Mossbauer spectroscopy: applications in chemistry, biology, industry, and nanotechnology 1118057244, 9781118057247, 9781118771976

Citation preview

€ MOSSBAUER SPECTROSCOPY

€ MOSSBAUER SPECTROSCOPY APPLICATIONS IN CHEMISTRY, BIOLOGY, AND NANOTECHNOLOGY

Edited by Virender K. Sharma, Ph.D. Göstar Klingelhöfer Tetsuaki Nishida

Copyright Ó 2013 by John Wiley & Sons, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: M€ ossbauer spectroscopy : applications in chemistry, biology, industry, and nanotechnology / [edited by] Virender K. Sharma, Ph.D., G€ ostar Klingelh€ ofer, Tetsuaki Nishida. pages cm Includes bibliographical references and index. ISBN 978-1-118-05724-7 (hardback) 1. M€ ossbauer spectroscopy. I. Sharma, Virender K., editor of compilation. II. Klingelh€ ofer, G€ ostar, 1956- editor of compilation. III. Nishida, Tetsuaki, 1950- editor of compilation. QD96.M6M638 2014 2013011056 5430 .6–dc23 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

We dedicate this book to the late Professor Attila Vertez, E€otv€os Lorand University, Budapest, Hungary

Contents Preface xix Contributors xxi

Part I

Instrumentation 1

Chapter 1 | In Situ M€ ossbauer Spectroscopy with Synchrotron Radiation on Thin Films

Svetoslav Stankov, Tomasz Sle˛zak, Marcin Zaja˛c, Michał Sle˛zak, Marcel Sladecek, Ralf R€ohlsberger, Bogdan Sepiol, Gero Vogl, Nika Spiridis, Jan Ła_zewski, Krzysztof Parlinski, and Jozef Korecki

3

1.1 1.2

Introduction 3 Instrumentation 4 1.2.1 Nuclear Resonance Beamline ID18 at the ESRF 5 1.2.2 The UHV System for In Situ Nuclear Resonant Scattering Experiments at ID18 of the ESRF 6 1.3 Synchrotron Radiation-Based M€ ossbauer Techniques 10 1.3.1 Coherent Elastic Nuclear Resonant Scattering 10 1.3.2 Coherent Quasielastic Nuclear Resonant Scattering 25 1.3.3 Incoherent Inelastic Nuclear Resonant Scattering 30 1.4 Conclusions 38 Acknowledgments 39 References 39

ossbauer Spectroscopy in Studying Electronic Spin and Valence Chapter 2 | M€ States of Iron in the Earth’s Lower Mantle

43

Jung-Fu Lin, Zhu Mao, and Ercan E. Alp 2.1 2.2

Introduction 43 Synchrotron M€ ossbauer Spectroscopy at High Pressures and Temperatures 44 2.3.1 Crystal Field Theory on the 3d Electronic States 46 2.3.2 Electronic Spin Transition of Fe2þ in Ferropericlase 47 2.3.3 Spin and Valence States of Iron in Silicate Perovskite 49 2.3.4 Spin and Valence States of Iron in Silicate Postperovskite 52 2.4 Conclusions 54 Acknowledgments 55 References 55

Chapter 3 | In-Beam M€ ossbauer Spectroscopy Using a Radioisotope Beam and a Neutron Capture Reaction

58

Yoshio Kobayashi 3.1 3.2

Introduction 58 57 Mn (!57Fe) Implantation M€ ossbauer Spectroscopy 61 3.2.1 In-Beam M€ ossbauer Spectrometer 61 3.2.2 Detector for 14.4 keV M€ ossbauer g-Rays 62 3.2.3 Application to Materials Science—Ultratrace of Fe Atoms in Si and Dynamic Jumping 62 vii

viii

CONTENTS

3.2.4 Application to Inorganic Chemistry 63 3.2.5 Development of M€ ossbauer g-Ray Detector 65 3.3 Neutron In-Beam M€ ossbauer Spectroscopy 66 3.4 Summary 66 References 67

Part II

Radionuclides 71

Chapter 4 | Lanthanides (151Eu and

155

Gd) M€ ossbauer Spectroscopic Study of Defect-Fluorite Oxides Coupled with New Defect Crystal Chemistry Model

73

Akio Nakamura, Naoki Igawa, Yoshihiro Okamoto, Yukio Hinatsu, Junhu Wang, Masashi Takahashi, and Masuo Takeda 4.1 4.2 4.3

Introduction 73 Defect Crystal Chemistry (DCC) Lattice Parameter Model 76 Lns-M€ ossbauer and Lattice Parameter Data of DF Oxides 79 ossbauer and Lattice Parameter Data of M-Eus (M4þ ¼ Zr, Hf, Ce, U, 4.3.1 151Eu-M€ and Th) 79 ossbauer and Lattice Parameter Data of Zr1yGdyO2y/2 80 4.3.2 155Gd-M€ 4.4 DCC Model Lattice Parameter and Lns-M€ ossbauer Data Analysis 84 4.4.1 DCC Model Lattice Parameter Data Analysis of Ce–Eu and Th–Eu 85 4.4.2 Quantitative BL(Eu3þO)-Composition (y) Curves in Zr–Eu and Hf–Eu 88 4.4.3 Model Extension Attempt from Macroscopic Lattice Parameter Side 89 4.5 Conclusions 92 References 93

Chapter 5 | M€ ossbauer and Magnetic Study of Neptunyl(þ1) Complexes

95

Tadahiro Nakamoto, Akio Nakamura, and Masuo Takeda 5.1 5.2 5.3 5.4

Introduction 95 237 Np M€ ossbauer Spectroscopy 96 Magnetic Property of Neptunyl Monocation (NpO2þ) 97 M€ ossbauer and Magnetic Study of Neptunyl(þ1) Complexes 98 5.4.1 (NH4)[NpO2(O2CH)2] (1) 98 5.4.2 [NpO2(O2CCH2OH)(H2O)] (2) 100 5.4.3 [NpO2(O2CH)(H2O)] (3) 101 5.4.4 [(NpO2)2((O2C)2C6H4)(H2O)3]H2O (4) 104 5.5 Discussion 106 ossbauer Relaxation Spectra 106 5.5.1 237 Np M€ 5.5.2 Magnetic Susceptibility and Saturation Moment: Averaged Powder Magnetization for the Ground jJz ¼ 4i Doublet 107 5.6 Conclusion 113 Acknowledgment 113 References 113

Chapter 6 | M€ ossbauer Spectroscopy of

161

Dy in Dysprosium Dicarboxylates

Masashi Takahashi, Clive I. Wynter, Barbara R. Hillery, Virender K. Sharma, Duncan Quarless, Leopold May, Toshiyuki Misu, Sabrina G. Sobel, Masuo Takeda, and Edward Brown 6.1 Introduction 116 6.2 Experimental Methods 117 6.3 Results and Discussion 117 Acknowledgment 122 References 122

116

ix

CONTENTS

Chapter 7 | Study of Exotic Uranium Compounds Using

238

U M€ ossbauer

Spectroscopy

123

Satoshi Tsutsui and Masami Nakada 7.1 7.2

Introduction 123 Determination of Nuclear g-Factor in the Excited State of 238U Nuclei 125 ossbauer Spectroscopy and Its Application 7.2.1 Background of 238 U M€ to Magnetism in Uranium Compounds 125 ossbauer and 235 U NMR Measurements of UO2 in the 7.2.2 238 U M€ Antiferromagnetic State 125 7.2.3 Determination of the Nuclear g-Factor in the First Excited State of 238 U 127 ossbauer Spectroscopy to Heavy Fermion 7.3 Application of 238U M€ Superconductors 127 7.3.1 Introduction of Uranium-Based Heavy Fermion Superconductors 127 7.3.2 Magnetic Ordering and Paramagnetic Relaxation in Heavy Fermion Superconductors 129 ossbauer Spectroscopy of Uranium-Based 7.3.3 Summary of 238 U M€ Heavy Fermion Superconductors 133 7.4 Application to Two-Dimensional (2D) Fermi Surface System of Uranium Dipnictides 134 7.4.1 Introduction of Uranium Dipnictides 134 7.4.2 Hyperfine Interactions Correlated with the Magnetic Structures in Uranium Dipnictides 135 ossbauer Spectroscopy of Uranium Dipnictides 137 7.4.3 Summary of 238 U M€ 7.5 Summary 137 Acknowledgments 138 References 138

Part III

Spin Dynamics 141

Chapter 8 | Reversible Spin-State Switching Involving a Structural Change

143

Satoru Nakashima 8.1 8.2

Introduction 143 Three Assembled Structures of Fe(NCX)2(bpa)2 (X ¼ S, Se) and Their Structural Change by Desorption of Propanol Molecules [23] 144 8.3 Occurrence of Spin-Crossover Phenomenon in Assembled Complexes Fe(NCX)2(bpa)2 (X ¼ S, Se, BH3) by Enclathrating Guest Molecules [25–27] 145 8.4 Reversible Structural Change of Host Framework of Fe(NCS)2(bpp)22 (Benzene) Triggered by Sorption of Benzene Molecules [29] 147 8.5 Reversible Spin-State Switching Involving a Structural Change of Fe(NCX)2(bpp)22(Benzene) (X ¼ Se, BH3) Triggered by Sorption of Benzene Molecules [30] 149 8.6 Conclusions 150 References 151

Chapter 9 | Spin-Crossover and Related Phenomena Coupled with Spin, Photon, and Charge Norimichi Kojima and Akira Sugahara 9.1 9.2

Introduction 152 Photoinduced Spin-Crossover Phenomena 153 9.2.1 LIESST for Fe(II) Complexes 153

152

x

CONTENTS

9.2.2 LIESST for Fe(III) Complexes 157 9.2.3 Recent Topics of Photoinduced Spin-Crossover Phenomena 160 9.3 Charge Transfer Phase Transition 161 9.3.1 Thermally Induced Charge Transfer Phase Transition 161 9.3.2 Photoinduced Charge Transfer Phase Transition 164 9.4 Spin Equilibrium and Succeeding Phenomena 168 9.4.1 Rapid Spin Equilibrium in Solid State 168 9.4.2 Concerted Phenomenon Coupled with Spin Equilibrium and Valence Fluctuation 173 References 175

Chapter 10

| Spin Crossover in Iron(III) Porphyrins Involving the Intermediate-Spin State

177

Mikio Nakamura and Masashi Takahashi 10.1 10.2

Introduction 177 Methodology to Obtain Pure Intermediate-Spin Complexes 178 10.2.1 Saddled Deformation 178 10.2.2 Ruffled Deformation 182 10.2.3 Core Modification 184 10.3 Spin Crossover Involving the Intermediate-Spin State 189 10.3.1 Spin Crossover Between S ¼ 3/2 and S ¼ 1/2 189 10.3.2 Spin Crossover Between S ¼ 3/2 and S ¼ 5/2 192 10.4 Spin-Crossover Triangle in Iron(III) Porphyrin Complexes 195 10.5 Conclusions 198 Acknowledgments 198 References 199

Chapter 11

| Tin(II) Lone Pair Stereoactivity: Influence on Structures and Properties and M€ ossbauer Spectroscopic Properties Georges Denes, Abdualhafed Muntasar, M. Cecilia Madamba, and Hocine Merazig 11.1 11.2 11.3

11.4

11.5

11.6

Introduction 202 Experimental Aspects 203 11.2.1 Sample Preparation 203 Crystal Structures 204 11.3.1 The Fluorite-Type Structure: A Typically Ionic Structure 204 11.3.2 Tin(II) Fluoride: Covalent Bonding and Polymeric Structure 205 11.3.3 The a-PbSnF4 Structure: The Unexpected Combination of Ionic Bonding and Covalent Bonding 207 11.3.4 The PbClF-Type Structure: An Ionic Structure and a Tetragonal Distortion of the Fluorite Type 207 Tin Electronic Structure and M€ ossbauer Spectroscopy 208 11.4.1 Tin Electronic Structure, Bonding Type, and Coordination 208 11.4.2 Using M€ ossbauer Spectroscopy to Probe the Tin Electronic Structure and Bonding Mode 211 Application to the Structural Determination of a-SnF2 213 11.5.1 History 213 ossbauer Spectroscopy to Determine that the Tin Positions 11.5.2 Using 119Sn M€ Used by Bergerhoff Were Incorrect 214 Application to the Structural Determination of the Highly Layered Structures of a-PbSnF4 and BaSnF4 216 11.6.1 History 216 11.6.2 Unit Cell of MSnF4 and Relationships with the Fluorite-Type MF2 217

202

xi

CONTENTS

11.6.3 M€ ossbauer Spectroscopy, Bonding Type, Crystal Symmetry, and Preferred Orientation 220 11.6.4 Combining All the Results: The a-PbSnF4 Structural Type 225 11.7 Application to the Structural Study of Disordered Phases 226 11.7.1 Disordered Fluoride Phases 226 11.7.2 Disordered Chloride Fluoride Phases 232 11.8 Lone Pair Stereoactivity and Material Properties 241 11.9 Conclusions 242 Acknowledgments 243 References 243

Part IV

Biological Applications 247

Chapter 12

| Synchrotron Radiation-Based Nuclear Resonant Scattering: Applications to Bioinorganic Chemistry

249

Yisong Guo, Yoshitaka Yoda, Xiaowei Zhang, Yuming Xiao, and Stephen P. Cramer 12.1 12.2

Introduction 249 Technical Background 250 12.2.1 Theoretical Aspects of NFS 250 12.2.2 Theoretical Aspects of SRPAC 252 12.2.3 Experimental Aspects of NFS and SRPAC 255 12.3 Applications in Bioinorganic Chemistry 258 12.3.1 Nuclear Forward Scattering 258 12.3.2 SRPAC 264 12.4 Summary and Prospects 269 Acknowledgments 269 References 269

Chapter 13

| M€ ossbauer Spectroscopy in Biological and Biomedical Research

272

Alexander A. Kamnev, Krisztina Kovacs, Irina V. Alenkina, and Michael I. Oshtrakh 13.1 Introduction 272 13.2 Microorganisms-Related Studies 273 13.3 Plants 276 13.4 Enzymes 280 13.5 Hemoglobin 281 13.6 Ferritin and Hemosiderin 283 13.7 Tissues 284 13.8 Pharmaceutical Products 286 13.9 Conclusions 286 Acknowledgments 287 References 287

Chapter 14

| Controlled Spontaneous Decay of M€ ossbauer Nuclei (Theory and Experiments)

292

Vladimir I. Vysotskii and Alla A. Kornilova 14.1 14.2

Introduction to the Problem of Controlled Spontaneous Gamma Decay 292 The Theory of Controlled Radiative Gamma Decay 293 14.2.1 General Consideration 293 14.3 Controlled Spontaneous Gamma Decay of Excited Nucleus in the System of Mutually Uncorrelated Modes of Electromagnetic Vacuum 295 14.3.1 Spontaneous Gamma Decay in the Case of Free Space 296 14.3.2 Spontaneous Gamma Decay of Excited Nuclei in the Case of Screen Presence 298

xii

CONTENTS

14.4

Spontaneous Gamma Decay in the System of Synchronized Modes of Electromagnetic Vacuum 302 14.5 Experimental Study of the Phenomenon of Controlled Gamma Decay of M€ ossbauer Nuclei 303 14.5.1 Investigation of the Phenomenon of Controlled Gamma Decay by Analysis of Deformation of M€ ossbauer Gamma Spectrum 303 14.6 Experimental Study of the Phenomenon of Controlled Gamma Decay by Investigation of Space Anisotropy and Self-Focusing of M€ ossbauer Radiation 309 14.7 Direct Experimental Observation and Study of the Process of Controlled Radioactive and Excited Nuclei Radiative Gamma Decay by the Delayed Gamma–Gamma Coincidence Method 311 14.8 Conclusions 314 References 314

Chapter 15

| Nature’s Strategy for Oxidizing Tryptophan: EPR and M€ ossbauer Characterization of the Unusual High-Valent Heme Fe Intermediates

315

Kednerlin Dornevil and Aimin Liu 15.1 Two Oxidizing Equivalents Stored at a Ferric Heme 315 15.2 Oxidation of L-Tryptophan by Heme-Based Enzymes 316 15.3 The Chemical Reaction Catalyzed by MauG 318 15.4 A High-Valent Bis-Fe(IV) Intermediate in MauG 319 15.5 A High-Valent Fe Intermediate of Tryptophan 2,3-Dioxygenase 320 15.6 Concluding Remarks 321 References 322

Chapter 16

| Iron in Neurodegeneration

Jolanta Gała˛zka-Friedman, Erika R. Bauminger, and Andrzej Friedman

324

16.1 16.2 16.3 16.4 16.5

Introduction 324 Neurodegeneration and Oxidative Stress 324 M€ ossbauer Studies of Healthy Brain Tissue 325 Properties of Ferritin and Hemosiderin Present in Healthy Brain Tissue 327 Concentration of Iron Present in Healthy and Diseased Brain Tissue: Labile Iron 328 16.6 Asymmetry of the M€ ossbauer Spectra of Healthy and Diseased Brain Tissue 330 16.7 Conclusion: The Possible Role of Iron in Neurodegeneration 331 References 331

Chapter 17

| Emission (57Co) M€ ossbauer Spectroscopy: Biology-Related Applications, Potentials, and Prospects

333

Alexander A. Kamnev 17.1 17.2 17.3 17.4

Introduction 333 Methodology 334 Microbiological Applications 336 Enzymological Applications 340 17.4.1 Choosing a Test Object 340 17.4.2 Prerequisites for Using the 57 Co EMS Technique 342 17.4.3 Experimental 57 Co EMS Studies 342 17.4.4 Two-Metal-Ion Catalysis: Competitive Metal Binding at the Active Centers 344 17.4.5 Possibilities of 57 Co Substitution for Other Cations in Metalloproteins 345 17.5 Conclusions and Outlook 345 Acknowledgments 345 References 346

xiii

CONTENTS

Part V

Iron Oxides 349

Chapter 18

| M€ ossbauer Spectroscopy in Study of Nanocrystalline Iron Oxides from Thermal Processes

351

Ji9rı Tu9cek, Libor Machala, Ji9rı Frydrych, Ji9rı Pechou9sek, and Radek Zbo9ril 18.1 18.2

Introduction 351 Polymorphs of Iron(III) Oxide, Their Crystal Structures, Magnetic Properties, and Polymorphous Phase Transformations 352 18.2.1 a-Fe2O3 353 18.2.2 b-Fe2O3 358 18.2.3 g-Fe2O3 360 18.2.4 e-Fe2O3 364 18.2.5 Amorphous Fe2O3 369 ossbauer Spectroscopy in Monitoring Solid-State 18.3 Use of 57Fe M€ Reaction Mechanisms Toward Iron Oxides 371 18.3.1 Thermal Decomposition of Ammonium Ferrocyanide—A Valence Change Mechanism 371 18.3.2 Thermal Decomposition of Prussian Blue in Air 374 18.3.3 Thermal Conversion of Fe2(SO4)3 in Air—Polymorphous Exhibition of Fe2O3 376 18.3.4 Nanocrystalline Fe2O3 Catalyst from FeC2O42H2O 376 18.4 Various M€ ossbauer Spectroscopy Techniques in Study of Applications Related to Nanocrystalline Iron Oxides 378 ossbauer Spectroscopy at Various Temperatures 378 18.4.1 57Fe Transmission M€ ossbauer Spectroscopy 379 18.4.2 In-Field 57Fe Transmission M€ ossbauer Spectroscopy 381 18.4.3 In Situ High-Temperature 57Fe Transmission M€ ossbauer Spectroscopy 383 18.4.4 57Fe Conversion Electron and Conversion X-Ray M€ 18.5 Conclusions 389 Acknowledgments 389 References 389

Chapter 19

| Transmission and Emission

57

Fe M€ ossbauer Studies on Perovskites

and Related Oxide Systems

393

Zoltan Homonnay and Zoltan Nemeth 19.1 19.2

Introduction 393 Study of High-TC Superconductors 394 19.2.1 Study of 57 Co-Doped YBa2Cu3O7d 395 19.2.2 Study of 57 Co-Doped Y1xPrxBa2Cu3O7d 397 19.3 Study of Strontium Ferrate and Its Substituted Analogues 401 19.3.1 Study of Sr0.95Ca0.05Co0.5Fe0.5O3d and Sr0.5Ca0.5Co0.5Fe0.5O3d 401 19.4 Pursuing Colossal Magnetoresistance in Doped Lanthanum Cobaltates 407 19.4.1 Emission M€ ossbauer Study of La0.8Sr0.2CoO3d Perovskites 408 19.4.2 Emission and Transmission M€ ossbauer Study of Iron-Doped La0.8Sr0.2FeyCo1yO3d Perovskites 411 References 413

Chapter 20

| Enhancing the Possibilities of

57

Fe M€ ossbauer Spectrometry to Study the Inherent Properties of Rust Layers Karen E. Garcıa, Cesar A. Barrero, Alvaro L. Morales, and Jean-Marc Greneche

20.1 20.2

Introduction 415 M€ ossbauer Characterization of Some Iron Phases Presented in the Rust Layers 416 20.2.1 Akaganeite 416

415

xiv

CONTENTS

20.2.2 Goethite 418 20.2.3 Magnetite/Maghemite 420 20.3 Determining Inherent Properties of Rust Layers by M€ ossbauer Spectrometry 421 20.3.1 Rust Layers in Steels Submitted to Total Immersion Tests 421 20.3.2 Rust Layers in Steels Submitted to Dry–Wet Cycles 424 20.3.3 Rust Layers in Steels Submitted to Outdoor Tests 426 20.4 Final Remarks 426 Acknowledgments 426 References 426

Chapter 21

| Application of M€ ossbauer Spectroscopy to Nanomagnetics

429

Lakshmi Nambakkat 21.1 21.2

Introduction 429 Spinel Ferrites 430 21.2.1 Microstructure Determination 430 21.2.2 Elucidation of Bulk Magnetic Properties in Nanoferrites Using In-Field M€ ossbauer Spectroscopy 434 21.2.3 Core–Shell Effect on the Magnetic Properties in Superparamagnetic Nanosystems 436 21.3 Nanosized Fe–Al Alloys Synthesized by High-Energy Ball Milling 441 21.3.1 Nanosized Al–1 at% Fe 442 21.4 Magnetic Thin Films/Multilayer Systems: 57Fe/AI MLS 446 21.4.1 Structural Characterization 447 21.4.2 DC Magnetization Studies 448 21.4.3 M€ ossbauer (CEMS) Study 451 21.5 Conclusions 452 Acknowledgments 453 References 453

Chapter 22

| M€ ossbauer Spectroscopy and Surface Analysis

455

Jose F. Marco, Jose Ramon Gancedo, Matteo Monti, and Juan de La Figuera 22.1 22.2

Introduction 455 The Physical Basis: How and Why Electrons Appear in M€ ossbauer Spectroscopy 456 22.3 Increasing Surface Sensitivity in Electron M€ ossbauer Spectroscopy 458 22.4 The Practical Way: Experimental Low-Energy Electron M€ ossbauer Spectroscopy 460 22.5 M€ ossbauer Surface Imaging Techniques 465 22.6 Recent Surface M€ ossbauer Studies in an “Ancient” Material: Fe3O4 466 Acknowledgment 468 References 468

Chapter 23

|

57

Fe M€ ossbauer Spectroscopy in the Investigation of the Precipitation of Iron Oxides Svetozar Music, Mira Ristic, and Stjepko Krehula 23.1 23.2 23.3 23.4

Introduction 470 Complexation of Iron Ions by Hydrolysis 470 Precipitation of Iron Oxides by Hydrolysis Reactions 472 Precipitation of Iron Oxides from Dense b-FeOOH Suspensions 480

470

xv

CONTENTS

23.5

Precipitation and Properties of Some Other Iron Oxides 483 23.5.1 Ferrihydrite 483 23.5.2 Lepidocrocite (g-FeOOH) 485 23.5.3 Magnetite (Fe3O4) and Maghemite (g-Fe2O3) 487 23.6 Influence of Cations on the Precipitation of Iron Oxides 490 23.6.1 Goethite 490 23.6.2 Hematite 495 23.6.3 Magnetite and Maghemite 496 Acknowledgment 496 References 497

Chapter 24

| Ferrates(IV, V, and VI): M€ ossbauer Spectroscopy Characterization

505

Virender K. Sharma, Yurii D. Perfiliev, Radek Zbo9ril, Libor Machala, and Clive I. Wynter 24.1 24.2 24.3

Introduction 505 Spectroscopic Characterization 506 M€ ossbauer Spectroscopy Characterization 508 24.3.1 Ferryl(IV) Ion 508 24.3.2 Ferrates(IV, V, and VI) 510 24.3.3 Case Studies 513 Acknowledgments 517 References 517

Chapter 25

| Characterization of Dilute Iron-Doped Yttrium Aluminum Garnets by M€ ossbauer Spectrometry

521

Kiyoshi Nomura and Zoltan Nemeth 25.1 Introduction 521 25.2 Sample Preparations by the Sol–Gel Method 523 25.3 X-Ray Diffraction and EXAFS Analysis 523 25.4 Magnetic Properties 525 25.5 M€ ossbauer Analysis of YAG Doped with Dilute Iron 526 25.6 Microdischarge Treatment of Iron-Doped YAG 528 25.7 Conclusions 531 Acknowledgments 532 References 532

Part VI

Industrial Applications 533

Chapter 26

| Some M€ ossbauer Studies of Fe–As-Based High-Temperature Superconductors

535

Amar Nath and Airat Khasanov 26.1 26.2 26.3 26.4

Introduction 535 Experimental Procedure 535 Where Do the Injected Electrons Go? 537 New Electron-Rich Species in Ni-Doped Single Crystals: Is It Superconducting? 538 26.5 Can O2 Play an Important Role? 539 Acknowledgment 541 References 541

Chapter 27

| M€ ossbauer Study of New Electrically Conductive Oxide Glass Tetsuaki Nishida and Shiro Kubuki 27.1

Introduction 542 27.1.1 Electrically Conductive Oxide Glass 542 27.1.2 Cathode Active Material for Lithium-Ion Battery (LIB) 543

542

xvi

CONTENTS

27.2

Structural Relaxation of Electrically Conductive Vanadate Glass 544 27.2.1 Increase in the Electrically Conductivity of Vanadate Glass 544 27.2.2 Cathode Active Material for Li-Ion Battery (LIB) 547 27.3 Summary 551 Acknowledgments 551 References 551

Chapter 28

| Applications of M€ ossbauer Spectroscopy in the Study of Lithium Battery Materials

552

Ricardo Alcantara, Pedro Lavela, Carlos Perez Vicente, and Jose L. Tirado 28.1 28.2

Introduction 552 Cathode Materials for Li-Ion Batteries 554 28.2.1 Layered Intercalation Electrodes 554 28.2.2 Phosphate Electrodes with Olivine Structure 554 28.2.3 Insertion Silicate Electrodes 555 28.3 Anode Materials for Li-Ion Batteries 556 28.3.1 Conversion Oxides 556 28.3.2 Tin Alloys and Intermetallic Compounds 558 28.3.3 Antimony Alloys and Intermetallic Compounds 560 28.4 Conclusions 561 Acknowledgments 561 References 562

Chapter 29

| M€ ossbauer Spectroscopic Investigations of Novel Bimetal Catalysts for Preferential CO Oxidation in H2

564

Wansheng Zhang, Junhu Wang, Kuo Liu, Jie Jin, and Tao Zhang 29.1 29.2

Introduction 564 Experimental Section 564 29.2.1 Catalyst Preparation 564 29.2.2 Catalytic Activity Test 565 29.2.3 M€ ossbauer Spectra Characterization 565 29.3 Results and Discussion 565 29.3.1 PtFe Alloy Nanoparticles Catalyst 565 29.3.2 Ir–Fe/SiO2 Catalyst 567 29.4 Conclusions 574 Acknowledgments 574 References 575

Chapter 30

| The Use of M€ ossbauer Spectroscopy in Coal Research: Is It Relevant or Not? Frans B. Waanders 30.1 30.2

Introduction 576 Experimental Procedures 577 30.2.1 M€ ossbauer Spectroscopy 577 30.2.2 SEM Analyses 577 30.2.3 XRD Analyses 577 30.2.4 Samples and Sample Preparation 577 30.3 Results and Discussion 578 30.3.1 M€ ossbauer Analyses of the As-Mined Samples 578 30.3.2 Weathering of Coal 578 30.3.3 Corrosion of Mild Steel Due to the Presence of Compacted Fine Coal 583 30.3.4 Coal Combustion 584 30.3.5 Coal Gasification and Resultant Products 587

576

xvii

CONTENTS

30.4 Conclusions 590 Acknowledgments 591 References 591

Part VII Chapter 31

Environmental Applications 593

| Water Purification and Characterization of Recycled Iron-Silicate Glass

595

Shiro Kubuki and Tetsuaki Nishida 31.1

Introduction 595 31.1.1 Water-Purifying Ability of Recycled Iron Silicate Glass 595 31.1.2 Iron Silicate Glass Prepared by Recycling Coal Ash 596 31.2 Properties and Structure of Recycled Silicate Glasses 596 31.2.1 Water-Purifying Ability of Recycled Silicate Glasses 596 31.2.2 Electromagnetic Property of Recycled Silicate Glasses 601 31.3 Summary 605 31.3.1 Water-Purifying Ability of Recycled Silicate Glasses 605 31.3.2 Electromagnetic Property of Recycled Silicate Glasses 606 References 606

Chapter 32

| M€ ossbauer Spectroscopy in the Study of Laterite Mineral Processing Eamonn Devlin, Michail Samouhos, and Charalabos Zografidis 32.1 Introduction 608 32.2 Conventional Processing 609 32.3 Microwave Processing 612 References 619

Index 621

608

Preface Five decades ago, the M€ossbauer concept was invented. Since then the M€ ossbauer spectroscopy has been applied in a wide range of fields including physics, chemistry, biology, and nanotechnology. The M€ ossbauer spectroscopy is still being applied vigorously in understanding the hyperfine interactions of electromagnetic nature. This is evident from a similar number of publications on the M€ ossbauer concept (
14,000/decade) in the last three decades. This book presents the current knowledge on the applications of M€ ossbauer spectroscopy. With this theme in the minds of editors, many experts were invited to contribute to the book on the use of the M€ossbauer effect in a number of subject areas. The editors also made sure that the contributors were from almost every region of the world (i.e., North America, South America, Europe, Africa, and Asia) in order to cover different aspects of the M€ ossbauer spectroscopy. In Chapters 1 and 2, an introduction is made to the synchrotron M€ ossbauer spectroscopy with examples. Examples include the in situ M€ ossbauer spectroscopy with synchrotron radiation on thin films and the study of deep-earth minerals. Investigations of in-beam M€ ossbauer spectroscopy using a 57 Mn beam at the RIKEN RIBF is presented in Chapter 3. This chapter demonstrates innovative experimental setup for online M€ ossbauer spectroscopy using the thermal neutron ossbauer spectroscopy of radionuclides is described in Chapters 4–7. Chapter 4 capture reaction, 56 Fe (n, g) 57 Fe. The M€ ossbauer structure and powder X-ray gives full description of the latest analysis results of lanthanides (151 Eu and 155 Gd) M€ diffraction (XRD) lattice parameter (a0) data of defect fluorite (DF) oxides with the new defect crystal chemistry (DCC) ossbauer and magnetic study of neptunyl(þ1) complexes, while Chapter 6 a0 model. Chapter 5 reviews the 237 Np M€ ossbauer spectrosdescribes the M€ ossbauer spectroscopy of organic complexes of europium and dysprosium. 238 U M€ copy is presented in Chapter 7. There are three chapters on spin-state switching/spin-crossover phenomena (Chapter 8– 10). Examples in these chapters are mainly on iron compounds, such as iron(III) porphyrins. The use of M€ ossbauer spectroscopy of physical properties of Sn(II) is discussed in Chapter 11. Chapters 12–17 are devoted to applications of the M€ ossbauer spectroscopy to the biological chemistry. Chapter 12 details the recent progress on the application of 57 Fe NFS, 57 Fe SRPAC, and 61 Ni SRPAC to bioinorganic chemistry. The future prospect of these techniques is also given. The role of M€ ossbauer spectroscopy in biological and biomedical research is described in Chapters 13 and 17. These chapters demonstrate how M€ ossbauer spectroscopy can be applied to study microorganisms, plants, tissues, enzymes, hemoglobin, ferritin, and hemosiderin. Chapter 15 deals with the M€ ossbauer characterization of high-valent iron intermediates in the oxidation of L-tryptophan by heme-based enzymes. Chapter 16 is focused on the use of M€ ossbauer spectroscopy to study iron in neurodegenerative diseases. Recent advances on studying iron and iron oxides using M€ ossbauer spectroscopy are described in Chapters 18–25. Chapter 18 discusses the nanocrystalline iron oxides, while Chapter 19 presents perovskite-related systems where emission M€ ossbauer spectroscopy contributes to exploring the structure and electronic or magnetic behavior of these ossbauer spectrometry to study iron phases in rust layers is described in Chapter 20. The materials. The use of 57 Fe M€ progress made on understanding bulk magnetic properties of nanosized powders of ferrites, mechanically alloyed/milled Fe–Cr–Al intermetallics, and a Fe–Al multilayer system is presented in Chapter 21. The application of surface M€ ossbauer spectroscopy to study very thin layers (a few atomic layers thick) of iron oxides is discussed in Chapter 22. Chapter 23 describes in detail the precipitation of iron oxides from aqueous iron salt solutions using M€ ossbauer spectroscopy. Chapter 24 is focused on the spectroscopic characterization of ferrates in high-valent oxidation states (þ4, þ5, and þ6). Chapter 25 deals with dilute iron-doped yttrium aluminum garnets. M€ ossbauer spectroscopy of materials of industrial interest is discussed in Chapters 26–30. Chapter 26 deals with Fe– As-based high-temperature superconductors. M€ ossbauer study of cathode active material for lithium-ion battery (LIB) and electrically conductive vanadate glass is presented in Chapter 27. More details on the applications of M€ ossbauer spectroscopy to LIB are given in Chapter 28. Chapter 29 is the example of applying M€ ossbauer spectroscopy to develop novel bimetal heterogeneous catalysts for preferential CO oxidation in H2. Chapter 30 shows the successful use of M€ ossbauer spectroscopy to identify and quantify the iron mineral phases of South African coal fractions. The last two chapters are mainly on the applications of M€ ossbauer spectroscopy to the environmental field, for example, describing the recycling process of iron-containing “waste” of silicate glasses, which is related to purification of polluted water xix

xx

PREFACE

(Chapter 31). The variables control in the laterite mineral processing using M€ ossbauer spectroscopy is another example (Chapter 32). Finally, the editors of the book would like to acknowledge contributions by late Professor Attila Vertez, E€ otv€ os Lorand University. In addition to studying fundamentals of M€ ossbauer spectroscopy, Attila applied the M€ ossbauer effect to various fields. One of the coeditors, Virender K. Sharma, met Attila in fall 2002 when he was visiting Budapest under the sustainability grant, received by Florida Tech, from the U.S. Department of States. During the visit, Attila was very kind to accept him in his group. Since then Virender had several interactions in Budapest and on one occasion in Melbourne, Florida. Because of the admiration for Attila, the M€ ossbauer community organized a special symposium titled “Chemical Applications of M€ ossbauer Spectroscopy,” honoring him at the American Chemical Society Spring Meeting at San Francisco in March 2010. In the summer 2010, Virender traveled to Budapest to present the “Salute of Excellence” from the American Chemical Society. It was heartening to see that leading chemists from Hungary, including the president of the chemistry division of the Hungarian Academy of Science and president of E€ otv€ os Lorand University, were present at that occasion. Attila will always be known as a great scientist with a gentleman touch and we will miss him dearly. This book is dedicated to late Professor Attila Vertez for his many accomplishments in M€ ossbauer spectroscopy.

VIRENDER K. SHARMA €FER €STAR KLINGELHo Go TETSUAKI NISHIDA

Contributors Ricardo Alc
antara, Laboratorio de Qu
ımica Inorg
anica, Universidad de C
ordoba, C
ordoba, Spain Irina V. Alenkina, Faculty of Physical Techniques and Devices for Quality Control, Institute of Physics and Technology, Ural Federal University, Ekaterinburg, Russian Federation Ercan E. Alp, Advanced Photon Source, Argonne National Laboratory, Argonne, IL, USA C
esar A. Barrero, Grupo de Estado S
olido, Facultad de Ciencias Exactas y Naturales, Universidad de Antioquia, Medell
ın, Colombia Erika R. Bauminger, Racah Institute of Physics, Hebrew University of Jerusalem, Jerusalem, Israel Edward Brown, Chemistry Department, Nassau Community College, Garden City, NY Stephen P. Cramer, Department of Applied Science, University of California-Davis, Davis, CA, USA Juan de la Figuera, Instituto de Qu
ımica F
ısica “Rocasolano”, CSIC, Madrid, Spain Georges D
en es, Department of Chemistry and Biochemistry, Concordia University, Montreal, Quebec, Canada Eamonn Devlin, Institute of Materials Science, N.C.S.R. “Demokritos”, Attiki, Athens, Greece Kednerlin Dornevil, Department of Chemistry, Georgia State University, Atlanta, GA, USA Andrzej Friedman, Department of Neurology, Faculty of Health Science, Medical University of Warsaw, Warsaw, Poland Ji9r
ı Frydrych, Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic Jolanta Gała˛zka-Friedman, Faculty of Physics, Warsaw University of Technology, Warsaw, Poland Jos
e Ram
on Gancedo, Instituto de Qu
ımica F
ısica “Rocasolano”, CSIC, Madrid, Spain Karen E. Garc
ıa, Grupo de Estado S
olido, Facultad de Ciencias Exactas y Naturales, Universidad de Antioquia, Medell
ın, Colombia Jean-Marc Greneche, LUNAM, Universit
e du Maine, Institut des Mol
ecules et Mat
eriaux du Mans, Le Mans Cedex, France Yisong Guo, Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA, USA Barbara R. Hillery, Chemistry Department, Nassau Community College, Garden City, NY Yukio Hinatsu, Department of Chemistry, Hokkaido University, Sapporo, Hokkaido, Japan Zolt
an Homonnay, Faculty of Science, Eotvos Lorand University, Budapest, Hungary Naoki Igawa, Quantum Bean Science Directorate, Japan Atomic Energy Agency, Tokai-mura, Naka-gun, Ibaraki, Japan Jie Jin, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China Alexander A. Kamnev, Laboratory of Biochemistry, Institute of Biochemistry and Physiology of Plants and Microorganisms, Russian Academy of Sciences, Saratov, Russian Federation Airat Khasanov, Department of Chemistry, University of North Carolina at Asheville, Asheville, NC, USA Yoshio Kobayashi, Department of Engineering Science, The University of Electro-Communications, Tokyo, Japan

xxi

xxii

CONTRIBUTORS

Norimichi Kojima, Graduate School of Arts and Sciences, The University of Tokyo, Meguro-ku, Tokyo, Japan J
ozef Korecki, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany Alla A. Kornilova, Moscow State University, Moscow, Russia Krisztina Kov
acs, Laboratory of Nuclear Chemistry, Institute of Chemistry, E€ otv€ os Lor
and University, Budapest, Hungary Stjepko Krehula, Division of Materials Chemistry, Rudjer Bo9skovi
c Institute, Zagreb, Croatia Shiro Kubuki, Department of Chemistry, Graduate School of Science and Engineering, Tokyo Metropolitan University, Hachioji, Japan Pedro Lavela, Laboratorio de Qu
ımica Inorg
anica, Universidad de C
ordoba, C
ordoba, Spain _ Jan Łazewski, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany Jung-Fu Lin, Department of Geological Sciences, Jackson School of Geosciences, The University of Texas at Austin, Austin, TX, USA Aimin Liu, Department of Chemistry, Georgia State University, Atlanta, GA, USA Kuo Liu, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China Libor Machala, Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic M. Cecilia Madamba, Department of Chemistry and Biochemistry, Concordia University, Montreal, Quebec, Canada Zhu Mao, Department of Geological Sciences, Jackson School of Geosciences, The University of Texas at Austin, Austin, TX, USA Jos
e F. Marco, Instituto de Qu
ımica F
ısica “Rocasolano”, CSIC, Madrid, Spain Leopold May, Chemistry Department, Nassau Community College, Garden City, NY Hocine Merazig, Laboratoire de Chimie Mol
eculaire, du Contr^ ole de l’Environnement et de Mesures PhysicoChimiques, D
epartement de Chimie, Facult
e des Sciences, Universit
e Mentouri, Constantine, Algeria Toshiyuki Misu, Department of Chemistry, Faculty of Science, Toho University, Funabashi, Chiba, Japan Matteo Monti, Instituto de Qu
ımica F
ısica “Rocasolano”, CSIC, Madrid, Spain Alvaro L. Morales, Grupo de Estado S
olido, Facultad de Ciencias Exactas y Naturales, Universidad de Antioquia, Medell
ın, Colombia Abdualhafed Muntasar, Department of Chemistry and Biochemistry, Concordia University, Montreal, Quebec, Canada Svetozar Musi
c, Division of Materials Chemistry, Rugjer Bo9skovi
c Institute, Zagreb, Croatia Masami Nakada, Advanced Science Research Center, Japan Atomic Energy Agency, Ibaraki, Japan Tadahiro Nakamoto, Department of Materials Characterization, Toray Research Center, Inc., Otsu, Shiga, Japan Akio Nakamura, Advanced Science Research Center, Japan Atomic Energy Agency, Tokai-mura, Naka-gun, Ibaraki, Japan Mikio Nakamura, Department of Chemistry, Faculty of Science, Toho University, Funabashi, Chiba, Japan Satoru Nakashima, Natural Science Center for Basic Research and Development, Hiroshima University, Higashi-Hiroshima, Japan Lakshmi Nambakkat, Department of Physics, University College of Science, Mohanlal Sukhadia University, Udaipur, Rajasthan, India

CONTRIBUTORS

xxiii

Amar Nath, Department of Chemistry, University of North Carolina at Asheville, Asheville, NC, USA Zolt
an N
emeth, Faculty of Science, Eotvos Lorand University, Budapest, Hungary Tetsuaki Nishida, Department of Biological and Environmental Chemistry, Faculty of Humanity-Oriented Science and Engineering, Kinki University, Iizuka, Japan Kiyoshi Nomura, School of Engineering, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Yoshihiro Okamoto, Quantum Beam Science Directorate, Japan Atomic Energy Agency, Tokai-mura, Naka-gun, Ibaraki, Japan Michael I. Oshtrakh, Faculty of Physical Techniques and Devices for Quality Control, Institute of Physics and Technology, Ural Federal University, Ekaterinburg, Russian Federation Krzysztof Parli
nski, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), EggensteinLeopoldshafen, Karlsruhe, Germany Ji9r
ı Pechou9sek, Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic Carlos P
erez Vicente, Laboratorio de Qu
ımica Inorg
anica, Universidad de C
ordoba, C
ordoba, Spain Yurii D. Perfiliev, Chemistry Department, Florida Institute of Technology, Melbourne, FL, USA Duncan Quarless, Chemistry Department, Nassau Community College, Garden City, NY Mira Risti
c, Division of Materials Chemistry, Rugjer Bo9skovi
c Institute, Zagreb, Croatia Ralf R€ ohlsberger, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), EggensteinLeopoldshafen, Karlsruhe, Germany Michail Samouhos, School of Mining and Metallurgical Engineering, National Technical University of Athens, Athens, Greece Bogdan Sepiol, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany Virender K. Sharma, Chemistry Department, Florida Institute of Technology, Melbourne, Florida, USA Marcel Sladecek, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), EggensteinLeopoldshafen, Karlsruhe, Germany
˛zak, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Michał Sle Karlsruhe, Germany
˛zak, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Tomasz Sle Karlsruhe, Germany Sabrina G. Sobel, Chemistry Department, Nassau Community College, Garden City, NY Nika Spiridis, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany Svetoslav Stankov, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), EggensteinLeopoldshafen, Karlsruhe, Germany Akira Sugahara, Graduate School of Arts and Sciences, The University of Tokyo, Meguro-ku, Tokyo, Japan Masashi Takahashi, Department of Chemistry, Faculty of Science, Toho University, Funabashi, Chiba, Japan Masuo Takeda, Department of Chemistry, Faculty of Science, Toho University, Funabashi, Chiba, Japan ordoba, C
ordoba, Spain Jos
e L. Tirado, Laboratorio de Qu
ımica Inorg
anica, Universidad de C
Satoshi Tsutsui, Research and Utilization Division, SPring-8/JASRI, Sayo-cho, Sayo-gun, Hyogo, Japan; Advanced Science Research Center, Japan Atomic Energy Agency, Ibaraki, Japan

xxiv

CONTRIBUTORS

Ji9r
ı Tu9cek, Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic Gero Vogl, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany Vladimir I. Vysotskii, Mathematics and Theoretical Radiophysics Department, Kiev National Shevchenko University, Kiev, Ukraine Frans B. Waanders, School of Chemical and Minerals Engineering, North West University, Potchefstroom, South Africa Junhu Wang, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China Clive I. Wynter, Chemistry Department, Nassau Community College, Garden City, NY Yuming Xiao, HPCAT, Advanced Photon Source, Argonne National Laboratory, Argonne, IL, USA Yoshitaka Yoda, Research and Utilization Division, SPring-8/JASRI, Kouto, Sayo, Hyogo, Japan Marcin Zaja˛c, Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany Radek Zbo9ril, Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic; Chemistry Department, Florida Institute of Technology, Melbourne, FL, USA Tao Zhang, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China Wansheng Zhang, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China Xiaowei Zhang, Photon Factory, KEK, 1-1 Oho, Tsukuba, Ibaraki, Japan Charalabos Zografidis, School of Mining and Metallurgical Engineering, National Technical University of Athens, Athens, Greece

PA R T I

INSTRUMENTATION

C H A P T E R 1

€ IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS SLE˛ZAK, MARCEL SLADECEK, SLE˛ZAK, MARCIN ZAJA˛C, MICHAŁ  SVETOSLAV STANKOV, TOMASZ  _ € HLSBERGER, BOGDAN SEPIOL, GERO VOGL, NIKA SPIRIDIS, JAN ŁAZEWSKI , RALF Ro   ZEF KORECKI KRZYSZTOF PARLINSKI, AND Jo Institute for Synchrotron Radiation, Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen, Karlsruhe, Germany

1.1 INTRODUCTION Soon after the first observation [1–3] by Rudolf M€ ossbauer in 1958 of nuclear resonant recoilless absorption and ossbauer effect became a well-established spectroscopic method for probing emission of g-rays from nuclei of 191 Ir, the M€ the electronic, magnetic, and dynamic properties of solids, liquids, and even gases [4]. The unprecedentedly high intrinsic ossbauer effect is determined by the natural linewidth of energy resolving power on the order of 10
13 offered by the M€ the nuclear resonant exited state. The relatively simple and easily accessible experimental setup, and the observation of the effect in isotopes of widely spread and technologically relevant elements such as iron and tin and their compounds stimulated the explosion of applications of the M€ ossbauer spectroscopy not only in solid state physics but also in fundamental physics [5], chemistry [6], geology [7], biology [8], industry [9], and many other fields. If a photon with an energy equal to the resonant energy impinges on the M€ ossbauer nucleus in its ground state, the photon may, with a probability given by the Lamb–M€ ossbauer factor fLM, excite the nuclear resonant level without an energy loss due to recoil. After the mean lifetime of this state, the nucleus returns back into its ground state either by emitting a photon or by ejecting an electron with probabilities 1/(1 þ a) and a/(1 þ a), respectively, where a is the coefficient of total internal conversion. For most of the M€ ossbauer isotopes, a is significantly greater that 1; therefore, the dominant mechanism for the de-excitation is internal conversion. The limited mean-free path of ejected electrons defines an escape depth of only few nanometers. By detecting the conversion electrons (conversion electron M€ ossbauer spectroscopy, CEMS), information about the electronic and magnetic properties of the materials’ surface can be derived. The high values of the cross section for nuclear resonant absorption and the small escape depths of the conversion electrons established CEMS as a standard technique for investigating surface layers of materials [10]. Moreover, by analyzing the energy of the ejected electrons depth-selective information can be retrieved [11,12]. This determined the vast range of applications of this technique to surface science and nanotechnology already soon after the first observation of the M€ ossbauer effect. The feasibility for in situ experiments on ultrathin 57 Fe films consisting of only one atomic layer of iron by the CEMS technique has been successfully demonstrated [13] in the mid-1980s. However, the relatively long data acquisition times M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

3

4

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

(15 h) needed for accumulating spectra with reasonable statistics by using a conventional radioactive source have limited further applications. A new era in the M€ ossbauer spectroscopy emerged in the year 1985 from the first observation of coherent elastic nuclear resonant scattering (NRS) of synchrotron radiation by the group of Erich Gerdau in Hamburg [14]. The demonstration that nuclear resonant experiments are indeed feasible by using synchrotron radiation instead of radioactive source was a tremendous success. However, only the advent of the third-generation synchrotron radiation sources in the middle 1990s along with the rapid development of high-resolution X-ray optics [15–19] and fast avalanche photodiode detectors (APDs) [20] established M€ ossbauer spectroscopy with synchrotron radiation as a standard technique for probing electronic, magnetic, and dynamic properties of materials [21]. The enormous brilliance of the Xray beams provided by insertion devices such as wigglers and undulators, which is by more than 10 orders of magnitude larger in comparison with that of the conventional radioactive sources, allowed for the investigation of samples containing very small quantities of the resonant isotope. This resulted in a significant expansion of applications to layered systems (thin- and ultrathin films, and multilayers), nanostructures (islands, clusters), and samples under extreme conditions, for example, very high pressures and temperatures [22]. The possibility to finely tune the energy of the photons using high-resolution monochromators (HRMs) has resulted in a new technique for direct determination of the phonon density of states of the resonant element—the nuclear inelastic scattering (NIS) [23–25]. Thus, in the same experimental setup, one is able to probe simultaneously hyperfine interactions and lattice dynamics of the sample. Investigation of well-defined nanostructures often requires that the preparation, characterization, and maintenance of the samples during experiments are performed under controlled, in most cases ultrahigh vacuum (UHV), conditions. Corresponding instrumentation [26,27] has been constructed and permanently installed at the nuclear resonance beamline ID18 [28] of the European Synchrotron Radiation Facility (ESRF) in Grenoble, France. This opened up new perspectives for in situ investigations of electronic and magnetic properties, vibrational dynamics, and diffusion phenomena by several nuclear resonant scattering-based techniques such as coherent elastic nuclear resonant scattering, coherent quasielastic nuclear resonant scattering, and incoherent inelastic nuclear resonant scattering/absorption. The aim of this chapter is to report on recent advances in the in situ M€ ossbauer spectroscopy with synchrotron radiation on thin films that became possible due to the instrumentation developments at the nuclear resonance beamline ID18 of the ESRF. After a detailed description of the beamline and of the UHV system for in situ experiments, a brief introduction into the basic NRS techniques is given. Finally, the application of these techniques to investigate magnetic, diffusion, and lattice dynamics phenomena in ultrathin epitaxial 57 Fe films deposited on a W(110) substrate is presented and discussed.

1.2 INSTRUMENTATION The Classical M€ ossbauer spectroscopy with a standard radioactive source has been performed in the energy domain, whereas the nuclear resonant scattering using synchrotron radiation measures the time-domain spectra. This implies fundamental differences in the experimental approach and the associated instrumentation. The Synchrotron radiation at third-generation sources is produced by insertion devices (wigglers or undulators) installed in the straight sections of the storage ring, where relativistic electrons (positrons) circulate the ring [29]. One of the prerequisites for performing NRS experiments is the timing mode of the storage ring operation. For nuclear resonant scattering applications, the storage ring is filled with equidistant electron (positron) bunches that produce intensive X-ray pulses having a width of about 100 ps. Especially suitable for NRS applications at the ESRF is the 16-bunch mode providing a time spacing of 176 ns between the bunches. The second condition that has to be fulfilled is the utilization of fast detectors and associated timing electronics. Avalanche photodiodes have been successfully introduced [20] for NRS applications due to their quick time response, high dynamic range, and relatively high quantum efficiency. The time resolution ranges from 0.1 to 1.0 ns and the efficiency from several percent to about 50% depending on the energy of the X-rays. The timing electronics is based on standard NIM modules, including constant fraction discriminators, gate generators, fast ADCs (analog digital converters), and MCAs (multichannel analyzers). A reference timing signal from the radio frequency system of the storage ring provides a synchronization of the electronics with the photon pulse arrival time. To further improve the performance of the APD detectors for nuclear resonant scattering applications, the highdegree suppression of the nonresonant radiation is essential. This is achieved by using high-resolution monochromators

5

1.2 INSTRUMENTATION

€ ssbauer Isotopes with the Natural Lifetime and Energy of the Nuclear Resonant Level, and Energy TABLE 1.1 Mo Resolution Provided by the Corresponding High-Resolution Monochromator at ID18 of the ESRF Isotope 57

Fe Fe 57 Fe 57 Fe 61 Ni 119 Sn 121 Sb 125 Te 129 I 149 Sm 151 Eu 161 Dy 57

Natural Lifetime (ns)

Resonant Energy (keV)

Energy Resolution (meV)

141.11 141.11 141.11 141.11 7.60 25.61 4.99 2.14 24.24 10.27 13.99 4.53

14.413 14.413 14.413 14.413 67.41 23.879 37.1298 35.493 27.890 22.502 21.542 25.651

6.4 3.1 1.8 0.50 120 0.65 1.2 1.1 1.0 0.76 1.6 0.9

based on asymmetrically cut Si single crystals with high-order reflections [19,30]. The role of the HRM in nuclear resonant scattering techniques is twofold: (1) it filters out a significant part of the nonresonant photons, thus preventing the detector in forward direction from overload during the intense X-ray pulses; and (2) the HRM provides the necessary instrumental energy resolution in the range of few millielectron volts to submillielectron volts for the purposes of the nuclear inelastic scattering experiments. A separate HRM has to be developed for each M€ ossbauer transition in order to provide the highest energy resolution, flux, and/or degree of polarization. Table 1.1 summarizes the isotopes that are currently accessible at ID18 of the ESRF for nuclear resonant scattering experiments with the natural lifetime and energy of the nuclear resonant level and the energy resolution of the corresponding high-resolution monochromator. 1.2.1 Nuclear Resonance Beamline ID18 at the ESRF A layout of the beamline for nuclear resonant scattering experiments ID18 of the ESRF in Grenoble is schematically shown in Fig. 1.1. It consists of a front end, two optical and three experimental hutches with adjacent control cabins. The front end accommodates three revolver-type undulators that allow one to switch between magnetic structures with 20 or 27 mm period. While the former provides photons with energy 14.413 keV on its fundamental, the latter is optimized for M€ ossbauer isotopes with lower/higher energies (Table 1.1). The first optics hutch (OH1) accommodates the primary slit system that shapes the synchrotron beam and defines its size, the high heat load monochromator (HHLM), further collimating optics, and beam intensity monitors (not shown in the figure). The HHLM [31] is a fixed exit double-crystal

FIGURE 1.1 A layout of the nuclear resonance beamline ID18 of the European Synchrotron Radiation Facility in Grenoble, France. u27 and u20 stand for the revolver-type undulators with 27 mm period for one structure and 20 mm period for the other. OH1 and OH2 are the first and second optics hutches devoted to the accommodation of HHLM, slit systems (S1, S2), CRL, HRMs, and FM. EH1, EH2, and EH3 stand for the experimental hutches, while CC1–CC3 denote the control cabins. DIFF is the six-circle diffractometer, UHV: the ultrahigh-vacuum system for in situ experiments, and CRYO: the cryomagnet system. X-ray beam monitors and attenuators are not shown.

6

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

Si(111) monochromator with a bandwidth of few electron volts. A cryogenic cooling allows for handling the heat load and ensures short- and long-term stability of the crystals positions. The second optics hutch (OH2) hosts various exchangeable high-resolution monochromators. For the resonant energy transition of 14.413 keV in 57 Fe, several possibilities exist offering various energy resolution and beam flux. The most suitable one has to be selected according to the needs of the particular experiment. A bending focusing monochromator (FM) for reducing the horizontal beam size to 100 mm is also available in OH2. A standard six-circle diffractometer for diffraction experiments in horizontal and vertical scattering geometries at ambient conditions is operational in the first experimental hutch (EH1). In addition, it can accommodate a furnace (300–1200 K) and continuous flow cryostat (10–300 K) allowing for NRS experiments in a large temperature range. The second experimental hutch (EH2) is devoted to experiments on samples at extreme conditions. It accommodates the UHV system [26] devoted to in situ experiments on thin films and nanostructures that is described in detail in the next section. A cryomagnet system offers external magnetic fields up to 8 T in horizontal direction, either perpendicular or parallel to the X-ray beam and a temperature range between 1.5 and 297 K. It is mounted on a two-circle heavy-duty goniometer used for sample tilting to an angular range of 5 both parallel and perpendicular to the direction of propagation of the X-ray beam. The setup is particularly suitable for experiments on surfaces, thin films, and multilayers. The sample holder is capable of carrying up to six samples. The third experimental hutch (EH3) offers enough space for user equipment, for example, portable ultrahighvacuum chambers. It is also used for setting up a high-resolution backscattering monochromator [32]. This type of HRM is used for M€ ossbauer isotopes with resonant energies above ca. 30 keV, that is, 125 Te (35.493 keV) and 121 Sb (37.1298 keV) [33,34]. An interlock system allows for carrying out an experiment in any of the experimental hutches while enabling free access to the others. The beam inside these hutches is either transported in shielded vacuum tubes or blocked by the upstream beam shutter. By this means, the experiment in one hutch can be performed in parallel with the preparation work in others. Computers and electronics to run the experiment are located in three separated control cabins (CC1–CC3) adjacent to the corresponding experimental hutches. Depending on the X-ray energy used in the experiments, compound refractive lenses (CRL), made of either Be, plastic, or Al [35], are employed for collimating the beam in order to match the beam divergence to the angular acceptance of the monochromators. In addition, CRLs serve to focus the beam to approximately 15  250 mm2 for the needs of particular experiments. These comprise, for example, high-pressure applications by using diamond anvil cells, or investigations of nanostructured samples deposited on a substrates where grazing incidence geometry has to be utilized. Alternatively, focusing of the X-ray beam down to about 5  10 mm2 is achieved by bending multilayer mirrors arranged in Kirkpatrick–Baez geometry. 1.2.2 The UHV System for In Situ Nuclear Resonant Scattering Experiments at ID18 of the ESRF The UHV instrument is permanently installed in the second experimental hutch of ID18 of the ESRF, which requires remote control of the entire system. Figure 1.2 presents schematically the instrument, while Fig. 1.3 displays a top-view photograph. The UHV setup with a base pressure of 1  10
10 mbar consists of a central distribution chamber equipped with a sample transfer mechanism and peripherally attached chambers as described in detail below. The preparation chamber, number 1 in Figs. 1.2 and 1.3, is equipped with two electron beam evaporators and an effusion cell. A four-pocket mini electron beam evaporator serves for deposition of metals from rods and crucibles from each pocket separately, as well as for codeposition of different combinations of the evaporants. A single-pocket e-beam source is used for deposition of the M€ ossbauer isotope 57 Fe. An effusion cell is available for evaporation of rare-earth  metals. A precise calibration of the deposition rate with a thickness reproducibility of 1 A is done by a quartz-balance monitor. The deposited structures can be characterized by low-energy electron diffraction (LEED) and Auger electron spectroscopy (AES). In this chamber, the samples can be cooled down to about 90 K and heated up to 2300 K by a multifunctional manipulator. The chamber for NRS experiments, number 2 in Figs. 1.2 and 1.3, is mounted on a two-circle goniometer that serves to align the sample and perform experiments with the focused synchrotron beam in grazing incidence scattering geometry. The manipulator provides contacts for temperature measurement, resistive and electron-bombardment heating, and feedthrough for cooling the sample down to 90 K by flow of liquid nitrogen. In addition, the sample holder can be rotated in the range of 180 degrees about an axis perpendicular to the sample surface, allowing for angular resolved studies. For the purpose of nuclear inelastic scattering experiments, an avalanche photodiode detector is brought close to the sample via a tube reaching into the chamber with a Be window at its end, as shown on the

7

1.2 INSTRUMENTATION

FIGURE 1.2 A schematic view of the UHV system with the preparation 1, NRS 2, distribution 3, storage 4, and load-lock 5 chambers (for simplicity all bypasses and evaporation sources are omitted). Number 6 shows the CF63 port where the portable chambers can be connected to the system. (Reproduced from Ref. 26 with permission of the American Institute of Physics.)

photograph in Fig. 1.4. The figure also shows the entrance and exit Be windows for the focused synchrotron radiation beam that impinges on the sample under grazing angles of few milliradians. It further shows the sample mounted on the sample station, and the Z-stage for the APD detector used for nuclear inelastic scattering experiments as well as the APDs for detecting the nuclear forward scattered radiation. The distribution chamber, number 3 in Figs. 1.2 and 1.3, connects the above-described chambers and serves for transferring the sample holders between them. In addition, a sample storage chamber, number 4 in Figs. 1.2 and 1.3, with

FIGURE 1.3 A top-side photograph of the UHV system with the preparation 1, NRS 2, distribution 3, storage 4, and load-lock 5 chambers. Chamber 6 shows the portable chamber for XRD, GISAXS, and XPCS experiments connected to the system. (Reproduced from Ref. 26 with permission of the American Institute of Physics.)

8

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.4 A site photograph of the sample station in the NRS chamber 2. Shown are the detectors (APD) for the nuclear forward scattering and the nuclear inelastic scattering experiments, sample, entrance and exit Be windows for the grazing incidence scattering geometry with the focused X-ray beam. During the nuclear inelastic scattering experiment, the distance between the APD (above the sample) and the sample is reduced to about 2 mm.

the capacity to store up to six sample holders, is an integrated part of the distribution chamber. The lowest position provides electrical contacts for resistive heating and temperature control allowing for annealing of the sample holders. The load-lock chamber, number 5 in Figs. 1.2 and 1.3, serves for introducing the sample holder into the UHV system. Two additional portable chambers are available for transferring samples to other beamlines of the ESRF. The portable chamber for NRS experiments (Fig. 1.5) was constructed in order to extend the in situ NRS experiments to thin films and nanostructures of noniron isotopes. (Formerly the end-station ID22N of the ESRF was optimized for the

FIGURE 1.5 A schematic view of the portable UHV chamber for NRS experiments. (Reproduced from Ref. 26 with permission of the American Institute of Physics.)

9

1.2 INSTRUMENTATION

FIGURE 1.6 A schematic view of the portable UHV chamber for wide-angle XRD, GISAXS, and XPCS experiments. (Reproduced from Ref. 26 with permission of the American Institute of Physics.)

nuclear resonances of 149 Sm, 151 Eu, and 161 Dy). This chamber is based on two crossed CF100 tubes with a sample station. The station provides electrical contacts to the sample holder for temperature control (K- or C-type thermocouples), resistive or electron-bombardment heating, and feedthroughs for sample cooling down to 90 K by flow of liquid nitrogen. The sample holder can be rotated in the range of 180 degrees around an axis perpendicular to the surface normal. This chamber is pumped by a double-flanged 100 l s
1 ion getter pump ending with a viewport. Entrance and exit windows for the X-ray beam are made out of a 100 mm thick Be foil diffusion bonded to CF63 flanges. Additional CF40 flanges are used for a viewport and a vacuum gauge. Similarly to the main NRS chamber, a CF63 tube with welded Be window on the bottom is mounted above the sample station on a stage with a linear transfer in order to reduce the distance between the APD detector and the sample. This chamber was employed to study magnetic properties and lattice dynamics of the extremely reactive Sm and Eu metallic films at the end-station ID22N. A portable chamber for wide-angle XRD, GISAXS, and XPCS experiments, shown in Fig. 1.6, is constructed to transfer the already prepared and characterized samples to other beamlines at the ESRF in order to apply these techniques in situ. The chamber is based on CF100 tubes including three CF40 flanges for a vacuum gauge, a viewport, and a simple evaporation source, allowing one to perform in situ deposition. Two diametrically mounted entrance and exit Xray Be windows with thickness of 200 mm are specially polished in order to minimize the small-angle X-ray scattering from the windows, which is an important issue for the XPCS spectroscopy. A Be dome for wide-angle X-ray diffraction (XRD) can be mounted on the top CF100 flange. By a linear manipulator, the sample holder is vertically transferred between the positions for small- and wide-angle X-ray scattering. Resistive heating and temperature control of the sample are provided via electrical feedthroughs on the manipulator. The chamber is pumped by a 75 l s
1 ion getter pump. By using this chamber, the morphology of ultrathin Fe films on MgO (001) [36,37] was systematically investigated employing GISAXS at beamline ID10A of the ESRF. The entire UHV setup is mounted on a support table with horizontal (2  102 steps mm
1) and vertical (2  105 steps mm
1) motorized movements to allow for precise adjustment of the sample to the synchrotron beam. The chamber for in-situ NRS experiments at grazing incidence geometry and for X-ray reflectivity measurements is mounted on a two-circle goniometer (1  105 steps deg
1 for an angular range of 5 ). The synchrotron beam can be conditioned to a sample spot size as small as 15  250 mm2 by compound refractive lenses or by 5  10 mm2 by Kirkpatrick–Baez focusing mirrors.

10

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES In this section, an introduction into the most common synchrotron radiation M€ ossbauer techniques is given emphasizing on the applications to thin films. A detailed elaboration of the methods can be found elsewhere [21,22,38]. In the following, atoms of the M€ ossbauer isotopes bound in a crystal lattice with single-line resonances (vanishing hyperfine interactions) are considered. If a photon with an energy matching that of the nuclear resonant transition impinges on a nucleus, the following absorption processes could take place: (i) the photon is resonantly absorbed by the nucleus without energy and momentum exchange with the crystal lattice, that is, elastic absorption. The probability for an elastic absorption is given by the Lamb–M€ ossbauer factor fLM. (ii) The photon is resonantly absorbed by the nucleus involving energy and momentum transfers with the crystal lattice, that is, inelastic absorption. The probability for this process is defined as 1
fLM. After the mean lifetime of the excited state the nucleus returns into the ground state. When a resonant photon is emitted (without energy and momentum exchange with the crystal lattice) and the system returns into a state that is indistinguishable from the state before the resonant absorption, this process is coherent scattering. The emission of conversion electrons or the fluorescent radiation leads to incoherent scattering since the system is moved back into a state that differs from the initial state. Each of these processes is considered separately below. Selected examples of relevant in situ experiments on thin films illustrating the capabilities of the techniques are presented and discussed. 1.3.1 Coherent Elastic Nuclear Resonant Scattering 1.3.1.1 Brief Theoretical Background In the case of coherent elastic scattering, a macroscopic ensemble of scattering atoms can be replaced by a continuous medium characterized by an index of refraction n [39]: n¼1þ

2p X ri M i ; k20 i

(1.1)

where Mi is the coherent forward scattering length of the ith atom and ri is the number density of atoms in the material. The sum runs over all atoms within the scattering volume. k0 is the wave number of the radiation. The wave amplitude in depth z of a slab of material is given by AðzÞ ¼ expðik 0 nzÞA0 :

(1.2)

For anisotropic media, the index of refraction depends on the polarization state of light and is represented by a 2  2 matrix [40]. In thatP case, the propagation of light in forward direction is described by the propagation matrix: F ¼ k0 þ f , with f ¼ ð2p=k0 Þ i ri Mi being the forward scattering matrix. Then (1.2) can be written in more general form: AðzÞ ¼ expðiFzÞA0 :

(1.3)

From (1.2) and (1.3) directly follows the relation between the index of refraction n and the forward scattering matrix f: n ¼ 1 þ f =k0 . The propagation matrix F describes the modification of the wave field A upon propagation from coordinate z to coordinate z þ dz. It is a multidimensional matrix with dimension given by the number of the open scattering channels. This formalism is successfully applied not only to describe the propagation of light in forward direction but also in case of X-ray diffraction from single crystals, and reflection from surfaces, thin films, and multilayers. While the structural arrangement of the scattering centers determines the dimension and the symmetries of F that are of importance for calculating the matrix exponential eiFz in (1.3), the interaction of the photons with the atoms is given by the atomic scattering length M. In order to account for its energy dependence and the polarization mixing M(v) is described by 2  2 matrix: MðvÞ ¼ EðvÞ þ NðvÞ. E(v) represents the nonresonant contribution of the electronic charge scattering, and N(v) contains the contributions from the resonant nuclear scattering processes. The electronic scattering length is then given by ½E ðvÞmn

  k0 ¼ ðem  en Þ
Zr 0 þ i s t ðvÞ ; 4p

(1.4)

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

11

where en and em are the polarization vectors, Z is the atomic number, r0 is the classical electron radius, and s t is the total absorption cross section. For the case of 2L-pole resonance, the resonant scattering length is given by [41] ½N ðvÞmn ¼

L   4pf R X ½en  Y LM ðk0 Þ Y LM ðk0 Þ  em F LM ðvÞ; k 0 M¼
L

(1.5)

fR < 1 describes the degree of elasticity of the scattering process, expressed by the Debye–Waller factor in the fastrelaxation limit and by the Lamb–M€ ossbauer factor in the slow-relaxation limit. The two dot products between the polarization basis vectors (en, em) and the vector spherical harmonics YLM(k0) describe the anisotropy of photon absorption and reemission, respectively. The energy dependence of the scattering process is contained in the functions FLM(v). These functions are the energy-dependent resonant strengths for transitions with a change of M in the magnetic quantum number. In the general case they are given by [41] F LM ðvÞ ¼

X a;h

pa pa ðhÞGx ðaMh; LÞ ; ½E ðhÞ
E ðaÞ
hv
iGðhÞ=2

(1.6)

where the sum runs over all initial (ground) states labeled by a and the intermediate excited states labeled by h. pa is the probability that the initial state a is occupied and pa(h) is the probability that the state h is initially unoccupied. E(a) and E (h) are the energies of the ground-state and excited-state levels, respectively. G x(aMh; L) is the partial resonance width of the transition between a and h with a change of M in the magnetic quantum number, G (h) denotes the full resonance width. Since G (h) 1–10 eV for electronic resonances, the scattering takes place on a timescale h=GðhÞ of 10
16 s that is essentially prompt. In contrast, the width of nuclear resonances is in the range of 10
6–10
12 eV; therefore, the scattering proceeds on comparatively long timescales. This is essential for separating both contributions one from another and for establishing the nuclear resonant scattering of synchrotron radiation as a time-domain spectroscopy. The expression (1.5) for the resonant scattering length can be expanded in powers of the unit vector m that defines the magnetic quantization axis of the atom in the sample. The resonant scattering length for an electric dipole transition (L ¼ 1) gets the form [42] ½N ðvÞmn ¼

 3  ðem  en Þ½F þ1 þ F
1 
iðem  en Þ  m½F þ1
F
1  þ ðem  mÞðen  mÞ½2F 0
F þ1
F
1  : 16p

(1.7)

For convenience, the subscript L is omitted. For the case of a magnetic dipole transition the role of the electric and magnetic fields of the radiation are interchanged and polarization vectors in Eq. (1.7) have to be transformed according to the rule e ! e  k0 , where k0 is a unit vector of the photon wave vector. The three terms in Eq. (1.7) represent different polarization dependences. The first term is not sensitive to the sample magnetization. Its polarization dependence given by ðem  en Þ is that of nonresonant charge scattering. The second term describes circular dichroism because it depends on the difference between the resonant scattering amplitudes Fþ1 and F
1. Since its polarization dependence is ðem  en Þ, it describes orthogonal scattering, for example, s ! p and p ! s. The third term is proportional to 2F 0
F þ1
F
1 and describes linear magnetic dichroism. Its polarization dependence allows for all scattering processes within the given polarization basis. For a linear polarization basis, as it is frequently used in case of scattering experiments with synchrotron radiation, the matrix elements can be explicitly written as i 3 h F þ1 þ F
1 þ ðp  mÞ2 ð2F 0
F þ1
F
1 Þ ; ½N ss ¼ 16p 3 ½
iðk0  mÞðF þ1 þ F
1 Þ
ðs  mÞðp  mÞð2F 0
F þ1
F
1 Þ; ½N sp ¼ 16p (1.8) 3 ½N ps ¼ ½iðk 0  mÞðF þ1 þ F
1 Þ
ðs  mÞðp  mÞð2F 0
F þ1
F
1 Þ; 16p i 3 h F þ1 þ F
1 þ ðs  mÞ2 ð2F 0
F þ1
F
1 Þ : ½N pp ¼ 16p These matrix elements express the strong polarization-mixing effects that are observed in resonant scattering from magnetized samples. The occurrence of optical activity crucially depends on the orientation of m relative to the incident

12

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.7 Experimental geometry used in nuclear resonant scattering of synchrotron radiation from thin films and surfaces. Q and f define the relative orientation of the incident wave vector k0 to the direction of a unidirectional magnetization m of the sample. s and p are the linear polarization basis vectors, and w is the incidence angle. (Reproduced from Ref. 22 with permission of Springer.)

wave vector and its polarization state. The off-diagonal elements describe the orthogonal scattering that turns incident s-polarization into p- polarization and vice versa. Equation (1.8) is often used to describe the polarization effects in nuclear resonant scattering experiments with synchrotron radiation. Figure 1.7 sketches the grazing incidence scattering geometry used for studies of thin films and surfaces. The set of polarization vectors (s, p), the polar and azimuthal angles (Q, f) describing the relative orientation of the wave vector k0 of the incident photons to the direction of the magnetization vector m of the sample are indicated. 1.3.1.2 Time Spectra of the Nuclear Resonant Scattering In the following discussion, the case of the nuclear resonant scattering from the 14.413 keV resonance of 57 Fe is considered. Due to its large absorption cross section, large Lamb–M€ ossbauer factor, and its relevance in many fields of natural sciences, this resonance is one of the most widely applied M€ ossbauer transitions. It is a magnetic dipole transition with spins Ig ¼ 1=2, Ie ¼ 3=2, magnetic moments mg ¼ 0:091mN , me ¼
0:153mN of the ground and excited states, respectively, and a natural lifetime t 0 ¼ 141 ns. In magnetic materials the spin-polarized 3d electrons create a spin polarization of the s-electrons via the exchange interaction that leads to a strong magnetic field at the nucleus with a magnitude of 33.3 T in the case of ferromagnetic a-Fe. This magnetic field lifts the degeneracy of ground and excited states, resulting in a Zeeman splitting of the nuclear energy levels. According to the dipole selection rule M ¼ me
mg ¼ 0; 1, six allowed energy transitions, corresponding to six separated resonances, are observed, as depicted in Fig. 1.8a. In the case of a pure magnetic hyperfine interaction, the energetic positions of the resonances are given by 

 mg me B ¼ E0
me ge
mg gg B;
mg E ¼ E 0
me Ie Ig

(1.9)

with mg and me being the magnetic quantum numbers and gg and ge the g-factors of the ground and excited states, respectively. In nonmagnetic materials the degeneracy of the ground and excited states could, in general, be partly lifted as a result of an electric field gradient (EFG) acting on the nucleus. The sources of the EFG are the distributed electric charges around the M€ ossbauer nuclei. These could be either charges of the neighboring ions or ligands surrounding the resonant atom (lattice/ligand contribution), or charges in partially filled valence orbitals of the M€ ossbauer atom (valence electron contribution). The EFG is a second-rank symmetrical tensor and in a coordinate system related to its principal axes V XX , V YY , and V ZZ , selected so that jV XX j  jV YY j  jV ZZ j. It is fully determined by the z-component V ZZ and by the asymmetry parameter h ¼ ðV XX
V YY Þ=V ZZ . From the definition of the asymmetry parameter, and the fact that the Laplace equation holds, that is, V XX þ V YY þ V ZZ ¼ 0 since the sources of the EFG are fully external for the nucleus, for h applies: 0  h  1. Assuming a purely electric quadrupole interaction the energy levels of the excited state splits to two levels,

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

13

FIGURE 1.8 Energy dependence of the functions FM in the case of nuclear resonant scattering from a magnetic dipole resonance between nuclear spins Ig ¼ 1/2 and Ie ¼ 3/2. (a) Pure magnetic hyperfine interaction. The energetic positions apply to the case of a-57 Fe with a magnetic hyperfine field of 33.3 T. The six dipole-allowed transitions decompose into three different polarization dependencies: the functions F
1, F0, and Fþ1 describe the scattering of right-circular, linear, and left-circular polarization, respectively. (b) Pure electric hyperfine interaction. (Reproduced from Ref. 22 with permission of Springer.)

resulting in two resonances (Fig. 1.8b) with energy positions given by eQV ZZ E ¼ E0  2

rffiffiffiffiffiffiffiffiffiffiffiffiffi h2 1þ ; 3

(1.10)

with Q being the quadrupole moment of the nucleus. For the particular case of nuclear resonant scattering, the functions FM defined by Eq. (1.6) with L ¼ 1 are given by the following expression [22,43]: F M ðvÞ ¼ K

X mi

 C 2 I g 1I e ; mi M ; hðv
vmi M Þ þ iG0 =2

(1.11)

where K is defined as K¼

2pf LM G0

; k 0 ð1 þ aÞ 2I g þ 1

(1.12)

with G0 being the natural linewidth of the transition. The sum runs over all ground-state levels with magnetic quantum numbers mi.  hv mi M is the resonant energy of the transition with the quantum numbers mi and M. The Clebsch–Gordan coefficients C2 Ig 1Ie ; mi M describe the relative strength of the transitions. Since the nuclear resonant scattering is a coherent elastic process it is impossible to identify the scattering atom in the sample. Instead, for each individual resonant nucleus there is a small probability that this nucleus is excited. The summation of all these small amplitudes gives the total probability amplitude for a photon to interact resonantly with the nuclei. If the incident radiation pulse is short compared to the nuclear lifetime t0, these probability amplitudes exhibit the same temporal phase. As a result, a collectively excited state is created, where a single excitation is coherently distributed over the resonant atoms of the sample [44]. The wave function of this collectively excited state is given by a coherent

14

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

superposition of states: 1 X ik0 ri e jCðk 0 Þi ¼ pffiffiffiffi jgijei i; N i

(1.13)

where jgijei i denotes the state in which the ith resonant atom at the position ri is in its excited state jei i. This collectively excited state is named nuclear exciton and exhibits remarkable optical properties resulting from the coherent superposition of states. Below some important implications of the nuclear exciton concept on the time spectrum of nuclear resonant scattering is given. A detailed elaboration of the formalism can be found elsewhere [42,45]. In a nuclear resonant scattering experiment all resonant levels of the M€ ossbauer nuclei in the sample are simultaneously excited by a short pulse of synchrotron radiation, creating the nuclear exciton. The time dependence of the delayed intensity emitted upon de-excitation of the nuclear exciton in forward direction is the time spectrum of nuclear forward scattering (NFS). In the absence of magnetic and electric fields in the sample, the M€ ossbauer spectrum consists of a single resonant line. The NFS time spectrum in this case is determined by an exponential decay of the nuclear exciton consisting of only one resonant frequency. In a semilogarithmic scale, the decay is a straight line with a slope defined by the lifetime of the excited state. Figure 1.9 illustrates a comparison between the M€ ossbauer transmission spectrum (a), NFS spectra in energy (b) and in time (c) domain, calculated for 0.2 mm thick stainless steel absorber (thin approximation holds, that is, self-absorption and multiple scattering effects are neglected) [46]. If hyperfine fields are acting on the nucleus, the degeneracy of the nuclear levels is lifted, leading to a splitting of the nuclear transition into six (magnetic hyperfine interaction, Fig. 1.8a) or two (electric hyperfine interaction, Fig. 1.8b) resonances. The superposition of wave’s amplitudes with frequency differences corresponding to the various resonant transitions leads to oscillations of the intensity in the temporal evolution of the exciton decay. These oscillations are referred to as quantum beats that contain information about the magnitude and orientation of the hyperfine fields. Figure 1.10 illustrates a comparison between the M€ ossbauer transmission spectrum (a), NFS spectra in energy (b) and in time (c) domain, calculated for the case of electric quadrupole interaction in 0.2 mm thick stainless steel absorber [46]. The time spectrum of NFS consists of quantum beats that result from coherent superposition of waves corresponding to the two possible resonant transitions. The quantum beat period is inversely proportional to the energy splitting of the resonant level, implying that small energy differences result in large quantum beat periods. A modulation of the delayed intensity additional to the quantum beats is very often encountered in the time spectra. Contrary to the quantum beats these oscillations, referred to as dynamical beats [46,47], are aperiodic and their periods increase with increasing time after excitation and decrease with increasing thickness of the sample. Figure 1.9d–f illustrates the influence of the sample thickness on the transmission M€ ossbauer spectrum, NFS energy spectrum, and NFS time spectrum, respectively. These spectra are calculated for a single-line stainless steel absorber with a thickness of 3 mm [46]. From the simulations it is obvious that in comparison to the thin absorber (Fig. 1.9a–c), an increase of the sample thickness leads to significant changes in the energy and time-domain NFS spectra, while the change in transmission M€ ossbauer spectrum is a moderate broadening of the absorption line due to resonant self-absorption [48].

FIGURE 1.9 € ssbauer transmisCalculated Mo sion spectra (a), nuclear forward scattering spectra in energy (b) and in time (c) domain for the case of single resonance in a 0.2 mm thick stainless steel foil 100% enriched in 57 Fe. (d), (e), and (f) are the corresponding spectra for a 3.0 mm thick stainless steel foil 100% enriched in 57 Fe. (Reproduced from Ref. 46 with permission of Kluwer Academic Publishers.)

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

15

FIGURE 1.10 € ssbauer transmisCalculated Mo sion spectra (a), nuclear forward scattering spectra in energy (b) and in time (c) domain for the case of a quadrupole doublet in a 0.2 mm thick stainless steel foil 100% enriched in 57 Fe. (d), (e), and (f) are the corresponding spectra for a 3.0 mm thick stainless steel foil 100% enriched in 57 Fe. (Reproduced from Ref. 46 with permission of Kluwer Academic Publishers.)

The NFS spectrum in energy domain exhibits a double-hump profile that results from a strong attenuation of the propagating wave through a thick absorber. At resonance, the incident wave and forward scattered wave exhibit about the same amplitudes and opposite phases leading to destructive interferences and minima in the transmitted intensity. As a result the double-hump profile around the resonance develops (Fig. 1.10e) [46]. The interferences between the corresponding propagating waves form the dynamical beat pattern in the NFS spectrum in time domain (Fig. 1.9f). In general, the time dependence of the nuclear forward scattering comprises both quantum and dynamical beats, as depicted in Fig. 1.10f for the case of electric quadrupole interaction. If the resonances are well separated, the time development of the amplitude of nuclear forward scattering from a sample of thickness d can be expressed as [22] AðtÞ ¼ dðtÞ


X i

pffiffiffiffiffiffi J 12 g t g i expðiE i t=h
t=2t0 Þ pffiffiffiffiffiffi i ; git

with g i ¼

k0 d f 0;i ; t0

(1.14)

where t0 is the natural lifetime of the excited nucleus, d is the thickness of the sample. J1 is the first-order Bessel function that describes the dynamical beats, while expðiEi t=h
t=2t 0 Þ stands for the quantum beats. The interplay between dynamical and quantum beats leads often to more complex beat structures referred to as hybrid beat. A detailed treatment of the hybrid beats is given by van B€urck et al. [49]. The time spectrum gets more complex when more than two resonant lines are present. This is demonstrated in Fig. 1.11a–c for magnetic dipole interaction in polycrystalline a-Fe foils [22]. In this case, all six transitions (Fig. 1.8a) contribute with different weights. While the quantum beats of the thin foil still exhibit a regular pattern, the beat pattern of the thick foil lacks any periodicity. This is a result of the superposition of quantum and dynamical beats from each of the different resonances forming a rather complex pattern. Nevertheless, this can be used as a “fingerprint” for a polycrystalline magnetic sample. An additional complexity arises if the hyperfine fields are not single valued but distributed over magnitude or direction. In many cases these distributions influence the time spectra in a characteristic way that allows one to identify them. A detailed analysis of the quantum beat patterns that arise from distributed magnetic fields is given by Shvyd’ko et al. [50]. Another phenomenon often observed in the NFS time spectra that follows from the collective coherent nuclear resonant excitation is the speedup of the initial decay compared to that of an isolated nucleus [51,52]. The time evolution of the radiative decay for an isolated nucleus is described by I ðtÞ ¼ I 0 expð
G0 t=hÞ;

(1.15)

with G0 ¼ Gg þ Ga being the natural linewidth, Gg the partial width for radiative decay, and Ga ¼ aGg the partial width for internal conversion decay. For the case of the nuclear exciton given in (1.13), the radiative decay width is significantly

16

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.11 Time spectra of nuclear forward scattering from a-Fe foil with various thicknesses. The polycrystalline foils are in magnetic state where all six transitions are allowed. (Reproduced from Ref. 22 with permission of Springer.)

increased by the effect of spatial coherence: Gg ¼ Gc þ G0g ;

(1.16)

where G0g is the partial width for spatially incoherent decay and Gc is the coherent decay width. In the absence of hyperfine splitting of the nuclear levels, Gc is given by [42]

Gcoh ¼

Z

f LM 2I e þ 1 Gg 2 2I g þ 1



2  dV 1
k 0  e 0 jSðk
k0 Þj2 ; ^

with

Gcoh 4pN

and

^

Gc ¼

Sðk
k0 Þ ¼

N X i¼1

(1.17)

expð
iðk
k0 Þ  r i Þ;

(1.18)



2  describes the where Sðk
k0 Þ is the structure factor of the ensemble of nuclei, and the factor 1
k 0  e 0 ^

^

polarization dependence. The integration over the solid angle V strongly depends on the dimensionality and the shape of the crystal. For a three-dimensional sample in forward scattering geometry, the following expression for the coherent enhancement holds: Gc ¼

rl2 d Gcoh ; 4p

(1.19)

where r is the density of resonant nuclei. Assuming l ¼ 0.1 nm and r ¼ 10 nm
3 results in a doubling of the decay rate already for sample thicknesses in the range of 10 nm. The linear dependence of Gc on the sample thickness has been experimentally demonstrated on 57 Fe foils [53]. The upper limit of the coherent enhancement is defined by photoabsorption in the sample and amounts to about Gc 1000Gg . The radiative decay, accounting for the speedup effect, is then given by I ðtÞ ¼ I 0 expð
ð1 þ xÞt=t0 Þ ; 1 with x ¼ rs 0 f LM d 4

(1.20)

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

17

FIGURE 1.12 Time evolution of nuclear forward scattering from iron foils (95% enriched in 57 Fe) of different thicknesses in a vertical magnetic field. Only the two Dm ¼ 0 transitions are excited. The solid lines are computations based on (1.15), while the dashed lines indicate the exponential decay of the envelope as calculated using (1.19). (Reproduced from Ref. 47 with permission of Kluwer Academic Publishers.)

where s 0 is the maximal absorption cross section at resonance for the given M€ ossbauer isotope. The speedup of the decay of nuclear exciton is illustrated in Fig. 1.12 [47,53]. The NFS time spectra for polycrystalline a-57 Fe foils, placed in a vertical magnetic field (only Dm ¼ 0 transitions are excited), are obtained for various foil thicknesses. The solid lines indicate the fits of the experimental data (open circles) based on (1.15), while the dashed lines are the calculated exponential decays taking into account the speedup effect by (1.19). The speedup of the nuclear decay at early times after excitation with increasing the foil thickness is clearly visible. Very large values of the speedup effect are obtained for thin films at grazing incidence scattering geometry [54–56], as it is demonstrated below. Another important factor that influences the shape of the time spectra of the nuclear forward scattering is the polarization of the incoming X-rays. Unlike the conventional M€ ossbauer source, the synchrotron radiation produced by an insertion device is completely polarized in nature with the electric field vector oscillating in the storage ring plane, that is, s-polarized, as depicted in Fig. 1.7. In case of the magnetic dipole hyperfine interactions (Fig. 1.8a), this polarized character might significantly simplify the measured spectrum and help for resolving sophisticated magnetic structures. Figure 1.13 schematically shows the corresponding angular emission characteristics for the transitions with M ¼ 0 and M ¼ 1 as polar diagrams. A thin film magnetized perpendicularly to the incident wave vector k0 is considered. The direction of the quantization axis given by m determines the orientation of the polar diagrams. For a magnetic (electric) dipole transition, the excitation probability is given by the value of the polar diagram along the direction of the magnetic field vector B (electric field vector E) of the incident wave. Thus, in case of the 14.413 keV magnetic dipole transition of 57 Fe, the M ¼ 0 transitions cannot be excited, if the magnetization m is in the plane of the sample (Fig. 1.12). This plays a decisive role on the shape of the time spectra that are recorded. The analysis of a given scattering problem becomes particularly simple in the frame of eigen-polarizations of the system that are obtained by diagonalization of the forward

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

18

FIGURE 1.13 Orientation of the vector spherical harmonics for a magnetic dipole transition in the case of an in-plane magnetization M. The magnetic field B of the incident electromagnetic wave is parallel to the surface normal. The M ¼ 0 transitions cannot be excited because there is no amplitude of the corresponding spherical harmonics along the direction of B. (Reproduced from Ref. 22 with permission of Springer.)

scattering matrix f. In a number of cases, a system of orthogonal eigen-polarizations for f can be found so that the following representation holds: f ðvÞ ¼ g f D ðvÞg
1 ;

(1.21)

where fD is a diagonal matrix, and the diagonalizing matrix g does not depend on v. Those are the cases where, for example, k0 and m are perpendicular or parallel to each other. Figure 1.14 illustrates the scattering matrices and typical time spectra for the particularly important and often encountered experimental geometries. The time spectra were calculated assuming a 2 nm thick 57 Fe film deposited on W substrate [22]. k0 and m are Perpendicular (k0  m ¼ 0). 3 fD ¼ 16p



F 1 þ F
1 0

0 2F 0



and

g¼



cosQ
sinQ

sinQ : cosQ

(1.22)

In this case, the linear polarizations are eigen-polarizations of the system. The scattering matrix f (v) is diagonal for w ¼ 0, Q ¼ 0 and w ¼ 0, Q ¼ p/2 (modulo p), which corresponds to the geometries sketched in Fig. 1.14b and c, respectively. k0 and m are Parallel or Antiparallel (k0  m ¼ 1). 3 fD ¼ 16p



F 1 0

0 F 1



and

g¼



1 i ; i 1

(1.23)

where the upper signs are for parallel orientation, the lower signs for antiparallel orientation. Here the circular polarizations are the eigen-polarizations of the system. The geometries shown in Fig. 1.14a and d lead to different scattering matrices because the off-diagonal elements in Eq. (1.8) change their sign if the relative orientation between k0 and m changes from parallel to antiparallel. Polycrystalline Directional Distributions of Magnetic Fields. Of particular relevance are situations where the directions of magnetic moments are distributed uniformly in space or over certain subspaces that are generated by magnetic anisotropies. In case of thin films, for example, the large magnetic shape anisotropy forces the magnetic moments to be aligned in the plane of the film. In case of unmagnetized multidomain films one often encounters a two-dimensional random distribution of magnetic moments. In these cases, the scattering amplitude can be calculated analytically by integration over the corresponding angle variables. The results are summarized in Fig. 1.14e–g. In all these cases, the scattering matrix N is diagonal in a linear polarization basis [57].

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

19

FIGURE 1.14 Time spectra of nuclear resonant scattering for selected orientations of the magnetic hyperfine field m relative to the incident wave vector k0. The scattering matrix is given in a linear polarization basis (s, p). The time spectra are calculated for a 2.0 nm thick, ferromagnetic a-57 Fe film on a tungsten substrate, assuming purely s-polarized incident radiation and unpolarized detection. Parts (a–c) display the results for a unidirectional magnetization of the sample. Part (d) results from the superposition of two magnetic sublattices in antiparallel alignment as it is the case in antiferromagnets, for example. Parts (e and f) display results for two-dimensional random distributions of spin directions. Part (g) shows the result for a 3D polycrystalline distribution. Since the scattering matrix in these cases is diagonal, only the diagonal elements are shown. The timescale in the right column covers one natural lifetime of the 14.4 keV transition, that is, 141 ns. The envelope of the time spectra (dashed line in the upper right figure) indicates that the time response of the film is considerably speeded up compared to the natural decay (solid straight line). (Reproduced from Ref. 22 with permission of Springer.)

It is obvious that the beat pattern in the spectra characteristically reflects the underlying spin structure. However, there are degenerate cases (b, d, f) in Fig. 1.14 where the time spectra are identical for different spin structures. For that reason, measuring a single time spectrum is not sufficient to determine the spin structure unambiguously. Instead, to avoid such degeneracies, a set of time spectra has to be taken at different sample orientations, that is, for various azimuthal angles w. In some cases the nuclear level splitting is determined by a pure electric hyperfine interaction. This holds, for example, for Fe in metallorganic complexes such as ferrocene [4], biomolecules [58,59], or for Fe monolayers (MLs) on surfaces such as W(110) [60]. The nuclear level scheme in such cases is shown in Fig. 1.8b together with the energy dependence of the functions FM. The nuclear scattering length matrix N(v) follows from the evaluation of Eq. (1.8)) with m being replaced by the unit vector VZZ that gives the direction of the principal axis of the electric field gradient tensor. The scattering matrices for different geometries together with the corresponding time spectra in a typical scattering experiment are shown in Fig. 1.15. The energy dependence of the functions F0, Fþ1, and F
1 for this case is shown in

20

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.15 Time spectra of nuclear resonant scattering for selected orientations of the principal axis VZZ of the electric field gradient relative to the incident wave vector k0. Calculations were performed for the same sample as in Fig. 1.14. The beat period in the time spectra corresponds to a quadrupole splitting as observed in the ferrocene molecule [4]. (Reproduced from Ref. 22 with permission of Springer.)

Fig. 1.8. Since Fþ1 ¼ F
1, it follows that there is no circular dichroism that is observed in the case of a magnetic hyperfine interaction. However, optical activity can occur due to linear dichroism, as illustrated in Fig. 1.14d. The following computer codes have been developed for simulating and fitting the nuclear resonant scattering spectra in both time and energy domain: CONUSS [61], MOTIF [62], EFFINO [63], SYNFOS [64], and REFTIM [65]. Since each of these codes has certain advantages and limitations in comparison to the others, one has to carefully select the suitable code for the given scientific task. The programs are freely available upon request to the authors. 1.3.1.3 Nuclear Resonant Scattering in Grazing Incidence Geometry The optical properties of a material are significantly modified when the energy of the X-rays approaches the nuclear resonance. This effect is even more pronounced at grazing incidence scattering geometry, applied to investigate nanosystems (thin and ultrathin films, multilyers, and nanostructures) [66]. An important figure of merit in nuclear resonant scattering experiments at grazing angles is the time-integrated signal that is reflected from the surface [67]. In that case, the nuclear delayed signal is large at angles where the electronic reflectivity varies rapidly with angle. Therefore, in case of a reflection from surface, a maximum in the delayed time-integrated signal is observed at the critical angle of total reflection. This is illustrated in Fig. 1.16a where the electronic and nuclear reflectivity of the surface of 57 Fe is shown around the critical angle of 3.8 mrad [22]. Figure 1.16b shows the situation for a 30 nm thick film of 57 Fe on Si substrate. Maxima in the delayed reflectivity occur at positions with the steepest negative slope in the electronic reflectivity. This phenomenon results from the formation of standing waves in the sample [22] and might lead to a significant enhancement of the radiative decay [56]. As mentioned in the beginning, the formalism that describes the wave field propagation in forward direction could, in general, be applied for nuclear reflection from ultrathin films. In that case, however, the thickness d of the scattering

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

21

FIGURE 1.16 Angular dependence of the delayed resonant reflectivity from surfaces and thin films (solid lines). Calculations have been performed (a) for the native surface of 57 Fe and (b) for a 30 nm thick film of 57 Fe on Si. The dotted lines are the corresponding electronic reflectivity curves. (c) Measured reflectivity from a 30 nm thick film of 57 Fe on Si. The top curve shows the electronic reflectivity, the bottom curve displays the delayed signal counted within a time window reaching from 12 to 170 ns after excitation. Solid lines are simulations. (Reproduced from Ref. 22 with permission of Springer.)

media has to be replaced with the effective thickness ~d given by d~ ¼ d j1 þ Rj2 =sinw;

(1.24)

where R is the reflectivity of the substrate. The term 1=sinw ffi 1=w describes the increase in path length for radiation that travels under an angle w through a film of thickness d. j1 þ Rj2 is the relative intensity of the standing wave that results from the superposition of incident and reflected waves. A detailed elaboration of the delayed time response from thin films can be found elsewhere [22]; here only some conclusions relevant to the nuclear reflection from thin films are given. (i) From (1.24), it directly follows that already for very thin layers the time response of the scattering exhibits a significant  speedup of the decay of nuclear exciton. For example, a monolayer of 57 Fe on W (d ¼ 2 A) illuminated at w ¼ 4 mrad corresponds to an effective thickness of ~d ¼ 0:2 mm, leading to a speedup of x ¼ 1.5. (ii) The maximum of the delayed intensity appears at the critical angle of the substrate. At this angle, the interference between incident and reflected wave forms a standing wave with an antinode exactly at the boundary. A similar effect has been observed for nuclear resonant scattering from thick Fe layers [67]. (iii) The highest scattering amplitude from the resonant film is achieved for the substrate with the lowest photoabsorption cross-section. This is illustrated in Fig. 1.17 where the time-integrated delayed intensity is plotted for one monolayer 57 Fe on different substrate materials [22]. The high penetration depth of the X-rays combined with the large values of the nuclear resonant absorption cross section and the isotopic selectivity of the M€ ossbauer effect define the nuclear resonant scattering with synchrotron radiation as a perfect tool for investigating the depth profile of the hyperfine interactions in thin films and multilayers. Two different approaches have been suggested and applied for obtaining depth-dependent information. The first method (Fig. 1.18a) relies on controlling the penetration depth by adjusting the angle of incidence of the X-ray beam. The penetration depth of X-rays varies strongly around the critical angle and decreases to few nanometers in the regime of total reflection. Figure 1.19 plots the electronic reflectivity (a) and the penetration depth (b) for grazing incidence

22

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.17 Angular dependence of delayed nuclear resonant scattering from 0.3 nm 57 Fe on various substrates: the intensity maximum is found at the critical angle of the substrate material. (Reproduced from Ref. 22 with permission of Springer.)

reflection from surfaces of Si, Pd, Fe2Cr2Ni at 14.4 keV. Above the critical angle, the radiation deeply penetrates into the material and the reflectivity drops according to w
4, with w being the angle of incidence. The radiation penetrates into the material even below the critical angle that leads to the existence of an evanescent wave in the film. The amplitude of this wave is exponentially damped as function of the film depth. The penetration depth of the evanescent wave is only a few nanometers, and strongly depends on the photoabsorption coefficient of the material, as depicted in Fig. 1.19b. This behavior suggests that by recording time spectra of nuclear scattering as a function of increasing grazing angle, one is able to probe hyperfine interactions in surface layers of increasing thickness.

FIGURE 1.18 Controlling the spatial origin of the reflected signal (a) via the penetration depth of the radiation or (b) via introducing of an isotopic probe layer in the region of interest within the sample. (Reproduced from Ref. 22 with permission of Springer.)

FIGURE 1.19 Calculated reflectivity (a) and penetration depth (b) for grazing incidence reflection from surfaces of Si, Pd, Fe2Cr2Ni (solid lines) at 14.4 keV. The dashed lines stand for Fe2Cr2Ni near the resonance at DE/G0 ¼ 2. (Reproduced from Ref. 22 with permission of Springer.)

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

23

FIGURE 1.20 Time spectra of grazing incidence nuclear resonant scattering from a 57 Fe film that was annealed in air at 170  C for 4 h. The angle of incidence is plotted on the right. (Reproduced from Ref. 68 with permission of Kluwer Academic Publishers.)

An example of such depth-sensitive hyperfine spectroscopy is shown in Fig. 1.20. Time spectra of nuclear resonant scattering at various grazing angles were recorded from an originally 20 nm thick 57 Fe film that was oxidized for 4 h at 170  C in air [68]. For the bottom three spectra, the penetration depth is comparable to the total thickness of the film. The fast beats are characteristic for a-Fe and a magnetic iron oxide, most likely Fe3O4. For the upper three spectra, the penetration depth was about 2–3 nm, corresponding to about 10 atomic layers. No magnetic interaction is seen in these spectra. The upper layer could probably be assigned to paramagnetic b-FeOOH and to superparamagnetic a-Fe2O3. The clear difference of the time spectra of nuclear scattering for the same sample at different grazing angles demonstrates the high sensitivity of this technique. Due to the short data collection times at third-generation synchrotron radiation sources, such investigations can be performed while external parameters are varied, that is, in situ. The second approach is based on deposition of ultrathin resonant probe layers in the regions of the films to be investigated (Fig. 1.18b). This method is used since a long time to study the depth dependence of magnetic properties in thin films by conversion electron M€ ossbauer spectroscopy [13,69–72]. Its application, however, was severely limited by the very low count rates obtained from such small amount of resonant material even by using strong radioactive sources. The enormous brilliance of the modern synchrotron radiation sources helped to overcome this difficulty and has opened new avenues in this field. It has been shown that the sensitivity of this method is in the range of a single monolayer [73] or even below [74]. Thus, hyperfine spectroscopy with the ultimate depth resolution of one atomic layer can be achieved. To reach this resolution, however, advanced sample preparation techniques such as molecular beam epitaxy have to be employed. This approach benefited significantly from the instrumentation development at the nuclear resonance beamline ID18 of the ESRF described in Section 1.2.2. An example is shown in Fig. 1.21. In order to get a precise depthselective information about the hyperfine interactions in Fe, time spectra of nuclear resonant scattering were recorded upon three epitaxial 56 Fe films on W(110) having a probe 57 Fe monolayer positioned at various depths. The following samples were investigated: 20 ML 56 Fe/1 ML 57 Fe referred to as a “surface” (S), 20 ML 56 Fe/1 ML 57 Fe/1 ML 56 Fe referred to as a “subsurface” (S
1), and 20 ML 56 Fe/3 ML 57 Fe/3 ML 56 Fe referred to as a “deep” (D) layers [75,76]. The time spectra of nuclear resonant scattering from samples (S
1) and (D) show a single-frequency quantum beat pattern described by a well-defined hyperfine magnetic field Bhf oriented parallel to the beam, that is, along the ½110 direction. This is consistent with the ½001 to ½1 10 switching of the easy magnetization axis observed in Fe/W(110) films with  thicknesses below 100 A [77]. In contrast, the time spectrum of sample (S) shows a modulation of the beat structure, which results from a finite electric field gradient caused by the broken translation symmetry at the surface, in agreement with the conversion electron M€ ossbauer experiment [13]. The possibility to perform nuclear resonant scattering experiments in situ has allowed for observation of a thicknessdriven spin reorientation transition in thin epitaxial Fe film on W(110) [78–80]. The time spectra of NRS were collected during continuous evaporation of 57 Fe on a freshly cleaned W(110) single-crystalline substrate oriented to 3.9 mrad with

24

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.21 Time spectra of nuclear resonant scattering from the Fe(110) surface (S), subsurface (S
1), and deep (D) layers measured  with  the X-ray beam along 110 direction. (Reproduced from Ref. 75 with permission of the American Physical Society.)

respect to the X-ray beam, where the maximum of the delayed intensity is found [81]. For the deposition of the first monolayer, the substrate was kept at 500 K to ensure a good nucleation [82], while for the rest of the film, with a final thickness of about 6 nm, the W(110) crystal was kept at 300 K. Figure 1.22 shows selected time spectra that were obtained with the wave vector of the X-ray beam being parallel to the ½110 direction. All spectra were consistently analyzed with CONUSS software [61]. The time spectrum of Fig. 1.22a reveals a regular beat pattern corresponding to a uniform magnetization of the Fe film along ½1 10 direction with the characteristic for bulk Fe value of the hyperfine field (33.5 T). At film thicknesses above  57 A the magnetization is already parallel to the ½001 direction (Fig. 1.22f). The spin reorientation transition is not abrupt;

FIGURE 1.22 The time spectra accumulated during continuous 57 Fe evaporation labeled with the corresponding Fe thicknesses are shown. The top and bottom spectra reflect, according to the theoretical fits (continuous line), homogenous magnetization states with the easy axis along   and ½001, respectively. 110 (Reproduced from Ref. 78 with permission of the American Physical Society.)

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

25

FIGURE 1.23 The magnetization structure during the thickness-induced spin reorientation transition for the Fe/W(110) system, as derived from the fits to the experimental data shown in Fig. 1.22 and by using a five-sublayer model. The sublayer ~
M ~ . (Reproduced from Ref. 78 with permission of the American Physical magnetization vectors are labeled as M 1 5 Society.) 

rather it extends over a large thickness range of about 6 A, corresponding to three monolayers. A consistent analysis of the time spectra shown in Fig. 1.22b–e supports the following model for thickness-driven spin reorientation transition. The film was divided into five sublayers of equal thicknesses and individual in-plane magnetization vectors oriented at an angle w with respect to the ½1 10 direction. Figure 1.23 shows schematically the orientation of the individual sublayer magnetizations derived from the analysis of the time spectra, for various film thicknesses. The onset of the transition was  noticed for a thickness of 51.6 A. The spin reorientation transition from ½110 to ½001 is initiated at the deepest layers, which switch first, while the magnetization of the remaining sublayers forms a fan-like structure. With increasing thickness, the magnetization of the subsequent sublayers rotates, and finally the transition is completed by the topmost surface layers. 1.3.2 Coherent Quasielastic Nuclear Resonant Scattering The discussions presented in the previous chapter assumed a static spatial atomic arrangement in the sample. In the general case, however, the time dependence of the atomic positions has to be taken into account. The atomic motions that are taking place at timescale comparable with the mean lifetime of the nuclear excited state may significantly influence the nuclear resonant scattering signal, and therefore that signal can be used to trace such dynamic processes. Below a brief review of the formalism that allows one to investigate atomic diffusion by nuclear resonant scattering experiments is presented. Extensive discussion and detailed elaboration of the theory of nuclear resonant scattering in the presence of diffusion can be found elsewhere [83–88]. The experimentally measured quantity in a scattering experiment is the dynamic structure factor S(q, v), describing the probability for the photon to acquire a momentum transfer q ¼ k1
k0 and an energy transfer v ¼ v0
v1 . In 1954, Van Hove demonstrated that the dynamic structure factor, known as the scattering function, relates to the probability G (r, t) for finding any particle at position r and time t when the particle was at r ¼ 0 and t ¼ 0 before by a spatial and temporal Fourier transforms [89]: Z 1 Sðq; vÞexpð
iðq  r
vtÞÞdq dv; (1.25) Gðr; tÞ ¼ ð2pÞ3 G(r, t) stands for the Van Hove correlation function. This relation underlies all modern scattering experiments. In a nuclear resonant scattering, the measured intermediate scattering function S(q, t) is defined as a single Fourier transform of the correlation function in space: Sðq; tÞ ¼

Z

Gðr; tÞexpð
iq  r Þdr:

(1.26)

In kinematic approximation, S(q, t) is determined from the measured nuclear resonant scattering intensity I(q, t): Iðq; tÞ ¼ I 0 ðq; tÞjSðq; tÞj2 ;

(1.27)

26

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

where I0 ðq; tÞ ¼ I0 L2 =4t0 expð
t=t 0 Þ, I0 being the scattering intensity under static conditions, and L is the effective thickness (dimensionless quantity defined as a product of the number of scattering atoms in the sample, the nuclear absorption cross section and the Lamb–M€ ossbauer factor) in case of a single-line scatterer. Assuming Markovian diffusion on a Bravais single-crystalline lattice, the intermediate scattering function is given by

t  Sðq; tÞ ¼ exp
aðqÞ ; t

(1.28)

where in the simplest case of jumps to nearest-neighbor (NN) sites only aðqÞ ¼

N 1X ð1
expð
iq  li ÞÞ; N i¼1

(1.29)

li are the jump vectors to N possible sites. a(q) is often called the “jump function” and could be considered as the structure factor of an ensemble of neighbor sites reached by the jumping atoms during the interaction time with the probing radiation. t is the residence time of the atoms on the lattice site and relates to the diffusion coefficient D via the Einstein–Smoluchovski equation:  2 r ; (1.30) D¼ 6t  2 r being the mean-square atomic displacement during the time t. The nuclear resonant scattered intensity in forward direction is obtained by inserting Eq. (1.28) into Eq. (1.27):  2 L2 t  t  I ðq; tÞ ¼ I 0 exp
(1.31) exp
aðqÞ  : 4t0 t0 t Taking into account the Heisenberg uncertainty principle Eq. (1.31) transforms into the expression I ðq; tÞ ¼ I 0

 L2 tðG0 þ Gd ðqÞÞ exp
; 4t0 h

(1.32)

with G0 ¼ h=t0 and Gd ðqÞ ¼ ð2h=tÞaðqÞ: From (1.32) it is obvious that in the presence of diffusion an effective broadening of the nuclear resonant line via Gd ðqÞ takes place leading to an accelerated decay of the nuclear exciton. The intuitive picture is that due to the diffusive motions of resonant atoms the spatial coherence between the scattered wave trains is partially destroyed, leading to increasing destructive interferences at latter times after the excitation. Figure 1.24 shows time spectra of nuclear forward scattering from a single crystal of FeAl measured at the indicated temperatures [85,86,90]. With increasing temperature one clearly observes an accelerated decay of the nuclear exciton. The maximal values of the diffusion coefficient D determined by this method are limited by the accessibility of early times after the excitation within which the fast decay can be observed. Assuming start time for the time spectra at 20 ns sets an upper limit for D < 10
12 m2 s
1 . By exploiting the orientation dependence of the nuclear resonant scattering signal, the diffusion mechanism can be elucidated at the atomic level. The orientation-dependent time-integrated intensity (ODIN) is given by

I ODIN ðqÞ /

Z1 0

Iðq; tÞdt /

1 : G0 þ Gd ðqÞ

(1.33)

An example for the case of FeAl is shown in Fig. 1.25. In a stoichiometric composition FeAl crystallizes in the B2 structure, as shown in the figure. An intriguing question is how exactly the diffusion of Fe atoms proceeds. Is it by a direct jump to another Fe site, despite of the relatively large distance (note that the nearest neighbors of the Fe atoms are only Al atoms) or via a short occupation of an Al site? In the latter case several independent jump sites have to be taken into account.

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

27

FIGURE 1.24 Time spectra of nuclear resonant scattering from a FeAl single crystal aligned with the ½110 direction along the wave vector of the photons for the indicated temperatures. (Reproduced from Ref. 90 with permission of Kluwer Academic Publishers.)

Then Eq. (1.33) takes the form "

2h X I ODIN ðqÞ / G0 þ v j a j ð qÞ t j

#
1

;

(1.34)

where vj denotes the weights of the individual jump paths and aj ðqÞ marks the corresponding jump functions as given by (1.29). Each solid point from the angular dependencies shown in Fig. 1.25 represents an integrated intensity of the time spectra recorded between 2 and 165 ns from FeAl kept at 1030  C, which is about 200  C below the melting point of the alloy. The angular dependence of the experimental data is very well pronounced. The solid lines mark the model calculations according to the corresponding atomic jump models plotted on the right side of the figure. In Fig. 1.25a–c, models assuming jumps to the first, second, and third nearest neighbors, respectively, are assumed. Figure 1.25d corresponds to a combination of jumps along ½110 and ½001 with the ratio of 1.9:1 [86,90] and it shows the best agreement with the experiment.

FIGURE 1.25 Orientation dependence of timeintegrated NFS from an FeAl single crystal at a temperature of 1030  C. Solid circles represent the measured data, while the solid lines are model calculations assuming the following different types of jumps: (a) jumps of the Fe atoms to nearest-neighbor sites, (b) jumps to second nearestneighbor sites, (c) jumps to third nearest-neighbor sites, and (d) a combination of ½110 and ½001 jumps at a ratio of 1.9:1. (Reproduced from Ref. 90 with permission of Kluwer Academic Publishers.)

28

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.26 (a) Diffusion of atoms in an iron monolayer on the (110) surface of W(110). 1,
1, 2,
2 designate possible vacancy positions at NN sites of the probe atom; N is a site outside the NN shell to which a NN atom can jump. (b) Diffusion of adatom A on top of an iron surface. (Reproduced from Ref. 98 with permission of the American Physical Society.)

Up to now diffusion processes in bulk solids have been considered. Thanks to the enormous brilliance of the thirdgeneration synchrotron radiation sources and the instrumentation developments described in detail in Section 1.2.2, dynamic phenomena in submonolayer Fe films were investigated by the NRS technique. For these pioneering experiments [60,91,92], an Fe film on W(110) was selected as a model system for investigation of the interplay between structure, hyperfine interactions, and dynamics. The NRS time spectra reveal a beat pattern resulting from the presence of an electric field gradient due to the broken translational symmetry at the surface [93,94]. The beat structure originates from a tilting of the EFG main axis from the surface normal. (Due to the polarization nature of the synchrotron radiation, the quadrupole splitting occurs only if the main axis of the EFG deviates from the surface normal [95], Fig. 1.14c.) To deduce quantitative information about the atomic jump rates and explain this behavior, the theory of Dattagupta and Schroeder [96,97] developed for bulk materials was further elaborated for surface applications. Two models are considered: vacancy diffusion and diffusion of an adatom [98]. Figure 1.26a schematically shows a (110) plane of the bcc lattice with the 57 Fe probe atom surrounded by NN atoms and one vacancy. Only nearest-neighbor interactions are taken into consideration restricting the calculations to self-diffusion. Diffusion of an adatom is considered as well with the corresponding scheme shown in Fig. 1.26b [98]. In Fig. 1.27, the NRS time spectra (left column) and the M€ ossbauer spectra (right column) simulated by the theory are plotted for the case of jump diffusion of the probe atom via vacancies. The quantum beats in the time spectra and the corresponding line splitting in the M€ ossbauer spectra are due to the broken translational symmetry at the surface and the presence of a vacancy. The outer doublet in the M€ ossbauer spectra is due to the EFG from the vacancy, while the inner one is caused by the asymmetry at the surface. (Note that from the inner doublet only the component at positive energies is excited due to the polarization of the synchrotron radiation, as discussed in Section 1.3.1.) An increase in the jump frequency leads to a broadening of the lines corresponding to an accelerated decay of the excited nuclear resonant level. Simulations assuming adatom diffusion are plotted in Fig. 1.28. The fundamental difference with the case of vacancy diffusion is that the nuclear resonant probe atom is at rest; instead other atoms at the surface are jumping, which results again in the fluctuation of the nearest-neighbor surrounding of the probe atom. As a consequence, by increasing the jump frequency, the accelerated decay is less pronounced, and at very high frequencies, the fluctuating electric field gradients are averaging out leading to a motional narrowing of the M€ ossbauer line and slow decay of the NRS signal. The surface atomic diffusion was experimentally investigated on a pseudomorphic monolayer of Fe(110) grown on atomically flat W(110) substrate [99]. The NRS time spectra were measured at an incidence angle of 5 mrad, where the maximum of the delayed scattered radiation was found. Figure 1.29a–e plots the data (open circles) obtained at 300, 570, 670, 770, and 870 K, respectively, while (f) shows the room-temperature spectrum after the experiment at 770 K. The delayed intensity decays essentially exponentially in time, which implies that there is no additional electric field gradient beyond the electric field gradient from the asymmetry inherent in the surface, that is, the surface is practically perfect and undisturbed. By stepwise increase in the temperature a beat appears at about 45 ns that become more pronounced with the temperature. The highest temperature of the experiment was 870 K in order to avoid destruction of the Fe monolayer by diffusion of Fe into the W substrate. When the temperature is lowered again to 300 K, the beat structure disappears and the original spectrum is recovered, proving that thermally activated defects are responsible for the beat. The best fit of the NRS spectra with the theory developed in Ref. 98 was achieved with a combination of quadrupole interaction energy QN of 21(2) neV (N indicates “normal to the surface”) caused by the broken translational symmetry at the surface and a fluctuating quadrupole interaction QD (D indicates “defect”) caused by inhomogeneities, that is, lattice defects. A dependence of QD on temperature was found with values varying between 70(5) neV at 570 K and 87(10) neV

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

29

FIGURE 1.27 € ssbauer emission energy spectra (right column) for jump Simulations of NRS time spectra (left column) and Mo diffusion of probe atoms via vacancies in a surface layer. The reason for the beat structure in the NRS spectra and for € ssbauer spectra is the quadrupole interactions due to (i) the broken translational symmetry the line splitting in the Mo € ssbauer spectra, the outer doublet is due to the electric field gradient at the surface and (ii) the vacancy. In the Mo from the vacancy. From the inner doublet caused by the asymmetry of the surface, only the component at positive energies is excited because of the polarization of the synchrotron beam. In the NRS spectra, quadrupole relaxation is accompanied by a marked acceleration of the intensity decay with increasing jump frequency of the probe atom € ssbauer spectra. (Reproduced from Ref. 98 with permission of the corresponding to line broadening in the Mo American Physical Society.)

at 770 K and an estimated 104(15) neV at 870 K. The quadrupole interaction QN of 21 neV caused by the anisotropy of the surface is about 30% lower compared to that reported in other studies [94]. The difference is attributed to the presence of gaseous adatoms, which cannot be avoided during the long-lasting conversion electron M€ ossbauer spectroscopy experiments even at UHV conditions [92]. The quadrupole interaction QD from the defects arises from the perturbation of the neighborhood of the probe atom upon arrival of a defect. The quadrupole interaction QD then enters into a joint Hamiltonian with the quadrupole interaction QN caused by the anisotropy of the surface [98]. Because of the in-plane polarization of the synchrotron beam, QN does not produce quantum beats as long as the main axis of the surface quadrupole tensor is normal to the surface. However, the quantum beats develop when a defect arrives in the neighbor shell of the probe atom. Then the surface quadrupole interaction adds up in energy to the quadrupole interaction from the anisotropy caused by the defect’s perturbation of the symmetry. From the resulting quadrupole interaction, that is, from the amplitude of the beat, the concentration of defects can be deduced. With increasing temperature the width of the beat with its minimum at about 45 ns and its maximum at about 80 ns becomes larger (is smeared out) due to the thermal fluctuation of the defect structure that brings about a relaxation of the quadrupole interaction. Furthermore, the intensity decay steepens. From the temperature dependence, the velocity of the defect motion can be deduced. This yields the jump frequency of the defects and its temperature dependence. The derived values for the concentration and jump frequencies of vacancies as well as the diffusion coefficients from the fitted time spectra are summarized in Table 1.2 [99].

30

€ 1 IN SITU MOSSBAUER SPECTROSCOPY WITH SYNCHROTRON RADIATION ON THIN FILMS

FIGURE 1.28 Simulations of NRS time spectra € ssbauer (left column) and Mo emission energy spectra (right column) for an immobile probe atom in a surface layer when an adatom is jumping on top of the surface layer. The essential difference to the spectra of Fig. 1.27 is the much gentler decay of the NRS spectra in time with the decay completely vanishing for fast adatom jumps. This corresponds to the motional narrow€ ssbauer spectra. ing of the Mo (Reproduced from Ref. 98 with permission of the American Physical Society.)

1.3.3 Incoherent Inelastic Nuclear Resonant Scattering In this process, energy and momentum exchange between the photon and the crystal lattice is involved. The nuclear resonant absorption could take place even if the energy of the incoming photon Eg differs from the resonant energy E0 by DE. In that case, the energy excess þDE is transferred to the crystal lattice for excitation of thermal lattice vibrations, that is, phonons, or the photon gains energy DE from annihilation of phonons. Figure 1.30 schematically shows the relevant energy levels involved in the process for the simplest case of an Einstein solid, with S0, S1, and S2 being the elastic interaction, one-, and two-phonons creation/annihilation, respectively. Experimentally one measures the energy dependence of the probability W(E) for nuclear inelastic absorption by detuning the energy of the incoming photons from E0 in the millielectron volts range by the high-resolution monochromator. One of the following channels is used for detecting the events of nuclear inelastic interaction: (i) nuclear resonant fluorescence (the resonant photon is absorbed and reemitted); (ii) emission of a conversion electron (discussed in Section 1.1); and (iii) atomic fluorescence radiation that follows the internal conversion. In most experiments, the fluorescent radiation is detected, that is, nuclear inelastic absorption is measured. An exception, for example, is 119 Sn having nuclear resonant energy 23.871 keV that is below the K-edge (29.200 keV) of the Sn atom; therefore, the nuclear resonant fluorescence is detected corresponding to nuclear inelastic scattering experiment [100,101]. Here we present a brief outline of the data treatment formalism, whereas detailed theoretical elaboration of the method can be found elsewhere [21,22,38,102,103]. The probability for nuclear inelastic absorption W(E), after normalization, is given by the following expression derived by Singwi and Sj€ olander [104]: "

WðEÞ ¼ f LM dðE Þ þ

1 X 1

#

S n ðE Þ ;

(1.35)

ossbauer factor, the Dirac function d(E) describes the elastic part of absorption (zero-phonon where fLM is the Lamb–M€ term), and the nth term of the series Sn(E) represents the inelastic absorption accompanied by creation and/or annihilation of n phonons. The following recursive iteration, known as the Fourier–Log decomposition method [105], is

€ 1.3 SYNCHROTRON RADIATION-BASED MOSSBAUER TECHNIQUES

31

FIGURE 1.29 NRS time spectra (open circles) measured on a pseudomorphic Fe monolayer on W(110) at the indicated temperatures. The solid lines are fits assuming the vacancy diffusion model (a–e). Part (f) shows the time spectrum recorded at room temperature after the experiment at 770 K. (Reproduced from Ref. 99 with permission of the American Physical Society.)

TABLE 1.2 Concentration c and Jump Frequency w of Vacancies and Diffusion Coefficient D as Derived from the Fits to the NRS Spectra (Fig. 1.29) for One Pseudomorphic Monolayer of Fe on W(110) T (K)

c (%)

w (MHz)

D (10
12 m2 s
1)

300 570 670 770 870

2 mm s 1) [45]. Thus, if the values of the quadrupole splitting parameter of TDO and MauG do represent protonated species, then there must be other contributing factors that can lower the value for these intermediates. Due to the nature of the reaction, it is possible that these enzymes are using secondary ligands to fine-tune the properties of the high-valent Fe intermediates. This can already be seen in the DFT calculations that predicted hydrogen bonding interactions from secondary ligands in the enzyme active site [32].

15.6 CONCLUDING REMARKS The molecular biochemistry exhibited by two structurally diverse heme-containing enzymes that oxidize either free or protein-bound tryptophan residues is fascinating. Mono- and bis-Fe(IV) intermediates have been trapped and characterized by EPR and M€ ossbauer spectroscopy from these enzymes [7,40]. In particular, the bis-Fe(IV) intermediate (cpd G)  found in MauG is unprecedented. The target Trp residues of substrate are remotely located from the hemes, 40 A from the heme that reacts with exogenous molecules. Thus, a long-range remote catalysis must take place in MauG. Recently, the cocrystal structure of MauG and its substrate protein preMADH has been reported. When the crystal is soaked in a solution containing H2O2, the TTQ product is formed in the substrate protein [27]. Thus, the proposed long-range remote catalysis is validated [46]. It should be noted that one of the indelible stories about spatial separation of the catalytic center and the site that stores oxidizing equivalent(s) is described in ribonucleotide reductase [47]. In that case, the catalytic center is located in the R1 subunit, while the oxidizing power, a protein-bound Tyr radical, is generated and stabilized by a metal center in the R2 subunit. When the substrate is loaded properly in the active site of the R1 subunit, a long-range radical transfer takes place and a thiol-based transient radical is generated at the expense of the Tyr radical for the nucleotide reduction. The chemistry found in MauG is much like that found in ribonucleotide reductase and is more straightforward in terms of the nature of remote catalysis [46]. The diheme cofactor in MauG reacts with exogenous oxidants, generates high-valent Fe intermediate, and then transmits oxidizing equivalents to the substrate protein. Unquestionably, cpd G or bis-Fe(IV) is a new natural strategy discovered thus far for harnessing the problem of substrates way too large for an enzyme to accommodate. In this case, a remote catalysis must take place. Thus, the unprecedented bis-Fe(IV) intermediate found in MauG occupies an entry position for future study of long-range remote catalysis by metalloenzymes.

322

15 NATURE’S STRATEGY FOR OXIDIZING TRYPTOPHAN

REFERENCES 1. H. Fujii, Coord. Chem. Rev. 2002, 226, 51–60. 2. D.L. Harris, Curr. Opin. Chem. Biol. 2001, 5, 724–735. 3. Y. Watanabe, High-valent intermediates, in Porphyrin Handbook, K.M. Kadish, K.M. Smith, R. Guilard, eds., Vol. 4, Academic Press, New York, 2000, pp. 97–117. 4. A. Gumiero, C.L. Metcalfe, A.R. Pearson, E.L. Raven, P.C.E. Moody, J. Biol. Chem. 2011, 286, 1260–1268. 5. M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Chem. Rev. 1996, 96, 2841–2888. 6. G. Lang, K. Spartalian, T. Yonetani, Biochim. Biophys. Acta 1976, 451, 250–258. 7. X. Li, R. Fu, S. Lee, C. Krebs, V.L. Davidson, A. Liu, Proc. Natl. Acad. Sci. USA 2008, 105, 8597–8600. 8. W.S. McIntire, D.E. Wemmer, A. Chistoserdov, M.E. Lidstrom, Science 1991, 252, 817–824. 9. N.M. Okeley, W.A. van der Donk, Chem. Biol. 2000, 7, R159–R171. 10. T.W. Stone, L.G. Darlington, Nat. Rev. Drug Discov. 2002, 1, 609–620. 11. R. Schwarcz, Curr. Opin. Pharmacol. 2004, 4, 12–17. 12. O. Kurnasov, V. Goral, K. Colabroy, S. Gerdes, S. Anantha, A. Osterman, T.P. Begley, Chem. Biol. 2003, 10, 1195–1204. 13. F. Forouhar, J.L. Anderson, C.G. Mowat, S.M. Vorobiev, A. Hussain, M. Abashidze, C. Bruckmann, S.J. Thackray, J. Seetharaman, T. Tucker, R. Xiao, L.C. Ma, L. Zhao, T.B. Acton, G.T. Montelione, S.K. Chapman, L. Tong, Proc. Natl. Acad. Sci. USA 2007, 104, 473–478. 14. S. Lee, S. Shin, X. Li, V.L. Davidson, Biochemistry 2009, 48, 2442–2447. 15. L. Chen, R.C. Durley, F.S. Mathews, V.L. Davidson, Science 1994, 264, 86–90. 16. L. Chen, M. Doi, R.C. Durley, A.Y. Chistoserdov, M.E. Lidstrom, V.L. Davidson, F.S. Mathews, J. Mol. Biol. 1998, 276, 131–149. 17. Y. Wang, X. Li, L.H. Jones, A.R. Pearson, C.M. Wilmot, V.L. Davidson, J. Am. Chem. Soc. 2005, 127, 8258–8259. 18. C.J. van der Palen, D.J. Slotboom, L. Jongejan, W.N. Reijnders, N. Harms, J.A. Duine, R.J. van Spanning, Eur. J. Biochem. 1995, 230, 860–871. 19. M.E. Graichen, L.H. Jones, B.V. Sharma, R.J. van Spanning, J.P. Hosler, V.L. Davidson, J. Bacteriol. 1999, 181, 4216–4222. 20. A.R. Pearson, S. Marimanikkuppam, X. Li, V.L. Davidson, C.M. Wilmot, J. Am. Chem. Soc. 2006, 128, 12416–12417. 21. Y. Wang, M.E. Graichen, A. Liu, A.R. Pearson, C.M. Wilmot, V.L. Davidson, Biochemistry 2003, 42, 7318–7325. 22. X. Li, L.H. Jones, A.R. Pearson, C.M. Wilmot, V.L. Davidson, Biochemistry 2006, 45, 13276–13283. 23. A.R. Pearson, T. De La Mora-Rey, M.E. Graichen, Y. Wang, L.H. Jones, S. Marimanikkupam, S.A. Agger, P.A. Grimsrud, V.L. Davidson, C.M. Wilmot, Biochemistry 2004, 43, 5494–5502. 24. P.R. Ortiz de Montellano, Cytochrome P450: Structure, Mechanism, and Biochemistry, 2nd edn, Plenum Press, New York, 1995. 25. J.E. Erman, L.P. Hager, S.G. Sligar, Adv. Inorg. Biochem. 1994, 10, 71–118. 26. R. Braaz, W. Armbruster, D. Jendrossek, Appl. Environ. Microbiol. 2005, 71, 2473–2478. 27. L.M.R. Jensen, R. Sanishvili, V.L. Davidson, C.M. Wilmot, Science 2010, 327, 1392–1394. 28. X. Li, M. Feng, Y. Wang, H. Tachikawa, V.L. Davidson, Biochemistry 2006, 45, 821–828. 29. R. Fu, F. Liu, V.L. Davidson, A. Liu, Biochemistry 2009, 48, 11603–11605. 30. P.G. Debrunner, M€ ossbauer spectroscopy of Fe porphyrins, in Iron Porphyrins, Physical Bioinorganic Chemistry Series, Vol. 4, A.B.P. Lever, H.B. Gray, eds., VCH Publishers, Inc., New York, 1989, pp. 137–234. 31. K.L. Stone, L.M. Hoffart, R.K. Behan, C. Krebs, M.T. Green, J. Am. Chem. Soc. 2006, 128, 6147–6153. 32. Y. Ling, V.L. Davidson, Y. Zhang, J. Phys. Chem. Lett. 2010, 1, 2936–2939. 33. J.T. Groves, R. Quinn, T.J. McMurry, M. Nakamura, G. Lang, B. Boso, J. Am. Chem. Soc. 1985, 107, 354–360. 34. E. Bill, V. Schunemann, A.X. Trautwein, R. Weiss, J. Fischer, A. Tabard, R. Guilard, Inorg. Chim. Acta 2002, 339, 420–426. 35. Y. Kotake, I. Masayama, Z. Physiol. Chem. 1936, 243, 237–244. 36. R. Gupta, R. Fu, A. Liu, M.P. Hendrich, J. Am. Chem. Soc. 2010, 132, 1098–1109. 37. E. Fukumura, H. Sugimoto, Y. Misumi, T. Ogura, Y. Shiro, J. Biochem. 2009, 145, 505–515. 38. L.W. Chung, X. Li, H. Sugimoto, Y. Shiro, K. Morokuma, J. Am. Chem. Soc. 2010, 132, 11993–12005. 39. A. Lewis-Ballester, D. Batabyal, T. Egawa, C. Lu, Y. Lin, M.A. Marti, L. Capece, D.A. Estrin, S.-R. Yeh, Proc. Natl. Acad. Sci. USA 2009, 106, 17371–17376. 40. R. Fu, R. Gupta, S. Wang, J. Geng, K. Dornevil, Y. Zhang, M.P. Hendrich, A. Liu, J. Biol. Chem. 2011, 286, 26541–26554. 41. C.E. Schulz, R. Rutter, J.T. Sage, P.G. Debrunner, L.P. Hager, Biochemistry 1984, 23, 4743–4754.

REFERENCES

42. 43. 44. 45. 46. 47.

T. Harami, Y. Maeda, Y. Morita, A. Trautwein, U. Gonser, J. Chem. Phys. 1977, 67, 1164–1169. O. Horner, J.L. Oddou, J.M. Mouesca, H.M. Jouve, J. Inorg. Biochem. 2006, 100, 477–479. R. Rutter, L.P. Hager, H. Dhonau, M. Hendrich, M. Valentine, P. Debrunner, Biochemistry 1984, 23, 6809–6816. R.K. Behan, M.T. Green, J. Inorg. Biochem. 2006, 100, 448–459. J.M. BollingerJr., M.L. Matthews, Science 2010, 327, 1337–1338. J. Stubbe, D.G. Nocera, C.S. Yee, M.C.Y. Chang, Chem. Rev. 2003, 103, 2167–2201.

323

C H A P T E R 1 6

IRON IN NEURODEGENERATION JOLANTA GAŁA˛ZKA-FRIEDMAN,1 ERIKA R. BAUMINGER,2 AND ANDRZEJ FRIEDMAN3 1

Faculty of Physics, Warsaw University of Technology, Warsaw, Poland Racah Institute of Physics, Hebrew University of Jerusalem, Jerusalem, Israel 3 Department of Neurology, Faculty of Health Science, Medical University of Warsaw, Warsaw, Poland 2

16.1 INTRODUCTION Although iron is essential for normal functioning of the human body, it may also play a deleterious role as a trigger of various diseases [1,2]. In this chapter, we will present results of studies on the role of iron in neurodegenerative diseases such as Parkinson’s disease (PD), Alzheimer’s disease (AD), and progressive supranuclear palsy (PSP). We report on studies whose main tool was M€ ossbauer spectroscopy, but also other complementary techniques such as electron microscopy, enzyme-linked immunosorbent assay (ELISA), and atomic absorption. In Section 16.2, we will present a short review on neurodegeneration and oxidative stress. In Sections 16.3 and 16.4, the properties of iron-binding compounds in the normal human brain will be discussed. In Sections 16.5 and 16.6, we will compare the properties of iron present in normal tissues with those from the structures affected by the disease process. Section 16.7 will summarize the possible role of iron in the pathological processes of some neurodegenerative diseases of the human brain.

16.2 NEURODEGENERATION AND OXIDATIVE STRESS Neurological diseases such as PD, AD disease, and many others are classified as neurodegenerative disorders. Neurodegeneration is a progressing process of the death of nervous cells in the central nervous system. In each of these diseases, the destruction in a specific area of the brain is the most prominent. This is substantia nigra (SN) for PD, the hippocampal cortex (HP) for AD, and globus pallidus (GP) and SN for a typical parkinsonism, such as PSP (Fig. 16.1). In all these structures, not only nervous cells but also glial cells are present. Several metabolic pathways may lead to the death of nervous cells. Only oxidative stress, which may be one of these pathways, is dealt with here. Chemical reactions, involving oxidation and reduction, permanently take place in all living cells. In these reactions, free radicals are produced and immediately annihilated by free radicals scavengers. The balance between production and inactivation of free radicals is crucial for the safety of the cells. In pathological conditions, such as oxidative stress, there might be either an overproduction or an insufficient inactivation of free radicals. An overproduction of free radicals may be mediated by the Fenton’s reaction, in which divalent iron plays an important role: Fe2þ þ H2 O2 ! Fe3þ þ OH
þ OH M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

324

€ 16.3 MOSSBAUER STUDIES OF HEALTHY BRAIN TISSUE

325

FIGURE 16.1 Magnetic resonance imaging of a normal human brain (material of the Department of Neurology, Medical University of Warsaw, Poland). Arrows show the location of GP—globus pallidus, HP—hippocampus (a—coronal section), and SN—substantia nigra (b—saggital section).

An excess of free radicals may destroy cell membranes, and thus lead directly to the death of cells [3]. Studies dealing with the role of iron, which may be involved in oxidative stress in the human brain, are therefore of ultimate importance.

€ 16.3 MOSSBAUER STUDIES OF HEALTHY BRAIN TISSUE There are about 4 g of iron in the human body. Sixty percent of this iron is found in hemoglobin, and the remaining 40% is found mainly in liver, spleen, heart, muscles, and brain. The distribution of iron in the brain differs between various brain structures. There are about 180 ng of iron per mg of wet tissue in substantia nigra and globus pallidus, and only about 50 ng of iron per mg of wet tissue in the hippocampal cortex [4–7]. The weight of tissue samples that can be used in MS studies is at most 1 g so that less than about 50–200 mg of natural iron (1–4 mg of Fe57) could be present in each sample. In order to study such low concentrations of iron, the MS measurements have to be performed in a highly iron-clean environment. This demands iron-free counter and cryostat windows and iron-free sample holders. In the M€ ossbauer studies reported here, M€ ossbauer spectra measured with a sample holder filled with distilled water proved that, indeed, the background, measured without tissues, was very tiny, effectively nonexistent, in these experiments [8]. M€ ossbauer spectra of three brain structures: SN, GP, and HP were obtained at 4.1 K, 90 K, and also at room temperature for lyophilized samples. All brain samples were obtained from autopsies performed within 48 h after death and were kept in liquid nitrogen until assayed or lyophilized. Formalin-fixed brain samples were not used, as preliminary measurements showed that this form of storage caused significant changes in the iron composition, which were proportional to the time the tissues had been stored in formalin [9]. All M€ ossbauer spectra of the measured tissues presented an asymmetric doublet at room temperature and at 90 K, and a sextet at 4.1 K. Typical M€ ossbauer spectra of SN, GP, and HP obtained at 90 K are shown in Fig. 16.2, and a typical M€ ossbauer spectrum of SN obtained at 4.1 K is shown in Fig. 16.3. The sextet observed at 4.1 K suggests that almost all iron atoms in these structures form clusters and interact magnetically at low temperatures. The above presented evolution of the M€ ossbauer spectra as a function of temperature is characteristic for spectra of small nanoparticles, in which the iron orders antiferromagnetically, but displays the phenomenon of superparamagnetism due to the small size of these clusters. This is the case for ferritin or hemosiderin, where the protein shells separate the small iron cores from each other. Also, the M€ ossbauer parameters obtained from the doublets at 90 K and at room temperature and the M€ ossbauer parameters obtained from the sextet at 4.1 K, are very similar to those observed for ferritin or hemosiderin (Table 16.1). These results strongly suggest that almost all the iron in globus pallidus, substantia nigra, and hippocampus is trivalent iron, bound to ferritin or hemosiderin as ferrihydrite. It was suggested [10] that iron in the brain may be bound to neuromelanin (NM) in small clusters. Iron was indeed detected in neuromelanin isolated from substantia nigra, yet it is not clear whether this iron was bound to NM in the brain or was attached to it during the isolation procedure [11]. M€ ossbauer parameters obtained from the spectra of isolated

326

16 IRON IN NEURODEGENERATION

1.000 HP

FIGURE 16.2 € ssbauer spectra of Typical Mo fresh frozen brain samples: HP, GP, and SN obtained at 90 K. The weight of each of the samples was about 1 g. The differences in the sizes of the effects are due to different concentrations of iron present in these structures. (Reproduced with permission from Ref. 32.)

Relative counting rate

0.999 0.998 1.000

GP

0.999 0.998 1.000

SN

0.999 0.998 –3

–2

0

–1

1

2

3

4

Velocity (mm s–1)

FIGURE 16.3 € ssbauer spectrum of A typical Mo fresh frozen substantia nigra obtained at 4.1 K. (Reproduced with permission from Ref. 32.)

Relative counting rate

1.0004 1.0002 1.0000 0.9998 0.9996 0.9994 0.9992 –10 –8

–6

–4

0 2 4 –2 Velocity (mm s–1)

6

8

10

NM were similar, yet not identical, to those obtained from ferritin. In the isolated NM, a magnetic sextet was observed at 90 K, indicating larger iron clusters than in ferritin [12]. The sextet observed at 4.2 K also showed a small quadrupole interaction that was not observed in ferritin (Table 16.1). In the spectra of the brain samples, no sextet was observed at 90 K and, therefore, only a small percentage of brain iron in whole tissues may be bound to neuromelanin [5]. However, as most of the iron in the brain is located in the glia and only a very small percentage of it is bound to neurons, iron in neurons may escape detection by MS measurements on whole tissues. This iron in neurons may be bound to neuromelanin or to another iron-storage compound different from ferritin.

€ ssbauer Parameters of Iron in Brain Samples (GP, SN, and HP), Ferritin, and Neuromelanin TABLE 16.1 Mo Sample

IS (mm s 1)

QS (mm s 1)

GP 90 K [2] GP 4.1 K [2] SN 90 K [2] SN 4.1 K [2] HP 90 K [2] HP 4.1 K [2] Ferritin 90 K [13] Ferritin 4.1 K [13] Neuromelanin 300 K [14] Neuromelanin 4.2 K [14]

0.46  0.02 0.48  0.02 0.47  0.01 0.48  0.04 0.47  0.01 0.49  0.02 0.47  0.02 0.49  0.02 0.37  ? 0.5

0.70  0.04 0.00  0.01 0.67  0.04 0.00  0.01 0.69  0.01 0.00  0.01 0.72  0.05 0.00  0.01 0.59  ? 0.26

IS: isomer shift relative to iron metal at room temperature; QS: quadrupole splitting; H: effective hyperfine magnetic field.

H (T) 49.5  0.5 49.2  0.5 49.8  0.5 49.4  0.5 49.5

327

16.4 PROPERTIES OF FERRITIN AND HEMOSIDERIN PRESENT IN HEALTHY BRAIN TISSUE

As mentioned above, the M€ ossbauer spectra of iron in all three brain structures studied, showed only ferritin-like, high-spin trivalent iron. There has been one spectrophotometry study that claimed about 75% of the iron in control SN (and about 50% of iron in parkinsonian SN) is divalent [15]. Such a high percentage of divalent iron could not escape detection by MS. Computer simulations, aimed at determining the concentration of divalent iron in the tissues that could escape detection by MS, showed this to be at most 5% [5]. Analysis of the procedure of the sample preparation for spectrophotometry lead us to the suspicion that the high concentration of divalent iron found in this experiment may have been due to the destruction of the protein shell of ferritin by hydrochloric acid and pepsin, which may have caused an efflux of iron from ferritin. It may be that this iron was then reduced from trivalent to divalent iron by ascorbic acid present in the cells. Indeed, tissues pretreated in this way and assessed by MS did show large concentrations of divalent iron [16].

16.4 PROPERTIES OF FERRITIN AND HEMOSIDERIN PRESENT IN HEALTHY BRAIN TISSUE Ferritin is an iron-storage protein, which keeps iron in a safe form within a protein shell. The protein shell of mammalian ferritin is composed of 24 subunits of heavy (H) and light (L) amino acid chains. The external diameter of the protein shell is about 12 nm, and the diameter of the internal cavity is about 7 nm. Inside this cavity up to 4500 ions of iron may be located. Most of the ferritin molecules are not full, but contain varying amounts of iron [17]. L and H chains of ferritin differ in amino acid sequences, mass, and function. H chains contain ferroxidase centers, where Fe(II) is oxidized to Fe(III) before moving into the protein cavity. In the cavity, iron is deposited as Fe(III) in the form of ferrihydrite. Usually, the cavity also contains variable amounts of inorganic phosphate [18]. L chains do not contain ferroxidase centers. They are involved mainly in building up the iron core of ferritin and keeping Fe(III) safely within the protein shell. The proportion between the number of H and L chains is different in different organs. In liver and spleen, L chains are predominant, while in heart and brain H chains dominate [17]. The concentrations of H and L chains obtained by ELISA studies using monoclonal antibodies (a generous gift from Professor Arosio) are shown in Table 16.2. Hemosiderin is a much less studied molecule. It is generally considered a product of the degradation of ferritin [17]. The diameters of the iron cores of ferritin and of hemosiderin were determined by electron microscopy on ferritin and hemosiderin isolated from the tissues. As presented in the table, the average diameters of the cores are significantly smaller in brain ferritin and hemosiderin than in liver ferritin and hemosiderin (3.1–3.7 nm versus 6.0 nm and 2.0  05 nm versus 3.5 nm, respectively) [19]. The smaller diameter of the iron cores of ferritin in brain compared to that in liver tissues fits well with the blocking temperatures measured by M€ ossbauer spectroscopy. The blocking temperature determined by MS for SN is about 15 K [20] compared to about 35–40 K for human liver [21]. Figure 16.4 presents the M€ ossbauer spectra obtained from SN at 20 K, 10 K, and 4.1 K, from which the blocking temperature was estimated. The above-presented properties of brain ferritin and hemosiderin suggest a different metabolism of iron in the brain and in organs such as the liver. As the H-ferritin subunits are involved more than L subunits in taking up and possibly also in releasing iron from ferritin, the significantly higher H/L ratio in the brain may possibly point to a higher turnover of iron in the brain [22].

TABLE 16.2 Concentration (ng mg

1

Wet Tissue) of H- and L-Ferritin Chains in Brain Tissues (GP, SN, HP), Liver, and Heart, Determined by ELISA and Diameter (nm) of the Iron Cores Determined by Electron Microscopy [2,12]

Structure

H-Ferritin

L-Ferritin

H/L Ratioa

Iron Core Diameter

GP SN HP Liver Heart

335  21 375  38 101  9 179  25 61  6

70.9  6.2 98  12 9.4  1.6 553  103 9.2  1.3

5.1  0.4 4.3  0.3 14.5  1.8 0.40  0.04 9.7  1.6

3.5  0.1 3.7  0.1 3.1  0.2 6.0  0.5 –

a

The H/L ratios were calculated separately for each sample and the mean values calculated from all the individual measurements are given in the table.

328

16 IRON IN NEURODEGENERATION

1.001 1.000

Relative counting rate

0.999

(a)

0.998 1.0000 0.9995

(b)

0.9990 e

1.0000 d

FIGURE 16.4 € ssbauer spectra obtained Mo from substantia nigra at 20 K (a), 10 K (b), and 4.1 K (c). (Reproduced with permission from Ref. 32.)

0.9998 0.9996

(c)

0.9994 –10 –8 –6 –4 –2

0

2

4

6

8 10

Velocity (mm s–1)

16.5 CONCENTRATION OF IRON PRESENT IN HEALTHY AND DISEASED BRAIN TISSUE: LABILE IRON It has been believed for a long time that an increase of the total iron concentration in specific structures of the brain is a primary cause of neurodegenerative diseases [15]. Some authors even suggested that Parkinson’s disease is a kind of brain siderosis [23]. However, comparing the results of studies assessing the concentration of total iron in specific structures affected in neurodegenerative diseases, such as iron in SN in PD and PSP, and iron in HP in AD, shows that this is still under dispute. The results obtained by various authors employing different techniques are summarized in Table 16.3. In all these studies, the measurements were performed on a small number of samples. As there are large differences in the concentration of iron between individuals, necessarily the errors in the obtained results are large. The first comparison between the iron concentration in SN of parkinsonian patients and the iron concentration in SN from brains of patients who died without clinical or pathological signs of neurodegenerative diseases (control), was

TABLE 16.3 Iron Concentrations (ng mg 1) in Substantia Nigra in PD and Control as Obtained by Various Methods in Fresh Frozen Samples

Method XRF [24] SPH [15] AA [25] ICP [26] Col [27] MS [20]

Iron in Control not stated 48  8 140  13 1159  379 ng mg 5600  400 ng mg 177  14

1 1

protein protein

No. of Samples ? 8 6 22 8 29

Iron in PD not stated 85  11 281  22 1813  846 ng mg 4600  300 ng mg 177  18

1 1

No. of Samples

protein protein

? 8 6 18 14 17

Ratio of Iron in PD to Iron in Control 2? 1.77  0.37 2.01  0.24 1.56  0.89 0.82  0.08 1.00  0.13

XRF: X-ray fluorescence; SPH: spectrophotometry; AA: atomic absorption; ICP: inductively coupled plasma spectroscopy; Col: colorimetry; MS: M€ ossbauer spectroscopy; PD: Parkinson’s disease.

329

16.5 CONCENTRATION OF IRON PRESENT IN HEALTHY AND DISEASED BRAIN TISSUE: LABILE IRON

TABLE 16.4 Iron Concentrations (ng mg 1) in Substantia Nigra in PD and Control as Obtained by Various Methods in Lyophilized Samples Method

Iron in Control

No. of Samples

Iron in PD

No. of Samples

Ratio of Iron in PD to Iron in Control

ICP [28] AA [29] MS [5]

580  60 613  56 890  70

9 12 1

780  60 653  55 644  60

8 9 1

1.34  0.17 1.07  0.13 0.93  0.18

ICP: inductively coupled plasma spectroscopy; AA: atomic absorption; MS: M€ ossbauer spectroscopy; PD: Parkinson’s disease.

performed by Earle in 1968 [24] using X-ray fluorescence (XRF). Earle reported a twofold increase of iron in parkinsonian SNs as compared to control. Some later studies also found a substantial increase in pathological tissues, whereas others did not find any difference in iron concentrations between control and PD. In Table 16.3, we summarize the ratio between the iron concentrations in SN of PD and control, as obtained by various methods from fresh frozen tissues. Some authors assessed the concentration of iron in lyophilized samples and the results cited also differ significantly. Table 16.4 presents the results of these studies. Such a wide range of results as shown in the tables is puzzling and we can only speculate what the reasons for such discrepancies could be. As mentioned above, the main reason may lie in the variety of concentration of iron in different individuals and the comparatively small numbers of samples studied with each technique. The twofold increase of iron concentrations in PD found by Earle may perhaps be due to the long-term storage of the control samples in formalin (exceeding 70 years for some samples), while the storage time for the PD samples was significantly shorter. Artifacts could also be caused by the preparation procedures of the samples. This may apply mainly to spectrophotometry [15], where it may be that during the preparation procedure, except for reducing iron, more iron is also lost from control than from PD samples. It is also possible that the variety of results is due to the inhomogeneity of the distribution of iron in SN tissues. In some studies, only very small parts of SN tissues were examined, which by chance may include more, or less, iron than the average concentration of iron in the whole SN. These possible artifacts are avoided by MS, as in M€ ossbauer studies whole SN samples, which do not require any pretreatment except freezing, are used. Preliminary M€ ossbauer studies were also performed on samples obtained from brains of patients who had died from PSP. PSP is an example of atypical parkinsonism. In this study, the concentration of total iron was assessed in SN and GP, as in this disease both structures are involved. The M€ ossbauer spectra of PSP samples showed a substantial increase of the concentrations of iron in both SN and GP compared to control. Table 16.5 summarizes the results of the assessments of the average concentrations of iron in control and PSP tissues. These results indicate that in PSP there is an increase of about 50% of the total iron concentrations in both SN and GP. In order to elucidate the possible role of iron in Alzheimer’s disease, we compared the concentrations of iron in hippocampal tissues obtained from AD and control brains. Figure 16.5 shows a M€ ossbauer spectrum obtained at room temperature of a lyophilized sample of hippocampal tissue from AD. The parameters obtained from the spectra of hippocampi are similar, within the experimental error, to the parameters obtained from SN and GP spectra. The only difference is in the size of the effect. The average iron concentration calculated from the spectra of 10 control HP was 45  10 ng mg 1 wet tissue. The concentration of iron assessed by atomic absorption in 10 AD HP was 66  13 ng mg 1 wet tissue. This difference between AD and control is not statistically significant [7]. It is important to note that this review concerns the total iron in the tissues, which is mainly bound to ferritin and, therefore, cannot trigger the Fenton’s reaction. In the Fenton’s reaction, only labile iron can participate and only the increase of the concentration of this labile iron may cause an overproduction of free radicals leading to oxidative stress.

TABLE 16.5 Average Iron Concentrations (ng mg 1) Obtained from n Samples of Fresh Frozen Control and PSP Tissues [6] Structure GP SN

PSP 257  19 (n ¼ 10) 301  26 (n ¼ 10)

Control 183  22 (n ¼ 12) 188  22 (n ¼ 9)

Iron in PSP/Iron in Control 1.40  0.20 1.60  0.23

330

16 IRON IN NEURODEGENERATION

FIGURE 16.5 € ssbauer spectrum of a lyophMo ilized sample of AD hippocampus obtained at room temperature. (Reproduced with permission from Ref. 7.)

Relative counting rate

1.0002 1.0000 0.9998 0.9996 0.9994 –4

–3

–2

–1 0 1 Velocity (mm s–1)

2

3

4

The assessment of the concentration of labile iron in parkinsonian and control SN was made by atomic absorption after eliminating all molecules that could be bound to large molecules, such as ferritin, and thus were bigger than 10 kDa. The concentration of the labile iron in PD SN was found to be significantly higher than in control (135  10 ng g 1 versus 76  5 ng g 1 wet tissue) [30]. It is important to emphasize that these iron concentrations are about 2000 times smaller than the total iron concentration and could not be detected by M€ ossbauer spectroscopy.

€ 16.6 ASYMMETRY OF THE MOSSBAUER SPECTRA OF HEALTHY AND DISEASED BRAIN TISSUE As mentioned above, all M€ ossbauer spectra of brain tissues obtained at 273 K and 90 K represent a doublet showing an asymmetry, as do also all M€ ossbauer spectra of ferritin. Good fits to M€ ossbauer spectra of ferritin demand a superposition of at least two doublets; however, the analyses of the spectra with two doublets are not unique and depend on initial parameters or assumptions. In order to quantify the asymmetry of the M€ ossbauer spectra of ferritin, the spectra were fitted by two lines of equal areas. As an example of asymmetrical spectra, the M€ ossbauer spectrum of pathological PSP GP obtained at 90 K is shown in Fig. 16.6 [2]. The asymmetry coefficient was defined as the ratio between the linewidth of the lower velocity line to that of the higher velocity line. Such a coefficient of asymmetry has only a phenomenological character and no obvious physical interpretation. However, by comparing the mean values of this coefficient for control and diseased tissues, statistically significant differences were found (Table 16.6). Three possible causes for the differences in the asymmetry were considered: differences in the concentrations of nonferritin-like iron, differences in the crystallinity of the iron cores of ferritin, and differences in composition of the iron core. As the concentration of labile iron is about 2000 times smaller than the total iron concentration, the difference in the concentration of labile iron between PD and control SN could not be the cause for this change in asymmetry. On the other hand, a difference in the protein shell structure of ferritin in control and disease-affected tissues could cause a different crystallinity of the iron cores. ELISA studies, summarized in Table 16.7, did in fact show different H/L ratios in the protein shells of ferritin in pathological SN for PD and in pathological HP for AD, and control samples.

FIGURE 16.6 € ssbauer spectrum of globus Mo pallidus from PSP brain obtained at 90 K and fitted with two singlets. (Reproduced with permission from Ref. 2.)

Relative counting rate

1.0002 1.0000 0.9998 0.9996 0.9994 0.9992 0.9990 0.9988 –4

–3

–2

0 1 –1 Velocity (mm s–1)

2

3

4

331

REFERENCES

€ ssbauer Spectra Obtained from n TABLE 16.6 Average Coefficients of Asymmetry of the Mo Samples of Fresh Frozen Control and Pathological Tissues

SN [8] GP [2]

Control

Pathology

1.06  0.02 (n ¼ 20) 1.08  0.02 (n ¼ 9)

PD 1.14  0.03 (n ¼ 9) PSP 1.12  0.03 (n ¼ 10)

TABLE 16.7 Concentration (ng mg 1) of H- and L-Ferritin in Fresh Frozen Samples of SN (Control and PD) and Hippocampal Cortex (Control and AD) [7] Structure

H-Ferritin

L-Ferritin

SN control SN PD HP control HP AD

375  38 534  71 101  9 397  29

98  12 52  8 92 29  5

No data on the structure of ferritin in pathological tissues from brains of patients who died with PSP are available. The ferritin core may also contain different amounts of other elements, such as phosphor or copper, in control and pathological tissues and this also could cause a different asymmetry in the M€ ossbauer spectra.

16.7 CONCLUSION: THE POSSIBLE ROLE OF IRON IN NEURODEGENERATION The results of all studies of iron and ferritin in the three brain structures—SN, GP, and HP, suggest that iron may play a role in all three neurodegenerative processes studied—Parkinson’s, Alzheimer’s, and PSP. It seems, however, that the exact mechanisms in which iron participates in these diseases are somewhat different. In PD, there is no increase of the total iron concentration, but only an increase of labile iron. This labile iron may trigger the Fenton chemistry leading to an overproduction of free radicals. This increase of labile iron concentration may be connected with a decrease of the concentration of L-ferritin that plays a crucial role in safekeeping of iron within the iron core of ferritin. As nothing is known about the labile iron in AD and PSP, we may only speculate about the way of action of iron. In AD, no significant increase of total iron was found, but a significant increase of the concentration of both H- and L-ferritin was detected. In preliminary measurements for PSP, an increase of about 50% of total iron compared to control was found in GP and SN, but nothing is known yet about any possible change in the structure of ferritin. Whether this is connected with an intensification of the Fenton chemistry, remains to be clarified. The higher asymmetry of the M€ ossbauer spectra of GP obtained from PSP brains compared to control, may represent a different crystallization or composition of the iron cores of ferritin in the pathological tissues. A similar higher asymmetry of the spectra was found in parkinsonian SN compared to control. It is important to stress that M€ ossbauer spectroscopy did not detect any divalent iron in all the samples measured. In a study using analytical transmission electron microscopy, the authors suggested that iron within the ferritin core in pathological SN and HP was present mainly as mixed ferric–ferrous iron oxides (magnetite-like or w€ustite) and not as ferrihydrite that is the main mineral in control brain ferritin [31]. M€ ossbauer spectra of magnetite or w€ustite differ from ferrihydrite spectra, but only ferrihydrite-like spectra were observed by MS in all pathological tissues. If other phases were present, they could be there only in minor quantities.

REFERENCES 1. R.B. Lauffer, ed., Iron and Human Disease, CRC Press, Boca Raton, 1992. 2. J. Galazka-Friedman, Hyperfine Interact 2008, 182, 31–44. 3. B. Halliwell, Pathol. Biol. 1996, 44, 6–13.

332

16 IRON IN NEURODEGENERATION

4. B. Hallgren, P. Sourander, J. Neurochem. 1958, 3, 41–51. 5. J. Gała˛zka-Friedman, E.R. Bauminger, A. Friedman, M. Barcikowska, D. Hechel, I. Nowik, Mov. Disord. 1996, 11, 8–16. 6. J. Galazka-Friedman, E.R. Bauminger, K. Szlachta, K. Schweitzer, Z. Wszolek, D. Dickson, A. Friedman, Acta. Phys. Pol. 2009, 115, 431–433. 7. J. Gała˛zka-Friedman, E.R. Bauminger, K. Szlachta, D. Koziorowski, R. Tomasiuk, A. Jaklewicz, Z.K. Wszolek, D. Dickson, K. Kaplinska, A. Friedman, Acta. Phys. Pol. 2011, 119, 81–83. 8. J. Galazka-Friedman, E.R. Bauminger, D. Koziorowski, A. Friedman, Biochim. Biophys. Acta 2004, 1688, 130–136. 9. E.R. Bauminger, M. Barcikowska, A. Friedman, J. Gała˛zka-Friedman, D. Hechel, I. Nowik, Hyperfine Interact 1994, 91, 853–856. 10. K. Jellinger, E. Kienzel, G. Rumpelmair, P. Riederer, H. Stachelberger, D. Ben-Shachar, M.B.H. Youdim, J. Neurochem. 1992, 59, 1168–1171. 11. L. Zecca, A. Stroppolo, A. Gatti, D. Tampellini, M. Toscani, M. Gallorini, G. Giaveri, P. Arosio, P. Santambrogio, R.G. Fariello, E. Karatekin, M.H. Kleinman, N. Turro, O. Hornykiewicz, F.A. Zucca, Proc. Natl. Acad. Sci. 2004, 101, 9843–9848. 12. J. Galazka-Friedman, E.R. Bauminger, A. Friedman, D. Koziorowski, K. Szlachta, Hyperfine Interact 2005, 165, 285–288. 13. E.R. Bauminger, S.G. Cohen, S. Ofer, A.E. Rachmilevitz, Proc. Natl. Acad. Sci. 1979, 76, 939–943. 14. M. Gerlach, A.X. Trautwein, L. Zecca, M.B.H. Youdim, P. Riederer, J. Neurochem. 1995, 65, 923–926. 15. E. Sofic, P. Riederer, H. Heinsen, H. Beckmann, G.P. Reynolds, G. Hebenstreit, M.B.H. Youdim, J. Neural Transm. 1988, 74, 199–205. 16. J. Gała˛zka-Friedman, A. Friedman, Acta Neurobiol. Exp. 1997, 57, 210–225. 17. P.M. Harrison, P. Arosio, Biochim. Biophys. Acta 1996, 1275, 161–203. 18. A. Treffry, P.M. Harrison, Biochem. J. 1978, 171, 313–320. 19. J. Galazka-Friedman, E.R. Bauminger, T. Tymosz, A. Friedman, Hyperfine Interact (C) 1998, 3, 49–52. 20. J. Galazka-Friedman, A. Friedman, E.R. Bauminger, Hyperfine Interact 2009, 189, 31–37. 21. S.H. Bell, M.P. Weir, D.P.E. Dickson, J.F. Gibson, G.A. Sharp, T.J. Peters, Biochim. Biophys. Acta 1984, 787, 227–236. 22. N.D. Chasteen, P.M. Harrison, J. Struct. Biol. 1999, 126, 182–194. 23. M.B.H. Youdim, D. Ben-Shachar, P. Riederer, Acta Neurol. Scand. 1989, 126, 47–54. 24. K.M. Earle, J. Neuropathol. Exp. Neurol. 1968, 27, 1–14. 25. P.D. Griffiths, A.R. Crossman, Dementia 1993, 4, 61–65. 26. V.M. Mann, J.M. Cooper, S.E. Daniel, K. Srai, P. Jenner, C.D. Marsden, A.H. Schapira, Ann. Neurol. 1994, 36(6), 876–881. 27. D.A. Loeffler, J.R. Connor, P.I. Juneau, B.S. Snyder, L. Kanaley, A.J. DeMaggio, H. Nguyen, C.M. Brickman, P.A. LeWitt, J. Neurochem. 1995, 65, 710–716. 28. D.T. Dexter, F.R. Wells, A.J. Lees, F. Agid, Y. Agid, P. Jenner, C.D. Marsden, J. Neurochem. 1989, 52, 1830–1836. 29. R.J. Uitti, A.H. Rajput, B. Rozdilsky, M. Bickis, T. Wollin, W.K. Yuen, Can. J. Neurol. Sci. 1989, 16, 310–314. 30. A. Wypijewska, J. Galazka-Friedman, E.R. Bauminger, Z.K. Wszolek, K.J. Schweitzer, D.W. Dickson, A. Jaklewicz, D. Elbaum, A. Friedman, Parkinsonism Relat. Disord. 2010, 16, 329–333. 31. C. Quintana, J.M. Cowley, C. Marhic, J. Struct. Biol. 2004, 147, 166–178. 32. J. Galazka-Friedman, E.R. Bauminger, K. Szlachta, A. Friedman, J. Phys. Condens. Matter 2012, 24, 244101.

C H A P T E R 1 7

€ EMISSION (57Co) MOSSBAUER SPECTROSCOPY: BIOLOGYRELATED APPLICATIONS, POTENTIALS, AND PROSPECTS

ALEXANDER A. KAMNEV Laboratory of Biochemistry, Institute of Biochemistry and Physiology of Plants and Microorganisms, Russian Academy of Sciences, Saratov, Russian Federation

17.1 INTRODUCTION €ssbauer spectroscopy, with the stable 57 Fe isotope most widely used up to now, The traditional absorption variant of Mo has had a very rich history of applications in a broad range of biology-related and biomedical fields, which started soon €ssbauer’s discovery. (Some earlier and more recent related representative publications are reviewed in after Rudolf Mo €ssbauer spectroscopy, the well-known main methodological Chapter 13[1].) For the 57 Fe absorption variant of Mo limitations are (i) a relatively low abundance of 57 Fe in the natural iron (2.19%), so that samples to be studied containing only traces of Fe have to be enriched with 57 Fe , and (ii) a relatively low sensitivity (i.e., with minimal sample requirements €ssbauer effect (i.e., recoilless absorption or of over 1 mM 57 Fe and a sample volume about 0.3–0.4 ml [2]). As the Mo emission of g-quanta) with a noticeable probability can generally be observed in a solid matrix only, solutions or liquid €ssbauer samples are commonly studied in a rapidly frozen state [3]. This methodology also allows ‘time-resolved’ Mo spectra to be obtained down to a millisecond time scale (using the rapid freeze-quench method [2]). Upon rapidly freezing a solution at a certain time point, which allows its structure to be conserved [3], virtually all ongoing (bio) €ssbauer study of the frozen system chemical processes in the system under study abruptly cease. Thus, a further Mo provides its ‘snapshot’ at the moment of freezing. €ssbauer spectroscopy (EMS; with the 57 Co radionuclide as the most widely used isotope) The emission variant of Mo is incomparably more sensitive and potentially very informative (for earlier and more recent reviews on or including EMSrelated topics see, for example, Refs. 4–10). It requires ca. 103–104-fold lower amounts of 57 Co in a sample under study (which, in this case, is used as a source of g-radiation), as compared with the 57 Fe absorption variant. Thus, with the theoretical specific activity of 57 Co of 8.5 Ci mg 1 (ca. 0.48 Ci nmol 1; see http://www.iem-inc.com/toolspa.html), less than microgram amounts of 57 Co in a sample may well be sufficient for measurements to obtain a good emission spectrum. It has to be mentioned that through the 1960s, there was some concern about the possibility for the M€ ossbauer isotopes with parent nuclei that undergo electron capture or converted isomeric transition (i.e., which lead to high Auger ionization) to be useful for studying matter [6,8]. This was due to some studies that reported disruption of some M€ ossbauer isotope-containing molecules induced by the Auger ionization process [6]. As witnessed by M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

333

334

€ 17 EMISSION (57Co) MOSSBAUER SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS

Professor Amar Nath [6], Nobel Prize winner Willard F. Libby’s comment on that point was, “What can you learn about a crystal vase using a sledge hammer?” Fortunately, this skepticism was not corroborated, as there are ways for the enormous energy (e.g., 6 keV on the 57 Fe nuclei after 57 Co decay), formed on the daughter atom after nuclear transformation, to be dissipated in a solid [8,11]. However, owing largely to the specific methodological difficulties of this sophisticated nuclear chemistry technique (particularly the necessity of using a radioactive isotope in samples under study), the total number of yearly EMS-related publications have tended to decline from a maximum of about a hundred papers per year (reached in the 1970s) down to ca. 20 papers appearing annually by the 21st century [12]. Cobalt as a microelement has a number of biochemical and physiological functions in various organisms (see, for example, Ref. 5 and references reported therein). Despite its evident importance in biology and biomedicine-related fields and the unique sensitivity of 57 Co EMS, applications of this powerful technique in life sciences have so far been featured by a limited number of reports. Several reports on vibration characteristics of the 57 Co-labeled eardrum, basilar membrane and cochlea of the inner ear, and the hammer of the middle ear of various mammals were published within about a decade after 1964 (see Ref. 13, pp. 384–386). Those experiments were based on the sensitivity of the M€ ossbauer  resonant system to Angstrom-range vibration amplitudes [14,15]. While in those earlier studies 57 Co was used merely as an external physical EMS-active label, some contemporary and subsequent 57 Co EMS investigations involved intrinsically biology-related cobalt-containing complexes, for example, 57 Co-labeled cobalamins (vitamin B12 and its analogs) [16–18], porphyrins [16,17], and phthalocyanines [16,17,19]; hemoglobins [20,21] and other heme complexes [22]; and 57 CoII complexes with various nucleotides [23–25]. In Ref. 26, the molecular dynamics in the 57 Co-doped protein concanavalin A was studied using 57 Co EMS. Note that the experiments with the system 57 Co-labeled vitamin B12 coenzyme þ the enzyme ethanolamineammonia lyase [17] represented the very first attempt to apply 57 Co EMS to an enzyme-related system, although in that ossbauer case solely the coenzyme (B12) was labeled with 57 Co. Unfortunately, all the emission spectra and the M€ parameters calculated therefrom for the initial 57 Co-labeled coenzyme B12 and for the system [labeled coenzyme þ enzyme] at various time points, up to the completion of the enzymic reaction, appeared to be virtually the same [17]. (It is ossbauer important to emphasize that a successful enzymic assay of the sample [57 Co]-coenzyme þ enzyme after M€ measurements [17] proved that any radiation damage that might have taken place was not sufficient to affect significantly the biological activity of the system.) The conclusion made in Ref. 17, stating that “it is unlikely that 57 Co emission M€ ossbauer spectroscopy will contribute greatly to the understanding of vitamin B12-dependent enzymes,” although related exclusively to that system, seems to have damped down the enthusiasm of other researchers, who could be interested in applying EMS to enzyme-related systems. Nevertheless, already our first 57 Co EMS experiments involving a 57 Co-doped enzyme, bacterial glutamine synthetase (GS) [27] (to be discussed in more detail below), where an enzyme with 57 Co-doped active centers was for the first time studied using 57 Co EMS, appeared to be successful and gave a new stimulus for further related investigations. Besides that, 57 Co EMS was earlier applied for monitoring the state of cobalt(II) in roots of water hyacinth Eichhornia crassipes [28], as well as in cells of a cyanobacterium (the blue-green alga Synechococcus vulcanus) [29] and Gram-negative bacteria (Escherichia coli [30] and, more recently, Azospirillum brasilense in the freeze-dried state [27,32] or in frozen aqueous suspensions (FAS) [31,32] after their contact with 57 CoII in solution). In this chapter, recent progress is reviewed considering the applications of the 57 Co EMS technique for studying 57 Co-doped enzyme samples as well as 57 CoII binding and transformations in bacterial cells. Further developments, potentials, and future prospects of the 57 Co EMS methodology are also briefly discussed with the aim to broaden the scope of possible applications of this fascinating technique, which has so far remained uncommon in biochemistry, microbiology, and related biological fields.

17.2 METHODOLOGY For the 57 Co radionuclide, the half-life period is 270 days. Its nuclear decay proceeding via electron capture gives the stable 57 Fe isotope and is accompanied by specific physical and chemical aftereffects (various details, consequences, and possible applications of 57 Co radioactive decay aftereffects with regard to 57 Co EMS measurements have been recently discussed, for example, in Refs. 5,6,8,10,11,33–36). The resulting recoil energy (ca. 4.6 eV) for the daughter 57 Fe nucleus is sufficiently low, so that the nucleogenic iron atoms do not shift from their positions, and in many cases their chemical state and environment reflect that of the parent 57 Co species.

335

17.2 METHODOLOGY

However, as a result of the concatenated processes of consecutively filling in the vacancies in inner electronic shells of the 57 Fe atom after the electron capture by the 57 Co nucleus (the so-called Auger cascade developing within 10 15 to 10 14 s), Auger electrons are ejected out of the 57 Fe atom, giving a series of its different, mainly short-lived charge states. By the moment when a 14.4-keV g-quantum is emitted by the nucleus (ca. 10 7 s after the electron capture), most of the Auger electrons return, giving common charge states of the daughter 57 Fe ions. Some of these charge states correspond to those of the parent 57 Co; however, the interaction of some Auger electrons with the chemical environment of the atom can result in the formation of (an)other charge state(s) of the latter. For example, for parent 57 CoII species, along with the corresponding daughter 57 FeII species, some portion of 57 FeIII species can be formed. These aftereffects, which inevitably complicate the emission spectra, yet can provide valuable additional information, for example, on the electron-acceptor properties of the proximal coordination environment around 57 Co [37,38]. For instance, the higher the electron-acceptor properties of the microenvironment of the nucleogenic 57 Fe atom formed upon decay of 57 CoII , the higher the yield of the aliovalent 57 FeIII form. Nevertheless, by the moment of emission of a 14.4 keV g-quantum, commonly a major or substantial part of the resulting 57 Fe ions regain the charge of the parent 57 Co ion, staying essentially in the same coordination microenvironment. Thus, the resulting substance under EMS study may be regarded as a 57 Fe complex in which the nucleogenic 57 Fe ion substitutes for the parent 57 Co ion (retaining both its charge and largely the geometry of the surrounding donor atoms of the ligands). As already mentioned, the recoil-free emission (as well as absorption) of g-quanta (the M€ ossbauer effect) is observed in solids only; therefore, samples of solutions or liquids are usually studied in the rapidly frozen state [3]. Rapid freezing often allows crystallization of the liquid (solvent) to be avoided, so that the structure of the resulting glassy solid matrix represents that of the initial solution. Upon freezing, any chemical reactions in the system as well as biochemical (metabolic) processes in live cells, tissues, or other biological samples cease at a certain point. Thus, freezing a biological system under study after different periods of time allows time-dependent metabolic transformations to be monitored [5,30–32]. On the other hand, in cells or tissues kept above the freezing point, any noticeable M€ ossbauer effect can be detected only for 57 Co species that remain within (or at the surface of) quasisolid parts (e.g., cell membranes) or in a very viscous medium [30]. A typical conventional experimental setup for measuring emission M€ ossbauer spectra is schematically shown in Fig. 17.1 [5]. The 57 Co-containing sample (which in EMS is the source of g-radiation) can be kept in a cryostat (e.g., in liquid nitrogen at T  80 K), whereas the absorber vibrates along the axis “source–absorber” at a constantacceleration value (with its sign periodically changing from þa to a) with a strictly defined time dependence of velocities (usually up to 10 mm s 1 relative to the sample), thus modifying the g-quanta energy scale according to the Doppler effect. Emission M€ ossbauer spectra are measured by placing a 57 Co-containing sample (as a source of g-radiation) in a sample holder of the spectrometer (if necessary, it may be kept in a specially designed cryostat with a window for g-rays) using a conventional constant-acceleration M€ ossbauer spectrometer combined with a PC-operated multichannel analyzer. The spectrometer is calibrated using a standard (e.g., a-Fe foil). In order to obtain suitable statistics, each spectrum is collected for some period of time, usually from a few hours up to several days, depending on the content of 57 Co in the sample and the M€ ossbauer effect [8,9]. Standard PC-based statistical analysis involves fitting the experimental data obtained (for EMS they are commonly converted into a form compatible with that of conventional absorption 57 Fe M€ ossbauer measurements, that is, with isomer shift values for the nucleogenic 57 FeII or 57 FeIII positive with regard to a-Fe) as a sum of Lorentzian-shaped lines using a least squares minimization procedure. The usual set of M€ ossbauer parameters is thus calculated from the experimental data (all isomer shift (d) values discussed in this chapter will be given relative to a-Fe at room temperature).

FIGURE 17.1 Scheme of experimental setup for measuring

57

€ ssbauer spectra. Co emission Mo

336

€ SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS 17 EMISSION (57Co) MOSSBAUER

17.3 MICROBIOLOGICAL APPLICATIONS Cobalt, being required as a trace element for both eukaryotes and prokaryotes, yet can induce toxic effects at higher concentrations (see Refs 5,27,39 and references therein). In particular, when present in the medium, CoII (as well as some other heavy metals) is rapidly accumulated in most bacterial cells by the fast and unspecific membrane-integral protein CorA belonging to the metal-inorganic-transport (MIT) family. In many microorganisms, cobalt(II) is involved in diverse enzymatic activities [5,27] and can be included in intracellular magnetosomes [5]. However, its primary binding to the cell surface in the course of purely chemical and virtually nonselective processes can take place in both live and dead microbial cells [5,27]. Cobalt attracts attention also owing to biogeochemical problems resulting from the microbially mediated migration of the radionuclide 60 Co from disposal sites [27,31]. Such processes can be facilitated by release of cobalt traces entrapped within ferric oxide minerals upon the microbial dissimilatory reduction of iron(III) [22] together with possible Co3þ species [31]. In view of the aforementioned, knowledge of microbiological transformations of cobalt is of importance both from the analytical viewpoint and for its chemical speciation. Virtually the first report on 57 Co EMS measurements in bacteria concerned the accumulation of 57 Co2þ complex with a bacterial hexadentate iron(III)-chelating agent, enterochelin (a cyclic trimer of N-2,3-dihydroxybenzoylserine) in cells of E. coli (strain AN 272) [30]. Measurements were performed both for frozen bacterial samples (at T ¼ 83 K) and above the freezing point (at þ3  C). Interestingly, in the latter case a very weak but distinguishable effect (0.03%, with a width about 3 mm s 1) was detected for 24 h, after which no effect was any longer observed. Disappearance of the effect was ascribed [30] to damage that might have occurred to the cells under the experimental conditions applied. For frozen bacteria, the effect was much higher and resulted in accumulating relatively good spectra within 4 days (no M€ ossbauer parameters were reported). By comparing the data obtained, it was assumed that about 4% of the 57 Co2þ complex absorbed by the bacteria were located within the cell membrane (a quasisolid cellular structure), thus giving some weak but noticeable effect without freezing. A similar EMS study was performed on 57 Co in thermophilic cyanobacteria (blue-green alga, Synechococcus vulcanus) incubated with carrier-free 57 Co2þ salt for 4 months at 55  C [29]. The cell suspension was centrifuged, and live cells were rinsed with 57 Co-free nutrient solution and frozen in liquid nitrogen, as in vivo measurements gave no appreciable effect. The ossbauer measurements on iron species in cells of the alga (cultivated data obtained were compared with absorption (57 Fe) M€ for up to 6 days with 57 FeIII –EDTA complex), which were totally different. The most appropriate fit of the emission spectrum presented in Ref. 29, comprising a superposition of two quadrupole-split components, yielded a major doublet (74% of the total spectral area) with the parameters (isomer shift d ¼ 0.27  0.01 mm s 1, quadrupole splitting DEQ ¼ 2.08  0.01 mm s 1 atliquidnitrogentemperature)similartothoseof 57 Co-dopedcobalamins[16]andrelatedcomplexes.Theminor quadrupole doublet (26% of the total spectral area; d ¼ 1.08  0.01 mm s 1, DEQ ¼ 2.66  0.01 mm s 1) was ascribed to a high-spin ferrous species not found in other 57 Co cobalamins. This component was assumed to be a result of biological reactions involving 57 Co in cyanobacterium cells. Note for comparison that the emission spectrum of the 57 CoII -doped cyanobacteria was found to be different ossbauer parameters for frozen roots of water hyacinth that from that of 57 Co in roots of water hyacinth [28]. The M€ had been cultivated for 2 weeks with 57 CoII salt (a quadrupole doublet of high-spin ferrous component with d ¼ 1.19  0.05 mm s 1, DEQ ¼ 2.66  0.05 mm s 1, 56% of the total spectral area, and an accompanying doublet with d ¼ 0.45  0.1 mm s 1; DEQ ¼ 1.6  0.1 mm s 1, 44% of the total spectral area, as a result of aftereffects) indicated that cobalt is possibly complexed with an EDTA chelating agent from a nutrient solution or another chelating compound of plant origin. Indeed, the M€ ossbauer parameters for 57 CoII –EDTA complex at pH 6.6 in frozen solution (d ¼ 1.17 1  0.05 mm s , DEQ ¼ 2.71  0.05 mm s 1, 80% of the total spectral area, and an accompanying doublet with d ¼ 0.6  0.1 mm s 1, DEQ ¼ 1.6  0.1 mm s 1, 20% of the total spectral area, as a result of aftereffects) [28] are very similar, except for the yield of the aliovalent ferric component (as a result of aftereffects). In its turn, the parameters of the latter are very close to those calculated from the absorption spectrum of the 57 FeIII –EDTA complex in frozen solution at pH 4 (d ¼ 0.46  0.03 mm s 1, DEQ ¼ 1.64  0.03 mm s 1). In the soil diazotrophic rhizobacterium A. brasilense (strain Sp245, reported to be tolerant to submillimolar concentrations of heavy metals, including cobalt(II), in the culture medium [32]), EMS studies were first performed on freeze-dried bacterial samples (measured at T ¼ 80 K; Fig. 17.2) [27]. The following experiments with the same bacteria were performed with live cells rapidly frozen after certain periods of time (2–60 min) of contact with 57 CoII , and EMS spectra were measured for frozen suspensions (without drying), which more closely represent the state of cobalt in the live cells [31,32] (Fig. 17.3) . M€ ossbauer parameters calculated from the experimental data are listed in Table 17.1.

17.3 MICROBIOLOGICAL APPLICATIONS

337

FIGURE 17.2 € ssbauer spectra of Emission Mo dried cells of Azospirillum brasilense (strain Sp245) that were incubated as live cells with 57 CoCl2 for 2 min (a) and 60 min (b) at ambient temperature, then rapidly frozen in liquid nitrogen and freeze-dried prior to EMS measurements (spectra measured at T ¼ 80 K; see also Table 17.1). The positions of spectral components (quadrupole doublets) composing the whole spectrum (solid envelope), obtained by fitting the experimental data (points), are shown above the spectra with square brackets; the same for other figures with emission spectra. (Adapted from Ref. 27.)

For each of the aforementioned cell samples, there were two EMS components corresponding to two chemical forms of high-spin 57 CoII . (The presence of the third component with the parameters typical for high-spin nucleogenic iron(III), stabilized after nuclear decay of the parent 57 Co2þ ion, is evidently a result of aftereffects of the 57 Co ! 57 Fe nuclear transformation, as CoIII is not expected to appear in such systems.) The presence of at least two major cobaltous forms (with different d and DEQ values, see Table 17.1) revealed in the spectra of cells may be related to the availability of different functional groups (also with possibly different donor atoms) as ligands at the bacterial cell surface (see, for example, Refs 31,32 and references therein). It is noteworthy that the main parameters (d, DEQ, and even G) of the corresponding (þ2)-forms for live bacterial cells measured in the dry state and in frozen aqueous suspension were found to be statistically indistinguishable or very close both for 2 min (see Table 17.1, sample 1; cf. frozen aqueous suspension (FAS) and dried biomass (DB)) and for 1 h of contact with 57 Co2þ (sample 2; cf. FAS and DB). This finding shows that freeze-drying does not significantly affect the chemical state of the cobalt(II) species bound at the surface or within the membrane of the bacterium, once their parameters remain virtually unchanged. There was some redistribution of the relative contents of the forms (featured by the corresponding S values; see Table 17.1) in going from the hydrated (FAS) state to the dry (DB) state (samples 1 and 2; note, in particular, the increased S values for 57 Fe3þ in the DB state). This might be related to the changes in the yield of the (þ3) form (resulting from aftereffects) in going from the aqueous medium to the dry biomass (e.g., due to possible changes in the outer coordination sphere of 57 Co2þ complexes occurring upon removal of excess water). Another contribution to this effect could partly be made by nonequal changes in the recoilless emission probability (M€ ossbauer– Lamb factor) for different forms upon drying. In Fig. 17.4, the M€ ossbauer parameters are plotted for different cobalt(II) forms in each sample and for various periods of contact (2 and 60 min) of the live bacteria with 57 CoII , as well as for dead bacterial cells (thermally killed by storing in the medium at 95  C in a water bath) and for the cell-free supernatant liquid [5,31]. The EMS data for different periods (2 min and 1 h) of contact of live bacteria with 57 Co2þ traces essentially differed in the corresponding QS values (see Table 17.1) for the two (þ2) forms [31]. This finding indicated that within an hour, after primary rapid adsorption onto the cell surface, cobalt(II) underwent further transformation, most probably occurring within the cell membrane. Nevertheless, the parameters for live bacteria after 2 min and for dead bacteria were found to be rather close (essentially

338

€ 17 EMISSION (57Co) MOSSBAUER SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS

FIGURE 17.3 € ssbauer spectra of frozen aqueous suspensions of live (a, b) and dead (c) cells of Azospirillum brasilense Emission Mo (strain Sp245) in the culture medium as well as of the cell-free supernatant liquid (d), that were incubated with 57 CoCl2 for 2 min (a) and 60 min (b–d) at ambient temperature and then rapidly frozen in liquid nitrogen (spectra collected at T ¼ 80 K; see also Table 17.1). (Adapted from Ref. 32.)

overlapping). This finding indicates that the mechanism of primary rapid Co2þ sorption by live cells of this strain is similar to the purely chemical binding process occurring at the surface of dead (thermally killed) cells, and is virtually not affected by such hydrothermal treatment. Some differences in the parameters for the stabilized (þ3)-form, resulting from aftereffects, for samples 1 (FAS) and 3 in Table 17.1, and its lower content in sample 3 might be related to some possible changes in the properties of cell-surface biopolymers induced by the hydrothermal treatment. Neither of the (þ2) components found in all of the samples, including cell-free supernatant liquid, corresponded to the aquo complex [57 Co(H2O)6]2þ, as the latter is featured by IS ¼ 1.3–1.4 mm s 1 and QS ¼ 3.3–3.4 mm s 1 in frozen aqueous solutions (see, for example, Ref. 3, Chapters 4 and 5). Note also that the parameters for the cell-free supernatant liquid (from which the bacterial cells had been removed by centrifugation) were clearly different from those for all the other samples (see Fig. 17.4). Thus, in the cell-free supernatant liquid the chemical state of Co2þ trace species (i.e., Co2þ complexes) was different from those in cell samples. In its turn, this shows that in the presence of bacterial cells, Co2þ traces are completely (and, considering sample 1, also rapidly, within 2 min) bound by the cells. In the cell-free culture solution, the most likely ligands that could bind Co2þ are phosphate and possibly malate anions (present in the initial standard growth medium at concentrations of the order of 10 2 M), ammonia (from NH4þ added to the initial medium at 5 mM), as well as various (probably acidic) exopolysaccharides dissolving from the cell surface [31,32]. Note that both malate and NH4þ are gradually consumed by the growing bacteria as sources of carbon and bound nitrogen. Thus, the majority of donor atoms in the first coordination spheres of 57 Co2þ complexes in the cellfree solution are likely to be represented by oxygen (including that of hydration water molecules) and probably nitrogen. It has to be admitted that direct identification of metal complexes in such sophisticated systems as bacterial cells using 57 Co EMS is so far difficult for several reasons. First, there is still lack of experimental EMS data on model cobalt biocomplexes [41], which could facilitate interpretation of EMS parameters. Second, various chemical species with structurally (with regard to the coordination symmetry and arrangement of donor atoms around the cation) and compositionally (with regard to the nature of donor atoms of the ligands) similar coordination microenvironments can give similar M€ ossbauer parameters. Also, in EMS, the Auger electrons formed during the Auger cascade inevitably

339

17.3 MICROBIOLOGICAL APPLICATIONS

€ ssbauer Parametersa Calculated from Emission Mo € ssbauer Spectroscopic Data for Aqueous TABLE 17.1 Mo Suspensions of Live and Dead Cells of Azospirillumbrasilense Sp245 in the 57 CoII -Containing Culture Medium as well as for the Cell-Free Supernatant Liquid, which were Incubated with 57 CoCl2 for Specified Periods of Time at Ambient Temperature and then Rapidly Frozen in Liquid Nitrogen (Spectra Measured at T ¼ 80 K) [27,31,32]) Sample

Stateb

Oxidation Statec

1. Live bacterial cells (frozen 2 min after adding 57 CoCl2 to the culture medium)

FAS

þ2 þ2 þ3 þ2 þ2 þ3 þ2 þ2 þ3 þ2 þ2 þ3 þ2 þ2 þ3 þ2 þ2 þ3

DB

2. Live bacterial cells (frozen 60 min after adding 57 CoCl2 to the culture medium)

FAS

DB

3. Dead bacterial cells (frozen 60 min after adding 57 CoCl2 )

FAS

4. Supernatant liquid (frozen 60 min after adding 57 CoCl2 )

FAS

d d (mm s 1)

DEQ e (mm s 1)

Gf (mm s 1)

S g (%)

1.26(1) 1.20(6) 0.45(5) 1.24(3) 1.14(3) 0.35(5) 1.26(1) 1.16(1) 0.24(2) 1.22(4) 1.00(5) 0.26(5) 1.24(1) 1.17(1) 0.33(3) 1.22(1) 1.21(1) 0.40(6)

3.00(3) 2.23(6) 1.0(1) 3.08(6) 2.35(9) 1.26(8) 2.89(2) 2.03(4) 1.40(3) 2.84(7) 2.03(9) 1.55(7) 3.00(2) 2.18(4) 1.39(6) 3.23(5) 2.48(3) 1.22(3)

0.69(3) 0.65(8) 1.2(1) 0.70(10) 0.83(13) 1.43(12) 0.78(2) 0.73(6) 1.13(6) 0.88(10) 0.5(2) 1.36(16) 0.73(2) 0.76(4) 1.4(1) 0.60(2) 0.80(3) 1.3(2)

44 20 36 19 23 58 51 20 29 38 8 54 44 27 29 24 48 28

a

Errors (in the last digits) are given in parentheses. FAS, frozen aqueous suspension (or frozen aqueous solution in case of sample 4); DB, dried biomass. c Oxidation states of the nucleogenic 57 Fe components stabilized after nuclear decay of the parent 57 CoII . d Isomer shift (relative to a-Fe) converted to the normal absorption convention. e Quadrupole splitting. f Full line width at half maximum (assumed to be equal for both lines of a doublet). g Relative resonant absorption areas (relative error 4%) of the relevant spectral components. b

FIGURE 17.4 € ssbauer parameters—isomer Mo shift (d, mm s 1; relative to a-Fe) and quadrupole splitting (DEQ, mm s 1)—plotted for different forms of 57 CoII in aqueous suspension of live cells of Azospirillum brasilense (strain Sp245) rapidly frozen after (1) 2 min and (2) 60 min of contact with 57 CoII , (3) dead cells (hydrothermally killed at 95  C), and (4) cell-free supernatant liquid (measured at T ¼ 80 K). (Adapted from Ref. 5.)

340

€ SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS 17 EMISSION (57Co) MOSSBAUER

€ ssbauer Parameters Calculated from 57 Co Emission Spectroscopic Data for Live or Dead Cells of A. TABLE 17.2 Mo brasilense (strain Sp7) Incubated with 57 CoCl2 for 2 min or 1 h and then Rapidly Frozen in Liquid Nitrogen (Measured at T ¼ 80 K), in Frozen Aqueous Suspension (s) or Dried (d) [40] Bacterial Cells

Multipleta

d b (mm s 1)

DEQ c (mm s 1)

S d (%)

Live (2 min), s

Doublet 1 Doublet 2 Doublet 1 Doublet 2 Doublet 1 Doublet 2 Doublet 1 Doublet 2 Doublet 1 Doublet 2 Doublet 1 Doublet 2

1.10 0.89 1.16 1.02 1.17 1.00 1.18 1.12 1.14 1.07 1.17 1.04

2.59 2.00 2.84 2.18 2.75 2.13 2.79 1.84 2.93 2.25 2.78 1.95

56 19 35 31 45 28 55 10 43 23 57 7

Live (1 h), s Dead (1 h), s Live (2 min), d Live (1 h), d Dead (1 h), d

Errors: for d, 0.02 mm s 1; for DEQ, 0.05 mm s 1; for S, 7 rel.%. a Main doublets corresponding to daughter FeII forms stabilized after the aftereffects). b Isomer shift (relative to a-Fe at room temperature). c Quadrupole splitting. d Relative resonant absorption area.

57

Co ! 57 Fe nuclear transition (the residual FeIII forms result from

influence to some extent the coordination environment of the nuclide, even if its chemical state remains unchanged, which, in particular, results in some noticeable line broadening in 57 Co emission spectra (as compared to absorption spectra of the corresponding 57 Fe compounds) [33–35]. Recently, comparative 57 Co EMS studies at T ¼ 80 K were carried out also for live and dead (hydrothermally treated) cells of another A. brasilense strain, Sp7, in frozen aqueous suspensions and as dried residues (in Table 17.2, the parameters of the main (þ2)-components are listed) [40]. Note that this bacterium represents a good model for studying its responses to ecological factors. Its strain Sp245 (see above) is an endophyte (capable of penetrating into and colonizing the plant root tissues), whereas strain Sp7 is an epiphyte (colonizing the root surface only). Thus, the two strains occupy different ecological niches and can show marked differences in behavior under similar conditions [32]. Live cells of strain Sp7 rapidly frozen 2 min and 1 h after their contact with 57 Co2þ (measured in frozen suspensions) also showed marked differences in their M€ ossbauer parameters, similarly reflecting metabolic transformations of 57 Co2þ occurring within an hour. However, the parameters for strains Sp245 [31] and Sp7 [40] (see Tables 17.1 and 17.2), obtained under similar conditions, also differ, implying dissimilarities in their metabolic responses to Co2þ detected earlier using FTIR spectroscopy [32].

17.4 ENZYMOLOGICAL APPLICATIONS 17.4.1 Choosing a Test Object The 57 Co EMS technique has recently been shown for the first time to be applicable to probing cation-binding sites of the enzyme active centers [27]. As the first example of a metal-dependent enzyme tested with 57 Co EMS [27,42], a bacterial glutamine synthetase (GS, EC 6.3.1.2; a key enzyme of nitrogen metabolism that is ubiquitous in all organisms from bacteria to humans) was used. For instance, in humans, GS is expressed in tissues, being involved in ammonia detoxification and interorgan nitrogen flux; its malfunctioning or inherited deficiency was found to lead to multiple disorders of organs and brain malformation [43]. There were logical reasons for choosing this particular enzyme, isolated from the soil diazotrophic bacterium A. brasilense, for testing the validity of data obtained by the EMS technique. A few main reasons are considered below. First, cobalt(II) is among its activating divalent cations (cofactors), along with Mg2þ and Mn2þ [44]. It was found that the enzyme could be obtained in a cation-free state (by removing the native cation(s) bound in the active centres using EDTA treatment with subsequent dialysis), in which it loses its biocatalytic activity in the absence of divalent cations in the

17.4 ENZYMOLOGICAL APPLICATIONS

341

medium. By adding Co2þ to the inactive metal-free GS it was possible to reinstate its activity, thereby proving that the added Co2þ cations bind in the active centres (the latter is prerequisite for the GS activity to be expressed) [44,45]. Thus, doping the cation-free GS with a precalculated amount of 57 Co2þ under physiologically similar conditions in principle allows 57 Co EMS to be applied. Second, this enzyme has two cation-binding sites (n1 and n2) at each of its 12 active centers (i.e., total 24 sites) per molecule [44]. From the structural point of view, a molecule of bacterial GS is a dodecamer formed from two hexameric rings of subunits (monomers) disposed face-to-face (Fig. 17.5a); each of the 12 active centers (having the shape of a ‘bifunnel’; Fig. 17.5b) is located between every two neighboring subunits within a ring. Site n1 has ca. 50-fold greater affinity for cations than site n2, which is related to their coordination modes and charge distribution [42]. In site n1, the cation is coordinated by three glutamic acid (Glu) residues (E131, E212, and E220), that is, by their three side-chain carboxylates (with possible additional binding of H2O), while in site n2 by two Glu (E129 and E357) and one histidine, H269 (i.e., one N-donor atom of the His heterocycle and two Glu carboxyls), and this structure is strictly conserved among different GSs [43].

FIGURE 17.5 Schematic presentation of (a) a top view of one of the two hexameric rings of a bacterial glutamine synthetase molecule (the position of an active center in an intersubunit space is shown by white circle); (b) a side-view of the active center (‘bifunnel’, ca. 1.5 nm wide and 4.5 nm high, with metal ions in two cation-binding sites, n1 € ssbauer spectrum and n2), which ATP and glutamate enter at opposite ends as shown; and (c) an emission Mo of cation-free GS from Azospirillum brasilense (in an average adenylylation state of 18%) doped with 57 Co2þ (measured in rapidly frozen aqueous solution; T ¼ 80 K) [42]. The subspectra featuring sites n1 (light-shaded doublet) and n2 (dark-shaded doublet) are indicated by accordingly marked square brackets above the spectrum (the third broader doublet of lower intensity is due to aftereffects of the 57 Co ! 57 Fe nuclear transformation [27]). Relative resonant absorption of g-quanta (in percent) by a standard absorber (K4Fe(CN)6
3H2O) is plotted against relative velocity of the absorber versus the 57 Co-containing sample (g-radiation source) [8], which corresponds to the energy scale according to the Doppler effect (1 mm s 1 corresponds to 48.1 neV), calibrated using a-Fe at room temperature. (Adapted from Ref. 43.)

342

€ SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS 17 EMISSION (57Co) MOSSBAUER

The cations in these two sites are 0.6 nm apart, have no common ligands (i.e., not bridged) and therefore, from the spectroscopic point of view, may be regarded as relatively independent from each other. Thus, sites n1 and n2, when both ossbauer parameters in an emission spectrum (i.e., occupied by 57 Co2þ cations, could be expected to exhibit different M€ giving distinguishable spectral components) owing to their essentially different coordination microenvironments. Moreover, since the affinity for cations is much higher in site n1, in case when the molar 57 Co2þ -to-GS ratio (x) is made higher than that required to saturate half of all the sites (i.e., primarily sites n1) but under the ‘saturation limit’ for both n1 þ n2 (i.e., under the condition 12 < x < 24), the resulting ratio of the areas for the subspectra (spectral components) corresponding to n1 and n2 could be expected to allow 57 Co2þ distribution between the sites to be assessed (Fig. 17.5c) [45]. 17.4.2 Prerequisites for Using the 57 Co EMS Technique In order for the enzyme, with its active centers doped with 57 Co2þ , to be useful in EMS, a correct analysis of the EMS data to be obtained is necessary. Thus, several conditions have to be observed [45]. First, it is important to make sure that the 57 Co2þ cations, substituting for the native metal(II) ions [44], are indeed bound within the active centers. If this is not so, the existence of numerous binding sites and, consequently, many forms of 57 CoII bound to functional groups of the protein macromolecule would render the EMS data hardly interpretable. Second, the replacement of the native activating cations (e.g., by natural Co2þ under identical conditions) ideally should not result in an irreversible deactivation of the enzyme. In the latter case, the correspondence between the 57 CoII form in the enzyme sample under study and the cobalt(II) form in the physiologically active enzyme would be doubtful. Finally, the quantity of the substituted 57 Co2þ should conform to the overall number of the cation-binding sites in the enzyme sample. Any excessive 57 Co2þ , binding to different functional groups of the protein beyond the active centers, would evidently lead to an unpredictable complication of the spectra. As the affinity for cations in the enzyme active centers is usually much higher than elsewhere on the protein globule, the above-mentioned conditions may well be feasible. Moreover, when the active center contains more than one cationbinding site with different affinities for the cation (but high, as compared to any other possible binding sites beyond active centers) and different coordination environments (as in the case with glutamine synthetase [43]), it may be expected that 57 Co EMS would allow one to obtain information not only on the chemical forms and coordination symmetry of the cobalt in each of the sites but also on its distribution between them. The above-discussed properties of A. brasilense GS were found to be suitable for using EMS in studying 57 Co2þ -doped enzyme preparations [5,7,27,45]. 17.4.3 Experimental 57 Co EMS Studies Analysis of the emission M€ ossbauer spectra of [57 Co]-cobalt(II)-doped GS (in a low average adenylylation state, 18%) both in rapidly frozen aqueous solution [27] and in the solid state [42] indeed revealed the presence of two forms of cobalt(II) with different affinities (in view of nonequal distribution of 57 CoII between the forms; Fig. 17.6) as well as with different coordination reflected by different M€ ossbauer parameters (Table 17.3) . However, each of the relevant forms in dissolved and solid enzyme showed similar (significantly overlapping) parameters. This is indicative of a close similarity of the cobalt(II) coordination structure in enzyme’s active centers both in solution and in the dry state, which reflects the well-known fact of stability of the enzyme upon drying. The values of isomer shifts (d ¼ 1.08 and 1.05–1.07 mm s 1 relative to a-Fe) and quadrupole splittings (DEQ ¼ 3.0–3.1 and 2.3–2.4 mm s 1; see Table 17.3) obtained for the two doublets allowed those components to be correlated with the two cation-binding sites in the GS active center (sites n2 and n1, respectively [42–44]). As mentioned above, these sites of bacterial GSs have different coordination environments, as well as a correspondingly lower (for site n2) and higher (for site n1) affinity to the cation. The latter difference is in line with the nonuniform distribution of 57 CoII between the spectral components (as the areas of quadrupole doublets 1 and 2 in each spectrum corresponding to different 57 CoII forms are significantly different; see Fig. 17.6). It may be reasonably supposed that the nucleogenic 57 FeII form with a lower DEQ ¼ 2.3–2.4 mm s 1 corresponds to the site with more symmetrically coordinated O-donor ligands, that is, site n1 (see Fig. 17.5b). In this case, the higher DEQ value (ca. 3.0–3.1 mm s 1) may reflect a lower coordination symmetry owing to different donor atoms including, along with oxygens, also the His-269 nitrogen (the d values are close in these cases). The relatively low IS values for both the 57 CoII forms in GS with a low (18%) adenylylation state of its subunits (IS ¼ 1.05 and 1.08 mm s 1 at T ¼ 80 K; Table 17.3) may indicate a tetrahedral (Td) symmetry of cobalt(II) coordination. In this case, the coordination mode of all the Glu residues must be monodentate, which is often observed for cation-binding

343

17.4 ENZYMOLOGICAL APPLICATIONS

FIGURE 17.6 € ssbauer spectra of Emission Mo cation-free glutamine synthetase (GS; average adenylylation state 18%) from Azospirillum brasilense Sp245 incubated with 57 Co2þ for 60 min at ambient temperature, measured (a) in rapidly frozen aqueous solution and (b) as a dried solid (T ¼ 80 K; see also Table 17.3). (Adapted from Ref. 42.)

sites in metalloproteins (note that there is also at least one water molecule as a ligand) [43]. For comparison, the d values reported for several high-spin carboxylate-rich FeII complexes (some with N-donor atoms, measured at T ¼ 4.2 K) are within the range (1.04–1.08)  0.02 mm s 1 for the Td geometry of the sites and (1.26–1.35)  0.02 mm s 1 for the octahedral (Oh) geometry [46]. Complexes of 57 CoII with anthranilic (o-aminobenzoic) acid and L-tryptophan, studied using 57 Co EMS in frozen aqueous solutions, showed, respectively, d ¼ 1.10  0.02 and 0.88  0.06 mm s 1, € ssbauer Parametersa for 57 Co2þ -Doped Glutamine Synthetase (GS, Average Adenylylation State TABLE 17.3 Mo € ssbauer Spectra in Rapidly 18%) from Azospirillumbrasilense Sp245 Calculated from Emission Mo Frozen Aqueous Solution and in the Solid State (Measured at T ¼ 80 K) [42] Sample

O.S.b

1. Cation-free GS incubated with 57 Co2þ for 60 min; frozen in liquid nitrogen (solution)

þ2 þ2 þ3 þ2 þ2 þ3

2. Cation-free GS incubated with 57 Co2þ for 60 min, frozen and dried (solid)

d c (mm s 1)

DEQ d (mm s 1)

Ge (mm s 1)

S f (%)

1.08(0.02) 1.05(0.02) 0.34(0.10) 1.08(0.02) 1.07(0.02) 0.39(0.10)

3.08(0.08) 2.39(0.06) 1.12(0.20) 2.97(0.06) 2.29(0.06) 1.37(0.20)

0.48(0.05) 0.75(0.08) 1.25(0.30) 0.52(0.12) 0.66(0.08) 1.19(0.27)

18(1) 60(1) 22(1) 24(1) 50(1) 26(1)

a

Relative errors are given in parentheses. Oxidation state for the nucleogenic 57 Fe components stabilized after nuclear decay of the parent c Isomer shift (relative to a-Fe; converted to the normal absorption convention). d Quadrupole splitting. e Full linewidth at half maximum. f Relative resonant absorption areas of the relevant spectral components. b

57

CoII .

344

€ SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS 17 EMISSION (57Co) MOSSBAUER

DEQ ¼ 2.71  0.05 and 2.8  0.1 mm s 1 [41], which is also within the range for high-spin Td coordination. Note that solid FeII anthranilate Fe(anthr)2 with an Oh coordination (with two bidentate carboxylates and two amino groups as ligands, that is, 4O þ 2N donor atoms) gave d ¼ 1.25  0.01 mm s 1 (at T ¼ 80 K) [41]. It may be expected that any intermediate coordination state of the activating cations, for example, occurring during substrate turnover in the course of the enzymatic reaction, could be distinguished using a rapid freeze-quench variant [2] applied to 57 Co EMS. An extremely important issue for the 57 Co EMS methodology is the study of aftereffects of the 57 Co ! 57 Fe radioactive decay process. For a complicated biocomplex, for example an enzyme, especially with more than one cationbinding site (n > 1) at the active center, each of the parent 57 CoII chemical species (with its own coordination microenvironment) would, in principle, give its own yield of the aliovalent 57 FeIII form (as a result of aftereffects). It is often the case when such daughter 57 FeIII lines in the source experiment have very close parameters and, being also relatively broad, virtually do not resolve in the spectrum and thus have to be fitted with a single doublet. Since the partial yields for the n aliovalent 57 FeIII forms remain unknown, the real distribution of 57 CoII between the n sites in the sample under study is therefore also unknown, even when the corresponding n lines for 57 FeII do resolve and can easily be quantitatively assessed. It has been shown that if a redistribution of 57 CoII between n sites can be experimentally achieved, then solving a system of equations based on the material balance and using the data of n experimental 57 Co EMS measurements allow both the partial yields of 57 FeIII (aftereffects) for each parent 57 CoII form and the real distribution of 57 CoII between the sites to be obtained [36]. This new methodology was applied to the aforementioned 57 CoII -doped bacterial glutamine synthetase (with n ¼ 2 different cation-binding sites at its active centers). For that, the experimental data were used on the redistribution of 57 Co2þ between the sites found upon adding Mn2þ together with 57 Co2þ as a result of their competitive binding [47]. Moreover, it was of interest to compare the calculated yields of 57 FeIII for the two enzyme sites (with coordinated carboxylic groups of three Glu amino acid residues and with two Glu þ one His residues, respectively) with those for some 57 CoII –(aromatic amino acid) complexes in their emission spectra [41]. It was found that some functional groups—potential electron acceptors (aromatic cycles within the ligands coordinated to 57 CoII ), located at various distances from parent 57 Co2þ ions (e.g., up to direct coordination via a N heteroatom), respectively, influence the yield of 57 FeIII due to aftereffects. This is in line with a virtually large yield of aliovalent 57 FeIII in emission spectra of 57 CoII –nucleotide complexes [23–25], in which the cation is directly bound to the N7 of the purine ring. Such studies facilitate understanding the appearance and regularities of aftereffects, which is necessary for adequately interpreting EMS data on sophisticated biocomplexes and for further successful biological and biochemical EMS applications. 17.4.4 Two-Metal-Ion Catalysis: Competitive Metal Binding at the Active Centers Besides GS (both type I and type II), a number of other metal-dependent enzymes including, for example alkaline phosphatases, endonucleases, DNA/RNA polymerases, and so on, require two metal cations bound in the active center for their catalytic activity [43]. It has, however, to be noted that while, for instance, in polymerases and nucleases the two cations are coordinated jointly by both a conserved Asp residue and the scissile phosphate, in GS the two cations are bound by a few separate amino acid residues (see Fig. 17.5b). Moreover, the GS substrates bind to different cations (glutamate to n1 and ATP to n2; NH4þ has its separate binding site close to n1). This is in line with a relatively large distance between the two cations in GS (0.6 nm), as compared to those in, for example, RNases (0.4–0.3 nm) [43]. For A. brasilense GS, the possibility of the simultaneous binding of two different metal cations at sites n1 and n2 has so far been unclear. As for the terminology (keeping in mind also the possible different modes of coordination of the cations both to protein residues and to substrates), two-metal-ion binding may be related to binuclear catalysis. Thus, binding of the same metal cation in both sites of an active center may be related to homobinuclear catalysis, whereas different metal cations binding at the two sites of an active center may be related to heterobinuclear catalysis. Preliminary comparative 57 Co EMS experimental studies were also performed on partly adenylylated GS from A. brasilense (in an intermediate adenylylation state of 44%) doped with 57 Co2þ alone or together with natural Mn2þ (the latter gives no contribution to emission spectra) [47]. Adding equimolar quantities of 57 Co2þ þ Mn2þ to the metal-free enzyme, as compared to 57 Co2þ alone, was found to result in a partial redistribution of 57 CoII between the sites, showing competitive binding of the two cations. This important finding, indirectly pointing to the possibility of heterobinuclear catalysis in bacterial GS, is in agreement with similar efficiency of the two cations in supporting the activity of partly adenylylated GS from A. brasilense [44]. Further 57 Co EMS studies are therefore warranted in order to compare not only the enzyme’s specificity for different cations, but also their ‘preference’ for each of the two different sites (n1 and n2),

ACKNOWLEDGMENTS

345

possible competitive binding, and redistribution within the GS active centers. Moreover, the coordination symmetry of 57 CoII at both sites in GS from A. brasilense was found to be altered by changing the GS adenylylation state [47], which provides information as to the mechanism of the enzyme activity regulation by covalent modification. 17.4.5 Possibilities of 57 Co Substitution for Other Cations in Metalloproteins Cobalt(II) acts as an activating cation or a cofactor in a variety of enzymes, which could thus be directly probed using 57 Co doping and EMS measurements. Moreover, Co2þ has also been shown to be applicable as an isostructural substitute, for example for Zn2þ in many zinc-containing proteins [43], and even to enhance enzyme’s beneficial properties (see, for example, Refs 48,49 and references therein). Besides the applicability of Co2þ as an optical (UV–Vis) probe that gives clearly different spectra for, for example, Td and Oh coordination, 57 Co2þ substitution for other cations can give a highly sensitive ‘snapshot’ of the coordination microenvironment and its fine structural changes when used in 57 Co EMS. Such possibility could greatly expand the areas of 57 Co EMS applications beyond only cobalt-dependent metalloproteins and enzymes, greatly enhancing the importance of the technique for biochemistry, molecular biology, and related fields.

17.5 CONCLUSIONS AND OUTLOOK In this chapter, besides an introduction to the topic and a brief consideration of earlier work, an overview is given on 57 Co EMS-assisted probing of the interaction of 57 Co2þ traces with live and dead bacterial cells in aqueous media. The M€ ossbauer parameters obtained in rapidly frozen suspensions of bacterial cells showed similarity for 57 CoII bound to dead biomass and for 57 CoII rapidly bound by the surface of live cells, reflecting the formation of similar chemical species. For live cells, however, the parameters of 57 Co emission spectra were found to change within an hour, which reflected ongoing metabolic transformations of the cation. The data obtained correlate with the recently discovered involvement of Co2þ in reactions with labile [Fe S] clusters during their de novo biosynthesis or repair in E. coli [39], presenting the molecular basis for Co2þ toxicity in bacteria, besides Co2þ-induced oxidative stress. The results obtained show that 57 Co EMS can provide unique information both for speciation bioanalysis and for the monitoring of radionuclide bioleaching and its environmentally important chemical transformations mediated by soil microorganisms. 57 Co EMS measurements have shown for the first time that this nuclear chemistry technique is sensitive (i) to differences in the coordination of CoII cations at different cation-binding sites within the enzyme active centers and (ii) to differences in the binding affinity of the sites to 57 CoII . Further recent studies have also revealed 57 Co EMS sensitivity (iii) to the effects of competitive binding of different activating cations inducing 57 CoII redistribution between the sites and (iv) to fine structural changes in 57 Co(II) coordination induced by covalent modifications of the enzyme molecule (related to its activity regulation). The results obtained are highly promising for studying the molecular mechanisms of enzyme– substrate biospecific interactions using the unique possibilities of EMS.

ACKNOWLEDGMENTS The author is deeply grateful to Professor Yu.D. Perfiliev and Dr L.A. Kulikov (M.V. Lomonosov Moscow State University, Moscow, Russia) as well as to Professors A. Vertes, E. Kuzmann, and Z. Homonnay (E€ otv€ os Lorand University, Budapest, Hungary) for long-lasting collaboration, hospitality, many stimulating discussions, and their invaluable help in experimental work. Support of the author’s studies, mentioned in this chapter, from the EC (INTAS Grant 96-1015), under the NATO “Security Through Science” Program (Collaborative Linkage Grants LST.CLG.977664, LST.NR.CLG.981092, and ESP. NR.NRCLG 982857; Expert Visit Grants LST.EV.980141 and CBP.NR.NREV 981748), under the Agreements on Scientific Cooperation between the Russian and Hungarian Academies of Sciences, in particular, for 2005–2007 (Protocol, Thematic Plan, Project 37), for 2008–2010 (Protocol, Thematic Plan, Project 45), and for 2011–2013 (Protocol, Thematic Plan, Projects 28 and 29), as well as from the Russian Academy of Sciences’ Commission for the Work with Young Scientists under the 6th Competition-Expertise of research projects (Grant # 205) and from the Russian Foundation for Basic Research (Grant # 13-04-01538a) is gratefully acknowledged.

346

€ SPECTROSCOPY: BIOLOGY-RELATED APPLICATIONS, POTENTIALS 17 EMISSION (57Co) MOSSBAUER

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

A.A. Kamnev, K. Kovacs, I.V. Alenkina, M.I. Oshtrakh, this book, 2013 (Chapter 13). C. Krebs, J.C. Price, J. Baldwin, L. Saleh, M.T. Green, J.M. BollingerJr., Inorg. Chem. 2005, 44, 742–757. A. Vertes, D. Nagy, eds., M€ossbauer Spectroscopy of Frozen Solutions, Akademiai Kiad o, Budapest, 1990. D.L. Nagy, Hyperfine Interact. 1994, 83, 9–13. A.A. Kamnev, J. Mol. Struct. 2005, 744–747, 161–167. A. Nath, M€ossbauer Effect Ref. Data J. 2007, 30, 107–109. Yu.D. Perfiliev, A.A. Kamnev, M€ossbauer Effect Ref. Data J. 2007, 30, 121–122. Yu.D. Perfiliev, Bull. Russ. Acad. Sci.: Phys. 2010, 74, 321–325. P. G€utlich, Y. Garcia, J. Phys.: Conf. Ser. 2010, 217, 012001. A. Nath, J. Nucl. Radiochem. Sci. 2010, 11, A1–A3. A. Nath, M.J. Prushan, J.G. Gilbert, J. Radioanal. Nucl. Chem. 2001, 247, 589–591. Editor’s comments, M€ossbauer Effect Ref. Data J. 2007, 30, 89–90. A. Vertes, L. Korecz, K. Burger, M€ossbauer Spectroscopy (Studies in Physical and Theoretical Chemistry, Vol. 5). Elsevier/Akademiai Kiad o, Amsterdam/Budapest, 1979. P. Hillman, H. Schechter, M. Rubinstein, Rev. Mod. Phys. 1964, 36, 360. L. Robles, W.S. Rhode, C.D. Geisler, J. Acoust. Soc. Am. 1976, 59, 926–939. A. Nath, M. Harpold, M.P. Klein, W. K€undig, Chem. Phys. Lett. 1968, 2, 471–476. D.J. Cardin, K.N. Joblin, A.W. Johnson, G. Lang, M.F. Lappert, Biochim. Biophys. Acta 1974, 371, 44–51. S. Tyagi, K. Inoue, A. Nath, Biochim. Biophys. Acta 1978, 539, 125–134. T.S. Srivastava, J.L. Przybylinski, A. Nath, Inorg. Chem. 1974, 13, 1562–1564. P.P. Craig, N. Sutin, Phys. Rev. Lett. 1963, 11, 460–462. T.S. Srivastava, S. Tyagi, A. Nath, Proc. Natl. Acad. Sci. USA 1977, 74, 4996–5000. L. Marchant, M. Sharrock, B.M. Hoffman, E. M€unck, Proc. Natl. Acad. Sci. USA 1972, 69, 2396–2399. A. Caubet, A. Labarta, R. Rodrıguez, A. Fernandez, E. Molins, V. Moreno, J. Tejada, Hyperfine Interact. 1986, 28, 773–776. A. Caubet, V. Moreno, A. Labarta, J. Tejada, J. Inorg. Biochem. 1990, 39, 173–186. A. Caubet, V. Moreno, A. Labarta, J. Tejada, Z. Naturforsch. 1993, 48, 98–106. E.R. Bauminger, I. Nowik, J. Yariv, Hyperfine Interact. 1990, 58, 2337–2341. A.A. Kamnev, L.P. Antonyuk, V.E. Smirnova, O.B. Serebrennikova, L.A. Kulikov, Yu.D. Perfiliev, Anal. Bioanal. Chem. 2002, 372, 431–435. S. Ambe, F. Ambe, T. Nozaki, Int. J. Appl. Radiat. Isot. 1985, 36, 7–11. S. Ambe, Hyperfine Interact. 1990, 58, 2329–2335. E. Giberman, Y. Yariv, A.J. Kalb, E.R. Bauminger, S.G. Cohen, D. Froindlich, S. Ofer, J. Phys. Colloq. 1974, 35(Suppl. 12), C6-371–C6374. A.A. Kamnev, L.P. Antonyuk, L.A. Kulikov, Yu.D. Perfiliev, BioMetals 2004, 17, 457–466. A.A. Kamnev, A.V. Tugarova, L.P. Antonyuk, P.A. Tarantilis, L.A. Kulikov, Yu.D. Perfiliev, M.G. Polissiou, P.H.E. Gardiner, Anal. Chim. Acta 2006, 573–574, 445–452. Yu.D. Perfiliev, V.S. Rusakov, L.A. Kulikov, A.A. Kamnev, K. Alkhatib, J. Radioanal. Nucl. Chem. 2005, 266, 557–560. Yu.D. Perfiliev, V.S. Rusakov, L.A. Kulikov, A.A. Kamnev, K. Alkhatib, Hyperfine Interact. 2006, 167, 881–885. V.M. Byakov, Yu.D. Perfiliev, L.A. Kulikov, S.V. Stepanov, Mosc. Univ. Chem. Bull. 2009, 64, 259–263. A.A. Kamnev, Seventh International Conference on Nuclear and Radiochemistry, August 24–29, 2008, Budapest, Hungary, Book of Abstracts, 2008, p. 48. Yu.D. Perfiliev, L.A. Kulikov, L.T. Bugaenko, A.M. Babeshkin, M.I. Afanasov, J. Inorg. Nucl. Chem. 1976, 38, 2145–2148. M.I. Afanasov, A.M. Babeshkin, L.A. Kulikov, Yu.D. Perfiliev, J. Radioanal. Nucl. Chem. Lett. 1985, 93, 37–42. C. Ranquet, S. Ollagnier-de-Choudens, L. Loiseau, F. Barras, M. Fontecave, J. Biol. Chem. 2007, 282, 30442–30451. A.A. Kamnev, A.V. Tugarova, K. Kovacs, E. Kuzmann, B. Bir o, P.A. Tarantilis, Z. Homonnay, Anal. Bioanal. Chem. 2013, 405, 1921– 1927. A.A. Kamnev, L.A. Kulikov, Yu.D. Perfiliev, L.P. Antonyuk, E. Kuzmann, A. Vertes, Hyperfine Interact. 2005, 165, 303–308. A.A. Kamnev, L.P. Antonyuk, V.E. Smirnova, L.A. Kulikov, Yu.D. Perfiliev, I.A. Kudelina, E. Kuzmann, A. Vertes, Biopolymers 2004, 74, 64–68.

REFERENCES

347

43. A.A. Kamnev, in Metal Ions in Biology and Medicine, Vol. 10, Ph. Collery, I. Maymard, T. Theophanides, L. Khassanova, T. Collery, eds., John Libbey Eurotext, Paris, 2008, pp. 522–527. 44. L.P. Antonyuk, V.E. Smirnova, A.A. Kamnev, O.B. Serebrennikova, M.A. Vanoni, G. Zanetti, I.A. Kudelina, O.I. Sokolov, V.V. Ignatov, BioMetals 2001, 14, 13–22. 45. A.A. Kamnev, L.P. Antonyuk, L.A. Kulikov, Yu.D. Perfiliev, E. Kuzmann, A. Vertes, Bull. Russ. Acad. Sci.: Phys. 2005, 69, 1561–1565. 46. S. Yoon, S.J. Lippard, J. Am. Chem. Soc. 2005, 127, 8386–8397. 47. A.A. Kamnev, L.P. Antonyuk, V.E. Smirnova, L.A. Kulikov, Yu.D. Perfiliev, E. Kuzmann, A. Vertes, FEBS J. 2005, 272 (Suppl. 1), 10. 48. E.M. Stone, E.S. Glazer, L. Chantranupong, P. Cherukuri, R.M. Breece, D.L. Tierney, S.A. Curley, B.L. Iverson, G. Georgiou, ACS Chem. Biol. 2010, 5, 333–342. 49. A. Srivastava, A.K. Sau, IUBMB Life 2010, 62, 906–915.

PA R T V

IRON OXIDES

C H A P T E R 1 8

€ MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

9 IL 9I PECHOU S9EK, AND RADEK ZBO R 9I FRYDRYCH, JI R 9 , LIBOR MACHALA, JI R JI R9I TUCEK

Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic

18.1 INTRODUCTION Iron oxides are being classified as materials with a very high application potential, especially when sizes of their particles fall below 100 nm [1]. Beside their pioneering utilization in the field of information storage [2], their application importance has been empowered by their recent introduction to various branches of medicine where they have been found as essential tools for diagnostic and treatment purposes [3–7]. Their dominant role in medicine applications springs from their eminent magnetic (e.g., superparamagnetism, strong magnetic response under external magnetic fields) and biochemical (e.g., nontoxicity, biocompatibility, biodegradability) properties they exhibit. In addition, iron oxides are frequently considered as model systems exploited for theoretical explanation of magnetism-based phenomena (e.g., quantum tunneling of magnetization, interparticle magnetic interactions) occurring solely in the nanoworld [8]. In general, the natural abundance of iron oxides is rich [2,9]. They can be commonly found in rocks and soils where they have diverse sizes and morphologies. On the other hand, it is not difficult to prepare them in a synthetic way; nowadays, numerous synthetic routes toward various iron oxide phases have been established enabling a sophisticated control of their size, morphological, structural, and magnetic aspects (for a review of chemical syntheses toward various iron oxide phases, see Ref. 2). Among various existing iron oxide-producing synthetic routes, thermally induced solidstate reactions are frequently exploited [10] due to a variety of suitable precursors for iron oxide synthesis, high yield of products, and possibility to get different iron oxide phases by controlling synthesis variables (temperature, reaction atmosphere, humidity, weight of iron-containing precursors, etc.). Iron oxide products (i.e., iron oxide nanoparticles) of solid-state reactions are, to some extent, predisposed to form agglomerates that can be eliminated by subsequent ultrasonification and coverage of iron oxide nanoparticles by a suitable surfactant. Application of a high temperature can also lead to undesirable broadening of particle size distribution owing to sintering of particles. Nevertheless, thermal syntheses represent simple and cost-effective methods to prepare relatively large amounts of iron oxides with desired properties and often in the form of nanoparticles.

M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

351

352

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

In order to distinguish and characterize various iron oxide phases, 57 Fe M€ ossbauer spectroscopy turns out to be a very powerful tool [11] allowing to monitor the local environment of the iron atoms in the crystal lattice. Hyperfine parameters, acquired from the analysis of the spectral line positions, such as the isomer shift (d), quadrupole splitting (DEQ), quadrupole shift (eQ), and hyperfine magnetic field (Bhf), yield important information on the electronic density, its ossbauer probed nucleus. Analyzing the widths of spectral lines, their symmetry, and magnetic properties at the 57 Fe M€ relative intensities, asymmetry of the spectrum, and temperature and field dependence of the hyperfine parameters also allows deducing valuable material characteristics. Valence and spin states of iron, quantification of nonequivalent iron sites in the crystal lattice, coordination of iron in its individual positions, level of ordering and stoichiometry, type of the magnetic ordering, orientation of the magnetic moments in external magnetic fields (i.e., spin canting and spin frustration), magnetic anisotropy, and magnetic transition temperature constitute the principal information that can ossbauer extracted from the temperature-dependent and in-field M€ ossbauer spectra. In the case of iron oxides, 57 Fe M€ spectroscopy, applied in a broad range of temperatures and intensities of an external magnetic field, provides a distinct separation of individual spectral components belonging to individual iron oxide phases due to their different crystal structure and magnetic behavior reflected in their characteristic temperature- and field-dependent M€ ossbauer hyperfine parameters [10]. ossbauer spectroscopy in the study of nanocrystalline iron(III) In this chapter, we report on exploitation of 57 Fe M€ oxide polymorphs and amorphous Fe2O3 synthesized through thermally induced solid-state reactions. In Section 18.2, various already described polymorphs of iron(III) oxide are discussed in details with respect to their structural and magnetic properties and how they are reflected in their characteristic M€ ossbauer spectra recorded in a broad range of temperatures and inductions of external magnetic field. Section 18.3 focuses on applicability of M€ ossbauer spectroscopy in monitoring solid-state reaction mechanisms that lead to iron(III) oxide phases. Finally, Section 18.4 provides an overview on various types and modifications of M€ ossbauer technique that enable to acquire a complex knowledge on physicochemical properties of nanosized application-promising iron(III) oxide phases.

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES, AND POLYMORPHOUS PHASE TRANSFORMATIONS Nowadays, six different crystalline forms of iron oxide are recognized (see Fig. 18.1) [2,10]. In iron oxides, iron can be in only trivalent (Fe3þ) state, only divalent (Fe2þ) state, or in both divalent and trivalent state. FeO (a black material mineralogically known as w€ ustite) contains only divalent iron and is very frequently nonstoichiometric with oxygen

FIGURE 18.1 Classification of iron oxides.

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

353

deficiency [2,12]. It exhibits a defective NaCl crystal structure (a cubic crystal structure with Fm3m space group) and enters an antiferromagnetic state below 210 K (depending on crystal defects). It is highly thermodynamically unstable and is often found as an important intermediate in the reduction of iron ores [2,12]. On the other hand, iron(III) oxide (or ferric oxide, Fe2O3) contains only trivalent iron and exists in both amorphous and crystalline form [2,10]. To date, four different crystalline forms of Fe2O3 have been already described: (i) a-Fe2O3; (ii) b-Fe2O3; (iii) g-Fe2O3; and (iv) e-Fe2O3. In all iron(III) oxide polymorphs, Fe3þ ion is in a high spin state (i.e., S ¼ 5/2). Finally, Fe3O4 (a black material mineralogically known as magnetite) contains both Fe2þ and Fe3þ. It possesses a cubic crystal structure (Fd3m space group) of an inverse spinel type establishing two crystallographically nonequivalent cation sites (i.e., tetrahedral and octahedral sites) forming two interpenetrating magnetic sublattices [2,13]. Fe3þ ions solely occupy the tetrahedral sites of the Fe3O4 crystal structure, whereas the octahedral sites are filled with both Fe2þ and Fe3þ ions in equal portion (i.e., Fe2þ/Fe3þ ¼ 0.5). Below 850 K, Fe3O4 enters a ferrimagnetic state manifested by relatively high values of spontaneous magnetization responsible for its strong magnetic response when exposed to an external magnetic field. From the historical aspects of magnetism, it is the oldest magnetic material studied. In addition, at 120 K (the Verwey transition), electron hopping mechanism between Fe2þ and Fe3þ ions, sitting at the octahedral sites of the Fe3O4 crystal structure, freezes invoking a change in electric properties of Fe3O4 (metal-to-insulator transition) [13,14]. The Verwey transition is also accompanied by a change in the Fe3O4 crystal structure from a high-temperature cubic to low-temperature monoclinic or triclinic crystal system. The Verwey transition temperature significantly depends on the size of Fe3O4 nanoparticles and degree of Fe3O4 nonstoichiometry [13–17]. Fe3O4 is frequently observed to be nonstoichiometric having a cation deficient Fe3þ sublattice [2]. a-Fe2O3, g-Fe2O3, and Fe3O4 are the most common phases of iron oxide found in nature, they exist in both nanosized and bulk forms and there is a diversity of synthetic routes to prepare them with desired sizes, particle size distribution, morphology, and physicochemical properties [2,10,18]. On the other hand, b-Fe2O3 and e-Fe2O3 are regarded as rare iron oxide phases, scarcely observed in nature and difficult to synthesize in laboratory as pure phases without other iron(III) oxide admixtures [2,10,19]. Due to their low surface energy, they exist only in the nanosized forms [20,21]. So far, a few reaction routes have been reported to prepare b-Fe2O3 and e-Fe2O3 phases [19,22–26], however, most of them provide low yields of b-Fe2O3 and e-Fe2O3 and lack the possibility to precisely control the size, particle size distribution, and morphology of final products. Furthermore, they are thermally unstable and easily transform to a-Fe2O3 upon thermal treatment [10]. Owing to the reaction products occurring in the below-mentioned solid-state thermal processes, only physical (i.e., structural and magnetic) properties and features of temperature-dependent zero-field and in-field M€ ossbauer spectra of Fe2O3 polymorphs are further discussed in details. 18.2.1 a-Fe2O3 a-Fe2O3 is a blood-red material and the oldest known iron oxide mineral [2]. From the thermodynamic viewpoint, a-Fe2O3 is extremely stable and is often found as the end member of temperature-induced phase transformations of other iron oxides [2,10]. It possesses a corundum-type (i.e., rhombohedrally centered hexagonal) crystal structure with hexagonally closed-packed arrays of oxygen ions stacked along the [001] direction leaving places for formation of the octahedral sites two-thirds of which are filled with Fe3þ ions (see Fig. 18.2) [2,10,27]. The arrangement of cations then constructs pairs of FeO6 octahedra when each octahedron shares edges with three neighboring octahedra in the same plane and one face with an octahedron in an adjacent plane along the c-axis. The crystal structure of a-Fe2O3 is described   within R 3c space group with lattice parameters a ¼ 5.034 A and c ¼ 13.752 A [2,10,27]. If exposed to pressure, the crystal    structure of a-Fe2O3 changes to an orthorhombic crystal family with a ¼ 4.580 A, b ¼ 4.950 A, and c ¼ 6.720 A [27]. When synthesized in a particle form, a rich variety of morphologies are observed such as plates and disks, rods, spheres, ellipsoids, rhombohedra, stars, and cubes. In cases when no additive is used a-Fe2O3 particles are often found to grow to rounded hexagonal plates and rhombohedra [2]. a-Fe2O3 is an n-type semiconductor and has the widest gap band (2.2 eV) of all the iron(III) oxide polymorphs reported so far [2]. In addition, it has been observed that a-Fe2O3 exhibits a very low hole mobility. Despite the wide bandgap, low quantum efficiency and high resistivity, a-Fe2O3 has been recently proposed as an electrode in the photoassisted electrolysis of water for hydrogen production [28,29]. It has been shown that the photoelectronic properties can be significantly improved if a-Fe2O3 particles are doped with Nb, Ge, Ti, Mo, Cr, and/or Si [30–33], are in the form of oriented needles placed on a conducting glass plate [34], and/or the pH of the electrolyte is increased [35]. However, its wide practical utilization is limited by a weak magnetization (see below) exhibited by a-Fe2O3. Recently, a-Fe2O3 has been found to be a suitable component in gas and humidity sensors [2].

354

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.2 (a) Schematic representation of the unit cell of a-Fe2O3 employing ball-stick model and (b) crystallographic structure of a-Fe2O3 represented by the cation polyhedra.

Besides, a-Fe2O3 is a magnetic material with two different magnetic orderings below its Neel temperature (TN  950 K) [2,10,27]. The magnetic state of a-Fe2O3 is driven by a strong magnetic interaction of an indirect exchange (i.e., superexchange) type acting among magnetic moments of Fe3þ ions via overlapping of Fe d-orbitals with p-orbitals of O2 ions situated between magnetically active Fe3þ ions (i.e., the antiferromagnetic interaction takes place across the shared faces of cation octahedron along the c-axis). In the basal plane, two interpenetrating sublattices exist [27]. At low temperatures (265 K in the case of bulk a-Fe2O3), a-Fe2O3 behaves as a pure antiferromagnetic material with Fe3þ magnetic moments lying along the c-axis (more precisely, lying along a line making an angle of 7 with the c-axis) and perpendicular to the basal plane of a-Fe2O3 crystal structure (see Fig. 18.3). In this magnetic state, Fe3þ magnetic moments of adjacent layers are exactly antiparallel leaving a zero net magnetic moment of the a-Fe2O3 magnetic structure. Furthermore, in the antiferromagnetic regime, Fe3þ magnetic moments are oriented along the electric field gradient axis. However, if the temperature exceeds a value of 265 K, the Fe3þ magnetic moments reorientate by 90 to the basal plane. Indeed, the Fe3þ magnetic moments do not lie in the basal plane but are rotated by a small angle (3 ) around the a-Fe2O3 c-axis. Thus, the magnetic structure is not perfectly antiferromagnetic (the two magnetic sublattices are not perfectly antiparallel with each other) and a slight canting gives rise to weak magnetic moment pointing along the c-axis. Therefore, at temperatures higher than 265 K, a-Fe2O3 is said to be a weakly ferromagnet (or canted antiferromagnet). This canting of Fe3þ magnetic moments is caused by the Dzyaloshinsky–Moriya antisymmetric interaction [36,37] and its existence is driven by the crystal fields generated by a favorable arrangement of O2 anions in the a-Fe2O3 crystal structure. This spin reorientation phenomenon is associated with the Morin transition [38] that separates the low-temperature pure antiferromagnetic regime and high-temperature weakly ferromagnetic state of a-Fe2O3. The Morin transition is regarded as the first-order thermodynamic transition that results from ceaseless competition between strong magnetic dipolar anisotropy (favoring the orientation of spins in the basal plane of the a-Fe2O3 crystal structure) and local ion anisotropy (i.e., spin–orbit interaction favoring the orientation of spins in the a-Fe2O3 rhombohedral [111] direction) terms having different thermal dependencies [39]. At the Morin transition, magnetic dipolar anisotropy equals to the local ion anisotropy leading to a change in the sign of the overall a-Fe2O3 magnetocrystalline anisotropy manifested by a change in the easy axis of the magnetic anisotropy. The temperature (TM) at which Morin transition takes place (for bulk a-Fe2O3, TM  265 K) depends on several parameters such as the size of particles (for nanoparticles with sizes below 20 nm, weakly ferromagnetic state is observed down to 4.2 K, see Fig. 18.4), lattice defects (e.g., low crystallinity, vacancies), substitution by foreign cations (if Al3þ ions with 8 mol% are

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

355

FIGURE 18.3 Arrangement of iron and oxygen atoms in the a-Fe2O3 structure where the circles represent Fe3þ ion positions and the dashed lines indicate planes within which oxygen atoms lie. The rhombohedral unit cell is drawn with dashed lines and the arrows mark the orientation of Fe3þ magnetic moments with respect to a coordination crystal system below the Morin transition temperature. (Adapted from Ref. 27 with permission of World Scientific Publishing Company.)

substituted, weakly ferromagnetic state is observed down to 4.2 K), presence of impurities, deviations from stoichiometry, surface effects, morphology, pressure, and application of external magnetic fields [2,9,10,27,40]. In addition, when the size of a-Fe2O3 nanoparticles drops below 10 nm, superparamagnetism appears due to a decrease in the particle relaxation time evoked by a decrease in the nanoparticle size (see Fig. 18.4) [40–43]. However, in the nanoparticle system with particle size distribution, the Morin transition is not sharp and takes place in the broad temperature interval when on lowering the temperature, a nanoparticle with a certain size enters the antiferromagnetic state at a given temperature corresponding to its size. In other words, weakly ferromagnetic and pure antiferromagnetic states coexist over a certain temperature interval the width of which reflects the particle size distribution exhibited by an assembly of a-Fe2O3 nanoparticles. At room temperature, the M€ ossbauer spectrum of a-Fe2O3 is composed of only one sextet (see Fig. 18.5a) [2,10] with the values of M€ ossbauer hyperfine parameters listed in Table 18.1. As the temperature is lowered, the transition

356

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.4 Schematic magnetic phase diagram of a-Fe2O3 as a function of particle size where P stands for the paramagnetic phase, WF for the weakly ferromagnetic phase, AF for the antiferromagnetic phase, SP for the superparamagnetic phase, and TM for the Morin temperature. The dashed areas represent region of coexistence of AF and WF phase and region of coexistence of WF and SP phase (from left to right). (Adapted from Ref. 9 with permission of Springer.)

from weakly ferromagnetic to pure antiferromagnetic state is indicated by a change in the value of d (d increases by 0.03 mm s 1), DEQ (DEQ changes from 0.20 mm s 1 for weakly ferromagnetic state to 0.40 mm s 1 for antiferromagnetic state), and Bhf (Bhf decreases by 0.9 T) [10]. As one can see, the Morin transition is well documented by a change in the value and the sign of the DEQ parameter that is associated with 90 spin reorientation phenomenon. In most cases, mainly due to the particle size distribution within the nanoparticle assembly, the Morin transition proceeds over a certain temperature interval and thus the Morin transition temperature is determined as a temperature at which the spectral areas of the sextets reflecting the antiferromagnetic and weakly ferromagnetic state equal to each other. Coexistence of the two sextets is manifested by an asymmetry of the overall spectral profile when the first and fifth lines

FIGURE 18.5 (a) Modeled typical zero-field € ssbauer room-temperature Mo spectrum of bulk a-Fe2O3 and (b) € ssbauer spectrum zero-field Mo of a-Fe2O3 nanosystem with a particle size distribution at the Morin transition temperature, showing the coexistence of weakly ferromagnetic and antiferromagnetic phase.

357

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

€ ssbauer Hyperfine Parameters of Four Crystalline Fe2O3 TABLE 18.1 Representative Room-Temperature Mo Polymorphs and Amorphous Iron(III) Oxide Iron (III) Oxide Polymorpha a-Fe2O3 b-Fe2O3 g-Fe2O3 e-Fe2O3

Amorphous Fe2O3

d  0.01 (mm s 1) 0.37 0.37 0.37 0.37 0.25 0.37 0.37 0.38 0.23 0.37

DEQ  0.01 (mm s 1) 0.20 0.72 0.96 0.00 0.00 0.25 0.25 0.03 0.15 1.03

Bhf  0.3 (T)

RA  1 (%)

Component

51.8 – – 50.5 50.0 44.5 44.6 38.7 25.6 –

100 75 25 62.5 37.5 25 25 25 25 100

Octahedral Fe3þ Octahedral d-site Fe3þ Octahedral b-site Fe3þ Octahedral Fe3þ Tetrahedral Fe3þ Octahedral A-site Fe3þ Octahedral B-site Fe3þ Octahedral C-site Fe3þ Tetrahedral D-site Fe3þ Octahedral Fe3þ

The isomer shift values are related to metallic iron. d is the isomer shift, DEQ denotes the quadrupole splitting, Bhf represents the hyperfine magnetic field, and RA stands for relative spectral area of individual spectral components a The values of the M€ ossbauer hyperfine parameters (especially Bhf) for a-Fe2O3 and g-Fe2O3 are frequently observed for bulk phases and may thus differ from those reported for nanosized counterparts.

of the overall sextet are more intense than the sixth and second lines of the overall sextet, respectively (see Fig. 18.5b). In this context, it has been proposed that in an a-Fe2O3 nanoparticle, its core exhibits a higher value of the Morin transition temperature than that corresponding to its surface [44]. Thus, weakly ferromagnetic phase progresses from the surface layers toward the nanoparticle core that is believed to be a result of a different magnetocrystalline anisotropy of nanoparticle surface layers compared to that present in the nanoparticle core. If the size of nanoparticles is reduced (typically below 10 nm), nanoparticle magnetic moment (i.e., superspin) begins to exhibit a temperature-activated oscillations of directions of superspin among orientations favored by nanoparticle magnetic anisotropy (for details, see Section 18.2.3). Thus, a-Fe2O3 nanoparticles behave in a superparamagnetic manner and if, at a given temperature and appropriate particle size, the relaxation time of superspin is shorter than the time window of the M€ ossbauer technique (10 8 s), a doublet component is detected. On lowering the temperature, a sextet may be restored at the so-called blocking temperature indicating a passage of the superspin from its superparamagnetic to blocked state. If monocrystalline a-Fe2O3 is exposed to an external magnetic field (Bext) in either weakly ferromagnetic or antiferromagnetic state, the profile of its in-field M€ ossbauer spectrum generally depends on the angle that Bext makes with the a-Fe2O3 antiferromagnetic axis [45]. In the parallel geometry when Bext is oriented along the g-rays direction, a sudden increase in the intensities of the second and fifth sextet line (A2,5) is observed if Bext, pointing along the a-Fe2O3 antiferromagnetic axis (i.e., c-axis), exceeds a value of 6.5 T (at 4.2 K) corresponding to the so-called spin-flop field (i.e., parallel-field transition when the antiferromagnetic axis of the atomic magnetic moments abruptly reorients from the caxis into the basal plane, but the atomic magnetic moments are canted to establish a net weakly ferromagnetic moment along Bext) [27]. Due to spin-flop phenomenon, characteristic of a-Fe2O3, Fe3þ magnetic moments, still preserving a nearly antiferromagnetic arrangement, reorientate so that they lie in a perpendicular direction with respect to the orientation of Bext. On the other hand, if Bext is oriented perpendicular to the c-axis and exceeds a value of 1.6 T (at 100 K), “screw rotation” (transverse-field) transition takes place when a ferromagnetic moment is observed to lie along Bext and the a-Fe2O3 antiferromagnetic axis tends, satisfying energy conditions, to move away from the c-axis toward the basal plane establishing a weakly ferromagnetic state [27]. The spin-flop phenomena also happen in an a-Fe2O3 nanoparticle but at lower Bext since the spin-flop field is significantly dependent on a particle size. It has been proposed that for an assembly of a-Fe2O3 nanoparticles with the antiferromagnetic axis pointing in all possible directions providing that for a given a-Fe2O3 nanoparticle Bext acts at an angle w with the c-axis that is smaller than an upper limit wC, we then observe that the antiferromagnetic axis flops into the basal plane and rotates toward the direction in that plane which is perpendicular to Bext [45]. The value of wC is dependent on the structural and morphological features of a-Fe2O3 nanosystem. On the contrary, if w > wC, no spin-flop takes place into the basal plane and the antiferromagnetic spin axis of such nanoparticles rotates away from the c-axis by a certain angle so that its direction is again perpendicular to Bext [45]. ossbauer spectroscopy, a-Fe2O3 can be recognized In both cases, A2,5 again increases. Thus, when employing in-field M€ from other iron(III) oxide polymorphs since its in-field spectral profile does not show any additional splitting due to Bext and, in the parallel experimental geometry, A2,5 increases when the spin-flop field characteristic for every a-Fe2O3

358

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.6 € ssbauer spectra of Modeled Mo mixture of a-Fe2O3 and g-Fe2O3 nanoparticles (mixture composed of 60 wt.% of a-Fe2O3 nanoparticles and 40 wt.% of g-Fe2O3 nanoparticles with sizes larger than 20 nm) at a temperature of 5 K (a) without an external magnetic field and (b) under an external magnetic field of 6.5 T, oriented along the g-rays direction.

nanoparticle is overcome. Different behaviors of A2,5 of a-Fe2O3 and g-Fe2O3 sextet components under Bext (for a M€ ossbauer spectrum response of g-Fe2O3 under Bext, see Section 18.2.3) can be used to distinguish them as demonstrated in Fig. 18.6. 18.2.2 b-Fe2O3 b-Fe2O3 is a rare and metastable crystalline polymorph of iron(III) oxide, existing only in a nanosized form, which was discovered by Bonnevie–Svendsen in 1956 (as a product of hydrolysis of FeCl3
6H2O) [10,22]. Its natural abundance has not been reported yet. The crystal structure of b-Fe2O3 displays two nonequivalent octahedral cation sites, denoted as b-site and d-site (see Fig. 18.7) [10,46]. b-Fe2O3 possesses a body-centered cubic crystal structure of a bixbyite type with Ia 3 space group. The b-site is describable by a nearly regular oxygen octahedron surrounding the cation, whereas the d site exhibits a rather irregular sixfold oxygen coordination around the cation. Thus, in the cubic unit cell with a ¼ 9.393 A, 24 Fe3þ ions occupy d-sites with the C2 symmetry and 8 Fe3þ ions are located at b-sites with the C3i symmetry [10,46]. All cation sites are filled in the b-Fe2O3 crystal structure. At room temperature, b-Fe2O3 is paramagnetic [23–25,47,48] in contrast to all other iron(III) oxide polymorphs that are well magnetically ordered at 300 K. The room-temperature paramagnetic behavior of b-Fe2O3, thus, enables to distinguish b-Fe2O3 from a-Fe2O3, g-Fe2O3, and e-Fe2O3. As the temperature is lowered, b-Fe2O3 enters an antiferromagnetic state at the Neel temperature (TN) reported in the range of 100–120 K [47,48]. As expected, the overall profile of the room-temperature M€ ossbauer spectrum of b-Fe2O3 consists of two distinguishable doublet components (see Fig. 18.8a) [10,24,47] with the values of M€ ossbauer hyperfine parameters listed in Table 18.1. The relative spectral areas of doublet components are found to be in the ratio of 3:1 (d-doublet/b-doublet) that corresponds with the ratio of Fe3þ ions in the two nonequivalent octahedral sites. In some cases, when it is not ossbauer spectrum with two doublet components due to increased possible to fit the room-temperature b-Fe2O3 M€ linewidths (driven by particle size or crystal defects), only one doublet is used with the value of d and DEQ equal to 0.37 ossbauer spectra are observed. Again, the and 0.75 mm s 1, respectively [48]. Below TN of b-Fe2O3, magnetically split M€ ossbauer spectra shows two sextet components belonging to b- and d-sites [47]. profile of low-temperature b-Fe2O3 M€ The two sextet components are well resolved due to significantly different values of Bhf at the two nonequivalent crystallographic positions (see Fig. 18.8b).

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

359

FIGURE 18.7 (a) Schematic representation of the unit cell of b-Fe2O3 employing ball-stick model and (b) crystallographic structure of b-Fe2O3 represented by the cation polyhedra.

If exposed to heat treatments, b-Fe2O3 very easily transforms to a-Fe2O3 at temperatures higher than 800 K [10,49,50]. This transformation is accompanied by a thermally induced growth of b-Fe2O3 nanoparticles that become thermally unstable if the size of b-Fe2O3 nanoparticles reaches a critical value at which the phase transformation is triggered (for details, see Section 18.2.4) [21].

FIGURE 18.8 Modeled typical zero-field € ssbauer spectra of b-Fe2O3 at Mo (a) room temperature and (b) at a temperature of 5 K.

360

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

18.2.3 g-Fe2O3 g-Fe2O3 is a red-brown material frequently occurring in nature as a weathering product of Fe3O4 or as a product of heating of oxide–hydroxides and hydroxides [2]. Besides, g-Fe2O3 is readily synthesized [2,10,18] and an enormous number of synthetic routes have been proposed how to prepare g-Fe2O3 nanoobjects (e.g., in the form of nanoparticles, powders, nanocomposites, films, etc.), each of them enabling a different degree of control of g-Fe2O3 particle size, particle size distribution, morphology, stoichiometry, phase purity, doping of foreign cations, and structural and magnetic features (for more details on various synthetic routes leading to g-Fe2O3 nanoobjects, see Refs. 2,18). Owing to its application-promising properties stemming especially from its interesting magnetic characteristics (e.g., superparamagnetism, strong magnetic response under external magnetic fields) that can be very easily tuned in the nanoworld, it is considered as the most important and prosperous iron(III) oxide phase that has already been exploited in various magnetism-based technological (e.g., magnetic fluids, gas sensors, magnetooptical devices, magnetocaloric refrigeration, etc.) and medical (e.g., contrast agent for nuclear magnetic resonance imaging, medical diagnosis, targeted drug delivery, magnetic-induced cancer therapy—hyperthermia, etc.) branches [18]. Within the scientific community working in the field of magnetic nanomaterials, g-Fe2O3 is widely recognized as a perspective magnetic nanomaterial having a very high application potential. Although its existence was announced at the end of the nineteenth century, g-Fe2O3 was first experimentally evidenced by Baudisch and Welo in 1925 (as a product of oxidation of Fe3O4) [51,52]; they were the first researchers to distinguish a-Fe2O3 and g-Fe2O3 on the basis of X-ray diffraction data. It has a cubic structure of an inverse spinel type  with a ¼ 8.351 A (see Fig. 18.9) [2,10,18]. Similarly as in Fe3O4, Fe3þ cations in the g-Fe2O3 crystal structure occupy two crystalographically nonequivalent sites, that is, tetrahedral (T) and octahedral (O) positions. In order to establish a neutral charge of the g-Fe2O3 crystal structure, some cation positions are left unfilled; vacancies usually occur at the octahedral sites. Thus, the g-Fe2O3 crystal structure is described being cation deficient fulfilling a stoichiometric formula given by (Fe3þ)A(Fe5/33þ&1/3)BO4, where & stands for vacancies. It follows that Fe3þ ions occupy the octahedral and tetrahedral sites in the ratio of 1:1.67. If this cation site-occupation ratio is found to be not satisfied, g-Fe2O3 is referred to as being nonstoichiometric indicating a presence of Fe2þ ions (“residue” of incomplete oxidation of Fe3O4 to g-Fe2O3) and/or possible migration of vacancies to the tetrahedral crystallographic sites. When refining an X-ray powder diffraction pattern of synthetic g-Fe2O3, superstructure diffraction lines may be observed implying ordering of vacancies and cations in the g-Fe2O3 crystal structure [10,18]. If vacancies are ordered, the highly symmetric cubic crystal symmetry is lost; the  g-Fe2O3 crystal structure then exhibits tetragonal symmetry and is describable by a tetragonal unit cell with a ¼ 8.330 A

FIGURE 18.9 (a) Schematic representation of the cubic unit cell of g-Fe2O3 (P4132 space group) employing ball-stick model and (b) cubic crystallographic structure of g-Fe2O3 represented by the cation polyhedra.

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

361

FIGURE 18.10 (a) Schematic representation of the tetragonal unit cell of g-Fe2O3 (P43212 space group) employing ball-stick model and (b) tetragonal crystallographic structure of g-Fe2O3 represented by the cation polyhedra.



and c ¼ 25.010 A within P43212 space group (see Fig. 18.10). If vacancies are partially ordered, g-Fe2O3 retains a cubic unit cell and its crystal structure falls into P4132 space group belonging to the cubic crystal family (see Fig. 18.9). In cases when vacancies are distributed at random over the octahedral crystallographic sites, g-Fe2O3 has a cubic unit cell and its crystal structure carries the same symmetry signs as those observed for Fe3O4, thus being described within Fd3m space group. The existence of the two crystallographically nonequivalent sites predestinates the nature of g-Fe2O3 magnetic behavior. Actually, they give rise to the two interpenetrating magnetic sublattices with T-site Fe3þ magnetic moments oriented in a perfect antiparallel manner to the direction of O-site Fe3þ magnetic moments [2,10,18]. Since the magnitude of T- and O-site Fe3þ magnetic moment is equal to 4.18 and 4.41mB (with mB being the Bohr magneton), respectively, g-Fe2O3 is then said to be a collinear ferrimagnet with a noncompensated antiparallel alignment of the Tand O-site Fe3þ magnetic moments that arises from strong antiferromagnetic T–O (intersublattice) superexchange interactions and weak T–T (intrasublattice) and O–O (intrasublattice) ones. In most cases, the direct determination of its Curie temperature (TC) is not possible due to its phase transformation to a-Fe2O3 on heating. Based on the extrapolation of g-Fe2O3 spontaneous magnetization, TC has been reported in the range from 780 to 980 K [2,10,18]. In general, the magnetic behavior of nanosized system is governed by finite-size and surface effects [53,54]. While finite-size effects arise due to a quantum confinement of electron motion along the confined dimensions of a nanomaterial, the emergence of surface effects is ascribed to an enhanced surface-to-volume ratio and a symmetry breaking at the surface layers of a nanomaterial leading to an increased surface anisotropy, surface spin disorder, spin frustration, and weakening of magnetic interactions among atomic magnetic moments lying at the nanoparticle surface layers and core. Beside these size-driven phenomena, magnetic properties of a nanomaterial are affected by strength of interparticle magnetic interactions, porosity, and presence of vacancies and defects. On a way to find the minimum energy of the nanosystem, interplay, action, and competition of all these relevant phenomena cause an emergence of diverse magnetic characteristics of the nanomaterial (nanosystem) anomalous to those exhibited by its bulk counterpart. As the size of g-Fe2O3 particles decreases, the number of magnetic domains within a particle is reduced and meeting the requirement of minimum particle magnetic energy, the multidomain state ceases to exist at a certain critical size of g-Fe2O3 particles (166 nm in diameter being estimated for spherical and noninteracting particles) below which they become single domain [55]. In the single-domain state, all atomic magnetic moments cooperate within a particle giving rise to a giant particle magnetic moment (i.e., superspin) that a particle carries. Under an assumption that particle

362

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

magnetic anisotropy is of a uniaxial character (fulfilled in most cases), easy axis of particle magnetization energetically favors two orientations of superspin separated by an anisotropy energy barrier the height of which depends on the magnetic anisotropy constant and particle size. Thus, superspin orientation is fixed in either of these two directions and superspin is said to be in a blocked state if no other energy contribution (e.g., thermal fluctuations, external magnetic field) compete with magnetic anisotropy energy. At a given temperature and under a zero external magnetic field, a reduction in the particle size results in a decline in the magnetic anisotropy barrier height causing thermal energy to be sufficiently enough to overcome it and activate a spontaneous reversal of superspin between the two energetically favorable orientations. This thermally activated behavior of superspin is known as superparamagnetism [55–57] and the time superspin spends in a direction favored by particle magnetic anisotropy is given by the relaxation time the value of which depends, beside other parameters, on the magnetic anisotropy barrier height and temperature. Thus, superparamagnetism is regarded as a result of competition between the magnetic anisotropy energy and energy of thermal fluctuations, the latter particle energy contribution dominating over the former one above the blocking temperature (TB). In general, TB, separating the superspin superparamagnetic state and low-temperature blocked state, is dependent on many factors including particle size, particle size distribution, morphology, external magnetic field, interparticle interactions, and so on. Moreover, due to the relaxation nature of superparamagnetism, its observation is related to the time window of the measuring technique emphasizing that the superparamagnetic relaxation of a given nanosystem, detected by an experimental technique with a characteristic measuring time, does not need to show up in the measurement by other experimental technique with different characteristic measuring time at the same temperature. In addition, the superparamagnetic behavior of an assembly of nanoparticles and a transition to its blocked state can be significantly influenced by presence of interparticle magnetic interactions the strength of which may induce a collective magnetic regime of nanoparticle superspins with features close to those reported for spin-glass systems [18,58,59]. Within the scientific community, nanoparticles of g-Fe2O3 are regarded as first model and experimentally most researched nanomaterials used for building the theory of superparamagnetic relaxation, effect of interparticle magnetic interactions on low-temperature state of nanoparticle systems, explanation of occurrence of collective magnetic excitations in the blocked state, theory of spin canting and spin frustration, and so on. Owing to the presence of the two crystallographically nonequivalent sites in the g-Fe2O3 crystal structure, one would expect two sextets forming the room-temperature g-Fe2O3 M€ ossbauer spectrum (see Fig. 18.11a) [10,18].

FIGURE 18.11 Modeled typical zero-field € ssbauer spectra of g-Fe2O3 at Mo (a) room temperature and (b) at a temperature of 5 K.

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

363

Instead, only one slightly asymmetric broad sextet is observed at room temperature implying that the M€ ossbauer hyperfine parameters for the tetrahedral and octahedral sites are almost the same in magnitude (see Table 18.1) [10,18]. If fitted with one sextet, the room-temperature values of the M€ ossbauer hyperfine parameters are as follows: d ¼ 0.31 mm s 1, DEQ ¼ 0 mm s 1, and Bhf ¼ 50.2 T (for g-Fe2O3 nanoobjects, especially the Bhf value may be smaller at 300 K) [10]. The sextet asymmetry is manifested by the first, second, and third line being more intense than the fourth, fifth, and sixth line, respectively. For stoichiometric g-Fe2O3, a ratio of spectral areas of tetrahedral-to-octahedral sextet is expected to be equal to 3:5; a deviation from this value indicates a certain nonstoichiometry of g-Fe2O3. As the temperature decreases, the asymmetry grows that, in some cases, makes possible to better distinguish the two sextet components (see Fig. 18.11b). For g-Fe2O3 nanoparticles (typically less than 15 nm in size), a doublet spectral component is observed indicating a superparamagnetic behavior of nanoparticle system with regard to the characteristic time of the M€ ossbauer technique. As the temperature decreases, the doublet gradually transforms to the sextet and the temperature interval of coexistence of the doublet and sextet components particularly depends on the particle size distribution, particle morphology, and degree of interparticle magnetic interactions. In addition, the Bhf temperature behavior of g-Fe2O3 nanosystem may differ from that reported for bulk g-Fe2O3 due to a presence of collective magnetic excitations arising from fluctuation of a particle superspin around the particle easy axis of the magnetization [18,60–62]. Beside this, the presence of collective magnetic excitations causes additional asymmetric broadening and shifts of individual lines in the g-Fe2O3 sextet. If the size of g-Fe2O3 nanoparticles is not uniform within their assembly, Bhf exhibits a distribution. Due to collective magnetic excitations, this distribution becomes asymmetric with longer tails spreading toward the lower field values and is, thus, responsible for asymmetrically broadened M€ ossbauer lines. If g-Fe2O3 is placed into Bext, the two sextet components are resolved as Bext adds to and subtracts from the Bhf value of the tetrahedral and octahedral sextet, respectively [10,18]. In other words, Bhf at the tetrahedral sites points in a parallel direction with respect to the orientation of Bext, whereas Bhf at the octahedral sites lies in an antiparallel direction with the orientation of Bext. Beside the values of the effective hyperfine magnetic field (Beff) at both sites, the degree of alignment of hyperfine fields can be monitored via A2,5. It is known that the relative spectral areas of six lines are given by 3:x:1:1:x:3, where x ¼ [4sin2(u)]/[1 þ cos2(u)] with u being the angle that the (effective) hyperfine magnetic field makes with the direction of propagation of the g-rays. In a parallel measuring geometry when Bext lies along the g-rays direction, the alignment of Beff to the direction of Bext is manifested by A2,5 ¼ 0. Thus, in the case of bulk g-Fe2O3 behaving as a Neeltype ferrimagnet, A2,5 ¼ 0 is expected as atomic magnetic moments belonging to tetrahedral and octahedral sites perfectly orientate to and against the direction of Bext, respectively (see Fig. 18.12a). However, if the size of g-Fe2O3 nanoparticles falls down below 10 nm, finite-size and surface effects cause a significant enhancement of other anisotropy contributions (especially surface magnetic anisotropy and local magnetic anisotropy) to the overall particle magnetic anisotropy, making thus certain atomic magnetic moments (especially those situated in the nanoparticle surface layers and/or around a defect) be more resistive to an aligning effect of Bext [10,18,63,64]. Thus, spin frustration evolves leading to a spin disorder. Certain atomic magnetic moments become canted and much stronger Bext is needed to overcome enhanced anisotropy terms and align them. The appearance of spin canting phenomenon is evidenced by nonzero A2,5 (see Fig. 18.12b) and its degree depends on particle size, particle size distribution, temperature, and interparticle magnetic interactions [18]. To explain the essence of spin canting, numerous works have dealt with g-Fe2O3 nanoparticle systems. As a result of this long-lasting theoretical and experimental endeavor, two different scenarios have emerged so far [18]. Some researchers claim that spin canting phenomenon is driven solely by surface effects, whereas the others believe that spin canting is entirely a demonstration of finite-size effects. If surface nature of spin canting is ossbauer spectra are fitted with three sextets: (i) the two sextets used for tetrahedral and adopted, the in-field g-Fe2O3 M€ octahedral sites, respectively, with A2,5 set to zero; and (ii) the third sextet with reduced value of Beff and nonzero A2,5 reflecting canted atomic moments lying in the nanoparticle surface layers. Employing the “core-shell” model, it is then straightforward to get the value of thickness of magnetically disorder layers. On the other hand, if a finite-size origin of ossbauer spectra. However, spin canting phenomenon is assumed, only two sextets are used to fit the in-field g-Fe2O3 M€ using the model of two sextets raises a debate how to correctly fit (respecting the physics behind the spin canting phenomenon) the intensities of the second and fifth lines belonging to the two sextet spectral components since these lines are not resolved frequently. One hypothesis supposes that for g-Fe2O3 nanoparticles, spin canting occurs at both crystallographic sites, whereas other theory claims that only one kind of sites (i.e., octahedral sites) suffers spin canting while, at the other sites (i.e., tetrahedral sites), spin canting does not take place. In order to clear it up, an approach based on bidimensional hyperfine field-canting angle distribution has been proposed when both effective hyperfine magnetic field and canting angle are left to vary and their distributions are then used to fit the in-field M€ ossbauer spectra of g-Fe2O3 nanoparticles exhibiting spin canting [10].

364

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.12 Modeled representative in-field € ssbauer spectra of g-Fe2O3 Mo nanosystem (a) without spin canting and (b) exhibiting spin canting (spectra at a temperature of 5 K and under an external magnetic field of 5 T, oriented along the g-rays direction).

In principle, if g-Fe2O3 is thermally treated, it transforms to a-Fe2O3 at a phase transformation temperature the value of which depends on particle size and related surface energy, particle size distribution, particle morphology, addition of foreign cations, exploitation of supporting matrix (i.e., silica in most cases), pressure elevation, conditions during and after formation of the natural sample, type of synthesis method, and impurity content [10,18]. Due to many factors affecting the onset of this phase transformation, a wide range of g-Fe2O3 ! a-Fe2O3 transformation temperatures has been reported from 600 to 1300 K [10,18]. As discussed in details in Section 18.2.4, under certain circumstances (suitable particle size, degree of g-Fe2O3 particle agglomeration, etc.) and in the certain temperature interval, a rare metastable e-polymorph of iron(III) oxide appears as an intermediate phase on the thermally induced transformation pathway from g-Fe2O3 to a-Fe2O3. 18.2.4 e-Fe2O3 e-Fe2O3 is a dark brown magnetic phase of iron(III) oxide [2,19]. It was discovered by Forestier and Guiot-Guillain in 1934 (as a product of thermal decomposition of Fe2O3
4BeO) but it started to be labeled as e-Fe2O3 by Schrader and Buttner later in 1963 (a product synthesized by an electric arc discharge of an iron oxide aerosol under oxidizing atmosphere) [19,65,66]. Its natural occurrence is very limited (e.g., biogenic nanoparticles in some plants, thermal decomposition product of almandine garnets, and iron-rich clays) and there is only a few of synthetic procedures described so far producing e-Fe2O3 either as a single phase (without experimentally detected traces of admixtures of other iron oxide phases) or with an experimentally detectable but, in some cases, insignificant portion of other Fe2O3 polymorphs [19]. Similarly as in the case of b-Fe2O3, e-Fe2O3 exists in the form of nanoobjects and to date, e-Fe2O3 nanoparticles of sphere and rod shapes, nanowires, and thin film have been reported [19]. However, e-Fe2O3 is difficult to synthesize as individual nanosized objects due to its significant thermal instability. The spatial restriction of nanoparticle growth is widely regarded as a crucial point for production of e-Fe2O3 nanoparticles [19]. Thus, it turns out that either a certain degree of agglomeration in the precursor powders and/or utilization of supporting medium (e.g., silica matrix) is highly required to produce e-Fe2O3. The use of silica matrix is advantageous since it offers a possibility of controlled preparation of e-Fe2O3 nanosized crystals. It has been proposed that the porous nature of the amorphous

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

365

FIGURE 18.13 (a) Schematic representation of the unit cell of e-Fe2O3 employing ball-stick model and (b) crystallographic structure of e-Fe2O3 represented by the cation polyhedra.

silica matrix provides nucleating sites for e-Fe2O3 nanoparticles and at the same time, significantly prevents their aggregation isolating nanoparticles from each other. Moreover, the confinement of e-Fe2O3 nanoparticles within the pores of supporting matrix increases their thermal stability. Unfortunately, all literature-presented synthetic routes yield e-Fe2O3 in small quantity and do not allow its large-scale production. The detailed overview of a series of synthetic routes leading to e-Fe2O3 can be found in the recent review work by Tuc9ek et al. [19]. e-Fe2O3 is an extraordinary nanomaterial exhibiting a giant room-temperature coercive field, millimeter-wave ferromagnetic resonance, and coupled magnetoelectric properties, that is, features that are not observed in any other simple iron oxide phase [19,67–69]. The combination of these nanomaterial characteristics predestinates e-Fe2O3 to exploit it in various applications based on high coercivity materials (e.g., recording media) and/or requiring coupled electric and magnetic material characteristics (e.g., electric/magnetic field tunable devices and multiple state memory elements) and/or involving absorption of electromagnetic waves with millimeter wavelengths (e.g., components in devices used for suppression of the electromagnetic interference and stabilization of the electromagnetic transmittance) [19]. However, currently, its practical utilization is obstructed by low yields, lack of precise control of resulting product size, mixed phases, and presence of impurities. By that time, the synthetic approach employing Group IIA metal ions and silica matrix represents the way to produce the purest e-Fe2O3 phase without any other iron(III) oxide polymorphs (especially a-Fe2O3 and/or g-Fe2O3) as admixtures since they have not been detected by various experimental techniques allowing the analysis of the sample phase composition including recently employed zero-field and in-field M€ ossbauer spectroscopy [19].   e-Fe2O3 crystallizes with an orthorhombic unit cell having lattice parameters a ¼ 5.072 A, b ¼ 8.736 A, and  c ¼ 9.418 A (see Fig. 18.13) [10,19]. The e-Fe2O3 crystal structure is described within Pna21 space group, is said to be polar and isomorphous to GaFeO3, AlFeO3, and k-Al2O3. The orthorhombic unit cell of e-Fe2O3 is composed of triple chains of edge-sharing FeO6 octahedra running along the a-axis that are connected to each other by sharing corners of the FeO6 octahedra leaving one-dimensional cavities that are filled by the chains of the corner-sharing tetrahedra [19]. Thus, the e-Fe2O3 crystal structure contains four independent crystallographically nonequivalent cation sites, that is, three different octahedral sites (i.e., A-, B-, and C-sites) and tetrahedral sites (i.e., D-sites). It has been found that all four cation coordination polyhedra (i.e., three different octahedra and one tetrahedron) exhibit a different degree of distortion. Thus, A- and B-sites are said to possess a distorted octahedral coordination, whereas the C-site is regarded having a regular octahedral coordination. These distortions, especially those occurring in the local environment of the octahedral A- and B-sites and tetrahedral D-site, play important role in understanding the magnetic properties of e-Fe2O3 since they are responsible for a generation of a nonzero orbital component to the total Fe3þ magnetic moment and consequently, emergence of significant spin–orbit coupling phenomenon in e-Fe2O3 [19,70]. Actually, the crystal structure of e-Fe2O3 involves similarities with both the g-Fe2O3 crystal structure (a cubic spinel structure composed

366

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.14 (a) Temperature dependence of magnetization (M) of e-Fe2O3 upon cooling and warming and (b) room-temperature hysteresis loop of e-Fe2O3. (Taken from Ref. 19 with permission of the American Chemical Society.)

of a tetrahedron of four-coordinated Fe3þ ion and an octahedron of six-coordinated Fe3þ ion) and the a-Fe2O3 crystal structure (a rhombohedral structure consisting of stacked sheets of octahedrally coordinated Fe3þ ions between two closed-packed layers of oxygens). The existence of four nonequivalent cationic sites in the e-Fe2O3 crystal structure predestinates its magnetic structure. e-Fe2O3 is described as a 4-sublattice magnetic material, characterized by four sublattice magnetizations exhibiting different temperature behaviors [19]. Based on the temperature behavior of the e-Fe2O3 overall magnetization, e-Fe2O3 is then said to be a Neel P-type ferrimagnet. In general, e-Fe2O3 exhibits complex magnetic properties that dramatically change at the two temperatures, that is, at 490 K (Curie temperature) and 110 K (see Fig. 18.14a) [19]. At 490 K, a transition from a paramagnetic to magnetically ordered state is observed. At room temperature, the magnetic moments (m300 K(Fei), i ¼ A, B, C, and D) of iron atoms point along the crystallographic a-axis with m300 K(FeA) ¼ 3.9mB, m300 K(FeB) ¼ 3.9mB, m300 K(FeC) ¼ 3.7mB and m300 K(FeD) ¼ 2.4mB [19]. In other words, the Fe3þ magnetic moments in the distorted octahedral A- and B-sites mutually cancel, and the net magnetization results from the uncompensated magnetic moments of the iron atoms in the tetrahedral D- and regular octahedral C-sites. Taking into account these values of magnetic moments of Fe3þ ions at different crystallographic sites, it is then easy to show that we get a net magnetization of 0.3mB per Fe3þ at 300 K. Such a low value of net magnetization offers two possible conclusions concerning the room-temperature magnetic ground state of e-Fe2O3 [19]. It behaves either as a collinear ferrimagnet or as a canted antiferromagnet but the agreement regarding this issue has not been reached within the scientific community yet. At room temperature, e-Fe2O3 displays a giant coercive field of 2 T (see Fig. 18.14b), the origin of which has not been satisfactorily explained. It seems to be a result of several characteristic features of e-Fe2O3 such as a large magnetocrystalline anisotropy (KMC  2–5  105 J m 3) driven by an orthorhombic crystal structure, a low saturation magnetization (MS  15–25 Am2 kg 1), an establishment of a single magnetic domain character due to conveniently sized nanoparticles, and a nonzero orbital component of Fe3þ magnetic moment [19,67,70]. At 110 K, other magnetic transition is observed accompanied by a collapse in the e-Fe2O3 coercivity and a significant reduction in the orbital component of Fe3þ magnetic moment. Again, two possible low-temperature magnetic regimes of e-Fe2O3 have been reported so far [19]. Some researchers claim that at 110 K, e-Fe2O3 passes from collinear ferrimagnetic state to some incommensurate magnetic structure, most probably of a square-wave-modulated origin, whereas the others propose that at 110 K, e-Fe2O3 undergoes magnetic transition from one canted antiferromagnetic state (characterized by a certain value of the canting angle) to another canted antiferromagnetic state (with a different value of the canting angle) with an emergence of the metamagnetic behavior at low temperatures. Independently on the nature of e-Fe2O3 magnetic states below and above 110 K, it is believed that this magnetic transition is not sharp but it

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

367

FIGURE 18.15 Modeled typical zero-field room€ ssbauer spectemperature Mo trum of e-Fe2O3.

involves several stages starting at 150 K and ending at 80 K. In this temperature interval, changes in the coordination of octahedral A- and tetrahedral D-sites are assumed to take place that arise simultaneously and/or as a consequence of emergence of the low-temperature magnetic regime. Thus, a magnetic transition at 110 K is accompanied by the lattice distortions and structural changes (i.e., a series of structural transformations) as evidenced by changes in the lattice parameters. This is closely related with the behavior of the orbital component of Fe3þ magnetic moment that suffers a significant reduction in this temperature interval. In addition, it has been demonstrated that the magnetic transition to the low-temperature magnetic regime is also accompanied by a spin reorientation phenomenon resembling characteristics of the Morin-like transition in a-Fe2O3. If foreign cations are substituted (Ga, Al, and In), the values of room-temperature coercive field and net magnetization are affected, providing thus possibility to tune the e-Fe2O3 magnetic properties and frequencies at which the ferromagnetic resonance then occurs (see Ref. 19 for details). ossbauer spectrum consists of four magnetically split components At room temperature, the profile of e-Fe2O3 M€ (see Fig. 18.15) [10,19] with the values of M€ ossbauer hyperfine parameters listed in Table 18.1. The ratio of spectral areas of four sextet components is expected to be equal to 1:1:1:1 taking into account the equal ion occupancies at all four crystallographic sites and assuming that all iron atoms in e-Fe2O3 have the same recoilless fraction. However, due to nearly the same values of M€ ossbauer hyperfine parameters of the two sextets belonging to A- and B-sites, e-Fe2O3 roomtemperature overall spectral profile is usually described by three sextets with a ratio of spectral areas of 2:1:1. Contrary to g-Fe2O3 for which, at 300 K and without an applied magnetic field, it is almost impossible to distinguish between the two sextets arising from the tetrahedral and octahedral sites, the separation of the tetrahedral and octahedral spectral components is clearly evident at 300 K, owing mainly to a significantly lower value of Bhf at the tetrahedral sites of the e-Fe2O3 crystal structure. Nonzero values of DEQ for sextets arising from octahedral A- and B-sites imply a slightly distorted surrounding of iron atoms sitting at the octahedral A- and B-sites, whereas a negligible value of DEQ for sextet from C-sites predicates a highly symmetric and undistorted local environment of iron atoms located at the octahedral Csites. Based on the value of DEQ for tetrahedral sextet, the tetrahedral D-sites also exhibit a certain degree of distortion but it is expected to be not so pronounced as that for the octahedral A- and B-sites. Below 150 K, a radical change in the

368

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

profile of e-Fe2O3 M€ ossbauer spectra is observed. In the temperature interval from 80 to 150 K, most of the M€ ossbauer hyperfine parameters of various iron sites deviate from the temperature dependences displayed at higher temperatures (e.g., unexpected changes in the temperature behaviors of Bhf for all sextets, DEQ for sextets belonging to octahedral Aand D-sites, significant line broadening, etc.) mainly due to a series of structural transformations and spin reorientation phenomenon accompanying the magnetic transition at 110 K (see Ref. 19 for details). Below 80 K, only two sextets ossbauer spectrum, when the first sextet arises from the three different are, in most cases, resolved in the e-Fe2O3 M€ octahedral sites and the second sextet corresponds to the tetrahedral sites. At 4.2 K, the fitting of the octahedral (O) sextet gives dO  0.49 mm s 1, DEQ,O  0 mm s 1, and Bhf,O  51.6 T, whereas for the tetrahedral (T) sextet, we find dT  0.31 mm s 1, DEQ,T  0 mm s 1, and Bhf,T  45.8 T [19]. Analyzing the temperature dependences of Bhf at all ossbauer technique individual sites of the e-Fe2O3 crystal structure demonstrates that within the timescale of the M€ (10 8 s), the magnetic moments of the iron atoms situated at the A- and B-sites behave as quantum spins contrary to the magnetic moments of the iron atoms located at the C- and D-sites that behave as freely rotating classical spins. ossbauer spectrum is more resolved since Bhf If e-Fe2O3 is exposed to Bext, its in-field room-temperature M€ increases for the A- and D-sites, whereas it decreases for the B- and C-sites [19]. Up to date, only a few of in-field ossbauer spectroscopy can be exploited to study the M€ ossbauer spectra of e-Fe2O3 have been reported. In-field M€ magnetic structure of e-Fe2O3 as it can reliably monitor the arrangement of Fe3þ magnetic moments under Bext. In the case of Fe3þ ions, Bhf is solely given by the negative Fermi contact term. Thus, the magnetic moment of an Fe3þ ion is ossbauer spectroscopy, applied in a oppositely oriented with respect to Bhf. When investigating e-Fe2O3, in-field M€ parallel and/or perpendicular geometry, did not proved unambiguously the magnetic ground state of e-Fe2O3 since either a collinear ferrimagnetic and/or canted antiferromagnetic regime have been derived from corresponding in-field M€ ossbauer spectra [19]. Thus, in-field M€ ossbauer spectroscopy applied for e-Fe2O3 has to be revisited in order to clear up this vague matter. When e-Fe2O3 is thermally treated, it transforms to a-Fe2O3 at temperatures falling in the range from 700 to 1300 K [19]. If confined in the pores of silica matrix with defined sizes, the thermal stability of e-Fe2O3 is enhanced shifting the transformation temperature up to 1700 K [19]. It is worth mentioning that the g-Fe2O3 ! a-Fe2O3 ! eFe2O3 transformation involves growth of Fe2O3 nanoparticles. It is known that the free energy (G) per volume (V) of different i-Fe2O3 phases (i ¼ a, b, g, e) and the energy barrier that must be overcome for the phase transformation to take place are main factors affecting what Fe2O3 nanosized polymorph is formed from a precursor and how it is transformed into various iron(III) phases [19,20,71]. G/V ratio takes into account the chemical potential (h) and the surface energy (s), that is, G/V ¼ h/y þ 6s/d, where y is the molar volume and d denotes the size of a nanoparticle. Then, one can easily derive that e-Fe2O3 can exist when the size of the Fe2O3 particle is 6y(s e s g)/(he hg) < d < 6y(s e s a)/(he ha), satisfying conditions that (i) ha < he < hg, (ii) s a > s e > s g, and (iii) (s e s g)/(he hg) > (s e s a)/(he ha). It thus follows that if nanoparticles of Fe2O3 grow large enough, the existence of e-Fe2O3 is no longer favored (see Fig. 18.16). In other words, the size reduction of the Fe2O3 particle enhances a contribution of the surface (or interface) energy to G that stabilizes e-Fe2O3 in the nanoscaled size. Recently, Sakurai et al. [21] succeeded to firstly observe a successive phase transformation including all the four crystalline polymorphs of Fe2O3 (i.e., the g-Fe2O3 ! e-Fe2O3 ! b-Fe2O3 ! a-Fe2O3 phase transformation) upon increasing the Fe2O3 particle size precisely. In the synthesis employing silica matrix as a size restrictor and FeSO4 as a precursor, the threshold sizes (diameters of spherical nanoparticles) at which the g ! e, e ! b, and b ! a phase transformations occur were estimated to be 8, 30, and 50 nm, respectively. FIGURE 18.16 Stability of individual polymorphs of Fe2O3 based on the calculated dependence of the free energy per volume (G/V) on the size (d) of the iron(III) oxide nanoparticles of particular polymorph (the dependences derived under the conditions of ha < he < hg , s a > s e > s g , and (s e s g )/(he hg ) > (s e s a)/(he – ha). (Taken from Ref. 19 with permission of the American Chemical Society.)

18.2 POLYMORPHS OF IRON(III) OXIDE, THEIR CRYSTAL STRUCTURES, MAGNETIC PROPERTIES

369

18.2.5 Amorphous Fe2O3 Iron(III) oxide exists in an amorphous form if the size of Fe2O3 nanoparticles is less than 5 nm in a diameter [10,72]. In 1976, an amorphous nature of Fe2O3 was first reported by van Diepen et al. [73] who synthesized it as a thin film by thermal decomposition of aerosol when FeCl3 solution in butylacetate was used as a precursor. Since that time, it has been shown that amorphous Fe2O3 has a significant application potential in the field of catalysis, as precursor for the multistep preparation of a magnetic fluid composed of superparamagnetic nanoparticles, in the area of magneto-optical sensors and magnetic devices (nanocomposites of amorphous Fe2O3 and SiO2), electrochemistry (as a component in a lithium intercalation cathode), humidity sensors, preparation of soft Fe(0) nanoparticles applicable in magnetic semiconductors and memory devices [72]. Up to date, several synthetic routes leading to amorphous Fe2O3 have been proposed; amorphous Fe2O3 can be synthesized in the form of nanoparticles forming nanopowders, nanocomposites, layers, and thin films [72]. In amorphous Fe2O3, Fe3þ ions, distributed over nonperiodical lattice, are surrounded by oxygen octahedra, each of them having a respective symmetry axis that is oriented in a random manner (i.e., absence of periodical lattice) [72]. Thus, all spatial orientations of Fe3þ coordination octahedron are equally probable. From the magnetism viewpoint, amorphous Fe2O3 is described as a speromagnetic material with the Neel temperature reported in the temperature interval from 80 to 100 K [72]. In the (super)paramagnetic regime, irrespective of the synthetic procedure, the values of the effective magnetic moment per iron atom (meff) are found in the range from 2.5 to 3mB. The experimental values of meff are lower than 5mB that is a value of meff theoretically expected. This reduction of meff has been ascribed to formation of magnetic clusters and an extremely low particle size with a high contribution of the surface anisotropy. Besides, some researchers claim that such a low value of meff reflects the essence of magnetism of amorphous Fe2O3 (i.e., speromagnetism, spin-glass-like behavior). While speromagnetic regime is expected to be a correct description of the low-temperature magnetic state of individual (noninteracting) amorphous Fe2O3 nanoparticles, spin-glass-like properties evolve as amorphous Fe2O3 nanoparticles begin to magnetically interact [72]. It turns out that due to strong interparticle magnetic interactions acting among amorphous Fe2O3 nanoparticles in a nanopowder, the lowtemperature magnetically ordered regime of amorphous Fe2O3 is well described in terms of spin-glass-like behavior as superspins of amorphous Fe2O3 nanoparticles undergo a random cooperative freezing at well-defined transition temperature [72]. When measuring hysteresis loop of amorphous Fe2O3, at room temperature, no hysteresis is observed. In addition, magnetization versus magnetic field curve does not saturate even at high applied magnetic fields. As the temperature decreases below the characteristic ordering temperature of amorphous Fe2O3, irrespective of degree of interparticle magnetic interactions, hysteresis loops retain a nonsaturation profile. This nonsaturation character of amorphous Fe2O3 hysteresis loops is considered as evidence that Fe3þ magnetic moments are randomly oriented within amorphous Fe2O3 nanoparticle establishing an arrangement typical of spin-glass and cluster-spin-glass systems with competing exchange interactions below the spin freezing temperature [72]. At room temperature, the M€ ossbauer spectrum of amorphous Fe2O3 shows a broadened doublet (see Fig. 18.17) [72] with the typical values of the M€ ossbauer hyperfine parameters listed in Table 18.1. However, the approach how to fit the room-temperature spectrum of amorphous Fe2O3 differs. The room-temperature spectral profile is fitted either by a broadened doublet employing non-Lorentzian lines (under an assumption of an electric field gradient distribution) or by two doublets originating from the nonequivalent nanoparticle surface and core iron atoms [72]. The large value of DEQ parameter implies a large deviation of the octahedral environment of the probed Fe nucleus from a spherical symmetry.

FIGURE 18.17 Modeled typical zero-field room€ ssbauer spectemperature Mo trum of amorphous Fe2O3.

370

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.18 Modeled representative € ssbauer spectra of amorphous Mo Fe2O3 at a temperature of 5 K (a) without an external magnetic field and (b) under an external magnetic field of 5 T, oriented along the g-rays direction.

The distribution of DEQ parameter then reflects variously distorted coordination octahedra of Fe3þ ions. Below the ossbauer spectra display a broadened sextet (see Fig. 18.18a) that arises ordering temperature of amorphous Fe2O3, M€ due to a random variation of angles between the local Bhf and axis of local symmetry within an amorphous Fe2O3 nanoparticle. Thus, low-temperature M€ ossbauer spectra of amorphous Fe2O3 are regarded as a superposition of sextets corresponding to all of these angles. In comparison to crystalline a-Fe2O3 and g-Fe2O3, amorphous Fe2O3 exhibits a lowest value of Bhf documenting its amorphous nature [72]. Below the characteristic ordering temperature of amorphous Fe2O3 and irrespective of degree of interparticle magnetic interactions, there is no significant difference between the zero-field and corresponding in-field M€ ossbauer spectra measured at the same temperature (see Fig. 18.18) [72]. In other words, no remarkable change in A2,5 is observed if amorphous Fe2O3 nanosystem is exposed to Bext. A value of A2,5 ¼ 2 is preserved even at high applied fields (up to 8 T). Thus, amorphous Fe2O3 is more often described as a speromagnetic material rather than an antiferromagnet. Speromagnetic regime frequently occurs in amorphous and disorder materials the atomic magnetic moments of which freeze in essentially random orientations where all of them are equally probable due to absence of the principal crystallographic axes energetically favored by magnetocrystalline anisotropy. Since amorphous materials are characterized by random magnetic anisotropy (i.e., it changes from one atomic magnetic moment position to another), the speromagnetic state is governed by a balance between exchange magnetic interactions and this site-dependent magnetic anisotropy below the speromagnet ordering temperature. Thus, speromagnetic state is typical manifestation of spin disorder in the atomic-nearest-neighbor scale when the neighboring atomic magnetic moments are not correlated in orientations. In speromagnet, no long-range magnetic order is established under Bext and a random spin arrangement is preserved even under high Bext. As all orientations of atomic magnetic moments are, more or less, still equally probable under Bext, A2,5 ¼ 2 is observed. It turns out that due to finite-size and surface effects, it becomes very difficult to distinguish between amorphous Fe2O3 and nanocrystalline Fe2O3 phases (especially a-Fe2O3 and g-Fe2O3). This problematic issue has been recently addressed in the review work by Machala et al. [72]. It has been proposed that based on M€ ossbauer and magnetization measurements, there are three important experimental markers enabling to distinguish amorphous Fe2O3 and eventually quantify its content in the mixture sample. These include (i) saturation magnetization and/or magnetization under a

€ 18.3 USE OF Fe57 MOSSBAUER SPECTROSCOPY IN MONITORING SOLID-STATE REACTION MECHANISMS

371

maximum applied magnetic field (typically in the interval from 5 to 10 T), (ii) saturation value of Bhf, and (iii) the behavior of A2,5 under Bext [72]. If nanopowdered amorphous Fe2O3 is thermally treated under air, g-Fe2O3 is found to be a primary crystallization product [72]. The crystallization temperature varies from 550 to 1000 K depending on the average particle size, particle size distribution, and specific surface area of amorphous Fe2O3 nanoparticles. The crystallization temperature is enhanced (up to 1000 K) in the case of amorphous Fe2O3–SiO2 nanocomposites due to the growth-preventive role of the silica matrix. When amorphous Fe2O3 in the form of film and layers is heated under air, a-Fe2O3 and b-Fe2O3 are observed as primary crystallization products [72]; the crystallization takes place at temperatures higher than 550 K. On the contrary, if thermally induced crystallization of nanopowdered amorphous Fe2O3 proceeds under inert atmosphere (i.e., N2), Fe3O4 is formed as a primary product at temperatures higher than 580 K [72]. Under the reduction atmosphere, amorphous Fe2O3 nanopowder has been found as a suitable precursor for production of a-Fe and/or Fe3O4 nanoparticles with a narrow particle size distribution being dispersed in the MgO matrix (i.e., controlled thermal reduction of amorphous Fe2O3 nanopowder with nanocrystalline Mg under a hydrogen atmosphere) [74].

€ 18.3 USE OF 57 Fe MOSSBAUER SPECTROSCOPY IN MONITORING SOLID-STATE REACTION MECHANISMS TOWARD IRON OXIDES Thermally induced solid-state syntheses represent simple and cost-effective methods to prepare a large amount of iron oxide-based (nano)materials. Crystal structure, magnetism, particle size, and morphology of iron oxide product can be easily controlled through reaction conditions (temperature, atmosphere, duration) as well as through precursor properties (crystal structure, particle size, and morphology). The preliminary but valuable information on decomposition temperature, number of decomposition steps or evolved gases gives thermal analysis. Nevertheless, a full understanding of the mechanism of the solid-state reaction requires applying more experimental techniques such as X-ray powder ossbauer spectroscopy, microscopic techniques, and so on. Just 57 Fe M€ ossbauer spectroscopy diffraction, 57 Fe M€ constitutes a sensitive, selective (it “sees” only Fe-containing phases) and nondestructive experimental tool for complex structural and magnetic characterization of iron-containing solid phases. Usually, the mechanisms of thermally induced solid-state reactions are studied “ex situ” that means a characterization of the sample after its preparation. Thus, 57 Fe M€ ossbauer spectroscopy provides both qualitative and quantitative information about the precursor, possible reaction intermediates as well as resulting iron oxide phases. Nevertheless, the analysis of the samples prepared ex situ might be affected by a cooling process or aging. Therefore, more objective monitoring of the phase composition during the reaction is established within “in situ” approach, which is based on a recording the experimental data directly during the reaction. From the technical point of view, it requires to put a temperature-controlled furnace together with the sample into the spectrometer and record the data as continuously as possible. The in situ approach will be described and explained in more detail in Section 18.4.3. In the following parts of Section 18.3, representative examples of M€ ossbauer studies on the thermally induced decompositions toward ferric oxide nanoparticles will be presented. 18.3.1 Thermal Decomposition of Ammonium Ferrocyanide—A Valence Change Mechanism A considerable prerequisite to get nanoparticles of ferric oxide by means of the thermally induced solid-state reaction requires a low decomposition temperature of the precursor preventing undesirable growth of the particles (sintering process). Ammonium ferrocyanide (AmF), (NH4)4[Fe(CN)6], presents a coordination chemical compound, which is ossbauer spectroscopy thermally unstable even at 70  C. The results from thermal analysis (see Fig. 18.19) and M€ confirmed that decomposition of AmF proceeds through several steps and finishes at 250  C by formation of Fe2O3 nanoparticles. AmF precursor contains divalent low-spin iron atoms showing high-symmetry octahedral surroundings, which is reflected by the singlet M€ ossbauer spectrum. Heat treatment of AmF at 70  C resulted to evolution of one NH4 group and thus the formation of ammonium ferricyanide, (NH4)3[Fe(CN)6], with trivalent low-spin iron atoms. The absence of NH4 group in each AmF molecule affects the original symmetry of iron atoms and leads to nonzero quadrupole splitting in the corresponding M€ ossbauer spectrum. An increase in temperature above 130  C induces ossbauer evolution of CN group and formation of ammonium pentacyanoferrate(II), (NH4)3[Fe(CN)5]. Its doublet M€ spectrum exhibits a higher isomer shift in comparison with that of (NH4)3[Fe(CN)6] owing to the absence of CN group in the inner coordination sphere that causes a decrease of electron density in the iron nucleus. While the reduction of NH4þ ions to neutral NH3 is accompanied by the oxidation of iron atoms, oxidation of two CN ions to neutral (CN)2

372

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.19 Differential scanning calorimetry of ammonium ferrocyanide in air (2.5  C min 1).

molecule is compensated by the reduction of iron atoms. Such a “valence change” mechanism taking place in an oxidative atmosphere (air) is remarkable. Heat treatment of AmF at 160  C is sufficient for evolution of all of ammonium ions from ossbauer spectrum of PB consists of two wellthe structure and formation of Prussian Blue (PB), Fe4[Fe(CN)6]3. The M€ resolved components corresponding to divalent and trivalent iron atoms in the inner and outer coordination sphere, respectively. A representative room-temperature M€ ossbauer spectrum of the sample containing all three mentioned phases is depicted in Fig. 18.20. It is worthwhile to mention that no oxygen participated in the above-described decomposition mechanism of AmF though carried out in air. Another increase in temperature above 250  C induced a participation of oxygen in the decomposition reaction, all CN groups were evolved and amorphous Fe2O3 was identified as a final decomposition product. M€ ossbauer characterization as well as magnetic properties of amorphous Fe2O3 have been described in Section 18.2.

FIGURE 18.20 € ssbauer spectrum and its Mo deconvolution into four components corresponding to three phases obtained by thermal decomposition of ammonium ferrocyanide.

€ 18.3 USE OF Fe57 MOSSBAUER SPECTROSCOPY IN MONITORING SOLID-STATE REACTION MECHANISMS

373

FIGURE 18.21 € ssbauer spectra of the PB sample with Fe2þ and Fe3þ spectral components. Low-temperature Mo

Isothermal heating of AmF at 160  C for 3 h in air followed by dissolution of the intermediate phases and separation of insoluble residuum gives a possibility to synthesize pure PB nanoparticles (10–50 nm). Such nanopowder forms a threedimensional hollow assembly of nanoparticles resulting in a large surface area of 394 m2 g 1. PB is a mixed valence hexacyanoferrate of stoichiometry described by Fe43þ[Fe2þ(CN)6]3, where Fe3þ is in a high-spin state with S ¼ 5/2 and Fe2þ is in a low-spin state with S ¼ 0. Fe2þ ion is coordinated to six CN ligands at C-end while Fe3þ is coordinated by the same six CN ligands but at the N atom. Crystal structure of PB is described by as a face-centered cubic (group Fm3m,  a ¼ 10.2 A) one with regular vacancies in the sites of ferrocyanide anions [75]. Room-temperature M€ ossbauer spectrum of PB contains a singlet (d ¼ 0.14 mm s 1) and a doublet (d ¼ 0.42 mm 1 1 s , DEQ ¼ 0.32 mm s ) of the low-spin Fe2þ and the high-spin Fe3þ cations in the PB structure, respectively. A symmetrical octahedral environments of Fe2þ cations give a singlet subspectrum with the zero value of DEQ, whereas regular vacancies in the sites of ferrocyanide anions produce an asymmetry in the Fe3þ coordination sphere giving the doublet subspectrum [76–79]. Changes in magnetic behavior of PB with temperature were monitored by low temperature and in-field M€ ossbauer measurements (see Figs. 18.21 and 18.22, Table 18.2). At T ¼ 5 K, a considerable broadening of spectral lines of the Fe3þ doublet is evident indicating an initiation of a magnetic transition. Below 5 K, a magnetically split subspectrum with broad spectral lines is observed, however, the singlet spectral component corresponding to Fe2þ atoms remains. Indeed, it is known that superexchange interactions just between the neighboring Fe3þ ions through the NC–Fe2þ–CN linkages in the fcc structure of PB lead to a ferromagnetic ordering at low temperatures and Fe2þ ions do not contribute to a magnetism [80]. A ferromagnetic ordering is clearly reflected in a character of the M€ ossbauer spectrum measured at T ¼ 2.7 K in an external magnetic field of 4 T applied parallel to the propagation of g-rays (see Fig. 18.22). Intensities of the second and the fifth sextet spectral lines are significantly reduced as a result of almost collinear orientation of magnetic moments with respect to a direction of the external magnetic field. However, a spin frustration distorting the collinearity of magnetic moments at a surface of small PB nanoparticles contributes to a nonzero intensity of the second and the fifth sextet spectral lines. Furthermore, we can surprisingly ossbauer spectrum. Explanation of the observe a doublet Fe2þ subspectrum instead of the singlet one in the in-field M€ nonzero quadrupole splitting takes into account a distortion of the cubic PB structure induced by a strong interaction of

374

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.22 In-field (B ¼ 4 T, T ¼ 2.7 K) € ssbauer spectrum of the PB Mo sample with the Fe2þ and Fe3þ spectral components.

the external magnetic field with atomic magnetic moments resulting in asymmetry surroundings of Fe2þ atoms. Such phenomenon is observable, to a certain extent, even without an application of the external magnetic field due to Fe3þ– Fe3þ magnetic moment interactions through the NC–Fe2þ–CN linkages below the magnetic ordering temperature. A slight broadening of the singlet subspectrum with decreasing temperature is observed and even a doublet with a small value of quadrupole splitting can be observed at 1.5 K (see Table 18.2). It is also worthwhile to mention that the spectral area ratio of Fe2þ/Fe3þ takes the value being closest to the ideal value of 3/4, corresponding to the ratio of divalent to trivalent iron atoms in the PB crystal lattice, just at the lowest temperature of 1.5 K (see Table 18.2). At such low temperatures, Lamb–M€ ossbauer factor (or f-factor related to probability of M€ ossbauer effect) is almost identical for Fe2þ 3þ in the inner and Fe in the outer coordination sphere of the PB structure. 18.3.2 Thermal Decomposition of Prussian Blue in Air As has been mentioned in Section 18.3.1, PB, Fe4[Fe(CN)6]3, was confirmed to be an intermediate product of thermal decomposition of ammonium ferrocyanide in air. Heat treatment of PB above 250  C in air gives a possibility to prepare various structural forms of iron(III) oxide [81]. The decomposition of PB in air at minimum decomposition temperature of 250  C allowed the synthesis of amorphous Fe2O3 nanoparticles with a controllable size (see atomic force microscopy images in Fig. 18.23). The larger the PB particles were, the larger the size of the formed particles of amorphous Fe2O3 was. The nanoparticles exhibited unique properties including a very low particle size, uniform size distribution, an extremely large surface area of up to 400 m2 g 1, and specific magnetic behavior. At 350  C, nanocrystalline cubic b- and g-Fe2O3 polymorphs were identified as the primary decomposition products and their molar ratio was found to be

€ ssbauer Spectra of the PB Sample Measured at Various Temperatures TABLE 18.2 Hyperfine Parameters of Mo T (K) 300 10 5 2.7 1.5

Fe Cation 2þ

Fe Fe3þ Fe2þ Fe3þ Fe2þ Fe3þ Fe2þ Fe3þ Fe2þ Fe3þ

dFe  0.01 (mm s 1) 0.14 0.42 0.06 0.53 0.07 0.73 0.09 0.45 0.08 0.50

DEQ  0.01 (mm s 1) – 0.32 – 0.35 – 0.32 – 0.12 0.30 0.14

Bhf  0.3 (T)

FWHM  0.01 (mm s 1)

RA  1 (%)

– – – – – – – 46.5 – 52.0

0.35 0.48 0.41 0.52 0.44 1.57 0.64 1.67 0.61 1.13

47 53 51 49 57 43 35 65 40 60

dFe is the isomer shift related to metallic iron; DEQ denotes the quadrupole splitting; Bhf represents the hyperfine magnetic field; FWHM stands for full width at half maximum of the spectral line; and RA stands for relative subspectrum area.

€ 18.3 USE OF Fe57 MOSSBAUER SPECTROSCOPY IN MONITORING SOLID-STATE REACTION MECHANISMS

375

FIGURE 18.23 AFM images of amorphous Fe2O3 nanoparticles prepared by thermal decomposition of smaller (left) and larger (right) particles of PB. Attached are corresponding distributions of vertical dimensions. (Taken from Ref. 81 with permission of the American Chemical Society.)

strongly dependent on the reaction conditions and/or particle size of the PB precursor. As it is evidenced from the lowtemperature (T ¼ 150 K) M€ ossbauer spectra in Fig. 18.24, the relative content of g-Fe2O3 increased with increasing PB particle size. The low-temperature measurement allowed to resolve clearly magnetically ordered g-Fe2O3 from paramagnetic b-Fe2O3 still being paramagnetic at 150 K. The influence of PB particle size on the polymorphous

FIGURE 18.24 € ssbauer spectra of samples obtained by thermal decomposition of larger (left) Low-temperature (T ¼ 150 K) Mo and smaller (right) particles of PB at 350  C in air. The doublet corresponds to b-Fe2O3 and the sextet to g-Fe2O3. (Taken from Ref. 81 with permission of the American Chemical Society.)

376

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

composition of ferric oxide was interpreted by the preferential formation of b-Fe2O3 from the surface layer and g-Fe2O3 from the bulk of PB particles. Such interesting phenomenon of simultaneous formation of two different phases from one chemical compound is based on worsened conditions for conversion gas evolution from the bulk of particles. Above 400  C, both b-Fe2O3 and g-Fe2O3 polymorphs undergo structural transformation to the most thermally stable a-Fe2O3. 18.3.3 Thermal Conversion of Fe2(SO4)3 in Air—Polymorphous Exhibition of Fe2O3 Similarly as in the case of PB, b- and g-Fe2O3 polymorphs were identified as primary decomposition products of iron(III) sulfate in air, however, the decomposition temperature of about 530  C is much higher than for PB. The main difference consists in the mechanism of polymorphous transformation of g-Fe2O3, which proceeds via e-Fe2O3 toward a-Fe2O3. e-Fe2O3 is known as an intermediate phase, which is obtained from g-Fe2O3 when reaction conditions prevented a growth of g-Fe2O3 particles and thus direct transformation to a-Fe2O3. In most cases, e-Fe2O3 was prepared under space-limited conditions, for example, in the pores of SiO2 matrix [21]. The presence of e-Fe2O3 in the powdered sample was observed very rarely and can be explained by specific course of the decomposition of ferric sulfate and following polymorphous transformations of ferric oxide. Particularly, the kinetic study using M€ ossbauer spectroscopy revealed that a compact layer of Fe2O3 nanoparticles is formed on the surface of ferric sulfate particles in the initial stage of the decomposition. The surface layer inhibits the conversion gas evolution from the bulk of particles and causes a temporary moderation of the decomposition as well as the polymorphous transformations of Fe2O3. While b-Fe2O3 is preferentially formed at the surface layers of particles, g-Fe2O3 with a vacant structure is found as a conversion product in the bulk of sulfate particles, where worsened diffusion conditions for the conversion gas prevail. Moreover, the compact surface layer inhibits a growth of g-Fe2O3 particles that allows the thermally induced transformation via e-Fe2O3 to a-Fe2O3 to take place instead of a direct transformation to a-Fe2O3. Therefore, all four iron(III) oxide polymorphs can be acquired by the thermal decomposition of ferric sulfate, but mostly in the form of a mixture. The relative content of the polymorphs is controllable through reaction conditions including temperature, diffusion conditions for the conversion gas evolution, and particle size of the precursor. Nevertheless, it is practically impossible to achieve a single b-Fe2O3 and/ or e-Fe2O3 phases. As an example, room-temperature M€ ossbauer spectrum of the sample prepared by the thermal ossbauer conversion of Fe2(SO4)3
5H2O at 590  C is shown in Fig. 18.25. Figure 18.25 demonstrates ability of M€ spectroscopy to distinguish between various structural forms of Fe2O3. A possible overlap of the doublets related to paramagnetic b-Fe2O3 and superparamagnetic g-Fe2O3 was resolved by the low-temperature measurement at 150 K, where g-Fe2O3 is already magnetically ordered but b-Fe2O3 is still in a paramagnetic state. 18.3.4 Nanocrystalline Fe2O3 Catalyst from FeC2O4
2H2O Iron(II) oxalate dihydrate, FeC2O4
2H2O, constitutes another iron-containing precursor, which decomposes at a relatively low temperature toward Fe2O3 nanoparticles. In the study of Hermanek et al. [82], the changes in the phase

FIGURE 18.25 € ssbauer Room-temperature Mo spectrum of the sample prepared by thermal conversion of Fe2(SO4)3 at 590  C. a-Fe2O3 (45%), b-Fe2O3 (20%), and nonequivalent Fe positions of e-Fe2O3 (35%) were clearly identified.

€ 18.3 USE OF Fe57 MOSSBAUER SPECTROSCOPY IN MONITORING SOLID-STATE REACTION MECHANISMS

377

FIGURE 18.26 € ssbauer spectra of samples prepared by isothermal decomposition of FeC2O4
2H2O at Room-temperature Mo 180  C for 6 h (a), 30 h (b), 48 h (c), and 100 h (d). (Taken from Ref. 82 with permission of the American Chemical Society.)

composition, crystallinity, and size of iron(III) oxide nanoparticles prepared by the calcinations of FeC2O4
2H2O at ossbauer spectra (see Fig. 18.26). We 180  C in air for various times were clearly reflected in the room-temperature M€ can see that the spectral area of the magnetically split component gradually increases at the expense of the superparamagnetic doublet. The sample prepared by heating for 12 h is characterized by a rather broad doublet spectrum of a non-Lorentzian shape, reflecting a quadrupole splitting distribution (see Fig. 18.26a). The character of low temperature and in-field M€ ossbauer spectrum unambiguously proved a presence of pure amorphous Fe2O3. The room-temperature M€ ossbauer spectra of the samples heated for longer periods of time (30, 48, 100 h) reveal, beside the previously discussed doublet superparamagnetic component, a magnetically split pattern fitted by two sextets, which was ascribed to a-Fe2O3 particles formed by the gradual growth (recrystallization) of the primarily formed a-Fe2O3 nanograins (see Fig. 18.26). The presence of the two sextets was ascribed to the extremely small (5–7 nm) a-Fe2O3 particles yielding different hyperfine parameters for surface and bulk Fe(III) ions. The sextet with a higher value of the hyperfine magnetic field, having the isomer shift of 0.37 mm s 1 and quadrupole shift of 0.21 mm s 1 parameters close to those reported for highly crystalline a-Fe2O3, has been ascribed to the “bulk Fe(III)” ions [10]. The second sextet with considerably broaden lines and exhibiting a substantially lower hyperfine magnetic field corresponds to the “surface Fe(III)” ions. The data mentioned above prove an occurrence of two processes during sample calcination: (i) crystallization of amorphous Fe2O3 to a-Fe2O3 nanograins and (ii) thermally induced sintering of these grains and growth of particles. The as-prepared iron(III) oxide nanopowders were tested as a catalyst in decomposition of hydrogen peroxide [82]. The present literature data show that the surface area represents the principal parameter affecting the catalytic efficiency of iron oxides in decomposition of hydrogen peroxide. Values of the rate constant were determined in dependence on the specific surface area of the samples. Surprisingly, amorphous Fe2O3 sample with the largest specific surface area of 400 m2 g 1 exhibited the lowest rate constant among the measured phases. On the other hand, the highest rate constant of 26.41  10 3 min 1 (g l 1) 1 was achieved for nanocrystalline a-Fe2O3 sample with a specific surface area of 337 m2 g 1. Such an unmonotonous behavior of the dependence revealed a key role of increasing crystallinity having a dominant role over the catalytic effect of the surface area.

378

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

€ 18.4 VARIOUS MOSSBAUER SPECTROSCOPY TECHNIQUES IN STUDY OF APPLICATIONS RELATED TO NANOCRYSTALLINE IRON OXIDES 18.4.1

57

€ ssbauer Spectroscopy at Various Temperatures Fe Transmission Mo

57

Fe M€ ossbauer spectroscopy [11,83–86] is based on the recoil-free resonance fluorescence of g-rays observed with Fe atomic nuclei. This sensitive technique enables to measure nuclear energy levels to an extremely high accuracy (up to 13–15 decimals). Therefore, even the slight variation of nuclear energy levels caused by hyperfine interactions between the electrons and the nucleus can be detected. Hyperfine interactions reflect changes in the electronic, magnetic, geometric, or defect structure as well as in the lattice vibrations, serving as a basis for a variety of analytical applications. The M€ ossbauer effect is observable only when the atoms are in a solid lattice or matrix and/or in a frozen solution. The M€ ossbauer–Lamb factor (f) represents a measure of the probability of recoil-free resonance fluorescence of g-photons, which primarily depends on the mean-square amplitude of the vibration of the M€ ossbauer atoms in the lattice. There are three basic hyperfine interactions including monopole, quadrupole, and magnetic dipole interactions. Generally, hyperfine interactions induce a shift and/or splitting of energy levels in the M€ ossbauer nuclei. Arrangement of transmission M€ ossbauer spectroscopy is depicted in the part of Fig. 18.34 (see Section 18.4.4). Shift and/or splitting of energy levels in the M€ ossbauer nucleus can be compensated by a Doppler modulation of g-photons. Modulation is provided by a movement of the radioactive source toward or from the absorber (iron-containing sample). Intensity of g-rays transmitted through the sample is measured by a detector placed in an opposite side. Therefore, resonant absorption of g-photons by an absorber results in a local minimum for a given velocity of the source in the spectrum. In comparison with backscattering geometry, transmission M€ ossbauer spectrum provides integral information coming from a whole volume of the sample. What is important when preparing the sample for the transmission measurement, is a determination of an optimal effective thickness of the sample for a given material. There is a competition between concentration of M€ ossbauer nuclei and nonresonant absorption of g-photons taking place for too thick samples. M€ ossbauer spectroscopy measurements have been carrying out under various conditions to which a sample is exposed, including low temperature, high temperature, high pressure, or external magnetic field. Low-temperature measurements are usually performed utilizing various kinds of cryostats filled with a cooling medium (liquid nitrogen, liquid helium, gaseous helium in the closed-cycle cryostat). A transmission M€ ossbauer spectrometer equipped with a cryomagnetic system allowing measurements at low temperatures and in external magnetic field is displayed in Fig. 18.27. General advantage of the low-temperature measurements lies in a lower diffusion broadening of spectral lines, and a higher f, both leading to a better resolution of spectral components. It is worth mentioning that the isomer shift mostly linearly increases with decreasing temperature. Such a phenomenon, which has to be taken into account when interpreting the spectra, arises from the second-order Doppler shift. The series of low-temperature M€ ossbauer measurements provides valuable information about temperature and temperature range of magnetic transitions including the special cases such as blocking temperature of superparamagnetic material, Verwey transition in Fe3O4, or Morin 57

FIGURE 18.27 € ssbauer specTransmission Mo trometer equipped by cryomagnetic system operating with an external magnetic field up to 10 T and temperatures down to 1.5 K.

€ 18.4 VARIOUS MOSSBAUER SPECTROSCOPY TECHNIQUES

379

FIGURE 18.28 Schematic representations of various hyperfine interactions affecting the nuclear levels of a € ssbauer-active nucleus. Mo (Adapted from Ref. 94 with permission of Cambridge University Press.)

temperature in a-Fe2O3. Concerning nanocrystalline iron(III) oxides, interpretation of the M€ ossbauer spectrum measured at a suitably selected temperature enables an easy identification and quantification of various Fe2O3 polymorphs. As an example, when measuring superparamagnetic g-Fe2O3 close to its blocking temperature, a broad distribution of hyperfine magnetic field might appear. On the other hand, low-temperature spectrum obtained, for example, at T ¼ 25 K might show a slightly asymmetric g-Fe2O3 sextet (owing to two structural positions of Fe atoms) but with a thin spectral lines and without a presence of the distribution. M€ ossbauer spectroscopy of frozen solutions [87] is regarded as a special case of utilization of low-temperature measurements. Nevertheless, it is recommended to enrich ossbauer isotope in order to have a sufficient concentration of M€ ossbauer nuclei. the solution by 57 Fe M€ € ssbauer Spectroscopy 18.4.2 In-Field 57 Fe Transmission Mo If an external magnetic field is applied to a sample and a transmission M€ ossbauer spectrum is recorded, we get an in-field M€ ossbauer spectrum. The combination of transmission M€ ossbauer spectroscopy and external magnetic field is then referred to as in-field (or high-field) M€ ossbauer spectroscopy [88–94]. It is well known that a M€ ossbauer-active nucleus senses its local environment via direct and indirect exchange mechanisms (see Fig. 18.28) [11,94]. Direct interactions include (i) interaction between atomic shell and nucleus of a M€ ossbauer-active atom (interaction of Type I), (ii) interaction between the nucleus of the M€ ossbauer-active atom and its ossbauer-active atom surrounding lattice (interaction of Type III), and (iii) interaction between Bext and nucleus of the M€ (interaction of Type V). On the other hand, indirect interactions involve (i) interaction between the lattice and atomic shell of the respective nucleus of the M€ ossbauer-active atom (interaction of Type II) and (ii) interaction between Bext and atomic shell of the respective nucleus of the M€ ossbauer-active atom (interaction of Type IV). Thus, Bext can act directly or indirectly on the nucleus of the M€ ossbauer-active atom and affects its nuclear levels. Beside this, Bext can magnetically induce a quadrupole interaction that frequently happens, for example, in the case of Fe2þ having a cubic environment. Interaction of Type IV and Type V can be effectively described by Hamiltonians in the form of [94] $ Bext
ge
~ S H EZ ¼ mB~

(18.1)

Bext
~ gN mN~ I;

(18.2)

and H JZ ¼

respectively, where mB is the Bohr magneton, ge denotes the electron g-factor, S represents the atomic spin, gN stands for the nuclear g-factor, mN is the nuclear magneton, and I denotes the nuclear spin. In most cases, Bext is generated by superconducting magnets or Bitter magnets. Nowadays, it is not a technological problem to achieve magnetic fields with inductions higher than 15 T, however, the presence of Bext brings enhanced requirements on the experimental setup. It becomes necessary to shield Bext to avoid its influence on a transducer and source (Bext induces an additional splitting of nuclear levels in the radioactive source). Thus, suitable shielding materials are highly required and the effect of Bext on a transducer can be, following our experience, partly eliminated with an

380

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.29 Schematic representations of (a) parallel and (b) perpendicular measuring geometry, most frequently used in in-field € ssbauer spectroscopy. Mo

exploitation of digital feedback control. In general, the orientation of Bext can be arbitrary at the place of the sample, however, a geometry when Bext points parallel or perpendicular to the direction of propagation of g-rays is most frequently used with regard to the construction and interpretation viewpoints (see Fig. 18.29). Furthermore, we assume magnetic materials (such as iron(III) oxide) with a dominant magnetic dipolar interaction (a quadrupole interaction is treated as a perturbation) and having no internal texture. It is known that in general, Bhf is given as a vector sum of several components including the Fermi contact term (BFC—core s-electrons being polarized by the magnetic moment (driven by 3d-electrons in iron) of the M€ ossbauer-active atom), orbital magnetic field (Borb—produced by orbital moment of the M€ ossbauer-active atom, given by orbital motion of electron present in a partly filled valence shell), dipolar magnetic field (Bdip—dipolar interaction of spins of electrons in 3d (iron) and 4f electronic shells), conduction electron magnetic field (Bcon), dipolar magnetic field from the neighboring atoms (Bdip,na), conduction electron polarization magnetic field (Bcon,na) caused by neighboring atoms, transferred hyperfine magnetic field (Btsf), and supertransferred hyperfine magnetic field (Bstsf) [11,94]. In the case of iron(III) oxide with Fe3þ ions, Bhf is given only by negative BFC (Borb and Bdip are negligible and/or equal to zero) and thus the Fe3þ magnetic moment is oriented in an ossbauer spectrum, we almost opposite direction to Bhf. If an iron nucleus is then exposed to Bext, analyzing the in-field M€ get the so-called effective hyperfine magnetic field (Beff) that is a result of a vector sum of Bhf and Bext. Thus, Beff makes an angle u with Bext (see Fig. 18.30). Using basic trigonometry, we arrive at u ¼ arccos½ðB2eff B2ext B2hf Þ=ð2Bext Bhf ފ. On the other hand, taking into account the well-known Clebsch–Gordan coefficients reflecting the transition probabilities between Zeeman-split excited and ground states, u is defined as an angle that Beff makes with the direction of g-rays propagation. Then, u ¼ arcsin{[6r/(4 þ 3r)]1/2} with r ¼ A2,5/A1,6 where A1,6 is the spectral area of the first and the sixth sextet line and A2,5 is the spectral area of the second and the fifth sextet line. These two relations can be, thus, used to estimate the value of u and, in the case of Fe3þ ions, determine the direction of Fe3þ magnetic moment under Bext. In-field M€ ossbauer spectroscopy can be used as a powerful tool to monitor the arrangement of magnetic moments of M€ ossbauer-active atoms under external magnetic fields and thus to distinguish between various magnetic states of matter [88,89,92–94]. Furthermore, in monocrystalline magnetically ordered samples, use of in-field M€ ossbauer spectroscopy provides possibility to determine the orientation of axis of resulting magnetically ordered spin system with regard to the crystallographically significant (magnetic anisotropy favored) crystal axes. Thus, in-field M€ ossbauer spectra reveal information on a relative alignment of magnetic moments of neighboring atoms or ions and directions of sample magnetic anisotropy [94]. In case of diamagnetic materials, Beff ¼ Bext, whereas for paramagnetic materials, Bext-induced magnetic splitting is observed if the relaxation of magnetic moments is slower compared to the characteristic measuring time of

FIGURE 18.30 Definition of effective hyperfine magnetic field (Beff) and its relation to hyperfine magnetic field (Bhf) and external magnetic field (Bext).

€ 18.4 VARIOUS MOSSBAUER SPECTROSCOPY TECHNIQUES

381

M€ ossbauer technique or Bext is sufficient to induce magnetization in a paramagnetic material. In magnetically ordered materials, the alignment of magnetic moments results from an energy competition between three terms, that is, ossbauer spectra then give information exchange energy, anisotropy energy, and energy from Bext. Magnetically split M€ on alignment of magnetic moments taking into account (i) relative line sextet intensities (parameters depending on a relative orientation of Bext with a direction of observation through polarization effects) and (ii) sextet line positions (parameters depending on a magnitude of Bext) [88]. This makes possible to distinguish ferromagnetic (no splitting of resonant lines in Bext, A2,5 ¼ 0 in the parallel experimental geometry), collinear (increase of A2,5 above certain value of Bext and certain Bext-orientation with respect to material antiferromagnetic axis, observation of jump spin reorientation phenomenon) and canted (increase of A2,5 above certain value of Bext and certain Bext-orientation with respect to material antiferromagnetic axis, observation of continuous spin reorientation phenomenon), antiferromagnetic, ferrimagnetic (splitting of resonant lines, A2,5 ¼ 0), and helimagnetic materials. Beside crystalline magnetic materials, in-field M€ ossbauer spectroscopy can be also exploited to study magnetically disorder materials such as spin glasses and/or amorphous materials and alloys exhibiting speromagnetic, asperomagnetic, and sperimagnetic arrangements (for details, see Refs. 88,89,92–94). If the sample is composed of a mixture of various polymorphs of iron(III) oxide, all spectral contributions do not have to be resolved in zero-field M€ ossbauer spectra due to similar magnitudes of M€ ossbauer hyperfine parameters of Fe2O3 polymorphs, especially in the case of their nanoobjects. Since all of Fe2O3 polymorphs exhibit different magnetic regimes, ossbauer spectra. This then if exposed to Bext, their different magnetic behaviors are reflected differently in in-field M€ enables to evaluate the purity of the sample and quantify the portion of individual phases in the sample. Zero-field M€ ossbauer spectra can be, thus, refitted taking into account information acquired from the analysis of corresponding infield M€ ossbauer spectra. For nanoobjects, spin disorder often occurs when placed into Bext. If present, the behavior of A2,5 deviate from that expected for bulk materials (with no spin disorder). A nonzero A2,5 can be then used to quantify the degree of spin disorder (through spin canting under Bext) and determine the portion of disordered magnetic moments within a nanoobject. Thus, M€ ossbauer spectroscopy in combination with external magnetic fields provides a complex physicochemical characterization of studied nanosized iron(III) oxides since it enables (i) more precise determination of phase compositions; (ii) fitting zero-field M€ ossbauer spectra on the basis of phase analysis obtained from in-field M€ ossbauer spectra; (iii) separation of individual contributions from various crystallographic sites; (iv) determination of direction of Beff and magnetic moments of Fe3þ ions in iron(III) oxides; and (v) study of spin-canting phenomenon and determination of various spin noncollinearities (e.g., spin frustration, spin-glass-like behavior, etc.). € ssbauer Spectroscopy 18.4.3 In Situ High-Temperature 57 Fe Transmission Mo The use of M€ ossbauer spectroscopy can be generally divided into ex situ and in situ. In ex situ M€ ossbauer spectroscopy, the samples are prepared outside the sample holder before performing the measurements. The in situ approach allows a possibility of direct monitoring of phase composition during physically induced processes such as temperature, pressure, or humid air. Moreover, the sample can be protected from ambient air to avoid possible chemical reactions of the sample ossbauer spectroscopy has, thus, shown to play a significant role in studying with air-H2O and/or CO2. The in situ M€ electrochemical processes [95–100], chemical reactions [101–104], catalysis [105,106], and in situ syntheses [107–110]. ossbauer spectroscopy (CEMS) In recent years, thin films were characterized by using in situ 57 Fe conversion electron M€ [111,112]. The measurements can be carried out under various conditions including high temperature, high pressure, and various atmospheres (e.g., inert, oxygen, and nitrogen). Importantly, M€ ossbauer spectra acquired by in situ M€ ossbauer spectroscopy measurements require proper precautions. First, the change in the phase composition during the measurements should be slow with respect to the time of recording of M€ ossbauer data. Second, M€ ossbauer spectra obtained under in situ conditions should be evaluated qualitatively because of expected changes in the character of the M€ ossbauer spectrum. Nevertheless, a periodic recording of M€ ossbauer spectrum enables to determine the kinetics of a process. Thirdly, in case of hightemperature measurements, the time needed for acquisition of a “high-quality” M€ ossbauer spectrum is negatively influenced by a lower value of the M€ ossbauer–Lamb factor [113]. The possible experimental setup to perform in situ high-temperature transmission M€ ossbauer spectra is shown in Fig. 18.31. A powdered sample is uniformly spread on quartz wool and put into a quartz tube. The tube is then inserted into the special furnace (see Fig. 18.31). The furnace was placed between source and detector of the M€ ossbauer spectrometer. The temperature inside the furnace was regulated accurately by PID (Proportional-Integral-Derivative)

382

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.31 € ssbauer The furnace for in situ Mo spectroscopy put between g-rays source and detector of the € ssbauer spectrometer. Mo

control. A tube providing access of flowing gases (e.g., N2, O2, humid air) is connected to furnace in order to have special atmospheric conditions if needed. ossbauer spectroscopy has been applied to study a thermal decomposition of The in situ high-temperature 57 Fe M€ potassium ferrate(VI) (K2FeO4). Potassium ferrate(VI) represents the most often studied and used chemical compound with hexavalent iron atoms. During the last decade, interest in K2FeO4 has been increasing empowered particularly due to its potential use in high energy density rechargeable batteries [114], in cleaner (“greener”) technology for organic synthesis [115], and in treatment of contaminants and toxins in water and wastewater [116,117]. This interest led to an effort to synthesize K2FeO4 in a large scale and with a high purity. There are wet (chemical and electrochemical) and dry (thermal) techniques to prepare solid K2FeO4. Wet techniques have drawback because of the reaction of K2FeO4 with water [118]. Comparatively, a dry technique is attractive as it can address difficulties associated with the wet techniques. In such a thermally induced solid-state process, potassium-containing salt and a nanocrystalline iron(III) oxide are heated together to obtain K2FeO4. However, the decomposition of ferrate(VI) occurs at elevated temperatures used in a thermal synthesis technique, which results in a low yield of K2FeO4 product. An increase in the K2FeO4 yield may be achieved by optimizing the temperature conditions under which prevention of the secondary decomposition could occur. This would, thus, require understanding the mechanism of the thermal decomposition of K2FeO4. The main reason for a realization of in situ high-temperature M€ ossbauer measurement on K2FeO4 sample was to explain instability of the primary decomposition product (KFeO2) at room temperature and under humid air conditions. Therefore, we have collected in situ high-temperature M€ ossbauer spectra at 190  C under static air conditions to understand the transformation process of K2FeO4 and to identify the iron-bearing conversion intermediates [119]. The M€ ossbauer spectrum of the sample heated at 190  C is shown in Fig. 18.32a, which consists of two components including a singlet (d ¼ 1.01 mm s 1) and a sextet with d ¼ 0.06 mm s 1, eQ ¼ 0.06 mm s 1, and Bhf ¼ 45.8 T. The singlet corresponds to nontransformed K2FeO4 while the isomer shift value of the latter subspectrum is much lower than that expected for an octahedral high-spin iron(III) compound. Using hyperfine parameters values, the subspectrum was assigned to potassium iron(III) oxide, KFeO2, where Fe3þ ions are tetrahedrally coordinated [120,121]. After cooling the sample to room temperature, and then exposing it to ambient air for 3 days, the M€ ossbauer spectrum at room temperature was totally different (see Fig. 18.32b). The M€ ossbauer spectrum showed the doublet with hyperfine parameters d ¼ 0.35 mm s 1 and DEQ ¼ 0.72 mm s 1, which are peculiar for (super)paramagnetic iron(III) oxides or oxyhydroxides. Overall, the interpretation based on the results from in situ high-temperature M€ ossbauer spectroscopy, in situ variable temperature X-ray powder diffraction and thermal analysis is that the K2FeO4 decomposes above 190  C in static air to potassium iron(III) oxide (KFeO2) and equimolar mixture of potassium oxide (K2O) and potassium superoxide (KO2). Upon sample cooling, KFeO2 reacts with CO2 and H2O present in air to give Fe2O3 nanoparticles and KHCO3 [119]. Controlled aging of KFeO2, thus, presents a simple method to prepare monodispersed superparamagnetic g-Fe2O3 nanoparticles. Recently, Machala et al. revealed [122] that potassium ferrate(VI) is, similarly as potassium iron(III) oxide, also instable in humid air and decomposes to KHCO3 and Fe(OH)3. The kinetic data obtained from in situ variable air-humidity

€ 18.4 VARIOUS MOSSBAUER SPECTROSCOPY TECHNIQUES

383

FIGURE 18.32 € ssbauer spectrum of potassium ferrate(VI) and (b) room-temperature (a) In situ high-temperature (190  C) Mo € ssbauer spectrum measured after cooling the potassium ferrate(VI) sample and exposing it to ambient air for 3 Mo days. (Adapted from Ref. 119 with permission of the American Chemical Society.)

M€ ossbauer spectroscopy showed a significant increase in the decomposition rate with increasing air humidity. Moreover, at lower humidity levels (55–70%), the authors [122] observed anomalous two-step kinetics of the ferrate(VI) transformation, with the first nearly linear slow step followed by the second significantly faster step, the beginning of which is dependent on the relative humidity (see Fig. 18.33). Such two-stage kinetics was explained by formation of a narrow amorphous layer of reaction products on the surface of the ferrate(VI) crystal. This layer probably prevents the access of gaseous reactants to the ferrate(VI) crystal for some time. It seems that crystallization of hygroscopic KHCO3 plays a key role in the change of the reaction kinetics as it probably causes destruction of the layer, thus allowing the massive access of water and carbon dioxide to the reaction system. 18.4.4

57

€ ssbauer Spectroscopy Fe Conversion Electron and Conversion X-Ray Mo

All of already mentioned M€ ossbauer techniques were arranged in so-called transmission geometry, where g-rays coming from 57 Co radioactive source pass through the whole axial length of the sample (see Fig. 18.34). In these cases, we can observe counting minima found at resonant velocities of the collected M€ ossbauer spectra. This is a result of the fact that resonant g-photons have the same emission probability in each direction (4p-angle), while the emission of nonresonant photons is most probable in direct line (source–absorber–detector) and contributes to the background counting. Contrary to it, backscattering M€ ossbauer spectroscopy (BMS) layout is different; the radiation is detected not behind, but in front of the specimen (taken from source point of view). Resonant maxima occur in the recorded M€ ossbauer spectra. Let us describe consecutive processes following nuclear resonant absorption of 14.41 keV g-photon (see Fig. 18.35) [123–126]. An excited nucleus can reach its ground state by recoilless emission of g-quantum with the same energy or through an emission of so-called conversion electron. In that event, excitation energy is transferred via internal electromagnetic coupling from nucleus to core-shell electron, which is “kicked out” from its binding and emitted. It has to be mentioned that emission of K-conversion electrons is approximately eight times stronger than emission of resonant g-photons. Escape kinetic energy of conversion electron corresponds to energy of excited state reduced by electron binding energy. A radius of released electron in material surrounding is very limited due to energetic losses during collisions. For example, 7.3 keV K-conversion electron does not get over 300 nm. There is a positive hole in the core shell (K-shell in the most cases) left after an escape of the electron. Filling of this gap takes place by an electron jump down from a higher energy level and excessive amount of energy is radiated in a form of characteristic X-rays (e.g., 6.3 keV Ka characteristic X-rays with a probing depth of 1–10 mm). Part of X-rays also undergoes an internal conversion joined with an Auger electron emission. Moreover, above-mentioned high-energy resonant transitions are accompanied with lowenergy Auger electrons, shake-off electrons, and true secondary electrons. These low-energy emissions represent 77% of all emitted electrons. Nonresonant photoelectrons, Compton electrons, and photons are regarded as major

384

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.33 Kinetics of K2FeO4 aging expressed as dependence of € ssbauer spectral area on relaMo tive humidity (a: lowest relative humidity; b: higher relative humidity levels; c: dependence of the second step decay rate on relative humidity). (Taken from Ref. 122 with permission of Wiley.)

contributors to the background counting in a M€ ossbauer spectrum, but fortunately, these types of radiation do not dominate the effective signal. All of mentioned irradiative as well as radiationless resonant events enable us to track the M€ ossbauer effect; thus, there is quite wide number of BMS conceptions based on a type of the detected radiation. These include CEMS, conversion X-ray M€ ossbauer spectroscopy (CXMS), backscattering g-ray M€ ossbauer spectroscopy (BGMS), Auger electron M€ ossbauer spectroscopy (AEMS), and low-energy electron M€ ossbauer spectroscopy (LEEMS). Both first referred spectroscopic approaches will be further discussed. 57 Fe CEMS and CXMS are excellent tools for surface investigation of iron-containing specimens. Their particular analytical potential lies in a depth selectivity and record sensitivity (1013–1014 of 57 Fe atoms cm 2). Thus, material surface layers, thin films, and core-shell powders can be characterized. These M€ ossbauer techniques can bring reliable information about phase composition and magnetic structure of materials, for example, about corrosive processes and their inhibitions or on properties of wrought surfaces. Both can serve as a supporting method in nanotechnology, material physics, chemistry, metallurgy, mineralogy, geology, and archeometry.

€ 18.4 VARIOUS MOSSBAUER SPECTROSCOPY TECHNIQUES

385

FIGURE 18.34 Scheme showing the difference in positioning of detector with respect to specimen and radiation source in the case of TMS and BMS.

Commonly, conversion electrons are counted using energetically independent technique called integral conversion electron M€ ossbauer spectroscopy (ICEMS). In this manner, conversion electrons are collected in a wide energy window from the whole 300 nm thick layer. Contrary to it, differential or depth selective (DCEMS) approach exploits a weight function TE(x) that calculates the probability that conversion electron originated in a depth of x reaches the surface with an energy of E. Narrow energy counting windows make possible to finely differentiate 300 nm layer into sublayers.

FIGURE 18.35 € ssbauer absorption (100%). The diagram summarizing consecutive processes taking place after an individual Mo

386

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.36 General view of CEMS 2010 spectrometer based on continuous gas flow counter.

DCEMS together with low-temperature and in-field measurements require much more sophisticated experimental equipment. Gas counters, scintillation detectors, electron multipliers (Channeltron, Ceratron), surface barrier silicon semiconductor detector, and electron energy analyzers belong to the most frequently used detectors applied in CEMS and CXMS, respectively. Differences among individual constructions can be found in Ref. 123. Commercially available version of CEMS/CXMS spectrometer is depicted in Fig. 18.36 [127]. Device is based on 2p proportional continuous gas flow counter for room-temperature zero-magnetic field measurements. Let us have a look at a few examples of CEMS usefulness. First, CEMS has been exploited in a study of phase composition of iron(III) oxide films deposited by ultrasonic spray pyrolysis (USP) onto FTO-coated glass (SnO2:F). Hematite (a-Fe2O3) films are developed with an aim to obtain semitransparent nanostructured photoelectrodes for direct photoelectrolysis of water by visible sunlight [128–132]. Iron(III) oxide phase occurrence was crucially dependent on substrate temperature during deposition process (pyrolysis run at a temperature range of 330–500  C). Phases other than a-Fe2O3 act as impurities in these systems and their presence decreases the photocatalytic activity of the system. Percentage phase content was measured by two independent methods, micro-Raman spectroscopy and CEMS, and obtained results were in very good agreement [132]. The most complex phase mixture, consisting of a-, b-, g- and ossbauer spectroscopy turned to be a superior method as 12% e-Fe2O3, was obtained at 440  C (see Fig. 18.37a). M€ content of e-Fe2O3 was also found. Pure a-Fe2O3 was not produced under 500  C (see Fig. 18.37b). Resulting temperature-dependent phase composition taken from CEMS spectra of the whole rank of specimens can be seen in Fig. 18.37c. Second, CEMS was used to observe Sn-doping act in FeOCl film—an intermediate during production of a-Fe2O3 film described in Ref. 133 (look at Fig. 18.38 to see an attractive morphology of both films). Tin atoms are thermally diffusing from FTO layer during the growth of the film. A small amount of dopant is sufficient to make observable changes in the isomer shift (see Fig. 18.39). Its value is shifted from a tabulated value of 0.40 mm s 1 [134] to 0.37 mm s 1. Tin atoms are diffused into hematite structure from SnO2:F. The doping was confirmed by XPS data. Note that a presence of appropriate dopant represents one of the crucial factors for obtaining good photocatalytic activity results.

€ 18.4 VARIOUS MOSSBAUER SPECTROSCOPY TECHNIQUES

387

FIGURE 18.37 (a) CEMS recording of the sample deposited by USP at T ¼ 440  C. It reveals a mixture of all known iron(III) oxide phases: 26% a-, 50% b-, 12% superparamagnetic g-, and 12% e-Fe2O3. (b) Sextet assigned to pure a-Fe2O3 (T ¼ 500  C). (c) Dependence of substrate phase composition on temperature. Percentage contents of individual phases were derived from proportions of spectral areas belonging to individual subspectra. (Adapted from Ref. 132.)

FIGURE 18.38 SEM visualization of (a) FeOCl intermediate and (b) final a-Fe2O3 film. Single crystals are assembled into cactus-fieldlike morphology with (a) compact and (b) porous sheets, respectively.

388

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

FIGURE 18.39 (a) CEMS spectra of undoped FeOCl film deposited onto uncoated glass substrate (d ¼ 0.40 mm s 1). (b) Sn-doped FeOCl film on FTO-coated glass (d ¼ 0.33 mm s 1). (Taken from Ref. 133 with permission of Elsevier.)

Finally, CEMS has been successfully employed in the course of investigation of surface crystallization of amorphous nanocrystalline metal alloy. Thin ribbons made up of this material are well applicable as transformer sheets or magnetic field shielding due to magnetically soft behavior [135,136]. Proper thermal treatment of these strips brings surface crystallization. It is notable that outside crystallized layer can limit undesirable vortex current inside transformer core. As it was already mentioned, main difference in CEMS and CXMS lies in the probing depth. By comparing CEMS and CXMS spectra (see Fig. 18.40) of the same sample, one can deduce that a slightly higher content of crystalline (Fe3O4) phase in the CEMS spectrum is caused by crystallization of top slice, while ribbon core stays unchanged. An axial depth of the sample is approximately 20 mm. Conversion X-ray with its probing depth up to 10 mm brings information from one-half of the overall thickness, while CEMS is limited only to 300 nm and thus CXMS data are far integral.

FIGURE 18.40 A comparison of (a) CEMS and (b) CXMS registration of identical amorphous metallic alloy that was thermally treated at T ¼ 370  C for 30 min. Both spectra differ in the phase content. A higher amount of crystalline component in the CEMS spectrum implies that crystallization is of a surface origin.

REFERENCES

389

18.5 CONCLUSIONS In this chapter, 57 Fe M€ ossbauer spectroscopy have been presented as a powerful technique exploited for studying thermal syntheses, identification, structural and magnetic characterization, and polymorphous transformations of iron ossbauer spectroscopy as “a local probing method” is very valuable particularly in the case of (III) oxide. 57 Fe M€ investigations of nanocrystalline iron(III) oxides and to distinguish between crystalline phases and amorphous form of ossbauer spectroscopy technique is not limited by particle size, which is the case of Fe2O3. It has been emphasized that M€ standardly used X-ray powder diffraction. Thanks to unusually short characteristic time of measurement and high ossbauer spectroscopy is a suitable method for the sensitivity to local environment of the probed iron nucleus, 57 Fe M€ study of the special phenomena observed in nanosized iron oxides such as superparamagnetism, collective magnetic excitations, spin disorders and spin frustrations, interparticle interactions effects, and/or structural transformationassisted magnetic transitions, that is, phenomena that are manifested characteristically in temperature- and fielddependent M€ ossbauer hyperfine parameters of particular iron(III) oxide. We have also discussed and illustrated a variability of M€ ossbauer spectroscopy experiments with respect to applied conditions and experimental arrangements. Nevertheless, to obtain a complex characterization of a (nano)material, a combination of M€ ossbauer spectroscopy with other experimental techniques such as X-ray powder diffraction, conventional magnetization measurements, transmission and/or scanning electron microscopy is highly encouraged.

ACKNOWLEDGMENTS This work has been supported by the Operational Program Research and Development for Innovations—European Regional Development Fund (CZ.1.05/2.1.00/03.0058 of the Ministry of Education, Youth and Sports of the Czech Republic), the Operational Program Education for Competitiveness—European Social Fund (CZ.1.07/2.3.00/20.0017 and CZ.1.07/2.3.00/20.0058 of the Ministry of Education, Youth and Sports of the Czech Republic), the projects of the Ministry of Education of the Czech Republic (1M6198959201 and MSM6198959218), and the project of the Academy of Sciences of the Czech Republic (KAN115600801).

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

A.-H. Lu, E.L. Salabas, F. Schuth, Angew. Chem. Int. Ed. 2007, 46, 1222–1244. R.M. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses, Wiley-VCH, Weinheim, 2003. Q.A. Pankhurst, J. Connolly, S.K. Jones, J.J. Dobson, J. Phys. D: Appl. Phys. 2003, 36, R167–R181. P. Tartaj, M.D. Morales, S. Veintemillas-Verdaguer, T. Gonzalez-Carreno, C.J. Serna, J. Phys. D: Appl. Phys. 2003, 36, R182–R197. C.C. Berry, A.S.G. Curtis, J. Phys. D: Appl. Phys. 2003, 36, R198–R206. A.K. Gupta, M. Gupta, Biomaterials 2005, 26, 3995–4021. C.J. Xu, S.H. Sun, Polymer Int. 2007, 56, 821–826. J. L. Dormann, D. Fiorani, E. Tronc, in Advances in Chemical Physics, Vol. 98, I. Prigogine, S.A. Rice, eds., Wiley, New York, 1997, pp. 283–494. R.E. Vandenberghe, E. De Grave, C. Landuydt, L.H. Bowen, Hyperfine Interact. 1990, 53, 175–196. R. Zboril, M. Mashlan, D. Petridis, Chem. Mater. 2002, 14, 969–982. N.N. Greenwood, T.C. Gibb, M€ossbauer Spectroscopy, Chapman & Hall, London, 1971. R.M. Hazen, R. Jeanloz, Rev. Geophys. 1984, 22, 37–46. F. Walz, J. Phys. Condens. Matter 2002, 14, R285–R340. J. Garcia, G. Subias, J. Phys. Condens. Matter 2004, 16, R145–R178. O. Ozdemir, D.J. Dunlop, B.M. Moskowitz, Geophys. Res. Lett. 1993, 20, 1671–1674. V.A.M. Brabers, F. Walz, H. Kronmuller, Phys. Rev. B 1998, 58, 14163–14166. G.F. Goya, T.S. Berquo, F.C. Fonseca, M.P. Morales, J. Appl. Phys. 2003, 94, 3520–3528. J. Tuc9ek, R. Zboril, D. Petridis, J. Nanosci. Nanotechnol. 2006, 6, 926–947.

390

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

J. Tuc9ek, R. Zbo9ril, A. Namai, S. Ohkoshi, Chem. Mater. 2010, 22, 6483. M. Gich, A. Roig, E. Taboada, E. Molins, C. Bonafos, E. Snoeck, Faraday Discuss. 2007, 136, 345–354. S. Sakurai, A. Namai, K. Hashimoto, S. Ohkoshi, J. Am. Chem. Soc. 2009, 131, 18299–18303. M. Bonnevie-Svendsen, Naturwissenschaften 1956, 45, 542. T. Gonzalez-Carreno, M.P. Morales, C.J. Serna, J. Mater. Sci. Lett. 1994, 13, 381–382. L. Ben-Dor, E. Fischbein, I. Felner, Z. Kalman, J. Electrochem. Soc. 1977, 124, 451–457. J.S. Lee, S.S. Im, C.W. Lee, J.H. Yu, Y.H. Choa, S.T. Oh, J. Nanoparticle Res. 2004, 6, 627–631. A. Kumar, A. Singhal, Nanotechnology 2007, 18, 475703. A.H. Morrish, Canted Antiferromagnetism: Hematite, World Scientific Publishing Company, Singapore, 1994. R.K. Quinn, R.D. Nasby, R.J. Baughman, Mater. Res. Bull. 1976, 11, 1011–1018. A. Kay, I. Cesar, M. Gratzel, J. Am. Chem. Soc. 2006, 128, 15714–15721. G.A. Somorjai, M. Salmeron, in Homogeneous and Heterogeneous Photocatalysis, Vol. 174, E. Pelizzetti, N. Sermone, eds., NATO ASI Series C, D. Seidel Publ. Co., Dordrecht, The Netherlands, 1986, pp. 445–478. Y.S. Hu, A. Kleiman-Shwarsctein, A.J. Forman, D. Hazen, J.N. Park, E.W. McFarland, Chem. Mater. 2008, 20, 3803–3805. J.A. Glasscock, P.R.F. Barnes, I.C. Plumb, N. Savvides, J. Phys. Chem. C 2007, 111, 16477–16488. A. Kleiman-Shwarsctein, Y.S. Hu, A.J. Forman, G.D. Stucky, E.W. McFarland, J. Phys. Chem. C 2008, 112, 15900–15907. J. Kiwi, M. Gr€atzel, J. Chem. Soc., Faraday Trans. I 1987, 83, 1101–1108. N. Beermann, L. Vayassieres, S. E. Lindquist, A. Hagfeldt, J. Electrochem. Soc. 2000, 147, 2456–2461. I. Dzyaloshinsky, J. Phys. Chem. Solids 1958, 4, 241–255. T. Moriya, Phys. Rev. 1960, 120, 91–98. F.J. Morin, Phys. Rev. 1950, 78, 819–820. J.O. Artman, J.C. Murphy, S. Foner, Phys. Rev. 1965, 138, A912–A919. R.D. Zysler, D. Fiorani, A.M. Testa, L. Suber, E. Agostinelli, M. Godinho, Phys. Rev. B 2003, 68, 212408. D.W. Strangwa, B.E. McMahon, R.M. Honea, E.E. Larson, Earth Planet. Sci. Lett. 1967, 2, 367. D.G. Rancourt, S.R. Julian, J.M. Daniels, J. Magn. Magn. Mater. 1985, 49, 305–316. M.D. Bentzen, J. van Wonterghem, S. Morup, A. Tholen, C.J.W. Koch, Philos. Mag. B 1989, 60, 169–178. O. Nikolov, T. Ruskov, T. Tomov, A. Toshev, Hyperfine Interact. 1988, 39, 409–417. E. Van San, E. De Grave, R.E. Vandenberghe, J. Magn. Magn. Mater. 2004, 269, 54–60. L. Ben-Dor, E. Fischbein, Acta Crystallogr. B 1976, 32, 667. E.R. Bauminger, L. Ben-Dor, I. Felner, E. Fischbein, I. Nowik, S. Ofer, Physica B 1977, 86–88, 910–912. Y. Ikeda, M. Takano, Y. Bando, Bull. Inst. Chem. Res. Kyoto Univ. 1986, 64, 249–258. D. Wiarda, T. Wenzel, M. Uhrmacher, K.P. Lieb, J. Phys. Chem. Solids 1992, 53, 1199–1209. D. Wiarda, G. Weyer, Int. J. Mod. Phys. B 1993, 7, 353–356. O. Baudisch, L.A. Welo, Naturwissenschaften 1925, 13, 749. R. Dronskowski, Adv. Funct. Mater. 2001, 11, 27–29. X. Batlle, A. Labarta, J. Phys. D: Appl. Phys. 2002, 35, R15–R42. Q.A. Pankhurst, R.J. Pollard, J. Phys. Condens. Matter 1993, 5, 8487–8508. D.L. Leslie-Pelecky, R.D. Rieke, Chem. Mater. 1996, 8, 1770–1783. L. Neel, Ann. Geophys. 1949, 5, 99–136. C.P. Bean, J. Appl. Phys. 1955, 26, 1381–1383. J.L. Dormann, R. Cherkaoui, L. Spinu, M. Nogues, F. Lucari, F. D’Orazio, D. Fiorani, A. Garcia, E. Tronc, J.P. Jolivet, J. Magn. Magn. Mater. 1998, 187, L139–L144. J.L. Dormann, D. Fiorani, R. Cherkaoui, L. Spinu, F. Lucari, F. D’Orazio, M. Nogues, E. Tronc, J.P. Jolivet, A. Garcia, Nanostruct. Mater. 1999, 12, 757–762. S. Mørup, H. Topsøe, Appl. Phys. 1976, 11, 63–66. S. Mørup, J. Magn. Magn. Mater. 1983, 37, 39–50. S. Mørup, F. Bødker, P.V. Hendriksen, S. Linderoth, Phys. Rev. B 1995, 52, 287–294. J.M.D. Coey, Phys. Rev. Lett. 1971, 27, 1140–1142. M.P. Morales, S. Veintemillas-Verdaguer, M.I. Montero, C.J. Serna, A. Roig, L. Casas, B. Martinez, F. Sandiumenge, Chem. Mater. 1999, 11, 3058–3064.

31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64.

REFERENCES

65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109.

391

H. Forestier, G.C. Guiot-Guillain, C. R. Acad. Sci. 1934, 199, 720. R. Schrader, G. Buttner, Z. Anorg. Allg. Chem. 1963, 320, 220–234. J. Jin, S. Ohkoshi, K. Hashimoto, Adv. Mater. 2004, 16, 48–51. M. Gich, C. Frontera, A. Roig, J. Fontcuberta, E. Molins, N. Bellido, C. Simon, C. Fleta, Nanotechnology 2006, 17, 687–691. S. Ohkoshi, S. Kuroki, S. Sakurai, K. Matsumoto, K. Sato, S. Sasaki, Angew. Chem. Int. Ed. 2007, 46, 8392–8395. Y.C. Tseng, N.M. Souza-Neto, D. Haskel, M. Gich, C. Frontera, A. Roig, M. van Veenendaal, J. Nogues, Phys. Rev. B 2009, 79, 094404. S. Ohkoshi, S. Sakurai, J. Jin, K. Hashimoto, J. Appl. Phys. 2005, 97, 10K312. L. Machala, R. Zboril, A. Gedanken, J. Phys. Chem. B 2007, 111, 4003–4018. A.M. van Diepen, T.J.A. Popma, J. Phys. Colloq. 1976, 37, 755–758. O. Schneeweiss, R. Zboril, N. Pizurova, M. Mashlan, E. Petrovsky, J. Tucek, Nanotechnology 2006, 17, 607–616. H.J. Buser, D. Schwaryenbach, W. Petter, A. Ludi, Inorg. Chem. 1977, 16, 2704–2710. E. Kroll, F.M. Winnik, R.F. Ziolo, Chem. Mater. 1996, 8, 1594. E. Reguera, J. Fernandez-Bertran, Hyperfine Interact. 1994, 88, 49–58. E. Reguera, J. Fernandez-Bertran, J. Radioanal. Nucl. Chem. 1995, 200, 443–455. E. Reguera, J. Fernandez-Bertran, J. Balmaseda, Transit. Met. Chem. 1999, 24, 648–654. A. Ito, M. Suenaga, K. Ono, J. Chem. Phys. 1968, 48, 3597. R. Zboril, L. Machala, M. Mashlan, V.K. Sharma, Cryst. Growth Des. 2004, 4, 1317–1325. M. Hermanek, R. Zboril, I. Medrik, J. Pechousek, C. Gregor, J. Am. Chem. Soc. 2007, 129, 10929–10936. R.L. M€ ossbauer, Z. Phys. 1958, 151, 124–143. G.K. Wertheim, M€ossbauer Effect, Academic Press, New York, 1964. V.I. Goldanskii, R.H. Herber, Chemical Application of M€ossbauer Spectroscopy, Academic Press, New York, 1968. U. Gonser, ed., M€ossbauer Spectroscopy, Topics in Applied Physics, Springer, Berlin, 1975. A. Vertes, D.L. Nagy, eds., M€ossbauer Spectroscopy of Frozen Solutions, Akademiai Kiad o, Budapest, Hungary, 1990. J. Chappert, J. Phys. Colloq. 1974, 35, C6-71–C6-88. J. Chappert, J. Teillet, F. Varret, J. Magn. Magn. Mater. 1979, 11, 200–207. J.M.D. Coey, Can. J. Phys. 1987, 65, 1210–1232. Q.A. Pankhurst, R.J. Pollard, in M€ossbauer Spectroscopy Applied to Magnetism and Materials Science, G.J. Long, F. Grandjean, eds., Plenum Press, New York, 1993, pp. 77–113. J.M. Greneche, Acta Phys. Slovaca 1995, 45, 45–55. C.E. Johnson, J. Phys. D: Appl. Phys. 1996, 29, 2266–2273. M.F. Thomas, C.E. Johnson, in M€ossbauer Spectroscopy, D.P.E. Dickson, F.J. Berry, eds., Cambridge University Press, Cambridge, UK, 1986, pp. 143–197. I. Dezsi C. Fetzer, Electrochem. Commun. 2007, 9, 1846–1849. J.D. Cashion, T.L. Greaves, Hyperfine Interact. 2005, 165, 5–16. L. Aldon, C.M. Ionica, P. E. Lippens, D. Larcher, J.-M. Tarascon, J. Olivier-Fourcade, J.-C. Jumas, Hyperfine Interact. 2006, 167, 729– 732. C. Gejke, E. Nordstrom, L. Franson, K. Edstrom, L. Haggstrom, L. Borjesson, J. Mater. Chem. 2002, 12, 2965–2970. R.A. Dunlap, O. Mao, J.R. Dahn, Phys. Rev. B 1999, 59, 3494–3500. S. Ghosh, W. Wen, R.C. Urian, C. Heath, V. Srinivasamurthi, W.M. Reiff, S. Mukerjee, V. Naschitz, S. Lichtb, Electrochem. SolidState Lett. 2003, 6, A260–A264. I. Mitov, S. Asenov, T. Tomov, D. Klissurski, J. Phys. Chem. C 2007, 111, 5389–5393. K. Lazar, W.D. Rhodes, I. Borbath, M. Hegedus, J.L. Margitfalvi, Hyperfine Interact. 2002, 139, 87–96. X. Ge, J.Y. Shen, H.L. Zhang, Sci. China, Ser. B Chem. 1996, 39, 53–63. B. Benaichouba, P. Bussiere, J.C. Vedrine, Appl. Catal. A 1995, 130, 31–45. J.W. Niemantsverdriet, W.N. Delgass, Top. Catal. 1999, 8, 133–140. T.R. Motjope, H.T. Dlamini, G.R. Hearne, J.N. Coville, Top. Catal. 2002, 71, 335–341. B. Kniep, A. Constantinescu, D. Niemeier, K.D. Becker, Z. Anorg. Allg. Chem. 2003, 629, 1795–1804. J. Cieslak, S.M. Dubiel, B. Sepiol, Solid State Commun. 1999, 111, 613–618. S.K. Xia, F.C.R. Assuncao, R.R. Avillez, S. Doyle, T. Wroblewski, E. Baggio-Saitovitch, J. Alloys Compd. 1997, 248, 246–250.

392

€ 18 MOSSBAUER SPECTROSCOPY IN STUDY OF NANOCRYSTALLINE IRON OXIDES FROM THERMAL PROCESSES

110. M.V. Badani, L.M. Eshelman, W.N. Delgass, R. Prins, M.A. Vannice, D. Suresh, W.C. Conner, V. Ponec, D.G. Blackmond, N. Homs M.M. Bhasin, Stud. Surf. Sci. Catal. 1993, 75, 1223–1234. 111. B.R. Cuenya, A. Naitabdi, E. Schuster, R. Peters, M. Doi, W. Keune, Phys. Rev. B 2007, 76, 094403. 112. K. Nomura, S.S. Sharma, Y. Ujihira, Nucl. Instrum. Methods Phys. Res., Sect. B 1993, 76, 357–359. 113. E. Kuzmann, S. Nagy, A. Vertes, Pure Appl. Chem. 2003, 75, 801–858. 114. K.E. Ayers, N.C. White, J. Electrochem. Soc. 2005, 152, A467–A473. 115. L. Delaude, P. Laszlo, J. Org. Chem. 1996, 61, 6360–6370. 116. V.K. Sharma, Water Sci. Technol. 2004, 49, 69–74. 117. V.K. Sharma, ed., Ferrates: Synthesis, Properties, and Applications in Water and Wastewater Treatment, Vol. 985, ACS Symposium Series, American Chemical Society, 2008. 118. H. Goff, R.K. Murmann, J. Am. Chem. Soc. 1971, 93, 6058. 119. L. Machala, R. Zboril, V.K. Sharma, J. Filip, O. Schneeweiss, Z. Homonnay, J. Phys. Chem. B 2007, 111, 4280–4286. 120. T. Ichida, Bull. Chem. Soc. Jpn. 1973, 4, 79–82. 121. J. Madarasz, R. Zboril, Z. Homonnay, V.K. Sharma, G.J. Pokol, J. Solid State Chem. 2006, 179, 1426–1433. 122. L. Machala, R. Zboril, V.K. Sharma, J. Filip, D. Janc9ık, Z. Homonnay, Eur. J. Inorg. Chem. 2009, 8, 1060–1067. 123. J. Frydrych, M. Mashlan, J. Pechousek, D. Jancik, Mossbauer Spectroscopy in Materials Science 2008, AIP Conference Proceedings, Vol. 1070, 2008, pp. 170–184. 124. J.R. Gancedo, M. Gracia, J.F. Marco, Hyperfine Interact. 1991, 66, 83–93. 125. K. Nomura, M. Miglierini, D. Petridis, Mossbauer Spectroscopy in Materials Science, Vol. 66, Springer, 1999, pp. 63–78. 126. S. Mitra, Applied Mossbauer Spectroscopy, Pergamon Press, Oxford, UK, 1992, p. 14. 127. J. Frydrych, M. Mashlan, R. Zboril, J. Pechousek, M. Hermanek, Acta Metall. Slovaca 2007, 13, 298–301. 128. K. Sivula, R. Zboril, F. Le Formal, R. Robert, A. Weidenkaff, J. Tucek, J. Frydrych, M. Graetzel, J. Am. Chem. Soc. 2010, 132, 7436–7444. 129. I. Cesar, K. Sivula, A. Kay, R. Zboril, M. Gr€atzel, J. Phys. Chem. C 2009, 113, 772–782. 130. F.L. Souza, K.P. Lopes, P.A.P. Nascente, E.R. Leite, Sol. Energy Mater. Sol. Cells 2009, 93, 362–368. 131. A. Kleiman-Shwarsctein, Y.S. Hu, A.J. Forman, G.D. Stucky, E.W. McFarland, J. Phys. Chem. C 2008, 112, 15900–15907. 132. J.A. Glasscock, P. Barnes, I.C. Plumb, N. Savvides, J. Phys. Chem. C 2007, 111, 16477–16488.I. Cesar, Solar photoelectrolysis of water with translucent nano-structured hematite photoelectrodes, Ph.D. Thesis, Ecole Polytechnique Federale de Lausanne, 2007. 133. J. Frydrych, L. Machala, M. Hermanek, I. Medrik, M. Mashlan, J. Tucek, J. Pechousek, V.K. Sharma, Thin Solid Films 2010, 518, 5916–5919. 134. S. Deki, Y. Aoi, J. Okibe, H. Yanagimoto, A. Kajinami, M. Mizuhata, J. Mater. Chem. 1997, 7, 1769–1772. 135. M. Miglierini, A. Lancok, M. Pavlovic, in Isiame 2008, E. Kuzmann, K. Lazar, eds., Springer, 2009, pp. 45–52. 136. M. Miglierini, A. Lancok, M. Pavlovic, Phys. Met. Metall. 2010, 109, 469–474.

C H A P T E R 1 9

TRANSMISSION AND EMISSION 57 € Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS METH  HOMONNAY AND ZOLTAN  NE ZOLTAN

Faculty of Science, Eotvos Lorand University, Budapest, Hungary

19.1 INTRODUCTION Binary oxides with the stoichiometry ABO3 often have the perovskite structure. The versatility of these oxides is mostly due to the multivalent and multispin metal ions that may occupy site B of the unit cell (Fig. 19.1) and combined with some covalent character of the metal oxygen bonds, the electric and magnetic properties of these oxides become unique. A large family of these oxides shows colossal magnetoresistance (CMR). In the case of the classical perovskite-related cuprate structure of YBa2Cu3O7 d (YBCO), superconductivity sets in under about 90 K, and further combining the perovskite lattice with inserted atomic layers, the critical temperature could be pushed up to 160 K. M€ ossbauer spectroscopy of these systems is rather challenging. For high-Tc superconductors, the greatest challenge is typically the lack of M€ ossbauer active nuclides in those compounds (not counting the lately discovered oxypnictides). Thus, iron or cobalt doping is required for a M€ ossbauer study. The comparison of emission and transmission results can be very beneficial. CMR materials often contain Fe or Co, and thus conclusions on their electric and magnetic properties are more direct. ossbauer studies either alone or Beyond the high-Tc superconductors (where doping is a must), emission M€ simultaneously with transmission studies can serve with new information on the system investigated. A 57 Co emission study, at the first glance, is just a simple mirror image of an 57 Fe transmission study; only the roles of the source and absorber are exchanged. However, it would be true only if the material investigated contained iron, and the dopant material would be 57 Fe instead of its mother nuclide 57 Co. Thus, the following important differences may be taken into account when emission results are interpreted or compared with transmission studies: (i) During the doping process, 57 Co will prefer an environment that matches its chemical “needs” best, and not necessarily the one that is ideal for iron. (ii) In Co-containing materials, 57 Co will occupy Co sites in the lattice. (iii) In Fe-containing materials, 57 Co will not necessarily occupy Fe sites, although the chemical similarity between iron and cobalt may facilitate it. M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

393

394

€ 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS

Oxygen Titanium (Site B)

FIGURE 19.1

Calcium (Site A)

The unit cell of the classical perovskite CaTiO3. The octahedral coordination environment of Site B is indicated.

(iv) Although the primary site occupancy of 57 Co is determined by its chemical identity as cobalt, after the EC decay, it is already iron (commonly referred to as “nucleogenic iron”), and the M€ ossbauer signal comes from an 57 Fe atom residing at a lattice site optimal for Co. (v) The EC decay of 57 Co may result in aftereffects; however, changing the lattice site (for another one best suited for iron) is not expected. Only electronic changes may be found due to slow relaxation of electronic excitations or competing reactions between lattice defects and mobile charges. These can result in detecting, for example, aliovalent charge states of iron. (vi) In the systems discussed in this chapter, electron mobility is high; therefore, the appearance of aliovalent charge states or any other species due to slow relaxation is not expected. Another important technical difference between transmission M€ ossbauer spectroscopy (TMS) and emission M€ ossbauer spectroscopy (EMS) studies is that 57 Co doping usually requires a much lower dopant concentration than in an analogue TMS study. This is valid for materials with high average atomic number (heavy matrix), and can be well understood on the basis of the electronic absorption of the 14.4 keV gamma rays. The absorption (and therefore the loss of the M€ ossbauer signal) is given by the following formula: I 1 ¼ I0

expð ms d s Þ expð ma d a Þ; ms d s

(19.1)

where I0 is the total emitted intensity in the source, I is the transmitted intensity that reaches the detector, ms and ma are the mass absorption coefficients in the source and in the absorber, respectively, ds is the surface density of the source, and da is the surface density of the absorber. Since this formula is asymmetric with respect to the mass absorption coefficients, a simple exchange of the matrices (of the source and the absorber) will result in a big difference in I/I0 if the absorption coefficients differ significantly. In simple words, the gamma rays, on an average, travel through only one half of the total thickness of the source, but through the total thickness of the absorber. Thus, if the material investigated represents a heavy matrix for which the gamma absorption is higher, an emission experiment is more beneficial from this point of view. This is the case for high-Tc superconductors. Since, in this case, either Co or Fe is an impurity, low doping level is crucial to get reliable information on the properties of the host material. In this chapter, we will focus on the application of M€ ossbauer spectroscopy on perovskite-related systems where emission M€ ossbauer spectroscopy had a special contribution in exploring the structure and electronic or magnetic behavior of these materials. Some typical examples are selected from the literature.

19.2 STUDY OF HIGH-TC SUPERCONDUCTORS The discovery of high-Tc superconductivity by Bednorz and M€uller in 1986 [1] generated an almost unprecedented rush in science to solve the mechanism of this phenomenon. Although the ultimate answer is yet to be found, M€ ossbauer spectroscopy contributed a lot to the understanding of the structure and properties of these compounds. The variability of high-Tc materials is huge and this itself presented a great challenge that if structure is not crucial then what else? Thus, one may say that the perovskite-related structure is just one example, and may not be the most important one; it is a fact that the highest critical temperature has been reached so far in this family of compounds (160 K). Emission M€ ossbauer studies have been focused on YBa2Cu3O7 d, with critical temperature Tc 91 K. Its structure has become well known (Fig. 19.2).

395

19.2 STUDY OF HIGH-TC SUPERCONDUCTORS

Cu(1) Ba

O(1) Cu(2) O(3)

Y O(2)

O(4) O(5)-vacancy

FIGURE 19.2 The unit cell of YBCO. The coordination environment of the potential doping sites for

57

Co is indicated.

Since there are no conveniently applicable M€ ossbauer elements in this compound, doping was necessary for any M€ ossbauer study. The heavy matrix and the danger of influencing the electronic or structural properties of the host compound both pointed to the preference of the emission technique, that is, 57 Co doping instead of 57 Fe doping. The measurements proved that there were some phenomena that could not be found in transmission experiments, but only in emission studies. This must be assigned to either the much lower dopant concentration needed (about one hundredths of that used in 57 Fe doping experiments) or the different oxygen coordination preferences of Co and Fe, or both. Emission M€ ossbauer studies have been already reviewed [2] in detail; thus, here we will discuss only some major findings and further selected results published since then for YBCO. 19.2.1 Study of 57 Co-Doped YBa2Cu3O7

d

57

In YBCO, Co substitutes predominantly at the chain (Cu(1)) site. Despite this restriction, M€ ossbauer spectra appear to be very complicated, mainly due to variable local oxygen arrangements around Co (and therefore the nucleogenic Fe) atoms situated at Cu(1). Furthermore, it is not only one particular spectrum that is complicated, but thermal treatments changing the oxygen stoichiometry of the compound make it even more complex. The emission M€ ossbauer spectrum of the fully oxygenated YBa2Cu3O7 d (i.e., when the compound is heated in oxygen at 900–950  C and then at 450  C, resulting in d0.02–0.03) has been analyzed into four main components representing different oxygen configurations around 57 Co [3]. The three major species designated as A, B, and C differ significantly in their quadrupole splittings. A slight deoxygenation of the compound (a loss of 0.1 stoichiometric unit of oxygen) results in conversion of B and C into A, showing that all the three species are genetically related and are situated at the chain site where vacant O(5) sites may be occupied by oxygen. Species A is four-coordinate and has a square planar configuration. Species B and C are both five-coordinate, with B having a distorted trigonal bipyramidal configuration, while C has a square pyramidal configuration. Species M could be assigned to the Cu(2) site. A summary of the parameters observed is given in Table 19.1. The low isomer shifts are remarkable, and are consistent with the metallic character of these oxides. Metallicity has also been demonstrated by deoxygenation of the compound, when the isomer shift of species A (the only stable € ssbauer Parameters Obtained for 57 Co-Doped YBCO in its Fully Oxygenated State [3] TABLE 19.1 Mo Doublet

d

A B C M

0.03 0.05 0.09 0.18

D

Relative Area

Species

2.01 1.00 1.55 0.61

0.51 0.21 0.19 0.09

Cu(1), square planar Cu(1), trigonal bipyramidal Cu(1), square pyramidal Cu(2)

396

€ 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS

FIGURE 19.3 Interconversion of species B and C in 57 Co-doped YBCO [3].

component under practically any condition) monotonously increased down to about an oxygen stoichiometry of 6.4 and then leveled off. It is known from independent measurements that YBCO is metallic above 6.4 and insulating below that. This finding was quite unique for the characterization of YBCO, and as far as the two extreme O stoichiometries are ossbauer studies. (Systematic oxygen concerned (O6 and O7), it was in perfect agreement with transmission 57 Fe M€ removal experiment has not been done with Fe-doped YBCO.) Nevertheless, the most interesting phenomenon found for 57 Co-doped YBCO and never observed for Fe-doped YBCO is connected with the temperature variation of the M€ ossbauer spectrum of fully oxygenated YBa2 Cuð57 CoÞ3 O7 d . As shown in Fig. 19.3, on raising the temperature of the material from 80 to 350 K, an interconversion of species B to species C takes place. The extent of the interconversion is visualized in Fig. 19.4. (Here species B0 is an impurity phase-related component that shows up when the sample is subjected to several oxygenation–deoxygenation cycles.) The phenomenon has been simply referred to as “B to C conversion.” The explanation of the interconversion is based on the fact that the structure of the Cu(1) O(4) chains in the YBCO lattice was found to be of zigzag type [4]; that is, the O(4) atoms are off-placed from the line of Cu(4) atoms by about 0.01 nm in an alternating manner. The amplitude of vibration of O(4) out of the chain was observed to be very

FIGURE 19.4 B to C conversion in fully oxygenated 57 Co-doped YBCO.

397

19.2 STUDY OF HIGH-TC SUPERCONDUCTORS

FIGURE 19.5 Structure of (nucleogenic) Fe–O species observed in fully oxygenated 57 Co-doped YBCO.

large, and since in the classical BCS theory of superconductors lattice vibrations are crucial for the formation of Cooper pairs as superconducting charge carriers, dynamics of the Cu(1) O(4) chains is of particular interest. This is valid even if superconductivity is supposedly assigned to the Cu(2) sheets and not to the Cu(1) chains. Any change (lattice distortion, valence fluctuation) in the chains may affect the Cu(2) sheets and may indirectly contribute to superconductivity. It was proposed that species B is a five-coordinate distorted trigonal bipyramidal, and C is a distorted square pyramidal species. Both of them are either high- or intermediate-spin 57 Fe4þ , although valence assignment in these systems with significant covalent character is rather dubious. The two species are separated by a stereochemical barrier, and there is a dynamic equilibrium between them that shifts toward C at higher temperatures. Such stereochemical nonrigidity in five-coordinate systems has been well known [5]. The structures of the three major species indicating the dynamic off-placement of the O(4) atoms are depicted in Fig. 19.5. It is assumed that when the normal zigzag structure of the Cu(1) O(4) chain is retained, the system is rather dynamic, that is, the chain vibrations are not hampered. However, off- placement of neighboring O(4) atoms to the same side of the chain must represent a more rigid structure and considerable damping of the chain vibration. One should add here that such species are obviously not present in the superconducting host material; however, the ossbauer probe, not the nucleogenic 57 Fe alone. whole species (i.e., “[FeO5]” or “[FeO4]”) may be considered as a M€ With this approach, drawing conclusions on the properties of high-Tc superconductors may be easier. 19.2.2 Study of 57 Co-Doped Y1 xPrxBa2Cu3O7

d

It was well known shortly after the discovery of YBCO as a high-Tc superconductor that almost any lanthanoid may be substituted for yttrium in its lattice and superconductivity is hardly affected; even Tc has a variation of only a few kelvins. A remarkable exception is praseodymium that has a deleterious effect on superconductivity and PrBa2Cu3O7 d is a simple insulator. The cause of this deleterious effect has been searched for extensively. Since it is commonly accepted that superconducting Cooper pairs form in the Cu(2) sheets in the lattice from holes of the Cu3dO2p hybrid band, and the Cu(2) sheet and the rare earth sheet are neighbors, overlap or even charge transfer may take place between them depending on the size of the rare earth cation. Charge transfer is plausible because unlike the other rare earths, Pr is ready to form Pr4þ ions (not only Pr3þ), thus pumping electrons to Cu(2) and neutralizing electron holes. This mechanism is referred to as hole filling. Let us note that 151 Pr M€ ossbauer measurements were in agreement with this theory [6]. Not reaching so far, and instead of assuming Pr to be totally ionic, allowing some covalency, it may be enough to assume that since Pr3þ is the largest cation among rare earth trivalent cations, the Pr4f orbitals may overlap with O2p

398

€ 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS

orbitals and the result is very close to that of hole filling [7]. The overlap may just alter the band structure so that electron holes become localized and thus they will not be able to form Cooper pairs [8]. Even weaker, yet very effective interaction is when the magnetic moment of the Pr3þ ion couples with Cooper pairs and breaks them apart (pair breaking). Unfortunately, all these theories deal with the Cu(2) sheets, and M€ ossbauer spectroscopy can serve with useful information only if there is some coupling between the chains and the sheets of the PrBCO/YBCO lattice. As we could see in the case of YBCO, the B to C conversion was an indication of the dynamics of the Cu(1) O(4) chains, and if the chain vibrations are intensive enough, this itself can be responsible for coupling, that is, for a chain–sheet interaction. Now it is interesting to see if there is any correlation between the observed B to C conversion and the gradual disappearance of superconductivity as the Pr content of YBCO increases. 57 Co emission studies were carried out on Y1 x Prx Ba2 Cu3 ð57 CoÞO7 d for x ¼ 0.10, 0.30, 0.45, and 1 [9,10]. The spectrum series for x ¼ 0.1 is shown in Fig. 19.6. The presence of B to C conversion is obvious. However, the extent of the transformation is reduced and the parameters of the various species are also altered a little bit. If we have a look at the whole series, a systematic change appears. The conversions are shown in Figs. 19.7 and 19.8. The pure PrBa2 Cu3 ð57 CoÞO7 d material did not show any interconversion between room temperature and 80 K; not even species B could be found. Table 19.2 shows the parameters of the conversion including enthalpy and entropy changes calculated from the temperature dependence of the equilibrium constants (relative spectral area were considered proportional to chemical concentrations). It is clear that B to C conversion gradually disappears and it seems to become complete at the Pr content critical to superconductivity (x ¼ 0.55) but, interestingly, not only the extent of conversion decreases but also the conversion enthalpy and entropy. If that is so, the species taking part in the conversion should become more and more alike as the Pr content increases, otherwise the energy difference between them cannot be understood.

FIGURE 19.6 Temperature dependence of the € ssbauer spectra of emission Mo Y0:9 Pr0:1 Ba2 Cu3 ð57 CoÞO7 d . (Reproduced from Ref. 10 with permission of Elsevier.)

399

19.2 STUDY OF HIGH-TC SUPERCONDUCTORS

FIGURE 19.7 B to C conversion in Y0:9 Pr0:1 Ba2 Cu3 ð57 CoÞO7 d . (Reproduced from Ref. 10 with permission of Elsevier.) 45 45

A Relative area (%)

Relative area (%)

35

B

30 25 20

C

15 10

I

35 30

I

25 20

C

15

5 50

B

40

40

100 150 200 250 300 350 400 450

Temperature (K)

10 100

A 150

200

250

300

350

400

Temperature (K)

FIGURE 19.8 B to C conversion in Y0:9 Pr0:3 Ba2 Cu3 ð57 CoÞO7 d (left) and Y0:55 Pr0:45 Ba2 Cu3 ð57 CoÞO7 Ref. 9 with permission of the American Physical Society.)

d

(right). (Reproduced from

Thus, the relevant M€ ossbauer parameters were collected and they are shown in Table 19.3. One can see that the isomer shifts of species B and C get closer to each other as the Pr content increases. Even the quadrupole splitting of B increases, although it does not reach that of C. Thus, the decreasing thermodynamical parameters of the conversion are consistent with the presumably changing spatial and electronic structure of the relevant species. TABLE 19.2 Thermodynamical Parameters of the B to C Conversion as a Function of the Pr Content of Y1 Pr Content (x) 0 0.10 0.30 0.45 1.00

x

Prx Ba2 Cu3 ð57 CoÞO7

d

DA(B)/A(B þ C)

DH (kJ mol 1)

DS (J molK 1)

0.63 0.56 0.32 0.17 –

31(4) 12(1) 7.3(5) 2.6(8) –

111(18) 50(8) 29(6) 6(3)

DA denotes the spectral area of the species indicated in parentheses, DA(B)/A(B þ C) measures the conversion ratio, DH is the conversion enthalpy, and DS is the conversion entropy.

400

€ STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER

TABLE 19.3 Pr Content (x) 0 0.10 0.30 0.45 1.00

€ ssbauer Parameters of the Observed Species in Y1 Mo [9,10] as a Function of the Pr Content d (A)

D (A)

0.03 0.02 0.03 0.01 0.08

2.00 1.97 1.95 1.98 1.86

d (B) 0.05 0.01 0.02 0.10 –

x

Prx Ba2 Cu3 ð57 CoÞO7

D (B)

d (C)

D (C)

0.98 1.00 1.02 1.06 –

0.09 0.10 0.10 0.12 0.11

1.54 1.46 1.48 1.56 1.56

d

at Room Temperature

d (M or I) 0.18 0.18 0.20 0.17 0.07

D (M or I) 0.59 0.35 0.32 0.47 0.91

Experimental errors are typically 0.02 mm s 1; however, they are higher for species I and M.

It is remarkable that species M could not be found in any Pr-containing sample, but species I was observed. Since there is no reason to believe that it is not a Cu(2)-related species, the sudden change in its M€ ossbauer parameters supports the assumed interaction between the Pr sheets and the Cu(2) sheets. Now the question arises as to how the substitution of Pr can affect the B to C conversion. The disappearance may be explained by either diminishing displacement of the O(4) atoms, or a transformation of the dynamic zigzag structure into a static one (i.e., when the chain becomes more rigid). Since the temperature of the B to C conversion did not change appreciably with the Pr content, the latter is unlikely. Thus, the lowering of the displacement of the chain oxygens should be explained. From this point of view, the structural data provided by neutron diffraction [11] and X-ray studies [12,13] are relevant, and show that the Cu(1) Cu(1) distance along the chain (i.e., lattice parameter b) increases with Pr substitution at the Y site. In YBa2Cu3O7 d, the exact position of the O(4) atoms is 0.015 nm off the chain, according to a high-resolution neutron diffraction study [4]. One can assume that this off-placement occurs because the Cu(1) Cu(1) distance is too short, that is, the optimal Cu(1) O(4) bond length is larger than one half of the Cu(1) Cu(1) distance. The consequence is that the O(4) atoms are forced into an off-chain position. The optimal Cu(1) O(4) bond length may be estimated from the displacement (0.015 nm) and from the Cu(1) Cu(1) distance. Using the data from Ref. 11, b ¼ 0.3883 nm (at 10 K), and thus an optimal bond length of 0.1947 nm is obtained. This means that if the Cu(1) Cu(1) distance were 2  0.1947 ¼ 0.3894 nm, the off-placement of the O(4) atoms would not be inevitable and the Cu O chains could acquire a straight configuration. In this case, the out-of-the-chain O(4) vibrations must be damped considerably. Furthermore, as a consequence of the M€ ossbauer results, in a straight chain, there is much less driving force for the formation of species B. Indeed, in PrBa2Cu3O7 d, b reaches 0.3918 nm [11], which is higher than twice the optimal Cu(1) O(4) bond length, and species B is absent (and of course, there is no B to C conversion) [14]. So, it could be concluded that the disappearance of the B to C interconversion is a sign of the straightening of the Cu(1) O(4) chain, and it may indicate that the out-of-the-chain vibration of the O(4) atoms can be important for the mechanism of high-Tc superconductivity in the 1–2–3 compounds. The idea is illustrated in Fig. 19.9. In addition to the proposed correlation between the B to C conversion and superconductivity, the analysis of the variation of the M€ ossbauer parameters was also beneficial in supporting the model of straightening of the chains. The isomer shifts of species A and C were found to be almost constant up to x ¼ 0.45; only a very slight decrease could be seen. Their quadrupole splittings were practically constant. These findings were consistent with the model, since the structure of these species should not depend on the offplacement of the O(4) atoms in the chain. The situation for species B is completely different, because the off-placement of the O(4) atoms determines the distortion of the trigonal bipyramid (Fig. 19.9). As the chain is straightened, the distortion increases (from the point of view of an ideal trigonal bipyramidal structure) and finally reaches the structure of C. This implies that during this process, the M€ ossbauer parameters of B should approach those of C, and that is exactly what happens (Table 19.3). This model allowed more discussion of the observed isomer shift values for species A, B, and C in correlation with the electronic structure of the chains. The first remarkable finding to be explained was the substantially smaller isomer shift for C than that for B. Since the coordination numbers for the two species are the same, and the valence state of the central iron atom must also be the same, the ligand sphere structure should be responsible for the difference. It was assumed that due to its square pyramidal structure, species C is an integral part of the metallic chain. Thus, the electronwithdrawing effect of conduction band of the chain results in low electron density on the 3d shell of Fe. In contrast, species B represents a severe perturbation of the chain, and its electron system is probably rather isolated. It means more localized d electrons at the central Fe atom, and consequently, a higher isomer shift. Nevertheless, it was pointed out that the isomer shift of species A is rather surprising, being much higher than that of species C, despite the lower coordination

401

19.3 STUDY OF STRONTIUM FERRATE AND ITS SUBSTITUTED ANALOGUES

x < x krit.

Cu(1) Ba

Cu(2)

Y vagy Pr

O(5) O(4) Species C

Species B

FIGURE 19.9

O(1)

Killing of the B to C conversion as a consequence of straightening of the Cu(1) O(4) chain by the effect of praseodymium in Y1 x Prx Ba2 Cu3 O7 d . The coordination sphere of only one selected Cu(1) is shown. The critical Pr concentration xcrit is 0.55. (Reproduced from Ref. 9 with permission of the American Physical Society.)

x > x krit.

Species C

Species C

number. It was assumed that the four-coordinate species A has a distorted square planar oxygen arrangement, presumably closer to a tetrahedral one, and it results in a partial structural isolation, that is, more localized d electrons for iron. The localization of the d electrons overcompensates the effect of the lower coordination number (as compared to C) and the higher isomer shift is explained. On the basis of the variation of M€ ossbauer isomer shifts of the major species A, B, and C with Pr content, it was concluded that only the isomer shift of species B varied significantly, but it is an intrinsic property of the varying structure of this species. Trying to find correlation with existing models of high-Tc superconductivity, it was concluded that if hole filling by Pr were to occur in the Cu(2) sheets, and if a fraction of that charge were to transfer onto the Cu(1)–O chains, then the delectron density would increase in the chains resulting in an increase of the isomer shifts of species A and C. It was not observed. These findings were consistent with models that restrict the effect of Pr to the superconducting sheets. However, the most striking result was the disappearance of the B to C conversion that was observed only in emission M€ ossbauer studies, and this lent more credence to models assuming an active role of the chains in the mechanism of highTc superconductivity. The breakthrough for explaining the real mechanism of high-Tc superconductivity and a final judgment on the contribution of emission M€ ossbauer spectroscopy to this challenging research field is yet to come.

19.3 STUDY OF STRONTIUM FERRATE AND ITS SUBSTITUTED ANALOGUES Strontium ferrate (SrFeO3) has the classic perovskite structure just like lanthanum cobaltate (LaCoO3). In SrFeO3, iron is tetravalent and strontium is bivalent. In LaCoO3, both lanthanum and cobalt are trivalent. Although these pristine compounds themselves are very interesting, some combinations of the two acquire even more exciting properties. Starting from SrFeO3, substituting Ba or Ca for Sr, and Co for Fe, the resultant material absorbs CO2 heavily at high temperatures and this is why these compounds have become subjects of several studies. It was found that absorption properties can be optimized by applying the appropriate ratio of these elements at the respective lattice sites. It was found that while SrFeO3 does not absorb CO2 readily, partial substitution of Fe by Co, and partial substitution of Sr by Ca results in significant increase in the CO2 absorption rate and capacity [15]. 19.3.1 Study of Sr0.95Ca0.05Co0.5Fe0.5O3

d

and Sr0.5Ca0.5Co0.5Fe0.5O3

d

The iron content of these materials gives an opportunity for the direct use of M€ ossbauer spectroscopy (TMS technique) as a tool to learn more about the CO2 absorption mechanism in these systems. It was reported on the basis of TMS

402

€ 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS

measurements [16] that the CO2 absorption capacities of Sr0.95Ca0.05Co0.5Fe0.5O3 d and Sr0.5Ca0.5Co0.5Fe0.5O3 d are markedly different, the latter with higher Ca content absorbed twice as much CO2 than the former. Not only the capacity but also the absorption rate showed similar differences. These differences could be understood partially on the basis of the M€ ossbauer results, showing different oxygen coordination environments of Fe in the two materials. However, since both iron and cobalt have variable valence states, it is important to know if they have the same or different chemical states in these compounds. EMS offers a unique way to probe the Co sites in these Co-containing perovskites. The 1:1 ratio of the two elements facilitates the comparison of the transmission and emission M€ ossbauer spectroscopy results. EMS study of Sr0.95Ca0.05Co0.5Fe0.5O3 d and Sr0.5Ca0.5Co0.5Fe0.5O3 d doped with 57 Co in pristine form and after different CO2 treatments is reported in [17]. 19.3.1.1 Sr0.95Ca0.05Co0.5Fe0.5O3 d Sr0.95Ca0.05Fe0.5Co0.5O3 d was prepared by the sol–gel technique [18] using citric acid, with a final sintering temperature of 1100  C. Since the oxygen stoichiometry reached by the end of the synthesis is critical, it was checked by cerimetry and found to be Sr0.95Ca0.05Fe0.5Co0.5O2.67  0.02. This indicates that the material should contain oxygen vacancies in the lattice. If those vacancies are randomly distributed or preferentially bound to either iron or cobalt, they may determine several properties including the CO2 absorption capability. This is exactly where a comparison of EMS and TMS studies can serve with useful information. X-ray diffractometry confirmed the cubic perovskite structure with (a ¼ 0.3862 nm). The preservation of the cubic structure despite the multiple substitutions at both site A and B may indicate that the distribution of the substituted elements is random, or, at least, significant long-range ordering does not occur. Still, the nearest ligand environment of Fe and Co can be different. Room-temperature TMS and EMS spectra can be seen in Fig. 19.10. Evaluation data are collected in Table 19.4. The difference between the two spectra is marked, and is obviously not just due to line broadening usually observed in EMS spectra. A three-doublet model was applied to fit both spectra using common linewidth within one evaluation, knowing that this can introduce some uncertainty in the evaluated parameters. This model resulted in the least difference between the TMS and EMS spectral data. For doublets D1 and D2, the emission and transmission parameters are fairly close to each other, while for the third doublet, they are markedly different (D3 and D30 ). This difference can be reasonably assigned to differing oxygen ligand environments for iron and cobalt in the lattice. Taking into account that the compound contains Co and Fe in equal concentrations, there are four border cases that may be considered for the distribution: (i) Fe and Co have a perfectly random distribution in the lattice (i.e., at site B). (ii) Fe and Co atoms have two-dimensional separation, that is, they are ordered in alternating Co and Fe layers. (iii) Fe and Co have a three-dimensional ordered structure with alternating occupancy of the B sites in all three directions of the lattice. (iv) The Fe and Co atoms are found in microscopic but separate pockets in the lattice (phase separation).

3500000 3480000 3460000

EMS

3440000

Counts

3420000 –2

0

2

0

2

3160000 3120000 3080000

FIGURE 19.10

3040000

EMS and TMS spectra of Sr0:95 Ca0:05 Fe0:5 Co0:5 O3 d recorded at room temperature.

3000000

TMS

–2

Velocity (mm s–1)

403

19.3 STUDY OF STRONTIUM FERRATE AND ITS SUBSTITUTED ANALOGUES

TABLE 19.4 Spectrum Evaluation Results of Sr0.95Ca0.05Fe0.5Co0.5O3 M€ ossbauer Technique

Subspectrum

TMS

D1 D2 D3 D1 D2 D30

EMS

d (mm s 1) 0.22 0.06 0.12 0.25 0.02 0.11

d

D (mm s 1) 0.66 0.65 0.16 0.56 0.46 1.49

in TMS and EMS Modes G (mm s 1)

Relative Abundance (%)

0.32

0.45

28.0 33.0 39.0 37.7 42.7 19.7

All these cases should be considered keeping in mind that in EMS measurements, one uses a nucleogenic 57 Fe probe with the coordination environment of the mother nuclide 57 Co preserved. Aftereffects are not considered due to high electron mobility in these perovskites. The nearest neighbor (NN) metal environment of the normal (TMS) and nucleogenic 57 Fe (EMS) is illustrated in Fig. 19.11. Following the reasoning in Ref. 17, it can be seen that in case of randomness, (i) the average number of NN Fe and Co ions is three for both normal and nucleogenic 57 Fe. Note that any distinction between “Fe environment” and “Co environment” is meaningless in this case, and TMS and EMS spectra may be identical. If Co and Fe atoms form alternating layers in the lattice (ii), the M€ ossbauer spectra should be different. As shown in Fig. 19.11, in the TMS case, the 57 Fe probe has four equatorial Fe neighbors and two axial Co neighbors. In the EMS case, the situation is reversed: there are four equatorial Co and two axial Fe neighbors. In case (iii), the 3D alternating occupancy at site B results in all-Co-neighbors in the TMS and all-Fe-neighbors in the EMS case. The situation is reversed in case (iv), when separation of Co-rich and Fe-rich regions is assumed. TMS and EMS spectra are expected to be different in both cases. Taking into account only the metal nearest neighbors is the first approximation because the M€ ossbauer spectra are primarily determined by the oxygen neighbors of the 57 Fe probe. Now it follows that, since it is assumed that the oxygen coordination of 57 Co is preserved for the nucleogenic 57 Fe, one has to take into account the different degrees of oxygen occupancy at lattice sites between two NN Fe atoms, between two NN Co atoms, as well as between NN Co and Fe atoms. On the basis of this, some difference between TMS and EMS spectra can also be expected in case (i) because the original 57 Co environment contains three Co–Co and three Co–Fe NNs while the 57 Fe environment contains three Co–Fe and three Fe–Fe NNs. Case (ii) is not very much affected by this new consideration, and there will be a sharp contrast between TMS and EMS spectra. In the 3D alternating structure (iii) there are only Fe–Co nearest neighbors in the whole lattice, so the TMS and EMS spectra may be identical. In this case, however, the fact of having Fe or Co NN (the first approximation) will certainly introduce some disparity. Finally, in the case of phase separation, (iv) the difference between TMS and EMS spectra is trivial, just as in our first approximation.

TMS

TMS

Random

3D alternating EMS

EC

EMS

EC

TMS

TMS

Layered

Phase separated EMS

EC

EMS

EC

FIGURE 19.11 Possible scenarios for the nearest neighbor cation distribution in Sr0:95 Ca0:05 Fe0:5 Co0:5 O3 d in € ssbauer experitransmission Mo ments using 57 Fe as a microprobe € ssbauer as well as in emission Mo measurements where nucleogenic 57 Fe (forming by electron capture (EC) decay from 57 Co) serves as microprobe. Oxygens of the lattice as well as alkaline earth ions are omitted for better clarity. (Full circles: Fe; empty circles: Co.).

404

€ STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER

On the basis of these considerations, three fairly reasonable assumptions could be settled: (1) The M€ ossbauer isomer shifts and quadrupole splittings are primarily determined by the number and spatial arrangement of coordinated oxygens; the identity of the nearest neighbor metal ion (Fe or Co) is of less importance. (2) The magnetic interaction as sensed by the Fe probe depends on the coordinated oxygens AND the nearest neighbor ions. (3) In EMS experiments, the relevant oxygen coordination environment of the Fe probe is that of the mother 57 Co nuclide, but for evaluating the magnetic interaction, the daughter 57 Fe matters. Whichever approximation is applied, in cases (i), (iii), and (iv) the symmetry of the environment of the 57 Fe nucleus is basically the same in the TMS and EMS experiments. The major difference is found in the relative number of NN Fe and Co and in the occupancy of the six connecting O sites. These differences should be reflected primarily in the isomer shift of the central 57 Fe probe. It is only case (ii) where the environment of the normal and nucleogenic 57 Fe significantly differs in symmetry (opposite tetragonal distortion), and this should be seen in the quadrupole splitting of the corresponding species. (The authors of Ref. 17 note here that one may consider a fifth case of Fe–Co distribution in the lattice, when Co and Fe atoms are arranged in alternating chains, but that case is symmetrically analogous to case (ii) only the EMS and TMS cases are exchanged.) Based on this argumentation, it was concluded that most probably, the Co and Fe atoms form alternating layers in the Sr0.95Ca0.05Fe0.5Co0.5O3 d perovskite lattice, and this is the reason of the markedly different quadrupole splitting of species D3 and D30 in the TMS and EMS spectra, respectively. It is noteworthy that SrFeO2.5 has the brownmillerite structure with alternating layers of octahedrally and tetrahedrally coordinated Fe3þ sites. When half of the Fe is replaced by Co that tends to acquire lower valence state and therefore lower coordination number than Fe, it is at least not very surprising to observe a layered structure. It is remarkable that the difference in the quadrupole splittings (between EMS and TMS experiments) is not observed in states D1 and D2. A possible explanation is that these states form as a consequence of charge disproportionation similar to those identified in Sr–ferrate analogues by others [19]. This process explains noninteger valence states in these systems and presumes an extended covalency of the lattice involving itinerant electrons, that is, a long-range interaction in the Fe sublattice. Such an interaction can reduce local asymmetries in the electron density distribution around the Fe nuclei. It appears that species D3 (D30 ) is sensitive to the local charge distribution because it corresponds to a local state, while species D1 and D2 are insensitive to it due to significant valence electron delocalization. The authors point out that observation of long-range magnetic order in these compounds can also be indicative of the local oxygen and metal environment of the M€ ossbauer probe atom. There was no sign of magnetism at room temperature. At 80 K, the TMS spectrum still did not show any magnetism, while the formation of a magnetic structure was under way in the EMS case [18]. Based on the observed tendency of magnetic ordering upon increasing number of oxygen vacancies in related compounds, this observation is practically in agreement with all models shown in Fig. 19.11. It is added that although it would be very difficult to make a quantitative statement about the tendency of magnetic ordering based purely on the relative number of Co–O–Fe and Fe–O–Fe exchanges, one may expect a more significant difference between the EMS and TMS experiments in the 3D alternating (iii) and the phase-separated (iv) models, and no difference at all in the random (i) model. Thus, this observation would also support the existence of Co and Fe layers in the lattice. It is noted, however, that the different oxygen occupancies complicate the picture, and it has to be taken into account. The authors also note that in case of a layered structure, one would anticipate lower than cubic symmetry for the lattice. It is possible that the lattice is composed of a mixture of tetragonal or orthorhombic microdomains, while macroscopically, the system appears to be cubic for XRD. This behavior is not unknown for perovskites. It has been shown by theoretical calculations for Fe-doped YBa2Cu3O7 d that a macroscopically tetragonal structure (as seen by XRD) can originate from a randomly oriented ensemble of orthorhombic microdomains [20]. These experiments clearly revealed that cobalt and iron are not randomly distributed in the perovskite lattice of Sr0.95Ca0.05Fe0.5Co0.5O3 d and their oxygen environments are different. Carbon dioxide absorption experiments confirmed this conclusion. About 0.43% m/m CO2 absorption resulted in the M€ ossbauer spectra shown in Fig. 19.12 [18]. Regarding the TMS spectrum, the appearance of magnetic order is obviously due to the absorption of CO2. The most likely mechanism of CO2 uptake at this level is the migration of stick-shaped CO2 molecules in channels of ordered O defects [15]. CO2 seems to play a similar role as the absence of oxygen: it blocks fast electron transfer between metal ions that inhibits magnetic fluctuations and helps form static long-range order of spins.

19.3 STUDY OF STRONTIUM FERRATE AND ITS SUBSTITUTED ANALOGUES

405

5680000 5660000 5640000

EMS

Counts

5620000 –10

0

10

32200000 32000000

TMS 31800000 –10

0

10

Velocity (mm s–1)

FIGURE 19.12 € ssbauer spectrum of Mo Sr0:95 Ca0:05 Fe0:5 Co0:5 O3 d recorded at room temperature in transmission (bottom) and emission (top) mode after absorption of CO2 at 650  C (0.43% m/m CO2 uptake).

The EMS spectrum was found to be basically different from the TMS one. The lines are broad (>1 mm s 1) but the spectrum could be evaluated by static approximation, that is, with Lorentzian doublets and sextets instead of a relaxational line shape. A very formal fit including one sextet and one doublet resulted in d ¼ 0.22 mm s 1 and D ¼ 1.25 mm s 1 for the doublet and d ¼ 0.29 mm s 1 and B ¼ 41.3 T with quadrupole shift e ¼ 0.15 mm s 1 for the sextet. CO2 absorption at higher temperature (950  C) where the uptake was 10 times as much, the spectra showed even more difference (Fig. 19.13) . The TMS spectrum regained the original shape it had before any CO2 absorption. This means that the ligand environment is the same as in the original compound. Now, since CO2 should also be there in the lattice, it has to be in the vicinity of cobalt. This situation can only be explained if some major phase separation takes place in the lattice. Fe should have only oxygen environment while Co must have CO2 neighbors also. SEM and X-ray microanalysis measurements verified the assumed phase separation. The X-ray maps of CoKa and FeKa radiation recorded by a wavelength dispersive spectrometer for the sample treated at 950  C revealed inhomogeneity in the Fe/Co ratio. ossbauer 19.3.1.2 Sr0.5Ca0.5Co0.5Fe0.5O3 d Sr0.5Ca0.5Co0.5Fe0.5O3 d has an orthorhombic structure and its M€ spectrum (either in EMS or in TMS mode) contains mostly magnetic sextets (Figs. 19.14 and 19.15) [17]. The chemical analysis showed an oxygen stoichiometry of 2.5 and this explains the presence of the two major sextets well known from the M€ ossbauer spectrum of the brownmillerite-structured antiferromagnetic CaFeO2.5. The sextet with higher hyperfine field and higher isomer shift corresponds to octahedrally coordinated sites, and the other one represents tetrahedrally coordinated sites (Table 19.5) . These sites form alternating layers in the lattice. The sextets in the spectrum of CaFeO2.5 have the same weight out of stoichiometric reason. In the EMS spectrum of Sr0.5Ca0.5Co0.5Fe0.5O2.5, however, the area ratio of the sextets is about 1 to 1.5 in favor of the tetrahedrally coordinated

2120000 2100000

EMS

2080000

Counts

2060000 -2

0

2

3450000 3420000 3390000

TMS

3360000 3330000 -2

0

Velocity (mm s–1)

2

FIGURE 19.13 € ssbauer spectrum of Mo Sr0:95 Ca0:05 Fe0:5 Co0:5 O3 d recorded at room temperature in transmission (bottom) and emission (top) mode after absorption of CO2 at 950  C (4.37% m/m CO2 uptake) [18].

406

€ STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER

FIGURE 19.14 € ssbauer spectra of Mo Sr0.5Ca0.5Fe0.5Co0.5O2.5 recorded in EMS mode at room temperature after synthesis (a), followed by thermal treatment in CO2 at 650  C (b), and after CO2 treatment at 825  C (c).

Counts

7.78x10 6 7.76x10 6 7.74x10 6 7.72x10 6 7.70x10 6

(a)

1.04x10 7 1.04x10 7 1.04x10 7 1.03x10 7

(b)

7.23x10 6 7.22x10 6 7.20x10 6 7.19x10 6

(c)

–20

–10

0

10

20

Velocity (mm s–1)

FIGURE 19.15 € ssbauer spectra of Mo Sr0.5Ca0.5Fe0.5Co0.5O2.5 recorded in TMS mode at room temperature after synthesis (a), followed by thermal treatment in CO2 at 650  C (b), and after CO2 treatment at 825  C (c).

Counts

3.40x10 6 3.30x10 6 3.20x10 6 3.10x10 6

(a)

2.19x10 6 2.16x10 6

(b)

2.13x10 6

1.78x10 7 1.77x10 7

(c)

1.76x10 7

–10

–5

0 5 Velocity (mm s–1)

10

nucleogenic 57 Fe, which means that the Co atoms have a strong preference for the lower coordination number. This preference is reconfirmed by the transmission M€ ossbauer spectrum of the compound that showed (somewhat less) preference for the octahedral lattice sites. This observation agrees well with chemical considerations on the valence state and coordination number preference of Fe and Co. The TMS and EMS spectra of Sr0.5Ca0.5Co0.5Fe0.5O2.5 also contained doublets with isomer shifts near to those of the sextets. They were assigned [17] to octahedral and tetrahedral species, which showed up due to microheterogeneity of the lattice. Microheterogeneity for these species meant either higher local oxygen content or a disordered defect structure that supposedly resulted in the lack of layered arrangements of oxygen vacancies necessary for the

€ ssbauer Parameters of the Two Major Sextets Measured in TMS and EMS Modes for TABLE 19.5 Mo Sr0.5Ca0.5Co0.5Fe0.5O2.5 Having the Structure of Brownmillerite

TMS EMS

Octahedral site Tetrahedral site Octahedral site Tetrahedral site

d (mm s 1)

2e (mm s 1)

B (T)

0.34 0.20 0.37 0.20

0.61 þ0.61 0.65 þ0.56

48.5 41.5 47.4 40.1

19.4 PURSUING COLOSSAL MAGNETORESISTANCE IN DOPED LANTHANUM COBALTATES

407

antiferromagnetic coupling. The assumption of the (thermally) disordered structure was supported by low-temperature measurement showing lower relative intensity of the paramagnetic doublets. ossbauer spectra, CO2 treatment of the compound showed very different effects in the transmission and emission M€ and actually confirmed the nonrandom distribution of Co and Fe in the lattice. Interestingly, despite the significant CO2 absorption, the emission spectra of Sr0.5Ca0.5Co0.5Fe0.5O2.5 showed hardly any change after the treatments at either 650 or 825  C where the CO2 uptake was 3 and 17%, respectively (Fig. 19.14). The transmission M€ ossbauer spectra revealed a completely different behavior. The effect of the CO2 treatment was clear. The ratio of the original sextets remained almost the same, but a new sextet appeared (Fig. 19.15). It was explained by the following reasoning. While ready absorption of CO2 in these perovskites can be due to the ordered oxygen vacancy arrangement, which is facilitated by the presence of Co, CO2 has a strong preference to react in a way that an almost pure Fe compound forms, and Co is retained in the brownmillerite structure: 2ðSr; CaÞ2 FeCoO5 þ CO2 ¼ ðSr; CaÞCO3 þ ðSr; CaÞFe2 O4 þ ðSr; CaÞ2 Co2 O5

(19.2)

The authors pointed out that an unequal decrease in the brownmillerite-type sextets was not found when the new sextet grew, and it was not seen in the emission experiment either. This apparent contradiction could be resolved by assuming that since the oxygen stoichiometry cannot decrease further down below 2.5 in the brownmillerite structure, if Fe leaves the lattice from the preferred octahedral sites, Co has to substitute it, and one observes octahedrally coordinated nucleogenic Fe. Thus, the relative intensity of the two sextets has to remain the same. In summary, one can see that the brownmillerite type Sr0.5Ca0.5Fe0.5Co0.5O2.5 gave a brilliant example of how parallel EMS and TMS M€ ossbauer experiments could provide very special structural information on a system.

19.4 PURSUING COLOSSAL MAGNETORESISTANCE IN DOPED LANTHANUM COBALTATES In the early 1990s, Kusters et al. [21], and a little later, Helmolt et al. [22] discovered the so-called CMR around the magnetic transition temperature in thin-layer Nd0.5Pb0.5MnO3 and La0.7Ba0.3MnO3 perovskites, respectively. The CMR effect of these compounds—more than 60%—exceeded all known magnetoresistance at the time. In 1994, Jin et al. [23], then in 1995, Xiong et al. [24] published results on thin-layer La0.67Ca0.33MnO3 and Nd0.7Sr0.3MnO3 showing more than 99.9% magnetoresistance (MR), that is, the resistance dropped by an order of magnitude of three and six for La0.67Ca0.33MnO3 and Nd0.7Sr0.3MnO3, respectively. This was a truly colossal effect; however, later on colossal magnetoresistance became a name of all MR arising around the Curie temperature, whatever the extent is. This huge change in conductivity with external magnetic field promises significant improvement in the field of, for example, data storage systems or sensors; although a few drawbacks, such as strong temperature dependence or the need of strong magnetic field, hinder its immediate use in electronics. In addition to manganates, analogous cobalt-based perovskites (e.g., La1 xSrxCoO3) were also found to show colossal magnetoresistance [25–27]. Although the CMR peak of these cobaltate perovskites was found to be less (around 10%) than that of manganates, lanthanum cobaltates with strontium content x < 0.2 showed much higher magnetoresistance at low temperatures. Modifying the original ABO3 perovskite structure by doping not only on the rare earth site (A) but also on the transition metal site (B), for example, with iron, results in significant changes in the electric and magnetic properties, thus in the magnetoresistance as well. Barman et al. [28] even reported that in La0.8Sr0.2Co1 yFeyO3 (0.025  y  0.3) perovskites not only at low temperatures (roughly below 50 K) but also above 150 K, what is more, even at room temperature, high magnetoresistance can be observed. For example, at iron content y ¼ 0.025, MR ¼ 90% could be measured at 300 K. However, room-temperature MR has not yet been successfully demonstrated by other groups. In order to better understand the mechanisms behind MR in cobaltate perovskites, first the determination of the highly dubious valence and spin states of both cobalt and iron ions, then a detailed description of the effect of iron doping on the local and bulk electric and magnetic properties are necessary. Cobaltate perovskites represent a well-established family of transition metal compounds with complex electronic and magnetic structures. The base system of the hereby presented materials, LaCoO3, is a frequent research topic due to its complicated spin transitions and the correlated metal-to-insulator and magnetic transitions driven by temperature, pressure, or partial exchange of lanthanum ions [29–33]. In contrast with manganate perovskites, where the spin state of

408

€ STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER

manganese ions is always high-spin state, cobalt can be easily converted between low-, intermediate-, or high-spin states for both trivalent and tetravalent forms, making the case of cobaltates even more intriguing. While LaCoO3 is a low-spin cobalt 3þ compound at low temperatures, replacing less than half of the trivalent lanthanum ions with divalent strontium oxidizes the CoO3 sublattice, and, due to the different sizes of the La3þ and Sr2þ ions, favors the higher spin state for trivalent cobalt ions [30]. Accordingly, the bulk properties of these La1 xSrxCoO3 perovskites change drastically with Sr content. Above a percolation threshold (about x  0.18) these perovskites show bulk ferromagnetism and metallic conductivity, while below this limit they are spin glass like and more resistive to electric current. However, all compositions exhibit complex magnetic structure, that is, a mixture of both paramagnetic and ferromagnetic regions [30], but it is still a question whether both the relative quantity and the size or the composition of the isomorphic magnetic clusters are responsible for the bulk behavior. On the other hand, besides the doping of lanthanum sites, cobaltate perovskites can also be doped by Fe, giving a further parameter to tune their electronic and magnetic properties. € ssbauer Study of La0.8Sr0.2CoO3 19.4.1 Emission Mo

d

Perovskites

ossbauer investigations on the cobalt site As a first step toward the understanding of B site doping in LaCoO3, detailed M€ with the most promising A site configuration of La0.8Sr0.2 were carried out. The pioneering works of Bhide et al. [34,35] clearly demonstrated the applicability of EMS on cobaltate perovskites presenting spectra of doped and undoped LaCoO3; however, these studies were not followed for a period of decades. As the interest in doped cobaltate perovskites boosted around the turn of the millennium due to the discovery of the CMR effect, the need for M€ ossbauer studies of these oxides became evident. The repeated measurements [36] showed ossbauer spectra of the La0.8Sr0.2CoO3 d (Figs. 19.16 and 19.17) resemble those published by that the 57 Co emission M€ Bhide et al. [35]; that is, between 78 and 191 K, they consist of a paramagnetic doublet and a magnetic sextet with broad lines, while above 218 K only an unresolved quadrupole-split M€ ossbauer spectrum is observed. The temperature at which 1.000 1.000

0.995

0.980

149 K

1.000

0.980 218 K

Relative transmission

Relative transmission

300 K

1.000

1.000

0.990 128 K

1.000

0.990 0.995 191 K

0.980 –8

–6

–4

–2

0

2

4

Velocity (mm

s–1)

6

78 K

8

FIGURE 19.16 Selected

57

€ ssbauer spectra of La0.8Sr0.2CoO3 d. Co emission Mo

–8

–6

–4

–2

0

Velocity (mm

2 s–1)

4

6

8

409

19.4 PURSUING COLOSSAL MAGNETORESISTANCE IN DOPED LANTHANUM COBALTATES

1.000

1.000

0.980 0.990

1.000

0.990

0.980

150 K

Relative transmission

Relative transmission

115 K

300 K

0.960

1.000

0.995

0.990

1.000

100 K

1.000

0.990

0.995 125 K

78 K

0.980 –8

–6

–4

–2

0

2

4

6

8

–8

–6

Velocity (mm s–1)

–4

–2

0

2

4

6

8

Velocity (mm s–1)

FIGURE 19.17 Selected

57

€ ssbauer spectra of La0.8Sr0.2CoO3 Co emission Mo

d

after oxygen removal.

the magnetic subspectrum disappears (Fig. 19.18) coincides with the first magnetic transition temperature and the CMR peak of this perovskite. [30,36] ossbauer spectra of The coexistence of the paramagnetic doublet and the magnetic sextet in the 57 Co emission M€ La0.8Sr0.2CoO3 d was discussed as a coexistence of two different environments of the investigated transition metal ions below the transition temperature [36]. The M€ ossbauer spectra showed that below about 218 K, only a part of the single paramagnetic subspectrum converts into magnetic sextet; the other part of the doublet (which diminishes with 0

as prepared y = 0 y = 0 after oxygen removal y = 0.025

sextet (% )

80

20

60

40

40

60

20

80

0

100

100

150

200 T (K)

250

300

% doublet

100

FIGURE 19.18 Area ratios between doublet and sextet components in the € ssbauer spectra of Mo La0.8Sr0.2Co1 yFeyO3 d perovskites.

410

€ 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS

FIGURE 19.19 Isomer shifts of doublets and € ssbauer spectra sextets in the Mo of La0.8Sr0.2Co1 yFeyO3 d perovskites.

decreasing temperature) remains magnetically unsplit till 78 K (the lowest temperature measured). Based on the magnetic cluster model, a theory describing the anomalous magnetic behavior of mixed perovskites [30], one can identify the magnetic subspectrum with the signal of nucleogenic iron ions originating from those cobalt ions that participated in the Sr-rich magnetically ordered clusters. Moreover, the broad lines of the magnetic subspectrum were fitted by the assumption of a small distribution of the hyperfine magnetic field, which may refer to a distribution of the size of the clusters. A distribution of hyperfine couplings in this perovskite was also observed by 59 Co NMR spectroscopy [37]. On the other hand, the paramagnetic fraction of the M€ ossbauer spectra can be attributed to the paramagnetic matrix of Co3þ 4þ ions (that is, the Sr- and Co -poor regions). Nemeth et al. also outlined that the M€ ossbauer isomer shifts of the sextets (dsextet) and doublets (ddoublet) were found to differ explicitly in every measured spectrum (Fig. 19.19) . While ddoublet was found to be between 0.46 and 0.24 mm s 1 for 78 K  T  300 K, values of dsextet were in the range of 0.32–0.21 mm s 1 between 78 and 191 K, and an average gap of about 0.15 mm s 1 could be found between ddoublet and dsextet measured at the same temperatures. The distinct isomer shifts refer to a slight difference between the valence electron densities of the nucleogenic iron ions. While the higher ddoublet refers to high-spin trivalent state, the lower dsextet values indicate a small decrease of the population of 3d electrons, which can arise from, for example, delocalization of valence electrons. This result is in good agreement with the presumption of the cluster model: while in the magnetically ordered conductive clusters the eg electrons of the Co3þ ions are delocalized between the trivalent and tetravalent cobalt ions (due to the double exchange process resulting in metallic conduction and ferromagnetic exchange interaction), in the paramagnetic matrix they remain more localized. Thus, the isomer shift referring to the ions in the ordered clusters should be indeed lower (representing a mixed 3þ and 4þ valence character) than that of the doublets referring mostly to the paramagnetic matrix. This finding is in agreement with the existence of an electronic phase separation in the investigated perovskites. With the help of this model, the authors gave a qualitative description of the observed magnetoresistance of La0.8Sr0.2CoO3 d. Moreover, the isomer shift of the quadrupolar doublet in the M€ ossbauer spectra above the Curie temperature is in good agreement with the values of ddoublet below TC, instead of becoming an average between the isomer shift values of the sextet and the doublet components at low temperatures. This suggests that the electron localization in the clusters above TC seems to prefer the (nucleogenic or dopant) iron ions. Otherwise, an additional doublet with lower isomer shift (corresponding to Fe4þ valence state) should have appeared above TC as well. ossbauer The consequences of oxygen removal in the octahedra in CoO3 sublattice have been studied before the M€ studies and the outcome was invariably that it is detrimental to ferromagnetism and, consequently, to the formation of ferromagnetic clusters [28,38,39]. That was the reason to study the effect of a controlled oxygen removal on the ossbauer measurements (Fig. 19.17) also showed the La0.8Sr0.2CoO3 d perovskite. The subsequent 57 Co emission M€ coexistence of ferromagnetic (FM) and paramagnetic (PM) subspectra below the magnetic transition temperature, qualitatively similar to the spectra observed in the unaltered parent compound [40]. In both cases, the FM/PM ratios change quite abruptly at the transition temperature (Fig. 19.18). A clear change of electron localization in the PM and FM phases could be well followed by the comparison of M€ ossbauer isomer shifts of the sextets (dsextet) and doublets (ddoublet) in the case of oxygen removal as well. However, the authors found that the relative area of the magnetic subspectrum is smaller in the case of the sample with higher oxygen deficiency than in the parent La0.8Sr0.2CoO3 d. [40] This can be

411

19.4 PURSUING COLOSSAL MAGNETORESISTANCE IN DOPED LANTHANUM COBALTATES

attributed mainly to the decreasing number of Co4þ ions, as the oxygen vacancies compensate for the extra electrons from La3þ to Sr2þ exchange. Additionally, the decrease in the long-range magnetic coherence length within the clusters was explained by the partial collapse of the double exchange process between Co3þ and Co4þ ions, which is caused by the missing intermediate oxygen ion. € ssbauer Study of Iron-Doped La0.8Sr0.2FeyCo1 yO3 19.4.2 Emission and Transmission Mo

d

Perovskites

In contrast to the modification of the number of oxygen vacancies, iron doping offers a more controllable way to alter the electronic and magnetic interactions in cobaltate perovskites. Thus, a series of experiments on doping various amounts of ossbauer iron at the cobalt site of La0.8Sr0.2CoO3 d were carried out and followed with the help of transmission M€ spectroscopy. 19.4.2.1 The Case of y  0.05 Iron Content A low doping value of y ¼ 0.025 57 Fe was found to show a very similar effect on the M€ ossbauer spectra of La0.8Sr0.2CoO3 d perovskites as in the case of oxygen vacancy creation: a coexistence of doublet and sextet in the M€ ossbauer spectra below the magnetic ordering temperature occurred with a distinguishing difference in isomer shifts representing the more insulating paramagnetic phase and the more conductive ferromagnetic phase, respectively [41]. However, further decrease in the amount of ferromagnetic phase, with respect to the undoped mother compound, was again outlined by the authors. This can be understood as the change of magnetic exchange interactions between cobalt ions due to the Co3þ/Fe3þ replacement. That is, iron doping was suggested to hinder the formation of magnetic clusters. A clear correlation between the above-described changes in M€ ossbauer spectra and the observed magnetoresistance was also pointed out. The absence of the MR peak at the Curie temperature in the 2.5% iron-doped sample, which was present in the as-prepared La0.8Sr0.2CoO3 d perovskite, was attributed to the magnetic clusters becoming too much separated to produce colossal magnetoresistance around TC. In contrast with the 2.5% doping, a La0.8Sr0.257 Fe 0.05Co0.95O3 d sample (that is, iron doping doubled) was found to show remarkably different behavior of magnetic transition in the recorded M€ ossbauer spectra [40]. The spectra observed between 120 K and room temperature showed a broad absorption peak that could be best fitted with a doublet showing a small (0.2 mm s 1) quadrupole splitting. Below about 120 K, a magnetic sextet developed gradually, referring to iron cations undergoing magnetic relaxation that slows down with decreasing temperature. The ratio of the paramagnetic component starts to fall with decreasing temperature at around Tc  120 K, and it disappears below about 65–70 K (Fig. 19.20) that coincides with the temperature where—according to AC susceptibility measurements— 1.00

1.00

0.95 0.90

0.99

0.85

67 K

300 K

1.000

0.95

0.995

0.90

125 K

0.85 1.00 0.98 0.96 0.94 0.92

57 K

0.990 1.00 0.99

100 K

40 K 0.98

1.00

1.00

0.98

0.99

0.96

–8

77 K –6

–4

–2

0

2

4

6

–8 –6 –4 –2 0

Velocity (mm s–1)

2

4

6

Relative transmission

Relative transmission

1.00

4.2 K 0.98 8 10

Velocity (mm s–1)

FIGURE 19.20 57

€ ssbauer spectra of La0:8 Sr0:2 57 Fe0:05 Co0:95 O3 Fe transmission Mo

d

perovskite at selected temperatures.

412

€ 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS

1.00

Relative transmission

0.98

FIGURE 19.21

0.96 0.94

290 K

1.00

0.99

4.2 K

57

€ ssbauer spectra Co emission Mo of La0:8 Sr0:2 57 Fe0:05 Co0:95 O3 d perovskite at two temperatures.

–10

–5

0

5

10

Velocity (mm s–1)

the spin dynamics changes. The fact that the iron magnetic moments reflect a tendency for the freezing of the relaxation of iron spins with decreasing temperature indicates the formation of spin or cluster glass phases, which is in accordance with the results of the AC magnetic susceptibility measurements of La0.8Sr0.257 Fe 0.05Co0.95O3 d displaying substantial blocking below T  65 K. At T ¼ 4.2 K, both EMS and TMS spectra were published. Both spectra (Fig. 19.20, 19.21) display a sextet with a distribution of hyperfine magnetic fields, although the average magnetic field of the EMS spectrum is higher than that of the TMS spectrum (46.2 versus 40.0 T, respectively). The difference of the magnetic distributions of the TMS and EMS spectra can be seen in Fig. 19.22. This was explained as follows: one may take into account that the EMS measurement corresponds to iron cations at cobalt sites, and so it reflects not only the local magnetic environment of the ironcontaining cobalt clusters but also those of the iron-free regions. The fact that the (originally) iron-free clusters can be characterized by higher hyperfine magnetic fields implies that their magnetic relaxation is slower because of their larger size. This suggests that due to iron substitution, relatively large cobalt clusters are broken up into smaller ones, giving further evidence to the cluster breaking effect of iron doping. 19.4.2.2 The Case of y  0.15 Iron Content Further doping with iron showed that the iron ions replacing cobalt break up the double exchange-based magnetic clusters [42]. It was clearly demonstrated by the M€ ossbauer spectra of La0.8Sr0.2FeyCo1 yO3 d perovskites with y ¼ 0.15, 0.2, and 0.3, where only a gradual transition from a high-temperature doublet to a broad, relaxed sextet phase occurred at constantly decreasing transition temperatures (Fig. 19.23) . This is in

TMS EMS

Weight of B

0.3

0.2

0.1

FIGURE 19.22 Hyperfine magnetic field distributions of the 57 Co EMS and 57 Fe € ssbauer spectra of TMS Mo La0:8 Sr0:2 57 Fe0:05 Co0:95 O3 d perovskite recorded at 4.2 K.

0.0 0

10

20

30

B (T)

40

50

413

REFERENCES

1.00

1.00

0.98 0.99

305 K

65 K

1.00

0.98 127 K 1.00

0.98 102 K 1.00

Relative transmission

Relative transmission

1.000 0.998 55 K 1.000

0.995

35 K

1.000

0.98

0.995 4,2 K

78 K –4

–2

0

Velocity (mm

2 s–1)

FIGURE 19.23 € ssbauer spectra of La0:8 Sr0:2 Fe0:3 Co0:7 O3 Mo

4

–12

–8

–4

0

Velocity (mm

d

4

8

12

s–1)

perovskite at selected temperatures.

a good accordance with the cluster model depicted magnetoresistance: the CMR peak was found to disappear with y  0.15 as the doping iron ions diminished the magnetic clusters in La0.8Sr0.2FeyCo1 yO3 d perovskites.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

J.G. Bednorz, K.A. M€uller, Z. Phys. 1986, B64, 189–193. A. Vertes, Z. Homonnay, eds., M€ossbauer Spectroscopy of Sophisticated Oxides, Akademiai Kiad o, Budapest, 1997. A. Nath, Z. Homonnay, Physica C 1989, 161, 205–208. M. Francois, A. Junod, K. Yvon, A.W. Hewat, J.J. Capponi, P. Strobel, M. Marezio, P. Fischer, Solid State Commun. 1988, 66, 1117– 1125. E.L. Muetterties, MTP International Review of Science, Vol. 9, M.L. Tobe, ed., Butterworths, London, 1972 p. 32. Z. Klencsar, E. Kuzmann, A. Vertes, P.C.M. Gubbens, A.M. van der Kraan, M. B odogh, I. Kotsis, J. Radioanal. Nucl. Chem. 2000, 246, 113–115. R. Fehrenbacher, T.M. Rice, Phys. Rev. Lett. 1993, 70, 3471–3474. A.I. Liechtenstein, I.I. Mazin, Phys. Rev. Lett. 1995, 74, 1000–1003. Z. Homonnay, Z. Klencsar, V. Chechersky, Gy. Vank o, M. Gal, E. Kuzmann, S. Tyagi, J.-L. Peng, R.L. Greene, A. Vertes, A. Nath, Phys. Rev. B, 1999, 59, 11596–11604. V. Chechersky, M. Gal, Z. Homonnay, Gy. Vank o, E. Kuzmann, S. Tyagi, R.L. Greene, A. Vertes, A. Nath. Physica C 1997, 277, 36–42. M. Guillaume, P. Allenspach, W. Henggeler, J. Mesot, B. Roessli, U. Staub, P. Fischer, A. Furrer, D.J. Trounov, J. Phys. Condens. Matter 1994, 6, 7963–7976. J.A. Hodges, G. le Bras, P. Bonville, P. Imbert, G. Jehanno, Physica C 1993, 218, 283–294. M. Takata, T. Takayama, M. Sakata, S. Sasaki, K. Kodama, M. Sato, Physica C 1996, 263, 340–343.

414

€ STUDIES ON PEROVSKITES AND RELATED OXIDE SYSTEMS 19 TRANSMISSION AND EMISSION 57Fe MOSSBAUER

14. Z. Homonnay, M. Gal, Gy. Vank o, E. Kuzmann, A. Vertes, Conference Proceedings, “ICAME-95”, Vol. 50, Ida Ortalli, ed., SIF, Bologna, 1996, pp. 551–554. 15. K. Nomura, Y. Ujihira, T. Hayakawa, K. Takehira, Appl. Catal. A. 1996, 137, 25–36. 16. G. Juhasz, Z. Homonnay, K. Nomura, T. Hayakawa, S. Hamakawa, A. Vertes, Solid State Ionics 2001, 139, 219–231. 17. Z. Homonnay, K. Nomura, G. Juhasz, E. Kuzmann, S. Hamakawa, T. Hayakawa, A. Vertes, J. Radioanal. Nucl. Chem. 2001, 255, 425– 429. 18. Z. Homonnay, K. Nomura, G. Juhasz, M. Gal, K. S olymos, S. Hamakawa, T. Hayakawa, A. Vertes, Chem. Mater. 2002, 14, 1127– 1135. 19. T.C. Gibb, J. Chem. Soc., Dalton Trans. 1985, 1455–1470. 20. G. Baumgartel, P.J. Jensen, K.H. Bennemann, Phys. Rev. B 1990, 42, 288–293. 21. R.M. Kusters, J. Singleton, D.A. Keen, R. McGreevy, W. Hayes, Physica B 1989, 155, 362–365. 22. R. von Helmolt, J. Wecker, B. Holzapfel, L. Schultz, K. Samwer, Phys. Rev. Lett. 1993, 71, 2331–2333. 23. S. Jin, T.H. Tiefel, M. McCormack, R.A. Fastnacht, R. Ramesh, L.H. Chen, Science 1994, 264, 413–415. 24. G.C. Xiong, Q. Li, H.L. Ju, S.N. Mao, L. Senapati, X.X. Xi, R.L. Greene, T. Venkatesan, Appl. Phys. Lett. 1995, 66, 1427–1429. 25. V. Golovanov, L. Mihaly, A.R. Moodenbaugh, Phys. Rev. B 1996, 53, 8207–8210. 26. R. Mahendiran, A.K. Raychaudhuri, Phys. Rev. B 1996, 54, 16044–16052. 27. S. Yamaguchi, H. Taniguchi, H. Takagi, T. Arima, Y. Tokura, J. Phys. Soc. Jpn. 1995, 64, 1885–1888. 28. A. Barman, M. Ghosh, S. Biswas, S.K. De, S. Chatterjee, Appl. Phys. Lett. 1997, 71, 3150–3152. 29. Gy. Vank o, J.-P. Rueff, A. Mattila, Z. Nemeth, A. Shukla, Phys. Rev. B 2006, 73, 024424. 30. J. Wu, C. Leigthon, Phys. Rev. B 2003, 67, 174408. 31. J. Wu, H. Zheng, J.F. Mitchell, C. Leigthon, Phys. Rev. B 2006, 73, 020404. 32. H.M. Aarbogh, J. Wu, L. Wang, H. Zheng, J.F. Mitchell, C. Leigthon, Phys. Rev. B 2006, 74, 134408. 33. D. Phelan, D. Louca, S. Rosenkranz, S.-H. Lee, Y. Qiu, P.J. Chupas, R. Osborn, H. Zheng, J.F. Mitchell, J.R.D. Copley, J.L. Sarrao, Y. Moritomo, Phys. Rev. Lett. 2006, 96, 027201. 34. V.G. Bhide, D.S. Rajoria, G. Rama Rao, C.N.R. Rao, Phys. Rev. B 1972, 6, 1021–1032. 35. V.G. Bhide, D.S. Rajoria, C.N.R. Rao, G. Rama Rao, V.G. Jadhao, Phys. Rev. B 1975, 12, 2832–2843.  36. Z. Nemeth, Z. Homonnay, F. Arva, Z. Klencsar, E. Kuzmann, J. Hakl, K. Vad, S. Meszaros, K. Kellner, G. Gritzner, A. Vertes, J. Radioanal. Nucl. Chem. 2007, 271, 11–17. 37. P.L. Kuhns, M.J.R. Hoch, W.G. Moulton, A.P. Reyes, J. Wu, C. Leighton, Phys. Rev. Lett. 2003, 91, 127202.  Cziraki, I. GerÅcs, M. K€ 38. A. oteles, A. Gabris, L. Pogany, I. Bakonyi, Z. Klencsar, A. Vertes, S.K. De, A. Barman, M. Ghosh, S. Biswas, S. Chatterjee, B. Arnold, H.D. Bauer, K. Wetzig, C. Ulhaq-Bouillet, V. Pierron-Bohnes, Eur. Phys. J. B 2001, 21, 521–526. 39. Y. Sun, X. Xu, Y. Zhang, Phys. Rev. B 2000, 62, 5289–5292. 40. Z. Nemeth, Z. Klencsar, E. Kuzmann, Z. Homonnay, A. Vertes, J.M. Greneche, B. Lackner, K. Kellner, G. Gritzner, J. Hakl, K. Vad, S. Meszaros, L. Kerekes, Eur. Phys. J. B 2005, 43, 297–303.  41. Z. Nemeth, Z. Homonnay, F. Arva, Z. Klencsar, E. Kuzmann, A. Vertes, J. Hakl, S. Meszaros, K. Vad, P.F. de Ch^atel, G. Gritzner, Y. Aoki, H. Konno, J.M. Greneche, Eur. Phys. J. B 2007, 57, 257–263. 42. Z. Klencsar, Z. Nemeth, E. Kuzmann, Z. Homonnay, A. Vertes, J. Hakl, K. Vad, S. Meszaros, A. Simopoulos, E. Devlin, G. Kallias, J.M.  Cziraki, S.K. De, J. Magn. Magn. Mater. 2008, 320, 49–59. Greneche, A.

C H A P T E R 2 0

ENHANCING THE POSSIBILITIES € OF 57Fe MOSSBAUER SPECTROMETRY TO STUDY THE INHERENT PROPERTIES OF RUST LAYERS SAR A. BARRERO,1 ALVARO L. MORALES,1 AND JEAN-MARC GRENECHE2 KAREN E. GARCIA,1 Ce 1 2

Grupo de Estado Solido, Facultad de Ciencias Exactas y Naturales, Universidad de Antioquia, Medellın, Colombia LUNAM, Universite du Maine, Institut des Molecules et Materiaux du Mans, Le Mans Cedex, France

20.1 INTRODUCTION Nowadays, it is widely accepted that the protective properties of steels against corrosion largely depend on the inherent properties of the rust layers [1–25]. Usually, the same iron phases in the rust are formed under a broad range of environmental conditions and steel composition, but their inherent properties (composition, morphology, structure, and so on) and their relative abundances can be very different. Therefore, a careful quantitative characterization of the iron phases composing the rust layers is an essential prerequisite to a complete understanding of the corrosion processes. One of the ossbauer spectrometry (VT-MS), principal characterization techniques used for this purpose is variable-temperature 57 Fe M€ which is able to distinguish Fe atoms with different structural and magnetic environments and with different valence states. According to Cook [11], VT-MS is the most important analytical technique for quantifying the corrosion properties of steel in different environments. The hyperfine parameters and recoil-free fractions for common iron phases have been tabulated [13,19,22]. Therefore, VT-MS allows identification of corrosion products, particularly their nanophase components, which are difficult to identify by using other techniques [11–13]. It is also able to determine the fraction of each iron phase present in the rust. The combination of transmission, gamma scattering, and conversion electron detection allows in-depth characterization of rust layer [12,13]. Another advantage of this technique is its nondestructive character [12]. The M€ ossbauer results are usually complemented by means of other techniques such as quantitative X-ray diffraction (XRD) ossbauer spectrometry analysis and infrared and Raman (and microRaman) spectroscopies. The possibilities of using 57 Fe M€ in corrosion research have been reviewed previously by different authors [11–13]. These works have clearly demonstrated that proper use of VT-MS allows the mechanisms associated with the deterioration processes to be better understood. Very recently it has been shown that with VT-MS it is also possible (i) to evaluate the protective ability index [3,4] of the rust layers [14,15] and (ii) to estimate the atomic fraction (a) of iron species coming from the steels that after the corrosion processes form part of the adherent rust [15–18].

M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

415

416

€ SPECTROMETRY TO STUDY THE INHERENT PROPERTIES 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER

In order to retrieve the inherent properties of the rust using VT-MS, it is essential to use proper fitting models to analyze the M€ ossbauer spectra (MS) of each iron phase. The methodology is rather simple for well-crystallized, stoichiometric, large-particle iron phases because the hyperfine structure results from quadrupolar components and/or magnetic sextets with narrow Lorentzian lines [19–22]. However, in the case of corrosion products, this analytical procedure becomes more complex: indeed, they consist of mixtures of different phases with nonstoichiometry, cation for iron substitutions and with crystalline grain, the size of which can be distributed from nanometer to micrometer range. Such features result in from broadening of M€ ossbauer lines due to combined static and dynamic effects, resulting from various local environments of the Fe probe and the presence of fast and slow superparamagnetic fluctuations induced by the grain-size distribution and their mutual interaction effects, respectively. In relation to corrosion studies, fitting procedures combining doublets and sextets with Lorentzian line shapes and distributions of magnetic hyperfine fields have been used [8,11,19,23]. Doublets and sextets with Voigtian line shapes have also been employed [24]. Another interesting fitting model, which allows direct determination of grain-size distribution of each iron phase and relies on the superparamagnetic phenomena, has been proposed by Mizoguchi et al. [9] and Okada et al. [25]. As we will notice below, in spite of the great number of investigations, there is poor agreement in the number of components needed to properly fit the MS and their physical origin in both the paramagnetic and the magnetically ordered regions for some of the iron phases presented in the rust layers. ossbauer spectrometry in corrosion research are reviewed in the case of In this work, the possibilities of using 57 Fe M€ three accelerated corrosion tests in chloride environments. Some suggestions for further research in this field are also proposed. Before introducing these examples, we will briefly discuss some models used to fit some of the iron phases in rust layers, such as goethite, akaganeite, and spinel phases. Additional information for other iron phases can be found elsewhere [11–13,19–22,26].

€ 20.2 MOSSBAUER CHARACTERIZATION OF SOME IRON PHASES PRESENTED IN THE RUST LAYERS 20.2.1 Akaganeite Room temperature M€ ossbauer spectrometry (RT-MS) of akaganeite has been fitted with either two or three quadrupolar components with poor agreement about their physical origins [27,28]. Murad [29] reported that the two quadrupolar components arise because of the increasing distortion of midwall to channel near iron sites. This interpretation has been based upon a superstructure concept, which was lately rejected by other authors [27]. Chambaere and de Grave [30] attributed the lower and higher quadrupolar components to iron sites with O3(OH)3 and O2(OH)4 coordination, respectively. However, recent investigations have suggested that only O3(OH)3 coordination takes place in akaganeite [31–33]. Rezel and Genin [34] have proposed that RT-MS must be fitted with three instead of two quadrupolar components, and they argued that the replacement of Cl ions by OH ions in the structure is an important factor to the presence of different iron sites. However, the replacement of Cl ions for OH ions needs further experimental support and is still a matter of controversy. The previously mentioned interpretations are based on a tetragonal unit cell for akaganeite. Post and Buchwald [31] proposed a monoclinic structure and related the two quadrupolar components with the two nonequivalent iron sites, Fe1 and Fe2, required by this unit cell. However, there is a problem with this interpretation because the number of iron ions per unit cell is equal to four for both crystallographic sites, that is, there are as many Fe1 sites as there are Fe2 sites per unit cell. If it is assumed that the Debye–Waller factor for both sites is similar to what it should be, then it is expected that the area ratio of both sites should be approximately equal to one, and not to about 35/65 ¼ 0.54, as usually estimated. Very recently, another interpretation based on the monoclinic structure and the chlorine content has been proposed by Barrero et al. [28]. They suggested that one Fe site is assigned to Fe3þ ions located close to chlorines and the other Fe site to the ions located close to vacant chlorine sites. It is noticed that since the works by Chambaere et al. [27,30] and Murad [29], the model involving two quadrupolar components is generally accepted as the most adequate way of fitting the room temperature M€ ossbauer spectrum of akaganeite (Fig. 20.1); however, as previously noticed, there is poor agreement about the physical origin of these components. Typical ranges of hyperfine parameters are listed in Table 20.1. Now, in relation to the magnetic ordered range below the Neel temperature of about 295 K, most works agree that three sextets are required to adequately fit the spectra. Most investigations based on the tetragonal structure agree in that the presence of Cl ions in the channels is the cause of the nonequivalent iron sites in the structure. According to Murad [29], the three components arise because of the increasing distortion of midwall to channel-near iron sites. A

€ 20.2 MOSSBAUER CHARACTERIZATION OF SOME IRON PHASES PRESENTED IN THE RUST LAYERS

417

€ ssbauer Spectra of Some TABLE 20.1 Range of Hyperfine Parameters Derived from the Room Temperature Mo Iron Phases [13,19,22,26] Component

d (mm s 1)

D, 2e (mm s 1)

Akaganeite

0.38–0.39 0.37–0.38 0.28–0.35 0.27–0.28 0.63–0.67

0.92–0.95 0.50–0.52 0.24 to 0.26 0–0.01 0.16 to 0.08

Goethite Magnetite

B (T)

f

– – 23–37.8 48.7–49.3 45.4–46.0

0.77–0.81 0.77–1.05 0.87–0.90

Transmission (a. u.)

d: isomer shift relative to a-Fe at room temperature, 2e: quadrupole shift, D: quadrupole splitting, B: magnetic hyperfine field, f: M€ ossbauer fraction relative to that of hematite [13,19].

A pure

FIGURE 20.1 –2

–1

0 Velocity (mm s–1)

1

2

€ ssbauer Room temperature Mo spectrum of akaganeite.

Transmission (a. u.)

different interpretation was put forward by Chambaere and De Grave [35], for whom two sextets are attributed to the Fe–O3(OH)3 and to the Fe–O2(OH)4 octahedral units and the third one is a contribution with fluctuating magnetization, whose intensity at absolute 0 K approaches zero. On the other hand, Rezel and Genin [34] proposed that two of the three magnetic components originate from the iron atoms that have chlorines as the first nearest neighbors, whereas the third component is due to iron atoms far from the chlorine ones. According to Oh et al. [19], the 77 K spectrum should be fitted with a distribution of hyperfine fields originating from the several magnetically nonequivalent iron sites. This way of fitting does not allow quantitative information on the crystallographic parameters and chlorine content to be directly retrieved. Stahl et al. [32] proposed that two of the three components are related to the two crystallographic sites required by the monoclinic space group, and that the third component arises because of the chlorine disorder. However, this work assumes that chlorine is subjected to a disorder, which clearly contradicts the results of Post et al. [31,33] and Weckler and Lutz [36], who proposed an ordering scheme with every third chlorine site vacant in a particular tunnel. Barrero et al. [28] also interpreted the spectra in terms of the monoclinic cell and chlorine content and used four sextets due to the presence of four nonequivalent iron sites arising from Fe13þ and Fe23þ located close to the Cl ions and Fe13þ and Fe23þ located near the vacant chlorine sites. In summary, below the Neel temperature, the M€ ossbauer spectra of akaganeite can be adequately fitted with either three or four sextets (Fig. 20.2). Typical hyperfine parameters of akaganeite (A) in the low temperature region are reported in Table 20.2.

A 77K

–9

–6

–3

0

3

Velocity (mm s–1)

6

9

FIGURE 20.2 € ssbauer spectrum of 77 K Mo akaganeite.

418

€ SPECTROMETRY TO STUDY THE INHERENT PROPERTIES 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER

€ ssbauer Spectra of Some Iron Phases TABLE 20.2 Range of Hyperfine Parameters Derived from the 77 K Mo [13,19,22,26] Component

d (mm s 1)

Akaganeite

0.50–0.52 0.46–0.49 0.46–0.51 0.28–0.32 0.37 0.49 0.83 1.03 0.96

Goethite Magnetite (4.2 K) [59]

D, 2e (mm s 1) 0.04 to 0.10 0.22 to 0.34 0.36 to 0.49 0.23 to 0.25 0.02 0.00 0.27 0.41 0.89

B (T)

f

47.0–48.4 46.1–46.7 41.6–44.2 23.0–48.9 50.1 52.2 49.8 48.2 35.9

0.86

0.80 0.88

d: isomer shift relative to a-Fe at room temperature, 2e: quadrupole shift, D: quadrupole splitting, B: magnetic hyperfine field, f: M€ ossbauer fraction relative to that of hematite [19].

Next, in relation to the magnetic structure, some works assume a collinear antiferromagnetism in akaganeite [19,20,29,30]. However, Coey suggested speromagnetism instead of antiferromagnetism in all iron oxyhydroxides [37]. Barrero et al. [28] proposed that the magnetic structure behaves as a system that consists of two asperimagnetic-like structures, which are antiferromagnetically coupled, because the disorder implies different local atomic neighboring. Supporting this idea, Millan et al. [38] recently reported the appearance of a small magnetic moment in akaganeite nanoparticles due to size effects and a deficient Cl site occupancy. These authors mentioned that the Cl ions play a role in magnetic exchange interactions and a perfect sublattice alignment is achieved only when the crystal lattice is saturated with Cl ions. Until now, we have reviewed M€ ossbauer investigations on pure akaganeites. However, in the corrosion process, the formation of akaganeite is affected by the simultaneous presence of cations and anions. Ishikawa et al. [39–41] studied the influences of cations, such as Ti4þ, Cr3þ, Cu2þ, and Ni2þ, and anions, such as SO42 , HPO42 , NO3 , and SiO32 , on the formation of akaganeite. This research group concluded that among the cations, Ti4þ is the one that more drastically affects both the size of structural domains and the crystalline state of akaganeite. The authors also showed that the crystallinity of akaganeite was appreciably reduced by adding SO42 and HPO42 , improved by adding SiO32 , but it was slightly influenced by NO3 . Kamimura et al. [42] have also studied the influence of cations (Cr3þ, Cu2þ, and Ni2þ) and anions (SO42 and NO3 ) on the formation of akaganeite. An important result is that the crystallinity of the powdered materials depends on the combination of cations and anions. For example, the XRD peaks of akaganeite are broadened when the sulfates are added as anions, but, on the contrary, there is no significant change in the XRD peaks when the nitrates are added. Remazeilles and Refait [43] reported that in order to obtain the formation of akaganeite, both iron (II) and chlorine concentrations are important parameters. By combining their results with those reported by Garcıa et al. [16,17], it is suggested that besides the necessary abundance of chlorine and iron, the presence of akaganeite also needs high-humidity conditions in corrosion processes for long periods of time. Barrero et al. [44] investigated the influence of Al3þ and Ti4þ ions on the formation of akaganeite when prepared by hydrolysis of FeCl3 solutions. Akaganeite was the only phase being formed independent of the type and concentration of the metallic cations; however, Ti4þ is the cation that more drastically affects the physical properties of akaganeite. Garcıa et al. [45] investigated the akaganeite formed in the presence of Al3þ, Cr3þ, Ti4þ, and Cu2þ ions and urea when it is obtained by the hydrolysis of FeCl3 solutions. They found that the properties of akaganeite are slightly affected by most of the alloying elements at low concentrations, and that, therefore, this compound when formed as corrosion product of weathering and carbon steels should exhibit very similar properties to those of synthetically pure obtained akaganeite. From all these features, it is suggested that M€ ossbauer spectra of akaganeite as a corrosion product can be described as in the case of pure akaganeite. 20.2.2 Goethite Goethite (a-FeOOH) (G) is a common mineral in soils and sediments on Earth [22], and has recently also been found on Mars [46]. It behaves as an antiferromagnet below its Neel temperature established at around 400 K when the grain size exceeds 100 nm. RT-MS of well-crystallized large-grain goethite consists of a sextet with symmetric narrow lines, whose hyperfine parameters (Table 20.1) are typical of Fe3þ in octahedral coordination (Fig. 20.3 [47]). As the average grain size and crystallinity decrease, the lines start to broaden asymmetrically toward the centroid of the spectrum and in a further

€ 20.2 MOSSBAUER CHARACTERIZATION OF SOME IRON PHASES PRESENTED IN THE RUST LAYERS

419

100 100 98 99 96 98

94

(a)

(e) 100

98

99

96

98

94

97

(b)

(f)

100

99

99

97

98

95

97

93

(c)

96

(g)

Transmission (%)

Transmission (%)

100

91

100

98

99 94 98 90

97 96

(d)

86

(h)

82 –10 –8 –6 –4 –2 0

2

4

Velocity (mm s–1)

6

8

–10 –8 –6 –4 –2 0

2

4

6

8 10

Velocity (mm s–1)

FIGURE 20.3 € ssbauer spectra of pure goethites with different mean crystallite dimensions in the (111) Room temperature Mo direction, MCD111: (a) 54.4 nm, (b) 44.3 nm, (c) 19.1 nm, (d) 17.8 nm; and Al-substituted goethites of about MCD111 ¼ 30 nm with different Al contents: (e) 4.7 at.%, (f) 9 at.%, (g) 12.4 at.%, (h) 15.7 at.% [47].

stage, the sextet starts to collapse and a doublet appears in the center. This process continues down to mean crystallite sizes of about 15 nm [48,49]. Oh et al. [8,19] and Cook [11] reported that the MS for goethites with grain sizes between 8 and 15 nm consist of a superparamagnetic doublet at 300 K and a magnetic sextet at 77 K. For crystallite sizes lower than 8 nm, a single superparamagnetic doublet occurs at both 300 and 77 K MS. The magnetic properties observed in goethite, which clearly affect the shape of the MS, are very complex and strongly depend on several structural and chemical properties. Barrero et al. [48,49] have attempted to incorporate in a unique phase diagram, the magnetic behavior as a function of temperature, surface area S (which is directly related to the grain size), and total water content DOH. DOH takes into account the nonstoichiometric parameter y (the chemical formula for nonstoichiometric goethite is given by a-Fe1 (y/3)O1 y(OH)1þy) and the adsorbed water content. Ideal goethite (DOH/S values above 0.50% m 2 g) with low S (large particle goethite) and perfect stoichiometry (y ¼ 0) is a collinear antiferromagnet below TN 400 K. As DOH/S decreases, and below TN, goethite exhibits a complex, including antiferromagnetism, weak ferromagnetism and a possible magnetic ordering of clusters. Like in akaganeite, the magnetic structure could alternatively behave as a system that consists of two antiferromagnetically coupled asperimagnetic-like structures. It is interesting to mention that for certain

420

€ SPECTROMETRY TO STUDY THE INHERENT PROPERTIES 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER

DOH values and below certain temperatures (e.g., 30 K), goethite can exhibit cluster glass-like ordering. As DOH/S decreases further, there are certain critical values, which strongly depend on grain size (surface area), for which goethite becomes superparamagnetic. Above certain S values, the magnetic properties are related to surface area properties. Now, goethite as a corrosion product can be formed in the presence of several cations that, when incorporated into crystallographic structure, makes the sample more defective. In the case of Al ions, Fig. 20.3e–h shows how the M€ ossbauer spectra change with Al incorporation into the crystallographic structure. It is important to emphasize that the hyperfine structure in MS spectra becomes much more complex and the fitting model has to be validated with further M€ ossbauer experiments versus both temperature and external magnetic field. Some typical hyperfine parameters at 77 K are listed in Table 20.2. 20.2.3 Magnetite/Maghemite For a proper discussion on the methodology of analysis of the RT-MS for spinel phases, we have to differentiate between the spectra of stoichiometric magnetite (Fe3O4), nonstoichiometric magnetite (Fe3 xO4) with 0 < x < 1/3, and maghemite (g-Fe2O3), that is, when x ¼ 1/3. Magnetite Fe3O4, which is probably the oldest known magnetic material, is a mixed valence Fe oxide. Its structure corresponds to that of an inverse spinel (written as AB2O4) resulting from A tetrahedral and B octahedral Fe units originating a ferrimagnetic order below an anomalously high Curie temperature of about 850 K. The large difference of the corresponding average Fe–O distances is associated with the occurrence of different oxidation states: above Verwey transition TV estimated at about 120 K [50], one observes an average valence value of 2.5þ per Fe ion in B sites, while A sublattice is concerned only with Fe3þ ions. Below TV, magnetite undergoes a sharp first-order metal–insulator transition associated with a decrease of the conductivity estimated at about two orders of magnitude and resulting from a charge ordering mechanism of Fe2þ ions on the B sublattice with the formation of charged (001) planes alternately occupied by ferric and ferrous ions. However, the physical properties that are correlated to the structure are strongly dependent on the total stoichiometry, that is, the electronic structure, and on the localization of vacancies and cationic ferric ions, that is, the magnetic structure. Consequently, RT-MS of ideal large grain size (>50 nm) Fe3O4 is normally fitted with two well-known distinct sextets, one for the Fe3þ ions on the tetrahedral site (Fe3þ)A, and the other one for the Fe2.5þ ions on the octahedral site (Fe2.5þ)B. The rapid hopping of electrons between pairs of Fe2þ and Fe3þ ions occurs only on the octahedral sites and leads to the formation of so-called Fe2.5þ ions [21,51,52]. The unit cell of ideal stoichiometric magnetite contains eight Fe3þ ions on the tetrahedral sites (or A sites, called Fe3þ A ) and 16 ions 3þ ) and the other are Fe (called Fe3þ on the octahedral sites (or B sites), eight of which are Fe2þ (called Fe2þ B B ) [22]. Magnetites with low particle sizes can also be fitted with two broad sextets, and sometimes it is necessary to introduce a broad doublet ascribed to the presence of superparamagnetic magnetite particles [53]. Some authors [54] have proposed three sextets instead of two, due to two possible directions of the electric field gradient principal axis with respect to that of the hyperfine field. However, for most purposes, a two-sextet fitting is adequate [21]. On the other hand, RT-MS of Fe3 xO4 can be fitted with three sextets: one for (Fe3þ)A, another for (Fe2.5þ)B, and the third one for (Fe3þ)B, due to unpaired Fe3þ ions on the octahedral site, which are not participating in the electron delocalization processes. However, because of the very similar hyperfine parameters for both the (Fe3þ)A and the ossbauer spectrum of Fe3 xO4 practically results in a similar spectrum of two sextets as for Fe3O4, (Fe3þ)B sites, the M€ although with a different subspectral area ratio [21]. Finally, RT-MS of g-Fe2O3 now consists of two Fe3þ sextets coming from the (Fe3þ)A and the (Fe3þ)B sites of this iron spinel phase. The hyperfine parameters for these sextets (S) are very similar to those for the (Fe3þ)A and the (Fe3þ)B sites of Fe3 xO4 (Table 20.1). It has been a long-standing problem to differentiate by RT-MS between nonstoichiometric magnetite (Fe3 xO4) on the one hand and the mixture of stoichiometric magnetite (Fe3O4) with maghemite (g-Fe2O3) on the other hand. In fact, both spectra are similar [21]. This is due to the fact that they exhibit very similar hyperfine parameters. Attempts to fit the MS at other temperatures and/or with applied external magnetic field have been proven to be unsuccessful [55]. Only under very special conditions [56–58], it is possible to differentiate between both cases, for example, mixtures of bulk oxides, because bulk g-Fe2O3 is readily distinguished from bulk Fe3O4 or Fe3 xO4. However, this procedure becomes more difficult when we are dealing with samples exhibiting nanometric sizes, nonstoichiometry, and cation for iron substitutions, as the samples in corrosion products because the first task is to distinguish the static contribution of local disorder and the dynamic one due to the presence of superparamagnetic relaxation phenomena, both giving rise to a nonhomogeneous broadening of M€ ossbauer lines. Nevertheless, in the case of magnetite nanoparticles, the effect of oxidation, if any, can be easily quantified through a core–shell structural model, as the surface/volume ratio is extremely high, contrary to larger particles.

€ 20.3 DETERMINING INHERENT PROPERTIES OF RUST LAYERS BY MOSSBAUER SPECTROMETRY

421

Below the Verwey transition, at about 120 K, the M€ ossbauer spectrum of magnetite can be properly decomposed into five components (Table 20.2), one related to Fe3þ ions on the tetrahedral sites and four corresponding to Fe3þ and Fe2þ ions on two nonequivalent octahedral sites [59,60]. This methodology was successfully used to analyze the MS at 77 K of a magnetite as a corrosion product [61,62].

€ 20.3 DETERMINING INHERENT PROPERTIES OF RUST LAYERS BY MOSSBAUER SPECTROMETRY We mentioned earlier that a careful quantitative characterization of the iron oxides and oxyhydroxides present in the rust layers allows (i) determining the protective ability of the rust layers, (ii) calculating the atomic fraction (a) of iron species coming from the steels that after the corrosion processes form part of the adherent rust, and (iii) understanding the mechanisms associated with the deterioration processes. Let us talk about each aspect a little bit more. To determine the protective characteristics of the rust layers adhered to the steel surfaces, we can use the ternary rust diagram proposed by Hara et al. [4]. This diagram compares the following parameters: (i) the protective ability index PAI ¼ G/(A þ L þ S) and (ii) the ratio (A þ S)/(A þ L þ S), where G, A, L, and S are the relative phase abundances of a-FeOOH, b-FeOOH, g-FeOOH, and spinel phases (magnetite/maghemite), respectively. Hara et al. [4] proposed a criterion for the classification of the rust layers as (i) protective if PAI > 1, (ii) nonprotective but active if PAI < 1 and (A þ S)/(A þ L þ S) > 0.5, and (iii) nonprotective but inactive if PAI < 1 and (A þ S)/(A þ L þ S) < 0.5. Although this criterion was defined for field experiments, it can also be applied in accelerated tests. Next, in order to calculate a, we need to measure the mass of the steel before the corrosion test, mi(S), the mass of the steel after the corrosion test taking into consideration the mass of the adherent rust, mf(S þ AR), and the mass of the steel after the corrosion test but without the rust, that is, the rust completely removed, mf(S). a can be calculated using the following formula [18]: 1 0P ! r j mj gj mf ðS þ ARÞ mf ðSÞ j C BP  (20.1) a¼
A; @  r j mj mi ðSÞ mf ðSÞ gFe ðSÞ j

where gFe(S) is the iron percentage in the steel composition, mj is the total mass of iron phase j, gj is the relative iron content in each iron phase j, and rj represents the relative abundance of each iron phase j in the AR. It is clear that the sum extends over all the iron phases present in the AR. The relative phase abundances, rj, can be determined using the ossbauer fraction fj is assumed to be equal to one, which is subspectral M€ ossbauer area of any iron component, Aj. If the M€ a good approximation particularly at low temperatures, then rj is proportional to nj, and we can compare the relative abundances of the different spectral components. Finally, a deep quantitative characterization of all corrosion products, which includes detailed identification and determination of relative phase abundances and physical properties, allows a better comprehension of possible mechanisms of formation of the iron phases. Taking these aspects into consideration, the potential for application of M€ ossbauer spectrometry to corrosion studies is now demonstrated for three accelerated corrosion tests in chloride environments: total immersion, dry–wet cycles, and outdoor tests. 20.3.1 Rust Layers in Steels Submitted to Total Immersion Tests In the first accelerated test, Colombian carbon steels (CS) and Colombian weathering steels (WS) were exposed for 42 days to total immersion tests in 0.6 M NaCl-containing solutions at room temperature and under continuous aeration [16]. The room temperature MS for the adherent rust of both steels were adjusted with three Lorentzian magnetic sextets and one Lorentzian quadrupolar doublet (Fig. 20.4). The magnetic hyperfine field values for the sextets components of about 49.0, 45.8, and 34.8 T, as well as their respective quadrupole shift values of about 0, 0, and 0.26 mm s 1, point to the presence of nonstoichiometric magnetite (S) and magnetic goethite (G). The presence of small crystallite maghemite coming from magnetite oxidation might be very small or absent. This interpretation is also supported by the quadrupole shift value of about 0.26 mm s 1, which rules out the presence of maghemite. On the other hand, the quadrupolar doublet suggests the presence of a mixture of lepidocrocite (L), akaganeite (A), and possibly superparamagnetic goethite (SPM-G). Next, Fig. 20.5 shows the 77 K MS for the adherent rust of both steels after 42 days

€ SPECTROMETRY TO STUDY THE INHERENT PROPERTIES 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER

Transmission (a.u.)

422

(a) CS-RT

–5

0 Velocity (mm s–1)

5

10

0 Velocity (mm s–1)

5

10

Transmission (a.u.)

–10

(b) WS-RT

–10

–5

FIGURE 20.4 € ssbauer spectra of rust from carbon steel (a) and weathering steel Room temperature Mo (b) submitted to total immersion tests in 0.6 M NaCl concentration for 42 days.

of total immersion. In line with the room temperature MS, the low temperature MS were fitted with groups of Lorentzian profiles, the first group of five sextets was assigned to S, the second group of three sextets (four sextets were also possible) to A, the third group of two sextets with broad lines to goethite G, and the fourth group of a doublet to L and possibly SPM-G. The proportions of each iron-based phase are estimated from the total absorption spectral areas, assuming the same values of recoilless factor for the different iron species. The adherent rust for both steels was found to be composed of L, A, G (minor components), and S (major component). Taking into consideration, the phase abundances and by using equation (20.1), a was estimated at 0.21 and 0.22 for CS and WS, respectively [16]. In a second work, Colombian CS and Japanese WS were submitted to total immersion in 0.05, 0.2, and 0.6 M NaCl-containing solutions during 2, 7, and 14 days, respectively [14,15,52]. The adherent rust for CS was composed of lepidocrocite, goethite, akaganeite, and nonstoichiometric magnetite, the last one being the most abundant phase. These compositions are rather similar to the ones obtained for the Colombian WS in the first work [16]. In contrast, the adherent rust for Japanese WS consisted of lepidocrocite, (the most abundant phase), goethite, akaganeite, and small amounts of maghemite. No magnetite was observed. A possible explanation for these results is the differing composition of the Colombian and Japanese WS. In fact, it was found that magnetite was formed in total immersion tests at 0.05 M  [NaCl]  0.6 M and from 2 to 42 exposure days, in the following range of steel compositions: 0.07  C  0.23; 0.01  Si  0.33; 0.46  Mn  0.72; 0.01  S  0.02; 0.01  P  0.02; 0.01  Ni  0.03; 0.01  Cr  0.48; 0.00  Al  0.40; 0.02  Cu  0.47; and Sb ¼ 0.01. Most alloying elements of Japanese WS in this study are in the mentioned ranges, but only Si (0.47), P (0.11), Ni (0.2), and Cr (0.71) are above these ranges. Therefore, in the higher concentration of these four elements should lay the reason for the different compositions formed in the

423

Transmission (a.u.)

€ 20.3 DETERMINING INHERENT PROPERTIES OF RUST LAYERS BY MOSSBAUER SPECTROMETRY

(a) CS-77K –10

–5

0

5

10

–1

Transmission (a.u.)

Velocity (mm s )

(b) WS-77K –10

–5

0 Velocity (mm s–1)

5

10

FIGURE 20.5 € ssbauer spectra of rust 77 K Mo from carbon steel (a) and weathering steel (b) submitted to total immersion tests in 0.6 M NaCl concentration for 42 days.

two WS. Next, taking into consideration Eq. (20.1), a was found to range between 0.30 and 0.72 for CS and between 0.20 and 0.60 for WS [15]. On the other hand, all PAI values for both steels were below one, which, compared to the high corrosion velocities well above 10 mm per year, are indicative of a nonprotective rust layer. The ratio (A þ S)/(A þ L þ S) showed a very different behavior for CS and WS. All values for the former are higher than 0.5 but below this value for the latter, indicating active and inactive types of rusts, respectively. Based on the work by Hara et al. [4], this fact suggests that the rust layer formed on CS has a low probability to evolve into a protective one in this type of environment, but that formed on WS possibly does. Now, we will concentrate on the analysis of the physical properties of the iron phases as derived from the VT-MS. In the case of magnetite, the room temperature magnetic hyperfine fields range between 48.8 and 49.1 T for the Fe3þ sites, whereas they vary between 45.3 and 46.0 T for the Fe2.5þ sites [16,52]. These values are generally lower than 49 and 46 T, which are the reference values for stoichiometric, pure magnetite. Therefore, the B hyperfine field range associated with the octahedral sites is broader than that of the tetrahedral sites, suggesting a greater sensitivity of the Fe2.5þ sites, which might be associated with the presence of defects, nonstoichiometry, and cation substitutions for iron into the Fe3 xO4 structure. On the other hand, the ratio, R, of the subspectral area of the Fe2.5þ component to the subspectral area of the Fe3þ component cannot be directly related to the oxidation parameter, x, of magnetite (Fe3 xO4) because besides the presence of vacancies, cation for iron substitution is also expected to occur in these phases. In general, the R values range between 0.75 and 1.35 [16,52]. The lower R values, in comparison to R ¼ 1.89, which corresponds to the value for ideal stoichiometric magnetite, indicate the nonstoichiometric character and the cation for iron substitution in all magnetites formed in these tests. In relation to goethite, the room temperature magnetic hyperfine fields are always below 38.0 T, which is the reference value of well-crystallized large-particle goethite [21,22]. The low hyperfine field values may be related to the small grain sizes as well as the alteration of the Fe neighboring due to anion and/or cation substitutions.

424

€ 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER SPECTROMETRY TO STUDY THE INHERENT PROPERTIES

M€ ossbauer spectrometry shows that all iron phases found in the corrosion experiments are normally highly defective, possibly due to multiple cation substitution for iron (in fact, there is a simultaneous dissolution of other alloying elements from the steels, besides iron, such as carbon, manganese, nickel, copper, and chromium), nonstoichiometry, broad distribution of nanometric grain sizes, the possible presence of multianionic substitution for oxygen (in fact, hydroxyl groups and vacancies could replace oxygen in some cases), and other iron phases that are simultaneously present. It is worth mentioning that Kwon et al. [63,64] have estimated realistic atomic scale structures of goethite, lepidocrocite, and akaganeite particles as corrosion products. Their results showed that the linkages of fundamental FeO6 octahedral units in the particles were deviated from the ideal crystal structure, that is, the iron oxyhydroxides were very defective, in good agreement with our results. Taking into consideration the physical properties, we can discuss possible mechanisms of deterioration based on the model by Kimura et al. [65–67]. In the nucleation stage, the dissolved metal (Me) reacts with oxygen and hydroxyl ions to form Me(O,OH)6 and/or MeOx units, thus resulting in the formation of Fe(OH)2 and probably metal-doped Fe(OH)2. For simplicity purposes, we will only use the term Fe(OH)2 to cover these two cases, that is, pure and metal-doped Fe(OH)2. In the growing stage, Me(O,OH)6 units in the Fe(OH)2 phase are transformed to form nanostructured Me(O,OH)6, which is like crystalline g-FeOOH. Finally, in the last stage, which involves adhesion, coagulation and ripening imply the evolution of this structure, depending on the environment and the alloy elements in the steel. A small amount of alloying elements can dramatically modify the evolution of the rust morphology and composition. The reactions involved in the three stages can be considered as the sum of nucleation and grain growth processes. The presence of metal ions can change these processes in such a way that nucleation is favored in comparison to growing, and therefore grains with small size are formed. Now, the effects of chloride ions on increasing the degradation processes can be understood in the following way. The number of positive or negative sites at the surface hydroxyl groups on the formed iron phases is affected by the pH of the solution [9,65–67]. The isoelectric point (iep) is the pH value at which positive and negative sites are present in equal number. In an environment with high chloride content (as in the present case), the pH is lower than the iep for adherent rust dominated with iron phases that predominantly contain Fe3þ ions. Therefore, the following reaction takes place [9]: >Fe–OH þ H3Oþ ! >Fe–OH2þ þ H2O, where >Fe–OH represents a surface site on rust particles. Therefore, the Cl ions are attracted to the surface of the rust grains, promoting further corrosion in wet periods. Finally, the duration of wetness determines the duration of the electrochemical processes or the percentage of the time in which the critical humidity is exceeded [68]. 20.3.2 Rust Layers in Steels Submitted to Dry–Wet Cycles In this test, Colombian WS and CS were exposed for 46 days to different chloride concentrations in dry–wet cycles [17]. The CS coupons were placed into four vessels containing solutions with 0.1, 0.6, 0.005, and 0.01 M NaCl. The WS coupons were placed into vessels with 0.1 and 0.6 M NaCl. The coupons were continuously rotated for a period of about 41 min, 13 min of which the samples were immersed in the solution, and the rest of the time, they were outside the solution at a controlled temperature ranging from 40 to 80  C, the lower temperature close to the solution and the higher close to the heating source [17]. The room temperature MS of the adherent rust were adequately adjusted by introducing two doublets and four sextets with Lorentzian line shapes (Fig. 20.6). The quadrupolar parameters suggest the presence of a mixture of L, SPM-G, and A, while two of the sextets are attributed the presence of S, and the others to magnetic G. Next, the 77 K M€ ossbauer spectra of the AR were adjusted with one quadrupolar doublet and three distributions of sextets, assuming some constraints during the fitting procedure (Fig. 20.7). The quadrupolar doublet is attributed to L while the magnetic components are assigned to A, S, and G with a broad distribution of particle sizes. From these relative abundances, and using Eq. (20.1), it was found that a ranged between 0.55 and 0.90 [17]. It is worth mentioning that at 0.6 M NaCl concentration for 46 days, the corrosion rate for CS and WS were of 3100 mm per year and 3686 mm per year, respectively. For comparison purposes, it is possible to see that the performance of both steels in such aggressive environments of high chloride content is very poor, but it is even three to five times worse than in the case of total immersion tests. These corrosion rates seem to be inversely related to the iron conversion factor, which is lower (about 0.21) in the immersion tests than in the dry–wet tests (about 0.80). The different behaviors should be related to the type of iron phases, their characteristics, and their relative amounts. The values obtained for a were always less than one, suggesting that no protective rust layer was formed in these experiments, in agreement with the high corrosion velocities as reported above. The PAI values for CS range between 0.26 and 0.72, whereas those for WS range between 0.43 and 0.5. These values were always below 1, which, compared to the high corrosion velocities well above 10 mm per year, are indicative of a nonprotective rust layer. On the other hand,

Transmission (a.u.) (a) CS-RT –10

–5

0

5

10

Transmission (a.u.)

Velocity (mm s–1)

FIGURE 20.6

(b) WS-RT –10

–5

0

10

Transmission (a.u.)

Velocity (mm

5 s–1)

€ ssbauer Room temperature Mo spectra of rust from carbon steel (a) and weathering steel (b) submitted to dry–wet cycles in 0.6 M NaCl concentration for 46 days.

(a) Scraped rust -77K –5

0 Velocity (mm s–1)

5

10

Transmission (a.u.)

–10

(b) Hit rust -77K –10

–5

0 Velocity (mm s–1)

5

10

FIGURE 20.7 € ssbauer spectra of 77 K Mo scraped (a) and hit (b) rusts from weathering steel submitted to dry–wet cycles in 0.6 M NaCl concentration for 46 days.

426

€ 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER SPECTROMETRY TO STUDY THE INHERENT PROPERTIES

the ratio (A þ S)/(A þ L þ S) ranged between 0.66 and 0.84 for CS and between 0.62 and 0.89 for WS. All these values are higher than 0.5, indicating an inactive type of rust. Based on the work by Hara et al. [4], this fact suggests that the rust layer formed on both steels has a probability to evolve into a protective one in this type of environment. 20.3.3 Rust Layers in Steels Submitted to Outdoor Tests In this test, CS and WS coupons were submitted to accelerated outdoor test by intermittent spraying, every second day, of a 0.6 M NaCl solution during periods of one and two years at University of Antioquia atmospheric station located at Medellın, Colombia [18]. The adherent rust was composed of L, A, G, and S in about the same amounts. We found that after two years of exposure, a was 0.93 and 0.76 for CS and WS, respectively [18]. On the other hand, all PAI values for CS and WS exposed for 1 and 2 years to OT were less than 1, indicating that a nonprotective rust layer was formed in all cases. The values of ratio (A þ M)/(A þ L þ M) were found above 0.5 in all tests for CS and WS. These results indicated that nonprotective and active rust layers were formed on both steels for 1 and 2 years of exposure time. Thus, chloride ions continuously attack the surface generating high corrosion rates and avoiding the development of a protective layer. According to Hara et al. [4], a PAI > 1 means that more than 50% of rust component is goethite, which is in agreement with a protective layer. Although no significant differences were observed in the PAI values of CS and WS, important differences in the corrosion velocities were detected. The average corrosion velocities for CS and WS after the first year of exposition were estimated at 120 mm per year and 70 mm per year, and for the second year at 240 mm per year and 46 mm per year, respectively. Thus, the corrosion rate increased for CS and decreased for WS. All these values were above 10 mm per year, which is the threshold value proposed by Hara et al. [4] as a criterion for the classification of the rust layers as nonprotective. Moreover, the data suggest that the rust layer formed on both steels has a low probability to evolve into a protective one in this type of environment.

20.4 FINAL REMARKS In this work, the potential for application of M€ ossbauer spectrometry to corrosion studies was demonstrated for three accelerated corrosion tests in chloride environments. This technique allowed retrieving maximum information from the inherent properties of the rust layers. With VT-MS, it was possible to identify and determine the relative iron phase abundances from which three parameters could be calculated: (i) a, (ii) (A þ S)/(A þ L þ S), and (iii) PAI. Moreover, with the physical properties retrieved from the analysis of the hyperfine parameters, it was possible to discuss prospective mechanisms of formation and therefore to contribute to the understanding of the deterioration progress. These studies showed that some corrosion product is lost and/or the conversion of metallic ions into iron oxides may be incomplete, and that the relative iron phase abundances pointed out to a nonprotective and active type of rusts. More work is required on other types and chemistry (composition and abundance of alloying elements) of steels, other environmental conditions, and different exposure times. More efforts are needed to improve the fitting models for nonstoichiometric and substituted iron oxides in rust layers.

ACKNOWLEDGMENTS The financial support given by CODI-Universidad de Antioquia (Sustainability Program for Solid State Group 2011–2012 and IN616CE), COLCIENCIAS (CENM under contract 043–2005) and ECOS NORD are greatly acknowledged.

REFERENCES 1. 2. 3. 4.

C. Leygraf, T. Graedel, Atmospheric Corrosion, Electrochemical Society Series, Wiley-Interscience, New York, 2000. H. Tamura, Corros. Sci. 2008, 50, 1872–1883. T. Kamimura, S. Hara, H. Miyuki, M. Yamashita, H. Uchida, Corros. Sci. 2006, 48, 2799–2812. S. Hara, T. Kamimura, H. Miyuki, M. Yamashita, Corros. Sci. 2007, 49, 1131–1142.

REFERENCES

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46.

47. 48.

427

T. Kamimura, S. Nasu, T. Tazaki, K. Kuzushita, S. Morimoto, Mater. Trans. 2002, 43, 694–703. S. Hoerle, F. Mazaudier, Ph. Dillmann, G. Santarini, Corros. Sci. 2004, 46, 1431. N. Nava, E. Sosa, J.L. Alamilla, C. Knight, A. Contreras, Corros. Sci. 2009, 51, 2652–2656. S.J. Oh, D.C. Cook, H.E. Townsend, Corros. Sci. 1999, 41, 1687–1702. T. Mizoguchi, Y. Ishii, T. Okada, M. Kimura, H. Kihira, Corros. Sci. 2005, 47, 2477–2491. M. Yamashita, H. Nagano, T. Misawa, H.E. Townsend, ISIJ Int. 1998, 38, 285–290. D.C. Cook, Hyperfine Interact. 2004, 153, 61–82. W. Meisel, J. Phys. (Paris) Colloq. 1980, 41, 63–70. A. Vertes, I. Czak o-Nagy, Electrochim. Acta 1989, 34, 721–758. F.R. Perez, C.A. Barrero, O. Arnache, L.C. Sanchez, K.E. Garcıa, A.R. Hight Walker, Physica B 2009, 404, 1347–1353. F.R. Perez, C.A. Barrero, K.E. Garcıa, Corros. Sci. 2010, 52, 2582–2591. K.E. Garcıa, A.L. Morales, C.A. Barrero, J.M. Greneche, Corros. Sci. 2006, 48, 2813–2830. K.E. Garcıa, C.A. Barrero, A.L. Morales, J.M. Greneche, Corros. Sci. 2008, 50, 763–772. C.A. Barrero, K.E. Garcıa, A.L. Morales, J.M. Greneche, Corros. Sci. 2011, 53, 769–775. S.J. Oh, D.C. Cook, H.E. Townsend, Hyperfine Interact. 1998, 112, 59–66. U. Schwertmann, R.M. Cornell, Iron Oxides in the Laboratory, 2nd edn, Wiley VCH, Weinheim, Germany, 2000. R.E. Vandenberghe, C.A. Barrero, G.M. da Costa, E. Van San, E. de Grave, Hyperfine Interact. 2000, 126, 247–259. R.M. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses, 2nd edn, Wiley-VCH, New York 2003. J.F. Marco, M. Gracia, J.R. Gancedo, M.A. Martın-Luengo, G. Joseph, Corros. Sci. 2000, 42, 753–771. M. Stratmann, K. Hoffmann, Corros. Sci. 1989, 29, 1329–1352. T. Okada, Y. Ishii, T. Mizoguchi, I. Tamura, Y. Kobayashi, Y. Takagi, S. Suzuki, H. Kihira, M. Itou, A. Usami, K. Tanabe, K. Masuda, Jpn. J. Appl. Phys. 2000, 39, 3382–3391. J.G. Stevens, A.M. Khasanov, J.W. Miller, H. Pollak, Z. Li, eds., M€ossbauer Mineral Handbook, M€ ossbauer Effect Data Center, The University of North Carolina at Asheville, USA, 2002. D.G. Chambaere, E. de Grave, R.L. Vanleerberghe, R.E. Vandenberghe, Hyperfine Interact. 1984, 20, 249–262. C.A. Barrero, K.E. Garcıa, A.L. Morales, S. Kodjikian, J.M. Greneche, J. Phys. Condens. Matter 2006, 18, 6827–6840. E. Murad, Clay Miner. 1979, 14, 273–283. D.G. Chambaere, E. de Grave, Phys. Status Solidi (a) 1984, 83, 93–102. J.E. Post, V.F. Buchwald, Am. Mineral. 1991, 76, 272–277. K. Stahl, K. Nielsen, J. Jiang, B. Lebech, J.C. Hanson, P. Norby, J. van Lanschot, Corros. Sci. 2003, 45, 2563–2575. J.E. Post, P.J. Heaney, R.B. Von Dreele, J.C. Hanson, Am. Mineral. 2003, 88, 782–788. D. Rezel, J.M.R. Genin, Hyperfine Interact. 1990, 57, 2067–2075. D.G. Chambaere E. de Grave, J. Magn. Magn. Mater. 1984, 44, 349–352. B. Weckler, H.D. Lutz, Eur. J. Solid State Inorg. Chem. 1998, 35, 531–544. J.M.D. Coey, J. Phys. Condens. Matter 1993, 5, 7297–7300. A. Millan, A. Urtizberea, E. Natividad, F. Luis, N.J.O. Silva, F. Palacio, I. Mayoral, M.L.R. Gonzalez, J.M.G. Calbet, P. Lecante, V. Serin, Polymer 2009, 50, 1088–1094. T. Ishikawa, R. Katoh, A. Yasukawa, K. Kandori, T. Nakayama, F. Yuse, Corros. Sci. 2001, 43, 1727–1738. T. Ishikawa, S. Miyamoto, K. Kandori, T. Nakayama, Corros. Sci. 2005, 47, 2510–2520. T. Ishikawa, T. Motoki, R. Katoh, A. Yasukawa, K. Kandori, T. Nakayama, F. Yuse, J. Colloid Interface Sci. 2002, 250, 74–81. T. Kamimura, S. Nasu, T. Segi, T. Tazaki, H. Miyuki, S. Morimoto, T. Kudo, Corros. Sci. 2005, 47, 2531–2542. C. Remazeilles, Ph. Refait, Corros. Sci. 2007, 49, 844–857. C.A. Barrero, K.E. Garcıa, A.L. Morales, J.M. Greneche, Physica B 2007, 389, 88–93. K.E. Garcıa, C.A. Barrero, A.L. Morales, J.M. Greneche, Mater. Chem. Phys. 2008, 112, 120–126. R.V. Morris, G. Klingelh€ ofer, C. Schr€ oder, D.S. Rodionov, A. Yen, D.W. Ming, P.A. de SouzaJr., I. Fleischer, T. Wdowiak, R. Gellert, B. Bernhardt, E.N. Evlanov, B. Zubkov, J. Foh, U. Bonnes, E. Kankeleit, P. G€ utlich, F. Renz, S.W. Squyres, R.E. Arvidson, J. Geophys. Res. 2006, 111, E02S13. C.A. Barrero, Ph.D. Thesis, University of Ghent, Belgium, 1997. C.A. Barrero, R.E. Vandenberghe, E. de Grave, Hyperfine Interact. 1999, 122, 39–46.

428

€ SPECTROMETRY TO STUDY THE INHERENT PROPERTIES 20 ENHANCING THE POSSIBILITIES OF 57Fe MOSSBAUER

49. C.A. Barrero, J.D. Betancur, J.M. Greneche, G.F. Goya, T.S. Berqu o, Geophys. J. Int. 2006, 164, 331–339. 50. E.J.W. Verwey, Nature 1939, 144, 327. 51. R.E. Vandenberghe E. de Grave, M€ossbauer Spectroscopy Applied to Inorganic Chemistry, G.J. Long, F. Grandjean, eds., Plenum Press, New York, 1989, pp. 59–182. 52. F.R. Perez, C.A. Barrero, A.R. Hight Walker, K.E. Garcıa, K. Nomura, Mater. Chem. Phys. 2009, 117, 214–223. 53. J.P. Urquijo, M.Sc. Thesis, Faculty of Exact and Natural Sciences, University of Antioquia, Colombia, September 2007 (in Spanish). 54. L. H€aggstr€ om, H. Annersten, T. Erickson, R. W€appling, W. Karner, S. Bjarman, Hyperfine Interact. 1977, 5, 201–214. 55. G.M. da Costa, E. de Grave, P.M.A. de Bakker, R.E. Vandenberghe, Clays Clay Miner. 1995, 43, 656–668. 56. T.J. Daou, S. Begin-Colin, J.M. Greneche, F. Thomas, A. Derory, P. Bernhardt, P. Legare, G. Pourroy, Chem. Mater. 2007, 19, 4494– 4505. 57. T.J. Daou, G. Pourroy, S. Begin-Colin, J.M. Greneche, C. Ulhaq-Bouillet, P. Legare, P. Bernhardt, C. Leuvrey, G. Rogez, Chem. Mater. 2006, 18, 4399–4404. 58. T.J. Daou, J.M. Greneche, G. Pourroy, S. Buathong, A. Derory, C. Ulhaq-Bouillet, B. Donnio, D. Guillon, S. Begin-Colin, Chem. Mater. 2008, 20, 5869–5875. 59. F.J. Berry, S. Skinner, M.F. Thomas, J. Phys. Condens. Matter 1998, 10, 215–220. 60. A.C. Doriguetto, N.G. Fernandes, A.I.C. Persiano, E. Nunes Filho, J.M. Greneche, J.D. Fabris, Phys. Chem. Miner. 2003, 30, 249–255. orquez, G. Perez-Alcazar, Hyperfine Interact. 2003, 148/149, 211–217. 61. C.A. Barrero, J.F. Rıos, A.L. Morales, A. Boh 62. C.A. Barrero, L.M. Ocampo, C.E. Arroyave, Corros. Sci. 2001, 43, 1003–1018. 63. S.-K. Kwon, S. Suzuki, M. Saito, Y. Waseda, Corros. Sci. 2006, 48, 3675–3691. 64. S.K. Kwon, S. Suzuki, M. Saito, T. Kamimura, H. Miyuki, Y. Waseda, Corros. Sci. 2006, 48, 1571–1584. 65. M. Kimura, H. Kihira, N. Ohta, M. Hashimoto, T. Senuma, Corros. Sci. 2005, 47, 2499–2509. 66. M. Kimura, T. Suzuki, H. Shigesato, H. Kihira, S. Suzuki, ISIJ Int. 2002, 42, 1534–1540. 67. M. Kimura, H. Kihira, Nippon Steel Technical Report No. 91, January 2005, pp. 86–90. 68. C. Guedes Soares, Y. Garbatov, A. Zayed, G. Wang, Corros. Sci. 2009, 51, 2014–2026.

C H A P T E R 2 1

€ APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

LAKSHMI NAMBAKKAT Department of Physics, University College of Science, Mohanlal Sukhadia University, Udaipur, Rajasthan, India

21.1 INTRODUCTION The study of magnetism and magnetic properties of materials has fascinated scientists ever since attraction between iron and lodestone (which is a naturally occurring magnetic material—magnetite) was noted by the Greek philosopher Thales of Miletus in sixth century B.C. The need to understand many physical phenomena ranging from the astronomical to the subatomic and the wide range of applications possible have helped in sustaining interest in magnetism over centuries. In recent times, requirements for miniaturized, low-cost, versatile devices with indispensable magnetic components have grown tremendously. To meet this ever-growing demand for functional materials with enhanced physical, chemical, and magnetic properties, many new materials—especially in the nanoregime—have been developed. At the nanolevel, materials exhibit special properties such as the formation of stable alloys using metals that are not miscible in the bulk [1], novel heterostructures capable of multifunctionality within the same nanoentity [2], extended solubility [3,4], new optical phenomena [5], and enhanced physical and chemical properties such as strength [6], conductivity, [7] and absorption [8,9] and are also amenable to the “tailoring” of these properties by a suitable choice of process parameters, size, and so on. In particular, the development of nanosized materials has led to a steep increase in research for developing and perfecting magnetic materials such as permanent magnets, soft magnets, and magnetoresistive and spintronics materials that meet specific requirements. The presence of disorder, vacancies, strains, clustering, granularity, and so on, easily realized in nanostructures, gives rise to a rich variety of magnetic effects and also a means of control. Novel methods of material preparation such as thin film deposition techniques, high-energy ball milling, and rapid solidification are now increasingly being used for the preparation of such materials with tailor-made properties. This growth has been matched with innovative characterization techniques since any material is useless from the technological and fundamental scientific points of view unless its properties are well understood. Conventional characterization techniques such as X-ray diffraction (XRD), neutron diffraction, transmission electron microscopy (TEM), and vibrating sample magnetometry (VSM) are now being supplemented by new techniques such as atomic force microscopy, magnetic force microscopy (MFM), and X-ray photoelectron spectroscopy (XPS) for understanding the fundamental properties of both bulk and nanostructured materials better. In addition to their technological relevance, nanogranular magnetic systems are also ideally suited for basic research necessary to understand the underlying phenomena and also for the investigation of several basic aspects of solid-state physics, such as superparamagnetism, metastable phase formation and its effect on magnetic properties, kinetics of nanocrystallization, and spin glass behavior. Thus, a combination of structural and magnetic characterization techniques is generally used to study M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

429

430

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

magnetic phenomena. M€ ossbauer spectroscopy is a versatile technique that can be used for elucidating these structures at the microscopic level. When it is coupled with bulk magnetization techniques such as DC magnetization, a complete picture of the magnetic interactions present in a material can be obtained since these two techniques probe the magnetic properties of the same material at different ranges of frequencies. This is because conventional magnetization measurements are done at frequencies ranging from about 1 to 1000 Hz while M€ ossbauer spectroscopy examines magnetic dynamics at much higher frequencies, for example, typically between 106 and 1011 Hz for 57 Fe nuclei. This chapter gives a few specific examples of the application of M€ ossbauer spectroscopy in studying mainly three types of magnetic nanostructures, namely, spinel ferrites, Fe–Al alloys, and Fe–Al films.

21.2 SPINEL FERRITES The spinel structure consists of a close packed fcc arrangement of oxygen ions and two interstitial sites with tetrahedral (A site) and octahedral (B site) symmetries. The type of metal ions (cations) on the A and B sites and relative strengths of the intersublattice interactions JAB and intrasublattice interactions JAA and JBB decide the magnetic properties of spinel ferrites. Of these, in most spinel ferrites, JAB is the strongest and JAA is the weakest. Since in most spinels, JAB, JAA, and JBB are also generally negative (i.e., antiferromagnetic), there exists a competition between intra- and intersublattice interactions, often resulting in frustration effects leading to a variety of magnetic properties depending on the site occupancy. The bulk magnetic properties exhibited by spinel ferrites are therefore intimately related to the relative strengths and types of dominant inter- and intrasublattice interactions. When the particle size is reduced to the nanolevel, contributions from disordering, vacancies, lowering of symmetry at surface, and so on also contribute to considerably modifying the trends seen in bulk. It has been generally observed that the accepted site preferences of cations established in bulk spinel ferrites need not necessarily hold in their nanosized counterparts. For example, the presence of Zn at the octahedral B site is nearly never realized in bulk spinel ferrites, whereas there are several reports wherein Zn is present at the B site in nanosized samples [10]. 21.2.1 Microstructure Determination The M€ ossbauer technique proves to be a powerful tool in elucidating both the structure and the magnetic properties as a function of cation substitution in a series where one cation is substituted for another or as a function of size [11]. As an example, in a series of Cr-substituted Co–Zn ferrite—CrxCo0.5 xZn0.5Fe2O4 for x ¼ 0.1–0.5 synthesized by the chemical coprecipitation technique followed by annealing at 300  C for 24 h, particles with sizes ranging from 7–3 nm are obtained. In this system, a systematic exponential reduction in size with increase in Cr content is also observed [12]. The relative areas occupied by Fe at the A and B sites in the M€ ossbauer spectra (Fig. 21.1) recorded at 20 K, which is well below the blocking temperature of all samples in the series, enable the determination of cation distribution in these samples. From the cation distribution, structural parameters such as the mean ionic radius of the tetrahedral A (rA) and octahedral B sites (rB) and oxygen parameter u can be theoretically calculated to provide a better insight into the observed hyperfine parameters. These estimates show that in this series, Cr does occupy the octahedral B site as is the trend in bulk ferrites [13]. In ferrites, if the metal ions occupying the tetrahedral A and octahedral B sites possess a magnetic moment, cooperative (ferrimagnetic) behavior is observed due to the antiferromagnetic A–B superexchange interaction mediated by oxygen ions. In general, since these interactions are indirect and cations at the A sites have cations at the B sites as nearest neighbors (nn) and vice versa, the A–B superexchange is the strongest one. The hyperfine magnetic field at the A/B sites is thus a function of the cation distribution at the nn B/A sites. In nanosized CrxCo0.5 xZn0.5Fe2O4, from Fig. 21.2 it can be seen that the values of the magnetic hyperfine fields at 20 K at both A and B sites are nearly equal for all concentrations of Cr, with a consistent decrease in the value at both sites with increase in Cr concentration. The cation distribution (Table 21.1) shows that Cr substitutes for Co at the B site alone, and since Cr moments order antiferromagnetically with the magnetic moments of Fe ions, substitution of Cr into the B site reduces the magnetization at the B site as a consequence of which the hyperfine field at the A site decreases. However, the cation configuration at the A sites does not show very significant changes across the series, and so the consistent decrease in the value of the hyperfine field at the B site cannot be explained in terms of a change in the configuration of the nearest neighbor. The next nearest neighbor to a B site cation is another B site cation, and so the observed systematic decrease in the hyperfine field at the B site with increase in Cr concentration is because of a stronger B–B interaction in this series. This is also supported by the much larger-than-ideal value of the oxygen parameter

431

21.2 SPINEL FERRITES

Transmission (%)

x = 0.5

x = 0.4

x = 0.3 x = 0. 2

x = 0.1 x = 0 –12

–9

–6

–3 0 3 Velocity (mm s–1)

6

9

12

FIGURE 21.1 € ssbauer spectra at 20 K of Mo CrxCo0.5 xZn0.5Fe2O4. (Reprinted with permission from Ref. 13. Copyright (2007) by the American Physical Society.)

A site B site

510

Hyperfine field (kOe)

500

490

FIGURE 21.2

480

470

460 0.0

0.1

0.2 0.3 Cr concentration

0.4

0.5

Variation of hyperfine fields at the A and B sites in CrxCo0.5 xZn0.5Fe2O4 as a func€ ssbauer spectra tion of x (from Mo at 20 K). (Reprinted with permission from Ref. 13. Copyright (2007) by the American Physical Society.)

u in this series that results in the displacement of oxygen ions in such a way that in the A–B interaction, the distance between A and O ions is increased while that between B and O is decreased, leading to a decrease in the A–A interaction accompanied by an enhancement of the B–B interaction. Often when nanoferrites are synthesized using chemical routes such as the coprecipitation method, which is very popular being quick, easy, and cheap, a-Fe2O3 is formed as an impurity. The quantity can be very small as to be barely TABLE 21.1 Cation Distributions in Spinel Structure of CrxCo05 xZn0.5Fe2O4 [13] x 0.0 0.1 0.2 0.3 0.4 0.5

A site

B site

Zn0.5Fe0.46Co0.04 Zn0.5Fe0.48Co0.02 Zn0.5Fe0.48Co0.02 Zn0.5Fe0.45Co0.05 Zn0.5Fe0.47Co0.03 Zn0.47Fe0.53Co0.00

Cr0.0Co0.46Fe1.54 Cr0.1Co0.38Fe1.52 Cr0.2Co0.3Fe1.52 Cr0.3Co0.2Fe1.55 Cr0.4Co0.1Fe1.53 Cr0.5Zn0.03Fe1.47

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

311

420

222

440

50

422

S4

220

100

400

432

0 100

S3

(I/I0 %)

50

FIGURE 21.3 X-ray diffractograms of Cu0.25Co0.25Zn0.5Fe2O4 prepared at various seeding temperatures (seeding temperature of S1—300, S2—333, S3—353, and S4—333 þ 368 K). a-Fe2O3 peaks have been marked with d. (Reprinted with permission from Ref. 14. Copyright (2009) by the Institute of Physics (IOP) publishing.)

h

100

S2 50

h

100

S1 h

50 h

h

h

h

0 20

40 2θ (degrees)

60

visible in the X-ray spectrum, but owing to its large f-fraction, the presence of a-Fe2O3 is clearly visible in the M€ ossbauer spectrum. However, even the presence of such impurities in small quantities can affect the bulk magnetic properties in the sample. For instance, Fig. 21.3 shows the XRD spectra of four different samples of Cu0.25Co0.25Zn0.5Fe2O4 synthesized using the chemical coprecipitation technique. The difference between the samples labeled S1, S2, S3, and S4 is the difference in the seeding temperature, maintained during the coprecipitation process. The seeding temperature refers to the temperature of the solution in the initial seeding stage when the precipitation of the hydroxides of the different cations occurs. From Fig. 21.3 it can be seen that sample S2 has very little a-Fe2O3 in it, estimated to be 0.7%, while it is about 12% in sample S1 [14]. The room-temperature (RT) M€ ossbauer spectra of these samples (Fig. 21.4) show the presence of a-Fe2O3 much more visibly than in the XRD spectra. Indeed, in sample S1, which has the largest content of a-Fe2O3, the spectrum of a-Fe2O3 is the most visible with the spinel part looking like smeared-out wriggles. The average particle sizes (D) for the samples were calculated using the Scherrer formula: D¼

0:9l ; B cos u

(21.1)

where l is wavelength of X-rays and B corresponds to the width of most intense peak in the XRD spectrum after correction for instrumental broadening. The hysteresis loops of these samples show all of them to be ferrimagnetic at room temperature with a systematic gradation in the magnetic parameters according to the content of a-Fe2O3 in it [14]. The sample with no a-Fe2O3 (S4) has an unusually high magnetic moment of 81 emu g 1. Despite being nanosized (40 nm in size), this magnetic moment is comparable with that of bulk CoFe2O4. This enhancement of the magnetic moment indicates the presence of sublattice spin interactions that are different from those normally observed in spinel ferrites. A TEM of this sample (S4) shows the presence of very small, nearly evenly sized particles of the order of 4 nm adhering to the surface of the larger, 40 nm sized particles (Fig. 21.5). Room-temperature M€ ossbauer spectrum of this sample (Fig. 21.4) is consistent with this observation and shows a superparamagnetic doublet along with two distinct sextets and a relaxation spectrum.

433

21.2 SPINEL FERRITES

100 SP doublet

A site B site

S4

99

relaxation spectral line

Transmission (%)

100 α-Fe203

S3

99

100

S2

99

99

S1 96

–10

–5

0

5

10

Velocity (mm s–1) FIGURE 21.4 € ssbauer spectra showing that a-Fe2O3 is the most in sample S1 and is totally absent in sample S4. A RT Mo superparamagnetic (SP) doublet is additionally present in sample S4 due to 4 nm sized particles that are adhered to the larger ones. (Reprinted with permission from Ref. 14. Copyright (2009) by the Institute of Physics publishing.)

(c)

( a)

1.0

Aver age si z e 39 nm

0.5

Probability

500 nm

0.0

(d) 0

1.0

20

80

60

40

( b) Aver age si z e 4 nm

10 nm

0.5 (e)

0.0 0

1

2

3

4

5

Particle size (nm)

6

7

10 nm

FIGURE 21.5 Particle size distribution and TEM micrographs for sample S4 (seeding temperature 333 þ 368 K). (a) and (b) are histograms for the larger and smaller particles, respectively. (c) is a low-resolution TEM micrograph. (d) and (e) show the smaller, 4 nm particles adhered to the larger one. (Reprinted with permission from Ref. 14. Copyright (2009) by the Institute of Physics (IOP) publishing.)

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

434

€ ssbauer Spectroscopy 21.2.2 Elucidation of Bulk Magnetic Properties in Nanoferrites Using In-Field Mo In bulk ferrites, the sense of magnetization between the A and B sublattices is opposite so that these are antiferromagnetically coupled, leading to an overall ferrimagnetic ordering with net moments depending on the cation distribution. The same holds true in nanosized particles when a magnetic field large enough to saturate is applied. In the absence of spin canting, the hyperfine field at the A site will be parallel and that at the B site will be antiparallel to the applied field [15]. A more complete picture can therefore be obtained using in-field M€ ossbauer spectroscopy in which a strong magnetic field is applied in a direction either along or perpendicular to the g-ray direction. The relative intensities of the second and fifth lines in the sextets are affected by the direction of the applied field. When the field is applied parallel to the g-ray direction, the intensity of the middle (2 and 5) peaks in a sextet becomes zero whereas it becomes four when the field is applied perpendicular to the g-rays [11,16]. Thus, the intensity of the middle peaks in the sextet becomes zero when the alignment of the spins in the sample is parallel or antiparallel to the g-rays, and is maximum, that is, four when the spins are aligned perpendicular to g-ray direction. The intensity ratio of peaks 1 and 3 (as also 6 and 4) is always 3 : 1. The magnetic splitting is determined by the total magnetic field BT at the nucleus, which is given by the vector sum of the hyperfine field, Bhf, and the applied field, BA. Thus, hyperfine field at the nucleus and the average canting angle both can be determined using the method described by Petitt and Forester [15]. The use of in-field M€ ossbauer spectroscopy for determining the hyperfine fields due to 57 Fe at different sublattices sites and the mutual spin orientations can help explain the bulk magnetic properties obtained through DC magnetization. The method is elaborated in the following example. The sample considered is the nanosized, impurity-free Cu0.25Co0.25Zn0.5Fe2O4, synthesized as described in the previous section. It has a bimodal size distribution with 4 nm sized particles adhered to larger, 40 nm sized ones. In-field measurements on this sample were done using a Janis OptiMag (USA) system in which a magnetic field of 5 T was applied along the direction of g-rays. Fig. 21.6 shows the M€ ossbauer spectra obtained at 5 K with and without the application of a magnetic field. The spectra consist of wellresolved sextets alone without the presence of any doublet or relaxation spectrum. The doublet observed in the RT spectrum can, therefore, be conclusively attributed to the 4 nm sized spinel particles, which are superparamagnetic at RT and have been observed in the TEM image of this sample. Initially, the low-temperature (LT) spectrum was fitted with two broad sextets corresponding to hyperfine fields of 56 and 55 T, which define the average hyperfine fields at Fe in the tetrahedral A and octahedral B sites. However, the individual linewidths are much broader than in the case of RT spectra and a careful examination of the LT spectrum shows a definite asymmetry in the peaks. The LT spectra of this sample have, therefore, been refitted for three unique sites corresponding to one A and two B sites arbitrarily designated as B1 and B2. The hyperfine field values for the spectrum without applied magnetic field are 55, 56.2, and 56.6 T (Fig. 21.6a) with

100 A site 98

B2 site

(b)

B1 site

Transmission (% )

96

FIGURE 21.6 € ssbauer spectra of sample S4 Mo at 5 K (a) without any applied external field (b) with an external field of 5 T applied along g-ray direction. (Reprinted with permission from Ref. 14. Copyright (2009) by the Institute of Physics (IOP) publishing.)

94

100 A site

(a)

B2 site B1 site

98

96

94 –15

–10

–5

0 Velocity (mm s–1)

5

10

15

435

21.2 SPINEL FERRITES

isomer shift values of 0.353, 0.49, and 0.58 mm s 1 with respect to metallic Fe for the A, B1, and B2 sites, respectively. It has generally been observed that in spinel ferrites, the sextet due to Fe at the B site is much broader than that at the A site and is due to the possibility of more than one B site because of the random distribution of cations over the A site [17]. In the present system, four cations—iron, copper, cobalt, and zinc are present, and so at any one particular site, the distribution of these cations is in a random fashion, leading to a probability of finding more than one unique A or B sites. In ferrites, the arrangement of cations at the A site affects the hyperfine fields at the B site and vice versa since the magnetic interaction between these sites is of the superexchange type, mediated by oxygen ions. Although such a random distribution of cations can occur at both sites, the B site is more affected by the difference that occurs in the cation distribution at the A site. This is so because the B site has six A sites as nearest neighbors whereas the A site has 12 B sites as nearest neighbors and so the possibility of formation of more than one distinct A site is 1/12th that of the B site [18]. On applying an external magnetic field of 5 T in the direction of g-rays, the hyperfine field at the A site increases to 60 T while those at the B1/B2 sites decrease to 51 and 52 T, respectively (Fig. 21.6b). The middle peaks (i.e, the second and fifth) do not disappear completely and, on the basis of isomer shift and hyperfine field values obtained in the RT spectrum, can be assigned to sextet B2. This shows that spins at A and B1 are collinear, while those at B2 are canted. The canting angle uB is estimated to be 34 with respect to g-ray direction and is calculated from the integrated intensity of the middle peaks (second and fifth) using the following expression: A2;5 4sin2 uB ; ¼ A1;6 3ð1 þ cos2 uB Þ

(21.2)

where, A2,5 is the area of the second or fifth peak and A1,6 is the area of the first or sixth peak in the M€ ossbauer spectrum. Using in-field M€ ossbauer spectroscopy, the spins on the sublattices A and B1 are conclusively shown to be collinear and aligned along and against the applied field, respectively, while those on B2 are canted at an angle of 34 . But since the hyperfine fields at both B1 and B2 sites are reduced on application of an external field, this would imply that the spins at the B2 sites are also aligned opposite to the magnetic field, albeit canted and thus not parallel to it, making the B–B interaction ferromagnetic. The in-field M€ ossbauer study thus clearly establishes the nature of exchange interactions among the various sublattices in sample S4. Accordingly, it can be concluded that while the A–B interaction is antiferromagnetic as is generally the case in ferrites and since the hyperfine field value of both the B sites decreases on application of an external magnetic field, the B–B interaction is ferromagnetic. This is rather unusual since, in general, the JBB interaction is negative, that is, antiferromagnetic, in spinel ferrites. The reason for both the enhanced and ferromagnetic B–B interactions observed in the Cr- and Cu-substituted Co–Zn ferrites is largely due to the fact that both Cr and Cu are Jahn–Teller cations [19,20], which causes a distortion especially when present at the octahedral sites. This is reflected in the unusually large values of the oxygen parameter u— estimated from M€ ossbauer measurements, in both types of samples. In contrast, in a nanosized spinel ferrite of nearly the same size with a composition Ni0.65Zn0.35Fe2O4, and synthesized by the sol–gel technique, Kumar et al. [21] have found a synchronization of surface and core magnetization contributions verified by in-field M€ ossbauer spectroscopy that has been used to determine the cation distribution and also to understand the saturation magnetization value of 80 emu g 1, which is larger than that in the corresponding bulk. Although their sample is nanosized (42 nm), the core in this case is shown to be large enough to mimic the effects seen in the bulk. Moreover, in this sample, magnetization contributions of the core and the surface spins are strongly coupled, which results in a magnetic moment that is 11% higher than in the corresponding bulk counterpart. Depending on the type of measurements made, the same system can show what appear to be seemingly contradictory results. For example, in a study of the magnetic properties in ball-milled Ni0.35Zn0.65Fe2O4, Ghosh et al. [22] have observed a cluster glasslike freezing even below the characteristic temperature Tf of 60 K of the sample. However, low-temperature M€ ossbauer studies (both with and without applied magnetic field) clearly show the sample to be ferrimagnetically ordered at 10 K. Thus, using this combination of DC magnetization, which measures the magnetic moment of the sample as a whole, and M€ ossbauer spectroscopy, which is sensitive to more short-ranged effects and therefore senses the local fluctuations in the magnetic hyperfine field, they conclude that there are ferrimagnetically ordered clusters within the grains of the nanosized samples. However, at low temperatures, a combination of spin frustration and competing sublattice interactions results in randomly oriented freezing of local magnetic moment of the clusters, making the system spin glasslike. The other possibility of the presence of a wide distribution of particle size and, in turn, a corresponding distribution in the blocking temperatures has also been categorically ruled out on the basis of the fully resolved sextets obtained at low temperatures in the M€ ossbauer spectrum.

436

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

21.2.3 Core–Shell Effect on the Magnetic Properties in Superparamagnetic Nanosystems In the nanoregime, contribution to the overall magnetic behavior from surface spins is considerable—increasing with decrease in the particle size to the extent of more than 50% of the spins being at the surface for sizes of 5 nm and less. Thus, in nanosized spinel ferrites, although on an average the cation distribution between available sites may be the same throughout the particle, two distinct magnetic phases exist due to spins on the surface and those in the core. Isolated noninteracting magnetic particles can be approximated to single large isolated spins described by superparamagnetism, which only hold good in very dilute systems. However, in real systems, the interparticle interactions become significant, leading to enhanced coercivity and exchange anisotropy [23]. As a result, it is possible to have a variety of magnetic configurations in such systems due to competing energy terms under different conditions of applied fields and temperature. As pointed out in the previous section, M€ ossbauer measurements in high external fields give information about the sense of sublattice magnetization and also enable the separation of core and shell contributions in ultrafine particles. To observe the individual contributions, a combination of methods such as DC magnetization studies that monitor the overall contributions and microscopic observations that are sensitive to much shorter range effects are necessary. It has been pointed out that spin canting of the surface layer in nanosized ferrimagnetic materials, which results in exchange interactions between the core and shell, is not easily realized in metallic ferromagnetic particles [24]. An equivalent mechanism in the case of ball-milled nanocrystalline Fe is the spin glasslike freezing at grain boundaries [25]. This indicates that the noncollinear arrangement of particle spins leading to spin canting originates from magnetic frustration of antiferromagnetically competing sublattices and mostly occurs at the surface. For this reason, a number of studies utilizing in-field M€ ossbauer spectroscopy have been done on nanosized spinel ferrites, which are ideal candidates for development/understanding of systems based on the core–shell exchange bias property [22,26–28]. In this section, studies on very fine particles of Ni-substituted Co–Zn ferrite using a combination of in-field M€ ossbauer spectroscopy coupled with DC magnetometry are used as an example of separating out core–shell contributions to the observed magnetic properties [29]. Single-phased nanosized particles of Ni0.25Co0.25Zn0.5Fe2O4 with an average size of 3 nm were prepared using the chemical coprecipitation technique. Room-temperature M€ ossbauer spectra of the sample show the two doublet pattern typical of superparamagnetic particles. M€ ossbauer spectra for this sample have been recorded at low temperatures under four different conditions—with no field applied (at 10 K), and field applied after cooling to 5 K with field directions parallel (ZFCparallel) and perpendicular (ZFCperpendicular) to g-rays. A fourth spectrum was obtained after cooling the sample down to 5 K with an applied field of 5 T, then recording with an applied field of 5 T perpendicular to g-ray direction (FCperpendicular) (Fig. 21.7). The spectrum obtained at zero applied field (Fig. 21.7a) is resolved into two sextets with average hyperfine fields of 52 and 49 T corresponding to Fe at the tetrahedral A and B sites, respectively. Cation distribution at the tetrahedral A and octahedral B sites obtained from the M€ ossbauer data on the basis of the area of the fitted sextets is (Zn0.5Fe0.5)A[Ni0.25Co0.25Fe1.5O4]B. Fitted parameters for core and shell are given in Table 21.2. Canting angles calculated from M€ ossbauer parameters for shell are given in Table 21.3. The experimental data for all three in-field M€ ossbauer spectra for nanosized Ni0.25Co0.25Zn0.5Fe2O4 can be satisfactorily fitted with four sextets. Since the overall relative transmission is only about 6%, to obtain physically realistic values for individual line intensities, the relative intensity x of the middle peaks (lines 2 and 5 of a sextet) with respect to the inner peaks, 3:x:1:1:x:3, has been constrained, as is necessary in many cases [23]. In all three spectra, two sets of sextets have a value of 2 for x, irrespective of the field direction. The other two sextets have a value of 4 for x when the field is applied perpendicular to the direction of g-rays and 0 when the field direction is parallel to g-rays. Comparing the relative intensities of the sextets and also the hyperfine fields, pairs of sextets in which the middle peak intensity is 0 or 4 also show a field shift exactly equal to double the applied field, less the demagnetization field. The total difference between the two sextets is therefore equal to 8 T. The observed value of contraction/expansion is 4 T, implying a demagnetizing field of 1 T [30]. The other pair has x values equal to 2, indicative of random spin arrangement corresponding to an average angle of 54 with respect to the applied field direction. The hyperfine field values are nearly the same as that obtained with no applied field. Interestingly, there is no significant difference in the obtained parameters of the ZFCperpendicular and FCperpendicular cases. M–H loop recorded at RT for this sample shows it to be superparamagnetic, as expected for particles of 3 nm size as in the present sample. The blocking temperature TB has been determined to be 123  1 K from the ZFC curve (Fig. 21.8b) recorded at a field of 50 Oe. The sextets that show a collinear arrangements of spins, that is, those with x ¼ 0 or 4 on application of a field parallel to the g-ray direction, can thus be assigned to the core and the other pair with x ¼ 2, to Fe at the surface. These results, particularly the indistinguishibility of FCperpendicular and ZFCperpendicular, show that the dominant contribution to the magnetization is from the shell and that the system can be compared to a model consisting of a well-

437

21.2 SPINEL FERRITES

0.99 0.96

(d) FC 5T perpendicular

5K

Transmission

1.00 0.98

(c)

ZFC 5T perpendicular

5K

ZFC 5T longitudinal

5K

No field

10 K

0.96 0.99 0.96

(b)

1.02 0.99 0.96

(a)

0.93 –10

–5

0 Velocity (mm s–1)

5

10

FIGURE 21.7 € ssbauer spectra with (a) no applied field (b) field applied along the g-ray direction (c) field Low-temperature Mo applied perpendicular to g-ray direction after cooling down to 5 K with no applied field (d) field applied perpendicular to g-ray direction after cooling down to 5 K in the presence of 5 T field. (Reprinted with permission from Ref. 29. Copyright (2009) by the American Physical Society.)

ordered core and a surface layer of canted spins in which, while the applied field is sufficient to order the core, it has no effect on the shell. In this model, the thickness of the surface layer t can be calculated from the particle size d using the following relation [16]: V canted ¼ 1


2

2t d

3

;

(21.3)

where Vcanted is the volume fraction of canted spins obtained from the fit. The thickness t of the surface layer is calculated to be 0.59 nm in the present system from M€ ossbauer studies, implying that almost 70% of Fe lies at the surface. Thus, the € ssbauer Fitting Using Core–Shell Model for Various Measurements TABLE 21.2 Parameters Obtained from Mo (Field Applied 5 T) [29] Type of Measurement Core Parameters ZFC (perpendicular) FC (perpendicular) ZFC (longitudinal) Shell Parameters ZFC (perpendicular) FC (perpendicular) ZFC (longitudinal)

Hyperfine Field (T) 0.03

Isomer Shift (mm s 1) 0.05

Width (mm s 1) 0.05

A

B

A

B

A

B

54.9 54.22 53.99

46.1 46 47.25

0.27 0.276 0.271

0.308 0.326 0.356

0.452 0.452 0.55

0.55 0.55 0.65

52.9 51.8 51.5

49.0 48.5 48.9

0.25 0.29 0.29

0.28 0.32 0.37

0.51 0.45 0.55

0.62 0.72 0.89

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

438

TABLE 21.3 Canting Effects from Shell for Various Measurements [29] u (Degree)a

f (Degree)b

Hhf (T)

Type of Measurement

A

B

A

B

A

B

ZFC (perpendicular) FC (perpendicular) ZFC (longitudinal)

54.7 54.7 58.7

54.7 54.7 59.9

50.2 48.8 49.1

52.0 51.6 51.6

59.4 59.5 63.7

50.2 50.2 55.2

a

Angle between gamma ray direction and the direction of total magnetic field. Angle between applied field and the magnetic hyperfine field.

b

core can be described by Neel’s model in which the spins at the A and B sites are collinear but aligned in opposite directions, whereas fields as large as 5 T do not affect the surface spins. Importantly, on the basis of area ratios of individual set of sextets, the average cation distribution of the core and shell is also obtained to be the same, showing that the average chemical composition at the core and shell is the same. From the value of TB and particle size V, the anisotropy energy K, the energy barrier DE for rotation on applying a field H, and anisotropy field Hk can be estimated using the following expressions: [31] 25k B T ; V
 H 2 ; DE ¼ KV 1 Hk K¼

(21.4) (21.5)

2K is the anisotropy field, and Ms is the saturation magnetization. Accordingly, Ms calculating these values using Eqs. (21.4) and (21.5), K is 1.08  107 erg cm 3, Hk is about 22 T, and DE at an applied field H of 5 T, used in the in-field M€ ossbauer studies, is about 17  10 14 erg. The large anisotropy fields in such extremely fine nanoparticles are because at these sizes, the particles are single domain in nature. Since fields much larger than 22 T would therefore be required for saturation, the saturation magnetization, Ms, has been estimated by extrapolating the M versus 1/H for 1/H tending to zero. To check for the presence of exchange biasing induced by the highly disordered surface, M–H loops (Fig. 21.9) have been recorded after cooling the sample down to 20 K in the presence of different values of applied fields termed as cooling field (CF). The M–H loop recorded at 20 K with zero applied field is closed and symmetric about both axes (Fig. 21.8a), while those recorded after field cooling are open due to the irreversibility caused by field cooling. where kB is the Boltzman constant, Hk ¼

(a) 10

(b)

FC ZFC 50 Oe

0.35

ZFC 20K

Moment (emu g–1)

Moment (emu g–1)

0.30 5

0

0.25

0.20

0.15

–5

FIGURE 21.8 (a) ZFC M–H curve at 20 K. (b) FC, ZFC measurements at applied field of 50 Oe. (Reprinted with permission from Ref. 29. Copyright (2009) by the American Physical Society.)

0.10 –10

0.05

0.00 –10000 –5000

0

5000 10000

Applied Field (Oe)

0

50

100 150 200 250 300

Temperature (K)

439

21.2 SPINEL FERRITES

10

10

0

0 CF 4000 Oe

Moment (emu g–1)

–10

10

10

0

0 CF 3000 Oe

–10

CF 7000 Oe

–10

10

10 0

0

–10

–10

CF 2000 Oe

10

CF 6000 Oe

10

0

0

–10 –10000

CF 10000 Oe

–10

CF 1000 Oe

–5000

0

–10

5000

CF 5000 Oe

10000 –10000

–5000

0

5000

10000

Applied Field (Oe)

FIGURE 21.9 M–H loops at 20 K recorded for various cooling fields. (Reprinted with permission from Ref. 29. Copyright (2009) by the American Physical Society.)

A systematic shift in the loop to the left is apparent on field cooling and originates from the phenomenon of exchange bias [32–35] associated with the exchange anisotropy created at the interface between antiferromagnetic (AFM) and ferromagnetic (FM) phases as first observed in nanosized oxidized Co nanoparticles [32]. Exchange bias is usually observed when a material having FM/AFM interfaces is cooled in the presence of a static magnetic field, through the Neel temperature (TN) of the AFM phase (with Curie temperature, Tc, of the FM phase larger than TN). After field cooling, the hysteresis loop of the FM–AFM system is shifted through a value called exchange bias field, HE, along the field axis generally in a direction opposite (negative) to the cooling field. The exchange bias field is usually taken as the shift of the center of the hysteresis loop along the field axis and the coercivity as the halfwidth of the loop. In the present system, exchange bias originates from the coupling between the ferromagnetic core and the disordered (antiferromagnetic) surface. The exchange bias field and coercivity can be obtained from the switching field on the left (HSL) and right (HSR) branches of the loop using the following equations [36]: HE ¼ HC ¼

H SR þ H SL 2 H SR

H SL 2

(21.6) (21.7)

HSR and HSL are those points at which the M–H loop cuts the field axis. To further confirm the nature of magnetic interactions, the isothermal remanence magnetization (IRM) curve Mr(H) and DC demagnetization (DCD) curve Md(H), both normalized to the saturation remanence, have been plotted at 20 K. From these curves, a Henkel plot has been obtained (Fig. 21.10a). In a perfectly noninteracting system of single-domain particles, the slope of the Henkel plot should be 2 and the experimental points would lie on the upper limit of this plot [37]. The types of interactions are characterized by means of dM plots (Fig. 21.10b) obtained using the following relation: dM ¼ M d ðHÞ

1 þ 2M r ðHÞ

(21.8)

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

440

(b) 0.00 –0.04

δM

–0.08 –0.12 –0.16 (a)

0

2500

5000

7500

10000

Applied field (Oe)

0.8

FIGURE 21.10 (a) Henkel plot and (b) dM plots obtained from remanence (IRM and DCD) curves. (Reprinted with permission from Ref. 29. Copyright (2009) by the American Physical Society.)

Md

0.4 0.0 –0.4 –0.8 0.0

0.2

0.4

0.6

0.8

1.0

Mr

Positive values of dM imply an exchange type of interaction that results in a magnetized state while a negative one indicates a demagnetized state attributable to the presence of dipolar type of interactions. The totally noninteracting nature of the ensemble is evident from the Henkel plot (Fig. 21.10a) of experimentally obtained points after remanence measurements (IRM and DCD), which has a slope of 1.82 and which almost coincides with the theoretical upper limit (Fig. 21.10a). The exchange coupling interactions between the core and shell are evaluated using dM plots (Fig. 21.10b). There is a shallow initial negative peak that then tends to zero after about 0.3 T. The negative peak is indicative of magnetostatic interactions that are demagnetizing like and favor antiferromagnetic coupling between the core and shell spins. There is no further change on application of larger fields, since, as discussed earlier, core alignment is complete and core–shell spin locking through antiferromagnetic coupling is strong. Results of field-cooled M–H curves together with those from remanence curves give a clear picture of the interaction mechanisms within each particle and those in the ensemble as a whole. Plots of variation of HC and HE (Fig. 21.11a and b) for different fields applied while cooling can be divided into two distinct portions—for CF—from 0 to 0.3 T and from 0.3 to 1 T. The value of the coercivity HC at an applied field of 0.1 T is slightly larger than its value after zero-field cooling. HE also first increases upto 0.3 T and then decreases. These results can be explained as follows: On cooling down to 20 K in the presence of a field of 0.1 T, most of the core spins get aligned. The spin frustration at the surface due to disordering or broken/dangling bonds leads to spin canting so that the spins at the surface are, on an average, randomized to attain a minimum energy configuration. This leads to a system in which the spins at the surface are effectively locked, leaving the surface unaffected on application of a field. The initial decrease in HC after field cooling on increasing the cooling field, CF, is because of magnetic softening induced in the system due to large applied fields. This is also reflected in the dM plot, which is demagnetizing like in this region. Constancy in the value of HC beyond a CF of 0.2 T clearly indicates that only the core is affected on application of field and that once the spins loosely bound to the surface have also aligned, further increase in the field has no effect on the system as evidenced by the very large anisotropy field Hk estimated for the present samples. The trend in HE (Fig. 21.11b) shows that for CF more than 0.3 T, with increase in the CF, the magnetic coupling between the field and the surface moment also increases, although the applied field is not sufficient to rotate the surface spins to align along the field direction. However, increasing the cooling field also leads to an increase in the energy at the core–shell interface or in other words, the system freezes through a configuration in which the core–shell energy is not minimized. The net effect is a reduction in HE, as observed in the present study and as reported by Bianco et al., in their study of Fe nanoparticles embedded in Fe oxide matrix [38]. Correspondingly, the value of remanence, þMr, increases from its value of 12 emu g 1 at zero applied field to 15 emu g 1, which does not change with further increase in the

441

21.3 NANOSIZED Fe–Al ALLOYS SYNTHESIZED BY HIGH-ENERGY BALL MILLING

HC (Oe)

(a)

2100

2000

1900

1800 0

3000

6000

9000

0

3000

6000

9000

(b)

HE (Oe)

360 240 120 0

(c) Mr (emu g–1)

5.0 4.5

FIGURE 21.11

4.0 3.5

0

3000

6000

Cooling field (Oe)

9000

(a) Coercive field (b) exchange bias field and (c) remanence as a function of the cooling field. (Reprinted with permission from Ref. 29. Copyright (2009) by the American Physical Society.)

applied field (Fig. 21.11c). Together, these results show that on application of field, the core spins get aligned while the surface is unaffected. The combination of in-field M€ ossbauer measurements and DC magnetization studies carried out in this sample clearly enables differentiation of core and shell contributions to the magnetic interactions. M€ ossbauer studies show that due to the very small sizes, most of the spins are at the surface and so the magnetic properties are dominated by surface spins that are nearly impervious to fields of the order of 5 T, as applied in the present studies. The system can be modeled as an ordered core with conventional collinear arrangement of spins at the A and B sites and a canted, highly frustrated surface, which thus constitute two separate magnetic phases, although the nominal chemical composition is the same. The effects observed are comparable to those obtained in nanogranular systems comprising ferromagnetic particles embedded in an antiferromagnetic matrix. The large volume fraction of surface spins completely isolates the cores so that the entire ensemble behaves as a system of nearly perfectly noninteracting particles. DC magnetization studies also establish the fairly strong exchange bias that exists between the core and shell.

21.3 NANOSIZED Fe–Al ALLOYS SYNTHESIZED BY HIGH-ENERGY BALL MILLING Intermetallic iron–aluminum (Fe–Al) systems are of particular interest because of their potential applications as structural and magnetic materials since these intermetallics generally possess high melting points, hardness, and, above all, excellent corrosion resistance even under extreme conditions of temperature/atmosphere. Also, it is well known that the mechanical and magnetic properties of these systems are intimately related to the microstructure and the degree of

442

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

disorder in the system, and that the Al content also influences the magnetic properties of these systems. Due to these reasons, bulk and nanocrystalline Fe–Al systems have been studied extensively and continue to be an important topic of research. In this section, the role of M€ ossbauer spectroscopy in elucidating the magnetic properties in nanosized Fe–Al alloys synthesized by either mechanical alloying (MA) or mechanical milling (MM) is discussed. MA is a nonequilibrium process that has become popular in recent years for the synthesis of nanostructured alloys through solid-state reactions. This process is now being widely used for the preparation of extended solid solutions, alloys of immiscible elements, quasicrystalline and amorphous phases, and different types of compounds and composites. The synthesized materials most often have nonequilibrium structures with a typical average size of 10 nm in MA materials. The large densities of constituent defects such as grain boundaries, interfaces, interphases, surfaces, antiphase boundaries, and antisite atoms lead to an observable broadening of the spectral lines in the M€ ossbauer spectra. The introduction of defects during milling, or their removal by annealing, may be conveniently followed quantitatively by M€ ossbauer spectrometry from the relative spectral areas. In the case of magnetic systems, it is most often possible to extract their hyperfine magnetic field distributions P(B), from the M€ ossbauer spectrum. Then, P(B)DB would give the fraction of probe atoms whose field is between B and B þ DB. However, it is generally difficult to interpret these hyperfine field distributions in detail because of the complexity of defect zones and in view of the perturbation they produce in surrounding grains. Magnetic relaxation phenomena, which are related to the nanocrystalline nature of the grains, may also alter the spectral shapes. Thus, care has to be taken while interpreting the M€ ossbauer spectra and hyperfine field distributions of mechanically alloyed nanostructured materials. However, a careful analysis of the hyperfine fields at Fe sites using M€ ossbauer spectroscopy can yield information about the nearest neighbor configuration of Fe atoms, grain boundaries, and defect structures. This can be correlated with structural and magnetic information obtained from XRD and VSM studies to gain a complete picture of the systematics of alloy formation and its influence on the magnetic properties of iron-containing, mechanically alloyed nanogranular magnetic systems. In this section, the evolution of magnetic properties in a mechanically alloyed system consisting of just 1 at% Fe with Al in which the entire sample becomes ferromagnetic when the grain size approaches the ferromagnetic exchange length in Fe is discussed [39]. 21.3.1 Nanosized Al–1 at% Fe A dilute Fe–Al system consisting of a mixture of pure Fe (95% enriched 57 Fe) and Al powders in the at% ratio 1:99 is chosen for this study [39]. Fe and Al powders are mixed in the required ratio and mechanically milled for various durations in a SPEX 8000 M mixer/mill using tungsten carbide vials and balls with a ball to powder ratio of 20:1. The hyperfine fields are determined using M€ ossbauer spectroscopy and the bulk magnetic properties using a VSM. Interestingly, it is seen that while an unmilled mixture of 1 at% Fe powder in Al does not exhibit bulk room-temperature ferromagnetism, the entire quantity of the powder becomes ferromagnetic, with soft magnetic properties, when the mixture is subjected to MA for 30 min. The average crystallite size of the as-milled sample calculated using the Scherrer formula is about 32 nm for 30 min of milling and reduces exponentially to 12 nm after 1000 min of milling. Since Fe and Al have common, overlapping peaks except for the intense (111) peak of Al, the alloying process is monitored by examining the integrated intensity of this peak. The variation of the percentage of integrated area of the (111) peak (Fig. 21.12) with respect to the total integrated area shows a steady decrease with milling time. The decrease in the intensity of the (111) Al peak can be attributed to the formation of Al–Fe solutions. However, the overlap of many Fe–Al alloy compositions with the more intense peaks of pure Al, combined with the low intensities of the new peaks makes it difficult to identify these phases unambiguously using XRD. The M€ ossbauer spectra of the as-milled samples at room temperature are shown in Fig. 21.13a and the relevant M€ ossbauer parameters are shown in Table 21.4. All the spectra show the presence of a hyperfine sextet and a quadrupole doublet. On fitting for discrete Lorentzians, the width of the magnetic sextets is seen to be large compared with the width of Fe inner peak (0.24 mm s 1), indicating a distribution in the hyperfine fields. The spectra were therefore refitted using Windows program [40] for hyperfine field distribution (Fig. 21.13b). Two subspectral components, corresponding to 33.0 and 14.0 T, are needed to fit the spectra satisfactorily. The ratio between the intensities of the first and third peaks of the sextets has been constrained to be three, while the ratio between the intensities of the second and third peaks, which gives information about the magnetic anisotropy of the sample, has been used as a free parameter in the fitting. Width of the individual peaks has been constrained to be that of the spectrum of Fe foil. The variation in the relative intensities of the second and third peak, obtained from the fitting varies from 3 to 2 as a function of milling time, indicating a decrease in anisotropy. The variation in the relative areas of the paramagnetic and ferromagnetic components (Table 21.4) suggests the formation of a minor solid solution phase initially. This Al-rich metastable phase begins to disappear after 60 min of milling,

443

21.3 NANOSIZED Fe–Al ALLOYS SYNTHESIZED BY HIGH-ENERGY BALL MILLING

Fractional intensity of (111) peak (%)

30

25

20

FIGURE 21.12 Fractional integrated intensity of (111) peak with respect to total integrated intensity in X-ray spectra for unmilled and samples milled for 30, 300, and 1000 min. (Reprinted with permission from Ref. 39. Copyright (2007) by Elsevier.)

15

10 0

200

400 600 Milling time (min)

800

1000

(a)

(b)

100

0.05

95

1000 min 0.00 0.05

100.0 99.5

600 minutes

0.00

600 min

100

0.05 0.00

300 min

96 100

0.05

98

180 min

0.00

96 100

0.05

98

300 minutes

P(H) Arbitrary units

98

Transmission (%)

1000 minutes

180 minutes

120 minutes

0.00

120 min

0.05

100

60 minutes 0.00 0.10

60 min

98 -10

-8

-6

-4

-2

0

2

4

6

8

10

100 99 –10

30 minutes

0.05 0.00

30 min –5

0 Velocity (mm s–1)

5

10

0

10

20

30

40

Field (T)

FIGURE 21.13 € ssbauer spectra of Al 1 at% Fe samples milled for various durations. The dots are (a) Room-temperature Mo experimental data and lines represent the fit. (Reprinted with permission from Ref. 39. Copyright (2007) by Elsevier.) € ssbauer spectra. (Reprinted with permission from Ref. 39. Copyright (2007) by (b) Hyperfine field distribution of Mo Elsevier.)

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

444

€ ssbauer Parameters: Quadrupole Splitting (QS), Isomer Shift (IS), Width of Quadrupole Peaks, TABLE 21.4 Mo € ssbauer Spectrum, and Average Hyperfine Percentage Area Occupied by Quadrupole Peaks in Mo Field for Al–1 at% Fe Powders Ball Milled for Different Times [39] Milling Time (min) 30 60 120 180 300 600 1000

Average Grain Size (nm)

QS (mm s 1)

ISa (mm s 1)

Width (mm s 1)

Area (%)

Average Fieldb (T)

32  4 22  4 17  2 16  3 15  3 13  2 12  2

0.42  0.005 0.58  0.015 0.58  0.015 0.58  0.01 0.52 0.005 0.45  0.01 0.44  0.01

0.18  0.005 0.19  0.005 0.11  0.005 0.12  .005 0.13  0.01 0.11  .005 0.20  0.01

0.24  0.01 0.36  0.01 0.31  0.01 0.3  0.01 0.38  0.015 0.42  0.02 0.83  0.01

21 14 8 8 10 26 51

29.3  2 25.4  2 30.8  1 30.5  2 29.9  1 25.4  2 16.3  3

a

Isomer shift quoted with respect to Fe-foil. From hyperfine field distribution.

b

as indicated by the increase in area of the ferromagnetic component evidencing that the iron that had gone into the formation of the metastable phase is again precipitated out during this time. However, further milling leads to the growth of a stable, nonmagnetic, Al-rich Fe–Al phase, which drastically reduces the magnetic component. The values of quadrupole splitting in all the samples are higher than that reported by Dunlap et al. for their ball-milled Al 2 at% Fe studies [41]. The quadrupole splitting obtained for their sample milled for 14 h is somewhat comparable with the value obtained in the 1000 min milled sample. At intermediate stages (60–300 min of milling) the quadrupole splitting is higher. The higher values of quadrupole splitting after about 60 min of milling are due to the much larger ball to powder ratio used in the present study, resulting in much more defects and strains in the lattice. Thus, a local disorder of this sort would increase the asymmetry of the Fe environment resulting in a larger value of quadrupole splitting. On longer milling, a simultaneous “self-annealing” effect also comes into play due to the high temperatures generated, which accounts for the lowered values of the quadrupole splitting for the 600 and 1000 min milled samples. The linewidths of the quadrupole doublets show a consistent increase with increase in milling time, which can be ascribed to the presence of sites with differing numbers of Fe nearest neighbors, leading to a distribution of the electric field gradients. From the values of the hyperfine field components, the low field peak (average value of 14.0 T) can be assigned to Fe with about 4–6 Al nn, and the main component at 33.0 T to Fe with 8 nn Fe (i.e., a-Fe) in a bcc lattice. Thus, the microscopic picture obtained from XRD and M€ ossbauer studies is that of a complex structure with four main phases— a-Fe, a disordered magnetic phase in which 4–6 atoms in the bcc Fe lattice are replaced by Al, a (nonmagnetic) Fe2Al5-like phase that grows with milling time after 300 min of milling, and the host Al nanocrystalline phase. Correspondingly, the magnetic structure is also expected to be complex. From Fig. 21.14 it is seen that the P(H) value (integrated intensity) in the magnetic hyperfine field distribution is almost constant for the low field component and is maximum for the high field component up to 600 min of milling. However, the intensity of the high field component in the hyperfine field distribution for the 1000 min milled sample

FIGURE 21.14 Variation of integrated P(H) with milling time for the high field component (33.0 T), low field component (14.0 T), and paramagnetic component in the hyperfine field distribution of € ssbauer spectra. the Mo (Reprinted with permission from Ref. 39. Copyright (2007) by Elsevier.)

Integrated P(H) (a.u.)

5

4

33 T

3

2 paramagnetic 1 14 T 0

200

400 600 Milling time (min)

800

1000

445

21.3 NANOSIZED Fe–Al ALLOYS SYNTHESIZED BY HIGH-ENERGY BALL MILLING

shows a decrease of about 60% compared with the 300 min milled sample, accompanied by a corresponding increase in the integrated intensity of the paramagnetic component. In comparison, the relative decrease in the saturation magnetization (Table 21.4) is about 62% for the 300 and 1000 min milled samples. This indicates that in this system, there is a direct correlation between the bcc a-Fe phase and saturation magnetization, leading to the conclusion that the magnetic constituent responsible for bulk magnetization is the bcc a-Fe phase. Although the exact distribution of the bcc Fe grains in the nonmagnetic matrix cannot be predicted from M€ ossbauer measurements alone, a few qualitative observations can be made. Processes such as mechanical alloying that cause severe plastic deformation are known to produce clustering at grain boundaries. It has been reported that magnetic iron atoms in Fe–Al mostly belong to grain boundaries in ball-milled samples [42]. In the present study, after up to 300 min of milling, most of the available iron atoms are in a magnetic environment. In a predominantly nonmagnetic environment (one Fe atom for 100 Al atoms on an average), this is possible only if the iron atoms are clustered together to form magnetic grains. The fact that the unmilled sample is nonmagnetic and the macroscopic magnetic properties improve with milling time up to 300 min suggests that during the initial hours of milling, rather than alloying with Al atoms, the reduction in size of the Fe grains is accompanied by a distribution of Fe grains in the Al matrix. This redistribution of Fe grains leads to a bulk magnetization. On the other hand, M€ ossbauer studies indicate that most of the Fe is in an a-Fe-like environment. Together, these results suggest that a large percentage of the Fe in the Al matrix has gone into grain boundaries. The strong intergranular exchange coupling between the Fe clusters at grain boundaries would then account for the observed spontaneous magnetization, even in the limit of a very dilute magnetic impurity in a nonmagnetic matrix, as in the present study. DC magnetization measurements in these samples show the evolution of bulk magnetic properties as a function of milling time (size). The variation of magnetization with applied field of samples milled for 30, 300, and 1000 min is shown in Fig. 21.15. All the samples show saturation below the maximum applied field of 1.4 T. VSM measurements show that the powders milled for 0, 10, and 20 min are paramagnetic while the samples milled for 30 min or more show soft ferromagnetic properties.

0.3

0.0 1000 min –0.3

Moment (emu g–1)

–1.0

–0.5

0.0

0.5

1.0

3

0

300 min

–3 –1.0

–0.5

0.0

0.5

1.0

3

FIGURE 21.15

0

30 min

–3

–0.5

0.0

Field (T)

0.5

M–H curves at room temperature for the Al 1 at% Fe samples milled for 30, 300, and 1000 min. (Reprinted with permission from Ref. 39. Copyright (2007) by Elsevier.)

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

446

The evolution of the bulk magnetic phase can be understood in terms of the ferromagnetic exchange length. It is known that the microstructure, especially the structural correlation length or the grain size, essentially determines the hysteretic properties of a ferromagnetic material. According to the random anisotropy model, when the grain size is less than the ferromagnetic exchange length, the magneto-crystalline anisotropy is averaged out and the grains align parallel due to exchange interaction [43]. The present system is initially characterized by randomly distributed ferromagnetic grains almost totally isolated from each other by the much larger quantity of a nonmagnetic phase (Al). As discussed above, a collective interaction of these isolated ferromagnetic grains requires that the structural correlation length becomes equal to that of the ferromagnetic exchange length, leading to an observable bulk magnetization. This is made possible when the grain sizes are reduced to the order of the ferromagnetic exchange length. Using the experimentally determined value of A in the equation L0 ¼

rffiffiffiffiffiffi A K1

(21.9)

and substituting the value of K1 to be that for a-Fe, the basic ferromagnetic correlation length L0 for this system is found to be 29 nm, which compares very well with the average grain size of the sample milled for 30 min [44]. At longer milling times, the size of the grains progressively becomes smaller than L0 and the hysteretic properties are improved. The coercivity HC, which is the applied field opposite in direction to the saturating field that is required to reduce the intensity of magnetization to zero, and the reduced remanence (Mr/Ms), which measures the squareness of the hysteresis loop and is related to the level of intergrain interactions, are seen to increase as the grain size decreases (Table 21.5). The steady increase in the squareness factor and coercivity for longer milling times points to an improved intergranular exchange interaction in this nanogranular system. In many multiphase nanocrystalline systems, the presence of paramagnetic phases between the ferromagnetic ones is known to lead to the enhancement of exchange interaction. This results in an enhancement of the coercivity, which would account for the observed trend of HC with grain size in the present series [45]. The combination of M€ ossbauer and DC magnetization studies conclusively shows that the ferromagnetic properties shown by 1 at% Fe mechanically alloyed with Al are purely an effect due to the reduction of size that enables the coupling of Fe moments when the grain size reduces to the order of the ferromagnetic exchange length.

21.4 MAGNETIC THIN FILMS/MULTILAYER SYSTEMS: 57Fe/AI MLS Thin film multilayer systems (MLS) consisting of alternate magnetic and nonmagnetic layers exhibit interesting magnetic phenomena not observed in bulk materials. The reduced dimensionality of these systems, together with the large number of interfacial and surface atoms, leads to surface enhanced magnetism. These artificially structured magnetic materials with reduced dimensionality are widely used in storage and sensor technology and have led to revolutionary changes in many areas of microelectronics. In particular, multilayered films consisting of pure Fe alternating with nonmagnetic layers such as Al exhibit good soft magnetic properties making them excellent candidates for possible use in magnetic recording [46]. These systems have

TABLE 21.5 Variation of Bulk Magnetic Parameters with Grain Size for Al–1 at% Fe Powders Ball Milled for Different Times [39] Milling Time (min) 30 60 120 180 300 600 1000

Average Grain Size (nm) 32  4 22  4 17  2 16  3 15  3 13  2 12  2

TC  5 (K)

HC (mT)

Mr (emu g 1)

Squareness (Mr/Ms)

meff (mB per Fe Atom)

A (pJ m 1)

kKi  104

833 667 642 613 553 528 524

3.4 5.8 7.4 11.2 11.3 13.9 25.7

0.13 0.12 0.29 0.37 0.45 0.09 0.17

0.022 0.026 0.059 0.113 0.117 0.145 0.195

1.3485 1.0846 1.0681 0.9442 0.9166 0.1439 0.2047

3.95 3.16 3.05 2.91 2.62 2.5 2.48

3.86 3.09 2.98 2.85 2.56 2.45 2.42

21.4 MAGNETIC THIN FILMS/MULTILAYER SYSTEMS: 57Fe/AI MLS

447

also been studied to understand, at a fundamental level, phenomena such as surface anisotropy and interlayer coupling [47]. The large difference in atomic radii of Fe and Al together with the different type of lattices of the two elements (bcc for Fe and fcc for Al) results in a complex equilibrium phase diagram. Fe–Al systems are also capable of forming a large variety of metastable nonequilibrium phases that can be achieved by using nonequilibrium preparation techniques such as ball milling and thin film deposition [48,49]. Moreover, the nonequilibrium nature of thin film deposition, attributable to chemical gradients at interfaces and influence of stresses, enables the formation of unconventional structural phases and enhanced solubility [50]. The properties of the resulting film are therefore very sensitive to the methods conditions of deposition, as well as thermal or magnetic annealing [51]. For instance, interfaces are sharper in as-deposited Fe–Al MLS prepared by thermal evaporation or sputtering as compared with those prepared by pulsed laser deposition [52]. During the deposition process itself or on heat treatments, interdiffusion at the interfaces, which can be quite large in an MLS of large multiplicity, leads to the formation of interfacial phases that would comprise a relatively large fraction of the sample and hence would determine the physical properties of the MLS. In multilayer systems containing Fe, if enriched Fe is used, the contribution of the interfaces to the overall magnetic properties can be understood using M€ ossbauer spectroscopy. Although the transmission mode is sometimes used [46], more often, conversion electron M€ ossbauer spectroscopy (CEMS) is used for films/multilayer structures. Taking the example of a 57 Fe=Al MLS, the use of CEMS combined with DC magnetization studies in characterizing the magnetic properties of the MLS is discussed here. 21.4.1 Structural Characterization 



An MLS consisting of a 10 bilayer stack of Fe and Al, that is, [Al (69 A)/57 Fe (26 A)]10 was deposited on Si(100) substrate   using the ion beam sputtering (IBS) technique [53]. Each bilayer consisted of a 26 A thick Fe and a 69 A thick Al layer that correspond to an average atomic composition of 1 : 2 of the deposited multilayer, since Fe and Al atomic densities are 0.85  1023 and 0.60  1023 atoms cm 3, respectively. A beam of argon ions was used to sputter 57 Fe-enriched Fe and Al targets using Kaufman type hot cathode ion source. A base vacuum of 1  10 7 Torr was achieved before deposition. The targets were mounted on rotary motion feedthrough to switch over from Fe to Al in order to deposit alternate layers. Calibration of deposition rate was done by individually sputtering Fe and Al for 10 min, which showed that the growth  rate of Fe and Al are 0.5 and 0.7 A s 1, respectively. The as-deposited multilayers were annealed at 623 K for 1 h and at 723 K for 4 h in a vacuum better than 1  10 6 Torr. Structural characterization of the MLS was done by X-ray reflectivity (XRR) and crosssectional transmission electron microscopy (XTEM). Magnetic properties were investigated by a combination of CEMS at RT and high temperature and RT DC magnetization studies. Verification of layer thickness for the as-deposited MLS was done using XRR measurement (Fig. 21.16a), which gave a bilayer periodicity of 9.3 nm and also enabled calculation of bilayer thickness and interface roughness. This  fitting shows an average individual layer thickness of 29 and 64 A for Fe and Al, respectively. A comparison of the electron density profile (EDP) obtained from fitting the experimental XRR data and EDP calculated with the assumption of sharp interfaces as a function of thickness (Fig. 21.16a inset) clearly shows that a significant amount of intermixing takes place in the as-deposited film. This is because of the large mobility of Al that results in the diffusion of Al into Fe and has also been reported in other studies of Fe–Al multilayer structures [52]. The presence of welldefined Bragg peaks in the XRR of samples annealed at 623 K for 1 h and 723 K for 4 h shows that while considerable mixing has occurred at interfaces in these samples, a layered structure is still present, consisting mostly of intermetallic Fe–Al layers between Al-rich regions. For direct observation and better understanding of the interfacial changes occurring during annealing, comparative XTEM investigations on as-deposited sample and that annealed at 723 K for 4 h were carried out employing imaging and selected area diffraction (SAD) modes. Before imaging, the multilayer interface was first aligned parallel to the electron beam. Fig. 21.17 shows the electron micrographs and SAD patterns of as-deposited and annealed Fe/Al MLS. The insets of Fig. 21.17 show the corresponding low-angle diffraction patterns. The difference in atomic number (Z) of the elements gives rise to the observed layer contrast and helps identify the Fe and Al layers. Since Fe (Z ¼ 26) has a much larger electron scattering form factor than Al (Z ¼ 13), in the bright-field image, lesser contribution from Fe layers than from Al layers makes the Fe layers look darker as compared with Al. While the interfaces between Fe and Al are well defined, there is a waviness along the interfaces due to intermixing of Al and Fe (Fig. 21.17a). The micrograph also shows that the bilayer period is almost constant throughout the MLS structure. All the diffraction rings shown in Fig. 21.17b have been indexed either by bcc Fe or by fcc Al. The reciprocal lattice vectors of Fe (110) (ring marked “2”) and Fe (220) (ring marked “5”), indicated by arrows in Fig. 21.17b, which are perpendicular to the film plane, are relatively stronger among rest of the vectors on the corresponding rings evidencing appreciable amount of texturing of the Fe grains along the (110) direction.

448

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

Experimental Simulated (c)

0.1

0.2

Reflectivity (a.u.)

(b)

0.1

0.2

(a)

XRR of 57 Fe=Al MLS (a) asdeposited (inset—thickness t versus scattered density profile Rho) (b) annealed at 623 K for 1 h (c) annealed at 723 K for 4 h. (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.)

Rho

FIGURE 21.16

t(Å) 0.1 q(Å)

0.2

All the visible diffraction rings corresponding to the annealed MLS (Fig. 21.17d) are indexed by the B2 type phase of Fe50Al50. This indicates complete alloying leading to the formation of nearly single phase B2 Fe50Al50. As indicated by arrows, the Fe50Al50 (110) vectors (with an angular distribution 10 ) normal to the film plane are very intense, indicating a (110) textured growth of Fe50Al50 after annealing, which is very clear from the electron micrograph corresponding to the annealed MLS (Fig. 21.17c). The TEM micrograph evidences a well-defined MLS structure of Fe50Al50 layers separated by thin layers (1 nm) with bright contrast after annealing. This bright thin layer can be attributed to unreacted, excess Al. The diffracted intensity from such a thin Al layer will be too weak to be detected. The contrast between the finally formed Fe50Al50 alloy layer and the thin Al layer is sufficient enough for reasonably good coherence and hence gives rise to Bragg peaks both in XRR and in low-angle electron diffraction (Fig. 21.17c inset). The occurrence of complete alloying at high temperature with an intact layer structure shows a rather unusual growth phenomenon. It should be noted that the two uppermost bilayers are destroyed, possibly due to oxidation during annealing. 21.4.2 DC Magnetization Studies The DC bulk magnetic properties of the MLS were studied using magnetization measurements at room and high temperatures on a vibrating sample magnetometer. RT magnetization (M–H) measurements of the as-deposited MLS with the plane of the film parallel to the applied magnetic field confirm that the direction of easy axis of magnetization is along that of the applied field (Fig. 21.18).

21.4 MAGNETIC THIN FILMS/MULTILAYER SYSTEMS: 57Fe/AI MLS

449

FIGURE 21.17 (a) Micrograph image of asdeposited MLS (inset—low-angle diffraction pattern of as-deposited MLS). (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.) (b) Wide-angle diffraction pattern of as-deposited MLS. (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.) (c) Micrograph image of annealed at MLS (inset—lowangle diffraction pattern of annealed MLS). (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.) (d) Wideangle diffraction pattern of MLS annealed at 723 K for 4 h. (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.)

Since the RT hysteresis loop shows the clear presence of two overlapping loops, individual loops were fitted using the following relation [54]: 
 2M s H  HC pS tan 1 ; (21.10) tan FðHÞ ¼ p HC 2 where F is the moment at field H, Ms the saturation magnetization, HC the coercivity, and S the squareness Mr/Ms (Mr being the remanent moment). Values of saturation magnetization, squareness, and HC for individual loops were varied to satisfactorily reproduce the experimentally obtained loops (Fig. 21.18). The fitted curves consist of a square loop and a nonsquare curve (Fig. 21.18 inset). The slight imperfectness in the overlap of the experimental and calculated curves can be because the fitting has been done for only two unique sets of parameters, while there can be different Fe-rich Fe–Al phases with close-lying values of parameters having an average value around the two unique sets that have been used in the fitting. The parameters obtained by fitting are listed in Table 21.6. In order to get more insight into the processes taking place at the interfaces during slow annealing and their influence on the corresponding magnetic properties, in situ M–H measurements were done at different intervals of temperature. TABLE 21.6 Experimental and Fitted Parameters of As-Deposited and at High-Temperature M–H Loops [53] Experimental

Fitted Nonsquare Loop

Fitted square loop

MS Mr HC HC (mT) Mr MS Mr MS T (K) HC (mT) (emu cm 3) (emu cm 3) (mT) ( ve) (þve) (emu cm 3) (emu cm 3) HC (mT) (emu cm 3) (emu cm 3) 293 323 373 423 443 463 473 483 513

0.955 0.879 0.702 0.630 0.556 0.493 0.440 0.311 0.171

137.98 101.21 128.9 90.93 88.04 86.18 72.33 72.44 57.36

83.639 85.274 82.100 78.533 71.314 68.161 63.230 54.800 26.932

0.8 0.7 0.5 0.2 0.3 0.2 0.2

0.8 0.7 0.4 0.2 0.1 0.1 0.1

85 35 80 20 20 20 05

17 7 16 4 4 4 1

0.10 0.9 0.8 0.7 0.6 0.5 0.4

75 75 75 75 75 70 65

75 75 75 75 75 70 65

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

450

100

100

RT

0

0

463K

–100 –60

–40

–20

0

20

40

–100 60 –60

100

0

M (emu cm–3)

–40

–20

0

20

40

60

100

60

–40

–20

0

20

40

60

–40

–20

0

20

40

60

–40

–20

0

20

40

60

473K

–60

483K

0 –50

–100 –40

–20

0

20

40

60

100

–60 50

0 423K

0

513K

–50

–100 –40

–20

0

20

40

60

–60 200

100 0 443K –100 –60

40

50

0 373K

–60

20

–50

–100

–60

0

50

0 323K

–60

–20

–40

0

–40

–20

0

20

40

60

–200

533K

–5000

0 –4

Applied magnetic field (X 10

5000

T)

FIGURE 21.18 Hysteresis loops of as-deposited 57 Fe=Al MLS at different temperatures. (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.)

Evolution of both loops for M–H curves at higher temperatures shows that, in general, Ms of the nonsquare loop becomes less with increasing temperature while that of the square loop is constant up to 443 K, after which it starts to decrease. The variation of Ms, Mr, and HC with temperature is shown in Table 21.6. With increase in temperature, the area of the nonsquare loop is observed to decrease steadily. Since increasing the temperature facilitates interdiffusion of Al and Fe, the nonsquare loop can be assigned to pure Fe. The coercivity of the nonsquare loop is smaller than that of the other but is still larger than that of pure a-Fe. This is due to exchange coupling between the two magnetic regions so that the observed coercive fields are not those of the bulk phases but a weighted average [33]. This would also account for the development of the wasp waist shape in the experimental hysteresis loops more noticeable in loops recorded at higher temperatures. Fitting of two subloops for the high-temperature M–H loops has shown that this shape cannot be accounted for by a simple difference in the coercivities of the two loops alone. Spin orientation in Fe film layers is very sensitive to the thickness and conditions of deposition [55]. For example,  isotropic magnetization in Fe–Al multilayers for a range of Fe layer thickness between 20 and 40 A has been reported,  while it is in plane for Fe thickness 10 A [56]. In the present MLS, considerable stress is developed at the interface due to  large mismatch in the lattice parameters of Fe and Al (2.866 and 4.049 A, respectively) as also between Fe and Fe3Al like regions that have lattice parameters about double that of Fe. These stresses can cause alignment of magnetic moments within the intermixed Fe–Al region evidenced by the square loop obtained in the M–H curve of the as-deposited sample.

21.4 MAGNETIC THIN FILMS/MULTILAYER SYSTEMS: 57Fe/AI MLS

451

€ ssbauer (CEMS) Study 21.4.3 Mo The magnetization studies clearly indicate that considerable intermixing at the interfaces has occurred in the as-deposited MLS and also after heat treatment. To identify the resulting Fe-containing phases and intermixing mechanisms at interfaces, CEMS spectra were recorded at room temperature. The spectrum of the as-deposited sample consists of a sextet and a doublet, while that of the sample after heat treatment has only a paramagnetic component (Fig. 21.19). The sextet is broad and can be considered to be an overlap of close-lying sextets, each due to Fe with different number of nearest neighbor Al, indicating a certain degree of interatomic mixing at the interfaces between Fe and Al layers. While fitting the CEMS spectra, the intensity ratio of first and third peaks (A13) of all subspectra has been constrained to be three, while that of the second to third peak, A23, has been left as a floating parameter to get information about the anisotropy of the two magnetic regions. Three major magnetic components with hyperfine fields corresponding to 32.3, 29, and 16.6 T with relative areas 15, 40, and 25%, respectively, are required to fit the CEMS spectrum satisfactorily. The subspectrum corresponding to 32.3 T has a linewidth of 0.33 (0.067) mm s 1 and A23 of 2.5 (0.017), indicating isotropic magnetization of this component. Subspectral components with hyperfine fields 29 and 16.6 T have linewidths of 0.63 (0.067) and 1.26 (0.067) mm s 1, respectively. The corresponding area ratios of the middle peaks, A23, are four for both, indicative of in-plane magnetization. The rather large linewidths of the 29 and 16.6 T components is because of considerable disorder that gives rise to slight variations in the number of nearest neighbor Al to Fe. The 32.3 T component can thus be assigned to a-Fe, while the 29.0 and 16.6 T components with relative area ratios 1:1.67 (0.005) show that Fe is in an environment with nearest neighbors Al comparable to disordered Fe3Al [57]. Even in bulk Fe3Al, the value of individual hyperfine fields and consequently, the average hyperfine field is dependent on the degree of ordering. Moreover, in the present sample, since Fe–Al alloy regions are formed at film interfaces, a rigid conformation of obtained hyperfine field parameters to that in the bulk is not expected. Studies by Kaptas et al. [46] on ultrathin Fe/Al MLS, by transmission M€ ossbauer spectroscopy, have shown that the formation of disordered Fe3Al is favored with as little as two atomic layers of Fe. The difference in the hyperfine field parameters compared to that in bulk Fe3Al obtained by them was attributed to a possible disorder and Al surplus. The CEMS study thus supports the conclusion that the nonsquare subloop obtained in the DC magnetization study is due to pure Fe and that the square subloop is attributable to a local Fe3Al like magnetic region. The central doublet (with relative area of 20%) of the as-deposited MLS has a quadrupole splitting of 0.50 (0.067) mm s 1 and an isomer shift of 0.19 (0.067) mm s 1 (with respect to metallic iron). The width of individual peaks (0.47 mm s 1) in the doublet is large compared to the width of metallic iron (0.24 mm s 1). Hence, the central quadrupole doublet also consists of a number of close-lying Lorentzians indicative of a distribution in the EFG. The values of the average quadrupole splitting and isomer shift of this central doublet are typical of nonmagnetic, Al-rich Fe–Al phases such as Fe2Al5 and FeAl2 [49].

(d)

Intensity (a.u.)

(a)

–10 –8 –6 –4 –2

0

2

4

6

8

(b)

–10 –8 –6 –4 –2

10 –10 –8 –6 –4 –2

0

2

4

6

8

10

FIGURE 21.19

(c)

0

2

4

6

8

10 –10 –8 –6 –4 –2

Velocity (mm s–1)

0

2

4

6

8

10

Room-temperature CEMS spectra of 57 Fe=Al MLS (a) as-deposited (b) after VSM high-temperature M–T measurement (c) annealed at 623 K for 1 h (d) annealed at 723 K for 4 h. (Reprinted with permission from Ref. 53. Copyright (2008) by the American Institute of Physics.)

452

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

CEMS spectrum of the MLS after high-temperature M–H study shows a broad singlet with a width of 0.67 (0.067) mm s 1 and an isomer shift value of 0.20 (0.03) mm s 1 with respect to metallic Fe. This broad singlet can be ascribed to the B2 Fe50Al50 phase with structural defects as indicated by the relatively large linewidth [58]. Since the hightemperature study consisted of slow heating lasting for about 6 h in all and also because Fe layers are much thinner than Al layers, it is evident that nearly all Fe atoms have been completely used up through interdiffusion. Further annealing up to 723 K does not bring about any change in the CEMS spectrum except for a narrowing of the central singlet since longer annealing leads to decreased vacancies and defects, resulting in better ordering as reflected in the reduction in linewidth to 0.40 mm s 1. The isomer shift also becomes 0.27 (0.03) mm s 1, indicating the formation of a thermodynamically favored, stable Fe50Al50 phase. Complementary characterization techniques used in studying the Fe/Al MLS have thus helped to identify the different types of distinct magnetic and nonmagnetic Fe–Al regions in the as-deposited film and also after heat treatment. The reason for the different phases formed can be explained in terms of the molecular dynamics simulation by Chan-Yeup et al. [58] who quantitatively investigated the interfacial features and the growth morphology of Fe–Al bimetallic films showing that the intermixing between the deposited Fe atoms on Al substrate and vice versa are not the same and give rise to two different types of interfaces. The conditions for initial deposition adopted in the present study, therefore favor the formation of a magnetic interface that is, on an average, Fe3Al like and a nonmagnetic, Al-rich Fe–Al interface. This can be attributed to the formation of two different types of interfaces due to deposition of Fe over Al and Al over Fe as has been observed in other systems such as Co on Fe and Al on Co [59]. In the present study, since Al and Fe have been deposited alternately to form the MLS, there can be two distinct interfaces and this is confirmed by room-temperature CEMS results of the as-deposited MLS, which show contributions from an Fe3Al like magnetic and FeAl2 like nonmagnetic regions. On annealing, the continuous diffusion of Fe/Al finally results in the formation of a stable, nonmagnetic, equiatomic Fe50Al50 phase. Although the temperature range for which the annealing was done favors the formation of Al3Fe, the formation of Fe50Al50 is more favored in the present case as was also observed by Checchetto et al. [60]. Formation of phases different from those predicted by the phase diagram in these films is because phase diagrams describe an equilibrium condition. Thin film formation is nonequilibrium in nature, in which a time dependence of the solid-state reaction is usually observed. Thus, in the thin-film regime, both the thermodynamic and kinetic processes together contribute to the formation of a particular phase. The heat of formation (DH) of the Al-rich phases Fe3Al (DH ¼ 27.9 kJ mol 1), Fe2Al5 (DH ¼ 27.0 kJ mol 1), and FeAl2 (DH ¼ 26.1 kJ mol 1) has values quite similar to that of the B2 phase (DH ¼ 25.1 kJ mol 1) [60]. These factors together indicate that the mixed interface layer during the deposition plays an important role for nucleation of the B2 Fe50Al50 phase. Crosssectional TEM micrograph and CEMS studies of the annealed MLS prove that prolonged heating even at higher temperatures after the formation of Fe50Al50 leaves the MLS intact. Once the B2 phase is nucleated, its distribution within the film is controlled by a reduction in the diffusion of Al atoms through the intermetallic layers already formed. The B2 Fe50Al50 phase thus acts as a diffusion barrier for the remaining pure Al layer, as confirmed by XRR, TEM micrograph images, and SAD data analysis of the sample after high-temperature study (Fig. 21.17b,c). The conditions of deposition and annealing adopted in this study for a Fe–Al ratio of nearly 1:2 result in the formation of a very stable, nearly pure intermetallic Fe50Al50/Al multilayer structure. The Fe50Al50 phase formed on annealing works as a barrier that prevents further diffusion of Al in the film. The formation of barrier layers thus addresses the problem of interface conduction electron scattering, which is a major challenge in developing a stable magnetic MLS where the effect of the intermetallic interfacial layers and conduction electron scattering at the interface have to be considered. This can be prevented by deposition of the magnetic layer on a nonmagnetic bulk of the same material since there would then be a perfect matching of the electronic potentials at the magnetic/nonmagnetic interface. The nonmagnetic bulk should be stable and not be amenable to diffusion. Ideally, therefore, a nonmagnetic multilayer structure with appropriate barrier layers, as has been developed in the multilayer system discussed here, would meet this requirement.

21.5 CONCLUSIONS In this chapter, using examples of mainly three types of magnetic materials, namely, nanosized powders of ferrites, mechanically alloyed/milled Fe–Cr–Al intermetallics, and a Fe–Al multilayer system, it has been demonstrated how M€ ossbauer spectroscopy is a powerful tool to understand the bulk magnetic properties in such nanosized systems. This is mainly because of the extreme sensitivity of the M€ ossbauer probe atom to short-range effects that get modified on a

REFERENCES

453

change in nearest neighbor configuration, type of nn ions, defects, strains, distortions, and so on. By a combination of room-temperature, low-temperature, and in-field M€ ossbauer techniques, structural information is obtained. Additionally, identification/contributions from various magnetic or chemical phases are made possible through a comparison of hyperfine field values, isomer shifts, relative spectral areas, peak widths, and so on. When used with other conventional techniques such as DC magnetization and X-ray diffraction, the origin of different types of magnetic interactions present in such complex structures can be completely elucidated.

ACKNOWLEDGMENTS I would like to acknowledge the financial support from the DST-FIST project of the Department of Physics, Mohanlal Sukhadia University, Udaipur, Rajasthan, India and also from the BRNS research project. I thank the UGC-DAE CSR, Indore, India, for use of many of their sophisticated facilities, especially TEM, CEMS, and low-temperature, in-field M€ ossbauer setup.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

Z. Zhang, T.M. Nenoff, J.Y. Huang, D.T. Berry, P.P. Provencio, J. Phys. Chem. C 2009, 113, 1155–1159. N. Zhao, L. Li, T. Huang, L. Qi, Nanoscale 2010, 2, 2418–2423. Y. Si, E.T. Samulski, Nano Lett. 2008, 8, 1679–1682. Y.M. Lvov, P. Pattekari, X. Zhang, V. Torchilin, Langmuir 2010, 27, 1212–1217. R.L. Klueh, N. Hashimoto, P.J. Maziasz, Scr. Mater. 2005, 53, 275–280. J.T. Groves, J. Inorg. Biochem. 2006, 100, 434–447. J. Garg, B. Poudel, M. Chiesa, J.B. Gordon, J.J. Ma, J.B. Wang, Z.F. Ren, Y.T. Kang, H. Ohtani, J. Nanda, G.H. McKinley, G. Chen, J. Appl. Phys. 2008, 103, 074301. L. Xi, Z. Wang, Y. Zuo, X. Shi, Nanotechnology 2011, 22, 045707. M. Auffan, J. Rose, O. Proux, D. Borschneck, A. Masion, P. Chaurand, J.L. Hazemann, C. Chaneac, J.P. Jolivet, M.R. Wiesner, A. Van Geen, J.Y. Bottero, Langmuir 2008, 24, 3215–3222. C. Upadhyay, H.C. Verma, S. Anand, J. Appl. Phys. 2004, 95, 5746–5751. J.M. Greneche, Hyperfine Interact. 2003, 148/149, 79–89. R.K. Sharma, O. Suwalka, N. Lakshmi, K. Venugopalan, A. Banerjee, P.A. Joy, Mater. Lett. 2005, 59, 3402–3405. R.K. Sharma, V. Sebastian, N. Lakshmi, K. Venugopalan, V.R. Reddy, A. Gupta, Phys. Rev. B 2007, 75, 144419. H. Bhargava, N. Lakshmi, V. Sebastian, V.R. Reddy, K. Venugopalan, A. Gupta, J. Phys.: D Appl. Phys. 2009, 42, 245003. G.A. Petitt, D.W. Forester, Phys. Rev. B 1971, 4, 3912–3923. E.C. Sousa, M.H. Sousa, H.R. Rechenberg, G.F. Goya, F.A. Tourinho, R. Perzynski, J. Depeyrot, J. Magn. Magn. Mater. 2007, 310, e1020–e1022. G.A. Sawatzky, F. Van Der Woude, A.H. Morrish, Phys. Rev. 1969, 187, 747–757. K. Krieble, T. Schaeffer, J.A. Paulsen, A.P. Ring, C.C.H. Lo, J.E. Snyder, J. Appl. Phys. 2005, 97, 10F101. J.D. Dunitz, I.E. Orgel, J. Phys. Chem. Solids 1957, 3, 20–29. J. Jiang, G.F. Goya, H.R. Rechenberg, J. Phys. Condens. Matter 1999, 11, 4063–4078. A.M. Kumar, K.H. Rao, J.M. Greneche, J. Appl. Phys. 2009, 105, 073919. B. Ghosh, S. Kumar, A. Poddar, C. Mazumdar, S. Banerjee, V.R. Reddy, A. Gupta, J. Appl. Phys. 2010, 108, 034307. L. Theil Kuhn, A. Bojesen, L. Timmermann, M. Meedom Nielsen, S. Mørup, J. Phys. Condens. Matter 2002, 14, 13551–13567. A. Labarta, X. Batlle, O. Iglesias, Surface Effects in Magnetic Nanoparticles, D. Fiorani, eds., Springer, New York, 2006, pp. 105–140. L.D. Bianco, D. Fiorani, A.M. Testa, E. Bonetti, L. Signorini, Phys. Rev. B 2002, 66, 174418. E.J. Choi, Y. Ahn, K.-C. Song, J. Magn. Magn. Mater. 2006, 301, 171–174. A. Narayanasamy, Pramana J. Phys. 2005, 65, 893–900. H.R. Rechenberg, E.C. Sousa, J. Depeyrot, M.H. Sousa, R. Aquino, F.A. Tourinho, R. Perzynski, Hyperfine Interact. 2008, 184, 9–14. N. Lakshmi, H. Bhargava, O.P. Suwalka, K. Venugopalan, V. Sebastian, V.R. Reddy, A. Gupta, Phys. Rev. B 2009, 136, 174425.

454

30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60.

€ 21 APPLICATION OF MOSSBAUER SPECTROSCOPY TO NANOMAGNETICS

B. Norman, G. Lee, Phys. Rev. 1964, 136, A133–A137. C.K. Ong, H.C. Fang, Z. Yang Y. Li, J. Magn. Magn. Mater. 2000, 213, 413–417. W.H. Meiklejohn, C.P. Bean, Phys. Rev. 1957, 105, 904–913. J. Nogues, I.K. Schuller, J. Magn. Magn. Mater. 1999, 192, 203–232. A.E. Berkowitz, K. Takano, J. Magn. Magn. Mater. 1999, 200, 552–570. J. Nogues, J. Sort, V. Langlais, V. Skumryev, S. Suri~ nach, J.S. Mu~ noz, M.D. Bar o, Phys. Rep. 2005, 422, 65–117. L. Sun, P.C. Searson, C.L. Chien, Phys. Rev. B 2005, 71, 012417. J. Garcıa-Otero, M. Porto, J. Rivas, J. Appl. Phys. 2000, 87, 7376–7381. L.D. Bianco, D. Fiorani, A.M. Testa, E. Bonetti, L. Savini, S. Signoretti, Phys. Rev. B 2004, 70, 052401. V. Sebastian, N. Lakshmi, K. Venugopalan, J. Magn. Magn. Mater. 2007, 309, 153–159. B. Window, J. Phys. E 1971, 4, 401–402. R.A. Dunlap, D.J. Lloyd, I.A. Christie, G. Stroink, Z.M. Stadnik, J. Phys. F 1988, 18, 1329–1341. L.F. Kiss, D. Kaptas, J. Balogh, L. Bujdos o, T. Kemeny, L. Vincze, Phys. Rev. B 2004, 70, 012408. G. Herzer, IEEE Trans. Magn. 1990, 26, 1397–1402. R.W. Gao, W.C. Feng, H.Q. Liu, B. Wang, W. Chen, G.B. Han, P. Zhang, H. Li, W. Li, Y.Q. Guo, W. Pan, X.M. Li, M.G. Zhu, X. Li, J. Appl. Phys. 2003, 94, 664–668. A. Hernando, T. Kulik, Phys. Rev. B 1994, 49, 7064–7067. D. Kaptas, J. Balogh, T. Kemeny, L.F. Kiss, L. Bujdaso, A. Kovacs, A. Hirata, I. Vincze, Phys. Rev. B 2007, 75, 014417. M. Carbucchio, M. Rateo, G. Ruggiero, M. Solzi, G. Turilli, J. Magn. Magn. Mater. 1999, 33, 196–197. J.L. Jordan, S.C. Deevi, Intermetallics 2003, 11, 507–528. V. Sebastian, N. Lakshmi, K. Venugopalan, Intermetallics 2007, 15, 1006–1012. D.C. Meyer, A. Gorbunov, W. Pompe, A.A. Levin, Cryst. Res. Technol. 2005, 40, 125–134. V. Tutovan, D. Luca, Appl. Surf. Sci. 1993, 65/66, 145–148. A.A. Levin, D.C. Meyer, P. Paufler, A. Gorbunov, A. Tselev, P. Gawlitza, J. Alloys Compd. 2001, 320, 114–125. S. Jani, V. Sebastian, N. Lakshmi, K. Venugopalan, V.R. Reddy, A. Gupta, N.P. Lalla, J. Appl. Phys. 2008, 104, 123907. G.R. Pattanaik, D.K. Pandya, S.C. Kashyap, J. Alloys Compd. 2001, 326, 260–264. S.L. Vysotski9l, S.S. Gel’bukh, A.S. Dzhumaliev, G.T. Kazakov, A.Yu. Filimonov, A.Yu. Tsyplin, Tech. Phys. Lett. 1999, 25, 79–82. R. Brajpuriya, S. Tripathi, A. Sharma, T. Shripathi, S.M. Chaudhari, Eur. Phys. J. B 2006, 51, 131–137. N. Lakshmi, K. Venugopalan, J. Varma, Phys. Rev. B 1992, 47, 14054–14061. C. Chan-Yeup, C. Yong-Chae, Mater. Lett. 2006, 60, 1063–1067. S.P. Kim, S.C. Lee, K.R. Lee, Y.C. Chung, J. Electroceram. 2004, 13, 315–320. R. Checchetto, C. Tosello, A. Miotello, G. Principi, J. Phys. Condens. Matter 2001, 13, 811–821.

C H A P T E R 2 2

€ MOSSBAUER SPECTROSCOPY AND SURFACE ANALYSIS

 GANCEDO, MATTEO MONTI, AND JUAN DE LA FIGUERA JOSE F. MARCO, JOSE RAMON

Instituto de Quımica Fısica “Rocasolano”, CSIC, Madrid, Spain

22.1 INTRODUCTION Surface analysis is a complex matter. The complexity starts by the definition of the term “surface” itself. Not everyone understands the same by surface. We often hear and speak about the surface of Earth or Mars, about the inner surfaces of a zeolite, about the existence of surface processes such as corrosion or adhesion, and so on. Depending on the field or the particular application, we talk about different things that have, however, a common characteristic. In all the cases, the surface is a border region that separates the solid (the bulk) from the environment (liquid or gaseous). It is the thickness of this border region, and what happens within this thickness, that determines the characteristics of many processes that are relevant in both basic science and technology. This thickness may extend from a few atomic layers (and thus be the subject of “surface science”) to a few nanometers (surface analysis) or a fraction of a micrometer (thin film analysis). Thus, in this chapter the term “surface” represents the external part of a solid having a thickness from a fraction of a monolayer to a few nanometers. The complexity mentioned above continues when one wishes to establish the limits of the term “analysis.” In a very broad sense, the term analysis is generally associated with composition. Knowledge of the composition of a surface is indeed extremely important to gain the understanding of many important processes, but sometimes more information is needed (structure, topography, etc.) to obtain a more complete description of the surface properties. Spectroscopic techniques are usually needed to obtain surface composition, diffraction techniques to inform about surface structure, and imaging techniques to characterize the surface topography. Therefore, in a general way, the analysis of a surface should include as many of those techniques as possible. This book is about M€ ossbauer spectroscopy. We might wonder how it can be applied to study surface layers a few nanometers thick. A surface analytical technique needs to meet two important requirements: (i) it should be sensitive, that is, it should be able to detect a small number of atoms (while a cubic centimeter of solid material contains roughly substitute 1023 atoms, a square centimeter of the same material contains a much smaller number of atoms, approximately 1015 atoms), and (ii) it should be specific, that is, it should be able to separate the contributions of the surface from those of the bulk. The techniques based on the detection of low-energy electrons are especially suited for the study of surfaces. This is due to the short electron inelastic mean free path in that energy range, which is within the range of a few nanometers [1]. Fortunately, during the nuclear deexcitation subsequent to the resonant nuclear absorption (M€ ossbauer effect), a large number of electrons of different energies are generated that can be used as a probe for surface analysis. The ability of M€ ossbauer spectroscopy to provide chemical, compositional, structural, and magnetic information makes it highly desirable to study surfaces that contain M€ ossbauer active isotopes. M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

455

€ 22 MOSSBAUER SPECTROSCOPY AND SURFACE ANALYSIS

456

After more than 50 years since the discovery of the M€ ossbauer effect, there exist a large number of reviews and book chapters devoted to the application of M€ ossbauer spectroscopy (and in particular to its variant integral conversion electron M€ ossbauer spectroscopy (ICEMS)) to the study of surfaces. Although we will talk briefly here about ICEMS, this chapter will be more focused on exploring the possibilities and the technical and practical difficulties of performing M€ ossbauer spectroscopy by means of the detection and analysis of resonant electrons of low or very low energy. Finally, we will also review some interesting developments in imaging M€ ossbauer techniques as well as the perspectives in the application of surface M€ ossbauer spectroscopy to the study of very thin layers (a few atomic layers thick) of iron oxides and, particularly, of Fe3O4, a material that is experiencing a renovated interest.

€ 22.2 THE PHYSICAL BASIS: HOW AND WHY ELECTRONS APPEAR IN MOSSBAUER SPECTROSCOPY After the resonant nuclear absorption, the excitation energy of the nucleus can be dissipated by two competing processes: gamma decay or internal conversion [2]. In an internal conversion process, the nucleus can directly impart its excess energy in a one-step process to an electron of an atomic shell ejecting it from the atom. The kinetic energy of the ejected electron may be relatively high since it is equal to the energy of the nuclear transition minus the binding energy of the electron in its corresponding atomic shell. It must be noted that internal conversion is a direct process and must not be confused with a second-order process, also possible but much less probable, in which a photon emission from the nucleus leads to the subsequent ejection of an electron from the atom. Internal conversion processes affect electrons of different atomic shells and thus we refer to them as K conversion, L conversion, M conversion, . . . electrons. Since the K electrons have a larger probability to venture within the nuclear region, the K conversion process is more probable than the L and M conversion processes. In the case of the 14.4 keV transition of 57 Fe, the probabilities of deexcitation by emitting a K, L, or M conversion electron are 81, 9, and 1%, respectively, and the corresponding probability of deexcitation by gamma emission is 9% [3]. As we have commented above, both gamma decay and internal conversion are competing processes, and for a given nuclear transition, the conversion coefficient is defined as a ¼ ðconversion probability=gamma emission probabilityÞ ¼ N e =N g ;

(22.1)

where Ne and Ng are the number of conversion electron and photons, respectively, observed per unit time. Table 22.1 collects the a values for the gamma transitions of some common M€ ossbauer isotopes. These values correspond to the total electron conversion yield. The ejection of an electron from a particular atomic shell leaves a hole that can be filled by other electrons in the atom. As a result, further emission in the form of Auger cascades or fluorescent X-rays takes place [1]. The probability of Auger emission or X-ray emission depends on the atomic number of the atom, the former being more likely for atoms having Z < 50. Besides the Auger cascades, there are other mechanisms that can lead to the emission of electrons. So, when an electron from an inner shell leaves the atom, a part of its energy can be invested in releasing an electron from an outer shell in a process that is known as shake-off [1]. The energy of shake-off electrons is usually small and in the case of the relaxation of 57 Fe they represent a significant fraction of the total number of electrons emitted (Table 22.2) [3]. In addition to resonant conversion, Auger cascades, and shake-off electrons, there are other secondary mechanisms that can lead to the generation of resonant electrons. Thus, electrons coming from the photoionization of an atom hit by resonant gamma or X-rays originated during the nuclear or atomic relaxation of the M€ ossbauer isotope (and usually denoted as GPE and XPE, respectively) are also electrons carrying resonant information and, consequently, contribute to the M€ ossbauer spectrum [4].

€ ssbauer Isotopes TABLE 22.1 Total Electron Conversion Coefficients (aT) for Some Transitions of Common Mo Isotope Transition energy (Eg, keV) aT

57

Fe 14.4 8.2

119

Sn 23.875 5.1

121

Sb 37.15 10

151

Eu 21.6 29

€ 22.2 THE PHYSICAL BASIS: HOW AND WHY ELECTRONS APPEAR IN MOSSBAUER SPECTROSCOPY

457

TABLE 22.2 Emission Probabilities for Electrons of Different Energies Emitted After the Deexcitation of a 57 Fe Nucleus [3] Electron Energy (eV) 0–15 29–100 50–800 5400–7000 7300 13000 14400

Electron Origin

Probability (%)

MMM, shake-off MMM, MMN, shake-off LMM KLL K conversion L conversion M conversion

23.5 33.9 19.8 8.6 12.5 1.4 0.2

It seems, then, that the processes following the nuclear resonant absorption, namely nuclear deexcitation and atomic relaxation, are quite complex. As a result, a large number of resonant electrons of different energies, and consequently able to travel different depths in the solid, are generated (Fig. 22.1 and Table 22.2). The variant of the M€ ossbauer spectroscopy technique arising from the detection of all the electrons emitted after a M€ ossbauer event, irrespectively of their energy, is called integral conversion electron M€ ossbauer spectroscopy and it has been the subject, as stated in Section 22.1, of numerous reviews (see Refs 5–9 and references therein). By depositing iron layers of different thicknesses on a stainless steel substrate, it can be demonstrated that the penetration depth of K conversion electrons is about 50 nm, while that of L conversion electrons extends up to approximately 330 nm [10]. This means that, when studying massive samples, ICEMS provides information up to a depth of 330 nm, although this information is heavily weighted between the surface and a depth of 50 nm since the emission probability of K conversion electrons is much higher than that of the L conversion electrons. This information depth is much larger than that characteristic of other more popular surface spectroscopic techniques such as X-ray photoelectron spectroscopy (XPS) or Auger electron spectroscopy (AES), where the information obtained comes mainly from the outermost 3 nm [1].

FIGURE 22.1 Schematic diagram showing the nuclear and atomic relaxation processes subsequent to the nuclear resonant absorption in 57 Fe that lead to the emission of electrons of different energies.

€ 22 MOSSBAUER SPECTROSCOPY AND SURFACE ANALYSIS

458

FIGURE 22.2 (a) ICEMS spectrum recorded from a 5 nm 57 Fe film deposited by evaporation on a Si wafer. The film is slightly oxidized at the surface and a doublet corresponding to a small-particle Fe3þ oxyhydroxide is observed besides the a-Fe contribution. (b) The Fe 2p XPS spectrum recorded from the same film that is completely dominated by the Fe3þ contribution of the oxyhydroxide.

Since the understanding of many chemical and physical problems depends on the behavior of the atoms confined precisely in this thin outermost part of the solids, it would be most interesting to devise procedures to enhance the surface sensitivity of M€ ossbauer spectroscopy. This will be discussed in the following section.

€ 22.3 INCREASING SURFACE SENSITIVITY IN ELECTRON MOSSBAUER SPECTROSCOPY We have to distinguish between two different situations:  The study of a very thin layer,1 from a fraction of monolayer to several atomic layers, on an inert (i.e., not M€ ossbauer active) substrate.  The study of a thin layer (3–5 nm in depth) on a thick or massive iron-containing substrate. The first situation necessarily implies the deposition of pure 57 Fe films to obtain a measurable signal and ultrahighvacuum (UHV) conditions to ensure film purity. Submonolayer resolution can be attained with ICEMS. This subject, and its application to the study of surface and interface magnetism, has been reviewed in detail [11]. Inthesecondsituation,theuseofICEMSwouldnotgivethedesiredresultssincetheinformationdepthextendsuptoabout 330 nm and an important contribution from the substrate would mask the contribution from the outermost layer. This is even observed with relatively thin samples. For example, Fig. 22.2(a) shows the ICEMS spectrum recorded from a 5 nm thick iron 1

In the rest of the chapter, we will restrict all our descriptions and discussions to

57

Fe, which is the most common M€ ossbauer isotope.

€ 22.3 INCREASING SURFACE SENSITIVITY IN ELECTRON MOSSBAUER SPECTROSCOPY

459

film,slightlyoxidizedatthesurface.Thespectrumisdominatedbyanintensesextetcorrespondingtoa-Feandaminordoublet that corresponds to the Fe3þ oxyhydroxide of the oxidized surface. On the contrary, the Fe 2p XPS spectrum recorded from the same sample [Fig. 22.2(b)] shows a very intense component that is characteristic of Fe3þ and a minor component that correspondstometalliciron.TheresultisillustrativesinceitdemonstratesthatXPSismuchmoresurfacesensitivethanICEMS. This is undoubtedly related to the energy of the electrons involved in both experiments. Since the Fe 2p photoelectrons of metallic iron have a much smaller kinetic energy (750–780 eV in the present case) than the conversion electrons, they are efficientlyattenuatedbytheoxidizedlayerandcontributeverylittletotheobservedXPSspectrum.Otherapproachesarethus necessary to increase the surface sensitivity of M€ ossbauer spectroscopy. The most direct one is to select the electron energy by means of an appropriate electron energy analyzer. Since electrons interact strongly with matter, they rapidly lose energy in their way from the point they have been generated to the surface and then the loss in energy can be correlated with depth. This depth-selective approach is known as DCEMS (depth-selective conversion electron M€ ossbauer spectroscopy). Although several electrostatic electron analyzers specifically designed for DCEMS have been described [12–20] and their ability to perform depth-selective analysis demonstrated (Fig. 22.3 shows a spherical electron analyzer built by the authors together with some representative DCEMS spectra [9,17]), the real situation is that none appears to be currently under operation. One of the most important limiting factors in DCEMS is that the experiments are extremely time consuming. Even with the use of sources of very high activity, samples rich in 57 Fe, and favorable analyzer designs in terms of transmission, luminosity, and acceptance angles, the number of electrons reaching the detector is small. It must be taken into account that in order to (a)

2+

Fe

α-Fe

2+

Fe

(c)

3+

Fe

16000

Effect (%)

130 7.3 keV 125 120 115 110 105 100

(b)

8000 6000 4000

Conversion K

10000

Auger KLM

Auger KLL

12000

Auger LMM

Number of counts

14000

2000 0

2000

4000 Energy (eV)

6000

8000

125 7.0 keV 120 115 110 105 100 120 6.7 keV 115 110 105 100 150 ICEMS 140 130 120 110 100 –6 –4

–2 0 2 Velocity (mm s–1)

4

6

FIGURE 22.3 (a) Electrostatic spherical analyzer for DCEMS. The vacuum chamber and the upper pole of the outer spherical surface have been removed to show the upper pole of the inner surface where the sample is located. (b) Electron spectrum recorded with this analyzer from a sample of 57 Co. The arrows indicate the energies at which the corresponding spectra on the right have been taken. (c) DCEMS spectra recorded at different electron energies from a 57 Fe thin film subjected to a corrosion process in a SO2-polluted atmosphere. The spectra show the peaks of the a-Fe substrate and Fe2þ and Fe3þ contributions corresponding to the corrosion layer. Note how the intensity of the a-Fe peaks increases with decreasing electron energy. An ICEMS spectrum of the same sample is shown for comparison purposes.

460

€ 22 MOSSBAUER SPECTROSCOPY AND SURFACE ANALYSIS

perform energy analysis with a reasonable energy resolution using an electron analyzer (i) a narrow window from the total electron energy spectrum has to be selected and (ii) the geometry is quite restrictive, as the electrons are collected within a solid angle much smaller than the favorable 2p solid angle used in conventional gas electron counters. This implies that only a small fraction of the total number of electrons generated is counted. Another important drawback of DCEMS is that the calculations needed to correlate the observed spectrum with the depth from which the electrons have been generated are not straightforward [6,21–23], even with carefully prepared, flat, homogeneous samples. In the case of “real” samples (e.g., corrosion samples) that might not be flat and that might have a heterogeneous composition, this task becomes almost impossible. A different approach that attracted considerable attention in the past [3,24–28], was nearly forgotten and has recently concentrated some interesting activity [29], is to perform surface M€ ossbauer spectroscopy by means of the detection of very low-energy electrons. This variant of the technique is known as ILEEMS (integral low-energy electron M€ ossbauer spectroscopy). As we have mentioned before, the deexcitation of the M€ ossbauer nucleus is accompanied by the emission of very low-energy electrons (mainly shake-off and low-energy Auger electrons). It has been estimated that electrons with energy below 15 eV are responsible for more than 50% of the total resonant signal. In fact, Monte Carlo simulations have shown that every single M€ ossbauer event is accompanied by 6.5 ionizations [26]. Of these, 1.5 correspond to resonant electrons of high energy (K and L conversion electrons and KLL, KLM, and KMM Auger electrons) and 5.0 to LMM, MMM, and MMN Auger electrons with energies lower than 100 eV and shake-off electrons with energies lower than 15 eV. It is also thought that these low-energy electrons are weighted with information from the topmost 5.0 nm of the specimen, which clearly opens the way to perform surface analysis with information depths similar to those characteristic of AES or XPS. In principle, detection of very low-energy electrons can be easily done using a channeltron as electron detector and applying an appropriate positive bias between the sample and the detector to carry out an appropriate energy filtering (we will provide more specific experimental details in the following sections). There is, however, a problem that is inherent to all the approaches mentioned here and it is how to improve the signal-to-noise ratio by separating the resonant electron contribution from the nonresonant electron contribution. The most important nonresonant contribution is that corresponding to photoelectrons emitted by the different elements in the sample and generated by the various gamma- and X-ray photons involved in the M€ ossbauer experiment. It must be ossbauer source emits not only 14.4 keV quanta, but also 6.3 keV X-rays and, even taken into account that the 57 Co M€ more abundantly, 122.0 and 136.3 keV gamma rays. The presence of low-Z elements in the sample helps to increase the signal-to-noise ratio as the probability of photoelectric effect is proportional to Z5. However, as pointed out in Ref. 11, the presence of iron itself in the sample can contribute to increase the nonresonant background with Fe 7.3 keV photoelectrons because their energy and that of the K conversion electrons are the same and cannot be filtered. Although the probability of production of 7.3 keV photoelectrons is two orders of magnitude smaller than that of K conversion electrons [11], their number increases with thickness. Another important nonresonant contribution comes from secondary electrons of very low energy ( 13). A solid Fe(III)–hydroxy complex was formed in strongly alkaline aqueous solutions. M€ ossbauer spectroscopy indicated that iron was highly symmetrical in this solid product, whereas XANES and EXAFS indicated that the  coordination of Fe(III) was 6, with a primary FeO bond distance of 2.03 A . Thus, it was concluded that the elementary building block in the solid product was FeðOHÞ6 3 . On the other hand, it was demonstrated that in the mother liquor the ossbauer highest coordination of Fe(III) was 4 (the formation of tetrahedral FeðOHÞ4  ). Kamnev et al. [17] also used M€ spectroscopy to investigate solid a-FeOOH and soluble products of Fe(III) in aqueous media at a high pH. The second step in the precipitation of solid phase by a slow or forced hydrolysis of iron(III)–salt solutions is olation, that is, the polymerization of hydroxy complexes and the formation of hydroxy bridges, Fe-(OH)–Fe. The olation process is reversible and depolymerization can be easily achieved by acidifying the solution. At the next step of polymerization there is a formation of oxo bridges  ¼ Fe OH þ HO Fe ¼!¼ Fe O Fe ¼ þH2 O

(23.4)

and this process end in the formation of Fe(III) oxyhydroxide or oxide. Slow or forced Fe(III) hydrolyses are accompanied by a gradual pH decrease. For this reason, the oxolation reaction could also be written as a deprotonation of hydroxy polymers  þ Hþ Fe !  Fe O Fe OH Fe

(23.5)

The polymerization of Fe(III) hydroxy complexes and the formation of a solid phase depend on the type and concentration of the Fe(III)–salt, pH, temperature, and reaction time. Spiro et al. [18–20] isolated the polymer from ossbauer spectroscopy. The a hydrolyzed Fe(NO3)3 solution and characterized it by different methods, including M€ empirical composition of this polymer corresponded to [Fe(OH)x(NO3)3x]n where x lies between 2.3 and 2.5 and n is of the order of 900. Segal [21] prepared a series of colloidal dispersions by adding sodium bicarbonate to Fe(NO3)3 solutions in the concentration range of 0.10–0.001 M. The ratio [OH]:[Fe3þ] varied. It was concluded that both unpolymerized and polymerized Fe3þ ions were present for [OH]:[Fe] < 2.5:1.

23.3 PRECIPITATION OF IRON OXIDES BY HYDROLYSIS REACTIONS Hydrolysis reactions in Fe(III)–salt solutions were extensively exploited to prepare iron oxides of desired properties [22]. The nature of anion plays an important role in the formation of solid hydrolytical products. Figure 23.1 shows principal solid products obtained by the hydrolysis of Fe(NO3)3, FeCl3, Fe(ClO4)3, or Fe2(SO4)3 aqueous solutions. The principal solid product of Fe(NO3)3 hydrolysis is a-FeOOH. Dousma and De Bruyn [23] investigated the hydrolyses of

473

23.3 PRECIPITATION OF IRON OXIDES BY HYDROLYSIS REACTIONS

FIGURE 23.1 Principal solid products obtained by the hydrolysis of Fe(NO3)3, FeCl3, Fe(ClO4)3, or Fe2(SO4)3 aqueous solutions.

Fe(NO3)3 solutions. Precipitates formed from 6.25  101 and 6.25  102 M Fe(NO3)3 solutions at 24 and 55  C showed the characteristic pattern of a-FeOOH, whereas in 6.25  103 M, Fe(NO3)3 solution at 24  C an amorphous precipitate was formed. a-Fe2O3 precipitated at 90  C. The same authors proposed (Fig. 23.2) the mechanism of hydrolysis–precipitation from Fe(NO3)3 solutions, consisting of the hydrolysis to monomers (A) and dimers (B), reversible, rapid growth of small polymers (C), the formation of slowly reacting large polymers (D), and polymerization and oxolation with the formation of a solid phase (E). XRD, FT-IR, and M€ ossbauer spectroscopies were used to characterize hydrolytical solid products obtained by the hydrolysis of 0.1 M Fe(NO3)3 aqueous solutions at 90  C in the presence of hexamethylenetetramine (HMTA) [24,25]. HMTA generates OH ions at an elevated temperature in accordance with the chemical reactions ðCH2 Þ6 N4 þ H2 O ! 4NH3 þ 6HCHO

(23.6)

NH3 þ H2 O ! NH4 þ þ OH

(23.7)

and

Three solid phases were detected: “amorphous” (low-crystalline ferrihydrite), a-FeOOH, and a-Fe2O3. The ratio of these fractions depended on the starting concentration of HMTA and the time of hydrolysis. Christensen et al. [26] made a suspension by adding NaOH solution to Fe(NO3)3 solution up to the pH 4.3 and left it to crystallize at room temperature for 23 years. After this time of crystallization, the pH of the suspension decreased to 2.85. XRD showed the pattern of a-FeOOH as a single phase. TEM images showed the needle-like and rhombohedral morphologies of a-FeOOH particles. The M€ ossbauer spectrum at 295 K showed a magnetically splitted spectrum and a central quadrupole doublet. This superparamagnetic doublet disappeared at 80 K. Barrero et al. [27] found by means of M€ ossbauer spectroscopy that a-FeOOH precipitated from Fe(II) solutions exhibited the superparamagnetic relaxation effect, whereas a-FeOOH precipitated from Fe(III) solutions exhibited a magnetic ordering of clusters. It was also found that the stirring during the precipitation from Fe(III) solution may affect the total water content and the magnetic behavior of a-FeOOH. De Grave et al. [28] used M€ ossbauer spectroscopy to investigate changes in nanosized a-FeOOH particles upon their hydrothermal treatment. The low-crystalline a-FeOOH particles grew larger with the increased autoclaving temperature, whereas changes in size of relatively well-crystallized a-FeOOH particles were insignificant at the same temperatures. The M€ ossbauer spectrum of a high-crystalline a-FeOOH is characterized by one sextet of spectral lines exhibiting the intensity ratio of 3:2:1:1:2:3. However, the spectral lines of precipitated a-FeOOH are as a

FIGURE 23.2 Proposed scheme of hydrolysis/ precipitation from an aqueous Fe(NO3)3 solution. (Reproduced from Ref. 23 with permission of Elsevier.)

474

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

TABLE 23.2 Hyperfine Magnetic Fields of the Most Common Iron Oxides, as € ssbauer Spectroscopy [22,31] Measured Using Mo Iron Oxide a-FeOOH

b-FeOOH g-FeOOH d0 -FeOOH

Fe3O4 g-Fe2O3 a-Fe2O3

Temperature RT 77 K 4.2 K 77 K 4.2 K 4.2 K RT 77 K 4.2 K RT RT 77 K RT 4.2 K

HMF (T) 38.2 50.0 50.6 46.0 48.2 46.0 42.0 53.0 53.5 49.0 46.0 50.0 52.6 51.7 54.2 53.3

rule broadened and deviate from the theoretical intensity ratio of 3:2:1:1:2:3 [29,30]. In the case of very fine a-FeOOH particles, one paramagnetic doublet can be obtained at RT. a-FeOOH particles smaller than 8 nm show a superparamagnetic type of the M€ ossbauer spectrum down to 77 K [31]. Hyperfine magnetic fields (HMFs) of the most common iron oxides are listed in Table 23.2. Cabral-Prieto et al. [32] precipitated nanosized a-FeOOH particles by the oxygenation of Fe(OH)2 suspension. Crystallite sizes of a-FeOOH calculated by using the Scherrer formula were in the range of 7–48 nm. Crystallites from 7 to 13 nm mainly exhibited superparamagnetic behavior and the specific area ossbauer spectrum and the 200 m2 g 1, whereas the crystallites from 23 to 48 nm showed the hyperfine magnetic M€ specific surface area 100 m2 g 1. Gotic and Music [33] investigated the precipitation of iron oxides from 0.1 M FeSO4 solutions containing varying concentrations of decomposing urea. The precipitation systems were aerated to oxidize Fe(II). In dependence on the experimental conditions, a-FeOOH, a-Fe2O3, and Fe3 xO4 were precipitated. The ossbauer spectrum of M€ ossbauer spectra showed only the formation of substoichiometric magnetite (Fe3 xO4). The M€ a-FeOOH showed spectral differences and the corresponding spectra could be well fitted only taking into account the distribution of HMF. b-FeOOH is a principal solid product of FeCl3 hydrolysis. Frayer et al. [34] prepared cube-shaped a-Fe2O3 particles by boiling 0.01 M FeCl3 solution, whereas cigar-shaped b-FeOOH particles were precipitated below 70  C. Matijevic and Scheiner [35] produced b-FeOOH or a-Fe2O3 particles from hydrolyzing FeCl3 solutions in dependence on the ossbauer spectroscopy to monitor forced hydrolyses concentrations of Fe3þ and Cl ions. Music et al. [36] used M€ (90  C) of Fe3þ ions in aqueous solutions containing nitrate, chloride, or sulfate anions. It was proposed that polymers formed from hydrolyzing 0.1 M Fe(NO3)3 solution included no NO3 ions, whereas the polymer precursors of b-FeOOH formed by a forced hydrolysis of 0.1 M FeCl3 solutions contained Cl ions in place of some OH ions. Bottero et al. [37] confirmed an earlier proposal [36] by using synchrotron measurements. Khoe and Robins [38] used potentiometric titrations to investigate polymerization during the hydrolysis of Fe3þ ions in solutions having the concentration in the range of 2  104 to 2  102 M. The authors also suggested the incorporation of Cl and SO4 2 ions into the structure of hydroxy polymers. The precipitates formed by forced hydrolysis at 120  C in Fe(NO3)3, FeCl3 and mixed Fe(NO3)3 þ FeCl3 solutions were investigated [39] using XRD, FT-IR, M€ ossbauer, and TEM. a-FeOOH and a-Fe2O3 were the phases precipitated from Fe(NO3)3 solution, whereas b-FeOOH and a-Fe2O3 were the typical products of FeCl3 hydrolysis. The hydrolysis of 0.2 M FeCl3 þ 0.1 M HCl yielded a-Fe2O3 with traces of a-FeOOH. In the mixed Fe(NO3)3 þ FeCl3 solutions, the phase composition of the hydrolytical product was determined by the concentration of the dominant Fe(III)–salt. The effect of HCl additions on the forced hydrolysis of FeCl3 solutions has been investigated by Music et al. [40]. b-FeOOH precipitated as a single phase from 0.01 M FeCl3 þ 0.01 M HCl solution after 1 and 3 days of forced hydrolysis at 90  C. In the presence of [HCl]  0.05 M, the precipitates also contained a small amount of a-FeOOH. Upon 35 days of forced hydrolysis a-Fe2O3 was formed as the end product of the precipitation process. It is generally accepted that in an aqueous medium, the transformation of b-FeOOH to a-Fe2O3

475

23.3 PRECIPITATION OF IRON OXIDES BY HYDROLYSIS REACTIONS

FIGURE 23.3 Characterization of the precipitates obtained by the forced hydrolysis of 0.1 M FeCl3 solutions with varying concentrations of HMTA (T ¼ 90  C; reaction time ¼ 5 h). (Reproduced from Ref. 44 with permission of Elsevier.)

proceeds through the dissolution/recrystallization mechanism [41,42]. This mechanism differs from the one in solid-state transformation b-FeOOH to a-Fe2O3 [43]. S9aric et al. [44] monitored chemical and microstructural changes in the precipitates formed by the forced hydrolysis of 0.1 M FeCl3 solution in the presence of HMTA. In the absence of HMTA, after 5 h of aging, b-FeOOH particles were formed as a single phase, exhibiting different morphologies, such as spindles, stars, X-, and Y shapes. With an increase in the starting concentration of HMTA up to 0.100 M, XRD showed a gradual broadening of diffraction lines of b-FeOOH, whereas at the starting concentration of HMTA ¼ 0.250 M, an amorphouslike peak belonging to low-crystalline two-XRD-line ferrihydrite was detected. This effect of HMTA addition was accompanied by a corresponding pH increase. Figure 23.3 shows the selected XRD patterns and M€ ossbauer spectra. The M€ ossbauer spectrum of the precipitate formed at the initial concentration of HMTA ¼ 0.250 M showed a quadrupole doublet at RT and the superposition of a quadrupole doublet and a collapsing sextet at liquid nitrogen temperature (LNT), which can be assigned to low-crystalline ferrihydrite. The b-FeOOH formation at pH values significantly higher than pH 2 was explained by an incomplete substitution of Cl ions in Fe(III)-hydroxy polymers with OH ions. TEM showed a gradual decrease in b-FeOOH particle size at the increased pH of the mother liquor (Fig. 23.4). For a long aging time (7 days), the phase transformations b-FeOOH to a-Fe2O3 and b-FeOOH ! a-FeOOH ! a-Fe2O3 were discussed in terms of the dissolution/recrystallization mechanism. Music et al. [45] investigated some factors influencing the forced hydrolysis of FeCl3 solutions. Phase composition, size, and shape of iron oxide particles generated by the forced hydrolysis of Fe3þ ions depended on FeCl3 and HCl concentrations, as well as the presence of 0.005 M quinine hydrogen sulfate (QHS). Upon 7 days of the forced hydrolysis of 0.01 M FeCl3 solution a-Fe2O3 precipitated, the mixture of b-FeOOH and a-Fe2O3 precipitated from 0.02 M FeCl3 solution, whereas b-FeOOH precipitated from 0.1 M FeCl3 solution. The forced hydrolysis of 0.1 M FeCl3 þ 0.1 M HCl solution yielded a-Fe2O3. In the presence of 0.005 M QHS, the formation of b-FeOOH and a-Fe2O3 was partially or completely suppressed in dependence on the experimental conditions for the forced hydrolysis of Fe3þ ions. The specific shapes of b-FeOOH particles (cigar-type, X-, Y-, and star shapes) were not obtained in the presence of 0.005 M QHS. The bundles of b-FeOOH needles with irregular ends were obtained instead, as a direct result of QHS presence during the precipitation process. Music et al. [46] used M€ ossbauer spectroscopy and TEM in the investigation of the formation of a-Fe2O3 particles in an aqueous medium. The hydrolyzing ossbauer and FT-IR FeCl3 solutions were partially neutralized with tetramethylammonium hydroxide (TMAH). M€ spectroscopies and TEM were used in the investigation of the precipitation of iron oxides by the forced hydrolysis of 0.1 M FeCl3 or 0.1 M Fe(NO3)3 solutions in decomposing urea [47]. It was shown [48] that under certain experimental conditions b-FeOOH could precipitate from FeCl3 solutions at the pH up to 9. At an increased pH, an amorphous fraction was also formed parallel to the already formed b-FeOOH. The amorphous fraction was unstable and transformed further to a-FeOOH and/or a-Fe2O3.

476

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

FIGURE 23.4 TEM images of the particles precipitated by the forced hydrolysis of 0.1 M FeCl3 solutions with varying concentrations of HMTA (T ¼ 90  C; reaction time ¼ 5 h). (Reproduced from Ref. 44 with permission of Elsevier.)

The RT M€ ossbauer spectrum of b-FeOOH is a superposition of two distinct doublets whose origin is related to the incorporation of chlorides into its crystal structure [49,50]. The origin of these two doublets was also discussed by other researchers [51–54]. Barrero et al. [55] focused on the M€ ossbauer spectrum of b-FeOOH. The RT M€ ossbauer spectrum was fitted with two doublets, whereas the low-temperature spectra were fitted with four sextets. Two sextets were assigned to two nonequivalent Fe3þ sites located close to the Cl ion, whereas the other two sextets were assigned to Fe3þ sites located at the vacant chloride sites. A crystal structure refinement of b-FeOOH nanoparticles, taking into account the monoclinic structure, was performed using the MAUD program [56,57]. The monoclinic symmetry requires the existence of two distinct iron octahedral sites, which is confirmed by the RT M€ ossbauer spectrum as a superposition of two quadrupole doublets. Barrero et al. [58] hydrolyzed at 70  C 0.1 M FeCl3 solutions containing varying concentrations of Al3þ and Ti4þ ions. b-FeOOH was formed as a single phase only. No significant variation in the M€ ossbauer and crystallographic parameters of b-FeOOH, precipitated in the presence of up to 10 mol% of Al3þ, was observed, hence it was concluded that only traces of Al entered the crystal structure of b-FeOOH. On the other hand, in the presence of 10 mol% of Ti4þ the decrease in average size and a reduction of the both unit cell and hyperfine parameters of b-FeOOH particles were observed. It was also found that forced FeCl3 hydrolysis in the presence of Al3þ, Cr3þ, or Cu2þ up to 10 mol% in decomposing urea yielded b-FeOOH [59]. Al3þ insignificantly altered crystallographic and M€ ossbauer parameters, whereas Cr3þ and Cu2þ ions changed very little the crystallographic parameters of b-FeOOH. Kamimura et al. [60] investigated the influence of Cr3þ, Cu2þ, and Ni2þ, as well as anions such as SO4 2 and NO3 on the forced hydrolysis of FeCl3 solutions. The XRD lines of b-FeOOH were significantly broadened when chromium(III) sulfate was present during the precipitation process. The M€ ossbauer spectrum at 78 K showed the superposition of an intense central quadrupole doublet and hyperfine magnetic components. At 5 K this doublet disappeared; however, the spectral lines were broader than those for pure b-FeOOH. It was concluded that Cr3þ ions were incorporated into b-FeOOH, but Ni2þ and Cu2þ were not. In the presence of SO4 2 ions they can partially substitute Cl ions and thus inhibit crystal growth and/or distort the b-FeOOH crystal structure. Yang et al. [61] prepared b-FeOOH nanoparticles 20–30 nm long and 3–5 nm in diameter using the microemulsion method. Possible changes in magnetic properties and the M€ ossbauer spectra of b-FeOOH in dependence on the preparation procedure and particle size were discussed. Funk et al. [62] investigated a series of iron(III) oxyhydroxide complexes kept in a solution as colloidal particles protected by different carbohydrate coatings. The M€ ossbauer spectra of iron sucrose, iron polymaltose, iron dextran, and iron dextrin complexes were investigated. Superparamagnetic blocking temperatures as determined by M€ ossbauer spectroscopy indicated the relative diameters of the aggregates to be between 2.5 and 4.1 nm. Three of the iron preparations had a b-FeOOH-like structure, whereas one had a ferrihydrite-like structure.

477

23.3 PRECIPITATION OF IRON OXIDES BY HYDROLYSIS REACTIONS

Aqueous Fe(ClO4)3 solutions also exhibit specific hydrolytical behavior during slow or forced hydrolysis. Hsu and Ragone [63] investigated the hydrolysis of three 0.01 M Fe(ClO4)3 solutions having R ¼ 0, 1, or 2, where R ¼ ½HCO3 Š=½Fe3þ Š. Experiments were performed at 25  C. Changes in these precipitation systems were monitored over the period of 8 months. At R ¼ 0, hydrolyzed particles grew from 650 nm after 9 days. Twenty-three percent of iron as FeOOH was found in the sediment upon 5 months of hydrolysis. At R ¼ 1 and R ¼ 2, a rapid hydrolysis and polymerization were observed immediately after preparation and colloidal suspensions were stable for at least 2 years. Hsu [64] has monitored the hydrolyses of 0.01 or 0.001 M Fe(ClO4)3 solutions seeded with brown iron(III) hydroxide sols. In two experiments, HClO4 was added to fresh Fe(ClO4)3 solutions to lower their degree of supersatuaration with respect to iron(III) hydroxide. The author concluded that the initial number of nuclei relative to the concentration of monomeric Fe(III) species was the key factor governing particle size distribution and the stability of colloidal suspensions. Hsu [65] and Hsu and Wang [66] analyzed the solid hydrolytical product of Fe(ClO4)3 solutions. For slow hydrolysis at RT a-FeOOH or a mixture of a-FeOOH and g-FeOOH were found, whereas the solid products of forced hydrolysis at 70  C were a-FeOOH and a-Fe2O3. Lorenz and Kempe [67] investigated the hydrolysis of Fe(ClO4)3 solutions in the presence of decomposing urea. a-Fe2O3, a-FeOOH, and g-FeOOH were found in the precipitates and the actual phase composition depended on the initial concentration of Fe(ClO4)3 and the molar ratio [Fe(ClO4)3]/[urea]. Murphy et al. [68] monitored aging at 20  C of partially neutralized Fe(ClO4)3 solutions. At an early stage of aging spherical polycations (1.5–3 nm), then nanosized rods were formed that were identified as a-FeOOH. Throughout the aging period these rods grew in length but not in width. Large lath-shaped g-FeOOH particles were detected in some cases. Ristic et al. [69] investigated the microstructural properties of g-FeOOH, a-FeOOH, and ossbauer spectroscopies and electron microsa-Fe2O3 particles produced by Fe(ClO4)3 hydrolysis using FT-IR and M€ copy. The kinetics of precipitation by hydrolysis from 0.005 to 0.01 M Fe(ClO4)3 solutions at RT were slow, which is in accordance with Hsu and Ragone [63]. Infrared spectroscopy is a very useful technique for analysis of oxyhydroxides a-FeOOH, b-FeOOH, and g-FeOOH phases, even in traces. On the other hand, the infrared spectroscopy results are less significant in the identification of oxide mixtures a-Fe2O3, g-Fe2O3, and Fe3O4 due to the overlapping of prominent IR bands. Figure 23.5 shows the FT-IR spectra of the precipitates obtained by slow hydrolysis in aqueous media at RT of 0.005 M Fe(ClO4)3 and 0.01 M Fe(ClO4)3, respectively, for 3 years. In the first case the hydrolytical product was identified as a pure g-FeOOH phase on the basis of an IR band at 1156 cm 1 and two sharp bands at 1021 and 745 cm 1. These IR bands could be assigned to g-FeOOH in accordance with the result of a group-theoretical analysis of the IR spectrum of g-FeOOH as reported by Weckler and Lutz [70]. The FT-IR spectrum of the precipitate produced by the slow hydrolysis of 0.01 M Fe(ClO4)3 solution also showed additional IR bands at 892 and 792 cm 1 that could be assigned to a small amount of a-FeOOH. The IR band at 1636 cm 1 was due to H2O bending vibrations and a band at 3420 cm 1 to H2O stretching vibrations, thus indicating adsorbed H2O (moisture). Figure 23.6 shows FE-SEM images of g-FeOOH particles. A lamellar microstructure is well visible. The hydrolysis of Fe(ClO4)3 solutions at 90 and 160  C yielded a-Fe2O3 as an end product. Figure 23.7a and b shows FE-SEM images of a-Fe2O3 particles produced by the hydrolysis of 0.1 M Fe(ClO4)3 þ 0.1 M HClO4 solution at 90  C for 21 days. Nearly monodisperse a-Fe2O3 particles were formed. Their formation was governed by the aggregation mechanism. For the same concentration conditions and hydrolysis at 160  C during 24 h the a-Fe2O3

FIGURE 23.5 FT-IR spectra of hydrolytical products obtained by a slow hydrolysis at room temperature of (a) 0.005 M Fe(ClO4)3 solution (g-FeOOH), and (b) 0.01 M Fe(ClO4)3 solution (g-FeOOH þ a small amount of a-FeOOH). (Reproduced from Ref. 69 with permission of Elsevier.)

478

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

FIGURE 23.6 Lamellar microstructure of g-FeOOH particles precipitated by a slow hydrolysis from (a, b, and c) 0.005 M Fe(ClO4)3 solution, or (d) 0.01 M Fe(ClO4)3 solution at room temperature. (Reproduced from Ref. 69 with permission of Elsevier.)

FIGURE 23.7 a-Fe2O3 particles precipitated by the forced hydrolysis from 0.1 M Fe(ClO4)3 þ 0.1 M HClO4 solution for 21 days at 90  C (a and b), and from 0.1 M Fe(ClO4)3 þ 0.1 M HClO4 solution for 1 day at 160  C (c and d). (Reproduced from Ref. 69 with permission of Elsevier.)

479

23.3 PRECIPITATION OF IRON OXIDES BY HYDROLYSIS REACTIONS

particles of different shapes and sizes were produced, as shown in Fig. 23.7c and d. These particles were also aggregates of ossbauer spectra corresponded to a-Fe2O3 as a single phase. smaller a-Fe2O3 particles. Their M€ Generally, the shape and size of a-Fe2O3 particles can be strongly influenced by the addition of inorganic ions, organic molecules, or polymers. For example, a forced hydrolysis of 0.1 M Fe(ClO4)3 solution at 160  C for 24 h yielded nanosized pseudospherical a-Fe2O3 particles (Fig. 23.8a). The addition of sodium polyanethol sulfonate at the start of the precipitation process changed completely the size and shape of particles [71]. Large hollow particles in the micrometer range were clearly visible (Fig. 23.8b). These hollow particles showed a substructure. The subunits were elongated

FIGURE 23.8 Pseudospherical a-Fe2O3 nanoparticles obtained by the forced hydrolysis of 0.1 M Fe(ClO4)3 solution for 24 h at 160  C (a); the addition of sodium polyanethol sulfonate at the beginning of the precipitation process caused the formation of hollow a-Fe2O3 particles in the micometer range (b); and the wall of hollow a-Fe2O3 particles shows a lateral aggregation of a-Fe2O3 particles (c). (Reproduced from Ref. 71 with permission of Elsevier.)

480

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

secondary particles laterally arrayed (Fig. 23.8c) and these subunits consisted of thin elongated primary a-Fe2O3 particles also laterally arrayed. XRD and M€ ossbauer spectroscopy showed that these particles obtained in the presence of a polyanetholsulfonate anion consisted of a-Fe2O3 as a single phase. IR spectroscopy showed that sulfate groups formed a bidentate bridging complex on the surface of a-Fe2O3 particles. The preferential adsorption of sulfate groups was responsible for the elongation of a-Fe2O3 particles along the crystallographic c-axis. During the precipitation process, the sulfonate groups were converted to sulfate groups. In relation to previously discussed anions, sulfates also show a specific influence on the precipitation of iron oxides by hydrolysis reactions. Dousma et al. [72] investigated the influence of sulfate anions on the formation of iron oxides using the titration of acidified 6.25  102 M Fe2(SO4)3 solution with NaOH solution. Precipitates were analyzed by XRD and M€ ossbauer spectroscopy. a-FeOOH, a-Fe2O3, and amorphous fractions were detected, in dependence on the instrumental technique used. The authors concluded that sulfates could promote or suppress the formation of iron oxides, depending on the experimental conditions. The precipitation of a jarosite-type compound was not observed. Matijevic et al. [73,74] investigated conditions for precipitation of monodisperse basic iron(III) sulfate particles from acidic solutions. The role of FeSO4þ complex in the formation of basic iron(III) sulfate was emphasized. S9ubrt et al. [75] investigated formation of iron oxides by the hydrolysis of Fe2(SO4)3 solution with decomposing urea. The hydrolysis of Fe2(SO4)3 solution was also conducted in the presence of decomposing hexamethylenetetramine. Schwertmannite, jarosite, a-FeOOH, and a-Fe2O3 were the phases found in the precipitates, with a phase composition depending on the variations of experimental conditions. Voigt and G€ obler [76] autoclaved 1 M Fe2(SO4)3 solution between 125 and 150  C. The hydrolytical product was hydronium jarosite, (H3O)Fe3(OH)6(SO4)2. At 250–300  C basic sulfate, Fe(OH)SO4 was formed, which further reacted with H2O producing a-Fe2O3. Music et al. [77] investigated the properties of the ossbauer precipitates formed by the hydrolysis of Fe2(SO4)3 solutions of varying concentrations at 90 or 120  C. XRD, M€ and FT-IR spectroscopies and electron microscopy were used as experimental techniques. The concentration regions of Fe2(SO4)3 salt solutions were determined for the precipitation of a-FeOOH or H3OFe3(OH)6(SO4)2 as a single phase. a-FeOOH particles exhibited superparamagnetic behavior. The hydrolysis of 0.1 M Fe2(SO4)3 solutions at 120  C ossbauer parameters recorded for produced H3OFe3(OH)6(SO4)2 and basic sulfate, Fe4(OH)10SO4. RT M€ H3OFe3(OH)6(SO4)2 corresponded to literature data [78] for this compound, dFe ¼ 0.38–0.39 mm s1 and D ¼ 1.00– 1.10 mm s1. Quadrupole splitting recorded for Fe4(OH)10SO4 varied between 0.868 mm and 0.525 mm s1. Gancedo and Martinez [79] published the following RT M€ ossbauer parameters for Fe4(OH)10SO4 as the corrosion product: dFe ¼ 0.35 0.03 mm s1 and D ¼ 0.55 0.03 mm s1. The observed D changes were probably affected by the nonstoichiometry of Fe4(OH)10SO4. Kandori et al. [80] investigated hydrolysis reactions in mixed Fe2(SO4)3 þ HCl solutions. At a relatively low concentration of Fe2(SO4)3 acicular or rod-like multidomain a-Fe2O3 particles were produced, whereas at higher Fe2(SO4)3 concentrations irregularly shaped H3OFe3(OH)6(SO4)2 particles were obtained. Acicular a-FeOOH particles were produced at pH  1.6, whereas H3OFe3(OH)6(SO4)2 particles were produced at pH  1.5. The properties of precipitates obtained by the hydrolysis of NH4Fe(SO4)2 solutions at 90 or 120  C were investigated by XRD, M€ ossbauer, and FT-IR spectroscopies [81]. The concentrations of NH4Fe(SO4)2 salt solutions were determined in relation to the precipitation of ammonium jarosite NH4Fe3(OH)6(SO4)2, or superparamagnetic a-FeOOH particles.

23.4 PRECIPITATION OF IRON OXIDES FROM DENSE b-FeOOH SUSPENSIONS Researchers mainly investigated the precipitation of iron oxides by hydrolysis reactions at relatively low concentrations of Fe3þ ions. This is understandable because the precipitation process is easier to control at relatively low concentrations of Fe3þ ions than at high concentrations of these ions. Monodisperse pseudocubic a-Fe2O3 particles (mean edge length 1.65 mm) were prepared by aging at 100  C for 8 days from condensed ferric hydroxide gel containing an excess of Fe3þ ions [82]. The mean diameter of a-Fe2O3 particles was decreased to 0.1 mm and the time required for the total conversion to a-Fe2O3 was shortened to less than 1 day as the excess concentration of Fe3þ ions was reduced to nominally zero. With addition of sulfate or phosphate anions, the pseudocubic a-Fe2O3 particles turned into peanut9 ic et al. [84] investigated type particles via an ellipsoidal shape, in dependence on the concentration of the additive [83]. Z microstructural changes in the particles formed during the transformation from b-FeOOH to a-Fe2O3 in dense aqueous suspensions. The concentrated FeCl3 solution was partially neutralized with a concentrated NaOH solution beforehand. The kinetics were conducted at 160  C. Between 30 and 60 min of autoclaving the dense suspension contained b-FeOOH as a single phase, whereas between 96 min and 72 h the precipitate consisted of a-Fe2O3, as shown by XRD and M€ ossbauer spectroscopy. A small amount or traces of a-FeOOH as an intermediate phase were detected by FT-IR,

23.4 PRECIPITATION OF IRON OXIDES FROM DENSE b-FEOOH SUSPENSIONS

481

FIGURE 23.9 Characteristic XRD patterns of the particles formed during the transformation of b-FeOOH to a-Fe2O3 in dense suspensions at 160  C for the starting concentration of (a) 0.1 M AAS and (b) 0.001 M AAS. The corresponding times of precipitation are denoted. AAS: ammonium amidosulfonate. (Reproduced from Ref. 86 with permission of Elsevier.)

which disappeared upon 6 h of autoclaving. A gradual formation of a-Fe2O3 double spheres with ring was observed. The double spheres with ring were formed by the aggregation mechanism. The orientation effect of a-Fe2O3 subunits in double spheres with ring was found. Upon 24 h of autoclaving two narrow size distributions of double spheres with ring were measured. The first narrow particle size distribution consisted of the particles 8 mm in length and 6.5 mm in width, where the width is a diameter of the particle’s ring. The second narrow particle size distribution consisted of smaller particles (also double spheres with ring) of dimensions 3.5 mm (length) and 2.5 mm (width). a-Fe2O3 subunits in the ring showed a tendency to make arrays parallel to the ring’s surface, that is, opposite to the a-Fe2O3 subunits in the sphere. The effect of temperature on the crystallization of a-Fe2O3 particles from dense b-FeOOH suspensions was also monitored by XRD, M€ ossbauer, FT-IR, and FE-SEM þ EDS [85]. Dense suspensions of very long laterally arrayed b-FeOOH fibrils were obtained at 90  C. Crystallization at 120  C between 18 and 72 h yielded monodisperse a-Fe2O3 particles of a shape similar to that of double spheres with ring. With a further rise in temperature up to 200  C, a-Fe2O3 subunits increased in size and the formation of big pores was observed. The increase in size of a-Fe2O3 subunits, as well as ossbauer and FT-IR spectra. Concentrated the increase in their crystallinity had an effect on the corresponding 57 Fe M€ FeCl3 solutions containing starting concentrations 0.1, 0.01, or 0.001 M of ammonium amidosulfonate (AAS) were partially neutralized with a concentrated NaOH solution and the dense suspensions so obtained were autoclaved at ossbauer spectra (20  C) of the particles 160  C for different times [86]. Figure 23.9 shows XRD patterns and Fig. 23.10 M€ formed during the transformation of b-FeOOH to a-Fe2O3 in dense suspensions at 160  C at the starting concentrations of (a) 0.1 M AAS and (b) 0.001 M AAS, respectively. The FT-IR spectra showed the presence of sulfate groups generated by the decomposition of amidosulfonate groups. It could be inferred that the sulfate groups were attached to the a-Fe2O3 surface as a bidentate ligand. Figure 23.11 shows the influence of AAS on the size and shape of the b-FeOOH fibrils (Fig. 23.11a and b) that were upon 4 h of autoclaving transformed to peanut-type a-Fe2O3 particles (Fig. 23.11c), and upon 72 h of autoclaving the interconnected (with the neck) double a-Fe2O3 coupola (Fig. 23.11d) was formed. The last a-Fe2O3 particles were also porous, consisting of linear chains of small a-Fe2O3 particles (also interconnected) that were directed from the center toward the surface of cupola. Figure 23.12a–c shows changes in the shape of a-Fe2O3 particles produced at different concentrations of AAS. At 0.001 M AAS for 24 h at 160  C double spheres with ring were visible, at

482

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

FIGURE 23.10 57 € ssbauer spectra (20  C) of the particles formed during the transformation of b-FeOOH to a-Fe2O3 in dense Fe Mo suspensions at 160  C for the starting concentrations (a) 0.1 M AAS and (b) 0.001 M AAS. The corresponding times of precipitation are denoted. AAS: ammonium amidosulfonate. (Reproduced from Ref. 86 with permission of Elsevier.)

FIGURE 23.11 FE-SEM images of the particles obtained during the transformation of b-FeOOH to a-Fe2O3 in dense suspensions by autoclaving at 160  C for (a) 1 h, (b) 1.5 h, (c) 4 h, and (d) 72 h. In these precipitation systems the starting AAS concentration was 0.1 M. AAS: ammonium amidosulfonate. (Reproduced from Ref. 86 with permission of Elsevier.)

23.5 PRECIPITATION AND PROPERTIES OF SOME OTHER IRON OXIDES

483

FIGURE 23.12 FE-SEM images of a-Fe2O3 particles obtained during the transformation of b-FeOOH to a-Fe2O3 in dense suspensions by autoclaving at 160  C for 24 h in the presence of (a) 0.001 M, (b) 0.01 M, (c) 0.1 M, and (d) 0.01 M (higher magnification) of ammonium amidosulfonate (AAS). (Reproduced from Ref. 86 with permission of Elsevier.)

0.01 M AAS there were no more rings, and at 0.1 M AAS interconnected double a-Fe2O3 cupolas were present. Figure 23.12d shows the microstructure of one broken a-Fe2O3 particle showing its internal microstructure. The effect of phosphates on the morphology and size of a-Fe2O3 particles crystallized from dense b-FeOOH suspensions was also investigated [87]. The reference precipitation system has already been discussed [84]. After 24 h of autoclaving at 160  C in the presence of 0.01 M Na2HPO4, a-Fe2O3 aggregates of different shapes were obtained instead of double spheres with rings. a-Fe2O3 subunits 20–25 nm in size were in laterally aggregated arrays. With a further increase in Na2HPO4 concentration up to 0.05 M, the crystallization of a-Fe2O3 was strongly suppressed. A strong elongation of b-FeOOH and a-FeOOH particles was observed. In b-FeOOH and a-FeOOH particles, phosphates were preferentially adsorbed on the surface of the crystallites, thus governing preferential crystal growth. As a consequence, very long wires and/or rods of iron(III)oxyhydroxide were obtained. At a high phosphate concentration, the surface precipitation of phosphates on FeOOH particles was possible, and in such way the dissolution of these particles was suppressed. For a-Fe2O3 particles the incorporation of phosphorus into a-Fe2O3 is suggested, taking into account the mechanism of crystal growth on iron(III) phosphate nuclei (probably amorphous) or some kind of iron(III) phosphate complexes serving as a template in the suspension. The incorporation of phosphorus into a-Fe2O3 particles retards the increase in their size, as well as that of the corresponding a-Fe2O3 subunits.

23.5 PRECIPITATION AND PROPERTIES OF SOME OTHER IRON OXIDES 23.5.1 Ferrihydrite Ferrihydrite is naturally present in various water systems, sediments and soils. It is also the product coming from mine wastes and acid mine drainage. Ferrihydrite plays an important role in the geochemical cycling of iron. Specifically, it is an

484

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

important sorbent for many metal cations and oxyanions. In reference literature, there is confusion because many authors describe amorphous ferric hydroxide, colloidal ferric hydroxide, Fe(OH)3, and similar; however, these descriptions actually correspond in many cases to ferrihydrite. Ferrihydrite is generally classified on the basis of Xray diffraction lines. For example, two-XRD-line ferrihydrite is a low-crystalline form, whereas six-XRD-line ferrihydrite is the high-crystalline form. The occurrence and properties of ferrihydrite were discussed by Jambor and Dutrizac [88]. In reference literature, different chemical formulae for ferrihydrite are given: Fe5HO8  4H2O [89], Fe5(O4H3)3 [90], Fe2O3  2FeOOH  2.6H2O [91], and 5Fe2O3  9H2O [92]. Generally, ferrihydrite consists of spherical particles varying several nanometers in size and this is accompanied by varying crystallinity. Due to low crystallinity the distributions of M€ ossbauer parameters are obtained rather than discrete values. At room temperature, the quadrupole doublet is measured and quadrupole splitting changes in dependence on ferrihydrite crystallinity. For example, two-XRD-line ferrihydrite may show a RT quadrupole splitting about 0.8 mm s1, whereas six-XRD-line ferrihydrite has a quadrupole splitting 0.7 mm s1 or less [92]. In the same work [92], Murad and Schwertmann observed the distribution of hyperfine magnetic splitting at 4 K with a HMF about 490 kOe. Fitting of the RT M€ ossbauer spectrum of ferrihydrite with the distribution of quadrupole splitting is more realistic than fitting of two quadrupole splitting sites [93], because in this way it does not imply the existence of two discrete iron sites. Cardile [94] attempted to demonstrate that the RT M€ ossbauer spectrum of ferrihydrite could be fitted for the superposition of three doublets corresponding to one tetrahedral and two octahedral iron sites. Murad et al. [95] investigated the effect of crystallinity on magnetic ordering in natural ossbauer spectroscopy [96,97]. Iron ferrihydrites. Soil iron oxide analogues were characterized by applied-field 57 Fe M€ oxides that exhibit similar zero applied-field M€ ossbauer patterns show distinctive patterns when a magnetic field is applied, due to the specific magnetic properties of each iron oxide. Madsen et al. [98] precipitated four-XRD-line ferrihydrite and investigated it by M€ ossbauer spectroscopy up to 5 K and in varying external magnetic fields up to 4 T. An average magnetic moment for 5 nm particles of ferrihydrite was estimated from the dependence of the hyperfine splitting of the M€ ossbauer spectra on external magnetic fields at 80 K. Coey and Readman [99] investigated natural ferric gel (Fe ossbauer techniques. This ferric gel was actually a natural (OH)3  0.9H2O) using XRD, ED, thermal, IR, magnetic, and M€ ferrihydrite. The authors introduced speromagnetism to explain the magnetic properties of a “ferric gel.” The concept of “speromagnetism” was criticized by Pankhurst and Pollard [100]. These authors [101] also investigated the structural and magnetic properties of two-XRD-line and six-XRD-line ferrihydrites. The authors concluded that Fe3þ ions were located in octahedral sites. The M€ ossbauer spectra indicated the presence of different magnetic states; ferrimagnetism in twoXRD-line ferrihydrite and antiferromagnetism in six-XRD-line ferrihydrite. The M€ ossbauer spectra were interpreted as two ferrimagnetic sublattices in two-XRD-line ferrihydrite and simple antiferromagnetism in six-XRD-line ferrihydrite. Stanjek and Weidler [102,103] investigated the effect of dry heating of two-XRD-line and six-XRD-line ferrihydrites on unit cell parameters, the surface area, and porosity. Greffie et al. [104] used HRTEM to investigate the difference between freeze-dried and untreated synthetic ferrihydrites. As prepared, the two-XRD-line ferrihydrite showed chain-like aggregates of 2–5 nm in particle size, whereas upon freeze-drying the two-XRD-line was transformed to three-XRD-line ferrihydrite and the presence of a-FeOOH phase was also observed. Chadwick et al. [105] prepared an iron oxide precursor by heating to dryness the Fe(NO3)3 solution, then heated so the prepared precursor in a laboratory oven. On the basis of M€ ossbauer spectroscopy, the authors concluded about the presence of ferrihydrite as an intermediate in the transformation iron precursor ! a-Fe2O3. Ristic et al. [106] precipitated low-crystalline ferrihydrite (two-XRD-line) and thermally treated the isolated precipitate at between 195 and 325  C. The XRD pattern of two-XRD-line ferrihydrite ossbauer spectra of two XRD lines of ferrihydrite after heating at 220 and is shown in Fig. 23.13, whereas the 57 Fe M€ ossbauer parameters are given in 245  C, respectively, are shown in Figs. 23.14 and 23.15. The corresponding M€ Table 23.3. The M€ ossbauer spectrum recorded at RT of sample prepared at 220  C showed a broadened quadrupole doublet, whereas HMF distributions were observed at 78 and 12 K, presumably due to nonuniform chemical ossbauer environments of Fe3þ and/or to the shape and size distributions of the a-Fe2O3 clusters. The RT M€ spectrum of sample heated at 245  C showed the superposition of a quadrupole doublet and a distributed sextet, the latter believed to be due to low-crystalline a-Fe2O3. At 78 K two sextets were resolved. The sharper one, with Hhf ¼ 527 kOe and a quadrupole shift 2eQ ¼ 0.12 mm s1, and the fractional area of RA ¼ 0.22, was due to a-Fe2O3, which suggests the coexistence of antiferromagnetic and weak-ferromagnetic spin states. The much broader magnetic component was assigned to ferrihydrite. At 12 K, hematite had Hhf ¼ 535 kOe and 2eQ ¼ 0.03 mm s1. The ferrihydrite sextet was still considerably broadened. The M€ ossbauer spectra at RT and 78 K recorded for sample obtained upon heating at 325  C corresponded only to the a-Fe2O3 fraction. For sample obtained by heating at 245  C the M€ ossbauer spectrum was recorded at 4.2 K in an applied field of 6 T. The ferryhydrite component clearly splitted into a high- and low-field contribution. The heating of two-XRD-line ferrihydrite caused a certain crystalline ordering and in the next stage the transformation to a-Fe2O3.

485

23.5 PRECIPITATION AND PROPERTIES OF SOME OTHER IRON OXIDES

FIGURE 23.13 XRD pattern of two-XRD-line ferrihydrite. (Reproduced from Ref. 106 with permission of Elsevier.)

Schwertmann and Murad [107] used XRD, M€ ossbauer, and TEM to analyze the crystallization products of ferrihydrite at RT for different pH values. Upon 3 years of aging, the maximum of a-Fe2O3 was found between pH 7 and 8, and the maximum of goethite at pH 4 and 12. Schwertmann et al. [108] suggested the crystallization of a-Fe2O3 within ferrihydrite aggregates in the presence of free or surface-bound water. Zhao et al. [109] prepared several ferrihydrite samples under different precipitation and drying conditions and recorded their M€ ossbauer, XRD, and iron K-edge X-ray absorption fine structure (XAFS) spectra. On the basis of these measurements, the authors suggested the presence of tetrahedral iron sites located on the surface of ferrihydrite particles. Andreeva et al. [110] used M€ ossbauer spectroscopy to monitor the transformation of ferrihydrite to a-FeOOH in an acid medium. It was concluded that ferrihydrite was transformed to a-FeOOH by the dissolution/recrystallization mechanism and that this reaction was facilitated by the presence of Fe2þ ions. 23.5.2 Lepidocrocite (g-FeOOH) Lepidocrocite is present in nature in many soils; it is, however, much less common than a-FeOOH or a-Fe2O3. The presence of g-FeOOH in soils is typically due to the precipitation process that includes oxygenation of reduced Fe2þ. The formation of g-FeOOH is suppressed by Al3þ ions, as well as high carbonate concentrations, so that it is not found in calcareous soils. g-FeOOH is a typical product of the early stages of atmospheric steel corrosion. In dependence on dry/ wet cycles it transforms to a-FeOOH and further on to magnetite and g-Fe2O3. Precipitated g-FeOOH is often used as a starting material for g-Fe2O3 production. For example, S9ubrt et al. [111] produced g-Fe2O3 by the dehydration of precipitated g-FeOOH, as shown by XRD and M€ ossbauer spectroscopy. Krishnamurti and Huang [112] investigated the precipitation of g-FeOOH by aeration of Fe(ClO4)2 solution at pH 7.5 in the presence of citrate ions. The presence of citrate promoted the formation of g-FeOOH at the expense of a-FeOOH and g-Fe2O3. Thus formed g-FeOOH was stable for a long time. An analogy was drawn with the formation of g-FeOOH in the presence of citrates in nature. The effect of chlorides [113] and phosphates [114] on the precipitation of g-FeOOH from Fe2þ solutions was also investigated. Wirnsberger et al. [115] used XRD, M€ ossbauer, and EXAFS measurements to investigate the restructuring process in the precipitation system of iron oxyhydroxide–surfactant (C12H25OSO3 Naþ). The precipitation systems were prepared by mixing FeCl3 solution with sodium dodecyl sulfate, then aging. It was concluded that at short aging times (1 or 2 days) the inorganic part imparted g-FeOOH features with predominantly corner-linked (Fe(OOH)6) units,

486

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

FIGURE 23.14 57 € ssbauer spectra of sample Fe Mo obtained on heating of two-XRDline ferrihydrite at 220  C. (Reproduced from Ref. 106 with permission of Elsevier.)

whereas at longer aging times the cross-linking within the inorganic part resulted in an increase of edge-sharing octahedra, thus making local structures similar to those of a-FeOOH and b-FeOOH. Dutta et al. [116] investigated in situ nucleation and growth of g-FeOOH nanocrystallites by polymeric supramolecular assemblies. The precipitation of green rust and its transformation to g-FeOOH [117], the precipitation of g-FeOOH by oxidation of Fe(OH)2 suspension [118], and acid dissolution of g-FeOOH [119] were also investigated. Jagminas et al. [120] electrochemically prepared g-FeOOH nanowires in alumina pores with the average diameter of 45 and 150 nm. Thus obtained g-FeOOH nanowires were characterized by M€ ossbauer spectroscopy up to 7.5 K. Generally, the RT M€ ossbauer spectrum of g-FeOOH shows a paramagnetic doublet (D ¼ 0.53 mm s 1) and hyperfine magnetic splitting at 4.2 K (Bhf ¼ 45.8 T) [31]. Murad and Schwertmann [121] observed a pronounced effect of g-FeOOH crystallinity on the corresponding M€ ossbauer spectra. Due to superparamagnetic relaxation and other particle size phenomena, the M€ ossbauer spectra of low-crystalline g-FeOOH may show the onset of magnetic ordering at the temperature range from the bulk Neel temperature of 77  1 to about 50 K. At 4.2 K, the distribution of HMF with a limiting upper field of 46 T was observed. De Grave et al. [122] also investigated the M€ ossbauer effect of low-crystalline g-FeOOH between 12 and 360 K. In a magnetically ordered region, the spectra displayed broad HMF distributions and, at approximately 20 K, an additional central quadrupole doublet was included in the fitting model.

23.5 PRECIPITATION AND PROPERTIES OF SOME OTHER IRON OXIDES

487

FIGURE 23.15 57 € ssbauer spectra of sample Fe Mo obtained on heating of two-XRDline ferrihydrite at 245  C [106], recorded at various temperatures in the absence of a magnetic field (top) and at 4.2 K in the applied magnetic field of 6 T (bottom). (Reproduced from Ref. 106 with permission of Elsevier.)

23.5.3 Magnetite (Fe3O4) and Maghemite (g-Fe2O3) The precipitation of Fe3O4 and g-Fe2O3 has been extensively investigated due to possible utilization of their magnetic and some other properties (magnetic inks, ferrofluids, catalysts, sensors, etc.). Fe3O4 and g-Fe2O3 particles, specifically of nanodimensions, have found important applications in biomedicine [123–129]. Precipitation methods were extensively used in the synthesis of Fe3O4 and g-Fe2O3 particles. Sugimoto and Matijevic [130] produced spherical magnetite particles with a mean diameter between 0.03 and 1.1 mm by oxidation of Fe(OH)2 with a nitrate ion at 90  C for several hours. If oxygen was present in addition to the nitrate ion, rod-like a-FeOOH particles also precipitated. Domingo et al. [131] also investigated the precipitation of magnetite particles by oxidation of Fe(OH)2 with nitrates. The same method was used by Spiers et al. [132] to prepare magnetite particles that were further

488

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

€ ssbauer Parameters of Samples Prepared by Heating of two-XRD-Line Ferrihydrite TABLE 23.3 The Relevant Mo at Different Temperatures Preparation of Sample

T (K)

FIT

220  C

295 78 13 295

DD SD SD DD SD SD1 SD2 SD1 SD2 SD1 SD2a

245  C (zero external field)

78 12 245  C (applied field)

4.2

325  C

295 78

S S

M€ ossbauer Parameters DEQ,m ¼ 0.62 Hhf,m ¼ 40.3 Hhf,m ¼ 50.8 DEQ,m ¼ 0.59 Hhf,m ¼ 50.2 Hhf,m ¼ 52.7 Hhf,m ¼ 45.6 Hhf,m ¼ 53.5 Hhf,m ¼ 50.5 Heff,m ¼ 53.6 Heff,m1 ¼ 54.7 Heff,m2 ¼ 47.8 Hhf ¼ 51.3 Hhf ¼ 53.8

dm ¼ 0.33 dm ¼ 0.46 dm ¼ 0.47 dm ¼ 0.33 dm ¼ 0.37 dm ¼ 0.48 dm ¼ 0.46 dm ¼ 0.47 dm ¼ 0.46 dm ¼ 0.48 dm ¼ 0.48 d ¼ 0.37 d ¼ 0.48

2eQ ¼ 0.02 2eQ ¼ 0.00 2eQ ¼ 0.21 2eQ ¼ 0.15 2eQ ¼ 0.02 2eQ ¼ 0.03 2eQ ¼ 0.00 2eQ ¼ 0.08 2eQ ¼ 0.02

RA ¼ 0.77 RA ¼ 0.23 RA ¼ 0.22 RA ¼ 0.78 RA ¼ 0.22 RA ¼ 0.78 RA ¼ 0.19 RA ¼ 0.81

2eQ ¼ 0.21 2eQ ¼ 0.38

Source: Reproduced from Ref. 106 with permission of Elsevier. The spectra were recorded in zero external field or applied field of 6 T (sample prepared at 245  C) [106]. Note: DD is doublet-distribution fit, SD sextet—distribution fit (one or two components), S is a single sextet fit; DEQ,m is the maximum-probability quadrupole splitting, dm is the maximumprobability center shift, Hhf,m is the maximum-probability hyperfine field, Heff,m is the maximum-probability effective hyperfine field (applied-field spectrum), RA is the fractional spectral area, which is proportional to the number of Fe species in the corresponding phase. a The probability-distribution profile of this hyperfine-field distribution has two well-defined maxima; center shifts and quadrupole shifts 2eQ are presumed to be the same.

treated in an ultrasound field to obtain much smaller particles. Vereda et al. [133] modified the precipitation method reported earlier [130] and obtained relatively uniform magnetite particles of an average 54 nm size. The influence of a chemical route on the synthesis of magnetite nanoparticles was also investigated by other researchers [134–138]. The influence of the pH on the formation of magnetite nanoparticles [139], as well as its effect on the oxidation of precipitated magnetite to g-Fe2O3 were investigated [140]. Stachan et al. [141] prepared ellipsoidal g-Fe2O3 particles of good uniformity. Their method was based on the forced hydrolysis of Fe(III)–salt aqueous solutions in the presence of phosphates. Thus obtained and isolated a-Fe2O3 particles were subjected to reduction/oxidation processes to produce g-Fe2O3. M€ ossbauer spectroscopy has found an important application in the characterization of magnetite and g-Fe2O3. Magnetite is an inverse spinel in which Fe3þ occupies the tetrahedral A-sites, whereas the B-sites are occupied by both Fe3þ and Fe2þ ions. The structural formula of magnetite could be written as Fe(III)IV[Fe(II)Fe(III)]VIO4. Magnetite shows Curie temperature of about 850 K. Below this temperature the spin arrangement on the A- and B-sites is antiparallel. Since the magnitudes of the spins on these sites differ, magnetite exhibits ferrimagnetic behavior. Above the Vervy transition at about 119 K, electron hopping takes place between Fe2þ and Fe3þ ions at the octahedral sites. The charge transfer is rapid enough for the nucleus to sense the average charge “Fe2.5þ” at the B-sites. Therefore, the RT M€ ossbauer spectrum of magnetite shows a superposition of two sextets. A greater width of the M€ ossbauer line for the B-site can be alternatively interpreted as a result of relaxation effects due to the rate of electron hopping or the existence of two magnetically nonequivalent iron ions at the B-sites. Harker and Pollard [142] recorded M€ ossbauer spectra of natural magnetite crystals at 4.2 K and in applied fields from 0 to 14 T. The spectra were fitted to at least five ossbauer spectra of natural octahedral Fe3þ sites and four octahedral Fe2þ sites. Berry et al. [143] also recorded M€ magnetite at 4.2 K and in applied fields from 2 to 13 T along the (111) direction. The spectra were fitted to five sextets corresponding to Fe3þ at tetrahedral A-sites, whereas the other four sextets were assigned to Fe3þ and Fe2þ at two nonequivalent octahedral B-sites. M€ ossbauer spectroscopy is a very useful technique to monitor changes in ossbauer spectroscopy to investigate the reduction stoichiometry of magnetite (Fe3 xO4). Gotic et al. [144] used M€ by H2 of commercial nonstoichiometric magnetite (Fe3 xO4) containing a small fraction (7 wt.%) of a-Fe2O3 to stoichiometric Fe3O4, then the reoxidation of Fe3O4 by O2. g-Fe2O3 can be obtained by the oxidation of Fe2þ ions in Fe3O4. The structure of g-Fe2O3 corresponds to a nonstoichiometric spinel where the vacancies at the B-sites make a compensation for the oxidized Fe2þ ions. Therefore, the formula of g-Fe2O3 could be alternatively written as Fe(Fe5/3[ ]1/3)O4, where [ ] refers to the vacancies.

23.5 PRECIPITATION AND PROPERTIES OF SOME OTHER IRON OXIDES

489

Gobe et al. [145] prepared magnetite nanoparticles 3–5 nm in size using the microemulsion method. The temperature-dependent M€ ossbauer spectra of these particles exhibited superparamagnetic behavior. Gotic and Music [146] reported a synthesis of iron oxides by autoclaving iron(III)–acetate/alcohol and iron(III)–acetate/acetic acid/alcohol solutions at 180  C for varying time periods. Magnetite was formed in iron(III)–acetate/ethanol, whereas a-Fe2O3 was formed in the iron(III)–acetate/acetic acid/ethanol system. The average crystallite sizes of 11.1 and 22.6 nm and the stoichiometry of Fe2.89O4 were found. In the iron(III)–acetate/octanol system, a mixture of magnetite and a-Fe2O3 was obtained with magnetite particles up to 22 nm and a-Fe2O3 up to 46.7 nm. In the iron(III)–acetate/acetic acid/ethanol precipitation system, nanosized a-Fe2O3 particles were formed exclusively. The corresponding mechanism of the formation of nanosized iron oxide particles was discussed. Matsuda [147] investigated the precipitation of magnetite starting from the solution of FeCl2/FeCl3 and urea for 5 or 15 h at 100  C in an applied magnetic field of 10,000 G. The M€ ossbauer spectrum of the precipitate formed for 5 h showed the presence of magnetite as a single phase, whereas after 15 h of heating in the M€ ossbauer spectrum a quadrupole doublet was additionally recorded, which was tentatively assigned to g-FeOOH, g-Fe2O3 or “Fe(OH)3.” Wang et al. [148] used the solvothermal precipitation of magnetite rods also with an external magnetic field. Biddlecombe et al. [149] hydrolyzed Fe(II) alkoxide precursor Fe(OBu0 )2(THF)2 and prepared magnetite or g-Fe2O3. The particle sizes, determined using the XRD data, were 19  2 nm for magnetite and ossbauer spectra showed the presence of Fe3 xO4 and a quadrupole doublet due to the 9  2 nm for g-Fe2O3. The M€ superparamagnetic behavior of g-Fe2O3. Magnetite nanoparticles were also prepared electrochemically using Fe electrodes [150]. The precipitate obtained at ossbauer spectrum. When the voltage between Fe electrodes was 5 V corresponded to Fe3 xO4 according to the M€ above 6 V the M€ ossbauer spectrum of precipitate additionally showed an Fe0 phase. The electrochemical activity of Fe3O4 particles sized 400, 100, or 10 nm in relation to Liþ and Naþ ions in rechargeable batteries was investigated [151]. The nanosize effect was observed. Fe3O4 particles 10 nm in size prepared by the precipitation method gave 190 mAh g 1 of the rechargeable capacity in the range of 2–3 V in the lithium salt electrolyte. The rechargeable capacity of 160 or 170 mAh g 1 with excellent capacity retention in sodium electrolyte was also obtained for nanoparticles of Fe3O4 and a-Fe2O3, respectively. Satula et al. [152] investigated Fe3O4/g-Fe2O3 nanoparticles prepared by the “wet” chemical method. M€ ossbauer spectroscopy was used in the characterization of magnetite particles [153,154], magnetite and o et al. [157] g-Fe2O3 particles [155] and g-Fe2O3 particles [156] produced by different synthesis routes. Szab prepared magnetite-loaded PVA hydrogels and characterized them by magnetic and M€ ossbauer measurements. On the basis of analysis of measured data, it was possible to estimate the size distribution of the magnetic cores of magnetite particles produced by chemical precipitation and built into a chemically cross-linked polyvinyl alcohol matrix. These findings were interpreted with a core-shell model applied to fine magnetite particles built into a highly swollen polymer network in a size range of 6–15 nm. Kim et al. [158] synthesized hydrothermally magnetite and g-Fe2O3 nanoparticles. These nanoparticles were used to extract the genomic DNA from bacterial cells. g-Fe2O3 nanoparticles showed a stronger binding affinity to the purified DNA than the magnetite nanoparticles. Isolated g-Fe2O3 nanoparticles (D < 30 nm) in a silica matrix were prepared [159] by heating at 400  C the gel formed by the hydrolysis of an ethanol solution of Fe(NO3)3  9H2O and tetraethylorthosilicate (TEOS). By changing the preparation conditions g-Fe2O3 nanoparticles sized 5–15 nm could be produced. However, when FeCl3  6H2O was used as the ossbauer starting salt, well-developed a-Fe2O3 particles were obtained in the final product instead of g-Fe2O3. M€ spectroscopy was used in the characterization of these g-Fe2O3 and a-Fe2O3 particles. g-Fe2O3 nanoparticles were also prepared by Zboril et al. [160] and Santos et al. [161] and investigated by M€ ossbauer spectroscopy. Tronc et al. [162–169] utilized the capabilities of M€ ossbauer spectroscopy in the investigation of g-Fe2O3 particles, specifically their magnetic behavior in relation to the degree of the aggregation of these particles, the type of matrix, and particle size. The method of g-irradiation has been utilized in the investigation of the precipitation of magnetite and g-Fe2O3. M€ ossbauer spectroscopy also found application in these investigations. The b-FeOOH nanoparticles 5 nm in size, dispersed in a mixture of isopropanol and water, were subjected to g-irradiation [170]. In the initial stage, b-FeOOH was transformed to a-FeOOH and at a dose of 64.3 kGy magnetite nanoparticles (54 nm) were obtained as a single phase. A similar work was published by Lee and Kang [171]. Du and Liu [172] prepared superparamagnetic g-Fe2O3 particles also by using the method of g-irradiation. Xu et al. [173] noticed changes in the morphology of g-Fe2O3 nanoparticles precipitated in the presence of a PEG polymer and in the field of g-irradiation. Gotic et al. [174] found that the concentration of starting chemicals (FeCl3/FeSO4), aeration/deaeration, high alkalinity, and g-irradiation are the factors that may influence the microemulsion synthesis of magnetite nanoparticles. A very strong effect of g-irradiation on the formation of magnetite by the microemulsion method was observed. At a

490

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

460 kGy dose, nanosized magnetite particles (5–20 nm) and a very small amount of a-FeOOH particles were produced. The formation of substoichiometric magnetite (Fe3 xO4) and not of g-Fe2O3 was suggested, this being in line with the strong reductive conditions induced by g-irradiation. Gotic et al. [175] also studied the mechanism of the formation of nanosized magnetite particles in different g-doses applied to microemulsions containing only Fe3þ precursor at the beginning of g-irradiation. Fe3O4 and g-Fe2O3 may precipitate during the corrosion of iron (steel) in aqueous media or in the atmosphere. Generally, the phase composition of rust depends on the chemical composition of aqueous media and the presence of alloying elements in iron (steel). In atmospheric corrosion dry/wet cycles also play an important role in the formation of rust. In dependence on time the phase composition of rust is changing and other forms of iron oxides can take part in corrosion processes, for example, ferrihydrite, Fe(OH)2, a-FeOOH, b-FeOOH, g-FeOOH, d-FeOOH, and a-Fe2O3. xþ n x Mixed FeII–FeIII hydroxy salts known as green rust of the general formula ½FeIIð1 xÞ FeIII x ðOHÞ2 Š . [(x/n)A , yH2O] , where 2 2 n A is an anion, for example, SO4 or CO3 , were observed as corrosion products. Their chemical behavior was systematically studied by Genin et al. [176,177]. Jolivet et al. [178–180] reviewed the precipitation chemistry of metal oxides.

23.6 INFLUENCE OF CATIONS ON THE PRECIPITATION OF IRON OXIDES Natural iron oxides are commonly formed by weathering of iron-rich primary minerals (silicates, spinels) in an environment where other metal cations are also present in these processes. For this reason, natural iron oxides may contain different amounts of various cations incorporated into their structure. Due to their abundance in the Earth’s crust as well as chemical and physical properties similar to those of iron, the most common metal found in natural iron oxides is Al [181–192], but the presence of many others, such as Mg [193,194], Si [195–197], Ti [194,198], V [199], Cr [186,192,200], Mn [192,194,200,201], Co [192,200], Ni [186,192,200,202,203], Cu [200], Zn [192,200], and Ge [204] has been observed. In order to compare synthetic iron oxides with their natural counterparts, or to study their properties (magnetic, electric, catalytic, adsorptive, electrochemical, gas sensing, and so on) appropriate for commercial applications, synthetic iron oxides with incorporated various metal cations have been prepared in laboratories. The properties of synthetic iron oxides with incorporated Mg [205,206], Al [96,183,188,207–243], Ti [206,244–247], V [248,249], Cr [250–262], Mn [263–274], Co [275–285], Ni [202,233,279,286–292], Cu [256,279,293–296], Zn [279,289,297,298], Ga [299–301], Y [302], Zr [303], Nb [304,305], Ru [306,307], Rh [308], Pd [309,310], Cd [279,311–313], In [314,315], Sn [206,245,316– 318], La [319], Ce [320–322], Nd [323], Pb [279,324], U [325] have been investigated. The incorporation of these cations into the structure of iron oxides could modify their thermal [219,270,282], dissolution [267,326], electric [327,328], magnetic [233,244,257,270,271,329], catalytic [272,292,296,301,304,305,322], adsorption [324,330], gas sensing [303,309,331,332], and other properties. Generally, the presence of various amounts of metal cations during the formation of iron oxides can yield the following results: (a) incorporation of these cations into the crystal structure of an iron oxide [205,206,218,245,261,267, 269,279,316], (b) modification of the size and shape of obtained particles [233,240,243,257,284,285,295,326,333], (c) formation of iron oxides different from those in the pure system [243,293,295,302,313], (d) formation of ferrites or other mixed iron–metal oxide phases [265,269,274,285,291], (e) formation of separated iron oxide and metal oxide phases, usually with incorporated “foreign” cations into the structure [242,265,314], and (f) formation of an iron oxide along with the formation of metal by reduction of metal cations present in the system [307,308,310,334].

23.6.1 Goethite M€ ossbauer spectroscopy has been used, along with other techniques, to investigate the effect of the presence of various metal cations in the precipitation system on the formation of goethite, a-FeOOH [243,262,274,285,291,295,298,307,310,313,334]. A collective characteristic of the M€ ossbauer spectra of goethites formed in the presence of foreign metal cations is the reduction of HMF due to the incorporation of cations into the structure of a-FeOOH and, in some cases, due to the reduction in crystallite size. A high level of incorporation and/or low crystallinity of goethite samples led to the collapse of the magnetic sextet and/or the presence of a superparamagnetic doublet in their RT M€ ossbauer spectra or even in M€ ossbauer spectra recorded at lower temperatures. A

23.6 INFLUENCE OF CATIONS ON THE PRECIPITATION OF IRON OXIDES

491

FIGURE 23.16 57 € ssbauer spectra, hyperfine magnetic field distributions, and selected FE-SEM images of a-FeOOH samples Fe Mo obtained after various aging times in a highly alkaline medium in the presence of TMAH. (Reproduced from Ref. 336 with permission of Elsevier.)

reduction in HMF due to the incorporation of foreign cations depends on the type of metal cation and the extent of its incorporation into the crystal structure of goethite. In the studies of the effect of foreign cations on the formation of iron oxides and their morphological and microstructural properties, it is important to establish the system of uniform iron oxide particles as a reference sample for comparison with other samples formed under the same experimental conditions in the presence of various amounts of foreign cations. One reference sample, consisting of monodispersed a-FeOOH particles, has been obtained in a highly alkaline medium (pH  13.5) using TMAH as a precipitating agent [335,336]. The sample obtained by autoclaving at ossbauer 160  C for 2 h consists of lath-like a-FeOOH particles 150–200 nm in length (Fig. 23.16, top left). The M€ spectrum of this sample shows a magnetic sextet of asymmetric lines characteristic of a-FeOOH samples of medium crystallinity [337] with an HMF distribution somewhat extended toward lower values and an average HMF of 35.3 T. Further autoclaving of this precipitation system up to 168 h resulted in the formation of the same iron oxide phase, a-FeOOH. However, the lateral arraying of a-FeOOH particles and the growth of larger particles at the expense of smaller ones resulted in the formation of thicker and wider particles (Fig. 23.16, bottom left). The M€ ossbauer spectrum of this sample (Fig. 23.16, right) shows a magnetic sextet of very narrow lines characteristic of a-FeOOH samples of high crystallinity [337] with a nearly symmetric HMF distribution and an average HMF of 37.1 T. 23.6.1.1 Influence of Divalent Cations The presence of Mn2þ ions in the described precipitation system [335,336] significantly affected the properties of thus formed goethite particles [274]. The particles formed in the presence of 7 mol% of Mn2þ ions were elongated to 250–300 nm (Fig. 23.17, top left), whereas an average HMF was reduced to 30.9 T due to the incorporation of manganese ions, probably in the oxidized form of Mn3þ (ionic radius  0.645 A [338], the same as the radius of Fe3þ), into the structure of goethite. The presence of higher concentrations of 2þ Mn ions resulted in a further elongation (up to approx. five times) of a-(Fe,Mn)OOH particles [274] and a reduction in

492

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

FIGURE 23.17 57 € ssbauer spectra and Fe Mo FE-SEM images of a-FeOOH samples synthesized in the presence of various amounts of Mn2þ ions. (Reproduced from Ref. 274 with permission of Elsevier.)

HMF (Fig. 23.18, bottom). The presence of 23 mol% of Mn2þ ions effected a change in the morphology of a-(Fe,Mn)OOH particles from laths to stars and a collapse of magnetic ordering (a doublet in the M€ ossbauer spectrum) (Fig. 23.17, middle). In the presence of 33 mol% of Mn2þ ions larger a-(Fe,Mn)OOH rods were formed along with smaller particles of Mn-ferrite (identified by characteristic M€ ossbauer spectrum) (Fig. 23.17, bottom). In other studies [256,287], a reduction in HMF in the M€ ossbauer spectra of Mn-substituted goethite, recorded at RT and LNT, has been observed. The contribution of the doublet in the RT M€ ossbauer spectrum has been increased. The reduction in HMF in Mn-substituted goethite was smaller than in the case of Al [264,287]. a-(Fe,Co)OOH nanowires were obtained in the conditions of reference sample modified only by an addition of Co2þ ions [285]. The elongation of Co-substituted goethite particles was higher at an increased Co2þ ion concentration. A reduction in HMF due to both the incorporation of Co2þ ions into the a-FeOOH structure and thinning of particles in a direction perpendicular to the crystallographic c-axis was observed by M€ ossbauer spectroscopy (Fig. 23.18, bottom). The formation of cobalt ferrite was observed by XRD, M€ ossbauer spectroscopy, and FE-SEM at the Co2þ content of 9 mol% or higher. In a similar study of the influence of cations on the properties of goethite, Diaz et al. [287] observed that the presence of Co2þ ions in the precipitation system reduced the HMF of goethite synthesized from Co-ferrihydrite in an alkaline medium. Cornell and Givanoli [276] studied the influence of Co2þ ions on the synthesis of goethite in similar alkaline conditions and observed the formation of thin Co-goethite particles at lower and Co-magnetite at higher Co2þ concentrations. The formation of Ni-substituted goethite in the presence of Ni2þ ions was observed by way of a significant reduction in HMF (Fig. 23.18, bottom) in corresponding M€ ossbauer spectra [291]. However, in contrast to Mn2þ or Co2þ ions, the 2þ presence of Ni ions has not significantly changed the size and shape of goethite particles. A ferrite phase was detected by XRD and M€ ossbauer spectroscopy in the sample formed at the Ni2þ content of 9 mol%. In other studies [202,287], the presence of Ni2þ ions in the precipitation system also reduced the HMF of goethite synthesized from Ni-ferrihydrite in an

493

23.6 INFLUENCE OF CATIONS ON THE PRECIPITATION OF IRON OXIDES

FIGURE 23.18 Average hyperfine magnetic field in metal-substituted goethites versus initial mole fraction in precipitation systems for some trivalent (top) and divalent (bottom) cations.

alkaline medium. The incorporation of Ni2þ ions into the structure of goethite caused a slight decrease in HMF in M€ ossbauer spectra recorded at 78 K, while other parameters remained unchanged [202]. Pure Cu-substituted goethite was formed at a low content of Cu2þ ions in the precipitation system (up to 3 mol%). A mixture of Cu-substituted goethite and Cu-substituted hematite was formed between 3 and 6 mol%. At the Cu2þ ion content of 7 mol% or higher, the Cu-substituted hematite was formed [295]. A reduction in HMF was observed in both the Cu-goethite (Fig. 23.18, bottom) and Cu-hematite. Cu-goethite particles were significantly elongated in comparison with reference sample, whereas Cu-hematite particles formed a characteristic rhombohedral shape. In other M€ ossbauer spectroscopic studies [256,296], the presence of Cu2þ ions in the precipitation system affected the reduction of HMF in goethite at RT and LNT, which indicates Fe-for-Cu substitution in the structure of a-FeOOH. The presence of Zn2þ ions in the precipitation system [298] affected the formation of elongated Zn-substituted a-FeOOH particles as a single phase up to 9 mol% Zn. At the Zn content of 11 mol% Zn-ferrite particles were formed besides elongated Zn-substituted a-FeOOH particles. At 13 mol% of Zn, three phases were detected by M€ ossbauer spectroscopy: Zn-substituted goethite, hematite, and Zn-ferrite. Similar to the case of Cu2þ ions, in the presence of Cd2þ ions Cd-goethite particles were formed at low concentrations of these ions (up to 2 mol%), a mixture of Cd-substituted goethite and Cd-substituted hematite was formed between 2 and 6 mol%, and at the content of Cd2þ ions of 7 mol% or higher, the Cd-substituted hematite was formed [313]. The distribution of HMF in Cd-goethite was broadened and, at a higher concentration of Cd2þ ions the distribution maximum was shifted to lower values. The influence of Cd2þ ions on the morphology of goethite and hematite particles was similar to the case of Cu2þ ions: elongated Cd-goethite and rhombohedral Cd-hematite particles were formed. 23.6.1.2 Influence of Trivalent Cations Due to the common presence of Al-substituted a-FeOOH in natural samples (well oxidized soils), the impact of Al-substitution on the properties of natural and synthetic goethite was

494

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

comprehensively investigated by M€ ossbauer spectroscopy [96,183–185,189,190,211,213,222,223,231,238,243]. A rapid decrease in HMF was observed with an increase in Al-substitution (Fig. 23.18, top) [183,211,243]. The decrease in HMF of the Al-substituted goethites was attributed to the drop in Neel temperature due to a dilution of the magnetic field by nonmagnetic Al3þ ions and, in some samples, smaller particle size in comparison with unsubstituted goethite. A complete collapse of the magnetic sextet to a broadened doublet in RT M€ ossbauer spectra was observed at 16 mol% [213], 13 mol % [183], 12 mol% [215], or 7.7 mol% [211] of Al-substitution, depending on sample crystallinity. At 77 K almost all Alsubstituted goethite samples showed a magnetic sextet in the M€ ossbauer spectra. A sample with a very high Al substitution (30 mol%) and of a very low crystallinity was needed to record a collapsed magnetic sextet [211,223]. All spectra of Al-substituted goethite recorded at 4.2 K were magnetic sextets with HMF linearly decreased with the level of Al-substitution [213]. Murad and Schwertmann [215] showed that even at 4.2 K HMF was reduced by both the Alsubstitution and the low crystallinity. Barrero et al. [231] showed that the Neel temperature, the saturation HMF, and the intersublattice exchange interaction in goethite decrease with a lower crystallinity and higher Al-content. In the presence of higher concentrations of Al ions in the precipitation system Al-substituted hematite was formed in addition to Alsubstituted goethite, depending on the temperature of the precipitation system (above 19 mol% at 60  C, 15 mol% at 135  C, and 13 mol% at 160  C) [213,243]. The influence of the V3þ ion presence during the formation of goethite from ferrihydrite in an alkaline medium was investigated using M€ ossbauer spectroscopy and other techniques by Kaur et al. [249]. The presence of V3þ ions reduced  HMF values in the M€ ossbauer spectra due to a substitution of Fe3þ by V3þ ions (ionic radius of 0.64 A almost equal to the radius of Fe3þ) in the goethite structure. Small quantities of hematite and superparamagnetic goethite were also formed. The presence of Cr3þ ions in the precipitation system [262] influenced the formation of thinner a-FeOOH lath-like particles. The average HMF was reduced to 31.4 T in the sample prepared in the presence of 9 mol% Cr (Fig. 23.18, top), indicating a significant incorporation of chromium ions into the structure of a-FeOOH. Further increase in the concentration of Cr3þ ions has not changed HMF. The remaining Cr3þ ions were incorporated into the structure of Crferrihydrite, as exemplified by the doublet in the M€ ossbauer spectrum and characteristic broad XRD line. In another study [254], the influence of the presence of Cr3þ ions during the formation of goethite from ferrihydrite in an alkaline medium was also investigated using M€ ossbauer spectroscopy. The presence of an appreciable concentration of chromium ions (7 wt.% and more) during the formation of goethite resulted in an asymmetric doublet in the M€ ossbauer spectrum recorded at 300 K due to a reduction in size of particles and incorporation of chromium ions into the structure of a-FeOOH. The spectra recorded at 77 K consisted of a magnetic sextet for all Cr-doped a-FeOOH samples, but HMF was significantly reduced at a higher chromium ion concentration. A similar reduction in HMF due to the incorporation of chromium ions into the structure of a-FeOOH and a reduction in particle size were observed in the case of lowcrystalline goethite synthesized from ferrous solution in a neutral medium [256].  Due to the similar ionic radius (0.62 A [338]) and the existence of isostructural oxyhydroxide (GaOOH), Ga3þ ions substituted Fe3þ in the structure of goethite in a broad concentration range [300]. Ga-for-Fe substitution in the structure of goethite affects the reduction in HMF and Neel temperature due to magnetic dilution by nonmagnetic Ga3þ ions and to the formation of small Ga-goethite particles. 23.6.1.3 Influence of Tetravalent Cations The influence of the Ti4þ ion presence during the formation of goethite from ferric solution in an alkaline medium was investigated by M€ ossbauer spectroscopy and other techniques [246]. The presence of Ti4þ ions partially suppressed the transformation of ferrihydrite to goethite and as a result a quadrupole doublet, corresponding to the low-crystalline phase, emerged in the M€ ossbauer spectrum. ossbauer spectroscopy by Sn4þ doped a-FeOOH samples of various crystallinity have been investigated using M€ Berry et al. [317]. The lines in the M€ ossbauer spectra recorded at room temperature were broadened and the HMF values were decreased due to Sn4þ ion incorporation and deviation from perfect crystallinity. The spectra recorded at 80 K were less affected by differences in crystallinity of samples, while the spectra recorded at 17 K were almost unaffected. 23.6.1.4 Influence of Platinum Group Metal Cations Monodispersed lath-like goethite particles described in Section 23.6.1 were used in the studies of the influence of the presence of platinum group metal cations (Ru3þ, Rh3þ, Pd2þ) during the formation of these particles on the properties of the obtained products [307,308,310,334]. The presence of these cations influenced the formation of hematite besides goethite in this system. Hematite/goethite ratio increased at a higher concentration of added metal cations in the system. The reduction of HMF in a-FeOOH formed in the presence of Ru3þ [307] and Rh3þ [308] indicates a significant incorporation of these cations into the structure of  a-FeOOH, which is possible due to the similar ionic radii (0.68 and 0.665 A, respectively [338]) and the same charge of

495

23.6 INFLUENCE OF CATIONS ON THE PRECIPITATION OF IRON OXIDES

FIGURE 23.19 57 € ssbauer spectra and selected FE-SEM images of iron oxide samples synthesized in the presence of 1.33 mol% Fe Mo of Rh3þ ions. (Reproduced from Ref. 308 with permission of Elsevier.) 



these cations and Fe3þ (ionic radius 0.645 A). The difference in ionic radius (0.86 A) and charge prevent the incorporation of a significant amount of Pd2þ ions into the structure of a-FeOOH [310]. The strong catalytic activity of platinum group metals has a significant impact on the properties of the solid products obtained in above described precipitation systems after longer hydrothermal treatment (9–72 h) at 160  C. In the course of hydrothermal treatment, the products of TMAH decomposition (methanol and trimethylamine) affect the reduction of metal cations and the formation of platinum group metal nanoparticles. These nanoparticles act as the catalyst for the reduction of Fe(III) to Fe(II) by the products of TMAH decomposition and the formation of characteristic Fe3O4 octahedra, which is followed by M€ ossbauer spectroscopy and FE-SEM (Fig. 23.19). 23.6.2 Hematite The influence of Al-substitution on the properties of hematite was comprehensively investigated by M€ ossbauer spectroscopy [212,216,220–222,228,234–236,239,242]. Substitution of Fe3þ by Al3þ ions in hematite influenced the reduction of HMF observed at various temperatures and Morin transition temperature [212,235,236]. Phase composition of ball-milled system a-Fe2O3-a-Al2O3 depends on the starting concentration of phases and on the milling velocity [239]. Al-doped a-Fe2O3 and Fe-doped a-Al2O3 were formed by ball milling and heating of mixtures of a-Fe2O3 and g-AlOOH [242]. ossbauer lines Substitution of Fe3þ ions by Cr3þ ions in hematite affected the reduction of HMF and broadening of M€ [253,255], as well as lowering of the Neel temperature [252,261]. The Morin transition was observed only for low substituted Cr-hematites—in samples with 4.3 mol% of substituted Cr or more the Morin transition no longer occurred [261]. Mn-for-Fe substitution in hematite influenced a reduction in HMF and has a considerable effect on the Morin transition—a weak ferromagnetic state was observed at 80 K for the substitution >4 mol% [264]. The presence of Zn2þ ions in the hydrothermal precipitation system of ferric solution in an alkaline medium resulted in the formation of Zn-doped a-Fe2O3 and ZnFe2O4 [297]. The spectrum corresponding to a-Fe2O3 was fitted to a component with a larger HMF (51.5 T), associated with Fe3þ ions that are not influenced by the presence of zinc, and a component with a smaller HMF (50.5 T), which can be assigned to Fe3þ ions that have Zn2þ ions in a close structural environment. The Zn2þ solubility limit in a-Fe2O3 was estimated at 2.88%. The Morin transition is completely ossbauer suppressed by the presence of zinc in a-Fe2O3. Hematite (sextet) and Zn-ferrite (doublet) were identified by M€ spectroscopy in the products obtained by the combustion of Fe(III)- and Zn(II)-oxinates [289]. 57 Fe and 151 Eu M€ ossbauer spectroscopies were used for the investigation of oxide phases in the system Eu2O3– Fe2O3 [339]. Due to a big difference in the ionic radii of Fe3þ (0.645 A) and Eu3þ (0.95 A), Eu-for-Fe substitution in

496

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

hematite has not been observed. Orthorhombic europium orthoferrite (EuFeO3) and cubic europium iron garnet (Eu3Fe5O12) were formed instead, depending on the initial conditions. Berry et al. [206] investigated the structural and magnetic properties of Sn-, Ti-, and Mg-substituted a-Fe2O3. Ti4þ ions in Ti-substituted hematite (10% Ti), which occupy both the substitutional and interstitial sites, reduced the HMF value determined from the M€ ossbauer spectrum recorded at 298 K. Sn4þ ions in Sn-substituted hematite (10% Sn), which occupy both the substitutional and interstitial sites, reduced the HMF and Neel temperature to 890–910 K. Mg2þ ions in Mg-substituted hematite (10% Mg), which occupy both the substitutional and interstitial sites, also affected the reduction in HMF and lowering of Neel temperature to 910–930 K, as compared with 950–960 K for pure a-Fe2O3. 23.6.3 Magnetite and Maghemite Incorporation of Mn2þ ions into the structure of magnetite was monitored by M€ ossbauer spectroscopy and XRD [272]. Mn-for-Fe substitution affected the increase of the unit cell parameters and reduction in the HMF. At a low content, Mn2þ ions substituted iron mainly in octahedral sites, whereas at higher concentrations substitutions at tetrahedral sites also occurred. The influence of the Co2þ ion concentration on the formation of magnetite by the oxidation of mixed Fe(II)–Co(II) hydroxide suspensions was investigated by M€ ossbauer spectroscopy and other techniques [283]. Higher Co2þ ion concentration affects the exchange of part of Fe3þ ions in the octahedral position with Fe2þ ions and the formation of a small quantity of metal Fe. Co-for-Fe substitution in the octahedral position affected a linear reduction in Curie temperature from 862 K for unsubstituted Fe3O4 to 816 K for Co0.8Fe2.2O4 due to the weaker intersublattice Co–Fe magnetic exchange interactions as compared with the Fe–Fe interactions [281]. A linear dependence of the isomer shift on Co-substitution was observed at tetrahedral and octahedral sites in the M€ ossbauer spectra of Co-substituted magnetites recorded above Curie temperature (at 860 K). HMF values at tetrahedral and octahedral sites increased with a higher Co-substitution. Co-for-Fe substitution in the octahedral position affected the emergence of various octahedral sites of Fe ions, depending on the number of Co ions in the immediate vicinity of the site [280,281]. The influence of the Ni2þ ion concentration on the formation of magnetite by the oxidation of mixed Fe(II)–Ni(II) hydroxide suspensions was investigated by M€ ossbauer spectroscopy and other techniques [288]. Higher Ni2þ ion concentration affected the exchange of part of Fe3þ ions in the octahedral position by Fe2þ ions. The incorporation of Ni2þ ions was higher in surface layers, and this would be a reason for the high stability to air oxidation of Fe3O4 formed in the presence of Ni2þ ions. The incorporation of Ni2þ ions into the structure of magnetite affected the decrease in intensity of the M€ ossbauer sextet corresponding to Fe2þ/3þ ions in octahedral positions in comparison with the sextet corresponding to Fe3þ ions [290]. Substitution of Fe3þ with Cd2þ and Zn2þ ions in tetrahedral sites of magnetite reduced the HMF of both the octahedral and tetrahedral iron [340]. Substitution with Cr3þ ions at octahedral sites in magnetite reduced the HMF values of iron at octahedral as well as at tetrahedral sites [250]. This substitution also reduced Curie temperature from 850 to 795 K [341]. Substitution of Fe3þ by Ga3þ ions in magnetite (Fe2.9Ga0.1O4) reduced Curie temperature from 850 to 785 K [341], whereas the substitution of Fe3þ by Sn4þ ions at octahedral sites of magnetite (Fe2.9Sn0.1O4) reduced Curie temperature from 850 to 770 K [318]. Incorporation of diamagnetic Zn2þ cations into the maghemite nanoparticles was investigated by Drofenik et al. [342] using M€ ossbauer spectroscopy and other techniques. Maghemite nanoparticles retained the lattice constant typical of maghemite and a relatively high magnetization per formula unit up to the composition close to zinc ferrite. Incorporation of Mn2þ ions into the structure of maghemite influenced an increase of the unit cell parameters and a reduction in the HMF observed in the M€ ossbauer spectra [272]. Substitution of Fe3þ by Al3þ ions in maghemite reduced the HMF [240] and Curie temperature (calculated from temperature dependence of M€ ossbauer spectra) from 948 K in unsubstituted maghemite to 730 K in sample containing 22 mol% Al [230]. Substitution of Fe3þ by Sn4þ ions at octahedral sites of maghemite (g-Fe1.9Sn0.1O3) caused a reduction in Curie temperature from 840–860 to 750–830 K [316,318]. The ossbauer spectra recorded in the presence of a position of Sn4þ and Ti4þ ions at octahedral sites was revealed by 57 Fe M€ longitudinal magnetic field, as well as by the Sn- and Ti-K-edge EXAFS [247].

ACKNOWLEDGMENT The authors wish to thank Elsevier Publ. Co. for permission to reproduce Figs. 23.2–23.19.

REFERENCES

497

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

B.O.A. Hedstr€ om, Ark. Kemi 1953, 6, 1–16. J.M. Kundsen, E. Larsen, J.E. Moreira, O. Faurskov Nielsen, Acta Chem. Scand. A 1975, 29, 833–839. J.E. Singley, J.H. Sullivan Jr., J. Am. Water Works Assoc. 1969, 61, 190–192. L. Ciavatta, M. Grimaldi, J. Inorg. Nucl. Chem. 1975, 37, 163–169. G.H. Khoe, P.L. Brown, R.N. Sylva, R.G. Robins, J. Chem. Soc., Dalton Trans. 1986, 1901–1906. W. Schneider, Comments Inorg. Chem. 1984, 3, 205–223. A. Vertes, M.R. Komor, M. Suba, Ann. Univ. Sci. Budap. Rolando E€otv€os Nominatae, Sect. Chim. 1971, 12, 79–93. A. Vertes, F. Parak, Acta Chim. Acad. Sci. Hung. 1972, 74, 293–313. A. Vertes, M. Suba, M.R. Komor, Ann. Univ. Sci. Budap. Rolando E€otv€os Nominatae, Sect. Chim. 1971, 12, 69–77. A. Vertes, M. Ranogajec-Komor, P. Gelencser, Acta Chim. Acad. Sci. Hung. 1973, 77, 55–67. A. Meagher, Inorg. Chim. Acta 1988, 146, 19–23. E. Conforto, H.R. Rechenberg, M. Jafelici Jr., J. Phys. Chem. Solids 1986, 47, 1179–1184. K. Kovacs, V.K. Sharma, A.A. Kamnev, E. Kuzmann, Z. Homonnay, A. Vertes, Struct. Chem. 2008, 19, 109–114. F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 3rd edn, Wiley, New York, 1972, p. 712. W.N. Perera, G. Hefter, Inorg. Chem. 2003, 42, 5917–5923. P. Sipos, D. Zeller, E. Kuzmann, A. Vertes, Z. Homonnay, M. Walczak, S.E. Canton, Dalton Trans. 2008, 5603–5611. A. Kamnev, B.B. Ezhov, N.S. Kopelev, Y.M. Kiselev, Y.D. Perfilyev, Electrochim. Acta 1991, 36, 1253–1257. T.G. Spiro, S.E. Allerton, J. Renner, A. Terzis, R. Bils, P. Saltman, J. Am. Chem. Soc. 1966, 88, 2721–2726. G.W. Brady, C.R. Kurkjian, E.F.X. Lyden, M.B. Robin, P. Saltman, T. Spiro, A. Terzis, Biochemistry 1968, 7, 2185–2192. B.A. Sommer, D.W. Margerum, J. Renner, P. Saltman, T. Spiro, Bioinorg. Chem. 1973, 2, 295–305. D.L. Segal, J. Chem. Tech. Biotechnol. 1984, 34A, 355–362. R.M. Cornell, U. Schwertmann, The Iron Oxides, Structure, Properties, Reactions, Occurrences and Uses, Wiley-VCH, Weinheim, 2003. J. Dousma, P.L. De Bruyn, J. Colloid Interface Sci. 1976, 56, 527–539. S. Music, A. S9aric, K. Nomura, S. Popovic, ACH-Models Chem. 1999, 136, 457–476. A. S9aric, S. Music, K. Nomura, S. Popovic, J. Mol. Struct. 1999, 480–481, 633–636. A.N. Christensen, T.R. Jensen, C.R.H. Bahl, E. DiMasi, J. Solid State Chem. 2007, 180, 1431–1435. C.A. Barrero, J.D. Betancur, J.M. Greneche, G.F. Goya, T.S. Berqu o, Geophys. J. Int. 2006, 164, 331–339. E. De Grave, R. Vochten, O. Quenard, E. Van San, H. Desseyn, A. Rousset, Nanostruct. Mater. 1999, 11, 493–504. S. Music, I. Czak o-Nagy, S. Popovic, A. Vertes, M. Tonkovic, Croat. Chem. Acta 1986, 59, 833–851. S. Music, S. Popovic, M. Gotic, Croat. Chem. Acta 1987, 60, 661–675. (a) E. Murad, J.H. Johnston, Iron oxides and oxyhydroxides, in M€ossbauer Spectroscopy Applied to Inorganic Chemistry, G.J. Long, ed., Plenum Publishing Corporation, New York, 1987, pp. 507–582. (b) E. Murad, in Properties and Behavior of Iron Oxides as Determined by M€ossbauer spectroscopy in Iron in Soils and Clay Minerals, J.W. Stucki, et al., eds., D. Reidel Publ. Co, 1988, pp. 309–350. A. Cabral-Prieto, A.A. Reyes-Felipe, M.G. Siles-Dotor, Nanostruct. Mater. 1998, 10, 311–326. M. Gotic, S. Music, J. Mol. Struct. 2007, 834–836, 445–453. R. Frayer, A. Gildawie, R. Paterson, Eighth International Congress on Electron Microscopy, Vol. 1, Canberra, 1974, pp. 708–709. E. Matijevic, P. Scheiner, J. Colloid Interface Sci. 1978, 63, 509–524. S. Music, A. Vertes, G.W. Simmons, I. Czak o-Nagy, H. Leidheiser Jr., J. Colloid Interface Sci. 1982, 85, 256–266. J.Y. Bottero, A. Manceau, F. Villeras, D. Tchoubar, Langmuir 1994, 10, 316–319. G.H. Khoe, R.G. Robins, J. Colloids Interface Sci. 1989, 133, 244–252. M. Gotic, S. Popovic, N. Ljube9sic, S. Music, J. Mater. Sci. 1994, 29, 2474–2480. S. Music, S. Krehula, S. Popovic, Mater. Lett. 2004, 58, 2640–2645. S. Go~ni-Elizalde, Ma.E. Garcia-Clavel, Ma.I. Tejedor-Tejedor, React. Solids 1987, 3, 139–154. T. Sugimoto, A. Muramatsu, J. Colloids Interface Sci. 1996, 184, 626–638. D.G. Chambaere, E. De Grave, Phys. Chem. Miner. 1985, 12, 176–184. A. S9aric, K. Nomura, S. Popovic, N. Ljubes9ic, S. Music, Mater. Chem. Phys. 1998, 52, 214–220. S. Music, S. Krehula, S. Popovic, Z. Skoko, Mater. Lett. 2003, 57, 1096–1102.

498

46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92.

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

S. Music, I. Czak o-Nagy, I. Salaj-Obelic, N. Ljube9sic, Mater. Lett. 1997, 32, 301–305. S. Music, M. Maljkovic, I. Czak o-Nagy, Mater. Lett. 1997, 31, 43–48. S. Music, M. Maljkovic, S. Popovic, ACH-Models Chem. 1999, 136, 299–316. D. Chambaere, A. Govaert, J. De Sitter, E. De Grave, Solid State Commun. 1978, 26, 657–659. D.G. Chambaere, E. De Grave, Phys. Stat. Sol. (a) 1984, 83, 93–102. J.H. Johnston, N.E. Logan, J. Chem. Soc., Dalton Trans. 1979, 13, 13–16. C. Childs, B. Goodman, E. Paterson, F. Woodhams, Aust. J. Chem. 1980, 33, 15–26. D.G. Chambaere, E. De Grave, R.L. Vanleerberghe, V.E. Vanderberghe, Hyperfine Interact. 1984, 20, 249–262. M. Ohyabu, U. Ujihira, J. Inorg. Nucl. Chem. 1981, 43, 3125–3129. C.A. Barrero, K.E. Garcia, A.L. Morales, S. Kodjikian, J.M. Greneche, J. Phys. Condens. Matter 2006, 18, 6827–6840. K.E. Garcia, A.L. Morales, C.A. Barrero, C.E. Arroyave, J.M. Greneche, Physica B 2004, 354, 187–190. MAUD (Material Analysis Using Diffraction), The program written by Luca Lutterotti, Universita di Trento, Italy, Version: 1.84, 2002. C.A. Barrero, K.E. Garcia, A.L. Morales, J.M. Greneche, Physica B 2007, 389, 88–93. K.E. Garcia, C.A. Barrero, A.L. Morales, J.M. Greneche, Mater. Chem. Phys. 2008, 112, 120–126. T. Kamimura, S. Nasu, T. Segi, T. Tazaki, H. Miyuki, S. Morimoto, T. Kudo, Corrosion Sci. 2005, 47, 2531–2542. Y.X. Yang, M.L. Liu, H. Zhu, Y.R. Chen, G.J. Mu, X.N. Liu, Y.Q. Jia, J. Magn. Magn. Mater. 2008, 320, L132–L136. F. Funk, G.J. Long, D. Hautot, R. B€uchi, I. Christl, P.G. Weidler, Hyperfine Interact. 2001, 136, 73–95. P.H. Hsu, S.E. Ragone, J. Soil Sci. 1972, 23, 17–31. P.H. Hsu, J. Soil Sci. 1972, 23, 409–419. P.H. Hsu, Clays Clay Miner. 1973, 21, 267–277. P.H. Hsu, M.K. Wang, Soil Sci. Soc. Am. J. 1980, 44, 143–149. M. Lorenz, G. Kempe, J. Signalaufz.-Mater. 1983, 11, 69–78. P.J. Murphy, A.M. Posner, J.P. Quirk, J. Colloid Interface Sci. 1976, 56, 298–311. M. Ristic, S. Music, M. Godec, J. Alloys Compd. 2006, 417, 292–299. B. Weckler, H.D. Lutz, Eur. J. Solid State Inorg. Chem. 1998, 35, 531–544. I. Opac9ak, M. Ristic, S. Music, Mater. Lett. 2010, 64, 2555–2558. J. Dousma, D. Den Ottelander, P.L. De Bruyn, J. Inorg. Nucl. Chem. 1979, 41, 1565–1568. E. Matijevic, R.S. Sapieszko, J.B. Melville, J. Colloid Interface Sci. 1975, 50, 567–581. R.S. Sapieszko, R.C. Patel, E. Matijevic, J. Phys. Chem. 1977, 81, 1061–1068. J. S9ubrt, J. Bohac9ek, V. S9tengl, T. Grygar, P. Bezdic9ka, Mater. Res. Bull. 1999, 34, 905–914. B. Voigt, A. G€ obler, Cryst. Res. Technol. 1986, 21, 1177–1183. S. Music, Z. Orehovec, S. Popovic, I. Czak o-Nagy, J. Mater. Sci. 1994, 29, 1991–1998. A. Leclerc, Phys. Chem. Mineral. 1980, 6, 327–334. J.R. Gancedo, M.L. Martinez, in Magnetic Resonance in Colloid and Interface Science, J.P. Fraissard, H.A. Resing, eds., Reidel, 1980, p. 371. K. Kandori, T. Shigetomi, T. Ishikawa, Colloids Surf. A: Physicochem. Eng. Asp. 2004, 232, 19–28. S. Music, S. Popovic, Z. Orehovec, I. Czak o-Nagy, J. Colloid Interface Sci. 1993, 160, 479–482. T. Sugimoto, K. Sakata, J. Colloid Interface Sci. 1992, 152, 587–590. T. Sugimoto, M.M. Khan, A. Muramatsu, Colloids Surf. A: Physicochem. Eng. Asp. 1993, 70, 167–169. 9 ic, M. Ristic, S. Music, J. Alloys Compd. 2008, 464, 81–88. M. Z 9 ic, M. Ristic, S. Music, Mater. Chem. Phys. 2010, 120, 160–166. M. Z 9 ic, M. Ristic, S. Music, J. Mol. Struct. 2009, 924–926, 235–242. M. Z 9 ic, M. Ristic, S. Music, J. Alloys Compd. 2008, 466, 498–506. M. Z J.L. Jambor, J.E. Dutrizac, Chem. Rev. 1998, 98, 2549–2585. K.M. Towe, W.F. Bradley, J. Colloid Interface Sci. 1967, 24, 384–392. F.V. Chukrov, B.B. Zvyagin, L.P. Ermilova, A.I. Gorshkov, Proceedings of the International Clay Conference, Madrid (Spain), 1972, pp. 333–341. J.D. Russel, Clay Miner. 1979, 14, 109–114. E. Murad, U. Schwertmann, Am. Mineral. 1980, 65, 1044–1049.

REFERENCES

93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137.

499

E. Murad, Hyperfine Interact. 1989, 47, 33–53. C.M. Cardile, Clays Clay Miner. 1988, 36, 33–53. E. Murad, L.H. Bowen, G.J. Long, T.G. Quin, Clay Miner. 1988, 23, 161–173. R.J. Pollard, C.M. Cardile, D.G. Lewis, L.J. Brown, Clay Miner. 1992, 27, 57–71. E. De Grave, G.M. Da Costa, L.H. Bowen, C.A. Barrero, R.E. Vandenberghe, Hyperfine Interact. 1998, 117, 245–270. M.B. Madsen, S. Morup, C.J.W. Koch, Hyperfine Interact. 1986, 27, 329–332. J.M.D. Coey, P.W. Readmann, Earth Planet. Sci. Lett. 1973, 21, 45–51. Q.A. Pankhurst, R.J. Pollard, J. Phys. Condens. Matter 1993, 5, 7301–7306. Q.A. Pankhurst, R.J. Pollard, Clays Clay Miner. 1992, 40, 268–272. H. Stanjek, P.G. Weidler, Clay Miner. 1992, 27, 397–412. P.G. Weidler, H. Stanjek, Clay Miner. 1998, 33, 277–284. C. Greffie, M. Amouric, C. Parron, Clay Miner. 2001, 36, 381–387. J.C. Chadwick, D.H. Jones, M.F. Thomas, C.J. Tatlock, M. Devenish, Hyperfine Interact. 1986, 28, 541–544. M. Ristic, E. De Grave, S. Music, S. Popovic, Z. Orehovec, J. Mol. Struct. 2007, 834–836, 454–460. U. Schwertmann, E. Murad, Clays Clay Miner. 1983, 31, 277–284. U. Schwertmann, J. Friedl, H. Stanjek, J. Colloid Interface Sci. 1999, 209, 215–223. J. Zhao, F.E. Huggins, Z. Feng, G. Huffman, Clays Clay Miner. 1994, 42, 737–746. D. Andreeva, I. Mitov, T. Tabakova, V. Mitrov, A. Andreev, Mater. Chem. Phys. 1995, 41, 146–149. J. S9ubrt, F. Hanousek, V. Zapletal, J. Lipka, M. Hucl, J. Thermal Anal. 1981, 20, 61–69. G.S.R. Krishnamurti, P.M. Huang, Soil Sci. Soc. Am. J. 1993, 57, 861–867. R.M. Taylor, Clays Clay Miner. 1984, 3, 175–180. J. Cumplido, V. Barr on, J. Torrent, Clays Clay Miner. 2000, 48, 503–510. G. Wirnsberger, K. Gatterer, H.P. Fritzer, W. Grogger, B. Pillep, P. Behrens, M.F. Hansen, C. Bender Koch, Chem. Mater. 2001, 13, 1467–1472. A.K. Dutta, G. Jarero, L. Zhang, P. Stroeve, Chem. Mater. 2000, 12, 176–181. U. Schwertmann, H. Fechter, Clay Miner. 1994, 29, 87–92. R. Lin, R.L. Spicer, F.L. Tungate, B.H. Davis, Colloids Surf. A: Physicochem. Eng. Asp. 1996, 113, 79–96. R.M. Cornell, R. Giovanoli, Clays Clay Miner. 1988, 36, 385–390. A. Jagminas, K. Maz9eika, E. Ju9ska, J. Reklaitis, D. Baltr^ unas, Appl. Surf. Sci. 2010, 256, 3993–3996. E. Murad, U. Schwertmann, Mineral. Mag. 1984, 48, 507–511. E. De Grave, R.M. Persoons, D.G. Chambaere, R.E. Vandenberghe, L.H. Bowen, Phys. Chem. Miner. 1986, 13, 61–67. M. Mitsumori, M. Hiraoka, T. Shibata, Y. Okuno, S. Masunaga, M. Koishi, K. Okajima, Y. Nagata, Y. Nishimura, M. Abe, K. Ohura, M. Hasegawa, H. Nagae, Y. Ebisawa, Int. J. Hyperthermia 1994, 10, 785–793. M. Forsman, Magnetism in gastroenterology, in Magnetism in Medicine, A Handbook, W. Andre€a, H. Nowak, eds., Wiley-VCH, 1988, pp. 430–445. S. Mornet, A. Vekris, J. Bonnet, E. Duguet, F. Grasset, J.-H. Choy, J. Portier, Mater. Lett. 2000, 4, 183–188. C. Lang, D. Sch€uler, D. Faivre, Macromol. Biosci. 2007, 7, 144–151. F.-J. Cheng, C.-H. Su, Y.-S. Yang, C.-S. Yeh, C.-Y. Tsai, C.-L. Wu, M.-T. Wu, D.-B. Shieh, Biomater. 2005, 26, 729–738. P.P. Macaroff, F.L. Primo, R. Bentes De Azevedo, Z.G.M. Lacava, P.C. Morais, A.C. Tedesco, IEEE 2006, 42, 3596–3598. F.L. Primo, L. Michieleto, M.A.M. Rodrigues, P.P. Macaroff, P.C. Morais, Z.G.M. Lacava, M.V.L.B. Bentley, A.C. Todesco, J. Magn. Magn. Mater. 2007, 311, 354–357. T. Sugimoto, E. Matijevic, J. Colloid Interface Sci. 1980, 74, 227–243. C. Domingo, R. Rodriguez-Clemente, M. Blesa, J. Colloid Interface Sci. 1994, 165, 244–252. K.M. Spiers, J.S. Forsythe, K. Suzuki, J.D. Cashion, J. Magn. Magn. Mater. 2007, 311, 97–100. F. Vereda, J. De Vicente, M. del Puerto Morales, F. Rull, R. Hidalgo-Alvarez, J. Phys. Chem. C 2008, 112, 5843–5849. G. Marinescu, L. Patron, D.C. Culita, C. Neagoe, C.I. Lepadatu, I. Balint, L. Bessais, C.B. Cizmas, J. Nanoparticle Res. 2006, 8, 1045– 1051. J. Salado, M. Insausti, I. Gil de Muro, L. Lezama, T. Rojo, J. Non-Cryst. Solids 2008, 354, 5207–5209. K.M. Lee, C.M. Sorensen, K.J. Klabunde, G.C. Hadjipanayis, IEEE Trans. Magn. 1992, 28, 3180–3182. J. Xu, H. Yang, W. Fu, K. Du, Y. Sui, J. Chen, Y. Zeng, M. Li, G. Zou, J. Magn. Magn. Mater. 2007, 309, 307–311.

500

138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180.

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

S. Franger, P. Perthet, J. Berthon, J. Solid State Electrochem. 2004, 8, 218–223. G. Gnanaprakash, S. Mahadevan, T. Jayakumar, P. Kalyanasundaram, J. Philip, B. Raj, Mater. Chem. Phys. 2007, 103, 168–175. J.R. Correa, D. Canetti, R. Castillo, J.C. Ll opiz, J. Dufour, Mater. Res. Bull. 2006, 41, 703–713. M. Stachen, M.P. Morales, M. Oca~na, C.J. Serna, Phys. Chem. Chem. Phys. 1999, 1, 4465–4471. S.J. Harker, R.J. Pollard, Nucl. Instrum. Meth. Phys. Res. B 1993, 76, 61–63. F.J. Berry, S. Skinner, M.F. Thomas, J. Phys. Condens. Matter 1998, 10, 215–220. M. Gotic, G. Ko9sc9ec, S. Music, J. Mol. Struct. 2009, 924–926, 347–354. M. Gobe, K. Kandori, K. Kon-No, A. Kitahara, J. Jpn. Soc. Colour Mater. 1984, 57, 380–385. M. Gotic, S. Music, Eur. J. Inorg. Chem. 2008, 966–973. K. Matsuda, Proc. Fac. Sci., Tokai Univ. 1984, 19, 69–73. J. Wang, Y. Wu, Y. Zhu, Mater. Chem. Phys. 2007, 106, 1–4. G.B. Biddlecombe, Y.K. Gun’ko, J.M. Kelly, S.C. Pillai, J.M.D. Coey, M. Venkatesan, A.P. Douvalis, J. Mater. Chem. 2001, 11, 2937–2939. L. Cabrera, S. Gutierrez, N. Menendez, M.P. Morales, P. Herrasti, Electrochim. Acta 2008, 53, 3436–3441. S. Komaba, T. Mikumo, N. Yabuuchi, A. Ogata, H. Yoshida, Y. Yamada, J. Electrochem. Soc. 2010, 157, A60–A65. nski, J. Kozubowski, Acta Phys. Pol. A 2008, 114, 1615–1621. D. Satula, B. Kalaska-Szostko, K. Szymanski, L. Dobrzy J. Wang, M. Yao, G. Xu, P. Cui, J. Zhao, Mater. Chem. Phys. 2009, 113, 6–9. L.-L. Wang, J.-S. Jiang, Nanoscale Res. Lett. 2009, 4, 1439–1446. A.B. Bourlinos, A. Bakandritsos, V. Georgakilas, V. Tzitzios, D. Petridis, J. Mater. Sci. 2006, 41, 5250–5256. A. Kong. H. Wang, J. Li, Y. Shan, Mater. Lett. 2007, 62, 943–945. D. Szab o, I. Czak o-Nagy, M. Zrinyi, A. Vertes, J. Colloid Interface Sci. 2000, 221, 166–172. Y.-S. Kim, K.-C. Kim, T.-W. Hong, Met. Mater. Int. 2010, 16, 225–228. F. Del Monte, M.P. Morales, D. Levy, A. Fernandez, M. Ocan9a, A. Roig, E. Molins, K. O’Grady, C.J. Serna, Langmuir 1997, 13, 3627–3634. R. Zboril, A. Bakandritsos, M. Mashlan, V. Tzitzios, P. Dallas, Ch. Trapalis, D. Petridis, Nanotechnology, 2008, 19(9), doi: 10.1088/ 0957-4484/19/9/095602. J.G. Santos, L.B. Silveira, A.C. Oliveira, V.K. Garg, B.M. Lacava, A.C. Tedesco, P.C. Morais, Hyperfine Interact. 2007, 175, 71–75. S. Brice-Profeta, M.-A. Arrio, E. Tronc, N. Menguy, I. Letard, C. Cartier Dit Moulin, M. Nogues, C. Chaneac, J.P. Jolivet, Ph. Sainctavit, J. Magn. Magn. Mater. 2005, 288, 354–365. E. Tronc, D. Fiorani, M. Nogues, A.M. Testa, F. Lucari, F. d’Orazio, J.M. Grenche, W. Wernsdorfer, N. Galvez, C. Chaneac, D. Mailly, J.P. Jolivet, J. Magn. Magn. Mater. 2003, 262, 6–14. H. Kachkachi, A. Ezzir, M. Nogues, E. Tronc, Eur. Phys. J. B 2000, 14, 681–689. E. Tronc, A. Ezzir, R. Cherkaoui, C. Chaneac, M. Nogues, H. Kachkachi, D. Fiorani, A.M. Testa, J.M. Grenche, J.P. Jolivet, J. Magn. Magn. Mater. 2000, 221, 63–79. J.L. Dormann, D. Fiorani, R. Cherkaoui, E. Tronc, F. Lucari, F. d’Orazio, L. Spinu, M. Nogues, H. Kachkachi, J.P. Jolivet, J. Magn. Magn. Mater. 1999, 203, 23–27. E. Tronc, Il Nuovo Cimento 1996, 18, 163–180. E. Tronc, P. Prene, J.P. Jolivet, F. d’Orazio, F. Lucari, D. Fiorani, M. Godinho, R. Cherkaoui, M. Nogues, J.L. Dormann, Hyperfine Interact. 1995, 95, 129–148. P. Prene, E. Tronc, J.P. Jolivet, J. Livage, R. Cherkaoui, M. Nogues, J.L. Dormann, Hyperfine Interact. 1994, 93, 1409–1414. S. Wang, H. Xin, Rad. Phys. Chem. 1999, 56, 567–572. D.K. Lee, Y.S. Kang, Mol. Cryst. Liq. Cryst. 2004, 424, 85–94. J. Du, H. Liu, J. Magn. Magn. Mater. 2006, 302, 263–266. F. Xu, X. Zhang, Y. Xie, X. Tian, Y. Li, J. Colloid Interface Sci. 2003, 260, 160–165. M. Gotic, T. Jurkin, S. Music, Colloid. Polym. Sci. 2007, 285, 793–800. M. Gotic, T. Jurkin, S. Music, Mater. Res. Bull. 2009, 44, 2014–2021. J.-M.R. Genin, A. Renard, Ch. Ruby, Hyperfine Interact. 2008, 186, 31–37, and references cited in. C. Ruby, A. Gehin, R. Aissa, J. Ghanbaja, M. Abdelmoula, J.-M.R. Genin, Hyperfine Interact. 2006, 167, 803–807. J.-P. Jolivet, S. Cassaignon, C. Chaneac, D. Chiche, E. Tronc, J. Sol.-Gel. Sci. Technol. 2008, 46, 299–305. J.-P. Jolivet, E. Tronc, C. Chaneac, C. R. Geosci. 2006, 338, 488–497. J.-P. Jolivet, C. Froidefond, A. Pottier, C. Chaneac, S. Cassaignon, E. Tronc, P. Euzen, J. Mater. Chem. 2004, 14, 3281–3288.

REFERENCES

181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229.

501

K. Norrish, R.M. Taylor, J. Soil. Sci. 1961, 12, 294–306. U. Schwertmann, R.W. Fitzpatrick, J. Le Roux, Clays Clay Miner. 1977, 25, 373–374. D.C. Golden, L.H. Bowen, S.B. Weed, J.M. Bigham, Soil. Sci. Am. J. 1979, 43, 802–808. J. Fleisch, R. Grimm, J. Gr€ubler, P. Gutlich, J. Phys. Colloq. 1980, 41, C1-169–C1-170. R.W. Fitzpatrick, U. Schwertmann, Geoderma 1982, 27, 335–347. Y. Tardy, D. Nahon, Am. J. Sci. 1985, 285, 865–903. U. Schwertmann, U. Latham, Geoderma 1986, 39, 105–123. M.A. Wells, R.J. Gilkes, R.R. Anand, Clay Miner. 1989, 24, 513–530. U. Schwertmann, L. Carlson, Soil Sci. Soc. Am. J. 1994, 58, 256–261. J. Friedl, U. Schwertmann, Clay Miner. 1996, 31, 455–464. E. Fritsch, G. Morin, A. Bedidi, D. Bonnin, E. Balan, S. Caquineau, G. Calas, Eur. J. Soil Sci. 2005, 56, 575–588. M. Landers, R.J. Gilkes, Appl. Clay Sci. 2007, 35, 162–172. M.E. Fleet, Acta Cryst. 1984, C40, 1491–1493. H. Okudera, K. Kihara, T. Matsumoto, Acta Cryst. 1996, B52, 450–457. N.G. Newbery, D.R. Peacor, E.J. Essene, J.W. Geissman, Contrib. Mineral. Petrol. 1982, 80, 334–340. H. Shimazaki, Resour. Geol. 1998, 48, 23–29. M. Ohkawa, M. Miyahara, E. Ohta, K. Hoshino, J. Miner. Petrol. Sci. 2007, 102, 182–193. M. Ohkawa, Y. Yamashita, M. Inoue, R. Kitagawa, S. Takeno, Miner. Petrol. 2000, 70, 15–23. U. Schwertmann, G. Pfab, Geochim. Cosmochim. Acta 1996, 60, 4279–4283. A. Manceau, M.L. Schlegel, M. Musso, V.A. Sole, C. Gauthier, P.E. Petit, F. Trolard, Geochim. Cosmochim. Acta 2000, 64, 3643–3661. E. Murad, U. Schwertmann, Am. Mineral. 1988, 73, 1395–1400. M.L.M. de Carvalho-e-Silva, C.S.M. Partiti, J. Enzweiler, S. Petit, S.M. Netto, S.M.B. de Oliveira, Hyperfine Interact. 2002, 142, 559–576. M.L. Carvalho-e-Silva, A.Y. Ramos, H.C.N. Tolentino, J. Enzweiler, S.M. Netto, M.C.M. Alves, Am. Mineral. 2003, 88, 876–882. L.R. Bernstein, G.A. Waychunas, Geochim. Cosmochim. Acta 1987, 51, 623–630. F.J. Berry, A. Bohorquez, C. Greaves, J. McManus, E.A. Moore, M. Mortimer, J. Solid State Chem. 1998, 140, 428–430. € Helgason, J. McManus, H.M. Palmer, R.T. Williams, J. Solid State Chem. 2000, 151, 157–162. F.J. Berry, C. Greaves, O. K. J onas, K. Solymar, Acta Chim. Hung. 1970, 66, 383–394. D.G. Lewis, U. Schwertmann, Clay Miner. 1979, 14, 115–126. U. Schwertmann, R.W. Fitzpatrick, R.M. Taylor, D.G. Lewis, Clays Clay Miner. 1979, 27, 105–112. M.V. Fey, J.B. Dixon, Clays Clay Miner. 1981, 29, 91–100. B.A. Goodman, D.G. Lewis, J. Soil Sci. 1981, 32, 351–363. E. De Grave, L.H. Bowen, S.B. Weed, J. Magn. Magn. Mater. 1982, 27, 98–108. S.A. Fysh, P.E. Clark, Phys. Chem. Miner. 1982, 8, 180–187. S.A. Fysh, P.M. Fredericks, Clays Clay Miner. 1983, 3, 377–382. E. Murad, U. Schwertmann, Clay Miner. 1983, 18, 301–312. E. De Grave, D. Chambaere, L.H. Bowen, J. Magn. Magn. Mater. 1983, 30, 349–354. V. Barron, J.L. Rendon, J. Torrent, C.J. Serna, Clays Clay Miner. 1984, 32, 475–479. D.G. Schulze, Clays Clay Miner. 1984, 32, 36–44. D.G. Schulze, U. Schwertmann, Clay Miner. 1984, 19, 521–539. E. De Grave, A.E. Verbeeck, D. Chambaere, Phys. Lett. 1985, 107A, 181–184. E. Murad, U. Schwertmann, Clays Clay Miner. 1986, 34, 1–6. D.D. Amarasiriwardena, E. De Grave, L.H. Bowen, Clays Clay Miner. 1986, 34, 250–256. E. Murad, L.H. Bowen, Am. Mineral. 1987, 72, 194–200. D.G. Schulze, U. Schwertmann, Clay Miner. 1987, 22, 83–92. P.G. Fazey, B.H. O’Connor, L.C. Hammond, Clays Clay Miner. 1991, 39, 248–253. R.J. Pollard, Q.A. Pankhurst, P. Zientek, Phys. Chem. Miner. 1991, 18, 259–264. H. Stanjek, U. Schwertmann, Clays Clay Miner. 1992, 40, 347–354. S. Bocquet, E. De Grave, J. Phys. Condens. Matter 1994, 6, 6825–6832. S. Popovic, M. Ristic, S. Music, Mater. Lett. 1995, 23, 139–142.

502

230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278.

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

G.M. da Costa, E. De Grave, R.E. Vandenberghe, Hyperfine Interact. 1998, 117, 207–243. C.A. Barrero, R.E. Vandenberghe, E. De Grave, Hyperfine Interact. 1999, 122, 39–46. A.C. Scheinost, D.G. Schulze, U. Schwertmann, Clays Clay Miner. 1999, 47, 156–164. M.A. Wells, R.W. Fitzpatrick, R.J. Gilkes, J. Dobson, Geophys. J. Int. 1999, 138, 571–580. U. Schwertmann, J. Friedl, H. Stanjek, D.G. Schulze, Clays Clay Miner. 2000, 48, 159–172. E. Van San, E. De Grave, R.E. Vandenberghe, H.O. Desseyn, L. Datas, V. Barr on, A. Rousset, Phys. Chem. Miner. 2001, 28, 488–497. G.M. da Costa, E. Van San, E. De Grave, R.E. Vandenberghe, V. Barr on, L. Datas, Phys. Chem. Miner. 2002, 29, 122–131. J. Majzlan, A. Navrotsky, B.J. Evans, Phys. Chem. Miner. 2002, 29, 515–526. C.A. Barrero, R.E. Vandenberghe, E. De Grave, A.L. Morales, G. Perez, Hyperfine Interact. 2003, 148/149, 337–344. L.F. C otica, S.C. Zanatta, S.N. de Medeiros, I.A. dos Santos, A. Paesano Jr. J.B.M. da Cunha, Solid State Ionics 2004, 171, 283–288. C. Sudakar, G.N. Subbanna. T.R.N. Kutty, J. Mater. Sci. 2004, 39, 4271–4286. A.J. Blanch, J.S. Quinton, C.E. Lenehan, A. Pring, Miner. Mag. 2008, 72, 1043–1056. M. Ristic, S. Popovic, S. Music, Croat. Chem. Acta 2009, 82, 397–404. S. Krehula, S. Music, Mater. Chem. Phys. 2010, 123, 67–76. N.E. Brown, A. Navrotsky, G.L. Nord Jr., S.K. Banerjee, Am. Mineral. 1993, 78, 941–951. F.J. Berry, C. Greaves, J. McManus, M. Mortimer, G. Oates, J. Solid State Chem. 1997, 130, 272–276. T. Ishikawa, H. Yamashita, A. Yasukawa, K. Kandori, T. Nakayama, F. Yuse, J. Mater. Chem. 2000, 10, 543–547. € Helgason, J.-M. Greneche, F.J. Berry, S. Mørup, F. Mosselman, J. Phys. Condens. Matter 2001, 13, 10785–10797. O. U. Schwertmann, G. Pfab, Geochim. Cosmochim. Acta 1994, 58, 4349–4352. N. Kaur, B. Singh, B.J. Kennedy, M. Gr€afe, Geochim. Cosmochim. Acta 2009, 73, 582–593. H.N. Ok, L.S. Pan, B.J. Evans, Phys. Rev. B 1978, 17, 85–90. U. Schwertmann, U. Gasser, H. Sticher, Geochim. Cosmochim. Acta 1989, 53, 1293–1297. utlich, J. Magn. Magn. Mater. 1995, 150, 277–283. R. Klinger, J. Ensling, H. Jachow, W. Meisel, E. Schwab, P. G€ S. Music, M. Lenglet, S. Popovic, B. Hannoyer, I. Czak o-Nagy, M. Ristic, D. Balzar, F. Gashi, J. Mater. Sci. 1996, 31, 4067–4076. R. Balasubramanian, D.C. Cook, M. Yamashita, Hyperfine Interact. 2002, 139/140, 167–173. T. Grygar, P. Bezdic9ka, J. De9dec9ek, E. Petrovsky, O. Schneeweiss, Ceram. Silikaty 2003, 47, 32–39. A.L. Morales, C.A. Barrero, F. Jaramillo, C. Arroyave, J.-M. Greneche, Hyperfine Interact. 2003, 148/149, 135–144. E. van den Pol, D.M.E. Thies-Weesie, A.V. Petukhov, P. Panine, G.J. Vroege, J. Phys. Condens. Matter 2008, 20, 404219. E.E. Sileo, A.Y. Ramos, G.E. Magaz, M.A. Blesa, Geochim. Cosmochim. Acta 2004, 68, 3053–3063. S. Suzuki, Y. Takahashi, T. Kamimura, H. Miyuki, K. Shinoda, K. Tohji, Y. Waseda, Corros. Sci. 2004, 46, 1751–1763. S. Suzuki, Y. Takahashi, M. Saito, M. Kusakabe, T. Kamimura, H. Miyuki, Y. Waseda, Corros. Sci. 2005, 47, 1271–1284. E.E. Sileo, D.P. Daroca, C.A. Barrero, A.L. Larralde, M.S. Giberti, C. Saragovi, Chem. Geol. 2007, 238, 84–93. S. Krehula, S. Music, J. Alloys Compd. 2009, 469, 336–342. W. Stiers, U. Schwertmann, Geochim. Cosmochim. Acta 1985, 49, 1909–1911. R.E. Vandenberghe, A.E. Verbeeck, E. De Grave, W. Stiers, Hyperfine Interact. 1986, 29, 1157–1160. R.M. Cornell, R. Giovanoli, Clays Clay Miner. 1987, 35, 11–20. M.H. Ebinger, D.G. Schulze, Clays Clay Miner. 1989, 37, 151–156. U.G. Gasser, R. N€uesch, M.J. Singer, E. Jeanroy, Clay Miner. 1999, 34, 291–299. A.C. Scheinost, H. Stanjek, D.G. Schulze, U. Gasser, D.L. Sparks, Am. Mineral. 2001, 86, 139–146. E.E. Sileo, M. Alvarez, E.H. Rueda, Int. J. Inorg. Mater. 2001, 3, 271–279. J. Lai, K.V.P.M. Shafi, K. Loos, A. Ulman, Y. Lee, T. Vogt, C. Estournes, J. Am. Chem. Soc. 2003, 125, 11470–11471. J. Lai, K.V.P.M. Shafi, A. Ulman, K. Loos, N.-L. Yang. M.-H. Cui, T. Vogt, C. Estournes, D.C. Locke, J. Phys. Chem. B 2004, 108, 14876–14883. L.C.A. Oliveira, J.D. Fabris, R.R.V.A. Rios, W.N. Mussel, R.M. Lago, Appl. Catal. A 2004, 259, 253–259. M. Alvarez, E.E. Sileo, E.H. Rueda, Chem. Geol. 2005, 216, 89–97. S. Krehula, S. Music, J. Alloys Compd. 2006, 426, 327–334. E. De Grave, R. Leyman, R. Vanleerberghe, Phys. Lett. 1983, 97A, 354–356. R.M. Cornell, R. Giovanoli, Clays Clay Miner. 1989, 37, 65–70. R.M. Persoons, E. De Grave, Solid State Commun. 1989, 72, 977–980. R.M. Persoons, E. De Grave, R.E. Vandenberghe, Hyperfine Interact. 1990, 54, 655–660.

REFERENCES

279. 280. 281. 282. 283. 284. 285. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 313. 314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 324.

503

J. Gerth, Geochim. Cosmochim. Acta 1990, 54, 363–371. E. De Grave, R.M. Persoons, R.E. Vandenberghe, P.M.A. Bakker, Phys. Rev. B 1993, 47, 5881–5893. R.M. Persoons, E. De Grave, P.M.A. de Bakker, R.E. Vandenberghe, Phys. Rev. B 1993, 47, 5894–5905. uesch, J. Berthelin, A.J. Herbillon, Clay Miner. 1996, 31, 465–476. U.G. Gasser, E. Jeanroy, N. Mustin, O. Barres, R. N€ T. Ishikawa, H. Nakazaki, A. Yasukawa, K. Kandori, M. Seto, Corros. Sci. 1999, 41, 1665–1680. K. Iwasaki, T. Yamamura, Mater. Trans. 2002, 43, 2097–2103. S. Krehula, S. Music, Clay Miner. 2008, 43, 95–105. R.M. Cornell, R. Giovanoli, W. Schneider, J. Chem. Tech. Biotechnol. 1992, 53, 73–79. C. Diaz, N.R. Furet, V.I. Nikolaev, V.S. Rusakov, M.C. Cordeiro, Hyperfine Interact. 1989, 46, 689–693. T. Ishikawa, H. Nakazaki, A. Yasukawa, K. Kandori, M. Seto, Mater. Res. Bull. 1998, 33, 1609–1619. C.A. Barrero, J. Arpe, E. Sileo, L.c. Sanchez, R. Zysler, C. Saragovi, Physica B 2004, 354, 27–34. M. Mohapatra, B. Pandey, C. Upadhyay, S. Anand, R.P. Das, R.P. Das, H.C. Verma, J. Magn. Magn. Mater. 2005, 295, 44–50. S. Krehula, S. Music, S. Popovic, J. Alloys Compd. 2005, 403, 368–375. W.F. de Souza, I.R. Guimar~aes, L.C.A. Oliveira, A.S. Giroto, M.C. Guerreiro, C.L.T. Silva, Appl. Catal. A 2010, 381, 36–41. K. Inouye, S. Ishii, K. Kaneko, T. Ishikawa, Z. Anorg. Allg. Chem. 1972, 391, 86–96. R.M. Cornell, R. Giovanoli, Polyhedron 1988, 7, 385–391. S. Krehula, S. Music, J. Mol. Struct. 2007, 834–836, 154–161. I.R. Guimaraes, A. Giroto, L.C.A. Oliveira, M.C. Guerreiro, D.Q. Lima, J.D. Fabris, Appl. Catal. B: Environ. 2009, 91, 581–586. € Helgason, R.C. Mercader, E.A. Moore, S.J. Stewart, P.G. Wynn, J. Solid State Chem. 2001, I. Ayub, F.J. Berry, R.L. Bilsborrow, O. 156, 408–414. 9 . Skoko, S. Popovic, J. Alloys Compd. 2006, 420, 260–268. S. Krehula, S. Music, Z S. Music, S. Popovic, M. Ristic, J. Mater. Sci. 1989, 24, 2722–2726. C.A. dos Santos, A.M.C. Horbe, C.M.O. Barcellos, J.B. Marimon da Cunha, Solid State Commun. 2001, 118, 449–452. E. Subramanian, J.-O. Baeg, S.M. Lee, S.-J. Moon, K.-J. Kong, Int. J. Hydrogen Energy 2009, 34, 8485–8494. M. Zhou, H. Luo, J. Alloys Compd. 1997, 255, 239–242. M. Sorescu, L. Diamandescu, A. Tomescu, S. Krupa, Physica B 2009, 404, 2159–2165. L.C.A. Oliveira, T.C. Ramalho, E.F. Souza, M. Goncalves, D.Q.L. Oliveira, M.C. Pereira, J.D. Fabris, Appl. Catal. B: Environ. 2008, 83, 169–176. L.C.A. Oliveira, F. Zaera, I. Lee, D.Q. Lima, T.C. Ramalho, A.C. Silva, E.M.B. Fonseca, Appl. Catal. A 2009, 368, 17–21. C.J. Capobianco, Am. Mineral. 1998, 83, 1152–1160. S. Krehula, S. Music, J. Alloys Compd. 2006, 416, 284–290. S. Krehula, S. Music, J. Mol. Struct. 2010, 976, 61–68. Y. Wang, F. Kong, B. Zhu, S. Wang, S. Wu, W. Huang, Mater. Sci. Eng. B 2007, 140, 98–102. S. Krehula, S. Music, J. Mol. Struct. 2009, 924–926, 201–207. X. Lin, R.C. Burns, G.A. Lawrance, Water, Air, Soil Pollut. 2003, 143, 155–177. E.E. Sileo, P.S. Solıs, C.O. Paiva-Santos, Powder Diffr. 2003, 18, 50–55. S. Krehula, S. Music, J. Alloys Compd. 2007, 431, 56–64. M. Ristic, S. Popovic, M. Tonkovic, S. Music, J. Mater. Sci. 1991, 26, 4225–4233. M. Sorescu, L. Diamandescu, D. Tarabasanu-Mihaila, J. Phys. Chem. Solids 2004, 65, 1719–1725. € Helgason, J. McManus, J. Mater. Chem. 1999, 9, 223–226. F.J. Berry, C. Greaves, O. € F.J. Berry, O. Helgason, A. Boh orquez, J.F. Marco, J. McManus, E.A. Moore, S. Mørup, P.G. Wynn, J. Mater. Chem. 2000, 10, 1643–1648. € Helgason, Hyperfine Interact. 2000, 126, 269–275. F.J. Berry, O. T.I. Bhuiyan, M. Nakanishi, Y. Kusano, T. Fujii, J. Takada, Y. Ikeda, Mater. Lett. 2007, 61, 3774–3777. M. Mohapatra, S.K. Sahoo, C.K. Mohanty, R.P. Das, S. Anand, Mater. Chem. Phys. 2005, 94, 417–422. N.C. Pramanik, T.I. Bhuiyan, M. Nakanishi, T. Fujii, J. Takada, S.I. Seok, Mater. Lett. 2005, 59, 3783–3787. F.J. Perez-Alonso, M. L opez Granados, M. Ojeda, T. Herranz, S. Rojas, P. Terreros, J.L.G. Fierro, M. Gracia, J.R. Gancedo, J. Phys. Chem. B 2006, 110, 23870–23880. T. Nagano, H. Mitamura, S. Nakayama, S. Nakashima, Clays Clay Miner. 1999, 47, 748–754. S. Mustafa, S. Khan, M.I. Zaman, S.Y. Husain, Appl. Surf. Sci. 2009, 255, 8722–8729.

504

325. 326. 327. 328. 329. 330. 331. 332. 333. 334. 335. 336. 337. 338. 339. 340. 341. 342.

23

57

€ Fe MOSSBAUER SPECTROSCOPY IN THE INVESTIGATION OF THE PRECIPITATION OF IRON OXIDES

M.C. Duff, J. Urbanik Coughlin, D.B. Hunter, Geochim. Cosmochim. Acta 2002, 66, 3533–3547. M.A. Wells, R.J. Gilkes, R.W. Fitzpatrick, Clays Clay Miner. 2001, 49, 60–72. T. Fujii, M. Kayano, Y. Takada, M. Nakanishi, J. Takada, Solid State Ionics 2004, 172, 289–292. W.B. Ingler Jr., S.U.M. Khan, Thin Solid Films 2004, 461, 301–308. Y. Kubota, H. Morita, Y. Tokuoka, Y. Imaoka, IEEE T. Magn. 1979, 15, 1558–1560. S. Mustafa, S. Khan, M.I. Zaman, Water Res. 2010, 44, 918–926. J.Z. Jiang, R. Lin, W. Lin, K. Nielsen, S. Mørup, K. Dam-Johansen, R. Clasen, J. Phys. D: Appl. Phys. 1997, 30, 1459–1467. O.K. Tan, W. Cao, Y. Hu, W. Zhu, Ceram. Int. 2004, 30, 1127–1133. Q. Liu, K. Osseo-Asare, J. Colloid Interface Sci. 2000, 231, 401–403. S. Krehula, S. Music, Croat. Chem. Acta 2007, 80, 517–527. S. Krehula, S. Popovic, S. Music, Mater. Lett. 2002, 54, 108–113. S. Krehula, S. Music, J. Cryst. Growth 2008, 310, 513–520. A. Goveart, C. Dauwe, P. Plinke, E. De Grave, J. De Sitter, J. Phys. Colloq. 1976, 37, C6-825–C6-827. R.D. Shannon, Acta Cryst. 1976, A32, 751–767. M. Ristic, S. Popovic, I. Czak o-Nagy, S. Music, Mater. Lett. 1996, 27, 337–341. H.N. Ok, B.J. Evans, Phys. Rev. B 1976, 14, 2956–2965. € Helgason, Hyperfine Interact. 2002, 139/140, 579–587. I. Ayub, F.J. Berry, O. M. Drofenik, M. Kristl, D. Makovec, Z. Jagli9cic, D. Han9zel, Ultrason. Sonochem. 2008, 15, 791–798.

C H A P T E R 2 4

FERRATES(IV, V, AND VI): € MOSSBAUER SPECTROSCOPY CHARACTERIZATION VIRENDER K. SHARMA,1 YURII D. PERFILIEV,1 RADEK ZBOR9 IL,1 LIBOR MACHALA,2 AND CLIVE I. WYNTER3 1

Chemistry Department, Florida Institute of Technology, Melbourne, FL, USA Regional Center of Advanced Technologies and Materials, Olomouc, Czech Republic 3 Chemistry Department, Nassau Community College, Garden City, NY, USA 2

24.1 INTRODUCTION Iron commonly exists as metallic iron (Fe(0)), ferrous (Fe(II)), and ferric (Fe(III)) forms in the natural environment. Minerals of ferrous and ferric oxides include w€ustite, hematite, maghemite, magnetite, and goethite (Table 24.1). Iron and iron oxide-based compounds are used in a wide range of applications such as pigments, magnetic recording media, catalysis, and magnetic fluids [1–3]. Amorphous iron oxides have shown potential in industrial and water purification applications [2,4,5]. In photocatalytic processes, a-Fe2O3 as an electrode transforms water into H2 and O2 [6]. In recent years, iron/iron oxides in the form of nanoparticles have demonstrated unique properties for many advanced technological applications [2,7]. Nanoparticles of iron and iron oxides in combination with oxygen and hydrogen peroxide are capable of oxidizing recalcitrant compounds [8–11]. Salts of hypoferrite and ferrite, given in Table 24.1, have also been synthesized because of their use as magnetic materials in the modern electronic industry. Examples include microwave devices, memory cores of computers, radar and satellite communications, and usage as permanent magnets [12–14]. Sodium and potassium salts of higher oxidation states of iron ranging from þ4 to þ6 have also been prepared [15–27] (Table 24.1). Compounds of þ4 and þ5 oxidation states of iron (Fe(IV) and Fe(V)) have interest in environmental and biological fields. Formation of ferryl species has been suggested in the atmospheric water droplets and photoassisted Fenton reaction [28,29]. The involvement of ferryl (FeIV O) and perferryl (FeV O) species as strong oxidants at enzymatic sites for aliphatic H abstraction and/or C C bond scission has been studied [30–35]. These species also play an important role in the oxygen activation and transfer reactions in heme and non-heme iron proteins; hence, studies on understanding their participation in the metabolic transformations in iron enzymes are being pursued [30–38]. These high-valent iron species have been synthesized and characterized including a model for iron(IV)-oxo complexes of heme and non-heme ligands and iron(VI) and iron(V) nitrido compounds [39–43]. In recent years, iron-tetraamidomacrocyclic ligand (Fe-TAML) catalysts have been developed, which in trace amounts activate hydrogen peroxide to generate reactive intermediates FeIV O and FeV O [43–49]. This system has been demonstrated to degrade numerous pollutants [50,51]. Examples include degradation of bisphenols, estrogens, and Orange II dye and inactivation of bacterial spores [50–53]. Toxicity tests of the oxidation reactions showed nontoxic products [50,51]. This system was successfully applied in the desulfurization of heavy oil and in the remediation of pulp and paper industry effluent [52,54]. M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

505

506

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

TABLE 24.1 Iron Oxide Compounds at Different Oxidation States of Iron Compound

Name

Mineral/Salt

FeO Fe2O3 Fe3O4 Fe2O3  H2O FeOOH FeO22 FeO2 FeO32 FeO44 FeO43 FeO42

Ferrous oxide Ferric oxide Ferrosoferric oxide Ferric oxide monohydrate Ferric oxyhydroxide Hypoferrite Ferrite Ferrate(IV) Ferrate(IV) Ferrate(V) Ferrate(VI)

ustite W€ Hematite, maghemite Magnetite Goethite Akaganeite Na2FeO2 NaFeO2, KFeO2 Na2FeO3 Na4FeO4 K3FeO4 Na2FeO4, K2FeO4

In the past two decades, salts of ferrate(VI) have been of great interest because of their role as oxidants and hydroxylating agents in industrial and water treatment processes, such as the development of a “super iron” battery, green chemistry synthesis, and non-chlorine oxidation for pollutant remediation [55–60]. An environmentally benign high energy density cathode using ferrate(VI) for batteries has been suggested [55,61]. For example, a Li-ion cathode consisting of the Na2FeO4 salt resulted in threefold higher reversible capacity [55,61]. Selective oxidations by ferrate(VI) were shown to synthesize organic compounds without producing toxic by-products [62]. The ferrate ion is a powerful oxidizing agent in aqueous media: under acidic conditions, the reduction potential of the ferrate ion is relatively high compared to that of other oxidants used as disinfectants in water treatment processes [60]. Oxidation reactions by ferrate(VI) were completed in shorter time periods than oxidations carried out by permanganate and chromate [57,63]. Iron, unlike chromium and manganese, is considered to be practically nontoxic; therefore, ferrate(VI) can make industrial processes more environmentally benign by achieving cleaner technologies for organic syntheses. Ferrate(VI) has shown potential in oxidizing emerging contaminants and toxins [59,64–69] and in deactivating harmful microorganisms in water treatment without generating harmful by-products [70,71]. Another advantageous property of ferrate(VI) is that its decomposition produces Fe(III), which itself is an excellent coagulant for removal of metals and radionuclides from contaminated water [66,72–74]. In this chapter, spectroscopy techniques as a tool to characterize the ferrates(IV, V, and V) are presented. 57 Fe M€ ossbauer spectroscopy, which is able to distinguish easily between various oxidation states of iron in compounds, is described in detail.

24.2 SPECTROSCOPIC CHARACTERIZATION A number of alkali and alkaline earth salts of ferrates(IV, V, and VI) have been synthesized using three main approaches: (i) wet chemical synthesis, (ii) electrochemical synthesis, and (iii) dry thermal synthesis [23]. In the solid phase, metal ferrates (IV) MFeO3 (alkali metals, Ba, Sr, Ag, and Pb), orthoferrates(IV) M2FeO4 (Sr and Ba), and Ba3FeO5 have been prepared [17,18,24,75]. Examples of ferrate(V) include Na3FeO4, K3FeO4, and Rb3FeO4 [16]. A number of ferrate(VI) salts have been prepared, which include M2FeO4 (M ¼ Li, Na, K, Rb, Cs, and Ag) and M0 FeO4 (M0 ¼ Ca, Ba, and Sr) [76–78]. Since K2FeO4 has a relatively high stability, it was used to prepare other alkali and alkaline earth ferrates(VI) [76,79–81]. Studies of powder X-ray and neutron diffraction on Na4FeO4 suggested that its crystals were in the triclinic system    P-1 with cell parameters a ¼ 8.4810 A , b ¼ 5.7688 A , and c ¼ 6.5622 A [17]. Crystals of Ba2FeO4 were nonmerohedrally    twinned with the monoclinic space group P21/n with a ¼ 6.034 A , b ¼ 7.647 A , and c ¼ 10.301 A [75]. Another barium salt   of ferrate(IV), Ba3FeO5, crystallized in the orthorhombic space group Pnma with a ¼ 7.7016 A , b ¼ 9.0920 A , and    c ¼ 7.8370 A [75]. Orthorhombic single crystals of K3FeO4 had space group Pnma with a ¼ 7.7016 A , b ¼ 9.0920 A ,  2 and c ¼ 7.8370 A [16]. Isolated crystals of alkali and alkaline salts of ferrate(VI) in the form of FeO4 showed space group Pnma too and there were four molecules in the orthorhombic unit cell [79,82]. However, an organic salt of FeO42 ,   [N(CH3)4]2FeO4, crystallized tetragonally (P4/nbm, a ¼ 11.010 A and c ¼ 10.902 A) [83]. The XANES spectra of ferrate(VI) have been determined (Fig. 24.1) [84]. As demonstrated in Fig. 24.1, the spectra could be used to determine

507

24.2 SPECTROSCOPIC CHARACTERIZATION

FIGURE 24.1 XANES spectrum of ferrate(VI) solution, as compared to iron(III) and iron(II) salts. (Reproduced from Ref. 84 with permission of Elsevier.)

the oxidation state and local geometry around iron atoms. A characteristic pre-edge feature of tetrahedrally coordinated ferrate(VI) was observed from both solid and solution samples [84]. The identification of þ2 and þ3 oxidation states of iron could also be achieved from more subtle changes in X-ray absorption edge energy (Fig. 24.1). Solid salts of ferrate(VI) have been characterized by Fourier transform infrared (FTIR) spectrometry [76,76– 78,81,82,85]. FTIR spectra showed an absorption band at 800 cm 1 in the infrared range with triplet splitting, which is distinct for K2FeO4 and BaFeO4 but less for SrFeO4 [82]. A FTIR spectrum of recently prepared CaFeO4, shown in Fig. 24.2, also had a characteristic absorption peak of FeO42 at 800 cm 1 and two slight shoulder peaks between 880 and 810 cm 1 [78]. A strong and broad absorption band centered around 3500 cm 1 and two strong absorption bands at about 1657 and 1623 cm 1 were also observed (Fig. 24.2) [78]. These bands may be due to bending and stretching vibrations of the water in the CaFeO4 sample. A sample artifact of NO3 might have resulted in the absorption peak at 800 cm 1. Spectra of ferrates(IV, V, and VI) are shown in Fig. 24.3 [25,86–88]. The [(P2O7)2FeIVO]6 complex had a peak at lmax ¼ 430 nm (e ¼ 1200 M 1 cm 1) (Fig. 24.3a) [87]. The spectra of FeO(OH)n2 n, [(P2O7)2FeIVO]6 , and LmFe(IV) were found to be similar. On lowering the pH, the 430 nm peak undergoes blueshift. The spectrum of ferryl(IV) (FeO2þ) includes a small broad peak around 320 nm (e320 nm  500 M 1 cm 1) with a continuum that grows in the UV region (Fig. 24.3b) [88,89]. The spectrum of ferrate(V) in alkaline solution had a maximum at 380 nm (e380 nm ¼ 1460 M 1 cm 1) (Fig. 24.3a), which undergoes a blueshift with decreasing pH. Ferrate(V) absorbs very strongly in the UV region (e270 nm  5000 M 1 cm 1) (Fig. 24.3b) [25]. The spectral characteristics of ferrate(V) are compatible with tetrahedral geometry and are similar to other high-valent oxyanions such as CrO42 and MnO42 . The aqueous ferrate(VI) solutions have a distinctive violet color similar to that of solutions of the permanganate ion. As shown in Fig. 24.3a, the alkaline solutions of the ferrate ion exhibited a maximum at 510 nm (e ¼ 1150  25 M 1 cm 1) and a shoulder between 275 and 320 nm. Electron paramagnetic resonance (EPR) measurements of K2FeO4 in alkaline solutions were also made to determine oxidation states and chemical bonding of ferrate(VI) [90]. This study suggested that d2 electrons occupy a triply degenerate orbital. However, later measurements with a single crystal of K2FeO4 concluded that the unpaired electrons occupy a doubly degenerate orbital [91,92]. Results also exhibited a marked change in g values with temperature from 2.14 at 20 K to 2.00 at 193 K. A significant difference between the absorption obtained from K2FeO4 and BaFeO4 was observed [90]. These results were in contrast with the expected pattern for covalent oxyanion with two unpaired

FIGURE 24.2 FTIR spectra of calcium ferrate (VI), CaFeO4. Spectra were measured by a Nicolet Nexus 670 Fourier transform infrared spectrophotometer in a conventional KBr pellet. (Adapted from Ref. 78 with permission of Elsevier.)

508

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

Fe(IV)

1200

(a)

Fe(VI)

ε (M–1 cm–1)

1000 800 Fe(V) 600 400 200 400

450

500

550

600

650

700

λ (nm) 14.0

(b)

FIGURE 24.3 Spectra of ferrates in visible (a) and UV (b) wavelength regions. (Adapted from Refs. 25 and 86–88 with permission of the American Chemical Society and Wiley.)

ε (103 M–1 cm–1)

12.0 10.0 8.0 Fe(V)

6.0

IV

2+

Fe O 4.0

Fe(VI)

2.0 0.0 200

220

240

260

280

300

320

340

λ (nm)

electrons. Recently, the powder and single-crystal EPR spectra of Fe(VI) in the host structures K2MO4 (M ¼ Cr, S, Se) were measured [92]. The calculated zero-field splitting parameters (g ¼ 1.99), D ¼ 0.10 cm 1, and E ¼ 0.01 cm 1 for Fe(VI)-doped K2CrO4 agreed nicely with earlier reported values [90,91]. The results confirmed that Fe(VI) isomorphically substitutes the tetrahedrally coordinated Cr(VI), S(VI), and Se(VI) cations in host compounds ossbauer characterization of ferrates. K2MO4. The next section discusses the M€

€ 24.3 MOSSBAUER SPECTROSCOPY CHARACTERIZATION 57

Fe M€ ossbauer spectroscopy technique has been widely applied to study the compounds of iron in high oxidation states. Various reviews have been written to utilize this technique [19,21,23,24,93,94]. Herein, we describe mostly the work carried out in the past decade. 24.3.1 Ferryl(IV) Ion

Aqua oxoiron(IV) compounds are considered as future alternatives to HO as the Fenton oxidant in environmental and catalytic chemistry. A ferryl complex, [(H2O)5FeIV O]2þ (species Z), has been generated by ozonization of [Fe (H2O)6]2þ by Pestovsky et al. [95]. The authors characterized it spectroscopically, theoretically, and chemically, and determined criteria for distinguishing between hydroxyl radicals and aqueous ferryl-oxo species. The proposed structure for Z, [(H2O)5FeIV O]2þ, calculated with the B3LYP functional and the 6-311G basis set with C2v symmetry is shown in Fig. 24.4. In addition to complex Z, the sample contained up to 50% [Fe(H2O)6]3þ as the decomposition product of Z during sample preparation and handling. The low-temperature (T ¼ 4.2 K) M€ ossbauer spectra of the sample prepared by a rapid freezing technique measured in various external magnetic fields applied parallel to the gamma rays are shown in

€ 24.3 MOSSBAUER SPECTROSCOPY CHARACTERIZATION

509

FIGURE 24.4 Proposed structure for Z, [(H2O)5FeIV O]2þ, calculated with the B3LYP functional and the 6-311G basis set with C2v symmetry. (Reproduced from Ref. 95 with permission of Wiley.)

Fig. 24.5. Expectedly, there are contributions from two species including a paramagnetic complex Z with integer electron spin and a high-spin ferric species with hyperfine parameters, which can be ascribed to [Fe(H2O)6]3þ. The solid lines fitted through the experimental data are spectral simulations (red for Z and blue for [Fe(H2O)6]3þ) based on the spin Hamiltonian [96].

FIGURE 24.5 € ssbauer spectra of a sample at Mo 4.2 K containing 250 mM ferryl (IV). The red lines show contribution from ferryl(IV), representing about 50% of total Fe. Hyperfine parameters for Z: D ¼ 9.7(7) cm 1, Ax/gnbn ¼ Ay/gnbn ¼ 20.3(3) T, DEQ ¼ 0.33 (3) mm s 1, and d ¼ 0.38(2) mm s 1. The contribution from [Fe(H2O)6]3þ is shown in blue. For the 8.0 T spectrum, the theoretical curves for ferryl(IV) and [Fe(H2O)6]3þ are added (black). (Reproduced from Ref. 95 with permission of Wiley.)

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

510

€ ssbauer Parameters of Ferrates(IV, V, and V) TABLE 24.2 Mo Compound Fe(IV) Na2FeO3 K2FeO3 Cs2FeO3 Na4FeO4

Ba2FeO4

Ba3FeO5

Fe(V) La2LiFeO6 K3FeO4 Fe(VI) Na2FeO4 K3Na(FeO4)2 K2FeO4

Rb2FeO4 Cs2FeO4 K2Sr(FeO4)2 BaFeO4

T (K)

d (mm s 1)

D (mm s 1)

2e (mm s 1)

H (T)

G (mm s 1)

298 298 298 6 25 295 85 155 225 295 78 90 140 190 240 295

0.08(2) 0.16(2) 0.11(2) 0.140 0.130 0.218 0.152 0.169 0.185 0.244 0.142 0.150 0.168 0.199 0.204 0.255

– 0.440 0.407 0.36 0.35 0.34 0.33 0.39 0.39 0.40 0.37 0.34 0.35

0.205 – –

26.24 – –

– – – – – – – – –

– – – – – – – – –

0.35(2) 0.35(4) 0.40(1) 0.28 0.30 0.30 0.35 0.37 0.37 0.34 0.32 0.31 0.30 0.30 0.30 0.29

293 RT

0.41 0.55

– 0.15

– –

– –

0.26 0.24

77 4.2 3.6 4.2 78 298 2.8 2.8 2.0 2.0 2.8

0.80 0.89 0.90 0.88 0.84 0.87 0.89 0.87 0.91 0.90 0.90

– 0.21 –

– – – 0 – – – – – – –

– – 14.7 14.4 – – 14.9a 15.1a 8.7a – 11.8a

0.32 – – – – 0.43 – – – – –

– – 0 0 0.14 – 0.16

Data were taken from Refs. 19,23,24,27, and 75. T is the temperature of measurement, d is the isomer shift related to metallic iron, D is the quadrupole splitting, eQ is the quadrupole shift, H is the hyperfine magnetic field, and G is the spectral linewidth. a At 2.8 K.

The spectra of complex Z exhibit paramagnetic hyperfine structure with a negative internal magnetic field (opposed to the applied field). The presence of the doublet measured in the field of 0.05 T for species Z confirms its integer electron spin and rules out an equilibrium mixture of two forms, such as [FeIV O]2þ/[FeIV(OH)2]2þ. Moreover, spin state of 2 instead of 1 can be deduced from the value of internal magnetic field (38 T) obtained from 8 T spectrum for species Z. 24.3.2 Ferrates(IV, V, and VI) The M€ ossbauer parameters of ferrates are given in Table 24.2. Na2FeO4 had the isomer shift value of 0.22 mm s 1, which reasonably agrees with other compounds of iron(IV) [96]. Examples include Ba0.5La1.5Li0.5Fe0.5O4 ( 0.17 mm s 1), Sr0.5La1.5Li0.5Fe0.5O4 ( 0.18 mm s 1), and Ca0.5La1.5Li0.5Fe0.5O4 ( 0.19 mm s 1) [97]. These compounds had a large quadrupole splitting at room temperature (1.10–1.27 mm s 1), because Fe(IV) ions are in distorted octahedral sites. Comparatively, ferrate(IV) had smaller quadrupole splitting due to tetrahedral coordination. The values of isomer shift and quadrupole splitting of Ba2FeO4 and Ba3FeO5 were also close to those of Na2FeO4 [75]. This confirms that these barium compounds have þ4 oxidation state of iron. Because of small quadrupole splitting, no charge disproportionations in the barium ferrates(IV) were observed down to 85 K.

€ 24.3 MOSSBAUER SPECTROSCOPY CHARACTERIZATION

Fe5+

511

Fe4+ KFeO2 α-Fe

–2.0

–1.0

0.0

v (mm s–1)

1.0

Absorption (%)

0.0

4.0

FIGURE 24.6 € ssbauer Room-temperature Mo spectra of ferrate(V) during the decomposition process. A doublet of ferrate(IV) and central lines of a-Fe and KFeO2 are also shown. (Reproduced from Ref. 21 with permission of Springer.)

8.

12.0

The M€ ossbauer spectrum of K3FeO4 is shown in Fig. 24.6 [21]. The spectra contain doublets of iron(V) and iron(IV) ions, both of which are tetrahedral. The M€ ossbauer parameters of iron(V) (Table 24.2) are in agreement with the values of isomer shift and quadrupole splitting of pentavalent iron introduced into K3MnO4 and V2O5 matrices [24,98]. As expected, the value of isomer shift of pentavalent iron in octahedral oxygen arrangement in La2LiFeO6 was smaller ( 0.41 mm s 1) [99] than that of the tetrahedrally coordinated Fe(V) in K3FeO4. Recently, iron(V) nitride complex [PhB(tBulm)3FeV N]BArF24 (PhB(t Bulm)3 ¼ phenyltris(3-tert-butylimidazol-2ylidene)borato; BArF24 ¼ B(3,5-(CF3)2C6H3)4 ) has been prepared [100]. This complex has a four-coordinate iron ion with a terminal nitride ligand. The characterization of the complex by M€ ossbauer and electron paramagnetic resonance spectroscopy demonstrated d3 iron(V) metal center in a low-spin (S ¼ 1/2) electron configuration. The decomposition of the complex by water at low temperature formed ammonia and iron(II) species. Most of the M€ ossbauer spectra were collected on salts of ferrate(VI) (Table 24.2). Table 24.2 demonstrates that isomer shifts of different compounds of ferrate(VI) vary very little ( 0.87 to 0.91 mm s 1). This indicates a weak influence of the counterions on iron bound in an oxygen tetrahedron, which is the main structural unit of all ferrates(VI). In recent years, hyperfine parameters of K2FeO4 at very low temperature (1.5–5 K) with and without external magnetic field have been determined (Table 24.3) [101]. Results of Table 24.3 indicate that the magnetic ordering temperature is

€ ssbauer Spectra of the Potassium TABLE 24.3 Hyperfine Parameters and Relative Spectrum Areas of the Mo Ferrate(VI) Sample T (K)

Bext (T)

Component

5 4

0 0

3.5 3 1.5 1.5

0 0 0 3

Singlet Singlet Sextet Sextet Sextet Sextet Sextet

d (mm s 1) 0.79 0.79 0.79 0.80 0.80 0.80 0.80

eQ (mm s 1) – – 0.04 0.03 0.04 0.04 0

Bhf (T)

RA (%)

– – 9.0 10.6 11.9 13.7 14.5a

100 64 36 100 100 100 100

Reproduced from Ref. 101 with permission from the American Chemical Society. T is the temperature of measurement, Bext is the external magnetic field, d is the isomer shift related to metallic iron, eQ is the quadrupole shift, Bhf is the hyperfine magnetic field, and RA is the relative spectrum area. a Effective magnetic field Beff ¼ Bext þ Bhf.

512

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

FIGURE 24.7 € ssbauer spectrum of Mo potassium ferrate(VI) measured at 1.5 K in an external field of 3 T (Bext jj g-ray). (Reproduced from Ref. 101 with permission of the American Chemical Society.)

around 3.8 K, in agreement with a previous study [79]. The hyperfine magnetic field increased from 10.6 to 13.7 T with decrease in the temperature from 3.5 to 1.5 K. The quadrupole shift parameters of sextet spectra were nonzero (Table 24.3), which suggest that the magnetic ordering induced a slight decrease of a high symmetry of the environment of iron(VI). Magnetic ordering was further explained by conducting in-field M€ ossbauer spectroscopy at temperatures below the Neel temperature, TN (Fig. 24.7). The lines of the spectrum had an intensity ratio close to 3:4:1:1:4:3, which clearly demonstrated a perfect antiferromagnetic ordering below the Neel temperature. Recently, M€ ossbauer parameters of an Fe(VI)-nitrido complex were also determined [42]. Interestingly, the isomer shift values of FeO42 ion are significantly lower than those of the iron(VI)-nitrido complex [(Me3cy-ac)FeN](PF6)2 (d ¼ 0.40 mm s 1). The iron(VI)-nitrido complex has an octahedral coordination and a strong iron–nitrogen multiple bond. Both geometry and electronic structure differences of iron(VI)-nitrido and iron(VI)-oxo compounds influence the variation in the isomer shifts of two species. Significantly, isomer shift values of ferrates decreased with increase in oxidation state (OS) of iron in ferrates (Fig. 24.8). The decrease was linear and can be expressed by Eq. (24.1). dðmm s 1 Þ ¼ 1:084

0:326  OS:

(24.1)

0.0

Na4FeO4

δ (mm s–1)

–0.2

–0.4 K3FeO4 –0.6

–0.8 K2FeO4

FIGURE 24.8 Relationship of isomer shifts with the oxidation state of iron in ferrates at 298 K.

–1.0 3.5

4.0

4.5

5.0

5.5

Oxidation state of Fe

6.0

6.5

€ 24.3 MOSSBAUER SPECTROSCOPY CHARACTERIZATION

513

ρ0 (a0–3) 11617 0.0 –0.2

11618

11619

11620

11621

11622

+4 Fe(IV)

–0.4

Fe(V) +5

δ (mm s–1)

–0.6 –0.8

+6

Fe(VI) –1.0 –1.2 –1.4

+7

Fe(VII)

+8

Fe(VIII)

FIGURE 24.9

–1.6 1.55

1.60

1.65

1.70

1.75

1.80

1.85

1.90

Calculated Fe—O bond length (Å)

Relationships of isomer shifts with the Fe O bond distances and electron density on the iron nucleus (r0) of ferrates.

Equation (24.1) was used to estimate isomer shifts of ferrate(VII) and ferrate(VIII) as 1.18 and 1.40 mm s 1, relative to a-Fe, respectively [102]. More recently, density function theory calculations were performed on ferrates ranging from þ4 to þ8 oxidation states of iron [102]. Figure 24.9 depicts relationship of isomer shift with Fe O bond length and electron density on the iron nucleus (r0). An increase in Fe O bond distance with decrease in oxidation states of iron from ferrate(VIII) to ferrate(IV) was determined. A positive linearity between isomer shift and bond distance of ferrate was found (Fig. 24.9). Conversely, values of isomer shifts showed a negative relationship with r0 (Fig. 24.9). An increase in the r0 with increase of the oxidation state of iron was expected. The relationship of Eq. (24.1) was used to understand the mechanism of decomposition and formation reactions of solid ferrates under different conditions, which are described below. 24.3.3 Case Studies 24.3.3.1 Decomposition of Na4FeO4 in Air In this study, Na4FeO4 with tetravalent iron atoms was exposed to atmospheric humidity and the decomposition was followed as a function of time by M€ ossbauer spectroscopy (Fig. 24.10) [17]. After 6 h, only one doublet corresponding to Na4FeO4 was seen. This phase partially turned into two additional phases of ferrate(VI) and Fe3þ in 12 h. In additional 12 h, Na2FeO4 fully decomposed and only two phases, ferrate(VI) and ossbauer spectroscopy showed three spectral components of two oxidation sates, Fe3þ, were observed. After 36 h, M€ ferrate(VI) and Fe3þ. Finally, two doublet components, both related to one oxidation state 3þ of amorphous Fe(OH)3 or FeO(OH), were obtained. The series of M€ ossbauer spectra confirms that the crystals of Na4FeO4 are highly hygroscopic and dissociate to Fe3þ and FeVIO42 in water (Eq. (24.2)). The hydrolysis of Fe3þ then gives amorphous Fe(OH)3 or FeO (OH) (Eq. (24.3)) [17]. 3Na4 FeO4 þ 8H2 O ! 12Naþ þ FeVI O4 2 þ 2Fe3þ þ 16OH

(24.2)

2FeVI O4 2 þ 5H2 O ! 2FeðOHÞ3 þ 32O2 þ 4OH

(24.3)

24.3.3.2 Decomposition of Iron(V) in Air The M€ ossbauer spectra of ferrate(V) during decomposition for 12 days have been studied [21]. An initial spectrum had an isomer shift value of approximately 0.55 mm s 1 for ferrate(V), which slowly disappeared in 12 days. The decomposition of ferrate(V) gave an intense line of iron(VI) (d ¼ 0.90 mm s 1) and unresolved wide line, which may be lines of þ3 and þ4 oxidation states of iron.

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

514

FIGURE 24.10 € ssbauer spectra of Na4FeO4 Mo recorded at 295 K, successively after 6, 12, 24, 36, and 48 h. (Reproduced from Ref. 17 with permission of Elsevier.)

24.3.3.3 Decomposition of K2FeO4 in Humid Air The decomposition of K2FeO4 under different humid conditions was studied with specially built experimental setup [103,104]. Three different humid conditions (55–60, 65–70, and 95–100%) were obtained by mixing as-generated humid air with natural air in an appropriate ratio. M€ ossbauer spectra were collected every hour to monitor the kinetics of the chemical transformation of K2FeO4. An example of the M€ ossbauer spectrum recorded between sixth and seventh hours of the aging at RH ¼ 65–70% humid conditions is shown in Fig. 24.11. A singlet of ferrate(VI) (dFe  0.9 mm s 1) disappeared with time to form a doublet subspectrum (dFe  0.35 mm s 1, DEQ  0.65 mm s 1) corresponding to a compound with a trivalent octahedrally coordinated iron. The values of the hyperfine parameters of the doublet suggested that the trivalent iron phase is iron(III) hydroxide. This phase was further confirmed by using X-ray diffraction (XRD), variable-temperature and in-field M€ ossbauer spectroscopy, magnetic measurements, thermogravimetry (TG) and differential scanning calorimetry (DSC), and scanning electron microscopy (SEM) techniques [101,104]. Furthermore, the SEM images of the phase revealed the formation of large KHCO3 crystallites whose surface was covered by ultrasmall X-ray amorphous iron(III) hydroxide nanoparticles in a high degree of agglomeration. The TG curve between 110 and 400  C of this phase gave the overall mass loss of 30 wt%, which is in good agreement with the theoretical value (29 wt%) calculated for decompositions of 2KHCO3 to K2CO3 and Fe(OH)3 to (1/2)Fe2O3 according to Eqs. (24.4) and (24.5). 2KHCO3 ! K2 CO3 þ CO2 þ H2 O

(24.4)

FeðOHÞ3 ! 12 Fe2 O3 þ 32H2 O

(24.5)

Equation (24.6) represents the net decomposition reaction of K2FeO4 under humid conditions. K2 FeO4 þ 2CO2 þ 52 H2 O ! 2KHCO3 þ FeðOHÞ3 þ 34O2

(24.6)

€ 24.3 MOSSBAUER SPECTROSCOPY CHARACTERIZATION

515

FIGURE 24.11 € ssbauer Room-temperature Mo spectrum recorded between sixth and seventh hours of the in situ measurement (RH ¼ 65–70%) starting from the potassium ferrate(VI) sample. (Reproduced from Ref. 104 with permission of Wiley.)

Overall, air humidity had a strong influence on the decomposition rate of K2FeO4 [104,105]. The decomposition rate of K2FeO4 changed from the order of days (at lower RH, 55–60%) to hours (at higher RH, 65–100%) [104]. Such results of aging under humid conditions provided important information on the storage of K2FeO4. 24.3.3.4 Decomposition of BaFeO4 in Dry and Humid Air The decomposition of BaFeO4 under sealed, exposed to air, and exposed to moist air environments has been observed using M€ ossbauer spectroscopy [105,106]. A typical spectrum of BaFeO4 sample after 75 days of shelf storage at 45  C is shown in Fig. 24.12 [106]. In addition to the original spectra of ferrate(VI) and final Fe3þ, a subspectrum of an intermediate state of crystalline ferrate(IV) was observed. Similar results were found in the case of slow disintegration of BaFeO4 under sealed or dry air conditions [105]. However, no intermediate of ferrate(IV) in the rapid decomposition of BaFeO4 under humid conditions has been

FIGURE 24.12 € ssbauer spectrum of BaFeO4 Mo sample after 75 days of shelf storage at 45  C. (Reproduced from Ref. 106 with permission of the Electrochemistry Society.)

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

516

FIGURE 24.13 € ssbauer spectra of a Mo decomposition product of K2FeO4 at different temperatures. (Reproduced from Ref. 101 with permission of the American chemical Society.)

detected [105]. The final phase of Fe3þ was determined as nanocrystalline Fe2O3. The decomposition of BaFeO4 may thus be expressed by Eqs. (24.7) and (24.8). 2BaFeO4 ! 2BaFeO3 þ O2

(24.7)

2BaFeO3 ! 2BaO þ Fe2 O3 þ 12O2

(24.8)

ossbauer spectra during the thermal 24.3.3.5 Thermal Decomposition of K2FeO4 in Static Air The in situ M€ decomposition of K2FeO4 are presented in Fig. 24.13 [101]. When the sample was heated, two components including a sextet of KFeO2 with d ¼ 0.06 mm s 1, eQ ¼ 0.06 mm s 1, and Bhf ¼ 45.8 T and a singlet of K2FeO4 with d ¼ 1.01 mm s 1 were seen at 190  C. In the M€ ossbauer spectra measured at 300, 420, and 590  C, only the sextet of KFeO2 was observed. The results clearly demonstrated that KFeO2 was the only iron-bearing phase during the thermal decomposition of K2FeO4 in air. The mechanism of thermal decomposition of K2FeO4 was further explored by using variabletemperature XRD, TG, and DSC techniques [101]. These techniques supported the stepwise thermal decomposition of K2FeO4 in static air containing CO2 (Eqs. (24.9) and (24.10)) to yield the net equation (Eq. (24.11)). K2 FeO4 ! KFeO2 þ 13 K2 O þ 13 KO2 þ 12O2 1 3

Net :

KO2 þ 16 CO2 ! 16 K2 CO3 þ 14O2

K2 FeO4 þ 16 CO2 ! KFeO2 þ 13 K2 O þ 16 K2 CO3 þ 34O2

(24.9) (24.10) (24.11)

24.3.3.6 Formation of Ferrates in the Solid-State Reaction of Fe2O3 and Na2O2 Unusual derivatives of iron in highest oxidation states were identified in the Na Fe O system upon the oxidation of iron(III) oxide with sodium peroxide [94]. It was shown that the first stage of the reaction of Na2O2 with Fe2O3 always yields a tetravalent iron derivative, which then undergoes a further reaction to give different products, depending on the reagent ratio in the system and temperature. For example, at a Na2O2:Fe2O3 molar ratio of 10:1, ferrate(IV) can be obtained almost without an admixture of other iron-containing phases (Fig. 24.14a). However, the product is unstable and gradually decomposes to give a number of compounds with iron in different oxidation states (Fig. 24.14b). It was suggested [2] that the decomposition follows the reduction and disproportionation mechanisms. Hence, the decomposition product is expected to contain derivatives of both iron(III) and iron in oxidation states higher than þ4. As shown in Fig. 24.14c, the weak isomer shifts of the iron derivatives in different oxidation states in the range of negative velocities become stronger and could be resolved satisfactorily. Analysis of M€ ossbauer spectra of good statistical quality was performed with the DISCVER (discrete versions of M€ ossbauer spectra) program [94]. At the early stage of analysis, the program identifies the maximal possible number of well-defined lines in the spectrum with a specified statistical quality and thus discerns a large number of known and unknown iron derivatives (phases) in samples of complex composition. Previously unknown highest oxidation states of iron from þ5 to þ8 were identified.

517

REFERENCES

FIGURE 24.14 Spectra and their representation by the DPS model for (a) sodium ferrate, (b) sodium ferrate decomposition products, and (c) the products of the reaction of Na2O2 and Fe2O3 at the molar ratio 100:1.

ACKNOWLEDGMENTS V.K. Sharma would like to thank the Center of Ferrate Excellence. V.K. Sharma also acknowleges the support of the United States National Science Foundation (CBET 1236331) for ferrate research. Yu.D. Perfiliev acknowledges the support by the Russian Foundation for Basic Research (project 09-03-01041-a). R. Zboril would like to acknowledge the supports from the Operational Program Research and Development for Innovations—European Social Fund (CZ.1.05/2.1.00/03.0058), the projects of the Ministry of Education of the Czech Republic (1M6198959201 and MSM6198959218), and the project of the Academy of Sciences of the Czech Republic (KAN115600801).

REFERENCES 1. 2. 3. 4.

R. Zboril, M. Mashlan, D. Petridis, Chem. Mater. 2002, 14, 969–982. L. Machala, R. Zboril, A. Gedanken, J. Phys. Chem. B 2007, 111, 4003–4018. S. Enthaler, K. Junge, M. Beller, Angew. Chem., Int. Ed. 2008, 47, 3317–3321. X. Teng, D. Black, N.J. Watkins, Y. Gao, H. Yang, Nano Lett. 2003, 3, 261–264.

518

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

5. 6. 7. 8. 9. 10. 11. 12. 13.

M. Zayat, F. del Monte, M.P. Morales, G. Rosa, H. Guerrero, C.J. Serna, D. Levy, Adv. Mater. 2003, 15, 1809–1812. K. Sivula, F.L. Formal, M. Gratzel, Chem. Mater. 2009, 21, 2862–2867. S. Laurent, D. Forge, M. Port, A. Roch, C. Robic, L. Elst, R.N. Muller, Chem. Rev. 2008, 108, 2064–2110. C. Lee, D.L. Sedlak, Environ. Sci. Technol. 2008, 42, 8528–8533. C.R. Keenan, D.L. Sedlak, Environ. Sci. Technol. 2008, 42, 6936–6941. A.L.T. Pham, C. Lee, F.M. Doyle, D.L. Sedlak, Environ. Sci. Technol. 2009, 43, 8930–8935. R.C.C. Costa, F.C.C. Mourab, J.D. Ardisson, J.D. Fabrisa, R.M. Lago, Appl. Catal. B: Environ. 2008, 83, 131–139. S. Kim, W. Wang, T. Iwaki, A. Yabuki, K. Okuyama, J. Phys. Chem. C 2007, 111, 10175–10180. N.G. Jovic, A.S. Masadeh, A.S. Kremenovi, B.V. Anti, J.L. Blanua, N.D. Cvjetianin, G.F. Goya, M.V. Antisari, E.S. Boin, J. Phys. Chem. C 2009, 113, 20559–20567. B.S. Randhawa, K. Gandotra, Ceram. Commun. 2009, 35, 157–161. M. Atanasov, Inorg. Chem. 1999, 38, 4942–4948. R. Hoppe, K. Mader, Z. Anorg. Allg. Chem. 1990, 586, 115–124. C. Jeannot, B. Malaman, R. Gerardin, B. Oulladiaf, J. Solid State Chem. 2002, 165, 266–277. I.G. Kokarovtseva, I.N. Belyaev, L.V. Semenyakova, Russ. Chem. Rev. 1972, 41, 929–937. N.S. Kopelev, in Moessbauer Spectroscopy of Sophisticated Oxides, V. Attila and H. Zoltan, eds., 1997, pp. 305–332. N.S. Kopelev, Y.D. Perfiliev, Y.M.J. Kiselev, Radioanal. Nucl. Chem. 1992, 162, 239–251. S.K. Dedushenko, Y.D. Perfiliev, A.A. Saprykin, Hyperfine Interact. 2008, 185, 197–202. Y.D. Perfiliev, E.M. Benko, D.A. Pankratov, V.K. Sharma, S.K. Dedushenko, Abstracts of Papers, 232nd ACS National Meeting, San Francisco, CA, USA, September 10–14, 2006, ENVR-019. Y.D. Perfiliev, V.K. Sharma, ACS Symp. Ser. 2008, 985(Ferrates), 112–123. Y.D. Perfliev, Russ. J. Inorg. Chem. 2002, 47, 611–619. B.H.J. Bielski, M.J. Thomas, J. Am. Chem. Soc. 1987, 109, 7761–7764. B.H.J. Bielski, Method Enzymol. 1990, 186, 108–113. N.S. Kopelev, Y.D. Perfiliev, Y.M. Kiselev, Russ. J. Inorg. Chem. 1990, 35, 1022–1025. J.J. Pignatello, D. Liu, P. Huston, Environ. Sci. Technol. 1999, 33, 1832–1836. J.J. Pignatello, E. Oliveros, A. MacKay, Crit. Rev. Environ. Sci. Technol. 2006, 36, 1–84. P.K. Grzyska, E.H. Appelman, R.P. Hausinger, D.A. Proshlyakov, Proc. Natl. Acad. Sci. USA 2010, 107, 3982–3987. S. Fukuzumi, Y. Morimoto, H. Kotani, P. Naumov, Y. Lee, W. Nam, Nat. Chem. 2010, 2, 756–759. J.T. Groves, J. Inorg. Biochem. 2006, 100, 434–447. K.D. Karlin, Nat. Chem. 2010, 2, 711–712. L. Que Jr., Acc. Chem. Res. 2007, 40, 493–500. X. Shan, L. Que Jr., J. Inorg. Biochem. 2006, 100, 421–433. S. Shaik, S. Cohen, Y. Wang, H. Chen, D. Kumar, W. Thiel, Chem. Rev. 2010, 110, 949–1017. J. Rittle, M.T. Green, Science 2010, 330, 933–937. W. Wang, M. Zhang, P. Buhlmann, L. Que Jr., J. Am. Chem. Soc. 2010, 132, 7638–7644. F.T.D. Oliveria, A. Chanda, D. Benerjee, X. Shan, S. Mondal, L. Que Jr., E.L. Bominaar, E. Munck, T.J. Collins, Science 2007, 315, 835–839. A. Mukherjee, M.A. Cranswick, M. Chakrabarti, T.K. Paine, K. Fujisawa, E. M€ unck, L. Que Jr., Inorg. Chem. 2010, 49, 3618–3628. A. Mukerjee, M. Martinho, E.L. Bominaar, E. Munck, L. Que Jr., Angew. Chem., Int. Ed. 2009, 48, 1780–1783. J.F. Berry, E. Bill, E. Bothe, S. George, B. Mienert, F. Neese, K. Wieghardt, Science 2006, 312, 1937–1941. A. Chanda, D. Popescu, D.T.D. Oliveira, E.L. Bominnar, A.D. Ryabov, E. Munck, T.J. Collins, J. Inorg. Biochem. 2006, 100, 606–619. A. Chanda, X. Shan, M. Chakrabarti, W.C. Ellis, D.L. Popescu, F.T.D. Oliveira, D. Wang, L. Que Jr., T.J. Collins, E. Munck, E.L. Bominaar, Inorg. Chem. 2008, 47, 3669–3678. A. Chanda, F. Tiago de Oliveira, T.J. Collins, E. Munck, E.L. Bominaar, Inorg. Chem. 2008, 47, 9372–9379. T.J. Collins, Acc. Chem. Res. 2002, 35, 782–790. T.J. Collins, S.K. Khetan, A.D. Ryabov, in Handbook of Green Chemistry: Green Catalysis, Vol. 1, P.T. Anastas, R.H. Crabtree, eds., Wiley-VCH, 2009, pp. 39–77. A. Ghosh, D.A. Mitchell, A. Chanda, A.D. Ryabov, D.L. Popescu, E.C. Upham, G.J. Collins, T.J. Collins, J. Am. Chem. Soc. 2008, 130, 15116–15126.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

REFERENCES

519

49. F. Tiago de Oliveira, A. Chanda, D. Banerjee, X. Shan, S. Mondal, L. Que Jr., E.L. Bominaar, E. Muenck, T.J. Collins, Science 2007, 315, 835–838. 50. N.W. Shappell, M.A. Vrabel, P.J. Madsen, G. Harrington, L.O. Billey, H. Hakk, G.L. Larsen, E.S. Beach, C.P. Horwitz, K. Ro, P.G. Hunt, T.J. Collins, Environ. Sci. Technol. 2008, 42, 1296–1300. 51. N. Chahbane, D. Popescu, D.A. Mitchell, A. Chanda, D. Lenoir, A.D. Ryabov, K. Schramm, T.J. Collins, Green Chem. 2007, 9, 49–57. 52. C.P. Horwitz, T.J. Collins, J. Spatz, H.J. Smith, L.J. Wright, T.R. Stuthridge, K.G. Wingate, K. McGrouther, ACS Symp. Ser. 2006, 921, 156–169. 53. R.T. Malecky, E. Beach, C.P. Horwitz, T.J. Collins, Abstracts of Papers, 232nd ACS National Meeting, San Francisco, CA, USA, September 10–14, 2006, ENVR-167. 54. S. Mondal, Y. Hangun-Balkir, L. Alexandrova, D. Link, B. Howard, P. Zandhuis, A. Cugini, C.P. Horwitz, T.J. Collins, Catal. Today 2006, 116, 554–561. 55. S. Licht, Energies 2010, 3, 960–972. 56. S. Licht, B. Wang, S. Ghosh, Science 1999, 285, 1039–1042. 57. L. Delaude, P. Laszlo, J. Org. Chem. 1996, 61, 6360–6370. 58. V.K. Sharma, Environ. Sci. Technol. 2010, 45, 5148–5152. 59. V.K. Sharma, J. Environ. Sci. Health A 2010, 45, 645–667. 60. V.K. Sharma, J. Environ. Manage. 2011, 92, 1051–1073. 61. S. Licht, Y. Wang, G. Gourdin, J. Phys. Chem. C 2009, 113, 9884–9891. 62. K.L. Southwell, Lib. Arch. Secur. 2003, 18, 39–49. 63. M.A. Khalilzadeh, A. Hosseini, H. Sadeghifar, P. Valipour, Acta Chim. Slov. 2007, 54, 900–902. 64. L. Hu, H.M. Martin, O. Arce-bulted, M.N. Sugihara, K.A. Keating, T.J. Strathmann, Environ. Sci. Technol. 2009, 43, 509–515. 65. C. He, X. Li, V.K. Sharma, S. Li, Environ. Sci. Technol. 2009, 43, 5890–5895. 66. Y. Lee, U.V. Gunten, Water Res. 2010, 44, 555–566. 67. C. Lee, Y. Lee, C. Schmidt, J. Yoon, U. von Gunten, Water Res. 2008, 42, 433–441. 68. V.K. Sharma, G.A.K. Anquandah, N. Nesnas, Environ. Chem. Lett. 2009, 7, 115–119. 69. V.K. Sharma, Chemosphere 2008, 73, 1379–1386. 70. V.K. Sharma, F. Kazama, H. Jiangyong, A.K. Ray, J. Water Health 2005, 3, 45–58. 71. V.K. Sharma, Water Sci. Technol. 2007, 55, 225–232. 72. Y. Lee, S.G. Zimmermann, A.T. Kieu, G.V. Gunten, Environ. Sci. Technol. 2009, 43, 3831–3838. 73. R.A. Yngard, V.K. Sharma, J. Filip, R. Zboril, Environ. Sci. Technol. 2008, 42, 3005–3010. 74. V.K. Sharma, Adv. Environ. Res. 2002, 6, 143–156. 75. J.L. Delattre, A.M. Stacy, V.G. Young, G.L. Long, R. Hermann, F. Grandjean, Inorg. Chem. 2000, 41, 2834–2838. 76. S. Licht, L. Yang, B. Wang, Electrochem. Commun, 2005, 7, 931–936. 77. X. Yu, S. Licht, J. Solid State Chem. 2008, 12, 1523–1540. 78. Z. Xu, J. Wang, H. Shao, Z. Tang, J. Zhang, Electrochem. Commun. 2007, 9, 371–377. 79. R.H. Herber, S. Johnson, Inorg. Chem. 1979, 18, 2786–2790. 80. S. Licht, X. Yu, ACS Symp. Ser. 2008, 985(Ferrates), 2–51. 81. X. Yu, S. Licht, J. Power Sources 2007, 171, 966–980. 82. R.J. Audette, J.W. Quail, Inorg. Chem. 1972, 11, 1904–1908. 83. M. Malchus, M. Jansen, Z. Anorg. Allg. Chem. 1998, 624, 1846–1854. 84. A.I. Tsapin, M.G. Goldfeld, G.D. McDonald, K.H. Nealson, B. Moskovitz, P. Solheid, K.M. Kemner, S.D. Kelly, K.A. Orlandini, Icarus 2000, 147, 68–78. 85. S. Licht, V. Naschitz, L. Halperin, N. Halperin, L. Lin, J. Chen, S. Ghosh, B. Liu, J. Power Sources 2001, 101, 167–176. 86. J.D. Rush, B.H.J. Bielski, J. Am. Chem. Soc. 1986, 108, 523–525. 87. J.D. Menton, B.H.J. Bielski, Radiat. Phys. Chem. 1990, 36, 725–733. 88. T. Loegager, J. Holcman, K. Sehested, T. Pedersen, Inorg. Chem. 1992, 31, 3523–3529. 89. F. Jacobsen, J. Holcman, K. Sehested, Int. J. Chem. Kinet. 1998, 30, 215–221. 90. A. Carrington, D.J.E. Ingram, D. Schonland, M.C.R. Symons, J. Am. Chem. Soc. 1956, 4701–4704. 91. R. Debuyst, F. Dejehet, D. Apers, J. Inorg. Nucl. Chem. 1975, 37, 338–340.

520

€ 24 FERRATES(IV, V, AND VI): MOSSBAUER SPECTROSCOPY CHARACTERIZATION

92. 93. 94. 95. 96. 97. 98. 99.

B. Wagner, D. Reinen, T.C. Brunold, H.U. Gudel, Inorg. Chem. 1995, 34, 1934–1942. Y.D. Perfiliev, L.A. Kulikov, Bull. Russ. Acad. Sci.: Phys. 2007, 71, 1264–1268. S.K. Dedushenko, Y.D. Perfil’eva, M.A. Chuevb, A.M. Afanas’ev, Russ. J. Inorg. Chem. 2010, 55(6), 942–949. O. Pestovsky, S. Stoian, E.L. Bominaar, X. Shan, E. Munck, L. Que Jr., A. Bakac, Angew. Chem., Int. Ed. 2005, 44, 6871–6874. F. Menil, J. Phys. Chem. Solids 1985, 46, 763–789. L. Fournes, Y. Potin, J.C. Grenier, G. Demazeau, M. Pouchard, Solid State Commun. 1987, 62, 239–244. S.K. Dedushenko, Y.D. Perfilev, D.E. Tcheboukov, D.A. Pankratov, Y.M. Kiselev, Mendeleev Commun. 1999, 5, 171–172. G. Demazeau, B. Buffat, F. Menil, L. Fournes, M. Pouchard, J.M. Dance, P. Fabritchnyi, P. Hagenmuller, Mater. Res. Bull. 1981, 16, 1465–1472. J.J. Scepaniak, C.S. Vogel, M.M. Khusniyarov, F.W. Heinemann, K. Meyer, J.M. Smith, Science 2011, 331, 1049–1052. L. Machala, R. Zboril, V.K. Sharma, J. Filip, O. Schneeweiss, Z. Homonnay, J. Phys. Chem. B 2007, 111, 4280–4286. O. Poleshchuk, N. Kruchkova, Y. Perfiliev, S. Dedushenko, J. Phys.: Conf. Ser. 2010, 217, 012041–012044. L. Machala, R. Zboril, V.K. Sharma, Z. Homonnay, AIP Conf. Proc. 2008, 1070, 114–121. L. Machala, R. Zboril, V.K. Sharma, J. Filip, D. Jancik, Z. Homonnay, Eur. J. Inorg. Chem. 2009, 1060–1067. I. Nowik, R.H. Herber, M. Koltypin, D. Aurbach, S. Licht, J. Phys. Chem. Solids 2005, 66, 1307–1313. K.E. Ayers, N.C. White, J. Electrochem. Soc. 2005, 152, A467–A473.

100. 101. 102. 103. 104. 105. 106.

C H A P T E R 2 5

CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY € MOSSBAUER SPECTROMETRY METH2  NE KIYOSHI NOMURA1 AND ZOLTAN 1 2

School of Engineering, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Faculty of Science, Eotvos Lorand University, Budapest, Hungary

25.1 INTRODUCTION Dilute magnetic semiconductors/insulators (DMS/DMI), which show ferromagnetism at room temperature, are very attractive for studying the basic mechanism of dilute magnetism and the potential application to spin transfer electronics, that is, spintronics. It was suggested that materials with high Curie temperature would be realized when magnetic impurities are incorporated into semiconductors with large bandgap [1]. In order to clarify the intrinsic dilute magnetism, it is necessary to prepare oxides doped with dilute magnetic ions that are not precipitated. The well-known superexchange or double-exchange interactions cannot be responsible for long-range magnetic order in the case of a few percent of doping magnetic cations. As a new approach to this phenomenon, Coey et al. proposed the so-called polaronic model for describing the local origin of dilute magnetism. In their model, shallow donor electrons and doping magnetic ions form bound magnetic polarons, which overlap to create a spin-split impurity band [2]. The Curie temperature in the mean-field approximation varies as (xd)1/2, where x and d are the concentrations of magnetic cations and donors, respectively. The magnetic phase diagram shows regions of semiconducting and metallic ferromagnetism, cluster paramagnetism, spin glass, and canted antiferromagnetism (Fig. 25.1). However, recent findings point out that nanomaterials such as ZnO, TiO2, SnO2, Al2O3, and HfO2 show very weak ferromagnetism even without doping magnetic ions [3]. The origin of magnetism of these oxides is considered to be due to defects existing on the surface and interface of nanograins. TiO2, ZnO, and SnO2 are transparent semiconductors with large bandgap. In our related studies, we have focused on Fe-doped SnO2 powders prepared by the sol–gel method [4], SnO2 films implanted with Fe ions [5], and the effect of dilution and clustering of Fe ions doped into SnO2 and TiO2 by nuclear inelastic scattering using synchrotron radiation [6]. Recently, we have studied SiO2 glass insulators implanted with Fe ions. It is found that Fe-implanted SiO2 glass has a potential as a transparent magnet although ion implantation often results in magnetic precipitates [7]. However, widening the range of dilute magnetic materials is essential in the process of understanding the basics of this phenomenon of unconventional magnetism.

M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

521

€ 522 25 CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY MOSSBAUER SPECTROMETRY

FIGURE 25.1 The magnetic phase diagram for dilute ferromagnetic semiconductors. The electrons are localized in the shaded area. xp and dp are the cation and donor polaron percolation thresholds, respectively. g is the ratio of the radius of the hydrogenic donor orbital to the Bohr radius [2].

Transparent and magnetic garnet of Y3Fe5O12 (YIG) is used for optical isolator and magnetic multiplier [8]. Its analogue Y3Al5O12 (YAG) is used for solid-state laser by small doping of, for example, Nb. Solid solutions of cubic Y3Al5 xFexO12 have been obtained over the entire composition range 0  x  5, although bulk ferromagnetic property has been shown only for relatively high iron concentrations (x  2.5). The ferrimagnetic structure develops by antiferromagnetic spin alignments between the tetrahedral (d) and octahedral (a) sublattices (the ratio is 24:16) of YIG. This is accompanied by M€ ossbauer spectra consisting of two sextets, with respect to the tetrahedral and octahedral sites, although two slightly different environments exist for the Fe ions in the octahedral site in YIG as well. Y3FexAl5 xO12 samples with lower iron content (x < 0.12) have been considered as paramagnets at room temperature [9]. There are no preferential sites of Fe ions [10]. Despite the previous assumptions, we were able to show an unprecedented iron doping dependence of roomtemperature magnetization in Y3Al5 xFexO12 (x ¼ 0.02, 0.1) and Y3(YAl4 x)B Fex O12 (x ¼ 0.04, 0.2, 0.4) samples prepared by the sol–gel method. The most interesting feature of this series of garnets was the appearance of bulk ferromagnetism at very dilute iron doping (x being less than 1% with respect to Al content) [11]. The sol–gel method provides uniform fine powders at lower heating temperatures and for shorter heating times than a conventional heating method of mixed oxides. The crystalline sizes of the powders are about 30–50 nm in diameter. X-ray diffraction (XRD) analysis of the prepared YIG and YAG solid solutions revealed garnet structure of A3B5O12 [11] even in the case of Y:(Al, Fe):O ¼ 4:4:12 starting ratio. In order to investigate where the excess Y atoms are located at, extended X-ray absorption fine structures (EXAFS) were measured. The results of investigating the electronic and magnetic properties of these garnets suggest that the oxygen defects and deformations produced may play an important role in bulk magnetism, in accordance with the proposed model of Coey et al. In order to confirm the phenomena, two samples with characteristically different iron doping ratios, Y3Al5 xFexO12 (x ¼ 0.02, 0.1), have been treated by a microdischarge between carbon felts (MD/CF), which provides the structure deformation and reduction by rapid heating and quenching [12]. The results showed that this treatment is effective for enhancing the bulk ferromagnetism. In order to understand the underlying physical and chemical background, M€ ossbauer spectrometry gave detailed and unique description of the doped iron ions. These phenomena and their M€ ossbauer analysis are explained in this chapter.

523

25.3 X-RAY DIFFRACTION AND EXAFS ANALYSIS

25.2 SAMPLE PREPARATIONS BY THE SOL–GEL METHOD Y3Al5 xFexO12 (x ¼ 0.02, 0.1) and Y3(YAl4 xFex)BO12 (x ¼ 0.04, 0.2, 0.4) were prepared by the sol–gel method. First, citric acid and ethylene glycol were added to solutions of Y, Al, and 57 Fe chloride. The mixed solutions with nominal compositions were stirred and evaporated at 353 K, thermally decomposed at 623 K in air, and finally calcined at 1173 K for 3 h in air. After the first series of measurements, all five samples were post-annealed at 1123 K for 1 h and then at 1223 K for 1 h in ambient atmosphere. The effect of heat treatment was investigated by X-ray diffraction, M€ ossbauer spectroscopy, and magnetization measurements. In order to further confirm what affects the dilute magnetism of these oxides, two garnet samples (x ¼ 0.02, 0.1) were treated by MD/CF, which could lead to the formation of crystal defects by quick heating and quenching [12]. After recording their X-ray diffractograms, magnetization data, and M€ ossbauer spectra, these samples treated by MD/CF were heated at high temperatures again.

25.3 X-RAY DIFFRACTION AND EXAFS ANALYSIS The samples with the nominal composition of YAl1 yFeyO3 (y ¼ 0.01, 0.05, 0.1) expecting perovskite structure were prepared by the sol–gel method and post-annealing. However, the X-ray diffractograms of these samples (using Cu Ka radiation) show a dominant garnet phase of A3B5O12 without the sign of perovskite or any other impurity phase as shown in Fig. 25.2. The resultant X-ray diffractograms show pure garnet phase in good agreement with the reference CIF file no. 2003066. While the A site is occupied only by yttrium, the B sites should be shared by aluminum, iron, and yttrium ions giving a composition of Y3(YAl4 xFex)BO12 (x ¼ 0.04, 0.2, 0.4), where the B in the superscript indicates that the YAl4 x ions occupy the B sites in the garnet structure. Compared to the conventional Y3Al5 xFexO12 garnets (x ¼ 0.02, 0.1) (Fig. 25.2, right), the presence of excess yttrium atoms increased the cubic cell parameter a by about 0.015 nm and altered the intensity ratios, whereas the cubic cell parameters of Y3Al5 xFexO12 garnets (x ¼ 0.02, 0.1) increased only by 0.003 nm.

(800)

(640) (552) (642)

x = 0.02

3

3x10

1st group: B Y3[YAl4-x ] FexO12

12.45 12.40

(840) (842) (664)

0

(532)

(211) (220)

3

1x10

(444)

3

2x10

(422) (431) (521) (440)

3

3x10

(321) (400)

(420)

12.50

2nd group: B Y3[Al5-x ] FexO12

12.35

x = 0.04

3

2x10

12.30

3

1x10

12.25

x = 0.1

3

5x10

a (Å)

Intensity (a.u.)

0 4 1x10

12.20 12.15

0 3 3x10

x = 0.2

3

2x10

12.10

3

1x10

12.05

0 3 3x10

x = 0.4

3

literature data for Y3Al5-xFexO12

12.00

2x10

3

11.95

1x10

0

0

10

20

30

40

50

60

70

80

1

2

3

4

5

Iron content, X

2θ (deg)

FIGURE 25.2 XRD patterns and lattice parameters of Y3Al5 xFexO12 (x ¼ 0.02, 0.1) and Y3(YAl4 xFex)BO12 (x ¼ 0.04, 0.2, 0.4).

€ 524 25 CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY MOSSBAUER SPECTROMETRY

FIGURE 25.3 Schematic crystal structures of YAG. Al(VI) and Al(IV) are located at octahedral site and tetrahedral site, respectively [13].

FIGURE 25.4 Radial structural functions of yttrium EXAFS spectra of Y3(YAl4 xFex)BO12 (x ¼ 0.04, 0.2, 0.4) (b, d, e) and Y3Al5 xFexO12 (x ¼ 0.02, 0.1) (a, c).

Magnitude of FT (Å–4)

The EXAFS of these samples were measured by recording the K absorption edge of yttrium and iron as well. However, due to the low concentration of iron ions, the iron EXAFS data are very poor in statistics and cannot be analyzed quantitatively. Generally, Y ions coordinated with 12 oxygen ions occupy the center of the garnet structure, and Al ions are located at octahedral and tetrahedral sites. The schematic structure of YAG is shown in Fig. 25.3. As the X-ray diffractograms showed only garnet phase in all of the investigated samples, it is plausible to presume that the excess Y ions in the case of the samples with cation ratio of Y:Al ¼ 1:1 should occupy the Al sites. As the yttrium EXAFS results showed, the main peaks of radial structural functions, which represent the neighboring atoms in several coordination spheres, have the same positions in all five samples (Fig. 25.4). As a main conclusion of the EXAFS measurements, this agrees well with the result of the XRD measurements that the basic structure is the same for all five samples, that is, they form garnet crystals. However, the spectra of the two groups with different Y and Al ratios differ somewhat. In the case where the yttrium ions are assumed to occupy the B site of the garnet structure as well, at least an extra peak appears in the Fourier  transformed EXAFS data in the region of the nearest cation ligand sphere (which is between about 2.5 and 3.5 A) at about  3 A. This can be explained either by the different ionic radii of Y3þ and Al3þ that result in a slightly different distance to the central yttrium ions or by the superposed EXAFS spectra of the yttrium ions at the B site. Although the exact identification of the peaks is rather difficult, it seems to be clear that the EXAFS data support the garnet structure with yttrium ions at B sites as well. Fe EXAFS spectra (Fig. 25.5) show that the environment of the iron ions is fairly similar in all five samples, indicating that in every sample a small number of iron ions occupy the same crystallographic site, which is supposed to be the B site.

0.10

0.05

c

a

de

b

0.00 0

2

4 R (Å)

6

8

525

25.4 MAGNETIC PROPERTIES

FIGURE 25.5 Normalized iron EXAFS spectra of Y3(YAl4 xFex)BO12 (x ¼ 0.04, 0.2, 0.4) (b, d, e) and Y3Al5 xFexO12 (x ¼ 0.02, 0.1) (a, c).

25.4 MAGNETIC PROPERTIES Magnetization measurements were taken with a VSM (vibrating sample magnetometer) at room temperature. Temperature dependence of magnetic properties with zero field cooling (ZFC) and field cooling (FC) modes was measured by a SQUID (superconducting quantum interference device) magnetometer. Pure Y3Al5O12 is a diamagnetic insulator, but slightly Fe-doped YAG such as Y3Al5 xFexO12 (x ¼ 0.02) and Y3(YAl4 xFex)BO12 (x ¼ 0.04) showed ferromagnetism at room temperature. With increasing Fe content above x  0.1, the samples showed no more ferromagnetic hysteresis, but only paramagnetism as seen in Figs. 25.6 and 25.12. The origin of

0.02

x = 0.04

0.00

–0.02 0.010

x = 0.2

m ( µB / Fe)

0.005 0.000 –0.005

as-prepared post-annealed

–0.010

x = 0.4 0.005

FIGURE 25.6

0.000 –0.005

–0.4

–0.2

0.0

µ0 H (T)

0.2

0.4

VSM curves of iron-doped yttrium aluminum oxides. In the case of x ¼ 0.04, the recorded VSM data after annealing are also shown. (Reproduced with permission of the American Chemical Society.)

€ 526 25 CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY MOSSBAUER SPECTROMETRY

x= 0.04 –5

x = 0.2 x = 0.4

5.0x10

8x10

–5

4.0x10

M (emu g–1)

1x10

–5

–6

–5

3.0x10

–5

2.0x10

M (emu g–1)

–5

1.0x10

0.0 0

6x10

100

150

200

250

300

T (K)

FIGURE 25.7 Temperature dependence of magnetization of Y3(YAl4
xFex)BO12 (x ¼ 0.04, 0.2, 0.4) with and without applied fields (50 Oe) (FC and ZFC). (Reproduced with permission of the American Chemical Society.)

50

–6

FC –6

4x10

ZFC

0

50

100

150

200

250

300

T (K)

ferromagnetism isnotlikelytheformationofmagneticimpuritiesor clusters, asitappearsonly attheloweriron doping ratesof x  0.04,andnotatx  0.1.Moreover,themagneticsaturationvaluesdecreasesignificantlyafterpost-annealing,indicatingthat the vacancies or structural defects may contribute to the bulk magnetism as well. In the case of the Y3(YAl4 xFex)BO12 samples, the magnetic susceptibility is significantly smaller than that of the Y3Al5 xFexO12 compounds; however, the disappearance of ferromagnetism with increasing iron content above x  0.1 is a common feature [11]. The temperature dependence of Y3(YAl4 xFex)BO12 (x ¼ 0.2, 0.4) measured by SQUID under zero field cooling and field cooling (50 Oe) showed the same paramagnetic behavior following the Curie–Weiss rule, whereas, for the sample with x ¼ 0.04, spontaneous magnetization was observed at lower temperatures than 50 K, but magnetization by ZFC increased with increase of temperature above 50 K as shown in Fig. 25.7. This behavior reflects a weak spin glass; that is, in the ZFC measurement, the extra energy gained from the elevating temperature helps magnetic moments overcome the blocking magnetic couplings.

€ 25.5 MOSSBAUER ANALYSIS OF YAG DOPED WITH DILUTE IRON M€ ossbauer spectra were obtained by transmission geometry at room temperature. The g-rays were provided by a CoðCrÞ source with an activity of 108 Bq. Isomer shifts are given relative to a-iron at room temperature. As shown in Fig. 25.8 for the investigated Y3(YAl4 xFex)BO12 samples (and for the two Y3Al5 xFexO12 samples in Fig. 25.13), the M€ ossbauer spectra show three distinctive features: two magnetic components, two paramagnetic components, and a broad magnetic relaxation. For a large number of iron ions (x  0.1), the intensity of magnetic components decreased and the intensity of paramagnetic components increased. The magnetic components can be connected to antiferromagnetic iron ions rather than the bulk ferromagnetism. If bulk ferromagnetism were due to the magnetic interactions between Fe ions, the bulk magnetic moments should be much larger than the observed value of m ffi 0.02 mB/Fe. The two magnetic components are due to occupation of both octahedral site (16a) and tetrahedral site (24d) by iron ions, in which spin moments are arranged antiparallel to each other. However, the hyperfine fields of dilute Fe-doped YAG (Bd ¼ 46 T, Ba ¼ 52 T) are larger than those of YIG (Bd ¼ 40 T, Ba ¼ 49 T [14]). After annealing, the intensity of the two magnetic components changed only very little, although the intensities of the paramagnetic and relaxation peaks are almost unperturbed. The two Fe3þ doublets can also be assigned to paramagnetic iron ions in different lattice sites, although the presence of oxygen vacancies or cation disorder can also change at the vicinity of iron atoms at the grain surface. The low-temperature spectra of the samples with x ¼ 0.04, 0.2, and 0.4 show the very same components as the corresponding room-temperature ones as shown in Fig. 25.9. The only difference is the anticipated increase of isomer 57

€ 25.5 MOSSBAUER ANALYSIS OF YAG DOPED WITH DILUTE IRON

527

1.00 1.00

0.98 0.96

x = 0.04 after postannealing

0.96

x = 0.04

–10

–5

0

5

10

1.0

1.0

0.9 0.8

0.94

Relative transmission

Relative transmission

0.98

0.8 –10

–5

–10

–5

0.6 0.7

x = 0.2

x = 0.4 0.4

–10

–5

0

5

–10

10

–5

v (mm s–1)

0

5

10

v (mm s–1)

FIGURE 25.8 € ssbauer spectra of Y3(YAl4 xFex)BO12 (x ¼ 0.04, 0.2, 0.4) before annealing and Y3(YAl4 xFex)BO12 (x ¼ 0.4) after Mo annealing. (Reproduced with permission of the American Chemical Society.)

1.00

Relative transmission

0.98

0.96

T = 300 K

1.00

0.98

FIGURE 25.9 0.96

T = 11 K –10

–5

0 v (mm

5 s–1)

10

Room-temperature and 11 K € ssbauer spectra of Mo Y3(YAl4 xFex)BO12. (Reproduced with permission of the American Chemical Society.)

€ 528 25 CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY MOSSBAUER SPECTROMETRY

shifts and magnetic fields. Considering the absence of drastic changes in M€ ossbauer spectra due to lowering of the temperature, it can be assumed that the Fe species of the two sextets may be incorporated into the lattice of YAG structure, as the lines of the sextets are more distinguished, and the broad magnetic component may be located in the interfacial sites. Thus, these results imply that the source of this component is more likely a structural inhomogeneity, like a nonhomogeneous distribution of iron ions in some part of the garnet system [11]. As indicated by the low saturation magnetization values, the inverse proportionality between iron content and ferromagnetism, the increased hyperfine magnetic field values, and the inconsistency between the effect of post-annealing on the M€ ossbauer spectra and the VSM data, the mechanism corresponding to the bulk ferromagnetism in these dilute garnets cannot be explained with the ferromagnetic and antiferromagnetic interactions between neighboring magnetic iron ions only. Instead, the model of dilute ferromagnetism in DMI materials should account for the observed bulk ferromagnetism. This model supposes the existence of overlapping locally bound magnetic polarons, consisting of doping magnetic iron ion(s) and a trapped donor electron. In our case, some magnetic iron ions nearby the crystal defects may be magnetically coupled to trapped electrons, and at an optimal concentration of these magnetic polarons (seemingly in the case of x  0.04 iron content and no post-annealing), the overlapping polarons can build up a well-detectable bulk ferromagnetic state.

25.6 MICRODISCHARGE TREATMENT OF IRON-DOPED YAG In order to investigate the role of crystal defects in dilute magnetism in more detail, a nonconventional method for creating more defects (MD/CF, see below) and a quite conventional technique for decreasing their number (annealing in air atmosphere) were used. The effect of modulation of crystal defects is presented by the recently developed method of microdischarge between carbon felts [12]. Carbon felts prepared at 2500  C have the following characteristics. The specific area is as large as 1.6 m2 g 1 and their small electric resistance is about 50 mV cm. Carbon felts with small diameters (about 1 mm) have so low dielectric resistance that they can absorb and reflect microwaves effectively. The discharge between carbon felts occurs easily with a small gap as shown in Fig. 25.10. Because of the high percentage of void and heat insulation, heat is accumulated between the carbon felts. Using the MD/CF method (at 500 W for 2 min), the samples treated only for a few minutes will have more oxygen defects. Two samples of Y3Fe0.02Al4.98O12 and Y3Fe0.1Al4.9O12, containing different iron concentrations and thus showing different level of room-temperature bulk magnetization, were treated with the MD/CF method. XRD patterns are shown in Fig. 25.11. While the constant linewidths prove no change of the crystalline size, the cell parameters decrease slightly. This can be understood as a vacancy-driven distortion of the original garnet structure. The recorded magnetic hysteresis curves of the samples can be separated into a ferromagnetic-like part and a linear increase with the external magnetic field. By a deeper examination of the resulting parameters (the saturation magnetization of the ferromagnetic portion, msat, and the slope of the linear portion), one can observe similar changes on the two samples following the MD/CF treatment, but different tendencies after the intermediate annealing procedures (Fig. 25.12). After the MD/CF treatment, the slope of the linear part decreases, while the ferromagnetic saturation magnetization increases for both samples. Even more surprisingly, heating in ambient atmosphere further strengthens the

FIGURE 25.10 Schematic picture of the furnace for microwave discharge generated between carbon felts and picture of carbon felts. (Reproduced with permission of the Japan Society of Powder and Powder Metallurgy.)

529

25.6 MICRODISCHARGE TREATMENT OF IRON-DOPED YAG

as-prepared after post-annealing after MD/CF

0.8

0.6

0.4

0.2

Intensity (a.u.)

Normalized intensity

1.0

32.0

32.5

33.0

33.5

34.0

34.5

2θ (deg)

as-prepared

FIGURE 25.11 X-ray diffractograms of Y3Al5 xFexO12 after preparation and after the MD/CF treatment. The inset shows the shift of the most intense line due to the MD/ CF treatment.

after MD/CF

10

20

30

40

50

60

70

80

2θ (deg)

ferromagnetic character of the x ¼ 0.1 sample, reaching to the level shown by the Y3Fe0.02Al4.98O12 sample, while it degrades back the saturation magnetization of the x ¼ 0.02 sample to about msat  0.005 mB/Fe. M€ ossbauer spectra are shown in Fig. 25.13. The most obvious change in the M€ ossbauer spectra of x ¼ 0.02 due to the MD/CF treatment is the appearance of new divalent iron ions (d ¼ 1.15 mm s 1, D ¼ 2.20 mm s 1) in addition to Fe2þ species with large quadrupole splitting values (d ¼ 1.27 mm s 1, D ¼ 3.62 mm s 1). They show unambiguously that the MD/CF treatment indeed increases the number of oxygen vacancies and thus the crystal deformation considerably. Moreover, the inclusion of more crystal defects is reflected clearly in the increased quadrupole splitting values of the Fe3þ doublets (d ¼ 0.29 mm s 1, D ¼ 0.80 mm s 1; d ¼ 0.30 mm s 1, D ¼ 1.53 mm s 1) in the Y3Fe0.1Al4.9O12 sample as well, in contrast to Fe3þ doublets (d ¼ 0.33 mm s 1, D ¼ 0.51 mm s 1; d ¼ 0.29 mm s 1, D ¼ 1.42 mm s 1) before the MD/CD treatment. The fact that the quadrupole splittings do not change after the MD/CF treatment in the case of

x = 0.1 0.02

0.00

0.00

–0.02

–0.02

0.010

0.010

0.005

0.005 0.000

0.000

x = 0.02

x = 0.1

No.1 No.2 No.3 No.4

–0.005 –0.010 –0.4

–0.2

0.0

µ0H (T)

0.2

0.4

No.1 No.2 No.3 No.4 –0.4

–0.2

0.0

µ0H (T)

0.2

0.4

–0.005 –0.010

m (µ B / Fe)

m (µ B / Fe) m (µ B / Fe)

No.1 No.4

No.1 No.4

m (µ B / Fe)

x = 0.02 0.02

FIGURE 25.12 Magnetic hysteresis curves of Y3Al0.98Fe0.02O12 (left) and Y3Al0.9Fe0.1O12 (right) measured after the different treatment steps; the insets show the derived magnetic moments after the subtraction of the linear paramagnetic part. Numbers 1, 2, 3, and 4 represent the subsequent treatments of as-prepared, postannealed, MD/CF, and second post-annealed.

€ 530 25 CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY MOSSBAUER SPECTROMETRY

x = 0.02

x = 0.1 1.0

1.00

0.8 0.95 0.6 1.0

0.95

0.8

b

0.6

0.90

1.0

1.00

0.8

0.95

c

0.90

Relative transmission

Relative transmission

a 0.90 1.00

0.6 1.0

1.00

0.8

0.95

d

0.6

0.90 –10

–5

0

v (mm s–1)

5

10

–10

–5

0

5

10

v (mm s–1)

FIGURE 25.13 € ssbauer spectra of Y3Al0.98Fe0.02O12 (left) and Y3Al0.9Fe0.1O12 (right) measured after the Room-temperature Mo different treatment steps: (a) as-prepared, (b) post-annealed, (c) MD/CF treatment, and (d) second post-annealing after MD/CF.

Y3Fe0.02Al4.98O12 agrees with the assumption that the change of the crystal defects occurs more likely around the aluminum ions and not around the iron ions [11]. Namely, in the case of x ¼ 0.02 iron doping, there are enough aluminumrich regions for crystal defects where they would not affect iron ions nearby, while this can hardly be the case for x ¼ 0.1 iron doping. A positive linear dependence between the magnetization and the applied magnetic field stems usually from paramagnetism. In our case, the observed magnetic susceptibility is considered to change mainly due to the competition of paramagnetic components (Fe-doped parts) and diamagnetic components (nondoped YAlO3). Comparing the magnetization data and the observed M€ ossbauer components, it can be seen that in the case of both investigated samples the MD/CF treatment causes dramatic loss in the slope of the magnetization curve. This fortifies that the bulk magnetic susceptibility originates from the paramagnetic iron ions, as the Y3Al5O12 host is diamagnetic. In both samples, the loss of the paramagnetic Fe3þ components is due to the conversion into paramagnetic Fe2þ, which has lower magnetic moment than the high-spin trivalent iron ion, thus resulting in lower susceptibility. Moreover, the broad component and the supposedly antiferromagnetic sextets in the M€ ossbauer spectra may also contribute to the bulk paramagnetic signal. As shown in Fig. 25.14, the lack of direct correspondence between any M€ ossbauer component and the ferromagnetic component of the bulk ferromagnetism suggests that the latter in these garnets may arise not only from the rare doped magnetic iron ions but also from the crystal defects. While the MD/CF treatment increased the ferromagnetic moment in both samples, the magnetic sextets did not increase in the M€ ossbauer spectra; in the case of x ¼ 0.02, the sextets even shrink to their half. On the other hand, the appearance of divalent iron ions clearly shows the increased number of oxygen vacancies. This implies that crystal defects indeed significantly affect the bulk ferromagnetic order in dilute ferromagnets, which fits exactly the polaronic theory of DMI materials [2]. When the number of doping iron ions is very small (in our case, x ¼ 0.02), the maximum of the ferromagnetic moment was already reached in the as-prepared sample with a moderate

531

CONCLUSIONS

0.010 –6

0.006 –6

4.4x10

0.004 –6

4.2x10

0.002

m sat (µ B / Fe)

0.008 –6

4.6x10

–6

4.0x10 100 90

Rel. area (%)

80

100 3+

x = 0.02

Fe doublets 3+ Fe sextets 3+ Fe singlets 2+ Fe doublets

70 60

x = 0.10

90 80 70 60

50

50

40

40

30

30

20

20

10

10

0

Rel. area (%)

Slope ((µ B / Fe) Oe–1)

4.8x10

0 1

2

3

Treatment number

4

1

2

3

4

Treatment number

FIGURE 25.14 € ssbauer The deduced parameters of the bulk magnetization (upper panels) and the relative areas of the Mo components (lower panels) for Y3Al0.98Fe0.02O12 (left) and Y3Al0.9Fe0.1O12 (right).

number of crystal defects, while in the case of x ¼ 0.1 the increased number of defects by the MD/CF treatment resulted in a much higher ferromagnetic moment as compared to the as-prepared state. In addition, the further increase of msat and the remnant magnetization due to the second post-annealing (after MD/CF) in the case of the Y3Fe0.1Al4.9O12 sample denotes that not only the number (reflected partly in the decrease of the Fe2þ component) but also the spatial distribution of the crystal defects with respect to the magnetic ions may essentially influence the formation of magnetic polarons as well. This can be observed in the quadrupole splitting of the Fe3þ doublets in Y3Fe0.1Al4.9O12 (d ¼ 0.33 mm s 1, D ¼ 0.73 mm s 1; d ¼ 0.32 mm s 1, D ¼ 1.56 mm s 1), which remains higher than that in the as-prepared case (d ¼ 0.33 mm s 1, D ¼ 0.51 mm s 1; d ¼ 0.28 mm s 1, D ¼ 1.44 mm s 1). Some Fe2þ (d ¼ 1.63 mm s 1, D ¼ 2.68 mm s 1) remains even after annealing in air. While those oxygen vacancies that are directly connected to iron ions converting them to divalent state almost disappear, those that are nearby the iron ions remain and possibly strengthen the magnetic polarons.

25.7 CONCLUSIONS Y3Al5 xFexO12 and Y3(YAl4 x)B Fex O12 garnets with x ¼ 0.02 and 0.1 and x ¼ 0.04, 0.2, and 0.4, respectively, were prepared by the sol–gel method. Unexpected bulk ferromagnetism was found, which shows a maximum at x ¼ 0.04 in Y3(YAl4 x)B Fex O12 and diminishes with increasing iron doping ratio until x ¼ 0.4. The ferromagnetism of the Y3Al4.95Fe0.02O12 and Y3(YAl3.96)B Fe0:04 O12 samples becomes weaker by annealing at high temperatures. When defects were induced by the MD/CF technique, the strengthening of bulk ferromagnetism was observed. Oxygen vacancy was considered to play an important role in room-temperature ferromagnetism through the iron ions in the dilute Fe-doped YAG garnets. The MD/CF treatment enhances the bulk magnetization together with the appearance of crystal defects and oxygen vacancies in these doped YAG compounds. Thus, the dilute magnetism is considered to be related to the interactions between doped irons and the electrons trapped in oxygen vacancies, that is, magnetic polarons.

€ 532 25 CHARACTERIZATION OF DILUTE IRON-DOPED YTTRIUM ALUMINUM GARNETS BY MOSSBAUER SPECTROMETRY

ACKNOWLEDGMENTS EXAFS measurements were supported by Dr. K. Sakurai and Dr. Mari Mizusawa at National Institute for Materials Science. Dr. Zoltan Nemeth is thankful to the Matsumae Foundation for supporting his cooperation with the Japanese partner.

REFERENCES 1. S.J. Pearton, C.R. Abernathy, G.T. Thaler, R. Frazier, F. Ren, A.F. Hebard, Y.D. Park, D.P. Norton, W. Tang, M. Stavola, J.M. Zavada, R.G. Wilson, Physica B 2003, 340–342, 39–47. 2. J.M.D. Coey, M. Venkatesan, C.B. Fitzgerald, Nat. Mater. 2005, 4, 173–179. 3. A. Sundaresan, R. Bhargavi, N. Rangarajan, U. Siddesh, C.N.R. Rao, Phys. Rev. B 2006, 74, 16130. 4. K. Nomura, C.A. Barrero, J. Sakuma, M. Takeda, Phys. Rev. B 2007, 75, 184411. 5. K. Nomura, H. Reuther, Hyperfine Interact. 2009, 191, 159–165. 6. I. Rykov, K. Nomura, J. Sakuma, C.A. Barrero, Y. Yoda, T. Mitsui, Phys. Rev. B 2008, 77, 014302. 7. K. Nomura, H. Reuther, J. Radioanal. Nucl. Chem. 2011, 287, 341–346. 8. K. Nomura, T. Hanai, R. Sadamoto, Y. Ujihira, T. Ryuo, M. Tanno, Hyperfine Interact. 1994, 84, 421. 9. M.A. Gilleo, S. Geller, Phys. Rev. 1958, 110, 73–78. 10. Y.F. Chen, K.T. Wu, Y.D. Yao, C.H. Peng,. K.L. You, W.S. Tse, Microelectron. Eng. 2005, 81, 329. 11. Z. Nemeth, K. Nomura, Y. Ito, J. Phys. Chem. C 2009, 113, 20044–20049. 12. H. Kurihara, T. Yajima, K. Nomura, J. Jpn. Soc. Powder Powder Metall. 2009, 56, 116–120. 13. L. Dobrzycki, E. Bulska, D.A. Pawlak, Z. Frukacz, K. Wozniak, Inorg. Chem. 2004, 43, 7656–7664. 14. J.J. van Loef, J. Appl. Phys. 1968, 39, 1258 (3 pages).

PA R T V I

INDUSTRIAL APPLICATIONS

C H A P T E R 2 6

€ SOME MOSSBAUER STUDIES OF Fe–As-BASED HIGH-TEMPERATURE SUPERCONDUCTORS AMAR NATH AND AIRAT KHASANOV Department of Chemistry, University of North Carolina at Asheville, Asheville, NC, USA

26.1 INTRODUCTION The discovery of superconductivity in Fe As-based compounds in early 2008 by Kamihara et al. [1] has generated much interest as is evident from hundreds of publications and several review articles [2–5]. In the parent compound BaFe2As2, the Fe atoms are arranged in a planar square and each atom is coordinated tetrahedrally to four As ions, resulting in layers of edge-sharing slightly distorted tetrahedra. Ba ions are sandwiched between Fe As layers (Fig. 26.1). At room temperature, BaFe2As2 acquires a tetragonal structure and exhibits paramagnetic behavior. On cooling, it changes to an orthorhombic structure at 140 K with antiferromagnetic alignment of Fe spins in the ab plane. It is believed that there is considerable orbital-dependent reconstruction of the electronic structure across the magnetostructural transition in pnictides [6–9]. Fe in BaFe2As2 has been substituted by several transition metals including Co, Ni, Cu, Mn, and Cr, and Ba by K [2–4,8–17]. The observed relationship between Co and Ni is particularly interesting. Ni with an additional electron to contribute to the band is found to be twice as effective as Co in inducing superconductivity. Measurements of Hall coefficient and thermoelectric effect exhibit clear evidence for increased injected electron density [3,4,16,17]. On the other hand, substitution of Cu failed to induce superconductivity. Again, K substitution of Ba induced superconductivity due to hole injection; however, Mn and Cr substitution for Fe failed to do so. It seems that the origin of superconductivity due to substitution of Co, Ni, and K is not purely electronic, and some other features such as Fe As bond lengths and As Fe As bond angles are also involved among other hitherto unknown features.

26.2 EXPERIMENTAL PROCEDURE The polycrystalline materials were prepared at FSU, where Ba, Fe, Ni (or Co), and As were mixed together and wrapped with Nb foil, and then sealed in stainless steel vial. The sealed samples were heat treated at 1120  C for 12 h and cooled to 900  C at the rate of 4  C h 1 and held at 900  C for 12 h and cooled to room temperature at the rate of 150  C h 1. The phase purity of the samples was checked by powder X-ray diffraction (XRD) and energy-dispersive X-ray analysis (EDX). M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

535

536

€ 26 SOME MOSSBAUER STUDIES OF Fe–As-BASED HIGH-TEMPERATURE SUPERCONDUCTORS

FIGURE 26.1 Crystal structure of BaFe2As2.

EDX indicated that the concentration of Co is somewhat lower than the formal values. Powder X-ray diffraction shows absence of any significant amount of impurity phases. The Tc values as determined by magnetic susceptibility measurements for 9 and 12% Co substituted materials are 25 and 22.5 K, respectively. The zero field cooled dc magnetic susceptibility of the 9% Co-doped sample is shown in Fig. 26.2, which shows a full shielding at low temperatures. ossbauer source for transmission mode spectroscopy. Pieces of the material were We used 57 CoðRhÞ as the M€ powdered and sandwiched between plastic disks. The M€ ossbauer spectrum for pristine BaFe2As2 shows a single line at room temperature and a magnetically split sextet at low temperatures, as expected (Fig. 26.2). On the other hand, Coand Ni-doped materials exhibit single lines at 78 K and ambient temperature. The observed M€ ossbauer parameters for BaFe2As2 and Ni-doped material, obtained by the least-squares fitting procedure, are in concordance with those

FIGURE 26.2 € ssbauer spectra of BaFe2As2. Mo

537

26.3 WHERE DO THE INJECTED ELECTRONS GO?

FIGURE 26.3 € ssbauer spectra of Mo Ba(Fe0.88Co0.12)2As2 [10].

€ ssbauer spectral parameters – Isomer Shift (IS, mm/s), Quadrupole Splitting (QS, mm/s), Hyperfine TABLE 26.1 Mo field (HI, kOe) for polycrystalline samples prepared at FSU. Observed linewidth 0.29–0.35 mm/s 300 K

BaFe2As2 Ba(Fe0.95Ni0.05)2As2 Ba(Fe0.93Ni0.07)2As2 Ba(Fe0.91Co0.09)2As2 Ba(Fe0.88Co0.12)2As2

78 K

IS

QS

IS

HI

0.42(1) 0.42(1) 0.41(1) 0.43(1) 0.42(1)

0.06(3) 0.10(3) 0.13(3) 0.01(3) 0.05(3)

0.53(1) 0.52(1) 0.52(1) 0.52(1) 0.52(1)

52.0(1)

QS 0.01(3) 0.14(3) 0.15(3) 0.08(3) 0.13(3)

reported in the literature for BaFe2As2 and Ba(Fe0.946Ni0.054)2As2 [9,12]. Remarkably, the spectra of Co- and Ni-doped materials are almost identical, and show no change in the M€ ossbauer isomer shift due to doping of the parent compound BaFe2As2 (Table 26.1). We attempt to rationalize this enigmatic observation by invoking the involvement of p-orbital rehybridization for Fe As and As As bonding.

26.3 WHERE DO THE INJECTED ELECTRONS GO? Interestingly, the isomer shift for the Co- and Ni-doped superconducting compound is the same as that of the parent undoped compound, namely, 0.52 mm s 1 at 78 K. It is widely believed that only the electrons in the d-orbitals in the Fe Fe chain participate in superconductivity. Then the question arises that if the electron is donated by Co2þ with d7 and Ni2þ with d8, to the Fe2þ band with d6, one should observe higher shielding from the d-shell and a decrease in s-electron density at the 57 Fe nuclei, resulting in a higher isomer shift. On the contrary, the isomer shifts for the superconducting compounds with 5 and 10% Ni or 9 and 12% Co substitution are the same as for the pristine BaFe2As2, namely, 0.52 mm s 1 at 78 K [10]. In a similar fashion, one would expect a lower isomer shift if holes are injected into the d-band. Kitao et al. [18] did not observe an observable change in isomer shift with 11% substitution of O by F in LaOFeAs nor did Rotter et al. [19] with substitution of 20% K for Ba. We are faced with the question as to why M€ ossbauer studies do not show any change in d-band population. Co and Ni dopants become an integral part of the band structure and electrons are donated by Co with d7 and Ni with d8 to Fe(II) with d6 with no change in the formal valence state. We rationalize

538

€ 26 SOME MOSSBAUER STUDIES OF Fe–As-BASED HIGH-TEMPERATURE SUPERCONDUCTORS

the apparent contradictions between M€ ossbauer measurements and the observation of increase of bulk electron density as exhibited in Hall and thermoelectric investigation by taking into account the involvement of p–d hybridization of Fe As bonding [20] and the associated contraction of the unit cell resulting in stronger s–p hybridization along the As As system [21]. Combination of these features could provide a conduit for the donated electrons to migrate to the As As system with itinerant electrons [10]. The proposed model would also help understand the threedimensional nature of superconductivity in the Fe As-based system [22].

26.4 NEW ELECTRON-RICH SPECIES IN Ni-DOPED SINGLE CRYSTALS: IS IT SUPERCONDUCTING? Single crystals of Ba(Fe.95Ni.05)2As2 were grown by Chengling Zang at the University of Tennessee, Knoxville, and by Huiqian Luo of the Institute of Physics, Beijing. The crystals were grown from Fe.95Ni.05As self-flux. The phase purity of the samples was checked by powder X-ray diffraction and energy-dispersive X-ray analysis. The M€ ossbauer spectra show an additional species with d of 0.57 mm s 1 and D of 0.62 mm s 1 at 77 K in addition to the regular singlet reported by us and another group [10,12] with d of 0.52 mm s 1 and D of 0.13 mm s 1 (Fig. 26.4). The abundance of the species with higher isomer shift varies from 15 to 65% from sample to sample, indicating that it is sensitively dependent on the conditions during growth. The higher d suggests a larger electron population in the d-shell shielding the s-electron density. One can ask whether it is phase impurity or an electronic phase. XRD and EDX do not show the presence of a secondary phase and therefore it has to be an electron-rich electronic phase. More importantly, is it superconducting? Magnetic susceptibility measurements are widely used to estimate the superconducting volume fraction of single crystals. Figures 26.5 and 26.6 are plots for crystals with 25 and 55% of the electron-rich species. The shielding is 100% in the former case and 95% in the latter. The crystal shape in the latter did not allow a very reliable measure of the volume, perhaps accounting for the less than 100% shielding. One cannot expect such robust magnetic shielding if the abundant

FIGURE 26.4 € ssbauer spectrum of Mo single-crystal BaFe1.9Ni0.1As2 [10].

FIGURE 26.5 Magnetic susceptibility of Ba(Fe0.91Co0.09)2As2 [10]: 25% electron-rich species present.

539

26.5 CAN O2 PLAY AN IMPORTANT ROLE?

FIGURE 26.6 Magnetic susceptibility of Ba(Fe0.95Ni0.05)2As2: 55% electron-rich species present.

electron-rich species were not superconducting. This procedure is widely used for estimating the abundance of superconducting fraction in a single crystal.

26.5 CAN O2 PLAY AN IMPORTANT ROLE? As we mentioned earlier, the parent compound BaFe2As2 has a tetrahedral structure and shows paramagnetic behavior at room temperature, while it transforms to an orthorhombic structure below 140 K. The structural change is somehow associated with drastic changes in electronic properties and the Fe spins are aligned in antiferromagnetic order. Therefore, we as well as other groups observed a singlet at room temperature and a magnetically split sextet at low temperature (Fig. 26.2). What we did not realize was that traces of adsorbed O2 were really responsible for arresting the magnetic state. When the observations were made at 90 K and different atmospheric pressures, a very interesting picture emerged (Fig. 26.7). The plots show differing degrees of magnetic relaxation. At 750 Torr, we see a fully developed sextet indicative of a static magnetic field; as the pressure decreases, one sees varying degrees of relaxation. Finally, at a

FIGURE 26.7 € ssbauer spectra of BaFe2As2 Mo at T ¼ 90 K and different atmospheric pressures: 1 mTorr (a), 35 mTorr (b), 40 mTorr (c), 160 mTorr (d), and 750 mTorr (e) [23].

540

€ 26 SOME MOSSBAUER STUDIES OF Fe–As-BASED HIGH-TEMPERATURE SUPERCONDUCTORS

FIGURE 26.8 € ssbauer spectra of BaFe2As2 at Mo low atmospheric pressure P < 1 mTorr and T ¼ 90 K (a), T ¼ 60 K (b), and T ¼ 6 K (c) [23].

reasonably good vacuum, a fast relaxation results in complete collapse of the sextet [23]. This is reminiscent of superparamagnetic behavior. At very low pressure, strictly limited amount of oxygen is available, forming tiny spin clusters that fluctuate rapidly along the easy directions at a rate faster than the Larmor frequency of the daughter 57 Fe (107 s 1), whereby the nuclei experience almost zero magnetic field. With availability of larger amounts of oxygen, the size of the spin clusters grows and the energy kT available at 90 K is not sufficient and the rate of fluctuation diminishes, resulting in partial relaxation. At still higher pressures, we observe a full-fledged sextet. The process is reversible; if one pumps air out, one observes rapidly relaxing spins. The relaxation can also be made to slow down by lowering the temperature. Cooling down from 90 K under good vacuum to 6 K, the rate of fluctuation decreases considerably (Fig. 26.8). Qualitatively speaking, the picture is the same for polycrystalline preparations of BaFe2As2 as well as gently shattered single crystals. However, some significant differences were observed. For instance, for the polycrystalline preparation, we repeatedly observed that the sample cooled to 6 K morphs from a singlet into a partially relaxed sextet after a couple of days (Fig. 26.9). This observation of hysteresis effect is very interesting and it also provides an additional clue to the origin of spin fluctuations. One can surmise that defects also play an important role. All the above features, namely,

FIGURE 26.9 € ssbauer spectra of BaFe2As2 at Mo T ¼ 6 K after 1 h (a) and after 48 h (b) [23].

REFERENCES

541

hysteresis, role of defects, preexisting or formed by grinding, and effect of some other extraneous species present, have been observed for spin crossover in Fe(II) complexes in the solid state [24–26]. Therefore, the observations discussed above allow us to propose the following model. The Fe(II) in BaFe2As2 is undergoing fluctuations between the intermediate-spin state exhibiting magnetism and the low-spin nonmagnetic state. The adsorbed O2 traps electrons and stabilizes the intermediate-spin magnetic state (Fig. 26.7). The rate of fluctuations, unsurprisingly, diminishes with decrease in temperature (Fig. 26.8). Defects can also arrest the intermediate-spin state albeit with hysteretic effect (Fig. 26.9).

ACKNOWLEDGMENT The authors gratefully acknowledge the generous support from NSF DMR-MRI 0922735.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

Y. Kamihara, T. Watanabe, M. Hirano, H. Hosono, J. Am. Chem. Soc. 2008, 130, 3296. K. Ishida, Y. Nakai, H. Hosono, J. Phys. Soc. Jpn. 2009, 78, 062001. P.C. Canfield, S.L. Bud’ko, Annu. Rev. Condens. Matter Phys. 2010, 1, 1. D. Mandrus, A. S. Sefat, M.A. McGuire, B.C. Sales, Chem. Mater. 2010, 22, 715. M.D. Lumsden, A.D. Christianson, J. Phys: Condens. Matter 2010, 22, 203203. J. Paglione, R.L. Greene, Nat. Phys. 2010, 6, 645. F. Wang, D.H. Lee, Science 2011, 332, 200. D. Kasinathan, A. Ormeci, K. Koch, U. Burkhardt, W. Schnelle, A. Leithe-Jasper, H. Rosner, New J. Phys. 2009, 11, 025023. M. Rotter, M. Tegal, D. Johrendt, I. Schellenberg, W. Hermes, R. P€ ottgen, Phys. Rev. B 2008, 78, 020503. A. Khasanov, S.C. Bhargava, J.G. Stevens, J. Jiang, J.D. Weiss, E.E. Hellstrom, A. Nath, J. Phys: Condens. Matter 2011, 23, 202201. P. Bonville, F. Rullier-Albenque, D. Colson, A. Forget, Eur. Phys. Lett. 2010, 89, 67008. I. Nowik, I. Felner, N. Ni, S.L. Bud’ko, P.C. Canfield, J. Phys.: Condens. Matter 2010, 22, 355701. A. Blachowski, K. Ruebenbauer, J. Zukrowski, K. Rogacki, Z. Bukowski, J. Karpinski, Phys. Rev. B 2011, 83, 134410. M. Alzamora, J. Munevar, E. Baggio-Saitovitch, S.L. Bud’ko, N. Ni, P.C. Canfield, D.R. Sanchez, J. Phys.: Condens. Matter 2011, 23, 145701. D.R. Parker, M.J. P. Smith, T. Lancaster, A.J. Steele, I. Franke, P.J. Baker, F.L. Pratt, M.J. Pitcher, S.J. Blundell, S.J. Clarke, Phys. Rev. Lett. 2010, 104, 057007. E.D. Mun, S.L. Bud’ko, N. Ni, A.N. Thaler, P.C. Canfield, Phys. Rev. B 2009, 80, 054517. F. Rullier-Albenque, D. Colson, A. Forget, H. Alloul, Phys. Rev. Lett. 2009, 103, 057001. S. Kitao, Y. Kobayashi, S. Higashitaniguchi, M. Saito, Y. Kamihara, M. Hirano, T. Mitsui, H. Hosono, M. Seto, J. Phys. Soc. Jpn. 2008, 77, 103706. unther, F. Schrettle, A. Loidl, D. Johrendt, M. Rotter, M. Tegel, I. Schellenberg, F.M. Schappacher, R. P€ ottgen, J. Deisenhofer, A. G€ New J. Phys. 2009, 11, 025014. P. Hansmann, R. Arita, A. Toschi, S. Sakai, G. Sangiovanni, K. Held, Phys. Rev. Lett. 2010, 104, 197002. T. Yildirim, Phys. Rev. Lett. 2009, 102, 037003. P. Vilmercati, A. Fedorov, I. Vobornik, U. Manju, G. Panaccione, A. Goldoni, A.S. Sefat, M.A. McGuire, B.C. Sales, R. Jin, D. Mandrus, D.J. Mannella, N. Singh, Phys. Rev. B 2009, 79, 220503. A. Khasanov, J. Jiang, E.E. Hellstrom, A. Nath, J. Phys.: Condens. Matter 2011, 23, 342201. M.S. Haddad, W.D. Federer, M.W. Lynch, D.N. Hendrickson, Inorg. Chem. 1981, 20, 131. P. Adler, A. Hauser, A. Vef, H. Spiering, P. G€ utlich, Hyperfine Interact. 1989, 47, 343. H. Toftlund, Coord. Chem. Rev. 1989, 94, 67.

C H A P T E R 2 7

€ MOSSBAUER STUDY OF NEW ELECTRICALLY CONDUCTIVE OXIDE GLASS

TETSUAKI NISHIDA1 AND SHIRO KUBUKI2 1 2

Department of Biological and Environmental Chemistry, Faculty of Humanity-Oriented Science and Engineering, Kinki University, Iizuka, Japan Department of Chemistry, Graduate School of Science and Engineering, Tokyo Metropolitan University, Hachioji, Japan

27.1 INTRODUCTION 27.1.1 Electrically Conductive Oxide Glass Most oxide glass has a three-dimensional network structure composed of network former (NWF), such as SiO2, B2O3, P2O5 and network modifier (NWM) such as Na2O, K2O, CaO, and so on. Several NWF constitutes the structural units like distorted SiO4 tetrahedra, distorted BO4 tetrahedra, distorted BO3 triangles, distorted PO4 tetrahedra, and distorted PO5 pyramids. Characteristic feature of glass is that it can dissolve a lot of metal ions or metal oxides as NWF or NWM; they occupy either “substitutional” NWF sites by sharing corner oxygen atoms or “interstitial” NWM sites. In such case, homogeneous glass samples can be obtained, otherwise precipitation of metal oxides occurs when the amount of “solute” exceeds the solubility limit of the glass matrix. It is well known that most oxide glasses like silicate glass, borate glass, and phosphate glass are insulator. In contrast, some oxide glasses containing TiO2, WO3, V2O5, and so on become a semiconductor, having an electrical resistivity (r) of 105107 V cm. Electric conductivity of vanadate (V2O5-based) glass is known to be caused by a step-by-step 3d-electron (polaron) hopping from tetravalent (VIV) or trivalent (VIII) to pentavalent vanadium (VV) atoms [1–3]. Asquenched vanadate glass generally has r of the order of 105107 V cm. It was found that additional heat treatment (e.g., an isothermal annealing) of potassium–iron–vanadate glass, K2O
Fe2O3
V2O5, at a given temperature between glass transition temperature (Tg) and crystallization temperature (Tc) showed a marked decrease in the r value from the order of 106–107 to 2.3 kV cm [1]. This could be explained by a structural relaxation of the network of potassium–vanadate glass that had a pseudochain structure. Structural relaxation is considered to accompany an increased number of carriers (polaron) in the conduction band, followed by an increased probability of electron hopping from VIV or VIII to VV atoms. M€ ossbauer study of K2O
Fe2O3
V2O5 glass measured after heat treatment showed a marked decrease in the quadrupole splitting (D) of FeIII, indicating a gradual increase in the local symmetry of distorted FeO4 tetrahedra that constituted the pseudochain structure with distorted VO4 tetrahedra and VO5 pyramids [1]. This is also the case for distorted VO4 tetrahedra of which corner oxygen atoms are shared by distorted FeO4 tetrahedra; distortion of FeO4 units could be viewed as that of VO4 units. This means that we can estimate the local distortion of VO4 tetrahedra from the D value of FeIII constituting distorted FeO4 tetrahedra at the neighboring sites. Value of D for FeIII can be an ultimate indicator for the local symmetry of FeIIIO4 tetrahedra, since the electric field gradient (EFG) at the nuclear site of FeIII

M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

542

543

27.1 INTRODUCTION

caused by five valence electrons (eqval) is negligible, that is, eqval ¼ 0. In such case, only the eq caused by adjacent atoms, eqlat, determines the e2qQ, in which Q is the nuclear quadrupole moment [4–6]; e2 qQ ¼ e2 qlat Q:

(27.1)

Electrical resistivity (r) of barium–iron–vanadate glass, for example, 20BaO
10Fe2O3
70V2O5, was decreased from the order of 106 to only 25 V cm after a heat treatment at a given temperature between Tg and Tc [2,3]. 20BaO
10Fe2O3
70V2O5 glass has a Japanese-registered trademark of NTA (nanotechnology assorted) glassTM. Heat treatment of NTA glassTM at a temperature higher than Tc brought about a further decrease of r down to several V cm [7]. 27.1.2 Cathode Active Material for Lithium-Ion Battery (LIB) LiCoO2 and LiNiO2 have been used as a cathode active material for Li-ion battery (LIB) of mobile phone, personal computer, and so on. LiFePO4 of olivine type has attracted much interest as a new cathode active material for LIB. LiFePO4 of olivine type has an outstanding characteristic feature that it is rare-earth free, less toxic, and highly abundant on the Earth [8,9]; the abundance of iron on the Earth amounts to 50,000 ppm (5%). M€ ossbauer spectra (Fig. 27.1) of LIB (coin cell type) in which FePO4 was used as a cathode active material showed a quantitative reduction of iron from FeIII to FeII after full discharge, while a reverse oxidation from FeII to FeIII was

FIGURE 27.1 € ssbauer spectra of “cathode” Mo material for LIB containing FePO4 as a cathode active material; (a) before discharge, (b) after discharge to 2.0 V, and (c) after charge to 4.0 V. (Reproduced from Ref. 11 with permission of Elsevier.)

544

€ 27 MOSSBAUER STUDY OF NEW ELECTRICALLY CONDUCTIVE OXIDE GLASS

confirmed after full charge [10,11]. Spectral change at the cathode can be expressed by redox of iron: FeIII PO4 þ Liþ þ e ) LiFeII PO4 during discharge;

(27.2)

LiFeII PO4 ) FeIII PO4 þ Liþ þ e during charge:

(27.3)

while

Behavior of Naþ ions during discharge and charge of sodium-ion battery (SIB) was also investigated by means of M€ ossbauer spectroscopy; intercalation (sorption) and deintercalation (desorption) of Naþ ions could be respectively confirmed as reduction and oxidation of iron in FeIIIPO4 [11].

27.2 STRUCTURAL RELAXATION OF ELECTRICALLY CONDUCTIVE VANADATE GLASS 27.2.1 Increase in the Electrically Conductivity of Vanadate Glass M€ ossbauer spectra of NTA glassTM, 20BaO
10Fe2O3
70V2O5, are illustrated in Fig. 27.2. Heat treatment (isothermal annealing) of this glass at 500  C showed a gradual decrease in D from 0.68 to 0.50 mm s 1. On the other hand, d of 0.38 mm s 1 proved to be independent of the heat treatment, reflecting that FeIII atoms have a coordination number of 4, constituting distorted FeO4 tetrahedra as NWF. These results suggest that distorted FeO4 tetrahedra constitute a threedimensional network structure (skeleton structure) of NTA glassTM, together with distorted VO4 tetrahedra and VO5 pyramids, as is the case for several vanadate glasses [1,2,5,6]. It is considered that distorted FeO4 tetrahedra, VO4 tetrahedra, and VO5 pyramids are randomly distributed in the three-dimensional network of NTA glassTM. Since distorted FeO4 tetrahedra share corner oxygen atoms with distorted VO4 tetrahedra, as described above, we can consider that distortion of FeO4 tetrahedra is essentially the same as that of VO4 tetrahedra, and hence we can estimate the distortion or local symmetry of VO4 tetrahedra from the magnitude of D. On the basis of this idea, local structure of several inorganic glasses containing iron or tin has been investigated so far [1,2,5,6].

100

initial, ∆ = 0.68mm s–1

99 100

30 min, ∆ = 0.54mm s–1

Transmission (%)

99 98 100

60 min, ∆ = 0.50mm s–1

99 100 99

120 min, ∆ = 0.51mm s–1

100

FIGURE 27.2 € ssbauer Room-temperature Mo spectra of 20BaO
10Fe2O3
70V2O5 glass (NTA glassTM) measured after annealing at 500  C for 0–180 min.

180 min, ∆ = 0.50mm s–1 99

98 –8

–6

–4

–2 0 2 4 Velocity (mm s–1)

6

8

545

27.2 STRUCTURAL RELAXATION OF ELECTRICALLY CONDUCTIVE VANADATE GLASS

Δ (mm s–1)

0.70

0.60

300 䉝

0.50

400 䉝

FIGURE 27.3

450 䉝

Plot of quadrupole splitting (D) of FeIII for 20BaO
10Fe2O3
70V2O5 glass measured at room temperature after isothermal annealing at 300–500  C for 0–180 min.

500 䉝

0.40

0

30

60

90

120

t (min)

150

180

210

Step-by-step decrease in D observed after heat treatment is illustrated in Fig. 27.3, from which we can understand that the decrease in D becomes pronounced with an increasing temperature. The decrease of D was effectively observed when NTA glassTM was annealed at a given temperature higher than Tg [3], especially when heated at temperatures higher than Tc [7]. This is ascribed to the structural relaxation of the glass network accompanying a precipitation of small amounts of crystalline particles of BaFe2O4, and so on, as illustrated in Fig. 27.4. Heat treatment at temperature lower than Tg results in a little change in D. These results indicate that heat treatment at temperatures higher than Tg is involved with a partial destruction of the weak chemical bonds and with a resultant rearrangement of the glass network, that is, a structural relaxation or an increase in the local symmetry of distorted FeO4 and VO4 tetrahedra. As described above, the decrease in D for FeIII simply indicates a decrease of eqlat, since eqval for FeIII is zero. DTA chart of NTA glassTM is given in Fig. 27.5, from which Tg and Tc of 312 and 376 C were estimated, respectively.

FIGURE 27.4 XRD pattern for 20BaO
10Fe2O3
70V2O5 glass measured after annealing at 370 and 500  C.

FIGURE 27.5 DTA chart for 20BaO
10Fe2O3
70V2O5 glass; Tg and Tc (peak) were assigned to be 312 and 376  C, respectively.

546

€ 27 MOSSBAUER STUDY OF NEW ELECTRICALLY CONDUCTIVE OXIDE GLASS

0.25

FIGURE 27.6 Current (I) versus voltage (V) curve measured for 20BaO
10Fe2O3
70V2O5 glass after annealing at 500  C for 300 min.

Voltage (V)

0.20 0.15 0.10

ρ : 6.3 Ω cm 0.05 0 0

0.002

0.004

0.006

0.008

0.010

Current (A)

Figure 27.6 illustrates a linear relationship between the current (of mA order) and the voltage (of a few hundreds mV order) observed in NTA glassTM after heat treatment at 500  C for 300 min. Specific resistivity of 6.3 V cm was calculated from the slope of the straight line by taking account the size of the glass sample. Electrically conductive NTA glassTM has been utilized for an electron-emitting needle for ionizer, antielectrostatic powder of 20 nm size, and for the material for FIB (focused ion beam) processing of submicron (several 10 nm) scale. A marked decrease in the r value or a marked increase in the specific conductivity (s) observed after isothermal annealing is depicted in Fig. 27.7. The heat treatment at 350  C, that is, a temperature between Tg and Tc, resulted in an increase of s from the order of 10 6 to 10 3 S cm 1. Much more increase in s was observed up to 10 2 and 2  10 1 S cm 1 (5 V cm) after heat treatment at 400 and 500  C, respectively, being higher than Tc. We can also understand from Fig. 27.7 that the increase in s occurs quickly when NTA glassTM was annealed at higher temperatures. A relationship between s and D of FeIII is illustrated in Fig. 27.8. We can understand from Fig. 27.8 that s increases along with a decrease of D of FeIII. This indicates that increase in the local symmetry of distorted FeO4 and VO4 tetrahedra is favorable for the polaron hopping from VIV (or VIII) to VV. Figure 27.8 also indicates that s becomes larger when annealed at temperatures higher than Tg and Tc. It is considered that polaron hopping is accelerated, and as a result, number of carrier occupying the conduction band is increased. This can be explained that the energy gap between the donor level and the conduction band is decreased. As a result, number of electron occupying the donor level will be increased. Conductivity of several annealed NTA glassTM was measured at different temperatures between 25 and 100  C. Different activation energies (Ea) were obtained for different types of NTA glassTM, depending on the heating condition. Small polaron hopping theory gives Ea values from the following equation: sT ¼ A expð Ea =kTÞ;

(27.4)

in which s is the conductivity at T, and k is the Boltzmann constant. The Ea for initial 20BaO
10Fe2O3
70V2O5 glass was estimated to be 0.38 eV. A gradual decrease in Ea was observed in annealed NTA glassTM, depending on the temperature

FIGURE 27.7 Room-temperature electrical conductivity (s) of 20BaO
10Fe2O3
70V2O5 glass measured after annealing at 350–500  C for 0–300 min.

547

27.2 STRUCTURAL RELAXATION OF ELECTRICALLY CONDUCTIVE VANADATE GLASS

10–1 500 oC, 60 min

10–2

400 oC, 120 min

σ (S cm–1)

400 oC, 60 min

10–3

370 oC, 60 min

400 oC, 90 min 400 oC, 30 min

350 oC, 120 min

10–4

FIGURE 27.8 350 oC, 60 min

10–5 initial

10–6 0.58

0.60

0.62 0.64 Δ (mm s–1)

0.66

0.68

A relationship between electrical conductivity (s) and quadrupole splitting (D) of FeIII for 20BaO
10Fe2O3
70V2O5 glass measured after isothermal annealing.

and time of annealing. Finally, it decreased to only 0.13 eV after annealing at 500  C for 60 min. A linear relationship between s and Ea is illustrated in Fig. 27.9, which shows that s is linearly increased with a decreasing value of Ea. This s versus Ea plot consists of two straight lines with different slopes; a small linear increase in s was observed with a decrease in Ea when annealed at 350  C for 30–90 min, while the slope of the straight line became steeper when annealed at 350  C for 120 min and at 370–500  C for 60 min. Precipitation of small amount of crystalline particles of several 10 nm size, as confirmed by XRD study (Fig. 27.4), seems to be favorable for the structural relaxation of the glass network involved with a marked increase of s. It is noteworthy that Ea of 0.13 eV is nearly 1/2 of the bandgap for GaSb (0.23 eV), 1/5 of that of Ge (0.68 eV), and 1/10 of that of Si (1.16 eV). Smaller Ea values obtained for annealed NTA glassTM will cause an increased number of carriers (electrons) populating in the donor level and conduction band. M€ ossbauer study of related NTA glassTM, xBaO
(90 x)V2O5
10Fe2O3, was reported elsewhere, in which s of “1 S cm 1” was observed after prolonged heat treatment at 500  C [12]. It proved that lower BaO and higher V2O5 contents were favorable for the higher electrical conductivity of NTA glassTM. Partial or total substitution of Liþ for Ba2þ was also investigated in NTA glassTM in order to elucidate the effect of additional ionic conductivity due to Liþ [13]. Ionic conduction due to Liþ was confirmed at temperatures higher than room temperature. By increasing the measuring temperature from room temperature to 150  C, s of 20Li2O
10Fe2O3
70V2O5 glass was increased from 3.2  10 5 to 2.2  10 3 S cm 1 due to additional Liþ-ion conduction, while NTA glassTM showed a small increase in s from 2.1  10 5 to 1.1  10 4 S cm 1 in the same temperature region [13]. 27.2.2 Cathode Active Material for Li-Ion Battery (LIB) M€ ossbauer spectroscopy can be a very effective tool for the development of energy-related materials such as electrode, solid-state electrolyte, sensor, and so on. M€ ossbauer study of cathode active material for LIB is very interesting, since the

FIGURE 27.9 Activation energy (Ea) for the electrical conduction obtained for 20BaO
10Fe2O3
70V2O5 glass after isothermal annealing at 350–500  C.

€ 27 MOSSBAUER STUDY OF NEW ELECTRICALLY CONDUCTIVE OXIDE GLASS

548

FIGURE 27.10 € ssbauer spectra of LiFeVPOx glass before (top, a) and after (top, b) isothermal annealing at 450  C for 120 min. Mo € ssbauer spectra of LiFeV0.5POx glass before (middle, a) and after (middle, b) isothermal annealing at 450  C Mo € ssbauer spectra of LiFePOx glass before (bottom, a) and after isothermal annealing at 550  C for for 120 min. Mo 120 min (bottom, b) [14].

chemical change at the cathode during the discharge and charge can be directly investigated as a redox reaction of iron, as described above. Heat treatment of Li2O
Fe2O3
V2O5
P2O5 glass (LiFeVPOx glass) resulted in a marked increase in the specific discharge- and charge capacities of LIB, together with a satisfactory cyclicity of discharge- and charge profiles [14]. M€ ossbauer spectra of LiFeVPOx glass (Fig. 27.10, top) showed a marked decrease in D from 0.99 to 0.50 mm s 1 after annealing at 450  C for 120 min, which is ascribed to a marked structural relaxation of FeO4 and VO4 tetrahedra, as ossbauer spectra of LiFeV0.5POx glass (Fig. 27.10, middle) observed in NTA glassTM and related vanadate glasses [4]. M€ showed a similar decrease in D from 0.96 to 0.55 mm s 1 after the same annealing as conducted for LiFeVPOx glass. Figure 27.10 provides with an important experimental proof that such structural relaxation is characteristic of “vanadate glass” units, since no change in D was observed in vanadate-free LiFePOx glass after isothermal annealing at 450  C for 120 min and even after annealing at 550  C for 120 min (Fig. 27.10, bottom). Figure 27.10 suggests that structural relaxation of “vanadate glass” units is brought about probably because of its “less rigid” or “flexible” glass network composed of VO4 tetrahedra and VO5 pyramids, in which FeO4 occupy substitutional sites of VO4. Divalent iron observed in LiFePO4 (Fig. 27.10, bottom) is characteristic of phosphate glass, in which FeIII was partially reduced to FeII during sample preparation, as generally observed in alkali phosphate glasses [5]. LiFeVPOx glass showed a marked increase in s from 10 6 to 10 3 S cm 1 after isothermal annealing at 450  C for 60 min (Fig. 27.11, ). This is ascribed to the structural relaxation of “vanadate glass” units in LiFeVPOx glass of which network is composed of FeO4, VO4, VO5, and PO4 units. In Fig. 27.11, LiFeV0.5POx glass (~) showed a similar increase in s from 10 7 to 10 4 S cm 1 after annealing at 450  C for 60 min or more. “Vanadate glass” units will be associated with the increase in s by three orders of magnitude. Vanadate-free LiFePOx glass, in contrast, showed no change in s after isothermal annealing at 450 and 550  C, as shown in Fig. 27.11 (&). These results indicate that structural relaxation of “Vanadate glass” units, as confirmed from M€ ossbauer study (Fig. 27.10, top), gives rise to an increase in s of LiFeVPOx glass from the order of 10 6 to 10 3 S cm 1 when annealed at 450  C for 60 min (Fig. 27.11, ). This is also the case for LiFeV0.5POx glass (Fig. 27.11, ~) that showed a similar decrease of D (Fig. 27.10, middle) with an increase of s from 10 7 to 10 4 S cm 1 after isothermal annealing at 450  C for 60 min. Figure 27.12 (solid line) illustrates discharge–charge profile measured for LIB of coin cell type, in which LiFeVPOx glass with s of 10 3 S cm 1 obtained after annealing at 450  C for 120 min was used as a cathode active material.





549

27.2 STRUCTURAL RELAXATION OF ELECTRICALLY CONDUCTIVE VANADATE GLASS

10–1

FIGURE 27.11

σ (S cm–1)

10–2

10

Electrical conductivity (s) of LiFeVPOx glass ( ), LiFeV0.5POx (~) measured at room temperature after isothermal annealing at 450  C for 0–240 min. Conductivity of vanadium-free LiFePOx glass measured after isothermal annealing at 550  C for 0 and 120 min (&) is also plotted for comparison [14].



10–3 –4

10–5 10–6

(annealed at 550 oC) 10–7

0

30

60

90

120

150

180

210

240

annealing time (min)

After the first discharge of the LIB from 3.5 to 2.0 V, the LIB showed a capacity of 145 mAh g 1. Capacity of LIB after second discharge was slightly decreased to about 140 mAh g 1, while it gradually increased to 150 mAh g 1 after additional discharge and charge. This is illustrated in Fig. 27.13 (solid line), in which capacities of LIB were estimated after 20 runs of discharge and charge. Capacity of LIB of another coin cell type, in which as-quenched (without annealing) LiFeVPOx glass was used as a cathode active material, is plotted with a dotted line in Fig. 27.12. A marked difference is observed between the dotted line and the solid line in Fig. 27.12; the potential in the former is rapidly decreased to 2.0 V, and the capacity is 50 mAh g 1 that is only one-third of the latter obtained for LIB with LiFeVPOx sample annealed at 450  C for 120 min. Figures 27.11– 27.13 prove that increase in the s of LiFeVPOx glass, used as a cathode active material, is very effective for the improvement of LIB. Heat treatment is not always restricted to the isothermal annealing used in the present study; isochronal annealing at different temperatures will also cause an increase in the s of LiFeVPOx glass involved with an increase in the capacity of LIB. Capacity of the LIB, plotted in Fig. 27.13 (top) after 20 runs of discharge and charge, indicates that electrically conductive LiFeVPOx glass with s of about 10 3 S cm 1 shows a satisfactory behavior as a rechargeable battery. It is noteworthy that discharge- and charge capacities of about 150 mAh g 1 were reproduced even after 20th discharge and charge. Similar tendency was observed in LIB in which as-quenched glass was used as a cathode active material (Fig. 27.13, bottom), although the capacity was only 50 mAh g 1. Electrochemistry of LIB before and after discharge/charge could be directly confirmed by M€ ossbauer spectroscopy. Figure 27.14 illustrates M€ ossbauer spectra of cathode active material containing LiFeVPOx glass annealed at 450  C for 120 min. Divalent iron with a distorted octahedral symmetry (d: 1.14 mm s 1, D: 2.17 mm s 1) was observed after discharge of LIB to 2.0 V (Fig. 27.14, top), together with a small amount (23%) of trivalent iron. It is expected that all the iron atoms will be reduced to FeII after discharge to less than 2.0 V, for example, to 1.5 or 1.0 V.

4

FIGURE 27.12

Voltage (V)

charge 3 discharge

2 Before annealing

After annealing

1

0

0

20

40

60

80

100

Capacity (mAh g –1)

120

140

160

Discharge–charge profile of LIB (coin cell) measured before (left, dotted line) and after annealing (right, solid line) the cathode active material, LiFeVPOx, at 450  C for 120 min. The coin cell consists of an electrolyte of 1 M LiPF6 mixed with EC/DMC and an anode of metallic Li. Capacity was measured under a current density of 0.2 mA cm 2 between 2.0 and 4.0 V [14].

550

€ 27 MOSSBAUER STUDY OF NEW ELECTRICALLY CONDUCTIVE OXIDE GLASS

charge

FIGURE 27.13 Discharge/charge capacity is plotted for 20 runs of LIB, in which LiFeVPOx glass annealed at 450  C for 120 min was used as a cathode active material [14]. Measurement was made under the same condition as described in Fig. 27.11.

Capacity (mAh g–1)

160

discharge

140

450oC, 120 min

120 100 80 60 40

initial

20 0 0

2

4

6

8 10 12 14 Cycle number

16

18

20

M€ ossbauer spectra of cathode active material containing LiFeVPOx glass showed trivalent iron (d: 0.43 mm s 1, D: 0.53 mm s 1) when measured after charge to 4.0 V (Fig. 27.14, bottom). All the iron will be oxidized to FeIII, if the LIB were charged to higher than 4.0 V (e.g., to 4.5 or 5.0 V). Spectral change of the cathode active material during discharge of the LIB can be explained by intercalation or sorption of Liþ and simultaneous electron transfer from the anode to the cathode, that is, LiFeIII VPO7 þ Liþ þ e ) Li2 FeII VPO7 :

(27.5)

A reverse reaction will take place during charge of the LIB, that is, deintercalation or desorption of Liþ and electron: Li2 FeII VPO7 ) LiFeIII VPO7 þ Liþ þ e :

(27.6)

Figure 27.14 evidently shows that M€ ossbauer spectroscopy could be an excellent probe for directly investigating the physicochemical behavior involved in discharge and charge of LIB through the redox of iron, that is, reduction of FeIII to FeII takes place during discharge, while oxidation of FeII to FeIII during charge. This means that we could successfully apply M€ ossbauer spectroscopy to the invention and innovation of LIB.

100

FIGURE 27.14 € ssbauer Room-temperature Mo spectra of LIB, in which LiFeVPOx glass annealed at 450  C for 120 min was used as a cathode active material. The spectrum was measured after discharge to 2.0 V (top) and after charge to 4.0 V (bottom) [14].

Transmission (%)

95

90 100

95

90

85 –8

–6

–4

–2 0 2 Velocity (mm s–1)

4

6

8

REFERENCES

551

27.3 SUMMARY M€ ossbauer spectroscopy is quite effective for the research and development of advanced materials that can contribute for the solution of resource, energy, and environmental issues. Electrical resistivity (r) of barium–iron–vanadate glass, for example, 20BaO
10Fe2O3
70V2O5, could be decreased from the order of 106 to only several V cm after heat treatment ossbauer spectra of annealed samples consist of a paramagnetic doublet due to at a given temperature above Tc. M€ distorted FeIII, of which D shows a marked decrease after annealing. This reflects an improved local symmetry of network-forming units like FeO4, VO4 units, and so on. M€ ossbauer spectroscopy is very useful for the invention and innovation of cathode active materials for LIB. We can directly investigate the electrochemical behavior of LIB through the redox of M€ ossbauer atoms included in the cathode active material. M€ ossbauer spectra of cathode active material containing LiFeVPOx glass consist of quadrupole doublet due to FeIII, of which D shows a marked decrease after isothermal annealing at a given temperature around Tc. Resistivity of LiFeVPOx glass is decreased by the annealing. M€ ossbauer spectra of cathode active materials evidently show that FeIII is II ossbauer reduced to Fe during discharge of LIB, as a result of intercalation or sorption of Liþ ions into the cathode. M€ spectra show oxidation of FeII to FeIII during charge of LIB, in accordance with deintercalation or desorption of Liþ from the cathode. The authors wish M€ ossbauer spectroscopy could be successfully utilized for the solution of energy and environmental problems so that our descendant should have healthy and peaceful life.

ACKNOWLEDGMENTS The authors are grateful to Profs. Shigeto OKADA and Jyun-ichi YAMAKI, Kyushu University for their kind help of the evaluation of LIB. They are also indebted to Yu YOSHIDA, Hiroki YASUMITSU, Satoshi ISODA, and Isao FURUMOTO, Kinki University for their help in the experiments.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

T. Nishida, J. Kubota, Y. Maeda, F. Ichikawa, T. Aomine, J. Mater. Chem. 1996, 6, 1889. K. Fukuda, A. Ikeda, T. Nishida, Solid State Phenom. 2003, 90–91, 215. T. Nishida, Japan Patent, No. 3854985, 2006. N.N. Greenwood, T.C. Gibb, M€ossbauer Spectroscopy, Chapman and Hall, London, 1971. Z. Homonnay, S. Music, T. Nishida, N.S. Kopelev, A. Vertes, M€ossbauer Spectroscopy of Sophisticated Oxides, A. Vertes, and Z. Homonnay, eds., Akademiai Kiad o, Budapest, 1997. F.E. Fujita, S. Nasu, T. Nishida, Y. Yoshida, Introduction to M€ossbauer Spectroscopy, -Principles and Applications-, Agne Gijutsu Center, Tokyo, 1999. T. Nishida, Japan Patent, No. 5164072, 2012. A.K. Padhi, K.S. Nanjundaswamy, J.B. Goodenough, J. Electrochem. Soc. 1997, 144, 1188. A.K. Padhi, K.S. Nanjundaswamy, C. Masquelier, S. Okada, J.B. Goodenough, J. Electrochem, Soc. 1997, 144, 1609. S. Okada, T. Yamamoto, Y. Okazaki, J. Yamaki, M. Tokunaga, T. Nishida, J. Power Sources 2005, 146, 570–574. T. Shiratsuchi, S. Okada, J. Yamaki, T. Nishida, J. Power Sources 2006, 159, 268–271. S. Kubuki, H. Sakka, K. Tsuge, Z. Homonnay, K. Sink o, E. Kuzmann, H. Yasumitsu, T. Nishida, J. Ceram. Soc. Jpn. 2007, 115, 776. S. Kubuki, Y. Tomota, R. Yoshimura, Z. Homonnay, K. Sink o, E. Kuzmann, T. Nishida, J. Phys Conf. Ser. 2010, 217(12026), 1. T. Nishida, Y. Yoshida, Y. Takahashi, S. Okada, J. Yamaki, J. Radioanal. Nucl. Chem. 2008, 275(2), 417–422.

C H A P T E R 2 8

€ APPLICATIONS OF MOSSBAUER SPECTROSCOPY IN THE STUDY OF LITHIUM BATTERY MATERIALS

  VICENTE, AND JOSE L. TIRADO RICARDO ALCANTARA , PEDRO LAVELA, CARLOS PEREZ

Laboratorio de Quımica Inorganica, Universidad de Cordoba, Cordoba, Spain

28.1 INTRODUCTION During the last decade, the use of spectroscopic techniques based on the M€ ossbauer effect to study lithium battery materials has increased dramatically. If we take into account that the first lithium-ion product emerged at the end of the eighties, there has been a 10-year delay before the scientific production on M€ ossbauer spectroscopy has been frequently used in the study of electrode materials. A simple search in SciFinder-CAS of published papers on “M€ ossbauer” and “lithium batteries” during the last 35 years reveals this tendency (Fig. 28.1). On the other hand, it is well known that not all chemical elements possess M€ ossbauer isotopes. This is an important limitation in the lithium battery field, as common electrode materials in commercial products contain Li, Co, C, O, or P, which have no M€ ossbauer nuclei. Fortunately, other valuable elements in the electrodes of commercial Li-ion batteries can be studied by this technique. Thus, Fe- and Sn-containing active electrode materials and the less commonly studied Ni have been the basis of thorough investigations. Moreover, many other elements, although not always involved in commercial Li-ion products, have been extensively investigated in lithium test cells and possess M€ ossbauer isotopes. Among these, Zn, Ge, Ru, Ag, Sb, Te, and iodine can be highlighted. This information is summarized in the form of a periodic table in Fig. 28.2. The nuclear transitions involved in M€ ossbauer spectroscopies (MS) are varied in the most relevant elements. In this way, 57 Fe and 119 Sn have simple quadrupolar transitions between I ¼ 3/2 and 1/2, although having a different sign in the changes of nuclear radii DR/R. In contrast, more complex transitions are expected for 61 Ni (I ¼ 5/2 ! 3/2 ) and 121 Sb (I ¼ 7/2þ ! 5/2þ). The materials that can be studied, thanks to the M€ ossbauer effect of the above-mentioned nuclei, are also varied. Both cathode and anode materials can be examined. Moreover, the electrochemical reactions in which they are involved may vary from intercalation to conversion and/or alloying. Table 28.1 shows some examples. 57 Fe MS provides useful information in the study of insertion cathodes, such as olivine LiFePO4, as well as layered solids structurally related to LiCoO2. 57 Fe MS is also useful to analyze anodes consisting of binary or ternary oxides for conversion reactions, or tin intermetallics that react with lithium by alloying processes. In the latter case, a multiisotope approach can be developed, due to the M€ ossbauer effect of both 57 Fe and 119 Sn nuclei. In this chapter, some of the examples are described in more detail. From this analysis, it should be clear that MS is a powerful tool to study electrode materials of advanced lithium-ion batteries. M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

552

553

28.1 INTRODUCTION

FIGURE 28.1 Published papers on € ssbauer” and “lithium “Mo batteries” during the last 35 years. Data from SciFinder-CAS.

Fe

K

Ni

Tc Ru Cs Ba La Hf Ta W

Re Os

Zn Ag

Ir

Ge Sn Sb Te

Kr I

Xe

Pt Au Hg

Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Th Pa

U Np Pu Am

FIGURE 28.2 € ssbauer elements for lithiumMo ion batteries. White box: € ssbauer Elements with known Mo isotopes. Gray box: Elements € ssbauer isotopes and with Mo tested in lithium batteries. Black € ssbauer box: Elements with Mo isotopes and used in commercial Li-ion batteries.

€ ssbauer Nuclei Used in the Study of Lithium-Ion Battery Materials TABLE 28.1 Examples of Mo M€ ossbauer Nucleus/Transition 57

Fe; I ¼ 3/2 ! 1/2 Ni; I ¼ 5/2 ! 3/2 119 Sn; I ¼ 3/2þ ! 1/2þ 121 Sb ;I ¼ 7/2þ ! 5/2þ 61

Positive Electrode LiFePO4; LixFeyNi1 yO2 Li1 xNi1þxO2; LixFeyNi1 yO2

Negative Electrode MFe2O4; FeC2O4 LixSn; MSny Li3Sb; MSby

554

€ 28 APPLICATIONS OF MOSSBAUER SPECTROSCOPY IN THE STUDY OF LITHIUM BATTERY MATERIALS

n

%

0 → 53.14

FIGURE 28.3

1 → 35.43

Different FeFenNi6 n environments in LiFexNi1 xO2 solids, and their probability according to a binomial distribution for x ¼ 0.1. Adapted from Ref. 5 with permission of Elsevier.

2 → 9.84 3 → 1.46 Total = 99.87

28.2 CATHODE MATERIALS FOR Li-ION BATTERIES 28.2.1 Layered Intercalation Electrodes Since the advent of the LiCoO2/C cell, several concerns have emerged on the toxicity and cost of cobalt. Similarly, the limited capacity (372 mAh g 1) and low voltage versus Li/Liþ for graphite prompted many researchers to look for alternative electrode materials in successive generations of Li-ion cells. For the cathode, the simplest choice is to partially or totally replace cobalt in layered material structurally related with the successful LiCoO2. In this way, LiNiO2 [1,2] and its solid solutions with LiCoO2 [3,4] were early considered as a potential option. However, the nickel compound has a more complex behavior, connected with possible nonstoichiometry and cation distribution, besides the existence of different deintercalated Li1 xNiO2 phases [1–4]. Partial replacement by iron was also evaluated. The room-temperature MS data of LiFexNi1 xO2 solids revealed a complex quadrupole split signal that could be reasonably well fitted to two doublets with isomer shift (IS) close to the expected value for Fe3þ in FeO6 coordination [5]. However, the values of quadrupole splitting (QS) differed notably, indicating a different environment around the iron ions. Each metal ion located between oxygen layers in LiFexNi1 xO2 is surrounded in-plane by other six closest metal neighbors. If we assume a random distribution of iron and nickel, the contribution of different environments such as FeFe6, FeFe5Ni, and FeFe4Ni2 (FeFenNi6 n) (Fig. 28.3) can be easily calculated according to the following equation: Pn ¼


 6 n x ð1 n

xÞ6 n :

(28.1)

For x ¼ 0.1, the three higher values resulting from Eq. (28.1) were obtained for n ¼ 0 (53.14%), n ¼ 1 (35.43%), and n ¼ 2 (9.84%). None of these values agrees with the relative intensities of the two doublets (86 and 14%) found experimentally for LiFe0.1Ni0.9O2. This discrepancy could imply a certain level of clustering of iron atoms, as it has been recently demonstrated for LiFexCo1 xO2 by using EPR data [6]. Simultaneously, Delmas et al. suggested that the M€ ossbauer spectra were fitted by considering a distribution of quadrupole splitting values. They also found that a statistical distribution did not fit to the experimental values, and a ossbauer spectroscopy study by G€utlich and coworkers bimodal distribution emerged [7]. A combined 61 Ni and 57 Fe M€ [8] unfolded the cationic site assignment in Li1 xNi1þxO2 and explained the ferromagnetic properties of this material because of the appearance of Ni2þ (S ¼ 1) among Ni3þ (S ¼ 1/2) in Ni3þO2 hexagonal planes. Besides the characterization of the pristine electrode materials, MS can give invaluable information about the electroactive atoms in used electrodes. In this way, it has been demonstrated that iron is also active in the doped LiNiO2 electrodes, and participates in a Fe3þ ! Fe4þ oxidation process during the charge process of lithium test cells [7]. 28.2.2 Phosphate Electrodes with Olivine Structure Due to its acceptable electrochemistry and low cost or toxicity, the compound LiFePO4 crystallizing with an olivinerelated structure is one of the most successful cathode materials in the short history of Li-ion batteries [9]. This material is already present in different commercial products, including energy storage for power tools and HEV [10].

28.2 CATHODE MATERIALS FOR Li-ION BATTERIES

555

FIGURE 28.4 € ssbauer spectra of (a) 57 Fe Mo (i) commercial LiFePO4 sample, (ii) ceramic LiFePO4 sample, and (iii) commercial sample heated at 750 (iii), 850 (iv), and 900  C (v) for € ssbauer spectra of 8 h. (b) 57 Fe Mo commercial LiFePO4 at different voltage values. Reproduced from Ref. 12 with permission of the Electrochemical Society.

The main drawback of this material is a low intrinsic electronic conductivity, which may lead to poor electrochemical performances. The synthesis procedure is also a sensitive point, as the oxidation of iron has to be avoided. A relatively simple method to deal with both the reducing synthesis conditions and the low conductivity of the product is to use carbon additives, also known as carbothermal process. Carbon additives have a two-fold effect. On the one hand, iron is protected against oxidation during the heating process. On the other hand, the use of carbon excess may provide an intimate mixture of LiFePO4 and carbon that has a much higher electronic conductivity than the phosphate alone. The carbothermal synthetic route may also have undesirable side effects. Thus, the carbothermal reduction of iron could yield iron phosphide contaminants, which are difficult to separate and compromise the 4 V capacity of the material. At this point, MS has been shown to be particularly effective in detecting and quantifying iron phosphide impurities and monitoring changes in the oxidation state of iron. The difference profiles of the M€ ossbauer spectra provide unequivocal identification of FeP and Fe2P phases (Fig. 28.4a) and their relative contributions depending on the temperature and other experimental conditions in which the synthetic work is carried out [11,12]. On the other hand, lithium deinsertion/insertion into LiFePO4 takes place through a two-phase reaction mechanism. Recent results indicate that by modifying the particle size and ion ordering (presence of defects and cation vacancies) in the material, the two-phase mechanism described at room temperature can be modified resulting in a single-phase mechanism [13]. 57 Fe MS has shown to be particularly effective to monitor the changes in the oxidation state of iron through charge–discharge cycles. It has been clearly demonstrated that the two oxidation states of iron involved in the electrochemical reaction coexist at intermediate depths of charge–discharge, and the reversibility of the process is also clear. From the hyperfine parameters, it was also shown that both Fe2þ and Fe3þ are basically in octahedral coordination and high-spin configuration throughout the process. Moreover, even in ZnO-coated materials, the same behavior is observed [12] (Fig. 28.4b). Unfortunately, for related solids such as LiCoPO4, an interesting 5 V cathode MS is not applicable. Other instrumental techniques such as Co K-edge XANES are better suited for the study of LiCoPO4, although they are outside the scope of this chapter. 28.2.3 Insertion Silicate Electrodes More recently than LiFePO4 and related phosphate compounds, the search for new cathode materials has been extended to lithium transition metal silicates. Again, these solids are characterized by a poor electric conductivity that made them being considered useless for this purpose. Probably as a consequence of the successful practical application of lithium iron

556

€ 28 APPLICATIONS OF MOSSBAUER SPECTROSCOPY IN THE STUDY OF LITHIUM BATTERY MATERIALS

phosphate with carbon additives and/or nanodispersion, the extension to silicates was expectable. Li2FeSiO4 has a lithium superionic conductor-related structure and gives a reversible electrochemical lithium extraction–reinsertion of 0.6 Li per formula at room temperature, in the 2.0–4.2 V range. Dominko [14] described the ideal half reaction leading to a maximum oxidation of iron to the þ3 oxidation state according to the following equation: Li2 FeSiO4 ! LiFeSiO4 þ Liþ þ e

(28.2)

Further extraction at higher voltages could be possible provided that a sufficiently stable electrolyte is found. By using nanoparticulate materials, a discharge capacity of 100 mAh g 1 at a current density of 2C has been reported [15]. The 57 Fe MS of fully lithiated Li2FeSiO4 consists of a single quadrupole split doublet ascribable to Fe2þ with IS close to 0.96 and QS to 2.50 mm s 1, which was interpreted in terms of only one local environment of iron [16]. Similarly, the spectrum of Li2FeSiO4 contains a doublet originating from Fe3þ ions, with IS close to 0.20 and QS to 0.60 mm s 1 [17]. Used electrodes in the fully charged and discharged states showed significant residual amounts of Li2FeSiO4 and LiFeSiO4. Magic angle M€ ossbauer measurements gave no indications of any texture in the sample, and thus the M€ ossbauer results confirmed 90% Fe2þ/Fe3þ conversion during charge–discharge [17].

28.3 ANODE MATERIALS FOR Li-ION BATTERIES 28.3.1 Conversion Oxides As mentioned above, the limited capacity (372 mAh g 1) and extremely low voltage versus Li/Liþ favoring electroplating reactions make the replacement of carbon electrodes in Li-ion cells necessary. One way to increase the capacity is the use of electrochemical reactions involving more than one electron per formula. A particular example of this type of reactions comes from the complete reduction to the metallic state of divalent or trivalent transition metal ions present in the structure of solids such as oxides or oxysalts. The potential to achieve this reduction is commonly close enough to the Li/Liþ pair to be considered as the negative electrode in Li-ion configurations avoiding electroplating phenomena. These materials, however, are yet far from true practical applications due to a series of concerns, especially a large irreversible capacity in the first cycle and high hysteresis in the charge–discharge branches. Among the binary oxides considered so far, cobalt oxides have a particularly good response [18]. A possible strategy to improve the electrochemical behavior while replacing the more expensive and toxic cobalt is the use of mixed oxides. In any case, the characterization of used electrode materials is complicated by the poorly crystalline nature of the reaction products, which limits the information provided by X-ray diffraction procedures. Then the use of 57 Fe MS offers unique possibilities of evaluating oxidation state, and chemical and magnetic local environment of iron atoms in the electrodes. ossbauer spectrum of well-crystallized ferrites such as CoFe2O4 or NiFe2O4 used as The room-temperature 57 Fe M€ starting materials [19–22] are commonly characterized by sextets related to their ferromagnetic character, while the MS spectra of electrodes discharged at about 0 V versus lithium undergo profound changes. The spectra of reduced electrodes show asymmetric profiles that can be fitted to two doublets (Fig. 28.4b). These signals have been interpreted in terms of a superparamagnetic material in which two different types of iron atoms are present. First, Fe0 atoms located at the core of the metallic nanometric particles show IS close to 0 mm s 1 and low QS, which agrees with an efficient electrochemical metal reduction and high-symmetry local environment, respectively. Then the surface atoms with higher IS and QS suggest either parallel metal oxidation of surface atoms in contact with the electrolyte, or a less efficient electrochemical reduction. The latter factor is expectable when the experimental capacities are lower than the theoretical value corresponding to full metal reduction. The loss of the hyperfine magnetic field in the reduced electrodes has its origin in a dramatic structural degradation produced during the discharge of lithium anode cells, which leads to a composite nanostructure in which nanometric iron particles are embedded in a paramagnetic Li2O matrix. The distances among adjacent particles are large enough to decrease the magnetic dipole interaction. Thus, the energy barrier to be overcome by the magnetization vector of a superparamagnetic iron particle decreases. Consequently, the magnetic moments in the nanometric iron particles are decoupled and the fluctuation of the magnetization vector in individual particles is favored. Additional evidence of these phenomena comes from the fact that the doublets become sextets on decreasing the recording temperature. The relative contribution of the sextets is coincident with that of the superparamagnetic

28.3 ANODE MATERIALS FOR Li-ION BATTERIES

557

FIGURE 28.5 57 € ssbauer spectra of cycled NiFe2O4 (600  C) electrodes after (a) one discharge and (b) 10 discharges. The spectra Fe Mo were recorded at 12, 120, and 200 K. Reproduced from Ref. 46 with permission of the American Chemical Society.

doublets. In turn, the relative contributions can be modeled in terms of a particular value of the thickness of the atomic surface layer on spherical particles of about 5 nm diameter, in agreement with electron microscopy observations. The blocking temperature is clearly influenced by the size of the iron domains, as electron spin cooperation in a large particle will counterbalance the thermal motion of the particle favoring the ferromagnetic character, while superparamagnetic relaxation will prevail in small particles. An additional relevant point is that the formation of bimetallic Ni–Fe or Co–Fe alloys cannot be inferred from the hyperfine field data. According to Jartych [23], the hyperfine field in body centered cubic Fe–Ni alloys increases by 0.94 T when one Ni atom substitutes an iron atom in the first coordination sphere and by 0.7 T, in the second sphere. A comparison of hyperfine field values for reduced FeFe2O4 (magnetite) and NiFe2O4 electrodes showed negligible differences (Fig. 28.5). After cell discharge to 0 V followed by a charge process to 3 V, the MS of the pristine sample is not recovered. Instead, two new doublets are visible with IS values ascribable to Fe3þ ions. The ultrafine nature of the oxidized particles prevents the ferromagnetic ordering to be retrieved. The origin of both signals with different QS values may be a consequence of a different spin state of the atoms in the structure of the oxide nanoparticles. Besides transmission MS, conversion electrons MS (CEMS) provides valuable information. 57 Fe CEMS shows significantly different spectra from the transmission mode. The observed average QS increases significantly, thus indicating that surface atoms located in anisotropic electric field mainly contribute to the CEMS data. This phenomenon is a definitive evidence of the different environment of iron atoms in the surface of the nanodispersed material found in the used electrodes [21]. ossbauer spectroscopy Finally, the behavior of these electrode materials on cycling can also be scrutinized by 57 Fe M€ in order to understand—and try to correct—undesirable phenomena such as capacity fading. The contribution of the superparamagnetic doublets in spectra recorded at a fixed intermediate temperature (200 K) increases on cycling, indicating a possible electrochemical grinding process that continuously disaggregates the Fe/Li2O composite material. This phenomenon could be responsible for the observed capacity fading [22].

558

€ 28 APPLICATIONS OF MOSSBAUER SPECTROSCOPY IN THE STUDY OF LITHIUM BATTERY MATERIALS

In recent years, the study of nanodispersed metal oxysalts as alternative conversion electrode material has unfolded a new group of compounds having interesting hybrid electrochemical responses [24–26]. Thus, the contribution of faradaic and nonfaradaic capacities makes these novel materials potentially attractive for both high-power and high-energy density applications. One of the first solids investigated in this group was the common and inexpensive anhydrous iron oxalate prepared in the form of nanoribbons by a reverse micelles procedure. The electrochemical reaction with lithium differs from conversion oxide electrodes by the formation of Fe/Li2C2O4 composites. Nevertheless, the properties of iron in the reduced and reoxidized electrodes resemble what was previously described for iron oxides. Thus, the quadrupole split Fe2þ signal in the spectra of iron oxalate transforms into two broadened doublets also ascribable to core and surface atoms in metallic nanoparticles. A notorious effect is that the reoxidation of the electrode does not recover the initial Fe2þ but Fe3þ is produced instead, as in the case of ferrites [25]. 28.3.2 Tin Alloys and Intermetallic Compounds Early reports revealed the electrochemical formation of Li–Sn intermetallics, which could provide theoretical capacities close to 1000 mAh g 1 in lithium cells [27]. A second impulse was received by a material patented by Fuji: tin composite oxides in which tin oxides were the starting material [28]. More recently, the commercialization of the NexelionTM battery by SONY gave new attraction to tin-based electrodes, with the special feature of being noncrystalline, which adds new value to the use of spectroscopic techniques. Since the late 1990s, a large number of papers have been published on the application of 119 Sn MS in the study of tin-based electrode materials. MS was found to be extremely useful in the analysis of the different steps commonly found during the electrochemical reaction of tin compounds with lithium: (i) reduction of tetra- or divalent tin atoms to the metallic state, followed by the most important step, (ii) a reversible formation of Li–Sn intermetallics. In order to perform a careful study of the various intermediate phases found during the electrochemical reaction with lithium, high-purity and stoichiometric Li–Sn intermetallics were prepared by nonelectrochemical means such as thermal or mechanical alloying and studied separately by 119 Sn MS. Dahn and coworkers [29] showed that the Li-rich compounds show less positive center shifts as a result of the smaller average number of Sn Sn bonds. Later, Jumas and coworkers [30] unfolded an IS–QS correlation diagram that can be used to identify the Li–Sn phases during the electrochemical process of new Sn-based materials. The decrease in isomer shift is commonly correlated with an increase in specific volume of the Li–Sn intermetallic (Fig. 28.6). As a result of the marked changes in specific volume, undesirable shrinking–expansion effects during cycling of lithium cells finally disintegrate the electrode with a tremendous loss of electrical contacts and electrode capacity. Among the different actions suggested to buffer volume changes and consequently improve capacity retention in tinbased electrodes, the use of additives and heteroatoms such as carbon, oxygen, and transition metals has given numerous examples of efficient performance of the electrodes. The interatomic interactions between tin and accompanying atoms can be successfully monitored by using 119 Sn MS. Thus, it is well established that tin isomer shift is affected by carbon and cobalt or lithium in different ways. In as early as 1984, Vertes and coworkers [31] elegantly showed that the electron density at the 5s orbital of the tin atom decreases as a consequence of electron transfer of the tin valence electrons to the conduction band of the cobalt–tin alloy. Further analysis was recently reported by showing that the decrease in IS with Co content is also affected by the temperature of reparation of Co–Sn intermetallics. Thus, it was shown that for a same stoichiometry CoSn3 a minimum in IS, which implies the maximum Sn–Co interaction, is observed at preparation temperature of 600  C [32]. On the contrary, the presence of carbon at the same temperature leads to a decrease in IS (Fig. 28.7). Even more important, the presence of heteroatoms may change the mechanism of electrochemical reaction with lithium. Composite electrodes ossbauer spectrum of the electrode discharged to 0 V having CoSn3 and carbon were obtained by ball milling. The 119 Sn M€ versus Li/Liþ showed IS and QS values that departed from the values reported for the maximum lithiated phase Li22Sn5, IS ¼ 1.837 mm s 1 and QS ¼ 0.27 mm s 1 [29]. Upon charge the barycenter IS was recovered but the signal was split into two doublets corresponding to Co-rich and Co-poor intermetallics [32]. An additional effect of great relevance in the electrochemical response of tin intermetallics is the role of nanodispersion. In this way, the maximum capacity achievable for CoSn electrodes was shown to increase significantly on decreasing the particle size (Fig. 28.8). For 10–20 nm, the capacity of the first discharge surpassed the theoretical capacity for CoSn. Moreover, for nanocrystalline CoSn, the reaction with lithium produces irreversible amorphization of the intermetallic compound. The presence of Co atoms avoids the formation of crystalline phases of LixSn, thus reducing the risks of enhanced volume changes. The two quadrupolar doublets found for micro-CoSn at about 0 V,

559

28.3 ANODE MATERIALS FOR Li-ION BATTERIES

Mean isomer shift (mm s–1)

2.6

β–Sn

2.5

LiSn

2.4

Li2Sn5

2.3 2.2 2.1

Li7Sn3

2.0

Li13Sn5 Li7Sn2

Li5Sn2

1.9 1.8

Li22Sn5

1.7 0

1

2 Li/Sn

3

4

5

Volume (mL) atom-gram of Sn

70 Li22Sn5

60

Li13Sn5 Li5Sn2

50 40

Li7Sn3

FIGURE 28.6

30 20 10

Li7Sn2

LiSn

Changes in isomer shift and specific volume of the Li–Sn intermetallics as a function of lithium content. Adapted from Ref. 52 with permission of Elsevier.

Li2Sn5

Sn 0.0

0.2

0.4

0.6

0.8

Li/(Li+Sn)

IS1 ¼ 1.85(3) and IS2 ¼ 2.05(3) mm s 1 were replaced by a single doublet signal in nano-CoSn with IS ¼ 1.93(3) mm s 1, indicative of a noncrystalline ternary LixCoSn phase as the main reaction product [33]. Finally, the electrochemical reaction of lithium with nano-CoSn3 prepared by a polyol method also leads to improved capacity retention as compared with pure tin. After complete discharge to 0.0 V, one narrow signal with IS ¼ 1.93 mm s 1 and QS ¼ 0.50 mm s 1 was observed, which again cannot be ascribed to Li22Sn5. Surprisingly, the value is close to the barycenter signal in the 119 Sn MS of the initial CoSn3, which was taken as indicative that the electron density change in tin

2.6

Center shift (mm s–1)

2.4 CoSn CoSn2 CoSn3 [CoSn3]0.6C0.4

2.2

2.0

×

+

×

×

× CoSn3+ etching + [CoSn3]0.6C0.4+ etching CoSn4 CoSn15

1.8

–100

+

0

100 200 300 400 500 600 700 800 900 1000 1100 Preparation temperature (°C)

FIGURE 28.7 Barycenter position of 119 Sn € ssbauer spectra for Mo tin-containing alloys and intermetallic compounds annealed at different temperatures. Reproduced from Ref. 32 with permission of Elsevier.

560

€ 28 APPLICATIONS OF MOSSBAUER SPECTROSCOPY IN THE STUDY OF LITHIUM BATTERY MATERIALS

3.0

CoSn 11 nm 24 nm 319 nm 319 nm (sieved < 50 µm)

Potential (V)

2.5 2.0 1.5 1.0

Expected capacity

0.5

FIGURE 28.8 Maximum capacity achievable for CoSn electrodes in lithium test cells as a function of particle size.

0.0 0

200

400

600

800

1000

1200

Q (mAh g–1)

atoms after reaction with Li is less extended for CoSn3 as compared with pure Sn, thus allowing a more stable electrode upon extended cycling [34]. ossbauer spectroscopies provides a unique For FeSn2, the possibility of a combined use of both 57 Fe and 119 Sn M€ situation. According to neutron diffraction and M€ ossbauer spectroscopy studies, bulk FeSn2 is antiferromagnetic below 378 K (Neel temperature), and a second magnetic transition occurs at 93 K. Between 378 and 93 K, the magnetic structure is collinear and characterized by ferromagnetic (100) planes, antiferromagnetically coupled along ossbauer the [100] direction. Below 93 K, FeSn2 becomes noncollinear antiferromagnetic [35]. However, the 57 Fe M€ spectrum of nanocrystalline FeSn2 recorded at 298 K shows a superparamagnetic singlet with the expected IS ¼ 0.51 ossbauer spectrum of nanocrystalline FeSn2 recorded at 298 K shows a (4) mm s 1. Moreover, the 119 Sn M€ broadened and low-intensity signal that also reflects a limited magnetic splitting at the tin nuclei [36]. Fe–Sn bonding with some covalent character reduces the electron densities at the nuclei and decreases the isomer shift value as compared with pure Sn [37]. The small particle size is probably the origin of a decrease in the recoil-free fraction of the tin atoms and the consequential decrease of signal intensity. Nanocrystalline FeSn2 displays interesting behavior in lithium test cells. Thus, superparamagnetic iron particles are formed during cell discharge and preserved after the recharge process. This behavior prevents particle aggregation and improves the cycling performance of FeSn2-based electrodes [36]. 28.3.3 Antimony Alloys and Intermetallic Compounds Almost simultaneous to tin, the theoretical capacity of pure antimony based on the formation of Li3Sb was known to be 660 mAh g 1 [38,39]. However, similar to tin, enhanced volume changes are found during the formation of Li–Sb compounds, leading to an important capacity fading upon cycling. Both the presence of other elements in the active material and nanostructuring contribute to reduce this phenomenon. For CoSb3, an initial capacity close to 500 mAh g 1 was experimentally observed [40], and then the capacity strongly decreased to 200 mAh g 1 after 20 cycles. The use of a nanostructured material improved cycling stability, and values over 400 mAh g 1 were reported after 20 cycles [41]. An additional improvement of the capacity retention upon cycling is obtained when multiwall carbon nanotubes (MWCTs) are added to CoSb3, especially after 20 cycles [42]. Concerning the application of M€ ossbauer effect to the study of CoSb3 electrodes, it should be noted that the 121 Sb spectral shape is complicated by the I ¼ 7/2þ ! 5/2þ nuclear transition and its natural linewidth. Nevertheless, the experimental M€ ossbauer spectrum of skutterudite is characterized by an IS value of 1.08(2) mm s 1 relative to the InSb [43,44]. By comparing this value to those plotted in Fig. 28.9, we observe that this shift is intermediate between those observed for elemental Sb and InSb where the Sb (5s) electrons are involved in covalent bonds. The QS value þ9.3(2) mm s 1 and the asymmetry factor g 0.65 suggest a strongly distorted environment. It corresponds to the geometrical distortion of the Sb(Co2, Sb2) tetrahedra and to the chemical difference between the antimony neighbors Co and Sb. During the first steps of lithiation, the hyperfine parameters suffer little changes, which have been interpreted in terms of an initial topotactic insertion of lithium up to Li0.2CoSb3 with a filled skutterudite structure. Further lithiation results in the splitting of the spectrum into two signals, ascribable to antimony atoms in Li3Sb and LixCoSby, respectively.

561

ACKNOWLEDGMENTS

1

Li3-xSb

55% 49%

53%

40%

δ (mm s–1)

0 45% –1 LiyMxSb

60%

–2

47% 51%

FIGURE 28.9

–3 Sb SnSb CoSb3 TiSb2 CrSb2

0

2

4 Li/Sb

6

10

11

Correlation between 121 Sb IS and the Li/Sb ratio. The percent numbers indicate the relative contributions of the different species. Reproduced from Ref. 47 with permission of Elsevier.

At the end of the first discharge, only the Li3Sb signal is present. An important phenomenon emerges during charge, as two components assigned to Li3Sb and LixCoSby can be observed again, and the relative intensity of the LixCoSby subspectra increases during lithium extraction. This complex process is a combination of insertion–conversion–alloying mechanism that has been found recently in different materials [44]. Antimonides with MSb2 stoichiometry have also been studied in lithium test cells, such as TiSb2 with u-Al2Cu structure [45], marcasite-CrSb2 [46,47], FeSb2 [48], and NiSb2 [49]. Fig. 28.9 summarizes the values of IS in 121 Sb MS as a function of the Li/Sb ratio for different antimonides. These results evidence the existence of two IS ranges, characteristic of LixMySb insertion compounds with IS between 0.7 and 2.4 mm s 1, and Li–Sb alloying at IS values around 1 mm s 1 [47]. Of special interest is SnSb, since both react with Li yielding intermetallics and both have isotopes showing M€ ossbauer effect. As revealed by 119 Sn MS, Li3Sb is formed first, then Li22Sn5 during the electrochemical reaction with Li. The pristine intermetallic compound SnSb is recovered after charging, with a reversible capacity of about 900 mAh g 1 for the first cycle [50]. Since the reversibility strongly depends on the availability of Sn atoms of the surface of Li2Sb particles to retrieve SnSb, the addition of Sn (i.e., Sn/SnSb) was used by Besenhard and coworkers to achieve a stable reversible capacity of about 600 mAh g 1 during at least 50 cycles [51].

28.4 CONCLUSIONS M€ ossbauer spectroscopy has become an irreplaceable tool in the study of electrode materials for Li-ion batteries. Relevant information is gathered on the changes in oxidation state and chemical environment of elements such as iron and tin, which are commonly present in various electroactive species. Particularly, the nature of nanocrystalline or amorphous solids involved in the electrochemical reactions, which are reluctant to be characterized by conventional X-ray diffraction methods, can be easily unfolded by M€ ossbauer spectroscopy. The use of M€ ossbauer spectroscopy leads to a better understanding of the reaction mechanism, which finally allows an optimized behavior of the electrode material. Notorious examples come from the study of nanodispersed materials, which provide large capacities and improved cyclability.

ACKNOWLEDGMENTS The authors would like to acknowledge the support of Spanish Ministry of Science and Innovation (MICINN), contract MAT2011–22753, Junta de Andalucıa Contract FQM-6017 (Research group FQM288), and European Institute ALISTORE.

562

€ 28 APPLICATIONS OF MOSSBAUER SPECTROSCOPY IN THE STUDY OF LITHIUM BATTERY MATERIALS

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.

H.G.S.R. Thomas, N.I.F. David, J.B. Goodenough, Mater. Res. Bull. 1985, 20, 1137. J.R. Dahn, U. von Sacken, C.A. Michal, Solid State Ionics 1990, 44, 87. T. Ohzuku, A. Ueda, N. Masatoshi, I. Yasunobu, H. Komori, Electrochim. Acta 1993, 38, 1159. R. Acantara, P. Lavela, J.L. Tirado, R. Stoyanova, E. Zhecheva, J. Electrochem. Soc. 1998, 145, 730. R. Alcantara, P. Lavela, C. Perez Vicente, J.L. Tirado, J. Olivier-Fourcade, J.C. Jumas, Solid State Commun. 2000, 115, 1. R. Alcantara, G. Ortiz, J.L. Tirado, R. Stoyanova, E. Zhecheva, S.V. Ivanova, J. Power Sources 2009, 194, 494. C. Delmas, G. Prado, A. Rougier, E. Suard, L. Fournes, Solid State Ionics 2000, 135, 71. utlich, Hyperfine Interact. 2002, 139, 107. V. Ksenofontov, S. Reiman, D. Walcher, Y. Garcia, N. Doroshenko, P. G€ A.K. Padhi, K.S. Nanjundaswamy, J.B. Goodenough, J. Electrochem. Soc. 1997, 144, 1188. http://www.a123systems.com/applications/hybrid-electric (July 20th, 2009). B. Le on, C. Perez Vicente, J.L. Tirado, Ph. Biensan, C. Tessier, J. Electrochem. Soc. 2008, 155, A211. C. Delmas, M. Maccario, L. Croguennec, F. Le Cras, F. Weill, Nat. Mater. 2008, 7, 665. M. Nakayama, S. Goto, Y. Uchimoto, M. Wakihara, Y. Kitajima, Chem. Mater. 2004, 16, 3399. R. Dominko, J. Power Sources 2008, 184, 462. S. Zhang, C. Deng, S. Yang, Electrochem. Solid-State Lett. 2009, 12, A136. M. Nadherna, R. Dominko, D. Hanzel, J. Reiter, M. Gaberscek, J. Electrochem. Soc. 2009, 156, A619. A. Nyten, S. Kamali, L. H€aggstr€ om, T. Gustafsson, J.O. Thomas, J. Mater. Chem. 2006, 16, 2266. P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, Nature 2000, 407, 496. R. Alcantara, M. Jaraba, P. Lavela, J.L. Tirado, J.C. Jumas, J. Olivier-Fourcade, Electrochem. Commun. 2003, 5, 16. P. Lavela, J.L. Tirado, J. Power Sources 2007, 172, 379. P. Lavela, J.L. Tirado, M. Womes, J.C. Jumas, J. Electrochem. Soc. 2009, 156, A589. C. Vidal-Abarca, P. Lavela, J.L. Tirado, J. Phys. Chem. C 2010, 114, 12828. E. Jartych, J. Magn. Magn. Mater. 2003, 265, 176. M.J. Arag on, C. Perez-Vicente, J.L. Tirado, Electrochem. Commun. 2007, 9, 1744. M.J. Arag on, B. Le on, C. Perez Vicente, J.L. Tirado, Inorg. Chem. 2008, 47, 10336. M.J. Arag on, B. Le on, C. Perez Vicente, J.L. Tirado, A.V. Chadwick, A. Berko, S.Y. Beh, Chem. Mater. 2009, 21, 1834. C.J. Wen, R.A. Huggins, J. Solid State Chem. 1980, 35, 376. Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, T. Miyasaka, Science 1997, 276, 1395. R.A. Dunlap, D.A. Small, D.D. MacNeil, M.N. Obravac, J.R. Dahn, J. Alloys Compd. 1999, 289, 135. F. Robert, P.E. Lippens, J. Olivier-Fourcade, J.C. Jumas, F. Gillot, M. Morcrette, J.M. Tarascon, J. Solid State Chem. 2007, 180, 339. J. Jaen, et al., Electrochim. Acta 1984, 29, 1119. G.F. Ortiz, R. Alcantara, I. Rodrıguez, J.L. Tirado, J. Electroanal. Chem. 2007, 605, 98. R. Alcantara, I. Rodrıguez, J.L. Tirado, ChemPhysChem 2008, 9, 1171. R. Alcantara, U. Nwokeke, I. Rodrıguez, J.L. Tirado, Electrochem. Solid-State Lett. 2008, 11, A209. B. Wodniecka, P. Wodniecki, A. Kulinska, A.Z. Hrynkiewicz, J. Alloys Compd. 2001, 321, 1. U. Nwokeke, R. Alcantara, J.L. Tirado, R. Stoyanova, M. Yoncheva, E. Zhecheva, Chem. Mater. 2010, 22, 2268. G. Trumpy, E. Both, C. Djega-Mariadassou, Phys. Rev. B 1970, 2, 3477. W. Weppner, R.A. Huggins, J. Electrochem. Soc. 1978, 125, 7. J. Wang, I.D. Raistrick, R.A. Huggins, J. Electrochem. Soc. 1986, 133, 457. R. Alcantara, F.J. Fernandez-Madrigal, P. Lavela, J.L. Tirado, J.C. Jumas, J. Olivier-Fourcade, J. Mater. Chem. 1999, 9, 2517. J. Xie, X.B. Zhao, G.S. Cao, S.F. Su, J. Electrochem. Soc. 2005, 152, A601. J. Xie, X.B. Zhao, G.S. Cao, M.J. Zhao, Electrochim. Acta 2005, 50, 2725. A. Kjekshus, D.G. Nicholson, T. Rakke, Acta Chem. Scand. 1973, 27, 1315. I. Devos, M. Womes, M. Heilemann, J. Olivier-Fourcade, J.C. Jumas, J.L. Tirado, J. Mater. Chem. 2004, 14, 1759. D. Larcher, L.Y. Beaulieu, O. Mao, A.E. George, J.R. Dahn, J. Electrochem. Soc. 2000, 147, 1703. F.J. Fernandez-Madrigal, P. Lavela, C. Perez Vicente, J.L. Tirado, J. Electroanal. Chem. 2001, 501, 205. L. Aldon, A. Garcia, J. Olivier-Fourcade, J.C. Jumas, F.J. Fernandez-Madrigal, P. Lavela, C. Perez Vicente, J.L. Tirado, J. Power Sources 2003, 119–121, 585.

REFERENCES

48. 49. 50. 51. 52.

563

J. Xie, X.B. Zhao, G.S. Cao, M.J. Zhao, Y.D. Zhong, L.Z. Deng, Mater. Lett. 2003, 57, 4673. C. Villevieille, C.M. Ionica-Bousquet, B. Ducourant, J.C. Jumas, L. Monconduit, J. Power Sources 2007, 172, 388. F.J. Fernandez-Madrigal, P. Lavela, C. Perez Vicente, J.L. Tirado, J.C. Jumas, J. Olivier-Fourcade, Chem. Mater. 2002, 14, 2962. M. Wachtler, M. Winter, J.O. Besenhard, J. Power Sources 2002, 105, 151. J.L. Tirado, Mater. Sci. Eng. R Rep. 2003, 40, 103.

C H A P T E R 2 9

€ MOSSBAUER SPECTROSCOPIC INVESTIGATIONS OF NOVEL BIMETAL CATALYSTS FOR PREFERENTIAL CO OXIDATION IN H2

WANSHENG ZHANG, JUNHU WANG, KUO LIU, JIE JIN, AND TAO ZHANG Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China

29.1 INTRODUCTION There is a wide interest in the development of novel heterogeneous catalysts that are applicable to the carbon monoxide (CO) preferential oxidation (PROX) in rich hydrogen atmosphere to improve the efficiency of fuel cells [1]. Among the reported heterogeneous catalysts for PROX reaction, supported bimetallic MFe (M ¼ Pt and Ir) nanoparticle catalysts are proved to be the most promising candidates [2–8]. A so-called “bifunctional mechanism” was proposed and it was pointed out that the mechanism was clearly different from that of the PROX reaction on supported single noble metal catalysts. It was also confirmed that noble metal forms the metallic clusters after H2 pretreatment or during the PROX reaction, whereas a large part of Fe exists as different state oxides even after H2 treatment [9]. The real information of iron active site during the PROX reaction is very important to understand the mechanism and design of an efficient catalyst. M€ ossbauer spectroscopy has been proven to be a useful tool for investigating the nature of supported metal catalysts, especially when these contain Fe, a commonly encountered M€ ossbauer isotope [10]. It is hoped that this can provide additional insight into the state of Fe active sites during the online or quasi in situ conditions. This chapter studies the quantitative relation between the real state of active sites and reaction condition and the possible mechanism for PROX reaction over the supported MFe catalysts by using the in situ Fe-M€ ossbauer spectra investigation.

29.2 EXPERIMENTAL SECTION 29.2.1 Catalyst Preparation Three kinds of novel alumina-supported PtFe, PtFe2, and PtFe3 alloy nanoparticles were prepared from Pt(acac)2 and Fe(acac)3 reduced by ethylene glycol in Ar atmosphere at 185  C [11,12]. The Ir–Fe bimetal catalysts used in this study M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

564

565

29.3 RESULTS AND DISCUSSION

were supported with SiO2. The Ir–Fe/SiO2 was prepared by coimpregnation as reported previously [13,14]. Briefly, SiO2 (SBET: 400 m2 g 1) was impregnated with an aqueous solution of chloroiridic acid and ferric nitrate, followed by drying at 80  C for more than 8 h and calcination at 300  C for 5 h in air. The loading amounts of Ir and Fe were 3 and 4.4 wt%, respectively, corresponding to an atomic Fe/Ir ratio of 5/1. To get a better signal-to-noise ratio in a relatively shorter ossbauer study, a desired amount of isotope 57 Fe(NO3)3 (which was obtained by dissolving 57 Fe foil period in the 57 Fe M€ into concentrated nitric acid) was used to replace a portion of Fe(NO3)3  9H2O in the impregnation procedure so that 57 Fe accounts for 40% of the total Fe amount in the Ir–Fe catalyst and 10% in the Fe/SiO2 catalyst. The activity tests for the 57 Fe-containing catalysts showed that the incorporation of isotope 57 Fe had no effect on the catalytic activity. 29.2.2 Catalytic Activity Test The catalytic activities were evaluated in a fixed-bed reaction system at atmospheric pressure. 100 mg of the catalyst diluted with 100 mg of SiC was loaded into a quartz reactor with the catalyst bed length of 10 mm. A total gas flow rate of 66.7 ml min 1 was employed, corresponding to a space velocity of 40,000 ml h 1 g-Cat 1. Prior to each test, the catalyst was reduced in situ with hydrogen at 300  C for 2 h. The feed gas mixture was composed of 2 vol% CO, 1 vol% O2, x vol% H2 (x ¼ 0–40), and balance He. In some cases, 10 vol% water was also introduced to the feed gas. The effluent gas was analyzed using an online gas chromatograph (Agilent GC-6890) equipped with a thermal conductivity detector. CO conversion (XCO) is calculated as XCO (%) ¼ {([CO]in–[CO]out) / [CO]in}  100. CO oxidation rate (mol h 1 g-Cat 1) is calculated as rCO ¼ XCO  YCO, in  Vgas / mcat, where mcat is the mass of the catalyst in grams, Vgas is the total molar flow rate in mol h 1, XCO is the conversion of CO, and YCO, in is the molar fraction of CO in the feed gas mixture. € ssbauer Spectra Characterization 29.2.3 Mo 

ossbauer spectra were recorded at room temperature using a Topologic 500 A spectrometer with a Quasi in situ57 Fe M€ proportional counter. 57 Co (Rh) moving in a constant acceleration mode was used as radioactive source. The Doppler velocity of the spectrometer was calibrated with respect to a-Fe foil. The spectra were fitted with appropriate superpositions of Lorentzian lines using the MossWinn 3.0i computer program [15]. In this way, M€ ossbauer parameters such as the isomer shift (IS), the electric quadrupole splitting (QS), the full linewidth at half maximum (FWHM), the magnetic hyperfine field (H), and the relative resonance areas of the different components of the absorption patterns (A) were determined. ossbauer measurement, a sample was loaded in a specially designed quartz reactor Before each quasi in situ57 Fe M€ similar to that described in [16], which enabled the measurements of a sample under various atmospheres without exposing to air. The quartz reactor has two stopcocks at the ends and a side cell made of polymethyl methacrylate (PMMA). First, the sample placed in the quartz reactor was reduced under 300  C for 2 h. After cooling to room temperature in He flow, both stopcocks were closed to seal the reactor filled with He gas. Then, the sample was moved into the side cell and a M€ ossbauer spectrum of the sample was measured. M€ ossbauer spectra were also collected for the catalysts after they were treated in the mixture containing 2 vol% CO, 1 vol% O2, and 0–40 vol% H2 in He in simulation to the reaction conditions.

29.3 RESULTS AND DISCUSSION 29.3.1 PtFe Alloy Nanoparticles Catalyst As can be known, the second additive can change the electric performance of a noble metal, so that the catalytic ability of the alloy catalyst will be improved a lot. In this section, we investigate the relation between the structural properties and the PROX catalytic performance of PtFe alloy nanoparticles catalysts supported with alumina. Table 29.1 shows the compositions of the samples used in this study. The weight percent of Pt was fixed as 3 wt% in the samples; three Fe/Pt atomic ratios (1/1, 2/1, and 3/1) of PtFe alloy nanoparticle catalysts were prepared. M€ ossbauer technique is effective for the identification of catalyst components in terms of active phase or active sites and search for correlations between these components and one or more of the catalytic properties [15,17–23]. Here, 57 Fe M€ ossbauer spectroscopy was employed to characterize the iron state in the as-prepared materials. Fig. 29.1 shows

566

€ 29 MOSSBAUER SPECTROSCOPIC INVESTIGATIONS OF NOVEL BIMETAL CATALYSTS FOR PREFERENTIAL CO

TABLE 29.1 Compositions of the Samples Used in this Study Weight Percent (wt%)

Pt/Fe Mole Ratio

PtFe nanoparticles 3% Pt/Al2O3 2.58% Fe/Al2O3 3% Pt–0.86% Fe/Al2O3 3% Pt–1.72% Fe/Al2O3 3% Pt–2.58% Fe/Al2O3

PtFe – – PtFe/Al2O3 PtFe2/Al2O3 PtFe3/Al2O3

100 98

Fe/Al2O3

Relative intensity (%)

96 100 98

PtFe3/Al2O3

96 100 99

PtFe2/Al2O3

98 100.0 99.5

FIGURE 29.1 57 € ssbauer spectra of the Fe Mo as-prepared alumina-supported PtFex alloy nanoparticles at room temperature, also including that of the as-prepared fcc disordered PtFe alloy nanoparticles and Fe/Al2O3.

PtFe/Al2O3

99.0 100.0 99.5 99.0 98.5

PtFe

–10

–5

0 Velocity (mm

5

10

s–1)

the 57 Fe M€ ossbauer spectra of the as-prepared alumina-supported PtFex alloy nanoparticles at room temperature. For a ossbauer spectroscopy. comparison, the PtFe (Pt/Fe ¼ 1/1) nanoparticles and Fe/Al2O3 were also characterized by 57 Fe M€ The spectrum of the PtFe nanoparticles was similar to that of the alumina-supported PtFe, PtFe2, and PtFe3 alloy nanoparticles, which mainly consist of one alloyed iron and one ferric quadrupole doublet. The spectrum of the Fe/Al2O3 sample exhibited a different pattern, which consists of one ferric and one ferrous quadrupole doublet. Small amounts of ferrous component were also observed in the spectra of the alumina-supported PtFe, PtFe2, and PtFe3 alloy nanoparticles. The ferrous component was possible to be produced by the addition of acidic alumina support since it was not observed in the spectrum of the Fe/Al2O3 sample. Taking into consideration on a very recent report that small monodispersed FeO nanoparticles were synthesized through reductive decomposition of Fe(acac)3 with oleic acid and oleylamine both as surfactants and solvents at 220–300  C [24], we can speculate that partial ferric irons were only reduced and stabilized in ferrous state due to the addition of g-Al2O3 even in the existence of a noble metal promoter such as Pt in the chemical process used in the present study. This is reasonable since the ethylene glycol and oleic acid or oleylamine are relatively soft reducers. The results also implied that a strong chemical interaction existed between the support of g-Al2O3 and the PtFe alloy nanoparticles. ossbauer parameters. The relative intensity of the ferrous species decreased Table 29.2 lists the evaluated 57 Fe M€ with the increase in Pt molar fractions contained in the as-prepared catalysts, that is, Fe2þ in PtFe/Al2O3 < PtFe2/ Al2O3 < PtFe3/Al2O3 < Fe/Al2O3, Pt molar fraction in PtFe/Al2O3 > PtFe2/Al2O3 > PtFe3/Al2O3 > Fe/Al2O3. It is known that a noble metal such as Pt can enhance the reduction of iron by intimately contacting with it in the supported bimetallic

567

29.3 RESULTS AND DISCUSSION

TABLE 29.2

57

€ ssbauer Parameters of PtFe Alloy Nanoparticles and Novel Alumina-Supported PtFex Alloy Fe Mo Nanoparticles at Room Temperature

Composition

Oxidation State of Iron

PtFe

Fe3þ Alloy Fe3þ Fe2þ Fe3þ Fe2þ Alloy Fe3þ Fe2þ Alloy Fe3þ Fe2þ Alloy

Fe/Al2O3 PtFe/Al2O3

PtFe2/Al2O3

PtFe3/Al2O3

IS (mm s 1)

QS (mm s 1)

RI (%)

0.32 0.35 0.32 1.17 0.30 1.15 0.35 0.33 1.16 0.33 0.32 1.05 0.18

0.96 0.34 0.88 2.06 0.91 1.99 0.35 0.87 2.02 0.32 1.04 1.95 0

50 50 82 18 50 3 47 75 10 15 60 15 25

IS, isomer shift; QS, electric quadrupole splitting; and RI, relative intensity.

catalysts [15]. The phenomenon was confirmed here by the 57 Fe M€ ossbauer technique. The IS and QS values of the alloyed iron were evaluated to be 0.35 and 0.35 mm s 1 for the PtFe/Al2O3 sample, 0.33 and 0.32 mm s 1 for the PtFe2/Al2O3 sample, and 0.18 and 0 mm s 1 for the PtFe3/Al2O3 sample, respectively. It means that the s-density at the iron nuclear position in the PtFe/Al2O3 and PtFe2/Al2O3 samples is smaller than that of the PtFe3/Al2O3 sample; however, the symmetry of the charge distribution at the iron nuclear position in the PtFe3/Al2O3 sample is higher than that of ossbauer results clearly indicated that the PtFe, PtFe2, and the PtFe/Al2O3 and PtFe2/Al2O3 samples [22]. These 57 Fe M€ PtFe3 nanoparticles supported on the alumina were still maintained in the alloyed state. Compared with the M€ ossbauer results of the catalysts with the same or similar composition prepared by the conventional method, the 57 Fe spectrum shapes of the Fe/Al2O3, PtFe/Al2O3, PtFe2/Al2O3, and PtFe3/Al2O3 samples are more similar to that reduced in H2 atmosphere above 400  C [15, 17, 18, 20]. As reported so far, all of the aluminasupported Fe and Pt–Fe catalysts prepared initially by the conventional method contained iron only in the ferric state; the ossbauer metallic iron and/or ferrous components appeared after the H2 or CO reduction, as determined by 57 Fe M€ spectroscopy [15]. XAFS characterization of the as-prepared Pt–Fe/zeolite catalyst, which was noted as “ion exchanged” catalyst by Kotobuki et al., also showed that Fe existed only as Fe3þ before the H2 pretreatment and the PROX reaction [16]. However, the as-prepared alumina-supported PtFex alloy nanoparticles in the present study contained iron not only in the ferric state but also in the alloyed iron and ferrous states. Furthermore, the oxidation states of iron in the asprepared samples were stable, which was also confirmed by the M€ ossbauer results of the PtFe nanoparticles kept in air before and after 1 month. The Fe2þ and alloyed iron existing in the catalysts should contribute to the higher PROX activity than the single Pt catalyst. Our catalytic performance results have proven this speculation [11]. The three bimetallic PtFex alloy nanoparticles supported on g-Al2O3 exhibited much higher improvements in both the CO conversion (40–120  C) and selectivity for CO2 at low temperature (40–80  C). The alumina-supported PtFe2 alloy nanoparticles showed the highest CO conversion at 40–100  C among the three alumina-supported bimetallic PtFex samples. 29.3.2 Ir–Fe/SiO2 Catalyst In this section, we systematically study the influence of H2 concentration on the oxidation state of Fe species and the CO ossbauer studies on PROX reaction [12,13] were not in situ and oxidation rate in Ir–Fe/SiO2 catalyst. Previous 57 Fe M€ could not exclude the effect of air on the iron state after H2 or PROX treatment. Nevertheless, our present study gives a ossbauer spectroscopy. In particular, by employing quasi more reliable result of iron state by employing quasi in situ 57 Fe M€ in situ 57 Fe M€ ossbauer spectroscopy, we give a quantitative analysis of different Fe species after Ir–Fe catalyst was subjected to treatments with various feed gas compositions. Correlating the relative amount of various Fe species with the reaction rate of PROX, we provide strong evidence that the Fe2þ in close contact with Ir particles is responsible for providing reactive oxygen in PROX reaction. To the best of our knowledge, this is for the first time that quantitative

568

€ 29 MOSSBAUER SPECTROSCOPIC INVESTIGATIONS OF NOVEL BIMETAL CATALYSTS FOR PREFERENTIAL CO

correlation between Fe2þ and the reaction rate has been established for the Fe-promoted catalysts in the PROX reaction. 29.3.2.1 Effect of H2 Concentration on CO Oxidation and the Adsorption of H2, CO and O2 The influence of H2 concentration on CO oxidation rate was reported earlier for gold-based catalysts [25,26] and MOx-promoted noble metal catalysts (under PROX conditions, i.e., H2 vol% > 50%) [27,28]. To gain a deep insight into the effect of H2 on the CO oxidation, we made a systematic study by varying the H2 volume fractions spanning from 0 to 40% in the feed gas. Fig. 29.2 illustrates CO conversions with the reaction temperature under different feed gas compositions over Ir–Fe/SiO2 and Ir/SiO2 catalysts. For Ir–Fe/SiO2 (Fig. 29.2a), CO conversion was low without the presence of H2. Even at an elevated temperature up to 220  C, the CO conversion was lower than 40%. When 2% H2 was present in the feed gas, however, the CO conversion was increased significantly, and the highest CO conversion of 85% was obtained at 140  C. Increasing the H2 concentration from 2 to 10% led to a further increase in low-temperature CO conversions and the temperature for maximum CO conversion shifting to 120  C. When the H2 concentration was further increased to 40%, the CO conversions below 100  C remained essentially the same as those in the case of 10% H2, while the CO conversions above 100  C had a remarkable decrease, which can be attributed to the competitive effect of H2 oxidation becoming pronounced with increasing H2 concentration and reaction temperature [25]. Our activity results on the Ir–Fe/SiO2 catalyst clearly showed that the presence of H2 indeed greatly promoted the low-temperature CO oxidation, and this promotional effect was closely related with the H2 concentration in the feed gas. For Ir/SiO2 catalyst (Fig. 29.2b), a positive effect of H2 on the CO conversion was also observed. However, compared with Ir–Fe/SiO2, this effect was much less remarkable, in particular at temperatures lower than 140  C. In addition, with an increase in the H2 concentration, the enhancement of CO conversion became smaller and smaller due to the competition of H2 oxidation. This trend was different from the case of Ir–Fe/SiO2. Kim and Lim [28] studied CO oxidation in a H2-rich mixture on Pt/Al2O3 catalyst and observed that the CO conversion did not change with H2 concentration from 10 to 50%, which was different from our observation on the Ir/SiO2 catalyst but in agreement with our result obtained on the Ir–Fe/SiO2 below 100  C. However, they did not consider the effect of H2 at lower H2 concentrations. To address the promotional role of Fe on Ir/SiO2 catalyst, we first study the adsorption behaviors of reactant molecules (CO, O2, and H2) by microcalorimetry [29] on the Ir–Fe/SiO2 in comparison with the Ir/SiO2. The results are shown in Table 29.3. The initial adsorption heat and the saturation uptake are 90 kJ mol 1 and 55 mmol g-Cat 1 for H2, and 140 kJ mol 1and 108 mmol g-Cat 1 for CO on Ir/SiO2 catalyst. This result indicates that CO is adsorbed more strongly than H2 on the Ir/SiO2 catalyst. In contrast with Ir/SiO2, the adsorption of either H2 or CO is greatly weakened on the Ir–Fe/SiO2 catalyst. Actually, H2 cannot be chemically adsorbed on the Ir–Fe/SiO2 catalyst at all, and the CO saturation uptake is only 20 mmol g-Cat 1, decreased by more than 80% compared with that on the Ir/SiO2. This result is consistent with PtFe catalyst where the addition of Fe causes a large decrease in CO adsorption [4]. On the other hand, O2 adsorption on the Ir–Fe/SiO2 catalyst is significantly improved with the saturation uptake increasing from 55 mmol g-Cat 1 on Ir/SiO2 to 240 mmol g-Cat 1 on Ir–Fe/SiO2. This result demonstrates that the addition of Fe facilitates the adsorption of oxygen, in good agreement with the results obtained by chemisorption under flow gas conditions [12].

100

(a)

80

FIGURE 29.2 CO conversions as a function of reaction temperature over Ir–Fe/SiO2 (a) and Ir/SiO2 (b) catalysts. The reaction feed gas contains various volume fractions of H2: ( ) 0%, (&) 2%, ( ) 10%, and (~) 40%; GHSV ¼ 40,000 ml h 1 gCat 1.



Î

CO conversion (%)

60 40 20 0 100 80

(b)

60 40 20 0 40

60

80

100

120

140

160

Temperature (°C)

180

200

220

240

569

29.3 RESULTS AND DISCUSSION

TABLE 29.3 The Saturation Uptakes and Initial Adsorption Heat of CO, O2, and H2 on Freshly Reduced Ir–Fe/SiO2 and Ir/SiO2 Catalysts H2 Absorbate 1

Differential heat (kJ mol ) Coverage (mmol g-Cat 1)

CO

O2

Ir–Fe/SiO2

Ir/SiO2

Ir–Fe/SiO2

Ir/SiO2

Ir–Fe/SiO2

Ir/SiO2

25 –

90 55

120 20

140 108

400 240

389 55

€ ssbauer Spectra of Ir–Fe/SiO2 and Fe/SiO2 Treated with H2 or O2 Fig. 29.3 displays the quasi in 29.3.2.2 Mo ossbauer spectra of Ir–Fe/SiO2 catalyst before and after H2 reduction, as well as reoxidation with O2. For situ57 Fe M€ comparison, the spectra of Fe/SiO2 sample pretreated under similar conditions are presented in Fig. 29.4. The corresponding M€ ossbauer parameters are listed in Tables 29.4 and 29.5. For the calcined Ir–Fe/SiO2 catalyst (Fig. 29.3a), the spectrum consists of a doublet and a sextuplet, which can be assigned to superparamagnetic and relatively large particles of a-Fe2O3, respectively. After reduction with H2 at 300  C for 2 h, the spectrum (Fig. 29.3b) is composed of a doublet (IS ¼ 0.58 mm s 1 and QS ¼ 0.71 mm s 1) for Fenþ (2 < n < 3), a doublet (IS ¼ 1.13 mm s 1 and QS ¼ 2.15 mm s 1) for Fe2þ [23], a sextuplet for Fe0, and a singlet for FeIr alloy. Niemantsverdriet et al. [30,31] studied the structure of FeIr/SiO2 catalyst (atomic ratio Fe/Ir ¼ 1/1) during CO hydrogenation by M€ ossbauer spectroscopy and also detected Fe0, Fe2þ, and FeIr alloy in the 400  C-reduced catalyst. In a similar study by Berry and Jobson [32], both Fe3O4 and Fe2þ were identified on the 270  C-reduced IrFe/SiO2 catalyst with low concentration of Ir. With an increase in the Ir content in the IrFe catalyst, Fe0 or FeIr alloy was also detected on the reduced catalyst. These results were basically in agreement with our present study. Clearly, the treatment with H2 at 300  C reduced Fe3þ species to some low-valence iron species. The reduced catalyst was subsequently treated with 1% O2 at 80  C for 1 h for investigation of the reactivities of different low-valence Fe species toward oxygen. The resulting spectrum (Fig. 29.3c and Table 29.4) was an overlap of Fe3þ (76%) and Fe0 (24%). The disappearance of Fenþ (2 < n < 3), Fe2þ, and FeIr alloy upon oxygen treatment strongly suggests that these low-valence Fe species are easily oxidized. It is interesting to note that there is still a significant amount of Fe0 retained in the Ir–Fe/SiO2 catalyst even after the oxygen treatment. It seems that only some outer layers of Fe0 particles were oxidized upon exposure to 1% O2 atmosphere at 80  C for 1 h, while the cores of the particles remained as Fe0. Wang et al. [33] pointed out that a critical size existed for which the iron could be fully oxidized, whereas for particles larger than the critical size (8 nm), an iron/iron oxide structure could be found. Different from Ir–Fe/SiO2, Fe/SiO2 still showed a typical Fe3þ doublet spectrum (Fig. 29.4a and Table 29.5) after reduction with H2 at 300  C. Only when reduced at 600  C, low-valence Fe species including Fenþ, Fe0, Fe2þ (1) (IS ¼ 1.10 mm s 1 and QS ¼ 2.25 mm s 1), and Fe2þ (2) (IS ¼ 1.06 mm s 1 and QS ¼ 1.69 mm s 1) could be detected (Fig. 29.4b). Moreover, the relative amount of Fe0 in the Fe/SiO2 catalyst was significantly lower than that in the

3+

Fe (mg) 3+

Relative intensity (%)

Fe (sp)

(a)

0

Fe n+ Fe 2+ Fe FeIr

(b) (c)

–10

–5

0 Velocity (mm s–1)

5

10

FIGURE 29.3 € ssbauer Quasi in situ 57 Fe Mo spectra of Ir–Fe/SiO2 catalyst (a) before and (b) after 100% H2 reduction at 300  C for 2 h, and (c) followed by 1% O2 treatment at 80  C for 1 h after (b).

570

TABLE 29.4

€ 29 MOSSBAUER SPECTROSCOPIC INVESTIGATIONS OF NOVEL BIMETAL CATALYSTS FOR PREFERENTIAL CO

57

€ ssbauer Parameters of Ir–Fe/SiO2 Catalyst Fe Mo

Treatment

IS (mm s 1)

Chemical State 3þ



Fe (sp) Fe3þ (mg) Fenþ (2 < n < 3) Fe0 Fe2þ FeIr Fe3þ Fe0

After calcination at 300 C for 5 h After 100% H2 reduction at 300  C for 2 h

1% O2 at 80  C for 1 h

0.33 0.36 0.58 0.00 1.13 0.11 0.34 0.00

QS (mm s 1)

MF (T)

RI (%)

0.87 0.40 0.71 0.00 2.15 – 0.94 0.00

– 52.0 – 32.9 – – – 33.0

87 13 20 43 26 11 76 24

IS, isomer shift; QS, electric quadrupole splitting; MF, magnetic field; RI, relative intensity; sp, superparamagnetic; and mg, magnetic. Uncertainty is 5% of reported value.

Fe

3+

Relative intensity (%)

(a)

FIGURE 29.4 57

€ ssbauer Quasi in situ Fe Mo spectra of Fe/SiO2 (a) after 100% H2 reduction at 300  C for 2 h, (b) after 100% H2 reduction at 600  C for 2 h, and (c) followed by 1% O2 treatment at 80  C for 1 h after (b).

TABLE 29.5

57

0

Fe n+ Fe Fe

2+

(1)

2+

(2)

Fe

(b) (c)

–10

–5

0 Velocity (mm s–1)

5

10

€ ssbauer Parameters of Fe/SiO2 Catalyst Fe Mo

Treatment

Chemical State 

After reduction at 300 C for 2 h After reduction at 600  C for 2 h

1% O2 at 80  C for 1 h

3þ

Fe Fenþ (2 < n < 3) Fe0 Fe2þ (1) Fe2þ (2) Fe3þ Fe2þ (1) Fe2þ (2)

IS (mm s 1)

QS (mm s 1)

MF (T)

RI (%)

0.34 0.53 0.00 1.10 1.06 0.35 0.98 0.95

0.87 1.38 0.00 2.25 1.69 0.92 2.40 1.73

– – 32.4 – – – – –

100 29 10 30 31 44 40 16

IS, isomer shift; QS, electric quadrupole splitting; MF, magnetic field; and RI, relative intensity. Uncertainty is 5% of reported value.

Ir–Fe/SiO2. This result indicates that the presence of Ir facilitates the reduction of Fe species, consistent with the previous H2-TPR results [13]. When the 600  C-reduced Fe/SiO2 sample was exposed to 1% O2 at 80  C for 1 h, Fe0 disappeared, while the amount of Fe3þ was increased (Fig. 29.4c) due to the oxidation of Fe0 and Fenþ. The complete oxidation of Fe0 in the 600  C-reduced Fe/SiO2 sample implies that the particle size of this part of Fe0 might be smaller than that in the reduced Ir–Fe/SiO2 catalyst [33]. To our surprise, the total amount of Fe2þ remained almost unchanged, although Fe2þ (1) increased at the expense of Fe2þ (2). This result strongly suggests that the low-valence Fe species in the Fe/SiO2, in particular Fe2þ, is more difficult to be reoxidized than those in the Ir–Fe/SiO2 catalyst. In other words, the Fe species in the Ir–Fe/SiO2 are highly reactive toward reduction by H2 and reoxidation by O2.

571

29.3 RESULTS AND DISCUSSION

€ ssbauer Spectra of Ir–Fe/SiO2 Treated with Reactant Gas Mixtures After reduction, the Ir–Fe 29.3.2.3 Mo ossbauer catalyst was subsequently treated with various reactant gas mixtures at 80  C for 1 h. The resulting 57 Fe M€ spectra are shown in Figs. 29.5 and 29.6, and the corresponding parameters are summarized in Table 29.6. Interestingly, all the low-valence iron components were retained when 1% O2 and more than 2% H2 were fed simultaneously (Fig. 29.5). Compared with the result obtained with 1% O2 treatment, the addition of H2 clearly stabilized Fenþ, Fe2þ, and FeIr alloy. Compared with the freshly reduced catalyst, the subsequent treatment with a mixture of H2 and O2 caused changes of the relative amounts of intermediate (Fenþ), ferrous, metallic, and alloyed state iron components. Moreover, the IS values of Fenþ also deceased with decreasing H2 volume fractions, that is, 0.58 mm s 1 (100%) > 0.48 mm s 1 (40%) > 0.42 mm s 1 (2%). This trend indicates that the intermediate iron component has the tendency toward a higher oxidation state (Fe3þ) with decreasing H2 concentration in the treatment gas mixture. When the freshly reduced catalyst was exposed to 2% CO at 80  C for 1 h, the M€ ossbauer spectrum was similar to that treated with a mixture of 2% H2 and 1% O2. In particular, the relative amounts of Fe0, Fe2þ, and FeIr alloy were almost identical in both cases. The only difference was the identification of Fe3þ species, although its IS value of 0.39 mm s 1 was larger than the typical ferric component identified in the present study. It was reported that CO could be dissociated on Fe0 sites, forming carbon and oxygen atoms [34]. The oxygen atoms would then react with iron atom forming iron oxide, resulting in higher iron oxidation states. When the reduced catalyst was exposed to a mixture of CO, O2, and H2 (0–40%) at 80  C for 1 h, that is, the typical reaction condition for CO oxidation or PROX (Fig. 29.6), both FeIr alloy and intermediate iron Fenþ disappeared completely. All the spectra could be well fitted into Fe0, Fe2þ, and typical Fe3þ in superparamagnetic a-Fe2O3. It can be seen in Table 29.4 that the amount of Fe0 remained steady around 28% for different H2 concentrations used in the feed 0

Fe

n+

Fe

2+

Relative intensity (%)

Fe

FeIr

2% CO

40% H2 + 1% O2

FIGURE 29.5 € ssbauer Quasi in situ 57 Fe Mo spectra of reduced Ir–Fe/SiO2 catalyst upon exposure to different atmospheres at 80  C for 1 h.

2% H2 + 1% O2 –10

–5

0 Velocity (mm s–1)

5

10

0

Fe

3+

Fe

2+

Relative intensity (%)

Fe

40% H2 + 1% O2 + 2% CO 10% H2 + 1% O2 + 2% CO 2% H2 + 1% O2 + 2% CO 1% O2 + 2% CO

–10

–5

0 Velocity (mm s–1)

5

10

FIGURE 29.6 € ssbauer Quasi in situ 57 Fe Mo spectra of reduced Ir–Fe/SiO2 catalyst upon exposure to reaction atmosphere containing different concentrations of H2 at 80  C for 1 h.

572

TABLE 29.6

€ 29 MOSSBAUER SPECTROSCOPIC INVESTIGATIONS OF NOVEL BIMETAL CATALYSTS FOR PREFERENTIAL CO

57

€ ssbauer Parameters of Ir–Fe/SiO2 Catalyst upon Exposure to Different Atmosphere Fe Mo

Treatment

Chemical State 

40% H2 þ 1% O2 at 80 C for 1 h

2% H2 þ 1% O2 at 80  C for 1 h

2% CO at 80  C for 1 h

40% H2 þ 2% CO þ 1% O2 at 80  C for 1 h 10% H2 þ 2% CO þ 1% O2 at 80  C for 1 h 2% H2 þ 2% CO þ 1% O2 at 80  C for 1 h 2% CO þ 1% O2 at 80  C for 1 h

nþ

Fe (2 < n < 3) Fe0 Fe2þ FeIr Fenþ (2 < n < 3) Fe0 Fe2þ FeIr Fe3þ Fe0 Fe2þ FeIr Fe3þ Fe0 Fe2þ Fe3þ Fe0 Fe2þ Fe3þ Fe0 Fe2þ Fe3þ Fe0 Fe2þ

IS (mm s 1) 0.48 0.00 1.13 0.14 0.42 0.00 1.10 0.12 0.39 0.00 1.10 0.15 0.35 0.00 1.12 0.35 0.00 1.10 0.36 0.00 1.10 0.34 0.00 1.11

QS (mm s 1)

MF (T)

RI (%)

0.73 0.00 2.16 – 0.73 0.00 2.19 – 0.72 0.00 2.14 – 0.92 0.00 2.16 0.92 0.00 2.16 0.92 0.00 2.15 0.95 0.00 2.13

– 33.0 – – – 33.1 –

29 40 19 12 41 29 23 7 49 23 21 7 54 28 18 56 27 17 60 26 14 67 29 4

– 32.8 – – – 32.9 – – 32.9 – – 32.9 – – 32.9 –

IS, isomer shift; QS, electric quadrupole splitting; MF, magnetic field; and RI, relative intensity. Uncertainty is 5% of reported value.

gas mixtures. At the same time, the amount of Fe3þ decreased from 67 to 54%, whereas the amount of Fe2þ increased from 4 to 18% with an increase of H2 concentration from 0 to 40%. 29.3.2.4 Possible Reaction Mechanism It is widely accepted that CO oxidation on an unpromoted noble metal catalyst follows a competitive Langmuir–Hinshelwood mechanism where CO, H2, and O2 are all adsorbed on the noble metal surface. In the case of Ir/SiO2 catalyst, the stronger adsorption of CO than H2 [13] would be favorable to the preferential oxidation of CO even under the presence of rich H2. At the same time, however, the strong adsorption of CO makes the Ir sites almost fully covered by CO at low temperatures, which leads to few sites available for oxygen adsorption [5]. In such a case, the CO conversions on the Ir/SiO2 at low temperatures are very low, as shown in Fig. 29.2b. To enhance the activity of low-temperature CO oxidation on Ir, the adsorption of CO on Ir sites must be weakened so that O2 has enough opportunities to adsorb on them. For this purpose, the second metal Fe is added. Both microcalorimetry and DRIFT studies have shown that the addition of Fe indeed weakens the adsorption of CO greatly [13]. In this case, the possibility of the reaction between adsorbed CO and O on the neighboring Ir sites, that is, competitive Langmuir–Hinshelwood reaction, should have been enhanced. However, the activity tests showed that the Ir–Fe/SiO2 catalyst was even less active than the Ir/SiO2 catalyst without the presence of H2. In other words, Fe did not impose any beneficial effect on CO oxidation without the presence of H2. This result clearly indicates that the promotional effect of Fe is associated with the presence of H2. Actually, the H2 concentration in the reaction stream ossbauer spectroscopic strongly influenced the distribution of different Fe species in the Ir–Fe/SiO2 catalyst. The 57 Fe M€ studies gave a quantitative analysis of the Fe species in the Ir–Fe/SiO2 catalyst treated with reaction gases containing different fractions of H2 (Table 29.6). To better correlate different Fe species with the catalytic activity, we plot the reaction rate and the concentration of Fe species in the reacted Ir–Fe/SiO2 catalyst as a function of the H2 concentration in the feed stream. As shown in Fig. 29.7, Fe3þ continuously decreases while Fe2þ increases with increasing the H2 concentration from 0 to 40%. Consistent with the change of Fe2þ amount, the reaction rate for CO oxidation increases linearly with the H2 concentration until 10% and then levels off at higher H2 concentrations. This result strongly suggests

573

29.3 RESULTS AND DISCUSSION

70 60

0.006 Fe 2+ Fe 3+ Fe r

50 40

0.004

30 20

0.002

10 0

0

10

20 30 H2 fraction (%)

40

0.000

r (mol h–1 gcat–1)

Concentration (%)

0

FIGURE 29.7 Plots of relative amounts of various Fe species determined by 57 € ssbauer spectra and CO Fe Mo oxidation reaction rates (r) against H2 volume fractions in the reaction stream.

that Fe2þ in the Ir–Fe/SiO2 has a close relation with the catalytic activity and, furthermore, the amount of Fe2þ in the Ir– Fe/SiO2 depends on the H2 concentration in the feed gas. Many researchers proposed that the middle valence metal oxide was responsible for O2 activation in the bimetallic catalysts. Kotobuki et al. [16] suggested that Pt formed the metallic clusters after H2 pretreatment or PROX reaction, whereas a large part of Fe existed as FeO after PROX reaction according to XAFS results. For PtSn bimetallic catalyst, the suggested mechanism involved O2 adsorption predominantly on Sn/SnOx islands on or adjacent to the active Pt sites [6]. Analogously, the redox changes between Cu2þ and Cuþ have been considered as the main steps in CO oxidation reaction mechanism for the Cu-based catalysts [35]. In agreement ossbauer spectra and the with these reports, the positive correlation between the amount of Fe2þ determined by M€ reaction rate also suggests that Fe2þ in the Ir–Fe catalyst acts as the adsorption site for O2. Different from the changes of either Fe3þ or Fe2þ, the amount of Fe0 remains constant at roughly 28% in the range of H2 concentrations investigated. This result indicates that this part of Fe0 did not participate in the PROX reaction. Actually, although the freshly reduced Ir–Fe/SiO2 catalyst contains 43% Fe0, the subsequent treatment with either 1% O2 or a mixture of 2% H2 and 1% O2 oxidized only a part of Fe0. In other words, there is always 25% Fe0 remaining intact in ossbauer the Ir–Fe/SiO2 catalyst. According to the literature [33] and the magnetic field value we detected in the 57 Fe M€ spectra, we assign this part of Fe0 to that encapsulated by the ferric oxide. Since Fe2þ might be responsible for providing reactive oxygen for CO oxidation, we now need to understand the ossbauer unique features of Fe2þ in the Ir–Fe/SiO2 catalyst. To address this issue, we performed a comparative M€ spectroscopic study on the Ir–Fe/SiO2 and Fe/SiO2 catalysts. The results demonstrate that Fe3þ in the former catalyst can be reduced with H2 much more easily than that in the latter catalyst, and the reduced Fe species including Fenþ, a portion of Fe0, Fe2þ, and FeIr alloy are easily reoxidized with O2. In agreement with the previous H2-TPR results [13], the M€ ossbauer results also show that Ir facilitates the reduction of Fe3þ and the microcalorimetric study indicates that the addition of Fe greatly enhances O2 adsorption. These results strongly suggest that the reduced (low valence) Fe species in the Ir–Fe/SiO2 catalyst could activate oxygen at low temperatures and be easily regenerated upon exposure to H2. Moreover, they must be in intimate contact with the Ir particles so that hydrogen spillover from Ir to Fe takes place easily. The disappearance of FeIr alloy upon exposure to CO oxidation (or PROX) atmosphere also implies that the active Fe2þ probably originates from the oxidation of FeIr alloy, which forms intimately contacted Fe2þ oxide and Ir particles in the reaction atmosphere. Analogously, Margitfalvi et al. [36] suggested that a reversible PtSn $ Sn4þ þ Pt interconversion took place in CO oxidation. Besides this part of Fe2þ arising from FeIr alloy, there may be another portion of Fe2þ contributing to PROX reaction as long as it is adjacent to the Ir particles. For example, the vanishing half amount of Fe0 upon exposure to the reaction atmosphere may be transformed into Fe2þ and then serves as the site for oxygen activation. However, Fenþ in the freshly reduced Ir–Fe/SiO2 catalyst is more probably transformed into Fe3þ than Fe2þ upon exposure to the reaction atmosphere, and therefore cannot be an active site for oxygen. Such a conclusion can be deduced from the decreasing tendency of IS values of Fenþ with decreasing H2 concentration in the treatment gas. In summary, when a reduced Ir–Fe/SiO2 catalyst is used for PROX reaction, the changes of Fe species in the catalyst may occur as follows: (1) the Fenþ is reoxidized to Fe3þ, which is inactive for oxygen activation; (2) the FeIr alloy is oxidized and segregated into Fe2þ and metallic Ir, which are active for CO oxidation; (3) the outer layers of large Fe0

574

€ 29 MOSSBAUER SPECTROSCOPIC INVESTIGATIONS OF NOVEL BIMETAL CATALYSTS FOR PREFERENTIAL CO

particles are oxidized into Fe2þ or Fe3þ while the cores remain intact as Fe0 and therefore do not take part in the PROX reaction; (4) the Fe2þ in intimate contact with Ir particles is also active for providing reactive oxygen for PROX, whereas the left Fe2þ may be oxidized to Fe3þ and therefore has little to do with the PROX reaction. Thus, for PROX reaction over the Ir–Fe/SiO2 catalyst, Ir acts as the site for CO adsorption while Fe2þ is responsible for providing the reactive oxygen; the reaction proceeds via a noncompetitive Langmuir–Hinshelwood reaction route. In the absence of H2, the lack of Fe2þ sites leads to a slow reaction rate for CO oxidation. The presence of H2, even in slight excess, can provide enough Fe2þ sites for the reaction and greatly promote the reaction. In addition to stabilizing Fe2þ in the Ir–Fe/SiO2 catalyst, H2 may promote the formation of OH groups on the catalyst surface. In fact, a slight discrepancy existing between the activity and the Fe2þ content at the point of 2% H2 suggests that enough amount of Fe2þ on the catalyst surface might not be the only factor determining the high activity of Ir–Fe/SiO2. In the absence of any promoters, a similar promotional effect of H2 was also observed on Au/Al2O3 catalyst. In that case, hydrogen is proposed to be directly involved in the reaction [37,38]. The positive effect of OH groups on the CO oxidation has also been claimed for several Pt-based catalysts [39,40] where the hydroxyl groups might be consumed during CO oxidation through the formation of active intermediate (–COOH) that decomposes to CO2 [41]. Therefore, we cannot exclude the possibility that OH groups formed by the reaction of H2 with O2 also act as oxidants for CO. Thus, the increase of OH groups by the addition of H2 in the feed might be another factor for promoting the CO oxidation rate.

29.4 CONCLUSIONS PtFe/Al2O3, PtFe2/Al2O3, and PtFe3/Al2O3 materials prepared from Pt(acac)2, Fe(acac)3, and ethylene glycol as a soft reducer in Ar at 185  C were confirmed to be more active for the PROX reaction as compared with that of Pt/Al2O3 material prepared by the same chemical route. The XRD and 57 Fe M€ ossbauer results indicate that PtFe, PtFe2, and PtFe3 nanoparticles supported on alumina are in high dispersion and exist as alloyed state. The oxidation states of iron contained in them are not only in the alloyed Fe0 state but also in the ferric and ferrous states as determined by 57 Fe M€ ossbauer spectroscopy. Compared with that of the PtFe alloy nanoparticles, the ferrous component appeared in 57 Fe M€ ossbauer spectra of the alumina-supported Fe, PtFe, PtFe2, and PtFe3 alloy nanoparticles, which implies that a strong chemical interaction exists between the nanoparticles and the g-Al2O3 support. The relative resonance area of the ferrous species decreased with the increase of Pt molar fractions contained in the as-prepared catalysts, indicating that Pt can enhance the reduction of iron by intimately contacting with it. The present study is essentially beneficial to obtaining a completed exploration on the catalytic mechanism of the PROX reaction over the bimetallic Pt–Fe catalyst. The states of Fe in Ir–Fe/SiO2 catalyst have great effects on the catalytic performance in PROX reaction, and the promotional role of Fe is always associated with the presence of H2 in the reaction stream. By employment of quasi in situ M€ ossbauer spectroscopy, a clear picture, which shows the changes of different Fe species with increasing the H2 concentrations from 0 to 40%, is obtained. The Fe3þ in the Ir–Fe/SiO2 catalyst is easily reduced to Fenþ (2 < n < 3), Fe0, Fe2þ, and FeIr alloy with the aid of Ir, and the reduced Fe species are also easily oxidized upon exposure to oxygen. When the reduced catalyst was exposed to reaction gas containing CO, O2, and 0–40% H2, FeIr alloy disappeared as a result of oxidation and formed intimately contacted Fe2þ oxide and Ir particles, which were active for PROX reaction. Meanwhile, Fenþ (2 < n < 3) was transformed into Fe3þ that was inactive for the reaction, and about half amount of Fe0 remained intact probably because it was encapsulated by ferric oxide. The amount of Fe2þ increases with increasing the H2 concentration in the reaction stream, well consistent with the change of CO oxidation rate, suggesting that Fe2þ acts as the active site for O2 activation. H2 promotes the reaction via maintaining a substantial amount of Fe existing as active Fe2þ. Meanwhile, H2 may also participate directly in the reaction by forming –OH groups.

ACKNOWLEDGMENTS Financial supports from Chinese Academy of Sciences for “100 Talents” program, the National Science Foundation of China (No. 20325620, No. 21003119) are greatly acknowledged. Prof. Jianyi Shen (Nanjing University) is thanked for helpful discussions.

REFERENCES

575

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

E.D. Park, D. Lee, H.C. Lee, Catal. Today 2009, 139, 280. A. Sirijaruphan, J.G. Goodwin Jr., R.W. Rice, J. Catal. 2004, 224, 304. M. Kotobuki, A. Watanabe, H. Uchida, H. Yamashita, M. Watanabe, J. Catal. 2005, 236, 262. X.S. Liu, O. Korotkikh, R. Farrauto, Appl. Catal A: Gen. 2002, 226, 293. A. Sirijaruphan, J.G. Goodwin Jr., R.W. Rice, J. Catal. 2004, 221, 288. uttner, S. Hackenberg, H.A. Gasteiger, R. Behm, Phys. Chem. Chem. Phys. 2001, 3, M.M. Schubert, M.J. Kahlich, G. Feldmeyer, M. H€ 1123. W.S. Zhang, A.Q. Wang, L. Li, X.D. Wang, T. Zhang, Catal. Today 2008, 131, 457. W.S. Zhang, A.Q. Wang, L. Li, X.D. Wang, T. Zhang, Catal. Letters 2008, 121, 319. Q. Fu, W.X. Li, Y.X. Yao, H.Y. Liu, H.Y. Su, D. Ma, X.K. Gu, L.M. Chen, Z. Wang, H. Zhang, B. Wang, X.H. Bao, Science 2010, 328, 28. F.J. Berry, M€ossbauer Spectroscopy Applied to Inorganic Chemistry, Plenum Press, New York, Vol. 1, 1984, p. 391. G. Gupta, M.N. Patel, D. Ferrer, A.T. Heitsch, B.A. Korgel, M. Jose-Yacaman, K.P. Johnston, Chem. Mater. 2008, 20, 5005. J. Yin, J.H. Wang, T. Zhang, X.D. Wang, Catal. Lett. 2008, 125, 76. W.S. Zhang, Y.Q. Huang, J.H. Wang, K. Liu, X.D. Wang, A.Q. Wang, T. Zhang, Int. J. Hydrogen Energy 2010, 35, 3065. K. Liu, A.Q. Wang, W.S. Zhang, J.H. Wang, Y.Q. Huang, T. Zhang, J.Y. Shen, J. Phys. Chem. C 2010, 114, 8533. M.A. Vannice, R.L. Garten, J. Mol. Catal. 1975/1976, 1, 201. M. Kotobuki, T. Shido, M. Tada, H. Uchida, H. Yamashita, Y. Iwasawa, M. Watanabe, Catal. Lett. 2005, 103, 263. J.W. Niemantsverdriet, J.A.C. van Kaam, C.F.J. Flipse, A.M. van der Kraan, J. Catal. 1985, 96, 58. A. Fukuoka, T. Kimura, N. Kosugi, H. Kuroda, Y. Minai, Y. Sakai, T. Tominaga, M. Ichikawa, J. Catal. 1990, 126, 434. L.W. Lin, W.S. Yang, J.F. Jia, Z.S. Xu, T. Zhang, Y.N. Fan, Y. Kou, J.Y. Shen, Sci. China B 1999, 42, 571. L. Stievano, Hyperfine Interact. 2000, 126, 101. J. Wang, K. Ozawa, M. Takahashi, M. Takeda, T. Nonami, Chem. Mater. 2006, 18, 2261. J.M.M. Millet, Adv. Synth. Catal. 2007, 51, 309. V.R. Balse, W.M.H. Sachtler, J.A. Dumesic, Catal. Lett. 1988, 1, 275. Y. Hou, Z. Xu, S. Sun, Angew. Chem., Int. Ed. 2007, 46, 6329. E. Quinet, F. Morfin, F. Diehl, P. Avenier, V. Caps, J.-L. Rousset, Appl. Catal. B 2008, 80, 195. L. Piccolo, H. Daly, A. Valcarcel, F.C. Meunier, Appl. Catal. B 2009, 86, 190. A. Sirijaruphan Jr., J.G. Goodwin, R.W. Rice, D.G. Wei, K.R. Butcher, G.W. Roberts, J.J. Spivey, Appl. Catal. A 2005, 281, 11. D.H. Kim, M.S. Lim, Appl. Catal. A 2002, 224, 27. L. Li, X. Wang, J. Shen, L. Zhou, T. Zhang, J. Therm. Anal. Calorim. 2005, 82, 103. L.M.P. van Gruijthuijsen, G.J. Howsmon, W.N. Delgass, D.C. Koningsberger, R.A. van Santen, J.W. Niemantsverdriet, J. Catal. 1997, 170, 331. J.W. Niemantsverdriet, A.M. van der Kraan, Surf. Interface Anal. 1986, 9, 221. F.J. Berry, S. Jobson, Hyperfine Interact. 1988, 41, 745. C.M. Wang, D.R. Baer, L.E. Thomas, J.E. Amonette, M.H. Engelhard, J.J. Antony, Y. Qiang, G. Duscher, J. Appl. Phys. 2005, 98, 094308. U. Seip, M.-C. Tsai, K. Christmann, J. K€uppers, G. Ertl, Surf. Sci. 1984, 139, 29. A. Martınez-Arias, M. Fernandez-Garıca, J. Soria, J.C. Conesa, J. Catal. 1999, 182, 367. J.L. Margitfalvi, I. Borbath, K. Lazar, E. Tfirst, A. Szegedi, M. Hegedás, S. GÅb€ ol€ os, J. Catal. 2001, 203, 94. C. Rossignol, S. Arrii, F. Morfin, L. Piccolo, V. Caps, J.-L. Rousset, J. Catal. 2005, 230, 476. E. Quinet, L. Piccolo, J. Daly, F.C. Meunier, F. Morfin, A. Valcarcel, F. Diehl, P. Avenier, V. Caps, J.-L. Rousset, Catal. Today 2008, 138, 43. O. Pozdnyakove-Tellinger, D. Teschner, J. Kr€ ohnert, F.C. Jentoft, A. Knop-Gericke, R. Schl€ ogl, A. Wootsch, J. Phys. Chem. C 2007, 111, 5426. S.H. Cho, J.S. Park, S.H. Choi, S.H. Kim, J. Power Sources 2006, 156, 260. J. Bergeld, B. Kasemo, D.V. Chakarov, Surf. Sci. 2001, 495, L815.

C H A P T E R 3 0

€ THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT? FRANS B. WAANDERS School of Chemical and Minerals Engineering, North West University, Potchefstroom, South Africa

30.1 INTRODUCTION Coal is fundamentally composed of the fossilized remains of plant debris that have undergone progressive physical and chemical alteration through geological time. Coal is highly complex and variable in nature with the inorganic matter dispersed in a random way in the form of mineral inclusions, dissolved salts, and inorganic matter often chemically associated with the organic structure. In the literature, general agreement exists that clays, sulfides, carbonates, and quartz are the most common minerals in coal [1], but the abundance in percentage of mineral matter is very variable from one coal source to another. During mining, transportation, and beneficiation, the coal breaks down into particles of variable nature—as pure coal containing macerals with varying composition, as particles containing mainly inorganic matter, and as a mixture of the above. The inorganic constituents are mainly minerals and according to Clark et al. [2] iron in coal is associated with sulfur in the minerals pyrite (FeS2), jarosite (KFe3(SO4)2(OH)6), a weathering product of pyrite, troilite, and pyrrhotite (both FeS), and non-sulfur-containing minerals such as ankerite (CaFe(CO3)2), illite (clay mineral), and siderite (FeCO3). The amounts of these minerals vary considerably in coals from diverse origin and in the South African case the most abundant mineral groups found in coal are clays (illite), carbonates (dolomite, calcite), sulfides (mainly pyrite), and quartz [3,4]. It is furthermore important to note that coal from the Southern Hemisphere differs from the coal found in the Northern Hemisphere. The speciation of iron and sulfur in coals has been extensively studied by M€ ossbauer spectroscopy and is well established. Due to the presence of iron in a large percentage of mineral matter appearing in coal, M€ ossbauer spectroscopy is a useful, and to a certain degree unique, analytical tool in the identification of iron-bearing minerals [5,6]. In this chapter, the aim is to identify and quantify the iron mineral phases present in South African coal fractions by the use of M€ ossbauer spectroscopy, in conjunction with various other analytical techniques. Because the atomic weight of the carbon content in coal is low, M€ ossbauer spectroscopy is a convenient, and to a certain degree unique, analytical tool in the identification of iron-bearing minerals in coal with iron contents as low as 1%. With an understanding of the iron mineral phases present in the as-mined coal, the fate of these minerals during transportation, weathering, oxidation, and combustion or gasification can be better understood.

M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

576

577

30.2 EXPERIMENTAL PROCEDURES

Detector

Data handling

Absorber

Source

Electronics

Constant velocity drive

FIGURE 30.1 Schematic representation of the experimental setup used.

30.2 EXPERIMENTAL PROCEDURES The apparatus and procedures used are described in the following sections. € ssbauer Spectroscopy 30.2.1 Mo A Halder M€ ossbauer spectrometer capable of operating in conventional constant acceleration mode, using a proporossbauer analyses. The tional counter, filled to 2 atm with Xe gas, and a 57 Co g-ray source were used for the M€ spectrometer was calibrated using a-Fe as the reference and a 50 mCi 57 Co source, plated into an Rh foil, was used to produce the g-rays. The samples were analyzed at room temperature and data collected in a multichannel analyzer (MCA) to obtain a spectrum of count rate against source velocity. A least-squares fitting program was used and by superimposing Lorentzian line shapes, the isomer shift, quadrupole splitting, and/or hyperfine magnetic field of each constituent were determined with reference to the centroid of the spectrum of a standard a-Fe foil at room temperature. The amount of each constituent present in a sample was determined from the areas under the relevant peaks. The typical experimental setup is shown in Fig. 30.1. Conversion electron M€ ossbauer spectra (CEMS) of the samples were measured at 295 K with the aid of the same Halder M€ ossbauer spectrometer capable of operating in conventional constant acceleration mode using a backscattertype gas flow detector. 30.2.2 SEM Analyses A FEI Quanta 200 ESEM scanning electron microscope, integrated with an Oxford INCA 400 energy-dispersive X-ray spectrometer, was used for the SEM analyses. 30.2.3 XRD Analyses For X-ray diffraction (XRD) analyses, the starting materials were characterized by conventional powder X-ray diffraction at room temperature and subsequent Rietveld analysis. The high-temperature XRD (HT-XRD) patterns were recorded in reflection mode in the temperature range 300–1500  C using a PANalytical X’Pert MPD diffractometer, equipped with a PANalytical X’Celerator detector. 30.2.4 Samples and Sample Preparation Representative samples from different coal mines in the South African coal fields were obtained. Samples from a power station and the dry bottom fixed-bed gasifiers used for the production of liquid fuels were also obtained. All the samples were thoroughly homogenized after measuring the bulk density, riffled, and split into equal quarters and sample sizes sufficiently large enough to handle in the laboratory were acquired. The samples were ground in an alumina ball mill to a particle size of less than 600 mm.

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

578

TABLE 30.1 Proximate Analyses of the Original Coal Samples Sample 1 2 3 4 5 6

Calorific Value (MJ kg 1)

Moisture (%)

Ash (%)

Volatile Matter (%)

Fixed Carbon (%)

Total Sulfur (%)

27.09 18.60 15.56 24.72 29.06 32.08

2.8 6.2 5.9 4.1 3.2 2.8

15.9 31.5 40.2 18.2 10.5 7.7

23.9 21.9 22.3 26.7 32.4 5.3

57.4 40.4 31.6 51.0 53.9 84.2

0.65 0.93 1.47 0.73 0.65 0.72

The proximate analysis of these samples was performed by using standard methods to measure the percentage of moisture (SABS 924, ISO 589), ash content (ISO 1171), and volatile matter (ISO 562). The difference between these three percentages and the mass of the original sample (100%) is referred to as the fixed carbon. The ash composition determination, in terms of its oxides, was conducted according to the ASTM D3682 standard. The proximate analyses of the original coal samples are shown in Table 30.1. It should be noted that these measurements are for the single samples investigated and not for bulk samples, as reported in Ref. 7.

30.3 RESULTS AND DISCUSSION In this section, the M€ ossbauer analyses of the as-mined coal samples and the weathering and combustion or gasification products will be discussed. € ssbauer Analyses of the As-Mined Samples 30.3.1 Mo Each coal sample represents the coal being produced from that particular mine and included samples from (1) the Grootegeluk mine in the Northern Province/Waterberg field, (2) Sigma and (3) New Vaal mines in the Free State/ Vereeniging field, (4) Delmas and (5) Optimum mines in the Mpumalanga/Witbank field, and (6) the Zululand mine in the Kwazulu-Natal/Ulundi field. In Fig. 30.2, the distribution of the different coal fields is shown and it should be noted that the coals from these various fields vary considerably, as can be seen from the proximate analyses shown in Table 30.1 and the M€ ossbauer spectra (e.g., Figs. 30.3, 30.4, 30.10, and 30.20a) and summarized M€ ossbauer parameters (Table 30.2). Typical M€ ossbauer spectra of coal samples are shown in Fig. 30.3a and b and the spectra of all the as-received samples consisted of one or two doublets, each characterized by the isomer shift (d), the quadrupole splitting (D), and a relative spectral area (%). From the M€ ossbauer spectra, it can be concluded that the iron-bearing minerals were pyrite, ankerite, illite, and jarosite. In the SEM analyses, the non-iron-bearing minerals quartz, clay, and carbonates were observed in addition to the iron-bearing minerals. The M€ ossbauer parameters obtained for the iron-bearing minerals are shown in Table 30.2 and are in agreement with those published in the literature [8]. It is clear from the data that distinctions between the different coal fields can be made. Sample 4 was taken from the stockyard pile of the Delmas mine in the Mpumalanga/Witbank coal field and has undergone some weathering, thus the appearance of the ferrous sulfate, a phenomenon that will be discussed in Section 30.3.2. It should also be noted that these results are obtained from just a few samples and already a pattern emerges and detailed analyses of an ongoing project will make an enormous amount of information available to assist in a better understanding of the iron minerals and their distribution in the coal and different coal fields. 30.3.2 Weathering of Coal Much of the coal weathering research is empirical and provides little understanding of the underlying chemistry involved and rarely provides adequate explanations for the behavior of coal during combustion, beneficiation, and processing. Generally, oxidation has a deleterious effect on coal behavior and the process most negatively affected by oxidation is carbonization. Utilization of weathered bituminous coal for coke making results in decrease in coke stability and coking

579

30.3 RESULTS AND DISCUSSION

FIGURE 30.2 A map of the different coal fields of South Africa with (a) representing the Waterberg field, (b) the Free State/Vereeniging field, (c) the Mpumalanga/Witbank field, and (d) the Kwazulu-Natal/ Ulundi field. (Adapted from Ref. 7.)

rate and increase in coke reactivity and fines generation. Coal oxidation adversely affects the yield and quality of liquids obtained in direct liquefaction and reduces the efficiency of coal cleaning processes, notably froth flotation. The effects of coal oxidation on combustion, except for reducing the calorific value somewhat, are of lesser importance [9]. The parameters affecting the natural oxidation of coal are far from fully understood. Because of the heterogeneous nature of coal, factors other than the coal’s organic matter and oxygen can affect the oxidation reaction, which may include the thermal chemistry, minerals, moisture, particle size, and surface area present for oxidation or weathering.

FIGURE 30.3 € ssbauer spectrum of the as-received sample containing pyrite and jarosite where the nonisometric (a) A typical Mo form of the doublet indicates that jarosite was present in the sample and (b) a spectrum that contained pyrite and € ssbauer parameters. illite with clearly different Mo

580

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

€ ssbauer Parameters of Various Coal Samples TABLE 30.2 Room-Temperature Mo Sample 1 2 3 4 5 6 Weathering sample Weathering sample Landau coal Landau coal and steel

Lethabo power plant fly ash

Laboratory-produced ash

AFT of coal Hematite Coal ash

Gasifier top 1/3 from top 2/3 from top

Bottom of gasifier

Mineral Pyrite Ankerite Pyrite Illite Pyrite Illite Pyrite Ferrous sulfate Pyrite Jarosite Pyrite Illite Pyrite Pyrite Ferrous sulfate Pyrite Steel substrate FeS Oxyhydroxide Fe2þ Fe3þ Hematite Fe2þ Fe3þ Hematite Pyrite Fe Fe2O3 Fe2þ Fe3þ Hematite Pyrite Pyrite Fe2þ Pyrite Fe2þ Fe3þ Fe2þ Fe3þ Hematite

d (0.01 mm s 1)

D (0.01 mm s 1)

0.33 1.17 0.30 1.04 0.30 1.02 0.30 1.61 0.32 0.20 0.31 0.82 0.31 0.31 1.63 0.32 0.02 0.98 0.34 0.93 0.38 0.34 0.89 0.41 0.35 0.28 0.02 0.37 0.85 0.32 0.36 0.23 0.23 1.05 0.23 1.07 0.70 1.06 0.68 0.33

0.58 1.59 0.62 2.52 0.61 2.40 0.59 2.87 0.63 0.96 0.62 2.40 0.59 0.60 2.88 0.59 0.01 0.18 0.58 2.11 1.01 0.08 2.12 1.02 0.07 0.63 0.00 0.20 1.88 0.57 0.16 0.58 0.59 2.39 0.58 2.29 1.10 2.22 1.10 0.08

H (T)

Relative Area (%)

– – – – – – – – – – – – – – – – 33.0 25.4 – – – 49.3 – – 49.5

58 42 75 25 77 23 76 24 90 10 78 22 100 Varying exposures

32.8 51.5 – – 51.5

49.7

100 54 10 36 27 38 35 39 38 23 80 20 100 20 8 72 100 52 48 37 52 11 56 26 18

d ¼ isomer shift relative to a-iron; D ¼ quadrupole splitting; H ¼ magnetic hyperfine field.

This initial study involves a study of the role of pyrite in coal weathering when finely divided coal is in contact with water for relatively short periods of time. Coal samples weighing 100 g were placed in 1 l glass beakers containing 500 ml distilled water at room temperature. M€ ossbauer spectra were recorded of the dry material (Fig. 30.4, Table 30.2, “weathering sample” containing only pyrite) and of the wet material after exposure times ranging from 2 to 8 days. The recorded M€ ossbauer spectra are shown in Figs. 30.4–30.8 and the parameters are shown in Table 30.2. The fine material was also exposed indoors to the atmosphere for about 8 months resulting in a large drop in the pyrite content of the coal (Fig. 30.9) and from the spectra it was possible to determine the quantity of unreacted pyrite. An important by-product of pyrite oxidation is sulfuric acid, as indicated by the following reaction: 2FeS2 þ 7O2 þ 2H2 O ! 2Fe2þ þ 4SO4 2 þ 4Hþ

(30.1)

581

30.3 RESULTS AND DISCUSSION

FIGURE 30.4 € ssbauer spectrum of as-mined Mo coal.

FIGURE 30.5 € ssbauer spectrum after Mo exposure time of 2 days.

FIGURE 30.6 € ssbauer spectrum after Mo exposure time of 4 days.

582

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

FIGURE 30.7 € ssbauer spectrum after Mo exposure time of 6 days.

FIGURE 30.8 € ssbauer spectrum after Mo exposure time of 8 days (pyrite doublet just visible).

1.1

FIGURE 30.9 Conversion of pyrite to ferrous sulfate (the pyrite content normalized to a value after 8-month exposure to the atmosphere).

Normalized pyrite content

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

0

1

2

3

4

5

6

7

Time exposed to distilled water (days)

8

9

583

30.3 RESULTS AND DISCUSSION

It can be expected that most major minerals in coal will undergo varying degrees of alteration as a result of exposure to acidic conditions. The role of sulfuric acid generated during pyrite oxidation may alter otherwise stable minerals into soluble minerals. Furthermore, the kinetics of the different processes will be strongly influenced by the “time of wetness,” the particle size of the coal, and the availability of oxygen. The present investigation relates to coal slurry in stagnant water, for which it was found that the pH dropped from the initial 7.0 to 5.5 after 8 days of contact with coal. The strong presence of ferrous sulfate is characteristic of stagnant conditions and will, due to the oxidation of the ferrous ion to ferric, affect the availability of oxygen for the pyrite oxidation reaction. 30.3.3 Corrosion of Mild Steel due to the Presence of Compacted Fine Coal During the transport of water-containing fine particle coal, significant corrosion can occur to steel structures, for example, railroad cars and ship cargo holds. The corrosion, occurring at the interface between the steel and the coal, as well as the compaction resulting from transport vibrations, also leads to sticking causing major problems during offloading. Equipment is frequently damaged when the cargoes have to be mechanically dislodged. Gardiner and Melchers [10,11] investigated the variation of the corrosion rate of mild steel as a function of the quantity of water in two types of fine black coal and the experimental evidence from this study suggests that the corrosion rate is maximized at a range of moisture common to all samples, that is, 60–80% of the maximum water holding capacity. Therefore, to assess the corrosivity of fine particle coal, it is beneficial to express the moisture as a percentage of its maximum water holding capacity as opposed to the traditional scale of volumetric or mass percentage. The influence of chloride and sulfate concentration, and to a lesser extent particle size, is dependent on the quantity of moisture. The laboratory experiments yielded similar corrosion rates to those measured within a bulk carrier (ship) cargo hold. Based on this, and the porous texture and dustiness of coal and iron ore cargo, it is proposed that the corrosion rate of mild steel in contact with coal or iron ore cargo is dominated by the fine particle material that is deposited from the larger particles. The coal used in this investigation originated from the Landau mine in the Mpumalanga/Witbank area and was specially selected due to the different sulfur content (0.6 and 3.0%, respectively) observed in the selected coal samples, the sulfur being present mainly in the mineral pyrite (see Fig. 30.10). The M€ ossbauer parameters are shown in Table 30.2. Fine coal samples were pressed into briquettes to simulate the compacting occurring during transportation and it was found that the amount of water contained in the coal declined by a factor of about 2 as compaction pressures increased from an initial pressure of 5 t to a final pressure of 25 t. The amount of water in the samples compacted at 5 t was 8.3% for the coal containing 3% pyrite and 11.9% for the coal containing only 0.6% pyrite. The briquettes pressed at a 5 t compaction pressure were then placed on cleaned steel coupons for predetermined times and CEMS mode M€ ossbauer spectroscopy was conducted on the steel samples with the resulting spectrum presented in Fig. 30.11. Corrosion of the steel was typically in the form of pitting and it was found that the sulfur content of the coal was an important factor in determining the corrosivity of the coal. M€ ossbauer spectroscopy of the corroded surfaces of the mild

FIGURE 30.10 € ssbauer spectrum of the coal Mo from the Landau mine containing 3% pyrite.

584

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

FIGURE 30.11 € ssbauer spectrum of the CEMS Mo corroded mild steel surface.

steel showed apart from iron hydroxides the presence of FeS, formed as a result of the reaction of pyrite (FeS2) with the steel surface, possibly according to the following reaction: FeS2 þ Fe ! 2FeS

(30.2)

Other corrosion products belong of the family of oxyhydroxides and have not been given attention so far. Although the FeS component is partly masked by the peaks from the steel substrate, a best fit was obtained when this component was added to the least fit program. 30.3.4 Coal Combustion During coal combustion, for example, in a power plant, the minerals associated with the coal are heated to melting temperatures and the resultant ash and fly ash can be investigated by means of M€ ossbauer spectroscopy, among the other spectroscopic analysis techniques. As a general rule, the decomposition of the abundant mineral species is depicted in Table 30.3. Fly ash samples and coarse agglomerated ash samples were obtained from the Lethabo power plant, which receives coal from the New Vaal mine (coal sample 3, Table 30.1). These samples and ash obtained after laboratory combustion at various temperatures of the coal fed to the power plant were ground to 180 mm and used as absorbers in the M€ ossbauer experiments. The combustion of the coal samples was done in a muffle furnace, under atmospheric

TABLE 30.3 Summary of Mineral Changes that Occur at Various Temperatures During Combustion Temperature

Effect Observed

50–110  C 130–180  C 200–400  C 300–600  C 340–530  C

Loss of absorbed water Dehydration of gypsum (endothermic) Dehydroxylation of Fe Dehydroxylation of Al Pyrite and siderite decomposition; clay mineral dehydroxylation and destruction Dehydroxylation of Ca and Mg Decomposition of Fe sulfates Quartz inversion (endothermic) Calcite and dolomite decomposition Continuation of clay mineral destruction and carbonate decomposition Continuation of clay mineral destruction; formation of spinel, mullite, corundum, and amorphous phases Solid-state reactions (mainly between CaO and silicates) Crystallization of amorphous silica to cristobalite, formation of corundum, spinels, mullite, Ca silicates, pyroxenes, and olivines, anhydrite decomposition, some reactions between phases, melting or solutions of different phases

350–620  C 470–810  C 570–575  C 680–915  C >800  C 850–1000  C 965–1000  C >1000  C

585

30.3 RESULTS AND DISCUSSION

FIGURE 30.12 € ssbauer spectrum of the fly ash from the Lethabo power plant and (b) the spectrum of coal sample 3 that (a) The Mo was heated to 1200  C.

conditions, in the temperature range 200–1000  C, in steps of 200  C. The mineral phases occurring in the coal and ash samples were also identified by means of SEM analyses. The M€ ossbauer spectrum of the fly ash sample obtained from the Lethabo power plant consisted of a Fe2þ and Fe3þ component, typical of silica glass. In addition to the glass formation, hematite (Fe2O3) was also observed (see Fig. 30.12). The same spectrum was observed for the agglomerated coarse ash sample, but with a slightly different Fe2þ/Fe3þ ratio. In the laboratory simulations, carried out on sample 3 from the New Vaal mine, it was clearly observed by M€ ossbauer spectroscopy how the amount of pyrite diminished as the combustion temperatures increased and how it altered until finally to the point where the iron was taken up in the glass and the hematite. At 200  C, the pyrite started to change, at 400  C already 60% had been altered to hematite, and at 600  C only 20% of the pyrite remained. At 800  C, most of the pyrite had been transformed and above a temperature of 1000  C the pyrite had fused completely into the glass and the oxide and thus similar products were formed in the fly ash from the power plant and the laboratory-produced ash as shown in Fig. 30.12, with M€ ossbauer parameters, given in Table 30.2, consistent with those found in the literature [8]. Decomposition of pyrite typically follows the path FeS2 ! Fe1:25 S ! Fe þ S2 þ O2 ! Fe2 O3

(30.3)

Pyrrhotite (Fe1.25S), an intermediate phase, may be present after decomposition of the pyrite and the final Fe species present is hematite, in addition to a glass phase containing both Fe2þ and Fe3þ. With the aid of the SEM analyses, the different components were further identified and the results are shown in Table 30.4. It is, however, important to note that both samples have relatively high iron, calcium, and potassium contents—components critical for low-temperature fusion of particles and a factor to be discussed in the next section. In the SEM element mapping of the aluminum, silicon, and calcium on a fly ash particle fusion point, calcium was found to be in a high concentration (see Fig. 30.13). This indicates that the high Ca content of the coal used results in fusion of small fly ash particles, leading to agglomeration and resulting in industrial problems encountered.

TABLE 30.4 Element Analysis (wt%) of the Industrial and Laboratory-Produced Coal Ash Element (Expressed as Oxide) SiO2 Al2O3 CaO K2O Fe2O3 TiO2 MgO Others (SO3, Na2O)

Industrial Fly Ash (wt%)

Laboratory-Produced Ash (wt%)

42.35 33.79 6.68 0.76 5.84 1.47 1.51 2.60

46.89 33.31 6.20 1.07 5.39 1.86 1.76 3.52

586

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

FIGURE 30.13 (a) A photomicrograph of a fusion point between two fly ash particles and the element maps for Al (b), Si (c), and Ca (d), indicating the higher Ca concentration on a fusion point.

The M€ ossbauer and SEM results obtained for the coal being combusted in the laboratory in the temperature range 200–1200  C show the transformation of iron minerals in the coal to ash via the following steps, in accordance with the observation made by Ram et al. [12]: FeS2 ðpyriteÞ ! Fe2 O3 ðhematiteÞ FeSO4
H2 Oðjarosite or iron sulfatesÞ ! Fe2 O3 ðhematiteÞ Fe2þ in clay ! Fe3þ phaseðsilica glassÞ The changes that occurred in the mineral matter during combustion of the coal, heated up to a temperature of 1500  C, were observed with HT-XRD techniques and the results are shown in Fig. 30.14. The pyrite started to decompose at a temperature of about 500  C to form hematite, consistent with the previously mentioned observations. It is also interesting to note that the a- to b-quartz phase change can be seen at a temperature of about 600  C and the transformation from b-quartz to tridymite at about 800  C, consistent with the literature [13]. After the HT-XRD experiments were conducted, a M€ ossbauer spectrum for each sample was obtained. The resulting M€ ossbauer spectrum yielded Fe-containing glass and hematite as shown in Fig. 30.12. The possible occurrence of the intermediate pyrrhotite phase, due to the decomposition of pyrite, was not observed. It is important to note that coal sample 3 (Table 30.1) had relatively high iron and calcium contents—components critical for low-temperature fusion of particles. The ash fusion temperature (AFT) of a coal source is mainly influenced by silica, iron, and sulfur with iron the controlling component in fouling and slagging of coal sources during combustion [12,14]. The standard ISO 540 method was utilized to demonstrate the fusion properties of the laboratory-prepared coal ash where the AFT was derived from the flow temperature (FT) value, which has been used in this study as a measure of the expected behavior of the ash bed during gasification. Mayoral et al. [14] stated that pyrite plays a major role in slagging due to formation of low-melting FeS–FeO intermediates together with the fluxing effect of aluminosilicates. Coal sample 3 was thus selected and various amounts of iron were added to determine the influence on the AFT of the coal. Under

587

30.3 RESULTS AND DISCUSSION

Muscovite Kaolinite Quartz

650

Pyrite Anatase

550

Hematite Mullite Cristobalite

Peak area (a.u.)

450 350 250 150 50 300 –50

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

Temperature (°C)

FIGURE 30.14 Relative intensities of the identified mineral phases in the sample as a function of temperature.

both oxidizing and reducing conditions, the AFT dropped drastically with an increasing amount of iron added, as can be seen from Fig. 30.15 [15]. In the M€ ossbauer spectra of the initial coal sample, mainly pyrite, as iron-containing mineral, was present, but with the addition of the iron, the iron sextet appeared and after combusting the sample at the temperature where the flow started to occur, the Fe2þ and Fe3þ glass phase and hematite were observed in the spectra (see Fig. 30.16). The M€ ossbauer parameters are shown in Table 30.2 and are consistent with the literature [8]. 30.3.5 Coal Gasification and Resultant Products The mineral matter associated with coal undergoes various transformations during the coal gasification process, with the temperature of transformation of pyrite depending on either the reducing or oxidizing conditions present in the gasifier. The temperatures may vary due to possible trace elements present in the pyrite structure [16,17].

FIGURE 30.15 AFT results of the centrifuge overflow sample and those where different amounts of iron were added under oxidizing conditions clearly showing the lowering of the AFT with increasing amounts of iron present.

588

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

FIGURE 30.16 € ssbauer spectrum of the coal sample with the addition of 2% Fe and (b) a spectrum of the sample after being (a) A Mo heated up to the ash fusion or flow temperature.

Optimization of the gasification process is always a necessity and, for example, agglomeration due to heat-induced transformation of the mineral matter inside the gasifier can cause various undesirable operational conditions, implying a loss in gas production [18–20]. The principal aim of this investigation was to determine the changes that the Fe-containing minerals and mineral associations undergo during gasification of coal. Representative samples were extracted from the gasifier and it was done at predetermined positions, after sufficient cooling was done and each of the samples was thoroughly homogenized, riffled, and split into equal quarters and sample sizes sufficiently large enough to handle in the laboratory. From Fig. 30.17, it is clear that the gasifier can be divided into various reactions zones. At the top of the gasifier where the coal enters the gasifier is the drying zone, which is followed by the reduction zone, then the oxidation zone, and at the bottom the ash is extracted, with the maximum temperature occurring almost two-thirds down the gasifier. In order to determine the temperature profile, vitrinite reflectance techniques were used, based on the ISO 7404-5 standard method, and the vitrinite reflectance analyses were performed on the whole range of the gasifier turn-out samples.

FIGURE 30.17 Graphical representation of a gasifier with the resultant temperature profile observed in the gasifier [19,20].

30.3 RESULTS AND DISCUSSION

589

FIGURE 30.18 € ssbauer spectra of (a) coal sample entering the gasifier, (b) at sample position 1/3 down the gasifier, (c) at The Mo sample position 2/3 down the gasifier, and finally (d) the sample retrieved at the ash grate of the gasifier [17].

In the coal feed samples, pyrite was the only Fe-containing mineral detected, while the pyrite changed gradually to form, in conjunction with the other minerals present in the coal, a Fe-containing glass and hematite at the bottom, or ash grate after combustion. From the M€ ossbauer spectra (Fig. 30.18), it can be seen that the first appearance of a Fe2þ phase occurred about one-third down the gasifier, which correlates to an average temperature of about 400  C and a surface temperature of about 600  C, consistent with literature values [21] for the temperature where the pyrite in the coal starts to decompose. The surface temperature, as measured for the coal particles via vitrinite reflectance, gave a good indication of the temperature the minerals experienced. Since the decomposition to pyrrhotite is relative fast [22], this phase was not observed in the current measurement. As the pyrite and the coal further descended in the process, it was subjected to higher temperatures and the last occurrence of pyrite was about two-thirds down the gasifier, where an average temperature of over 800  C and a surface temperature of about 1000  C were measured. This indicates that the pyrite mainly decomposed in the pyrolysis zone of the gasifier. In the oxidation zone, at the bottom part of the gasifier oxidation of the iron starts to play a role and hematite was formed in the same temperature environment as was found in the pyrolysis zone (i.e., an average temperature of over 800  C and a surface temperature of 1000  C). In the ash bed, the highest temperatures occurred with peak temperatures well above 1400  C. At these high temperatures, the SiO2 and Al2O3 present in the coal agglomerate and form the final glass observed. Thus, the pyrite changed gradually to form, in conjunction with the SiO2 and Al2O3 present in the coal, Fe-containing glass and hematite at the bottom, or ash grate of the gasifier. FactSage modeling was also conducted to try and solve the problem regarding the fact that the formation of pyrrhotite was not observed in any of the experiments. For this purpose, a phase diagram of the changes that can occur when pyrite is subjected to increasing temperatures was constructed using the FactSage modeling technique and the results are shown in Fig. 30.19. From this diagram, it is clear that the partial oxygen pressure plays an important role in the products that can form under these conditions. To form, for example, hematite from the pyrite at a temperature of about 420  C, the oxygen partial pressure must be approximately log10(pO2)  22 bar. FactSage modeling of possible pyrite changes under combustion conditions with a constant oxygen pressure shows that pyrite starts to decompose at a temperature of around 300  C to pyrrhotite, which stays stable up to about 1200  C, after which a Fe slag starts to form (see Fig. 30.20a). Under more oxidizing conditions, hematite forms at a temperature of about 600  C (see Fig. 30.20b), which is a bit higher than expected.

590

€ 30 THE USE OF MOSSBAUER SPECTROSCOPY IN COAL RESEARCH: IS IT RELEVANT OR NOT?

FeS2 - O2 - Fe mole FeS2/(Fes2+Fe) = 1 2000

1605

T (c)

1210

815

FIGURE 30.19 FactSage phase diagram constructed for the changes occurring in pyrite when heated under changing partial oxidizing conditions.

420

25 –55

–44

–33 –22 log10(p(O2) (bar)

–11

0

FIGURE 30.20 FactSage modeling showing that (a) under constant partial oxygen pressures pyrrhotite forms while (b) under a partial oxygen pressure of log10(pO2)  22 bar, hematite will form.

30.4 CONCLUSIONS It is evident that M€ ossbauer spectroscopy can be a handy tool in differentiating between various coal fields on the basis of the Fe mineral composition of the coal. It should, however, be noted that a detailed analysis of all coal seams needs to be done to determine the run-of-mine coal iron mineral composition. The technique is also suitable for determining the iron components and the presence of sulfur associated with the iron-bearing minerals as part of the impurities found in coal. Of the minerals present in coal, the Fe-bearing minerals observed via M€ ossbauer spectroscopy constitute the major component and are believed to govern, to a great extent, the polluting behavior of coal during combustion. It is clear that the iron mineralogy, as established by M€ ossbauer spectroscopy, can be used not only to monitor the oxidation during coal combustion, but also to reconstruct the probable conditions that prevail. It is, in conjunction with other analytical tools such as XRD and SEM analyses, a powerful diagnostic tool that can be applied in coal research and can definitely be regarded as a relevant research technique. As such, accurate characterization of the ash-forming constituents, especially the Fe-bearing minerals and possible clay minerals present, is desirable not only to predict the slagging/fouling characteristics of that coal, but also for

REFERENCES

591

knowledge of other mineral phases, to provide a positive clue to the transformations of these minerals during combustion and/or gasification.

ACKNOWLEDGMENTS The author would like to express his gratitude to all students who helped making this contribution possible. Without their dedication none of the results discussed would have been possible to publish.

REFERENCES 1. B. Alpern, J. Nahuys, L. Martinez, Mineral matter in ashy and non-washable coals—its influence on chemical properties, Comun. Serv. Geol. Portugal 1984, 70(2), 299–317. 2. T.M. Clarke, B.J. Evans, C. Wynter, H. Pollak, S. Taole, D. Ratcliffe, Investigation of Virgin Coals and Coals Subjected to a Mild Acid Treatment, Hyperfine Interact. 1998, 112, 227–230. 3. R.M.S. Falcon, C.P. Snyman, An Introduction to Coal Petrography: Atlas of Petrographic Constituents in the Bituminous Coals of Southern Africa. The Geological Society of South Africa, Johannesburg, South Africa, 1986,123 pp. 4. C. Van Alphen, AFT modification—CCSEM characterisation of coal feed and gasifier ash, Van Alphen Consultancy, South Africa, October 2004, pp. 1–24. 5. P.A. Montano, Application of M€ ossbauer spectroscopy to coal characterisation and utilisation, in M€ossbauer Spectroscopy and Its Chemical Applications, Advances in Chemistry, Vol. 194, J.G. Stevens, G.K. Shenoy, eds., American Chemical Society, Washington, DC, 1981, pp. 135–175. 6. G.J. Long, J.G. Stevens, eds., Industrial Applications of the M€ossbauer Effect, Plenum Press, New York, 1986, 796 pp. 7. H.J. Pinheiro, ed., Bulletin 113: A techno-economic and historical review of the South African coal industry in the 19th and 20th centuries and analyses of coal product samples of South African collieries, 1998–1999, Department of Minerals and Energy and the SABS Coal and Minerals Services, 1999, 97 pp. 8. J.G. Stevens, A.M. Khasanov, J.W. Miller, H. Pollak, Z. Li, eds., M€ossbauer Mineral Handbook, M€ ossbauer Effect Data Centre, University of North Carolina, Asheville, NC, 1998, 527 pp. 9. F.E. Huggins, G.P. Huffman, G.R. Dunmyre, M.J. Nardozzi, M.C. Lin, in J.A. Moulijn, F. Freek Kapteijn, eds., Proceedings of the First International Symposium on Coal Science, April 28–May 1, 1986, Rolduc, The Netherlands, Elsevier, Amsterdam, 1986. 10. C.P. Gardiner, R.E. Melchers, Corrosion of mild steel by coal and iron ore, Corros. Sci. 2002, 44, 2665–2673. 11. C.P. Gardiner, R.E. Melchers, Corrosion of mild steel in porous media, Corros. Sci. 2002, 44, 2459–2478. 12. L.C. Ram, P.S.M. Tripathi, S.P. Mishra, M€ oessbauer spectroscopic characterization and studies on their transformations of ironbearing minerals during combustion of coals: correlation with fouling and slagging, Fuel Process. Technol. 1995, 42, 47–60. 13. C. Klein, B. Dutrow, Manual of Mineral Science (After J.D. Dana), 23rd edn, Wiley, 2007, 673 pp. 14. M.C. Mayoral, M.T. Izquierdo, J.M. Andres, B. Rubio, Mechanism of interaction of pyrite with hematite as simulation of slagging and fireside tube wastage in coal combustion. Thermochim. Acta 2002, 390, 103–111. 15. C.B. Prinsloo, A study into the fundamental understanding of iron-transformations and the effect of iron as fluxing agent on Highveld fine coal sources during gasification, M.Eng. dissertation, North West University, 2008, 96 pp. 16. F.B. Waanders, E. Vinken, A. Mans, A.M. Mulaba-Bafubiandi, Iron minerals in coal, weathered coal and coal ash—SEM and M€ ossbauer results, Hyperfine Interact. 2003, 148–149, 21–29. 17. F.B. Waanders, J.R. Bunt, Transformation of the Fe-mineral associations in coal during gasification, Hyperfine Interact. 2006, 171, 287–292. 18. A. Govender, Determination and statistical evaluation of the effect of minerals and mineral associations in specific dense medium fractions on ash fusion temperature, M.Sc. dissertation, NWU, Potchefstroom, 2005, 83 pp. 19. J.R. Bunt, A new dissection methodology and investigation into coal property transformational behaviour impacting on a commercial-scale Sasol-Lurgi MKIV fixed-bed gasifier, Ph.D. thesis, North West University, 2006, 165 pp. 20. J.C. Van Dyk, Manipulation of coal feed to gasification in order to increase the AFT of the coal, Ph.D. thesis, NWU, Potchefstroom, 2006, 147 pp. 21. F.B. Waanders, The mineral changes occurring in lignite coal during gasification, Presentation at ISIAME2012, Dalian, China, September 2012. ossbauer and FT-IR spectroscopy during natural burning. 22. D. Bandyopadhyay, Study of kinetics of iron minerals in coal by 57 Fe M€ Hyperfine Interact. 2005, 163(1–4), 167–176.

PA R T V I I

ENVIRONMENTAL APPLICATIONS

C H A P T E R 3 1

WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS SHIRO KUBUKI1 AND TETSUAKI NISHIDA2 1 2

Department of Chemistry, Graduate School of Science and Technology, Tokyo Metropolitan University, Hachioji, Japan Department of Biological and Environmental Chemistry, Faculty of Humanity-Oriented Science and Engineering, Kinki University, Iizuka, Japan

31.1 INTRODUCTION Environmental problems, such as increase in the amount of waste, water contamination, and green house effect, have become serious at a global scale. The fourth Intergovernmental Panel on Climate Change (IPCC) reported that increases in the average temperature and in the sea level were expected to be 2.4–6.4  C and 0.26–0.59 m, respectively, in the worst case of twenty-second century [1]. It is important for us to establish a “sustainable society” by developing processes for recycling of wastes and purification of polluted water. In order to solve these environmental problems, the authors have investigated a relationship between the structure ossbauer spectroscopy. In this chapter, recent and properties of iron-containing “waste” of silicate glasses by using 57 Fe-M€ studies are described on the recycling process of iron-containing “waste” of silicate glasses that are very effective for the purification of polluted water. The studies proved that iron-containing “waste” of silicate glasses had an interesting electromagnetic phenomenon. 31.1.1 Water-Purifying Ability of Recycled Iron Silicate Glass The compliance rate in a closed water system (e.g., lake or pond) in Japan is reported to be 53%, which is much lower than that in an open water system such as river, (92.3%), or ocean and coast (76.4%) [2]. The low compliance rate in closed water systems has not improved in the last two decades, showing a difficulty in maintenance, despite that purification of wastewater in a closed water system is one of the most critical environmental problems. Although “water-purifying minerals” are commercially available, their ability has not been elucidated scientifically. Kubuki and coworkers developed a new water-purifying system in which porous silicate glass obtained from waste bottle glass was utilized. Water quality could be improved to the chemical oxygen demand (COD) and dissolved oxygen (DO) levels of 25 and 8.3 mg l 1, respectively [3,4]. The final disposal amount of domestic wastes in Japan has reduced from 235 to 119 g per person during the last decade. The Ministry of Environment reported that the damp yard would become full in 18 years [2]. Gasification–fusion furnaces introduced in several cities in Japan established a waste combustion system. The disadvantage of this system is M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

595

596

31 WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS

that the running cost is five times higher than that of the earlier facilities. Due to these reasons, wastes should be recycled into new functional materials. It is well known that FeII plays a significant role in the Fenton method that is effective for water purification [5]. Leupin and Hug reported that FeII was effective for reducing hazardous AsIII in the ground water [6]. Iron-containing silicate glass can be a good candidate for effective water purification, because the solubility of iron in glass and the FeII/FeIII ratio can easily be controlled by changing the composition and preparation conditions of the glass sample. Although the local structure of iron-containing silicate glass has been studied by using M€ ossbauer spectroscopy [7–9], the relationship between the structure and the water-purifying ability has not been investigated so far. In this chapter, recycling of combustion ash, discarded from a municipal garbage combustion plant, and application to water purification are described. It is proven that waste glass recycled from combustion ash is a very effective waterossbauer, X-ray diffraction (XRD), COD, and induced coupled plasma (ICP) purifying material. Measurements of 57 Fe-M€ are used for characterizing several waste glasses containing different amounts of Fe2O3. 31.1.2 Iron Silicate Glass Prepared by Recycling Coal Ash Recycling of waste materials is very effective for saving energy and preserving the global environment. Coal ash is generally composed of about 90% fly ash and 10% bottom ash or clinker ash. Fly ash is well known to be composed of round particles with diameters of several micrometers. About 10 million tons of fly ash is exhausted from the electric power station and the like in Japan every year. It is utilized as a raw material for cement or mortar in the fields of architecture, civil engineering, and so on. Samples of fly ash studied in this chapter were provided by Kyushu Electric Power Station, Fukuoka Prefecture, Japan. Evaluation of local structure and electrical or magnetic properties of fly ash-recycled glass ossbauer spectroscopy, X-ray fluorescence (FARG), prepared with fly ash and Fe2O3, was investigated by means of Fe-M€ analysis, and DC four-probe method for electrical conductivity (s). Heat treatment of FARG was carried out in order to estimate the effect of structural relaxation and crystallization on the electrical or magnetic properties [10].

31.2 PROPERTIES AND STRUCTURE OF RECYCLED SILICATE GLASSES 31.2.1 Water-Purifying Ability of Recycled Silicate Glasses Iron-containing soda-lime silicate (ISLS) glass, with a composition of 0.5xNa2O  0.5xCaO  5Fe2O3  (95 x)SiO2, in which x ¼ 20 and 50 (mass%), was prepared by the conventional melt-quenching method. Enriched isotope of 57 Fe2 O3 (57 Fe ¼ 95:54%) was used when preparing the samples for M€ ossbauer measurement. Room-temperature M€ ossbauer spectra of ISLS glass with “x” of 20 and 50 measured before and after the leaching test are shown in Fig. 31.1 [11]. M€ ossbauer spectra are decomposed into two paramagnetic doublets due to FeII and FeIII with d values of 0.91–0.97 (0.03) and 0.27–0.28(0.01) mm s 1, respectively. D values of 1.96–2.18 (0.06) and 0.91–0.96 (0.02) mm s 1 were estimated for FeII and FeIII, respectively. Absorption area of FeII was decreased from 26.7 to 11.6% when Na2O plus CaO content was increased from 20 to 50 mass%. These M€ ossbauer parameters are comparable to those obtained for sodalime silicate glass with a composition of 17Na2O  13CaO  65SiO2  2Fe2O3  3Al2O3, that is, d ¼ 0.97 mm s 1, D ¼ 1.96 mm s 1, and A ¼ 24.7% for FeII (Td), and d ¼ 0.23 mm s 1, D ¼ 1.14 mm s 1, and A ¼ 75.3% for FeIII (Td) [12]. After ossbauer spectra decreased from 26.7 to 17.9% and leaching ISLS glass in water at 30  C, absorption area for FeII in the M€ from 11.6 to 0% when “x” was 20 and 50, respectively. These results indicate that FeII ion in ISLS glass is easily oxidized into FeIII when immersed into water. It is noted that leaching of hazardous heavy metal sludge and diatomaceous earth results in the precipitation of g-FeOOH, as confirmed by the XRD patterns [13]. In the case of ISLS glass, however, no XRD pattern was detected due to g-FeOOH, Fe(OH)3, a-FeOOH, b-FeOOH, and others. ICP analysis showed that the amount of dissolved iron was increased stepwise from 3.6 to 4.7, 12.1, 8.7, and 13.6 mg l 1 after 10 day leaching of ISLS glass with “x” ¼ 20, as shown in Fig. 31.2 [11]. Dissolution of Naþ and Ca2þ, which are weakly bound to the glass skeleton as a network modifier (NWM), also increased from 10.9 to 39.1 mg l 1 and from 9.1 to 36.7 mg l 1, respectively. Dissolution rate of cations (D in mass%) was calculated by: Da ¼ ðmass of dissolved ion“a”Þ=ðmass of ion“a”in ISLS glassÞ  100:

(31.1)

From Eq. (31.1), values of DFe, DNa, and DCa for ISLS glass with “x” ¼ 20 were estimated to be 9.7, 13.2, and 12.8%, respectively. In the case of ISLS glass with “x” ¼ 50, DFe, DNa, and DCa values were estimated to be 24.5, 87.3, and 93.5%,

597

100

Fe3+

99

(a-1) –6 – 4 –2 0 2 4 Velocity (mm s–1)

6

Transmission (%)

Transmission (%)

Fe2+

100 Fe3+

98 97 –6

98

(b-1) –4

–2 0 2 4 Velocity (mm s–1)

6

101 100 Fe3+

99 98 97 96 (b-2)

(a-2) –4

Fe3+

99

97 –6

101

99

Fe2+

100

Fe2+

Transmission (%)

Transmission (%)

31.2 PROPERTIES AND STRUCTURE OF RECYCLED SILICATE GLASSES

–2 0 2 4 Velocity (mm s–1)

6

–6

–4

–2 0 2 4 Velocity (mm s–1)

6

FIGURE 31.1 € ssbauer spectra of ISLS glasses Mo with “x” ¼ (a) 20 and (b) 50. Spectra (1) and (2) were respectively measured before and after the leaching test for 10 days with distilled water. (Reproduced from Ref. 11 with permission of Springer.)

respectively. It is noted that the DFe value became 2.5 times larger when the Na2O þ CaO content increased from 20 to 50 mass%, due to defragmentation or depolymerization of the glass network. Absorption spectroscopy showed that dissolution of FeII from ISLS glass having “x” ¼ 20 was increased from 0.26 to 0.43, 0.68, 0.84, and 0.93 mg l 1 after 10 day leaching at 30  C, while dissolution of FeII was not confirmed when “x” was 50. These results indicate that dissolution rate of FeII in ISLS glasses can be easily controlled by the Na2O þ CaO content, and that ISLS glass having “x” ¼ 20 is the preferable FeII provider to water. In Fig. 31.3, a decrease in ORP value was observed from 30.0 to 197.7, 196.2, 197.0, 191.4, and 196.8 mV after the leaching of the ISLS glass with “x” ¼ 20 [11]. At the same time, pH increased from 6.1 to 10.0, 10.0, 10.1, 9.9, and 10.0 (Fig. 31.3). On the other hand, a much smaller ORP value of 284.7 mV and larger pH of 11.6 were observed when “x” was 50 [11]. It is noteworthy that a decrease in ORP and an increase in pH are more remarkable when “x” is 50. Because the solubility product constant (Ksp) of Fe(OH)2 is reported to be 1  10 15 [14], it seems that the dissolved amount of Fe2þ in ISLS glass with “x” ¼ 50 was suppressed under the high pH value of 11.6 after 10 day leaching. COD value was determined by redox titration method using KMnO4 in accordance with the analysis method of JISK010217. As shown in Fig. 31.4, a decrease in the COD value was observed from 100 to 8.6% after 10 day leaching of ISLS glass with “x” ¼ 20, while COD decreased from 100 to 85.3% when “x” was 50. These results indicate that slight dissolution of cations from ISLS glass was effective for water cleaning, as observed in ISLS glass with “x” ¼ 20. It is

Concentration (mg l –1)

1000

FIGURE 31.2 100

10

1 0

2

4 6 8 Time (days)

10

12

Concentrations of total iron (square), Naþ(circle), and Ca2þ (triangle) in the filtrate after the leaching test with distilled water. Solid and open symbols marks refer to the result obtained for ISLS glasses with “x” ¼ 20 and 50, respectively. (Reproduced from Ref. 11 with permission of Springer.)

31 WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS

12

50

11

0 –50

pH

10

–100 9 –150

FIGURE 31.3 8

Change of pH (triangle) and ORP (circle) measured after the leaching test of ISLS glasses with “x” ¼ 20 (black) and 50 (white). (Reproduced from Ref. 11 with permission of Springer.)

–200

7 6

–250 0

2

4

6 8 Time (days)

Relative concentration (%)

Decrease in relative COD value observed for (a) blank, ISLS glass with “x” ¼ (b) 20 and (c) 50.

–300 12

10

100

FIGURE 31.4

ORP (mV)

598

(c)

80 60

(a)

40 20

(b)

0 0

2

4

6

8

10

12

Time (days)

considered that larger COD values were observed for ISLS glass with “x” ¼ 50 probably because of high pH and low ORP values. Large amounts of Fe, Na, and Ca ions dissolved during the leaching and resultant high pH are not preferable for the microorganisms to propagate effectively. As a result of chemical analysis, the composition of discarded ash from Ube municipal garbage combustion plant (Ube Slag: US) was determined to be SiO2—38.4, CaO—28.5, Al2O3—14.9, Fe2O3—5.54, P2O5—3.34 mass%, and (Na2O ossbauer spectrum of the original US before melting with Fe2O3 is depicted in Fig. 31.5 þ K2O þ MgO) for the rest. M€ [15]. The spectrum can be analyzed into two doublets: one for the FeIIO4 tetrahedra with d ¼ 0.99 (0.01) mm s 1, D ¼ 1.88 (0.02) mm s 1, G ¼ 0.78 (0.03) mm s 1, and A ¼ 71.9 (2.2)%, and the other for FeIIIO4 tetrahedra with

2+

FIGURE 31.5 € ssbauer spectrum of an Mo exhausted ash from waste treatment plant in Ube city [15]. (Yamaguchi Prefecture, Japan) (Reproduced from Ref. 15 with permission of Springer.)

Transmission (%)

100.0

Fe 3+ Fe

99.5

99.0 –10

–5 0 5 Velocity (mm s–1)

10

599

31.2 PROPERTIES AND STRUCTURE OF RECYCLED SILICATE GLASSES

(a)

100

Fe

3+

2+

Fe

100 (b)

Transmission (%)

99 100

(c)

99 100 (d)

99 100

(e) 99

FIGURE 31.6 € ssbauer spectra of US (a) 0, (b) Mo 5, (c) 10, (d) 15, and (e) 20 glasses. (Reproduced from Ref. 15 with permission of Springer.)

98 97 –10

–5

0

5

10

Velocity (mm s–1)

d ¼ 0.33 (0.03) mm s 1, D ¼ 0.97 (0.07) mm s 1, G ¼ 0.78(0.03) mm s 1, and A ¼ 28.1 (1.6)%. XRD patterns of US showed no indication of crystal peaks. M€ ossbauer spectra of silicate glass recycled with different Fe2O3 contents ranging from 0 to 20 mass% (US0-20 glass) are given in Fig. 31.6 [15]. Each spectrum could be analyzed into two doublets due to FeIIO4 and FeIIIO4 tetrahedra. A small increase in d was observed for FeIII from 0.26 (0.01) to 0.30 (0.01) mm s 1 when Fe2O3 content was increased from 0 to 20%, while an identical D value of 1.25 (0.01) mm s 1 was obtained in every sample. An identical d value of 1.00 (0.02) mm s 1 was obtained for FeII. A gradual decrease was observed for A of FeII from 25.0 (1.3) to 5.7 (1.6)%. By comparing Figs. 31.5 and 31.6, we can observe a drastic decrease in A of FeII in the latter, indicating that FeII is oxidized to FeIII when the melt of US is quenched. ossbauer spectra were measured at In order to calculate Debye temperature (QD) of US0 and US20 glasses, M€ temperatures between 77 and 285 K, as illustrated in Fig. 31.7 [15]. A relationship between the recoil-free fraction, f (T), and QD can be expressed by the following equation [16]:  Z f ðTÞ ¼ exp ð 3E R =2kQD Þð1 þ 4ðT=QD Þ

Transmission (%)

3+

Fe

100 99

(a)

100 99

(b)

100 99

(c)

100 99

(d)

100 99

(e)

100 99

(f)

100 99 98 –10

QD

xdx=ðex 0

 1ÞÞ ;

(31.2)

2+

Fe

(g)

–5 0 5 Velocity (mm s–1)

10

FIGURE 31.7 € ssbauer spectra of US20 glass Mo measured at (a) 77, (b) 100, (c) 140, (d) 175, (e) 210, (f) 250, and (g) 285 K. (Reproduced from Ref. 15 with permission of Springer.)

600

31 WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS

300 blank US0 US5 US10 US15 US20

COD (mg l–1)

250 200 150 100

FIGURE 31.8 Decrease in COD of artificial drain by stirring with US glasses in “stage 1”. (Reproduced from Ref. 15 with permission of Springer.)

50 0

0

2

4

6

8

10

12

Time (days)

where ER, T, and k are the recoil energy (0.002 eV), temperature (K), and Boltzmann constant (8.62  10 5 eV), respectively. The QD values were calculated to be 287 (14.1) K for FeIII, 246 (15.1) K for FeII, and 275 (10.5) K for the total iron (FeII plus FeIII) in case of US0 glass, while 445 (17.3) for FeIII, 224 (19.4) for FeII, and 407 (12.2) K for (FeII and FeIII) in US20 glass. These results show that both FeII and FeIII are more rigidly bound to oxygen atoms in US20 glass than in US0 glass [15]. Johnson et al. reported comparable QD values of 312 K for FeIII and 268 K for FeII in (Fe2O3)0.05(Na2O)0.30(SiO2)0.65 glass [17], proving that the QD is affected by Fe2O3/SiO2 ratio in iron-containing silicate glass. Nishida and coworkers reported that QD values of several oxide glasses could be classified into two groups, that is, those higher than 280 K and others less than 270 K [18]; the former values are obtained when FeIII or FeII play a role of network former (NWF), while the latter values when FeIII or FeII play a role of network modifier (NWM). We conclude that iron occupies NWM sites in both US0 and US20 glasses, since QD value was estimated to be 246 K in the former and 224 K in the latter [15]. A decrease in COD was observed in a sample composed of 10 g of US glass, 1 ml of active sewage, and 300 ml of artificial drain, after being stirred for 10 days (stage 1), as shown in Fig. 31.8 [15]. A remarkable decrease in the COD was observed after two day stirring of US20 glass from 250 to 36 mg l 1, while the COD for the blank (USblank) decreased slightly from 250 to 154 mg l 1. These results show that iron-containing US glass promotes the decomposition of organic materials. Concentrations of iron and phosphate ion in the artificial drain determined by the ICP are shown in Fig. 31.9 [15]. Iron concentration in “stage 1” increased from 0.144 to 0.353 mg l 1 with an increasing amount of Fe2O3 in the US glass. This result indicates that Fe2O3 content in the US glass can regulate the dissolution of iron. We can consider that iron primarily forms FeII species occupying NWM sites, because QD values of FeII site (246 K for US0 glass and 224 K for US20 glass) are smaller than those of FeIII (287 K for US0 glass and 445 K for US20 glass). It is noted that the concentration of phosphate ion in the “stage 1” becomes low with an increasing concentration of Fe2O3 in the waste glass. This tendency is clear in the US20 glass, in which dissolution of phosphate ion is suppressed by adding Fe2O3 into the waste glass; high chemical durability of iron (FeIII) phosphate is responsible for the low dissolution rate of iron.

Change of the concentrations of phosphate ion (left bar) and total Fe ion (right bar) in the artificial drain after stirring with US glasses in “stage 1”. (Reproduced from Ref. 15 with permission of Springer.)

phosphate ion (mg l–1) total Fe ion (mg l–1)

0.5

40

0.4

30

0.3

20

0.2

10

0.1 0.0

0 Blank

US0

US5 US10 US15 US20

Total Fe Ion (mg l–1)

FIGURE 31.9

Phosphate Ion (mg l–1)

50

601

31.2 PROPERTIES AND STRUCTURE OF RECYCLED SILICATE GLASSES

31.2.2 Electromagnetic Property of Recycled Silicate Glasses X-ray fluorescence analysis was used for determining the composition of each FARG [10]. For example, composition of the FARG prepared with 88 mass% of fly ash and 12 mass% of Fe2O3 was determined to be SiO2—58.8, Al2O3— 19.5, Fe2O3—12.5, CaO—3.4, Na2O—1.6, K2O—1.4, MgO—1.1, and others—1.7 mass%. Crystallization temperature (Tc) was determined by differential thermal analysis (DTA) conducted at a heating rate of 10 K min 1, using powdered a-Al2O3 as a standard. DTA of FARG revealed endothermic peaks starting at around 1000  C, accompanied by a plateau at around 1100  C and a minimal peak at around 1220  C. A distinct exothermic peak due to crystallization was observed at around 1320  C. Taking account of these DTA results, isothermal annealing was conducted at 1100  C. M€ ossbauer spectra of FARG containing 5–20% of Fe2O3 are illustrated in Figs. 31.10 and 31.11 [10]. They show that ossbauer spectra magnetic hyperfine structure (hfs) appears when Fe2O3 content is equal to or higher than 14%. M€ evidently prove that this hfs is due to a ferrimagnetic Fe3O4 having an inverse spinel structure [16,18–20]. Taking account of the fact that hfs of Fe2O3 appears when its particle size is larger than about 10 nm [19], we can consider that the diameter of ferrimagnetic Fe3O4 particles is comparable to or larger than about 10 nm. It is also noted that FARG becomes sensitive to a magnet in accordance with the appearance of hfs in the M€ ossbauer spectrum. Internal magnetic field, Bint, for the hfs was estimated to be 47.8, 46.7, and 44.8 T when the Fe2O3 contents were 14, 15, and 20 mass%, respectively [10]. It is known that Bint for crystalline Fe3O4 is 49.0 or 49.1 T [19,20] at room temperature. Hence, it is considered that a decrease in Bint, observed along with an increase in the Fe2O3 content, is ascribed to a decrease in the particle size of Fe3O4 nanocrystals precipitated in the FARG. Second outermost peaks (hfs) of Fe3O4 are associated with a fast electron hopping from FeII to FeIII at the octahedral sites (B) occurring with a relaxation time (t) of 1.1 ns [19]. An electron hopping at site B is known to take place at temperatures higher than the Verwey temperature of 120 K [16,18–20]. M€ ossbauer spectra of FARG containing less than 14% of Fe2O3 show only doublets due to FeIII and FeII, as illustrated in Figs. 31.10 and 31.11 [10]. It is considered that the reduction of FeIII to FeII is caused by carbon and other reducing atoms remaining in the fly ash. M€ ossbauer parameters of FeIII in FARG are summarized in Table 31.1 [10]. Values of d for III the doublet of Fe are in the range of 0.32–0.35 mm s 1. They are typical of FeIII with distorted tetrahedral symmetry

FIGURE 31.10 € ssbauer Room-temperature Mo spectrum of FARG containing 5, 10, 15, and 20 mass% of Fe2O3. (Reproduced from Ref. 10 with permission of Springer.)

602

31 WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS

FIGURE 31.11 € ssbauer Room-temperature Mo spectrum of FARG containing 11, 12, 13, and 14 mass% of Fe2O3. (Reproduced from Ref. 10 with permission of Springer.)

[16,18]. D values larger than 1.2 mm s 1 indicate a large distortion of FeO4 tetrahedra. d values of 1.02–1.11 mm s 1 for FeII indicate a distorted octahedral symmetry [16,18]. D values of 1.94–2.22 mm s 1 for FeII are smaller than those observed for typical distorted tetrahedral FeII [16,18]. Linewidth (G) values for the iron in oxide glasses are generally larger than 0.4 mm s 1 [16,18]. G values obtained for FeIII and FeII, that is, 0.79–0.85 and 0.78–0.93 mm s 1, respectively, are characteristic of the iron present in oxide glasses [16,18]. We can conclude from these M€ ossbauer parameters that FeIII atoms are randomly distributed at the site of NWF, that is, at essentially the same sites as SiIV or AlIII constituting a three-dimensional network of FARG, while FeII atoms occupy the site of NWM, as do Ca2þ, Naþ, Kþ, and so on in several oxide glasses.

€ ssbauer Parameters for the Doublet in FARG [10] TABLE 31.1 Mo Fe2O3 (mass%) 5 5 10 10 11 11 12 12 13 13 14 14

d (mm s 1)

D (mm s 1)

G (mm s 1)

A (%)

Species

1.02 0.34 1.03 0.35 1.11 0.34 1.09 0.32 1.10 0.35 1.07 0.33

2.22 1.34 2.14 1.20 1.98 1.21 2.00 1.26 1.94 1.26 1.94 1.36

0.89 0.79 0.93 0.83 0.81 0.84 0.78 0.84 0.80 0.85 0.69 0.71

57.6 42.4 49.0 51.0 41.1 58.9 39.8 60.2 39.8 60.2 40.0 60.0

FeII FeIII FeII FeIII FeII FeIII FeII FeIII FeII FeIII FeII FeIII

Source: Reproduced from Ref. 10 with permission of Springer. d: isomer shift, D: quadrupole splitting, G: linewidth, A: absorption area.

603

31.2 PROPERTIES AND STRUCTURE OF RECYCLED SILICATE GLASSES

Electrical conductivity (s in S cm 1) was determined from the value of resistivity (R in ohm) and from the size of the glass sample, that is, s¼R

1

A

1

 l;

(31.3)

where A is the cross section (in cm2) and l the length (in cm) between the two electrodes. Heat treatment of several FARG at 1100  C resulted in a remarkable increase in the s from the order of 10 8 to the order of 10 6 S cm 1, as illustrated in Fig. 31.12 [10]. A maximal s value was observed in every sample after heat treatment for 60 min, irrespective of the Fe2O3 content. A maximal s value of 2.2  10 6 S cm 1 was observed in the FARG containing 13% Fe2O3 [10]. The original s value of this sample was 3.9  10 8 S cm 1 [10]. It is noted that no hfs was observed in the M€ ossbauer spectra before heat treatment (Fig. 31.11). The increase in the s is ascribed to a decreased distortion of FeO4, SiO4, and AlO4 tetrahedra, which causes an increased probability of electron hopping from FeII to FeIII of Fe3O4 nanoparticles. A similar increase in s was recently observed in 25K2O  65V2O5  10Fe2O3 and 15BaO  70V2O5  15Fe2O3 glasses [21,22] after being annealed at temperatures close to their individual Tc. A decreased distortion of FeO4 and SiO4 tetrahedra causes an increased probability of electron hopping associated with an increase in s, as described above. Figure 31.12 indicates that an additional heat treatment of FARG for 90 and 120 min resulted in a drop of s from the order of 10 6 to 10 7 S cm 1, for example, to 1.7  10 7 S cm 1 when Fe2O3 content was 13 mass% [10]. They are still higher than that of as-cast glass (3.9  10 8 S cm 1). Drop of s is due to partial crystallization, as observed in some vanadate glasses; 25K2O  65V2O5  10Fe2O3 glass showed a decrease in s due to the precipitation of insulating KV3O8 particles [21]. These experimental results indicate that structural relaxation, that is, a decrease in the distortion of the structural units of the glass network, causes an increased probability of electron hopping from FeII to FeIII. On the other hand, crystallization of FARG results in a drop of s probably because of cleavage of the network; electron hoping is suppressed due to depolymerization or destruction of the glass network. M€ ossbauer spectroscopy is very effective for the local structural change of glass. Room-temperature M€ ossbauer spectra of FARG after heat treatment are illustrated in Figs. 31.13 and 31.14 [10]. When the Fe2O3 content of FARG was 12 mass%, hfs with Bint values of 48.1 and 48.2 T appeared after heat treatment at 1100
C for 30 and 60 min, respectively. In the case of FARG containing 15 mass% of Fe2O3, Bint values of 48.1 and 46.4 T were observed after heat treatment for 30 and 60 min, respectively. ossbauer parameters of FARG heat Change of the distortion of FeO4 tetrahedra is well reflected in D values. M€ treated at around Tc are summarized in Table 31.2 [10]. When the Fe2O3 content was 12 mass%, D values of FeIII decreased remarkably from 1.26 to 1.16 and 1.17 mm s 1 after heat treatment for 30 and 60 min, respectively (Table 31.2 [10]). When Fe2O3 content was 15 mass%, D values of FeIII decreased from 1.34 to 1.20 and 0.94 mm s 1 after heat treatment at 1100  C for 30 and 60 min, respectively (Table 31.2 [10]). A decrease in the D values of FeIII indicates an

Electrical conductivity (10–6 S cm–1)

2.5

2.0

1.5

1.0

FIGURE 31.12 0.5

0

30

60

90

Annealing time (min)

120

Change of the electrical conductivity of FARG containing Fe2O3 of 12 (), 13 (&), 14 (~), and 15 mass% () caused by heat treatment at 1100  C for different times. (Reproduced from Ref. 10 with permission of Springer.)

604

31 WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS

FIGURE 31.13 € ssbauer Room-temperature Mo spectrum of FARG containing 12 mass% of Fe2O3 after heat treatment at 1100  C for 0 (top), 30 (middle), and 60 min (bottom). (Reproduced from Ref. 10 with permission of Springer.)

increased local distortion of FeIIIO4 tetrahedra, since FeIII has a symmetric 3d electron configuration in the outermost orbital, and hence only the steric configuration of the neighboring oxygen atoms determines the magnitude of D. A similar phenomenon was observed for the change of D for FeII; it decreased from 2.00 to 1.84 mm s 1 when Fe2O3 content was 12 mass%, while it decreased from 2.09 to 1.64 mm s 1 when Fe2O3 content was 15 mass% [10]. These smaller D values of FeII indicate a decreased distortion of the FeIIO6 octahedra occupying NWM sites.

FIGURE 31.14 € ssbauer Room-temperature Mo spectrum of FARG containing 15 mass% of Fe2O3 after heat treatment at 1100  C for 0 (top), 30 (middle), and 60 min (bottom). (Reproduced from Ref. 10 with permission of Springer.)

605

31.3 SUMMARY

€ ssbauer Parameters for the Doublet in FARG Annealed at 1100
C for Different Times [10] TABLE 31.2 Mo Fe2O3 (mass%)

t (min)

d (mm s 1)

D (mm s 1)

G (mm s 1)

A (%)

Species

0 0 30 30 60 60 0 0 30 30 60 60

1.09 0.32 0.86 0.15 0.85 0.15 1.02 0.34 0.83 0.15 0.81 0.14

2.00 1.26 1.86 1.17 1.84 1.16 2.09 1.34 1.78 1.20 1.64 0.94

0.78 0.84 0.87 0.76 0.85 0.76 1.03 1.05 1.00 0.60 1.04 0.79

39.8 60.2 37.9 62.1 37.5 62.5 45.5 54.5 71.8 28.2 73.7 26.3

FeII FeIII FeII FeIII FeII FeIII FeII FeIII FeII FeIII FeII FeIII

12 12 12 12 12 12 15 15 15 15 15 15

Source: Reproduced from Ref. 10 with permission of Springer. t: annealing time, d: isomer shift, D: quadrupole splitting, G: linewidth, A: absorption area.

The effect of heat treatment is also reflected in the M€ ossbauer parameters, as summarized in Table 31.2 [10]. d value for FeIII decreased from 0.32 to 0.15 mm s 1 and from 0.34 to 0.15 or 0.14 mm s 1 when Fe2O3 contents were 12 and 15 mass%, respectively [10]. Decrease in d is ascribed to an increased s-electron density at the FeIII nucleus. A similar decrease in d was observed for FeII occupying NWM sites, that is, from 1.09 to 0.86 and 0.85 mm s 1 and from 1.02 to 0.83 and 0.81 mm s 1 when the Fe2O3 contents were 12 and 15 mass%, respectively [10]. Shorter FeIII O or FeII O bond length and a larger overlapping of 4s-orbitals of FeIII and FeII atoms with 2p-orbital of oxygen atoms will be favorable for the 3d-electron hopping from FeII to FeIII by way of oxygen atom. ossbauer spectra becomes close to that of FeIII, as It is noteworthy that the peak position of FeII in the M€ illustrated in Figs. 31.13 and 31.14 (Table 31.2 [10]). These experimental results indicate that electron hopping ossbauer lifetime of 10 7 s. The electron from FeII to FeIII takes place with a timescale close to that of the 57 Fe-M€ II III hopping from Fe to Fe is accelerated after structural relaxation, as confirmed from a marked increase in the s (Fig. 31.12).

31.3 SUMMARY 31.3.1 Water-Purifying Ability of Recycled Silicate Glasses A relationship was found between the water-purifying ability and the local structure of ISLS and US glasses: (1) After a 10 day leaching test, decrease in the absorption area of FeII was observed in the M€ ossbauer spectra of both 10Na2O  10CaO  5Fe2O3  75SiO2 and 25Na2O  25CaO  5Fe2O3  45SiO2 glasses, showing that FeII ions are easily dissolved into water since they were loosely bound to the glass network. (2) Due to the effect of dissolved cations, the ORP of 10Na2O  10CaO  5Fe2O3  75SiO2 glass decreased from 30.0 to 195.8 mV. The value of pH increased from 6.0 to 10.0 as a result of the dissolution of FeII. A much larger decrease in the ORP from 30.0 to 284.7 mV and a large increase in pH from 6.0 to 11.6 were observed after leaching of 25Na2O  25CaO  5Fe2O3  45SiO2 glass. These results suggest that both ORP and pH effectively reflect the dissolution of Na2O þ CaO. (3) It is concluded that iron-containing silicate glass such as 10Na2O  10CaO  5Fe2O3  75SiO2 glass can be a useful water-purifying material having high FeII providing ability. (4) Debye temperature revealed that FeII atoms in US glass occupy the site of NWM, while FeIII atoms occupy the site of NWF. Debye temperature of US glass became higher along with the introduction of Fe2O3, indicating that iron atoms (FeII and FeIII) are more rigidly bound to oxygen atoms. It is concluded that FeII atoms occupy NWM site, being loosely bound to oxygen.

606

31 WATER PURIFICATION AND CHARACTERIZATION OF RECYCLED IRON-SILICATE GLASS

(5) Decrease in COD value could be associated with dissolution of the iron in US glass. US20 glass showed a marked effect on the reduction of COD. A step-by-step increase in the dissolution of FeIII was observed in the leaching test of US glass in the artificial drain. (6) It is concluded that iron-containing inorganic glasses prepared from household garbage cause a marked reduction of phosphate ions and COD value. Iron-containing inorganic glasses can be utilized as a decreasing agent of phosphate ions in wastewater. 31.3.2 Electromagnetic Property of Recycled Silicate Glasses A relationship was found between the electrical conductivity and local structure of FARG: (1) (2) (3) (4)

Fly ash can be recycled to an electrical conducting and ferrimagnetic material. FARG prepared with fly ash and Fe2O3 has s value of 10 8  10 6 S cm 1. When nanocrystals of magnetite are precipitated in a glassy phase, FARG becomes a magnetic material. FeIII atoms in FARG occupy the site of NWF in a three-dimensional glass network, as do SiIV and AlIII, whereas FeII atoms occupy the sites of NWM. (5) Heat treatment of FARG at 1100  C causes a structural relaxation, that is, a decrease in the distortion of FeO4, SiO4, and AlO4 units, accompanying a marked increase in the s .

The authors have proven that iron-containing waste silicate glass and glass–ceramic have an excellent ability for €ssbauer spectroscopy is an water purification, and can be an excellent semiconductor and magnetic material. 57 Fe -Mo extremely powerful tool for characterizing these iron-containing waste silicates because both water-purifying ability and €ssbauer electromagnetic property are strongly affected by the chemical behavior of iron. Finally, we expect that Mo spectroscopy can contribute to the general society by solving several environmental problems.

REFERENCES 1. Ministry of the Environment of Japan, Abstract of IPCC Fourth Assessment Report (http://www.env.go.jp/earth/ipcc/4th/ar4syr. pdf) 2007, p. 35 [in Japanese]. 2. Ministry of the Environment of Japan, Kankyo Hakusho (http://www.env.go.jp/policy/hakusyo/h22/pdf/full.pdf) 2010, pp. 25–28 [in Japanese]. 3. S. Kubuki, N. Kawakami, H. Matsumura, T. Fujisato, T. Nishida, M. Fukagawa, J. Technol. Educ. 2004, 11(2), 77–82 [in Japanese]. 4. J. Tamaki, S. Kubuki, T. Nishida, Bull. Fac. Hum. 2006, 7–12 [in Japanese]. 5. F.C.C. Moura, M.H. Araujo, R.C.C. Costa, J.D. Fabris, J.D. Ardisson, W.A.A. Macedo, R.M. Lago, Chemosphere 2005, 60, 1118– 1123. 6. O.X. Leupin, S.J. Hug, Water Res. 2005, 39, 1729–1740. 7. C.R. Kurkjian, J. Non-Cryst. Solids 1970, 3, 157–194. 8. J.M.D. Coey, J. Phys. (Paris) 1974, 35, C6–C82. 9. G.H. Frischat, G. Tomandl, Glastech. Ber. 1969, 42, 182–185. 10. T. Nishida, M. Tokunaga, Y. Sugata, S. Kubuki, J. Radioanal. Nucl. Chem. 2005, 266, 171–177. 11. S. Kubuki, K. Akagi, H. Sakka, Z. Homonnay, K. Sink o, E. Kuzmann, H. Kurimoto, T. Nishida, Hyperfine Interact. 2009, 192, 31–36. 12. T. Nishida, M. Seto, S. Kubuki, O. Miyaji, T. Ariga, Y. Matsumoto, J. Ceram. Soc. Jpn. 2000, 108(3), 245–248. 13. T. Nishida, S. Kubuki, Hyperfine Interact. (C) 2003, 5, 313–316. 14. K. Shinra, T. Shono, I. Masuda, Kiso Bunseki Kagaku (Basic Analytical Chemistry), Sankyo Shuppan, Tokyo, 1999, pp. 49–51 [in Japanese]. 15. S. Kubuki, N. Kawakami, T. Kamikawa, M. Fukagawa, T. Nishizumi, T. Nishida, Z. Homonnay, E. Kuzmann, Hyperfine Interact. 2005, 166, 429–436. 16. Z. Homonnay, S. Music, T. Nishida, N.S. Kopelev, A. Vertes, in M€ ossbauer Spectroscopy of Sophisticated Oxides, A. Vertes, Z. Homonnay, eds., Akademiai Kiado, Budapest, 1997. 17. J.A. Johnson, C.E. Johnson, D. Holland, A. Mekki, P. Appleyard, M.F. Thomas, J. Non-Cryst. Solids 1999, 246, 101–114.

REFERENCES

607

18. E. Fujita, S. Nasu, T. Nishida, Y. Yoshida, in Introduction to the M€ ossbauer Spectroscopy—Principle and Applications, E. Fujita, ed., Agne Gijutu Center, Tokyo1999, pp. 200–230 [in Japanese]. 19. N.N. Greenwood, T.C. Gibb, in M€ ossbauer Spectroscopy, Chapman and Hall, London, 1971, pp. 246–254. 20. H. Sano, M. Katada, in M€ ossbauer Spectroscopy, Gakkai Shuppan Center, Tokyo, 1996, pp. 76–88 [in Japanese]. 21. T. Nishida, J. Kubota, Y. Maeda, F. Ichikawa, T. Aomine, J. Mater. Chem. 1996, 6, 1889–1896. 22. K. Fukuda, A. Ikeda, T. Nishida, Solid State Phenom. 2003, 90–91, 215–220.

C H A P T E R 3 2

€ MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING EAMONN DEVLIN1 , MICHAIL SAMOUHOS2, AND CHARALABOS ZOGRAFIDIS2 1 2

Institute of Materials Science, N.C.S.R. “Demokritos,” Attiki, Athens, Greece School of Mining and Metallurgical Engineering, National Technical University of Athens, Athens, Greece

32.1 INTRODUCTION Nickel trends in the world market have shown a rising demand over the last decade, with global annual nickel consumption growth greater than 3%, and a constrained supply. World primary refined nickel production of 1.44 Mt in 2010 was anticipated to increase to 1.60 Mt in 2011 and 1.74 Mt in 2012 to satisfy projected increases in consumption [1]. Although a decline in global consumption and an overcapacity is currently predicted for 2013, long term demand for Nickel is expected to rise. While nickel is used in a range of industries including engineering, electrical and electronics, automobile and automobile components, and batteries, most nickel, 65%, is used in the manufacture of stainless steels, with a further 20% in other steel and nonferrous alloys, often for high-specification applications. Primary nickel is produced from two different ores, lateritic and sulfidic. Lateritic ores are normally extracted by opencast mining of ore deposits in layers at varying depths below the surface, while sulfidic ores are mined from underground. Approximately 70% of global land-based nickel resources are contained in laterites, but they are responsible for only about 40% of the global nickel production. Production has been dominated by sulfide ores, as they are easier to process, through conventional mining, smelting, and refining, compared to laterite ores that require intensive hydrometallurgical processing. Thus, laterite ores require substantially more energy and chemicals to produce than the sulfide nickel. Nevertheless, as the global demand for nickel is rising, and many of its uses limit high rates of recycling, a progressive shift to nickel laterite projects in the global nickel industry is underway despite the high environmental cost that it entails. Offsetting this disadvantage are the facts that these ores are found close to the surface and are cheaper to access, and that they contain the valuable element cobalt [2]. The processing costs do mean that major sustainability challenges such as energy and greenhouse emissions remain of paramount importance to the nickel sector [3]. In connection with these challenges, new processing techniques are being sought. One of these techniques is the use of microwave radiation [4]. In this chapter, we will look at the use of M€ ossbauer spectroscopy (MS) as a tool in monitoring traditional and microwave processing of laterite ores in Greece. M€ ossbauer spectroscopy has been used as a characterization tool in a wide range of mineral processes including ilmenites, chalcopyrites, pyrrhotites, bauxites, as well as laterites to some extent [5–9]. The M€ ossbauer technique yields detailed information on the phases present, their composition, structure, and their relative amounts. In the case of laterite processing in particular, in addition to the above useful data, it gives critical information on the overall degree of metal reduction. M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

608

32.2 CONVENTIONAL PROCESSING

609

To get some background, let us quickly review laterite ore processing. The three types of nickel laterite deposits of economic interest for nickel are limonitic, intermediate, and saprolitic (or garnieritic). The main mineralogical phase of the saprolitic ores is garnierite (Ni,Mg)6Si4O10(OH)8, which has a high magnesia and low iron content, with a Ni grade of 1.5–3.5%. Limonitic laterites contain iron oxides as the main mineralogical phase—goethite (a-FeOOH), hematite (Fe2O3), or magnetite (Fe3O4)—thus having a higher iron content and a lower Ni grade (1–2%). Laterite ores are classified into three classes: (i) Class A—garnieritic laterites (Fe < 12% and MgO > 25%); (ii) Class B—limonitic laterites (high Fe content, 15–32% or >32% and MgO < 10%); and (iii) Class C—intermediate laterite ores that lie between garnieritic and limonitic ores (Fe 12–15% and MgO 25–35% or 10–25%). Greek nickeliferous laterite deposits represent 90% of nickel laterite reserves in the European Union and are classified as either type B or C laterites. Reviews of pyrometallurgical processing of laterite ores have been carried out by Bergman [10], Taylor [11], Simons [12], and Diaz [13], among others. The two main stages of standard pyrometallurgical treatment of Greek nickeliferous laterites for ferronickel production are (1) drying, preheating, and controlled prereduction of laterite ores with solid fuel reducing agents and fuel oil in rotary kilns (R/Ks) for the production of a calcine; and (2) smelting reduction of the calcine in open-bath submerged-arc electric furnaces (E/Fs) for the production of Fe–Ni alloy with 12–16% Ni [14]. The reductive reactions of iron, nickel, and cobalt oxides (Fe2O3, FeO, NiO, and CoO) in the calcine with the remaining carbon in the E/Fs are endothermic, and the energy consumed is about 16% of the total energy in the smelting reduction step [14]. Thus the calcine’s reduction degree is an extremely significant parameter for the electric energy requirements in the E/Fs. The chemical characteristics and the granulometry of the calcine also constitute critical parameters for the E/Fs’ energy requirements. It is in stage 2 that microwave technology can be applied, providing an alternative energy source for the laterite calcination and prereduction [15,16]. The use of microwave energy minimizes the duration of the calcination– prereduction process, and can produce a more reduced calcine product, reducing the smelting process energy demand. A range of characterization techniques must be used to monitor the key variables in the processes (conventional processing and microwave heating), in addition to M€ ossbauer spectroscopy, which gives us critical information on the reduction degree.

32.2 CONVENTIONAL PROCESSING To monitor the daily variations in the calcination process in a standard industrial pyrometallurgical process, daily 2–3 kg samples were obtained over a period of 20 days. The calcine samples were produced by a selected R/K and taken from the charging pipes of the E/F that was fed exclusively by the R/K. The calcine samples were immediately cooled to room temperature using nitrogen gas at a high flow rate to avoid reoxidation phenomena. In addition, daily slag samples of approximately 0.5 kg were taken over the same time period from the same E/F. The following parameters were monitored:  Reduction degree (RD) of the calcine, calculated based on iron speciation determined by M€ ossbauer spectroscopy [17].  Chemical analysis of the calcine and slag samples by atomic absorption spectrometry (AAS) and LECO induction furnace for the determination of carbon and sulfur content.  Temperature values of calcine and slag, measured by laser pyrometer.  Melting temperature of slag samples, determined by triangular diagrams.  Grain size analysis of the calcine samples, experimentally determined by the use of standard test sieves.  Fe and SiO2 content (%) of the laterite feed in the R/K within the examined time period based on the R/K operational data.  E/F energy consumption index (kWh T 1 of calcine).  Chemical analysis of nickel in the Fe–Ni alloy.  E/F electrode consumption index (millimeters of slipping per megawatthour per 24 h). We will present the M€ ossbauer data to show its usefulness in monitoring various stages of the process. Transmission M€ ossbauer spectroscopy was carried out on powder samples at room temperature with a Co57(Rh) source in constant acceleration mode using a thin iron foil for calibration.

610

€ 32 MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING

0.0

Absorbtion (%)

–0.5 –1.0 –1.5 Fit Data Fe2.5+ Ox Fe2+ NM Fe2.5+ NM Fe3+ Ox

–2.0 –2.5 –3.0 –10

–5

0 Velocity (mm s–1)

5

10

0.0

FIGURE 32.1 € ssbauer Room-temperature Mo spectra of samples 4 (upper) and 14 (lower). Ox: Magnetic iron oxides. NM: nonmagnetic iron containing phases. (Reproduced from Ref. 21 with permission of Wiley.)

Absorbtion (%)

–0.5

–1.0

–1.5

Fit Data Fe2.5 Ox Fe 2+ NM Fe2.5 NM Fe3+ Ox

–2.0

–10

–5

0 Velocity (mm s–1)

5

10

Previous laboratory studies conducted on the roasting reduction of Greek laterite samples have revealed that iron in the calcine is mainly found in the form of hematite (a-Fe2O3), nonstochiometric magnetite (Fe3O4), and complex iron– silica phases, such as fayalite (2FeO.SiO2) [18]. In Fig. 32.1, the M€ ossbauer spectra of two representative samples from days 4 and 14 (#4, #14), with different degrees of reduction, are presented. The significantly higher nonmagnetic ferrous content of sample 4 is clearly visible. Ferrous and ferric components are observed along with intermediate states (Fe2.5þ) that represent iron sites such as the B site in magnetite, which is a mixture of Fe2þ and Fe3þ iron states. Nonmagnetic phases and magnetic oxides (hematite, nonstochiometric magnetite) are observed. No metallic a-iron is detected. The different subspectral components are graphically distinguished in Fig. 32.1 on the basis of the iron oxidation state. Thus, the M€ ossbauer spectroscopy yields an accurate quantitative analysis of the content of the mineral phases present that allows us to determine the degree of reduction (RD ¼ (Fe2þ/Fetot)%), Fig. 32.2 [19].

FIGURE 32.2 Reduction degree of the calcine samples. (Reproduced from Ref. 21 with permission of Wiley.)

611

32.2 CONVENTIONAL PROCESSING

FIGURE 32.3 Reduction degree of calcine versus temperature. (Reproduced from Ref. 21 with permission of Wiley.)

Correlations between the reduction degree and other parameters such as the calcine temperature, determined by the use of an optical pyrometer, can be obtained (Fig. 32.3). It should be noted that the temperature measurements of the calcine samples are made at the E/F freeboard and correspond to the temperature of the calcine fed into the charging pipes of the E/F. The temperature of the calcine fed into the E/F is about 100  C lower than that of the calcine exiting the R/K. It is clear that an increase of the calcine temperature favors the reduction. Previous work on the reducibility of Greek nickeliferous laterites has shown the importance of the calcine temperature in rotary kiln roasting reduction [20]. As hematite reduction can take place above 570  C, increasing the R/K temperature improves both the reduction degree and the reduction rate. We also see, in Fig. 32.4, that when the calcine reduction degree increases, we obtain a decrease in the calcine’s carbon content. Both temperature and calcine carbon content are critical parameters for the smelting reduction procedure in the E/F. A higher calcine temperature corresponds to reduced electric energy requirements both for smelting and for the endothermic reductive reactions of iron and nickel oxides of the calcine in the E/F. Thus, we see the crucial role that MS can play in generating RD values that can be correlated with various control parameters to obtain better process control. For example, limiting the calcine carbon contents results in fewer operational problems and a quieter smelting reduction process, as smaller volumes of reduction gases are generated through the slag in the E/F. Examining the variation of reduction degree with Nickel content reveals another important correlation. A lower calcine carbon content due to higher reduction results in a lower reduction degree of the iron oxides inside the E/F, Fig. 32.5. This results in an increased nickel content in the Fe–Ni alloy and an increased iron content in the slag. Furthermore, increasing the iron content in the slag reduces the slag resistivity, with the result that the electrodes are lifted more out of the slag and the electrode current increases [21]. The decreased contact between the electrode and the slag leads to decreased electrode consumption, which is a crucial factor in the economics of the smelting reduction process. In Fig. 32.6, we clearly see the effect of reduction degree on electrode slipping. Thus, we see how the reduction degree, obtained from M€ ossbauer data, yields information bearing on critical economic parameters. The optimization of open-bath submerged-arc E/F operation for laterite smelting reduction, in terms of energy consumption, electrode consumption, and limitation of the operational problems, is a multiparametric problem. The temperature and the reduction degree of the calcine feeding material, the iron and silica content of the calcine and slag, the ore grain size of the calcine, and the temperature of the slag, all constitute critical parameters. M€ ossbauer spectroscopy can yield vital information to help with the understanding and modeling of such systems.

FIGURE 32.4 Reduction degree of calcine versus residual carbon content. (Reproduced from Ref. 21 with permission of Wiley.)

612

€ 32 MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING

FIGURE 32.5 Ni content in the Fe–Ni produced in E/F versus calcine reduction degree. (Reproduced from Ref. 21 with permission of Wiley.)

FIGURE 32.6 Electrodes slipping versus calcine reduction degree. (Reproduced from Ref. 21 with permission of Wiley.)

Mathematical modeling of the roasting reduction–smelting reduction process using comprehensive sets of such raw data should be the goal for optimization of the metallurgical method.

32.3 MICROWAVE PROCESSING The use of microwaves has huge potential for mineral processing arising from several clear advantages of the microwave technique. These include the direct transfer of heat to the target, with little thermal loss to the surrounding area, and the reduction of heat transfer problems associated with poor thermal conductivity. Microwaves are a clean and controllable energy source that facilitate continuous materials processing. Implementation of this technique means addressing technical problems that sometimes occur, for example, inhomogenous sample heating and financial issues associated with energy and equipment costs. Microwave radiation generates heat rapidly inside absorptive materials, and the heat then spreads via conduction through the material. The prime absorption mechanisms are dipolar rotation and electron/anion resistivity that produce ohmic heating (i.e., conductive currents due to the movement of ion or electron charges) [22]. In nonmagnetic media, microwave heating at 2.45 GHz is determined by the temperature-dependent complex permittivity [e (T)] and, in particular, by its imaginary part [e00r ðT Þ], which is a measure of the electromagnetic energy conversion into heat. e ðT Þ ¼ e0r ðT Þ

je00r ðT Þ;

(32.1)

where e ðT Þ is the complex permittivity, e0r ðT Þ is the relative real permittivity or relative dielectric constant, and e00r ðT Þ is the relative imaginary permittivity or relative dielectric loss. The average power (PL) absorbed by a material per unit volume is estimated by
 1 PL ¼ ve0 e00r ðT ÞjEint j2 W cm 3 ; 2

(32.2)

613

32.3 MICROWAVE PROCESSING

Drying rate (g mm–2 sec x 10–7)

60 Microwave Drying at 750 W Conventional Drying at 140°C

50 40

FIGURE 32.7

30

Microwave drying rates at 750 W and conventional drying rates of limonitic nickeliferous laterite ores as a function of moisture fraction. (Reproduced from Ref. 23 with permission of the Canadian Institute of Mining, Metallurgy, and Petroleum.)

20 10 0 0.0

0.1

0.2

0.4 0.5 0.6 0.7 – Moisture content (x)

0.3

0.8

0.9 1.0

where v is the angular frequency (2pf, in rad s 1), Eint is the internal electric field in the sample (V m 1), and e0 is the permittivity of free space. It is clear from the foregoing equations that the dielectric properties, and especially e00 , determine the rate of microwave heating and can be used to predict the successful application of microwave heating for a pyrometallurgical process. In the last decade, a great deal of work has been published on the application of this technique to the pyrometallurgical processing of oxide ores. A review of this work by Pickles gives a good indication of the range of applications of the technique at all stages of processing: drying, calcination, sintering, reduction and smelting, slag reduction, segregation, and the processing of electric arc furnace dust [4]. A single example of the power of this technique can be seen in its use in the drying stage (Fig. 32.7). Greatly improved drying rates are achieved due to the different heating and moisture transport properties associated with the microwave method, for example, direct microwave heating of the moisture content rather than conventional heating of the entire sample mass. For the microwave processing study, we used hematitic nickeliferous laterite ore from the Lokrida area, one of the main laterite sources in Greece. A sample mass of about 50 kg was collected from the homogenized ore heaps and was pulverized using an LM2 Labtechnics pulverizing mill. A 5 kg charge grinding with 2 min of milling was sufficient to mill the total mass of laterite to a 16 mesh ( 1.18 mm) particle size. The reducing agent applied was lignite (30.1% Cfix). The bulk laterite sample was analyzed by X-ray fluorescence using a Xepos Spectro bench instrument (Table 32.1). Iron oxide (40.7 mass%) and silica (30.6 mass%) are the major chemical phases with noticeable amounts of MgO, Al2O3, and Cr2O3 also present. The metallic nickel content is about 1 wt%. XRD of the raw bulk laterite and the carbothermic reduced laterite samples was carried out using a Bruker D8 Focus analyzer. The XRD pattern of the raw bulk laterite ore is presented in Fig. 32.8. It shows that the iron oxide is present almost entirely in the form of hematite. In addition to quartz, the other important phases are clinochlore, montmorillonite, and calcite. The dominant hematite phase, together with the absence of nickel–silicate minerals, classifies the laterite ore as type C. The carbothermic reductions of the laterite samples were done using a Ceralink ThermWave 1.3 microwave furnace working at 2.45 GHz. Temperature measurements of the top surface of the powder samples were done using an optical pyrometer Raytek MX4, located at the end of a 20 cm long and 4 cm in diameter bronze 2.45 GHz choke tube, attached to the center of the metal furnace wall. The optical pyrometer was connected to a PC and measurements were recorded every 0.1 s. A real-time temperature versus time diagram was produced using the appropriate software. The variation of

TABLE 32.1 XRF Chemical Analysis of Bulk Hematitic Laterite Ore (wt%) Species

SiO2

Fe2O3

MgO

Cr2O3

CaO

Al2O3

NiO

K2O

L.O.I

30.6

40.7

4.5

4.0

3.2

4.8

1.3

0.8

8.0

Source: Reproduced from Ref. 19 with permission of Elsevier.

€ 32 MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING

614

200 1. SiO2 - Quartz 2. Fe2O3 - Hematite 3. (Mg,Fe,Al)6(Si,Al)4O10(OH)8 - Clinochlore 4. (Na,Ca)0.3(Al,Mg)2Si4O10(OH)2 xH2O Montmorillonite 5. FeOOH - Goethite 6. KAl2Si3AlO10(OH)2 - Muscovite 7. CaCO3 - Calcite 8. Mg3(OH)2Si4O10 - Talc

1

Intensity (cps)

150

100 1 2

50

3

3

2 1

4

2

334

66 8

3

12

33

5

1 1,2

1

1

7

2 2

2

1 2

3

1

2

1

3

0 5

10

15

20

25

30

35

40

45

50

55

60

65

70

2θ

FIGURE 32.8 X-ray pattern of the bulk hematitic laterite ore. (Reproduced from Ref. 19 with permission of Elsevier.)

microwave power was achieved employing a proportional power controller connected to the furnace magnetron. Powder samples were placed, without applying pressure, in a cylindrical alumina crucible with r ¼ 4 cm, h ¼ 6 cm and d (thickness) ¼ 3.3 mm, (r ¼ 7.5, h ¼ 10, and d ¼ 3.3 mm crucible was used in the case of the 477 g sample), which was positioned at the center of the furnace on a low-density porous alumina platform. Measurements of the real and imaginary permittivities of hematitic laterite samples were done using the cavity perturbation technique at the Microwave Properties North laboratory in Canada [24]. The apparatus consists of a cylindrical cavity connected to a network analyzer and a conventional resistance furnace. The materials, in powder or pellet form, are placed in a long, slim quartz sample holder tube, and are then loaded vertically into the system Fig. 32.9, whereupon the furnace temperature is increased in steps. After a soak at each temperature step, the samples are rapidly moved down into the cavity for measurement (2 s duration), and then rapidly returned to the furnace. The method is based on measuring the difference in the cavity resonant frequency and Q factor between the empty sample holder and the holder with sample. The dielectric constants are derived from the resonant frequency and Q factor changes produced by the sample, and are

FIGURE 32.9 Schematic diagram of the TM0n0 cavity system (in cross section) showing the linear actuator with the quartz sample holder and a sample located on the y axis in the center of the cavity. (Reproduced from Ref. 19 with permission of Elsevier.)

615

32.3 MICROWAVE PROCESSING

0

–4

Fit Data

(d)

Absorption (%)

0

–4 0

–4

Fe2+ metallic Fe(Ni)

Fe3+ (oxide) Fe2+/3+(oxide) alpha-Fe

(c)

(b)

0

(a) –4 –10

–8

–6

–4

–2

0

2

4

6

8

10

Velocity (mm s–1)

FIGURE 32.10 € ssbauer spectra of The Mo reduced laterite samples at 800 W, with two times the stoichiometric carbon addition and 11.75 g sample mass at 120 (a), 240 (b), 360 (c), and 480 (d) s. (Reproduced from Ref. 19 with permission of Elsevier.)

dependent on the sample and cavity shapes and volumes. The laterite samples were measured in powder form in a stagnant air atmosphere. M€ ossbauer spectroscopy is again used as a tool to assess the phase composition and the degree of reduction of the samples with processing. For example, the evolution of iron oxide phases during the reduction of a 10 g laterite sample with twice the stoichiometric addition of carbon at 800 W is presented in Fig. 32.10, where the spectra of reduced laterites at 120 (a), 240 (b), 360 (c), and 480 (d) s are recorded. The relative amounts of iron in each phase are directly determined from their subspectral areas and are given in Table 32.2. Divalent iron is found in nonmagnetic silicate mineral phases and in the magnetic nonstoichiometric magnetite phase. The presence of Fe2þ is already significant (42%) after 120 s, considering that the Fe2þ content in the initial laterite is zero (all the iron is in the form of hematite), while the metallic iron (i.e., bcc Fe(Ni)) recovery reaches 11%. Metallic iron–nickel alloy is also present as a nonmagnetic phase, probably nonmagnetic g-FeNi that can form for 10at% Ni alloys, the approximate composition of the FeNi metallic phase. Scanning electron microscopy data combined with EDS analysis on these samples [19] have shown that the reduced samples consist of a silicate matrix within which the phases FeNi, spinel, Mg–ferrosillite, and chromite are dispersed. A layer of metallic iron (FeNi) adjacent to an iron oxide phase, which is normally observed in the case of conventionally carbothermic reduced laterites, is not observed. Particles of the Fe(Ni) alloy are visible in the processed materials, Fig. 32.11. The slightly negative isomer shift observed for the nonmagnetic metallic phase agrees with published values for g-FeNi. Another possibility is that a nonmagnetic component could arise from superparamagnetic nanoparticles of a-Fe(Ni). In some other samples, lowtemperature measurements do indicate the existence of a-FeNi superparamagnetic nanoparticles. In whichever of these two phases the iron is present, a or g, this iron is metallic, that is, completely reduced. At 240 and 360 s, the reduction process progresses as can be seen by the further elimination of hematite, Fe3þ oxide, and the respective increase of

TABLE 32.2 Iron-Containing Phase Content (wt%) of the Reduced Laterite Samples Samples a b c d

Fe3þ oxide

Fe2þ

Fe2þ3þ Oxide

Nonmagnetic Fe(Ni)

a-Fe(Ni)

31 9 4 13

42 72 69 29

16 5 2 7

7 4 6 18

4 10 19 34

Source: Reproduced from Ref. 19 with permission of Elsevier.

616

€ 32 MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING

FIGURE 32.11 A spherical Fe(Ni) alloy particle 0.7 mm in diameter formed after 240 s of microwave reduction. (Reproduced from Ref. 19 with permission of Elsevier.)

a-Fe(Ni). After 480 s of reduction, the Fe2þ content decreases, indicating its reduction to a-Fe(Ni), which has increased significantly. It should be noted that during the reduction process, the content of Fe2þ/3þ oxide, which refers to a nonstoichiometric magnetite phase, is low. The nonmagnetic Fe2þ component is in the form of the iron silicate minerals fayalite, ferrosilite, and kilchoanite that have been detected using X-ray diffraction. No significant w€ustite content is indicated. Finally, the continuing presence of Fe3þ oxide—probably hematite—even after 8 min of microwave reduction is likely due to the temperature heterogeneity of the sample (Fig. 32.12), indicating that the external surface of the laterite has not reduced. In Fig. 32.13, the reduction degrees of 10 g laterite samples with stoichiometric addition of carbon (1.75 g of lignite) at three microwave power levels (200, 400, and 800 W) are presented. It is apparent that power supply has a great effect on the reduction degree and rate. A reduction degree of 60% has been achieved at 600 s using an input power of 800 W, while only a 32% reduction degree has been achieved in the same time using 200 W. The experimental data in the case of double the stoichiometric addition of carbon (3.5 g of lignite) are presented in Fig. 32.14. These data show that the reduction degree is significantly affected by the amount of the reducing agent (lignite) added. The doubling of carbon content results in an increase in reduction degree from 60 to about 70% at 600 s and 800 W power. In Fig. 32.15, the effect of laterite–lignite mass on the reduction degree of laterite at an input power of 800 W and two times the stoichiometric addition of carbon is presented. By increasing the laterite’s mass from 10 g (13.5 g of mixture) to 365.5 g (493.5 g of mixture), a strong decrease of the reduction degree is observed, from about 70 to 25%. The effect on the reduction degree of sample mass should be correlated with the temperatures achieved for the respective samples under a given microwave radiation, Fig. 32.16. In Fig. 32.17, differential thermal analysis (DTA) data (obtained using a TG Labys-DSC system in the temperature range 25–970  C with a 10  C min 1 heating rate and in a stagnant air atmosphere) for the laterite–lignite mixture are compared with both the real (e0 ) and the imaginary (e00 ) permittivities at 2.45 GHz in the temperature range 25–970  C. The dielectric constant data present a significant increase in the values of e0 and e00 , noticeable after 300  C. The increase, especially in the case of the imaginary permittivity (e00 ), is sharp and acquires its maximum value at 585  C (from 1.8 at 300  C to 5.2 at 585  C). Given that the formation of the magnetite phase occurs at 572.5  C, and that the Curie point of magnetite is 585  C, the concurrence of the temperature in which e00 maximizes with the magnetite Curie point (585  C) should be noted. Several research studies on the measurement of the dielectric properties of magnetite, in powder and pellet forms, have shown a dramatic increase in e00 at the Curie temperature (from 5 at 300  C to about 35 at 585  C) [25,26]. Between 585 and 980  C, the value of e00 gradually decreases as the volume of magnetite decreases, it being reduced to w€ ustite in this temperature interval. Thus, the dielectric constant of the laterite–lignite mixture increases in the temperature range of 450–750  C. However, these results may not be directly related to the actual dielectric properties of the mixture under microwave heating conditions. The extremely high rate of temperature increase during microwave heating cannot be achieved during the cavity perturbation measurement. Thus, the high temperatures developed in the core of the sample in the case of microwave heating result in the formation of phases (e.g., ferrosilicates), which are not formed in the cavity perturbation

617

32.3 MICROWAVE PROCESSING

FIGURE 32.12 Infrared images of the crucible containing laterite–lignite mixtures. Lateral images at (a) 120, (b) 240, and (c) 360 s of microwave heating and a vertical image (d) at 120 s. (Reproduced from Ref. 19 with permission of Elsevier.)

instruments. Nevertheless, the data on the dielectric properties of laterite–lignite mixture give a useful order of magnitude indication of their evolution in the temperature range between 25 and 980  C. From the study of the microwaved materials we can deduce the following (1) The laterite–lignite mixture is susceptible to microwave radiation and can reach a temperature of 900  C after 120 s of heating using a power of 800 W. The heating behavior of the mixture is strongly affected by the power

Reduction degree (%)

70 200 W

60

400 W

50

800 W

40

FIGURE 32.13

30

Reduction degree as a function of time and microwave power supply for a laterite–lignite mixture sample of 11.75 g (stoichiometric carbon addition). (Reproduced from Ref. 19 with permission of Elsevier.)

20 10 0 0

100

200

300 400 Time (s)

500

600

700

€ 32 MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING

618

Reduction degree (%)

80

FIGURE 32.14 Reduction degree as a function of time and microwave power supply for a laterite–lignite mixture sample of 13.50 g and double the stoichiometric carbon addition. (Reproduced from Ref. 19 with permission of Elsevier.)

70

200 W

60

400 W 800 W

50 40 30 20 10 0 0

100

200

300 400 Time (s)

500

600

700

Reduction degree (%)

80

FIGURE 32.15 Logarithmic scale diagram of reduction degree as a function of sample mass for a laterite–lignite mixture with two times the stoichiometric carbon addition at 800 W. (Reproduced from Ref. 19 with permission of Elsevier.)

70 60 50 40 30 20 10 0 10

100 Mass (g)

1000

1000 m = 47 g

900

m = 117.5 g

FIGURE 32.16 Microwave heating of laterite– lignite mixtures (stoichiometric addition of carbon) of various masses at 800 W. (Reproduced from Ref. 19 with permission of Elsevier.)

Temperature (°C)

800

m = 493.5 g

700 600 500 400 300 200 100 0 0

100

200

300

400

500

600

700

800

Time (s)

supplied and the mass of the sample. A power setting of 800 W is insufficient to heat laterite–lignite mixture samples with masses over 47 g effectively. (2) The mineralogical analysis of laterite-reduced samples after 120, 240, 360, and 480 s of heating indicates the rapid reduction of hematite to magnetite, metallic iron, and iron silicate phases. Also, the occurrence of temperature heterogeneity is supported by the presence of materials that are not stable at the same temperature conditions (e.g., after 240 s of heating, although the metallic iron–nickel alloy has been formed, the calcite phase has still not been totally calcined). (3) The comparison of DTA data of laterite–lignite heating with the dielectric properties shows an important increase in the value of the imaginary permittivity (e00 ) after 300  C, which maximizes at 585  C (from 1.8 at

619

REFERENCES

Dielectric constants

7

12 10 8 6

6 5 4

4 2 0 –2

3 2

ε΄ 5xε''

1

DTA

0 0

100

200

300

400

500

600 o

Temperature ( C)

700

800

900

–4 –6 1000

Heat flow (µV)

16 14

8

FIGURE 32.17 Comparison of the real and imaginary permittivities (e0 and e00 ) of the laterite–lignite mixture at 2.45 GHz with the DTA data. (Reproduced from Ref. 19 with permission of Elsevier.)

300  C to 5.2 at 585  C). It is the Curie point of magnetite, which forms at 572  C, as the DTA curve demonstrates. The values of the measured dielectric constants do not accurately correspond to the “real” respective constants of the mixture while being heated with microwaves, because the heating rates are much higher and a number of different mineralogical phases are formed where topically high temperatures are achieved. However, the measured constants approach the “real” values in parts of the sample that were heated relatively smoothly with microwaves, that is, the external surface. (4) The scanning electron microscopy combined with EDS analysis showed that the reduced samples consist of a silicate matrix within which the phases FeNi, spinel, Mg–ferrosillite, and chromite are dispersed. A layer of metallic iron (FeNi) adjacent to an iron oxide phase, which is normally observed in the case of conventionally carbothermic reduced laterites, was not observed. (5) The reduction degree of laterite was measured using M€ ossbauer spectroscopy and was examined as a function of the reducing time, the added carbon content, and the mass of sample. It was demonstrated that the optimum conditions in case of a 10 g reduced laterite sample were two times the stoichiometric addition of carbon and a power supply of 800 W. Under these conditions, a 70% reduction degree has been achieved after 480 s of heating. The reduction degree is strongly affected by the laterite–lignite sample mass. A power supply of 800 W was not efficient to heat a 493.5 g laterite–lignite sample effectively. In this case, only a 25% reduction degree was achieved after 480 s. In these studies, we can see the power of M€ ossbauer spectroscopy for shedding light on many important aspects of mineral processing in this example of laterite processing.

REFERENCES 1. International Nickel Study Group Press Release, Lisbon, Sept 2011. 2. D.A. Davli, G.W. Bacon, C.R. Osborne,“The past and the future of nickel laterites”. PDAC (Prospectors and Developers Association of Canada) International Convention, March 7–10, 2004, Canada, pp. 1–27. 3. G.M. Mudd,“Nickel Sulfide Versus Laterite: The Hard Sustainability Challenge Remains.” Proceedings of the 48th Annual Conference of Metallurgists, August 2009, Canadian Metallurgical Society, Sudbury, Ontario, Canada, 2009, 4. C.A. Pickles, Microwaves in extractive metallurgy: Part 2: A review of applications, Miner. Eng. 2009, 22(13), 1102–1111. ossbauer and X-ray diffraction study of the ilmenite reduction process in a 5. B. Saensunon, G.A. Stewart, R. Pax, A combined 57 Fe-M€ commercial rotary kiln, Int. J. Miner. Process. 2008, 86(1–4), 26–32. 6. D. Bandyopadhyay, R.M. Singru, A.K. Biswas, Study of the roasting of chalcopyrite minerals by 57 Fe M€ ossbauer spectroscopy, Miner. Eng. 2000, 13(8–9), 973–978. 7. A. Navarra, J.T. Graham, S. Somot, D.H. Ryan, J.A. Finch, M€ ossbauer quantification of pyrrhotite in relation to self-heating, Miner. Eng. 2010, 23(8), 652–658.

620

€ 32 MOSSBAUER SPECTROSCOPY IN THE STUDY OF LATERITE MINERAL PROCESSING

8. E. Murad, Characterization of a standard bauxite and its deferration products by M€ ossbauer spectroscopy, Miner. Eng. 2005, 18(9). 9. V. Anila, R.M. Singru, V.N. Sharma, S. Chander, M€ ossbauer study of the treatment of nickeliferous laterite ore in argon and hydrogen/water mixtures, Hydrometallurgy 1981, 7, 329–337. 10. R.A. Bergman, Nickel production from low-iron laterite ores: process descriptions, CIM Bull. 2003, 96(1072), June–July, 127–138. 11. A. Taylor, ALTA 2000 Nickel/Cobalt-6: Seminar Proceedings Review of Nickel/Cobalt Laterite Processes, Alta Metallurgical Services, Perth, West Australia, 2000. 12. C.S. Simons, Proceedings of a Symposium on The Production of Nickel: Extractive Metallurgy—Past, Present and Future, Extractive Metallurgy of Nickel and Cobalt, The Metallurgical Society, C.P. Tyroler, C.A. Landolt, eds., 117th TMS Annual Meeting, Phoenix, Arizona, 1988, pp. 91–134. 13. C.M. Diaz, C.A. Landolt, A. Vahed, A.E.M. Warner, J.C. Taylor, Proceedings of a Symposium on Extractive Metallurgy of Nickel and Cobalt, The Metallurgical Society, C.P. Tyroler, C.A. Landolt, eds., 117th TMS Annual Meeting, Phoenix, Arizona, 1988. 14. E.N. Zevgolis, et al., EPD Congress on Energy Requirements in Nickeliferous Laterite Treatment, 12–16 March 2006, San Antonio, Texas, 2006, pp. 487–496. 15. J. Ma, C.A. Pickles, Microwave segregation process for nickeliferous silicate laterites, Can. Metall. Quart. 2003, 42(3), 313–326 16. C.A. Pickles, Microwave heating behaviour of nickeliferous limonitic laterite ores, Miner. Eng. 2004, 17(6), 775–784. 17. ASTM, Standard E277 Test method for total iron in iron ores by stannous chloride reduction and dichromate titration, 1988. 18. E. Zevgolis, C. Zografidis, I. Halikia, E. Devlin, Roasting reduction study of Greek nickeliferous laterites, 138th TMS Congress, San Francisco, California, 2009. 19. M. Samouhos, M. Taxiarchou, R. Hutcheon, E. Devlin, Microwave reduction of a nickeliferous laterite ore, Miner. Eng. 2012, 34, 19–29. 20. N.E. Zevgolis, C. Zografidis, T. Perraki, E. Devlin, Phase transformations of nickeliferous laterites during preheating and reduction with carbon monoxide, J. Therm. Anal. Calorim. 2010, 100, 133–139. 21. K. Karalis, C. Zografidis, A. Xenidis, S. Tabouris, E. Devlin, Contribution to the energy optimization in the pyrometallurgical treatment of Greek nickeliferous laterites, in Third International Symposium on High-Temperature Metallurgical Processing, T. Jiang, J.-Y. Hwang, P. Masset, O. Yucel, R. Padilla, G. Zhou, eds., Wiley, Hoboken, NJ, 2012. 22. C.A., Metaxas, J.M.R., Meredith, Industrial Microwave Heating, Peter Peregrinus, London, 1983. 23. C.A. Pickles, Dielectric drying of nickeliferous limonitic laterite ores, in Proceedings of Nickel and Cobalt 2005—Challenges in Extraction and Production, J. Donald, R. Schonewille, eds., Calgary, AB, Canada August 21–24, 2005, pp. 515–530. 24. R. Hutcheon, M. De Jong, F. Adams, A system for rapid measurements of RF and microwave properties up to 1400  C, J. Microw. Power Electromagn. Energy 1992, 27(2), 87–92. 25. M. Hotta, M. Hayashi, K. Nagata, High temperature measurement of complex permittivity and permeability of Fe3O4 powders in the frequency range of 0.2 to 13.5 GHz, ISIJ Int. 2010, 51(3), 491–497. 26. Z. Peng, J.-Y. Hwang, J. Mouris, R. Hutcheon, X. Huang, Microwave penetration depth of in materials with non-zero magnetic susceptibility, ISIJ Int. 2010, 50(11), 1590–1596.

Index Adipato (adi) complex, 116 Akaganeite, 416–418 collinear antiferromagnetism, 418 crystallinity, 418 Debye–Waller factor, 416 hyperfine parameters, 417, 418 properties, 418 room temperature M€ ossbauer spectrum, 417 three sexters for M€ ossbauer spectra, 416, 417 x-ray diffraction peaks, 418 Alkaline earth metal fluorides, 202 Aluminous silicate perovskite, 43 Aluminum photomicrograph of, 586 SEM element mapping, 585 Ankerite (CaFe(CO3)2), 576 b–g Anticoincidence detector, 60 APD. see Avalanche photodiode detectors (APDs) Aqueous Fe(ClO4)3 solutions, 477 As–As bonding s–p hybridization, 538 Ash fusion temperature (AFT) centrifuge overflow sample, 587 of coal source, 586 Atomic force microscopy, 374, 429 Atomic jumping, 60 Auger electrons, 338 Auger electron spectroscopy (AES), 6, 457, 460, 464 Auger ionization process, 333 Avalanche photodiode detectors (APDs), 4, 7–9, 44, 257, 268 Averaged powder magnetization, for ground doublet, 107 Azospirillum brasilense, 341, 343, 344 emission M€ ossbauer spectra, 337 Bacterial glutamine synthetase molecule, 341 BaFe2As2 atmospheric pressures, 539, 540 crystal structure, 536

low-spin nonmagnetic state, 541 M€ ossbauer spectroscopy, 536 Ba(Fe0.91Co0.09)2As2, magnetic susceptibility, 538 ossbauer Ba(Fe0.88Co0.12)2As2, M€ spectroscopy (MS), 537 Ba(Fe0.95Ni0.05)2As2, magnetic susceptibility, 539 BaFe1.9Ni0.1As2 single-crystal, 538 BaFe2O4 crystalline particles, 545 BaFeO4 sample, 515 20BaO
10Fe2O3
70V2O5 glass activation energy, 547 current vs. voltage, 546 DTA chart, 545 electrical conductivity (s) and quadrupole splitting, 547 room-temperature electrical conductivity, 546 XRD pattern, 545 Barium ferrates(IV), 510 Barium–iron–vanadate glass electrical resistivity, 543 BaSnF4, structural determination historical perspectives, 216–217 M€ ossbauer spectroscopy, 220 neutron powder diffraction, 225 sp3d2 hybridization, and VSEPR rules, 225–226 tetragonal symmetry, 225 BCS theory, superconductors, 397 Benzene-enclathrated Fe(NCS)2(bpp)2
2 (benzene), 147 conformers of bpp, 147 magnetic measurements with SQUID, 149 packing view, 148 projections of Fe(NCS)2(bpp)2 to ab plane, 148 reversible structural changes, triggered by sorption of benzene molecules, 147–149 symmetry change around iron, 148 x-ray diffraction analysis, 147 Be window, 6, 7 Bifunctional mechanism, 564

Biology-related and biomedical applications, of M€ ossbauer spectroscopy, 272 enzymes, 280–281 ferritin, and hemosiderin, 283–284 hemoglobin, 281–283 microorganisms-related studies, 273–276 pharmaceutical products, 286 controlling iron state, and impurity level, 286 iron-containing, 286 iron–polygalacturonate complex, characterization, 286 plant tissues, 276 cucumber roots, spectroscopic studies, 278–279 spectroscopic studies of iron uptake, and translocation, 276–277 tissues containing iron-storage proteins, 284–286 1,3-Bis(4-pyridyl)propane, 147 Bohr magneton, 361 Boltzmann constant, 546 Borate glass, 542 bpa: 1,2-Bis(4-pyridyl)ethane, 144 bpp: 1,3-Bis(4-pyridyl)propane, 144 Bravais single-crystalline lattice, 26 Brown’s model, 209 Ca2þ concentrations, 597 CaIrO3- type structure (Cmcm), 52 Calcine temperature residual carbon content, 611 in rotary kiln roasting reduction, 611 vs. electrodes slipping, 612 vs. temperature, 611 Calcium photomicrograph of, 586 SEM element mapping, 585 Calcium ferrate(VI) FTIR spectra of, 507 Calcium silicate perovskite, 43 Carbon felts microwave discharge, 528

M€ ossbauer Spectroscopy: Applications in Chemistry, Biology, and Nanotechnology, First Edition. Edited by Virender K. Sharma, G€ ostar Klingelh€ ofer, and Tetsuaki Nishida. Ó 2013 John Wiley & Sons, Inc. Published 2013 by John Wiley & Sons, Inc.

621

622

Carbon monoxide (CO) preferential oxidation (PROX), in rich hydrogen atmosphere, 564 Carbothermal reduction, of iron, 555 Cation–cation bond (CCB), 95 Cation-free glutamine synthetase emission M€ ossbauer spectra, 343 Cd-goethite particles, formation, 493 CEMS M€ ossbauer spectrum, 3, 59, 584 CEMS 2010 spectrometer, 386 Charge transfer phase transition (CTPT), 153, 161 photoinduced, 164–168 thermally induced, 161–164 Chemical oxygen demand (COD), 595–598, 600, 606 artificial drain, 597 redox titration method, 597 Chemical shift, 44 Clebsch–Gordan coefficients, 380 CMR. see Colossal magnetoresistance (CMR) Coal ash recycling, iron silicate glass prepration, 596 uses, 596 Coal combustion, 584 Coal oxidation, 579 Coal samples, room-temperature M€ ossbauer parameters of, 580 Coals research coal combustion, 584–587 experimental setup, 577 gasification, 587–590 iron/sulfur, speciation, 576 mild steel, corrosion of, 583–584 M€ ossbauer spectrometer, 577 resuts, 578, 581 sample preparation, 577–578 schematic representation, 577 SEM analyses, 577 weathering of, 578–583 XRD analyses, 577 57 Co-cobalt(II)-doped GS emission M€ ossbauer spectra, 342 COD. see Chemical oxygen demand (COD) 57 Co-Doped Y1-xPrxBa2Cu3O7-d, studies, 397–401 57 Co emission studies, 398, 399 emission M€ ossbauer spectra, Y0.9Pr0.1Ba2Cu3(57CO)O7-d, 398 Cu3dO2p hybrid band, 397 M€ ossbauer parameters, room temperature, 400 praseodymium (Pr)

INDEX

effects, 397 ionic properties, 397, 398 PrBCO/YBCO lattice, 398 57 Co emission M€ ossbauer spectra, 335 measurements, 334, 336, 340, 344 spectroscopic data, parameters, 340 Coherent quasielastic nuclear resonant scattering, 25–29 57 CoII-doped cyanobacteria, 336 57 Co ion implantation, 60 57 Co-labeled cobalamins, 334 57 Co-labeled vitamin B12 coenzyme, 334 Colossal magnetoresistance (CMR), 393, 407–413 analogous cobalt-based perovskites, La1-xSrxCoO3 (See La0.8Sr0.2CoO3–d) cobaltate perovskites, 407 iron-doped La0.8Sr0.2FeyCo1-yO3-d perovskites, 411 (See also Irondoped La0.8Sr0.2FeyCo1-yO3-d) La0.7Ba0.3MnO3 perovskites, 407 manganate perovskites, 407, 408 materials, 393 Nd0.5Pb0.5MnO3 perovskites, 407 59 Co NMR spectroscopy, 410 Controlled radiative gamma decay, 293 within the framework of two-level approximation at excited nucleus consideration, 293 Hamiltonian operator, 293 nonstationary Schr€ odinger equation, 294 solutions of system, for cases, 294–295 wave function of whole system, 294 Controlled radioactive, and excited nuclei radiative gamma decay direct experimental observation, and study of process by delayed gamma–gamma coincidence method, 311–314 detecting of effect of controlling time of radiation nuclear decay, 311 influence of distributed resonance screen made of 119Sn, 314 optimization of decay controlling system parameters, 313 spontaneous decay of radioactive  Co(57 Fe) nucleus, 313 structure of controlled, and uncontrolled nuclear transitions, 312 Controlled spontaneous gamma decay excited nucleus in system of mutually uncorrelated modes of electromagnetic vacuum, 295–296 spontaneous gamma decay

in case of free space, 296–298 in case of nonresonant screen presence, 301–302 of excited nuclei in case of screen presence, 298 experimental study of phenomenon, by investigation of space anisotropy and self-focusing of M€ ossbauer radiation, 309–311 experimental study of phenomenon, for M€ ossbauer nuclei, 303–308 of nucleus in the case of resonant (M€ ossbauer) screen presence, 298–301 in system of synchronized modes of electromagnetic vacuum, 302–303 CONUSS program, 50 Conversion electron M€ ossbauer spectroscopy (CEMS), 447, 557, 577 Conversion x-ray M€ ossbauer spectroscopy (CXMS), 384 registration, 388 Coordination polymers of lanthanides (Ln), 116 Corrosion process, 418 Corrosion rate, of mild steel, 583 CoSn electrodes, in lithium test, 560 57 Co substitution cations, in metalloproteins, 345 Coulomb excitation, 58, 59 Coulomb excitation and recoil implantation M€ ossbauer effect (CRIME), 59 Coulomb interaction, 152 between d electrons, 152 Crystal field splitting energy, 47 Crystal field theory, 46 on 3d electronic states, 46–47 Cubic fluorite (F) structure of MO2, 74 Curie temperature, 521 Curie–Weiss law, 107, 173 Cyano-bridged coordination polymer complexes, 117 DCC model lattice parameter, 84–85 data analysis of Ce–Eu and Th–Eu, 85 DCC model, 86, 87 derivation, 85–88 model IV, 87–88 model extension attempt from macroscopic lattice parameter side, 89–92 quantitative BL(Eu3þ–O)-composition (y) curves, 88, 89

623

INDEX

DC magnetization, 435, 436 curve, 439, 440 magnetic moment, 435, 436 0D, 1D–3D network structure, 95 Debye velocity, 34 b-decay, 59 57 Mn, 60 Decomposition, of abundant mineral species, 584 Defect crystal chemistry (DCC) lattice parameter model, 76–79 Defect formation, 60 DF oxides, 74 Dicarboxylate anions, 116 Diffusion coefficients, 29 Diffusion mechanism, 59 Dilute Fe-doped YAG hyperfine fields, 526 M€ ossbauer analysis of, 526 Dilute magnetic semiconductors/ insulators (DMS/DMI), 521 magnetic phase diagram, 522 magnetic properties, 525–526 DISCVER (discrete versions of M€ ossbauer spectra) program, 516 Disordered phases, structural studies, 226 disordered chloride fluoride phases, 232–241 M€ ossbauer spectroscopy, and bonding, 234, 236 sample preparation, 232, 234 tin sublattice strength vs. solid solution composition, 236–241 x-ray diffraction, and double disorder, 232–236 disordered fluoride phases, 226–232 lone pair stereoactivity, and tin coordination, 226–230 positional disorder, and orientational disorder, 230–232 Dissolved oxygen (DO) levels, 595 DNA/RNA polymerases, 344 DOS program, 35 DPS model spectra, representation, 517 Dried biomass (DB), 337 161 Dy in dysprosium dicarboxylates, 116 absorption area and number of methylene groups, relationship between, 119 common bands in infrared spectra, 117 experimental methods, 117 M~ ossbauer spectroscopy, 116, 118 Nowick–Wickman relaxation model, simulation, 121 polymeric dimensional state, relaxation time criteria, 121

spectral parameters, 118 temperature dependence, 119–120 Dysprosium–transition metal complexes, 117 EDTA treatment, 340 Eigen-polarizations, 17 Einstein–Smoluchovski equation, 26 Electrical conductivity, 44, 603 Electrically conductive vanadate glass structural relaxation, 544 vanadate glass, electrically conductivity, 544 Electric field gradient (EFG), 12, 542 Electromagnetic isotope separation online (ISOL) technique, 59 Electron capture (EC) decay, 59 Electronic absorption, formula, 394 Electronic scattering length, 10 Electronic spin transition of Fe2þ in ferropericlase, 47–49 Electron M€ ossbauer spectroscopy, 458–460 contribution, 460 depth-selective conversion electron M€ ossbauer spectroscopy (DCEMS), 459 electron spectrum, 459 electrostatic spherical analyzer, 459 spectra, 459 electron energy analyzer, 459 increasing surface sensitivity, 458–460 integral low-energy electron M€ ossbauer spectroscopy (ILEEMS), 460 diagram schematic, 461 spectra recorded at different bias voltages from a 5 nm thick 57 Fe film, 463 spectrum 5nm thick 57Fe film, 462 vs. ILEEMS, 464 problems, 460, 462 study of a thin layer, 458–460 The Fe 2p XPS spectrum, 458, 459 study of a very thin layer, 458–460 deposition of pure 57Fe films, 458 Electrons appearance, in M€ ossbauer spectroscopy Auger electron spectroscopy (AES), 457 conversion coefficient, a, 456 electrons emission probabilities, deexcitation of a 57Fe Nucleus, 457 emitted electrons detection, 457 M€ ossbauer spectroscopy, 457 nuclear and atomic relaxation process, diagram, 457

physical basis, 456, 458 resonant nuclear absorption, 456 total electron conversion coefficients, 456 x-ray photoelectron spectroscopy (XPS), 457 Electron-withdrawiing effect, 400 Emission channeling, 59 Emission M€ ossbauer spectroscopic data, 338, 339 Emission (57Co) M€ ossbauer spectroscopy, 333 biochemical/physiological functions, 334 enzymological applications, 340–345 cations in metalloproteins, 345 EMS studies, 342–344 test object, choosing, 340–342 two-metal-ion catalysis, 344–345 isotope-containing molecules, 333 methodology, 334–335 microbiological applications, 336–340 radionuclide, 333 vitamin B12 coenzyme, 334 Energy-dispersive x-ray analysis (EDX), 535 Ethylene glycol, in Ar atmosphere, 564 European Synchrotron Radiation Facility (ESRF), 4 Europium(III) dicarboxylates, 116 EXAFS analysis, 523 Experimental low-energy electron M€ ossbauer spectroscopy, 460–165 bias voltage, 462, 463 intensity of surface component relation, 463 channeltron, electron detector, 460, 461 channeltron detection efficiency curve, 461 dependence of number of counts recorded per unit of time, 463 important parameter, 462 electrostatic energy analyzers, 460 transmission, ICEMS, and ILEEMS spectra acicular g Fe2O3 nanoparticles, 464 Fe0.33NbTiP3O12, 464, 465 FactSage modeling, 589 under constant partial oxygen pressures pyrrhotite, 590 diagram, 590 FARG room-temperature M€ ossbauer spectrum, 601, 602, 603, 604, 605

624

Fe-As-based high-temperature superconductors experimental procedure, 535–537 injected electrons, 537–538 Ni-doped single crystals new electron-rich species, 538–539 O2, adsorbtion, 539–541 Fe–As bonding, p–d hybridization, 538 Fe electrodes, magnetite nanoparticles, 489 Fe EXAFS spectra, 524 FeFenNi6-n environments, in LiFexNi1-xO2 solids, 554 Fe(III)–salt solutions, hydrolysis reactions, 472 Fe ion, artificial drain, 600 FeIr alloy CO oxidation, exposure, 574 57 Fe M€ ossbauer parameters H2 volume fractions, 573 Ir–Fe/SiO2 catalyst, 572 quasi in situ, 570, 571 57 Fe M€ ossbauer spectroscopy technique ammonium ferrocyanide, thermal decomposition, 371–374 decomposition BaFeO4 in dry/humid air, 515–516 iron(V) in air, 513–514 K2FeO4 in humid air, 514–515 of Na4FeO4 in air, 513 ferrates formation Fe2O3/Na2O2, solid-state reaction, 516–517 ferrates(IV, V, and VI), 510–513 ferryl(IV) ion, 508–510 iron oxide, 470 nanocrystalline Fe2O3 catalyst, 376–377 nanocrystalline iron oxides consecutive processes, 383–388 in-field transmission, 379–381 in situ high-temperature, 381–383 at various temperatures, 378–379 solid-state syntheses, 371 thermal conversion Fe2(SO4)3 in air, 376 thermal decomposition K2FeO4 in static air, 516 prussian blue in air, 374–376 use of, 371–377 Fe(NCX)2(bpa)2 (X¼S, Se) assembled structures, 144 powder x-ray diffraction study, 145 structural change by desorption of propanol molecules, 144–145 time dependence of 57Fe M€ ossbauer spectra, 145 Fe(NCX)2(bpp)2
2(benzene), 149 57 Fe M€ ossbauer spectrum, 149–150

INDEX

powder x-ray diffraction pattern, 149 reversible spin-state switching involving structural changes, triggered by sorption of benzene molecules, 149–150 spin-crossover phenomenon, 150 Fe(Ni) alloy particle, microwave reduction, 616 Fe–Ni production, in E/F vs. calcine reduction degree, 612 Fe(NO3)3 solutions characterization of, 475 hydrolysis/precipitation, 473 particles precipitation, TEM images, 476 Fenton method, 596 Fe2O3 AFM images, 375 characteristic ordering temperature, 370 hysteresis loops, 369 individual polymorphs, stability of, 368 iron(II) oxalate dihydrate, 376 modeled typical zero-field roomtemperature, 369, 370 M€ ossbauer hyperfine parameters, 369 nanoparticles, 369 polymorphs, room-temperature m€ ossbauer hyperfine parameters, 357 a-Fe2O3, 353–358 FE-SEM images of, 482 hydrolysis, 478, 479 modeled M€ ossbauer spectra, 358 monocrystalline, 357 particle size distribution, 356 phosphorus, incorporation of, 483 room temperature, 355 schematic representation of, 354 magnetic phase diagram of, 356 shape and size, 479 spin-flop field characteristic, 357 temperature, effect of, 481 e-Fe2O3 iron(III) oxide, brown magnetic phase, 364 magnetization, temperature dependence, 366 room temperature, 367 schematic representation of, 365 g-Fe2O3 low-temperature, 375 magnetic behavior, 361 M€ ossbauer spectra, 363 schematic representation, 360, 361 zero-field M€ ossbauer spectra, 362 FeOCl film, 386

b-FeOOH characteristic XRD patterns, 481 FeCl3 hydrolysis, 474 FE-SEM images of, 482 XRD lines, 476 g-FeOOH particles iron oxyhydroxide–surfactant, 485 lamellar microstructure, 476 precipitation process, 485 in situ nucleation, 486 a-FeOOH samples Co2þ ions, 492 FE-SEM images, 491, 492 [Fe(pyrazine){Pt(CN)4}] complex, 160 photoinduced spin conversion, between LS and HS states, 160–161 spin-crossover transition, thermally induced, 160 Ferrates Fe-O bond distances/electron density on iron, 513 isomer shift values of, 512 oxidation state (OS) of iron, 512 parameters, 510 spectroscopic characterization, 506–508 Ferrate(VI) solution, XANES spectrum, 507 Ferrimagnetic phase transition, 173 Ferritin, 327 H/L-ferritin chains, 327 M€ ossbauer spectroscopy, 328 properties, 327 Ferromagnetic clusters, 411 Ferryl(IV), 509 Ferryl species, formation of, 505 Fe/SiO2 catalyst, 57Fe M€ ossbauer parameters, 570 Fe2(SO4)3
2H2O isothermal decomposition, 377 ossbauer Fe2(SO4)3 room-temperature M€ spectrum, 376 Field cooling (FC) modes, 525 Flow temperature (FT) value, 586 Fly ash, M€ ossbauer spectrum, 585 Focusing monochromator (FM), 6 Free energy, 153 Galy–Andersson’s model, 209 Gamma decay, 292 controlled spontaneous, problems of, 292 models describing process of spontaneous decay, 292 limitations and solutions, 293 Gasification–fusion furnaces, 595 Gasification process, optimization, 588

625

INDEX

Gasifier graphical representation of, 588 Glu amino acid residues, 344 Glutamic acid (Glu) residues, 341, 342 Glutamine synthetase (GS), 334, 343 Goethite (a FeOOH), 418–420, 421, 423, 426, 490–495 collinear antiferromagnet, 419 crystallographic structure, 420 divalent cations, influence of, 491–493 formation, 494 hyperfine parameters, 417, 418 magnetic properties, 419 nonstoichiometric, 419 platinum group metal cations, influence of, 494–495 room temperature M€ ossbauer spectrum, 419 stoichiometric, 419 tetravalent cations, influence of, 494 trivalent cations, influence of, 493–494 Grazing incidence scattering geometry, 34 Healthy and diseased brain tissue, asymmetry of M€ ossbauer spectra, 330–331 Healthy brain tissue, M€ ossbauer studies, 325–327 Heavy fermion superconductors, 127 magnetic ordering and paramagnetic relaxation in, 129–133 Heisenberg uncertainty principle, 26 Hematite (Fe2O3), 585 Eu-for-Fe substitution, 495 oxidizing conditions, 589 Heme (iron protoporphyrin IX) proteins, 315 characterization of intermediate states, and chemical significances, 316 chemical stability and other advantages, 316 store two oxidizing equivalents, nature’s strategies, 316 Hemosiderin M€ ossbauer spectroscopy, 328 properties, 327 Hexamethylenetetramine (HMTA), 473 High-pressure diamond anvil cell (DAC), 44 High-resolution monochromators (HRMs), 4 High-resolution neutron diffraction studies Cu(1)—Cu(1) distance, 400 High-Tc superconductors, 393–401 57 Co-doped Y1-xPrxBa2Cu3O7-d (See 57Co-doped Y1-xPrxBa2Cu3O7-d)

highest critical temperature, 394 M€ ossbauer probe, 397 out-of-the-chain vibrations, 400 studies, 394–401 YBa2Cu3O7-d (YBCO) (See YBa2Cu3O7-d (YBCO)) High-valent Fe intermediates., 315 HS–LS transition, 152 Hund’s rule, 152, 175 Hydrolytical products, FT-IR spectra, 477 Hyperfine magnetic field, 380 ICEMS. see Integral conversion electron M€ ossbauer spectroscopy (ICEMS) Imidazole, 177 In-beam M€ ossbauer spectroscopy (IBMS), 58 INES program, 35 In-field M€ ossbauer spectroscopy, 436 spectra, nanosized Ni0.25Co0.25Zn0.5Fe2O4, 436 Infrared (IR) spectra of Eu malonate, 116 in situ M€ ossbauer spectroscopy, 382 Instrumentation, 4 detectors, 4 isotopes, accessible at ID18, 5 monochromators, 4–5 nuclear resonance beamline, 5–6 radioactive source, 4 UHV system for in situ nuclear resonant scattering, 6–9 Integral conversion electron M€ ossbauer spectroscopy (ICEMS), 385, 456–458, 462–465, 467 spectrum from a 5nm 57Fe film deposited by evaporation on a Si wafer, 458 Intergovernmental Panel on Climate Change (IPCC), 595 Intermediate-spin complexes, pure, 178 core modification, 184–189 ruffled deformation, 182–184 saddled deformation, 178–182 Ir–Fe bimetal catalysts, 564 Ir–Fe/SiO2 catalyst, PROX reaction, 573, 574 Iron, 43, 44 concentrations, 597 in substantia nigra, 328–329 experimental M€ ossbauer spectroscopic results, 53–54 labile, 328, 330 possible role in neurodegeneration, 331 preliminary M€ ossbauer studies, 329 concentrations in control and PSP tissues, 329

present in healthy and diseased brain tissue, 328 reduced laterite samples, phase content of, 615 simplified decay scheme, 59 spin and valence states, in silicate postperovskite, 52–54 spin transition in ferropericlase, 55 states in both perovskite and postperovskite, 55 total spin momentum of 3d electronics, 55 Iron-containing soda-lime silicate (ISLS) glass, 596 COD values, 598 dissolution rate of cations, 596 network modifier (NWM), 596 room-temperature M€ ossbauer spectra of, 596, 597 Iron-doped La0.8Sr0.2FeyCo1-yO3-d, 411–413 case of y0.05 iron content, 411, 412 case of y0.15 iron content, 412, 413 emission and transmission M€ ossbauer study, 411–413 Iron-doped YAG microdischarge treatment, 528–531 Iron(III)–hydroxy complexes formation of, 470 Iron(III) oxide, polymorphs, 353, 381 amorphous Fe2O3, 369–371 crystal structures, magnetic properties, 352 a-Fe2O3, 353–358 b-Fe2O3, 358–359 e-Fe2O3, 364–368 g-Fe2O3, 360–364 Iron(III) porphyrin complexes axial ligands for, 177 deformation modes, 178 molecular structures of Fe(Por)I, 186 M€ ossbauer and EPR spectra, 187 properties, 177 spin-crossover triangle in, 195–198 spin state, controll of, 177 Iron oxide, precipitation crystallographic systems, 471 dense b-FeOOH suspensions, 480–483 57 Fe M€ ossbauer spectroscopy technique, 470 hydrolysis reactions, 470–480 hyperfine magnetic fields (HMFs), 474 influence of cations goethite, 490–495 hematite, 495–496 maghemite, 496 magnetite, 496

626

Iron oxide, precipitation (Continued ) properties of ferrihydrite, 483–485 lepidocrocite (g-FeOOH), 485–487 maghemite (g-Fe2O3), 487–490 magnetite (Fe3O4), 487 Iron oxides classification of, 352 compounds, oxidation states, 506 Iron silicate glass prepared by recycling coal ash, 596 Iron-tetraamidomacrocyclic ligand (Fe-TAML) catalysts, 505 Iron(V) nitride complex, 511 Isomer shift (IS), 537 ISO 7404-5 tandard method, 588 Isothermal remanence magnetization (IRM) curve, 439, 440 dM plots, 439, 440 Henkel plot, 440 Jahn–Teller distortions, 52 Jarosite, 579 Jump function, 26 K2FeO4 aging, kinetics, 384 decomposition product, 516 electron paramagnetic resonance (EPR) measurements of, 507 thermal decomposition, 516 Kr€ oger–Vink notation, 75 Kynurenic acid, 316 Lamb–M€ ossbauer factor, 3, 12, 26 Landau mine, 583 Langmuir–Hinshelwood reaction, 572 La0.8Sr0.2CoO3-d area ratios between doublet and sextet components La0.8Sr0.2Co1–yFeyO3–d perovskites, 409 57 Co emission M€ ossbauer spectra after oxygen removal, 408, 409 emission m€ ossbauer study, 408–411 isomer shifts of doublet and sextet of La0.8Sr0.2Co1-yFeyO3-d perovskites, 410 M€ ossbauer isomer shifts of sextets (dsextet) and doublets (ddoublet), 410 Laterite–lignite mixture dielectric constant of, 616 DTA data of, 618 infrared images of, 617 logarithmic scale diagram, 618 microwave heating of, 618

INDEX

microwave power supply, 617, 618 real and imaginary permittivities, 619 Laterite mineral processing, 608 conventional processing, 610–612 microwave processing, 613–619 nickel production, from lateritic/sulfidic ores, 608 pyrometallurgical processing, 609 thermal effect, 610 LIB. see Li-ion batteries (LIB) LIESST. see Light-induced excited spinstate trapping (LIESST) LiFePO4 lithium deinsertion/ insertion, 555 LiFeVPOx glass, 548 electrical conductivity, 549 Light-induced excited spinstate trapping (LIESST), 153 for Fe(II) complexes, 153–157 for Fe(III) complexes, 157–159 highest critical temperature, 160 Li-ion batteries (LIB) anode materials alloys/intermetallic compounds, 558–560 antimony alloys, 560–561 conversion oxides, 556–558 cathode active material, 543–544, 547–550 cathode materials, 543, 554 insertion silicate electrodes, 555–556 layered intercalation electrodes, 554 phosphate electrodes, with olivine structure, 554–555 discharge/charge capacity, 549, 550 electrochemistry of, 549 isotopes, 553 of mobile phone, 543 physicochemical behavior, 550 room-temperature M€ ossbauer spectra of, 550 Linear polarization, 11 Lipkin’s sum rules, 35 Liquid nitrogen temperature (LNT), 475, 492, 493 Li3Sb signal, 561 Li–Sn intermetallics, 558 isomer shift, 559 Lithium-ion battery, nucleus/ transition, 553 Ln complexes, of dicarboxylates, 116 Ln3þ/M4þ ionic radii ratio (x0026A;x00070;x00070;), 75 Lns-M€ ossbauer and lattice parameter data of DF oxides, 79 151 Eu-M€ ossbauer, and lattice parameter data, 79

DF-type Ce–Eu, 80 F-type U–Eu, 79–80 Hf–Eu, 80 P-type stabilized Zr–Eu, 80 Th–Eu, 80 155 Gd-M€ ossbauer, and lattice parameter data, 80–84 Lns-M€ ossbauer data analysis, 84–85 Lone pair stereoactivity, and material properties, 241–242 Low-energy electron diffraction (LEED), 6 Low-energy electrons technique, 455 Magnesium silicate perovskite, 49 Magnetic anisotropy, 362 Magnetic cluster model, 410 Magnetic dipole hyperfine interactions, 17 Magnetic force microscopy (MFM), 429 Magnetic susceptibility, and saturation moment, for ground doublet, 107 averaged powder moment, 110–113 calculated paramagnetic relaxation spectra of Np for different fluctuation rates, 109 comparison of energy-level, 110 free ion model, 108–109 ising-type magnetic moment, 109–111 comparison of experimental and theoretical T plots, 112 experimental and theoretical magnetization curves, 112 for O¼Np¼O axes collinear crystals, 112 with/without high-temperature approximation, 111–112 polycrystalline samples fast in STYCAST, 107 Magnetite/maghemite, 353, 420, 421 formation of Co2þ ion concentration, 496 Ni2þ ion concentration, 496 Zn2þ ion concentration, 496 hyperfine parameters, 417, 418 magnetic properties, 420 nanoparticles, 489 nonstoichiometric, 420 vs, stoichiometric, 420 stoichiometric, 420 superparamagnetic magnetite particles, 420 three sextets, 420 Verwey transition, 420, 421

627

INDEX

Magnetization, 11, 18, 24, 25, 35, 101, 104, 111, 113, 173, 367, 435, 445, 465, 531, 556 Manganese, 59 Markovian diffusion, 26 Metal-inorganic-transport (MIT) family, 336 Metallicity, 395, 396 Microdischarge between carbon felts (MD/CF), 522 Microenvironment, 345 Microwave processing bulk hematitic laterite ore x-ray pattern of, 614 drying rates, 613 heat generation, 612 hematitic nickeliferous laterite ore, 613 XRF chemical analysis, 613 laterite mineral processing, 613–619 TM0n0 cavity system, 614 Minerals pyrite (FeS2), 576 57 Mn beam characteristics, 60 57 Mn decay, 59 57 Mn (!57Fe) implantation M€ ossbauer spectroscopy, 61 application to inorganic chemistry, 63–65 exotic localized Fe molecules, 64–65 unusual high Fe oxidation states, 63 application to materials science, 62 detector for 14.4 keV M€ ossbauer g-rays, 62 g-ray detector, development of, 65–66 in-beam M€ ossbaue spectrometer, 61, 62 Morin transition, 354, 355, 495 temperature, 356 M€ ossbauer effect, 3, 58, 60, 133, 378, 552, 561 M€ ossbauer g-ray detector, development, 65, 66 M€ ossbauer–Lamb factor, 378 M€ ossbauer parameters, isomer shift, 339 M€ ossbauer surface imaging techniques, 465, 466 absorption, 465, 466 application, 465 coss section, M€ ossbauer microscope, 466 magnetic resonance imaging (MRI), 465 multicapillary x-ray lens, 466 operating principle, 465 working, 465, 466

M€ ossbauer transitions, in actinide isotopes, 124 Naþ concentrations, 597 Nafion, 471 Nanocrystalline CoSn, 558 Nanoregime, properties, 429 Nanosized Al-1 at % Fe, 442–446 fractional integrated intensity of (111) peak, 443 hyperfine field distribution of M€ ossbauer spectra., 442, 443 M–H curves at room temperature, 445 M€ ossbauer parameters, 444 properties, 442 quadrupole splitting, 444 room-temperature M€ ossbauer spectra, 442, 443 self-annealing effect, 444 variation of bulk magnetic parameters, with grain size, 446 variation of integrated P(H) with milling time, for various field component, 444 Nanosized Fe-Al alloys coercivity HC, 446 hysteretic properties, ferromagnetic material, 446 calculation, 446 Nanosized Fe–Al alloys, 441–446 ball-milled samples, 445 bcc Fe grains, 445 bulk Fe-Al systems, 442 DC magnetization measurements, 445 intermetallic Fe-Al systems, 441 mechanical alloying (MA), alloy sysnthesis, 442 M€ ossbauer studies results, 445 nanocrystalline Fe-Al systems, 442 nanosized Al–1 at % Fe, 442–446 (See Nanosized Al-1 at% Fe) unmilled samples, 445 Nanotechnology, 3 Neel P-type ferrimagnet, 366 Neel’s model, 438 Neel temperature, 416, 418, 439, 494 Neptunyl(þ1) complexes, 95 M€ ossbauer and magnetic study, 98–106 (NH4)[NpO2(O2CH)2], 98–100 [(NpO2)2((O2C)2C6H4)(H2O)3]
H2O, 104–106 [NpO2(O2CCH2OH)(H2O)], 100–101 [NpO2(O2CH)(H2O)], 101–104 Neptunyl monocation (NpO2þ) magnetic property, 97–98 Network former (NWF), 600 Network modifier (NWM), 600 Neurodegeneration, 324–325

Neutron in-beam M€ ossbauer spectroscopy, 66 Nickel production from lateritic/sulfidic ores, 608 Ni-ferrihydrite, 492 N-methyl-D-aspartic acid (NMDA) receptors, 316 Nonmagnetic materials, 12 237 Np M€ ossbauer relaxation spectra, 106–107 conditions for ferro- and paramagnetic states, 106 Kramers ion of lanthanide elements, 107 Wickman’s relaxation model illustration, 107 Zeeman pattern, 106 237 Np M€ ossbauer spectroscopy, 96–97 NTA (nanotechnology assorted) glassTM, 543, 544, 546, 547 Nuclear forward scattering (NFS), 249 Nuclear inelastic interaction, 30 Nuclear inelastic scattering, 4, 6 experiment, 33 Nuclear inelastic spectrum, 33 Nuclear resonance beamline ID18, 5–6 Nuclear resonant scattering (NRS), 3, 4, 13, 23, 28, 30, 225, 256, 269 delayed, angular dependence, 22 experimental geometry used in, 12 intensity in forward direction, 26 time spectra, 19, 20, 27 Nuclear transitions, M€ ossbauer spectroscopy (MS), 552 Off-placement, O(4) atoms, 399, 400 Orange II dye, 505 Orientation-dependent time-integrated intensity (ODIN), 26 Original coal samples, proximate analyses, 578 ORP pH, changes, 598 Orthorhombic distortion, 202 Orthorhombic europium orthoferrite (EuFeO3), 496 Oxalato complex, 116 Oxidative stress, 324–325, 329 Oxide glass, 542 Oxygen vacancies, 529 Parallel-plate avalanche, 61 a-PbSnF4, structural determination historical perspectives, 216–217 M€ ossbauer spectroscopy, 220, 224 neutron powder diffraction, 225 sample orientation, 223 sp3d2 hybridization, and VSEPR rules, 225

628

a-PbSnF4, structural determination (Continued ) tetragonal symmetry, 225 x-ray powder diffraction, 221, 222 Perovskite structure, 52, 407 binary oxides, 393 silicate, 43 strontium ferrate (SrFeO3), 401, 402 x-ray diffraction, 523 Phase transitions, 203 PHOENIX program, 35 Phonon density, 4 Phonon DOS of Fe films, 36 PHONON program, 37 Phosphate glass, 542 Photoinduced spin-crossover phenomena, 153 LIESST for Fe(II) complexes, 153–157 LIESST for Fe(III) complexes, 157–159 Photoinduced spin transition, 153 Plastic scintillation counters, 61 Polarization, 17 Porphyrin analogues, 179–180 Potassium ferrate(VI) sample in situ measurement, 515 Preferential CO oxidation, in H2 catalyst preparation, 564–565 catalytic activity test, 565 Ir–Fe/SiO2 catalyst, 567–568 calcined catalyst, 569 57 Fe M€ ossbauer parameters, 570 H2 concentration, effect of, 568–569 M€ ossbauer spectra characterization, 565 PtFe alloy nanoparticles catalyst, 565–567 PROX reaction Ir–Fe/SiO2 catalyst, 573, 574 Prussian blue (PB) formation of, 372 hyperfine parameters, 374 Pyridine, 177 Pyrite, 579 to ferrous sulfate, conversion, 582 Pyrrhotite (Fe1.25S), 585 Quadrupole splitting (QS), 44, 75, 228, 261, 273, 339, 395, 484, 537, 542, 578 Quinine hydrogen sulfate (QHS), 475 Radiative decay, 15, 16, 20, 292 Raman spectroscopy, 160 Rapid spin equilibrium in solid state, 168–172 g-Rays, 3, 58, 59 Recycled iron silicate glass water-purifying ability of, 595–596

INDEX

Recycled silicate glasses characterization of, 596 electromagnetic property of, 601–605 Fe2O3 contents, 599 water-purifying ability of, 596–600 Relative absorption area (RAA), 75 Resonant scattering length, 11 Reverse-LIESST, 156 Rh3þ ions iron oxide samples synthesis, 495 RI beam irradiation, 60 Room temperature M€ ossbauer spectrometry (RT-MS) Akaganeite, 416, 417 Goethite, 418 magnetite/maghemite, 420 Rust layers, 421–426 a calculation, formula, 421 classification, 421 iron oxides and oxyhydroxides, quantitative characterization, 421 M€ ossbauer fraction fj, 421 relative phase abundances, rj, 421 steels, 421–426 (See also Steels) ternary rust diagram, 421 Rutherford backscattering, 59 Shannon’s ionic radii, 75 Silicate glass, 542 recycling (See Recycled silicate glass) Silicate postperovskite spin and valence states of iron in, 52–54 Silicon photomicrograph of, 586 SEM element mapping, 585 Single-molecule magnet (SMM), 113 a-SnF2, structural determination, 213 historical perspectives, 213–214 work of Bergerhoff, 214–216 Sodium ferrate decomposition products, 517 Sodium ferrate, DPS model, spectra/ representation, 517 Sodium-ion battery (SIB), 544 Soil diazotrophic rhizobacterium A. brasilense, 336 Sol–gel method sample preparations, 523 Y3Fe5O12 (YIG), 522 Spin-crossover phenomenon, 143, 153 in assembled complexes Fe (NCX)2(bpa)2 (X¼S, Se, BH3) by enclathrating guest molecules, 145–147 coordination angle of anion, 147 2D grid structure, 146 57 Fe M~ ossbauer spectra, 146

transition temperatures in 1D structure, 147 x-ray structural analysis, 145 contribution of S¼3/2 in Fe (OETPP), 180 data of five-coordinate Fe carrying anionic axial ligand, 180 involving intermediate-spin state, 189 spin crossover between S¼3/2 and S¼1/2, 189–191 spin crossover between S¼3/2 and S¼5/2, 192–195 M€ ossbauer spectrum of [Fe (OETPP), 181 ortho/meta-H signals, movements, 180 photoinduced, for [Fe(pyrazine) {Pt(CN)4}](pyrazine¼ C4H4N2), 153 temperature dependence of effective magnetic moments, 181 Spinel ferrites, 230–241 Cr-substituted Co–Zn ferrite (CrxCo0.5-xZn0.5Fe2O4), 430 cation distributions, 431 tetrahedral A sites, 430, 431 tetrahedral B sites, 430, 431 intersublattice interactions JAB, 430 intrasublattice interactions JAA and JBB, 430 microstructure determination, 430–433 M€ ossbauer technique, 430 spectra of Cr0.25Co0.25Zn0.5Fe2O4, with/without external field of 5T, 434 structural parameters calculation, 430 variations of hyperfine fields at A and B sites in CrxCo0.5-xZn0.5Fe2O4, 431 nanoferrites, elucidation of bulk magnetic properties using in-field M€ ossbauer spectroscopy, 434–435 ball-milled Ni0.35Zn0.65Fe2O4, 435, 436 canting angle uB, calculation, 435 example, impurity-free Cr0.25Co0.25Zn0.5Fe2O4, 434, 436 interactions among sublattice in samples, 435 low-temperature (LT) spectrum, 434 M€ ossbauer measurements, 435 real-time (RT) spectrum, 434, 435 nanoferrites synthesis, coprecipitation method, 431

629

INDEX

average paritcle size (D), Scherrer formula, 432 a Fe2O3 impurity, 431, 432 particle size distribution, 433 room-temperature M€ ossbauer spectrum, 433 TEM micrographs, 433 XRD spectra of Cr0.25Co0.25Zn0.5Fe2O4, 432 superparamagnetic nanosystems, effect on magnetic properties, 236–241 antiferromagnetic (AFM), 339 canting effects, 436, 438 coercivity, calculation, 439 core-shell effect, 436, 437 dilute system vs. real system interparticle interactions, 236–241 exchange bias field, calculation, 439 ferromagnetic (FM), 439 HC, coercive field plot, 440, 441 HE, exchange bias field plot, 440, 441 low-temperature M€ ossbauer spectra, 437 M-H loops, 438, 439 parameters calculation, formula, 438 parameters, M€ ossbauer Fitting Using Core-Shell Model, 437 remanence, function of coling field, 441 room-temperature M€ ossbauer spectra, 43 single-phased nanosized particles, 436 ZFC curve, blocking temperature TB, 436, 438 Spin equilibrium, and valence fluctuation antiferromagnetic interaction between HS state, 175 (n-C4H9)4N[FeIIFeIII(mto)3], 173 concerted phenomenon coupled with, 173 effective magnetic moment, 173, 175 57 Fe M€ ossbauer spectra, 173, 174 ferrimagnetic phase transition, 175 inverse magnetic susceptibility, 173 schematic feature of rapid spin equilibrium in FeIIIO3S3 site, 175 Spin-pairing transition, 51 Spin-state isomers, 177 SRPAC. see Synchrotron radiation-based perturbed angular correlations (SRPAC) Stannous fluoride SnF2, 202, 203 lead-containing phases, 203, 204

Steels Colombian carbon steels (CS), 421–424 composition, 421 77K M€ ossbauer spectra total immersion tests, 423 M€ ossbauer spectra, room temperature Dry-Wet cycles, 425 total immersion tests, 422 Colombian weathering steels (WS), 421–423 composition, 421 77K M€ ossbauer spectra hit rusts, Dry-Wet cycles, 425 scraped rust. Wey-Dry cycles, 425 total immersion tests, 423 M€ ossbauer spectra, room temperature dry-wet cycles, 425 total immersion tests, 422 degradation process, chloride ions, 424 dry-wet cycles, 424–426 Japanese weathering steels (WS), 422–423 composition, 422 Lorentzian profiles, 422, 423, 425 mechanisms of deterioration, kimura et al., 424 metal-doped Fe(OH)2, 424 outdoor tests, 426 total immersion tests, 421–424 adherent rust, 422 Strontium ferrate (SrFeO3), 401–407 properties, 401 Sr0.95Ca0.05Co0.5Fe0.5O3-d vs. Sr0.5Ca0.5Co0.5Fe0.5O3-d, 401, 402 Sr0.5Ca0.5Co0.5Fe0.5O3-d , study, 405–407 brownmillerite-structured antiferromagnetic CaFeO2.5, 405 CO2 absorption, 407 EMS spectra, 406 microheterogeneity, 406 M€ ossbauer Parameters of the two major sextets, 406 TMS spectra, 406 Sr0.95Ca0.05Co0.5Fe0.5O3-d, study, 401–405 assumptions on the basis of TMS and EMS experiments, 404 carbon dioxide absorption experiments, M€ ossbauer spectra, 405 CO2 absorption spectra, 405 3D alternating models, 404

difference in TMS and EMS spectra, 403, 405 differing oxygen ligand environments for Fe and Co in the lattice, 402 EMS spectra, 402–404 long-range magnetic order, 404 nearest neighbor (NN) cation distribution, normal (TMS) and nucleogenic 57 Fe (EMS), 403 orthorhombic microdomains, 404 perovskite lattice, 402, 404 phase-separated models, 404 quadrupole splittings difference, 403, 404 relative number of NN Fe and Co, 404 synthesis, 402 TMS spectra, 402–404 x-ray diffractometery studies, 402 SrFeO2.5 brownmillerite structure, 404 Succinato (suc) complex, 116 Suface M€ ossbauer studies biomedical application, 466, 467 characterization, 467 discovery, 466 Fe3O4, an ancient material, 466–468 ICEMS invetigation, 467 ICEMS spectra ILEEMS investigation, 467 ILEEMS surface sensitivity vs. ICEMS surface sensitivity, 467, 468 magnmetic film application, 467 Superconducting sheets, 401 Superconductor isomer shifts, 400, 401 Super iron battery, 506 Surface analysis, 455 Surface analytical technique, requirements, 455 Synchrotron M€ ossbauer spectroscopy (SMS), 44 extreme P–T conditions, 44 simulated time spectra, 46 spectra evaluation, 44–46 Synchrotron radiation-based M€ ossbauer techniques, 4, 10 brief theoretical background, 10–12 coherent elastic nuclear resonant scattering, 10 coherent quasielastic nuclear resonant scattering, 25–30 incoherent inelastic nuclear resonant scattering, 30–38 nuclear resonant scattering in grazing incidence geometry, 20–24 time spectra of nuclear resonant scattering, 12–20

630

Synchrotron radiation-based perturbed angular correlations (SRPAC), 250 applications in bioinorganic chemistry, 258 57 Fe NFS of oxidized Pf D14C Fd, 261–262 57 Fe NFS of oxidized Rc FdVI, 260–261 57 Fe NFS on oxidized Fe protein, 262 57 Fe NFS on the resting state MoFe protein, 263–264 [4Fe-4S] ferredoxin from Pyrococcus furiosus, 259 [2Fe-2S] ferredoxin VI from Rhodobacter capsulatus, 258–259 nitrogenase, 259–260 nuclear forward scattering, 258 detectors and detection methods, 256–257 electric hyperfine interactions, 252, 254 experimental aspects of NFS and SRPAC, 255 experimental setup, general aspects, 255 experimental setup in NFS and SRPAC, differences, 257–258 magnetic hyperfine interactions, 250–252, 254–255 model complexes examined by, 264 57 Fe SRPAC of Fe2(S2C3H6)(CO)6, 264–265 57 Fe SRPAC of Ni ferrite, 265–267 61 Ni SRPAC of Ni ferrite, 267–268 monochromators, 255–256 sample requirements, 258 as scattering variant of TDPAC, 250 technical background, 250 theoretical aspects, 252–254 theoretical aspects of NFS, 250 Tanabe–Sugano diagram, 152 TDPACs. see Time differential perturbed angular correlations (TDPACs) Tetramethylammonium hydroxide (TMAH), 475 TG Labys-DSC system, 616 Thermal hysteresis, 152 Thermal stability, 60 Thin film multilayer systems (MLS), 446–453 DC magnetization studies, 448–450 experimental and fitted parameters of as-deposited and at hightemperature M–H Loops, 449

INDEX

experimental hysteresis loops, 450 hysteresis loops of as-deposited 57 Fe/Al MLS at different temperatures, 450 nonsquare loop, coercivity, 450 RT hysteresis loop, formula, 449 spin orientation, Fe film, 450 deposition process, 447 57 Fe/AI MLS, 446–453 M€ ossbauer (CEMS) study, 451, 452 annelaing conditions, 452 annelaing effects, 452 as-deposited MLS, 451, 452 close-lying Lorentzians, 451 complementary characterization techniques, 452 crosssectional TEM micrograph, 452 deposition conditions, 452 heat formation effects, 452 high-temperature study, 452 hyperfine fields, 451 magnetization studies, 451 room-temperature CEMS spectra of 57Fe/Al MLS, 451 ultrathin Fe/Al MLS, studies, 451 properties, 446, 447 structural characterization, 447, 448 electron density profile (EDP), 447 electron micrographs and SAD patterns of as-deposited and annealed Fe/Al MLS, 447, 449 ion beam sputtering (IBS) technique, 447 TEM micrograph evidences, 448 XRR measurement of 57Fe/Al MLS, 447, 448 XTEM investigations, direct observation, 447 Time differential perturbed angular correlations (TDPACs), 249, 250, 253, 269 Tin-containing alloys, barycenter position, 559 Tin electronic structure, and M€ ossbauer spectroscopy, 208 bonding type, and coordination, 208 chemical isomer shift d, 212 geometry agreement with VSEPR model, 209–211 quadrupole splitting, 212–213 Sn2þ ion, 208 spin system of 119Sn nuclide, 211–212 tin(II) covalently bonded, 208 Tin(II) fluoride SnF2, 202 covalent bonding, and polymeric structure, 205–207 crystal structures, 204

experimental aspects, for pure phases, 203 forming mixed fluorides with alkaline earth metal fluorides, 202 lead-containing phases transitions, 203 M€ ossbauer spectroscopy, 204 PbClF-type structure, 207–208 a-PbSnF4 structure, 207 phase characterization and elemental analysis, 204 projection of structures, 206 pure phases, 203 react with BaCl2
2H2O in aqueous solutions, precipitate function of, 203 sample preparation, 203 spectrometer equipped with helium closed-cycle refrigerator, 205 typically ionic structure, 204–205 Transition metal ion, 152 Transmission electron microscopy (TEM), 393, 429 vs. emission M€ ossbauer spectroscopy (EMS), 393, 394 Transmission 57Fe M€ ossbauer studies, 396 L?-Tryptophan oxidation by heme-based enzymes, 316–318 chemical reaction catalyzed by MauG, 318–319 high-valent BIS-Fe(IV) intermediate in MauG, 319–320 oxidation/oxygenation of free, and protein-bound Trp, 317 utilization of two c-type hemes by MauG to perform, 317 tryptophan 2,3-dioxygenase (TDO) ligand-bound structure, 317 Tryptophan 2,3-dioxygenase (TDO), 316 high-valent Fe intermediate, 320–321 ligand-bound structure, 317 Two-XRD-line ferrihydrite XRD pattern of, 485 Ube municipal garbage combustion plant, chemical analysis, 598 UHV system, 6–9 APD detector, 7, 8 Be windows, 7 distribution chamber, 7 Kirkpatrick–Baez focusing mirrors, 9 load-lock chamber, 8 low-energy electron diffraction, 6 manipulator, 6 NRS chamber, 8 nuclear inelastic scattering, 6 portable chambers, 8–9

631

INDEX

preparation chamber, 6 quartz-balance monitor, 6 sample holders, 8 sample storage chamber, 7 support table, 9 Ultrahigh vacuum (UHV), conditions, 4 Ultrasonic spray pyrolysis (USP), 386 238 U M€ ossbauer spectroscopy, 123 application for, 124 heavy fermion superconductors, 127 magnetism in uranium compounds, 125 decay scheme, 124 energy spectrum of 242PuO2 source for, 124–125 nuclear g-factor in excited state of 238U nuclei determination of, 125, 127 235 U NMR measurements of UO2 in antiferromagnetic state, 125–127 UNILAC accelerator, 59 Uranium-based heavy fermion superconductors, 127–129 observed hyperfine field, 133 temperature dependence of spectra, 134 238 U M€ ossbauer spectroscopy, 133–134 ossbauer Uranium dipnictides, 238U M€ spectroscopy, 134 application to two-dimensional (2D) fermi surface system, 134–137 characteristic physical properties, 134 crystal electric field (CEF) contributions, 136 de Haas–van Alphen effect, 134 hyperfine interactions correlated with, 135–136 isomer shift, 135 Neel/Debye temperatures, 136 nuclear quadrupole interactions, 136 orbital occupancy, 134 parameters, 136 plots of hyperfine coupling constant, at uranium nuclei, 137 spectra, 135 Uranium isotopes, 124 US20 glass measurement, M€ ossbauer spectroscopy (MS), 599

Valence shell electron pair repulsion (VSEPR) model, 209 Vanadate glass, 548 electrically conductivity, 544–547 Variable-temperature 57Fe M€ ossbauer spectrometry (VT-MS), 415, 416 advantage, 415 corrosion experiments, 424 physical properties of iron phases, 423 Vibrating sample magnetometer (VSM), 429 iron-doped yttrium aluminum oxides, 525 magnetization measurements, 525 Vibrating sample magnetometry (VSM), 429 Vibrational entropy, 38 VICKSI accelerator, 59 Vienna ab initio simulation package (VASP), 37 VO5 pyramids, 542 VO4 tetrahedra, 544 Waste materials, recycling, 596 Waste treatment plant, 598 Water-purifying minerals, 595 Wet chemical method, 489 W€ustite, 352 XPCS spectroscopy, 7, 9 X-ray absorption fine structure (XAFS), 74, 75, 77, 90, 567, 573 iron K-edge, 485 X-ray absorption near-edge spectroscopy (XANES), 49, 472, 506, 507, 555 X-ray diffraction (XRD), 7, 10, 44, 48, 52, 73, 75, 90, 128, 145, 147, 207, 218, 232, 404, 429, 442, 480, 484, 522, 523, 535, 561, 599, 613, 616 Rietveld refinement, 52 spectrum CrxCo0.5-xZn0.5Fe2O4, 432 X-ray emission spectroscopy (XES), 49–52, 55 X-ray photoelectron spectroscopy (XPS), 386, 429, 457, 458, 460, 464, 490

X-ray studies, for Cu(1)—Cu(1) distance, 400 YAG. Al(VI), schematic crystal structures, 524 YAG doped, M€ ossbauer analysis of, 526 Y3Al0.98Fe0.02O12 bulk magnetization, 531 Y3Al5-xFexO12 XRD patterns and lattice parameters, 523 YBa2Cu3O7-d (YBCO), 393, 395–401 B to C species interconversion, 396, 398–400 critical temperature, 394 emission M€ ossbauer spectrum, YBCO fully oxygenated state, 396 interconversion, species B and C, 396 major species, 395 M€ ossbauer parameters obtained, YBCO fully oxygenated state, 395 nucleogenic Fe—O three major species, structure, 397 straightening of the Cu(1)—O(4) chain, 400, 401 structure of unit cell, 395 thermodynamic parameters, as function of Pr content of Y1-xPrxBa2Cu3(57Co)O7-d, 399 vs. Fe-doped YBCO, 396 Y0.9Pr0.1Ba2Cu3(57CO)O7-d, 399 Y0.55Pr0.45Ba2Cu3(57CO)O7-d, 399 Y3Fe5O12 (YIG) magnetic garnet/transparent, 522 sol–gel method, 522 Yttria-stabilized zirconia (YSZ), 74 Yttrium EXAFS spectra ions, 524 radial structural functions, 524 Y3(YAl4-xFex)BO12 magnetic susceptibility, 526 magnetization, temperature dependence, 526 normalized iron EXAFS spectra, 525 Zero field cooling (ZFC), 525