Surface Enhanced Raman Vibrational Studies at Solid Gas Interfaces (Springer Tracts in Modern Physics) 3540134166, 9783540134169

Citation preview

Springer Tracts in Modern Physics 104

Editor: G. H6hler Associate Editor: E.A. Niekisch Editorial Board: S. FI0gge H. Haken J. Hamilton H. Lehmann W. Paul

Springer Tracts in Modern Physics 79 Elementary Particle Physics 80* Neutron Physics

With contributions by E. Paul, H. Rollnick, P. Stichel

With contributions by L. Koester, A. Steyerl

81 Point Defects in Metals I: Introductions to the Theory (2nd Printing) By G. Leibfried, N. Breuer 82 Electronic Structure of Noble Metals, and Polariton-Mediated Light Scattering With contributions by B. Bendow, B. Lengeler 83 Electroproduction at Low Energy and Hadron Form Factors By E. Amaldi, S. P. Fubini, G. Furlan 84 Collective Ion Acceleration With contributions by C. L. Olson, U. Schumacher 85 Solid Surface Physics

With contributions by J. HOIzl, F. K. Schulte, H. Wagner

86 Electron.Positron Interactions

By B. H. Wiik, G. Wolf

87 Point Defects in Metals I1: Dynamical Properties and Diffusion Controlled Reactions With contributions by P. H. Dederichs, K. Schroeder, R. Zeller 88 Excitation of Plasmons and Interband Transitions by Electrons

By H. Raether

89 Giant Resonance Phenomena in Intermediate-Energy Nuclear Reactions By F. Cannata, H. Uberall 90* Jets of Hadrons

By W. Hofmann

91 Structural Studies of Surfaces With contributions by K. Heinz, K. M011er,T. Engel, and K. H. Rieder 92 Single-Particle Rotations in Molecular Crystals

By W. Press

93 Coherent Inelastic Neutron Scattering in Lattice Dynamics 94

By B. Dorner

Exciton Dynamics in Molecular Crystals and Aggregates With contributions by V. M. Kenkre and P. Reineker

95 Projection Operator Techniques in Nonequilibrium Statistical Mechanics By H. Grabert 96 Hyperfine Structure in 4d- and 5d-Shell Atoms

By S. B0ttgenbach

97 Elements of Flow and Diffusion Processes in Separation Nozzles By W. Ehrfeld 98 Narrow-Gap Semiconductors With contributions by R. Dornhaus, G. Nimtz, and B. Schlicht 99 Dynamical Properties of IV-VI Compounds With contributions by H. Bilz, A. BussmannHolder, W. Jantsch, and P. Vogl 100" Quarks and Nuclear Forces

Editedby D. C. Fries and B. Zeitnitz

101 Neutron Scattering and Muon Spin Rotation With contributions by R. E. Lechner, D. Richter, and C. Riekel 102 Theory of Jets in Electron-Positron Annihilation By G. Kramer 103 Rare Gas Solids With contributions by H. Coufal, E. LL~scher, H. Micklitz, and R.E. Norberg 104 Surface Enhanced Raman Vibrational Studies at Solid/Gas Interfaces By I. Pockrand 105 Two-Photon Physics at e÷e - Storage Rings

By H. Kolanoski

* denotes a volume which contains a Classified Index starting from Volume 36.

Iven Pockrand

Surface Enhanced Raman Vibrational Studies at Solid/Gas Interfaces With 60 Figures

Springer-Verlag Berlin Heidelberg New York Tokyo 1984

Dr. Iven Pockrand Dr~.gerwerk AG, Postfach 1339 D-2400 L0beck 1, Fed. Rep. of Germany

Manuscripts for publication should be addressed to:

Gerhard H6hler Institut for Theoretische Kernphysik der Universit~.tKarlsruhe Postfach 6380, D-7500 Karlsruhe 1, Fed. Rep. of Germany Proofs and all correspondence concerning papers in the process of publication should be addressed to:

Ernst A. Niekisch Haubourdinstrasse 6, D-5170 J~lich 1, Fed. Rep. of Germany

ISBN 3-540-13416-6 Springer-Verlag Berlin Heidelberg New York Tokyo ISBN 0-387-13416-6 Springer-Verlag New York Heidelberg Berlin Tokyo Library of Congress Cataloging in Publicatfon Data. Pockrand, Iven, 1943- Surface enhanced Raman vibrational studies at solid/gas interfaces. (Springer tracts in modern physics; 104) Bibliography: p. 1. Raman effect, Surface enhanced. 2. Surfaces (Physics) 3. Surface chemistry. I. Title. II. Series. QC1.S797 voL 104 [QC454.R36] 539s [530.4'1] 84-5387

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically those of translation, reprinting, reuse of illustrations, broadcasting, reproduction by photocopying machine or similar means, and storage in data banks. Under § 54 of the German Copyright Law where copies are made for other than private use, a fee is payable to ,,Verwertungsgesellschaft Wort", Munich. © by Springer-Verlag Berlin Heidelberg 1984 Printed in Germany The use of registered names, trademarks, 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. Offset printing and bookbinding: BrL~hlsche Universit~,tsdruckerei, Giessee 2153/3130-54321 0

Preface

Many molecules adsorbed on appropriately prepared metal surfaces display a Raman scattering cross section which is several orders of magnitude greater than the corresponding quantity for the isolated molecule. This e f f e c t , surface enhanced Raman scattering (SERS), which was discovered eight years ago, opened the very interesting path to Raman vibrational spectroscopy of sub-monolayer quantities of adsorbates, whose study had formerly been thought to be without prospects because of the insuff i c i e n t s e n s i t i v i t y of ordinary Raman scattering. This book comprehensively reviews surface enhanced Raman vibrational studies of solid/gas interfaces. I t b r i e f l y illuminates the current state of understanding of SERS as inferred from relevant experimental r e s u l t s . Emphasis is put on the presentation and evaluation of SER vibrational data from various molecules adsorbed on metal surfaces, in p a r t i c u l a r s i l v e r and the other noble metals. In addition, app l i c a t i o n s of SERS to problems in t r i b o l o g y and catalysis as well as related surface enhanced phenomena l i k e enhanced nonlinear optical effects or infrared absorption are described. SER studies of metal electrodes and c o l l o i d a l suspensions are not treated since these are summarized in several other reviews. I hope that t h i s volume w i l l be a useful help for surface s c i e n t i s t s interested in vibrational spectroscopy of adsorbates and act as a stimulus for future work and progress in the f i e l d . Much of the work presented in the book has been performed during a four year stay at the "Physikalisches I n s t i t u t I I I " of the University of D~sseldorf. I would l i k e to thank Prof. A. Otto for the exciting times spent at his i n s t i t u t e and many stimulating, f r u i t f u l , and c r i t i c a l discussions. The s k i l f u l technical assistance of J. Liebetrau in the experimental work performed at DUsseldorf is highly appreciated. I am also indebted to Dr. J. Billmann for a careful reading of the manus c r i p t , to Mrs. B. Derks for the accurate execution of the drawings, and to Mrs. C. LUtjens for the fast and e f f i c i e n t processing of the manuscript. Many of my colleagues supported the work on t h i s review a r t i c l e by sending information and/or preprints prior to publication. I would l i k e to thank F. Adrian, A. Campion, R. Chang, A. Creighton, J. Demuth, S. Efrima, M. Kerker, P. Liao, H. Metiu, A. Nitzan, M. P h i l p o t t , G. Schatz, H. Seki, D. Tevault, J. Tsang, H. Ueba, K. Ushioda, R. Van Duyne, D. Weitz, T. Wood, and H. Yamada. F i n a l l y , I would l i k e to thank my wife Petra and my l i t t l e daughter Friederike, without whose patience with my almost permanent absence from home during the formation of the review t h i s book would never have been completed. LUbeck, June 1984

Iven Pockrand

Contents

1.

Introduction

2.

Fundamentals of Surface Enhanced Raman Scattering

3.

4.

.....................

2.1

Basic Experimental Observations .....................................

2.2

Theoretical

2.3

Present State of Understanding ......................................

Concepts ................................................

Experimental .............................................................

1 6 7 11 17 19

3.1

Arrangements ........................................................

19

3.2

Sample P r e p a r a t i o n and C h a r a c t e r i z a t i o n

20

Pyridine Adsorption 4.1

.............................

.....................................................

Coldly Evaporated S i l v e r

Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Features

....................................

26 27

4.1.1

General S p e c t r a l

4.1.2

Coverage Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

4.1.3

Annealing Behaviour ..........................................

37

4.1.4

Excitation

39

4.1.5

Comparison o f Results from Various Experiments . . . . . . . . . . . . . . .

Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

46

4.2

C o l d l y Evaporated Copper and Gold Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

4.3

Surfaces Prepared w i t h Various Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . .

51

4.3.1

Silver

51

4.3,2

Other M a t e r i a l s

4.4 5.

.............................................................

.......................................................

Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Hydrocarbon Adsorption 5.1

5.2

5.3

..............................................

.................................................

0 p e n - C h a i n H y d r o c a r b o n s on S i l v e r

...................................

53 54 59 59

5.1.1

Alkanes ......................................................

59

5.1.2

Ethylene .....................................................

60

5.1.3

P r o p y l e n e and B u t y l e n e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

73

5.1.4

Acetylene ....................................................

76

C y c l i c Hydrocarbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83

5,2,1

Benzene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83

5.2.2

Benzene D e r i v a t i v e s

86

..........................................

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

87 VII

6. C a r b o n M o n o x i d e E x p o s u r e a n d C a r b o n a c e o u s D e p o s i t s 6.1

...................

89

Adsorbed Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

90

6.2

Carbonaceous " I m p u r i t y "

94

6.3

Amorphous Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

96

6.4

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

7. O x y g e n E x p o s u r e

Deposits ....................................

........................................................

Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

99

7.1

Silver

7.2

Other . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

105

7.3

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

106

8. Water A d s o r p t i o n

........................................................

9. O t h e r A d s o r b a t e s

........................................................

101

107 112

9.1

D i a t o m i c Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

112

9.2

Azabenzenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

113

9.3

Pyridine Derivatives

114

................................................

9.4

Polymer and L a n g m u i r - B l o d g e t t C o a t i n g s

9.5

Dye Molecules

9.6

List

..............................

.......................................................

o f Systems S t u d i e d so Far . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I 0 . S e l e c t e d A p p l i c a t i o n s a n d R e l a t e d S u r f a c e E n h a n c e d Phenomena . . . . . . . . . . .

114 115 115 118

10.1

Tribology

..........................................................

118

10.2

Catalysis

..........................................................

119

10.3

Other Surface Enhanced Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11. S u m m a r y a n d O u t l o o k Appendix: References

....................................................

Recent Developments and Results

.................................

..................................................................

120 124 127 133

Subject ]ndex ................................................................

155

Material Index

159

VIII

...............................................................

List of Abbreviations

AES

Auger Electron Spectroscopy

AIS

Atom I n e l a s t i c Scattering

ATR

Attenuated Total Reflection

CMA

Cylindrical Mirror Analyzer

DIRS

Disorder Induced Raman Scattering

EELS

Electron Energy Loss Spectroscopy

FWHM

Full Width at Half Maximum

IETS

I n e l a s t i c Electron Tunneling Spectroscopy

IRAS

Infra - Red Absorption Spectroscopy

IRTS

Infra - Red Transmission Spectroscopy

LEED

Low Energy Electron Spectroscopy

ML

Mono-Layer

NIS

Neutron I n e l a s t i c Scattering

OMA

Optical Multichannel Analyzer

SER

Surface Enhanced Raman

SERS

Surface Enhanced Raman Scattering

TDS

Thermal Desorption Spectroscopy

UHV

Ultra High Vacuum

UPS

UV- Photoemission Spectroscopy

XPS

X - Ray Photoemission Spectroscopy

A~

Work function

L

Langmuir

(1L

D

Debye

( i D = 10-18 esu)

= 10-6 Torr s)

JX

1. Introduction

Vibrational spectroscopy has been employed for many years to study the structure and bonding of molecules. As each bond has i t s own, characteristic frequency /1-4/ vibrational spectra and molecular structure are related. Infrared absorption and Raman arrangements have most frequently been used for vibrational studies / 5 - 8 / . Other techniques l i k e i n e l a s t i c scattering of electrons /9,10/, neutrons /11,12/, or atoms /10/, which require more refined experimental set-ups, have found considerably less broad spreading as analytical tools. Molecules are usually perturbed upon adsorption on solid surfaces. Bond strengths and/or structure may change, new species may be formed due to dissociative adsorption or surface promoted reactions between d i f f e r e n t adsorbed molecules. Surface vibrational spectroscopy can provide s i g n i f i c a n t information on these changes. Character and concentration of the adsorbed species as well as adsorption geometry and s i t e might be extracted from the data. To f a c i l i t a t e evaluation, vibrational spectroscopy is usually backed by other surface sensitive techniques l i k e , e.g., u l t r a v i o l e t photoemission spectroscopy (UPS), low energy electron d i f f r a c t i o n (LEED), thermal desorption spectroscopy (TDS), or work function measurements (A~). Several experimental techniques have been developed to study vibrations of adsorbed molecules. Infrared transmission (IRTS) or infrared r e f l e c t i o n absorption spectroscopy (IRAS) /13,14/ and electron energy loss spectroscopy (EELS) /15,16/ have found widespread popularity. As outlined in detail in a recently published book /17/, neutron and atom i n e l a s t i c scattering (NIS and AIS) as well as i n e l a s t i c electron tunneling spectroscopy (IETS) /18/ are also becoming established as usef u l , surface sensitive techniques. Raman spectroscopy, however, did not a t t r a c t the attention of surface s c i e n t i s t s ( u n t i l recently, see below), although this method combines several advantages in a unique way. A resolution of 1 cm-1 and a free spectral range of 100-4000 cm-1 are e a s i l y obtainable, solid/gas surfaces under high pressure or s o l i d / e l e c t r o l y t e interfaces may be investigated in s i t u , and, f i n a l l y , one may get additional information from depolarization measurements. However, the s e n s i t i v i t y of Raman scattering is poor and has been believed to be in general i n s u f f i c i e n t for vibrational studies of adsorbed molecules. To i l l u s t r a t e this fact l e t us estimate the i n t e n s i t y which is i n e l a s t i c a l l y scattered from a

monolayer of molecules adsorbed on a perfectly r e f l e c t i n g metal surface. The detected Raman i n t e n s i t y IRaman is given by /19/: (1)

IRaman = 4~ (da/dQ)nNAQTsTo

where (da/d~) is the d i f f e r e n t i a l Raman cross section, ~ the collected solid angle, n the f l u x of incident photons, N the density of adsorbed molecules, A the i l l u m i nated area, Q the quantum e f f i c i e n c y of the detector, and Ts and To are the transmission of the spectrometer and the collecting optics, respectively. The factor four considers the influence of the p e r f e c t l y r e f l e c t i n g metal on the incident as well as on the Raman scattered photons. For a Lorentzian l i n e shape of the vibrational mode (halfwidth F), the integrated i n t e n s i t y IRaman can be converted into a peak i n t e n s i t y I peak by /19/: Raman iPeak Raman = IRaman/(~F)

(2)

For the specific case of adsorbed pyridine CsHsN one has N = 5.1014 molecules/cm 2 /20/ and a r e l a t i v e l y large (gas phase) Raman cross section of 3.3.10 -30 cm2/(sterad molecule) for the symmetric breathing vibration /21,22/. Using 200 mW of 514.5 nm Ar-ion laser radiation focused down to A = 3.10 -2 cm2, and assuming ~ = I sterad, Q = 0.15, Ts.T o = 5.10 -3 , and a photon counting system which records a l l m u l t i p l i e r pulses, one expects an integrated Raman i n t e n s i t y of at best 15 cts/s. This gives a peak i n t e n s i t y of ~ 2 cts/s (F = 4 cm- I , spectrometer bandpass 2 cm-l), which is not p a r t i c u l a r l y encouraging i f one thinks of surface vibrational studies. Indeed, early investigations yielded Raman signals only from strong Raman scatterers on high surface area adsorbents /23/ or from r e l a t i v e l y thick films (~ 5 nm) of strong scatterers on s i l v e r films /24/. Thinner overlayers may be detected, i f the background i n t e n s i t y in the Raman spectra is s u f f i c i e n t l y low. This has been shown for pyridine on Ag(llO) /25/ and N i ( l l l )

/26/. Figure la displays the Raman spectrum of

a sample coated with roughly three layers of pyridine. A peak i n t e n s i t y of ~ 10 cts/s in rough agreement with the expected value has been measured. When using a more elaborate experimental arrangement, Raman spectra from less than a monolayer of adsorbed molecules on N i ( l l l )

/27/ or Ag(111) (Fig. ib, /28/) can be taken. These re-

sults open the very interesting path to Raman vibrational studies of adsorbates on well characterized single crystal surfaces (with, however, s t i l l

moderate s e n s i t i v -

i t y i f compared to, e.g., EELS). As seen from ( I ) , IRaman can be increased when the f l u x n or, correspondingly, the electromagnetic f i e l d strength at the s i t e of the Raman scatterer is increased. This has been accomplished in the early seventies for thin organic layers by incorporating these films into suitable, layered structures /29-32/, so that guided l i g h t modes in the f i l m or plasmon surface polaritons at the film/metal interface can

(a) :~

7[

(b) 992cm-1

100

1037cn~1 996cm-1

Roman shift Fig. 1. Ordinary Raman spectra from pyridine on s i l v e r , a) ~ 3 layers on Ag(110), Ts'= 150 K; 250 mW of 514.5 nm r a d i a t i o n , 2 cm-1 bandpass;~ b) ~ 1 monolayer on A g ( l l l ) , T~ = 110 K; I000 mW of 514.5 nm r a d i a t i o n , I0 cm- I bandpass ( a f t e r / 2 8 / ) . Symmetric ~992/996 cm- I ) and antisymmetric breathing mode (1030/1037 cm-1)

propagate (these optical modes are discussed i n , e . g . , / 3 3 / ) ,

Raman spectra of good

q u a l i t y have been recorded, when a t h i n f i l m or interface mode was resonantly exc i t e d . The influence of long wavelength, extended surface plasmon polaritons on the Raman scattering from molecules on h i g h l y r e f l e c t i n g metal surfaces (e.g. Ag) has subsequently been investigated in some more d e t a i l , Besides the resonant enhancement of the i n c i d e n t f i e l d the c a l c u l a t i o n s generally consider also the resonant emission of the Raman scattered photons via plasmon surface p o l a r i t o n s . For the attenuated t o t a l r e f l e c t i o n (ATR) configuration /34-36/ as well as f o r grating surfaces /37-41/ enhancements of the Raman scattered i n t e n s i t y of 103- 106 have been calculated under favourable conditions (the ATR technique is described i n , e . g . , / 3 3 / , plasmon surface p o l a r i t o n s on gratings are treated in, e . g . , / 4 2 / ) .

Experi-

mentally observed enhancements are usually much smaller, between ~ 5 and ~ 100 (/43-48/; only recently a rather large f a c t o r of ~ 4.104 has been estimated from an ATR-Raman study /49/. Grating surfaces have e s p e c i a l l y been used in tunnel j u n c t i o n structures /50-54/). Nevertheless, the enhancement brought about by e x c i t a t i o n of surface waves may render possible or f a c i l i t a t e

surface Raman v i b r a t i o n a l

studies in certain cases. As t h i s technique is only applicable to molecules adsorbed on certain materials with appropriately corrugated surface or incorporated into an ATR c o n f i g u r a t i o n , i t did not f i n d , however, widespread a t t e n t i o n and i n t e r e s t in the community of surface s c i e n t i s t s . This was d i f f e r e n t , when two research groups independently reported a giant enhancement (105- 106 ) of the Raman cross section of pyridine on s i l v e r electrodes /55,56/ ( a c t u a l l y , s i m i l a r Raman spectra from pyridine on Ag electrodes had been published e a r l i e r /57-59/; however, these authors did not r e a l i z e the unusual enhancement). Enhanced Raman signals were only observed a f t e r a proper a c t i v a t i o n of the electrode by an o x i d a t i o n - r e d u c t i o n cycle. Soon a f t e r the f i r s t

report of sur-

face enhanced Raman scattering (SERS) from s i l v e r electrodes SERS was also observed from molecules on s i l v e r c o l l o i d s /60/ and on s i l v e r / g a s (vacuum) interfaces /61/. I t became apparent, that the surfaces of only certain metals were SERS "active" 3

(group Ib mainly), which had to be pretreated appropriately or prepared under spec i a l conditions (the significance of t h i s SERS " a c t i v i t y " w i l l be outlined l a t e r ) . Moreover, not a l l ,

although many~ molecules adsorbed on SERS active surfaces d i s -

played enhanced Raman scattering equally w e l l . Several mechanisms have been proposed to contribute to SERS ( i n c l u d i n g e x c i t a t i o n of extended plasmon surface polaritons as mentioned above). The various models have been extensively discussed in several review a r t i c l e s /62-68/ and a recently published book on SERS /69/. Therefore they w i l l be only b r i e f l y exposed in Chapt. 2 of t h i s volume. Some basic experimental facts and the t h e o r e t i c a l concepts, which are presently accepted by most groups active in the f i e l d , w i l l as well be summarized in t h i s chapter. The experimental s i t u a t i o n is less comprehensively reviewed. Early experimental results from s o l i d / e l e c t r o l y t e interfaces are contrasted with t h e o r e t i c a l concepts in /19,70/, applications of Raman spectroscopy in electrochemistry are discussed in / 7 1 / , and some experimental observations from s o l i d / e l e c t r o l y t e as well as solid/gas interfaces are l i s t e d in /72/. Several recent a r t i c l e s /73-76/ discuss selected, relevant observations in connection with actual theoretical developments. This review summarizes SER experimental studies from solid/gas interfaces. The impact of experimental facts on the theoretical discussion w i l l be displayed, but no e f f o r t is made to comprehensively appraise theoretical concepts ( f o r t h i s the interested reader is referred to /66-69/). Rather, surface enhanced Raman v i b r a t i o n a l spectra from various molecules on metal surfaces w i l l be evaluated in some d e t a i l . Therefore the volume addresses also the surface s c i e n t i s t , who is not p a r t i c u l a r l y i n terested in the d e t a i l s of the theoretical discussion, but rather wants to be i n formed of the applications and potential of surface enhanced Raman scattering as a surface a n a l y t i c a l t o o l . The paper is organized as follows. A f t e r a b r i e f survey of basic experimental facts, proposed models, and the present state of the theory discussion (Chapt. 2), some d e t a i l s of the experimental techniques w i l l be i l l u m i n a t e d in Chapt. 3. A rather detailed analysis of SER data from pyridine on metals w i l l be presented in Chapt. 4. The relevance of some results with respect to t h e o r e t i c a l concepts w i l l be accentuated. SER v i b r a t i o n a l spectra from hydrocarbons~ carbon monoxide and carbonaceous species, oxygen, and water adsorbed to metals are discussed in Chapts. 5-8. Data from other, less f u l l y investigated adsorbate/metal systems are summarized in Chapt. 9. Relative broad room w i l l be given to results from molecules on "coldl y " evaporated f i l m s , since these usually display the most detailed spectra ("coldl y " evaporated films are characterized in Chapt. 3). In Chapt. 10 some applications of SERS to problems i n , e . g . , t r i b o l o g y or c a t a l y t i c a c t i v i t y of metal surfaces are presented. F i n a l l y , momentary problems and state of the a r t are reflected in Chapt. II.

In an outlook, future c a p a b i l i t i e s and l i m i t a t i o n s of Raman spectroscopy as a

surface a n a l y t i c a l tool are displayed.

This a r t i c l e almost t o t a l l y ignores the very i n t e r e s t i n g and important SERS work on metal electrodes and c o l l o i d a l suspensions. The reader, who is also i n terested in these aspects of SERS, is referred to another review /77/.

2. Fundamentals of Surface Enhanced Raman Scattering

In t h i s chapter we give an overview on the present experimental as well as theor e t i c a l s i t u a t i o n of SERS. No attempt is made to exhaustively quote a l l related work, and I apologize to those, whose work did not find the attention i t deserves. A few points require special comments. F i r s t l y , much confusion has been i n t r o duced into the f i e l d by experimental papers, whose results or interpretations were not c a r e f u l l y enough cross checked. I shall express scepticism, whenever i t is necessary, i . e . when results could not be reproduced. Secondly, a v a r i e t y of observations has been c l a s s i f i e d as SERS, often without elaborating the specific properties of the system under investigation. The "giant" enhancement (105-106 ) at appropriately pretreated s i l v e r electrodes /55,56/, the weak e f f e c t (enhancement 5- i00) when resonantly exciting plasmon surface polaritons at optical gratings /46/ or in an ATR configuration /43,44/, as well as Raman spectra from adsorbed molecules on for instance s i l i c a supported Ni catalysts /78/ have a l l been simply labeled SERS. To the outsider not f a m i l i a r with the f i e l d t h i s may have suggested one single enhancement mechanism s i m i l a r l y working in quite d i f f e r e n t systems (which is a wrong picture). Thirdly, theoretical concepts developed for special configurations l i k e gratings or isolated metal spheres have sometimes been i n t r o duced so, as i f they were capable to explain a l l or almost a l l aspects of SERS in every system. Unfortunately, the s i t u a t i o n is more complicated. F i n a l l y , I would l i k e to remind of the "pre SERS" Raman work on adsorbed molecules. Numerous papers report on (ordinary) Raman studies of molecules physi- or chemisorbed on Ni single crystal surfaces (/79/, see also / 2 7 / ) , oxide surfaces l i k e s i l i c a or alumina /80, 81/, supported metal catalysts /82/, or metal electrodes /83,84/. Several review a r t i c l e s summarize these investigations /59,85,86/. The important messages from these studies are: ( i ) only an extremely careful q u a n t i t a t i v e evaluation of scattered i n t e n s i t i e s allows to safely decide, whether an observed weak Raman signal is surface enhanced or not, and ( i i )

laser Raman spectroscopy - ordinary or en-

hanced - can provide valuable information on adsorbed molecules.

2.1

Basic Experimental Observations

As was already evident from the very f i r s t SER studies at s i l v e r electrodes, only samples activated by an anodic oxidation-reduction cycle exhibited strongly enhanced Raman signals from adsorbed molecules (/55-57/; enhancement factor 105-106). The pretreatment ( a c t i v a t i o n ) has been shown /87,88/ to change the surface topography: SERS active s i l v e r electrodes are rough. The roughness scale "important" for SERS is s t i l l

a matter of debate /64,67,89/. There i s , however, agreement that some kind

of roughness is a necessary prerequisite for SERS /64/. SERS is not r e s t r i c t e d to s i l v e r electrodes. Enhanced Raman signals have been reported for several other activated s i l v e r interfaces or s p e c i a l l y prepared systems: -

mechanically polished, p o l y c r y s t a l l i n e s i l v e r sheets measured in a i r (/90/; no q u a n t i t a t i v e estimation of the enhancement factor for cyanide deposited by immersion in alkaline KCN solution) s i l v e r island films vapour deposited on glass (/62,91/; enhancement factors of 105 have been observed for adsorbed i s o n i c o t i n i c acid /92/, p-nitrobenzoic acid /93/, and pyridine /94/) s i l v e r aqueous sol p a r t i c l e s of dimensions comparable to or less than the wavelength of l i g h t (/60/; for c i t r a t e ions adsorbed to s i l v e r p a r t i c l e s of 42 nm diameter an enhancement factor of 6.105 has been measured /95/) s i l v e r optical gratings with periods comparable to the wavelength (/46/; a weak enhancement of ~ 30 for thin polystyrene coatings /47/ and of ~ 102 for pyridine /48/ due to resonant excitation of plasmon surface polaritons has been observed) p o l y c r y s t a l l i n e s i l v e r f o i l s cleaned and probably roughened by Ar-ion bombardment in UHV (/96,97/; for adsorbed pyridine enhancement factors of 103-105 /96/ and ~ 103 /97/ have been estimated, where the f i r s t value is uncertain because of d i f f i c u l t i e s

in measuring the dosing rate)

photochemically roughened s i l v e r surfaces with roughness features of t y p i c a l l y 50 nm l a t e r a l extensions (/98/; an enhancement factor of ~ 5.104 for pyridine has been observed) coldly evaporated s i l v e r f i l m s , i . e . thick s i l v e r films evaporated on substrates cooled to t y p i c a l l y 120 K (/61/; Raman signals from adsorbed pyridine display an enhancement of ~ 104 /99/) AI-AI203-Ag tunnel junctions evaporated on rough CaF2 films or on optical gratings (/51/; for 4-pyridine-carboxaldehyde at the AI203-Ag interface an enhancement of ~ 20 for junctions on gratings has been estimated due to resonant excitation of plasmon surface polaritons; CaF2 roughened structures have not been evaluated q u a n t i t a t i v e l y ) . The enhancement factors given above have usually been determined f o r the strongest l i n e of the adsorbate (for pyridine, this is the symmetric ring breathing v i b r a t i o n ) . The q u a l i t y of measured spectra in terms of peak i n t e n s i t i e s and signal

i

L

r

I

I

]

Ic4ts/s00

cts/s

E

i

i

1100

I

Fi 9. 2. SER signals from 0.1 monolayer of pyridine on Ag. Left: photochemicallYlroughened surface (488 nm r a d i a t i o n , 8 cm- bandpass; a f t e r / 9 8 / ) . Right: c o l d l y evaporated f i l m (170 mW of 514.5 nm r a d i a t i o n , 3 cm- I bandpass; a f t e r / i 0 0 / ) For both cases an enhancement factor of ~ I04 has been estimated

i

1000 1100 1000 Roman shift ( cm-1 )

to noise r a t i o is rather d i f f e r e n t for d i f f e r e n t systems, even when comparable enhancement factors are estimated. The spectra displayed in Fig. 2 have been recorded under s i m i l a r experimental conditions. They show surface enhanced Raman signals from roughly a tenth of a monolayer of pyridine on photochemically roughened s i l v e r /98/ and on c o l d l y evaporated s i l v e r f i l m s /100/. Either signal has been estimated to be about four orders of magnitude enhanced. Figure 2 might indicate a too o p t i m i s t i c valuation of enhancement factors in some cases. Well prepared, smooth single crystal surfaces do not enhance the Raman signal from adsorbed molecules beyond that what is expected from Fresnel equations (/25,28/, pyridine on Ag). A weak enhancement of ~ 4.102 reported for pyridine on Ag(lO0) /101/ might be the combined r e s u l t of residual roughness as o u t l i n e d in / 2 8 / , of the f l a t

10

I

i

i

i

I

I

I

i

L peak intensity 2 xl0 3 cts/s

~6 o

I

I

2 0

~

3000

2000 Roman shift (cm-1)

tGO0

Fig. 3. SER spectra from c o l d l y evaporated Ag films exposed to 3 L of ethylene. Upper trace: sample exposed and measured at 120 K. Lower trace: sample annealed to 260 K (~ iK/min I , recooled to 120 K, exposed and measured. 200 mW of 514.5 nm rad i a t i o n , 4.5 cm- bandpass. A f t e r /109/

metal surface c o n t r i b u t i o n given by the Fresnel equations /102/, and of uncertaint i e s in the evaluation procedure. SER l i n e s from adsorbates are accompanied by a continuous background scattering which extends beyond 4000 cm-1 /90,103,104/. The background is also observed from SERS active surfaces without adsorbed molecules / 1 0 5 - 108/. Hence i t is an i n t r i n s i c property of the metal. For s i l v e r electrodes, both, background and SER l i n e s , i n crease with increasing a c t i v a t i o n /104/. Therefore the two phenomena may have important features in common, as assumed in /106,107/. The c o r r e l a t i o n of background and SER i n t e n s i t y i s , however, not always observed as shown in Fig. 3 f o r ethylene on c o l d l y evaporated s i l v e r films /109/. The background has been interpreted as luminescence /110,111/ due to roughness assisted, r a d i a t i v e decay of e l e c t r o n i c exc i t a t i o n s from a continuum of states /112,113/. Ordinary Raman selection rules are relaxed in SERS. IR allowed v i b r a t i o n s of centrosymmetric molecules, which are subject to the p r i n c i p l e of mutual exclusion, have been detected in SER spectra ( f o r instance pyrazine on s i l v e r electrodes /114,115/). Even s i l e n t modes were observed with appreciable i n t e n s i t y (e.g., benzene on s i l v e r films /116/). The breakdown of selection rules has been a t t r i b u t e d to the change of symmetry upon adsorption /114,116/ or, a l t e r n a t i v e l y , to the large e l e c t r i c f i e l d gradients which e x i s t near a metal surface /116-118/. In /67/ yet another explanation based on the "charge t r a n s f e r picture of the chemical c o n t r i b u t i o n to SERS" (see below) is given. Relative SER l i n e i n t e n s i t i e s d i f f e r in general from the corresponding values of ordinary scattering from the isolated molecule, i . e . mode selective enhancement is observed /55/. The r e l a t i v e SER i n t e n s i t i e s vary with electrode potentials /19, 88/ and with the wavelength of the e x c i t i n g l i g h t /19,119/. They are d i f f e r e n t f o r the same molecule adsorbed to d i f f e r e n t , SERS active metal substrates /120- 123/, and they are d i f f e r e n t f o r d i f f e r e n t l y prepared SERS active surfaces of the same metal. The most s t r i k i n g example for the l a t t e r is the SER signal of the C-H s t r e t c h ing v i b r a t i o n s of pyridine and of other molecules on s i l v e r .

I t is comparable in

i n t e n s i t y to the strong breathing mode signal for activated electrodes / 5 5 / , whereas i t is roughly two orders of magnitude smaller f o r c o l d l y evaporated films (/108/; see also Chapt. 4). Overtones and combination bands are absent or only weakly pronounced /124/. SER lines are depolarized, even i f the corresponding l i n e s of the isolated molecule are strongly polarized /64/. SER e x c i t a t i o n spectra do not show sharp resonances, Only slow v a r i a t i o n s or broad maxima have been observed. E x c i t a t i o n p r o f i l e s are d i f f e r e n t for electrodes /19,46,119,125-127/, f o r c o l l o i d a l dispersions /60,95,128- 131/ and m a t r i x - i s o l a t e d p a r t i c l e s /132,133/, and for vapour-deposited films in vacuum /94,119,123,134-136/ and island films /93,137,138/. E x c i t a t i o n spectra and t h e i r impact on t h e o r e t i c a l concepts w i l l be discussed in some d e t a i l in Chapt. 4.

Besides s i l v e r , which is s t i l l

the most widely used material for SER studies,

appropriately prepared Cu and Au surfaces are SERS active as well under red l i g h t illumination (/123,139,140/; SER signals disappear for e x c i t a t i o n wavelengths 570 nm /123/). Other highly r e f l e c t i n g materials l i k e the a l k a l i metals l i t h i u m /133/, potassium /141/, and sodium /142/ e x h i b i t also SERS. Preliminary results from A1 films /143,144/ require further careful experiments to establish the degree of the enhancement (very recently: a weak enhancement of ~ 103 has been reported for p-nitrobenzoic acid on aluminum p a r t i c l e arrays /145/). Alloying gold to s i l v e r quenches the SER signal f o r green/blue l i g h t e x c i t a t i o n , but not for red /146/; 5% Pd in Ag also quenches the enhancement below the l i m i t of detection /147,148/. Other reports of SERS from low r e f l e c t i v i t y metals l i k e Hg /149,150/, Cd /151/, Pd /152-154/, Pt /152-157/, Ti /154/, and Ni /153,154,157-161/ should be valuated very c r i t i c a l l y .

Some results could not be reproduced by other groups (e.g. Hg,

Cd /162/), and some might be interpreted in terms of ordinary resonance or pre-resonance Raman scattering rather than in terms of SERS (J2 on Pt, Pd /152/). Because of the high area surface of supported metal catalysts the r e l a t i v e l y weak signals from Ni, Pt /156,158-161/ might be j u s t ordinary Raman scattering (independent of the interpretation these Raman vibrational studies of supported metal catalysts yielded interesting r e s u l t s ) . F i n a l l y , SERS has been reported for pyridine on metal oxides (NiO /152/, TiO2 /154/), for iodine on a semiconductor electrode (TiO 2, /163/), and for molecular oxygen on an organic single crystal (polydiacetylene, /164/). The l a t t e r has been attributed to adsorption induced resonance Raman scattering. Molecular oxygen forms a complex with polydiacetylene with a well defined electronic t r a n s i t i o n at 2.39 eV /164/. I t is interesting within the context of the present discussion of SERS mechanisms (/67/ and Sect. 2.2) that this t r a n s i t i o n involves a s i g n i f i c a n t degree of charge transfer. The former two results as well as the studies on low r e f l e c t i v i t y metals require more careful experimental work to unambiguously clear the s i t u a t i o n . There seems to be no l i m i t a t i o n in molecules which e x h i b i t SER5. Enhanced Raman signals have been observed from simple adsorbates l i k e , e.g., halide ions /165/ and complicated molecules l i k e pyridine derivatives or nucleic acid components (/166, 167/; a l l on s i l v e r electrodes). However, the magnitude of the enhancement might be d i f f e r e n t for d i f f e r e n t molecules on the same surface as has recently been demonstrated for coadsorbed CO and N2 /133/. There is a " f i r s t layer effect" in surface enhanced Raman scattering: molecules in the f i r s t

layer often show a much stronger enhancement than those in consecutive

layers /99,136/. The e f f e c t might be r e s t r i c t e d to s p e c i f i c a l l y adsorbed molecules in the f i r s t

10

layer as is assumed within the concept of SERS active sites /67,87/.

2.2

Theoretical

Concepts

The extensive t h e o r e t i c a l work on surface enhanced Raman scattering is summarized in several reviews. A survey of early concepts is presented in / 6 3 / , more recent work is b r i e f l y i l l u m i n a t e d in /102/. A detailed, c r i t i c a l

valuation of some SER

models can be found in /66/ (electromagnetic effects at various SERS active surfaces), in /67/ ("electromagnetic" and "chemical" c o n t r i b u t i o n s to SERS), and in

168/. The i n e l a s t i c a l l y scattered i n t e n s i t y IRaman f o r an isolated molecule may be w r i t t e n as /168/: 4 2 F2 IRaman ~ Ws.l~l 9

(3)

Here ms is the Stokes frequency, F the e l e c t r i c f i e l d strength of the i n c i d e n t rad i a t i o n , and ~ a component of the Raman tensor /168/ (phenomenologically, ~ describes the normal coordinate d e r i v a t i v e of the p o l a r i z a b i l i t y of the molecule). IRaman is usually very small, much smaller than the e l a s t i c a l l y scattered Rayleigh i n t e n s i t y /168/. I t may be up to six orders of magnitude stronger when the i n c i d e n t frequency is in resonance with a real t r a n s i t i o n of the molecule (resonance Raman e f f e c t , see, e . g . , /169/). Upon adsorption on a metal surface, the e l e c t r o n i c states of a molecule which shows only ordinary scattering may be perturbed such as to allow f o r resonance Raman s c a t t e r i n g . This "adsorption induced resonance Raman e f f e c t " led to the p r e d i c t i o n of enhanced Raman scattering from molecules near a metal surface in /170/, which has been published before the discovery of SERS ( i n /170/ i n teraction between excited molecular states and surface plasmon modes is thought to perturb the molecule). An " e f f e c t i v e " Raman tensor % f f may take into account such e f f e c t s . More general, for an adsorbed molecule (3) has to be modified to IRaman - ~ . l ~ e f f l

2 (F2.GL).Gs 9

(4)

Now ~ e f f contains any change of the molecular p o l a r i z a b i l i t y upon adsorption or, more e x a c t l y , describes the p o l a r i z a b i l i t y d e r i v a t i v e of the adsorbate/adsorbent system. GL and GS account for "electromagnetic" effects: the e l e c t r i c f i e l d at the s i t e of the adsorbed molecule and the Stokes emission might both be amplified by the presence of the metal. I shall f i r s t

discuss electromagnetic e f f e c t s . These are small on f l a t metal

surfaces. Due to interference e f f e c t s , the local f i e l d as well as the Stokes emission f i e l d might each have up to twice the value of the corresponding q u a n t i t y f o r the isolated molecule /171/. Since the local f i e l d is almost perpendicular to the surface, " s e l e c t i o n rules" appear as o u t l i n e d in /102/ (see also /172/). This might allow to determine adsorption geometries ( s i m i l a r l y , EELS (dipole scattering) and IRAS are governed by selection rules / 1 7 / ) .

The evanescent f i e l d s of extended plasmon surface polaritons /33/ can give rise to stronger electromagnetic effects, As mentioned in Chapt. I , these interface modes can resonantly be excited in an ATR configuration /33/ or at a grating surface /42/. For the former, a resonance enhancement of the i n t e n s i t y of the local e l e c t r i c f i e l d at a s i l v e r surface of ~ 250 /34/ and an enhanced Stokes emission on the prism side of ~ 300 /36/ has been calculated. This gives an overall enhancement of ~ 7.104. For optical s i l v e r gratings, the corresponding values are ~ 104 for GL /39/ and ~ 5-102 f o r GS /173/ r e s u l t i n g in a t o t a l enhancement of ~ 5.106 . Taking into account radiat i v e damping of surface p o l a r i t o n s , a considerably smaller GL of ~ 25 has been calculated (/174/; GS should be affected s i m i l a r l y ) . As is obvious from a comparison to measurements (see Chapt. i ) , most calculations overestimate the plasmon surface p o l a r i t o n c o n t r i b u t i o n to the enhancement. S i m i l a r l y , Raman scattering from molecules on isolated metal p a r t i c l e s (e.g. on d i l u t e d c o l l o i d a l dispersions) is enhanced by electromagnetic resonances ( l o c a l i z e d surface plasmons). Calculations w i t h i n the R a y l e i g h - s m a l l - p a r t i c l e l i m i t have been performed for spheres /175- 178/ and spheroids /179/, rigorous electrodynamic calculations f o r spheres /180,181/ and, very recently, f o r prolate e l l i p s o i d s /182/. Numerical evaluations generally assume bulk optical properties f o r the small metal p a r t i c l e s . Only in /183/ the size dependence of the imaginary part of the d i e l e c t r i c function due to surface scattering has been taken into account. D i e l e c t r i c environment effects have been studied in /184,185/. Depending on the shape and the dimensions of the p a r t i c l e , t o t a l enhancement factors of ~ 102- 106 have been calculated for silver particles. I f a rough metal surface is modeled by an ensemble of isolated hemispheroids protruding from a p e r f e c t l y conducting plane, the same formalism as for isolated spheroids may be used to calculate enhancement factors (/186- 188/; note, that only a perpendicular resonance exists for t h i s configuration because of the image e f f e c t of the i d e a l l y conducting plane). For sharp, needle-like protrusions enhancement factors up to 1011(!)

for molecules on the t i p of the structure have been calcu-

lated due to resonant e x c i t a t i o n of surface plasmons and the l i g h t n i n g rod e f f e c t , i . e . the concentration of the e l e c t r i c f i e l d at parts of the surface with extreme curvature. A s l i g h t l y more r e a l i s t i c case has been treated in /189/. Here e l e c t r o magnetic resonances of an isolated hemispherical bump protruding from the plane boundary of a halfspace with the same d i e l e c t r i c function as the bump material have been studied. Numerical results are presented for a p a r t i c u l a r p o s i t i o n of the scattering molecule only. The calculations mentioned so far neglect i n t e r a c t i o n between the metal p a r t i c l e s or the bumps on the surface. Because of the long range of the elctromagnetic f i e l d s of the resonance, t h i s is usually a crude approach, C o l l e c t i v e i n t e r a c t i o n s have been treated with Maxwell-Garnett theory /136,190,191/. Within t h i s frame, the op-

12

t i c a l properties of metal spheres in a d i e l e c t r i c matrix are described by an effect i v e d i e l e c t r i c function, which contains - besides the d i e l e c t r i c function of the metal and the environment - only the volume f r a c t i o n ( f i l l i n g

factor) of the metal

/192,193/. A rough, bumpy surface is modeled by a t r a n s i t i o n layer, whose e l e c t r o magnetic resonance is then given by the Maxwell-Garnett approach /190/. No absolute numerical r e s u l t s have been presented. For the simple case of two metal spheres, electromagnetic i n t e r a c t i o n leads to the appearance of two resonances, whose s p l i t t i n g depends on the i n t e r p a r t i c l e distance, and to a substantial enhancement of the f i e l d between the spheres /194/. In a d i f f e r e n t approach the dipole moments induced in the metal p a r t i c l e s or bumps are treated as point dipoles. Dipole-dipole coupling between randomly dist r i b u t e d small p a r t i c l e s in a d i e l e c t r i c host broadens the electromagnetic resonance and s h i f t s i t to the red (with respect to the Maxwell-Garnett r e s u l t ; /195/). The broadening leads to a decrease of GL as well as of Gs. As shown in /196/, the transverse c o l l e c t i v e electromagnetic resonance of a square array of uniformly shaped obl a t e e l l i p s o i d s on glass gives a t o t a l enhancement factor of ~ 3.106 f o r molecules uniformly adsorbed on the e l l i p s o i d s ,

In t h i s case, which is regarded as represen-

t a t i v e f o r an island f i l m , the c o n t r i b u t i o n of the image dipoles to the t o t a l f i e l d has also been taken into account. I t has been pointed out /196/, that the transverse resonance w i l l be inhomogeneously broadened due to randomly d i s t r i b u t e d sizes, shapes, o r i e n t a t i o n s , and spacings of the islands in an actual evaporated f i l m , which may reduce the enhancement by two orders of magnitude. The c a l c u l a t i o n s have recently been extended to spheroids of any shape in ordered square l a t t i c e s or on random positions, and to square arrays of spheroids of random shape /197/. Effects of r e t a r d a t i o n , r a d i a t i v e damping, as well as f i n i t e size contributions to the d i e l e c t r i c response of the island f i l m were discussed. An i n t e n s i t y enhancement (GL) of I - 2

orders of

magnitude was estimated. A r e a l , rough metal surface may be described by a random d i s t r i b u t i o n of metal hemispheroids on a p e r f e c t l y conducting f l a t plane /198,199/. I f dipolar coupling between the protrusions is taken into account, GL is calculated to ~ 102 for Ag (/198/; average over the whole surface). Small scale, randomly d i s t r i b u t e d roughness may be treated with f i r s t

order perturbation theory (Born approximation) as has been

done f o r e l a s t i c Rayleigh scattering / 2 0 0 - 2 0 3 / .

The approach breaks down for 6

3 nm, where 6 is the rms-value of the roughness amplitude /204/. The roughness i n duced increase of the r a d i a t i o n from an o s c i l l a t i n g dipole r e l a t i v e to the f l a t surface has been estimated to ~ i0 w i t h i n t h i s model /204/ ( s i l v e r ; ~ = 3nm; ~ = 2 nm, which is at the l i m i t of v a l i d i t y of the Born approximation; ~: c o r r e l a t i o n length). With a d i f f e r e n t approach somewhat l a r g e r enhancement factors (102- 103 ) have recently been calculated (/205/; 6 = 15nm, ~ = 40nm). Several other approaches /35,93,206/ to q u a n t i t a t i v e l y estimate the e f f e c t of roughness on e x c i t a t i o n and emission of a Raman dipole on a metal surface are treated 13

in /67/. Here the interested reader w i l l f i n d a comprehensive, c r i t i c a l

discussion

of proposed models for the electromagnetic enhancement. Let us now b r i e f l y touch the "molecular enhancement" mechanisms, which are contained in the e f f e c t i v e p o l a r i z a b i l i t y ~ e f f of (4). The "image f i e l d " model /171, 207-213/ considers the influence of the image f i e l d on the adsorbate p o l a r i z a b i l i t y ( i n /213/, the influence of other adsorbed molecules and t h e i r images on the f i e l d at a given molecule has also been taken into account). The adsorbed molecule is usually treated as a point dipole located at a certain distance R from a sharp metal boundary. The e f f e c t i v e p o l a r i z a b i l i t y d e r i v a t i v e ~eff of the system (dipole plus image dipole) then varies r a p i d l y with the metal-adsorbate separation and may be h i g h l y peaked w i t h i n a small i n t e r v a l of distances. For s i l v e r , an enhancement fact o r of ~ 107 f o r R = 1.41 ~ has been calculated, which drops by more than three orders of magnitude when moving the molecule by only 0.1 ~ to R = 1.50 ~ /211/. More r e a l i s t i c , refined image f i e l d models /206,214-218/, which take i n t o account the f i n i t e molecular size, (and/or) spatial dispersion of the metal d i e l e c t r i c response, (and/or) the continuous v a r i a t i o n of the electron density across the i n t e r f a c e , or use a coupled-state quantum formalism / 2 1 9 - 221/, y i e l d considerably smaller enhancement factors. The status is at present s t i l l

uncertain, since d i f f e r e n t groups e s t i -

mate enhancement factors of ~ I /216/ and 104 /214/ f o r apparently s i m i l a r systems and approaches /212/. A second group of models considers the i n t e r a c t i o n of the v i b r a t i n g ion cores of the adsorbed molecule with the electrons of the metal /39,112,132,222-226/. A l l are based on the idea, that p a r t i c i p a t i o n of the h i g h l y polarizable metal electrons in the Raman process may enhance the cross sections. The ion cores may i n t e r a c t with the metal electrons via coulomb forces /112/ and thus modulate the e l e c t r o n i c polarizability

at the surface giving rise to so called "Raman r e f l e c t i o n " /222,223/.

For chemisorbed molecules t h i s mechanism may be accompanied by v i b r a t i o n a l l y modulated charge t r a n s f e r to and from the metal into the molecule /132,224,225/, which also modulates the e l e c t r o n i c p o l a r i z a b i l i t y .

Yet another mechanism is investigated

in /39/. Here i t is assumed, that the motion of the molecular i o n i c charges modulates the surface b a r r i e r potential f o r tunneling of metal electrons to the molecul a r s i t e . This, on the other hand, modulates the surface charge density induced by the e x c i t i n g f i e l d , which r e s u l t s in the emission of Raman Stokes photons. Enhancement factors of I 0 - i00 /102/ due to v i b r a t i o n a l modulation of metal electrons have been estimated f o r f l a t surfaces. F i n a l l y , we b r i e f l y touch models which may be summarized under "adsorption i n duced resonance Raman s c a t t e r i n g " . Within t h i s frame i t is assumed, that ( i ) the e l e c t r o n i c states of the molecule are perturbed by i n t e r a c t i o n with the metal, and/or ( i i )

an additional t r a n s i t i o n from metal states below the Fermi level to

the lowest unoccupied molecular level becomes possible so as to allow f o r ordinary resonant Raman scattering. Early papers /170,209,227,228/ focused on the coulomb 14

i n t e r a c t i o n between molecule and metal. This process is of long range, i . e . not res t r i c t e d to the f i r s t

layer of adsorbed molecules. S h i f t and broadening of the mole-

cular l e v e l s were estimated by using a formalism, which was developed to describe the properties of an o s c i l l a t i n g dipole close to a metal surface /229/. F i n i t e molecular size and nonlocal metal response have been included in a recent treatment of v i b r a t i o n a l properties of diatomic molecules on metals /230/. Another, rather special mechanism - formation of a surface complex upon adsorption with a new o p t i c a l t r a n s i t i o n in the v i s i b l e /126/ - gives enhanced Raman scattering only f o r molecules in d i r e c t contact with the metal. The same holds for the equally rather special s i t u a t i o n discussed in /231/. The importance of charge transfer e x c i t a t i o n s f o r SERS /65, 232- 238/ (case ( i i )

from above) has recently been discussed in d e t a i l /67/. This

mechanism requires chemisorption of the adsorbate, i . e . is a short range e f f e c t . Along with other processes mentioned e a r l i e r /132,224,225/, which also involve charge t r a n s f e r from and to the metal, i t is usually called the "chemical" c o n t r i b u t i o n to SERS. The role of charge t r a n s f e r e x c i t a t i o n s in SERS may best be i l l u s t r a t e d by the approach of /236/. Here i t is assumed that the lowest unoccupied level of the molecule is broadened to a resonance upon chemisorption due to p a r t i a l f i l l i n g

of t h i s

o r b i t a l by metal electrons. The t r a n s i t i o n of electrons from metal states below the Fermi energy to about the maximum in the density of states of the molecular resonance gives r i s e to a weak resonance in the Raman cross section. Enhancement factors of 50 f o r a " t y p i c a l case" of chemisorption on s i l v e r have been estimated /236/. The magnitude of any chemical c o n t r i b u t i o n to SERS depends most l i k e l y on adsorpt i o n geometry and environment f o r a given molecule/metal system (experimental e v i dence f o r t h i s is discussed in / 6 7 / ) . Within the concept of SERS active sites /64, 87, 239/ i t has been proposed, that the chemical e f f e c t is p a r t i c u l a r l y strong, i f the molecule is adsorbed to sites of atomic scale roughness ( t h i s concept is often also addressed as the "adatom model" / 8 7 / ) . Adsorption induced resonance Raman models, the c o n t r i b u t i o n of local e l e c t r o n i c e x c i t a t i o n s to SERS, and the role of adatoms have independently also been discussed by a Russian group /226,240- 245/. For several other i n t e r p r e t a t i o n s of the enhancement mechanism, which have been developed to explain experimental r e s u l t s from rather special systems, the interested reader is referred to the o r i g i n a l l i t e r a t u r e / 2 4 6 - 2 5 0 / . ic approaches to SERS of / 2 5 1 - 2 5 3 / (ab i n i t i o

F i n a l l y , we mention the microscop-

Hartree-Fock c l u s t e r c a l c u l a t i o n s )

and of /219-221,254/ (coupled molecule-surface plasmon formalism). Various proposed mechanisms and calculated enhancements are summarized and divided into four classes in Table I. "Electromagnetic" (or " c l a s s i c a l " ) mechanisms are usua l l y of long range, i . e . not r e s t r i c t e d to the f i r s t

layer of adsorbed molecules.

Surface corrugation (roughness) is necessary: except in the ATR c o n f i g u r a t i o n . Their magnitude depends on the d i e l e c t r i c properties of the metal, but they should work equally well f o r a l l adsorbates. The l i s t e d " f i e l d " effects are only important f o r small metal-molecule separation ( s a y ~ 1 nm). They do not need surface roughness. 15

.

Orchid" mechanisms

?

roughness mediated excitation and scattering of plasmon surface polaritons, nonperturbive approach GL.GS ~ 102; /205/

roughness mediated, near f i e l d driven Stokes emission ( s t a t i s tical surface roughness) Gs ~ 103; /206/

GL-GS

collective electron resonances, optical conduction resonances ("bumpy" surface)

vibrational modulation of metal surface polarizability via coulomb interaction ("Raman reflectivity") (E ~ 103; /102,223/)

resonance Raman scattering due to chemisorptien induced charge transfer excitation E: 10-102; /236/) E: 10- 103; /237/)

vibrational modulation of plasmon surface polariton resonance (by modulation of tunneling barrier potential) (E ~ 104; grating /39/)

"surface chemistry" effects: formation of complexes, radicals with new electronic properties E: ?

"advanced" image field models ((E: 1-104; /212/))

.

collective electron resonances (island films) GL.Gs ~ 104; /196/ GL ~ 102; /197/

.

vibrational modulation of small particle resonance (by charge injection/withdrawal) ((E ~ 108; /132/))

Chemical. mechanisms

image f i e l d models including renormalization of upper molecular level (E ~ 106; /228/)

.

small particle plasmon resonances (colloids) GL'GS 106; /179/

.

vibrational modulation of charge increase of surface area by gratransfer to the metal p h i t i c carbon overlayers (E: I0-102; /225/) ((E ~ 104- 105; /249/))

.

"simple" image dipole effect ((E ~ 107; 1211/))

Field . mechanisms

extended plasmon surface polariton (ATR, optical grating) GL.GS ~ 104; /36/

.

Enhancement of system polarizab i l i t y by metal-molecule i n t e r action involving charge transfer

.

Enhancement of system polarizab i l i t y by f i e l d mediated metalmolecule interaction

.

Local f i e l d and Stokes emission enhancement by plasmon type resonances

"Electromagnetic . ("Classical") mechanisms

Table i. Proposed enhancement mechanisms, Enhancement f a c t o r s E, GL, GS have been c a l c u l a t e d in t h e q u o t e d articles. Mechanisms p l a c e d on t h e s e p a r a t i o n between two c l a s s e s c o n t a i n e l e m e n t s o f e i t h e r c l a s s ( f u r t h e r e x p l a n a t i o n s in t h e t e x t )

"Chemical" mechanisms require contact of metal and molecule, i . e . chemisorption. They may be especially pronounced at sites of atomic scale roughness, i . e . at socalled "SERS active s i t e s " . Contrary to electromagnetic mechanisms, they are quite individual for every adsorbate/metal system. "Orchid" mechanisms might contribute to the enhancement in special situations (see, e . g . , / 2 4 9 / ) , but are c e r t a i n l y of limited utility

for the general i n t e r p r e t a t i o n of SERS. The enhancement factors

given in Table I have been estimated by the d i f f e r e n t groups for a "typical situation" ( s i l v e r , green l i g h t e x c i t a t i o n ) . They are set into brackets, i f rather unr e a l i s t i c parameters or only a crude theoretical approach have been used. Two brackets are used, i f e i t h e r holds. Generally, numerical estimations tend to s t a r t from highly idealized systems and therefore often y i e l d too large values with l i t t l e

con-

nection to the experimental situation. We note that many mechanisms of Table i may work simultaneously for appropriately prepared surfaces resulting in very large ( t h e o r e t i c a l ) enhancement factors. A c r i t i c a l valuation of various models and numerical estimations may be found elsewhere (e.g. /67/).

2.3

Present State of Understanding

I t is now generally accepted that several processes may contribute to the overall enhancement of the Raman signal from adsorbed molecules. Electromagnetic enhancement always contributes to SERS, i f the surface morphology and d i e l e c t r i c properties of the metal allow the excitation of not too strongly damped surface plasmon resonances. Long range electromagnetic effects play a major role for s u i t a b l y roughened surfaces

/48,98/ as has been soundly demonstrated by spacer experiments /255-258/. There is also clear experimental evidence for an additional short range f i r s t

layer e f f e c t

/48,98,99,136/. This contribution might be p a r t i c u l a r l y pronounced for or even res t r i c t e d to s p e c i a l l y adsorbed molecules, i . e . molecules on certain SERS active sites /67,87,99,259,260/. The nature of these active sites is unclear. Atomic scale roughness might be of importance /67/, at least for certain systems (e.g. for p y r i dine on Ag, see Chapt. 4). Strong SER signals are expected, i f several enhancement mechanisms work simultaneously as for instance for pyridine on coldly evaporated s i l v e r films (see Chapt. 4). Each adsorbate/adsorbent system has to be treated i n d i v i d u a l l y . The share of the various mechanisms contributing to the overall enhancement might be quite d i f f e r e n t for d i f f e r e n t adsorbates on the same surface or f o r the same adsorbate on d i f f e r e n t l y prepared surfaces. The r e l a t i v e weak pyridine signal from s i l v e r gratings /48/ and photochemically roughened s i l v e r (/98/; Fig. 2) is presumably mainly caused by a weak long range electromagnetic effect. Coldly evaporated f i l m s , on the other hand, do not e x h i b i t long range electromagnetic enhancement (Chapt. 4). The pronounced f i r s t

layer e f f e c t of the strong "surface" pyridine signal from these sur17

faces is probably caused by a chemical and a short range electromagnetic effect. For low r e f l e c t i v i t y materials l i k e nickel and palladium, f i n a l l y , any electromagnetic e f f e c t is c e r t a i n l y of l i t t l e

importance.

Electromagnetic mechanisms are in principle understood. They explain, why s u i t a b l y roughened surfaces of metals of high r e f l e c t i v i t y are the best enhancers. Qualitat i v e l y , measured SER e x c i t a t i o n p r o f i l e s (Chapt. 4), mode selective enhancement /261/, and breakdown of selection rules /116-118/ may be understood within a "class i c a l " frame. The q u a n t i t a t i v e description of real, SERS active systems is in many cases, however, s t i l l

marginal because of the crudeness of the models and the l i m i t e d

information on the surface morphology from the experiment. We emphasize that many, but not a l l , aspects of SERS can be understood on a purely electromagnetic basis (/67/, Chapt. 4). As so f a r appreciably enhanced Raman signals from LEED-clean, single c r y s t a l l i n e , smooth surfaces have not been observed /25,28/, a major c o n t r i bution of the f i e l d effects l i s t e d in Table 1 to SERS is doubtful. The chemical effect in SERS i s , however, well established (see, e.g., /67/ and Chapt. 4). As i t depends on the details of the metal-molecule i n t e r a c t i o n , i t may be sensitive to the adsorbate and the adsorption s i t e . Molecules bonded to certain defect sites are often subject to a p a r t i c u l a r l y strong chemical enhancement (e.g. pyridine on Ag; /67/ and Chapt. 4). The d e t a i l s of the chemical mechanism are s t i l l

a matter of

debate. Currently, photon driven charge transfer excitations /262/ at sites of atomic scale roughness /263/ are thought to play a major role /67,264,265/. Q u a l i t a t i v e l y , chemical effects can account for many experimental observations (breakdown of selection rules, mode and species selective enhancement, etc. /67/; as in real systems usually chemical and electromagnetic effects contribute to SERS, i t i s , however, very difficult

to disentangle the r e s p o n s i b i l i t i e s of e i t h e r mechanism). Quantitative

theoretical evaluations are extremely complicated and represent presently hardly more than crude order of magnitude estimations. To understand the d e t a i l s of any chemical mechanism in SERS means to understand chemisorption, which s t i l l

requires very much

experimental and theoretical work. Nevertheless, i t seems worthwile to use SERS as surface analytical tool. As long as the d e t a i l s of the enhancement mechanism are s t i l l however, to be taken when interpreting SER spectra.

18

unknown, extreme care has,

3. E x p e r i m e n t a l

3. I

Arrangements

Standard optical and vacuum equipment can be used f o r SER studies. A typical experimental set-up is sketched in Fig. 4. Radiation from an Ar- or Kr-ion laser is cleaned from plasma l i n e s by means of a laser f i l t e r

monochromator,

polarized par-

a l l e l to the plane of incidence by a p o l a r i z a t i o n r o t a t o r , and focused on the sample by a c y l i n d r i c a l lens to a l i n e focus of t y p i c a l l y 0.1.3 mm2. The angle of i n cidence is set to maximize e x c i t a t i o n e f f i c i e n c i e s (~ 75o to the normal / 2 4 / ; often also conventional backscattering geometry, i . e . perpendicularly i n c i d e n t l i g h t ,

is

used). The power i n c i d e n t on the sample is t y p i c a l l y 100 mW. The scattered l i g h t

high power

LEED

lens %o UHVchamber spectrometer and II ~~ S ~ -L ~\(o 0.5 L, 140 K), pyridine molecules are forced into an upright orientation. This high coverage compressional phase is even more weakly bound to the metal via the nitrogen lone-pair o r b i t a l and should therefore desorb below 210 K. In another paper /355/, desorption of condensed multiple pyridine layers from Ag(llO) is observed at ~ 190 K, whereas a chemisorbed nitrogen-bonded species is s t i l l

present at the surface in sub-monolayer amounts at 275 K. S i m i l a r l y ,

pyridine is adsorbed with i t s aromatic ring perpendicular to the surface on Cu(110) /356/ and Ir(111) /357/ at room temperature. L i t t l e is known on the bonding of pyridine to i r r e g u l a r metal surfaces such as coldly evaporated f i l m s , i . e . of bonding to defect sites. For m e t a l l i c catalysts, pyridine is a r e l a t i v e l y toxic substance /316/. Due to the nitrogen lone-electron pair i t seems to block active center~ by forming a r e l a t i v e l y stable bond to these sites /314,315/.

4. I

Coldly

Evaporated

Silver

Films

L~.I.1 General Spectral Features Figure 7 displays surface enhanced Raman spectra from coldly evaporated s i l v e r films exposed to ~ 0.2 L of pyridine and deuterated pyridine at 120 K /358/ (an exposure of 0.2 L corresponds to ~ 0 . i monolayer coverage /267/). An enhancement factor of several 104 is estimated by comparing SER l i n e i n t e n s i t i e s to corresponding values in ordinary spectra from a thick pyridine layer condensed on a SERS inactive s i l v e r surface at 120 K (Fig. 7c and /99/; note, that this estimation neglects the f l a t metal surface contribution as well as any electromagnetic e f f e c t of the inactive surface due to residual roughness). Different vibrations experience evidently d i f ferent enhancements. This mode specific behaviour is p a r t i c u l a r l y clear for the C-H 27

%2

o' 9

(a)

i~o-b-6--~

i

:;~o-~-~

I

,~'oo

*

~o-- 1~oo ~ - ' ~ - - - ~,oo '

' ~

' '

X3.3

-9 ~E 9

:32,oo

C

~.

(b)

x 3

o6 "~3

,

.

~,soo

.,I

~6po

,

,2,o0

,

.~

,

'

~

~.oo

(c)

2800

3200

1600

1200

800

~00

0

1600

1200

800

~00

0

3

o

2200

'

2OO 'O

'

Roman shlfl (cm -1]

Fig. 7. Raman spectra from vapour deposited s i l v e r films. (a): coldly evaporated f i l m exposed to 0.2 L of pyridine; (b): l i k e (a), but exposed to 200 L; (c): f i l m condensed at room temperature and exposed to 200 L of pyridine at 120 K; (d): l i k e (a), but exposed to deuterated ~yridine. All spectra have been taken with 200 mW of 514.5 nm radiation and 4 cm-~ bandpass. After /358/

stretching vibrations around 3000 cm-1, which are only weakly pronounced in SER spectra from coldly evaporated s i l v e r films /108/ (adsorbed hydrocarbons behave s i m i l a r l y , see Chapt. 5). Upon deposition of further pyridine layers, SER l i n e int e n s i t i e s decrease and several new peaks appear (/99/ and Fig. 7b; ~ 102 layers of pyridine). Spectral positions of the new features are v i r t u a l l y identical to those of corresponding vibrations from a thick layer on a SERS inactive s i l v e r surface (Fig. 7c and Fig. 8). They are assigned to scattering from "bulk" pyridine, i . e . pyridine in the second and consecutive layers as well as pyridine physisorbed to Ag /67,119,267/. On the other hand, low coverage SER lines are due to "surface" pyridine, which are molecules c~misorbed to s i l v e r , probably on certain "active" sites /67,119,267/. Bulk and surface signals overlap for most vibrations, for instance for ~6 at 1034 cm- I (Fig. 8). A few modes allow, however, a separation of either contribution because of stronger chemisorption induced s h i f t s of the vibrational energy. For example, bulk signals are observed at 993 cm-1 (wI , symmetric ring breathing) or 607 cm- I (~3' planar ring deformation), whereas corresponding surface signals appear at 1003 cm-1 and 621 cm- I respectively (Fig. 8b; note, that the l i n e frequencies of surface pyridine s h i f t somewhat with coverage, see also Sect. 4.1.2). A comparison of bulk and surface pyridine i n t e n s i t i e s reveals the pronounced "first

layer effect" in SERS from coldly evaporated s i l v e r films /99,123,267/: only

surface pyridine is subject to the full enhancement of ~ 10 4.

28

I

,i

I

I

I I r

30611~ L

I

I

I

29~I~I 29~1

I

~

306g

I

/ ~'. q , ~

~G

I

302Z,

2952

u~

%

~0

I

(c)

E

-~,,a

31~.-v

3200

I

I

I

3061

I

30~

3077i 3086 II 30&O r3025 309511~J ~.4 2992 ' ,,,

I

I

3000

3100

1-

2900

lu4u

iuuu

~m)

650

5DO

Raman shift (cm-1) Fig. 8. Details of Raman spectra from vapour deposited s i l v e r f i l m s . (a) c o l d l y evaporated f i l m exposed to 0.2 L of p y r i d i n e ; (b): ~ike (a), but exposed to 200 L; (c): room temperature deposited f i l m exposed to 2.10 L of pyridine. Same experimental conditions as for Fig. 7 (except for (c): 2 cm-1 bandpass). A f t e r /358/

The spectra displayed so f a r (Figs. 7 and 8) have been recorded by using 514.5 nm e x c i t a t i o n . Other e x c i t a t i o n frequencies lead to s i m i l a r SER spectra from surface p y r i d i n e (Fig. 9). Note, however, that the r e l a t i v e l i n e i n t e n s i t i e s depend on e x c i t a t i o n wavelength. Modes of large v i b r a t i o n a l energy (e.g. ~4 and v5) become more prominent when changing the e x c i t a t i o n from red to blue /123/. This points to mode

~ 676.Z,nm .

.

.

- 159, ~c

o x

.

.

1006

J

12I 5 1035 ,35,~

,,;.,om',,,

.

/

62~

.

168

, , J

l'~176176 '

"

'

/

.w/

'/

:,- 1C

| C -_=

,

Z,

57.9nm 12;,

' ~5'oo

'

1oos

lo;o

/ '

Raman shift (cm "1)

s~O ........ "

/

Fig. 9. SER spectra of c o l d l y evaporated s i l v e r f i l m s exposed to 0.2 L of pyridine f o r d i f f e r e n t exc i t a t i o n wavelengths as indicated. Incident power was 60 mW (676.4 rim), I00 mW (514.5 nm), and 40 mW (457.9 ?m); bandpass was set to 4.5 cm-~ for a l l spectra. A f t e r /358/

29

specific e x c i t a t i o n p r o f i l e s for surface enhanced Raman scattering (see Sect. 4.1.4). A close look at the SER spectra allows the following statements: i)

Almost a l l pyridine skeletal fundamentals are observed and can be assigned /99,

108,119/. There i s , however, no indication of a metal-pyridine stretching v i b r a t i o n

/108/. ii)

Several low energy features at 73 cm-1, 112 cm-1, and 161 cm-1 (Fig. I0) re-

f l e c t an i n t r i n s i c property of coldly evaporated s i l v e r films. They are present without adsorbed molecules, do not change upon pyridine exposure, and are consider-

ably weakened upon annealing the sample to room temperature. The structures have been a t t r i b u t e d to disorder induced Raman scattering (DIRS) from bulk acoustical phonons within the penetration depth of the l i g h t (/100,108/ and Sect. 3.2). iii)

Combination bands and overtones are weakly pronounced /358/. The f i r s t over-

tone of wI is detected with % 1% i n t e n s i t y of the fundamental (Fig. I I ) . iv)

The features at 690 cm-1 and 1050 cm- I in the spectra of Fig. 7 (marked by an -1

arrow) and the broad peak at 2108 cm

.

in Fig. I I are due to adsorbed impurities

(see Chapts. 5 - 7 ) . v)

The l i n e at 1026 cm- I , marked by a star in Fig. 7a, is only observed a f t e r py-

r i d i n e exposure, but not always (see Fig. 8a). I t s i n t e n s i t y seems to be connected with the strength of impurity lines. The breathing vibration (Wl) of pyridine bonded

|0

[

I

i

8 E

2 (a) >- 2.~.!b)

f

31r 1 "6 21-

2oos

.-I

X

~1

-

0 i

i 2000 Roman shift ( c m "1)

O

400

I 2100

,

Fi 9. 11. Overtone of breathing mode ~I in SER spectrum from coldl~ evaporated s i l ver f i l m exposed to 0.2 L of pyridine (peak i n t e n s i t y of fundamental was 2700 c t s / s ) . 200 mW of 514.5 nm radiation, 4 cm- I bandpass. After /358/

~

L........ -r "~, I 300 200 lOO o Romon shift (crn-I )

Fig. 10. Low energy features of Raman spectra. (a): coldly evaporated f i l m , unex~ ( b ) : l i k e (a), exposed to 0.2 L of pyridine; (c): l i k e (b), a f t e r annealing to room temperature; (d): l i k e (a), but warmed to 220 K and recooled to 120 K to . increase the i n t e n s i t y of the Raman features (170 mW of 514.5 nm radiation, 0.6 cm- I bandpass). (a), (b), and (c) have been recorded with 200 mW of 514.5 nm radiation and 2 cm-1 bandpass. After /100/ and /108/ 30

Table 2. V i b r a t i o n a l energies o f p y r i d i n e in v a r i o u s systems [ i n cm-1; number in parentheses a f t e r each v i b r a t i o n gives the i n t e n s i t y r e l a t i v e to b r e a t h i n g mode (Vl) i n t e n s i t y which is set to I00 ( f o r (g) to 1 0 ) ] . (a) Neat l i q u i d p y r i d i n e , a f t e r /331,332/; (b) 1 M aqueous s o l u t i o n o f p y r i d i n e , a f t e r / 5 5 , 1 2 3 / ; (c) complexed p y r i d i n e AgCIO4.2Py, IR study, a f t e r / 3 4 1 / ; (d) t h i c k l a y e r on SERS i n a c t i v e s i l v e r surface in UHV, a f t e r / 3 5 8 / ; (e) p y r i d i n e on SERS a c t i v e Ag f i l m in UHV (0.2 L exposure), a f t e r / 1 0 8 , 1 2 3 , 3 5 8 / ; ( f ) l i k e (e), but exposed to 200 L, a f t e r / 3 5 8 / ; (h) p y r i d i n e - d 5 on SERS a c t i v e Ag (g) neat l i q u i d p y r i d i n e - d 5, a f t e r / 3 3 0 , 3 3 1 / ; f i l m in UHV (0.2 L exposure), a f t e r / 3 5 8 / Mode

(a) Liquid

(b) Aqueous Solution

(c) Complex

Solid

Symmetry

C5H5N

C5H5N

AgCIO4.2Py

C5H5N

v21,A2

374

(d)

(0)

378

Ce) SERS on O.1ML C5H5N

(f) Ag IOOML C5H5N

(g) Liquid C5D5N

(h) SERSon Ag O.1ML CsD5N

380

(2) 382

(30)

329 (1)

326

(I)

(1) 413

(3) 412

(32)

371 (1)

373

(2)

(3) 623 (20) 621 607

(68) (16)

582 (3) 601 (18)

{2) 654

(34)

625 (6) 626

(2)

(i)

v27,B2

405

(1)

409

(4)

412

(s)

~3' A1

605

(3)

618 (15)

641

(m) 607

v12,B1

652

(6)

654 (30)

651

(w) 652 (10) I 652

~26,B2

700

(0)

697/700

(s)

706

(0) 696

(3)

707

(28)

530 (i)

530

(4)

~23,B2

749

(0)~ 756

(3)

749/754

(s)

755

(0) 749

(4)

753

(28)

567

557

(0)

v25,82

886

(1) 890

(3)

889

(i)

(1) 887

(11)

762 (4)

772

(8)

v20,A2

(886)

690 (5)

696

(0)

v24,82

942

(0)

950

(4)

944

v22,A2

981

(4)

980

(4)

990

798 (0) 1 812

(i)

V l ' AI

992 (i00

410

(w) 897 (w) 958 (vw

986

880

(0)

942

(4) 944

(2)

972

(2) (965

{12) (4))

(I00) 1004 (i00) 1005/1012 (m) 996 (100) 1006 (I00) 1003 993 (115) (46) 1037 (20) 1033 (170)

823 (5)

962 (i0)

975 (100)

1006 (7) 1007 (10)

v6 ' AI

1030 (74) 1037 (87)

1037

(s)!1037

~8 ' A1

1068

(I) i071 (22)

1068

(s) 1059

~17,81

1085

(0)

Vl6,B I

1148

(i) 1154 (17)

1156

v 5 , AI

1218

(6) 1221 (41)

1218 1224

(vw) 1227 (10) 1215 (59) 1216 (116) (m)

886 (5) 889 (14)

Vll,Bl

(1218)

1233

(vw)i1217

(2)

908 (5)

v15,B1

1375

(0) 1362

(i)

1361

(vw)i1360

(0) 1355

(i) 1356

(8)

vls,Bl

1439

(0) 1447

(4) 1440/1449 (s)

1442

(1) I1444

~9' AI

1482

(2) 1491 (19)

1482

(m) 1488

(1) 1480

(2) 1481

~14,BI

1572

(4) 1579 (33)

1573

(w) 1576 (I0) 1572

v4 ' AI

1583

(6) 1597 (38) 1597/1607(m,s) 1586

~Io,AI

3036

(2)

Vl3,Bl

(3036)

v2 ' AI

3054 (26) 3076

v7 ' AI

(3054)

v19,B I

3083

(2)

1067 (w) 1150

(1) 1069 (i) (3) 1150

I

(7)

1070

(28)

(3) 1149

(39)

3037

(1) 3033

3025

(2)

3066

(9)

3077

(1)

823

829

(2)

833 (5)

839

(2)

(887)

909

(6)

1324

(1)

(10)

1301 (0) 1306

(i)

(16)

1340 (1) 1341

(9)

(4) 11573 (103) I1542 (4) 1575

(7)

1591 (9) 1593 (59) 1583

3061 (17) 3061

I?)

(6) 1057

1322

(44) 11530 (6) 1555 (51) (74)

(4) 3038 (- 20) 2254 (6) 2250 3024 (1) 3059

{35)

2285

(2)

2285

(0)

(40) 2293 (i0) 2290

(2)

2270 (2) 2266

(2)

3089 (~ 3) ]2293) 31

to impurity sites is presumably responsible for this peak (see also Sect. 4.1.3). The l i n e at 1026 cm- I is quite strong in SER spectra from pyridine on activated s i l v e r electrodes for potentials positive (~ I V) to the point of zero charge /57, 247/. Here i t has t e n t a t i v e l y been assigned to Lewis-coordinated pyridine /57/ or adsorbed pyridinium cations PyH+ /246/. Table 2 summarizes the pyridine SER l i n e energies and i n t e n s i t i e s . The mode sel e c t i v e enhancement is c l e a r l y seen when comparing A1 mode i n t e n s i t i e s of surface pyridine (column e) and neat pyridine (column a). Vibrational energies of surface pyridine are generally shifted to higher values with respect to neat pyridine. The s h i f t is most pronounced for some planar ring modes (~3,~i,~4), but does not exceed 20 cm-1. This points to weak perturbation of the adsorbed molecule, i . e . weak chemisorption /20/. Similar l i n e s h i f t s are observed when pyridine is coordinated to Ag in metal-pyridine complexes (/341/ and column c in Table 2), or bonded to Agx- or CUx-Clusters ( x ~ 3 ) in an argon matrix /359/. They have been explained in terms of coupling with low frequency vibrations~ p a r t i c u l a r l y with the metal-pyridine stretching /339,347/, or, a l t e r n a t i v e l y , with changes of the electron d i s t r i b u t i o n in the molecule resulting in stronger chemical bonds in the ring system /334/. For our purposes, the correct interpretation of the s h i f t s of complexed pyridine is less important than the fact that they e x i s t . The s i m i l a r i t i e s of vibrational features of surface pyridine and metal-pyridine complexes may allow to speculate on the adsorption geometry of the former. Bonding to surface sites of "certain a c i d i t y " via the nitrogen lone-pair o r b i t a l should be involved, an orientation as proposed in /20/ (low coverage phase) seems reasonable. The question of orientation and bonding of

, r

o

x

9

5.8

ix,o

4,1 1.4

c

1600

1590

15801230 1220

1210

1200

1010

1000

990

I ~

625

620

610

600

Romon shift ( cm-1) Fig. 12. Line shape of some SER lines from coldly evaporated s i l v e r films exposed to 0.2 L of pyridine. Additional peaks in (c) are ordinary lines from room temperature deposited films exposed at 120 K to 103 L (3.3times enlarged) and 105 L (lOtimes enlarged, 0.5 cm- I bandpass). Spectra have been taken with 200 mW of 514.5 nm radiation and 4 cm- I bandpass. Indicated halfwidth (FWHM) has been corrected for spectremeter response. After /358/ 32

surface pyridine and hence of the character of the SERS active species is addressed in somewhat more detail below. SER lines of surface pyridine display quite d i s t i n c t shapes /358/ (Fig. 12): ~3 and w5 are almost symmetrical, whereas Wl and ~4 show a pronounced asymmetry (note, that the breathing mode Wl in ordinary spectra is symmetrical, as expected; due to convolution by the spectrometer function measured peaks resemble Gaussian rather than Lorentzian p r o f i l e , Fig. 12c). SER lines from adsorbates on coldly evaporated films are frequently asymmetrical. In general, a slow increase on the low energy side of the l i n e is accompanied by a steeper decrease on the high energy side. An interpretation of the l i n e shape w i l l be given in Chapt. 5, where ethylene adsorption is discussed. 4.1.2

Coverage Dependence

Pyridine exposures as low as 10-2 L corresponding to roughly one per cent of a monolayer coverage r e s u l t in e a s i l y detectable Raman signals from coldly evaporated Ag films /99/. Line i n t e n s i t y , shape, and spectral position vary with coverage, and new lines develop. The l a t t e r has b r i e f l y been touched in the preceeding section (see Fig. 7). The change of spectral features with exposure is shown in detail for the breathing mode region in Fig. 13. The bulk pyridine ~I l i n e emerges from the slope of the surface pyridine l i n e at ~ 2 L exposure and is detected as a d i s t i n c t peak f o r ~ 5 L. I t is much stronger than expected for ordinary Raman scattering, much weaker, however, than the surface pyridine signal for 0.2 L exposure. I t s i n t e n s i t y does not measurably increase f o r exposures between ~ 5 L and ~ 30 L (in f a c t , i t decreases s l i g h t l y ) . I t starts to grow for exposures ~ 30 L caused by ordinary scattering, analogous to pyridine condensed on inactive s i l v e r surfaces /25/. Hence bulk -I pyridine signals at 993 cm are weakly enhanced, and the e f f e c t is restricted to molecules in the immediate vicinity of the s i l v e r surface. The development with exposure of the breathing mode i n t e n s i t i e s depends on the excitation wavelength (Fig. 14). Even for 2.104 L exposure, bulk signals do not exceed surface pyridine signals for red e x c i t a t i o n , and the i n t e n s i t y of ~6 (overlapping bulk and surface signal) is always smaller than that of Vl" Blue and green e x c i t a t i o n leads to features s i m i l a r to ordinary spectra from thick layers on SERS inactive surfaces /25/. Here w6 is~ the strongest mode for intermediate exposures (Fig. 14, 200 L). Recalling the ~

dependence of ordinary scattering and the SERS

excitation p r o f i l e of the breathing modes with i t s peak in the red (see Sect. 4.1.4), the i n t e r p r e t a t i o n is straightforward: r e l a t i v e l y small ordinary and large SER signals combine for 676.4 nm e x c i t a t i o n , whereas the opposite is the case for blue (green) e x c i t a t i o n . The SER l i n e of the breathing mode from surface pyridine broadens and s h i f t s to s l i g h t l y smaller energy with coverage (Fig. 13). Other vibrational modes behave 33

676./- nm I

2

_=

I

1

51~.5nm

/,57.gnm

i

0

u

0

x

0 ,F ~ 1010

7000

990

Romanshift ( cm-1}

Fig. 13

1040

1000

1040 1000 Reman shift(cm-I}

Fiq. 14

Fig. 13. SER spectra of symmetric breathing vibration from coldly evaporated Ag films exposed to various amounts of pyridine as indicated. 200 mW of 514.5 nm radiat i o n , 2 cm- I bandpass. The bare to the l e f t of each spectrum represents 100 cts/s. After /358/ Fig. 14. Development of breathing modes with exposure f o r pyridine on coldly evap~ g f i l m . Spectra have been taken with 65 mW (676.4 nm), 200 mW (514.5 nm), and 75 mW (457.9 nm). Bandpass was 4 cm- I for a l l spectra. Circles: surface pyridine, dots: bulk pyridine. After /358/

s i m i l a r l y (Fig. 15). The variations of l i n e width and spectral position are most prominent f o r exposures, which correspond to roughly monolayer completion. Besides these changes, one observes a transformation from the c h a r a c t e r i s t i c , asymmetrical SER l i n e shape of ~I (0.2 L) into a more symmetrical l i n e with exposure (200L; Fig. 13). A detailed i n t e r p r e t a t i o n of these observations is d i f f i c u l t ,

since l i t t l e

is

known of pyridine adsorption on the i r r e g u l a r surface of SERS active coldly evaporated s i l v e r films. As the variations are most pronounced when completing a monolayer, interaction with adjacent adsorbed molecules seems to be involved. As mentioned, the discussion of the variation of the l i n e shape is postponed to the ethylene/Ag system (Chapt. 5). A l i n e a r increase of the SER i n t e n s i t y is observed for very small exposures 0 . 1 L (Fig. 16). When this increase is extrapolated to greater coverage and when 34

~

1

r-'il

,

{{ {{

v3

el

I

I

I

{{ '5

li

1590~

I |

{

I

I

t

{,}

~,{

I

I

I

I

I

I

I

t

~1

,

f

,

c

I

I

I I

,

l

!

I

I

I ]

10-1 100 101 exposure(L)

10-2

I

{{{ I.

}

I

C

I

I

{

i

, I

1s98

iI

i

102

10-2

103

10-1 100 101 exposure (L}

102 103

Fig. 15

lo-3

10-2 '

coverage ( layers 100 101 102

lo-1

[

i

I

i

I

j"

IOt

'

|

103

J

'

I

'

10/. I

f"

/Y'/

103 u

m

g

E 101 i0 i

,

I

10-2.

,

I

10-1

,

I

,

I

100 101 exposure {L}

,

I

102

,

]

103

J

I

10L

Fig. 16 Fig. 15. Shift of spectral position ( l e f t ) and broadeninq of line width (right) as ~ o n of exposure for three SER lines from surface pyridine on coldly evaporated silver films. Arrow marks exposure equivalent to monolayer formation. After

/358/

Fig. 16. Peak intensity of some pyridine Raman lines Fiile-dC]n symbols: pyridine on coldly evaporated film squares: ~6 (all from surface pyridine); triangles: vl bols: Vl of pyridine on SERS inactive Ag surface /25/ triangles: polycrystalline slug). Lines are guides to of Echem, Ebulk, and Esurf see text

as a function of exposure. /99/ [dots: ~i; rhombs: ~3; (bulk pyridine)]. Open sym(rhombs: (110) single crystal; the eye. For an explanation

35

the data for ordinary Raman scattering from SERS inactive surfaces are extended to smaller coverage, the d i f f e r e n t slopes of the two lines point to a roughly 30 per cent smaller sticking c o e f f i c i e n t of m u l t i l a y e r pyridine compared to surface pyridine. This agrees with results of other investigations (/98,101/; note, that the upper scale in Fig. 16 neglects this difference). Taking the difference into account a total enhancement Esurf of ~ 104 f o r the symmetric breathing vibration of surface pyridine is estimated, which is s l i g h t l y smaller than the less accurate value given above. Saturation of the SER signal from surface pyridine is observed for ~ 0.3 L exposure. Upon further exposure, the signal decreases by up to roughly a factor of 18 for ~ 30 L, before the i n t e n s i t y starts to increase again. Other surface pyridine l i n e s , f o r example ~3' e x h i b i t a s i m i l a r coverage dependence (Fig. 16). Note, however, that the i n t e n s i t y decrease for ~3 a f t e r saturation is smaller than for ~I" This is due to the fact that peak i n t e n s i t i e s rather than integrated intens i t i e s are plotted in Fig. 16. The l i n e shape of w3 broadens less than that of w1 with exposure (see Fig. 15). Taking integrated values, either l i n e i n t e n s i t y drops by roughly an order of magnitude a f t e r saturation. A s l i g h t l y d i f f e r e n t exposure dependence is observed for overlapping bulk and surface pyridine signals. The a n t i symmetric breathing mode ~6 shows maximum i n t e n s i t y at a larger dose ( 0 . 6 L ) ,

and

subsequently decreases by only a factor of ~ 4 (Fig. 16). Both effects are caused by the bulk pyridine contribution to the overall i n t e n s i t y of w6, which does not vary appreciably between ~ 2 L and = i0 L as outlined above for wI (see also Fig. 16; the observed s l i g h t i n t e n s i t y decrease is interpreted in Sect. 4.1.4). The short range enhancement Ebulk of the bulk pyridine signal is estimated to > 30 for the symmetric breathing vibration (Fig. 16). The value represents a lower l i m i t , since the bulk wI i n t e n s i t y used for the estimation is c e r t a i n l y from less than a monolayer of adsorbed molecules. Assuming that this mechanism also amplifies the surface pyridine signal, an additional e f f e c t must be responsible for the ~ 300 times stronger enhancement for t h i s species. For reasons, which w i l l become clear l a t e r , this factor is called Echem in Fig. 16. F i n a l l y we note, that the l i n e a r exposure dependence of the ~I i n t e n s i t y from pyridine on Ag(llO) or on SERS inactive p o l y c r y s t a l l i n e f o i l s down to ~ 3 layers coverage (Fig. 16; /25/) excludes short range, "smooth surface" enhancements of ~ 3. Very recent measurements showed the l i n e a r development of the i n t e n s i t y also in the sub-monolayer region /28/, which leaves no space for any "smooth surface" enhancement of ~ i nm range, i . e . for any measurable image f i e l d effect.

36

4.1,3

Annealing Behaviour

Annealing to room temperature i r r e v e r s i b l y destroys the enhancement properties of coldly evaporated s i l v e r films /99,100,134/. The temperature v a r i a t i o n of the background i n t e n s i t y , of the Rayleigh scattered l i g h t , and of the symmetric breathing mode of pyridine has been discussed in detail in /100,239/. In these experiments the sample was warmed to room temperature with = I K/min. Figures 17 and 18 summarize the annealing behaviour of Raman signals from adsorbed pyridine. The peak int e n s i t y ~i of surface pyridine (0.2 L exposure) f i r s t

increases with temperature,

exhibits a maximum at ~ 210 K, and then decreases. The l i n e disappears at ~ 270 K. I t cannot be restored by recooling to 120 K and re-exposing to pyridine /100/. The

105

I

i

I

i

I

105

l

I

r

I

"1 -~.~ 1~

T~" ~

~

V"

-IJ

i ~,~\

~ 103

I021F-

A-~A

l

.~" lo3.~..x g

A

101I~-~/i L- I i 120 150 180 210 2z.o2?0 temperoture (K) _F~g. i7

i

\4

'i

~ 1026cm- 1

1i ]011

I

I

I

I

120 150 180 210 240 270 temperclture{K}

Fig. 18

Annealing of Raman peak i n t e n s i t i e s from breathina mode v I of pyridine i l l e d - i n symbols: coldly evaporated f i l m (dots: O.2-L, surface pyridine; triangles: 200 L, surface pyridine; rhombs: 200 L, bulk pyridine). Open symbols: SERS inactive, room temperature deposited f i l m exposed to 200 L (rhombs: Vl; t r i angles: v6)" Lower l e f t inset shows temperature v a r i a t i o n of the Rayleig~ scattered l i g h t from the SERS inactive sample. 200 mW of 514.5 nm radiation, 4 cm-• bandpass, and ~ I K/min temperature variation. Lines are guides to the eye. After /358/ Temperature v a r i a t i o n of various SER peak i n t e n s i t i e s from coldly evaporated Ag films exposed to 0.2 L of pyridine ( f i l l e d - i n symbols; lines are guides to the eye). Open rhombs: measured ~nnealing of "impurity" l i n e at 1050 cm- I , 200 mW of 514.5 nm radiation, 4 cm- bandpass, and = 1K/min temperature v a r i a t i o n . After /358/ 37

coldly evaporated s i l v e r f i l m has l o s t i t s SERS a c t i v i t y i r r e v e r s i b l y . As shown in / I 0 0 / and in agreement with TDS studies /355/, the decrease of i n t e n s i t y for T 210 K is not due to desorption of pyridine. When the SERS active s i l v e r f i l m is coated with a thick pyridine layer, neither bulk nor surface pyridine Raman signals of ~1 change between 120 K and 175 K (Fig. 17). The solid pyridine overlayer apparently prevents any annealing. M u l t i l a y e r pyridine, i . e . layers beyond the f i r s t ,

desorb at 175 K. This follows from the dis-

appearance of the bulk pyridine signal from i n a c t i v e Ag surfaces, the behaviour of the Rayleigh scattered i n t e n s i t y (lower l e f t inset in Fig. 17), and the pressure increase in the vacuum chamber at t h i s temperature. Above 175 K, surface pyridine signals from the t h i c k l y coated active sample grow much f a s t e r than those from the sample exposed to only 0.2 L. They eventually approach the l a t t e r at ~ 240 K. The bulk pyridine i n t e n s i t y from physisorbed molecules in d i r e c t contact with s i l v e r increases also for T ~ 180 K, but peaks already at ~ 205 K, and is f i n a l l y l o s t at 220 K. This is presumably due to desorption of the species responsible for the bulk pyridine l i n e for T ~ 180 K, Desorption of a weakly bonded species has been observed in this temperature range /20,97/. I n t e n s i t i e s of various surface pyridine lines anneal s i m i l a r l y as shown for wI and the ring deformation modes ~3 and w4 in Fig. 18 (0.2 L exposure). Note, however, that the increase between 120 K and 210 K is weaker for ~4 (factor of 3) than for w3 and wI (factor of 5 - 6 ) .

Only the l i n e at 1026 cm-1 behaves d i f f e r e n t l y . I t s

i n t e n s i t y drops immediately upon warming from 120 K and i t disappears at ~ 180 K. Some impurity lines display a quite s i m i l a r i n t e n s i t y v a r i a t i o n (e.g. the l i n e at 1050 cm-1 in Fig. 7; open rhombs in Fig. 18). This corroborates the t e n t a t i v e assignment of the l i n e at 1026 cm- I to ~i of pyridine bonded to impurity sites (Sect. 4.1.i). Spectral position and halfwidth of SER lines may also vary upon annealing. Variations are i n s i g n i f i c a n t f o r samples exposed to 0.2 L of pyridine. A l i n e width independent of temperature and a very small l i n e s h i f t to larger vibrational energy have been observed (Fig. 19). This j u s t i f i e s the use of peak i n t e n s i t i e s in Figs. 17 and 18. I t does, however, not hold for coldly evaporated s i l v e r films exposed to 200 L of pyridine. After desorption of m u l t i l a y e r pyridine at 175 K, a considerable decrease of the l i n e width and appreciable blue s h i f t of the breathing vibration of surface pyridine is observed (Fig. 19). Bands of other modes change s i m i l a r l y . I t is remarkable, that l i n e parameters vary in the same way with coverage, where high coverage data correspond to those at ~ 180 K and low coverage data to those at ~ 250 K (compare Figs. 19 and 15). This suggests a common explanation which w i l l be discussed l a t e r . In closing we note, that the d i f f e r e n t l i n e widths of the breathing vibration of surface pyridine from thick and thin overlayers can explain only part of the corresponding i n t e n s i t y difference at 120 K (Fig. 17).

38

~~-~i--

IO~

'~

IlXt

}}}}}

"~ 100~

}}

12

o 1002

A

i0:E

VI

Q.

120

150

,1

6

=o 4= I

I

180 210 240 temperoture (K)

270

Fi 9. 19. Variation of spectral position and l i n e width of ~1 from surface p y r i d ine with temperature. Solid l i n e s : from sample exposed to 0.2 L (measured); dots: from sample exposed to 200 L. Arrow i n dicates desorption temperature of m u l t i layer pyridine. 200 mW of 514.5 nm rad i a t i o n , 4 cm-1 bandpass, and ~ I K/min temperature v a r i a t i o n . A f t e r /358/

The observed effects may r e f l e c t a complicated simultaneous acting of several processes. Surface and bulk defects as well as small scale surface roughness in h i g h l y disordered c o l d l y evaporated s i l v e r f i l m s anneal with temperature ( i n general not simultaneously; Sect. 3.2). This affects SER i n t e n s i t i e s via the density of possible SERS active sites and the q u a l i t y of electromagnetic resonances. In addit i o n , spectral p o s i t i o n and strength of the l a t t e r depend on overlayer thickness. F i n a l l y , geometry and density of adsorbed molecules may vary with temperature by, for instance, desorption of weakly bonded species, i . e . of physisorbed molecules. This may a f f e c t v i b r a t i o n a l i n t e r a c t i o n of adsorbed molecules as well as the dens i t y of the "relevant" species, i . e . of surface pyridine. The c o n t r i b u t i o n s of the various processes to the i n t e n s i t y v a r i a t i o n of SER signals with temperature are discussed in Sects. 4.1.4 and 4.4.

4.1,4

Excitation Spectra

Raman e x c i t a t i o n spectra are p a r t i c u l a r l y useful to elaborate various contributions to SERS. They may provide information on the intermediate e l e c t r o n i c states of the scattering process as well as on the surface topography of the metal via the local f i e l d strength which a f f e c t s the Raman scattered i n t e n s i t y . A v a r i a t i o n of approp r i a t e experimental parameters may allow to discriminate c o n t r i b u t i o n s of d i f f e r e n t processes to the e x c i t a t i o n p r o f i l e . Hence the share of d i f f e r e n t enhancement mechanisms might be estimated from such i n v e s t i g a t i o n s . The procedure employed to obtain e x c i t a t i o n p r o f i l e s has been described in d e t a i l in /119/. Thick pyridine layers condensed on SERS i n a c t i v e s i l v e r surfaces served as standard. An exposure of 1.9 L of p y r i d i n e was assumed to form a monolayer of 0.5 nm thickness ( f o r d e t a i l s see /267/).

39

Raman excitation spectra from coldly evaporated s i l v e r films exposed to pyridine e x h i b i t resonance-like p r o f i l e s (Fig. 20, /119/). For the breathing v i b r a t i o n wI , the broad resonance peaks at ~ 2.15 eV (FWHM: ~0.5 eV). An i n t e n s i t y r a t i o on and o f f resonance of ~ i00 is estimated from the shape of the p r o f i l e . Similar resonances are observed for other pyridine lines as well as for surface enhanced Raman lines of other adsorbates (Fig. 20; /119,281,360/). Independent of the adsorbed species, the observed maxima s h i f t to shorter wavelength with increasing vibrational energy. This is summarized in Fig. 21 for d i f f e r e n t lines of various adsorbed mole-

wave(ength of incident radiation (nm) 700 600 500

2,6i

I

I

I

500

~2.A E

E

2 2~

/

\

il/ ~/,~,,j!/, j )~..,,,"'/

I

'

7

~,

",A .\

' # " - ' 2.8

' A 2.6 energy of incident photons (eV)

Fig, 20

~so

2.2~-

~

2.1P 2.OF

/

I

I

t000 2000 energy of vibration (cm -1)

I

3000

Fig. 21

Fig, 20. SER excitation p r o f i l e s from molecules on coldly evaporated s i l v e r films. Symmetric breathing (dots) and C-H stretching mode (rhombs) of surface pyridine (0.2 L), and symmetric scissors mode (triangles) of ethylene (36 L). Lines are guides to the eye. After /119,281/ Fig. 21. Spectral position of resonance maximum as a function of vibrational energy for various adsorbates on s i l v e r films. Dots: pyridine lines (0.2 L ) ; t r i a n g l e s : oxygen ~ines (340 L); rhombs: ethylene lines (36 L); square: "carbon monoxide" l i n e (i.8.10 L; see Chapt. 6). Curve has been calculated (see t e x t ) . After /281/

cules. Here the resonance maximum has been plotted against vibrational energy (the data are from excitation p r o f i l e s plotted against the energy of the incident photons; when the resonances are plotted as a function of the Stokes energy, the spectral pos i t i o n of the resonance maximum is almost independent of the vibration /281/). Note, that the i n t e n s i t i e s of d i f f e r e n t lines for given wavelength of the incident radiation as displayed in Fig. 20 cannot be compared, since the curves have been normalized to give the same i n t e n s i t y for 568,2 nm excitation ( i n t e n s i t y of C-H stretching mode attenuated by a factor of f i v e ) . SER i n t e n s i t i e s increase continuously with increasing wavelength of the incident radiation for pyridine on Cu and Au (Fig. 22; /123/). The data suggest a threshold 40

E

700

wavelengthof )ncident radiation (nm) 600 500

700

wavelengthof incidentradiation (nm) 600 50O i

[

q \

i

/i

i

\\

2/,

2.6

~

i I,I", ',.. \. r,.,.,,

E

/

8

.... ,,,,,

1.8

energy of incident photons (eV)

f o r SERS from these m a t e r i a l s

O.06L. 0.2 L 0.6 L 6 L 60 L-

\

i '"//'-"\I f""~'~'...

F i g . 22. SER e x c i t a t i o n p r o f i l e s f o r v I o f p y r i d i n e on Ag ( d o t s , 0.2 L ) , on Cu ( t r i a n g l e s , 2 L ) , and on Au (rhombs, 2 L). Note e n l a r g e m e n t o f Cu and Au d a t a , Lines are guides t o the eye. A f t e r / 1 2 3 , 2 8 1 /

I

---o----~---""...... ...... 9 -..--i....

2B

22

energy of incident photons (eV)

F i g . 23. SER e x c i t a t i o n p r o f i l e s from s u r f a c e p y r i d i n e on Ag f o r v a r i o u s exposures as i n d i c a t e d (symmetric b r e a t h ing v i b r a t i o n ) . Lines are guides to the eye. A f t e r / 2 6 7 /

a t ~ 2.4 eV and an e x c i t a t i o n

profile

the a c c e s s i b l e wavelength range (> 700 nm). Q u a l i t a t i v e l y ,

maximum o u t s i d e

the r e s u l t s

to those from e l e c t r o d e s u r f a c e s where g o l d a l s o e x h i b i t s

relatively

are s i m i l a r small i n t e n s i -

ties /122/. I n c r e a s i n g exposure leads t o c h a r a c t e r i s t i c of surface pyridine

(Fig.

23; a l l

results

changes o f the e x c i t a t i o n

d i s p l a y e d in t h i s

mono|oyers

10-1 100 101 T ' ' ','"'i ' ' ','"'i ' ' '-'"i ..... ~1~0

2010-2

10.2

>-7t+%

'o>

~

o; ,

10-1

..i...J

,I

1.6

~

0.4

10-1

, ,,,,,,i

,

mono(oyers

100

','1

'

101

' ','"'1

,

~/~'~

, ,,,,

(b):

16~0 |

"~ ii1,...t

'

profiles

s e c t i o n are from s u r -

.

100 exposure

.

,,h,..i

1~ ,

101

(L)

,,h.,I

102

I

/

+

. ,,.,,i

..................... 4,4 ,

lO-1

, .h,,d

100

9 . .n,.I

exposure ( L )

lO1

.

9, h . . I

1

lO2

3. 24. S p e c t r a l p o s i t i o n (a) and i n t e n s i t y (b) o f the resonance maximum in the SER e x c i t a t i o n p r o f i l e s as a f u n c t i o n o f exposure (~1 o f s u r f a c e p y r i d i n e ) . F i l l e d in data p o i n t s in (b) are from e v a l u a t i o n o f i n t e g r a t e d i n t e n s i t i e s i n s t e a d o f peak intensities. Curves have been c a l c u l a t e d (see t e x t ) . A f t e r / 2 6 7 / 41

face pyridine i f not otherwise stated). The resonance s h i f t s from ~ 590 nm (2.12 eV, 0.06 L) to ~ 680 nm (1.82 eV, 60 L). The i n t e n s i t y at maximum increases f o r small coverage, e x h i b i t s an extremum at 0.6 L exposure, and drops again. The spectral pos i t i o n of the resonance as a function of exposure is displayed in Fig. 24a (as in Fig. 16, the upper scale - thickness of pyridine overlayer - neglects the difference of the s t i c k i n g c o e f f i c i e n t of m u l t i l a y e r and surface p y r i d i n e ) . Note, that only 6L exposure (~ 3 layers corresponding to 1.5 nm thickness) are s u f f i c i e n t to displace the resonance by ~ 90% of i t s f i n a l s h i f t for very t h i c k coatings. The i n t e n s i t y at maximum of the resonance ceases to change considerably f o r exposures above 20 L (Fig. 24b). The maximum of t h i s quantity is observed f o r ~ I L corresponding to about h a l f a monolayer pyridine coverage. For t h i c k overlayers, the r e l a t i v e i n t e n s i t y of equivalent surface and bulk pyridine v i b r a t i o n s depends on the e x c i t a t i o n wavelength. The r a t i o I b u l k / Isurfac e for three v i b r a t i o n a l modes is displayed in Fig. 25. Whereas t h i s q u a n t i t y is almost independent of e x c i t a t i o n wavelength for v3' i t increases with e x c i t a t i o n energy for Vl and e x h i b i t s a maximum around 520 nm f o r v4" The influence of annealing on SER e x c i t a t i o n p r o f i l e s of the pyridine breathing v i b r a t i o n (~1' 0.2 L) is shown in Fig. 26. The resonance s h i f t s to shorter wave-

wavelength of incident radiation 700. 600 500 (nm}

wavelength of incident radiation(nm) ?00 600 500 2/, . . . . . . . 0.6

.~_ g

-

2.0

:

~ 1.6

-u~ 1.6

..........~ ~ ' ~ . . ....~" ,..~

02

"~ as := 1,6

Fig. 25

0.4

8

'

'

2.0

'

'

'

'

'

'

''~2

2,2 2/, 2.6 energy of incident photons (eV)

\

~

/

"~=1,2~-//'~i~',.

.

4

0.8"

~t /

i

i

~

[

I

t

I

I

~

'

1.8 2,0 2,2 2./. 2.6 2.8 energy at incident photons (eV)

Fi9. 26

Fig. 25. I n t e n s i t y r a t i o of corresponding bulk and surface pyridine l i n e s as a function of the energy of the i n c i d e n t photons (SERS active f i l m ; 200 L exposure). Rhombs: ~3; dots: Vl; t r i a n g l e s : v 4, Lines are guides to the eye. A f t e r /119/ Fig. 26. Annealing of SER e x c i t a t i o n p r o f i l e s from pyridine on c o l d l y evaporated vsi ~ i l m (symmetric breathing v i b r a t i o n 0.2 L). Dots, f u l l curve: Ts = 120 K; t r i a n g l e s , dotted curve; warmed up to 209 K and recooled to 120 K; rhombs, dashed curve: warmed up to 246 K, recooled to 120 K. Dashed-dotted curve and arrows (pos i t i o n of maximum) have been calculated from absorption spectra (see t e x t ) . A f t e r /281/ 42

length with increasing temperature. The i n t e n s i t y at maximum f i r s t

increases (up to

210 K) and then drops rapidly. In passing we note, that part of the excitation p r o f i l e studies have recently been repeated /360/: the results e s s e n t i a l l y agree with those displayed here. The observations can be explained by assuming a mainly electromagnetic o r i g i n of the e x c i t a t i o n p r o f i l e resonances. In this picture, e x c i t a t i o n of surface plasmon type resonances in "appropriate" roughness features of the surface (bumps) leads to enhanced Raman scattering /190,191,361,362/ (see also Chapts. 2 and 3). I t is assumed that "cold" evaporation creates the "appropriate" surface roughness, i . e . bumps of suitable shape and size. As this is a property of the metal, s i m i l a r e x c i t a t i o n prof i l e s are expected for d i f f e r e n t adsorbates (see Fig. 20; the s h i f t of the resonance is explained below). Because of the optical properties of Cu and Au /363/, electromagnetic resonances in these metals are strongly damped above ~ 2.5 eV, which explains the results of Fig. 22, especially the threshold behaviour. Increasing the r e f r a c t i v e index of the medium adjacent to the rough surface causes a red s h i f t of the electromagnetic resonance (see, e.g., /364/), which is reflected by the excitation spectra displayed in Fig. 23 (see also Fig. 24). The s h i f t contributes a factor of ~ 3 to the 18fold decrease of the Vl peak i n t e n s i t y with coverage for exposures 0.2 L (Fig. 16, e x c i t a t i o n wavelength 514.5 nm; according to the results of Fig. 23 the shape of the i n t e n s i t y versus exposure curve should change with e x c i t a t i o n wavelength, which is indeed observed /358/). The blue s h i f t of the excitation prof i l e resonance upon annealing (Fig. 27) is explained with a transformation of a high density of small bumps on the coldly evaporated films to a lower density of somewhat larger bumps /118,281/. I f the rough surface is modeled in a s i m p l i f y i n g , crude approach by an ensemble of isolated, non-interacting spheroids, q u a n t i t a t i v e comparison of some experimental results with theoretical predictions is possible. F i r s t l y , the spectral dependence of SER i n t e n s i t i e s can be related to the o p t i cal absorption A(m) and the d i e l e c t r i c function ~ = El + i E 2 of the metal by /93/:

ISERS -

~(mL)'A(m L) ~(ms)'A(m S) mL.E2(~L) mS.~2(mS )

(5)

where mL and uS are the frequencies of the incident and Stokes photons. A(m) may be extracted from r e f l e c t i v i t y measurements (/239,281/; see also Chapt. 3). The r e l a tive reflectivity

[1-R(T)/R(295 K)] is a measure of the a d d i t i o n a l optical absorp-

tion of coldly evaporated films with respect to annealed films (Rayleigh scattering neglected, for d e t a i l s see /239/). A(m) is approximated by this quantity, and excitation p r o f i l e s for d i f f e r e n t vibrational energies are calculated with the help of (5). The maxima of the calculated p r o f i l e s follow the solid l i n e in Fig. 21, which nicely reproduces the experimentally observed trend. Within the used approach, the

43

s h i f t of the excitation p r o f i l e resonance with vibrational energy is the consequence of a rather broad absorption p r o f i l e of coldly evaporated films and the rapid variation of the s i l v e r d i e l e c t r i c function in the frequency region of i n t e r e s t (a det a i l e d discussion is given in /281/). S i m i l a r l y , the annealing behaviour of excitation p r o f i l e s (Fig. 26) may be calculated from corresponding r e l a t i v e r e f l e c t i v i t y spectra (Fig. 5 and /281/) by using (5). The dashed-dotted l i n e in Fig. 26 is the calculated spectral dependence of ISERS (no parameters except the height of the curve have been f i t t e d ) . We find reasonable agreement between theory and experiment, Quantitative agreement between calculated and measured s h i f t of the excitation prof i l e maxima with annealing temperature i s , however, poor (arrows in Fig. 26 mark the calculated peaks in ISERS). In addition, r e l a t i v e r e f l e c t i v i t y spectra (Fig. 5) do not show an equivalent to the increase of the SER i n t e n s i t y on annealing to 209 K. This might be explained with partial masking of the "SERS relevant absorption" in the r e f l e c t i v i t y spectra by other absorption processes /281/ and/or the influence of effects not considered so f a r , e.g. a temperature dependent density of SERS act i v e molecules (sites) /267/ on the s i l v e r surface / i 0 0 / . Secondly, the spectral s h i f t of electromagnetic resonances in spheroids due to

confocal d i e l e c t r i c overlayers of f i n i t e thickness can be calculated with a formula derived in /365/. Corresponding results for prolate e l l i p s o i d s f i t

the experimental

data quite well (solid l i n e in Fig. 24a). The details of the calculation are presented elsewhere /267/. We only note here, that the dimensions of the e l l i p s o i d s (bumps) have to be ~ i - 2

nm in order to f i t

the experimental data. This is a con-

sequence of the fast saturation of the resonance s h i f t with pyridine overlayer thickness. Following /175/, we estimate a decrease of the electromagnetic enhancement by 10 for the second layer of adsorbed molecules compared to the f i r s t .

Coldly evap-

orated s i l v e r films e x h i b i t a short-range electromagnetic enhancement (as assumed in /366/) in contrast to some other s i l v e r surfaces investigated (e.g. /255/ and Sect. 4.3). T h i r d l y , the coverage dependence of the i n t e n s i t y at resonance (Fig. 24b) presumably r e f l e c t s the density of SERS active molecules on the s i l v e r surface /267/. What are SERS active molecules? As mentioned e a r l i e r , pyridine adsorbs in two configurations on s i l v e r /20/: a low coverage, e s s e n t i a l l y ~-bonded species (phase I ) , and a high coverage, nitrogen lone-pair bonded species (phase I I ) . As we can trace the SER signal of pyridine down to very small exposures /99/, ~-bonded molecules on certain active adsorption sites as discussed in, for instance, /239/ must be ident i c a l with surface pyridine (see also Sect. 4.4; note, that only part of phase I molecules constitutes the SERS active species). These are subject to the full enhancement (~ 104 , Fig. 16), the electromagnetic (~ 102; Ebulk in Fig. 16) as well as the chemical contribution (~ 102; Echem in Fig. 16). I f we assume that phase I I molecules feel e s s e n t i a l l y only the electromagnetic enhancement and show wI at 993 cm-1 l i k e bulk pyridine, the data in Fig. 24b mirror the exposure dependence 44

of the density of surface pyridine molecules (see also /268/), This density may be estimated in a simple approach under the following assumptions /267/: ing p r o b a b i l i t y for an incident molecule is unity; face molecules adsorb as phase I species;

(iii)

( i ) the s t i c k -

( i i ) on clean parts of the sur-

a molecule, which adsorbs on a sur-

face pyridine covered part of the surface, either starts to build the second layer [probability (I-S)]

or squeezes into the f i r s t

layer and adsorbs as phase I I species

[ p r o b a b i l i t y S; ( 1 - S ) accounts for the fact, that even for m u l t i l a y e r coverage we still

must have some surface pyridine molecules (Fig. 24b); these might be molecules

on selected SERS active sites (details unknown)]; into the f i r s t

( i v ) when a molecule squeezes

layer, i t moves a second, phase I molecule into the upright orienta-

tion (hence the density of phase I I is twice that of phase I ; see also /20/). The d i f f e r e n t i a l equation, which describes the development of the fraction of saturation coverage of surface pyridine molecules (|

as a function of integrated expo-

sure (E), is then given by /267/: @(E) = ( I - S ) { 1 - [ i -

(6)

ES/(I-S)]exp(-E)}

For S = 0.84 (solid) and S = 0.95 (dotted) numerical results are displayed in Fig. 24b. The curves f i t

the experimental data quite well for sub-monolayer coverages.

The agreement is poor for high coverage, probably due to a change of the resonance properties neglected here. Several further observations can be taken as support /267/ for the picture described by (6), for instance the s p l i t t i n g of the N Is peak for pyridine exposed s i l v e r f o i l s as measured by XPS /260,367/ and i t s r e l a t i o n to SER features. Note, that "ordinary", T-bonded pyridine, i . e . molecules on f l a t parts of the surface, does not contribute to low coverage SER spectra in our interpretation which is in agreement with corresponding results from Ag(111) /268/. Although the electromagnetic model explains most features of SER excitation prof i l e s quite reasonably, there remain d i f f i c u l t i e s .

Main problem i s , that optical

properties of very small bumps (~ 2 nm dimensions) are only crudely described by continuum electromagnetic procedures and that our i n t e r p r e t a t i o n allows an only moderate wavelength dependence of the chemical contribution to the enhancement. Moreover, the mode specific spectral dependence I b u l k / Isurfac e (Fig. 25) s t i l l

needs

to be explained. A l t e r n a t i v e l y , an interpretation of the measured excitation prof i l e s based on optical excitations involving charge transfer between pyridine and localized surface electronic states /368/ has been proposed /67,264/. In closing t h i s section we note, that recently published SER excitation profiles from coldly evaporated films on an island structure /94/ and from films condensed at 15 K /369/ are in f a i r agreement with the results presented here.

45

q.l.5

Comparison of Results from Various Experiments

In Fig. 27 we compare the SER features in the region of the strong breathing vibrations from pyridine on various c o l d l y evaporated s i l v e r f i l m s . Mono- as well as m u l t i l a y e r spectra from t h i n Ag films l a i d down on a s i l v e r optical grating (Fig. 27a) are s i m i l a r to the corresponding spectra from c o l d l y evaporated f i l m s displayed in Fig. 8. They are, however, considerably less intense than equivalent measurements at t h i c k films (Figs. 8 and 27e, / 5 0 / ; the spectra shown in Fig. 27a have been recorded by resonantly e x c i t i n g plasmon surface polaritons at the p e r i o d i c a l l y corrugated surface). Figure 27b shows SER spectra from t h i c k Ag films deposited on quartz substrates at 13 K, where part of the substrate had been coated by a s i l v e r island structure. Weak signals from ~ 3 layers of pyridine have been observed a f t e r annealing the exposed sample to ~ 70 K /94,370/ [ t h i s so called "low temperature anneal" /370/ was necessary to see the f u l l

l

i

]

i

I

I

I

1

peak strength; the e f f e c t has been a t t r i b u t e d

I

Fig. 27. Comparison of SER spectra from pyridine on various c o l d l y evaporated f i l m s .

I~U'~

1oo4

m3o

[

(b)

g91

1003

1006

(a): 2 nm of Ag evaporated on s i l v e r grating (A = 800 nm, h = 50 nm) at 120 K; coverage of 0.5 (bottom) and 3.5 monolay~rs of p y r i d i n e ; 50 mW of 514.5 nm rad i a t i o n , 8 cm- bandpass; a f t e r / 5 0 / ; (b): 30 nm of Ag deposited at 150 K; lower curves: exposed at 13 K and "low temperature annealed" at 70 K (bottom: f i l m on quartz substrate, top: f i l m on Ag island s t r u c t u r e ) ; upper curve: f i l m on Ag island structure a f t e r "high temperature anneal" at 200 K, recorded at 150 K; 18.1014 molecules per cm2 coverage; 150 mW of 530.9 nm r a d i a t i o n ; a f t e r / 9 4 / ; (c): lower spectrum: 15 nm of Ag evaporated on s i l v e r substrate at 180 K and exposed to 2 L (uncorrected) of p y r i d i n e ; 300 mW of 488.0 nm r a d i a t i o n , 7 cm- I bandpass; a f t e r /271/; upper spectrum: 50 nm of Ag deposi t e d on copper substrate at I00 K, 5• monolayers of pyridine coverage; 150 mW of 488.0 nm r a d i a t i o n , 7 cm-1 bandpass; a f t e r /366/;

991

1032

(d): t h i c k Ag f i l m deposited on polished copper substrate at 15 K, condensation of t h i c k (250- I000 nm) pyridine overlayer; top: sample at 15 K, bottom: sample annealed to 200 K; 250 mW of 647.1 nm r a d i a t i o n , 5 cm-1 bandpass; a f t e r /369/; (e): t h i c k Ag f i l m evaporated on polished Cu substrate at 120 K, 0.2 L of pyridine exposure; sample annealed to 210 K; 200 mW of 514.5 nm r a d i a t i o n , 4 cm-1 bandpass; a f t e r /358/. I 1060

46

I

I

L

I

I

I

1020 980 Rarnon shift ( cm-1}

I

The bars to the l e f t of the spectra represent 102 c t s / s , except for (e), where i t represents 103 cts/s

to thermally stimulated, i r r e v e r s i b l e movement to and/or r e o r i e n t a t i o n of pyridine molecules at active sites /370/; annealing of defect sites in the bulk of evaporated f i l m s might also c o n t r i b u t e via electromagnetic enhancement mechanisms, which are stronger f o r less disturbed layers (see Chapt. 3 ) ] .

I t is somewhat s u r p r i s i n g , t h a t , -1

besides the surface pyridine l i n e , only a weak bulk p y r i d i n e l i n e at ~ 990 cm

.

is

observed (overestimation of coverage?). A f t e r a "high temperature anneal" /94/ the spectra are f u r t h e r enhanced (Fig. 27b) s i m i l a r to those discussed in Sect. 4,1.3. Within experimental accuracy, peak positions of ~i and ~6 are i d e n t i c a l to those of pyridine on t h i c k c o l d l y evaporated Ag films l a i d down at 120 K and annealed to 210 K (Fig. 27e). Figure 27c shows spectra from roughly a monolayer of p y r i d i n e on a t h i n Ag f i l m deposited at 180 K /271/ and from 5• i layers on a t h i c k f i l m deposited at i00 K /366/. Although spectral features are only weakly pronounced, the add i t i o n a l l i n e at ~ 991 cm-1 f o r the t h i c k e r coating is c l e a r l y seen. F i n a l l y , t h i c k pyridine layers (250- 1000 rim) l a i d down on t h i c k s i l v e r films deposited on a polished copper block at 15 K /369/ e x h i b i t spectra s i m i l a r to samples prepared at 120 K (compare Fig. 27d to Fig. 27e and Fig. 8). Annealing to 200 K (desorption of m u l t i l a y e r pyridine) has a s i m i l a r e f f e c t as discussed in Sect. 4.1.3. SERS from t h i c k pyridine layers on s i l v e r f i l m s deposited and exposed at l i q u i d He temperature has been reported e a r l i e r /133/. Here spectral features of surface pyridine are very s i m i l a r to corresponding results from Ag deposited at 120 K (see Table 2 and Table 3 in /133/). In summary, t h i c k c o l d l y evaporated s i l v e r films seem to be stronger enhancers than t h i n f i l m s . For sub-monolayer pyridine coverage, SER spectra are dominated by a species with v I at 1003- 1006 cm-l, and v 6 at 1030- 1036 cm- I is r e l a t i v e l y weakly pronounced. A t h i r d l i n e at 991-993 cm-1 develops with increasing coverage. The i n t e n s i t y of v 6 grows f a s t e r with coverage than that of ~1 and matches or exceeds the l a t t e r f o r a coverage of several layers. These f a i r l y consistent experimental data can be understood w i t h i n the frame o u t l i n e d in the preceeding sections, i . e . can be interpreted in terms of s t r o n g l y enhanced surface and weakly enhanced bulk pyridine contributions. For the sake of completeness we b r i e f l y mention i n v e s t i g a t i o n s on films evaporated or sputtered and exposed at room temperature /153,157/. Spectra of rather low q u a l i t y e x h i b i t weak features at ~ 1009 cm- I and ~ 1035 cm-1 ( s i m i l a r features are observed from Pt, Pd, T i , or Ni films /154/, where Ni shows a much stronger signal than Ag!). To the opinion of the author, careful cross checks of the experimental conditions are necessary before any sound conclusion can be drawn from these somewhat unique r e s u l t s .

47

4.2

Coldly

Evaporated

C o p p e r a n d Gold Films

SER spectra from various coldly evaporated noble metal films (group Ib) exposed to 0.2 L (Ag) and 2 L (Au,Cu) of pyridine are displayed in Fig. 28 /123/. Corresponding l i n e i n t e n s i t i e s from s i l v e r and copper samples are comparable, whereas the gold sample exhibits only v I with a roughly 30times smaller i n t e n s i t y . In f a c t , a f t e r evaporation and exposure we did not observe any l i n e from Au. The spectrum shown in Fig. 28 has been recorded a f t e r warming the sample to 210 K and recooling to 120 K, a procedure known to increase SER i n t e n s i t i e s from Ag /100/ (very recently, a somewhat more intense spectrum from Au films has been reported /369/; l i k e Cu, but unl i k e Ag, the " q u a l i t y " of SERS active Au films depends c r i t i c a l l y

on evaporation

conditions; see /273/ for Cu). The r e l a t i v e SER l i n e i n t e n s i t i e s of pyridine on Cu

8

i

Ag

E 2

Jo

I

1215 j

159,: ~

0

i

I

loo5

/ 1035~ 1067' H

I

,941 -~'

i

i

?48

i

i

699 ~2z" , )

[ J

t ~"

/.11 I~'

~

J

I

I

x

2 0

1218 Io4di A ,o,,~Jl ~,o

16oi "-"-r-

.....

--"-r" ~ - ,

~

633

F

A

,,/

.,---/

Z'

" ""--'T

"

'

~

,

o/ Q~

a,

,s'oo

'

,;oo

Roman shift

' ( cm -1)

s;o

'

Fig. 28. SERS from pyridine on coldly evaporated Ag (0.2 L), Cu (2 L), and Au films (2 L). 60 mW of 676.4 nm radiation, 4 cm- I bandpass. Au f i l m has been warmed to 210 K and recooled to 120 K before measurement. After /123/

are d i f f e r e n t from those on Ag as is e a s i l y seen from the i n t e n s i t y r a t i o of ~4 (~ 1600 cm-1) and ~i (~ 1000 cm-1

Fig. 28)

SER l i n e positions and r e l a t i v e i n t e n s i t i e s of pyridine on Cu and Ag films are compared to corresponding data from electrode surfaces in Table 3 /123/. Spectral features are almost identical for Cu films and electrodes. Compared to s i l v e r , lines from pyridine on Cu films are only s l i g h t l y shifted to greater energies, which suggests s i m i l a r bonding on both metals. I n t e r e s t i n g l y , SER data from Cu samples approach those of the copper pyridine complex. For red e x c i t a t i o n , l i n e i n t e n s i t i e s of high energy modes (e.g. ~4) are more pronounced for copper than for s i l v e r . However, exciting s i l v e r samples with green or blue radiation (Table 3 and Fig. 9), r e l a t i v e l i n e i n t e n s i t i e s s i m i l a r to those from Cu are observed (red l i g h t excitat i o n ) , Changes of r e l a t i v e l i n e i n t e n s i t i e s with excitation wavelength have been explained with mode specific excitation profiles (Sect. 4.1.4 and /119,360/). The s h i f t of the resonance-like p r o f i l e s to greater excitation energy with vibrational 48

Table 3. Vibrations of pyridine in various systems. (a): on activated s i l v e r electrode at -0.6 VSCE, 647.1 nm excitation, after /123,371/; (b): on SERS active s i l v e r film in UHV (0.2 L exposure), 676.4 nm excitation, after /123/; (c): complexed pyridine, Cu(PY)2Ni(CN)4, IR study, after /347/; (d): on activated copper electrode at -0.6 VSCE, 647.1 nm excitation, after /123,371/; (e): on SERS active copper film in UHV (2 L exposure), 676.4 nm excitation, after /123/

(a) Mode Symmetry

(b)

(c)

SERS from Ag Electrode I C5H5N

Film C5H5N

IR Complex Cu(PY)2Ni(CN) 4

(d) (e) SERS from Cu Electrode Film C5H5N

C5H5N

385

(3)

(m)

422

(14)

421

(9)

640

(m)

635

(67)

633

(32)

650

(w)

652

(12)

(7)

689 (vs)

699

(9)

701

(2)

(6)

753

755

(4)

946

(6)

950

(2)

v21,A2

391

(5)

v27,B2

419

(12)

411

(7)

435

v3 " A1

635

(55)

624 (27)

Vl2,Bl

651

(9)

~26,B2

698

(6)

699

v23,B 2

753

(4)

748

v25,B 2

872

(3)

943

(6)

(s)

868 (vw)

v20,A 2 v24,92

941

(5)

949

(m)

v22,A2 v I , A1

1013 (100)

lOO5 (lOO)

1017

(s)

1013 (I00)

~6 ' AI

1036

(20)

1035 (19)

1043

(s)

1041

(19)

1040

(4)

v8 ' AI

1067

(17)

1067

1068

(s)

1067

(31)

1068

(8)

1154 (s)

1156

(7)

1219

(s)

1218

(64)

Vll,B I

1241

(s)

v15,B I

1360

(w)

1359

(2)

(7)

I010 (I00)

lO88 (w)

Vl7,Bl Vl6,Bl

1157

(3)

~5 ' A1

1216

(34)

1215

(8)

1218 (21)

v18,Bl

1449

(2)

1449 (vs)

1448

(8)

v9 ' A1

1486

(3)

1487

(s)

1486

(8)

1485

(2)

Vl4,B I

1574

(3)

1575

(m)

1571

(8)

1573

(2)

v4 ' A1

1602

(46)

1609 (vs)

1603

(77)

1594

(11)

1601 (33) 49

energy causes high energy l i n e s to be less prominent for red e x c i t a t i o n (Ag). As the e x c i t a t i o n p r o f i l e maximum for Cu is s h i f t e d to longer wavelength with respect to Ag (Fig. 22, / 1 2 3 / ) , one might expect q u a l i t a t i v e l y s i m i l a r r e l a t i v e i n t e n s i t i e s for Cu (red e x c i t a t i o n ) and Ag samples (green, blue e x c i t a t i o n ) , as one works in e i t h e r case on the high energy side of the e x c i t a t i o n p r o f i l e resonance. Indications f o r t h i s trend can be extracted from Table 3. Note, that weak signals from pyridine on Au and Cu have also been detected with green e x c i t a t i o n , but not with blue /123/. Another remarkable feature is the rather low i n t e n s i t y of w6 from Cu samples (Fig. 28 and /123/) which is not understood at present. A recent comparat i v e SER study of pyridine on Ag, Cu, and Au deposited at 15 K /369/ has confirmed the results from /123/ o u t l i n e d above. Recently, coverage dependence and annealing behaviour of SER features from pyridine on c o l d l y evaporated Cu films have been studied in d e t a i l /273/. Q u a l i t a t i v e l y , s i m i l a r effects as with s i l v e r have been found (Sects. 4.1.2 , 3 ) . However, spect r a l features are more stable against annealing. This might be explained with the lower m o b i l i t y of Cu surface atoms compared to Ag /274,275/. In f a c t , Cu films l a i d down at room temperature or samples annealed t o / a t room temperature /273,369/ may still

be SERS active. Figure 29 displays Raman spectral features from pyridine on

f i l m s deposited at ~ 130 K, at ~ 230 K, and at room temperature, respectively. A l l films have been exposed and measured at 120 K. In contrast to c o l d l y evaporated f i l m s , samples prepared at room temperature display pronounced bulk pyridine peaks ( w i t h i n experimental accuracy at the same energy as for s i l v e r ) . The spectra displayed in Fig. 29 mirror presumably the d i f f e r e n t surface topography and bulk pro-

2F i0~ Jl~ 993 ]

2O

101o

|

1032 [990

,o39/J o__L_._L 0.~ -

05' 1~ 10/'01~ 1080 1000 Rarnan shift (cm-I )

50

SER spectra from Cu films deposited at ~ 130 K (bottom), ~ 230 K ( c e n t r a l ) , and at room temperature (top). All samples are exposed to 2 L of pyridine at 120 K and measured at t h i s temperature. 190 mW of 647.1 nm r a d i a t i o n , 4 cm- I bandpass. As spectra are from d i f f e r e n t experimental runs, absolute i n t e n s i t i e s cannot be compared. A f t e r /273/

perties of films prepared at d i f f e r e n t temperatures. The magnitude of both, electromagnetic and chemical contributions to SERS, change with deposition temperature. The l a t t e r , because point defect related, SERS active sites (Sects. 4.1.4 and 4.4) anneal at higher temperatures, the former, because the density of bulk defects is smaller and the shape and density of surface bumps might be more favourable for films prepared at higher temperature (see also Sects. 4 . 1 . 3 , 4 ) .

Obviously, Cu sam-

ples prepared at ~ 230 K give the best SERS performance: both, surface (chemical and electromagnetic enhancement) and bulk pyridine lines (only electromagnetic enhancement), are very strong. A somewhat more detailed discussion may be found elsewhere /273/. Like Cu, gold samples may e x h i b i t enhanced Raman signals at room temperature /369/. Here more experimental work is necessary. Besides the noble metals of group Ib, only coldly evaporated sodium /142/ and l i t h i u m /133/ films (see Chapt. 5) as well as aluminum samples /144/ display SERS. The l a t t e r does not show, however, any characteristic l i n e a f t e r pyridine exposure /144/.

4.3

4.3.1

Surfaces

Prepared

with Various

Techniques

Silver

Raman spectral features in the breathing mode region for pyridine on various s i l v e r surfaces are compared in Fig. 30. Approximately three layers of pyridine deposited on island films at 13 K and annealed to 70 K ("low temperature anneal", /94/) display bulk (990 cm" I ) and surface (I000 cm- I ) pyridine ~1 peaks with approximately the same i n t e n s i t y . The antisymmetric stretching vibration ~6 is almost as strong as ~ I ' which indicates bulk as well as surface contributions to t h i s l i n e . Only about 150 cts/s peak i n t e n s i t y have t y p i c a l l y been observed. Upon annealing to 200K, a l l features except a small peak at ~ 1006 cm- I disappear /94/. For thin as well as thick layers on s i l v e r optical gratings no surface pyridine l i n e is seen, and ~6 is rather weakly pronounced (Fig. 3Oh, /48/). The spectra have been taken by resonantly exciting plasmon surface polaritons. The coverage dependence of the Raman features points to a pronounced f i r s t

layer e f f e c t . Signals from the f i r s t

layer are ~ 100

times stronger than those from subsequent layers /48,268/. Note the overall weakness of the signals (peak i n t e n s i t y ~ 50 c t s / s ) . The enhancement of Raman bands from pyridine on iodine roughened Ag (Fig. 30c) has been a t t r i b u t e d to long range electromagnetic effects caused by surface roughness features of ~ 50 nm l a t e r a l dimension /98/. Two d i f f e r e n t adsorbed states corresponding to the f i r s t

layer and succeeding

layers, respectively, with overlapping ~6 are believed to contribute to the spectra. Again, note the weakness of the spectra with peak count rates of ~ 50 cts/s. The sputter-cleaned s i l v e r surfaces used in /97/ also lead to r e l a t i v e l y weak signals. 51

I I

L I

L '1oo~

199~

~,o3o"/"~/~

'

Fig. 30. Comparison of Raman spectra from pyridine on various s i l v e r substrates. (a): ~ 3 layers on island f i l m (~ i00 nm l a t e r a l dimensions); 150 mW of 530.9 nm r a d i a t i o n ; a f t e r / 9 4 / ; 1032

1002 . .990

c

(b): ~ 1 (bottom) and ~ 25 (top) layers on p e r i o d i c a l l y corrugated A g ( l l l ) surface (A = 1000 rim, corrugation depth ~ I00 nm); 150 mW of 514.5 nm r a d i a t i o n , 6 cm-~ bandpass; a f t e r / 4 8 / ; (c): ~ i (bettom) and ~ 4 (top) layers on photochemic a l l y roughened s i l v e r surface; several hundred mW (?) of 488.0 nm r a d i a t i o n , 8 cm-1 bandpass; a f t e r / 9 8 / ;

I

(d): ~ 0.3 (bottom) and ~ 1.2 (top) layers on s p u t t e r cleaned Ag f o i l ; 100 mW of 514.5 nm radiation~ 6 cm- I bandpass; a f t e r / 9 7 / ; (e): ~ 1.5 (bottom) and ~ i0 (top) layers on "smooth" Ag(lO0); i00 mW of 514.5 nm r a d i a t i o n , 5.5 cm- I bandpass; a f t e r /101/; ( f ) : ~ i monolayer on smooth A g ( l l l ) ; i W of 514.5 nm r a d i a t i o n , i0 cm- I bandpass; a f t e r /372/. 1060

1020 980 Roman shift (cm-1)

The bars to the l e f t of the spectra represent 50 cts/s ( a - d ) and 1 ct/s ( e , f )

Sub-monolayer and ~--monolayer spectra do not d i f f e r very much. Both e x h i b i t wI peaks at 990 cm- I as well as at 1002 cm-1 along with a ~6 l i n e at 1032 cm-1 of somewhat smaller i n t e n s i t y . A l l features disappear upon annealing the sample to ~ 200 K /97, 269/. The two ~1 l i n e s have been interpreted as being due to pyridine adsorbed to two d i f f e r e n t sites. M u l t i l a y e r Raman scattering was not observed below an exposure of 330 L (uncorrected). F i n a l l y , mono- and m u l t i l a y e r spectra from pyridine on Ag(lO0) (Fig. 30e, /101/) are q u a l i t a t i v e l y s i m i l a r to results from chemically roughened samples (Fig. 30c). The i n t e n s i t y i s , however, much smaller (~ 2 cts/s peak i n t e n s i t y ! ) . The enhancement for the surface pyridine l i n e has been estimated to ~ 400 /101/. These results disagree with recent experiments on A g ( l l l ) ,

Ag(110),

and Ag(lO0) surfaces (Fig. 3Of, /372/). Here only a sin21e ~I peak at 993 cm-1 along with w6 at 1034 cm- I has been observed. The Raman i n t e n s i t y increases l i n e a r l y from sub-monolayer to m u l t i l a y e r coverage and the depolarization r a t i o is low. This points to ordinary Raman scattering /372/. I t has been argued / 2 8 / , that the l i n e at 1004 cm- I in the spectra of /101/ (Fig. 30e) is caused by pyridine on spe52

cial s i t e s , which are usually not available on c a r e f u l l y prepared single crystal surfaces (presumably defect sites l i k e steps, kinks, adatoms or vacancies as suspected in /372/). A comparison of the ordinary spectrum from A g ( l l l ) (Fig. 3Of) to the spectra displayed in Fig. 30a-d suggests a weak enhancement o f ~ 102 f o r the l a t t e r (note, that this crude estimation neglects differences in the experimental procedures). The SER spectra ( a ) - (d) in Fig. 30 have some features in common. They display, even f o r monolayer or less coverage, e i t h e r both, the bulk as well as the surface pyridine l i n e , or only the bulk pyridine Vl signal. The bands are roughly two orders of magnitude weaker than corresponding lines from thick coldly evaporated films (Fig. 27e), and the strong domination of the surface pyridine l i n e f o r low coverage is not observed. A small density or absence of SERS active sites and hence of surface pyridine in combination with electromagnetic enhancement due to surface roughness (grating, island) can q u a l i t a t i v e l y account for the observed features. The comparison of Fig. 27 and Fig. 30 provides additional evidence, that the strong enhancement (electromagnetic plus chemical) is r e s t r i c t e d to surface pyridine only.

4.3.2

Other Materials

The report of appreciably enhanced Raman signals (~ 105 ) from pyridine on a drop of mercury /149/ has recently caused much excitement. Because of i t s optical properties and i t s smooth surface, no electromagnetic enhancement is expected from mercury. Spectra from the drop in gaseous or l i q u i d pyridine or benzene have been compared to spectra recorded without the mercury drop. With mercury present, about 20 times larger i n t e n s i t y of ~1 has been observed, from which the enhancement given above was estimated (Vl is found at 992 cm-l; for l i q u i d benzene the i n t e n s i t y is only doubled with the drop present). There are, however, several unsuccessful attempts by other groups to reproduce the results /67/. I t is the opinion of the author, that the results reported in /149/ are either experimental artefacts and/or are misinterpreted. They should not be included into the general SERS discussion. In /153,154,157/ Pd, Pt, Ti, or Ni films evaporated or sputtered at room temperature have been exposed to saturated pyridine vapour for 1 h, evacuated for 30 min at 10-5 Torr, and investigated. Weak features at pyridine frequencies in Raman spectra of rather moderate q u a l i t y have been interpreted as surface enhanced lines. We think, that the presented data do not j u s t i f y this conclusion. More experiments to separate ordinary Raman contributions along with a detailed, q u a n t i t a t i v e evaluation of i n t e n s i t i e s is necessary for sound conclusions. Like the mercury r e s u l t , the observations reported in /153,154,157/ should not enter the general discussion of SERS at present.

53

4.4

D i s c u s s i o n and C o n c l u s i o n s

From the rich body of SER studies of pyridine adsorbed to metal/vacuum interfaces several conclusions can safely be drawn. Three species can be distinguished in the spectra from s i l v e r , which is best i l l u s t r a t e d with the symmetric breathing vibration ~I" ( i ) Surface pyridine displays wI at 1003- 1006 cm- I . Only t h i s species is subject to the strong enhancement (~ 104 for coldly evaporated s i l v e r f i l m s ) , which is composed of an electromagnetic (~ 102 ) and a chemical (# 102 ) part. Surface pyridine are molecules chemisorbed to certain SERS active sites at the s i l v e r surface (see below). Bulk pyridine consists of two species. ( i i ) Molecules bonded to SERS inactive parts of the metal surface display wI at 990-993 cm-1, which indicates physisorption. These feel e s s e n t i a l l y only the electromagnetic part of the enhancement (a weak chemical contribution cannot be excluded). ( i i i )

wI is found between

991- 996 cm- I f o r m u l t i l a y e r pyridine, which overlaps the range given under ( i i ) . There is some evidence, that ~I of m u l t i l a y e r pyridine is s l i g h t l y shifted to greater energies compared to ( i i )

/358/. M u l t i l a y e r pyridine exhibits e i t h e r only ordinary

Raman scattering or weak SERS caused by long range electromagnetic effects (gratings, island f i l m s ) . Short range electromagnetic enhancement as observed from coldly evaporated films is e s s e n t i a l l y r e s t r i c t e d to ( i i ) and, of course, ( i ) . We take the opportunity to emphasize, that, following common usage, the label "SERS active" for a s i t e or adsorbed molecule refers to the chemical contribution to SERS. SERS inactive molecules (molecules on inactive sites) may well display Raman features surface enhanced by electromagnetic effects as outlined. I t follows, that intense SER spectra can only be expected from substrates, which simultaneously e x h i b i t ( i ) a high density of SERS active s i t e s , ( i i ) appropriate optical properties ( i . e . small damping), and ( i i i )

suitable surface topography to

support electromagnetic resonances. This is reflected by the f a c t , that SERS so far

has convincingly been demonstrated only for appropriately prepared metals of high reflectivity (Ag, Au, Cu, L i , Na). Coldly evaporated films obviously give the best performance: these meet apparently ( i ) to ( i i i )

in a unique way. On the other hand,

s i l v e r gratings or island films give r i s e to weaker SER spectra featuring mainly bulk pyridine: the density of SERS active sites is small on these surfaces, whereas ( i i ) and ( i i i )

are matched.

On s i l v e r , SERS active sites are stable only at low temperatures (~ 220 K). They anneal at room temperature. Moreover, SERS a c t i v i t y depends on the adsorbed molecule, i . e . is molecule specific: methane and ethane on otherwise SERS active surfaces do not show SERS /133/. Certain bonding properties or adsorption geometries to the active sites are presumably necessary for the chemical e f f e c t in SERS. In other words, SERS a c t i v i t y is a property of the entire complex, adsorbed molecule plus active s i t e . Therefore, SER features may change with, for instance, pyridine coverage, because

54

bonding properties and hence the density of SERS active surface pyridine may change, although the density of active metal sites is not affected (Sect. 4.1.2). For coldly evaporated f i l m s , the electromagnetic contribution to SERS is responsible f o r the observed e x c i t a t i o n p r o f i l e resonances. As expected within t h i s interpretation, the resonances s h i f t with coverage to the red and upon annealing s l i g h t l y to the blue. Note, that there is room for only a rather f l a t wavelength dependence of the chemical contribution to SERS within t h i s concept. Variations of SER spectral features and i n t e n s i t i e s for d i f f e r e n t l y prepared surfaces (Figs. 27, 30) are caused by several effects. The electromagnetic and chemical share of the total enhancement depends on surface preparation. The density of SERS active sites is influenced by the surface treatment, and the scale of supra-atomic roughness features and hence the range of the electromagnetic enhancement is d i f ferent for d i f f e r e n t surfaces. The remarkable decrease of SER i n t e n s i t i e s from surface pyridine on coldly evaporated films for exposures ~ 0.5 L (Fig. 16) is due to two e f f e c t s , the red s h i f t of electromagnetic resonances with coverage, which is also responsible f o r the s l i g h t decrease of the bulk pyridine signal from physisorbed molecules, and the change of the surface pyridine density (Sects. 4 . 1 . 2 , 4 ;

the den-

s i t y of active metal surface sites does not change). The annealing behaviour is more complicated: desorption of molecules may be involved. The increase of the surface pyridine signal between 180 K and 220 K is presumably mainly caused by physisorbed pyridine molecules, which s t a r t to migrate at the surface and e i t h e r f i n a l l y desorb or are trapped at vacant active sites in t h i s temperature range. The l a t t e r process increases the surface pyridine density. Other contributions to the increase of the surface pyridine signal may come from the blue s h i f t of the e x c i t a t i o n p r o f i l e resonance with temperature (Fig. 26) and an orientation of surface pyridine to more favourable adsorption geometries (hindered at low temperature and/or by neighbouring molecules). Several effects contribute to the decrease of the SER signal f o r T 210 K (Fig. 26). Most l i k e l y , annealing of active sites plays a leading role. Annealing of supra-atomic scale roughness (bumps), slow desorption of c~emisorbed pyridine, or change of bonding properties may also contribute. In closing this part we note, that s i m i l a r conclusions have recently been drawn in another paper /373/. I t remains an important open question: what is the nature of SERS active s i t e s ? There is strong evidence that atomic scale roughness is involved /67/. As coldly evaporated surfaces are poorly defined, i t i s , however, d i f f i c u l t

to extract de-

t a i l s from experiments. Early attempts to demonstrate the importance of atomic scale roughness by deposition of ~ monolayer amounts of s i l v e r at low temperature on inactive s i l v e r surfaces were unsuccessful /271,373,374/. Studies of appropriately prepared single crystal faces with known defects and defect densities seem to be more promising. Such experiments are j u s t about to s t a r t /372/. Despite appreciable lack of knowledge, available SER vibrational data and UPS results /20/ may allow some preliminary conclusions on the p y r i d i n e / s i l v e r system. The ideas contain, however, 55

some degree of speculation and have to be cross checked by suitable experimental studies. Coldly evaporated s i l v e r films have presumably a (111) f i b e r texture /277,375/ with various kinds of surface defects. Figure 31 schematically displays a surface with d i f f e r e n t imperfections. These include i r r e g u l a r i t i e s l i k e adatoms, vacancies, clusters, kinks, steps, and dislocations and adsorbed or incorporated impurities.

p'? '

o-,d

r.,..~ '---V..,,%

..

t

.2

~'//////~2~7//"/"

bump

/kink :~

/ />"'~-~

~

~/"~r

v_a,.c._an_cY _ stepvacancy~ subsurf.clce

impurities

" y'i,/ ~, - w / ~r~ ~

/

2 nm depth) and molecules w i t h i n the c a v i t i e s are subject to an electromagnetic enhancement (~ 3.104 ) due to c a v i t y resonances /781/. This conclusion is doubted in /783,785/. For island f i l m s , the magnitude of the plasmon resonance contribution to SERS has been determined to be % 103 /784/. I t was concluded, that other mechanisms provide an additional enhancement of ~ 103 for the investigated system. The chemical contribution to SERS is further discussed in /786-801/. For the ground-state charge transfer model /224/ an enhancement factor of 102- 103 is e s t i mated /787/. I t is pointed out, that many characteristic SER features (e.g. specif i t y to adsorption s i t e , appearance of forbidden bands) can be understood within this model. I t is remarkable, that the model predicts a f l a t excitation p r o f i l e . Due to p a r t i c i p a t i o n of excitonic or interband excitations in the scattering process, Raman signals from molecules adsorbed on semiconductor surfaces may be enhanced /788,789/. Experimental evidence for t h i s e f f e c t has been presented /756, 790-792/. The role of charge transfer excitations in SERS (excited-state charge 128

transfer model /236-238/)

is discussed in various recent experimental and t h e o r e t i -

cal papers /793-801/. P a r t i c u l a r l y noteworthy is the statement given in /801/, that advanced models of SERS should comprise both aspects, ground-state and excited-state charge transfer (these are termed "vibrational driven hopping" and "coherent tunneling" charge transfer in /801/). The assumption, that the chemical contribution to SERS is p a r t i c u l a r l y strong at or r e s t r i c t e d to defect sites (concept of SERS active s i t e s ) , is supported by various recent results from mainly electrode surfaces /802808/, although controversial views s t i l l

e x i s t /802,809,810/.

Time dependent Hartree-Fock calculations have been used to determine the Raman enhancement of Li n - H 2 clusters /811/. The calculated p o l a r i z a b i l i t y d e r i v a t i v e enhancement of 103- 104 may be partitioned into three terms, an electromagnetic part, a term arising from the modulation of the cluster metal o r b i t a l s by the vibration of the adsorbate, and a part involving charge transfer between metal and adsorbate. Laser beam induced photodecomposition effects and t h e i r impact on the interpretation of SER features from mainly electrode surfaces are discussed in /812-818/.

Experimental.

Optical properties, structure and surface roughness of coldly evap-

orated films and t h e i r r e l a t i o n to SERS are discussed in /819-824/. The influence of the size dependence of the d i e l e c t r i c function on the optical absorption of i s land films is investigated in /825/. Electronic and vibrational features of matrixisolated clusters and small p a r t i c l e s (noble metals, s i l i c o n ) are studied in /826832/.

Pyridine Adsorption.

Angle resolved UPS measurements from pyridine on Pd(111) at

room temperature suggest bonding to the surface through both the nitrogen atom and the ring plane /833/. Therefore an adsorption geometry with the ring plane t i l t e d with respect to the surface plane is proposed, which is s i m i l a r to the high coverage phase of pyridine on A g ( l l l )

(see /20/ and Chap. 4).

A detailed SER study of pyridine on copper (colloids) is presented in /834/. SER signals from pyridine on ~ palladium hydride (electrode) are discussed in /757/. Continuing e a r l i e r work /153,154,157/, Raman scattering from pyridine adsorbed to various vacuum evaporated metal films (room temperature, 10-5 Torr) as well as to semiconductor surfaces is studied in /835/. An enhancement factor of ~ 104 has been estimated for pyridine on Ag, Pd, and Ni (strongest e f f e c t for Ni). The conclusions are in c o n f l i c t with those of other groups (see, e.g., /272/ and Sect. 3.2).

Hydrocarbon Adsorption.

The bonding and surface chemistry of hydrocarbons and hydro-

carbon fragments on metals has recently been reviewed /836-839/. Numerous new investigations t r e a t adsorption and surface reactions of ethylene / 8 4 0 - 8 4 9 / , propylene / 8 5 0 - 8 5 2 / , and acetylene /841,842,853-856/ on metal surfaces (mainly t r a n s i t i o n metals of group V I I I , but also s i l v e r /844,847,852/). Decomposition of adsorbed unsaturated hydrocarbons into hydrogen-impoverished species has 129

frequently been studied /837,840,848,851,854/. These investigations are interesting with respect to the results discussed in Sects. 5.1.2 and 5.1.4, where annealing induced changes of SER spectra from coldly evaporated s i l v e r films exposed to ethylene and acetylene have been explained with formation of such species. The role of oxygen pre-exposure for adsorption of ethylene on s i l v e r is investigated in /844, 847/. An enhanced interaction was observed even at 77 K /847/. SER spectra of ethylene on coldly evaporated s i l v e r films obtained by a Russian group /857/ are in l i n e with those displayed in Sect. 5.1.2. A detailed SER study of aminobenzoic acid on s i l v e r (colloids) is presented in /858/. Oarbon Monoxide Exposure and Carbonaceous Deposits. The continuing i n t e r e s t in carbon monoxide interaction with metal surfaces is reflected by the great number of papers s t i l l

published in t h i s f i e l d . State of the a r t overviews on bonding, vibra-

tional features, and surface reactions are given in /746,859,860/. Vibrational spectra and t h e i r r e l a t i o n to the structure of CO adlayers on various single crystal metal surfaces are discussed in /861-869/, coverage dependent s h i f t s and changes of the l i n e shape of vibrational bands are treated in /746,866,870-872/. Studies of carbon monoxide interaction with copper /873- 878/ and s i l v e r /844,878,879/ accentuate the role of defect sites /875-877/ and l a t e r a l interaction of the v i b r a t ing molecules /875,877,879/ for the interpretation of ~he observed band frequencies and shapes. In agreement with SER data (Sect. 6.1), a downshift of the CO stretching frequency from 2140 cm- I to 2110 cm-1 with increasing coverage on s i l v e r has been measured and attributed to dipole-dipole coupling between the adsorbed molecules /879/ (IR study of matrix-isolated Ag clusters). Due to i t s importance for heterogeneous c a t a l y s i s , thermally (or otherwise) activated fragmentation of adsorbed hydrocarbons or carbon monoxide and formation of carbonaceous layers on metal surfaces has been the subject of several recent studies (e.g. /837,854,880-882/).

Even at low temperature (140 K) cyclotrimerization of

acetylene to benzene has been observed on P d ( l l l ) /854/. The dependence of Raman spectral features on structure and microtexture of carbon films is discussed in /883/. Oxygen Exposure.

Interaction of oxygen with group V I I I metal surfaces is treated

in several recent papers /860-862,884-891/.

Chemisorbed molecular species have

been found to co-exist with atomic oxygen also on Pd(lO0) /884/ and Rh(lO0) /890/ (T ~ i00 K). The role of defect sites (steps, kinks) for oxygen adsorption is investigated in /861,862,889,892/. Several studies are concerned with oxygen adsorption on p o l y c r y s t a l l i n e s i l v e r films /847,893/ or f o i l s /894,895/ and s i l v e r single crystal surfaces /844,896-901/. Uncertainty in interpreting the enormous downshift of WO_0 of molecular oxygen on Ag(llO) compared to the gas phase value s t i l l

persists /896,897/. A t e n t a t i v e i n t e r -

pretation in terms of s i n g l e t oxygen analogous to the system 02/Cu(llO ) /644/ has

been presented /896/. A very small sticking c o e f f i c i e n t has been found for oxygen adsorption on A g ( l l l ) /898/ corroborating e a r l i e r results /651,652/. Whereas no mmolecularly adsorbed oxygen was found on A g ( l l l ) at I00 K /846/, chemisorbed 02 has been detected on cesium- and potassium-dosed s i l v e r f o i l s even at room temperature /894/ as well as on p o l y c r y s t a l l i n e f o i l s of Ni, Cu, Ag, and Pt at 80 K /895/. In f a i r agreement with EELS data from Ag(llO), vibrational bands at ~ 240 cm- I , 314 cm- I , and 630 cm- I have been reported for Ag f o i l s /895/. With respect to Raman studies of coldly evaporated A1 films exposed to oxygen (Sect. 7.2), the IR investigation of matrix reaction products of oxygen and ozone with aluminum atoms /902/ is interesting. Water Adsorption.

A detailed theoretical analysis of the O-H stretching band of

l i q u i d water and ice is given in /903/. Several recent investigations t r e a t water adsorption and interaction with adsorbed oxygen atoms on various metal surfaces / 9 0 4 - 9 1 1 / . Formation of OH species has been reported to s t a r t at temperatures as low as 80 K /910/ (oxygen pre-coated Ag(lO0) surface). Water monomers have been observed on Cu(lO0) and Pd(lO0) at i0 K /911/ (for small exposures % 0.4 L). They s t a r t to cluster when warming the sample to only 20 K. The necessity of working with electrolytes of high s a l t concentration to observe SERS from water on s i l v e r electrodes /666,667/ is a t t r i b u t e d to an increased density of active sites at high s a l t concentration as well as to a more complete hydrogen bond disruption in the Helmholtz layer /912/ ( i . e . a higher density of water-halide complexes). In /913/ i t is pointed out, that the absence of SER features from normal water in electrode spectra may be explained with purely electromagnetic arguments. Other Adsorbatee.

Recently, SER spectra from nitrogen and carbon dioxide on coldly

evaporated s i l v e r films deposited at 120 K and cooled down to ~ 40 K have been observed /914/. Nitrogen displays a single vibrational l i n e at 2321 cm- I close to the gas phase value in agreement with e a r l i e r results from samples deposited at I i K /133/ (on coldly evaporated copper films a band at 2282 cm-1 has been reported /783/). Four bands are observed a f t e r CO2 exposure at 653 cm- I ( v 1 ' 6 0 - C - 0 ) ' 1278/1371 cm-1 [probably Fermi resonance of 2wI with ~2 (~sO-C-O)]' and 2343 cm-1 (~3'~aO-C-O) respectively. The measured frequencies are close to EELS data from Ag(110) /661/ and the gas phase values / i / . The importance of chemical effects for SERS is revealed by a study of Raman scattering and luminescence i n t e n s i t i e s from crystal v i o l e t on smooth and roughened films of Ag and Au /915/. I t is emphasized, that Raman scattering by adsorbed molecules should be viewed as scattering by the entire adsorbate/substrate complex. This point is also evident from a SER study of meso-tetraphenylporphine on Ag in a layered structure /916/.

131

F i n a l l y , enhanced Raman scattering from crystal vibrational modes of an antimony f i l m l a i d down on a s i l v e r island substrate has been reported /917/ (enhancement factor: ~ 20). Selected Applications and Related Surface Enhanced Phenomena.

SERS has been used

( i ) to study formation of NO2 and NO3 on the pre-oxidized surface of ag powder catal y s t s a f t e r exposure to NO and NO2/N204 /918/, ( i i )

to determine the extent of

charge transfer between metal surfaces and chemisorbed molecules from s h i f t s of vibrational frequencies /919/, ( i i i )

to characterize silver-modified n-GaAs(lO0)

photoelectrodes /920/, ( i v ) to obtain structural information on adsorbed amphiphilic molecules which are used to a l t e r w e t t a b i l i t y and surface tensions at l i q u i d / s o l i d interfaces /921/, and (v) to investigate the oxide layer on Ag and Cu smoke p a r t i cles /922/. For basic fuchsin molecules placed at varying separation from a Ag island f i l m by means of a SiOx spacer layer, maximum luminescence enhancement of 200 has been observed f o r a separation of ~ 2.5 nm /923,924/. The r e s u l t is explained with the competition between local electromagnetic f i e l d enhancement and loss of e x c i t a t i o n by radiationless energy transfer to the metal. Picosecond fluorescence relaxation measurements of rhodamine 6G on s i l v e r island films are described in /925/. Enhanced infrared absorption from monolayer species on s i l v e r films in an ATR arrangement is investigated in /926/. An interesting r e - i n t e r p r e t a t i o n of early IR transmission experiments on coldly evaporated copper films exposed to CO /574/ is given in /927/. The observed Fano-type l i n e shape of the C-O stretching band has been assigned to interference of continuous e-h-pair excitations in the metal (excited via surface defects by the IR radiation) and the discrete vibrational excitations. For surfaces with a high density of defects such as coldly evaporated copper f i l m s , the i n d i r e c t e x c i t a t i o n of the vibrational mode via infrared e-h-pair excitations has been postulated to be stronger than the d i r e c t photon-vibration interaction for CO on smooth surfaces. Recent experimental results and theoretical developments concerning surface enhanced nonlinear optical processes are described in /928-936/. Besides those effects mentioned in Sect. 10.3, enhanced four-wave mixing has found some i n t e r e s t /931,933/.

132

References

i

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

G. Herzberg: Infrared and Raman Spectra of Polyatomic Molecules (Van Nostrand, New York 1945) E.B. Wilson, Jr., P.C. Decius, P.C. Cross: Molecular Vibrations (McGraw- H i l l , New York 1955) H.C. Allen, Jr., P.C. Cross: Molecular Vib-Rotors (Wiley, New York 1963) N.B. Col thup, L.H. Daly, S.E. Wiberley: Introduction to Infrared and Raman Spectroscopy (Academic Press, New York 1975) R.J.H. Clark, R,E. Hester (eds.): Advances in Infrared and Raman Spectroscopy (Heyden, London 1978) L.J. Bellamy: The Infrared Spectra of Complex Molecules (Chapman and Hall, London 1980) N.L. Alpert, W.E. Kaiser, H.A. Szymanski: IR, Theory and Practice of Infrared Spectroscopy (Plenum, New York 1970) A. Anderson (ed.): The Raman Effect (Dekker, New York 1971) S.C. Brown (ed.): Electron-Molecule Scattering (Wiley, New York 1979) K. Takayanagi, N. Oda (eds.): Proc. of the Xlth Int. Conf. on the Phys. of Electronic and Atomic Collisions (North-Holland, Amsterdam 1979) J.W. White: "Neutron Scattering Spectroscopy in Relation to Electromagnetic Methods", in Molecular Spectroscopy, ed. by P. Hepple (Institute of Petroleum, London 1972) p. 199 S.W. Lovesey, T. Springer (eds.): Dynamics of Solids and Liquids by Neutron Scattering, Topics in Current Physics, Vol. 3 (Springer, Berlin, Heidelberg, New York 1977) L.H. L i t t l e : Infrared Spectra of Adsorbed Species (Academic Press, New York 1966) M.L. Hair: Infrared Spectroscopy in Surface Chemistry (Dekker, New York 1967) H. Ibach (ed.): Electron Spectroscopy for Surface Analysis, Topics in Current Physics, Vol. 4 (Springer, Berlin, Heidelberg, New York 1977) H. Ibach, D.L. Mills: Electron Energy Loss Spectroscopy and Surface Vibrations (Academic Press, New York 1982) R.F. Willis (ed.): Vibrational Spectroscopy of Adsorbates, Springer Series in Chemical Physics, Vol. 17 (Springer, Berlin, Heidelberg, New York 1980) T. Wolfram (ed.): Inelastic Electron Tunneling Spectroscopy, Springer Series in Solid State Sciences, Vol. 4 (Springer, Berlin, Heidelberg, New York 1978) R.P. Van Duyne: "Laser Excitation of Raman Scattering from Adsorbed Molecules on Electrode Surfaces", in Chemical and Biochemical Applications of Lasers, ed. by C.B. Moore, Vol. 4 (Academic Press, New York 1979) p. 101 J.E. Demuth, K. Christmann, P.N. Sanda: Chem. Phys. Lett. 76, 201 (1980) H.W. Schr~tter, H.W. Kl~ckner: "Raman Scattering Cross Sections in Gases and Liquids", in Raman Spectroscopy of Gases and Liquids, ed, by A. Weber, Topics in Current Physics, Vol. 11 (Springer, Berlin, Heidelberg, New York 1979) p. 123 J.G. Skinner, W.G. Nilsen: J. Opt. Soc. Am. 58, 113 (1968) P.J. Hendra, E.J. Loader: Trans. Faraday Soc. 67, 82 (1971) R.G. Greenler, T.L. Slager: Spectrochim. Acta 29A, 193 (1973) I. Pockrand, A. Otto: Applic. Surf. Sci. 6, 362 (1980) 133

26 Chih-Cong Chou, C.E. Reed, J.C. Hemminger, S. Ushioda: J. Electron Spectrosc. Relat. Phenom. 29, 401 (1983) 27 A. Campion, J.K. Brown, V.M. Grizzle: Surf. Sci. 115, L 153 (1982) 28 A. Campion: J. Electron Spectrosc. Relat. Phenom. 29, 397 (1983) 29 Y. Levy, C. Imbert, J. Cipriani, S. Racine, R. Dupeyrat: Opt. Commun. 11, 66 (1974) 30 J. Cipriani, S. Racine, R. Dupeyrat, H. Hasmonay, M. Dupeyrat, Y. Levy, C. Imbert: Opt. Commun. 11, 70 (1974) 31 J. Cipriani, H. Hasmonay, Y. Levy, S. Racine, M. Dupeyrat, R. Dupeyrat, C. Imbert: Japan. J. Appl. Phys. 14, Suppl. i - 4 , 93 (1975) 32 M. Menetrier, R. Dupeyrat, Y. Levy, G. Imbert: Opt. Commun. 21, 162 (1977) 33 H. Raether: "Surface Plasma Oscillations and Their Applications", in Physics of Thin Films, ed. by G. Hass, M.H. Francombe, R.W. Hoffmann, Vol. 9 (Academic Press, New York 1977) p. 145 34 Y.J. Chen, W.P, Chen, E. Burstein: Phys. Rev. Lett. 36, 1207 (1976) 35 W.H. Weber, G.W. Ford: Opt. Lett. 6, 122 (1981) 36 K. Sakoda, K. Ohtaka, E. Hanamura: Sol. State Commun. 41, 393 (1982) 37 J.R. Kirtley, S.S. Jha, J.C. Tsang: Sol. State Commun. 35, 509 (1980) 38 S.S. Jha, J.R. Kirtley, J.C. Tsang: "Raman Intensities from Molecules Adsorbed on Rough Metal Surfaces", in Proc. Vllth Int. Conf. on Raman Spectroscopy, Ottawa, 1980; ed. by W.E. Murphy (North-Holland, Amsterdam 1980) p. 356 39 S.S. Jha, J.R. Kirtley, J.C. Tsang: Phys. Rev. B22, 3973 (1980) 40 H. Numata: J. Phys. Soc. Jap. 51, 2575 (1982) 41 N. Garcia: Opt. Commun. 45, 307 (1983) 42 I. Pockrand: J. Phys. D: Appl. Phys. 9, 2423 (1976) 43 B. Pettinger, A. Tadjeddine, D.M. Kolb: Chem. Phys. Lett. 66, 544 (1979) 44 R. Dornhaus, R.E. Benner, R.K. Chang, I. Chabay: Surf. Sci. 101, 367 (1980) 45 H.W.K. Tom, C.K. Chen, A.R.8o de Castro, Y.R. Shen: Sol. State Commun. 41, ~49 (1982) 46 A. Girlando, J.G. Gordon I I , D. Heitmann, M,R. Philpott, H. Seki, J.D. Swalen: Surf. Sci. lOl, 417 (1980) 47 A. Girlando, M.R, Philpott, D. Heitmann, J.D. Swalen, R. Santo: J. Chem. Phys. 7~, 5187 (1980) 48 P.N. Sanda, J.M. Warlaumont, J.E. Demuth, J.C. Tsang, K. Christmann, J.A. Bradley: Phys. Rev. Lett. 45, 1519 (1980) 49 S. Ushioda, Y. Sasaki: Phys. Rev. B27, 1401 (1983) 50 J.C. Tsang, J.R. Kirtley, T.N. Theis: J. Chem. Phys. 77, 641 (1982) 51 J.C. Tsang, J.R. Kirtley, J.A. Bradley: Phys. Rev. Lett. 43, 772 (1979) 52 J.C. Tsang, J.R. Kirtley, T.N. Theis: Sol. State Commun. 35, 667 (1980) 53 J.R. Kirtley, T.N. Theis, J.C. Tsang: Appl. Phys. Lett. 37, 435 (1980) 54 J.R. Kirtley, J.C. Tsang, T.N. Theis, S.S. Jha: "Surface Enhanced Raman Scattering from Tunnel Junction Structures", in Proc. Vllth Int. Conf. on Raman Spectroscopy, Ottawa, 1980; ed. by W.E. Murphy (North- Holland, Amsterdam 1980) p. 386 55 D.L. Jeanmaire, R.P. Van Duyne: J. Electroanal. Chem. 84, I (1977) 56 M.G. Albrecht, J.A. Creighton: J. Amer. Chem. Soc. 99, 5215 (1977) 57 M. Fleischmann, P.J. Hendra, A.J. McQuillan: Chem. Phys. Lett. 26, 163 (1974) 58 A.J. McQuillan, P.J. Hendra, M. Fleischmann: J. Electroanal. Chem. 65, 933 (1975) 59 P.J. Hendra, M. Fleischmann: "Raman Spectroscopy at Surfaces", in Top,as in Surfaae Chemistry, ed. by E. Kay, P.S. Bagus (Plenum, New York 1978) p. 373 60 J.A. Creighton, C.G. Blatchford, M.G. Grant: J. Chem. Soc. Faraday II 75, 790 (1979) 61 T.Ho Wood, M.V. Klein: J. Vac. Sci. Technol. 16, 459 (1979) 62 E. Burstein, C.Yo Chen, S. Lundquist: "Giant Raman Scattering by Molecules Adsorbed on Metals: An Overview", in Li#ht Scattering in Solids, ed. by J.L. Birman, H.Z. Cummins, K.K. Rebane (Plenum, New York 1979) p. 479 63 T.E. Furtak, J. Reyes: Surf. Sci. 93, 351 (1980) 64 A. Otto: Applic. Surf. Sci. 6, 309 (1980) 65 H. Ueby, S. Ichimura, H. Yamada: Surf. Sci. 119, 433 (1982) 66 H. Metiu: "Surface Enhanced Spectroscopy", Progr. Surf. Sci., to be published 134

67 A. Otto: "Surface Enhanced Raman Scattering: "Classical" and "Chemical" Origins", in Light Scattering in Solids IV, ed. by M. Cardona, G. GUntherodt, Topics in Applied Physics, Vol. 54 (Springer, Berlin, Heidelberg, New York 1984) p. 289 68 K. Arya, R. Zeyher: "Theory of Surface Enhanced Raman Scattering" in L