advances in catalysis, volume 48

ADVANCES IN CATALYSIS

VOLUME 48

Advisory Board

M. CHE

Paris, FranceD.D. ELEY

Nottingham, EnglandG. ERTL

Berlin/Dahlem, Germany

V.B. KAZANSKY

Moscow, RussiaW.M.H. SACHTLER

Evanston, Illinois, USAR.A. VAN SANTEN

Eindhoven, The Netherlands

K. TAMARU

Tokyo, JapanJ.M. THOMAS

London/Cambridge, England

H. TOPSØE

Lyngby, DenmarkP.B. WEISZ

State College, Pennsylvania, USA

ADVANCES INCATALYSIS

VOLUME 48

Edited by

BRUCE C. GATES

University of California

Davis, California, USA

HELMUT KNOZINGER

University of Munich

Munich, Germany

Amsterdam · Boston · Heidelberg · London · New York · Oxford

Paris · San Diego · San Francisco · Singapore · Sydney · Tokyo

ELSEVIER B.V.

Sara Burgerhartstraat 25

P.O. Box 211, 1000 AE

Amsterdam, The Netherlands

ELSEVIER Inc.

525 B Street, Suite 1900

San Diego, CA 92101-4495

USA

ELSEVIER Ltd

The Boulevard, Langford Lane

Kidlington, Oxford OX5 1GB

UK

ELSEVIER Ltd

84 Theobalds Road

London WC1X 8RR

UK

q 2004 Elsevier Inc. All rights reserved.

This work is protected under copyright by Elsevier Inc., and the following terms and conditions apply to its use:

Photocopying

Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher

and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or

promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to

make photocopies for non-profit educational classroom use.

Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK: phone (+44) 1865 843830, fax (+44) 1865

853333, e-mail: [email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.else-

vier.com/locate/permissions).

In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive,

Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing

Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44)

20 7631 5500. Other countries may have a local reprographic rights agency for payments.

Derivative Works

Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or

distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and

translations.

Electronic Storage or Usage

Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of

a chapter.

Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any

means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher.

Address permissions requests to: Elsevier’s Rights Department, at the fax and e-mail addresses noted above.

Notice

No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability,

negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein.

Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be

made.

First edition 2004

Library of Congress Cataloging in Publication Data

A catalog record is available from the Library of Congress.

British Library Cataloguing in Publication Data

A catalogue record is available from the British Library.

ISBN: 0-12-007848-1

ISSN (Series): 0360-0564

W1 The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in

The Netherlands.

Contents

CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv

ROBERT L. BURWELL, Jr. (1912–2003) . . . . . . . . . . . . . . . . . . . . . . . . . xix

Active Sites and Reactive Intermediates in Titanium Silicate

Molecular Sieves

P. Ratnasamy, D. Srinivas and H. Knozinger

I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

II. Active Sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

II.A. State and Framework Coordination of Ti . . . . . . . . . . . . . . . . . 9

II.A.1. Diffraction Techniques . . . . . . . . . . . . . . . . . . . . . . . . 10

II.A.2. Influence of Particle Size . . . . . . . . . . . . . . . . . . . . . . . 12

II.A.3. UV–Visible Spectroscopy . . . . . . . . . . . . . . . . . . . . . . 12

II.A.4. Photoluminescence Spectroscopy . . . . . . . . . . . . . . . . . 15

II.A.5. X-Ray Absorption Spectroscopy . . . . . . . . . . . . . . . . . 15

II.A.6. Vibrational Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 18

II.A.7. EPR Spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

II.B. Surface Acidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

II.B.1. Brønsted Acid Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

II.B.2. Lewis Acid Sites and Expansion of

Coordination Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . 28

III. Oxo-Titanium Species and Reactive Intermediates . . . . . . . . . . . . . . 33

III.A. UV–Visible Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

III.B. Vibrational Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

III.C. X-Ray Absorption Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 39

III.D. Cyclic Voltametry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

III.E. EPR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

IV. Computational Investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

V. Catalytic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

V.A. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

V.B. Reactions Using H2O2 as Oxidant . . . . . . . . . . . . . . . . . . . . . . 56

V.B.1. General Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

v

V.B.2. H2O2-Catalyzed Reactions in the Homogeneous

Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

V.C. Epoxidation on Titanium Silicate Molecular Sieves . . . . . . . . . 60

V.C.1. General Features of Epoxidations. . . . . . . . . . . . . . . . . 60

V.C.2. Yields and Stereospecificities. . . . . . . . . . . . . . . . . . . . 62

V.C.3. Diffusional Constraints . . . . . . . . . . . . . . . . . . . . . . . . 62

V.C.4. Influence of Ti-Silicate Structure . . . . . . . . . . . . . . . . . 65

V.C.5. Epoxidation Catalyzed by Mesoporous Titanium

Silicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

V.C.6. Influence of Alkene Structure . . . . . . . . . . . . . . . . . . . 70

V.C.7. Dialkenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

V.C.8. Epoxidation in the Presence of Other Oxidizable

Functional Groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

V.C.9. Hydroxyl-Assisted Epoxidation . . . . . . . . . . . . . . . . . . 72

V.C.10. Diastereoselectivity in Epoxidations . . . . . . . . . . . . . . 74

V.C.11. Side Reactions During Epoxidation . . . . . . . . . . . . . . 75

V.C.12. Influence of pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

V.C.13. Epoxidation with Alkyl Hydroperoxides . . . . . . . . . . . 80

V.C.14. Epoxidation of Alkenes Containing Carbonyl Groups . . 81

V.C.15. Epoxidation Using Urea–H2O2 Adduct . . . . . . . . . . . 82

V.C.16. Epoxidation Using Dioxygen . . . . . . . . . . . . . . . . . . . 83

V.D. Hydroxylations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

V.D.1. General Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

V.D.2. Hydroxylation of Aliphatic Compounds . . . . . . . . . . . . 85

V.D.3. Hydroxylation of Aromatic Compounds . . . . . . . . . . . . 89

V.E. Oxidation of Nitrogen-Containing Compounds . . . . . . . . . . . . . 90

V.F. Oxidation of Sulfur-Containing Compounds . . . . . . . . . . . . . . . 93

V.G. Oxidation of Oxygen-Containing Compounds . . . . . . . . . . . . 100

V.G.1. Alcohols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

V.G.2. Ethers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

V.G.3. Phenols. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

V.G.4. Ketones, the Baeyer–Villiger Oxidation. . . . . . . . . . . 102

V.H. CyN Cleavage Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

V.I. Acid-Catalyzed Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

V.I.1. Beckmann Rearrangement . . . . . . . . . . . . . . . . . . . . . . 106

V.I.2. Synthesis of Polycarbonate Precursors. . . . . . . . . . . . . . 106

V.I.3. Transesterification of Esters . . . . . . . . . . . . . . . . . . . . . 110

V.I.4. Carbon–Carbon Bond Formation Reactions . . . . . . . . . 110

Contentsvi

V.I.5. Formation of Pinacols . . . . . . . . . . . . . . . . . . . . . . . . . 114

V.I.6. Oxidative Dehydrogenation . . . . . . . . . . . . . . . . . . . . . 115

V.J. Photocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

V.J.1. Photocatalytic Degradation of Pollutants . . . . . . . . . . . . 116

V.J.2. Photocatalytic Synthesis . . . . . . . . . . . . . . . . . . . . . . . . 120

V.J.3. deNOx Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

V.K. Influence of Solvents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

V.L. Influence of Silylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

VI. Structure-Activity Correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

VI.A. Structure of Titanium Species and Activity . . . . . . . . . . . . . . 127

VI.B. Titanium-Oxo Species and Activity . . . . . . . . . . . . . . . . . . . 128

VII. O–O Bond Cleavage and Product Selectivity. . . . . . . . . . . . . . . . . 137

VII.A. General. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

VII.B. Epoxidation of Alkenes . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

VIII. Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Appendix A. Fingerprint Features for Ti Isomorphous Substitution

in TS-1 Titanosilicates . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Appendix B. Characteristics of the Oxo-Titanium Species Generated

on TS-1 on Contact with Aqueous H2O2 . . . . . . . . . . . . . . 143

Appendix C. Synthesis of Titanium Silicate Molecular Sieves . . . . . . . . 143

C.1. TS-1, TS-2, Ti-ZSM-48, Ti-MWW, and Ti-MMM-1. . 144

C.2. Ti-Beta Zeolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

C.3. Ti-Containing HMS, MCM-41, and MCM-48. . . . . . . 147

C.4. Ti-SBA-15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

C.5. Ti-TUD-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

Electron Microscopy and the Materials Chemistry of Solid Catalysts

John Meurig Thomas and Pratibha L. Gai

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

II. Electron Microscopy (EM) Methods . . . . . . . . . . . . . . . . . . . . . . . . 176

II.A. Electron Microscopy in Catalysis . . . . . . . . . . . . . . . . . . . . . . 177

II.B. Imaging in the Electron Microscope . . . . . . . . . . . . . . . . . . . . 178

II.C. TEM Imaging Method Using Diffraction Contrast. . . . . . . . . . 179

II.D. Theoretical Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

III. High-Resolution Transmission Electron Microscopy . . . . . . . . . . . . 181

Contents vii

III.A. Conditions Required for Optimizing HRTEM Images . . . . . 182

III.B. Development of HRTEM. . . . . . . . . . . . . . . . . . . . . . . . . . 184

III.C. Elucidation of the Structures of Meso- and Microporous

Catalysts by HRTEM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

III.C.1. L-Type Zeolite Catalysts . . . . . . . . . . . . . . . . . . . . 185

III.C.2. Metal-Substituted Aluminum Phosphate

(MAPO-36) Microporous Catalysts . . . . . . . . . . . . 186

III.C.3. High-Silica Microporous SSZ-48 Catalysts . . . . . . 187

III.C.4. Intergrowths in Zeolite Catalysts: Coherent,

Recurrent, and Random. . . . . . . . . . . . . . . . . . . . . 188

IV. Chemical Composition Analysis with the Analytical

Electron Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

V. Scanning Transmission Electron Microscopy . . . . . . . . . . . . . . . . 193

VI. Recent Advances in Ultra-High Resolution, Low-Voltage

Field Emission Scanning Electron Microscopy and

Extreme FESEM in Catalysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

VII. Cathodoluminescence Imaging for Elucidation of

Electronic Structures of Catalysts. . . . . . . . . . . . . . . . . . . . . . . . . 195

VIII. Recent Advances in In Situ Atomic Resolution-Environmental

Transmission Electron Microscopy (ETEM) Under Controlled

Reaction Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

VIII.A. In Situ Investigations of Gas–Solid Reactions and

Active Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

VIII.B. Illusrative Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

VIII.B.1. In Situ Gas–Catalyst Reactions

at the Atomic Level . . . . . . . . . . . . . . . . . . . . . 201

VIII.B.2. Atomic-Resolution ETEM of Butane

Oxidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

VIII.B.3. Atomic-Resolution ETEM of Nanorods . . . . . . 210

VIII.C. Advances in In Situ Wet-Electron Microscopy

Technique (Wet-ETEM) for Probing Solid Catalysts

Under Liquid Environments . . . . . . . . . . . . . . . . . . . . . . 210

IX. Environmental Scanning Electron Microscopy . . . . . . . . . . . . . . . 212

X. Electron Tomography: Three-Dimensional Electron

Microscopy Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

X.A. The Topography and Location of Nanoparticles

in Supported Catalysts; BSE and HAADF . . . . . . . . . . . . . . 213

X.B. Pinpointing the Location of Nanoparticles Supported

on Nanoporous Solids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

Contentsviii

XI. Energy Filtered Transmission Electron Microscopy

and Elemental Maps of Solid Catalysts Using EFTEM . . . . . . . . . 218

XII. Other Significant Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

XIII. Critical Evaluations of the Methods and Challenges . . . . . . . . . . . 220

XIV. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Chemistry and Technology of Isobutane/Alkene Alkylation Catalyzed

by Liquid and Solid Acids

Andreas Feller and Johannes A. Lercher

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

II. Alkylation Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

II.A. Overall Product Distribution . . . . . . . . . . . . . . . . . . . . . . . . . 234

II.B. Initiation Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

II.C. Alkene Addition and Isomerization . . . . . . . . . . . . . . . . . . . . 239

II.D. Hydride Transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

II.E. Oligomerization and Cracking . . . . . . . . . . . . . . . . . . . . . . . . 247

II.F. Self-Alkylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

II.G. Product and Acid Degradation . . . . . . . . . . . . . . . . . . . . . . . . 251

II.H. Pathways to Allylic and Cyclic Compounds . . . . . . . . . . . . . . 251

II.I. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

III. Physical–Chemical Phenomena Influencing the Reaction. . . . . . . . . 252

III.A. Properties of Liquid Acid Alkylation Catalysts . . . . . . . . . . . 253

III.B. Properties of Zeolitic Alkylation Catalysts . . . . . . . . . . . . . . 255

III.B.1. Adsorption and Diffusion of Hydrocarbons . . . . . . . . 255

III.B.2. Brønsted Acid Sites . . . . . . . . . . . . . . . . . . . . . . . . . 256

III.B.3. Lewis Acid Sites and Extra-Framework Aluminum . . 260

III.B.4. Silicon/Aluminum Ratio . . . . . . . . . . . . . . . . . . . . . 261

III.B.5. Metal Ions in Ion-Exchange Positions. . . . . . . . . . . . 263

III.B.6. Structure Types of Zeolites . . . . . . . . . . . . . . . . . . . 264

III.C. Other Solid Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

III.C.1. Sulfated Zirconia and Related Materials . . . . . . . . . . 267

III.C.2. Heteropolyacids. . . . . . . . . . . . . . . . . . . . . . . . . . . . 268

III.C.3. Acidic Organic Polymers . . . . . . . . . . . . . . . . . . . . . 269

III.C.4. Supported Metal Halides . . . . . . . . . . . . . . . . . . . . . 270

III.D. The Influence of Process Conditions . . . . . . . . . . . . . . . . . . . 271

III.D.1. Reaction Temperature . . . . . . . . . . . . . . . . . . . . . . . 272

Contents ix

III.D.2. Alkane/Alkene Ratio and Alkene Space Velocity . . . 274

III.D.3. Alkene Feed Composition . . . . . . . . . . . . . . . . . . . . 276

IV. Industrial Processes and Process Developments. . . . . . . . . . . . . . . . 278

IV.A. Liquid Acid-Catalyzed Processes . . . . . . . . . . . . . . . . . . . . . 278

IV.A.1. Sulfuric Acid-Catalyzed Processes . . . . . . . . . . . . . . 278

IV.A.2. Hydrofluoric Acid-Catalyzed Processes . . . . . . . . . . 281

IV.B. Solid Acid-Catalyzed Processes . . . . . . . . . . . . . . . . . . . . . . 283

IV.B.1. UOP Alkylenee Process . . . . . . . . . . . . . . . . . . . . . 285

IV.B.2. Akzo Nobel/ABB Lummus AlkyCleane Process . . . 286

IV.B.3. LURGI EUROFUELw Process. . . . . . . . . . . . . . . . . 286

IV.B.4. Haldor Topsøe FBAe Process . . . . . . . . . . . . . . . . . 287

V. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

Catalytic Conversion of Methane to Synthesis Gas by

Partial Oxidation and CO2 Reforming

Yun Hang Hu and Eli Ruckenstein

I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

II. Partial Oxidation of Methane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

II.A. Hot Spots in Catalyst Beds . . . . . . . . . . . . . . . . . . . . . . . . . . 301

II.B. Minimizing O2 Purification Costs. . . . . . . . . . . . . . . . . . . . . . 306

II.C. Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312

II.D. Reaction Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

II.D.1. Changes in Catalyst During Reaction . . . . . . . . . . . . . 315

II.D.2. Which is the Primary Product, CO or CO2? . . . . . . . . 316

II.D.3. CHx Species and Rate-Determining Steps . . . . . . . . . . 318

II.D.4. Comparison of Reactions on Reduced and

Unreduced Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . 320

III. CO2 Reforming of Methane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

III.A. Carbon Formation on Metal Surfaces . . . . . . . . . . . . . . . . . . 321

III.B. Critical Issues Related to Carbon Deposition . . . . . . . . . . . . . 322

III.C. Supported Noble Metal Catalysts . . . . . . . . . . . . . . . . . . . . . 323

III.D. Non-Noble Metal Supported Catalysts . . . . . . . . . . . . . . . . . 324

III.D.1. Ni/Al2O3 Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . 325

III.D.2. Ni/SiO2 Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . 327

III.D.3. Ni/La2O3 Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . 328

III.D.4. Ni/ZrO2 Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . 330

III.D.5. Other Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331

Contentsx

III.E. MgO-Containing Solid-Solution Catalysts . . . . . . . . . . . . . . . 332

III.E.1. Characteristics of MgO-Containing Solid-Solution

Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332

III.E.2. Highly Effective MgO-Containing Solid-Solution

Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333

IV. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347

Contents xi

This Page Intentionally Left Blank

Contributors

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

ANDREAS FELLER, Institut fur Technische Chemie, Technische Universitat

Munchen, D-85747 Garching, Germany (229)

PRATIBHA L. GAI, DuPont, Central Research and Development Laboratories,

Experimental Station, Wilmington, DE 19880-0356, USA and also at Department

of Materials Science, University of Delaware, Newark, DE 19716, USA (171)

YUN HANG HU, Department of Chemical Engineering, State University of New

York at Buffalo, Buffalo, NY 14260, USA (297)

H. KNOZINGER, Department Chemie-Physikalische Chemie, Universitat

Munchen, Butenandt Strasse, 5-13, Haus E, D-81377 Munchen, Germany (1)

JOHANNES A. LERCHER, Institut fur Technische Chemie, Technische Universitat

Munchen, D-85747 Garching, Germany (229)

P. RATNASAMY, National Chemical Laboratory, Pune 411008, India (1)

ELI RUCKENSTEIN, Department of Chemical Engineering, State University of

New York at Buffalo, Buffalo, NY 14260, USA (297)

D. SRINIVAS, National Chemical Laboratory, Pune 411008, India (1)

JOHN MEURIG THOMAS, Davy Faraday Research Laboratory, The Royal

Institution of Great Britain, 21 Albemarle Street, London, United Kingdom and

also at Department of Materials Science, Cambridge CB2 1QY, UK (171)

xiii

Preface

The forty-eighth volume of Advances in Catalysis includes a description of a new

and increasingly well understood class of catalysts (titanosilicates), a review of

transmission electron microscopy and related methods applied to catalyst

characterization, and summaries of the chemistry and processes of isobutane-

alkene alkylation and partial oxidation and CO2 reforming of methane to

synthesis gas.

Ratnasamy, Srinivas, and Knozinger provide an incisive review of recent

advances in the understanding of titanosilicate catalysts, which have generated

intensive research activity and already found industrial application for

hydroxylation of phenol to hydroquinone and catechol. This chapter comp-

lements one by Notari in Volume 41 of Advances in Catalysis. The application of

physical and computational methods has resulted in a detailed understanding of

the nature and coordination state of titanium ions and functional groups such as

OH on dehydrated titanosilicate molecular sieves. Tetrapodal (Ti(OSi)4) and

tripodal (Ti(OSi)3OH) structures have been identified, and the interactions of

these active sites with oxidant/reactant molecules during catalysis lead to the

formation of oxo intermediates. The authors analyze the properties of the

catalysts that influence the activity and selectivity of these sites and the reaction

intermediates, showing, for example, that O–O bond cleavage can occur

heterolytically or homolytically, with the relative rates determining product

selectivities. The review includes a compilation of reactions catalyzed by

titanosilicates, including epoxidations, hydroxylations, oxidations of nitrogen-

and oxygen-containing organic compounds, and acid-catalyzed and photocata-

lytic reactions. The results lead to correlations between catalyst structure and

activity of titanium sites and reactivity of oxo-titanium intermediates.

Thomas and Gai contribute an exhaustive review of advanced methods of

electron microscopy, highlighting the techniques that provide the most insight

into the understanding of solid catalysts. The techniques comprise high-

resolution real-space imaging, electron crystallography, powerful scanning

probe methods, and electron energy loss spectroscopy. Recent developments in

electron tomography permit the three-dimensional imaging of catalytic materials

at the nano scale, and environmental cells make possible the direct in-situ probing

of the dynamics of catalytic reactions at the atomic scale. The authors emphasize

the complementarity of electron microscopy and other physical characterization

tools (including sum frequency generation, scanning tunneling microscopy,

and X-ray absorption spectroscopy) and the accompanying capabilities for

xv

elucidation of the nature of solid catalysts in the electron microscope, including

determination of the number and nature of crystallographic phases; electronic

properties such as oxidation states of particular atoms and the electronic structure

of the solid; coordination of atoms to neighboring atoms; locations of active sites;

mechanisms of the release of structural oxygen and of the creation of defects; and

the accommodation of catalyst non-stoichiometry.

Feller and Lercher present a critical and insightful assessment of alkylation of

isobutane with light alkenes, summarizing both the chemistry and processes.

Alkylation is gaining in importance as aromatics and methyl-tertiary-butyl ether

in motor fuels are limited by environmental concerns. Increasingly, the branched

alkane products of alkylation are regarded as superior gasoline components. The

authors build from the well-known chemistry of acid-catalyzed hydrocarbon

conversion, using concepts such as those of carbenium ion stability and reactivity

to elucidate patterns of the complex parallel and consecutive reactions.

Considering both liquid-phase alkylation catalyzed by hydrofluoric acid and

sulfuric acid, they draw contrasts between the two classes of processes and assess

the interplay between the chemistry and effects of physical properites such as

viscosity and the solubility of hydrocarbons in acid phases, which illuminate

issues such as mixing and dispersion in the reactors, where the reactions occur

near liquid-liquid interfaces. Feller and Lercher also consider solid-catalyzed

alkylation, providing a critical review of process developments and the role of

zeolite catalysts. The fundamental chemistry of zeolite-catalyzed alkylation is

essentially identical to that occurring in acidic solutions, but key differences

between liquid and solid catalysts result from differences in individual reaction

steps originating from the variety of possible structures and distributions of acid

sites in the solid catalysts; the sensitivity to a particular parameter depends

strongly on the catalyst. All the acids deactivate by the formation of unsaturated

polymers, which are strongly bound to the acid. Liquid acid-catalyzed alkylation

is a mature technology, but solid acid-catalyzed alkylation now has been

developed to a point where it eliminates most of the drawbacks of the liquid acid

processes and can compete with them economically. Catalyst regeneration by

hydrogen treatment is the method of choice for the solid catalysts.

Hu and Ruckenstein present a review of the catalytic production of synthesis

gas from methane by partial oxidation and CO2 reforming. This chapter

complements that by Rostrup-Nielsen et al. in Volume 47 of the Advances, which

provides an in-depth review of the chemistry and technology of steam reforming

of hydrocarbons, with some information about CO2 reforming as well. Hu and

Ruckenstein present results of catalyst testing experiments, chemical reaction

engineering analysis, and determination of reaction networks, addressing the

issue of whether CO2 and H2O are the primary products and whether CO is

formed from CO2 or H2O and CH4 or directly from CH4 and O2. The rapid heat

generation that results when the partial oxidation of methane produces some CO2

Prefacexvi

leads to hot-spot formation in fixed-bed reactors and potentially hazardous

operation and difficulty in process control. Process options include the

application of fluidized bed reactors to flatten the temperature gradients and

processes that eliminate hot spots by combining the exothermic partial oxidation

with the endothermic CO2 reforming or steam reforming. The partial oxidation

requires an air separation unit, and a major research goal is to make the process a

commercial reality by reducing the cost of air separation, for example, by using

O2-permeable ceramic membrane reactors in which air could be used without

pre-separation. CO2 reforming of methane is in prospect an attractive technology

because it converts two greenhouse gases into useful chemicals. Catalyst

deactivation, a consequence of carbon deposition, constitutes the greatest

challenge in this process. Although noble metal catalysts are less sensitive to

carbon deposition, Ni-containing catalysts have attracted the most research

interest, and some are reported to have both high activity and stability. A solid

solution catalyst offers high activity, selectivity, and stability by inhibiting

carbon deposition and catalyst sintering.

B.C. GATES

Preface xvii

Robert L. Burwell, Jr.

1912–2003

Robert L. Burwell, Jr., Ipatieff Professor Emeritus of Chemistry at Northwestern

University, passed away at his home in Williamsburg, VA, on May 15, 2003. He

will be remembered by his many friends, colleagues, and students as a learned

gentleman of high moral standard, a dedicated educator, a thorough and brilliant

researcher in heterogeneous catalysis, and a leading figure in guiding the catalysis

community.

Robert Burwell was born May 6, 1912. He graduated from St. John’s College in

1932 and received his Ph.D. in 1936 from Princeton University under the guidance

of Sir Hugh Taylor. After three years as a chemistry instructor at Trinity College,

in 1939 he joined the Chemistry Department at Northwestern University. During

World War II, having enlisted, he worked at the Naval Research Laboratory

(1942-1945). After the war, he returned to the chemistry faculty at Northwestern

where he served until his retirement in 1980. He was Chair of the Chemistry

Department from 1952 until 1957 and in 1970 succeeded Herman Pines as Ipatieff

Professor, holding this position until his retirement. Later, as Ipatieff Professor

Emeritus, he continued his research and intellectual activities for another decade.

In 1994, he moved to Virginia with Elise, his wife of more than sixty years.

xix

To those who knew him personally, Burwell was not only an imposing

intellect, but a warm, deeply caring, pleasant person, and a complicated

individual with many facets. For instance, while wise and judicious, he

nevertheless conducted himself with a great sense of humor and wit. Any

whom he favored soon realized he could engage in lively conversation on

practically any subject. Many of his coworkers also remembered him for his

perceptive scientific advice and suggestions. Often in seminars, students felt that

they learned more about a subject from Burwell’s probing questions than from

the seminar itself. His family remembered him also as a caretaker extraordinaire.

His devotion to his beloved Elise, particularly during the last year of her life, will

be remembered by all.

During his career, Robert Burwell published more than 170 original research

articles. He was among the first scientists who understood the critical connection

between general chemistry and catalysis. He introduced and popularized

concepts that are now familiar and even commonplace within the entire catalysis

community. His research themes centered around elucidation of reaction

mechanisms, the nature of surface intermediates, and characterization of active

sites of solid catalysts. He was well known for the use of H-D exchange for such

studies. Using this technique, he identified the importance of 1,2-diadsorbed

alkane on noble metal surfaces in the exchange and the hydrogenation reaction,

and the irreversibility in the adsorption of alkene during hydrogenation. [J. Amer.

Chem. Soc. 148, 6272 (1960); Acc. Chem. Res. 2, 289 (1969); Catal. Rev.-Sci.

Eng. 7, 25 (1972)]. He established the “rollover” mechanism for cyclic

hydrocarbons in these reactions [J. Amer. Chem. Soc. 79, 5142 (1957)], and

the term “surface organometallic zoo.” He carefully documented the importance

of surface coordination unsaturation in catalysis by metal oxides [Adv. Catal. 20,

1 (1969)] and developed new catalysts of unusual activities by deposition of

organometallic complexes on alumina and silica, and by modifying silica

surfaces [J. Amer. Chem. Soc. 97, 5125 (1975); J. Catal. 52, 353 (1978); J. Amer.

Chem. Soc. 107, 641 (1985)]. Together with colleagues John Butt and Jerome

Cohen, he completed one of the most comprehensive series of characterizations

of supported noble metal catalysts, starting with the paper J. Catal. 50, 464

(1977) and concluding with the paper J. Catal. 99, 184 (1986).

Burwell’s contributions to the scientific community include service on the

governing body of the North American Catalysis Society from 1964 to 1977 as

Director, Vice President, and, from 1973 until 1977, President. From 1955 until

1984 he served the International Congress on Catalysis, as a member of the Board

of Directors; as U.S Representative; Vice President; and President (1980-84). He

chaired the Gordon Research Conference on Catalysis in 1957 and was Associate

Editor (1984-88) and a member of the Editorial Board of Journal of Catalysis. He

served on National Research Council committees, IUPAC committees, the

Petroleum Research Fund Advisory Board, the National Science Foundation

Robert L. Burwell, Jr. (1912–2003)xx

Chemistry Advisory Board, and others. Professor Burwell was a long-time

consultant for Amoco Oil Company and was a consultant for the World Book

Encyclopedia.

His many scientific contributions and their industrial applications were

recognized by the awards and honors he received. They include the American

Chemical Society Kendall Award in Colloid and Surface Chemistry in 1973, the

American Chemical Society Lubrizol Award in Petroleum Chemistry in 1983,

and the Alexander von Humboldt Senior Scientist Award. The Robert L. Burwell

Lectureship Award of the North American Catalysis Society was established in

recognition of his outstanding contributions to catalysis. Professor Burwell was

also known for the first short course in heterogeneous catalysis, which he taught

for several years with Michel Boudart.

Robert Burwell’s influence on the catalysis community goes beyond his

science to his sharing of his many cultural interests with his colleagues, friends,

and post-doctoral and graduate students.

Harold Kung

Kathleen Taylor

Gary Haller

Polly Burwell Haynes

Lou Allred

Robert L. Burwell, Jr. (1912–2003) xxi

Active Sites and Reactive

Intermediates in Titanium Silicate

Molecular Sieves

P. RATNASAMY* and D. SRINIVAS

National Chemical Laboratory, Pune 411008, India

and

H. KNOZINGER*

Department Chemie-Physikalische Chemie, Universitat Munchen, Butenandt Strasse,

5-13, Haus E, D-81377 Munchen, Germany

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

II. Active Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

II.A. State and Framework Coordination of Ti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

II.A.1. Diffraction Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

II.A.1.1. X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

II.A.1.2. Neutron Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

II.A.2. Influence of Particle Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

II.A.3. UV–Visible Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

II.A.4. Photoluminescence Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

II.A.5. X-Ray Absorption Spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

II.A.6. Vibrational Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

II.A.7. EPR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

II.B. Surface Acidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

II.B.1. Brønsted Acid Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

II.B.2. Lewis Acid Sites and Expansion of Coordination Sphere . . . . . . . . . . . . . . 28

III. Oxo-Titanium Species and Reactive Intermediates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

III.A. UV–Visible Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

III.B. Vibrational Spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

III.C. X-Ray Absorption Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

III.D. Cyclic Voltametry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

III.E. EPR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

IV. Computational Investigations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

ADVANCES IN CATALYSIS, VOLUME 48 Copyright q 2004 Elsevier Inc.ISSN: 0360-0564 DOI 10.1016/S0360-0564(04)48001-8 All rights reserved

*Corresponding author.

E-mail address: [email protected] (P. Ratnasamy); [email protected]

(H. Knozinger).

P. Ratnasamy, D. Srinivas and H. Knozinger / Adv. Catal. 48 (2004) 1–169

V. Catalytic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

V.A. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

V.B. Reactions Using H2O2 as Oxidant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

V.B.1. General Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

V.B.2. H2O2-Catalyzed Reactions in the Homogeneous Phase . . . . . . . . . . . . . . 58

V.C. Epoxidation on Titanium Silicate Molecular Sieves. . . . . . . . . . . . . . . . . . . . . . . 60

V.C.1. General Features of Epoxidations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

V.C.2. Yields and Stereospecificities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

V.C.3. Diffusional Constraints. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

V.C.4. Influence of Ti-Silicate Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

V.C.5. Epoxidation Catalyzed by Mesoporous Titanium Silicates . . . . . . . . . . . . 67

V.C.6. Influence of Alkene Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

V.C.7. Dialkenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

V.C.8. Epoxidation in the Presence of Other Oxidizable Functional Groups . . . . 72

V.C.8.1. Alkenes and Alcohol Functions . . . . . . . . . . . . . . . . . . . . . . . . 72

V.C.8.2. Alkenes and Alkanes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

V.C.9. Hydroxyl-Assisted Epoxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

V.C.10. Diastereoselectivity in Epoxidations . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

V.C.11. Side Reactions During Epoxidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

V.C.12. Influence of pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

V.C.13. Epoxidation with Alkyl Hydroperoxides . . . . . . . . . . . . . . . . . . . . . . . . 80

V.C.14. Epoxidation of Alkenes Containing Carbonyl Groups . . . . . . . . . . . . . . 81

V.C.15. Epoxidation Using Urea–H2O2 Adduct . . . . . . . . . . . . . . . . . . . . . . . . . 82

V.C.16. Epoxidation Using Dioxygen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

V.D. Hydroxylations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

V.D.1. General Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

V.D.2. Hydroxylation of Aliphatic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . 85

V.D.3. Hydroxylation of Aromatic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . 89

V.E. Oxidation of Nitrogen-Containing Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . 90

V.F. Oxidation of Sulfur-Containing Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

V.G. Oxidation of Oxygen-Containing Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

V.G.1. Alcohols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

V.G.2. Ethers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

V.G.3. Phenols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

V.G.4. Ketones, the Baeyer–Villiger Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . 102

V.H. CyN Cleavage Reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

V.I. Acid-Catalyzed Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

V.I.1. Beckmann Rearrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

V.I.2. Synthesis of Polycarbonate Precursors. . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

V.I.3. Transesterification of Esters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

V.I.4. Carbon–Carbon Bond Formation Reactions. . . . . . . . . . . . . . . . . . . . . . . . 110

V.I.5. Formation of Pinacols. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

V.I.6. Oxidative Dehydrogenation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

V.J. Photocatalysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

V.J.1. Photocatalytic Degradation of Pollutants . . . . . . . . . . . . . . . . . . . . . . . . . . 116

V.J.2. Photocatalytic Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

V.J.3. deNOx Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

V.K. Influence of Solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

V.L. Influence of Silylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

VI. Structure-Activity Correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127


VI.A. Structure of Titanium Species and Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

VI.B. Titanium-Oxo Species and Activity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

VII. O–O Bond Cleavage and Product Selectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

VII.A. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

VII.B. Epoxidation of Alkenes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

VIII. Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Appendix A. Fingerprint Features for Ti Isomorphous Substitution

in TS-1 Titanosilicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Appendix B. Characteristics of the Oxo-Titanium Species Generated on TS-1

on Contact with Aqueous H2O2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

Appendix C. Synthesis of Titanium Silicate Molecular Sieves . . . . . . . . . . . . . . . . . . . . . . . 143

C.1. TS-1, TS-2, Ti-ZSM-48, Ti-MWW, and Ti-MMM-1. . . . . . . . . . . . . . . . . 144

C.2. Ti-Beta Zeolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

C.3. Ti-Containing HMS, MCM-41, and MCM-48. . . . . . . . . . . . . . . . . . . . . . 147

C.4. Ti-SBA-15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

C.5. Ti-TUD-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

This review is a summary and critical analysis of recent advances in the understanding of

(a) the nature and coordination state of Ti ions and other functional groups (such as OH)

on dehydrated titanium silicate molecular sieves, (b) the type and structure of the oxo

intermediates generated by the interaction of these active sites with oxidant/reactant molecules

during catalytic reactions, and (c) the factors that influence the reactivity and selectivity of

these active sites and reaction intermediates. In the dehydrated state, most of the Ti4þ ions

have the tetrapodal (Ti(OSi)4) or the tripodal (Ti(OSi)3OH) structure. On contact with H2O2,

titanium oxo species, Ti(O2H) and Ti(O2z2), respectively, are formed. On reaction with organic

reactants, O–O bond cleavage in these titanium oxo species occurs in a hetero- or homolytic

manner. Product selectivity is determined by the relative importance of these two modes of

O–O cleavage. Factors such as the coordinative environment of titanium, substituents on the

O–O bond (H or alkyl), temperature, solvent, nature of the organic reactant, etc. influence the

mode of O–O cleavage. Correlations between the structure and catalytic activity of titanium

sites and oxo-titanium intermediates are also described. q 2004 Elsevier Inc.

Abbreviations

TS-1 and TS-2 microporous titanium silicate molecular sieves with MFI

and MEL structures, respectively

Ti-beta (Ti-b) large-pore titanium silicate with BEA structure

Ti-MCM-41 and Ti-MCM-48 titanium-containing Mobil composite materials/mesopor-

ous composite materials of type 41 (hexagonal array of

pores) and 48 (cubic array of pores)

Ti " MCM-41 titanium grafted on MCM-41

Ti-HMS titanium containing hexagonal mesoporous silica material

Ti-ZSM-48 titanium containing one-dimensional 10-ring zeolite

Ti-ZSM-12 Mobil Corporation’s one-dimensional large-pore

(12-membered ring) zeolite

ZSM-5 Zeolite Socony Mobil constructed from five-membered-

ring building units

P. Ratnasamy, D. Srinivas and H. Knozinger / Adv. Catal. 48 (2004) 1–169 3

Ti-SBA-15 mesoporous, titanium-containing silica self-assembly-15

(with uniform, hexagonal, tubular channels) synthesized by

using a triblock organic copolymer as a template

Ti-MMM titanium-containing microporous mesoporous material

Ti-MWW titanium silicate with MWW structure

Ti-TUD-1 three-dimensionally randomly connected mesoporous silica

ETS-4 and ETS-10 Engelhard Corporation titanium silicate molecular sieves

MST amorphous mesoporous silica-titania

VS-2 vanadium-containing silicalite with MEL topology

Sil silicalite

XRD X-ray diffraction

UV–visible ultraviolet–visible

DRUV diffuse reflectance ultraviolet

FTIR Fourier transform infrared

NIR near infrared

EXAFS extended X-ray absorption fine structure

XANES X-ray absorption near-edge structure

XAS X-ray absorption spectroscopy

XAFS X-ray absorption fine structure

EPR electron paramagnetic resonance (also known as electron

spin resonance (ESR))

NMR nuclear magnetic resonance

LMCT ligand to metal charge transfer

DFT density functional theory

HP aqueous H2O2

TBHP tert-butyl hydroperoxide

UHP urea–H2O2 (1:1) adduct

FCC fluidized catalytic cracking

TEOS tetraethyl orthosilicate

TEOT tetraethyl orthotitanate

TBOT tetrabutyl orthotitanate

TMAOH tetramethylammonium hydroxide

TEAOH tetraethylammonium hydroxide

TPAOH tetrapropylammonium hydroxide

TPABr tetrapropylammonium bromide

DDA dodecylamine

TEA triethanolamine

CTABr cetyltrimethylammonium bromide

DH0 gas-phase dissociation enthalpy

DE energy required for gas-phase heterolytic cleavage

TOF turnover frequency (moles of reactant converted per mole

of active catalyst species per unit time)

SEM scanning electron microscopy

TPD temperature programmed desorption

n frequency

Dn shift in peak position in frequency units

l spin–orbit coupling constant

D energy gap between pxg and p

yg orbitals of oxygen

E energy separation between 3sg and 1pgx orbitals of oxygen

gxx; gyy; and gzz principal g-values


pKa negative logarithm of acidity constant

rip ion pair separation

DEipðsolventÞ energy of heterolytic cleavage in a solvent

DEsolv solvation energy

1 dielectric constant

m dipole moment of the solvent

a radius of a spherical cavity formed by solvent molecules

surrounding an ion pair

e charge of the electron

I. Introduction

Hugh Taylor’s landmark postulate in 1925 that particular atoms or groups of

atoms on the surfaces of solids are the active sites responsible for the catalytic

activity and selectivity laid the foundation for catalysis by design (1,2). Once the

active sites for a particular reaction are identified, one can, in principle, design

and prepare an optimal catalyst wherein the constituents of the active sites are

laid out to meet the needs of that reaction. The design and preparation of

aluminosilicate-containing zeolite catalysts wherein the Al ions (the active sites)

are located in different shape-selective channels and cavities (as per the needs

of the reaction) is an illustration of the further development and beneficial

consequences of Taylor’s postulate (1,2) in the area of acid-catalyzed reactions.

Similarly, solid catalysts containing supported bimetallic nanoparticles that are

highly active and selective for the hydrogenation of specific organic functional

groups can now be tailor made (3,4).

The discovery by Taramasso et al. (5), in 1983, of a titanosilicate zeolite with

the MFI structure (titanium silicate-1, TS-1), active in oxidation reactions, raised

hopes of a similar achievement in the catalysis of oxidation reactions by solids.

Since 1983, many titanosilicate molecular sieves containing Ti ions in various

structural and geometric locations have been synthesized and their physical,

chemical and catalytic properties investigated (TS-2 (6,7), Ti-ZSM-48 (8),

Ti-beta (9–15), Ti-ZSM-12 (16), Ti-MCM-41 (17–19), Ti-HMS (19–21),

Ti-MCM-48 (22), Ti-MSU (23,24), Ti-SBA-15 (25–27), Ti-MMM (28–30), Ti-

MWW (31) and Ti-TUD-1 (32)). TS-1 was one of the earliest classes of

molecular sieves containing a transition metal cation (Ti4þ) in framework

positions and possessing remarkable activity and selectivity for partial oxidation

of organic reactants by aqueous H2O2. Such molecular sieves containing a redox

metal cation (such as Ti4þ, Fe3þ, or V3þ) in framework positions have an

enormous potential in shape-selective oxidation reactions, similar to the predomi-

nant role of their aluminosilicate analogs in acid-catalyzed reactions. However,

in comparison with the enormous literature on the structure and dynamics of

the acidic active sites in aluminosilicate zeolites (both Brønsted and Lewis acid


sites), our knowledge of the identity and structure of the active sites on these

titanosilicates, the configuration of the reaction intermediates formed by their

interaction with the oxidant/reactant molecules, and the reaction mechanism is

far from adequate.

An excellent overview of the early work (up to 1995) by Notari (33) and a

discussion of the state and coordination of titanium ions in titanium silicates by

Vayssilov (34) are already available. During the 1980s and 1990s, the main

technical issues that dominated the research in this area were the confirmation

of the isomorphous substitution of titanium in the MFI lattice of TS-1 and

the development of fingerprints for distinguishing samples of TS-1 with good

catalytic activity. These were characterized by the crystalline MFI XRD pattern;

small (,0.5 mm) particles; infrared/Raman bands at 960 and 1125 cm21; sharp

peaks at 210 nm in the UV region; the absence of significant absorption in the

250–400 cm21 region; the absence of other elements (such as Fe, Al, B, etc.);

and intense yellow color upon addition of aqueous H2O2. Substitution of Ti for

Si in other molecular sieve frameworks (both silicate and phosphate) and the

discovery of new catalytic applications were other areas of worldwide research.

Since the reviews of this area by Notari (33) and Vayssilov (34) in the mid-

1990s, significant advances have been made in the charaterization of these

materials by use of FTIR and resonance Raman vibrational spectroscopies

(35–45), EXAFS and XANES (35,43,46–49), EPR (50–54), NMR (55) and

UV–visible (55–57) spectroscopies as well as computational chemistry (41,48,

58,59,61–63). An informative review of the molecular structural characteristics

and physical chemical properties of titania–silica catalysts was published by

Gao and Wachs (64). There is a consensus now that tetrahedrally coordinated,

isolated Ti4þ ions in the MFI framework of TS-1 zeolite are the precursors of

the active sites for many selective oxidations.

Although a coherent picture of the identity and structure of the surface groups

on TS-1, TS-2, and, to some extent, Ti-MCM-41 is slowly emerging, the function

and role of these surface Ti and OH groups during catalytic oxidation reactions

is far from clear. Active sites are usually formed by the interaction of the solid

surface with the reactant molecules during the catalytic reaction (1,2). This is

especially true in oxidation catalysis with H2O2. Do the tetrahedrally coordinated

Ti ions present on the “free” surface preserve their tetrahedral coordination on

interaction with H2O2? In recent years, advances in in situ spectroscopic tech-

niques have added considerably to our knowledge of the structure of the active

sites and the nature of reaction intermediates on TS-1 and Ti-MCM-41 during

catalysis (35–57). Results of these investigations suggest that the coordination

number of the Ti ions expands from tetra- to penta- and 6-fold coordination on

contact with H2O, H2O2, reactant, and solvent molecules. The latter are probably

more relevant in the quest for the active site. Related to the nature of the titanium

species present during the catalytic reaction is the structure of the oxo


intermediate formed from H2O2 on contact with the titanium ion. Here again, in

situ EPR spectroscopic investigations carried out recently (51,52,54) in the

presence of H2O2, (H2 þ O2), H2O, NH3, and organic reactants (such as alkenes,

alcohols, and aromatic compounds) have revealed significant information about

the peroxo- and superoxo-species that are probably the reactive intermediates that

influence selectivity in the various oxidation reactions.

In contrast to the significant progress that has been made in the structural

and scientific investigations of TS-1 during the past two decades, and,

notwithstanding the enormous potential of such a novel class of selective

oxidation catalysts in the chemical and petrochemical industry, their commercial

utilization in industrial plants has been rather disappointing. This is especially so

when the applications are compared with the major commercial process

breakthroughs and dozens of industrial plants using the Al-MFI analogs during

a similar period after their discovery (applications include hydrodewaxing of

petroleum fractions, production of ethyl benzene, xylene isomerization, methanol

to gasoline conversion, use as FCC additives for production of alkenes, etc.) (65).

Only one world-scale commercial plant (for hydroxylation of phenol to

dihydroxy benzenes) (66) and a large pilot plant (for the ammoximation of

cyclohexanone) using TS-1 are reported to be in operation so far (67,68). Apart

from the higher cost of manufacture of TS-1 (the current price is about US $100/

kg), another major constraint has been the necessity to use H2O2 in stoichiometric

quantities, rather than molecular oxygen, as the oxidant. Because H2O2 itself is

rather expensive, its use can be commercially justified only for the manufacture

of high-value products (say, those costing more than US $2/kg), thereby

excluding the majority of bulk and petrochemicals.

High-valued fine chemicals (used in the pharmaceutical, agrochemical,

flavors, and perfumery industries) are, however, usually complex molecules

too large to enter the pores of the MFI structure in TS-1. This was one of

the driving forces for attempts, worldwide, to synthesize titanosilicate and

titanophosphate molecular sieves with large and mesoporous structures. Such

materials (such as Ti-beta, Ti-MCM-41, and Ti-SBA-15, for example) do not

have the geometric constraints of TS-1. Unfortunately, even though significant

success has been attained in the synthesis of such materials, they are not found to

be as chemoselective as TS-1 in oxidation reactions using aqueous H2O2 as the

oxidant. Their structural stability is also less (especially with regard to leaching

of the Ti ions). They are more suitable when alkyl hydroperoxides are used as

the oxidant, thereby lacking the advantages of inherent process simplicity and

environmental advantages that ensue when aqueous H2O2 is used.

Why is TS-1 more chemoselective than Ti-beta and Ti-MCM-41 (17–19)

even though Ti4þ ions are isolated and in near-tetrahedral locations in all

of them? Are differences in hydrophobicity/hydrophilicity between TS-1 and

the large/mesoporous material the only factors responsible for the lower


chemoselectivity of the latter? During the past few years, in situ XAFS

investigations (46–48) have revealed that although Ti4þ ions have 4-fold

coordination, in TS-1 and Ti-MCM-41, most of the Ti ions in the former have a

closed tetrapodal Ti(OSi)4 structure, whereas those in the latter have an open

tripodal Ti(OSi)3(OH) structure.

Parallel diffuse reflectance UV (DRUV) and EPR spectroscopic investigations

(51,52,54) have provided evidence that the nature of the oxo intermediates formed

on contact with H2O2 depends on the intrinsic local structure and environment

of the Ti ions. The tetrapodal structures seem to generate oxo species the con-

centrations of which correlate with selectivity in the epoxidation of alkenes.

The structure of the titanium peroxo and superoxo species formed on the

surface during the catalytic reaction influences the scission of the O–O bond in

H2O2 (homolytic vs. heterolytic). The oxo ion/radical formed during such

scission, in turn, determines the selectivity in oxidation reactions. Recent XAFS

(46–48) and Raman (39,42) spectroscopic investigations indicate that a side-on

bound O2 species is formed on interaction of H2O2 with TS-1. In situ UV and

EPR spectroscopic measurements also suggest (51,52) that at least some of

them exist as titanium superoxide ion radicals. Such species can initiate a

radical reaction pathway for the oxidation reaction. It is possible that, depending

on the type of oxo species and the consequent O–O bond scission, two different

mechanisms may be operative on TS-1: one involving the heterolytic O–O

bond dissociation, acting, for instance, in the epoxidation of alkenes, and a

second involving the homolytic O–O bond dissociation, acting in the oxidation

of alkanes and side chains in alkyl aromatics (66).

Although attempts have been made to replace the aqueous H2O2 oxidant with

a mixture of H2 and O2 in the presence of metals such as palladium and gold

(69–74), the observed catalytic activities are much lower. But selectivities of

99% for propene oxide formation from propene were observed by Haruta and

coworkers (73) with Au-containing TS-1 catalysts. In situ EPR investigations

(54) have shown that similar oxo species are generated in reactions using

H2 þ O2 instead of H2O2, thereby suggesting the exciting feasibility of designing

efficient Ti-silicate-containing partial oxidation catalysts which can use H2 þ O2

instead of the more expensive H2O2 as the oxidant.

The main objective of this review is to summarize and critically analyze recent

advances made in the characterization and catalytic properties of titanium silicate

molecular sieves after the reviews of Notari (33) and Vayssilov (34) in 1996 and

1997, respectively. Of special interest are

(1) the nature and coordination state of Ti ions and other functional groups (such

as OH) on the “free” surface of titanosilicates,

(2) the type and structure of the active sites and oxo intermediates generated

by interaction of these surface groups with oxidant/reactant molecules during

catalysis, and


(3) the factors that influence the reactivity and selectivity of these active sites

and reaction intermediates.

It is hoped that the better understanding of the active sites and reaction inter-

mediates will lead to the design of superior solid titanium-containing selective

oxidation catalysts.

II. Active Sites

Although many micro- and mesoporous titanosilicate-containing oxidation

catalysts have been synthesized and their catalytic properties studied extensively

since 1983, detailed information about surface structure and active sites is

available mainly for TS-1 and, to a limited extent, Ti-MCM-41. The surface

structures of titanosilicates can be described in terms of (i) the state and framework

coordination of Ti and (ii) surface –OH groups present in the form of silanols and

titanols. All these structural characteristics together influence the catalytic activity

and selectivity. In this section, the various parameters affecting the surface

structure and the methodologies adopted to quantify and distinguish the surface

properties of the titanosilicate molecular sieves are discussed. The reviews by

Notari (33) and Vayssilov (34) give excellent accounts of the early structural work

done up to about 1995. During this period, the main subjects of investigation were

(i) the state and extent of framework coordination of Ti ions, (ii) the presence,

nature, and influence of extra-framework titanium, (iii) the influence of impurities

(such as Al, Fe, B, etc.), (iv) the types of surface acidic sites, (v) the influence of

surface hydrophobicity/hydrophilicity on catalytic activity and selectivity, and

(vi) the dependence of product distribution on crystal size.

II.A. State and Framework Coordination of Ti

According to Pauling’s criterion, Ti4þ cannot normally be included in frame-

work positions in the silicate structure as its ionic radius is too large. Titanium

compounds with tetrahedral geometry are scarce, as highly stable hexacoordi-

nated complexes are more stable. However, the flexibility of the MFI framework

(for example, for the reversible orthorhombic $ monoclinic transformation)

or the fact that it tolerates the trigonally coordinated B atom in B-substituted

ZSM-5, allows for such a substitution (5). But because of the differences in the

ionic radii, the coordination about Ti cannot be perfect tetrahedral, but instead

is pseudotetrahedral. Moreover, in small crystals of dimension of about 0.1 mm,

of TS-1, for example, even the silicate lattice will contain many defects (Si–OH

groups) and, hence, can accommodate some additional strain in accepting the

larger Ti ions in tetrahedral positions.


As Ti is incorporated in the silicate lattice, the volume of the unit cell expands

(consistent with the flexible geometry of the ZSM-5 lattice) (75), but beyond a

certain limit, it cannot expand further, and Ti is ejected from the framework,

forming extraframework Ti species. Although no theoretical value exists for such

a maximum limit in such small crystals, it depends on the type of silicate structure

(MFI, beta, MCM, mordenite, Y, etc.) and the extent of defects therein, the latter

depending to a limited extent on the preparation procedure. Because of the

metastable positions of Ti ions in such locations, they can expand their geometry

and coordination number when required (for example, in the presence of

adsorbates such as H2O, NH3, H2O2, etc.). Such an expansion in coordination

number has, indeed, been observed recently (see Section II.B.2). The strain

imposed on such 5- and 6-fold coordinated Ti ions by the demand of the

framework for four bonds with tetrahedral orientation may possibly account for

their remarkable catalytic properties. In fact, the protein moiety in certain

metalloproteins imposes such a strain on the active metal center leading to their

extraordinary catalytic properties (76).

II.A.1. Diffraction Techniques

II.A.1.1. X-Ray Diffraction. The X-ray patterns of silicalite-1 and TS-1

demonstrate a change from the monoclinic structure of the former to

orthorhombic when Ti4þ is introduced into the silicalite framework (5). The

Rietveld analysis of Millini et al. (75) demonstrates a linear dependence of the

lattice parameters and unit cell volume on the extent of Ti substitution in

silicalite-1 and constitutes confirmatory evidence for the location of Ti in

framework positions. Millini and Perego (77) concluded that the upper limit for

incorporation of Ti in the TS-1 framework is about 2.5%.

XPS (78–80) and XANES (81–84) data indicate that in the as-synthesized and

calcined state all the Ti ions in titanosilicates are in the þ4 oxidation state.

II.A.1.2. Neutron Diffraction. There are 12 crystallographically distinct T sites

in the orthorhombic structure of silicalite (MFI type), as illustrated in Fig. 1. The

exact location of the Ti atoms in TS-1 could not be determined unambiguously by

X-ray diffraction, even on the basis of high-quality synchrotron data (85–87).

The first evidence for non-random siting of Ti atoms was obtained by neutron

diffraction (85,87,88). It is complicated to determine the preferred Ti substitution

sites in TS-1 because of the low concentration of titanium (less than 2.5 Ti atoms

per unit cell) and the presence of silicon vacancies. Although the neutron

scattering length of titanium is quite different from that of silicon, it remains

difficult to determine a multiple Ti site substitution among the 12 possible ones.

Hijas et al. (87) concluded from their neutron diffraction results that Ti is

distributed among only four or five of the 12 sites, with Ti occupying T3(0.30),


T7(0.34), T8(0.92), T10(0.41), and T12(0.50), where the numbers in parentheses

represent the estimated site occupancies for the 2.57 total Ti atoms per unit cell

of the particular sample. Investigating a TS-1 sample with a Si:Ti atomic ratio

of 39:1, Henry et al. (85) applied a combination of single and multiple data set

Rietveld analyses exploiting the scattering length contrast between the different

titanium isotopes and silicon. They succeeded in determining the silicon vacancy

and titanium site substitution distribution. Both distributions were found to be

non-random, with Ti preferentially substituting three of the 12 crystallographi-

cally independent framework sites, namely, T8, T10, and T3 (in the order of

decreasing Ti content), and silicon vacancies being located at two framework

sites, T1 and T5. Although not identical with that reported by Hijas et al. (87),

this titanium siting agrees reasonably well with it. In contrast, Lamberti et al. (88)

concluded from their neutron diffraction data that T6, T7, and T11 are the sites

most populated by Ti.

The debate about the origin of the discrepancies in these results is ongoing

(85,88). Very likely, the preparation procedures of TS-1 have a significant

influence on the Ti site distribution, and it was argued that kinetics rather

than thermodynamics controls the framework formation and stability (85,87)

(Section V.C.3).

Fig. 1. The structure of orthorhombic form of silicalite-1 (MFI type) showing the 12 crystallo-

graphically distinct T sites. The oxygen atoms are omitted for clarity [Reprinted from Henry et al. (85)

with permission. Copyright (2001) American Chemical Society].


II.A.2. Influence of Particle Size

A useful “fingerprint” of an active TS-1 catalyst is the particle size of the

titanosilicate (,0.4 mm). Although the particle size influences the catalytic

activity of all molecular sieves, it is especially so in the case of TS-1 and due care

should be exercised in comparing samples varying in particle size (89,90).

II.A.3. UV–Visible Spectroscopy

Additional evidence of isolated Ti ions in tetrahedral locations in the silicate

lattice comes from the diffuse reflectance UV band indicative of a charge transfer

process in isolated Ti(OSi)4 or Ti(OSi)3(OH) units from the ligand oxygen to an

unoccupied orbital of the central Ti ion (82,84,91). This band occurs at 210 nm

for TS-1 and TS-2, at 220 nm for Ti-MCM-41 (51,52), and at 205–220 nm

for Ti-beta(F) that was synthesized in a fluoride medium (13). TS-1 (and other

titanosilicates) sometimes also contain Ti ions in other coordination states

(usually six) and in non-framework locations. The latter exhibit a broad

absorption in the region about 270–290 nm. If the Ti content is high, a separate

titania phase is also observed. Large anatase particles have an absorption

maximum at 330 nm, and rutile absorbs at about 400 nm. Amorphous TiO2–SiO2

shows a band at 290 nm (possibly penta- or hexacoordinated Ti). The blue shift

from 330 nm (anatase) to 210 nm (TS-1) is due to isolation of the Ti ion in the

silicate matrix and the change in coordination (from 6 to 4). These spectral

differences among Ti ions in various environments can be related to different

Ti–O–Si bond angles at the Ti sites (92). An increase of the angle will shift

the bridging oxygen hybridization from sp3 to sp2 and eventually to sp, favoring

a p-electron donation into the empty orbitals of Ti in Td symmetry. As a con-

sequence, the non-bonding “e” level of Td will split into a bonding “ep” level and

an empty anti-bonding “epp” level (LUMO). Because it is this LUMO that is

involved in the ligand-to-metal charge transfer (LMCT) responsible for the UV

band, the enlargement of the Ti–O–Si angle (as a result of a change from 6-fold

to 4-fold coordination, for example) will lead to a blue shift of the LMCT band, as

indeed has been observed experimentally. On the basis of XANES data, Gleeson

et al. (47) inferred two types of tetrapodal structures, one having three 1408

Ti–O–Si angles and one 1608 Ti–O–Si angle and the other having only two

1408 Ti–O–Si angles but two 1608 Ti–O–Si angles (Fig. 2).

These structures should, in principle, show LMCT transitions at two different

positions. Except for TS-1, data representing these angles for other titanosilicates

are not available. Such data would be useful in determining the influence of

the Ti–O–Si angle on the ease of hydrolysis of the Ti–O–Si bond, which is

crucially important for the stability and, hence, utility of the material in

catalytic applications.


Table I illustrates the utility of DRUV–visible data in determining the surface

structures involving Ti. Samples of TS-1 were prepared by three different methods

or treatments. Samples 1 and 2 were prepared by conventional hydrothermal

synthesis and sample 3 by synthesis in a fluoride medium. TS-2 was synthesized

as reported (7). At least five bands could be discerned by deconvolution (Fig. 3),

at 205, 228, 258, 290, and 330 nm. Band 1 at 205 nm is assigned to tetrahedral,

tetrapodal Ti present in TS-1, TS-2, and Ti-beta. Band 5 at 330 nm is assigned to an

TABLE I

Diffuse reflectance UV-visible data of titanosilicate samples

Titanosilicatea Deconvoluted bands and assignments: lmax, nm (relative intensity, %)

Band 1

(Ti(OSi)4)

Band 2

(Ti(OH)(OSi)3)

Band 3

(Ti(OH)(H2O)(OSi)3)

Band 4 (Ti(OH)2

(H2O)2(OSi)2)

Band 5

(Anatase-like)

TS-1

(Sample 1)

206 (85) 228 (8) 258 (6) 293 (1) Nil

TS-1

(Sample 2)

203 (72) 228 (10) 255 (8) 288 (5) 328 (5)

TS-1

(Sample 3)

206 (78) 229 (11) 260 (7) 293 (4) Nil

TS-2 201 (58) 229 (13) 255 (24) 288 (5) Nil

Ti-MCM-41 207 (27) 227 (49) 263 (8) 290 (16) Nil

Adapted from Shetti et al. (93).a All the titanosilicates (TS-1 (Si/Ti ¼ 33), TS-2 (Si/Ti ¼ 30) and Ti-MCM-41 (Si/Ti ¼ 35)) except

TS-1 (sample 3) were synthesized by the conventional pre-hydrolysis method (see Appendix C).

Sample 3 was synthesized in the fluoride medium.

Fig. 2. Schematic representations of the two different tetrapodal environments: Model A,

characterized by 3 Ti–O–Si angles of 1408 and 1 at 1608; Model B, characterized by 2 Ti–O–Si

angles of 1408 and 2 at 1608 [Reproduced from Gleeson et al. (47) by permission of the PCCP Owner

Societies].


anatase—such as phase. Band 2 at 228 nm is probably best assigned to tetrahedral,

tripodal Ti (present in all the samples, with the maximum amount in Ti-MCM-41).

Bands 3 and 4 are probably best attributed to penta- and hexacoordinated open Ti

structures in which Ti is attached to ligands such as H2O.

Fig. 3. Experimental and deconvoluted DRUV–visible spectra of TS-1 ðSi=Ti ¼ 33Þ and TS-2

ðSi=Ti ¼ 30Þ samples prepared by various methods/treatments. Deconvoluted bands are representated

by 1–5 [from Shetti et al. (93)].


II.A.4. Photoluminescence Spectroscopy

Because of the high sensitivity of Ti-containing luminescence centers to their

local environments, photoluminescence spectroscopy can be applied to discrim-

inate between various kinds of tetrahedral or near-tetrahedral titanium sites,

such as perfectly “closed” Ti(OSi)4 and defective “open” Ti(OSi)3(OH) units.

Lamberti et al. (49) reported an emission spectrum of TS-1 with a dominant

band at 495 nm, with a shoulder at 430 nm when the sample was excited at

250 nm. When the excitation wavelength was 300 nm, the emission spectrum

was characterized by a dominant band at 430 nm with a shoulder at 495 nm.

These spectra and their dependence on the excitation wavelength clearly indicate

the presence of two slightly different families of luminescent Ti species, which

differ in their local environments, in agreement with EXAFS measurements

carried out on the same samples.

When photoluminescence spectra were recorded for a Ti(OSi(CH3)3)4 model

compound, upon excitation at 250 nm only one emission band was detected (at

500 nm), which was assigned to a perfect “closed” Ti(OSi)4 site. The excitation

of these species is considered to be a LMCT transition, O22Ti4þ ! (O2Ti3þ)p,

and the emission is described as a radiative decay process from the charge

transfer state to the ground state, O2Ti3þ ! O22Ti4þ. Soult et al. (94) also

observed an emission band at 499 nm, which they attributed to the presence of

a long-lived phosphorescent excited state. The emission band at 430 nm of TS-1

was tentatively assigned to a defective “open” Ti(OSi)3(OH) site (49).

Ti-beta at 77 K exhibits a photoluminescence spectrum at about 465 nm (95).

The excitation was at 260 nm. Addition of H2O and CO2 quenches the photo-

luminescence, H2O being more effective than CO2 (Fig. 4). The lifetime of the

charge transfer excited state was also shortened by such additions, indicating that

H2O and CO2 interact with the Ti4þ ions in both the ground and excited states.

Recently, Gianotti et al. (96) reported photoluminescence and DRUV spectra

of pure siliceous MCM-41 and Ti-MCM-41 containing Ti4þ species anchored to

the inner walls of the siliceous MCM-41. They observed complex luminescence

signals and concluded that these could be used for a clear distinction of the

emission of tetrahedral Ti4þ ions from those of silica surface centers.

II.A.5. X-Ray Absorption Spectroscopy

A distinctive feature of Ti4þ ions in tetrahedral coordination is the intense

XANES peak at 4969 eV (39,97). The position and intensity of the pre-Ti K edge

peaks can throw significant light on the coordination number and correspond-

ing concentrations of surface Ti ions. The pre-edge intensity arising from the

transition between the core level (in this case 1s) to an unoccupied or a partially

occupied level (3d, which is unoccupied, because Ti4þ is a d0 system) is known


to be sensitive to the symmetry of the coordination environment. Ti4þ ions in

octahedral positions show low intensity (because the corresponding A1g ! T2g

and A1g ! Eg transitions are symmetry-forbidden) and those in tetrahedral

positions show the maximum intensity. Penta-coordinated Ti4þ ions (square

pyramidal, for example) exhibit intermediate values. Rutile and anatase, in which

all the Ti ions are in 6-fold coordination, exhibit three low-intensity peaks.

Titanium complexes, some of which are known from single crystal XRD data to

incorporate Ti4þ ions in Td positions, or well-synthesized samples of TS-1 exhibit

an intense peak in the pre-edge region (Fig. 5), the intensity of which should be

proportional to the Ti content of the sample. When the intensities of pre-edge

peaks of samples containing varying amounts of Ti are normalized to the

absorption edge jump (i.e., to the respective total amount of absorbing Ti atoms

contained in the sample), the resulting values are invariant, as shown in the inset

in Fig. 5, thus demonstrating the proportionality between pre-edge peak intensity

and the amount of Ti in a given sample.

Difficulties may arise when a sample contains Ti ions in more than one type

of location (the usual case). An intense peak representative of tetrahedral Ti

(the majority species) can then also include contributions from minor quantities

of Ti in 5- and 6-fold coordination (34). In particular, such species are observed

if the samples are not fully dehydrated or contain larger amounts of Ti. EXAFS

investigations of TS-1 (98,99) and TS-2 (81,100) indeed showed the presence of

Fig. 4. (a) The photoluminescence spectrum of Ti-beta(OH) and the effects of the addition of CO2

((b) 0.5 mmol CO2/g) and H2O ((c) and (d) 0.1 and 0.5 mmol H2O/g, respectively) molecules on the

photoluminescence spectrum. Measurements were made at room temperature with excitation at 260 nm

[Reproduced from Yamashita et al. (95) with kind permission of Kluwer Academic Publishers].


6-coordinated Ti in addition to the tetracoordinated Ti species. This technique

is not sensitive enough to discriminate between mixtures of this predominant

species with other oxidic tetrahedral species (101,102). DFT calculations (103)

indicated the possible coexistence of various oxidic tetrahedral structures, as

the difference in energy between them was very small (about 20 kJ/mol). Well-

prepared, dehydrated, titanocene grafted on MCM-41 (104) and TS-1 (47,49)

catalysts contained mainly the tetrahedral, tripodal (in Ti-MCM-41) and

tetrapodal structures (in TS-1) as the most plausible of the averaged structures.

Blasco et al. (13) observed single sharp and intense pre-edge peaks for

calcined dehydrated Ti-beta silicates which were synthesized in either an OH2

(Ti-beta(OH)) or F2 (Ti-beta(F)) medium, suggesting the uniformity of the

tetrahedral Ti species in these materials. Rehydration affected the pre-edge

peak, resulting in a decrease of the intensity, a shift of the peak position to higher

energy, and a peak broadening. The effects of rehydration were more noticeable

for samples synthesized in an OH2 medium, and it was concluded that the degree

of interaction of titanium with water was strongly influenced by the hydrophobic/

Fig. 5. XANES spectrum of a typical TS-1 sample in vacuum. Inset: intensity of the pre-edge peak

(spectra normalized to the edge jump) for samples with various Ti contents. Because the height of the

edge jump is proportional to the Ti content, the intensity of the normalized pre-edge is invariant

(within experimental uncertainty) with Ti concentration [Reprinted from Ricchiardi et al. (41) with

permission. Copyright (2001) American Chemical Society].


hydrophilic character of the zeolitic framework. The XANES spectra of hydrated

Ti-beta(F) were consistent with the presence of Ti in either 4 or 5-fold

coordination, indicating the strong adsorption of one water molecule per Ti

atom. This result was confirmed independently by adsorption measurements.

In contrast, the XANES spectrum of hydrated Ti-beta(OH) was consistent with

a mixture of 5 and 6-fold coordinated Ti atoms, suggesting the preferred

adsorption of one or two water molecules per Ti atom, as supported by

independent adsorption measurements.

II.A.6. Vibrational Spectroscopy

In addition to the characteristic XRD patterns and photoluminescence, UV–

visible and X-ray absorption spectra, another fingerprint thought to indicate

lattice substitution of titanium sites was the vibrational band at 960 cm21, which

has been recorded by infrared and Raman spectroscopy (33,34). Although there is

some controversy about the origin of this band, its presence is usually character-

istic of a “good” TS-1 catalyst, although it turned out to be experimentally

extremely difficult to establish quantitative correlations between the intensity of

the 960 cm21 band and the Ti content of a Ti silicate and/or its catalytic activity.

The band at 960 cm21 was already reported in the original TS-1 patent (5)

and attributed to the presence of isomorphously substituted Ti in the silicate

lattice. It was shown later that an analogous band in the 960–970 cm21 range

also characterizes other Ti silicates, namely, TS-2, Ti-ZSM-48, Ti-beta, and

Ti-MCM-41 (34). This band was attributed in early work to a Si–O stretching

vibration in a Si–O–Ti group (91) and later to a titanyl TiyO group (105). The

attribution of the band to the presence of Ti in the silicate matrix was based on

the argument that Ti-free silicates would not show any vibrational modes in the

950–970 cm21 region. However, this reasoning is not entirely valid, because

the presence of bands in this region, although they are weak, has been reported

for the Raman spectra of pure silicalite-1 (106) and for the infrared spectra of

crushed silica, alkali silicates, and silica gels (107–109). Therefore, Camblor

et al. (110) assigned the band at 960 cm21 to the stretching vibration of Si–O2

groups. An analogous band was also observed in the spectra of zeolites with high

concentrations of defects. The observation of an oxygen isotope effect (9,10,111)

and the absence of a hydrogen isotope effect were considered consistent with

this band assignment. However, it was recently demonstrated that in Ti-beta

synthesized by the fluoride route there is no noticeable hydrolysis of Ti–O–Si

bonds (13). Consequently, bands near 960 cm21 cannot be attributed to Si–OH

defects, which are essentially absent from these zeolites. It was, therefore,

concluded (112) that the stretching of Si–O bonds in Si–O–Ti groups is the

major contribution to the absorption in this region in Ti silicates, in agreement

with previously reported results (91,112,113).


Boccuti et al. (91) interpreted the 960 cm21 band on the basis of a consider-

ation of the effect of a TiO4 unit on the vibrational modes of a neighboring SiO4

tetrahedron. The Si–O stretching mode was expected to shift to lower wave-

numbers because of the higher ionicity of the Ti–O bond (Si–Od2zzzTidþ). The

quantum chemical (SCF) calculations of de Man and Sauer (62) suggested that

the 960 cm21 band can be interpreted as an antisymmetric stretching mode of

the Si–O–Ti bridge in a Ti(OSi(OH)3)4 unit in which Ti is tetracoordinated.

Ricchiardi et al. (41) pointed out that these band assignments may be considered

as coincident because they describe the same physical mode on the basis of

different building units.

Su et al. (114), in an investigation of a wide variety of silicotitanates by Raman

spectroscopy, concluded that for titanosilicates containing isolated TiO6 units, a

strong band at 960 cm21 indicative of the [(O3Si–O)]d2–[(TiO5)]dþ stretching

mode will dominate the spectra. In contrast, Smirnov and van de Graaf (115),

applying molecular dynamics techniques, calculated the vibrational spectrum of a

periodic model of TS-1 containing TiO4 tetrahedra and supported the localized

Ti–O–Si nature of the 960 cm21 vibration. They also emphasized that the Si–O

and Ti–O bands are not equivalent and that the Si–O stretching makes the greater

contribution to the vibration, consistent with previous conclusions (41,91).

Further support for the direct relationship of the 960 cm21 band to the presence

of 4-coordinated Ti atoms in the framework of TS-1 came from the photo-

luminescence investigations of Soult et al. (94). At 12 K, an emission band

was observed at 490 nm, which was unequivocally attributed to titanium

(Section II.A.4). This band showed a resolved vibrational structure of

966 ^ 24 cm21, which clearly demonstrates that Ti is involved in the

corresponding vibrational mode.

This relationship was recently questioned by Li et al. (40,116) when they

reported the observation of bands at 490, 530 and 1125 cm21 in the UV-excited

(244 nm) Raman spectra of TS-1. Bands at 1085 and 1110 cm21 were also

observed for Ti–SiO2 prepared by chemical grafting (117) and for Ti-MCM-41

(118), respectively. Raman bands near 1120 cm21 in addition to the 960 cm21

band had been reported earlier for TS-1 by Scarano et al. (113) and Deo et al.

(119), who used conventional Raman spectroscopy (NIR excitation), and later by

Bordiga et al. (39), who used UV–visible- and NIR-excitation. Li et al. (40,116)

were the first to show that the bands of TS-1 at 490, 530, and 1125 cm21 and

the corresponding bands of Ti-MCM-41 were resonance-enhanced when the

Raman spectra were excited in the UV (244 nm) in the wavelength region of

the O22Ti4þ ! O2Ti3þ LMCT absorption (band at 220 nm; see Section II.A.3),

whereas the 960 cm21 band was not resonance-enhanced. On the basis of

this observation, the authors concluded that the oscillator responsible for the

960 cm21 band cannot be located in the immediate vicinity of the Ti atom.

Consequently, they also proposed that the three resonance-enhanced bands at


490, 530, and 1120 cm21 were the real fingerprint for the presence of Ti in

the framework. The three bands were assigned to the bending, symmetric, and

antisymmetric stretching modes of a Ti–O–Si unit (116).

Unfortunately, the different selection rules that apply to resonant and normal

Raman scattering were not taken into account in this spectral interpretation.

In the following, it is shown that the conclusions and assignments mentioned

above have to be modified on the basis of symmetry considerations as discussed

by Ricchiardi et al. (41).

Figure 6 reproduces the Raman spectra in the region 800–1200 cm21 reported

by these authors for pure silicalite (sample 1) and for two TS-1 samples, 3 and 5,

which contain 1.4 and 3.0 wt% TiO2. The spectra shown in Fig. 6a were recorded

with a Fourier transfrom (FT) Raman spectrometer at an excitation wavelength

of lexc ¼ 1064 nm (9398 cm21), whereas those shown in Fig. 6b were excited

with a UV–laser line at lexc ¼ 244 nm (40,984 cm21). With each excitation

wavelength, the pure silicalite gives rise to weak bands at 975 and 1085 cm21 and

a complex band centered near 800 cm21. In the FT-Raman spectra of the dehy-

drated TS-1 samples (Fig. 6a), a band is clearly visible at 960 cm21, the intensity

of which increases with TiO2 content.

This band is not to be confused with the silicalite band that is observed

at 975 cm21. In addition, a band appears at 1125 cm21, the intensity of which,

although relatively low, also grows with the TiO2 content. Hence, both bands

Fig. 6. Raman spectra of sample 1 (Ti-free silicalite), and samples 3, and 5 (TS-1 with TiO2 wt%

being 2 and 3, respectively). (a) Spectra collected with a l ¼ 1064 nm (9398 cm21) excitation.

(b) Spectra collected with a l ¼ 224 nm (40,984 cm21) excitation. Inset: UV–DRS spectrum of

sample 5. Vertical line indicates the position of the excitation wavelength l used for collecting the

sample reported in part (b). Vertical dotted lines are placed at 960 cm21. Spectra of both parts have

been vertically shifted for clarity [Reprinted from Ricchiardi et al. (41) with permission. Copyright

(2001) American Chemical Society].


may be considered as fingerprints of the Ti incorporation into the silicalite

framework. In contrast, the UV–excited Raman spectra (Fig. 6b) show a weak

band at 960 cm21 and a very strong band at 1125 cm21, suggesting a resonance

enhancement of this vibration, but not of the 960 cm21 band, consistent with the

observations reported by the group of Li (40,116–118).

The requirements for Raman resonance that must be fulfilled are the following

(120,121): (a) total symmetry of the vibrations with respect to the absorbing

center, and (b) same molecular deformation induced by the electronic and vibra-

tional excitations. Quantum chemical calculations (41) of the vibrational freque-

ncies and the electronic structure of shell-3 cluster models allowed the assignment

of the main vibrational features, as shown in Fig. 7. The 1125 cm21 band is

unequivocally assigned to the symmetric stretching of the TiO4 tetrahedron.

Vibrations of the TiO4 tetrahedron, achieved via in-phase, anti-symmetric

stretching vibrations of the four-connected Ti–O–Si oscillators, are outlined

in Fig. 8b. Considering the electronic structure of the Ti moiety and the symmetry

of this mode, it is the only vibration that fulfills the resonance Raman selection

rules (a) and (b) above. This vibrational mode can be described equivalently as

the in-phase stretching of the four Si–O bonds surrounding Ti. The 960 cm21

band is assigned to the antisymmetric stretching mode of the TiO4 unit, which can

Fig. 7. Calculated vibrational frequencies for the Ti[OSi(OH)3]4 model, classified following the

symmetries of the T–O–T unit (upper part) or according to the symmetries of the TO4 unit (lower part)

[Reprinted from Ricchiardi et al. (41) with permission. Copyright (2001) American Chemical Society].


be described as the out-of-phase-antisymmetric stretching of the four connected

Ti–O–Si oscillations or as the out-of-phase stretching of the four Si–O bonds

surrounding the Ti atom (Fig. 8c). This vibrational mode does not fulfill the

resonance Raman selection rules (a) and (b) above and is, therefore, not expected

to be resonance-enhanced, consistent with the experimental results (Fig. 6).

On the basis of these assignments, the two bands must be associated with the

presence of isolated Ti atoms in tetrahedral coordination within the silicalite

framework. Consequently, a quantitative linear correlation between the TiO2

content and the intensities of both the infrared and Raman bands at 960 cm21 is

expected—and this is indeed observed, as shown in Fig. 9b.

Furthermore, both the resonant (Fig. 6b) and non-resonant (Fig. 6a) Raman

spectra give a constant value for the ratio of the intensity of IR band at 1125 cm21

to that at 960 cm21 ðIð1125Þ=Ið960ÞÞ ratio of 0.25 and 11, respectively, for

samples with varying TiO2 contents. This result suggests that the two bands

should be related to two different spectroscopic manifestations of the same

phenomenon, namely, incorporation of Ti in the silicalite framework (41).

II.A.7. EPR Spectroscopy

Electron paramagnetic resonance (EPR) spectroscopy is yet another diagnostic

tool for the detection of isomorphous substitution of Ti. Its sensitivity is very

high, and investigations can be performed with samples even with very low

contents of paramagnetic species. The spectra and g parameters are sensitive to

the local structure and associated molecular distortions. Hence, it is an ideal tool

to characterize Ti in titanosilicates. Ti in the þ 4 oxidation state in titanosilicates

is diamagnetic and hence EPR-silent. Upon contacting with CO or H2 at elevated

Fig. 8. (a) Definition of symmetric and antisymmetric stretching modes of the T–O–T bridges. (b)

Symmetric stretching of the central tetrahedron, achieved through in-phase antisymmetric stretching of

the four connected Ti–O–Si bridges. (c) One of the antisymmetric stretching modes of the central

tetrahedron, achieved through out-of-phase antisymmetric stretching of the Ti–O–Si bridges

[Reprinted from Ricchiardi et al. (41) with permission. Copyright (2001) American Chemical Society].


temperatures, the Ti ions are reduced from a diamagnetic þ 4 (3d0) to a param-

agnetic, EPR-active þ 3 (3d1) oxidation state. Tuel et al. (122) and Zecchina

et al. (123) used this technique to differentiate Ti3þ ions from framework and

extraframework precursors. Later, Kevan and co-workers (124–129) investi-

gated TS-1 and Ti-MCM-41 reduced with g-radiation. This method is, however,

valid only if the reduced structure retains a structure memory of the precursor.

Recently, Srinivas and Ratnasamy (130,131) reported a detailed EPR investi-

gation of Ti3þ in titanosilicate molecular sieves, TS-1, Ti-MCM-41, ETS-4, and

ETS-10 (Fig. 10). Ti4þ was reduced to Ti3þ by dry hydrogen. Only one type of

Ti3þ species (I) was identified when the sample was reduced at 673 K. However,

reduction at 873 K revealed two non-equivalent Ti3þ ions (species I and II) in

TS-1 and Ti-MCM-41 (Table II). ETS-4 and ETS-10 contained only one type

of Ti3þ ion in octahedral positions. In agreement with the other spectroscopic

investigations (XAS and UV), EPR gave evidence for the presence of two types

of tetrahedral Ti (tetrapodal and tripodal) structures in TS-1 and Ti-MCM-41,

differing in their reducibility (130,131). The EPR g-parameters (Table II) indicate

that Ti3þ ions in TS-1 and Ti-MCM-41 have a tetragonally elongated Td

Fig. 9. (a) Infrared spectra of outgassed thin pellets of Ti-free silicalite (curve 1) and TS-1 with

increasing Ti content x (curves 2–5). Spectra were normalized by means of the overtone bands

between 1500 and 2000 cm21 (not shown) and vertically shifted for clarity. The thick horizontal line

represents the fwhm of the 960 cm21 band for sample 2. By assuming that this band has a constant

fwhm for any x; the absorbance W obtained is plotted as the ordinate in panel b, where the band has the

same fwhm as in curve 2 (horizontal thin lines). (b) Intensity W of the 960 cm21 infrared band

(normalized absorbance units) as a function of x (full squares) and corresponding Raman counts

(open squares) [Reprinted from Ricchiardi et al. (41) with permission. Copyright (2001) American

Chemical Society].


geometry whereas those in ETS-4 and ETS-10 have a tetragonally compressed

Oh geometry.

The reducibility of Ti (monitored by formation of Ti3þ) varied with the type of

silicate structure. The spectra normalized (with respect to the Ti atoms in TS-1)

indicate that the overall signal intensity of Ti3þ ions decreases in the following

order: ETS-10 . ETS-4 q TS-1 at 673 K and ETS-4 . ETS-10 . Ti-MCM-

41 . TS-1 at 873 K. Apparently, it is more difficult to reduce Ti in a tetrahedral

coordination geometry (as in Ti-MCM-41 and TS-1) than in an octahedral

geometry (as in ETS-10 and ETS-4). The intensity of the Ti3þ signals increased

with an increase in the reduction temperature (673–873 K). The g-values are

sensitive to the silicate structure (Table II). Whereas both the Ti3þ species (I and

II) in TS-1 are characterized by axial symmetry, species I has axial symmetry,

and II has rhombic symmetry in Ti-MCM-41. In each structure, gk , g’. In the

case of ETS-10, gk . g’; and for ETS-4, gk , g’:The investigations also showed that counterions and additives also influence

the redox properties. ETS-10 samples were exchanged with Csþ ions to examine

Fig. 10. EPR spectra (at 77 K) of Ti3þ generated by contacting TS-1, Ti-MCM-41, and ETS-10

with dry H2 at 873 K, and ETS-4 at 673 K. Signals denoted by an asterisk correspond to superoxo

radical species generated by further reaction of Ti3þ with O2 [from Bal et al. (130)].


the interaction of extraframework ions with titanium. The exchanged samples

(ETS-10(Cs)) were then reduced with dry H2 at 673 K. The spectrum of ETS-10

containing Naþ/Kþ ions is characterized by axial g values with gk . g’: After

exchange of the cations with Csþ, the spectrum corresponded to rhombic

g-values with gzz , gxx; gyy (Fig. 11) and the overall Ti3þ signal intensity

decreased by a factor of about three.

A platinum (0.05 wt%)-impregnated ETS-10(Cs) sample showed spectra

similar to that of ETS-10(Cs) ðgzz , gxx; gyyÞ; except that the Ti3þ signal intensity

increased by a factor of about 2.4 compared with that of the ETS-10(Cs) sample.

Although the reduction in Ti3þ intensity by Cs is attributed to greater stabili-

zation of Ti4þ ions by the more basic and larger Cs atoms, the increase in the

intensity induced by platinum is attributed to better activation of the reductant

molecules (H2) by platinum and the consequently greater reduction of Ti4þ

to Ti3þ. In other words, both cesium and platinum influence the reducibility

TABLE II

EPR spin Hamiltonian parameters (at 77 K) of Ti3þ in titanosilicate molecular sieves generated by

reduction with dry hydrogen

Sample Reduction

temperature (K)

Species gk g’ gzz gxx gyy

ETS-10 873 1.969 1.942

673 1.966 1.941

ETS-10(Cs) 673 1.869 1.944 1.959

ETS-10(Cs)-Pt 673 1.870 1.943 1.959

ETS-4 673 1.863 1.930

ETS-4-Pt 673 I 1.870 1.920

II 1.863 1.930

TS-1 873 I 1.930 1.956

II 1.916 1.956

673 I 1.930 1.956

TS-1-Pt 673 I 1.931 1.955

Ti-MCM-41 873 I 1.902 1.958

II 1.894 1.938 1.974

Ti-MCM-41-Pt 873 I 1.906 1.958

II 1.894 1.938 1.974

Adapted from Bal et al. (130).


of Ti. Similar enhancements in Ti3þ signal intensity of TS-1 (by 3 times) and

Ti-MCM-41 (by 1.35 times) were observed when the titanosilicates were impreg-

nated with platinum.

II.B. Surface Acidity

II.B.1. Brønsted Acid Sites

In addition to the Ti, hydroxyl groups constitute a second class of surface

functional groups on dehydrated samples that can be of importance in catalytic

reactions. The presence of a large number of Si–OH groups on the surfaces

of all the titanosilicates is apparent from the intense absorption in the 3200–

3800 cm21 region of the infrared spectra. The experimental evidence of surface

Fig. 11. EPR spectra of Ti3þ (at 77 K) showing the influence of Cs exchange and platinum

impregnation on the intensity and g-parameters of Ti3þ signals in ETS-10 reduced in dry H2 at 673 K

(signals denoted by an asterisk correspond to superoxide radical species generated by secondary

reactions by Ti3þ interaction with O2) [from Bal et al. (130)].


Ti–OH groups, on the other hand, is scarce. Titanol groups on Ti-grafted

MCM-48 (132) and TS-1 (133) have been claimed to absorb at about

3676 cm21. In the case of TS-1, the 3676 cm21 band was not observed (133) on

the free dehydrated surface, but instead only as a result of contact with H2O2

and photoirradiation. TS-1 typically contains a high density of framework

defects (Si vacancies) generating internal, hydrogen-bonded hydroxyl groups

(silanols as well as possibly titanols acting as potential weak Brønsted acid

sites) (49,134). The infrared spectra in the O–H stretching region of dehydrated

TS-1 and pure silicalite are, therefore, very similar to each other and char-

acterized by broad bands, which do not allow an easy discrimination between

titanols and silinols (43,44,135,136). The presence of acidity in TS-1 was

inferred from typical acid-catalyzed reactions, such as the formation of diols in

epoxidation reactions (137), rearrangement of cyclohexanone oxime to capro-

lactam (138,139), and the cycloaddition of CO2 to epoxides (140), the latter two

not involving the use of H2O2 during the reaction. Although there is no doubt

about the presence of functional acid sites on dehydrated TS-1 (and other

titanosilicates), their type (Brønsted or Lewis), structure and concentration have

not yet been conclusively established. Of course, acidity can be generated,

in situ, during oxidation reactions in the presence of H2O2, because the peroxide

proton-donor group, generated by coordination of H2O2 to the titanium sites,

can be quite acidic (111). But, as noted earlier, there is evidence for the

occurrence of acid-catalyzed reactions on TS-1 even in the absence of H2O2

(138–140). However, results of earlier investigations of the acidity of TS-1

have to be viewed with caution because of inadequate appreciation of the

influence of impurities (such as Fe, Al, B, etc.) and non-framework Ti ions in

generating surface Brønsted acidity on these materials.

The Brønsted acid strength of the hydroxyl groups on dehydrated TS-1

was tested by measuring the wavenumber shift DnOH of the O–H stretching

bond induced by hydrogen bonding with probe molecules (141,142), viz., CO

(135,143), acetonitrile (100,136,141), tert-butylnitrile (141), and pyridine (44).

The O–H stretching spectra of TS-1 and pure silicalite resulting from the

adsorption of the probe molecules were practically identical for all probes. For

example, the O–H stretching band was found at 3390 cm21 for silicalite-1 and

at 3400 cm21 for TS-1 upon contact with acetonitrile. The corresponding

wavenumber shift is very close to the shifts of 300–330 cm21 reported for

amorphous silica after adsorption of acetonitrile (64,144). Brief outgassing

caused the almost complete disappearance of the band due to hydrogen bonding,

without leaving evidence of the presence of other components indicating that

the OH groups on TS-1 were not more acidic than those on silicalite-1. The

main conclusion was that the presence of Ti in the silicalite lattice does not

generate new OH groups or does not induce detectable Brønsted acidity in the

Si–OH groups of the silicalite (135,139).


Conclusions, some of them contrary to the above, were reached more recently

by Zhuang et al. (145) from a combination of 31P and 1H MAS NMR spectro-

scopy of adsorbed trimethylphosphine. These authors found not only Lewis acid

sites (vide infra), but also Brønsted acid sites in TS-1 (145). They claimed that

the 1H, 29Si MAS NMR spectra and the resonance related to Brønsted acid

sites in the 31P MAS NMR demonstrated clearly that the “presence of Ti in the

framework results in the formation of a new OH group, titanols, which is more

acidic than the silanols of silicalite-1 (145)”. The peak at 4.3 ppm in the 31P MAS

NMR spectra was assigned to a ((CH3)3P–H)þ complex arising from the inter-

action of (CH3)3P with Brønsted acid sites present on TS-1. The origin of this

proton is not clear at present, especially because the 1H MAS NMR spectra of

the same TS-1 samples did not differ significantly from those of silicalite-1 (145);

the latter, when free from impurities, is not known to be a Brønsted acid.

In conclusion, dehydrated TS-1 (and presumably other titanosilicates) most

likely does not have Brønsted acid centers. The observed activity for acid-

catalyzed reactions that yield undesired side products is, therefore, inferred to be

created under reaction conditions in the presence of aqueous H2O2 (vide infra).

II.B.2. Lewis Acid Sites and Expansion of Coordination Sphere

Although there are doubts about the existence of Brønsted acid sites on TS-1 and

related materials, there is strong evidence that Lewis acid sites are present on the

surface of dehydrated TS-1. The significant activity of TS-1 and of Ti-MCM-41

in the cycloaddition of CO2 to epoxides to give cyclic carbonates (140), a reaction

typically catalyzed by Lewis acids such as AlCl3, SbF5, etc., lends strong support

to the inference of the existence of Lewis acid sites on their surfaces.

Infrared spectroscopic evidence of Lewis acidity comes from recent spectra of

CH3CN adsorbed on TS-1 (136). In the liquid state, the C–N stretching vibration

is characterized by a doublet at 2294 and 2254 cm21, which is caused by Fermi

resonance (144). Upon interaction with electron-withdrawing groups, these

frequencies are shifted to higher values (146–148). When CH3CN is adsorbed

on silicalite-1, the bands shift to 2297 and 2263 cm21. The slight shift to higher

energy was attributed to hydrogen bonding with the silanol groups that act as

weak electron-withdrawing centers from the nitrile nitrogen lone pair. In the

case of TS-1, two doublets were observed, the first at 2313 and 2291 cm21 and

the second at 2290 and 2256 cm21. The band at 2256 cm21 and one of the bands

in the 2290 cm21 region decrease in intensity faster than the others upon out-

gassing as a result of the desorption of hydrogen bonded acetonitrile from

the Si–OH sites. The positions of the other (more stable) doublet (2313 and

2290 cm21) is similar to that found in the spectrum of anatase, TiO2, on which

two Lewis-bonded species, characterized by two doublets at 2315 and 2290 cm21

and 2304 and 2274 cm21 were observed earlier (149) and assigned to CH3CN


attached to tetra- and penta-coordinated Ti4þ ions. The observation of a similar

doublet (at about 2313 and 2290 cm21) in the case of both TS-1 and anatase on

adsorption of CH3CN suggests that Ti4þ ions in TS-1 also possess Lewis acidity

similar to that in anatase.

The detailed interpretation of the C–N stretching region of CH3CN is

relatively complex because of the Fermi resonance between the C–N stretching

fundamental mode n2 and the combination mode of the C–C stretching and

symmetric CH3 deformation modes that leads to the doublet mentioned above

(146). Therefore, the use of CD3CN as a probe molecule is preferred, as this

has only a single C–N stretching band (at 2259 cm21) in the free molecule.

This band shifts to higher wavenumbers when the molecule forms a coordination

bond (146). Bonino et al. (44), therefore, tested the Lewis acid centers in TS-1

with the infrared spectra of adsorbed CD3CN in comparison with those observed

for CD3CN on pure silicalite-1. The corresponding spectra are shown in Fig. 12.

On adsorption of CD3CN on silicalite-1, a C–N stretching band grows

at 2276 cm21 at low equilibrium pressure followed by a second band at

2265 cm21 as the pressure increases. These bands are attributed to CD3CN that

is hydrogen bonded to SiOH groups (Section II.B.1) and physically adsorbed

molecules, respectively. The same two bands are detected when CD3CN is

adsorbed on TS-1, together with an additional band at 2302 cm21 which char-

acterizes the most stable adsorbed species. The high C–N stretching frequency

signals the highest adsorption bond energy, with the CD3CN molecule being

coordinated to a Ti4þ ion:

ð1Þ

The inset in Fig. 12 shows the effect of the adsorption of CD3CN on the

960 cm21 framework band of TS-1, which clearly shifts to higher wave-

number with increasing CD3CN loading. This observation is a strong evidence

of the tetrahedral Ti4þ ions in the silicalite framework acting as Lewis acid sites,

which can undergo an expansion of their coordination sphere from a coordina-

tion number of four to a coordination number of five, as indicated in Eq. (1).

Bonino et al. (44) reported supporting evidence for the Lewis acid character of

the tetrahedral Ti4þ ions by using pyridine as an alternative probe. Furthermore,

quantum chemical calculations were fully consistent with the conclusions drawn

from the infrared spectra of the adsorbed probe molecules.

Zecchina et al. (135) were unable to detect coordination of CO on Ti4þ centers

at 77 K. A possible explanation for the apparent discrepancy between this result

and those stated above may be the steric shielding of the tetrahedral Ti4þ by


the oxygen ligands despite the larger size of Ti4þ relative to Si4þ. At 77 K, the

vibrational motions of the TiO4 moiety are likely frozen, and the oxygen ligands

may, therefore, not allow a close approach of the very weak base CO to the Ti4þ

center. In contrast, the stronger bases acetonitrile and pyridine may overcome the

steric barrier at the temperature of the experiments (room temperature).

Infrared spectra of pyridine adsorbed on dehydrated TS-1 and Ti-MCM-41 of

comparable Ti content indicated the presence of only Lewis acid sites (Fig. 13).

The infrared absorptions at 1595 and 1445 cm21 are attributed to hydrogen-

bonded pyridine (Si/Ti–OHzzzpyridine) and those at 1580 and 1485 cm21 to

pyridine bonded to weak Lewis acid sites (Fig. 12). Brønsted sites, if present,

Fig. 12. Background-substracted spectra at increasing coverage of CD3CN on TS-1 (top) and

silicalite-1 (bottom), n(CN) region. The spectra obtained at high CD3CN coverages are reported with

the bold line. The inset reports the perturbative effect of CD3CN on the 960 cm21 band; the pure TS-1

spectrum is reported with a dotted line, although the bold line reports the spectrum obtained at high

CD3CN coverage [Reprinted from Bonino et al. (44) with permission. Copyright (2003) American

Chemical Society].


should show pyridinium ion peaks at 1639 and 1546 cm21, and strong Lewis

acid sites should give rise to bands at 1623 and 1455 cm21 (141,150,151).

The infrared bands disappeared as temperatures were increased beyond 398 K

for TS-1 and 523 K for Ti-MCM-41, indicating higher acid strength in the

latter than in the former titanosilicate. Furthermore, the number of acid sites

(estimated from infrared peak intensities) is higher on Ti-MCM-41 than on TS-1.

The temperature-programmed desorption of NH3 from these samples showed

a desorption peak maximum at 448 K (Fig. 14). The peak is broader and more

asymmetric when the sample is Ti-MCM-41. The amount of NH3 desorbed is

1.3 times higher for Ti-MCM-41 than for TS-1.

With the Ti4þ ions acting as Lewis acid centers, a strong interaction with

ammonia and water with these centers is expected. There is in fact abundant

spectroscopic evidence for the coordination of NH3 and H2O molecules to

tetrahedral Ti4þ centers and for the corresponding expansion of their

coordination spheres.

Figure 15 shows the modification in the UV–visible spectra of TS-1, initially

in vacuo, upon interaction with H2O (152). Evidence of the interaction of NH3, a

stronger base, is also shown. The LMCT band (mentioned in Section II.A.3)

undergoes a red shift of the edge as a result of the increase of the coordination

sphere about Ti4þ ions. In TiO2, in which Ti is surrounded octahedrally by six

O atoms in its first coordination sphere, the Ti4þO22 ! Ti3þO2 LMCT is also

red shifted to lower wavenumbers (32,000 cm21). A stronger perturbation

is obtained upon dosing of NH3, but the line shape of the UV–visible curve is

Fig. 13. FTIR spectra of pyridine adsorbed on dehydrated TS-1 and Ti-MCM-41 [from Srinivas

et al. (152)].


similar. It was, therefore, concluded (152) that the four-coordinated, framework

Ti species in dehydrated samples of TS-1 increase their coordination number

(to 5 or 6) on interaction with H2O (or NH3), thus forming Ti(H2O)xO4 (or

Ti(NH3)xO4) species with x ¼ 1 or 2.

Bolis et al. (43) reported volumetric data characterizing NH3 adsorption on

TS-1 that demonstrate that the number of NH3 molecules adsorbed per Ti atom

under saturation conditions was close to two, suggesting that virtually all Ti

atoms are involved in the adsorption and have completed a 6-fold coordination:

Ti(NH3)2O4. The reduction of the tetrahedral symmetry of Ti4þ ions in the

silicalite framework upon adsorption of NH3 or H2O is also documented by

a blue shift of the Ti-sensitive stretching band at 960 cm21 (43,45,134), by a

decrease of the intensity of the XANES pre-edge peak at 4967 eV (41,43,134),

and by the extinction of the resonance Raman enhancement of the 1125 cm21

band in UV–Raman spectra (39,41). As an example, spectra in Figs. 15 and

16 show the effect of adsorbed water on the UV–visible (Fig. 15), XANES

(Fig. 16a), and UV–Raman (Fig. 16b) spectra of TS-1.

Appendix A summarizes what we believe to be the basic “fingerprint” features

for the isomorphous substitution of Ti in silicate-1 lattice.

Fig. 14. Temperature programmed desorption of NH3 profiles of TS-1 and Ti-MCM-41 [from

Srinivas et al. (152)].


III. Oxo-Titanium Species and Reactive Intermediates

Although the identification of tetrahedrally coordinated, tetra- and tripodal Ti4þ

ions on the surface of titanosilicates, as the likely active sites in reactions that

require Lewis acidity, seems convincing, the structure and role of the sites active

in catalytic oxidation, presumably oxo-titanium species, formed by the inter-

action of H2O2 (or H2 þ O2) with these surface Ti ions, are not clear. In recent

years, this problem has been investigated by FTIR (133), Raman (39,40),

XANES (46–48), electronic (54–57), and EPR (51–54) spectroscopies. This is

one of the areas in which major progress has been made since the reviews of

Notari (33) and Vayssilov (34). Zecchina et al. (153) recently summarized some

of the salient features of this progress.

Fig. 15. UV–visible spectra of a TS-1 catalyst in vacuo (solid line) and upon interaction from the

gas phase with H2O (dashed line) and NH3 (dotted line) [from Armaroli et al. (136)].


III.A. UV–Visible Spectroscopy

The color of an aqueous solution of Ti4þ in H2O2 depends on the pH, being orange

in acidic solutions, yellow in neutral solutions, and colorless in strongly alkaline

solutions. The yellow species contains one peroxy group for each Ti ion (154).

The formation of a yellow color when TS-1 is brought in contact with H2O2 and its

disappearance during the hydrocarbon oxidations has been known for a long time.

DRUV–visible spectroscopy has confirmed the formation, upon contact of TS-1

with H2O2/H2O solutions, of a new LMCT band at about 385 nm (26,000 cm21,

Fig. 17) corresponding to a charge transfer from the peroxide moiety to the Ti

center (42). Hence, this UV–visible light-absorbing species (a peroxo moiety

interacting with framework Ti ions) must be involved in the oxidation reaction.

The yellow color produced by aqueous H2O2 progressively loses its color with

time (153) (Fig. 17). The intensity, however, is nearly restored upon addition of

pure H2O to the system, and this observation highlights the cooperative role of

water in the stabilization of the Ti(O2) complex.

III.B. Vibrational Spectroscopy

Vibrational frequencies of some titanium peroxo complexes and of solids

containing peroxo and/or superoxo species are summarized in Table III.

The three infrared vibrations of the triangular peroxo group in the C2v structure

Fig. 16. Effect of soaking TS-1 with water on the XANES (a) and UV–Raman (b) spectra: dried

TS-1 (solid line); soaked TS-1 (dotted line). The inset in part (a) reports the k3-weighted, phase-

uncorrected Fourier transforms of the corresponding EXAFS spectrum [Reprinted from Ricchiardi

et al. (41) with permission. Copyright (2001) American Chemical Society].


typically appear in the regions 800–950 cm21 (n(O–O)), and 500–650 cm21

(n(M–O) symmetric and antisymmetric stretching) (155,156), the exact band

positions being strongly dependent on the nature of the central atom. The O–O

stretching mode of superoxo groups has been detected in the range of 1020–

1220 cm21 for the typical end-on configuration on CoO–MgO solid solutions

(157). Although O2z2 species have been detected on titanium-containing silicalites

Fig. 17. Evolution of the UV–visible spectra of a TS-1 catalyst brought in contact with an aqueous

solution of H2O2 as a function of time: 1 min, 4, and 8 h (curves 1, 2, and 3, respectively). Curve 4

shows the effect of H2O dosage on the catalyst sample after acquisition of spectrum 3 [Reproduced

from Zecchina et al. (153) with kind permission of Kluwer Academic Publishers].

TABLE III

IR spectroscopy of peroxo and superoxo species

Compound Dioxygen species n(O–O) n(M–O)s,as Oxygen source Reference

(Pic)2TiO2HMPA O222 895 575, 615 H2O2 (155)

(OEP)TiO2 O222 895 595, 635 H2O2, O2 (155)

Ca12Al10Si4O35 O222 895 O2 (156)

Ca12Al10Si4O35 O2z2 1075 O2 (156)

Pic, pyridine-2-carboxylate; HMPA, hexamethylphosphoric triamide; OEP, octaethylporphyrin.


by EPR spectroscopy (Section III.E), the corresponding O–O stretching vibration

has, to the best of our knowledge, never been reported. The lack of such reports

may possibly be a consequence of the low sensitivity of infrared and Raman

spectroscopy and an overlap of the O–O stretching band with the 1125 cm21

band of TiO4 tetrahedra.

Infrared absorption of an unstable hydroperoxo species had been observed at

230 K by Tozzola et al. (63). A peak at 886 cm21, strongly overlapping the peak

at 877 cm21 attributed to physisorbed H2O2, was attributed to TiOOH (h1; end-

on coordination), although a band at 837 cm21 was assigned to anionic triangular

Ti(O2) (side-on coordination).

Lin and Frei (133), upon loading of aqueous H216O2 into TS-1 and removal

of the solvent by evacuation, detected a peroxidic O–O stretch absorption at

837 cm21 and a broad band at 3400 cm21 by infrared difference spectroscopy.

The former absorption shifted to 793 cm21 when aqueous H218O2 was loaded in

TS-1 instead of H216O2 (Fig. 18). No bands were observed at 837 or 3400 cm21

with the same loading of H2O2 on silicalite-1.

Lin and Frei (133) assigned the 3400-cm21 band (Fig. 18) to hydrogen-bonded

OH groups of TiOOH, and the two infrared bands were suggested to originate

from a side-on hydroperoxo species (h2-Ti(O2H) interacting with frame-

work Ti (Scheme 1). The large red shift of the O–O stretching band (from

877 cm21 for physisorbed H2O2 to 837 cm21 for the strongly attached species)

was claimed to be a result of the hydroperoxo group’s being covalently linked

to the Ti center (133). This h2-Ti(O2H) group was found to be indefinitely

stable at room temperature. It was suggested that the exposure of dehydrated

TS-1 to H2O2 led (133) to the conversion of the tetrapodal framework Ti to

(SiO)3TiOOH (Scheme 1).

Fig. 18. Infrared difference spectra before and after loading of H216O2 (curve a) and H2

18O2 (curve

b) into TS-1 followed by 12 h evacuation (1025 mbar) [Reprinted from Lin and Frei (133) with



The very large bandwidth and red shift of nOH of the hydroperoxo group was

postulated to be evidence of hydrogen bonding to the oxygen of the Si–OH

moiety formed by cleavage of the Ti–O–Si linkage (Scheme 1). In the case of

the tripodal framework (SiO)3Ti–OH centers, substitution of OH by OOH rather

than opening of Si–O–Ti bridges was thought to occur. Hence, independent of

whether H2O2 reacts with tetra- or tripodal framework Ti, the result is the same,

namely, the formation of a TiOOH moiety adjacent to a Si–OH group. When the

DRUV difference spectrum of the H2O2-loaded TS-1 sample was recorded after

photolysis at 355 nm, it showed clearly the growth of a LMCT band with a

maximum at about 360 nm and a tail extending to 550 nm (Fig. 19).

Scheme 1.

Fig. 19. Diffuse reflectance difference spectrum of the LMCT absorption upon 355 nm photolysis

of TS-1/TiOOH molecular sieve (20 min at 45 mW cm22) [Reprinted from Lin and Frei (133) with



The red shift of this band from its position in the dehydrated sample (Section

II.A.3) is attributed to the increase of the coordination sphere about Ti4þ ions and

is similar to the changes observed on adsorption of H2O and of NH3 (153). The

simultaneous observation of the 837 and 3400 cm21 bands in the infrared region

(attributed to peroxidic O–O, Fig. 18) and the 360 nm band in the DRUV spectra

(attributed to octahedrally coordinated Ti4þ ions, Fig. 19) further confirms that

the Ti4þ ions in the side-bound Ti(O2H) species are indeed 6-fold coordinated.

When the H2O2-loaded TS-1 sample was irradiated with 355-nm light of a

Nd:YAG laser or the visible emission of a conventional tungsten source, photo-

dissociation of TiOOH was observed (133). The 837 and 3400 cm21 bands (and

the corresponding 18O substitutes) diminished in intensity (Fig. 20).

The loss of the 837 and 3400 cm21 bands was accompanied by the growth of

bands at 3676 cm21 (assigned to O–H), 1629 cm21 (assigned to the bending mode

of H2O), and 960 cm21 (assigned to Si–O–Ti), indicating at least partial restora-

tion of the original coordination environment of the metal center (Scheme 1). The

net result of the photodissociation is the disproportionation of TiOOH to TiOH

and O and the further condensation of this TiOH with adjacent SiOH to regenerate

Ti–O–Si and H2O. The lack of Ti leaching in TS-1 during catalytic oxidations

was attributed to such recondensation of the Ti–O–Si linkages.

The structure of the peroxide species in the TS-1 catalyst was also investigated

by resonance Raman spectroscopy (39,42). Interaction with H2O2 caused (i) a

reduction and blue shift (to 976 cm21) of the 960-cm21 band, (ii) a quenching

of the 1125 cm21 band in the UV–Raman spectrum as a result of the breakdown

of the tetrahedral symmetry, (iii) the appearance of a strong and sharp band

at 875 cm21 (attributed to O–O stretching in physically adsorbed H2O2), and

(iv) the appearance of a strong and complex new feature centered at 618 cm21.

The 618 cm21 band was assigned to a resonance Raman enhanced vibration

mode of the titanium peroxo complex. On the basis of the similarity between

the spectroscopic features in both the UV–visible and Raman spectra of

(NH4þ)3(TiF5O2)32 and TS-1/H2O2 systems, Bordiga et al. (42) concluded that

Fig. 20. Infrared difference spectra before and after 20 min irradiation (with 355 nm light

(45 mW cm22)) of aqueous H216O2 loaded TS-1 molecular sieves [Reprinted from Lin and Frei (133)

with permission. Copyright (2002) American Chemical Society].


the species responsible for the 385 nm LMCT band is a side-on titanium peroxo

species which is also characterized by a Raman mode at 618 cm21. The presence

of side-on (O2) attachment in the TiF5(O2) molecular unit of (NH4þ)3(TiF5O2)32,

in particular the Ti(O2) fragment, is known (42).

III.C. X-Ray Absorption Spectroscopy

XANES and EXAFS spectroscopies were applied by Zecchina et al. (153) to

investigate the changes in coordination of the framework Ti ions in TS-1 on

contact with H2O, NH3, and a mixture of H2O þ H2O2 (Fig. 21). There is a

progressive reduction in the pre-edge intensity on going from H2O to NH3 to

H2O þ H2O2, indicating the transition from four to six coordination (Section

II.B.2). Their EXAFS results suggested the formation of a strongly adsorbed

side-on peroxo complex in which both the O atoms are located at a Ti–O

distance of 2.01 A. Presumably, the formation of this complex is accompanied

Fig. 21. XANES spectra of TS-1 catalyst in vacuo and upon interaction with H2O (from the liquid

phase), NH3 (from the gas phase), and H2O/H2O2 (liquid solution) [Reproduced from Zecchina et al.

(153) with kind permission of Kluwer Academic Publishers].


by the hydrolysis of one or even two Ti–O–Si bonds and the total

deprotonation of H2O2 (153).

Ti(O2) and Ti(O2H) species formed on Ti " MCM-41 during reaction were

studied by using XANES and EXAFS measurements and density functional

theory (DFT) (36,46,48,104). Investigating the nature of titanium sites on

catalysts obtained by grafting titanocene dichloride on MCM-41 (Ti " MCM-41),

the authors found that in the “free”, dehydrated state, these sites consist mostly of

Ti4þ–OH groups tripodally anchored to the silica via covalent bonds to oxygen.

In addition to these tripodal, single-site, titanol centers, there were also bipodal

Ti4þ centers present in the as-prepared Ti " MCM-41 catalysts. Their proposed

models of the tetrahedral tri- and bipodal species are illustrated in Scheme 2.

There were no signs of Ti–O–Ti linkages, nor of any titanyl (TiyO) groups, nor

of a three-, five-, or six-coordinated species. Under reaction conditions when

cyclohexene and tert-butylhydroperoxide (TBHP) were brought in contact with

these catalysts, there was a decrease in the pre-edge intensity of the XANES,

in comparison with the intensities characterizing the calcined and dehydrated

catalysts, indicating that the coordination about the Ti ions increases on contact

with the oxidant/reactant. Considering both the intensity and position of the

pre-edge peak (the energy position of the peak after interaction with the TBHP

was slightly higher), the authors ruled out the presence of a five-coordinated

Ti species. The expansion in coordination was from four to six. Furthermore,

whereas four of the surrounding oxygen atoms are at distances strictly com-

parable to those in the pristine surface structure (about 1.81 A), in the reactive

state there are two additional oxygen atoms situated farther away (2.2–2.4 A).

The EXAFS data characterizing the (catalyst þ TBHP þ alkene) system also

indicated that there are at least three Ti–O distances close to 1.83 A (a slight

expansion compared to the “free” surface), and two of the other three oxygen

distances were between 2.2 and 2.4 A. From among different models of the

titanium oxo species investigated, the authors concluded that the Ti-h2-OOR

Scheme 2.


and Ti-h1-OOR structures (where R is H or alkyl) gave the best fits between the

experimental and computed EXAFS data (Fig. 22).

III.D. Cyclic Voltametry

The presence of two types of titanium sites in TS-1 (tetra- and tripodal) was also

suggested by the cyclic voltametry experiments of Bodoardo et al. (158). The

tripodal Ti(OSi)3(OH) showed a redox couple at 0 V and the tetrapodal Ti(OSi)4

Fig. 22. Best fit between experimental results and computed EXAFS employing the full multiple

scattering method. The model is depicted in the bottom right figure [Reprinted from Thomas and

Sankar (104) with permission. Copyright (2001) American Chemical Society].


a redox couple at 20.6 V, indicating that the electron density is higher in the

tripodal than in the tetrapodal structure. The higher electron density at Ti, in turn,

will increase the electron density at the O–O bond attached to it, facilitating the

cleavage of the latter. The ease of cleavage of the O–O bond will influence the

mode of its cleavage, homo- or heterolytic. Product selectivity in H2O2-catalyzed

reactions of course depends strongly on the mode of cleavage (homo or

heterolytic) of the O–O bond, as discussed in detail in Section VI.

III.E. EPR Spectroscopy

Superoxide species, O2z2, were observed by Zhao et al. (50) by EPR spectroscopy

on contact of TS-1 with H2O2. Two types of superoxides were identified, a major

species with gzz ¼ 2:0236; gyy ¼ 2:0100; and gxx ¼ 2:0091; and a minor species

differing only in its gzz value which was 2.0270 in contrast to 2.0236. The major

signal was assigned to superoxides on framework titanium sites and the weaker

signal to those on dispersed, extra-framework titanium sites. The superoxide

attached to the framework Ti was also less stable, decomposing completely

within a few hours. The second signal, assigned to the superoxide on non-

framework Ti, was more stable. When a drop of phenol in acetone solution was

wetting TS-1, the lines of the superoxide species on framework Ti disappeared

and a new intense signal attributed to phenoxy radicals appeared. It was suggested

that the appearance of the phenoxy radical along with the disappearance of the

superoxide on framework titanium sites provided direct support for a free radical

mechanism of oxidation.

The formation of paramagnetic oxygen species as a result of interaction of

H2O2 or H2 þ O2 with titanosilicates was also investigated by Ratnasamy et al.

(51,52,54) using a combination of UV–visible and EPR spectroscopies. The

diamagnetic peroxo/hydroperoxo species (TiO2H) could be discerned by their

UV–visible spectra, and the concentration of the paramagnetic superoxo species

(Ti(O2z2)) was independently estimated from their EPR spectra. Two types of

Ti4þ-superoxo species, A and B (A being preponderant), were detected in TS-1

and Ti-beta. Ti-MCM-41 contained mainly species B (Fig. 23). An additional

species, C, was detected upon interaction of TS-1 with the (H2O2 þ urea) adduct

or palladium impregnated TS-1 (Pd-TS-1) with H2O2. EPR spectroscopy also

provided evidence, for the first time, for the in situ generation of similar oxo

species in reactions using H2 þ O2 instead of H2O2 as the oxidant. The titanium

sites adjacent to Pd ions (in Pd-TS-1) behave magnetically differently from the

other Ti ions, generating a greater variety of superoxo species. Pd (as expected)

was found to facilitate the reducibility of Ti4þ ions and promoted the formation

of the diverse titanium oxo species at lower temperatures (about 323 K). In the

absence of H2, exposure of TS-1, Ti-MCM-41, Pd-TS-1, or Pt-TS-1 to O2 alone


does not generate the superoxo species. When Pd(Pt)-TS-1 samples were brought

in contact with H2 þ O2, Ti4þ was reduced to Ti3þ by H2 (Fig. 24). The Ti3þ ion

(characterized by its typical EPR spectrum) generates Ti(O2z2) species on

interaction with O2. This reduction and reoxidation of Ti ions, which requires

473 K or higher temperatures in TS-1, is facilitated by Pd or Pt and even occurs at

323 K (Fig. 24). The superoxo species generated are more of A (and A0) types

(Table IV and Fig. 24). The extent of Ti4þ reduction and Ti(O2z2) formation

depends on the Pd content, with the concentration of the paramagnetic titanium

oxo species reaching maximal values at 2 wt% Pd (54).

There has been an attempt to estimate the relative concentrations of the two

superoxo and hydroperoxo species (54) by deconvolution into two bands of the

broad UV–visible band observed on reaction of titanosilicates with aqueous

Fig. 23. EPR spectra (at 210 K) of titanosilicates interacting with aqueous H2O2; the gzz region at

higher gain (£ 5) is shown. The peaks corresponding to A0, A, and B-type Ti-superoxo species are

indicated [(from Srinivas et al. (52)].


H2O2 or non-aqueous urea–H2O2 adducts (Fig. 25). Bands I and II were

attributed to the charge transfer transitions associated with Ti(O2z2) superoxide

and Ti(O2H) hydroperoxo/peroxo species, respectively. The position and relative

intensity of these two bands are different in TS-1 and Pd-TS-1. The intensity

ratio (Ti(O2H))/Ti(O2z2)) was higher for Pd-TS-1 than TS-1. In the spectrum

of Ti-MCM-41, these bands overlapped with those assigned to the H2O2-free

solid. The conversion energy for the hydroperoxo–superoxo transformation was

estimated from the DRUV–visible band positions in (TS-1 þ H2O2), (Pd-TS-

1 þ H2O2), and (TS-1 þ (urea þ H2O2)) to be 38.8, 46.0, and 56.4 kJ/mol,

respectively. At 298 K, for the (TS-1 þ H2O2) system, the Ti(O2H)/Ti(O2z2)

ratio was found to be 0.66.

A comparative value of this ratio was also computed from EPR measurements

(52). The line labeled “theoretical” passing through the origin in Fig. 26 was

computed on the assumption that all the Ti ions in the sample react with H2O2

forming only the paramagnetic superoxo species. The line labeled “experi-

mental” in Fig. 26 shows that the intensity of the EPR signal varies linearly with

Fig. 24. EPR spectra of Ti(O2z2) and Ti3þ ions at 80 K. (a) Pd(2)-TS-2 þ H2O2; (b) Pt(0.015)-

TS-1 þ H2 þ O2 (treated at 673 K); (c) Pd(2)-TS-1 þ H2 þ O2 (treated at 323 K); and (d) TS-

1 þ H2 þ O2 (treated at 673 K). For clarity, spectra (c) and (d) are shown at four and five times the

actual gain. Spectral regions corresponding to Ti(O2z2) and Ti3þ ions are marked [from Shetti et al. (54)].


TABLE IV

EPR parameters (at 77 K) for the superoxo-Ti(IV) species generated on titanosilicates by

contacting with aqueous H2O2 (HP), urea-H2O2 adduct (UHP) and (H2 þ O2)

Systema Species gzz gyy gxx D (cm21)b

TS-1 þ HP A 2.0264 2.0090 2.0023 11203

B 2.0238 2.0090 2.0023 12558

Ti-MCM-41 þ HP B 2.0244 2.0095 2.0031 12217

Pd(2)-TS-1 þ HP A0 2.0309 2.0100 2.0350 9440

A 2.0276 2.0100 2.0350 10672

A00 2.0265 2.0100 2.0350 11157

B0 2.0255 2.0100 2.0350 11638

B 2.0245 2.0100 2.0350 12162

C 2.0220 2.0100 2.0350 13705

TS-1 þ UHP A0 2.0300 2.0101 2.0035 9747

A 2.0275 2.0101 2.0035 10715

B 2.0242 2.0101 2.0035 12329

C 2.0206 2.0101 2.0035 14754

Ti-MCM-41 þ UHP B 2.0232 2.0096 2.0046 12919

TS-1 þ H2 þ O2 A 2.0265 2.0080 2.0010 11157

Ti3þ 1.930 1.956 1.956

Pd(2)-TS-1 þ H2 þ O2 A0 2.0340 2.0092 2.0022 8517

A00 2.0295 2.0092 2.0022 9926

B 2.0241 2.0092 2.0022 12385

Ti3þ 1.928 1.953 1.953

Pt(0.015)-TS-1 þ H2 þ O2 A0 2.0300 2.0080 2.0012 9747

A000 2.0295 2.0080 2.0012 9890

B 2.0241 2.0080 2.0012 12385

Ti3þ 1.931 1.955 1.955

Adapted from Shetti et al. (54).a Pd(2)-TS-1 and Pt(0.015)-TS-1 correspond to TS-1 samples impregnated with 2 wt% Pd and 0.015

wt% Pt, respectively.b D is the energy separation between the oxygen pg

x and pgy orbitals.


Fig. 25. DRUV–visible spectra of TS-1, TS-1 þ H2O2, TS-1 þ urea–H2O2, and Pd(2)-TS-1þ

H2O2. Bands characterizing superoxo (I) and hydroperoxo (II) species are marked. Experimental (—),

simulated (– – –), and deconvoluted oxo-titanium bands (–·–·–) are shown [from Shetti et al. (54)].

Fig. 26. Total EPR signal intensity as a function of Ti content in TS-1 samples [Srinivas et al. (52)].


the Ti content in the various TS-1 samples. This line, however, does not pass

through the origin (Fig. 26). If all the Ti ions in TS-1 had formed the para-

magnetic Ti-superoxo species, the experimental line would have passed through

the origin and coincided with the theoretical line. All the Ti ions in the chosen

samples (Si/Ti ¼ 30, 60, and 80) were isolated and in framework positions

(as shown by XRD, FTIR, and UV–visible analyses). Thus, they are expected to

interact with H2O2 and form either paramagnetic superoxo or diamagnetic

peroxo-Ti species. Consequently, it is concluded that only a fraction of the Ti

ions form paramagnetic superoxo-Ti species and the rest form diamagnetic

hydroperoxo/peroxo-Ti species. From the difference in the theoretical and experi-

mental EPR intensity values (Fig. 26), the amounts of Ti-hydroperoxo and

Ti-superoxo species were estimated to be 45 and 55%, respectively, at 80 K. This

estimate of the (Ti(O2H)/Ti(O2z2) ratio ¼ 45/55 ¼ 0.82 is in reasonable

agreement with the value of 0.66 based on DRUV data.

An additional, independent estimate of the concentration of paramagnetic

superoxo and diamagnetic hydroperoxo-/peroxo-titanium species was made

from magnetic susceptibility measurements using a Lewis coil force magneto-

meter (52). The gram-susceptibility of Ti in TS-1 þ H2O2 was estimated to be

5.5 £ 1026 emu/g, which corresponds to an effective magnetic moment of

0.79 B.M. If all the Ti ions in the sample had formed superoxo species upon

interaction with H2O2, the effective magnetic moment should have been

1.73–1.78 B.M. The concentration of superoxo-Ti species is, thus, about 45%

of the total Ti, comparable to the values found by EPR (55%) and electronic

spectroscopies. The remaining fraction is, presumably, the diamagnetic

hydroperoxo-/peroxo-Ti species.

H2O2 can be a potential source of many radicals (e.g., OH, O2H, etc.).

However, EPR spectroscopy did not reveal the presence of any of these radicals,

indicating that their concentrations are not very significant. They may be highly

unstable. Thus, their contribution to the total magnetic susceptibility

is apparently negligible.

The conversion of hydroperoxide/peroxide to superoxide is a one-electron

redox reaction and requires the presence of transition metals having accessible

multiple oxidation states as in biological iron or manganese clusters (e.g.,

Fe(II, III, IV) clusters of monooxygenase or the Mn(II, III, IV) clusters of

photosystems). Ti is usually not reduced at ambient temperatures. The various

possibilities that could facilitate the transformation of hydroperoxo/peroxo to

superoxo species are as follows:

1. Homolysis of H2O2 to HOz radicals, which react with hydroperoxo-Ti species

to form superoxo-Ti and H2O:

H2O2 ! 2HOz ð2Þ

Ti–OOH þ HOz ! TiðOz22 Þ þ H2O ð3Þ


Formation of HOz radicals by decomposition of H2O2 on contact with titanium

silicates increases with temperature. At 77 K, this decomposition is less

probable.

2. The second possibility is the dismutation of two superoxo ions to yield the

peroxo species.

Oz22 þ Oz2

2 ! O222 þ O2 ð4Þ

Again, even if mobile superoxide ions were present in the material, they would

not be able to diffuse at the low temperatures used for the EPR experiments

(190–77 K).

3. The third possibility for the conversion of the superoxide to the peroxide

is the homolytic opening of a cyclic peroxo species (more precisely,

Ti4þ(O222) to Ti3þ(O2

z2)), as proposed by Notari (33). Formation of Ti3þ

species was indeed observed in the presence of a base, such as NaOH

(spectrum not shown), but in neutral or acidic conditions, the Ti3þ species was

not observed. Either their concentration, if they were formed, was very low or

they were short-lived.

4. The concentration of the Ti(O2z2) species is solvent dependent. Thus, the

solvent (or H2O) may play the role of a redox partner.

The HOz radicals, generated from the decomposition of H2O2, perhaps cause

the hydroperoxo/peroxo to superoxo conversion. The superoxo species (with the

O–O stretching absorption near 1120–1150 cm21) could not be seen in the FTIR

spectrum (63), perhaps because of the dominant stretching and bending modes of

water in the same region.

Although the Ti(O2H) hydroperoxide may be reasonably identified with the

corresponding species derived from infrared–Raman and XAFS spectroscopies

mentioned above, the nature of the paramagnetic superoxide ion-radical,

Ti(O2z2), seen in the EPR spectra, merits more elaboration. Shetti et al. (54)

proposed tentative structures A, B, and C arising from the tetrahedral TiO4 units

upon interaction of the sample with H2O2 (Scheme 3). Species A was postulated

to arise from the framework substitutional sites in the MFI lattice and B and

C from the defect sites. The free O2z2 radical, with a 2P ground state, has a

(1sg)2(1su)2(2sg)2(2su)2(3sg)2(1pu)4(1pg)3 electronic configuration. Interaction

with Ti removes the degeneracy of the HOMO pg into pgx and pg

y orbitals with

an energy gap of D: Neglecting the second-order terms, the g value expressions

(when l , Dp E) may be written as follows (159):

gzz ¼ ge þ 2l=D ð5Þ

gyy ¼ ge þ 2l=E; ð6Þ

andgxx ø ge; ð7Þ


where ge ¼ 2:0023; l is the spin–orbit coupling constant (135 cm21 for oxygen),

and E is the energy separation between 3sg and 1pgx orbitals. The gzz value of the

superoxo anion is sensitive to the oxidation state, coordination number, and local

geometry of the cation to which it is coordinated. (Ti–(O2z2) distances also

influence the gzz parameter. The stronger the Ti–O bond, the lower the gz value of

the superoxo anion. Using the above expressions and the experimental gzz value,

Shetti et al. (54) estimated the separation between the pxg and p

yg orbitals ðDÞ

(Table III). The gzz values of various (Ti–(O2z2)) species decrease in the order

A . B . C. The D (O2z2) values for the A type species lie in the range of 8520–

11,200 cm21. Accordingly, the electron density in the O–O bond increases in the

order A , B , C. Because this electron is added into the antibonding orbital, the

strength of the O–O bond may be expected to decrease in the order A . B . C

(Scheme 3). The O–O bond strength (in the oxo-Ti intermediate) is expected to

play a significant role in influencing the nature of its cleavage (homolytic vs.

heterolytic). Appendix B is a list of some of the major characteristics of the

titanium oxo species generated on TS-1 as a result of contact with H2O2.

IV. Computational Investigations

Significant progress has been made in the last few years in theoretical investi-

gations of the geometry and coordination number of Ti ions in TS-1 and

Ti-MCM-41, both in the dehydrated state and after interaction with H2O2 or

TBHP (48,59–63,103). When such investigations are combined with X-ray

Scheme 3.


absorption, infrared, UV–visible, Raman, and other spectroscopic results

described in Sections II and III, an integrated picture of the structural identity

of the active sites and reactive intermediates involved in the catalytic reactions of

titanosilicates emerges.

The various spectroscopic techniques had revealed that Ti4þ ions in TS-1,

Ti-beta and, Ti-MCM-41 are 4-coordinate in the dehydrated state. Tetrapodal

Ti(OSi)4 and tripodal Ti(OH)(OSi)3 are the main Ti species. Upon exposure

to H2O, NH3, H2O2, or TBHP, they increase their coordination number to 5 or 6.

On samples in which the Ti4þ has been grafted onto the silica (referred to as

Ti " MCM-41), a dipodal Ti species (Ti(OH)2(OSi)2) may also be present. As a

result of interaction with the oxidant ROOH (R ¼ H, alkyl), the formation of

h1- and h2-peroxo (Ti–O–O2), hydroperoxo (Ti–OOH), and superoxo (TiO2z2)

species has been observed experimentally (Section III). A linear correlation

between the concentration of the h2-hydroperoxo species and the catalytic activity

for propene epoxidation has also been noted from vibration spectroscopy (133).

Computational methods, especially DFT, have been used to elucidate the

structure of the oxo-titanium species and their interactions with reactants such

as ethene and NH3 (48,60). From a combined DFT and EXAFS investigation,

Barker et al. (48) recently proposed that 6-coordinate hydrated Ti(h1-OOR)

and (h2-OOR) complexes, where R ¼ H or tert-butyl, are the oxygen-donating

species in peroxide/Ti " MCM-41 mixtures. The computed structural features of

the h1- and h2-species are given in Table V. A schematic illustration of the two

structures in the case of TBHP/Ti-MCM-41 is given in Fig. 27. Figure 28 shows

the calculated energetic pathways from the bare active site and isolated peroxide

to the h1 and h2 reactive oxo-intermediates. The calculated activation barriers are

in each case about 40 kJ/mol. In addition to the monodentate h1-Ti–OOH and

bidentate h2-Ti(O2H) complexes, a third type of oxo-intermediate h1-Ti(O2H2)

complex was also calculated to be feasible. The structures of these three Ti-

peroxo intermediates are shown in Fig. 29. (The calculations were done starting

from the model of the tripodal Ti (Ti(OH)(OSi)3), as this was the predominant

species in Ti " MCM-41. Similar calculations, more realistic for TS-1 and

Ti-beta, starting from the tetrapodal Ti(OSi)4 will be of interest.)

If the h1- and h2-hydroperoxo species are the oxygen-donating entities, the

mode of their interaction with reactants such as alkenes is of interest. Cora et al.

(59) claimed, on the basis of a Mullikan population analysis, that the electron-

rich alkene double bond will preferentially interact with the most electrophilic

oxygen atom, which was identified to be the one closest to Ti in the hydro-

peroxo species (h1-TiOOH), because it has a lower net negative charge.

Following a frontier orbital approach and comparing the energies of the HOMO

and LUMO of the oxo intermediates with that of ethene, the authors found that

for both h1 and h2 structures, the interaction between the LUMO of the catalyst

and the HOMO of the alkene was, as expected, energetically more favorable


TABLE V

Calculated and refined EXAFS parameters for six-coordinate Ti-h 2(OOH) and Ti-h 1(OOH) species in peroxide/surface grafted Ti " MCM-41 mixtures

Cluster Ti–O distance

(A)

Ti–Si distance

(A)

Ti–O–Si (Ti–O–OH)

angle (8)

Eformation

(Calculated)a

(kJ/mol)

R-factor

(EXAFS)

Calculateda EXAFS parameter Calculateda EXAFS parameter Calculateda EXAFS parameter

Ti-h 2(OOH) 1.92 1.91 3.35 3.38 151.3 160 245 16.02

2.25P 2.20 3.32 3.30 145.5 148

1.83Si 1.83Si 3.28 3.21 143.0 139

1.80Si 1.83Si (81.9) (80)

1.80Si 1.83Si

2.26W 2.43

Ti-h 1(OOH) 2.24 2.20 3.31 3.28 145.3 144 2102 16.18

1.97P 1.97 3.34 3.38 146.8 152

1.81Si 1.83Si 3.34 3.39 151.5 163

1.84Si 1.83Si (117.3) (120)

1.81Si 1.83Si

2.35W 2.43

Adapted from Barker et al. (48). Superscript characters: P, Ti–peroxide bond length; Si, Ti–OSi bond length; W, The Ti–O bond distance of Ti to water

molecule. Calculated by the BP86/DZVP procedure employing a larger model cluster extending three-coordination spheres from the central Ti ion.a Eformation ¼ Etotal (“extended” Ti-h 1(OOH) þ other products) 2 Etotal (“extended” tripodal TiIV cluster þ H2O2 þ 2H2O).

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

95

1

than the inverse interaction of the LUMO of the alkene and the HOMO of

the catalyst. Further, the LUMO–HOMO gap for propene was approximately

50 kJ/mol lower than for ethene, suggesting a higher reactivity of propene, as

indeed was observed experimentally (Section V). Figure 30 illustrates this

interaction for the three p-peroxo species. In each case the starting geometries

for modeling were obtained by orienting the ethene molecule so that its HOMO

overlaps with the LUMO of the catalyst. The interaction of ethene with all the

three peroxo species is exothermic.

In the case of the side-bound h2 intermediate, the interaction was initiated

(in the calculations) by positioning the double bond parallel to the peroxide

Fig. 27. Ti-peroxo species in TBHP/Ti " MCM-41 catalysts. All distances (DFT calculated values

and experimental parameters (in parentheses)) shown are in A [Reproduced from Barker et al. (48) by

permission of the PCCP Owner Societies].

Fig. 28. Calculated energetic pathways from the bare active site and isolated peroxide to

the h1- (left) and h2- intermediates (right) [from Cora et al. (59)].


molecule, because the OH ligand hinders other directions of attack of ethene

molecule on the peroxidic oxygen closest to Ti. Optimization of this structure

leads to an alcohol-type functionality (Fig. 30b), which the authors suggested

(59) to be possibly responsible for the formation of the diol products observed

experimentally (Section V).

The Ti4þ distribution in TS-1 has also been studied by computational methods

(34,62,160–163). The actual location of the Ti atoms in the framework of

titanosilicates is difficult to determine experimentally because of the low Ti

content (Section II), and information obtained from theoretical methods is,

therefore, of considerable interest. In the orthorhombic MFI structure, substi-

tution can take place at 12 crystallographically different tetrahedral (T) sites

(T1–T12) (Fig. 1 and Section II.A.1.b). In the monoclinic MFI framework,

the mirror symmetry is lost and 24 crystallographically different T sites can be

distinguished (Fig. 31) (160).

Although all computational investigations that have been reported confirm

that Ti atoms are incorporated in the framework at regular Ti-sites, there is still

controversy about the exact siting of the Ti atoms in the MFI structure. De Man

and Sauer (62) by ab initio investigations found only small subsitution energy

differences among the various T sites, and this result implies that Ti atoms

are distributed over all the lattice positions rather than being located at one

preferred T-site. Using a combination of Metropolis Monte Carlo method and

molecular mechanics calculations, Njo et al. (160) concluded that the Ti atoms

are indeed distributed over all the crystallographically different lattice positions

rather than located at one preferred site. The distribution, however, is not equal

or random. In Fig. 32 the Ti occupancies per unit cell for the orthorhombic and

monoclinic structure are shown (160). In the orthorhombic structure, T12 is

preferred, whereas in the monoclinic structures T2 is preferred. The framework

symmetry (orthorhombic/monoclinic) is apparently related to both the location

of the Ti atoms and the Ti loading. Njo et al. (160) also computed the occupancy

of the different T sites at different loadings (Fig. 33). At all Ti loadings up

Fig. 29. Geometry-optimized structure of the three stable Ti-peroxo intermediates: (a) h1-

monodentate complex, (b) h2-bidentate complex, and (c) h1-O2H2 complex [from Cora et al. (59)].


to 2.5 Ti atoms per unit cell, the experimentally determined upper limit for

incorporation of Ti in lattice positions, the T2 and T12 sites were preferred. A

200-atom cluster study of Ti-siting in TS-1 by Atoguchi and Yao (162) using the

ONIOM method (164), however, suggested that the most stable Ti substituted

Fig. 30. Calculated initial and final states for the interaction of “ethene” with (a) theh1-, (b) theh2-,

and (c) h1-O2H2 Ti-peroxo intermediate [from Cora et al. (59)].


T sites were T9 and T10 sites—if thermodynamics controls the structure of

Ti-containing MFI zeolite. The stability sequence of T sites was found to be

T9 . T10 . T12 . T1 . T6 . T5 . T3. The exact location of Ti ions in TS-1

is still controversial. There are no similar investigations for other Ti silicates.

V. Catalytic Properties

V.A. Introduction

The catalytic activity of the titanosilicate molecular sieves, especially those

of TS-1, TS-2, Ti-beta and Ti-MCM-41 has been investigated extensively

Fig. 32. Ti distribution per unit cell over the crystallographically different T-sites for the ortho-

rhombic structure (T1–T12, white) and monoclinic (T1–T12, stripes; T13–T24, black) structures

[Reprinted from Njo et al. (160) with permission. Copyright (1997) American Chemical Society].

Fig. 31. Crystallographically different T-sites in MFI. T1 (T2,…,T12) and T13 (T14,…,T24) are

related by a mirror plane in orthorhombic MFI [Reprinted from Njo et al. (160) with permission.

Copyright (1997) American Chemical Society].


(33,165–169). When a tetravalent ion, such as Ti4þ, replaces, the Si4þ in a

silicate lattice isomorphously, the generation of Brønsted acidity is not anti-

cipated. In fact, no experimental evidence exists for a purely Brønsted acid-

catalyzed reaction in a well-synthesized and pure sample of TS-1 and in the

absence of H2O2. Lewis acid-catalyzed reactions can, of course, occur because of

the coordinatively unsaturated Ti ions, as mentioned above (Section II.B).

The enormous interest in these materials is, however, due to their remarkable

catalytic activities in oxidation reactions using the environmentally benign

aqueous H2O2 as the oxidant.

V.B. Reactions Using H2O

2as Oxidant

V.B.1. General Features

Oxidations of organic reactants using H2O2 as an oxidant have been known for a

long time (170). Although H2O2 is a weak acid ðpKa ¼ 11:6Þ and a mild oxidant,

a small amount of HOþ may be present in equilibrium with H2O2 solutions,

especially at low pH:

H2O2 þ HþO H2O þ HOþ ð8Þ

The major use of H2O2 as an oxidant arises from its ability to insert an oxygen

atom in an organic molecule (alkene, alkane, aromatic hydrocarbon, etc.) in

the presence of some catalysts. In reactions using H2O2 as an oxidant, the type of

Fig. 33. Population of crystallographically different T-sites for various Ti loadings: (i) one Ti atom

per unit cell (white), (ii) one Ti atom per double unit cell (dotted), and (iii) eight Ti atoms per unit cell

(striped) [Reprinted from Njo et al. (160) with permission. Copyright (1997) American Chemical

Society].


cleavage of the O–O bond (in H2O2) plays a crucial role in determining the

product distribution. A homolytic cleavage generating radicals (such as HOz)

usually leads to a product distribution different from the one that arises by

heterolytic cleavage (generating HOþ and HO2, for example). The gas-phase

dissociation enthalpy, DH0; for O–O homolytic cleavage in H2O2 is 205 kJ/mol

(171). The O–O bond is considerably weakened if H is replaced by electron-

donating alkyl groups as in ROOH (R ¼ alkyl), the bond dissociation enthalpy

being only 180 kJ/mol for the homolytic cleavage of the O–O bond in CH3OOH

(171). A heterolytic cleavage of the O–O bond, HOOH ! HOþ þ HO2,

requires a considerably higher dissociation enthalpy if the emerging ions are

not stabilized. The enthalpy for the heterolytic O–O cleavage of H2O2 into

HOþ and HO2 is 1252 kJ/mol (171) in the gas phase. The corresponding value

for CH3O–OH ! CH3Oþ and HO2 is 775 kJ/mol. The situation is, however,

different in solution. Heterolytic cleavage requires less energy if the dissociated

ions form an ion pair in solution at a distance less than rip; separating the effective

charge centers. Then, the energy of heterolytic cleavage in a solvent, DEip

(solvent) is given (171) by Eq. (9)

DEipðsolventÞ ¼ DE 2 e2=rip 2 DEsolv; ð9Þ

where DE is the energy required for gas-phase heterolytic cleavage, rip ¼ 2:65 �A

(172,173), and DEsolv is the solvation energy given by

DEsolv ø 14:39ðð12 1Þ=ð21þ 1ÞÞm2=a3; ð10Þ

1 is the dielectric constant and m the dipole moment of the solvent, and, a is

the radius of a spherical cavity formed by solvent molecules surrounding the

ion pair.

With a ¼ 3:5 �A (173), the solvation energy of a typical hydrocarbon solvent

ð1 ¼ 2Þ is about 45 kJ/mol (171). This energy will increase if the dielectric

constant of the solvent is higher. Hence, as the dielectric constant/dipole moment

of the solvent is progressively increased, the heterolytic fission of the O–O

bond (in H2O2, TBHP, etc.) will be favored over homolytic fission. Because the

latter generates radical intermediates and the heterolytic fission produces ionic

products, it is likely that the oxidation reaction mechanism and product dis-

tribution will depend to some extent on the choice of the solvent, as indeed has

been observed experimentally (vide infra). Homolytic decomposition increases

at higher temperatures, especially temperatures above about 333 K. Radical

pathways, hence, play a greater role in influencing product selectivity at higher

temperatures and in non-polar solvents.


V.B.2. H2O2-Catalyzed Reactions in the Homogeneous Phase

Reactions with H2O2 may be divided into two classes arising from the homolytic

vs. heterolytic cleavage of the O–O bond (173). In homolytic catalysis, the

oxygen-centered radicals are intermediates; the participation of concerted

processes in heterolytic catalysis precludes paramagnetic intermediates. Product

selectivity is usually higher in the latter class. Transition metal cations in low

oxidation states, such as Cu1þ, Ti3þ, V2þ, Cr2þ, and Fe2þ, catalyze the homolytic

route, although those in higher oxidation states, such as Mo6þ, W6þ, V5þ, and

Ti4þ, catalyze the heterolytic cleavage.

The one-equivalent, homolytic scission of peroxides may be either reductive

(Eq. (11)) or oxidative (Eq. (12)):

HOOH þ Mnþ ! HOz þ HO2 þ Mðnþ1Þþ ð11Þ

HOOH þ Mnþ ! HOOz þ Hþ þ Mðn21Þþ ð12Þ

An alternate homolytic cleavage is the following:

HOOH ! 2HOz: ð13Þ

The reductive cleavage (Eq. (11)) is more common. TBHP can also undergo

preferential reductive cleavage to the alkoxyl radical:

ROOH þ Cu1þ ! ROz þ Cu2þðOHÞ: ð14Þ

The oxidative cleavage may be illustrated as follows:

ROOH þ Co3þ ! ROOz þ Co2þ þ Hþ: ð15Þ

Hydroxy radicals are intermediates in the reaction of Ti3þ and H2O2 (175). This

system is also capable of hydroxylation of aromatics and alkanes but, in contrast

to reactions with Fenton’s reagent (Fe2þ þ H2O2, reductive, homolytic cleavage,

Eq. (11)), only non-chain processes are possible, because Ti4þ is not usually an

oxidant. Hence, relatively high selectivities are feasible.

Heterolytic catalysis is promoted by W6þ, Ti4þ, Cr3þ, V5þ, and many Mo6þ

complexes. These complexes do not normally react with peroxides. However, in

the presence of electron-rich molecules, such as alkenes, amines, sulfides, etc.,

oxygen insertion in the reactant occurs. For example,

M– ðROOHÞ þ alkene ! ROH þ epoxide; M ¼ Mo;Cr;V;Ti;W ð16Þ

These catalytic reactions are distinguished from the homolytic reactions in that

no evidence exists for paramagnetic intermediates. The epoxidation is stereo-

specific, trans- and cis-alkenes yielding trans- and cis-epoxides, respectively.

Under the same conditions, complexes of Cu, Mn, and Fe give no yields or


poor yields of epoxides because they decompose ROOH rapidly into radicals.

High yields of epoxides and, especially, the stereospecificity of the reaction are

compatible only with a heterolytic mechanism in which the active epoxidizing

agent delivers an electrophilic oxygen species from a hydroperoxide-metal

complex to the reactant in a concerted manner; there is no free rotation of the

C–C bond during this process. The high yields of epoxides in one case (Mo6þ,

V5þ, Cr6þ, and Ti4þ) and the low yields in the other case (Fe2þ, Cu1þ, Co2þ,

Cr2þ) suggest that the epoxidation of the alkene by heterolytic cleavage and

oxygen insertion and the homolytic decomposition of ROOH (R ¼ H, alkyl)

are competing processes (176). The selectivity to epoxide is determined by the

relative rates of reaction of the catalyst-hydroperoxide complex with the alkene

(Eq. (16)) in competition with its homolytic decomposition (Eq. (12)). The

oxidation potential of the metal ion (in the complex) and its Lewis acidity may be

expected to influence the relative rates of Eqs. (12) and (16). The redox potentials

of some transition metals are given in Table VI; the heterolytic pathway is likely

to be preferred for reaction on Ti4þ-silicalite.

In the epoxidation step (Eq. (16)), the main function of the catalyst is to

withdraw electrons and reduce the electron density at the peroxide O–O bond,

making it more susceptible to attack by nucleophiles such as alkenes. In this

process, the M ion acts as a Lewis acid. Active epoxidation catalysts are

usually strong Lewis acids and relatively weak oxidants in their highest

oxidation state (to avoid one-electron oxidative decomposition of the peroxide

as per Eq. (13). (177). The Lewis acidity of M, in turn, is influenced by its

coordinating ligands. The hetero- vs. homolytic O–O cleavage is also affected

by the substituent on the hydroperoxide; electron-donating tert-alkyl groups on

the peroxide moiety tend to favor the homolytic cleavage of the O–O bond,

whereas electron-withdrawing substituents such as acyl groups facilitate O–O

bond heterolysis. In other words, homolytic O–O bond cleavage is facilitated

when more electron density resides on the O–O bond of the M–OOR (R ¼ H,

alkyl) intermediate.

TABLE VI

Redox potentials of transition metal ions in aqueous solutions

Reaction E0 (V) Reduction H2O2 decompostion

Co(III) þ e ! Co(II) þ1.82 Easy Fast

V(V) þ e ! V(IV) þ1.00 Moderate Moderate

Fe(III) þ e ! Fe(II) þ0.77 Moderate Moderate

Ti(IV) þ e ! Ti(III) 20.37 Difficult Difficult


In the field of enzyme catalysis, heme-proteins such as cytochrome P450, for

example, exhibit both types of O–O bond cleavages in organic hydroperoxides

and peroxy acids (178). Heterolytic cleavage of HOOH/ROOH yields H2O

or the corresponding alcohol, ROH and a ferryl-oxo intermediate (Scheme 4).

Homolytic O–O bond cleavage results in the formation of a hydroxyl (HOz) or an

alkoxyl (ROz) radical and an iron-bound hydroxyl radical.

V.C. Epoxidation on Titanium Silicate Molecular Sieves

V.C.1. General Features of Epoxidations

Epoxidation reactions in the liquid phase have been reviewed by Sawaki (179)

and more recently by Arends and Sheldon (180), and those occurring in the

presence of solid catalysts by Dusi et al. (181). Because H2O2 is only a mild

oxidant, its use in alkene epoxidation requires the application of appropriate

catalysts. The catalytic epoxidation using H2O2 and tungstic acid, for example,

proceeds via the formation of peroxytungstic acid. Aqueous conditions are

usually not appropriate for epoxidations, because epoxides are prone to undergo

acid-catalyzed hydrolysis. In alkene epoxidation with alkyl hydroperoxides

catalyzed by various metal complexes of Ti, Mo, and V in the liquid phase, two

alternate pathways, A and B in Scheme 5, each involving a metal alkyl peroxide

complex, have been accepted in the literature (182). Mechanism A involves

an electrophilic O transfer to alkene. Mechanism B involves a five-membered

dioxametallocyclopentane. For the particular case of vanadium, the alkylperoxy

Scheme 4.


complexes were isolated and pathway B was supported by the fact that the

relative rates were correlated with the coordinating ability of alkenes.

The operating pathway seems, however, to change as a result of changes in

the metals, ligands, and solvents (182). Early transition metals, such as Ti, for

example, seem to prefer path A (182). Prior to the discovery of TS-1, amorphous

Ti–SiO2 was the best known solid catalyst for the epoxidation of propene (183)

using alkyl hydroperoxides, offering an alternate route to the homogeneous

catalytic Halcon/ARCO process (184). However, the catalyst was unstable in the

presence of H2O.

In the overall reaction, ethylbenzene and propene are converted with oxygen to

styrene, propene oxide, and H2O. The epoxidizing agent is ethylbenzene hydro-

peroxide. Sheldon et al. (185) attributed the catalytic activity to site isolation of

Ti4þ on the silica surface, preventing the formation of TiO2 domains, and to the

enhanced Lewis acidity of Ti4þ resulting from electron withdrawal by the Si–O-

ligands. The reaction mechanism is assumed to involve the Ti-alkyl peroxo

groups (Ti–OOR).

Propene oxide is also manufactured by the chlorhydrin route (186):

CH3 –CHyCH2 þ HOCl ! CH3 –CHðOHÞ–CH2Cl; ð17Þ

CH3 –CHOH–CH2Cl þ base ! CH3CHðOÞCH2 þ baseðHClÞ: ð18Þ

Scheme 5.


The chlorhydrin route is also used in the manufacture of epichlorohydrin from

allyl chloride (187):

CH2yCH–CH2Cl þ HOCl ! CH2ðOÞCH–CH2Cl þ HCl: ð19Þ

The direct conversion of propene to its epoxide, in near quantitative yields,

with aqueous H2O2 will be environmentally more benign. One of the unique

features of TS-1 as a solid oxidation catalyst is its ability to utilize aqueous

H2O2 as the oxidant for such conversions. This ability of TS-1 derives from

the fact that silicalite-1 is hydrophobic, in contrast to the hydrophilic amor-

phous Ti–SiO2. Consequently, hydrophobic reactants, such as alkenes, are

preferentially adsorbed by TS-1, thus precluding the strong inhibition by H2O

observed with amorphous Ti–SiO2.

Unfortunately, for economic reasons and in the absence of compelling

environmental legislation, the process for manufacture of propene oxide using

TS-1 and H2O2 is not very attractive and is not yet in commercial practice.

Worldwide efforts are underway to develop this process by using H2O2

generated in situ (from H2 þ O2) or (secondary/tertiary alcohol þ O2). Metal-

loaded TS-1 structures are the likely catalysts (Section V.C.16). Titanosilicate

molecular sieves, especially those with large pores and mesopores, however,

offer great potential in the fine chemicals industry (for manufacture of drug

intermediates, fragrances, agrochemicals, etc.), as the reactant molecules are

larger and the economics allows the use of the more expensive H2O2 as the

oxidant. Most of these large-pore and mesoporous materials need to use the alkyl

hydroperoxides (such as TBHP) rather than aqueous H2O2 as the oxidant (see,

however, Section V.F).

V.C.2. Yields and Stereospecificities

Lower alkenes such as ethene, propene, and butenes are epoxidized in high

yields (.95%) in the presence of TS-1 catalyst by aqueous H2O2 (33). The

stereochemical configuration is retained in the case of butenes; cis-but-2-ene

gives exclusively the cis-epoxide, and trans-but-2-ene gives exclusively the

trans-epoxide. These high epoxide yields and retention of stereochemical

configuration argue against the homolytic decomposition of the O–O bond of the

Ti(O2H) intermediate and support a heterolytic mechanism.

V.C.3. Diffusional Constraints

As expected, although TS-1 is more active and selective in the epoxidation of

linear alkenes (such as hex-1-ene and dodec-1-ene), the large-pore Ti-beta is

more active in the case of the bulkier cyclohexene (TON of 14 vs. 1 for TS-1) and

cyclododecene (TON of 20 vs. 5; Table VII) (11).


TABLE VII

Diffusional constraints in selective oxidation of alkenes over Euro-TS-1 and Ti-Beta

Alkene Catalyst Reaction

time (h)

Turnover

(mol/mol Ti)

H2O2 Product selectivity

(%)

Glycol

ethers

Conversion (%) Selectivity (%) Epoxide Glycol

Hex-1-enea TS-1b 3 50 98 80 96 – 4

Ti-betac 3 12 80 80 12 8 80

Cyclohexenea TS-1 3 1 – – 100 – –

Ti-betad 3.5 14 80 83 – – 100

Dodec-1-enee TS-1 3.5 110 83 68 77 23 –

Ti-betaf 3.5 87 80 87 – 100 –

cyclododecenee TS-1 4 5 26 26 66 34 –

Ti-betaf 3.5 20 47 71 80 20 –

Adapted from Corma et al. (11).a Reaction condition: catalyst, 0.2 g; alkene, 33 mmol; H2O2/alkene (mol) ¼ 0.082; solvent (methanol), 23.57 g; temperature ¼ 333 K, tr ¼ 4 h.b Euro-TS-1 (1.7 wt% of Ti given as TiO2).c Ti-beta (Ti/(Ti þ Si) ¼ 0.044, TiO2 (wt%) ¼ 5.7, TiO2/Al2O3 ¼ 244).d Ti-beta (Ti/(Ti þ Si) ¼ 0.040, TiO2 (wt%) ¼ 5.2, TiO2/Al2O3 ¼ 210).e Reaction condition: catalyst, 0.2 g; alkene, 33 mmol; H2O2/alkene (mol) ¼ 0.258; solvent (ethanol), 23.57 g; temperature ¼ 353 K; tr ¼ 4 h. Some

oxidation of ethanol was observed at these reaction conditions, which was taken into account to calculate H2O2 conversion and selectivity. H2O2

selectivity(%) ¼ (mol alkene oxidized/mol H2O2 converted) £ 100.f Ti-beta (Ti/(Ti þ Si) ¼ 0.018, TiO2 (wt%) ¼ 2.4, TiO2/Al2O3 ¼ 111).

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

96

3

The influence of catalyst particle size and morphology in phenol hydroxylation

is shown in Table VIII and confirms the diffusional constraints in this

reaction also.

A novel strategy for overcoming the diffusional limitations associated with

the pore size of TS-1 without sacrificing the advantages of its hydrophobicity

was demonstrated by Schmidt et al. (188). These authors impregnated a sample

of carbon black (particle diameter 18 nm) with a clear solution of tetrapropyl-

ammonium hydroxide, water, and ethanol. After evaporation of the ethanol, the

carbon particles were impregnated with a 20% excess (relative to the incipient

wetness value) of a mixture of tetraethyl orthotitanate and tetramethylortho-

silicate. The composition of the resultant synthesis gel was 20 TPA2O:TiO2:

100SiO2:200H2O, and the resultant zeolite concentration was about 20%. TS-1

was then obtained by conventional hydrothermal synthesis from this inorganic

gel–carbon matrix system. Finally, carbon was removed by calcination at

823 K. The resulting sample of TS-1 had a Si/Ti atomic ratio of 110, a high

crystallinity, and an average crystallite size of about 1.5 mm, and it exhibited

mesoporosity (about 20 nm in diameter dispersed throughout the crystal).

The advantage of this “mesoporous” TS-1 over samples prepared by the

conventional route is illustrated in Fig. 34. The two samples behave similarly for

the oxidation of linear reactant oct-1-ene. But a marked difference was observed

for the oxidation of bulkier cyclohexene. Because of the absence of diffusional

constraints, the catalytic epoxidation activity in the “mesoporous” TS-1 enhanced

by almost an order of magnitude for the oxidation of the bulkier cyclohexene.

TABLE VIII

Influence of textural properties of TS-1 samples on phenol hydroxylation activity

Sample Average particle

sizea (mm)

Morphologya R0b Conversionc

(%)

Selectivityd

(%)

Yielde

(%)

1 0.2 Cubic 10.2 50 95 93

2 0.3 Cauliflower 9.00 44 93 92

3 5.0 Coffins 1.07 6 15 40

4 10.0 Coffins 0.46 2.5 8 18

Adapted from van der Pol et al. (89). Reaction conditions: catalyst, 0.5 g; phenol, 10 g; solvent

(acetone), 10 mL; 35 wt% H2O2, 2 mL (added at the beginning of the reaction); temperature ¼ 353 K.a Estimated using SEM.b R0 ¼ initial reaction rate of dihydroxy benzene formation (mol/m3 s).c Conversion ¼ H2O2 conversion at t ¼ 1 h.d Selectivity ¼ (moles dihydroxybenzene/moles of reacted H2O2) £ 100% at t ¼ 1 h.e Yield ¼ (moles dihydroxy benzene/moles of H2O2 added) £ 100% at complete H2O2 conversion.


V.C.4. Influence of Ti-Silicate Structure

The greater activity of Ti-beta (vs. TS-1) in the oxidation of the bulky

cyclohexane was noted in the previous section. Table IX provides a comparison

of the conversion and epoxide selectivity in the reaction catalyzed by TS-1 and

three large-pore/mesoporous Ti-silicates in the epoxidation of a single, linear

allyl alcohol (pentenol).

Fig. 34. Ratio of product concentrations [sum of epoxide and secondary products; (a) from oct-1-

ene and (b) from cyclohexene] obtained with mesoporous and conventional TS-1 as a function of the

contact time. The results show that the mesoporous TS-1 has a similar activity for oct-1-ene epoxidation

as conventional TS-1. However, the mesoporous TS-1 is significantly more active for cyclohexene

epoxidation [Reproduced from Schmidt et al. (188) by permission of the Royal Society of Chemistry].

TABLE IX

Influence of titanosilicate structure on epoxidation of pentenol with H2O2

Catalyst Temperature (K) Pentenol conversion (%) Epoxide selectivity (%)a Reference

Ti-MCM-41 323 32 19 (81) (273)

Ti-MCM-48 323 32 21 (79) (273)

Ti-beta 343 42 89 (11) (195)

TS-1 323 ndb 76 (24) (193)

a Numbers in parentheses indicate the selectivities to the corresponding unsaturated carbonyl

compounds.b nd, no data available.


The higher conversion in the presence of Ti-beta is probably a result of the

higher temperature (343 vs. 323 K). Diffusional constraints cannot account for

the observed differences in selectivity. Ti-beta and TS-1 are distinctly more

selective than the mesoporous material. Recalling that tetrapodal titanium sites

are more predominant in the former two molecular sieves although tripodal

titanium sites are the major surface species over the latter mesoporous material

(Section II), we infer that the data indicate that high epoxidation selectivity

is probably correlated with the presence of tetrapodal structures in these two

molecular sieves. This correlation is discussed in Section VI.

The epoxidation of hex-1-ene catalyzed by Ti-beta samples synthesized

in the conventional, basic medium (Ti-beta(OH)) is compared in Table X

with that catalyzed by a sample synthesized in a fluoride-containing medium

(Ti-beta(F)) (13). The latter was more hydrophobic. Results for the reaction

catalyzed by TS-1 are also included in Table X. Ti-beta(F) is superior to TS-1

for reaction in acetonitrile solvent. The most significant difference between

Ti-beta(F) and Ti-beta(OH) is in their selectivities. Although the selectivity to

the epoxide for reaction in acetonitrile is always very high, regardless of the

zeolite; for reaction in methanol, Ti-beta(F) is more selective than Ti-beta(OH)

(76.6 vs. 54.9%, Table X). Both Ti-beta samples are, however, less selective

than TS-1 for reaction in methanol.

The lower activity of Ti-beta(OH) in the epoxidation of an alkene containing

a polar head (oleic acid, Table XI) was attributed by Blasco et al. (13) to the

different adsorption properties of the two catalysts. A strong adsorption of oleic

acid through the polar head on the relatively more hydrophilic Ti-beta(OH)

TABLE X

Epoxidation of hex-1-ene catalyzed by Ti-containing zeolites: influence of method of preparation

Catalyst TiO2 (wt%) Solvent Hex-1-ene

conversiona

Epoxide

selectivity (%)

H2O2

selectivity (%)

TONb

Ti-beta(F) 2.86 CH3CN 41.2 100 99.7 43.1

Ti-beta(OH) 2.78 CH3CN 40.3 100 76.6 53.4

TS-1 2.18 CH3CN 25.5 100 76.5 39.1

Ti-beta(F) 2.86 CH3OH 26.8 76.6 97.9 30.7

Ti-beta(OH) 2.78 CH3OH 25.4 54.9 90.1 24.2

TS-1 2.18 CH3OH 46.6 97.6 96.7 94.5

Adapted from Blasco et al. (13). Reaction conditions: catalyst, 0.1 g; hex-1-ene, 16.5 mmol; solvent,

11.8 g; H2O2, 4.1 mmol; temperature ¼ 323 K; time ¼ 2 h.a Percentage of maximum.b Initial turnover number (moles of converted alkene/moles of Ti £ hours).


would make the oxidation of the double bond in the middle of the hydrocarbon

chain more difficult.

V.C.5. Epoxidation Catalyzed by Mesoporous Titanium Silicates

Although the mesoporous materials, such as Ti-MCM-41, have lower intrinsic

epoxidation selectivity than TS-1 and Ti-beta, they must nevertheless be used as

catalysts for reactions of large molecules typical in the fine chemicals industry.

It is, therefore, interesting to elucidate how these ordered mesoporous materials

compare with the earlier generation of amorphous titania–silica catalysts.

Guidotti et al. (189) recently compared Ti-MCM-41 with a series of amorphous

titania–silica catalysts for the epoxidation of six terpene molecules of interest in

the perfumery industry (Scheme 6). Anhydrous TBHP was used as the oxidant

because the catalytic materials are unstable in water. The physical character-

istics of these catalysts are compared in Table XII.

It was observed that no leaching of Ti occurs during the catalytic reaction

in the anhydrous medium. The acidity of the catalysts (which gave rise to many

side products) was evaluated by a comparison of their reaction rates in the

acid-catalyzed conversion of citronellol into isopulegol (Scheme 7). The acidity

of the catalysts decreased in the following order: A . C . D . B ø E. The

catalytic activity and epoxidation selectivities are compared in Table XIII.

TABLE XI

Epoxidation of oleic acid over Ti-beta prepared in fluoride (F) and alkali (OH) medium

Catalyst TiO2 wt % Acid conversiona Epoxide selectivity H2O2 selectivity

Ti-beta(F) 2.52 31.2 100 67.6

Ti-beta(OH) 2.78 20.2 100 24.8

Adapted from Blasco et al. (13). Reaction conditions: catalyst, 30 mg; oleic acid, 1 mmol; CH3CN,

2 mL; H2O2, 0.25 mmol; temperature, 323 K; time, 8 h.a Percentage of maximum.

Scheme 6.


The results led to the following conclusions:

1. With regard to the specific activity, the mixed oxide catalyst, E, showed the

best performance of all reactants, 1 and 6 being exceptions. For the latter,

A and B performed better.

2. Epoxidation of alkeneic reactants is faster on titanium-grafted silicates (such

as A, B and C) than on the coprecipitated titanosilicates (such as D and E).

This difference was attributed to the fact that on extra-framework titanium-

grafted silicates, the catalytically active sites are virtually all exposed and

accessible, whereas on the coprecipitated material some of them may be

buried within the silicate walls and, thus, cannot adsorb reactant molecules.

3. Most of the side product formation was caused by the oxidation of the

alcohol function, as expected.

4. When the OH group in the reactant is absent or far from the double bond

(reactants 6 and 1, respectively), the Ti-grafted materials displayed the

best activity values. When the OH group is in the proximity of the CyC

TABLE XII

Ti composition and textural characteristics of titanium silicates

Sample Method of preparation Ti loading,

calcined

samples

(wt%)

Specific

surface

area

(m2/g)

Total pore

volume

(mL/g)

Mean pore

diameter

(nm)

(A) Ti-MCM-41 Ordered, Ti-grafted, mesoporous

silica

1.88 861 0.53 2.4

(B) Ti-SiO2 Amorphous, Ti-grafted, porous

silica (Grace Davison 62)

1.75 303 1.10 12.8

(C) Ti-SiO2 Ti-grafted silica (Aerosil 380,

Degussa)

1.78 268 nd nd

(D) MST Amorphous, mesoporous titania-

silica (co-precipitation)

1.84 454 0.38 4.6

(E) TiO2-SiO2 Commercial, amorphous, porous

mixed oxide (Grace)

1.40 303 1.16 12.7

Adapted from Guidotti et al. (189); nd, not determined.

Scheme 7.


bond, the promotion effect of the OH group (hydroxyl-assisted epoxidation,

see Section V.C.9) prevails and the differences in activities between the

various catalysts become smaller.

5. The epoxide selectivity did not depend noticeably on the gross structural

features of the catalyst. For instance, the selectivity in the epoxidation of 4 is

about 85% on all solids (Table XIII).

6. As long as the pore diameters are large enough for easy entry and exit of

reactant and product molecules, the catalyst porosity features do not have a

significant influence on the epoxidation activity. In a comparison between

two epoxidation catalysts obtained by grafting Ti(iso-PrOi)4 on MCM-41

and an amorphous silica gel, respectively, the former showed a lower

activity (189).

7. A significant absorption band in the 300–350 nm region of the DRUV

spectra indicated that samples B and C, which contained significant amounts

of Ti–O–Ti oligomeric sites in octahedral coordination (Fig. 35), have good

catalytic activity.

The authors postulated that on these materials “complete site isolation is not

mandatory in order to have active and selective titania–silica epoxidation

catalysts”. The 100% selectivity of the dinuclear, silica-supported

TABLE XIII

Comparative catalytic activities (turnover numbers and selectivity (in parentheses))

of ordered Ti-MCM-41 (A) and amorphous titania–silica (B–E) catalysts in the

epoxidation of unsaturated cyclic terpenes (1–6) using anhydrous TBHP

Terpenes Catalyst

A B C D E

1 44a (51)b 37 (60) 29 (57) 22 (58) 28 (53)

2 38 (61) 37 (80) 31 (88) 23 (65) 44 (90)

3 43 (64) 44 (84) 40 (81) 40 (74) 59 (71)

4 36 (80) 38 (82) 32 (88) 19 (84) 45 (89)

5 40 (73)c 45 (84) 43 (83) 30 (83) 52 (75)

6 30 (90)c 33 (89) 32 (92) 19 (85) 25 (75)

Adapted from Guidotti et al. (189). Reaction conditions: catalyst, 50 mg; substrate,

1 mmol; TBHP: terpene (mol) ¼ 1:1; solvent, CH3CN; VTOT mix., 10 mL; temperature,

363 K; time, 24 h; magnetic stirring (ca. 800 rpm). Textural properties of the catalysts (A–

E) are given in Table XII. Structures of the substrates (1–6) are shown in Scheme 6.a TON, turnover number after 24 h ([mol converted terpene]/[mol Ti]).b Selectivity to monoepoxide after 24 h (%).c Selectivity to endocyclic monoepoxide after 24 h (%).


(xSiO)2TiOTi(OO-t-Bu)4 species, prepared by the grafting route, in the epoxida-

tion of cyclohexene (190) was cited as additional support for the above argument.

V.C.6. Influence of Alkene Structure

Epoxidation of alkenes with terminal CyC bonds is faster than that of alkenes

with internal CyC bonds when the reaction is catalyzed by TS-2 (Table XIV).

Fig. 35. Diffuse reflectance UV–visible spectra of Ti-MCM-41 (A), Ti–SiO2 Davison (B), Ti–

SiO2 Aerosil (C), MST (D), and TiO2–SiO2 Grace (E) [from Guidotti et al. (189)].

TABLE XIV

Epoxidation of various alkenes over TS-2: influence of alkene structure

Hex-1-ene Hex-2-ene Hex-3-ene Oct-1-ene Dodec-1-ene Cyclohexene

Conversion

(mol%)

92.0 81.2 72.0 56.4 28.8 40.2

Epoxide selectivity

(%)

73.5 69.0 76.5 66.3 50.0 54.3

Adapted from Kumar et al. (165). Reaction condition: catalyst (TS-2; Si/Ti ¼ 29), 0.1 g; reactant,

1.0 g; H2O2/substrate ¼ 1.1; solvent (CH3CN), 10 g; temperature, 333 K; time, 6 h.


The rate also decreases with an increase in the chain length of the alkene

molecule (hex-1-ene . oct-1-ene . dodec-1-ene). Although the latter phenom-

enon is attributed mainly to diffusion constraints for longer molecules in the

MFI pores, the former (enhanced reactivity of terminal alkenes) is interesting,

especially because the reactivity in epoxidations by organometallic complexes

in solution is usually determined by the electron density at the double bond,

which increases with alkyl substitution. On this basis, hex-3-ene and hex-2-ene

would be expected to be more reactive than the terminal alkene hex-1-ene.

The reverse sequence shown in Table XIV is a consequence of the steric

hindrance in the neighborhood of the double bond, which hinders adsorption

on the electrophilic oxo-titanium species on the surface. This observation

highlights the fact that in reactions catalyzed by solids, adsorption constraints

are superimposed on the inherent reactivity features of the chemical reaction

as well as the diffusional constraints.

The epoxidation rates of various alkenes relative to hex-1-ene on Ti-beta

with H2O2 and TBHP are summarized in Table XV. In the absence of diffusional

constraints, the branched alkenes are more reactive than the linear ones (see also

Section V.C.13).

V.C.7. Dialkenes

Selective epoxidation of one of the double bonds in dialkenes is of practical

interest (Table XVI). Although monoepoxides predominate at low H2O2 con-

centrations, the diepoxides are also formed at higher concentrations. The diallyl

epoxides of bisphenol A are major intermediates in the adhesives industry, and

their synthesis in solid-catalyzed reactions in an eco-friendly manner remains

a challenge.

TABLE XV

Relative reaction rates for epoxidation between different alkenes

and hex-1-ene on Ti-beta with H2O2 and TBHP

Oxidant Oct-1-ene Dec-1-ene 4-m-Pent-1-ene 1-m-Cyclohex-1-ene

H2O2a 0.70 0.60 1.45 1.22

TBHPb 0.52 0.37 0.49 1.09

Adapted from Corma et al. (191).a Reaction conditions (H2O2 oxidant): catalyst, 0.2 g; Alkene, 33 mmol; H2O2

(35 wt%), 0.8 g; solvent (CH3OH), 23.6 g; temperature, 323 K; time, 2 h.b Reaction conditions (TBHP oxidant): catalyst, 0.3 g; alkene, 25 mmol; TBHP,

6.25 mmol; solvent (CH3CN), 10 g; temperature, 323 K; time, 5 h.


V.C.8. Epoxidation in the Presence of Other Oxidizable Functional Groups

V.C.8.1. Alkenes and Alcohol Functions. Although TS-1 and other titanosi-

licates oxidize alcohols to the corresponding aldehydes and ketones, the rates are

suppressed in the presence of compounds containing CyC bonds. CH3OH, for

example, is not oxidized at all during epoxidations of alkene reactants. Higher

alcohols, however, are partially oxidized. The oxidation of unsaturated alcohols

in the presence of TS-1 is shown in Table XVII (193).

When the double bond has no substituents, as in allyl alcohol, but-3-ene-

1-ol, or 2-methylbut-3-ene-1-ol, only the epoxide is formed; but when the

double bond has substituents, the epoxidation rate is decreased and ketone and

aldehyde products are formed from the oxidation of the OH group. This effect

is more pronounced with a greater degree of substitution of the reactant.

Because the double bond and the OH group are part of the same molecule, this

difference must arise from the different abilities of the functional groups to

coordinate and react at the Ti center. The terminal double bond, sterically less

hindered, interacts strongly with titanium, preventing coordination of the com-

peting OH group. Because of steric hindrance, this interaction is weaker in

substituted alkenes, allowing the OH group to undergo oxidation (190).

V.C.8.2. Alkenes and Alkanes. When oct-1-ene was oxidized by H2O2/TS-1

in the presence of n-hexane, under conditions that would lead to the oxidation of

each if it were used separately, epoxidation occurred preferentially (103). This

result is probably an evidence of the greater nucleophilicity and, hence, coordi-

nating ability of the alkene.

V.C.9. Hydroxyl-Assisted Epoxidation

Hydroxyl-assisted epoxidation using TS-1/H2O2 is chemo- and stereoselective

(165). Thus, when cyclopent-2-en-1-ol or cyclohex-2-en-1-ol was treated with

TABLE XVI

Epoxidation of dialkenes catalyzed by TS-1

Alkene Solvent T (K) H2O2

conversion

Yield based

on H2O2 (%)

Epoxide selectivity

(%)

Mono Di

Butadiene tert-Butyl alcohol 293 98 85 85 15

Diallyl carbonate Methanol 338 95 50 93 5

Diallyl ether Methanol 338 96 60 90 4

Adapted from Romano et al. (192); diene/H2O2 ¼ 2.5.


aqueous H2O2/TS-1, the corresponding epoxides were obtained in 75–80%

yields. Cyclohexenol gave the cis I as the major product (90%) (where epoxide

and OH are cis to each other), and the trans II as a minor product (10%) (where

the epoxide and OH are trans to each other) (Scheme 8). Cyclopentenol also

TABLE XVII

Oxidation of unsaturated alcohols in the presence of TS-1: effect of alkene

structure on selectivity

Reactant Product yield (mol/mol of Ti)

Ketone/aldehyde Epoxide

0 19

0 16

0 30

31 95

37 4

7 27

43 65

44 141

98 94

18 10

75 17

Adapted from Tatsumi et al. (193). Reaction conditions: TS-1 (Si/Ti ¼ 52),

0.01 g; reactant, 2.5 mL; H2O2 (30% aq. solution), 2.5 mL; temperature, 323 K;

time, 3 h.


behaved similarly. In addition to the epoxide, other products resulting from

oxidation of the OH group and cleavage of the epoxide were also detected.

As a further example of a hydroxyl-assisted epoxidation, geraniol and nerol

bearing two isolated CyC double bonds were regioselectively epoxidized with

TS-1 at the 2-position (near the OH group), as reported by Kumar et al. (195).

On the basis of these results, Kumar et al. (195) proposed that the transition state

of the epoxidation of allylic alcohols involves coordination of the alcoholic

functional group to the Ti active site and that the double bond interacts with one

of the peroxidic oxygen atoms, not with the titanium site (Scheme 9).

The epoxidation of a bulky reactant such as alpha-terpineol was accomplished

with Ti-beta as the catalyst. The initially formed epoxide was rearranged to cineol

alcohol, as shown in Scheme 10 (18,196,197).

Even as large a molecule as cholesterol was epoxidized in the presence of

Ti-MCM-41 catalyst (198). An epoxide selectivity of 53% at 48% conversion

was achieved. The oxidation of the OH group and allylic oxidations were

important side reactions.

V.C.10. Diastereoselectivity in Epoxidations

Epoxidation of allyl alcohols can generate two isomers, the threo- and erythro-

epoxides (Schemes 11 and 12). Control of the relative amounts of the two isomers

Scheme 8.

Scheme 9.


is crucial in the synthesis of many compounds of interest in the fine chemicals

industry. The results of Adam et al. (199,200) for reactions catalyzed by TS-1 and

Ti-beta are summarized in Table XVIII.

As expected, TS-1 was not active for the bulky reactants (Table XVIII, entries

9–14). The diastereoisomeric ratios evidencing catalysis by TS-1 and Ti-beta

are broadly similar to those of the homogeneous system Ti(OPr)4-TBHP and

chloroperbenzoic acid. A transition state for the active species analogous to

the structure of peracid epoxidations was, therefore, suggested (199), involving

interaction of the alcoholic functional group with the peroxo oxygen atom by

hydrogen bonding.

V.C.11. Side Reactions During Epoxidation

On titanosilicate molecular sieves, especially non-TS-1 materials, the epoxides

formed react further to form glycols, glycol ethers, and even products arising

from the further rearrangemnent of the epoxide. Thus, in the epoxidation of

styrene by H2O2/TS-1, the epoxide rearranged efficiently into phenylacetalde-

hyde (165). No or very little acetophenone was produced, phenylacetaldehyde

being the sole or major product. The high regioselectivity for phenylacetaldehyde

was attributed to the stabilization of the benzyl cation (165). Although high

(epoxide þ phenylacetaldehyde) selectivities (85–90%) were obtained for reac-

tion in the presence of acetone, alcoholysis occurred to a great extent (45%) in the

presence of methanol solvent, producing mono glycol ethers.

Scheme 10.

Scheme 11.


Allyl alcohol epoxidation with TS-1/H2O2 and the subsequent epoxide ring

opening reaction by water or the organic solvent was investigated thoroughly by

Hutchings et al. (201–203). Although very high selectivities to epoxides were

observed at low conversions and temperatures, ether diols, resulting from the

nucleophilic epoxide ring opening by the alcohol, were the major products at

temperatures above 338 K. Scheme 13 was proposed for epoxide ring opening by

polar solvent molecules. It was shown (201–203) that the Ti-peroxo complex is

more acidic than TS-1 alone (without H2O2), and it is mainly this complex that

catalyzes the solvolysis reaction. When Brønsted acid sites were deliberately

introduced into TS-1 (by partial introduction of Al3þ in the framework), the

epoxide was not found among the reaction products because it was rapidly

converted to the ether diol solvolysis products.

V.C.12. Influence of pH

Since acidity (Lewis or Brønsted) impacts adversely on the yield of epoxides,

Clerici and Ingallina (204) added basic compounds in low concentrations to

TS-1 catalysts during epoxidation of alkenes to inhibit the oxirane ring opening

and enhanced the epoxide yields. A comprehensive investigation of the influ-

ence of pH on product selectivity in epoxidation of allylalcohol, allylchloride,

and styrene catalyzed by various titanosilicates was reported recently by Shetti

et al. (205).

Although conversion of allyl alcohol catalyzed by TS-1 decreased from

95.3% (at pH ¼ 3.5) to 22.2% (at pH ¼ 8.5), epoxide selectivity increased from

86.8 to 100% (Table XIX). The H2O2 efficiency decreased markedly at high pH.

Most of the H2O2 probably decomposed to H2O and O2 at high pH. Ti-MCM-41

exhibited lower activity than TS-1. Changes in pH did not affect conversions

significantly when reaction was catalyzed by Ti-MCM-41. To investigate the

Scheme 12.


TABLE XVIII

Diastereoselective epoxidation of allylalcohols

Reactant Catalyst/oxidant/solvent

(TS-1/UHP/CH3COCH3)

Ti-beta/H2O2 (85%)/

CH3CN

Diastereomeric ratio (threo: erythro)

60:35 62:38

55:45 56:44

65:35 64:36

87:13 91:9

81:19 89:11

95:5 95:5

90:10

80:20 93:7

No epoxide 58:42

No reaction 95:5

No reaction 70:30

(Continued)


influence of cations present in solution, the epoxidation of allyl alcohol was

carried out with TS-1 catalyst at pH ¼ 8 in the presence of various alkali metal

and alkaline earth compounds (205). Catalytic activity increased in the following

order: Liþ , NH4þ , Naþ , Kþ , Csþ and Mg2þ , Ca2þ , Ba2þ. Epoxide

selectivity followed the reverse order; Csþ exhibited 100% allyl alcohol con-

version but only 76.7% epoxide selectivity (Table XX, Run number 5).

The influence of pH on epoxidation of styrene with aqueous H2O2 catalyzed by

TS-1 was also investigated. Conversion of styrene decreased, and styrene oxide

selectivity increased marginally at high pH values (Table XXI).

Scheme 13.

TABLE XVIII

Continued

Reactant Catalyst/oxidant/solvent

(TS-1/UHP/CH3COCH3)

Ti-beta/H2O2 (85%)/

CH3CN

No reaction 88:12

No reaction 15:85

No reaction 70:30

Adapted from Adam et al. (199, 200).


TABLE XIX

Epoxidation of allyl alcohol and allyl chloride—influence of pH

Run no. Catalyst Reactant pH TOF Olefin conversion

(mol%)

H2O2

efficiency

Epoxide selectivity

(mol%)

Initial/ before

H2O2 addition

After H2O2

addition

At the end

of the reaction

1 TS-1 AA 4.5 3.5 4.2 18.9 95.3 100 86.8

2 TS-1 AA 5.5 3.5 4.2 18.7 94.4 100 87.4

3 TS-1 AA 7.0 5.5 5.8 17.7 89.2 95 92.8

4 TS-1 AA 8.0 5.7 5.9 16.0 80.7 87 100

5 TS-1 AA 9.0 5.9 6.2 7.1 35.6 46 100

6 TS-1 AA 10.0 8.5 8.0 4.4 22.2 21 100

7 TS-1 in runs 3

and 4 reused

AA 7.8 5.5 5.7 14.8 74.4 97 96.8

8 TS-1-Na(8) AA 8.0 5.7 5.9 16.2 81.6 89 100

9 TS-1-Na(10) AA 10.0 7.8 8.0 5.9 29.7 20 100

10 Ti-MCM-41 AA 6.8 – – 2.1 10.4 100

11 Ti-MCM-41 AA 8.0 – – 2.3 11.4 100

12 TS-1 AC 6.8 3.2 3.0 19.3 97.2 97 73.8

13 TS-1 AC 7.0 3.8 3.8 19.8 100 100 79.7

14 TS-1 AC 8.0 4.3 4.2 19.8 100 100 81.6

Adapted from Shetti et al. (205). Reaction conditions: catalyst (TS-1: Si/Ti ¼ 33; Ti-MCM-41: Si/Ti ¼ 52), 0.1 g; reactant, 0.5 g; CH3OH, 10 g; H2O2 (50%

aqueous), 0.9 mL; H2O2/allylalcohol, 2.0; temperature, 333 K; time, 8 h; AA, allyl alcohol; AC, allyl chloride. TOF, moles reactant converted per mol of Ti

per hour. Catalysts used in run nos. 8 and 9 were prepared by impregnating TS-1 with Naþ ions with initial pH being 8 and 10, respectively.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

97

9

V.C.13. Epoxidation with Alkyl Hydroperoxides

Although aqueous H2O2 is an efficient oxidant with TS-1 and Ti-beta, catalyst

stability and conversion are not as good when Ti-MCM-41 or other hydrophilic,

mesoporous Ti-silicate molecular sieves are used as catalysts. The behavior of the

mesoporous materials resembles the Shell catalyst, amorphous Ti–SiO2. TBHP

is a better oxidizing agent than H2O2 in this case. Although mesoporous materials

do not match the epoxide selectivity and H2O2 efficiency of TS-1 for small

TABLE XXI

Influence of pH on styrene epoxidation over TS-1

Run no. pH Conv.

(mol%)

Styrene

oxide

Methylated

diol

Diol Benzaldehyde Phenyl

acetaldehyde

Others

1 6.8 39.9 35.9 46.0 1.1 13.8 0.3 2.9

2 7.0 42.3 35.0 45.5 0.8 15.1 0.5 3.1

3 8.0 35.2 40.9 40.7 1.1 15.5 0.3 1.5

4 9.0 27.3 45.0 37.4 0.8 16.8 0.0 0.0

5 11.0 4.4 66.6 5.9 0.0 18.3 9.2 0.0

Adapted from Shetti et al. (205). Reaction conditions: TS-1: 0.1 g; styrene, 0.898 g; CH3OH, 10 g;

H2O2 (50%), 0.9 mL; H2O2/styrene, 2; temperature, 333 K; time, 8 h.

TABLE XX

Effect of alkali and alkaline ions on the epoxidation of allylalcohol

Run no. Alkali/Alkaline

earth ions

TOF Conversion

(mol%)

Epoxide selectivity

(mol%)

1 Liþ 2.4 11.9 100

2 NH4þ 9.7 48.6 100

3 Naþ 17.0 85.8 91.7

4 Kþ 18.7 94.4 79.1

5 Csþ 19.8 100 76.7

6 Mg2þ 12.5 63.0 100

7 Ca2þ 18.8 94.7 88.5

8 Ba2þ 18.6 94.1 75.0

Adapted from Shetti et al. (205). Reaction conditions: catalyst (TS-1; Si/Ti ¼ 33),

100 mg; allyl alcohol, 0.5 g; CH3OH, 10 g; H2O2 (50%), 0.9 mL; H2O2/

allylalcohol ¼ 2.0; temperature, 333 K; run time, 8 h; pH, 8.0. TOF, moles of allyl

alcohol converted per mol of Ti per hour.


molecules, they are superior to it in the epoxidation of bulky alkenes (11,13,193,

196,199). Ti-beta, in contrast to TS-1, is considerably more active in the

epoxidation of allylic alcohols highly substituted at the CyC bond (compare

reactants 3 and 4 in Table XXII). The accessibility of the CyC bond to the

titanium oxo centers is apparently not seriously hindered by the alkyl substituents

in reaction catalyzed by Ti-beta.

V.C.14. Epoxidation of Alkenes Containing Carbonyl Groups

In homogeneous systems, electron-withdrawing groups such as CyO, when con-

jugated with the alkene double bond, retard the epoxidation as the delocalization

TABLE XXII

Epoxidation of allylalcohols with titanosilicates and H2O2

Reactant Catalyst T (K) Conversion

(%)

Epoxide

selectivitya (%)

Productivityb

(mmol/g/h)

Ti-beta 343 42 89 (11) 28

TS-1 323 nd 76 (24) 19

Ti-beta 343 nd .90 3.9c

TS-1 333 nd 96 21

Ti-beta 343 nd .90 2.5c

TS-1 323 nd 100 3

Ti-beta 343 nd 85 (11) 8.5c

TS-1 323 nd 90 (10) 3.3

Ti-beta 343 nd 96 25

TS-1 323 nd 82 (18) 9

Ti-beta 343 nd 74 (26) 12

TS-1 303 54 100 73

Adapted from Dusi et al. (181). Note: nd, no data available.a Numbers in parentheses indicate the selectivities to the corresponding unsaturated carbonyl

compounds.b Amount of oxygenated products, related to unit amount of catalyst and unit time.c Based on epoxide and triol formed.


of the p-electrons reduces the electron density at the double bond. Ratnasamy

and Kumar (206) found that the main products formed in the oxidation of acrolein

and methacrolein were the corresponding acids from the reaction at the carbonyl

end (.90%); only little epoxide was obtained.

V.C.15. Epoxidation Using Urea–H2O2 Adduct

The epoxide selectivity in the TS-1/aqueous H2O2 system is reduced because

of the formation of isomerized and/or cleaved secondary products because the

oxirane ring is quite prone to hydrolysis in the presence of water. To circumvent

this problem, an anhydrous source of H2O2, namely, urea–H2O2 adduct (UHP),

which slowly releases anhydrous H2O2 into the solution, was successfully

employed by Laha and Kumar (207) to enhance epoxide selectivities, even

in the difficult case of styrene to styrene oxide (Table XXIII). The formation

of side products, diols (by hydrolysis), phenylacetaldehyde (by rearrangement

of the epoxide), and benzaldehyde (by C–C bond cleavage), have all been

significantly reduced when UHP was used as the oxidant.

TABLE XXIII

Effect of different oxidants on epoxidation of styrene and allylbenzene catalyzed by TS-1 and TS-2

Reactant Catalyst Oxidanta Conversion

(mol%)

TONb Product distribution (mol%)a

EP PAD BD Diols

Styrene TS-1 HP 56 13.4 5 44 29 22

U þ HP 65 15.6 81 8 7 4

UHP 71 17.0 87 5 7 1

TS-2 HP 57 14.1 7 42 28 23

U þ HP 62 15.8 80 8 8 4

UHP 67 17.3 85 6 7 2

Allylbenzene TS-1 HP 60 12.7 58 – – 42

U þ HP 68 14.4 95 – – 5

UHP 70 14.8 98 – – 2

Adapted from Laha and Kumar (207). Reaction conditions: reactant:oxidant (mol) ¼ 4; solvent,

acetone; reactant:acetone (wt/wt) ¼ 1; reaction time (h) ¼ 12 h; catalyst wt ¼ 20 wt% of the reactant;

T ¼ 313 K.a EP, epoxy allylbenzene or styrene oxide; PAD, phenylacetaldehyde; BD, benzaldehyde; Diols, 3-

phenyl-1,2-propanediol or styrene diol, including some high-boiling products; HP, H2O2 (45 wt%

aqueous); U þ HP, urea and H2O2 mixture (1:1 mol ratio); UHP, urea–H2O2 adduct.b TON, moles of H2O2 converted for producing epoxide þ secondary products per mole of Ti.


V.C.16. Epoxidation Using Dioxygen

One of the major developments in the preceding decade in the area of epoxidation

catalyzed by titanosilicates is the attempt to generate H2O2 in situ by a mixture of

H2 þ O2 catalyzed by Pd/Pt-TS-1 (69–71,208–210) or Au-TS-1 (74).

The strategy was to generate H2O2 from H2 þ O2 catalyzed by the noble

metals and react it with the alkenes (especially propene) in the presence of TS-1

catalyst to produce the epoxide. Intimate contact between the metal and TS-1

and consequently a high dispersion of the metal on the hydrophobic TS-1 surface

is needed. The latter is difficult to achieve and especially to maintain. Catalyst

deactivation was a major problem (71). In addition to propene epoxide, the

by-products included methyl formate (from the methanol solvent), acetone,

acrolein, acrylic acid, and methylated glycols (71). An interesting observation in

most of the investigations (69–71,74,208) was that although the yields were

low, the propene selectivity to the epoxide was .99%; the yields were low as a

consequence of the low hydrogen and oxygen efficiencies in the production

of H2O2 (74). The in situ generation of H2O2 at the precious metal site is

probably rate-determining in this reaction (208). Catalyst deactivation was also

a problem. Meiers et al. (69) found that the formation of propene oxide in the

presence of Pd–Pt-TS-1 was favored when a high fraction of palladium is

present as Pd2þ species and small palladium clusters, whereas fully reduced

palladium and large clusters favored propene reduction to propane. The fraction

of Pd2þ was increased by autoreduction of the complex incorporating tetramine

ligands [(Pd(NH2)4]2þ was the precursor for Pd-TS-1) in the absence of

hydrogen in the reduction medium; calcination of the dried sample in N2 at

523 K was adequate to reduce the Pd ions. Reaction temperatures .423 K or

calcinations in air led to palladium cluster agglomeration on the external TS-1

surface and thus to decreasing epoxide yields and selectivities. Addition of

minor amounts of platinum also drastically increased the fraction of Pd2þ

species in comparison to the Pd0 species (69). Although no epoxidation of

propene occurred in the catalysis by TS-1 with H2 þ O2 as the oxidant, a 5.3%

yield of propene oxide was obtained with a 1% wt Pd–0.1%wt Pt-TS-1 catalyst

under the same conditions. Of course, the yield was much higher (39%) in the

TS-1/H2O2 system (70). Although higher yields have been reported (up to 12%

propene oxide (69)), they are still much lower than those obtained with H2O2.

V.D. Hydroxylations

V.D.1. General Features

Titanosilicate molecular sieves, especially TS-1, are active in the hydroxylation

of both alkanes and aromatic compounds (33,165) when H2O2 is used as


the oxidant. The manufacture of hydroquinone and catechol, in nearly equal

quantities, from phenol and H2O2 with TS-1 catalyst is in commercial practice.

The substitution or insertion of an oxygen atom into C–H bonds is not easy, and

the applied reagents have to be strongly electrophilic oxidants or radical species.

C–H hydroxylations can be classified broadly into two reaction types (179): the

first type is the insertion of a singlet oxygen atom (1O) into C–H bonds from

electrophilic oxidants as in Eq. (20).

ð20Þ

The second type is the hydroxylation by a triplet oxygen atom (3O) and

involves radical intermediates via H abstraction:

ð21Þ

The hydroxylation of C–H bonds by radicals, in contrast to the case of

electrophilic oxidants, leads to alcohols without retention of stereochemical

configuration. H2O2, activated by strong acids (superacids (211), HF–BF3 (212),

AlCl3 (213), and CF3COOH (214)) have been used for the hydroxylation of

aromatic compounds. These acid-catalyzed hydroxylations cannot be applied for

aliphatic reactants because the hydroxylated products are more reactive than the

starting compounds and, hence, they are oxidized further.

Radical hydroxylation of hydrocarbons by autooxidation yields alcohols (major

products), ketones, and acids (minor products). Cyclohexanol, for example, is

formed in 90% yield from cyclohexane and peroxyacetic acid (215). The high

-ol/-one ratio at low conversions can sometimes be used as a partial diagnostic

tool to distinguish between the radical and electrophilic pathways. The

predominant reaction of electrophilic radicals, such as HOz, ROOz, and CH3z is

H-atom abstraction from reactants (S–H) or peracids, as exemplified by the

following:

Xz þ S–H ! XH þ Sz ð22Þ

Sz þ HOz ! SOH ð23Þ

Xz þ H–OOCOR ! XH þ RCOz

3 ð24Þ

Thus, the generation of these radicals leads to the hydroxylation of S–H.

The reactive hydroxyl radicals can be produced by the radiolysis of water or


the reduction of H2O2:

H2O ! H2Oþ ! H3Oþ þ HOz ð25Þ

H2O2 þ Fe2þ ! HO2 þ HOz þ Fe3þ ð26Þ

H2O2 þ Ti3þ ! HO2 þ HOz þ Ti4þ ð27Þ

V.D.2. Hydroxylation of Aliphatic Compounds

Linear alkanes have been hydroxylated in the 2-, 3-, and 4-positions to give

secondary alcohols and ketones in the presence of TS-1 catalyst (216,217) with

good selectivities based on alkanes and H2O2 (Table XXIV).

The alcohols are intermediates in the formation of ketones. Isomerization of

the products is not observed. Hydroxylation at the 2-position is favored over

that at the 3-position, and the latter is preferred over hydroxylation at the 4-

position. Solubility and concentration in the reaction medium, intrazeolite

diffusion of the reactants, steric hindrance at the reactive carbon center, and

C–H bond strength influence the reactivity and H2O2 selectivity (Table XXIV).

The advantage of the large-pore Ti-beta over TS-1 in the oxidation of bulky

alkane molecules is shown by the results in Table XXV.

Table XXVI shows the results of a competitive experiment in which hydroxy-

lation of an equimolar mixture of n-hexane and another alkane (alkane II)

TABLE XXIV

Oxidation of n-alkanes in 95% methanol

Hydrocarbon Selectivity

based on H2O2 (%)a

2/3 ratiob Product distribution (mol%)

2-ol 3-ol 4-ol 2-one 3-one 4-one

Propane 35 66.2 33.8

n-Butane 69 55.0 45.0

n-Pentane 82 4.5 34.3 16.1 47.4 2.1

n-Hexane 86 2.6 32.1 25.9 39.8 2.0

n-Heptane 75 1.9 33.7 29.2 6.2 28.1 2.8 Trace

n-Octane 63 2.6 30.1 20.5 12.5 32.8 3.0 1.0

n-Decane 56 1.1 11.5 20.5 36.2 16.5 4.5 10.8

From Notari (33).a Represents the moles of oxygenated products obtained per 100 moles H2O2 reacted.b Ratio between 2- and 3-compounds.


TABLE XXV

Comparative activity of Ti-beta and Euro-TS-1 for selective oxidation of different alkanes

Alkane Catalysta Turnover

(mol/mol Ti)

H2O2 Product

selectivity (%)

Conv. (%) Sel. (%)b -ol -one

n-Hexane TS-1 48.5 77 100 91.5 8.5

Ti-beta 0.5 11 32 55.0 45.0

3-Methylpentane TS-1 0.7 6 19 88.9 11.1

Ti-beta 0.8 17 29 84.8 15.2

Cyclohexane TS-1 –b – – – –

Ti-beta 2.3 22 51 98.9 1.1

Methylcyclohexane TS-1 –b – – – –

Ti-beta 5.2 29 88 92.8 0.9

Adapted from Corma et al. (11). Reaction condition: catalyst, 0.2 g; alkane, 33 mmol; solvent

(CH3OH), 23.57 g; H2O2/alkane ¼ 0.082 mol/mol; temperature, 333 K; reaction time ¼ 4 h.

Catalyst: Euro-TS-1 (1.7 wt% TiO2); Ti-beta (5.2 wt% TiO2, TiO2/Al2O3 ¼ 210; Ti/(Ti þ

Si) ¼ 0.040).a H2O2 selectivity (%) ¼ (mol alkane oxidized/mol H2O2 converted) £ 100.b Activity below detection limit.

TABLE XXVI

Competitive oxidation of equimolar mixtures of n-hexane and another alkane

(alkane II) over TS-2 using H2O2 as oxidant

Alkane II Critical

diameter (nm)

Conversion (mol%) n-Hexane/alkane II

conversion

n-Hexane Alkane II

3-MP 0.55 7.8 2.8 2.8

2,2-DMB 0.61 8.2 1.7 4.8

Cyclohexane 0.60 12.3 1.8 6.8

n-Hexane 0.43 18.9 – –

From Kumar et al. (165). Reaction conditions: catalyst (TS-2; Si/Ti ¼ 77);

reactant, 1 g; reactant/H2O2 (mol), 3; solvent (CH3CN), 10 g; temperature,

353 K; time, 8 h; 3-MP, 3-methyl pentane; 2,2-DMB, 2,2-dimethyl butane.


with varying critical diameter, selected from 3-methylpentane (3-MP) or

2,2-dimethylbutane (2,2-DMB), or cyclohexane was carried out in the presence

of TS-2 with dilute H2O2. As the size of the competing alkane II increases, its

relative conversion (vis-a-vis n-hexane) decreases, the reactivity order being

n-hexane . 3-MP . 2,2-DMB . cyclohexane. From the point of view of

chemical reactivity in unconstrained or homogeneous catalytic systems, the

reverse trend is expected. Further, although the critical diameters of 2,2-DMB

and cyclohexane are comparable (0.60 and 0.61 nm, respectively), 2,2-DMB

competes better with n-hexane than with cyclohexane. Apparently, not only the

size but also the shape and/or conformation of the reactants may play a role

in competitive hydroxylations; the results highlight the importance of steric

factors in the adsorption process. Similar results were obtained with TS-1

catalyst.

In contrast to their vanadosilicate analogues, the titanosilicate molecular sieves

do not hydroxylate the terminal primary carbon in n-alkanes. Ramaswamy et al.

(218,219) found that when n-hexane was hydroxylated under identical conditions

in the presence of TS-2 or VS-2 (VS-2 is a vanadium analogue of TS-2), the

distribution of products was as follows:

TS-2 : hexan-2-ol ð52%Þ . hexan-3-ol ð48%Þ ðno activation at 1-positionÞ

VS-2 : hexan-2-ol ð45%Þ . hexan-3-ol ð42%Þ . hexan-1-ol ð13%Þ

Furthermore, the -ol/-one ratio was also higher when the catalyst was TS-2

(0.77) than when it was VS-2 (0.36). The pathways for reaction catalyzed by

the titano- and vanadosilicates are probably different. The absence of hydroxy-

lation of the primary C–H bond and the higher -ol/-one ratio when the catalyst

is the titanosilicate is significant. Because the homolytic bond dissociation

energies decrease in the order primary C–H . secondary C–H . tertiary C–H

bonds, radical pathways involving C–H bond homolysis almost always show a

marked preference for the functionalization of tertiary and secondary C–H

bonds (220). The preference for secondary C–H bonds and the high -ol/-one

ratios when the catalyst is TS-2 suggest that radical pathways are involved in

the hydroxylation of alkanes with TS-2. In fact, Khouw et al. (221) had earlier

proposed a possible mechanism for alkane hydroxylation catalyzed by TS-1

which proceeds via homolytic Hz abstraction from R–H by a Ti(O2H) group

which may have some superoxo-like character (Scheme 14). This Hz abstraction

generates an alkyl radical, Rz, and is accompanied by reduction of Ti4þ to Ti3þ.

A subsequent homolytic O–O bond cleavage occurs to form the C–O bond.

In support of the above mechanism, the following results may be mentioned:

(i) superoxo radicals have indeed been observed in oxidation reactions catalyzed

by titanosilicates (51,52,54,131,205,222); (ii) Ti4þ ions are reduced to Ti3þ

in the presence of reducing agents such as CO (122), H2, and hydrocarbons


(51,52,130,131), or at high pH (205); (iii) the preference of secondary over

primary C–H bonds in the hydroxylation of alkanes; and (iv) the high -ol/-one

ratios in the oxidation of cyclohexane. The titanyl group (TiyO) proposed by

Khouw et al. (221), has, so far, not been observed experimentally during

oxidation catalyzed by titanosilicates.

The hydroxylation of octane and cyclohexane catalyzed by Ti-MMM-1,

a mixed- phase material (TS-1 and Ti-MCM-41) containing both micro- and

mesopores, with aqueous H2O2 was reported by Poladi et al. (223). Ti-MMM-1

was found to be more active and selective in these hydroxylations than either

Ti-MCM-41 or TS-1; the yield of alcohol was higher (Table XXVII).

The detailed crystallographic and textural structure of this mixed phase

material is not clear. It seems likely that the higher activity (conversion) is a

consequence of the presence of mesopores (of the MCM-41 phase) leading into

the micropores (of the MFI phase); these mesopores would enhance the diffusion

of the reactants deep into the crystallites while simultaneously preserving the

advantages of the microporous MFI phase (such as higher intrinsic activity and

selectivity). In the absence of the mesopores of the MCM-41 phase, a significant

portion of the interior of the crystallite would have been inaccessible to the

reactants. Similarly, the high selectivity for alcohols, the primary oxidation

product, is a consequence of their faster diffusion out of the solid crystallite

through the mesopores. In the absence of the mesopores, the alcohol molecules

diffusing more slowly through the pores of the MFI phase would undergo further

oxidation to the ketone before emerging from a catalyst particle. The advantages

of a mixed phase catalyst are thus evident. One major advantage of Ti-MMM-1

is that it allows the application of aqueous H2O2 as the oxidant. Apparently most

Scheme 14.


of the catalysis occurs in the TS-1 phase, which, being hydrophobic, is quite

stable in aqueous media. The role of the MCM-41 phase is mainly to facilitate

the transport of reactants and products to and from the active sites of TS-1. Other

mesoporous titanosilicates suffer from their instability in an aqueous medium,

and therefore, have to be used with TBHP or other alkyl hydroperoxides, with

the attendant environmental problems. Hence, if the hydrothermal stability,

absence of titanium leaching, and catalytic superiority of this mixed phase

material is validated thoroughly, it will be a significant addition to the family of

titanosilicate-containing oxidation catalysts.

V.D.3. Hydroxylation of Aromatic Compounds

The selective hydroxylation, in the presence of aqueous H2O2, of aromatic

hydrocarbons such as benzene, toluene, and xylene to phenol, cresols, and

xylenols, respectively, occurs easily on TS-1 (33,165,224). Again, a significant

contrast between TS-2 and VS-2 in the oxidation of toluene is that when the

catalyst is the former, only aromatic ring hydroxylation takes place, although

when the catalyst is VS-2, the side chain C–H bonds are also hydroxylated (165,

218,219,225,226) (Table XXVIII).

When the alkyl substituent contains secondary C–H bonds, both ring and side

chain oxidation at the secondary C–H bond occur. Thus, ethylbenzene gives

TABLE XXVII

Comparative activity of mixed phase Ti-MMM-1 with TS-1 and Ti-MCM-41 for the oxidation of

cyclohexane and n-octane

Conversiona

(mol%)

Ketone(s)

(mol%)

Alcohol(s)

(mol%)

Othersb Ketone:alcohol

Cyclohexane

Ti-MMM-1 9.2 35.1 54.7 13.2 0.64

TS-1 4.2 26.4 27.6 46.0 0.96

Ti-MCM-41 1.9 9.8 17.0 73.2 0.58

n-Octane

Ti-MMM-1 19.8 14.5 80.8 4.7 0.18

TS-1 13.3 10.3 80.3 9.4 0.13

Ti-MCM-41 2.9 21.5 52.7 25.8 0.41

Adapted from Poladi and Landry (223).a Conversion ¼ (moles of alkane converted/total moles of alkane) £ 100.b Includes diols and diones.


ethyl phenols (40%), acetophenone (56%), and 2-phenyl ethanol (4%). Mono-

substituted benzenes with electron-donating groups (such as phenol, toluene,

etc.) undergo rapid hydroxylation (mainly in the ortho and para positions),

although those containing electron-withdrawing groups (such as Cl, NO2, etc.) do

not react so facilely (165). Similarly, bulky substituents, such as tert-butyl, retard

the reaction because of the steric restriction imposed by the pore size of the TS-1.

An increased selectivity for phenol in the oxidation of benzene by H2O2 with

TS-1 catalyst in sulfolane solvent was attributed to the formation of a bulky

sulfolane–phenol adduct which cannot enter the pores of TS-1. Further oxidation

of phenol to give quinones, tar, etc. is thus avoided. Removal of Ti ions from the

surface regions of TS-1 crystals by treatment with NH4HF2 and H2O2 was also

found to improve the activity and selectivity (227). The beneficial effects of

removal of surface Al ions on the catalytic performance of zeolite catalysts for

acid-catalyzed reactions have been known for a long time.

V.E. Oxidation of Nitrogen-Containing Compounds

As expected from the Lewis acidity of Ti4þ, the titanosilicates strongly adsorb

and oxidize basic nitrogen-containing compounds with a lone pair of electrons

localized on the N atom. By contrast, nitrogen oxides (NOx) and nitro compounds

TABLE XXVIII

Hydroxylation of aromatics over TS-2 and VS-2 molecular sieves

Benzene Toluene

TS-2 VS-2 TS-2 VS-2

Conversion (mol%) 51.3 21.6 39.6 35.1

Products (mol%)

Phenol 88.0 90.0 – –

p-Benzoquinone 9.0 7.0 – –

o-Cresol – – 36.0 20.0

p-Cresol – – 59.0 17.0

Benzyl alcohol – – – 8.0

Benzaldehyde – – – 52.0

Others 3.0 3.0 5.0 3.0

Adapted from Kumar et al. (165). Reaction conditions: catalyst (TS-2:

Si/Ti ¼ 77; VS-2: Si/V ¼ 79), 0.1 g; reactant, 1 g; reactant/H2O2 (mol) ¼ 3.0;

solvent (CH3CN), 10 g; temperature, 333 K; time ¼ 8 h.


(both aliphatic and aromatic) are not reactive in the TS-1/H2O2 system; nitro-

benzene, for example, is not oxidized to nitrophenols. The following oxidations

occur: (i) NH3 to NH2OH (14); (ii) primary amines to oximes (Table XXIX,

Scheme 15) (228); (iii) secondary amines to nitrones (229); (iv) tert-amines to

the corresponding nitrogen oxides (33); and (v) anilines to azoxybenzenes (230):

NH3 þ H2O2 ! NH2OH þ H2O ð28Þ

Scheme 15.

TABLE XXIX

Oxidation of primary amines catalyzed by TS-1

Amine Solvent Conversion Oxime selectivity H2O2 efficiency

CH3NH2 CH3OH 40 88 90

CH3NH2 CH3OHa 3 0 0

n-C3H7NH2 CH3OH 32 73 86

i-C3H7NH2 CH3OH 38 77 88

i-C3H7NH2 t-BuOHb 29 74 85

i-C3H7NH2 t-BuOHc 31 84 90

C6H11NH2 CH3OH 3 33 8

C6H11NH2 t-BuOH 3 32 8

C6H5CH2NH2 CH3OH 20 82 55

Adapted from Reddy and Jacobs (228).a Reaction without catalyst.b t-BuOH, tert-butyl alcohol.c Reaction over TS-2.


R1R2CH–NH2 ! R1R2CH–NHOH ! R1R2CH–NO ! R1R2CyNOH ð29Þ

R1R2CH–NHR3 ! R1R2CyNðOÞR3 ð30Þ

R3N ! R3NO ð31Þ

C6H5NH2 ! C6H5NðOÞyN–C6H5 ð32Þ

The oxidation of NH3 to NH2OH forms the basis of a process for the

ammoximation of cyclohexanone to the oxime because the NH2OH formed in

solution readily reacts with the ketone (non-catalytically) to give the oxime (231).

Table XXX (165) illustrates the conversions and selectivites obtained for a

few typical ketones and aldehydes. The ammoximation of aldehydes is faster

than that of ketones. The oxime selectivity is also higher. The ammoximation

of cyclohexanone by this method offers a more eco-friendly alternative route

to the cyclohexanone oxime intermediate for the production of Nylon-6. The

current route coproduces large quantities of ammonium sulfate and involves the

use of hazardous chemicals such as oleum, halides, and oxides of nitrogen.

One of the major problems in all the ammoximation processes using aqueous

H2O2 þ TS-1 with NH3 is that, under the basic conditions (pH $ 10) prevailing

during the reaction, some of the lattice Si ions of the zeolite structure in TS-1 are

leached into solution, leading to catalyst destruction. This leaching is a common

characteristic of all silicates. Innovative catalyst formulations and process

modifications are needed to overcome this problem.

TABLE XXX

Ammoxidation of carbonyls over TS-1 (Si/Ti ¼ 29)

Reactant Time (h) Conversion (mol%) Selectivity (%)

Acetone 6.0 79.7 98.1

Hex-3-one 4.0 70.4 98.1

Methylisobutyl ketone 3.0 98.0 99.5

Cyclohexanone 4.0 98.2 96.4

p-Tolualdehyde 2.0 97.0 97.7

Benzaldehyde 2.5 97.0 99.4

Adapted from Kumar et al. (165). Reaction conditions: catalyst (TS-1; Si/Ti ¼ 29),

1.5 g; reactant, 10 g; reactant: H2O2:NH3 ¼ 1:1.2:2.0; solvent (tert-butanol), 40 g;

temperature, 343 K.


V.F. Oxidation of Sulfur-Containing Compounds

Similar to nitrogen compounds, electron-rich sulfur compounds, such as the

sulfides, with the lone pair of electrons on the sulfur atom, are oxidized to

sulfoxides and, further, to sulfones by the H2O2/titanosilicate sytem (218,232,

233). Table XXXI (232) illustrates typical conversions and product selectivities

for various sulfides for the reactions catalyzed by TS-1. Bulky sulfides such as

alkyl, phenyl sulfides are relatively unreactive because of their steric exclusion

from the pores of TS-1. Diphenyl sulfide could not be oxidized at all. As the

diffusivity and, hence, the conversion of the sulfide decreases, the further oxida-

tion of the primary product (sulfoxide) becomes more competitive, leading to

increased formation of the corresponding sulfone (Table XXXI):

R2S ! R2SO ! R2SðOÞ2 ð33Þ

Promising results in the oxidation of sulfides with mesoporous SBA-15 type

titanium silicates with hydrolytic stability in aqueous H2O2 were obtained by

Trukhan et al. (233). Their structural and textural parameters are given in Table

XXXII along with those of Ti-MMM, a mesoporous, mesophase material of the

MCM-41 type (29,229). The oxidation of methylphenyl sulfide (MPS) was

chosen as a test reaction. The SBA-15 samples had a highly ordered hexagonal

arrangement of mesopores (with a diameter about 11 nm). XPS, XANES, and

DRUV spectra indicated (234) that most of the Ti4þ ions in the Ti-SBA-15

(Fig. 36) and Ti-MMM samples are in an octahedral environment. Ti ions in

Ti-SBA-15 are present both as oligomerized titanium-oxygen species and as

segregated TiO2 (anatase) particles. The presence of anatase in Ti-SBA-15

containing 7.17 wt% Ti was also confirmed by Raman spectroscopy (Fig. 37) by

the strong peak at 145 cm21 characteristic of anatase. The absence of this Raman

peak in the spectrum of Ti-MMM (containing 1.9 wt% Ti) indicated that the Ti

ions in it are more dispersed than those in Ti-SBA-15. One difference between

TABLE XXXI

Oxidation of sulfides with H2O2 catalyzed by TS-2

Reactant Conversion (%) Selectivity (%)

Sulfoxide Sulfone

CH3–S–CH3 100 97 3

C2H5–S–C2H5 100 85 15

C6H5–S–CH3 98 78 22

C6H5–S–C2H5 70 75 15

Adapted from Reddy et al. (232).


TABLE XXXII

Structural and textural parameters of Ti-SBA-15 catalysts

Sample no. Ti content

(wt %)

Si/Ti

(atomic ratio)

pHa Structural parameters Textural parameters

Unit cell

parameter (nm)

FWHMb Specific surface area

(m2/g)

Specific mesopore

volume (cm3/g)

Mesopore

diameter (nm)

Wall

thicknessc

Mesopore External

1 2.05 38 3.18 12.25 0.054 573 30 1.34 10.6 1.7

2 4.00 19 2.61 12.46 0.061 619 37 1.40 10.9 1.6

3 7.17 10 2.78 12.84 0.033 514 44 1.10 10.9 2.0

Ti-MMMd 1.89 39 9.00 4.23 0.110 1260 29 0.90 3.45 0.8

Adapted from Trukhan et al. (234).a pH in the final mixture.b FWHM, full width at half maximum of the (100) reflection.c Calculated from the equation unit cell parameter ¼ mesopore diameter þ wall thickness.d Mesoporous mesophase material of the MCM-41 type.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

99

4

the Ti-MMM and Ti-SBA-15 samples is that, as a consequence of the greater wall

thickness in the latter (1.6–2.0 vs. 0.8 nm, Table XXXII), a greater fraction of the

Ti ions in Ti-SBA-15 are inaccessible to the reactants, as was confirmed by

infrared spectra of CO adsorbed on these samples (Fig. 38). Three types of bands

Fig. 37. Ambient-temperature Raman spectra of Ti-MMM, Ti-SBA-15 (samples 1–3), and TiO2

(anatase); p , plasma line [from Trukhan et al. (234)].

Fig. 36. UV–visible diffuse reflectance spectra and elemental analysis data for Ti-SBA-15:

(1) sample 1; (2) sample 2; (3) sample 3; and (4) sample 1 after treatment with 30% H2O2 [from

Trukhan et al. (234)].


were observed (Fig. 38), at 2137 cm21 (physically adsorbed CO), 2153 cm21

(complexes of CO with Si–OH groups), and 2179 cm21 (CO on Ti4þ). The

2179 cm21 band is clearly seen only for Ti-MMM, indicating that the surface

concentration of Ti4þ is considerably higher for Ti-MMM than for Ti-SBA-15

with the same Ti content.

The catalytic activities of Ti-MMM, Ti-SBA-15, and TS-1 are compared in

Table XXXIII (234). The activities of these titanoslicates for MPS oxidation are

in the order Ti-MMM . Ti-SBA-15 . TS-1. The catalytic activity was found to

correlate with the rate of H2O2 decomposition in the absence of the organic

reactant (Fig. 39). Ti-MMM on which H2O2 decomposed (to H2O and O2) faster

(curve b) was also more active in the oxidation of the sulfur-containing

compounds (Table XXXIII).

Among the Ti-SBA-15 samples, the activity decreased in the order, sample

1 . sample 2 . sample 3. The intensity of the broad band in the 200–350 nm

DRUV spectra of these samples also follows the same order (Fig. 36) and is a

rough measure of the dispersion of Ti in the sample. The higher catalytic

activity of Ti-MMM was ascribed to its greater surface Ti concentration.

Fig. 38. Infrared spectra of adsorbed CO for samples with similar titanium contents: (1) Ti-MMM

and (2) Ti-SBA-15 (sample 1) [from Trukhan et al. (234)].


Fig. 39. H2O2 conversion profiles: (a) for reaction catalyzed by Ti-SBA-15 (sample 1, 30 mg) and

(b) by Ti-MMM (33 mg). Reaction conditions: H2O2, 1.29 mmol; Ti, 0.013 mmol; MeCN, 3 ml;

T ¼ 353 K [from Trukhan et al. (234)].

TABLE XXXIII

Thioanisole (MPS) oxidation with 30% aqueous H2O2 over Ti-SBA-15 and other Ti,

Si-catalysts

Catalyst Time

(h)

MPS conversion

(%)

Product distribution (%)

Sulfoxide Sulfone

None 1 6 – –

1 1 48 72 28

1 (second cycle) 1 47 73 27

1 (third cycle) 1 50 72 28

2 1 26 73 27

2 3 53 65 35

3 1 29 71 29

Ti-MMMa 0.5 100 76 24

TS-1 (Ti, 2.54 wt%) 1 29 79 21

Adapted from Trukhan et al. (234). Reaction condition: MPS, 0.1 M;

[MPS]/[H2O2] ¼ 1/1.1; CH3CN, 3 mL; Ti, 6 £ 1023 mmol, 292 K. Structural and

textural properties of Ti-SBA-15 (1–3) and Ti-MMM catalysts are given in Table

XXXII.a [MPS]/[H2O2] ¼ 1/1.3.


Contrary to what was observed with Ti-MMM (29), no loss of catalytic activity

was observed after the recycling of the Ti-SBA-15 catalyst (Table XXXIII), a

result that confirms the hydrolytic stability of the Ti-SBA-15 materials. It was

verified that there was no leaching of Ti during the catalytic reaction (by hot

filtration of catalyst and testing the filtrate for catalytic activity) (29). Elemental

analysis after the catalytic runs confirmed that the total Ti content remained

the same for Ti-MMM (236), Ti-SBA-15, and TiO2–SiO2 mixed oxides (30). A

comparison of the DRUV spectra recorded before and after the treatment with

aqueous H2O2 indicated that, in contrast to the observations for Ti-MMM (236),

Ti-MCM-41 (237), TiO2–SiO2 mixed oxides (30), and TS-1 (228), there was

no change in the Ti-SBA-15 (221). The higher hydrolytic stability could not

be attributed to a lower hydrophilicity of Ti-SBA-15 because the specific H2O

adsorption capacity was similar for both Ti-MMM and Ti-SBA-15 (Fig. 40).

We emphasize that the above results have been observed only in the oxida-

tion of sulfides and phenols, reactions known to follow radical mechanisms.

A thorough investigation of the catalytic potential of the materials in other

oxidation reactions (epoxidation, hydroxylations, etc.) is warranted.

One of the major challenges in the petroleum industry today is the removal

of sulfur compounds, especially refractive ones such as 4,6-dimethyldibenzo-

thiophene (DMDBT), from petroleum fractions such as diesel to concentrations

,5–10 ppm from the current values of 50–500 ppm. The current technology

is hydrodesulfurization catalyzed by cobalt–nickel–molybdenum sulfides at

high pressures. Reducing sulfur concentratios in diesel fuels below 5–10 ppm

Fig. 40. Water adsorption on Ti-SBA-15 (180 mg) and Ti-MMM (202 mg) [from Trukhan

et al. (234)].


will impose a heavy economic penalty as a consequence of the high H2 partial

pressures that will be required to remove the DMDBTs.

Hulea et al. (238) demonstrated the ability of Ti-beta and Ti-HMS to oxidize

the thiophenic compounds to their corresponding sulfoxides and sulfones (with

H2O2 as the oxidant), which are then removed by conventional liquid–liquid

separation technology. The use of high-pressure equipment and the consumption

of large quantities of H2 can be avoided by this route. TS-1, as expected, exhibits

low activity as a consequence of the restricted access of DMDBT to the active

sites. Both Ti-beta and Ti-HMS catalysts exhibited high activities for the removal

of sulfur compounds from kerosene by mild oxidation with H2O2 (238). The best

results were obtained with acetonitrile as the polar solvent, because the oxidized

compounds (sulfoxides and sulfones) were fully soluble in this solvent (and they

are only partially soluble in ethanol and water) (Table XXXIV). During the

chemical treatment, the oxidized organic sulfur compounds (such as the sulf-

oxides and sulfones of dibenzothiophene and DMDBT) transfer completely to the

polar solvent, which is immiscible with kerosene. The oxidized product is then

recovered from the solvent, and the latter is recycled to the oxidation reactor.

TABLE XXXIV

Influence of catalyst and nature of solvent on the sulfur removal from kerosene (T ¼ 343 K)

Catalyst Solvent Reaction time

(h)

Phase Sulfur

(ppm)

Sulfur removal

(%)

- Acetonitrile Extraction Kerosene 1220 7.0

Ti-HMS Acetonitrile 9 Kerosene 190 85.5

Ti-HMS Acetonitrile 9 Acetonitrile 2500

Ti-beta Acetonitrile 5 Kerosene 80 94.0

Ti-beta Acetonitrile 5 Acetonitrile 2300

Ti-beta Ethanol 5 Kerosene 390 70.2



Ti-beta Ethanol 24 Kerosenea 80 94.0

Ti-beta Ethanol 24 Ethanol 1800

Ti-beta Water 10 Kerosene 840 36.0

Ti-beta Water 10 Kerosene 300 77.1

Ti-beta Water 10 Water 450

Adapted from Hulea et al. (238).a Kerosene washed with acetonitrile.


V.G. Oxidation of Oxygen-Containing Compounds

V.G.1. Alcohols

The oxidation of primary alcohols to aldehydes and secondary alcohols to

ketones proceeds smoothly on TS-1 and Ti-beta. On TS-1, because of diffusion

constraints, the oxidation rate decreases with reactant chain length, and linear

alcohols are oxidized faster than branched and cyclic alcohols, contrary to the

trends observed in homogeneous systems (198). By analogy with transition

metal complexes, it has been supposed (111) that intermediates such as that

illustrated in Scheme 16 can be responsible for the oxidation of alcohols with

H2O2. In the absence of diffusional constraints, Ti-beta exhibits (240) activity

and selectivity trends similar to those observed in homogeneous systems. Rates

increase with chain length, and cyclic/branched alcohols are more reactive than

linear alcohols. When alkyl substituents are introduced near the carbon atom

bearing the OH group, the reactivity of the molecule decreases, the decrease

being more pronounced when the number of such alkyl groups is increased.

These results are in agreement with the cyclic intermediate proposed in Scheme

16 and reflect the importance of the steric restrictions to form the transition state

complex at the Ti sites on the reactivity of molecular sieves. The apparent

activation energy was the same (70 kJ/mol) for both TS-1 and Ti-beta, indicating

that the oxidation of alcohol proceeds on both catalysts through similar cyclic

intermediates (239,240).

V.G.2. Ethers

The oxidation of both linear and cyclic ethers to the corresponding acids and

lactones by aqueous H2O2 as catalyzed by TS-1 and TS-2 was reported by

Sasidharan et al. (241) (Scheme 17 and Table XXXV). The titanosilicates

exhibited significantly better activity (about 55% conversion) and selectivity

(98%) than chromium silicates, although vanadium silicates totally failed to

catalyze the reaction. Such conversions are usually accomplished using either

stoichiometric amounts of chromium trioxide, lead tetraacetate, or ruthenium

tetroxide as oxidants (242) or catalytic amounts of RuO4 in the presence of

Scheme 16.


hypochlorite or periodate (243). The use of solid catalysts such as TS-1 has

significant environmental and economic advantages.

V.G.3. Phenols

When aromatic compounds containing a phenolic OH group are brought in

contact with titanosilicates in the presence of H2O2, two reactions are possible:

the first is the hydroxylation of the aromatic ring to give diphenols (Section

V.D). When the electron density in the ring is high (as in polyalkyl phenols)

and the ortho- and/or para position (with respect to the OH group) is vacant,

the formation of ortho- or para-benzoquinone also occurs. Indeed, in the

hydroxylation of phenol to catechol and hydroquinone, one of the major side

products (and the main cause of the tar formation) is the formation of benzo-

quinones and products derived from them. The benzoquinones of polyalkyl-

benzenes are starting materials for many products in the photographic and

fine chemicals industries. Trukhan et al. (234) reported the oxidation of 2,3,-

6-trimethylphenol (TMP) to trimethylbenzoquinone (TMBQ) catalyzed by

Ti-SBA-15, Ti-MMM, or TS-1 with aqueous H2O2 used as a reactant

(Table XXXVI). The Ti-SBA-15 samples with higher Si/Ti ratios, which

according to their diffuse reflectance UV spectra have higher dispersions of

Scheme 17.

TABLE XXXV

TS-1 catalyzed oxidation of various ethers with 30% H2O2

Reactant Product Yield (%)a

Dibutyl ether Butyric acid 54

Benzyl methyl ether Benzoic acid 65

Tetrahydrofutan g-Butyrolactone 55

Tetrahydropyran d-Valerolactone 42

Dihydropyran d-Valerolactone 40

1,4-Dioxan Keto-1,4-dioxane 5

Sasidharan et al. (241).a Isolated yield and the rest is essentially unreacted ether.


titanium species, exhibited a higher catalytic activity. The higher catalytic

activity of Ti-MMM was also thought to arise from the higher dispersion of Ti

in Ti-MMM. Apart from TMBQ, the main byproduct was the C–C coupling

dimer, 2,20,3,30,6,60-hexamethyl-4,40-biphenol. A small amount of the C–O

coupling dimer was also found. Experiments with fast catalyst filtration at the

reaction temperature confirmed (Fig. 41) that no further reactant conversion

occurred in the filtrate after catalyst removal, indicating that the oxidation takes

place on the catalyst surface and is a true heterogeneous process.

V.G.4. Ketones, the Baeyer–Villiger Oxidation

Baeyer–Villiger (BV) oxidation, induced by a peroxy acid or a H2O2/Lewis

acid system, organometallics, and metalloenzymes is an important reaction

for synthesizing lactones or esters from ketones. Bhaumik et al. (244) reported

that TS-1 is an efficient catalyst for BV oxidation of cyclic and aromatic ketones

(such as cyclohexanone and acetophenone, respectively) (Scheme 18, Tables

XXXVII and XXXVIII). Conversions and yields were higher in the absence

of any solvent in the triphase (solid catalyst along with two immiscible

liquid reactants (ketone þ aqueous H2O2). The addition of a few drops of H2SO4

increased the yield of the BV products. The titanium peroxo species, a Brønsted

acid stabilized by the presence of protic solvent was proposed by the authors to be

responsible for the BV reaction. In accordance with this proposal, Wang et al.

(245) later found that the Brønsted acid HZSM-5(Al) was also more active than

TS-1 in BV oxidation of cyclopentanone to d-valerolactane. The conversions

of the ketone and yield of the lactone were 47 and 15% for HZSM-5 vs. 35 and

10% for TS-1.

TABLE XXXVI

2,3,6-Trimethylphenol (TMP) oxidation with 30% aqueous H2O2

Catalyst Time (h) TMP conv. (%) TMBQ yield (%)

None 6 0 0

Ti-SBA-15 (38) 6 57 43

Ti-SBA-15 (19) 6 43 29

Ti-SBA (10) 6 31 30

Ti-MMM (39) 0.4 100 77

TS-1 (33.4) 6 14 8

Adapted from Trukhan et al. (234). Reaction conditions: TMP, 0.1 mol;

TMP/H2O2, 0.28; CH3CN, 3 mL; temperature, 353 K; Ti, 1.3 £ 1022 mmol.

Values in parentheses refer to Si/Ti ratios. TMBQ, trimethylbenzoquinone.


Fig. 41. TMP oxidation catalyzed by Ti-SBA-15, after filtration of the catalyst (full squares) and

without filtration (open squares) [from Trukhan et al. (234)].

Scheme 18.


TABLE XXXVIII

Baeyer–Villiger rearrangement and hydroxylation of acetophenone catalyzed by TS-1/H2O2 system

System Phasea Conv.

(%)

Product selectivities (mol%)b

PA o-HAP p-HAP PH CA HQ AA

TS-1/H2O2/Hþ Tri 31.0 49.7 16.6 16.0 7.0 1.0 1.1 8.6

TS-1/H2O2 Tri 7.0 27.0 2.8 5.6 12.6 7.4 12.3 32.3

TS-1/H2O2/Hþ Bi 6.1 61.0 – – 4.6 10.8 4.4 19.0

TS-1/H2O2 Bi – – – – – – – –

Blank/H2O2/Hþ Bi 5.5 31.8 6.9 - 24.9 2.8 3.5 30.1

Blank/H2O2/Hþ Mono – – – – – – – –

Adapted from Bhaumik et al. (244). Reaction conditions: reaction time, 12 h; reactant:H2O2 ¼ 1:1;

catalyst (TS-1, Si/Ti ¼ 29), 20 wt% with respect to reactant; temperature, 353 K.a Tri: solid catalyst þ two immisible liquid phases (organic reactant þ H2O2 in water); bi: solid

catalyst þ one homogeneous liquid phase (organic reactant þ aqueous H2O2 þ CH3CN as co-

solvent).b PA, phenyl acetate; o-HAP, o-hydroxy acetophenone; p-HAP, p-hydroxy acetophenone; PH,

phenol; CA, catechol; HQ, hydroquinone; AA, acetic acid.

TABLE XXXVII

Oxidation of cyclohexanone catalyzed by TS-1

System Phasea Conv.

(mol%)

Product selectivity (mol%)

1-Capro-

lactone

Hydroxy-

ketone

Diketone Cyclohexene

þ Epoxide

TS-1/H2O2/Hþ Tri 64.0 45.2 17.0 14.0 23.8

TS-1/H2O2/Hþ Bi 30.2 28.4 25.5 31.0 15.1

TS-1/H2O2 Tri 31.0 19.6 31.3 33.6 15.5

TS-1/H2O2 Bi 5.0 – 64.0 36.0 –

Adapted from Bhaumik et al. (244). Reaction conditions: reactant:H2O2 ¼ 1:1; catalyst (TS-1,

Si/Ti ¼ 29), 20 wt% with respect to reactant; temperature, 353 K.a Tri: solid catalyst þ two immisible liquid phases (organic reactant þ H2O2 in water); bi: solid

catalyst þ one homogeneous liquid phase (organic reactant þ aqueous H2O2 þ CH3CN as co-

solvent).


V.H. CyN Cleavage Reactions

Titanium silicate molecular sieves not only catalyze the oxidation of CyC

double bonds but can be successfully employed for the oxidative cleavage of

carbon–nitrogen double bonds as well. Tosylhydrazones and imines are oxidized

to their corresponding carbonyl compounds (243) (Scheme 19). Similarly,

oximes can be cleaved to their corresponding carbonyl compounds (165). The

conversion of cyclic dienes into hydroxyl ketones or lactones is a novel reaction

reported by Kumar et al. (165) (Scheme 20). Thus, when cyclopentadienes, 1,3-

cyclohexadiene, or furan is treated with aqueous H2O2 in acetone at reflux

temperatures for 6 h in the presence of TS-1, the corresponding hydroxyl ketone

or lactone is obtained in moderate to good yields (208).

V.I. Acid-Catalyzed Reactions

Acid catalysis by titanium silicate molecular sieves another area characterized by

recent major progress. Whereas only two categories of acid-catalyzed reactions

(the Beckmann rearrangement and MTBE synthesis) were included in the review

by Notari in 1996 (33), the list has grown significantly since then. In view of the

presence of weak Lewis acid sites on the surfaces of these catalysts, they can be

used for reactions that require such weak acidity.

Scheme 19.

Scheme 20.


V.I.1. Beckmann Rearrangement

The transformation of oximes to lactams (the Beckmann rearrangement) was

one of the earliest such acid-catalyzed reactions to be reported with TS-1 (138)

and TS-2 (247) catalysts. The rearrangement of cyclohexanone oxime to

1-caprolactam proceeds with high selectivity in the presence of TS-1, with high

catalyst stability (138,247).

V.I.2. Synthesis of Polycarbonate Precursors

Recently, Srivastava et al. (248, 249) reported the novel application of TS-1 and

Ti-MCM-41 in the synthesis of polycarbonate precursors such as cyclic

carbonates and dimethyl/diphenyl carbonates, avoiding toxic chemicals such as

phosgene or CO. With either TS-1 or Ti-MCM-41, cyclic carbonates were

prepared in high yields by cycloaddition of CO2 to epoxides such as epichloro-

hydrin, propene oxide, and styrene oxide at low temperatures and pressures

(Scheme 21, Table XXXIX). Although TS-1 and Ti-MCM-41 showed similar

activity for epoxides of smaller dimensions (such as epichlorohydrin and propene

oxide) (compare runs 1 and 3 and 5 and 7, Table XXXIX), Ti-MCM-41 was more

active for cycloaddition of CO2 to the larger styrene epoxide (compare runs 9

and 10, Table XXXIX). Although most of the experiments reported in Table

XXXIX were conducted with CH2Cl2 as solvent, similar (or better) yields were

obtained, even in the absence of any solvent (runs 3, 6, and 10, Table XXXIX).

However, the product was slightly colored. At higher temperatures/pressures/

reaction periods (e.g., 413 K, 24 bar, and 24 h), HPLC analyses showed

the formation of methanol-insoluble solid aliphatic polycarbonates. Apparently

the cyclic carbonate monomer had polymerized to give polycarbonates under

the influence of the weak acidity of the TS-1 system. In addition to the main cyclic

carbonate, the side products in the case of epichlorohydrin included 3-chloro-1,

2-propanediol, and 3-chloropropanaldehyde.

The cyclic carbonate could also be synthesized directly from the alkenes in

the same reactor by reacting the alkenes in the presence of Ti-MCM-41 with a

mixture of an epoxidizing agent (such as H2O2 or tert-butyl hydroperoxide) and

Scheme 21.


TABLE XXXIX

Synthesis of cyclic carbonates from epoxides and CO2

Run no. Catalyst Co-catalyst Temperature

(8C)

Run time

(h)

Epoxide Conv. of epoxide

(mol%)

TOF Selectivity for cyclic

carbonate (mol%)

1 TS-1 DMAP 120 4 EC 85.4 790 92.6

160 4 EC 94.2 872 97.0

2 TS-1 (3rd recycle) DMAP 120 4 EC 77.0 713 90.4

3 TS-1 (no solvent) DMAP 120 4 EC 89.6 829 97.5

4 TiMCM-41 DMAP 120 4 EC 78.8 938 84.0

5 TS-1 DMAP 120 6 PO 66.8 412 84.6

160 6 PO 94.0 580 83.0

6 TS-1 (no solvent) DMAP 120 6 PO 77.6 719 88.1

7 TiMCM-41 DMAP 120 6 PO 63.7 758 91.2

8 TS-1 DMAP 120 6 BO 76.6 354 70.9

9 TS-1 DMAP 120 8 SO 44.7 166 45.5

10 TiMCM-41 DMAP 140 10 SO 98.1 584 73.1

11 TiMCM-41(no solvent) DMAP 140 10 SO 100 595 82.0

From Srivatsava et al. (248). Reaction conditions: catalyst (TS-1: Si/Ti ¼ 36, Ti-MCM-41: Si/Ti ¼ 46), 100 mg; co-catalyst, 0.0072 mmol; epoxide,

18 mmol; CH2Cl2, 20 mL; CO2, 6.9 bar. DMAP: N,N-dimethylaminopyridine; EC: epichlorohydrin; PO: propylene oxide; SO: stytene oxide; BO:

a-butylene oxide; TOF: turnover frequency (moles epoxide converted per mole of Ti per hour.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

07

CO2 (Table XL). A conversion of 54.6% and cyclic carbonate selectivity of

55.6% were obtained when allylchloride was the reactant. Some ring-hydrolyzed

products were also detected. With styrene, a conversion of 50.4% and cyclic

carbonate selectivity of 26% were obtained. When the reaction was conducted

with TiMCM-41 as the catalyst and TBHP as the oxidizing agent, the conversions

of alkenes to epoxides (stage 1) were lower (allylchloride conversion ¼ 13.3%

and styrene conversion ¼ 44%), but the further conversion of epoxide formed

during the reaction to cyclic carbonate (stage 2) was almost 100% (Table XL).

As expected, TiMCM-41, with its larger pore diameter, was more active and

selective than TS-1 for the cycloaddition of CO2 to the epoxide (stage 2, rows 2

and 4, Table XL).

Aromatic polycarbonates are currently manufactured either by the interfacial

polycondensation of the sodium salt of diphenols such as bisphenol A with

phosgene (Reaction 1, Scheme 22) or by transesterification of diphenyl carbonate

(DPC) with diphenols in the presence of homogeneous catalysts (Reaction 2,

Scheme 22). DPC is made by the oxidative carbonylation of dimethyl carbonate.

If DPC can be made from cyclic carbonates by transesterification with solid

catalysts, then an environmentally friendlier route to polycarbonates using CO2

(instead of COCl2/CO) can be established. Transesterifications are catalyzed

by a variety of materials: K2CO3, KOH, Mg-containing smectites, and oxides

supported on silica (250). Recently, Ma et al. (251) reported the transesterifica-

tion of dimethyl oxalate with phenol catalyzed by Sn-TS-1 samples calcined

at various temperatures. The activity was related to the weak Lewis acidity of

Sn-TS-1 (251).

TABLE XL

Synthesis of cyclic carbonates from alkenes: epoxidation-cum-cycloaddition

Catalyst Alkene Oxidizingagent

Stage 1: alkeneto epoxide

Stage 2: epoxide tocyclic carbonate

Alkeneconversion toepoxide (%)

Epoxideselectivity

(%)

Epoxideconversion

(%)

Cyclic carbonateselectivity

(%)

TS-1 Allyl chloride H2O2 54.6 100.0 92.5 55.6

TS-1 Styrene H2O2 50.4 89.0 49.2 26.0

TiMCM-41 Allyl chloride TBHP 13.3 100 100 100

TiMCM-41 Styrene TBHP 44.0 93.1 97.2 83.4

Adapted from Srivatsava et al. (248). Runs with TS-1 (Si/Ti ¼ 36; 400 mg) were carried out with

26.2 mmol alkene, 0.0072 mmol DMAP, 14.7 mmol 50% H2O2 and CO2 (6.9 bar) in acetone (20 mL).

Runs with TiMCM-41 (Si/Ti ¼ 46; 100 mg) were carried out with 8 mmol alkene, 0.0036 mmol

DMAP, 8 mmol 40% TBHP in CH2Cl2 and CO2 (6.9 bar) in acetonitrile (6.4 g).


The transesterifications of chloropropene carbonate and propene carbonate

with methanol and phenol catalyzed by TS-1, Ti-MCM-41, and TiO2 (Table XLI)

have been reported (248). Neither TiO2 nor TS-1 showed any activity in the

transesterification reactions. Ti-MCM-41 catalyzed the reaction with a high

selectivity for DMC (86%). Ti-MCM-41 also catalyzes the transesterification of

cyclic carbonates with phenols (Table XLI).

TABLE XLI

Transesterification of cyclic carbonates with CH3OH and phenol catalyzed by Ti-MCM-41

Cyclic carbonate ROH Cyclic carbonate

conversion (mol%)

DMC selectivity

(mol%)a

DPC selectivity

(mol%)a

Chloropropylene carbonate CH3OH 26.5 86.2

Propylene carbonate CH3OH 5.1

Propylene carbonate C6H5OH 58.9 24.4

Adapted from Srivatsava et al. (248). Reaction conditions: for reactions with methanol (3.2 g)—

catalyst (TiMCM-41: Si/Ti ¼ 46), 400 mg; cyclic carbonate, 1.36 g; temperature, 393 K, reaction

time ¼ 2 h. For reactions with phenol (4.7 g) reaction time ¼ 17 h and rest all are the same.a Balance is phenyl ether.

Scheme 22.


V.I.3. Transesterification of Esters

Transesterification is a crucial step in several industrial processes such as

(i) production of higher acrylates from methylmethacrylate (for applications

in resins and paints), (ii) polyethene terephthalate (PET) production from

dimethyl terephthalate (DMT) and ethene glycol (in polyester manufacturing),

(iii) intramolecular transesterifications leading to lactones and macrocycles,

(iv) formation of alkoxy esters (biodiesel) from vegetable oils, and (v) co-

synthesis of dimethyl carbonate (an alkylating agent, octane booster, and

precursor for polycarbonates) and ethene glycol from ethene carbonate and

methanol (252,253).

Other than mineral acids and bases, compounds such as metal alkoxides

(aluminum isopropoxide, tetraalkoxytitanium, (RO)Cu(PPh3)n, PdMe(OCHCF3

Ph(dpe)), organotin alkoxides, etc.), non-ionic bases (amines, dimethylaminopyr-

idine, guanidines, etc.), and lipase enzymes also catalyze these transformations

(252). Tatsumi et al. (254) reported the synthesis of dimethylcarbonates from

ethene carbonate and methanol using K-TS-1 as a solid base catalyst. The trans-

esterification of dimethyl oxalate with phenol has also been reported recently

(251). TS-1 and Ti-MCM-41 catalyze transesterification reactions of aliphatic

esters selectively (152). Acidity measurements (infrared spectra of adsorbed

pyridine and TPD of NH3) had revealed the presence of only weak Lewis acid

sites on these samples. Catalytic activity was found to parallel the acid strength.

Both increased in the order TS-1 , Ti-MCM-41 , amorphous TiO2–SiO2. TS-1

catalyzed the transesterifications (Tables XLII and XLIII) of linear esters (ethyl-

acetoacetate and diethylmalonate), but failed for cyclic esters such as propene

carbonate. Ti-MCM-41 and amorphous TiO2–SiO2 were found to be superior

for the cyclic esters (Tables XLIV and XLV). The catalysts could be recycled

without any loss in activity/selectivity.

V.I.4. Carbon–Carbon Bond Formation Reactions

The Mukaiyama-type aldol reactions (255) between silyl enol ethers and

aldehydes to give b-hydroxy esters/aldols provide a facile method for C–C bond

formation. They are facilitated by a variety of Lewis acids, including TiCl4,

SnCl4, and ZnCl4, used in either stoichiometric or catalytic amounts under

homogeneous conditions. A few solid catalysts, such as Nafion-117, zeolite Ca–

Y, montmorillonite clay, and SiO2–Al2O3, have also been reported to be active

for these reactions (256). Sasidharan and Kumar (257) recently investigated the

Mukaiyama-type reactions with a variety of metallosilicates including TS-1 and

Al-free Ti-beta. Michael addition reactions of silyl enol ethers with various

a,b-unsaturated carbonyl compounds were also investigated with these catalysts.

In the Mukaiyama aldol reaction of methyl trimethylsilyl dimethylketene acetal


TABLE XLII

Transesterification of ethylacetoacetate with various alcohols (ROH) over TS-1

Entry ROH Transester product Conv. (mol%) Product yield (%)

1 95.6 92.9

2 100 87.1

3 97.6 90.7

4 99.2 85.0

5 96.2 84.3

(Continued)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

11

TABLE XLII

Continued

Entry ROH Transester product Conv. (mol%) Product yield (%)

6 96.4 95.3

7 86.4 69.5

8 83.1 66.9

9 CH3(CH2)7CHyCH(CH2)7CH2OH 87.6a

From Srinivas et al. (152). Reaction conditions: catalyst (TS-1; Si/Ti ¼ 33), 130 mg; ethylacetoacetate, 5 mmol; ROH, 15 mmol; temperature ¼ 383 K, run

time ¼ 4 h.a Isolated yield.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

12

(silyl enol ether) with benzaldehyde (Scheme 23) catalyzed by various metallo-

silicates, TS-1 and Ti-beta gave the highest yields (85–87%) of the product

b-hydroxy ester (aldol) (Table XLVI). The number of turnovers for different

isomorphously substituted metallosilicates followed the order Ti . Sn . V .

Al. Table XLVII illustrates the 1,4-Michael addition of various a,b-unsaturated

carbonyl compounds with silyl enol ether (Scheme 24). The reactions were

carried out in the absence of H2O or H2O2. The product yields mentioned in

Tables XLVI and XLVII are isolated yields; the selectivity for the aldols as well

as the Michael addition products was always 100%, regardless of conversion, and

no side products were observed. Among the various solvents investigated,

tetrahydrofuran was found to be the best. The authors attributed the excellent

activity of TS-1 and Ti-beta in the aldol condensation and Michael addition

reactions to the “oxophilic Lewis acidity” of Ti4þ ions (257).

TABLE XLIII

Transesterification of diethyl malonate with various alcohols catalyzed by TS-1

ROH Conversion (%) Selectivity (%) Products distribution (%)

Mono Di

n-Propanol 97.5 98.4 26.6 73.4

n-Butanol 99.3 97.0 16.0 84.0

n-Butanolb 95.8 100 46.4 53.6

n-Butanol (recycle I)a 95.0 100 45.9 54.1

n-Butanol (recycle II)a 94.4 100 46.6 53.4

n-Hexanol 99.7 100 10.9 89.1

n-Octanol 100 100 82.6 17.4

Isobutanol 95.6 96.3 58.0 42.0

Cyclohexanol 100 100 34.4 65.6

Benzyl alcohol 84.2 88.8 39.0 61.0

From Srinivas et al. (152). Reaction conditions: catalyst (TS-1; Si/Ti ¼ 33), 130 mg; diethyl

malonate, 5 mmol (0.8 g); ROH, 15 mmol; temperature, 383 K; run time, 12 h.a Reaction conditions are same except the temperature, 353 K.


V.I.5. Formation of Pinacols

The name “pinacol” denotes vicinal diols with four alkyl groups; when all the

alkyls are methyl, it is called pinacol (CH3)2C(OH)–C(OH)(CH3)2. These

compounds are the starting materials for the manufacture of many pesticides,

pharmaceuticals, fragrances, photographic chemicals, and crop protection

chemicals. They are usually made by dihydroxylation of alkenes by OsO4 or

KMnO4. Both of these toxic reagents are used in stoichiometric quantities.

Another strategy to make these 1,2-diols is reduction of aldehydes and ketones

with reactive metals such as Na, Mg, or Al. But many side products are formed as

a result of coupling reactions.

Sasidharan et al. (258) reported the formation of pinacols from alkenes cata-

lyzed by various titanosilicates. Aluminum-free Ti-beta exhibited better activity

and pinacol selectivity than TS-1, TS-2, Ti-MCM-22, or mesoporous Ti-MCM-41

(Table XLVIII). The side products included the pinacolone, alcohol, and

oligomers. The epoxide was the initial product, which underwent acid-catalyzed

nucleophilic ring-opening by H2O molecules to give the pinacol (Scheme 25).

TABLE XLIV

Comparative activity of TS-1, Ti-MCM-41 and amorphous TiO2–SiO2 in transesterification of (a)

ethylacetoacetate with benzyl alcohol and allylalcohol and (b) diethylmalonate with allylalcohol

Catalyst Benzyl alcohol Allylalcohol

Ester conversion

(mol%)

Transester

yield (%)

Ester conversion

(mol%)

Transester

yield (%)

(a) Ester–ethylacetoacetate (Run time ¼ 4 h)

TS-1 86.4 69.5 83.1 69.8

Ti-MCM-41 93.7 90.2 85.2 84.5

Amorphous TiO2-SiO2 95.2 91.9 87.3 86.1

Monotransester

selectivity (%)

Ditransester

selectivity (%)

(b) Ester–diethylmalonate (Run time ¼ 12 h); alcohol–n-butanol

TS-1 58.2 57.4 59.0 41.0

Ti-MCM-41 66.5 65.6 87.2 12.8

Amorphous TiO2-SiO2 68.6 66.7 89.2 10.8

Adapted from Srinivas et al. (152). Reaction conditions: ester, 5 mmol; alcohol, 15 mmol; catalyst,

130 mg; temperature, 383 K.


V.I.6. Oxidative Dehydrogenation

The oxidative dehydrogenation of propane to give propene catalyzed by TS-1,

Ti-beta, Ti-MCM-41, TiO2-silicalite-1, or others was investigated by Schuster

et al. (259). TS-1 was the best catalyst, with a selectivity of 82% for propene at a

propane conversion of 11% (Fig. 42). Sulfation of TS-1 by H2SO4 prior to the

reaction increased the conversion to 17%, with a selectivity of about 74%.

Although conversion of propane was higher on Ti-beta and Ti-MCM-41,

selectivity for propene was much lower; CO2 was the main product. Lewis acid

sites were considered to be the major active sites (259).

TABLE XLV

Transesterification of cyclic propylenecarbonate with different alcohols and

phenol catalyzed by titanosilicates

Titanosilicate ROH Reaction

time (h)

Conversion of

propylene

carbonate (mol%)

Selectivity of

transester

product (mol%)a

TS-1 Methanol 2 Nil –

TiMCM-41 Methanol 2 5.1

Phenol 8 58.9 24.4

Amorphous TiO2-SiO2 Methanol 4 71.4 48.2

8 86.0 51.2

Ethanol 8 73.0 61.8

Propanol 12 86.3 69.4

n-Butanol 12 85.0 73.4

n-Hexanol 12 49.0 61.5

Adapted from Srinivas et al. (152). Reaction conditions: catalyst (TS-1 or TiMCM-41), 400 mg;

propylene carbonate, 1.36 g, ROH (alcohol, 3.2 g; phenol, 4.7 g); temperature, 393 K. Reaction

conditions: catalyst (amorphous titanosilicate), 400 mg; propylene carbonate, 1.02 g (0.01 mol);

ROH, 0.1 mol; temperature, 423 K.a Balance is the corresponding ether.

Scheme 23.


V.J. Photocatalysis

V.J.1. Photocatalytic Degradation of Pollutants

The oxidation of small concentrations of aromatic compounds in industrial

effluents using UV radiation and catalysts such as TiO2 is gaining in importance

(260). The hydroxyl radicals generated on TiO2 under UV irradiation are the

agents of photodegradation. To increase the efficiency of the process, the TiO2

has been dispersed on SiO2 (261). Titanosilicates such as TS-1 and Ti-beta have

two inherent advantages as photodegradation catalysts: (i) they are hydrophobic

and, hence, adsorb selectively the aromatic pollutants from aqueous effluents,

thereby facilitating the photocatalytic efficiency for charge transfer from the

catalyst to the pollutants; (ii) the high surface area and atomic dispersion of Ti

enable an efficient use of the metal. Kang et al. (262) compared TiO2 and two

samples of TS-2 (TS-2 and TS-2h) in the photodegradation of various phenols.

TABLE XLVI

Activities of various metallosilicates for Mukaiyama aldol reaction of benzaldehyde

with silyl enol ether

Catalyst Si/M ratio

(product)

Particle size

(mm)

Micropore volume

(mL/g)

Yield

(%)

TONa

Ti-ZSM-5 (TS-1)b 33.5 0.1–0.2 0.138 85.0 9.2

Sn-ZSM-5b 73.5 0.2–0.4 0.132 25.0 7.35

V-ZSM-5b 86.0 0.4–0.6 0.135 10.0 1.9

H-ZSM-5b 40.0 0.3–0.5 0.153 - -

Ti-betab 43.0 0.2–0.3 0.269 87.0 11.6

Al-betab 26.7 0.3–0.4 0.274 29.0 1.3

Sn-ZSM-12b 78.0 1–2 0.169 15.0 4.5

Na–Y 2.5–3.0 0.5–0.7 0.343 - -

La–Yc 2.5–3.0 0.5–0.7 0.292 37.0 6.5d

Re–Yc 2.5–3.0 0.5–0.7 0.270 50.0 7.3d

Zn/ZSM-5c 40.0 0.3–0.5 0.141 34.0 4.4e

Adapted from Sasidharan and Kumar (257). Reaction conditions: catalyst, 150 mg; methyl trimethyl-

silyl dimethylketene acetal (silyl enol ether), 10 mmol; benzaldehyde, 10 mmol; dry THF as dispersion

medium, 10 mL; temperature, 333 K; reaction time, 18 h. Yield refers to the isolated product yield.a Moles of product per mole of metal per hour.b The metal atom is substituted in the tetrahedral position.c La ¼ 2.3 wt%; combination of all the rare-earth metals ¼ 2.85 wt% and Zn ¼ 2.63 wt%.d TON based on rare-earth metals.e TON based on Zn.


The surface areas of the three catalysts were 58 (TiO2), 360 (TS-2), and 550

(TS-2h) m2/g, respectively. UV-irradiation of solutions (1024 M) containing 4-

chlorophenol (4-CP) in the presence of suspended TiO2, TS-2, or TS-2h yielded

time-dependent spectra from which the concentration of unconverted 4-CP was

estimated. Figure 43 is a plot of the relative concentration of 4-CP as a function

TABLE XLVII

Michael addition of various a,b-unsaturated carbonyl compounds to silyl enol ether

catalyzed by Ti-beta and TS-1

a,b-Unsaturated carbonyl compounds Product Product yield (%)a

Ti-beta TS-1

Methyl methacrylate, (2a) 3a 53.0 47.0

Ethyl methacrylate, (2b) 3b 41.0 35.0

2-Ethylhexyl acrylate, (2c) 3c 39.0 36.0

2-Hydroxyethyl methacrylate, (2d) 3d 41.0 39.0

Methyl vinyl ketone, (2e) 3e 45.0 49.0

Cyclohexenone, (2f) 3f 39.0 36.0

2-Methylcyclohexenone, (2g) 3g 35.0 33.0

Adapted from Sasidharan and Kumar (257). Reaction conditions: catalyst, 150 mg; methyl

trimethylsilyl dimethylketene acetal (silyl enol ether), 10 mmol; a,b-unsaturated carbonyl

compounds, 10 mmol; dry THF, 10 mmol; reaction temperature, 333 K; reaction time, 14 h.

Structures of a,b-unsaturated carbonyl compounds (2a–2g) and products (3a–3g) are shown

in Scheme 24.a Isolated yield by column chromatography and the rest is unconverted starting material.

Scheme 24.


of irradiation time for the three catalysts. The activity decreases in the order

TS-2h . TS-2 . TiO2. Notwithstanding the lower surface Ti concentration (by

about 19%) and the larger band gap of the TS-2 catalysts relative to TiO2, the

photodecomposition rate is enhanced on TS-2 and TS-2h. The greater photo-

reactivity was attributed to the increased adsorption of 4-CP resulting from

TABLE XLVIII

Formation of pinacol over various titanium-silicates

Catalysta Conv.

(mol%)

H2O2 selectivity

(%)

Product selectivity (%)

Epoxide Pinacol Pinacolone DMBb Othersb

Ti-beta (43) 55.3 80.1 1.3 92.9 1.9 4.3 0.5

Ti-Al-beta (40) 51.2 76.5 1.1 82.6 3.7 15.6 0.4

TS-1 (33) 39.2 61.5 3.9 88.0 1.3 1.0 5.9

TS-2 (46) 21.2 57.0 4.0 83.6 1.2 1.9 9.0

Ti-MCM-22 (51) 22.6 54.5 3.4 86.0 2.0 5.4 5.1

Ti-MCM-41 (50) 48.2 65.0 1.6 96.3 1.1 0.7 0.3

Adapted from Sasidharan and Kumar (258). Reaction conditions: 2,3-dimethyl-2-butene, 10 mmol;

H2O2 (31 wt% aqueous solution), 10 mmol; catalyst, 20 wt% with respect to substrate; water (as

dispersion medium), 5 mL; temperature, 333 K; reaction time, 6 h.a The figures in the parentheses represent the Si/Ti ratios.b DMB, 2, 3-dimethyl-2-butanol and “others” include oligomers.

Scheme 25.


Fig. 42. Catalyst screening for the oxidative dehydrogenation of propane to propene. T ¼ 823 K;

molar ratios C3H8/O2/N2/H2O ¼ 5/25/25/45; GHSV ¼ 1300 h21; mcat ¼ 1:4 2 8:0 g; vcat ¼ 5 ml

[from Schuster et al. (259)].

Fig. 43. Time dependence of the relative concentration of 4-CP at 225 nm of illuminated 4-

CP aqueous solutions in the presence of TiO2, TS-2, and TS-2h catalysts in suspension [from Kang

et al. (262)].


the greater hydrophobic surface areas of TS-2 and TS-2h as well as their greater

total surface areas relative to TiO2.

V.J.2. Photocatalytic Synthesis

Reduction of CO2 with H2O to give useful chemicals using sunlight is one of

the holy grails in solar energy-to-fuels and chemicals conversion. Towards this

goal, Anpo et al. (263) used Hg lamp radiation (l . 280 nm) to reduce CO2

with H2O to CH4 and CH3OH at 328 K using titanosilicate molecular sieves,

TS-1, Ti-MCM-41, and Ti-MCM-48 (Fig. 44). The order of reactivity was

Ti-MCM-48 . Ti-MCM-41 . TS-1 . TiO2. The Ti-containing zeolites led to

the formation of considerable amounts of the CH3OH, although the formation of

the CH4 was found to be the major reaction on bulk TiO2 (Fig. 44). Although

both Ti-MCM-41 and Ti-MCM-48 are mesoporous, the pore geometry is three-

dimensional in the latter and one-dimensional in the former. Addition of Pt

onto Ti-MCM-48 increased its photocatalytic activity. However, only the for-

mation of CH4 is promoted, being accompanied by a decrease in the CH3OH

yields (Fig. 45). Anpo et al. (263) proposed that CO2 is reduced to CO and

subsequently to C radicals although H2O photodecomposes to H and OH

radicals. Reaction of OH and H with the carbon species yields CH3OH and CH4,

respectively (263).

The mechanism of CO2 photoreduction in TS-1 with methanol as the electron

donor was also investigated by Ulagappan and Frei (264), who used in situ FTIR

Fig. 44. Yields of CH4 and CH3OH in the photocatalytic reduction of CO2 with H2O on TiO2

powder: (a) TS-1; (b) Ti-MCM-41; (c) Ti-MCM-48; and (d) zeolite catalysts [from Anpo et al. (263)].


spectroscopy. The reaction was induced by 266-nm excitation of the Ti4þ–

O22 ! Ti3þ–O2 ligand-to-metal charge transfer transition of the framework

center. HCO2H, CO, and HCO2CH3 were the observed products. The CO

originates from secondary photolysis of HCO2H, although HCO2CH3 is formed

by the spontaneous Tischenko reaction of HCHO, which is the initial oxidation

product of methanol. HCO2H is the primary 2-electron reduction product of CO2

at the Ti centers, a result that suggests that C–H bond formation occurs in the

initial steps of CO2 activation.

V.J.3. deNOx Reactions

TS-2 exhibited high photocatalytic activity (with a 75-W high-pressure Hg

lamp) for the direct decomposition of NO into N2 and O2 and N2O at 275 K

(265), with a high selectivity (76%) for the formation of N2. The yields (in

mmol/g of TiO2 h) of N2 and N2O were 12 and 4, respectively. In the case of

isolated Ti ions in 4-fold coordination present in TS-2, charge transfer excited

complexes (Ti3þ–O2)p are formed under UV irradiation. Electron transfer

from Ti3þ to the p-antibonding orbital of NO takes place, and simultaneously

the electron transfer from the p-bonding orbital of another NO into the hole-

trapped center (O2) occurs. These electron-transfer processes lead to the direct

decomposition of two sets of NO on the (Ti3þ–O2) species, to selectively form

N2 and O2 (265). On the other hand, when Ti ions are present in an aggregated

Fig. 45. The effects on Pt-loading on the yields of CH4 and CH3OH in the photocatalytic reduction

of CO2 with H2O on Ti-MCM-48 zeolite catalyst: (a) Ti-MCM-48; (b) Pt-loaded Ti-MCM-48

(0.1 wt% Pt); and (c) Pt-loaded Ti-MCM-48 (1.0 wt% Pt) [from Anpo et al. (263)].


form (as in anatase), the photoformed holes and electrons are rapidly separated

from each other. This separation prevents the simultaneous activation of two NO

molecules on the same active site, resulting in the formation of N2O and NO2

instead of N2 and O2 (265).

V.K. Influence of Solvents

Solvents are usually used to keep both reactants and products in a single phase.

Apart from enabling the proper mixing of the reactants, solvents can also affect

conversions and product selectivities through interaction with the active sites

and the transition state. The influence of the dielectric constant of the solvent on

the mode of cleavage of the O–O bond in H2O2 (hetero- vs. homolytic cleavage)

and consequently on product distribution was mentioned above (Section V.B).

The influence of solvents on oxidation reactions catalyzed by TS-1 had been

investigated by both experimental (111,266,267) and theoretical (63,268,269)

methods. Atoguchi and Yao (267) examined the effect of solvents (various

mixtures of H2O and CH3OH) on the oxidation of phenol catalyzed by TS-1 both

experimentally (Table XLIX) and by DFT calculations for cluster models made

up of the Ti center having the tetrahedral structure, Ti(OSiH3)4, a H2O2, and a

solvent molecule. Water addition to methanol increases the dielectric constant of

the reaction medium and accelerates the catalytic oxidation of phenol (increasing

the conversion from 43.6 to 70.2%). The amount of dihydroxy benzenes increases

from 4.3 to 6.6 mmol (Table XLIX).

Additional results of the enhancement in phenol conversion (to dihydroxy

benzenes) and oxidation of allyl alcohol (to glycidol and allylic oxidation

products) catalyzed by TS-1 in various solvents are illustrated in Fig. 46. In

solvents with high dielectric constants, the heterolytic cleavage of the O–O bond

TABLE XLIX

Phenol oxidation over TS-1 in H2O and methanol mixture solvent

CH3OH:H2O

(wt%)

Dielectric

constant (1)

Phenol

conversion (%)

Hydroquinone þ

catechol (mmol)

Hydroquinone/

catechol

Selectivity

(%)a

85.6: 14.4 39.2 43.64 4.33 1.99 91.97

77.0: 23.0 43.1 51.67 5.06 1.82 91.76

42.8: 57.2 58.9 70.16 6.58 1.44 87.24

Adapted from Atoguchi and Yao (267). Reaction conditions: catalyst, 0.2 g; phenol, 1.0 g; 30%

aqueous H2O2, 1.2 g; solvent, 5 g; temperature, 349 K; time, 3 h.a Selectivity (%) ¼ {(produced hydroquinone þ catechol)/(consumed phenol)} £ 100 (mol/mol).


is probably dominant when the partially or totally ionic intermediates are

stabilized by the solvent. Solvents can influence the reactivity of the hydroperoxo

titanium species Ti(O2H) either through changes in the dielectric constant of

the reaction medium (as discussed in Section V.B.1) or by specific coordination

to the Ti center. Care should be taken to distinguish between the two effects in

investigations of solvent effects. Furthermore, molecules such as acetone and

acetonitrile are oxidized by H2O2, forming 2-hydroxy-2-hydroperoxy propane

and peroxyimidic acid CH3–CyNH(OOH), respectively (269–270), affecting

the rate of the oxidation and H2O2 selectivity.

An interesting observation reported in Table XLIX is the increase in the

hydroquinone/catechol ratio from 1.44 to 1.99 when the dielectric constant of

the medium is decreased from 58.9 to 39.2 by addition of methanol to water. A

similar increase in the hydroquinone/catechol ratios was also observed in phenol

hydroxylation catalyzed by TS-1 (266) in dioxane-water and tert-butyl alcohol-

water mixtures. The para/ortho ratio increased nearly 10-fold when 10% dioxane

was added to water. Similarly, the para/ortho ratio more than doubled (1.3–3.0)

when 10% tert-butyl alcohol was added to water. An opposite trend, namely,

a decrease in the para/ortho ratio from 1.4 to 0.6, was observed when 10%

formamide ð1 ¼ 108Þ was added to water. Because of geometric constraints in

the MFI pores, catechol is expected to be formed more easily on the external

surface of TS-1 crystallites than in the pores (91). Hydroquinone, less spatially

demanding, can form in the TS-1 channels. A greater coverage of the hydrophobic

Fig. 46. Influence of solvent dielectric constant (logarithm (ln) values) on (a) phenol hydroxyla-

tion [data taken from Thangaraj et al. (266)] and (b) epoxidation of allyl alcohol catalyzed by TS-1

[data from Wu and Tatsumi (229)].


external surface of the TS-1 crystals by the less polar solvents having lower

dielectric constants (alcohol and dioxane in alcohol–water and dioxane–water

mixtures, respectively) will suppress the formation of catechol.

The influence of solvents of varying polarity in the epoxidation of allyl

alcohol catalyzed by TS-1, Ti-beta, and Ti-MWW is shown in Table L (229).

The surface of TS-1 is hydrophobic. Hydrophilic molecules such as H2O do not

compete with the reactant, allyl alcohol, for either diffusion in the pores or

coordination at the Ti site. On the contrary, hydrophobic solvent molecules do

compete with reactant molecules for adsorption on the hydrophobic TS-1

surface. Hence, the conversion of the reactant is higher in hydrophilic H2O than

in hydrophobic isopropanol. No steric constraints in the pores are anticipated for

any of the above solvents. Ti-beta is relatively more hydrophilic and, hence, it is

not surprising that H2O inhibits the conversion strongly. In general, large-pore

and mesoporous Ti-silicates behave similarly to amorphous Ti–SiO2 catalysts

in this respect. Hence, TBHP is a better oxidant than aqueous H2O2 for such

materials. Ti-MWW is even more active and selective. Conversion follows a

trend similar to that observed for TS-1, except that it is the highest when

CH3CN is the solvent. Selectivity to the epoxide is high for reaction catalyzed

by either TS-1 or Ti-MWW. The selectivity of Ti-beta is low, especially in

solvents of high coordinating ability such as the higher alcohols.

V.L. Influence of Silylation

One of the reasons for the low selectivity of the mesoporous Ti silicates is their

surface hydrophilicity, which is caused by the presence of a large number of

surface Si–OH and Ti–OH groups. Because these mesoporous materials are

better suited than TS-1 to the oxidation of large, bulky molecules, the passivation

of these OH groups (e.g., by silylation) may improve catalyst activity and

selectivity. Attempts have been made to reduce the concentrations of such OH

groups by silylating them with various alkyl silanes (Table LI) (273).

The treatment leads to a significant improvement in alkene conversion in

cyclohexene epoxidation in the case of Ti-MCM-41 and Ti-MCM-48 (273).

Although epoxide selectivity improved in the former case, there was a decrease in

the latter. In the case of hexane oxidation, silylation did not improve the

conversion. An enhancement in the number of turnovers and selectivity for the

epoxide on silylation was also observed in the cyclohexene epoxidation with

TBHP catalyzed by Ti-SBA-15 (Table LII) (274). Ti-SBA-15 was claimed to be

thermally more stable than Ti-MCM-41. Ti leaching was absent.

A better understanding of the changes in surface structure during silylation is

needed before the potential advantages of silylation of these mesoporous

materials are realized. A potential pitfall in silylation reactions is the silylation of


TABLE L

Epoxidation of ally alcohol with H2O2 in various solvents

Solvent Ti-MWW (Si/Ti ¼ 46) (mol%) TS-1 (Si/Ti ¼ 36) (mol%) Ti-beta (Si/Ti ¼ 42) (mol%)

AA conv. Prod. Sel.a H2O2 AA conv. Prod. Sel.a H2O2 AA conv. Prod. Sel.a H2O2

Gly. Others Conv. Eff. Gly. Others Conv. Eff. Gly. Others Conv. Eff.

MeCN 87.0 99.9 0.1 87.9 99.0 26.8 82.6 17.3 28.5 94.1 13.9 75.4 24.6 18.4 75.5

Water 82.3 99.9 0.1 84.3 97.6 34.6 96.0 4.0 36.6 94.5 2.8 92.6 7.4 9.6 29.2

MeOH 34.5 75.7 24.3 35.9 96.1 34.2 86.6 13.4 36.2 94.5 16.7 42.0 58.0 21.6 77.3

EtOH 32.5 91.0 9.0 33.0 98.5 24.4 94.6 5.4 29.8 81.8 15.1 59.5 40.5 28.6 52.8

1-PrOH 30.1 96.0 4.0 37.5 80.3 12.6 95.6 4.4 16.1 78.6 – – – – –

Acetone 41.5 96.7 3.3 42.5 97.6 31.0 92.8 7.2 36.6 84.7 11.9 41.4 58.6 26.3 45.2

Dioxane 27.8 96.0 4.0 28.6 97.2 – – – – – 5.2 78.3 21.7 6.5 80.0

Adapted from Wu and Tatsumi (229). Reaction conditions: catalyst, 70 mg; allyl alcohol (AA), 10 mmol; H2O2, 10 mmol; solvent, 5 mL; temperature,

333 K; time, 0.5 h.a Gly, glycidol; others, solvolysis products, glycerol and alkyl glycerol ethers, etc.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

25

TABLE LI

Effect of trimethylsilylation on catalytic activity of Ti-containing mesoporous molecular sieves

Catalysta Conv.

(mol% of max.)

TON

(mol/(mol Ti))

Selectivity (%) H2O2 decom

(%)

Alcohol Ketone Epoxide Diol

Cyclohexene

Ti-MCM-41 (nonsil)b 0.72 5.4 30.0 15.2 0 54.7 57.6

Ti-MCM-48 (nonsil)c 2.1 6.1 26.7 32.8 4.7 35.7 61.9

Ti-MCM-41(sil)b 13.3 112.1 14.4 21.0 13.9 50.7 0

Ti-MCM-48 (sil)b 38.5 120.9 21.3 17.0 2.2 59.4 0

2-ol 3-ol 2-one 3-one

Hexane oxidation

Ti-MCM-41 (nonsil)c 0 0 – – – – 74.7

Ti-MCM-41 (sil)c 0.06 0.5 40.5 59.5 0.0 0.0 0.0

Ti-MCM-41 (sil)d 0.2 6.8 22.0 22.9 31.0 24.1 97.3

Ti-MCM-48 (nonsil)c 0 0 – – – – 75.0

Ti-MCM-48 (sil)b 0.17 0.52 45.4 54.6 0 0 20.2

Adapted from Tatsumi et al. (273).a Ti-MCM-41 (non-sil) (Si/Ti ¼ 123, SBET ¼ 1015 m2/g, pore diameter ¼ 2.32 nm, pore volume ¼ 0.88); Ti-MCM-41 (sil) (Si/Ti ¼ 139, SBET ¼ 879 m2/g,

pore diameter ¼ 1.90 nm, pore volume ¼ 0.82); Ti-MCM-48 (non-sil) (Si/Ti ¼ 47, SBET ¼ 1048 m2/g, pore diameter ¼ 2.32 nm, pore volume ¼ 0.91); Ti-

MCM-48 (sil) (Si/Ti ¼ 51, SBET ¼ 839 m2/g, pore diameter ¼ 1.90 nm, pore volume ¼ 0.71).b Reaction conditions: catalyst, 50 mg; reactant, 25 mmol; H2O2, 5 mmol; temperature, 323 K; time, 3 h.c Catalyst, 50 mg; reactant, 25 mmol; H2O2, 5 mmol; temperature, 323 K; time, 2 h.d Catalyst, 50 mg; reactant, 100 mmol; H2O2, 20 mmol; temperature, 353 K; time, 16 h.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

26

the Ti–OH groups in the tripodal Ti centers. Such Ti–OH groups play an

essential role in the formation and reactivity of the titanium oxo groups

(Sections III and IV). Their elimination by silynation will lead to a reduction in

the number of active sites. Elimination of the Si–OH groups without affecting the

Ti–OH groups is difficult and may account for some of the conflicting results of

silynation reported in the literature.

VI. Structure-Activity Correlations

The majority of the titanium ions in titanosilicate molecular sieves in the

dehydrated state are present in two types of structures, the framework tetrapodal

and tripodal structures. The tetrapodal species dominate in TS-1 and Ti-beta, and

the tripodals are more prevalent in Ti-MCM-41 and other mesoporous materials.

The coordinatively unsaturated Ti ions in these structures exhibit Lewis acidity

and strongly adsorb molecules such as H2O, NH3, H2O2, alkenes, etc. On inter-

action with H2O2, H2 þ O2, or alkyl hydroperoxides, the Ti ions expand their

coordination number to 5 or 6 and form side-on Ti-peroxo and superoxo complexes

which catalyze the many oxidation reactions of NH3 and organic molecules.

VI.A. Structure of Titanium Species and Activity

Attempts have been made to find correlations between the types and concentra-

tions of the various surface groups and titanium oxo complexes, on the one hand,

TABLE LII

Influence of silynation on epoxidation of cyclohexene with TBHP over Ti-SBA-15

Sample Si/Ti Conversion

(mol%)

TON

(mol/mol Ti)


Oxide Diolsa Allylicb

Ti-SBA1 80 1.9 76 83.9 6.5 9.7

Ti-SBA2 27 2.8 38 86.9 6.8 6.3

Ti-SBA1-silc 87 21.1 843 97.0 2.0 1.9

Ti-SBA2-silc 28 29.8 437 96.0 2.4 1.6

Adapted from Wu et al. (274). Reaction conditions: catalyst, 0.05 g; cyclohexene, 30 mmol, TBHP

(70%), 30 mmol; CH3CN, 10 mL; temperature, 333 K; time, 2 h.a 1,2-Cyclohexanediols.b Products of allylic oxidation: 2-cyclohexen-1-ol and 2-cyclohexen-1-one.c Trimethylsilylated by refluxing in hexamethyldisilazane/toluene for 2 h.


and their catalytic activity and selectivity on the other. The concentration of

framework Ti ions was the earliest structural parameter to be related to the

catalytic activity of TS-1. The original patent of Taramasso et al. (5) itself

claimed that the intensity of the 960-cm21 band, indicative of the concentration

of Ti in framework positions, is related to the catalytic activity. Later, Thangaraj

et al. (275) observed that the catalytic activity of TS-1 in phenol conversion was

proportional to the molar ratios x ( ¼ Ti/(Si þ Ti)) at low Ti concentrations

ðx , 0:02Þ and suggested that these Ti ions are responsible for the observed

catalytic activity. A similar conclusion was also reached by Mantegazza et al.

(276), who observed that at low Ti concentrations the activity of TS-1

(represented by the turnover number) in ammonia oxidation, cyclohexanone

ammoximation, and propene epoxidation was proportional to the mole fraction of

Ti in the framework (276). There is, hence, a consensus that, on TS-1 (and

probably Ti-beta), tetrapodal Ti ions in framework tetrahedral positions are

responsible for the catalytic activity. On Ti-MCM-41 (and probably other similar

mesoporous materials), the XANES/XAFS investigations of Thomas and Sankar

(104) show that the tripodal Ti centers are responsible for catalytic activity in the

conversion of cyclohexene to its epoxide with TBHP as the oxidant. From their

in situ XAFS data, these authors concluded that during the catalytic reaction the

original four-coordinated Ti4þ centers in the tripodal species expand their

coordination sphere to six (Section V.C.5).

Chaudhari et al. (277) had observed a linear dependence of H2O2 selectivity on

Ti content in Ti-MCM-41 in the hydroxylation of 1-naphthol to 1,2-dihydroxy

naphthalene with aqueous H2O2 (Fig. 47). Both XAS and EPR results had

indicated the presence of mainly the tripodal titanium sites on Ti-MCM-41. As

a consequence of the large surface area of the material, these sites are well

dispersed, leading to the linear dependence of catalytic activity on Ti content.

Such detailed structural information about surface Ti species is not available

for other Ti–SiO2 mesoporous materials. The results of Guidotti et al. (189)

(Section V.C.5) indicate that catalytic reactions on these materials involving

peroxide are complex processes and other titanium oxo species may also be

involved.

VI.B. Titanium-Oxo Species and Activity

If the tetra- and tripodal Ti structures and the titanium oxo species derived from

these structures in the presence of ROOH (R ¼ H, alkyl) are involved as active

sites and reaction intermediates, the next step beyond their identification is to

seek correlations between the structure and concentrations of these titanium oxo

species and catalytic activity and selectivity. Clerici and Ingallina (204) were the

first to propose the Ti(O2H) group as the active site of alkene epoxidation by


H2O2 in TS-1. On the basis of the observed solvent and acid/base effects on the

kinetics and yield in alkene epoxidation in various alcohols, an end-on (1) group

with a simultaneously coordinated alcohol group was envisioned as the reactive

intermediate.

A direct correlation between the concentration of the titanium oxo species and

epoxidation activity was proposed by Lin and Frei (133). Loading TS-1/H2O2

with propene after evacuation, they observed by FTIR difference spectroscopy

the loss of the bands characterizing propene (at 1646 cm21) and TiOOH (at 837

and 3400 cm21). Figure 48 is the infrared difference spectrum recorded imme-

diately after loading the propene on TS-1/H2O2; Fig. 49 includes the spectra

recorded 80 and 320 min later.

The disappearance of the propene bands was not noticed when H2O2 (and

consequently TiOOH) was not present. After 80 min, the product spectrum

included bands at 830, 895, 1372, 1409, 1452, 1460 and 1493 cm21. The product

spectrum was similar to that obtained when a sample of propene oxide was loaded

onto TS-1. The rate of decay of the 837-cm21 absorption (O–O vibration of

TiOOH) was accompanied by the growth of the infrared bands of the product.

These observations led Lin and Frei to conclude that the TiOOH group was

Fig. 47. Catalytic selectivity as a function of Ti content in Ti-MCM-41 for 1-naphthol

hydroxylation with aqueous H2O2. H2O2 selectivity (mol%) ¼ (number of moles of H2O2 utilized in

product (1, 4-naphthoquinone, 1,4-dihydroxynaphthalene and 1,2-dihydroxynaphthalene) formation/-

number of moles of H2O2 fed) £ 100 [data from Chaudhari et al. (277)].


the active species in alkene epoxidation catalyzed by TS-1. When propene oxide

was brought in contact with a sample of TS-1 containing the TiOOH species,

propionaldehyde was formed by rearrangement. No such rearrangement of the

epoxide occurred (133) in the absence of the TiOOH, indicating that it is the

protonic acidity of TiOOH and not the Lewis acidity of the Ti ions in TS-1 that

is responsible for this acid-catalyzed rearrangement. Although dehydrated TS-1

does not contain Brønsted acid sites, such sites are apparently created during its

interaction with H2O2. The Lewis acid sites on TS-1 are probably deactivated by

the water present in the reaction medium.

Fig. 48. Infrared difference spectrum recorded immediately after loading of 6.5 mbar propene gas

into TS-1 molecular sieve containing TiOOH. Although the main peaks originate from adsorbed

C3H6, the small shoulders of the bands at 1443, 1646, 2980, and 3081 cm21 are attributed to gas-phase

propene [Reprinted from Lin and Frei (133) with permission. Copyright (2002) American Chemical

Society].

Fig. 49. FTIR difference spectrum recorded 80 min (trace a) and 320 min (trace b) after loading

of TS-1/TiOOH molecular sieve with 6.5 mbar of propene at room temperature [Reprinted from Lin

and Frei (133) with permission. Copyright (2002) American Chemical Society].


In an attempt to quantify the relationship between the TiOOH groups and

the yield of propene oxide from the extinction coefficients of the latter’s 1409-

and 1493-cm21 bands, it was determined that 0.6 mol of the epoxide formed per

mole of framework Ti center in the molecular sieve. That is, at least 60% of all

framework Ti (80% of the surface-exposed Ti) is converted to TiOOH upon

reaction with H2O2. The consumption of the TiOOH species during the oxygen

insertion into propene was also independently confirmed by the loss in intensity

of its LMCT band at 360 nm when the catalyst was brought in contact with

propene at room temperature (Fig. 50).

In contrast to propene, ethene, with its less electron-rich CyC bond, did not

react at room temperature in the dark with TS-1 and instead required excitation

of the UV–visible LMCT absorption at 360 nm to activate the TiOOH group for

electrophilic oxygen transfer to form the epoxide. Again, the formation of the

products, ethene oxide (at 871 cm21) and acetaldehyde (at 1353 and 1724 cm21)

was accompanied by the loss of the TiOOH peaks at 837 and 3400 cm21 and the

concurrent growth of the 3676- and 1629-cm21 bands assigned to Ti–OH and

H2O, respectively (133). Direct evidence for O transfer from TiOOH to ethene

was sought from the 18O isotope frequency shifts of ethene epoxide when a

Ti18O18OH moiety (generated from TS-1 and H218O2) was used. The epoxide

product, C2H418O, was isotopically pure, confirming that the oxygen atom in the

epoxide indeed originated from the TiOOH species.

Fig. 50. Diffuse reflectance spectra recorded (a) before and (b) after 20 min of thermal reaction of

propene in TS-1/TiOOH molecular sieve at room temperature [Reprinted from Lin and Frei (133) with



To probe the origin of acetaldehyde in ethene oxidation, ethene oxide was

admitted to the (TS-1/H2O2) system containing TiOOH groups. The formation

of acetaldehyde was negligible even under the influence of UV–visible irradia-

tion. Hence, the significant amount (10%) of acetaldehyde formed in the reaction

of ethene with TS-1/H2O2 could not have been the product of the further

reaction of ethene oxide. It is rather a primary product of oxidation at the

vinylic carbon atom.

Zhao et al. (50), on the basis of the appearance of the phenoxy radical

(detected by EPR spectroscopy) simultaneously with the disappearance of the

framework Ti-superoxide species resulting from contact of phenol with TS-1/

H2O2, correlated the concentration of the superoxide with catalytic activity for

phenol oxidation (Section III.E). Srinivas et al. (52) recently attempted to

correlate the relative EPR intensities of individual Ti-superoxides (A0, A, B, and

C) in the various titanosilicates with their chemoselectivities in styrene oxida-

tion (Sections II.A.7 and III.E). The relative concentration of A0 þ A was related

to styrene oxide (SO) selectivity (Fig. 51). Both the intensity of (A0 þ A)

Ti-superoxo signals and the selectivity for styrene oxide (SO) were higher in the

case of TS-1 than Ti-beta (Fig. 51). The yield of non-selective products (phenyl

acetaldehyde and benzaldehyde) correlates with the concentration of the (B þ C)

oxo species. Similarly, the concentration of the (B þ C) oxo species is higher

in methanol solvent than in acetonitrile, in parallel with the greater formation of

the non-selective products in the former than in the latter. It was also found that

the styrene epoxide concentration was higher when the total EPR signal

intensity was lower.

On the basis of these results, Srinivas et al. (52) suggested that EPR-inactive

hydroperoxo/peroxo titanium species are probably responsible for epoxidation,

although superoxo-titanium is responsible for the side reactions. The predomi-

nant formation of the epoxide at low temperatures and the non-selective products

observed when the temperature was raised were ascribed to the greater stability of

the hydroperoxo/peroxo-titanium species at lower temperatures and the relatively

high stability of the superoxo species at elevated temperatures. Additional

support for the greater involvement of the hydroperoxide in epoxidation comes

from investigations of the Pd-TS-1 system. The hydroperoxo/superoxo ratio

(0.73) observed when Pd-TS-1 is brought in contact with H2O2 was noted in

Section III.E (Fig. 25). Correspondingly, the selectivity for the epoxide in the

oxidation of propene catalyzed by Pd-TS-1 with H2O2 generated in situ from H2

and O2 was also high (99%) (Section V.C.16). The EPR signal intensity of the

titanium oxo species in Ti-MCM-41 was lower (52) when tert-butyl hydroper-

oxide in n-decane (rather than aqueous H2O2) was used as the oxidant, suggesting

that a majority of the oxo-titanium is in the EPR-silent hydroproxo/peroxo form

when reaction occurs in n-decane solvent.


A similar conclusion was also reached by Sankar et al. (46), who used EXAFS/

DFT techniques. From the selective decrease in the EPR intensity of the A type

superoxo species during the epoxidation of styrene and allyl alcohol (Fig. 52),

Srinivas et al. (52) concluded that these types of oxo species are preferentially

consumed during the reaction.

The correlation between the concentration of the superoxide species, A and B,

and catalytic activity is further illustrated in Tables LIII and LIV. A TS-1 sample

(without any trace of anatase) as well as another one containing some anatase

were prepared by the method of Thangaraj et al. (138) (with some minor

modifications). A sample of TS-1 (fluoride) was prepared in a fluoride medium.

Fig. 51. Correlation between the intensity of Ti-superoxo ([A0 þ A] and [B þ C]) signals and

selectivity for styrene oxide and non-selective products in the styrene epoxidation reaction. The effects

of titanosilicates, oxidants, and solvent on the correlation are depicted [from Srinivas et al. (52)].


The three TS-1 catalysts with similar Ti contents have cuboidal morphology with

comparable particle sizes of 0.2–0.3 mm (as shown in SEM pictures, Fig. 53).

The EPR spectra of the samples in contact with aqueous H2O2 (46%) (Fig. 54)

indicate that the ratio of the A to B superoxo species in various TS-1 samples

increases in the order TS-1 (fluoride) , TS-1 (with anatase) , TS-1 (without

anatase). Catalytic activity for phenol hydroxylation and allyl alcohol epoxi-

dation (Table LIII) was found to parallel the A/B ratio of the oxo-Ti species

(TS-1(fluoride) , TS-1 (with anatase) , TS-1 (without anatase)).

Catalytic activity in benzene hydroxylation (Table LIV), on the other hand,

followed the total concentration of the various superoxo species, which increased

in the order TS-1 (with anatase) , TS-1 (without anatase) , TS-1 (fluoride).

The total concentration of the superoxo species was obtained from the integrated

intensity of all the EPR signals representing superoxo species. This intensity in

various solvents increases in the order acetone , methanol p water.

The picture that emerges from the results summarized above is the following:

H2O2 reacts with the titanium centers on TS-1 and other titanosilicates to generate

the titanium oxo species (hydroperoxo and superoxo). At room temperature and

Fig. 52. EPR spectra recorded at 90 K. (a) TS-1 þ aqueous H2O2. (b)–(d) TS-1 þ H2O2 þ

styrene reacted at 333 K for 5, 10, and 20 min, respectively, and (e) TS-1 þ H2O2 þ allyl alcohol

reacted at 333 K for 25 min. Asterisk represents signal caused by a styrene-derived radical formed

during the reaction [from Shetti et al. (93)].


TABLE LIII

Catalytic activities of TS-1 samples (Si/Ti ¼ 33; particle size ¼ 0.2–0.3 mm) prepared by different methods

Catalyst Epoxidation of Allyl Alcohol (AA)a Phenol hydroxylationb,c

AA conversion

(mol%)

Product selectivity

(mol%)

Phenol conversion

(mol%)

Product selectivity

(mol%)

Glycidol -diol Catechol Hydroquinone

TS-1 (without anatase) 96.1 96.3 3.7 12.9 (16.2) 43.7 (24.9) 56.4 (75.2)

TS-1 (with anatase) 89.6 97.5 2.5 11.1 (17.3) 49.0 (23.6) 51.0 (76.4)

TS-1 (fluoride) 32.9 97.1 2.9 3.3 (13.6) 40.7 (22.4) 59.4 (77.6)

a Reaction conditions (epoxidation of AA): catalyst, 100 mg; AA, 8.6 mmol; H2O2 (aq. 46%), 17.2 mmol; acetone, 10 g; temperature, 333 K; time, 8 h.b Reaction conditions (phenol hydroxylation): catalyst, 100 mg; phenol, 10 mmol; H2O2 (aq. 32.8%), 3.33 mmol; solvent (acetone or methanol), 4.2 mL;

temperature, 333 K; H2O2 addition over 1.5 h; reaction time, 5.25 h (after H2O2 addition).c Values in parentheses correspond to the results in methanol solvent.

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

35

higher temperatures, there is an interconversion of the two types of oxo-species

(Section III.E). In alkene epoxidation the hydroperoxide reacts with the alkene to

give the epoxide (133). In view of the direct correlation observed between the

concentration of the (A þ A0) superoxo species and selectivity for the styrene

epoxide (Fig. 51), these two types of superoxides (A and A0, respectively) are

perhaps transformed more easily into the hydroperoxides than the others (B and

C, respectively). The side products probably arise from the reaction of either

or both of the B and C groups of superoxides. The more recent calculations and

in situ EPR results of Shetti et al. (54) suggest that the A and A0 superoxides are

attached to tetrapodal Ti, although the B and C species are coordinated to tripodal

titanium sites. The formation of TiOOH by both the tetra- and tripodal Ti is also

supported by the FTIR spectroscopic results of Lin and Frei (133).

TABLE LIV

Catalytic activity in benzene hydroxylation of TS-1 samples (Si/Ti ¼ 33;

particle size ¼ 0.2–0.3 mm) prepared by different methods

Catalyst Benzene conversion

(mol%)


Phenol Catechol Hydroquinone

TS-1 (without anatase) 33.8 64.1 16.0 19.9

TS-1 (with anatase) 23.4 72.4 12.9 14.8

TS-1 (fluoride) 36.1 63.0 16.3 20.8

Reaction conditions: catalyst, 100 mg; benzene, 19.2 mmol; H2O2 (aq. 32.8%), 9.6 mmol;

solvent (water), 7.5 g; temperature, 333 K; H2O2 addition in one lot; reaction time, 2 h.

Benzene conversion in methanol, acetone and acetonitrile solvents is negligible.

Fig. 53. SEM photographs of TS-1 samples, without anatase (left); with a trace amount of anatase

(center); and from a fluoride medium (right) [from Shetti et al. (93)].


VII. O–O Bond Cleavage and Product Selectivity

VII.A. General

It is known from the homogeneous catalytic oxidations by metal complexes

and biological oxidations by metalloenzymes that the type of cleavage of the

O–O bond in the active oxygenated metal species formed during the oxidation

reactions plays a crucial role in determining the product pattern. The breaking

of the O–O bond in Ti–O–O or the various other titanium oxo species dis-

cussed in Section III will also be determined by similar structural considerations

and influence product selectivities. Electron-donating or withdrawing ligands

either on the Ti atom (such as OSi, OH, or H2O) or the peroxo moiety (such as

the alkyl group in TBHP) can influence the scission of the O–O bond. In other

words, the type of Ti site (tetra-, tripodal, etc.) or oxidant (H2O2, TBHP)

influences the homolytic vs. heterolytic cleavage. The open structured, tripodal

titanium sites form penta- or hexa-coordinated species such as Ti(OSi)3

(H2O)2(OH) more easily than the closed tetrapodal Ti structures (vide supra).

The coordinated water and OH groups enhance electron density at Ti center and

the O–O bond, favoring homolytic O–O bond cleavage and zOH radical forma-

tion. Hence, systems having the tripodal Ti(OSi)3(OH) sites in preponderance

Fig. 54. EPR spectra showing the differences in the types of superoxo species generated on

various TS-1 samples prepared by different methods after contacting with aqueous H2O2 [from Shetti

et al. (93)].


(such as Ti-MCM-41 and possibly other Ti-mesoporous material) are likely to

cleave the O–O homolytically and generate greater amounts of radicals than

catalysts with predominantly tetrapodal Ti(OSi)4 sites (such as TS-1). Similarly,

alkyl hydroperoxides, when used as oxidants, are more likely to cleave the O–O

bond in TiOOR complexes homolytically than H2O2 in TiOOH. This may be

one of the reasons for the greater selectivities observed with TS-1 that uses

aqueous H2O2 as the oxidant than with Ti-MCM-41 that uses alkyl hydroper-

oxides as the oxidants.

VII.B. Epoxidation of Alkenes

In the epoxidation of alkenes, as was discussed above (Section V.C), TS-1

produces mainly the epoxide, although Ti-MCM-41 and similar mesoporous

materials produce, in addition, significant amounts of side products including

those derived from allylic CH activation. Adam et al. (278), exploring the factors

that influence the allylic CH oxidation vs. epoxidation in the oxidation of

2-cyclohexenol by Cr- and Mn-salen complexes in the liquid phase, found that

although manganese salens were selective for epoxidation, the chromium analo-

gues selectively gave allylic CH oxidation. Iodosobenzene was the oxygen

source. The authors interpreted the chemoselectivity in terms of the electron

transfer (for manganese salens) vs. the hydrogen abstraction mechanisms (for

the chromium salens) (Scheme 26).

When the reactant is cyclohexene, in the first step of Scheme 26, the direct

hydrogen abstraction for the allylic oxidation (path 1) competes with the electron

transfer (from the alkene to the M-oxo complex) for the epoxidation (path 2).

Because the manganese complex is more readily reduced than the chromium

Scheme 26.


complex, the authors speculated that the higher reduction potential of the

manganese complex (relative to the chromium complex) favors electron transfer

from the cyclohexene reactant to the metal catalyst and thus allows competetive

epoxide formation to take place. Conversely, for the more difficult to reduce,

electron-rich chromium complex, allylic oxidation by hydrogen abstraction

(path 1) is favored.

In the titanosilicate system, cyclic voltametric measurements had indicated

(Section III.D) that the electron density at the tripodal sites is higher than at the

tetrapodal sites. Hence, by analogy with the chromium and manganese complexes,

we may expect the tripodal sites to favor hydrogen abstraction and allylic CH

oxidation, although electron transfer and epoxidation occur preferentially on the

tetrapodal sites.

A tentative mechanism involving the heterolytic cleavage of the O–O bond

along with electron transfer from the alkene to the electrophilic oxygen of the

Ti(O2H) complex is shown in Scheme 27.

In the envisaged titanium oxo complex, the Ti atom is side-bound to the

peroxy moiety (O2H), consistent with all the spectroscopic results mentioned in

Section III; in Scheme 27, between the two O atoms that are side-bound to Ti4þ,

the O atom attached to both the Ti and H atoms is expected to be more electro-

philic than the O atom attached to only the Ti atom and is likely to be the site of

nucleophilic attack by the alkene double bond. The formation of the Ti–OH

group (and not the titanyl, TiyO, as proposed by Khouw et al. (221)) after the

epoxidation and its subsequent condensation with Si–OH to regenerate the

Ti–O–Si links had been observed (Section III.B) by FTIR spectroscopy by Lin

and Frei (133). Because this is a concerted heterolytic cleavage of the O–O

bond, high epoxide selectivity and retention of stereochemistry may be expected,

as indeed has been observed experimentally (204).

The transition state in the above scheme differs from the cyclic titanium peroxo

complex proposed earlier (217). In the earlier mechanism, any of the two peroxo

oxygens in the Ti–O–O–H (bound end-on) could have been inserted into the

CyC bond, and accordingly two isomers would be possible. They have never

Scheme 27.


been observed (33). In Scheme 27, on the other hand, the oxygen attached to both

Ti4þ and the proton will be relatively more electrophilic to accept the electron

from the CyC bond. Our mechanism bears similarities to those proposed in the

homogeneous catalysis literature (170) for reactions catalyzed by peroxyacids,

RC(O)OOH. The Ti4þ replaces, formally, the acylium cation, RCOþ. When

instead of H2O2 an alkyl hydroperoxide (such as tert-butyl hydroperoxide) is

used, the titanium oxo species that is generated may be Ti(O2R) (R ¼ alkyl). As

a consequence of the electron-donating effect of R, it is unlikely that the

oxygen atom attached to it acquires an electrophilic character. Hence, it is the

other oxygen atom attached to the metal that is more electrophilic and is,

therefore, attached to the CyC bond forming epoxide, as shown in Scheme 5.

We emphasize that the above mechanism is strictly valid only for H2O2 and

alkyl hydroperoxide epoxidations of alkenes catalyzed by TS-1 and Ti-MCM-

41. In view of the observation of similar titanium oxo species when H2 þ O2

are brought in contact with TS-1 or Ti-MCM-41 (54), similar conclusions may

be drawn for that system as well. A radical mechanism involving the TiyO

groups had been proposed earlier by Khouw et al. (221) for the hydroxylation

of alkanes. No spectroscopic investigation of the TS-1/H2O2/alkane has yet

been reported.

VIII. Conclusions and Outlook

Significant progress has been achieved in the preceding few years in the study

of titanosilicate molecular sieves, especially TS-1, TS-2, Ti-beta, and Ti-

MCM-41. In the dehydrated, pristine state most of the Ti4þ ions on the surfaces

of these materials are tetrahedrally coordinated, being present in either one of

two structures: a tetrapodal (Ti(OSi)4) or a tripodal (Ti(OSi)3OH) structure.

The former predominates in TS-1, TS-2, and Ti-beta, and the latter is prominent

in Ti-MCM-41. The Ti ions are coordinatively unsaturated and act as Lewis

acid sites that coordinatively bind molecules such as H2O, NH3, CH3CN, and

H2O2. Upon interaction with H2O2 or H2 þ O2, the Ti ions form titanium oxo

species. Spectroscopic techniques have been used to identify side-bound

hydroperoxo species such as Ti(O2H) and superoxo structures such as Ti(O2z2)

on these catalysts.

These titanium oxo species oxidize various organic reactants. Direct con-

firmations of the participation of these titanium oxo species in the oxidation

reactions have been obtained by infrared and EPR spectroscopies (54,133). The

infrared absorption (133) or EPR (54) signal intensity of the titanium oxo species

decreased simultaneously with an increase in the infrared or EPR signal

intensities characterizing reaction products.


Although TS-1 in its dehydrated state is not a Brønsted acid, the hydroperoxo

species Ti(O2H) generated as a result of its interaction with H2O2 has Brønsted

acidity and catalyzes reactions such as the isomerization of epoxides to aldehydes

(for example, propene oxide to propionaldehyde). Hence, although oxidation by

H2O2 is the predominant reaction catalyzed by these materials, side reactions

attributed to the Brønsted acidity of the Ti(O2H) group can also occur, decreasing

the selectivity for the desired oxidation product. In the absence of H2O2, these

titanium silicates are weak Lewis acids and catalyze reactions such as the

rearrangement of cyclohexanone oxime to 1-caprolactam or the cycloaddition of

CO2 to epoxides to yield cyclic carbonates.

A large number of oxidation reactions of a variety of reactants have been

reported to be catalyzed by titanosilicate molecular sieves (Section V). The

transition from the laboratory to the factory will undoubtedly happen in some of

the cases. Because of the high price of H2O2, most of the novel applications are

likely to be in the area of fine chemicals rather than commodity or bulk materials.

Attempts have already been made to find substitutes for H2O2 or to generate H2O2

in situ from H2 þ O2 or alcohol þ O2. Metals such as platinum, palladium, gold,

etc. supported on TS-1 have been explored as catalysts. The strategy was to

synthesize the H2O2 on the metal and use it in turn to catalyze the oxidation

reaction on the titanosilicate. The main difficulty has been the efficient synthesis

of H2O2; only low H2 and O2 efficiencies have been encountered in the synthesis

of H2O2, rendering the process economically unviable. An alternate approach is

to generate H2O2 in situ from the oxidation of alcohols (such as isopropanol or

anthraquinol) with O2:

Alcohol þ O2 ! ketone þ H2O2: ð34Þ

The ketone can be hydrogenated in a separate reactor and recycled. This is the

current route for the manufacture of H2O2 using anthraquinone–anthraquinol.

The technological and economic advantages of combining the two processes

(H2O2 synthesis and oxidation of organic reactants) in one reaction zone are not

clear. To overcome the limitations of the MFI pore structure of TS-1 in oxidizing

large molecules, Ti-beta, Ti-MCM-41, and other large and mesoporous materials

have been investigated. The results have been mixed. Although the rates of

the oxidation reaction have been enhanced (by the absence of diffusional con-

straints), attaining high selectivity for the desired oxidation product has been

more elusive. Identifying, designing, and synthesizing the appropriate titanium

oxo species on the surface of large-pore or mesoporous Ti-silicates while

simultaneously increasing their hydrophobicity will be necessary to obtain the

high selectivity characteristic of TS-1. There will be an increasing focus on the

standardization of the synthesis procedures of these novel materials and charac-

terizing modifying their physicochemical and catalytic properties in the coming


years. Appendix C includes a list of some of the recent advances and publications

regarding the synthesis of titanium silicate molecular sieves.

Acknowledgements

PR thanks the Alexander von Humboldt Foundation for a visiting Fellowship to

Munich.

Appendix A. Fingerprint Features for Ti Isomorphous

Substitution in TS-1 Titanosilicates

See Table A1.

TABLE A1

Characterization technique Fingerprint feature

XRD MFI structure; orthorhombic (Pnma space group at room

temperature) to monoclinic (P21=n space group at low

temperatures) structural phase transition

UV (diffuse reflectance) Intense band at 210–220 nm (O(2p) ! Ti(3d) charge transfer

transition)

XAS Intense Ti pre-edge peak (1s ! 3d) at about 4969 eV

EPR No signal (diamagnetic þ 4 oxidation state of Ti); contact with

CO or H2 (at elevated temperatures (773 K)) generates

paramagnetic Ti3þ species

UV resonant Raman Strong bands at 490, 530 and 1125 cm21 (due to bending,

symmetric stretching and asymmetric stretching vibrations of

Ti–O–Si, respectively) when excited at 244 nm

UV photoluminescence Emission bands at 495 and 430 nm with the corresponding

excitation bands at 250 and 300 nm, respectively

XPS Ti2p core level spectrum at 460.0 ^ 0.2 eV (due to þ 4

oxidation state of tetrahedral Ti; higher energy shift in binding

energy by ,1.5 eV compared to TiO2 anatase) (caution: highly

dispersed Ti in silica matrices (Ti . 2%) can produce a similar

high energy shift; this shift is also claimed to depend on the

large number of Si atoms in the second coordination shell of Ti)

Infrared and Raman Band at 960 cm21 assigned to Ti–O–Si vibration (Caution:

Si–OH and defect sites in silicalites also show this feature).


Appendix B. Characteristics of the Oxo-Titanium Species

Generated on TS-1 on Contact with Aqueous H2O2

See Table B1.

Appendix C. Synthesis of Titanium Silicate Molecular Sieves

The review of Notari (33) covers the synthesis methodologies of titanium silicate

molecular sieves available up to 1996. The reviews of Corma (279) and

subsequently of Biz and Occelli (280) describe the synthesis of mesoporous

molecular sieves. An informative article on the preparation of TS-1 was reported

recently by Perego et al. (68). In this section we list some of the recent develop-

ments in the synthesis of micro and mesoporous titanosilicate molecular sieves.

TABLE B1

Technique Characteristic feature

Visual appearance

(color)

Yellow

Diffuse reflectance

UV–visible

A labile charge transfer band at about 385 nm (25,800 cm21) in

neutral H2O2 solutions and a relatively more stable band at

350 nm (28,500 cm21) in alkaline H2O2 solutions

Vibrational spectroscopy

(infrared and

Raman/resonance Raman)

Reduction and blue shift of characteristic Si–O–Ti band (at

960 cm21) to 976 cm21 and quenching of 1125 cm21 band in

resonance Raman spectrum when excited with 442 and 1064 nm

laser radiation

Strong, complex feature at 618 cm21 in resonance Raman

spectrum when excited with 442 nm radiation

Infrared-weak and Raman-intense absorption at about

880–890 cm21 in neutral H2O2 and at about 840 cm21 in

alkaline H2O2 solutions

Large bandwidth, red-shifted infrared band corresponding to

hydrogen bonded OH groups at 3400 cm21.

XAS Significant reduction in the pre-edge intensity indicating

increase in the coordination number of Ti

EPR Labile, rhombic type spectrum corresponding to Ti-superoxo

species; spectral features sensitive to the type of silicate

structure, temperature, solvent and pH

Magnetism Partly paramagnetic.


C.1. TS-1, TS-2, Ti-ZSM-48, Ti-MWW, and Ti-MMM-1

Taramasso et al. (5) had originally reported two methods for the

hydrothermal synthesis of TS-1. The first method (mixed alkoxide method)

involves the preparation of a solution of mixed alkoxides of titanium and

silica (preferably ethoxides) followed by hydrolysis with alkali-free solution

of tetrapropylammonium hydroxide (TPAOH), distillation of the alcohol and

crystallization of the resulting gel at 448 K. In the second method (dissolved

or hydrolyzed titanium method) a soluble tetrapropylammonium peroxo-

titanate species was prepared initially and then colloidal SiO2 (Ludox AS-40)

was added. This entire operation had to be carried out at 278 K. The TS-1

samples obtained by these two synthesis routes differed, particularly because

of the presence of impurities such as Al3þ usually present in colloidal

silica (33).

Later, Thangaraj et al. (275,281) developed a novel, improved route ( pre-

hydrolysis method) for the preparation of good quality TS-1 samples. In this

method the silica source (tetraethyl orthosilicate; TEOS) in iso-propanol was first

hydrolyzed with 20% aqueous TPAOH solution prior to the (dropwise) addition

of titanium butoxide in dry iso-propanol under vigorous stirring. Crystallization

was done statically at 443 K for 1–5 days and the solid was calcined at 823 K

for 10 h. The TS-1 samples thus obtained exhibited high catalytic activity in

hydroxylation reactions.

Another method (known as the wetness impregnation method) originally

reported by Padovan et al. (282,283) used a SiO2–TiO2 coprecipitated dry gel

which was impregnated with an aqueous solution of TPAOH and crystallized

under autogeneous pressure. At a high concentration of the base, dissolution of

the oxides occurs, followed by crystallization in the presence of TPAOH. This

method offers the advantage of requiring relatively small amount of TPAOH. But

the catalyst obtained was poorly active as a consequence of the impurities present

in the starting material.

In an attempt to produce TS-1 at low cost, alternative, cheaper sources of Ti

and Si and other bases such as binary mixtures of (tetrabutylammonium and

tetraethylammonium hydroxides), (tetrabutylphosphonium and tetraethylpho-

sphonium hydroxides), (tetrapropylammonium bromide and ammonia, water,

hexanediamine, n-butylamine, diethylamine, ethylenediamine, or triethanola-

mine) in place of TPAOH have been used (284–294). TS-1 was synthesized in

the presence of fluoride ions but the material thus formed contained extraframe-

work Ti species (295–297).

Kumar et al. (298–300) reported a method wherein the crystallization time is

significantly reduced. They found that addition of a small amount of oxyanion

(e.g., H3PO4) to the TS-1 synthesis gel enhances the nucleation and crystallization


rates. By this promoter-induced synthesis method the overall crystallization time

was reduced by about five times.

Ahn et al. (301) and subsequently Prasad et al. (302,303) reported the rapid

synthesis of highly crystalline TS-1 by microwave irradiation technique with

yields exceeding 90%. The synthesis, which requires 1–2 days by the conven-

tional heating methods of Taramasso et al. (5) and Thangaraj et al. (275,281),

was achieved within 30 min. In the synthesis reported by Ahn et al. (301), a

SiO2–TiO2 cogel ðSi=Ti ¼ 50Þ prepared by a two-step acid/base sol–gel process

was dried overnight at 383 K and subsequently ground to give a fine powder

which was dry impregnated by adding TPAOH solution. The impregnated gel

was then heated with microwaves (500 W; 443 K) to obtain the crystalline

powder. Prasad et al. (303) prepared the gel ðSi=Ti ¼ 10Þ following the pre-

hydrolysis synthesis method and then heated by microwaves (800 W; 448 K).

Approximately 12–14 bar autogeneous pressure was developed during the

synthesis. The catalysts prepared by the microwave technique showed activity

similar to those prepared by the conventional heating methods.

In an attempt to reduce the amount of expensive TPAOH template, Khomane

et al. (304) used a non-ionic surfactant, Tween 20, in the TS-1 synthesis. Their

method required only a small amount of TPAOH. Highly crystalline TS-1 samples

(0.15 mm size) showing good activity for octane epoxidation were obtained.

Similar procedures adopted for the synthesis of TS-1 (the mixed alkoxide

method, dissolved titanium method, pre-hydrolysis method, wetness impreg-

nation method, and promoter induced synthesis method) were also used for the

synthesis of TS-2. Tetrabutylammonium hydroxide (TBAOH) instead of TPAOH

was used as the template (6,7,305–308).

Ti-ZSM-48 was prepared by the dissolved titanium method using fumed silica

(Cabosil), TBOT, H2O2, and diaminooctane (309–310). Ti-ZSM-48 was also

prepared using hexamethonium hydroxide base and by the pre-hydrolysis

method (311).

A titanosilicate with MWW structure (Ti-MWW) reported by Wu and Tatsumi

(228) was claimed to be more active than TS-1 in the epoxidation of linear

alkanes. Ti-MWW was synthesized in two steps. The first step consists of hydro-

thermal synthesis of Ti-containing MWW lamellar precursors using piperidine

as a structure-directing agent and boric acid as a crystallization support agent.

The second step was to treat the precursors in HNO3 or H2SO4 solutions under

reflux for removing the extraframework titanium species together with a part of

the framework boron.

The diffusional properties of TS-1 catalysts could be modified by the synthesis

of nanosized TS-1 (by the recently developed confined space synthesis method),

but the separation of the finely crystalline catalyst from the product mixture is

difficult. The procedure of Jacobsen and co-workers (188) for the synthesis of a

mesoporous TS-1 overcomes this problem. In a typical synthesis of mesopous


TS-1 (mesoporosity ,20 nm, 0.3–1.2 mm size), carbon black pearls 700w

(Carbot Corp., average particle diameter ¼ 18 nm (ASTM D-3249)) were

impregnated by the incipient wetness method with a clear solution of TPAOH,

water, and ethanol. After evaporation of ethanol, the carbon particles were

impregnated with 20% excess (relative to incipient wetness) of a mixture of

TEOT and TEOS. Aging for a minimum of 3 h at room temperature and heating

at 453 K for 72 h yielded the solid product, which was isolated, and the carbon

black was removed by controlled combustion in air at 523 K for 8 h.

A similar development in this direction is the synthesis of a mixed-phase

material containing both micro- and mesopores (Ti-MMM-1) (223). This

material was synthesized by the addition of organic templates for mesopores

(cetyltrimethylammonium bromide, CTABr) and micropores (tetrapropylammo-

nium bromide, TPABr) at staggered times and the variation of the temperature of

a single reaction mixture. Ti-MMM-1 is more selective (for oxidation of

cyclohexane and of n-octane) than either Ti-MCM-41 or TS-1. The powder X-ray

diffraction pattern indicates that the material contains both MCM-41 and MFI

structures. The mixed phase contains framework Ti species and more atomic

order within its walls than Ti-doped MCM-41.

C.2. Ti-Beta Zeolite

Large-pore Ti-beta (pore diameter ,0.4–1 nm) was synthesized by direct

hydrothermal synthesis, wetness impregnation, and by secondary synthesis

methods (9,10,12,14,196,312–318). It was thought initially that cations such

as Al3þ are essential for the crystallization of beta-zeolite. Most of the early

methods gave low zeolite yields, together with inefficient use of the expensive

structure-directing agent (tetraethylammonium cation). Futhermore, the intrinsic

activity of these materials was lower than that of TS-1 for small reactant

molecules. The lower activity was found to be caused by Al3þ ions, a high density

of connectivity defects (resulting in extreme hydrophilic properties), and a higher

acidity of framework Ti species. Although Al-free Ti-beta zeolite could be

synthesized by the use of dealuminated zeolite-beta seeds at high pH, the product

(Ti-beta(OH)) contained a high density of Si–OH groups with a hydrophilic

surface (12,13).

Blasco et al. (12,13) developed a novel method for the synthesis of Al-free Ti-

beta zeolite in a fluoride medium. The Ti-beta zeolite thus obtained (Ti-beta(F))

was free of connectivity defects and was hydrophobic. The typical unseeded

synthesis of Al-free Ti-beta zeolite (Ti-beta(F)) involves hydrolysis of TEOS in

aqueous solutions of TEAOH (35%) and H2O2, followed by hydrolysis of TEOT

and evaporation of ethanol and water. The water lost in the evaporation and


an appropriate amount of HF (48%) are then added and the reaction mixture

crystallized while tumbling the autoclaves (60 rpm) at 413 K.

C.3. Ti-Containing HMS, MCM-41, and MCM-48

Tanev et al. (19) prepared titanium-substituted hexagonal mesoporous silica (Ti-

HMS) by adding Ti(iso-OC3H7)4 and Si(OC2H5)4 dissolved in a mixture of

ethanol–isopropanol to an aqueous solution of dodecylamine (DDA) and HCl.

Aging of the resulting gel for 18 h at ambient temperatures afforded the

crystalline as-synthesized Ti-HMS sample, which was then calcined in air at

923 K for 4 h. Ti-MCM-41 was prepared in a similar manner except for using

quaternary ammonium ion template [C16H33N(CH3)3]þ (CTMAþ) (with

counterion Br2) as a replacement of DDA (19). Corma et al. (17) reported the

preparation of Ti-MCM-41 by use of amorphous silica (Aerosil 200 Degussa), an

aqueous solution of tetramethylammonium hydroxide (25% TMAOH, K þ

Na , 5 ppm,), an aqueous solution of hexadecyltrimethylammonium bromide

(CTABr), and titanium isopropoxide at 408 K under static conditions (14 h).

Maschmeyer et al. (319) prepared Ti-containing MCM-41 by grafting titanocene

to the surface of silica walls (Ti " MCM-41). In contrast to the situation in Ti-

MCM-41, the Ti ions in Ti " MCM-41 are at the surface, mostly having the

tripodal tetrahedral structure. In Ti-MCM-41, part of the Ti is substituted in the

silica lattice and resides within the walls. In an improved procedure, Corma et al.

(320) reported that the structural order of MCM-41 is superior when Si(OCH3)4

is used as the silica source in place of Si(OC2H5)4. Ti-MCM-41 prepared by

the above methods exhibited a lower efficiency in the utilization of H2O2 (for

formation of the epoxide) in alkene oxidation than either TS-1 or Ti-beta. The

hydrophilic/hydrophobic properties of Ti zeolites influence their catalytic activity

and selectivity. The activity of Ti-MCM-41 catalysts was enhanced by silylation

of the surface (273,321,322).

Ti-MCM-48 (surface area ¼ 1000–1450 m2/g, pore volume ¼ 0.8–1.1 cm3/g,

pore diameter ¼ 2.4–2.7 nm) was synthesized by hydrothermal and postsyn-

thetic grafting techniques from cationic alkylammonium surfactants (22,25,323).

C.4. Ti-SBA-15

Morey et al. (25) synthesized Ti-SBA-15 with uniform tubular channels

(surface area ¼ 600–900 m2/g, pore volume ¼ 0.6–1.3 cm3/g, average pore

diameter ¼ 6.9 nm) by direct and postsynthesis methods by using triblock

copolymers, poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) in


TABLE C1

Synthesis of titanosilicate molecular sieves

Titanosilicate Synthesis methodology, composition and

improvements

Si/Ti Crystallite size (nm)/

morphology

References

TS-1 (MFI) Mixed alkoxide method. Hydrothermal synthesis

using tetraethylorthosilicate (TEOS) as the source

of Si, tetraethyltitanate (TEOT) as the source of Ti,

tetrapropylammonium hydroxide (TPAOH) as

structure directing agent (template), base and

distilled water

90–30 Parallelepipeds with

rounded edges

(5)

Dissolved titanium method. Hydrothermal syn-

thesis using tetrapropylammonium peroxytitanate

(prepared from TEOT, distilled water, 30%

aqueous H2O2, and 25% aqueous TPAOH) as the

source of Ti and colloidal silica (Ludox AS-40) as

the source of Si and TPAOH as template. All

additions done at 278 K

90–30 Parallelepipeds with

rounded edges

(5)

Preparation using TiCl2, 14% aqueous TPAOH,

30% colloidal silica, and demineralized water

Microspheres of

diameter 5–1000 mm

(284)

Preparation at low pH using fluoride ions as

mineralizing agent

(295,296)

Wetness impregnation method (282,283)

Prehydrolysis method. The Si source (TEOS) in

dry iso-propyl alcohol is hydrolyzed with 20%

aqueous TPAOH prior to addition of Ti source,

Ti(OBu)4. Gel composition: SiO2:xTiO2:0.36-

TPA:35H2O ðx ¼ 0–0:10Þ; the synthesis time is

reduced considerably (1–5 days at 433 K com-

pared to 6–30 days at 448 K, as reported in the

original patent (5))

$10 Cuboid (,1 mm) (275,281)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

48

Prehydrolysis method. Synthesis using binary

mixtures of tetrabutylammonium and tetraethy-

lammonium hydroxides instead of TPAOH

(285)

Influence of TPAOH/TEAOH and TPAOH/NH4-

OH ratio on the rate of crystallization and

crystallite size investigated

(286)

Prehydrolysis method. Synthesis using binary

mixtures of tetrabutylphosphonium hydroxide and

tetraethylphosphonium hydroxide instead of

TPAOH as base and template; TEOS and TBOT

are sources of Si and Ti, respectively. Molar gel

composition, SiO2:xTiO2:0.4 (x0TEPOH þ (1 2

x0)TBPOH):30H2O ðx ¼ 0–0:02Þ;

temperature ¼ 443 K and synthesis time ¼ 4 days

Ovate shaped crystals

(when x0 ¼ 0); hexago-

nal prisms (when

x0 ¼ 0:25–0:5) (2–

3 mm)

(287)

Influence of nature of silicon and titanium

alkoxides on the incorporation of Ti

(288)

Wetness impregnation method (325,326)

Prehydrolysis method. Synthesis under stirring

(250 rpm; 453 K, 5 days) using TPABr and

hexanediamine instead of TPAOH and other alkali

media, TEOS and TBOT are sources of Si and Ti.

Gel composition: SiO2:0.01TiO2:0.3C6-

DN:0.1TPABr:50H2O

24–76 Elongated prisms

,7 £ 2.5 £ 0.5 mm

(289)

Prehydrolysis method. Synthesis using SiO2

instead of silica alkoxides. Gel composition:

SiO2:xTiO2:0.4TPAOH:35H2O; 0 , x , 0:03

50–86 Hexagonal prisms/t-

winned conffin shaped

particles (10–19 mm)

(290)

(Continued)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

49

TABLE C1

Continued


improvements


morphology

References

Prehydrolysis method. Investigation of influence

of added oxyanions such as phosphate, perchlorate,

arsenate, chlorate, bromate, etc. on rate of

crystallization. The overall crystallization time in

the presence of additives reduced by about five

times compared to the conventional prehydrolysis

method (7,8)

30–80 0.1–0.2 mm (298–300)

Synthesis using TPABr as structure-directing

agent and ammonia, water, hexanediamine, n-

butylamine, diethylamine, ethylenediamine, or

triethanolamine as base (seeds of TS-1 were added

to get smaller crystallites and 100% crystallinity)

(327)

Synthesis of “fibrous” titanosilicate 2.5 mm length and

aspect ratio

(length/diameter) ¼ 50–

70

(328)

Synthesis using TiF4 (as the source of Ti), TEOS,

TPAOH, and distilled water. Gel composition:

SiO2:xTiO2:0.4TPA:30H2O, 0 , x , 0:05

45–90 Round shaped particles

(0.3 mm diameter)

(291,292)

Prehydrolysis method. Crystallization without

evaporating the alcohol in the conventional

synthesis (7,8)

(293)

Preparation by gas–solid isomorphous substitution

of Ti4þ for Si4þ and hydrothermal crystallization

using TPABr as template

(329)

Preparation using TiCl3 as source of Ti: influence

of pH (11.6–9.7)

Crystallite size 0.1–

4 mm

(330)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

50

Mixed alkoxide method. Synthesis using TPAOH

and HF and wetness-impregnation method using

TPABr and NH4F

(297)

Mixed alkoxide method. Preparation using ethyl-

silicate-40 (ES-40) as the cheaper, cost-effective

source of Si. Gel composition: SiO2:0.03-

TiO2:0.33TPA:35H2O

33 0.1–0.2 mm (294)

Template-impregnated SiO2–TiO2 xerogels.

SiO2–TiO2 cogel prepared via a two-step acid/-

base sol–gel process. Gel obtained dried overnight

383 K, ground to fine powder and dry impregnated

by adding 1.6 g of TPAOH (20% aq. solution) per

1 g of xerogel and heated in microwave environ-

ment. Crystalline product dried at 383 K and

calcined at 823 K for 5 h (crystal yield .90%)

50 Round shaped particles

,0.5 mm

(301)

Prehydrolysis method. Synthesis under microwave

irradiation; gel composition: SiO2:xTiO2:0.36-

TPAOH:35H2O, x ¼ 0:03–0:11; reaction

temperature ¼ 448 K, power input ¼ 800 W, 12–

14 bar autogeneous pressure, crystallization

time ¼ 20–90 min

10–33 0.3–1.2 mm (302,303)

Prehydrolysis method. Synthesis using small

amount of TPAOH template in the presence of

Tween 20, a non-ionic surfactant. Gel compo-

sition: 0.03TiO2:SiO2:0.12TPAOH:0.0009Tween

20:0.88IPA:14.45H2O. Crystallized at 433 K for

18 h under autogeneous pressure

33 0.15 mm (304)

TS-2 (MEL) Mixed alkoxide method using TBAOH as structure

directing agent

(6)

(Continued)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

51

TABLE C1

Continued


improvements


morphology

References

Prehydrolysis method using tetraethylorthosilicate,

titanium tetrabutoxide, and tetrabutylammonium

hydroxide. Gel composition: SiO2:xTiO2:0.2-

TBAOH:20H2O, x ¼ 0:14–0:0055; 443 K, 2–7

days

(7,305)

Synthesis using TBPOH as templating agent. Only

a maximum of 1.1 Ti/unit cell can be incorporated

in the framework

(286)

Wetness-impregnated SiO2–TiO2 xerogels Elliptical particles

(,1 mm)

(307)

Synthesis based on hydrolyzed titanium alkoxides

with H2O2. Gel composition: SiO2:xTiO2:0.88-

TBAOH:99H2O:25x H2O2. Crystallization at

449 K

25 Ovate type crystals

(2 mm)

(8,308)

Ti-ZSM-48 Prehydrolysis method. TEOS, TBOT, hexametho-

nium hydroxide template; 473 K, 7 days, crystal-

lization by rotation (40 rpm)

36–60 0.2–0.3 mm Spherical

random agglomerates of

small needle shaped

crystals 5–15 mm diam-

eter containing needles

of 0.2–1 mm long and

diameter 0.1 mm

(309,310)

Hydrolyzed titanium oxide method using fumed

silica as Si source

24–111

Synthesis using hexamethonium hydroxide 49 (311)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

52

Ti-Beta (BEA;) [Ti–Al]-beta (Si/Al # 150): Prehydrolysis

method–conventional method using amorphous

silica (Arosol 200), tetraethyl titanate, sodium

aluminate/aluminium nitrate as sources of Si, Ti,

and Al, respectively. Crystallization at 408 K by

rotation (60 rpm); zeolite yield #7%.

(9,10,110)

Cogel method by impregnating TiO2–SiO2 cogel

with TEAOH solution in the presence of some

amount of aluminium ions. Crystallization at

408 K while tumbling the autoclave (60 rpm).

Zeolite yields ,29%; Si/Al ¼ 300. Requires

lesser amount of TEAþ ions than classical

prehydrolysis method

(312,313)

Seeding technique. Al-free Ti-beta obtained by use

of dealuminated zeolite-beta seeds

(12)

Fluoride method. Al-free Ti-beta: synthesis from a

reaction mixture containing TEAOH and fluoride

ions (HF) at near-neutral pH. Gel composition:

TiO2: 60SiO2:32.9NEt4OH:32.9HF:20H2O:457.5

H2O. Crystallization at 413 K with rotation of the

autoclave (60 rpm)

50 (13,314)

Al-free Ti-beta: Direct synthesis (196,315,316)

(Continued)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

53

TABLE C1

Continued


improvements


morphology

References

Dry gel conversion method. 0.58 g of TBOT

suspended in distilled water (4.0 g) to which was

added 2 g of H2O2 (31 wt%). Mixture was stirred

for 1 h, leading to solution A. Solution B prepared

by dissolving anhydrous NaAlO2 (0.0124 g) and

0.015 g of NaOH in 8 g of TEAOH (40 wt% in

water) and stirred for 1 h. Solution B added to

solution A, stirred during heating at 353 K to

dryness. Dried powder with composition SiO2:

TiO2:Al2O3:Na2-

O:TEAOH ¼ 304:10:0.46:1.55:132.5) transferred

to an autoclave where water as a source of steam

was pored into the bottom. Crystallization carried

out in steam first at 403 K (96 h) and then at 448 K

(18 h) under autogeneous pressure. The recovered

product was washed, dried (308 K, 10 h), and

calcined (793 K, 10 h). The resulting Ti-beta was

treated with 1-M H2SO4 at room temperature

(12 h), washed, dried, and again calcined at 793 K

for 5 h in the flowing air. By using colloidal silica

(ST-40, 40 wt% SiO2, Nissan) instead of fumed

silica, Ti-beta with higher crystallinity was

synthesized. The molar composition of the gel was

SiO2:TiO2:Al2O3:Na2O:

TEAOH ¼ 310:10:0.52:12:135

,30 (318)

ETS-10/-4

(Zorite structure)

Synthesis with TiCl3 and without any organic

template

(331–336)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

54

Influence of various organic bases (R ¼

pyrrolidine, tetramethylammonium chloride, tet-

raethylammonium chloride, tetrapropylammonium

chloride, 1,2-diaminoethane, and 1,2-diaminohex-

ane) on crystallization of ETS-10. Synthesis using

Na2SiO3·n H2O, TiCl4, NaOH, KOH, and distilled

water. Gel composition: 40R:52Na2O:42K2-

O:20TiO2:100SiO2:7030H2O. pH ¼ 10.5–12.9.

Crystallization at 473 K for 2–30 days

(337,338)

ETS-10: Hydrothermal synthesis using TiO2 (P25,

200 A particle size, Degussa), 40% colloidal silica,

KF, and NaOH and crystallization at 473 K for 2

days. Gel composition: 1.0 M2O:TiO2: 2–

8SiO2:5–50H2O. ETS-4: hydrothermal synthesis

using NaF instead of KF (used in ETS-10) and

crystallization at 473 K for 44 h. Gel composition:

1.0 M2O:TiO2:1.2–6SiO2:5–50H2O

0.6–1 mm (339)

Synthesis of ETS-10, both in the presence and in

the absence of seeds of ETS-4 and using TiCl4 as

the source of Ti

(340)

ETS-10 synthesis using organic templating agents

(R) such as choline chloride and bromide salt of

hexaethyl diquat-5, sodium silicate, TiCl3 (15%

solution in HCl), NaOH, and KF·2H2O. Gel

composition: 1.14R2O:3.7Na2O:0.95K2O:

TiO2:5.71SiO2:171.9 or 256.9 H2O. Crystallization

at 473 K for 5–7 days

Cuboid or wheat-shaped

agglomerated crystals of

2–4 mm

(341)

(Continued)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

55

TABLE C1

Continued


improvements


morphology

References

Synthesis of ETS-10 using TiCl3 and crystalline

TiO2 (anatase) as Ti sources. Gel composition:

4Na2O:1.5K2O:TiO2:5.5SiO2:125H2O. Crystalli-

zation at 503 K for 24 h

,25 mm crystals (342)

ETS-10 synthesis from gels containing TiF4 and

TiO2

(343)

Ti-MCM-41 Synthesis mixtures prepared using amorphous

silica (Aerosil 200, Degussa), 25% aq. TMAOH,

aqueous solution of hydroxide and bromide of

hexadeciltrimethylammonium. Source of Ti was

TEOT. Gels with following molar compositions

were prepared: Si/Ti ¼ 60, (CTMA)2O:TMA2-

O ¼ 0.67, (TMA)2O:SiO2 ¼ 0.13,

H2O:(TMA)2O ¼ 188

60 Pore size ¼ 2 nm; sur-

face area ¼

936 m2/g

(17)

Silylation of surface of Ti-MCM-41. Synthesis gel

composition: SiO2:0.015 TEOT:0.26

CTABr:0.26TMAOH:24.3 H2O

66 (321)

Trimethylsilylation: Ti-MCM-41 prepared from

TEOS, TBOT, and CTMACl with molar gel

composition SiO2:0.01TiO2:0.6CTMA:0.3NMe4-

OH:60H2O was silylated with Me3SiCl and

(Me3Si)2O

139 (123 before

silylation)

Pore diameter ¼ 1.9 nm

(2.32 nm before silyal-

tion); pore

volume ¼ 0.82 mL/g

(0.88 mL/g before sily-

lation, surface

area ¼ 139 m2/g

(123 m2/g before silyla-

tion)

(273)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

56

One-step synthesis with methylated silicons:

synthesis of organo-silica containing Ti-MCM-41

carried out with gels having following molar

compositions: (1 2 x)Si(OCH3)4:xCH3Si(OC2-

H5)3:0.26TMAOH:0.15CTABr:24.3 H2O:yTEOT,

where x ¼ 0:15–0:35 and y ¼ 0:0166–0:0075:

After crystallization, the solid was first treated with

0.05-M H2SO4 in ethanol and then with 0.15-M

HNO3 in heptane-ethanol

60–133 (320)

Ti-HMS Synthesis by acid hydrolysis in alcohol solution of

mixture of TEOS and Ti(iso-OC3H7)4 in dodecy-

lamine

100 Pore diameter ,2.8 nm (19)

20–160 (20)

Synthesis using Gemini surfactant (bromide salt of

[C18H37(CH3)2N–C12H24–N(CH3)2C18H37]2þ

14.3 and 33.3 (21)

50 and 100 (22)

Ti-MCM-48 Direct hydrothermal synthesis. Prepared using

titanium isopropoxide (triethanolaminato) and

TEOS as the sources of Ti and Si, respectively, and

the Gemini-type surfactant 18–12–18 or cetyl-

benzyl dimethylammonium chloride (CBDAC) as

a template. In the grafting method, silicious MCM-

48 first prepared and then the dry surface grafted

with titanium isopropoxide

Pore diameter ¼ 2.6 nm,

Surface area ¼ 1296

m2/g (1093 m2/g for

grafted material)

(25)

(Continued)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

57

TABLE C1

Continued


improvements


morphology

References

Ti-SBA-15 Grafting method. SBA-15 prepared first using the

amphiphilic triblock copolymer poly(ethyleneox-

ide)–poly(propyleneoxide)–poly(ethyleneoxide)

(EO–PO–EO) as template and TEOS as Si source.

The composition was 2 g copolymer:0.021 mol

TEOS: 0.12 mol HCl:3.33 mol H2O. The solid was

calcined at 600 K for 4 h to remove the copolymer.

Ti in the form of titanium isopropoxide was grafted

onto the dehydrated surface of SBA-15

Pore diameter ¼ 6.3 nm,

surface area ¼ 518

m2/g, pore

volume ¼ 0.68

(25)

Direct synthesis under microwave heating.

Ti-substituted SBA-15 prepared using TEOS and

TiCl4 as sources of Si and Ti and the triblock

copolymer EO–PO–EO as structure-directing

agent. The gel was crystallized during heating in a

microwave environment

5–40 Mesopore size ¼ 7.3–

7.6 nm, specific surface

area ¼ 767–844 m2/g,

external surface

area ¼ 15–26 m2/g),

mesopore

volume ¼ 0.78–

0.95 cm3/g

(25)

Incipient wetness method. For every 1 g of SBA-

15, varying amounts of titanium isopropoxide in

10 g of ethanol were used for impregnation. The

titanium concentration in the solution varies from

0.05 to 5 M, depending on the desired titanium

loading. The impregnated material was dried and

calcined at 723 K for 5 h.

0.6–36

(XPS)

Pore size ¼ 4.2–5.1 nm,

specific surface

area ¼ 690–997 m2/g,

volume ¼ 0.81–

1.17 cm3/g

(27)

P.

Ratn

asamy

,D

.S

riniv

asan

dH

.K

nozin

ger

/A

dv

.C

atal.4

8(2

00

4)

1–

16

91

58

an acidic medium. The direct synthesis of Ti-SBA-15 molecular sieves under

microwave-hydrothermal conditions has considerably reduced the crystallization

times (27). Kevan and co-workers (26,324) prepared SBA-15 incorporating Ti by

incipient-wetness impregnation with titanium isopropoxide in ethanol followed

by calcination.

C.5. Ti-TUD-1

The mesoporous materials reported above are usually prepared from relatively

expensive surfactants. Some of them have poor hydrothermal stability. Further-

more, the MCM-41 host structure has a one-dimensional pore system with

consequent pore blockage and diffusion limitations. Shan et al. (32) reported the

synthesis of a three-dimensional and randomly connected mesoporous titano-

silicate (Ti-TUD-1, mesopore wall thickness ¼ 2.5–4 nm, surface area ,700–

1000 m2/g, tunable pore size ,4.5–5.7 nm) from triethanolamine (TEA). Ti-

TUD-1 showed higher activity (about 5.6 times) for cyclohexene epoxidation

than the framework-substituted Ti-MCM-41. Its activity was similar to that of the

Ti-grafted MCM-41(32).

Compositions of the synthesis gel and other physical characteristics of titan-

ium silicate materials obtained in various synthesis methodologies are listed in

Table C1.

References

1. Taylor, H.S., Proc. R. Soc. A 108, 105 (1925).

2. Taylor, H.S., J. Phys. Chem. 30, 145 (1926).

3. Thomas, J.M., Johnson, B.F.G., Raja, R., Sankar, G., and Midgley, P.A., Acc. Chem. Res.

36, 20 (2003).

4. Thomas, J.M., Raja, R., Johnson, B.F.G., O’Connell, T.J., Sankar, G., and Khimyak, T.,

Chem. Commun. 1126 (2003).

5. Taramasso, M., Perego, G., Notari, B., US Patent No. 4,410,501 (1983) to Snamprogetti S.

p. A.

6. Belussi, G., Carati, A., Clerici, M.G., Esposito, A., Millini, R., Buonomo, F., Belg, Patent

No.1,001,038 (1989) to Eniricerche S. p. A., Snamprogetti S. p. A., EniChem. S. p. A.

7. Reddy, J.S., Kumar, R., and Ratnasamy, P., Appl. Catal. 58, L1 (1990).

8. Serrano, D.P., Hong-Xin, L., and Davis, M.E., J. Chem. Soc., Chem. Commun. 745 (1992).

9. Camblor, M.A., Corma, A., Martinez, A., and Perezpariente, J., J. Chem. Soc., Chem.

Commun. 589 (1992).

10. Camblor, M.A., Corma, A., and Perezpariente, J., Zeolites 13, 82 (1993).

11. Corma, A., Camblor, M.A., Esteve, P., Martinez, A., and Perezpariente, J., J. Catal. 145,

151 (1994).

12. Camblor, M.A., Costantini, M., Corma, A., Gilbert, L., Esteve, P., Martinez, A., and

Valencia, S., Chem. Commun. 1339 (1996).


13. Blasco, T., Camblor, M.A., Corma, A., Esteve, P., Guil, J.M., Martinez, A., Perdigon-

Melon, J.A., and Valencia, S., J. Phys. Chem. B 102, 75 (1998).

14. Davis, R.J., Liu, Z., Tabora, J.E., and Wieland, W.S., Catal. Lett. 34, 101 (1995).

15. van der Waal, J.C., Kooyman, P.J., Jansen, J.C., and van Bekkum, H., Micropor. Mesopor.

Mater. 25, 43 (1998).

16. Tuel, A., Zeolites 15, 236 (1995).

17. Corma, A., Navarro, M.T., and Perezpariente, J., J. Chem. Soc., Chem. Commun. 147

(1994).

18. Blasco, A.T., Corma, A., Navarro, M.T., and Perezpariente, J., J. Catal. 156, 65 (1995).

19. Tanev, P.T., Chibwe, M., and Pinnavaia, T.J., Nature 368, 321 (1994).

20. Gontier, S., and Tuel, A., Zeolites 15, 601 (1995).

21. Sudhakar Reddy, J., Dicko, A., and Sayari, A., in “Synthesis of Porous Materials: Zeolites,

Clays and Nanostructures” (M. Occelli and H. Kessler, Eds.), p. 405–415. Marcel Dekker,

New York, 1997.

22. Koyano, K.A., and Tatsumi, T., Chem. Commun. 145 (1996).

23. Bagshaw, S.A., Prouzet, E., and Pinnavaia, T.J., Science 269, 1242 (1995).

24. Bagshaw, S.A., Di Renzo, F., and Fajula, F., Chem. Commun. 2209 (1996).

25. Morey, M.S., O’Brien, S., Schwarz, S., and Stucky, G.D., Chem. Mater. 12, 898 (2000).

26. Luan, Z., and Kevan, L., Micropor. Mesopor. Mater. 44/45, 337 (2001).

27. Newalkar, B.L., Olanrewaju, J., and Komarneni, S., Chem. Mater. 13, 552 (2001).

28. Kholdeeva, O.A., Derevyankin, A.Yu., Shmakov, A.N., Trukhan, N.N., Paukshtis, E.A.,

Tuel, A., and Romannikov, V.N., J. Mol. Catal. A: Chem. 158, 417 (2000).

29. Trukhan, N.N., Derevyankin, A.Yu., Shmakov, A.N., Paukshtis, E.A., Kholdeeva, O.A.,

and Romannikov, V.N., Micropor. Mesopor. Mater. 44/45, 603 (2001).

30. Trukhan, N.N., Romannikov, V.N., Paukshtis, E.A., Shmakov, A.N., and Kholdeeva,

O.A., J. Catal. 202, 110 (2001).

31. Wu, P., Tatsumi, T., Komatsu, T., and Yashima, T., J. Phys. Chem. B 105, 2897 (2001).

32. Shan, Z., Jansen, J.C., Marchese, L., and Maschmeyer, Th., Micropor. Mesopor. Mater.

48, 181 (2001).

33. Notari, B., Adv. Catal. 41, 253 (1996).

34. Vayssilov, G.N., Catal. Rev.—Sci. Engng 39, 209 (1997).

35. Bolis, V., Bordiga, S., Lamberti, C., Zecchina, A., Carati, A., Rivetti, F., Spano, G., and

Petrini, G., Micropor. Mesopor. Mater. 30, 67 (1999).

36. Ricchiardi, G., de Man, A.J.M., and Sauer, J., Phys. Chem. Chem. Phys. 2, 2195 (2000).

37. Damin, A., Bordiga, S., Zecchina, A., and Lamberti, C., J. Chem. Phys. 117, 226 (2002).

38. Damin, A., Bordiga, A., Zecchina, A., Doll, K., and Lamberti, C., J. Chem. Phys. 118,

10183 (2003).

39. Bordiga, S., Damin, A., Bonino, F., Ricchiardi, G., Zecchina, A., Tagliapietra, R., and

Lamberti, C., Phys. Chem. Chem. Phys. 5, 4390 (2003).

40. Li, C., Xiong, G., Liu, J., Ying, P., Xin, Q., and Feng, Z., J. Phys. Chem. B 105, 2993

(2001).

41. Ricchiardi, G., Damin, A., Bordiga, S., Lamberti, C., Spano, G., Rivetti, F., and Zecchina,

A., J. Am. Chem. Soc. 123, 11409 (2001).

42. Bordiga, S., Damin, A., Bonino, F., Ricchiardi, G., Lamberti, C., and Zecchina, A.,

Angew. Chem., Int. Ed. 41, 4734 (2002).

43. Bolis, V., Bordiga, S., Lamberti, C., Zecchina, A., Carati, A., Rivetti, F., Spano, G., and

Petrini, G., Langmuir 15, 5753 (1999).

44. Bonino, F., Damin, A., Bordiga, S., Lamberti, C., and Zecchina, A., Langmuir 19, 2155

(2003).


45. Damin, A., Bonino, F., Ricchiardi, G., Bordiga, S., Zecchina, A., and Lamberti, C., J. Phys.

Chem. B 106, 7524 (2002).

46. Sankar, G., Thomas, J.M., Catlow, C.R.A., Barker, C.M., Gleeson, D., and Kaltsoyannis,

N., J. Phys. Chem. B 105, 9028 (2001).

47. Gleeson, D., Sankar, G., Catlow, C.R.A., Thomas, J.M., Spano, G., Bordiga, S., Zecchina,

A., and Lamberti, C., Phys. Chem. Chem. Phys. 2, 4812 (2000).

48. Barker, C.M., Gleeson, D., Kaltsoyannis, N., Catlow, C.R.A., Sankar, G., and Thomas,

J.M., Phys. Chem. Chem. Phys. 4, 1228 (2002).

49. Lamberti, C., Bordiga, S., Arduino, D., Zecchina, A., Geobaldo, F., Spano, G., Genoni, F.,

Petrini, G., Carati, A., Villain, F., and Vlaic, G., J. Phys. Chem. B 102, 6382 (1998).

50. Zhao, Q., Bao, X., Wang, Y., Lin, L., Li, G., Guo, X., and Wang, X., J. Mol. Catal. A:

Chem. 157, 265 (2000).

51. Chaudhari, K., Srinivas, D., and Ratnasamy, P., J. Catal. 203, 25 (2001).

52. Srinivas, D., Manikandan, P., Laha, S.C., Kumar, R., and Ratnasamy, P., J. Catal. 217,

160 (2003).

53. Bonoldi, L., Busetto, C., Congiu, A., Marra, G., Ranghino, G., Salvalaggio, M., Spano, G.,

and Giamello, E., Spectrochim. Acta 58A, 1143 (2002).

54. Shetti, V.N., Manikandan, P., Srinivas, D., and Ratnasamy, P., J. Catal. 216, 461 (2003).

55. Carati, A., Flego, C., Massara, E.P., Millini, R., Carluccio, L., Parker, W. Jr., and Bellussi,

G., Micropor. Mesopor. Mater. 30, 137 (1999).

56. Krijnen, S., Sanchez, P., Jakobs, B.T.F., and van Hooff, J.H.C., Micropor. Mesopor.

Mater. 31, 163 (1999).

57. Grieneisen, J.L., Kessler, H., Fache, E., and Le Govic, A.M., Micropor. Mesopor. Mater.

37, 379 (2000).

58. Zicovich-Wilson, C., and Dovesi, R., J. Mol. Catal. A: Chem. 119, 449 (1997).

59. Cora, F., Alfredsson, M., Barker, C.M., Bell, R.G., Foster, M.D., Saadoune, I., Simperlar,

A., and Catlow, C.R.A., J. Solid State Chem. 176, 496 (2003).

60. Munakata, H., Oumi, Y., and Miyamoto, A., J. Phys. Chem. B 105, 3493 (2001).

61. Vayssilov, G.N., and van Santen, R.A., J. Catal. 175, 170 (1998).

62. de Man, A.J.M., and Sauer, J., J. Phys. Chem. 100, 5025 (1996).

63. Tozzola, G., Mantegazza, M.A., Ranghino, G., Petrini, G., Bordiga, S., Ricchiardi, G.,

Lamberti, C., Zulian, R., and Zecchina, A., J. Catal. 179, 64 (1998).

64. Gao, X., and Wachs, I.E., Catal. Today 51, 233 (1999).

65. Roberts, M.W., Catal. Lett. 67, 65 (2000).

66. Belussi, G., and Perego, C., in “Handbook of Heterogeneous Catalysis” (G. Ertl,

H. Knozinger and J. Weitkamp, Eds.), Vol. 5, p. 2329. Wiley, New York, 1997.

67. Petrini, G., and Leofanti, G., Mantegazza, M.A., and Pignataro, F., in “Green Chemistry:

Designing Chemistry for the Environment” (P.T. Anastas and T.C. Williamson, Eds.),

ACS Symposium Series No. 626, p. 33. American Chemical Society, Washington, DC,

1996.

68. Perego, C., Carati, A., Ingallina, P., Mantegazza, M.A., and Bellussi, G., Appl. Catal. A:

Chem. 221, 63 (2001).

69. Meiers, R., Dingerdissen, U., and Holderich, W.F., J. Catal. 176, 376 (1998).

70. Meiers, R., and Holderich, W.F., Catal. Lett. 59, 161 (1999).

71. Jenzer, G., Mallat, T., Maciejewski, M., Eigenmann, F., and Baiker, A., Appl. Catal. A:

Gen. 208, 125 (2001).

72. Stangland, E.E., Stevens, K.B., Andres, R.P., and Delgass, W.N., J. Catal. 191, 332

(2000).

73. Qi, C., Akita, T., Okumura, M., and Haruta, M., Appl. Catal. A: Gen. 218, 81 (2001).


74. Nijhuis, T.A., Huizinga, B.J., Makkee, M., and Moulijn, J.A., Ind. Engng Chem. Res. 38,

884 (1999).

75. Millini, R., Massara, E.P., Perego, G., and Bellussi, G., J. Catal. 133, 220 (1992), and 137,497 (1992).

76. Holm, R.H., Kennepohl, P., and Solomon, E.I., Chem. Rev. 96, 2239 (1996).

77. Millini, R., and Perego, G., Gazz. Chim. Ital. 126, 133 (1996).

78. Reddy, J.S., and Sayari, A., Stud. Surf. Sci. Catal. 94, 309 (1995).

79. Hasegawa, Y., and Ayame, A., Catal. Today 71, 177 (2001).

80. Contarini, S., van der Heide, P.A.W., Prakash, A.M., and Kevan, L., J. Electron Spectrosc.

Relat. Phenom. 125, 25 (2002).

81. Trong On, D., Bonneviot, L., Bittar, A., Sayari, A., and Kaliaguine, S., J. Mol. Catal. 74,

233 (1992).

82. Zecchina, A., Bordiga, S., Lamberti, C., Ricchiardi, G., Scarano, D., Petrini, G., Leofanti,

G., and Mantegazza, M., Catal. Today 32, 97 (1996).

83. Bordiga, S., Coluccia, S., Lamberti, C., Marchese, L., Zecchina, A., Boscherini, F., Buffa,

F., Genoni, F., Leofanti, G., Petrini, G., and Vlaic, G., J. Phys. Chem. 98, 4125 (1994).

84. Bordiga, S., Boscherini, F., Coluccia, S., Genoni, F., Lamberti, C., Leofanti, G., Marchese,

L., Petrini, G., Vlaic, G., and Zecchina, A., Catal. Lett. 26, 195 (1994).

85. Henry, P.F., Weller, M.T., and Wilson, C.C., J. Phys. Chem. B 105, 7452 (2001).

86. Lamberti, C., Bordiga, S., Zecchina, A., Carati, A., Fitch, A.N., Artioli, G., Petrini, G.,

Salvalaggio, M., and Marra, G.L., J. Catal. 183, 222 (1999).

87. Hijas, C.A., Jacubinas, R.M., Eckert, J., Henson, N.J., Hay, P.J., and Ott, K.C., J. Phys.

Chem. B 104, 12157 (2000).

88. Lamberti, C., Bordiga, S., Zecchina, A., Artioli, G., Marra, G., and Spano, G., J. Am.

Chem. Soc. 123, 2204 (2001).

89. van der Pol, A.J.H.P., Verduyn, A.J., and van Hooff, J.H.C., Appl. Catal. A: Gen. 92, 113

(1992).

90. Wilkenhoner, U., Langhendries, G., van Laar, F., Baron, G.V., Gammon, D.W., Jacobs,

P.A., and van Steen, E., J. Catal. 203, 201 (2001).

91. Boccuti, M.R., Rao, K.M., Zecchina, A., Leofanti, G., and Petrini, G., Stud. Surf. Sci.

Catal. 48, 133 (1989).

92. Le Noc, L., Trong On, D., Solomykina, S., Echchahed, B., Beland, F., Cartier di Moulin,

C., and Bonneviot, L., Stud. Surf. Sci. Catal. 101A, 611 (1996).

93. Shetti, V.N., Srinivas, D., Ratnasamy, P., Unpublished results.

94. Soult, A.S., Poore, D.D., Mayo, E.I., and Stiegman, A.E., J. Phys. Chem. B 105, 2687

(2001).

95. Yamashita, H., Ikeue, K., Takewaki, T., and Anpo, M., Top. Catal. 18, 95 (2002).

96. Gianotti, E., Yoshida, H., Dellarocca, V., Marchese, L., Martra, G., and Coluccia, S., Res.

Chem. Intermed. 29, 681 (2003).

97. Lamberti, C., Bordiga, S., Zecchina, A., Vlaic, G., Tozzola, G., Petrini, G., and Carati, A.,

J. Phys. 7, C2–851 (1997).

98. Behrens, P., Felsche, J., Vetter, S., Schulz-Ekloff, G., Jaeger, N.I., and Niemann, W.,

J. Chem. Soc., Chem. Commun. 678 (1991).

99. Bonneviot, L., Trong On, D., and Lopez, A., J. Chem. Soc., Chem. Commun. 685 (1993).

100. Trong On, D., Bittar, A., Sayari, A., Kaliaguine, S., and Bonneviot, L., Catal. Lett. 16, 85

(1992).

101. Marchese, L., Maschmeyer, T., Gianotti, E., Coluccia, S., and Thomas, J.M., J. Phys.

Chem. B 101, 8836 (1997).

102. Marchese, L., Gianotti, E., Dellarocca, V., Maschmeyer, T., Rey, F., Coluccia, S., and

Thomas, J.M., Phys. Chem. Chem. Phys. 1, 585 (1999).


103. Sinclair, P.E., Sankar, G., Catlow, C.R.A., Thomas, J.M., and Maschmeyer, T., J. Phys.

Chem. 4232 (1997).

104. Thomas, J.M., and Sankar, G., Acc. Chem. Res. 34, 571 (2001).

105. Huybrechts, D.R.C., Buskens, P.L., and Jacobs, P.A., J. Mol. Catal. 71, 129 (1992).

106. Jacobs, P.A., “Proceedings of DGMK Conference on Selective Oxidations in

Petrochemistry, Germany” 1992, p. 171.

107. Decottiguier, M., Phalippai, J., and Zarzycki, J., J. Mater. Sci. 13, 7605 (1978).

108. Soda, R., Bull. Chem. Soc. Jpn 34, 1491 (1961).

109. Duran, A., Sesna, C., Fornes, V., and Fernandez-Navarro, J.M., J. Non-Cryst. Solids 82, 69

(1986).

110. Camblor, M.A., Corma, A., and Perez-Pariente, J., J. Chem. Soc., Chem. Commun. 557

(1993).

111. Bellussi, G., Carati, A., Clerici, M.G., Maddinelli, G., and Millini, R., J. Catal. 133, 220

(1992).

112. Bellusi, G., and Rigutto, M.S., Stud. Surf. Sci. Catal. 85, 177 (1994).

113. Scarano, D., Zecchina, A., Bordiga, S., Geobaldo, F., Spoto, G., Petrini, G., Leofanti, G.,

Padovan, M., and Tozzola, G., J. Chem. Soc., Faraday Trans. 89, 4123 (1993).

114. Su, Y., Balnics, M.L., and Bunker, B.C., J. Phys. Chem. B 104, 160 (2000).

115. Smirnov, K.S., and van de Graaf, B., Micropor. Mater. 7, 133 (1996).

116. Li, C., Xiong, G., Xin, Q., Lin, J., Ying, P., Feng, Z., Li, J., Yang, W., Wang, Y., Wang,

G., Liu, X., Lin, M., Wang, X., and Min, E., Angew. Chem., Int. Ed. 38, 2220 (1999).

117. Yang, Q., Wang, S., Lu, J., Xiong, G., Feng, Z., Xin, Q., and Li, C., Appl. Catal. A: Gen.

194/195, 507 (2000).

118. Yin, J., Feng, Z., Xu, L., Li, M., Xin, Q., Lin, Y., and Li, C., Mater. Chem. 13, 1994

(2001).

119. Deo, G., Turex, A.M., Wachs, I.E., Huybrechts, D.R.C., and Jacobs, P.A., Zeolites 13, 365

(1993).

120. Nishimura, Y., Hirakawa, A.Y., and Tsuboi, M., in “Advances in Infrared and Raman

Spectroscopy” (R.J.H. Clark and R.E. Hester, Eds.). Heydon & Sons, London, 1978.

121. Carey, P.R., “Biochemical Applications of Raman and Resonance Raman Spectro-

scopies.” Academic Press, New York, 1982.

122. Tuel, A., Diab, J., Gelin, P., Dufaux, M., Dutel, J.-F., and Ben Taarit, Y., J. Mol. Catal. 63,

95 (1990).

123. Zecchina, A., Spoto, G., Bordiga, S., Ferrero, A., Leofanti, G.G., and Padovan, M., Stud.

Surf. Sci. Catal. 69, 251 (1991).

124. Prakash, A.M., and Kevan, L., J. Catal. 178, 588 (1998).

125. Luan, Z., and Kevan, L., J. Phys. Chem. B 101, 2020 (1997).

126. Prakash, A.M., Kurshev, V., and Kevan, L., J. Phys. Chem. B 101, 9794 (1997).

127. Prakash, A.M., Sung-Suh, H.M., and Kevan, L., J. Phys. Chem. B 102, 857 (1998).

128. Prakash, A.M., Kevan, L., Zahedi-Niaki, M.H., and Kaliaguine, S., J. Phys. Chem. B 103,

831 (1999).

129. Zhu, Z., Hartmann, M., Macs, E.M., Czernuszewicz, R.S., and Kevan, L., J. Phys. Chem.

B 104, 4690 (2000).

130. Bal, R., Chaudhari, K., Srinivas, D., Sivasanker, S., and Ratnasamy, P., J. Mol. Catal. A:

Chem. 162, 199 (2000).

131. Chaudhari, K., Bal, R., Srinivas, D., Chandwadkar, A.J., and Sivasanker, S., Micropor.

Mesopor. Mater. 50, 209 (2001).

132. Morey, M.S., O’Brian, S., Schwarz, S., and Stucky, G.D., Chem. Mater. 12, 898 (2000).

133. Lin, W., and Frei, H., J. Am. Chem. Soc. 124, 9292 (2002).


134. Bordiga, S., Damin, A., Bonino, F., Zecchina, A., Spano, G., Rivetti, F., Bolis, V.,

Prestipino, C., and Lambert, C., J. Phys. Chem. B. 106, 9892 (2002).

135. Zecchina, A., Spoto, G., Bordiga, S., Padovan, M., Leofanti, G., and Petrini, G., Stud. Surf.

Sci. Catal. 65, 671 (1991).

136. Armaroli, T., Milella, F., Notari, B., Willey, R.J., and Busca, G., Top. Catal. 15, 63 (2001).

137. Huybrechts, D.R.C., Vaesen, I., Li, H.X., and Jacobs, P.A., Catal. Lett. 8, 237 (1991).

138. Thangaraj, A., Sivasanker, S., and Ratnasamy, P., J. Catal. 137, 252 (1992).

139. Ichihashi, H., and Sato, H., Appl. Catal. A: Gen. 221, 359 (2001).

140. Srivastava, R., Srinivas, D., and Ratnasamy, P., Catal. Lett. 91, 133 (2003).

141. Knozinger, H., (G. Ertl, H. Knozinger and J. Weitkamp, Eds.), Handbook of

Heterogeneous Catalysis, Vol. 2, p. 707. Wiley-VCH, Weinheim, 1997.

142. Paushkits, E.A., and Yurchenko, E.N., Russ. Chem. Rev. 52, 42 (1983).

143. Manoilova, O.V., Dakka, J., Sheldon, R.A., and Tsyganenko, A.A., Stud. Surf. Sci. Catal.

94, 163 (1995).

144. Knozinger, H., in “The Hydrogen Bond” (P. Schuster, G. Zundel and C. Sandorfy, Eds.),

Vol. III, p. 1265. North-Holland, Amsterdam, 1976.

145. Zhuang, J., Yan, Z., Liu, X., Liu, X., Han, X., Bao, X., and Mueller, U., Catal. Lett. 83, 87

(2002).

146. Knozinger, H., and Krietenbrink, H., J. Chem. Soc., Faraday Trans. 71, 2421 (1975).

147. Redijk, J., Zaur, A.P., and Groeneveld, W.L., Rec. Trav. Chim. 86, 1127 (1967).

148. Bertran, J.F., La Serna, B., Doerffel, K., Dathe, K., and Kabish, G., J. Mol. Struct. 95, 1

(1982).

149. Busca, G., Saussey, H., Saur, O., Lavalley, J.C., and Lorenzelli, V., Appl. Catal. 14, 245

(1985).

150. Barzetti, T., Selli, E., Moscotti, D., and Forni, L., J. Chem. Soc., Faraday Trans. 92, 1401

(1996).

151. Farnoth, W.E., and Gorte, R.J., Chem. Rev. 95, 615 (1995).

152. Srinivas, D., Srivastava, R., Ratnasamy, P., Catal. Today (2004) (in press).

153. Zecchina, A., Bordiga, S., Spoto, G., Damin, A., Berlier, G., Bonino, F., Prestipino, C.,

and Lamberti, C., Top. Catal. 21, 67 (2002).

154. Connor, J.A., and Ebsworth, E.A.V., in “Advances in Inorganic Chemistry and

Radiochemistry” (H.J. Emeleus and A.G. Sharpe, Eds.), p. 287. Academic Press, New

York, 1964.

155. Mimoun, H., in “The Chemistry of Functional Groups, Peroxides” (S.J. Patai, Ed.), p. 463.

Wiley, New York, 1983.

156. Fujita, S., Suzuki, K., Ohkawa, M., Mori, T., Jida, Y., Miwa, Y., Masuda, H., and

Shimada, S., Chem. Mater. 15, 255 (2003).

157. Giamello, E., Sojka, Z., Che, M., and Zecchina, A., J. Phys. Chem. 90, 6084 (1986).

158. Bodoardo, S., Geobaldo, F., Penazzi, N., Arrabito, M., Rivetti, F., Spano, G., Lamberti, C.,

and Zecchina, A., Electrochem. Commun. 2, 349 (2000).

159. Che, M., and Tench, A., Adv. Catal. 32, 1 (1983).

160. Njo, S.L., van Koningsveld, H., and van de Graaf, B., J. Phys. Chem. B 101, 10065 (1997).

161. Oumi, Y., Matsuba, K., Kubo, M., Inui, T., and Miyamoto, A., Micropor. Mesopor. Mater.

4, 53 (1995).

162. Atoguchi, T., and Yao, S., J. Mol. Catal. A: Chem. 191, 281 (2003).

163. Zentys, A., and Catlow, C.R.A., Catal. Lett. 22, 251 (1993).

164. Maseras, F., Chem. Commun. 1821 (2000), and references therein.

165. Kumar, P., Kumar, R., and Pandey, B., Synlett 289 (1995).

166. Ramaswamy, M., and Roesky, H.W., Angew. Chem., Int. Ed. Engl. 36, 477 (1997).


167. Sheldon, R.A., Arends, I.W.C.E., and Lempers, H.E.B., Heterogeneous catalysts for liquid

phase oxidations (supported reagents and catalysis in chemistry). Spec. Publ. R. Soc.

Chem. 216, 37 (1998).

168. Sheldon, R.A., Arends, I.W.C.E., and Lempers, H.E.B., Collect. Czech. Chem. Commun.

63, 1724 (1998).

169. Clerici, M.G., in “Fine Chemicals through Heterogeneous Catalysis” (R.A. Shelden and

H. van Bekkum, Eds.), p. 538, Wiley-VCH, Weinheim, Germany, 2001.

170. Patai, S. (Ed.), “The Chemistry of Peroxides.” Wiley, New York, 1983.

171. Cremer, D., in “The Chemistry of Peroxides” (S. Patai, Ed.), p. 43. Wiley, New York,

1983, Chapter 43.

172. Nangia, P.S., and Benson, S.W., J. Phys. Chem. 83, 1138 (1979).

173. Benson, S.W., and Nangia, P.S., Acc. Chem. Res. 12, 223 (1979).

174. Kochi, J.K., “Organometallic Mechanisms and Catalysis.” Academic Press, New York,

1978, p. 50.

175. Gunther, K., Filby, W.G., and Eiben, K., Tetrahedron Lett. 251 (1971).

176. Sheldon, R.A., and van Doorn, J.A., J. Catal. 34, 242 (1974).

177. Sharpless, K.B., and Akashi, K., J. Am. Chem. Soc. 98, 1986 (1976).

178. Nam, W., Han, H.J., Oh, S.-Y., Lee, Y.J., Choi, M.-M., Han, S.-Y., Kim, C., Woo, S.K.,

and Shin, W., J. Am. Chem. Soc. 122, 8677 (2000).

179. Sawaki, Y., in “The Chemistry of Hydroxyl, Ether and Peroxide Groups” (S. Patai, Ed.),

Suppl. E, Vol. 2, p. 587. Wiley, New York, 1993.

180. Arends, I.W.C.E., and Sheldon, R., Topics Catal. 19, 133 (2002).

181. Dusi, M., Mallat, T., and Baiker, A., Catal. Rev.—Sci. Engng 42, 213 (2000).

182. Bartok, M., and Schneider, G., in “The Chemistry of Double-Bonded Functional Groups,

Part 2” (S. Patai, Ed.), p. 1229. Wiley, New York, 1997.

183. Sheldon, R.A., Van Doorn, J.A., Schran, C.W.A., and De Jong, A.J., J. Catal. 31, 438

(1973).

184. Sheldon, R.A., in “Applied Homogeneous Catalysis with Organometallic Compounds”

(B. Cornils and W.A. Hartman, Eds.), Vol. 1, p. 411. VCH, Weinheim, 1996.

185. Sheldon, R.A., Arends, I.W.C.E., and Lempers, H.E.B., in “Supported Reagents and

Catalysts in Chemistry” (B.K. Hodnett, A.P. Kybett, J.H. Clark and K. Smith, Eds.), p. 37.

Royal Society of Chemistry, Limerick, 1997.

186. Sheldon, R.A., and van Santen, R.A. (Eds.), “Catalytic Oxidation: Principles and

Applications.” World Scientific, Singapore, 1995.

187. “Industrial Organic Chemicals: Starting Materials and Intermediates: An Ullmann’s

Encyclopedia,” Vol. 7, p. 4135, Wiley-VCH, Weinheim, Germany, 1999.

188. Schmidt, I., Krogh, A., Wienberg, K., Carlsson, A., Brorson, M., and Jacobsen, C.J.H.,

Chem. Commun. 2157 (2000).

189. Guidotti, M., Ravasio, N., Psaro, R., Ferraris, G., and Moretti, G., J. Catal. 214, 242 (2003).

190. Bouh, A.O., Rice, G.L., and Scott, S.L., J. Am. Chem. Soc. 121, 7201 (1999).

191. Corma, A., Esteve, P., Martınez, A., and Valencia, S., J. Catal. 152, 18 (1995).

192. Romano, U., Esposito, A., Maspero, F., Neri, C., and Clerici, M.G., Chim. Ind. 72, 610

(1990).

193. Tatsumi, T., Yako, M., Nakamura, M., Yuhara, Y., and Tominaga, H., J. Mol. Catal. 78,

L41 (1993).

194. Bhaumik, A., Kumar, R., and Ratnasamy, P., Stud. Surf. Sci. Catal. 84C, 1883 (1994).

195. Kumar, R., Pais, G.C.G., Pandey, B., and Kumar, P., J. Chem. Soc., Chem. Commun. 1315

(1995).

196. van der Waal, J.C., Rigutto, M.S., and van Bekkum, H., Appl. Catal. A: Gen. 167, 331

(1998).


197. Corma, A., Navarro, M.T., Perez-Pariente, J., and Sanchez, F., Stud. Surf. Sci. Catal. 84,

69 (1994).

198. Vercruysse, K.A., Klingeleers, D.M., Colling, T., and Jacobs, P.A., Stud. Surf. Sci. Catal.

117, 469 (1998).

199. Adam, W., Corma, A., Reddy, T.I., and Renz, M., J. Org. Chem. 62, 3631 (1997).

200. Adam, W., Kumar, R., Reddy, T.I., and Renz, M., Angew. Chem., Int. Ed. Engl. 35, 880

(1996).

201. Hutchings, G.J., and Lee, D.F., J. Chem. Soc., Chem. Commun. 1095 (1994).

202. Hutchings, G.J., Lee, D.F., and Minihan, A.R., Catal. Lett. 33, 369 (1995).

203. Hutchings, G.J., Lee, D.F., and Minihan, A.R., Catal. Lett. 39, 83 (1996).

204. Clerici, M.G., and Ingallina, P., J. Catal. 140, 71 (1993).

205. Shetti, V.N., Srinivas, D., and Ratnasamy, P., J. Mol. Catal. A: Chem. 210, 171 (2004).

206. Ratnasamy, P., and Kumar, R., Stud. Surf. Sci. Catal. 97, 367 (1995).

207. Laha, S.C., and Kumar, R., J. Catal. 208, 339 (2002).

208. Laufer, W., Meiers, R., and Holderich, W.F., J. Mol. Catal. A: Chem. 141, 215 (1999).

209. Sato, A., Oguri, M., Tokumaru, S., Miyake, T., JP Patent 08269029A (1996).

210. Miller, U., Lingelbach, P., Bassler, P., Harder, W., Eller, K., Kohl, V., Dembowski, J.,

Rieber, N., Fischer, M., EP Patent 7,724,91 A1 (1995).

211. Olah, G.A., and Ohnishi, R., J. Org. Chem. 43, 865 (1978).

212. Olah, G.A., Fung, A.P., and Keumi, T., J. Org. Chem. 46, 4305 (1981).

213. Kurz, M.E., and Johnson, G., J. Org. Chem. 36, 3184 (1971).

214. Olah, G.A., and Ernst, T.D., J. Org. Chem. 54, 1204 (1989).

215. Ogata, Y., and Tomizawa, K., J. Org. Chem. 43, 261 (1978).

216. Clerici, M.G., Appl. Catal. 68, 249 (1991).

217. Huybrechts, D.R.C., Parton, R.F., and Jacobs, P.A., Stud. Surf. Sci. Catal. 65, 225 (1991).

218. Ramaswamy, A.V., and Sivasanker, S., Catal. Lett. 22, 239 (1993).

219. Ramaswamy, A.V., Sivasanker, S., and Ratnasamy, P., Micropor. Mater. 2, 451 (1994).

220. Sen, A., in “Applied Homogeneous Catalysis with Organometallic Compounds,” 2nd

Edition. p. 1226 Wiley, New York, 2002.

221. Khouw, C.B., Dartt, C.B., Labinger, J.A., and Davis, M.E., J. Catal. 149, 195 (1995).

222. Geobaldo, F., Bordiga, S., Zecchina, A., Giamello, E., Leofanti, G., and Petrini, G., Catal.

Lett. 16, 109 (1992).

223. Poladi, R.H.P.R., and Landry, C.C., Micropor. Mesopor. Mater. 52, 11 (2002).

224. Thangaraj, A., Kumar, R., and Ratnasamy, P., Appl. Catal. 55, L1 (1990).

225. Rao, P.R.H.P., Ramaswamy, A.V., and Ratnasamy, P., J. Catal. 137, 225 (1992).

226. Rao, P.R.H.P., and Ramaswamy, A.V., Appl. Catal. A: Gen. 93, 123 (1993).

227. Bianchi, D., Balducci, L., Bortolo, R., Aloisio, D., Ricci, M., Tassinari, R., and Ungarelli,

R., Angew. Chem., Int. Ed. 42, 4937 (2003).

228. Reddy, J.S., and Jacobs, P.A., J. Chem. Soc., Perkin. Trans. 1, 2665 (1993).

229. Wu, P., and Tatsumi, T., J. Catal. 214, 317 (2003).

230. Sonawane, H.R., Pol, A.V., Moghe, P.P., Biswas, S.S., and Sudalai, A., J. Chem. Soc.,

Chem. Commun. 1215 (1994).

231. Zecchina, A., Spoto, G., Bordiga, S., Geobaldo, F., Petrini, G., Leofanti, G., Padova, M.,

Mantegazza, M., and Roffia, P., in “Proceedings of the 10th International Congress on

Catalysis” (L. Guczi, F. Solymosi and P. Tetenyi, Eds.), p. 719. Akademiai Kiado,

Budapest, 1993; Stud. Surf. Sci. Catal. 75, 719 (1993).

232. Reddy, R.S., Reddy, J.S., Kumar, R., and Kumar, P., J. Chem. Soc., Chem. Commun. 84

(1992).

233. Kumar, R., Reddy, J.S., Reddy, R.S., and Kumar, P., in “Selective Oxidation in Petro-

chemistry” (J. Weitkamp, Ed.), DGMK Conference, p. 367, 1992.


234. Trukhan, N.N., Romannikov, V.N., Shmakov, A.N., Vanina, M.P., Paukshtis, E.A.,

Bukhtiyarov, V.I., Kriventsov, V.V., Danilov, I.Y., and Kholdeeva, O.A., Micropor.

Mesopor. Mater. 59, 73 (2003).

235. Hutter, R., Mallat, T., and Baiker, A., J. Catal. 153, 177 (1995).

236. Kholdeeva, O.A., Trukhan, N.N., Vanina, M.P., Romanikov, V.N., Parmon, V.N., Bialon,

J.M., and Jerzebski, A.B., Catal. Today 75, 203 (2002).

237. Hagen, A., Schueler, K., and Roessner, F., Micropor. Mesopor. Mater. 51, 23 (2002).

238. Hulea, V., Fajula, F., and Bousquet, J., J. Catal. 198, 179 (2001).

239. van der Pol, A.J.H.P., and van Hooff, J.H.C., J. Appl. Catal. A: Gen. 106, 97 (1993).

240. Corma, A., Esteve, P., and Martınez, A., Appl. Catal. A: Gen. 143, 87 (1996).

241. Sasidharan, M., Suresh, S., and Sudalai, A., Tetrahedron Lett. 36, 9071 (1995).

242. Harrison, I.T., and Harrison, S., J. Chem. Soc., Chem. Commun. 752 (1966).

243. Clarlsen, P.H.J., Katsuki, T., Martin, V.S., and Sharpless, K.B., J. Org. Chem. 46, 3936

(1981).

244. Bhaumik, A., Kumar, P., and Kumar, R., Catal. Lett. 40, 47 (1996).

245. Wang, Z.B., Mizusaki, T., Sano, T., and Kawakami, Y., Bull. Chem. Soc. Jpn 70, 2567

(1997).

246. Joseph, R., Sudalai, A., and Ravindranathan, T., Tetrahedron Lett. 35, 5493 (1994).

247. Reddy, J.S., Ravishanker, R., Sivasanker, S., and Ratnasamy, P., Catal. Lett. 17, 139 (1993).

248. Srivastava, R., Srinivas, D., and Ratnasamy, P., Catal. Lett. 91, 133 (2003).

249. Srivasta, R., Srinivas, D., and Ratnasamy, P., in “Proceedings 14th International Zeolite

Conference” (E.W.J. van Steen, L.H. Callanan, M. Claeys and C.T. O’Connor, Eds.),

p. 2703, Produced by: Document Transformation Technologies, Cape Town, South Africa

(2004).

250. Bhanage, B.M., Fujita, S.I., Ikushima, Y., Torri, K., and Arai, M., Green Chem. 5, 71

(2003).

251. Ma, X., Guo, H., Wang, S., and Sun, Y., Fuel Proc. Technol. 83, 275 (2003).

252. Otera, J., Chem. Rev. 93, 1449 (1993).

253. Schuchardt, U., Sercheli, R., and Vargas, R.M., J. Braz. Chem. Soc. 9 (1998).

254. Tatsumi, T., Watanabe, Y., and Koyano, K.A., Chem. Commun. 2281 (1996).

255. Mukaiyama, T., Bano, K., and Narasaka, K., J. Am. Chem. Soc. 96, 7503 (1974).

256. Kawai, M., Onaka, M., and Izumi, B., Bull. Chem. Soc. Jpn 61, 1237 (1988).

257. Sasidharan, M., and Kumar, R., J. Catal. 220, 326 (2003).

258. Sasidharan, M., Wu, P., and Tatsumi, T., J. Catal. 209, 260 (2002).

259. Schuster, W., Niederer, J.P.M., and Hoelderich, W.F., Appl. Catal. A: Gen. 209, 131 (2001).

260. Hanmann, M.M., in “Photodegradation of Water Pollutants”. CRC Press, New York, 1995.

261. Anpo, M., Yamashita, H., Ichihashi, Y., Fujii, Y., and Honda, M., J. Phys. Chem. B 101,

2632 (1997).

262. Kang, M.G., Park, H.S., and Kim, K.-J., J. Photochem. Photobiol. A: Chem. 149, 175

(2002).

263. Anpo, M., Yamashita, H., Ikeue, K., Fujii, Y., Zhang, S.G., Ichihashi, Y., Park, D.R.,

Suzuki, Y., Koyano, K., and Tatsumi, T., Catal. Today 44, 327 (1998).

264. Ulagappan, N., and Frei, H., J. Phys. Chem. A 104, 7834 (2000).

265. Yamashita, H., Ichihashi, Y., Zhang, S.G., Matsumura, Y., Souma, Y., Tatsumi, T., and

Anpo, M., Appl. Surf. Sci. 121/122, 305 (1997).

266. Thangaraj, A., Puthoor, L., and Sivasanker, S., Indian J. Chem. A 33, 255 (1994).

267. Atoguchi, T., and Yao, S., J. Mol. Catal. A: Chem. 176, 173 (2001).

268. Tantanak, D., Vincent, M.A., and Hillier, I.H., J. Chem. Soc., Chem. Commun. 1031

(1998).

269. Sinclair, P.E., and Catlow, C.R.A., J. Phys. Chem. B 103, 1084 (1999).


270. Sauer, M.V.C., and Edwards, J.O., J. Phys. Chem. 75, 3004 (1971).

271. Bach, R.D., and Knigh, J.W., Org. Synth. 60, 63 (1981).

272. McIsaac, J.E., Ball, R.E., and Behrman, E.J., J. Org. Chem. 36, 3048 (1971).

273. Tatsumi, T., Koyano, K.A., and Igarashi, N., Chem. Commun. 325 (1998).

274. Wu, P., Tatsumi, T., Komatsu, T., and Yashima, T., Chem. Mater. 14, 1657 (2002).

275. Thangaraj, A., Kumar, R., Mirajkar, S.P., and Ratnasamy, P., J. Catal. 130, 1 (1991).

276. Mantegazza, M.A., Petrini, G., Spano, G., Bagatin, R., and Rivetti, F., J. Mol. Catal. A:

Chem. 146, 223 (1999).

277. Chaudhari, K., Bal, R., Srinivas, D., Chandwadkar, A.J., and Sivasanker, S., Micropor.

Mesopor. Mater. 50, 209 (2001).

278. Adam, W., Harold, M., Hill, C.L., Saha-Moller, C.R., and Eur, J. Org. Chem. 941 (2002).

279. Corma, A., Chem. Rev. 97, 2373 (1997).

280. Biz, S., and Occelli, M.L., Catal. Rev. Sci. Engng 40, 329 (1998).

281. Thangaraj, A., Eapen, M.J., Sivasanker, S., and Ratnasamy, P., J. Catal. 12, 943 (1992).

282. Padovan, M., Leofanti, G., Roffia, P., EP Patent No. 311983 (1989).

283. Padovan, M., Genoni, F., Leofanti, G., Petrini, G., Trezza, G., and Zecchina, A., Stud.

Surf. Sci. Catal. 63, 431 (1991).

284. Bellussi, G., Buonomo, F., Esposito, A., Clerici, M., Romano, U., Notari, B., US Patent

No. 4, 701,428 (1987).

285. Tuel, A., Ben Taarit, Y., and Naccache, C., Zeolites 13, 454 (1993).

286. Tuel, A., and Ben Taarit, Y., Micropor. Mater. 1, 179 (1993).

287. Tuel, A., and Ben Taarit, Y., Zeolites 14, 272 (1994).

288. Tuel, A., and Ben Taarit, Y., Appl. Catal. A. Gen. 110, 137 (1994).

289. Tuel, A., Zeolites 16, 108 (1996).

290. Gontier, S., and Tuel, A., Zeolites 16, 184 (1996).

291. Tuel, A., Teissier, R., EP Patent Appl. No. 665188 A 1 (1995).

292. Jorda, E., Tuel, A., Treissier, R., and Kervennal, J., Zeolites 19, 238 (1997).

293. Tuel, A., Catal. Lett. 51, 59 (1998).

294. Sabde, D.P., Hegde, S.G., and Dongare, M.K., J. Mater. Chem. 10, 1365 (2000).

295. Guth, J.L., Kessler, H., Higel, J.M., Lamblin, J.M., Patarin, J., Seive, A., Chezeou, J.M.,

and Way, A., ACS Symp. Ser. 398, 176 (1989).

296. Delmotte, L., Soulard, M., Guth, F., Seive, A., Lopez, A., and Guth, J.L., Zeolites 10, 778

(1990).

297. Grieneisen, H., Kessler, H., Fache, E., and Le Govic, A.M., Micropor. Mesopor. Mater.

37, 379 (2000).

298. Kumar, R., Bhaumik, A., Ahedi, R.K., and Ganapathy, S., Nature 381, 298 (1996).

299. Bhaumik, A., Ganapathy, S., and Kumar, R., Stud. Surf. Sci. Catal. 113, 225 (1998).

300. Kumar, R., Mukherjee, P., Pandey, R.K., Rajmohanan, P., and Bhaumik, A., Micropor.


301. Ahn, W.S., Kang, K.K., and Kim, K.Y., Catal. Lett. 72, 229 (2001).

302. Kulkarni, S.K., Prasad, M.R., Kamalakar, G., Raghavan, K.V., Prasad, P.S.S., Rao, K.N.,

US Patent No. 6,387,348 B1 (2002).

303. Prasad, M.R., Kamalakar, G., Kulkarni, S.J., Raghavan, K.V., Rao, K.V., Sai Prasad, P.S.,

and Madhavendra, S.S., Catal. Commun. 3, 399 (2002).

304. Khomane, R.B., Kulkarni, B.D., Paraskar, A., and Sainker, S.R., Mater. Chem. Phys. 76,

99 (2002).

305. Reddy, J.S., and Kumar, R., J. Catal. 130, 440 (1991).

306. Uguina, M.A., Serrano, D.P., Ovejero, G., van Griekson, R., and Camacho, M., Zeolites

18, 368 (1997).

307. de Lucas, A., Rodrıguez, L., and Sanchez, P., Appl. Catal. A: Gen. 180, 375 (1999).


308. de Lucas, A., Rodrıguez, L., and Sanchez, P., Chem. Engng Res. Des. 78, 136 (2000).

309. Reddy, K.M., Kaliaguine, S., and Sayari, A., Catal. Lett. 23, 169 (1994).

310. Reddy, K.M., Kaliaguine, S., Sayari, A., Ramaswamy, A.V., Reddy, V.S., and Bonneviot,

L., Catal. Lett. 23, 174 (1994).

311. Tuel, A., and Ben Taarit, Y., Zeolites 15, 164 (1995).

312. Sexton, R.J., Crocco, G.L., Zajacek, J.G., Wijesekea, K.S., EP Patent No. 659685 (1994).

313. Camblor, M.A., Costaitini, M., Corma, A., Gilbert, L., Esteve, P., Martinez, A., and

Valentia, S., Appl. Catal. 133, L185 (1995).

314. Blasco, T., Camblor, M.A., Corma, A., Esteve, P., Martinez, A., Prieto, C., and Valencia,

S., Chem. Commun. 2367 (1996).

315. van der Waal, J.C., Lin, P., Rigutto, M.S., and van Bekkum, H., Stud. Surf. Sci. Catal. 150,

1093 (1997).

316. Reddy, J.S., and Sayari, A., J. Chem. Soc., Chem. Commun. 23 (1995).

317. Dartt, C.B., and Davis, M.E., Appl. Catal. A: Gen. 143, 53 (1996).

318. Jappar, N., Xia, Q., and Tatsumi, T., J. Catal. 180, 132 (1998).

319. Maschmeyer, T., Rey, F., Sankar, G., and Thomas, J.M., Nature 378, 159 (1995).

320. Corma, A., Jorda, J.L., Navarro, M.T., and Rey, F., Chem. Commun. 1899 (1998).

321. Corma, A., Domine, M., Gaona, J.A., Jorda, J.L., Navarro, M.T., Rey, F., Perez-Pariente,

J., Tsuji, J., McCulloch, B., and Nemeth, L.T., Chem. Commun. 2211 (1998).

322. D’Amore, M.B., and Schwarz, S., Chem. Commun. 121 (1999).

323. Morey, M., Davidson, A., and Stucky, G.D., Micropor. Mesopor. Mater. 6, 99 (1996).

324. Luan, Z., Maes, E.M., van der Heide, P.A.W., Zhao, D., Czernuszewicz, R.S., and Kevan,

L., Chem. Mater. 11, 3680 (1999).

325. Uguina, M.A., Ovejero, G., van Griechen, R., Serrano, D.P., and Camacho, M., J. Chem.

Soc., Chem. Commun. 27 (1994).

326. Uguina, M.A., Serrano, D.P., Ovejero, G., van Grieken, R., and Camacho, M., Appl. Catal.

A: Gen. 124, 391 (1995).

327. Guo, G.L., Zhang, X., and Wang, X., Stud. Surf. Sci. Catal. 112, 499 (1997).

328. Jung, K.T., Hyum, J.H., Shul, Y.G., and Kim, D.S., Zeolites 19, 161 (1997).

329. Wang, X.S., and Guo, X.W., Catal. Today 51, 45 (1999).

330. Gao, H., Lu, W., and Chen, Q., Micropor. Mesopor. Mater. 34, 307 (2000).

331. Kuznicki, S.M., US Patent No. 4,938,939 (1990).

332. Kuznicki, S.M., Thrush, A.K., EP Patent No. 0405978 A1 (1990).

333. Kuznicki, S.M., US Patent No. 4,853,202 (1989).

334. Chapman, D.M., and Roe, A.L., Zeolite 10, 730 (1990).

335. Anderson, M.W., Terasaki, O., Ohsuna, T., Phillippou, A., Mackay, S.P., Ferreira, A.,

Rochca, S., and Lidin, S., Nature 367, 347 (1994).

336. Anderson, M.W., Terasaki, O., Ohsuna, T., Malley, P.J.O., Phillippou, A., Mackay, S.P.,

Ferreira, A., Rochca, J., and Lidin, S., Phil. Mag. B 71, 813 (1995).

337. Valtchev, V.P., J. Chem. Soc., Chem. Commun. 1435 (1994).

338. Valtchev, V.P., and Mintova, S., Zeolites 14, 697 (1994).

339. Liu, X., and Thomas, J.K., Chem. Commun. 1435 (1996).

340. Das, T.Kr., Chandwadkar, A.J., Budhkar, A.P., Belhekar, A.A., and Sivasanker, S.,

Micropor. Mesopor. Mater. 4, 195 (1995).

341. Das, T.Kr., Chandwadkar, A.J., Budhkar, A.P., and Sivasanker, S., Micropor. Mesopor.

Mater. 5, 401 (1996).

342. Rocha, J., Ferreira, A., Lin, Z., and Anderson, M.W., Micropor. Mesopor. Mater. 23, 253

(1998).

343. Yang, X., Paillaud, J.-L., van Bruekelen, H.F.W., Kesser, H., and Duprey, E., Micropor.



Electron Microscopy and the Materials

Chemistry of Solid Catalysts

JOHN MEURIG THOMAS*Davy Faraday Research Laboratory, The Royal Institution of Great Britain,

21 Albemarle Street, London W1S 4BS, UK

and also at

Department of Materials Science, Cambridge CB2 1QY, UK

and

PRATIBHA L. GAI*DuPont, Central Research and Development Laboratories, Experimental Station,

Wilmington, DE 19880-0356, USA

and also at

Department of Materials Science, University of Delaware, Newark, DE 19716, USA

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

II. Electron Microscopy (EM) Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

II.A. Electron Microscopy in Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

II.B. Imaging in the Electron Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

II.C. TEM Imaging Method Using Diffraction Contrast . . . . . . . . . . . . . . . . . . . . . . . 179

II.D. Theoretical Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

III. High-Resolution Transmission Electron Microscopy. . . . . . . . . . . . . . . . . . . . . . . . . . 181

III.A. Conditions Required for Optimizing HRTEM Images . . . . . . . . . . . . . . . . . . . 182

III.B. Development of HRTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184


Catalysts by HRTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

III.C.1. L-Type Zeolite Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

III.C.2. Metal-Substituted Aluminum Phosphate (MAPO-36)

Microporous Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

III.C.3. High-Silica Microporous SSZ-48 Catalysts . . . . . . . . . . . . . . . . . . . . . 187

III.C.4. Intergrowths in Zeolite Catalysts: Coherent,

Recurrent, and Random . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

IV. Chemical Composition Analysis with the Analytical Electron Microscope. . . . . . . . . . 191

V. Scanning Transmission Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

ADVANCES IN CATALYSIS, VOLUME 48 Copyright q 2004 Elsevier Inc.ISSN: 0360-0564 DOI 10.1016/S0360-0564(04)48002-X All rights reserved

*Corresponding Addresses.

J. M. Thomas and P. L. Gai / Adv. Catal. 48 (2004) 171–227

VI. Recent Advances in Ultra-High Resolution, Low-Voltage Field Emission

Scanning Electron Microscopy and Extreme FESEM in Catalysis . . . . . . . . . . . . . . . . 195

VII. Cathodoluminescence Imaging for Elucidation of Electronic

Structures of Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

VIII. Recent Advances in In Situ Atomic Resolution-Environmental

Transmission Electron Microscopy (ETEM) Under Controlled Reaction Conditions. . . 196

VIII.A. In Situ Investigations of Gas–Solid Reactions and Active Sites . . . . . . . . . . . 196

VIII.B. Illustrative Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

VIII.B.1. In Situ Gas–Catalyst Reactions at the Atomic Level . . . . . . . . . . . 201

VIII.B.2. Atomic-Resolution ETEM of Butane Oxidation . . . . . . . . . . . . . . . 203

VIII.B.3. Atomic-Resolution ETEM of Nanorods . . . . . . . . . . . . . . . . . . . . . 210

VIII.C. Advances in In Situ Wet-Electron Microscopy Technique (Wet-ETEM)

for Probing Solid Catalysts Under Liquid Environments . . . . . . . . . . . . . . . . 210

IX. Environmental Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

X. Electron Tomography: Three-Dimensional Electron Microscopy Imaging . . . . . . . . . . 212

X.A. The Topography and Location of Nanoparticles in Supported

Catalysts; BSE and HAADF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

X.B. Pinpointing the Location of Nanoparticles Supported on Nanoporous Solids. . . . 218

XI. Energy Filtered Transmission Electron Microscopy and Elemental Maps of

Solid Catalysts Using EFTEM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

XII. Other Significant Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

XIII. Critical Evaluations of the Methods and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . 220

XIV. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

No other method rivals electron microscopy (EM) in the wealth of structural (atomic,

nanoscopic, microscopic, and mesoscopic), topographic, and electronic information that

it provides in the characterization of solid catalysts such as those used commercially, for

laboratory trials or model studies: EM provides deep insights into the structure of solid

catalysts—their precursors, active sites, and expired or regenerated forms—as well as vital clues

to their mode of operation. In some important instances it serves as the only trustworthy means of

determining the structure and composition of a catalyst. After a brief update on the significance

of recent advances in EM techniques which allow (i) the probing of catalysts at atomic

resolution, (ii) electron crystallography, and (iii) the determination of the chemical compositions

of catalysts, we illustrate these achievements with specific examples. These include (a) pin-

pointing the location and topography of nanoparticle catalysts; (b) constructing elemental maps

(and compositional distributions) of solid catalysts; (c) in situ investigations of active sites and

reaction processes at the atomic level; (d) elucidating the nature of intergrowths (coherent,

recurrent, and random) of closely similar structures within a supposed new catalyst; (e) identi-

fying atoms (or small groups of atoms) of high atomic number supported on high-area solids;

and (f) characterizing nanoparticles on uneven supports. In (a) and (e) the recently developed

technique of electron tomography plays a crucial role. q 2004 Elsevier Inc.

Abbreviations

A absorption

a unit cell dimension of crystal along a-axis

ADF annular dark field


AEM analytical electron microscopy

A angstrom units

b unit cell dimension of crystal along b-axis

BET Brunauer, Emmett, Teller surface area

b angle between a- and c-axes in the crystal unit cell

BF bright field

c unit cell dimension of crystal along c-axis

CA concentration of element A in a compound AB

CB concentration of element B in a compound AB

CBDP convergent beam electron diffraction

CCD charge-coupled device

Cs coefficient of spherical aberration of the electron microscope objective lens

CTF contrast transfer function

Df objective lens defocus value

Df ðSÞ Scherzer defocus value

e electron charge

E electron energy

ECELL environmental cell

ED electron diffraction

EDX energy dispersive X-ray spectroscopy

EELS electron energy loss spectroscopy

EFTEM energy-filtered transmission electron microscopy

ELNES electron energy loss near-edge structure

EM electron microscopy (or microscope)

EPMA electron probe microanalysis

ESEM environmental scanning electron microscopy

ETEM environmental-TEM

EXELFS extended energy loss fine structure

F fluorescence

FESEM field emission scanning EM

FðuÞ electron envelope function

FT Fourier transform

g gram

GIF Gatan imaging filter

L Green’s function

HAADF high-angle ADF

HREM or HRTEM, high-resolution TEM

HRSTEM high-resolution scanning TEM

HVEM high-voltage EM

IA background-subtracted peak intensity of element A

IB background-subtracted peak intensity of element B

Iðx; yÞ image intensity of sample in the image plane, k, rate constant

KAB sensitivity factor in analysis for elements A and B in compound AB

kcal kilo calories

l wavelength of electrons

LVSEM low-voltage SEM

m electron mass

MA maleic anhydride

m meter

mm micrometers

J. M. Thomas and P. L. Gai / Adv. Catal. 48 (2004) 171–227 173

mbar millibar

mm millimeters

mrad milliradians

mol mole

nm nanometers

PEELS parallel EELS

f0 amplitude of electron wave incident on sample

fg amplitude of scattered electron wave

SEM scanning EM

s interaction constant

sinðxÞ contrast transfer function (CTF)

Cðx; yÞ electron wave function at exit face of sample, with incident electrons

along z-direction

CðrÞ electron wave function at the spatial coordinate, r

SMSI strong metal–support interactions

STM scanning tunneling microscopy

t sample thickness

ti time

TEM transmission electron microscopy

u scattering angle of electrons in radians

XAFS X-ray absorption fine structure


XRE X-ray emission

V volume of crystal unit cell

Vðx; yÞ thickness-projected crystal potential

VðrÞ crystal potential at the spatial coordinate, r

WDS wavelength dispersive X-ray spectroscopy

WPO weak phase object

Z atomic number

I. Introduction

Most commercial catalysts are powdered solids that consist of one or two distinct

phases (or polyphasic aggregates) or of supported metallic components on high-

area supports of a quite different composition (such as oxides, chacolgenides or

halides). Table I is a list of elements present in typical catalysts. A wide range

of techniques has been developed (1–4) to characterize the composition and

structure of surfaces of model catalysts, such as single-crystals of metals, alloys

or oxides. These techniques include low-energy electron diffraction (ED), sum-

frequency generation, and polarized reflection–absorption infrared spectroscopy

and others that are usually inapplicable in the characterization of commercial

catalysts and of no value in determining structural, electronic, or compositional

information for functioning catalysts.

Insofar as most solid catalysts are concerned, characterization entails, inter alia,

the determination of surface composition; the number and nature of distinct


crystallographic phases; electronic properties of the catalyst (encompassing such

information as the oxidation states of particular atoms, especially those at active

sites) and their coordination to surrounding atoms; the location of active sites;

reaction mechanisms; the mode of release of structural oxygen; and accommo-

dation of the catalyst non-stoichiometry (3,4). In the growing field of nanoporous

solids (used as catalysts or catalyst supports), the atomic structure of the

framework (5), as well as the nature of its nanoporosity, needs to be determined.

For the elucidation of these properties, electron microscopy (EM), used in one

or more of its many modern variants — high-resolution (real-space) imaging, or

as a means of effecting electron crystallography, or as a powerful scanning probe

instrument, or as an electron energy loss spectroscopic (EELS) tool — is of

unrivalled value. No other single tool yields such a wealth of diverse information

concerning solid catalysts and their surfaces. The sophistication, reliability, and

ease of operation of electron microscopes have increased enormously since their

early applications, which included channelling of metallic particles across the

surfaces of graphite (6), and a range of physico chemical problems have been

solved (7–10).

In contrast, mass spectrometers, for example, are very powerful tools, but the

information they yield is largely compositional. Likewise, laser-based spectro-

scopic tools (such as laser-induced Raman or infrared (IR) spectroscopy) yield

insights that are largely related to bonding and site environment. Scanning probe

methods, especially STM, provide great detail and high resolution concerning

atomic arrangements at surfaces (even under in situ conditions), but they yield

essentially no information about atomic composition and diffraction.

In addition to the information enumerated above that is important in the

characterization of catalysts, we also require as much knowledge as possible

TABLE I

A selection of typical commercial and viable new solid catalysts

Elements present in the catalyst Process catalyzed

Fe, K (Al, Si, O) Synthesis of ammonia

Mo (W), S, Co (Ni) Hydrodesulfurization

V, P, O Selective oxidation of butane

Co (Mn), Al, P, O Oxyfunctionalization of alkanes

La, Pt, Al, Si, O Cracking of hydrocarbons

Pt, Re (Ir), Al (Si), O Naphtha reforming

Ti, Si, O Alkene epoxidation

Si (P), Mo (W), O, Cs (Na), Co, Al, P, O Dehydration of alkanols


about the electronic states of individual atoms, the electronic (band) structure of

the solid and — for specific active sites in, say, oxide catalysts — the statics and

dynamics of bonding of the atoms that constitute these sites. So far as the last

named desideratum is concerned, X-ray absorption fine structure (XAFS) is the

prime technique of choice (11–13). But such is the progress that has recently

been made (12,14) in electron energy loss near-edge structure (ELNES) analysis

using electron microscopes equipped with the appropriate electron spectrometers

that there are real prospects for retrieval of information equivalent to that which

XAFS (15) yields from micro- and nano-regions of a catalyst, in EM studies.

In the following sections, we summarize some of the most advanced and novel

EM methods that are playing pivotal roles in the understanding of solid catalysts.

We then proceed to demonstrate the veracity of the claims made above about the

unique power of EM in catalyst characterization. The reader is also directed to a

series of up-to-date authoritative reviews pertaining to EM and catalysis

contained in Ref. (16). In particular, there are comprehensive reviews of energy-

filtered TEM (EFTEM), which has advantages in constructing element-image

maps of specimens under consideration, including solids of catalytic interest such

as carbon nanotubes (17) and the development of in situ atomic resolution-ETEM

for direct probing of dynamic catalytic reactions at the atomic scale (18).

II. Electron Microscopy (EM) Methods

The use of EM (except in the special case of SEM) demands that the catalyst,

whether mono-or multi-phasic, be thin enough to be electron transparent. But, as

we show below, this seemingly severe condition by no means restricts its

applicability to the study of metals, alloys, oxides, sulfides, halides, carbons, and

a wide variety of other materials. Most catalyst powder preparations and

supported metallic catalysts, provided that representative thin regions are

selected for characterization, are found to be electron transparent and thus

amenable to study by EM without the need for further sample preparation.

In recent years, increasing use has been made of in situ methods in EM—as is

true of other techniques of catalyst characterization such as IR, Raman, and NMR

spectroscopy, or X-ray diffraction. Although the low mean-free path of electrons

prevents EM from being used when model catalysts are exposed to pressures

comparable to those prevailing in industrial processes, Gai and Boyes (4)

reported early investigations of in situ EM with atomic resolution under

controlled reaction conditions to probe the dynamics of catalytic reactions. Direct

in situ investigation permits extrapolation to conditions under which practical

catalysts operate, as described in Section VIII.

Most applications of EM to catalysis take advantage of high-resolution

transmission EM (HRTEM) instruments, and the structures of an ever-increasing


number of molecular sieve catalysts have been determined by HRTEM. Scanning

transmission EM (STEM) instruments, however, as well as sophisticated variants

of conventional SEM, which are ideal for determining both the morphology and

the composition of exterior layers of solid catalysts in a spatially highly resolved

fashion, plays a significant role in the characterization of catalysts and related

materials such as precursor gels or supports.

The modern-day analytical EM (AEM) is capable of achieving a multiplicity

of functions: information pertaining to structure and/or phase purity comes via

ED patterns and real-space images; composition on the other hand emerges from

the electron-stimulated X-ray emission (XRE) peaks or from EELS. And because

of advances in the technology of energy dispersive detectors for XRE spectra

and in parallel processing of EEL spectra, commercial EMs are now routinely

equipped with those two powerful analytical capabilities. They are also equipped

with more sensitive means of recording, digitizing (and processing by, for

example, Fourier transform (FT) and various filtering procedures) transmission

images of a sample.

One of the significant instrumental advances has been in the field of detection

and recording of diffracted or focused electrons. One of the difficulties is the

occurrence of electron-beam damage in EMs (19), and low electron-dose imaging

methods are required to eliminate it. The traditional electron microscope quality

film first gave way to TV recording (with an improved sensitivity and a slightly

inferior dynamic range). But then came the image plate (IP) and the slow scan

charge-coupled device (CCD) (20), each possessing very high sensitivities

(2 £ 10214 and 5 £ 10214 C cm22) and dynamic ranges of 4.0 and 2.5 orders of

magnitude, respectively. This revolutionary improvement in detection now makes

it possible to deploy novel electron crystallography (21,22) to solve the crystal

structures of microscopic samples such as siliceous mesoporous materials in a

manner analogous to conventional X-ray crystallography using direct methods.

The key difference, however, apart from the inability of the X-ray crystallography

to cope with the minute specimens now solvable by electron crystallography, is

that, with so-called mesoporous (open-structure) solids such as the SSZ-48 silica

family, it is in principle impossible to determine the details of the pore topology

(when the pore diameters are in the range 1–20 nm) using X-ray crystallography,

when the framework silica that constitutes the filigree arrangement of pores is

structurally disordered (as demonstrated by solid-state NMR).

II.A. Electron Microscopy in Catalysis

Traditional approaches to explore catalysts are generally based on indirect chemical

and spectroscopic methods. Constructions of structural or mechanistic models of

reactions on the surfaces of complex catalysts based on such methods often provide


incomplete or inadequate pictures of the processes involved. EM is providing

important insights into changes in the atomic structure and chemistry of reactions

that profoundly influence catalytic properties. These have prompted the develop-

ment of new catalytic materials, the solution of complex structures, and also the

optimization of catalytic properties by delicate control of the solid structures.

In this chapter, we outline some of the most significant recent developments in

EM methods, including in situ EM techniques for probing catalysis and active

sites at the atomic level, the imaging conditions required to obtain the local fine

structure, and the chemistry of the catalysts. We also briefly discuss limitations

and future trends.

II.B. Imaging in the Electron Microscope

Electrons undergo scattering as a result of the beam-sample interactions. An

essential feature of EM is diffraction. Crystals (samples) diffract electrons

according to Bragg’s law. The diffraction pattern thus formed may be regarded as

the FT of the crystal, and hence an inverse FT in the objective lens forms the image.

With high-energy electrons ($100 kV) incident on a sample, a number of signals

are emitted, which can be used for structural and chemical analyses (Fig. 1). These

signals result from elastically and inelastically scattered transmitted electrons,

characteristic X-rays, and back-scattered and secondary electrons (SEs).

In the operation of a conventional transmission electron microscope (CTEM),

the electron beam generated by a filament passes through a condenser lens

system, and the collimated beam is then incident upon the sample. Scattered rays

from the same point are brought to focus in the image formed by the objective

lens (Fig. 2a). The associated signals are illustrated schematically in Fig. 2b. The

characteristics of the objective lens (its spherical aberration coefficient, Cs; and

the accelerating voltage (wavelength of electrons, l) determine the image

resolution. Parallel electron beams interfere in the back focal plane (bfp) of the

objective lens to form a diffraction pattern. The information in the image is

present in the diffraction pattern originating from the same region of the sample.

The relationship between the image and the diffraction pattern is that of direct

(real) and reciprocal space. These are mutually complementary in the

interpretation of structural characteristics of the sample. The intermediate lens

can bring into focus either the image or the electron diffraction (ED) pattern

(through a change in its focal length) onto the focal plane of the projector lens

system, which magnifies the image on the screen. The point (or interpretable)

resolution, d; depends on the wavelength of the electron beam, l; and the

spherical aberration coefficient Cs and is given by a simple relationship:

d , 0:64C1=4s l3=4 ð1Þ


To improve the resolution, one, therefore, minimizes Cs (with aberration

correction almost to zero) and increases the electron energy. However, electron

energy spread and stability issues are also critical as the resolution is improved.

II.C. TEM Imaging Method Using Diffraction Contrast

In TEM diffraction contrast imaging, the Bragg condition is satisfied for a single

diffracted beam (23). The interpretable resolution depends on the size of the

objective aperture (i.e., it is diffraction limited) and can be of the order of 1 nm. If

the objective aperture includes only the diffracted beam corresponding to the

incident electron beam direction (primary beam containing the direct transmitted

electrons), a bright-field (BF) image is obtained. The contrast is produced as a

consequence of differences in electron intensities scattered into Bragg reflections

from different areas of a thin sample. In imaging, if only scattered electrons are

included, a dark-field (DF) image is formed.

Fig. 1. Schematic of the information from elastically and inelastically scattered electrons during

the electron beam–sample interactions.


The diffraction contrast technique is very useful in determining the nature of

defects or lattice imperfections of catalysts. The technique can be used to analyze

dislocations in catalysts by determining components of their displacement vector

(called the Burgers or shear vector, b) in the three crystallographic dimensions

and to define the three-dimensional geometry of defects. (In the HRTEM method

described below, which gives a planar image, calculations may be necessary to

ascertain the component of the displacement vector of the defect normal to the

plane of projection.) Defects such as dislocations play a key role in governing

the properties of catalysts, and understanding their nature is critical in the

optimization of catalytic properties.

There are established criteria for obtaining b by using diffraction contrast (23).

Briefly, the dislocation intensity (contrast) is mapped in several Bragg reflections

(denoted by vector, g) by tilting the crystal to different reflections and deter-

mining the dot product of the vectors g and b (called the g·b product analysis).

Fig. 2. (a) Ray diagram in the electron microscope under imaging (microscopy) conditions. E:

electron source; C: condenser lens; S: sample; O: objective lens; bfp: back focal plane of O; I:

intermediate lens; P: projector lens. (b) Structural imaging, diffraction and compositional

functionalities of TEM.


The reflections include a particular g in which the dislocation is invisible

(i.e., g·b ¼ 0 when b is normal to the reflecting plane). With these criteria in

diffraction contrast, one can determine the character of the defect, e.g., screw

(where b is parallel to the screw dislocation line or axis), edge (with b normal to

the line), or partial (incomplete) dislocations. The dislocations are termed screw

or edge, because in the former the displacement vector forms a helix and in

the latter the circuit around the dislocation exhibits its most characteristic feature,

the half-plane edge. By definition, a partial dislocation has a stacking fault on

one side of it, and the fault is terminated by the dislocation (23–25). The nature

of dislocations is important in understanding how defects form and grow at a

catalyst surface, as well as their critical role in catalysis (3,4).

We now briefly review some theoretical aspects of transmission ED using

high-energy electrons based on an electron wave mechanical formulation of the

dynamical theory of contrast.

II.D. Theoretical Procedures

The steady-state wave function CðrÞ describing electrons with energy E moving

in a crystal potential VðrÞ obeys the Schrodinger equation:

72CðrÞ þ 8p2meðE þ VðrÞÞCðrÞ ¼ 0 ð2Þ

Where m and e are the electronic mass and charge, respectively, and r is the spatial

coordinate (23).

To interpret electron micrographs and diffraction patterns, it is essential

to understand electron scattering mechanisms occurring through the crystal. In

kinematical theory of ED contrast, the amplitude of a scattered electron wave

ðfgÞ is a small fraction of the amplitude of the incident wave ðf0Þ and the

kinematical theory is valid only for thin crystals. In thicker crystals, kinematical

theory breaks down because of multiple scattering effects, and the dynamical

theory incorporating Bloch wave functions should be used instead. Intensity

(contrast) calculations for specific defects located at a particular depth in a crystal

of thickness t can be performed by using the two-beam approximation in the

dynamical theory of ED, or more accurately by using the many-beam theory for

thicker crystals, with the inclusion of absorption effects.

III. High-Resolution Transmission Electron Microscopy

One of the most powerful methods of direct structural analysis of solids is

provided by HRTEM, whereby two or more Bragg reflections are used for

imaging. Following Menter’s first images of crystal lattice periodicity (26) and


the early theoretical work by Cowley and Moodie (27), the power of

experimental HRTEM in the experimental determination of real-space structures

of complex inorganic solids that were not amenable to conventional techniques of

structure determination (e.g., X-ray and neutron methods) was elucidated by

Cowley and Iijima (28), Anderson (29), and Thomas (30). In contrast to

conventional diffraction techniques, HRTEM provides localized real-space

information—at the atomic level—concerning the bulk and surface properties of

solids, as well as the corresponding chemical information and the ED information

in reciprocal space.

Because atomic scattering amplitudes for electrons are approximately 104–105

times as large as they are for X-rays and neutrons, it follows that, with electrons

as probes, structural information may be obtained from single crystals of almost

nanoscale dimensions. To illustrate this point, we note that the best attainable

X-ray performance (with synchrotron sources) requires crystal dimensions of

2 £ 2 £ 2 mm3. Because of the strong interactions between the electron beam and

atoms in a sample, only some 104 unit cells of sample (corresponding to masses

of sample as little as 10218 g) are required to yield significant HRTEM images

and diffraction patterns.

In HRTEM, very thin samples can be treated as weak-phase objects (WPOs)

whereby the image intensity can be correlated with the projected electrostatic

potential of crystals, leading to atomic structural information. Furthermore, the

detection of electron-stimulated XRE in the electron microscope (energy

dispersive X-ray spectroscopy, or EDX, discussed in the following sections)

permits simultaneous determination of chemical compositions of catalysts to the

sub-nanometer level. Both the surface and bulk structures of catalysts can be

investigated.

The micrograph or the image obtained on an EM screen, photographic film,

or (more commonly today) a CCD is the result of two processes: the interaction

of the incident electron wave function with the crystal potential and the

interaction of this resulting wave function with the EM parameters which

incorporate lens aberrations. In the wave theory of electrons, during the

propagation of electrons through the sample, the incident wave function is

modulated by its interaction with the sample, and the structural information is

transferred to the wave function, which is then further modified by the transfer

function of the EM.

III.A. Conditions Required for Optimizing HRTEM Images

The HRTEM requires samples that are electron transparent (normally, a few

tens of nanometers in thickness). As described in preceding section, during

the interaction of the electron beam with a crystal specimen, electrons are


scattered by the interactions with the inner potential of the crystal. The

objective lens of a microscope serves as a kind of Fourier transformer. The

diffraction pattern formed at the bfp of the objective lens is further Fourier

transformed to yield the image of the specimen. The theory of HRTEM tells

us that, because the objective lens is imperfect (being characterized by size,

aberration (spherical and chromatic), and defocus effects, some fundamental

information about the specimen structure is lost. Electron-sample interactions

result in phase and amplitude changes in the electron wave. The contrast of

images in HRTEM (for example, in atomic-scale imaging) is a result of

phase contrast caused by phase shifts (changes) of diffracted electron beams

by the scattering, in combination with the objective lens effects. Amplitude

changes are small.

For a thin enough crystal, the WPO approximation is used, which is based on

the assumption that the electron wave is modulated only in phase (phase contrast)

and not in amplitude. The image intensity is then linearly related to the projected

potential distribution of the sample (similar to the charge density) along the

direction of incidence of the electron beam and can be expressed in terms of

the crystal structure. The phase contrast is produced by the phase modulation of

the incident electron wave when it is transmitted through the crystal potential

of the sample. The propagation of a plane electron wave traversing a thin sample

is thus treated as a weak (scattering) phase object. The wave function Cðx; yÞ at

the exit face of a thin sample can be written as follows:

Cðx; yÞ ¼ expðisVðx; yÞÞ ð3Þ

and for a very thin crystal, Eq. (3) can be approximated as

Cðx; yÞ ¼ 1 þ isVðx; yÞ ð4Þ

where Vðx; yÞ is the thickness projected crystal potential and s is the interaction

constant, which is a function of the electron wavelength and energy (31). The

image intensity, Iðx; yÞ at the image plane of the objective lens results from

two-dimensional Fourier synthesis of the diffracted beams (square of the FT of

the waves at the exit face of the crystal), modified by a phase contrast transfer

function factor (CTF or sin x which is dependent on the objective lens

parameters and incident electrons. These are given by Scherzer (32) in Eqs. (5)

and (6) as follows:

Iðx; yÞ , 1 2 2sVðx; yÞ p FTðsin xÞ ð5Þ

Where p is a convolution integral and FT is the Fourier transform. The phase-

contrast imaging performance of an HRTEM is controlled by sin x; which

contains the basic phase-contrast sinusoidal terms modified by an envelope


function, FðuÞ; which is due to the partial coherence of the electron beam:

sin x ¼ FðuÞsin½ð2p=lÞðDfu 2=2 2 Csu4=4Þ� ð6Þ

where u is radial scattering angle, Df is objective lens defocus value, and F

depends on the coherence conditions of the incident beam. CTF is a quantitative

measure of the trustworthiness of the lens in recording a reliable image. Directly

interpretable structure images are recorded near the Scherzer defocus, defined as

Df ðSÞ ¼ 2C1=2s l1=2: At this defocus, the image can be directly related to

the two-dimensional projected potential of the specimen, with dark regions

corresponding to columns of heavier atoms. This is illustrated for a Ge-silicalite

(GeSiO4) in Fig. 3. Beyond the point resolution, calculations to match

experiments are required.

In the following section we discuss the progress in HRTEM instrumentation.

III.B. Development of HRTEM

To improve the point resolution, a number of important home-built instruments

operated at higher voltages (,500–600 keV) were developed during the 1970s.

However, these were in-house, highly specialized instruments that experienced

some difficulties in operation. (Some were built at substantial cost and had

difficulty meeting the theoretical resolution limit specifications, and some lacked

a proper goniometer stage for tilting the samples.)

The breakthrough in wide applications of HRTEM came with the

development of the first state-of-the-art medium-voltage (200 kV) HRTEM by

Fig. 3. HRTEM atomic structure image of germanium silicalite (GeSiO4) in which there are

channels of aperture diameter 0.55 nm running along the [010] direction. Inset shows the 5- and

6-membered smaller apertures that are circumjacent to larger (0.55 nm) channels (5).


Boyes, Gai and coworkers at the University of Oxford (in association with

JEOL Ltd) (33) and by Thomas and coworkers at the University of Cambridge

(also in association with JEOL Ltd) (34,35). The key points of these

developments were that the instrument had a resolution similar to that of the

best home-built HRTEM instruments (,2 A) at a small fraction of the cost, and

it came in a user-friendly package, achieving the full theoretical performance

routinely while fitting in a standard laboratory and requiring no special

buildings. Incremental improvements in resolution (,1.3–1.6 A) were

achieved later with the development of a 400-kV HRTEM (36).

The state-of-the-art HRTEM has achieved very high resolutions of ,1.7–

2.3 A and ,1.3–1.6 A at operating voltages of 200 and 400 keV, respectively,

providing information at the atomic level. New high-voltage (1 MeV) and high-

resolution commercial instruments have also been built, and a point resolution

of ,1 A has been reported (37). Aberration-corrected commercial HRTEM

instruments are becoming available (38).

On-line digital processing techniques are also available to quantify HRTEM

images. Quantification of the HRTEM image interpretation is checked by

matching experimental images with complementary multi-slice image simu-

lations using the n-beam dynamical theory of ED (27,39). Variations in image

detail can be computed as a function of sample thickness, electron wavelength,

and lens characteristics (spherical and chromatic aberrations and focusing

conditions) (3,4,40–42).


Catalysts by HRTEM

As described in the preceding section, there are fundamental and practical

difficulties that require great caution in the interpretation of HRTEM images. The

electron beam-sample interactions lead to multiple scattering (dynamical) effects

that are quite complex but can be simulated. These are especially important in

understanding the structures and shapes of nanoscale catalysts on supports (40).

Furthermore, the image information is limited by electron lens aberrations.

Efforts are in progress to minimize or eliminate corruption of the image by

spherical aberration and chromatic aberration by aberration-free EMs and energy

filtering; these are described in Sections XI and XIII.

III.C.1. L-Type Zeolite Catalysts

A convenient approach in HRTEM is to record a series of images at different

settings of the objective lens defocus and as a function of sample thickness.


A trustworthy result can be obtained by comparing the observed image to

a simulated image, as illustrated in Fig. 4a and b for an L-type (LTL) zeolite

catalyst (5). HRTEM shows that the structure remains unchanged from the

surface (outermost layer) to the bulk. The simulated image (inset) of the surface

of the LTL zeolite—terminating with cancrinite cages which are major structural

components (Fig. 4b)—matches the observed HRTEM image. There is little

evidence of contraction normal to the catalyst surface.

III.C.2. Metal-Substituted Aluminum Phosphate (MAPO-36)

Microporous Catalysts

Microporous catalysts such as MAPO-36 (43,44), which are excellent for

selective oxidation of hydrocarbons (45), are highly beam-sensitive. Yet HRTEM

Fig. 4. (a) HRTEM image of zeolite LTL along the [001] direcction. On the extreme (top) left is a

schematic drawing of the framework of the idealized LTL structure. Next to it is a computed image,

which is almost indistinguishable from the observed HREM image. This comparison demonstrates that

the extent of structural distortion at the surface and immediate sub-surface region of the zeolite is less

than about 5%. The cancrinite cages (Fig. 2(b)) are clearly visible at the outermost surface (side wall)

(5). (b) Schematic diagram of a cancrinite cage, which is a major structural component of zeolite LTL.


yielded images that even show high-symmetry crystallographic directions that

unmistakably reveal (Fig. 5, with the computed image inset) 12-ring channel

systems (similar to those in zeolite LTL (and also ALPO-5), the structure of

which was solved by X-ray diffraction (44)). The results are consistent with

results of gas adsorption measurements. The crystal symmetry and approximate

values of the unit dimensions of the MAPO-36 catalyst were determined by

HRTEM and ED patterns (43). These data provided a plausible structural model;

the resulting simulated XRD pattern closely resembles the experimental pattern

measured at high temperatures. The structure was then refined by the use of

distance least-squared and energy minimization techniques, and excellent

agreement was obtained between the experimental and simulated XRD patterns

at both high and low temperatures (44).

III.C.3. High-Silica Microporous SSZ-48 Catalysts

ED intensity data collected by using a HRTEM and CCD detector reveal

a monoclinic crystal structure having the following unit-cell dimensions:

Fig. 5. HRTEM image of MAPO-36 showing well-defined large apertures. The inset shows the

computed image where the outline of the 12-rings is clearly visible (top and bottom are with and

without taking beam damage into account) (5,43).


a ¼ 11:19 �A; b ¼ 4:99 �A; c ¼ 13:65 �A; and b ¼ 100:78 ðV ¼ 748:6 �A3Þ:Reflections with normalized structure factors between 0.65 and 10 were used

in the structural solution by the direct methods (5). The phases obtained were

used to generate a three-dimensional potential map that readily revealed the

seven tetrahedrally coordinated silicon atoms in the asymmetric unit and five of

the 14 oxygen atoms. The resulting structure is shown in Fig. 6.

III.C.4. Intergrowths in Zeolite Catalysts: Coherent, Recurrent, and Random

One of the earliest direct bonuses of imaging zeolitic catalysts by HRTEM

was the discovery (10) that the nominally phase-pure ZSM-5 (structure code

MFI) contained sub-unit-cell coherent intergrowths of ZSM-11 (MEL). It soon

became apparent (46) that, depending on the mode of synthesis of these

and other pentasil (zeolitic) catalysts, some specimens of ZSM-5 contained

recurrent (regular) intergrowths of ZSM-11. It also emerged that intergrowths

of offretite and erionite are features of both nominally phase-pure erionite and

of pure offretite and of many members of the so-called ABC-6 family of

zeolites (47).

Fig. 6. Structural model of the SSZ-48 crystal structure, showing the projected positions of the

organic template within the pores, of SSZ-48 (5).


All this served as a prelude to the clarifying work (48–50) that showed that

faujasite (FAU) and its hexagonal analogue (EMT) (Fig. 7) exhibit a strong

tendency to form coherent intergrowths (5). And so, by HREM direct imaging,

many hitherto puzzling problems concerning the structure of zeolitic catalysts

were unambiguously resolved. For example, some zeolites claimed to be new on

the basis of powder X-ray diffractograms (and usually published in the patent

literature) turned out not to be new structures but rather intergrowths (of various

kinds) of FAU (cubic) and EMT (hexagonal), as revealed by HRTEM (Fig. 8).

Fig. 7. Diagram illustrating the building units and structural relationship between the FAU and

EMT frameworks. (a) Two (111) layers type K (A) and L (B) in twin orientation. (b) Hexagonal [100],

(or cubic [1 2 1 0]) views of A and B. (c) Cubic FAU framework occurs when only type A or B stack.

(d) Hexagonal EMT framework which occurs when A and B stack in alternation.


The stacking is shown in Fig. 9. This is the situation pertaining to ZSM-3, ZSM-

20, and ECR-30, for example.

High-resolution EM also showed that the synthetic zeolitic catalyst ZSM-23

(MTT) is a recurrently twinned version of the synthetic zeolite theta-1 (TON)

(51). It is noteworthy that the elucidation of the structures of zeolite beta, for a

long time an enigma and problematic for X-ray crystallographers, came only

through the application of HRTEM (50).

Fig. 8. HRTEM image of FAU/EMT intergrowths viewed along the [110] direction. The stackings

ABC… and AB… correspond to the FAU and EMT end-member structures, respectively (5).

Fig. 9. Schematic of FAU and EMT intergrowth structures.


IV. Chemical Composition Analysis with the

Analytical Electron Microscope

Chemists who study solids are aware of the fact that microstructures of solids

profoundly influence and control their properties. AEM at high resolution permits

both the analysis of the elemental composition of a solid and its structure under

high-resolution conditions. HRTEM (with high spatial resolution microdiffrac-

tion) provides high-resolution structure images, including structural defects

such as dislocations or internal boundaries, in parallel with direct experimental

measurements of local chemical composition from small areas—especially for

heterogeneous solids (52–56). Microcomposition analysis in the EM using

electron-stimulated characteristic X-rays is a well-known technique, and EFTEM

serves a very similar purpose.

EDX, in which X-ray intensities are measured as a function of the X-ray

energy, is the common method for chemical composition analysis in the electron

microscope. In EDX, interaction of a beam of high-energy electrons with an

inner-shell electron of the sample atom results in the ejection of a bound inner-

shell electron from the attractive field of the nucleus in the sample atom, leaving

the atom in an excited state with an electron shell vacancy. De-excitation by

transition from an outer shell involving a change in the energy state of the atom

between sharply defined levels produces X-rays (or Auger electrons),

characteristic of elements in the sample.

Stoichiometric variations in compositions of a material and of surface layers

can be revealed by AEM. Because a relatively small amount of scattering occurs

through a thin HRTEM specimen, X-rays are generated from a volume that is

considerably less than in the case of electron microprobe analysis (EPMA). For

quantitative microanalysis, a ratio method for thin crystals (57) is used, given by

the equation:

CA=CB ¼ KABIA=IB ð7Þ

where CA and CB are the concentrations of the elements A and B and IA and IB are

the background-subtracted peak intensities for A and B, respectively; typically, a

few dozen crystals are analyzed. The sensitivity factor KAB is determined by

using appropriate standards. For bulk materials, more complex correction

procedures are required and account is taken of the atomic number ðZÞ; absorption

ðAÞ and X-ray fluorescence ðFÞ: Thus, AEM provides real-space imaging and

crystallographic and microcompositional information on a very fine scale.

Furthermore, AEM can be used to obtain partial occupancies of cation sites (and,

under some conditions, anion sites). In cases for which elemental peaks overlap,

wavelength dispersive X-ray spectroscopy (WDS) may be used to advantage.


Spatial mapping of the distribution of particular elements in catalysts typified

by those listed in Table I is readily conducted by EM. Chemical variations of

entire crystals in a sample can be obtained by analyzing X-ray intensities from

elements across a line or over an area in the sample. The latter (two-dimensional

scanning) is known as X-ray elemental mapping. Elemental maps recorded in an

analytical HRTEM from MAPO catalysts (e.g., Zn–aluminum phosphate, Fig. 10)

indicate a uniform distribution of the elements. Similarly, Fig. 11 shows an X-ray

elemental map for GeSiO4 silicalite (Fig. 3), indicating a uniform distribution of

Ge and Si. Quantification of intensities in X-ray maps can provide relative

amounts of the elements (but care is required when peak-overlaps occur).

Examples of elemental mapping of transition metal ion distributions in

framework-substituted ALPO catalysts determined by EFTEM are described in

Section X.

Fig. 10. X-ray elemental map in the electron microscope of metal-substituted aluminophosphate

(MAPO-36 (with M ¼ Zn)) catalyst. The map shows a uniform distribution of the elements in the

sample.


V. Scanning Transmission Electron Microscopy

Crewe et al. (58) pioneered STEM as a structural and an analytical tool. STEM,

which is capable of acquiring signals that are difficult to obtain by other methods,

is essentially a combination of SEM and TEM. In STEM, electrons are focused

on a spot with a diameter less than 0.8–1 nm by a “probe-forming” lens (Fig. 12).

The STEM detector collects scattered electrons and generates picture points by

scanning the focused electron spot on the sample via a pair of deflection

coils, and the resulting signal variation constitutes the image. Noteworthy is

the excellent microanalytical capability of high-resolution STEM (HRSTEM)

Fig. 11. X-ray elemental map in the electron microscope of GeSiO4 catalyst (shown in Fig. 3). The

map illustrates a uniform distribution of Ge and Si.

Fig. 12. Schematic of the information from HRSTEM, DF, and high-angle annular dark-field

(HAADF) microscopy.


(including modern TEM/STEM instruments) equipped with a field-emission gun

(FEG-STEM), especially in the context of catalyst characterization. By use of

sub-nanometer electron probes with high-electron currents, chemical analyses of

catalysts (in addition to high-resolution imaging and element mapping) may be

effected at the sub-attogram (10218 g) level.

An important aspect of HRSTEM is Z-contrast (or atomic number) imaging. It

exploits the fact that electrons scattered at high angles (.30 mrad) obey

Rutherford’s scattering law; the scattering cross-section is proportional to Z2;where Z is the atomic number. Moreover, the scattered electron wave is

predominantly incoherent, so that images formed by using a high-angle annular

dark-field detector (HAADF) (or “Rutherford” detector) do not show the

complicating contrast changes associated with coherent scattering, as occurs in

BF images (formed from Bragg-scattered electrons). HAADF images are directly

interpretable, and the technique is tailor-made for detecting clusters of

catalytically active metals such as Pt, Pd, or Ru clusters (including bimetallics)

on light supports such as zeolites. Isolated atoms or small cluster of heavy atoms

(such as Pt) have been clearly identified by HRSTEM (59) as shown in Fig. 13.

Since Crewe’s work, there have been significant advances in Z-contrast

imaging, following suggestions by Howie and coworkers (60–62). For example,

Treacy et al. (61) and Pennycook et al. (62) imaged very small particles in

catalysts by using high-angle Rutherford scattering contrast. Using the HAADF

imaging technique in the STEM, low concentrations of dopants (,1 at.%) in

semiconductors and in zeolites have been demonstrated (63). Other spectroscopies

Fig. 13. Uniform bifunctional platinum-loaded zeolite catalyst. Large white dots (Pt) are ,0.5 nm

in diameter.


are also possible in the STEM. In Section X, we illustrate recent advances in

three-dimensional mapping of nanocatalysts using HAADF.

VI. Recent Advances in Ultra-High Resolution, Low-Voltage

Field Emission Scanning Electron Microscopy and

Extreme FESEM in Catalysis

A new ultra-high-resolution low-voltage field emission scanning electron

microscopy (HR-LV FESEM) instrument with a 0.5-nm probe at 30 kV (and

2.5 nm at 1 kV), integrated with high-sensitivity EDX, was designed by Boyes

(64) to explore high-resolution imaging and chemical microanalysis in

reflection from bulk samples. The instrument is equipped with an optimized

high-brightness cold-FEG, combined with a very low-aberration condenser

objective final lens. The low voltages allow investigations of uncoated, non-

conducting samples (e.g., ceramic catalyst supports). Low-voltage electron

probes (,5 kV) generally yield inherently better SE images, making

HRLVSEM a powerful tool in catalytic science. This advance is particularly

important because it has made possible high-resolution surface analysis from

bulk catalyst samples, and the resolution gap that previously existed between

the SEM and many of the STEM applications has been bridged. Furthermore, a

novel approach to FESEM design by Boyes (65,66) integrates new levels of

low-voltage image resolution (,1 nm at 1 keV) with greatly improved

sensitivity for EDX elemental microanalysis; chemical imaging at new levels

of spatial resolution down to ,100 nm; and, in favorable cases, resolution

limits of 1–10 nm, while retaining the advances of robust and representative

bulk samples (mm in extent). These powerful capabilities are markedly

improving our understanding of catalysts (4).

VII. Cathodoluminescence Imaging for Elucidation of Electronic

Structures of Catalysts

Cathodoluminescence imaging uses photons emitted from a sample area

irradiated by a scanning electron beam for understanding point defect

concentrations and promoter distributions in working catalysts (67). When an

energetic electron scatters inelastically, an electron from the (filled) valence band

can be promoted to the (empty) conduction band, creating an electron/hole pair.

On recombination, the excess energy is released as a photon, the wavelength of

which is well defined by the band-gap transition. The cathodoluminescence

technique is powerful for determining the local electronic structures of catalysts.


It is diagnostic of electronic/chemical state, is sensitive to point defects, and can

be used to probe the distribution of promoters in catalytic oxides (67). Examples

include effects of the distribution of antimony in Sb–SnO2 catalysts (used for

selective hydrocarbon oxidation) on the electronic structure of the catalyst and

mapping of point defects in titania catalysts.

VIII. Recent Advances in In Situ Atomic Resolution-

Environmental Transmission Electron Microscopy (ETEM)

Under Controlled Reaction Conditions

VIII.A. In Situ Investigations of Gas–Solid Reactions and Active Sites

Catalysis is a dynamic process, and deeper insights into its phenomenology are

extractable from in situ measurements than from characterizations of catalysts

before and after catalysis. A number of notable in situ experiments have relied on

modifications of standard TEM operations under vacuum. The main functions of

the EM depend on a high-vacuum environment, and the pressure in a TEM is

usually of the order of 1027–1026 mbar. Because the influence of the reaction

environment on the structure and activity of a catalyst is critical (3), the high-

vacuum environment of a conventional EM is inappropriate for investigating a

catalytic reaction, as are characterizations of catalysts in post-reaction

environments (e.g., when the catalyst has been taken out of the reaction

environment and cooled to room temperature).

With the gas reaction cell or an environmental cell (ECELL), controlled

chemically reducing atmospheres and oxidizing atmospheres can be maintained

in the EM, and a wide range of gases and vapors can be used. The development of

the methods is described in the following sections.

Early in situ ETEM experiments have been well documented by Hashimoto

et al. (68), Swann and Tighe (69), and Butler and Hale (70). In the development

of ECELLs, window cells have been used to contain gases, solvent vapors, and

hydrated samples (71,72). However, these cells present problems in reliably

sustaining a large pressure difference across a window that is thin enough to

permit electron penetration. Generally, window cells are not suitable for heating

systems. Below, we describe alternative methods used recently to investigate

gas–solid and solution–solid reactions in the ETEM.

The complications of windows can be avoided by substituting small apertures

above and below the sample to restrict the diffusion of gas molecules while

allowing penetration of the electron beam. Typically, pairs of apertures are added

above and below the sample, with differential pumping lines attached between

them. In the early in situ experimentation, an ECELL system (69) could be

inserted inside the EM column vacuum between the objective lens pole pieces.


The gas reaction chamber and the objective aperture assembly occupied the gap

between the upper and lower objective pole pieces, leading to a gas reservoir

around the sample. Such ECELL systems were a major step forward in scientific

capability, being used by Gai et al. (3,73–78), Doole et al. (79), Crozier et al.

(80), and Goringe et al. (81) to characterize catalysis. Other developments for

catalytic studies include an ex situ reaction chamber attached externally to the

column of a TEM, for example, by Parkinson and White (82) and Colloso-Davila

et al. (83). Reactions were carried out in the ex situ chamber (and not in situ), and

the sample was cooled to room temperature and inserted into the column of the

TEM (without exposure to the atmosphere) under vacuum. Baker et al. (84) used

ETEM at gas pressures of a few mbar with limited resolution, and, in these

experiments, representative higher gas pressures were not employed.

Gai (73) developed in situ high-voltage EM (HVEM) to meet the demands of

realistically high gas pressures and temperatures (up to 1273 K) for catalysis,

performing the first in situ investigations of selective hydrocarbon oxidation

reactions catalyzed by metal oxides at high pressures (,1 bar) and operating

temperatures. The results provided insights into the fundamental role of defects at

the catalyst surface in selective oxidation catalysis. With this system, image

resolution was improved from ,1–2 nm at .100 mbar to 0.5 nm at lower gas

pressures of ,30 mbar. This in situ HVEM development thus laid the foundation

for the development of in situ atomic-resolution ETEM (85–90).

The quest to probe gas molecule–solid catalyst reactions directly at the atomic

level resulted in the pioneering development of in situ atomic resolution-ETEM

by Gai and Boyes (87–90), who demonstrated that catalysis at atomic resolution

was possible under controlled dynamic reaction conditions of gas pressure

of a few mbar and elevated temperatures (91,92). In this development, a new

approach was taken to design the ETEM instrument, which is dedicated to

ECELL operations; the ECELL is permanently mounted and integrated with the

HRTEM. The design is based on a computer-controlled Philips CM30T TEM/

STEM system with a proven high-resolution (crystal lattice imaging) perfor-

mance. Furthermore, the whole EM column, and not just the region around

the sample, was redesigned for the ECELL functionality, and a custom set of

objective lens pole pieces incorporating radial holes was designed for the first

stage of differential pumping (with no deleterious effect on imaging).

In the atomic-resolution ETEM, the alignment and excellent atomic resolution

(0.2 nm) of the microscope were maintained with the ECELL facilities, even with

sample temperatures exceeding 973 K and small amounts of gas (at mbar

pressures) flowing through the ECELL. The relatively large apertures in the cell

provide useful angles of diffraction and allow some convergent beam diffraction

pattern (CBDP) analysis with a dynamic STEM probe. The regular, smaller

objective apertures can be used inside the ECELL for diffraction contrast

experiments to determine the nature of defects.


In the development of Gai and Boyes (87,88,90), the ECELL, atomic-

resolution (HRTEM), STEM, hot stage and PEELS/Gatan imaging filter (GIF)

functionalities were combined in a single instrument. The combination is required

to aid simultaneous dynamic structure and composition of the reactor contents.

ETEM is thus used as a “nanolaboratory” with multi-probe measurements.

Design of novel reactions and nanosynthesis are possible. The structure and

chemistry of dynamic catalysts are revealed by atomic imaging, ED, and

chemical analysis (via PEELS/GIF), while the sample is immersed in controlled

gas atmospheres at the operating temperature. The analysis of oxidation state in

intermediate phases of the reaction and, in principle, EXELFS studies are

possible. In many applications, the size and subsurface location of particles

require the use of the dynamic STEM system (integrated with ETEM), with

complementary methods for chemical and crystallographic analyses.

The basic geometry of the novel atomic resolution-ETEM design is a four-

aperture system, in pairs above and below the sample, but the apertures are now

mounted inside the bores of the objective lens pole pieces (rather than between

them, as in previous designs). Regular microscope apertures are mounted in

bushes in each pole piece. The controlled-environment ECELL volume is the

regular sample chamber of the microscope. Differential pumping between the

apertures is achieved by using molecular drag pumps (MDP) and turbo-molecular

pumps (TMP). This design permits high gas pressures in the ECELL sample

region while maintaining high vacuum in the rest of the ETEM (Fig. 14).

A conventional reactor-type gas manifold system enables the inflow of gases into

the ECELL of ETEM, and a sample hot stage allows samples to be heated. A mass

spectrometer is included for gas analysis.

For dynamic atomic resolution, a few millibars of gas pressure are used in the

ECELL. Higher gas pressures (up to a substantial fraction of a bar) are possible,

but they compromise the resolution (as a consequence of multiple scattering

effects of the electron beam through thicker gas layers). A video system

connected to the ETEM facilitates digital image processing and real-time

recording of dynamic events. The instrument and a schematic of the accessories

are shown in Fig. 15a and b, respectively. In in situ ETEM experiments, very low-

electron dose techniques (with doses well below the threshold for structural

damage) are used. The signal is amplified by a low light-level TV camera. The in

situ data are checked in a parallel blank calibration experiment, with the beam

switched off for this in situ reaction and the sample exposed to the beam only to

record the reaction end point. In situ experiments are then confirmed by

comparisons with data from calibration experiments. The aim is completely non-

invasive characterization under benign conditions. Electronic image shift and

drift compensation help to stabilize high-resolution images for data recording on

film or with real-time digitally processed video. Under carefully simulated


conditions close to those in practical reactors, data from in situ ETEM can be

directly related to structure-activity relationships in technological processes.

Because of the small amounts of solid reactant in the microscope sample, analyses

of reaction products are performed with larger samples in a microreactor operating

under similar conditions, and these are used for microstructural correlations.

Several conditions must be met for successful ETEM investigations. Thin,

electron-transparent samples are necessary—this requirement can usually be met

with most catalyst powders. Ultrahigh-purity heater materials and sample grids

capable of withstanding elevated temperature and gases are required (such as

those made of stainless steel or molybdenum). The complex nature of catalysis

with gas environments and elevated temperatures requires a stable design of the

ETEM instrument to simulate realistic conditions at atomic resolution.

Fig. 14. Schematic of the basic geometry of the aperture system and objective lens pole pieces

incorporating radial holes for differential pumping system in the novel atomic resolution-ETEM

design of Gai and Boyes (85–90) to probe catalysis at the atomic level.


The design of in situ atomic-resolution environmental cell TEM under

controlled reaction conditions pioneered by Gai and Boyes (87,89) has been

adopted by commercial TEM manufacturers, and latter versions of this in situ

instrument have been installed in a number of laboratories. In situ atomic

resolution-ETEM data demonstrated by Gai et al. (85–90) have now been

reproduced by researchers in laboratories using commercial instruments;

examples include investigations of promoted ruthenium and copper catalysts

in various gas environments (93) and detailed investigations of Ziegler–Natta

catalysts (94).

Fig. 15. (a) Novel atomic resolution-ETEM (87) and (b) schematic of various components for

imaging, chemical analysis and diffraction under catalyst operating conditions.


VIII.B. Illustrative Examples

VIII.B.1. In Situ Gas–Catalyst Reactions at the Atomic Level

Nanophase platinum catalysts supported on rutile TiO2 are of technological

interest in environmental pollution control and methane reforming (95). Strong

metal–support interactions of a reacting metal particle can lead to catalyst

deactivation (96). Such phenomena can be examined in atomic-resolution

ETEM. An ETEM investigation of sintering of Pt/TiO2 prepared by impregnation

of TiO2 with a solution of a platinum salt, is shown in Fig. 16; Fig. 16a shows the

catalyst containing finely dispersed platinum on TiO2; Fig. 16b shows in situ

ETEM of dynamic catalyst activation in H2 at 573 K, and Fig. 16c shows the same

particle of plainum (P) under dynamic conditions in H2 at ,723 K. The 0.23-nm

atomic lattice spacings are clearly resolved in the platinum particle (P) in H2

at the elevated temperatures. The dynamic image (Fig. 16c) shows that the

particle is faceted; SMSI deactivation with a growth of an amorphous titanium

oxide monolayer on the particle is observed (indicated at the area marked

Fig. 15. Continued


by a larger arrow), along with the development of nanometer-scale single-crystal

clusters of platinum with ,0.2-nm lattice spacings, without overlayers (indicated

by a smaller arrow in Fig. 16c). The H2 is a key contributor to this process.

The results provide insights into the platinum dispersion, and the role of

Fig. 16. Nanophase Pt/TiO2 catalysts: (a) finely dispersed Pt/TiO2 at room temperature. (b) In situ

dynamic catalyst activation in hydrogen imaged at 3008C. The (111) lattice atomic spacings (0.23 nm)

are clearly resolved in the platinum metal particle (P) under the controlled reaction conditions. (c) The

same particle of platinum (P) imaged at 4508C, also in H2. Catalyst deactivation with growth of the

support oxide monolayer indicated by a larger arrow, and the development of nm-scale single-crystal

clusters of platinum metal (which show no coating as they emerge) with ,0.2-nm lattice spacings

indicated by smaller arrow (87).


temperature and particle size in the strong metal support interactions. A range of

conditions and the dynamic rearrangement of the microstructure can be followed

in each in situ experiment.

In situ dynamic surface structural changes of catalyst particles in response to

variations in gas environments were examined by ETEM by Gai et al. (78,97).

In studies of copper catalysts on alumina, which are of interest for the water gas

shift reaction, bulk diffusion of metal particles through the support in oxygen

atmospheres was shown (78). The discovery of this new catalyst diffusion

process required a radical revision of the understanding of regeneration processes in

catalysis.

Bimetallic (98) and alloy catalysts (97), of interest for hydrogenation reactions,

have been investigated in in situ characterizations of methanol synthesis from CO

and H2 in the presence of novel Cu–Pd alloy catalysts supported on carbon; the

results show surface segregation of palladium on the catalyst particles in CO

atmospheres, but surfaces with equal amounts of copper and palladium when the

atmosphere is H2 (97).

VIII.B.2. Atomic-Resolution ETEM of Butane Oxidation

The selective oxidation of n-butane to give maleic anhydride (MA) catalyzed by

vanadium phosphorus oxides is an important commercial process (99). MA is

subsequently used in catalytic processes to make tetrahydrofurans and agricultural

chemicals. The active phase in the selective butane oxidation catalyst is identified

as vanadyl pyrophosphate, (VO)2P2O7, referred to as VPO. The three-

dimensional structure of orthorhombic VPO, consisting of vanadyl octahedra

and phosphate tetrahedra, is shown in Fig. 17, with a ¼ 1:6594 nm,

b ¼ 0:776 nm, and c ¼ 0:958 nm (100), with (010) as the active plane (99).

Conventional crystallographic notations of round brackets ( ), and triangular

point brackets k l, are used to denote a crystal plane and crystallographic

directions in the VPO structure, respectively. The latter refers to symmetrically

equivalent directions present in a crystal.

In situ ETEM has met the formidable goal of revealing atomic structures of

active sites; a mechanism for the release of catalyst structural oxygen; and

the means for accommodation of anion deficiency in the butane oxidation catalyst

(85,89). In situ ETEM and parallel chemical reactivity tests of calcined and

activated VPO catalysts ((010) face), carried out with a continuous fixed-bed as

well as with a pulse microreactor (101), were performed with the catalyst in butane,

and alternatively in N2, or steam and 1.5% butane in air. Figures 18a and b show

the (010) lattice image of the well-ordered VPO at room temperature and the

corresponding ED, respectively. The structural model is superimposed, with dark

regions corresponding to the heavier atoms. The ED shows some of the Bragg

reflections. Figure 19a and b illustrates a sequence of in situ ETEM images


Fig. 17. Structure of complex (VO)2P2O7 in (010), viewed down the b-axis. Vanadium octahedra

and phosphate tetrahedra link together forming a three-dimensional network. Front (bold) and back

(faint) layers are shown.


with the catalyst in 20% butane in He (5 mbar) at room temperature and at the

operating temperature of ,663 K, respectively. The dynamic surface structural

development (a consequence of the catalyst anion loss) in butane with the

formation of extended defects along the k201l direction is illustrated in Fig. 19b.

The corresponding ED (inset, Fig. 19b) shows streaking along the k201l direction.

The image in Fig. 19b is enlarged in Fig. 19c, showing a dislocated lattice with

terminating lattice planes and the presence of partial dislocations (defects) in (201)

lattice planes. The two partial dislocations, P1 and P2 (arrowed), are close to the

catalyst surface (shown at S in profile, with the projection of the structure along

the electron beam direction), bounding a stacking fault associated with them.

Fig. 18. (a) Atomic structure image of VPO and (b) electron diffraction (ED) at room temperature.


The streaking in the ED provides important evidence of the structural disorder

attributed to the defects in (201) planes. This means that anions in (201) planes,

located between vanadyl octahedra and phosphate tetrahedral, are involved in the

alkane oxidation reaction. The disorder attributed to the catalyst anion loss is

revealed only in (201) lattice planes, thus excluding all other planes in the crystal

structure.

These findings, coupled with the results of detailed diffraction contrast

experiments (85,89), show that the defects are formed by glide shear; the lattice is

Fig. 19. (a) In situ atomic resolution ETEM image of (010) VPO in n-butane at room temperature

with electron diffraction (ED); (201) lattice plane (0.63 nm) spacings and other lattice planes are

resolved (201 reflection is arrowed). (b) In situ direct imaging of dynamic atomic motion of reacting

VPO in n-butane at ,3908C. (c) Enlarged image of (b). The (201) lattice displacements (disturbing

the periodicity) due to the reaction are close to the surface S. The resulting defects P1 and P2 are

formed by novel glide shear and the lattice is not collapsed. The corresponding ED (inset) shows

diffuse streaks along k201l (arrowed) (4).


not collapsed (Fig. 20). The sheared (transformed) structure creates regions of

extended glide plane defects. The defect regions (at P1 and P2 in Fig. 19c) lead to

structural regions akin to metaphosphate (PO3)n groups. The dynamic atomic

studies show that only a few monolayers of the catalyst are involved in butane

catalysis (89).

The results showing disorder along the k201l direction illustrate that in the

catalyst–adsorbate interaction, lattice oxygen loss leads to the formation of

Fig. 20. (a) Active sites observed by in situ atomic-resolution ETEM: structural modification of

VPO in n-butane along k201l indicates the presence of in-plane anion vacancies (active sites in the

butane oxidation) between vanadyl octahedra and phosphate tetrahedra. (b) Projection of (010) VPO

(top) and generation of anion vacancies along k201l in n-butane. V and P are denoted. Bottom: model

of novel glide shear mechanism for butane oxidation catalysis; the atom arrowed (e.g., front layer)

moves to the vacant site leading to the structure shown at the bottom.


coplanar anion vacancies between vanadyl octahedra and phosphate tetrahedra

(Fig. 21a). Extended defects are introduced along the k201l direction. They show

that the release of structural oxygen in the oxidation catalysis is accompanied by

a novel glide shear mechanism in which a few surface layers of the oxide undergo

a structural transformation by glide shear to accommodate the surface misfit

resulting from anion vacancies formed during the reaction (shown schematically

in Fig. 21b). This mechanism explains the release of structural oxygen and the

preservation of active Lewis acid sites at the surface without changes in the bulk

structure of the catalyst.

This mechanism is of fundamental importance in the understanding of solid-

state heterogeneous catalytic oxidation processes. The glide defect regions are

Fig. 21. (a) The nature of the glide shear plane defects in three-dimensional projection and (b) in

one layer of idealized structure, showing the novel glide shear process and the formation of glide shear

plane defects. Filled circles are anion vacancies. (c) Schematic of glide shear. Glide defects

accommodate the misfit at the interface between catalyst surface layers with anion vacancies (filled

circles) and the underlying bulk (85,89).


not readily revealed in XRD because of overlap of the peaks from the defective

regions and the VPO matrix; atomic-resolution ETEM has been crucial to

unraveling the reaction mechanism. The positively charged anion vacancy sites

preserved by glide shear at the catalyst surface can be readily available for alkane

activation (by accepting electrons) and for exchange with gaseous oxygen. In

partial oxidation in 1.5% butane/air, the alkane catalysis and the catalyst

regeneration are possible, as validated by parallel reaction chemistry, shown in

Table II (101). Pseudo-first-order rate constants ðkÞ for the disappearance of

butane were measured with a microreactor and a larger amount of the catalyst

(,1 g) at 633 K. The constants are normalized to T ¼ 633 K assuming an

activation energy of 25 kcal/mol and are shown in the second column of Table II.

By varying the volumetric flow-rate of gas and constant times ðtiÞ; k is obtained

by fitting the reactor data to the classical first-order rate expression (101),

dðbutaneÞ=dti ¼ 2kðbutaneÞ ð8Þ

The conversion of butane is based on the difference in the moles between the feed

and the products. Intrinsic rate constants, shown in the third column of Table II,

are based on BET surface areas (m2/g) measurements (101).

Samples 1 and 2 correspond to VPO treated in steam for 92 and 312 h,

respectively. Samples 3 and 4 are N2-treated and activated base VPO catalysts,

respectively. MA capacities represent the total amount of MA liberated by

reduction in 1.5% butane/N2 at the reaction temperature. Table II shows that the

base and N2-treated catalyst have nearly equal activities in the presence of air in

the reactant stream and continue to operate.

TABLE II

Continuous fixed-bed microreactor measurements

Sample k

(rate)

(s21)

k

(intrinsic)

(g/m2 s)

1.5% butane/air; % selectivity at 1.5% butane/N2;

maleic capacity

(micromol/g catalyst)20%

conversion

40%

conversion

60%

conversion

1 2.75 0.110 78 75 71 2.22

2 3.07 0.134 79 78 73 1.23

3 3.35 77 77 74 6.06

4 3.39 ^ 0.22 0.113 82 80 77 4.95 ^ 0.58

Samples 1–4 correspond to VPO treated in steam for 92, 312 h, in N2 and activated base catalysts,

respectively. k; are pseudo-first-order rate constants for the disappearance of butane. The constants are

measured in a microreactor on a larger amount (,1 g) of catalyst at 633 K. k (intrinsic) are based on

the BET surface area.


The novel glide shear mechanism revealed by ETEM and correlations with

activity (89,101) show that glide shear is a key to effective butane oxidation

catalysis. Investigations of the reduction of other oxide catalysts have also shown

that the glide shear mechanism and temperature regimes where glide shear

operates are beneficial for optimal catalyst performance (3,4). Catalysts can

accommodate anion deficiencies without collapse of the crystal lattice and

continue to operate, lengthening the catalyst life under optimized butane/air

ratios. The work has led to the development of improved catalysts for the butane

oxidation process, by incorporation of promoters to induce selective glide

transformations (89,101). Earlier in situ EM investigations correlated with

reaction chemistry (3,4,52,102–105) have shown that crystallographic shear

plane defects produced by the well-known crystallographic shear mechanism,

which eliminate super-saturation of anion vacancies (resulting from the reduction

of oxides) by shear and lattice collapse, are secondary to catalysis. That is,

crystallographic shear planes are consequences of oxide reactions and not the

origins of catalytic activity (3).

VIII.B.3. Atomic-Resolution ETEM of Nanorods

Nanowires and nanorods with high-aspect ratios have generated interest because

of their potential applications in the next generation of nano and molecular

electronics and in catalysis (106). They are being developed as potential supports

for organic molecules (for applications in molecular electronics) and catalysts.

Investigations of surface atomic structure by HRTEM and ED from single gold

nanorods have provided the first direct evidence of the stabilization of the highly

unstable (110) surface by surfactant molecules of cetyl trimethylammonium

bromide (107). In situ heating experiments in an atomic-resolution ETEM in an

atmosphere of N2 (Fig. 22) demonstrated that the rods are stable at elevated

temperatures (18).

VIII.C. Advances in In Situ Wet-Electron Microscopy Technique

(Wet-ETEM) for Probing Solid Catalysts Under

Liquid Environments

Many hydrogenation and polymerization reactions in the chemical industry are

carried out with liquid-phase reactants. An example is the hydrogenation of

aliphatic dinitriles to produce diamines (108,109), which are subsequently

converted with adipic acid in solution and polymerized to produce linear

polyamides, including nylon 6,6. Recently, the development of wet-environmental

transmission electron microscopy (wet-ETEM) for direct nanoscale probing of


reactions between solid catalysts and reactants in the liquid phase—at reaction

temperatures—has been reported (110). Using a liquid-feed holder with an injector

system (similar to those used in chromatography), it is possible to inject pulses of

the liquid into the ECELL under appropriate gaseous environments. The gas

Fig. 22. ETEM at 1808C in N2, illustrating the stability of gold nanorods, for nanoelectronics and

catalysis applications. Gold atomic layers and surface atomic structures are visible. Surface of gold

nanorod at room temperature showing twin defect lamellae on the atomic scale. They indicate

interaction of the surfactant with the (110) surface forming twins to accommodate the shape misfit

between the two.


manifold of the ETEM allows the flow of gases in the ECELL, and catalytic

hydrogenation and polymerization reactions can be followed at operating

temperatures. The wet-ETEM has been used to discover alternative, low-

temperature routes for the heterogeneous hydrogenation of liquid-phase

adiponitrile using novel nanocatalysts consisting of Co–Ru on TiO2 followed

by polymerization (110).

The approach used in these experiments is different from that with window

cells, which are generally not compatible with heating (71,72). The advances in

characterization with liquid-phase reactants may lead to new opportunities for

high-resolution imaging of a wide range of solution–solid and solution–gas–

solid reactions in the chemical and biological sciences.

IX. Environmental Scanning Electron Microscopy

Following early ETEM investigations using environmental cells, environmental

scanning electron microscopy (ESEM) has been developed for characterization

of surface effects of “bulk” SEM samples in the presence of gaseous or wet

environments (111–114). The method has been applied to the examination of

food, wool fibers (111), and polymers (112) and in the conservation of cultural

properties (113). Recently, fuel cell catalysts have been characterized using a

low-voltage ESEM with a resolution capability of ,2 nm (114).

X. Electron Tomography: Three-Dimensional Electron

Microscopy Imaging

There is a growing need for ultra-sensitive methods for determining the size,

elemental composition, precise location, spatial distribution, and detailed

morphology of nanoparticles anchored to high-area supports. In catalysis and

fuel cell technology, many different high-area (and generally low-atomic-

number) supports are employed, such as silica, alumina, and magnesia, as well as

graphitic, amorphous, or adamantine carbons and thermally stable polymers.

Furthermore, in many other areas of nanotechnology and biology, information

about three-dimensional morphology and understanding of the spatial distri-

bution and composition of nanoparticles are important.

As shown above, the size and distribution of minute particles are conveniently

investigated by high-resolution STEM with a HAADF detector (60,63). The

intensity in HAADF images is a monotonic function of the sample thickness and

atomic number, a pre-requisite for the electron tomography experiments

described below.


Electron tomography has been used in biology (115) to investigate the three-

dimensional structure of macromolecules and cells. Recently, the approach has

been applied to zeolites using conventional BF-TEM (116). Whereas

conventional transmission electron microscopic images are essentially two-

dimensional projections of the object—the structural features are superimposed

upon one another in the direction of the electron beam—in tomography, by

contrast, one acquires projections of the object as viewed from different

directions, and then one merges them computationally into a three-dimensional

reconstruction, the tomogram. For electron tomography, a series of images must

be recorded at successive tilt angles using a signal which must be a monotonic

function of the projected thickness of the sample (115,117). A schematic

diagram illustrating the acquisition of a series of tilted projections and the

reconstruction of a three-dimensional object (a magnetite nanocrystal from a

magnetotactic bacterium) is shown in Fig. 23 (116,117). The novel use of

HAADF-STEM to determine the three-dimensional structure of a supported

metal nanocatalyst at a spatial resolution of ,1 nm has been demonstrated for

Pd–Ru nanocatalysts supported on mesoporous silica (117,118). The goal was

achieved by tilting the sample to a series of different and finely spaced angles of

two-dimensional projection. In the same way as was used with the established

X-ray tomography methods, the information in the series of individual two-

dimensional images is analyzed to yield a detailed three-dimensional

construction of the structure, with the full resolution of the process (in this

case, 0.8 nm and potentially even higher resolution). The images obtained by

the use of HAADF-STEM signal are directly interpretable.

X.A. The Topography and Location of Nanoparticles in Supported

Catalysts; BSE and HAADF

Many catalysts consist of heavy (high-Z) atoms such as platinum, palladium,

ruthenium, or alloys (binary or ternary) and bimetallic variants of these elements,

supported on low-Z, high-area solids such as carbon, alumina, silica, or magnesia.

The metal particles are rendered readily visible by HAADF imaging, as described

above, and when a series of two-dimensional images is recorded (117,118) at a

succession of closely spaced tilt angles, tomographic information is retrieved.

Moreover, by using back-scattered (Rutherford) imaging, as pioneered by Gai

and Boyes (4), even more refined information may be gleaned about the spatial

distribution and topography of such nanocatalysts.

Back-scattered electrons (BSE), i.e., those scattered to angles greater than 908,

also yield sharp images of nanoparticles containing .100 atoms of high-Z

materials distributed over low-Z supports, again because they obey Rutherford’s

scattering law. BSE scattering may be thought of as reverse Rutherford


scattering, although the exact form of the experimental BSE scattering is

modified by the high (.908) scattering angle and by the bulk specimen

environment. High-angle scattered electrons recorded using a STEM equipped

with a HAADF detector and an SEM equipped with a BSE detector (Fig. 24) offer

an essentially incoherent signal, and images are monotonically dependent on the

atomic number of the sample and its thickness.

Typical examples of Rutherford-scattered imaging of nanoparticles of a

commercially important Pd/C catalyst recorded with (a) a BSE detector in a

field emission scanning electron microscope as well as (b) a STEM HAADF

image of the same 5% Pd/C sample, recorded in the same instrument, are shown

Fig. 23. Schematic diagram illustrating the acquisition of a series of titled projections and

reconstruction of the three-dimensional object (118,119).


in Fig. 25a and b, respectively (119). The strong Z-dependence is apparent in

the images, which show enhanced contrast from the nanoparticles. It is clear

that high-spatial resolution (,1 nm) is achievable in the FESEM, here operated

at 30 kV, and similar images are obtained in either BSE or forward scattered

(HAADF) mode. It is a simple matter to identify small particles in thin sections

by EDX methods. With a bulk (electron opaque) sample, the sensitivity of the

BSE method in the nanometer range (and of EDX on the sub-micron scale)

increases at medium to low voltages (with some limit set by instrumental

parameters). The mixing of the SEM-BSE signal, primarily for higher-Z particle

imaging—with a component of SE imaging, for lower Z support topography—

together with the use of medium to low beam energies, may prove to be the

optimum combination in the SEM (65,109).

We now illustrate the HAADF images of the Pd/C catalyst. Figure 26a shows a

single image from a series of successive tilt angles from þ60 to 2548 (119).

Figure 26b shows the images of the same sample where each image represents the

projection of the reconstructed three-dimensional structure (119,120). In these

images, the reconstruction was obtained using a back-projection approach, shown

schematically in Fig. 23. The data of Fig. 26 demonstrate the power of the

technique for monitoring changes in the three-dimensional distribution of

supported nanocatalysts.

The examples shown in the preceding paragraphs illustrate that combined use

of HAADF imaging and BSE imaging, both using Rutherford-scattered electrons,

Fig. 24. Schematic of BSE and HAADF detector geometry.


is powerful in recording images of nanoparticle catalysts supported on irregular

and thick carbonaceous supports. The incoherent scattering process ensures that

images are ideal for electron tomography and the reconstruction of three-

dimensional nanoparticle distributions (119). These studies show the consider-

able potential of the method in the analysis of nonuniform catalysts and similar

nanostructured systems. The images also illustrate that the HAADF and BSE

approach (in which images are directly interpretable) may be superior to

conventional BF-TEM and BF-STEM methods for catalysts, because of reduced

exposure of the samples to the electron beam. In conventional TEM, for example,

the large beam currents used can quickly damage the sample. BSE imaging can

also be simple and effective in the study of surface-loaded nanocatalysts on bulk

supports (employed in many industrial reactions), compared to conventional

TEM or STEM analysis, which requires electron-transparent samples.

Fig. 25. (a) SEM-BSE image and (b) STEM-HAADF image of palladium nanocatalysts on a

carbon support (119).


Fig. 26. (a) STEM-HAADF image acquired from the Pd/C sample shown in Fig. 25. (b) Animation

of the three-dimensional reconstruction of the object in (a) (119).


X.B. Pinpointing the Location of Nanoparticles Supported on

Nanoporous Solids

An exciting area of modern heterogeneous catalysis involves the production of

highly dispersed bimetallic nanoparticles (such as Cu4Ru12, Pd6Ru6, Ru10Pt2,

and Ru6Sn) distributed over the interior surfaces of mesoporous silica (the pore

diameter of which may be determined in the range of 2–20 nm by the prepara-

tion conditions). Such highly dispersed nanoparticles function effectively as

catalysts for a variety of solvent-free reactions, especially the hydrogenation of

organic molecules (121,122). Provided due care is taken in their preparation,

individual nanoparticles (of 10–15 A diameter) may be anchored to the inner

walls of the porous silica (Fig. 27). Figure 27a shows an HRTEM image of a

hexagonal array of nanopores in silica and Fig. 27b shows a schematic of the

interior of the single pore of silica. Evidence that the individual nanoparticles

are situated as depicted in Fig. 27d emerges from images such as the HAADF

image of Fig. 27c, which, as described in the preceding section, is an example

of Z-contrast imaging whereby elements of high-atomic number ðZÞ show

up readily against a background of low-Z elements. Indeed, because of a Z2

dependence on electron scattering cross-section of elements (described in

Section V), one Pt atom scatters as strongly as about 100 oxygen atoms or 32

silicon atoms (in conformity to the Rutherford scattering law). Images such as

that of Fig. 27c, coupled with electron tomography (123), show that the

nanoparticles are indeed anchored to the walls of the pores, and with the pore

diameter being so large there is ample room for reactant and product molecules

to diffuse in the pores.

XI. Energy Filtered Transmission Electron Microscopy and

Elemental Maps of Solid Catalysts Using EFTEM

Recent advances in elemental mapping of solid catalysts have been accomplished

by the use of EFTEM (124), as exemplified by the distribution of transition metal

ions in framework-substituted aluminophosphate, which are good shape-selective

and regio-selective oxidation catalysts (43,44,121). With up to about 4 at.% of the

Al3þ ions isomorphously replaced by either Co3þ or Mn3þ, giving oxyfunctio-

nalization catalysts for alkanes (122), it is important to know how uniformly these

ions (the active sites) are distributed. This is rapidly done by using a solid-state

detector to record the electron-stimulated XRE spectra characteristic of the ion, as

shown in the example of Fig. 10. Energy-filtered (EF) EM in various modes yields

the element distribution maps for light as well as heavy elements (124). Even

mixed-valence states in catalytic solids may be charted by electron-filtered EM


(125). In the case of silica-encapsulated bimetallic catalysts, one can establish

from precisely coincident element maps taken with Ru K-emission and Pd K-

emission X-rays that the individual nanoparticles retain their structural integrity

and are indeed nanoparticles such as Ru6Pd6 (or Ru12Cu4, (63)). EF images (for

example, those obtained using oxygen K-loss peaks or nitrogen K-loss peaks,

which are centered around 530 and 400 eV, respectively, or even plasmon-loss

peaks) are also instructive in revealing the distributions of light elements in

catalytic solids (14).

Fig. 27. (A) HRTEM micrograph of a typical hexagonal array of nanopores in silica (diameter

10 nm). (B) Computer graphic representation of the interior of a single pore of the silica showing

pendant silanol groups. (C) HAADF (see text) showing the distribution of anchored Ru6Sn

nanoparticles within the nanopores of the siliceous host. (D) Computer graphic illustration of the

Ru6Sn nanoparticles superimposed on an enlargement of the electron micrograph shown in (C). (After

Ref. (122b)).


XII. Other Significant Trends

The electron crystallography method (21) has been used to characterize three-

dimensional structures of siliceous mesoporous catalyst materials, and the three-

dimensional structural solutions of MCM-48 (mentioned above) and of SBA-1,

-6, and -16. The method gives a unique structural solution through the Fourier sum

of the three-dimensional structure factors, both amplitude and phases, obtained

from Fourier analysis of a set of HRTEM images. The topological nature of the

siliceous walls that define the pore structure of MCM-48 is shown in Fig. 28.

XIII. Critical Evaluations of the Methods and Challenges

The advanced EM methods described in this chapter are critical to the

fundamental understanding of the nanostructure and chemistry of chemically

and physically complex solid catalysts. These methods uniquely determine the

nature, atomic structure and crystallography of defect structures (disorder) at

catalyst surfaces in the reaction. These include whether defects result from

vacancies or interstitials, the nature of point defects associated with surface Lewis

or Bronsted acidity or basicity, their diffusion in the catalytic reaction, growth of

extended defects, and specific crystallographic planes and lattice displacements

(Burgers vector) involved in these processes. The nature of defects is, therefore, of

critical importance to the catalyst performance, in the hydrocarbon activation and

catalyst regeneration processes. Bulk diffraction methods such as X-ray

diffraction simply average data from larger areas, and scanning probe methods

(for which chemical composition and diffraction information are difficult and

deficient, respectively) require specialized sample preparations and are not

Fig. 28. Schematic illustration of the siliceous wall and channel structure of the mesoporous solid

known as MCM-48 (based on the results given in Ref. (122b)).


readily applicable to commercial catalysts. In EM, careful experimentation is

required along with understanding of the ED phenomena. We now address some

of the challenges and opportunities in the methods described in the chapter.

Conventional HRTEM operates at ambient temperature in high vacuum and

directly images the local structure of a catalyst at the atomic level, in real space.

In HRTEM, as-prepared catalyst powders can be used without additional sample

preparation. The method does not normally require special treatment of thin

catalyst samples. In HRTEM, very thin samples can be treated as WPOs, whereby

the image intensity can be correlated with the projected electrostatic potential of

the crystal, leading to the atomic structural information characterizing the

sample. Furthermore, the detection of electron-stimulated XRE in the EM

permits simultaneous determination of the chemical composition of the catalyst.

Both the surface and sub-surface regions of catalysts can be investigated.

However, care must be taken to use a very low-dose electron beam to avoid

beam damage to the sample. This is especially important in molecular sieve and

zeolite catalysts, which have extraordinary tendency to become amorphous under

prolonged exposure to the electron beam. This limitation has been overcome by

using high electron accelerating voltages in the EM (e.g., 200–300 kV instead of

100 kV), to minimize the inelastic collisions that are primarily responsible for the

structural degradation, along with better vacuum in the EM. For the new class of

ALPO catalysts, high-resolution CCD, because of their ability to record digital

images with very low-incident electron doses, are becoming increasingly

common to image catalysts and avoid sample damage.

In HRTEM of complex structures, image simulations are necessary to correlate an

experimental image with theory. Calculations are especially needed for images from

thicker samples, from the latest FEG HRTEMs and very high-voltage electron

microscopes. Electron crystallography, incorporating HRTEM, ED, and compu-

tational methods are powerful in determining the three-dimensional structure of

complex zeolites and molecular sieve structures which are not amenable to X-ray

diffraction. The approach offers opportunities in identifying the fine structure of

zeolite catalysts and metal promoters in particular positions in the catalysts.

Challenges include the determination of the three-dimensional structures of point

and extended defects on the surfaces of these materials during catalysis.

In supported catalysts, particle visibility may be a challenge if the support

thickness exceeds a certain value. This statement is applicable to both amorphous

and crystalline supports. Particles can be viewed in plan view or in the surface

profile mode. In the former, the contrast from nanoparticles can be obscured by

the support contrast (40). Surface profile imaging can be employed for thicker

industrial supported catalysts in which particles are visible only when they are

near the edge of the support. Investigations can provide insights into the structure

and shape of the nanoparticles even when the fraction of the particles near the

edges of the support is small. Out-of-focus imaging and image processing


methods are also helpful in gleaning structural information from supported

nanocatalysts. Calculations carried out by Gai et al. (40) show that in spherical-

aberration ðCsÞ corrected (ideal) electron microscopes, the particle visibility is

dramatically improved.

Lens aberrations (imperfections) yield limited spatial and spectral resolution in

EM. Sample thickness also affects the achievable resolution. HRTEM with

selected-area ED is especially useful in providing insights into the disorder and

ordering of anion vacancies in oxide catalysts in oxidation catalysis. To image

oxygen atom columns in an oxide using conventional HRTEMs with Cs; thin

samples are oriented down the exact crystallographic zone axis, and the imaging

requires appropriate defocus conditions. For example, imaging of oxygen atom

columns in high-temperature cuprates has been demonstrated (126,127).

Challenges for EM technology are, therefore, to achieve the development of

spherical (and chromatic) aberration-free electron microscopes to improve the

spatial and analytical resolution. Abberation-corrected HRTEM and STEM

instruments have been reported (128,129). Recent work using Cs-corrected EM

shows oxygen atom column imaging in perovskite ceramics (38). Thus,

aberration-corrected EMs are becoming routinely available.

The aforementioned development of in situ atomic-resolution environmental

TEM (ETEM) as a multifunctional “nanolaboratory” has enabled the determi-

nation of the structure and chemistry of catalysts including active site

configurations by atomic imaging, ED, and chemical analyses during catalysis.

Low-electron beam currents (well below the threshold for sample damage) are

employed, and the signal is amplified and recorded via a low-light level television

camera and a video system. In addition, blank experiments are performed without

the electron beam, and the beam is switched on for only a few seconds to record

the final state of the material. The results are then compared with those of in situ

experiments performed with very low electron doses to confirm the validity of the

in situ experiments. Under these controlled experimental conditions, beam

damage to the catalyst is not observed, and ETEM data can be directly related to

structure–property relationships and reaction kinetics in technological processes.

Time- and temperature-resolved experiments can be carried out. In situ ETEM

thus helps to reduce the time and costs involved in scaling up laboratory

experiments to industrial conditions. Because the method operates under

dynamic catalyst operating conditions, caution should be exercised in maintain-

ing the reaction environment, temperature regimes, and imaging. At present,

atomic-resolution in ETEM is possible with a few mbar of gas pressures. Higher

gas pressures (up to 1 bar) are possible, but the resolution is compromised at

higher pressures because of the absorption of electrons by thicker gas layers. In

gas–catalyst experiments, the coverage of the catalyst with the reactant-derived

species is crucial, and this is more important than the presence of high gas

pressures in the ECELL (gas reaction cell or microreactor), or around the sample.


Aberration-corrected ETEM/STEM (130) is expected to offer superior (sub-

atomic) resolution under catalytic reaction conditions; furthermore, it will

provide improved flexibility for tilting the sample to different crystallographic

orientations to allow understanding of the geometry of surface structural changes,

enable the use of complex sample stages, and perhaps higher gas pressures.

STEM uses a very small probe scanned in a raster across the sample. The method

provides many analytical signals, including HAADF and EELS, and offers several

advantages over conventional TEM. In HAADF, highly incoherent high-angle

scattering electrons are employed (Rutherford scattering), and the method is

sensitive to the atomic number of the atoms (Z-contrast). The HAADF signal

removes the complexity of conventional bright-field scattering in TEM

and associated diffraction complications and allows the direct interpretation

of results. Three-dimensional electron tomography using HAADF-STEM

(Z-contrast) is powerful in determining the structures of supported nanocatalysts.

The results are achieved by tilting the catalyst sample to a series of different and

finely spaced angles, and the images are reconstructed. Current challenges of

STEM include resolution; delivery of adequate current in the 0.2-nm probe in EDX

chemical analysis at the atomic level; beam damage to the sample; and sample

stability. Pulsing the electron beam onto the sample can be helpful in increasing the

sample stability. Aberration-corrected STEM can be helpful in obtaining high

probe currents for chemical analysis. In three-dimensional electron tomography, it

may be challenging to obtain enough tilt for the sample and reconstruction of three-

dimensional images of nanoparticles on irregular (and thick) supports. Wide gap

lenses with aberration correction may be able to provide adequate tilt range and

resolution. Electron beam damage to the sample is a fundamental issue in STEM,

and careful experimentation to ensure the stability of the sample is required.

In low-voltage, high-resolution SEM (LVSEM) of catalysts, a spatial

resolution 0.5 nm at 1 kV and more current in electron probes for high-precision

microchemical analysis are being sought. Challenges in LVSEM of catalysts

include control of the sample charging and preservation of sample stability. In

ESEM, challenges and opportunities include improved resolution and micro-

analysis with better sensitivity and accuracy.

XIV. Conclusions

Several general conclusions are drawn concerning the status of EM as a

supremely versatile tool in the study of the materials chemistry of catalysts. First,

it is no longer necessary to regard EM as a tool for model studies (131–133). The

triumphant exploitation of the environmental cell in HRTEM marks the dawn of

a new era in probing dynamic catalysis (4,87–95). Second, EM techniques, as has


recently been illustrated by Rupprechter (134), may be smoothly integrated with

parallel investigations (e.g., of polycrystalline, nanoparticle platinum, palladium,

and rhodium) by vibrational (sum frequency generation) spectroscopy and

scanning tunnelling microscopy. Thus, for example, with alumina-supported

rhodium nanoparticles, it was explicitly demonstrated that high-index faces (low-

coordinated sites) are preferred for hydrogenolysis catalysis. Extrapolating

Rupprechter’s results and recognizing the vast new possibilities that are now

possible (thanks to the arrival of intense near-IR femtosecond laser pulses) in

time-resolved in situ measurements (135,136), one may reasonably expect further

major advances in studies of polycrystalline rather than just single-crystal

surfaces. Finally, electron crystallography (21) and electron tomography

(117–119) are important new developments in the study of catalysts.

Acknowledgements

We thank our colleagues Osamu Terasaki, Edward Boyes, Paul Midgley, Robert

Raja, Frank Gooding, Leland Hanna, Kostantinos Kourtakis, Gopinath Sankar,

Matthew Weyland, and Brian Johnson for their friendly cooperation.

References

1. Thomas, J.M., and Lambert, R.M. (Eds.), “Characterization of Catalysts.” Wiley,

Chichester, 1980.

2. Thomas, J.M., and Thomas, W.J., Principles and Practice of Heterogeneous Catalysis.

Wiley-VCH, Weinheim, 1997.

3. Gai, P.L. (Gai-Boyes, P.L.), Catal. Rev.—Sci. Engng 34, 1 (1992).

4. Gai, P.L., and Boyes, E.D., “Electron Microscopy in Heterogeneous Catalysis.” Insitute of

Physics Publishing, Bristol, UK, 2003.

5. Thomas, J.M., Terasaki, O., Gai, P.L., Zhou, W., and Gonzalez-Calbet, J., Acc. Chem. Res.

34, 583 (2001).

6. Thomas, J.M., Evans, E.L., and Bahl, O.P., Surf. Sci. 8, 473 (1967).

7. Thomas, J.M., Inorganic Chemistry: Towards the 21st Century, 211, p. 445. ACS

Publication, 1983.

8. Thomas, J.M., Millward, G.R., and Bursill, L.A., Phil. Trans. R. Soc. A300, 43 (1981).

9. Terasaki, O., Millward, G.R., and Thomas, J.M., Proc. R. Soc. A395, 153 (1984).

10. Thomas, J.M., and Millward, G.R., J. Chem. Soc. Chem. Commun. 1380 (1982).

11. Stern, E.A., and Siegel, R.W., Curr. Opin. Solid State Mater. Sci. 4, 321 (1999).

12. (a) Redlich, P., Curr. Opin. Solid State Mater. Sci. 4, 325 (1999); (b) Thomas, J.M.,

Jefferson, D.A., and Egerton, R.F., Chem. Brit. 17, 514 (1981).

13. Sankar, G., and Thomas, J.M., Top. Catal. 8, 1 (1999).

14. Egerton, R.F., Top. Catal. 21, 185 (2002).

15. Thomas, J.M., and Sankar, G., Acc. Chem. Res. 34, 571 (2001).

16. Gai, P.L., and Thomas, J.M. (Eds.), Top. Catal. 21, 107 (2002).

17. Kirkland, A.J., and Sloan, J., Top. Catal. 21, 139 (2002).


18. Gai, P.L., Top. Catal. 21, 161 (2002).

19. Heinemann, H., and Poppa, H., Ultramicroscopy 17, 213 (1985).

20. Ohnishi, N., and Hiraga, K., J. Electron Microsc. 45, 85 (1996).

21. Carlsson, A., Kaneda, M., Sakamoto, Y., Terasaki, O., Ryoo, R., and Joo, S.H., J. Electron

Microsc. 48, 795 (1999).

22. Wagner, P., Terasaki, O., Ritsch, S., Nery, J.G., Zones, S., Davis, M.E., and Hiraga, K.,

J. Phys. Chem. B103, 8245 (1999).

23. Hirsch, P.B., Howie, A., Nocholson, R.B., Pashley, D., and Whelan, M.J., “Electron

Microscopy of Thin Crystals.” Kruger, Florida, 1977.

24. Gai, P.L., and Howie, A., Phil. Mag. A30, 939 (1974).

25. Gai, P.L., and McCarron, E., Science 247, 553 (1990).

26. Menter, R., Proc. R. Soc. A236, 119 (1952).

27. Cowley, J.M., and Moodie, A.F., Acta Cryst. 10, 609 (1957).

28. Cowley, J.M., and Iijima, S., Z. Naturforsch. 27a, 445 (1972).

29. Anderson, J.S., J. Chem. Soc. Dalton Trans. 1107 (1973).

30. Thomas, J.M., Phil. Trans. R. Soc. A277, 251 (1974).

31. Cowley, J.M., “Diffraction Physics.” North-Holland, Amsterdam, 1981.

32. Scherzer, O., J. Appl. Phys. 20, 20 (1949).

33. Boyes, E.D., Watanabe, E., Gai, P.L., Naruse, M., Jenkins, M.L., Hutchison, J.L., and

Skarnulis, A., Inst. Phys. (IOP, London, UK), Conf. Ser. 52, 445 (1980).

34. Thomas, J.M., Jefferson, D.A., and Millward, G.R., JEOL News 23E, 7 (1985).

35. Jefferson, D.A., Thomas, J.M., Millward, G.R., Harrison, A., and Brydson, R.D., Nature

323, 428 (1986).

36. Boyes, E.D., Hutchison, J., and Watanabe, E., JEOL News 24E (1986).

37. Ruhle, M., Ultramicroscopy 56, 1 (1994).

38. Jia, C.L., Lentzen, M., and Urban, K., Science 299, 870 (2003).

39. O’Keefe, M.A., Acta. Cryst. A29, 389 (1973).

40. Gai, P.L., Goringe, M.J., and Barry, J.C., J. Microsc. 142, 9 (1986).

41. Datye, A., and Smith, D.J., Catal. Rev.—Sci. Engng 34 (1992).

42. Yacaman, M.J., and Avalos-Borja, M., Catal. Rev.—Sci. Engng 34 (1992).

43. Gai, P.L. (Gai-Boyes, P.L.), Thomas, J.M., Wright, P.A., Jones, R.H., Natarajan, S., Chen,

J., and Xu, R., J. Phys. Chem. 96, 8206 (1992).

44. Wright, P.A., Natarajan, S., Thomas, J.M., Gai-Boyes, P.L., Jones, R.H., and Chen, J.,

Angew. Chem. 31, 1472 (1992).

45. Thomas, J.M., Raja, R., Sankar, G., and Bell, R.G., Acc. Chem. Res. 34, 191 (2001).

46. Millward, G.R., J. Chem. Soc. Faraday Trans. II 79, 1075 (1983).

47. Millward, G.R., Ramdas, S., and Thomas, J.M., Proc. R. Soc. A399, 57 (1985).

48. Thomas, J.M., Audier, M., Klinowski, J., Bursill, L.A., and Ramdas, S., Faraday Discuss.

72, 345 (1981).

49. Terasaki, O., Ohsuma, T., Alfredsson, V., Bovin, J., Watanabe, D., and Anderson, M.W.,

Chem. Mater. 5, 452 (1993).

50. Vaughan, D., Treacy, M.M.J., and Newsam, J.M., in “Guidelines for Mastering the

Properties of Molecular Sieves” (D. Barthomeu, Ed.), p. 96. Pleasure Press, New York,

1999.

51. Thomas, J.M., Millward, G.R., White, D., and Ramdas, S., J. Chem. Soc. Chem. Commun.

434 (1988).

52. Gai, P.L., Bart, J.C., and Boyes, E.D., Phil. Mag. A45, 531 (1982).

53. Gai, P.L., and Boyes, E.D., IOP (London) Conf. Ser. 61, 215 (1981).

54. Boyes, E.D., and Gai, P.L., in “Analytical Electron Microscopy” (R. Geiss, Ed.), p. 71.

San Francisco Press, San Francisco, 1981.


55. Thomas, J.M., Ultramicroscopy 8, 13 (1982).

56. Boyes, E.D., Goringe, M.J., and Muggridge, B., J. Microsc. 127, 321 (1982).

57. Cliff, I., and Lorimer, G., J. Microsc. 103, 203 (1975).

58. Crewe, A.V., Issacson, M., and Johnson, D.E., Rev. Sci. Inst. 42, 411 (1971).

59. Treacy, M.M.J., and Rice, S.B., Ultramicroscopy 7, 11 (1981).

60. Howie, A., J. Microsc. 117, 11 (1979).

61. Treacy, M.M.J., Howie, A., and Pennycook, S.J., Inst. Phys. Conf. Ser. 52, 261 (1980).

62. Pennycook, S.J., J. Mol. 20, 345 (1983).

63. Ozkaya, D., Zhou, W., Thomas, J.M., Midgley, P.A., Keast, V., and Hermans, S., Catal.

Lett. 60, 113 (1999).

64. (a) Boyes, E.D., Proc. ICEM-13 (Paris) 1, 51 (1994); (b) Boyes, E.D., Inst. Phys. (IOP,

London, UK) Conf. Ser. 168, 115 (2001).

65. Boyes, E.D., Microsc. Microanal. 6, 307 (2000).

66. Boyes, E.D., Adv. Mater. 10, 1277 (1998).

67. Boyes, E.D., Gai, P.L., and Warwick, C., Nature 313, 666 (1985).

68. Hashimoto, H., Naiki, T., Eto, T., and Fujiwara, K., Jpn. J. Appl. Phys. 7, 946 (1968).

69. Swann, P., and Tighe, N., Jernkont. Ann. 155, 251 (1971).

70. Butler, P., and Hale, F., “Dynamic Experiments.” North-Holland, Amsterdam, 1981.

71. Doube, D.D., Mater. Sci. Engng 12, 29 (1973).

72. Parsons, D.F., Adv. Biol. Med. Phys. 15, 161 (1974).

73. Gai, P.L., Phil. Mag. 43, 841 (1981).

74. Gai, P.L., Kristall. Technik 14, 1385 (1979).

75. Gai, P.L., Phil. Mag. 48, 359 (1983).

76. Gai, P.L., J. Solid State Chem. 49, 25 (1983).

77. Gai, P.L., and Labun, P.A., J. Catal. 94, 79 (1985).

78. Gai, P.L., Smith, B.C., and Owen, G., Nature 348, 430 (1990).

79. Doole, R.C., IOP Conf. Ser. 119, 161 (1991).

80. Crozier, P., Sharma, A., and Datye, A., Microsc. Microanal. 4, 228 (1998).

81. Goringe, M.J., Bithell, E., and Doole, R.C., Faraday Discuss. 105, 85 (1996).

82. Parkinson, G.R., and White, D., Proc. XI Int. Congr. Electron Microsc. 331 (1986).

83. Collozo-Davila, C., Marks, L., and Landree, E., Microsc. Microanal. 1, 219 (1995).

84. Baker, R.T., and Chludzinski, J., Carbon 19, 75 (1981).

85. Gai, P.L., and Kourtakis, K., Science 267, 661 (1995).

86. Boyes, E.D., and Gai, P.L., Mater. Res. Soc. 404, 53 (1996).

87. Boyes, E.D., and Gai, P.L., Ultramicroscopy 67, 219 (1997).

88. Gai, P.L., and Boyes, E.D., in “In situ Microscopy in Materials Research” (P.L. Gai, Ed.).

Kluwer, Boston, 1997.

89. Gai, P.L., Acta Cryst. B53, 346 (1997).

90. Gai, P.L., Adv. Mater. 10, 125 (1998).

91. Haggin, J., C&E News (Am. Chem. Soc.) 73, 39 (1995).

92. Jacoby, M., C&E News (Am. Chem. Soc.) 80, 26 (2002).

93. Hansen, T.W., Wagner, J., Hansen, P.L., Dahl, S., Topsøe, H., and Jacobsen, C., Science

294, 1508 (2001).

94. Oleshko, V., Crozier, P., Cantrell, R., and Westwood, A., J. Electron Microsc. 51, S27 (2002).

95. Gai, P.L., Top. Catal. 8, 97 (1999).

96. Tauster, S., Fung, S., and Garten, J.L., J. Am. Chem. Soc. 100, 170 (1978).

97. Gai, P.L., and Smith, B.C., Ultramicroscopy 34, 17 (1990).

98. Sinfelt, J.H., “Bimetallic Catalysts.” Wiley, New York, 1985.

99. Centi, G., Catal. Today 16(1) (1993).

100. Gorbunova, Yu., and Linde, S., Dokl. Acad. SSSR 245, 584 (1979).


101. Gai, P.L., Kourtakis, K., Coulson, R., and Sonnichsen, G.C., J. Phys. Chem. 101, 9916

(1997).

102. Gai, P.L., J. Solid State Chem. 104, 119 (1993).

103. Gai, P.L., Curr. Opin. Solid State Mater. Sci. 4, 63 (1999).

104. Gai, P.L., in “Catalyst Materials for High Temperature Processes” (K. Ramesh, M.

Misono and P.L. Gai, Eds.), p. 127. American Ceramic Society, Westerville, OH, USA,

1997.

105. Gai, P.L., Curr. Opin. Solid State Mater. Sci. 5, 371 (2001).

106. Service, R.F., Science 294, 2442 (2001).

107. Gai, P.L., and Harmer, M.A., Nano Lett. 2, 771 (2002).

108. De Bellefon, C., and Foilloux, P., Catal. Rev.—Sci. Engng 36, 459 (1994).

109. Gai, P.L., Kourtakis, K., and Ziemecki, S., Microsc. Microanal. 6, 335 (2000).

110. Gai, P.L., Microsc. Microanal. 8, 21 (2002).

111. Danilotos, G., in “In situ Microscopy in Materials Research” (P.L. Gai, Ed.), p. 13.

Kluwer, Boston, 1997.

112. Donald, A., Proc. EUREM 12, 1 (2000).

113. Doehne, E., in “In situ Microscopy in Materials Research” (P.L. Gai, Ed.), p. 45. Kluwer,

Boston, 1997.

114. Boyes, E.D., “Frontiers in Electron Microscopy Conference,” Berkeley, CA, USA, 2003.

115. Crowther, R.A., de Rossier, D., and Klug, A., Proc. R. Soc. A317, 319 (1970).

116. Koster, A.J., and de Jong, K.P., J. Phys. Chem. B104, 9368 (2000).

117. Midgley, P.A., Weyland, M., Thomas, J.M., and Johnson, B.F.G., Chem. Commun. 907

(2001).

118. Weyland, M., Midgley, P.A., and Thomas, J.M., J. Phys. Chem. B 105, 7882 (2001).

119. Midgley, P.A., Weyland, M., Thomas, J.M., Gai, P.L., and Boyes, E.D., Angew. Chem. Int.

Ed. 41, 3804 (2002).

120. For further details an animated description of electron tomographic work on three-

dimensional nanostructure analysis based on Rutherford scattering (117–119) see http://

www.hrem.msm.cam.uk/~mw259/Work/Tomo.html.

121. Raja, R., Angew. Chem Int. Ed. 40, 4638 (2001).

122. (a) Thomas, J.M., Raja, R., Sankar, G., Johnson, B.F.G., and Lewis, D.W., Chem. Eur. J.

7, 2973 (2001); (b) Thomas, J.M., and Terasaki, O., Top. Catal. 21, 155 (2002).

123. Weyland, M., Top. Catal. 22, 175 (2002).

124. Thomas, P., and Midgley, P.A., Top. Catal. 22, 107 (2002).

125. Wang, Z., Bentley, J., and Evans, N., J. Phys. Chem. B 103, 751 (1999).

126. Gai, P.L., Subramanian, M.A., Gopalakrishnan, J., and Boyes, E.D., Physica C 159, 801

(1989).

127. Gai, P.L., and Thomas, J.M., Supercond. Rev. 1, 1 (1991).

128. Haider, M., Kabius, K., Uhelmann, S., Urban, K., and Rose, H., Microsc. Microanal. 7,

902 (2001).

129. Batson, P., and Krivanek, O., Nature 423, 211 (2003).

130. Gai, P.L., and Boyes, E.D., “Frontiers in Electron Microscopy Conference,” Berkeley,

CA, USA, 2003.

131. Freund, H.-J., Baumer, M., and Kuhlenbeck, H., Adv. Catal. 45, 412 (2000).

132. Poppa, H., Catal. Rev.—Sci. Engng 35, 359 (1993).

133. Henry, C.R., Surf. Sci. Rep. 31, 235 (1998).

134. Rupprechter, G., Phys. Chem. Chem. Phys. 3, 4621 (2001).

135. Hess, C., Wolf, M., and Bonn, M., Phys. Rev. Lett. 85, 4321 (2000).

136. Thomas, J.M., Faraday Discuss. 122, 395 (2002), and references cited therein.


Chemistry and Technology of

Isobutane/Alkene Alkylation

Catalyzed by Liquid and Solid Acids

ANDREAS FELLER1 and JOHANNES A. LERCHER

Institut fur Technische Chemie, Technische Universitat Munchen,

D-85747 Garching, Germany

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

II. Alkylation Mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

II.A. Overall Product Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

II.B. Initiation Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

II.C. Alkene Addition and Isomerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

II.D. Hydride Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

II.E. Oligomerization and Cracking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

II.F. Self-Alkylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

II.G. Product and Acid Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

II.H. Pathways to Allylic and Cyclic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

II.I. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

III. Physical–Chemical Phenomena Influencing the Reaction . . . . . . . . . . . . . . . . . . . . . . . 252

III.A. Properties of Liquid Acid Alkylation Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . 253

III.B. Properties of Zeolitic Alkylation Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

III.B.1. Adsorption and Diffusion of Hydrocarbons . . . . . . . . . . . . . . . . . . . . . . 255

III.B.2. Brønsted Acid Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

III.B.3. Lewis Acid Sites and Extra-Framework Aluminum . . . . . . . . . . . . . . . . 260

III.B.4. Silicon/Aluminum Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

III.B.5. Metal Ions in Ion-Exchange Positions . . . . . . . . . . . . . . . . . . . . . . . . . . 263

III.B.6. Structure Types of Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

III.C. Other Solid Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

III.C.1. Sulfated Zirconia and Related Materials . . . . . . . . . . . . . . . . . . . . . . . . 267

III.C.2. Heteropolyacids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268

III.C.3. Acidic Organic Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

III.C.4. Supported Metal Halides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270

III.D. The Influence of Process Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

III.D.1. Reaction Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272


E-mail address: [email protected] address: CS CLEAN SYSTEMS AG, Fraunhoferstr. 4, 85732 Ismaning, Germany.

A. Feller and J. A. Lercher / Adv. Catal. 48 (2004) 229–295

III.D.2. Alkane/Alkene Ratio and Alkene Space Velocity. . . . . . . . . . . . . . . . . . 274

III.D.3. Alkene Feed Composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

IV. Industrial Processes and Process Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

IV.A. Liquid Acid-Catalyzed Processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

IV.A.1. Sulfuric Acid-Catalyzed Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

IV.A.2. Hydrofluoric Acid-Catalyzed Processes . . . . . . . . . . . . . . . . . . . . . . . . . 281

IV.B. Solid Acid-Catalyzed Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

IV.B.1. UOP Alkylenee Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

IV.B.2. Akzo Nobel/ABB Lummus AlkyCleane Process. . . . . . . . . . . . . . . . . . 286

IV.B.3. LURGI EUROFUELw Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

IV.B.4. Haldor Topsøe FBAe Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

V. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

This contribution is an in-depth review of chemical and technological aspects of the

alkylation of isobutane with light alkenes, focused on the mechanisms operative with both liquid

and solid acid catalysts. The differences in importance of the individual mechanistic steps are

discussed in terms of the physical–chemical properties of specific catalysts. The impact of

important process parameters on alkylation performance is deduced from the mechanism. The

established industrial processes based on the application of liquid acids and recent process

developments involving solid acid catalysts are described briefly. q 2004 Elsevier Inc.

Abbreviations

ASO acid-soluble oil

DMH dimethylhexane

EFAL extra-framework aluminum

H0 Hammett acidity function

k rate constant

LHSV liquid-hourly space velocity (m3olefin/(m3

catalyst h))

OSV olefin space velocity (kgolefin/(kgcatalyst h))

P/O paraffin/olefin ratio (mol/mol)

r reaction rate

RE rare earth

RON research octane number

T temperature (K)

TMP trimethylpentane

TOS time on stream

TS transition state

WHSV weight-hourly space velocity (kgolefin/(kgcatalyst h))

I. Introduction

Alkylation of isobutane with C3–C5 alkenes in the presence of strong acids leads

to the formation of complex mixtures of branched alkanes, called alkylate, which

are excellent blending components for gasoline. Alkylate has a high octane


number and a low Reid vapor pressure, and is free of aromatics, alkenes, and

sulfur. The worldwide production capacity as of the end of 2001 was approxi-

mately 74 million tons/year (1). Because of increasing strictness of the clean air

regulations in the EU and the USA and restrictions of the contents of alkenes,

sulfur, and aromatics (particularly benzene) in gasoline, the production of

alkylate is expected to increase. Furthermore, the planned phase-out of methyl-

tertiary-butyl ether (MTBE), a high-octane-number oxygenate, will boost the

demand for alkylate to meet the requirements for reformulated gasoline (2).

Alcohols such as ethanol, that could conceivably replace the ethers, suffer from a

very high blending vapor pressure when mixed into gasoline, thus limiting their

usefulness. Therefore, it is expected that the demand for alkylation catalysts will

increase by 5% per year up to the year 2003, with an estimated total catalyst value

for 2003 of $340 million (3).

The alkylation unit in a petroleum refinery is situated downstream of the fluid

catalytic cracking (FCC) units. The C4 cut from the FCC unit contains linear

butenes, isobutylene, n-butane, and isobutane. In some refineries, isobutylene is

converted with methanol into MTBE. A typical modern refinery flow scheme

showing the position of the alkylation together with an acid regeneration unit is

displayed in Fig. 1.

In the 1930s, Ipatieff’s group at Universal Oil Products discovered that

isoalkanes react with alkenes in the presence of strong acids to give saturated

hydrocarbons under relatively mild conditions. The acids initially tested

were AlCl3/HCl and BF3/HF (4). Soon, the first processes were commercia-

lized (5). The early alkylation plants utilized sulfuric acid, but the need for

high-octane-number aviation gasoline spurred by World War II led to the

construction of plants based on HF as catalyst, which are more flexible

regarding the feed alkenes. The first HF alkylation process units were built in

1942 by Phillips as wartime emergency units (6). The importance of alkylate

increased steeply, and the daily production of alkylate then reached 5 million

gallons; during the Korean War in 1952 the production rate was already 14

million gallons/day, and in the beginning of the 1980s, with the phase out of

leaded gasoline in the USA, it increased to an estimated 50 million gallons/

day (7). From the 1960s to about 1986, the relative importance of plants

using HF increased relative to those using H2SO4 (8). Now, nearly equal

amounts of alkylate are produced on a worldwide basis by each of the two

processes (1).

Both H2SO4 and HF catalysts suffer from substantial drawbacks. Anhydrous

HF is a corrosive and highly toxic liquid with a boiling point close to room

temperature. Tests in the Nevada desert showed that, if released into the

atmosphere, HF forms stable aerosols, which drift downwind at ground level

for several kilometers. In 1987, the accidental release of gaseous HF in Texas

City resulted in emergency treatment for several hundred people (9). Therefore,

A. Feller and J. A. Lercher / Adv. Catal. 48 (2004) 229–295 231

Fig. 1. Process units in a modern refinery.

A.

Feller

and

J.A

.L

ercher

/A

dv

.C

atal.4

8(2

00

4)

22

9–

29

52

32

refineries with HF alkylation plants are under pressure to install expensive

mitigation systems minimizing the dangers of HF leaks. Moreover, authorities

in many industrialized countries have ceased to license new HF alkylation

plants.

Sulfuric acid is also a corrosive liquid, but not volatile, making its handling

easier. Its major disadvantage is the high acid consumption in the alkylation

process, which can be as much as 70–100 kg of acid/ton of alkylate. The spent

acid contains water and heavy hydrocarbons and has to be regenerated, usually by

burning. The cost of such a regenerated acid is about 2–3 times the market price

for sulfuric acid (10). About one-third of the total operating costs of alkylation

units using H2SO4 can be attributed to acid consumption (11). The sulfuric acid-

catalyzed process is more sensitive than the other towards the feed alkenes; C3

and C5 alkenes generally lead to higher acid consumption and lower octane

numbers of the alkylate (12). Equipment corrosion, transport, and handling

hazards and environmental liability associated with the disposal of spent acid are

disadvantages of both the processes.

For more than 30 years, these issues have provided the driving force for

research in industry and academia to find suitable replacements for the existing

liquid acid catalysts. Zeolites, being non-corrosive, non-toxic, and rather

inexpensive, seemed to be promising candidates, especially after they were

successfully installed as cracking catalysts. In the late 1960s, two groups, those of

Garwood and Venuto of Mobil Oil (13) and Kirsch, Potts, and Barmby of Sun

Oil (14), did pioneering work on rare earth exchanged faujasitic zeolites. Later,

other zeolites were also examined. In general, all large-pore zeolites are active

alkylation catalysts, giving product distributions similar to those characteristic of

the liquid acids, but their unacceptably rapid deactivation was and still is the

obstacle to commercialization.

Other materials that have been investigated include sulfated zirconia,

Brønsted and Lewis acids promoted on various supports, heteropolyacids, and

organic resins, both supported and unsupported. On the whole, these materials

also deactivate rapidly, and some of them also exhibit environmental and

health hazards.

The technology and chemistry of isoalkane–alkene alkylation have been

thoroughly reviewed for both liquid and solid acid catalysts (15) and for solid

acid catalysts alone (16). The intention of this review is to provide an up-to-

date overview of the alkylation reaction with both liquid and solid acids as

catalysts. The focus is on the similarities and differences between the liquid

acid catalysts on one hand and solid acid catalysts, especially zeolites, on the

other. Thus, the reaction mechanism, the physical properties of the individual

catalysts, and their consequences for successful operation are reviewed. The

final section is an overview of existing processes and new process developments

utilizing solid acids.


II. Alkylation Mechanism

Since the discovery of alkylation, the elucidation of its mechanism has attracted

great interest. The early findings are associated with Schmerling (17–19), who

successfully applied a carbenium ion mechanism with a set of consecutive and

simultaneous reaction steps to describe the observed reaction kinetics. Later,

most of the mechanistic information about sulfuric acid-catalyzed processes was

provided by Albright. Much less information is available about hydrofluoric acid

as catalyst. In the following, a consolidated view of the alkylation mechanism

is presented. Similarities and dissimilarities between zeolites as representatives

of solid acid alkylation catalysts and HF and H2SO4 as liquid catalysts are high-

lighted. Experimental results are compared with quantum-chemical calculations

of the individual reaction steps in various media.

II.A. Overall Product Distribution

Table I gives the compositions of alkylates produced with various acidic catalysts.

The product distribution is similar for a variety of acidic catalysts, both solid and

liquid, and over a wide range of process conditions. Typically, alkylate is a

mixture of methyl-branched alkanes with a high content of isooctanes. Almost all

the compounds have tertiary carbon atoms; only very few have quaternary carbon

atoms or are non-branched. Alkylate contains not only the primary products,

trimethylpentanes, but also dimethylhexanes, sometimes methylheptanes, and a

considerable amount of isopentane, isohexanes, isoheptanes and hydrocarbons

with nine or more carbon atoms. The complexity of the product illustrates that no

simple and straightforward single-step mechanism is operative; rather, the

reaction involves a set of parallel and consecutive reaction steps, with the

importance of the individual steps differing markedly from one catalyst to

another. To arrive at this complex product distribution from two simple molecules

such as isobutane and butene, reaction steps such as isomerization, oligomeriza-

tion, b-scission, and hydride transfer have to be involved.

The distributions of products within a certain carbon number fraction are

far from equilibrium. In the C8-fraction, for example, the dimethylhexanes would

be thermodynamically favored over the trimethylpentanes, but the latter are

predominant. The distribution within the trimethylpentanes is also not

equilibrated. 2,2,4-TMP would prevail at equilibrium over the other TMPs,

constituting 60–70% of the product, depending on the temperature. Furthermore,

2,2,3-TMP as the primary product is found in less than equilibrium amounts.

Qualitatively, the same statement is valid for the other carbon number

distributions. Products with a tertiary carbon atom in the 2-position dominate

over other isomers in all fractions.


TABLE I

RON values of various alkanes and the C5þ composition of isobutane/butene alkylates produced with various acids in laboratory

scale/pilot-plant scale reactors

Component

(wt%)

Research octane number Catalyst

H2SO4

(T ¼ 528 K, P=O ¼ 5)

HF

(T ¼ ?; P=O ¼ 12)

RE-FAU

(T ¼ 348 K, P=O ¼ 7)

Sulfated zirconia

(T ¼ 275 K, P=O ¼ 15)

Isopentane 93.0 1.2 1.8 6.8 24.0

n-Pentane 61.8 0 0.1 0 0

2,2-Dimethylbutane 91.8 0 0 0 0.8

2,3-Dimethylbutane 104.3 1.5 1.4 4.8 4.3

2-Methylpentane 73.4 0.2 1.4 4.8 3.5

3-Methylpentane 74.5 0.1 0.1 0.7 1.7

n-Hexane 24.8 0 0 0 0

2,2-Dimethylpentane 92.8 0 1.3 0 0.1

2,4-Dimethylpentane 83.1 0.6 1.3 3.5 5.5

2,2,3-Trimethylbutane 112.1 0.1 0 0.2 0.3

3,3-Dimethylpentane 80.8 0 0 0 0.3

2,3-Dimethylpentane 91.1 0.6 0.6 1.7 1.8

2-Methylhexane 42.4 0 0.1 1.7 1.0

3-Methylhexane 52.0 0 0.2 0.3 0.7

(Continued)

A.

Feller

and

J.A

.L

ercher

/A

dv

.C

atal.4

8(2

00

4)

22

9–

29

52

35

TABLE I

Continued

Component

(wt%)

Research octane number Catalyst

H2SO4

(T ¼ 528 K, P=O ¼ 5)

HF

(T ¼ ?; P=O ¼ 12)

RE-FAU

(T ¼ 348 K, P=O ¼ 7)

Sulfated zirconia

(T ¼ 275 K, P=O ¼ 15)

2,2,4-Trimethylpentane 100 30.2 48.7 23.8 25.5

n-Heptane 0 0 0 0 0

2,2-Dimethylhexane 72.5 0 0 0 0.4

2,4-Dimethylhexane 65.2 1.2 2.9 1.1 0.8

2,5-Dimethylhexane 55.5 2.0 2.1 10.1 0

2,2,3-Trimethylpentane 109.6 0.8 1.1 10.1 11.0


2,3-Dimethylhexane 71.3 1.7 2.1 3.0 0.9

2-Methylheptane 21.7 0 0 0 0


3,4-Dimethylhexane 76.3 0.2 0.2 1.0 0.4

3-Methylheptane 26.8 0 0 0 0

Octenes .90 0 0 0.3 1.3

C9þ <80–85 5.4 2.9 7.5 3.3

Data taken from Ref. (20) for H2SO4, Ref. (21) for HF, and Ref. (22) for sulfated zirconia; RE-FAU, unpublished data.

A.

Feller

and

J.A

.L

ercher

/A

dv

.C

atal.4

8(2

00

4)

22

9–

29

52

36

The overall reaction is highly exothermic. Depending on the product

composition, 82–93 kJ/mol of reacted isobutane are liberated (23).

II.B. Initiation Steps

The alkylation reaction is initiated by the activation of the alkene. With liquid

acids, the alkene forms the corresponding ester. This reaction follows

Markovnikov’s rule, so that the acid is added to the most highly substituted

carbon atom. With H2SO4, mono- and di-alkyl sulfates are produced, and with

HF alkyl fluorides are produced. Triflic acid (CF3SO2OH) behaves in the same

way and forms alkyl triflates (24). These esters are stable at low temperatures and

low acid/hydrocarbon ratios. With a large excess of acid, the esters may also be

stabilized in the form of free carbenium ions and anions (Reaction (1)).

ð1Þ

The esters differ from each other in stability. To decompose the isopropyl

ester, higher temperatures and higher acid strengths are needed than for

decomposition of the s-butyl ester. It is claimed that the resulting carbenium ions

are stabilized by solvation through the acid (25–27). Branched alkenes do not

form esters. It is believed that they are easily protonated and polymerized (28).

In zeolites, the adsorption of an alkene will lead to a surface alkoxide and not

to an adsorbed carbenium ion. The alkene is “solvated” by the basic surface

oxygen atoms of the zeolite, and the solvation is similar to that by water in

aqueous solutions. Depending on the basicity of the surface oxygen atoms,

proton transfer to adsorbed alkenes results in the formation of more or less

covalent surface alkoxides rather than carbenium ions (29,30). Ab initio

quantum-chemical calculations representing a cluster modeling the zeolitic acid

site (29,31) showed that the alkene first forms a p-complex with the acidic site.

This transforms via a carbenium ion-like transition state into the alkoxide. The

transition state has a much higher positive charge than the alkoxide, and it forms

a cyclic species with both oxygen atoms and an aluminum atom of the zeolite.

The final alkoxide will not bind to the oxygen to which the hydrogen was

bonded but instead to one neighboring it. The involvement of both oxygen

atoms and the “switching” between them is characteristic of hydrocarbon

transformations on zeolitic acid sites (32). An illustrative energy diagram is

depicted for the isobutylene protonation in Fig. 2.

More recent calculations representing propene chemisorption, however,

showed the sensitivity of the system to the surrounding zeolite structure. The

calculated energies were found to depend strongly on the relaxation of the zeolite


unit cell size and its shape (33). Experimentally, monomeric alkoxides are

difficult to investigate. Because of their high reactivities, alkenes tend to

oligomerize, so that mainly dimerized species were detected upon adsorption of

isobutylene and of n-butenes on zeolites (34,35).

In their experiments with perdeuterioisobutane on various zeolites, Engel-

hardt and Hall (36) found the carbenium ions to be metastable reaction

intermediates. The lifetime of an intermediate was concluded to depend on the

acid strength.

The direct protonation of isobutane, via a pentacoordinated carbonium ion, is

not likely under typical alkylation conditions. This reaction would give either a

tertiary butyl cation (trimethylcarbenium ion) and hydrogen, or a secondary

propyl cation (dimethylcarbenium ion) and methane (37–39). With zeolites, this

reaction starts to be significant only at temperatures higher than 473 K. At lower

temperatures, the reaction has to be initiated by an alkene (40). In general, all

hydrocarbon transformations at low temperatures start with the adsorption of the

much more reactive alkenes, and alkanes enter the reaction cycles exclusively

through hydride transfer (see Section II.D).

When n-butenes are used, the initiation produces a secondary carbenium

ion/butoxide. This species may isomerize via a methyl shift (Reaction (2)) or

accept a hydride from isobutane to form the tertiary butyl cation (Reaction (3)).

Isobutylene forms the tertiary cation directly.

ð2Þ

Fig. 2. Potential energy profile and structure of final alkoxide for the adsorption of isobutylene on a

high-silica zeolite according to Ref. (29).


ð3Þ

The skeletal rearrangement needed in reaction (2) has to go through a transition

state, which resembles a primary carbenium ion, for which an activation energy

of about 130 kJ/mol has been calculated (41). In zeolites and presumably in the

liquid acids also, this reaction does not proceed under alkylation conditions.

Another possibility is the addition of a butene molecule to the secondary butyl

cation, giving a 3,4-dimethylhexyl cation, which can be freed via hydride transfer

from isobutane and form the tertiary butyl cation in this way. This route seems to

play only a minor role, as no significantly higher dimethylhexane selectivities

during the initial reaction phase have been reported. At the same time, n-butane is

formed in substantial amounts at this stage, confirming the importance of this

initiating step.

In the reaction with sulfuric acid and n-butenes or propene, only minor

amounts of n-butane or propane are observed. Only little isobutane is consumed

in the initial phase, whereas the alkenes react immediately (42). In this case, the

alkenes first oligomerize to form conjunct polymers. These polymers are also

called acid-soluble oil (ASO) or red oil, because they are found in the acid

phase and exhibit a dark red color. This oil is a complex mixture of highly

branched hydrocarbons with single and conjugated double bonds and rings

containing five and six carbon atoms. The individual compounds have

molecular weights in the range of 265–360 (43). They can abstract a hydride

from isobutane, forming a tertiary carbenium ion (8,44). When the reaction

is started with sulfuric acid that already contains some ASO, a better alkylate is

produced than with fresh acid (45), and the initiation period, which is

characterized by low yield and product quality, is markedly reduced (46).

The importance of the conjunct polymers is discussed below.

II.C. Alkene Addition and Isomerization

Once the tertiary cations have been formed, they can undergo electrophilic

addition to alkene molecules (Reaction (4)). The addition is exothermic and

contributes most of all the reaction steps to the overall heat of reaction. It has

been proposed (24) that instead of the alkenes, the corresponding esters are added

to the carbenium ions, restoring the acid in this way (Reaction (5)). The products

of both potential steps are the same.

ð4Þ


ð5Þ

In the case of the butene isomers, the addition will lead to different isooctyl

cations, depending on the isomer and the type of carbenium ion. The reactions

involving s-butyl ions are likely to be negligible for liquid acid catalysts and of

minor importance for zeolites.

2-Butene as the feed alkene would thus—after hydride transfer—give 2,2,3-

TMP as the primary product. However, with nearly all the examined acids,

this isomer has been observed only in very small amounts. Usually the main

components of the TMP-fraction are 2,3,3-, 2,3,4-, and 2,2,4-TMP, with the

selectivity depending on the catalyst and reaction conditions. Consequently, a

fast isomerization of the primary TMP-cation has to occur. Isomerization through

hydride- and methyl-shifts is a facile reaction. Although the equilibrium

composition is not reached, long residence times favor these rearrangements

(47). The isomerization pathways for the TMP isomers are shown schematically

in Fig. 3.

Using 1-butene as the feed alkene in most cases does not lead to

dimethylhexanes as expected, but also to a mixture of TMPs. These are formed

in a rapid isomerization of the linear butenes, almost to equilibrium com-

positions, in which the 2-butenes are strongly favored. On the other hand, some

of the DMH-isomers produced in 2-butene alkylation also stem from a rapid

isomerization of the feed.

Not all acids are equally active isomerization catalysts. With zeolite H-BEA,

nearly identical selectivities are achieved when the feed is 1-butene instead of

2-butene (48). In general, even mildly acidic zeolites are excellent catalysts for

double-bond shift isomerization. Sulfuric acid also produces nearly identical


alkylates with 1- and 2-butene (45,49). Hydrofluoric acid, on the other hand, is

known to produce substantial amounts of DMHs from 1-butene (21,50).

Aluminum chloride also shows low rates of butene isomerization (18,51). It

seems unlikely that under mild alkylation conditions skeletal rearrangements,

which could isomerize TMP-cations into DMH-cations (and methylheptyl

cations), occur to a large extent. This type of isomerization has a much higher

true activation energy than hydride and methyl shifts.

Theoretically, even the direct alkylation of carbenium ions with isobutane is

feasible. The reaction of isobutane with a t-butyl cation would lead to 2,2,3,3-

tetramethylbutane as the primary product. With liquid superacids under

controlled conditions, this has been observed (52), but under typical alkylation

conditions 2,2,3,3-TMB is not produced. Kazansky et al. (26,27) proposed the

direct alkylation of isopentane with propene in a two-step alkylation process. In

this process, the alkene first forms the ester, which in the second step reacts

with the isoalkane. Isopentane was found to add directly to the isopropyl ester

via intermediate formation of (non-classical) carbonium ions. In this way, the

carbenium ions are freed as the corresponding alkanes without hydride transfer

(see Section II.D). This conclusion was inferred from the virtual absence of

propane in the product mixture. Whether this reaction path is of significance in

conventional alkylation processes is unclear at present. HF produces substantial

amounts of propane in isobutane/propene alkylation. The lack of 2,2,4-TMP in

the product, which is formed in almost all alkylates regardless of the feed (53),

implies that the mechanism in the two-step alkylation process is different from

that of conventional alkylation.

Fig. 3. Possible hydride- (<H) and methyl-shifts (<CH3) between the individual TMP isomers.


II.D. Hydride Transfer

Intermolecular hydride transfer (Reaction (6)), typically from isobutane to an

alkyl-carbenium ion, transforms the ions into the corresponding alkanes and

regenerates the t-butyl cation to continue the chain sequence in both liquid acids

and zeolites.

ð6Þ

Hydride transfer is the crucial step in the reaction sequence. It ensures the

perpetuation of the catalytic cycle and leads to the exclusive formation of

saturated compounds. In general, the hydride transfer between alkanes and alkyl

cations is the elementary step responsible for chain propagation of acid-catalyzed

transformations of hydrocarbons (54). Hydride transfer between tertiary carbon

atoms is much faster than that between secondary carbon atoms. Although

hydride transfers involving secondary alkyl cations take place in aluminum

halide systems (55), they are too slow to be observed in sulfuric acid (56). In

general, hydride transfer is accelerated by neighboring groups, which encourage

the stabilization of the resulting ion (57).

Investigations of hydride transfer in the gas phase (58–61) showed that the

reaction proceeds without a substantial activation energy. Its reaction rate was

found to exhibit two regimes, i.e., fast kinetics at low temperatures and slow

kinetics at high temperatures. This behavior was explained by a consecutive

mechanism proceeding through two reaction steps. It involved the formation of a

loose complex between the ion and the neutral alkane, which reacts to form a

tight complex having a bridging hydride between the two fragments. The rates of

different hydride transfer reactions between different carbenium ions and

different alkanes were found to depend on the reaction enthalpy and steric

factors involving van der Waals interactions between the approaching ion and

hydrogen and methyl groups on the adjacent carbon atom next to the tertiary

carbon atom. Steric hindrance in tertiary–tertiary hydride transfer reactions was

also established in the liquid phase employing superacidic catalysts (62). These

steric restrictions are presumably responsible for the low selectivity to the

primary product 2,2,3-TMP observed with all acids. Hydride- or methyl-shifts are

much more likely than hydride transfer to a difficult-to-access carbon atom

bearing the positive charge. Note that the precursor carbenium ions of the most

abundant TMPs have their charge centers next to the chain end at a tertiary

carbon atom (Fig. 3).

There are substantial differences between gas-phase and liquid-phase hydride

transfer reactions. In the latter, the hydride transfer occurs with a low activation

energy of 13–17 kJ/mol, and no carbonium ions have been detected as

intermediates when secondary or tertiary carbenium ions were present (25).


These differences were explained by solvation effects in the liquid phase. The

carbenium ions are more efficiently stabilized by solvation than carbonium ions,

because the former have unsaturated trivalent carbon atoms. In this way, the

energy barrier between the (solvated) carbenium ion and the carbonium ion

transition state increases.

In zeolites, this barrier is even higher. As discussed in Section II.B, the lower

acid strength and the interaction between the zeolitic oxygen atoms and the

hydrocarbon fragments lead to the formation of alkoxides rather than carbenium

ions. Thus, extra energy is needed to transform these esters into carbonium ion-

like transition states. Quantum-chemical calculations of hydride transfer between

C2–C4 adsorbed alkenes and free alkanes on clusters representing zeolitic acid

sites led to activation energies of approximately 200 kJ/mol for isobutane/tert-

butoxide (29), 230–305 kJ/mol for propane/sec-propoxide, and 240 kJ/mol for

isobutane/tert-butoxide (32), 130–150 kJ/mol for ethane/ethene (63), 95–

105 kJ/mol for propane/propene, 88–109 kJ/mol for isobutane/isobutylene, and

110–118 kJ/mol for propane/isobutylene (64). In the last two references, the

carbonium ions were not found to be transition states but instead to be

energetically high-lying reaction intermediates. The authors claimed that these

carbonium ions exist as intermediates when the charge is delocalized and not

accessible to framework oxygen. The carbonium ions decompose directly into

the alkene and alkane, without forming alkoxides. Thus, the activation energies

are about a 100 kJ/mol lower than those calculated in the other mentioned

references, because covalent bonds do not have to be broken to reach the

transition state. Note that the activation energy is lowest in tertiary–tertiary

hydride transfer. In a study by Nowak et al. (65) activation energies for hydride

transfer between isobutane and tertiary and secondary acceptor cations were

compared with activation energies of isomerization steps between tertiary

carbenium ion species. The energy for tertiary–tertiary hydride transfer was

comparable to the energy of the isomerization, whereas the energy for tertiary–

secondary hydride transfer was almost twice as high.

Another study of ethane/ethene hydride transfer was performed to investigate

the influence of the Si/Al ratio and different levels of coverage of the acid sites

(66). The zeolite was modeled to represent the chabazite structure. It was found

that the electrostatic effects increase with decreasing Si/Al ratio, but they are

important only when the interaction between the zeolite and the adsorbed species

is clearly ionic. High coverage led to a destabilizing effect on the carbonium ions

due to repulsion between neighboring ions. The authors inferred that the

electrostatic forces are just one of many effects being of importance in zeolite-

catalyzed hydrocarbon reactions. Figure 4 summarizes the different calculated

potential energy profiles for the hydride transfer reaction in different media.

Experimental results characterizing hydride transfer in zeolites are scarce, as it

is a secondary reaction, which cannot be observed directly. Data from kinetics


measurements of cracking reactions of 2,2,4-TMP on USY zeolite gave values

for the apparent energies of activation of 47 kJ/mol lumped for all the hydride

transfer reactions that were occurring (67). A more detailed investigation of

isobutane cracking gave values of 64 kJ/mol for isobutane/propyl, 76 kJ/mol for

isobutane/n-butyl, and 62 kJ/mol for isobutane/isopentyl hydride transfer (40).

An earlier investigation by the same group led to higher values, namely, 81 kJ/

mol for isobutane/propyl, 67 kJ/mol for isobutane/n-butyl, and 125 kJ/mol for

isobutane/isopentyl hydride transfer (68). Even when average heats of adsorption

(ca. 40 kJ/mol) are added to the measured apparent energies to estimate the true

activation energies, these values are lower than the calculated values. Clearly, the

theoretical calculations overestimate the energy barrier. The overestimation is

speculated to be a consequence of incorrect modeling of the acid strength

(deprotonation energy, basicity of the lattice oxygen atoms) in the zeolitic cluster

used for the calculation.

It has been proposed that hydride transfer in zeolites requires the presence of

two adjacent Brønsted acid sites (69). In light of the above-mentioned theoretical

examinations and also adsorption isotherms of 1-butene and n-butane on USY

zeolites with various aluminum content (70), this proposition seems unlikely.

The reaction enthalpy of the hydride transfer step usually has a low absolute

value. Whether hydride transfer is exo- or endothermic depends on the stability

(evidenced by the heat of formation) of the involved carbenium ions. Branched

carbenium ions are more stable than linear ones. Longer carbenium ions are

more stable than shorter ones. Replacement of a long-chain carbenium ion by

Fig. 4. Potential energy profiles for the isobutane/t-butyl cation hydride transfer reaction in various

media (25,64).


a short-chain alkane to give a short-chain ion is endothermic, as exemplified by

the transfer of a hydride from isobutane to C8 carbenium ions.

With both liquid acid catalysts, but presumably to a higher degree with sulfuric

acid, hydrides are not transferred exclusively to the carbenium ions from

isobutane, but also from the conjunct polymers (44,46,71). Sulfuric acid

containing 4–6 wt% of conjunct polymers produces a much higher quality

alkylate than acids without ASOs (45). Cyclic and unsaturated compounds,

which are both present in conjunct polymers, are known to be hydride donors

(72). As was mentioned in Section II.B, these species can abstract a hydride from

isobutane to form the t-butyl cation, and they can give a hydride to a carbenium

ion, producing the corresponding alkane, for example the TMPs, as shown in

reactions (7) and (8).

ð7Þ

ð8Þ

In this way, the conjunct polymers serve as a reservoir of hydride ions. Under

some conditions, the polymers are a source of hydride ions, but they accept these

ions under other conditions. Substantial amounts of the saturated products are

supposedly formed via this route with sulfuric acid. In zeolites, species similar to

conjunct polymers also form. The heavy hydrocarbon molecules, which

deactivate the catalyst by pore blocking or by site blocking, are generally

termed “soft coke” or “low-temperature coke”, because of the absence of

aromatic species.

Only scant information is available about the influence of coke formation on

the alkylation mechanism. It has been proposed that, similar to the conjunct

polymers in liquid acids, heavy unsaturated molecules participate in hydride

transfer reactions. However, no direct evidence was given for this proposition

(69). In another study, the hydride transfer from unsaturated cyclic hydrocarbons

was deduced from an initiation period in the activity of NaHY zeolites; complete

conversion of butene was achieved only after sufficient formation of such

compounds (73).

In a series of investigations of the cracking of alkanes and alkenes on Y

zeolites (74,75), the effect of coke formation on the conversion was examined.

The coke that formed was found to exhibit considerable hydride transfer activity.

For some time, this activity can compensate for the deactivating effect of the

coke. On the basis of dimerization and cracking experiments with labeled 1-

butene on zeolite Y (76), it is known that substantial amounts of alkanes are

formed, which are saturated by hydride transfer from surface polymers. In both

liquid and solid acid catalysts, hydride transfer from isoalkanes larger than


isobutane may occur, especially from isopentane, which sometimes is used as

feedstock. However, no data are available providing information about the

significance of hydride transfer reactions with higher hydrocarbons.

Hydride transfer from alkenes was also proposed to occur during sulfuric

acid-catalyzed alkylation modified with anthracene (77). Then the butene loses a

hydride and forms a cyclic carbocation intermediate, yielding—on reaction with

isobutene—trimethylpentyl cations. This conclusion was drawn from the obser-

vation of a sharp decrease in 2,2,3-TMP selectivity upon addition of anthracene

to the acid.

Fast hydride transfer reduces the lifetime of the isooctyl cations. The molecules

have less time to isomerize and, consequently, the observed product spectrum

should be closer to the primary products and further from equilibrium. This has

indeed been observed when adamantane, an efficient hydride donor, was mixed

with zeolite H-BEA as the catalyst (78). When 2-butene/isobutane was used as

the feed, the increased hydride transfer activity led to considerably higher 2,2,3-

TMP and lower 2,2,4-TMP selectivities, as shown in Fig. 5.

Fig. 5. Changes in TMP selectivities with the use of adamantane (5 wt%) as an additive in a H-

BEA catalyst at 30 min TOS (P=O ¼ 10; OSV ¼ 0:2 h21, T ¼ 348 K) (78).


II.E. Oligomerization and Cracking

The overall product distribution is governed by the relative rates of alkene

addition and hydride transfer. With all acids, alkene addition is a much more

facile reaction than hydride transfer. With sulfuric acid, n-butene oligomerization

was found to be four times faster than hydride transfer (79). With zeolites, de

Jong et al. (80) reported oligomerization to be two orders of magnitude faster

than hydride transfer, whereas Simpson et al. (81) reported it to be three orders of

magnitude faster. With too low internal alkane/alkene ratios the alkenes will

oligomerize before they can be removed via hydride transfer. This is the key

problem in solid acid-catalyzed alkylation. A polymer will build up, which will

finally block the acid sites. With liquid acids, the conjunct polymers help in

maintaining a high hydride transfer activity. However, when the concentration

reaches a critical level, the acid strength will be too low for producing high-

quality alkylate. For this reason, in a continuous process, a stream of used acid

has to be constantly replaced by fresh acid to maintain the optimum level of acid

strength. The route to oligomerization products (sometimes also called multiple

alkylate) is depicted in Fig. 6. The rate constant kA defines the rate of alkene

addition, kB the hydride transfer rate, and kC the rate of deprotonation. The rate

ratio rB=ðrA þ rCÞ is the critical parameter that determines whether the catalyst

will effectively catalyze alkylation or deactivate quickly through multiple

alkylation/oligomerization reactions. High ratios can be achieved with low

alkene concentrations (as would be achieved in a backmixed reactor) and

maximized hydride transfer rates (a property of the catalyst).

Hydrocarbons with up to 16 carbon atoms are detected in a typical alkylate

(82). With the liquid acids, it was found that the oligomerization rate is higher for

isoalkenes than for linear alkenes (49). The same is true for solid acids (14,83).

Because of their tertiary carbon atoms, isobutylene and isopentene obviously

react more easily with carbenium ions. This point can be inferred from the reverse

reaction, b-scission (see below), which is fastest for reactions of tertiary cations

to give tertiary cations. In oligomerization experiments, the following pattern of

Fig. 6. Pathway to oligomerization products with the corresponding rate constants. Adapted from

Ref. (81).


reactivity of alkenes was found: isobutylene q n-butenes . propene . ethene.

This order can be readily explained by the relative stabilities of the carbenium

ions involved (84).

Not only are products with carbon numbers that are multiples of four are

formed, but so also are C5–C7 and C9, C10, and higher hydrocarbons. Cracking is

invariably associated with oligomerization. The heavy cations formed by

oligomerization have a tendency to fragment, forming C4–C16 cations and

alkenes, according to the b-scission rule, as depicted schematically in Reaction

(9) for a dodecyl cation cracking to give an isopentene and a heptyl cation.

ð9Þ

The isopentene produced will either be protonated or be added to another

carbenium ion. With a butyl cation, this would lead to a nonyl cation. The

resultant carbenium ion fragment can accept a hydride and form a product

heptane, or it can possibly add a butene to form a C11 cation. With hydride

transfer, another alkane with an odd number of carbon atoms is produced. Just

this example is sufficient to show the huge variety of possible reactions. By

means of gas chromatographic analysis, Albright and Wood (82) found about

100–200 peaks in the C9–C16 region, regardless of the alkene and acid

employed. A similar number of products can be observed for solid acid-catalyzed

alkylation.

In general, oligomerization and cracking products exhibit lower octane

numbers than the TMPs. Average research octane number (RON) values of 92–

93 for C5–C7 and of 80–85 for C9–C16 have been reported (8). Parts of the

octane fraction also stem from oligomerization/cracking reactions. It is believed

that substantial amounts of the dimethylhexanes are produced via this route (79),

especially when isobutylene is the feed alkene (71). Isobutylene tends to

oligomerize quickly. Hence, it produces higher amounts of light and heavy ends

and cannot isomerize to 1-butene to produce DMHs in this way. Some of the

TMPs also will be produced through oligomerization/cracking pathways (20).

Concentrations of more than 20 wt% of TMPs in the C6þ fraction have been

observed in isopentane/2-pentene alkylation (53). The TMPs cannot be produced

via simple alkylation or self-alkylation with this feed. It has been proposed that

oligomerization/cracking constitutes the main route to alkylation products (16),

but this proposition fails to explain the usually high selectivity to the TMPs. To

form trimethylpentanes, some specific precursors would have to build up in high

concentrations, which is rather unlikely.

Hydrocracking experiments under ideal conditions provided kinetics informa-

tion characterizing the b-scission step. On the basis of this work, a classification


of various types of b-scission has been introduced (85). Fragmentations starting

from a tertiary carbenium ion and giving a tertiary cation (type A) are very rapid.

Fragmentations involving secondary and tertiary ions (type B) are slower than

tertiary–tertiary b-scissions, but faster than secondary–secondary b-scissions

(type C). The slowest mode is the cracking of a secondary ion to give a primary

ion (type D). From the typical low reaction temperatures and the product

composition of a typical alkylate, which consists almost exclusively of branched

hydrocarbons, it can be concluded that only type A b-scissions occur at

significant rates. Furthermore, protolytic cracking of alkanes via a carbonium ion

mechanism is highly unlikely under typical alkylation conditions. Hydrogen or

methane, which are characteristic products of such cracking, are not found in the

alkylate. At low temperatures, the cracking of alkanes is initiated by traces of

alkenes in the feed (also see Section II.B).

In general, oligomerization is an exothermic reaction (and therefore the

reverse, b-scission, is an endothermic reaction). Quantum-chemical calculations

of the b-scission step on a zeolite represented by a cluster model were performed

to estimate activation energies. For tertiary–secondary fragmentations, values

in the range of 234–284 kJ/mol and for secondary–secondary values in the range

of 288–314 kJ/mol (32) and 217–275 kJ/mol (86) were calculated. Here, the

activation energy of the reverse reaction was reported to be 71 kJ/mol less than

that of the forward reaction. Evaluation of alkane conversion experiments with

USY zeolite as catalyst, in general, provided much lower values than these (40,

67); average apparent activation energies for secondary–tertiary and secondary–

secondary b-scission steps were estimated to be approximately 115 kJ/mol. The

values for tertiary–tertiary b-scission given in the two references differed

between 66 and 102 kJ/mol. In an older study by the same authors (68), values for

b-scission and oligomerization were given; tertiary–tertiary b-scission was

characterized by an activation energy of 184 kJ/mol and the reverse reaction by a

value of 105 kJ/mol. Tertiary–secondary b-scission was found to be character-

ized by an activation energy of 84 kJ/mol and the reverse reaction by a value of

71 kJ/mol. The corresponding values for secondary–secondary b-scission were

found to be 130 and 33 kJ/mol. As for hydride transfer, the calculated values are

significantly greater than the measured values (plus the heat of adsorption),

presumably as a consequence of an underestimation of the acid strength.

II.F. Self-Alkylation

With hydrofluoric acid (23,50), and to a lesser degree also with zeolites (14,81,

87–89), a significant fraction of the product stems from self-alkylation, which is

sometimes also termed hydrogen transfer. The importance of this mechanism

depends on the acid, the alkene, and the reaction temperature. Self-alkylation


reactivity increases with molecular weight and the degree of branching of the

feed alkene (90). Generally, sulfuric acid is less active for self-alkylation than

hydrofluoric acid. Only when pentenes or higher alkenes are used is self-

alkylation significant with sulfuric acid (49,91). In Fig. 7, the mechanism is

displayed for an exemplary isobutane/2-butene feed.

The crucial step in self-alkylation is decomposition of the butoxy group into a

free Brønsted acid site and isobutylene (proton transfer from the t-butyl cation to

the zeolite). Isobutylene will react with another t-butyl cation to form an isooctyl

cation. At the same time, a feed alkene repeats the initiation step to form a

secondary alkyl cation, which after accepting a hydride gives the t-butyl cation

and an n-alkane. The overall reaction with a linear alkene CnH2n as the feed is

summarized in reaction (10):

2i-C4H10 þ CnH2n ! i-C8H18 þ CnH2nþ2 ð10Þ

With propene, n-butene, and n-pentene, the alkanes formed are propane,

n-butane, and n-pentane (plus isopentane), respectively. The production of

considerable amounts of light n-alkanes is a disadvantage of this reaction route.

Furthermore, the yield of the desired alkylate is reduced relative to isobutane

and alkene consumption (8). For example, propene alkylation with HF can give

more than 15 vol% yield of propane (21). Aluminum chloride–ether complexes

also catalyze self-alkylation. However, when acidity is moderated with metal

chlorides, the self-alkylation activity is drastically reduced. Intuitively, the

formation of isobutylene via proton transfer from an isobutyl cation should be

more pronounced at a weaker acidity, but the opposite has been found (92). Other

properties besides acidity may contribute to the self-alkylation activity. Earlier

publications concerned with zeolites claimed this mechanism to be a source of

hydrogen for saturating cracking products or dimerization products (69,93).

However, as shown in reaction (10), only the feed alkene will be saturated, and

dehydrogenation does not take place.

Fig. 7. Self-alkylation mechanism, depicted with 2-butene as the feed alkene.


II.G. Product and Acid Degradation

It has been found that C7–C9 isoalkanes react with strong acids to produce a

low-quality alkylate and conjunct polymers (94). In the presence of conjunct

polymers, highly branched isoalkanes might re-enter the reaction cycle by the

reverse of reaction (8). Oligomerization/cracking will then lead to inferior

products. This problem affects alkylation by both HF and H2SO4. It is unclear

whether this side reaction is of importance with zeolites under alkylation

conditions. On the zeolite H-FAU at temperatures as low as 373 K, 2,2,4-TMP

undergoes cracking into isobutane and isobutylene, with significant coke

formation (95).

A problem that is characteristic of sulfuric acid-catalyzed alkylation is its

capability to oxidize hydrocarbons. H2SO4 decomposes in the presence of

isoalkanes to form water, SO2, and alkenes. This is a slow process, and so it

occurs predominantly when the acid is in contact with hydrocarbons for a longer

period. Higher temperatures favor the formation of SO2 (10). Some irreversible

reactions between acid and hydrocarbons also take place during alkylation.

Sulfone, sulfonic acid, and hydroxy groups have been detected in conjunct

polymers produced with H2SO4 as the catalyst (8,96). Kramer (97) reported that

2,3,4-TMP, after an induction period, is converted into a mixture of lower alkanes

(with a high fraction of isobutane) and isomerized octanes. The reaction was

initiated by the reduction of sulfuric acid to SO2 with the formation of carbenium

ions. In a subsequent paper by Kramer (98), more information about the reaction

of selected branched alkanes with sulfuric acid led to the conclusion that SO2 is

produced only during the initiation reaction. All subsequent reactions are

conventional carbenium ion type reactions. Alkanes with a higher degree of

branching show higher rates of degradation. Pure isobutane was found to react

with sulfuric acid at 298 K. The acid was slowly reduced to SO2, with isobutane

forming carbenium ions undergoing subsequent reactions. With traces of alkenes

in the feed, however, acid reduction was not observed (99).

II.H. Pathways to Allylic and Cyclic Compounds

The conjunct polymers formed during liquid-phase alkylation contain single and

conjugated double bonds and five- and six-rings. The residue on zeolitic catalysts

is highly branched, containing double bonds and conjugated double bonds and

possibly also five- and six-rings (73,88,100,101). The H/C ratio is about 1.8

(102), similar to that of conjunct polymers. In general, it is believed that at

temperatures below 473 K, coking of acidic catalysts mainly involves conden-

sation and rearrangement steps. Aromatic compounds are usually not formed

under such mild conditions (95). Extending these results to typical alkylation


reaction conditions, we expect that several alkene molecules will oligomerize and

crack or deprotonate to form a large and branched alkene. This alkene might

transfer a hydride to another carbenium ion and thus form an alkenyl carbenium

ion, which can desorb via proton transfer as a diene (Reaction (11)). Further

hydride transfer leads to a dienylic cation, which easily rearranges into an alkyl-

substituted ring (Reaction (12)) via a 1,5-cyclization and subsequent hydride and

methyl shifts.

ð11Þ

ð12Þ

The resultant cycloalkenyl carbenium ions, especially the cyclopentenyl

cations, are very stable (103,104) and can even be observed as free cations

in zeolites (105,106). These ions can oligomerize further and, within zeolites,

irreversibly block the acidic hydroxyl groups. With liquid acids, the oligomers

will dilute the acid and thus lower its acid strength.

II.I. Summary

Figure 8 summarizes the main reactions occurring during alkylation. Dimeriza-

tion and oligomerization reactions are more important with zeolitic catalysts on

acidic sites with lower acid strengths (Section III.B.2) or with severely diluted

liquid acids (Section III.A). Hydride transfer from conjunct polymers is more

important with sulfuric acid, and self-alkylation activity is more significant with

hydrofluoric acid. Repeatedly going through the alkylation cycle without hydride

transfer (multiple alkylation) and through the dimerization cycle without proton

transfer (oligomerization) leads to the formation of heavy compounds, which will

react further via cracking, hydride or proton transfer, and cyclization. As long as

the catalyst shows sufficient hydride transfer activity, all alkenes will react, and

only saturated products will leave the reaction cycles.

III. Physical–Chemical Phenomena Influencing the Reaction

As was pointed out, the chemistry of the alkylation reaction can be explained by a

set of mechanistic steps that are similar and in some cases the same for all

the different acids examined. However, the importance of each step varies with


the catalyst and reaction conditions. The understanding of these parameters is

thus of utmost importance. This is especially true for the solid acid catalysts.

They can be synthesized and modified in a nearly infinite number of ways to

influence in a complex and subtle manner the alkylation performance. In Section

III.A, the chemical and physical properties of the individual alkylation catalysts

and how they affect the mechanism are reviewed, and concomitantly the

influence of process parameters, such as temperature, alkane/alkene ratio, and

residence time on the reaction is assessed.

III.A. Properties of Liquid Acid Alkylation Catalysts

In the liquid acid-catalyzed processes, the hydrocarbon phase and the acid phase

are only slightly soluble in each other; in the two-phase stirred reactor, the

hydrocarbon phase is dispersed as droplets in the continuous acid phase. The

reaction takes place at or close to the interface between the hydrocarbon and

the acid phase. The overall reaction rate depends on the area of the interface.

Larger interfacial areas promote more rapid alkylation reactions and generally

result in higher quality products. The alkene is transported through the

hydrocarbon phase to the interface, and, upon contact with the acid, forms an

acid-soluble ester, which slowly decomposes in the acid phase to give a solvated

Fig. 8. Concerted alkylation mechanism including alkylation, “self-alkylation”, cracking,

dimerization, and hydride transfer via isobutane and via conjunct polymers.


carbenium ion or the alkene. Isobutane can react at the interface or be transported

into the acid phase and react there. The most important parameters determining

the ease of formation of a large reaction zone are the viscosity and the solubility

of hydrocarbons in the acid. These properties differ substantially for sulfuric and

hydrofluoric acid.

Under typical alkylation conditions, the viscosity of sulfuric acid is two orders

of magnitude higher than that of hydrofluoric acid, and the solubility of isobutane

is approximately 30 times lower. The relatively high solubility of isobutane in

HF, together with a high interfacial area, ensures high isobutane/alkene ratios

in the acid and, thus, high hydride transfer rates and relatively low selectivity

for the formation of undesired products from oligomerization/cracking and

isomerization and for the formation of conjunct polymers. Consequently, sulfuric

acid/hydrocarbon phases have to be mixed much more vigorously than hydro-

fluoric acid/hydrocarbon phases to obtain a high-quality alkylate. For the same

reason, hydrofluoric acid-catalyzed processes can operate at lower residence

times and higher temperatures than sulfuric acid-catalyzed processes. Using

sulfuric acid with isobutane/2-butene in a laboratory reactor, Li et al. (107) found

that increasing the agitator speed from 1000 to 3000 rpm increased the product

RON from 86 to 94. Albright (11) discerned a minimum of four types of droplets

in acid/hydrocarbon dispersions. The droplets differ in size and in the concen-

trations of reactants and products. The formation (and the separation) of acid/

hydrocarbon emulsions depends on the temperature, the composition of the acid,

and the acid/hydrocarbon ratio (108).

Sulfuric acid is a somewhat stronger acid than hydrofluoric acid. The values of

the Hammett acidity function H0 for the water-free acids are 214.1 for H2SO4

and 212.1 for HF. It is, however, interesting to note that the maximum alkylate

quality obtained with sulfuric acid is not achieved with the highest acidity, but

with acid containing 1–1.5 wt% water and 4–5 wt% ASOs (96). Water reduces

the acidity to a greater extent than hydrocarbon diluents. Besides their hydride

transfer capabilities, the ASOs act as surfactants, increasing the interfacial area.

When the concentration of diluents exceeds a certain level, the acid strength

is too low to produce a high-quality alkylate. Sulfuric acid of 60–80 wt%

concentration catalyzes only alkene oligomerization. The acid strength is too low

for catalysis of the more demanding reactions hydride transfer and b-scission

(27). A relatively sharp transition between oligomerization and alkylation

activity has been measured with sulfuric acid at H0 values between 28.0 and

28.5 (109). If such low-acidity values occur in an alkylation reactor, oligomer-

ization reactions become so predominant that the acid strength cannot be

maintained and the plant is said to be in an acid runaway condition.

The same principles regarding the acidity can be applied to hydrofluoric acid-

catalyzed alkylation, which is more sensitive towards water, so that the feed must

be thoroughly dried before entering the reactor. Furthermore, the acid dilution by


hydrocarbons is greater as a consequence of their higher solubility in HF (15).

Employing triflic acid modified with water or trifluoroacetic acid, Olah et al.

(110) found the best alkylation conditions at an acid strength of about H0 ¼

210:7 for both systems. Pure triflic acid (with H0 ¼ 214:1) produced mainly

cracked compounds. Diluted triflic acid with H0 . 210:7 favored oligomeriza-

tion. The same research group tested different liquid acids diluted with liquid

carbon dioxide. Although very strong acids such as triflic acid produce higher

quality alkylate upon dilution with CO2, sulfuric acid (being less strongly acidic

than triflic acid) performed better without CO2 (111). The different H0 values

observed for the transition from alkylation to oligomerization with sulfuric and

triflic acid suggest that the acid strength is not the only factor determining the

reactivity of the carbenium ions.

III.B. Properties of Zeolitic Alkylation Catalysts

Zeolite molecular sieves are widely used as solid acid catalysts or catalyst

components in areas ranging from petroleum refining to the synthesis of

intermediates and fine chemicals (112,113). An important reason for their

widespread use is the flexibility they offer regarding the tailoring of the

concentration and nature of catalytically active sites and their immediate

environments. We note that discrimination between chemical and structural

aspects works well at a conceptual level, but one faces quite severe limitations as

soon as one tries to separate the contributions of the two effects. The complexity

arises because the chemical properties of a particular molecular sieve are

connected with its framework density.

III.B.1. Adsorption and Diffusion of Hydrocarbons

One of the major characteristics of acidic zeolites that sets them apart from

the liquid acids is their selective and strong chemisorption of unsaturated

compounds. Because of the high polarity of the zeolitic surface, especially in

aluminum-rich zeolites, polar molecules will be preferentially adsorbed. This

property is clearly demonstrated by the high water uptake capacity of zeolite X,

which exceeds 25 wt%. Furthermore, the electrostatic field in the zeolite pores

enhances the adsorption of polarizable molecules (114). Thus, although the

concentration of alkenes in the liquid phase might be low, they will preferentially

adsorb in the zeolite pores, so that in the pore system the alkene concentration

will be considerably higher resulting in much higher relative rates of oligo-

merization vs. hydride transfer, as was discussed in Section II.E. This selective

adsorption is the major reason why zeolites deactivate rapidly if no special

measures are taken to minimize the alkene concentration close to the acidic sites.


Nevertheless, the adsorption of alkenes can differ substantially from one zeolite

to another, even for one type of zeolite, depending on the concentration of

framework aluminum and the modification procedure (70).

Also typical for molecular sieves is the increasing heat of adsorption of

hydrocarbons with increasing chain length (115). Each carbon atom contributes

equally to the total heat of adsorption. This value depends on the zeolite pore size

and shape, so that different adsorption enthalpies are measured for different

zeolites. Increasing framework density (number of T-atoms per unit volume,

where T refers to Si or Al) leads to increased heats of adsorption (116,117).

Protons add another constant value (which depends on the chemical composition)

to the overall heat of adsorption, as represented in Fig. 9A and B. This

phenomenon is responsible for different apparent activation energies for a given

reaction type found with hydrocarbons of different chain lengths. The actual

intrinsic activation energies (as well as the corresponding pre-exponential

factors) are nearly independent of chain length (118). Assuming the relationship

between chain length and adsorption enthalpy to be linear over a wide range,

relative desorption rates for various hydrocarbons can be calculated for a given

temperature. Thus, using the data for H-FAU and a temperature of 348 K,

the desorption of a C12 molecule is four orders of magnitude slower than that of

an C8 molecule, and that of a C16 is eight orders of magnitude slower and that of

a C20 12 orders of magnitude slower (Fig. 9C). These huge differences give one

a sense of the difficulties of removing heavy products from the zeolite surface

using purely adsorption/desorption arguments. Once such a heavy molecule has

formed, it is unlikely to desorb.

III.B.2. Brønsted Acid Sites

Zeolites exhibit a considerably lower proton (acid site) concentration than liquid

acids. For example, 1 g of H2SO4 contains 20 £ 1023 moles of protons, whereas

1 g of zeolite HY, with a Si/Al atomic ratio of five, contain no more than 3 £ 1023

moles of protons. (Note that this is a crude approximation of the acidic sites

available for catalysis, because it assumes that with both materials all protons are

available and catalytically active.) Moreover, 1 g of H2SO4 occupies far less

volume (i.e., 0.5 cm3) than the equivalent mass of zeolite (4–6 cm3).

In contrast to liquid acids, zeolites encompass different populations of sites that

differ substantially in their nature and strength. Liquid acids with a given

composition have a well-defined acid strength. This is not the case for zeolites.

Depending on the type of zeolite, its aluminum content, and the exchange

procedure used in its preparation, Brønsted and Lewis acid sites with a wide

range of strengths and concentrations are present. To summarize the effects of all

parameters influencing the acidity of zeolites is beyond the scope of this review.


Fig. 9. Effect of the chain length of hydrocarbons on the adsorption enthalpy and rates of desorption. (A) Hydrocarbon in interaction with zeolite

framework. Methyl groups interact with the framework oxygen; protons exhibit an additional attractive force. (B) Heat of adsorption as a function of carbon

number for zeolites MFI and FAU in the acidic and non-acidic form. (C) Relative desorption rates of a C12, C16, and C20 alkane compared to octane at

348 K. Values calculated from the linear extrapolation of the heat of adsorption values shown in (B).

A.

Feller

and

J.A

.L

ercher

/A

dv

.C

atal.4

8(2

00

4)

22

9–

29

52

57

The different reaction steps in alkylation require different minimum acid

strengths to be effectively catalyzed. Double bond isomerization is catalyzed

even by weak acid sites. Even a fully deactivated zeolite retains some activity for

isomerization of butenes (119,120). Dimerization/oligomerization also does not

require strong acidity, as was concluded from a study of a series of USY zeolites

with different unit cell sizes. Correlations between the acidity and the alkylation

performance revealed that the acid strength required for performing the different

reactions occurring during alkylation decreases in the order cracking .

alkylation (addition of butene to a tertiary butyl) . dimerization (addition of a

butene to a secondary butyl) (121). A comparison between the iso-structural H-

SAPO-37 and H-FAU as alkylation catalysts showed that the H-FAU has a much

higher relative concentration of strong acid sites than the H-SAPO-37. Therefore,

the H-SAPO-37 mainly catalyzed dimerization, with a small amount of 3,4-DMH

as the most abundant saturated compound, whereas the H-FAU produced mainly

TMPs (122).

The lifetime of a zeolitic alkylation catalyst depends on the concentration of

Brønsted acid sites. This has been shown by Nivarthy et al. (78), who used a

series of zeolites H-BEA with varied concentrations of back-exchanged sodium

ions. The sodium decreased the concentration of Brønsted acid centers, which led

to a concomitant decrease in the measured catalyst lifetime during alkylation.

However, there are contrasting opinions about the acid strength required for

optimum alkylation performance with zeolites. Hydride transfer is the step that

determines the product quality and the catalyst lifetime. Thus, it is crucial to

know which conditions favor a high hydride transfer rate. From the above-

presented investigations, it can be concluded that stronger sites are necessary to

effectively catalyze hydride transfer. Stocker et al. (123) synthesized and tested

EMT and FAU samples with enhanced Si/Al ratios of 3.5 (made by use of crown

ethers as templates). They explained the better performance of H-EMT relative to

that of H-FAU by its higher ratio of strong-to-weak Brønsted acid sites.

Dealumination of the H-FAU led to better results because of additional small

numbers of very strong acid sites. No direct demonstration was given to support

this opinion (124). La-exchange of H-EMT led to a slightly better performance

than that of H-EMT. This improvement was also attributed to a higher ratio of

strong-to-weak Brønsted acid sites (125). A similar conclusion was derived by

Corma et al. (126), who compared USY, MOR, BEA, ZSM-5, and MCM-22. The

relative decrease of activity for the formation of TMPs with time on stream was

observed to depend on the concentration of strong Brønsted acid sites in the fresh

zeolite. Diaz-Mendoza et al. (127) studied commercial REY, USY, and BEA

samples. In contrast to the aforementioned groups, they found Brønsted acid sites

with intermediate strengths to be the appropriate sites for maintaining good

alkylation performance.


It is well established that with time on stream the sites strong enough for

catalyzing hydride transfer deactivate first. In the first phase of operation and

deactivation, the catalyst produces a mixture of saturated isoalkanes, usually with

(nearly) complete butene conversion, and in the second phase, alkenes, mainly

octenes, are produced at a substantially lower butene conversion. The product in

this second phase resembles the product observed with the weakly acidic H-

SAPO-37. The mixture of butene isomers found in the product stream in the

second phase is close to the equilibrium composition. A typical example is

depicted in Fig. 10. Several investigations of zeolite- and other solid acid-

catalyzed alkylation obviously have been performed in the oligomerization

regime (129–134). As a consequence of insufficient acidity or an inappropriate

choice of reaction conditions, the catalysts examined in these investigations

produced mainly oligomerization products and only small amounts of true

alkylate. Unequivocal conclusions can be drawn neither about the alkylation

mechanism nor on the true alkylation activity of the tested materials under these

circumstances.

The characterization method employed in nearly all the above-mentioned

investigations for measuring the strength of acid sites was pyridine adsorption/

desorption monitored by IR spectroscopy. Pyridine forms the pyridinium ion

on Brønsted acid sites and binds to Lewis acid sites by forming coordination

complexes. Heating the sample with the adsorbed pyridine gives characteristic

desorption curves for pyridine bound to Brønsted or Lewis acid centers. From

such data, Brønsted acid/Lewis acid site ratios at a given temperature and

“strong” to “weak” acid site ratios can be calculated and correlated with

Fig. 10. Typical time on stream behavior of a CeY zeolite alkylated in a fixed-bed reactor (128).


the catalytic performance. Strong and weak acid sites here are defined by the

amount of pyridine that remains adsorbed at a certain temperature—the higher

the temperature, the stronger the bonding and the stronger the acid site. These

temperatures are chosen in a rather arbitrary manner, the upper limit of which is

typically restricted by the IR cell and the zeolite itself (673–823 K). The strong

acid sites are not of uniform strength, likely exhibiting a broad distribution, up to

a minute amount of “very strong” acid sites, which are difficult to detect because

of their low concentrations. Whether such very strong acid sites are responsible

for an enhanced alkylation activity is not yet determined.

III.B.3. Lewis Acid Sites and Extra-Framework Aluminum

Lewis acid sites in zeolites originate from a partial destruction of the framework.

During the modification procedure, which is necessary to transform the parent

material into its acidic form, part of the aluminum present in the framework is

removed from its positions in the crystalline framework (usually during

calcination in a water-containing atmosphere, i.e., high-temperature hydrolysis

of Si–O–Al bonds) to give extra-framework aluminum species (EFAL). Some of

the species formed in this way exhibit Lewis acidity. Another source of Lewis

acid sites is metal ions in cation-exchange positions. However, most of these

metals exhibit weaker Lewis acidity than aluminum species. Lewis acid sites do

not catalyze the alkylation reaction, but their presence undoubtedly influences the

performance of zeolitic catalysts in alkylation. It has been claimed that the

presence of strong Lewis acid sites promotes the formation of unsaturated

compounds (127). The favored production of unsaturated carbenium ions with

increased Lewis acidity was also evidenced by Flego et al. (100) investigating the

deactivation of a La–H-FAU zeolite in isobutane/1-butene alkylation. Increasing

the catalyst activation temperature led to higher Lewis acid site concentrations,

which increased the formation of mono- and di-enylic carbenium ions. Besides

the ability to increase the rate of formation of unsaturated compounds, Lewis acid

sites have been claimed to increase the alkane/alkene ratio close to the Brønsted

acid sites, through the adsorption/desorption equilibrium of the alkenes on the

Lewis acid sites. The increased alkene concentration accelerates oligomerization

and leads to premature catalyst deactivation (78). EFAL species also influence

the acidity of neighboring Brønsted acid sites. Corma et al. (135) examined

zeolite H-BEA, which they had exposed to several post-synthesis treatments to

change the framework and extra-framework composition. On the basis of the

combined reaction and characterization data, the authors concluded that some

cationic EFAL species compensate the framework charge, and other condensed

EFAL species block a fraction of the Brønsted acid sites, thus lowering their

concentration. On the other hand, these authors suggested a synergistic effect of

dispersed cationic EFAL species and framework hydroxyls to form Brønsted acid


sites of enhanced strength. A further study by the same group (136) showed that

in samples with a high framework aluminum concentration, the removal of EFAL

was detrimental to the catalytic performance, whereas in the samples with low

framework aluminum content the catalytic activity increased as a result of the

removal of EFAL. The fact that mild steaming enhances the strength of Brønsted

acid sites is known from other hydrocarbon reactions, such as cracking and

isomerization of alkanes and disproportionation of toluene. Selective poisoning

of Brønsted acid sites with cesium has shown that only a minute amount of very

strong sites is present in mildly steamed samples. However, these sites are

responsible for a drastic increase in activity (118). Residual sodium also exhibits

a poisoning effect on very strong Brønsted acid sites. Small amounts of sodium

were found to eliminate highly acidic centers created by the interaction of EFAL

with protonic sites (137).

III.B.4. Silicon/Aluminum Ratio

The influence of the Si/Al ratio on the catalytic performance is discussed

primarily in terms of effects of changes in the concentration and acid strength of

the protonic sites. The electrostatic forces induced by the presence of framework

aluminum are often neglected. With increasing aluminum concentration in the

framework (i.e., with lower Si/Al ratio), the total concentration of protonic acid

sites increases. On the other hand, it is believed that the strengths of the acid sites

decrease with increasing aluminum concentration. At high aluminum concen-

trations the thermal stability of the zeolites in their protonic forms is also reduced,

facilitating the formation of extra-framework species (138). Examining a series

of ultrastable Y zeolites, Corma et al. (121) found the catalyst with the lowest Si/

Al ratio to be best in time-on-stream behavior and TMP selectivity. With a

decreasing Si/Al ratio, the ratio of stronger to weaker acid sites increased and was

correlated with the alkylation/oligomerization selectivity ratio measured with the

samples. The same trend was found by de Jong et al. (80), who also tested a series

of ultrastable Y zeolites in a semi-batch reactor. These authors also tested a

zeolite BEA with a Si/Al ratio of 15 that performed better than the Y zeolites.

They postulated that a decrease of the Si/Al ratio in BEA also should lead to a

superior catalyst associated with a higher Brønsted acid site concentration.

Weitkamp and Traa (139) also accentuated this hypothesis.

Some investigations have focused on the influence of the Si/Al ratio in zeolite

BEA. Corma et al. (140) used various BEA samples synthesized with different

Si/Al ratios and found a higher thermal stability towards dealumination with

increasing Si/Al ratio. The most stable catalyst was also the most active one.

Weitkamp et al. (141) compared the selectivities of four H-BEA samples with

Si/Al ratios ranging from 12 to 90. The octane number selectivities ran through

a maximum at a Si/Al ratio of 19, whereas the TMP/DMH ratio decreased


continuously with the Si/Al ratio. Loenders et al. (142) tested BEA samples

with framework Si/Al ratios ranging from 13 to 77, reporting that the individual

acid sites perform an identical number of catalytic turnovers before

deactivation, independent of the acid site density. They claimed that the only

way to enhance the activity and stability of zeolite BEA for isobutane alkylation

is to increase the aluminum content of BEA nanoparticles. The only reported

investigation of zeolite BEA with a Si/Al ratio lower than nine was performed

by Yoo and Smirniotis (143), utilizing H-BEA synthesized with Si/Al ratios

between 6 and 30. In contrast to what was postulated earlier, the zeolites

exhibited a maximum in the catalytic lifetime when the Si/Al ratio was between

8.5 and 15. The hydrogen transfer activities measured separately with n-hexane

as the reactant were comparable for all the samples up to a Si/Al ratio of 15.

The authors concluded that the BEA with the highest aluminum content

performed worse than the other samples with the same hydride transfer activity

as a consequence of the lower crystallinity and micropore volume, which were

inherent to the synthesis procedure for aluminum-rich zeolite BEA. In a patent

assigned to Mobil Oil (144), three BEA samples, with Si/Al ratios of 7.3, 16.0,

and 18.5, were compared. The only detail given about the alkylation perfor-

mance was the TMP/(C8-TMP) ratio, which was observed to increase with

decreasing Si/Al ratio, which is suggestive of the superiority of the material

with the low Si/Al ratio.

Notwithstanding some obviously contradictory results in the literature, the data

summarized above can be summarized as follows: the general trend is that high

aluminum contents are beneficial for the alkylation performance. This inference

is supported by results from cracking experiments with zeolites having various

Si/Al ratios. The bimolecular hydride transfer step is favored in materials with

low Si/Al ratios (54,145,146). Thus, zeolites with low Si/Al ratios should exhibit

better time-on-stream behavior than those with high Si/Al ratios.

Zeolite X is the large-pore zeolite with the highest aluminum content

possible. The first investigations of zeolite-catalyzed alkylation were done on

this material (13,147). Weitkamp, comparing highly cerium-exchanged Y and

X zeolites, found the CeX zeolite to exhibit twice the lifetime of CeY zeolite as

a consequence of the higher concentration of acid sites (148). In light of these

findings, it is surprising that only a small number of investigations have been

devoted to this material. As the purely protonic form of zeolite X is unstable,

polyvalent metal ions have to be introduced to induce acidity (Section III.B.5).

A variety of di- and tri-valent metals have been examined, with and without

additional ammonium exchange (149–151). Rare earth elements, especially

lanthanum, obviously are best suited to the goal, producing highly acidic and

thermally stable catalysts. LaCaX zeolite has also been proposed as an

excellent isobutane/ethene alkylation catalyst (152,153). Falsely, the authors


attributed the excellent performance to superacidic centers with a narrow

acidity distribution.

III.B.5. Metal Ions in Ion-Exchange Positions

Rare earth exchanged faujasites (REHY and REUSY) are widely used in the

FCC process (138). Aqueous ion exchange with rare earth salts in faujasites

leads to removal of ions in the supercage only, because a bulky hydration

sphere around the ions is larger than the six-membered ring of the sodalite cage,

so that the ions do not enter these cages. Calcination removes the waters of

hydration, and the naked cation is able to move into the sodalite cage, forming

cationic polynuclear hydroxy complexes (154–156). These species impart

thermal and hydrothermal stability to the material. Rare earth exchanged

zeolites exhibit considerable Brønsted acidity resulting from hydrolysis of the

hydrated rare earth ions (157,158). This principle works with most polyvalent

metals, and the rare earth elements induce the highest acidity and best stability

(159–161).

Besides zeolite X (discussed in Section III.B.4), zeolite Y is the one that has

been the subject of most investigations of cation exchange. Researchers at Sun

Oil Company extensively explored rare earth exchanged Y zeolites (14). On the

basis of their work described in patents (151,162–166), it can be concluded that

partially rare earth exchanged faujasites are more active catalysts than the purely

protonic forms. The importance of quantitative removal of sodium from zeolite

was demonstrated. Chu and Chester (119,120) compared variously modified Y

zeolites; REHY zeolite (RE is rare earth) gave the highest yield and the best

product quality. Dealumination of REHY zeolite did not improve its

performance. USY and REUSY zeolites were both characterized by low

conversion and yield, and there were not significant differences between the two.

In their work on EMT and FAU zeolites, the SINTEF group (125,167) compared

H- and La-exchanged samples and showed that a partially La-exchanged catalyst

is superior to both fully La-exchanged and pure H-form samples. H-EMT

contains the highest total number of Brønsted acid sites as measured with

pyridine adsorbed at 423 K. The partially La-exchanged sample (51%

exchanged) has twice as high a concentration of strong Brønsted acid sites (as

measured by the pyridine retained at 823 K) as the pure H-EMT and also a lower

concentration of Lewis acid sites. The increase in acid strength has been

rationalized by a withdrawal of electrons from the Lewis-base framework oxygen

atoms through polyvalent lanthanum cations in the sodalite cages. This electron

withdrawal effect is supposed to be similar to the action of EFAL species

in steamed zeolites. The abstraction of electrons weakens the O–H bond and

thus increases the proton-donor strength of the OH group (156). In a patent

assigned to Mobil Oil (168), rare earths exchanged into zeolite ZSM-20


(intergrowth between FAU and EMT) are claimed to improve conversion and

selectivities to TMPs. REZSM-20 was also claimed to perform better than

REHY. In another Mobil patent (169), REY and REUSY zeolites were compared;

the REY zeolite exhibited a slightly higher alkylate quality, whereas the REUSY

zeolite gave a slightly higher conversion. The subtle differences in the reports are,

however, difficult to evaluate as detailed characterizations of the materials are

lacking most of the time. In light of the strong influence of the concentration of

Brønsted and Lewis acid sites, a judgment about which of the materials has the

best properties is not possible.

III.B.6. Structure Types of Zeolites

Only large-pore zeolites exhibit sufficient activity and selectivity for the

alkylation reaction. Chu and Chester (119) found ZSM-5, a typical medium-pore

zeolite, to be inactive under typical alkylation conditions. This observation was

explained by diffusion limitations in the pores. Corma et al. (126) tested HZSM-5

and HMCM-22 samples at 323 K, finding that the ZSM-5 exhibited a very low

activity with a rapid and complete deactivation and produced mainly dimethyl-

hexanes and dimethylhexenes. The authors claimed that alkylation takes place

mainly at the external surface of the zeolite, whereas dimerization, which is less

sterically demanding, proceeds within the pore system. Weitkamp and Jacobs

(170) found ZSM-5 and ZSM-11 to be active at temperatures above 423 K. The

product distribution was very different from that of a typical alkylate; it contained

much more cracked products; trimethylpentanes were absent; and considerable

amounts of monomethyl isomers, n-alkanes, and cyclic hydrocarbons were

present. This behavior was explained by steric restrictions that prevented the

formation of highly branched carbenium ions. Reactions with the less branched

or non-branched carbenium ions require higher activation energies, so that higher

temperatures are necessary.

MCM-22, with a larger pore volume than ZSM-5, revealed behavior inter-

mediate between what was observed for large- and medium-pore zeolites (126).

Unverricht et al. (141) also examined MCM-22; at 353 and 393 K, it was found to

produce mainly cracked products and dimethylhexanes and to deactivate rapidly.

MCM-36 gained considerable interest that is evidenced by the patent literature

(171–174). MCM-36 is a pillared zeolite based on the structure of MCM-22.

Ideally, it should contain mesopores between layers of MCM-22 crystallites. This

structure was found to be much more active and stable than MCM-22 (175).

Alkane cracking experiments with zeolites having various pore dimensions

evidenced the preference of monomolecular over sterically more demanding

bimolecular pathways, such as hydride transfer, in small- and medium-pore

zeolites (146).


In contrast to the product distributions observed for medium-pore zeolites, the

product distributions observed for large-pore zeolites resemble those of typical

alkylates. However, within the distribution, significant differences are observed.

It is difficult to separate the influence on the alkylation reaction of the structure

from the influences of other properties, mainly the acid site strength and

concentration. Undisputable results may be achieved only if all but one

parameters are held constant. Yoo et al. (176) compared USY, BEA, MOR, LTL,

and ZSM-12 zeolites with Si/Al ratios between 20 and 34 (achieved either by

direct synthesis or by various leaching techniques) and acid site densities

between 0.4 and 0.7 mmol/g. These structure types were chosen because they

represent three-, two-, and one-dimensional zeolites. The authors claimed that the

influence of most properties influencing the performance—besides the

structure—was minimized. Zeolite BEA exhibited the best time-on-stream

behavior with respect to lifetime and TMP selectivity. ZSM-12 also showed a

long lifetime, but it catalyzed oligomerization instead of alkylation. USY, MOR,

and LTL were found to deactivate quickly, with LTL retaining a surprisingly

stable TMP selectivity at low conversions. No heavy coke molecules were found

in zeolites BEA and ZSM-12. The authors concluded that zeolites without

periodic expansions (i.e., without larger voids that connect channels) do not allow

extensive coke formation and hence deactivate relatively slowly. Unfortunately,

no details about the concentrations and strengths of the acid sites in the samples

were given. The finding that zeolite BEA does not produce significant amounts of

coke is at variance with results of other research groups. For example, Nivarthy

et al. (48) calculated values of about 14 wt% of deposit formed on H-BEA

zeolites. In the aforementioned investigation by Corma et al. (126), USY, BEA,

and MOR were compared with ZSM-5 and MCM-22. The three large-pore

zeolites exhibited similar C8-selectivities but different behavior with time on

stream. The differences were attributed to differences in the acidities of the

samples. In a comparative investigation of the acidity of zeolites with low Si/Al

ratios (zeolites BEA, ZSM-20, Y, and dealuminated USY), the acid strength

was found to decrease in the following order: H-BEA . H-USY . H-ZSM-20.

H-Y (177).

In another article by Corma et al. (178), ITQ-7, a three-dimensional large-pore

zeolite, was tested as an alkylation catalyst and compared with a BEA sample

of comparable Si/Al ratio and crystal size. The ratio of the selectivities to 2,2,4-

TMP and 2,2,3-TMP, which have the largest kinetic diameter of the TMPs, and

2,3,3-TMP and 2,3,4-TMP, which have the lowest kinetic diameter, was used as a

measure of the influence of the pore structure. Lower (2,2,4-TMP þ 2,2,3-TMP)/

(2,3,3-TMP þ 2,3,4-TMP) ratios in ITQ-7 were attributed to its smaller pore

diameter. The bulky isomers have more spacious transition states, so that their

formation in narrow pores is hindered; moreover, their diffusion is slower. The

hydride transfer activity, estimated by the dimethylhexane/dimethylhexene ratio,


was found to be lower in ITQ-7 than in H-BEA. This observation was also

explained by the smaller pore diameter, because the acidities of the two different

zeolites were found to be similar. Nivarthy et al. (179) compared the three large-

pore zeolites H-BEA, H-FAU, and H-EMT; the lifetimes of the zeolites were

found to depend on the Brønsted acid site concentration. H-BEA, with the lowest

Brønsted acid site concentration, was characterized by the shortest lifetime and

H-EMT, with the highest concentration, the longest lifetime. Significant

differences were observed in the TMP distribution. H-BEA exhibited a very

high 2,2,4-TMP selectivity, which was attributed to a lower rate of hydride

transfer vs. isomerization of the precursor carbenium ions. An exceptionally high

2,2,4-TMP selectivity is characteristic of zeolite BEA. Although with most other

zeolites the selectivities vary depending on the conditions employed, BEA

always produces high yields of 2,2,4-TMP.

The research group at SINTEF (123,124,180) dedicated a series of papers

to the examination of FAU and EMT zeolites, comparing them in their H-

and La-exchanged form with and without dealumination. EMT was always

superior to FAU. The alkylate yield, expressed as mass of alkylate produced

divided by the catalyst mass, was higher for the EMT samples. EMT also

produced a greater amount of trimethylpentanes than the FAU samples. The

differences between the two materials were discussed in terms of the slightly

larger supercage in EMT, which is claimed to reduce the steric constraints on

the bulky transition states for hydride transfer, and in terms of acidity, with

EMT samples exhibiting a higher concentration of Brønsted acid sites

retaining pyridine at high temperatures. A comparison of La-EMT, La-FAU,

and La-BEA revealed that the La-BEA performed worse than the two other

materials, both in terms of alkylate yield and selectivity (167), but the lack of

information about the acidity of the samples prevents a detailed evaluation of

this report.

Recently, mesoporous aluminosilicates with strong acidity and high

hydrothermal stability have been synthesized via self-assembly of alumino-

silicate nanoclusters with templating micelles. The materials were found to

contain both micro- and mesopores, and the pore walls consist of primary

and secondary building units, which might be responsible for the acidity and

stability (181). These materials were tested in isobutane/n-butene alkylation

at 298 K, showing a similar time-on-stream behavior to that of zeolite BEA.

No details of the product distribution were given.

The patent literature discloses alkylation performances of several additional

structure types. A Mobil patent (182) describes the use of VTM-A, a pillared

titanosilicate of the MCM-27 family. The catalyst produced about 80 wt% of

octanes under relatively mild conditions (OSV ¼ 0:05 h21, P=O ratio ¼ 20).

A number of patents describe the use of MCM-36. MCM-49, which is closely

related to MCM-22, has also been tested as an alkylation catalyst. In general,


these materials require a relatively high reaction temperature to be sufficiently

active, which inevitably leads to high cracking and high DMH selectivities (172,

183–187).

III.C. Other Solid Acids

III.C.1. Sulfated Zirconia and Related Materials

A variety of solid acids besides zeolites have been tested as alkylation catalysts.

Sulfated zirconia and related materials have drawn considerable attention

because of what was initially thought to be their superacidic nature and their well-

demonstrated ability to isomerize short linear alkanes at temperatures below

423 K. Corma et al. (188) compared sulfated zirconia and zeolite BEA at reaction

temperatures of 273 and 323 K in isobutane/2-butene alkylation. While BEA

catalyzed mainly dimerization at 273 K, the sulfated zirconia exhibited a high

selectivity to TMPs. At 323 K, on the other hand, zeolite BEA produced more

TMPs than sulfated zirconia, which under these conditions produced mainly

cracked products with 65 wt% selectivity. The TMP/DMH ratio was always

higher for the sulfated zirconia sample. These distinctive differences in the

product distribution were attributed to the much stronger acid sites in sulfated

zirconia than in zeolite BEA, but today one would question this suggestion

because of evidence that the sulfated zirconia catalyst is not strongly acidic, being

active for alkane isomerization because of a combination of acidic character and

redox properties that help initiate hydrocarbon conversions (189). The time-on-

stream behavior was more favorable for BEA, which deactivated at a lower rate

than sulfated zirconia. Whether differences in the adsorption of the feed and

product molecules influenced the performance was not discussed.

In a subsequent publication (22), two sulfated zirconia samples with different

sulfate loadings were reported as alkylation catalysts with isobutane/2-butene

feed at temperatures between 263 and 323 K. The sample with the higher sulfur

loading was slightly more active in the initial reaction phase, and the rates of

deactivation were similar for the two catalysts. The alkylation/cracking ratio

increased with decreasing reaction temperature. 2,2,4-TMP was the dominant

octane isomer under all conditions and less dimethylhexanes and octenes were

produced than with the zeolitic catalysts. In another investigation by the same

authors, sulfate-doped ZrO2, TiO2, and SnO2 were prepared by various sulfation

and activation procedures. The acidity decreased in the order SO422/

ZrO2 . SO422/TiO2 . SO4

22/SnO2, which was reflected in the cracking activities

of the samples. All the oxides showed considerable sensitivity towards the modi-

fication procedure, each with a different optimum. All the samples deactivated


rapidly and additionally lost significant fractions of the sulfur that was originally

present (190).

Satoh et al. (191) also compared several sulfated metal oxide catalysts, which

were tested with gas-phase reactants at 273 K. This is an inappropriate procedure,

however, because most of the products are liquid under these conditions and

remain in the catalyst pores. The authors of an investigation with pulsed gas-

phase reactants for alkylation with sulfated zirconia catalyst also concluded that

at temperatures below 323 K the TMPs could not desorb from the pores. Raising

the temperature to just 373 K led to dehydrogenation of isobutane (192,193).

Other investigations of alkylation of gas-phase reactants with sulfated zirconia

were reported by Das and Chakrabarty (194) and Guo et al. (195,196). Working

with liquid-phase reactants and employing relatively mild conditions, Xiao et al.

(197) were able to extend the lifetime of a sulfated zirconia catalyst to more than

70 h. In the initial phase, the catalyst produced 80 wt% cracked products, but the

value fell to less than 20 wt% after 30 h TOS with an increase in TMP selectivity

to more than 60 wt%. Within the TMPs, 2,2,4-TMP selectivities became higher

than 60 wt%. Platinum-promoted sulfated zirconia and tungstated zirconia were

found to be much less active alkylation catalysts.

An interesting variation on sulfated metal oxide type catalysts was presented

by Sun et al. (198), who impregnated a dealuminated zeolite BEA with titanium

and iron salts and subsequently sulfated the material. The samples exhibited a

better time-on-stream behavior in the isobutane/1-butene alkylation (the reaction

temperature was not given) than H-BEA and a mixture of sulfated zirconia and

H-BEA. The product distribution was also better for the sulfated metal oxide-

impregnated BEA samples. These results were explained by the higher

concentration of strong Brønsted acid sites of the composite materials than in

H-BEA.

III.C.2. Heteropolyacids

Heteropolyacids are strongly acidic non-porous solids. Salts of these acids

containing large cations, such as Csþ, Kþ, Rbþ, and NH4þ, exhibit surface areas in

the order of 150 m2/g. Supporting heteropolyacids on highly porous carriers

provide a method to increase the surface area. This was done by Blasco et al.

(199), who used 12-tungstophosphoric acid on silica, on a high-surface-area

amorphous aluminosilicate, and on all-silica mesoporous MCM-41. These

materials were tested for isobutane/2-butene alkylation at 306 K. The acid

supported on silica performed best, with high initial activity and selectivity to

trimethylpentanes. Heteropolyacids supported on the aluminosilicate interacted

strongly with the support, which decreased the acidity, thus leading to lower

activity and selectivity. Heteropolyacids on MCM-41 were observed to partially

block the pores of the support, so that a fraction of the acid was inaccessible


to the reactants. This effect of pore blocking could be decreased by the use of a

MCM-41 with a larger pore diameter. All the materials deactivated rapidly.

Gayraud et al. (200) tested potassium salts of 12-tungstophosphoric acid with

various potassium loadings to modify acidity and porosity. The samples were

tested at sub- and super-critical conditions. Samples with high potassium content

exhibited better time-on-stream behavior and selectivities than others. The

authors claimed that high acid site density was detrimental for the alkylation

reaction, leading to increased oligomerization activity. This conclusion contra-

dicts the generally accepted notion that high acid site densities enhance the

alkylation activity. The results can be better explained by the decrease in surface

area with decreasing potassium content, which was found to vary from 156 to

50 m2/g.

Cesium salts of 12-tungstophosphoric acid have been compared to the pure

acid and to a sulfated zirconia sample for isobutane/1-butene alkylation at room

temperature. The salt was found to be much more active than either the acid or

sulfated zirconia (201). Heteropolyacids have also been supported on sulfated

zirconia catalysts. The combination was found to be superior to heteropolyacid

supported on pure zirconia and on zirconia and other supports that had been

treated with a variety of mineral acids (202). Solutions of heteropolyacids

(containing phosphorus or silicon) in acetic acid were tested as alkylation

catalysts at 323 K by Zhao et al. (203). The system was sensitive to the hetero-

polyacid/acetic acid ratio and the amount of crystalline water. As observed in

the alkylation with conventional liquid acids, a polymer was formed, which

enhanced the catalytic activity.

III.C.3. Acidic Organic Polymers

Nafion-H, a perfluorinated sulfonic acid resin, is another strongly acidic solid

that has been explored as alkylation catalyst. Rørvik et al. (204) examined

unsupported Nafion-H with a nominal surface area of 0.2 m2/g (surface area of

a swellable polymer is difficult to define) in isobutane/2-butene alkylation at

353 K and compared it with a CeY zeolite. The zeolite gave a better alkylate and

higher conversion than Nafion-H, which produced significant amounts of octenes

and heavy-end products. The low surface area of the resin and questions about

the accessibility of the sulfonic acid groups probably make the comparison

inadequate.

To increase the surface area, the resin can be supported on porous carriers, or it

can be directly incorporated into silica by a sol–gel preparation technique. Both

methods have been used by Botella et al. (205), who compared several composite

Nafion/silica samples with varying surface areas and Nafion loadings for

isobutane/2-butene alkylation at 353 K. Furthermore, supported and unsupported

Nafion samples were used. As expected, the unsupported resin with its low


surface area performed only poorly. The composite materials showed an

optimum performance at intermediate surface areas, which was explained by an

interaction between the sulfonic groups of the resin and the silanol groups of the

silica, decreasing the acid strength of the resin. The supported resin showed

activity and selectivity similar to that of the composite material of the same

Nafion content. Changes in the temperature from 305 to 353 K showed that the

material produces oligomers at low temperatures and saturated products at higher

temperatures. This led the authors to claim the catalyst to have an acid strength in

the range characteristic of zeolite BEA and lower than that of sulfated zirconia.

Such a ranking, however, seems obsolete because sulfated zirconia has been

shown to exhibit only a moderate acid strength (Section III.C.1).

III.C.4. Supported Metal Halides

Supported metal halides gained considerable attention as candidate alkylation

catalysts, and at least three companies tested them in pilot plants (206).

Chlorinated alumina, obtained by reacting alumina with hydrogen chloride,

is a highly Brønsted-acidic porous solid. This material is related to the Friedel-

Crafts catalyst aluminum chloride, which was one of the first catalysts tested

in alkylation. Similar catalysts are used in commercial alkane isomerization

plants. A series of chlorinated alumina samples modified with Liþ and Naþ ions

was prepared and tested by Clet et al. (207) for isobutane/2-butene alkylation

at 273 K. The purpose of the cation addition was to moderate the acidity of

the material. It was shown that the cations prevent excessive cracking, and

the time-on-stream behavior is superior to that of the unmodified sample.

The improvement was attributed to a selective annihilation of very strong

acid sites by the cations. The degradation of 2,2,4-TMP on these catalysts at

273 K was also investigated. 2,2,4-TMP was found to be surprisingly reactive

under these conditions and gave a product resembling an alkylate—but with

more dimethylhexanes and light- and heavy-end products. Emphasis was placed

on the explanation of the rearrangement steps for producing dimethylhexanes

and cracked products, but the initiation of the 2,2,4-TMP degradation was

not discussed. These catalysts are also described in a patent application (208).

A similar type of catalyst including a supported noble metal for regeneration

was described extensively in a series of patents assigned to UOP (209–214).

The catalysts were prepared by the sublimation of metal halides, especially

aluminum chloride and boron trifluoride, onto an alumina carrier modified with

alkali or rare earth-alkali metal ions. The noble metal was preferably deposited

in an eggshell concentration profile. An earlier patent assigned to Texaco (215)

describes the use of chlorinated alumina in the isobutane alkylation with higher

alkenes, especially hexenes. TMPs were supposed to form via self-alkylation.

Fluorinated alumina and silica samples were also tested in isobutane alkylation,


but were found to produce mainly heavy-end products under the conditions

employed (216).

Patents assigned to Mobil (217) describe the use of boron trifluoride supported

on several porous carriers. BF3 supported on silica was found to exhibit a slightly

higher performance with added water in the alkylation of a mixed alkene feed

at 273 K. It was also shown that self-alkylation activity was considerably lower

than that with HF as catalyst. Another patent (218) describes the use of a pillared

layered silicate, MCM-25, promoted with BF3 to give a high-quality alkylate at

temperatures of about 273 K. BF3 was also supported on zeolite BEA, with

adsorbed water still present (219). This composite catalyst exhibited low butene

isomerization activity, which was evident from the inferior results obtained with

1-butene. At low reaction temperatures, the product quality was superior to that

of HF alkylate.

Triflic acid has also been supported on a porous silica carrier (220). The

authors emphasized the importance of a strong interaction between the acid and

the support to prevent leaching of the acid. In pulsed liquid-phase isobutane/

1-butene alkylation experiments at 298 K, the catalysts produced a very high-

quality alkylate, made up almost exclusively of isooctanes. With silanol groups

on the silica surface or with added water, triflic acid was found to form a

monohydrate that was firmly grafted to the silica surface.

III.D. The Influence of Process Conditions

The choice of appropriate reaction conditions is crucial for optimized

performance in alkylation. The most important parameters are the reaction

temperature, the feed alkane/alkene ratio, the alkene space velocity, the alkene

feed composition, and the reactor design. Changing these parameters will induce

similar effects for any alkylation catalyst, but the sensitivity to changes varies

from catalyst to catalyst. Table II is a summary of the most important parameters

employed in industrial operations for different acids. The values given for

zeolites represent best estimates of data available from laboratory and pilot-scale

experiments.

Two points are emphasized: (i) zeolites can be successfully operated at the

same or higher severities (with respect to P/O (feed alkane/alkene) ratio and OSV

(alkene space velocity)) than the liquid acids; (ii) the productivities of zeolite

catalysts (i.e., the total amount of alkylate produced per mass of catalyst) are

roughly the same as of that of sulfuric acid. If the intrinsic activities of zeolites

(which have 0.5–3 mmol of acid sites per gram) are compared with that of

sulfuric acid (which has 20 mmol of acid sites per gram), zeolites outperform

sulfuric acid. Nevertheless, the price of a zeolite catalyst and the high costs of


effective regeneration set high hurdles for a competition with sulfuric acid-

catalyzed processes.

III.D.1. Reaction Temperature

The reaction temperature affects both the chemistry of alkylation through the

activation energies of the individual reaction steps and the solubility/adsorption

and diffusion of products and reactants. With sulfuric acid, the viscosity is also

strongly influenced by the temperature. Dispersion effects, such as too low

interfacial areas between acid and hydrocarbons, thus set the lower temperature

limit with sulfuric acid. Temperatures below 277 K inhibit the separation of acid

from the hydrocarbon phase and lead to acid carryover from the acid settler

downstream of the reactor. At temperatures exceeding 291 K, polymerization

reactions dominate, leading to increased acid consumption and low octane

numbers (12). The higher solubility of isobutane in HF and its lower viscosity

allow higher isobutane consumption rates to be applied with HF. Therefore, HF

can be operated at higher temperatures, resulting in higher reaction rates. This

operation also reduces the refrigeration costs. Instead of a true refrigeration

system, cooling water can be used. Nevertheless, the product quality is higher

when the operation is at the lower temperature limit. With increasing

temperature, the rates of side reactions increase. Oligomerization/cracking is

of greater importance at higher temperatures, reducing the selectivity to

trimethylpentanes.

TABLE II

Typical values of important process parameters for alkylation

Catalyst

HF H2SO4 Zeolites

Reaction temperature (K) 289–313 277–291 323–373

Feed alkane/alkene ratio (mol/mol) 11–14 7–10 6–15

Alkene space velocity (kgAlkene/kgAcid h) 0.1–0.6 0.03–0.2 0.2–1.0

Exit acid strength (wt%) 83–92 89–93 –

Acid per reaction volume (vol%) 25–80 40–60 20–30

Catalyst productivity (kgAlkylate/kgAcid) 1000–2500 6–18 4–10

The numbers for the liquid acids are taken from Refs. (12,23,221). As zeolites are not used in

industrial alkylation process, the given values represent the judgment of the authors extracted from

laboratory and pilot scale data obtained in a slurry reactor.


Zeolites, in principle, operate at significantly higher reaction temperatures than

the liquid acids. The need for higher temperatures is attributed to the lower acid

strengths of zeolites or the lack of solvation, resulting in higher activation

energies for the individual reaction steps. Efficient mobility in the zeolite

micropores also requires higher temperatures. The optimum temperature is in the

range 323–373 K, with the exact value likely depending on the individual

sample. The problem of the optimum reaction temperature is often overlooked in

test for comparison of various catalysts. Testing catalysts at sub-optimum

temperatures leads to false conclusions about the true alkylation performance.

Nivarthy et al. (48) found an optimum for zeolite H-BEA at 348 K, at which

temperature the highest octane selectivity and the highest TMP/DMH ratio were

achieved. At lower temperatures, oligomerization dominated, and at higher

temperatures, cracking reactions dominated. Kirsch et al. (14) tested various rare

earth exchanged Y zeolites at temperatures 298–373 K. A sample with 0.2 wt%

residual sodium had a optimum temperature around 313 K, and a sample with

1.0 wt% sodium performed best at 353 K.

Taylor and Sherwood (222) examined the influence of several process

parameters on the performance of a USY zeolite. The catalyst was tested at 311,

339, and 367 K. The TMP selectivity decreased steadily with increasing

temperature, and the longest lifetime was achieved at 339 K. Pronounced effects

on the product selectivities were also observed by Corma et al. (140), who used

a H-BEA catalyst at 323 and 353 K. At the higher temperature, the activity was

higher, as indicated by the increased conversion. The selectivity to cracked

products increased drastically, and the C9þ selectivity also increased with

temperature. Within the TMP fraction, 2,2,4-TMP increased significantly

with temperature. Feller et al. (89) performed a detailed investigation of the

influence of the reaction temperature in the range of 313–403 K on the per-

formance of a LaX zeolite. The catalyst lifetime was found to depend strongly

on the reaction temperature, with an optimum at 348 K. The product quality

was highest at low temperatures; with increasing temperatures, increasingly

more cracked and heavy compounds were produced. The TMP/DMH ratio

declined with temperature. The selectivity phenomena can be explained by the

relative rates of the individual reaction steps. b-Scission (and presumably also

alkene addition) are characterized by higher activation energies than hydride

transfer. Increases in temperature consequently lead to higher relative rates of

secondary products from multiple alkylation and cracking. Cracked products

are favored over multiple alkylation products, because the activation energy

is higher for b-scission than for alkene addition, which is the (exothermic)

reverse reaction.

The bad performance of zeolites at low reaction temperatures is most likely a

consequence of the hindered diffusion of bulky molecules under such conditions.

The catalyst will be prematurely deactivated by pore blocking. These diffusion


problems are the reason why several research groups tried to overcome a buildup

of heavy molecules in the catalyst pores by employing supercritical conditions. A

supercritical reaction medium should combine liquid-like density with high

oligomer solubilities and gas-like transport properties. Under such conditions, the

bulky molecules that otherwise would deactivate the catalyst are supposedly

more efficiently removed from the catalyst pores. The feed itself can be employed

as a supercritical medium, but the critical point of isobutane is 408 K and

36.5 bar. Performing the alkylation reaction under these conditions leads to

excessive cracking. The catalyst stays active for longer times than in the con-

ventional operation, but it produces cracked products and especially substantial

amounts of alkeneic products (130,131). To overcome the problems associated

with the high critical temperature of isobutane, carbon dioxide has been used as a

diluent to reduce the critical temperature. The results presented by Clark and

Subramaniam (223) show that a stable conversion can indeed be maintained

with a 10-fold excess of carbon dioxide at 323 K and 155 bar. However, the

conversion was very low (,20 wt%), and the product contained only minor

amounts of trimethylpentanes. Similar results were reported by Santana and

Akgerman (224). Ginosar et al. (225), testing a variety of supercritical solvents

and a variety of solid acids, came to the conclusion that working under super-

critical conditions generally does not improve the alkylation performance.

A temperature-programmed-oxidation analysis of samples coked under super-

critical conditions revealed that the carbonaceous deposits are very similar in

concentration and oxidizability to coke produced under liquid-phase reaction

conditions. The slight changes were related to a smaller amount of coke on the

outer surface of the zeolite (226).

III.D.2. Alkane/Alkene Ratio and Alkene Space Velocity

Rates of reaction are influenced by reactant concentrations, determined by the

feed composition, and temperature. The crucial parameter that determines a high

alkylate quality and a low acid consumption is the ratio of rates of hydride

transfer and oligomerization. This ratio should be as high as possible. Increasing

the isobutane concentration minimizes undesired reactions and acid consumption

by increasing the probability that the carbenium ion will react with an isobutane

molecule to form the desired product via hydride transfer rather than undergoing

oligomerization with other alkenes. The ratio of rates of hydride transfer to

oligomerization is primarily influenced by two process parameters: the feed

alkane/alkene (P/O) ratio and the alkene space velocity (OSV, which is appro-

ximately proportional to the reciprocal of the average residence time). The P/O

ratio determines the concentration of isobutane in the reactor and thereby the rate


of hydride transfer. The P/O ratio also sets the product concentration, which

affects the rates of the product degradation reactions.

Another point might be of importance, although no quantitative data are

available to assess it. Ideally, feed entering the reactor should be instan-

taneously mixed with the acid. The conversion of the alkene in the reactor is

usually complete, so that the internal P/O ratio might be 1000:1 or even higher.

In the case of incomplete mixing, the alkene concentration will be higher at

some positions in the reactor and consequently lead to higher rates of

oligomerization and acid consumption than would occur if mixing were perfect.

With high feed P/O ratios, the detriment of incomplete mixing will be

minimized. Thus, increasing the P/O ratio increases alkylate quality and yield

and decreases acid consumption. On the other hand, at high P/O ratios, more

isobutane has to be recycled, which leads to increased separation costs. A

balance has to be found to optimize the economic performance of the unit. The

OSV determines the production rate of alkylate, so that high OSV would be

economically favored, but this is limited by high acid consumption, low octane

number, and high rates of formation of heavy-end products at high values of

OSV. When the catalyst is sulfuric acid, more esters are introduced into the

products than when the catalyst is hydrofluoric acid, and these products corrode

down-stream equipment (221). As a first approximation for sulfuric acid-

catalyzed n-butene alkylation, an increase in OSV of 0.1 vol/(vol £ h) leads to a

decrease in RON of about 1, and an increase in the P/O from 8 to 9 leads to an

increase in RON of 0.15 (12). The above-mentioned higher solubility of

isobutane in HF allows higher space velocities in HF plants, although they are

usually operated at higher P/O ratios.

In principle, the same rules hold true when zeolitic alkylation catalysts are

used. A detailed study of the influence of PO and OSV on the performance of

zeolite H-BEA in a backmix reactor was reported by de Jong et al. (80). The

authors developed a simple model of the kinetics, which predicted catalyst

lifetimes as a function of P/O and OSV. Catalyst lifetime (which is equivalent to

the catalyst productivity, the reciprocal of acid consumption) increased with

increasing P/O ratio and decreasing OSV. Furthermore, the authors persuasively

demonstrated the superiority of a backmix reactor over a plug flow reactor.

Qualitatively similar results were obtained by Taylor and Sherwood (222)

employing a USY zeolite catalyst in a backmix reactor. The authors stressed the

detrimental effect of unreacted alkene on the catalyst lifetime and product

quality. Feller et al. (89) tested LaX zeolites in a backmix reactor and found the

catalyst productivity to be nearly independent of the OSV within the examined

OSV range. At higher values of OSV, the catalyst life was shorter, but in this

shorter time the same total amount of product was produced. The P/O ratio had

only a moderate influence on the catalyst performance.


III.D.3. Alkene Feed Composition

Propene, 1-butene, 2-butene, isobutylene, and normal- and isopentenes can be

used as feedstocks in alkylation. Depending on the catalyst, they give different

alkylate qualities and yields with differing acid consumptions. Only linear

butenes give a fairly low acid consumption in sulfuric acid-catalyzed processes.

All the other alkenes lead up to three times higher acid consumption (12).

Hydrofluoric acid consumption is nearly independent of the feed alkene (227).

The low double bond isomerization activity of HF leads to higher production of

dimethylhexanes when 1-butene is the feed alkene. The high self-alkylation

activity of HF is responsible for a high fraction of TMPs in the alkylate when

alkenes other than butenes are used.

Table III provides a comparison of alkylate compositions for both the liquid

acid-catalyzed reactions with various feed alkenes. The data show that H2SO4

produces a better alkylate with 1-butene, whereas HF gives better results with

propene or isobutylene. The products from 2-butene and also from pentenes (not

shown in Table III) are nearly the same with either acid.

Zeolites have also been tested with feed alkenes other than butenes. Daage and

Fajula (229) reported an investigation of isobutane/propene alkylation with a

CeY zeolite with 13C-labeled feed molecules. The products could be grouped into

those formed by three classes of reactions: dimerization leading to C6 products,

alkylation leading to C7 products, and self-alkylation leading to C8 and also to C7

products. Investigations by Guisnet et al. (69,93) comparing 2-butene and

propene as feed alkenes with a USY zeolite catalyst gave similar results. Self-

alkylation was slower by a factor of two than the alkylation of isobutane with

propene and faster by a factor of two than the dimerization of propene. The

conversion in isobutane/propene alkylation was considerably lower than in

isobutane/2-butene alkylation. A comparative study of zeolite H-BEA-catalyzed

alkylation of isobutane with 2-butene, propene, and ethene was published by

Nivarthy et al. (230). The reactivity of the alkenes decreased in the order 2-

butene . propene . ethene. Here, the products could also be grouped into those

formed by dimerization, alkylation, and self-alkylation. Dimerization is

especially important with ethene, forming n-butenes, which react in the normal

way to give octanes. The distribution within the C8 fraction was almost the same

when ethene was used instead of 2-butene. Ethene exhibits a low reactivity

because it can form only primary carbenium ions, which requires high activation

energies. Ethene is reactive with AlCl3/HCl, but not with sulfuric acid or

hydrofluoric acid.

Early investigations of zeolite REHX as a catalyst were done with ethene as

the feed alkene (13). At 300 K, the product was mainly hexanes, whereas at

temperatures as high as 422 K, isopentane dominated, with hexanes and octanes

being the other main products. KTI developed a process which utilizes ethene


from FCC off-gases to produce alkylate with a zeolite catalyst having a

“dimerization function” (231). The catalyst disclosed consists of a RECaX

zeolite impregnated with palladium as the “dimerization function” (152).

Operated at temperatures in the range of 323–343 K, the catalyst produces a

high yield of octanes and almost no hexanes (153,232,233). Chlorided alumina

was also tested as a catalyst for isobutane/ethene alkylation at temperatures

between 273 and 373 K. Catalyst stability was better at low temperatures than at

high temperatures. Hexanes constituted the main product fraction, especially

at high P/O ratios (234). Thermodynamically, hexanes are strongly favored over

octanes and higher molecular weight products (235).

TABLE III

Compositions of alkylates obtained with various feed alkenes and various acid

catalysts (50,228)

Component (wt%) Feed alkene and employed acid catalyst

Propene Isobutylene 2-Butene 1-Butene

HF H2SO4 HF H2SO4 HF H2SO4 HF H2SO4

C5

Isopentane 1.0 3.8 0.5 10.0 0.3 4.2 1.0 4.7

C6

Dimethylbutanes 0.3 0.8 0.7 0.8

Methylheptanes 04.2

0.25.2

0.24.6

0.34.4

C7

2,3-Dimethylpentane 29.5 50.4 2.0 2.6 1.5 1.4 1.2 1.5

2,4-Dimethylpentane 14.3 20.8 0 3.9 0 2.4 0 2.6

C8

2,2,4-Trimethylpentane 36.3 4.4 66.2 28.7 48.6 30.6 38.5 30.5

2,2,3-Trimethylpentane 0 0 1.9 0.9

2,3,4-Trimethylpentane 7.5 12.8 22.2 19.1

2,3,3-Trimethylpentane 4.03.7

7.123.1

12.941.6

9.739.1

Dimethylhexanes 3.2 1.7 3.4 9.5 6.9 9.0 22.1 11.0

C9þ products 3.7 11.0 5.3 17.1 4.1 6.3 5.7 6.2


IV. Industrial Processes and Process Developments

This section is a review of alkylation process technology. The processes in which

liquid acids are used are all mature technologies and described briefly here.

Information about process developments with solid acid catalysts is also

presented.

IV.A. Liquid Acid-Catalyzed Processes

All the processes require intensive mixing of acid and hydrocarbon phases

to form emulsions. The droplets have to be small enough to give a sufficiently

large phase boundary area, but they also have to ensure a quick separation

in the settler downstream of the reactor to prevent degradation reactions.

Because of the high viscosity of sulfuric acid, mixing is more of a problem

with sulfuric acid than with hydrofluoric acid. In all sulfuric acid-catalyzed

processes impellers have to be employed. In hydrofluoric acid processes the

hydrocarbons are typically injected through nozzles, which are sufficient for

effective dispersion.

Because the alkylation reaction is exothermic, a considerable amount of

process heat has to be removed. As HF-catalyzed processes operate at

temperatures between 289 and 313 K, the reactors can be cooled with water.

H2SO4-catalyzed processes operate at temperatures between 277 and 291 K

(Table II) and therefore require more complex cooling systems, which typically

utilize the processed hydrocarbon stream itself.

The feed hydrocarbons, which come from the FCC or from the etherification

unit of a petroleum refinery, usually have to be treated before entering the

alkylation unit. They contain water, butadienes, and sulfur- and nitrogen-

containing compounds and—when coming from an etherification unit—traces of

oxygenates.

The general treatment of the hydrocarbon stream leaving the alkylation reactor is

similar in all processes. First, the acid and hydrocarbon phases have to be separated

in a settler. The hydrocarbon stream is fractionated in one or more columns to

separate the alkylate from recycle isobutane as well as from propane, n-butane, and

(sometimes) isopentane. Because HF processes operate at higher isobutane/alkene

ratios than H2SO4 processes, they require larger separation units. All hydrocarbon

streams have to be treated to remove impurity acids and esters.

IV.A.1. Sulfuric Acid-Catalyzed Processes

Two licensors now offer sulfuric acid alkylation units. The one with the higher

market share is Stratco, with its Effluent Refrigerated Sulfuric Acid Alkylation


Process (12). The reactor is a horizontal pressure vessel called Contactore and

containing an inner circulation tube, a heat exchanger tube bundle to remove the

heat of reaction, and a mixing impeller in one end. The hydrocarbon feed and

recycle acid enter on the suction side of the impeller inside the circulation tube.

This design ensures the formation of a fine acid-continuous emulsion. The high

circulation rate prevents significant temperature differences within the reactor.

The reactor is shown schematically in Fig. 11.

A portion of the emulsion flows to the settler, where the hydrocarbon phase is

separated from the acid phase. The hydrocarbon phase is expanded and partially

evaporated. The cold two-phase hydrocarbon effluent is passed through the

cooling coils of the contactor reactor and takes up the heat of reaction as it

undergoes evaporation. To increase the efficiency of the cooling system, propane

is co-fed to the reactor. The gaseous hydrocarbons are sent to a refrigerant

compressor and separated from excess propane in a depropanizer column. The

acid leaving the settler is recycled into the reactor, with a small stream of fresh

acid continuously replacing the equivalent stream of spent acid. To increase

product quality and reduce acid consumption, the reactor can be staged with

respect to the acid flow; the acid can be passed through up to four contactor

reactors with each reactor being fed with fresh hydrocarbons.

The spent acid strength is maintained at about 90 wt% H2SO4. The molar

isobutane/alkene feed ratio ranges from 7:1 to 10:1. Typical operating alkene

space velocities (LHSV) range from 0.2 to 0.6 h21 (corresponding to WHSVs

from 0.06 to 0.19 h21). The optimum reaction temperatures range from 279 to

283 K, but some units are operated at temperatures up to 291 K.

Fig. 11. Stratcow Contactore reactor used in sulfuric acid-catalyzed alkylation (12).


The second licensor of sulfuric acid-catalyzed alkylation processes is

ExxonMobil, with the stirred auto-refrigerated process (221), a technology

formerly licensed by Kellogg. In this process, the reactor consists of a large

horizontal vessel divided into a series of reaction zones, each equipped with a

stirrer (Fig. 12). The alkene feed is premixed with recycle isobutane and fed in

parallel to all mixing zones, and the acid and additional isobutane enter only the

first zone and cascade internally to the other zones. The heat of the reaction is

removed by evaporating isobutane plus added propane from the reaction zones.

Thus, no cooling coils are necessary in this type of process. To minimize

any increase in temperature along the reaction zones, the vessel is divided

into two pressure stages, with the second stage operating at a lower pressure to

decrease the boiling point of the hydrocarbon mixture. The vapors are sent to

the refrigeration section, where they are compressed, condensed, and returned

to the reactor as recycle refrigerant. To prevent a buildup of propane in the

refrigeration section, a slipstream is withdrawn and separated in a depropanizer.

The liquid stream is separated in a settler, from which the acid phase is recycled

into the reactor.

Because of its large reactor volume, the auto-refrigerated process can operate

at very low alkene space velocities of about 0.1 h21 LHSV (WHSV ca. 0.03 h21).

This design helps in increasing the octane number of the product and lowering

acid consumption. The reaction temperature is maintained at about 278 K to

minimize side reactions. Spent acid is withdrawn as 90–92 wt% acid. The

isobutane concentration in the hydrocarbon phase is kept between 50 and

70 vol%.

Fig. 12. ExxonMobil auto-refrigerated alkylation process. Adapted from Ref. (221).


Stratco offers a process called Alkysafee, proposing the conversion of an

existing HF alkylation unit to use H2SO4 for approximately the same cost as

installing an effective HF mitigation system. The process reuses the reaction and

distillation sections from the existing unit. Refrigeration is carried out with a

closed-loop packaged propane refrigeration section. Emulsion pumps and static

mixers have to be installed to provide the required mixing. Stratco claims the

production of similar or even increased quality alkylate as compared to that of the

former HF plant.

The process flow of the converted unit is similar to that of the time tank units

built between 1938 and 1958 (15,227). In this process, the hydrocarbons are

brought in contact with the acid in a large non-cooled pipe close to the entrance

of a centrifugal pump, which provides mixing and emulsification. The heat of

reaction is removed in a chiller (utilizing propane as a refrigerant) situated

downstream of the pump. The emulsion then flows into the time tank, which is a

large vertical vessel containing baffles. Although these units produced a high-

octane alkylate, they were successively shut down or changed over to different

types because of high costs of operation.

IV.A.2. Hydrofluoric Acid-Catalyzed Processes

ConocoPhillips offers a process using a non-cooled riser-type reactor (Fig. 13).

The hydrocarbon mixture is introduced through nozzles at the bottom and along

the length of the riser (236). The acid is injected at the bottom. The reactor

contains perforated trays, which help to maintain a high dispersion of the

hydrocarbons in the acid phase. The reaction mixture enters the settler, from

which the acid is withdrawn at the bottom and then cooled in a heat exchanger

with cooling water to remove the heat of the reaction. The cold acid is then fed

back into the reactor. The acid flow is driven by gravity. The hydrocarbons in the

settler are routed to the fractionation section, with an overhead stream of propane

and HF, a side stream of isobutane, another side stream of n-butane, and a

bottoms stream of alkylate leaving the section. The HF is separated from propane

in an HF stripper. The acid is regenerated by distillation to remove ASO and

water. Typical process parameters are temperatures of about 297 K, molar

isobutane/alkene ratios of about 14–15, and acid concentrations of 86–92 wt%.

At the heart of the UOP HF alkylation unit is a vertical reactor-heat exchanger,

shown in Fig. 14. The isobutane–alkene mixture enters the shell of the reactor

through several nozzles, and HF enters at the bottom of the reactor. The reaction

heat is removed by cooling water, which flows through cooling coils inside the

reactor. After phase separation in the settler, the acid is recycled to the reactor.

The hydrocarbon phase together with a slipstream of used acid and makeup

isobutane is sent to the “isostripper”, where the alkylate product, n-butane, and

isobutane are separated. The isobutane is recycled to the reactor. During normal


operation, the acid is distilled with the product, so that no external regeneration is

necessary. An additional acid regeneration column is still needed, however, for

startup, or when feed contamination occurs.

As a reaction to the pressure imposed on refiners operating HF processes,

licensors developed safety systems to reduce the inherent risks. Among the

mitigation systems are high-volume water sprays to “knock down” an acid cloud,

a low acid inventory, and a rapid acid de-inventory system. HF modifiers, which

reduce the volatility and the aerosol-forming tendency of HF, are also offered.

ConocoPhillips together with Mobil developed an HF modifier technology

named ReVape to reduce the volatility of the acid. It is claimed that a 60–90%

reduction in airborne acid release relative to that of the unmodified acid is

Fig. 13. ConocoPhillips HF alkylation reactor (236).


achieved. The modifier does not undergo a chemical reaction with the acid.

The additive is separated from the alkylate by extraction and recycled within the

alkylation unit. Furthermore, the ASO has to be separated from the additive. The

additive most likely is based upon sulfones. ConocoPhillips claims that when

using the additive the acid concentration can be lowered to 60 wt%.

UOP in a joint venture with ChevronTexaco developed an additive technology

named Alkade. The additive is based on HF salts of amines, which form liquid

“onium” polyhydrogen fluoride complexes with HF, reducing the vapor pressure

of the catalyst; 65% to more than 80% aerosol reduction is claimed with this

additive. As in the ReVape technology, additional separation columns have

to be installed. Both additives are claimed to increase the product octane

number, especially when propene, isobutylene, and pentenes are employed in

the feedstock.

IV.B. Solid Acid-Catalyzed Processes

Processes based on solid acids are not operated on an industrial scale. However,

several companies are developing processes or already offering technology for

licensing. The overall process scheme is similar to that of a liquid acid-based

process, except for the regeneration section, which is necessary with all solid acid

Fig. 14. UOP HF alkylation reactor.


catalysts. In principle, three regeneration methods have been examined closely:

(1) As in FCC, the hydrocarbons can be burned off the catalyst surface. This

requires a catalyst with extreme temperature stability, which only ultrastable

zeolites achieve. Moreover, as the alkylation process is exothermic and

conducted at low to moderate temperatures, large amounts of process heat

have to be removed.

(2) The catalyst can be treated with a solvent to extract hydrocarbon deposits.

The most straightforward solvent to use is isobutane, which has been shown

to restore catalytic activity only partially. Supercritical solvents have been

tested, but they also lead to only partial restoration of the activity. Super-

critical alkylation to remove the deposits in situ has been shown in Section

III.D.1 to be less effective. It is unlikely that this method of operation will

lead to a competitive process.

(3) The most promising regeneration method and the one that is used in all true

solid acid-catalyzed process developments is a hydrogen treatment at both

reaction and elevated temperature. This typically requires the incorporation

of a hydrogenation function, for example a noble metal, in the catalyst. The

regeneration mechanism depends on the temperature: at low temperatures

(,373 K), highly unsaturated species, which block the acid sites but not the

pores, are hydrogenated. At higher temperatures, hydrocracking of long-

chain alkanes and other hydrocarbons that are too bulky to leave the pores is

the predominant reaction. The fragments formed in this process easily desorb

and leave the pore system.

Although substantial research was devoted to plug-flow reactors, they are not

a good choice for large-scale operation. To achieve a high internal isobutane/

alkene ratio (.200), an enormous amount of isobutane has to be recycled.

Nevertheless, a plug-flow reactor remains attractive because of the simplicity of

its design and operation. When the alkene feed is introduced over the whole

length of the reactor, very low isobutane/alkene ratios can be avoided. However,

in a true fixed-bed reactor the inlet zones would nevertheless suffer from the

higher alkene concentration and deactivate prematurely.

A more appropriate type of reactor would be a backmixed slurry reactor, with

the catalyst suspended in the liquid. Such a system, however, also has obvious

disadvantages, such as the more complex design necessary for suspending the

solid in the liquid and for solid/liquid separation. These disadvantages may be

compensated by intrinsically higher isobutane/alkene ratios (a consequence of the

backmixing), which lower catalyst consumption. Another advantage of a slurry

reactor is the possibility to withdraw spent catalyst for regeneration. In fixed-bed

reactors, the bed can only be regenerated as a whole, so that multiple swing

reactors are necessary for uninterrupted production. Moreover, the attainment of

isothermal operation in slurry reactors is better than in fixed-bed reactors.


IV.B.1. UOP Alkylenee Process

UOP offers the Alkylenee process (237) utilizing a vertical riser reactor. A

process scheme is shown in Fig. 15. The pretreated alkene feed is mixed with

recycle isobutane and injected into the riser together with freshly reactivated

catalyst. Both flow concurrently upward in the riser, where the reaction occurs. At

the top of the riser the catalyst particles are disengaged and sink down into the

reactivation zone. The hydrocarbons flow out through the top of the reactor vessel

to the fractionation section, where they are separated into alkylate, n-butane,

isobutane, and light ends including hydrogen. The recycle isobutane is cooled

before re-entering the riser. The reactivation zone is a packed bed with the

catalyst slowly moving downward in a low-temperature stream of isobutane

saturated with hydrogen. Unsaturated molecules on the catalyst are claimed to be

hydrogenated and desorbed from the catalyst surface. The reactivation zone leads

to the bottom of the riser, where the cycle starts again. The catalyst reactivation is

not complete, and so a small slipstream of catalyst is withdrawn and directed to a

reactivation vessel, where the catalyst is regenerated in a semi-batch or batch

mode at elevated temperature in a circulating hydrogen stream. The composition

of the catalyst, which UOP refers to as HAL-100e, has not been disclosed. In

several patents the use of an alumina-supported AlCl3 catalyst modified with

alkali metal ions and a Ni, Pd, or Pt hydrogenation function is mentioned (see, for

example, Ref. (214)). Obviously, traces of halogen compounds are leached out

Fig. 15. UOP Alkylenee solid acid-catalyzed alkylation process (237).


of the catalyst, because a product treatment section is necessary. This would

additionally imply that a makeup halogen source is required. The alkene feed has

to be extensively treated to remove di-alkenes, sulfur-, oxygen-, and nitrogen-

containing compounds. The process operates at temperatures of 283–313 K and

at a molar isobutane/alkene ratio of 6–15. No information is available concerning

the alkene space velocity. It is interesting that typical alkene conversions are

between 93–100%, which most likely is a consequence of very short contact

times in the riser reactor. The alkylate RON is claimed to be as high as what is

attained with the existing technology.

IV.B.2. Akzo Nobel/ABB Lummus AlkyCleane Process

Akzo Nobel and ABB Lummus recently started a solid acid-catalyzed

alkylation demonstration plant at a Fortum refinery in Finland (238). The

reactor type used in the so-called AlkyCleane process has not been disclosed.

However, the process utilizes serial reaction stages with distributed alkene feed

injection for high internal isobutane/alkene ratios. The reactor type is claimed

to achieve a high degree of mixing to reduce alkene concentration gradients

throughout the reactor. Multiple reactors are used, which swing between

reaction and regeneration. As in the Alkylenee process, two regeneration

phases with different severities are employed. A mild regeneration at reaction

temperature and pressure with hydrogen dissolved in isobutane is performed

frequently (far before the end of the theoretical catalyst lifetime). When

necessary, the catalyst is fully regenerated at 523 K in a stream of gas-phase

hydrogen. Presumably, each reactor is in (mild) regeneration mode far longer

than in reaction mode.

The catalyst is reported to be a “true solid acid” without halogen ion addition.

In the patent describing the process (239), a Pt/USY zeolite with an alumina

binder is employed. It was claimed that the catalyst is rather insensitive to feed

impurities and feedstock composition, so that feed pretreatment can be less

stringent than in conventional liquid acid-catalyzed processes. The process is

operated at temperatures of 323–363 K, so that the cooling requirements are less

than those of lower temperature processes. The molar isobutane/alkene feed ratio

is kept between 8 and 10. Alkene space velocities are not reported. Akzo claims

that the alkylate quality is identical to or higher than that attained with the liquid

acid-catalyzed processes.

IV.B.3. LURGI EUROFUELw Process

LURGI and Sud-Chemie AG are developing a solid acid-catalyzed alkylation

process termed LURGI EUROFUELw. The reactor is derived from tray

distillation towers. Isobutane and suspended catalyst enter at the top of the


tower, and the alkene with premixed isobutane is introduced in stages (Fig. 16).

The evolved heat of reaction is most likely dissipated by the evaporation of the

reaction mixture. Thus, the temperature is controlled by the overall pressure and

the composition of the liquid. The catalyst–reactant mixture is agitated by the

boiling mixture of alkylate and isobutane. At the bottom of the column, the

catalyst is separated, and the majority of the alkylate/isobutane mixture is fed into

the separation section. Isobutane is recycled and mixed with the catalyst, which is

fed into the top of the reaction column. Intermittently, the catalyst is exposed to

hydrogen-rich operating conditions to minimize accumulation of unsaturated

compounds on its surface. Infrequent regeneration occurs in a proprietary section

at elevated temperatures.

The catalyst is faujasite derived, with a high concentration of sufficiently

strong Brønsted acid sites and a minimized concentration of Lewis acid sites. It

also contains a hydrogenation function. The process operates at temperatures of

about 323–373 K with a molar isobutane/alkene ratio between 6 and 12 and a

higher alkene space velocity than in the liquid acid-catalyzed processes.

Preliminary details of the process concept have been described (240).

IV.B.4. Haldor Topsøe FBAe Process

Haldor Topsøe’s fixed-bed alkylation (FBAe) technology is a compromise

between liquid and solid acid-based processes. It applies a supported liquid-phase

catalyst in which liquid triflic (trifluoromethanesulfonic) acid is supported on a

porous material (206,241). The acid in the bed is concentrated in a well-defined

catalyst zone, in which all the alkylation chemistry takes place: at the upstream

Fig. 16. LURGI EUROFUELw solid acid-catalyzed alkylation process (240).


end of the catalyst zone, ester intermediates are formed, which are soluble in the

hydrocarbons and are transported into the acid zone. Here, they react to form the

products and free acid. Thus, the active zone slowly migrates through the bed in

the direction of the hydrocarbon flow, as shown in Fig. 17. The spent acid can be

withdrawn from the reactor without interrupting the production. The acid is

regenerated in a proprietary acid recovery unit, which produces some oil as a by-

product. The products have to be treated to remove trace amounts of acid.

Reaction temperatures are in the range of 273–293 K. The reactor is operated

adiabatically, and the reaction heat is removed by a cooled reactor effluent

recycle (Fig. 18). The process is claimed to be robust against feed impurities.

Feed drying, however, is recommended.

Fig. 17. Reaction zone in Haldor Topsøe’s FBAe alkylation process (206).

Fig. 18. Haldor Topsøe’s FBAe alkylation process. Adapted from Ref. (206).


V. Conclusions

The foregoing review of the alkylation mechanism and the influence of the catalyst

type and reaction conditions show that, in essence, the chemistry is identical with

all the examined acid catalysts, liquid and solid. Differences in the importance of

individual reaction steps originate from the variety of possible structures and

distributions of acid sites of solid catalysts. Changing process parameters induces

similar effects with each of the catalysts; however, the sensitivity to a particular

parameter depends strongly on the catalyst. All the acids deactivate by the

formation of unsaturated polymers, which are strongly bound to the acid.

Liquid acid-catalyzed processes are mature technologies, which are not

expected to undergo dramatic changes in the near future. Solid acid-catalyzed

alkylation now has been developed to a point where the technology can compete

with the existing processes. Catalyst regeneration by hydrogen treatment is the

method of choice in all the process developments. Some of the process

developments eliminate most if not all the drawbacks of the liquid acid processes.

The verdict about whether solid acid-catalyzed processes will be applied in the

near future will be determined primarily by economic issues.

References

1. Stell, J., Oil Gas J. 99(52), 74 (2001).

2. Anonymous, Oil Gas J. 98(13) (2000).

3. Anonymous, Oil Gas J. 98(9) (2000).

4. Ipatieff, V.N., and Grosse, A.V., J. Am. Chem. Soc. 57, 1616 (1935).

5. Ipatieff, V.N., and Pines, H., US Patent 2,122,847 (1938).

6. Hutson, T. Jr., and McCarthy, W.C., in “Handbook of Petroleum Refining Processes”

(R.A. Meyers, Ed.), p. 1/23. McGraw-Hill, New York, 1986.

7. Pines, H., Chemtech. March, 150 (1982).

8. Albright, L.F., Chemtech. June, 40 (1998).

9. Albright, L.F., Oil Gas J. 26, 70 (1990).

10. Furimsky, E., Catal. Today 30, 223 (1996).

11. Albright, L.F., Chemtech. July, 46 (1998).

12. Kinnear, S., “STRATCO Alkylation Seminar,” Phoenix, AZ, 1998.

13. Garwood, W.E., and Venuto, P.B., J. Catal. 11, 175 (1968).

14. Kirsch, F.W., Potts, J.D., and Barmby, D.S., J. Catal. 27, 142 (1972).

15. Corma, A., and Martinez, A., Catal. Rev.—Sci. Engng 35, 483 (1993).

16. Weitkamp, J., and Traa, Y., in “Handbook of Heterogeneous Catalysis” (G. Ertl,

H. Knozinger and J. Weitkamp, Eds.), Vol. 4, p. 2039. VCH, Weinheim, 1997.

17. Schmerling, L., J. Am. Chem. Soc. 67, 1778 (1945).

18. Schmerling, L., J. Am. Chem. Soc. 68, 275 (1946).

19. Schmerling, L., Ind. Engng Chem. 45, 1447 (1953).

20. Albright, L.F., Spalding, M.A., Faunce, J., and Eckert, R.E., Ind. Engng Chem. Res. 27,

391 (1988).


21. Hutson, T. Jr., and Logan, R.S., Hydrocarbon Process. September, 107 (1975).

22. Corma, A., Martinez, A., and Martinez, C., J. Catal. 149, 52 (1994).

23. Gary, J.H., and Handwerk, G.E., in “Petroleum Refining—Technology and Economics”

(L.F. Albright, R.N. Maddox and J.J. McKetta, Eds.), p. 142. Marcel Dekker, New York,

1979.

24. Hommeltoft, S.I., Ekelund, O., and Zavilla, J., Ind. Engng Chem. Res. 36, 3491 (1997).

25. Frash, M.V., Solkan, V.N., and Kazansky, V.B., J. Chem. Soc. Faraday Trans. 93, 515

(1997).

26. Kazansky, V.B., Abbenhuis, H.C.L., van Santen, R.A., and Vorstenbosch, M.L.V., Catal.

Lett. 69, 51 (2000).

27. Kazansky, V.B., Catal. Rev.—Sci. Engng 43, 199 (2001).

28. Albright, L.F., Spalding, M.A., Nowinski, J.A., Ybarra, R.M., and Eckert, R.E., Ind.

Engng Chem. Res. 27, 381 (1988).

29. Kazansky, V.B., Catal. Today 51, 419 (1999).

30. Gorte, R.J., and White, D., Topics Catal. 4, 57 (1997).

31. Kazansky, V.B., in “Handbook of Heterogeneous Catalysis” (G. Ertl, H. Knozinger and

J. Weitkamp, Eds.), Vol. 2, p. 740. VCH, Weinheim, 1997.

32. Rigby, A.M., Kramer, G.J., and van Santen, R.A., J. Catal. 170, 1 (1997).

33. Rozanska, X., Demuth, T., Hutschka, F., Hafner, J., and van Santen, R.A., J. Phys. Chem.

B 106, 3248 (2002).

34. Ishikawa, H., Yoda, E., Kondo, J.N., Wakabayashi, F., and Domen, K., J. Phys. Chem. B

103, 5681 (1999).

35. Kondo, J.N., Ishikawa, H., Yoda, E., Wakabayashi, F., and Domen, K., J. Phys. Chem. B

103, 8538 (1999).

36. Engelhardt, J., and Hall, W.K., J. Catal. 151, 1 (1995).

37. Mota, C.J.A., Esteves, P.M., Ramirez-Solis, A., and Hernandez-Lamoneda, R., J. Am.

Chem. Soc. 119, 5193 (1997).

38. Hogeveen, H., Gaasbeek, C.J., and Bickel, A.F., Rec. Trav. Chim. 88, 703 (1969).

39. Olah, G.A., and Olah, J.A., in “Carbonium Ions” (G.A. Olah and P.R. Schleyer, Eds.), Vol.

2, p. 715. Interscience, New York, 1970.

40. Sanchez-Castillo, M.A., Agarwal, N., Miller, C., Cortright, R.D., Madon, R.J., and

Dumesic, J.A., J. Catal. 205, 67 (2002).

41. Boronat, M., Viruela, P., and Corma, A., J. Phys. Chem. A 102, 982 (1998).

42. Stewart, T.D., and Calkins, W.H., J. Am. Chem. Soc. 70, 1006 (1948).

43. Miron, L., and Lee, R.J., J. Chem. Engng Data 8, 150 (1963).

44. Albright, L.F., and Li, K.W., Ind. Engng Chem. Process Des. Dev. 9, 447 (1970).

45. Li, K.W., Eckert, R.E., and Albright, L.F., Ind. Engng Chem. Process Des. Dev. 9, 441

(1970).

46. Hofmann, J.E., and Schriesheim, A., J. Am. Chem. Soc. 84, 953 (1962).

47. Gorin, M.H., Kuhn, C.S., and Miles, C.B., Ind. Engng Chem. 38, 795 (1946).

48. Nivarthy, G.S., He, Y., Seshan, K., and Lercher, J.A., J. Catal. 176, 192 (1998).

49. Albright, L.F., and Kranz, K.E., Ind. Engng Chem. Res. 31, 475 (1992).

50. Shah, B.R., in “Handbook of Petroleum Refining Processes” (R.A. Meyers, Ed.), p. 1/3.

McGraw-Hill, New York, 1986.

51. Pines, H., Grosse, A.V., and Ipatieff, V.N., J. Am. Chem. Soc. 64, 33 (1942).

52. Olah, G.A., and Prakash, G.K.S., in “The Chemistry of Alkanes and Cycloalkanes”

(S. Patai and Z. Rappoport, Eds.), p. 609. Wiley, London, 1992.

53. Durrett, L.R., Taylor, L.M., Wantland, C.F., and Dvoretzky, I., Anal. Chem. 35, 637

(1963).

54. Cumming, K.A., and Wojciechowski, B.W., Catal. Rev.—Sci. Engng 38, 101 (1996).


55. Bartlett, P.D., Condon, F.E., and Schneider, A., J. Am. Chem. Soc. 66, 1531 (1944).

56. Deno, N.C., Peterson, H.J., and Saines, G.S., Chem. Rev. 60, 7 (1960).

57. Wojciechowski, B.W., and Corma, A., “Catalytic Cracking—Catalysts, Chemistry, and

Kinetics.” Marcel Dekker, New York, 1986, p. 5.

58. Ausloos, P., and Lias, S.G., J. Am. Chem. Soc. 92, 5037 (1970).

59. Meot-Ner, M., and Field, F.H., J. Chem. Phys. 64, 277 (1976).

60. Meot-Ner, M., J. Am. Chem. Soc. 109, 7947 (1987).

61. Sunner, J.A., Hirao, K., and Kebarle, P., J. Phys. Chem. 93, 4010 (1989).

62. Olah, G.A., Mo, Y.K., and Olah, J.A., J. Am. Chem. Soc. 95, 4939 (1973).

63. Boronat, M., Viruela, P., and Corma, A., Phys. Chem. Chem. Phys. 2, 3327 (2000).

64. Boronat, M., Viruela, P., and Corma, A., J. Phys. Chem. A 102, 9863 (1998).

65. Nowak, A.K., Mesters, C.M.A.M., Rigby, A.M., and Schulze, D., Preprints. Div. Petrol.

Chem., Am. Chem. Soc. 41, 668 (1996).

66. Boronat, M., Zicovich-Wilson, C.M., Corma, A., and Viruela, P., Phys. Chem. Chem.

Phys. 1, 537 (1999).

67. Beirnaert, H.C., Alleman, J.R., and Marin, G.B., Ind. Engng Chem. Res. 40, 1337 (2001).

68. Yaluris, G., Rekoske, J.E., Aparicio, L.M., Madon, R.J., and Dumesic, J.A., J. Catal. 153,

54 (1995).

69. Guisnet, M., and Gnep, N.S., Appl. Catal. A 146, 33 (1996).

70. Corma, A., Faraldos, M., Martinez, A., and Mifsud, A., J. Catal. 122, 230 (1990).

71. Hofmann, J.E., and Schriesheim, A., J. Am. Chem. Soc. 84, 957 (1962).

72. Nenitzescu, C.D., in “Carbonium ions” (G.A. Olah and P.R. Schleyer, Eds.), Vol. 2, p. 463.

Interscience, New York, 1970.

73. Schollner, R., and Holzel, H., Z. Chem. 15, 469 (1975).

74. Reyniers, M.-F., Beirnaert, H., and Marin, G.B., Appl. Catal. A 202, 49 (2000).

75. Reyniers, M.-F., Tang, Y., and Marin, G.B., Appl. Catal. A 202, 65 (2000).

76. Weeks, T.J. Jr., and Bolton, A.P., J. Chem. Soc. Faraday Trans. 170, 1676 (1974).

77. Lutsyk, A.I., and Suikov, S.Y., Petrol. Chem. 35, 412 (1995).

78. Nivarthy, G.S., Seshan, K., and Lercher, J.A., Micropor. Mesopor. Mater. 22, 379 (1998).

79. Lee, L., and Harriott, P., Ind. Engng Chem. Process Des. Dev. 16, 282 (1977).

80. de Jong, K.P., Mesters, C.M.A.M., Peferoen, D.G.R., van Brugge, P.T.M., and de Groot,

C., Chem. Engng Sci. 51, 2053 (1996).

81. Simpson, M.F., Wei, J., and Sundaresan, S., Ind. Engng Chem. Res. 35, 3861 (1996).

82. Albright, L.F., and Wood, K.V., Ind. Engng Chem. Res. 36, 2110 (1997).

83. Schollner, R., and Holzel, H., J. Prakt. Chem. 317, 694 (1975).

84. Martens, J.A., and Jacobs, P.A., Stud. Surf. Sci. Catal. 137, 633 (2001).

85. Weitkamp, J., and Ernst, S., Stud. Surf. Sci. Catal. 38, 367 (1988).

86. Frash, M.V., and van Santen, R.A., Topics Catal. 9, 191 (1999).

87. Daage, M., and Fajula, F., Bull. Soc. Chim. Fr. 5/6, 153 (1984).

88. Pater, J., Cardona, F., Canaff, C., Gnep, N.S., Szabo, G., and Guisnet, M., Ind. Engng

Chem. Res. 38, 3822 (1999).

89. Feller, A., Guzman, A., Zuazo, I., and Lercher, J.A., J. Catal. 224, 80 (2004).

90. Hofmann, J.E., J. Org. Chem. 29, 1497 (1964).

91. Graves, D.C., Kranz, K., and Millard, J., Prepr. Div. Petrol. Chem., Am. Chem. Soc. 39,

398 (1994).

92. Roebuck, A.K., and Evering, B.L., Ind. Engng Chem. Prod. Res. Dev. 9, 76 (1970).

93. Cardona, F., Gnep, N.S., Guisnet, M., Szabo, G., and Nascimento, P., Appl. Catal. A 128,

243 (1995).

94. am Ende, D.J., and Albright, L.F., Ind. Engng Chem. Res. 33, 840 (1994).

95. Guisnet, M., and Magnoux, P., Appl. Catal. A 212, 83 (2001).


96. Albright, L.F., Spalding, M.A., Kopser, C.G., and Eckert, R.E., Ind. Engng Chem. Res. 27,

386 (1988).

97. Kramer, G.M., J. Org. Chem. 32, 920 (1967).

98. Kramer, G.M., J. Org. Chem. 32, 1916 (1967).

99. Otvos, J.W., Stevenson, D.P., Wagner, C.D., and Beeck, O., J. Am. Chem. Soc. 73, 5741

(1951).

100. Flego, C., Kiricsi, I., Parker, W.O. Jr., and Clerici, M.G., Appl. Catal. A 124, 107 (1995).

101. Feller, A., Guzman, A., Zuazo, I., Barth, J.O., and Lercher, J.A., J. Catal. 216, 313 (2003).

102. Weitkamp, J., and Maixner, S., Zeolites 7, 6 (1987).

103. Deno, N.C., in “Carbonium Ions” (G.A. Olah and P.R. Schleyer, Eds.), Vol. 2, p. 783.


104. Sorensen, T.S., in “Carbonium Ions” (G.A. Olah and P.R. Schleyer, Eds.), Vol. 2, p. 807.


105. Nicholas, J.B., and Haw, J.F., J. Am. Chem. Soc. 120, 11804 (1998).

106. Yang, S., Kondo, J.N., and Domen, K., Catal. Today 73, 113 (2002).

107. Li, K.W., Eckert, R.E., and Albright, L.F., Ind. Engng Chem. Process Des. Dev. 9, 434

(1970).

108. Albright, L.F., Ind. Engng Chem. Res. 40, 4032 (2001).

109. He, M.Y., and Min, E., Catal. Today 63, 113 (2000).

110. Olah, G.A., Batamack, P., Deffieux, D., Torok, B., Wang, Q., Molnar, A., and Surya

Prakash, G.K., Appl. Catal. A 146, 107 (1996).

111. Olah, G.A., Marinez, E., Torok, B., and Prakash, G.K.S., Catal. Lett. 61, 105 (1999).

112. Holderich, W., and Jacobs, P.A., Stud. Surf. Sci. Catal. 137, 821 (2001).

113. Corma, A., and Garcia, H., Catal. Today 38, 257 (1997).

114. Ramachandran, S., Lenz, T.G., Skiff, W.M., and Rappe, A.K., J. Phys. Chem. 100, 5898

(1996).

115. Eder, F., and Lercher, J.A., Zeolites 18, 75 (1997).

116. Janchen, J., Stach, H., Uytterhoeven, L., and Mortier, W.J., J. Phys. Chem. 100, 12489

(1996).

117. Eder, F., and Lercher, J.A., J. Phys. Chem. 100, 16460 (1996).

118. Haag, W.O., Stud. Surf. Sci. Catal. 84, 1375 (1994).

119. Chu, Y.F., and Chester, A.W., Zeolites 6, 195 (1986).

120. Weitkamp, J., in “Proceedings of the Fifth International Zeolite Conference” (L.V.C.

Rees, Ed.), p. 858. Heyden, London, 1980.

121. Corma, A., Martinez, A., and Martinez, C., J. Catal. 146, 185 (1994).

122. Mostad, H.B., Stocker, M., Karlsson, A., and Rørvik, T., Appl. Catal. A 144, 305 (1996).

123. Stocker, M., Mostad, H., and Rørvik, T., Catal. Lett. 28, 203 (1994).

124. Stocker, M., Mostad, H., Karlsson, A., Junggreen, H., and Hustad, B., Catal. Lett. 40, 51

(1996).

125. Rørvik, T., Mostad, H.B., Karlsson, A., and Ellestad, O.H., Appl. Catal. A 156, 267

(1997).

126. Corma, A., Martinez, A., and Martinez, C., Catal. Lett. 28, 187 (1994).

127. Diaz-Mendoza, F.A., Pernett-Bolano, L., and Cardona-Martinez, N., Thermochim. Acta

312, 47 (1998).

128. Weitkamp, J., and Ernst, S., “Proceedings of the 13th World Petroleum Congress,” 1992,

p. 315.

129. Zhuang, Y., and Ng, F.T.T., Appl. Catal. A 190, 137 (2000).

130. Nakamura, I., Ishida, S., and Fujimoto, K., Preprints. Div. Petrol. Chem., Am. Chem. Soc.

40, 512 (1995).


131. Fan, L., Nakamura, I., Ishida, S., and Fujimoto, K., Ind. Engng Chem. Res. 36, 1458

(1997).

132. Chu, W., Zhao, Z., Sun, W., Ye, X., and Wu, Y., Catal. Lett. 55, 57 (1998).

133. Baronetti, G., Thomas, H., and Querini, C.A., Appl. Catal. A 217, 131 (2001).

134. Ramos-Galvan, C.E., Dominguez, J.M., Sandoval-Robles, G., Mantilla, A., and Ferrat, G.,

Catal. Today 65, 391 (2001).

135. Corma, A., Martinez, A., Arroyo, P.A., Monteiro, J.L.F., and Sousa-Aguiar, E.F., Appl.

Catal. A 142, 139 (1996).

136. Corma, A., Martinez, A., and Martinez, C., Appl. Catal. A 134, 169 (1996).

137. Fritz, P.O., and Lunsford, J.H., J. Catal. 118, 85 (1989).

138. Scherzer, J., Catal. Rev.—Sci. Engng 31, 215 (1989).

139. Weitkamp, J., and Traa, Y., Catal. Today 49, 193 (1999).

140. Corma, A., Gomez, V., and Martinez, A., Appl. Catal. A 119, 83 (1994).

141. Unverricht, S., Ernst, S., and Weitkamp, J., Stud. Surf. Sci. Catal. 84, 1693 (1994).

142. Loenders, R., Jacobs, P.A., and Martens, J.A., J. Catal. 176, 545 (1998).

143. Yoo, K., and Smirniotis, P.G., Appl. Catal. A 227, 171 (2002).

144. Agaskar, P.A., and Huang, T.J., US Patent 5,824,835 (1998).

145. Corma, A., and Orchilles, A.V., Microp. Mesop. Mater. 35-36, 21 (2000).

146. Wielers, A.F.H., Vaarkamp, M., and Post, M.F.M., J. Catal. 127, 51 (1991).

147. Garwood, W.E., Leaman, W.K., and Plank, C.J., US Patent 3,251,902 (1966).

148. Weitkamp, J., Stud. Surf. Sci. Catal. 5, 65 (1980).

149. Kashkovskii, V.I., Galich, P.N., Patrilyak, K.I., and Royev, L.M., Petrol. Chem. 31, 480

(1991).

150. Galich, P.N., Tsupryk, I.N., and Vasil’yev, A.N., Petrol. Chem. 37, 332 (1997).

151. Kirsch, F.W., Barmby, D.S., and Potts, J.D., US Patent 4,300,015 (1981).

152. Mortikov, E.S., Plakhotnic, V.A., and Dolinsky, S.E., German Patent DE 197 45 548 A1

(1999).

153. Dolinsky, S.E., Plakhotnik, V.A., Mortikov, E.S., and Bachurikhin, A.L., Prepr. Div.

Petrol. Chem., Am. Chem. Soc. 46, 246 (2001).

154. Park, H.S., and Seff, K., J. Phys. Chem. B 104, 2224 (2000).

155. Scherzer, J., Bass, J.L., and Hunter, F.D., J. Phys. Chem. 79, 1194 (1975).

156. Carvajal, R., Chu, P.J., and Lunsford, J.H., J. Catal. 125, 123 (1990).

157. Bolton, A.P., J. Catal. 22, 9 (1971).

158. Cheetham, A.K., Eddy, M.M., and Thomas, J.M., J. Chem. Soc., Chem. Commun. 1337

(1984).

159. Ward, J.W., J. Catal. 13, 321 (1969).

160. Ward, J.W., J. Colloid Interf. Sci. 28, 269 (1968).

161. Ward, J.W., J. Catal. 14, 365 (1969).






167. Mostad, H., Stocker, M., Karlsson, A., Junggreen, H., and Hustad, B., Stud. Surf. Sci.

Catal. 105, 1413 (1997).

168. Chester, A.W., and Chu, Y.F., US Patent 4,377,721 (1983).

169. Buchanan, J.S., and Huang, T.J., PCT Patent WO 97/20787 (1997).

170. Weitkamp, J., and Jacobs, P.A., in “New Frontiers in Catalysis” (L. Guczi, F. Solymosi

and P. Tetenyi, Eds.), Vol. B, p. 1735. Akademiai Kiado, Budapest, 1993.

171. Chu, C.T.W., Husain, A., Keville, K.M., and Lissy, D.N., US Patent 5,516,962 (1996).


172. Husain, A., PCT Patent WO 94/03415 (1994).

173. Child, J.E., Huss Jr., A., Krambeck, F.J., Ragonese, F.P., Thomson, R.T., and Yurchak, S.,

US Patent 5,073,665 (1991).

174. Huss Jr., A., Kirker, G.W., Keville, K.M., and Thomson, T.J., US Patent 4,992,615 (1991).

175. He, J.Y., Nivarthy, G.S., Eder, F., Seshan, K., and Lercher, J.A., Microp. Mesop. Mater.

25, 207 (1998).

176. Yoo, K., Burckle, E.C., and Smirniotis, P.G., Catal. Lett. 74, 85 (2001).

177. Borade, R.B., and Clearfield, A., J. Phys. Chem. 96, 6729 (1992).

178. Corma, A., Diaz-Cabanas, M.J., Martinez, C., and Valencia, S., “Proceedings of the 13th

International Zeolite Conference, Montpellier,” 2001, p. 275.

179. Nivarthy, G.S., Feller, A., Seshan, K., and Lercher, J.A., Stud. Surf. Sci. Catal. 130, 2561

(2000).

180. Rørvik, T., Mostad, H., Ellestad, O.H., and Stocker, M., Appl. Catal. A 137, 235 (1996).

181. Zhang, Z., Han, Y., Xiao, F.S., Qiu, S., Zhu, L., Wang, R., Yu, Y., Zhang, Z., Zou, B.,

Wang, Y., Sun, H., Zhao, D., and Wei, Y., J. Am. Chem. Soc. 123, 5014 (2001).

182. Huang, T.J., and Kresge, C.T., US Patent 5,498,817 (1996).

183. Husain, A., Huss, A., Klocke, and D.J., Timken, H.K.C., PCT Patent WO 93/07106

(1993).

184. Chu, C.T., Husain, A., Huss Jr., A., Kresge, C.T., and Roth, W.J., US Patent 5,258,569

(1993).

185. Harandi, M.N., Beech Jr., J.H., Huss Jr., A., Ware, R.A., and Husain, A., US Patent 5,625,

113 (1997).

186. Husain, A., Huss Jr., A., and Rahmim, I.I., US Patent 5,475,175 (1995).

187. Huss Jr., A., Le, Q.N., and Thomson, R.T., US Patent 5,326,922 (1994).

188. Corma, A., Juan-Rajadell, M.I., Lopez-Nieto, J.M., Martinez, A., and Martinez, C., Appl.

Catal. A 111, 175 (1994).

189. Farcasiu, D., Ghenciu, A., and Li, J.Q., J. Catal. 158, 116 (1996).

190. Corma, A., Martinez, A., and Martinez, C., Appl. Catal. A 144, 249 (1996).

191. Satoh, K., Matsuhashi, H., and Arata, K., Appl. Catal. A 189, 35 (1999).

192. Gore, R.B., and Thomson, W.J., Appl. Catal. A 168, 23 (1998).

193. Chellappa, A.S., Miller, R.C., and Thomson, W.J., Appl. Catal. A 209, 359 (2001).

194. Das, D., and Chakrabarty, D.K., Energy Fuels 12, 109 (1998).

195. Guo, C., Yao, S., Cao, J., and Qian, Z., Appl. Catal. A 107, 229 (1994).

196. Guo, C., Liao, S., Qian, Z., and Tanabe, K., Appl. Catal. A 107, 239 (1994).

197. Xiao, X., Tierney, J.W., and Wender, I., Appl. Catal. A 183, 209 (1999).

198. Sun, M., Sun, J., and Li, Q., Chem. Lett. (1998), 519 (1998).

199. Blasco, T., Corma, A., Martinez, A., and Martinez-Escolano, P., J. Catal. 177, 306 (1998).

200. Gayraud, P.Y., Stewart, I.H., Derouane-Abd Hamid, S.B., Essayem, N., Derouane, E.G.,

and Vedrine, J.C., Catal. Today 63, 223 (2000).

201. Okuhara, T., Yamashita, M., Na, K., and Misono, M., Chem. Lett. (1994), 1451 (1994).

202. de Angelis, A., Ingallina, P., Berti, D., Montanari, L., and Clerici, M.G., Catal. Lett. 61, 45

(1999).

203. Zhao, Z., Sun, W., Yang, X., Ye, X., and Wu, Y., Catal. Lett. 65, 115 (2000).

204. Rørvik, T., Dahl, I.M., Mostad, H.B., and Ellestad, O.H., Catal. Lett. 33, 127 (1995).

205. Botella, P., Corma, A., and Lopez-Nieto, J.M., J. Catal. 185, 371 (1999).

206. Hommeltoft, S.I., Appl. Catal. A 221, 421 (2001).

207. Clet, G., Goupil, J.M., Szabo, G., and Cornet, D., Appl. Catal. A 202, 37 (2000).

208. Nascimento, P., Szabo, G., and Milan, A., US Patent 6,225,517 (2001).

209. Kojima, M., and Kocal, J.A., US Patent 5,391,527 (1995).


210. Funk, G.A., Hobbs, S.A., Oroskar, A.R., Gembicki, S.A., and Kocal, J.A., US Patent 5,

523,503 (1996).

211. Zhang, S.Y.F., Gosling, C.D., Sechrist, P.A., and Funk, G.A., US Patent 5,675,048 (1997).

212. McBride, T.K., Bricker, M.L., and Steigleder, K.Z., US Patent 5,744,682 (1998).

213. Kocal, J.A., and Oroskar, A.R., US Patent 5,849,977 (1998).

214. McBride, T.K., Bricker, M.L., and Steigleder, K.Z., US Patent 5,883,039 (1999).

215. Herbstman, S., Cole, E.L., and Estes, J.H., US Patent 4,138,444 (1979).

216. Mantilla-Ramirez, A., Ferrat-Torres, G., Dominguez, J.M., Aldana-Rivero, C., and

Bernal, M., Appl. Catal. A 143, 203 (1996).

217. Chou, T.S., Huss Jr., A., Kennedy, C.R., Kurtas, R.S., and Tabak, S.A., PCT Patent WO

90/00534 (1990).

218. Huss Jr., A., and Johnson, I.D., US Patent 5,221,777 (1993).

219. Chou, T.S., Huss, Jr., A., Kennedy, C.R., and Kurtas, R.S., PCT Patent WO 90/00533

(1990).

220. de Angelis, A., Flego, C., Ingallina, P., Montanari, L., Clerici, M.G., Carati, C., and

Perego, C., Catal. Today 65, 363 (2001).

221. Ackerman, S., Chitnis, G.K., and McCaffrey, D.S. Jr., Prepr. Div. Petrol. Chem., Am.

Chem. Soc. 46, 241 (2001).

222. Taylor, R., and Sherwood, D.E. Jr., Appl. Catal. A 155, 195 (1997).

223. Clark, M.C., and Subramaniam, B., Ind. Engng Chem. Res. 37, 1243 (1998).

224. Santana, G.M., and Akgerman, A., Ind. Engng Chem. Res. 40, 3879 (2001).

225. Ginosar, D.M., Thompson, D.N., Coates, K., and Zalewski, D.J., Ind. Engng Chem. Res.

41, 2864 (2002).

226. Querini, C.A., Catal. Today 62, 135 (2000).

227. Jones, E.K., Adv. Catal. 10, 165 (1958).

228. Cupit, C.R., Gwyn, J.E., and Jernigan, E.C., Petrol. Chem. Engng 33, 47 (1961).

229. Daage, M., and Fajula, F., Bull. Soc. Chim. Fr. 5/6, 160 (1984).

230. Nivarthy, G.S., Feller, A., Seshan, K., and Lercher, J.A., Microp. Mesop. Mater. 35/36, 75

(2000).

231. Schweitzer, E.J.A., Barendregt, S., and van der Oosterkamp, P.F., “Proceedings of the

Petrotech 98,” Bahrain, 81, 1998.

232. Dolinsky, S.E., and Plakhotnik, V.A., “Proceedings of the 13th International Zeolite

Conference, Montpellier,” 2001, p. 272.

233. Dolinskii, S.E., Subbotin, A.N., Lishchiner, I.I., Plakhotnik, V.A., and Mortikov, Y.S.,

Petrol. Chem. 37, 447 (1997).

234. Cornet, D., Goupil, J.M., Szabo, G., Poirier, J.L., and Clet, G., Appl. Catal. A 141, 193

(1996).

235. Goupil, J.M., Poirier, J.L., and Cornet, D., Ind. Engng Chem. Res. 33, 712 (1994).

236. http://www.fuelstechnology.com/soft_processoverview.htm, 2000.

237. Black, S.M., Gosling, C.D., Steigleder, K.Z., and Shields, D.J., “Proceedings of the NPRA

Annual Meeting, San Antonio, TX, USA,” 2000.

238. D’Amico, V.J., van Broekhoven, E.H., Nat, P.J., Nousiainen, H., and Jakkula, J.,

“Proceedings of the NPRA Annual Meeting, San Antonio, TX, USA” 2002.

239. van Broekhoven, E.H., Mas Cabre, F.R., Bogaard, P., Klaver, G., and Vonhof, M.,

US Patent 5,986,158 (1999).

240. Buchold, H., Dropsch, H., and Eberhardt, J., “Proceedings of the 17th World Petroleum

Congress, Brazil, 2002,” Vol. 3, p.189, Institute of Petroleum, London, 2003.

241. Sarup, B., Hommeltoft, S.I., Sylvest-Johansen, M., and Søgaard-Andersen, P., “Proceed-

ings of the DGMK-Conference Catalysis on Solid Acids and Bases,” Berlin 1996, p. 175.


Catalytic Conversion of Methane to

Synthesis Gas by Partial Oxidation

and CO2 Reforming

YUN HANG HU and ELI RUCKENSTEIN

Department of Chemical Engineering, State University of New York at Buffalo,

Buffalo, NY 14260, USA

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

II. Partial Oxidation of Methane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

II.A. Hot Spots in Catalyst Beds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

II.B. Minimizing O2 Purification Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

II.C. Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312

II.D. Reaction Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

II.D.1. Changes in Catalyst During Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

II.D.2. Which is the Primary Product, CO or CO2? . . . . . . . . . . . . . . . . . . . . . . . 316

II.D.3. CHx Species and Rate-Determining Steps . . . . . . . . . . . . . . . . . . . . . . . . 318

II.D.4. Comparison of Reactions on Reduced and Unreduced Catalysts . . . . . . . . 320

III. CO2 Reforming of Methane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

III.A. Carbon Formation on Metal Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

III.B. Critical Issues Related to Carbon Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 322

III.C. Supported Noble Metal Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

III.D. Non-Noble Metal Supported Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324

III.D.1. Ni/Al2O3 Catalysts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

III.D.2. Ni/SiO2 Catalysts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327

III.D.3. Ni/La2O3 Catalysts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328

III.D.4. Ni/ZrO2 Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330

III.D.5. Other Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331

III.E. MgO-Containing Solid-Solution Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332

III.E.1. Characteristics of MgO-Containing Solid-Solution Catalysts. . . . . . . . . . 332

III.E.2. Highly Effective MgO-Containing Solid-Solution Catalysts . . . . . . . . . . 333

IV. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

The preparation of synthesis gas from natural gas, which is the most important step in the

gas-to-liquid transformation, has attracted increasing attention in the last decade. Steam

reforming, partial oxidation, and CO2 reforming are the three major processes that can be

employed to prepare synthesis gas. Because steam reforming was reviewed recently in this

series [Adv. Catal. 47 (2002) 65], this chapter deals only with the latter two processes.


Y. H. Hu and E. Ruckenstein / Adv. Catal. 48 (2004) 297–345

The history of the development of methane conversion to synthesis gas is summarized as an

introduction to the partial oxidation of methane, which is reviewed with emphasis on hot spots

in reactors, major developments in the reduction of O2 separation costs, and reaction

mechanisms. The various catalysts employed in CO2 reforming are examined, with emphasis

on inhibition of carbon deposition. q 2004 Elsevier Inc.

Abbreviations

DRIFT diffuse reflectance infrared Fourier transform

EDS energy dispersive X-ray spectrometer

FTIR Fourier transform infrared

GHSV gas hourly space velocity

MIEC mixed ionic/electronic conductors

MS mass spectrometer

R reaction rate

SPARG sulfur passivated reforming

TEM transmission electron microscopy

TG/DTG thermal gravimetric/differential thermal gravimetric

TPD temperature-programmed decomposition

TPH temperature-programmed hydrogenation

TPO temperature-programmed oxidation

TPR temperature-programmed reduction

TPSR temperature-programmed surface reaction

WHSV weight hourly space velocity


w/o water-in-oil

XPS X-ray photoelectron spectroscopy

I. Introduction

In the 1930s, Standard Oil of New Jersey (1) was the first company to employ on

a commercial scale the indirect conversion of methane, the main component of

natural gas, via steam reforming to give synthesis gas, which is a mixture of H2

and CO, with the H2/CO ratio depending on the reactant composition. CO2 is

also formed in synthesis gas production, and sulfur compounds are present as

impurities. Synthesis gas can be used as a feedstock for numerous chemicals and

fuels and as a source of pure hydrogen or carbon monoxide.

The steam reforming process is widely employed today (2). The reaction

CH4 þ H2O ! CO þ 3H2; DH0298 ¼ 206 kJ mol21 ð1Þ

is expensive because of its endothermic nature, the requirement for low space

velocities, and the high H2/CO ratio (3/1), which is unsuitable for synthesis of

methanol or the long-chain hydrocarbons made in the Fischer–Tropsch process.


The other two main processes for conversion of methane into synthesis gas are

partial oxidation and CO2 reforming. In the 1940s, Prettre et al. (3) first reported

the formation of synthesis gas by the catalytic partial oxidation of CH4

CH4 þ12

O2 ! CO þ 2H2; DH0298 ¼ 236 kJ mol21 ð2Þ

They used a Ni-containing catalyst. In contrast to steam reforming of methane,

methane partial oxidation is exothermic. However, the partial oxidation requires

pure oxygen, which is produced in expensive air separation units that are

responsible for up to 40% of the cost of a synthesis gas plant (2) (in contrast, the

steam reforming process does not require pure oxygen). Therefore, the catalytic

partial oxidation of methane did not attract much interest for nearly half a century,

and steam reforming of methane remained the main commercial process for

synthesis gas manufacture.

CO2 reforming,

CH4 þ CO2 ! 2CO þ 2H2; DH0298 ¼ 247 kJ mol21 ð3Þ

was investigated as early as 1888 (4). Although this process, like steam reforming,

is also endothermic, it produces synthesis gas with a lower H2/CO ratio than steam

reforming, and is, therefore, suitable for the Fischer–Tropsch synthesis of long-

chain hydrocarbons (5). Furthermore, it can be carried out with natural gas from

fields containing large amounts of CO2, without the pre-separation of CO2 from

the feed. Because CO2 is a greenhouse gas that causes warming of the earth

and climate change, there are incentives for reducing its concentration in the

atmosphere (6). CO2 reforming of methane may provide a practical method for

consumption of the two greenhouse gases—CH4 and CO2. Unfortunately, no

industrial technology for CO2 reforming of methane has yet been developed,

because no effective, economic catalysts have been discovered (7); furthermore,

high energy costs may be another drawback preventing commercialization.

When the conventional Ni-containing catalyst for steam reforming was used for

CO2 reforming, carbon deposits formed on the catalyst, which deactivated rapidly,

at least in the absence of steam. A high molar ratio of CO2 to CH4 ($3) could be

used to reduce the carbon deposition by inhibiting CO disproportionation, but the

selectivity to synthesis gas was found to become much lower than that for the

stoichiometric CO2 reforming (CO2/CH4 ¼ 1, molar). Therefore, the inhibition of

carbon deposition without extra cost and loss of catalyst performance constitutes a

major challenge for CO2 reforming of methane.

In the 1980s, the oxidative coupling of methane to give ethylene and ethane

was reported by Keller and Bhasin (8), whose discovery prompted numerous

attempts to convert methane directly—and not only to ethylene and ethane (8),

but also to methanol and formaldehyde (9) (Table I). Research on oxidative

coupling of methane was motivated by results showing that the methane was

Y. H. Hu and E. Ruckenstein / Adv. Catal. 48 (2004) 297–345 299

converted into hydrocarbons with higher boiling points, which can be more

economically transported than methane; the transportation issue is important

because substantial reserves of methane are located in remote places.

The reported results show that Li/MgO, with or without promoters, is the best

known catalyst (11). However, Pasquon (12) reported that the best result obtained

in long-run tests has been a C2þ yield of only 15% for methane conversions of

15–40%, at 1270–1370 K and a pressure of 1–2 bar, when a 5–10 CH4/O2

molar ratio was used. In the early 1990s, a consensus emerged that it would be

very difficult to achieve a significantly better result than that mentioned above for

the oxidative coupling to become an economical industrial process. The reason is

that the formation of CO2, rather than of more desirable products (ethylene,

ethane, methanol, and formaldehyde), is favored thermodynamically (Table I)

when the reactions of methane and oxygen become fast enough to be of practical

interest (typically at temperatures exceeding 973 K).

Consequently, in the early 1990s, interest in the direct processes decreased

markedly, and the emphasis in research on CH4 conversion returned to the

indirect processes giving synthesis gas (13). In 1990, Ashcroft et al. (13) reported

some effective noble metal catalysts for the reaction; about 90% conversion of

methane and more than 90% selectivity to CO and H2 were achieved with a

lanthanide ruthenium oxide catalyst (L2Ru2O7, where L ¼ Pr, Eu, Gd, Dy, Yb or

Lu) at a temperature of about 1048 K, atmospheric pressure, and a GHSV of

4 £ 104 mL (mL catalyst)21 h21. This space velocity is much higher than that

employed by Prettre et al. (3). Schmidt et al. (14–16) and Choudhary et al. (17)

used even higher space velocities (with reactor residence times close to 1023 s).

TABLE I

Gibbs free energy change, DG0; for methane transformation reactions (10)

Reaction DG0 (kcal mol21)

400 K 600 K 800 K 1000 K

CH4 þ12

O2 !12

C2H4 þ H2O 234.6 235.1 235.8 236.4

CH4 þ14

O2 !12

C2H6 þ12

H2O 218.4 217.1 215.8 214.5

CH4 þ12

O2 ! CH3OH 225.4 223.0 220.5 218.0

CH4 þ O2 ! HCHO þ H2O 269.0 270.0 270.8 271.2

CH4 þ 2O2 ! CO2 þ 2H2O 2191.3 2191.3 2191.3 2191.3

CH4 þ12

O2 ! CO þ 2H2225.0 233.9 243.1 252.5

CH4 þ H2O ! CO þ 3H2 28.6 17.3 5.5 26.5


An advantage of the high space velocities is the relatively low mass-transfer

resistances associated with them.

The catalytic partial oxidation of methane to CO is exothermic, and even a low

conversion to CO2 generates a large amount of heat, which leads to significant

temperature gradients (hot spots) in the reactor; the temperature may vary by

several hundred kelvin over a distance of only 1 mm from the hot spot. Because

the partial oxidation is a fast reaction, it is difficult to remove the heat from the

reactor as fast as it is generated, particularly from a large-scale reactor. As a

result, the process is potentially hazardous and can lead to explosions. The partial

oxidation process requires a pure oxygen feed and, therefore, a unit to prepare

oxygen by air separation. Therefore, one of the major research goals for making

the catalytic partial oxidation a commercial reality is to reduce the cost of the air

separation.

The reaction pathways for the partial oxidation reaction are still debated.

According to one interpretation, CO2 and H2O are the primary products, and CO

is formed by the reaction of CO2 or H2O with CH4; according to another

interpretation, CO is produced directly by the reaction of CH4 with O2.

In summary, major challenges in the partial oxidation of methane are: (1)

designs to avoid excessive thermal gradients (hot spots) in the catalyst bed; (2)

reduction of the cost of O2 separation; and (3) elucidation of the reaction

pathways as a step toward improved catalyst design.

The purpose of this chapter is to provide a critical assessment of the literature

regarding the partial oxidation of methane and the CO2 reforming of methane,

with emphasis on the following challenging areas: hot spots, O2 separation cost,

and the issues of reaction pathways and catalyst selection; we also address the

issue of carbon deposition in the CO2 reforming of methane. The reason why we

review these two reactions together is that they have many common

characteristics, including the catalysts, the products, and CH4 as reactant.

II. Partial Oxidation of Methane

II.A. Hot Spots in Catalyst Beds

In the early 1990s, several papers (17–20) reported that one can reach CO

and H2 concentrations in excess of those expected at thermodynamic

equilibrium by operating the CH4 oxidation reaction at exceptionally high

space velocities (GHSV ¼ 52,000 mL (g catalyst)21 h21) in a fixed-bed

reactor. The following catalysts were employed: Ni/Yb2O3 (18), Co/rare

earth oxide (19), Co/MgO (20), and Ni/Al2O3 (17). However, the actual

reaction temperatures (21) could have been much higher than those reported

(17–20). By using an optical pyrometer, Lunsford et al. (22) found that,


during the catalytic oxidation of methane to CO and H2, the combination of

a high space velocity, an exothermic reaction, and an active catalyst (Ni/

Yb2O3) gave rise to steep temperature gradients (hot spots). Furthermore, the

temperature of the hot spot was greater by as much as 573 K than the

temperature measured with a thermocouple located at a distance of only

1 mm from the hot spot in the catalyst bed. If a temperature lower than that

of the hot spot is used to calculate the equilibrium concentrations of CH4,

CO, CO2, and H2, one can draw the conclusion that the concentrations of

CO and H2 exceeded their thermodynamic equilibrium values. However, if

the true maximum (hot spot) temperature is used in the calculation, the

observed concentrations are found to be somewhat less than those predicted

at equilibrium. Indeed, using a careful temperature measurement method, in

which a thermocouple end contacted just the top surface of the catalyst bed,

Hu et al. (23,24) found that the CH4 conversion in the presence of Ni/Al2O3

catalyst was less than that predicted by thermodynamic equilibrium.

Furthermore, Hu and Ruckenstein (25) observed hot layers (thinner than

1 mm) in NiO/MgO solid solution catalysts and in NiO/Al2O3 and NiO/SiO2

catalysts during the partial oxidation of methane in a fixed-bed reactor. The

hottest layers were located at the top of the bed of the NiO/MgO and NiO/Al2O3

catalysts, but they were observed to move down and then up for the NiO/SiO2

catalyst bed. The down-and-up movement resulted in an oscillatory temperature

of the NiO/SiO2 catalyst at a given position in the bed (Fig. 1), which was

absent when the catalyst was NiO/MgO or NiO/Al2O3 (Fig. 2).

The different temperature behaviors of the three catalysts were attributed to the

different strengths of the interactions between the metal oxide and the support.

Temperature-programmed reduction (TPR) experiments with 4% H2 in argon

indicated that the initial reduction temperature was about 3308C for 13.6 wt%

NiO/SiO2, which is near that of pure NiO (about 3008C) (26). In contrast, for

13.6 wt% NiO/Al2O3 the initial reduction temperature was high (6708C) and no

marked reduction peak could be detected even at 8008C for 13.6 wt% NiO/MgO.

These results clearly indicate that there are weak interactions between NiO and

SiO2 and much stronger interactions between NiO and Al2O3 and between NiO

and MgO.

The weak interactions in Ni/SiO2 might have been responsible for the

temperature oscillation by allowing a facile redox behavior of the active nickel

sites, namely, the oxidation of Ni0 to NiO by O2 and the reduction of NiO to Ni0

by CH4. The strong interactions characteristic of NiO/Al2O3 and NiO/MgO were

inferred (25,26) to inhibit in part the redox behavior of the nickel sites. In the case

of NiO/SiO2, according to this interpretation, the freshly reduced NiO located at

the inlet of the bed became highly active, causing a hot layer to be generated. The

high temperature of this hot layer resulted in sintering of the nickel particles,

which led to the loss of activity. Therefore, the reaction is inferred to have taken


place in the neighboring section of the catalyst. As a result, a hot layer propagated

downward in the reactor. However, the sintered nickel particles were re-dispersed

on the SiO2 support when they were reoxidized by O2, because the oxygen

concentration is high when the reaction of CH4 with O2 does not take place. After

a certain time, the reoxidized layer near the entrance was again reduced by CH4

and became active again, resulting in a hot layer. The following part of reoxidized

nickel on SiO2 can be reduced rapidly by H2 and CO generated near the entrance

of the reactor. The redox of the Ni/SiO2 catalyst constitutes a cycle of

deactivation and reactivation in each part of the catalyst. The hot layer moved

downward in the bed during the time required for the reduction of the entrance

layer. Consequently, the time scale of the oscillations was determined by the time

scale of the reduction–oxidation process.

Recently, such a temperature oscillation was also observed by Zhang et al.

(27,28) with nickel foils. Furthermore, Basile et al. (29) used IR thermography to

monitor the surface temperature of the nickel foil during the methane partial

oxidation reaction by following its changes with the residence time and reactant

concentration. Their results demonstrate that the surface temperature profile was

strongly dependent on the catalyst composition and the tendency of nickel to be

oxidized. Simulations of the kinetics (30) indicated that the effective thermal

conductivity of the catalyst bed influences the hot-spot temperature.

Fig. 1. Relationship between catalyst temperature and reaction time in methane partial oxidation

catalyzed by Ni/SiO2 (temperature of the gas phase: (a) 1019 K, (b) 899 K, (c) 809 K, (d) 625 K). The

reaction was carried out in a fixed-bed reactor (a quartz tube of 2 mm inside diameter) at atmospheric

pressure. Before reaction, the feed gas was allowed to flow through the catalyst undergoing heating of

the reactor from room temperature to 1073 K at a rate of 25 K min21 to ignite the reaction, and then the

reactant gas temperature was decreased to the selected value. Reaction conditions: pressure, 1 atm;

catalyst mass, 0.04 g; feed gas molar ratio, CH4/O2 ¼ 2/1; GHSV, 90,000 mL (g catalyst)21 h21) (25).


Researchers have attempted to minimize thermal gradients, for example, by

using fluidized-bed reactors (31–33). Olsbye et al. (31) investigated methane

partial oxidation in a fixed bed and in a fluidized-bed reactor with a 1.5 wt% Ni/

Al2O3 catalyst operated at 973 K, with a feed flow rate of about 400

(STP) mL min21 (CH4/O2/N2/H2O ¼ 2/1/2/0.5, molar) and a catalyst volume

of 17 mL. They observed that the maximum temperature difference was only

282 K in the fluidized bed, but 423 K in the fixed bed, indicating that the fluidized-

bed reactor is a good heat exchanger because of the rapid mixing of the fluid and

the catalyst.

Another way to minimize the temperature gradient (34–40) is to combine the

exothermic partial oxidation with an endothermic reaction. Ioannides and

Verykios (34) developed a novel reactor consisting of a ceramic tube with metal

catalyst films deposited on the inner and outer surfaces. The CH4/O2 feed enters

into the tube, and a large fraction of the heat generated by the methane

combustion reactor is transferred through the tube wall towards the outer catalyst

film, where an endothermic reforming reaction takes place. With this design,

the temperature in the combustion zone is controlled and the hot spots are

significantly reduced in magnitude.

Fig. 2. Relationship between catalyst temperature and reaction time for reaction catalyzed by

Ni/Al2O3(- - -) and Ni-MgO solid solutions (—); temperature (K) of the gas phase: (a) 1019; (b) 899:

(c) 809; (d) 625. The reaction was carried out in a fixed-bed reactor (a quartz tube of 2 mm inside

diameter) at atmospheric pressure. Before reaction, the feed gas was allowed to flow through the

catalyst undergoing heating of the reactor from room temperature to 1073 K at a rate of 25 K min21 to

ignite the reaction, and then the reactant gas temperature was decreased to the selected value. Reaction

conditions: pressure, 1 atm; catalyst mass, 0.04 g; feed gas molar ratio, CH4/O2 ¼ 2/1; GHSV,

90,000 mL (g catalyst)21 h21) (25).


Coupling of the endothermic CO2 reforming of methane with the exothermic

catalytic partial oxidation of methane can, in addition to overcoming the hazard

of overheating, also provide a control of the H2/CO ratio and thus the selectivity

for various Fischer–Tropsch synthesis products. Aschcroft et al. (35) carried out

this combination of reactions with an Ir/Al2O3 catalyst, obtaining synthesis gas

yields of up to 90% (Table II). However, they found that when nickel-containing

catalysts were used, carbon deposits were formed rapidly, except when an excess

of CO2 was used. Choudhary et al. (41,42) reported that a NiO–CaO catalyst for

15 h exhibited a conversion .95%, with 100% CO selectivity and .90% H2

selectivity, without catalyst deactivation caused by carbon deposition. Further-

more, Ruckenstein and Hu (37) found that the reduced NiO/MgO catalyst

provided a high activity and selectivity, as well as excellent stability in the

combination process, even when no excess of CO2 was used. They carried out the

combined reaction catalyzed by each of the following: a NiO/MgO solid solution,

NiO/Al2O3, and NiO/SiO2. A CH4 conversion of about 90% and selectivities to

CO and H2 of about 98% were achieved at 1063 K and a GHSV of

90,000 mL (g catalyst)21 h21 (O2/CO2/CH4 ¼ 14.5/26.9/58.6) when a reduced

NiO/MgO solid-solution catalyst was used. Almost no change in activity or

selectivity occurred during 50 h of reaction. Compared with the reduced NiO/

MgO, the reduced NiO/SiO2 and NiO/Al2O3 catalysts provided lower activities

and stabilities. Furthermore, Ruckenstein and Hu (37) observed a decrease in the

CH4 conversion with increasing space velocity, whereas during the partial

oxidation alone, because of the hot spots, it would have increased (43). This

observation implies that the coupling can, indeed, control the thermal behavior of

TABLE II

Results of catalytic reactions with mixtures of CH4, O2, and CO2 of different compositions in the

presence of 1 wt% Ir/Al2O3 at 1050 K (35)

Feed composition (mol%) CH4 converted

(%)

CO2 converted

(%)

H2 yield

(%)

CO yield

(%)

CH4 CO2 O2

64.4 3.5 32.1 92 9 89 86

59.4 20.0 20.6 87 83 81 86

58.3 23.7 18.0 84 83 81 84

58.0 28.0 14.0 83 90 79 85

49.8 48.8 1.4 91 87 91 89

Total gas hourly space velocity, 2 £ 104 mL (mL catalyst)21 h21; pressure, 1 atm. In all the cases, the

oxygen conversions were .99.7%.


the reactor. Ruckenstein and Wang (39) found that the Co/MgO solid solution is

also an effective catalyst for the combined reaction.

Steam reforming of methane, which like CO2 reforming is endothermic, has also

been combined with the exothermic partial oxidation of methane (44–47). This

combination process is usually called “autothermal reforming”, because no heat

addition is required for the reforming reaction. For example, ExxonMobil (44–46,

48–50) extended its experience with fluidized-bed catalytic cracking to the

synthesis gas production, developing a process in which the steam reforming was

combined with partial oxidation of the natural gas in a single fluidized-bed reactor.

II.B. Minimizing O2

Purification Costs

Although the partial oxidation of methane with air as the oxidant would at first

seem to be a potential alternative to the steam-reforming process, the downstream

processing requirements in the conventional process do not tolerate nitrogen

(because the cost of compression of synthesis gas diluted by nitrogen to pressures

.20 atm, which is necessary for downstream industrial processes, is high), and,

therefore, pure oxygen must be used. An important advance in the direction

of making air a feedstock resulted from the use of an inorganic membrane reactor

(51–75). The reactor consists of a dense ceramic membrane (made from mixtures

of ionic and electronic conductors, such as SrFeCo0.5Ox (51)) that is permeable

only to oxygen; application of this reactor can, in principle, reduce the entire

synthesis gas process to a single step, allowing elimination of the oxygen

plant and decreasing the total cost of the synthesis gas production by 25–40%.

For this reason, this inorganic membrane process has attracted significant

commercial interest.

Solid electrolytes are materials that exhibit high ionic conductivities (76). If a

solid electrolyte is a pure ionic conductor, the transference number for ions is

two or more orders of magnitude greater than that for electrons. Yttria-stabilized

zirconia, a pure ionic conductor, is the classical solid electrolyte for solid-state

transport of oxygen. However, a system based on a classical solid electrolyte for

ionic oxygen transport requires electrodes to transfer the electrons to the reduction

interface from the oxidation interface (Fig. 3a). In contrast, the perovskites of the

ABO3 type (with the CaTiO3 structure) with dopants in the A and/or B sites, called

mixed ionic/electronic conductors (MIEC), provide high conductivities for both

oxygen ions and electrons (54–62) (Fig. 3b). The MIEC membrane can be used for

the O2 separation without electrodes. The driving force for the overall oxygen

transport is the gradient of the oxygen partial pressure across the membrane (77).

The dissociation and ionization of oxygen to generate oxygen ions, by capturing

the electrons provided by accessible surface electronic states, occur at the oxide

surface at the high-pressure feedside. The flux of oxygen ions and the reverse flux


of electronic charge carriers across the MIEC membrane constitute a charge-

compensation process. The individual oxygen ions from the high-pressure

feedside separate from their electrons and recombine again, at the low-pressure

permeateside, to form O2 molecules that are released into the permeate stream.

Therefore, because of its ability to conduct both oxygen ions and electrons, the

MIEC membrane can operate without electrodes attached to the oxide surface and

without external circuitry.

Extensive research has been carried out with the acceptor-doped perovskite

oxides with the generic formula La12xAxCo12yByO32d (where A ¼ Sr, Ba, or Ca

and B ¼ Fe, Cu, or Ni) (77). Teraoka et al. (54,55,63) were the first to report very

high oxygen fluxes through the cobalt-rich perovskites that can become highly

oxygen anion defective at elevated temperatures and reduced oxygen partial

pressures. The oxygen-ion conductivity in these perovskites can be 1–2 orders of

magnitude greater than those of stabilized zirconias at elevated temperatures,

although in the usual ranges of temperature and oxygen partial pressure, the

electronic conduction of the perovskite remains predominant (78,79).

In the early 1990s, Balachandran et al. (51,64,65) of the Argonne National

Laboratory, in collaboration with Amoco (now part of BP), investigated the partial

oxidation of methane using membrane materials consisting of Sr–Fe–Co–O

mixed oxides with the perovskite structure, which have high oxygen

permeabilities. In their experiments (51,66), the membrane tubes, which were

Fig. 3. Oxygen transport in solids. O2 is dissociated and ionized at the reduction interface to give

O22 ions, which are transferred across the solid to the oxidation interface, at which they lose the

electrons to return back to O2 molecules that are released to the stream. (a) In the solid electrolyte cell

based on a classical solid electrolyte, the ionic oxygen transport requires electrodes and external

circuitry to transfer the electrons from the oxidation interface to the reduction interface; (b) in the

mixed conducting oxide membrane, the ionic oxygen transport does not require electrodes and

external circuitry to transfer the electrons to the reduction interface from the oxidation interface,

because the mixed conductor oxide provides high conductivities for both oxygen ions and electrons.


prepared from an electronic/ionic conductor powder (Sr–Fe–Co–O) by a plastic

extrusion technique, were investigated for their performance in the quartz reactor

sketched in Fig. 4. The quartz reactor supports the ceramic membrane tube with hot

Pyrex seals. A Rh-containing reforming catalyst was located adjacent to the tube.

In this reactor, air could be used directly, because the membrane itself carried out

the separation of oxygen from air. The electrons of the membrane combine with

the oxygen from air to generate oxygen anions. The ions migrate through the

membrane, from the air side to the methane side. At the methane side, the electrons

are stripped from the ions, which are thus converted into oxygen atoms that

combine with methane to form the synthesis gas. The freed electrons migrate back

to the air side of the membrane, generating fresh oxygen anions, and so on. The

experimental results show that the performance of the membrane was strongly

dependent on the composition of the material. The most promising material had the

composition SrFeCo0.5Ox. This membrane operated in a partial oxidation reactor

for more than 1000 h at 1123 K (Fig. 5), whereas other mixed-oxide membranes

fractured rapidly. A methane conversion of 98% with a 90% CO selectivity was

thus achieved. Another advantage of the membrane reactor is that the process does

not involve the handling of potentially explosive CH4/O2 mixtures.

Other early contributions to the membrane processes for partial oxidation of

methane include the following: (a) the La0.2Sr0.8Fe0.8Cr0.2Ox membrane of

Standard Oil Company at Ohio (now part of BPAmoco) (67), which remained

stable for more than 1000 operating hours at 1373 K, and (b) a brownmillerite

membrane with the general composition A2B2O5 (where A and B were not

disclosed), consisting of a layer of BO6 octahedra sharing vertices with a layer of

BO4 tetrahedra (68), which was tested for more than 3000 operating hours at

1173 K and 1 atm with a CO selectivity .96% and a CH4 conversion .80%.

A group at Worcester Polytechnic Institute (69) also investigated the partial

Fig. 4. Configuration of a ceramic membrane reactor for partial oxidation of methane. The

membrane tube, with an outside diameter of about 6.5 mm and a length of up to about 30 cm and a

wall thickness of 0.25–1.20 mm, was prepared from an electronic/ionic conductor powder (Sr–Fe–

Co–O) by a plastic extrusion technique. The quartz reactor supports the ceramic membrane tube

through hot Pyrex seals. A Rh-containing reforming catalyst was located adjacent to the tube (51).


oxidation of methane to give synthesis gas using a mixed-conducting La(12x)-

AxFe0.8Co0.2O32d perovskite dense membrane reactor at 1123 K, in which the

oxygen was separated from air and simultaneously fed into the methane stream.

The steady-state oxygen permeation rates for membranes in non-reacting air/

helium experiments were in the sequence La0.2Ba0.8Fe0.8Co0.2O32d . La0.4

Ba0.6Fe0.8Co0.2O32d . La0.4Ca0.6Fe0.8Co0.2O32d . La0.4Sr0.6Fe0.8Co0.2O32d.

By packing a 5% Ni/Al2O3 catalyst directly on the reaction-side surface of the

membrane, the researchers obtained a fivefold increase in O2 permeation and a

fourfold increase in CH4 conversion. The oxygen, which was continuously

transported from the air side, appeared to stabilize the membrane interior, and the

reactor could be operated for up to 850 h (69,70).

Recently, Li et al. (71) demonstrated a promising application of a Ba0.5Sr0.5

Co0.8Fe0.2O32d membrane for oxygen separation characterized by a high

permeation flux (1.1 mL cm22 min21 at 1123 K) and stability (leak-free during

partial oxidation). A membrane reactor, prepared from a Ba0.5Sr0.5Co0.8Fe0.2

O32d membrane (Fig. 6), was applied successfully to the partial oxidation of

methane (with LiLaNiOx/g-Al2O3 containing 10 wt% Ni as catalyst, located on

the top of the membrane) at 1148 K for about 500 h without failure, with a

methane conversion .97% and a CO selectivity .95% (Fig. 7) (72). A novel

dense catalytic membrane reactor, prepared from the stable conducting

perovskite BaCo0.4Fe0.4Zr0.2O32d and the catalyst LiLaNiO/g-Al2O3 also

Fig. 5. Methane conversion and oxygen flux during partial oxidation of methane in a ceramic

membrane reactor. Reaction conditions: pressure, 1 atm; temperature, 1173 K, feed gas molar ratio,

CH4/Ar ¼ 80/20; feed flow rate, 20 mL min21 (NTP); catalyst mass, 1.5 g; membrane surface area,

8.4 cm2 (51).


demonstrated excellent performance for partial oxidation (73). This membrane

reactor was characterized by a short induction period (2 h), high CH4 conversion

(98%) and CO selectivity (about 99%), and excellent stability (more than 2200

operating hours) at 1123 K.

Since 1997, to accelerate the membrane technology towards commercialization,

two major alliances have been formed, one comprising Amoco (now part of BP),

Praxair, Statoil, Sasol, and Philips, and the other (a US Department of Energy

Fig. 6. Configuration of a ceramic membrane reactor for partial oxidation of methane. The

membrane disk was prepared by pressing Ba0.5Sr0.5Co0.8Fe0.2O32d oxide powder in a stainless steel

module (17 mm inside diameter) under a pressure of (1.3–1.9) £ 109 Pa. The effective area of the

membrane disk exposed to the feed gas (CH4) was 1.0 cm2 (72).


cost-shared project) made up of Air Products, Arco (now part of BP), Argonne

National Laboratory, Babcock and Wilcox, Ceramate (Salt Lake City), Chevron-

Texaco, Eltron Research, Norsk Hydro, Pacific Northwest National Laboratory,

Pennsylvania State University, and the University of Pennsylvania (74).

Notwithstanding the extensive research, there are still hurdles to overcome

(80–82). Although the mixed conducting membranes offer high oxygen fluxes,

they are mechanically and chemically less stable than the traditional stabilized

zirconias. Furthermore, the integration of a ceramic membrane into large-scale

production units will be difficult, because the ceramics break easily and are not

easily manufactured without microscopic voids and fractures. It is also difficult to

connect them to other, more flexible materials such as steel pipes. These critical

issues represent major challenges to the commercialization of MIEC membrane

reactors for the partial oxidation of methane.

Therefore, a team, led by the University of Alaska-Fairbanks, was formed

to study these practical issues (75), including the composition of the ceramic

membrane, seals that would join the ceramic and metal materials, membrane

performance, and development of a ceramic that would resist warping and

fracturing at the high temperatures of the conversion process.

Another way to eliminate the oxygen plant is to react a metal oxide with

methane to yield the synthesis gas in a fluidized-bed reactor (83–86).

Experiments have shown that copper oxide readily oxidizes methane to carbon

monoxide and hydrogen with high selectivity at a temperature of about 1200 K

and that the reduced CuO can be reoxidized with air. Lewis et al. (83–86)

Fig. 7. Methane conversion, CO selectivity, and oxygen flux through the ceramic membrane

during the partial oxidation of methane in a ceramic membrane reactor (see Fig. 6). Reaction

conditions: temperature, 1148 K; catalyst, 300 mg of LiLaNiOx/g-Al2O3; air flow rate, 300 mL min21

(NTP); feed gas molar ratio, CH4/He ¼ 1/1; feed flow rate, 42.8 mL min21 (NTP) (72).


proposed a process using two interconnected fluidized beds—a reactor for the

hydrocarbon oxidation by the metal oxides (Step 1) and a regenerator for the

reoxidation of the reduced metal oxide by air (Step 2). The major advantage of this

process is that air can be used directly without pre-separation. A high conversion

of about 95% and a selectivity of 90% were thus achieved (83–86). However,

metal oxide sintering during the reduction–oxidation cycles could be a difficulty.

II.C. Catalysts

In the 1940s, Prettre et al. (3) reported the formation of synthesis gas via the

catalytic partial oxidation of CH4 catalyzed by a 10 wt% refractory supported

nickel, at temperatures between 973 and 1173 K. Thermodynamic equilibrium

corresponding to the catalyst bed exit temperature was achieved under all

conditions investigated. In 1970, Huszar et al. (87) examined the effect of

diffusion on methane partial oxidation catalyzed by a single grain of Ni/mullite

catalyst in the temperature range of 1033–1173 K and examined the ignition and

extinction characteristics of this catalyst. They observed that the nickel catalyst

deactivated in an oxidative environment but could recover on reduction. In 1984,

Gavalas et al. (88) investigated the effects of the calcination temperature, pre-

reduction, and feed ratio on the reaction of CH4/O2 mixtures catalyzed by NiO/a-

Al2O3 at 843–1033 K. However, under their experimental conditions, the main

products were CO2 and H2O.

Since 1990, researchers (89–148) have continued to examine nickel-contain-

ing catalysts for the partial oxidation of methane, and they also started to use

noble metals as catalysts. In 1990, Ashcroft et al. (13) reported a methane

conversion of about 90% and more than 90% selectivity to CO and H2 at 1043 K,

atmospheric pressure, and at the high GHSV of 4 £ 104 mL (mL catalyst)21 h21

for a reaction catalyzed by lanthanide ruthenium oxides, such as Pr2Ru2O7,

Eu2Ru2O7, Gd2Ru2O7, Dy2Ru2O7, or Lu2Ru2O7. In 1992, Hickman and Schmidt

(14) used platinum monoliths to achieve high selectivities to CO and H2 in the

partial oxidation of methane. In the following 10 years, various noble metal

catalysts have been examined (Table III) (89–106). Compared with the non-

noble metal catalysts, the noble metals exhibit high stability with excellent

activities and selectivities. The major drawback of the noble metal catalysts is

their high cost, which restricts their potential use in industrial processes.

Non-noble metal catalysts, particularly those containing nickel, have also

been investigated extensively since 1990. Lunsford et al. (107) examined a

25 wt% Ni/Al2O3 catalyst in the temperature range 723–1173 K. Carbon

monoxide selectivities approaching 95% and virtually complete conversion of

the methane were achieved at temperatures above 973 K. The authors observed

that, under their operating conditions, the calcined catalyst bed consisted of


TABLE III

Noble metal catalysts for partial oxidation of methane

Metal Support References

Rh Al2O3 (89,94,102,104,106,126)

SiO2 (101)

MgO (101)

a-Al2O3 monolith (16,105)

Pt Al2O3 (89,91,99,126,148)

CeO2/Al2O3 (91)

MgO (95)

ZrO2 (148)

CeO2 (99)

CeO2/ZrO2 (148)

a-Al2O3 monolith (16)

Pt sponge (103)

Pd Al2O3 (89,126)


Ir TiO2 (90,92,96)


Al2O3 (126)

Eu2O3 (126)

Ru SiO2 (43,93)

Al2O3 (10,93,126,127)

YSZ (yttria-stabilized zirconia) (93)

TiO2 (93,97,100)


Pr2O3 (13)

Sm2O3 (13)

Eu2O3 (13)

Gd2O3 (13)

Tb2O3 (13)

Dy2O3 (13)

Tm2O3 (13)

Yb2O3 (13)

Lu2O3 (13)

Re a-Al2O3 monolith (16)


three regions, NiAl2O4 (upstream, section), NiO þ Al2O3 (middle section), and

reduced Ni/Al2O3 (downstream section). In the upstream section of the reactor,

the CH4/O2/He feed contacted NiAl2O4, which exhibited only a moderate

activity for the complete oxidation of methane to CO2 and H2O. The next

section of the reactor contained NiO þ Al2O3, which catalyzed the complete

exothermic oxidation of methane to CO2. Because of the complete consumption

of O2 in the second section, the third (downstream) section of the catalyst bed

consisted of a reduced Ni/Al2O3. The formation of the CO and H2 products,

corresponding to thermodynamic equilibrium at the temperature of the bed

exit, occurred in this section, as a result of the reforming reactions of CH4

with CO2 and H2O produced during the complete oxidation reaction catalyzed

by the NiO/Al2O3.

Choudhary et al. reported a high conversion of CH4 and high selectivities to

CO and H2 with Ni/CaO (17a), Ni/Al2O3 (17b), NiO-rare earth oxide (108), and

Co/rare earth oxide catalysts (19). Hu et al. (23) used a Ni/Al2O3 catalyst for the

adiabatic partial oxidation of methane. The nickel- or cobalt-containing catalysts

exhibited high activities and selectivities to synthesis gas from CH4/O2 mixtures.

The major problem encountered with these non-noble metal catalysts is their

relatively low stability (109–111). The main causes of the deactivation of the

catalysts are carbon deposition and metal sintering in the catalyst. Nevertheless,

numerous effective nickel-containing catalysts have been developed by

incorporation in suitable supports (111–116), such as La2O3 (111), MgO (112,

113), SrTiO3 (114), and CeO2 (115); effective promoters (117–119), including

La2O3 (117,118), Li2O (118), and iron oxide (119); and novel preparation

methods (120–125), such as a solid phase crystallization method (120), a sol–gel

method (122), and a citrate method (125). However, because the high stabilities

reported for these effective nickel-containing catalysts were based on short-term

tests (,100 h), it is unclear how stable these catalysts will be in long term tests

(.1000 h), which is the first step that any candidate catalyst for commercializa-

tion must pass.

II.D. Reaction Pathways

In the last decade, numerous attempts have been made to understand the

mechanism of the partial oxidation of methane (3,13–15,17,25,37,97,107,

128–137,142–148). Mechanistic investigations of the partial oxidation are

still challenging, because this exothermic reaction is very fast and causes

extremely high catalyst temperature rises, so that the usual methods of

investigation are unsuitable.

Two kinds of pathways have been suggested: (i) a combustion-reforming

pathway, in which CO2 and H2O are the primary products, and CO and H2 are


formed by their reactions with CH4 (3,14,97,107,128,132,133); and (ii) a

pyrolysis pathway, in which CO is the primary product formed by the pyrolysis of

methane, CH4 ! CHx þ ð2 2 12

xÞH2; followed by the oxidation of carbon-

containing species to give CO without the pre-formation of CO2 (17,129,130,

134–137). Thus, the major questions regarding these reactions are: (1) is CO or

CO2 the primary product? (2) What is the rate-determining step? (3) What are the

intermediate species? (4) How does the state of the catalyst change during

reaction? We review these issues in the following sections.

II.D.1. Changes in Catalyst During Reaction

The catalyst surface structure depends on the reactants in contact with it. During

steady-state experiments, the catalyst surface may reach an equilibrium with the

reactants at various positions in the reactor, and so steady-state methods provide

little information about the surface state of the catalyst. On the other hand, pulse

methods, in which a small amount of reactants is injected into the reactant stream,

do not affect the surface of the catalyst significantly during a single pulse.

Therefore, during the first pulse, the reaction can be attributed to the original state

of the catalyst. As additional pulses are introduced, the catalyst surface gradually

changes. Therefore, changes in the selectivities and conversions as a function of

the number of pulses are indicative of the changes in the catalyst. Thus, Hu and

Ruckenstein (134) determined the selectivities and conversions as a function of

the number of pulses of CH4 and O2 using mass spectrometry to analyze the

products. The catalyst was an unreduced NiO/La2O3 or one reduced in H2. As

shown in Fig. 8, when the catalyst was unreduced, the CH4 conversion increased

gradually with the number of pulses, reaching the constant value of about 18%

after the ninth pulse. When the catalyst was initially reduced, the CH4 conversion

was the greatest for the first pulse and after the ninth pulse reached the same

constant value as for the initially unreduced catalyst. The change of the CO

selectivity with the number of pulses of CH4 and O2 was found to be similar to

that observed for the CH4 conversion. This comparison indicates that the initial

oxide and reduced states of the catalyst changed towards the same working state

as the number of CH4/O2 pulses increased. In other words, the oxide state of the

catalyst was partially reduced during catalysis, and the reduced catalyst was

partially oxidized during catalysis. Presumably, a redox equilibrium was finally

attained between the catalyst and the reactant stream. Furthermore, the curves of

oxygen coverage on a reduced Ni/La2O3 catalyst with time during a pulse of

CH4/O2 indicated that oxygen-containing species were easily generated on the

reduced catalyst, and that, after a pulse of CH4 and O2 had reacted completely,

oxygen-containing species were still present on the catalyst surface, hence that

the reduced catalyst had been partially oxidized (129). This inference is

consistent with the X-ray diffraction data of Lunsford et al. (107), which showed


that both reduced and oxidized nickel were present in a Ni/Al2O3 catalyst used for

CH4 oxidation. Because the reactor was a fixed bed, the change in the catalyst

resulting from the interactions between the catalyst and the stream containing

reactants and products were non-uniformly distributed along the catalyst bed

(138). As the catalyst was reduced, the CH4 conversion increased. This result

implies that the reduced nickel is more active than the oxidized nickel for CH4

activation, and that in the reaction between CH4 and the lattice oxygen of NiO,

the CH4 conversion increased when NiO was partially reduced (139,140).

Campbell et al. (141) reported that the reaction probability of methane on NiO

films is significantly lower than that on a clean Ni(100) surface. Furthermore,

results of experiments with deuterium-methane pulses showed that CH4 easily

dissociates into CHx and H on a reduced nickel-containing catalyst, whereas such

a dissociation cannot take place on the catalyst in the oxidized state (131). One

can, therefore, conclude that the reduced nickel, which might be a zero valent

nickel, constitutes the principal active site for the partial oxidation of methane.

II.D.2. Which is the Primary Product, CO or CO2?

To discriminate between the combustion-reforming mechanism and the pyrolysis

mechanism, one must clarify whether CO or CO2 is the primary product

(or whether both are). Typically, to discriminate between primary and secondary

Fig. 8. CH4 conversion as a function of the number of CH4/O2 pulses for partial oxidation of CH4

catalyzed by Ni/La2O3. Reaction conditions: temperature, 873 K; catalyst, 20 mg of 20 wt% Ni/La2O3

loaded in a fixed-bed flow reactor; feed gas, 0.9 mL CH4/O2 (molar ratio 2/1) in each pulse; carrier

gas, helium (flow rate,100 mL min21) (134).


products in a conventional continuous flow reactor, one changes the reactant–

catalyst contact time by changing the space velocity of the reactant stream,

expecting the primary products but not the secondary products to be observed in

the limit, as the contact time approaches zero. However, this method is not

straightforwardly suitable for the partial oxidation of methane, because when the

space velocity of the reactants is markedly changed, it is difficult to maintain the

catalyst temperature unchanged as a consequence of the large differences in heats

of reaction at different space velocities. Therefore, pulse methods were used to

determine the dependences of the CO and CO2 selectivities on the residence time,

by changing the carrier gas flow rate. Because the amount of reactant in a pulse

was small, no significant differences in the catalyst temperature resulted from

injection of the pulses. Pulse reaction experiments with the Ni/La2O3 catalyst

(134) showed that when the space velocity of the carrier gas was changed from

257,000 to 400,000 mL (g catalyst)21 h21, the selectivity to CO increased

gradually from 41 to 46% at 873 K, whereas that for CO2 decreased. This

comparison implies that CO formation is favored by short residence times,

consistent with the suggestion that CO is the primary product and CO2 a

secondary product, formed from CO. This observation supports the pyrolysis

mechanism by which CO is generated by the oxidation of C, formed via the CH4

dissociation on the Ni/La2O3 catalyst.

Furthermore, pulse transient response experiments (129) show that the initial

time at which CO was detected is shorter than that of CO2 by about 0.2 s for a

CH4/O2 (2/1) pulse at 773 K, in the presence of a reduced Ni/La2O3 catalyst.

This result indicates that either the CO generation occurs earlier, or the CO

desorption is faster than that of CO2. However, the response curve to a pulse of

pure CO and the response curve to a pulse of pure CO2 were similar when a

reduced Ni/La2O3 catalyst was used. Consequently, the delay of the CO2

generation relative to that of CO is not caused by desorption, but by the earlier

generation of CO. Therefore, the combination of experiments demonstrates that

CO is the primary product.

Shen et al. (142) used an isotopic transient technique and XPS to investigate

the partial oxidation of CH4 to synthesis gas on a Ni/Al2O3 catalyst at 973 K. The

results show that CH4 can decompose easily and quickly to give H2 and NixC on

the reduced catalyst, and that NixC can react rapidly with NiO, formed by the

oxidation of nickel by O2 to give CO or CO2, depending on the relative

concentration of NixC around NiO on the catalyst surface. The conclusion drawn

by the authors (142) was not only that H2 and CO are primary products in the

partial oxidation of CH4, but also that most of the CO2 is also the primary product

of the surface reaction between NixC and NiO. In contrast, the kinetics results of

Verykios et al. (143) indicated that the reaction on the Ni/La2O3 catalyst mainly

takes place via the sequence of total oxidation to CO2 and H2O, followed by


the reforming reactions to give synthesis gas, whereas CO formation by the direct

route was observed at very low oxygen partial pressures.

The partial oxidation of methane to synthesis gas on a Ru/TiO2 catalyst was

examined by combining non-steady-state and steady-state isotopic transient

experiments with in situ DRIFT spectroscopy (144). The authors showed that the

primary product of the reaction was CO, which resulted from the surface reaction

between carbon and adsorbed atomic oxygen on metallic Ru sites, with CO2

being formed by the oxidation of CO on the oxidized sites. The unique ability of

the Ru/TiO2 to catalyze the direct formation of H2 and CO was attributed to its

high resistance to oxidation under the conditions of partial oxidation of methane.

However, Weng et al. (145) proposed that on Ru/SiO2 the dominant pathway

to synthesis gas is via the sequence of total oxidation of CH4 followed by

the reforming of the unconverted CH4 by CO2 and H2O; the prevalence of this

pathway can be attributed to the high oxygen affinity of ruthenium.

It was reported (146) that the pathway of partial oxidation of methane on a

rhodium-containing catalyst depends strongly on the support material. On the

basis of a pulsed reaction and temperature-jump measurements, Nakagawa et al.

(146) proposed that, on Rh/TiO2 and Rh/Al2O3 catalysts, the endothermic

decomposition of CH4 to H2 and deposited carbon or CHx first takes place at the

upstream end of the catalyst bed, followed by the oxidation of the deposited

carbon or the CHx species to COx. However, on Rh/SiO2, synthesis gas was

produced by a two-step pathway consisting of a highly exothermic complete

oxidation of methane to H2O and CO2, followed by the endothermic reforming of

methane by H2O and CO2. In contrast, in situ time-resolved IR spectroscopy

showed that on the Rh/SiO2 catalyst, the synthesis gas was formed principally by

the direct oxidation of CH4—hence CO was the primary product (145). CO and

H2 produced as primary products were also observed for the reaction catalyzed by

a rhodium sponge at temperatures from 873 to 1023 K (147).

It was suggested that the partial oxidation of methane on Ir/TiO2 (146), Pt/

Al2O3, Pt/ZrO2, and Pt/Ce–ZrO2 (98) takes place by a two-step pathway,

consisting of a highly exothermic complete oxidation of methane to H2O and

CO2 followed by the endothermic reforming of methane with H2O and CO2.

Thus, in summary, we infer that there is not a simple answer to the question

“which is the primary product, CO or CO2?”. The complexity might suggest that

the reaction pathway depends not only on the catalyst composition but also on the

reaction conditions.

II.D.3. CHx Species and Rate-Determining Steps

Although numerous authors (17,129,130,134–137) suggested that the dis-

sociation of methane constitutes the first step in the methane oxidation by the

pyrolysis mechanism, it is important to provide direct evidence for the formation


of the CHx species. Deuterium-methane pulses were used to obtain such evidence

(131). The experiments showed that only CH4 and CD4 free of CHxDy (x þ y ¼ 4;x . 0; y . 0) were present in the product obtained in the presence of the

unreduced nickel-containing catalyst, whereas besides CH4 and CD4, CHxDy was

detected in the product obtained with the reduced catalyst (3% CH4, 31% CH3D,

38% CH2D2, 20% CHD3, and 8% CD4) (131). Whereas 28% of the methane (CH4

and CD4) in the feed gas was converted to CO and CO2, a larger amount (65%) of

the methane participated in the isotopic exchange reaction on the reduced

catalyst. This result demonstrates that on the reduced catalyst the isotopic

exchange reaction was faster than the partial oxidation. In contrast, on the

unreduced catalyst, no isotopic exchange reaction occurred. Hence, CH, CH2,

and CH3 species are inferred to form on the reduced catalyst, demonstrating that

the reaction follows the pyrolysis mechanism. Furthermore, on the reduced

catalyst, the amount of methane involved in the exchange between CH4 and CD4

was found to be greater than that involved in the conversion to CO and CO2. This

observation demonstrates that the exchange between CH4 and CD4 was faster

than the conversion of methane to CO and CO2, and hence that the methane

cleavage cannot be the rate-determining step. It is inferred that, instead, the rate-

determining step is the reaction of the CHx species with oxygen. Considering the

oxidation of surface carbon as the rate-determining step, Hu and Ruckenstein

(129) obtained an activation energy of 30.5 kcal mol21 for the partial oxidation

from the concentration curves of the C and O species on the nickel-containing

catalyst against reaction time. This activation energy is consistent with the

theoretical value of 33 kcal mol21 obtained for the CðsÞ þ OðsÞ! COðsÞ reaction

on Ni(111) (149), and this result provides additional support that the oxidation of

the surface carbon species is rate-determining.

Deuterium effects were also used to identify the rate-determining step for the

partial oxidation (137,150–152). By replacing the reactants CH4 þ O2 with

CD4 þ O2, Tang et al. (150) examined the deuterium isotope effect in the partial

oxidation of methane on Pt/a-Al2O3 in the temperature range of 823–923 K by

using pulses of reactant and analysis of the product by mass spectrometry. No

deuterium isotope effect was observed for CH4 conversion, whereas the CO

formation exhibited a normal deuterium isotope effect, indicating that the surface

reaction between the adsorbed hydrocarbon species and adsorbed oxygen species

to give CO may be a relatively slow step. In contrast, on the rhodium-containing

catalysts at 973 K, normal deuterium isotope effects were observed for methane

conversion and CO yield, but no effect on the CO selectivity was detected (137).

For the partial oxidation of methane to synthesis gas catalyzed by Ru/TiO2 at

903 K, Elmasides and Verykios (144) also found deuterium isotope effects on the

CH4 consumption rate ðRH=RD ¼ 1:6Þ and CO formation rate ðRH=RD ¼ 1:9Þ; but

no effect on CO2 formation. Therefore, they suggested that the processes of CH4

consumption and CO production are affected by the C–H cleavage of CH4 and,


consequently, are slow or rate-determining steps. However, it is worthwhile to

note that, to identify the rate-determining steps, one must ensure that the

deuterium isotope effects are obtained under non-equilibrium reaction con-

ditions. This restriction applies because the equilibrium partial oxidation of

methane also involves deuterium isotope effects on CH4 conversion and CO

formation (153), which do not depend on the rate-determining steps.

II.D.4. Comparison of Reactions on Reduced and Unreduced Catalysts

Results of isotopic pulse experiments showed that different mechanisms occurred

on the reduced and unreduced nickel catalysts (130). Hu and Ruckenstein (130)

demonstrated that the reaction on the unreduced catalyst involved gas-phase or

weakly adsorbed CH4 and strongly adsorbed or lattice oxygen (an Eley–Rideal

mechanism). Furthermore, these authors found that the methane conversion on

the unreduced catalyst took place predominantly (18%) by its reaction with the

lattice oxygen of the catalyst and also by its reaction with the oxygen from the gas

feed stream (12%). This comparison implies that the reaction with the lattice

oxygen was more facile than that with the gas-phase oxygen. According to the

above results, it is reasonable to suggest that CH4 is oxidized mainly by the

oxygen of the lattice, which is replenished by the oxygen from the gas phase. This

mechanism can be expressed as follows:

CH4ðgÞ þ 4NiO ! CO2ðgÞ þ 2H2OðgÞ þ 4Ni ð4Þ

2Ni þ O2ðgÞ! 2NiO ð5Þ

Using a temperature-programmed surface reaction (TPSR) technique, Li et al.

(154) showed that this complete oxidation of methane took place on the NiO

catalyst during the CH4/O2 reaction. Weng et al. (145) used in situ microprobe

Raman and in situ time-resolved IR spectroscopies to obtain a relationship

between the state of the catalyst and the reaction mechanism. These authors

showed that RuO2 in the Ru/SiO2 catalyst formed easily at 873 K in the presence

of a CH4/O2/Ar (2/1/45, molar) mixture and that the dominant pathway to

synthesis gas was by the sequence of total oxidation of CH4 followed by

reforming of the unconverted CH4 by CO2 and H2O. Thus, these results indicate

that the oxidation of methane takes place principally by the combustion

mechanism on the oxidized form of this catalyst.

Hu and Ruckenstein’s results (130) showed that on the reduced nickel-

containing catalyst, the reaction took place by a Langmuir–Hinshelwood

mechanism involving adsorbed CH4 and oxygen species. Furthermore, they

indicated that a slow dynamic redox process consisting of lattice oxygen

formation and its reduction by carbon species was at least partly responsible for

the CO formation.


In summary, one can conclude that even for one catalyst, the oxidation of

methane follows different reaction pathways on the reduced and oxidized forms.

Because the state of the metal in the catalyst depends on reaction conditions and

the reactant and catalyst compositions (155), the oxidation of methane might

follow different mechanisms under different conditions even for the same type of

catalyst (156). Wang and Ruckenstein (152) indeed found that the mechanisms of

the partial oxidation of methane to synthesis gas on rhodium-containing catalysts

were dependent on both the metal loading and temperature. At low metal

loadings (e.g., 0.05 wt%), the combustion-reforming mechanism was responsible

for the reaction, whereas at high loadings (e.g., 1.0 wt%) a combination of the

combustion-reforming and pyrolysis–oxidation mechanisms predominated at

low temperatures (#773 K), and the pyrolysis–oxidation mechanism became

dominant at high temperatures ($923 K). Froment et al. (157) also found that, on

Rh/Al2O3, the oxidation products under oxidative conditions were CO2 and H2O,

whereas the selectivities towards CO and H2 rose to almost 100% as the

conditions became more reductive.

III. CO2 Reforming of Methane

Deactivation of supported metal catalysts by carbon or coke formation, which has

its origin in the CH4 dissociation and/or CO disproportionation, is the most serious

problem hindering the application of the CO2 reforming of methane. Attempts to

overcome this limitation have focused on the development of improved catalysts.

III.A. Carbon Formation on Metal Surfaces

In the CO2 reforming of methane, carbon formation can occur via two possible

pathways: CH4 decomposition and CO disproportionation (the Boudouard

reaction). Carbon formation by CH4 decomposition is a structure-sensitive

reaction (158,159). Specifically, the Ni(100) and Ni(110) surfaces are more

active in the decomposition of CH4 to carbon than the Ni(111) surface (158).

The CO disproportionation,

2CO ¼ C þ CO2; DH0298 ¼ 2172 kJ mol21 ð6Þ

is an exothermic reaction favored at temperatures below 973 K. Measurable rates

of carbon deposition occur in the presence of cobalt, iron, and nickel catalysts at

temperatures above 623 K (159). The form of carbon on metal surfaces generated

during this reaction depends on the reaction conditions; amorphous and

filamentous carbons predominate in the lower temperature range of 623–873 K


(160–163), and a graphitic structure predominates at 973 K or higher

temperatures (160,164–166). The diffusion and segregation of carbon are also

dependent on the metal surface structure. For example, the carbon on Ni(110) can

diffuse more readily into the bulk than that on Ni(100) (159). Furthermore, the

carbon adsorbed on the smaller metal particles diffuses with more difficulty than

that on the larger particles (167). The structure-sensitivity of carbon formation

provides the possibility for inhibition of the carbon deposition by modification of

the catalyst surface structure.

III.B. Critical Issues Related to Carbon Deposition

Thermodynamic considerations (5,168,169) suggest operation at high CO2/CH4

ratios (.1) and high temperatures to minimize carbon formation in the CO2

reforming of methane. However, from an industrial viewpoint, it is desirable to

operate at lower temperatures and with a CO2/CH4 (or H2O/CH4) ratio near unity.

Such an operation requires a catalyst that kinetically inhibits the carbon

formation under conditions that are thermodynamically favorable for carbon

deposition. The noble metals and nickel were found to be highly active catalysts

(170). Although the noble metals are characterized by much less carbon

deposition than others (35), their high cost makes them unsuitable for large-scale

applications. In terms of cost, nickel appears to be the most suitable catalyst.

However, thermodynamic investigations indicated that the nickel-containing

catalysts are prone to carbon deposition in CO2 reforming, resulting in catalyst

deactivation (5). The inhibition of carbon deposition on the catalyst constitutes

the greatest challenge in CO2 reforming.

Two main properties of a catalyst affect the carbon deposition: surface

structure and surface acidity (171,172). Evidence that the structure has a strong

influence on carbon formation is provided by data showing that carbon formation

is more difficult on Ni(111) than on Ni(100) or Ni(110) (159). One method of

inhibiting carbon deposition is to control the size of the ensembles of metal atoms

on the surface, because the ensembles necessary for carbon formation are larger

than those needed for CH4 reforming (173). Thus, by controlling the nickel

particle size, one can control the carbon deposition. For example, strong

adsorption of sulfur can be used to influence ensemble size, and the suppression

of carbon deposition on nickel catalysts by sulfur passivation was commercia-

lized in the SPARG process (174,175). Sulfur passivation is attributed to

the control of the size of the active metal ensembles because sulfur preferentially

eliminates the larger ensembles. It has also been noted that carbon deposition can

be attenuated or even suppressed when the metal is supported on a metal oxide

with a strong Lewis basicity (176–179). This suppression occurs because the

high Lewis basicity of the support increases the ability of the catalyst to


chemisorb CO2 in the CO2 reforming of methane and H2O in the steam reforming

of methane, and these species react with carbon to form CO, resulting in

decreased net carbon formation.

III.C. Supported Noble Metal Catalysts

Inui (180) and Rostrup-Nielsen et al. (175) reported that the amount of carbon

deposited on metal catalysts decreases in the order Ni q Rh . Ir ¼ Ru . Pt ø Pd

at 773 K and Ni . Pd ¼ Rh . Ir . Pt q Ru at 923 K. Thus, the noble metals

exhibit higher selectivities for a carbon-free operation than nickel. Nevertheless,

carbon deposition does also occur on noble metals.

The above sequence also depends on the nature of the support (35,175,

181–184). ZrO2 has been widely used as support for platinum because of

the lower rate of carbon formation than with other supports (185–190). Bitter

et al. (188) observed that the rate of carbon formation decreased in the sequence

Pt/Al2O3 q Pt/TiO2 . Pt/ZrO2. Furthermore, the authors found that carbon

formation (most likely from methane) rather than sintering is the main cause of

the deactivation of the platinum-containing catalyst (Fig. 9). The high stability of

the zirconia-containing catalysts is probably associated with the strong Pt–Zrnþ

interactions, which reduce the carbon formation during reaction by promoting the

CO2 dissociation (189). It was suggested that the catalytic activity is determined

Fig. 9. CO2 conversions in the CO2 reforming of CH4 catalyzed by Pt/ZrO2 (V), Pt/TiO2 (B), and

Pt/g-Al2O3 (O). Each catalyst contained 0.5 wt% Pt. Before reaction, the catalyst was reduced in

flowing H2 at 1125 K for 1 h. Reaction conditions: temperature, 875 K; feed gas molar ratios,

CO2=CH4=Ar=N2 ¼ 4:2=4:2=7:5=1:0; GHSV, 32,000 mL (g catalyst)21 h21 (188).


by the available Pt–ZrO2 perimeter (186). On Pt/ZrO2, methane is decomposed

on the metal to give CHx (the average value of x is 2) and H2. The principal

pathway to CO2 reduction occurs by the initial formation of a carbonate close to

the metal-support boundary. The carbon on the metal reduces the carbonate to

formate, which decomposes rapidly to CO and surface hydroxyl groups. The

hydroxyl groups recombine to form water or react further with methane to

generate CO and hydrogen (steam reforming). When the rate of methane

decomposition and carbonate reduction are in balance, the catalytic activity

remains stable.

In contrast, the activity of supported rhodium catalysts is determined

principally by the concentration of accessible surface Rh atoms, which catalyze

methane decomposition, followed by CO2 reduction (186). As a result, the

support plays a minimal role in the rhodium-containing catalysts.

The promoters also have a significant effect on carbon deposition. It was found

that the bimetallic Pt–Au/SiO2, Pt–Sn/SiO2, and Pt–Sn/ZrO2 catalysts exhibited

less carbon deposition during CO2 reforming of CH4 than the respective

monometallic platinum catalysts (191), probably because of the formation of

alloys. Vanadium oxide also plays a promoting role in the Rh/SiO2 catalyst at

temperatures of 723–773 K (192). Vanadium oxide enhances the catalytic

activity of Rh/SiO2 and decreases the carbon deposition. This benefit was

attributed to the formation of a partial VOx overlayer on the rhodium surface,

which decreases the sizes of the accessible ensembles of Rh atoms, making some

of them too small for coke formation; new sites at the Rh–VOx interface that are

considered to activate CO2 dissociation were also created. The addition of cerium

or lanthanum resulted in a significant improvement in the stability of Pt/ZrO2, with

no decrease in either CH4 or CO2 conversion (193). Temperature-programmed

oxidation (TPO) data showed that although the total amount of carbon deposited

on the Ce-promoted Pt/ZrO2 catalyst was not less than that on the unpromoted

catalyst, these deposits were eliminated at much lower temperatures, indicating

the ability of the catalyst to self-clean its active sites. The La-promoted catalyst

also exhibited a much lower carbon deposition than the unpromoted catalyst.

III.D. Non-Noble Metal Supported Catalysts

In CO2 reforming, most of the reported research has been focused on non-noble

metal catalysts, particularly nickel, because nickel has activity and selectivity

comparable to those of noble metals, at much less cost. However, thermodynamic

investigations indicated that the nickel-containing catalysts are prone to carbon

deposition in CO2 reforming, resulting in catalyst deactivation (5). Therefore, an

important challenge is to increase the resistance of nickel-containing catalysts to

deactivation by carbon deposition.


III.D.1. Ni/Al2O3 Catalysts

Alumina is one of the most commonly used supports for nickel catalysts (111,

178,194–204). Ni/Al2O3 exhibits carbon deposition (180) that depends on the

catalyst structure, composition, and preparation conditions.

Chen and Ren (205) observed that the carbon deposition was markedly

suppressed if NiAl2O4 was formed during pretreatment. This suppression might

be the result of a strengthening of the Ni–O bond in NiAl2O4 when compared to

that in the NiO crystal (206). The stronger Ni–O bond increases the difficulty of

reduction of Ni2þ to Ni0, resulting in smaller nickel crystallites on the catalyst

surface. These nickel crystallites, which are smaller than the size necessary for

carbon deposition, decrease the carbon formation (195). Kim et al. (194,207)

noted that, in comparison with the alumina-supported nickel catalyst prepared by

the conventional impregnation method, Ni/Al2O3 catalysts prepared from aerogel

alumina exhibited remarkably low coking rates, which the authors associated

with the high dispersion of the metal particles. A similar observation was made

by Osaki et al. (208). The authors suggested that the Ni–O–Al bonds formed in

aerogels, which resulted in fine nickel particles after H2 reduction, contributed to

both the high activity and low carbon deposition (208).

A water-in-oil (w/o) microemulsion method was also effective in the

preparation of a Ni/Al2O3 catalyst with good stability and low carbon deposition

(209). Hayashi et al. (209) demonstrated that, although their conventionally

impregnated catalyst deactivated with time-on-stream as a result of severe

coking, the catalyst prepared by a w/o microemulsion method maintained its

activity for 50 h, generating little coke for a CO2/CH4 molar ratio . 1.4.

Furthermore, it was found that, at 1088 K and 21 atm pressure, a fresh nickel- and

magnesium-containing hydrotalcite clay-derived catalyst provided the same

performance as the commercial Ni/Al2O3 or Ni/MgAl2O4 catalysts, whereas

under more severe operating conditions, the clay-derived catalysts exhibited

superior activity and stability (210).

A dependence of the amount of carbon deposition on the nickel loading was

observed for Ni/Al2O3 catalysts (197). For example, a 1-wt% Ni/Al2O3 exhibited

much less carbon deposition than a 13.6 wt% Ni/Al2O3 catalyst (197).

Many promoters have been used to improve the performance of Ni/Al2O3

catalysts. The effect of the basic oxides of Na, K, Mg, and Ca on Ni/Al2O3 was

examined by a number of authors (178,203,211–213). They found that these added

oxides markedly decrease the carbon deposition. The kinetics results showed that

the added metal oxides changed the reaction order in CH4 from negative to positive

and that in CO2 from positive to negative. This observation implies that the surface

of a nickel catalyst incorporating basic metal oxides is abundant in adsorbed CO2,

whereas the surfaces devoid of these oxides are abundant in adsorbed CH4 (178).

The coverage of nickel with CO2 is most likely unfavorable to CH4 decomposition


and, as a result, the carbon deposition is decreased. Wang and Lu (214) also

observed that Na2O or MgO promoters decreased the carbon deposition on Ni/

Al2O3 catalysts (Fig. 10). However, these promoted Ni/Al2O3 catalysts were

characterized by lower activities and significant deactivation. Hence, it is inferred

that the deactivation of the NaO- or MgO-promoted Ni/Al2O3 catalysts was not

principally caused by carbon deposition.

Choi et al. (215) examined the effect of Co, Cu, Zr, and Mn as promoters of Ni/

Al2O3 catalysts. They found that, in comparison with the unmodified Ni/Al2O3

catalysts, those modified with Co, Cu, and Zr exhibited slightly improved

activities, whereas other promoters reduced the activity. The Mn-promoted

catalyst provided a remarkable reduction in coke deposition with only a small

reduction in catalytic activity. Furthermore, Seok et al. (216) noted that the

manganese addition to Ni/Al2O3 led to a partial coverage of the surface of nickel

by patches of MnOx, which promoted the adsorption of CO2. Both the partial

coverage of the nickel surface with MnOx and the promoted CO2 adsorption

appear to be responsible for the decreased carbon deposition on Ni/MnO–Al2O3

catalysts. Mo can also improve the stability of Ni/Al2O3 by reducing the carbon

Fig. 10. Carbon deposition on nickel-containing catalysts at 973 K as determined by TGA. Before

reaction, the catalysts were reduced at 1073 K for 3 h. Reaction conditions: temperature, 973 K; feed

gas molar ratio, CO2=CH4 ¼ 1=1; GHSV, 144,000 mL (g catalyst)21 h21 (214).


deposition (217). Noble metal (Ru or Pd) addition to supported nickel catalysts

resulted in a marked improvement in both activity and stability (218).

Rare earth metals have also been used to promote Ni/Al2O3 catalysts. Slagtern

et al. (219) tested Ni/Ln/Al2O3 (Ln ¼ rare earth mixture) catalysts containing

0.15 wt% Ni for their lifetimes (60–600 h) in a fluidized-bed reactor at 1073 K

and 1 atm. The catalyst with a rare earth content of 1.7 wt% Ln was more active

and stable than the unpromoted catalyst, and more active than a catalyst

containing 8.5 wt% Ln. Furthermore, it was found that nickel sintering was

initially the major cause of deactivation, with coking becoming increasingly

important at longer times on stream (.60 h). The catalyst with 1.7% Ln had a

higher initial nickel dispersion than the catalyst devoid of Ln. However, the

higher activity of the promoted catalyst than of the unpromoted catalyst could not

be fully explained by this difference. Neodymium also promotes Ni/Al2O3

catalysts, by reducing the carbon deposition (220). CeO2 was also found to have

an effect on the Ni/Al2O3 catalyst (221,222). Although CeO2 is not a suitable

support for nickel because of the strong metal–support interaction, which

reduces the catalytic activity, it can have a positive effect on the catalytic activity,

stability, and suppression of carbon deposition when used as a promoter of Ni/

Al2O3 catalysts (221,222). A loading of 1–5 wt% CeO2 was found to be the

optimum. The use of CeO2 as a promoter for the nickel catalysts decreases the

strength of the interactions between the nickel oxide and support, resulting in an

increase in the reducibility of the nickel oxide and a higher nickel dispersion. The

stability and reduced coking characteristic of CeO2-promoted catalysts can be

attributed to the redox properties of CeO2, which can react directly with carbon-

containing species to generate CO and CeOx, followed by the reoxidation of

CeOx by CO2 back to CeO2 (221).

III.D.2. Ni/SiO2 Catalysts

The deactivation of Ni/SiO2 catalysts during the CO2 reforming of methane was

examined as a function of various operating parameters (223). The two principal

causes of catalyst deactivation, nickel sintering and carbon deposition, were

shown to depend strongly on the pretreatment conditions. Kroll et al. (224) noted

that for the Ni/SiO2 catalyst, nickel carbide-like layers, formed during the very

initial period of the run, provided the active phase for CO2 reforming. However,

when the carbon formation, which takes place at equilibrium with gaseous CH4,

became faster than the oxidation of the carbon with the oxygen adspecies formed

by carbon dioxide activation, carbon deposition occurred. The carbon deposition

depended strongly on the nickel loading (197). It was found that a 13.6 wt% Ni/

SiO2 catalyst exhibited a greater carbon deposition than a 1 wt% Ni/SiO2. A

physical mixture of SiO2 and nickel minimized the amount of deposited carbon

(225), and a physical mixture of Al2O3 and nickel generated a greater amount of


carbon deposition (197). This comparison indicates that the Al2O3 surface

promotes carbon deposition. CaO also affects Ni/SiO2 catalysts by decreasing the

dispersion of the nickel phase (226).

III.D.3. Ni/La2O3 Catalysts

Zhang and Verykios (227) reported a Ni/La2O3 catalyst which exhibited a higher

activity and higher long-term stability for CO2 reforming of methane to synthesis

gas than Ni/Al2O3 and Ni/CaO catalysts. As shown in Fig. 11, although the initial

rate of reaction on Ni/g-Al2O3 was higher than that on Ni/CaO, probably as a

consequence of the higher dispersion of nickel in the former catalyst, the

deactivation rate of Ni/g-Al2O3 was higher than that of Ni/CaO. In contrast, the

rate of reaction on a Ni/La2O3 catalyst increased significantly with time on stream

during the initial 2–5 h of reaction, and then tended to remain unchanged with time

Fig. 11. CO formation rates determined from reactant conversions and product selectivities in

a fixed-bed flow reactor for CO2 reforming of CH4. The catalysts were nickel supported on

La2O3, g-Al2O3, or CaO. Each catalyst contained 17 wt% Ni. Before reaction, the catalyst was

reduced in flowing H2 at 773 K for at least 5 h and then at 1023 K for 2 h. Reaction conditions:

pressure, 1.0 atm; temperature, 1023 K; feed gas molar ratio, CH4=CO2=He ¼ 2=2=6; GHSV,

1,800,000 mL (g catalyst)21 h21 (227).


on stream for 100 operating hours. In these experiments, low conversions of CH4

and CO2 were observed at a very high space velocity (227,228). However, when

higher CH4 and CO2 conversions (about 75 and 80%, respectively) were obtained

by reducing the space velocity, the Ni/La2O3 catalyst exhibited deactivation

(Fig. 12) (228). Other researchers also observed the deactivation of Ni/La2O3

catalysts at high CH4 and CO2 conversions (229). The higher stability of the catalyst

at low reactant conversions might have occurred because high concentrations of

unreacted CO2 inhibited carbon deposition by the reaction CO2 þ C ¼ 2CO.

Ruckenstein and Hu (230) investigated the role of the anions NO32 or Cl2

(used in the catalyst preparation by impregnation of the unreduced Ni/La2O3) in

carbon deposition on Ni/La2O3 catalysts. The unreduced Ni/La2O3 catalyst,

prepared from nickel nitrate, was characterized by a high initial CO yield but a

low stability; in contrast, the unreduced Ni/La2O3 catalyst, prepared with

chloride, had a high stability. This stabilization probably occurred because a

stable lanthanum chloride inhibited the formation of large ensembles of nickel

atoms, which are necessary for carbon deposition.

The preparation method also affects the Ni/La2O3 catalysts (231). The

conversions of CH4 and CO2 in the CO2 reforming of CH4 catalyzed by Ni/La2O3

Fig. 12. Conversions of CH4 and CO2 and selectivities for formation of CO and H2 as a function of

time on stream for CO2 reforming of CH4 catalyzed by 17 wt% Ni/La2O3. Before reaction, the catalyst

was reduced in flowing H2 at 773 K for at least 5 h and then at 1023 K for 2 h. Reaction conditions:

pressure, 1 atm; temperature, 1023 K; feed gas molar ratio, CH4=CO2 ¼ 1=1;GHSV is unknown (228).


prepared by a sol–gel technique were significantly higher than those catalyzed by

Ni/La2O3 prepared by wet impregnation. TG/DTG experiments confirmed that

the amount of carbon deposited in the former case was smaller than in the latter

case. It is inferred that the difference can be attributed to the uniform dispersion

of nanoscale nickel particles in the sol–gel-generated Ni/La2O3 catalyst.

III.D.4. Ni/ZrO2 Catalysts

The suitability of zirconia-supported nickel catalysts for the CO2 reforming

reaction was investigated with emphasis on the stability of the catalysts under

conditions favorable for carbon formation (232). It was found that at temperatures

between 993 and 1053 K, the ZrO2-supported catalysts with lower nickel loadings

(,2 wt%) were more stable than those with higher nickel loadings for a

stoichiometric CO2/CH4 ratio. Furthermore, two forms of deposited carbon were

observed in the less stable catalysts, and only one form was observed in the more

stable ones. Carbon deposits were formed on the reduced catalyst at a very high

rate during the TPSR (233). The amount of deposited carbon remained constant

on the catalyst during reaction at 973 K (233), consistent with the inference that

the initially formed carbon acted as a reaction intermediate that transformed CO2

into CO. Even with catalysts having high nickel loadings, catalyst lives

without significant deactivation were achieved for 30 h at 1023 K and for 20 h

at 1123 K (234).

Li et al. (235) found that promoters can affect the Ni/ZrO2 catalysts. Among

the Ni/ZrO2 catalysts promoted with oxides of lanthanum, cerium, or

manganese, Ni/La–ZrO2 exhibited the highest activity, whereas Ni/Ce–ZrO2

and Ni/Mn–ZrO2 were characterized by low carbon depositions during reaction.

Furthermore, Ni–Mg/ZrO2 exhibited the highest activity and stability. It was

inferred that the promotion by magnesium can be attributed to increasing

dispersion of nickel and to an enhancement in the interaction between CO2 and

the catalyst.

Lercher et al. (236) reported that Ni/ZrO2 catalysts with small sizes of metal

particles (2–3 nm) exhibited high stability. The small particles prevented the

formation of carbon filaments. The stabilities of Ni/ZrO2 catalysts were also

dependent on the preparation method. Wei et al. (237) reported that the Ni/ZrO2

catalyst prepared from large Zr(OH)4 particles deactivated rapidly. In contrast, a

catalyst with a high metal loading of nickel (27 wt%), obtained by impregnating

ultra-fine Zr(OH)4 particles (6 nm) with nickel nitrate, exhibited a high and stable

activity for CO2 reforming without deactivation by carbon deposition.

The activity of this catalyst for CO2 reforming of CH4 at 1030 K, with a CH4/

CO2 ¼ 1:1 molar feed rate of 24,000 mL (g catalyst)21 h21 did not deactivate

for 600 h, but exhibited oscillations in the CH4 conversion between 80 and

85% (Fig. 13). Comparing their best Ni/ZrO2 catalyst with the NiO/MgO


solid-solution catalyst of Fujimoto et al. (238), Wei et al. concluded that their

Ni/ZrO2 catalyst exhibited higher activity. But the activity of their best Ni/ZrO2

catalyst is much lower than that of Ruckenstein and Hu’s NiO/MgO solid-

solution catalysts (239) (see Section III.E for more details).

III.D.5. Other Catalysts

Carbon deposition is much greater on Co/Al2O3 catalysts than on Ni/Al2O3 (240).

The presence of MgO markedly decreased the carbon deposition on the surface of

the cobalt catalyst (241). The role of MgO may be attributed to the formation of

strongly adsorbed CO2 species, which can easily react with the deposited carbon,

thus preventing catalyst deactivation (241).

Osaki et al. (242) compared the catalytic performance of MoS2 and WS2 with

that of Ni/SiO2. The CO2 reforming of methane on MoS2 or WS2 catalysts was

characterized by much lower reaction rates than that on the nickel catalyst,

although the sulfides prevented carbon deposition during the reforming reaction.

Completely different rate equations were obtained for the metal disulfide and

nickel catalysts. The positive reaction order in CH4 partial pressure and the

negative order in CO2 partial pressure characteristic of the sulfide catalyst are

Fig. 13. CH4 conversion in the CO2 reforming of CH4 catalyzed by Ni/ZrO2. Before reaction, the

catalyst was reduced in flowing H2/N2 (1/9, molar ratio) at 973 K for 3 h. Reaction conditions:

pressure, 1 atm; temperature, 1030 K; feed gas molar ratio, CH4=CO2 ¼ 1=1; GHSV,

24,000 mL (g catalyst)21 h21 (237).


contrasted with the negative order in CH4 partial pressure and positive order in

CO2 partial pressure characteristic of the nickel catalyst. These observations

suggest that the surface of the sulfide catalyst was abundant in adsorbed CO2,

whereas the surface of the nickel catalyst was abundant in adsorbed CH4. The

coverage with CO2 can be considered to be the principal cause of the suppression

of carbon deposition on the sulfide catalysts.

The Haldor Topsøe firm developed the SPARG process for the CO2/CH4

reforming, in which the conventional nickel-containing steam reforming catalyst

was modified to reduce its coke-forming propensity, by the continuous addition of

small amounts of sulfur to the feed gas during operation (174,175,243). However,

the passivation process led to a lower catalytic activity and required high operating

temperatures as a consequence of the sulfur poisoning of the active sites.

To develop effective catalysts for the CO2 reforming of methane, other supports

were also used for nickel catalysts, including perovskite (244), Y zeolite (245,246),

5A zeolite (247), high-silica ZSM-5 zeolite (248), and AlPO4 (tridymite) (249).

In summary, the development of non-noble metal catalysts has been focused

on nickel-containing catalysts (because nickel has an activity and a selectivity

comparable to those of noble metals at much less cost); on finding effective

promoters, selecting suitable supports; and on improving preparation methods.

Although some nickel-containing catalysts appear to be effective in short-term

tests for CO2 reforming, their long-term stability and tolerance for impurities,

which are important in industrial applications, are not yet clear.

III.E. MgO-Containing Solid-Solution Catalysts

III.E.1. Characteristics of MgO-Containing Solid-Solution Catalysts

MgO is a basic metal oxide and has the same crystal structure as NiO. As a result,

the combination of MgO and NiO results in a solid-solution catalyst with a basic

surface (171,172), and both characteristics are helpful in inhibiting carbon

deposition (171,172,239). The basic surface increases CO2 adsorption, which

reduces or inhibits carbon-deposition (Section III.B). The NiO–MgO solid

solution can control the nickel particle sizes in the catalyst. This control occurs

because in the solid solution NiO has strong interactions with MgO and, as

indicated by TPR data (26), the former oxide can no longer be easily reduced.

Consequently, only a small amount of NiO is expected to be reduced, and thus

small nickel particles are formed on the surface of the solid solution, smaller than

the size necessary for coke formation. Indeed, the nickel particles on a reduced

16.7 wt% NiO/MgO solid-solution catalyst were too small to be observed by

TEM (171). Furthermore, two additional important qualities stimulated the

selection of MgO as a support: its high thermal stability and low cost (250,251).


Like NiO, CoO and FeO are characterized by the same crystal structure as

MgO and have comparable lattice parameters, and, hence, can form CoO/MgO

and FeO/MgO solid solutions. Therefore, it was expected that CoO/MgO and

FeO/MgO would inhibit carbon deposition and metal sintering, just as Ni/MgO

does, resulting in high stability (171).

III.E.2. Highly Effective MgO-Containing Solid-Solution Catalysts

In 1989, Gadalla and Sommer (252) reported that a solid-solution NiO/MgO

(1:1.35) catalyst prepared by precipitation can inhibit the carbon deposition in the

CO2 reforming of methane; however, they obtained a low CO2 conversion (66%),

a low H2 selectivity (79%), and a low CO selectivity (77%), even at the very low

WHSV of 3714 cm3 (g catalyst)21 h21 with a CH4/CO2 (1/1, molar) feed gas and

the high temperature of 1200 K. Their relatively high CH4 conversion was partly

a consequence of homogeneous gas-phase reactions that occurred under their

conditions. Indeed, the authors found extensive carbon deposits plugging the

reactor upstream and downstream of the reaction zone.

In 1992, Fujimoto et al. (176) reported results for the CO2 reforming of

methane catalyzed by NiO/MgO prepared by coprecipitation of the hydroxides

from aqueous solutions of nickel acetate and magnesium acetate with K2CO3 at

333 K; the coprecipitates were dried at 393 K for 12 h and calcined at 1223 K for

20 h. Although they did not mention that the NiO/MgO, which had the

composition Ni0.03Mg0.97O, was a solid solution, it was surely a solid solution

because it was calcined at a high temperature, as the authors (253) later reported.

Fujimoto et al. (176) observed that the NiO/MgO catalyst had a low stability,

suggested to be a consequence of carbon deposition. Although they added CaO

to the NiO/MgO to increase the stability, this addition decreased the activity

tremendously. Takayasu et al. (254,255) also noted a deactivation of the NiO/

MgO catalysts, caused by the formation of carbonaceous deposits. In 1994,

Swaan et al. (256) reported that a 3 wt% Ni/MgO catalyst had a low activity.

They suggested that the stabilization of Ni2þ ions in the MgO matrix was

responsible for the limited reducibility of the nickel observed experimentally and

for the formation of an active phase for the reforming reaction. Because of

these results, NiO–MgO solid-solution catalysts did not attract much interest at

that time.

In 1995, Ruckenstein and Hu reported a highly efficient 16.7 wt% NiO/MgO

solid-solution catalyst for CO2 reforming of methane, which was prepared by

impregnation and was calcined at 1073 K for 1.5 h (239). It exhibited almost

100% conversion of CO2, .91% conversion of CH4, and .95% selectivities

to CO and H2 at 1063 K, atmospheric pressure, and the very high space velocity

of 60,000 mL (g catalyst)21 h21 for a CH4/CO2 molar ratio of 1 (Fig. 14) (239).


The conversions and selectivities remained unchanged during the entire reaction

time employed (120 h), indicating that the reduced NiO/MgO catalyst had a high

stability (Fig. 14).

In contrast to MgO, the other alkaline-earth oxides, such as CaO, SrO, and

BaO, were found to be poor supports for NiO, as they provided catalysts with low

activities, selectivities, or stabilities (Fig. 14) (239). Although the reduced

NiO/Al2O3 catalyst provided high initial conversions (CH4, 91%; CO2, 98%) and

selectivities (.95% for both CO and H2), it was characterized by the fastest

carbon deposition, which led to the complete plugging of the reactor after only

6 h of reaction (197). The reduced Ni/TiO2 catalyst gave relatively low initial

Fig. 14. CH4 conversion (a) and CO yield (b) in the CO2 reforming of CH4 catalyzed by reduced

16.7-wt% NiO/alkaline earth metal oxides. Before reaction, each catalyst was reduced in flowing H2 at

773 K for 14 h. Reaction conditions: pressure, 1 atm; temperature, 1063 K; feed gas molar ratio,

CH4=CO2 ¼ 1=1; GHSV, 60,000 mL (g catalyst)21 h21 (239).


conversions of CH4 and CO2 (41 and 67%, respectively), which decreased with

increasing reaction time (197). It seems reasonable to conclude that the excellent

catalytic performance of NiO/MgO should be attributed to the formation of a

solid solution (257).

The conversions and selectivities characteristic of NiO/MgO solid-solution

catalysts were found to be dependent on their composition, preparation

conditions, and even the properties of the MgO (257–259). Furthermore, the

authors found that high and stable CO yields (.95%) occurred with NiO/MgO

catalysts having NiO contents between 9.2 and 28.6 wt% (258). No activity was

observed, however, for a NiO content of 4.8 wt%. At the high NiO content of

50 wt%, the CO yield decreased from 91 to 53% after 40 h, and the catalyst

became black, because of carbon deposition, after about 50 h of reaction. In

contrast, the other NiO/MgO solid-solution catalysts maintained their initial

color, and no carbon deposition was detected by TEM even after 120 h of reaction

(171). It was, therefore, inferred that too small amounts of NiO in the NiO/MgO

catalysts provided too-small numbers of Ni sites, and too-high amounts provided

numerous nickel metal particles that could easily sinter, generating large particles

that facilitated carbon deposition. Furthermore, the MgO surface area, pore size

distribution, and lattice parameters were observed to affect significantly the

performance of NiO/MgO solid-solution catalysts (259). An unsuitable MgO can

lead to a low initial conversion and a long induction time (259).

In 1997, Fujimoto et al. (260–262) reported new results for the CO2 reforming

of methane catalyzed by the Ni0.03Mg0.97O solid solution used by them in 1992

(176) and by bimetallics containing in addition small amounts of platinum,

palladium, or rhodium (molar ratio of M/(Ni þ Mg) was varied between

0.7 £ 1024 and 3.2 £ 1024, where M ¼ Pt, Pd, or Rh) (260). The Ni0.03Mg0.97O

solid-solution catalyst provided a low CO yield (about 215 mmol (g catalyst)21

s21, i.e., 38%) at a space velocity of 44,800 mL (g catalyst)21 h21 even at

1123 K. However, the addition of a noble metal promoted both the activity and

the stability at 773 K. The optimum noble metal loading was obtained for M/

(Ni þ Mg) < 2.1 £ 1024 (molar ratio). Temperature-programmed hydrogen-

ation (TPH) of the carbonaceous species formed during the catalytic reaction

indicated that the resistance of the Ni0.03Mg0.97O solid-solution catalyst to carbon

deposition was retained by the bimetallic catalysts as well (260). Furthermore,

TEM and EDS observations provided evidence of the formation of Pt–Ni alloy

particles (260). Temperature-programmed decomposition (TPD) data obtained

with CH4 suggested that CH4 decomposition was the rate-determining step on

Ni0.03Mg0.97O and that the CH4 decomposition was accelerated by alloy

formation (260). The improved stability of the catalyst was attributed to the

increased catalyst reducibility caused by noble metal promotion.

Fujimoto et al. (253) also found that the water treatment of the Ni0.03Mg0.97O

solid-solution catalyst increased the catalytic activity and stability for CO2


reforming of CH4. This promoting effect was inferred to be the consequence of a

structural rearrangement of the solid solution by the formation of nickel and

magnesium hydroxides (253). Furthermore, Fujimoto et al. (261–263) reported

that even Ni0.03Mg0.97O has a fairly good stability in the CO2 reforming of

methane. The excellent anti-coking performance of the reduced NiO/MgO solid

solution catalyst can be attributed to the high dispersion of the reduced nickel

species, the basicity of the support surface, and the nickel-support interactions.

From the above results, one can conclude that different NiO/MgO solid-

solution catalysts can have very different catalytic performances. For example,

Fujimoto et al.’s Ni0.03Mg0.97O solid-solution catalyst exhibited relatively

low activities. To reach about 82% conversion of CH4 in the presence of

this Ni0.03Mg0.97O catalyst, the space velocity had to be reduced to

18,670 mL (g catalyst)21 h21 at 1123 K (Fig. 15) (238). In contrast, Ruckenstein

and Hu’s NiO/MgO catalysts have very high activities (.91% conversion of

CH4 and .95% selectivities of CO and H2 at the space velocity of

60,000 mL (g catalyst)21 h21 at 1063 K) (Fig. 14) (239). Hu and Ruckenstein

(239,257,259) noted that the properties of the MgO, such as its surface area, pore

size distribution, and crystal structure, have important effects on the NiO/MgO

solid-solution catalysts. They found that the MgO supplied by Aldrich, which has

Fig. 15. CH4 conversions in the CO2 reforming of CH4 in the presence of nickel-containing

catalysts. Before reaction, the catalyst was reduced in flowing H2 at 1123 K for 14 h. Reaction

conditions: pressure, 1 atm; temperature, 1123 K; feed gas molar ratio, CH4=CO2 ¼ 1=1; GHSV,

18,670 mL (g catalyst)21 h21 (238).


a surface area of about 50 m2 g21 with nano-pores (10 –100 nm) and

nano-crystals (about 20 nm) (171,172,257), was a suitable support material for

MgO-containing solid-solution catalysts with very high activity and selectivity as

well as high stability.

Recently, Ruckenstein and Wang (264–266) also successfully developed

excellent CoO/MgO solid-solution catalysts for CO2 reforming of methane.

They reported that Co/MgO exhibited a good catalytic performance with a CO

yield of 93% and a H2 yield of 90% at the high space velocity of

60,000 mL (g catalysts)21 h21 and 1163 K, which remained unchanged during

50 h of investigation (264). In contrast, Co/CaO, Co/SrO, and Co/BaO each

provided low CO yields, and Co/CaO also had a low stability. The results indicate

that the CoO/MgO catalysts are characterized by performances similar to those of

NiO/MgO.

In summary, the basicity and the strong NiO–MgO interactions in binary

NiO/MgO solid solution catalysts, which inhibit carbon deposition and catalyst

sintering, result in an excellent catalytic performance for CO2 reforming. The

characteristics of MgO play an important role in the performance of a highly

efficient NiO/MgO solid-solution catalyst. Moreover, the NiO/MgO catalyst

performance is sensitive to the NiO content: a too-small amount of NiO in the

solid solution leads to a low activity, and a too-high amount of NiO to a low

stability. CoO/MgO solid solutions have catalytic performances similar to those

of NiO/MgO solid solutions, but require higher reaction temperatures. So far, no

experimental information is available regarding the use of a FeO/MgO solid

solution for CH4 conversion to synthesis gas.

IV. Conclusions

Synthesis gas production from natural gas, the most important step in the gas-to-

liquid process, can account for at least 60% of the integrated cost of the total gas-

to-liquid plant. The catalytic partial oxidation of methane provides a fast process

for the synthesis gas production. However, several challenges still remain

regarding this process. Large temperature gradients in the reactors (hot spots),

which are the result of a combination of a high space velocity and an exothermic

reaction, could make the process hazardous and difficult to control in industrial-

scale operations. Current technical options to solve this problem include

fluidized-bed reactors, in which the temperature of the mixed catalyst is almost

uniform, and combined processes that eliminate hot spots by combining the

exothermic partial oxidation with the endothermic CO2 reforming (or steam

reforming). High O2 separation costs represent the greatest challenge facing the

partial oxidation process. The main focus of research aimed at overcoming this


limitation is on O2-permeable ceramic membrane reactor processes, in which air

can be used directly. The membrane processes, which obviate the O2 separation

plant, could reduce the process cost by 25–40%. However, the O2 permeation

and stability of such membranes still need improvement. The complexity of the

reaction mechanisms, which depend on the catalyst composition and degree of

reduction, as well as on reaction conditions, can cause great difficulties in the

process design and control.

CO2 reforming of methane is an attractive technology because it converts two

greenhouse gases into useful chemicals. The deactivation of the catalyst, caused

by carbon deposition, constitutes the greatest challenge in this process. Although

noble metal catalysts are less subject to carbon deposition, nickel-containing

catalysts have attracted the most research interest, in part because of the relatively

low cost of nickel. In the preceding 10 years, several types of nickel-containing

catalysts with high activities and stabilities have been reported. For example,

nickel-containing solid solution catalysts have very high activity, selectivity, and

stability; and they inhibit carbon deposition and catalyst sintering. Because CO2

reforming of methane is a strongly endothermic process, the development of new

methods to provide less expensive energy constitutes attractive goal for future

research related to CO2 reforming.

References

1. Byrne, J.P.J., Gohr, R.J., and Haslam, R.T., Ind. Engng Chem. 24, 1129 (1932).

2. Rostrup-Nielsen, J.R., Sehested, J., and Nørskov, J.K., Adv. Catal. 47, 65 (2002).

3. Prettre, M., Eichner, C., and Perrin, M., Trans. Faraday Soc. 43, 335 (1946).

4. Lang, J., Z. Phys. Chem. 2, 161 (1888).

5. Gadalla, A.M., and Bower, B., Chem. Engng Sci. 43, 3049 (1988).

6. Dyrssen, D., and Turner, D.R., in “Carbon Dioxide Chemistry: Environmental Issues”

(J. Paul and C.M. Pradier, Eds.), p. 317. Athenaeum Press, Cambridge, 1994.

7. Puskas, I., CHEMTECH 24, 43 (1995).

8. (a) Keller, G.E., and Bhasin, M.M., J. Catal. 73, 9 (1982); (b) Lee, J.S., and Oyama, S.T.,

Catal. Rev.—Sci. Engng 30, 249 (1988); (c) Amenomiya, Y., Birss, V.I., Goledzinowski,

M., Galuszka, J., and Sanger, A.R., Catal. Rev.—Sci. Engng 33, 163 (1990).

9. (a) Pitchai, R., and Klier, K., Catal. Rev.—Sci. Engng 28, 13 (1986); (b) Foster, N.R.,

Appl. Catal. A: Gen. 19, 1 (1985); (c) Hall, T.J., Hargreaves, J.S.J., Hutchings, G.J.,

Joyner, R.W., and Taylor, S.H., Fuel Process. Technol. 42, 151 (1995).

10. Kiennemann, A., Petrol. Technol. 29, 355 (1990).

11. Guczi, L., van Santen, R.A., and Sarma, K.V., Catal. Rev.—Sci. Engng 38, 249 (1996).

12. Pasquon, I., Plenary lecture at EUOPACAT-1, Montpellier, France, September 12–17,

1993.

13. Ashcroft, A.T., Cheetham, A.K., Foord, J.S., Green, M.L.H., Grey, C.P., Murrell, A.J., and

Vernon, P.D.F., Nature 344, 319 (1990).

14. Hickman, D.A., and Schmidt, L.D., J. Catal. 138, 267 (1992).

15. Hickman, D.A., and Schmidt, L.D., Science 259, 343 (1993).


16. Torniainen, P., Chu, X., and Schmidt, L.D., J. Catal. 146, 1 (1994).

17. (a) Choudhary, V.R., Rajput, A.M., and Prabhakar, B., Catal. Lett. 15, 363 (1992);

(b) Choudhary, V.R., Rajput, A.M., and Prabhakar, B., J. Catal. 139, 326 (1993).

18. Choudhary, V.R., Rajput, A.M., and Rane, V.H., J. Phys. Chem. 96, 8686 (1992).

19. Choudhary, V.R., Rajput, A.M., and Rane, V.H., Catal. Lett. 16, 269 (1992).

20. Choudhary, V.R., Sansare, S.D., and Mamman, A.S., Appl. Catal. A: Gen. 90, L1 (1992).

21. Chang, Y.F., and Heinemann, H., Catal. Lett. 21, 215 (1993).

22. Dissannayake, D., Rosynek, M.P., and Lunsford, J.H., J. Phys. Chem. 97, 3644 (1993).

23. Hu, Y.H., Au, C.T., and Wan, H.L., Chin. Sci. Bull. 40, 303 (1995).

24. Hu, Y.H., Au, C.T., and Wan, H.L., Kexue Tongbao 40, 430 (1995).

25. Hu, Y.H., and Ruckenstein, E., Ind. Engng Chem. Res. 37, 2333 (1998).

26. Hu, Y.H., and Ruckenstein, E., Ind. Engng Chem. Res. 38, 1742 (1999).

27. Zhang, X.L., Hayward, D.O., and Mingos, D.M.P., Catal. Lett. 83, 149 (2002).

28. Zhang, X.L., Mingos, D.M.P., and Hayward, D.O., Catal. Lett. 72, 147 (2001).

29. Basile, F., Fornasari, G., Trifiro, F., and Vaccari, A., Catal. Today 64, 21 (2001).

30. Wolf, D., Hohenberger, M., and Baerns, M., Ind. Engng Chem. Res. 36, 3345 (1997).

31. Olsbye, U., Tangstad, E., and Dahl, I.M., in “Natural gas conversion II.” p. 303, Elsevier,

Amsterdam, 1994.

32. Bharadwaj, S.S., and Schmidt, L.D., J. Catal. 146, 11 (1994).

33. Roberts, M.J., AIChE 1997 Spring National Meeting, Houston, TX, March 9–13, 1997.

34. Ioannides, T., and Verykios, X.E., Catal. Lett. 47, 183 (1997).

35. Ashcroft, A.T., Cheetham, A.K., Green, M.L.H., and Vernon, P.D.F., Nature 352, 225

(1991).

36. Inui, T., Saigo, K., Fujii, Y., and Fujioka, K., Catal. Today 26, 295 (1995).

37. Ruckenstein, E., and Hu, Y.H., Ind. Engng Chem. Res. 37, 1744 (1998).

38. O’Connor, A.M., and Ross, J.R.H., Catal. Today 46, 203 (1998).

39. Ruckenstein, E., and Wang, H.Y., Catal. Lett. 73, 99 (2001).

40. Mo, L.Y., Zheng, X.M., Huang, C.J., and Fei, J.H., Catal. Lett. 80, 165 (2002).

41. Choudhary, V.R., Rajput, A.M., and Prabhakar, B., Catal. Lett. 32, 391 (1995).

42. Choudhary, V.R., Rajput, A.M., and Prabhakar, B., Angew. Chem. Int. Ed. 33, 2104 (1994).

43. Matsumura, Y., and Moffat, J.B., Catal. Lett. 24, 59 (1994).

44. Goetsch, D.A., and Say, G.R., US Patent 4877550 (1989).

45. Eberly, P.E., Goetsch, D.A., Say, G.R., Vargas, J.M., and Argas, J.M., US Patent 4888131

(1989).

46. Fiato, R.A., Long, D.C., Say, G.R., Taylor, J.H., Say, G., and Long, D., US Patent

5143647 (1992).

47. Choudhary, V.R., and Mamman, A.S., Appl. Energy 66, 161 (2000).

48. Clavenna, L.R., Davis, S.M., Fiato, R.A., and Say, G.R., US Patent 5348717 (1994).

49. Say, G.R., and Tayor, J.H., US Patent 5421840 (1995).

50. Long, D.C., EP Patent 673877 (1997).

51. Balachandran, U., Dusek, J.T., Mieville, R.L., Poeppel, R.B., Kleefisch, M.S., Pei, S.,

Kobylinski, T.P., Udovich, C.A., and Bose, A.C., Appl. Catal. A: Gen. 133, 19 (1995).

52. Jin, W., Gu, X., Li, S., Huang, P., Xu, N., and Shi, J., Chem. Engng Sci. 55, 2617 (2000).

53. Jin, W.Q., Li, S.G., Huang, P., Xu, N.P., Shi, J., and Lin, Y.S., J. Membr. Sci. 166, 13 (2000).

54. Teraoka, Y., Zhang, H.M., Furukawa, S., and Yamozoe, N., Chem. Lett. 1743 (1985).

55. Teraoka, Y., Nobunaga, T., and Yamazoe, N., Chem. Lett. 503 (1988).

56. Mazanec, T.J., Cable, T.L., and Frye, J.G. Jr., Solid State Ion. 53111 (1992).

57. Gur, T.M., Belzner, A., and Huggins, R.A., J. Membr. Sci. 75, 151 (1992).

58. Cable, T.L., European Patent 0 399 833 A1 (1990).

59. Cable, T.L., European Patent 0 438 902 A2 (1991).


60. Wiley, J.B., and Poeppelmeier, K.R., J. Solid State Chem. 88, 250 (1990).

61. Eror, N.G., and Balachandran, U., J. Solid State Chem. 40, 85 (1981).

62. Balachandran, U., and Eror, N.G., J. Phys. Chem. Solids 44, 231 (1983).

63. Teraoka, Y., Nobunaga, T., Okamoto, K., Miura, N., and Yamazoe, N., Solid State Ion. 48,

207 (1991).

64. Balachandran, U., Dusek, J.T., Maiya, P.S., Mieville, R.L., Kleefisch, M.S., and Udovich,

C.A., Catal. Today 133, 19 (1997).

65. Balachandran, U., Dusek, J.T., Picciolo, J.J., Maiya, P.S., Ma, B., Mieville, R.L.,

Kleefisch, M.S., and Udovich, C.A., Prepr. ACS Div. Fuel Chem. 42, 591 (1997).

66. Balachandran, U., Ma, B., Maiya, P.S., Mieville, R.L., Dusek, J.T., Picciolo, J.J., Guan, J.,

Dorris, S.E., and Liu, M., Solid Sate Ion. 108, 363 (1998).

67. Mazanec, T.J., Cable, T.L., Frye, J.G., and Kliewer, W.R., US Patent 5,306,411 (1994).

68. Schwartz, M., While, J.H., Myers, M.G., Deych, S., and Sammells, A.F., Prepr. ACS Div.

Fuel Chem. 42, 596 (1997).

69. Tsai, C.Y., Dixon, A.G., Moser, W.R., and Ma, Y.H., AIChE J. 43, 2741 (1997).

70. Tsai, C.Y., Dixon, A.G., Ma, Y.H., Moser, W.R., and Pascucci, M.R., J. Am. Ceram. Soc.

81, 1437 (1998).

71. Li, G.T., Shao, Z.P., Xiong, G.X., Dong, H., Cong, Y., and Yang, W.S., J. Inorg. Mater.

17, 1041 (2002).

72. Dong, H., Shao, Z.P., Xiong, G.X., Tong, J.H., Sheng, S.S., and Yang, W.S., Catal. Today

67, 3 (2001).

73. Tong, J.H., Yang, W.S., Cai, R., Zhu, B.C., and Lin, L.W., Catal. Lett. 78, 129 (2002).

74. DOE Techline, May 20,1997.

75. Oil & Gas Journal, Page 62, May 3, 1999.

76. Riess, I., in “The CRC Handbook of Solid State Electrochemistry” (P.J. Gellings and

H.J.M. Bouwmeester, Eds.), p. 223. CRC Press, Boca Raton, FL, 1997.

77. Bouwmeester, H.J.M., and Burggraaf, A.J., in “Fundamentals of Inorganic Membrane

Science and Technology” (A.J. Burggraaf and L. Cot, Eds.), p. 435. Elsevier, Amsterdam,

1996.

78. Teraoka, Y., Nobunaga, T., Okamoto, K., Miura, N., and Yamazoe, N., Solid State Ion. 48,

207 (1991).

79. Teraoka, Y., Zhang, H.M., Okamoto, K., and Yamazoe, N., Mater. Res. Bull. 23, 51 (1988).

80. Mazanec, T.J., Solid State Ion. 70/71, 11 (1994).

81. Steele, B.C.H., Mater. Sci. Engng B 13, 79 (1992).

82. Steele, B.C.H., Curr. Opin. Solid State Mater. Sci. 1, 684 (1996).

83. Lewis, W.K., Gilliland, E.R., and Reed, W.A., Ind. Engng Chem. 41, 1227 (1949).

84. Lewis, W.K., and Gilliland, E.R., US Patent 2,603,608 (1952).

85. Lewis, W.K., US Patent 2,631,934 (1953).

86. Read, W.A., DSc Dissertation, Massachusetts Institute of Technology, 1948.

87. Huszar, K., Racz, G., and Szekely, G., Acta Chim. Acad. Sci. Hungary 70, 287 (1971).

88. Gavalas, G.R., Phichitcul, C., and Voecks, G.E., J. Catal. 88, 54 (1984).

89. Kikuchi, E., and Chen, Y., Natural gas conversion V. Stud. Surf. Sci. Catal. 119, 441

(1998).

90. Elmasides, C., Ioannides, T., and Verykios, X.E., Natural gas conversion V. Stud. Surf.

Sci. Catal. 119, 801 (1998).

91. Yan, Q.G., Chu, W., Gao, L.Z., Yu, Z.L., and Yuan, S.Y., Natural gas conversion V. Stud.

Surf. Sci. Catal. 119, 855 (1998).

92. Nakagawa, K., Ikenaga, N., Suzuki, T., Kobayashi, T., and Haruta, M., Appl. Catal. A:

Gen. 169, 281 (1998).


93. Boucouvalas, J., Efstathiou, A.M., Zhang, Z.L., and Verykios, X.E., Natural gas

conversion IV. Stud. Surf. Sci. Catal. 107, 435 (1997).

94. Laa, J.C., Berger, R.J., and Marin, G.B., Catal. Lett. 43, 63 (1997).

95. Soick, M., Buyevskaya, O., Hohenberger, M., and Wolf, D., Catal. Today 32, 163 (1996).

96. Nakagawa, K., Suzuki, T., Kobayashi, T., and Haruta, M., Chem. Lett. 1029 (1996).

97. Boucouvalas, Y., Zhang, Z.L., and Verykios, X.E., Catal. Lett. 40, 189 (1996).

98. Mattos, L.V., Oliveira, E.R., Resende, P.D., Noronha, F.B., and Passos, F.B., Catal. Today

77, 245 (2002).

99. Pantu, P., and Gavalas, G.R., Appl. Catal. A: Gen. 223, 253 (2002).

100. Elmasides, C., Kondarides, D.I., Neophytides, S.G., and Verykios, X.E., J. Catal. 198, 195

(2001).

101. Wang, H.Y., and Ruckenstein, E., J. Catal. 186, 181 (1999).

102. Wang, H.Y., and Ruckenstein, E., Catal. Lett. 59, 121 (1999).

103. Mallens, E.P.J., Hoebink, J.H.B.J., and Marin, G.B., Catal. Lett. 33, 291 (1995).

104. Wang, D., Dewaele, O., Groote, A.M., and Froment, G.F., J. Catal. 159, 418 (1996).

105. Hickman, D.A., and Schmidt, L.D., Catal. Lett. 17, 223 (1993).

106. Hofstad, K.H., Hoebink, J.H.B.J., Holmen, A., and Marin, G.B., Catal. Today 40, 157 (1998).

107. Dissanayake, D., Rosynek, M.P., and Lunsford, J.H., J. Catal. 132, 117 (1991).

108. Choudhary, V.R., Rane, V.H., and Rajput, A.M., Catal. Lett. 22, 289 (1993).

109. Drago, R.S., Jurczyk, K., Kob, N., Bhattacharyya, A., and Masin, J., Catal. Lett. 51, 177

(1998).

110. Claridge, J.B., Green, M.L.H., Tsang, S.C., York, A.P.E., Ashcroft, A.T., and Battle, P.D.,

Catal. Lett. 22, 299 (1993).

111. Tsipouriari, V.A., Zhang, Z., and Verykios, X.E., J. Catal. 179, 283 (1998).

112. Wang, J.K., Hu, Y.H., Weng, W.Z., Zheng, B.L., and Wan, H.L., Acta Chim. Sinica 54,

869 (1996).

113. Tang, S., Lin, J., and Tan, K.L., Catal. Lett. 51, 169 (1998).

114. Takehira, K., Shishido, T., and Kondo, M., J. Catal. 207, 307 (2002).

115. Zhu, T.L., and Flytzani-Stephanopoulos, M., Appl. Catal. A: Gen. 208, 403 (2001).

116. Lago, R., Bini, G., Pena, M.A., and Fierro, J.L.G., Third world congress on oxidation

catalysis. Stud. Surf. Sci. Catal. 110, 721 (1997).

117. Cao, L., Chen, Y., and Li, W., Natural gas conversion IV. Stud. Surf. Sci. Catal. 107, 467

(1997).

118. Liu, S.L., Xiong, G.X., Sheng, S.S., Miao, Q., and Yang, W.S., Natural gas conversion V.

Stud. Surf. Sci. Catal. 119, 747 (1998).

119. Wang, J.G., Liu, C.J., Zhang, Y.P., Zhu, X.L., Zou, J.J., Yu, K.L., and Eliasson, B., Chem.

Lett. 1068 (2002).

120. Shiozaki, R., Andersen, A.G., Hayakawa, T., Hamakawa, S., Suzuki, K., Shimizu, M., and

Takehira, K., Third world congress on oxidation catalysis. Stud. Surf. Sci. Catal. 110, 701

(1997).

121. Lezaun, J., Gomez, J.P., Blanco, M.D., Cabrera, I., Pena, M.A., and Fierro, J.L.G., Natural

gas conversion V. Stud. Surf. Sci. Catal. 119, 729 (1998).

122. Zhang, Y.H., Xiong, G.X., Sheng, S.S., and Yang, W.S., Catal. Today 63, 517 (2000).

123. Nichio, N., Casella, M., Ferretti, O., Gonzalez, M., Nicot, C., Moraweck, B., and Frety, R.,

Catal. Lett. 42, 65 (1996).

124. Lee, K.M., and Lee, W.Y., Catal. Lett. 83, 65 (2002).

125. Hayakawa, T., Harihara, H., Andersen, A.G., Suzuki, K., Yasuda, H., Tsunoda, T.,

Hamakawa, S., York, A.P.E., Yoon, Y.S., Shimizu, M., and Takehira, K., Appl. Catal. A:

Gen. 149, 391 (1997).


126. Claridge, J.B., Green, M.L.H., Tsang, S.C., York, A.P.E., Ashcroft, A.T., and Battle, P.D.,

Catal. Lett. 22, 299 (1993).

127. Poirier, M.G., Trudel, J., and Guary, D., Catal. Lett. 21, 99 (1993).

128. Boucouvalas, Y., Zhang, Z.L., and Verykios, X.E., Catal. Lett. 27, 131 (1994).

129. Hu, Y.H., and Ruckenstein, E., J. Catal. 158, 260 (1996).

130. (a) Hu, Y.H., and Ruckenstein, E., J. Phys. Chem. B 102, 230 (1998); (b) Hu, Y.H., and

Ruckenstein, E., Acc. Chem. Res. 36, 791 (2003).

131. Hu, Y.H., and Ruckenstein, E., J. Phys. Chem. A 102(1), 10568 (1998).

132. Au, C.T., Liao, M.S., and Ng, C.F., J. Phys. Chem. A 102, 3959 (1998).

133. Papp, H., Schuler, P., and Zhuang, Q., Catal. Lett. 3, 299 (1996).

134. Hu, Y.H., and Ruckenstein, E., Catal. Lett. 34, 41 (1995).

135. Au, C.T., Liao, M.S., and Ng, C.F., Chem. Phys. Lett. 267, 44 (1997).

136. Buyevskaya, O.V., Walter, K., Wolf, D., and Baerns, M., Catal. Lett. 38, 81 (1996).

137. Au, C.T., and Wang, H.Y., J. Catal. 167, 337 (1997).

138. VanLooij, F., and Geus, J.W., J. Catal. 168, 154 (1997).

139. Au, C.T., Hu, Y.H., and Wan, H.L., Catal. Lett. 27, 199 (1994).

140. Au, C.T., Hu, Y.H., and Wan, H.L., Catal. Lett. 36, 159 (1996).

141. Campbell, R.A., Lenz, J.P., and Goodman, D.W., Catal. Lett. 17, 39 (1993).

142. Li, C., Yu, C.C., and Shen, S.K., Catal. Lett. 75, 183 (2001).

143. Tsipouriari, V.A., Zhang, Z., and Verykios, X.E., J. Catal. 179, 283 (1998).

144. Elmasides, C., and Verykios, X.E., J. Catal. 203, 477 (2001).

145. Weng, W.Z., Yan, Q.G., Luo, C.R., Liao, Y.Y., and Wan, H.L., Catal. Lett. 74, 37 (2001).

146. Nakagawa, K., Ikenaga, N., Teng, Y.H., Kobayashi, T., and Suzuki, T., J. Catal. 186, 405

(1999).

147. Mallens, E.P.J., Hoebink, J.H.B.J., and Marin, G.B., J. Catal. 167, 43 (1997).

148. Au, C.T., Ng, C.F., and Liao, M.S., J. Catal. 185, 12 (1999).

149. Shustorovich, E., Adv. Catal. 37, 101 (1990).

150. Tang, S., Lin, J., and Tan, K.L., Catal. Lett. 55, 83 (1998).


152. Wang, H.Y., and Ruckenstein, E., J. Phys. Chem. 103, 11327 (1999).

153. Hu, Y.H., and Feeley, J.S., AIChE J. 49, 3253 (2003).

154. Ji, Y.Y., Li, W.Z., Xu, H.Y., and Chen, Y.X., Catal. Lett. 71, 45 (2001).

155. Hayakawa, T., Andersen, A.G., Shimizu, M., Suzuki, K., and Takehira, K., Catal. Lett. 22,

307 (1993).

156. Ruckenstein, E., and Hu, Y.H., Appl. Catal. A: Gen. 183, 85 (1999).

157. Wang, D.Z., Dewaele, O., DeGroote, A.M., and Froment, G.F., J. Catal. 159, 418 (1996).

158. Schouten, S.C., Gijzeman, O.L.J., and Bootsma, G.A., Bull. Soc. Chim. Belg. 88, 541

(1979).

159. Bartholomew, C.H., Catal. Rev.—Sci. Engng 24, 67 (1982).

160. Baker, R.T.K., and Harris, P.S., in “Chemistry and Physics of Carbon” (P.L. Walker Jr.,

Ed.), Vol. 14, p. 83. Marcel Dekker, New York, 1979.

161. Baukloh, W., Chatterjee, B., and Das, P.P., Trans. Indian Inst. Metals 4, 271 (1950).

162. Kehrer, V.J., and Leidheiser, H. Jr., J. Phys. Chem. 58, 550 (1954).

163. Rostrup-Nielsen, J.R., J. Catal. 27, 343 (1972).

164. Renshaw, G.D., Roscoe, C., and Walker, P.L., J. Catal. 22, 394 (1971).

165. Grenga, H.E., and Lawless, K.R., J. Appl. Phys. 43, 1508 (1972).

166. Renshaw, G.D., Roscoe, C., and Walker, P.L., J. Catal. 18, 164 (1970).

167. Eizenberg, M., and Blakely, J.M., Surf. Sci. 82, 228 (1979).

168. Reitmeier, R.E., Atwood, K., Bennet, H.A. Jr., and Baugh, H.M., Ind. Engng Chem. 40,

620 (1948).


169. White, G.A., Roszkowski, T.R., and Stanbridge, D.W., Hydrocarbon Process. 54, 130

(1975).

170. Tokunaga, O., and Ogasawara, S., React. Kinet. Catal. Lett. 39, 69 (1989).

171. (a) Ruckenstein, E., and Hu, Y.H., Chem. Innovat. 30, 39 (2000); (b) Hu, Y.H., and

Ruckenstein, E., J. Catal. 184, 298 (1999); (c) Hu, Y.H., and Ruckenstein, E., J. Catal.

163, 306 (1996); (d) Hu, Y.H., and Ruckenstein, E., Langmuir 13, 2055 (1997).

172. Hu, Y.H., and Ruckenstein, E., Catal. Rev.—Sci. Engng 44, 423 (2002).

173. Rostrup-Nielsen, J.R., Stud. Surf. Sci. Catal. 68, 85 (1991).

174. (a) Udengaar, N.R., Hansen, J.-H.B., Hanson, D.C., and Stal, J.A., Oil Gas J. 90, 62

(1992); (b) Dibbern, H.C., Olesen, P., Rostrup-Nielsen, J.R., Totrrup, P.B., and

Udengaard, N.R., Hydrocarbon Process. 65, 71 (1986).

175. Rostrup-Nielsen, J.R., and Hansen, J.H.B., J. Catal. 144, 38 (1993).

176. Yamazaki, O., Nozaki, T., Omata, K., and Fujimoto, K., Chem. Lett. 1953 (1992).

177. Zhang, Z.L., and Verykios, X.E., Catal. Today 21, 589 (1994).

178. Horiuchi, T., Sakuma, K., Fukui, T., Kubo, Y., Osaki, T., and Mori, T., Appl. Catal. A:

Gen. 144, 111 (1996).

179. Kim, G.J., Cho, D.-S., Kim, K.-H., and Kim, J.-H., Catal. Lett. 28, 41 (1994).

180. Inui, T., Catalysis, Vol. 16, p. 133, The Royal Society of Chemistry, Cambridge,

2002.

181. Perera, J.S.H., Couves, J.W., Sankar, G., and Thomas, J.M., Catal. Lett. 11, 219 (1991).

182. Qin, D., and Lapszewicz, J., Catal. Today 21, 551 (1994).

183. Solymosi, F., Kustan, Gy., and Erohelyi, A., Catal. Lett. 11, 149 (1991).

184. Erdohelyi, A., Fodor, K., and Solymosi, F., Stud. Surf. Sci. Catal. 107, 525 (1997).

185. Bitter, J.H., Hally, W., Seshan, K., van Ommen, J.G., and Lercher, J.A., Catal. Today 29,

349 (1996).

186. Bitter, J.H., Seshan, K., and Lercher, J.A., J. Catal. 176, 93 (1998).

187. O’Connor, A.M., and Ross, J.R.H., Catal. Today 46, 203 (1998).

188. Bitter, J.H., Seshan, K., and Lercher, J.A., J. Catal. 183, 336 (1999).

189. Souza, M.M.V.M., Aranda, D.A.G., and Schmal, M., Ind. Engng Chem. Res. 41, 4681

(2002).

190. Chen, Y.Z., Liaw, B.J., and Lai, W.H., Appl. Catal. A: Gen. 230, 73 (2002).

191. Stagg, S.M., and Resaco, D.E., Stud. Surf. Sci. Catal. 111, 543 (1997).

192. Sigl, M., Bradford, M.C.J., Knozinger, H., and Vannice, M.A., Top. Catal. 8, 211 (1999).

193. Stagg-Williams, S.M., Noronha, F.B., Fendley, G., and Resasco, D.E., J. Catal. 194, 240

(2000).

194. Kim, J.H., Suh, D.J., Park, T.J., and Kim, K.L., Appl. Catal. A: Gen. 197, 191 (2000).

195. Bhattacharyya, A., and Chang, V.W., Stud. Surf. Sci. Catal. 88, 207 (1994).

196. Halliche, D., Bouarab, R., Cherifi, O., and Bettahar, M.M., Catal. Today 29, 373 (1996).

197. (a) Ruckenstein, E., and Hu, Y.H., J. Catal. 162, 230 (1996); (b) Hu, Y.H., and

Ruckenstein, E., J. Phys. Chem. B 101, 7563 (1999).

198. Wang, S., and Lu, G.O.M., Appl. Catal. B: Environ. 16, 269 (1998).

199. Yan, Z.-F., Ding, R.-G., Song, L.-H., and Qian, L., Energy Fuels 12, 1114 (1998).

200. Wang, S., and Lu, G.Q., Appl. Catal. A: Gen. 169, 271 (1998).

201. Wang, S., and Lu, G.Q., Ind. Engng Chem. Res. 38, 2615 (1999).

202. Ito, M., Tagawa, T., and Goto, S., Appl. Catal. A: Gen. 177, 15 (1999).

203. Xu, Z., Li, Y.M., Zhang, J.Y., Chang, L., Zhou, R.Q., and Duan, Z.T., Appl. Catal. A: Gen.

210, 45 (2001).

204. Wang, S.B., and Lu, G.Q., Energy Fuels 12, 1235 (1998).

205. Chen, Y.G., and Ren, J., Catal. Lett. 29, 39 (1994).

206. Sridhar, S., Sichen, D., and Seetharaman, S., Z. Metallkd. 85, 9 (1994).


207. Kim, J.H., Suh, D.J., Park, T.J., and Kim, K.L., Natural gas conversion V. Stud. Surf. Sci.

Catal. 119, 771 (1998).

208. Osaki, T., Horiuchi, T., Sugiyama, T., Suzuki, K., and Mori, T., Catal. Lett. 52, 171

(1998).

209. Hayashi, H., Murata, S., Tago, T., Kishida, M., and Wakabayashi, K., Sekiyu Gakkaishi—

J. Jpn Petrol. Inst. 44, 334 (2001).

210. Bhattacharyya, A., Chang, V.W., and Schumacher, D.J., Appl. Clay Sci. 13, 317 (1998).

211. Mori, T., Osaki, T., Horiuchi, T., Sugiyama, T., and Suzuki, K., Catalyst deactivation

1999. Stud. Surf. Sci. Catal. 365 (1999).

212. Lemonidou, A.A., Goula, M.A., and Vasalos, I.A., Catal. Today 46, 175 (1998).

213. Osaki, T., and Mori, T., J. Catal. 204, 89 (2001).

214. Wang, S.B., and Lu, G.Q.M., J. Chem. Technol. Biotechnol. 75, 589 (2000).

215. Choi, J.S., Moon, K.I., Kim, Y.G., Lee, J.S., Kim, C.H., and Trimm, D.L., Catal. Lett. 52,

43 (1998).

216. Seok, S.H., Han, S.H., and Lee, J.S., Appl. Catal. A: Gen. 215, 31 (2001).

217. Quincoces, C.E., de Vargas, S.P., Grange, P., and Gonzalez, M.G., Mater. Lett. 56,

698 (2002).

218. Crisafulli, C., Scire, S., Maggiore, R., Minico, S., and Galvagno, S., Catal. Lett. 59, 21

(1999).

219. Slagtern, A., Olsbye, U., Blom, R., Dahl, I.M., and Fjellvag, H., Appl. Catal. A: Gen. 165,

379 (1997).

220. Li, W.Y., Feng, J., and Xie, K.C., React. Kinet. Catal. Lett. 64, 381 (1998).

221. Wang, S.B., and Lu, G.Q., Appl. Catal. B: Environ. 19, 267 (1998).

222. Xu, G.L., Shi, K.Y., Gao, Y., Xu, H.Y., and Wei, Y.D., J. Mol. Catal. A: Chem. 147,

47 (1999).

223. Kroll, V.C.H., Swaan, H.M., and Mirodatos, C., J. Catal. 161, 409 (1996).

224. Kroll, V.C.H., Delichure, P., and Mirodatos, C., Kinet. Catal. 37, 698 (1996).

225. Takayasu, O., Takegahara, Y., and Matsuura, I., Energy Conserv. Mgmt 36, 597 (1995).

226. Quincoces, C.E., de Vargas, S.P., Diaz, A., Montes, M., and Gonzalez, M.G., Natural gas

conversion V. Stud. Surf. Sci. Catal. 119, 837 (1998).

227. Zhang, Z.L., and Verykios, X.E., J. Chem. Soc., Chem. Commun. 71 (1995).

228. Zhang, Z., and Verykios, X.E., Appl. Catal. A: Gen. 138, 109 (1996).

229. Wang, S.B., and Lu, G.Q., Energy Fuels 12, 248 (1998).

230. Ruckenstein, E., and Hu, Y.H., J. Catal. 161, 55 (1996).

231. Liu, B.S., and Au, C.T., Catal. Lett. 85, 165 (2003).

232. Hally, W., Bitter, J.H., Seshan, K., Lercher, J.A., and Ross, J.R.H., Catalyst deactivation

1994. Stud. Surf. Sci. Catal. 167 (1994).

233. Li, X.S., Chang, J.S., Lee, F.K., and Park, S.E., React. Kinet. Catal. Lett. 67, 383 (1999).

234. Li, X.S., Chang, J.S., and Park, S.E., React. Kinet. Catal. Lett. 67, 375 (1999).

235. Li, X.S., Chang, J.S., Tian, M.Y., and Park, S.E., Appl. Organomet. Chem. 15, 109 (2001).

236. Lercher, J.A., Bitter, J.H., Halley, W., Niessen, W., and Seshan, K., in “Proceedings of the

11th International Congress on Catalysis” (J.W. Hightower, W.N. Delgass, E. Iglesia and

A.T. Bell, Eds.), Stud. Surf. Sci. Catal., 101, p. 463. Elsevier, Amsterdam, 1996.

237. Wei, J.M., Xu, B.Q., Li, J.L., Cheng, Z.X., and Zhu, Q.M., Appl. Catal. A: Gen. 196,

L167 (2000).

238. Tomishige, K., Yamazaki, O., Chen, Y., Yokoyama, K., Li, X., and Fujimoto, K., Catal.

Today 45, 35 (1998).


240. Osaki, T., Masuda, H., Horiuchi, T., and Mori, T., Catal. Lett. 34, 59 (1995).


241. Guerrero-Ruiz, A., Sepulveda-Escribano, A., and Rodriguez-Ramos, I., Catal. Today 21,

545 (1994).

242. Osaki, T., Horiuchi, T., Suzuki, H., and Mori, T., Appl. Catal. A: Gen. 155, 229 (1997).

243. Dibbern, H.C., Olesen, P., Rostrup-Nielsen, J.R., Tottrup, P.B., and Udengaard, N.R.,

Hydrocarbon Process. 65, 71 (1986).

244. Hayakawa, T., Suzuki, S., Nakamura, J., Uchijima, T., Hamakawa, S., Suzuki, K.,

Shishido, T., and Takehira, K., Appl. Catal. A: Gen. 183, 273 (1999).

245. Chang, J.S., Park, S.E., and Lee, K.W., Stud. Surf. Sci. Catal. 84, 1587 (1994).

246. Park, S.E., Nam, S.S., Choi, M.J., and Lee, K.W., Energy Conserv. Mgmt 36, 573 (1995).

247. Luo, J.Z., Gao, L.Z., Ng, C.F., and Au, C.T., Catal. Lett. 62, 153 (1999).

248. Chang, J.S., Park, S.-E., and Chon, H., Appl. Catal. A: Gen. 145, 111 (1996).

249. Choudhary, V.R., Uphade, B.S., and Mamman, A.S., Micropor. Mesopor. Mater. 23,

61 (1998).

250. Suzuki, I., Okajima, S., and Seya, K., J. Phys. Earth 27, 69 (1979).

251. Arai, H., and Machida, M., Catal. Today 10, 81 (1991).

252. Gadalla, A.M., and Sommer, M.E., J. Am. Ceram. Soc. 72, 683 (1989).

253. Chen, Y., Tomishige, K., and Fujimoto, K., Chem. Lett. 999 (1997).

254. Takayasu, O., Soman, C., Takegahara, Y., and Matsuura, I., Stud. Surf. Sci. Catal. 88,

281 (1994).

255. Takayasu, O., Hongo, N., and Matsuura, I., Stud. Surf. Sci. Catal. 77, 305 (1993).

256. Swaan, H.M., Kroll, V.C.H., Martin, G.A., and Mirodatos, C., Catal. Today 21, 545 (1994).

257. (a) Hu, Y.H., and Ruckenstein, E., Catal. Lett. 43, 71 (1997); (b) Ruckenstein, E., and

Hu, Y.H., Catal. Lett. 51, 183 (1998).



260. Chen, Y., Tomishige, K., Yokoyama, K., and Fujimoto, K., Appl. Catal. A: Gen. 165, 335

(1997).

261. Tomishige, K., Yamazaki, O., Chen, Y., Yokoyama, K., Li, X., and Fujimoto, K., Catal.

Today 45, 35 (1998).

262. Tomishige, K., Himeno, Y., Matsuo, Y., Yoshinaga, Y., and Fujimoto, K., Ind. Engng

Chem. Res. 39, 1891 (2000).

263. Chen, Y.G., Tomishige, K., and Fujimoto, K., Appl. Catal. A: Gen. 161, L11 (1997).

264. Ruckenstein, E., and Wang, H.Y., Appl. Catal. A: Gen. 204, 257 (2000).

265. Ruckenstein, E., and Wang, H.Y., J. Catal. 205, 289 (2002).

266. Ruckenstein, E., and Wang, H.Y., Catal. Lett. 73, 99 (2001).


Index

1-butene, and alkylation isomerization,

240

2-butene, and alkylation isomerization,

240

4,6-Dimethyldibenzothiophene

(DMDBT), 98–99

5A Zeolite, 332

A

Acceptor-doped perovskite oxides,

307–308

Acid catalysis, 105–116

Acid runaway condition, 254

Acid-soluble oil (ASO), 239

Acid strength

and alkylation reaction, 256–260

in TS-1, 26–28

Activation energy

and adsorption by zeolites, 256

of cracking, 249

and hydride transfer, 243

Active sites. see Titanosilicate surface

structures

Adamantane, 246

AEM (analytical EM), 177

Alcohol oxidation, 100

Aliphatic compound hydroxylation,

85–89

Alkanes, cracking, 249

Alkene epoxidation, 70

and O–O bond cleavage, 138–140

Alkene space velocity (OSV), 274–275

AlkyCleanTM, 286

Alkyl hydroperoxides, 80–81

Alkylation, 235–236t

about, 230–233

and alkane/alkene ratio, 274–275

alkene addition, 239–241

and alkene feed composition,

276–278

and alkene space velocity,

274–275

and aluminum content, 262

of carbenium ions with isobutane,

241

and coke formation, 245–246

cracking, 247–249

hydride transfer, 242–246

initiation steps, 237–239

isobutylene protonation, 238f

isomerization, 239–241

isomerization catalysts, 240–241

isomerization for TMP isomers,

241f

oligomerization, 247–249

pathways to allylic/cyclic com-

pounds, 251–252

and process parameters, 272t

product distribution, 234–237

reaction rate and interface area,

253–254

reaction temperature, 272–274

refinery process unit, 232f

self-alkylation, 249–250

solid-acid processes, 283–288

sulfated zirconia, 267–271

Alkylation as industrial process

ConocoPhillips HF-catalyzed

process, 281–282

347

ExxonMobil auto-refrigerated

alkylation, 280

Haldor Topsøe FBATM process,

287–288

LURGI EUROFUEL, 286–287

ReVapTM alkylation process, 282

Stratco Alkysafe, 281

sulfuric-acid as catalyst, 278–281

UOP Alkad process, 283

UOP AlkyleneTM, 285–286

Alkylation catalysts. see also Zeolites

acid strength, 257–258

heteropolyacids, 268–269

liquid acid properties, 253–255

Nafion-H, 269–270

rare earth exchanged zeolites,

263–264

and silicon/aluminum ratio, 261–263

sulfated metal oxides, 268

supported metal halides, 270–271

VTM-A, 266

ALPO catalysts, 221

AlPO4, 332

Aluminum, and alkylation, 260–261

Aluminum chloride, and isomeriza-

tion, 241

Ammoximation, 92

Analytical electron microscope

(AEM), 191–192

Aromatic compound, hydroxylation,

89–90

Aromatic polycarbonates, 108

Atomic resolution ETEM, 197–200

of butane oxidation, 203–210

of nanorods, 210

Autothermal reforming of methane,

306

B

Ba0.5Sr0.5Co0.8Fe0.2O32x membrane

reactor, 309

Back-scattered electrons (BSE),

213–217

Baeyer-Villiger (BV) oxidation, 102,

104t

BEA

and silicon/aluminum ratio,

261–262

time-on-stream behavior, 265

Beckmann rearrangement, 106

Benzene hydroxylation,

134–135

Benzoquinones, 101

b-scission, 247, 248–249

Boron trifluoride, 271

Boudouard reaction, 321

Bright-field (BF) image, 179

Brønsted acid sites

and catalytic action of titanosilicate

molecular sieves, 56

creation during TS-1 and H2O2

interaction, 130

in TS-1, 26–28

in zeolites, 256–260

Brownmillerite membrane, 308

BSE and HAADF detector geometry,

215f

Butane oxidation, 203–210

C

CvN cleavage reactions, 105

Carbon deposition, with Ni/Al2O3

catalysts, 325–327

Carbon dioxide

as product of partial oxidation of

methane, 316–318

Carbon monoxide

as product of partial oxidation of

methane, 316–318

Carbon monoxide disproportionation,

321

Index348

Catalysts. See also reactions by name

to locate references to catalysts

for these reactions, 177–178

deactivation by metal-support

interactions, 201

defects and TEM analysis, 180

dislocation types, 181

glide shear plane mechanism,

208–210

high-silica microporous SSZ-48

type, 187–188

HRTEM analysis of L-type zeolites,

185–186

HRTEM analysis of MAPO-36

microporous type, 186–189

intergrowths in zeolites, 188–190

in situ analysis, 196–200

solid in liquid environments,

210–212

titanosilicate molecular sieves, 56,

132–136

Catalysts (for conversion of methane)

and carbon deposition in CO2

reforming, 322–323

changes during partial oxidation of

methane, 315–316

Co/Al2O3, 331

MgO-containing catalysts, 332–337

MoS2, 331

Ni/Al2O3, 324

Ni/ZrO2, 330–331

noble metal types for CO2

reforming, 323–324

for partial oxidation of methane,

312–314

reduced and oxidized for partial

oxidation of methane, 320–321

WS2, 331–332

Cathodoluminescence imaging,

195–196

Ceramic membrane reactor, 308f

configuration for partial oxidation of

methane, 310f

Chlorided alumina, and alkene feed

composition, 277

Chlorinated alumina, 270

Co/Al2O3 catalyst, 331

CO2 photoreduction, 120–121

CO2 reforming of methane

carbon deposition, 299

carbon deposition pathways,

321–322

catalyst and carbon deposition,

322–323

development of processes, 299


Ni/Al2O3 catalysts, 325–327

noble metal catalysts, 323–324

SPARG process, 332

Cobalt catalysts for partial oxidation of

methane, 314

ConocoPhillips HF-catalyzed process,

281–282

Coordination sphere, expansion of

TS-1, 31–32, 38

Copper catalysts, 200, 203

Cracking

and alkylation, 248–249

and reaction temperature, 272–273

CTEM (conventional transmission

electron microscope), 178

Cyclic carbonates, 107t–108t

transesterification, 109

Cyclic voltametry of oxo-titanium

species, 41–42

Cyclohexane hydroxylation, 88

Cytochrome P450, 60

D

Dark field (DF) image, 179, 193f

DFT, and titanosilicate surface

structure, 50

Index 349

Dialkene epoxidation, 71–72t

Diamagnetic peroxo/hydroperoxo

species, 42

Diethyl malonate, 113t

Diffuse reflectance UV-visible

spectroscopy

and concentrations of superoxo and

hydroperoxo species, 44

for oxo-titanium species, 37

TS-1 (titanium silicate-1), 46f

titanosilicate, 13t

titanosilicate surface structure, 14f

Dimethyl terephthalate (DMT), 110

Dimethylcarbonates, 110

Dimethylhexanes, 234, 248

Dioxygen, 83

Diphenyl carbonate (DPC), 108

Dislocation types, 181

DMDBT, 98–99

E

ECELLs, 196–200, 222

Edge dislocations, 181

EDX (energy dispersive X-ray

spectroscopy), 182, 191

EFAL (extra-framework aluminum

species), 260–261

Effluent Refrigerated Sulfuric Acid

Alkylation Process (Stratco),

278–279

EFTEM (energy-filtered TEM), 176

Electron crystallography, 220

Electron diffraction (ED), 174

Electron microscopy (EM)

aberration correction, 222–223

analytical electron microscopy

(AEM), 191–192

application to catalysis, 176–177

beam damage, 222

in catalysis, 177–178

cathodoluminescence imaging,

195–196

challenges, 220–223

charge-coupled device, 177

diffraction patterns, 178

electron-beam damage, 177

electron tomography, 212–218

energy-filtered TEM, 218–219

environmental scanning electron

microscopy (ESEM), 212

ETEM, 196–200

HRLVSEM, 195

image plate (IP), 177

imaging, 178–179

lens aberrations, 222

methods, 176–181

point resolution, 178

ray diagram, 180f

resolution, 222

sample preparation, 176

spatial mapping, 192

STEM (scanning

transmission EM), 177,

193–195, 222

TEM imaging methods, 179–181

theoretical procedures, 181

wet-ETEM, 210–212

Electron tomography

about, 212–218

BSE and HAADF, 213–217

nanoparticle location, 218

ELNES (electron energy loss

near-edge structure), 176

EMT

acid strength and catalyst role in

alkylation, 258

intergrowths in, 189

and rare earth exchange, 263–264

SINTEF research, 266

Energy

Gibbs free energy for methane

transformations, 300t

Index350


potential energy profiles for hydride

transfer, 244f

Energy-filtered transmission electron

microscopy, 218–219

Enthalpy

of adsorption on zeolites, 256–257f

of hydride transfer step in

alkylation, 244

of O–O cleavage in H2O2, 57

Environmental scanning electron

microscopy, 212

Environmental transmission electron

microscopy, 196–200

Epichlorohydrin, 62

Epoxidation

and alkene structure, 70–71

alkenes and alcohol functions,

72–73t

alkenes and alkanes, 72

alkenes and O–O bond cleavage,

138–140

alkenes containing carbonyl groups,

81–82

with alkyl hydroperoxides, 80–81

allyl alcohol, 80t, 125t

catalyzed by mesoporous titanium

silicates, 67–70

and concentration of titanium oxo

species, 129

of cyclohexene and silylation, 127t

dialkenes, 71–72t

diastereoselectivity, 74–75, 77–78t

diffusional constraints, 62-63t

general features, 60–62

of hex-1-ene, 66t

hydroperoxide involvement, 132

hydroxyl-assisted, 72–74

of oleic acid, 67t

and pH, 78–80t

reaction rates for alkenes, 71t

side reactions, 75–76

stereospecificities, 62

and Ti-silicate structure, 65–67

of unsaturated cyclic terpenes, 69t

using dioxygen, 83

using urea-H2O2, 82

yields, 62

EPR spectroscopy, oxo-titanium

species, 42–49

Ester transesterification, 110

ETEM, 196–200

Ethane/ethene hydride transfer, 243

Ethene oxidation, 131–132

Ether oxidation, 100–101

Ethylacetoacetate, 111–112, 114t

ETS-10

synthesis, 154–156

vibrational spectroscopy, 25

Euro-TS-1, diffusional constraints, 63t

EXAFS

oxo-titanium species, 39–41

and titanosilicate surface structure,

50–51t

ExxonMobil auto-refrigerated

alkylation process, 280

F

FAU zeolites

and rare earth exchange, 263–264

SINTEF research, 266

Faujasite, 189, 263–264

FEG HRTEM, 221

FEG-STEM, 194

Fluorinated alumina, 270

G

g·b product analysis, 180–181

Glide shear mechanism, 208–210

H

H-BEA

and Brønsted acid sites, 258

Index 351

and EFAL species, 260–261


and silicon/aluminum ratio,

261–262

H-EMT, acid strength and catalyst role

in alkylation, 258

H-FAU, acid strength and catalyst role

in alkylation, 258

H-SAPO-37, 257

H2O2, and stabilization of Ti(O2)

complex, 34

H2O2

anhydrous source, 82

as catalyst in homogeneous phase,

58–60

conversion using Ti-SBA-15 and

Ti-MMM, 97t

as oxidant, 56–57

and oxo-titanium species, 33

replacement in TS-1, 8

HAADF, 193f, 194

combined with BSE, 215

STEM-HAADF image, 217f

and topography of nanoparticles,

213–217

Haldor Topsøe FBATM process,

287–288

Heterolytic catalysis, 58, 137–138

Heteropolyacids, 268–269

High-angle annular dark-field

(HAADF) miscroscopy,

193f–194

High-resolution STEM (HRSTEM),

193–194

High-voltage EM, 197

HMCM-22, 264

Homolytic catalysis, 58, 137–138

HRLVSEM, 195

HRTEM (high-resolution transmission

EM), 176–177

aberration-corrected, 222

description, 181–182

development, 184–185

germanium silicate, 184f

intergrowths in zeolites, 188–190

L-type zeolite catalysts, 185–186

MAPO-36 microporous catalysts,

186–189

of MAPO catalysts, 192

nanopores in silica, 219f

optimizing images, 182–183

sample preparation, 221

samples as weak phase objects

(WPO), 182

SSZ-48 catalysts, 187–188

Hydride transfer in alkylation process,

242–246

and acid strength, 258

from alkenes, 246

energy barrier, 243

of ethane/ethene, 243

gas-phase and liquid-phase,

242–243

potential energy profiles, 244f

reaction enthalpy, 244–245

in zeolites, 243–244

Hydrocracking, 248–249

Hydrofluoric acid (HF)

and alkylation initiation, 237

drawbacks as catalyst, 231, 233,

251

as isomerization catalyst, 241

strength and alkylation product

quality, 254–255

Hydrogen transfer, 249

Hydroperoxo Ti species, 36–37

transformation to superoxo species,

47–49

Hydroxyl-assisted epoxidation,

72–74

Hydroxylation

advantages of mixed-phase catalyst,

88

of aliphatic compounds, 85–89

Index352

aromatic compounds, 89–90

general features, 83–85

I

In situ ETEM, 203

gas–catalyst reactions, 201–203

gas–solid reactions, 196–200

VPO in n-butane, 206f

Inorganic membrane reactor, 306

Ir/Al2O3 catalyst, 305

IR spectroscopy

and irradiation of TS-1 (H2O2

loaded), 38–39

and Lewis acidity in TS-1, 28


Ti-MMM, 96f

Ti-SBA-15, 96f

Isoalkanes, and sulfuric acid as

alkylation catalyst, 251

Isobutane, and hydride transfer, 242

Isobutylene

oligomerization, 248

and self-alkylation, 250

Isobutylene protonation, 238f

Isopentane alkylation, 241

ITQ-7, 265–266

K

Ketones, 102

L

La-promoted catalysts, 324

La0.2Sr0.8Fe0.8Cr0.2Ox membrane,

307–308

La(12x)AxFe0.8Co0.2O32x perovskite

dense membrane reactor, 309

Lanthanide ruthenium oxide, 300

Laser-based spectroscopy, 175

Lewis acid sites

and catalytic action of titanosilicate

molecular sieves, 56

deactivation by water, 130

in TS-1, 28–32

in VPO, 208

in zeolites, 256, 260–261

Lewis basicity, and carbon deposition

in CO2 reforming, 322

Li/MgO, 300

Ligand-to-metal charge transfer

(LMCT)

and coordination sphere expansion

in TS-1, 31–32

in DRUV of oxo-titanium species,

37f

and titanosilicate surface structure,

12

Low-temperature coke, 245

Low-voltage, high resolution SEM,

223

LURGI EUROFUEL, 286–287

M

MAPO-36 microporous catalysts,

186–189

MAPO catalysts

HRTEM (high-resolution trans-

mission EM) analysis, 192

Mass spectrometry, 175

MCM-22, 264–265

MCM-25, 271

MCM-36, 264, 266

MCM-41, 268–269

MCM-48, 220f

MCM-49, 266

Mesoporous aluminosilicates, 266

Mesoporous titanium silicates,

epoxidation reactions, 67–70

Mesoporous TS-1, 64

influence of silylation, 124, 126t

Index 353

Metal-substituted aluminum phosphate

(MAPO-36) microporous


HRTEM (high-resolution

transmission EM) analysis, 192

Methane to synthesis gas. see also CO2

reforming of methane; partial

oxidation of methane

development of processes, 298–301

Gibbs free energy for reactions, 300t

Methyl-tertiary-butyl ether (MTBE),

231

Methylheptanes, 234

Metropolis Monte Carlo method, and

Ti4þ distribution, 53


Michael addition reactions, 110, 113,

117t

MIEC (mixed ionic/electronic

conductors) membrane, 306–311

Molecular sieves

analysis by EM, 221

early investigations, 5

MTBE, 231

Mukaiyama-type aldol reactions,

110, 116t

Mulliken population analysis, and

titanosilicate surface structure, 50

Multiple alkylate, 247

N

n-butenes, and alkylation initiation,

238–239

Nafion-H, 269–270

Nanoparticles

location in nanoporous solids, 218

Nanophase Pt/TiO2 catalysts, 202f

Nanoporous solids, 175

Nanorods, 210–211f

Ni/Al2O3 catalysts, 325–327

Ni/La2O3 catalysts, 328–330

Ni/SiO2 catalysts, 327–328

Ni/ZrO2 catalysts, 330–331

Nickel

catalyst changes during partial

oxidation of methane, 316

catalysts for partial oxidation of

methane, 312–314

particle size and carbon deposition,

322

reduced and oxidized for partial

oxidation of methane, 320–321

and thermal gradients in partial

oxidation of methane, 302, 305

NiO/Al2O3 catalyst, 305

NiO/MgO catalyst, 305

NiO/SiO2 catalyst, 305

Nitrogen-containing compound

oxidation, 90–93

NO decomposition, 121–122

O

O–O bond cleavage

in alkene epoxidation, 138–140

Octane hydroxylation, 88

Octane number, 235–236

for oligomerization and cracking

products, 248

Oligomerization

and alkylation, 247–249

and reaction temperature, 272–273

and strength of acid catalyst, 258

and triflic acid, 255

ONIOM method, and titanosilicate

surface structure, 54–55

OSV (alkene space velocity), 274–275

Oxidation

of alcohols, 100

Baeyer-Villiger (BV) oxidation,

102

cyclohexanone, 104t

Index354

of ethers, 100–101

influence of solvents, 122–124

of n-alkanes, 85t

nitrogen-containing compounds,

90–93

of phenols, 101–102

sulfur-containing compounds,

93–99

of TMP, 103f

Oxidative coupling of methane,

299–300

Oxidative dehydrogenation, 115, 119t

Oxo-titanium species

characteristics on TS-1 with

aqueous H2O2, 143t

concentrations of superoxo and

hydroperoxo species, 43–44, 47

cyclic voltametry, 41–42

EPR spectroscopy, 42–49

formation, 33

free radical oxidation mechanism,

42

H2 þ O2 as oxidant, 42

O–O stretch, 36

paramagnetic oxygen species, 42

peroxide species structure, 38–39

role in epoxidation reactions,

132–133

structure and activity, 128–136

transformation to superoxo species,

47–49

UV-visible spectroscopy, 34–35f

vibrational spectroscopy, 34–39

X-ray absorption spectroscopy,

39–41

Oxygen separation methods, 306–312

P

P/O (paraffin/olefin) ratio, 274–275

Paramagnetic superoxo-titanium

species, 47

Partial dislocations, 181

Partial oxidation of methane

catalyst changes during reaction,

315–316

catalyst composition and tempera-

ture profile, 303


ceramic membrane reactor, 308f,

310f

CHx species formation, 318–320

deuterium isotope effects, 319

development of processes, 299

fluidized-bed reactor, 310–311

hot spots in catalyst beds,

301–306

inorganic membrane process, 306

La0.2Sr0.8Fe0.8Cr0.2Ox membrane,

307–308

lanthanide ruthenium oxide catalyst,

300–301

major challenges, 301

minimizing thermal gradients, 304

Ni/La2O3 catalysts, 328–330

Ni/SiO2 catalysts, 327–328

oxygen purification methods,

306–312

primary product determination,

316–318

rate-determining steps, 318–320

reaction pathways, 301, 314–315

reaction temperatures, 301

with reduced and unreduced


temperature and catalyst

composition, 303

using reduced and oxidized


Pd/C catalyst, 214

Perdeuterioisobutane, 238

Perovskite-based oxygen transport,

306–311

Peroxo-titanium species, 34

Index 355

structure, 52f

Phenol oxidation, 101–102

Photocatalysis and degradation of

pollutants, 116–120

Photocatalytic synthesis, 120–121

Pillared layered silicate, 271

Pinacols, 114, 118t

Platinum-containing catalysts, 323

Polycarbonate precursors, 106

Polyethene terephthalate (PET),

110

Propane, in two-step alkylation,

241

Propene epoxidation, 129–131

Propene oxide, 61–62

Pyridine, and acid strength

measurement, 259

R

Raman spectra

selection rules, 21–22

Ti-MMM, 95t

and TS-1 peroxide species structure,

38–39

Rare earth exchanged faujasites

(REHY, REUSY), 263–264

Rare earth metals, as promoters for

Ni/Al2O3 catalysts, 327

Red oil, 239

Redox potentials of transition metal

ions, 59t

REHX, 276

ReVapTM alkylation process, 282

Rhodium catalysts, 324

RON (research octane number)

for oligomerization and cracking

products, 248

values of alkanes, 235–236t

Ruthenium catalysts, 200

Rutherford detector, 194

S

Sb–SnO2 catalysts, 196

SBA-15 type titanium silicates, 93

Scanning probe methods, 175

Screw dislocations, 181

Self-alkylation, 249–250

Silica nanopores, 219f

Silicalite-1 orthorhombic structure, 11f

Silicon/aluminum ratio, 261–263

Silylation, 124–127

SMSI deactivation, 201

Soft coke, 245

SPARG process, 322, 332

Spectroscopic analysis of titanosilicate

surface structures


and Lewis acidity in TS-1, 28–32

photoluminescence spectroscopy,

15

UV-visible spectroscopy, 12–14



15–18

Sr–Fe–Co–O mixed oxide

membranes, 307–308

SSZ-48 catalysts, 187–188

Steam reforming of methane, 298

combined with partial oxidation,

306

STEM-HAADF image, 217f

STEM (scanning transmission EM),

177, 193–195

aberration-corrected, 222

Stoichiometry, and AEM, 191

Stratco Alkysafe process, 281

Stratco Contactor reactor, 279f

Sulfated zirconia, 233, 267–271

Sulfur compound removal, 98–99

Sulfuric acid (H2SO4)

and alkylation initiation, 237

drawbacks as catalyst, 231, 233, 251

Index356

in industrial processes, 278–281

as isomerization catalysts, 240

and oligomerization, 247


strength and alkylation product

quality, 254

Superoxo titanium, 34

and catalytic activity, 132–133

EPR data, 45t

EPR spectroscopy, 42

transformation from hydroperoxo/

peroxo species, 47–49

Synthesis of titanium silicate

molecular sieves

confined space method, 145

dissolved (hydrolyzed) titanium

method, 144, 148

ETS-10, 154–156

microwave irradiation technique,

145

mixedalkoxidemethod,144,148,151

prehydrolysis method, 144, 149–150

reduced crystallization time, 144

Ti-beta, 153–154

Ti-HMS, 147, 157

Ti-MCM-41, 147, 154–156

Ti-MCM-48, 147, 157

Ti-SBA-15, 147, 158

Ti-ZSM-48, 152

TS-2, 151–152

using TiF4, 150

wetness impregnation method, 144,

148

T

Tar formation, 101

TEM

diffraction contrast technique, 180

energy-filtered, 218–219

imaging methods, 179–181

sample damage, 216

Temperature

and alkylation reaction, 272–274

and partial oxidation of methane,

301, 303

Thioanisole (MPS) oxidation, 97t

Ti-beta

and alcohol oxidation, 100

diffusional constraints, 62, 63t

relative selectivity, 65–67

synthesis, 146–147, 153–154

Ti-HMS synthesis, 147, 157

Ti-MCM-41, 67

catalytic activity, 69t

catalytic selectivity and Ti content,

129f

influence of silylation, 124

photocatalytic synthesis, 120–121

structure and activity, 128

synthesis, 147, 154–156

Ti composition and textural

characteristics, 68t

and transesterification

reactions, 110

Ti-MCM-48, 124

synthesis, 147, 157

Ti-MMM-1, 88

and H2O2 conversion, 97t

IR spectroscopy, 96f

Raman spectra, 95t

synthesis, 146

Ti-MWW synthesis, 145

Ti-SBA-15, 93

and H2O2 conversion, 97t

IR spectroscopy, 96f

Raman spectra, 95t

structural and textural parameters,

94t

synthesis, 147–148, 158

UV-visible spectroscopy, 95t

Ti-ZSM-48 synthesis, 145, 152

Ti4þ

coordination number, 50

Index 357

distribution in TS-1, 53

Time on stream

behavior of BEA, 265

behavior of CeY zeolite, 259f

and catalyst temperature in partial

oxidation of methane, 304f

for CO formation in CO2 reforming

of methane, 328f–329f

for partial oxidation of methane,

303f

Ti(O2H) activity, 128–129

Titanium peroxo species, 8

Titanium silicate molecular sieves

acid-catalyzed reactions, 105–116

active sites, 6–7, 9–33

CvN cleavage reactions, 105

catalytic properties, 56, 132–136

commercial application, 7, 62

computational investigations,

49–55

early investigations, 5–6

epoxidation. see Epoxidation

hydroxylations. see Hydroxylations

neutron diffraction, 10–11

NO decomposition, 121–122

O–O bond cleavage, 137–138

oxidation reactions. see Oxidation

photodegradation of pollutants,

116–120

silylation, 124–127

structure and activity, 127–128

synthesis, 143–146

Ti composition and textural

characteristics, 68t

Titanium superoxo species, 8

Titanosilicate surface structures


particle size, 12

photoluminescence spectroscopy,

15

Ti tetrahedral geometry, 9

UV-visible spectroscopy, 12–14



15–18

X-ray diffraction, 10

TMP (trimethylpentane), 101, 234

isomerization pathways, 241f

oligomerization, 248

Toluene oxidation, 89–90

Tomography, 212–218

using HAADF-STEM, 223

Transesterification

of cyclic carbonates, 109

of diethyl malonate, 113t

of esters, 110

of ethylacetoacetate, 111–112, 114t

Triflic acid

as alkylation catalyst, 271

and oligomerization, 255

Trimethylpentanes. see TMP

(trimethylpentanes)

TS-1 (titanium silicate-1)

Brønsted acid sites, 26–28

chemoselectivity, 7

coordination sphere expansion,

31–32

diffusional constraints, 64

discovery, 5

DRUV data, 13t–14f

fingerprint features for Ti

isomorphous substitution, 142t

and oxidation of amines, 91t

photocatalytic synthesis, 120–121

Raman spectra, 20f

relative selectivity, 65–67

SEM photographs, 136f

and transesterification reactions, 110

XANES spectrum, 17f

TS-2 (titanium silicate-2)

aliphatic compound hydroxylation,

87

epoxidation of alkenes, 70t

oxidation of sulfides, 93t

Index358

synthesis, 145, 151–152

and toluene oxidation, 89–90

U

Ultra-high resolution low-voltage field

emission scanning electron

microscopy (HRLVSEM), 195

UOP Alkad process, 283

UOP AlkyleneTM process, 285–286

USY zeolite

and alkene feed composition, 276

cracking, 249

and reaction temperature, 273

UV-visible spectroscopy

and adsorbed water on TS-1, 32–33f

and Lewis acidity in TS-1, 31–32

oxo-titanium species, 34–35f

and surface structure of TS-1, 12–

14

for TiSBA-15, 95t

V

Vanadium oxide, 324

Vanadyl pyrophosphate (VO)2P2O7,

203

active sites, 207f

electron diffraction, 205f

glide shear plane defects, 208–209

structure, 204f

Vibrational spectroscopy,


VPO, 203

active sites, 207f

electron diffraction, 205f

solid-state heterogeneous catalytic

oxidation processes, 208–209

structure, 204f

VS-2

n-haxane hydroxylation, 87

and toluene oxidation, 89–90

VTM-A, 266

W

Wavelength dispersive X-ray

spectroscopy (WDS), 191

Weak phase objects (WPO), 182–183

Wet-ETEM, 210–212

WS2 catalyst, 331–332

X

X-ray absorption spectroscopy, 39–41

X-ray elemental mapping, 192

XAFS (X-ray absorption fine

structure), 176

XANES spectrum

and adsorbed water on TS-1, 33f


Z

Z-contrast imaging, 194

Zeolite X, 262–263

Zeolite Y, 263–264

and CO2 reforming of methane, 332

Zeolites

acid strength and catalyst role in

alkylation, 258

advantages in alkylation, 233

and alkene feed composition, 276

and alkylation initiation, 237, 239

analysis by EM, 221

Brønsted acid sites, 256–260

characteristics as molecular sieve,

255–256

electron microscopic image, 194f

enthalpy of adsorption, 257f

HRTEM analysis of L-type,

185–186

and hydride transfer, 243–244

intergrowth analysis by HRTEM,

188–190

as isomerization catalysts, 240

Index 359

Lewis acid sites, 260–261

and reaction temperature, 273


structure and types, 264–267

vibrational spectroscopy, 18

Ziegler-Natta catalysts, 200

Zirconia-based oxygen transport, 306

Zirconia-containing catalysts, 323

Zirconia-supported nickel catalyst,

330–331

ZSM-11, 264

ZSM-12, 264

ZSM-20, and rare earth exchange,

263–264

ZSM-5, 264–265

ZSM-5 zeolite, 332

Index360

advances in catalysis, volume 48

Documents