

Summary and Outlook

Over the last few decades, advances in nano-material engineering and improvements in in situ and operando detection methodologies have significantly enhanced our understanding of catalytic materials and reaction mechanisms. At the same time, many studies have exemplified the dynamic nature of catalyst materials responding to reaction conditions (reactant/ product concentration, pressure, and temperature), highlighting the necessity for operando studies to firmly establish structure vs. activity relationships towards rational design and optimization of next generation catalysts. These technological advances, however, are often not sufficient to sense active chemical species and catalytic active sites due to their low populations compared to other inactive or irrelevant chemical species and material structures. MES allows both adding detection selectivity towards what we wish to monitor (active sites/structures) and boosting sensitivity drastically through the averaging scheme and the subsequent mathematical treatment, PSD. There are numerous application examples demonstrating the power of MES in studying catalytic reaction mechanisms [15], and this approach is increasingly gaining popularity over the past decade due to its necessity as well as its versatile character to combine with any analytical methods. The important aspect which is not covered in this chapter is the cell design to perform MES experiments. As described above, various stimulus types can be used and, among them, concentration modulation is the most common type of experiment. It is very important to design spectroscopic cells so that one can perform relevant studies (i.e., operando, with simultaneous reactivity measurement) and exert a stimulus with a desired shape (e.g., sinusoidal or square wave). Flow-through cells are commonly used for concentration modulation, and it is of critical importance to characterize the flow and mixing behavior to precisely evaluate chemical information. For reaction kinetic analysis, it is mandatory to characterize in situ and operando cells in depth by understanding convection and diffusion, thus their nonideal behavior as reactors and their influences on the signal responses [10, 30–33]. When a stimulus is created by using substrates of chemically identical nature such as isotopes [34, 35], it is possible to combine MES with detailed kinetic studies using

975

the scheme of steady-state isotopic transient kinetic analysis (SSITKA) [36] to elucidate molecular interactions, reaction mechanisms, and reaction kinetics. It is also important to highlight one major challenge with MES on disentangling overlapping bands in the phase domain. In theory, from the in-phase angles, one can differentiate kinetics of the bands with a distinct chemical nature; however, this fails largely when bands overlap because they show mixed in-phase angles of the underlying bands. It is possible to enhance the kinetic resolution of the bands with different chemical origins by making use of high-frequency demodulation (Fig. 42.4); however, it is not straightforward to obtain chemically and kinetically pure spectra by MES. To solve this problem, one can use multivariate spectral analysis such as multivariate curve resolution (MCR) [37, 38]. Recently, combining the advantages of MES for sensitivity enhancement and disentangling power of multivariate spectral analysis has been demonstrated [39, 40]. Further applications of these new methodologies, their theoretical development, as well as more insightful and detailed MES studies are expected and are necessary to maximally make use of these powerful methodologies for mechanistic and kinetic studies of catalytic systems.

References 1. Weckhuysen, B.M.: In Situ Spectroscopy of Catalysts, vol. 36. American Scientific Publishers (2004) 2. Urakawa, A., Bürgi, T., Baiker, A.: Sensitivity enhancement and dynamic behavior analysis by modulation excitation spectroscopy: principle and application in heterogeneous catalysis. Chem. Eng. Sci. 63, 4902–4909 (2008) 3. Rasmussen, S.B., Portela, R., Bazin, P., Avila, P., Banares, M.A., Daturi, M.: Transientoperandostudy on the NH3/NH4+ interplay in V-SCR monolithic catalysts. Appl Catal B. 224, 109–115 (2018) 4. Kopelent, R., van Bokhoven, J.A., Szlachetko, J., Edebeli, J., Paun, C., Nachtegaal, M., Safonova, O.V.: Catalytically active and spectator Ce3+ in Ceria-supported metal catalysts. Angew. Chem. Int. Ed. 54, 8728–8731 (2015) 5. Crandall, J.K., Conover, W.W., Komin, J.B., Machleder, W.H.: Allene epoxidation. Isolation of reactive intermediates from hindered allenes. J. Org. Chem. 39, 1723–1729 (1974) 6. Proulx, G., Bergman, R.G.: Synthesis, structures, and kinetics and mechanism of decomposition of terminal metal azide complexes: isolated intermediates in the formation of imidometal complexes from organic azides. Organometallics. 15, 684–692 (1996) 7. Grdadolnik, J.: Infrared difference spectroscopy Part I. Interpretation of the difference spectrum. Vibr Spectrosc. 31, 279–288 (2003) 8. Baurecht, D., Fringeli, U.P.: Quantitative modulated excitation Fourier transform infrared spectroscopy. Rev Sci Instr. 72, 3782–3792 (2001) 9. Bürgi, T., Baiker, A.: In situ infrared spectroscopy of catalytic solidliquid interfaces using phase-sensitive detection: enantioselective hydrogenation of a pyrone over Pd/TiO2. J. Phys. Chem. B. 106(41), 10649–10658 (2002). https://doi.org/10.1021/ jp0255987 10. Urakawa, A., Wirz, R., Bürgi, T., Baiker, A.: ATR-IR flow-through cell for concentration modulation excitation spectroscopy: diffusion

42

976 experiments and simulations. J. Phys. Chem. B. 107(47), 13061–13068 (2003) 11. Caliandro, R., Chernyshov, D., Emerich, H., Milanesio, M., Palin, L., Urakawa, A., van Beek, W., Viterbo, D.: Patterson selectivity by modulation-enhanced diffraction. J. Appl. Crystallogr. 45, 458–470 (2012). https://doi.org/10.1107/s0021889812011569 12. Schwarzott, M., Lasch, P., Baurecht, D., Naumann, D., Fringeli, U.P.: Electric field-induced changes in lipids investigated by modulated excitation FTIR spectroscopy. Biophys. J. 86(1), 285–295 (2004). https://doi.org/10.1016/s0006-3495(04)74104-1 13. Borges Ordoño, M., Yasumura, S., Glatzel, P., Urakawa, A.: Synergistic interplay of Zn and Rh-Cr promoters on Ga2O3 based photocatalysts for water splitting. Phys. Chem. Chem. Phys. 20(36), 23515–23521 (2018). https://doi.org/10.1039/c8cp03987a 14. van Beek, W., Emerich, H., Urakawa, A., Palin, L., Milanesio, M., Caliandro, R., Viterbo, D., Chernyshov, D.: Untangling diffraction intensity: modulation enhanced diffraction on ZrO2 powder. J. Appl. Crystallogr. 45, 738–747 (2012). https://doi.org/10.1107/ s0021889812018109 15. Müller, P., Hermans, I.: Applications of modulation excitation spectroscopy in heterogeneous catalysis. Ind. Eng. Chem. Res. 56, 1123–1136 (2017) 16. Urakawa, A., Van Beek, W., Monrabal-Capilla, M., GalánMascarós, J.R., Palin, L., Milanesio, M.: Combined, modulation enhanced X-ray powder diffraction and Raman spectroscopic study of structural transitions in the spin crossover material [Fe (Htrz)2(trz)] (BF4). J. Phys. Chem. C. 115(4), 1323–1329 (2011). https://doi.org/10.1021/jp107206n 17. Urakawa, A., Bürgi, T., Baiker, A.: Kinetic analysis using squarewave stimulation in modulation excitation spectroscopy: mixing property of a flow-through PM-IRRAS cell. Chem. Phys. 324, 653–658 (2006) 18. Srinivasan, P.D., Patil, B.S., Zhu, H.D., Bravo-Suarez, J.J.: Application of modulation excitation-phase sensitive detection-DRIFTS for in situ/operando characterization of heterogeneous catalysts. React. Chem. Eng. 4(5), 862–883 (2019). https://doi.org/10.1039/ c9re00011a 19. Ferri, D., Matam, S.K., Wirz, R., Eyssler, A., Korsak, O., Hug, P., Weidenkaff, A., Newton, M.A.: First steps in combining modulation excitation spectroscopy with synchronous dispersive EXAFS/ DRIFTS/mass spectrometry for in situ time resolved study of heterogeneous catalysts. PCCP. 12, 5634 (2010) 20. König, C.F.J., van Bokhoven, J.A., Schildhauer, T.J., Nachtegaal, M.: Quantitative analysis of modulated excitation X-ray absorption spectra: enhanced precision of EXAFS fitting. J. Phys. Chem. C. 116, 19857–19866 (2012) 21. Ferri, D., Newton, M.A., Di Michiel, M., Chiarello, G.L., Yoon, S., Lu, Y., Andrieux, J.: Revealing the dynamic structure of complex solid catalysts using modulated excitation X-ray diffraction. Angew. Chem. Int. Ed. 53, 8890–8894 (2014) 22. Chernyshov, D., van Beek, W., Emerich, H., Milanesio, M., Urakawa, A., Viterbo, D., Palin, L., Caliandro, R.: Kinematic diffraction on a structure with periodically varying scattering function. Acta Cryst. A67, 327–335 (2011) 23. Chiarello, G.L., Ferri, D.: Modulated excitation extended X-ray absorption fine structure spectroscopy. PCCP. 17, 10579 (2015) 24. König, C.F.J., Schildhauer, T.J., Nachtegaal, M.: Methane synthesis and sulfur removal over a Ru catalyst probed in situ with high sensitivity X-ray absorption spectroscopy. J. Catal. 305, 92–100 (2013) 25. Marchionni, V., Newton, M.A., Kambolis, A., Matam, S.K., Weidenkaff, A., Ferri, D.: A modulated excitation ED-EXAFS/ DRIFTS study of hydrothermal ageing of Rh/Al2O3. Catal. Today. 229, 80–87 (2014)

A. Urakawa et al. 26. Ferri, D., Newton, M.A., Nachtegaal, M.: Modulation excitation X-ray absorption spectroscopy to probe surface species on heterogeneous catalysts. Top. Catal. 54, 1070–1078 (2011) 27. Marchionni, V., Ferri, D., Kröcher, O., Wokaun, A.: Increasing the sensitivity to short-lived species in a modulated excitation experiment. Anal. Chem. 89, 5801–5809 (2017) 28. Marchionni, V., Nachtegaal, M., Ferri, D.: Influence of CO on dry CH4 oxidation on Pd/Al2O3 by operando spectroscopy: a multitechnique modulated excitation study. ACS Catal. 10, 4791–4804 (2020) 29. Marchionni, V., Kambolis, A., Nachtegaal, M., Kröcher, O., Ferri, D.: High energy X-ray diffraction and IR spectroscopy of Pt/Al2O3 during CO oxidation in a novel catalytic reactor cell. Catal. Struct. React. 3, 71–78 (2016) 30. Meier, D.M., Urakawa, A., Baiker, A.: Polarization-modulation infrared reflection-absorption spectroscopy affording time-resolved simultaneous detection of surface and liquid phase species at catalytic solid-liquid interfaces. Analyst. 134(9), 1779–1780 (2009). https://doi.org/10.1039/b911151d 31. Urakawa, A., Bürgi, T., Schläpfer, H.-P., Baiker, A.: Simultaneous in situ monitoring of surface and gas species and surface properties by Modulation Excitation PM-IRRAS: CO oxidation over Pt film. J. Chem. Phys. 124, 054717 (2006) 32. Bieri, M., Bürgi, T.: Probing enantiospecific interactions between proline and an L-glutathione self-assembled monolayer by modulation excitation ATR-IR spectroscopy. J. Phys. Chem. B. 109(20), 10243–10250 (2005). https://doi.org/10.1021/jp050197n 33. Patil, B.S., Srinivasan, P.D., Atchison, E., Zhu, H.D., Bravo-Suarez, J.J.: Design, modelling, and application of a low void-volume in situ diffuse reflectance spectroscopic reaction cell for transient catalytic studies. React. Chem. Eng. 4(4), 667–678 (2019). https://doi.org/10. 1039/c8re00302e 34. Maeda, N., Meemken, F., Baiker, A.: Insight into the mechanism of the preferential oxidation of carbon monoxide by using isotopemodulated excitation IR spectroscopy. ChemCatChem. 5(8), 2199–2202 (2013). https://doi.org/10.1002/cctc.201300172 35. Pavelko, R.G., Choi, J.-K., Urakawa, A., Yuasa, M., Kida, T., Shimanoe, K.: H2O/D2O exchange on SnO2 materials in the presence of CO: operando spectroscopic and electric resistance measurements. J. Phys. Chem. C. 118(5), 2554–2563 (2014). https://doi.org/ 10.1021/jp4108766 36. Ledesma, C., Yang, J., Chen, D., Holmen, A.: Recent approaches in mechanistic and kinetic studies of catalytic reactions using SSITKA technique. ACS Catal. 4(12), 4527–4547 (2014). https://doi.org/10. 1021/cs501264f 37. Voronov, A., Urakawa, A., van Beek, W., Tsakoumis, N.E., Emerich, H., Rønning, M.: Multivariate curve resolution applied to in situ X-ray absorption spectroscopy data: an efficient tool for data processing and analysis. Anal. Chim. Acta. 840, 20–27 (2014). https://doi.org/10.1016/j.aca.2014.06.050 38. Urakawa, A.: Trends and advances in operando methodology. Curr. Opin. Chem. Eng. 12, 31–36 (2016). https://doi.org/10.1016/j. coche.2016.02.002 39. Alcaraz, M.R., Aguirre, A., Goicoechea, H.C., Culzoni, M.J., Collins, S.E.: Resolution of intermediate surface species by combining modulated infrared spectroscopy and chemometrics. Anal. Chim. Acta. 1049, 38–46 (2019). https://doi.org/10.1016/j.aca.2018. 10.052 40. Witzke, M.E., Almithn, A., Coonrod, C.L., Triezenberg, M.D., Hibbitts, D.D., Flaherty, D.W.: In situ methods for identifying reactive surface intermediates during hydrogenolysis reactions: C-O bond cleavage on nanoparticles of nickel and nickel phosphides. J. Am. Chem. Soc. 141(42), 16671–16684 (2019). https://doi.org/ 10.1021/jacs.9b06112

42

Modulation Excitation Spectroscopy (MES)

977

42 Atsushi Urakawa is Professor of Catalysis Engineering at the Department of Chemical Engineering, Delft University of Technology (The Netherlands). His research team combines fundamental and applied approaches for catalysis research and focuses on the rational development of heterogeneous catalysts and processes aided by in situ and operando methodologies.

Davide Ferri received his PhD in chemistry at the ETH Zurich in 2002. He is currently head of the group for Applied Catalysis and Spectroscopy at the Paul Scherrer Institut (Switzerland). His interests span from environmental catalysts to liquid-phase catalyzed reactions and the development of in situ/operando experiments. He uses vibrational spectroscopies in combination with X-ray based methods.

Rob Jeremiah G. Nuguid currently works at Linde GmbH (Germany). He obtained his PhD degree from École polytechnique fédérale de Lausanne and Paul Scherrer Institut (Switzerland), where he used time-resolved spectroscopy to uncover the mechanisms of catalytic reactions.

Case Study 1: Modulation Excitation Spectroscopy (MES) Rob Jeremiah G. Nuguid

43

and Davide Ferri

Contents 43.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 979

43.2

State-of-the-Art Spectroscopic Studies on Selective Catalytic Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 979

43.3

Beyond the Steady-State: Advantages of Modulated Excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amplification of Weak Signals and Resolution of Peaks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discrimination Between Active and Responsive Sites . . . Detection of Intermediates Species . . . . . . . . . . . . . . . . . . . . . . .

43.3.1 43.3.2 43.3.3

980 980 983 985

of MES. Coupled to vibrational methods such as infrared or Raman spectroscopy, MES can amplify weak but dynamic signals, resolve overlapping peaks, discriminate between active and merely responsive sites, and enable the detection of intermediate species. Thanks to these improved features, MES can extend our molecular view of SCR to unprecedented levels that are not attainable using steady-state measurements.

43.4

Considerations on the Selection of Modulation Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 987

Keywords

43.5

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 988

ME · PSD · DRIFT · Raman · SCR · Vanadium · IR · Iron · Copper · Titania · Transient response · Zeolite

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 988

43.1

Introduction

Abstract

Selective catalytic reduction (SCR) is an emission control technology that has continuously remained at the forefront of research using spectroscopy. While steady-state studies significantly advanced our mechanistic knowledge of SCR in the past, they proved to be not the best-suited approach to elucidate the reaction cycle conclusively. Nowadays, transient-response experimentation – especially modulated excitation spectroscopy (MES) – is rapidly emerging as an alternative due to its improved sensitivity and analytical capabilities. Here we present a selection of recent SCR-related studies that clearly illustrates the advantages R. J. G. Nuguid Energy and Environment Research Division, Paul Scherrer Institut, Villigen, Switzerland École polytechnique fédérale de Lausanne (EPFL), Institute for Chemical Sciences and Engineering, Lausanne, Switzerland e-mail: [email protected] D. Ferri (*) Energy and Environment Research Division, Paul Scherrer Institut, Villigen, Switzerland e-mail: [email protected]

While the principles of modulation excitation spectroscopy (MES) have already been discussed, the goal of this companion chapter is to present how MES has been applied to improve the molecular-level understanding of catalytic reactions using spectroscopic experiments. To illustrate this point, we have selected as a case study the selective catalytic reduction (SCR) of NOx in the presence of NH3 – a reaction that enjoys commercial success in the exhaust gas aftertreatment of diesel-powered vehicles and thermal power plants. Over the years, the SCR mechanism has been debated [1–3], and the reaction cycles over various catalysts continue to be redrawn as new spectroscopic evidence emerges.

43.2

State-of-the-Art Spectroscopic Studies on Selective Catalytic Reduction

All proposed SCR mechanisms share a common, unifying theme [4]. The overall reaction cycle can be roughly divided into three parts: (1) an acid-base part where NH3 adsorbs on

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_43

979

980

the acid sites present on the catalyst surface; (2) a reduction part that is accompanied by the reaction of NO with the activated NH3 molecule; and (3) a reoxidation part where molecular O2 or lattice oxygen regenerates the active site. This general scheme is valid for the two most industrially relevant families of SCR catalysts (i.e., vanadia-based and zeolite-based catalysts). One of the most polarizing issues is the involvement of Lewis and Brønsted-Lowry acid sites in the reaction. For zeolite-based catalysts, the assignment of the active site is straightforward because the redox-active component, usually Fe or Cu, possesses only Lewis acid character. Nonetheless, the role of Brønsted-Lowry acid sites to supply reactants to the redox center cannot be neglected. In the case of vanadiabased catalysts, the interpretation is complicated by the fact that the active VOx site can adopt both Lewis and BrønstedLowry functionalities depending on the temperature and gas environment. For instance, terminal vanadyl groups (V¼O) acting as Lewis acid sites can be hydroxylated and transform into Brønsted-Lowry sites in the presence of water. The arguments in support of Lewis acid sites as the SCR-active species include their higher thermal stability and higher turnover frequency (TOF) upon reaction with NO [1, 5]. On the other hand, the higher abundance of Brønsted-Lowry acid sites under reaction conditions and the correlation of their surface concentration with the catalytic activity play in favor of Brønsted-Lowry acid sites as the active sites [2]. While NH3 is generally acknowledged to adsorb strongly on any SCR catalyst, it is still not clear whether NO also adsorbs (albeit weakly) or not. For vanadia-based catalysts, most of the studies suggest an Eley-Rideal-type mechanism where NH3 is the only reactant that is present on the catalyst surface, with molecular NO just reacting from the gas phase [6–8]. This is supported by many IR studies wherein only NH3 was observed at the catalyst surface under reaction conditions. For zeolitic catalysts, however, the mechanism may be temperature-dependent. Previous studies have speculated that at low temperatures, NO-derived species may also form on the catalyst surface and could react with pre-adsorbed NH3 in a Langmuir-Hinshelwood fashion [9, 10]. Due to the highly oxidizing nature of Fe and Cu, NO may even transform into nitrate species and accumulate on the surface. Raman spectroscopy was also used as a direct probe of the molecular structure of the catalyst during SCR conditions, especially for V2O5/TiO2-based materials. Due to the strong active phase-support interaction, vanadia appears exclusively as surface amorphous species until the monolayer coverage is reached, which is typically at 7–8 VOx unitsnm2 [11, 12]. The signal corresponding to ν(V¼O), the fundamental Raman-active stretch mode of vanadyl groups, is very sensitive to the gas environment to which the sample is exposed. In particular, the presence of H2O and NH3 causes a redshift,

R. J. G. Nuguid and D. Ferri

indicating molecular interaction [13]. NO had no apparent effect on the energy of this signal, thus further supporting the IR data showing that NO does not adsorb strongly. For zeolitic SCR catalysts, Raman studies are rather scarce and often center on using the technique as a probe of the framework stability after hydrothermal aging [14]. In situ X-ray absorption spectroscopy, as well as ultraviolet-visible spectroscopy and electron paramagnetic resonance contributed to confirming the redox cycling of the active metal during the SCR reaction [15–17]. Most studies theorize that the simultaneous presence of both NO and NH3 is needed to drive the reduction of the active site. The redox couples responsible for the reaction are V5+/V4+, Fe3+/Fe2+, and Cu2+/Cu+ for the catalysts based on vanadium, iron, and copper, respectively. Arguably, steady-state spectroscopic studies have already driven our molecular understanding of SCR to the highest possible level. In order to extract further molecular information and clarify the finer details of the SCR mechanism, transientbased methodologies should be implemented in spectroscopic experiments. MES belongs to this approach. The following examples using earlier results on SCR illustrate how the added sensitivity imparted by MES can reveal insights that fall beyond the scope of conventional spectroscopy.

43.3

Beyond the Steady-State: Advantages of Modulated Excitation

43.3.1 Amplification of Weak Signals and Resolution of Peaks Figure 43.1a shows two selected diffuse reflectance infrared (DRIFT) spectra of Fe-ZSM-5 obtained in a modulation experiment consisting of ten NO pulses in a flow of NH3, O2, and H2O at 250  C [18]. The spectra were selected at the end of the NH3/O2/H2O and NO/NH3/O2/H2O pulses. At first glance, the two spectra appear practically identical: they are dominated by the signals of adsorbed NH3, which is evident in the stretching (3500–2500 cm1) and the bending (1750–1250 cm1) regions. Two types of adsorbed NH3 species can be distinguished: (1) Lewis NH3 coordinated directly with Fe2+ ions and (2) Brønsted-Lowry NH4+ species linked to Al-O-Si groups of the zeolite framework. In the N-H stretching region of the steady-state spectra, these two species appear at different frequencies, but the peaks corresponding to NH3 and NH4+ are not well-resolved. Furthermore, at this temperature, adsorbed NO species are also expected at around 2150 cm1 (as nitrosyl, NO+) yet their presence appears to be obscured by NH3, which can interact more strongly with the catalyst sites due to reasons of basicity and hydrogen bonding strength. Because of the coordination of NH3 with the zeolite framework, various negative signals

43

Case Study 1: Modulation Excitation Spectroscopy (MES)

Figure 43.1c shows the result of such averaging for five consecutive adsorption/desorption cycles. Here, the NO+ peak is clearly visible and we can also begin to discern between the Lewis NH3 and Brønsted-Lowry NH4+ species at around 1600 and 1430 cm1, respectively, but their spectral features are not yet completely resolved. The bending region of the averaged spectra is still quite noisy. Further averaging over ten repeated cycles enhances the quality, but only marginally (Fig. 43.1d). This indicates that the improvement obtained from averaging cycles reaches rapidly a new equilibrium level, and that simple spectral subtraction is not sufficient to extract molecular information from these rather complex datasets. The averaged data over the repeated cycles of measurement can be processed using the phase sensitive detection (PSD) concept shown in the previous chapter (Eq. 43.1) to convert the raw IR spectra, measured in the time domain, to the corresponding set of spectra in the phase domain.

a) NO+NH3 NH3 0.1 a.u.

b)

Absorbance (a.u.)

981

0.01 a.u.

c)

0.01 a.u.

d)

ðT

 PSD 2 PSD I φ I ðtÞ sin kωt þ φ ¼ dt T

0.01 a.u. 4000

3500

3000

2500

2000

ð43:1Þ

0

1500

1000

Wavenumber (cm–1)

Fig. 43.1 (a) Selected diffuse reflectance IR spectra of 2.6 wt% Fe-ZSM-5 at 250  C obtained in a modulation experiment consisting of alternate pulses of 1000 ppm NO, 1000 ppm NH3, 2 vol% H2O, 5 vol % O2 balanced in Ar and of 1000 ppm NH3, 2 vol% H2O, 5 vol% O2 balanced in Ar as indicated. (b–d) Difference spectra under NO+NH3 and NH3 obtained for one cycle, five cycles, and ten cycles

indicate the perturbation of the framework itself (2000 and 1890 cm1), the Brønsted-Lowry sites (3602 cm1), and the silanol groups (3740 and 3650 cm1). The most straightforward way to uncover molecular information from these data is to perform a mathematical subtraction of the NH3 spectrum from the NH3 þ NO spectrum. This should yield the signals contributed solely by NO species and groups affected by NO adsorption. Indeed, in the difference spectrum shown in Fig. 43.1b, the expected broad NO+ peak at 2150 cm1 starts to emerge. The NH3 signal, however, still appears as a wide peak in the stretching region. Furthermore, the overall spectral quality is rather poor, especially in the bending region where no peak can be discerned unambiguously. The adsorption/desorption of NO could be repeated multiple times while recording IR spectra, in order to obtain averaged spectra with higher signal-to-noise ratio.

Analogous to a Fourier transformation, this process amplifies the signal from the species that responded to the perturbation (i.e., NO pulsing in this case), and filters out the signal from those which are not affected [19, 20]. Hence, a species giving rise to a strong signal in the time-resolved spectra will be absent in the phase-resolved spectra if it remained static during the pulse duration. The reverse is true for chemical species that exhibited weak signals in the time-resolved spectra but nonetheless generated a substantial response to the repeated pulses compared to static signals. Figure 43.2b shows a selected phase-resolved spectrum obtained from PSD of a series of time-resolved spectra measured during NO pulsing in a gas feed containing NH3. Essentially, these time-resolved spectra belong to the modulation experiment shown in Fig. 43.1a. Here, the primary NO+ peak that was obscured in Fig. 43.1a is now visible, thereby confirming the presence of transient NO-derived species on the catalyst surface at 250  C. Furthermore, NH4+ (~3380 and ~3270 cm1) and NH3 (~3180 cm1) species appear as clearly separate peaks, suggesting the involvement of Lewis and Brønsted-Lowry sites in the SCR process. The spectral quality of the bending region also markedly improved compared to the difference spectra shown in Fig. 43.1b–d, making most of the peaks that were previously present in the time-resolved spectra discernable. It should also be noted that the maximum intensity of the

43

982

R. J. G. Nuguid and D. Ferri

Absorbance (a.u.)

(a)

(c)

0.1 a.u.

(b)

(d)

0.001 a.u. 3380

0.1 a.u.

0.001 a.u.

3270 3180

3650 1365 4000

3500

3000 2500 2000 Wavenumber (cm–1)

1500

1000

4000

3500

3000 2500 2000 Wavenumber (cm–1)

1500

1000

Fig. 43.2 (a) Time-resolved spectra and (b) selected phase-resolved spectrum of 2.6 wt% Fe-ZSM-5 obtained in a modulation experiment consisting of 1000 ppm NO pulses in a gas feed of 1000 ppm NH3, 2 vol % H2O, 5 vol% O2 balanced in Ar at 250  C. (c) Time-resolved spectra

and (d) selected phase-resolved spectrum obtained in a modulation experiment consisting of Ar pulses in a gas feed of 1000 ppm NH3, 2 vol% H2O, 5 vol% O2 balanced in Ar at 250  C. Adapted from [18]

signals in Fig. 43.2 is 1–3 mabs units higher compared to the intensity measured in the time-resolved spectra clearly indicating the advantage of MES with respect to amplification of weak signals. It is also worth noting that not all of the peaks present in the time-resolved spectra appeared also in the phase-resolved spectra. For instance, the broad peak at ca. 2770 cm1 which is ascribed to physisorbed NH3 (e.g., adsorbed molecules in excess of the monolayer) [21] was almost completely filtered out by the PSD treatment. This suggests that this species was not affected significantly by the NO pulses. Random noise, which does not vary with the same frequency as the stimulation, is also eliminated in the same manner. It is obvious from Fig. 43.2b that the level of noise of the spectrum is much lower compared to that of Fig. 43.1b–d. The behavior of selected peaks in the time domain can explain how this signal enhancement or attenuation occurs in the phase domain. Figure 43.3a, b shows the evolution of

the signals assigned to NO+ and physisorbed NH3, respectively, during the entire duration of the experiment. The absolute intensity of the physisorbed NH3 signal was greater than that of NO+ and correspondingly, physisorbed NH3 dominated the time-resolved spectra (Fig. 43.2a). However, the relative changes that it underwent during the NO pulsing were about four times smaller than the changes experienced by the latter. Hence, the signal contribution of physisorbed NH3 to the phase-resolved spectra was diminished correspondingly. It is also evident that not all of the peaks in Fig. 43.2 varied synchronously. In particular, hydroxyl (-OH; 3650 cm1) and sulfate (-SO4 from the ion-exchange procedure; 1365 cm1) groups varied asynchronously (out-of-phase) with NH3 and NO+ suggesting that they are titrated upon reactant adsorption and liberated upon reactant desorption. To ascertain that the presence of the peaks observed in the phase-resolved spectra was due to the SCR reaction and not

43

Case Study 1: Modulation Excitation Spectroscopy (MES)

a)

0.001 a.u.

Absorbance (a.u.)

0.12 a.u.

b)

0.001 a.u. 0.28 a.u.

0

200

400

600 Time (s)

800

1000

1200

Fig. 43.3 Time-resolved signals of (a) NO+ (2150 cm1) and (b) physisorbed NH3 (2770 cm1) as a result of pulsing 1000 ppm NO in a gas feed of 1000 ppm NH3, 2 vol% H2O, 5 vol% O2 balanced in Ar on 2.6 wt% Fe-ZSM-5 at 250  C

because of experimental issues (e.g., pressure drop during valve switching), a control experiment was performed in which Ar was pulsed instead of NO on a catalyst sample that was equilibrated with NH3 at 250  C. The resulting phase-resolved spectrum is displayed in Fig. 43.2c. As opposed to the experiment where NO was pulsed in the NH3 feed to the cell, no significant phase-resolved signal corresponding to either NO or NH3 was observed. This is because Ar does not interact with the catalyst. Furthermore, this also shows that NH3 was strongly adsorbed on the catalyst surface and no significant desorption occurred at the selected temperature. A recent study conducted on Cu-SSZ-13 provides another striking example of signal amplification due to MES [15]. During the SCR reaction cycle, Cu species change oxidation state from +2 to +1. While other techniques such as x-ray absorption spectroscopy or UV-visible spectroscopy are more suited to follow this redox cycling behavior, infrared spectroscopy can provide an indirect approach to probe Cu2+ species. In sufficiently low-loaded Cu-SSZ-13 catalysts, the Cu sites maintain atomic dispersion and coordinate individually with Si–O–Al bridges, forming Cu2+–OH species. These Cu2+–OH units represent the fully oxidized form of the Cu species and serve as the starting point for most of the proposed SCR reaction mechanisms. The ν(O-H) vibrational mode of the Cu2+–OH units is IR-active and provides an indication of the presence of Cu2+ species in the experiment. Figure 43.4 shows the IR spectra of a Cu-SSZ-13 sample as a result of NO pulsing in a gas feed containing NH3 and O2. The high-frequency region between 3750 and 3550 cm1 is dominated by the ν(O-H) signal from different species. These include isolated silanol groups (Si–OH), Cu2+–OH units, and hydroxyl bridges (Si–(OH)+–Al; Brønsted-Lowry

983

acid sites). Due to the large differences in the charge and reduced masses of the atoms in the vicinity of the O–H bond, the vibrational modes appear as well-resolved peaks in the spectra. Furthermore, the peaks are negative as the reactant molecules coordinate to them (similar to the case of Fe-ZSM-5 previously discussed). The absolute value of the peak height in the time-resolved spectra can be used as an estimate of the surface concentration of the species. It is clear that Si–(OH)+–Al units are the most abundant hydroxyl moieties. They are followed by charge-neutral Si–OH groups, which also serve as part of the zeolite framework. The least abundant species are the Cu2+–OH units, which could reflect the low Cu content of the catalyst. In summary, the timeresolved peak heights followed the trend Si–(OH)+– Al > Si–OH > Cu2+–OH. In contrast, the phase-resolved spectra showed a different picture as Cu2+–OH displayed the greatest changes. From a mechanistic point of view, this result is intuitive and important. The Cu2+–OH species are the active sites in the SCR reaction and are therefore very sensitive to the gas-phase concentration of the reactants. When the NO supply was switched on, the Cu2+–OH species became less negative as they were freed up from NH3 coordination and the SCR products desorbed. In the other half period where NO was cut off, NH3 started to adsorb once again in vacant Cu2+ sites, and the Cu2+–OH signal became more negative again. Interestingly, Si–(OH)+–Al groups were also significantly modulated, pointing out their role in the adsorption and mobilization of adsorbed NH3 molecules. In contrast, Si–OH groups did not appear to modulate significantly. This implies that while Si–OH can coordinate with NH3 molecules, they are just spectators in the SCR process. The amplification of the weak signals of the Cu2+–OH groups was made possible only due to the sensitivity offered by the PSD treatment.

43.3.2 Discrimination Between Active and Responsive Sites In modulation experiments where the SCR reaction is repeatedly turned on and off, any species that changes with the stimulation is catalytically relevant and is involved to some extent to the over-all SCR process. However, this does not necessarily mean that it is the active species; this additional confirmation has to be obtained separately. For example, both Lewis-bound NH3 and Brønsted-Lowry-bound NH4+ were present in the phase-resolved spectrum in Figs. 43.2 and 43.4. However, we know that only Lewis acid sites (i.e., vacant Fe2+ and Cu2+ centers) are active for SCR because the reaction involves a redox step that can be only satisfied by a transition metal. Adsorbed NH4+ on Al-O-Si bridges is not active for SCR. This is confirmed by the finding that metal-

43

H .N ad s

2+

Cu + ad s

Absorbance (a.u.)

N

O+

3

Is ol [C ate u d Br 2+( SiO i d OH H gi – ng )] + A N l-O H + 4 H Cu io 2+ n s ad an s. d N H

Fig. 43.4 Time-resolved and phase-resolved spectra of Cu-SSZ-13 obtained in a modulation experiment consisting of alternate pulses of 500 ppm NO in a gas feed of 500 ppm NH3, 10 vol% O2 balanced in Ar at 250  C. (Reproduced from [15] with permission from the Royal Society of Chemistry)

3

R. J. G. Nuguid and D. Ferri

.N O Cu 2+ ad s N H . NH 4 + io 3 ns Cu

984

0.2 a.u.

0.002 a.u. 3500

3000

free zeolites possessing only Brønsted-Lowry sites have negligible deNOx activity even at elevated temperature. Hence, the presence of NH4+ in the phase-resolved spectra supports the occurrence of NH3 transfer from one site to another; that is, adsorbed NH4+ present elsewhere in the zeolite can spillover to Fe2+ or Cu2+ sites and be activated. V2O5/TiO2-based SCR catalysts provide another illustration of this concept. In this system, V2O5 is present as surface VOx species and serves as the active site while TiO2 is the high-surface-area support. Figure 43.4a shows the timeresolved Raman spectra of V2O5/TiO2 obtained in a modulation experiment consisting of NH3 pulses in a gas feed containing NO. Under these conditions, NO conversion through the SCR reaction changed repeatedly between a maximum and a minimum (non-zero) value. The peaks at 395, 515, and 635 cm1 originate from the lattice vibrations of TiO2 in the anatase polymorph [22]. The weak peak at 1030–1020 cm1 stems from the fundamental stretch mode, ν(V¼O), of terminal VOx groups. After demodulation, VOx and TiO2 peaks appeared in the phase-resolved spectra (Fig. 43.6c), which indicates that both the active phase and the support are involved in the SCR process [23]. Expansion of the ν(V¼O) region of the Raman spectrum revealed that in the time-resolved spectra (Fig. 43.5d) this peak appears rather broad because under the experimental conditions there is a wide distribution of vanadyl species in different states of coordination, each one giving rise to a slightly different frequency. Furthermore, the peak maximum was centered at 1025 cm1, which is the expected frequency when VOx is in contact with adsorbates such as NH3 and

2000 2500 Wavenumber (cm–1)

1500

1000

H2O [13, 23]. However, in the phase-resolved spectra (Fig. 43.5d), the vanadyl peak appeared narrower, which suggests that not all of the VOx species contribute to the SCR activity. The peak in the phase-resolved spectra was centered at 1031 cm1, which corresponds to coordinatively unsaturated vanadyl species. Therefore, the VOx species responding to the NH3 pulses and the variation in NO conversion – and most probably the active sites – are the vacant Lewis VOx sites. The intensity of the ν(V═O) peak in the phase-resolved spectra correlated with the proportion of vanadyl sites participating in the SCR reaction. Hence, the confirmation that coordinatively unsaturated VOx species are the active site for SCR can be obtained from Arrhenius-type relations where the maximum intensity of the VOx peak in the phaseresolved spectra obtained from ME experiments at various temperatures is compared with the corresponding NO conversion (Fig. 43.6). The comparable activation energies obtained from the two sets of data confirmed that the species associated with the ν(V═O) signal changing in response to the NH3 modulation are effectively involved in the SCR reaction. In a complementary experiment wherein NH3 is pulsed in a gas feed without NO, no SCR reactivity can be observed and only NH3 adsorption/desorption can occur. The timeresolved and phase-resolved spectra for this experiment are shown in Fig. 43.7a, b. The time-resolved spectra appear similar to those obtained in the experiment where NH3 is pulsed in NO, with the exception that the ν(V¼O) was significantly red-shifted. However, the phase-resolved

43

Case Study 1: Modulation Excitation Spectroscopy (MES)

a)

985

a)

100 a.u.

100 a.u.

b)

Intensity (a.u.)

Intensity (a.u.)

1025

b)

5 a.u.

1031

5 a.u.

43 1200

1000

800

600

400

1200 1100 1000 900 800 700 600 Raman shift (cm–1)

Raman shift (cm–1)

Fig. 43.5 Raman spectra of 2 wt% V2O5/TiO2 obtained while pulsing 500 ppm NH3 in a gas feed of 1000 ppm NO, 2 vol% H2O, 5 vol% O2 balanced in Ar at 250  C: (a) full time-resolved spectra and (b) corresponding phase-resolved spectra. The enlarged spectra of the VOx region in the time and phase domains are given as insets. (Reprinted (adapted) with permission from [23]. Copyright 2019 American Chemical Society)

3.0

ln k (a.u.)

–1.0 Ea = 38 ± 4 kJ·mol–1

2.5

–1.5 2.0 –2.0 –2.5 –3.0

Ea = 43 ± 6 kJ·mol–1

1.5 1.0

Maximum peak intensity (a.u.)

3.5

–0.5

0.5 0.0019 0.0020 0.0021 0.0022 0.0023 0.0024 1/T (K–1)

Fig. 43.6 Arrhenius-type relationship for 2 wt% V2O5/TiO2 using catalytic activity data and phase-resolved peak intensity. (Reprinted (adapted) with permission from [23]. Copyright 2019 American Chemical Society)

spectra showed only the signal from the VOx species, with only minor contribution from TiO2. This result agrees well with the observation that NH3 binds strongly to TiO2 and that this stable interaction renders TiO2 only weakly responsive to the pulses. During the first few pulses before the quasiequilibrium state was established, TiO2 was most probably saturated with NH3 molecules so strongly bound that they did not desorb under these experimental conditions. On the other hand, VOx is less likely to retain NH3 molecules because of its role as the active phase; it should adsorb reactant molecules sufficiently but also be able to release the products. As a

500

400

300

Fig. 43.7 Raman spectra of 2 wt% V2O5/TiO2 obtained at 250  C while pulsing 500 ppm NH3 in a gas feed of 2 vol% H2O, 5 vol% O2 balanced in Ar: (a) full time-resolved spectra; and (b) corresponding phase-resolved spectra. (Reprinted (adapted) with permission from [23]. Copyright 2019 American Chemical Society)

consequence, the signal of VOx appeared in the phaseresolved spectra. The role of TiO2 in the SCR process could now be revealed using the results of these two experiments. When NO is present, there is a continuous consumption of NH3 bound to the VOx sites as the SCR reaction is activated. The replenishment of NH3 could either take place from the gas phase (during the NH3 pulses) or from the NH3 bound to TiO2. However, the presence of TiO2 in the phase-resolved spectra in Fig. 43.5b suggests that the spillover of NH3 from the support to the active site takes place actively. From a molecular viewpoint, this is plausible because pre-adsorbed molecules are much closer to the active sites and are easier to abstract than those approaching from the gas phase. In contrast, when NO is not present, this NH3 transfer mechanism is not significant because NH3 is not actively consumed at the active sites; rather, only desorption occurs. This observation rationalizes the poor intensity of TiO2 peaks in the phaseresolved spectra in Fig. 43.7b. ME-Raman spectroscopy therefore provides evidence that TiO2 acts as a reservoir of pre-adsorbed NH3 molecules that can be mobilized to the active sites to be activated.

43.3.3 Detection of Intermediates Species Proving the formation of intermediates is pivotal in evaluating the validity of proposed mechanisms. SCR is believed to proceed through the formation of a nitrosamide intermediate, NH2NO, which proceeds consequent to the reduction of the active metal [1, 4]. As with many other intermediate species, NH2NO has an extremely low thermodynamic

986

R. J. G. Nuguid and D. Ferri

stability and decomposes readily into the products, N2 and H2O. Indeed, the lifetime of NH2NO can be as short as 1011 s [24], preventing its accumulation in significant amounts for spectroscopic detection under steady-state conditions. NH2NO exhibits molecular vibration at 1490 cm1 due to the NO stretching mode [1]. In steady-state spectra, however, low levels of this species are certainly masked by the broad signal of NH4+ species at 1435 cm1 (Fig. 43.8a). It is only after applying PSD that the expected signal of NH2NO emerged from the large contribution of NH4+ species (Fig. 43.8b) [25]. Closer inspection of Fig. 43.8b reveals that the Lewisbound NH3 signal (1603 cm1) varies asynchronously with respect to that of NH2NO, already hinting to the appearance of one species when the other is consumed. This becomes clearer when the phase angle dependence of the two species is plotted as shown in Fig. 43.8c. When two species evolve with different kinetics, then there should be a phase shift between their signals. As a corollary, multiple signals contributed by the same species should behave similarly. Furthermore, the phase angle is related to the time domain in such a way that the higher the phase angle maximum of a species is, the sooner it appears during the modulation period [26]. Here it is clearly shown that upon the introduction of NO (i.e., start of the pulse), NH2NO is formed abruptly but then decreases immediately as it most probably decomposes into the products. In addition, the formation of NH2NO is concomitant with the decrease of NH3, which is consistent with their respective roles as intermediate and reactant, respectively. Although this information can also be obtained by analyzing the evolution of the IR signals of the species in

a)

1425

0.1 a.u.

the time domain, which is similar to what has been presented in Fig. 43.3, the noise level will be higher. In a similar way, the Cu–N(¼O)–NH2 intermediate was shown to form on Cu-SSZ-13 under SCR conditions [15]. Theoretical calculations postulate that this species exhibits IR-active vibrational modes at 1431 and 1258 cm1 as a result of N¼O and N-N stretches, respectively. Figure 43.4 shows that the time-resolved spectra feature a peak at 1460 cm1, which could be attributed to NH4+ species bound to Brønsted–Lowry sites. There appears to be a corresponding peak around this region in the phase-resolved spectra. However, a closer inspection reveals that the corresponding band is red-shifted to 1436 cm1, which suggests that the observed responsive species are probably not NH4+ species. Another peak was identified at 1270 cm1 that could be associated with NH3 molecules adsorbed on Lewis acid sites. Yet the corresponding demodulated peak is centered at 1258 cm1, amounting to a red shift of more than 10 cm1. To ascertain the identities of these peaks, the phase angular dependence of the peaks was analyzed as shown in Fig. 43.9. Figure 43.9 clearly shows that there is only a small angular delay between the signals at 1436 and 1258 cm1, thereby suggesting that these peaks are contributed by the same species – most probably the Cu–N(¼O)–NH2 intermediate species proposed by the theoretical calculations. The signal from NH3 bound to Lewis acid sites at 3182 cm1 showed a significant phase delay from those of the intermediate. This observation further strengthens the claim that the NH3 peak at 1270 cm1 in the time-resolved spectra was not the one responsible for the peak at 1258 cm1 in the phase-resolved spectra.

c) 1.0

b)

1488

0.001 a.u.

1604 1900

1800

1700

1600 1500 1400 Wavenumber (cm–1)

Normalized absorbance

Absorbance (a.u.)

1604 0.5

0.0 1488 cm–1 1604 cm–1

–0.5

–1.0 1300

1200

Fig. 43.8 DRIFT spectra of 2 wt% V2O5/10 wt% WO3/TiO2 obtained while pulsing 500 ppm NO in a gas feed of 500 ppm NH3, 5 vol% O2 balanced in Ar at 250  C: (a) full set of time-resolved spectra and (b) of

0

60

120

180 240 Phase angle (°)

300

360

phase-resolved spectra. (c) Phase angular dependence of the signals assigned to the nitrosamide intermediate (1488 cm1) and of NH3 coordinated to Lewis acid sites (1604 cm1)

43

Case Study 1: Modulation Excitation Spectroscopy (MES)

Normalized absorbance

1.0

1258 cm–1 1436 cm–1 3182 cm–1

0.5

0.0

–0.5

–1.0 0

60

120 180 240 Phase angle (°)

300

360

Fig. 43.9 Phase angular dependence of the signals assigned to the nitrosamide intermediate (1258 and 1436 cm1) and NH3 on Lewis acid sites (3182 cm1) of Cu-SSZ-13 in a modulation experiment consisting of alternate pulses of 500 ppm NO in a gas feed of 500 ppm NH3, 10 vol% O2 balanced in Ar at 250  C. (Reproduced from [15] with permission from the Royal Society of Chemistry)

43.4

Considerations on the Selection of Modulation Experiment

The experiments described so far were of different nature (i.e., modulation of NO concentration and modulation of NH3 concentration). The question arises on which type of modulation – essentially which pulse sequence – is going to deliver the most useful information. The final choice of which perturbation to apply largely depends on the expected response of the system and the tool used to quantify that response. Controlled variations in pH, temperature, pressure, or incident wavelength can all potentially elicit a signal change. Concentration modulation is by far the most widely used in studies concerning heterogeneous catalysts. In this approach, one (or sometimes more than one) of the reactants is pulsed intermittently while the others are kept in constant supply as spectra are collected. As catalysis is a surface phenomenon, the active sites will have to interact with the reactants before a chemical transformation occurs. Hence, changes in the reactant concentration will ultimately trigger changes in the chemical state of the catalytically relevant sites, allowing them to be selectively detected. Three major reactants are involved in the standard SCR reaction, namely, NH3, NO, and O2. By repeatedly cutting off and reintroducing one of these reactants, the SCR reaction is effectively turned off and on. Modulation of one reactant or the other can produce different or no additional information, depending on the selected perturbation [27]. The decision of which reactant concentration to modulate in order to obtain the most significant information from the modulation

987

experiment can be also taken based on the specific technique to be applied in reason of the different type of information that the specific technique addresses. IR spectroscopy can reveal the presence of adsorbed reactants and intermediates. Under SCR conditions, NH3 populates the surface of the catalyst, with NO and O2 both reacting from the gas phase. Furthermore, NH3 can form two types of surface species, depending on which site it coordinates to: NH4+ on Brønsted-Lowry sites and NH3 on Lewis acid sites. In order to determine which of the two species is active, a modulation experiment concerning NO can be applied under constant NH3 and O2. During the “SCR on” half-period, NO will selectively titrate either NH3 or NH4+ while during the “SCR off” half-period, the sites from which NH3 has been consumed, will have a chance to be replenished from the NH3 in the gas phase. The constant presence of O2 is also needed during this “SCR off” part because it ensures that most of the vanadyl species are in the V5+ state, which is where NH3 is activated. Hence, NO modulation is a useful tool to combine with IR spectroscopy to study the active sites of SCR. In contrast, NH3 modulation will probably not reveal the same depth of information. Since NO can interact only weakly with the catalyst, whatever NO adsorbs on the catalyst will be displaced by NH3 in the reactant pulse. The reverse approach is true for a complementary experiment using Raman spectroscopy, which in the case of SCR probes rather the catalyst structure than the adsorbed species. Vanadyl species with different degrees of coordination provide slightly different Raman signals. In particular, the presence of adsorbed species red-shifts the main vanadyl peak because of the decreased V¼O bond strength. Under SCR conditions, the vanadyl species exhibit varying degrees of coordination, resulting in a broad Raman peak for ν(V¼O). If NO is pulsed in the NH3 flow, all of the observable vanadyl species will remain in the NH3-coordinated state. As soon as they react with NO, they will be populated again with NH3 due to the strong acid-base interaction. This will not reveal additional information beyond what is delivered by the steady-state approach. For this reason, NH3 pulsing is the intuitive choice to observe changes in the distribution of vanadyl species and potentially reveal which is then responsible for the SCR activity. Yet another applicable technique is UV-visible spectroscopy, which delivers information on the oxidation state of the active metal species. For a modulation experiment of this type, O2 concentration would be the preferred stimulation while keeping NH3 and NO in constant supply. Without oxygen, the vanadyl sites are reduced under the joint presence of NH3 and NO. The second half of the catalytic cycle is initiated only when O2 is pulsed, reoxidizing the vanadyl sites to re-start the cycle. While NH3 or NO modulation could also cause oxidation state changes, only O2 pulsing can do so with the largest fraction of vanadyl species.

43

988

43.5

R. J. G. Nuguid and D. Ferri

Summary

The inherent complexity of the SCR reaction makes it an interesting case study for MES. We have shown that this experimental technique can amplify the presence of weakly adsorbed yet responsive species such as NO+ and at the same time filter out signals from species such as physisorbed NH3 that are not catalytically relevant but overwhelm the steadystate spectra. Demodulation is also capable of resolving overlapping peaks, as in the case of the stretching modes of NH3 and NH4+. Another important advantage of MES is its ability to isolate the signal of the active/responsive species (e.g., coordinatively unsaturated VOx sites) from the stationary and catalytically inactive species. Lastly, MES makes the detection of intermediate and short-lived species possible – a significant step in confirming proposed reaction mechanisms. In the mentioned examples, the increased sensitivity of MES provided molecular information and mechanistic insights that transcend the boundaries of conventional spectroscopy under steady-state conditions. Although the mechanism of SCR is still far from finalized, MES and other transientbased methods provide a promising route toward clarifying the open and still-debated questions about the reaction.

References 1. Ramis, G., Busca, G., Bregani, F., Forzatti, P.: Fourier transforminfrared study of the adsorption and coadsorption of nitric oxide, nitrogen dioxide and ammonia on vanadia-titania and mechanism of selective catalytic reduction. Appl. Catal. 64, 259–278 (1990) 2. Topsøe, N.-Y.: Mechanism of the selective catalytic reduction of nitric oxide by ammonia elucidated by in situ on-line Fourier transform infrared spectroscopy. Science. 265, 1217 (1994) 3. Lai, J.-K., Wachs, I.E.: A perspective on the selective catalytic reduction (SCR) of NO with NH3 by supported V2O5-WO3/TiO2 catalysts. ACS Catal. 8, 6537–6551 (2018) 4. Janssens, T.V.W., Falsig, H., Lundegaard, L.F., Vennestrøm, P.N.R., Rasmussen, S.B., Moses, P.G., Giordanino, F., Borfecchia, E., Lomachenko, K.A., Lamberti, C., Bordiga, S., Godiksen, A., Mossin, S., Beato, P.: A consistent reaction scheme for the selective catalytic reduction of nitrogen oxides with ammonia. ACS Catal. 5, 2832–2845 (2015) 5. Zhu, M., Lai, J.-K., Tumuluri, U., Wu, Z., Wachs, I.E.: Nature of active sites and surface intermediates during SCR of NO with NH3 by supported V2O5-WO3/TiO2 catalysts. J. Am. Chem. Soc. 139, 15624–15627 (2017) 6. Amiridis, M., Wachs, I., Deo, G., Jehng, J.-M., Soung Kim, D.: Reactivity of V2O5 catalysts for the selective catalytic reduction of NO by NH3: influence of vanadia loading, H2O, and SO2. J. Catal. 161, 247–253 (1996) 7. Centeno, M.A., Carrizosa, I., Odriozola, J.A.: In situ DRIFTS study of the SCR reaction of NO with NH3 in the presence of O2 over lanthanide doped V2O5/Al2O3 catalysts. Appl Catal B. 19, 67–73 (1998) 8. Soyer, S., Uzun, A., Senkan, S., Onal, I.: A quantum chemical study of nitric oxide reduction by ammonia (SCR reaction) on V2O5 catalyst surface. Catal. Today. 118, 268–278 (2006) 9. Grossale, A., Nova, I., Tronconi, E.: Role of nitrate species in the “NO2-SCR” mechanism over a commercial Fe-zeolite catalyst for SCR mobile applications. Catal. Lett. 130, 525–531 (2009)

10. Leistner, K., Mihai, O., Wijayanti, K., Kumar, A., Kamasamudram, K., Currier, N.W., Yezerets, A., Olsson, L.: Comparison of Cu/BEA, Cu/SSZ-13 and Cu/SAPO-34 for ammonia-SCR reactions. Catal. Today. 258, 49–55 (2015) 11. Went, G.T., Oyama, S.T., Bell, A.T.: Laser Raman spectroscopy of supported vanadium oxide catalysts. J. Phys. Chem. 94, 4240–4246 (1990) 12. Roozeboom, F., Mittelmeijer-Hazeleger, M.C., Moulijn, J.A., Medema, J., De Beer, V.H.J., Gellings, P.J.: Vanadium oxide monolayer catalysts. 3. A Raman spectroscopic and temperatureprogrammed reduction study of monolayer and crystal-type vanadia on various supports. J. Phys. Chem. 84, 2783–2791 (1980) 13. Giakoumelou, I., Fountzoula, C., Kordulis, C., Boghosian, S.: Molecular structure and catalytic activity of V2O5/TiO2 catalysts for the SCR of NO by NH3: in situ Raman spectra in the presence of O2, NH3, NO, H2, H2O, and SO2. J. Catal. 239, 1–12 (2006) 14. Su, W., Li, Z., Peng, Y., Li, J.: Correlation of the changes in the framework and active cu sites for typical Cu/CHA zeolites (SSZ-13 and SAPO-34) during hydrothermal aging. Phys. Chem. Chem. Phys. 17, 29142–29149 (2015) 15. Greenaway, A.G., Marberger, A., Thetford, A., Lezcano-González, I., Agote-Arán, M., Nachtegaal, M., Ferri, D., Kröcher, O., Catlow, C.R.A., Beale, A.M.: Detection of key transient Cu intermediates in SSZ-13 during NH3-SCR deNOx by modulation excitation IR spectroscopy. Chem. Sci. 11, 447–455 (2020) 16. Negri, C., Signorile, M., Porcaro, N.G., Borfecchia, E., Berlier, G., Janssens, T.V.W., Bordiga, S.: Dynamic CuII/CuI speciation in Cu-CHA catalysts by in situ diffuse reflectance UV–vis-NIR spectroscopy. Appl. Catal. A. 578, 1–9 (2019) 17. Gao, F., Walter, E.D., Karp, E.M., Luo, J., Tonkyn, R.G., Kwak, J.H., Szanyi, J., Peden, C.H.F.: Structure-activity relationships in NH3-SCR over Cu-SSZ-13 as probed by reaction kinetics and EPR studies. J. Catal. 300, 20–29 (2013) 18. Nuguid, R.J.G., Ferri, D., Kröcher, O.: Design of a reactor cell for modulated excitation Raman and diffuse reflectance studies of selective catalytic reduction catalysts. Em. Control Sci. Technol. 5, 307–316 (2019) 19. Urakawa, A., Bürgi, T., Baiker, A.: Sensitivity enhancement and dynamic behavior analysis by modulation excitation spectroscopy: principle and application in heterogeneous catalysis. Chem. Eng. Sci. 63, 4902–4909 (2008) 20. Baurecht, D., Fringeli, U.P.: Quantitative modulated excitation Fourier transform infrared spectroscopy. Rev. Sci. Instrum. 72, 3782–3792 (2001) 21. Zecchina, A., Marchese, L., Bordiga, S., Pazè, C., Gianotti, E.: Vibrational spectroscopy of NH4+ ions in zeolitic materials: an IR study. J. Phys. Chem. B. 101, 10128–10135 (1997) 22. Ohsaka, T., Izumi, F., Fujiki, Y.: Raman spectrum of anatase, TiO2. J. Raman Spectrosc. 7, 321–324 (1978) 23. Nuguid, R.J.G., Ferri, D., Marberger, A., Nachtegaal, M., Kröcher, O.: Modulated excitation Raman spectroscopy of V2O5/TiO2: mechanistic insights into the selective catalytic reduction of NO with NH3. ACS Catal. 9, 6814–6820 (2019) 24. World Scientific Publishing Company: The Chemical Dynamics and Kinetics of Small Radicals. World Scientific Publishing Company, Singapore (1996) 25. Marberger, A., Ferri, D., Elsener, M., Kröcher, O.: The significance of Lewis acid sites for the selective catalytic reduction of nitric oxide on vanadium-based catalysts. Angew. Chem. Int. Ed. 55, 11989–11994 (2016) 26. Marchionni, V., Ferri, D., Kröcher, O., Wokaun, A.: Increasing the sensitivity to short-lived species in a modulated excitation experiment. Anal. Chem. 89, 5801–5809 (2017) 27. Maeda, N., Meemken, F., Hungerbühler, K., Baiker, A.: Spectroscopic detection of active species on catalytic surfaces: steady-state versus transient method. Chimia. 66, 664–667 (2012)

43

Case Study 1: Modulation Excitation Spectroscopy (MES)

989

43 Rob Jeremiah G. Nuguid currently works at Linde GmbH (Germany). He obtained his PhD degree from École polytechnique fédérale de Lausanne and Paul Scherrer Institut (Switzerland), where he used time-resolved spectroscopy to uncover the mechanisms of catalytic reactions.

Davide Ferri received his PhD in chemistry at the ETH Zurich in 2002. He is currently head of the group Applied Catalysis and Spectroscopy at the Paul Scherrer Institut (Switzerland). His interests span from environmental catalysts to liquid phase catalyzed reactions and development of in situ/operando experiments. He uses vibrational spectroscopies in combination with X-ray based methods.

Case Study 2: Modulation Excitation Spectroscopy (MES) Sebastian Collins

and Laura Briand

Contents

Abstract

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biocatalyzed Kinetic Resolution of Racemic Profens with Lipases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanism of the Enzymatic Kinetic Resolution of Racemic Profens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental Evidences of the Acyl-Enzyme Intermediate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

993

44.2 44.2.1 44.2.2 44.2.3 44.2.4 44.2.5

Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attenuate Total Reflection Infrared Spectroscopy . . . . . . . Isotopic Exchange H-D of the Enzymes with D2O . . . . . . MES Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MCR-ALS Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

994 994 994 995 996 996

44.3 44.3.1

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 997 Effect of the Nature of the Liquid Environment in the Secondary Structure of Lipases . . . . . . . . . . . . . . . . . . . . 997 MES-PSD Approach for the Molecular Recognition of an Acyl-Enzyme Intermediate . . . . . . . . . . . . . . . . . . . . . . . . . 999

44.1 44.1.1 44.1.2 44.1.3

44.3.2 44.4

44

991 992 993

Conclusions and Future Perspectives . . . . . . . . . . . . . . . . . . 1000

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1001

S. Collins (*) Instituto de Desarrollo Tecnológico para la Industria Química (INTEC), CONICET-UNL, Santa Fe, Argentina Chemical Engineering Faculty – Institute for Technological Development of Chemical Industry (INTEC), National University of Litoral – National Scientific and Technical Research Council, Santa Fe, Argentina e-mail: [email protected] L. Briand Centro de Investigación y Desarrollo en Ciencias Aplicadas – Dr Jorge J. Ronco CINDECA-CCT La Plata-CONICET, Buenos Aires, Argentina Chemistry Faculty – Center of Research and Development in Applied Sciences (CINDECA), National University of La Plata UNLP – National Scientific and Technical Research Council (CONICET), La Plata, Argentina e-mail: [email protected]

In this case of study, we provide examples of the use of infrared spectroscopy in attenuated total reflection (ATR) to investigate enzymes at work. An optimized ATR flowthought cell was developed, which allows performing modulation excitation spectroscopy with phase-sensitive detection (MES-PSD) experiments in a controlled way. The reversible effect of solvents with different hydrophobicity on the secondary structure of a Thermomyces lanuginosus lipase (TTL) was demonstrated by an analysis of the amide I band. Finally, the lipase B of Candida antarctica (CALB), active to the enantioselective esterification of ketoprofen with alcohols, was studied. MES-PSD analysis allowed to univocally identifying signals belonging to the interaction between ketoprofen and the lipase, that is the acyl-enzyme intermediate. Keywords

MES-PSD · ATR-FTIR · Lipases · Acyl-enzyme complex · CALB

44.1

Introduction

Catalytic processes at liquid/catalyst interfaces are highly important in a number of chemical and biochemical systems. Biocatalyzed reactions for synthetic transformations at industrial scale, and particularly for enantioselective transformations, are of key importance in modern green chemistry [1]. Spectroscopic studies using in situ or operando methods in the presence of the solvent are essential in order to postulate reliable reaction pathways with meaningful kinetic data. Understanding of reaction mechanisms and structure-activity relationships under working conditions is of central importance to perform a rational design of new and optimized biocatalytic systems [2]. In this matter, rigorous operando experimentation has been developed in the recent years,

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_44

991

992

where the simultaneous measurement of conversion and selectivity, alongside with the acquisition of spectroscopic data for the catalyst under steady-state condition, was made [3]. Thus, increasing efforts to implement cells/microreactors to spectroscopically assess the reactant/catalyst interface, while simultaneously measuring activity and selectivity, have been carried out by several research groups [4–6]. In previous works, an optimized flow-through attenuated total reflection (ATR) infrared (IR) spectroscopy cell for spectrokinetic analysis was developed and fully characterized, which showed that it is possible to obtain intrinsic kinetic information of surface-catalyzed reactions in liquid phase [7, 8]. Nevertheless, when investigating reaction mechanisms under steady state, it is virtually impossible to selectively discriminate between signals of true intermediates from signals of solvents, spectator species, and/or the catalyst itself, e.g., the background, due to the static nature of the measurements. Moreover, the static signals are usually much more intense than those from the intermediates, making very difficult the identification and tracking of intermediates. Then, transient experiments have been developed and increasingly used for the analysis of reaction intermediates by perturbing a catalytic system under working conditions. Thus, accurate background compensation can be carried out by difference spectroscopy. In this matter, two distinct alternatives have come out for external stimulation, which additionally can supply kinetic information of the reaction: (i) relaxation technique, e.g., parameter jump, and, more recently, (ii) modulated excitation, e.g., periodic parameter change. The latter can only be applied when the system responds reversibly to a periodic external excitation, for instance, in the absence of catalyst deactivation. In this chapter, modulation excitation spectroscopy (MES) coupled with phase-sensitive detection (PSD) analysis was thoroughly explained [9–11]. In this case of study, we present examples of characterization of enzymes and detection of intermediate species of biocatalyzed reactions using ATRFTIR with MES-PSD spectroscopy. In order to achieve this goal, some specific tools must be developed. First, an optimized flow-through ATR cell was designed, constructed, and characterized. This ATR cell was mounted in an experimental setup that allows performing transient and modulated (MES) experiments. Second, to investigate the secondary structure of enzymes (proteins in general), amide I signal is employed. This infrared band arises around 1700–1600 cm
1, e.g., in the same range of the bending mode of water (H2O). Then, to avoid this spectra interference, D2O exchange in the enzyme must be performed previously to any ATR in situ, operando, or MES experiment. We present here fundamental experimental procedures and results showing the importance of D2O exchange. Finally, examples of MES experiments are presented, which allows investigating changes in the

S. Collins and L. Briand

secondary structure and identify acyl-enzyme intermediates under reaction. A special procedure to resolve highly overlapped MES-PSD spectral signals based on a chemometric technique is described and employed to analyze small, reversible, changes in amide I band.

44.1.1 Biocatalyzed Kinetic Resolution of Racemic Profens with Lipases Lipases (triacylglycerol hydrolase, EC 3.1.1.3) are a family of enzymes broadly applied in organic chemistry due to their ability to catalyze reactions of hydrolysis and synthesis such as esterification, transesterification, interesterification, and alcoholysis among others [12–16]. Due to their high activity, substrate-, regio-, and stereo-specificity along with their mild reaction conditions, these enzymes are chosen for application in several industries: food, cleaning, pharmaceutical, biodiesel, cosmetic, and paper. Pharmaceutical industries apply lipase enantioselectivity for preparing different drugs and fine chemicals containing a chiral center [12, 15–22]. Lipases are one of the most suitable enzymes for kinetic resolution of profens due to their wide substrate specificity, their ability to recognize chirality, and they do not require labile cofactors. Profens (2-arylpropionic acids) constitute an important group of racemic nonsteroidal anti-inflammatory substances (NSAIDs). Their pharmacological activity resides on the S-enantiomer. These are widely used for relieving pain and fever, and for the treatment of inflammatory diseases, such as rheumatoid arthritis and osteoarthritis. The R-enantiomer often depicts unwanted gastrointestinal side effects mainly in the stomach and duodenum, such as pyrosis, dyspepsia, gastritis, or diarrhea. The appearance of gastric or duodenal mucosal injuries in chronically treatments are of major concern since those erosions and ulcers lead to bleeding or perforations [17, 18]. The clinical investigations concerning the medical use of dexenantiomers provide evidences that the S-enantiomers might be a healthier choice than the racemic counterpart since a lower uptake is needed for a faster and reliable effect. This observation has been the motivation of numerous investigations toward the synthesis of pure dexketoprofen and dexibuprofen between other enantiopure pharmaceuticals. The investigations regarding the kinetic resolution of racibuprofen and rac-ketoprofen with lipases are summarized chronologically within the references [17] and [18, 23]. The authors present the various acyl acceptors, reaction conditions (molar ratio of substrates, temperature, and others), cosolvents, conversion, and enantiomeric excess of the esterification of the profens with alcohols. The immobilized lipase B of Candida antarctica known as Novozym ® 435 is the most widely investigated even though some problems such

44

Case Study 2: Modulation Excitation Spectroscopy (MES)

993

enzyme desorption under certain conditions and mechanical fragility of the polymeric support, accumulation of hydrophilic compounds, and support dissolution in some organic media are well documented in the literature [24].

formed in the interaction of ketoprofen with the serine residue of the active site of CALB proposed by Toledo et al. [30]. Particularly, a molecular mechanics calculation in the framework of the Chem3D Cambridge Soft was performed to explore the steric interactions between ketoprofen and the serine hydroxyl with the presence of one water molecule interacting with ketoprofen through hydrogen bonding. Additionally, DFT calculations allowed obtaining FTIR bands of the acyl-enzyme model that somehow paved the way to undoubtedly identify the intermediate through modulated infrared spectroscopy MES-PSD as will be discussed in the following section.

44.1.2 Mechanism of the Enzymatic Kinetic Resolution of Racemic Profens Undoubtedly, the literature demonstrates that lipases are the most studied biocatalysts in the kinetic resolution of racemic compounds. In fact, unveiling of the mechanism involved in the catalytic behavior of CALB is the core of the following sections. Thermomyces lanuginosus lipase (TTL) and the lipase B of Candida antarctica (CALB) are composed of 314 and 317 amino acid residues and a molecular mass of about 33 kDa [24, 25]. Similar to other serine α/β-hydrolases, a Ser-His-Asp (S146-H258-A201 for TLL and S105-H224D187 for CALB) catalytic triad partially covered with a lid and an oxyanion hole formed by threonine (T40) and glutamine (Q106) are responsible for the catalytic activity of lipases [25]. The mechanism of action of CALB has been described as bi-bi ping-pong with the formation of two tetrahedral intermediates and an acyl-enzyme complex, and a competitive substrate inhibition by the alcohols used as acyl acceptors as depicted in Fig. 44.1 [26]. This scheme shows an ester R1OOR2 that enters the active site and is adsorbed onto the oxyanion hole formed by T40 and Q106. Then, the hydrogen atom of serine S105 is transferred to the histidine H224 residue of the catalytic triad and the Oγ nucleophilically attacks the carbonyl group of the substrate to form the first tetrahedral intermediate that is stabilized by the oxyanion hole. The acyl-enzyme complex is formed when the hydrogen atom in histidine H224 is transferred to the oxygen atom to release an alcohol R2OH. The acyl-enzyme complex suffers a nucleophilic attack by either water in the case of a hydrolysis or alcohol for the esterification reactions. Then, a second tetrahedral intermediate is formed. Finally, the product is released and the lipase is regenerated [27–29]. Figure 44.2 shows a model of the acyl-enzyme intermediate

44 44.1.3 Experimental Evidences of the Acyl-Enzyme Intermediate The literature shows few investigations (dating in the 90’) devoted to the identification of the acyl-enzyme even though this species is a key intermediate to validate the mechanism of a biocatalytic process. The acyl-enzyme complex has been determined experimentally through infrared spectroscopic techniques in the hydrolysis of β-lactam antibiotics by enzyme β-lactamase and ornithine acetyl transferase (OAT) and in the hydrolysis reaction of trans-cinnamoyl imidazole by α-chymotrypsin [31–35]. Early studies of the hydrolysis of trans-cinnamoyl imidazole with α-chymotrypsin have been performed through attenuated total reflectance ATR Fourier transform infrared spectroscopy [31]. The experiments involved a stable transcinnamoyl-α-chymotrypsin acyl-enzyme species prepared in advance and placed on a germanium crystal surface as a thin film for investigation. Wilkinson et al. [32] detected four acyl-enzyme species (conformers) in the hydrolysis of methicillin with β-lactamase. The reaction was performed in IR cuvettes while acquiring the time-resolved infrared spectra in a static kind of approach [32–34]. Similarly, to the case discussed above, the acyl-enzyme is stable and its concentration remains constant until the substrate is exhausted and then diminishes with a first-order rate constant of 9.1  10
3 s
1.

O O

O O O

N

Q106 S105

O T40 HH H N

O

O R1

O

R2

Q106

O– H N D187

H224

H

R1 O O S105 H O N O– H N

H

O

N

O

O

D187

O

O H H N T40 R2

Q106

S105

H224

H

O R1 O

Q106

O– H N

H

Fig. 44.1 The ping-pong bi-bi mechanism of Candida antarctica lipase B

O O

O R1

H2O

O

S105

D187

H

R1

O

H

Q106 S105

H –

O HN D187

O T40 HH H N O

O

N

H224

N

O H

O– H N

N

H224

O

O H H N T40

N

O

O D187

O H H N T40

N

R2-OH

O

N

H224

994

S. Collins and L. Briand

Ser Complex at 1749 cm–1

CH2 O

H O H

O C

H

N

H

O

O C

CH3 H

Ser Complex at 1734–1736 cm–1

CH2 O

H O O C

H

H

N

H

O

active site produces a stable amide bond with the acylenzyme complexes. This modification of the active site allowed investigating thioester and ester intermediates in the synthesis of valinomycin catalyzed with a cysteine protease and a thioesterase using tandem mass spectrometry. Back to infrared spectrometry, the most important advancement in the identification of unstable acyl-enzyme species developed in the last years is the modulation excitation spectroscopy (MES) with phase-sensitive detection (PSD). This methodology allowed to univocally distinguishing the signals belonging to the acyl-enzyme species due to the interaction between ketoprofen and the lipase B of Candida antarctica as will be discussed in the following section [30].

O C

44.2

CH3

Experimental Setup

H

44.2.1 Materials Ser Complex at 1756 cm–1

CH2 H O C

H

O

O

H

N

H

O

O C

CH3 H

Fig. 44.2 Complexes between ketoprofen and serine interacting with the surroundings, the acyl-enzyme intermediate with one molecule of water and their infrared signals. The purple color in C-O-C shows the new bond

Ornithine acetyl transferases (OATs) are enzymes that catalyze the reversible transfer of an acetyl group to/from N-α-acetylated substances such as L-ornithine and L-glutamate via an acyl-enzyme intermediate. Again, the acyl-enzyme complex of OAT is relatively stable, therefore is an excellent candidate to be studied with infrared spectroscopy in a static way [35]. More recently, alternative methodologies based on solidstate 17O NMR have been developed to investigate unstable acyl-enzyme intermediates [36]. In this context, a freezedrying intermediate species between 17O-labeled N-acyl imidazole substrates and a serine protease (α-chymotrypsin) was studied. Besides the spectroscopy-based methods to detect such intermediates, the biological approach is also interesting. In this sense it is worth mentioning the investigation of Huguenin-Dezot et al. that incorporated 2,3-diaminopropionic acid (DAP) into recombinant proteins, via expansion of the genetic code [37]. The replacement of cysteine or serine residues with DAP within the catalytic

Highly pure Candida antarctica lipase B, CALB (35,500 g/ mol) was purchased from Sigma-Aldrich (10.9 U/mg). Thermomyces lanuginosus (TTL) was kindly provided by Novozymes (Brasil). Papain was purified to mass spectrometry degree from dry latex of Carica papaya fruits obtained in Jujuy, Argentina. (R/S)-ketoprofen (Parafarm, 99.8%, batch 030718 000928/004), ethanol (Carlo Erba), methanol (Carlo Erba, 99.8%), carbon tetrachloride (Merck, 99.5%), and deuterium oxide D2O (Cambridge Isotope Laboratories) were used.

44.2.2 Attenuate Total Reflection Infrared Spectroscopy Attenuated total reflection (ATR) infrared (IR) spectroscopy is an appropriate tool for the investigation of reaction pathways in liquid(reactive)/solid(catalyst) systems, because it provides the detection of species adsorbed on a catalyst under reaction conditions [38]. Catalysts are usually deposited on the internal reflection elements (IRE) as films (e.g., metal film) or as layers of powders or enzymes, and they are exposed to the liquid-phase reactants. In order to obtain quantitative information, that is, determining intrinsic reaction rates, the chemical engineering aspects of an ATR flow-through cell must be developed. Particularly, mass transport into the ATR cell has to be considered. With this aim, we have designed and characterized an optimized flow-through ATR microfluidic cell, which transports dynamics in transient-flow experiments governed by a convection-diffusion mechanism [7, 8]. The ATR cell (Fig. 44.3), with a total volume of 60 μL, has linear-shaped entrance and exit ports close to the extremes of the cell, which

44

Case Study 2: Modulation Excitation Spectroscopy (MES)

995

avoid dead-volume zones and enable uniform fluid velocity profile across the cell. A 45 trapezoidal (80  10  4 mm) ZnSe crystal was used, providing ten internal reflections at the liquid/crystal interface (penetration depth ¼ 1.66 μm for n2 ¼ 1.4 at 1000 cm
1) and a cutoff spectral range at 650 cm
1. The temperature of the cell was controlled by a heating jacket connected to a thermostatic bath. To avoid fluctuation of the reactant temperature, the glass flasks containing the liquid reactants (solutions) are immerse in a thermostic bath, and all the tubes, as well the three-way valve, are heated with a heating wire (nichrome). This setup ensures the temperature control of the reactants into the cell. Figure 44.4 presents a scheme of the experimental setup. The cell was mounted onto an ATR attachment (Pike Technologies) inside the sample compartment of the FTIR spectrometer (Thermo-Electron, Nicolet 8700 with a cryogenic MCT detector). The bench of the spectrometer was continuously purged with dried air (Parker Balston FTIR purge gas generator) to eliminate CO2 and water vapor contributions to the Fig. 44.3 Scheme of the optimized flow through ATR cell

spectra. Time-resolved ATR-FTIR spectra were recorded at a resolution of 4 cm
1 (up to 1 spectrum/3 s). Liquids were pumped using a pulse-free peristaltic pump (Ismatec ICP4) located at the end of the cell. A pneumatically actuated threeway valve controlled by a computer software allows switching the solution streams. Repetitive square-wave stimulations with flow rates from 0.5 to 2 mL/min and modulation frequencies from 50–200 mHz could be generated into the ATR cell. More details of the experimental setup can be found elsewhere [7].

44.2.3 Isotopic Exchange H-D of the Enzymes with D2O Before the ATR-FTIR experiments, the isotopic exchange of water molecules in the enzyme by D2O molecules must be performed in order to allow the investigation of the amide I

Zero-volume fitting

Top body Holes for screw

Cavity to host the O-ring

Drilled jacket

Mirror ZnSe Lower body

Fig. 44.4 Scheme of the ATR experimental setup: (1) ATR cell; (2) FTIR spectrophotometer; (3) peristaltic pump; (4) pneumatic actuated valve; (5) reactants; (6) thermostatic bath; (7) wastes; and (8) PC

5

8 4 6 3 1

7 2

FTIR

44

996

signal (1700–1600 cm
1) without the interference of the bending vibration of O-H species at 1640 cm
1. A very important issue corresponds to the kinetic of the isotopic exchange upon exposing enzymes to D2O. To determine the completeness of the isotopic exchange, the D2Opapain (as a model enzyme) exchange was investigated by time-resolved ATR infrared spectroscopy. The papain sample was deposited as a thin film on the ZnSe crystal, and IR spectra were collected once every 5 min during the first 60 min and once every hour for 24 hs. The H-D isotopic exchange of the water molecules in the highly pure Thermomyces lanuginosus (TTL) and the lipase B of Candida antarctica (CALB) was performed by dissolving 5 mg of enzyme in 500 μL of D2O (D > 99%). The mixture was incubated for 24 h and then dispersed over a ZnSe crystal order to prepare a homogeneous film. Then, the cell was closed and purged overnight with dry air at room temperature in order to obtain a dry film. The thickness of the film can be controlled by the concentration of the dispersed enzyme and performing successive deposition-drying cycles. For the present studies, the prepared films possess a thickness lower than the penetration depth of the IR beam in order to have access to both, the signals from the reactant/solvent and the enzyme/intermediates [7, 10].

44.2.4 MES Experiments Modulation excitation spectroscopy (MES) experiments were carried out for investigating the effect of solvents with different hydrophobicity on the secondary structure of TTL and to identify the acyl-enzyme intermediate on CALB. The effect of the solvents on the secondary structure of a TTL was investigated through MES experiments by alternatively flowing 2-propanol (Log P ¼ 0.3) and n-heptane (Log P ¼ 7.2) located in a thermostatic bath (333 K) at a flow rate of 0.5 mL/min. Activated molecular sieve (3 Å) was introduced into the flasks to remove possible water impurities. MES experiments were also performed in order to identify acyl-enzyme intermediate. Experiment was started by varying the inlet composition from carbon tetrachloride (Carlo Erba HPLC grade) to rac-ketoprofen 0.16 M in CCl4, using the desired modulation frequency at 318 K. After allowing at least five modulation periods to adjust the system to the external perturbation the recording of the spectra was started. Spectra were acquired every 3 s, during each c-MES period, using reactants exchange frequencies from 1.7–33 mHz. Liquids were provided from two separate glass bottles. Phase-sensitive detection (PSD) analysis of the spectra was performed according to the method developed by Baurecht and Fringeli [9].

S. Collins and L. Briand

44.2.5 MCR-ALS Procedure As stated before, applications of MES-PSD has gained popularity to investigated heterogeneous catalysis mechanisms, as well as enzymatic reactions. Usually, spectroscopic data gathered from different intermediate species, reaction sites, or secondary structure of proteins show highly overlapped bands, which represent a challenge when individual target signals must be identified. Then, the combination of MES-operando and signal processing by PSD with chemometric analysis could improve the capability to obtain valuable quantitative information. Automated data processing procedures, such as least-square fitting of data with linear combination (LC) of reference spectra and principal component analysis (PCA), are powerful tools. However, these approaches have several limitations, particularly LC requires standards and PCA often provides mixed components that are not easily interpretable. In this context, multivariate curve resolution–alternating least squares (MCR-ALS) is a widespread iterative soft modeling method that performs decomposition of a data matrix into two submatrices that comprise chemically meaningful information for the compounds involved in the system, i.e., pure spectral profiles and the varying intensity [39, 40]. Recently, a novel and systematic chemometric procedure was implemented to model MES-PSD infrared data comprising highly overlapped bands [41]. The strategy involves successive MCR-ALS steps and allowed extracting spectral and kinetic information of highly overlapped PSD signals. This procedure was firstly developed using simulated data for a consecutive first-order kinetic. The simulated spectral signals covered the cases from noiseless and nonoverlapped spectral bands to highly overlapped spectral bands with high noise level. One of the important accomplishments reached from this analysis was that, even though MCRALS tends to be operator dependent, the presented strategy can be applied without the necessity of introducing modifications. Figure 44.5 shows a general data-processing proceedings: (i) first, the original MES-PSD data (a) is transformed by means of computing the absolute value (b); (ii) PCA loadings are calculated, then the first MCR-ALS resolution is performed in the absolute MES-PSD data by using purest variables as initial estimates obtaining the spectral and phase lag profiles (c); and (iii) a second MCR-ALS is performed in the original MES-PSD data (a) using the previously optimized spectral profiles to achieve spectral and phase lag physically recognizable profiles (d). This MCR-ALS strategy was validated to resolve highly overlapped MES-PSD infrared spectra collected from an ATR-FTIR study of the adsorption-desorption dynamic of oxalic acid on titanium dioxide [41]. The chemometric

44

Case Study 2: Modulation Excitation Spectroscopy (MES)

a)

997

b)

Q(O-D) ~2472

Absolute value MCR-ALS II

d)

(iii)

(ii)

MCR-ALS I

c)

Absorbance (a.u.)

(i)

Amide II ~1400 Amide III 1312 G(O-D) 1206 Q(N-D) 1073

Q(O-H)/(N-H) 3378 18 h

Amide I 1649

44 18 00 16 00 14 00 12 00 10 00 80 0

25 00

30 00

35 00

40 00

1h

Wavenumber (cm–1)

Fig. 44.5 Scheme of the MCR-ALS data processing for MES-PSD data: (i) original MES-PSD data (a) is transformed by means of computing the absolute value (b); (ii) the first MCR-ALS resolution is performed in the transformed MES-PSD data (b) by using purest variables as initial estimates obtaining the phase lag (c, top) and spectral (c, bottom) profiles; and (iii) a second MCR-ALS is performed in the original MES-PSD data (a) using the previously optimized spectral profiles (c, bottom) to achieve the final spectral (d, bottom) and phase lag (d, top) profiles

resolution allowed identifying the individual species that are involved in the system, as well as obtaining kinetic information in terms of phase lag. It was concluded that this procedure for MES-PSD data resolution is an effective, robust, and powerful tool to obtain relevant information of a complex spectroscopic systems. This procedure will be employed for the amide I resolution of individual components when investigating the reversible changes of secondary structure of a lipase.

44.3

Results

44.3.1 Effect of the Nature of the Liquid Environment in the Secondary Structure of Lipases The ability of lipases to catalyze reactions in organic solvents, or low water environments, is well established. As well, water content plays a fundamental role in structure and function of these enzymes. Water retains the catalytically active three-dimensional conformation and contributes to structural integrity, active site polarity, protein stability, and flexibility [42, 43]. Additionally, water can influence lipases enantioselectivity by binding to substrate-binding pockets

Fig. 44.6 Infrared spectra obtained every 60 min for 18 h of a liquid mixture of papain in D2O in the in situ ATR cell at room temperature

(stereospecificity pockets) [44]. The nature of the solvent and water activity directly influences the structure and flexibility of the active site of lipases B [45]. Nonpolar solvents, such as octane, allow active site hydration through water molecules which freely move in the enzymatic surface [46]. Molecular dynamics studies evidenced that in systems with low water content, polar solvents (such as tetrahydrofuran and acetonitrile) remove water and penetrate the active site of the enzyme. Moreover, enzymes are usually freeze-dried (lyophilized) and then suspended in organic solvents, which can affect the catalytic activity of these systems and their secondary structure [47]. Fundamental structure-activity relationships of enzymes are obtained through the investigation of the secondary structure of the proteins. The infrared band from the amide I vibration, absorbing near 1650 cm
1, is produced mainly from the C-O stretching, and a combination of out-of-phase C-N stretching vibration and the NH in-plane bend [48]. The amide I band composition of polypeptides and proteins is sensitive to secondary structure and it is most commonly used for secondary structure analysis [48–51]. However, the infrared spectra of such band are greatly influenced by the presence of water that is inherent to the proteins’ structure. In this context, the exchange of water and hydrogen by D2O appears as an ineludible requisite to obtain reliable spectroscopic information. Now a question rises about the extent of the isotopic exchange of a protein and in this sense previous investigations on the progress of the H/D isotopic exchange of papain (as a model enzyme) provided evidences of the evolution of the various amide bands and the completeness of the exchange [52]. Figure 44.6 shows the in situ ATR-FTIR spectra of papain in contact with deuterium

998

S. Collins and L. Briand

oxide at room temperature up to 18 h. An intense band ascribed to amide A and B due to the stretching vibrations of O-H and N-H is centered at 3378 cm
1. Additionally, the characteristic stretching and bending vibrations of the O-D species at ~2472 cm
1 and 1206 cm
1 are observed. The signal of low intensity of amide I appears at 1649 cm
1 while the broad band at about 1400 cm
1 is assigned to amide II. This signal arises due to the vibrations of the N-H and C-N species of the protein. The former is easily exchanged by deuterium atoms shifting the signal toward low frequency, decreasing its intensity and enhancing the resolution of the amide II band. The Fig. 44.7 presents a closer look to the evolution of that signal upon H/D exchange. The broad signal splits into two signals at 1404 cm
1 and 1369 cm
1 upon progression of deuteration and the signal ascribed to the amide III becomes apparent at 1311 cm
1. A complete isotopic exchange is reached after 10 h. At this point, it is suitable to perform the deconvolution of the amide I signal in its various elements to obtain reliable information of the secondary structure of the protein.

Table 44.1 shows the assignments of the infrared signals (both average and extreme wave numbers) observed for the various components of the amide I band upon H/D isotopic exchange. The contribution of the α-helix and random/disordered structures correspond to the signals at 1652 cm
1 and 1645 cm
1, respectively. The contribution of the β-sheet structure corresponds to the signals appearing at 1630 cm
1 and 1679 cm
1 wave numbers. Similarly, the contribution of the β-turn structure appears at 1664 cm
1, 1666 cm
1, and 1676 cm
1. Finally, a signal at 1612 cm
1 corresponds to β aggregates [23, 30, 49]. MES ATR-FTIR was employed to investigate the effect of solvents with different hydrophobicity on the secondary structure of a Thermomyces lanuginosus (TTL). The secondary structure, after the complete exchange with D2O, was studied using the amide I band in order to avoid the interference of the strong band at 1645 cm
1 of H2O. Time-resolved spectra were measured during cyclic exchange between n-heptane (log P ¼ 7.2) and 2-propanol (log P ¼ 0.3) within periods of 60, 120, and 240 s. Figure 44.8 shows time-

a) Amide III

Absorbance (a.u.)

1404

1369 1311

0.25 δ(N-D) 1073

δ(O-D)

18 h

1h

b) 0.005 1800

1600

1400

1200

1000

800

Wavenumber (cm–1)

0.2

Fig. 44.7 Evolution of the infrared bands of the peptide linkage from 2 to 18 h of the isotopic exchange with D2O. The spectrum obtained after 60 min of the isotopic exchange was subtracted of the successive spectra

1700 1680 1660 1640 1620 1600 1580

Table 44.1 Assignment for amide I’ band position to secondary structure [48–51] Secondary structure α-Helix β-Sheet Unfold/Aggregated Disordered/Random

Band position of amide I’ (D2O) cm
1 Average Extremes 1652 1642–1660 1630 1615–1638 1679 1672–1694 1612 1605–1625 1645 1639–1654

3500

3000

2500 2000 Wavenumber (cm–1)

1500

1000

Fig. 44.8 ATR-FTIR spectra (a) in the time domain and (b) phase domain after PSD for the periodic exchange of neat n-heptane and neat 2-propanol, flow rate of 1.2 mL/min, modulation equals to 8.33 mHz and 24 spectra per cycle. The residence time of the cell (τ) equal to 2.4 s

44

Case Study 2: Modulation Excitation Spectroscopy (MES)

resolved and phase-resolved spectra in the amide I region, calculated for the fundamental frequency (k ¼ 1). Due to the highly overlapped signals in the amide I region, the resolution of bands was obtained by applying a multivariate curve resolution–alternating least squares (MCR-ALS) procedure developed specifically to analyze phase-resolved spectra [41]. Figure 44.9 shows the resolved components at their maximum amplitude. The spectra correspond to the components of the amide I signals that are detected during the modulation of the hydrophobic and hydrophilic environments in the MES experiment. It comes clear that the PSD spectra in the amide I region has three components, which can be assigned to α-helix and intermolecular hydrogen-bonded antiparallel β-sheets (at about 1657 cm
1 and 1618 cm
1 in the red spectra, respectively), and β-sheet and/or random structures (at 1627 cm
1, 1638 cm
1, and 1647 cm
1 in the blue spectra). A decrease in the intensity and an enhancement of the resolution of the amide I have been observed previously upon progression of the H/D exchange of papain due to the perturbation of the hydrogen bonds and van der Waals interactions between the enzymatic residues [52]. It is well known that the substrate access to the active catalytic triad of TLL is regulated by conformational changes of a helical lid located over the active site [53]. The “closed” conformation prevails in aqueous media (low log P) and, therefore, the lipase displays very low catalytic activity. However, in hydrophobic (organic, high log P) media or in the presence of an organic-aqueous mixture, the “open” form of the lipase allows the interaction of the substrate with the active site and a higher activity is observed. Even though the MES ATR-FTIR experiment is exposing the TTL enzyme to alternating hydrophilic and hydrophobic

j = 330° 1640 j = 290° 1657 1618

1700

1680

1660

1640

1620

1600

1580

Wavenumber (cm–1)

Fig. 44.9 Spectral profiles reached after applying the MCR-ALS chemometric procedure to the region of amide I

999

solvents, there is an interface between 2-propanol and n-heptane until one of the solvents is completely replaced. That interface might provide the appropriate activation (by modifying the hydrogen-bridged bonds and van der Walls interactions) to modify the enzymatic conformation and “open” the helical lid. Somehow, this situation mimics the so-called interfacial activation of the lipases under reaction in an organic-aqueous media. In the present experiment, the change in the amide I bands is in phase with the hydrophobic solvent. Finally, it must be emphasized that the intensity change during the exposure to the solvents is around 2% of the amide I band, which highlights the very high sensitivity of the MES-PSD experiments.

44.3.2 MES-PSD Approach for the Molecular Recognition of an Acyl-Enzyme Intermediate Some of us provided evidences of the acyl-enzyme species involved in the interaction of rac-ketoprofen with the lipase B from Candida antarctica through in situ time-resolved ATR-FTIR both in static and transient conditions [30]. Modulation excitation spectroscopy with phase-sensitive detection (MES-PSD) allowed to univocally distinguishing the signals belonging to the interaction between ketoprofen and the lipase from the strong background signals as will be discussed later on. The powerfulness of the experimental evidence and theoretical DFT analysis enlightened the presence of key intermediates in the esterification and hydrolysis biocatalyzed with lipases [23, 30, 54]. The experiment starts by placing a film of highly purified CALB on the zinc selenide crystal of the ATR cell and allowed to exchange with deuterium oxide at 318 K overnight. It is worth noticing that the commercial CALB lipase (c.a. Lipozyme®) possesses a series of additives that readily interferes in infrared spectra. Previous investigations reported by Llerena Suster et al. proved the presence of a nonsoluble fraction of cellular debris, nucleic acids, benzoate, and sorbate species and a mixture of three proteins [55]. This observation makes the purification of the lipase a key step in order to assure a reliable spectroscopic investigation of the interaction enzyme substrate. In general and briefly, the purification involves the separation of the nonsoluble fraction of the crude extract through centrifugation. Then the centrifuged sample is treated with size exclusion chromatography to obtain an enzymatic sample free of nucleic acids, benzoate, and sorbate species. Moreover, 82% of the protein was retained with a similar specific activity as the crude extract which ensures the integrity of the catalytic triad [55]. The experiments were performed in the ATR setup both in static and dynamic fashions as mentioned before.

44

1000

S. Collins and L. Briand

0.02 Q(C=O) Ketoprofen

1710 1664

0.01

Q(C=O) Acyl

1736 1757

Q(COC) Ketoprofen

1196 1028

0.00

–0.01 1757 –0.02 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 Wavenumber (cm–1) 1800 1780 1760 1740 1720 1700 1680 1660 1640 1620 1600 Wavenumber (cm–1)

Fig. 44.10 MES ATR-FTIR spectra in the phase domain of the interaction of rac-ketoprofen with the lipase B of Candida antarctica. Halfcycle spectra (from 180 to 360 ) is presented in the figure to facilitate the observation of the spectral changes. Reaction conditions: racketoprofen (0.16 M in CCl4) to carbon tetrachloride, 0.5 mL/min, exchange frequency 33 mHz, 318 K, 1 spectrum/3 s

Interestingly, the time-resolved IR was strongly influenced by the speciation of ketoprofen due to the increased concentration of the profen in contact with the enzyme during the 68 min experiment. In fact, the signals belonging to monomeric, linear, and cyclic dimeric forms of rac-ketoprofen (νC¼O at 1792 cm
1, 1772 cm
1, and 1742 cm
1) were detected [30, 56]. There is no doubt that an improved methodology that avoids the modification of the substrate concentration during the experiment was a key in this research. In this context, a modulation experiment was performed by varying the feeding composition from carbon tetrachloride rac-ketoprofen (0.16 M in CCl4) using the desired modulation frequency. Spectra were acquired every 3 s, during each c-MES period, using reactants exchange frequencies from 1.7–33 mHz. The spectra in the phase domain and a closer look of the signals in the range 1800–1600 cm
1 are presented in Figs. 44.10 and 44.11. The IR signals belonging to the enzyme film were eliminated from the spectra and the periodic change in the signals at 1710 and 1664 cm
1 due to dissolved ketoprofen is observed. Two new bands are seen at 1757 and 1736 cm
1. These signals change synchronically, but with a phase lag, with the ketoprofen flow into the cell, and therefore can be ascribed as product of the interaction with the lipase. This observation correlates with theoretical DFT calculations of the potential infrared signals, from the various interactions between the enzyme and the substrate, which were published previously by the authors [30]. Those theoretical calculations allow to correlate the presence of an infrared signal at 1757 cm
1 to the acyl-enzyme complex formation with a simultaneous release of a water molecule

Fig. 44.11 Spectral deconvolution of the infrared signals in the 1800–1600 cm
1 range

(formed between the hydroxyls of the profen and serine) that is interacting with the complex through hydrogen bonds [30]. This interaction resembles an ester between the profen and the amino acid serine of the catalytic triad. Additionally, ketoprofen might interact through three H-bonding (one at the carbonyl group and two in the OH group) with the hydroxyl of serine (without reaction) and other adjacent NH and OH species of the catalytic active site. DFT investigation of this intermediate provided a new signal at 1736 cm
1 that has also been detected in the c-MES experiments. Moreover, the experiments show new signals in the range 1300 cm
1 and 1000 cm
1 (see Fig. 44.10). According to the DFT calculations, it is possible to ascribe the bands at 1028 cm
1 and 1196 cm
1 to the stretching vibrations of the C-O-C bond [30]. Additionally, the amide I signal of the lipase (after the experiments described above) was analyzed in order to obtain the contribution of each component of the secondary structure and obtain insights on the effect of the substrate on the protein structure. The results indicate the decrease observed in β-sheets and the increase in β-turns by the interaction of the acyl-enzyme complex. Toledo et al. reported similar finding in the esterification of rac-ketoprofen with short-chain alcohols catalyzed with the immobilized lipase B of Candida antarctica [23, 57].

44.4

Conclusions and Future Perspectives

The present case study is the first report about the investigation of the effect of the nature of the solvent on the enzymatic secondary structure and the experimental identification of the

44

Case Study 2: Modulation Excitation Spectroscopy (MES)

species formed between rac-ketoprofen and the catalytic triad of an α/β-hydrolase through in situ MES ATR-FTIR techniques. Those tools provided key information of the various types of interactions of monomeric ketoprofen with the serine and residues surrounding the catalytic triad of a lipase that is the genesis of the active acyl-enzyme intermediate. There are undoubtful experimental and theoretical evidences that the acyl-enzyme species vibrates at 1756 cm
1. It is worth noticing that modulated ATR-FTIR spectroscopy provided the right experimental conditions to maintain a constant concentration of the monomeric form of the pharmaceutical avoiding the formation of nonreactive polymeric species. Additionally, the effect of hydrophobic and hydrophilic liquid environments on the conformation of the protein was assessed contacting a lipase with solvents of very different log P in a modulated fashion. Preliminary results evidence the decoupling of the various components of the amide I signal due to the modification of the hydrogen bridge bonds and van der Waals interactions in between the enzymatic residues. This fundamental knowledge at a molecular level is a key step to understand the interfacial activation of lipases. The acyl-enzyme species is the key intermediate in the kinetic resolution of profens as commented before. As a second step within the mechanism, the alcohol would interact with the acyl-enzyme complex in the esterification process. According to theoretical studies, the interaction of the alcohol with the acyl-enzyme provides a tetrahedral intermediate that conducts toward the ester. Experimental evidences of the existence of such an intermediate would unleash a longlasting debate about the influence of the structure of the alcohols (short vs long chain; primary vs secondary alcohols) in the catalytic performance of enzymes. In this context, the MES PSD technique is being applied in the investigation of the intermediate species, the secondary structure of lipases, and the progress of the esterification of racemic profens with glycerol to generate valuable prodrugs. It is expected that the MES-PSD technique presented here will allow a deeper description of biocatalyzed processes and structure-activity relationships in order to design optimized biocatalytic materials. Acknowledgments The authors acknowledge the financial support of Consejo Nacional de Investigaciones Científicas y Técnicas CONICET (projects PIP 11220130100171 and 11220130100086) and Universidad Nacional de La Plata (project 11-X898).

References 1. Faber, K.: Biotransformations in Organic Chemistry. Springer International Publishing (2018) 2. Bañares, M.A., Guerrero-Perez, M.O., Fierro, J.L.G., Cortez, G.G.: Raman spectroscopy during catalytic operations with on-line

1001 activity measurement (operando spectroscopy): a method for understanding the active centres of cations supported on porous materials. J. Mater. Chem. 12, 3337–3342 (2002) 3. Meunier, F.C.: The design and testing of kinetically-appropriate operando spectroscopic cells for investigating heterogeneous catalytic reactions. Chem. Soc. Rev. 39, 4602–4614 (2010) 4. Bolivar, J.M., Eisl, I., Nidetzky, B.: Advanced characterization of immobilized enzymes as heterogeneous biocatalysts. Catal. Today. 259, 66–80 (2016) 5. Jeremiah, R., Nuguid, G., Ferri, D., Kröcher, O.: Design of a reactor cell for modulated excitation raman and diffuse reflectance studies of selective catalytic reduction catalysts. Emission Control Sci. Tech. 5, 307–316 (2019) 6. Aguirre, A., Collins, S.E.: Design of an optimized DRIFT cell/ microreactor for spectrokinetic investigations of surface reaction mechanisms. 481, 100628. https://doi.org/10.1016/j.mcat.2018. 07.003 7. Aguirre, A., Kler, P.A., Berli, C.L.A., Collins, S.E.: Design and operational limits of an ATR-FTIR spectroscopic microreactor for investigating reactions at liquid–solid interface. Chem. Eng. J. 243, 197–206 (2014) 8. Aguirre, A., Berli, C.L.A., Collins, S.E.: ATR-FTIR spectrokinetic analysis of the CO adsorption and oxidation at water/platinum interface. Catal. Today. 283, 127–133 (2017) 9. Baurecht, D., Fringeli, U.P.: Quantitative modulated excitation Fourier transform infrared spectroscopy. Rev. Sci. Instrum. 72, 3782–3792 (2001) 10. Baurecht, D., Porth, I., Fringeli, U.P.: A new method of phase sensitive detection in modulation spectroscopy applied to temperature induced folding and unfolding of RNase A. Vib. Spectrosc. 30, 85–92 (2002) 11. Urakawa, A., Bürgi, T., Baiker, A.: Sensitivity enhancement and dynamic behavior analysis by modulation excitation spectroscopy: principle and application in heterogeneous catalysis. Chem. Eng. Sci. 63, 4902–4909 (2008) 12. Barbosa, O., Ariza, C., Ortiz, C., Torres, R.: Kinetic resolution of (R/ S)-propranolol (1-isopropylamino-3-(1-naphtoxy)-2-propanolol) catalyzed by immobilized preparations of Candida antarctica lipase B (CAL-B). New Biotechnol. 27, 844–850 (2010) 13. Kapoor, M., Grupta, M.N.: Lipase promiscuity and its biochemical applications. Process Biochem. 47, 555–569 (2012) 14. Durand, J., Lecomte, J., Baréa, B., Piombo, G., Dubreucq, E., Villeneuve, P.: Evaluation of deep eutectic solvents as new media for Candida antarctica B lipase catalyzed reaction. Process Biochem. 47, 2081–2089 (2012) 15. Romero, M.D., Calvo, L., Alba, C., Daneshfar, A., Ghaziaskar, H.S.: Enzymatic synthesis of isoamyl acetate with immobilized Candida antarctica lipase in n-hexane. Enzym. Microb. Technol. 37, 42–48 (2005) 16. Wu, A.C., Wang, P.Y., Ling, Y.S., Kao, M.F., Chen, J.R., Ciou, J.F., Tsai, S.W.: Improvements of enzyme activity and enantioselectivity in lipase-catalyzed alcoholysis of (R,S)-azolides. J. Mol. Catal. B Enzym. 62, 235–241 (2010) 17. José, C., Toledo, M.V., Briand, L.E.: Enzymatic kinetic resolution of racemic ibuprofen: past, present and future. Crit. Rev. Biotechnol. 36, 891–903 (2016) 18. Toledo, M.V., Briand, L.E.: Relevance and bio-catalytic strategies for the kinetic resolution of ketoprofen towards dexketoprofen. Crit. Rev. Biotechnol. 38, 778–800 (2018) 19. Morrone, R., Antona, N.D., Lambusta, D., Nicolosi, G., Biocatalyzed irreversible esterification in the preparation of S-naproxen: J. Mol. Catal. B Enzym. 65, 49–51 (2010) 20. Kim, M.G., Lee, E.G., Chung, B.H.: Improved enantioselectivity of Candida rugosa lipase towards ketoprofen ethyl ester by a simple two-step treatment. Process Biochem. 35, 977–982 (2000)

44

1002 21. Ciou, J.F., Wang, P.Y., Wu, A.C., Tsai, S.W.: Lipase-catalyzed alcoholytic resolution of (R,S)-flurbiprofenyl azolides for preparation of (R)-NO-flurbiprofen ester prodrugs. Process Biochem. 46, 960–965 (2011) 22. Sontakke, J.B., Yadav, G.D.: Kinetic modeling and statical optimization of lipase catalyzed enantioselective resolution of (R,S)-2pentanol. Ind. Eng. Chem. Res. 50, 12975–12983 (2011) 23. Toledo, M.V., José, C., Collins, S.E., Ferreira, M.L., Briand, L.E.: Towards a green enantiomeric esterification of R/S-ketoprofen: a theoretical and experimental investigation. J. Mol. Catal. Enzym: B. 118, 52–61 (2015) 24. Ortiz, C., Ferreira, M.L., Barbosa, O., dos Santos, J.C.S., Rodrigues, R.C., Berenguer-Murcia, A., Briand, L.E., Fernández-Lafuente, R.: Novozym 435: the “perfect” lipase immobilized biocatalyst? Cat. Sci. Technol. 9, 2380–2420 (2019) 25. Uppenberg, J., Hansen, M.T., Patkar, S., Jones, T.A.: The sequence, crystal structure determination and refinement of two crystal forms of lipase B from Candida antarctica. Structure. 2, 293–308 (1994) 26. Bornscheuer, U.T., Kazlauskas, R.J.: Hydrolases in Organic Synthesis: Regio- and Stereoselective Biotransformations, 2nd edn. WileyVCH, Germany (2006) 27. Martinelle, M., Hult, K.: Kinetics of acyl transfer reactions in organic media catalyzed by Candida Antarctica lipase B. Biochim. Biophys. Acta. 1251, 191–197 (1995) 28. Paiva, A.L., Balcão, V.M., Malcata, F.X.: Kinetics and mechanism of reactions catalyzed by immobilized lipases. Enzym. Microb. Technol. 27, 187–204 (2000) 29. Arroyo, M.: Síntesis enantioselectivas catalizadas por lipasas microbianas. An. R. Soc. Esp. Quim. 1, 19–24 (2000) 30. Toledo, M.V., Llerena Suster, C.R., Ferreira, M.L., Collins, S.E., Briand, L.E.: Molecular recognition of an acyl-enzyme intermediate on the lipase B from Candida antarctica. Cat. Sci. Technol. 7, 1953–1964 (2017) 31. Swedberg, S.A., Pesek, J.J., Fink, A.L.: Attenuated total reflectance Fourier transform infrared analysis of an acyl-enzyme intermediate of α–chymotrypsin. Anal. Biochem. 186, 153–158 (1990) 32. Wilkinson, A.-S., Ward, S., Kania, M., Page, M.G.P., Wharton, C.W.: Multiple conformations of the acylenzyme formed in the hydrolysis of methicillin by Citrobacter freundii β-lactamase: a time-resolved FTIR spectroscopic study. Biochemistry. 38, 3851–3856 (1999) 33. Hokenson, M.J., Cope, G.A., Lewis, E.R., Oberg, K.A., Fink, A.L.: Enzyme-induced strain/distortion in the ground-state ES complex in β-lactamase catalysis revealed by FTIR. Biochemistry. 39, 6538–6545 (2000) 34. Wilkinson, A.-S., Bryant, P.K., Meroueh, S.O., Page, M.G.P., Mobashery, S., Wharton, C.W.: A dynamic structure for the acylenzyme species of the antibiotic aztreonam with the Citrobacter freundii β-lactamase revealed by iInfrared spectroscopy and molecular dynamics simulations. Biochemistry. 42, 1950–1957 (2003) 35. Iqbal, A., Clifton, I.J., Bagonis, M., Kershaw, N.J., Domene, C., Claridge, T.D.W., Wharton, C.W., Schofield, C.J.: Anatomy of a simple acyl intermediate in enzyme catalysis: combined biophysical and modeling studies on ornithine acetyl transferase. J. Am. Chem. Soc. 131, 749–757 (2009) 36. Tang, A.W., Kong, X., Terskikh, V., Wu, G.: Solid-state 17O NMR of unstable acyl-enzyme intermediates: a direct probe of hydrogen bonding interactions in the oxyanion hole of serine proteases. J. Phys. Chem. B. 120, 11142–11150 (2016) 37. Huguenin-Dezot, N., Alonzo, D.A., Heberlig, G.W., Mahesh, M., Nguyen, D.P., Dornan, M.H., Boddy, C.N., Schmeing, T.M., Chin, J.W.: Trapping biosynthetic acyl-enzyme intermediates with encoded 2,3-diaminopropionic acid. Nature. 565, 112–118 (2019)

S. Collins and L. Briand 38. Bürgi, T., Baiker, A.: Attenuated total reflection infrared spectroscopy of solid catalysts functioning in the presence of liquid-phase reactants. Adv. Catal. 50, 227–283 (2006) 39. Tauler, V.R.: Multivariate curve resolution applied to second order data. Chemom. Intel. Lab. Syst. 30, 133–146 (1995) 40. Ruckebusch, C., Blanchet, L.: Multivariate curve resolution: a review of advanced and tailored applications and challenges. Anal. Chim. Acta. 765, 28–36 (2013) 41. Alcaraz, M.R., Aguirre, A., Goicoechea, H.C., Culzoni, M.J., Collins, S.E.: Resolution of intermediate surface species by combining modulated infrared spectroscopy and chemometrics. Anal. Chim. Acta. 1049, 38–46 (2019) 42. Kumar, A., Dhar, K., Kanwar, S.S., Arora, P.K.: Lipase catalysis in organic solvents: advantages and applications. Biol. Proceed. Online. 18, 1–11 (2016) 43. Adlercreutz, P.: Immobilisation and application of lipases in organic media. Chem. Soc. Rev. 42, 6406–6436 (2013) 44. Léonard, V., Fransson, L., Lamare, S., Hult, K., Graber, M.: A water molecule in the stereospecificity pocket of Candida antarctica lipase B enhances the enantioselectivity towards 2-pentanol. Chembiochem. 8, 662–667 (2007) 45. Wedberg, R., Abildskov, J., Peters, G.H.: Protein dynamics in organic media at varying water activity studied by molecular dynamics simulation. J. Phys. Chem. B. 116, 2575–2585 (2012) 46. Yang, L., Dordick, J.S., Garde, S.: Hydration of enzyme in nonaqueous media is consistent with solvent dependence of its activity. Biophys. J. 87, 812–821 (2004) 47. Dong, A., Meyer, J.D., Kendrick, B.S., Manning, M.C., Carpenter, J.F.: Effect of secondary structure on the activity of enzymes suspended in organic solvents. Arch. Biochem. Biophys. 334, 406–414 (1996) 48. Barth, A.: Infrared spectroscopy of proteins. Biochim. Biophys. Acta. 1767, 1073–1101 (2007) 49. Kong, J., Yu, S.: Fourier transform infrared spectroscopic analysis of protein secondary structures. Biochim. Biophys. Acta. 39, 549–559 (2007) 50. Barth, A., Zscherp, C.: What vibrations tell us about proteins. Q Rev Biophys. 35, 369–430 (2002) 51. Heinz, F., Werner, M.: Infrared spectroscopy of proteins. In: Chalmers, J.M., Griffiths, P.R. (eds.) Handbook of Vibrational Spectroscopy, vol. 5, pp. 3749–3775. Wiley, Chichester (2002) 52. Llerena-Suster, C.R., José, C., Collins, S.E., Briand, L.E., Morcelle, S.R.: Investigation of the structure and proteolitic activity of papain in aqueous miscible organic media. Process Biochem. 47, 47–67 (2012) 53. Khan, F.I., Lang, D., Durrani, R., Huan, W., Zhao, Z., Wang, Y.: The lid domain in lipases: structural and functional determinant of enzymatic properties. Front. Bioeng. Biotechnol. 5(16), 1–13 (2017) 54. Foresti, M.L., Galle, M., Ferreira, M.L., Briand, L.E.: Enantioselective esterification of ibuprofen with ethanol as reactant and solvent catalyzed by immobilized lipase: experimental and molecular modeling aspects. J. Chem. Technol. Biotechnol. 84, 1461–1473 (2009) 55. Llerena Suster, C.R., Briand, L.E., Morcelle, S.R.: Analytical characterization and purification of a commercial extract of enzymes: a case study. Colloids Surf. B Biointerfaces. 121, 11–20 (2014) 56. Champeau, M., Thomassin, J.-M., Jérôme, C., Tassaing, T.: Solubility and speciation of ketoprofen and aspirin in supercritical CO2 by infrared spectroscopy. J. Chem. Eng. Data. 61, 968–978 (2016) 57. Toledo, M.V., José, C., Collins, S.E., Bonetto, R.D., Ferreira, M.L., Briand, L.E.: Esterification of R/S-ketoprofen with 2-propanol as reactant and solvent catalyzed by Novozym ® 435 at selected conditions. J. Mol. Catal. Enzym: B. 83, 108–119 (2012)

44

Case Study 2: Modulation Excitation Spectroscopy (MES)

Sebastian E. Collins received his PhD in chemistry from Universidad Nacional del Litoral (Santa Fe, Argentina) in 2005. He is currently a full professor and a researcher of the National Scientific and Technical Research Council (CONICET). He works in the development of new spectroscopic tools for in situ and operando investigations of (bio) catalytic reactions.

1003

Laura E. Briand received her PhD from Universidad Nacional de La Plata (Buenos Aires, Argentina) in 1993. She is a scientist of the National Scientific and Technical Research Council CONICET at the Center of Research and Development in Applied Sciences in La Plata, Argentina. She is currently involved in the application of molecular spectroscopy in biocatalysis and biotransformation processes involved in the kinetic resolution of pharmaceuticals.

44

Temperature-Programmed (TP) Techniques Jih-Mirn Jehng

, Israel E. Wachs

45

, and Michael Ford

Contents 45.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1006

45.2

Description of Temperature-Programmed Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benefits from Characterization of Catalysts . . . . . . . . . . . . Limitations of Temperature-Programmed Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of Method to Other Techniques . . . . . . . . . . .

1007 1007

Description of Temperature-Programmed Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Single Crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Powders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1008 1008 1008 1009

45.2.1 45.2.2 45.2.3 45.2.4 45.3 45.3.1 45.3.2 45.3.3 45.4

1006 1006 1007

45.4.4 45.4.5 45.4.6

Applications to Catalyst Structure-Activity Relationships: Methods and Case Studies . . . . . . . . . . . Thermogravimetric Analysis (TGA) and Differential Thermogravimetric Analysis (DTG) . . . . . . . Temperature-Programmed Decomposition . . . . . . . . . . . . . . Temperature-Programmed Oxidation (TPO) and Differential TPO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Temperature-Programmed Reduction (TPR) . . . . . . . . . . . . Temperature-Programmed Desorption (TPD) . . . . . . . . . . . Temperature-Programmed Surface Reaction (TPSR) . . .

45.5

Summary/Conclusion/Future Outlook . . . . . . . . . . . . . . . 1027

45.4.1 45.4.2 45.4.3

1009 1009 1009 1012 1012 1015 1016

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027

Abstract

Temperature-programmed techniques are extremely informative and provide fundamental surface information J.-M. Jehng (*) Department of Chemical Engineering, National Chung Hsing University, Taichung, Taiwan, Republic of China e-mail: [email protected] I. E. Wachs · M. Ford Operando Molecular Spectroscopy & Catalysis Laboratory, Department of Chemical & Biomolecular Engineering, Lehigh University, Bethlehem, PA, USA e-mail: [email protected]; [email protected]

about the number of surface sites, chemical nature of the surface sites, catalytic reaction mechanisms, and surface kinetics for the rate-determining-steps for all types of solid catalysts (bulk oxides/metals, mixed oxides/metal alloys, supported metals/metal oxides, and zeolites/molecular sieves). The ability to address such a wide range of problems for all types of solid catalysts makes temperatureprogrammed techniques among the most versatile methods in catalysis research. In this chapter, the temperature-programmed techniques will be systematically reviewed by the fundamental methods, instruments, and applications. In Sect. 45.2, the description of the temperature-programmed methods is divided into theory, benefits, limitations, and comparison with other techniques. The temperature-programmed instruments, under ultrahigh vacuum and gas flow conditions, and its historical development are described in Sect. 45.3. In Sect. 45.4, the relationship of the structural information and reactivity for different catalyst cases has been established by applying various temperature-programmed methods such as Thermogravimetric Analysis-Differential Thermogravimetric Analysis (TGA-DTG), Temperature Programmed Decomposition (TPD), Temperature Programmed Oxidation/Reduction (TPO/TPR), Temperature Programmed Desorption (TPD), and Temperature Programmed Surface Reactions (TPSR). The advantage of temperature-programmed techniques is that they can distinguish between multiple kinetic processes occurring during temperature programming. This kinetic feature, and its quantitative capability, makes temperature-programmed techniques a very powerful characterization method. Keywords

Thermogravimetric analysis-differential thermogravimetric analysis (TGA-DTG) · Temperatureprogrammed decomposition (TPD) · Temperatureprogrammed oxidation/reduction (TPO/TPR) ·

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_45

1005

1006

J.-M. Jehng et al.

Temperature-programmed desorption (TPD) · Temperature-programmed surface reactions (TPSR) · CO-TPR · NH3-TPD · CH3OH-TPSR · Water-gas shift (WGS) reaction · Selective catalytic reduction (SCR) reaction

45.1

Introduction

Temperature-programmed techniques are a group of methods that employ the stimulus of rapidly increasing temperature that yields responses (detected as changes in heat (uptake/loss), weight (uptake/loss), or evolution of gases) providing information about a material’s physical/chemical properties and/or its catalytic reaction products [1, 2]. The thermal analytical techniques can be classified as follows according to the properties of the material the experiment provides: (1) energy changes (Differential Thermal Analysis (DTA) and Differential Scanning Calorimetry (DSC)), (2) weight changes (Thermogravimetric Analysis (TGA) and Differential Thermogravimetric Analysis (DTG)), (3) evolved volatiles (Temperature Programmed Decomposition (TPD)), and (4) gases evolved or consumed from surface and bulk reactions (Temperature Programmed Reduction (TPR), Temperature Programmed Oxidation (TPO), Temperature Programmed Desorption (TPD), and Temperature Programmed Surface Reaction (TPSR)). Temperature-programmed techniques are the most versatile characterization methods because they can be applied to all types of solid catalytic materials and environmental conditions (e.g., vacuum or in presence of gases).

45.2

Description of TemperatureProgrammed Methods

45.2.1 Theory The temperature-programming methods apply a rapid linear increase in temperature (typically ~1–30 K/s or ~1–20 K/min to conducting and poorly conducting solid materials, respectively). The type of detector or sensor determines the nature of the information collected. The heat uptake or loss (e.g., phase transition, exothermic event, or endothermic event) taking place during temperature programming is quantitatively monitored with a calorimeter. The weight changes (e.g., material decomposition (TPD), reduction (TPR), or oxidation (TPO)) occurring during temperature programming are quantitatively monitored with a gravimetric balance. The weight changes corresponding to loss or evolution of a gaseous component can also be detected with an online thermal conductivity detector (TCD) or mass spectrometer (MS) that monitors the consumption or evolution of gases. When

multiple products are evolved into the gas phase during temperature-programmed decomposition/desorption (TPD) and temperature-programmed surface reaction (TPSR) studies, it is necessary to use an online MS to distinguish among the different molecules being evolved into the gas phase. A special feature of temperature-programmed techniques is that they contain kinetic information that corresponds to the first derivative of the change in heat, weight, or gaseous evolution with respect to time or temperature. Rate of change ¼ dðheat or weightÞ=dT

ð45:1aÞ

Rate of change ¼ dðheat or weightÞ=dt:

ð45:1bÞ

Monitoring the evolved gases also corresponds to the rates of evolution since Rate of evolution ¼ dðweightÞ=dT

ð45:1cÞ

Rate of evolution ¼ dðweightÞ=dt:

ð45:1dÞ

Consequently, the gas consumption or evolution already provides the first derivative, which is the rate of the event. A significant advantage of temperature-programmed techniques is that they can distinguish between multiple kinetic processes occurring during temperature programming (e.g., multiple reaction steps that may produce the same gas or different gases). This kinetic feature, and its quantitative capability, makes temperature-programmed techniques a very powerful characterization method. Kinetic parameters such as Eact and pre-exponential factors (v) can also be estimated from temperature-programmed techniques by application of the Redhead Equation [3]. For events that follow first-order kinetics, the Tp (maximum peak temperature in the temperature programmed measurement), heating rate, Eact, and v are related by Eq. 45.2. Given that the equation is independent of Ci (initial surface concentration), the equation can be applied to all kinetic processes. The Eact value can be calculated from knowledge of the experimentally determined Tp and heating rate (β (K/s)) with the assumption that v(1) has a theoretical value of 1013/s.

Redhead Equation for first  order process :

Eact ¼ RT 2p

vð1Þ  : βe

Eact RT p

ð45:2Þ

Once Eact is calculated with the assumption of v(1) ¼ 1013/ s, the first-order rate constant for the kinetics can be calculated from Eq. 45.3. k

ð 1Þ

¼ 10

13

expE ½ act =RT  1=s:

ð45:3Þ

45

Temperature-Programmed (TP) Techniques

1007

For events that follow second-order kinetics, the Tp, heating rate, v, and Ci (initial surface concentration) are related by Eq. 45.4. Second-order kinetics require knowledge of the value of Ci. The Ci value can readily be quantified for TPD/TPSR processes by saturating the surface with a monolayer of the adsorbed probe molecule (e.g., H2, O2, CO, etc.). The Ci value during the TPD/TPSR process is then determined from knowledge of the Cmonolayer value. The Eact value can be calculated from knowledge of the experimentally determined Tp, heating rate, and Ci with the assumption that v(2) has a theoretical value of 102 cm2/s.

Redhead Equation for second  order :

Eact vð2Þ C ¼  i : Eact RT 2p βe RT p

ð45:4Þ

45

Once Eact is calculated with the assumption of v ¼ 102 cm2/s, the second-order rate constant for the kinetics can be calculated from Eq. 45.5. (2)

k

ð2Þ

¼ 10

2

2

expE ½ act =RT  cm =s:

calibrated detectors (TCD, MS, etc.) allow for determining the surface coverage of adsorbates and concentration of species involved in multiple kinetic events (e.g., distinguishing between similar components with different binding energies: desorption or reduction of oxygen on the surface and in the bulk of a catalytic material, reduction of surface metal oxide phases vs. crystalline metal oxide nanoparticles, etc.). The kinetic aspect of temperature-programmed techniques allows for identifying the rate-determining step of many catalytic reactions. The ability to apply temperature-programmed techniques to all types of catalysts, under reactive environments, and also provide kinetic information makes temperatureprogrammed techniques highly informative characterization methods for catalysis research.

ð45:5Þ

Additionally, the rate constants are usually related to the rate-determining steps of catalytic reactions. More detailed descriptions of surface kinetics equations are given elsewhere [4].

45.2.2 Benefits from Characterization of Catalysts A major benefit of the temperature programmed techniques is the simplicity and relatively low cost of the required experimental instrumentation. For TPO and TPR, it is just necessary to have an inexpensive TCD detector usually present on gas chromatographs and GCs are also equipped with temperature programming capabilities. For TPD and TPSR experiments involving multiple gases, an online MS is required. Differential thermal analysis and DSC require a calorimeter to detect heat loss (exotherms) and heat gains (endotherms). A sensitive balance for monitoring weight gains and losses is required for TGA and DTA. The costs for an MS, calorimeter, and mass balance are not as exorbitant compared to the sophisticated modern spectroscopic instrumentation. Combining temperature-programmed techniques with spectroscopic methods (e.g., Raman, IR, XAS, etc.), however, is extremely powerful since such advanced instruments can simultaneously provide fundamental catalyst structure-activity/selectivity relationships. The quantitative aspect of TGA (mass loss and gain and the rates of mass loss and gain) and

45.2.3 Limitations of Temperature-Programmed Techniques The temperature-programmed techniques may be challenged by catalytic materials in powder form with extremely low surface area (1 m2/g) that would limit the amount of gas that can adsorb on the catalyst for TPD and TPSR studies. One way to get around the low surface area issue with catalysts of low surface area is to flow one or all the reactants over the catalyst during TPD and TPSR studies. An exception to the low surface area issue, however, is found for single crystal studies under ultrahigh vacuum (UHV) conditions with highly sensitive MS that are able to successfully perform TPD and TPSR measurements. This is achieved by placing the surface of the single crystal next to the MS. The temperature-programmed techniques are limited to gas-solid studies and cannot be applied to liquid-solid studies.

45.2.4 Comparison of Method to Other Techniques Conventional temperature-programmed studies do not provide structural information about the catalyst bulk and surface phases containing the catalytic active sites. With the aid of isotopic labeling (e.g., D, 18O, and 13C) and an online MS, however, the surface intermediates, rate-determining steps, and reaction pathways can also be established. Structural information about the catalyst bulk and surface phases can be achieved by combining temperature-programmed techniques with structural characterization methods such as IR, Infrared Absorption Spectroscopy (IRAS), Raman, X-ray Photoelectron Spectroscopy (XPS), X-ray Absorption Spectroscopy (XANES and EXAFS), and Secondary Ion Mass

1008

J.-M. Jehng et al.

Spectroscopy (SIMS). Such studies are starting to appear in the literature as TP-IRAS [5–8], TP-Raman [8–11], TP-XPS [12], TP-XANES [13], and TP-SIMS [14].

Sample manipulator Electron gun

45.3

Description of TemperatureProgrammed Instruments

Gas doser UHV

45.3.1 Single Crystals Single crystals consist of well-defined, single-phase crystallographic materials that terminate with specific surface faces (e.g., 111, 110, 100, etc.). Most commonly, metallic single crystals are employed for TPD and TPSR studies because of their excellent thermal conductivity that allows for rapid sample heating (1–30 K/s). The single crystals are usually heated from the rear with the adsorbed gases desorbing from the front, which is directed at the MS as shown in Fig. 45.1. The UHV conditions allow for the desorbed gases to be directly detected by the MS without any complications such as gas phase collisions or reactions. This chapter, however, will only focus on powdered materials since such materials find the widest use as catalysts and sorbents.

Sample handling chamber Sample

Vacuum pump

Fig. 45.1 Schematic of model single crystal catalyst in UHV chamber

Microbalance

45.3.2 Powders Powders, both as crystalline and amorphous phases, can be investigated with temperature-programmed techniques under all environmental conditions (UHV, pressure, and temperature). Most studies with powders, however, are performed under pressure and temperature with flowing gases. For Thermogravimetric Analysis (TGA) and Differential Thermogravimetric Analysis (DTG), experiments are performed by placing the catalyst sample in a cup that is connected to a sensitive balance, with the gases flowing over the cup. This arrangement, as shown in Fig. 45.2, minimizes the gas solid interaction because the gases do not flow through the catalyst powder as in plug flow fixed-bed reactors. Nevertheless, information about the change in mass (weight) and rate of mass change (weight) is provided. Processes that result in mass increases are adsorption and TPO, while processes that lead to mass decreases are TPR, TP Decomposition, TP Desorption, and TPSR. It is quite impressive that the modern, highly sensitive mass balances are able to detect adsorption and desorption events of submonolayer amounts of molecules on the surface of catalytic materials. Commercial temperatureprogrammed systems are also available that combine mass changes (TGA) and calorimetric changes (Differential Thermal Analysis (DTA) and Differential Scanning Calorimetry

Mass spectrometer

Reference basket

Furnace with temperature controller Sample basket

Gas flow maniflod

Fig. 45.2 Schematic of Thermogravimetric Analysis (TGA) and Differential Thermogravimetric Analysis (DTG) system with powder catalysts

(DSC)). DTA measures the temperature difference between the sample and the reference upon the thermal processing, and DSC measures the heat flow between the sample and reference in order to remain at the same the temperature upon the thermal processing. DTA and DSC monitor thermal events primarily from phase changes, decomposition of bulk phases, and adsorption/desorption from surfaces.

Temperature-Programmed (TP) Techniques

1009

45.3.3 History Temperature-programmed techniques first appeared with the publication that introduced the concept of TGA ~1900 [15]. TGA was the only temperature-programmed technique until the introduction of differential microcalorimeters (DTA and DSC) in the 1960s [16]. With the introduction of small online mass spectrometers in the 1960s, it also became possible to extend the investigation to the kinetics of adsorption/desorption rates as well as surface chemical reactions [4, 17]. The introduction of Temperature-Programmed Reduction (TPR) to the study of catalysts was developed by J.W. Jenkins in 1975 [18, 19]. Most recently, the addition of multiple probes (e.g., IR, IRAS, Raman, XPS, XANES, and SIMS) have allowed the simultaneous monitoring of surface sites, surface intermediates, and desorbing gas phase molecules from the catalyst surface as well as the kinetics of these processes during temperature programming with online fast detectors (e.g., MS, IR, etc.) [5–14].

Applications to Catalyst StructureActivity Relationships: Methods and Case Studies

45.4.1 Thermogravimetric Analysis (TGA) and Differential Thermogravimetric Analysis (DTG)

Rate of evolution ¼ dðweightÞ=dT

ð45:1aÞ

Rate of evolution ¼ dðweightÞ=dt

ð45:1bÞ

The TGA-DTG results from Graphene Oxide (GO) are shown in Fig. 45.3. The experiment was performed with a PIKE ATR-600 instrument. A thermal analysis initiated the heating program from 100 to 800  C with a ramping rate of 10  C/min under flowing 50 cc/min of air. From GO-DTG curve, 3 stages of weight loss are observed in the thermal diagram. The first stage in the 150–350  C region is due to the loss of hydroxyl, carbonyl, carboxylic, and epoxide functional groups, and remaining water molecules on the GO surface. Then, the second stage from 450 and 550  C involves the pyrolysis of the amorphous carbon species. Finally, a weight loss appearing from 550 to 700  C is caused by the decomposition of the carbon skeleton [24].

45.4.2 Temperature-Programmed Decomposition

Thermal gravimetric experiments monitor the weight changes of a material as a function of temperature programming. Materials that undergo a bulk phase transformation (e.g., TiO2 (anatase) ➔ TiO2 (rutile)) do not undergo weight changes but do involve energy changes that can be detected with DSC. Weight changes do take place during material bulk decomposition processes (Ag2CO3 ➔ 2Ag þ CO2 þ 1/2O2 [20], 4CH3COOAg þ 7O2 ➔ 4Ag þ 8CO þ 6H2O [21], NH4VO3 ➔ V2O5 [22], α-PtO2 ➔ Pt þ O2 [23], etc.) in inert or vacuum environments and are easily detected with the highly sensitive modern TGA instruments. Differential Thermogravimetric Analysis (DTG) is the first derivative of the weight change with temperature (Eq. 45.1a) or time (Eq. 45.1b), respectively, Rate of change ¼ dðweightÞ=dT

ð45:1aÞ

Rate of change ¼ dðweightÞ=dt

ð45:1bÞ

that represent the rate of the weight loss. One disadvantage of TGA/DTG measurements is that the powder is placed in a cup, and when flowing gases are employed the gases do not flow through the bed of powder. It is, thus, more common

TGA-DTA (or DSC) is a synchronous thermal analyzer that can characterize multiple thermal properties of samples in a single experiment. The TGA measures the weight loss or gain of the sample related to decomposition, oxidation, and any other physical or chemical changes. The corresponding DTA indicates whether the thermal process is endothermic or

100 80

0.0 203

–0.5

530

60 590 40

–1.0

Graphene oxide First derivative

–1.5

20 0 100

200

300

400 500 600 Temperature (°C)

700

Derivative of weight loss (%)

45.4

with flowing gases to monitor the amount and rate of gas evolution during temperature programming with an online thermal conductivity detector (TCD) or mass spectrometer (MS). Monitoring the evolved gas also corresponds to the rates of evolution since

Weight loss (%)

45

–2.0 800

Fig. 45.3 TGA-DTG result for graphene oxide (ramp rate of 10  C/min under flowing 50 cc/min of air)

45

1010

J.-M. Jehng et al.

ð45:2Þ

Ag2 O ! 2Ag þ § O2

ð45:3Þ

The decomposition of Ag2CO3 has been investigated with TGA-DTA (or DSC) with a Shimadzu Co. DTG 50M [20]. The TGA-DTA experiments were performed by using ~15 mg of sample placed on a platinum pan (6 mm diameter  3 mm height). The temperature-programmed experiments were performed with variable heating rates β (0.5  β  10 K/min) in flowing N2 (or CO2) at 80 cc/min. The TGA-DTA results from six commercial Ag2CO3 samples grouped into A (Type III), B (Type I), and C (Type II) are shown in Fig. 45.5 from 300 to 750 K [20]. It appears that all the DTA spectra contain an endothermic event at lower temperature (400–500 K) that is characteristic of the thermal decomposition of Ag2CO3 (step (2)) and a second endothermic event at higher temperature (600–700 K) characteristic of the decomposition of Ag2O to metallic Ag (step (3)). The TG-DTA (Rigaku Co., TG8120) was equipped to an online

Microbalance

Sample temperature, TS

Reference temperature, TR

Fig. 45.4 Schematic of Differential Thermal Analysis (DTA) and Differential Scanning Calorimetry (DSC) system with powder catalysts. Differential Thermal Analysis (DTA) measures ΔT ¼ TS  TR as a function of TS during a temperature-programmed process. Differential Scanning Calorimetry (DSC) measures the heat flux (ΔT/t) to compensate the energy loss or gain of the sample as a function of TS during temperature-programmed process

0.1 DTA 0.0

0

TG

–5

–0.1

ΔT (e.m.f.) (mV)

Ag2 CO3 ! Ag2 O þ CO2

Sample chamber

Endo.

Thermal Decomposition of Silver Carbonate (Ag2CO3) Silver carbonate thermally decomposes by the following steps:

Heating element

Relative mass change (Δm/m0) (%)

exothermic where no mass loss occurs for phase transformations such as melting, crystallization, and glass transitions. The combined TGA-DTA experiment monitors the (i) weight change as a function of temperature (and/or time) under a controlled gas environment and temperature, and (ii) temperature difference between the sample temperature (TS) and the reference temperature (TR) as a function of TS within the temperature-programmed region. The simultaneous DSC measures the absorbed or released energy of the sample when a phase transformation, glass transition, or chemical reaction occurs during the thermal process. The DSC arrangement includes two calorimeteric cells, one for the sample and the other for the reference without the sample. When a sample undergoes a phase transformation during the temperature-programmed process, it absorbs or releases heat. The compensator measures how to increase or decrease the heat flow to maintain the temperature of the sample (TS) and the reference (TR). The relationship between the heat flux of the sample and reference is scanned within the thermal process, and the DSC curve is plotted to perform qualitative or quantitative analysis as a function of TS. The TGA-DSC measurement can determine the glass transition temperature, melting point, crystallization temperature, specific heat, thermal stability, oxidation stability, reaction heat, dynamic analysis, etc. The schematic diagram of a TGA-DTA (or DSC) system with powder catalysts is shown in Fig. 45.4.

–0.2

–10 –15 –20 –25 300

A(type III) B(type I) C(type II) 400

500

600

700

T (K)

Fig. 45.5 TG-DTA experiments for samples A, B, and C (mo ¼ 15 mg) at β ¼ 5 K min1 under flowing N2 (80 cc/min). (Reproduced from ref. [20] [Copyright (2013) with the permission from Elsevier])

quadrupole MS (Anelva Co., M-200QA) to monitor the evolved gases during the temperature-programmed studies. The TGA-DTA-MS experiments used ~2.0 mg of sample loaded into a platinum pan (5 mm diameter  2.5 mm height) with a heating rate of 5 K/min and flowing He (200 cc/min), and the mass spectrometer signals of O2 (m/z ¼ 32) and CO2

45

Temperature-Programmed (TP) Techniques

1011

1.5 1.0

A (type III) 1st heating

Endo.

0.0 6.0

Heat flow (mW)

Intensity (a.u.) (10–10 A)

0.5

B (type I)

4.0 2.0 0.0 6.0

2.0 mW

3rd heating

C (type II)

4.0

2nd heating

m/z = 32 m/z = 44

2.0 0.0 300

400

500

600

700

45

T (K) 300

Fig. 45.6 Mass spectra (m/z ¼ 32 and m/z ¼ 44) of the evolved gases during thermal decomposition of samples A, B, and C (mo ¼ 2 mg) at β ¼ 5 K/min under flowing He (200 cc/min). (Reproduced from ref. [20] [Copyright (2013) with the permission from the American Chemical Society])

(m/z ¼ 44) gases were simultaneously monitored. The TGADTA-MS results from the six commercial Ag2(CO3) samples grouped into A (Type III), B (Type I), and C (Type II) are shown in Fig. 45.6 from 300 to 750 K [20]. The results are consistent with the above proposed steps for thermal decomposition of Ag2CO3: Only CO2 was detected at the lower decomposition temperature (400–500 K), and O2 or O2/CO2 was observed at the higher decomposition temperature (600–700 K). The appearance of CO2 for the Type II and III samples at higher decomposition temperature is due to the different synthesis methods and impurities of the starting Ag2(CO3) samples [20]. The DSC (Shimadzu Co., DSC60) experiments were performed with ~10.0 mg of sample sealed in an aluminum pan (5 mm diameter) under flowing CO2 (50 cc/min). The cyclic heating (β ¼ +5 K/min) from 300 K (or above 300 K) to 495 K and cooling (β ¼ 5 K/min) below 495 K of sample A under the CO2 environment were performed in order to determine the phase transformations of Ag2CO3 as a function of temperature, and the DSC results are shown in Fig. 45.7. Under flowing CO2 during the initial heating from 300 to 495 K, two overlapping endothermic DSC peaks are indicating the phase transformation of RT-Ag2CO3 to β-Ag2CO3 at about 460 K and β-Ag2CO3 to α-Ag2CO3 at about 469 K. The α-Ag2CO3 phase transforms back to the β-Ag2CO3 phase upon cooling from 495 to 420 K as reflected with a sharp exothermic peak at about 460 K, and the phase transformation of β-Ag2CO3 to α-Ag2CO3 transition occurs at about 469 K. The β $ α transition cycles were controlled by the heating and cooling thermal processes. The phase

350

400

450

500

T (K)

Fig. 45.7 DSC experiments of sample A (mo ¼ 10.0 mg) during cyclic heating (β ¼ 5 K/min) and cooling (β ¼ 5 K/min) in flowing CO2 (50 cc/min). (Reproduced from ref. [20] [Copyright (2013) with the permission from the American Chemical Society])

transformation of β-Ag2CO3 to RT-Ag2CO3 possesses a broad exothermic DSC peak occurring at ~370 K. On the third heating process, the phase transformation of RT-Ag2CO3 to β-Ag2CO3 starts at about 447 K with a broad endothermic DSC peak which is 13 K lower than that of the first heating process. The temperature difference of the reversible transition between the RT-Ag2CO3 and β-Ag2CO3 phase is obviously larger than that of the reversible transition between β-Ag2CO3 and α-Ag2CO3 phase. The bulk phases of the Ag samples were determined with high-temperature XRD (HT-XRD) experiments with in situ XRD diffractometer (Rigaku Co., RINT 2200 V) equipped with a programmable heating chamber (Rigaku Co., TC20) under flowing 100 cc/min of CO2 at every stage of the thermal processes [20].

Thermal Decomposition of Ammonium Metavanadate (NH4VO3) The decomposition of NH4VO3 was examined with TGA (Du Pont, Model 2200) employing a heating rate of 10  C/ min in flowing N2 (60 cc/min), and the TGA-DTG results are shown in Fig. 45.8 as a function of temperature [25]. Multiple DTG peaks appear at 160, 195, and 260  C indicating the phase transformations of NH4VO3 during the temperatureprogrammed process. Initially, the NH4VO3 phase transformed to the (NH4)2V4O11 phase at 160  C and further decomposed to the NH4V3O8 phase at 195  C, and the phases were determined by also studying the decomposition process

1012

J.-M. Jehng et al.

200 0.5

0.1

80 DTG 70

0

200

400

–0.1

T (°C)

Fig. 45.8 TG-DTG analysis of the thermal decomposition of NH4VO3 under flowing N2. (Reproduced from ref. [25] [Copyright (1990) with the permission from The Royal Society of Chemistry])

with in situ Raman spectroscopy [25]. The phase transformation of NH4V3O8 to V2O5 is completed above 300  C, and the DTG curve shows no further weight loss until the end of the experiment (see Fig. 45.8).

45.4.3 Temperature-Programmed Oxidation (TPO) and Differential TPO TPO measurements are temperature-programmed studies performed under a flowing O2/inert gas environment where a component of the material becomes oxidized (V2O3 þ O2 ➔ V2O5) or is evolved into the gas phase (e.g., C þ O2➔ CO2) and the weight changes are monitored with TGA and DTG. Oxidation studies with thermal gravimetric analysis (TGA) and differential thermal analysis (DTA) are routinely used to characterize the thermal properties of carbonaceous materials such as graphite [26] and carbon nanotubes (CNTs) [27]. The oxidation rate and onset temperature (also called ignition temperature) of carbonaceous materials by O2 are strongly dependent on the active surface area [28–30] and crystallographic structure [31]. The oxidative stability of carbon nanotubes is influenced by defect sites in graphite [30] and nanotube diameters [32]. It appears that various types of carbon such as amorphous carbon, nanocarbon capsule, and CNT possess different oxidation characteristics that allow them to be distinguished with TPO and DTPO [33]. Alternatively, the amount and rate of O2 consumption and gas evolution during the combustion process can also be followed with an online mass spectrometer (MS) because of its ability to discriminate between different gases. A TCD detector may not be ideal for TPO and DTPO measurements during combustion because it does not distinguish between evolution of O2, CO, and CO2.

Without H2 With H2

634 °C

160

30 20

570 °C

140 Weight (%)

0.3

90

1st derivative mass (%)

Mass (%)

100

40 a b

180

10

120

0

100

–10 a b

80 60

Without H2 With H2

–20 –30

40

–40

20

–50

0

0

Deriv. weight (%)

TG

–60 100 200 300 400 500 600 700 800 900 Temperature (°C)

Fig. 45.9 TGA results for oxidation of CNTs grown with CH4 and CH4/H2 addition. (Reproduced from ref. [33] [Copyright (2006) with the permission from Elsevier])

The CNTs were synthesized by two different methods to examine the effect of the presence of hydrogen on CH4 pyrolysis to form CNTs. The following procedures were used. The nano-intermetallic MgNi alloy catalyst was first reduced with flowing H2 gas (50 cm3/min) at 500  C for 1 h, and the temperature was then raised to 670  C for growing CNTs by introducing either flowing CH4 (100 cm3/min) or a mixed CH4/H2 gas (CH4/H2 ¼ 2:1) for 1 h. The TGA (weight of sample – left scale) and DTA (rate of weight loss – right scale) results during combustion of the CNTs are presented in Fig. 45.9. The CNTs formed with CH4/H2 begin to combust at ~600  C while oxidation of the CNTs formed with only CH4 already initiated at ~500  C. This indicates that CNTs formed during CH4 pyrolysis are more thermally stable in the presence of a H2 cofeed. This is further reflected by the derivatives of the weight loss curves that represent the rate of combustion of the CNTs showing that pyrolysis in CH4 results in a maximum combustion rate at ~570  C and pyrolysis in CH4/H2 exhibits a maximum combustion rate at 634  C. The total amount of CNTs combusted, however, is greater for CNTs synthesized in CH4/H2 (92%) than synthesized in CH4 (68%). These TGA/DTA findings demonstrate that the CNTs formed during pyrolysis in CH4 and CH4/H2 are not identical and can be chemically probed by TPO-TGA and TPO-DTA.

45.4.4 Temperature-Programmed Reduction (TPR) TPR is one of the most extensively employed thermoanalytical techniques and is performed by flowing a reducing gas (e.g., H2/inert, CO, CH4, etc.) to reduce the material

45

Temperature-Programmed (TP) Techniques

(e.g., V2O5 þ 2 H2 ➔ V2O3 þ 2 H2O, V2O5 þ 2 CO ➔ V2O3 þ 2 CO2, 2 Fe3O4 þ 2 CH4 ➔ Fe6C2 þ 4 H2O). TPR is capable of distinguishing between multiple reducible species present in the material since each species may have different reduction properties. H2-TPR and CO-TPR measurements can be conducted with TGA, DTA, TCD, or MS analysis. For the more complex reduction with CH4 that involves the presence of more than one type of gas molecule, MS would be the detector of choice.

1013

503 VOx/Al2O3 493

30VAl

H2-TPR The nature of the supported VOx sites on an Al2O3 support was chemically probed by H2-TPR with an online TCD sensor and the reduction spectra, representing the consumption of H2, as a function of vanadia loading on the alumina support are presented in Fig. 45.10 [34]. The Al2O3 support does not undergo reduction, and only VOx-containing catalysts exhibit a reduction signal. The supported 1VAl catalyst exhibits one reduction peak at ~550  C that shifts progressively to lower temperatures with increasing vanadic loading up to 15VAl (~466  C). For higher vanadia loading, however, the H2-TPR peak shifts progressively to higher temperatures with increasing vanadia loading. Corresponding Raman and UV-vis analysis of the supported VOx/Al2O3 catalysts revealed that below 20% VOx the vanadia is present as two-dimensional surface VOx sites that progressively oligomerize with surface coverage [35]. The decreasing H2-TPR peak (Tp) with increasing vanadia loading below 20% VOx, which corresponds to monolayer surface coverage, reveals that oligomeric surface VOx sites are more reducible than isolated surface VOx sites. The increase in the H2-TPR Tp values with vanadia loading above 20% VOx corresponds to the presence of V2O5 nanoparticles (NPs) residing on top of the surface VOx monolayer [35]. The more sluggish reduction of the V2O5 NPs than the surface VOx sites is related to the slower reduction kinetics of the oxygen sites in the bulk lattice of the V2O5 NPs. The H2-TPR is able to distinguish between the different reducibility characteristics of isolated surface VOx sites, oligomeric surface VOx sites, and V2O5 NPs with the following ease of reduction: oligomeric surface VOx sites > isolated surface VOx sites > V2O5 NPs. The influence of the oxide support on supported VOx catalysts (ZrO2, TiO2, Al2O3, and SiO2) on the reducibility of the VOx phase was chemically probed with H2-TPR. The H2-TPR peak maxima (Tp) and the H/V ratios as a function of vanadium oxide loading are indicated in Table 45.1. The Tp value of the supported vanadium oxide catalysts is a strong function of the oxide support reflecting a strong support effect on the reduction of the supported vanadia phase. The surface vanadium oxide phases on the oxide supports for 1% VOx primarily consist of isolated surface VOx sites, and their reduction is a strong function of temperature (ZrO2 (370  C)  TiO2 (425  C)  SiO2 (520  C) < Al2O3

469 25VAl 466

45

20VAl 15VAl

467

10VAl 505 5VAl

525

3VAl

550

1VAl Al2O3

200

300

400

500

600

700

Temperature (°C)

Fig. 45.10 The H2-TPR results of supported V2O5 on Al2O3 as a function of V2O5 loading. The experiments were run in flowing 10% H2/Ar (80 cc/min) with a heating rate of 10  C/min using about 20 mg of catalyst and monitored by TCD detector. (Reproduced from ref. [34] [Copyright (2006) with the permission from Elsevier])

(540  C)) [35, 36]. Decreasing Tp values indicate that the kinetics for reduction are faster. With increasing vanadia loading, the Tp value shifts due to both oligomerization of the surface VOx sites and the presence of larger amounts of H2O in the gas phase that retard the reduction kinetics as well as the presence of V2O5 NPs at the higher loadings above monolayer surface coverage [37]. The number of H atoms consumed were also quantified, and the H/V ~2 indicates that one O atom was removed from each VOx site that corresponds to the reduction of the initial V+5 to V+3. Interestingly, the H/V ratios for TiO2 and ZrO2 at the lower loadings were slightly greater than 2 reflecting slight reduction of these oxide supports during the H2-TPR experiments [38]. In contrast, the H/V ratios for the Al2O3- and SiO2-supported vanadia catalysts are slightly less than 2 reflecting that these oxide supports are not being reduced during H2-TPR and that they are not allowing complete reduction of the surface V+5 sites to V+3 sites. It is, thus, not surprising that the turnover

1014

J.-M. Jehng et al.

Table 45.1 Maximum reduction temperature (Tmax) and Hatom consumed per Vatom of supported vanadia catalysts Hatom/Vatom 2.1 2.0 1.9 2.2 2.1 2.0 2.0 1.6 1.6 1.8 1.8 1.7 1.9 1.9 1.9 1.9 2.0 1.9

frequency (TOF) values for all redox reactions by supported V2O5/ZrO2 and V2O5/TiO2 catalysts are significantly greater than the supported V2O5/Al2O3 and V2O5/SiO2 catalysts [38]. The influence of dispersed zirconia on the reduction characteristics of supported VOx/ZrO2/SiO2 catalysts was examined with H2-TPR, and the spectra are presented in Fig. 45.11 as a function of ZrO2 loading. Characterization of the supported ZrO2/SiO2 catalysts with Raman and UV-vis spectroscopy revealed that the supported zirconia phase is present as surface ZrOx sites on the SiO2 support [39]. Depositing the surface VOx sites, confirmed with Raman, on the ZrO2/SiO2 decreased the reduction Tp from 526 to 495  C indicating that (1) the surface VOx sites are interacting with the surface ZrOx site, (2) the VOx-ZrOx interaction enhances the reduction of the surface VOx sites (lower Tp), and (3) the surface ZrOx sites possess different characteristics than the crystalline bulk ZrO2 (monoclinic) phase. The influence of the addition of sodium oxide to the supported ReOx/TiO2 catalyst upon the H2-TPR profile is shown in Fig. 45.12 [40]. The interaction of sodium oxide with the surface ReOx sites was confirmed by Raman spectroscopy and found to strongly retard the reduction of the surface ReOx sites on the TiO2 support as shown in Fig. 45.13 (reflected by the ~200  C increase in Tp values). The pronounced effect of sodium oxide on the supported ReOx/TiO2 catalyst system confirms that (1) sodium is directly interacting with the surface ReOx sites and (2) sodium retards the reduction kinetics.

CO-TPR The CO-TPR results of the Fe2O3, Cr2O3Fe2O3, and CuOCr2O3Fe2O3 catalysts were shown in Fig. 45.13

526 522 5% V2O5/SiO2 H2-consumption (arb. unit)

Tmax 370 380 430 425 450 456 470 540 520 505 495 495 500 520 520 526 540 560

515 5% V2O5/1% ZrO2/SiO2 507

5% V2O5/5% ZrO2/SiO2

495

5% V2O5/10% ZrO2/SiO2 5% V2O5/15% ZrO2/SiO2

433

5% V2O5/ZrO2 200

300

400 500 Temperature (°C)

600

700

Fig. 45.11 H2-TPR profiles of 5% V2O5 on SiO2, ZrO2, and ZrO2/SiO2 as a function of ZrO2 loading. (Reproduced with the permission from ref. [39] [Copyright (1999) American Chemical Society])

Effect of sodium on 3.1% Re2O7/TiO2 433

428 TCD signal (arbitrary units)

Catalysts 1% V2O5/ZrO2 2% V2O5/ZrO2 5% V2O5/ZrO2 1% V2O5/TiO2 3% V2O5/TiO2 5% V2O5/TiO2 6% V2O5/TiO2 1% V2O5/Al2O3 3% V2O5/Al2O3 5% V2O5/Al2O3 10% V2O5/Al2O3 20% V2O5/Al2O3 25% V2O5/Al2O3 1% V2O5/SiO2 3% V2O5/SiO2 5% V2O5/SiO2 10% V2O5/SiO2 15% V2O5/SiO2

3% Na2O

2% Na2O

401

1% Na2O 224 0% Na2O

100

200

(u2)

300 400 Temperature (°C)

500

600

Fig. 45.12 TPR profile of the sodium-doped on Re2O7/TiO2. (Reproduced with the permission from ref. [40] [Copyright (1998) The Royal Society of Chemistry])

8.0 × 10–8 176 °C 6.0 × 10–8

137 °C

4.0 × 10–8 2.0 × 10–8

300

ZSM-5 fresh (44 mg) Na-ZSM5 (68 mg)

420

CO2 MS signal

1.0 × 10–7

Fe2O3 Cr-Fe Cu-Cr-Fe

225

1.2 × 10–7

1015

211

Temperature-Programmed (TP) Techniques

TCD signal

45

100 150 200 250

0.0 50

100 150 200 250 300 350 400 450 500 Bed temperature (°C)

Fig. 45.13 CO-TPR experiments of Fe2O3, CrFe, and CuCrFe catalysts. The experiments were performed in the following procedures: (1) dehydrated under 10% O2/Ar at 350  C for 1 h, (2) activated the catalyst under WGS reaction at 350  C for 90 min, (3) the reactor was flushed by He and then cooled down to 80  C, and (4) a flow of 30 mL/min 10% CO/Ar was then introduced, and the temperature was ramped up to 450  C at a rate of 10  C/min, and analyzed the products by an online quadrupole mass spectrometer. (Reproduced with permission from ref. [41] [Copyright (2016) American Chemical Society])

[41]. The formation of CO2 was classified into two reaction pathways: (1) CO reacts with surface oxygen species possessing low temperature formation; (2) CO reacts with bulk lattice oxygen species possessing high temperature formation. The addition of Cr promoter into iron oxide has no effect on the peak of the low temperature production of CO2 which remains the same weak and broad peak at about 176  C in the region of 100–250  C as the iron oxide itself. However, the strong CO2 production peak has been retarded to higher temperature above 275  C. The addition of the Cr promoter does not affect the low temperature CO2 peak involving surface region oxygen but retards the onset of the removal of oxygen from the bulk lattice of the catalyst. The addition of Cu into the Cr2O3Fe2O3 catalyst restructures the Cr-Fe surface to form Cu nanoparticles and Cu-Fe interfacial sites, shifts the low-temperature peak to 137  C, and also enhances the interaction of the bulk lattice oxygen with CO to increase the reactivity of the WGS reaction [41]. TPR is widely used to characterize the catalyst’s reducibility given its high sensitivity to the reduction properties of a material’s composition, additives (promoters and poisons), and specific oxide support. In addition, it has the potential to distinguish between different sites that may be present in the same material as well as quantify the concentration of each site (represented by the area under the TPR peak).

100

200

300 400 Temperature (°C)

500

600

45 Fig. 45.14 NH3-TPD profiles of the ZSM-5 and Na-ZSM-5. The experiments were performed with the following procedures: (1) calcined the catalyst at 500  C for 1 h in flowing of 10% O2/Ar to remove surface water molecules and other impurities, (2) cooled to 100  C and switched to Ar gas to flush out the oxygen molecules, (3) a ~2000 ppm NH3 balanced in He was introduced into the reactor to start the adsorption for 30 min, and flushed with Ar gas for additional 30 min to remove the physical adsorbed NH3, and (4) NH3-TPD was initiated and monitored with an online TCD detector at a 10  C/min ramping rate from 100 to 600  C in flowing He gas

45.4.5 Temperature-Programmed Desorption (TPD) NH3-TPD TPD can be applied to determine surface characteristics of materials by monitoring the desorbing molecules during temperature programming with an online TCD or MS. For example, desorption of basic molecules is used to chemically probe the surface acid sites of a catalytic material. The NH3-TPD method was applied to probe the surface acidity of the hydrogen and Na forms of the ZSM-5 (Si/Al ¼ 15) zeolite. A 60–100 mg of catalysts was placed in a U-type quartz cell using an Altamira AMI-200 coupled with TCD detector. Ammonia was adsorbed on the dehydrated ZSM-5 catalysts at 100  C to minimize the presence of physisorbed molecular ammonia that desorbs between 60 and 100  C. The zeolite was subsequently temperature programmed in flowing He and the desorbing NH3 monitored with an online TCD with the NH3-TPD spectra presented in Fig. 45.14. The broad and overlapping NH3 desorption peaks can arise from the (i) nonuniformity of the surface of the catalysts, (ii) diffusional limitations present for the high-surface area zeolites, and (iii) readsorption of NH3. For the H-ZSM-5 zeolite, two ammonia desorption peaks are present at 225 and 420  C that correspond to weakly and

1016

J.-M. Jehng et al.

strongly chemisorbed ammonia species on surface acid sites. Upon addition of the basic Na, the NH3-TPD peak from the stronger acid sites are essentially absent, and only ammonia desorption from weaker acid sites is present. The relative NH3-TPD peak area of the Na-ZSM-5 and H-ZSM-5 is calculated to be 0.76 indicating the decrease of the total amount of the adsorbed ammonia upon the presence of Na on the surface. The NH3-TPD study demonstrates the sensitivity of the TPD technique to subtle changes in the surface properties of catalytic materials.

O2-TPD Supported Ag/α-Al2O3 catalysts are employed in the industrial manufacture of ethylene epoxide by selective oxidation of ethylene. This has generated much interest into the nature of the oxygen species present in Ag catalytic materials [42–45]. The O2-TPD spectra from α-Al2O3 that have been oxidized at 350  C are shown in Fig. 45.15. The α-Al2O3 does not release any oxygen, but several O2 desorption peaks are present from the supported Ag/α-Al2O3 catalyst. The low-temperature O2 desorption peak at ~240  C is associated with desorption of surface oxygen species, and the high-temperature O2 desorption peaks at ~347 and 394  C are desorption of oxygen originating from lattice atomic oxygen. Given that the ethylene oxidation reaction by supported Ag/α-Al2O3 catalysts is performed at temperatures of ~220–250  C, it is concluded that the surface oxygen species desorbing at ~240  C are the active oxygen species involved in the ethylene oxidation reaction.

O2 (m/e=32) MS signal (arbitrary units)

243

300

45.4.6 Temperature-Programmed Surface Reaction (TPSR) The most informative temperature programmed method is TPSR since it chemically probes the surface chemistry of catalytic active surface sites. The TPSR experiments can be performed by two methods: (1) initially adsorbing the reactive molecule before initiating temperature programming, which is ideal for molecules that can readily adsorb on the surface, or (2) flowing the reactive molecule if it is difficult to adsorb on the catalyst surface at modest temperatures (e.g., alkanes, CO, CO2, SO2, NO, etc.). Such experiments usually involve a gas flowing through the fixed-bed of catalyst powder that also serves to bring the gases to a downstream online detector as shown in Fig. 45.16. For reactions, the detector of choice is an MS because multiple products are usually formed and only an MS can distinguish between different gas molecules in real time. An additional advantage of the MS detector is that it can also distinguish between isotopes such as hydrogen (H2, HD, and D2), oxygen (16O2, 16O18O, and 18O2), and water (H216O/ H218O, HD16O/HD18O, and D216O/D218O). Such isotopic labeling allows for obtaining fundamental insights about rate-determining steps and reaction mechanisms. Additionally, TPSR spectra also contain information about the number of surface sites (Ns), which corresponds to the area under the desorption curve.

Number and Types of Surface Sites (Ns) Methanol readily adsorbs on surfaces as surface methoxy (CH3O*) intermediates that allow determining the number of surface sites present on oxides and metals (TGA, IR, etc.). Furthermore, the surface CH3O* also provides information about the surface chemistry of the surface sites making methanol a “smart chemical probe molecule” [46]. On surface redox sites, methanol will yield redox products (formaldehyde (HCHO), methyl formate (CH3OOCH), and methylal (H2C(OCH3)2), on surface acid sites methanol will form

347 394 15% Ag/D-Al2O3 Ventilation D-Al2O3

100

200

300

400

500

600

700

Sample

Mass spectrometer

Temperature (°C) Mass flow meters

Fig. 45.15 O2-TPD profile of α-Al2O3 and supported Ag/α-Al2O3 catalyst after the following procedure: (1) the catalyst was oxidized with 10% O2/Ar at 350  C for 1 h, (2) cooled to 50  C under 10% O2/ Ar, then flushed with Ar for 30 min, and (3) temperature programmed from 50 to 600  C with a 10  C/min ramping rate with flowing He

Furnace

Fig. 45.16 Schematic diagram of temperature-programmed system with fixed-bed of powder catalyst and flowing gases

45

Temperature-Programmed (TP) Techniques

1017

dimethyl ether (CH3OCH3), and on basic sites methanol will produce CO2. Therefore, knowledge of the number of surface methoxy species formed during methanol chemisorption is key for the determination of the number of surface redox, acid, and basic active sites on a catalytic material. The technique can be applied to bulk metal /metal oxide, bulk mixed metal oxide/ metal alloy, and supported metal oxide/metal catalysts in order to establish the number of surface sites per unit surface area (Ns). The number of chemisorbed surface methoxy species was calculated from the weight gain of the catalysts by the TGA under the specified conditions for maximum surface methoxy production, and subtracting the contribution of the water molecules formed and desorbed during the methanol chemisorption process. Methanol chemisorption on a series of supported molybdenum oxide catalysts with monolayer surface MoOx coverage is presented in Table 45.2. With the exception of the supported MoO3/SiO2 catalyst that does not have monolayer coverage, the CH3O*/ Mo ratio varied from ~0.2 to 0.4 indicating that only a fraction of the surface MoOx sites adsorb methanol because methanol forms one surface methoxy per 3 Mo atoms on the surface molybdenum oxide species due to repulsive interactions between surface alkoxy groups bound to adjacent surface MoOx oxide sites [47]. The CH3O*/Mo ratio is also independent of the coordination of the surface MoOx sites. In addition, the number of surface sites for the bulk MoO3, Fe2O3, Fe2(MoO3)4, and supported MoO3/Fe2O3 catalysts were also determined by integrating the HCHO/ CH3OH-TPSR spectra and are listed in Table 45.3. The CH3OH-TPSR experiments were performed with the AMI-200 temperature-programmed system (Zeton-Altamira Instruments) equipped with an online mass spectrometer (Dycor DyMaxion, Ametek Process Instruments). Typically, about 10–100 mg of catalyst was loaded in a U-type quartz tube and pretreated at 400  C for 40 min to dehydrate the sample. The catalysts subsequently cooled in flowing dry air to ~100  C and then switched to a flowing He stream to flush out any residual gas phase O2 for 30 min. Methanol

adsorption was achieved by flowing 2000 ppm CH3OH/He for 30 min, and the system was then purged with flowing He for additional 60 min. Afterward, the catalysts were heated at a constant heating rate of 10  C/min to 490  C in flowing He (30 cc/min). The effluent gases from the quartz tube reactor were analyzed with an online mass spectrometer, which was linked via a capillary tube, as a function of catalyst temperature. Bulk MoO3 contains about 0.9 μmol/m2 of active redox sites. The significantly lower value of bulk MoO3 is related to the anisotropic morphology and the preferential chemisorption of methanol at the edge sites of the platelets [48]. The stoichiometric Fe2(MoO4)3 mixed oxide is isotropic and possesses a higher surface density about 6.1 μmol/m2 of active redox sites. The presence of surface MoO3 monolayer on the Fe2O3 support dramatically decreases the surface density of surface active sites, but the nature of the surface sites changes from acidic to redox property.

Supported Metal Oxide Catalysts The CH3OH-TPSR spectra for the supported 18% MoO3/ Al2O3 catalyst containing about monolayer surface MoOx coverage are shown in Fig. 45.17 [49]. The first molecule to desorb is CH3OH (Tp ~ 160  C) that originates from Lewisbound CH3OH and formation of CH3OH by adjacent of two CH3O* species (2 CH3O* ➔ CH3OH þ HCHO). The primary reaction product is HCHO (Tp ~ 250  C) indicating that Table 45.3 The formation of surface MoO3 monolayer on Fe2O3 changes the nature of the Fe2O3 surface and possesses similar surface properties as the Fe2(MoO3)4 mixed oxide Sample (calcination temperature in  C) Fe2O3 (350)

BET surface area (m2/g) 23

MoO3 (400) Fe2(MoO4)3 (500) 2.3% MoO3/Fe2O3 (350)

2 12 23

Ns (# of active sites) (μmol/m2) 3 (CH3OH) 1.4 (redox) 3.4 (acid) 0.9 (redox) 6.1 (redox) 1.6 (redox)

Table 45.2 Surface molybdenum oxide structures and methanol chemisorption data reproduced from ref. [48] Catalyst 3% MoO3/ZrO2 6% MoO3/ Nb2O5 6% MoO3/TiO2 2% MoO3/MnO 3% MoO3/Cr2O3 18% MoO3/ Al2O3 4% MoO3/NiO 5% MoO3/SiO2

Mo molecular structure [47] Oh + Td (polym.) Oh (polym.)

Mo surface density [47] μmole/ m2 5.31 7.64

Ns (μmole methoxy/ m2 1.25 2.10

Methoxy adsorbed Per Mo atom 0.24 0.28

Oh (polym.) Not determined Not determined Oh + Td (polym.)

7.64 7.50 7.42 6.97

3.07 1.14 2.87 2.78

0.40 0.15 0.39 0.40

Not determined Td (isolated)

7.32 0.70

1.47 0.95

0.20 1.35

Copyright (2000) with the permission from Elsevier Oh octahedral coordination, Td tetrahedral coordination, polym. polymerized

45

1018

J.-M. Jehng et al.

1.60E–008 188

60

m/e 18

1.40E–008

50 40

MS signal (A.U.)

Mass spectral intensity (au)

70

m/e 32

30 m/e 28

20

m/e 30

10 m/e 45

0 –10

0

50

100

150 200 250 Temperature (°C)

1st cycle

Bulk V2O5 sequential HCHO/CH3OH

1.20E–008 2nd cycle 1.00E–008 3rd cycle 8.00E–009 6.00E–009

300

350

400

Fig. 45.17 CH3OH-TPSR spectra of the 18% MoO3/Al2O3 catalyst after methanol adsorption at 100  C and temperature ramping to 400  C. The desorbing products were monitored online with the MS: methanol (m/z ¼ 32), formaldehyde (m/z ¼ 30), water (m/z ¼ 18), dimethyl ether (m/c ¼ 45), and products of methanol and formaldehyde cracking in the mass spectrometer (m/z ¼ 28). (Reproduced from ref. [49] [Copyright (2000) with the permission from Elsevier])

the surface MoOx sites have redox character. The minor reaction product is dimethyl ether (DME) (Tp ~ 225  C) from surface acid sites (either from acid sites from exposed AlOx or surface MoOx sites). Carbon dioxide is not a significant product reflecting the absence of basic surface sites. Water is continuously produced with temperature reflecting its ability to be held up by adsorption/desorption processes because of its strong interaction with the catalyst surface. The TPSR experiments were carried out with the following procedure: The catalyst ( 10 mg in powder form) was placed in the sample pan of a microbalance (Cahn R.G.UHV), where it was pretreated by heating to 400  C at 1  108 Torr followed by oxidation at the same temperature in 10 Torr of oxygen. Chemisorption was performed by exposing the pretreated sample to 1000 mTorr of methanol vapor for 16 min. The amount of adsorbed methanol was measured gravimetrically as the stable weight after pumping out at room temperature up to 108 Torr. Adsorption experiments were performed at 100  C. After saturation of the surface was reached, the sample was heated up to 400  C and the effluent gases were continuously monitored with a quadrupole mass spectrometer [49].

Bulk Oxide Catalysts Bulk V2O5 The production of formaldehyde (FA) during CH3OH-TPSR from the bulk V2O5 catalyst is presented in Fig. 45.18 as a function of several TPSR cycles without exposure to gas phase-molecular O2 between each cycle. The formation of HCHO from the surface methoxy species on redox sites

100

200

300 400 Temperature (°C)

500

Fig. 45.18 CH3OH-TPSR spectra from bulk V2O5 after several cycles. The experiments were performed with the following procedures: (1) The catalyst was dehydrated under dry air at 350  C for 1 h, (2) cooled down to 50  C under flowing dry air, then switched to He to flush the residual gas phase oxygen for additional 30 min, (3) methanol adsorption was achieved by flowing 2000 ppm CH3OH/He for 30 min, and the system was then purged with flowing He for another 30 min, (4) the catalysts were heated at a constant heating rate of 10  C/min to 500  C in flowing He (30 mL/min), and the gases exiting from the quartz tube reactor were analyzed with an online MS. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

always occurs with a Tp of ~188  C, and the peak temperature does not change position with each subsequent reducing cycle with CH3OH. The relatively constant Tp of ~188  C suggests that a high concentration of active surface V5+ redox sites are present on the bulk V2O5 powder. The ability to maintain the surface V5+ sites fully oxidized implies that oxygen atoms from the bulk lattice are able to easily diffuse to the surface and reoxidize the reduced surface V cations created by the CH3OH-TPSR experiment. This reveals that CH3OH oxidation reaction involves the significant participation of bulk lattice oxygen atoms and proceeds via the Marsvan Krevelen mechanism. The decrease in the total quantity of HCHO product is associated with some sintering of this thermally fragile material during the TPSR experiments [50]. Bulk MoO3 The sequential CH3OH-TPSR spectra from the bulk MoO3 catalyst are shown in Fig. 45.19. Unlike the response observed for bulk V2O5, the FA Tp value for MoO3 continuously increases with each reduction cycle (~196 ➔ 221  C). This change is analogous to the reduction of surface V5+ to V4+/V3+ observed by Vohs et al. [51, 52], by combining CH3OH-TPSR and XPS analysis, and reflects the reduction of the surface Mo species with each cycle. The ease of reduction of the surface Mo6+ sites to lower oxidation states with each cycle reflects the inability or sluggishness of the bulk lattice oxygen from MoO3 to diffuse to the surface and

Temperature-Programmed (TP) Techniques

Bulk MoO3 sequential HCHO/CH3OH 213 2nd cycle 221 3rd cycle

1.20E–008

1.60E–007

196 1st cycle

1.00E–008

MS signal (A.U.)

1019

8.00E–009 6.00E–009 4.00E–009

432 HCHO (m/e = 30)

1.20E–007 1.00E–007 8.00E–008 6.00E–008 4.00E–008 2.00E–008

2.00E–009

Bulk TeO2

1.40E–007 MS signal (A.U.)

45

DMM (m/e = 45) DMM (m/e = 75) MF (m/e = 60)

438

0.00E+000 100

0.00E+000 100

150

200 250 300 Temperature (°C)

350

400

Fig. 45.19 CH3OH-TPSR spectra from bulk MoO3 after several cycles. The experiments followed the same procedures as described in Fig. 45.12. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

4.00E–008

Bulk Nb2O5

150

299

MS signal (A.U.)

3.00E–008 MeOH (m/e = 31)

DME (m/e = 45)

2.00E–008

HCHO (m/e = 30) DMM (m/e = 75) MF (m/e = 60)

1.00E–008

0.00E+000 100

150

200 250 300 Temperature (°C)

350

400

Fig. 45.20 CH3OH-TPSR spectra from bulk Nb2O5. The experiments followed the same procedures as described in Fig. 45.12. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

fully reoxidize the reduced MoOx sites. The reduction in the total amount of FA formed with each cycle is due to sintering of this thermally fragile metal oxide and the creation of surface-reduced MoOx sites that decrease methanol chemisorption [50]. Bulk Nb2O5 The CH3OH-TPSR spectra for the Nb2O5 support are shown in Fig. 45.20. The formation of dimethyl ether (DME, CH3OCH3) with a Tp about 300  C originates from surface methoxy reactions over surface acidic sites. These are weak acidic sites since pyridine adsorption IR measurements did not detect any surface acidic sites on the calcined Nb2O5

300 400 200 Temperature (°C)

486 490 MeOH (m/e = 31) 500

Fig. 45.21 CH3OH-TPSR spectra from bulk TeO2. The experiments followed the same procedures as described in Fig. 45.12. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

[53, 54]. Desorption of CH3OH with a Tp about 150  C is primarily due to molecularly adsorbed methanol and is also obtained when the adsorption temperature is further decreased toward room temperature. The high temperature tail of the CH3OH curve originates from surface methoxy (CH3O*) species that become rehydrogenated by surface hydroxyls to form CH3OH during the TPSR experiment. No formation of HCHO from surface redox sites is observed [55, 56]. Thus, the Nb2O5 support contains weak surface Lewis acid sites that convert the surface methoxy (CH3O*) species to DME and also back to CH3OH. Bulk TeO2 The CH3OH-TPSR spectra for the bulk TeO2 catalyst are presented in Fig. 45.21. The selective formation of HCHO indicates that surface TeOx sites exclusively possess redox character. The surface TeOx sites, however, are the least active sites since its Tp value of 432  C for reaction of the surface methoxy species is dramatically higher than that for V2O5 (188  C), MoO3 (196  C), and Nb2O5 (299  C). Bulk Mixed Oxides Bulk mixed oxide catalysts find wide application as oxidation catalysts as well as other catalytic reactions. The combination of metal oxides usually has been found to give rise to improved catalytic performance. As indicated above, surface methoxy species on V2O5 are oxidized to HCHO at 188  C while the corresponding surface methoxy species on MoO3 become oxidized to HCHO at 196  C, reflecting the slightly greater redox activity of surface VOx sites than surface MoOx sites. The binary bulk Mo0.6V1.5Ox mixed metal oxide, however, produces HCHO from surface methoxy species at Tp ~173  C (see Fig. 45.22), which is slightly lower than that for

45

1020

J.-M. Jehng et al.

3.00E–008 139

Mo0.6V1.5Ox MeOH (m/e = 31)

1.00E–008

2.50E–008

MS signal (A.U.)

MS signal (A.U.)

1.50E–008

173 HCHO (m/e = 30)

5.00E–009

Mo1.0V0.3Te0.16Nb0.12Ox

133

MeOH (m/e = 31)

2.00E–008 1.50E–008

173

1.00E–008 146 5.00E–009

DME (m/e = 45) 0.00E+000

DME (m/e = 45)

0.00E+000 50

100

150

200 250 300 Temperature (°C)

350

400

Fig. 45.22 CH3OH-TPSR spectra of binary bulk Mo0.6V1.5Ox, mixed oxide catalyst. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

5.00E–009 139

Mo1.0V0.5Te0.16Ox MeOH (m/e = 31)

4.00E–009 MS signal (A.U.)

HCHO (m/e = 30)

3.00E–009

175 HCHO (m/e = 30)

2.00E–009

DME (m/e = 45) 1.00E–009 0.00E+000 50

100

150

200 250 300 Temperature (°C)

350

400

Fig. 45.23 CH3OH-TPSR spectra of tertiary bulk Mo1.0V0.5Te0.16Ox mixed oxide catalysts. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

bulk V2O5 and MoO3. The shift of the HCHO Tp value from 188/196 to 173  C from the V2O5/MoO3 to the binary bulk Mo0.6V1.5Ox mixed metal oxide, respectively, reveals a synergistic interaction between surface VOx and MoOx sites. Similar results were observed for the tertiary bulk Mo1.0V0.5Te0.16Ox mixed metal oxide system that yields HCHO at a Tp of ~175  C (see Fig. 45.23). The surface of the quaternary bulk Mo1.0V0.3Te0.16Nb0.12Ox mixed metal oxide was also found to yield HCHO (Tp at about 173  C), as well as DME (Tp at about 146  C), as shown in Fig. 45.24. The most active redox site on these bulk mixed oxides is surface V5+Ox sites, and the temperature shift from 188  C for V2O5 to ~174  C for the bulk mixed oxides during oxidation of the surface methoxy species suggests that the adjacent MoOx sites are promoting the surface V5+Ox sites. The

50

100

150

200 250 300 Temperature (°C)

350

400

Fig. 45.24 CH3OH-TPSR spectra of quaternary bulk Mo1.0V0.3Te0.16Nb0.12Ox mixed oxide catalysts. (Reproduced with the permission from ref. [50] [Copyright (2005) American Chemical Society])

presence of TeOx does not affect the Tp value for surface methoxy oxidation to HCHO that either reflects (i) the very low activity of surface TeOx sites (Tp ¼ 432  C), (ii) its inability to promote the surface VOx sites (redox Tp constant at ~174  C), or (iii) the absence of surface TeOx sites for the bulk Mo1.0V0.5Te0.16Ox mixed metal oxide catalyst. The presence of NbOx did not affect the redox Tp value for the bulk Mo1.0V0.3Te0.16Nb0.12Ox mixed oxide, but introduced surface acid sites (DME at 146  C). The same redox Tp value with and without NbOx indicates that NbOx sites are not promoting the surface VOx sites. The addition of NbOx, however, results in a much lower acidic Tp value at 146  C than from bulk Nb2O5 at 299  C. This suggests that surface NbOx sites are probably not directly responsible for the surface acidity but may indirectly promote the surface acidity of adjacent VOx and MoOx sites that are known to form DME in this temperature range [50]. The CH3OH-TPSR spectra have shown to be able to provide fundamental surface information about the nature (elements and their oxidation states), surface chemistry, and reactivity of the active surface sites present on the outermost surface layer of bulk Mo-V-Te-Nb-O mixed metal oxides. The Tp values are sensitive to each specific cation, its surface chemistry (redox, acid, or basic character), its oxidation state (Tp increase with extent of reduction), and specific reactivity (V > Mo; V5+ > V4+ > V3+). This latter characteristic can also be applied to determine if a Mars-van Krevelen reaction mechanism over bulk mixed oxides proceeds via bulk or surface lattice oxygen. The chemical nature of the active surface sites is reflected in the methanol oxidation reaction products formed (e.g., HCHO from surface redox sites and DME from surface acidic sites) [50].

45

Temperature-Programmed (TP) Techniques

1021

Bulk Fe2(MoO4)3 Mixed Oxide The CH3OH-TPSR measurements were performed to investigate the surface chemical properties of Fe2O3, MoO3, supported MoO3 on Fe2O3, and bulk Fe2(MoO4)3 catalysts. The CH3OH-TPSR spectra from Fe2O3 are presented in Fig. 45.25a, and the primary reaction product is DME (Tp ~ 242  C) from surface acid sites and some FA (Tp ~ 187  C) from surface redox sites. In addition, CH3OH also desorbs (Tp ~ 188  C) and a trace amount of CO/CO2 (Tp ~ oC) desorbs from surface basic sites. The large amounts of DME, low amounts of FA, and trace of CO/CO2 reveal that the surface of Fe2O3 mostly consists of surface acid sites (Lewis acid sites) [57]. The desorption of CH3OH is related

a)

to desorption of chemisorbed molecular CH3OH at Lewis acid sites and the recombination of surface CH3O* and H* species [57]. For the MoO3 catalyst, the CH3OH-TPSR (Fig. 45.25b) indicates that the major product is FA (Tp ~ 205  C) from surface redox sites, with small amounts of CO/CO2 (Tp ~ 220  C) from surface basic sites and essentially no DME from surface acid sites. This indicates that the surface of MoO3 mainly consists of surface redox sites. The CH3OH-TPSR spectra from the bulk Fe2(MoO4)3 mixed oxide catalyst are presented in Fig. 45.25c and are quite similar to the CH3OH-TPSR spectra from MoO3. The primary product from Fe2(MoO4)3 is FA (Tp ~ 209  C); small amounts of CO/CO2 (Tp ~ 220  C) are formed and DME is

45

b)

188 °C

205 °C

Fe2O3

CH3OH

Signal intensity (a.u.)

Signal intensity (a.u.)

242 °C DME

187 °C HCHO

MoO3

HCHO

CH3OH CO

CO

CO2 DME

CO2 200

Signal intensity (a.u.)

c)

250 300 Temperature (°C) 209 °C HCHO

CH3OH

350

400

150

d)

300

190 °C

Fe2(MoO4)3

350

400

2.3% MoO3/Fe2O3

HCHO

CO2

CO

CO2 CO

DME 200

250

Temperature (°C)

CH3OH

150

200

Signal intensity (a.u.)

150

DME 250 300 Temperature (°C)

350

400

Fig. 45.25 CH3OH-TPSR spectra of (a) bulk Fe2O3; (b) bulk MoO3; (c) Fe2(MoO4)3 (Mo/Fe ¼ 1.5); and (d) 2.3% MoO3/Fe2O3. The experiments were performed with the following procedures: (1) The catalyst was dehydrated under dry air at 400  C for 40 min; (2) cooled down to 100  C under flowing dry air, then switched to He to flush the residual gas phase oxygen for additional 30 min; (3) methanol adsorption was

150

200

250 300 Temperature (°C)

350

400

achieved by flowing 2000 ppm CH3OH/He for 30 min, and the system was then purged with flowing He for another 60 min; and (4) the catalysts were heated at a constant heating rate of 10  C/min to 490  C in flowing He (30 mL/min). The gases exiting from the quartz tube reactor were analyzed with an online MS. (Reproduced from the Journal of Catalysis [58] with permission from Elsevier)

1022

Source of Oxygen Involved in Oxidation Reactions The source of oxygen involved in the rate-determining-step of oxidation reactions by oxide catalysts is an important aspect of such catalysts: catalyst bulk lattice (Mars-van Krevelen) or gas phase molecular O2 (LangmuirHinshelwood). To probe for the possible participation of bulk lattice oxygen for methanol oxidation, a series of sequential CH3OH-TPSR experiments were conducted in the absence of gas phase-molecular O2, both during TPSR and in between each TPSR cycle. If an oxide catalyst is able to efficiently provide bulk lattice oxygen for the CH3OHTPSR cyclic studies, then the amount of oxygen being provided will be constant and the Tp value will not change. If an oxide catalyst is not able to efficiently provide bulk lattice oxygen for the CH3OH-TPSR cyclic studies, then the amount of oxygen being provided will decrease and/or the Tp value will increase. If an oxide catalyst is not able to provide bulk or surface lattice oxygen for the CH3OH-TPSR studies, then bulk lattice oxygen is not being provided for methanol oxidation and any redox products could only be provided by the participation of gaseous molecular O2. Experiments to distinguish between the participation of gas phase molecular O2 and lattice oxygen from the oxide catalysts have typically employed the expensive 18O2 isotope that is not required with TPSR studies since TPSR can be conducted with the presence and absence of gas phase molecular O2. Comparison of the TPSR spectra from experiments with and without gas phase molecular O2 can inform on the roles of gas phase molecular O2 and lattice oxygen from the catalyst. Supported Metal Oxide Catalysts The supported 5% V2O5/TiO2 catalyst consists of monolayer surface vanadia coverage on the TiO2 support, and the

CH3OH-TPSR spectra from these catalysts in the presence and absence of gas phase molecular O2 are presented in Fig. 45.26 [59]. In the absence of gas phase-molecular O2 in flowing He, the methanol oxidation redox reaction to FA exhibits a Tp ~ 191  C. The same Tp value is also found when the experiment is repeated in the presence of gas phasemolecular O2. The simultaneous Tp values indicate that the rate-determining-step of surface CH3O* oxidation to FA proceeds by a Mars-van Krevelen mechanism employing surface lattice oxygen from the surface VOx sites and that the only role of gas phase molecular O2 is to rapidly reoxidize the reduced VOx sites [60]. Unlike bulk oxide catalysts that can store bulk lattice oxygen, supported VOx catalysts can only store a limited amount of surface oxygen that is associated with the surface VOx sites. This is demonstrated by the sequential CH3OHTPSR in the absence of gas phase molecular O2 from the supported 3% V2O5/ZrO2, consisting of ¾ monolayer of a surface VOx on the ZrO2 support, as shown in Fig. 45.27. The first cycle is the CH3OH-TPSR from the fully oxidized catalyst with the FA Tp ~159  C; Tp values sequentially increase with repeated CH3OH-TPSR cycles, and the amount of FA production decreases. The increase in Tp values reflects the reduction of the initial V+5 sites to lower oxidation states (V+4 and V+3), and surface V+5 sites are much more active than surface V+4/V+3 sites for methanol oxidation [59]. The

Tp = 191 °C

4 × 10–7

5% V2O5/TiO2 Tads = 100 °C

H2CO MS signal (a.u.)

essentially absent. The similarity between the CH3OH-TPSR of MoO3 and Fe2(MoO4)3 catalysts and the different products from CH3OH-TPSR from Fe2O3 and Fe2(MoO4)3 suggest that the Fe2(MoO4) catalyst is surface enriched with MoOx. To test the hypothesis that the Fe2(MoO4)3 catalyst is surface enriched with MoOx, a supported 2.3% MoO3/ Fe2O3 catalyst was synthesized and confirmed to have a surface MoOx monolayer on the Fe2O3 support. The CH3OH-TPSR spectra from the supported MoO3/Fe2O3 catalyst are presented in Fig. 45.25d and are almost identical to the CH3OH-TPSR from the Fe2(MoO4)3 catalyst. The absence of any DME from acid sites associated with surface FeOx sites and dominance of the FA product confirms that the Fe2(MoO4)3 catalyst is indeed surface enriched with MoOx [58]. The slightly lower FA Tp value from MoO3/Fe2O3 (Tp ~ 190  C) than Fe2(MoO4)3 (Tp ~ 209  C) indicates that the surface MoOx redox are slightly more active for the supported MoO3/Fe2O3 catalyst than that for the Fe2(MoO4)3 mixed oxide catalyst.

J.-M. Jehng et al.

5% O2/He

1 × 10–7 He

100

200

300

400

500

Temperature (°C)

Fig. 45.26 CH3OH-TPSR spectra of 5% V2O5/TiO2 with/without the presence of gas phase O2. (Reproduced with permission of Royal Society of Chemistry (Great Britain), from ref. [59]; permission conveyed through Copyright Clearance Center, Inc.)

45

Temperature-Programmed (TP) Techniques

1023

decrease in total FA production demonstrates that surface V+5 sites are more efficient in chemisorbing and oxidizing methanol to FA. Bulk Mixed Oxide Catalysts Methanol oxidation to formaldehyde by the bulk Fe2(MoO4)3 mixed oxide catalyst is a major commercial chemical process [61]. The reaction mechanism for CH3OH oxidation to FA by the bulk Fe2(MoO4)3 catalyst was investigated with

3% V2O5/ZrO2 HCHO/CH3OH in He Tads : 50 °C

1.00E–008

MS intensity (A.U.)

8.00E–009

159 1st cycle

6.00E–009

168 2nd cycle

4.00E–009

185 3rd cycle 185 4th cycle 185 5th cycle

2.00E–009

CH3OH-TPSR to determine the source of oxygen involved in the rate-determining-step of the reaction (the oxidative dehydrogenation of the surface CH3O* reaction intermediate) [62]: catalyst bulk lattice (Mars-van Krevelen) or gas phase-molecular O2 (Langmuir-Hinshelwood). Sequential CH3OH-TPSR experiments were performed with the bulk Fe2(MoO4)3 mixed oxide catalyst in the absence of gas phase-molecular O2 and are shown in Fig. 45.28 [58]. The Tp value for production of FA from the redox surface sites remained at ~208  C independent of the number of cycles of CH3OH-TPSR demonstrating that the oxygen involved in the oxidation of the surface CH3O* intermediate was being supplied by the catalyst bulk lattice that maintained the surface MoOx sites fully oxidized as Mo+6. The amount of fully oxidized surface MoOx sites was also constant indicating the efficient reoxidation of the surface MoOx sites by the bulk oxygen lattice. Interestingly, the CH3OH-TPSR study also reveals that MoOx on the surface of the bulk Fe2(MoO4)3 mixed oxide catalyst follows the Mars-van Krevelen mechanism while bulk MoO3 does not since the FA Tp value increases with each CH3OH indicating lack of full reoxidation of the surface MoOx sites.

0.00E+000

100

200

300 400 Temperature (°C)

500

Fig. 45.27 CH3OH-TPSR spectra of 3% V2O5/ZrO2 as a function of partial prereduction (achieved by performing sequential experiments in flowing He without catalyst reoxidation). (The first cycle corresponds to the fully oxidized catalyst republished with permission of Royal Society of Chemistry (Great Britain) from ref. [59]; permission conveyed through Copyright Clearance Center, Inc.)

Fe2(MoO4)3

208 °C

Signal intensity (a.u.)

Cycle1 Cycle2 Cycle3 Cycle4

150

200

250 300 Temperature (°C)

350

400

Fig. 45.28 Cyclic HCHO/CH3OH-TPSR spectra from the bulk Fe2(MoO4)3 (Mo/Fe ¼ 1.5) catalyst (the catalyst was not oxidized between the TPSR experiments). (Reproduced from the Journal of Catalysis from ref. [58] with permission from Elsevier)

Reaction Mechanisms Water-Gas Shift (WGS) Reaction (CO þ H2O $ H2 þ CO2) The TPSR transient kinetic studies for H2 production during HT-WGS by the Cr2O3Fe2O3 mixed oxide catalyst were investigated with TPSR. The Cr2O3Fe2O3 catalyst was prepared by coprecipitation method and consists of 8 wt % Cr2O3 and 92 wt % Fe2O3. The evolution of CO2 and H2 from CO þ H2O-TPSR and HCOOH-TPSR are compared in Fig. 45.29 [63]. The evolution of CO2 and H2 from HCOOH decomposition was investigated since it proceeds via the surface HCOO* intermediate, which has been proposed as a reaction intermediate during the WGS reaction. For HCOOH decomposition, the evolution of the CO2 and H2 reaction products initiates at the same temperature and follows the exact same kinetics as expected for decomposition of a common surface reaction intermediate (HCOO*), which is the rate-determining-step [64]. In contrast, the production of CO2/(CO þ H2O) begins at a much earlier temperature than H2/(CO þ H2O) formation because CO oxidation by surface O* is the first reaction step. The kinetics for evolution of CO2/(CO þ H2O) and H2/(CO þ H2O) above 250  C are not the same with more CO2 being initially formed than H2. Furthermore, the kinetics for CO2 and H2 evolution from CO þ H2O-TPSR are also different than found for the kinetics for CO2 and H2O production from the decomposition of HCOOH. The TPSR findings demonstrate that (i) the WGS-associated reaction mechanism through a common surface reaction intermediate,

45

1024

J.-M. Jehng et al.

Normalized mass intensity

1.0 0.8 CO2(HCOOH-TPSR) 0.6 H2(HCOOH-TPSR) 0.4

CO2(CO+H2O-TPSR)

0.2

H2(CO+H2O-TPSR)

0.0 –0.2 100

200

300 400 Bed temperature (°C)

500

Fig. 45.29 HCOOH-TPSR and CO þ H2O-TPSR experiments on CO2 and H2 production over the Cr2O3Fe2O3 catalyst. (Reproduced with the permission from ref. [63] [Copyright (2016) American Chemical Society])

such as surface HCOO* species, is not supported by the current findings and (ii) the current findings are only consistent with the WGS reaction proceeding through a redox or regenerative mechanism. The new insights suggest the following redox reaction mechanism for the HT-WGS reaction by chromium-iron mixed oxide catalysts. 

CO þ O $ CO2 



ð45:1Þ

CO2 $ CO2 þ



ð45:2Þ





ð45:3Þ

H2 O þ $ H2 O 









H2 O þ $ OH þ H 





ð45:4Þ



ð45:5Þ



ð45:6Þ

OH þ $ O þ H H þ H $ H2 þ 2



The oxidation of CO by surface O* appears rather straightforward, but isotopic oxygen studies showed rapid oxygen scrambling that also implicates the presence of surface carbonates (CO3*) during the HT-WGS [65]. The surface carbonates may just be formed by complexation of the CO2 product with surface O* and not directly involved in the HT-WGS reaction [66]. The details of the elementary steps involved in water decomposition during HT-WGS are not completely clear at present because formation of H2 must involve several reaction steps such as reactions (45.4)–(45.6). The current findings also suggest that activation of surface hydroxyls to yield H2 involves formation of surface vacant sites by CO oxidation. It appears that the HT-WGS shift reaction is much more complex involving multiple

elementary steps than the originally conceived reaction steps for the associative reaction mechanism. In conclusion, the evolution of CO2 and H2 from CO þ H2O-TPSR with activated Cr2O3Fe2O3 catalysts has, for the first time, been able to provide experimental evidence that the HT-WGS reaction follows a redox mechanism where the catalyst surface is alternatively reduced by CO and reoxidized by H2O [63]. The TPSR studies were carried out using an Altamira Instruments system (AMI-200) connected to a Dycor Dymaxion mass spectrometer (DME 200MS). A ~30 mg of catalyst was placed into a U-type fixed-bed quartz reactor and supported by quartz wool. The catalyst was first dehydrated under 10% O2/Ar at 350  C for 1 h, then switched to a mixed gas of 10% CO/Ar (10 cc/min) and He (30 cc/min) flowing through a water saturator at room temperature to perform the WGS activation at 350  C for 1 h. The catalyst was cooled to 110  C under flowing Ar for 15 min. The CO and H2O reactants (same as the WGS activation composition) and the HCOOH reactant (flowing He through a saturator containing formic acid at room temperature) were introduced into the reactor, separately, and heated with a constant ramping rate of 10  C/min from 110  C to 500  C. Selective Oxidation of C3H6 to C3H4O (Acrolein) C3H6-TPSR experiments have been undertaken to investigate the participation of gas phase molecular O2 and lattice O* oxygen during propylene oxidation to acrolein with the aid of isotopic 16O-18O labeling. These C3H6-TPSR experiments were performed with an Altamira Instruments temperature programmable system (AMI-200) equipped with an online quadrupole MS (Dycor Dymaxion, DME200MS). Typically, 200 mg of the supported V2O5/Nb2O5 catalyst was loaded into a U-shaped quartz tube and initially calcined at 450  C in flowing air for 60 min to remove adsorbed moisture and any combustible impurities that may be present. The pretreated catalyst was subsequently cooled down in flowing air to 110  C, and the flowing gas stream was switched to the ultrahigh purity He upon further cooling to 70  C for 30 min. Finally, propylene was adsorbed on the catalyst surface by flowing 2.96% C3H6/He for 30 min. Acrolein was found to only be produced when gas phasemolecular O2 is present during the propylene-dissociative chemisorption step. This suggests that surface oxygen supplied by gaseous molecular O2 is required to oxidize the surface H* to H2O in order to prevent the reversible hydrogenation of the surface C3H5* intermediately back to propylene, which is consistent with a Langmuir-Hinshelwood reaction mechanism. The presence of gaseous molecular O2 does not affect the surface allyl oxidation step since the same surface kinetics (same Tp value) occur in the presence and absence of gaseous molecular O2 during the TPSR reaction (see Fig. 45.30). Furthermore, only acrolein-16O is formed in the presence of gaseous molecular 18O2 during propylene

45

Temperature-Programmed (TP) Techniques

1025

Acrolein/C3H6 V2O5/Nb2O5 (8.4 V/nm2) Tads 70°C

MS intensity (a.u.)

198

100

200

a. step 1: He, step 2: He b. step 1: O2, step 2: O2 c. step 1: O2, step 2: He d. step 1: He, step 2: O2

300

400

500

Temp (°C)

MS intensity (a.u.)

Fig. 45.30 Influence of gaseous molecular O2 upon acrolein formation from propylene oxidation during C3H6- TPSR spectroscopy. (Reproduced from ref. [10])

Acrolein/C3H6 V2O5/Nb2O5 (8.4 V/nm2) Tads 70°C 18 16 18 198 Acr- O in step 1: O2, step 2: O2 Acr-16O in step 1: 16O2, step 2: 18O2 Acr-18O in step 1: 18O2, step 2: 16O2 Acr-16O in step 1: 18O2, step 2: 16O2

100

200

300 Temp (°C)

400

500

Fig. 45.31 Influence of gaseous molecular 18O2 upon C3H6-TPSR spectroscopy. Experimental conditions: (e) 16O2 in step 1 and 18O2 in step 2; (f) 18O2 in step 1 and 16O2 in step 2. (Reproduced from ref. [10])

adsorption and surface allyl oxidation steps. This observation shows that the oxygen inserted into the surface allyl intermediate originates from the catalyst lattice oxygen via a Marsvan Krevelen mechanism (see Fig. 45.31) [10]. The above findings reveal that the selective oxidation of propylene to acrolein proceeds via a Langmuir-Hinshelwood mechanism during the dissociative chemisorption step (both C3H6 and O2 must be present) and a Mars-van Krevelen reaction mechanism during the surface allyl oxygen insertion step (catalyst lattice oxygen provides the oxygen). This is the first time that a combined Langmuir-Hinshelwood-Mars-van Krevelen (L-H-M-V-K) reaction mechanism has been found to be operative for a selective oxidation reaction [10].

Selective Catalytic Reduction (SCR) of NO with NH3 Temperature programmed studies were performed using an Altamira AMI-200 system equipped with an online quadrupole mass spectrometer (Dycor Dymaxion DME200MS). Total 50 mg of catalyst was loaded into a U-type quartz tube and initially treated in flowing 5% O2/ He to 500  C (30 ml/min; heating rate of 10  C/min) for 60 min and cooled to 50  C. After the O2 treatment, different sets of isotopically labeled SCR reaction mixtures were introduced (a) 35 ml/min of 2000 ppm NH3/He, 35 ml/min of 2000 ppm 14N16O/He and 5 ml/min of 5% 16O2/He; (b) 35 ml/min of 2000 ppm ND3/He, 35 ml/min of 2000 ppm 14N16O/He and 5 ml/min of 5% 16O2/He; and (c) 35 ml/min of 2000 ppm NH3/He, 35 ml/min of 2000 ppm 14N16O/He and 5 ml/min of 5% 18 O2/He and (d) 35 ml/min of 2000 ppm NH3/He, 35 ml/min of 2000 ppm 15N18O/He and 5 ml/min of 5% 16 O2/He), then the sample was heated up to 500  C at a rate of 10  C/min. For the [14N16O þ NH3 þ 16O2]-TPSR after H218O pretreatment, the catalyst was exposed to 20 ml/min of He bubbling through a H218O (Sigma Aldrich, ISOTEC; 99% chemical purity; and 97% atom purity) saturator for 1 h at 400  C before cooling down in flowing helium to 50  C [67]. The supported 1%V2O5/5%WO3/TiO2 (1V5WTi) catalyst, which is known to exhibit higher SCR activity than the supported 1%V2O5/TiO2 (1VTi) and 5% WO3/TiO2 (5WTi) catalysts, also appears to be more active for N2O formation [68]. A series of temperature-programmed experiments with isotopically labeled molecules was performed to obtain fundamental insights about the reaction pathways for N2O formation during SCR with a supported 1% V2O5-5% WO3/ TiO2 catalyst consisting of surface VOx and WOx sites on the titania support. The MS signals of N2O produced during [15N18O þ 14NH3 þ 16O2]–TPSR are presented in Fig. 45.32. Only the isotopomers 14N15N16O (m/z ¼ 45) and 14N15N18O (m/z ¼ 47) were evolved demonstrating that N2O is formed by reaction between one NO molecule and one NH3 molecule as was previously reported. The absence of 14N216O (m/z ¼ 44) and 14N218O (m/z ¼ 46) rules out an N2O reaction pathway proceeding via oxidation of two 14NH3 molecules [67]. Additional isotope labeling TPSR experiments were designed to obtain further insights into the oxygen sources, gas phase-molecular O2, and catalyst lattice oxygen, for the N2O formation pathways. The [14N16O þ 14NH3 þ 18 O2]–TPSR and the [14N16O þ 14NH3 þ 16O2]–TPSR on H218O-pretreated catalyst (known to replace 16O of the surface VOx sites by 18O) [68] formed both the N216O and N218O isotopomers during the two experiments. The percentage of N218O among all N2O products, N218O/(N216O þ N218O), produced in the 400–500  C temperature range is presented in Fig. 45.33 as a function of temperature [67]. During [15N18O þ 14NH3 þ 16O2]–TPSR and [14N16O þ 14 NH3 þ 18O2]–TPSR, the percentage of N218O remained

45

1026

J.-M. Jehng et al.

a) 3 × 10–9

14N 16O 2

2 × 10–9

a) 100

(m/z = 44)

80

1 × 10–9

b) 3 × 10–9

60 40

150

200

250

300

400

14N15N16O

2 × 10–9 MS signal (a.u.)

350

450

500

1 × 10–9 0 100

c) 3 × 10–9

150

200

250

300

350

15N 16O 2

2 × 10–9

400

450

500

or N218O (m/z = 46)

1 × 10–9 0 100

d) 3 × 10–9

150

200

250

300

350

400

14N15N18O

2 × 10–9

20

(m/z = 45)

450

Percentage of N218O (%)

0 100

500

(m/z = 47)

420

440

460

480

500

420

440

460

480

500

480

500

80 60 40 20 0 400 c) 100 80 60 40

1 × 10–9

20

150

200

250 300 350 400 Bed temperature (°C)

450

0 400

500

Fig. 45.32 MS signals of 14N216O, 14N15N16O,15N216O, 14N218O, and 14 15 18 N N O during [15N18O þ 14NH3 þ 16O2]–TPSR for supported 1V5WTi catalyst. (Reproduced from ref. [67] [Copyright (2017) with the permission from Elsevier])

constant at 50% and 30%, respectively. The absence of 16 18 O O, 16O2, and 14N18O isotopomers demonstrates that no significant NO or O2 scrambling takes place under SCR reaction conditions. The rate-determining-step (rds) for N2O formation by the supported 1V5WTi catalyst can also be evaluated by probing for Kinetic Isotope Effects (KIE) during TPSR with judiciously designed isotopically labeled experiments: (1) [14N16O þ NH3 þ 16O2]–TPSR as reference; (2) [14N16O þ ND3 þ 16O2]–TPSR to evaluate N-H bond breaking; (3) [15N18O þ 14NH3 þ 16O2]–TPSR to evaluate N-O bond breaking; (4) [14N16O þ 14NH3 þ 18O2]–TPSR to evaluate O-O bond breaking; and (5) [14N16O þ 14NH3 þ 16 O2]–TPSR after H218O treatment to evaluate V-O bond breaking. The formation of N2O with the different isotopic labels is presented in Fig. 45.34 [65] and follows quite similar trends with the only exception when ND3 is the reactant. This TPSR experiment reveals that in the presence of the ND3 the N2O formation is significantly retarded and requires higher temperatures to proceed, demonstrating a kinetic isotope effect involving N-H/N-D bond breaking as the rate-determining-step. The N2O evolution at different temperatures also allows determining the influence of the isotopomers upon the apparent activation energy for N2O formation within the 450–485  C range.

420

440 460 Temperature (°C)

Fig. 45.33 Percentage of N218O evolved among N2O reaction products during (a) [15N18O þ 14NH3 þ 16O2]–TPSR; (b) [14N16O þ 14NH3 þ 18 O2]–TPSR; and (c) [14N16O þ 14NH3 þ 16O2]–TPSR after H218O pretreatment of the supported 1V5WTi catalyst. (Reproduced from ref. [67] [Copyright (2017) with permission from Elsevier])

5 × 10–9 4 × 10–9

N2O MS signal (a.u.)

0 100

0 400 b) 100

3 × 10–9

[14N16O + NH3 + 16O2] – TPSR [14N16O + ND3 + 16O2] – TPSR [15N18O + NH3 + 16O2] – TPSR [14N16O + NH3 + 18O2] – TPSR [14N16O + NH3 + 16O2] – TPSR after H218O pretreatment

2 × 10–9 1 × 10–9 0 100

150

200

250 300 350 400 Temperature (°C)

450

500

Fig. 45.34 MS signals of N2O formed during TPSR on 1V5WTi with various isotopically labeled reactants. (Reproduced from ref. [67] [Copyright (2017) with permission from Elsevier])

A series of TPSR experiments with isotopically labeled molecules (18O2, H218O, 15N18O, and ND3) allowed direct demonstration of many new fundamental insights about the formation pathway and rate-determining-step of the N2O side

45

Temperature-Programmed (TP) Techniques

reaction during SCR of NO with NH3 by the V2O5/WO3/ TiO2 catalysts: (1) surface VO4 species are the catalytic active sites while surface WOx species are not active and only act as promoters, (2) formation of N2O involves one NH3 molecule and one NO molecule, (3) NO, molecular O2, and lattice oxygen from the catalyst all contribute oxygen to the formation of N2O ( 50%, 30%, and 20%, respectively), and (4) the rate-determining-step was for the first time shown to involve breaking of the ammonia N-H bond.

45.5

Summary/Conclusion/Future Outlook

Temperature-programmed techniques are extremely informative and provide fundamental surface information about the number of surface sites, chemical nature of the surface sites, catalytic reaction mechanisms, and surface kinetics for the rate-determining steps for all types of solid catalysts (bulk oxides/metals, mixed oxides/metal alloys, supported metals/ metal oxides, and zeolites/molecular sieves). The ability to address such a wide range of problems for all types of solid catalysts makes temperature-programmed techniques among the most versatile methods in catalysis research. The combination of spectroscopy with temperatureprogrammed techniques opens up the ability to perform operando spectroscopy measurements (operando is a Latin word meaning operating and has been coined for studies that simultaneously combine spectroscopy and online reactivity/ selectivity measurements during relevant catalytic reaction conditions) [69, 70]. Such cutting edge studies allow simultaneously obtaining information about the catalyst bulk (structure and oxidation states), catalyst surface sites (structure and oxidation states), surface reaction intermediates, surface reaction pathways, surface kinetics, and rate-determining-step. Such fundamental molecular level studies are just beginning to be reported, and many more operando temperature-programmed techniques and studies will be reported in the coming years. Acknowledgments I. E. Wachs would like to thank the funding from the Department of Energy, Basic Energy Sciences (FG02-93ER14350). The travel and short-term research grant (MOST- 108-2918-I-005 -007) is gratefully appreciated by J. M. Jehng.

References 1. Cheng, S.E.D. (ed.): Handbook of Thermal analysis and Calorimetry: Application to Polymers and Plastics, vol. 3. Elsevier, Amsterdam (2002) 2. Auroux, A. (ed.): Calorimetry and Thermal Methods in Catalysis Springer Series in Material Science, vol. 154. Springer, Heidelberg (2013) 3. Redhead, P.A.: Thermal desorption of gases. Vacuum. 12, 203–211 (1962)

1027 4. Masel, R.I.: Principles of Adsorption and Reaction on Solid Surfaces Wiley Series in Chemical Engineering. Wiley, New York (1996) 5. Topsoe, N.-Y., Topsoe, H., Dumesic, J.A.: Vanadia/titania catalysts for selective catalytic reduction (SCR) of nitric oxide by ammonia: I. Combined temperature programmed in-situ FTIR and on-line mass spectroscopy studies. J. Catal. 151, 226–240 (1995) 6. Sato, S., Ukisu, Y., Ogawa, H., Takasu, Y.: Infrared reflection absorption spectroscopy, X-ray photoelectron spectroscopy and temperature-programmed desorption study on the absorption and decomposition of Fe(CO)5 over silver surfaces. J. Chem. Soc. Farad. Trans. 24, 3527–3780 (1998) 7. Bolina, A.S., Wolff, A.J., Brown, W.A.: Reflection absorption infrared spectroscopy and temperature programmed desorption investigations of the interaction of methanol with a graphite surface. J. Chem. Phys. 122, 044713 (2005) 8. Elmay, Y., Le Brech, Y., Delmotte, L., Dufour, A., Brosse, N., Gadiou, R.: Characterization of Miscanthus pyrolysis by DRIFTs, UV Raman spectroscopy and mass spectrometry. J. Analy. Appl. Pyrolysis. 113, 402–411 (2015) 9. Litorja, M., Haynes, C.L., Haes, A.J., Jensen, T.R., Van Duyne, R.P.: Surface-enhanced raman scattering detected temperature programmed desorption: optical properties, nanostructure, and stability of silver film over SiO2 nanosphere surfaces. J. Phys. Chem. B. 105, 6907–6915 (2001) 10. Zhao, C., Wachs, I.E.: An operando Raman, IR, and TPSR spectroscopic investigation of the selective oxidation of propylene to acrolein over a model supported vanadium oxide monolayer catalyst. J. Phys. Chem. C. 112, 11363–11372 (2008) 11. Xu, J., Deng, Y.-Q., Luo, Y., Mao, W., Yang, X.-J., Han, Y.-F.: Operando Raman spectroscopy and kinetic study of low-temperature CO oxidation on alpha-Mn2O3 nanocatalyst. J. Catal. 300, 225–234 (2013) 12. Brender, P., Gadiou, R., Rietsch, J.-C., Fioux, P., Dentzer, J., Ponche, A., Vix-Guterl, C.: Characterization of carbon surface chemistry by combined temperature programmed desorption with in situ X-ray photoelectron spectrometry and temperature programmed desorption with mass spectrometry analysis. Anal. Chem. 84, 2147–2153 (2012) 13. Pastore de Lima, A.E., de Oliveira, D.C.: In situ XANES study of cobalt in Co-Ce-Al catalyst applied to steam reforming of ethanol reaction. Catal. Today. 283, 104–109 (2017) 14. Souda, R.: A temperature-programmed time-of-flight secondary ion mass spectroscopy study of intermixing of amorphous ethanol and heavy-water films at 15–200 K. J. Chem. Phys. 122, 134711 (2005) 15. Wunderlich, B.: Thermal Analysis in Encyclopedia of Materials: Science and Technology, 2nd edn, pp. 9134–9141 (2001) 16. Watson, M., O’Neill, J.: Differential Microcalorimeter, U.S. Patent 3,263,484, issued August 2, 1966; filed April 4, (1962) 17. Madix, R.J.: The application of flash desorption spectroscopy to chemical reactions on surfaces: temperature programmed reaction spectroscopy. Crit. Rev. Solid States Mater. Sci. 7(2), 143–152 (1978) 18. Robertson, S.D., McNicol, B.D., de Baas, J.H., Kloet, S.C., Jenkins, J.W.: Determination of reducibility and identification of alloying in copper-nickel-on-silica catalysts by temperature-programmed reduction. J. Catal. 37(3), 424–431 (1975) 19. Golunski, S.E., Jenkins, J.W.: A tribute. Platinum Metals Rev. 52(4), 249–250 (2008) 20. Koga, N., Yamada, S., Kimura, T.: Thermal decomposition of silver carbonate: phenomenology and physicogeometrical kinetics. J. Phys. Chem. C. 117, 326–336 (2013) 21. Nakano, M., Fujiwara, T., Koga, N.: Thermal decomposition of silver acetate: physico-geometrical kinetic features and formation of silver nanoparticles. J. Phys. Chem. C. 120, 8841–8854 (2016) 22. Twu, J., Shih, C.-F., Guoa, T.-H., Chen, K.-H.: Raman spectroscopic studies of the thermal decomposition mechanism of ammonium metavanadate. J. Mater. Chem. 7(11), 2273–2277 (1997)

45

1028 23. Zhenshenga, J., Chanjuan, X., Qingmei, Z., Feng, Y., Jiazheng, Z., Jinzhen, X.: Catalytic behavior of nanoparticle α-PtO2 for ethanol oxidation. J. Mol. Catal. A Chem. 191, 61–66 (2003) 24. Liu, W.-J., Kao, T.-W., Dai, T.-M., Jehng, J.M.: Ni-based nanocomposites supported on graphene nano sheet (GNS) for supercapacitor applications. J. Solid State Electrochem. 18(1), 189–196 (2014) 25. Twu, J., Shih, C.-F., Guo, T.-H., Chen, K.-H.: Raman spectroscopic studies of the thermal decomposition mechanism of ammonium metavanadate. J. Mater. Chem. 7(11), 2273–2277 (1997) 26. Jiang, W., Nadeau, G., Zaghib, K., Kinoshita, K.: Thermal analysis of the oxidation of natural graphite – effect of particle size. Thermochim. Acta. 351(1–2), 85–93 (2000) 27. Bom, D., Andrews, R., Jacques, D., Anthony, J., Chen, B., Meier, M.S.: Thermogravimetric analysis of the oxidation of multiwalled carbon nanotubes: evidence for the role of defect sites in carbon nanotube chemistry. Nano Lett. 2(6), 615–619 (2002) 28. Laine, N.R., Vastola, F.J., Walker Jr., P.L.: The importance of active surface area in the carbon-oxygen reaction. J. Phys. Chem. 67(10), 2030–2034 (1963) 29. Jones, L.E., Thrower, P.A., Walker Jr., P.L.: Reactivity and related microstructure of 3d carbon/carbon composites. Carbon. 24(1), 51–59 (1986) 30. Walker Jr., P.L.: Carbon: an old but new material revisited. Carbon. 28(2/3), 261–279 (1990) 31. Welham, N.J., Williams, J.S.: Extended milling of graphite and activated carbon. Carbon. 36(9), 1309–1315 (1998) 32. Yao, N., Lordi, V., Ma, S.X.C., Dujardin, E., Krishnan, A., Treacy, M.M.J.: Structure and oxidation patterns of carbon nanotubes. J. Mater. Res. 13(9), 2432–2437 (1998) 33. Chen, C.-M., Dai, Y.-M., Huang, J.G., Jehng, J.-M.: Intermetallic catalyst for carbon nanotubes (CNTs) growth by thermal chemical vapor deposition method. Carbon. 44, 1808–1820 (2006) 34. Martinez-Huerta, M.V., Gao, X., Tian, H., Wachs, I.E., Fierro, J.L. G., Bañares, M.A.: Oxidative dehydrogenation of ethane to ethylene over alumina-supported vanadium oxide catalysts: relationship between molecular structures and chemical reactivity. Catal. Today. 118(3–4), 279–287 (2006) 35. Tain, H., Ross, E.I., Wachs, I.E.: Quantitative determination of the speciation of surface vanadium oxides and their catalytic activity. J. Phys. Chem. B. 110(19), 9593–9600 (2006) 36. Jaegers, N.R., Wan, C., Hu, M.Y., Vasiliu, M., Dixon, D.A., Walter, E., Wachs, I.E., Wang, Y., Hu, J.Z.: Investigation of silica-supported vanadium oxide catalysts by high-field 51V magic-angle spinning NMR. J. Phys. Chem. C. 121(11), 6246–6254 (2017) 37. Arena, F., Frusteri, F., Martra, G., Coluccia, S., Parmaliana, A.: Surface structures, reduction pattern and oxygen chemisorption of V2O5/SiO2 catalysts. J. Chem. Soc. Faraday Trans. 93, 3849–3854 (1997) 38. Deo, G., Wachs, I.E.: Reactivity of supported vanadium oxide catalysts: the partial oxidation of methanol. J. Catal. 146, 323–334 (1994) 39. Gao, X., Fierro, J.L.G., Wachs, I.E.: Structural characteristics and catalytic properties of highly dispersed ZrO2/SiO2 and V2O5/ZrO2/ SiO2 catalysts. Langmuir. 15, 3169–3178 (1999) 40. Mitra, B., Gao, X., Wachs, I.E., Hirt, A.M., Deo, G.: Characterization of supported rhenium oxide catalysts: effect of loading, support and additives. Phys. Chem. Chem. Phys. 3, 1144–1152 (2001) 41. Zhu, M., Rocha, T.C.R., Lunkenbein, T., Knop-Gericke, A., Schlögl, R., Wachs, I.E.: Promotion mechanisms of iron oxidebased high temperature water–gas shift catalysts by chromium and copper. ACS Catal. 6(7), 4455–4464 (2016) 42. Kim, Y.-C., Park, N.-C., Shin, J.-S., Lee, S.R., Lee, Y.J., Moon, D.J.: Partial oxidation of ethylene to ethylene oxide over nanosized Ag/α-Al2O3 catalysts. Catal. Today. 87, 153–162 (2003)

J.-M. Jehng et al. 43. Sung, Y.-E., Lee, W.Y., Rhee, H.-K., Lee, H.-I.: The effect of oxygen on the chemisorption on polycrystalline silver surface. Korean J. Chem. Eng. 6, 300–305 (1989) 44. Backx, C., de Groot, C.P.M., Biloen, P., Sachtler, W.M.H.: Interaction of O2, CO2, CO, C2H4 and C2H4O with Ag(110). Surf. Sci. 128, 81–103 (1983) 45. Grant, R.B., Lambert, R.M.: A single crystal study of the silvercatalysed selective oxidation and total oxidation of ethylene. J. Catal. 92, 364–375 (1985) 46. Badlani, M., Wachs, I.E.: Methanol: a “smart” chemical probe molecule. Catal. Lett. 75, 137–149 (2001) 47. Briand, L.E., Farneth, W.E., Wachs, I.E.: Quantitative determination of the number of active surface sites and the turnover frequencies for methanol oxidation over metal oxide catalysts: I. Fundamentals of the methanol chemisorption technique and application to monolayer supported molybdenum oxide catalysts. Catal. Today. 62, 219–229 (2000) 48. Farneth, W.E., Ohuchi, F., Staley, R.H., Chowdhry, U., Sleight, A.W.: Mechanism of partial oxidation of methanol over MoO3 as studied by temperature-programmed desorption. J. Phys. Chem. 89(12), 2493–2497 (1985) 49. Hu, H., Wachs, I.E.: Catalytic properties of supported molybdenum oxide catalysts: in situ Raman and methanol oxidation studies. J. Phys. Chem. 99(27), 10911–10922 (1995) 50. Wachs, I.E., Jehng, J.-M., Ueda, W.: Determination of the chemical nature of active surface sites present on bulk mixed metal oxide catalysts†. J. Phys. Chem. B. 109, 2275–2284 (2005) 51. Wong, G.S., Concepcion, M.R., Vohs, J.M.: Oxidation of methanol to formaldehyde on vanadia films supported on CeO2(111). J. Phys. Chem. B. 106, 6451–6455 (2002) 52. Wong, G., Vohs, J.: An XPS study of the growth and electronic structure of vanadia films supported on CeO2(111). Surf. Sci. 498, 266–274 (2002) 53. Datka, J., Turek, A.M., Jehng, J.-M., Wachs, I.E.: Acidic properties of supported niobium oxide catalysts: An infrared spectroscopy investigation. J. Catal. 135, 186–199 (1992) 54. Jehng, J.-M., Turek, A.M., Wachs, I.E.: Surface modified niobium oxide catalyst: synthesis, characterization, and catalysis. Appl. Catal. A. 83, 179–200 (1992) 55. Tatibouet, J.M.: Methanol oxidation as a catalytic surface probe. Appl. Catal. A. 148, 213–252 (1997) 56. Tatibouet, J.M., Lauron-Pernot, H.: Transient isotopic study of methanol oxidation on unsupported V2O5: mechanism of methylal formation. J. Mol. Catal. A Chem. 171, 205–216 (2001) 57. Burcham, L.J., Briand, L.E., Wachs, I.E.: Quantification of active sites for the determination of methanol oxidation turn-over frequencies using methanol chemisorption and in situ infrared techniques. 2. Bulk metal oxide catalysts. Langmuir. 17, 6175–6184 (2001) 58. Routray, K., Zhou, W., Kiely, C.J., Grünert, W., Wachs, I.E.: Origin of the synergistic interaction between MoO3 and iron molybdate for the selective oxidation of methanol to formaldehyde. J. Catal. 275, 84–98 (2010) 59. Wachs, I.E.: Catalysis science of supported vanadium oxide catalysts. Dalton Trans. 42, 11762–11769 (2013) 60. Mars, P., van Krevelen, D.W.: Oxidations carried out by means of vanadium oxide catalysts. Chem. Eng. Sci. 3, 41–59 (1954) 61. Soares, A.P.V., Portela, M.F., Klennemann, A.: Methanol selective oxidation to formaldehyde over iron-molybdate catalysts. Catal. Rev. Sci. Eng. 47, 125–174 (2005) 62. Routray, K., Briand, L.E., Wachs, I.E.: Is there a relationship between the M¼O bond length (strength) of bulk mixed metal oxides and their catalytic activity? J. Catal. 256, 145–153 (2008) 63. Zhu, M., Wachs, I.E.: Resolving the reaction mechanism for H2 formation from high-temperature water–gas shift by chromium–iron oxide catalysts. ACS Catal. 6, 2827–2830 (2016)

45

Temperature-Programmed (TP) Techniques

64. Poulston, S., Rowbotham, E., Stone, P., Parlett, P., Bowker, M.: Temperature-programmed desorption studies of methanol and formic acid decomposition on copper oxide surfaces. Catal. Lett. 52, 63–67 (1998) 65. Zhu, M., Wachs, I.E.: Determining number of active sites and TOF for the high-temperature water gas shift reaction by iron oxide-based catalysts. ACS Catal. 6, 1764–1767 (2016) 66. Kalamaras, C.M., Olympiou, G.G., Efstathiou, A.M.: The water-gas shift reaction on Pt/γ-Al2O3 catalyst: operando SSITKA-DRIFTSmass spectroscopy studies. Catal. Today. 138, 228–234 (2008) 67. Zhu, M., Lai, J.-K., Wachs, I.E.: Formation of N2O greenhouse gas during SCR of NO with NH3 by supported vanadium oxide catalysts. Appl. Catal. B Environ. 224, 836–840 (2018) 68. He, Y., Ford, M.E., Zhu, M., Liu, Q., Tumuluri, U., Wu, Z., Wachs, I.E.: Influence of catalyst synthesis method on selective catalytic reduction (SCR) of NO by NH3 with V2O5-WO3/TiO2 catalysts. Appl. Catal. B Environ. 193, 141–150 (2016) 69. Guerrero-Pérez, M.O., Bañares, M.A.: Operando Raman study of alumina-supported Sb–V–O catalyst during propane ammoxidation to acrylonitrile with on-line activity measurement. Chem. Commun. 1292–1293, 1292 (2002) 70. Bañares, M.A., Guerrero-Perez, M.O., Fierro, J.L.G., Cortez, G.G.: Raman spectroscopy during catalytic operations with on-line activity measurement (operando spectroscopy): a method for understanding the active centres of cations supported on porous materials. J. Mater. Chem. 12, 3337–3342 (2002)

1029

Israel E. Wachs received his PhD from Stanford University and started his career at Exxon Research & Engineering Company. He subsequently transitioned to academia and holds the Endowed G. Whitney Snyder Chair in the Department of Chemical & Biomolecular Engineering at Lehigh University. His research focuses on operando molecular spectroscopy of mixed oxide catalysts and development of their structure-activity relationships.

Michael Ford received his PhD from the University of East Anglia. After postdoctoral research at Colorado State University, he pursued a career in organic process discovery and development at Air Products and Chemicals. Upon retiring, he joined Israel Wachs’ Operando Molecular Spectroscopy & Catalysis Research group as a research associate, focusing on the synthesis, characterization, and applications of oxide catalysts.

Jih-Mirn Jehng received PhD from Lehigh University in 1990. He is working on design, synthesis, and characterization of heterogeneous catalysts with related chemical reactions such as partial oxidation, pollution abatements, alternative energy, and photocatalysis reactions. Recently, he is focusing on CO2 photocatalytic reduction to alcohols over nanostructure composite catalysts.

45

46

Calorimetry Techniques Lishil Silvester

, Quentin Touloumet

, and Aline Auroux

Contents 46.1 46.1.1 46.1.2 46.1.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operation Modes in Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . General Calorimetry Applications . . . . . . . . . . . . . . . . . . . . . . Measurement of Heats of Interactions . . . . . . . . . . . . . . . . . .

1031 1032 1033 1033

46.2 46.2.1 46.2.2 46.2.3 46.2.4

Differential Scanning Calorimetry (DSC) . . . . . . . . . . . . Heat Flux DSC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Power Compensated DSC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calvet DSC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catalytic Applications of DSC and Coupled DSC/TGA Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1034 1034 1035 1035

46.3

Calorimetry-Volumetry (Gas Adsorption Calorimetry) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039

46.4 46.4.1 46.4.2 46.4.3

Liquid Phase Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Titration Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Immersion Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reaction Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

46.5

Single Crystal Adsorption Calorimetry (SCAC) . . . . 1053

46.6

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056

1036

1045 1045 1047 1049

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056

Abstract

Calorimetry is quite an old but complex technique that is capable of directly measuring the heat associated with any physical transformations or chemical reactions. In the past few decades, the calorimetry techniques underwent drastic technical improvements in their design that help to investigate in details the catalytic properties and reaction L. Silvester Laboratoire des Multimatériaux et Interfaces – UMR CNRS 5615, Université Claude Bernard Lyon1, Villeurbanne CEDEX, France e-mail: [email protected] Q. Touloumet · A. Auroux (*) Institut de Recherches sur la Catalyse et l’Environnement de Lyon – UMR CNRS 5256, Université Claude Bernard Lyon1, Villeurbanne CEDEX, France e-mail: [email protected]; [email protected]

mechanisms, the two core areas in heterogeneous catalysis. Since calorimeters are designed to measure the heats associated with gas-solid, gas-liquid, and liquid-liquid interactions, they provide deep insight on the catalyst properties, kinetics and reaction energetics. This chapter gives a detailed overview of the basic principles of different calorimeters and their operation modes along with the examples of their catalytic applications particularly during the last one decade. Some of the major calorimetry techniques covered in this chapter include: Differential scanning calorimetry (DSC), Calorimetry-Volumetry (Gas adsorption calorimetry), Liquid phase calorimetry, and Single crystal adsorption calorimetry (SCAC). Keywords

Differential Scanning Calorimetry · CalorimetryVolumetry · Gas adsorption calorimetry · Liquid phase calorimetry · Titration calorimetry · Immersion calorimetry · Isothermal reaction calorimetry · Isoperibolic reaction calorimetry · Single Crystal Adsorption Calorimetry

46.1

Introduction

The term “Calorimetry” is derived from the Latin word “calor,” meaning heat and the Greek word “με
τρoν” (metron), meaning measure. As the term calorimetry indicates, it is a process to measure the heat (q) in any physical change (melting, crystallization, phase transition, etc. . .) or any chemical reactions (adsorption/desorption, reduction, oxidation, decomposition, etc. . .). Any device that is used to measure the heat flow involved in physical and chemical processes is commonly known as a “Calorimeter”. All calorimeters function on the basis of thermodynamics and law of conservation of energy (dU ¼ dq þ dw), i.e., energy (U ) transfer can take place only in the form of heat (q) and work

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_46

1031

1032

L. Silvester et al.

(W). Theory of almost all calorimeters is based on the calorific equation of Tian: P¼C

dT þ K ðT T 0 Þ dt

ð46:1Þ

where P ¼ power (W), C ¼ Heat capacity (J.K1), K ¼ thermal conductance (W.K1), (T  T0) ¼ temperature difference. For a calorimeter with a given time constant ‘τ’, the measurement is carried out either in the adiabatic way (C measurement) or in the quasi-static way (K measurement); τ ¼ C/K. Hence, a calorimeter mainly consists of a measurement chamber with a sample cell/crucible and a sensor like thermocouples, thermopiles, resistance wire, or Peltier elements that are capable of precisely determining the ‘C’ value or ‘K’ value and calculate the heat associated with the sample. Different types of calorimeters that are currently employed are given in Table 46.1. Due to the technological differences among calorimeters, there are various operation modes (temperature, programming rate, effective volume, sensitivity) and different types of applications (specific heat, reaction heat, gas-solid interaction) that are briefly discussed below.

46.1.1 Operation Modes in Calorimetry Some of the common usage modes in calorimeters are explained below.

Temperature Range The temperature in the calorimeters can lie in the range from around 200  C to 1500  C. The experiments related to fusion and solidification of organic liquids, measurements

of transitions of elastomers and silicones, and physisorption are performed at very low temperatures. High temperature mode is used in calorimetry for studies related to phase/ structure change of inorganic solids, solid decomposition, gas-solid interactions (oxidation, reduction), specific heats, and phase transitions.

Measurement Under Pressure In order to carry out the experiments under pressure, the specially designed calorimetric cells that are capable to adapt under pressure will be employed. The pressure arises in a calorimetric cell especially during experimental conditions involving the vapor pressure of the investigated samples, the sample decomposition vapors, and catalytic reactions in gas flow. However, the pressure inside the cell can be controlled during most of the experiments and it is even possible to pre-set the pressure before the experiment begins. Heating Rate Calorimetric measurements can be conducted in either isothermal or temperature-scanning mode. Heating rate is an important factor as it influences both the duration and the accuracy of a calorimetric experiment. Both duration and accuracy generally opposes to one another, i.e., precise heat measurements are usually difficult if the heating rate of the sample is very rapid. During high heating rate, the sample will never attain a thermal equilibrium with its environment. A high temperature programming rate is useful to investigate the behavior of a sample over a broad temperature range. However, a low heating rate is recommended for the accurate measurement of the specific heat or enthalpy associated with the solid transformations.

Table 46.1 Common calorimetric techniques, principles and examples Calorimetry technique Adiabatic calorimetry

Working principle Works under zero heat exchange conditions (no heat gain or loss)

Isothermal calorimetry

Heat exchange occurs between sample and surroundings at constant temperature

Heat exchange calorimetry

Heat exchange occurs between sample and surrounding thermostat while temperature of former or latter is kept constant Temperature variations of the sample are measured while temperature of the surrounding jacket fluid is kept constant

Isoperibolic calorimetry

Measure Enthalpy changes during physical process like crystallization, mixing, dilution, etc. . . Thermal runaway, thermal stability, and energy during reaction/decomposition Enthalpy changes and kinetics of a chemical reaction Heat of adsorption Enthalpy of reactions Specific heat capacities of liquids Enthalpy of reactions

Examples of calorimeters Accelerating rate calorimeter (ARC), adiabatic pressure Dewar calorimeter (ADC), bomb calorimeter, reaction calorimeter, power compensated DSC

Reaction calorimeter, adsorption calorimeter, titration calorimeter

Reaction calorimeter

Differential reaction calorimeter

46

Calorimetry Techniques

1033

Effective Sample Volume Since the amount or volume of the analyzed solid can have a great impact on the diffusion phenomena and the reactivity, the amount of the material should be chosen carefully during a calorimetric study. For example, a small amount of sample must be used during a measurement with high programming rate in order to avoid thermal gradients. Sensitivity Sensitivity is an essential parameter at a given temperature range in calorimetric experiments. Absolute sensitivity and relative sensitivity per unit of volume are two widely used sensitivity terms in calorimetry.

46.1.2 General Calorimetry Applications Most of the calorimeters are generally designed to measure specific heats, reaction heats, and heat of interactions whose basic principles are briefly explained below:

Specific Heat Measurements Three methods are used to measure the specific heats in a calorimeter: (1) continuous temperature scanning, (2) discontinuous temperature scanning (increments), and (3) dropping samples between T1 and T2. A continues temperature scanning works on the principle of the equation: Cp ¼ ðdH=dT Þp =ðdT=dtÞp

ð46:2Þ

where Cp ¼ (Calorimetric signal  Scanning rate). Two tests must be performed under identical experimental conditions. In the first test, both measuring and reference cells are empty and in the second test, the measuring cell contains sample but the reference cell is empty. The specific heat of the sample is measured from the difference in the calorimetric plots from two tests. Principle behind the discontinuous temperature scanning is based on the equation: h

ðH Þp

itn t0

h itn ¼ Cp ðT Þp t0

ð46:3Þ

where Cp ¼ ΔH=ΔT (ΔH is the area under heat curve and ‘ΔT’ is the difference in temperatures). A temperature increment is generally programmed between two temperatures ‘T1’ and ‘T2’ and the rise in temperature of the sample for this increase of temperature (ΔT ) is measured. As in continuous scanning, two tests are necessary also in discontinuous scanning method to measure the specific heat with accuracy.

This method is capable of measuring specific heats of both solids and liquids. The principle of the dropping sample method is to allow the sample to fall from a region with temperature ‘T1’ to a region with temperature ‘T2’ within the calorimeter. The heat associated with the temperature rise of the sample is measured and compared to heat associated with a reference sample (e.g., alumina, platinum) analyzed under identical conditions. The precision of this method rely on the measurement accuracy of the temperatures (T1 & T2).

Measurement of Reaction Heats In a heterogeneous reaction, the gaseous reactants are preheated and brought into contact with the solid catalyst placed inside a calorimetric cell. The reaction heat and the progress of the reaction are simultaneously measured by coupling the calorimeter to a gas chromatograph or a mass spectrometer. Hence, the heat of reaction associated with a specific product or a consumed reactant can be measured for a reaction in a steady state. During complex reactions with more than one product formation, the heat of conversion of the reactants to corresponding end products can be calculated from the system of equations deducted from the variations in the reaction kinetic parameters.

46.1.3 Measurement of Heats of Interactions Interactions between solid-liquid and solid-gas are the widely studied interactions using calorimetry. In solid-liquid adsorption, a solid is allowed to come in contact with a solution of adsorbent in increasing concentrations. By investigating the solid-liquid interactions in calorimetry, the parameters such as surface area of the solid, thermoporometry (porosity of adsorbents), and liquid adsorption can be determined. The solid-gas interactions are important in the field of heterogeneous catalysis as this interaction involves a wide range of energies according to the nature of the reactants. The phenomena of physical adsorption (characterized by low energies) and chemical adsorption (generally characterized by higher energies, specific to the established bond type) can be distinguished by calorimetric measurements of heats from solid-gas interactions. Calorimeters that obey one of the above-mentioned usage modes and calorimetric principles (in Table 46.1) and employed in the investigation of catalyst properties and/or catalytic reactions are detailed below. Most of the examples of catalysts/catalytic applications of calorimeters referred in this chapter are the works performed during the last one decade from the year 2010 (with few exceptions).

46

1034

46.2

L. Silvester et al.

Differential Scanning Calorimetry (DSC)

Differential scanning calorimeter is a calorimetric technique that measures the heat flow in and out of the sample relative to a reference. DSC normally measures the heat flux between a crucible containing the sample and a reference crucible during a controlled temperature ramping. Generally, both the sample and the reference are heated/cooled linearly while they are maintained at almost same temperature throughout the temperature ramping. The heat flux for a given sample at temperature ‘Ts’ is represented by equation: dqs dh C dT ¼ þ s s dt dt dt

ð46:4Þ

where ‘dh/dt’ is the heat flux produced by the sample transformation or the reaction and ‘Cs’ is the heat capacity of the sample including container. The heat flux ‘dqs/dt’ is exchanged to a thermostatic block at temperature ‘Tp’ through a thermal resistor having resistance ‘R’ according to following relation:   Tp  Ts dqs ¼ R dt

ð46:5Þ

It is evident from the above relation that any temperature perturbation of the thermostatic block will considerably affect the calorimetric measurements. In order to avoid this issue, a symmetrical design is used in which an identical crucible is placed on the detector at same temperature ‘Tp’ of thermostatic block. The difference of heat flux is measured between the sample and reference: dq dqs dqr dh C dT C dT ¼  ¼ þ s s r r dt dt dt dt dt dt

ð46:6Þ

where ‘Cr ’ is the heat capacity of the reference including container and ‘Tr ’ is the temperature of the reference. Now the heat flux can be represented by relation:

46.2.1 Heat Flux DSC

dq ðT r  T s Þ ¼ R dt or by its derivation: Rd2 q dT r dT s ¼  dt dt Rdt2

ð46:7Þ

By combining Eqs. (46.6) and (46.7), the characteristic equation of the calorimetric measurement can be obtained: ðCs  Cr ÞdT p RCs d2 q dq dh ¼ rþ  dt dt dt dt2

transformations or reactions. This value is also related to reaction kinetics and the shape of the DSC curve can give an indication on the rate of reaction. If ‘dh/dt’ value is zero (no transformations or reactions), the heat capacity of the sample can be retrieved from eq. (46.8). Hence, specific heat of any kind of material can be measured using DSC. During isothermal tests in DSC, the temperature of thermostatic block (Tp) remains steady and hence the parameter ‘dTp/dt’ is zero. In case if there is small variation in the ‘Tp’ value, the resulting thermal effect can be minimized if the heat capacities of the sample (Cs) and the reference (Cr) employed are similar. From a typical DSC curve, it is possible to deduce some important information such as the temperature of transformation, the heat (Q) of transformation, and the rate at which the transformation occurs. Figure 46.1 shows an example of a DSC curve in which the heat of transformation/reaction is obtained by integrating the area under the curve from the start of the peak at time ‘ti’ (corresponding to temperature Ti) until the end of the peak at time ‘tf ’ (corresponding to temperature Tf). The kinetics of transformation is the other important information that is directly associated to the DSC peak. As the DSC signal corresponds to a heat flow rate, the rate of the transformation can be directly measured from the shape of the peak. For example, the sharp and the broad peaks indicate a higher and a lower rate of reaction, respectively. The rate of the reaction (dh/dtmax) reaches the maximum value at the top of the DSC peak. Hence, the above-mentioned information obtained from the DSC curve can be used in many models to determine the kinetic parameters of a transformation. DSC’s are widely employed in both academics and industries due to its simplicity, rapid measurement ability, comparatively cheaper price, and availability. The principles of three widely used differential scanning calorimeters: Heat flux, Power compensated and Calvet type are described briefly below.

ð46:8Þ

The value of ‘dh/dt’ is positive for endothermic transformations or reactions and its value is negative for exothermic

In a heat flux DSC, the difference in heat flow rate into a sample and a reference is analyzed as a function of temperature while their temperature varied (heated or cooled) at a constant rate. Figure 46.2 shows the schematics of a Heat Flux DSC as an example. Heat Flux DSC consists of the sample holder, reference holder, the heat resistor, the heat sink, and the heater. Heat from the heater will flow through heat sink and heat resistor before reaching the sample and the reference. The heat difference between the sink and the holders is directly proportional to the heat flow. Since the heat capacity of the sink is much larger than the heat capacity of the sample, the heat loss or gain by the sample during endothermic or exothermic processes is compensated by the heat

46

Calorimetry Techniques

1035

Fig. 46.1 A typical DSC curve Thermal flux dq/dt (mW)

Exo

Temperature (°C)

Area = Q (heat expressed in mJ) Tf

Ti ti

tf

Time (s)

46 Fig. 46.2 Schematics of a heat flow DSC (adapted from HITACHI High-Tech)

Reference

Sample

Heat sink Heater Heat driver

CPU

Thermocouple Temp. control Heat resistor Amplifier Thermocouple

sink and maintains a constant temperature difference between the sample and the reference. The difference in the amount of heat between the sample and the reference is proportional to the temperature difference of both holders measured using the thermocouples connected to the holders.

Amplifier

Temp. recording Temp. difference (heat flux) recording

compensated by adjusting the power input to the sample and reference furnaces. The energy required to do adjust this power is used to measure the enthalpy or heat capacity changes in the sample relative to the reference.

46.2.3 Calvet DSC 46.2.2 Power Compensated DSC Power-compensated DSC is designed to measure precisely the enthalpy change of the sample as a function of time. The temperatures of the sample and reference are controlled independently using separate, identical furnaces (Fig. 46.3). In this DSC, temperature difference between a sample and a reference during an endothermic/exothermic process is

The working principle of Calvet DSC is same as that of heat flux DSC. As aforementioned for heat flux DSC, the difference in heat flow rate into a sample and a reference is analyzed as a function of temperature while their temperature is varied (heated or cooled) at a constant rate in Calvet DSC. A general schematic of Calvet type DSC is shown in Fig. 46.4. Compared to point source detectors used in heat

1036

L. Silvester et al.

flux and power compensated DSC’s that are only 25–30% efficient in capturing the heat, the detector in Calvet type DSC is highly efficient (~90–95%) and is based on a three dimensional fluxmeter sensor consisting of highly thermal conductive thermopile (series of thermocouples) completely surrounding the crucible. An added advantage of the 3D sensor in Calvet type DSC is that the calibration is independent of atmosphere or gas pressure hence eliminating the

Sample

Reference

Platinium resistance

need to recalibrate the instrument during the changes in any gas flow rates, pressure, or composition of the gas. DSC’s have been employed in various applications such as polymers (glass transitions), pharmaceuticals (phase diagram, polymorphism, purity, thermal stability), safety processes (transition, decomposition under high pressure), energy (heat storage, hydrogen storage), and catalysis (thermal stability, activation energy, kinetics, sintering, coke deposition). Since, this book chapter is meant for the calorimetric applications in the field of catalysis, only catalysis related applications of DSC’s will be discussed. Most often DSC is coupled with a thermogravimetric analysis (TGA) to measure weight and heat simultaneously that help to gain more detailed mechanisms related to catalysts or catalytic reactions. A simple schematic of a Calvet type TGA-DSC unit is shown in Fig. 46.5. In a TGA-DSC apparatus, the weight gain/loss and the heat associated with the weight change are recorded versus time and temperature.

Heaters

46.2.4 Catalytic Applications of DSC and Coupled DSC/TGA Unit Heat sink

Fig. 46.3 Simple representation of power compensated DSC

Heatflux

Thermal stability of catalysts is an important parameter as it is related to their structural changes including decomposition and phase transformations. Therefore, catalysts stability is generally investigated before employing them in any catalytic reactions conducted at temperatures other than ambient. Though, thermogravimetry (TGA) is a simple technique that

Transducers Microbalance

DSC

Sample

Reference Thermostated

Calorimetric

Block

Fig. 46.4 Pattern of a three dimensional cylindrical Tian-Calvet DSC [1]

Fig. 46.5 DSC coupled to a TGA microbalance

46

Calorimetry Techniques

1037

helps to learn the thermal stability of catalysts, DSC (or DSC/ TGA unit) is used to have more insights on the mechanisms (organic precursor removal, structural changes, phase transition, decomposition) associated with the thermal treatment of the catalysts. For example, the thermal stability of the ionic liquid immobilized silica gel was investigated using a DSC coupled to TGA (NETZSCH STA 409PC) as the weight loss and associated heat can be related to the loss of the organic components attached to the silica surface [2]. Results suggested that the ionic liquid immobilized silica gel is quite stable until 350  C after which an endothermic peak was observed at 410  C in DSC due to decomposition of the ionic liquid component. Similarly, DSC/TGA have been employed to illustrate the calcination processes of PMMA colloidal crystal template and ionic liquid, which are vital steps in the removal of the polymeric spheres and formation of active site in ionic liquid based Mo/TiO2 catalyst [3]. It showed that almost all PMMA template and ionic liquid decomposed completely at around 350  C (Fig. 46.6). The DSC/TGA results suggested that the heating rate and air velocity are very important in the process of calcination. Also, lower heating rate (1  C.min1) is advisable to attain proper melting and decomposition of PMMA template, while higher heating rate was found accelerating the decomposition process and tend to produce a large amount of gas, unfavorable to the formation of 3D ordered catalyst structure. DSC was also used to perform the kinetic study of a decomposition reaction. For example, electrolytic graphite oxide (EGO) exhibited exothermic peak in DSC due to decomposition of oxygen containing group and the activation energy of the decomposition reaction was estimated by applying Kissinger equation which described the relationship between heating rate and peak temperature [4]. The decomposition experiments were performed at different heating rates (5, 10, 15, and 20  C.min1) to attain a maximum

73.5%

PMMA

Weight loss (%)

80 –2 60 40

–3

20 0 200

300 400 500 Temperature (°C)

600

Temperature difference (PV/mg)

–1

100

–4 700

Fig. 46.6 Exothermic heat and weight change from DSC/TGA due to precursor removal during calcination of Mo/TiO2 based catalyst [3]

temperature of 400  C (Fig. 46.7a). The activation energy of EGO decomposition was calculated using the equation:  ln

V T 2P



  AR E ¼ ln  a Ea RT P

ð46:9Þ

where ‘V’ is heating rate, ‘TP’ is the temperature at which the heat flow reaches maximum, ‘Ea’ is activation energy, ‘A’ is frequency factor, and ‘R’ is universal gas constant. Activation energy (Ea) and frequency factor  (A) are calculated from the linear regression of the  ln

V

T 2P

vs 1/TP plot (Fig. 46.7b).

In other study, DSC is used to study the heat changes related to structural evolution of carbon nanotubes during purification [5]. Also, DSC enables to determine the suitable calcination temperature for the complete removal of organic template/surfactant used to prepare macroporous LaMnO3 catalysts [6]. K. Chen et al. studied the impact of phase transition of VO2 catalysts in oxidative desulphurization of dibenzothiophene [7]. They employed DSC/TGA unit (TA SDT Q600) to study the phase transition temperature of VO2 catalysts and found that phase transition temperature of monoclinic VO2 to rutile VO2 decreased with increasing cell volume for different VO2 catalysts. Also, the appearance of endothermic peak suggested the occurrence of structural phase transition from monoclinic VO2 to tetragonal VO2. A DSC/TGA aparatus (NETZSCH STA 449F3) was employed in a study describing the effect of morphology of Birnessite type MnO on its catalytic activity [8]. During thermal treatment, highly active nano flower MnO catalyst exhibited large endothermic peaks at 75  C and 125  C which are attributed to physisorbed and interlayer water in birnessite nanostructure. An exothermic DSC peak was observed at around 491  C due to the phase transformation from the layered structure of nanoflower MnO to the α-MnO2. When the temperature was further increased, two endothermic peaks appeared at 864  C and 949  C due to the structural transformations from MnO2 to Mn2O3 and from Mn2O3 to Mn3O4, respectively. DSC also helps to study the influence of dopants on the catalyst deactivation. Jin et al. used DSC to study the tolerance of the Mn-Ce/TiO2 catalyst toward SO2 poisoning during the selective catalytic reduction of NOx [9]. DSC/TGA results suggested that Ce addition could reduce thermal stabilities of the sulfate species covered on the catalyst surface, thereby improving the SO2 tolerance of the Ce modified catalysts. Hence, DSC has been employed to study the stability of the catalysts used in different reactions such as oxidation [8, 10], desulfurization [3, 7], combustion [6], acetalization [2], esterification [11], reforming [12, 13], methanation [14], etc. . .. The above examples clearly suggested the need of differential scanning calorimetry in studying the thermal stability (decomposition, structural changes, phase transition) of most of the catalysts.

46

1038

L. Silvester et al.

a)

b) 9.2 20°C/min

9.0

15°C/min

In (Tp2/V)

Heat flow (W/g)

8.8

10°C/min

8.6 8.4 8.2 8.0

5°C/min

Ea = 115.6 kJ.mol–1

7.8 0.0043

100

200

300

0.0044

0.0045

0.0046

0.0047

0.0048

0.0049

1/Tp (K2)

400

Temperature (°C)

Fig. 46.7 Decomposition of electrolytic graphite oxide in DSC: (a) Dynamic mode DSC thermograms with systematic variation in heating rate and   V (b) plot of  ln T 2 as a function of 1/TP [4] P

a) 100 Weight loss (%)

96 92 88 Ni-La/MgAl2O4

84

Ni/MgAl2O4 80

b)

4

2

0

Ni-La/MgAl2O4 Ni/MgAl2O4 100

200

300

400

500

600

700

DSC (mW/mg)

Apart from determining the thermal stability of catalysts, DSC has been successfully employed to study the activation energy, enthalpy and kinetics associated with different catalysts and catalytic reactions. For example, differential scanning calorimetry helps to study the catalytic properties in CO methanation (to synthetic natural gas) process that need a feasible catalyst with coke resistant and anti-sintering behavior [15]. DSC/TGA helped to identify different types of carbon species deposited on the non-promoted and lanthanum promoted Ni/MgAl2O4 catalysts used in CO methanation (Fig. 46.8). Results of spent catalysts indicated that only active nickel carbide species (DSC peak < 450  C) were formed on La promoted catalysts while filamentous/graphitic carbon species (DSC peak > 600  C) that cause catalysts deactivation were also observed on non-promoted catalyst [15]. DSC has been widely used to study the catalyst deactivation in different reforming reactions (methane, dry, ethanol, bio-oil) resulting from the deposition of different carbon species [12, 13, 16–19]. Different carbon species including amorphous carbon, graphitic carbon, nanotubes, and graphene formed on the surface of reforming catalysts exhibited exothermic peaks in the DSC at different temperatures. DSC/TGA studies demonstrated that some of these carbon species results in catalysts deactivation while others have no deactivating behavior. In other study, DSC analyses of the spent catalysts after used in a xylose cyclo-dehydration

–2 800

Temperature (°C)

Fig. 46.8 DSC/TGA of the spent Ni/MgAl2O4 and La doped Ni/MgAl2O4 catalysts [15]

reaction exhibited exothermic features at 200  C due to the deposited organic matter and an endothermic peak below 200  C attributed to physisorbed water and volatiles [20].

46

Calorimetry Techniques

Also, DSC results indicated that carbon deposition was lower on the composite catalyst due to favorable competitive adsorption effects and hence improved performance of this catalyst. Differential Scanning calorimetry allow determining the dehydrogenation/hydrogenation ability and kinetics of the hydride solids exhibiting excellent hydrogen storage capacities [21–23]. The decomposition of the hydrides and their re-hydrogenation generally cause variations in mass and heat flow which can be characterized by the DSC/TGA apparatus. Multiple hydrogenation/dehydrogenation cycles were performed in a DSC to study the stability, the low-pressure hydrogen absorption, and the cycling behavior of VTiCr/ MgH2 [21]. DSC measurements were also used to determine the enthalpies of decomposition/desorption of the LiAlH4/ single-walled carbon nanotube composite system [22]. Huang et al. investigated the non-isothermal dehydriding performance of the graphene supported Ni/MgH2 systems by differential scanning calorimeter [23]. For the first time, Eggenhuisen et al. reported the DSC analysis of nanoparticle induced constrictions by studying the delayed freezing of water in the constricted pores [24]. The freezing and melting behavior of water confined in the empty mesoporous silica supports and the catalysts was recorded using differential scanning calorimetry. The percentage of blocked or open pores was quantified from the melting enthalpy values obtained from DSC experiments conducted at very low temperatures up to 70  C. The same group used DSC for the quantitative analysis of entrance size in mesoporous materials like SBA-16 and FDU-12 with varying cage and entrance sizes [25]. They demonstrated that DSC is a powerful technique to analyze the cage meso-structures and can be applied to most of the materials showing constraint porosities. Recently, high-throughput (HT) methodologies are gaining interests in the field of catalysis as it not only enables a significant time, energy, and money savings but also reduce the time lag to industrialize a lab-scale production/reaction. J. Loskyll et al. reported a high-throughput calorimetry system to measure the heat of reaction and the catalytic activity of various heterogeneous catalysts in a fast, reliable and reproducible manner using a simultaneous TGA/DSC thermal analyzer [26]. They coupled the gas outlet of the TGA/DSC with a mass spectrometer via a heated capillary in order to acquire additional data. The catalytic activity of several transition and noble metal catalysts were screened using oxidation of carbon monoxide as a test reaction. TGA/DSC simultaneous analyzer allowed rapid sequential screening of about 70 catalysts per day and hence can be used as a powerful calorimetric tool in catalyst research revealing in-situ gravimetric data of the catalyst in addition to the heat of reaction.

1039

In short, DSC and DSC/TGA techniques are inevitable part in characterizing the catalysts and to study the stability, phase transformation, carbon deposition, deactivation behavior, kinetics, constraint porosities, etc. . ..

46.3

Calorimetry-Volumetry (Gas Adsorption Calorimetry)

Adsorption of probe or reactive gas molecules onto the catalyst surfaces can help us to gain more insights on the nature of gas-solid interactions and also give valuable information about the properties of the adsorbent surface. Thus gas adsorption calorimetry is of primary importance in the fields of catalysis. Adsorption is an exothermic process and when gas molecules interact with the solid surfaces heat is evolved that is measured by a calorimeter. The released heat is related to the bounding energy between the adsorbed species and the adsorbent and hence to the nature of the bonds and to the chemical reactivity of the solids surfaces. From a number of techniques used to study these gas-solid interactions, only a few provide information about the strength of chemisorption itself. The determination of the differential heats evolved by a suitable calorimetry-volumetry technique, when known amounts of gas probe molecules are adsorbed on catalytic surface is the most suitable and accurate method which allows the determination of the number, strength, and energy. This technique consists of a microcalorimeter coupled to a volumetric measurement device to measure the heat of adsorption and corresponding amount of reactants adsorbed onto solids/catalysts. It is widely used to study the gas-solid interactions during the adsorption of reactant molecules either discontinuously (pulse mode) or continuously. Gas adsorption calorimetry is the major calorimetry-volumetry technique employed in the detailed characterization and quantification of active sites in the catalysts. In gas adsorption calorimetry, gaseous reactants are allowed to interact with the solid catalyst surfaces and the heats evolved from the interactions are measured using calorimeter. Gases are more often introduced into the catalyst placed in the calorimeter cell in pulse mode, i.e., by successive doses of known amount of gases via volumetric setup. The schematics of the gas adsorption calorimetric apparatus along with the volumetric lines are shown in Fig. 46.9. The gas adsorption calorimetry-volumetry apparatus consists of mainly two measuring parts: (i) a volumetric line to measure the quantities of adsorbed gases and (ii) a differential micro-calorimeter to measure the heat flow. The tubes and valves (R0, R1, & R2) in volumetric line (of constant volume) help to dose a known volume of gas reactants and the differential gauge monitor the pressure of each dosage before and after adsorption (Fig. 46.9). Hence, for each dose of known volume, the quantity of gas adsorbed is calculated from the differences in

46

1040

L. Silvester et al.

Fig. 46.9 Schematic of a coupled calorimetry-volumetry apparatus [27]

R4

Pretreatment furnace 673 K – 1073 K Ionisation gauge

R2 Pump

Pressure gauge

R0

R1

R3

Trap (dry ice)

NH3

CO2

T = const. Calorimeter + Amplifier

monitored pressure and the corresponding heat evolved is recorded by differential micro-calorimeter containing a sample cell and a reference cell. Before each dosing, the volumetric line section and the line connected to calorimeter section are evacuated using a vacuum pump that allows attaining residual pressure as low as 0.13 mPa. During a gas adsorption calorimetric experiment, the catalyst is first heated at certain temperature at vacuum in the pre-treatment furnace connected to the system (Fig. 46.9). After pre-treatment, the sample cell is connected to calorimetric unit and the amount of gas is introduced sequentially by dosing onto the sample. Further, the sample cell is allowed to attain thermal equilibrium before introducing the next dose. The dosing is continued until the surface of the catalyst is completely covered by gaseous reactants and no considerable heat evolution observed with increase in pressure/dosage. After adsorption is complete, desorption is achieved by evacuation of the sample using vacuum pump and all the reversibly adsorbed gaseous reactants are removed. A second adsorption is performed to calculate the amount of reversibly and irreversibly adsorbed gas. The calorimetric and volumetric raw data are monitored and stored directly in a PC, such as differential heat and pressure. The raw data collected until the full coverage of the investigated solid surface will provide the information about the sample properties that can be expressed by different plots as shown in Fig. 46.10. Figure 46.10a shows the volumetric isotherms in which the amount of gas adsorbed is plotted as a function of equilibrium pressure (na ¼ f(P)) for adsorption cycle (I), followed by a desorption by pumping at

the same temperature and then re-adsorption cycle (II). The irreversibly adsorbed volume (Virr) that characterizes the strong sites of the catalyst can then be calculated as the difference between the adsorption volume and the re-adsorption volume at a given equilibrium pressure. The corresponding calorimetric isotherms (Qint ¼ f (P)) are shown in Fig. 46.10b. Since, both reversible and irreversible adsorption mechanisms are observed with most of the catalysts during the first adsorption and desorption cycle, a second adsorptiondesorption cycle is successively performed to determine the integral heats related to the irreversible and reversible processes. Depending on whether the reversible and irreversible adsorption processes occurs simultaneously or step-by-step, the points in the isotherms of first adsorption and desorption cycles may coincide or deviate at different pressure regions. However, the isotherms of second adsorption and desorption cycles coincide in all conditions as the curves from second cycle is the integral heats from only reversible adsorption. Hence, the integral heat related to reversible adsorption (Qint(rev)) can be directly obtained from the second adsorption/desorption isotherm and the integral heat corresponding to irreversible process (Qint(irr)) can be evaluated subtracting the heat values from first and second adsorption cycles. It is also possible to investigate the rate and kinetics of adsorption process from the normal or logarithmic plots of heat flow as a function of time. The plot of integral heats (Qint) as a function of the adsorbed quantities (na) is shown in Fig. 46.10c. This representation helps in the determination of coverage ranges

46

Calorimetry Techniques

Fig. 46.10 Data obtained from calorimetry-volumetry adsorption experiments

1041

a)

b)

na

Qint I

II

Virr

P

P

c)

d)

Qint

Qdiff

46 na

na

e) dna/dQdiff

Qdiff

with constant heat of adsorption, those for which the evolved heat is a linear function of coverage. Figure 46.10d shows the differential heat (Qdiff ¼ Qint/@ na), i.e., the molar heat of each dose of adsorbate, as a function of na. The ratio of the amount of heat evolved for each increment to the number of moles adsorbed in the same period is equal to the average value of the differential enthalpy of adsorption in the interval of the adsorbed quantity. The differential heat plots as a function of surface coverage is generally represented as histograms. Differential heats of chemisorption usually decrease with increase in the surface coverage. The way in which the profile of differential heat changes with increasing surface coverage depends on nature of adsorbate and adsorbent. For example, a large decrease in the heat even at low surface coverage indicates the surface heterogeneity of a solid catalyst. Finally, a plot of energy distribution spectra in which –dna/dqdiff is expressed as a function of qdiff is shown in Fig. 46.10e. The area under this curve represents the number of molecules that are

adsorbed with a given evolved heat. This kind of representation is not as accurate as the other aforementioned plots. Although it is not possible to determine the nature of adsorbed species, or even to distinguish between different kinds of adsorbed species from the calorimetric data, the variation of the differential heats of adsorption with coverage quite clearly shows energy distribution of surface active sites with respect to a given adsorbate and their varying reactivity on given adsorbents. Figure 46.11 shows some of the information that can be realized from kinetic data processing, notably the peaks of evolved heat and the evolution of pressure on the adsorbent during the adsorption. From the logarithmic scale plots of heat flow, both fast and slow adsorption phenomena can be clearly distinguished and the half reaction times corresponding to each increment can be calculated (Fig. 46.11a, b). When the evolution of adsorption heat stops, the heat flow peak becomes linear with time axis. However, in certain cases, a rapid release of heat followed

1042

b) Heat flow signal

a) Log (heat flow signal)

Fig. 46.11 Processing of kinetic data in adsorption calorimetry. Evolution of (a) logarithmic heat flow, (b) adsorption heat, (c) integral heat, and (d) adsorption gas pressure with the time

L. Silvester et al.

Slow

Fast

t

d)

c)

t1/2

t0

t

P

Qint

Pressure on adsorbate

100

t

by an exponential decrease is accompanied by a slow adsorption phenomenon that is detectable from the tail of the calorimetric curves (Fig. 46.11b). Evolution of the integral heat as a function of time (Qint, t) during the pulse adsorption by injection is an alternative way to represent the adsorption kinetics (Fig. 46.11c). If the evolution of heat during adsorption is comparatively slower than the response time of a calorimeter, it is possible to obtain kinetic data. The evolution of adsorbate gas pressure can be also used to study the adsorption kinetics (Fig. 46.11d). For example, if the time necessary to reach the equilibrium of adsorption in the gaseous phase (tequilibrium) and the time required for the corresponding calorimetric peak to return to baseline (t0) coincides then the kinetically prevailing process can be inferred as an activated chemisorption. However, if the value of tequilibrium is much lower than the value of t0, the slow kinetic phenomenon can be attributed to surface reactions following rapid adsorption. In order to gain better insight on the adsorption mechanism and the location of adsorbing catalytic surface sites, volumetric and calorimetric data can be combined with the thermokinetics study of the heat evolution. The shapes of individual thermograms help to identify the simultaneous processes exhibiting different kinetics during adsorption as well as the variation in the kinetics with surface coverage. A change in thermokinetic parameter (τ) can be used to monitor the kinetics of heat release during adsorption. During an adsorption, the calorimetric signal increases until the maximum of each adsorption peak and then decreases exponentially with the time of adsorption. This can be expressed in

tmax

tequilibrium

t

the form D ¼ Dmet/τ, where ‘D’ and ‘Dm’ are the deviation at time ‘t’ and the maximum deviation of a calorimetric signal, respectively. The thermokinetic parameter ‘τ’, also known as time constant, can be thus calculated from the inverse of the slope of a straight line obtained by plotting logD as a function of time. Variations of the thermokinetic parameter during an adsorption can also be plotted as a function of the amount of adsorbed probe. Having seen the different information that can be explored using gas adsorption calorimetry, we can further examine some examples of its usage in catalytic applications. The gas adsorption calorimetry is employed mainly to study the acidic and/or basic properties of the solids that have been used as acidic, basic, or multifunctional catalysts. While NH3 is mostly used as the probe gas to determine the acidic sites, CO2 or SO2 is often used to evaluate the basic sites on the catalyst surface. Some of the catalytic reactions that use the aforementioned gas adsorption calorimetric technique include selective oxidation [28, 29], oxidative coupling [30–32], dehydration [33–35], hydrolysis [33, 36], condensation [37–40], trans-esterification [41], and isomerization [42] reactions. Selective or partial oxidation of methanol to dimethoxymethane (DMM) required catalysts with adequate and balanced amount of both redox and acid sites. Hence, the ammonia adsorption microcalorimetry is employed to precisely evaluate number of acid sites, sites strength, and strength distribution of the acid sites on the catalysts capable of producing DMM. Zhao et al. reported that the sulfated binary titania-based catalysts exhibit good performance in

46

Calorimetry Techniques

1043

methanol oxidation to DMM [28, 29]. They employed heat flow calorimeter (C80 from Setaram) linked to a conventional volumetric apparatus having a Barocel capacitance manometer for pressure measurements. The differential heats of adsorption were measured as a function of coverage by repeatedly introducing small doses of ammonia gas onto the catalyst until an equilibrium pressure of about 66 Pa was reached. Authors found that the initial heats of ammonia adsorption on sulfated binary titania-based catalysts lies in the range of 160–190 kJ.mol1 suggesting considerable acidic behavior of catalysts (Fig. 46.12). In addition, the differential heats (Qdiff) exhibited a continuous decrease with ammonia coverage, revealing the heterogeneous acid strength distribution of the catalysts. NH3 adsorption isotherms show that equilibrium was not reached at low ammonia coverage on CrOx-titania catalysts suggesting a surface reaction between chromium oxide and ammonia (Fig. 46.12). Hence, ammonia adsorption microcalorimetry helped to evaluate the acidic behavior of the catalysts and to further correlate the acidic properties with reactivity of catalysts in selective oxidation of methanol. Ammonia adsorption calorimetry is often used to study the acidic behavior of catalysts employed in various dehydration reactions like fructose dehydration and glycerol dehydration [33, 34]. Vladislav et al. used Tian–Calvet type C80 microcalorimeter to evaluate the acidic properties of three different hierarchical zeolites (USY, Beta and ZSM-5) in fructose dehydration reaction [34]. They found that alkaline treatment in USY, Beta, and ZSM-5 zeolites not only created mesopores but also resulted in alterations in their acidic behavior. Though the amount of acidic sites was almost preserved after alkaline treatment, their distributions in Fig. 46.12 Differential heats of adsorption as a function of NH3 coverage on sulfated titania-based catalysts [29]

strength were modified. ZSM-5 zeolite proved to be the most resistant structure, in terms of relative acidity strength preservation, which is evident from NH3 calorimetric assessments. Ammonia adsorption measurements depicted large decrease of acid sites strength due to treatment in USY zeolite, which exhibited a substantial decrease of the concentration of acid sites in the 100–150 kJ.mol1 region of isotherm, accompanied by an increase in the concentration of weaker sites. 5-hydroxymethylfurfural selectivity was also found to be dependent on the acid sites strength of the investigated zeolites. Acrolein production by oxidative coupling of methanol and ethanol is currently a hot topic due to the industrial value of acrolein. The oxidative coupling proceeds through two steps: oxidation of methanol and ethanol to corresponding aldehydes over redox catalyst and aldolization of aldehydes to acrolein over acid-base catalysts. This reaction is performed either using a multifunctional catalyst in a single reactor or using different catalysts in two separate consecutive reactors. However, an acid-base catalyst has a very important role in improving the yield and selectivity toward acrolein. Hence, Lilić et al. employed NH3 and SO2 adsorption calorimetry techniques to understand the role of the acidity and basicity of the catalysts, respectively, in acrolein production and to clarify the reaction pathway [30–32]. Authors performed aldolization reaction over the silicasupported oxide, Mg-Al oxide, and La-Ce doped FeMo catalysts whose acid-base properties are evaluated by NH3 and/or SO2 adsorptions. NH3 was used as a basic molecule to quantify both Lewis- and Brønsted-acidic sites whereas SO2 titrated basic sites consisting mostly of surface oxygen ions. The volumetric isotherms and differential heats of NH3

Differential heat (kJ mol–1)

200

25CrTiS 25MnTiS 25FeTiS 25CoTiS 25MoTiS 25VTiS (data from ref. [14]) TiS

150

100

50

0 0

200

400

600

800

NH3 coverage (Pmol g–1)

1000

1200

1400

46

1044

a) NH3

600 Probe molecule uptake (Pmol g–1)

Fig. 46.13 (a) Volumetric isotherms & (b) differential heats of NH3 and SO2 adsorption as a function of surface coverage on different catalysts screened for acrolein production [30]

L. Silvester et al.

SO2

500

400

300

MgO Al2O3 MgO/Al2O3 CHT Mg/Al 0.06 CHT Mg/Al 0.3 CHT Mg/Al 0.5 CHT Mg/Al 3 ZrO2

200

100

b)

0.5

0.3

MgO Al2O3 MgO/Al2O3 CHT Mg/Al 0.06 CHT Mg/Al 0.3 CHT Mg/Al 0.5 CHT Mg/Al 3 ZrO2

0 0.1 0.1 Equilibrium pressure (Torr) 250

NH3 Differential heat (kJ mol–1)

0.7

0.3

0.5

0.7

SO2

200

150

100

50

0 600

400 200 Vmid NH3 (Pmol g–1)

and SO2 adsorptions on various catalysts used are shown as an example in Fig. 46.13a, b. Authors succeeded in finding a correlation between the catalyst reactivity quantified from aldolization reaction and the optimum base/acid site ratio, amount and strength of sites evaluated from adsorption calorimetry. They suggested that appropriate ratio of basic to acidic sites is required for the catalyst to show better activity in acrolein production. Also, presence of too high amount of strong basic sites drastically increases COx production. Basicity has an important role in the reactivity of catalysts employed in condensation reactions. CO2 adsorption calorimetry is one among the few techniques that help in

0

200 400 Vmid SO2 (Pmol g–1)

600

determining the amount, strength, and strength distribution of basic sites on catalysts used in condensation reactions. Leon et.al used CO2 adsorption microcalorimetry to study the correlation between the basic behavior of MgO based catalysts and their acetone condensation reactivity [37]. These authors used same calorimetric technique to study the influence of aluminum substitution in the hydrotalcitederived mixed oxide catalysts on their basicity and further demonstrate the relation between basicity and ethanol condensation products [40]. CO2 calorimetry also assists to study influence of rehydration and reconstruction of hydrotalcite catalysts on their basicity and condensation activity and

46

Calorimetry Techniques

further to deduct plausible reaction aldolization mechanism [39]. CO2 adsorption calorimetry has also been used to study the properties and reactivity of the hybrid catalysts in basic catalyzed reactions such trans-esterification and isomerization [41, 42]. Though temperature programmed desorption (TPD) of CO has been generally employed in determining the dispersion of active metal species in a catalyst, the reactivity of the catalyst depends also on the binding energy of the dispersed active sites. It has been known for a long time that an ideal catalyst should not have either too strong or too weak adsorption (or binding energy) for the reactant. Hence, CO adsorption calorimetry is the best tool to precisely determine the heat of adsorption of CO on active metal particles. In order to have some insight on the homogeneity/heterogeneity of the active sites on a catalyst, the CO adsorption microcalorimetry was conducted on dehydrated and reduced Cu–Y zeolites [43]. Zeolites with small alkali metal co-cations showed a decrease in heat of CO adsorption with increase in Cu content suggesting a transition from Cu+ monocarbonyl to dicarbonyl formation. Also, the heat of adsorption of CO is different for Cu+ ions localized at different sites in a zeolite structure. Hence, the kinetics of the CO adsorption on zeolites were correlated to the changes in the heat of CO adsorption resulted from their structural modifications. Adsorption calorimeter (Setaram C-80 II) is used to determine the CO adsorption heats related to Ru and Potassium promoted Ru graphite catalysts reduced at different temperatures and correlate the calorimetric results with Fischer-Tropsch reactivity [44]. Addition of potassium promotes more efficiently charge transfer to Ru and increase contribution of the M-C π bond (by larger π back-donation from the metal atoms to antibonding molecular orbitals of the CO molecule) resulting in higher CO adsorption heats. The CO adsorption heats of K promoted catalysts increases with increase in the reduction temperature. This is because the residual chloride anions on the catalysts, that neutralize the electronic promotion of potassium by potassium chloride formation, are progressively removed by increasing the reduction temperature and the effect of potassium promotion is revealed. Further, these K promoted catalysts with higher CO adsorption heats exhibited higher light olefins selectivity and lower methane production in the Fischer–Tropsch reaction. Tanksale et al. performed microcalorimetric experiments (Setaram BT2.15 calorimeter) using CO as probe molecule to determine the heat of adsorption of Ni and promoted Ni catalysts [45]. They suggested that the addition of noble metals to the Ni catalyst results in decreasing of initial heat of CO adsorption that may also avoids poisoning of the active metal sites. Hence, the promoted Ni catalysts with low binding energy of CO are expected to favor WGS reaction during the reforming reaction leading to higher H2 selectivity.

1045

46.4

Liquid Phase Calorimetry

As name implies, liquid phase calorimetry technique is a method in which the heat evolved from the liquid-solid or liquid-liquid interactions is measured while a liquid (a probe or a reactant) is usually allowed to interact with a solid or another liquid. The liquid titration calorimeter, liquid immersion calorimeter and reaction calorimeter are the three main liquid phase calorimetric techniques employed to determine the heat resulted from titration, adsorption, and reaction respectively.

46.4.1 Titration Calorimetry Liquid titration calorimetry is an isothermal calorimetry in which the surface sites of the solids, usually suspended in a solvent, is titrated using a specific liquid probe/indicator molecule. Though, liquid titration technique has some limitations in deducting gas-solid phenomena and surface Hammett acidity/basicity functions, it is still widely used to study the surface of solid catalysts with very strong surface acidity/ basicity. Moreover, this technique is employed to characterize the catalysts used in liquid phase reactions as it helps to study the variations in the effective acid-base properties of a catalyst in real test conditions, for example, at different solvent atmospheres in liquid phase reactions. Liquid titration calorimetry has also been used for determination of heats of adsorption of different adsorbates on adsorbents. Since the titration calorimeter is capable of measuring only the heat evolved during the adsorption/reaction in a liquid phase, it is often coupled with UV-Visible spectrometer to quantify the amount adsorbed on a solid corresponding to the measured heat. A scheme of the Tian-Calvet type differential heat flow calorimeter coupled to a UV-Visible spectrometer is shown in Fig. 46.14. A programmable syringe pump is used to introduce successive pulses of known amounts of probe molecule solution on the solid maintained at constant temperature during the experiment. The syringe pump is linked to the calorimeter by capillary tubes. The UV-Visible spectrometer, used to monitor the changes in concentration of the probe molecule, is coupled to calorimetric cell by an optic fiber. From the equilibrium concentration of the solution obtained after each dose of probe solution, adsorption isotherms can be constructed. The output from a typical titration microcalorimetric experiment is shown in Fig. 46.15. Each pulse injection of known amount of a solution of probe molecule using syringe pump gives a specific exothermic peak related to the heat of adsorption. This heat is a result of the exothermic enthalpy of adsorption and the endothermic enthalpy of displacement

46

1046

L. Silvester et al.

of adsorption. The enthalpy effects originating from dilution and mixing phenomena can be neglected due to the differential design of the calorimeter. The intensity of adsorption peaks decreases successively and reaches a plateau until all the available solid surface sites are titrated by the liquid probe.

Syringe pump Optic fiber

UV-Vis spectrometer

Magnetic stirrer

Fig. 46.14 Schematic of a TITRYS calorimeter connected to UV-Vis spectrometer (adapted from Setaram KEP technologies)

Fig. 46.15 The output exothermic peaks of titration calorimetric experiment for successive doses (adapted from Setaram KEP technologies)

Determination of Effective Acid-Base Properties For microcalorimetric determination of the strength of solid acids in solution, the usual probe molecules such as phenylethylamine, n-butylamine, aniline, etc are used to titrate acid sites [46–48]. In order to study effective basicity of solids using titration calorimetry, benzoic acid is generally used as a probe molecule [46]. Figure 46.16a shows the differential heats of benzoic acid adsorption in water as a function of the adsorbed amount (Na) of probe molecule for the different mesoporous carbon (MC) materials [46]. These differential heats represent the strength of the effective basic sites as the heat evolved in titration calorimetry is the sum of the enthalpy of displacement of the solvent and the enthalpy of adsorption. The amounts of benzoic acid adsorbed (Na) were obtained from the combined values of isotherms (Amount adsorbed vs Equilibrium concentration in solution) measured from UV-visible spectroscopy. Figure 46.16b shows the integral heats as a function of the amounts of benzoic acid introduced on different mesoporous carbon while n-decane or water is used as solvent. The integral heats obtained from benzoic acid adsorption in n-decane are lower than in water which indicating the influence of solvent on the heat evolved and hence the amount of effective basic sites on different carbon solids. Similarly, phenylethylamine is used as a probe to perform calorimetric titration of the acid sites on mesoporous carbon [46].

Heat flow m (mW) Exo 5

4

3

–1.560 J

–0.986 J –0.914 J

–0.904 J –0.880 J

2

0.851 J

–0.822 J –0.774 J –0.659 J

1

0 –1.795 J –1.063 J –0.927 J –0.900 J –0.880 J –0.866 J –0.834 J –0.768 J –0.681 J 2 4 6 8 10 12 14 16 18 Time (h)

Calorimetry Techniques

1047

Differential heat (kJ/mol)

a) 80

70

Activated carbon

MC

70

Zeolite beta, Si/Al = 43

N-MC

60

C3N4-MC

50 40 30 20 10

Differential heats (kJ mol–1)

46

60

Zeolite beta, Si/Al = 12.5

50 40 30 20

0 0.0000

0.0005

0.0010

0.0015

Na (mol/g)

0.0000

0.0002

0.0004 Na (mol

0.0006

0.0008

0.0010

g–1)

b) 70 Fig. 46.17 Differential heats as a function of amount of nicotine adsorbed on different solids [49]

N-MC in water 60

MC in water

Qint (J/g)

50

C3N4-MC in water

40 N-MC in decane 30 C3N4-MC in decane

20

MC in decane 10 0 0

1000 2000 3000 Benzoic acid introduced (Pmol/g)

4000

Fig. 46.16 (a) Differential heat as a function of amount of benzoic acid adsorbed and (b) the integral heats as a function of the amounts of benzoic acid introduced while titrating mesoporous carbons [46]

Determination of Heat of Adsorption It is widely recognized that adsorption processes provide a feasible method for the removal of pollutants and heavy metals from the environment. Since, the adsorption involves heat that is a representative of combined physical, chemical, and sometime biological processes, a microcalorimetric technique can help to determine the thermodynamic characteristics of the adsorption of the pollutants/heavy metals in a solution on a solid surface. Rakic et al. used titration calorimeter (TITRYS, from Setaram) to measure the heats of nicotine adsorption on different zeolites and activated carbon in aqueous phase [49]. The decreasing trend in the differential heats revealed as a function of the amounts of nicotine adsorbed (Na) indicates interaction of the nicotine molecules with heterogeneous solid surfaces (Fig. 46.17). Based on the calorimetric results, the nicotine adsorption capacities of different solids could be identified. Yan et.al used titration calorimetric technique (ITC-200 calorimeter; MicroCal Co., USA) to investigate the thermodynamics in the sorption behavior of the heavy metals on

extracellular polymeric substance (EPS) during the waste water treatment [50]. Authors succeeded to determine the stoichiometric number, binding constant, enthalpy, and entropy of binding from the calorimetric data. Trivalent cations were more competitive than divalent cations for heavy metal ion binding because they formed complexing bonds. Results of this study provide detailed insight into the interaction mechanism between EPS and heavy metal and hence in the heavy metal removal from waste water.

46.4.2 Immersion Calorimetry When a solid is immersed into a non-reacting liquid, a certain amount of heat is evolved. The enthalpy of immersion (ΔHimm) is defined as the enthalpy change when a solid is immersed into a wetting liquid in which it does not dissolve or react. Immersion calorimetry is a liquid phase calorimetric technique that measures the enthalpy of immersion. Enthalpies of immersion of a given solid into different liquids are usually different because they are also related to the specific interaction between the solid surface and the immersion liquid in addition to the available surface area. Immersion calorimetry is important in catalysis as it is capable of determining some of the important parameters of catalysts such as surface hydophilicity/hydrophobicity [51], oxygen content [51, 52], micropore area and size [53, 54], catalyst surface – reactivity correlations [55], and surface acidity/ basicity [56]. Nowadays, immersion calorimetry experiments are generally carried out in Tian-Calvet type calorimeter (e.g., C80D Setaram). A simple scheme of immersion calorimeter is shown in Fig. 46.18. In a typical experiment, weighed amount of sample is taken in a glass bulb with a tiny brittle end that is heated and out-gassed under very low pressure of

46

1048

L. Silvester et al.

Fig. 46.18 Schematic of a basic immersion calorimeter

Isothermal jacket Thermopiles Solid Bulb Liquid

around 105 torr. The bulb is then sealed and attached to a stainless steel rod through a closed stainless steel fitting. This rod-bulb assembly is introduced into the calorimeter cell containing about 3–5 mL of the wetting liquid, and carefully sealed. Then, the whole system is introduced inside the calorimeter block and allowed to attain temperature equilibration between the sample set-up and the calorimeter. After reaching thermal equilibrium, the steel rod is gently pressed down to break the brittle tip of the bulb hence allowing the liquid to enter the bulb and wet the solid. The corrected enthalpy of immersion(ΔHimm) is calculated by integrating the calorimetric signal, after eliminating the value of heats due to (i) the breaking of the tip (exothermic) and (ii) evaporation of the immersion liquid necessary to fill the empty volume of bulb with the vapor (endothermic). Both the aforementioned thermal effects were previously calibrated by performing blank experiments with empty glass bulbs of different volumes. Barroso-Bogeat et al. used immersion calorimetry to perform physico-chemical characterization of activated carbon–metal oxide (AC-MO) photocatalysts [51]. The results of immersion calorimetry for the AC-MO catalysts provide valuable information concerning the dependence of the hydrophobicity and hydrophilicity of the samples on their preparation method. Total surface area of the catalysts accessible to benzene and water can be calculated from their immersion enthalpies by following expressions:

Stot ðC6 H6 Þ ¼

ΔHimm ðC6 H6 Þ 0:114

Stot ðH2 OÞ ¼

ΔH imm ðH2 OÞ 0:114

ð46:10Þ

ð46:11Þ

where ΔHimm(C6H6) and ΔHimm(H2O) are the immersion enthalpies into benzene and water and respective total surface area [Stot (C6H6)] and [Stot (H2O)] of the catalysts. Also immersion calorimetry into benzene and water allowed estimating the total content of surface oxygen for the solids as follows:   0:21  ΔH ðC H Þ  ΔH ðH OÞ 1 imm 6 6 imm 2 ½O mmol:g ¼ 10 ð46:12Þ Though not widely studied, immersion calorimetry has also been used to characterize the surface acidic and basic groups on solids. Lopez-Ramon et al. obtained a correlation between the net enthalpy of neutralization (difference of immersion enthalpies of NaOH and H2O) and the number of acidic sites (in terms of NaOH titration) for different carbon solids (Fig. 46.19) [56]. In a recent study, the immersion calorimetry into liquids of different kinetic diameter was employed to evaluate the textural properties (pore size distribution and pore shape) in

46

Calorimetry Techniques

1049

Fig. 46.19 Correlation between the net enthalpy of neutralization by NaOH and the number of acidic sites obtained from NaOH titration [56]

140 120 –'iH(NaOH)net, (J g–1)

AZ46-0 AZ46-24

100 80

AZ46-5 BV46-OX

AZ46-3

60 40

UO2-OX KF-1500

20 0

AZ46-0 BV-46H25-OX H-25 0.0

0.5

1.0

1.5

2.0

meq NaOH

2.5

3.0

g–1

46 ultra-microporous carbon materials [52]. Vernimmen et al. used immersion calorimetry to evaluate the catalytic performance of titanosilicate catalysts in the epoxidation of cyclohexene [55]. For the first time, immersion calorimetry, with the same substrate molecule as in the catalytic test reaction, is used as an extra tool to deduce the catalytic performance. A good correlation is obtained between the catalytic results and immersion calorimetry data. Hence, immersion calorimetry can be used as an auxiliary characterization technique to have a better insight on the interaction between the solid catalyst and the substrate in the fields of heterogeneous catalysis.

46.4.3 Reaction Calorimetry Unlike most of the calorimeters that are mainly used to measure the heats (directly or indirectly) associated to solid catalysts, the reaction calorimetry measures the heat released during the catalytic reactions. Hence, it is capable of directly providing the information regarding the thermochemistry and the kinetics of a chemical reaction that are crucial parameters in thermal process safety and process development. Reaction calorimetry represents a unique technique to gather information about both the chemical and the physical processes during reactions that are accompanied by heat effects. Also, reaction calorimetry represents a differential kinetic analysis method, as the rate of heat flow is proportional to the conversion rate (mol.s1) for most chemical reactions that can be expressed by following mathematical expression:

q_ React ðtÞ  r ðtÞV r

ð46:13Þ

where, ‘q_ React ’ is the reaction heat-flow rate determined by a calorimeter, ‘r’ the rate of reaction (mol.m3.s1), and ‘Vr ’ the reaction volume (m3). All the above-mentioned variables are a function of time and vary as the reaction progresses [57]. Though, the isothermal reaction calorimetry is mostly preferred for scale-up of a desired synthesis reaction including kinetic and thermodynamic analysis, the non-isothermal calorimetry experiments are also performed during safety analysis in order to investigate undesired decomposition reactions. However, non-isothermal experiments contain information about the temperature-dependent chemical reaction system that is obviously larger compared to an isothermal measurement. Hence, non-isothermal calorimetric measurements are possible only if sophisticated evaluation methods were available as the information density for the complex non-isothermal reactions is often too large to evaluate using the common analysis methods. Hence, the heat flow reaction calorimeters that work either in the isothermal conditions (e.g., RC1e from Mettler) or under the isoperibolic conditions (e.g., differential reaction calorimeter (DRC) from Setaram) will be discussed. The schematic of an isothermal reaction calorimeter is shown in Fig. 46.20a. In an isothermal reaction calorimeter, the temperature of the reaction mixture (Tr) is kept constant throughout the reaction by cooling or heating the fluid in the reactor jacket. Whereas, in the isoperibolic reaction calorimetry the temperature of the thermostat jacket fluid (Tj) is kept constant, leaving the reaction mixture to evolve. The heat flow rate (q_ flow ) through a reactor wall of an isothermal

1050

L. Silvester et al.

Fig. 46.20 Schematics of (a) isothermal reaction calorimeter and (b) isoperibolic DRC (adapted from Setaram KEP technologies)

a)

Cooling liquid Calibration / compensation heater Tj,OUT

Tr q·Flow Tj,IN

Tj

Reactor content Reactor jacket

b)

Dosing funnel

Motor

Motor Calibration probe

TDOS: Temperature of introduced reagent

Temperatures and 'T measurements Reagent

Optional probe

Reaction medium

Thermostated jacket

Reference Thermostat

reaction calorimeter can be expressed by following steady state equation:   q_ flow ¼ UA T r  T j

ð46:14Þ

Thermostated fluid

where ‘A’ is the total heat-transfer area (m2) and ‘U’ is the overall heat-transfer coefficient (W.m2.K). Parameters ‘U’ and ‘A’ generally vary during the measurement of chemical reaction as they depend on physical properties of the reaction

46

Calorimetry Techniques

mixture/jacket fluid and reactant mixture volume that change during a reaction measurement. Hence, these two parameters have to be calibrated often by means of a calibration heater (see Fig. 46.20a) before and after a reaction experiment as such a calibration is not possible during the course of a reaction.

Isothermal Reaction Calorimeters Almost all reaction calorimeters consist of a reaction vessel and a surrounding jacket filled with circulating fluid that helps to carry heat away from the reactor (Fig. 46.20a). According to the measurement and control principles associated with the reaction calorimeters, they are classified into following four categories: Heat-Flow Reaction Calorimeter Most of the commercially used reaction calorimeters are based on the heat-flow principle in which the temperature of the cooling liquid inside the jacket (Tj) will be adjusted to control the temperature of the reactor content (Tr) (Fig. 46.20a). The heat-flow rate (q_ Flow ) between the reactor content and the cooling liquid is determined by measuring the temperature difference between them. The temperature signal is then converted into a heat flow signal using a heat-transfer coefficient value previously determined using a calibration heater. The flow rate of the cooling liquid circulating through jacket should be high enough to have a fast control of the ‘Tr ’. Heat-Balance Reaction Calorimeter The heat-balance principle was first introduced by Meeks [58]. In a heat-balance calorimeter, temperature of the reactor content (Tr) is maintained by adjusting the cooling liquid temperature (Tj). The heat-flow rate (q_ Flow) is then determined by measuring the difference between inlet temperature (Tj, IN) and outlet temperature (Tj, OUT) of the calorimeter jacket and the mass flow of the cooling liquid (Fig. 46.20a). Heat capacity of the cooling liquid and the heat-flow signal is determined directly without any prior calibration. Power-Compensation Reaction Calorimeter The power-compensation principle was first introduced by H. M. Andersen [59, 60]. In the power-compensation calorimetry, a compensation heating device is inserted directly into the reactor content whose power can be adjusted to control the temperature of the reactor content (Tr) (Fig. 46.20a). As the electrical heaters are not generally designed for cooling purpose, the compensation heaters serve to maintaining a constant temperature difference between the reactor jacket and the reactor content. Thus, “cooling” of the reactor content is always achieved by reducing the power of the compensation heater. The heat flow rate is determined directly from the power consumption of the

1051

compensation heater. An external cryostat is employed to maintain constant temperature of the cooling liquid (Tj). Peltier Reaction Calorimeter First reaction calorimeter using Peltier elements was designed by F. Becker [61, 62]. The temperature of the reactor content (Tr) in a Peltier reaction calorimeter is controlled by varying the power of the Peltier elements. Unlike the power-compensation calorimeter in which cooling cannot be achieved by electrical heating device, the Peltier elements in Peltier reaction calorimeter are capable to perform both the cooling and the heating. The heat flow occurs through the Peltier element and heat flow rate (q_ Flow ) is calculated based on the required electrical power and the measured temperature gradient over the Peltier elements. Similar to power compensated calorimeter, an external cryostat is used to control the temperature of the cooling liquid (Tj). The N-oxidation of alkylpyridines has been widely employed as a model reaction for the development of calorimetric methods for studying complexity involved in highly exothermic reactions. Wang et al. used an isothermal reaction calorimeter (RC1e Mettler-Toledo) to deduce a model that can be used for inherently safer reactor design in homogeneous tungstic acid catalytic hydrogen peroxide oxidation processes [63, 64]. The calorimeter consisted of a 1.2 L pressurized glass reactor, an anchor stirrer, temperature sensor, calibration heater, and FTIR probe. Reaction kinetics data of the N-oxidation reaction and the competing H2O2 decomposition reaction were investigated over a wide range of catalyst concentration, temperature, and H2O2 feeding rate condition using isothermal reaction calorimeter. Under the given conditions of the calorimetric measurements, reaction protocols including the amount of catalysts, the reaction temperature and the dosing rate to reaction mixture were optimized to obtain a maximum product yield. Further, a comprehensive kinetics model was developed accounting for the catalyst solubility effect and the dynamic reversible catalyst activation/deactivation that fits very well with calorimetric experimental data. Reaction calorimeters have also been used to evaluate the process safety in industrial reactions. In order to identify the inherent hazards of a highly exothermic Ritter reaction, a process safety evaluation was performed using reaction calorimeter (RCe1 Mettler Toledo) under adiabatic and isothermal conditions [65]. Based on experimental results using adiabatic and isothermal calorimeters, lot-wise addition of acetonitrile to a mixture of 1,3 dimethyladamantane and sulfuric acid at ~38  C is recommended to perform controlled and safer Ritter reaction. Though reaction calorimeters have been widely used in isothermal conditions, the performance of a reaction calorimeter was successfully demonstrated for the thermo-kinetic reaction analysis of n-butanol esterification under non-isothermal conditions [66]. Reaction

46

1052

L. Silvester et al.

calorimeter can be also used to precisely determine the specific heat capacities of liquids that are used as solvents in different reactions [67].

Isoperibolic Differential Reaction Calorimeter For weakly exo- or endo- thermal reactions, the temperature change of ‘Tr ’ in an isoperibolic reaction calorimeter is rather small and comparable to an isothermal reaction calorimeter. Moritz and co-workers introduced an isoperibolic reaction calorimeter that is inserted into a compensation heater controlled heat-flow device [68]. The varying heat-transfer coefficient during a reaction measurement in isoperibolic calorimeter does not influence the calculated heat flow rate of that particular reaction. Later, Setaram developed a completely new design for an isoperibolic differential reaction calorimeter (DRC). The schematic of a new isoperibolic differential reaction calorimeter is shown in Fig. 46.20b. In a DRC, the working reactor and the reference reactor are double-enveloped spherical reactors connected in parallel. The reactor surroundings of both reactors are maintained at constant temperature with 0.01  C stability by a continuous flow of a heat-carrier fluid through the thermostat jackets. DRC continuously monitor a temperature difference (ΔT ) between the working reactor and the reference reactor (Fig. 46.20b). Reactions are performed inside the working reactor, whereas the reference reactor is filled with a solvent that possesses similar physical properties as that of the reaction mixture. All phenomena such as the temperature fluctuations of heat carrier, the heat losses through top of the reactor, and the heat related to the stirring that interfere with the actual heat of reaction can be corrected by using the reference reactor. Hence, presence of the reference reactor in DRC enables the correction of the perturbations of the system and determines the actual heat of reaction and heat capacity. The overall heat balance in a DRC can be expressed as follows:   qr ¼ UAðT R1 T R2 Þ þ mR CpR  CpiÞ 

dðT R1  T R2 Þ _ pdos ðT R1 T dos Þ þ mC dt

ð46:15Þ

If the heat transfer coefficient (U ), heat transfer area (A), the heat capacities of the reactor contents (mRcpR), and the heat capacities of the inserts (cpi) are known, the heat flow rate of the reaction can be measured from the temperature difference (TR1  TR2) between both reactors. For a reaction performed at specific conditions, the values of the aforementioned parameters (U, A, mRcpR, cpi) are determined prior by a calibration that is generally performed using a joule effect probe at same conditions but in the absence of any feed/ reaction. The term ‘mcpdos (TR1  Tdos)’ corresponds to the

heat evolved if the temperature of the reaction feed is different from that of the reaction mixture. Calibration of the differential reaction calorimeter is performed using a Joule effect probe made up of a special alloy. The power can be set up to 10 W during calibration (qc). If the reactor is not fed and no reaction occurs during the calibration, the heat balance equation becomes:   qc ¼ UAðT R1 T R2 Þ þ mR CpR  CpiÞ 

dðT R1  T R2 Þ dt

ð46:16Þ

The heat-transfer characteristics can be obtained by integration of the calorimetric signal over time: 1 ð

qc dt UA ¼ ð 1 0

0

ðT R1  T R2 Þdt

ð46:17Þ

After switching off the calibration heater, the specific heat capacity of the reactor’s contents can be determined by evaluating the thermal relaxation. The integration of Eq. 46.16 gives: q ðT R1  T R2 ÞðtÞ ¼ ðT R1  T R2 Þðt0Þ þ c UA  ðtt0 Þ  1 e τ where time constant τ ¼

mr CpR þCpi UA

ð46:18Þ

; Cpi is the heat capacity of

the inserts whose value is already determined using a solvent with known specific heat capacity. Since DRC is operated as a differential calorimeter, it is important to maintain the symmetry of both measure and reference reactors during the reactions and hence the questions regarding the loss of symmetry arise. As in Eq. 46.15, the calorific sensitivity (UA) values in both measure and reference reactors is assumed to be same. However, this term may vary during a reaction either due to the change of the heat-exchange area of the measure reactor resulting from addition of a reactant in semi-batch operation or due to a viscosity change in the measure reactor resulting from the chemical reaction. The possible influence of volume increase and viscosity change was investigated by the addition of water and poly(ethyleneglycol) [69]. The results obtained by DRC are within the commonly accepted tolerance of 15% hence confirming that the approximation made in Eq. 46.15 is valid for this calorimeter. Although, DRC is an isoperibolic differential calorimeter commercialized by Setaram around two decades ago, not

46

Calorimetry Techniques

1053

many works have been conducted with this system. However, some classical reactions performed in DRC demonstrated that results can be obtained more rapidly and low cost (smaller amount of raw materials) compared to normal reaction calorimeters [69, 70]. DRC enabled to accurately obtain the enthalpy of reactions of acetic anhydride hydrolysis and esterification of propanoic anhydride by butanol [70]. Hence, DRC allows to measure the enthalpy of a very fast reaction (acetic anhydride hydrolysis) and to obtain reliable results even for a slow reaction (esterification) that requires a good thermal sensitivity to distinguish the low heat release rate at the end of experiment from the measurement noise. The dissolution process of natural phosphate solids in acid mixtures and the influence of parameters like solution concentration, reaction temperature and solid-liquid ratio, particle size and stirring speed on the dissolution rate were investigated using DRC [71, 72]. Heats evolved during the NaBH4 hydrolysis and hydrogen production catalyzed by different metals are investigated using DRC (Setaram) coupled with hydrogen flow meter [73, 74]. Figure 46.21a presents the differential temperature evolution recorded for the addition of NaBH4 solution on the platinum catalyst. Heat evolved during the hydrolysis reaction decreases continuously, indicating the deactivation due to a self-poisoning of the Pt catalyst during the reaction [73]. On the other hand the differential heat due to NaBH4 hydrolysis increases and

Differential temperature (°C)

a) 3.0 2.5 2.0 1.5 1.0 0.5 0.0

0

3000

6000

–0.5

9000

12,000

Time (s)

Differential temperature (°C)

b) 0.9 0.6 0.3 0.0

0

3000

6000

9000

12,000

15,000

18,000

–0.3 –0.6

Time (s)

Fig. 46.21 Evolution of the differential temperature versus time for the addition of NaBH4 solution on to (a) Pt-based & (b) Ni-based catalysts [73]

reaches a plateau for Ni based catalysts indicating the redox mechanism of the surface catalyst particles and higher stability of Ni catalyst compared to Pt (Fig. 46.21b). The catalyst performance of Co nanoparticles in NaBH4 hydrolysis investigated using the “multi-addition” calorimetric method enabled determining the energy associated with the formation of surface boride species and activity of the catalyst [74].

46.5

Single Crystal Adsorption Calorimetry (SCAC)

Though, the temperature programmed desorption (TPD) and the equilibrium adsorption isotherm analyses have been applied to determine the heats of adsorption of molecules on single crystal surfaces, they both have a limitation that the adsorbate molecule must adsorb and then desorb reversibly without undergoing any partial dissociation. Three decades ago, David King’s group developed first single crystal adsorption microcalorimeter that is capable of measuring the heat of adsorption on single crystal surfaces even if the molecules adsorb irreversibly and undergo partial dissociation [75, 76]. Furthermore, Charles T. Campbell and team developed a highly advanced single crystal calorimeter that is widely used in catalytic applications in present days [77–79]. It uses a novel method of heat detection which has several advantages over previous designs considering the abilities to easily prepare samples, use thicker samples, use broad range of substrates, and work at much lower temperatures. Also, accurate calorimetric measurements of heats of adsorption on clean single crystals can be achieved at room temperature with the help of a unique molecular beam and method for correction of absorbed radiation from its hot effusion cell. This SCAC is capable of measuring the heat of adsorption associated with the larger molecules like benzene, ethylene, etc. . . possessing low vapor pressure in addition to the lighter molecules like CO, H2, O2, CO2 etc. . . [78]. The general schematic of the advanced single crystal calorimeter is shown in Fig. 46.22 [79]. The SCAC unit is separated into a preparation chamber (1) and an adsorption/ reaction chamber (2) because of spatial limitations of the ultra-high vacuum (UHV) chamber where calorimetric detector is located. Both chambers are separated by a gate valve (3) and manipulators and a translational rod (4) are used to transfer the sample between the chambers and the microcalorimeter (5). Physical vapor deposition is used to prepare a complex model catalyst which is realized by two electron beam evaporators (9, 10). The evaporator fluxes are calibrated using a quartz crystal microbalance (QCM) (11). A differentially pumped ion gun (6) and a low energy electron diffraction (LEED)/Auger electron spectroscopy (AES) system (7) are employed aiming to clean and characterize the single crystal and supported catalysts. In order to reduce the

46

1054

L. Silvester et al.

background pressure rise during preparation, a gas dosing system (8) was implemented. The sample can be cooled to 90 K using liquid nitrogen and heated up to 1300 K. Sample is heated by electron bombardment from a flat filament of a commercial halogen bulb. A pyrometer (12) is used to measure temperature during sample preparation in the range between 450 and 1300 K. Temperature-programmed desorption (TPD) can be performed in the preparation chamber and analyzed by means of a mass spectrometer (13). The adsorption/reaction chamber is equipped with quadrupole mass spectrometer (QMS) and wobble-stick to perform sticking probability measurements simultaneously with microcalorimetry. Like all other calorimetric techniques, SCAC also required prior calibration. Since SCAC use a molecular beam, absolute energy calibration of the SCAC beam is performed using light pulses from a He–Ne laser [79]. In order to realize calibration, the laser and the molecular beam should have the same spatial and temporal profiles. Figure 46.23 shows a schematic representation of the energy calibration unit that is installed in the SCAC experimental

setup (see Fig. 46.22). Using a lens system (1), the laser light is spread and directed to one of six optical neutral density filters mounted in a rotatable wheel (2). This rotatable wheel permits laser power attenuation with a desired % transmission value. The attenuated laser beam is then passed through a window (3) into the molecular beam source, where it falls onto a prism (4) that is temporarily placed directly in the beam path. The prism reflects laser light down the molecular beam path and splits the laser beam into pulses by a beam chopper (5) in the same way as the molecular beam (Fig. 46.23). Hence, it is possible to compare the detector response from a real adsorption measurement to the calorimetric signal obtained from the heat by laser pulses, provided that the full heat deposition take place on a time scale shorter than the characteristic detection time of SCAC. The overall accuracy of the adsorption energy (kJ per mole adsorbed) measured with SCAC is in the range of ~7–9% for a single measurement adsorption energy versus surface coverage using model oxide-supported catalysts and high sticking coefficient gases that do not stick on the chamber walls. In the adsorption energy measurement, the main source of the

(13): Mass spectrometer (6): Ion gun Wobble stick

Window for reflectivity setup

(5): Microcalorimeter

(8): Gas doser

(11): QCM (4): Magnetic transfer rod

QMS Heating

(3): Gate valve

Molecular beam (2): Adsorption/reaction chamber

(9,10): Two evaporators and (12): Window for pyrometer

(7): LEED AES

(1): Preparation chamber

Fig. 46.22 Schematic overview of the experimental setup: (1) preparation chamber, (2) adsorption/reaction chamber, (3) gate valve, (4) magnetic transfer rod, (5) microcalorimeter, (6) ion gun, (7) optics for LEED

and AES, (8) gas doser, (9) and (10) two metal evaporators, (11) QCM, (12) port for pyrometer, and (13) mass spectrometer [79]

(2): Filter wheel

(4): Prism

Shutter

(5): Chopper

Aperture HeNe laser (1): Lens system

Mirror

(3): Window

Beam aperture Inside the molecular beam source

Fig. 46.23 Scheme of the laser calibration system consisting of components: (1) lens system, (2) wheel carrying the neutral density filters, (3) chamber window, (4) prism, and (5) chopper [79]

Calorimetry Techniques

1055

errors is to exactly determine the laser energy absorbed by the sample per pulse, which depends on several factors such as the photodiode sensitivity, the pulse length, and the sample reflectivity. Palladium nanoparticles are of great interest as catalytic materials and particle size of the Pd nanoparticles is one of the most important parameters that determine their catalytic performance. Single crystal adsorption micro-calorimetry enables to draw a correlation between the particle size and the catalytic activity by measuring the heat evolved by the adsorption of reactant molecules. For example, FischerWolfarth used SCAC to investigate the influence of particle size on the adsorption heat of carbon monoxide on welldefined Pd nanoclusters ranging from 1.8 to 8 nm and on the extended Pd_111_surface [80]. The CO binding energy decreases by about 20–40 kJ.mol1 with decreasing the Pd particle size (Fig. 46.24). The reason for such decrease in the enthalpy of CO adsorption is possibly due to weak chemisorptive interaction on Pd cluster with contracted lattice parameter and reduced van der Waals attraction from less polarizable small Pd particles. These findings can be of great help in further understanding the catalytic performance of these Pd nanoparticles in various reactions. The same authors reported similar trends in the heat of adsorption of CO on Pt (111) as a function of CO surface coverage at 130 K and 300 K [79]. Single crystal adsorption microcalorimetry is also used to study the thermodynamic stability and energetics of adsorbed intermediates that are critical parameters of a catalyst selectivity and activity. These parameters are generally determined by density functional theory (DFT) calculations without much experimental support. Carey et.al used SCAC to Fig. 46.24 Adsorption heat of CO on Pd_111_ and on Pd/Fe3O4 / Pt_111_ model catalyst plotted as a function of CO surface coverage at 300 K [80]

investigate the energetics of adsorbed methyl and methyl Iodide on Ni(111) [81]. Adsorbed methyl (CH3(ad)), the simplest example of all alkyl intermediates, is formed during catalytic reactions such as combustion, partial oxidation, reforming, methanation, Fischer–Tropsch, and several fuel cell reactions. Comparison study of the energetics for methyl obtained from SCAC and DFT calculations showed that DFT method routinely underestimates the bond energy of methyl. They found that the bond energy of CH3-Ni(111) is stronger than the reported value for CH3-Pt(111) bonds and hence explains the greater activity of Ni catalysts for hydrogenolysis in comparison to Pt catalysts. Similarly, Zhao et al. reported SCAC investigation of the heat of formation for adsorbed formate that is considered as an important intermediate on catalysts surfaces for many catalytic reactions [82]. They directly measured the heat of molecular adsorption and the dissociative adsorption of formic acid monomers onto Ni(111) at different temperatures (120–240 K) by SCAC and extracted the enthalpy of formation and the bonding enthalpy of adsorbed formate on Ni (111). At 240 K the heat of adsorption is more exothermic than heat of adsorptions at lower temperatures (210 K & 155 K) in calorimetric experiments due to the heat of dissociative adsorption of formic acid, leading to adsorbed bidentate formate and hydrogen species (Fig. 46.25). They suggested that intermediate heat of adsorption observed at 210 K is perhaps due to the dissociative adsorption making monodentate formate and adsorbed hydrogen or due to the formation of a co-adsorbed mixture of bidentate formate and molecularly adsorbed formic acid (Fig. 46.25). The lower heat of adsorption at 155 K and 120 K is because formic acid molecularly adsorbs on Ni(111) and does not undergo

180

1.8 Pd(111) 7 Å Pd/Fe3O4 0.3 Å Pd/Fe3O4

Heat of adsorption (kJ mol–1)

160 140 120

1.6 1.4 1.2

100

1.0

80

0.8

60

0.6

40

0.4

20

0.2

0

0 0

0.15

0.3

0.45

0.6

Number of adsorbed CO molecules (cm–2)

0.75×107

Heat of adsorption (eV)

46

46

L. Silvester et al.

Differential heat of adsorption (kJ/mol)

1056

160 140

240 K HCOOH/Ni(111)

120 100

210 K

80 120 K 60

155 K

40 0.0

49.9 ± 2.0 kJ/mol 0.1

0.2

0.3 0.4 1.6 1.7 1.8 1.9 2.0 Coverage (ML)

Fig. 46.25 Heats of adsorption of HCOOH on clean Ni(111) at 120, 155, 210, and 240 K as a function of total formic acid coverage that permanently adsorbed [82]

any dissociation. Also, the experimental bond energy determined by calorimeter is 25.6–57.6 kJ.mol1 greater than those found by DFT methods indicating the fact that the DFT calculations systematically underestimated the bond strength of formate to Ni(111). For the first time, Silbaugh et al. reported an experimentally determined energetics and mechanistic insights of oxidation of methanol and formic acid on Pt(111) using single crystal adsorption calorimetry [83]. Though energies of all the adsorbed intermediates and activation energies of all the elementary steps have been estimated for methanol and formic acid oxidation reactions on Pt(111) by density functional theory (DFT), no experimental measurements to validate these energy diagrams and mechanisms have been conducted previously. They suggested a new reaction pathway that involves the key intermediate monodentate formate species while bidentate formate only being a secondary species that slows the reaction rate.

46.6

Conclusions

This chapter was intended to present the basic principles of different calorimetric techniques including experimental setups and their applications during the last one decade in the field of catalysis. The operation modes of four main calorimetry techniques: (i) Differential scanning calorimetry (DSC), (ii) Calorimetry-Volumetry (Gas adsorption calorimetry), (iii) Liquid phase calorimetry, and (iv) Single crystal adsorption calorimetry were detailed along with the examples of their catalytic applications. DSC’s are one of the widely used calorimetry due to its simplicity, rapid measurement ability, comparatively cheaper price, and availability. Most often, a DSC is coupled to

thermogravimetry (TGA) to gain more detailed mechanisms on the catalyst evolution during reactions. DSC or DSC/TGA systems are used to study the thermal stability, carbon deposition, activation energy, enthalpy, and kinetics associated with different catalysts and catalytic reactions. A calorimetry-volumetry technique consists of a calorimeter coupled to a volumetric measurement device enables to simultaneously measure the heat of adsorption and corresponding amount of reactants adsorbed onto solids/catalysts. Gas adsorption calorimetry is the main calorimetryvolumetry technique employed to study the gas-solid interactions during the adsorption of reactant molecules either discontinuously (pulse mode) or continuously. Hence, adsorption of probe molecules like NH3, CO2, SO2, CO, etc. is used in gas adsorption calorimetry for detailed characterization and quantification of the catalytically active sites (acidic, basic and redox sites). In a liquid phase calorimetric technique, the heat evolved from the liquid-solid or liquid-liquid interactions is measured. There are three widely used liquid phase calorimetric techniques: liquid titration calorimeter, liquid immersion calorimeter, and reaction calorimeter that measure the heats resulted from titration, adsorption, and reaction, respectively. Titration calorimetry uses probe molecules like benzoic acid, phenylethylamine, etc. to titrate the acidic and basic sites in a catalyst placed in a solution. It can also measure the heats associated with the adsorption of heavy metals and other pollutant molecules on different solids in aqueous phase. Immersion calorimetry is capable of determining the surface hydophilicity/hydrophobicity, oxygen content, micropore area and size, catalyst surface – reactivity correlations and surface acidity/basicity. The reaction calorimetry is also a liquid phase calorimetry that is used to measure the heat released during a catalytic reaction and is capable to directly determine the thermochemistry and kinetics associated with a chemical reaction. Single crystal adsorption calorimeter (SCAC) can accurately measure heats of adsorption on clean single crystal surfaces. SCAC is capable of determining the heat of adsorption associated with both smaller and larger molecules like CO, H2, O2, CO2, benzene, ethylene, etc. Hence, it is employed to study the thermodynamic stability and energetics of adsorbed intermediates that are critical parameters in a catalytic reaction.

References 1. Schick, C.: Glass transition under confinement-what can be learned from calorimetry. Eur. Phys. J. Spec. Top. 189(1), 3–36 (2010) 2. Miao, J., Wan, H., Shao, Y., Guan, G., Xu, B.: Acetalization of carbonyl compounds catalyzed by acidic ionic liquid immobilized on silica gel. J. Mol. Catal. Chem. 348(1–2), 77–82 (2011) 3. Du, Y., Zhou, L., Liu, Z., Guo, Z., Wang, X., Lei, J.: Designed formation of mesoscopical order of ionic liquid-based meso/

46

Calorimetry Techniques

macroporous Mo/TiO2 materials for high-performance oxidative desulfurization. Chem. Eng. J. 387, 124056 (2020) 4. Zhang, G., Wen, M., Wang, S., Chen, J., Wang, J.: Insights into thermal reduction of the oxidized graphite from the electro-oxidation processing of nuclear graphite matrix. RSC Adv. 8(1), 567–579 (2018) 5. Ratkovic, S., Peica, N., Thomsen, C., Bukur, D.B., Boskovic, G.: Thermal stability evolution of carbon nanotubes caused by liquid oxidation. J. Therm. Anal. Calorim. 115(2), 1477–1486 (2014) 6. Liu, Y., Dai, H., Du, Y., Deng, J., Zhang, L., Zhao, Z., Au, C.T.: Controlled preparation and high catalytic performance of threedimensionally ordered macroporous LaMnO3 with nanovoid skeletons for the combustion of toluene. J. Catal. 287, 149–160 (2012) 7. Chen, K., Liu, N., Zhang, M., Wang, D.: Oxidative desulfurization of dibenzothiophene over monoclinic VO2 phase-transition catalysts. Appl. Catal. B Environ. 212, 32–40 (2017) 8. Hou, J., Li, Y., Mao, M., Ren, L., Zhao, X.: Tremendous effect of the morphology of birnessite-type manganese oxide nanostructures on catalytic activity. ACS Appl. Mater. Interfaces. 6(17), 14981–14987 (2014) 9. Jin, R., Liu, Y., Wang, Y., Cen, W., Wu, Z., Wang, H., Weng, X.: The role of cerium in the improved SO2 tolerance for NO reduction with NH3 over Mn-Ce/TiO2 catalyst at low temperature. Appl. Catal. B Environ. 148–149, 582–588 (2014) 10. Yao, Y., Yang, Z., Zhang, D., Peng, W., Sun, H., Wang, S.: Magnetic CoFe2 O4 – graphene hybrids: facile synthesis, characterization, and catalytic properties. Ind. Eng. Chem. Res. 51(17), 6044–6051 (2012) 11. Boro, J., Thakur, A.J., Deka, D.: Solid oxide derived from waste shells of Turbonilla striatula as a renewable catalyst for biodiesel production. Fuel Process. Technol. 92(10), 2061–2067 (2011) 12. Li, H., Cheng, S., He, Y., Javed, M., Yang, G., Yang, R., Tsubaki, N.: A study on the effect of PH value of impregnation solution in nickel catalyst preparation for methane dry reforming reaction. ChemistrySelect. 4(31), 8953–8959 (2019) 13. Jabbour, K., Saad, A., Inaty, L., Davidson, A., Massiani, P., El Hassan, N.: Ordered mesoporous Fe-Al2O3 based-catalysts synthesized via a direct “one-pot” method for the dry reforming of a model biogas mixture. Int. J. Hydrog. Energy. 44(29), 14889–14907 (2019) 14. Janke, C., Duyar, M.S., Hoskins, M., Farrauto, R.: Catalytic and adsorption studies for the hydrogenation of CO2 to methane. Appl. Catal. B Environ. 152–153, 184–191 (2014) 15. Liu, Q., Sun, J., Feng, Q., Ji, S., Wang, Z.: A La-promoted Ni/MgAl2O4 catalyst with superior methanation performance for the production of synthetic natural gas. Catal. Today. 339, 127–134 (2020) 16. Ochoa, A., Valle, B., Resasco, D.E., Bilbao, J., Gayubo, A.G., Castaño, P.: Temperature programmed oxidation coupled with in situ techniques reveal the nature and location of coke deposited on a Ni/La2O3-Al2O3 catalyst in the steam reforming of bio-oil. ChemCatChem. 10(10), 2311–2321 (2018) 17. Amin, R., Liu, B., Ullah, S., Biao, H.Z.: Study of coking and catalyst stability over CaO promoted Ni-based MCF synthesized by different methods for CH4/CO2 reforming reaction. Int. J. Hydrog. Energy. 42(34), 21607–21616 (2017) 18. Homsi, D., Rached, J.A., Aouad, S., Gennequin, C., Dahdah, E., Estephane, J., Tidahy, H.L., Aboukaïs, A., Abi-Aad, E.: Steam reforming of ethanol for hydrogen production over Cu/Co-Mg-Albased catalysts prepared by hydrotalcite route. Environ. Sci. Pollut. Res. 24(11), 9907–9913 (2017) 19. Liu, J., Peng, H., Liu, W., Xu, X., Wang, X., Li, C., Zhou, W., Yuan, P., Chen, X., Zhang, W., Zhan, H.: Tin modification on Ni/Al2O3: designing potent coke-resistant catalysts for the dry reforming of methane. ChemCatChem. 6(7), 2095–2104 (2014) 20. Lima, S., Antunes, M.M., Fernandes, A., Pillinger, M., Ribeiro, M.F., Valente, A.A.: Catalytic cyclodehydration of xylose to furfural in the presence of zeolite H-beta and a micro/mesoporous beta/TUD-

1057 1 composite material. Appl. Catal. A Gen. 388(1–2), 141–148 (2010) 21. Zhou, C., Fang, Z.Z., Sun, P., Xu, L., Liu, Y.: Capturing low-pressure hydrogen using V Ti Cr catalyzed magnesium hydride. J. Power Sources. 413, 139–147 (2019) 22. Ismail, M., Zhao, Y., Yu, X.B., Ranjbar, A., Dou, S.X.: Improved hydrogen desorption in lithium alanate by addition of SWCNT–metallic catalyst composite. Int. J. Hydrog. Energy. 36(5), 3593–3599 (2011) 23. Huang, H.X., Yuan, J.G., Zhang, B., Zhang, J.G., Zhu, Y.F., Li, L.Q., Wu, Y., Zhou, S.X.: Improvement in the hydrogenationdehydrogenation performance of a Mg–Al alloy by graphene supported Ni. Int. J. Hydrog. Energy. 45(1), 798–808 (2020) 24. García-Mota, M., Bajdich, M., Viswanathan, V., Vojvodic, A., Bell, A.T., Nørskov, J.K.: Importance of correlation in determining electrocatalytic oxygen evolution activity on cobalt oxides. J. Phys. Chem. C. 116(39), 21077–21082 (2012) 25. Eggenhuisen, T.M., Prieto, G., Talsma, H., de Jong, K.P., de Jongh, P.E.: Entrance size analysis of silica materials with cagelike pore structure by thermoporometry. J. Phys. Chem. C. 116(44), 23383–23393 (2012) 26. Loskyll, J., Stoewe, K., Maier, W.F.: Infrared thermography as a high-throughput tool in catalysis research. ACS Comb. Sci. 14(5), 295–303 (2012) 27. Dussault, L., Dupin, J.C., Dumitriu, E., Auroux, A., Guimon, C.: Microcalorimetry, TPR and XPS studies of acid–base properties of NiCuMgAl mixed oxides using LDHs as precursors. Thermochim. Acta. 434(1–2), 93–99 (2005) 28. Zhao, H., Bennici, S., Shen, J., Auroux, A.: Nature of surface sites of V2O5–TiO2/SO42- catalysts and reactivity in selective oxidation of methanol to dimethoxymethane. J. Catal. 272(1), 176–189 (2010) 29. Zhao, H., Bennici, S., Shen, J., Auroux, A.: Influence of the host oxide of sulfated-titania catalysts on partial oxidation methanol reaction. Appl. Catal. A Gen. 385(1–2), 224–231 (2010) 30. Lilić, A., Wei, T., Bennici, S., Devaux, J.-F., Dubois, J.-L., Auroux, A.: A comparative study of basic, amphoteric, and acidic catalysts in the oxidative coupling of methanol and ethanol for acrolein production. ChemSusChem. 10(17), 3459–3472 (2017) 31. Lilić, A., Bennici, S., Devaux, J., Dubois, J., Auroux, A.: Influence of catalyst acid/base properties in acrolein production by oxidative coupling of ethanol and methanol. ChemSusChem. 10(9), 1916–1930 (2017) 32. Borowiec, A., Lilić, A., Morin, J.-C., Devaux, J.-F., Dubois, J.-L., Bennici, S., Auroux, A., Capron, M., Dumeignil, F.: Acrolein production from methanol and ethanol mixtures over La- and Ce-doped FeMo catalysts. Appl. Catal. B Environ. 237, 149–157 (2018) 33. Gervasini, A., Carniti, P., Marzo, M., Auroux, A.: Effect of the K+, Ba2+, and Nd3+ addition to Nb2O5 on intrinsic and effective acidity in relation to biomass reactions. J. Catal. 296, 143–155 (2012) 34. Rac, V., Rakić, V., Stošić, D., Otman, O., Auroux, A.: Hierarchical ZSM-5, beta and USY zeolites: acidity assessment by gas and aqueous phase calorimetry and catalytic activity in fructose dehydration reaction. Microporous Mesoporous Mater. 194, 126–134 (2014) 35. Dalil, M., Carnevali, D., Edake, M., Auroux, A., Dubois, J.-L., Patience, G.S.: Gas phase dehydration of glycerol to acrolein: coke on WO3/TiO2 reduces by-products. J. Mol. Catal. Chem. 421, 146–155 (2016) 36. Karaki, M., Karout, A., Toufaily, J., Rataboul, F., Essayem, N., Lebeau, B.: Synthesis and characterization of acidic ordered mesoporous organosilica SBA-15: application to the hydrolysis of cellobiose and insight into the stability of the acidic functions. J. Catal. 305, 204–216 (2013) 37. León, M., Faba, L., Díaz, E., Bennici, S., Vega, A., Ordóñez, S., Auroux, A.: Consequences of MgO activation procedures on its catalytic performance for acetone self-condensation. Appl. Catal. B Environ. 147, 796–804 (2014)

46

1058 38. Kikhtyanin, O., Bulánek, R., Frolich, K., Čejka, J., Kubička, D.: Aldol condensation of furfural with acetone over ion-exchanged and impregnated potassium BEA zeolites. J. Mol. Catal. Chem. 424, 358–368 (2016) 39. Wang, Z., Fongarland, P., Lu, G., Essayem, N.: Reconstructed La-, Y-, Ce-modified MgAl-hydrotalcite as a solid base catalyst for aldol condensation: investigation of water tolerance. J. Catal. 318, 108–118 (2014) 40. León, M., Díaz, E., Vega, A., Ordóñez, S., Auroux, A.: Consequences of the iron–aluminium exchange on the performance of hydrotalcite-derived mixed oxides for ethanol condensation. Appl. Catal. B Environ. 102(3–4), 590–599 (2011) 41. Fabiano, D.P., Hamad, B., Cardoso, D., Essayem, N.: On the understanding of the remarkable activity of template-containing mesoporous molecular sieves in the transesterification of rapeseed oil with ethanol. J. Catal. 276(1), 190–196 (2010) 42. Souza, R.O.L., Fabiano, D.P., Feche, C., Rataboul, F., Cardoso, D., Essayem, N.: Glucose–fructose isomerisation promoted by basic hybrid catalysts. Catal. Today. 195(1), 114–119 (2012) 43. Frolich, K., Bulanek, R., Frydova, E.: Energetic heterogeneity of Cu +-carbonyl complexes in Y zeolites: kinetics of CO desorption studied by FTIR combined with CO adsorption microcalorimetry. Microporous Mesoporous Mater. 171, 185–195 (2013) 44. Gonzalo-Chacón, L., Almohalla, M., Gallegos-Suarez, E., GuerreroRuiz, A., Rodríguez-Ramos, I.: Effects of the reduction temperature over ex-chloride Ru Fischer–Tropsch catalysts supported on high surface area graphite and promoted by potassium. Appl. Catal. A Gen. 480, 86–92 (2014) 45. Tanksale, A., Beltramini, J., Dumesic, J., Lu, G.: Effect of Pt and Pd promoter on Ni supported catalysts—a TPR/TPO/TPD and microcalorimetry study. J. Catal. 258(2), 366–377 (2008) 46. Cai, J., Bennici, S., Shen, J., Auroux, A.: The acid–base properties of nitrogen-containing mesoporous carbon materials. Microporous Mesoporous Mater. 212, 156–168 (2015) 47. Dragoi, B., Dumitriu, E., Guimon, C., Auroux, A.: Acidic and adsorptive properties of SBA-15 modified by aluminum incorporation. Microporous Mesoporous Mater. 121(1–3), 7–17 (2009) 48. Gervasini, A., Auroux, A.: Combined use of titration calorimetry and spectrofluorimetry for the screening of the acidity of solid catalysts in different liquids. Thermochim. Acta. 567, 8–14 (2013) 49. Rakić, V., Damjanović, L., Rac, V., Stošić, D., Dondur, V., Auroux, A.: The adsorption of nicotine from aqueous solutions on different zeolite structures. Water Res. 44(6), 2047–2057 (2010) 50. Yan, P., Xia, J.-S., Chen, Y.-P., Liu, Z.-P., Guo, J.-S., Shen, Y., Zhang, C.-C., Wang, J.: Thermodynamics of binding interactions between extracellular polymeric substances and heavy metals by isothermal titration microcalorimetry. Bioresour. Technol. 232, 354–363 (2017) 51. Barroso-Bogeat, A., Alexandre-Franco, M., Fernández-González, C., Gómez-Serrano, V.: Physico-chemical characterization of activated carbon–metal oxide photocatalysts by immersion calorimetry in benzene and water. J. Therm. Anal. Calorim. 125(1), 65–74 (2016) 52. Carvajal-Bernal, A.M., Gomez-Granados, F., Giraldo, L., MorenoPirajan, J.C.: Application of the sips model to the calculation of maximum adsorption capacity and immersion enthalpy of phenol aqueous solutions on activated carbons. Eur. J. Chem. 8(2), 112–118 (2017) 53. Grau-Marin, D., Silvestre-Albero, J., Jardim, E.O., Jagiello, J., Betz, W.R., Peña, L.E.: Evaluation of the textural properties of ultramicroporous carbons using experimental and theoretical methods. Carbon. 157, 495–505 (2020) 54. Silvestre-Albero, J.: Characterization of microporous solids by immersion calorimetry. Colloids Surf. Physicochem. Eng. Asp. 187–188(1–3), 151–165 (2001) 55. Vernimmen, J., Guidotti, M., Silvestre-Albero, J., Jardim, E.O., Mertens, M., Lebedev, O.I., Van Tendeloo, G., Psaro, R.,

L. Silvester et al. Rodríguez-Reinoso, F., Meynen, V., Cool, P.: Immersion calorimetry as a tool to evaluate the catalytic performance of titanosilicate materials in the epoxidation of cyclohexene. Langmuir. 27(7), 3618–3625 (2011) 56. Lopez-Ramon, M.V., Stoeckli, F., Moreno-Castilla, C., CarrascoMarin, F.: On the characterization of acidic and basic surface sites on carbons by various techniques. Carbon. 37(8), 1215–1221 (1999) 57. Zogg, A., Stoessel, F., Fischer, U., Hungerbühler, K.: Isothermal reaction calorimetry as a tool for kinetic analysis. Thermochim. Acta. 419(1–2), 1–17 (2004) 58. Meeks, M.R.: An analog computer study of polymerization rates in vinyl chloride suspensions. Polym. Eng. Sci. 9(2), 141–151 (1969) 59. Andersen, H.M.: Isothermal kinetic calorimeter applied to emulsion polymerization. J. Polym. Sci. [A1]. 4(4), 783–791 (1966) 60. Andersen, H.M.: Polymerization rates by calorimetry. II. J. Polym. Sci. [A1]. 7(10), 2889–2896 (1969) 61. Becker, F.: Thermokinetische Meßmethoden: Thermokinetische Meßmethoden. Chem. Ing. Tech. 40(19), 933–947 (1968) 62. Becker, F., Walisch, W.: Meßgröße eine Temperaturänderung oder eine andere, der Temperatur proportionale Größe ist. Hieraus ergeben sich für die Konstruktion des Kalorimeters einige wesentliche neue Gesichtspunkte. Zunächst geht wie bei allen isothermen Verfahren die Wärme- kapazität. 46, 279–293 (1965) 63. Sullivan, M., Rodosta, T., Mahajan, K., Damiani, D.: An overview of the Department of Energy’s CarbonSAFE Initiative: moving CCUS toward commercialization. AICHE J. 66, e16855 (2019) 64. Wang, J., Huang, Y., Wilhite, B.A., Papadaki, M., Mannan, M.S.: Toward the identification of intensified reaction conditions using response surface methodology: a case study on 3-methylpyridine N -oxide synthesis. Ind. Eng. Chem. Res. 58(15), 6093–6104 (2019) 65. Veedhi, S., Babu, S.R.: Process safety evaluation to identify the inherent hazards of a highly exothermic Ritter reaction using adiabatic and isothermal calorimeters. Org. Process. Res. Dev. 17(12), 1597–1602 (2013) 66. Richner, G., Neuhold, Y.-M., Hungerbühler, K.: Nonisothermal calorimetry for fast thermokinetic reaction analysis: solvent-free esterification of n-butanol by acetic anhydride. Org. Process. Res. Dev. 14(3), 524–536 (2010) 67. Alozie, D.-M., Courtes, P., Ha, B., Prat, D.: Determination of accurate specific heat capacities of liquids in a reaction calorimeter, by statistical design. Org. Process. Res. Dev. 15(6), 1412–1419 (2011) 68. Erwin, S., Schulz, K., Moritz, H.-U., Schwede, C., Kerber, H.: Increased reactor performance versus reactor safety aspects in acrylate copolymerization. Chem. Eng. Technol. 24, 305–311 (2001) 69. André, R., Bou-Diab, L., Lerena, P., Stoessel, F., Giordano, M., Mathonat, C.: A new reaction calorimeter for screening purposes during process development. Org. Process. Res. Dev. 6(6), 915–921 (2002) 70. Nogent, H., Le Tacon, X.: The differential reaction calorimeter: examples of use. J. Loss Prev. Process Ind. 16(2), 133–139 (2003) 71. Soussi-Baatout, A., Brahim, K., Khattech, I., Kamoun, L., Jemal, M.: Thermochemical and kinetic investigations of the phosphoric attack of Tunisian phosphate ore. J. Therm. Anal. Calorim. 131(3), 3121–3132 (2018) 72. Lachkar-Zamouri, O., Brahim, K., Bennour-Mrad, F., Khattech, I.: Attack of Tunisian phosphate ore by a mixture of sulfuric and phosphoric acid: thermochemical study by means of differential reaction calorimetry. Adv. Mater. Phys. Chem. 08(10), 411–427 (2018) 73. Garron, A., Bennici, S., Auroux, A.: In situ generated catalysts for NaBH4 hydrolysis studied by liquid-phase calorimetry: influence of the nature of the metal. Appl. Catal. A Gen. 378(1), 90–95 (2010) 74. Andrieux, J., Swierczynski, D., Laversenne, L., Garron, A., Bennici, S., Goutaudier, C., Miele, P., Auroux, A., Bonnetot, B.: A multifactor study of catalyzed hydrolysis of solid NaBH4 on cobalt nanoparticles: thermodynamics and kinetics. Int. J. Hydrog. Energy. 34(2), 938–951 (2009)

46

Calorimetry Techniques

75. Borroni-Bird, C.E., Al-Sarraf, N., Andersoon, S., King, D.A.: Single crystal adsorption microcalorimetry. Chem. Phys. Lett. 183(6), 516–520 (1991) 76. Borroni-Bird, C.E., King, D.A.: An ultrahigh vacuum single crystal adsorption microcalorimeter. Rev. Sci. Instrum. 62(9), 2177–2185 (1991) 77. Stuckless, J.T., Frei, N.A., Campbell, C.T.: A novel single-crystal adsorption calorimeter and additions for determining metal adsorption and adhesion energies. Rev. Sci. Instrum. 69(6), 2427–2438 (1998) 78. Ajo, H.M., Ihm, H., Moilanen, D.E., Campbell, C.T.: Calorimeter for adsorption energies of larger molecules on single crystal surfaces. Rev. Sci. Instrum. 75(11), 4471–4480 (2004) 79. Fischer-Wolfarth, J.-H., Hartmann, J., Farmer, J.A., FloresCamacho, J.M., Campbell, C.T., Schauermann, S., Freund, H.-J.: An improved single crystal adsorption calorimeter for determining gas adsorption and reaction energies on complex model catalysts. Rev. Sci. Instrum. 82(2), 024102 (2011) 80. Fischer-Wolfarth, J.-H., Farmer, J.A., Flores-Camacho, J.M., Genest, A., Yudanov, I.V., Rösch, N., Campbell, C.T., Schauermann, S., Freund, H.-J.: Particle-size dependent heats of adsorption of CO on supported Pd nanoparticles as measured with a single-crystal microcalorimeter. Phys. Rev. B. 81(24), 241416 (2010) 81. Carey, S.J., Zhao, W., Frehner, A., Campbell, C.T., Jackson, B.: Energetics of adsorbed methyl and methyl iodide on Ni(111) by calorimetry: comparison to Pt(111) and implications for catalysis. ACS Catal. 7(2), 1286–1294 (2017) 82. Zhao, W., Carey, S.J., Morgan, S.E., Campbell, C.T.: Energetics of adsorbed formate and formic acid on Ni(111) by calorimetry. J. Catal. 352, 300–304 (2017) 83. Silbaugh, T.L., Karp, E.M., Campbell, C.T.: Energetics of methanol and formic acid oxidation on Pt(111): mechanistic insights from adsorption calorimetry. Surf. Sci. 650, 140–143 (2016)

Lishil Silvester earned his PhD in heterogeneous catalysis from Université Lille1, France in 2013. He was a postdoctoral fellow at the Texas A&M University at Qatar and at the Ecole Centrale de Lille, France. He is currently a post-doctoral researcher in Université Claude Bernard Lyon1. His research interests include heterogeneous catalysis, porous hybrid materials, energy storage, calorimetry, and biomass valorization.

1059

Quentin Touloumet obtained his master’s degree in Green Chemistry from Université Toulouse III – Paul Sabatier. He is currently a PhD student in Université Claude Bernard Lyon1. His research interests include energy storage, calorimetry, porous hybrid materials, and catalysis.

46

Aline Auroux is currently Emeritus CNRS Research Director at the Institut de Recherches sur la Catalyse et l’Environnement de Lyon, France. She has published more than 346 articles and 17 book chapters mostly in the domain of catalysts characterization using calorimetry techniques. She is currently Secretary General of the French Chemical Society (SCF), Chair of EuChemS Chemistry and Energy Division.

Case Study: Calorimetry

47

Felipe Polo-Garzon

Contents 47.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1061

47.2

Strength and Surface Density of Acid/Base Sites . . . . . . 1062

47.3

Interaction of Reactants with the Surface . . . . . . . . . . . . . . 1064

47.4

Turnover Frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1064

47.5

Transient Catalyst Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1067

47.6

Elucidating a Reaction Mechanism . . . . . . . . . . . . . . . . . . . . . 1067

47.7

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1068

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1069

surface reconstruction of a set of perovskites impacts the density of active sites to a greater extent than the intrinsic reactivity of the active site. At the end of this chapter, fundamental insights into catalyst transient behavior during isomerization of n-butane, and the reaction mechanism for conversion of 4-methylpentan-2-ol, are presented. Keywords

Adsorption microcalorimetry · Acid-base catalysis · Mixed metal oxides · Surface reconstruction · Dehydration/dehydrogenation · Turnover frequency · Catalyst deactivation

Abstract

This chapter presents the use of adsorption microcalorimetry to study acid-base catalysis over mixed metal oxides. Understanding the acid-base properties of catalysts is of paramount importance for steering product selectivity in a wide variety of catalytic industrial processes. In particular, catalytic conversion of biomass into biofuels and chemicals demands comprehensive understanding of the acid-base properties of the catalyst to selectively convert functional groups. Herein, the use of probe molecules to measure the strength and density of acid and basic sites on mixed metal oxide surfaces is introduced. Further, the challenges when characterizing mixed metal oxides, presenting surface reconstruction, are explained. An example displays how combining adsorption calorimetry with ab initio calculations explains the activation of a reactant on the surface of a mixed metal oxide. Acid-base catalysis over nanoshapes is studied to unveil that different nanoshapes lead to different extents of surface reconstruction, which ultimately affect acid-base catalysis. The use of steady-state isotopic transient kinetic analysis (SSITKA), in combination with adsorption microcalorimetry, unveils that F. Polo-Garzon (*) Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA e-mail: [email protected]

47.1

Introduction

Acid-base catalysis refers to the catalytic conversion of chemical compounds into select products, where the product selectivity depends on the acid/base characteristics of the catalyst surface. By the end of the twentieth century, 127 industrial processes (dehydration, isomerization, alkylation, etherification, amination, cracking, etc.) involved acid-base solid catalysts. From the catalysts used in these industrial processes, 30% were oxides or complex oxides [1]. Conversion of macromolecules with oxygen functionalities derived from biomass represents a big area for acid-base catalysis, with the production of biodiesel as an attractive target in the search for independence from fossil fuel resources [2]. In addition, conversion of small sugars (glucose, fructose) into key intermediates (e.g., 5-hydroxymethylfurfural) for the production of biofuels and chemicals is an acid-base-catalyzed process [3]. Glycerol production goes hand in hand with biodiesel production, as it is a byproduct of this process that can be acid-base-catalyzed into useful chemicals [4, 5]. The synthesis of fuel additives, agrochemicals, antioxidants, plastics, pharmaceuticals, and fine chemicals also involves acid-base catalysis [6].

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_47

1061

1062

F. Polo-Garzon

47.2

Strength and Surface Density of Acid/ Base Sites

In the case of mixed metal oxides, their surface composition can be vastly different from their bulk composition. The crystallographic facet exposed at the surface, the surface segregation of cations, and the existing nonideal defect sites (kinks, steps, and corners) are factors that can affect the acidbase properties of the catalyst. It is evident that bulk characterization of mixed metal oxides does not provide meaningful

information about surface acidity/basicity. Therefore, a common practice is to rely on probe molecules to measure the strength and density of acid/base sites. Acidic molecules like CO2 and SO2 are used to probe basic sites, whereas basic molecules like NH3 or pyridine are used to probe acidic sites. Perovskites are a type of mixed metal oxides of the general formula ABO3, where A is an alkali, alkaline, or lanthanide metal, and B is a transition metal. Adsorption microcalorimetry studies have shown that the acid/base properties of perovskites can vary vastly depending on the cations in the structure, despite the cations having the same oxidation state (+2 for A, and +4 for B) in the structures studied: SrTiO3, BaTiO3, SrZrO3, and BaZrO3 (see Fig. 47.1a, b) [7]. For instance, when surface coverage approaches zero, the adsorption strength of NH3 on acidic sites can differ up to ~20 kJ/mol (Fig. 47.1a),

a) Heat of adsorption (kJ/mol)

100

b)

SrTiO3 BaTiO3 SrZrO3 BaZrO3

80 60 40 20 0

0

1 2 Surface coverage (μmol/m2)

140 Heat of adsorption (kJ/mol)

For mixed metal oxides, surface acidity relates to the effective positive charge at the surface, promoted by the metal cations (M+δ), whereas surface basicity relates to the effective negative charge at the surface induced by oxygen anions. In other words, acid sites on the surface withdraw electrons from the intermediate species during the catalytic process, whereas basic sites donate electrons. Although the definition of basicity and acidity of a catalyst surface seems straightforward at first glance, it is important to consider that mixed metal oxides contain multiple types of cations placed on surface defects (kink, steps, and corners). Therefore, the acid character of these cations varies with their location at the surface even if the metal atom is the same. On the surface of the catalyst, oxygen atoms with different electronic charges and coordination number are also encountered, creating a variety of basic sites. Drawing relationships between acid-base surface properties and catalytic performance must be done with care as one must consider factors such as the density of surface sites, the strength of the sites, and whether acid and basic sites work in tandem to unlock a specific catalytic cycle. Adsorption microcalorimetry is a powerful technique capable of providing quantitative data on the density and strength of acid/base sites. It is true that pulse chemisorption can provide information on the density of surface sites, and temperature-programmed desorption can provide indirect information on the strength and heterogeneity of the surface sites based on the desorption temperature of the titrant molecule. Nonetheless, adsorption microcalorimetry provides additional information: It relates the adsorption strength of the surface sites with the partial coverage of the surface and directly measures the adsorption strength in terms of specific heat released upon adsorption (e.g., kJ/mol) (Fig. 47.6). In addition, in order to test the adsorption strength on active sites, temperature-programmed desorption provides limited information if the adsorbate decomposes when the temperature increases. In this study case, on the use of adsorption microcalorimetry for acid-base catalysis over mixed metal oxides, various uses of this powerful technique are presented: (1) Titration of acid/base sites with probe molecules, (2) interaction of reactants with the catalyst surface, (3) estimation of turnover frequencies (TOF), (4) understanding transient catalyst behavior, and (5) elucidation of a reaction mechanism.

SrTiO3 BaTiO3 SrZrO3 BaZrO3

120 100 80 60 40 20 0

0

1

2

3

Surface coverage

4

(μmol/m2)

c)

SrTiO3

Acid site (µmol/m2) 0.42

Base site (µmol/m2) 3.47

Ratio of base to acid site 8.26

BaTiO3

0.29

1.25

4.31

SrZrO3

0.35

2.78

7.94

BaZrO3

0.26

2.11

8.12

Catalyst

Fig. 47.1 (a, b) Heat of adsorption of (a) NH3 and (b) CO2 on perovskite catalysts measured at 150 and 30
C, respectively; and (c) surface density of acid and basic sites on perovskites. (Reprinted with permission from ref. [7]. Copyright 2017 American Chemical Society)

47

Case Study: Calorimetry

1063

and the adsorption strength of CO2 on basic sites can differ up to ~20 kJ/mol (Fig. 47.1b) for the set of perovskites studied. Further, the density of sites on one perovskite can be up to approximately three times the density of sites on another perovskite (e.g., SrTiO3 has a 3.47 μmol basic site per m2, BaTiO3 has 1.25 μmol basic sites per m2) as shown in Fig. 47.1c. An important note to make is that the strength (heat released upon adsorption) and density of acid/basic sites depends on the probe molecule chosen to conduct the experiment. In addition, the temperature at which the titration is performed influences the density of adsorption sites and the heat released upon adsorption, as probe molecules overcome desorption barriers more easily as the temperature is increased. In the ideal scenario, the molecules used to characterize the basicity and acidity of the surface are also reactants in the catalytic reaction of interest. However, some molecules are prohibited from being used in adsorption microcalorimetry equipment due to their low-vapor pressure, given that it would be difficult to evacuate the system. It has been recently shown that surface reconstruction of perovskites can affect their surface termination, and thus the acid/base surface sites [8]. Treatment in O2 at temperatures above 500
C induces the segregation of Sr at the surface of SrTiO3, leading to a more basic surface. Treatment of SrTiO3 in

HNO3 exposes Ti at the surface, reducing the strength and density of basic sites (Fig. 47.2a), and increasing the strength and density of acid sites (Fig. 47.2b). Implications of the change in the surface termination of SrTiO3 for acid-base catalysis were shown using the conversion of 2-propanol as a probe reaction. 2-propanol dehydrogenates to produce acetone over basic sites and dehydrates to produce propene over acid sites (Fig. 47.2c). As shown in Fig. 47.2d, the single metal oxides TiO2 and SrO deliver propene selectivities of 95% and 15%, respectively. The surface termination of SrTiO3 can be manipulated to obtain three different catalyst surfaces: STO(HNO3),400
C (SrTiO3 treated in HNO3 and then calcined at 400
C), STO400
C (SrTiO3 calcined at 400
C), and STO550
C (SrTiO3 calcined at 550
C). These surfaces rank from the most to the least acidic as: STO(HNO3),400
C > STO400
C > STO550
C, and have Sr-surface concentrations of 28, 40, and 59%, respectively. These catalysts deliver propene selectivities of approximately 87, 40, and 28% (STO(HNO3),400
C, STO400
C, and STO550
C, respectively). It is clear that the conditions of the microcalorimetry study do not resemble operando conditions; yet, findings can be easily correlated to product selectivity, and enrich the understanding of the impact that the surface reconstruction of mixed metal oxides has on acid-base catalysis.

ΔHads,CO2 (kJ/mol)

a) 350

c) SrO

250

STO

200

STO(HNO3)

150

Propene –H2O

300

OH O –H2

2-propanol

TiO2-disk

Dehydrogenation (basic sites)

Acetone

100 50

TiO2-disk400 °C STO(HNO3),400 °C STO400 °C STO550 °C SrO400 °C

0

d)

b) 120 100

100

80 60 40 20 0 0

1

2

3

4

5

6

Surface coverage (μmol/m2)

7

Propene selectivity (%)

ΔHads,NH3 (kJ/mol)

Dehydration (acidic sites)

80 270 °C 265 °C 260 °C 255 °C 250 °C

60 40 20 0 0.0

Fig. 47.2 (a, b) Heat of adsorption of (a) CO2, and (b) NH3 on SrO, STO, STO(HNO3), and TiO2-disk catalysts measured at 30
C (CO2 and 2-propanol) and 150
C (NH3). All samples were treated at 550
C before adsorption microcalorimetry measurements. NH3 adsorption on

0.5 1.0 Sr/(Sr+Ti)

SrO was not detected; (c) 2-propanol conversion over acidic and basic sites; and (d) propene selectivity at steady-state during 2-propanol conversion. (Reprinted with permission from ref. [8]. Copyright 2017 Wiley Company)

47

1064

F. Polo-Garzon

47.3

Interaction of Reactants with the Surface the trend of 2-propanol adsorption on the perovskites studied

Besides characterizing the acidity/basicity of a catalyst surface using probe molecules, understanding the interaction of the reactants/products with the surface can help explain catalytic performance. Continuous-flow adsorption microcalorimetry studies are challenging, and batch-type experiments are more widely used. The in situ treatment of the catalyst, followed by microcalorimetry experiments, yet provides extremely valuable information. The fact that the heat released upon adsorption at different coverages is measured makes adsorption microcalorimetry results particularly suitable for comparison with ab initio computations. While studying the conversion of 2-propanol over a set of perovskites (SrTiO3, BaTiO3, SrZrO3, and BaZrO3) [7], the adsorption energy calculated using DFT and the adsorption energy measured via adsorption microcalorimetry were compared. As shown in Fig. 47.3a, DFT slightly overpredicted the adsorption strength of 2-propanol but clearly reproduced

a)

0

Experimental DFT calculated - Fixed surface DFT calculated - Relaxed surface

Adsorption enthalpy of 2-propanol (kJ/mol)

–30 –60 –90 –120 –150

Active site displacement (Å)

b) 0.6

0.4

0.2

SrZrO3 0.8

BaZrO3 SrTiO3 0.9

BaTiO3 1.0

Tolerance factor

Fig. 47.3 (a) Correlation of the adsorption energy of 2-propanol with the tolerance factor of the bulk perovskites for experimental adsorption energy, DFT-computed adsorption energy on a relaxed A surface, and DFT-computed adsorption energy on an unrelaxed (or fixed) A surface; (b) correlation of active site displacement on the A-terminated surface (that is, oxygen atom relaxation from DFT geometry optimization) with the tolerance factor. (Reprinted with permission from ref. [7]. Copyright 2017 American Chemical Society)

(from strongest to weakest: SrZrO3, SrTiO3 ≈ BaZrO3, and BaTiO3) when the surface atoms were allowed to relax during the simulations. In this case, adsorption microcalorimetry provided validation of the computational results. Next, computations of the unrelaxed perovskite structures were performed and evidently deviated from the experimental results. Thus, it was concluded that the relaxation of the cations on the surface had a greater impact on the adsorption strength of 2-propanol than the identity of those cations. However, the extent of surface relaxation upon 2-propanol adsorption, measured by the “active site displacement” (Fig. 47.3b), was correlated to the relative size of the cations in the structure, measured via the Goldschmidt tolerance factor (tolerance factor ¼ (rA þ rO)/[21/2(rB þ rO)]). The active site displacement is defined as the spatial movement of the surface oxygen atom that interacts with the hydrogen atom of the alcohol group in 2-propanol. In this work, the quantification of energy released upon adsorption of a reactant via microcalorimetry is paired with DFT simulations to comprehend the role of the relative size of the cations in the perovskite structure on surface relaxation and stabilization of the reactant, 2-propanol.

47.4

Turnover Frequencies

Different catalyst synthesis procedures, and surface reconstruction under reaction conditions, are factors that can limit the ability to compare catalytic performances reported by different research laboratories. Therefore, reporting TOF (rates per active site), along with rates per surface area, can provide information about the reactivity of surface sites and their density. To quantify the density of surface sites, it is common practice to rely on the titration of the surface sites with the reactant molecule, rather than relying on the theoretical atomic density of pristine ideal surfaces. The dehydrogenation of ethanol was studied over nanoshapes of SrTiO3 (cubes, dodecahedra, truncated cubes, and cubes etched in HNO3) [9]. Nanoshapes have different crystallographic facets exposed: Nanocubes are terminated with the (001) surface, dodecahedra are terminated mostly with (110), and truncated cubes have both (001) and (110) facets at the surface (see Fig. 47.4a–d). During catalytic measurements for ethanol conversion to acetaldehyde at 300
C, basic (NH3 and 2,6-di-tert-butylpyridine (DTBP)) and acidic molecules (CO2 and SO2) were cofed with ethanol. The reduction of the reaction rate, while cofeeding basic or acidic titrants separately, confirmed that acid-base pair sites are required for the reaction to proceed. The acid-base pairs were quantified via adsorption microcalorimetry of acetic acid, and these adsorption sites were used to calculate TOF (Fig. 47.4e). It was found that the TOF for acetaldehyde synthesis (dodecahedra > etched cube > truncated cube > cube) is correlated

Case Study: Calorimetry

1065

a)

e)

b) 50 nm

50 nm

(110)

(001)

(001)

200 nm

c)

200 nm

d) 50 nm 50 nm

10

Acetaldehyde formation rate (Pmol/Pmolpair.min)

47

Dodecahedra 8 6

Etched cubes

4 Truncated 2

Cubes

(110)

(001)

200 nm

1

2 3 4 Ethanol pressure (kPa)

g)

150 Cubes Truncated Dodecahedra Etched cubes

125 Heat of adsorption (kJ/mol)

0

100 75 50

5

6

Cubes Truncated Dodecahedra Etched cubes

250 Heat of adsorption (kJ/mol)

f)

0

200 nm

200

47

150

100

25 50 0 0.0

0.3

0.6

0.9

1.2

1.5

1.8

Surface coverage (Pmol/m2)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Surface coverage (Pmol/m2)

Fig. 47.4 (a–d) SEM and TEM images of SrTiO3 nanoshapes: (a) cubes, (b) truncated, (c) dodecahedra, and (d) etched cubes STO. The inset shows the schematic drawing of each SrTiO3 nanocrystal. (e) Initial acetaldehyde formation rate as a function of ethanol pressure

over shape-controlled SrTiO3 nanocrystals. Reaction conditions: 300
C, 50 mg of catalyst, 40 mL/min Ar. (f, g) Adsorption microcalorimetry experiments: (f) NH3 and (g) CO2. (Reprinted with permission from ref. [9]. Copyright 2018 American Chemical Society)

with the strength and density of acid sites titrated with NH3 (Fig. 47.4f), and it is inversely correlated with the strength of basic sites titrated with CO2 (Fig. 47.4e). No correlation was found between the catalytic performance and the facets exposed at the surface of the nanocatalysts; however, different nanoshapes led to different surface reconstructions after treatment in oxygen at 550
C (pretreatment used before catalytic measurements). Therefore, the nanoshape provides a “pre-stage” for different surface reconstructions after thermal treatment [9]. TOF (e.g., min1) can be estimated by dividing the reaction rates per surface area (e.g., mol/min.m2) by the density of adsorption sites (e.g. mol/m2), although this estimation presents some shortcomings: The adsorption sites are usually not measured under operando conditions, and TOF would be reported per adsorption site not per active site. A more rigorous strategy to measure rates per active intermediate

(approach to active site) is the implementation of steadystate isotopic transient kinetic analysis (SSITKA). After showing that treatment of SrTiO3 nanoshapes in O2 at 550
C led to enrichment of Sr at the surface (Fig. 47.5a, b) and creation of step-sites (not shown), Bao et al. [10] used SSITKA to discern the impact of surface reconstruction on the catalytic rates per active site for 2-propanol conversion. The reaction rates per surface area were considerably different among the nanoshapes after being pretreated in O2 at 550
C (Fig. 47.5c). In addition, different pretreatment temperatures in O2 showed to impact the rates per surface area over dodecahedra (Fig. 47.5d). Interestingly, TOFs at 280
C per active intermediate (TOFSSITKA) for 2-propanol conversion over SrTiO3 nanocrystals were the same, independent of the nanoshape of the catalyst and the pretreatment temperature (Fig. 47.5e). Thus, it was concluded that the difference in rates per surface area was rooted on the density of the active

1066

F. Polo-Garzon

c)

Cube, 550C Truncated, 550C Dodecahedra, 550C

1 × 10–4 Reaction rateIPA (mmol/m2·s)

0.4 0.6 0.4 Cube_NT Truncated_NT Dodecahedra_NT

Surface

4

260

1 × 10–4

270

280

290

0 –4

TOF SSITKA TOF Calorem.

TOF Theoretic

–8 0 –4

5 × 10–5

TOF SSITKA

TOF Calorem.

–8

f)

0C 75

55 od D

od D

e._

e._

c._

200 Cube_550C Truncated_550C Dodecahedra_550C Dodecahedra_750C Dodecahedra_900C

150

Enthalpy (kJ/mol)

un

300

Reaction temperature (°C)

55

50

290

_5

280

be

270

Cu

260

0C

C

0C

TOF Theoretic

0 250

300

Reaction temperature (°C)

e)

NT 550C 750C 900C

2 × 10–4

250

Bulk

Ln(TOF) [unit: Ln (s–1)]

Reaction rateIPA (mmol/m2·s)

d) 2 × 10–4

0

3

0C

2 Depth (nm)

2 × 10–5

90

1

4 × 10–5

e._

0

6 × 10–5

od

0.0

Dodecahedra Cube Truncated

D

0.2

8 × 10–5

Tr

b)

Sr/(Sr+Ti) ratio

0.6

Acetone

0.8

Propene

1.0

a)

100

50

0 0

2

4 6 8 Surface coverage (μmol/m2)

Fig. 47.5 (a, b) Surface Sr/(Sr þ Ti) cation intensity ratio in SrTiO3 nanocrystals as a function of probing depth using low-energy ion scattering (LEIS): (a) after 550
C treatment; (b) without treatment; (c) steady-state reaction rate for 2-propanol conversion for different nanoshapes. Pretreatment: 550
C in O2. Reactor feed: 0.5 μL/min of 2-propanol in 40 mL/min of Ar; (d) steady-state reaction rate for

10

2-propanol conversion over dodecahedra at different pretreatment temperatures in O2. Reactor feed: 0.5 μL/min of 2-propanol in 40 mL/min of Ar; (e) TOF at 280
C of nanoshapes after different pretreatment temperatures in O2; and (f) heat of adsorption of 2-propanol measured at 30
C. (Reprinted from ref. [10], Copyright (2020), with permission from Elsevier)

47

Case Study: Calorimetry

sites, not their intrinsic reactivity. Comparison of TOFSSITKA and TOFCalorimetry (TOFCalorimetry ¼ rate per surface area / density of adsorption sites for 2-propanol via microcalorimetry) clearly shows that the density of adsorption sites exceeds the density of active sites (Fig. 47.5e), and therefore, TOFCalorimetry is an underprediction. Nonetheless, estimation of TOF using adsorption sites from microcalorimetry (TOFCalorimetry) is better than using the theoretical density of adsorption sites from a pristine surface (TOFTheoretic). Theoretical calculation of adsorption sites does not account for the heterogeneity of the surface and the repulsion effect of the adsorbates.

47.5

Transient Catalyst Behavior

Many catalysts undergo induction periods and deactivation periods. During these transient periods, the fundamental questions are the following: Did the nature of the active sites change? Did sites become unblocked/blocked? In the previous section, it was shown that measuring TOF provides a way to discern whether the nature or the density of the active sites change. If the intrinsic reactivity per active sites does not change, TOF should not change. In some cases, some of the surface sites become blocked by surface intermediates and by-products, and thus the overall reaction rate and product selectivity are transient up to the point blockage of surface sites reaches a steady-state. Adsorption microcalorimetry provides a way to titrate the changes in the active sites during transient conditions. A clever experimental setup has been developed to perform catalytic reactions and perform adsorption microcalorimetry experiments in situ (not operando) [11]. In this setup, a capillary tube is introduced downward into a calorimetry cell, and the end of the capillary is placed in close proximity to the catalyst bed. The capillary tube delivers the reactant mixture. The product mixture flows upward in the cell and is analyzed online. The reaction under study was the isomerization of n-butane over sulfated zirconia, which produced mostly isobutane (selectivity  88%) at 105
C when performed in a plug-flow reactor. Despite the microcalorimetry design consisting of flow on a catalyst bed, instead of flow through a catalyst bed (plug-flow reactor), catalytic performance with time-on-stream (TOS) for n-butane isomerization was remarkably similar. Thus, the setup in the calorimetry cell can be used to understand transient behavior observed in the plug-flow reactor. The experimental procedure to study the transient behavior of the catalyst design consisted of performing the catalytic reaction in the modified microcalorimeter cell for a certain TOS, then the cell was evacuated and adsorption microcalorimetry of the reactant (nbutane) and product (isobutane) was performed. The surface sites of the sample were characterized after activation, after 60 min of reaction, and after 120 min, as at these times the

1067

catalytic performance presents drastic differences. Results showed that the amount of adsorption sites decreased with TOS. In addition, the adsorption strength of both n-butane and isobutane decreased for coverages in the range 0.002–0.02 mmol/g, whereas the adsorption strength held constant (40 kJ/mol) for coverage in excess of 0.02 mmol/ g. The sites with adsorption strength between 60 and 40 kJ/mol present on the fresh sample disappeared during the induction period of n-butane isomerization and are presumed to be sites where reaction chain carriers were created. Heats of adsorption for coverages below 0.002 mmol/g were difficult to interpret due to the scatter of the data. Thus, it was clearly shown that surface sites that strongly adsorb the reactant n-butane and the product isobutane become blocked as the reaction proceeds, and that the surface sites with weaker adsorption carry the reaction forward. Although this modified calorimeter cell is very useful to study the dynamic nature of active sites as the reaction evolves, it can provide limited information when the reaction temperature is not far above the adsorption temperature during microcalorimetry studies. This limitation arises from the fact that surface reactions can occur upon adsorption of the molecules during microcalorimetry experimentation.

47.6

Elucidating a Reaction Mechanism

As shown earlier in this section, the conversion of 2-propanol provides information on the acid/basic character of the catalyst surface, and it is widely used due to the facile product detection and quantification. However, the conversion of 2-propanol does not provide information on the strength of the surface sites. Other probe reactions provide a richer pool of information, for instance, the product selectivity during the conversion of 4-methylpentan-2-ol provides information on the density and strength of basic and acid sites. The possible products include 1- and 2- alkenes (dehydration), and ketones (dehydrogenation). Based on the density of basic/acid sites and their strength, three reaction mechanisms for dehydration can be present: E1, E2, and E1cB. Further, alkene isomerization and ketone condensation reactions can be related to the strength of the surface sites [12]. In the E1 mechanism, a carbocation intermediate is formed. This mechanism occurs when the concentration of acid sites is high when compared with the concentration of basic sites, as the acid sites attack the -OH group of the reactant first, creating a carbocation. When a proton is released, an internal double bond is formed. Strong acid sites lead to various isomers. The E2 mechanism occurs when the strength and density of acid sites is similar to the strength and density of basic sites. In this mechanism, C-H and C-OH bond breaking is concerted, producing an alkene with an internal double bond. The E1cB mechanism occurs when the density of acid and basic sites is similar, but basic

47

1068

F. Polo-Garzon

Catalyst MgNiAl (0.22) MgNiAl (0.47) MgNiAl (4.05)

Basic sites (μmol/g) 7

Acid sites (μmol/g) 146

Basic sites/acid sites 0.05

Strong basic sites/strong acid sites 0.6

388

637

0.61

2.0

206

448

0.46

1.4

O C O

1 O C O

O O C

O C O

O O C

2

3

sites are stronger. The C-H bond breakage occurs first, forming a carbanion; next, OH is released to form an alkene with a terminal double bond (1-alkene). Meloni et al. [12] studied the conversion of 4-methylpentan-2-ol over a series of MgNiAl(x) mixed oxides (x is Ni/Mg molar ratio). Using adsorption microcalorimetry, they characterized the concentration and strength of acid/basic sites. In Table 47.1, the “strong basic sites” and “strong acid sites” are sites with adsorption energy between 110 and 150 kJ/mol. This range of adsorption energy was determined by plotting dn/d (ΔHads) versus ΔHads, where n is the uptake of the probe molecule and ΔHads is the heat released upon adsorption. The appearance of peaks on this plot means the existence of a group of relatively homogenous sites. For NH3 adsorption, a peak at ΔHads ¼ 120 kJ/mol was observed. For CO2 adsorption, a peak at ΔHads ¼ 150 kJ/mol with a tail toward lower adsorption energies was observed. Hence, it was concluded that “strong” acid or basic sites have adsorption energies between 110 and 150 kJ/mol. The reaction was performed under mild conditions (200
C and contact time τ ¼ 0.01 h.gcat/galcohol) and under harsh conditions (350
C and contact time τ ¼ 0.52 h.gcat/ galcohol). Under mild conditions and independent from the sample, only dehydration products were observed: 4-methylpent-1-ene (1-alkene, 9–7%), cis and trans 4-methylpent-2-ene (2-alkene, 27–23%), and skeletal isomers of C6-alkenes (68–65%). The fact that the concentration of acid and basic sites is unbalanced (Table 47.1, column “basic sites/ acid sites”) rules out the E2 and E1cB mechanisms. The low selectivity toward the 1-alkene supports the nonexistence of the E1cB mechanism. The E1 dominates since acid sites prevail over basic sites. Further, the existence of isomerization products indicates that the acid sites are strong. Under harsh conditions, both dehydration and dehydrogenation occurred over the samples, and samples presented different product selectivity. The sample with the lowest Ni/Mg ratio, MgNiAl(0.22) showed mostly dehydration to 2-alkene, with some other products (dehydration: 1-alkene, C6 isomers, and dehydrogenation: ketone). MgNiAl(0.47) presented formation of mostly heavy ketones within the first 3 h of TOS, but

O

C

O

Heat of adsorption (kJ/mol)

Table 47.1 Density of basic and acid sites, measured via CO2 and NH3 adsorption, respectively [12]

1 2 3

Surface coverage (µmol/m2)

Fig. 47.6 Sketch of CO2 adsorption microcalorimetry on a mixed metal oxide

later on the production of the ketone (4-methlypentan-2-one) was the major product. MgNiAl(4.05) produced mostly the ketone, with some heavy ketones. Under harsh conditions, the stronger basic or acid sites are more relevant for reaction. In MgNiAl(0.47) and MgNiAl(4.05), the concentration of basic sites prevails over the acidic sites, given the obtained dehydrogenation products, and the ratio “strong basic sites / strong acid sites” (Table 47.1). In particular, the condensation reaction (to produce heavy ketone) in MgNiAl(0.47) suggests strong basic sites. The decreasing heavy ketone selectivity with TOS indicates that the strong basic sites become poisoned by strongly adsorbed products. In conclusion, the relative density of basic and acid sites, along with the relative density of strong basic sites and strong acid sites, was a key input from adsorption microcalorimetry to understand reaction selectivity for the conversion of 4-methylpentan-2-ol. This type of combined microcalorimetry-kinetic study allows rational design of catalysts for acid-base catalysis in general.

47.7

Summary

Adsorption microcalorimetry is an exceptional technique to gain information about the density and strength of surface acid/base sites. However, these properties are dependent on the molecule adsorbed. The heterogeneous nature of mixed metal oxide surfaces, due to surface defects and segregation of cations, makes it challenging to accurately identify number and strength of basic/acid sites. However, adsorption microcalorimetry provides insight into the heterogeneity of surface sites, and their strength, by unveiling the relation between adsorption energy of a molecule versus surface coverage. Relating the density and strength of acid/basic sites with catalyst performance unlocks fundamental understanding of product selectivity. With regard to the adsorption microcalorimetry of reaction species, measuring the density of surface sites enables the

47

Case Study: Calorimetry

estimation of TOF, a critical value to compare the intrinsic reactivity of different catalyst samples. Further, adsorption of reaction species helps understand transient catalyst behavior. Although adsorption microcalorimetry is typically performed in situ (not operando), modified microcalorimeter cells allow for studying the state of catalytic sites at different TOS. The transient study of catalytic sites unveils the surface sites that prevail under steady-state conditions. This technique, coupled with other spectroscopic techniques, ab initio simulations and kinetic analysis, helps achieve a comprehensive understanding of the highly heterogeneous surfaces of mixed metal oxides for acid-base catalysis.

References 1. Tanabe, K., Hölderich, W.F.: Industrial application of solid acid–base catalysts. Appl. Catal. A Gen. 181(2), 399–434 (1999) 2. Sani, Y.M., Daud, W.M.A.W., Abdul Aziz, A.R.: Activity of solid acid catalysts for biodiesel production: a critical review. Appl. Catal. A Gen. 470, 140–161 (2014) 3. Kourieh, R., Rakic, V., Bennici, S., Auroux, A.: Relation between surface acidity and reactivity in fructose conversion into 5-HMF using tungstated zirconia catalysts. Catal. Commun. 30, 5–13 (2013) 4. Behr, A., Eilting, J., Irawadi, K., Leschinski, J., Lindner, F.: Improved utilisation of renewable resources: new important derivatives of glycerol. Green Chem. 10(1), 13–30 (2008) 5. Katryniok, B., Paul, S., Capron, M., Dumeignil, F.: Towards the sustainable production of acrolein by glycerol dehydration. ChemSusChem. 2(8), 719–730 (2009) 6. Polo-Garzon, F., Wu, Z.: Acid–base catalysis over perovskites: a review. J. Mater. Chem. A. 6(7), 2877–2894 (2018) 7. Foo, G.S., Polo-Garzon, F., Fung, V., Jiang, D.-E., Overbury, S.H., Wu, Z.: Acid–base reactivity of perovskite catalysts probed via conversion of 2-propanol over titanates and zirconates. ACS Catal. 7(7), 4423–4434 (2017)

1069 8. Polo-Garzon, F., Yang, S.-Z., Fung, V., Foo, G.S., Bickel, E.E., Chisholm, M.F., Jiang, D.-E., Wu, Z.: Controlling reaction selectivity through the surface termination of perovskite catalysts. Angew. Chem. Int. Ed. 56(33), 9820–9824 (2017) 9. Foo, G.S., Hood, Z.D., Wu, Z.: Shape effect undermined by surface reconstruction: ethanol dehydrogenation over shape-controlled SrTiO3 nanocrystals. ACS Catal. 8(1), 555–565 (2018) 10. Bao, Z., Fung, V., Polo-Garzon, F., Hood, Z.D., Cao, S., Chi, M., Bai, L., Jiang, D.-E., Wu, Z.: The interplay between surface facet and reconstruction on isopropanol conversion over SrTiO3 nanocrystals. J. Catal. 384, 49–60 (2020) 11. Wrabetz, S., Yang, X., Tzolova-Müller, G., Schlögl, R., Jentoft, F.C.: Characterization of catalysts in their active state by adsorption microcalorimetry: experimental design and application to sulfated zirconia. J. Catal. 269(2), 351–358 (2010) 12. Meloni, D., Sini, M.F., Cutrufello, M.G., Monaci, R., Rombi, E., Ferino, I.: Characterization of the active sites in MgNiAl mixed oxides by microcalorimetry and test reaction. J. Therm. Anal. Calorim. 108(2), 783–791 (2012)

47

Felipe Polo-Garzon received his BS from Universidad del Valle (2011, Colombia), and his PhD from Clemson University (2015, US). After obtaining his doctorate, he joined Oak Ridge National Laboratory (US) as a postdoctoral fellow and later became an R&D Staff. His research deals with the development of reactivity descriptors for heterogeneous catalysis by using in situ/operando characterization techniques, kinetic analysis, and computational tools.

Part VIII Soft Operando

Chemometrics and Process Control

48

Dirk Engel and Clemens Minnich

Contents 48.1

Why Chemometrics? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1073

48.2 48.2.1 48.2.2 48.2.3 48.2.4

Mechanistic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Basics of Mechanistic Methods . . . . . . . . . . . . . . . . . . . . . . . . . Workflow of Peak Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . Workflow of Spectral Hard Modeling . . . . . . . . . . . . . . . . . . Workflow of Univariate Calibration . . . . . . . . . . . . . . . . . . . .

48.3 48.3.1 48.3.2

Statistical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1083 Basics of Statistical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1083 Workflow of PLS Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084

48.4

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1090

1074 1074 1076 1078 1080

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1090

Abstract

Measurement techniques for operando monitoring of catalytic chemical processes are reaching new levels of maturity. However, the primary outputs of monitoring devices, e.g., raw optical spectra, require multivariate interpretation or processing to extract useful information. Chemometric methods can achieve this in different ways, either in a mechanistic way using physicochemical principles or using statistical approaches. The concepts behind them are presented and compared, and the three most prominent methods – peak integration, spectral hard modeling, and projection to latent structures (PLS regression, also known as partial least squares) – are demonstrated in case studies using mid-infrared, Raman, and nuclear magnetic resonance (NMR) spectroscopy. The underlying principles are transferrable to other techniques as well if certain constraints remain fulfilled. The workflow style of the chapter is expected to efficiently assist users which are new to the field in solving their challenges in multivariate spectral analysis. D. Engel · C. Minnich (*) S-PACT GmbH, Aachen, Germany e-mail: [email protected]; [email protected]

Keywords

Chemometrics · Spectral analysis · Mechanistic methods · Statistical methods · Spectral hard modeling

48.1

Why Chemometrics?

Operando monitoring of catalytic processes has become increasingly important due to the high requirements of efficiency and specificity. The more a process is pushed to its limits, the more urgent the need for a monitoring approach which enables real-time control of the related process conditions. One challenge is the design and implementation of appropriate monitoring hardware, including light/field sources, signal transmitters like optical fibers, localized detection devices like probes or flow cells, and signal processing units. However, the generated primary output typically represents just a raw image of the process (e.g., a raw optical spectrum) which requires additional interpretation or extraction of the relevant process parameters. The output of this next-level data interpretation can be of various nature: • Qualitative, e.g., identification of species • Semiquantitative, e.g., prediction of relative trends • Fully quantitative, e.g., prediction of absolute values, referenced with a lab method A qualitative analysis commonly identifies categorical features (categories, group names, text identifiers, etc.) of a sample while a quantitative analysis predicts numerical features. The prediction of numerical features can deliver any value on a continuous scale (with a given uncertainty); the identification of a categorical feature gives one value out of a predefined set, commonly accompanied by a probability to qualify the reliability of the result. Categorical features play a minor role in real-time process analytics and are not dealt with below.

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6_48

1073

1074

D. Engel and C. Minnich

Whatever the nature of the analysis output will be, it is typically used as an input either for mechanistic or kinetic studies, i.e., in an explorative way, or in known environments to predict the current status of a process and to use this information as a basis for real-time process control. To convert the raw sensor data into useful information, data analysis approaches are applied which are typically summarized as chemometrics. This term does not stand for a specific analysis method but for the pool of methods to approach the primary data with the aim of extracting processrelevant information. From the authors’ experience, a selection of three fundamental approaches is sufficient to address most of the applications where quantitative information should be extracted from spectral data: 1. Peak integration optionally followed by univariate calibration 2. Spectral hard modeling optionally followed by univariate calibration 3. Projection to latent structures (PLS) involving multivariate calibration After introducing the basics of each approach, the practical workflow is demonstrated on use cases for mid-infrared (MIR), nuclear magnetic resonance (NMR), and Raman spectra. The examples and figures are created with the help of analysis software PEAXACT (S-PACT GmbH, Germany) which makes all the listed methods available. This chapter is expressly not intending to be a compendium for chemometrics but is meant as a guideline to the practical user of spectroscopic tools for studying chemical processes.

48.2

Mechanistic Methods

48.2.1 Basics of Mechanistic Methods The most robust and therefore most preferable way of extracting quantitative information from a spectrum is by using a mechanistic analysis method. The term mechanistic refers to a rigorous description of a studied system based on causal relationships and fundamental laws of physics that govern the system. In spectroscopy, this “system” is the spectrum of a mixture, which, as physics tells us, is proportional to the concentration of the mixture components. This proportionality applies to every single point in the spectrum, even though different components have peaks at different spectral positions. Mathematically, the total spectral signal S at spectral position x can be expressed by the sum product of two factors SðxÞ ¼ c1 ∙ s1,ref ðxÞ þ c2 ∙ s2,ref ðxÞ þ . . . þ ck ∙ sk,ref ðxÞ where ci represents the molar concentration or quantity of each mixture component i, and si, ref represents the spectral

landscape or reference shape of how each component contributes to the mixture spectrum. The equation holds for most spectroscopic techniques and can be derived from respective physical principles, with different origins of the shape factors sref: • For transmission/absorption spectroscopy, the equation can be derived from the Beer-Lambert law, which relates the attenuation of light to the properties of the material it travels through. sref originates from the molar attenuation coefficient. However, the mechanistic equation does not hold for near-infrared spectroscopy, where the reference shape of each component contains vibrational overtones and combination bands which strongly depend on the physicochemical environment. These additional sources of variance are often better modeled by statistical methods. • For Raman spectroscopy, the equation can be derived from first principles of the Raman effect, which describes inelastic scattering of light at molecules. sref originates from the Raman scattering cross-section. • For NMR spectroscopy, the equation can be derived from quantum mechanical calculations describing the effect of the molecule structure on the local magnetic field experienced by a nucleus. sref originates from electron currents in the bonds of a molecule generating local magnetic fields counteracting the outer static magnetic field. The abovementioned equation is the motivation for all mechanistic analysis methods in quantitative spectroscopy since the quantity of each mixture component is encoded in the intensity of every point of the spectrum. Just analyzing a single spectral point (the height of a peak) is highly discouraged though, because it would capture the full uncertainty at that point, including noise, local artifacts, or physical distortions of the signal related to temperature or molecular interactions. To be more robust against such influences, methods like peak integration and spectral hard modeling use many points of the spectrum to first calculate an intermediate variable – the component area – from which component properties can be predicted afterward. Prediction requires further steps to model the relationship between areas and concentrations, a procedure better known as model training or calibration. Using component areas, which literally are componentspecific areas under the spectrum, also has a physical motivation which is related to the limited resolution of most detectors: What should be an intensity spike at a very distinct position is instead recorded as an intensity distribution across multiple points. These distributed peaks are typically shaped like the Voigt distribution, a convolution of Gaussian and Lorentzian distributions, and can be anything from broad bands as in near-infrared and UV-VIS spectroscopy to moderate widths as in mid-infrared, Raman, and low-field NMR

48

Chemometrics and Process Control

spectroscopy, and to the very thin peaks of high-field NMR spectroscopy. Although exceptions are possible, mechanistic methods are more often used for spectra with sharp peaks. Peak integration methods use the mechanistic equation as follows: If one could find a spectral region where only one component contributes to the total signal, that signal would tell the component’s concentration. This can be exploited by numerically integrating the signal of one or more nonoverlapping component-specific peaks to calculate the component area Ai ¼ ci ∙ ai,ref, which corresponds to the concentration except for an unknown reference area ai, ref to be determined by calibration. Obviously, peak integration requires that one can find at least one nonoverlapping characteristic peak per component. Spectral hard modeling methods follow a different path: If one knew all the component-specific peak landscapes sref one could calculate all the unknown component concentrations by a procedure called model fitting. This can be exploited by modeling each of the component-specific landscapes by a group of peak functions, and then fitting the model to a measured spectrum. Alternatively, the spectral landscape can be modeled by measured pure component spectra, but such a model would be rather stiff compared to the flexibility provided by parameterized peak functions. For example, a function having a parameter for the peak position can be fitted to measured spectra even if the measured peak shifts around. Fitting the hard model results in component areas, but in contrast to peak integration, these areas originate from the fitted peak functions and therefore correctly account for overlapping peaks. Still, the conversion to concentrations requires calibration. Calibration involves reference samples, i.e., mixtures for which the concentration is already known, either from other quantitative methods or by specifically creating mixtures of known composition. In theory, a single reference sample would suffice to calculate the unknown reference areas aref, but such a single-point calibration is discouraged because it does not allow for the calculation of diagnostic properties like uncertainties or outlier statistics. Instead, multiple reference samples should be used, and univariate regression should be employed to calculate aref in a least squares sense. An important advantage in the calibration of a mechanistic model – compared to a statistical model – is its capability for extrapolation, i.e., using the calibrated model to predict concentrations outside the range of reference concentrations it is trained with. This can actively be used to reduce the number of reference samples and make calibration more costefficient, but it is particularly interesting for reactive systems when reference mixtures of certain compositions are difficult or impossible to obtain. Even without calibration, component areas already provide semiquantitative information about the monitored

1075

process. For example, one can tell whether a component is missing if its area is zero. Furthermore, the ratio of areas calculated for a component from two spectra is equal to the c A concentration ratio i,Sample2 ¼ i,Sample2 , so that one can easily ci,Sample1

Ai,Sample1

analyze the relative change of concentrations along a series of mixtures. It must be noted that comparing areas of two different components is not possible, because due to different references aref, each component’s area has a different scale. An exception exists for NMR spectra, where reference areas only differ by the number of component-specific nuclei n, so that the concentration ratio of any two components can be c A n calculated by i ¼ i j. In all other cases, calibration would cj

Aj ni

be needed to convert areas of different components onto a common concentration scale. An important requirement for both calibrated and calibration-free analyses is that for all measurements of the same experiment the reference areas aref must stay constant. If fulfilled, they cancel out in ratios or can be determined once by calibration. If not fulfilled, component area ratios of any two spectra would be incomparable, and individual calibrations would be needed for every single spectrum, which would render any quantitative analysis impossible or at least impractical. As the reference areas depend as well on experimental settings as on environmental properties, the following should be considered to keep them constant: • All measurements of one experiment, including associated calibration measurements, should be performed with the same instrument setup and with same acquisition settings, including, e.g., exposure time, accumulation count, laser power, fiber optic probes, length of fiber optic cables, optical path length, and receiver gain. • The spectral shapes sref and thus the reference areas aref might depend on the temperature. Identifying this dependency is difficult though, as molar concentrations also depend on the temperature. Therefore, if possible, measurements should be performed at a constant temperature. If not possible, one should conduct a test and record spectra of the same mixture at different temperatures. If spectra change only by a constant factor, this is likely due to the temperature dependency of the concentration or density and can be ignored. If some spectral regions change differently than others, this is likely due to the temperature dependency of sref, and, in the worst case, requires a separate analysis per temperature. • The spectral shapes sref might also depend on the mixture composition. This dependence is visible in the form of peak shifts or changes of the peak shape and is mostly observed in liquids due to molecular interactions. It is known, however, that the integral shape is invariant to these interactions, i.e., the reference areas aref remain

48

1076

unchanged. Therefore, for peak integration, one must only choose integration limits wide enough such that shifting peaks always stay inside. For spectral hard modeling, one must only choose a model flexible enough such that the model gets correctly fitted to the changed spectrum. In practical applications it sometimes is impossible to assure measurements with constant values of aref. For example, one might be forced to measure calibration samples and process samples on different instruments, or the studied process might be nonisothermal. But even with experimental settings kept constant, and environmental conditions controlled tightly, there may still be other factors (except for the concentration) that cause the spectrum to change, such as aging of the equipment or fouling. The effect on the spectrum can be additive, i.e., in the form of an unwanted background signal, or multiplicative, i.e., in the form of unknown signal scaling. Both effects would unintentionally contribute to the calculated component areas and therefore must be removed by proper signal pretreatment. Removing the spectral background can be achieved by subtracting a straight line or some other baseline function from the signal. Removing the effect of unknown scaling is trickier though, but generally involves signal standardization techniques such as dividing each spectrum by the peak area of an internal standard. In addition to baseline removal and standardization, many more pretreatment operations exist, and their application always aims for either suppressing unwanted signal or emphasizing useful signal, but whenever possible one should rather try to record clean spectra in the first place.

48.2.2 Workflow of Peak Integration Due to its simplicity for mid-infrared, Raman, NMR, and other kinds of spectroscopy with sharp peaks, peak integration has certainly become the most established analysis method. Its workflow is as follows: First, one must identify at least one characteristic nonoverlapping peak per component, e.g., by comparing the pure component spectra. In some cases, a small peak overlapping a much broader peak of another component may still be a viable choice if the broad peak can be treated as background signal and gets removed appropriately. Components for which no nonoverlapping characteristic peak exists may still be analyzed by spectral hard modeling or other methods. Secondly, one chooses integration limits for characteristic peaks, and, if necessary, removes unwanted background signal within the limits. One could choose limits and baselines for each spectrum individually, but in the interest

D. Engel and C. Minnich

of automation one should rather define them once in such a way that works for all samples to be analyzed. This concept is known as an integration model; it requires a little foresight to choose limits wide enough to always include shifting peaks but at the same time narrow enough to not accidentally include adjacent peaks. Similarly, a baseline function must be defined to work with all samples, e.g., a straight line from the left to the right integration limit. Finally, the specified integration model should be tested and potentially adjusted by means of a few samples with different compositions. A graphical display of the integrated areas is helpful in this step to evaluate whether component areas get integrated as intended. The actual numerical integration is often carried out with the help of software tools but can also be approximated by simply taking the sum of all signal intensities within the integration limits. After the integration model has been tested, it can be applied to spectra of the process under study to calculate component areas which then can be utilized for a semiquantitative analysis of relative concentrations.

Case 1: Carbonate Selectivity in CO2 Utilization Scientists in the catalysis research lab CAT Catalytic Center at RWTH Aachen University, Germany, aim to produce large-scale polymers using CO2 as a feedstock in processes of the industrial stakeholder Covestro. Promising products identified are functionalized polyethers, i.e., copolymers of glycols like 1.2-ethanediol (ethylene glycol) or 1.3propanediol (propylene glycol) with CO2. These materials serve as substitutes or comonomers for the generation of polyurethanes from diisocyanates. Of specific interest is the substitution of carbon building blocks from fossil sources by CO2 which would otherwise be released to the atmosphere – the so-called carbon capture and utilization. As CO2 is not readily reacting with its reaction partners, commonly a catalyst is required to activate the molecule for chemical conversion [5]. Propylene oxide (PO) is the most prominent epoxide involved in these processes. Depending on the process conditions and on the nature of the catalyst, PO and CO2 can either be converted in the presence of Cr- or Co-salen complexes in an equimolar reaction to yield an alternating polycarbonate, or with epoxide excess to yield polyethercarbonates (PEC) containing both ether and carbonate linkages in the polymer backbone (Fig. 48.1). The copolymerization is performed in semi-batch mode, charging an oligomeric polyethylene glycol (PEG) as a starter chain in pressurized CO2. Upon addition of propylene oxide, carbonate moieties are built up. Although the selection of

48

Chemometrics and Process Control

1077

catalyst mainly affects the polymer structure, it does also affect the generation of cyclic propylene carbonate (cPC) as a by-product. To avoid undesired material loss into this side reaction, close monitoring of the process is desired. Mid-infrared spectroscopy has been identified as a means to track the reaction in real time, and thanks to probe technology the measurements can be performed in situ. From the detection and quantification of the carbonates and of the residual propylene oxide, several performance parameters can be assessed: PO conversion and related safety indications, the loss of PO and CO2 into the undesired cyclic propylene carbonate, and the buildup kinetics for carbonate linkages in the desired polyethercarbonate.

O

mn R

O *

O

O

*

R

m n Polyethercarbonate (PEC) O

R

+ n CO2

Catalyst

+

CO2

O

Catalyst

O

O Cyclic propylene carbonate (cPC)

Fig. 48.1 Copolymerization of epoxides with CO2 to yield polyethercarbonate (blue) with the cyclic carbonate as side product (red)

Fig. 48.2 Spectral surface from a semi-batch copolymerization of CO2 with propylene oxide. Evolving C¼O signals from the newly formed carbonates are visible around 1800 cm1

All mid-infrared (MIR) spectra are measured with a Bruker MATRIX-MF process spectrometer, based on Fourier transform (FT) technology using a liquid nitrogen (LN2) cooled MCT (mercury cadmium telluride) detector. Inline monitoring is achieved by mounting a fiber-optical diamond probe for spectra acquisition in attenuated total reflectance (ATR) mode into the pressurized stainless steel reactor. Inline spectra from the process (Fig. 48.2) show characteristic spectral features in the entire fingerprint region. New signals appear in the C¼O range (1850–1700 cm1) which are characteristic for the two carbonate species PEC and cPC formed in the chemical reaction. These signals are sufficiently separated and specific for quantification based on peak integration and thus for the monitoring of catalytic selectivity (Fig. 48.3). The excellent separation of the characteristic signals in this range prefers this range over the 1300 cm1 range, which displays stronger signals but with significant peak overlap. Not only are the integration limits clearly identifiable, but the baseline definition for the integration area is also obvious. Here, a linear fit baseline cuts the lower peak offset between the local minima (Fig. 48.4). Once the two integration ranges are specified, application of the model to spectra of a semi-batch process provides component areas (Fig. 48.5) and allows a semiquantitative assessment of the carbonate formation. The output of component areas is already sufficient to study selectivity between catalyst types, different process conditions and other parameter changes. It should be noted again that component areas of the two species must not be compared to each other because they are scaled differently.

1

Absorbance

0.8 0.6 0.4 0.2 0 6 4 2 Time (h)

1800

1600

1400

1200

Wavenumber (cm–1)

1000

800

48

1078

D. Engel and C. Minnich

b)

Absorbance

0.3

1

0.8

0.8

0.6

0.6

0.2

Absorbance

x = 1812

cPC (eq.)

0.4

3

x = 1743

2.5 2

0.4

PEC (eq.)

a)

1.5 0.4 0.1

0.2 1 0.2

0

0 1850

1800

1750

1700

1850

1800

1750

1700

Wavenumber (cm–1)

Wavenumber (cm–1)

Fig. 48.3 Peak evolution with concentration changes for (a) cPC and (b) PEC. Square inlays: preview of intensity versus concentration. Note the slight shift of the peak maximum with increasing concentration

0.2

8 Component area (a.u.)

Absorbance

0.15

0.1

0.05

0

1850

1800 1750 Wavenumber (cm–1)

1700

6 PEC cPC

4

2

0 0

2

4

6

8

Time (h)

Fig. 48.4 Visualization of the integration ranges for the two carbonate species

Fig. 48.5 Component area profiles along a semi-batch experiment for the two carbonates

48.2.3 Workflow of Spectral Hard Modeling Whenever peak integration is not applicable because of overlapping peaks, the next logical step is to use spectral hard modeling methods. Here the modeling goal is to build a mathematical model according to the mechanistic equation, using groups of peak functions to resemble the peak landscapes sref of all mixture components. Several ways exist to generate such a hard model; the most straightforward one is indirect hard modeling (IHM) [1, 2]. The name refers to the fact that the hard model of the complete mixture is built indirectly with the help of measured pure component spectra. Other hard modeling techniques exist for when pure components cannot be measured, for example, hard modeling factor analysis (HMFA) which extracts the pure component hard

models from a set of mixture spectra [3]. Spectral hard modeling requires specialized software to perform the individual operations. In detail, the workflow of IHM is as follows: First, one should identify the spectral region of interest by looking at spectra of the pure components and finding regions of least overlap. Although hard models can deal with overlapping peaks, improved performance is to be expected when reducing the model to less overlapping regions. In the trivial case when no peaks overlap one could of course use peak integration methods instead. Next, a spectral model is created for the peak landscape of every mixture component; a step which can be automated

48

Chemometrics and Process Control

by fitting peak functions to measured pure component spectra (peak fit). This reduces the modeling complexity to only two decisions: the kind of peak functions to use and the number of functions to fit. The kind of peak function on the one hand must be suitable for the type of spectroscopy, e.g., for mid-infrared, Raman, and NMR spectroscopy, either the Voigt distribution function or the less complex pseudo-Voigt function – a linear combination of a Gaussian and a Lorentzian distribution function – is a physically motivated function to describe peak shapes. The number of functions on the other hand is dictated by the observed peaks within the selected spectral region, though a higher priority should be laid on fitting large peaks while tiny peaks may not be fitted at all. To put the decision about the number of fitted peaks on a more objective basis, a goodness-of-fit criterion such as the root mean squared (RMS) spectral residuals between measured spectrum and fitted model should be used. Eventually, the hard model of the full mixture is the mechanistic equation, with sref being represented by componentspecific groups of flexible peak functions. Finally, the flexibility of the hard model should be tested by fitting it to a few spectra of different mixtures. Here, too, RMS residuals can be used to assess the goodness of fit. Also, a graphical display of both the mixture spectrum and the fitted model is helpful to evaluate whether the fitting works as intended. Ensuring that the hard model fits the shape of the spectra correctly is important, because only then the component areas calculated from the peak functions also correctly correspond to the component quantities, so that the model can be applied reliably to analyze unknown mixtures [4].

Case 2: Continuous Lithiation Reaction Benchtop low-field nuclear magnetic resonance (LF-NMR) spectrometers are bringing this powerful technique close to real-time process monitoring because they overcome the limitations of traditional high-field instruments: voluminous setup, liquid-helium-cooled superconductors, deuterated solvents, and long acquisition times. The drawback is an increased peak width and more overlapping peaks, which in many cases renders peak integration impossible and demands for spectral hard modeling. The application of LF-NMR spectroscopy in continuous flow is a promising step toward the establishment of this technology as a routine tool in process monitoring. A standard PTFE tube or glass flow cells can serve as a flow cell inside the magnetic core of the benchtop NMR device. A project consortium involving the German Federal Institute for Materials Testing (Berlin, Germany) evaluated a commercial-scale modular pilot plant for tests under industrial conditions within the European Union’s Horizon 2020 project CONSENS [6]. The reaction under consideration

1079

is the pharmaceutically relevant aromatic coupling of aniline with o-fluoro-nitrobenzene (oFNB) to give 2-nitrodiphenylamine (NDPA), initiated by organolithium reagents (Fig. 48.6). The product is generated in its lithiated form (Li-NDPA). Close monitoring of this reaction in compact continuous flow setups is crucial to guarantee a consistent product quality, to make sure that the safety constraints of the process are not violated, and at the same time to drive the process to its economically optimal operating point. This mainly concerns the starting materials and intermediates aniline, lithiated aniline, and o-fluoro-nitrobenzene which can easily be distinguished from the product. All spectra are acquired with a Magritek Spinsolve benchtop NMR instrument (43.32 MHz 1H frequency) as free induction decay (FID). Automated subsequent FID processing – zero filling, apodization, Fourier transformation, and phase correction – is established first, followed by baseline subtraction, spectral alignment, and range selection. The NMR signals in the relevant aromatic protons range are unique for all interesting components but are overlapping significantly (Fig. 48.7a, b). Therefore, spectral hard modeling (of the analysis software PEAXACT) is used to analyze mixture spectra of the reaction experiments. Three reference solutions of the pure components are prepared in tetrahydrofuran (THF), and the spectral signal in the relevant aromatic ppm range (which is free of THF) is fitted by peak functions (Fig. 48.8). The decision about the number of fitted functions is a trade-off between minimizing the RMS residuals, i.e., reducing the difference between model and spectrum, and overfitting, i.e., the fitting of irrelevant noise. To support the decision, it is helpful to either measure pure component spectra with very high signal-tonoise ratio, or to estimate the amplitude of noise from replicate measurements. The resulting pure component models are combined to a mixture model which is subsequently fitted to all mixture spectra. For an example spectrum, the comparison of the flexible fit (Fig. 48.9a) with a hypothetic linear fit (Fig. 48.9b) shows the need for the nonlinear adjustments of peak parameters, resulting in a significant reduction of the spectral residuals. Here, the improvement of the fit especially by adjusting the position of the large peaks between 7.5 and

NO2

NO2

NH2 +

F 2 LiHMDS THF

NLi +

LiF

Fig. 48.6 Coupling of aniline (red) with o-fluoro-nitrobenzene (yellow) to give 2-nitrodiphenylamine (blue) in tetrahydrofuran (THF). Reaction is initiated with an organolithium reagent (LiHMDS)

48

1080

a)

D. Engel and C. Minnich

b)

3

2

Intensity

Intensity

1.5 2

1

1 0.5 0

0 8.5

8

7.5

7

6.5

6

8.5

8

7.5

7

6.5

6

Chemical shift (ppm)

Chemical shift (ppm)

Fig. 48.7 (a) Pure component LF-NMR spectra of aniline (red), oFNB (yellow), and Li-NDPA (blue). (b) Typical mixture spectra. Only the aromatic region of the spectra is shown

exist in this case as one would see for poor-quality spectra, and slightly decreasing RMS values toward the end of the batch suggest that the model performs better for high conversion, which is beneficial as one aims for high conversion in the continuously operated process setup later. For further kinetic studies of the reaction, a transformation of component areas into absolute concentration is desired. We come back to this case study in the next section and perform the necessary univariate calibration.

a)

b)

48.2.4 Workflow of Univariate Calibration

c)

8.5

8

7.5

7

6.5

6

Chemical shift (ppm)

Fig. 48.8 Pure component models for (a) aniline (13 peaks), (b) oFNB (19 peaks), and (c) Li-NDPA (26 peaks). All peaks are fitted to reference solutions of the materials in THF

7.0 ppm is visible by eye already, but also the residuals plot is a useful display of the fit quality. The calculated component areas serve as relative concentration measures and nicely display the progress of the reaction and the buildup of the product (Fig. 48.10a). A constant component area for the two substrates indicates the finalization of the reaction when reaching full conversion or an equilibrium state. In addition to component areas, the analysis delivers RMS residuals per spectrum (Fig. 48.10b). No obvious outliers

Univariate calibration aims at modeling the one-to-one relationship between a component’s concentration and its area by training the model with a set of reference samples. In theory, the relationship should be linear, but in practice there are several sources of errors to deal with, such as random errors in component areas caused by spectral noise, a systematic bias in component areas caused by structurally imperfect integration or hard models, as well as uncertainties or outliers in the provided reference data. Therefore, the relationship is often modeled by a low-order polynomial, and statistical regression is used to minimize the differences between actual reference concentrations and concentrations predicted by the polynomial model. The calibration workflow is as follows: The first step is to design and generate a training set of reference samples. The number of training samples depends on the complexity of the mixture, but for mechanistic methods it is typically in the order of 5–50 when dealing with liquid mixtures of two to five components. For gas mixtures fewer samples are often sufficient. More important than the sample count is a good and uniform coverage of the interesting concentration space. Using the model to predict

Chemometrics and Process Control

Intensity

a)

1081

b)

400 300

Intensity

48

200

300 200 100

100

0

0 8

7.5

7

6.5

6 Residuals

8.5 Residuals

48

400

50 0 –50 8.5

8

7.5

7

6.5

6

8.5

8

7.5

7

6.5

6

8.5

8

7.5

7

6.5

6

50 0 –50 Chemical shift (ppm)

Chemical shift (ppm)

Fig. 48.9 (a) Fit of a spectral hard model (red) to a mixture spectrum, Li-NDPA contribution highlighted (bold blue). (b) Hypothetic linear fit yielding large spectral residuals throughout the fitted range

and pick the one with the best predictive performance. The validation can be based on the following objective criteria:

a) 300 oFNB LiNDPA Aniline

Component area (a.u.)

250 200 150 100 50 0

5

10

15 20 Sample index

25

30

10

15 20 Sample index

25

30

RMS spectral residuals

b) 10

5

0 5

Fig. 48.10 Component area profiles (a) and RMS spectral residuals (b) of a lab-scale semi-batch experiment

concentrations outside the trained space is possible, but one should try to avoid it whenever possible. Next, it is recommended to perform several regressions of alternative polynomials (linear, quadratic, etc.), validate each,

• The root mean squared error (RMSE) between actual and predicted concentrations is an estimate of the standard uncertainty of predicted values. It is the strongest indicator for the predictive performance of the calibration model. The error should be as small as possible. • The mean error (bias) between actual and predicted concentrations is an estimate of the systematic error in predicted values. The bias should be as close to zero as possible. • The coefficient of determination R2 is the proportion of the variance in concentrations that can be explained by the polynomial. The value should be close to one. • The intercept and slope of the recovery function, which is the linear trend line between predicted and actual concentrations. It must not be confused with the regression function (calibration curve), which is the polynomial relationship between component areas and actual concentrations. The recovery function should be as close as possible to the identity line, i.e., a function with an intercept of zero and a slope of one. All these criteria tend to improve with increasing polynomial degree, but so does the risk of overfitting. Therefore, the most appropriate degree is a compromise between a number as small as possible and an RMSE as small as acceptable. An even better strategy for validating the calibration model is by calculating the criteria based upon an independent test set of reference samples. Test samples are not supposed to be used for the regression but are just predicted and compared to the known actual concentrations. The

1082

D. Engel and C. Minnich

resulting root mean squared error of prediction (RMSEP) is a measure for the robustness of the calibration when it comes to analyzing new unknown samples. Typical test samples would be, e.g., from a different instrument or sensor, for a different temperature, for a different lot of feedstock material, or for a new concentration range. An acceptable compromise for when no independent test samples can be provided are cross-validation techniques. Cross-validation uses training samples to assess how prediction results will be generalized to an independent test set by partitioning the training set into two subsets, performing the regression on one subset, and validating with the other. Multiple rounds of cross-validation, e.g., 10, should be performed using different partitions, so that the resulting root mean squared error of cross-validation (RMSECV) gives a proper estimate of the predictive performance for unknown samples. A graphical representation of RMSE values, the regression function, or the recovery function are often used to compare calibration alternatives and support the decision for a polynomial degree. This decision, which needs to be made for each component separately, is the final step of calibration. After calibration, the model can be used to predict concentration from unknown samples, and, when proper statistics software is used, also to calculate associated prediction uncertainties. The prediction uncertainty defines a range around the predicted value within which the true value lies with a specified probability. Ranges for approximately 68% (standard uncertainty) and 95% (expanded uncertainty) are common. When no statistics software is used, the average standard uncertainty can be roughly approximated by the RMSE (ideally RMSEP or RMSECV), and the average expanded uncertainty is roughly twice the standard

Case 2: Continued The training set for the calibration is generated from a lab-scale semi-batch experiment in which the reacting mixture was not only circulated through the inline low-field NMR device, but additionally transferred to a high-resolution NMR instrument for referencing purposes. Four mixture compositions in the target range, each with three replicate acquisitions, were used as training samples for the calibration step. From the same experiment, three additional mixture compositions along the conversion profile with three acquisitions each served as an independent test-set for the model validation. This allows to use the RMSEP as a decision criterion for selecting the polynomial degree of the calibration curve. As no significant improvement of RMSEP with increasing polynomial degree is observed, a straight line without intercept (“simple”) is the recommended selection in this case (Fig. 48.11a). A cross-check with the calibration curve confirms that the straight line is a suitable fit to both training and test samples (Fig. 48.11b). The detailed view into the recovery plot (Fig. 48.12) shows that the model recovers the given true values for the oFNB content without locally over- or underpredicting on average, confirmed by the recovery line basically matching the identity line (ideal case). However, individual samples at high concentration show discrepancies between prediction and true value, indicating an increased uncertainty in this range.

b) 0.5

0.01

0.4 RMSEP RMSECV RMSE

0.005

0 Simple

Linear

Quadratic

Cubic

oFNB (mol/L)

RMSE (mol/L)

a) 0.015

uncertainty. As all prediction results are subject to uncertainty, predicted values are considered complete only when accompanied by a statement of the associated uncertainty.

0.3

0.2

Training samples Test samples Calibration curve Extrapolated: y = 1.76164x

Function 0.1

R2

1 0.995

0

0.99 Simple

Linear

Quadratic

Function

Cubic

0

0.1

0.2

Component area

Fig. 48.11 (a) Evolution of RMSE values with increasing polynomial degree of the calibration curve, (b) selected calibration curve

0.3

48

Chemometrics and Process Control

1083

i.e., when the rigorous causal relationship between the quantity and the shape contribution to the spectrum is, at least partially, unknown. Typical use cases are:

Predicted oFNB (mol/L)

0.5 0.4 0.3 0.2

Training samples Test samples Recovery line Identity line

0.1 0

0

0.1

0.2

0.3

0.4

0.5

Difference

True oFNB (mol/L) 0.02 0 –0.02 0

0.1

0.2

0.3

0.4

0.5

True oFNB (mol/L)

• Near-infrared spectroscopy, where the components’ reference shapes are unknown and not necessarily scaling linearly with the components’ quantities. • The quantity of interest is not a concentration but, e.g., a mixture property, whose shape contribution to the spectrum is unknown. • The mixture contains unknown components. To cope with the lack of information, statistical methods like principal component analysis (PCA) or projection to latent structures (PLS, also known as partial least squares) [7] use data-driven techniques to learn from a large set of reference spectra how to decompose the spectral signal S into a quantity contribution t and shape contribution pref: SðxÞ ¼ t1 ∙ p1,ref ðxÞ þ t2 ∙ p2,ref ðxÞ þ . . . þ tk ∙ pk,ref ðxÞ

Fig. 48.12 Recovery and difference plot for oFNB

The trained model is applied in real time to a continuously operated flow setup with attached NMR inline analysis. Process conditions like the total flow rate, stoichiometry, raw material quality, or temperature are varied to study the effect on the process performance. After each variation, equilibration of the process is attended until steady-state operation is reached. The predicted concentration profiles show excellent repeatability of consecutive spectra in each steady-state operation point (Fig. 48.13a). Consecutive predicted values deviate significantly less than 0.01 mol/L, whereas expanded uncertainties for a 95% level of confidence are in the range of 0.02 mol/L. The relatively high prediction uncertainties must be tolerated in this case, as they are a direct result of the seen deviation in the training data. For all analyzed process spectra, outlier probabilities can be calculated by putting the RMS spectral residual of each fitted spectrum into perspective to the distribution of RMS values from training samples. Outlier probabilities way below 1 confirm that there are no spectral outliers among the process spectra (Fig. 48.13b).

48.3

Statistical Methods

48.3.1 Basics of Statistical Methods Another approach to extracting information from spectra is to use statistical methods. Such methods should be considered only after ruling out the applicability of mechanistic methods,

The equation looks like the mechanistic equation in the previous section. However, the individual summands no longer represent real chemical components but so-called latent factors or principal components (PC), where t and pref are referred to as PC scores and loadings. Loadings can be interpreted as independent sources of variance that contribute to the total signal and that are common to all reference spectra. They are ranked from most to least important, i.e., the first few loadings typically explain most of the spectral variance in the dataset while the rest explains unimportant sources such as spectral noise. Scores can be interpreted as the amount by which each source of variance contributes to a specific spectrum. The exact meaning of scores in terms of real physical or chemical properties can often not be interpreted, but it is indeed not important. Important for predicting a mixture property is only that it has a certain correlation with the spectral signal, because then the property is responsible for one or more sources of variance and accordingly relates to one or more scores. The unknown relationship between scores and the property of interest is then established by multivariate calibration. In almost all cases, calibrating PLS scores is superior to calibrating PCA scores. The two methods use different algorithms for ranking the sources of variance, and PCA only considers spectral variance while PLS also considers variance in the property of interest. Consequently, PLS principal components are ranked by their importance for calibrating the property, so that rather than a question of which individual scores to calibrate, it simply becomes a question of up to which rank. The rank must be chosen high enough to capture all relevant sources of variance, yet low enough to not capture noise.

48

1084

D. Engel and C. Minnich

a) oFNB (mol/L) Li-NDPA (mol/L) Aniline (mol/L)

0.2

Predicted

0.15

0.1

0.05

0

0

0.5

1

1.5

1

1.5

Time (h)

Outlier probability

b)

0.6 0.4 0.2 0 0

0.5 Time (h)

Fig. 48.13 (a) Predicted concentrations with expanded uncertainty and (b) outlier probability based on the quality of fit

An important condition for the robust application of PC-based methods is that the sources of variance identified from reference samples, i.e., the loadings pref, also apply to future process spectra for which predictions should be made. This requires a representative set of reference spectra which captures all the relevant variance in the system, so that the statistical method can learn about all sources of variance. Predicting a property from an unseen spectrum is then a matter of calculating the new scores by which the reference loadings contribute to the new spectrum, and converting the scores using the established calibration. However, if the new spectrum contains any new source of variance, statistical methods cannot extrapolate into the unknown and the prediction is likely going to fail. Typical sources of variance to keep in mind are the chemical components and their concentration ranges, the spectrometer and its measurement settings, the temperature and other environmental conditions, as well as detector noise and

other random effects. Ideally, any factor one is not interested in should be kept constant to reduce the complex task of providing representative reference samples. Shifting peaks and other nonlinear peak variations are another source of variance. This is problematic because the statistical equation is linear with respect to the loadings, so that nonlinear effects can only be approximated by increasing the rank of the model. In case of strongly shifting peaks, a more appropriate method, if applicable, might be spectral hard modeling. If the PLS model gets sufficiently validated, though, the technique can be applied for versatile scenarios, which is one of the reasons for its widespread use.

48.3.2 Workflow of PLS Calibration Projection to latent structures and multivariate calibration are two operations that most chemometric software packages

48

Chemometrics and Process Control

treat as one. The union, known as PLS regression or PLS calibration, is reasonable because both operations require the same reference data for training. In the PLS part, references are used to identify and rank principal components, and in the multivariate calibration part, references are used to model the one-to-many relationship between the property of interest and PC scores. The workflow is in many ways identical to univariate calibration, starting with the design and measurement of reference samples, followed by the actual calibration for alternative ranks, and finished by validating the alternatives to find the one with best predictive performance. A major difference to univariate calibration is the number of reference samples needed for training. The number depends on the factors that are responsible for the variance in the mixture spectra. When dealing with mixtures of two to five components it is typically in the order of 20–200. The sample count can easily grow large because the training set should cover all factor combinations, e.g., different levels of concentrations of the individual components, different temperatures, and different instruments. It is important to remember that data-driven statistical methods like PLS cannot reliably be used to extrapolate, i.e., to predict features outside the trained space. Therefore, good coverage of the relevant factors is key, and creating reference samples usually dominates the time spent on the entire calibration procedure. To save some time for the preparation of reference samples one should conduct a design of experiments (DOE) which aims for an optimal way to cover all known physical factors in their respective ranges with the least amount of training samples. Designing the training set guarantees a uniform coverage of all relevant factors and avoids overrepresentation of very similar samples, e.g., too many samples from the typical operating point of a continuous process. The actual PLS calibration is comparatively quick when performed with appropriate statistics software. It is recommended to perform several calibrations for alternative PLS ranks (typically between 1 and 20) and validate each using the same objective criteria and graphical tools mentioned earlier: the root mean squared error (RMSE) between actual and predicted feature value, the mean error (bias), the coefficient of determination R2, and the recovery plot. In addition, a graphical display of the PLS loadings is helpful to determine the rank that marks a transition from relevant sources of variance to noise. Validation with an independent test set of references, or at least cross-validation, is crucial for statistical models. Their lacking ability for extrapolation and the easy risk of overfitting by choosing a rank too high demands for a sound testing of the model’s robustness against unknown samples. Like in univariate calibration, the most appropriate rank is a compromise between a number as small as possible to avoid overfitting and an RMSE as small as acceptable. And like in univariate calibration, statistics software can use this

1085

information about calibration errors afterward to calculate uncertainties for predicted values. The steps above are the minimum requirement for PLS calibration. Several optional steps exist for improving the predictive performance of the model: • Spectra can be pretreated to remove unimportant sources of variance and emphasize relevant variance. In addition to baseline removal, for PLS it is common to artificially increase variance by using first- or second-order derivatives. • Prior to calibration, the spectral region can be reduced to those points which show a high correlation with the to-becalibrated feature. • Prior to calibration, outliers toward the set of reference spectra can be removed. • Starting from an initial calibration, the model can be iteratively refined, e.g., by filtering the spectral region even further or by removing outliers toward the model. A complete PLS calibration including a demonstration of the optional improvement steps is demonstrated in the following example.

Case 3: Nutrients, Metabolites, and Cell Parameters in Mammalian Cell Culture In recent years, spectroscopic analysis methods have proven their use especially in biocatalytic processes, both in fermentation processes to manufacture bio-based materials, or in cell culture for biopharmaceuticals production. In these cases, a multitude of molecular constituents potentially appears in the spectra, e.g., carbohydrates and amino acids as nutrients, metabolites like organic acids, and very specific product proteins in extremely low concentrations below the typical detection limits of optical spectroscopy. However, a reliable correlation can be expected between the nondetectable ultralow concentration of target molecules and the detectable sources of spectral variance, e.g., the signals of the key nutrients, key metabolites, and total biomass carbohydrates and proteins. Accordingly, the application of mechanistic methods for the trace components can be excluded, whereas statistical methods are a promising approach to quantification. The presented example originates from a mammalian cell culture, used to produce a specific target protein. Cells are fed with sugars as the key nutrient (“carbon source”), and various protein mixtures like yeast extract are used as nitrogen sources for the cell growth. Currently, Raman spectroscopy is one of the favorite techniques when it comes to bioprocess monitoring. This is mostly because the sensitivity to water is extremely low, so that even in the presence of large amount of water the dissolved analytes can still be detected. Additionally, bioprocesses are comparably slow with batch times of

48

1086

days to weeks, allowing for long acquisition times over several minutes to generate spectra with a high signal-tonoise ratio. Measurements are performed with a Kaiser Raman RXN 2 analyzer operated at 785 nm excitation wavelength, equipped with four probes to achieve simultaneous monitoring of four fermenters operated in parallel. The biocompatible probes, made from FDA-certified materials for body and window, are inserted through standard PG 13.5 ports and cleaned and sterilized in place before inoculation. The following data are available: • Continuously acquired Raman spectra representing two cultivation runs; first run performed in four parallel identical batches, second run containing three identical batches plus one with modified process conditions. • Reference values of nutrient concentration (determined on a BioProfile analyzer) and cell density (determined on a NucleoCounter, expressed in 106 cells per mL) obtained by sampling the fermenters every 2–3 days.

D. Engel and C. Minnich

derivative and a standard normal variate (SNV) normalization to emphasize the variance in the spectra (Fig. 48.16). With appropriate pretreatments applied to the spectra, a principal component analysis (PCA) supports the identification of potential spectral outliers. The PCA is conducted on the entire dataset, and scores in principal components #1 and #3 are selected for the assessment. Samples are color-coded batch wise, so that each batch is visualized as a trajectory in the 2D space (Fig. 48.17). The equivalence of all batches can be recognized (except for #7), without isolated points indicating outliers to the dataset. A confidence ellipsoid allows to visualize 95% coverage of the samples. The selected samples are then used to calculate the correlation between spectral intensity and the feature of interest at each spectral point. When this correlation is plotted it provides a preview of the most promising spectral regions to include into the model (Fig. 48.18). Here, the alcohol and carbohydrate range of the spectrum (1200–850 cm1) shows patterns of pronounced correlation with the nutrient content, positive as well as negative, which

The available dataset is large enough to include sufficient variation into the training step (120 samples), and at the same time leaving sufficient data for a sound test-set validation (49 samples):

50

20

0

Fig. 48.14 Assignment of the samples to the training (green) and test (blue) set. Batch 5 (red) remaining for separate validation

4

× 106 15

3 2

10

1

5

0 3000

2000

1000

Raman shift (cm–1)

Fig. 48.15 Raw Raman spectra from a typical cultivation batch

Batch time (d)

This selection procedure keeps five out of eight batches in the training set (#1, 3, 4, 7, and 8), two are intended to be the regular test-set (#2 and 6), and one remains as an extra challenge to the model at low concentrations (#5, Fig. 48.14). A quick glance into the raw spectra shows the need for baseline removal as the spectral background (which is increasing over time) is unlikely to be relevant for the prediction of the target parameters (Fig. 48.15). The signals visible by eye (e.g., around 3000, 1600, and 400 cm1) can be assigned to unspecific C–H contributions, water, and the probe optics and are therefore excluded from further considerations. The remaining fingerprint range (1600–500 cm1) requires the application of a first-order

30

10

Intensity

• Both runs are represented in the training and in the test set to prove run-to-run transfer. • Training and test samples cover the different probes to allow probe-to-probe transfer; full batches are defined as test batches. • Since many low-nutrient samples from the final batch phases are available, their number is reduced by excluding the lowest concentrated samples. Thus, biasing of the training set by focusing too much on the variance in the final phase of the batch is avoided. • The low concentration batch (#5, 20 samples) is reserved for model revalidation under new process conditions.

Nutrient (mM)

40

48

Chemometrics and Process Control

1087

Fig. 48.16 Plot displaying pretreated Raman spectra for eight cultivations, colored by nutrient content to highlight ranges of related variations

48

40 35

25 20 15

Nutrient (mM)

30

10 5

1600

1500

1400

1300

1200

1100

1000

900

800

700

600

500

Raman shift (cm–1)

100

6 0

4 3 2

–5

Correlation with nutrient

7 Batch no.

Score on PC 3 (5.27%)

8 5

50 0 –50 –100

1 –20

–10

0

10

1100

20

Score on PC 1 (61.79%)

6 RMSE (mM)

both can be exploited by PLS. The data are filtered accordingly to only include these correlative regions – this selection will be refined later. Next, alternative PLS calibrations are calculated up to a rank of 15, assuming that the optimal number of relevant principal components is somewhat smaller. Judging from the RMSEP, rank 7 is a viable candidate giving an RMSEP 99.99%)

20 True Predicted cell count (10^6/mL)

10

0

–10

Outlier probability

b)

48.4

2

4

6

8 Batch time (d)

10

12

4

6

8 Batch time (d)

10

12

1 0.99 0.98 0.97

2

Conclusion

Combining a powerful measurement technology with an intelligent quantitative data analysis forms a smart sensor which easily allows desired and unexpected insights into the physicochemical phenomena of catalytic processes. For the spectroscopic real-time monitoring of processes, materials, or catalysts, a versatile toolbox to access quantitative information is available. However, chemometrics cannot be done by hand on a calculator, but requires specialized software which supports the user in the analysis workflow. Depending on the desired process parameters, the step of extracting information requires more direct or indirect approaches, and the corresponding variety of methods is available to assist the user in achieving this. However, a clear preference should be given to the mechanistic approaches before considering statistical methods. • Peak integration is the most simple and straightforward way to generate a quantitative model even from few spectra. The lack of nonoverlapping peaks is a showstopper. • Spectral hard modeling solves the overlap and nonlinear changes of peaks while still maintaining the benefits of a mechanistic approach. It has proven valuable especially in extrapolation to new process conditions and model maintenance.

• PLS regression is extremely versatile since it does not require explicit appearance of the target parameter’s spectral characteristics in the spectra. Its limitation lies in the extrapolation of the model to conditions not covered by the training data. From the authors’ experience, these three methods give access to most of the applications in spectroscopic process monitoring. Few applications remain that would require more specialized approaches. The interested reader is referred to the respective primary literature.

References 1. Alsmeyer, F., Koß, H., Marquardt, W.: Indirect spectral hard modeling for the analysis of reactive and interacting mixtures. Appl. Spectrosc. 58, 975–985 (2004). https://doi.org/10.1366/ 0003702041655368 2. Alsmeyer, F., Marquardt, W.: Automatic generation of peak-shaped models. Appl. Spectrosc. 58, 986–994 (2004). https://doi.org/10. 1366/0003702041655421 3. Kriesten, E., Alsmeyer, F., Bardow, A., Marquardt, W.: Fully automated indirect hard modeling of mixture spectra. Chemom. Intell. Lab. Syst. 91, 181–193 (2008). https://doi.org/10.1016/j.chemolab. 2007.11.004 4. Kriesten, E., Mayer, D., Alsmeyer, F., Minnich, C.B., Greiner, L., Marquardt, W.: Identification of unknown pure component spectra by indirect hard modeling. Chemom. Intell. Lab. Syst. 93, 108–119 (2008). https://doi.org/10.1016/j.chemolab.2008.05.002

48

Chemometrics and Process Control

1091

5. Langanke, J., Wolf, A., Hofmann, J., Böhm, K., Subhani, M.A., Müller, T.E., Leitner, W., Gürtler, C.: Carbon dioxide (CO2) as sustainable feedstock for polyurethane production. Green Chem. 16, 1865–1870 (2014). https://doi.org/10.1039/c3gc41788c 6. Kern, S., Wander, L., Meyer, K., Guhl, S., Gottu Mukkula, A.R., Holtkamp, M., Salge, M., Fleischer, C., Weber, N., King, R., Engell, S., Paul, A., Pereira Remelhe, M., Maiwald, M.: Flexible automation with compact NMR spectroscopy for continuous production of pharmaceuticals. Anal. Bioanal. Chem. 411(14), 3037–3046 (2019). https://doi.org/10.1007/s00216-019-01752-y 7. Martens, H., Næs, T.: Multivariate Calibration. Wiley, New York (1989)

48

Clemens Minnich is a studied chemist, and holds a PhD in engineering from RWTH Aachen University, Germany. He has been using in situ spectroscopy (mid-IR, UV-VIS, and Raman) ever since and brought this experience into the foundation of S-PACT in 2010. His key responsibility is the development of chemometric models for customer applications all over the process industries and in related R&D activities.

Dirk Engel is a studied process engineer from RWTH Aachen University, Germany. His academic research focused on rigorous mathematical descriptions of chemical systems, including spectroscopic data. Cofounder of S-PACT in 2010, he is responsible for method and software development of the PEAXACT software for quantitative analysis of spectroscopic data.

Index

A Abbe diffraction limit equation, 702 Aberration, 412 correctors, 413 Aberration-corrected spectrometer, 136 ab-initio calculations, 822 ab initio XANES spectra calculations, 633 ab-plane, 425 Absorbance, 240, 265, 273 Absorption, 265 coefficient, 243 edge, 676, 679 rate constant, 274 Absorption-contrast, 675, 677 Accelerated degradation (ADT) tests, 575 Acetylene, 56 hydrochlorination, 420 Acid-base catalysis, 1061 Acidic behavior, 1043 Acidic/electrophilic molecules, 23 Acidic environment, 427 Acid sites, 795, 980 strength, 1043 Acoustic modes, 6 Acrylonitrile, 424 Activated carbon (AC), 832 Activated carbon fiber (ACF), 832 Activation energy (Ea), 984, 1037 Active phase(s), 339 dispersion, 640 Active site(s), 267, 681, 683, 980 titration, 915 Active species/structure, 970 Active zone, 900, 907 Acyl-enzyme complex, 993, 994, 1000 Acyl enzyme intermediates, 994 Acyl enzyme species, 993 Adenine, 179 Adiabatic calorimetry, 1032 Adsorbates, 984 Adsorbed formate, 1055 Adsorbed methyl, 1055 Adsorption/desorption, 981 Adsorption kinetics, 1042 Adsorption microcalorimetry, 1065 Advanced light source (ALS), 605 Aerogels, 829 Ag catalytic materials, 1016 27 Al, 759 Al2O3, 224 Al2O3/ZnO(0001)-Zn model, 479

Alcohol dehydration, 421 Algebraic reconstruction tomography (ART), 436 Alkaline environments, 427 Alkane(s), 138 isomerization, 421, 1067 27 Al MAS NMR, 765, 767, 768 Alpha-phase molybdenum carbide (α-MoC), 450 AlPO4, 816 AlPO4-5, 819 AlPO4-8, 818 AlPO4-18, 819 Alumina, 7, 11, 12, 17, 23, 28, 135 Aluminophosphate molecular sieves, 818 Aluminum-containing zeolite materials, 888 Ambient conditions/in situ condition CoMoO4, 86 controlled environment, 87 D band, 88 dehydration, 87 F2g mode, 88 2LO mode, 88 moisture, 87 MoOx, 87 Raman spectra, 87 reoxidation, 86 solid-state interaction, 86 sulfidation, 86 V2O5/CeVO4, 88 Ambient pressure studies, 71 Ambient pressure X-ray photoelectron spectroscopy (AP-XPS) active catalytic sites, 347–349 compositional restructuring, 352, 354 surface adsorbates, 349–352 Amide I, 997, 999, 1000 band position, 998 Amines, 20 Ammonia, 20 Ammonia adsorption microcalorimetry, 1042 Ammonium Metavanadate (NH4VO3), 1011 Ammoxidation, 423 Amorphous, 413, 672 Amorphous silica, 8, 20, 22 crystalline silicates, 8 Amplitude, 215 1-and 2-alkenes (dehydration), 1067 Ångstrom scale, 255 Annealing temperature, 66 Anti-Stokes Raman scattering, 136, 190 Anti-Stokes scattering, 76 Apparent rate constant, 911, 915, 918, 919, 925, 929 Applying bias, 367

© Springer Nature Switzerland AG 2023 I. E. Wachs, M. A. Bañares (eds.), Springer Handbook of Advanced Catalyst Characterization, Springer Handbooks, https://doi.org/10.1007/978-3-031-07125-6

1093

1094 Archaeology, 825 Argon, 81 ion, 81 Ar-ion laser, 81 Aromatic hydrocarbons, 139 Arrhenius-type relations, 984 Artificial intelligence, 683 Associated mechanism/formate, 954 Astigmatism, 412 Asynchronous, 972 Atom counting, 439 Atomic distribution, 643 Atomic force microscopy (AFM), 84, 170, 194 Atomic number contrast, 360 Atomic number (Z) dependent contrast, 450 Atomic probe microscope, 162 Atomic resolution electron tomography, 439 “Atop” platinum hydrides, 47 Attenuated total reflection (ATR), 992, 994–996, 999 infrared spectroscopy, 95, 994, 995 Attenuated total reflection Fourier-transform infrared (ATR-FT-IR), 250, 992, 995–999 Au/α-MoC catalyst, 452–454 Au clusters, 418 Auger electron spectroscopy (AES), 64, 218, 604, 1053 Au nanoparticles, 394, 395 Auto-correlation peak, 768 Automotive catalysis, 614–616 Average and surface EXAFS, 666 Average and surface structures differ significantly, 666 Average coordination geometry, 666 Average nanoparticle composition, 663 Average of atoms, 666

B Backdonation, 61 Backscattered electrons (BSE), 360 Band bending, 313, 325 Bandgap narrowing effect, 309 Bandpass filter, 163 Baseline correction/background subtraction, 206 Battery, 573, 574 Battery materials, 387 Beer-Lambert law, 630, 704, 705 Beer’s law, 54 Belousov-Zhabotinsky chemical oscillator, 852 Benchtop low-field nuclear magnetic resonance (LF-NMR) spectrometers, 1079 Bending magnet, 593, 594 sources, 590 Bending of the surface band structure, 301 Benzoic acid adsorption, 1046 Benzophenone, 310 Benzophenone ketyl radicals, 310 Beta, 142 β-O-4 linkages, 874 β-O-4 model, 880 BET method, 828 Bias, 1081 Bidentate nitrates, 36, 41 Bilinearity deviation, 633 Bimetallic, 679 alloys, 421 catalysts, 62, 660, 827 systems, 643

Index Bimetallic substrates alloys, 181 benzene, 181 core-shell structures, 181 cyclohexane, 181, 182 DFT, 182 hydrogenation, 181 palladium-catalyzed reactions, 182 platinum-catalyzed reactions, 182 STM topography, 182 Binding energies, 219 Biocatalytic processes, 993 Biomass, 1061 Biosensors, 826 Biphasic structure, 683 Biphenyl, 177 “Birth, life and death” of catalysts, 636 Bismuth molybdate, 100 Blends of polymers, 823 “Blind-source separation” method, 634 Bloch equation, 870 Blyholder model, 61 Boehmite, 6 Boltzmann distribution, 758 Bond distances, 661 Bond energy of methyl, 1055 σ-Bonding, 24 Bond strength of formate, 1056 Borosilicalites, 816 Bottom-up particle preparation, 636 Bragg-Brentano (reflection) geometry, 527, 528 Bragg CDI (BCDI), 681 Bravais lattices, 522, 523 Bremsstrahlung, 389 Bridge bound CO, 222 Bright field imaging (BF-STEM), 384 Brightness, 411 Brønsted acid/base sites, 12 Brønsted acid catalysis, 14 Brønsted acidic proton, 22 Brønsted acidity, 16 Brønsted acid sites, 779 Brønsted basicity, 311 Brønsted-Lowry acid sites, 797, 980 B1u symmetry, 57 Bulk CuO
Cr2O3
Fe2O3 mixed oxide, 488 Bulk Fe2(MoO4)3 mixed oxide, 1021 Bulk-like surface oxide, 64 Bulk metal catalysts, 490, 491 Bulk mixed metal oxide catalysts, 487–489 Bulk mixed oxide catalysts, 1019 Bulk MoO3, 1018 Bulk Nb2O5, 1019 Bulk oxide catalysts, 423 bulk V2O5, 1018 Fe2(MoO4)3 mixed oxide, 1021 mixed oxide catalysts, 1019, 1020 MoO3, 1018 Nb2O5, 1019 TeO2, 1019 V2O5, 1018 Bulk TeO2, 1019 Bulk V2O5, 1018 Bunsenite, 7 1,3-Butadiene, 141 1-Butene, 141

Index C C1 feedstock, 272 C3H6-TPSR experiments, 1024 Calcination, 140, 422 Calorimeter, 1031 Calorimetric isotherms, 1040 Calorimetry definition, 1031 effective sample volume, 1033 heating rate, 1032 heats of interactions, 1033 reaction heats, 1033 sensitivity, 1033 specific heat measurements, 1033 temperature range, 1032 Calorimetry-volumetry, 1039–1045, 1056 Calvet DSC, 1035 Candida antarctica, 994 lipase B, 994 Capillary reactor, 608 Carbanion, 1068 Carbide, 746 Carbocation intermediate, 1067 Carbon/coke, 123 Carbonate ions, 23 Carbonate mechanism, 954 Carbon capture and utilization, 1076 Carbon dioxide, 23, 343 Carbon dioxide reduction reaction (CO2RR) Ag, 204 Ag nanovoid, 204 multi-electron transfer process, 204 operando, 205, 206 operando SERS, 204 SERS, 206 silver, 205 Carbon materials, 831 Carbon monoxide, 22 cationic centres, 24 isotopically labeled 13CO, 24 low temperature CO adsorption, 24 metal oxide surfaces, 24 metal sites, 24 room temperature experiments, 24 Ru metal particles, 26 surface carbonyls, 25 surface science studies, 25 zeolitic internal surfaces, 24 Carbon nanotubes (CNTs), 1012 Carbonyl displacement, 776 Catalysis, 286, 287, 289 Catalysis, UV-Vis spectroscopy electrocatalysis, 259 heterogeneous catalysis, 254–255 homogeneous catalysis, 256–259 photocatalysis, 259 Catalyst(s), 410, 671, 757–759 advantages, 77 black-body radiation, 78 CCD detectors, 78 deactivation, 442, 1038 definition, 77 external standards, 79 fluorescence, 79 ignition, 854 instrumentation, 78

1095 internal standards, 79 laser irradiation, 78 laser sources, 78 laser spot size, 78 laser wavelengths, 78 particle, 672, 679 pellet, 674 Raman inactive, 79 structure, 987 Catalyst preparation, 636 colloidal particles, 638 supported catalysts, 645 Catalyst stabilization and deactivation, 649 Catalyst structure-activity relationships MoO3, 100 (MoVW)5O14, 100 propene oxidation, 100 real-time experiments, 100 Catalyst structure and composition vs activity, 649 Catalyst treatments butane dehydrogenation, 99, 100 ceria, 98 coke formation, 99 high temperatures, 97 oxidation TPO, 97 reduction condition/oxidation condition, 97 UV Raman, 98 V/θ-Al2O3, 98 V/Al2O3, 99 V2O5, 97 VOx on silica, 97 Catalytic active site, 1000 Catalytically active form, 419 Catalytic filamentous carbons (CFC), 832 Catalytic reactions, without plasmon contribution bimetallic substrates, 181 cobalt-phthalocyanine (CoPC), 181 cobalt-tetraphenylporphyrin (CoTPP), 180 CoPc surface, 181 metal phthalocyanines (PC), 180 phthalocyanine, 180 Catalytic reactions, with plasmon contribution chemical interface damping (CID), 173 chemical reactions, 174–177 definition, 171 hot carriers, 172–174 hot electron-hole pairs, 172–173 hot electrons, 172 plasmonic dephasing, 172 plasmon life time, 173 plasmon pump, 173 thermal effects, 173–174 thermalized electrons, 172 Catalytic system, 739 Catalytic triad, 993 Cation exchange, 819, 821 CAU-1-X derivatives, 830 Cement, 829 Ceria, 11, 145 CH4 ➔ CH3OH conversion, 343 Challenging nuclei, 788 Channeling effect, 451 Charge carrier dynamics, 323–326 Charge carrier trapping, 323 Charge-coupled detectors (CCD), 82 Charge coupled devices (CCD), 91, 190

1096 Charge generation, 305 Charge separation, 312 of electrons and holes, 312 Charge transfer complexes, 239 Charge transfer excited triplet state, 303 Charge transfer transition, 268 Chelating nitrates, 39, 40 Chemical enhancement (CT), 152 Chemical gradients, 673 Chemical imaging, 671 Chemical-interface damping (CID), 172 Chemical mechanism (CT) charge transfer, 155 chemical SERS effect, 155 HOMO, 155 LOMO, 155 Raman-scattering efficiency, 155 Chemical response, 746 Chemical shift, 814, 828 Chemical shift anisotropy (CSA), 817, 818, 825 Chemical tomography, 683 Chemometric(s), 1074 analysis, 276 and multivariate analyses, 251–254 4-Chloro-phenylisocyanine, 182 Chromatic aberration, 413, 414 Chromatographic effects, 942 Chromyl, 14 CHx species, 343 C-Hyd-VSiβ zeolite, 307 cis/trans-2-butene, 141 Classical Nucleation Theory (CNT), 636 Clays, 824 Closed cell, 441 Cluster(s), 69 computing, 698 c-MES, 996 CMK-3 carbon material, 832 13 C NMR spectra, 859 CNPt-V to CNPt-Pt, 665 Co/TiO2, 103 CO2 activation, 60 CO2 adsorption calorimetry, 1044 CO2 electroreduction, 375 CO2 hydrogenation, 340, 341 to methane, 10 CO adsorption, 221 calorimetry, 1045 Cobalt, 577 binding energy, 1055 catalysts, 949, 951 hydrogenation, 936, 944, 947, 948, 951, 953 molecules, 394, 395 nanoparticles, 9 oxidation, 222, 340, 745 particle size, 951 as probe molecule, 324 Cobalt-based perovskite catalysts, 579, 580 Cobalt phthalocyanine (CoPc), 180, 181 Cobalt-tetraphenylporphyrin (CoTPP), 180, 272 Coefficient of determination, 1081 Coherent anti-Stokes Raman scattering (CARS) microscopy, 84 Coherent diffraction imaging, 575, 681 Coherent X-ray, 672 Coherent X-ray diffraction imaging (CXDI), 714 Coke formation, catalyst deactivation

Index butane, 141 carbonaceous deposits, 141 dehydroaromatization, 143 excitation lasers, 142 graphitic carbon, 142 methanol dehydration, 143 reactor geometry, 142 Ti-species, 144 Coke formation, 277, 651 Coke species, 277 Coking, 823 Cold FEG, 411 Column-by-column fashion, 418 Column-wise augmented (CWA) data strategy, 634, 636 Combined XAS/FTIR/XRD, 750 Complementary characterization, 752 Complex mixed oxide, 425 Complex structures, 818 Composites, 819, 828 Computational fluid dynamics (CFD), 114 Computational modeling, 125, 768, 771 Computational techniques, 120 Computed tomography (CT), 573, 672, 690, 691, 696, 697 Concentration modulation, 970, 975, 987 Confocal Raman, 78 Co-Ni/α-MoC catalyst, 453 Contact time, 592 Continuous flow, 577 Continuous-flow hyperpolarization (CF-HP) methods, 824 Continuously stirred tank reactor (CSTR), 937–940, 943, 944 Contrast reversal, 416 Conventional temperature programmed studies, 1007 Convergence semi-angle, 411 Convergent beam electron diffraction (CBED), 415 Conversion of 2-propanol, 1063 Conversion of 4-methylpentan-2-ol, 1067 Coordination number, 661, 665, 1062 Coordinatively under-saturated (CUS), 224 Coordinatively unsaturated vanadyl species, 984 Co-phthalocyanine (CoPc), 180 CO-poisoning, 223 Copolymerization of epoxides, 1077 Copper, 266, 267, 590, 980 Copper-based catalysts, 267 Copper
chromium
iron mixed oxide catalyst, 488 Copper ion-exchanged chabazite zeolites (Cu-CHA), 35 Core electron shells, 602 Core-loss EELS, 390 Core-shell, 678 structures, 663 Corundum, 7, 12 CO-TPR, 1014 Covalent organic frameworks (COFs), 270 Covalent oxides, 12 C1 processes, 339 Cr alloy formation, 662 Cross-validation techniques, 1082 Cryptophane, 826 Crystal field geometry, 239 Crystal field theory (CFT), 239 Crystalline, 672 solids, 522–525 Crystalline non-conducting solids active vibrational modes, 6 aluminas, 6 asymmetric molecules, 6

Index crystalline solid, 6 factor group, 6 inactive modes, 6 NiAl2O4 spinel, 7 NiO phase, 7 smallest Bravais cell, 6 space group, 6 Crystallinity, 818 Crystallographic facets, 1064 Crystallography, 541, 542, 552, 556, 557 CT excited state, 306 Cu/Al2O3/ZnO(0001)-Zn model, 477 Cu/CeO2, 341 Cu/ZnO/Al2O3 catalysts, 592 Cu/ZnO(0001)-Zn model, 478 Cubic perovskite, 7 Cu-chabazite, 34 Cucurbiturils, 160 Cu mordenite (Cu-MOR) catalysts, 645 Cumulative variance (CVE), 275 CuO
Cr2O3
Fe2O3 catalysts, 487, 1014 CuO
Cr2O3
Fe2O3 mixed oxide catalyst, 487 Curie law, 872 Cu-SSZ-13, 983 Cu-Zeolite, 893 Cycling conditions, 745, 748 Cyclopentadiene, 141 Cyclovoltammetry, 178

D Damage threshold, 429 Damping, 215 Dark-field imaging, 384 Data collection strategy, 673 3D characterisation, 439 d-d charge transfer (CT) transitions, 266 Deactivation mechanisms, 654 Deactivation periods, 1067 Dead-layer model, 301 Dealumination, 679, 768 Debye-Scherrer (transmission) geometry, 528, 529 Debye-Waller modifications, 633 Decay, 313 Decomposition reaction, 1037 Dehydrated catalysts, 788 Dehydrated materials, 773 Dehydration, 1061, 1068 Dehydroaromatization, 777 Dehydrogenation, 742, 1068 process, 44 d0 electronic structure, 821 Delithiation process, 776 Demodulation, 970, 984 index, 970 DeNOx reaction, 102 Density functional theory (DFT), 71, 201, 239, 248, 250–253, 439, 944, 952, 954, 1055 ambient/ex-situ conditions, 121 calculations, 120, 121, 666, 1000 intricate pore channels, 120 methods, 547, 549 MoOx moieties, 122, 123 zeolite catalysis, 121 zeolites, 120 Depletion region, 326

1097 Depth-of-afield (DOF), 437 Depth of focus, 415 Depth sectioning, 373, 437 Design of experiments (DOE), 1085 Desorption of surface oxygen species, 1016 Detection limit, 370 Detection system, spectrometer, 82 Detector, 82, 192 Determination of the number of catalytic sites, 664 Determination of the turnover rates, 664 Deuteration, 12, 998 Dibenzo(1,2)dithiine-3,8-diamine (D3ATP), 177 Dichroic beam splitter, 163 Diesel oxidation, 373 Difference EXAFS, 660, 661, 663–666, 668 Difference spectra, 973 Difference spectroscopy, 968 Difference spectrum, 660, 663, 664, 981 Difference XAS, 661 Differential heat, 1041 Differential micro-calorimeter, 1039 Differential scanning calorimetry (DSC), 1006–1011, 1056 calvet DSC, 1035, 1036 catalytic applications of, 1036 coupled DSC/TGA unit, 1037, 1039 curve, 1034, 1035 heat flux, 1034 power-compensated DSC, 1035 Differential thermal analysis (DTA), 1006–1010, 1012, 1013 Differential thermogravimetric analysis (DTG), 1006, 1008, 1009 Differential TPO, 1012 Diffraction, 971, 972, 974 limit, 413, 702 Diffraction-limited probe size, 411 Diffusive reflectance (DR), 268 spectroscopy, 241, 243, 252, 313 technique, 5 Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), 95, 266, 740, 742, 744, 746–748 Diffuse reflectance infrared (DRIFT) spectra, 980, 981, 986 Diffuse-reflectance infrared spectroscopy (DRIFTS), 690 Diffusion limitations, 903 5,5-Dimethyl-1-pyrroline (DMPO), 877 Dimethyl ether (DME), 678 Diode laser, 81 Direct formation, 925 Direct methane to methanol (DMTM) conversion, 645 Dispersion, 418, 660, 664 Dispersive optics, 593 Displacement field, 683 Dose, 427 Double frequency sweep (DFS), 778 Double-quantum MAS NMR (DQ-MAS), 768 3-D printed cell, 584 3D reconstruction, 434, 435 DRIFT experiments, 34 DRIFT/XAS/MS, 743 Dry reforming of methane, 342 4D-STEM techniques, 456 DUT-8(Ni), 830 d-π* electron backdonation, 24 dx electronic structure, 822 Dynamic(s), 305 of charge carriers, 312 of hole transfer, 313 of MOF linkers, 801

1098 Dynamic(s) (cont.) of photocatalysis, 312 of photogenerated electrons and holes, 312 redox, 749 restructure, 644 Dynamic nuclear polarization (DNP), 778, 790, 804, 882

E Edge filter, 136 EFA (Evolving Factor Analysis) method, 628, 631 Effective acid-base properties, 1046 Effective sample volume, 1033 Effect of temperature on photoluminescence spectra, 314–316 Egg-shell distribution, 674 Electric field, 815 Electric-field gradient (EFG), 759, 821, 829 Electrocatalysis, 259, 578 catalyst, 198 definition, 198 OERS, 198 operando, 198 Electrocatalysts, 587, 638 Electrocatalytic energy conversion Cu-based materials, 199 HER, 200 identification of active sites, 200, 201 MoSx materials, 200 ORR, 199 SECM, 199 SG/TC SECM mode, 200 storage, 198 water electrolysis, 198, 199 Electrocatalytic processes, 618 Electro-catalytic study, 587 Electrochemical cells, 196, 197 Electrochemical conversions, 773 Electrochemical cycling, 578 Electrochemical processes covalently bound Nile Blue, 179 cyclovoltammetric conditions, 178 diffusion processes, 179 EC-AFM-TERS, 179 EC-STM-TERS, 179–180 EC-TERS, 178 electrical potential, 178–180 metal phthalocyanine, 180 molecular desorption, 179 Nile Blue, 178 protonation reactions, 179 redox reactions, 178–180 substrate potential, 180 water cleavage, 179–180 Electrochemical reduction, 582 Electrochemical systems, 579 Electrochemistry, 196, 578 Electrode, 178, 196 electrolyte interface, 190, 196 Electrolyte, 196, 584 Electrolytic graphite oxide (EGO), 1037 Electromagnetic effect (EM), 152 Electromagnetic hot spots, 155 Electron affinity, 311 Electron-atom scattering, 566 Electron backscatter diffraction (EBSD), 371–372

Index Electron density, 683 Electronegativity, 301, 311 Electron energy-loss spectroscopy (EELS), 384, 389–392, 415 Electron–hole recombination, 324 Electronic, geometric and surface properties, 752 Electronic and structural properties, 740, 742 Electronic excited singlet state (S1), 297 Electronic transitions, 265 Electron-matter interactions, 415 Electron microscopy, 681 Electron paramagnetic resonance (EPR), 306 frequency, 778 information from analysis, 889–891 signal intensity, 888 time resolved EPR spectral intensity, 891 Electron paramagnetic resonance (EPR) spectroscopy in situ electrochemistry and, 881, 882 future of, 882, 883 hardware, 872–874 metals and free radicals, 880, 881 spin labeling and isotopic substitution, 878, 879 spin trapping, 875, 877 theory, 870, 871 transition metal centers, 888–889 Electron-phonon scattering, 172 Electron transfer, 577 Electro(photo)catalysis, 163 Electroreduction of CO2 (CO2RR), 270–272 Elemental analysis, 369–371 Element sensitivity of XAS, 643 Eley-Rideal-type mechanism, 980 Ellipsoidal mirror, 136 Emission spectrum, 303 Emitting materials, 301 Enantiomers, 992 Energetics, 1055 Energy-dispersive, 573 polychromators, 566 X-ray, 608 Energy-dispersive spectroscopy (EDS), 370–371 Energy dispersive X-ray absorption fine structure (ED-XAFS), 125 Energy dispersive X-ray spectroscopy (EDS), 383, 389–392 Energy levels, 415 Energy of the unfilled Pt 5d orbitals, 666 Energy separation, 304 Energy spin states, 759 Energy storage, 773, 774, 776 Energy transfer, 299, 666 Energy transfer process, 305 Enhancement factors (EF), 193 Enthalpy of immersion, 1047 Enthalpy of neutralization, 1048 Environmental holders, 382, 386–388, 396–403 Environmental STEM, 442 Environmental TEM, 441 Environmental transmission electron microscopes (ETEM), 382, 384–386, 392–397 ESEM, 375 Esterification, 853 Ethanol, 28 Ethylene, 56 epoxidation, 375 oligomerization, 857 oxidation, 1016 propylene EPR rubber, 823 Ethylidyne, 56, 57

Index Euler angles, 760 Eulerian angles, 156 Even-odd effect, 227 Evolving Factor Analysis, 275 Excitation, 296, 302 Excited triplet state, 299, 314 Exciton binding energy (Eb), 316 Exhaust gas after-treatment, 979 Ex-situ and in-situ/operando modes, 613 Ex situ Electron Nuclear Double Resonance (ENDOR), 880 Ex situ TEM, 382 Extended X-ray absorption fine structure (EXAFS), 332, 398, 566, 570, 575, 583, 592, 604, 708, 740, 743–746, 748, 750 spectra, 973 Extended X-ray fine structure, 602–605 equation, 604–605 fit, 604 Extra-framework, 768 Extra-framework aluminum (EFAL), 763, 796

F Factor analysis (FA), 748 Faujasite zeolite, 9 Fe+1 compound, 872 Fe-Mo-O catalysts, 97 FeN4, 429 Fermi level, 172 Fermi resonance phenomenon, 38 Ferrierite (Fer), 818 Ferromagnetism, 870 FEXRAV, 584 Fe-ZSM-5, 980 Fiber optic coupling, 136 Fiber-optic probes, 82 Fiber optics spectroscopy, 245 Filled and unfilled states, 666 Filtered back projection (FBP), 698 Fingerprint spectrum, 267 Finite-difference time-domain (FTDT), 182 Finite difference time domain-finite element method (FDTD-FEM), 176 First-order rate constant, 939 Fischer-Tropsch catalysts, 588 Fischer-Tropsch synthesis (FTS), 90, 103, 480, 532, 533, 674, 678, 713, 728, 951 catalysis, 616, 619 catalyst particle, 610 Fitting the difference spectra, 665 Five dimensional (5D), 673 Fixed-bed quartz-microreactor, 86 Fixed bed reactor, 100 Fixed energy X-ray absorption voltammetry (FEXRAV), 582 Flexibility, 829, 830, 997 Fluid catalytic cracking (FCC), 8, 679, 683, 708 particle, 610 Fluorescence, 296, 298 detection, 577 mode, 741 Fluorescence-yield EXAFS, 584 Fluorescence-yield XAS, 581, 582, 584 energy-dispersive polychromator, 568 perovskite, 568 photo-catalysis, 567 planar silicon diode (PIPS) detector, 568 self-absorption effects, 568 spatial dimension, 568

1099 step-scan approach, 568 step-scan methods, 567 time-resolution, 568 time-resolved, 567 XANES and EXAFS, 568 o-Fluoro-nitrobenzene, 1079 Focused ion beam scanning electron microscopy (FIB-SEM), 696, 699, 716 Focusing lens, 82 Formation kinetics of zeolites, 819 Formation of Pt-O bonds, 661 Fourier transform, 663, 970, 981 Raman, 78 technology, 1077 Fourier transform infrared (FTIR) gas cell, 113 Fourier transforms of extended X-ray absorption fine structure (FT-EXAFS), 306 Fourth-generation synchrotrons, 683 Fragmentation model, 679 Franck-Condon principle, 297 Free electron gas, 163 Free electron lasers (FELs), 606 Free induction decay (FID), 758, 1079 Frequency, 215 FT-IR spectroscopy BM23/ID24, 35 Brønsted sites, 40 Cu-CHA catalyst, 36 CuI sites, 39 Cu-nitrates, 35, 36 EXAFS evidence, 36 experimental set-ups, 34 heterogeneous catalysts, 33 in situ, 33 isotopic exchange, 41 isotopic labelled 15NO, 41 operando, 33 reaction temperature, 40 reactor cell, 34 FT-Raman spectroscopy, 83 Fuel cell, 677 configurations, 578 Fuel hydrogen, 201 Full-field approach, 573 Full-field imaging, 699–701, 703, 710 Full width at half maximum (FWHM), 57 intensity, 525

G Gas adsorption, 861 calorimetry, 1039, 1042 Gas chromatograph (GC), 113, 241, 592 Gas-liquid-solid processes, 854, 855 Gas-liquid-solid system, 112 Gas phase hold-up, 941 Gas-solid catalysis, 567, 591 Gas-solid interactions, 1039 Gas-solid processes, 855–857 Gas-solid system, 112 Gaussian distribution function, 1079 g-C3N4, 326 Geometric aberration, 411, 413, 414 Geometry and electronic properties, 665 Glass transition, 823 Glassy polymers, 824

1100 Graham’s law, 913 Grains, 676 Granular catalyst beds, 854 Graphene monolayer, 68 Graphene oxide (GO), 1009 Graphical processing units (GPU), 698 Grating monochromator, 82 Gyromagnetic ratio, 758, 763

H Hard coke, 143 Hard X-rays, 691, 693, 704 sample all atoms, 666 Harmonics, 4 H/D exchange, 999 H/D isotopic exchange, 997 of water molecules, 996 Heat-balance reaction calorimeter, 1051 Heat exchange calorimetry, 1032 Heat flow, 1034 reaction calorimeter, 1049, 1051 Heat flux DSC, 1034, 1035 Heating, 427 rate, 1032 Heats of interactions, 1033 Heavy atom effect, 298 HeCd laser, 81 Hectorite, 824 Helium/Neon, 81 Helium-cadmium (HeCd) laser, 135 α-Helix, 998 HEROS (High Energy Resolution Off-Resonant Spectroscopy), 568 Hertzian dipole, 193 Heterogeneity, 286–287 Heterogeneous catalysis, 71, 254–255, 449, 741, 787, 833 ethoxy groups, 28 gas-phase spectra, 29 in situ studies, 27 spectrum, 27 Heterogeneous catalysts, 242, 638 automotive catalysis, 614–616 hydrogenation catalysis, 616, 617, 619 operando Raman-GC, 113 operando spectroscopy, 113 photons, 111 solid-fluid interaction, 113 spectral requirements, 114 Heterogeneous catalytic hydrogenation, 849 Heterogeneously catalysed reactions applications, 126 carbon formation, 123 DFT methods, 125 operando experimental techniques, 126 Raman spectra, 125 Heterogenous catalysis, 387 Heterojunction, 326 Heteronuclear correlation experiments (HETCOR), 792 Heteropolyacid, 12, 825 Heteropolyoxometalate salts, 825 Hexagonal channels, 423 Hgh-spin coordination, 583 1 H-1H dipolar coupling, 765 Hierarchically structured materials, 832 High-angle annular dark field (HAADF), 410, 450 High-angle annular dark-field (HAADF)–STEM, 384

Index High-energy transitions, 238 Higher harmonics, 974 Highest occupied molecular orbital (HOMO), 155, 172 High-field NMR experiments, 795 High-field NMR spectroscopy catalyst characterization, 757–759 energy storage, 773, 774, 776 high Field 27Al MAS NMR, 765, 767, 768 low-natural abundance, gamma nuclei, 776–778 for MOFs, 798–801, 803, 804, 806 quadrupolar nuclei, 759, 760, 762–765 supported catalysts, 793–795 vanadium oxide characterization, 769, 771–773 of zeolites, 795, 796 Highly ordered substrates aggregation stages, 158 colloidal silver, 158 nanoparticle supercrystals, 161 NIR, 158 optical field, 158 polarization, 158 SERS enhancement, 158 whispering gallery modes, 159 High resolution-transmission electron microscopy (HR-TEM), 742 High-sensitivity low-energy ion scattering spectroscopy (HS-LEIS) active component, 466–468 AMn2O5 (A ¼ Sm, Bi) oxides, 480 catalytically active surface sites, 470–474 catalytic reaction, 462 depth information, 465 FeMoTeO mixed oxides, 480 FTS, 480 fundamentals, 462–465 location and orientation, 462 Mo-V-O and Mo-V-Te-Nb-O phases, 480 nanocatalysts, 473–476 outermost surface layer, 462 Pd(111) surface, 481 photocatalyst systems, 481 pre-treatments, 465, 466 quantification, 465 solid catalysts, 461 strong metal-support interaction, 475–480 supported bimetal catalysts, 466–472 High spatial resolution, 430 High temperature water-gas shift (HT-WGS), 487 High-throughput calorimetry, 1039 Hindered translational modes, 6 2 H NMR technique, 800 H2O2, 853 Hollow cathode technology, 136 Holographic notch/edge filter, 82 Holography, 700 Homemade quartz cell, 34 Homogeneous catalysis, 255–259 Host-guest interactions, in MOFs, 804 Hot carriers, 172–174 Hot-electron based catalysis, 172 Hot electrons, 171–172 Hot holes, 171–172 Hot spots, 170, 193 H-Rho, 818 H-SAPO-34, 272, 273 H-SSZ-13, 272, 273 H2-TPR, 1013, 1014 HT-WGS, 488

Index Hybrid material chemical affinity, 160 definition, 160 electrostatic charge, 160 host-guest interaction, 160 hybrid platform, 160 optical accumulators, 161 porous silica, 160 Hybrid statistical-simulation method, 439 Hydrides, 745 Hydrocarbon(s), 951 acetylene hydrogenation, 57 ethylidyne, 59 gas-phase and surface species, 57 heterogeneous catalysis, 55 D2h point group, 57 hydrogenation of acetylene, 57 oligomers, 57 propylidyne moieties, 60 reactions, 98 turnover frequencies, 60 UHV, 57 Hydrocarbon pool (HP) hypothesis, 272 Hydrodesulfurization (HDS), 429 Hydrogenation, 679, 741, 742, 744 catalysis, 616, 617, 619 ethylene, 857, 863 α-methylstyrene, 854, 855 1-octene, 855, 858, 863 propene, 857 Hydrogen bonding activity, 13 Hydrogen evolution (HER), 583 Hydrogen evolution reaction (HER), 259 Hydrogen storage, 774 Hydroprocessing catalyst, 92 Hydrotalcite, 12 Hydrothermal degradation, 679 Hydroxyl groups, 20 Hydroxyl (-OH) species, 115 Hyperfine, 871 Hyperpolarization, 849 Hyperpolarized 129Xe gas, 864 Hyperpolarized xenon (HP-Xe), 814, 824 Hyphenated Raman characterization techniques, 125 Hysteresis, 830 H-ZSM-5, 1015 I Ice-nucleation, 227 Identification of active sites, 650, 664 Identify the surface composition, 663 Ilmenite, 7 Image contrast, 360, 852 Image resolution, 363–364 Image simulation, 431 Immersion calorimetry, 1047 Incoherent inelastic neutron scattering (INS) spectroscopy, 42 Indirect hard modeling (IHM), 1078 Induction periods, 1067 Industrial solid acid catalysts, 16 Inelastic, 190 mean free path, 566 scattering, 415 Inelastic neutron scattering (INS), 742 heterogeneous catalysis, 500–502 principles, 499, 500

1101 Infinite layer thickness, 245 Infrared (IR), 741, 750, 751 Raman method, 592 spectra, 981, 987, 997 Infrared (IR) spectroscopy, 53, 246, 992 catalytic phenomena, 4 crystalline non-conducting solids, 6 emission, 5 probe molecules, 16 pure catalyst powders, 10 reflection, 5 solids characterization, 5 surface chacracterization of catalyst, 10 surfaces, 3 vibrational modes, 5 Inorganic halide perovskite quantum dot (IPQD), 315 In-phase angle(s), 975 analysis, 970 “In-phase” condition, 970 Insertion device-based resources, 594 In situ, 869–873, 875–882, 975 cells, 740 characterization, 341 FT-IR spectroscopy, 43, 334 gas experiments, 387 heating, 433 investigations, 327 liquid experiments, 387 manipulation, 394, 396, 397 MES ATR-FTIR techniques, 1001 operando, 671 and operando Raman spectroscopy, 125 and operando spectroscopy, 613 photoluminescence spectroscopy, 334 SEM, 375 series, 442 spectra, 112 study, 306–312 synchrotron XRD, 663, 666 transmission electron microscopy, 382 UV-Vis and IR spectroscopies, 328 X-ray absorption spectroscopy, 980 In situ/Operando measurement γ-Fe2O3, 91 iron-based catalyst, 90 metal oxide catalysts, 89 operando Raman cells, 90 propane ammoxidation, 89 Sb-O catalyst, 89 structure-selectivity relationship, 91 In-situ Raman spectroscopy, 422 heterogeneous catalysts, 112, 113 laser sources, 112 operando, 112 Raman cell, 112 second-generation characterization, 112 techniques, 112 In-situ (S)TEM experiments, 455 In situ X-ray diffraction (XRD), 661 Instrumental advances, Raman spectroscopy fluorescence effect, 83 increase sensitivity, 83 spatial resolution, 84 Integral analysis, 690 Integral heats, 1040 Integrated differential phase contrast (iDPC), 433

1102 Integration Model, 1076 Intensity, 416 Interaction volume, 364 Interfacial activation, 999 Interfacial clustering, 683 Interferometer, 82 Intermediate, 992, 1001 Intermediate-voltage, 413 Intermetallic alloy, 661, 662, 665, 667 Internal conversion, 300 Internal reflection elements (IRE), 994 Internuclear distance, 297 Intersystem crossing, 298 Intrinsic, 900, 921, 924, 929 activities, 420 radiationless transition, 309 Inverse Laplace transform (ILT) method, 943 Inversion symmetry, 216 Ionization damage, 426 Ionization potential, 301 Ionizing radiation, 590 IR absorption spectroscopy, 6 IRMOF-1, 830 Iron, 574, 579, 581, 980 molybdate, 96 Iron phthalocyanine (FePc), 180 Irreversible loss, 206 Irreversible reaction, 911, 912 IR spectroscopy, 324, 741 Isocyanide, 182 Isomerization of 1-octene, 859 of n-butane, 1067 Isoperibolic calorimetry, 1032 Isoperibolic differential reaction calorimeter, 1052 Isoperibolic reaction calorimetry, 1049 Isostatic polypropylene, 679 Isothermal calorimetry, 1032 Isothermal reaction calorimeters, 1049, 1051, 1052 129 Isotope, 814 Isotope effect, 955, 957 Isotopic, 916 Isotopically labelled SCR, 1025 Isotopic exchange, 995 Isotopic labeling, 1007, 1016 Isotopic shift, 14 Isotropic chemical shifts, 762, 765, 768, 771, 773 Isotropic radiation scattering, 245 Iterative methods, 698

K Kaolin, 8 Kaolinite, 8 KBr pressed discs, 5 K-edge spectrum, 602 Keggin-type polyoxometalate, 893, 894 Ketone condensation, 1067 Ketones (dehydrogenation), 1067 Ketoprofen, 994, 1001 Kinetic, 577, 583, 972, 973, 986 of adsorption/desorption, 1009 analysis, 973 coherency, 926 curves, 331 information, 997

Index modelling, 327, 331 Petals, 902, 915, 925 resolution, 992 of transformation, 1034 Kinetic isotope effects (KIE), 944, 951, 954, 1026 Kirkendall effect, 399, 400, 402 Knight shift, 777 Knock-on Damage, 426 Knudsen, 900, 901, 903, 904, 906 diffusion, 904 Kramer-Heisenberg-Dirac (KHD), 132 Krypton ion, 81 Kubelka-Munk-Schuster (KMS) theory, 243 Kubelka-Munk theory, 80

L Lab-based techniques, 613 Lab-based X-ray diffraction (XRD), 606 Lab-based X-ray spectroscopy methods, 606, 607 Lab-on-a-chip type microreactors, 861 Laboratory based spectrometers, 594 Laboratory-based X-ray spectrometers, 587 Lambert-Beer’s law, 240 LaMer model, 636 Langmuir-Hinshelwood, 1022 fashion, 980 mechanism, 926 Langmuir-Hinshelwood-Mars-van Krevelen (L-H-M-V-K) reaction mechanism, 1025 Langmuir-type adsorption isotherm, 330 Larmor frequency, 758 Laser(s), 191 Raman spectroscopy, 80 system, 216 Lattice vibrations, 984 Lewis acidity, 16 Lewis acids, 18 sites, 980, 983, 986, 987 Lifetimes, 279, 298, 304 Light polarization, 218 Light scattering, 190 LiH powder, 774 LiNbO3, 7 Linear combination fitting (LCF), 614, 636 Linear Pt-hydrides, 45, 48 Lipase(s), 992, 993, 997, 999, 1001 Lipase B of Candida antarctica (CALB), 993, 994, 996, 999 Lipid bilayers, 226 Liquid cells, 195 Liquid nitrogen temperature, 314 Liquid phase calorimetric technique, 1056 Liquid phase calorimetry, 1056 effective acid-base properties, 1046 heat of adsorption, 1047 immersion calorimetry, 1047 isoperibolic differential reaction calorimeter, 1052 isothermal reaction calorimeters, 1051, 1052 reaction calorimetry, 1049, 1051 titration calorimetry, 1045 Liquid-solid processes, 853 Lithiation/delithiation, 574 LMCT transitions, 266 Loadings, 1083 Localized surface plasmon (LSP), 170 Localized surface plasmon resonance (LSPR), 152, 153, 193

Index Local structure, 661 Location of nanoparticles, in pores, 373 Longitudinal optical mode (LO), 7 Long-range order, 661 Lorentzian distribution function, 1079 Low-dose STEM-HAADF, 433 Low energy electron diffraction (LEED), 64, 218, 1053 Low energy ion scattering spectroscopy (LEIS) bulk metal catalysts, 490, 491 bulk mixed metal oxide catalysts, 487–489 photocatalysts, 485–487 supported metal oxide catalysts, 489, 490 Lowest unoccupied molecular orbital (LUMO), 155, 172 Low gyromagnetic ratio, 777 Low-γ nuclei, 776 Low-temperature, 419 aqueous-phase reforming of methanol (APRM), 450–452 CO oxidation, 419 WGS reaction, 420, 452–454 Low-voltage, 429 STEM imaging, 427 LTA-type zeolite, 121 LTA zeolite, 510 Luminescent probe molecules, 296

M Machine learning, 698 algorithms, 683 Machine learning (ML)-based approaches, 404 Macromolecular crystallography, 590 Magic-angle spinning, 762 Magnetic field gradients, 852 Magnetic resonance imaging gas-liquid-solid processes, 854, 855 gas-solid processes, 855–857 homogeneous vs. heterogeneous catalysis, 852, 853 liquid-solid processes, 853 operating reactor, 854 signal enhancement, 858–861 thermometry, 862, 863 Magneto-photoluminescence, 317 Mahalanobis distance (MD), 1088 Mammalian cell culture, 1085 Mars-van Krevelen, 1022, 1023 mechanism, 1018 redox mechanism, 9 Mass flow controllers (MFC), 944 Mass spectrometer (MS), 116, 241 Mass spectrometry (MS), 113, 592, 593, 746 Mass transport, 579 MCM-41, 828 MCM-48, 828 Mean free path, 815 Mean free space, 815 Mean residence time, 947, 950 Measurement under Pressure, 1032 Mechanistic analysis method, 1074 Mechanistic methods, 1075, 1080, 1083, 1085 ME experiments, 984 MEMS-based heating, 441 4-Mercaptopyridine (4-MPY) molecules, 177 Mercury-arc lamp, 80 MES ATR-FTIR spectra, 1000 Mesopores, 819

1103 Mesoporous silica, 367 Metabolites, 1085 Metakaolin, 8 Metal(s), 14 cluster, 439, 826 dispersion, 639 hydride, 741 particles, 828 phthalocyanines, 180 single crystal, 219 tetraphenylporphyrin, 180 Metal distribution within bimetallic particles, 646 formation of phases and interaction with support, 643 Metal-loaded MOF, 804 Metal-organic, 161 Metal-organic frameworks (MOFs), 309, 749, 751, 777, 798, 829, 830 nanosheets, 799 Metal-oxygen stretching, 14 Metal-to-ligand charge-transfer (MLCT) band, 258 Methane, 121, 339 activation and conversion, 341, 343, 344 dry reforming, 63 reforming reactions, 63 Methanol, 14 chemisorption, 1017 synthesis, 340, 341, 592 Methanol adsorption, Pd blue shift, 60 formate, 60 hydrogen-bonded clusters, 59 redshift, 59 Methanol to gasoline (MTG), 503, 542 Methanol to hydrocarbons (MTH), 141 conversion, 542 Methanol-to-olefins (MTO), 272–279, 542 reaction, 242, 247, 254 Methylbenzenium, 143 1‑Methylnaphthalene, 135 MgCl2 Ziegler-Natta catalyst, 777 MgO, 777 Mg vanadates, 12 Michelson interferometer, 163 Microelectromechanical system (MEMS), 386, 727 reactor, 607, 608 technology, 455 Microfluidic cell, 375, 994 Microporosity, 825 Microporous nanocomposites, 825 Microporous polymer networks, 824 Microreactors, 861 Micro-spectroscopy, 241, 245, 256 Microtomography, 691, 694, 702 Mid-infrared (MIR) spectra, 1077 Migration, 305 Miller indices, 523, 524 Missing-wedge, 437 Mixed-linker MOFs, 803 Mixed-metal MOFs, 804 M-N-C catalysts, 429 Mo/HZSM-5, 143 Mo/SiO2 catalyst, 102 Mo/ZSM5 carburized catalyst, 124 Mo6+/SiO2, 328, 330–332 Model catalyst, 63 Modulated ATR-FTIR spectroscopy, 1001

1104 Modulation excitation spectroscopy (MES), 651, 968–972, 975, 979, 980, 982, 983, 988, 994 definition, 94 experiments, 996 modulation and phase sensitive detection, 968, 970 spin-crossover material, 95 stimulation, 94 time-resolved Raman spectra, 96 Modulation excitation spectroscopy with phase sensitive detection (MES-PSD), 992, 996, 997, 999 acyl enzyme intermediate, 993 attenuate total reflection infrared spectroscopy, 994 biocatalyzed kinetic resolution of racemic profens with lipases, 992 enzymatic kinetic resolution of racemic profens, 993 isotopic exchange H-D of the enzymes with D2O, 995 liquid environment in the secondary structure of lipases, 997 materials, 994 MCR-ALS procedure, 996 MES experiments, 996 molecular recognition of an acyl enzyme intermediate, 999, 1000 Modulation experiments, 973 Modulation frequency, 969 Modulation, of liquid feed, 855 Modulation period, 972 MOF-808, 749 Mo(Hcitrate)2O11, 92 Moiré pattern, 68 Molecular crystals, 825 Molecular orbital theory (MOT), 239 Molecular vibration, 986 Molybdenum carbide-supported metal catalysts, 450–455 Molybdenum disulfide (MoS2), 429 Molybdenyl, 14 moment, 910, 911, 920 Monochromatic irradiation, 245 Monochromator, 192, 302, 414 Monodentate, 36 nitrates, 38, 41 Monolayer, 24 Molybdenum Disulfide, 429 MoS2, 431 Monolayer equivalent (MLE), 62 Monomeric vanadia, 772 Mono-tungstate species, 422 Montmorillonites pillared, 824 MoO3, 12 Mo-oxide single site heterogeneous catalysts, 308 MoOx moieties, 121, 122 Mordenite, 142 Morphology, 418 Morse potential energy curves, 297 Mo-V-Nb-Te Mixed Oxide Catalysts, 423 Mo-V-Te-Nb, 425 MRI-compatible reactor, 857 M1 structure, 424 Multicomponent SSITKA, 946 Multi-length, 671 Multi-modal, 672 imaging, 677 Multinuclear NMR spectroscopy, 788 Multiple attenuated reflection, 5 Multiple pools in parallel for a first-order irreversible reaction, 940 Multiple pools in series for a first-order irreversible reaction, 939 Multi-quantum magic angle spinning (MQMAS), 759, 768 Multi-technique characterization, 739 “MultiTRACK” Operando Raman system, 116

Index Multivariate calibration, 1084 Multivariate curve regression methods, 626 Multivariate Curve Resolution (MCR), 275, 975 Multivariate curve resolution with alternating least square (MCR-ALS) analysis, 626–636, 996, 999 bilinearity model, 632 initial estimation, 631 PCA-SVD determination, 629 physically and chemically constraints, 627 rank deficiency, 635 Multivariate MOFs, 801 Multivariate spectral analysis, 975 Mutual exclusion rule, 5 N NaA zeolite, 819 Na-mordenite, 818 Na-NiY zeolite, 821 Nanoantennas, 159 Nanocatalysts, 383, 390, 392 Nanoclusters, 68 Nanocolloids, 157 Nanocrystals, 681 Nanoparticles (NPs), 156, 193, 289, 290, 292, 674 interior, 660 Nanoporous crystals, 433 Nanoscale, 681 spatial resolution, 94 Nanoscale catalytic reactions anti-Strokes, 177 dimerization, 177 DMAB, 174–176 pATP, 175, 176 pNTP, 174–176 TERS, 174, 175 thermal effects, 177 triple bond, 177 Nanoshapes, 1064 Nanostructural characterisation, 421 Nanostructure, 434 Nanostructured substrates, 194 Nanotomography, 694 Naphthalene, 138 NaY, 822 Na-ZSM-5, 1016 Nd:YAG laser, 81 Near ambient pressure X-ray photoelectron spectroscopy (NAP-XPS), 60, 250 Near atmospheric pressure (NAP) XPS CO2 hydrogenation and methanol synthesis, 340, 341 CO oxidation, 340 methane activation and conversion, 341, 343, 344 technical issues, 338, 339 Near-edge X-ray absorption fine structure (NEXAFS), 566, 602, 603 Nearest-neighbor distance, 436 Near-field optical spectroscopy, 174 Near-field optical techniques, 183 Near-field Raman spectroscopy, 183 Near-infrared (NIR), 158 Neutron diffraction, 544, 546 heterogeneous catalysis, 506–509 principles of, 505, 506 Neutron scattering advantages and disadvantages, 498, 499 definition, 494 elements, 495, 496

Index imaging, 509–511 INS, 499–502 instrumentation, 495–497 modeling, 497, 498 neutron diffraction, 505–509 properties, 494, 495 QENS, 502–505 SANS, 508, 510 types of, 494 n-fold coordinated hydrides, 45 NH3 adsorption isotherms, 1043 NH3-TPD method, 1015 Ni/CeO2(111), 343 Ni/α-MoC catalyst, 453 Ni/SiO2 methanation catalysts, 617 Ni(0) nanoparticles, 881 Ni (111), 343 NiCu alloyed particles, 643 Ni-Fe based catalysts, 199 Nile Blue, 178 NiO, 587 Niobyl, 14 0.2% NiO/NaTaO3, 485 Ni(OH)2/NiOOH redox couple, 199 Nitrosamide intermediate, 985 Nitrosonium NO+ ion, 37 NO+ ions, 38 NO/O2 reactivity, Cu-CHA chelating nitrate, 38 Cu/Al ratio, 37 CuII-(N,O), 37 speciation, 38 temperature, 37 variable temperature, operando, 38, 39 Noble metal nanoparticles, 636 Noble metal promoters, 640 Nocera catalyst, 882 NO conversion, 984 Non-classical nucleation theory, 636 Nonframework aluminum, 822 Nonlinear optical crystals, 81 Non-noble metal, 427 Non-oxidative dehydrogenation of propane (DH), 269 Non-steroidal anti-inflammatory substances (NSAIDs), 992 Normal coordinate analysis (NCA), 267 Normalization, 206 Notch filter, 82 NOx emission, 266 n-type bases, 16 Nuclear magnetic resonance (NMR), 544, 548, 757, 787 catalyst support surfaces, 789, 790, 792 crystallography method, 788 spectroscopy, 198 studies of zeolites, 795 Number of photons, 304 Numerical fitting, 923 Nutrients, 1085

O Octahedral, 583 17 O enriched MOFs, 800 O-H bond, 227 ν(OH) stretching, 37 Olefin(s), 141 polymerization method, 23

1105 Oligomeric species, 772 Oligomeric vanadia, 772 Oligomerization, 773 Operando, 267, 869, 872, 879, 880, 967, 975 approach, 645 characterization approach, 620 DRIFTS, 42, 742 investigation, 740 measurements, 741 methodology, 740 mode, 853 operation, 590 Raman measurement, 86 Raman techniques, 104 spectroscopy, 27, 593 time-resolved Raman spectroscopy, 92 tomographic methods, 574 XAS, 582, 590, 593 XES, 567 X-ray-based techniques, 190 X-ray microscopy and ptychography, 620 Operando computed tomography (CT) XANES lateral (2D) and depth-profiling (3D) information, 575 Nafion membrane, 575 oxidation state, 575 Pt speciation, 575 redox, 575 three-dimensional, 575 tomographic methods, 575 two-dimensional XAS imaging approach, 574 Operando electrochemical Raman spectroscopy (OERS), 190 cathodic reactions, 197 cell design, 197 conditions, 196 coupling/electrochemistry, 196 OERS, 196 spectroelectrochemical flow-cell, 196 Operando FT-IR spectroscopy DRIFT mode, 45 dynamic, 44 H2 concentration, 44 hydrogen concentration, 45 hydrogen coverage, 45 kinetics, 45 Pt-H species, 44, 45 reconstruction, 45, 46 Operando Raman spectroscopy, 89 activation flow, 117 carbon formation, 117 FTIR, 116, 117 FTIR and Raman spectra, 118, 119 heterogeneous catalysts, 114 in-situ Raman band intensity, 117 in-situ Raman spectra, 116 P-ODH reaction, 117, 118, 120 supported metal oxide catalysts, 114, 115 V/ZrO2 samples, 117 Operando X-ray absorption spectroscopy (XAS), 567 bending magnet, 591 brilliance, 590, 593 bypassing, 592 carbon-carbon coupling catalysis, 570 electron-atom scattering, 566 fluorescence, 566 fluorescence-yield XAS, 566–568 flux, 590, 593

1106 Operando X-ray absorption spectroscopy (XAS) (cont.) gas-solid catalysis, 567 gradients, 570 hard X-ray resources, 585 history, 565 hydrogen peroxide, 571 inelastic mean free path, 566 isothermal, 591 kinetic data, 591 kinetic probe, 566 laboratory-based reactor systems, 593 laboratory based spectrometers, 567 laboratory-scale reactors, 569 laboratory XAS instruments, 587, 590 leaching and mass transport, 570 micro-seconds timescale, 575 nanoparticles, 570 non-focus, 571 oxidation state, 566 palladium, 570 palladium hydride phases, 578, 579 photo-catalytic process, 567 photo-electric effect, 566 photo-electro catalysis, 583 photoelectron, 566 resources, 587 sample environments, 591 soft, 585 soft X-ray approach, 566, 584, 586, 587 soft X-ray region, 567 solvent selection guides, 570 space-and time-resolved XAS, 570 space-resolved XAS, 570 spatially-resolved operando XAS, 569 spatio-temporal mapping, 570 “step-scan” approach, 566 synchrotron X-rays, 566 tender, 566, 584 tender X-ray, 585 time-resolved, 566 “traditional” heterogeneous, 567 transmission, 566 two-dimensional XAS imaging, 573, 574 unoccupied electronic states, 567 Optical microscopy, 286 Organic impurities, 9 Organic linkers, 800 Orientation, 676 Ornithine acetyl transferase (OAT), 993, 994 Orthorhombic phase, 425 Orthorombic perovskite, 7 Ortho xylene, 22 Ostwald ripening, 397 O2-TPD spectra, 1016 Outermost surface layer, 462, 463, 465, 467, 468, 472, 473, 475, 480, 481 Outlier probability value, 1088 Overlapped spectral bands, 996 Oxidation of CO, 332, 419 Oxidation of methane, 681 Oxidation processes, 745 Oxidation reactions, 1022 Oxidation state, 566, 574, 577, 584, 677 Oxidative coupling of methane, 672 Oxidative dehydrogenation (ODH), 92, 426 of propane, 267, 268

Index Oxide, 347, 349, 350 supports, 792 Oxide-supported catalysts, 305 Oxygen evolution reaction (OER), 198, 259, 579, 580, 587 Oxygen isotope pulsing, 923 Oxygen reduction reaction (ORR), 427 electrocatalysis, 201 intermediates, 202 Pt facets, 202 SERS measurements, 202 SHINERS, 202, 203 Oxygen storage capacity (OSC), 575 Oxygen to vanadium LMCT, 268

P Pair distribution function (PDF), 672, 696, 711, 713 p-aminothiophenol (pATP), 174, 176 p-and s-polarized spectra, 67 Parahydrogen (p-H2), 860 Parahydrogen-induced polarization (PHIP), 860 Parallax, 675 Paramagnetic, 872, 878, 880 cations, 822 effects, 777 materials, 870 Paramagnetism, 870 Parametric methods, 943 Para xylene, 22 Partial least squares, 1083 Particle migration-and-coalescence (PMC), 397 phenomena, 396–399 Particle sintering, 650 Passivated Implanted Planar Silicon (PIPS), 568 PCA-SVD determination 2D-score plots, 628 1D-score trajectory plots, 628 scree plot, 628 PC scores, 1083 Pd carbide, 748 PdCx, 746 PDF-CT, 673, 674 Pd nanoparticle, 221, 745, 748, 777 Pd-TWC, 749 Peak fit, 1079 Peak integration, 1076–1077 Pecan shell-based biochar, 825 Peltier reaction calorimeter, 1051 Pentagonal pyramidal units, 423 Pentahedrally-coordinated alumina sites, 767 Pentahedrally-coordinated aluminum species (AlP), 767 Periodicities, 414 Periodic perturbation, 968 Perovskites, 1062 Peroxymonosulfate (PMS), 877 Perturbation, 987 Petrochemical, 124 Phase angle, 969, 970, 972, 986 Phase angular dependence, 972 Phase delay, 970 Phase domain, 970, 982 Phase-resolved data, 971–974 Phase-resolved spectra, 972, 982, 983 Phase-resolved spectrum, 981 Phase sensitive detection (PSD), 969–974, 981–983, 986, 994

Index Phase shift, 986 Phase transition, 831, 1037 4-(phenylazo)phenoxy]hexane-1-thiol, 183 Phenylisocyanide, 182 Phosphines, 20 Phosphorescence, 296, 298 Photoacoustic techniques, 5 Photocatalysis, 225, 259, 296, 485–487, 575, 577, 618, 638 Photocatalytic decomposition of NO, 307 Photo-catalytic process, 567 Photocatalytic reactions, 300 Photocatalytic reactivity, 302 Photocatalytic reduction of CO2 with H2O, 307 Photochemical reaction, 135, 296 Photochemistry, 296 Photodecomposition of H2O, 314 Photo-electric effect, 566 Photoelectrochemical water splitting, 400, 403 Photoelectrochemistry, 403 Photoelectron, 566 Photo-excitation, 577 Photo-generated charge carriers, 305 Photoinduced-desorption, 311 Photoluminescence (PL) absorption spectrum, 297 benzophenone, 310 Bronsted basicity, 311 carbon nitride samples, 309, 310, 312 effect of magnetic fields, 317 effect of temperature, 314–316 energy migration, 305 energy transfer, 305 excitation energy, 297–302 Franck-Condon principle, 297 graphitic carbon nitride (g-C3N4), 326, 328 instrumentation, 302 intensities and lifetimes, 323 Mo-oxide single-site containing samples, 308, 310 photocatalytic process, 311–314 quantum efficiency, 304 quenching, 332 radiative lifetime, 304, 305 reactant interactions, 324–326 sample preparation, 302, 303 spectral parameters, 303 Stern-Volmer expression, 304, 305 surface chemical phenomena, 305, 306 time-resolved PL spectroscopy, 324, 326 Ti-oxide single-sites, 306–308 TMI, 327–334 ultrafast time-resolved PL spectra, 305 vibration structure, 297 V-oxide species, 307, 309 wavelength and spectral shape, 303, 304 Photoluminescent material, 304 Photons, 151 Photo-PROX reaction, 332–334 Photothermal, 5 Physical modelling, 439 Pilatus, 568 Ping Pong Bi Bi mechanism, 993 Plasmonic catalysis, 174 Plasmonic nanoparticles, 171 Plasmonic peaks, 390 Plasmonic platform, 153 Plasmonic spectral shaping effect (PSSE), 207

1107 Plasmonic substrates, SERS bio-detection, 157 dielectric constant, 158 highly ordered substrates, 158 localized surface plasmon response (LSPR), 158 metallic nanoparticles, 157 reducing agent, 157 Plasmon-induced catalysis hot carrier generation, 172 hot electron, 172 hot electron-hole formation, 173 LSP, 171 Plasmon induced nanoscale reactions bond cleavage, 177–178 (de)protonation, 177 dimerization, 174–177 polymerization, 177 triple bond formation, 177 Platinum hydrides, 45 Platinum oxo-species, 433 Plug-flow, 592 p-mercaptoaniline (pMA), 176 p-nitrothiophenol (pNTP), 174–176 Poisoning effect, 679 Polarization-dependent RAIRS (PD-RAIRS), 55 Polarization modulation, 67 Polarization
modulation infrared reflection absorption spectroscopy (PM-IRRAS), 95 Polarization modulation RAIRS (PM-RAIRS), 54 Polarization rotator, 81 poly(2-hydroxyethyl methacrylate) (pHEMA), 226 Polyacenic semiconductor (PAS), 825 Polycyclic aromatic hydrocarbons (PAH), 138, 142 Polyethercarbonate, 1077 Polyethylene, 679, 823 fibers, 824 Polyhedral oligomeric silsequioxane (POSS), 144 Polymer electrolyte fuel cells (PEFCs), 610, 618 Polymer electrolyte membrane (PEM) fuel cell catalyst, 573, 578 Polymeric carbon nitride (CN), 309 Polymeric semiconducting photocatalysts, 315 Polymerization, 679 Polymers, 226, 823 Polyoxometalates, 777 Polystyrene, 142 Poly-tungstate species, 422 Polyurethane with poly(dimethylsiloxane) (PDMS), 226 Polyvinylchloride (PVC), 823 Populations, 419, 422 Pore structure, 679 Porous carbon adsorbents, 832 Porous glass, 828 Porous molecular crystals, 825 Porphyrins, 270 potassium superoxide (KO2), 878 Powders, 1008 Power-compensated DSC, 1035 Power-compensation reaction calorimeter, 1051 Power X-ray diffraction (PXRD), see X-ray diffraction (XRD) p,p’-dimercaptoazobenzene (DMAB), 174 p,p’-dimercaptobenzene, 174–176 Practical membrane electrode assembly, 677 Prediction uncertainties, 1082 Pre-edge feature, 577 Preferential CO oxidation (PROX), 879

1108 Pre-nucleation clusters, 637 Pressure drops, 592 Pressure gap, 441, 929 Principal component(s) (PC), 1083 Principal component analysis (PCA), 146, 453, 628, 996, 1083 Probe gas, 1042 Probe microscopy (SPM), 84 Probe molecules acid probes, 23 adsorption, 16 basicity, 24 basic probes, 16 carbon monoxide, 24 extraframework material, 22, 23 hydrogen bonding, 22 intensity of the IR band, 20 Lewis acid sites, 19 Lewis acid strength, 17 location of acid sites, 21 molecular probes, 21 pore structure, 19 protonic zeolites, 20 spinel structure, 20 surface reconstruction, 20 UV-Vis spectroscopy, 246–248 Probe size, 414 Profens (2-arylpropionic acids), 992 Progression, 303 Projected potential, 416 Projection to latent structures (PLS), 1083, 1085, 1087, 1088 calibration, 1085 regression, 1085 Promoter atoms, 430 Propane, 423 dehydrogenation, 660, 661 oxidation, 426 Propylene epoxidation, 144 Propylene oxide (PO), 1076 Protein, 998, 999, 1001 stability, 997 Pt/Al2O3 catalyst, 34, 48 Pt/α-MoC catalyst, 450–452 Pt alloy catalysts, 660 Pt-based metallic systems, 741 PtCo3, 436 Pt complexes, 433 ν(Pt-H) absorption bands, 42 Pt-hydride, 43, 47, 741, 744 Pt-hydride species, dynamic behavior DRIFT/XAS/MS, 47 H2 concentration, 43 hydrogenation reaction, 46 hydrogen coverage, 43 hydrogen splitting, 41 isotopic shift, 44 molecular hydrogen, 41 morphological reconstruction, 48 Pt/Al2O3 catalyst, 42, 43 Pt-hydrides, 42 toluene hydrogenation, 47 Pt nanoparticles, 326, 741, 742, 744, 745 Pt-NaY, 826 [Pt(NH3)4]2+ complexes, 433 Pt-rich core and a Pt3V shell, 666 Ptychographic X-ray computed tomography (PXCT), 723–725 Pulsed field gradient (PFG) technique, 858

Index Pulse sequence, 987 Pump/probe, 903, 904, 922, 923, 927–930 Pump-probe-spectra, 221 Pump-probe techniques, 618 Pure catalyst powders baseline slope, 10 bridging OH’s, 12 combination modes of lattice vibrations, 11 cut off, 11 hydroxyl groups, 12 impurity, 12 large particle size powders, 11 light scattering, 10 metal-oxygen “double” bonds, 14 methoxy groups, 14 overtones of bulk vibrational modes, 12 probe molecules, 14 residual carbonates, 12 strong Brønsted acidity, 14 sulphate impurities, 12 surface sulphate, 14 triply bridging OH’s, 12 4′-(pyridin-4-yl)biphenyl-4-yl)-methanethiol (PBT), 182 Pyridine, 14, 16–20, 22

Q Q band, 270 Q band is the a1u(π)-eg(π*) transition, 270 3QMAS, 773 5QMAS spectra, 773 Quadrupolar, 759, 760, 763, 765 broadening, 777 coupling constant, 760 interaction, 788 nuclei with half integer spin, 788 splittings, 829 Quadrupolar Carr-Purcell-Meiboom-Gill (QCPMG) sequence, 779 Quadrupole mass spectrometer (QMS), 1054 Quantitative analysis, 157, 432 Quantitative compositional analysis, 422 Quantitative image simulation, 418 Quantitative kinetics, 592 Quantitative STEM, 439 Quantitative STEM-HAADF imaging, 439 Quantum efficiency, 303 α-Quartz, 8 Quartz crystal microbalance (QCM), 196, 1053 Quarupolar lineshape, 763 Quasi-elastic neutron scattering (QENS) heterogeneous catalysis, 503–505 principles of, 502, 503 Quasi-equilibrium, 925 state, 985 Quasi steady-state, 925, 968, 969 Quench, 298 Quenching, 329 of fluorescence, 309 of the PL, 301 process, 305 species, 304 Quick-scanning, 577 monochromators, 566 Quinone methide radicals, 874 QXAS equipment, 608

Index R rac-ketoprofen, 996, 1000 Radial distribution function, 973 Radiationless decay, 298 Radiative decay, 298 Radiative lifetime, 304 Radical, 873–877, 880 Raman-active, 79 Raman cell, 85 Raman effect, 190 Raman imaging, 92 Raman-mass spectroscopy, 124 Raman microscopy, 94 Raman scattering (RS), 151, 190 cross-section, 152 definition, 152 electric dipole moment, 77 electric field, 77 fluorescence, 77 inelastic light-scattering process, 76 internuclear separation, 77 oscillating dipole, 152 polarizability, 77 process, 153 Raman active, 77 spectroscopy, 5, 6 virtual state, 152 wavenumber shift, 76 Raman scattering theory adsorption, 135 anti-stokes, 132 A-term, 133 B-term, 133 characteristics, 133 coke formation, 134 crystalline MoO3, 134 C-term, 133 dispersed Mo species, 134 FTIR, 132 isooctane, 135 Jablonski diagram, 132 oscillator strength, 133 overtone progressions, 133 polarizability derivative, 133 resonance enhancement, 133 scatterers, 132 Stokes Raman, 132 sum-over-states formalism, 132 vibration coupling, 133 Raman spectra, 111, 112 Raman spectrometers, 191 Raman spectroscopy, 111, 980, 984, 985, 987 background signals, due to fluorescence, 191 bulk information, 80 catalyst, 77 catalyst preparation, 96 CCDs, 192 coprecipitation process, 96 crystal structure, 80 definition, 76, 190, 207 dehydrogenation of propane, 102 dispersive devices, 192 ED-XANES, 103 elastic and inelastic scattering processes, 190 excitation source, 80 ideal catalytic reactors, 86

1109 in situ cell design, 85 in situ experiment, 84 instrumentation, 80 light source, 191 magnification, 192 multimodal analysis, 102 non-dispersive instruments, 192 operando conditions, 190 particle size, 80 principles, 207 Raman data sets, 104 resolution, 76, 192 rotational cell, 85 sample holder/detection system, 80 scattering, 190 selection rules, 79 spectral range, 76 spectrometer, 191 surface structure, 80 UV lights, 81 weak Raman scatterer, 79 Raman spectrum, 113 Rank, 1083 deficiency, 633–636 Rate constants, 329 Rate-determining steps, 1007, 1016, 1023, 1026 Rayleigh scattering, 76, 77, 136, 152, 190 Ray spectro-ptychography, 683 Reaction calorimetry, 1049 Reaction cells anthracene, 137 decomposition, 137 H-USY zeolite, 137 Linkam, 138 magnetic coupling, 137 membrane pump, 138 numerical aperture, 137 oxidative dehydrogenation, 138 photochemical degradation, 136 power density, 137 sample heating, 136 Reaction heats, 1033 Reaction kinetics, 577, 592 Reaction mechanism, 1023 Reaction pathways, 1007 Reactor, 672 Readsorption, 941 Real-space reconstruction, 436 Real-time control, 1073 Reciprocity, 415 Recombination of photo-formed electrons and holes, 309 Recovery function, 1081 Redhead Equation, 1006 Redistribution of atoms in bimetallic particles, 650 Redox center, 980 Redox mechanism, 954 Redox reaction mechanism, 1024 Reduction-oxidation, 673 Reflection absorption infrared spectroscopy (RAIRS), 53 absorption factor, 54 adatoms/clusters, 71 adsorbed CO, 66 ambient, 56 atop CO, 67 bimetal, 63 Blueshifts, 62

1110 Reflection absorption infrared spectroscopy (RAIRS) (cont.) bridge site CO, 68 bridging site CO, 63 broad spectral range, 71 chamber pumping system, 55 CO2 formation, 65 CO adsorption, 61, 62, 64, 66, 69 CO interaction, 63 CO oxidation, 64 DFT, 67 environmental chemistry, 61 FTIR, 63 graphene moiré unit cell, 69 herringbone reconstruction, 67 hydrocarbons, case studies, 55 integrated peak area, 66 IR band, 64 LEED, 55 liquid-solid interface, 54 nanoclusters, 70 P-and R-branch, 68 PEM, 55 PM-RAIRS, 55 polarizer, 55 probe molecule, 67 red-shifted, 71 spectra, 69 sticking probability, 62 STM, 67, 70 superstructure, 69 thermocouple, 55 TPD, 66 UHV, 53 vibrational frequencies, 53 XPS, 63, 71 zero oxidation state, 66 Regeneration strategies, 650 Rejection filters, 82 Resonance effect, 41 Resonance enhanced Raman spectroscopy, 146 Resonance Raman scattering (RRS), 132 Resonance Raman (rR) spectroscopy, 248 Resonance scattering condition, 83 Resonant, 215 Resonant inelastic X-ray scattering (RIXS), 660, 664, 665, 695 Resonant inelastic X-ray spectroscopy (RIXS), 587 Resonant Raman enhancement, 178 Reverse water gas shift, 63 Reversible adsorption, 911, 912, 915, 917, 918, 920, 922, 927 Reversible reactions, 940 Rh/TiO2 model, 476 Rhombohedral perovskite, 7 Rietveld method, 525, 548, 553, 555 Ronchigram, 413 Ruthenium, 822 Rutile and anatase TiO2, 311

S Sample generation/tip collection (SG/TC) mode, 199 Sample holder, 82 Sample illumination/light collection system, 80 Sample illumination, 81 Sample topography, 360 Sampling depth, 373 s-and p-polarized components, 54

Index s-and p-polarized infrared radiation, 54 s-and p-polarized spectra, 55 SAPO-34, 142 SAPO-37, 816 SAPO-5, 816 SBA-15, 828 Scanning 3D XRD, 676 Scanning confocal electron microscope (SCEM), 437 Scanning electrochemical microscopy (SECM), 197 Scanning electron microscopy (SEM), 204 of liquids, 375 in a STEM, 367 Scanning near-field optical microscopy (SNOM), 94 Scanning probe imaging, 699–701, 703, 707, 709 Scanning probe microscopies, 194 Scanning probe microscopy, 174 Scanning transmission electron microscopy, 382, 410, 450 single atom catalysts (SACs), 449 Scanning transmission electron microscopy (STEM) high-angle annular dark field (HAADF) imaging Au/α-MoC catalyst, 452–454 bulk mixed oxide catalysts, 423–426 carbon-based electrocatalysts, 427–429 depth-sectioning, 437–439 2D MoS2-based catalysts, 429–433 electron-matter interactions, 415–419 electron probe, 411–415 formation, 410, 411 heating, 427 ionization damage, 426 knock-on damage, 426 Ni/α-MoC and (Co-Ni)/α-MoC, 452–455 Pt/α-MoC catalyst, 450–452 quantitative imaging, 439, 440 supported metal catalysts, 418–421 tomography, 434–437, 439 working environment, 441, 442 WOx/ZrO2 solid acid catalyst, 421–423 zeolite type materials, 433–435 Scanning tunneling microscopy (STM), 70, 194 Scattering, 265, 672 Scherrer’s equation, 525, 527 Schottky field emission gun (FEG), 411 SCR mechanism, 979 SE and BSE detectors, 360 Secondary electrons, 360 Secondary structure, 996 Secondary well defined porosity, 819 Second-derivative photoluminescence spectra, 308 Second-order quadrupolar perturbation, 762 Segregation of cations, 1068 Selection rules, 300 Selective catalytic reduction (SCR), 35, 95, 120, 266, 489, 614, 615, 772, 1025–1027 adsorption/desorption of NO, 981 Cu-SSZ-13, 983 discrimination between active and responsive sites, 983–985 DRIFT of Fe-ZSM-5, 980 Fourier transformation, 981 intermediates species, 985–986 modulation, 987 overall reaction cycle, 979 phase domain, 982 phase-resolved spectra, 982, 983 phase-resolved spectrum, 981 phase sensitive detection (PSD), 981

Index Raman spectroscopy, 980 signal enhancement, 982 time-resolved spectra, 981 Selective oxidation of C3H6 to C3H4O, 1024 Selective oxidation of CO in rich H2, 308 Selective oxidation of surface atoms, 666 Selectivity, 970 1s electron, 602 Self-absorption, 675 artefacts, 707 Self-assembled monolayer (SAM), 159, 175, 183, 227 Self-supporting pressed disks, 10 Semiconducting catalyst, 300, 301 Semiconducting materials, 300 Semiconductor photocatalysts, 310 S-enantiomers, 992 Sensitive scanning, 84 Sensitivity, 970, 1033 SERS-active catalysts, 206 Setups, 740, 751 Shale, 829 Shape, 682 function, 681 β-sheet structure, 998 Shekhtman Reactivities, 910–912 Shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS), 195 SHINERS method, 202 Si/Al ratios, 815 Si3N4, 586, 587 Signal enhancement, 982 Signal-to-noise ratio (SNR), 68, 430, 758, 851, 969 Silica/Zeolite synthesis Fe-ZSM-35, 140 microporosity, 140 NMR, 140 oxidation/isomerization, 140 Silica, 8, 22 Silica-alumina system, 8 Silica-supported transition metal ions, 327 Silicon drift detector (SDD), 389 Silicon nitride, 441 Silver carbonate (Ag2CO3), 1010 SIMPLISMA (SIMPLe to use Interactive Self-modeling Mixture Analysis) method, 146, 628, 631 Simultaneous algebraic reconstruction technique (SART), 436 Simultaneous iterative reconstruction technique (SIRT), 436 Sine-wave stimulus, 970 Single, 679 Single-atom alloy (SAA), 57, 67 Single atom catalysts (SACs), 449, 646 Single crystal(s), 1008 diffraction, 543, 546 electrode, 221 surfaces, 1053 Single crystal adsorption calorimetry (SCAC), 1053, 1056 Single-molecule fluorescence microscopy 2D Gaussian function, 286 reactivity and heterogeneity, 286, 287 restructuring and switching, 287, 288 scalable parallel screening, 290–292 spatial and temporal catalysis cooperativity, 292 super-resolution catalysis imaging, 288–290 Single-pellet reactor models, 854 Single pool first-order irreversible reaction, 937 Singlet-triplet transitions, 300

1111 Singular value decomposition (SVD) method, 628 Sintering, 70, 668 Site coverage, 951, 954, 957 Size-dependent, 745 Skeletal IR spectra, 5, 6 Small-angle neutron scattering (SANS), 508, 510 Small angle X-ray scattering (SAXS), 696 Small loss of surface Pt-Pt and Pt-Cr bonds, 661 SnO2, 586 SO2 adsorption calorimetry, 1043 Sodium aluminate, 12 Soft coke, 143 Soft-landed, 439 Soft versus hard X-rays, 606–608 Soft X-ray region, 567 Soft X-ray XAS, 593 Sol-gel techniques, 9 Solid acid, 421 Solid catalyst amorphous solids, 8 binders, 9 diordered solids, 8 far IR studies, 5 fundamental vibrations, 5 impurities in catalysts, 9 irreducible representation, 8 spent, 8 unsupported metal nanoparticles, 9 Solid-electrolyte interface (SEI) composition, 774 Solid oxide fuel cell, 672 Solid-state laser, 81 Solid state NMR, 788 Solvation structures, 776 Soret band, 270 Source size, 411 Sources of variance, 1083 Spatially-resolved catalyst characterization, 691 Spatially resolved NMR spectroscopy (MRS), 850, 855 Spatially resolved Raman microscopy AFM, 95 AFM-TERS, 94 definition, 92 spatial distribution, 94 Spatially resolved XAS methods, 608–610, 612 Spatial resolution, 139, 411, 573 definition, 701 in heterogeneous catalysis, 702–703 Spatio-temporal resolution, 683 Specific heat measurements, 1033 Spectator(s), 945 species, 47, 968, 969 Spectral analysis, 1073, 1089 Spectral deconvolution, 1000 Spectral hard modeling methods, 1075, 1078, 1079 Spectral residuals, 1079 Spectroelectrochemical cell, 882 Spectrokinetic analysis, 992 Spectrometer, 82, 216 Spectrophotofluorometers, 302 Spectroscopic methods, 216 Spectroscopy, see High-field NMR spectroscopy Spectroscopy cell, 216 Spherical Aberration, 412, 413 Spillover, 985 Spin-exchange optical pumping (SEOP), 814 Spin Hamiltonian, 870

1112 Spin-lattice relaxation, 777, 871 Spinning disk, 136 Spinning sideband(s), 765 patterns, 773 Spin number, 762 Spin-orbit coupling, 300 Spin trapping, 875 Splitting/decomposition of water, 309 Spontaneous emission, 298 Square-wave shape, 970 Square-wave stimulus, 970, 974, 995 SSITKA-neutron scattering, 946 States of coordination, 984 Statistical methods, 1083 Steady-state isotopic transient kinetic analysis (SSITKA), 917, 975, 1065 advantages, 936 chromatographic effects, 942 CO hydrogenation on Al2O3 supported Co catalysts, 948, 950 combing DFT and transient and steady-state modeling to study reaction mechanism of CO, 952, 954 combining SSITKA-DRIFT for WGS, 954, 955, 958 CSTR, 943 discrimination between different reaction mechanisms, 947, 948 flowsheet of a setup, 936 ILT method, 943 kinetic parameters for higher hydrocarbons in CO hydrogenation, 951 reaction systems, 937 single pool first-order irreversible reaction, 937, 939 SSITKA-FTIR, 944, 946 Steady-state spectroscopic studies, 980 STEM-EELS, 427 elemental mapping, 452 STEM-HAADF Tomography, 435 STEM in an SEM, 367–368 Step-scan methodology, 577 Stereospecificity pockets, 997 Stern
Volmer plots, 332 Stern-Volmer equation, 330 Stern-Volmer expression, 305 Stimulation, 968 Stimulus, 968–970, 972, 973, 975 Stokes/anti-Stokes scattering, 131 Stokes Raman scattering, 190 Stokes scattering, 77, 152 Strain, 676, 681 Strategy of borrowing SERS activity, 195 Stretching modes, 12 Strong adsorption sites, 818–823 Strong ligand-to-metal charge transfer (LMCT), 102 Strong metal-support interaction (SMSI), 452, 616 Structural defects, 818 Structural information, 418 Structure–activity relationship, 788, 991 Structure-performance relationship, 273, 644 Structure-property relationships, 410 Structure-reactivity relationships, 591 Styrene, 142, 529, 530 Sub-Å, 415 Sub-critical size, 637 Subparticle, 288–290 Subtracting the oxidized from reduced EXAFS spectrum, 661 Subtraction spectrum, 14

Index Sulfated-zirconia (SZ), 91 Sulphate species, 9 Sum frequency generation (SFG) spectroscopy adsorption properties, 225 asymmetry, 216 atmospheric pressure, 222 broadband, 217 case studies, 218 CO adsorbed, 220 combined studies, 222 components, 216 D2O, 224 definition, 214 DFT, 224 electric field, 215 energy transfer, 221 frequencies, 214 gold surface, 227 heterogeneous catalysis, 227 hydrocarbon chains, 226 hyperpolarizability, 215 ice, 227 incidence angles, 217 interfaces, 213 interfacial structure, 226 laser, 215 laser pulses, 217 linear CO, 222 lineshape, 216 liquid water, 227 microscopy, 218 model catalysts, 222 molecular ordering, 227 molecular orientation, 219 morphology, 222 nonlinear optical process, 214 oxide surfaces, 224, 225 phase, 216 polarization dependent, 226 ppp, 218 Pt(111) and Pd(111), 219 SAMs, 227 scanning, 217 second order, 215 selection rules, 214 self-assembled monolayers, 226 setup, 218 signaling intensity, 215, 216 solid surfaces, 219 super-cooled water, 227 supported metal nanoparticles, 221 surface roughening, 222 surfaces, 213 surface-specificity, 219 susceptibility, 214 tilt angle, 226 time-resolved, 217 titled molecules, 220 UHV, 218 UHV grown, 221 water, 227 Supercritical conditions, 829 Superparamagnetism, 870 Super Photon Ring-8 (SPring-8), 605 Support, 984

Index Supported alumina catalysts, 793 Supported catalysts, 767 active phase dispersion, 642 metal distribution, formation of phases, 643 metal distribution, within bimetallic particles, 643–644 Supported metal oxide catalysts, 114, 489, 490, 1017, 1022 Supported metals, 826 Supported oxide catalysts, 423, 793 Supported vanadium oxide catalysts, 115, 120 Surface adsorbates, 349–352 Surface and bulk structures, 666 Surface and particle interior, 662 Surface atoms, 663 Surface catalytic active sites, 485 Surface chemistry, 341, 352, 1016 Surface compositions, 661 Surface coverage, 915, 916, 918, 925, 929 Surface dipole selection rule, 54 Surface-enhanced Raman scattering (SERS), 84 active substrates, 156, 157 applications, 164–166 cross-sections, 151 CT, 155 definition, 152 electromagnetic effect, 153–155 field enhancement, 155 imaging, 163 kinetics, 164 surface selection rules, 156 vibrational states, 151 Surface-enhanced Raman spectroscopy (SERS) chemical enhancement, 194 EF, 194 EH, 193 NP, 193 signal intensity, 192 substrates/fabrication, 194 surface plasmons, 193 surface-specific experiments, 195 TERS, 194, 195 Surface-enhanced resonance Raman scattering (SERRS), 155 Surface EXAFS, 660 Surface facets, 372, 375 Surface hydration/hydroxylation, 324 Surface intermediates, 900, 902, 911, 915, 925, 940, 944, 947, 951, 955, 1007, 1009 Surface of Al2O3, 310 Surface plasmon resonances, 152 Surface Pt3Cr monolayer, 662 Surface reactions, 337 Surface reconstruction, 668, 1063–1065 Surface relaxation, 1064 Surface science, 3 Surface-selective labelling, 790 Surface sensitive, 660, 661, 664 Surface sites, 1009, 1016 Surface species, 28 Surface states, 324 Surface structure, 660, 664 Surface termination, 1063 Surface transformation, 925 Surface traps, 324 Surfactants, 157 Swiss light source (SLS), 605 Symmetrical strong hydrogen bonding, 22

1113 Synchronous, 972 Synchrotron, 587, 590, 672, 681 Synchrotron-based NAP-XPS, 337 Synchrotron-based X-ray techniques, 102 Synchrotron radiation characteristics of, 695–696 sources, 605 Synchrotron X-rays, 566 resources, 593 Synthesis gas, 678

T Tauc plots, 268 Temperature-dependence of the Debye-Waller factor σ, 631 Temperature-dependent PL spectra, 316 Temperature-dependent process, 314 Temperature-dependent XAS data set, 633 Temperature effects, 303, 314–316 Temperature gradients, 592 Temperature maps, 852, 862, 864 Temperature programmed decomposition (TPD), 1005–1009 ammonium metavanadate (NH4VO3), 1011 silver carbonate (Ag2CO3), 1010 Temperature programmed desorption (TPD), 55, 65, 246, 1005, 1006, 1045 NH3-TPD method, 1015 O2-TPD, 1016 Temperature programmed oxidation (TPO), 1006, 1012 reduction (TPO/TPR), 1005 Temperature-programmed Raman spectroscopy, 97 Temperature programmed reduction (TPR), 97, 1006, 1009 CO-TPR, 1014 H2-TPR, 1013, 1014 Temperature programmed surface reaction (TPSR), 1005, 1006 Temperature programmed techniques catalysts, 1007 history, 1009 limitations, 1007 powders, 1008 single crystals, 1008 theory, 1006 thermogravimetric analysis, 1009 Temperature range, 1032 Temporal analysis of products (TAP) active zone configurations, 906 adsorption mechanism, 921, 922 collisions in a diffusion reactor, 904, 905 competitive and inhibited adsorption, 922 diffusion/reaction systems, 908, 909 experimental concepts, distinctions, 902, 903 experimental studies of, 915–917, 919–921 gas and solid, in kinetic measurement, 907, 908 general reaction model analysis, 923–925, 927 instrument configuration, 900, 901 Mars-van Krevelen mechanism, 923 model-free analysis of pulse response data, 910 moment-based quantities, 910, 911 pressure gap, 928–930 reaction mechanism, 927 Shekhtman reactivities, 911, 912 standard diffusion curve, 905, 906 time-dependent analysis of rate and concentration, 912, 913, 915 Terahedrally coordinated V-oxide single oxide species, 314

1114 Terminal group orientation, 227 TERS-cyclic voltammograms, 179 Test set, 1081 Tetraethylorthosilicate (TEOS), 140 Tetragonal perovskite, 7 Tetrahedral aluminum, 763 Tetrahedral V5+-oxide species, 307 Tetramethylorthosilicate (TMOS), 140 Thermal effects Arrhenius equation, 173 hot carriers, 173 near-field heating, 173 near-field temperature, 173 power densities, 173, 174 Thermal emission, 134 Thermal stability, 1037 Thermogravimetric analysis (TGA), 1006–1009, 1011, 1012, 1017 Thermogravimetric analysis-differential thermogravimetric analysis (TGA-DTG), 1005, 1009, 1011 Thermogravimetry (TGA), 1036 Thermokinetics, 1042 Thermomyces lanuginosus (TTL), 993, 994, 996, 998 Thickness, 416 Thin zone, 906–908, 911, 912, 916, 926, 929 Third-generation synchrotrons, 590, 591, 594, 673 Thoria, 11 Three-dimensional (3D), 434 Three way catalysts (TWCs), 745 Through-focal series, 437 Ti:Sapphire, 135 Tikhonov-Fredholm (T-F) method, 943 Tilt-series, 436 Ti-MCM-41, 306 Ti-MCM-48, 306 Time-of-flight (TOF) information, 861 Time-resolved ATR-FTIR, 999 Time-resolved infrared spectra, 993 Time-resolved photoluminescence spectroscopy, 324, 326 Time-resolved Raman coke, 92 definition, 91 ethanol reaction, 92 oligomerized surface structure, 92 resonance enhancement, 93 VOx species, 92 Time-resolved spectra, 971, 981 Time-resolved spectroscopic techniques, 313 Time-resolved X-ray absorption spectroscopy (TR-XAS), 625 in catalysis, 613–614 catalyst in operation, 644–654 catalyst preparation, 636–646 electrocatalytic processes, 618 electromagnetic radiation interaction, 602–604 EXAFS equation, 604–605 heterogeneous catalysis, 614–616 photocatalysis, 618 spectral versus time resolution, 608, 609 X-ray source, 605–607 TiO2, 225 TiO2-based catalysts, 777 TiO2 nanoparticles (NPs), 324 reactant interactions, 324–326 time-resolved PL spectroscopy, 324, 326 TiO2 photocatalysts, 311 TiOx/Pt(111) surface, 476

Index Ti-oxide single-site species, 306 Tip-enhanced Raman scattering (TERS), 78, 84, 94, 190, 194 AFM, 170 atomic force microscopy (AFM), 173 catalyst/probe, 183 catalytic reactions, 174–177, 181–183 cis-trans isomerization, 183 complexation reactions, 175–176 configurations, 170 definition, 161 deprotonation, pyridine, 177 electrochemical reactions, 178–180, 183 in situ, 170 localized protonation, 177 malachite green, 177 Raman signals, 170 resonance Raman, 177 scanning tunneling microscopy (STM), 170 setups, 170 spectra, 183 spectral database, 184 spectral fluctuations, 175 TERS probes, 170–171 tracking catalytic reactions, 171 transmission mode, 170 UHV-TERS, 162 van der Waals interactions, 163 Titania, 11, 12 Titania-supported vanadium oxides, 771 Titanium silicalite, 143 Titration calorimetry, 1045 TO/LO splitting, 5, 7 Toluene, 143 Tomographic reconstruction algorithms, 698 Tomography, 404, 434, 573, 672 Total electron-yield, 586 Total X-ray fluorescence-yield, 587 Training set, 1080 Transfer, 305 of polarization, 824 Transformation pathways, 970 Transient absorption, 312 Transient-based methodologies, 980 Transient catalyst behavior, 1067 Transient experiments, 992 Transient mode, 268 Transient operando XAS experiments, 614 Transition metal ions (TMI), 238, 327–334 Transition metal oxide, 296 nanoparticles, 613 Transmission/absorption technique, 5 Transmission electron microscopy (TEM), 410, 693 EDS, 389–392 EELS, 389–392 environmental holder, 386–388, 396–403 ETEM, 384–386, 392–397 in situ experiment, 401, 402, 404 principles, 382–384 Transmission spectroscopy, 241–243 Transmission XAS, 567 Transmission X-ray microscopy (TXM), 610, 700 Transmittance, 265 Transverse optical mode (TO), 7 Trapped holes, 314 Traveling chemical waves, 852 Trickle bed reactors, 854

Index Tri-coordinated aluminum species, 798 Triple monochromator, 82 Triple-quantum MAS NMR, 768 True shift in the energy, 666 TR-XAS with Raman spectroscopy, 651 Tunable laser systems, 136 Tungstated-zirconia, 91 Turnover frequency (TOF), 938, 939, 950–952, 954, 956, 1062 Turnover rates, 665 β-Turn structure, 998 Two-dimensional high-field NMR, 768 Two-dimensional XAS imaging, 573 σ-type bases, 16 π-type bases, 16

U Ultrafast time-resolved PL spectra, 305 Ultrafast time-resolved PL spectroscopy, 323 Ultrafast transient absorption spectra (TA), 310 Ultrafast X-ray spectroscopy of catalysts, 618 Ultrahigh-field, 764 Ultra-high magnetic field, 795 Ultrahigh vacuum (UHV), 53, 218 Ultrahigh vacuum-TERS (UHV-TERS), 162 Ultraviolet (UV) Raman, 78 Ultraviolet-Visible spectroscopy (UV-Vis) Ultraviolet-visible (UV-Vis) spectroscopy, 240–241, 249, 257, 259, 261 basic principles, 238–240 chemometrics and multivariate analyses, 251–254 complementing data interpretation with density functional theory, 251 coupling UV-Vis spectroscopy, analytical methods, 248–251 diffuse reflectance spectroscopy, 243–245 electrocatalysis, 259 electroreduction of CO2, 270–272 fiber optics spectroscopy, 245 heterogeneous catalysis, 254–255 homogeneous catalysis, 255–259 micro-spectroscopy, 245 MTO, 272–279 NH3-SCR, vanadium/copper catalysts, 266, 267 photocatalysis, 259 probe molecule UV-Vis spectroscopy, 246–248 propane dehydrogenation, 267–270 sample preparation, mode of measuring and catalytic reactors, 241–245 transmission spectroscopy, 241–243 Uncertainty, 1081 “Unconventional” Raman spectroscopy, 112 Under-coordinated promoters, 430 Unit assembly, 425 Univariate calibration, 1080–1082 Unoccupied electronic states, 567 UV Raman spectroscopy, 83 UV resonance Raman spectroscopy (UVRRS), 83 application, 132, 140 BBO, 135 bipodal, 145 catalyst deactivation, 140, 141 catalysts, 138 catalyst treatments, 140 ceria crystals, shape, 146 collection times, 139 definition, 131 dehydration, 140

1115 ethanol, 144 fiber bundle, 136 fluorescence, 132 hydrogen reduction, 140 instrumentation, 135 microscope objective, 135 multimodal operation, 146 nanoshaped, 146 normal coordinate analysis, 145 operando, 136 polymer size, 141 Raman scattering, 132 reaction cells, 136 reflecting microscope objective, 135 spatial resolution, 139 speciation, 140 spectrum, time required to obtain, 138 steady state conditions, 139 supported metal oxide, 139 supported vanadium oxide, 145 titania, 143, 144 tripodal, 145 V2O5 forms, 145 vanadia, 140 UV-vis, 741, 772 UV-visible spectrometer, 1045 UV-visible spectroscopy, 987 UV-Vis-NIR spectrophotometer, 153 UV-Vis-NIR spectroscopy, 238–241, 243 UV-Vis spectroscopy, 146, 604 UV-wavelength lasers, 84

V Valence orbitals, 665 Valence-to-core (V2C) XES, 567 Vanadium, 980 catalysts, 266, 793, 980 Vanadium oxide, 135, 139 catalysts, 308, 769, 772, 1013 species, 116 Vanadyl, 14 groups, 980 Variable importance in projection (VIP), 1088 Vibrational deactivation, 300 Vibrational excited states ν1, ν2, 297 Vibrational fine structure, 303 Vibrational spectroscopy, 42, 76, 213, 593 excitation, 4 fundamental transitions, 4 isotopically labeled molecules, 5 overtone and combination modes, 4 polarizability, 5 structure, 4 TO/LO splitting, 5 vibrational ground state, 4 Vibrational structure, 297, 298 Vinyl chloride, 420 Vinylidene (CCH2), 57 (VO)2P2O7, 12 Voltammetry, 578, 584, 593 Volumetric isotherms, 1040 V2O5, 12 VOx/Al2O3 catalysts, 1013 VOx species, 88 V2O5/TiO2 catalyst, 95, 489

1116 V2O5-WO3/TiO2 catalysts, 489, 490 1% V2O5-5% WO3/TiO2 catalysts, 490 VOx/ZrO2/SiO2 catalysts, 1014 VPI-5, 818 VT 13C NMR, 806 VT 17O SSNMR, 806

W Water electrolysis, 196 Water gas shift (WGS), 420 Water-gas shift reaction (WGS), 339, 529, 531, 955, 958, 1015, 1023 Water photocatalytic splitting, 326 Wavelength-dispersive spectroscopy (WDS), 370–371 Weak phase object, 416 Weak Phase Object Approximation (WPOA), 416 Whiteline, 973 Wide angle X-ray scattering (WAXS), 633 Widefield fluorescence microscopy, 163 WO3, 12 Wolframyl groups, 14 Wolframyl species, 14 5% WO3/TiO2 catalysts, 489 Wulff constructions, 436

X XANES-CT, 676, 677 XAS/DRIFTS/MS, 742, 745, 749 XAS/FTIR/XRD, 749 μ-XAS probes, 573 129 Xe NMR, 804 131 Xe NMR, 829 Xenon chemical shift anisotropy, 817 complex structures, 818 high pressure, 816 hyperpolarized xenon, 814 influence of cations, 821, 822 influence of temperature, 817 mixtures of zeolites and catalysts, 818, 819 nonframework aluminum, 822 strong adsorption sites, 819, 821, 822 zeolite crystallization, 819, 820 zeolites, 815, 816 Xenon-xenon interactions, 814, 832 Xe-surface interactions, 815 X-ray(s), 671 coherence, 593 diffractogram, 525, 527 dose, 590, 591 induced reduction, 574 optics, 694 ptychography, 679, 683 scattering, 575, 593 spectro-ptychography, 575 X-ray absorption fine structure (XAFS), 397 X-ray absorption near edge structure (XANES), 332, 566, 568, 577, 583, 584, 590, 602, 603, 683, 708, 740, 742–745, 748, 750 X-ray absorption spectroscopy (XAS), 95, 332, 570, 587, 591, 593, 602, 660, 676, 690, 691, 741, 744, 749 in catalysis, 613–614 data, 973, 974 methodology, 620

Index spatially resolved, 608–610 theory, 608 XAS combined with DRIFTS and MS, 741–748 XAS combined with transmission FT-IR and X-ray diffraction, 749–751 X-ray-based characterization methods, 613 X-ray CT, 694, 697, 702, 703, 715 principle, 697 X-ray diffraction (XRD), 80, 95, 251, 424, 672, 690, 693, 695, 698, 703, 704, 706, 708, 711, 713, 717, 725, 740, 741, 746, 748–750, 972, 974 bulk characterization technique, 519 crystalline solids, 522–525 definition, 521 developments, 534–537 diffractogram, 525, 527 history, 519, 520 in-situ and operando characterization, 527–533 limitations, 532, 534 source of X-rays, 521, 522 X-ray diffraction computed tomography (XRD-CT), 672–674, 678, 679 X-ray emission spectroscopy (XES), 567, 593, 695 X-ray energy-dispersive spectroscopy (XEDS), 421 X-ray excited optical luminescence (XEOL), 568 X-ray fluorescence (XRF), 676, 693, 694 detector, 706 X-ray fluorescence-CT (XRF-CT), 672 X-ray fluorescence spectroscopy (XRF), 604 X-ray imaging, 690, 693–696, 699, 703, 704, 716, 718, 724, 725, 727 fourth generation and diffraction-limited synchrotron radiation sources, 729–731 full-field imaging, 699–700 imaging in 2D vs. tomography in 3D, 696–699 in situ and operando tomography, 725–731 laboratory X-ray sources and synchrotron light sources, 694–696 principles of tomographic reconstruction, 697–698 principles of X-ray tomography, 696–697 scanning probe imaging, 699–701 scientific resources for tomography data, 698–699 spatial resolution and length scale, 701–703 tomographic reconstruction algorithms, 698 XANES tomography, 719–723 X-ray ptychographic microscopy and tomography, 722–726 X-ray microscopy (XRM), 610, 690, 691, 694, 696, 710, 716, 719, 720, 727, 731 absorption contrast imaging, 704–707 chemical imaging, 704 diffraction contrast imaging, 711–717 energy-resolved XAS imaging, 708–713 fluorescence contrast imaging, 706–708 high spatial resolution, 703 high time resolution, 704 phase contrast imaging, 713–719 X-ray photoelectron spectroscopy (XPS), 198, 337, 584

Y Y-procedure, 903, 910, 912–914, 918, 920, 925–927

Z Z-contrast, 451 Zeeman effect, 758, 870 Zeeman level, 762

Index Zeeman splitting, 758 Zeolite(s), 20, 120, 138, 272, 273, 275, 276, 278, 279, 433, 590, 678, 679, 683, 871, 980, 983, 984 and catalysts, 818, 819 characterisation, 542–544 framework, 795 identification, 541 in situ studies, 549–551 materials, 778 MTH conversion, 542 operando catalytic studies, 551, 552 organic molecules, 544, 546, 547

1117 PDF analysis, 556 post mortem, 547–549 synthesis, 140 time and space resolved studies, 552–557 Ziegler-Natta catalyst, 679 Zinc MOF, 874 Zirconia, 11, 14 ZnO/CuOx/Cu(111), 341 ZrO2 film, 60 ZSM-22, 142 ZSM-5, 816, 819 catalysts, 1015