Catalysis: A Review of Chemical Literature: Vol 16 (Specialist Periodical Reports)

Cat a Iysi s Volume 16

Catalysts & Catalysed Reac ReadershipAcademics and lndust involved in the design, development application of catalysts

Effective, rapid, currentaw Visual, graphical abstracts new developments Chemical structures, reacti conditionsand catalytic re! Monthly updates 200+ abstractsper issue sel Comprehensive indexing b; Searchableonline via thew Free site-wideonline acces! Catalysts 81 Catalysed Reactions is a m; providing graphical abstracts of new deve 100 primary journals. Coverage includes homogeneous, heterogeneous and biocai areas such as chiral catalysts, polymeris: clean catalytic methods. The multidiscipli Reactions encompasses research from organic and physical chemistry, and in bioc Abstracts are selected by a respected intt Catalysts & CatalysedReactionsis a sing research developments in catalysis.

12 issues Print (includes FREE site-wide online lSSN 1474-9173 €3gs.oo/SsgG.oo

For further de

WWW.'

A Specialist Periodical Report

Catalysis Volume 16 A Review of Recent Literature Senior Reporter James J. Spivey, Department of Chemical Engineering, North Carolina State University, Raleigh, North Carolina, USA

Reporters J. Agrell, Royal Institute of Technology, Stockholm, Sweden M. Baems, Institute for Applied Chemistry Berlin-Adlershof, Berlin, Germany 0. Buyevskaya, Institute for Applied Chemistry Berlin-Adlershof, Berlin, Germany M. Faghihi, Vaxjo University, Vaxjo, Sweden B. Gevert, Chalmers University o f Technology, Goteborg, Sweden S.-W. Ham, Kyungil University, Kyungsan, Korea B. Harrysson, Nynas Naphthenics AB, Nynashamn, Sweden W. Hoelderich, University of Technology, Aachen, Germany T.Inui, Air Water Inc., Osaka, Japan S. Jaris, Royal Institute of Technology, Stockholm, Sweden F. Kollmer, University of Technology,Aachen, Germany S.W. Lin, Tokyo Institute of Technology, Yokohama, Japan B. Lindstrorn, Royal Institute of Technology, Stockholm, Sweden 1 . 4 . Nam, Pohang University of Science and Technology, Pohang, Korea L.J. Pettersson, Royal Institute of Technology, Stockholm, Sweden M. Sanati, Vaxjo University, Vaxjo, Sweden C. Song, Pennsylvania State University, University Park, Pennsylvania, USA W. Ueda, Hokkaido University, Sapporo, Japan

RSmC ROYAL SOCIETY OF CHEMISIRY

ISSN 0-85404-224-5 ISBN 0140-0568

0The Royal Society of Chemistry 2002 All rights reserved Apart from any fair dealing for the purpose of research or private study, or criticism or review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988, this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission in writing of The Royal Society of Chemistry, or in the case of reprographic reproduction only in accordance with the terms of the licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of the licences issued by the appropriate Reproduction Rights Organization outside the UK Enquiries concerning reproduction outside the terms stated here should be sent to The Royal Society of Chemistry at the addressprinted on this page. Published by The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 OWF, UK Registered Charity Number 207890 For further information see our web site at www.rsc.org Typeset by Computape (Pickering) Ltd, Pickering, North Yorkshire, UK Printed and bound by Athenaeum Press Ltd, Gateshead, Tyne &Wear

Catalysis research continues to address both the traditional areas of chemical synthesis and fuel production, but with increased emphasis on environmentally benign processes and products. This volume reflects the interest in both. The reviews provided here include hydrogen production, upgrading of fuels to minimize emissions, as well as new catalysts and processes for chemical synthesis - each with attention to the environmental impact of catalyst activity and selectivity. Mehri Sanati and Mostafa Faghihi (Vaxjo University, Sweden), Bjorn Harrysson (Nynas Naphthenics AB, Stockholm, Sweden), Bjorje Gevert (Chalmers Gothenburg, Sweden), and Sven Jaris (KTH, Stockholm, Sweden) provide a review of hydrodearomatization, which is important in fuel processing. This reaction is typically carried out with other hydrotreating reactions (e.g . , hydrodesulfurization). Hydrodearomatization is essential in order to improve the fuel quality and minimize undesirable emissions. One key area for research is sulfur tolerance, particularly for the noble metal-based catalysts. The authors explore the reactivity of a wide range of catalysts, and address the hydrodearomatization of mono- and multi-ring aromatic compounds. Wolfgang Hoelderich and Felix Kollmer (University of Technology, RWTH, Germany) examine catalytic oxidations for fine chemical synthesis, which typically use oxidizing agents other than dioxygen, such as N20 and H202. The authors point out the important differences between oxidation reactions (and catalysts) needed to produce bulk chemicals, compared to those needed to produce fine chemicals: operating temperature, reactor design. In another review from KTH, Johan Agrell, Bird Lindstrom, Lars Pettersson, and Sven Jaris review the catalytic generation of hydrogen from methanol. This is one of the options for generation of hydrogen from liquid fuels for stationary and mobile applications, fuel processing, and fuel cells. Methanol can be used to generate hydrogen by several reactions: decomposition to CO and hydrogen, steam reforming, and partial oxidation (as well as combinations of these reactions). The authors explore the different catalysts needed for each of these reactions, as well as recent industrial research activity. Tomoyuki Inui (Air Water Inc., Osaka, Japan) also addresses reforming of hydrocarbons for syngas and hydrogen production. His review focuses on the reaction of methane with CO2, oxygen and/or steam. There are significant research needs in this well-studied area, including coke formation and reactor desigdkinetics. Both noble metal and conventional Ni-based catalysts are reviewed, as well as new synthesis techniques. Olga Buyevskaya and Manfred Baerns (Institute for Applied Chemistry V

vi

Preface

Berlin-Adlershof, Germany) provide a thorough review of the oxidative functionalization of ethane and propane. This review concentrates on the oxidative dehydrogenation of these two compounds to produce the corresponding olefins. In addition, the selective oxidation of ethane and propane to produce directly higher value oxygenates such as acetic acid, acrolein and acrylic acid is addressed. Wataru Ueda (formerly at Science University of Tokyo in Yamaguchi; now at Hokkaido University, Sapporo, Japan) and Sui Wen Lin (Tokyo Institute of Technology, Japan) also address selective oxidation of lower alkanes epoxidation, coupling and dehydrogenation reactions, for example. Their review focuses on the use of metal halide catalysts for these reactions. These catalysts can increase the activity and selectivity of selective oxidation reactions. Particular emphasis is given to the layered metal chloride catalysts for selective oxidations. Sung-Won Ham (Kyungil University, Korea) and In-Sik Nam (Pohang University of Science and Technology, Korea) discuss the selective catalytic reduction of NO, using ammonia over conventional vanadium-based catalysts, zeolite catalysts, and exploratory catalysts based on novel titania formulations and pillared clays. A comprehensive review of poisoning and deactivation is also provided, along with a study of the mechanisdkinetics and reactor modeling. Finally, Chunshan Song (Penn State University, USA) reviews the conversion of polycyclic hydrocarbons into specialty chemicals over zeolites. Until recently, the conversion of these compounds into useful products was not widely studied. These compounds are used in advanced polymers, but the large-scale production of the corresponding monomers requires selective catalysts for the conversion of polycyclic aromatic hydrocarbons. This review focuses on the various zeolites that can be modified to produce these compounds. I am working with the authors of Volume 17 to prepare reviews of topics that are at the leading edge of catalysis research. I look forward to bringing this next volume to you. As always, comments are welcome. James J. Spivey Department of Chemical Engineering NC State University Raleigh, NC 27695 USA jjspivey @ncsu.edu

Contents

Chapter 1 Catalytic Hydrodearomatization By Mehri Sanati, Bjorn Harrysson, Mostafa Faghihi, Borje Gevert and Sven Jar&

1

1

1 Introduction 2 Hydrogenation of Mono-, Di-, Tri-, Multiring and Mixtures of Aromatic Compounds 2.1 Reactivity in Hydrogenation Reactions 2.2 Thermodynamics of Hydrogenation 2.3 Reaction Pathways and Kinetics of Hydrogenation 2.3.1 Reaction Network 2.3.2 Reaction Order 2.3.3 Activation Energies 2.4 Catalysts and Nature of Catalytic Sites

3 27 30 31 31 31 31 32

3 Industrial Aspects

37

4 Summary and Conclusion

37

5 References

38

Chapter 2 Heterogeneously Catalysed Oxidations for the Environmentally Friendly Synthesis of Fine and Intermediate Chemicals: Synergy between Catalyst Development and Reaction Engineering By Wolfgang I;: Hoelderich and Felix Kollmer

43

1 Introduction

43

2 Activation of the Oxidant by Noble Metal: BASF Citral Process

45

3 Activation of the Oxidant by Metal Oxide: Direct Oxidation of P-Picoline

46

Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 vii

...

Contents

Vlll

4 Activation of the Oxidant by Lewis Acid Sites: Direct Hydroxylation of Benzene

49

5 Activation of the Oxidant by H2 over a Bifunctional Catalyst: Epoxidation of Propylene by a Mixture of H2 and 0 2

52

6 Activation of the Oxidant by Immobilized Homogeneous Complexes: Stereoselective Epoxidation of Olefins

57

7 Activation of H202 over Solid Acid Catalysts: Baeyer-Villiger Oxidation of Cyclopentanone

61

8 Summary

63

9 Acknowledgements

64

10 References Chapter 3 Catalytic Hydrogen Generation from Methanol By Johan Agrell, Bird Lindstrom, Lars J. Pettersson and Sven G. Jaris

64 67

1 Introduction 1.1 Fuel Options for Hydrogen Generation 1.1.1 Methanol Synthesis 1.2 Methanol Conversion Processes 1.2.1 Decomposition 1.2.2 Steam Reforming 1.2.3 Partial Oxidation 1.2.4 Combined Reforming

67 68 68 69 69 69 70 71

2 Decomposition of Methanol 2.1 Introduction 2.2 Copper-based Catalysts 2.2.1 Reaction Pathway and By-product Formation 2.2.2 Reaction Mechanism 2.2.3 Promotional Effects 2.2.4 Catalyst Stability 2.2.5 Active Species 2.2.6 Calcination Temperature and Catalyst Copper Content 2.3 Precious Metal Catalysts 2.3.1 Catalyst Activity and Product Distribution 2.3.2 The Support 2.3.3 Reaction Pathway and Mechanism

71 71 75 75 76 78 79 79 80 80 80 81 83

ix

Contents

2.4 Other Catalyst Materials 2.4.1 Nickel-based Catalysts 2.4.2 Hydrogen Storage Alloys 2.4.3 Other Metal Oxides

84 84 84 84

3 Steam Reforming of Methanol 3.1 Introduction 3.2 Copper-based Catalysts 3.2.1 Reaction Pathway 3.2.2 By-product Formation 3.2.3 Steam-Methanol Ratio and Water-Gas Shift 3.2.4 Reaction Mechanism 3.2.5 Kinetic Models 3.2.6 Catalyst Composition 3.3 Other Transition Metal-based Catalysts 3.3.1 Reaction Pathway 3.3.2 Catalyst Composition

84 84 87 87 88 89 89 90 91 95 95 95

4 Partial Oxidation of Methanol 4.1 Introduction 4.2 Copper-based Catalysts 4.2.1 Reaction Pathway 4.2.2 By-product Formation 4.2.3 Reaction Mechanism 4.2.4 Reaction Temperature 4.2.5 Oxygen Partial Pressure 4.2.6 Catalyst Composition 4.2.7 Catalyst Stability 4.2.8 The Oxidation State of Copper 4.2.9 Cu-ZnO Synergistic Effects 4.3 Palladium Catalysts 4.3.1 Catalytic Behaviour 4.3.2 PdZn Alloy 4.3.3 The Support 4.4 Combined Reforming over Copper-based Catalysts 4.4.1 Background 4.4.2 Overall Reaction 4.4.3 Oxygen and Steam Partial Pressures 4.4.4 Reaction Temperature 4.4.5 Catalyst Stability and the Oxidation State of Copper 4.4.6 By-product Formation 4.4.7 Kinetics

96 96 99 99 100 100 101 101 102 103 103 103 104 104 105 105 105 105 106 106 107

5 Methanol Reforming for Fuel Cell Applications 5.1 Introduction

109 109

107 107 107

Contents

X

5.2 Partial Oxidation verms Steam Reforming Technology 5.2.1 Steam Reforming 5.2.2 Partial Oxidation 5.2.3 Combined Reforming 5.3 CO Clean-up 5.3.1 Water-Gas Shift 5.3.2 Selective CO Oxidation 5.3.3 Methanation 5.3.4 Palladium Membranes 5.4 Industrial Activities 5.4.1 Argonne National Laboratory 5.4.2 Arthur D Little 5.4.3 Ballard 5.4.4 DaimlerChrysler 5.4.5 Ford 5.4.6 General Motors (GM) 5.4.7 Haldor Topsrae 5.4.8 International Fuel Cell Corporation (IFC) 5.4.9 Johnson Matthey 5.4.10 Toshiba 5.4.1 1 Toyota 5.4.12 Volkswagen (vw)-Siid Chemie 5.4.13 Volvo 5.4.14 Wellman-ECN 5.4.15 Case Study I: The Johnson Matthey HotSpotTMReactor 5.4.16 Case Study 11: The Argonne APOR Fuel Processor

109 109 110 111 111 112 112 112 113 113 114 114 115 115 115 116 116 116 117 117 117 117 118 118 118 119

6 Intellectual Property

120

7 Concluding Remarks

126

8 Acknowledgements

126

9 References

127

Chapter 4 Reforming of CH4 by C02,02 and/or H20 By Tomoyuki Inui 1 Introduction 1.1 Recent Research 1.2 Significanceof the Subject and Disadvantage to be Overcome 1.3 Indispensable Requisites of Catalytic Technologies for Moderating CO, Problems and/or Producing: Svneas

133 133 133 134 135

Contents

xi

2 Comparison of the Intrinsic Activity of Ni and Noble Metals Belonging to Groups 8-10

135

3 Performance of Ni Catalyst Supported on Different Kinds of Support and the Effect of Additives

137

4 Prevention of Coke Formation by Combining Alkaline-earth or Rare-earth Oxides

139

5 Attempts to Avoid Equilibrium-limited Conversion

141

6 Kinetic and Mechanistic Studies

141

7 High Ni-loading Catalyst Supported on ZrO2 Prepared by a Novel Method

142

8 Towards Rapid Reforming of Methane - Catalytic Partial Oxidation of Methane to Syngas

143

9 Perfomance of Ni-based Composite Catalyst to Carry Out the Ultra-rapid C02 Reforming of Methane

144

10 Application to Other Reactions

148

11 Summary of the Review

148

12 Acknowledgement

150

13 References

150

Chapter 5 Oxidative Functionalization of Ethane and Propane By Olga I/. Buyevskaya and Manfred Baerns

155

1 Introduction

155

2 Mechanisms of Ethane and Propane Activation

155

3 Oxidative Dehydrogenation of Ethane 3.1 Catalyst Types and Reactor Concepts 3.1.1 (Mixed) Oxides or Supported Catalytic Materials Consisting of Reducible Metal Oxides 3.1.2 Non-reducible (Mixed) Oxides 3.1.3 Catalytic Materials Containing Noble Metals 3.2 Ethane Dehydrogenation in the Presence of COZ 3.3 Summary on ODE

156 158 159 161 163 166 167

Contents

xii

4 Oxidative Dehydrogenation of Propane 4.1 Mixed Oxides or Supported Catalytic Materials Consisting of Reducible Metal Oxides 4.1.1 Catalyst Development 4.1.2 Mechanistic and Characterization Studies 4.2 Non-reducible (Mixed) Oxides 4.2.1 REO-based Catalysts 4.2.2 Boron-containing Oxide Catalysts 4.3 Catalytic Materials Containing Noble Metals 4.4 Reactor Concepts and Modes of Operation

168

5 Partial Oxidation of Ethane and Propane to Oxygenates 5.1 Ethane to Acetic Acid 5.1.1 Catalysts 5.1.2 Fundamentals and Kinetics 5.1.3 Reactor Concepts 5.2 Propane to Acrylic Acid and Acrolein 5.2.1 Direct Oxidation of Propane to Acrylic Acid 5.2.2 Direct Oxidation of Propane to Acrolein

180 180 180 181 185 186 186 190

6 Other Reactions

192

7 Summary and Outlook

193

8 Acknowledgements

193

9 References

193

Chapter 6 Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation By Wataru Ueda and Sui Wen Lin

168 168 172 176 176 177 178 178

198

1 Introduction

198

2 Gas-phase Seletive Oxidation in Halogen-containing Systems 2.1 Epoxidation 2.2 Redox System 2.3 Gas-phase Halogen Promoters 2.4 Oxidative Reaction of Alkanes 3 Catalytic Properties of Halogen-containing Metal Oxides 3.1 Oxidative Coupling of Methane 3.2 Oxidative Dehydrogenation of Ethane 3.3 Oxidative Dehydrogenation of C3+

200 200 202 202 203 204 204 206 209

4 Structural Design of Layered Metal Chloride Oxide Catalysts

210

...

Contents

xi11

4.1 4.2 4.3 4.4 4.5

Structural Background Layer Structure-Activity Relationship Structural Tuning in Local Chlorine Environment Structural Tuning of Metal Oxide Layer Halogen-involved Alkane Activation Mechanism

21 1 217 220 226 229

5 Conclusions

230

6 Acknowledgement

230

7 References

231

Chapter 7 Selective Catalytic Reduction of Nitrogen Oxides by Ammonia By Sung- Won Ham and In-Sik Nam

236

1 Introduction

236

2 Reactions Involved

238

3 SCR Catalysts 3.1 V2Os-based Catalysts 3.2 Zeolite Catalysts 3.3 Newly Proposed Catalysts

239 239 243 244

4 Catalyst Poisoning and Deactivation 4.1 Sulfur Tolerance of SCR Catalysts 4.2 SO2 Oxidation

247 248 250

5 Reaction Mechanism and Kinetics

252

6 Characteristics of an SCR Catalyst 6.1 Surface Acidity 6.2 State of the Vanadia Phase

254 254 255

7 Reactor Modeling and Process Characteristics 7.1 Low Pressure-drop Reactor 7.2 SCR Process Characteristics

258 258 26 1

8 Alternative Technology to NH3-SCR

262

9 Conclusions and Future Prospects 10 References

264 265

xiv

Contents

Chapter 8 Recent Advances in Selective Conversion of Polycyclic Hydrocarbons into Specialty Chemicals over Zeolites By Chunshan Song

272

1 Introduction

272

2 Selective Synthesis for Polycyclic Specialty Chemicals 2.1 Ring-shift Isomerization 2.2 Shape-selective Alkylation of Naphthalene 2.3 Shape-selective Alkylation of Biphenyl 2.4 Conformational Isomerization 2.5 Selective Hydrogenation of Naphthalene 2.6 Regio-selective Hydrogenation of Quinoline and Naphthol

275 276 279 287 288 29 1

3 Effects of Mordenite Dealumination on Naphthalene Isopropylation 3.1 Characterization of the Mordenite Catalysts 3.2 Selectivity and Activity of the Catalysts 3.3 Structure-Selectivity Relationships 3.4 Effect of Reaction Conditions and Catalyst Loading 3.5 Effects of Mordenite Dealumination on Further Conversion of 2,6-DIPN 3.6 Conformational Analysis of Shape-selective Isopropylation 3.7 Factors Affecting Shape-selective Naphthalene Alkylation

292 293 294 299 303 304 307 310 312

4 Considerations on Environmentally Friendly Synthesis of Chemicals over Zeolites

315

5 Conclusions

317

6 Acknowledgements

318

7 References

319

1 Catalytic Hydrodearomatization BY MEHRI SANATI, BJORN HARRYSSON, MOSTAFA FAGHIHI, BORJE GEVERT, AND SVEN JARAS

1

Introduction

Hydroprocessing of various feeds for the production of fuels is extensively practised in the petroleum industry, and to some extent in coal liquefaction and in the upgrading of synthetic fuels and lubricant oils. Another promising area where hydroprocessing can be applied is the development of renewable non-fossil fuels (pyrolitic bio-oil) for the elimination of the oxygen-containing molecules and the improvement of the H/C ratio. Hydroprocessing reactions occur on the active sites of the catalysts. Also, a suitable pore size distribution of the catalysts is required to ensure the access of reactant molecules to the active sites. The catalysts used in hydroprocessing consist of a molybdenum catalyst that is supported on a high surface area carrier in the 100-300 m2/g range, most commonly alumina, and is promoted by either cobalt or nickel. The concentration by weight of the metal is usually 1-4% for Co and Ni, and 8-16% for Mo.' The catalysts are active in the sulfided state, being either presulfided or sulfided on stream with a sulfur containing feed. Monometallic (Pt or Pd) and bimetallic (Pt-Pd) catalysts of noble metal supported on y-A1203are known to be highly active in the hydrogenation of aromatics under mild conditions. However, noble metal catalysts are easily poisoned by a small amount of sulfur; severe pre-treatment of the feedstock is needed to reduce sulfur to a few ppm. Recent studies have dealt with how to improve the activity of these catalysts and their sulfur tolerance, e.g. by adding a second transition metal or using different support The typical feedstock for laboratory tests is usually either a mixture of model mono-compounds and/or a mixture of different aromatic hydrocarbons. 12-13 In industrial feeds, however, several types of aromatics are present, whose hydrogenation activities differ considerably. The composition and concentration of various nitrogen and sulfur compounds also significantly influence the activity. The process is normally carried out in a trickle-bed reactor at an elevated temperature and hydrogen pressure. In the case of severe deactivation, an ebullating bed reactor might be used but this type of reactor is not suitable due to back-mixing when a high conversion is needed. The specific characteristic of a trickle bed reactor is that a part of the catalytic surface is covered by a liquid Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 1

2

Catalysis

and the other part by a gas. In the common set up, the liquid phase flows downwards through the reactor concurrently with a gas phase that partly consists of vaporized compounds. The temperature and pressure ranges for the hydrogenation of aromatic hydrocarbons in a liquid phase batch reactor were reported to be 450-700 K and 3.5-17 MPa, re~pectively.'~-~~~*~-'~ Hydroprocessing catalysts are quite versatile, exhibiting activity for a number of important reactions. Those of major interest in hydroprocessing that might be referred to as hydrorefining correspond to removal of heteroatoms; hydrodesulfurization (HDS), hydrometallization (HDM), hydrodenitrogenation (HDN) and hydrodeoxygenation (HDO). These reactions involve hydrogenolysis of C-heteroatom bonds. The removal of sulfur and nitrogen is necessary to meet environmental limits. Sulfur may also cause problems with catalyst poisoning and corrosion. HDN is needed to avoid catalyst poisoning of acid sites and improving stability in lube oils. An important reaction in petrochemical industry and refineries is hydroconversion, which enables a change in the molecular weight and structure of organic molecules. Examples are hydrogenation (HYD) and hydrodearomatization (HDA). When oil is hydrotreated, the reduction of aromatic compounds competes with the removal of sulfur and nitrogen. The purpose of hydrotreating (in the latter sense) is to improve the stability and quality of the product. The reduction of aromatic compounds, especially polyaromatics, gives a higher stability to the product, as well as affecting the solubility and colour of the product. Aromatics in fuels not only lower the quality and produce undesired exhaust emissions, they also have potential hazardous and carcinogenic effects.26 Thus polyaromatic compounds are removed to meet health and environmental regulations. The growing understanding of health hazard associated with these emissions is leading to limitation in the use of aromatics in both Europe and the United States.27 The process to make cleaner fuels that are more environmentally friendly is often accompanied by desulfurization and hydrodearomatization. Decreasing the aromatic content increases the cetane number in diesel fuel. Two approaches, a single-stage process and a two-stage process, have been proposed for distillate fuels (particularly diesel fuels) to meet these strict standards for diesel fuels. The single-stage process combines severe hydrodesulfurization and hydrogenation using a single conventional sulfided CoMo, NiMo or NiW catalyst. In order to reach the necessary aromatic saturation the H2 pressure needs to be substantially higher than the H2 pressure at which current hydrodesulfurization units operate.28 The two-stage system uses a conventional hydrotreating catalyst in the first reactor and a noble metal catalyst in the second; this yields a low aromatic diesel stream at moderate hydrogen p r e s s ~ r e . ~ ~ ~ ~ ~ This latter system is highly active for the reduction of aromatics but is very susceptible to sulfur poisoning; the sulfur concentration at the inlet of the second reactor must be reduced to a few parts per million.31 Thus, the use of these catalysts depends strongly on severe pre-treatment conditions, unless the sulfur tolerance can be greatly improved for the noble

1: Catalytic Hydrodearomatization

3

metal catalyst. A number of recent studies have attempted to address this problem by developing catalysts with a high resistance to the sulfur poisoning and at the same time retaining a high hydrogenation activity. In spite of the large number of articles published in recent years, the subject has been widely reviewed. The catalytic aspects of the hydrogenation were discussed by Krylov and N a ~ a l i k h i n a Special .~~ attention to the preparation methods was discussed in more detail by P. Grange and X. Vanhaeren.27 A comprehensive review of the hydrodeoxygenation, with particular focus on upgrading of bio-oils, was published by F u r i m ~ k y .Catalyst ~~ deactivation during hydroprocessing, including the adverse effects of the 0-compounds, was reviewed by Furimsky and M a s ~ o t h . ~ ~ In this review, the primary focus is on the most recently reported work in the literature for both basic and industrial aspects of hydrodearomatization reactions. It is an extension and update of recent studies dealing with the aromatic reduction in different petrochemical feedstocks. These reviews, which have recently appeared in literature, provide comprehensive information regarding hydrodear~matization.~~. 35-37 A comprehensive review of the reactions during hydroprocessing has been published by Topsoe et al.

2

Hydrogenation of Mono-, Di-, Tri-, Multiring and Mixtures of Aromatic Compounds

In recent years an increasing awareness of the use of aromatics contained in different feedstocks, especially distillate fuel (in particular diesel and gas oils), with respect to the adverse effects of undesired emissions and potential health risks, has received considerable attention. In addition, a high aromatic content is associated with poor fuel quality, giving low cetane number in diesel fuel and a high smoke point in jet fuel. To date, a number of the model compounds that are representative of components in industrial feeds, have been extensively studied on several catalysts. These include both unsupported and y-A1203 supported hydrogenation catalysts, using the conventional CoMo, NiMo, NiW, and platinium group metals (including ruthenium, rhodium, palladium and platinum). On all catalysts, the rate of hydrogenation generally increases with the number of aromatic rings present, i.e. a low rate of hydrogenation is observed for monoaromatic rings such as benzene.' The greater reactivity for hydrogenation with higher fused ring systems, such as naphthalene and anthracene, is due to the fact that the resonance energy of the second ring of these multiple compounds is less than for benzene.35 Table 1 of this review shows the recent related publications on hydrodearomatization and the catalytic systems, reaction conditions and product selectivities for these studies. The choice of model compounds were often the mono-aromatics compounds or a mixture of the aromatics in order to simulate a composition similar to the industrial feedstock in refinery.

T = 195-250 "C,P = 5 MPa, WHSV Toluene and 4.5 h-l, aromatic reactivity follows: naphthalene mixture (20% Mono and 2% naphthalene >> toluene > tetralin. This order was related to a decrease in Di) and the amount of the resonance energy per atomic ring, sulfur was varied from difference in the .n-electron cloud 0.065 to 0.070 (DBT). density in the aromatic ring, as well as inductive effect of the methyl group

2000 PtPd/Si02-A1203, Pt/Si02-A1203 and Pd/Si02-A1203

Oil fraction containing 32-40 vol YOaromatics (20.5 Mono, 17.3 Di, 2.2 Tri) and 172-474 ppm sulfur

T = 300 "C,P = 4.9 MPa and LHSV 1.5 h-l, the long-term stability test under industrial operating conditions demonstrated the excellent stability of the catalyst

PtPd/Si02-A1203catalyst shows the highest activity for the hydrogenation of mono- and di- in the presence of S, Pt/ Si02-A1203has the lowest activity. The excellent activity of PtPd catalyst was attributed to the electron - deficient platinum species (site isolation) of isolated Pt cluster on the Pd surface

Enhancement of catalytic activity, which depended on the Pd/(Pt+Pd) weight ratio and reached a maximum of aromatic hydrogenation of about 0.7 Pd/(Pt+Pd) weight ratio. Hydrated mono, di, tri was about 30,4 and 3 vol YO,respectively

4

3

2

Radioacti~e~~S-labeled DBT was used to measure of catalytic HDY activity of phenanthrene, which was monitored by a change in unreacted [3sS]DBTto form [3sS]H2S

Phenanthrene + 1 wt YOdibenzothiophene (DBT)

RefW

38

Products I remarks

Recycle solvent in coal The tetra-aromatic, for example tetrahydronaphthalene, formed much liquefaction, more than decahydronaphthalenes naphthalenes: 15.8, phenanthrenes 3.25 anthracene 2.3, pyrene 3.9 (wt %)

Substrate

2000 Pt-Pd/Si02-Al2O3

2000 Ni-MolAlZO3, support T = 290 to 330 "C,150 t/d pilot plant, y-alumina cylindrical solvent dimethyl disulfide continuously grain added at a conc. 0.5% wt. The hydrogenation rate increases in the order naphthalenes > phenanthrenes > anthracenes > pyrenes 2000 y-alumina-supported T = 260 and 280 "C,P = 5 MPa, Pt-Pd HYD activities over the bimetallic catalyst (Pt-Pd) were higher than that of the mono-metalliccatalysts together

Reaction conditions I remarks

Hydrogenation of polyarornatics

Year Catalyst

Table 1

P

Cyclohexane and methylcyclopentane isomer, a total selectivity to CH+MCP above go%, the concentration of MCP increased withreaction temperature for both catalysts

T = 200-300 "C, P = 3.0 MPa, Pt/Beta has a higher sulfur resistance than Pt/ WOx-ZrO2. The lower sulfur resistance of P W Z r could be explained by a strong interaction of part of the Pt with surface W6+

7

EXAFS (X-ray absorptionfine structure) for characterisation of the catalysts, structure, interaction between Pt and Pd in the Pt-Pd/Si02-Al,03 catalyst. The amount of Pt (0.5 wtY0)-Pd/SiO~-A1~0~ was sufficient for obtaining optimum catalytic performance. Regarding the activities of aromatic hydrogenation, it was assumed that the Pd sites dispersed on Pt particle were responsible for the high hydrogenation activity

Fixed bed reactor, 200-300 "C, total pressure 4.9 MPa, LHSV 1.5 h- and hydrogen to feed oil ratio 500 NYl. EXAFS characterization of the catalysts

1999 PtPd/Si02-A1203, Pt/Si02-A1203 and Pd/Si02-A1203

LCO (middle distillate refinery product) / SRLGO feed obtained from refinery

6

Dependency of solvent effect on the hydrogenation activity. Toluene conversion in MeOH, EtOH, 1-PrOH (alcohol as a solvent), Ru/A1203 was 95, 100 and 100, respectively. Toluene conversion in MeOH, EtOH, 1-PrOH (alcohol as a solvent), Pt/A1203 was nil

5

39

T = 120 "C, P = 6 MPa, ratio substrate : Mono-aromatic solvent = 3 g : (6-0.75 g), reaction time compounds, toluene, phenol, etc. was 2 h

Benzene, n-heptane / benzene mixture (25 wt YObenzene), in the presence (200 ppm) and absence of sulfur

The dependency of the catalytic activity on thesupport has been considered

Alumina support provides a higher Black oil catalytic activity than the hydrotreatment hydrogenation of aromatics, under reaction conditions as used in industry

1999 Ru/A1203and Pt/ A1203

2000 Mo-Ni supported on y-A1203,or hydrated titanium dioxide (HTD), or palygorskitemontmorillonite clay (PMC) 2000 Pt/WO,-ZrO2, (12.7 wt YOW) and Pt/Beta (Si/Al= 12)

s3 z8'

b

ul

6' 3

i;.

se

3

B3

=L

L

* *

40

The role of hydrogen partial pressure in the hydrogenation of benzene, toluene and three isomeric forms of xylene over Ni/Si02; TOF decreases in the order benzene>toluene>p-xylene>m-xylene> o-xylene, representing the order of increasing stability and decreasing reactivity, of the adsorbed x complex

Fixed bed glass reactor (i.d. = 15 mm), Benzene, toluene, o-xylene, rn-xylene atmospheric pressure, temperature range 120 "C 5 T 5 250 "C, space and p-xylene velocity 2 x lo3 h-l, hydrogen to pressure was varied from 1.9. 9.6. MPa at constant aromatic pressure 0.004 MPa using dry nitrogen as the diluent; when aromatic pressure was varied 0.001-0.006 MPa, the hydrogen pressure was constant, 0.094 MPa and again nitrogen as the make-up gas

1999 Ni/SiO;?

9

Improve selectivity of benzene hydrogenation to cyclehexene. Ru-B/Si02 (amorphous) exhibited better selectivitytowards cyclohexene than the corresponding RdSiO2

Selective hydrogenation of benzene

Autoclave with a magnetic stirrer and 1000 rpm stirring, T = 150 "C, P = 4 MPa

1999 Ru-B/Si02 amorphous and Ru / Si02

Ref (s) 8

Benzene, xylene, mesitylene

T = 220 "C, P = 6 MPa, WHSV 2 h-l, H2 / hydrocarbon volume ratio 800, catalyst with particle size 40-60 mesh, the feed was a mixture containing 20 wt% hydrocarbon with 80 wt% n-hexane

1999 Pt (or Pd)/ y-A1203 Pt (or Pd)/ PLC, PLC is pillared clay which are known as crosslinked smectites, a new class of molecular sieve-like materials with a large pore size

Products / remarks Both catalysts, Pt (or Pd)/y-A1203and Pt (or Pd)/PLC showed the same saturation activity for benzene, although the selectivityfor cyclohexane Pd/Al-PLC is lower than Pt/Al-PLC. Generally, Pt (or Pd)/ PLC show much higher catalytic activity for the hydrogenation of large aromatic molecules, large pore structure and weak acidity

Substrate

Reaction conditions / remarks

(contd.)

Year Catalyst

Table 1

o\

-

Benzene, toluene

Phenanthrene

T = 200-400 “C,P = 5 MPa, a dual noble metal catalyst Pt(2)-Pd( 10) showed similar hydrosulfurization results and better hydrogenaton compared with a Co-Mo catalyst

1999 Alumina-supported Pt and Pd, Pt (or Pd)/ A1203,presulfided (mixture of 5% H2S in H2) Co-Mo catalyst

1998 Modified catalysts Micro-catalytic stainless steel reactor, L = 10 cm, d = 6 mm, T = 50-250 “C, containing 0.35% Pt-Al203, introducing flow rate H2 20 cm3/ min a second metal Ir, Rh, Re and U and fluorination and chlorination with different halogen contents of 1.3 and 6 wtYo

Distillatefuels

Low temperature catalytic hydrogenation processes over noble metal catalyst without addition of dibenzothiophene

1999 Two stage, new design catalyst concept in the hypothetical stage, First Ni-Mo or Co-Mo. Second noble-metal, which will be supported as bimodal on zeolite with two pore openings, < 5 A and in laqge pore opening <6A

Benzene to cyclohexane, toluene to methylcyclohexane, respectively.Toluene possesses higher activity than benzene on the monometallic catalysts, the benzene hydrogenation activity of the bimetallic catalysts follows the order PtRh > PtIr > PtRe > Pt > PtU, the corresponding order for toluene hydrogenation follows PtRh > PtRe > PtIr > Pt > PtU

Dihydrophenanthrene, tetrahydrophenanthrene, octahydrophenanthrene perhydrophenanthrene Pt-Pd catalyst showed better performance in hydrogenation of phenanthrene in presence of 1 wt% DBT than the conventional Mo-based catalyst

Multiple catalyst bed to achieve deep hydrosulfurization and hydrogenation, first stage ‘HDS’ Ni-Mo or Co-Mo catalyst, second stage ‘HDY’ over noble-metal catalyst supported on dealuminated mordenite

43

Tetralin with benzothiophene as a model compounds for sulfur poisoning (0- 1000 ppm sulfur content)

1998 Pt/yA1203

Continuous down flow fixed-bed reaction system, WHSV ranging from 2.0 to 12 (g of feed/h:g of catalyst), investigation of the relation between sulfur-poisoningand catalytic properties of Ptly-A1203catalysts for aromatic hydrogenation. T = 270 "C,total pressure 3.2 MPa

n-Propylbenzene alone; n-propylbenzene and hydrogen sulfide; n-propylbenzene, hydrogen sulfide and ammonia; n-propylbenzene, hydrogen sulfide, ammonia and water

1998 Conventional NiMoI Batch autoclave reactor, pressure 6.9 yAlzO3,3.2 wt% NiO MPa, T = 330,350 and 375 "C and 15.4 wt% Moo3, presulfiding with a 1: 9 H2S-H2 mixture at 325 "C, 8 h

Industrial conditions, a-A1203 Benzene supported nickel showed the best activity, with 40% nickel concentration and optimum metal area of 10.8 m2/g

1998 Ni on different supports

Substrate

Reaction conditions I remarks

(contd,)

Year Catalyst

Table 1

44

45

46

The reaction is moderately reduced by sulfide and severely inhibited by ammonia, addition of water did not further affect the aromatic hydrogenation

The material balance is about 98% for the catalytic performance tests. CO chemisorption and extended X-ray absorption h e structure (EXAFS) spectroscopy revealed that the decline of the reaction rate is caused by the formation of PtS and the reduction of PtS was achieved by hydrogen reactivation

RefW

Hydrogenated cyclic compounds

Products I remarks

00

1. Dimethylcyclohexane, ethylbenzene, decalins, tetralin 2. Perhydroanthracene 3. Benzene, cyclohexane, hexylbenzene, hexylcyclohexane, cyclohexylbenzene, bicyclohexyl, dodecahydro-, octahydro-, hexahydro-, tetrahydrocarbazole 4. Benzene, cyclohexane, hexylbenzene, hexylcyclohexane, cyclohexylbenzene, bicyclohexyl, biphenyl, tetrahydro-, hexahydrodibenzothiophene 5 . Different hydrocarbons etc. The results showed that the higher sulfur tolerance of the bimetallic Pd-Pt was achieved when Pd-Pt was supported on the highest acidity zeolite, the sulfur tolerance decreasing when the acidity of support decreases

Naphthalene Anthracene Carbazole Dibenzothiophene Coal-derived oil

1. 2. 3. 4. 5.

For the proupose of the hydrogenation of aromatics

Characterization of the catalysts by different technique, i. e. EXAFS measurement

1998 Pt and Pd, zeolitesupported catalysts, (Pd : Pt mole ratio of 4 : 1). The sulfidation of the catalysts was done in a 1000 ppm H2S/N2stream

Tetralin, trans- and cis- decalins. All samples demonstrate high hydrogenation activity in the first step when tetralin is formed. In the second step, trans- and cis- decalins are formed, the tungsten catalysts are two times more active than the molybdenum catalysts. Mo for mild hydrogenation, while tungsten is superior for deep hydrogenation

Autoclave, P = 10 MPa HZ, T = 430 "C The addition of phosphoric acid to the catalyst was an attempt to change the chemical properties of the catalyst, in order to increase activity of heteroatom removal

Naphthalene

1998 Phosphoric acidpromoted Mn203-Ni0

1998 (P)NiMo/Ti-HMS T = 325 "C, total pressure 4.4 MPa, and (P)NiW/Ti-HMS, 7 h, naphthalene dissolved in nphosphorus hexadecane concentation was 0.2 wt% P205,NiMo/ Ti-HMS, NiW/ Ti-HMS, sulfidation mixture (H2 15 vol YO H2S)

10

(contd.)

50

51

1-Methylnaphthalene, Examination of the hydrogenation of 1-methylnaphthalene was used to coal-derived middle quantify the conversion rate of the distillate hydrogenation of the aromatics contents in the coal-derived middle distillate At low Ru content (Ru 0.2) cyclohexane is the hydrogenation product. As the Ru content increases cyclohexane formation decreases and hydrogenolysis product (methane) will be formed

Ref($1 49

Products I remarks Cumene hydrogenation yielded isopropylcyclohexane is the only product, Hydrogenation of cumene, in the absence of sulfur. 5 wt% cumene and 95 wt% tetradecane Mo2C possesses higher activity than noble metal catalyst, the thermodynamic equilibrium conversion was 99.99%, indicating neither Mo2C nor Ptly-Al203 achieve equilibrium conversion. MoS2 is not an active catalyst for hydrogenation

Substrate

First-order reaction with respect to aromatic and partial pressure of hydrogen. P=4-12MPa, T=400"C

Benzene Catalytic test were performed in a continuous flow fixed bed microreactor, atmospheric (0.10 MPa) pressure, T = 250 "C, H2 flow = 60 cm3/ min, weight of catalyst = 20-40 mg, the conversion level was kept below 10%

1998 Rdalumina catalysts (0.21-5.1 1 wt% Ru)

Simultaneous hydrogenation, 5 wt% cumene and 95 hydrosulfurization and wt% tetradecane, in hydrooxygenation were observed with the absence of sulfur minimal deactivation of Mo2C up to 30 ppm for sulfur, 2000 ppm oxygen and 5 wt % cumene, Ptly-A1203 deactivated immediately upon addition of sulfur. P=5.1 MPa, T=250"C, 5.3 pmol s-l (5 cm3h-l)

Reaction conditions I remarks

1998 Ni-Mo/alumina

Year Catalyst

Table 1

0

w

1997 Pt/Y-zeolite

Reaction conditions were close to realistic industrial conditions, temperature dependency with a max. in the reaction rate at 325 "C, the pressure dependency of the rate with respect to the ratio of the toluene/H2 was sensitive to the level of the sulfur in feed Toluene

Methylcyclohexenes and methylcyclohexane

Higher acidity of Pt/MCM-41 has been favoured over the higher activity of Pt/A1203,for aromatics hydrogenation

Naphthalene in n-hexadecane was used to simulate the aromatic in diesel fuels

1997 Pt/MCM-41, Pt/A1203 T = 180"C, P = 4.2 MPa, concurrent down-flow trickle bed reactor, the metal dispersion on MCM-4lwas higher than that on A1203

Main components of the hydrogenation product, Naphthalene to decalins and tetralin; acenaphthene to tetrahydroacenaphthene; phenanthrene to dihyrophenanthrene, tetrahydrophenanthrene, octahydrophenanthrene; anthracene to dihydroanthracene, terahydroanthracene and octahydroanthracene; fluoranthene to tetrahydrofluoranthene; and pyrene to dihydropyrene

Toluene, solvent was Methylcyclohexenesand either in mixture of methylcyclohexane water (33.4 mol%) and methanol (66.6 mol%) or methanol

A light fraction of Trickle bed reactor P = 9.8 MPa, T = 350 "C, hydrogen anthracene oil dm3 s - l , the main flow 4.0 x dissolved in toluene purpose of the experiment was to obtain information on the effectiveness factor for both the wetted and dry zones of the catalyst

1997 Ruthenium supported T = 303 "C, atmospheric pressure (0.10 MPa), a semi-batch magnetically on silica and silicastirred reactor with a coolingjacket active charcoal contained 5 wt YOof the metal by sol-gel method

1998 Ni-Moly-AlzO3, catalyst was presulfided by a mixture of 10% of H2S in H2

53

52

11

13

Naphthalene in the Decahydronaphthalene absence and present of Tetrahydronaphthalene sulfur (Benzothiophene), n-tridecane was used as a solvent

Tetralin, in presence of H2S, or dimethyl disulfide in solution in n-heptane

T = 200 "C, in a microautoclave reactor, mordenite-supported Pt and Pd catalysts are more active than Y zeolite supported Pt and Pd, respectively. Y zeolite supported catalysts afford higher yield of cisdecalin. Mordenite-supported catalysts give higher yield of trans-decalin.The addition of sulfur decreases the activity of all catalysts tested

T = 300 "C, the activity for RuKYd was higher than for NiMoIalumina. The activity of the RuKYd was dependent on the sulfidation method, the catalyst sulfided by dimethyl disulfide was less active than when sulfided by H2S/H2mixture

Cyclohexene, 1-hexene Cyclohexane, hexane Competitive hydrogenation of cyclohexene and \-hexene pore, with a constriction of 5 A yielded higher rate of hydrogenation for 1-hexene. Other catFlysts supported on a constriction (7 A) and activated coal display comparable rates for both reactants. Gas-phase fixed-bed reactor, T = 100 "C, WHSV = 25-50 h-'

1997 Pt and Pd supported on mordenite (HM38), a Y zeolite (HY), A1203 and Ti02

1997 Sulfided Ru on dealuminated Y zeolite (RuKYd), NiMo/alumina

1997 Pt supported on carbon fibers; small constriction (5 A), large constriction (7 A)

ProductsI remarks

Substrate

Reactionconditions I remarks

(contd.)

Year Catalyst

Table 1

56

55

54

Ref(s)

N

c

To develop high sulfur tolerant Pt catalysts for hydrogenation of aromatics, the choice of catalyst support and/or the finding of the second metal (for the formation of bimetallic interactions, so the sulfur adsorption decreases) is necessary Tetralin, cis- and trans-decalin, both MCM-41 and USY supports showed the greatest sulfur resistance. Ptf (A1)MCM41 presented the highest activity for hydrogenation of aromatics in LCO feedstock, especially at low temperature (300 "C)

Normal sulfur poisoning, tetralin with 1000 ppm sulfur, T = 270 "C, P = 3.2 MPa. With more severe conditions, 19 atm and 2000 pprn sulfur, WHSV 4.8 (g of feed/h:g of catalyst). Continuous fixed bed reactor

Batch reactor, (reaction conditions for 1. Hydrotreatment of naphthalene + 200 naphthalene hydrogenation) T = 225-275 "C, total P = 5.0 MPa, PPm (DBT) dibenzothiophen solvent was n-decane; LCO hydrogenation, T = 300-350 "C, 2. light cycle oil (LCO) (400 ppm total P = 5.0 MPa, WHSV = 4 hsulfur, 70 wt 'YO aromatic)

1997 Wy-A1203

1997 Pt/ (two MCM-41 supports, differing in their chemical composition) Wamorphous mesoporous silicaalumina (MSA Si/Al= 100) Pt supported on a commercially amorphous silicaalumina (ASA) WUSY zeolite Ptfy-AlzO3 PtfSiOZ The nominal Pt = 0.5- 1 wt%

w

c ,

61

Naphthalene dissolved cis-Decalin, trans-decalin The selectivity of cis-decalin was studied in n-hexadecane for HYD purpose

1997 Pt/A1203,Pt/MCM-41 Concurrent down flow trickle bed reactor, T = 180 "C, P = 4.2 MPa, LHSV = 2.8 h- l , H2/liquidfeed 800 mVml, the hydrogenation activity of PdMCM-41 is higher than that of PdA1203, but cis-decalin selectivity was lower than with Pt/A1203

-

60

-

59

Tetralin in presence of From HDY of tetralin the major product The hydrogenation activity was very large amounts of H2S was cis- and trans-decalin high in the presence of H2S and (1.85%) roughly 300 times the activity (expressed per metal atom) of an industrial NiMo/ A1203 catalyst. The active phase, which consists of cluster of ca. 50 Ru-S, is located in the zeolite framework

From HDY of toluene the major product was methylcyclohexanewith a 50% selectivity, the others were ethylcyclopentane,methylcyclopentane and dimethylcyclopent anes; From HDY of tetralin the major product was cis- and trans-decalin with a 90% selectivity, small quantities of methylindanes and methylcyclopentanes were observed

Tetralin and toluene in the presence of 1.9% H2S

1997 Ruthenium sulfide supported in a Y zeolite, HYd (commercial dealuminated) and KYd (prepared from commercial dealuminated HYd zeolite) and subsequent sulfidation

-

Microreactor system, for toluene HYD, T = 280-390 "C and P 4.5 MPa; for tetralin HDY, T=250-300 "C, P 4.5 Mpa. For tetralin HDY, the specific activities were found to decrease in the order: RuHYd > RuHY > RuKHYd >> NiMo/ A1203> RuKY For toluene HYD, the specific activities were decreased in the order: RuHYd > RuKHYd > RuHY >> RuKY NiMo/A1203

Ref(s)

1997 Ru dispersed in a series of zeolites with various acidic properties; HY, KY (prepared from Nay), KHYd (prepared from commercial dealuminated HYd zeolite) and subsequent sulfidation

ProductsI remarks

Substrate

Reaction conditions / remarks

(contd.)

Year Catalyst

Table 1

P

b-

Batch reactor, T = 310-350 "C, P = 5.0-8.0 Mpa. This study focused on the difference in kinetics, between powdered and pellet catalysts in liquidphase HDY reactions of the aromatic ring. The inhibitory effect of 1-methylnaphthalene was clearly observed for the powdered catalyst

1997 Ni/SiOz

Gas-phase hydrogenation, T = 120-250 "C, the hydrogen pressure was constant, 0.095 MPa; where the pressure of each aromatic was varied in the range 0.001-0.006 MPa, the TOF of the three xylenes increased in the order o-xylene < rn-xylene < p-xylene, over the entire temperature range studied

1997 Ruthenium supported Autoclave, initial P = 10 MPa (H2), T = 430 "C on mixed oxides (Mn203-Zn0, Mn203-Ni0, Mn203-La203) with different molar compositions. The supports were loaded with 0.1,0.2, 0.5 and 1.0 wt ruthenium

1997 Powdered, cylindrical and trilobed commercial Co-Moly-Al203

I. Etthylcyclohexane, decalin, tetralin, others I I. Per-, octa-,tetr ahydr o-anthracene, others 111. Different aromates

Stereoisomeric product mixtures of the saturated dimethylcyclohexane

0-,m-and p-xylene

Methyltetralins and methyldecalins

I. Naphthalene 11. Anthracene 111. Coal-derived oil

1-Methylnaphthalene diluted in a C14-C16 normal paraffin mixture

64

63

62

c #

5' 3

g

'i$

2 8

k

k

is

' 5 i2

The analysis of product samples from long-term stability tests (25 days), collected periodically, found that aromatic conversion ranges from 3- 10 wt%, depending on precursors and pretreatment conditions

Commercial diesel

Synthetic crude The aromatics conversion was 29% for Pt distillates from on pillared interlayered clay (PILCj Canadian oil sands of alumina and 44% for Pt/Y-zeolite varying sulfur content, high > 100 ppm and low < 10 ppm, monoaromatics content 31 mass%; diaromatics 8 mass% and triaromatics 0.8 mass%

The bimetallic catalysts possess a relatively high selectivity for aromatic reduction and other hydrotreating processes, T = 340 "C,P 2 4 MPa

T = 310 "C, P = 7.0 MPa, for kinetic investigation the temperature was 320, 340 and 360 "C. The enhanced activity of Pt/Y-zeolite-alumina catalyst is attributed to unique structure of the support, producing well-dispersed Pt metal clusters

1996 Pt-PdlrAl203

1996 Pt on pillared interlayered clay (P1LC)-alumina and Y-zeolite-alumina

-

- -

cis- and trans-dimethylcyclohexane;the distribution of stereoisomerswas dependent on the reactant pressure as well as on temperature. The stereoisomers were governed by the adsorptioddesorption kinetics of an intermediate cyclic olefin; the configuration of the final product depended on the orientation of the olefin double bond upon adsorption

0-, p-xylene

Gas-phase hydrogenation, from the kinetic investigation suggested a noncompetitive adsorption of hydrogen and aromatic compounds on the catalyst surface. Thermodynamics suggested the stepwise hydrogen addition to the aromatic molecule

1997 NilAl2O3

~

Products I remarks

~~~

Substrate

~

Reaction conditions I remarks

(contd)

Year Catalyst

Table 1

67

66

65

Ref(s)

m

c

69

70

71

The relative rate of toluene and benzene HDY with an equimolar gas mixture showed that toluene was less reactive than benzene on Pd catalysts, whereas the value for toluene HDY indicates that is favoured on Pt catalysts The catalysts showed activity and selectivity to HDY reaction of aromatics

Naphthalene, decalin and several products with the same formula C10H18

The reaction parameters such as metal Methyl benzoate loading, temperature and solvents effect have been studied

Tetralin Flow reactor, T = 300 "C, P = 7 MPa (total). DMDS and H2S were used to study the effects of the sulfidation method and its dependencies on hydrogenation activity. It was found that the H2S method was the most suitable

1996 Sulfied ruthenium on KY zeolite

1996 Pd/Si02-A1203 (0.22-1.64 wt0/0Pd), Pdq-Al203 (1.67 wt YOPd), Ptlq-Al203 (0.78 wt% Pt) and commercial P d C (3.61wtYo Pd)

1996 Pt supported on inorganic polymer, silicapolysulfoalumoxane, with different metal loading

68

The main reaction product was the completely hydrogenated cycloalkane

Di- and tri-substituted T = 95-125 "C, P = 2-4 MPa. The aromatics, such as oactivity of different substituent positions decreased in the order para > m-and p-xylenes,pcymene and meta > ortho, the trisubstituted mesitylene benzene (mesitylene) had a lower reaction rate than the disubstituted compounds (xylenes) Benzene and toluene T = 80 "C, P = 0.007 MPa (total aromatic partial pressure) and 0.095 MPa (hydrogen partial pressure)

1996 Commercial preactivated catalyst, nickel-alumina

(contd.)

Reaction conditions / remarks Substrate

1996 Platina on mesoporous aluminosilicate (Al-MCM-4 1) support with varying WA1 ratios using different aluminum sources (aluminum isopropoxide, pseudo boehmite and aluminum sulfate)

A 30 ml stainless-steel tubing bomb I Naphthalene batch reactor was used. The reactor I1 Phenanthrene was heated in a fluidized sand batch under vertical shaking (240 cycles/ min), P=7-10.5 MPa H2, T = 200-300 "C. The catalyst was active for hydrogenation of large aromatics, the activity was significantly different depending on the synthesis conditions, especially with respect to the source of aluminium

1996 Ammonium T = 350,400 and 450 "C, P = 7.2 MPa, 1. Naphthalene tetrathiomolybdate as hydrogedreactant molar ratio (5: l), 2. Phenanthrene a catalyst precursor microautoclave reactor, the rate of the 3. Pyrene reaction is most important for bicyclic compounds, as ring size increases, kinetic play a less important role and thermodynamics become the driving force for the outcome of the reaction

Year Catalyst

Table 1

I Decalin tetralin I1 Tetra-, di-, octa-, tetradecahydrophenanthrene

1. Tetralin the only hydrogenation product 2. Primary product is dihydrophenanthrene and tetrahydrophenanthrene. Octahyrophenanthrene is a hydrogenated product of the primary products 3. The major product at all temperatures for pyrene hydrogenation is dihydropyene. Secondary hydrogenation products are tetrahydropyrene and hexahydropyrene

Products/ remarks

w

73

72

Ref

Benzene, toluene

T=30°C P=O.lOMPa

T = 350°C P = 6 MPa. The hydrogenation was performed over a large range of H2S partial pressures. The Nualumina is nearly inactive

1996 Silica-supported carboxymethylcellulose platinum complex (Si02-CMCPt> 1996 Zeolite-supported noble metal catalysts

1996 Mo, Ni, Ni-Mo and Ni-Mo-P/alumina

Toluene in the presence of H2S

Low temperature hydrogenation of aromatics in the absence and presence of DBT

Tetralin, cis- and trans-decalins and others. The equilibrium conversion decreases with increasing temperature. Tetralin becomes the dominant product at higher temperatures

Naphthalene, the solvent used was tridecane. The hydrogenation reaction was started when bezothiophene was added to the solvent

Tubing bombs, T = 200 and 280 "C, P = 6.99 MPa H2. The Pd/TiOZ showed higher sulfur resistance in comparison with the other catalysts

1996 Pf/A1203,PdA1203, Pd/Ti02 and NiMo/ A1203(commercial)

Methylcyclohexane

Cyclohexane and methylcyclohexane

Tetralin cis- trans-decalins

Naphthalene in hexadecane (as a solvent) was used. In the reaction with sulfided catalysts, an additional 1.5 wt.% Dimethyl disulfide was added as a sulfur source

Tubing bomb microreactor, T = 310 "C, P = 6.90 MPa cold hydrogen pressure. The Pt-promoted catalysts were more active than the original catalysts in the oxide forms, whereas the activities of the sulfide catalysts both in promoted and nonpromoted were similar

1996 Pt, Ru and Ir promotion on three commercial A1203supported catalysts (NiMo, CoMo and NiW)

78

77

76

Reaction conditions I remarks

The hydrogenation has been carried out under pressure of different coke oven gases = 55 vol.% H2 instead of pure hydrogen, influence of reaction time and temperature were studied T = 300-450 "C, P = 11-25 MPa, 16 h

Mathematical calculation of hydrogenation

Kinetic investigation of gas-phase hydrogenation, T = 130-220 "C, differential fixed bed reactor, hydrogen pressure was varied over 0.04-0.09 MPa and xylene pressure over 0.01-0.035 MPa, helium was a makeup to adjust the flow to 10 mmollmin

Gas-phase hydrogenation, T = 145-220 "C,m-xylene partial pressure was varied over 0.01-0.035 MPa and hydrogen partial pressure was in the range 0.4-0.9 bar

1996 Commercial nickelmolybdenum (3 wt% NiO, 15 wt?h Moo3 on alumina) and Pd (5 wt% on alumina)

1996

1996 Ni/A1203

1996 17 wt .Yo Ni/A1203

(contd.)

Year Catalyst

Table 1

Dihydrophenanthrene, tetra hy drophenanthrene, octahydrophenanthrene; the maximum yield was at 370 "C

ProductsI remarks

m-Xylene

0-, p-Xylene

The distribution of stereoisomers depended only on temperature, mixtures of stereoisomeric products of the saturated dimeth y lcyclohexane. Rate maximum for saturated product was at about 405 K

Main product, cis- and transdimethylcyclohexaneswas formed in non-equilibrium ratios; the kinetic results showed that the reaction proceeds through consecutive hydrogen addition steps on the catalyst surface and cyclic olefin was intermediate product

Benzene, naphthalene Cyclohexane, decalin and pyrene

Phenanthrene

Substrate

82

81

80

79

Ref0)

0

h,

84

Cyclohexane, the data obtained from benzene hydrogenation were found to be in very good agreement with the percentages of metal exposed Completely hydrogenated cycloalkane and trace amounts of cycloalkenes, hydrogenation rate decreased with increasing length of the substituent in the benzene ring

Benzene The anionic exchange on a low and high specific surface area ceria, exhibits the same amount of exposed metallic Rh atoms, hence the high surface area ceria contain 50% more Rh

Benzene, toluene, T =95-125 "C, P = 2-4 MPa. A ethylbenzene, cumene model on competitive adsorption of hydrogen and the aromatic compound fitted the experimental data. The hydrogenation rate was based on a sequential addition mechanism of adsorbed hydrogen to the aromatic nucleus

1996 0.15-0.33 wt % Rhl CeOz catalysts

1996 Commercial preactivated catalyst of nickel-alumina (Ni 16.6 wt%)

85

83

Tetralin and two isomers of decalin, tetralin is identified as a primary product, decalins are identified as non-primary products

Naphthalene Batch reactor at 350 "C and 17 MPa, kinetic investigation. Vanadium sulfide deposits led to decrease hydrogenation rate in the naphthalene network; the results showed a sequential hydrogenation of naphthalene to form tetralin which was hydrogenated to give decalin

1996 Ni-Mo/yAlzO3 enriched with various amounts of nickel and vanadium by contact with solutions of the respective meta1 naphthenates. CS2 was added to the mixture so that the catalysts was presulfided

Naphthalene dissolved The catalytic activity of Pt-aluminium borate was higher than that of Ptlyin n-hexadecane A1203,but its cis, trans decalin selectivity is lower than that of Ptly-Al203 catalyst owing to the higher acidity; too much boron (A1:B = 8) degrades the hydrogenation activity

P=5.17 MPa, LHSV=2.8 h-'

Continuous fixed-bed reactor, T = 270 "C, P = 1.2-3.3 MPa, kinetic investigation of sulfur on the deactivation of catalysts

1996 Pt-aluminium borate

1995 Pt/y-A1203

Tetraline

1. Tetralin

A Langmuir-Hinshelwoodreaction model, which was based on a chemisorption scheme with irreversible surface reaction control for tetralin and reversible surface reaction for sulfur poisoning was proposed to describe the deactivation model of the catalysts

5. Dodecahydrochrysene 6. Dodecahydro-l,2-benzanthracene

4. Octahydrophenanthrene

3. 1,2,3,4,5,6,7,8,-octahydroanthracene

2. 9,lO dihydroanthracene

1. Naphthalene + hexane 2. Anthracene with lithium diisopropylamide 3. Anthracene with potassium bistrimethylsilylamide 4. Phenanthrene with the lithium diisopropylamide 5. Chrysene 6. 1,Zbenzanthracene

88

87

86

Ref (s)

1996 Lithium Autoclave, T = 250 "C, 4- 18 h, P = 7 MPa diisopropylamide Potassium bistrimethylsilylamide

Year Catalyst

Reaction conditions I remarks

~

Products I remarks

(contd.) Substrate

Table 1

t 4

h)

1995 Rh and Ni organometallic complexes anchored on USY zeolites

Mild reaction conditions, 80 "C and 0.6 MPa of HzO, the strong cooperative effect between the zeolite surface and the transition metal surface was thought to be responsible for the enhancement of hydrogenation reaction

1995 Ptlalumina-aluminum P = 6.8 MPa, T = 350 "C,trickle bed reactor, LHSV = 2.8 hphosphate

Tetralin, decalin, Walumina-aluminium phosphate had a better hydrogenation activity and lower cis-decalin selectivity than Wy-A1203, due to the higher acidity of the support

Trickle bed reactor, P = 5.17 MPa, T =240 "C

1995 Pt/A1203-aluminium phosphate

Benzene, toluene, a-methylstyrene

Total conversion of benzene, toluene and a-methylstyrene using zeolite containing the Rh complex was achieved after reaction times of 6, 18 and 24 h respectively. The Ni complex was less active than the corresponding Rh complex. The positive influence of the zeolite was attributed to an increase in the concentration of reactants inside the pores

Naphthalene dissolved The result showed that the catalyst had a in n-hexadecane better hydrogenation activity and lower cis-decalin selectivity than Pt/y-A1203 catalyst due to the higher acidity of the support

A solution of naphthalene in nhexadecane was used to simulate the aromatic in diesel fuels

Pt-Pd combination catalyst showed the highest sulfur resistance. The interpretation of the spectroscopic investigation was that the role of Pd in enhancing sulfur resistance was due to decreasing the electron density of Pt and thereby inhibiting the adsorption of H2S

Hydrogenation feed contain 1000 ppm Tetralin + 1000 ppm sulfur and sulfur, H2/hydrocarbon = 2.7, straight run distillate T =280 "C,P = 2.62 MPa diesel

1995 Pt/y-A1203,modified by adding a second metal, Co, Mo, Ni, Re, Ag and Pd

90

16

15

89

w h,

T = 200-260 "C, P = 1.7-8.7 MPa and LHSV 1.5-8.0 liquid hourly space velocity, trickle bed reactor

1995 Naphthalene in inert solvent n-hexadecane

1. o-xylene 2. Naphthalene 3. Phenanthrene 4. Pyrene 5. Anthracene 6. Chrysene

Batch autoclave, T = 350 "C,P = 6.9 MPa H2, presulfiding at 400 "Cfor 135 min. in 10% H2S in H2. Cyclohexane was used as a solvent. Focus on the relationship between molecular structure and hydrogenation reactivity in heavy oil processing. Equilibrium ratios were much larger than unity for benzenic, larger than unity for the naphthenic and smaller than unity for the PHE

1995 Presulfided CoMoI A1203

PtlA1203

Substrate

Reaction conditions I remarks

(contd.)

Year Catalyst

Table 1

92

91

Ref0)

Tetralin and cis- and trans-decalin. The reaction was sequential, i. e. naphthalene hydrogenated to tetralin, followed by sequence hydrogenation to cis- and transdecalin. The apparent activation energies for hydrogenation of tetralin to cis- and trans-decalin were found to be 9.88 and 7.25 kcallmol, respectively

m1. cis-trans-l,2-dimethylcyclohexane, xylene, p-xylene, cis- and trans- 1,3dimethyl- and -1,4dimethylcyclohexanes 2. Tetralin and cis-trans-decalins 3. Di-, tetra-, octa- and perhydrophenanthrene 4. Di-, tetra-, hexa-, deca- and perhydropyrene 5 . Di-, tetra-, octa- and perhydroanthracene 6. Di-, tetra-, hexa-, octa-, dodeca- and perhydrochrysene

Products I remarks

Formation of surface sulfide on the catalyst

The synergistic effects between Ni(Co) Benzene and Mo and between metal loading and acidity of zeolite was investigated. The synergistic effect occurs at a Ni/ (Ni+Mo) ratio of about 0.4-05 and was related to the formation of Ni-Mo-S phases within the structure of zeolite

1995 Molybdenum nitride

1995 Ni-Mo sulfide catalysts supported on zeolites (HY), conventional Ni-Mo/ A1203

Vacuum gas oil

Stirred autoclave having a Parr 4561 Selective assembly system, T = 150-250 "C, P up hydrogenation of to about 6.90 MPa. The hydrogenation benzene in gasoline was carried out in a biphasic system of water and gasoline. The benzene was selectively solubilized in the water for further hydrogenation in the presence of the water-soluble catalyst mixture

1995 Catalyst comprising a mixture of catalyticallyactive mixture of water-soluble, organometallic compounds. Catalyst (l), M[L],[X], wherein M is a metal selected from Cr, Fe, Co, Ni, Mo, Ru, Rh, Pd, Ta, W, Re, Os, Ir, Pt, La and Ce; L is an aromatic hydrocarbon; X is halogen; x and y are integers from 1 to 10. Catalyst (2), tris(tripheny1phosphine)rhodium(1)halide or tris(tripheny1phosphine)ruthenium(1)halide

Cyclohexane, Ni-Mo sulfide in the zeolite is highly dispersed and responsible for the high hydrogenation activity, which was comparable to the conventional Ni-Mo/A1203

The combination of both catalysts was found to produce conversion in excess of 40% which was much more than either of the catalysts separately

95

94

93

r,

3

6'

5

G-

$5

B

Q.

?

k

%

s

2

..

Naphthalene 1995 Mordenite-supported The isomers of trans- and cis-decalin Pt catalysts takes place on acid sites 1995 17 wt.% Ni/A1203 Differential microreactor, atmospheric Ethylbenzene pressure, T = 130- 190 "C, ethylbenzene partial pressure was varied from 0.01 MPa to 0.035 MPa, hydrogen partial pressure was kept in the range of 0.04-0.09 MPa 1995 Molybdenum Propylbenzene Total P = 5 MPa (hydrogen at 4.14 oxynitride MPa, cyclohexane at 0.85 bar and propylbenzene at 0.06 MPa), T=267-397 "C

1995 NiMo on Y zeolite

The carbon atoms on the surface and Benzene in the bulk of Pd particles of the catalyst showed a capacity for chemisorption and hydride formation o-Xylene T = 380 "C , P = 6 MPa, various amount of o-xylene (5, 10 and 20 mol% was added to n-heptane)

1995 PdIC

Substrate

Reaction conditions I remarks

(contd.)

Year Catalyst

Table 1 Ref(s)

Kinetic investigation of ethylbenzene hydrogenation. Rate maximum was obseved at 160- 175"C, depending on the concentration ratio of the reactant, main hydrogenation product was ethylcyclohexane Propylcyclohexane,characterization through, chemical analysis. TPD of propylbenzene and preadsorbed benzene at room temperature and successive catalytic runs indicated that, during catalytic runs, the surface of the catalyst will be chemically modified; an oxycarbonitride is formed without decrease of the specific surface area

101

100

96 The main product was cyclohexane, the turnover frequency in benzene hydrogenation was a structure-insensitive reaction 97 The rate of formation of m- and p isomers allow an estimation of the acid activity of a bifunctional catalyst. The hydrogenation activity was defined as the rate of formation of all Cg naphthenic compounds, which is proportional to the content of Ni and Mo of the catalyst and independent of the zeolite content Tetralin, trans- and cis-decalin 98,99

ProductsI remarks

a

h,

I : Catalytic Hydrodearomatization

27

2.1 Reactivity in Hydrogenation Reactions. - The recent reactivity studies have been reviewed by Moreau and G e n e ~ t e ,Girgis ~ ~ and Gates36 and Stanislaus and Cooper.37 In these reviews, the reactivity of aromatic compounds was defined as the overall conversion of aromatic compounds to fully and/or partially hydrogenated products. The hydrogenation reactivity of aromatic hydrocarbon was affected by the following factors: 0

0

0

0

0

0 0

0

0

aromaticity; the aromatic character of a molecule is a measure of its degree of unsaturation and on its thermodynamic stability the total aromaticity; generally given by total resonance energy which is defined as the value obtained by subtracting the actual energy of the molecule from that of the most stable contributing structure102-10s partial resonance energy; as the number of fused aromatic rings was increased, the resonance stabilisation energy per aromatic ring was decreased hydrogenation reactivity related to geometric modification of model compounds the contribution of the structural and geometrical effects of organic molecules to the reactivity; this is taken into account where interaction between the molecule and the catalyst surface was an important parameter differences in the n- electron cloud density in alkylated aromatics electronic effects; when the model compounds was substituted by alkyl or aryl groups, the slight differences observed in reactivity was accounted for in terms of electronic effect the presence of bulkier substituents; a significant effect on reactivity was assumed to be due to the steric effects. These effects have been most important when fused multiring aromatics were hydrogenated relationship to the reaction rate constant

Unlike olefin hydrogenation, high hydrogen pressures are required to effect ring saturation in aromatics hydrogenation. This is partly due to the low reactivity of the aromatic structure as a result of resonance stabilization of the conjugated system and partly due to limitations determined by the thermodynamic equilibrium at the pressures and temperatures employed. Therefore, most studies on the reactivity of aromatics have been conducted at pressures and temperatures that favour low equilibrium concentrations of aromatics. Relative hydrogenation reactivates of one ring in the multi-aromatic model compounds over a sulfided NiMo/A1203 were correlated to the rate constant by Moreau et al.,35 the correlation showed the following order: benzene (1) < phenanthrene (4) < naphthalene (20) < anthracene (40) where the numbers are relative rate constants. The reactivity of the aromatic compounds was correlated to the aromaticity of the rings.35-106 The total aromaticity is generally given by resonance energy (RE);lo5 resonance energy increases with the number of aromatic rings, independently of the presence or the absence of heteroatoms in the rings.35

28

Catalysis

Table 2

Hydrogenation of mono- di- and tri-aromatic hydrocarbons, positions for addition reaction in di- and tri-aromatic

Hydrogenat ion reaction

Total resonance energy (kcal mol-')

Resonance energy1 ring (kcal mol-')

36-40

40

59-75

28

1

9

+H2=Q

71-105

10

10

The magnitude of resonance energy per ring was less for naphthalene compared with benzene, and consequently the hydrogenation rate was low for benzene.lo6The low aromatic character of one of the rings in the naphthalene molecule is experimentally shown by its ability to undergo addition reactions across 1,2-positions, and the corresponding positions are 9,10 for phenanthrene and anthracene,lo7the behaviour is shown in Table 2. The increase in resonance energy with the angularity of the system,105e.g. anthracene and phenanthrene which are isomeric hydrocarbons (containing the same total number of aromatic rings) resulted in lower reactivity for phenanthrene. It is commonly argued that the presence of methyl substituents on the benzene ring stabilizes the adsorbed n complex with the resultant introduction of a higher energy barrier to aromatic ring hydrogenation. The hydrogenation rate in the presence of the noble metal catalysts decreased in order benzene > toluene > p-xylene > m-xylene > o-xylene under moderate reaction conditions, confirming this hypothesis. The bonding of a molecule on the surface of solid catalysts depends on the local electron density of states on the adsorbing metal atoms. Aromatic compounds are bonded to the solid surface of the catalysts via n-bonds involving an electron transfer from the aromatic ring to the unoccupied d-metal orbitals.lo8 Since the n-electron cloud density in toluene is higher than that of benzene, it would be expected that toluene is more strongly adsorbed on the metal surface. Consequently, a strong interaction between aromatic

1: Catalytic Hydrodearomatization

29

compounds and the metal atoms will lead to a reduction in the hydrogenation rate of the former. An increase in electron density cloud with the addition of another methyl groups can account for the lower reactivity of xylene. For metal sulfide catalysts, such as Ni-W-S/A1203 and Ni--Mo-S/A1203, the reactivity pattern of aromatic and alkylated aromatic compounds is the opposite.37In other words, addition of methyl groups to aromatic compounds gives rise to an enhancement of the reactivity of these molecules for hydrogenation; the corresponding hydrogenation rate for the same homologous series (as stated above) will decrease in the order benzene < toluene < p-xylene < rn-xylene < o-xylene over moderate reaction conditions. Both the number and position of the substituted groups in benzene ring affected the hydrogenation activity.68The observed reaction rate of a trisubstituted benzene (mesitylene) over a commercial pre-activated catalyst particles of nickel-alumina was lower than the reaction rate of the disubstituted (xylene) benzene? The activity of the different substituent positions decreased in the order para > meta > ortho.68The lowest reactivity of ortho-isomers has been attributed to steric effects.37 Steric hindrance of neighbouring methyl groups in ortho-positions had a more significant effect on the formation of n-bonded complexes. Pondi and V a n n i ~ observed e~~ that the relative rate of toulene and benzene hydrogenation with an equimolar gas mixture for the Pd supported catalysts in A1203and Si02-A1203 was less than unity. A value less than unity indicated that toluene was less reactive than benzene on Pd catalyst^.^^ In the same experiment a threefold higher value for relative rate was reported for Pt catalysts; the authors interpretation was that toluene hydrogenation was favoured on Pt.69 For the three catalytic systems, Pt/Si02-A1203, Pd/Si02--A1203, and PtPdl Si02-A1203, under moderate operating conditions and in the presence of 113 ppm sulfur, the order of aromatic reactivity was as the following sequence: naphthalene >> toluene > tetralin.4 This reactivity sequence was related to a decrease in the resonance energy per aromatic ring as well as to differences in the n-electron cloud density in the aromatic ring because of the inductive effect of the methyl group in t01uene.~The bimetallic catalyst exhibited higher hydrogenation activity for those three aromatic compounds. This enhancement in the aromatic hydrogenation activity in presence of sulfur was attributed to the synergetic effect between Pt and Pd.4 Of course, there are a variety of interpretations concerning the noble bimetallic effect on catalytic activity, but obviously the greater reactivity is a result of the presence of an addition effect, as opposed to a synergetic effect. In agreement with the literature data, in addition to the influence of resonance energy and stereochemistry, additional factors such as choice of the catalyst metal and reaction conditions have a significant effect on the reactivity of the aromatic molecules. Hydrogenation activity on a commercial Ni-alumina catalyst decreased with increasing length of the substituent in the benzene ring, thus benzene appeared to be the most reactive.85

Catalysis

30 100

iE

80

W

60 I ,

0

40

"

20 0 150

200

250

300

350

Pt (3)-Pd (10)

400

450

Temperature ('C) Figure 1

Conversion of PHE hydrogenation as a function of temperature on several catalysts (fromref 42)

The rate of hydrogenation of derivatives of benzene varies with the nature, number and position of the substituents. The ratio KTIB (toluene to benzene) of the adsorption coefficients is very large on ruthenium, about 9-10 on rhodium and about 1 on palladium. This behaviour was interpreted as an effect of a considerable electron transfer from methyl groups of the aromatics to the metal. Large KTIBratios are associated with the electron-deficient character of the metal.log Pyrene, fluoranthene, anthracene and fluorene were hydrogenated over a Pt, Pd, Rh and Ru catalyst in a study by Sakanishi et al. 110Also here the results showed a change in reactivity order depending on the catalyst used. However, the difference in reactivity can not be explained only by the structural properties of aromatic compounds. The reactivity will be influenced by the properties of the catalysts. The appearance of differences in reactivity depended on varying the type of catalyst and diversities recorded from the different reported works confirmed that more research is needed with a special focus on the reactivity aspects.

2.2 Thermodynamics of Hydrogenation. - One important property of the hydrodearomatization reaction is that the maximum equilibrium conversion occurs at temperatures close to or even below the temperatures needed for the hydrodearomatization rate to be of industrial interest and this is clearly seen in Figure 1. To the right of the maximum of the curves the equilibrium rate is limiting the hydrogenation process and to the left, the hydrogenation rate is limited by the temperature.2 This is why catalysts from the platinum group with a high activity are preferred rather than less active sulfided catalysts. Keane and Patterson40 included equilibrium curves for benzene, toluene and o-xylene at 0.1 MPa.

1 :Catalytic Hydrodearornat iza tion

31

Djega-Mariadassou et al. lol present equilibrium data for ethylbenzene at 5.0 MPa total pressure and a hydrogen pressure of 4.1 MPa.

2.3 Reaction Pathways and Kinetics of Hydrogenation. - The reaction network for hydro-dearomatization has been studied by several groups and will be presented below. Reaction order, activation energies and thermodynamic values will be presented. 2.3.1 Reaction Network. The hydrogenation of a one-ring system can be influenced by the presence of other functional groups. The reaction in this case will start with either hydrogenation of the functional group or hydrogenation of the aromatic ring. Takagi et aZ.6 have studied the effect of solvent for the hydrogenation of benzyl alcohol. This makes it possible to control the selectivity in the reaction. When hydrogenating monoaromatics the major product will be the hydrogenated ring, but also some cyclic mono alkenes are also formed.'" It is this alkene that according to Salmi et aZ.65981385 undergoes 'roll-over' in the adsorbed state and two different cis and trans isomers are formed. In the same way naphthalene is hydrogenated first to tetralin and finally to decalin in both the cis and trans f0rms.83992Huang et al.92present a figure, originating from Weitcamp, l2 claiming that several other intermediates exist. They present indications of the existence of octalin, as this compound is a necessary intermediate for the formation of the isomerized products. Qian et al.42 presented phenanthrene (an aromatic with three rings) dearomatization using both PdPt catalysts and sulfided CoMo catalysts. Total pressure was kept at 5 MPa with decalin as solvent and 1% aromatic, at various temperatures. Hydrogenation of phenanthrene proceeds through a network of reversible reactions. The middle ring was the most likely to be hydrogenated first. 2.3.2 Reaction Order. Keana hydrogenated xylenes over a nickel catalyst on silica. The turnover frequency (TOF) followed the following general expression.

TOF = kPxmPnH2 where the reaction order for the xylenes varied from 0.1 at 393 to about 0.44 at 530°C. The value of n for the hydrogen pressure increased with temperature from 0.7 to 2.3. 2.3.3 Activation Energies. Huang and Kang92 and Corma et aZ.58 provide a table of older results on apparent activation energies for hydrogenation of aromatic hydrocarbons, mainly benzene. Discussion of activation energies is given by Keane and P a t t e r s s ~ n . ~ ~ * ~ The true activation energy can be calculated from the observed apparent activation energy by subtracting the sum of adsorption energies of the aromatic and the hydrogen. The adsorption energies provide the major contribution to the true activation energy. Adsorption energies for an aromatic are in the range 100-140 kJ mol-'.

32

Catalysis

2.4 Catalysts and Nature of Catalytic Sites. - Because of more stringent environmental regulations, greater attention has been focused on the reduction of aromatics in distillate fractions. Also important is the use of deep catalytic hydrogenation of the aromatic compounds to produce petrochemicals, e.g. conversion of benzene into cyclohexane in the synthesis of caprolactam. The choice of catalyst type varies according to the application, nature of the feedstocks and desired activity/selectivity of the reactions. Conventional hydrotreating catalysts containing sulfided mixed oxides (NiMo, NiW, CoMo) can only accomplish moderate levels of aromatic saturation under typical hydrotreating conditions. Because of the thermodynamic limitations in hydrodearomatization, a deep level of aromatic saturation has not been achieved by increasing operation severity (high temperature and high Hz pressure). Among the conventional catalysts, NiMo and NiW are preferred for aromatic saturation. NiW catalysts have the highest activity for aromatic hydrogenation at low hydrogen sulfide partial pressures,37but their use has been limited due to the higher cost. Models of the active phases and structural aspects of these mixed oxides (NiMo, NiW, CoMo) have been recently reviewed by Topsoe and co-workers. Transition metals, as well as platinium-group metals, have been used as the catalyst for this type of hydrogenation. Platinum-group metals have the advantage of high hydrogenation activity at low temperature. The catalyst activity for the hydrogenation of benzene was reported to be in order Rh > Ru >> Pt >> Pd >> Ni > C0.113 Three categories of catalysts have been studied for hydrodearomatization in literature. The first category is a mono-metallic noble-metal catalyst. Obviously, these catalysts will be used in the hydrogenation processes in the absence of sulfur in the feedstocks. The second category of catalysts consists of two noble metals and is known as a bimetallic noble-metal catalyst. Several combinations of Pt, Pd, Rh and Re have been proposed. The third is a combination of the conventional hydrotreating catalysts promoted by a noble metal; these catalysts consist of three metals. An attribute of the noble-metal based catalysts is that they are active at lower temperatures; thereby they are the preferred catalysts for deep aromatic saturation. The main drawback of noble-metal hydrogenation catalysts is that they are poisoned by small amounts of sulfur and nitrogen organic compounds present in the feed. However, because in the industrial feeds several heteroaromatic compounds are present, considerable attention has been paid recently to developing catalysts with the high hydrogenation activity while maintaining strong resistance to poisoning by the small amounts of sulfur- and nitrogencontaining compounds in the feed stream. Several attempts have been made to minimize the poisoning effects of heteroatoms. The resistance to poisoning by heteroatoms is improved by adding a second noble metal to Pt, changing the particle size, or changing the acid-base properties of the carrier. With respect to the latter, it is reported that the sulfur tolerance of the noble metal catalysts can be greatly enhanced by using acidic supports such as zeolites, whereas less

I : Catalytic Hy drodearomat iza tion

33

acid supports such as Si02-Al203 can generate moderate sulfur resistance.28.3 1,37,41,114- 115 The studies performed by Vannice et al.69J16-117found that the activity of the Pt or Pd catalyst for benzene and toluene hydrogenation was much higher when loaded on Si02-Al203 compared with an A1203-supported catalyse. The enhancement of the activity was explained by the more acidic nature of the Si02-Al2O3 as a support. The exceptional performance of the PtPd catalyst in aromatic hydrogenation was interpreted in terms of the electron-deficient platinum species (isolated Pt cluster on Pd surface) in the resultant PtPd particle^.^ It has been reported in literature118y119 that the high sulfur tolerance and higher activity of zeolite-supported Pt or Pd catalysts arises from the formation of an electron-deficient cluster of metal particles, Pt6+ or Pds+ upon the hydrogenation of the aromatic compounds. The close contact between the strong acidic support and the small cluster of Pt or Pd atoms makes it possible for the electron to be withdrawn from the noble metal thereby creating an electron-deficient metal particle. 120~121In the case of zeolite-supported bimetallic PtPd catalysts, it seems that the molar ratio of the PtRd is seldom greater than unity. Also the role of the acidic support is to modify the active site, which is the noble metal. The catalytic cracking reaction of the acidic support, under hydrogenation conditions, is undesired as this lowers the selectivity towards the desired hydrodearomatization reaction. The hypotheses that the deactivation effect is related to acidic support has been proven by using a less acidic carrier. In these studies, the Si02-Al203 carrier was replaced by the more acidic support such as zeolites in order to minimise the effect of the cracking activity, which accelerate the catalytic deactivation by coke and results in enhanced yields of naphtha and g a ~ . ~ . ~ The sulfur resistance and saturation activity of the aromatic catalysts was improved by using MCM-41 materials as a support for noble metal catalyst^.^^ The higher saturation activity of naphthalene on Pt/MCM-41, in comparison with the Pt/A1203,was attributed to the medium-strong acidity of the MCM41 support and its higher metal d i s p e r s i ~ n . ~ ~ Different s t u d i e have ~ ~ ~been ~ ~devoted ~ to the nature of the active phase on the supported mono-noble metals such as ruthenium or rhodium in hydrodearomatization. The ruthenium or rhodium sulfide phase was identified as the active phase for aromatic hydrogenation. Aromatics hydrogenation in the presence of H2S shows that a weak Pt (or Pd)-S sulfide bond is formed.2 It was further found that the synergetic effect observed on the bimetallic PtPd was not as significant as those achieved with typical CoMo catalysts.2 The observed activity in Figure 1 indicates clearly the absence of synergetic effect between the bimetallic Pt and Pd. However, the activity results revealed the presence of the additional effect from the bimetallic combination of PtPd. The transition metal catalysts used in hydrodearomatization, the role of

34

Catalysis

preparation, the nature of the metal precursor and support and the ultimate dispersion of the active metal phase and their bearing on the catalytic data were carefully reviewed by Krylov et aZ.32 3

Industrial Aspects

The aim of this section is to illustrate the importance of aromatic saturation in the refining industry. There were a number of commercial processes having aromatic saturation as their major objective before the growing interest in removal of aromatics in diesel began in 1991. For example, aromatic saturation is used in the production of kerosene and jet fuel for smoke point i m p r ~ v e r n e n t . ' ~ ~A- ' ~ high ~ smoke point due to a low aromatic content improves the quality of a flame, producing a greater flame without smoke and d e p 0 ~ i t s .A l ~high ~ smoke point reduces the effects of the radiation from the jet flame on exposed mechanical parts in a jet engine. The removal of aromatics is essential in the production of lubricants and solvents to both control the technical performance of the product and the impact on health and environment. A low aromatic content improves the viscosity index (VI) of engine oil. The lower the VI, the lower the variation of the viscosity with temperature.126 In 1993 the European Community decided to label lubricants based on the DMDS extractable content of the lubricant according to the Institute of Petroleum method IP 346. The polyaromatics in the lubricant are extracted in to a DMDS phase and the 3% limit for the DMDS extractable content is based on a correlation with skin cancer in skin painting tests on mice. 127 In the production of food/medical grade oils, a complete reduction of aromatics according to the demands in the FDA, DAB or BS pharmacopoeia is needed. The production of medical white oil represents the most severe process condition used in hydrorefining. The first extraction stage and the first hydrotreating stage remove the major part of the nitrogen, sulfur and the aromatics. The intermediate product will meet the demands on the less refined product technical white oil. A low space velocity, high pressure and noble metal catalyst is needed in the second hydrotreating stage to overcome the thermodynamic limitations for the reduction of the remaining aromatics in the medical white oil. The production routes for lubricants and food/medical grade oil by extraction and acid/clay treatment have gradually been replaced by catalytic hydroprocessing routes. 128-130 It is an established practice in fuel refining to pretreat FCC feed stocks. One of the main objectives in the hydrotreating of FCC-feed stock is to reduce sulfur to minimize sulfur oxide release during regeneration of the catalyst and to meet the sulfur specification of the products. A more severe hydrotreating of the FCC-feed and a change of the catalyst from a CoMo catalyst to a catalyst containing NiMoP, a higher pressure, and a lower space velocity will increase the reduction of the nitrogen and aromatic content in the FCC-feed. The hydrogenation of aromatic and nitrogen components in FCC feed stocks

I :Catalytic Hydrodearomat iza t ion

Table 3

35

Swedish and Californian diesel specijications

Class

Class I

Sweden Class 11

Class 111

California >50 000 bpsd

Density, kgm-3 Sulfur, ppm Nitrogen, ppm Aromatics, vol% Cetane index Cetane number IBP, "C FBP, "C

800-820 10

800-820 50

800-830 500

5 50

10 47

25 47

180 300

180 310

180 330

830-860 500 10 10 48 170-215

305-350

decreases the coke deposited on the catalyst and improves the selectivity to more valuable products.1279131-132 Aromatic saturation is also desired in residue desulfurization for transportation fuels and in hydrocracking to increase the hydrogedcarbon ratio. Hydrogenation of aromatics is desired in the hydrotreating of straight run gas oil to improve the cetane number in diesel, and is even more necessary because the use of other low quality sources of various diesel feedstocks as straight run gas oil, visbreaker gas oil, coker gas oil and light cycle oil is increasing.133 The Swedish (1991) and the California (1993) authorities were the first to introduce legislation limiting the aromatic content in diesel. Table 3 shows the Swedish and Californian diesel specifications. The motivation for the legislation is that the reduction of the aromatic content and sulfur content results in reduced particulate and NO, emissions from the diesel engine. The specifications for fuel oil and diesel will become even more stringent for Europe and USA. 134-135 Asian countries are already moving towards similar tighter specifications.136 Changes in fuel specifications have been the dominant force for the development and introduction of new catalysts and processes for hydrotreating. Several new processing schemes have been introduced since 1991. Stricter specifications will continue to require improvements in performance. As highlighted above, thermodynamic considerations will limit the use of higher reaction temperatures in deep hydrogenation of aromatics. In current commercial processes for deep hydrogenation of aromatics, a two-stage process is used. In the first stage the sulfur and nitrogen contents are reduced by using a NiMo catalyst on alumina. Some aromatic saturation does also occur during this first step, but the main objective in the first step is to reduce sulfur and nitrogen. This allows for the use of a noble-metal catalyst at low reaction temperatures in the second stage after an intermediate gas separation and stripping in between the two stages.137Examples of licensors of this twostage technology are Shell, Topsoe, Lummus-Criterion and IFP. An option with the Synsat process from Criterion-Lummus is to remove the H2S and NH3 containing by-product gases within a single reactor. The Synsat 1339136~138-140

Catalysis

36 Fresh Feed

~

Recycle gas Make-upH2

Recvcle

Gas to VapodLiquid Separation and Product to stripping

Reactor 1ene;th

Figure 2

(a) SynSat reactor; (b) temperature and hydrogen partial pressure proJles

reactor, shown in Figure 2, also includes an optional counter-current gas flow in the bottom catalyst bed to further reduce the H2S content, increase the hydrogen partial pressure and maintain the temperature profile. 38 The catalysts commonly used in the second stage for aromatic saturation is a sulfur-resistant Pd-Pt, noble-metal catalyst on an acidic zeolite. Much attention has been focused on ways to increase the sulfur tolerance of noble metal catalysts. l4 A comparative in 1992 by Cooper et al. shows that a two-stage operation with an inter-stage stripper and sulfur-tolerant noble metal catalyst is the optimum for deep aromatic saturation. It has to be mentioned that diesel meeting the tightening specifications can also be obtained by hydrocracking or mild hydrocracking where the catalyst also contains an acidic function as well as the hydrogenating-dehydrogenating function.143The acidic function introduces a shift in boiling range and a ring opening. The ring opening results in a further improvement of the cetane number after a prior saturation of the aromatic ring. Hydrocracking is outside 3

339

1: Catalytic Hydrodearomatization

37

the scope of the present paper but there are some processes for deep aromatic saturation including an option for ring opening, boiling range shift or/and a cold flow improvement, e.g. the Akzo-Fina CFI and the SynShift processes.

4

Summary and Conclusion

In this paper we have reviewed the literature regarding the state of aromatic hydrogenation processes including the kinetic, thermodynamic, catalytic chemistry and industrial aspects. It is possible to produce a diesel fuel with a low aromatic content, high cetane number, uncoloured and with a low sulfur content. The low sulfur content is a necessity since the effective dearomatization catalysts are sensitive to sulfur. In the recent years several solutions to the severe sulfur deactivation of the dearomatization catalyst have been presented. However, the sulfur tolerance has to be further increased. The interesting temperatures are around 340 "C. At higher temperatures, the equilibrium limits the yield. At 370 "C, uncatalytic hydrogen donor reactions are possible and aromatics are formed from hydrogenated products. Higher temperatures can be used, depending on the hydrogen partial pressure. The higher the pressure the higher the temperature that can be used. If a more active catalyst is found, a lower temperature could be used, making it possible to use lower hydrogen partial pressure. This would lead to large savings in the production of low aromatic compounds. In the near future, it would be of interest to develop more basic knowledge on the industrial process. This will reduce the cost for dearomatization. There is a need for experiments with more industrial relevant model compounds and to find equilibrium data for model compounds under industrial conditions. There is also a need for experiments with industrial feeds, e.g. feedstocks to FCC and light cycle oils from FCC that can be used as diesel fuel. This could achieve a valuable increase of the cetane number, especially in the USA. Other interesting oils are base oils for lube stock. The sulfur resistance by using a zeolite carrier has been demonstrated. However, it is of great interest to increase this resistance or to develop completely new concepts since most feedstocks do contain significant levels of sulfur. Deep desufurization before dearomatization is costly. The need for more knowledge of selective hydrogenation 1 benzene --+ cyclohexene (not cyclohexane). 2 competitive reduction of benzene in gasoline by hydrogenation of a light fraction from catalytic reformer to reduce benzene in gasoline. 3 hydrogenation of specific aromatic compounds in competition with other aromatics in the feed. 4 developing catalysts with a high hydrogenation activity while at the same time maintaining strong resistance to poisoning of sulfur in the form of hydrogen sulfide and benzothiophenes.

Catalysis

38

5 !studying the influence of possible poisons like hydrogen sulfide, ammonia,

and water on the kinetic of hydrodehydroaromatization on one ring and two ring and multiple aromatic model compounds (the major portion of the aromatic content in diesel feedstocks) in the production of low aromatic diesel fuels. 6 more research to develop both sulfur- and nitrogen-resistant catalysts, because in industrial feeds, several heteroaromatics are present. 7 studying new methods for hydrogenation of aromatics like under superzritical conditions. I

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

14 15 16 17 18

19 20 21 22 23 24

References H. Topsare, B.S. Clausen and F.E. Massoth, in Hydrotreating Catalysis, Springer, Berlin, 1996. T. Kabe, W.H. Qian, Y. Hirai, L. Li, and A. Ishihara, J. Catal., 2000,190, 191. T. Fujikawa, K. Idei, T. Ebihara, H, Mizuguchi and K. Usui, Appl. Catal., 2000, 192, 253. R.M. Navarro, B. Pawelec, J.M. Trejo, R. Mariscal and J.L.G. Fierro, J. Catal., 2000,189,184. M.A. Arribas, F. Marquez and A. Martinez, J. Catal., 2000,190,309. H. Takagi, T. Isoda, K. Kusakabe and S. Morooka, Energy Fuels, 1999,13,1191. T. Fujikawa, K. Tsuji, H. Mizuguchi, H. Godo, K. Idei and K. Usui, Catal. Lett., 1999,63,27. W-Q. Liu, L. Zhao, G-D. Sun and E-Z. Min, Catal. Today, 1999,51,135. S . Xie, M. Qiao, H. Li, W. Wang and J-F. Deng, Appl. Catal. A, 1999,176,129. H. Yasuda, N. Matsubayashi, T. Sat0 and Y. Yoshimura, Catal. Lett., 1998, 54, 23. P. Kluson, J. Had, Z. Belohlav and L. Cerveny, Appl. Catal. A , 1997,149, 331. S.C. Korre, M.T. Klein and R.J. Quann, Ind. Eng. Chem. Res., 1995,34, 101. J.J. Llano, R. Rosal, H. Sastre and F.V. Diez, J. Chem. Technol. Biotechnol., 1998,72, 74. A.V. Spare and B.C. Gates, Ind. Eng. Chem. Process Des. Dev., 1981,20,68. T.C. Huang and B.C. Kang, Ind. Eng. Chem. Res., 1995,34,2349. T.C. Huang and B.C. Kang, Ind. Eng. Chem. Rex, 1995,34,2955. W.H. Wiser, S. Singh, S.A. Qader and G.R. Hill, Ind Eng. Chem. Prod Res. Dev., 1972,9, 350. W.H. Wiser, DOE Report DEO/ET/14700-9, Department of Energy, Washington DC, 1982, pp. 39-41. J. Shabtai, L. Velusmany and A.G. Oblad, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1978,23, 107. C.S. Huang, K.C. Wang and H.W. Haynes, in Liquid Fuels from Coal, ed. R.T. Ellington Academic Press, New York, 1977, pp. 63-78. M.J. Girgis and B.C. Gates, Ind Eng. Chem. Res., 1994,33, 1098. A.T. Lapinas, M.T. Kelin, B.C. Gates, A. Macris and J.E. Lyons, Ind Eng. Chem. Res., 1987,26, 1026. H.P. Stephenes and R.J. Kottenstette, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1983,28, 161. M.J. Girgis and B.C. Gates, Ind. Eng. Chem. Rex, 1994,33, 2301.

1:Catalytic Hydrodearomatizat ion

25 26

27 28 29 30

31 32 33 34 35 36 37 38 39 40 41 42 43 44

45 46 47 48 49 50 51

52 53 54 55

56

39

R. Rosal, V.F. Diez and H. Sastre, Ind. Eng. Chem. Res., 1992,31, 1007. C.S. Nessel, D.T. Coker, A.G. King and D.L. Mumford, in Symposium on Worldwide Perspectives on the Manufacture, Characterization and Application of Lubricant Base Oils presented before the Division of Petroleum Chemistry, Inc. 213th National Meeting, ACS, San Francisco, CA, 1997, April 13-17, pp. 25125. P. Grange and X. Vanhaeren, Catal. Today, 1997,36,375. H. Yosuda and Y. Yoshimura, Catal. Lett., 1997,46,43. B.H. Cooper, A. Stanislaus and P.N. Hannerup, Am. Chem. SOC.Div. Petrol. Chem., Preprints, 1992,37,41. P. Sogaard-Andersen, B.H. Cooper and P.N. Hannerup, in AM-92-50, NPRA Annual Meeting, March 1992. B.H. Cooper and B.B.L. Donnis, Appl. Catal. A., 1996,137,203. O.V. Krylov and M.D. Navalikhina, Russ. J. Appl. Chem., 1997,70, 1236. E. Furimsky, Appl. Catal. A , 2000,199, 147. E. Furimsky and E.F. Massoth, Catal. Today, 1999,52,381. C. Moreau and P. Geneste, Factor affecting the reactivity of organic model compounds, in Hydrotreating Reactions, Theoretical Aspects of Heterogeneous Catalysts, ed. J. Moffat, van Nostrand Reinhold, New York, 1990, pp. 256-306. M.J. Girgis and B.C Gates., Ind. Eng. Chem. Res., 1991,30, 2021. A. Stanislaus and B.H. Cooper, Catal. Rev. Sci. Eng., 1994,36,75. M. Kouzu, K. Koyama, M. Oneyama, T. Aramaki, T. Hayashi, M. Kobayashi, H. Itoh and H. Hattori, Fuel, 2000,79,365. M.A. Lure, 1.2. Kurets, D.F. Kushnarev, A.A. Malyuchenko and F.K. Schmidt, Kinet. Catal., 2000,41, 57. M.A. Keane and P.M. Patterson, Ind. Eng. Chem. Res., 1999,38, 1295. C. Song, Chemtech, March, 1999, p. 26. W. Qian, Y. Yoda, Y. Hirai, A. Ishihara and T. Kabe, Appl. Catal. A, 1999, 184, 81. A-G.A. Ali, A.L. Ali, S.M. Aboul-Fotouh and A.K. Aboul Gheit, Appl. Catal. A , 1998,170,285. S.N. Chaudhuri, S.K. Nath and D.S. Majumdar, Studies in Surface Science and Catalysis Recent Advances in Basic and Applied Aspects of Industrial Catalysis, Vol. 113, p. 793, Elsevier, Amsterdam, 1998. S.A. Ali, Fuel Process. Technol., 1998,55,93. J.R. Chang and S.L. Chang, J. Catal., 1998,176,42. T. Halachev, R. Nava and L. Dimitrov, Appl. Catal. A , 1998: 169,111. M. Yamamoto, 0.Nishimura and T. Kotanigawa, Appl. Catal. A , 1998,174,41. B. Dhandapani, T. St. Clair and S.T. Oyama, Appl. Catal. A , 1998,168,219. M . Machida, Y. Takeda and H. Hattori, Sekiyu Gakkaishi - J. Jpn. Petroleum Insti., 1998,41,285. P. Betancourt, A. Rives, R. Hubaut, C.E. Scott and J. Goldwasser, Appl. Catal. A , 1998,170,307. Y.W. Chen and C.P. Li, J. Chin. Inst. Chem. Eng., 1997,28,375. H. Bergem, E.A. Blekkan and A. Holmen, Hydrotreat. Hydrocrack. Oil Fractions, 1997,106,391 C. Song and A.D. Schmitz, Energy Fuels, 1997,11,656. J.L. Lemberton, M. Cattenot, V. Kougionas, M. Mhaouer, J.L. Portefaix, M. Breysse and G. Perot, Hydrotreat. Hydrocrack. Oil Fractions, 1997, 106, 529. M.V. Landau, S.B. Kogan, D. Tavor, M. Herskowitz and J.E. Koresh, Catal. Today, 1997,36,497.

40

Catalysis

57 58 59

J.R. Chang, S.L. ChangandT-B. Lin, J. Catal., 1997,169,338. A. Coma, A. Martinez and V. Martinez-Soria, J. Catal., 1997,169,480. M. Breysse, M. Catteno, V. Kougionas, J.C. Lavalley, F. Mauge, J.L. Portefaix and J.L. Zotin, J. Catal., 1997,168, 143. B. Moraweck, G. Bergeret, M. Catteno, V. Kougionas, C. Geantet, J.L. Portefaix, J.L. 2 0 t h and M. Breysse, J. Catal., 1997,165,45. Y-W. Chen and C. Li, J. Chin. Inst. Chem. Eng., 1997,28,375. S . Ishigaki and S . Goto, J. Chem. Eng. Jpn., 1997,30,60. T. Kotanigawa, M. Yamamoto and T. Yoshida, Appl. Catal. A , 1997,164,323. M.A. Keane, J. Catal., 1997,166, 347. S . Smeds, D. Murzin and T. Salmi, Appl. Catal. A , 1997,150, 115. C-A. Jan, T-B. Lin and J-R. Chang, Ind. Eng. Chem. Res., 1996,35,3893. N. Kimbara, J-P. Charland and M.F. Wilson, Ind. Eng. Chem. Res., 1996, 35, 3874. S. Toppinen, T-K. Rantakyla, T. Salmi and J. Aittamaa, Ind. Eng. Chem. Res., 1996,35,4424. D. Poondi and M.A. Vannice, J. Catal., 1996,161,742. S.J. Su, S.F. Mao, C.W. Chen, M.Y. Huang and Y.Y. Jiang, Polym. Adv. Technol., 1996,7,707. M. Mhaouer, J.L. Lemberton and G. Perot, Catal. Today, 1996,29,241. R.P. Dutta and H.H. Schobert, Catal. Today, 1996,31,65. K.M. Reddy and C. Song, Catal. Today, 1996,31,137. H.S. Joo and J.A. Guin, Fuel Process. Technol., 1996,49, 137, S.D. Lin and C. Song, Catal. Today, 1996,31,93. L.M. Tang, M.Y. Huang and Y.Y. Jiang, Chin. J. Polym. Sci., 1996,14, 199. C.S. Song and A.D. Schmitz, Meeting Abstract ofACS, 1996, ISSN 0065-7727. N. Marchal, S. Mignard and S . Kasztelan, Catal. Today, 1996,29,203. D.C. Braekman, A. Fontana, P. Laurent and P. Lolivier, Fuel, 1996,75, 579. R. Sen, R. Sumathy and C.N.R. Rao, Theochem. - J. Mol. Struct., 1996, 361, 211. ISSN 0166-1280. S. Smeds, T. Salmi and D. Murzin, Appl. Catal. A , 1996,145,253. S . Smeds, D. Murzin and T. Salmi, Appl. Catal. A , 1996,141,207. M. Yumoto, S.G. Kukes, M.T. Klein and B.C. Gates, Ind. Eng. Chem. Res., 1996,35,3203. F. Fajardie, J-F. Tempere, G. Djega-Mariadassou and G. Blanchard, J. Catal., 1996, 163, 77. S. Toppinen, T.K. Rantakyla. T. Salmi and J. Aittamaa, Ind. Eng. Chem. Res., 1996,35, 1824. S. Yang and L.M. Stock, Energy Fuels, 1996,10, 1181. T.C. Huang and B.C. Kang, Chem. Eng. J., 1996,63,27. J.F. Chiou, Y.L. Huang, T.B. Lin and J.R. Chang, Ind. Eng. Chem. Res., 1995, 34,4277. T.B. Lin, C.A. Jan and J.R. Chang, Ind. Eng. Chem. Res., 1995,34,4284. A. Coma, M. Iglesias and F. Sanchez, Catal. Lett., 1995,32, 313. S.C. Korre, M.T. Klein and R.J. Quann, Ind. Eng. Chem. Res., 1995,34, 101. T.C. Huang and B.C. Kung, Ind. Eng. Chem. Res., 1995,34,1140. U.S. Gill, C.W. Fairbridge, B.A. Farnand and E.C. Castellanos, US Patent 5 457 252, 1995, Oct. 10. J.A.R. Van Veen, J.K. Minderhoud, J.G. Buglass, P.W. Lednor and L.T. Thompson, Mat. Res. SOC.Symp. Proc., 1995,368,51.

60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94

1 :Catalytic Hydrodearoma t iza t ion 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114

115 116 117 118 119 120 121 122 123 124 125 126 127

41

J. Leglise, J.M. Manoli, C. Potvin, G. Djega-Mariadassou and D. Cornet, J. Catal., 1995,152,275. N. Krishnankutty and M.A. Vannice, J. Catal., 1995,155, 327. M. Guisnet, C. Thomazeau, J.L. Lemberton and S. Mignard, J. Catal., 1995, 151, 102. Div. Fuel Chem. Prepr., A.D. Schmitz, G. Bowers and C. Song, Am. Chem. SOC. 1995,40, 930. W-C. Lai and C. Song, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1995,40,1018. S . Smeds, D. Murzin and T. Salmi, Appl. Catal. A , 1995,125,271. G.D.-Mariadassou, M. Boudart, G. Bugli and C. Sayag, Catal. Lett., 1995, 31, 41 1. M.J. Cook, A.R. Katritzky and P. Linda, in Adv. Heterocycl, Chem. 1974, 17, 255. T.L. Gilchrist, in Heterocyclic Chemistry, Wiley, New York, 1985, p. 37. A. Moyano and J.C. Paniagua, J. Org. Chem., 1986,51,2250. G.W. Wheland, in Resonance in Organic Chemistry, Wiley, New York, 1955, pp. 75,98-100,104. C. Moreau, C. Aubert, R. Durand, N. Zmimita and P. Geneste, Catal. Today, 1988,4, 117. J. March, in Advanced Organic Chemistry, Wiley, New York, 1985, p. 39-40. J.J. Rooney, J. Mol. Catal., 1985,31, 147. P. Kluson and L. Cerveny, Appl. Catal. A, 1995,128, 13. K. Sakanishi, M. Ohira, I. Mochida, H. Okazaki and M. Soeda, Bull. Chem. SOC. Jpn., 1989,62, 3994. K.G. Knudsen, B.H. Cooper and H. Topsfie, App. Cat. A., 1999,189,205. A.W. Weitcamp, Adv. Catal., 1968,18, 1. H. Greenfiel and N.Y. Ann, Acad. Sci, 1973,214,233. P. Gallezot, in Catalysis by Zeolites, eds. B. Imelik, C. Naccache, Y. Ben Taarit, J.C. Verdine, G. Coudurier and H. Praliaud, Elsevier Science, Amsterdam, 1980, Vol. 5, p. 227. Div. Petrol. Chem. H. Yasuda, T. Sat0 and Y. Yoshimura, Am. Chem. SOC. Prepr., 1997,42, 580. S.D. Lin and M.A. Vannice, J. Catal., 1993,143, 539. S.D. Lin and M.A. Vannice, J. Catal., 1993,143, 554. P. Gallezot, Catal. Rev. Sci. Eng., 1979, 20, 121. S.T. Homeyer, Z. Karpinski and W.M.H. Schatler, J. Catal., 1990,123,60. R.A. Dalla Betta and M, Boudart, in Proc. 5th Int. Congr. on Catalysis 1973, Vol. 2, p. 1329, North Holland. R.A. Dalla Betta, M. Boudart, P. Gallezot and R.S. Weber, J. Catal., 1981, 69, 5 14. G.H. Unzelman, AM-87-33, NPRA Annual meeting, March 1987. Arosat process, Lummus, Hydrocarbon Processing, Refinery Process Handbook, 1982, p.108. Smoke point improvement, IFP, Hydrocarbon Processing, Refinery Process Handbook, 1982, p. 109. J.F. Le Page, M. Derrien and A. Haddad, in Proceedings ofthe Seventh World Petroleum Congress, April 1967, Mexico. J.F. Le Page, Applied Hetrogeneous Catalysis, Editions Technip, 1987, pp. 357-466. Concawe report, 1994, 94/51.

42

Catalysis

128 P.J. Polanek, D.J. Artrip and G Kons, FL-96-112, NPRA Annual meeting, March, 1996. 129 G.L. Everett and A. Suchanek, AM-96-37, NPRA Annual meeting, March, 1996. 130 A. Sequeira, Jr, Lubricant base oil and wax processing, Chemical Industries, 1994, v. 60. 131 V.F. Bavaro, J.F. Pagel, G.K. Collins and M.L. Gray, Hydrocarbon Eng., October 1999, 50-52, 54-55. 132 P.H. Desai, C.W. Stanger, J.W.M. Sonnemans and F.W. van Houtert, AM-8333, NPRA annual meeting, March 1983. 133 J.P. van den Berg, J.P. Lucien, G. Germaine and G.L.B. Thielemans, Fuel Process. Technol., 1993,35, 119. 134 EC Directive 981701EC Directive, Official Journal L 350, 2811211998 p . 0058-0068. 135 Proposed heavy-duty engine and vehicle standards and highway diesel fuel sulfur control requirements, EPA420-F-00-022. 136 Oil Gas J., August 7, 1995, p. 63. 137 W.H.J. Stork, Hydrotreat. Hydrocrack. Oil Fractions, 1997, 106,49. 138 A.J. Suschanek and G.L. Hamilton, AM-91-35, NPRA Annual meeting, March, 1991. 139 A.J. Suschanek, D. Dave, A. Gupta, H. van Stralen and K. Karlsson, NPRA Annual meeting, March, 1993. 140 S. Kasztelan, Proceedings of the 14th World Petroleum Congress, John Wiley & Sons. 141 C. Song, Chemtech, March, 1999,26. 142 B.H. Cooper, A. Stanislaus and P.N. Hannerup, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1991,36,41. 143 J.K. Minderhoud and J.A.R. van Veen, Fuel Process. Technol., 1993,35, 87. 144 H.W. Homan Free, T. Schochaert and J.W.M. Sonnemans, Fuel Process. Technol., 1993,35, 111. 145 G.L. Hamilton and D. Dave, Hydrocarbon Eng., October 1997,56.

3

Heterogeneously Catalysed Oxidations for the Environmentally Friendly Synthesis of Fine and Intermediate Chemicals: Synergy between Catalyst Development and Reaction Engineering BY WOLFGANG F. HOELDERICH AND FELIX KOLLMER

1

Introduction

Environmental concerns and regulations have increased in the public, political and economical world over the last 15 years because quality of life is strongly connected to a clean environment. Also in recent years increasingly stringent environmental constraints have led to a great interest in the application of new catalytic oxidation methods for synthesizing fine and intermediate chemicals. Although the manufacture of bulk and finelintermediate chemicals have some common features, there are some typical differences which have important drawbacks on the environmental assessment of these processes (Figure 1). In the bulk chemical industry for manufacturing, e.g. ethylene oxide, formaldehyde, phthalic anhydride, environmentally unacceptable processes have already been replaced by cleaner catalytic procedures, whereas fine and intermediate chemicals are still widely produced via traditional stoichiometric oxidations. However, organic chemists still employ stoichiometric oxidants such as dichromatelsulfuric acid, chromium oxides, permanganates, periodates, osmium oxide or hazardous chlorine causing high salt freights and heavy metal containing dump unable to be recycled, as well as relatively expensive hydroperoxides, alkylperoxides and peroxycarbonic acids. In the case of bulk chemicals heterogeneously catalysed processes have been developed employing cheap and abundantly available molecular oxygen. Due to their thermal instability fine chemicals often must be produced in the liquid phase at moderate temperatures. They are generally complex and multifunctional and regio- and stereoselectivity also play an important role. The reactors of the choice are batch or semi batchwise operated multipurpose units. Bulk chemicals, mostly consisting of relatively small, thermostable molecules, can be produced in continuously operated fixed bed or fluidized bed reactors in the gas phase allowing much higher space time yields, thus minimizing investment costs. Especially in the case of fine chemicals this is due to the relatively small scale Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 43

Catalysis

44

Differences in oxidation reactions of bulk and finehtermediate chemicals Sustainable and green processes Molecular oxygen Small molecules Temperature stable Higher temperature Fixed bed reactor Gas phase One purpose unit Continuous production Good E-factor

Figure 1

versus versus versus versus versus versus versus versus versus versus

Traditional stoichiometric oxidants Inorganic/organic peroxides Bulkier molecules Temperature sensitive Lower temperature Slurry reactor Liquid phase Multi purpose unit Batchwise production Bad E-factor

Differences in the production of bulk andfinelintermediate chemicals

of the production (< 1000-10000 t/a), the complex synthesis routes and the short development time to meet market demands leading to tremendously higher weight ratios of unwanted by-products per target substance, so-called E-factors, (25-100) compared to manufacturing bulk chemicals (< 1- 9 . l Still this effect is tolerable from an economic point of view because of the high value added, but it is getting increasingly difficult to carry out industrial scale oxidations in such a manner. There is currently a general trend to develop clean and eco-efficient catalytic oxidation processes which minimize the generation of unwanted and harmful by-products. Because of their versatility, their ease of separation, the lack of corrosion problems, the long life time and regenerability, solid catalysts for heterogeneously catalysed direct oxidations in the liquid phase as well as in the gas phase can contribute to solutions of the above mentioned problems. These catalysts offer a wide variety of different sites for the activation of clean oxidants such as molecular oxygen, hydrogen peroxide and nitrous oxide, which in some cases is available as an off-gas. In the case of intermediate chemicals the realistic target is extensive use of molecular oxygen in gas-phase reactions, allowing cost and eco-efficient manufacturing as known from the production of bulk chemicals. For fine chemicals and specialties this will rather be the exception due to the reasons outlined above, but in this case the replacement of stoichiometric oxidants by H202, in liquid phase oxidations is a cheap and realistic alternative. The heterogeneous catalysts applied thereby can be oxometal or peroxometal species, supported noble metals, heteropolyacids, noble metal pyrochlore oxides, metal containing layered double hydroxides, immobilized complexes (ship-in-the-bottle concept) and, last but not least, metal substituted molecular sieves (for an extensive review see e.g. ref. 2). In the following we present some examples which illustrate that catalyst development closely joined with reaction engineering aspects lead to sustainable production methods of high value added chemicals.

2: Heterogeneously Catalysed Oxidationsfor Fine and Intermediate Chemicals

2

45

Activation of the Oxidant by Noble Metal: BASF Citral Process

Gas phase oxidation of ethylene to ethylene oxide (EO) over supported silver catalysts is one of the most prominent examples for the use of noble metals in oxidation catalysis and one of the important industrial scale processes (8-10 x lo6 da). Ethylene is converted to EO either with air (Scientific Design, UCC) or with mixtures of oxygen and methane (Shell process) in tube bundle reactors at temperatures of 200-300 "C with addition of inhibitors, such as 1,2dichloroethane or vinyl chloride. The yields per pass are approximately 50%. The catalyst is a silver loaded (7-20%) aluminium oxide with a very low specific area (<2m2/g) promoted with alkali salts.3 Thereby the oxygen is activated by silver; other oxidation processes using noble metals such as Pt, Pd, Au, Ir, Ru are also known.2 In the following we will describe a 'similar' process technology that can also be applied for regioselective oxidation of much larger and more temperature sensitive reactive molecules, such as isopren01.~9~ Citral is a very valuable intermediate for the production of a- and p-ionone, which have importance as fragrances possessing smell of violet and are used in perfumes and household cleaners or as building units for carotenoids and vitamin A. It can be isolated from fruits. However, this method is impractical because the demand is several thousand tons per year. The conventional industrial route uses P-pinene as starting material and chlorine as oxidant. In a five-step procedure P-pinene is converted to citral in low yields and chlorinated side products are formed. These drawbacks can be overcome by the new environmentally benign BASF citral process. The BASF route (Scheme 1) starts from cheaply available isobutene and formaldehyde. In this condensation reaction 3-methylbutenol (isoprenol) is formed. A part of this product is isomerized by shifting the double bond, and the other part is oxidized to the corresponding aldehyde. Afterwards a Cope rearrangement of these compounds takes place to form citral with more than 95% overall yield. The regioselective oxidation of isoprenol to isoprenal is a remarkable environmentally benign process step. Patent literature reports that alcohols can readily be converted to the corresponding aldehydes over supported and massive metallic copper, silver and gold catalysts and alloys

Scheme 1

BASFprocess for the production of citral

46

Catalysis

thereof (see ref. 4 and references therein). The best results have been obtained with coated inert shapes, e.g. silica spheres where metallic silver (> 10wt.%) has been applied by flame spraying. In a continuous flow fixed bed reactor at 300-600 "C and an extremely short residence time of about 0.001 s., undiluted isoprenol is oxidized in the presence of pure oxygen surprisingly almost without deep oxidation. The process is based on the use of a short fixed bed reactor temperature controlled by a salt melt in order to run the reaction more or less isothermally and with an ultra short residence time.2 These facts and an efficient quench allow the production of several thousand tons/a of citral in an extremely high selectivity, even though oxygen is used as oxidant at high temperatures in a gas phase reaction. Another approach to carry out oxidation reactions of sensitive compounds such as isoprenol to isoprenal could be the use of honeycomb structured carrier dopedhmpregnated with noble metals. In such cases also very short residence times can be achieved due to the almost complete lack of pressure drop. But such a new technology has to be developed in a way to overcome the difficulties which may occur in catalyst preparation, in temperature control etc. In conclusion this example demonstrates that even a sensitive chemical such as isoprenol can be oxidized in very high yields, and how catalyst development, process conditions and reactor design have to meet each other. Simply by changing the construction of a conventional multitubular reaction system and the modification of a silver catalyst the oxidation of an intermediate chemical can be carried out in a similar way to the production of a bulk chemical in an environmentally benign and sustainable way.

3

Activition of the Oxidant by Metal Oxide: Direct Oxidation of f%Picoline

Nicotinic acid is an important intermediate for pharmaceuticals and serves as a provitamin in food additives for animal feeding. It is produced via the LONZA process or the DEGUSSA process. The Lonza process is the oxidation of 2-methyl-5-ethylpyridine by using nitric acid. This process suffers under the separation of nicotinic acid, a high amount of harmful salts due to the neutralization of nitric acid, and NO, formation as well as the loss of valuable framework carbon in the form of CO2. A two-step reaction is the Degussa process: hydrolysis of P-cyanopyridine produced by the ammonoxidation of P-picoline.9 The disadvantages are the high investment costs (two reactors, distillation) and the low yield. Selective vapor phase oxidation of P-picoline catalysed by vanadium titanium oxide catalysts has been described in the past8-13 The temperature and the molar oxygen/water/ P-picoline-ratio were varied over a wide range from 25 to 450°C and between 40/80/1 and 41/470/1, and air was used as oxygen source. All authors describe a binary vanadiaharrier system as the essential basis for their catalysts using Ti02-carriers with a BET surface area of 10-50 m2/g impregnated with amounts of 1-10 wt.% V205.14-17 The BET

2: Heterogeneously Catalysed Oxidationsfor Fine and Intermediate Chemicals

Scheme 2

47

SimpliJiedreaction network for the gas phase oxidation of /3-picoline

surface of the final catalyst was kept below 50 m2/g. Further sources examining the vapor phase oxidation of organic compounds confirm that anatase has a positive effect on the vanadia layer although it is not active in this reaction itself (see, e.g. ref. 12). The by-products of this reaction are pyridine-3-carbaldehyde as reaction intermediate and C02, CO and HCN as thermodynamically stable products. 12-14 According to our experiments, deep oxidation mainly takes place via oxidation of the substrate and the desired product, whereas pyridine-3carbaldehyde proved to be remarkably stable.14Blank tests in the empty tube with P-picoline/H20 and nicotinic acidH20 mixtures showed conversions of up to 40% with 100% selectivity to COXand HCN, especially in the presence of high excesses of water. Pyridine-3-carbaldehyde was stable in the empty tube and converted to the corresponding acid in the presence of the catalyst. From these observations we developed the simplified reaction scheme (Scheme 2). The precise control of residence time will be crucial for this reaction in order to avoid the consecutive oxidation of nicotinic acid. The selectivity towards nicotinic acid could be tremendously improved by appropriate process design, thereby avoiding the parallel reaction. It is very important that the procedure is carried out by separate feeding of j3-picoline and water as well as the mixing of both components just at the catalyst bed and a rapid quenching of the product stream after leaving the catalyst bed.l 8 Concerning the catalyst development we used a series of Ti02-carriers of the anatase modification with different BET-surfaces, impregnating them with adequate amounts of vanadia (Table 1). The catalytic performance of those catalysts depicted in Figure 2 clearly states that the increase of original Ti02-carrier surface area enhances activity and selectivity of the catalyst at identical reaction conditions. First, due to the higher surface area of the carrier material, a higher amount of vanadia could be impregnated upon them. Nevertheless it has to be considered that other parameters of these catalysts were changed, too. Furthermore, the V205 and thereby the active sites can be much better dispersed on the larger surface of the catalyst carrier. This higher dispersion should be maintained, even though the BET surface of the final catalysts is drastically reduced (Table 1). Regarding the differing chemical composition of the Ti02-materials the way

Catalysis

48

Table 1

VlTi-oxide catalysts with various BET-surface of the carriers

Catalyst

Ti02-carrier

BET surface carrier [m2/g]

V20, content [wt.%]

BET surface catalyst [m2/g]

A B C

Ti02-P25 Anatase B Anatase C

50 130 265

273 17 19

25 65 40

100

, 300

1

A

Figure 2

B

C

Relation between variation of BET surface of the carrier and the yield of nicotinic acid. Reaction conditions: T = 265-275 "C, WHSV= 0.04 gl(gKa, h), 02/H20//3-Pic= 35/55-70/1 (molar)

they were produced has to be considered. The Ti02-P25 material has been produced by burning TiCb, yielding an anatase material with a slight amount of rutile formation. In contrast the Ti02 materials Anatase B and Anatase C were synthesized by the sulfate process, yielding a higher surface area and Brsnsted acid sites due to sulfate residue which was the subject of further investigations (Table 2). As the specific values confirm, the carriers can be compared only with respect to their sulfate content while the other relevant parameters remain constant. Figure 3 clearly demonstrates that the total sulfate content has not such a strong influence on conversion and selectivity as the actual surface area of the Ti02-carrier does. It can be observed that the reaction temperature where the maximum yield is obtained is shifted with a higher sulfate content from 275 "C to approximately 255"C, i.e. the higher the sulfate content the better the activity. These experiments show clearly that it is necessary to add P-picoline and water separately to the catalyst bed in order to avoid the parallel reaction to the deep oxidation products. This can be managed by appropriate reactor design. The main parameter which improves the performance is the original surface area of the Ti02-carrier7which might be related to the sulfate content. Thereby the Brsnsted acid centers promote the activation of oxygen resulting in lowering the optimum of the reaction temperature. Therefore it is possible to improve the yield of nicotinic acid to values up to 98%.'*

2: Heterogeneously Catalysed Oxidations for Fine and Intermediate Chemicals

49

Table 2

VITi-oxide catalysts with various sulfate contents of the support materials

Catalyst

TiO2-carrier

BET surface Ti01 [m2/g]

Sulfate content V20,- content [wt.%] [wt.-Yo]

C C*

Anatase C Anatase C*

275 260

095

19

40

1 9 5

22

35

100.0

BET surface cat. [m2/g]

I

240

250

260

270

280

290

300

TIT

Figure 3

4

Temperaturedependenceof Catalyst Cand C* WHSV = 0.04-0.05 gl(gKath), 02lH20@-Pic = 35I55l1 (molar)

Activation of the Oxidant by Lewis Acid Sites: Direct Hydroxylation of Benzene

There are several processes for the production of hydroxy-aromatics. For production of phenol the cumene process is mainly used nowadays. The disadvantage of this three-step process is the low conversion at every step in order to keep the selectivity high. At the last step, the decomposition of cumene hydroperoxide, equal molecular amounts of phenol and acetone are produced. The economical efficiency of this process is strongly dependent on the market price of the inevitable by-product acetone. Therefore efforts to develop a new route to phenol free of acetone are of interest. Oxidizing a mixture of cyclohexanonelcyclohexanol with nitric acid to adipic acid yields equimolar amounts dinitrogen oxide, which could serve as a cheap feedstock for direct hydroxylation of benzene. Phenol in turn can easily be converted to a mixture of cyclohexanonelcyclohexanolwhich again serves as feedstock for oxidation to adipic acid (Scheme 3). This process, first proposed by Solutia, formerly Monsanto, is a fine example of eco-efficiency and s~stainabi1ity.l~ Solutia announced a commercial plant to go on stream by the end of 1999 or the beginning of 2000,20though it seems this project is postponed. Since 1983 there has been a great deal of research on the hydroxylation of

Catalysis

50

0 Scheme 3

HNO, ---+

ccooH COOH

2 '

+ N,O

Integrated adipic acid phenol production

0

5

10

15

20

25

Time of steaminglh

Figure 4

The ratio of the band intensities representing pyridine adsorbed at Lewis acid sites andpyridine adsorbed at Brmsted acid sites respectively as a function of steaming time (steamed 550 "C,300A H 2 0 )

benzene with N 2 0 to p h e n 0 1 . ~ l -G.I. ~ ~ Panov et al. found a correlation of the amount of iron in the ZSM-5 zeolite and the conversion of benzene with N20 to p h e n 0 1 . ~ ~They - ~ ~ suggested that only the coordinatively unsaturated iron atoms in extra-framework iron hydroxide complexes, formed at the thermal treatment step after the synthesis of the zeolite, are catalytically active. The idea of our work was to create coordinatively unsaturated extra-framework aluminium (EFA) by mild steaming of H-[Al]ZSM-5 zeolites avoiding collapse of the zeolite structure. The Lewis acid sites created thereby should activate the N20 and catalyse the hydroxylation of b e n ~ e n e . ~ ~ ~ ~ ' The steaming of the H-[All MFI-zeolite having a Si02/A1203 = 28 occurs at 550 "C and a water partial pressure of 300 mbar. The period of steaming varied in order to find a dependency of steaming duration and catalytic properties. FTIR spectra of pyridine loaded samples of H-[All-ZSM-5 after various steaming times demonstrate the increase of the intensity of the band at 1455 cm-', which represents pyridine adsorbed at Lewis acid sites.30 From such measurements we derived a quantitative estimation about the ratio of the band intensities representing pyridine adsorbed at Lewis acid sites and pyridine adsorbed at Brransted acid sites respectively (Figure 4).

2: Heterogeneously Catalysed Oxidations for Fine and Intermediate Chemicals

$,

96

30

0

'5

51

92

90

! 0

1

I

I

5

10

15

CO 20

25

time of steamingh

Figure 5

Conversion (0) and selectivity treated samples

(a) to phenol obtained with hydrothermally

This ratio increases due to steaming and dealumination. After 3-7 h of steaming the further increase of this ratio is insignificant. For a longer steaming time the Lewis to Brarnsted ratio is rather constant with only a slight decrease after 7 h of steaming. This slight decrease of the ratio is probably caused by agglomeration of the extra-framework aluminium (EFA) to clusters exposing less Lewis acid sites than numerous smaller EFA species. Such an agglomeration was already observed as an effect of ~ t e a m i n g .These ~ ~ . ~FT-IR ~ investigations were supported by MAS NMR measurement^.^^*^ Such H[AlIZSM-S zeolite catalysts were used for the conversion of benzene (25 mol%) and N20 (75 mol%) to phenol at 350 "C and WHSY= 1 h-' (Figure 5). In the case of steamed samples the conversion increased from 15% to nearly 30% and the selectivities from 94% to 98% (Figure 5) in the same manner as the ratio of band intensities representing Lewis and Brernstedt acid sites (Figure 4). The increase of conversion as a function of steaming time can be explained with an increase of Lewis acid sites created by EFA. There is a direct correlation between the amount of EFA created (Figure 4) and the conversion observed (Figure 5). With respect to the mechanism, we believe the N20 adsorbs at the electron-rich nitrogen on the Lewis-acid site weakening the N20 bond and at a certain stage cleaving this bond. Also in the case of Fezeolite^^^-^^ and of Ga zeolites29the Lewis acid site can be responsible for the activation of N20. Despite the excellent activity and selectivity of these catalyst systems, all the research groups observed a rapid decline in activity due to coking, which is a serious obstacle for commercializing this process. Concerning the reaction engineering aspects, Solutia Inc. claims to have improved the lifetime of a MFI catalyst, enabling the performance of this reaction in an adiabatic fixed bed reactor operated with periodical regeneration by burning of coke deposits, but thereby the use of a high excess of benzene and expensive recyclization is needed.34 Another reactor concept could be a circulating fluidized bed unit including

Catalysis

52

continuous regeneration of the spent catalyst analogous to the FCC process. This concept enables excellent temperature control in the reactor as well as in the regenerator and can be run in a steady state mode thus facilitating the down stream process control, e.g. distillation columns. Rooks et al. propose a circulating process with riser-downer concept analogous to the maleic anhydride process of D ~ P o n t ,where ~ ~ , in ~ ~the riser reactor the catalyst is loaded with surface oxygen, which then reacts with benzene in the downer reactor. The advantage of this concept is the high level of safety by avoiding hot spots and explosive atmospheres. In conclusion it could be demonstrated that N 2 0 can be activated by Lewis acid sites. The right reactor system for the commercial process is still under discussion, but Solutia plans to run it in an adiabatic fixed bed.37-39 5

Activation of the Oxidant by Hz over a Bifunctional Catalyst: Epoxidation of Propylene by a Mixture of H2and O2

The direct oxidation of propylene to propylene oxide (PO) is one of the most challenging subjects in catalysis in order to avoid the disadvantages of the currently employed chlorohydrine process, ARC0 and SHELL processes, such as the formation of CaC12 contaminated with toxic chlorinated compounds and of the inevitable coupling products styrene and MTBE, respectively. The invention of titanium silicate-1 (TS-1) by Taramasso et al. opened a new route for the synthesis of propylene oxide according to the Enichem technology using hydrogen peroxide as oxidant Because of the relatively high and steadily changing cost of hydrogen peroxide this process has not been commercialized up to date. In order to overcome this economic obstacle the use of hydrogen peroxide generated in situ is currently under research. Different approaches are employed for the in situ formation of hydrogen peroxide. The first approach makes use of the conventional anthraquinone (AQ) process for the production of hydrogen peroxide by feeding propylene and oxygen to the oxidation stage of anthrahydroquinone (AHQ). Contacting AHQ with oxygen leads to the formation of hydrogen peroxide, which is then consumed in the oxidation of propylene to PO catalysed by TS-1.42 The second c o n ~ e p tis~based ~ ? ~ on the oxidation of propylene by a H2-02 gas mixture over a precious metal containing titanium silicalite in the liquid phase, i.e. the combination of the DuPont process for the production of H20245with the Enichem process.41In a third approach Haruta et al. use a highly dispersed Au/Ti02 catalyst in a gas-phase reaction.46 However, the reported propylene conversions are below 2% whereas the selectivity is very high (> 95%). Since the generation of hydrogen peroxide seems to be the rate determining step, we investigated opportunities to increase the PO yield by varying the precious metal species on a TS-1 catalyst, using a mixture of palladium and platinum for the in situ formation of hydrogen peroxide, studying the reduction procedure, the addition of alkali salts and the influence of the reaction procedure in batch and semi continuous manner r e s p e ~ t i v e l y . ~ ~ - ~ ~ .40741

2: Heterogeneously Catalysed Oxidationsfor Fine and Intermediate Chemicals

Table 3

53

Effects of the reduction methods on the PO-yield over 1% Pdl 0.1% Ptl TS-1Conditions: 0.2 g catalyst, 15 g MeOH, 5 g H20, 3 h, 43 "C, 10 gpropylene, 7 bar H2 (59 mmol), 15 bar N2 (138 mmol), 10 bar O2 (92 mmol) ~

~~

Reduction

PO yield p!]

N2,150 "C 5 ~01%H2 + 95 VO~YO N2,150 "C 5 ~ 0 1 %H2 + 95 ~ 0 1 %N2,300 "C H2,5OoC H2,300 "C Air, 500 "C/5 vol% H2 + 95 vol%N2, 50 "C

5.1 3.5 2.5 2.2 1.4 0.8

The effect of different reduction methods on the catalytic performance was studied by reducing Pd/Pt impregnated TS-1 under an atmosphere of H2, 5% H2 in N2 or N2 at temperatures ranging from 50 to 350°C. The resulting PO yields are displayed in Table 3. Two trends, one related to the gaseous medium used and one related to the temperature, can be observed. PO yield and PO selectivity are favored by reduction under a pure nitrogen atmosphere, i.e. these catalysts are autoreduced as the thermal decomposition of the precious metal amine complexes used for impregnation creates locally a reducing atmosphere. Such a reduction behavior was already found by W. Sachtler et al.? Pd2++ NH3 + PdO + 2H+ + O.5N2 + O.5H2 These autoreduced samples exhibited a peak in the PO yield of 5.1% at a reduction temperature of 150 "C. Adding hydrogen to the reduction medium causes a strong decrease of the PO yield and PO selectivity. Catalysts reduced with pure hydrogen gave the lowest yield and selectivity. More propane is formed because a high concentration of Pdo is formed under hydrogen r e d u ~ t i o n . ~PO * $formation ~~ is favored by 'mild' reduction conditions, i. e. no calcination step prior to reduction, reduction temperatures around 150 "C and a very low hydrogen concentration or no hydrogen at all (autoreduction). Though we were able to improve the PO yield by varying the reduction methods we could not suppress the propane formation sufficiently. The platinum loading of 0. lwt.% offered an explanation for the low PO selectivities as platinum is known for its hydrogenation activity. In order to verify this assumption we conducted a series of experiments over TS-1 catalysts with lwt.% palladium and varying platinum loadings: 0 wt.%, 0.01 wt.%, 0.02wt.%, 0.1 wt.%. The impregnated catalysts were autoreduced under nitrogen atmosphere at 150 "C. We found that by adding as little as 0.01 or 0.02 wt% Pt to the 1 wt% Pd/TS-1 catalyst the yield increased to 10.3 or 11.7%, re~pectively.~~ The increase in yield was accompanied by an increase in the selectivity, i.e. 41 .O% and 46.0%, respectively. However a platinum loading of 0.1 wt% leads to a decrease in yield and selectivity.

Catalysis

54

6o

Figure 6

c lo

t8

150°C

150°C

Pd(II) content and PO yield as a function of the reduction method, catalyst I wt.% Pd + 0.I wt.% Pt on TS-I,reaction conditions: 0.2 g catalyst, 43 “C, 15 g MeOH, 5g H20, H2IO2IN2IC,H,= 15:15:15:1 (molar)

An ESCA study was carried out to assess the oxidation state of the Pd clusters as a function of reduction methods and platinum loading. Binding energies of the Pt species could not be measured because the Pt signal of the platinum containing catalysts was too weak. The results of the ESCA study are compiled in ref. 49. Binding energies for three different Pd species were found. Binding energy (BE) and Pd species were correlated according to Neumand2 and Briggs and Seah.53Binding energies ranging from 335.3 to 335.5 eV were assigned to PdO, which is in agreement with the BE of Pdmetal(335.0 eV). The exposure to air shifted the BE to higher values. No species in the state of Pdo were found on two samples 1 wt% Pd + 0.1 wt% Pt/TS- 1 autoreduced under N2 at 150°C and reduced with 5 vol% H2/N2 at 150”C, respectively. Instead values for BE were in the range 336.1-336.2 eV, which is close to the BE for PdO. Since the presence of PdO cannot be explained by a calcination step we assume that this effect is due to a prolonged exposure of Pdo species to air. Every catalyst examined was covered by Pd species with BE ranging from 337.2 to 337.8 eV. The same BE value was observed for a sample impregnated with 1 wt% Pd that was not reduced prior to ESCA analysis. The BE value for this Pd species cannot be easily attributed to Pd binding energies found in the literature. For reasons of simplicity we denote this species as ‘Pd(I1)’. As such ‘Pd(I1)’ species were present in all catalysts, we use their fraction given in mol% to describe the oxidation state of Pd and relate it to the reduction method and the platinum loading. Figure 6 demonstrates the effect of the reduction method on the PO yield and the fraction of the ‘Pd(I1)’ species for TS-1 catalysts with 1 wt% Pd + 0.1 wt% Pt. Three catalysts were reduced at 150 “C with N2, 5 vol% H2/N2 or H2 and the fourth catalyst was first calcined under air at 500 “C and subsequently reduced with H2 at 250°C. The fraction of the ‘Pd(I1)’ species, that was found to be the highest for the autoreduced catalyst (56%), decreased drastically to 28% for the catalyst reduced with 5 vol% H2/N2 or H2. The calcined and

2: Heterogeneously Catalysed Oxidations for Fine and Intermediate Chemicals

55

40

20 5

T

lo O I

0

I

0.02

!

0.04

I

0.06

I

0.08

I 0

0.1

Pt contenthid.-%

Figure 7

Fraction of Pd(II) and yield of PO as a function of the Pt loading, Pd loading always 1 wt.%, reduction 150 "Cunder N2, reaction conditions: 0.2 g catalyst, 43 "C,15 g MeOH, 5 g H 2 0 , H2IO,IN2IC,H,= 15:15:15:1 (molar)

reduced sample contained the least amount of 'Pd(I1)' species (18%). As expected the more severe the reduction conditions the more Pd was reduced to Pdo. However, it is remarkable that the Pd species even on the calcined and reduced sample was not completely reduced to Pdo although calcination and reduction conditions should create a fully reduced catalyst. Concerning the effect of platinum the addition of little amounts of Pt has a strong impact on the oxidation state of Pd and is accompanied by an improvement of the catalytic performance (Figure 7), which is consistent with the results of the optimization of the reduction method. This observation is also in accordance to the results of Grosser et al. who found enhanced activity for H202 formation by adding low amounts of Pt to palladium catalyst^.^^ Higher loadings of Pt, however, lead to a saturation effect concerning the oxidation state of the Pd species, but cause decreasing PO yields because of the Pt-catalysed hydrogenation of propylene. The formation of H202 from H2 and 0 2 is favored by the addition of promoters like HCl, H2SO4 and NaBr or NaC1.54This effect can be observed in the direct synthesis of PO as well, whether or not the salt was dissolved in the reaction mixture or the catalyst was impregnated prior to the reaction.50 Among the different salts, especially alkali halides, e.g. NaBr, showed positive results. It was found that both the conversion of propylene and selectivity of PO show a maximum with respect to the salt concentration in the solution. In the case of the Pd/Pt/TS-1 catalyst impregnated with NaBr, the selectivity could be dramatically increased up to 85% and the conversion could be maintained at an almost constant level of 15-20% (Figure 8). However, these investigations have been carried out in a semi continuous procedure as explained in the following paragraph. Because of the simplicity of the experimental set-up most of the catalyst

Catalysis

56 loo

1

7-

80 .

without NaBr

Figure 8

0,1% NaBr

0,SK NaBr

1,0% NaBr

Influence of impregnation the catalyts with salt; conditions: 0.2 g Cat., 15 g MeOH, 5 g HZO, 43 “C,120 min;feed: 1.3 llh H2, 1.3 llh 0 2 , 1.4 llh N2, 0.1 llh Propene; 7 bar; catalyst: I wt??Pd + 0.1 wt?? PtlTS-I, reduced at 150 “C

development and optimization was done in a batch system. Running the reaction in such a manner the selectivity towards PO never exceeded 50%, because of the high initial pressure of propylene and H2, leading preferably to the formation of propane. Carrying out the reaction in a semi continuous setup, i.e. liquid phase and catalyst in a steady phase, continuous gas supply, however, allows isobaric conditions over the whole reaction time and thus lower reactant pressures compared to the batch system and provides more effective dispersion of the reactants in the liquid phase, minimizing mass transfer limitations. While obtaining comparable conversions of about 20%, the selectivity could be improved from 40% to 60% (Figure 9), i.e. a relative improvement in the yield of 50%. In conclusion, the physical characteristics and the catalytic performance of Pd/Pt/TS-1 catalysts strongly depend on the reduction method and the platinum loading. The formation of PO was found to be favored by a high fraction of the ‘Pd(I1)’ species whereas fully reduced Pd and large clusters favored propane production. The fraction of ‘Pd(I1)’ was increased by autoreduction of the tetraamine ligand of the noble metal in the absence of hydrogen in the reduction medium. Reduction temperatures above 150 “C, or calcination, lead to cluster agglomeration on the external TS-1 surface and thus to decreasing PO yields and PO selectivities. Adding minor amounts of platinum also drastically increased the fraction of ‘Pd(I1)’ species in comparison to the Pdo species. Since platinum also catalyses the hydrogenation of propylene to propane there is an optimum Pt loading to achieve maximum PO yields and selectivies in the range 0.01-0.02 wt% Pt. Addition of small amounts of alkali halides led to an improvement of PO yield because of the promoting and stabilizing effect for the formation of hydrogen peroxide. By applying a more suitable reactor design such as a semi continuous process the yield could be further improved. The next step must be to develop continuous reactor systems for direct synthesis of PO.

2: Heterogeneously Catalysed Oxidations for Fine and Intermediate Chemicals

Batch

Figure 9

6

-

semi-

57

cow.

Comparison batch and semi-continuous modus; conditions: 0.2 g Cat., 15 g MeOH, 5 g H20,43 "C,120 min; batch: 10 gpropene, 7 bar H2,lO bar 0 2 , 15 bar N2; semi-continuous: 1.3 llh H2, 1.3 llh 0 2 , 1.4 llh N2, 0.1 llhpropene, 7 bar; catalyst: 1% Pd + 0.01% PtlTS-1, reduced at 150 "Cunder N2

Activation of the Oxidant by Immobilized Homogeneous Complexes: StereoselectiveEpoxidation of Olefins

Due to the steadily increasing demand for optically pure compounds in the pharmaceutical and agrochemical field the use of chiral catalysts has become a powerful tool in synthetic organic c h e m i ~ t r yNumerous .~~ attempts have been made at the immobilization of homogeneous chiral catalyst^.^^-^^ In this regard, the 'ship-in-a-bottle' approach dating back to 1977 offers several advantages over homogeneous or conventional heterogeneous catalytic systems where the metal complex is attached to a solid surface by covalent or ionic b ~ n d .The ~ , main ~ ~ feature of the 'ship-in-a-bottle' catalyst is the hostguest interaction which is neither covalent nor ionic. The guest is retained in the zeolite matrix by restrictive pore openings and will, in principle, keep all properties of the homogeneous complex. The superiority of these catalysts to homogeneous systems is based on their easy separation from the reaction mixture and, thus, their recyclability and environmental compatibility.66 Furthermore, it is likely that the zeolitic host bestows size and shape selectivity to the catalyst as well as a stabilizing effect on the organometallic complex since, due to the site isolation of the single complexes, multimolecular deactivation pathways such as formation of poxo- or peroxo-bridged species will be rendered i m p ~ s s i b l e . However, ~ ~ . ~ ~ even a large-pore zeolite such as zeolite Y, whose structure consists of almost spherical 12 A supercages interconnected tetrahedrally through smaller apertures of 7.4 A in diameter, is limited regarding the size of guest molecules by the space available in these cavities, whereas the use of mesoporous molecular sieves, such as USY zeolites and MCM-41 suffers from leaching because the pores are open to the external surface. The (sa1en)manganese-catalysed oxidation of olefins is currently investigated by various groups, such as Jacobsen and c o - w o r k e r ~ . ~ ~Immobilization ~~~ 368969

58

Figure 10

Catalysis

Procedure to generate the novel host with mesopores completely surrounded by micropores

of the Jacobsen catalyst is an attractive target. However, its outstanding activity, selectivity and chiral induction are accompanied by several disadvantages, such as quick deactivation, difficult separation and salt formation due to the use of sodium hypochlorite as oxidant. Attempts have been made to immobilize the Jacobsen catalyst, but its entrapment in the zeolitic space has presented a problem so far, with the average faujasite supercage cavity being too small for such a spacious complex. Thus only complexes of the less bulky bissalicylidene-1,2-~yclohexanediamine have been successfully occluded in fa~jasites.~~?~~ A new approach has been developed in our institute for creating of mesopores surrounded by micropores in a faujasite matrix which are spacious enough to accommodate the Jacobsen ligand, (R,R’)-(N,N)-bis(3,5-di-tertbutylsalicy1idene)-1,2-~yclohexanediamine,but with narrow pore openings to In catalytic tests this catalyst the external surface to prevent showed promising results, e.g. for activity, selectivity and chiral induction in the epoxidation of a-pinene and (R)-(+)-limonene. For the generation of mesopores in the zeolitic framework NaX and NaY zeolites were subjected to the combined dealumination steps with Sic& and H 2 0 outlined in Figure 10. This dealumination procedure leads to mesopores that are completely surrounded by microporous space. During treatment with silicon tetrachloride, a dealumination method first reported by Beyer et al. ,71 the faujasite’s framework aluminium is isomorphously replaced by silicon maintaining the microporous structure. The reaction is self-terminating due to the precipitation of NaA1C14 in the outer pores of the zeolite crystal.72The temperature to conduct the reaction (523 K) is relatively mild. Therefore it can be assumed that only the outer layers of the zeolite crystal are dealuminated by the treatment with Sick. During the successive ion-exchange the chloro-aluminium complexes are extracted and the zeolite is converted into the ammonium form. The zeolite is then steam-dealuminated. After this hydrothermal treatment the amount of framework aluminium has considerably decreased to a high degree of dealurnination with a Si02lA1203 ratio ranging from 125 to 190 for both zeolites X and Y. Using this dealumination procedure highly dealuminated faujasite zeolites are obtained regardless of the parent material. Final treatment of the host material with hydrochloric acid removes some of the extra-framework

2: Heterogeneously Catalysed Oxidations for Fine and Intermediate Chemicals

R'

59

R'

salen-1: R' = H, R~= H

salen-3: R' = t-Bu, R2= t-Bu

salen-2: R'

= t-Bu, R2= t-Bu

salen-fi:R' = t-Bu, R2= Me

salen-3: R'

= t-Bu, R2= Me

salen-2:R' = t-Bu, R2= H

salend: R' = t-Bu, R2= H

Figure 11

Salen complexes based on (R,R ) -cyclohexanediamine

Figure 12

Salen complexes based on (R,R ) -diphenylethylenediamine

aluminium in order to make room for guest molecules. The overall procedure is described elsewhere in detail.70 To introduce the desired transition metal cation the zeolitic host was ionexchanged for 24 hours at 353 K with a solution of the corresponding transition metal salt. The ion-exchanged zeolite was then dried under vacuum at 823 K (heating rate: 1 Wmin) and cooled under inert atmosphere. The synthesis of the salen ligands in the mesopores was conducted at room temperature under an inert atmosphere. To the ion-exchanged zeolitic material was added an amount of the optically pure diamine in dichloromethane in slight excess (1.2 equiv.) to the metal content. After stirring for 24 hours the appropriate amount of aldehyde in dichloromethane was added to the slurry and stirring was continued for another 24 hours. The mixture was then transferred into an extraction thimble and was soxhlet-extracted with dichloromethane and toluene, respectively, until the solvent remained colorless. Figures 11 and 12 show the salen complexes that have been entrapped in the mesopores of the new host materials and identified with FTIR measurements, thermogravity, XRD, and wet a n a l y s i ~ . ~ ~ . ~ ~ In addition to investigating the transition metal salen complexes' location within the host material nitrogen adsorption experiments were conducted with the empty zeolitic host and with three different immobilized Mn salen complexes. The results are summarized in Table 4. In comparison to the host material there can be detected a decrease in the BET surface area and the micro- and mesoporous volumes with all three 'ship-in-a-bottle' catalysts. For the Mn(sa1en-1)SIB catalyst, which hosts the smallest among the three entrapped organometallic compounds, the decrease of microporous space is the most prominent, indicating that this complex is mainly located in the zeolite's micropores. In contrast to this, the Mn(sa1en-5) complex, the most

60

Table 4

Catalysis

Nitrogen adsorption investigations

Sample

Host Mn(salen-1)SIB Mn(sa1en-2)SIB Mn(sa1en-5JSIB

BET surface area

t-plot microporous volume

[m2/sl

[cm3/sl

BJH-rnesoporous volume [cm3/gl

522.2 393.0 396.2 373.7

0.21 1 0.153 0.158 0.166

0.113 0.102 0.098 0.064

COOH

(-)-a-pinenoxide

(-)-a-pinene

Scheme 4

Oxidation of ( - )-a-pinene

bulky of the guest molecules, seems to occupy mainly the mesoporous space of its host. According to the nitrogen adsorption data of the Mn(salen2)SIB catalyst, the guest occupies both micropores and mesopores of the host material, suggesting that major parts of the ligands, e.g. the tert.-butylsubstituted aromatic groups are located in the intrazeolitic openings into the microporous space. These catalysts have been tested in the stereoselective epoxidation of R-(+)limonene and ( - ) - ~ - p i n e n e . ~Here ~ only the epoxidation of (-)-a-pinene as depicted in Scheme 4 is considered. The oxidant applied in the reaction is somewhat similar to the one introduced by Mukaiyama et aZ.73 and was favored over the system used by Jacobsen and c o - ~ o r k e r sThe . ~ ~major benefit of this system is that undesirable salt formation can be avoided by the use of environmentally benign molecular oxygen at RT instead of NaOCl as oxidant at 0 "C.Due to its inert nature and the excellent solubility of molecular oxygen in fluorous solvents fluorobenzene has been chosen as solvent. The combination of a transition metal, an organic ligand and a reductant also creates an effective oxygenation system in an aerobic reaction.75 Figure 13 shows that the (salen-2) complexes of Fe and V and the Co(sa1en5) complex retained their catalytic properties upon entrapment in the host materials. In contrast to the Mn(sa1en-2) complex, which loses only some of its epoxide selectivity upon immobilization, the corresponding Co and Cr complexes show an additional decrease in stereoselectivity as well. Strikingly, the immobilized Co(sa1en-5) complex achieved with 100% conversion, 96% selectivity and 91% d.e., even better results in the epoxidation of (-)-a-pinene than its homogeneous counterpart. However, it is worth notifying that among the

2: Heterogeneously Catalysed Oxidationsfor Fine and Intermediate Chemicals

61

1oOOA 90.A 80%

7ooA

conversior

60.A

SClCCtiViQ

% 5ooA 40%

30.A

20% 1099

0%

Figure 13

Epoxidation of (-)-a-pinene in the presence of various homogeneous and immobilized transition metal (salen) complexes, * = homogeneous complex. reaction conditions: I0 ml CdH5F, 4.6 mmolpivalic aldehyde, 1.85 mmol ( -)-a-pinene, 25 mg catalyst, 30 bar 0 2 , RT

(salen-2) complexes neither the homogeneous nor the occluded Jacobsen complex catalysed the epoxidation of ( -)-a-pinene best. We can conclude that a comparison of the respective catalytic results of these new heterogeneous catalysts and their homogeneous counterparts showed that the entrapment of the organometallic complex was achieved without considerable loss of activity and selectivity. The immobilized catalysts are reusable and do not leach. The oxidation system applies only 0 2 at RT instead of sodium hypochloride at 0 "C. The use of pivalic aldehyde for oxygen transformation via the corresponding peracid results in the formation of pivalic acid which has to be separated from the reaction mixture which is at the present state a disadvantage. The best results so far - 100% conversion, 96% selectivity and 91% de - were achieved with the immobilized Cobalt(sa1en-5) complex in the epoxidation of (-)-a-pinene. 7

Activation of H202 over Solid Acid Catalysts: Baeyer-Vier-Oxidation of Cyclopentanone

The invention of titanium silicate-1 (TS-1) by Taramasso et al. opened a variety of liquid phase oxidations using aqueous hydrogen peroxide which has been reviewed extensively in the past (see, e.g. ref. 76). But such reactions can also be performed over a number of other heterogeneous catalysk2 An example is the acid-catalysed Baeyer-Villiger oxidation of cyclopentanone to 8-valerolactone. Traditional oxidizing agents for this type of reaction include

62

Catalysis 100

90 80 70 60 $ 50

40

30 20 10

0

Figure 14

Comparison of tested acid catalysts using dilute H202 as an oxidant; 1.05 equiv. H 2 0 2 , loading 87 mmol cyclopentanone per g catalyst, 70 "C, atmospheric pressure, data after six hours

peracids such as trifluoracetic acid,77perbenzoic acid78and rn-chloroperbenzoic acid.79 These reagents, however, present the disadvantage of having to be prepared by using highly concentrated hydrogen peroxide, which has been removed from commerce recently because of the dangers involved in transport and handling. Several alternative procedures fill this gap, for example oxidations with molecular oxygen, catalysed by metal complexes and using aldehydes as c o - o x i d a n t ~ .A~ ~ heterogenation *~~ of the reaction has been achieved using polymer-supported peracids as oxidants which can be prepared from commercially available polymer resins and hydrogen peroxide.81 Heterogeneous catalysis of the Baeyer-Villiger reaction with hydrogen peroxide is possible in the presence of zeolite catalyst, albeit in a reaction system consisting of an organic solvent and the co-partners.82Recent results, however, show that such reactions can also be performed using diluted hydrogen peroxide in an aqueous ~ y s t e m . ~ ~ ? ~ ~ Acidic functionality is regarded as most important for the Baeyer-Villiger oxidation according to the mechanism proposed by C~-iegee.~~ Therefore several types of acidic materials have been tested as catalysts, including acetic acid as a homogeneous catalyst, organic ion-exchange resins, such as Amberlyst@ 15 and Amberlyst@ XN1010, Nafion@/Silicacomposite materials and also various zeolitic materials, such as HZSM-5, HMOR, USY and HBEA with Si/Al ratios of 34, 8, 55 and 11 respectively (Figure 14). Besides HMOR, all the catalysts showed an improved yield compared to the blank test, the Amberlyst@ 15, Amberlyst@XN 1010 and HBEA being the most active. In a further series, Amberlyst@ 15 was found to be the most selective catalyst at a given conversion, i.e. a selectivity of 80% and 35% conversion. By further

2: Heterogeneously Catalysed Oxidationsfor Fine and Intermediate Chemicals

63

optimizing reaction conditions yields of up to 70% 6-valerolactone can be obtained . Monitoring conversion and selectivity as a function of reaction time and looking more closely at the by-products formed, a simplified reaction network was proposed.83 Besides 6-valerolactone, significant amounts of glutaric acid and dicyclopentylidene peroxide were found, whereas the amount of C 0 2 was negligible even under high temperatures and excesses of hydrogen peroxide. Oxidation of 6-valerolactone mainly yielded glutaric acid. Dicyclopentylidene peroxide was isolated from the reaction mixture and again used as substrate, showing no conversion to 6-valerolactone, thus indicating it was formed via a parallel pathway. It is assumed that electrophilic attack on the carbonyl group to be the initial step in both reaction pathways. Protonation or polarization of cyclopentanone by acid centers of the catalyst allow an attack by hydrogen peroxide or a peracid, forming an intermediate which then reacts to 6-valerolactone. Polarization of hydrogen peroxide could lead to an intermediate that can react in a dimerization to dicyclopentylidene diperoxide. Bearing in mind the excellent activity of sulfonic acid containing materials, persulfonic acid groups can result from reaction of hydrogen peroxide with the acid groups, allowing a Criegee mechanism to take place. In conclusion, cyclopentanone can be effectively converted to 6-valerolactone with a diluted aqueous hydrogen peroxide solution in a Baeyer-Villigertype reaction, i.e. without the use of organic solvents. Though zeolite beta also shows good results, the best catalyst was found to be Amberlyst@ 15, an organic ion exchange resin. This kind of reaction is simply carried out in a batch reactor and no extensive chemical engineering development work is needed.

8

Summary

Catalyst and process development has to be seen as a unit. The examples described above demonstrate the potential of sustainable and green processes for replacing traditional technology in intermediate and fine chemistry which are not acceptable from an ecological and economical point of view. Heterogeneous catalysts offer a variety of different sites for activation of nonstoichiometric (02, N20) or environmentally benign stoichiometric (H202) oxidants. Thus the strategies, needs and targets for the future catalyst and process development include: 0

0

0

Optimization of atomic efficiency and E-factor, 100% yield or selectivity is the target,' new oxidation processes using environmentally benign oxidants, e.g. 0 2 , H202, H2/02, N20, thus avoiding inorganic and organic peroxides as well as chlorine, catalyst and process development have to be seen as a joint challenge to be faced and solved in integrated way, e.g. extensive use of Mars-van Krevelen

64

0

0

0

0

0

0

0

0

0

Catalysis

principle via riserdowner concept (maleic anhydride production of DuPont), development of catalytic membrane, monolithic and wall reactors, reproducibility of oxidation catalyst preparation; this is especially true for oxide catalysts synthesized by precipitation methods, development of redox catalysts with acid and basic properties, i.e. oxidation reactions in the presence of materials with acid or basic or acidbase sites, activation of oxygen by hydrogen, i.e. in situ production of H202, development of multifunctional catalysis, i.e. combining several individual reaction steps in one step, development of direct synthesis routes saving energy and starting material, e.g. oxidation of aromatic compounds (‘OnePot’ reactions), Extensive use of by-products in other syntheses, e.g. use of the exhaust gas N20, Use of regenerable and biodegradable resources, such as sugar and starch, Direct C-H activation, i.e. using natural gas and alkanes, e.g. CH30H from CH4, CH3COOH from CH4 and C02, CH3COOH from C2H6,CH2=CHCN and CH2=CHCH0from C3Hs, Development of true chiral heterogeneous catalysts, i.e. immobilization of chiral homogeneous catalysts is not a real solution, Development of heterogeneous catalysts for treatment of polluted water, wet air oxidation.

In oxidation catalysis for the synthesis of fine and intermediate chemicals using heterogeneous catalysts there is still much more to be investigated than in acid base catalysis. 9

Acknowledgements

The authors express their sincere thanks to Aventis Research and Technology (former HOECHST AG), DEGUSSA AG and the state North-Rhine Westfalia for the generous financial support. 10 1 2 3

4 5 6 7

References R.A. Sheldon, Chem. Ind., 1997, 12. W.F. Hoelderich, Stud, SurJ Sci. Catal., 1993,75, 127. S . Rebsdat and D. Mayer, Ullmann’s Encyclopedia of Industrial Chemistry, electronic release, 6th Edn., Wiley VCH, 1998. W. Aquila, H. Fuchs, 0. Worz, W. Ruppel and K. Halbritter, BASF AG, US 6 013 843,2000. A. Stocker, 0. Marti, T. Pfammatter, G. Schreiner and S. Brander, Lonza, DE 2 046 556,1970. H. Beschke, H. Friedrich, K.-P. Miiller and G. Schreyer, Degussa AG, DE 2 5 17 054, 1975. T. Liissling, H. Schaefer and W. Weigert, Degussa AG, DE 1 948 715, 1969.

2: Heterogeneously Catalysed Oxidationsfor Fine and Intermediate Chemicals

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

49

65

S.T. Lundin and S. Jaris, DE 2 647 712, 1976. S. Jaris and S.T. Lundin, J. Appl. Chem. Biotechnol., 1977,27,499. R. Yokoyama and K. Sawada, US 3 803 156,1969. E. Alkaeva, T. Andrushkevich, G. Zenkovets and M. Makarenko, Boreskov Institute of Catalysis, EP 747 359, 1995. M. Gasior, J. Haber and T. Machej, Appl. Catal., 1987,33, 1. T.V. Andrushkevich, G.Y. Popova, E.M. Alkaeva, M.G. Marenko and G.A. Zenkovets, Russ. Chem. Ind., 1996,28, 25. G. Prim, D. Heinz and W.F. Hoelderich, Chem. Eng. Sci., submitted. J. Haber, Catal. Today, 1994,20, 11. P. Ruiz and B. Delmon, Catal. Today, 1994,20, 17. G. Busca, Catal. Today, 1994,20,61. D. Heinz, W.F. Hoelderich, S. Krill, W. Boeck and K. Huthmacher, J. Catal., 2000, 192, 1. A.K. Uriarte, M.A. Rodkin, M.J. Gross, A.S. Kharitanov and G.I. Panov, Stud. Surf: Sci. Catal., 1997,110,857. K. Tanabe and W.F. Holderich, Appl. Catal. A , 1999,181, 399, M. Iwamoto, J. Hirata, K. Matsukami and S. Kagawa, J. Phys. Chem., 1983,87, 903. E. Suzuki, K. Nakashiro and Y. Ono, Chem. Lett., 1988,953. M. Gubelmann and P.J. Tirel, Rhone-Poulenc Chimie, US 5 001 280, 1991. R. Burch and C. Howitt, Appl. Catal. A , 1993,103, 135. G.I. Panov, A.S. Kharitanov and V.I. Sobolev, Appl. Catal. A , 1993,98, 1. A.S. Kharitanov, et al., Zeolites, 1995, 15, 253. V.I. Sobolev, G.I. Panov, A.S. Kharitanov, V.N. Romannikov and A.M. Volodin, Kinet. Catal., 1993,34,887. G.I. Panov, V.I. Sobolev and A.S. Kharitanov, J. Mol. Catal., 1990,61, 85. M. Hafele, A. Reitzmann, D. Roppelt and G. Emig, Appl. Catal. A , 1997,150, 153. J.L. Motz, H. Heinichen and W.F. Holderich, Stud. Surf. Sci. Catal., 1997, 105, 1053. J.L. Motz, H. Heinichen and W.F. Holderich, J. Mol. Catal. A, 1998,136, 175. J. Dwyer, F.R. Fitch, G. Qin and J.V. Vickermann, J. Phys. Chem., 1982,86,4574. L.D. Fernandes, et al., Zeolites, 1994, 14,533. C.W. Rooks, A.K. Uriarte and M.J. Gross, Solutia Inc., US 5 892 132, 1999. R.M. Contractor, Appl. Catal. B, 1995,6, N3. R.M. Contractor, et al., Stud. Surf. Sci. Catal., 1994,82,233. Chem. Week, 1997,1, 11. Chem. Eng. Prog., 1997,93, 16. Chem. Eng. (Int. Ed.), 1998,105,27. M.G. Clerici, G. Belussi and U. Romano, J. Catal., 1991,129, 159. M. Taramasso, G. Peregro and B. Notari, Snamprogetti, US 4 410 501, 1983. P. Ingallina, M.G. Clerici, L. Rossi and G. Belussi, Stud. Surf: Sci. Catal., 1995, 92, 31. A. Sato, M. Oguri, M. Tokumaru and T. Miyake, Tosoh, JP 269 029, 1996. U. Muller, et al., BASF AG, DE patent application 4425672 Al, 1996. L.W. Grosser and J.-A.T. Schwartz, DuPont de Nemours, US 4832938,1989. M. Haruta and T. Hayashi, Shokubai, 1995,37,72. W. Laufer, R. Meiers and W. Hoelderich, J. Mol. Catal. A., 1999,141, 215. R. Meiers and W.F. Hoelderich, Catal. Lett., 1999,59, 161. R. Meiers, U. Dingerdissen and W.F. Holderich, J. Catal., 1998,176, 376.

66

Catalysis

50

W. Laufer, R. Meiers and W.F. Hoelderich, in Proceedings of the 12th International Zeolite Conference, ed. M.M. Treacy, B.K. Marcus, M.E. Bisher and J.B. Higgins, Material Research Society, Warrendale, Pennsylvania, 1999, p. 1351. W.M.H. Sachtler and Z. Zhang, in Advances in Catalysis, Vol. 39, ed. D.D. Eley and P.B. Weisz, Academic Press, New York, 1993, p. 129. M. Neumann, PhD Thesis, University of Frankfurt, 1995. D. Briggs and M.P. Seah, Practical Surface Analysis, Auger and X-Ray Photoelectron Spectroscopy, Wiley, Chichester, 1990. A.I. Dalton and R.W. Skynner, Air Products and Chemicals, US 4 379 778, 1983. R.A. Sheldon, Chirotechnology: Industrial Synthesis of Optically Active Compounds, Marcel Dekker Inc., New York. G.J. Kim and S.H. Kim, Catal. Lett., 1999,57, 139. J.A. Mayoral, J.M. Fraile, J.I. Carcia and J. Massam, J. Mol. Catal. A: Chem., 1998,136,47. P. Piaggio, et al., New J. Chem., 1998, 1167. I.F.J. Vankelecom, D. Tas, R.F. Parton, V. Van de Vyver and P.A. Jacobs, Angew. Chem., 1996,108, 1445. A. Corma, M. Iglesias, F. Mohino and F. Sanchez, J. Urganomet. Chem, 1997, 544,147. A. Corma, M.J. Sabater, A. Domenech, V. Fornes and H. Garcia, Chem. Commun., 1997,1285. C. Heinrichs and W.F. Hoelderich, Catal. Lett., 1999, 58,75. W. Kahlen, H.H. Wagner and W.F. Holderich, Catal. Lett., 1998,54, 85. G. Schulz-Ekloff, G. Meyer, D. Wohrle and M. Mohl, Zeolites, 1984,4, 30. B.V. Romanovskii and V.Y. Zakharov, Khim., 1977,18,142. P.A. Jacobs, R. Parton and D. de Vos, in Zeolite Microporous Solids: Synthesis, Structure, and Reactivity, Kluwer Academic Publishers, 1992, p. 555. K.J. Balkus Jr., A.K. Khanmamedova, K.M. Dixon and F. Bedioui, Appl. Catal. A , 1996,143,159. E.N. Jacobsen, in Catalytic Asymmetric Synthesis, ed. I. Ojima, VCH, New York, 1993, p. 159. T. Katsuki, J. Mol. Catal. A , 1996,113, 87. C. Schuster and W.F. Hoelderich, Catal. Today, 2000,60, 193. H.K. Beyer and I.M. Belenykaja, Stud. SurJ Sci. Catal., 1980,5,203. H.K. Beyer, P.J. Grobet and P.A. Jacobs, Zeolites, 1986,6,47. T. Mukaiyama, T. Yamada, K. Imagawa and T. Nagata, Chem. Lett., 1992,2231. E.N. Jacobsen, A.R. Muci and N.H. Lee, Tetrahedron Lett., 1991,32, 5055. J.-P. BCguC, F.F. Barbier, K.S.K.S. Ravikumar and D.D. Bonner-Delpon, Tetrahedron, 1998,54,7457. P.B. Venuto, Microporous Mater., 1994,2,297. W.F. Sager and A. Duckworth, J. Am. Chem. Soc., 1955,77,188. S.L. Friess and J. Farnham, J. Am. Chem. SOC.,1950,72,5518. G.R. Krow, Org. Reactions, 1993,43,251. C. Bolm, K. Schlinghoff and K. Weickhardt, Angew. Chem., Int. Ed. Engl., 1994, 33, 1848. C.S. Pande and N. Jain, Synth. Comrnun., 1989,19, 1271. C.D. Chang and S.D. Hellring, Mobil Oil Corporation, US 4 870 192, 1996. J. Fischer and W.F. Hoelderich, Appl. Catal. A , 1999,180,435. J. Fischer, G.P. Schindler, D. Arntz and W.F. Hoelderich, Degussa AG, DE 19745442, 1999.

51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84

3 Catalytic Hydrogen Generation from Methanol BY JOHAN AGRELL, BARD LINDSTROM, LARS J. P E ~ E R S S O N AND SVEN G. JARAS

1

Introduction

Hydrogen is used in vast quantities in the chemical industry for production of various bulk, fine and specialty chemicals, in food processing, for fuel production in refineries, in the steel industry and also directly as a fuel, e.g. in the space programmes. The largest portion of hydrogen in the world is manufactured at ammonia production units and consumed on site in the process. Other large consumers are the processes for methanol and hydrogen peroxide production. In order to meet stricter fuel quality standards there is a continuous growth in hydrogen demand in the petroleum industry. From being producers the refineries are now net consumers of hydrogen. The content of aromatics and sulfur in gasoline and diesel is decreasing.2 This combined with the expected future use of more heavy crude oils creates an increased need for hydroprocessing. Hydrodesulfurisation, hydrodearomatisation, hydrodenitrogenation and hydrodemetalisation are all processes that require large amounts of hydrogen. Due to the tightening air quality standards in the world this trend is going to be even more pronounced in the first decade of the new century. Hydrogen can be produced from both fossil and renewable sources. The largest quantities are manufactured from natural gas. However, in a future socalled hydrogen society it can be produced by electrolysis of water using solar energy. In this case it can clearly be viewed as a sustainable source of energy. Hydrogen is the cleanest fuel available and ideally produces only water during combustion, which makes it an interesting alternative for decreasing anthropogenic C 0 2 emissions. The most important driving force for using hydrogen in automotive applications is the potential of obtaining low emissions of hazardous compounds. Hydrogen can be used in internal combustion engines3 or in fuel cell engine^.^ However, storing hydrogen on board a vehicle poses many concerns regarding safety and handling and can affect customer acceptance in a negative way. Hydrogen can be stored as compressed gas at high pressures, as liquid at cryogenic temperatures, in metal hydrides or chemically in a hydrogen-containing compound. The last method is superior to the other three in terms of energy d e n ~ i t yMethanol .~ is one of the largest bulk chemicals in the world and an excellent hydrogen carrier, which can be readily converted into hydrogen at moderate temperatures. Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 67

68

Catalysis

This review focuses on various catalytic methods to generate hydrogen from methanol. We will present the different options available: decomposition, steam reforming and partial oxidation. A special discussion is devoted to combined reforming, which is a combination of steam reforming and partial oxidation. This process is especially interesting, since it can be selfsustaining regarding heat. When operated close to adiabatic conditions this process is normally referred to as autothermal reforming. The last part of the paper concentrates on methanol reforming for fuel cell applications. In this part we compare the different hydrogen generation techniques and give the requirements for a reformer in a fuel cell vehicle. A discussion of various gas clean-up technologies is included. Industrial activities are given special attention and the last section of this review is a summary of patents in the area. 1.1 Fuel Options for Hydrogen Generation. - There are several candidates for chemical storage of hydrogen. The most obvious ones are the petroleumderived fuels gasoline and diesel, which both have a well developed infrastructure. Natural gas, ethanol, liquefied petroleum gas (LPG), aviation and marine fuels are other alternatives. Natural gas has a high hydrogen-to-carbon ratio, but has the clear disadvantage of being a gas at ambient temperatures. Natural gas in liquefied form (LNG) or compressed form (CNG) are both associated with various problems. The petroleum-derived fuels require high reforming temperatures (800-900 "C) and, furthermore, there is a great risk of deactivating the catalyst by coking. The latter fact is due to the carboncarbon bonds in the fuel. Ethanol is an interesting candidate, since it can be produced from renewable sources and also has a high hydrogen content. However, the reforming reaction must be performed at higher temperatures compared to methanol6 and several undesirable by-products are f ~ r m e d . ~ Methanol is a primary alcohol with a high hydrogen-to-carbon ratio, which can be produced from renewables, such as biomass. Fuel methanol is sulfurfree. Furthermore, the molecule contains no carbon-carbon bonds and is rather easily converted to a hydrogen-rich gas.

1.1.1 MethanoZ Synthesis. Methanol is produced from synthesis gas, which is primarily manufactured via steam reforming of natural gas. The first largescale process was started up by BASF AG in Germany in 1923, using a sulfurresistant catalyst composed of zinc oxide and chromia.8 The process operated at 320-450°C and 25-35 MPa. The feed gas was produced by gasification of coal. In 1966, Imperial Chemical Industries (ICI) developed a low-pressure process where copper was the active material in the ~ a t a l y s tCO, . ~ COz and hydrogen, derived from natural gas by steam reforming, are reacted over a Cu/ZnO/A1203catalyst at 250°C and 5-10 MPa according to the reactions below: CO + 2H2 -+ CH30H (1)

CO2 + 3H2

CH30H (1) + H20 (1)

AHo= -128kJmol-'

AHo = - 131 kJ mol-'

3: Catalytic Hydrogen Generationfrom Methanol

69

The relatively mild reaction conditions lower production costs, because centrifugal compressors can be used. The reaction mechanism below was described by Chinchen et al.: lo c02 + c02*

H2 +2H* C02* + H*

+

HCOO*

HCOO* + 3H* co+o* H2+0*

--+

--+

CH30H + O*

c02

+ H20

Methanol formation from CO is usually negligible, which is quite the opposite of what was first believed when this process was launched. The reason for this is that the oxidation of CO to C02 by surface-bound oxygen is a very fast reaction. Consequently, CO is only the main reactant when the copper surface is free of adsorbed oxygen as in a CO2-free CO/H2 mixture. The ICI process still dominates the methanol world market. However, new and innovative processes are evolving. Recently, the US Clean Coal Technology programme, which is funded by the US Department of Energy, has presented a new process based on a liquid phase reaction system. 1.2 Methanol Conversion Processes. - In principle, there are four routes available for catalytic hydrogen generation from methanol: (i) methanol decomposition, (ii) steam reforming, (iii) partial oxidation and (iv) combined reforming (see Figure 1). All four reactions can be carried out at moderate temperatures (200-300 "C) over transition metal catalysts, such as copper and palladium. Especially CdZnO-based catalysts have received considerable attention, due to their use in the low-pressure methanol synthesis process. We will focus our discussion on the reaction pathways, mechanisms and kinetics of all three processing options. The differences in the catalytic functions of copper and precious metal catalysts will also be elucidated, as well as the role of various supports and promoters.

1.2.1 Decomposition. Methanol decomposition has been studied for decades. The reaction is the reverse of methanol synthesis from H2 and CO: CH30H (1) 4 CO + 2H2

AW = +128 kJ mol-'

The reaction is highly endothermic and well-suited for recovery of waste heat and/or to increase the heating value of the methanol fuel, for instance in combustion engines. However, the process is less suitable for supplying hydrogen to fuel cells, which are poisoned by CO. 1.2.2 Steam Reforming. Steam reforming of methanol is a well-developed

Catalysis

70 Decomuosition

Steam Reforming

(l>

3Hz +

Q.0

Partial Oxidation CHsOH

0

2 H,+ CO,

Q
(m) Combined Reforming at Autothermal conditions

FJoHe H2 0

e

2.5 \+ CO,

Q=O

Figure 1

Methanol conversionprocesses

process with a high efficiency, as hydrogen is formed from both methanol and water: CH30H (1) + H20 (1) + C02 + 3H2

AW= +131 kJ mol-'

Yields are typically high, with hydrogen concentrations of 70-75%. However, the reaction is highly endothermic and requires external heat exchange. Steam reformers are well suited for operation under stationary conditions where high hydrogen production capacities are needed. 1.2.3 Partial Oxidation. Partial oxidation of methanol is generally carried out with air or oxygen as oxidant:

CH30H (I) + @2

+

COz + 2H2

AW = - 155 kJ mol-'

The overall thermal efficiency of a system based on partial oxidation technology is somewhat compromised by the exothermicity of the reaction. However, there is no need for external heat exchange and, due to the high reaction rate, a partial oxidation reformer can be made compact, light-weight and dynamically responsive. The theoretical maximum hydrogen concentration in the product gas is 67% when using oxygen and 41% when using air as oxygen source.

3: Catalytic Hydrogen Generationfrom Methanol

71

1.2.4 Combined Reforming. Combined reforming, i.e. the combination of steam reforming and partial oxidation, is an attractive option. The general equation is given below: (s+p)CH30H (1) + sH20 (1) + ip02 + (s+p)CO2 + (3s+2p)H2 where s and p are stoichiometric coefficients for steam reforming and partial oxidation, respectively. The process is sometimes referred to as oxidative steam reforming or autothermal reforming, when operated close to adiabatic conditions. The main advantage of this option is that it combines the best features of steam reforming and partial oxidation. By changing the composition of the reactant feed, the process can be carried out at a wide range of operating conditions, spanning from highly endothermic to highly exothermic. Hence, the process can easily adapt to the requirements of the system at any given point of operation. This hydrogen generation technique has gained an increased interest for use in mobile applications.

2

Decomposition of Methanol

2.1 Introduction. - Several different types of catalysts have been studied for methanol decomposition to CO and H2. Among these, copper, 12-22 nicke123-25 and precious metals such as p l a t i n ~ m ~ and ~ 9p~a~l l a d i ~ m ~are ~ -reported ~~ to be effective. Some examples of references on decomposition of methanol are presented in Table 1. By-product formation may be quite severe in certain cases, depending on parameters such as reaction temperature, catalyst composition and preparation technique. Much work has been concerned with the role of the support. Alumina and silica are commonly used, but other supports such as titania and zirconia have also been investigated. Being an endothermic reaction, methanol decomposition has received growing attention for the recovery of heat from the engine exhaust in methanol-fuelled vehicles, in which decomposed methanol is fed to the engine (on-board reforming). Neat methanol or methanol-gasoline blends have also been suggested as alternative automobile fuels. However, this is associated with disadvantages, such as a lower heating value per unit volume, difficulty in starting the engine using neat methanol and a tendency of phase separation of methanol-gasoline blends.l 5 The use of decomposed methanol eliminates these problems and provides a cleaner fuel and increased fuel efficiency. Furthermore, utilisation of the heat in the engine exhaust leads to increased overall efficiency. Methanol decomposition to syngas is also applicable to recovery of waste heat of around 200°C from industries, thereby increasing the heating value of the f ~ e 1 .Moreover, ~ ~ - ~ ~decomposed methanol may be utilised as an on-site source of CO and/or H2 for chemical processes and material processing or serve as a supplemental fuel for gas turbines at peak demand.12914*16 Catalysts, which are active at low temperatures, are important for the development of heat recovery systems based on methanol decomposition. 12,26927*30

12y14-16,25,27,30,32,34

1998

1999

1999

2000

2000

2000

3 wt% Pd/CeO2

Fixed-bed flow reactor (6 mm i.d.); 0.20 g cat. 20% CH30H/Ar (4.8 dm3h-l) 160-220 "C; atmospheric pressure Prereduction: 20% H2/Ar, 300 "C, 1 h Fixed-bed flow reactor; 6 mm i.d. quartz tube; Matsumura et al.33 3 wt% Pd/Ce02 0.2 g cat. 25% CH30H/Ar (4.0dm3 h-l); 180 "C Prereduction: 20%H2/Ar, 300 "C, 1 h Shishido et al.32 Fixed-bed Pyrex tube flow reactor (8 mm i.d.); Pd, Pd-Cu, Cu, Zn, Pd-Zn 0.3 g cat. 40% CH30H/N2(GHSV = 10 dm3h-' g supported on Al, Mg and cat- l) 200-300 "C; atmospheric pressure Mg-A1 oxides Prereduction: 10% H2/N2, 350 "C Clement et al.22 CdSi02 Gas-phase flow set-up; 100 mg cat. Cu/ZnO/Si02 2.5%CH30H/He (27 ml min-'); 154 "C; 2 bar Prereduction: H2,330-475 "C Fisher and Bell13 5.7 wt% Cu/Si02 Low dead-volume IR cell; 2 cm catalyst disk of 0.2 mm thickness (75 mg) 32.6 wtY0 Zr02/Si02 1% CH30H/Ar (60 ml min-') Cu/Zr02/Si02 (5.7 wt% Cu, 30.5 wt% Zr02) 50-250 "C; atmospheric pressure Prereduction: 10%H2/He, 250 "C, 8 h Imamura et al.26 1-5 wt% Pt on CeO2, Tubular flow reactor; 1 ml cat. 3.8% CH30H/Ar (SV = 4200 h-l) Zr02, Al2O3, Si02/A1203, 150-500 "C; atmospheric pressure CaO and MgO; 3 wt% Rh, Pd or Ru on Ce02 Prereduction: H2,300 "C, 1 h Usami et Fixed-bed cont. flow reactor; 0.5 g cat. 15 wt% Pd supported on 15% CH30H/Ar (5 dm3 h-l); 200-300 "C Zr02, Pr2O3, Ce02, Fe304, TiOz, Si02 and ZnO Prereduction: 10% H2/Ar,400 "C, 1 h

Shen and Matsumura28

Reaction conditions

2000

Catalyst

AuthorlReJ

A selection of references on catalytic methanol decomposition ( I 990-2000)

Year

Table 1

Kinetics and reaction mechanism; Pt-CeO2 interactions; TEM, ESCA and EXAFS studies Support effects; cationic Pd species; Pd-support interactions

ZnO promotional effects; FT-IR study; Cu-ZnO interface Mechanistic study; ZrO2 promotional effects; in situ FT-IR study

Catalyst surface properties; EXAFS, XPS and TPD studies Catalyst preparation from synthetic anionic clays by solid phase crystallisation

Pd-Ce02 interactions; catalyst preparation by deposition-precipitation

Description

E

Y5-

9

Influence of support pore size

LOW activation energy over catalysts prepared by deposition-precipitation

Fixed-bed cont. flow reactor; 0.5 g cat. 25% CH30H/Ar (12 dm3 h-l) 200 "C; atmospheric pressure Prereduction: 17% H2/Ar Fixed-bed flow reactor; 6 mm i.d. quartz tube; 0.50 g cat. 20% CH30WAr (2.5 dm3h-l) 250 "C; atmospheric pressure Prereduction: 10%H2/Ar, 500 "C, 1 h Fixed-bed flow reactor; 6 mm i.d. quartz tube; 0.05-0.20 g cat.; 25% CH30WAr 180-240 "C; atmospheric pressure Prereduction: 20% H2/Ar, 400 "C, 1 h

Matsumura et al.23 2 and 10 wt% Ni on porous glass; 10 wt% Ni/Si02

Matsumura et al. 25 Si02-supported Ni and Pt

Matsumura et al. 30 2 wt% Pd/ZrOz

1997

1997

High Ni-content catalysts

Fixed-bed cont. flow reactor; 1.O g cat. 25% CH30H/Ar (12 dm3h-l); 250 "C Prereduction: 17%H2/Ar, 500 "C, 1 h

Matsumura et al.24 5-50 wt% Ni/Si02

1995

Effect of catalyst pretreatment; in situ XRD study

16 mm 0.d. Inconel reactor; 2.6 g cat.; 250 "C Prereduction: 5% H2/N2,250 "C, 3 h

10,20 and 25 wt% Cu/ZnO

Cheng16

Multicomponent catalysts; role of constituents; reaction and XRD studies

16 mm 0.d. Inconel reactor; 1.8 g cat. CH30H (35 ml min-l); 250 "C; 2 bar Prereduction: 5% H2/N2,250 "C, 3 h

Cu and Cu/ZnO catalysts promoted by Cr, Ba, Mn and Si oxides

1995

1995

Catalyst deactivation and regeneration

16 mm 0.d. Inconel cont. flow reactor; 2 g cat. Undiluted CH3OH; 275 "C

Cu/Cr oxide catalysts promoted by Mn, Ba and Si

1995

1998

Promotional effect of alkali additives

Glass reactor (13.6 mm i.d.); 0.6 g cat. Undiluted liquid CH30H (0.2 ml min- l) 225-275 "C; atmospheric pressure Prereduction: 17% H*/Nz, 250 "C, 3 h

Cu-Cr-Mn oxide catalyst (Nissan Girdler G-89) promoted by 1-4 wt% Li, Na or K

Cheng et al. l 2

1998

PdCe alloy formation

Fixed-bed cont. flow reactor; 1 g cat. 20% CH30H/Ar (2.5 dm3 h-l); 180-220 "C Prereduction: 10%H2/Ar, 325 "C, 1 h

5-35 wt% Pd/Ce02

Usami et al. 29

1998

w

4

3

0,

8

s

2

L

%

? n b' Ei e 5. n

Clarke et a1.21

Imai et al.39

1994

1990

FeTi, Feo.9Mno. ITi, CaNi5, LaN&.7A10.3, Mg2Ni and Mg2Cu

7 wt% CdSi02

Idem and Cu-A1 oxide catalysts B a k h ~ h i ~ O ? ~ ~ (0-57.3 wt% CU)

1994

Catalyst

AuthorlRef:

(contd.)

Year

Table 1 ~~

~

Catalyst composition and calcination temperature; Cu20 formation; acid site distribution

Description

Packed bed Pyrex reactor (6.5 mm i.d.) CH30H, 54.7 torr, in N2 and He (20 ml min150-350 "C; atmospheric pressure Prereduction: H2, 350 "C,3 h l)

Hydrogen storage alloys; methanol and ethanol decomposition

Low dead-volume IR cell (0.4 cm3)containing 50 mg Reaction mechanism; surface intermediates; catalyst pressed into a disk of 2 cm diameter TPD-IR study

Stainless steel microreactor (10 mm i.d.); 2 g cat. CH30H (WHSV = 26.4 h- l ) diluted with N2 170-250 "C; atmospheric pressure Prereduction: H2, 300 "C

Reaction conditions

2

3: Catalytic Hydrogen Generationfrom Methanol

75

Especially for cold start in automotive applications, where space in the engine compartment is limited, high-activity catalysts are needed. Many studies have been carried out at temperatures above 300°C. However, the lowtemperature activity is not sufficiently high, which is why catalyst development remains a major field of research. Methanol decomposition is the reverse of methanol synthesis from CO and H2, a reaction that has been carried out commercially over copper-based catalysts for decades. However, the mechanism is still not fully understood. By studying methanol decomposition on copper surfaces, information can be obtained also about intermediate surface species during methanol synthesis. In fact, a significant part of the work concerned with interactions between methanol and copper surfaces has been aimed at gaining further understanding of the methanol synthesis reaction. Recent work, however, has been more application oriented. In a previous paper from our laboratory, Pettersson and SjO~trOrn~~ reviewed the use of decomposed methanol as a fuel in automobile engines. The reader is referred to that work for references prior to 1990. The following discussion describes more recent work (1990-2000) and we will focus our discussion on catalyst materials designed specifically for methanol decomposition. 5727

2.2 Copper-based Catalysts. - 2.2.I Reaction Pathway and By-product Formation. CdZnO-based catalysts are active for methanol decomposition at temperatures between 200 and 300 "C, producing mostly CO and H2. However, the reaction may yield a mixture of by-products, including methyl formate (CH30CHO), dimethyl ether (CH30CH3), C02, methane and water. 15,169253 A significant amount of methyl formate is produced at low temperatures and low conversions, suggesting that it is the primary product at low temperatures. Gas-phase methyl formate is formed according to the following reaction: 12-22736

37

2CH30H (1) --+ CH30CH0 (g) + 2H2

AH" = +120 kJ mol-'

At higher temperatures, methyl formate is further decomposed into CO by either of the following two r e a c t i o n ~ : ' ~ 7 ' ~ 3 ~ ~ CH30CH0 (g) -+CO + CH30H (g) CH30CHO (g)

--+

2CO + 2H2

AHo = +46 kJmol-' AHo = +136 kJ mol-'

The following reaction is also plausible:36 CH30CHO (g) -+ C02 + CH4

AHo = - 111 kJ mol-'

Small amounts of dimethyl ether (DME) are also produced over copper: 2CH30H (1) -+ CH30CH3(g) + H 2 0 (1) AHo = +7 kJ mol-' Production of DME is affected by the acidity of the active sites. Especially over alumina and chromia-containing catalysts, methanol dehydration is much

76

Catalysis

more significant, due to the acidity of these metal oxides.15-26,36 DME formation increases with reaction t e m p e r a t ~ r eIdem . ~ ~ and Bakhshi20found that the acid site distribution on calcined Cu-A1 catalysts varied drastically with copper concentration and calcination temperature, hence affecting the extent of DME formation. Small amounts of C02 may be formed, either in the water-gas shift (WGS) reaction: 15926y36

CO + H20 (8) -+ C02 + H2

AHo = -41 kJ mol-

or via the following reaction, which requires the presence of oxygen species on the catalyst surface:l 3 CH30H + O* + COZ + 2H2 + * where * denotes an active site on the catalyst. Methane formation increases with increasing temperature due to methanation: 15926*36

CO + 3H2 4 CH4 + H20 (1)

AHo = -250 kJ mol-'

2.2.2 Reaction Mechanism. Much attention has been aimed at understanding the interactions of methanol with copper surfaces. The work is motivated, in part, by the use of copper-based catalysts in methanol synthesis, where formate is believed to be a key intermediate. Adsorption of formic acid produces formate species. Hence, some work has also focused on formic acid adsorption on copper surfaces.l7 Moreover, the interaction of methanol with oxide and metal surfaces represents a good test of surface chemistry.37 The reaction mechanism for methanol decomposition may be quite complex. In situ infrared (IR) spectroscopy has given insight into the nature and type of intermediates that are present on the catalyst surface, as well as increased understanding of the actual mechanism involved. CO, C 0 2 , H2 and methyl formate are the predominant products during methanol decomposition on copper s ~ r f a c e sCO . ~and ~ ~H2 ~ are ~ ~formed ~ ~ by the decomposition of methyl formate, while C02 and H2 are produced by decomposition of formate species, which requires the presence of surface oxygen, e.g. a preoxidised copper surface or dissociative adsorption of water in the feed.17 There appears to be general agreement that methanol decomposition over copper catalysts occurs via either of these two routes. Exposure of silica-supported copper to methanol leads to formation of methoxy species and adsorbed hydrogen on both copper and silica by dissociative adsorption of methanol: 3 ~ 1 7 , 2 1

CH30H (g) + 2 * -+ CH30*+ H* where * denotes an active site on the catalyst surface. The presence of preadsorbed oxygen enhances the formation of methoxy species and stabilises them to further decomposition. However, formaldehyde species may be formed by dehydrogenation of the methoxy groups: CH30*+ * + CH20*+ H*

77

3: Catalytic Hydrogen Generationfrom Methanol

H H\

I

/H

/"\PH3

? Figure 2

?

Methyl formate surface species

Surface formaldehyde can desorb to produce gas-phase formaldehyde:

CH20* + CH20 (8) + * Subsequently, two routes are possible, one forming CO and H2 via a methyl formate intermediate and the other producing C 0 2 and H2 by decomposition of formate intermediates. 2.2.2.1 The methyl formate route. On reduced copper surfaces, methyl formate is formed when methanol is present in the gas phase: l 3

CH30*+ CH20* + H2CO*O*CH3 H2CO*O*CH3 + * + HCO*O*CH3 + H* The methyl formate species present on the catalyst surface are depicted in Figure 2. Gas-phase methyl formate may be produced by desorption of surface methyl formate:

HCO*O*CH3 + CH30CHO (g) + 2 * CO and H2 result from the decomposition of methyl formate: HCO*O*CH3 + 2CO (8) + 2H2 (g) + 2 * HCO*O*CH3 + CO (8) + CH30H (g) + 2 * 2.2.2.2 The formate route. On a preoxidised copper surface, or in absence of methanol in the gas phase, formaldehyde is oxidised to formate species via a

.

,

~

~

~

~

~

l

39

7921

~

~

.

h

~

~

~

~

~

CH20* + O* + CH2(0)2* Methylenebis(oxy) is dehydrogenated to either monodentate or bidentate formate:

CH2(0)2* + HCOO* + H* HCOO* + * 4 HC(0)2* which decomposes to C02 and H2:

HCOO* + C02 (g) + $ 3 2 (g) + * HC(0)2* 4 C02 (8) + #H2 (g) + 2 * Formation of gas-phase C 0 2 requires a source of oxygen. Therefore, it is

~

~

78

Catalysis

important to consider different possible sources of oxygen on the catalyst surface. For instance, oxygen may remain on the surface due to incomplete reduction or may originate from small amounts of water in the feed, which dissociates to provide surface oxygen. Another possibility is the reaction with lattice oxygen according to the Mars-van Krevelen mechanism. 2.2.2.3 Mechanism over Zr02-promoted copper catalysts. Fisher and Bell13 studied methanol decomposition over reduced Zr02/Si02 and Cu/Zr02/Si02 catalysts. When Zr02-containing catalysts are exposed to methanol, most of the surface species are associated with zirconia. In the presence of copper, methoxy species on zirconia are dehydrogenated to formaldehyde and oxidised to formate, which decomposes to CO, C 0 2 and H2. Methyl formate and DME are also observed. In absence of copper, methoxy species on zirconia decompose only very slowly. It was suggested that the Cu/ZrO2-containing catalyst is bifunctional in nature, copper and zirconia having different but complementary roles. l 3 Metallic copper is proposed to provide sites on which hydrogen atoms, formed by dehydrogenation of surface species on zirconia, can recombine and desorb as H2. Thus, methanol decomposition over Cu/Zr02/Si02is proposed to occur primarily on zirconia, while the primary function of copper is that of hydrogen removal by reverse spillover of atomic hydrogen from zirconia. 2.2.3 Promotional Effects. 2.2.3.1 Chromia. Enhanced catalytic activity of copper catalysts for methanol decomposition can be obtained by addition of chromium oxide.15Cu/Cr weight ratios between 1:l and 3:l are recommended, but the ratio does not affect catalyst activity to any great extent. Chromia also helps to maintain the catalyst stability. The promotional effect is not fully understood, but it is believed that chromia helps to disperse or support the active copper phase. There is, however, a disadvantage associated with using chromia as dopant. Its presence lowers the selectivity to CO by dehydration of methanol to DME, caused by the acidity of chromia. 2.2.3.2 Silica. A few percent of silica (typically 2-4 wt%) significantly enhances the activity of copper based methanol decomposition catalysts.l5 Xray diffraction (XRD) studies have suggested that silica increases the dispersion of metallic copper in the working catalyst and decelerates the formation of crystalline CuO (and CuCr204) during calcination. The promotional effect of silica reaches an optimum at about 2.5 wt% with respect to methanol conversion. The CO selectivity, on the other hand, remains more or less unaffected by the silica concentration in the catalyst. 2.2.3.3 Zinc oxide. The industrial methanol synthesis catalyst, which contains zinc oxide (Cu/ZnO/A1203), has been studied for methanol decomposition. However, the role of ZnO during methanol decomposition is not ~ 1 e a r . l ~ Some report on a promotional effect, while others do not. Clement et a1.22 studied the effect of reduction temperature on the activity of unpromoted and ZnO-promoted Cu/Si02 catalysts in methanol decomposition. They found that the reaction rate over unpromoted Cu/Si02 was constant with increasing catalyst reduction temperature, whereas the ZnO-promoted catalyst showed a

3: Catalytic Hydrogen Generationfrom Methanol

79

strong decrease in activity, accompanied by an increased selectivity to methyl formate. FT-IR experiments showed that, under reaction conditions, methoxy species were stabilised on the Cu/ZnO/Si02 catalyst, which had been reduced at high temperatures. Since stabilisation of methoxy species above 200 "C only occurs on oxidised copper surfaces, it was suggested that copper cations were present in the catalyst reduced at high temperatures (600-750 K). It was further proposed that these stabilised copper sites are located at the Cu-ZnO interface. The stabilisation of methoxy species could explain the lower methanol decomposition activity and the higher selectivity to methyl formate over the promoted catalyst reduced at higher temperature. 2.2.3.4 Alkali additives. The activity of a Cu-Cr-Mn oxide catalyst is enhanced by addition of small amounts of alkali metals.12 The promotional effect decreases in the following order: K > Na > Li. Although alkali addition decreases the surface area, the copper dispersion significantly increases, leading to enhanced activity and increased CO and H2 selectivities, mostly due to decreased formation of methyl formate. The formation of DME due to dehydration of methanol is clearly suppressed by the addition of alkali. For example, the amount of DME has been observed to decrease by 80% when 2% K is added to a Cu-Cr-Mn oxide catalyst.12 However, the activity does not always increase monotonically when increasing the concentration of alkali in the catalyst. Recommended additions of alkali are in the 1-4 wt% range. 2.2.3.5 Multicomponent copper catalysts without zinc oxide. The preferred constituents of a high-activity methanol decomposition catalyst are Cu-CrMn-Si or Cu-Cr-Ba-Si. ChengI5 studied methanol decomposition over a series of Cu-Cr based catalysts doped with a few percent of Si, Mn and one or more alkaline earth metals, such as Ba. The product distribution was similar to that of Cu-ZnO based catalysts, but in absence of ZnO the catalyst stability was significantly better. 2.2.4 Catalyst Stability. In spite of their high activity in the methanol synthesis reaction, Cu-ZnO-based catalysts do not perform well in methanol decomposition during long-term on-stream operation. The catalysts exhibit high initial activity for methanol conversion but quickly deactivate, typically in the first few hours of operation.l5.l6 The deactivation may occur due to partial dissolution of zinc into the copper lattice. Addition of chromia is known to increase catalyst stability. In addition, the activity of copper catalysts can be greatly enhanced by eliminating the ZnO component. ZnO-free catalysts exhibit much more stable initial activity, making them more active during extended on-stream operation. Cheng14studied the deactivation and regeneration of multicomponent Cu-Cr oxide-based catalysts. These catalysts slowly deactivated due to a carbonaceous deposit, but the initial activity could be completely restored by an oxidative treatment, followed by reduction in H2 to restore the metallic copper state. 2.2.5 Active Species. It is believed that metallic copper (CuO) is the active

species in methanol decomposition.1 5 9 1 6 Chengls characterised spent copper

80

Catalysis

catalysts by ESCA and found that copper is mostly in the Cuo state. No Cu+ was observed, but X-ray diffractograms of spent CdZnO catalysts revealed the presence of a CuZn alloy, which readily forms at higher temperatures. If metallic copper was the active species, alloy formation would decrease the activity for methanol decomposition due to the partial dissolution of zinc into the metallic copper lattice. This is consistent with the low activity observed over ZnO-containing catalysts, compared to ZnO-free copper catalysts. 2.2.6 Calcination Temperature and Catalyst Copper Content. The calcination temperature affects the activity of copper-based catalysts. High temperatures generally cause lower activity. Traditionally, the same pretreatment conditions have been used for catalysts of different composition. Cheng,16 however, monitored the structural evolution of CdZnO catalyst precursors of varying composition during pretreatment by in situ XRD. It was evident that the temperature at which deactivation occurs depends on the catalyst composition and decreases with increasing copper content. The temperature at which crystalline CuO phase starts forming also depends on the catalyst composition and decreases with increasing copper content. The CO yield was higher over 10 wt% CdZnO than over its 20 wt% counterpart within the entire range of examined calcination temperatures (250-450 "C). Deactivation of the two catalysts occurred after calcination at 400 and 350 "C, respectively. However, there was not a direct correlation between activity and loss of surface area. The surface area decreased monotonically, while the activity dropped abruptly. Idem and Bakh~hi~O9~~ prepared Cu-A1 catalysts and studied the influence of catalyst composition and calcination temperature on copper dispersion and catalytic activity. The methanol conversion during methanol decomposition at 200 and 250 "C increased with increasing copper concentration up to about 25 wt%, where it reached a plateau. Copper dispersion increased with calcination temperature and decreasing copper concentration, but the increase in methanol conversion with copper dispersion was not significant. The turnover numbers for methanol conversion, however, decreased as the copper dispersion increased. As the high copper areas were obtained by calcination at high temperature, at which CuA1204 is formed, entrapment of metallic copper in the spinel structure after reduction may explain the lower turnover numbers.

2.3 Precious Metal Catalysts. - 2.3.1 Catalyst Activity and Product Distribution. 2.3.1.1 Platinum catalysts. Imamura et a1.26 studied methanol decomposition over ceria-supported platinum catalysts and found that complete methanol conversion was reached at 250°C over 1 wt% Pt/CeO,. Total methanol conversion and unchanged product selectivity could be maintained for 5 h at 250°C. However, the selectivities to H2 and CO decreased sharply with increasing temperature above 250 "C, where methane formation was initiated. Methane production increased up to 400"C, where it reached a maximum and subsequently declined. Increasing C02 production was observed above 300 "C, also reaching a peak value at 400 "C. Consequently, the CO and H2 selectivities increased again at temperatures exceeding 400 "C. The activity

3: Catalytic Hydrogen Generationfrom Methanol

81

increased almost linearly with increasing platinum load up to 3 wt%. Between 3 and 5 wt%, it was more or less constant. Over the most active catalyst, methanol was fully converted at 230°C with high selectivities to H2 and CO, 99.2 and 94.6%, respectively. 2.3.1.2 Palladium catalysts. Usami et aZ.27 found that 15 wt% palladium catalysts supported on Zr02, Pr203 and Ce02 displayed reasonably high activity for methanol decomposition at 200 "C, with CO selectivities higher than 98% between 200 and 250°C, regardless of support. At 300°C, the CO selectivity decreased due to by-product formation. Methane was the main byproduct over Pd/Zr02, while C02 dominated over Pd/PrzO3 and Pd/Ce02. TiO2- and Si02-supported catalysts exhibited negligible by-product formation, but significantly lower methanol conversion. A Pd/ZnO catalyst displayed very limited activity and only above 300°C. The same authors pointed out the effect of catalyst preparation technique. The methanol conversion over coprecipitated Pd/Zr02 catalysts increased with increasing palladium content and reached a peak value at approximately 20 wt%. At 5 wt% palladium, the methanol conversion was comparable over a co-precipitated and an impregnated catalyst. At 15 wt% palladium, however, the impregnated catalysts exhibited much lower conversion. It seems that the co-precipitation technique is more suitable than the impregnation technique when preparing supported palladium catalysts with a high palladium load. 2.3.1.3 By-product formation. Methane, C 0 2 and water are the main byproducts during methanol decomposition over precious metal catalysts. Methane by-product is formed by the reaction between CO and H2, so-called methanation: CO + 3H2 + CH4 + H20 C 0 2 is probably formed in the water-gas shift (WGS) reaction: CO + H20 --+ C02 + H2 As the WGS reaction is slightly exothermic, it may go in the reverse direction at high temperatures, producing water. Thus, steam reforming of methane may occur (see below), an endothermic reaction favoured by high temperatures. This would explain the decline in methane formation observed at higher temperatures:26 CH4 + H20 + CO + 3H2 Over alumina-supported catalysts, dimethyl ether production may be quite severe, which is probably due to the presence of acid sites:26 2CH30H + CH30CH3 + H20 2.3.2 The Support. The catalytic activity of precious metals is strongly affected by the nature and type of support, suggesting that interactions between the metal and the support are important. There is some disagreement in the literature regarding the most suitable choice of support. For instance, platinum supported on Ce02 exhibited the highest activity for methanol decomposition

82

Catalysis

in a series of precious metal catalysts (Pt, Pd and Rh) supported on various metal oxides (Ce02, A1203,Zr02, MgO, Si02/A1203and Ca0).26On the other hand, the catalytic activity decreased in the following order over palladium catalysts supported on various supports: Pd/ZrOz > PdPrzO3 > Pd/Ce02 > PdFe304 > Pd/Ti02 > Pd/Si02 > Pd/Zn0.27 2.3.2.1 Alumina. Wickham et aZ.38studied the deactivation at temperatures above 400°C of a series of catalysts consisting of palladium on modified alumina supports. The effect of Li, Na, K, Rb, Cs, Ba and La modifiers was investigated. The Li-, Na- and Ba-modified materials deactivated substantially during methanol decomposition compared to their K-, Rb-, Cs- and Lamodified counterparts and the unpromoted catalyst. The former displayed a greater amount of C 0 2 relative to CO desorbed than the latter, indicating carbon build-up on the catalyst surface. It was found that the Li-modified catalyst exhibited significant coking, a feature that was not observed over the La-modified catalyst. However, the authors found no correlation between support acidity or particle size and the degree of CO dissociation. 2.3.2.2 Ceria. Ceria is often used as a catalyst support because of its redox behaviour and electronic properties. For instance, ceria is a key component in the automotive three-way exhaust catalyst, where it is known to prevent sintering of precious metals by stabilising their dispersed state, to promote CO oxidation, water-gas shift and to act as an oxygen reservoir. Interestingly, several researchers have reported a strong metal-support interaction (SMSI), i. e. migration of cerium on the metal surface and/or metal e n c a p s ~ l a t i o n . ~ * - ~ ~ There are some different opinions in the literature regarding the state of precious metals supported on ceria, i.e. whether the metal particles aggregate on the surface or interact with ceria and penetrate into the bulk. Strong interactions between platinum and the bulk ceria, indicating penetration of platinum into the ceria bulk and possible Pt-0-Ce bond formation, have been found in ceria-supported platinum catalysts prepared by co-precipitation.26 Matsumura and c o - ~ o r k e r studied s ~ ~ the influence of preparation technique of ceria-supported palladium catalysts. They report that Pd/Ce02 prepared by the co-precipitation method, using sodium carbonate as precipitant, is more active than Pd/Ce02 precipitated using sodium hydroxide. Cationic palladium species were observed in the former catalyst and PdCe alloy was suggested to be the active species during methanol decomposition. The promotional effect of cationic palladium species formed by interaction with the support has also been reported e l s e ~ h e r e .The ~ ~ .Matsumura ~~ g 1 - 0 ~also ~ ~found ~ 9 that ~ ~ using the deposition-precipitation technique for preparation of Pd/Ce02 results in strong metal-support interactions. These catalysts were highly active for methanol decomposition (160-220 "C), compared to conventional impregnated samples. There were cationic palladium species in the former catalyst after reaction, while only metallic palladium could be found in the latter. 2.3.2.3 Zirconia. Zirconia has been reported as a suitable support for palladium, ruthenium, rhodium and platinum catalysts. Matsumura and c o - w ~ r k e r s *observed ~ ~ ~ ~ an interaction between palladium and zirconia in catalysts prepared by the deposition-precipitation technique

3: Catalytic Hydrogen Generationfrom Methanol

83

that affects the catalytic activity. An apparent activation energy (E,) of 80 kJ mol-' for methanol decomposition was observed over 2 wt% Pd/Zr02 prepared by deposition-precipitation. Over a similar catalyst prepared by impregnation, the activation energy was 100 kJ mol- I . Small cationic palladium particles that interact strongly with the support are more active than metallic palladium. Thus, smaller particles and a closer contact between the metal and the support is desired. Matsumura and colleagues27observed a mean Pd crystallite size as small as 4 nm in Pd/Zr02 catalysts prepared by co-precipitation. A considerable fraction of the particles were incorporated in the support, which resulted in strong Pd-support interactions. X-ray photoelectron spectroscopy (XPS) analyses suggested the presence of a positively charged palladium species, most likely Pd+, and the authors speculated that its presence is preferable for low-temperature methanol decomposition. 2.3.3 Reaction Pathway and Mechanism. The decomposition of methanol over palladium is roughly comprised of the following three steps:30

1 dissociative adsorption of methanol to form surface methoxy groups 2 decomposition of the surface methoxy groups to surface CO and H2 3 desorption of surface CO and H2

-

Dissociative adsorption of methanol on platinum occurs readily: CH30H (g) + 2 *

CH30*+ H*

where * denotes an active platinum site. Step (2), i.e. the abstraction of a hydrogen atom from the methoxy group, is generally believed to be the ratedetermining step,26represented by the following sequence: CH30* + H* + H2CO* + H2 + * CH30* + * -+H2CO* + H* 2CH30* + CH30CHO (8) + H2 (g) + 2 * CH30* -+H2CO* + i H 2 (g) Surface methoxide groups are converted into surface formaldehyde, producing H2. Formaldehyde, in turn, is readily decomposed into CO and H2. A methyl formate intermediate is also plausible, which may be further decomposed into CO by either of the following two reactions: CH30CHO + CO + CH30H CH30CHO -+2CO + 2H2 As discussed in Section 2.3.1.3, methane may be formed in the reaction between CO and H2. However, the formation of methane by decomposition of the methyl formate intermediate, producing C02, is also possible:

CH30CHO + C02 + CH4

84

Catalysis

2.4 Other Catalyst Materials. - 2.4.I Nickel-based Catalysts. Nickel-based catalysts have received some attention in the methanol decomposition reaction by Matsumara and collaborator^,^^-^^ among others. For instance, Ni/Si02 catalysts prepared by the sol-gel technique were studied and the catalytic activity for methanol decomposition was found to increase with increasing nickel content up to 40 wt%. The activity of Ni/Si02 prepared by the conventional impregnation technique also increased with increasing nickel load, but only up to 10 wt% where it reached a plateau. The former catalyst was significantly more active than the latter. Methanol was decomposed mainly to CO and H2 at 25OoC, but methane and water were also detected as by-products.24 The decomposition of methanol over nickel dispersed on porous glass (Si02, B2O3 and NazO) of narrow pore size distribution (4, 19 and 45 nm) has also been in~estigated.~~ It was noted that the catalyst in which the crystallite size of nickel was close to the pore diameter of the glass (19 nm) was the most active. Furthermore, the relationship between catalytic activity and the location of nickel particles on the support was investigated. 2.4.2 Hydrogen Storage Alloys. Methanol decomposition has also been studied over hydrogen storage alloys.39Each alloy displays a characteristic selectivity depending on its composition. Titanium-containing alloys are active for the formation of formaldehyde, while nickel-containing alloys are active towards CO formation. Methyl formate is produced over magnesiumxopper alloy. 2.4.3 Other Metal Oxides. Aas et studied methanol decomposition by TPD and XPS over Ti02, SrTi03 and SrO metal oxides. Titania was partially reduced during methanol adsorptioddesorption and exhibited predominantly methane formation at low coverages. Methoxy species were proposed to be the major surface intermediate. The formation of CO and H2 by decomposition of formate species was suggested. The catalytic behaviour of SrTi03 was similar to that of Ti02. On SrO, CO and H2 formation coincided with minor production of CO2 and H20, indicating that formate was the dominating intermediate on this surface. 3

Steam Reforming of Methanol

3.1 Introduction. - Methanol steam reforming is a well-developed and commercialised process. It can be carried out over copper-based catalysts with high selectivities to H2 and C02.40-54Group 8-10 metals have also received some a t t e n t i ~ n . ~ OSome * ~ ~ -examples ~~ of references on steam reforming of methanol are presented in Table 2. Under favourable conditions, the product stream contains up to 75% H2 and 25% C02 on a dry basis. The high H2 production capacity is the main argument for choosing the steam reforming process over partial oxidation, which theoretically produces a maximum H2 content of 67% when pure oxygen is used as oxidant. If air is used instead, the maximum H2 content decreases to about 40%.

Conventional flow reactor H20/CH30H = 111 180-300 "C; atmospheric pressure Prereduction: 10% H2/N2, 180-250 "C, 1 h

1

1994 Iwasa et al.55

1997 Shen et ~

1997 Takezawa and Iwa~a~~

.

Pd/ZnO, Pd/ZrO2, Pd/SiOZ and Pd. 10 wtY0 Pd for all combinations

Various ~ ~ Cu/ZnO catalysts

Conventional flow reactor 220 "C; atmospheric pressure H20/CH30H = 1:1 (0.47 s res. time) Prereduction: 4% H2/He

40 mg cat.; H20/CH30H= 1:1 in N2 (100 ml min-') 140-290 "C; atmospheric pressure Prereduction: 3% H2/He, 210 "C, 1 h

Supported copper and Conventional flow reactor Group 9 and 10 H20/CH30H = 1 1 metals (Ni, Pt, Pd, Rh) 0.47 s residence time Atmospheric pressure

1999 Breen and Ross40 Multiple combinations Conventional flow system; 0.1 g cat. of Cu, Zn, Al, Zr, La H20/CH30H = 1.3:1 (38.6 ml min- ') 140-345 "C; atmospheric pressure and Y Prereduction: 5% H2/N2, 240 "C, 4 h

2000 Lindstrom CdCr, Cu/Zn and and Pettersson60 Cu/Zr on A1203

A1203-supported Conventional flow reactor Cu/Cr, Cu/Zn, Cu/Zr, H20/CH30H = 1:l; Space vel.: 17 000 h-' Cu/Cr/Zn, Cu/Cr/Zr 180-300 "C; atmospheric pressure Prereduction: 10% H2/N2, 180-240 "C, 2 h

2001 Lindstrom and Pettersson6'

Reaction conditions

Catalyst

A selection of references on methanol steam reforming

Year Author/reJ:

Table 2

A study on the difference of the catalytic functions of Group 9 and 10 metals and copper. The study also includes dehydrogenation of methanol A study on the catalytic surface structure of copper-based catalysts and the effects on the selectivity

A detailed study on the effects of the support for palladium-based catalysts

co

HCOOH, COY CH30CHO

A detailed investigation on the performance of copper containing catalysts and a suggested reaction route

HCOOH, COY CH30CHO HCOOH, COY CH30CHO

A study on the performance of various copper-based catalysts supported on alumina

An investigation concerning the Performance of binary and ternary alumina-supported copper-based catalysts

Description

co

co

By-products

f+

k

n*

2

h' Q

P

*

Cu/Zn/Al alloys of various mass combinations

Pd on various supports, Pd content at 1 wt% for all combinations

1994 Miyao et al. 53

1993 Iwasa et al. 57

Conventional flow system 150-250 "C; atmospheric pressure

Conventional flow reactor H20/CH30H = 111 0.23-3.29 s residence time 200-300 "C

Conventional flow reactor H20/CH30H= 1:1 (0.47 s res. time) 220 "C; atmospheric pressure

Reaction conditions

1982 Takezawa et al. 54 Cu/Si02catalyst with Conventional flow system H20/CH30H = 1:l in N2 a variety of copper loadings (96 ml min-') 220 "C; atmospheric pressure

1985 Takahashi et al.50 10 wt% Cu/Si02 1 wt% Pt/Si02

Catalyst

(contd)

Year Authorlref

Table 2

CH3OCHO

co,

HCOOH,

HCOOH, co, CH30CH0, HCHO

co

co

By-products

The effect of copper loading is studied and how it affects the selectivity and conversion for the reforming of methanol

A study, which describes the different reaction paths for the reforming of methanol over palladium and copper-based catalysts

A detailed study of catalytic effects on numerous supports including SiOz, A1203, La2O3, Nd2O3 and Nb2O5

An investigation of the catalytic behaviour of various alloys for the steam reforming reaction

Description

m

a0

3: Catalytic Hydrogen Generationfrom Methanol

87

H2 can be produced by steam reforming of a variety of other fuels, such as ethanol, gasoline and natural gas, and the reasons for choosing methanol are given in Section 1.1. In this section, a detailed review of the literature concerning methanol steam reforming for automotive applications is presented. Stationary applications have been omitted as the energy requirements and outlet gas qualities are different. The high energy requirement is one of the main technological bottlenecks for the integration of a steam reforming system in a mobile application. The reaction is highly endothermic and the need to produce steam further increases the energy demand of the process. Nevertheless, there are several commercial endothermic methanol-reforming solutions. Some of these are reviewed in Section 6 . The reaction path and kinetics of the steam reforming reaction are firmly directed by the catalytic material used. Therefore, the following discussion is divided into two parts: (i) copper-based catalysts, especially the traditional low-temperature shift catalyst based on Cu/ZnO/A1203, and (ii) materials containing other transition metals supported on various metal oxides. 3.2 Copper-based Catalysts. - 3.2.I Reaction Pathway. Two different reaction pathways for steam reforming of methanol over copper have been suggested in the literature. The first one is a decomposition-water-gas shift (WGS) sequence, while the other proceeds via methanol dehydrogenation to methyl formate. 3.2.1.1 The decomposition-WGS reaction scheme. In the decompositionWGS reaction scheme, the overall reaction for methanol steam reforming is considered to occur through methanol decomposition, followed by the watergas shift reaction, in which C02 and H2 are produced from CO formed in the decomposition step: CH30H + H20 + C02 + 3H2 CH30H + CO + 2H2 CO + H20 -+ C02 + H2 This reaction sequence has been adopted by several authors for Cu/ZnO/ A1203catalysts, for instance by Santacesaria and Carra45 and Amphlett and c o - ~ o r k e r s-43*52 . ~ ~ The latter studied the steam reforming reaction over the BASF K3- 110 catalyst, a commercial low-temperature water-gas shift catalyst consisting of Cu/ZnO/A1203(40/40/20 wt%). A reaction model presented by Peppley et aL41 indicated that CO is produced separately from C02, rather than simultaneously. The authors also came to the conclusion that the active sites for methanol decomposition are distinct from the sites at which the steam reforming and WGS reactions take place. 3.2.1.2 The methyl formate reaction scheme over Cu/ZnO/A1203. Jiang et aZ.44951studied the S3-85 Cu/ZnO/A1203catalyst from BASF, containing 50 wt% ZnO, compared to 40 wt% in the K3-110 catalyst. They proposed a set of reactions based on unstable intermediates. The WGS reaction, along with the

88

Catalysis

production of CO through methanol decomposition, was ruled out. Instead, methyl formate was proposed as the only stable intermediate and by-product present. The suggested reaction path concurs with the route proposed by Breen who observed small amounts of methyl formate during steam and reforming over copper on a variety of supports (i.e. ZnO, Zr02, A1203 and combinations thereof). The reaction route is presented below: 2CH30H + CH30CHO + 2H2 CH30CHO + H20 + HCOOH + CH30H HCOOH + C02 + H2 In the first step, methanol is dehydrogenated. The methyl formate intermediate is then hydrolysed to formic acid, which decomposes to C02 and H2. Breen and Ross40 argued that the reaction path was similar over all the studied copper catalysts, irrespective of support. 3.2.1.3 The methyl formate reaction scheme over CdSi02. Takezawa and c o - w o r k e r ~ ,who ~ ~ -studied ~ ~ ~ ~ CdSi02 ~ catalysts, also proposed a reaction path including methyl formate at high temperatures: CH30H + HCHO + H2 HCHO + CH30H + CH30CHO + H2 CH30CHO + H20 + CH30H + HCOOH HCOOH + C02 + H2 At low temperatures, however, the methyl formate intermediate was suggested to be absent, i.e. the second and third steps above reduce to: HCHO + H2O + HCOOH + H2 Takezawa and IwasaS6found that the CdSi02 catalyst is not affected by an increase in steam-methanol ratio, thus indicating that the Si02 supports provide a different reaction path than A1203. 3.2.2 By-product Formation. Steam reforming of methanol over copper catalysts is generally carried out with high selectivities to H2 and C02. In certain cases, however, CO and methyl formate formation may be significant. Formic acid and traces of formaldehyde and dimethyl ether have also been observed by some r e ~ e a r c h e r s . ~ ~ ~ ~ ~ ~ ~ * ~ Jiang et aL4 noted that no CO is produced below 25OoC, whereas CO formation occurs to an appreciable extent at temperatures above 300 "C. CO has been explained to be a secondary product of the steam reforming reaction as a result of reverse water-gas shift, occurring when methanol is almost completely converted:40 C02 + H2 + CO + H20 The level of methyl formate is generally low due to the reaction with water to C 0 2 and HZ:

3: Catalytic Hydrogen Generationfrom Methanol

89

CH30CHO + H20 -+ CH30H + HCOOH HCOOH + C02 + H2 3.2.3 Steam-Methanol Ratio and Water-Gas Shift. In order to maximise the H2 yield, steam reforming is generally carried out with steam in excess of the stoichiometrically required amount. This also favours the water-gas shift reaction. In other words, CO clean-up by the shift reaction can be integrated in the reforming reactor, increasing the H2 content (see also Section 5.3). By optimising the reaction conditions for WGS, the need for additional clean-up can be significantly reduced. However, the use of excess steam is disadvantageous from an energy perspective. Moreover, if the steam-methanol ratio is too high, the system will be diluted, hindering the reforming reaction from taking place. There is some disagreement in the literature regarding the extent of the water-gas shift reaction during methanol steam reforming. Jiang et aL4 found that the Cu/ZnO/A1203catalyst gave high conversion for WGS when this was carried out alone. However, the reaction did not occur when methanol was present, i.e. the presence of methanol was found to hinder the WGS reaction from taking place. In a study by Amphlett et on the other hand, the authors came to the conclusion that the low-temperature shift catalyst (Cu/ZnO/A1203) was effective in the removal of CO from the reformed gas also in the presence of methanol. The work focused on the poisoning effects of CO on the fuel cell and various methods reducing the CO content in the product stream. It was concluded that the WGS reaction was slow and that a large reactor would be required for an effective clean-up and that increasing the reactor temperature would shift the equilibrium making the WGS reaction unfavourable. Peppley et aL41 were also unable to find any supporting evidence to the claim made by Jiang and co-workers.44 Takahashi et aZ.,” however, found that the Cu/SiO2 catalyst does not promote the WGS reaction. The claim by some authors that the process is 100% selective for C02 and that the contribution from the WGS reaction can be neglected may be appropriate for industrial applications, in which the CO production may be considered negligible. However, for automotive applications where very low levels of CO can severely poison the anode of the PEM fuel cell the decomposition and WGS reactions must be taken into consideration. 3.2.4 Reaction Mechanism. Jiang et aL4 suggested a reaction mechanism for steam reforming on a copper surface. It should be noted that these authors have adopted the reaction pathway based on a methyl formate intermediate. The mechanism is given below:

CH30H + CH30* +H* CH30* + CH20*+ H* H20 + H20* H2+2H*

Catalysis

90

2CH20* + CH30CHO* CH30CHO* + H20* 4 HCOOH* + CH30H HCOOH* + H2 + C02 where * represents an adsorption site on the copper surface. It was suggested that the dehydrogenation reaction controls the rate of steam reforming. Furthermore, Langmuir-Hinshelwood modelling indicated that extraction of hydrogen from methoxy groups on the surface is rate-determining. 3.2.5 Kinetic Models. In 1999, Peppley et al.41342published two papers containing an extensive kinetic model of the methanol steam reforming reaction over the commercial BASF K3- 110 Cu/ZnO/A1203catalyst. The experiments were carried out using stoichiometric conditions, i.e. steam and methanol at a 1:l molar ratio. In the first part, the following rate prediction system for the various components of the steam reforming reaction network was pre~ented:~' rH2

= ( 3 r R + 2 r D + rw) S A

rC02= (rR+ rW)SA -rH20= rCO= (rD

(rR+rW)SA - rW)SA

- r C H 3 O H = (rR + r D ) S A

(mol s-' kg-')

(34

(mol s-' kg-')

(3.2)

(mol s-l kg-')

(3.3)

(mol s- kg-

(3-4)

l)

(mol s-l kg-')

(3.5)

where S, is the specific surface area and YR, rwand r D are the rates of steam reforming, water-gas shift and decomposition, respectively. . ~ ~a more In the second part of their kinetic study, Peppley et ~ 1performed detailed investigation of the reaction rates of the three main reactions during steam reforming. The derived model was based on the findings of Diimpelmann,59who proposed a direct route for the steam reforming reaction, as well as a route for the WGS reaction involving the formation of formaldehyde as an intermediate. The authors proposed a detailed surface reaction network, in which the WGS reaction was incorporated in the steam reforming reaction, thus including all side reactions. The rate laws for the main reaction steps of the steam reforming reaction, as presented in the paper, are listed below: Methanol steam reforming:

Methanol decomposition:

3: Catalytic Hydrogen Generationfrom Methanol

91

Water-gas shift:

where pi is the partial pressure of component i, ki the rate constant of component i and Ki the equilibrium constant of reaction i or adsorption coefficient for surface species i. K' is the composite parameter. The results above were verified by applying a temperature scanning method to study the kinetics of the steam reforming reaction.43 It was concluded that the mechanism for the steam reforming reaction could be derived from a Langmuir-Hinshelwood mechanism, as performed by Peppley et al.,42and that the reaction site for the steam reforming reaction differs from that of the WGS reaction. Santacesaria and Carra45 also proposed a kinetic model for the system, presented below:

where

Jiang et aLS1suggested a simple rate law for the steam reforming reaction: rsR

= kQe-105kJ/mol/RTp Q . 2 6

CH30H

p Q . 0 3 p-0.2 H20 H2

(3.10)

A rate expression for the conversion of CO during steam reforming was suggested by Amphlett et al.:58

A selection of references involving kinetic modelling is presented in Table 3 . 3.2.6 Catalyst Composition. In 1999, Breen and Ross40 performed a study to investigate the influence of zirconia addition to copper-based catalysts. Various types of catalysts were tested, i.e. Cu-Zn, Cu-Zr, Cu-Zn-Zr, Cu-ZnA1 and Cu-Zn-Zr-A1 materials with varying mass compositions. Steam reforming experiments were carried out with a molar steam to methanol ratio of 1.3:l. The authors found that an increase in the copper content of C d Z r 0 2 catalysts increased the conversion and the selectivity. It was further observed that the activity decreased notably for a Cu/Zr02 catalyst (30/70 wto/o) at temperatures above 345°C. The selectivity to C 0 2 was 100% for all of the

Cu/ZnO/A1203 BASF K3-110

Asprey et al.43

Breen and Ross40 Multiple Reaction sequence model combinations of Cu, Zn, Al, Zr, La and Y

Amphlett et al.58

Amphlett et a1.52 Cu/ZnO/A1203low temperature shift catalyst

Jiang et al.51

Jiang et al.

Iwasa et al.57

1999

1999

1996

1994

1993

1993

1993

Pd-based catalysts

Cu/ZnO/A1203 BASF S3-85

A detailed study on the derivation of the rate laws for the individual species in the steam reforming reaction

Description

Kinetic model for the reforming of methanol including CO production

Kinetic model for CO clean-up for PEM fuel cells

Detailed kinetic model for all intermediates

Reaction sequence model

Rate model for the steam reforming reaction

A detailed study on the catalytic effects of supports with palladium as the active material

The model presented in this paper assumes that no CO is formed and is based on kinetic data

A study on the order of the reaction rate, which includes a rate model and a claim that methanol hinders the WGS reaction

A study investigatingthe kinetic parameters of reforming including the effects of temperature and pressure

A detailed model for the conversion of CO to C02 including kinetic parameters

A detailed investigation of the performance of copper-containing catalysts and a suggested reaction route

The study presents kinetic data for all intermediates from the sequence proposed by Peppley et al. using a temperature scanning method

Detailed rate law for all reaction steps This is a continued investigation of a previous study, in which detailed rate laws for the steam reforming, decomposition and WGS reactions are presented

Rate model for the steam reforming Cu/ZnO/A1203 BASF S3-85 and reaction Harshaw Cu-0203-T

Cu/ZnO/A1203low temperature shift catalyst

Cu/ZnO/A1203 BASF K3-110

Peppley et al.42

1999

Basic rate law for all reaction species

Cu/ZnO/A1203 BASF K3-110

Peppley et aL41

1999

Model

Catalyst

Authorlref.

Some examples of references on the kinetics of methanol steam reforming

Year

Table 3

3: Catalytic Hydrogen Generationfrom Methanol

93

94

Catalysis

catalysts tested at 225"C, with the exception of Cu-Zn combinations with copper contents above 80 wt%, where the reaction delivered CO concentrations exceeding 3%. The performance of the CdZnO catalysts was better than that of the Cu/Zr02 catalysts. However, the best results over the entire temperature range were obtained over the Cu/ZnO/Zr02/A1203catalyst. It was also noted that addition of zinc to CulZr02 catalysts increases the copper dispersion, as well as the activity. Furthermore, Cu-Zn-Zr catalysts were more active than Cu-Zn-A1 catalysts, indicating that zirconia is a more effective support for copper and zinc than alumina. The addition of alumina to Cu-Zn-Zr increased the stability, which was further improved by addition of yttria. Yttria-containing catalysts were also more active than their lanthanacontaining counterparts. In our own laboratory, Lindstrom and Pettersson6' performed a study on CdCr, CulZn and CdZr supported on y-alumina. A commercial catalyst, G66-B from Sud-Chemie, was used as a reference. The experiments were performed at an equimolar methanol-to-steam ratio. The authors concluded that the addition of chromium improved the performance of the copper-based catalyst at temperatures above 230 "C. The authors61 also tested various promoted copper catalysts supported on y-A1203, both binary and ternary combinations. The promoters used were chromium, zinc and zirconium. The catalysts were evaluated at 30% excess of steam and a space velocity of 17 000 h-l. It was concluded that an increase of the copper content of the catalyst will increase the hydrogen concentration in the product gas. Chromium promoted catalyst activity, especially at high temperatures, while zirconium showed a positive effect at low temperatures. The performance was greatly enhanced when adding a second promoter. Furthermore, the authors reported that alumina-supported ternary Cu/Cr/Zn and Cu/Cr/Zr catalysts were more selective than their binary counterparts, Cu/ Cr, CdZn and CdZr. Lindstrom and Pettersson62 recently conducted a study regarding the deactivation of a methanol steam reforming catalyst caused by poisonous substances or thermal effects. The catalyst activity was greatly affected by the addition of low concentrations of sulfur and chlorine. Sulfur was more detrimental to the commercial Cu/ZnO/A1203 catalyst than chlorine. The exposure of the catalyst to temperatures above 340°C caused an irreversible loss of activity and the C02 selectivity also decreased when the catalyst was subjected to high temperatures. The results in this study indicate that care must be taken in order to prevent any contact with either chlorite or sulfur in a fuel cell application. Shen et aL4* investigated how the preparation method of Cu-Zn catalysts affects the surface area and the catalytic activity for the steam reforming process. The highest methanol conversion was obtained for a 1:1 combination. However, rather low values were obtained in the experiments with methanol conversions not exceeding 13% when operating at 220 "C. Takezawa et ~ 1 investigated how the copper loading of Cu/SiO2 catalysts affects the selectivity and conversion for steam reforming of methanol. The authors found that for

.

~

3: Catalytic Hydrogen Generationfrom Methanol

95

catalysts with low copper loading there was considerable methyl formate formation, while over catalysts with high copper loading, C 0 2 and H2 production was favoured. Miyao et aLS3studied catalysts derived from Cu-Zn-A1 alloys in the steam reforming reaction and found that the alloys with an even mass distribution of active material and support were superior to catalysts with a high content of active material and low content of support. The best results were obtained for a catalyst with a 30/30/40 (wt%) composition, over which the authors reported a selectivity of 99.6% for the formation of C02. There seems to be general agreement that metallic copper is the active species in methanol steam r e f ~ r m i n g . ~ y ~ * * ~ ~

3.3 Other Transition Metal-based Catalysts. - The catalytic functions of copper differ drastically from Group 9- 10 transition metals in the conversion of alcohols. Takezawa and I ~ a s conducted a ~ ~ a study on the behaviour of Group 9-10 metals supported on Si02 for the steam reforming of methanol and concluded that the reaction takes place by a different mechanism than over copper-based catalysts. It was found that the performance of the Group 9- 10 transition metal catalysts was much poorer than that of copper-based catalysts. The difference in the catalytic behaviour was ascribed to the difference in reactivity of formaldehyde species. 3.3.1 Reaction Pathway. Iwasa et aLS7suggested that the reaction over PdlZnO catalysts proceeds according the pathway below:

CH30H

--+

HCHO + H2

HCHO + H20 4 HCOOH + H2 HCHO + CH30H --+ CH30CHO + H2 CH30CHO + H20 + CH30H + HCOOH HCOOH --+ C02 + H2 Formic acid is formed either directly from formaldehyde or via a methyl formate intermediate. Finally, formic acid decomposes to H2 and C02. Takezawa and colleaguessO~ss suggested the following route for the steam reforming of methanol over Pt/Si02 catalysts: CH30H HCHO

--+

--+

CO + H20

HCHO + H2

CO + H2 --+

C02 + H2

3.3.2 Catalyst Composition. Takezawa and collaboratorss6~s7have studied Group 9-10 metals (palladium, platinum, rhodium and nickel) supported on a large variety of metal oxides, including A1203,Si02, ZrOz, ZnO, Cr2O3, La203, Mn02, MgO, Nb205, Nd2O3 and Hf02. The selectivity of silica-supported catalysts was found to be poor for the steam reforming process when compared

96

Catalysis

to Cu/Si02, which produces H2 and C 0 2 very selectively.The best performance was obtained over the Pt/Si02 catalyst, which also exhibits high selectivity in the WGS reaction, as demonstrated by Takahashi and c o - w ~ r k e r sNever.~~ theless, CO selectivities of up to 25% were observed during steam reforming. The same researchers also found that the product distribution during steam reforming over palladium catalysts varied drastically with the composition of the support. The only catalyst which gave a reasonable activity and selectivity for steam reforming was Pd/ZnO. The performance of this catalyst was comparable to the traditional low-temperature shift catalysts. Over the other catalysts, methanol decomposition dominated with a high selectivity to CO. In 1995, Iwasa et ~ 1 concluded . ~ ~ that reduction of Pd/ZnO at high temperature produces PdZn alloy, which enhances the activity of the catalyst for the steam reforming reaction. It was found that CO was produced in parallel with C02, suggesting that the steam reforming reaction occurs together with the decomposition of methanol. These findings support the kinetic studies presented previously for copper-based catalyst^.^' -43945346952

4

Partial Oxidation of Methanol

4.1 Introduction. - Up to the mid-l980s, steam reforming was the only process successfully demonstrated for production of H2-rich gas streams nearly free of CO, using methanol as fuel. As previously described, steam reforming produces a mixture composed mainly of H2 and C02. However, due to the endothermicity of the reaction, heat needs to be supplied externally by combustion of a fraction of the fuel-methanol or some other fuel. In addition, the need to produce steam makes this process even less attractive. Partial oxidation, in the absence of steam, offers several a d ~ a n t a g e s . ~ ~ - ~ l The reaction route is exothermic, i. e. thermodynamically favourable, and uses air or oxygen instead of steam as oxidant. Moreover, the partial oxidation reaction displays a higher reaction rate than steam reforming. Under favourable reaction conditions, partial oxidation with oxygen yields H2 and C 0 2 almost quantitatively at a 2:l ratio. If air is used as oxidant, the reformate gas will also contain about 40% nitrogen. Prereduced Cu/ZnO-based catalysts exhibit high activity for partial oxidation at temperatures in the 200-300 "C range.65-67*71*72 Precious metal catalysts have also received some a t t e n t i ~ n . ~ ~ Some @ ? ~ examples ~ - ~ ~ of references on partial oxidation of methanol are presented in Table 4. There are some significant differences in the catalytic functions of copper and palladium catalysts with respect to product distribution, by-product formation and the effect of oxygen partial pressure. Copper favours partial oxidation products, whereas CO formation is quite severe over palladium. A variety of by-products have been observed over copper, while only trace amounts of formaldehyde have been detected at low methanol conversions over palladium catalysts. There is also a more pronounced influence of oxygen partial pressure on the catalytic behaviour of copper catalysts.

CuO/ZnO/A1203 (BASF K3-110) 40 wt% CUO 40 wt% ZnO 20 wt% A1203

Cu/ZnO/A1203 35-40 wt% CU 45-50 wt% Zn 10-20 wt% A1

Cu/ZnO/A1203 Cu/Zn01ZrO2/Al2O3 (Cu + Zn)/(Al + Zr) = 1.6-3.8 (atomic ratio)

Reitz et al. 7o

Velu et a/.65

Velu et aE.9099'

2000

1999

1999

Fixed bed flow reactor (4 mm i.d.); 0.09-0.1 mg cat. 1.6-2.5 ml h-' liq. H20/CH30H(1.3:l) 10-20 ml min-' air; 43 ml min-' Ar 180-290 "C; atmospheric pressure Prereduction: H2, 300 "C, 3 h

Packed bed microreactor (4 mm i.d.); 0.1 g cat. 1.5-2.0 ml h-' liq. CH3OH; 10-20 ml min-' air; 40 ml min- Ar; 02/CH30H = 0.2-0.4 200-245 "C; atmospheric pressure Prereduction: H2, 300 "C, 2 h

Fused silica microreactor; 10-50 mg cat. 1O-4O% CH3OH; 1-12% 0 2 ; 1O-3O% H20 100 ml min-' 180-225 "C

Microreactor (4 mm i.d.); 0.1 g cat. WHSV = 5-50 h-' (based on CH30H) 220-300 "C; 100-1000 kPa Prereduction: HZ,450 "C, 1 h

Copper-alumina (commercial)

Newson et al.92

2000

Combined reforming; extremely low CO-levels; XRD, DRS, TPR and EPR studies

Partial oxidation; catalysts derived from CuZnAl-layered double hydroxides

Combined reforming; oxidation state of copper; kinetics

Combined reforming; autothermal conditions; kinetics

Packed bed microreactor (6 mm id.); 0.5 g cat. 8.7% CH30H (230 ml min-'); 02/CH30H = 0.5; 230-300 "C; atmospheric pressure Prereduction: 10Y0H2/N2,300 "C, 0.5 h

2% Pd/ZnO

Agrell et al. 73

2000

Partial oxidation; catalyst preparation in microemulsion; Cu-ZnO interactions Partial oxidation; catalyst preparation in microemulsion; Pd particle size

Packed bed microreactor (6 mm id.); 50 mg cat. 8.7% CH30H (230 ml min- '); 02/CH30H= 0.1-0.5; 160-330 "C; atmospheric pressure Prereduction: 10Y0H2/N2,250 "C, 1 h

30-70 wt% CdZnO Cu/ZnO/A1203(ICI 83-3 M)

Agrell et al. 71

2000

Description

Catalyst Reaction conditions

AuthorlRef:

Some examples of references describing partial oxidation and combined reforming of methanol

Year

Table 4

rg

4

2

2

k

m

xq

-s

? n

Microreactor; 50 mg cat.

Huang and Chred7

Huang and Wang72

1988

1986

0-100 wt% Cu/ZnO

35.7 wt% Cu/ZnO

Packed bed microreactor (6.5 mm id.); 0.2 g cat. 2.8 ml h-' liq. H20/CH30H;25 ml min-' Ar or 02/Ar. H20/CH30H = 1; 02/CH30H = 0-0.166 220-290 "C; atmospheric pressure Prereduction: 75% H2/Ar, 300 "C, 1 h

Packed bed microreactor (1 or 3 mm i.d.); 1.5-50 mg cat. 0.078 ml min- liq. H,O/CH,OH; 40 ml rnin-' Ar or OdAr 210-250 "C; atmospheric pressure Prereduction: H2, 260 "C

02/CH30H = 0.06 200-230 "C; atmospheric pressure Prereduction: 10% H2/N2,230 "C, 1.5 h

4 rnl h-' liq. CH30H, 12 ml min-' air;

Alejo et al. 66

1997

Cu/ZnO; Cu/ZnO/A1203 20-70 wt% CU 30-80 wt% Zn 5-15 wt% A1

Stainless steel tubular reactor (6.3 mm id.); 0.2 and 0.5 g cat. 21.2% CH30H (100 ml min-'); 02/CH30H = 0.3 and 0.5 230-270 "C; atmospheric pressure Prereduction: 10% H-JNz, 300 "C, 0.5 h

1 and 5 wt% PdZnO; Cubeiro and F i e r r 0 ~ ~ 3 ~ 1 wt% PdZrO2; Cu/ZnO/A1203(BASF 3110)

Reaction conditions

1998

Catalyst

AuthorlRef:

(contd.)

Year

Table 4

Combined reforming

Combined reforming; reaction mechanism; kinetics

Partial oxidation; catalyst composition and lifetime; N20 as oxidant

Partial oxidation; PdZn alloy formation; TPR, XRD and XPS studies

Description

3: Catalytic Hydrogen Generationfrom Methanol

99

The complete combustion of methanol to water and C 0 2 is another important topic of research. Methanol is one of the world’s largest commodity chemicals, finding uses in many industrial applications, and has been suggested as an alternative fuel in automobiles. It is important to control the emissions of unburned methanol from industrial processes and vehicles. Most existing catalysts for combustion of methanol contain precious metals or silver supported on alumina.76For instance, palladium,77r h o d i ~ m ,p~l ~a ,t ~i n~~ m ~ ~ 9 ~ ~ and goldg0catalysts have been studied. Other catalyst systems under investigation include nickel oxide.81.82Although an important subject, the complete oxidation of methanol will not be discussed in detail in this review.

4.2 Copper-based Catalysts. - 4.2.1 Reaction Pathway. The route for partial oxidation of methanol over copper catalysts may be quite complex. Several different reactions involving methanol are catalysed by copper, e.g. steam reforming, partial oxidation, decomposition, water-gas shift and other oxidation reactions, such as the combustion of CO and H2. The H2 selectivity exhibits a strong dependence on methanol conversion, suggesting that oxidation and reforming reactions occur conse~utively.~~ Some insight into the reaction pathway may be gained by studying the production of formaldehyde, a compound which can be formed by partial oxidation of methanol over a copper-based catalyst: 2CH30H + 5 0 2 + 2HCHO + H2 + H20 At temperatures above 200°C, formaldehyde decomposes to CO and H2. Hence, the product has to be rapidly quenched in order to avoid further decomposition.67 In a recent paper, Velu et aZ.65 studied the partial oxidation of methanol over Cu/ZnO/A1203 catalysts. The authors noted that the reaction in absence of oxygen gave a mixture of formaldehyde, CO and H2, with traces of methyl formate. Consequently, the reaction scheme below was proposed, a route already suggested by Huang and Wang:72 CH30H + 5 0 2 + H2 + CO + H20 CH30H -+ 2H2 + CO CO + H20 + C02 + H2 In essence, the overall reaction for partial oxidation: CH30H + i

0 2 -+ 2H2 + C02

consists of all three reactions above, while the conventional steam reforming route includes only the decomposition and water-gas shift (WGS) reactions. In fact, the decomposition-WGS reaction sequence has been suggested by several authors for the steam reforming of methanol over copper catal y s t ~ . ~ ~Based - ~ ~on* the ~ ~discussion * ~ ~ above, it may be suggested that the steam reforming route is part of the partial oxidation reaction ~ c h e m e . ~ ~ 3 ~ ~ . ~ Alejo et aZ.66came to that conclusion, i.e. that at least part of the H2 and C02

100

Catalysis

are not formed in a single step by partial oxidation, but as secondary products from other reactions. They speculated that the water-gas shift reaction most likely contributes to formation of H2 and C02 at the expense of CO and water. 4.2.2 By-product Formation. H2 and C02 are the main and desired products in the partial oxidation of methanol. Other compounds that may be expected out of the reactor are CO and water, as well as unconverted methanol. Formaldehyde, formic acid, methyl formate, dimethyl ether (DME) and methane are other possible by-products in reactions involving methanol over copper catalysts. During partial oxidation over copper-zinc catalysts, Alejo et aZ.66 observed none of the above by-products. In our own laboratory, however, we have observed trace amounts of formaldehyde over C ~ d Z n 0 Velu . ~ ~ et aZ.,65 who studied catalysts derived from Cu-Zn-A1 layered double hydroxides (LDH), concluded that precursors containing hydroxycarbonates other than LDH as major phase produced considerable amounts of DME. DME formation over Cu/ZnO/A1203may be ascribed to the residual acidity of A1203in the basic CdZnO

4.2.3 Reaction Mechanism. The following reaction mechanism has been suggested for production of formaldehyde by partial oxidation of methanol on a copper surface:83 CH30H (8) + O* + *

-+

CH30*+ OH*

CH30H (g) + OH* -+ CH30*+ H 2 0 2CH30* + 2 * -+ 2CH20* + 2 H* 2H* + H2 (g) + 2 * where * denotes an active site on the catalyst surface. Methoxide was reported as being the most abundant surface species. Since very little methanol is adsorbed on an oxygen-free copper surface, this mechanism can be completely ascribed to the partial oxidation of methanol. Considering the similarities between the production of formaldehyde and H2 by methanol partial oxidation, the mechanism below has been adopted for H2 production over Cu/ zn0:67 0 2

(8) + 2 * + 20"

CH30H (g) + 0"+ *

-+

CH30*+ OH*

CH30H (g) + OH* -+ CH30* + H20 2CH30* + 2CH20* + H2 (g) CH20*

-+

CO (g) + H2 (g) + *

co (g) + o*+ c02 (g) + *

3: Catalytic Hydrogen Generationfrom Methanol

101

The first step consists of the dissociative chemisorption of oxygen on copper. In the following two steps, methanol undergoes dissociative adsorption on copper to form methoxy species (CH30*). In the next two steps, surface methoxide forms formaldehyde, which is then decomposed to CO and H2. The last two reactions combined form the water-gas shift reaction. However, at high surface concentrations of oxygen, the oxidation of CO and H2 would be the predominant reactions. As in formaldehyde production, methoxide was assumed to be the most abundant surface intermediate.67 In recent insight into the rate limiting step of the partial oxidation reaction over copper-zinc was gained by a study of the kinetic isotope effect (KIE). It was noted that production of H2 from CH30H and CH30D over prereduced CdZnO at 215 "C was quite similar with respect to selectivities and KIE data. Hence, no KIE was observed, indicating that the 0-H bond in the methanol molecule does not participate in the rate-limiting step of methanol conversion into H2. 4.2.4 Reaction Temperature. During partial oxidation, both methanol conversion and product distribution depend on the reaction temperature. The methanol conversion generally increases with reaction temperature. The H2 selectivity follows a similar trend, increasing at the expense of water, although the influence of temperature on H2 yield is much less pronounced at higher oxygen to methanol ratios. The C02 selecetivity, on the other hand, decreases due to a greater influence of methanol d e c o m p ~ s i t i o n . ~ ~ * ~ ~ ~ ~ * ~ ~ ~ Alejo et aZ.66 studied the partial oxidation of methanol over copper and found that at an oxygen to methanol ratio of 0.06, the reaction ignited at 215 "C and the rates of methanol and oxygen conversion strongly increased with temperature to produce H2 and C 0 2 selectively. The oxygen conversion reached 100% at 220°C. The rate of CO formation was low throughout the examined temperature interval and the water formation rate decreased at temperatures above 2 15 "C. In our own laboratory, we studied the partial oxidation of methanol over CdZnO catalysts prepared by microemulsion technique.71At a molar oxygen to methanol ratio of 0.1, the reaction ignited at 185 "C, which should be compared to 215 "C over catalysts prepared by conventional co-precipitation.66*71Oxygen was fully converted at 190"C over the microemulsion catalysts. The high activity was discussed in terms of a strong Cu-ZnO interaction. Over CdZnO/Al2O3,methanol conversions of 40-60% have been observed at 200 "C, with virtually no CO formation. With increasing temperature, however, the CO production rate increases rapidly at the expense of COZ.~'

4.2.5 Oxygen Partial Pressure. The partial pressure of oxygen is also very important in determining the product distribution and the reaction rate during partial oxidation of methano1.65~66*68*6g~71 The methanol conversion increases with increasing oxygen-methanol ratio. In general, excess methanol, i e . a molar oxygen to methanol ratio well below

102

Catalysis

0.5, favours partial oxidation products, while oxygen concentrations close to or above stoichiometry favour combustion products. When operating under stoichiometric conditions, the product gas contains a mixture of H2, C02, water and CO, due to the complete combustion of a fraction of the methanol. At high oxygen-methanol ratios, the C 0 2 selectivity remains high. However, this occurs at the expense of a decreased H2 selectivity, with a consequent increase in the water selectivity, due to the loss of H2 through oxidation: H2 + @2

+

H20

Decomposition products predominate when methanol is in great excess. In complete absence of oxygen in the feed, methanol decomposition occurs over copper-based catalysts. Therefore, it can be concluded that partial oxidation and decomposition occur simultaneously when the oxygen partial pressure is maintained well below stoichiometry.68 F i e r r ~noted ~ ~ that at constant temperature (215°C) the rates of methanol consumption and H2 and C02 formation over Cu/ZnO increase upon increasing the oxygen partial pressure from 0 to 0.05 bar. Upon a further increase up to 0.10 bar, there is a sharp decrease in methanol conversion as well as in H2 and C02 formation. By further increasing the oxygen pressure to 0.20 bar, there is no significant change in activity, i.e. there is an optimum oxygen pressure, at which the H2 formation rate reaches a maximum. The catalyst does not recover its initial activity on again decreasing the oxygen concentration. The irreversibility suggests that the copper surface is oxidised at oxygen pressures exceeding 0.05 bar, thus deactivating the catalyst for H2 formation. The oxidation of the catalyst surface was evidenced by X-ray photoelectron spectroscopy (XPS). The unreduced catalyst was inactive for H2 production, yielding mainly water and C02. 4.2.6 Catalyst Composition, There is a general agreement that there exists an optimum composition of copper-zinc in the catalyst between 1:2 and 2:3 (weight ratio), at which good catalytic performance is obtained for the partial oxidation of m e t h a n 0 1 . ~ ~For ~ ~ ~instance, ~ ’ ~ Alejo and systematically varied the Cu:Zn ratio of Cu/ZnO and Cu/ZnO/A1203catalysts and found a direct correlation between catalytic activity for partial oxidation and metallic copper surface area. In the series of catalysts, the one containing Cu:Zn at a 2:3 ratio displayed the highest metallic copper area, as well as the best performance. This is comparable to the composition of the commercial low-pressure methanol synthesis catalyst. An increased copper dispersion in copper-zinc catalysts can be obtained by the addition of y-alumina, a high-surface area metal oxide. For instance, Velu et aZ.65 studied methanol partial oxidation over Cu/ZnO/A1203catalysts and found that the catalytic activity depended on the (Cu+Zn)/Al ratio. The activity could be correlated with the metallic copper surface area, which was dependent on copper dispersion, which in turn depended on the catalyst composition. Copper surface area and dispersion increased with decreasing (Cu + Zn)/Al ratio. Hence, the catalytic activity increased in the same order.

3: Catalytic Hydrogen Generationfrom Methanol

103

4.2.7 Catalyst Stability. Copper-zinc oxide catalysts are known to deactivate quickly during time on-stream. Addition of alumina improves the catalyst stability, but has a slightly inhibiting effect on methanol conversion, whereas enhanced H2 and C02 selectivities are obtained. This was demonstrated by Alejo et al.,66 who found that a copper-zinc catalyst without alumina rapidly deactivated after 20 h of operation at 230 "C, while a catalyst containing 5 wt% aluminium (Cu/Zn/Al=40:55:5) could withstand 110 h of operation at the same temperature with no significant change in activity or selectivity. In order to obtain the increased stability by addition of alumina, but avoid the penalty in methanol conversion, Velu et aZ.65prepared Cu-Zn-A1 ternary oxides obtained by thermal decomposition of Cu-Zn-A1 hydroxycarbonates, containing hydrotalcite as major phase. Catalyst stability tests were carried out at 200°C for 24 h. After a small initial deactivation observed over most catalysts, accompanied by a small drop in H2 selectivity, the activity and selectivity remained more or less unaffected during 24 h of operation. 4.2.8 The Oxidation State of Copper. There has been great controversy regarding the nature of the active site in reactions involving methanol over copper-based catalysts.72Alejo et aZ.66 observed that if copper-zinc catalysts are not reduced prior to reaction, a very low activity for partial oxidation methanol is obtained, forming mainly C02 and water. Catalyst characterisation by XPS and Auger spectroscopy revealed that the active catalyst surface during partial oxidation consisted of metallic copper, and in some cases very small amounts of Cu2+. These researchers suggested that metallic copper is active for production of H2 and C02, whereas Cu+ favours formation of water and CO, and Cu2+shows low activity, producing mainly C02 and water. This is in full agreement with observations during attempted partial oxidation over unreduced CuO/ZnO/A1203, where C02 and water were the main products, due to complete oxidation of methanol?" However, recent XPS investigations of prereduced CdZnO catalysts after exposure to 02/CH30H mixtures (1:2 molar) at 250°C and 60 mbar have indicated that surface Cu+ species are involved in the transformation of methanol to H2 and C02.84 4.2.9 Cu-ZnO Synergistic Effects. In the literature, there is an ongoing discussion regarding the exact nature of possible Cu-ZnO interactions in the low-pressure methanol synthesis catalyst (Cu/ZnO/A1203).A synergistic effect has been reported by several researchers and some explanations of the mechanism are (i) an active site located at the Cu-ZnO interface, (ii) a modification of the electronic properties of copper by contact with zinc oxide, (iii) a stabilisation of the active crystallographic planes of copper by zinc oxide or (iv) a spillover-mechanism between copper and zinc oxide.8y9*85-88 In our own laboratory, we studied Cu/ZnO catalysts prepared in water-inoil microemulsions.71~84~89 A combination of techniques, i.e. N20 chemisorption, XRD, TPR-TPO and TEM, were used to correlate the physicochemial properties of the catalysts with catalytic activity. High activity for partial oxidation of methanol was observed, although the copper dispersion was lower

104

Catalysis

in the microemulsion catalysts than in conventionally co-precipitated catalysts. The high activity was not expected considering the low number of exposed copper sites. Therefore, the existence of a strong interaction between copper and the zinc oxide lattice was suggested. The microemulsion technique is known to produce materials that are homogeneous at the nanometre scale with copper and zinc in close proximity, thus enhancing possible Cu-ZnO synergies. It is suggested that catalyst redox properties, and the ability to form surface Cu+ species, are important characteristics for high activity in the partial oxidation reaction.

4.3 Palladium Catalysts. - Group 10 noble metals, such as palladium and platinum, are active in methanol transformation reactions, although they are less selective than copper in steam reforming, yielding primarily the decomposition p r o d u ~ t sThe . ~differences ~ ~ ~ ~ in~ catalytic ~ ~ ~ performance ~ of copper and palladium catalysts have been ascribed to the difference in reactivity of formaldehyde intermediates over the respective surfaces. However, in studies by Takezawa and ~ o l l e a g u e s , ~exceptional ~5~~ performance of PdZnO for steam reforming of methanol has been reported, a reaction otherwise conducted with high selectivity over copper catalysts. 4.3.1 Catalytic Behaviour. Cubeiro and F i e r r 0 ~ ~conducted 9~ catalytic tests

over PdZnO and Pd/Zr02 catalysts at molar oxygen to methanol ratios of 0.3 and 0.5 and temperatures between 230 and 270°C. H2, C02, CO, water, unconverted methanol and traces of formaldehyde at low methanol conversions were the only products detected. The oxygen consumption was complete in the entire temperature interval and methanol conversions reached 40-80%. At a molar oxygen to methanol ratio of 0.3, a H2 selectivity of 96% could be obtained at a methanol conversion of 70%. Upon increasing the partial pressure of oxygen, higher methanol conversions were required to obtain the same H2 selectivity. C02 formation dominated, although a small proportion of CO was observed. The methanol conversion and product distribution exhibit a strong dependence on oxygen to methanol ratio and reaction t e m p e r a t ~ r e . At ~ ~low $~~~~ conversions, methanol combustion dominates, the molar ratio of oxygen/ methanol consumed being higher than 0.5. As the conversion increases and oxygen is completely converted, the participation of steam reforming becomes more important. When increasing the reaction temperature, the methanol conversion increases with a simultaneous increase in H2 selectivity at the expense of water. The selectivity of C02 also increases, although this increment was less marked. Since the oxygen consumption is complete over a wide temperature range, this selectivity trend suggests some contribution from steam reforming caused by water appearing from combustion of a fraction of the methanol in the feed. Work in our own laboratory has shown that the palladium particle size has a profound effect on the distribution of carbon oxides.73In general, large palladium particles favour CO formation.

3: Catalytic Hydrogen Generationfrom Methanol

105

4.3.2 PdZn Alloy. Takezawa and c o - w ~ r k e r sfound ~ ~ * ~that ~ reduction at high temperatures greatly improved the catalytic performance of PdZnO catalysts for methanol steam reforming, due to the formation of PdZn alloy. It was suggested that formaldehyde, which was formed over the alloy-containing catalysts, was effectively attacked by water and transformed into H2 and C 0 2 . On metallic palladium, on the other hand, formaldehyde was decomposed to H2 and CO. PdZn alloy was also observed by Cubeiro and Fierro during partial oxidation.63@Catalytic systems of 1-5% palladium supported on ZnO were studied and PdZn alloy was observed upon reduction already at moderate temperatures. The relative ease with which PdZn alloy was formed in a 1% Pd/ ZnO catalyst was explained in terms of a strong Pd-ZnO interaction due to a small palladium particle size. 4.3.3 The Support. Cubeiro and F i e r r 0 ~ ~observed 9~ some interesting differences in the product distribution over the ZnO- and Zr02-supported palladium catalysts, irrespective of the large difference in surface area. While a Pd/Zr02 catalyst displayed a slightly higher methanol conversion than its Pd/ZnO counterpart, H2 and CO selectivities were lower. The differences became even more pronounced at higher oxygen to methanol ratios and higher methanol conversions. It was proposed that steam reforming may occur over PdZnO when all oxygen is consumed, i.e. the reaction proceeds via consecutive oxidation-steam reforming steps. This appears not to be the case over Pd/ ZrO2. The higher rate of CO formation over PdZrO2 suggests that the decomposition reaction occurs to a much greater extent.

4.4 Combined Reforming over Copper-based Catalysts. - 4.4.I Background. In 1986, Huang and Wang72proposed a new route for production of H2 from methanol, based on the combined reforming of methanol over a copper-based catalyst. This route was exothermic and has, as will be shown further on, some additional advantages over the conventional steam reforming route. It was speculated that addition of oxygen in the feed during methanol steam reforming may result in modifications of the copper-zinc catalyst surface, causing an increase in the rate of methanol decomposition and, consequently, enhanced steam reforming activity. By adding oxygen or air to the feed of water and methanol, the proposed process is a combination of partial oxidation and steam reforming. The process is also known as oxidative steam reforming or - when operated under close to adiabatic conditions - autothermal reforming. In recent years, the number of publications on this topic has rapidly increased. Several research groups are intensively studying the process, both on the catalytic and the reaction engineering scale, to produce H2 for stationary and mobile fuel cell applications. In essence, high H2 production rates can be obtained by simultaneously feeding oxygen, steam and methanol to a copper-based catalyst. Under appropriate conditions, H2 selectivities are very high and CO is virtually nonexistent. Additionally, by combining partial oxidation and steam reforming,

106

Catalysis

the ratio of the reactants can be varied and chosen such that the overall reaction is thermally neutral or moderately exothermic. In other words, the heat necessary to sustain the steam reforming process is supplied by oxidation reactions. 69370

4.4.2 Overall Reaction. The overall equation for combined reforming of methanol is given below:

where s and p are stoichiometric coefficients for steam reforming and partial oxidation, respectively. Velu et al.90991recently studied combinatorial reforming over Cu/ZnO/A1203 catalysts with and without Zr02. The ratio of H2 production to methanol consumption was found to be around 2.5. Newson and c o - ~ o r k e r s who ,~~ studied autothermal reforming over a copper-alumina catalyst, measured a 2.2 ratio of H2 production to methanol consumption. As mentioned previously, the net heat of reaction can be varied from endothermic to thermally neutral and exothermic by varying the ratio of the reactants. 4.4.3 Oxygen and Steam Partial Pressures. Without oxygen present in the gas phase, only steam reforming can occur if steam and methanol are fed to a completely reduced copper surface. On a partially oxidised surface, however, both partial oxidation and steam reforming can occur. Similarly, with oxygen present in the gas phase, the extent of steam reforming will depend on the partial pressure of oxygen. In the early work by Huang and Wang,72 the addition of oxygen to methanol steam reforming over copper-zinc catalysts was studied. It was observed that as the amount of added oxygen increased, the rate of H2 production increased. However, the H2 to C02 ratio decreased, due to the greater influence of oxidation reactions. The CO to C02 molar ratio was generally below 0.015, even without oxygen present, indicating that the equilibrium of the water-gas shift reaction was approached. Additionally, with oxygen present, CO may be directly oxidised to C02 and the shift reaction may be less important. It was also observed that the rate of H2 production by partial oxidation generally was higher than that of the steam reforming route over a wide range of catalyst compositions. Velu et ~ 1studied . ~ the~ influence of both oxygen and steam partial pressures on the catalytic performance during combined reforming. It was found that the rates of methanol consumption and H2 production increase with increasing oxygen-methanol ratio up to about 0.3 (molar) and then level off. The H2 selectivity decreases with increasing oxygen-methanol ratio. In other words, the steam reforming reaction dominates at lower oxygen-methanol ratios, while mainly partial oxidation occurs at higher ratios. The selectivity of COz increases with increasing oxygen-methanol ratio, at the expense of H2 selectivity. Based on their investigation, the molar oxygen-methanol ratio should be kept at approximately 0.2 for optimum catalyst performance. Furthermore,

3: Catalytic Hydrogen Generationfrom Methanol

107

the effect of steam-methanol ratio was studied by keeping the oxygenmethanol ratio constant at 0.2. Upon an increase in steam-methanol ratio from 0.4 to 1.3, the rate of H2 production and turnover frequency (TOF) both increased. The ratio of H2 production to methanol consumption was around 2 at lower partial pressures of steam, increasing to around 2.6 at a steammethanol ratio of 1.3, which was found to be the optimum. 4.4.4 Reaction Temperature. By carrying out combined reforming over CuZn-A1 and Cu-Zn-Al-Zr oxide catalysts, Velu et al.90991report having produced a virtually CO-free reformate and achieving complete methanol conversion at 230 "C. For pure steam reforming, complete methanol conversion was reached at 290°C. It was observed that due to the high methanol conversion, the H2 production rate was about twice as high in the case of combined reforming, as compared to pure steam reforming or partial oxidation. In pure steam reforming CO was formed at temperatures exceeding 260 "C, while it started to form at 200 "C during partial oxidation. The authors further found that addition of alumina increases catalyst stability, but the activity decreases with increasing aluminium content. The best catalytic activity was obtained over catalysts containing zirconia instead of alumina.

4.4.5 Catalyst Stability and the Oxidation State of Copper. Reitz and colleag u e ~ ~studied ~ , ~an~ unreduced , ~ ~ commercial CuO/ZnO/A1203 catalyst and found a strong correlation between catalyst performance and the oxidation state of copper. In the presence of oxygen, when copper is in the Cu2+ state, combustion of methanol forming C02 and water was predominant, exhibiting only minor H2 formation. When all oxygen was consumed, the catalyst was reduced and the endothermic steam reforming reaction proceeded with a marked increase in H2 production. Moreover, thermal deactivation was found to occur with time on-stream. At 225 "C and a molar oxygen to methanol ratio of 0.28, 80% of the activity was lost in 20 h. Both methanol and oxygen conversion decreased, but there was little change in the distribution of products. 4.4.6 By-product Formation. CO is the major by-product observed during combined reforming of methanol. Huang and Wang72 observed trace amounts of methyl formate over copper-zinc catalysts, while Reitz and cow o r k e r ~detected ~~ some formic acid over a fresh, unreduced CuO/ZnO/A1203 catalyst. 4.4.7 Kinetics. Huang and Chred7 have presented a kinetic model for the combined reforming of methanol over a copper-zinc oxide catalyst. The model assumes methoxide to be the most abundant surface species and the dissociative adsorption of methanol to be the rate-determining step. The authors employed the simple relationship

overall rate = partial oxidation rate + steam reforming rate in order to calculate the partial oxidation conversion from combined reforming

Catalysis

108

measurements at 210°C. At such a low temperature, the contribution from steam reforming is small. At higher temperatures, where the contribution from steam reforming is much more significant, the methanol conversion for combined reforming increased from below that of steam reforming to well above it upon an increase in contact time. The increase in conversion was accompanied by a complete consumption of the available oxygen. For this reason, only experimental data at 210 "C were used to derive a rate equation, using the Langmuir-Hinshelwood-Hougen-Watsonexpression. The kinetic model agreed well with experimental results. The partial oxidation reaction rate was second order in methanol and had a pronounced maximum with respect to oxygen concentration. The rate equation is presented below:

At 210 "C, the kinetic parameters were found to be:

k = 2.7 x 1013cm7.5m01-l.~min-' g cat-' Kl = 3.4 x 10l2cm7.5m ~ l - ~ . ~

K2 = 1.5 x lo3 cm1.5m ~ l - ~ . ~ More recently, Newson et aLg2studied the kinetics of autothermal reforming of methanol. The authors developed a kinetic model derived on the basis of six kinetic constants, based on isothermal measurements over a commercial copper-alumina catalyst. The reaction system consisted of (i) DME formation, (ii) methanol decomposition, (iii) WGS, (iv) steam reforming, (v) partial oxidation and (vi) the complete oxidation of H2. However, it was simplified as it was found that the water-gas shift reaction was slow in comparison to the others, and the H2 oxidation was mass transfer limited over the catalyst used. The apparent activation energy for partial oxidation was found to be 65 kJ mol-' and the turnover frequency (TOF) at 250 "C was estimated to be 460 min-'. These results may be compared to those of Alejo et aZ.,66 who found that the apparent activation energies and TOFs over Cu/ZnO catalysts varied with Cu:Zn composition. High activation energies and TOFs were observed over catalysts with low copper content (e.g. 482 kJ mol- and 200 min- over 30 wt% Cu/ZnO), whereas a trend towards constant values was observed at higher copper contents (e.g. 71 kJ mol-' and 160 min-' over 70 wt% Cu/ ZnO). The simultaneous variation in activation energy and TOF was explained in terms of possible Cu-ZnO interactions, depending on the catalyst composition. Catalysts with lower copper content contained smaller copper particles. The smaller the particle size, the larger the Cu-ZnO interface and any promotional effects arising therefrom. Reitz et aL70 studied the combined reforming over an unreduced CuO/ZnO/ A1203catalyst. For temperatures between 180 and 225 "C, the kinetics of the reaction were found to follow the power-law expression below:

'

3: Catalytic Hydrogen Generationfrom Methanol

109

The pre-exponential factor (Ao) was calculated as 6.0 x 10' mol min-' g cat-' kPa-0.22 and the apparent activation energy (E,) was reported to be 115 kJ mol-'. 5

Methanol Reforming for Fuel Cell Applications

5.1 Introduction. - Methanol reforming for H2 production may be used in a variety of commercial and industrial applications. Production of a nearly COfree reformate for fuel cell vehicles is one example, which has attracted increasing attention in recent years In simple terms, a fuel cell is an electrochemical device that continuously converts chemical energy from an external fuel and oxidant directly to electrical energy. The type of fuel cell, which is currently under consideration by the automotive industry as an alternative to the combustion engine, is the polymer electrolyte membrane fuel cell (PEMFC), also known as the proton exchange membrane fuel cell or the solid polymer fuel cell (SPFC). In the PEM fuel cell, H2 is dissociated over a platinum-based catalyst at the anode. The protons are conducted through the polymeric membrane, while the electrons travel through an external circuit. By feeding oxygen or air to the cathode, water and heat are produced: Anode:

H2 + 2H+ + 2e-

Cathode:

2H+ + $ 0 2 + 2e-

H20

-+

In a fuel cell vehicle, careful design and operation of the methanol reformer is necessary for optimum performance with respect to reformate composition and system efficiency. The mobility aspect in particular requires engineering for fast start-up and dynamic transient response. Furthermore, there is a need to eliminate external utility and power supplies as far as possible. Selecting an appropriate catalyst for a fuel cell powered vehicle introduces criteria exceeding the traditional chemical process requirements of high selectivity and conversion. The requirements for implementing a catalytic solution in automotive applications are presented in Table 5 .4796-99

5.2 Partial Oxidation versus Steam Reforming Technology. - The steam reforming reaction is endothermic and requires a large heat input, making the process both capital and energy intensive. Much can be gained by turning to partial oxidation or combined reforming technology. However, there are advantages and disadvantages associated with all three types of processes. 5.2.I Steam Reforming. Steam reforming yields higher H2 concentrations than partial oxidation by producing H2 from both methanol and water in the feed. Being an endothermic reaction, surplus energy from other parts of the system

Catalysis

110

Table 5

Requirements for implementing a catalytic solution in automotive appli~ation.P-~~-~~

Catalyst requirements

Reformer system requirements

High activity and selectivity Low ignition temperature Resistance to poisoning, coking and sintering Uniform composition of outlet gas independent of operating temperature Resistance to temperature cycling Good mechanical properties Good heat transfer characteristics Low production cost

Fast start-up Dynamic response Low pressure drop Compact

Figure 3

Light-weight High efficiency Good heat transfer characteristics Low production cost

Steam reforming system design

can be usefully recycled, promoting overall system efficiency. However, the reaction is slow and requires a large reactor and long residence time^.^^^^^^^^^^ The steam reforming system is presented in Figure 3. A steam reformer in a fuel cell vehicle requires external combustion of recycled exhaust from the fuel cell anode, a fraction of the fuel-methanol or some other fuel to generate heat for the reaction, which is provided via indirect heat transfer. Whereas recycling of anode off-gases promotes system integration and overall efficiency, there is a penalty in transient response, i.e. a time lag between energy requirement and availability. In other words, start-up is slow, the transient response poor and the reactor heavy and bulky due to the weight and volume of heat-exchange components.102~103

5.2.2 Partial Oxidation. Partial oxidation is a fast process, resulting in a small reactor size, rapid start-up and dynamic response. There is direct transfer of heat generated by oxidation of methanol in the catalyst bed. Moreover, the reactor can be made even more compact and light-weight, due to the absence of heat-exchanger parts. However, being an exothermic reaction, it can lead to low vehicle efficiency if the generated heat is wasted. External heat exchange is not required in the partial oxidation system, although some sort of catalytic afterburner is recommended to avoid emissions of H2 to the The partial oxidation system is described in Figure 4. atmosphere. 9771009103

3: Catalytic Hydrogen Generationfrom Methanol

Figure 4

111

Partial oxidation system design

5.2.3 Combined Reforming. Combined reforming technology combines the best features of both processing methods - the speed and compactness of partial oxidation with the high efficiency of steam reforming. By supplying a feed of methanol, water and air, the two processes can be made to sustain each other and reforming can occur under almost autothermal conditions. The balance between the two reactions can be tuned to match the power demand of the system at a given time. Less steam is needed by introducing air as oxidant, reducing the weight and size of the fuel processor, as well as the energy needed to vaporise the feed and bring the vapour to the reaction temperature. As both endothermic and exothermic reactions occur on the same catalyst, heat transfer occurs over very short distances and heat transfer limitations are minimised.96,97,100,102,103 The reformate from a steam reformer contains 70-75% HZ. In a combined reformer, the H2 concentration is somewhat lower, due to the use of air as oxidant. This difference in H2 concentration leads to a penalty in open circuit fuel cell p ~ t e n t i a l . ~Moreover, ~ $ ' ~ ~ the high mass flow of reformate through the fuel cell anode chamber in a combined system causes an additional problem. Being more diluted, it will be significantly greater than in a steam reformer system consuming the same amount of H2. As a consequence, wider flow channels are required and the power density of the fuel cell decreases.

CO Clean-up. - In most practical cases, the H2 feedstock for a PEM fuel cell will be obtained by on-board reforming of a hydrogen-rich fuel, such as methanol. A small amount of CO, typically 0.5-1%, is formed irrespective of methanol conversion process. The reformate gas also contains C02, water and usually small amounts of by-products and unconverted methanol. CO-levels in excess of a few ppm will poison the PEM fuel cell by blocking the active sites of the platinum-based anode electrocatalyst, thus hampering the fuel cell performance. Therefore, CO removal is essential. Possible strategies include oxidation, reduction (methanation), adsorption or the use of palladium membranes. Adsorption, however, typically requires an unacceptably large volume of adsorbent and will not be discussed in further detail. The operating 5.3

112

Catalysis

temperature of the clean-up step should preferably be compatible with the outlet temperature of the reformer (200-300°C) and the inlet temperature of the PEM fuel cell (90-1 10 "C). A thorough investigation into the kinetics and reaction paths of the reforming reaction needs to be carried out in order to determine the required size of the clean-up step. 5.3.1 Water-Gas Shift. Water-gas shift (WGS)is used to convert CO and water to C 0 2 and H2.1W-lo9The process not only reduces the amount of CO, but also increases the yield of H2. Water-gas shift typically constitutes a first cleanup step, which needs to be followed by secondary CO removal in order to reach ppm levels. Being exothermic, the reaction is suppressed at higher temperatures. There is a disadvantage in terms of the considerable amount of energy required for evaporation of additional water to drive the equilibrium towards low CO levels. The reaction has been extensively studied in the low(180-250 "C) and mid-temperature range (220-350 "C) over copper-based catalysts.'04-lo6 A1203and Cr203-containing CdZnO catalysts have been used commercially since the early 1960s. Several manufacturers, e.g. BASF, ICI, Sud-Chemie and Haldor Topwe, sell water-gas shift catalysts commercially.

5.3.2 Selective CO Oxidation. Selective CO oxidation, sometimes referred to as preferential oxidation, involves oxidation of CO in the presence of excess ~~:110-115

co + $ 0 2 + c02 Major efforts have been focused on alumina-supported platinum catalysts.113-115 This type of catalyst can effectively catalyse the oxidation of CO in presence of excess H2, but the reaction temperature and the rate of oxygen addition need to be carefully controlled. The stoichiometrically required amount of oxygen to convert 1% CO in the reformate is 0.5%. However, higher oxygen concentrations are generally needed, leading to an increased risk for an incident explosion and/or combustion of H2: H2 + 4 0 2 --+ H20 Exceptional selectivity for CO oxidation is essential for purification of the fuel cell feed stream with a minimum loss of H2. In addition, both CO and H2 oxidation reactions are highly exothermic and it is essential to remove heat from the reactor in order to control the activity and selectivity of the process. Other catalyst systems have been reported to be more efficient than the conventional Pt/A1203 catalyst, for instance platinum supported on zeolites' l 2 and ruthenium and rhodium supported on alumina. 139' l4 Nearly complete CO conversion can be achieved at temperatures as low as 100°C. Gold is another precious metal that has been investigated. Base metal catalysts, including l4 copper, have also received some attention.

'

'

5.3.3 Methanation. Reduction of CO by methanati~n"~-"~ appears to be an attractive alternative. In this process, CO is converted into methane by reaction with H2:

3: Catalytic Hydrogen Generationfrom Methanol

113

CO + 3H2 + CH4 + H20 The reaction has been studied over alumina-supported nickel, ruthenium and rhodium catalysts.' l 6 9 l l 7 However, methanation is not suitable for removal of large amounts of CO. As each converted CO molecule consumes three H2 molecules, the system will suffer a considerable H2 penalty. In addition, the presence of C02 in the gas stream may severely compromise the process by consuming large amounts of H2: C02 + 4 H2 + CH4 + 2 H20 5.3.4 Palladium Membranes. Dense palladium and palladium-silver alloy membranes exhibit infinite H2 permselectivity. They may be used for bulk separation to produce ultra-pure H2 for the fuel cell or in a membrane reactor incorporating the reforming catalyst to carry out simultaneous reaction and separation with enhanced methanol conversion.120-125 Several membrane configurations are possible, but the most common option for high performance, long lifetime and minimum cost is to support a pm-thin film of palladium on a porous tubular substrate. Available membrane preparation techniques include electroless plating, chemical vapour deposition (CVD) and magnetron sputtering.125 Porous alumina and stainless steel are widely used supports. A difference in H2 partial pressure between the feed side and the permeate side provides the driving force for separation. The pressure differential is generally achieved by operating the upstream side of the membrane at high pressure and/or by using a sweep gas on the permeate side. The use of a sweep gas, however, lowers the concentration of H2 in the feed stream to the fuel cell. H2 remaining in the retentate may be combusted externally to provide heat for the methanol reforming reaction. The main advantage of using palladium membrane technology is that ultrapure H2 is obtained. However, the process is somewhat compromised by both high material costs and a high operating pressure, typically in the region of 20 bar. By operating at high pressure, the size of the methanol reformer can be reduced, but methanol conversion is suppressed. Moreover, if air is used as oxidant, the additional need for a high-pressure compressor lowers the overall efficiency of the system. Some issues of poor sealing and membrane lifetime also remain to be resolved.

Industrial Activities. - Great research and development efforts are focused on development of automotive fuel reformers. In the USA, the Department of Energy (DOE) has launched a programme on fuel cell development for automotive applications with a total funding of 50 million USD.126 In the commercial sector, investments into fuel cell technology have markedly increased during the last years. The number of patents in this area also reflects this development (see Section 6 ) . Major chemical and automotive companies have recently entered into a cooperation agreement to facilitate the introduction and commercialisation of

5.4

114

Catalysis

methanol fuel cell ~ e h i c 1 e s . The l ~ ~ companies involved are BASF, BP Amoco, DaimlerChrysler, Methanex, Statoil and XCELLSIS. The goal is to establish a joint position after examining any health, safety, environmental and infrastructure issues associated with the use and introduction of methanol fuel cell vehicles. Two important alliances promoting fuel cell development are the California Fuel Cell Partnership (CaFCP) and the Partnership for a New Generation of Vehicles (PNGV). The latter do not focus on a specific propulsion system. The CaFCP is a voluntary alliance working to demonstrate and promote awareness of fuel cell technology. The alliance includes automobile manufacturers (DaimlerChrysler, Ford, General Motors, Honda, Hyundai, Nissan, Toyota, Volkswagen), energy providers (Atlantic Richfield, BP Amoco, Shell, and Texaco), fuel cell and fuel cell engine companies (Ballard Power Systems, International Fuel Cells, XCELLSIS), and government agencies (the California Air Resources Board, California Energy Commission, South Coast Air Quality Management District, US Department of Energy, and US Department of Transportation). Some examples of the industrial activities is discussed in this section and they are based on available information at the time of submitting this review (October 2000). For more in-depth information consult the papers by Kalhammer et a1.128and Pettersson and W e ~ t e r h o l m . ~ ~ 5.4.1 Argonne National Laboratory. The Argonne National Laboratory is working actively to create a reformer for fuel cell app1i~ations.l~~ From reports of their own tests, it seems that they are using a commercial copper-zinc oxide catalyst, which has properties that maximises the H2 production while minimising the production of by-products such as CO. It is also reported that tests are being performed on reforming of hydrocarbons, such as toluene and iso-octane, with relatively good results.130The fuel conversion process is called partial oxidation. Yet, in the reaction formulae that are presented, it is clear that a combination of partial oxidation and steam reforming takes place. The Argonne National Laboratory also has a key role in the US Department of Energy programme on reformer development, where they evaluate participating projects. A detailed description of the Argonne APOR fuel processor is presented in a separate case study (Section 5.4.16).

5.4.2 Arthur D Little. Through its subsidiary Epyx Corporation, Arthur D Little Corporation has developed a multi-fuel reformer funded by the US Department of Energy.128 The reactor system operates as a three-stage process. In the first part of the process, partial oxidation takes place, converting methanol into H2, CO and C02. In the second stage, the reformate is reacted with water vapour over a catalyst to promote the water-gas shift reaction converting CO into CO2 and releasing even more H2. In the last stage of the process the remaining CO is reacted with air in a catalytic reactor producing additional C02. Epyx has recently merged with the Italian fuel cell company De Nora Fuel Cells to form Nuvera Fuel Cells. The new company

3: Catalytic Hydrogen Generationfrom Methanol

115

will design and develop fuel processors, PEM fuel cell stacks and integrated fuel cell systems for stationary and transportation applications. 5.4.3 Ballard. At a glance, Ballard is involved in the construction of most fuel cell vehicles today. Their largest investment, however, lies in the Ballard/ DaimlerChrysler/Ford alliance which has created three jointly-owned companies: XCELLSIS Fuel Cell Engines, Ecostar Electric Drive Systems and Ballard Automotive. XCELLSIS Fuel Cell Engines is responsible for developing, manufacturing and commercialising fuel cell engines for buses, cars and trucks using fuel cells manufactured by Ballard Power Systems. XCELLSIS was formerly known as DBB Fuel Cell Engines. Thus, most of the intellectual property is registered under the former name.'31-133 Ecostar Electric Drive Systems is responsible for developing electric drive systems for electric vehicles. Ballard Automotive is responsible for selling fuel cells and fuel cell engines for vehicles to car manufacturers around the world. Ballard has also registered an integrated fuel cell for generating power in submarines. 134 However, Ballard's involvement in the development of fuel cell-based vehicles does not end with the joint ventures above. Ballard is currently developing fuel cells for the following automotive corporations: General Motors, Volkswagen, Honda and Nissan. 5.4.4 Daimler Chrysler . Since 1997, the DaimlerChrysler Corporation has been developing a family of fuel cell vehicles under the working title NECAR (New Electric Car). The fourth prototype is currently being tested (NECAR 4). This zero emission vehicle, which is based upon a Mercedes Benz A-class, can reach speeds up to 145 km h-l. At this stage, the fuel cells are powered by liquid H2. However, the aim of the project is to have a fully developed on-board reforming system. DaimlerChrysler is also involved in several joint ventures with Ballard and Ford. DaimlerChrysler is the automotive corporation with the largest patent holdings in reforming technology, ranging from reactors and clean-up devices to entire systems for automotive propulsion systems based on methanol reforming technology. It is also the leading company in terms of registering patents for the steam reforming of m e t h a n ~ l , ' ~ranging ~ - ~ ~from ~ endothermic reactors to full turnkey solutions for automotive reforming systems (see Section 6 for patent details). One example of a reforming reactor consists of three serially arranged reactors, each filled with catalytic material. 14* The centre reactor stage is heated and the inlet reactor has a heat exchange connection with the outside stage. The reformer is designed for the reforming of methanol by water vapour and the three stages are implemented to increase the productivity and reduce the formation of by-products that are detrimental to the fuel cell. 5.4.5 Ford. The Ford Motor Company is one of the leading global participants

116

Catalysis

in the fuel cell vehicle development. Ford is also working in the automotive fuel cell alliance with DaimlerChrysler and Ballard Power Systems. Mazda, which belongs to the Ford group has recently demonstrated a fuel cell vehicle (Premacy FC-EV) with a methanol reformer. The Premacy FC-EV has been developed in co-operation with Ford’s TH!NK group. 5.4.6 General Motors (GM) . GM’s German subsidiary OPEL has developed a fuel cell concept car based upon the reforming of methanol. The prototype, an OPEL Zafira, is reported to have zero emissions of SO2 and NO, and the C 0 2 output is expected to be 50% of an internal combustion engine. GM has also developed an electric vehicle, EVl (Electric Vehicle l), based solely upon batteries. The company also holds some patents on reformer technology. GM has designed a thermally integrated two-stage reformer for methanol,147 including a heat exchanger. The reactor utilises a gaseous heat transfer medium to pre-heat the inlet gases. The reformer uses the products from the primary reforming step as the heat transfer medium before entering the second stage, in which preferential oxidation takes place. The recirculation of heat by the gaseous products is a common characteristic for the compact reformers presented in this paper. Delphi Automotive Systems is a former subsidiary of General Motors (GM), now working independently with DaimlerChrysler and GM. Delphi is developing a reforming process using gasoline rather than methanol for the production of H2.128 5.4.7 Haldor Topsge. The Haldor Topsare Corporation in Denmark has patented a reactor for the specific purpose of endothermic gas phase react i o n ~ The . ~ ~reactor ~ has a tubular design, in which the required heat is transferred indirectly by heat exchange through the walls. The heat is obtained by allowing a medium to condense on the tube walls. The catalyst is placed in the reactor vessel’s main chamber in such a way that it spreads around the pipes in the vessel. The condensing gas then supplies the endothermic reaction with energy. There are two standard problems, which are solved with this design: (i) high heat transfer area and (ii) high catalyst volume relative to the reactor volume. Haldor Topsare has also been involved in several projects within the European non-nuclear energy programme, Joule-111. Haldor Topsare’s role in these programmes is usually as a developer of catalyst and reformer technology. 5.4.8 International Fuel Cell Corporation (IFC). IFC has developed a compact

reformer vessel with the original intent to serve as a feeder to a compact H2rich fuel ~ 1 a n t . l ~ The ’ vessel, however, could be implemented into the use in on-board reforming. The rather unique part of this invention is that it contains an annular burner, which is surrounded by the catalyst beds. The use of multiple beds enables the vessel to double the effectiveness of the catalyst. The heat in the vessel travels radially, which makes it possible to heat the catalyst

3: Catalytic Hydrogen Generationfrom Methanol

117

both from the inside and outside at the same time, without adding an extra heat source. 5.4.9 Johnson Matthey. The Johnson Matthey Corporation has invented its own fuel processing reformer technology, called the HotSpotTM reactor system.97The reformer system has been specifically designed for the on-board conversion of methanol to H2 and C02. According to G ~ l u n s k i ,the ~~ HotSpotTMreactor can produce fifty times its own volume of H2 per minute. The system combines both steam reforming and partial oxidation, using the partial oxidation reaction for cold start to generate the required heat for running the endothermic steam reforming reaction.97The HotSpotTMreactor system is the result of the evolution of a patented reformer solution from 1987.150This reactor is described in detail in a separate case study (see Section 5.4.15). 5.4.10 Toshiba. Toshiba has developed its own vessel for the catalytic conversion of methanol to HZ, which operates at temperatures as low as 205°C.151 The vessel operates in two stages: (i) the conversion of methanol into H2, C02 and CO gas in the first stage, followed by (ii) the water-gas shift reaction, converting CO into C02 and H2. The design solution chosen by Toshiba uses all of the exothermic energy from the water-gas shift reaction as an energy source for the endothermic conversion of methanol to H2. The result of this utilisation implies that the steam reforming reaction can be accelerated.

5.4.11 Toyota. Toyota has developed reformers for on-board applications using both gasoline and methanol for conventional engines. The Toyota system is a hybrid construction, which utilises the heat from the exhaust gases of the combustion engine to heat the catalyst bed in the reformer. The reformer operates with a combination of partial oxidation and decomposition.12* Recently, Toyota and GM have reached agreement regarding fuels for fuel cell vehicles, including technology collaboration with the ExxonMobil Corporation. The joint research activities of ExxonMobil, GM and Toyota are also focusing on propulsion system development, fuel processing technology, fuel cell R&D and H2 storage materials. 5.4.12 Volkswagen (VW) - Siid Chemie. Sud-Chemie and VW have jointly developed a compact reactor for on-board reforming.152 It was originally meant to function as a pre-reformer for a methanol combustion engine as it was found that the combustion engine worked more efficiently on reformed methanol than on pure methanol. In the original design the reforming was to take place on a Pt/Ti02 and Pt/Ce02 ~ a t a 1 y s t . The l ~ ~ energy required for the endothermic reforming reaction was to be supplied from the exhaust gases from the combustion engine. The reformed gases contained mainly H2 and CO. By adding water to reformed gas, the water-gas shift reaction could take place and C 0 2 and additional H2 could be produced. However, it was found that the produced amount of CO always exceeded the amount of C02. This

118

Catalysis

reactor could be set to perform autothermally with the addition of air. The main problem with this system is the high concentration of CO, which makes it impossible to use with PEM fuel cells at this stage. It has therefore been necessary for Sud-ChemieNW to re-design the applied catalyst to promote a conversion to C02 as the favourable product rather than the current route to

co.

5.4.13 Volvo. The Volvo Corporation is one of the most experienced contractors in the field of alternative fuel systems for the traditional combustion engine. Volvo has participated in several industrial projects for on-board reforming including a joint development project with Volkswagen 0, Johnson Matthey and ECN (the Netherlands Energy Research Foundation). Volvo is currently developing a reforming system in co-operation with the Department of Chemical Engineering and Technology at KTH - the Royal Institute of Technology in Stockholm. 5.4.14 Wellman - ECN. Wellman CJB (WCJB) is a company working in the

areas of O2 and H2 generation and purification. Longer-term development projects include thin film membrane H2 purifiers and multifuel processors for low temperature fuel cell applications, using a variety of feedstocks including alcohols, hydrocarbons and ethers. Various reformer configurations are being studied, i.e. packed beds and coated corrugated plates with catalytic or direct flame burning of fuel cell off-gases and/or feedstock/air. The Netherlands Energy Research Foundation (ECN) develop technology among other things for efficient, clean and sustainable use of fossil fuels. In the MERCATOX projectlS3 (Methanol Reforming and Catalytic Oxidation), part of the EC Joule-I11 program, WCJB and ECN have been working together to develop catalysts and technology for an integrated catalytic methanol steam reformer and high temperature selective oxidation gas clean-up system. An integrated approach has been adopted, using a metal-supported catalytic system combining the endothermic methanol steam reforming with the exothermic combustion of methanol or H2 in the fuel cell anode off-gases, thus ensuring good heat transfer characteristics. By using a heat-exchanger concept, inherent problems encountered in packed bed configurations resulting in poor performance, such as hot and cold spots, can be avoided. 5.4.15 Case Study I: The Johnson Matthey HotSpotTM Reactor. The HotSpotTM reactor was developed at Johnson Matthey in the mid-1980s. Jenkinslso describes the original version in a patent from 1987. More recently, Golunski and collaborators97~100~101 and Gray and PetchlS4 have described improvements of the system. The name HotSpotTMrefers to a highly localised reaction zone near the point of injection, which is deliberately incorporated in the catalyst bed. The temperature of the hot spot is stable around 600 "C, while temperatures in the remainder of the reactor are in the 200-300°C range.96 The reactor can be operated under close to autothermal conditions, obtaining almost complete single-pass fuel conversion and high H2 yields. The tech-

3: Catalytic Hydrogen Generationfrom Methanol

119

nology is most advanced for methanol processing, but the concept is also applicable to hydrocarbon fuels. The standard HotSpotTMmodule is a cylindrical 245 cm3 canister, producing up to 750 dm3 h-I of H2. Scale-up is accomplished by combining several reactors in parallel, feeding them from a central manifold block. Having a modular design, characteristics such as start-up and transient behaviour for the overall system are similar to those of a single reactor. An 8-unit HotSpotTM system has a volume of 6 dm3, weighs 8.8 kg and produces up to 6000 dm3 h-' of H2. The largest current HotSpotTMreactor (1999) has a power output of 18 kW, assuming that 1000 dm3 h-' of H2 will produce 1 kW of PEMFC power. In the HotSpotTMreactor, the liquid reactants are vaporised and mixed with air inside the manifold, which also functions as a heat exchanger, allowing surplus heat in the system to be utilised for vaporisation. For fast start-up, a lean feed is used. According to the inventors, full output can be achieved in less than a minute when feeding vaporised methanol and air into a cold reactor. Using a feed composition intended for autothermal operation, a cold reactor can reach steady-state in less than 5 minutes. Under endothermic operating conditions, heat input can be obtained by combustion of H2 in the anode off-gases. However, this requires external combustion and circulation of exhaust in the manifold. Methanol conversions are typically 99- 100%. In absence of steam, the reactor produces a reformate containing 41% H2 (dry basis). By introducing water and reducing the amount of air, the H2 concentration can reach 58% (dry basis). The amount of CO produced is highly dependent on reaction temperature. The first few seconds after start-up, CO formation is negligible, nevertheless reaching levels of 2-3% under autothermal steady-state operation. Multi-stage catalytic clean-up lowers the CO-level to below 10 ppm during steady-state and below 100 ppm during transients. The system does not produce large spikes of CO at any time during normal operation. 5.4.16 Case Study 11: The Argonne APOR Fuel Processor. Kumar and colleagues102.'03at Argonne National Laboratory have reported on the development of a catalytic partial oxidation reformer for fuel cell vehicles - the socalled APOR (Argonne National Laboratory partial oxidation reformer). The original concept of the system is found in a patent from 1993.155 Incorporating partial oxidation of methanol, the system is said to be faststarting and responsive to transients, as well as compact and light-weight. In absence of a separate methanovwater vaporiser, air-preheater, external heat exchanger and recycle loop for anode exhaust from the fuel cell, the design is fairly simple. The system requires no external heating or cooling and can be controlled by varying the feed rates of methanol, water and air to respond to the fluctuating power demand of the vehicle. The operation of a 10 kW bench-scale APOR has been demonstrated. The reactor was cylindrical in shape (D=75111111, L=600 mm) and contained a commercial copper-zinc catalyst. The catalyst required no prereduction and could be heated to 600°C without any significant loss in activity. Both pellet

120

Catalysis

and honeycomb catalysts were tested. Methanol and water are injected through an ultrasonic nozzle at the top of the reactor, into a down-flowing air stream. By exothermic methanol oxidation reactions, the temperature at the inlet of the catalyst bed is almost 500°C. As steam reforming, WGS and decomposition reactions occur, the temperature decreases and reaches about 200°C at the reactor outlet. The oxygen to methanol ratio is maintained slightly higher than what is required for thermal neutrality due to heat losses from the reactor and since the products leave at a temperature higher than the feed temperature. The reformate consists of more than 50% H2 and about 1% CO. Based on bench-scale results, the weight and volume of a 50 kW APOR system were estimated to be less than 35 kg and 25 dm3, respectively. 6

Intellectual Property

The protection of solutions for the direct conversion of primary alcohols to H2 started to expand in the first part of the 1980s. The methods described in most of the inventions prior to 1989 deal with decomposition of methanol to H2 and CO and the assignees are mostly chemical companies, rather than automotive, and the applications are not directly linked to transportation solutions. In the beginning of the 1990s, patents concerning steam reforming and CO clean-up began to surface. The reactor systems presented involved fuel cells and direct links to mobile applications could be seen. From 1997 and to the present, automotive corporations have begun to register entire solutions of reforming systems for automotive applications, especially DaimlerChrysler (previously Daimler-Benz), which holds numerous patents in the area. Some of the more recent patents are modifications of previous inventions. DaimlerChrysler has also systematically protected parts of a reforming system prior to patenting the entire system. However, several other companies have also chosen to re-invent a previous patent by adding small modifications to the original in order to extend the time span of protection. The protection of an idea exposes the research carried out, which may explain why some automotive developers have abstained from registering their ideas. However, the amount of patents held by DaimlerChrysler in the area of reforming makes it difficult for other developers to intrude on their patents. Some of the inventions patented by various companies contain similar solutions with a few simple alterations. In Table 6 descriptions of inventions in the area of reforming are presented. The inventions vary from detailed reactor solutions, catalyst make-up for reforming and clean-up, as well as entire systems for automotive reforming of methanol including the fuel cells in which the electric energy is generated. In the beginning of the 21st century, developers of fuel cells such as Ballard have started to protect entire reformer systems rather than the fuel cell only, which has been the company’s development profile. This suggests that the developers of fuel cells are moving towards turnkey solutions, which is very common in the chemical and the mobile communications industry.

156 157

147

The invention describes method and apparatus for supplying hydrogen from reformed methanol to the fuel cell The invention describes a catalytic system for the reforming of methanol focusing on the placement of the catalyst and the reaction conditions

Method of supplying fuel gas to a fuel cell

Reformer

US 6,045,933/Honda Giken Kogyo, Japan

EP 968,958mGK Insulators Ltd., Japan

Thermally integrated two-staged The invention describes a detailed apparatus for the reforming EP 977,293/General Motors Corporation, USA methanol reformer and method of methanol focusing on the problems with the endothermicity of operation

2000

2000

2000

135

This invention describes a method for operating a reformer with focus on the elimination of CO in the reformate

Process for operating a system for the water vapour reforming of methanol

US 6,048,4731 DaimlerChrysler AG, Germany

2000

133

US 6,074,770ElBB Fuel Cell Engines and BASF AG, both of Germany

2000

Process for operating a methanol The invention is a detailed description for operating and reforming apparatus running a methanol reforming system

134

Integrated fuel cell electric power The invention describes a compact system for the production US 6,063,s 1S/Ballard Power Systems Inc., USA generation system for submarine of electric power from a steam reformer applications

2000

132

The patent describes a method for the pre-aging of catalyst in order to prevent a loss of catalyst activity once the reforming has started

Method for treatment of methanol reforming catalyst

US 6,083,863/DBB Fuel Cell Engines, Germany

2000

131

The invention describes a methanol reforming system with a drop catching system for the protection of the catalyst

Multi-stage reforming reactor using a drop catching element

US 6,090,499/DBB Fuel Cell Engines, Germany

2000

Re$

Title

Description

Patent numberlassignee

Patents on methanol reformer systems

Year

Table 6

5

$ s Q

33

63

8

s CJ s

Q-

k

A

zis-.

6

?

Patent numberlassignee

US 5,989,503/ DaimlerChrysler AG, Germany

US 5,984,986/ DaimlerChrysler AG, Germany

US 5,955,395/MercedesBenz AG, Germany

US 5,942,346lUniversity of Chicago, USA

US 5,935,277/DaimlerBenz AG, Germany

US 5,928,614/DaimlerBenz AG, Germany

US 5,922,291/DaimlerBenz AG, Germany

US 5,904,9 13/DaimlerBenz AG, Germany

1999

1999

1999

1999

1999

1999

1999

1999

(contd.)

Year

Table 6

The invention describes a system for the reforming of methanol and process regulations for keeping a constant low co output

The invention is a description of a reactor solution for the reforming of methanol including heat transfer solutions

Description

The invention describes a reforming system with a filter to prevent material from the catalyst to enter the fuel cell without interrupting the operation

The invention describes a reactor designed to handle the endothermic steam reforming by internal heat exchange

The invention describes a reforming reactor for mobile applications and a system for filling the reactor with catalyst pellets

The invention describes a combined reactor system as it is designed for methanol, water and air. The system is specifically designed for use in automotive applications

The invention describes a reforming system in which special Process for obtaining a high hydrogen, low-carbon-monoxide attention has been paid to the removal of carbon monoxide gas

Reforming reactor system and particle filter usable therefore

Reforming reactor particularly for the water reforming of methanol

Reforming reactor particularly for the water vapour reforming of methanol

Methanol partial oxidation reformer

Preparation and use of Pt/zeolite The invention describes a process for preparing a Ptlzeolite material for the removal of carbon monoxide catalyst material for removing carbon monoxide

Process for operating a system for the water reforming of methanol

Process and apparatus for methanol reforming

Title

142

141

140

139

158

138

137

136

Ref

N N

c

143

160

145 161

163 164

The invention describes a reactor for selective oxidation, which introduces the oxidising gas at several points while controlling the exothermic reaction The system described in this invention is a reforming system for methanol to produce a CO-free H2-rich gas stream The apparatus enclosed in the invention is aimed at automotive applications with a focus on obtaining low carbon monoxide levels The invention describes an apparatus for two stage steam reforming of methanol to eliminate CO The invention is practically identical to US Patent 6,045,933 with the exception of the membrane, which was added in the later patent

Method and apparatus for selective oxidation of carbon monoxide

Reformer

Process and apparatus for methanol reforming

Two-stage reforming of methanol

Method for supplying a fuel gas to a fuel cell assembly

Two stage reformer of methanol The invention describes a reactor system, which operates in two separate stages to optimise conversion and selectivity

Fuel cell containing a reforming The invention describes a high temperature fuel cell with a reforming catalyst with improved resistance to deactivation catalyst

The invention describes a method for the production of a Copper containing catalyst for low temperature shift conversion copper, zinc, aluminum and potassium catalyst for low-temperature conversion of C 0 2 in hydrogen-rich gas

EP 967,174mGK Insulators Ltd., Japan

US 5,772,707/DaimlerBenz AG, Germany

US 5,753,194/Daimler Benz AG, Germany

US 5,714,276/Honda, Japan

US 5,672,629/DaimlerBenz AG, Germany

Controlled CO preferential US 5,637,41 YGeneral Motors Corporation, USA oxidation

Fuel cell system for transportation applications

US 5,874,05 l/DaimlerBenz AG, Germany

US 5,248,566/US Department of Energy, USA

US 5,246,79 UJohnson Matthey, UK

US 5,128,307/BASFAG, Germany

1999

1999

1998

1998

1998

1997

1997

1993

1993

1992

162 155

The invention describes a method for oxidising CO in the presence of H2 without consuming it The system describes a combined reactor system and fuel cell system for automotive applications

146

144

159

The invention describes a catalyst for the conversion of methanol to hydrogen

One step conversion of methanol to hydrogen and C02

US 5,904,880/Exxon Chemical patents Inc., USA

1999

w

h,

c

z

3

23

6' 3

2

?3!

k

*B

Q

's-

EP 409,5 17/Mitsui Toatsu Process for the decomposition Chemicals Inc., Japan of methanol

US 5,045,297/Du Pont Inc., USA

US 5,075,268/Agency of Industrial Science and Technology, Japan

US 4,913,842/Mitsubishi Gas Chemical Company, Japan

US 4,946,667/Engelhard Corporation, USA

US 4,978,513/Lonza Ltd., Catalyst for the oxidation of Switzerland carbon compounds

US 4,826,798/Du Pont Inc., USA

1991

1991

1991

1990

1990

1990

1989

C02 calcination of methanol dissociation catalysts

Method for steam reforming methanol to hydrogen

Steam reforming of methanol

Regeneration method for methanol reforming catalyst

Selective oxidation of carbon monoxide in a mixture

US 4,986,978/Soci:tC Process for reforming impure Anonyme pour L’Etude et methanol L’Exploitation des Procedes George Claude, France

1991

Title

Patent numberlassignee

(contd.)

Year

Table 6 ~~

~

The patent describes a method for improving the calcination of copper-containing catalysts for the decomposition of methanol to C 0 2

The invention describes a method for the production of Pd/Zr02 catalysts for the oxidation of carbon compounds, especially carbon monoxide

The invention describes an apparatus for steam reforming of methanol for usage in electric power producing systems

The invention describes a method for the steam reforming of methanol, in which organic pollutants are removed. The invention also presents various methods for the preparation of steam reforming catalysts

The invention describes a method for regenerating copper containing methanol-reforming catalysts using a mixture of oxygen in nitrogen

The invention proposes a method for the selective oxidation of CO in a mixture of organic compounds over platinum and palladium catalysts

A method for producing a high CO/H2 ratio is proposed in this invention using a chromium catalyst

The invention describes a method and apparatus for producing pure hydrogen from methanol containing impurities such as propanol and butanol

Description

~~~

171

170

169

168

167

166

165

Ref:

!2?

w

178

Catalytic hydrogen generator for The invention describes a partial oxidation reformer for hydrogen 150 use with methanol generation which later evolved to the HotSpotTMreactor 179

180 181

The patented process describes a method for the conversion of alcohols to CO and H2 in the absence of steam using metal carbides as catalysts The invention describes a method and catalyst for the dissociation of methanol for application in a combustion engine The invention describes a catalyst preparation method and usage for the decomposition of methanol to CO and H2

Process for conversion of alcohols to gaseous products

Process for reforming of alcohols

Alcohol dissociation process and a reactor therefore

Methanol dissociation using a chromium manganese catalyst

US 4,676,972Btandard Oil Company, USA

EP 217,532/Johnson Matthey, UK

US 4,632,774/Standard Oil Company, USA

EP 102,845/CONOCO Inc., USA

US 4,407,238/CONOCO Inc., USA

1987

1987

1986

1984

1983

The invention describes the production of copper-containing catalysts for the decomposition of methanol

177

176

The invention describes a method for the cracking of methanol to produce a hydrogen-rich gas using Group 9 and 10 catalysts. The invention presents results from several tested catalysts

175

Methanol reforming process and The invention describes a method for reforming methanol and apparatus for practicing it a process of circulation to increase the hydrogen content of the produced gas

US 4,780,300/Mitsubishi Process for reforming methanol Jukogyo Kabushiki, Japan

1988

The invention prescribes a method for regenerating copper-containing catalysts by oxidising the catalyst and reducing the catalyst

174

US 4,670,187/Societe Chimiques de la Grande, Belgium

Regeneration of methanol dissociation catalysts

US 4,855,267/Du Pont Inc., USA

1989

The invention describes a method for the production of Ni-containing catalyst for the decomposition of methanol to CO and H2

173

1987

EP 324,618/Mitsubishi Gas Catalyst composition for Chemical Company, Japan decomposition of methanol

1989

Process for the production of The invention describes an apparatus for the steam hydrogen by catalytic reforming reforming of methanol to produce a CO free hydrogen-rich of methanol with water vapor product gas

US 4,840,783finstitut Francais du Petrole, France

1989

' + Q

zR'

sis

Catalysis

126

7

Concluding Remarks

This review focuses on methanol as a hydrogen carrier mainly for automotive applications. The question whether methanol will be used as the new motor fuel in this millennium is not easy to answer and is beyond the scope of this review. From a strictly technological point of view it seems that methanol is the best solution among the fuel options available. Some of the most important features are its high hydrogen-to-carbon ratio, the absence of carbon-carbon bonds and the relative ease with which it is converted into hydrogen. Furthermore, fuel methanol is sulfur-free. Using fuels with a well developed infrastructure, such as gasoline or diesel, lowers the cost in the distribution chain compared to introducing a new fuel. However, the efficiency of an onboard reformer system will suffer from using fuels, which require complicated clean-up and pre-treatment steps. Methanol can be produced from several different feedstocks, both fossil and renewable. The resources of natural gas are still abundant and the chemical industry could easily adapt to meet a fast increase in methanol demand. Moreover, there is currently a production overcapacity of methanol in the world. We expect to see more of combined reformer systems in the future, i.e. systems combining steam reforming and partial oxidation. Using solely partial oxidation or steam reforming will not be an optimal solution in an automotive application. During the course of writing this review we have observed a need for the development of tailor-made catalysts for combined reforming. Material development is still an issue and it is a great challenge for the research community to meet the high requirements of the transport sector. Most likely, we will experience a great number of studies focusing on promotional effects of various compounds. Another important factor is the thermal stability of the catalyst, especially if it is set to operate in a combined system with great changes in operating temperatures. Another key issue that has to be addressed is the lifetime of the catalyst. An automotive system must probably function for at least 5 years or 80000 km. The customer will not tolerate frequent changes of the reformer system. In conclusion, reformer development is very important. In this area there is plenty of room for innovative reactor design. The material issues in conjunction with heat transfer considerations are the key points in the development process. The thermal management of the entire system along with the system design is crucial for the success of methanol reforming technology. A sophisticated system integration, which addresses the efficiency of both the reformer and the total system, is required to operate a fuel cell vehicle according to the stringent customer requirements.

8

Acknowledgements

The authors gratefully acknowledge financial support from the European Union (Contract No JOE3-CT97-0049), Volvo Technological Development

3: Catalytic Hydrogen Generationfrom Methanol

127

Corporation, the Swedish National Energy Administration and KTH - Royal Institute of Technology.

9 1 2 3 4 5 6 7

8 9 10 11

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

28

References J.N. Armor, Appl. Catal. A , 1999,176, 159. K.G. Knudsen, B.H. Cooper and H. Topsare, Appl. Catal. A, 1999,189,205. M.A. DeLuchi, Int. J. Hydrogen Energy, 1989,14,81. T. Klaiber, J. Power Sources, 1996,61, 61. L.J. Pettersson, PhD Thesis, Royal Institute of Technology, Stockholm, Sweden. ISRN KTHIKETIR-4-SE, 1994. J.C. Amphlett, S. Leclerc, R.F. Mann, B.A. Peppley and P.R. Roberge, Proc. 33rd Intersoc. Energy Conv. Eng. Conf., 1998, p. 269I1. B. Lindstrom and L.J. Pettersson, Proc. XIII Int. Symp. on Alcohol Fuels, Stockholm, Sweden, July 3-7, Vol. 11, Chap. 24,2000. E. Fiedler, G. Grossmann, B. Kersebohm, G. Weiss and C. Witte, in Ullmann’s Encyclopedia of Industrial Chemistry, ed. B. Elvers, S. Hawkins and G. Schulz, vol. A16, VCH Verlagsgesellschaft,Weinheim, Germany, 1990, p. 465. K.C. Waugh, Cutal. Today, 1992,15, 51. G.C. Chinchen, K. Mansfield and M.S. Spencer, CHEMTECH, 1990,20,692. W.R. Brown, D.P. Drown, and F.S. Frenduto, Commercial-Scale Demonstration of the Liquid Phase Methanol (LPMEOHTM) Process, Public Design Report, United States Department of Energy, DE-FC22-92PC90543,2000. W.-H. Cheng, C.-Y. Shiau, T.H. Liu, H.L. Tung, J.-F. Lu and C.C. Hsu, Appl. Catal. A , 1998,170,215. I.A. Fisher and A.T. Bell, J. Catal., 1999,184, 357. W.-H. Cheng, Appl. Catal. B, 1995,7, 127. W.-H. Cheng, Appl. Catal. A, 1995, 130, 13. W.-H. Cheng, Mater. Chem. Phys., 1995,41, 36. S . Poulston, E. Rowbotham, P. Stone, P. Parlett and M. Bowker, Catal. Lett., 1998, 52, 63. Y. Sakata, K. Domen, K. Maruya and T. Onishi, Appl. Surf Sci., 1989,35, 363. G. Hakvoort and X. Xiaoding, J. Therm. Anal., 1991,37, 1269. R.O. Idem and N.N. Bakhshi, Ind. Eng. Chem. Res., 1994,33,2047. D.B. Clarke, D.-K. Lee, M.J. Sandoval and A.T. Bell, J. Cutal., 1994, 150, 81. M. Clement, Y. Zhang, D.S. Brands, E.K. Poels and A. Bliek, Stud. Surf Sci. Catal., 2000,130,2123. Y. Matsumura, K. Kuraoka, T. Yazawa and M. Haruta, Catal. Today, 1998,45, 191. Y. Matsumura, N. Tode, T. Yazawa and M. Haruta, J. Mol. Catal. A , 1995, 99, 183. Y. Matsumura, K. Kagawa, Y. Usami, M. Kawazoe, H. Sakurai and M. Haruta, Chem. Commun., 1997,657. S . Imamura, T. Higashihara, Y. Saito, H. Aritani, H. Kanai, Y. Matsumura and N. Tsuda, Cutal. Today, 1999,50,369. Y. Usami, K. Kagawa, M. Kawazoe, Y. Matsumura, H. Sakurai and M. Haruta, Appl. Catal. A , 1998,171, 123. W.-J. Shen and Y. Matsumura, J. Mol. Catal. A , 2000,153, 165.

128 29 30

31 32 33

34 35 36 37 38

39 40 41 42 43

44 45 46 47 48 49 50 51 52 53 54 55 56 57

58 59

Catalysis Y. Usami, K. Kagawa, M. Kawazoe, Y. Matsumura, H. Sakurai and M. Haruta, Stud. SurJ Sci. Catal., 1998,118, 83. Y. Matsumura, M. Okumura, Y. Usami, K. Kagawa, H. Yamashita, M. Anpo and M. Haruta, Catal. Lett., 1997,44, 189. X. Montagne, R. Boulet, E. Freund and J.C. Lavalley, Stud. S u r - Sci. Catal., 1989,48,695. T. Shishido, H. Sameshima, T. Hayakawa, S. Hamakawa, E. Tanabe, K. Ito and K. Takehira, Stud. SurJ Sci. Catal., 2000,130,2117. Y. Matsumura, Y. Ichihashi, Y. Morisawa, M. Okumura and M. Haruta, Stud. SurJ Sci. Catal., 2000,130,2315. L. Pettersson and K. Sjostrom, Combust. Sci. Technol., 1990,71, 129. L. Pettersson and K. Sjostrom, Combust. Sci. Technol., 1991,80,265. R.O. Idem and N.N. Bakhshi, Ind. Eng. Chem. Res., 1994,33,2056. N. Aas, T.J. Pringle and M. Bowker, J. Chem. SOC.,Faraday Trans., 1994, 90, 1015. D.T. Wickham, B.W. Logsdon, S.W. Cowley, and C.D. Butler, J. Catal., 1991, 128, 198. H. Imai, T. Tagawa and K. Nakamura, Appl. Catal., 1990,62,348. J.P. Breen and J.R.H. Ross, Catal. Today, 1999,51, 521. B.A. Peppley, J.C. Amphlett, L.M. Kearns and R.F. Mann, Appl. Catal. A , 1999, 179, 21. B.A. Peppley, J.C. Amphlett, L.M. Kearns and R.F. Mann, Appl. Catal. A, 1999, 179, 31. S.P. Asprey, B.W. Wojciechowski and B.A. Peppley, Appl. Catal. A , 1999, 179, 51. C.J. Jiang, D.L. Trimm, M.S. Wainwright and N.W. Cant, Appl. Catal. A , 1993, 97, 145. E. Santacesaria and S . Carra, Appl. Catal., 1983,5,345. K. Takahashi, N. Takezawa and H. Kobayashi, Appl. Catal., 1982,2,363. H. Kobayashi, N. Takezawa and C. Minochi, J. Catal., 1981,69,487. G.-C. Shen, S.4. Fujita, S. Matsumoto and N. Takezawa, J. Mol. Catal. A , 1997, 124, 123. J.P. Breen, F.C. Meunier and J.R.H. Ross, Chem. Commun., 1999,2247. K. Takahashi, H. Kobayashi and N. Takezawa, Chem. Lett., 1985,6,759. C.J. Jiang, D.L. Trimm, M.S. Wainwright and N.W. Cant, Appl. Catal. A , 1993, 93,245. J.C. Amphlett, K.A.M. Creber, J.M. Davis, R.F. Mann, B.A. Peppley and D.M. Stokes, Int. J. Hydrogen Energy, 1994,19, 131. K. Miyao, H. Onodera and N. Takezawa, React. Kinet. Catal. Lett., 1994,53,379. N. Takezawa, Kobayashi, A. Hirose, M. Shimokawabe and K. Takahashi, Appl. Catal., 1982,4, 127. N. Iwasa, S. Masuda, N. Ogawa and N. Takezawa, Appl. Catal. A , 1995, 125, 145. N. Takezawa and N. Iwasa, Catal. Today, 1997,36,45. N. Iwasa, S. Kudo, H. Takahashi, S. Masuda and N. Takezawa, Catal. Lett., 1993,19, 211. J. C. Amphlett, R.F. Mann and B.A. Peppley, Int. J. Hydrogen Energy, 1996,21, 673. R. Diimpelmann, PhD. Thesis, Eidgenossischen Technischen Hochschule, Zurich, 1992 (in German).

3: Catalytic Hydrogen Generationfrom Methanol

60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93

129

B. Lindstrom and L.J. Pettersson, Proc. 2000 Fuel Cell Seminar, Portland, Oregon, Oct. 30-Nov. 2, 2000, p. 325. B. Lindstrom and L.J. Pettersson, Int. J. Hydrogen Energy, 2001,26,923. B. Lindstrom and L.J. Pettersson, Catal. Lett., 2001,74,27. M.L. Cubeiro and J.L.G. Fierro, Appl. Catal. A , 1998,168, 307. M.L. Cubeiro and J.L.G. Fierro, J. Catal., 1998,179, 150. S . Velu, K. Suzuki, and T. Osaki, Catal. Lett., 1999,62, 159. L. Alejo, R. Lago, M.A. Peiia and J.L.G. Fierro, Appl. Catal. A , 1997,162,281. T.-J. Huang and S.-L. Chren, Appl. Catal., 1988,40,43. M.A. Peiia, J.P. Gomez and J.L.G. Fierro, Appl. Catal. A , 1996,144,7. J.L.G. Fierro, Stud Surf Sci. Catal., 2000,130, 177. T.L. Reitz, S. Ahmed, M. Krumpelt, R. Kumar and H.H. Kung, Stud. Surf: Sci. Catal., 2000,130, 3645. J. Agrell, K. Hasselbo, K. Jansson, S.G. Jaris and M. Boutonnet, Appl. Catal. A., 2001,211,239. T.-J. Huang and S.-W. Wang, Appl. Catal., 1986,24,287. J. Agrell, G. Germani, S.G. JarHs and M. Boutonnet, to bepublished. L. Pettersson and K. Sjostrom, Proc. VII Int. Symp. on Alcohol Fuels, Paris, Oct. 20-23, 1986, Complementary Publication, 77. L. Pettersson and K. Sjostrom, Int. J. Hydrogen Energy, 1991,16, 671. H.-B. Zhang, W. Wang, Z.-T. Xiong and G.-D. Lin, Stud. Sut$ Sci. Catal., 2000, 130,1547. H.Y.H. Chan, C.T. Williams, M.J. Weaver, C.G. Takoudis, J. Catal., 1998, 174, 191. M.P. Zum Mallen and L.D. Schmidt, J. Catal., 1996,161,230. R.W. McCabe and D.F. McCready, J. Phys. Chem., 1986,90,1428. M. Haruta, A. Ueda, S. Tsubota and R.M. Torres Sanchez, Catal. Today, 1996, 29,443. S.G. Christoskova, M. Stoyanova, M. Georgieva and D. Mehandzhiev, Appl. Catal. A , 1998, 173,95. S.G. Christoskova, M. Stoyanova, N. Danova and 0. Argirov, Appl. Catal. A, 1998,173,101. I.E. Wachs and R.J. Madix, J. Catal., 1978,53,208. J. Agrell, M. Boutonnet, I. Melian-Cabrera, R. Navarro and J.L.G. Fierro, manuscript under preparation. G.J.J. Bartley and R.Burch, Appl. Catal., 1988,43, 141. W.R.A.M. Robinson and J.C. Mol, Appl. Catal., 1988,44, 165. R. Burch, R.J. Chappell and S.E. Golunski, J. Chem. SOC.Faraday Trans. I, 1989,85,3569. R. Burch, S.E. Golunski and M.S. Spencer, J. Chem. SOC.Faraday Trans., 1990, 86, 2683. J. Agrell, K. Hasselbo, S. Jaris and M. Boutonnet, Stud. Surf:Sci. Catal., 2000, 130,1073. S . Velu, K. Suzuki and T. Osaki, Chem. Commun., 1999,23,2341. S . Velu, K. Suzuki, M. Okazaki, M.P. Kapoor, T. Osaki and F. Ohashi, J. Catal., 2000,194,373. E. Newson, P. Mizsey, T. Truong and P. Hottinger, Stud. Surf:Sci. Catal., 2000, 130,695. S . Velu, K. Suzuki and T. Osaki, Recent Reports of the 12th Int. Congress on Catalysis, Granada, Spain, July 9- 14,2000.

130

Catalysis

94

T.L. Reitz, P.L. Lee, K.F. Czaplewski, J.C. Lang, K.E. Popp and H.H. Kung, J. Catal., 2001,199, 193. T.L. Reitz, S. Ahmed, M. Krumpelt, R. Kumar and H.H. Kung, J. Mol. Catal. A , 2000,162,275. J.W. Jenkins and E. Shutt, Platinum Metals Rev., 1989,33, 118. S. Golunski, Platinum Metals Rev., 1998,42,2. S. Ahmed and R. Kumar, I995 Automotive Technology Development Contractors’ Coordination Meeting, Dearborn, MI, Oct. 23-27, 1995. L.J. Pettersson and R. Westerholm, Int. J. Hydrogen Energy, 2000,26,243. N. Edwards, S.R. Ellis, J.C. Frost, S.E. Golunski, A.N.J. van Keulen, N.G. Lindewald and J.G. Reinkingh, J. Power Sources, 1998,71, 123. I. Carpenter, N. Edwards, S. Ellis, J. Frost, S. Golunski, N. van Keulen, M. Petch, J. Pignon and J. Reinkingh, SAE Paper 1999-01-1320, 1999, 173. S. Ahmed, R. Doshi, S.H.D. Lee, R. Kumar and M. Krumpelt, Proc. 32nd Intersoc. Energy Conv. Eng. Conf., 1997, p. 843. R. Kumar, S. Ahmed and M. Krumpelt, Electric Hybrid Vehicle Technol., 1996, 96, 123. Y. Li, Q. Fu and M. Flytzani-Stephanopoulos, Appl. Catal. B, 2000,27, 179. A.A.G. Lima, M. Nele, E.L. Moreno and H.M.C. Andrade, Appl. Catal. A , 1998, 171, 31. C. Rhodes, G.J. Hutchings and A.M. Ward, Catal. Today, 1995,23,43. D. Andreeva, V. Idakiev, T. Tabakova, A. Andreev and R. Giovanoli, Appl. Catal. A, 1996,134,275. F. Boccuzzi, A. Chiorino, M. Manzoli, D. Andreeva and T. Tabakova ,J. Catal., 1999,188,176. M.V. Twigg (Ed.), Catalyst Handbook, 2nd ed., Wolfe Publishing Co., London, 1989. K. Sekizawa, S. Yano, K. Eguchi and H. Arai, Appl. Catal. A , 1998,169,291. R.M. Torres Sanchez, A. Ueda, K. Tanaka and M. Haruta, J. Catal., 1997,168, 125, H. Igarashi, H. Uchida, M. Suzuki, Y. Sasaki and M. Watanabe, Appl. Catal. A , 1997,159,159. C.D. Dudfield, R. Chen and P.L. Adcock, J. Power Sources, 2000,85,237. S.H. Oh and R.M. Sinkevitch, J. Catal., 1993, 142,254. R. Barrett, C.J. Read and R.S. Weber, Proc. 1998 Fuel Cell Seminar, Nov. 16-19, Palm Springs, CA, 1998, p. 590. G. Colsman, Industrial Processing Optimization of Fuel Gas Production for Fuel Cells, Technical Report No JUEL-3 127, Forschungszentrum Jiilich (IEV), Jiilich, Germany, 1995. B.S. Baker, J. Huebler, H.R. Linden and J. Meek, US Patent 3,615,164, 1971. H. Oshiro, K. Nagaya, K. Mitani and S. Tsuchiyama, Proc. 1996 Fuel Cell Seminar, Orlando, FL, Nov. 17-20, 1996, p. 319. J.E. Cuzens, J. Mauzey and R.R. Woods, Proc. 1998 Fuel Cell Seminar, Nov. 16-19, Palm Springs, CA, 1998, p. 234. R.E. Buxbaum, Sep. Sci. Technol., 1999,34,2113. K. Ledjeff-Hey, V. Formanski and J. Roes, Proc. 1998 Fuel Cell Seminar, Nov. 16-19, Palm Springs, CA, 1998, p. 345. J. Han, I. Kim and K.-S. Choi, J. Power Sources, 2000,86,223. Y.-M. Lin, G.-L. Lee, and M.-H. Rei, Catal. Today, 1998,44, 343. Y.-M. Lin and M.-H. Rei, Int. J. Hydrogen Energy, 2000,25,211.

95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124

3: Catalytic Hydrogen Generationfrom Methanol

131

125 J. Shu, B.P.A. Grandjean, A. van Neste and S . Kaliaguine, Can. J. Chern. Eng., 1991,61,1036. 126 S.G. Chalk, J. Milliken, J.F. Miller and S.R. Venkateswaran, J. Power Sources, 1998,71,26. 127 BASF AG, Press Release, P 435, Sept. 12,2000. 128 F. Kalhammer, P.R. Prokopius, V.P. Roan and G.E. Voecks, Status and Prospects of Fuel Cells as Automobile Engines. A Report of the Fuel Cell Technical Advisory Panel, State of California Air Resources Board, Sacramento, CLA, 1998. 129 R. Kumar, Fuel Cellsfor Transportation, TOPTEC, 1995. 130 S . Ahmed, M. Krumpelt, R. Kumar, S.H.D. Lee, J.D. Carter, R. Wilkenhoener and C . Marshall, Proc. 1998 Fuel Cell Seminar, Palm Springs, CA, Nov. 16-19, 1998, p. 242. 131 R. Autenrieth and D. Heil, US Patent 6,090,499, 2000. 132 R. Autenrieth and 0. Krumberger, US Patent 6,083,863,2000. 133 R. Autenrieth, A. Christen, M. Schuessler and 0. Krumberger, US Patent 6,074,770,2000. 134 D.G. Epp and R.M. Baumert, US Patent 6,063,515,2000. 135 M. Denda, C. Duelk, R. Gerberich and N. Wiesheu, US Patent 6,048,473,2000. 136 N. Wiesheu, D. zur Megede and R. Autenrieth, US Patent 5,989,503, 1999. 137 N. Wiesheu, C. Duelk, R. Autenrieth and U. Benz, US Patent 5,984,986, 1999. 138 R. Andorf, W. Maunz, C. Plog and T. Stengel, US Patent 5,955,395, 1999. 139 R. Autenrieth, D. Heil, W. Weger and U. Benz, US Patent 5,935,277, 1999. 140 R. Autenrieth and D. Heil, US Patent 5,928,614, 1999. 141 J. Hanfling, D. Heil, R. Autenrieth and N. Wiesheu, US Patent 5,922,291, 1999. 142 G. B o b , T. Staneff and J. Steinwandel, US Patent 5,904,913, 1999. 143 D. Heil, B. Strobel, N. Wiesheu, D. zur Megede and U. Benz, US Patent 5,874,051, 1999. 144 N. Wiesheu, D. zur Megede and R. Autenrieth, US Patent 5,772,707, 1998. 145 D. Heil and U. Benz, US Patent 5,753,194, 1998. 146 D. Heil and U. Benz, US Patent 5,672,629, 1997. 147 G.W. Skala, D.J. Hart-Predmore, W.H. Pettit and R.L. Borup, Eur. Patent 977,293,2000. 148 I. Ivarsen, Eur. Patent 149,162, 1984. 149 D.F. Szydlowski, Eur. Patent 275,549, 1990. 150 J.W. Jenkins, Eur. Patent 217,532, 1987. 151 T. Eguchi, US Patent 5,221,524, 1991. 152 A. Konig, Eur. Patent 201,670,1986. 153 P.J. de Wild and M.J.F.M. Verhaak, Catal. Today, 2000,60, 3. 154 P.G. Gray and M.I. Petch, Platinum Metals Rev., 2000,44, 108. 155 R. Kumar, S. Ahmed, M. Krumpelt and K.M. Myles, US Patent 5,248,566, 1993. 156 T. Okamoto, US Patent 6,045,933,2000. 157 F. Abe, N. Noda and J. Suzuki, Eur. Patent 968,958,2000. 158 S. Ahmed, R. Kumar and M. Krumpelt, US Patent 5,942,346, 1999. 159 H. Sun, US Patent 5,904,880, 1999. 160 F. Abe, N. Noda and J. Suzuki, Eur. Patent 967,174, 1999. 161 T. Okamoto, US Patent 5,714,276, 1998. 162 M.A. Meltser and M.M. Hoch, US Patent 5,637,415, 1997. 163 J.M. Fisher and P.S. Bennett, US Patent 5,246,791, 1993. 164 H. Wanjek, C. Witte, M.J. Sprague, M. Irgang, W. Hoelderich and W.D. Mross, US Patent 5,128,307, 1992.

132

Catalysis

165 166 167 168 169

R. Dupont and G. Simonet, US Patent 4,986,978,1991. T. Kiyoura, T. Jimbo, Y. Kogure and K. Kanaya, Eur. Patent 409,517,1991. C. Bonifaz and D.R. Corbin, US Patent 5,045,297, 1991. M. Kurashige, M. Takiguchi and N. Matsuo, US Patent 5,075,268, 1991. M. Yoneoka, T. Furuta, Y. Yamamoto, S. Nozaki, K. Tanaka, K. Watabe and T. Ikarashi, US Patent 4,913,842, 1990. B.S. Beshty, US Patent 4,946,667, 1990. A. Baiker and D. Gasser, US Patent 4,978,513, 1990. W.-H. Cheng, US Patent 4,826,798, 1989. D.V. Quang, M. Davidson, D. Durand and P. Courty, US Patent 4,840,783, 1989. N. Isogai, M. Takagawa, K. Watabe, M. Yoneoka and K. Yamagishi, Eur. Patent 324,618, 1989. W.-H. Cheng, US Patent 4,855,267, 1989. N. Yokoyama, T. Imai, H. Fujita, M. Murakami, Y. Miyairi and M. Tamai, US Patent 4,780,300, 1988. J. Schurmans, H. Neukermans, J. Quibel and R. Dupont, US Patent 4,670,187, 1987. L.J.Velenyi and T.J. Mazanec, US Patent 4,676,972, 1987. J.R. Fox, F.A. Pesa and B.S. Curatolo, US Patent 4,632,774, 1986. H. Yoon, Eur. Patent 102,845, 1984. H. Yoon, US Patent 4,407,238, 1983.

170 171 172 173 174 175 176 177 178 179 180 181

4 Reforming of CH4 by COa, O2 and/or H20 BY TOMOYUKI INUl

1

Introduction

1.1 Recent Research. - Steam reforming of methane has been established and industrialized for some time, and a fairly small amount of fundamental research work on the catalysis of this reaction has been reported in recent The main reason for this trend might be that catalytic steam reforming of methane is highly endothermic, as shown in Eq. (3), which is a summation of Eqs. (1) and (2), and the rate-determining step is the heat transfer from the reactor wall to the catalyst bed. The relevant reactions are: CH4 + H20 + 3H2 + CO,

AH" = +206 kJ mol-

(1)

CO + H20 + H2 + C02,

AH"= -41 kJmol-'

(2)

CH4 + 2H20 + 4H2 + C02,

AH"=+165 kJmol-'

(3)

The effect of the development of improved catalysts on the total efficiency is relatively minor. Recent interest in steam reforming of CH4 seems to be in the membrane r e a c t ~ r , ~in- which ~ hydrogen can be selectively removed; the reaction equilibrium shifts to the products, resulting in higher conversion of CH4 even at a lower temperature range than that in usual steam reforming processes. The other advantage of the membrane reactor system is that COX-freehydrogen can be obtained directly. This is favourable for use in fuel cells, which is now attracting attention as the ideal energy system in the near future. However, in general, since the reaction rate in the membrane reactor is strictly limited by the magnitude of the permeation rate through the membrane layer, a sufficiently high reaction rate to be industrially practicable cannot be expected. On the other hand, a recent increase in papers concerning C02 reforming of methane, Eq. (4), and partial oxidation of methane, Eq. (5), to synthesize syngas has occurred in the past decade. More than 80% of these papers have appeared in recent years. CH4 + C02 + 2H2 + 2C0,

AH" = +247 kJ mol-I

(4)

CH4 + 11202 + 2H2 + CO,

AH" = -38 kJ mol-'

(5)

C 0 2 reforming of CH4 itself is not effective for the purpose of direct mitigation Catalysis, Volume 16 0The Royal Society of Chemistry, 2002

133

134

Catalysis

of C 0 2 accumulation, because C 0 2 needs an equivalent mole of methane for the reaction. Nevertheless, this subject has been the focus of a great deal of attention in international congresses on CO2 mitigation and/or utilization.6-16 This subject has intriguing new aspects, as enumerated in Section 1.2. Reflecting this situation, several elaborate reviews on this subject have been provided by in 1993, Rostrup-Nielsen'* in 1994, Ross et a l l 9 in 1996, Halmann and Steinberg20in 1999 and Bradford and Vannice2' in 1999, who summarized as many as 190 papers. The most recent 16 review papers were published in 1997. The focuses of this review are the fundamental aspects such as activation of CH4 and C02, carbon deposition, kinetics, and reaction mechanisms. Catalyst reaction engineering of autothermal reforming and coreforming such as C02 + 0 2 and C02--H20-02 are not discussed. From the viewpoint of rapid reaction rates reforming, which is applicable to industrial practice, reaction engineering is indispensable, and most recently the number of papers on this subject has rapidly i n ~ r e a s e d . Therefore, ~ ~ - ~ ~ in this review, papers reported at the major international conferences on C 0 2 conversion after 1998 to the present will be also included, although some of them are still at the status of 'in press'. 1.2 Significance of the Subject and Disadvantage to be Overcome. - In C02 reforming of CH4, expensive hydrogen is not needed to reduce C02. Syngas can be synthesized by dry gas conditions, which allows simple reaction processes and saves energy for use in steam reforming of CH4. Syngas, having a lower H2/C0 molar ratio, which is preferable for the use in some FT synthesis using iron-based catalysts, can be directly synthesized from methane and C 0 2 or directly from natural gas containing a considerable concentration of C02, sometimes more than 25 mol%, which is produced at some natural gas fields such as the Gulf of Thailand, Indonesia, and Australia. Synthesis of valuable compounds from two inexpensive resources, C02 and CH4, can be carried out. While conventional processes of steam reforming of CH4 lead to up to 2 mol% of unreacted methane in the product, in a commercial process of CO2 reforming of CH4 unconverted CH4 could be as low as 0.05 molYk7 H20-CO2 co-reforming and partial oxidation reforming can be carried out using a highly active catalyst for C02 reforming. An advanced reforming process, which involves in situ heat supply by catalytic combustion, can be developed. Direct control of appropriate ratios of H2/CO/C02 to couple with downstream reactions could be achieved directly in a one-pass operation. On the other hand, more rapid coke formation than in steam reforming can be anticipated by the use of conventional reforming catalysts, such as supported Ni-based materials. Coke formed by Eqs. (6) and (7) cannot be removed by the reaction (8), which occurs in steam-reforming with the use of excess steam. 2 c o + c + c02 (6) CH4 + C + 2H2 (7) C+H20-+CO+H2

(8)

4: Reforming of CH4 by C02, 0 2 andlor H20

135

1.3 Indispensable Requisites of Catalytic Technologies for Moderating COz Problems and/or Producing Syngas. - Before reviewing the subject of C02 reforming of methane or natural gas, general considerations on C 0 2 conversion must be examined and then rational criteria to evaluate the work can be developed. A high conversion rate is the most important criterion for treating large quantities of C02 and natural gas. High selectivity to targeted compounds is also necessary from the viewpoint of production economy. Since coke deposition and the accompanying deactivation of the catalyst is a general problem, minimizing coke formation during the catalytic reaction is also an indispensable element in maintaining high reaction rates. Low cost for catalyst preparation and heat conduction during the reaction must not be overlooked. Key factors to promote rapid rates include suitable pore structures to allow faster mass transfer for the feed and products. In order to compensate for the highly endothermic reforming reaction, the combination of an exothermic reaction on the same catalyst surface would maintain the proper catalyst temperature and the reaction rates. Other factors include proper utilization of the hydrogen spillover effect, which might keep the catalyst surface in a very reduced state, preventing coke formation and maximizing the space-time yield of syngas. A realistic target is 4 x lo3 moll- h- at a space velocity of at least 200 x 103 h-1. 2

Comparison of the Intrinsic Activity of Ni and Noble Metals Belonging to Groups 8-10

A comparison of the specific activities among platinum-group metals such as Pt, Ir, Rh, Pd and Ru, and ion-group metals, Ni, Co, and Fe, has been made Rostrup-Nielsen for C02 reforming as single metal-component and Hansen28compared the catalytic activity among Ni, Ru, Rh, Pd, Ir, and Pt supported on MgO, and concluded that replacing steam by C02 has no significant impact on the reforming mechanism. The order of activity for these metals was similar to each other as follows: Ru, Rh> > Ir> Ni, Pt % Pd for steam reforming and Ru, Rh 2 Ni> > Ir> Pt > Pd for C 0 2 reforming. Rates of carbon formation in the CH4/H2= 95/5 feed were compared by TGA measurements, and a striking contrast among catalysts was observed as shown in Figure 1.28 The order of deposited coke was; Ni> > > Rh> Ir = Ru> Pt % Pd at 500 "C and Ni> Pd = Rh> Ir> Pt> > Ru at 650 "C. As a result, they concluded that Rh and Ru show high selectivity for carbonfree operation, but because of the cost of these metals, they cannot be expected to impact on the total reforming catalyst market. The results were somewhat different for different supports, although the essential conclusion that Rh, Ru, and Ni are highly active catalyst metals for this reaction is the same.29-33 Detailed studies on Rh40-47 and Ir43348catalysts have been made, using different kinds of supports such as, A1203,Ti02, Si02 and MgO. Erdohelyi et aZ.40 showed that the order of catalytic activity was the same as described

Catalysis

136 Pd

/ :I’ ,I

0

I

5

/

10

Ni

---- 650°C - 500°C

-_ rt

__----

15

20

25

Time on stream (min) Figure 1

Rates of carbonformation on various single component catalyst supported on MgU. CH41H2= 9515, TGA measurements (Reproduced by permission from J. Catal., 1993,144,38)

above. Other than these supports, yttria-stabilized zirconia (YSZ),41942 Z~--Si02,4~ and zeolites (Nay and NaZSM-5)44were also tested. Ru* and Pt53-61catalysts were also studied extensively because of their high activity and minimal coke formation in the C02 reforming. Ru catalysts have been studied on y-A1203,13350-51 1pA1203,~~ Ti02,49 Si02,51 carbon graphite,50 La203,52Y20352active carbon,13 Zr02,52 and zeolite Y.37 Pt catalysts have been studied on A1203,53359y-A1203,54958 Ti02 , Zr02,54-60cr20357and Si02.57Bradford and Vannice et aZ.49showed that the order of turnover frequencies of Ru catalyst supported on different supports for C 0 2 reforming of CH4 in the absence of heat and mass transfer effects was T i O p A1203>> C. On the other hand, the most widely studied support for Pt catalyst has been Zr02. One reason is the lower rate of coke formation on Pt/Zr02 than on other s ~ p p o r t sAmong . ~ ~these ~ ~studies, ~ ~ O’Connor ~ ~ ~ ~and ~ Ross59 ~ ~ ~observed that by the addition of a low concentration of 0 2 , as shown in Figure 2, coke deposition was avoided completely. Deposited coke might be smoothly combusted on the catalyst surface by the dissociated oxygen due to oxygen spillover from the gas phase to the active sites.62 A limited number of Co catalysts have been studied using carbon and Si0263 and a l ~ m i n aas~ the * ~supports. ~ Carbon deposition is much more promoted on Co/A1203 catalyst than on Ni/A1203,64The presence of MgO markedly reduced the carbon deposition on the surface of the Co catalyst.63The role of the MgO may be ascribed to the formation of strongly adsorbed C 0 2 species, which can easily react with the deposited carbon, preventing catalyst deactivat i ~ nHowever, . ~ ~ this prevention of coke deposition is still not complete. 3937749-52

54957

4: Reforming of CH4 by CO,,0 2 andlor H 2 0

137

70 60 -

GI

CO, reforming

tEo '"I Y

W

40

s5

20 30

1

-101

0

3 vol% 0, added

'

'

'

'

'

'

'

'

'

'

'

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

Time onstream (h) Figure 2

Weight changes recorded during the reaction at 800 "Cunder CO, reforming conditions and with 3 VOPA0 2 added to the feed (from Ref 59) (Reproduced by permission from Catal. Today, 1998,46,203)

CO2 reforming of methane on MoS2 and WS2 catalysts showed much lower reaction rates than that of a Ni catalyst, although the disulfides did not lead to carbon deposition during the reforming reaction. The observed reaction kinetics suggested that the surface of the disulfide catalysts was abundant in adsorbed C02, resulting in suppression of decomposition of CH4, while the surface of Ni catalyst was abundant in adsorbed CH4.66967

3

Performance of Ni Catalyst Supported on Different Kinds of Support and the Effect of Additives

From a practical viewpoint, relatively inexpensive Ni has been the focus of efforts to improve catalysts for COz reforming of methane. Rh and Ru have received less attention. One of the most popular supports used for Ni catalyst is alumina.34~39~68-76 In order to avoid Ni particle growth, NiAl spinel was tested as the catalyst. Stability increased at the expense of high activity. The effect of additives to Ni catalyst supported on a-A1203upon activity has also been tested;69 however, no enhancement was observed as shown by the following order of activity: Ni, Ni-Co, Ni-Ce> Ni-Cu> > Ni-Fe. investigated the effect of noble metal (Ru or Pd) addition Crisafulli et to supported Ni catalysts for the reaction of C02 reforming of methane. The strong improvement in the activity (7-fold) and stability (deactivation rate decreased by a factor of 1/7) was observed in the case of 2NiM.6Ru, which was prepared by the use of Ru nitrate [Ru (NO)(N03)3]. When Ru chloride (RuC13) was used instead of the nitrate, the reaction was suppressed. Since H2 uptake increased markedly (3.5 times) by the addition of Ru, they concluded that Ni-Ru clusters, with the surface mainly covered by Ni, were formed, i.e.

Catalysis

138

300

400

500

600

700

Temperature (K) Figure 3

TPR profiles of mono- and bimetallic Ni-Ru catalysts (Reproduced by permission from Catal. Lett., 1999,59,21)

the metallic dispersion of Ni increased. However, such a direct relation between the apparent increase of H2 uptake and dispersion of metals is quite doubtful, because a nHhM (hydrogen atom uptakenlmetal atom) ratio much higher than 1 is often obtained in the case of supported noble metals.78 Furthermore, if the surface of the bimetallic clusters is mainly covered by Ni, significant carbon deposition during the reaction would be anticipated. Judging from the average deactivation rate due to coke formation [as expressed by wt% increase per hour (Ohoh) for 2Ni and 2Ni0-0.6Ru catalysts, these had rates of 2.8 and 0.4, respectively] and the comparison of TPR profiles for Ni, Ru and the Ni-Ru catalysts (Figure 3), the main reason of the low coke deposition and high activity is the hydrogen spillover from Ru79-87to Ni. Coke precursors are hydrogenated by the spillover of hydrogen, resulting in a lower deactivation rate and a higher reaction rate. Carbon deposition during C 0 2 reforming was markedly suppressed by the addition of alkaline oxide such as Na20 and K20, and alkali earth-metal oxides such as MgO, and CaO, but as shown in Figure 4 the activity of Ni decreased in the order of None % Na20> MgO> > CaO> K20.73 The effect of Ni precursor for a Nily-A1203 catalyst was tested using Ni(N03)2, NiC12, and Ni acetylacetonate (Ni(C5H702)2).However, the latter two precursors gave considerably lower activities than Ni (N03)2.74 Supports other than alumina used for Ni catalysts were Si02,15,83perovskite,84 zeolite 5A,36 high silica ZSM-5 zeolite85and AlP04.86In the case of ZSM-5 support,85coke deposition was remarkably diminished with the addition of K and Ca, and stability was confirmed up to 140 h, although the spacevelocity of the experiment was fairly low, 4.4 x lo4 - 6 x lo4 h-l. Choudhary et aZ.86investigated partial oxidation of methane to syngas with or without simultaneous C 0 2 and steam reforming reaction over Ni/A1PO4. As discussed

139

4: Reforming of CH4 by C02, O2 andlor H 2 0 -7 n

Ei

Na,O

7 I

8

v

-8

No metal oxid

n

8

E

-9

E

0

. c) I

Y

g -10

v

a

I

-11

0.9

Figure 4

1

1.1

1.2

1.3~10-3

Reaction rate of CO2-reforpning at 973 K over the NiIAl203 catalysts with and without basic metal oxide. (0) NiIA1203 catalyst, (a)NiIA1203 catalysts with I0 wPAbasic metal oxides (from Ref: 73) (Reproduced by permission from Appl. Catal. A: Gen., 1996,144, 11 1)

later, this kind of reaction would be one of the promising directions for practical operation. In general, besides ordinary effects of the support properties, such as surface area, porosity, and pore diameter, in the C 0 2 reforming of CH4, acidity and basicity of the support strongly affect coke deposition during the reforming reaction. Coke deposition is markedly decreased by the addition of MgO or La203. The reason is considered to be that acidic C 0 2 strongly adsorbs on these basic materials and forms carbonates, which may decompose at high temperature to produce CO or provide oxygen species. The latter reacts with carbon species accumulated on or in the vicinity of Ni crystallites due to CH4 cracking to produce CO. However, too strong basicity decreases catalytic activity by covering the active sites with strong and excess amounts of carbonate. 4

Prevention of Coke Formation by Combining Akaline-earth or Rare-earth Oxides

The effect of the addition of alkaline earth oxides or rare-earth oxides to supported Ni catalysts on the prevention of coke formation and consequent enhancement in methane to syngas conversion has recently been recognized, and a number of studies have focused on this.6-9*87-106 Gadalla et al. compared various commercial Ni catalysts supported on different kinds of supports and found that MgA1204, calcium aluminate or Ca0-Ti02-A1203 gave high conversion. Takayasu and Matsuura et al. 87988

699989

140

Catalysis

100 90 n

80

E

70 60

p

50

8

40

E

xw 30 20 10 0

0

20

40

I

I

I

60

80

100

120

Time on stream (h) Figure 5

Conversion of CH4 in CO, reformingfor reduced NiOlalkaline earth metal oxides: 790°C, CH4IC0, = 1 :1, I atm, SV60000 h-' (from Ref. 90) (Reproduced by permission from Appl. Catal. A: Gen., 1995,133, 149)

reported that Ni supported on MgO exhibited stable activity for C02 reforming of methane compared with other supports, and simple mechanical mixing of Ni/MgO with other supports, such as SO2, markedly prevented deactivation. The stability was confirmed over 600 h on stream. Choudhary et aZ.91found that precoating with MgO for the commercial low surface area porous catalyst carriers was also effective. Ruckenstein and Hugoreported that MgO inhibits the disproportionation reaction, Eq. (6), over Ni, probably due to the formation of a NiO-MgO solution, as a result of the similar crystalline structures of NiO and MgO. The same conclusion was drawn by Bradford and Vanniwg2 The unique characteristics of NiO-MgO solid solutions have been studied in detail, especially by Ruckenstein and H ~ ~ ~ and 9 Tomishige ~ ~ - and ~ Fuji~ 9 mot0 et aZ.96998,1019102 As shown in Figure 5, Ruckenstein and Hugoshowed a striking contrast between the stability of Ni/MgO and other catalysts. Ni/MgO maintained a constant conversion level to 12 h, whereas other catalysts lost their activity abruptly . The problem with Ni/MgO is its low intrinsic activity. At a space velocity of 6 x lo4 ml g-' h-l, the conversion almost coincided with equilibrium; however, when space velocity was increased 4.2 times, the conversion decreased from 95 to 60% at 790°C. Tomishige et d g 8 also showed the stable catalytic performance of a Ni0.03Mg0.97 catalyst. At a space velocity of 18 670 1 kg-' h-l, methane conversion was ca. 82% for 100 days, although Ni (3 mol%)/A1203 lost its activity within one day. However, at a space velocity of 224000 1 kg-' h-l, methane conversion decreased to a level below 11%,lo2 i.e. the space-time yield of CO was as low as 66.7 mol kg- h- l .

~

~

4: Reforming of CH4 by C02, 0 2 andlor H20

5

141

Attempts to Avoid Equilibrium-limited Conversion

To avoid the low equilibrium conversion for reaction (4) at lower temperatures, a membrane reactor was tested by using suitable catalysts to minimize coke formation, including Rh or Ru'07 and Ni/La203 or Ni0.03Mg0.970.lo* Ponelis and Zyllo7 carried out the C 0 2 reforming of CH4 using a membrane reactor. This consisted of a commercially supplied porous ceramic membrane provided by Membralox Co. and a reforming catalyst of 5 wt% Ru supported on Si02. Operating conditions were 500 "C, C02/CH4molar ratio 1, at 100 and 200 kPa, and space velocities of less than 8000 h-l. Conversion of CH4 was 50%, whereas the equilibrium conversion is 15% under these reaction conditions. However, above such space velocities the conversion decreased. Prabhu et aZ.lo8 studied C02 reforming of CH4 using the above mentioned Ni solid solution and Ni/La203. Incorporation of a Vycor glass membrane in the reactor resulted in higher conversions than obtained in an identical plug flow reactor. H2 transport through the membrane limits the performance of the mol membrane reactor, and a maximum yield of hydrogen was 3.1 x cm-' s - l , which is too low a rate compared with the criteria mentioned in Section 1. In general, the conversion exceeded the equilibrium conversion by using membrane reactors; however, because H2 transport through the membrane limits the performance of the membrane reactor, an extremely low space velocity must be used and a correspondingly very low space-time yield of hydrogen was obtained.

6

Kinetic and Mechanistic Studies

Few studies on the mechanism of the C02/CH4 reforming reaction have been conducted, mainly because the objective of most studies has been to develop high performance catalysts to maximize conversion without coke formation. Most studies are focused on catalyst development. In one study, the role of La203 was shown by kinetic-isotope measurements to involve strong chemisorption of C02.1°9 It has also been shown that a Ni/TiO2 catalyst exhibited the highest activity both in kinetic studies by pulse surface reaction rate analysis' lo and evaluation based on the specific turnover frequency. However, both of these studies were conducted at low conversions and the results are not directly related to the performance at high conversions. Transient kinetic techniques' l 2 and measurement of isotope effect' l 3 generally have the same limitation. A comprehensive summary of kinetic data before 1997, including turnover frequency, partial pressure dependence and apparent activation energy for C 0 2 reforming of methane, is presented in the review by Bradford and Vannice.21They pointed out that since many of the kinetic data appearing in the literature were obtained under low space velocity conditions, these give near equilibrium conversions and can be altered by the reverse reactions, i.e. methanation of CO. Although numerous kinetic models have been proposed

142

Catalysis

to describe the C02 reforming of CH4, the most consistent model currently available is as follows. At least the reverse dissociation of CH4 to yield CH, species and H2 and nondissociative adsorption of C02 on the support occur. In the metal-support interfacial region, H-promoted C02 dissociation and the reaction of CH, species with OH (or 0)species to yield CH,O species occur. After this review only a few papers122**13 have appeared. These studies are different from ordinary kinetic analysis; Schuurman et aZ. l 2 and Osaki et al. l3 introduced the steady-state isotopic kinetics techniques (SSITK) and the isotope effect calculation method, respectively. The formers' mechanistic conclusions on the basis of the transient kinetics for Ni/Si02 catalyst are that CH4 activation proceeds via reaction between gas phase and surface, forming adsorbed carbon and gaseous hydrogen. The dissociative adsorption of C02 would lead to surface oxygen and CO. The surface carbon monomers react with surface oxygen atoms and also form CO. For Ru/Si02 catalyst a different mechanism was induced. CH4 activation is similar to that for Ni/SiO2; however, in contrast with Ni, the gaseous C02 reacts directly with the adsorbed carbon species to give two molecules of CO, and no oxygen is left on the surface. The latter authors113 induced the conclusion that the step of CH,O + CO + x/2H2 is rate determining for C02 reforming of CH4 on a supported Ni catalyst. These analytical techniques would give more precise insight to differentiate the catalytic properties of different kinds of metals for C 0 2 reforming of CH4.

7

High Ni-loading Catalyst Supported on ZrOz Prepared by a Novel Method

The reason for low coke formation on the Ni-MgO solid solution is that a very low concentration of Ni (0.03 mol%) is highly dispersed and anchored in the matrix of magnesium oxide crystal, and therefore Ni metallic clusters are not exposed on the catalyst surface. This results in less deactivation at the expense of intrinsic catalytic activity. Recently, Z.X. Cheng and Q.M. Zhu et aZ.16,38 reported that a high metal loading of Ni (27 wt%), prepared by a unique impregnation on the Zr02, exhibited higher activity than a Ni-MgO solid solution catalyst98 even at higher space velocity, without deactivation due to coke formation. The ultra-fine Zr(OH)4 (6 nm, 160 m2 g-') particles designated as Zr(0H)-B were prepared by drying under supercritical ethanol (516 K, 6.3 MPa) of an alcogel in an autoclave for 1 h, followed by release of the pressure in flowing N2 to reduce the temperature. The alcogel was prepared from the chlorine-free wet precursor of Zr(OH)4 by exchanging water for ethanol several times at room temperature. The Zr(OH), precursor was prepared beforehand by addition of 0.17 M ZrOC12 solution into a 2.5 M ammonia water solution with pH = 9- 11, followed by removal of chlorine. The catalytic activity of this catalyst for C02 reforming of CH4 at 1030 K, with a feed (CHdC02 = 1:1) rate of 24000 ml g-cat-' h-' did not deactivate for 600 h, although self-oscillation

4: Reforming of CH4 by CO,,0 2 andlor HZO

143

of CH4 conversion between 80-85% was observed. The performance was markedly affected by the preparation method of Zr02. Zr(OH)4-A, prepared by using water, had a particle size of 18 nm, and the Ni catalyst loaded on this support precursor deactivated rapidly in C02 reforming of methane. This indicates that degree of Ni dispersion (0.06 and 0.09 for the samples A and B, respectively) sharply affects on the coke deposition. The result of this indicated the potential for high and stable activity of a well-designed Ni/Zr02 catalyst. 8

Towards Rapid Reforming of Methane - Catalytic Partial Oxidation of Methane to Syngas

In order to increase the rate of oxidation of methane to syngas and mitigate the problems of coke deposit, most recently selective catalytic partial oxidation 14- 128 This of methane to produce syngas has been studied exten~ively.~~.' reaction has attracted attention due to its moderate exothermicity, as expressed in Eq. (9), and suitable H2/CO ratio of products for the synthesis of methanol and Fischer-Tropsch type reactions. CH4 + 4 0 2 -+ 2H2 + CO,

AH" = -36 kJmol-'

(9)

This reaction has been denoted as CPO or OMS by de Smet et a1.117 and Beretta et aZ.,118or Au et a1.,l19 respectively; however, these abbreviations are still not widely used. To differentiate from the oxidative coupling of methane (OCM), here the reaction expressed as Eq. (9) is denoted as COMS for the abbreviation of Catalytic Oxidation of Methane to Syngas. In the most recent study on COMS, a single component of a platinum-group metal is used as catalyst. Rh catalyst is the most popular. Supports include y-A1203, 8,122 a-Al2O3,124 MgO, alumina-magnesia, 126 Ti02,121 he~aaluminate,'~~ La203,122,127 Y203127Ta205,127and SO2.120 Pt supported on Ce02--Zr02,125Pt gauze,127,128 Pt-10% Rh gauze,123Pt-Co supported on C L - A ~ and ~ O Ir~ supported ~'~ on Ti02121were also tested. Only one paper116on a Ni catalyst supported on y-A1203with additives of alkaline or alkaline earth oxides has been reported. Ni-based multi-component catalysts81.82developed by the author et aZ. will be explained in some detail in the following section. In general, the reaction is carried out at atmospheric pressure and very high space velocities such as 50 x lo4 - 72 x lo4 ml g-cat- h-l, corresponding to contact times of 7-5 ms.81,82,114,120,122,124,127 Rh supported on y-A1203,a-A1203or MgO exhibited a very stable activity with a high conversion level for or 120 h114on stream; however, in the case of Si02 support,120 the activity rapidly deteriorated by coke deposition. Hexaaluminate as support for Rh showed a less activity and concluded that the selectivity to CO/H2. Ruckenstein and Wang' reason for the high and stable activity of the catalyst Rh supported on MgO is the formation of a stable composite metal oxide as MgRh204, and the catalytic function of Rh is properly moderated. 14y1

12091229127

14912091229127

144

Catalysis

Au et aZ. l9 conducted a theoretical study on the comparison between single component catalysts such as Rh, Ru, Ir, Os, Pd, and Pt, and concluded that the most efficient catalyst for methane dissociation is Rh, which results in a high CH4 conversion. This roughly supports the experimental data mentioned above. The results obtained for Rh supported on alumina or magnesia catalyst meet the criteria for space-time yield of syngas mentioned in Section 1.3. However, in principle, COMS has a fatal disadvantage that, unlike steam reforming, the produced hydrogen is derived only from methane, and hydrogen in inexpensive water cannot be utilized, and also expensive oxygen is required. Moreover, complete oxidation of methane to H 2 0 and C02 competes with COMS, especially at higher conversion levels, causing decrease in selectivity to H2.117*1239124 This is also a common problem in the methane oxidative coupling to synthesize C2 compounds. Furthermore, Rh is extremely expensive and therefore, from the view point of industrialization, the use of Rh as the main catalyst component is impractical.

9

Performance of Ni-based Composite Catalyst to Carry Out the Ultra-rapid CO2 Reforming of Methane

Two decades ago Inui et aZ.79 developed a highly active catalyst for C02 methanation. The catalyst was composed of Ni-La203-R~ with the atomic ratios of 100:20:3 supported on spherical silica particles having a meso-macro bimodal pore structure. La203 and Ru could be replaced by Ce2O3 and Rh, respectively, The catalyst allowed C02-CO comethanation with complete conversion at considerably lower temperature than the conventional CO methanation cata1y~t.I~~ The role of each catalyst component was elucidated. Lanthanide oxides increase the adsorption of C02 by their weak basicities and dispersion of Ni, and Ru or Rh promote rapid hydrogen spillover during the reaction. The spillover hydrogen keeps the Ni catalyst surface in a reduced state and prevents coke deposition.80This catalyst concept was applied to the rapid purification of hydrogen containing a low concentration of 0 2 to produce ultra-high purity hydrogen using a Ni-Ce203-Pt catalyst supported alumina-washcoated ceramic fiber plate.130 This catalyst also exhibited high performance in methanol to syngas conversion. These kinds of catalysts showed high activity in both C02 reforming81and partial-oxidation reforming of methane.82.134This kind of catalyst was then applied to ultra-rapid C 0 2 reforming of methane. Even at a considerably high space velocity of 73 000 h-l (contact time of 49.3 ms), conversions of methane and C 0 2 on the Ni-Ce203-Pt catalyst reached the equilibrium conversion at a range of catalyst temperature from 350 to 650 "C, corresponding to conversions from 0 to 80% for a feed gas composed of 10% CH4-10% C02-80% N2.81However, as shown in bar @ in Figure 6,I4l the conversion decreased to 1/3 at a high space velocity 730000 h-' or contact time of 4.93 ms on this catalyst. The 14781,82?132-141

4: Reforming of CH4 by COZ, 0 2 andlor H20

145

s u

41.5

Ccl

0

30 20

10

0 0.5

Kind of Catalyst Figure 6

Synergistic effect of Rh-modijied Ni-Ce2 O r P t four-component catalyst on C02 reforming of methane Catalyst composition: @ = I 0 WPA Ni-6.0 WPACe2O3; Q=l.Owt!?A Pt; 0 = 0 . 2 wt!?ARh; @ = @ + Q = 1 0 WPANi-6.0 WPA CezO3-1.0 WPAPt; 0 = 0 + 0 = 1 0 wPA N i - 6 . 0 WPA Ce203-0.2 wPARh; @ = 0 + 0 + 0 = 1 0 wt?? N i - 6 . 0 wt!?!? Ce2O3 - 1.0 wt!?!? Pt - 0.2% Rh Catalyst support: Alumina-wash coated ceramicj b e r plate Feed gas (voPA):I0 CH4 - 10 C02 - 80 N2, S V: 730,000h - I , CT.. 4.93 ms; furnace temp.:600 C Numerals on bar graph: CH4 conversion PA) (Reproduced by permission from Res. Chem. Intermed., 1998,24,373)

Ni-Ce203-Pt catalyst was then modified with very low concentration of Rh. The Rh-modified Ni-based four-component catalyst, bar 0 in Figure 6, exhibited high reaction rate. The conversion increased to the equilibrium conversion, even under such high space velocities due to two-stage hydrogen spillover from Rh to Pt and then from Pt to Ni. The conversion on catalyst 0 was several percent lower than equilibrium. This is attributed to the decrease of the catalyst-bed temperature from the furnace temperature due to the large endothermic reaction heat .81 To compensate for the large endothermic reaction heat, partial oxidation reforming of methane (Eq. 9) or complete combustion (Eq. 10) were combined with C02 reforming CH4 + 202 + C02 + 2H20,

AH773K= - 800 kJ mol-'

(10) of methane, and much higher conversion beyond the equilibrium conversion was achieved134This principle was then applied to C02-H20 coreforming of methane as Eq. (11), and a very high space-time yield of H2, 12 190 or syngas 17070 moll-' h-' was obtained.'34 6 C H 4 + C 0 2 + H 2 0 + 2 0 2+ 13H2+7CO, AH773K=+383kJmol-I (11)

This concept was then extended to use other combustible hydrocarbons. Ethane and propane-are components of natural gas, and these hydrocarbons

Catalysis

146

Q

(36.2%),16,800 moVl h

........................... .. .. .. .. .. .. .. .. .. .. . Q .................................................. .........................

20,360 moU * h (80.8%),3,160 moVl h

............................... Q .............................................................

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

24,870 moVl h 0

5,000

10,OOO

15,000

20,000

25,000

Space-timeyield of syngas (moVI*h) Figure 7

Effect of catalytic oxidation of ethane or propane on the C02 reforming of methane at two kinds of space velocities 70 000 and 730 000 h-' Catalyst: The Rh-modifiedfour-component catalyst Gas composition: 035% CH4 - lO?? C02 - 55% N2; Q 35% CH4 - 1005% C02 - 5% CzH6 - 1 7.5% 0 2 - 32.5% N2; 0 35% CHd - 1 O?? C02 - 3.3% C3Ha - 16.5% 0 2 - 35.2% N2 Catalyst-bed temperature: 700 C (furnace temperature: 500 C); pressure: 1 atm (Reproduced by permission from Res. Chem. Intermed., 1998,24, 373) O

are very easily combustible on the catalyst while generating high combustion heat as shown in Eqs. (12) and (13). C2H6 + 3 . 5 0 2 -+2C02 + 3H20,

AH773~ = - 1424 kJ

C3H8 + 5 0 2 --+ 3co2 + 4H20,

AH773= ~

mol-'

(12)

-2040 kJ mol-'

(13) Since catalytic combustion rates of C2+ hydrocarbons are much higher than those for methane combustion and reforming reactions of methane, reactions (12) and (13) progress predominantly from a considerably lower temperature range around 300-350 "C. However, after rapid elevation of catalyst-bed temperature due to the combustion heat, at that temperature reforming of methane immediately participates. The total reaction heats are then shown as Eqs. (14) and (1 5). (~+5y)CH4+ xCO2 +yC2H6 + 3 . 5 ~ 0 + 2 (2~+7y)CO+ (2~+13y)H2, AH773K = 259x-24Oy kJ mol-1 (14) (X+72)CH4 + x C O + ~z C ~ H + 5~ ~ AH773~ = 259x-3752 kJ mol-

0+ 2

(2~+10z)CO+ ( 2 ~ 1+8z)H2,

(15) In fact, as shown in Figure 7141 the reaction progressed smoothly and the effect of this in situ heat supply gave a marked increase of conversion, and consequently space-time yield of syngas.138 By controlling the feed composition, a syngas having an appropriate composition to fit the successive reactions

4: Reforming of CH4 by C02, 0 2 andlor H20 Hydrocarbons in feed gas

&)

147

CH4 conv.

-56

44.2

+30

55.3

+7 1

78.3

CHp

+

C2H6

I

I

I

I

I

I

0

5,000

10,000

15,000

20,000

25,000

I

30,000

STY (rnoyl-h) Figure 8

Results of steam reforming of methane enhanced by on-site heat supply with combination of catalytic combustion of easily combustible hydrocarbons Catalyst: Ni-Ce2OrPt-Rh catalyst, in a unit of honeycomb type catalyst by cross-type packing: ( a ) Catalyst bed temperature - furnace temperature, furnace temperature 700 C; Feed gas: CHd 30 mop?, H 2 0 30 mop?, $xed; ( b ) N2 40 mop?; ( c ) c2H6 3.0-02 10.5-N2 26.5 (mop?); ( d ) C3H8 3.0-02 15.0-N2 22.0 (moPA); ( e ) , ( f ) H2 originated from c&, C3Ha added (conversions of C2H6 and C3H8 were 1OoOA); SV 730,000 h-I (CT 4.93 ms) (from Ref. 142)

would be obtained without the CO-H20 shift reaction or C02-H2 partial reduction reaction. The same concept was applied to the steam reforming of methane, and a marked effect on the conversion of methane to syngas could be obtained142as shown in Figure 8. Since no coke formation occurs even Ni-based catalyst due to the vigorous hydrogen spillover effect and catalytic combustion by oxygen, on ultra-high conversion rate of C 0 2 reforming and H 2 0 reforming of methane was achieved at considerably lower temperature with little deactivation, even under medium pressure conditions. Industrial application of this new catalytic process for H20 reforming of natural gas is now successfully progressing in Air Water Inc., Japan, and detailed reports will be presented in the near future. Recently, Choudhary et aZ.86reported on partial oxidation of methane to syngas with and without simultaneous C02 and H 2 0 reforming reaction using NdAlP04 as catalyst at a space velocity of 47 000 ml g-cat - h- l . They26also report on C02-HZ0 coreforming of Cox Nil -,O supported on macroporous silica-alumina precoated with MgO. More recently, Song25proposed a concept of tri-reforming of natural gas (oxy-C02-steam reforming) to utilize C02 in flue gas. They made a computational analysis of the energy implication of this reaction under pressurized condition^.^^ Even in conventional steam reforming processes, combustion of a part of the methane feed can be used, and is sometimes called autothermal reforming.

Catalysis

148

However, this involves homogeneous partial oxidation and/or combustion of methane before entering the catalytic reforming reactor. On the other hand, in the cases of Inui et al. 134-142 and Choudhary et al. ,86 the oxygen is consumed almost completely in catalytic combustion andlor COMS, i. e. the oxidation heat is directly generated on the same catalyst surface where the reforming reaction take place. This compensates for the endothermic reaction heat of the reforming reactions. The combustion products, H20 and C02, are used as part of the source for reforming of unconverted methane. Therefore, this new type of combined reaction would be differentiated from 'autothermal reforming', and should be called rather simply and clearly C02-02-co-reforming, H2-02co-reforming, and C02-H20-02-co-reforming. 10

Application to Other Reactions

Most conventional multi-step conversion of natural gas to synthesize valuable compounds takes place in independent reactors, and the products obtained in each reactor are introduced to the next reactor after separating unnecessary compounds such as unconverted feed materials and by-products, which hinder the desired reactions. These complexities increase the accumulated expenses of the total process. On the other hand, if these sequential processes were conducted by one-pass operation using multi-step reactors connected in series, the total process becomes simple, helping to reduce costs. To realize these one-pass operations, the catalyst for the use in each reactor should be developed to have highly selective performance for the required reaction and high tolerance against unnecessary components in the feed gas. In order to increase the value of C02 conversion, this kind of consideration is necessary. Therefore, although Section 10 is simply one option in the light of this chapter, accumulated knowledge for one-pass operations to synthesize various high value compounds from C02 as the source is briefly introduced here. By using C02 reforming of methane in combination with steam reforming and oxidation reactions, syngas having an appropriate ratio of hydrogen and carbon oxides can be obtained with a high efficiency. The syngas can then be introduced directly into a second reactor connected in series. The second reactor, containing a methanol synthesis c a t a l y ~ t or ' ~ethanol ~ ~ ~ synthesis ~ ~ ~ ~ catalyst,143produces methanol or ethanol. The total effluent gas from the methanol synthesis reactor can be directly introduced into the third reactor, in which zeolite catalysts are contained, producing low aromatic, high octane gasoline or light olefins with high selectivities.147140 11

Summary of the Review

The essential requirements of catalytic technologies for both effective utilization of natural gas and efficient decrease in C 0 2 discharge are described based

4: Reforming of CHd by C02, 0 2 andlor H20

149

on the recent papers on methane to syngas conversion. Among the various requirements, high conversion rates are needed to react the tremendous amount of feed gas. In the case of methane conversion, high tolerance against deactivation due to coke deposition is most important. In spite of many advantages, C02 reforming of methane has the following general problems. Unlike steam reforming, the coke deposited by the disproportionation of CO and the cracking of methane in C02 reforming cannot be volatilized by the reaction with steam. Based on this mechanism of coke formation, numerous attempts to reduce or prevent coke deposit have been carried out by many researchers. Many of these studies compare the effect of using different kinds of catalyst components and supports. Taking into account the catalytic performance and cost of catalyst materials, Ni supported on the alkaline-earth oxides and rare-earth oxides, especially magnesia have been studied extensively. Sufficient stability for magnesia supported catalysts was observed, but the conversion rate is not high. More recently, it has been reported that a high loading Ni catalyst supported on zirconia prepared by a novel process exhibited higher activity with lower coke deposit than the nickel-magnesia solid solution catalyst. Steam-C02 co-reforming and other approaches involving membrane reactors have also been investigated. This method has advantages that the reaction can be operated at much lower temperatures and carbon oxides-free hydrogen can be obtained. However, these results are limited by the intrinsic performance of the catalysts used and limits of diffusion across the membrane. Therefore these materials may be limited to use as smaller scale devices for hydrogen production. The most recent trend in the research is the use of O2 and H20 simultaneously with C02 in methane reforming. Even a low concentration of 0 2 in the feed gas dramatically decreases coke deposition. The addition of 0 2 to the reactants brings about an oxidation heat which compensates for the large endothermic heat of the reforming reaction, resulting in a marked enhancement in the reaction rate of the reforming. Simultaneous use of H20 gives an appropriate H2/C0, molar ratio directly to match the down stream use of the syngas produced. A typical approach, in the light of this, has been carried out by Inui et al. A four-component Ni-based composite catalyst was designed which highly promoted the hydrogen spillover properties. High concentrations of hydrogen on the catalyst surface prevented coke deposition and achieved an extraordinarily high reaction rate and a low rate of deactivation by limiting the coke on the active catalytic sites. The large endothermic reaction heat of the reforming of methane could be compensated by simultaneous combination of catalytic combustion of a low concentration of easily combustible hydrocarbons added in the feed. In order to overcome coke deposition in the C02 reforming of methane, Ru and Rh catalysts were studied. Simultaneously, the effect of catalyst support was investigated, and it was found that alkaline-earth or rare-earth oxides suppress coke deposition. Recently, Ni supported on MgO and Ni-MgO solid 14981,82~132-141

150

Catalysis

solution catalysts have been studied extensively due to their characteristically low coke deposition. Another approach is based on a Ni-based catalyst combined with small concentration of Ce203 modified with very small concentrations of Pt and Rh, which is designed to maximize hydrogen spillover. In this review, studies on C 0 2 reforming and related reactions over the past decade were surveyed and summarized to contribute to the further improvement of catalysts and reaction processes for the reforming of methane and/or natural gas.

12

Acknowledgement

Most of the studies cited from the literature, in which the author’s name (T. Inui) is involved, were performed in the author’s laboratory at Kyoto University before the author’s retirement. I acknowledge all the co-authors. The author also acknowledges all the publishing companies, who have permited publication of the following Figures in this chapter. Reference numbers shown below are the same as the numbers shown in the Reference section in this chapter. Figure 1 is drawn from Ref. 28 (Academic Press). Figure 2 is drawn from Ref. 59 (Elsevier Science). Figure 3 is drawn from Ref. 77 (J.C. Baltzer AG). Figure 4 is drawn from Ref. 73 (Elsevier Science). Figure 5 is drawn from Ref. 90 (Elsevier Science). Figures 6 and 7 are drawn from Ref. 141 (John Wiley & Sons Ltd.).

13 1 2 3 4 5 6 7

8

9 10 11

References J.R. Rostrup-Nielsen and I. Alstrup, Catal. Today, 1999,53, 31 1. T.V. Choudhary and D.W. Goodman, Catal. Lett., 1999,59,93. K. Hou, M. Fowles and R. Hughes, Catal. Today, 2000,56, 13. E. Kikuchi, Y. Nemoto, M. Kajiwara, S. Uemiya and T. Kojima, Catal. Today, 2000,56, 75. E. Kikuchi, Catal. Today, 2000,56,97. I. Matsuura and 0. Takayasu, Proc. Intern. Symp. on Chemical Fixation of Carbon Dioxide, Nagoya, 1991, pp. 247-252. K. Seshan and J.A. Lercher, in Carbon Dioxide Chemistry: Environmental Issues, ed. J. Paul and C.-M. Prodier, Royal Society of Chemistry, Cambridge, 1994, pp. 16-19. T. Inui, in Carbon Dioxide Chemistry: Environmental Issues, ed. J. Paul and C.-M. Prodier, Royal Society of Chemistry, Cambridge, 1994, pp. 64-68. 0. Takayasu, Y. Takegahara and I. Matsuura, Energy Convers. Mgmt., 1995,36, 597. T. Inui, H. Hara, T. Takeguchi, K. Ichino, J.B. Kim, S. Iwamoto and S.B. PU, Energy Convers. Mgmt., 1997,38, Suppl. S385. 0. Takayasu, F. Sato, K. Ota, T. Hitomi, T. Miyazaki, T. Osawa and I. Matsuura, Energy Convers. Mgmt., 1997,38, Suppl., S391.

4: Reforming of CH4 by C02,0 2 andlor H20

12 13 14 15 16

17 18 19 20 21 22 23

24

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

151

S.E. Park, J.S. Chang, H.S. Roh, M. Anpo and H. Yamashita, Stud. Surf:Sci. Catal., 1998,114,395. P.F. Aparicio, B.B. Baeza, A.G. Ruiz and I.R. Ramos, Stud. Surf:Sci., Catal., 1998,114,399. T. Inui, Green Gas Control Technology, ed. P. Riemer, B. Eliasson and A. Wokaun, Elsevier, Amsterdam, 1999, pp. 331-336. A.D. Schmitz and T. Yoshida, Green Gas Control Technology, ed. P. Riemer, B. Eliasson and A. Wokaun, Elsevier, Amsterdam, 1999, pp. 355-360. Z. Cheng, X. Zhao, J. Liu and Q. Zhu, Green Gas Control Technology, ed. P. Riemer, B. Eliasson and A. Wokaun, Elsevier, Amsterdam, 1999, pp. 359-384. J.M. Fox 111, Catal. Rev. -Sci. Eng., 1993,35(2), 169. J.R. Rostrup-Nielsen, Natural Gas Conv. 11, ed. Curry-Hyde and R.F. Howe, Elsevier, Amsterdam, 1994, p. 25. J.R.H. Ross, A.N.J. van Keulen, M.E.S. Hegarty and K. Seshan, Catal Today, 1996,30,193. M.M. Halmann and M. Steinberg, Green Gas Carbon Dioxide Mitigation, Sci. Tech., Lewis Publishers, Boca Raton, 1999, p. 315. M.C.J. Bradford and M.A. Vannice, Catal. Rev. - Sci. Eng., 1999,41(1), 1. T. Inui, Preprints Fifth Intern. Confer. C02 Utilization, Karlsruhe, Sept. 1999, p. 20. Full paper, J. Appl. Organometal. Chem. 2001,15(2), 87. T. Tagawa, M. Ito and S. Goto, Preprints Fifth Intern. Confer. COz Utilization, Karlsruhe, Sept. 1999, p. 32. Full paper, J. Appl. Organometal. Chem. 2001, 15(2) 127. T. Inui, Preprints Symp. C02 Conversion and Utilization in Refinery and Chem. Process, ACS 219th Annual Meeting, San Francisco, 2000, pp. 113-117. Full paper ACS Symp. Series, in press. C. Song, ibid., pp. 159- 163. Full paper, ACS Symp. Series, in press. V.R. Choudhary, A.S. Mamman, B.S. Uphade and R.E. Babcolk, ibid, p. 164-167, Full paper, ACS Symp. Series, in press. W. Pan and C. Song, ibid., pp. 168-171, Full paper, ACS Symp. Series, in press. J.R. Rostrup-Nielsen and J.-H.B. Hansen, J. Catal., 1993, 144, 38. L. Basini and D. Sanfilippo, J. Catal., 1995, 157, 162. D. Qin, J. LapszewiczandX. Jiang, J. Catal., 1996, 159, 140. S. Wang and G.Q. Lu, Energy Fuels, 1996,10,896. P. Ferreira-Aparicio, A. Guerrero-Ruiz and I. Rodriguez-Ramos, Appl. Catal. A: Gen., 1998,170, 177. M.C.J. Bradford and M.A. Vannice, Catal. Today, 1999,50,87. V.A. Tsipouriari and X.E. Verykios, J. Catal., 1999, 187, 85. F. Arena, F. Frusteri and A. Parmaliana, Appl. Catal., A: Gen., 1999, 187, 127. J.Z. Luo, L.Z. Gao, C.F. Ng and C.T. Au, Catal. Lett., 1999,62, 153. U.L. Portugal Jr., C.M.P. Marques, E.C.C. Araujo, E.V. Morales, M.V. Giotto and J.M.C. Bueno, Appl. Catal., A: Gen., 2000,193,173. J.M. Wei, B.Q. Xu, J.L. Li, Z.X. Cheng and Q.M. Zhu, Appl. Catal., A: Gen., 2000,196, L169. J.H. Kim, D.J. Suh, T.J. Park and K.L. Kim, Appl. Catal., A: Gen., 2000, 197, 191. A. Erdohelyi, J. CserCnyi and F. Solymosi, J. Catal., 1993,141,287. Z.L. Zhang, V.A. Tsipouriari, A.M. Efstathiou and X.E. Verykios, J. Catal., 1996, 158, 51.

152 42 43 44 45 46 47 48 49 50 51

52 53 54 55

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78

Catalysis

A.M. Efstathiou, A. Kladi, V.A. Tsipouriari and X.E. Verykios, J. Catal., 1996, 158, 64. M.F. Mark and W.F. Maier, J. Catal., 1996,164, 122. R.N. Bhat and W.M.H. Sachtler, Appl. Catal. A: Gen., 1997,150,279. P. Gronchi, P. Centola and R. Del Rosso, Appl. Catal. A: Gen., 1997,152,83. H.Y. Wang, Appl. Catal. A: Gen., 1997,155,239. P.F. Aparicio, M.F. Garcia, A.G. Ruiz and I.R. Ramos, J. Catal., 2000, 190, 296. A. Erdohelyi, K. Fodor and F. Solymosi, Stud. Surf:Sci. Catal., 1997,107,525. M.C.J. Bradford and M.A. Vannice, J. Catal., 1999,183,69. P. Ferreira-Aparicio, C. Marquez-Alvarez, I. Rodriguez-Ramos, Y. Schuurman, A. Guerrero-Ruiz and C. Mirodatos, J. Catal., 1999,184,202. Y. Schuurman, C. Marquez-Alvarez, V.C.H. Kroll and C. Mirodatos, Catal. Today, 1998,46,185. N. Matsui, K. Anzai, N. Akamatsu, K. Nakagawa, N. Ikenaga and T. Suzuki, Appl. Catal, A: Gen., 1997,179,247. F. Solymosi, G. Kutsan and A. Erdohelyi, Catal. Lett., 1991,11, 149. J.H. Bitter, W. Hally, K. Seshan, J.G. van O m e n and J.A. Lercher, Catal. Today, 1996,29,349. A.N.J. van Keulen, K. Seshan, J.H.B.J. Hoebink and J.R.H. Ross, J. Catal., 1997,166,306. A.N.J. van Keulen, M.E.S. Hegarty, J.R.H. Ross and P.F. van den Oosterkamp, Stud. Surf:Sci. Catal., 1997,107, 537. M.C.J. Bradford and M.A. Vannice, J. Catal., 1998,173, 157. J.H. Bitter, K. Seshan and J.A. Lercher, J. Catal., 1998,176, 93. A.M. O’Connor and J.R.H. Ross, Catal. Today, 1998,46,203. S.M. Staag, E. Romeo, C. Padro and D.E. Resasco, J. Catal., 1998,178, 137. J.H. Bitter, K. Seshan and J.A. Lercher, J. Catal., 1999,183, 336. T. Inui, Y. Ono, Y. Takagi and J.B. Kim, Appl. Catal., A: Gen., 2000,202:2,215. A. Guerrero-Ruiz, A. Sepulveda-Escribano, I. Rodriguez-Ramos, Catal. Today, 1994,21,545. T. Osaki, H. Masuda, T. Horiuchi and T. Mori, Catal Lett., 1995,34,59. Y. Lu, C. Yu, J. Xue, Y. Liu and S. Shen, Chem. Lett., 1997, 515. T. Osaki, T. Horiuchi, K. Suzuki and T. Mori, Catal Lett., 1995,35,39. T. Osaki, T. Horiuchi, K. Suzuki and T. Mori, Appl. Catal. A: Gen., 1997, 155, 229. A. Bhattacharyya and V.W. Chang, Stud. Surf: Sci. Catal., 1994,88,207. D. Halliche R. Bouarab, 0. Cherifi and M.M. Bettahar, Catal. Today, 1996, 29, 373. E. Ruckenstein and Y.H. Hu, J. Catal., 1996,162,230. S . Wang and G.Q.M. Lu, Appl. Catal. B: Environmental, 1998,16,269. Z.-F. Yan, R.-G. Ding, L.-H. Song and L. Qian, Energy Fuels, 1998,12, 1 1 14. T. Horiuchi, K. Sakuma, T. Eukui, Y. Kubo, T. Osaki and T. Mori, Appl. Catal. A: Gen., 1996,144,111. S . Wang and G.Q. Lu, Appl. Catal. A: Gen., 1998,169,271. S . Wang and G.Q. Lu, I.E. C. Res., 1999,38,2615. M. Ito, T. Tagawa and S. Goto, Appl. Catal. A: Gen., 1999,177, 15. C. Crisafulli, S. Scire, R. Maggiore, S. Minico and S. Galvagno, Catal. Lett., 1999,59,21. P.A. Sermon and G.C. Bond, Catal. Rev., 1974,8, 117.

4: Reforming of CH4 by C02,0 2 andlor H z 0

79 80

81 82

83 84 85 86

87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112

113 114

153

T. Inui, M. Funabiki, M. Suehiro and T. Sezume, J. Chem. SOC.,Faraday Trans. I , 1979,75,787. T. Inui, M. Funabiki and Y. Takegami, J. Chem. Soc., Faraday Trans. I . , 1980, 76,2237. T. Inui, Stud. Surf: Sci. Catal., 1993,77, 17. T. Inui, K. Fujioka, W. Tanakulrungsank and T. Takeguchi, Stud. Surf. Sci. Catal., 1993,77, 369. V.C.H. Kroll, H.M. Swaan and C. Mirodatos, J. Catal., 1996,161,409. T. Hayakawa, S. Suzuki, J. Nakamura, T. Uchijima, S. Hamakawa, K. Suzuki, T. Shishido and K. Takehira, Appl. Catal. A: Gen., 1999,183,273. J.S. Chang, S.-E. Park and H. Chon, Appl. Catal., A: Gen., 1996,145, 1 1 1. V.R. Choudhary, B.S. Uphade and A.S. Mamman, Microporous Mesoporous Mater., 1998,23,61 . A.M. Gadalla and B. Bower, Chem. Eng. Sci., 1988,43,3049. A.M. Gadalla, Chem. Eng. Sci., 1989,44,2825. 0 .Takayasu, C. Soman, Y. Takegahara and I. Matsuura, Stud. Surf: Sci. Catal., 1994,88,281. E . Ruckenstein and Y.H. Hu, Appl. Catal. A: Gen., 1995,133, 149. V.R. Choudhary, B.S. Uphade and AS. Mamman, Catal. Lett., 1995,32,387. M.C.J. Bradford and M.A. Vannice, Appl. Catal. A; Gen., 1996,142, 73. Y.H. Hu and E. Ruckenstein, Catal. Lett., 1996,35, 145. E. Ruckenstein and Y.H. Hu, Appl. Catal. A: Gen., 1997,154, 185. Y.H. Hu and E. Ruckenstein, Catal. Lett., 1997,43, 71. Y.-G. Chen, K. Tomishige and K. Fujimoto, Chem. Lett., 1997,999. V.R. Choudhary, B.S. Uphade and A.S. Mamman, Appl. Catal. A: Gen., 1998, 168,33. K. Tomishige, 0. Yamazaki, Y. Chen, K. Yokoyama, X. Li and K. Fujimoto, Catal. Today, 1998,45,35. A.A. Lemonidou, M.A. Goula and I.A. Vasalos, Catal. Today, 1998,46, 175. E. Ruckenstein and Y.H. Hu, I. E. C.,Res., 1998,37, 1744. K. Tomishige, Y.-G. Chen and K. Fujimoto, J. Catal., 1999, 181,91. Y.-G. Chen, K. Tomishige, K. Yokoyama and K. Fujimoto, J. Catal., 1999, 184, 479. E. RuckensteinandY.H. Hu, J. Catal., 1996,161, 55. A. Slagtern, U. Olsbye, R. Blom and I.M. Dahl, Stud. Surf: Sci. Catal., 1997, 107,497. S . Wang and G.Q. Lu, Energy Fuels, 1998,12,248. S. Wang and G.Q. Lu, Appl. Catal. B: Environ., 1998,19,267. A.A. Ponelis and P.G.S. van Zyl, Stud. Surf: Sci. Catal., 1997,107, 555. A.K. Prabhu, R. Radhakrishnan and S.T. Oyama, Appl. Catal. A: Gen., 1999, 183,241. Z. Zhang and X.E. Verykios, Catal. Lett., 1996,38, 175. T. Osaki, T. Horiuchi, K. Suzuki and T. Mori, J. Chem. SOC. Faraday Trans., 1996,92,1627. M.C.J. Bradford and M.A. Vannice, Appl. Catal. A: Gen., 1996,142,97. Y. Shuurman, V.C.H. Kroll, P. Ferreira-Aparicio and C. Mirodatos, Catal. Today, 1997,38, 129 , T. Osaki, H. Fukaya, T. Horiuchi, K. Suzuki and T. Mori, J. Catal., 1998, 180, 106. H.Y. Wang and E. Ruckenstein, Catal. Lett., 1999,59, 121.

154

Catalysis

115 116

S. Tang, J. Lin and K.L. Tan, Catal. Lett., 1999,59, 129. S . Liu, G. Xiong, W. Yang, L. Xu, G. Xiong and C. Li, Catul. Lett., 1999, 63, 167. C.R.H. de Smet, M.H.J. de Croon, R.J. Berger, G.B. Marin and J.C. Schouten, Appl. Catal., A: Gen., 1999,187, 33. A. Beretta, G. Groppi, L. Majocchi and P. Forzatti, Appl. Catal., A: Gen., 1999, 187,49. C.T. Au, C.F. Ng and M.S. Liao, J. Catal., 1999,185, 12. H.Y. Wang and E. Ruckenstein, J. Catal., 1999,186, 181. K. Nakagawa, N. Ikenaga, Y. Teng, T. Kobayashi and T. Suzuki, J. Catal., 1999,186,405. E. Ruckenstein and H.Y. Wang, J. Catal., 1999,187, 151. D.L. Iordanoglou, A.S. Bodke and L.D. Schmidt, J. Catal., 1999,187,400. L. Majocchi, G. Groppi, C . Cristiani, P. Forzatti, L. Basini and A. Guarinoni, Catal. Lett., 2000,65,49. P. Pantu, K. Kim and G.R. Gavalas, Appl. Catal., A: Gen., 2000,193,203. P.S. Maiya, T.J. Anderson, R.L. Mieville, J.T. Dusc Kj.J. Picciolo and V. Balachandran, Appl. Catal., A; Gen., 2000,196,65. E. Ruckenstein and H.Y. Wang, J. Catal., 2000,190, 32. M. Fathi, F. Monnet, Y. Schuurman, A. Holman and C. Mirodatos, J. Catal., 2000,190,439. T. Inui, M. Funabiki and Y. Takegami, Ind. Eng. Chem., Prod. Res. Dev., 1980, 19, 385. T. Inui, Y. Miyamoto and Y. Takegami, Stud. S u r - Sci. Catal., 1983,17, 181. Y. Nakazaki and T. Inui, Ind. Eng. Chem., Res., 1989,28, 1285. T. Inui, K. Fujioka, Y. Fujii, T. Takeguchi, H. Nishiyama, M. Inoue and W. Tanakulrungsank, Stud. Surf:Sci. Catal., 1994,81, 361. T. Inui, T. Takeguchi and J.-B. Kim, Trans. Mat. Res. SOC.Jpn., 1994,18A, 345. T. Inui, K. Saigo, Y. Fujii and K. Fujioka, Catal. Today, 1995,26,295. T. Inui, Catal. Today, 1996,29, 329. T. Inui, S. Yoshida, K. Saigo and S. Iwamoto, Catal. Today, 1997,35, 171. T. Inui, Appl. Surf: Sci., 1997,121/122,26. T. Inui, K. Ichino, I. Matsuoka, T. Takeguchi, S. Iwamoto, S.-B. Pu and S . Nishimoto, Korean J. Chem. Eng., 1997,14,441. T. Inui, K. Saigo, K. Ichino, T. Takami, T. Takeguchi, S. Iwamoto and Y . Fujii, Ceram. Trans., 1997,73, 39. T. Inui, Catal. Today, 1999,51,361. T. Inui, Res. Chem. Intermed., 1998,24, 373. T. Inui, T. Takeguchi and I. Matsuoka, Curr. Topics Catal., 1999,2,73. T. Inui, T. Yamamoto, M. Inoue, H. Hara, T. Takeguchi and J-B. Kim, Appl. Catal. A: Gen., 1999,186, 395.

117 118

119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138

139 140 141 142 143

5 Oxidative Functionalization of Ethane and Propane BY OLGA V. BUYEVSKAYA AND MANFRED BAERNS

1

Introduction

Much research effort is currently devoted to the chemical conversion of light alkanes to valuable chemicals. In particular, the increased demand for light olefins requires new catalytic routes comprising novel catalytic materials and modes of operation. Activities in the field of direct catalytic oxidative conversion of low-molecular-mass alkanes to petrochemical feedstocks and petrochemicals are reviewed elsewhere.1-5 For catalyst development, fundamental knowledge about the interaction between reactants and catalytic surfaces is essential. Such knowledge is also beneficial in applying combinatorial methods in the search for new materials as has been previously s h ~ w nIn. ~the ~ present ~ contribution, particular attention is given to recent developments (1998-2000) in the conversion of ethane and propane to their olefins by oxidative dehydrogenation and to their oxygenates, i.e. acetic acid, acrylic acid and acrolein by partial oxidation.

2

Mechanisms of Ethane and Propane Activation

The selective oxidation of low alkanes suffers because of the competing reaction involving their complete oxidation. Further progress towards any commercial realization of such processes requires further research and development work including a thorough understanding of the catalytic mechanisms and kinetics as a basis for identifying factors determining the selectivityconversion relationship. For metal oxide catalysts, three main types of mechanisms have been proposed which are based on the type of oxygen species involved in alkane activation.2 Primary steps for the respective schemes of selective alkane transformation are presented in Figure 1. In the Mars-van-Krevelen-redox mechanism, oxygen of the metal oxide takes part in the reaction by abstracting and oxidizing the hydrogen from the alkane (case A). The OH groups that are formed are then removed from the surface by dehydration. In the formation of alkenes and total oxidation Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 155

Catalysis

156

C A) Redoxmechanism

B) Activation by adsorbed oxygen

32n+2

CnHZn+, + Me0,H

MeO,

0

CnH2,+ MeO,-, + H20 0.50,i

MeO,

CnHZn+l + Me0,-OH

Meox-oad.

0

CnH2,+ MeO, + H 2 0 MeO, + 0.50,

C) Activation by lattice oxygen (no redoxmechanism)

Figure 1

CnH2n+1 + Me0,H

MeO,

0 0.502

CnH2,+ MeO, + H,O

Primary reaction steps of the oxidative dehydrogenation of alkanes on metal oxides for three different mechanistic assumptions A, B and C (Reproduced by permission from Catal. Today, 1998,45, 13)

products, the alkane reduces the catalyst surface. The catalyst is subsequently re-oxidized by gas-phase oxygen. In case B oxygen participates via its adsorbed state; hydrogen is abstracted forming OH groups on the catalytic surface which are removed by dehydration. The active surface oxygen is then restored by oxygen adsorption from the gas phase. In case C , strongly bound lattice oxygen abstracts hydrogen which is then removed via oxidation with gas phase oxygen. The functions of the catalytic material along with the desired and undesired properties, as well as the oxides which fulfil such requirements, are summarized for cases A to C in Table 1. Based on the different hypotheses for the initial surface processes of alkane activation three main groups of metal oxides can be defined. This classification agrees with metal oxide catalysts which have been reported in the literature for the oxidative dehydrogenation of ethane and propane as described in the subsequent Sections 3.1 and 4.1. 3

Oxidative Dehydrogenation of Ethane

Oxidative dehydrogenation of ethane (ODE) is an endothermic thermal pyrolysis, and has been the applying different catalysts in a temperature range under isothermal and autothermal conditions in monolith and membrane reactors.

alternative to the highly subject of many studies from 623 to ca. 1273 K fixed-bed, fluidized-bed,

0

0

0

0

medium 0-Me binding energy: enthalpy of oxide formation AfH per oxygen atom in the range -400 to -200 kJ mol- l; reference: V2O5 (-310 kJmol-l)

Properties to be avoided

activation of the alkane high &Me binding energy by strongly bound lattice (AfH per oxygen atom oxygen < -400 kJ mol- ') low activity towards oxygen adsorption

CaO ( - 636), SrO (- 590), Y203 (- 639, La203(- 598), Gd203 (- 607), Ce02 ( - 544), Tho2 ( - 613), Nd203 ( - 603), Sm2O3 ( -608)d)

V205 (- 310), Nb205( - 380), Moo3 ( -248), W 0 3 (- 28 l), MnO2 ( -261), Fez03 ( - 274), C0304 ( - 220), NiO ( -240), ZnO (- 35 l), CdO (- 260), Ga203 (- 363), In203(- 309), Ge02 (- 290), SnO2 ( -290)

Oxides fu&lling requirement$' (AfH per 0 atom, kJ mol-')

Be0 (- 598), MgO ( - 602), A1203(- 559), Sc2O3 (- 636), ZrO2 (- S O ) , Ti02 (-470), high rate of oxygen isotopic Si02 ( - 453), HfOz ( - 559), exchangec Ta205 ( - 409), B2O3 ( -41 8)d

formation of adsorbed high activity towards oxygen species activating oxygen adsorption: alkane high rate of oxygen isotopic exchangeb low activity of lattice low and medium 0-Me oxygen towards alkane binding energy: AfH per oxygen atom conversion > -400 kJ mol-'

product formation with participation of oxygen from the catalyst suppression of total oxidation caused by weakly bound lattice oxygen

Required properties

a

oxides which are stable under oxidizing conditions A(oidationG < O ) only are considered; rates of oxygen isotopic exchange from ref. 8 are used; oxides on which lg(rate)3500 > 10 are considered; oxides on which lg(rate)3500< 10 are considered (cf: ref. 8); own measurements on oxygen adsorption in the TAP reactor.

activation by lattice oxygen (no redoxmechanism)

Case C

activation by adsorbed oxygen

Case B

0

0

Catalyst function

& 2 G 9m

Q

rp

s

2. s

G.

F*

!s231

z

-.

Summary on catalyst functions, properties required or to be avoided for dvferent assumptions on reaction mechanism (cJ Y ? Figure I )

redox-mechanism

Case A

Mechanism

Table 1

Catalysis

158

3.1 Catalyst Types and Reactor Concepts. - ODE catalysts reported in the scientific literature can be divided into three main groups: (A) (mixed)oxides or supported catalytic materials consisting of reducible metal oxides, (B) non-reducible (mixed)oxides, (C) catalytic materials containing of noble metals. The various types of catalysts reveal in general different behaviour and require different reaction conditions, in particular reaction temperature, for their optimal performance. Most catalysts that are active at low temperatures ( T < 873 K) belong to group A. Many of these catalysts consist of oxides which fulfil the requirements of the redox mechanism (see Table 1). Different studies confirm the occurrence of a redox mechanism on catalysts consisting of transition metal oxides. The occurrence of a redox mechanism is well established for V-containing catalysts. For example, Valenzuela et al. confirmed this mechanism for the oxidative ethane dehydrogenation over mixed Ni-V-Sb oxide^.^ Le Bars and co-workers performed pulse experiments with ethane over V205/A1203catalysts and estimated the reactivity of lattice oxygen in the formation of ethylene.lo Most of the catalysts from second group are based on rare-earth metal oxides (REO) corresponding to case B (c$ Figure 1 and Table 1). There are also some papers dealing with other catalytic compositions consisting of nonreducible oxides (e.g. B203/A1203, lithium-magnesium oxide12). For the catalytic systems consisting of non-reducible oxides, the formation of gasphase ethyl radicals is in most cases the decisive step, and optimal reaction temperatures are generally above 873 K. At this temperature, catalytic and gas-phase activation can be important. In group C noble metals (preferentially Pt) are deposited on monoliths; on these catalysts the formation of ethylene occurs at temperatures higher than 1073 K.’’9l6 To improve ethylene selectivity and yield in ODE, halides, in particular chlorine-containing compounds, have been added to the reaction mixture and to the catalyst. For example, Conway et al. obtained ethylene yields of about 57% (S= 76%) over a Dy203/Li+-MgO-C1- catalyst at 834 K after 25 h.17The use of a BaC12N203catalyst allowed Dai et al. to obtain 53% yield of ethylene ( S = 74%) at 913 K.18 An ethylene yield of 68% ( S = 83%) was achieved by Wang et al. using a LiC1/6%S04/Zr02catalyst at 923 K, which was, however, reduced to 46% after 15 h of catalyst operation.” An ethylene yield of 69% (S= 82%) was achieved by Dai et al. using a YBa2CU307-0.2#0.16 catalyst at a In the authors’ opinion, F- and relatively high reaction temperature of 953 Cl-containing catalysts are durable although the time-on-stream behaviour was tested only within 40 h. Based on the results of different physico-chemical characterizations the authors assumed that the effect of halide on this catalyst is not due to its direct participation in the reaction but due to the incorporation of the halide ions into the YBa2Cu3O7-8lattice which enhances the activity of l 3 9 l 4

5: Oxidative Functionalization of Ethane and Propane

159

lattice oxygen. A similar conclusion was derived by Dai and co-workers for SrFeO3_&1, and Lal -,Sr,Fe03 -sX, (X = F, Cl) catalysts; the incorporation of halide ions into oxygen vacancies in perovskite-type oxides was proposed to be responsible for the catalytic activity of these systems.21.22 The use of halides in the ODE reaction results, however, in an increased corrosion and in a contamination of ethylene with chorine compounds, which is not desired for many subsequent applications, e.g. polyethylene synthesis. Moreover, the use of halide-containing catalysts suffers in general from leaching, for example as HCl, since water formed by complete oxidation and oxidative dehydrogenation is always present in the reaction mixture. Due to these facts the ODE using halides is not considered further in this paper. 3.I . 1 (Mixed) Oxides or Supported Catalytic Materials Consisting of Reducible Metal Oxides. A number of mixed oxides as well as heteropolyacids consisting of reducible (mostly transition) metals have been examined in a fixed bed reactor (see Table 2). An advantage of this type of catalyst is the use of relative low reaction temperatures (< 873 K) at which a pure heterogeneous reaction pathway prevails. Among the catalysts of this group Mo-V-Nb, Mo-V-Sb and Mo-V-Ta oxide systems showed high activity at low temperatures (623-673 K); an ethylene yield of 38% ( S = 65 YO)was obtained by Thorsteinson et al. on a M00.73V0.18Nb0.090, catalyst at 623 K.23Simulation of the ODE reaction over a Mo-V-Nb-0 catalyst was performed by Al-Sherehy et aZ.24for a fixed bed reactor of 20 m length (di,=26.6 cm) at P = 2 0 atm and T=540 K using kinetics reported by Thorsteinsom et al.23 The authors showed that ethylene yield of 10.8% with a selectivity of 73.6% can be obtained under these conditions. By applying oxygen feed distribution with an injection every 1 m an increase of ethylene yield to 17.4%with increased selectivity of 93.7%can be achieved. A distributed oxygen feed can also overcome the runaway problem as shown by the simulation of temperature profiles. A hydrothermally synthesized Mo6V2Te10, catalyst reported recently by Ueda and Oshihara also revealed activity at the low temperature of 613 K resulting in an ethylene yield of 13% (S = 72.5%).25 Potassium salts of 12-molybdophosphoric acid (Keggin-type heteropolycompounds) were tested as ODE catalysts by Cavani et al.26A high selectivity of 91% corresponding to an ethylene yield of 13.7% was obtained for a K2P1.2M010W1SblFelCro.5Ce0.750n catalyst at 743 K, showing a stable performance for 100 h. Le Bars et al. reported ethylene selectivity of 60% at ethane conversions of 28% over a V205(5wt%)ly-A1203catalyst at 823 K.l0 The improved catalytic properties of vanadialy-alumina catalysts in ODE, as compared to vanadial silica and bulk vanadium pentoxide, was related to the dispersion and coordination of the vanadium cations assumed as active sites and only to a lesser degree to the development of their acidic properties. The active sites of the V205/y-A1203catalysts were described as dimers of two tetrahedrally coordinated and oxygen-bridged vanadium cations which are reduced by

9 30 35 9 9 151 131 30 21 15 15 46 30 30 45.5 50 12 30 70.5 61.7 56.7 56.7 56.7 6 10 16 27 18 182 101 15 16.3 8.5 8.5 6.7 15 15 18.2 25 6 6 41.5 36.2 28.3 28.3 28.3 623 613 743 823 863 873 923 953 973 873 873 1123 903 913 140 213 873 923 193 198 143 103 1143

Catalytic results on the oxidative dehydrogenation of ethane (ODE)

M00.73VO.1 8 ~ 0 . 0 9 0 ~ MosV2Te10x K2p1.2Mo10W1SblFelCr0.5Ce0.750n v205(5 ~t%)/y-A1203 VOx/a-Ti phosphate NiMo04 (a-phase) BaC03 + Ba3(V04)2 SrFe02.81 SrCeO.5YbO.5 0 2 . 7 5 LdCaO Li/La/CaO SrO/La203 H0203 30 mol% BaON205 SmN%,028Po.o4o(autothermal) Sro, Ndl La (autothermal) Li (3 wt%)/MgO Nao.064CaOx 2.3 wtY0 Pt monolith Sn-Pt (7: 1) monolith Pt/a-A1203(fluidized bed) a-A1203(fluidized bed) Agla-AlzO3 (fluidized bed)

Catalyst

Table 2

0.12" 0.12" 6 0.6" 1.4 3.6 3.6 0. 1" 0.6" 0.6" 0.33" 0.45" 12 0.2" 4.9 5b 1Sb 1Sb 1 .Sb

-

10.6 0.6* 5.8 58 18.2 15 28 17 30.7 34 54.8 70 48.8 47.4 68 40.5 62.9 67 79 39.3 13.4 82 87 90 83 90

65 72.5 91 60 70 64.7 76.5 50.5 70 75.0 93.8 82 47.8 68.1 68 71 74.8 78.1 70 66 67 67 67 56 19.4 42.8 46 56 29.4 10.5 57 57 60 55.6 60.3

44.5

37.7 13.2 13.7 16.8 12 19.8 26 27.7 49 36.6

23 25 26 10 27 28 29 21 30 31 31 32 41 18 33 35 12 37 15 16 38 38 39

m 0

c

161

5: Oxidative Functionalization of Ethane and Propane

ethane more easily than tetrahedral vanadate monomers and octahedrally coordinated vanadium cations of bulk vanadium pentoxide. From pulsing ethane over the catalyst it was estimated that the reduction degree of 3.8 oxygen atoms per vanadium atom on the vanaddalumina catalyst is higher than the theoretical reduction degree corresponding to the reduction of V5+ cations to V4' cations (0.5 oxygen atom per vanadium atom) and to V3+ cations (1.0 oxygen atom per vanadium atom) as a result of the intrinsic activity of y-alumina. After 25 pulses of ethane, 4% of lattice oxygen atoms from y-alumina have been consumed.lo A new type of supported vanadia ODE catalysts in which the vanadium oxide was incorporated in a layered a-Ti phosphate substrate was described by Santamaria-Gonzalez et al.27 The results of in situ Raman spectroscopy revealed that the incorporated vanadium oxide interacts under reaction conditions with the phosphate layer forming an a-VOP04 phase being more active than dispersed V2O5 species. An ethylene yield achieved on this catalyst type amounted to 12% (S= 70%) at 863 K. A Ni molybdate phase examined by Kaddouri et al. resulted in 20% yield (S=65%) at 873 K.28 Reaction temperatures above 920 K were required to achieve ethylene yields of 26-28% when combinations of transition and alkali earth metal oxides (Ba-V-0, Sr-Fe-0) were used as catalyst^.^^.^' Thus, the ODE results reported for catalysts consisting of reducible metal oxides showed that, except in the work of Thorsteinson et al.,23ethylene yields less than 30% were obtained using this catalyst type. 3.1.2 Non-reducible (Mixed) Oxides. To obtain high ethylene yields on this type of catalyst without being doped with halides reaction temperatures 2 873 K are required. A significant portion of catalysts from this group is based on the rare-earth oxides (REO). An ethylene yield of approximately 49% (S=70%) was obtained by Velle et al. on Srl.0Ce0.5Yb0.502.75 at 973 K.30 Lithium-doped lanthanum-calcium oxide (Li:La:Ca = 1:1:2.5) was described as an effective catalyst resulting in C2H4 yields of 48.2% (S= 89.2%) at 893 K and GHSV= 1000 h-1.31Using a Ba-Y-0 system Dai et al. obtained a yield of 43%.18The dehydrogenation of ethane was studied by Choudhary et al. at 923 to 1123 K using a diluted SrO/La203catalyst; ethane conversions of approximately 20% (SC2H4 = 75%) and 68% (SC2H4 = 82%) were obtained at the low and high temperature, re~pectively.~~ The authors propose that the reaction was initiated on the catalyst; however, no results are given on ethane oxidation in the absence of a catalyst under the same conditions. This is an important issue, since at these high temperatures, particularly at 1123 K, non-catalytic reactions prevail. As shown by different authors, a significant non-catalytic oxidative dehydrogenation occurs above 873 K. In ODE on REO-based catalysts, e.g. Ln203/Sr0 (Ln=Sm, La, Nd) and Na-P/Sm203, ignition of the reaction mixture occurred in continuous-flow operation using undiluted catalyst particles at fuel-rich condition^.^^-^^ After ignition a sharp temperature increase occurred at the beginning of the catalyst layer; the temperature was then mostly well above 873 K, which is the border 3~14933

162

Figure 2

Catalysis

Temperature projles in an autothermal reactor after ignition of the reaction mixture at a preheating temperature of 773 K on a Sro.ILal.oNdl.~O, catalyst for different space velocities (C2H6/021N2 = 21111) (Reproduced by permission from J. Catal., 2001,197,43)

between catalytic and homogeneous gas-phase oxidation. A characteristic temperature profile observed in the reactor is presented in Figure 2. The maximum temperature measured in the catalyst bed was 100-300" higher than that of preheated feed gas. Oxygen conversions amounted to >98-99%. Reaction products were C1-C4 hydrocarbons and COX;no oxygenates were detected. It should be noted that the observed temperatures are strongly determined by the reactor configuration and conditions applied; the pattern, however, is characteristic for this type of catalyst. Ignition temperatures amounted to 723-753 K for Sm(La, La-Nd)-Sr-0 catalysts and to 973-983 K for SmNao.028Po.ol40, catalysts. Ethylene yields of 44-46% at conversions above 50% were obtained at short contact times of the order of 30-40 ms. An increase of temperature in the catalyst layer was achieved by increasing the preheat temperature of the feed gas or by increasing oxygen concentration resulting in increasing ethane conversion. Since the oxygen conversions amounted to 95-96% at the lowest temperature, an increase in ethane conversion with rising temperature was mostly due to thermal pyrolysis. An increase of ethylene yield was obtained without significant loss in selectivity. The performance of Sr,Lal.oNdl.oOy catalysts (x = 0, 0.001, 0.01, 0.1 and

5: Oxidative Functionalization of Ethane and Propane

163

1.O) was studied for their potential application as ignitors in the autothermal oxidative dehydrogenation of ethane.35 Ethylene formation occurred as a result of catalytic ignition of the reaction mixture and of the subsequent thermal pyrolysis of ethane due to the significant temperature increase; all catalysts showed comparable performance under ignition conditions. The optimization of the reaction conditions was carried out on a Sro.lLal.oNdl.oO, catalyst by varying space velocity (1 x lo4 - 8 x lo4 h- l), preheat temperature of the feed gas (373-873 K), and ethane-to-oxygen ratio (2-5) in the absence and presence of 15 vol.% steam in the feed. Space velocity was a significant factor determining the maximal temperature in the ignition zone and hence the extent of thermal pyrolysis. The highest ethylene yield amounting to 56% ( S = 71%) was achieved at GHSV = 8 x lo4 h-' ( ~ ~ ~ ~ms), - 4C2Hd02/ 5 N2 = 2/1/1 and a preheat temperature of 873 K corresponding to a maximum reaction temperature of 1213 K. The addition of steam resulted in some decrease of the maximum temperature leading to a decrease in the ethylene yield by 2%. No significant effect of the steam addition on product distribution was observed. Another group of non-reducible ODE catalysts is alkali-earth oxides doped with alkali oxides. Morales and Lunsford reported on an ethylene yield of 29% ( S = 75%) over a Li/MgO catalyst.12 The authors proposed that reaction occurs via the formation of gas-phase ethyl radicals. The use of a fixed bed of Li/MgO catalyst encompassed by a porous ceramic membrane has been investigated in ODE by Coronas et al.36 Using this system propane conversions above 40% with ethylene selectivities of 80-84% were obtained at reaction temperatures >923 K. Oxygen distribution along the catalyst bed due to membrane did not, however, improve the ethylene selectivity compared to the fixed bed reactor. Recently Kondratenko et al. studied ODE over NazO/CaO catalysts with different Na20 contents.37 Particular attention was given to the oxygen interaction with catalytic surfaces which was investigated in the TAP reactor and by means of contact potential difference (CPD) measurements. Modelling of TAP-oxygen responses in the temperature range 673-873 K showed that a reversible dissociative adsorption via a molecular precursor provides an accurate description of the measured transient responses for all catalysts. CPD results confirmed the transformation of the molecular adsorbed oxygen species to the atomic oxygen. Rate constants and activation energies of the elementary reaction steps of oxygen adsorption, desorption, dissociation and association were estimated from the TAP oxygen responses. Based on these kinetic parameters the steady-state surface coverages with molecular ( 0 0 2 ) and atomic oxygen species (00) for different oxygen partial pressures and temperatures were simulated and used to explain the catalytic data. It was found that changes in C2H4 selectivity with Na doping were mostly determined by the ratio of 0d OO2as illustrated in Figure 3. 3.1.3 Catalytic Materials Containing Noble Metals. The synthesis of ethylene from ethane in the presence of oxygen is carried out at high temperatures

Catalysis

164

50'

0.0

*

' ' 0.1 0.2 I

'

I

0.3

s

I

0.4

'

0.5

ratio of @@ /, 0 2

Figure 3

Selectivity of ethyleneformation over Na2OlCaO catalysts in the ODE versus steady-state ration of as simulated for different oxygen partial pressures in the absence of ethane (Reproduced by permission from J. Mol. Catal. A: Chem., 2000,158, 199)

above 1073 K and contact times in the range of milliseconds using noble metal The containing catalysts, was proposed by Schmidt and co-workers. best results achieved on this type of catalyst which was applied in the form of a coated monolith or as a fluidized bed of A1203-supportedcatalyst are given in Table 2. Schmidt et al. reported ethylene yields of 53-57% (S=66-70%) at contact times in the order of milliseconds over Pt- and Pt-Sn monoliths at ca. 1193 K in the autothermal m ~ d e . ' The ~ ' ~heat ~ which is set free after ignition was used to sustain the further reactions. Heterogeneous ignition was achieved by heating reactants up to about 500 K. The catalytic materials were deposited on ceramic a-A1203 monolith disks ( D = 17 mm, L = 10 mm) by impregnation with Pt, Pt-Me (Me=Cu, Sn, Ag, Mg, Ce, Ni, La, Co, Au), Rh and Ag. The reaction temperatures depended on the flow rates and on the C2Hd02-ratio and were for Pt loaded monoliths 1143, 1193 and 1273 K for C2Hd02 = 2, 1.7 and 1.4, respectively. The flow rates varied from 2 to 12 lsTp min (13-79 cm s- l ) and the overall pressure was set to 1.4 bar. The highest selectivities (70%) and conversions (80%) were found for a Pt monolith at a C2Hd02 ratio of about 1.7 (20% N2 and 4.5 lsTp min-').15 Although very high reaction temperatures were used, the influence of the gas-phase reaction in the formation of olefins was considered negligible and a heterogeneous mechanism was proposed.15 In a recent paper, Huff et al. analysed the ODE reaction over Ptcontaining monoliths including the contribution of gas-phase reactions.40 The authors performed the simulation using kinetic parameters for the gas-phase 15~16938939

5: Oxidative Functionalization of Ethane and Propane

165

400 n L

t cn E=I 3

o! a

300.

200.

(P

. I

5 P

100.

0. 0

Figure 4

3 6 9 Time (milliseconds)

12

Calculated partial pressures for non-catalytic gas-phase ODE as a function of contact time (2 1STp min-', 1.2 atm, C2HslO2 = 1.2, 20% N2 dilution, inlet temperature of 1073 K ) (Reproduced by permission from J. Catal., 2000,191,46)

and surface reactions reported in the literature. The results of such a simulation considering only gas-phase reactions along the monolith channels are presented in Figure 4. The reactants are almost completely converted in about 6 ms and no significant reaction takes place beyond about 9 ms. Results of simulation including both surface and gas-phase reactions showed good agreement between experimental and calculated conversions and selectivities. A critical part of the analysis of Huff and co-workers accounts for the heat effects and the resulting reactor-temperature gradient.40 Similarly to earlier findings of Buyevskaya and Baerns for REO-based catalysts,33the authors concluded that the main features of ODE on platinum-containing monoliths can be understood in terms of a sequential process in which exothermic surface oxidation reactions are followed by a combination of endothermic and exothermic gasphase reactions. The rapid rise in temperature along the monolith leads to the nearly exclusive formation of ethylene (and acetylene) via gas-phase homogeneous reactions. The role of the catalyst is therefore the initiation of reactions that supply the heat for the subsequent gas-phase reactions. The ODE reaction over a-A1203 (85-1OOpm) impregnated with Pt, Ag, Ni, Pd, Pt-Au and Pt/Zr02was examined by Bharadwaj and Schmidt in a fluidized bed r e a ~ t o r .In~ comparison ~,~~ to a monolith reactor, 6-8% higher selectivities for ethylene at equal conversions were found in the fluidized bed reactors. The main difference between these two reactors is the transport of heat. Due to the circulation of the catalyst particles in the fluidized bed nearly isothermal conditions were achieved. In the monoliths the heat transfer occurs upstream only via heat conduction and radiation.

166

Catalysis

3.2 Ethane Dehydrogenation in the Presence of Cot. - Utilization of carbon dioxide in the dehydrogenation of ethane to ethylene has been reported in some papers. Nakagawa et aZ,42studied the dehydrogenation of ethane in the presence of carbon dioxide over several metal oxides and found that gallium oxide is effective, resulting in an ethylene yield of 18.6% ( S = 95%). Liu et al. reported ethylene yields of 52% ( S = 97%) and 63% ( S = 91%) at 1073 and 1123 K, respectively over a Na2W04-Mn/Si02 catalyst using a C2H6:C02= 1:l mixture.43 Byproducts CO and CH4 were formed with a selectivity of 6%. The authors speculated that the reaction occurs via dissociation of C02 on the catalyst surface forming active oxygen species which are responsible for the ethane activation and dehydrogenation to ethylene. No comparative experiments were carried out in the absence of C02 at similar conditions. Thus, it is not possible to estimate the involvement of nonoxidative ethane dehydrogenation which occurs to a significant extent at 1073-1 123 K as applied in this study. Xu et aZ. performed the ethane dehydrogenation with C02 over a chromia supported on Silicalite-2 zeolite and doped with Mn and La oxides.44Ethylene yield of 55% (S=86%) was obtained at 1073 K. Contrary to Liu et al.43the authors proposed that the main pathways consist of two consecutive reactions: C2H6 + C2H4 + H2

(1)

CO2 + H2 + CO + H20

(2)

Ethane is first dehydrogenated to ethylene; then, the hydrogen formed reacts with C02 in reverse water-gas shift reaction to CO and water. Thus, hydrogen is removed from the reactants, enhancing ethane conversion. For chromium oxide supported on sulfated silica,Wang et al.45reported an ethylene yield of 55% ( S = 87%) at 923 K using a feed gas consisting of 10% C2&, 50% C02 and 40% N2 Stability test of ca. 6 h showed that the catalyst deactivates. The main reactions occurring when using C 0 2 were proposed to be as follows: C2H6 + C2H4 + H2

(3)

C2H6 + C02 -+ C2H4 + CO + H20

(4)

C2H6 +2C02 + 4CO + 3H2

(5)

C02 + H2 -+ CO + H20

(6) The authors assumed that C02 acts as a mild oxidant according to Eq. 4 for ethylene formation in addition to the direct ethane dehydrogenation (Eq. 3). One can, however, assume that ethylene is produced only by dehydrogenation (Eq. 3). Supported chromium oxides are known as commercial dehydrogenation catalysts. C02 is involved in shifting the equilibrium of reaction (1) further towards ethylene due to consumption of H2 by C02 according to Eq. 6, as also proposed by Xu et aZ.44 The carbon deposits formed from complete ethane dehydrogenation can react with C02 in the reverse Boudourd reaction,

5: Oxidative Functionalization of Ethane and Propane

Table 3

167

Best ethylene yields and selectivities for different processes of ethane dehydrogenation

React ionlcatalyst

Autothermal ODE on REO-based catalyst^^^^^^ Autothermal ODE on Pt-Monolith16 Thermal pyrolysis of ethane136(commercial process) Non-catalytic ODE33 ~~

a coil

1 140- 1213 1193-1 198 923-953"

46-56 45-57 55-57

68-71 65-66 85-88

1076

45

66

~~

inlet temperature.

which occurs to a significant extent above 873 K. The latter reaction contributes to slow catalyst deactivation compared to the absence of C 0 2 .The authors did not give any comparison of the initial behaviour of ethane dehydrogenation with and without COz, which might supply important information with respect to the reaction pathways and the role of CO2 in the ethane dehydrogenation reaction. 3.3 Summary on ODE. - For the ODE reaction, different reactor concepts and modes of operation have been examined. Autothermal processing is attractive due to its high energy efficiency: the heat generated by reaction is used to sustain thermal reactions after ignition and oxygen consumption.'6,33-35The best results reported for autothermal ODE are given in Table 3 along with data for the non-catalytic thermal and oxidative pyrolysis. For all types of operation, similar ethylene yields were achieved. Oxidative pyrolysis and autothermal ODE resulted, however, in 15-20% lower selectivities than thermal pyrolysis. Although similar ethylene yields can be obtained in the oxidative pyrolysis and autothermal ODE, no ignition can be achieved below 973 K in the absence of catalyst.33 The ignition of the reaction is an important catalyst function. The use of catalysts (noble metal, REO-based) allows the operation in an energy efficient manner. After ignition, reaction temperatures can be sustained at the desired level without significant external heat input. The contact times for both types of catalysts (Pt-containing vs. REO-based) are in the range of milliseconds requiring only a small volume of the autothermal reactor. The advantage of RE0 is related to the high thermal stability of those materials. Thus, the effect of sintering which takes place when using supported Pt, e.g. in the form of a Pt-monolith at high temperatures, can be avoided in long-term operation. In general, autothermal operation using different catalysts looks promising for the production of ethylene from ethane, From an economic point of view yields of at least 70% are required; thus improvements are still necessary.

Catalysis

168

4

Oxidative Dehydrogenation of Propane

There are some well established processes (e.g. Star, Catofin, Oleflex, LindeBASF) for the production of propene from propane by catalytic non-oxidative dehydrogenation over supported Pt or chromia-alumina catalyst^.^^ Propene yields of 27-35% with high selectivities of 80-90% can be achieved operating at 1.1-8 bar. At subatmospheric operation (33-50 kPa, Catofin process) the yield can be increased up to 53% ( S = 82%). The advantages of these processes are high propene selectivities and a co-production of hydrogen. The disadvantages are related to the requirement of frequent catalyst regeneration due to coke formation, thermodynamically limited propene yield and high endothermicity. Due to the latter factors the development of direct oxidative dehydrogenation of propane (ODP) is of high interest. A large number of catalysts of different types have been investigated but, contrary to the ODE reaction, propene yields in the ODP reaction did not in general exceed 10-20%. Also more efforts compared to the ODE were devoted to understanding the factors determining the selective pathway of the ODP reaction as illustrated by numerous mechanistic and characterization studies. Similar to ODE, most of the ODP catalysts reported in the scientific literature can be divided in three groups: (A) mixed oxides or supported catalytic materials consisting of reducible metal oxides; (B) non-reducible (mixed) oxides; ( C ) catalytic materials consisting of noble metals. Contrary to ODE, a major part of the ODP catalysts belongs to group A. These are catalysts which mostly consist of transition metal oxides and operate by a redox-mechanism. Among the group B catalysts, there are materials on which oxygen participates in propane activation via its adsorbed state (cf. Figure 1, case B). Almost no attention has been paid in the literature on ODP to this case, which is mostly associated with REO-based catalysts, although this type of alkane activation was proposed for the oxidative coupling of methane and for the oxidative dehydrogenation of ethane. Another example of the non-reducible ODP catalysts are boron oxide-containing oxide materials on which propane activation via formation of propyl radicals takes place with the participation of strongly bound lattice oxygen and further reactions occur in the gas phase (cf Figure 1, case C). Recent publications dealing with the development of ODP catalysts, mechanistic investigations as well as reaction engineering aspects are described below. Catalytic data over different ODP catalysts are given in Table 4. 4.1 Mixed Oxides or Supported Catalytic Materials Consisting of Reducible Metal Oxides. - 4.I . I CataZyst Development. Among catalytic materials of group A, the following show the best catalytic performance: vanadium oxides on different supports, magnesium vanadates and Co and Ni molyb-

19 6 6 27 5 15 4 30 40 40 54 40 10

M&.92M00x V205/A1@3 KN205/A1203 lO%(K/Mo = 0.07)/Si02+Ti02

a

s; SZ - sulfated zirconia.

P-NiMoOd2.O%Ca K-Mo/MgO-y-A1203 V0.22Mg0.47MOO.1 1G%.20°x VOX(2.8wt.%)/MCM-48 SrlLaSO, Pt/y-A1203 B2O3 (30 wtYo)/A1203 Ni-Li/SZb

pC3H8 kPa

7 6 6 13 2.5 18 4 10 20 20 23 20 10

kPa

PO2

773 880 859 823 823 823 823 773 773 953 1141 823 873

T K

0.08 0.06 0.003" 2 0.5

4.4 0.02 0.02 1.3" 5a 0.1-0.36 0.7

gsml-'

'I:

Catalytic results on the oxidative dehydrogenation of propane (ODP)

Catalyst

Table 4

-

17.4 70.9 41 32 50 13 26.4 15.7 32.8 53.5 57 42 20.8

%

X(C&)

-

58.0 25.4 39.3 66 66 84.2 44.7 57.4 53.0 20.3 23 40 52.1

%

s(c&) 10 18 16 21 30 11 12 9.0 17 11 13 17 11

YO

Y(C&)

62 61 6 7 82 88 84 63

76 59 59 60

Refs.

42 9ra

@a

g'

!2

s* $

5-. % 2-

P

?

Catalysis

170

dates. V-Mg-0 catalysts have been studied most extensively. These catalysts have various pure phases (e.g. Mg2V208, Mg2V207, MgV206), as well as their mixtures, depending on the V-to-Mg ratio. A maximum propene yield reported over a V-Mg-0 catalyst is 23.9% (S=38%).52 The yields obtained over molybdates of various transition metals were generally not higher than 15%. Several papers described the use of magnesium molybdates as catalysts for the ODP r e a ~ t i o n . ~The ~ - excess ~ ~ of molybdenum relative to MgMo04 was found to be important for improved performance. Special attention has also been given to Mg-V-Sb-oxide catalysts; the composition Mg4V2Sb20, was identified as the best catalytic material among various alternative~.~~ Some recent publications dealing with the development of catalysts consisting of transition metal oxides which appeared since 1998 are described below in more detail; the respective catalytic data are given in Table 4. Viparelli et al. studied ODP over niobium and vanadium oxides supported on high surface area Ti02.58 The most selective catalyst was Nb/Ti02. Addition of vanadia led to an increase of activity. It was concluded that the interaction between vanadium and niobium modifies the surface acidity: the redox properties were, however, less affected by this interaction. Among different samples, the highest propene yield of 7.7% ( S = 36.6%) was obtained over a niobia-titania catalyst. It is generally known that alkali doping has a beneficial effect on selectivity. ODP over alumina-supported vanadia doped with potassium and manganese was studied by Ermini et al.59 It was found that K-doping poisons the acid sites of the support resulting in increasing propene selectivity coinciding with a loss of activity while Mn-doping had no positive effect on both activity and selectivity. A maximum propene yield of 18% ( S = 25.4%) was obtained over undoped V205(17.2%)/A1203catalyst. On I(N2O5/Al2O3,the highest propene yield amounted to 16% ( S = 39’%0).~~ Watson and Ozkan have examined ODP over K/Mo catalysts supported on SilTi mixed oxides prepared by a ‘one-pot’ sol-gellcoprecipitationtechnique on which high propene yields (up to 30%) were obtained at K/Mo ~ 0 . using 6 diluted feed mixtures (5% propane).60 With increasing propane concentration in the feed gas (up to 26%), the maximum propene yield was 21%. The catalysts revealed well-dispersed Mo species on the surface either in an octahedral molybdenum oxide matrix or in a tetrahedral K-molybdate matrix. The promotion effect of potassium combined with the unique state of titania in the Si/Ti mixed oxide support were used to explain the high yields. The influence of potassium and samarium on molybdenum supported on MgOly-Al203 was studied by Abello et al. The highest propene yield amounted to 11.8% ( S = 44.7%) over a K-doped catalyst at 823 K.61 The propene selectivity of nickel molybdates in the ODP reaction was found to be improved by Ca addition. On P-NiMo04/2.0%Ca, a propene yield of 11% was achieved with a selectivity of 84%.62 Nickel supported on sulfated zirconia (SZ) was tested in the ODP reaction by Wang et al.63 A maximum propene yield of 10.8% ( S = 52.1%) was achieved over a Ni-Li/SZ catalyst. 174947-51

5: Oxidative Functionalization of Ethane and Propane

171

In our own work, a combinatorial and evolutionary approach was applied for development of ODP catalysts consisting of transition metal oxides.6@The application of combinatorial and high throughput screening methods in heterogeneous catalysis is a new and rapidly developing area (see e.g. references in refs. 6, 64).Due to the large number of catalytic formulations to be screened, the concepts for the creation of libraries of catalytic compositions are of high importance, in particular for multicomponent formulations. We showed that optimization methods comprising evolutionary algorithms appear promising for finding catalytic compositions from a broad basis of single compounds considered to be important for the target reaction. The method has been described in detail e l ~ e w h e r e . The ~ , ~ ~procedure supported by a software code uses the following steps: 1st Step: Initialization of the 1st generation of catalytic materials in a stochastic manner based on pre-defined primary components; 2nd Step: Parallel preparation and testing of catalytic materials of the 1st generation, evaluation of catalyst quality; 3rd Step: Creation of the 2nd generation based on the catalytic results of the 1st generation using mutation and cross-over methods, preparation and testing catalytic materials; Subsequent steps: Repetition of 2nd and 3rd steps for the next generations.

For the proof of principle of the combinatorial and evolutionary methodology, the following compositions: 'A,,- 2Ax,- 3Ax,- . . . - n - 1 A,-,- nAxn- '0, where 'A is, e.g., one of the elements V, Mg, B, Mo, La, Mn, Fe, Ga and xiis the stoichiometric coefficient ( e x i = l), were prepared and tested for the ODP reaction. Four generations wkre created with the total number of tested catalytic materials being 224. As a result of the optimization procedure an increase in the propene yield was achieved with an increasing number of generations (see Figure 5). Most of the combinations showing good performance consisted of V, Mg, Mo, Ga or V, Mg, Ga only. The highest C3H6 yield amounted to 9.0% ( S = 57.4%) in the 3rd generation on V0.22Mg0.47M00.11Gao.200~.In the 4th generation, the mean propene yield increased but no improvement of the best propene yield was obtained. The results can be certainly improved by screening further generations of catalytic materials. Some of the best catalytic materials which were developed by this method as well as additional materials consisting of V, Mg and Ga were characterized by XRD, XPS and EPR.7 A correlation between catalytic performance and the M g N ratio on the surface was found. An increase in propene yield was achieved by increasing the M g N ratio. A maximum propene yield of 13.8% (S=57.5%) was obtained on a Mg-V-0 sample revealing a MgN ratio of 9.3 in the near surface layers. From fundamental study it was concluded that a good catalytic performance for ODP is achieved by an optimal concentration of the active VO, species which

Catalysis

172

'

2

3 ' 4

catalyst ranking

Figure 5

9

10'

Best propene yields achieved in each generation in the order of decreasing catalyst quality6 (Reproduced by permission from Appl. Catal., 2000,200,63)

have to be well dispersed on the surface. Based on the structural knowledge obtained, VOJ2.8 wt.%)/MCM-48 and VOX(2.8wt.%)/MCM-41 catalysts with a high dispersion of vanadia were tested giving a maximum propene yield of 17.4% (S= 53.0%) and 14.9% ( S = 49. YO), re~pectively.~ 4.1.2 Mechanistic and Characterization Studies. Numerous studies were undertaken for the ODP reaction which aimed at understanding the nature of active sites and the factors determining the selective pathway. Most of these studies dealt with V-containing oxide materials.5*47,66967 Some selected aspects of recent work on the reaction mechanisms over different catalysts are reviewed below. Khodakov et al. examined the effects of different supports (Al2O3, Si02, Hf02, Ti02, and 25-02)on the structure and catalytic behaviour of supported vanadia in the ODP reaction. over a wide range of vanadium surface densities (0.5- 15.0 VOx/nm2).68They found that support surfaces predominantly covered with polyvanadate structures or small V2O5 clusters containing V-0-V or V=O linkages lead to high oxidative dehydrogenation rates and selectivities. Based on the analysis of the rate coefficients it was concluded that sites required for oxidative dehydrogenation also catalyse the undesired combustion of propene. Chen et al. measured kinetic isotopic effects (KIEs) over a V205(10wt. YO)/ Zr02 catalyst using undeuterated (CH3CH2CH3), perdeuterated (CD3CD2CD3) and selectively deuterated propane (CH3CD2CH3)m o l e ~ u l e s . ~ ~ Their aim was to identify the kinetic significance of C-H bond activation steps and the position of the C-H involved in the oxidative dehydrogenation reaction. It was found that C-H bond dissociation is a kinetically relevant step for propane primary dehydrogenation and combustion as well as for consecu-

5: Oxidative Functionalization of Ethane and Propane

Figure 6

173

Proposed reaction pathway for propane reaction over VOJZrO, catalysts (Reproduced by permission from J. Catal., 2000,192, 197)

tive propene combustion. It was confirmed that the methylene C-H bond is activated in the rate-determining steps for propane dehydrogenation and combustion. The consecutive propene combustion involves activation of an allylic C-H bond. The KIE for propane oxidative dehydrogenation (2.8) being smaller than the maximum KIE (6) expected for propane thermal dehydrogenation indicated the participation of lattice oxygen. The difference in the KIE value for propane primary dehydrogenation and combustion reactions reflects that different lattice oxygen species are involved in these two reactions. The pathways proposed in this work for the activation of propane and propene are summarized in Figure 6. Creaser et al. investigated the kinetics and the mechanism of ODP over a V-Mg-0 catalyst using various transient technique^.^^ Their results supported a redox mechanism in which propane and intermediate products react with lattice oxygen, reducing the catalyst surface, which is reoxidized by gas-phase oxygen. In non-steady-state operation, propene selectivity was superior in the absence of gas-phase oxygen as compared to steady-state operation when

174

Catalysis

oxygen and propane were fed simultaneously. Successive pulsing of C3H8 reduced V5+ in the magnesium orthovanadate phase to V2+. The authors showed that, besides adsorbed oxygen being an important source of total combustion, lattice oxygen also resulted in total oxidation products down to levels of reduction of 50%. Furthermore, the authors investigated the formation of carbon-containing species. Upon interruption of the feed, reactive carbon species remaining on the catalyst were oxidized by a succeeding oxygen pulse. Temperature-programmed oxidation (TPO) experiments showed that the catalyst used in steady-state operation revealed mainly strongly bound carbon deposition which did not, however, affect catalyst activity. An important point regarding the formation of active sites is related to reconstruction phenomena; hereby intermediate phases may be stabilized which determine the catalytic behaviour. For the ODP reaction over V-Mg-0, Pantazidis et al. showed that the presence of a continuous and reversible phenomenon of ordeddisorder of a vanadium overlayer plays a key role in the catalytic perf~rmance.~~ Structural characterization of a V-Mg-0 catalyst was carried out by Burrow et al. using in situ EXAFS, in situ XRD and HREM.72 The authors showed that an ‘as prepared’ catalyst consisted of magnesium orthovanadate (Mg3V205) particles, magnesium oxide and a disordered vanadium-containing overlayer deposited on the MgO. When the catalyst was exposed to typical ODP reaction conditions at 773 K a change in vanadium oxidation state from 5+ to 3+ occurred and Mg3V205 transformed to a cubic spinel type structure with a lattice parameter of 8.42 These changes are reversible on exposure to air at 773 K. HREM showed that the overlayer on MgO changes from a disordered state to a weakly ordered structure under exposure to normal reaction conditions while pure propane (strongly reducing conditions) induced pronounced structural ordering of the overlayer, forming a cubic spinel (MgV204) phase in parallel epitaxy with the MgO support. The near-surface regions of bulk Mg3V205 decomposed under reaction conditions to a mixture of MgO crystallites and MgV204; strong reduction caused a complete conversion to MgV204.72 The activity of the framework and extraframework vanadium species in the V-containing aluminophosphate VAPO-5 was studied by Okamoto et al. 73 Results of X-ray diffraction, electron spin resonance spectroscopy and temperature-programmed reduction together with catalytic data confirmed that that framework vanadium is the most active and selective species for propene formation. For a V2041 wt.0/)/Ti02 catalyst consisting of VO, species attached to the Ti02 surface, transient experiments in vacuum at 766 K using the TemporalAnalysis-of-Products (TAP) reactor system confirmed the reaction of propane with lattice oxygen.74 Pulsing propane over the oxidized surface resulted in propane conversion of 36%; propene selectivity was 27%. Selectivity was, however, improved with an increasing degree of surface reduction. A propene selectivity of 67% (&3,8=7%) was achieved at a relatively high degree of surface reduction (0= 0.48 where 0 is the ratio of catalyst oxygen removed by pulsing propane to the total amount of vanadium atoms). The degree of

A.

5: Oxidative Functionalization of Ethane and Propane

175

surface reduction, which is affected in steady state by the propane-to-oxygen ratio, is a key factor determining both propane conversion and propene selectivity in the ODP over a V2Os/TiO2 catalyst. For magnesium molybdate catalysts, an excess of molybdenum relative to MgMo04 was found to be important for improved performance. Different explanations of this phenomenon were proposed. Cadus and co-workers detected only MgMo04 and Moo3 phases; they concluded that excess molybdena leads to a synergistic cooperation between these phases.54 On the contrary, Lee et al. concluded that the benefit of excess molybdenum is due to active MOO, clusters formed on the surface of M ~ M o OTheir ~ . ~conclusions ~ were based on experiments in which molybdena supported on inactive magnesium molybdates was further modified by treating the surface with acidic and basic solutions. Two more recent publications deal with the formation of active species in magnesium molybdates catalyst^.^^.^^ Miller et al.75 examined pure and mixed molybdenum-containing phases (Moo3, MgMo04, MgMo207) and concluded that excess molybdenum (as MoQ3 or MgMo207) combined with MgMo04 resulted in changed redox properties leading to improved catalytic performance. All of the active and selective catalysts tested in this study consisted of P-MgMo04 and another crystalline phase which could not be identified by XRD. A strong correlation was found between catalyst performance and TPR results. Each of the active and selective catalysts has a reduction peak centered at about 823 K which was not present for inactive materials, including P-MgMo04. The authors suggested that the active catalysts share a common phase or species which is not detected by X-ray diffraction. Magnesium molybdates prepared from an aqueous solution of magnesium nitrate and ammonium paramolybdate at various pH values (pH = 1.5-6.6) revealed after calcinations below 823 K the formation of three phases, a-MgMo04, P-MgMo04 and Mg2M03011,as was shown Yoon et aZ.76 The catalyst prepared at pH = 5.7 showed the highest activity for the oxidative dehydrogenation of the alkane as well as the strongest acidity. By XPS measurements, an excess amount of Mo compared to Mg was observed over the active catalysts. The authors assumed that the excess molybdenum species is present as molybdate and creates acidic sites over the catalyst surface. Concerning the effect of alkali doping there are some findings in the literature which confirm that the reduction behaviour as well as acidity depends on the extent of alkali doping. For K-Mo-O supported on SdTi mixed oxides, a shift of the reduction maxima to higher temperatures was observed by Watson and Ozkan; the authors assumed that potassium suppresses the reducibility of the molybdenum species at low WMo ratios and stabilizes the Mo(V) oxidation state.60 ODP reaction network and the respective kinetics were studied by Grasselli and c ~ - w o r k e r s . For ~ ~ -a~Nio~5C00~sMo0q/Si02 ~ catalyst, propene is the sole primary product and the abstraction of a methylene hydrogen from propane is the rate limiting step.77Kinetic studies over a Mg4VZSb0, catalyst revealed that propane is oxidized in parallel reactions to propene, CO and C02. All three reactions exhibited Langmuir-type dependencies on propane concentration.

176

Catalysis

Propene was then oxidized in parallel to acrolein, CO and C02. The partial oxidation of both propane and propene are zero-order, whereas deep oxidations are half-order in oxygen. Two different oxygen species, i.e. lattice and adsorbed oxygen, were proposed to take part in partial and deep ~ x i d a t i o n . ~ ~ , ~ ~ Thus, from above studies it can be concluded that the understanding of ODP catalysis has certainly advanced; nevertheless, the suppression of nonselective reaction steps depending on the different oxygen species or sites respectively needs still to be resolved. The development of a stable catalyst leading to propene yields higher than 20% at high propene selectivities and near-to-practice conditions is still a challenging task.

4.2 Non-reducible (Mixed) Oxides. - 4.2.I REO-based Catalysts. Contrary to the ODE only few studies dealt with the ODP using REO-based catalysts. Zhang et al. studied fluorine promoted oxides of Ce, Sm, Nd and Y." The catalytic performance of pure Ce02 and Sm2O3, on which CH4, C2H4 and C02 were the main products at 773 K was compared with CeF3-promoted samples without discussion, however, of the reaction mechanism. Oxidative dehydrogenation of propane over rare-earth orthovanadates was recently reported by Au and Zhang;sl the catalytic performance was related to the redox properties of these catalysts. R E 0 (Sm2O3, La2O3, Nd2O3) doped with strontium oxide, sodium phosphate or calcium oxide were studied by Buyevskaya and Baerns.82 Oxygen adsorption and the involvement of adsorbed oxygen species in propane activation on this type of catalysts were investigated by means of pulse experiments in vacuum using a TAP-reactor system. It was confirmed that high propane conversions on REO-based catalysts can be obtained only in the presence of short-lived surface oxygen species formed by the adsorption of gasphase oxygen. Results for the dependence of propane conversion and product selectivities on the amount of adsorbed oxygen over Sm2O3 at 823 K are presented in Figure 7. An increase in propane conversion from 7 to 67% was observed with increasing surface coverage of short-lived adsorbed oxygen species. Propene selectivities amounted to 21 -26% in the whole range of surface coverages. The formation of C-C bond scission products (C2H4, CH4) was observed only in the presence of adsorbed oxygen. Continuous flow experiments on Na-P/Sm203 (Na:P:Sm = 2:1 :700) and LnzO3/SrO (Ln = Sm, La, Nd; Ln:Sr = 5:l) catalysts at 1 bar showed that due to high catalyst activity towards oxygen adsorption, the reaction mixture was ignited at 693-723 K. The heat produced by the oxidative conversion of propane was sufficient to sustain the reactions after ignition with the preheat of the feed gas. Olefin yields of 21-30.5% ( S = 53.3-56.8%) were obtained at a short contact time of 0.06 g s ml-' and Tm,,=913-983 K. Propene selectivities of only ca. 20% were observed in the whole range of the reaction conditions indicating that C-C splitting is prevailing when adsorbed oxygen participates in the reaction. Although REO-based catalysts are promising for ODE, their use in the case of propane is limited due to significant formation of ethylene besides propene.

5: Oxidative Functionalization of Ethane and Propane

177

70 I 60

.-0UJ c

20 10

0

5

I0

15

20

25

30

35

oads, /I 014 atoms Figure 7

Propane conversion andproduct selectivities as a function of the amount of adsorbed oxygen consumed in the reaction with propane in sequential pulsing of 0 2 and C3H8 (At = 0.2 s) over Srn203 at 823K (Reproduced by permission from Catal. Today, 1998,42,315)

4.2.2 Boron-containing Oxide Catalysts. ODP catalysts consisting of boria supported on A1203,Ti02, Zr02 and MgO were proposed in our g r o ~ p . ~ ~ - ~ ~ The highest propene yield was 22% ( S = 45-48Y0) on B2O3 (30 wt.%)/Al203; in addition, yields to CI-C3 oxygenates of up to 8% were achieved (S= 2 8 Y 0 ) . ~ ~ Based on the results of pulse experiments with C3D8 and l 8 0 2 in the TAP reactor, it was concluded that the suppression of both dissociative adsorption of propane and oxygen led to the improvement of propene selectivity and yield upon addition of B2O3 to y-A1203.83Propane activation via redox mechanism or by adsorbed oxygen was excluded on this type of catalyst. Boridalumina catalysts most likely possess Lewis acidity, which determines the activation of propane. The initiation involves the formation of an alkyl radical which is released into the gas phase due to diminished surface oxidation and reacts with oxygen forming a peroxy radical. The reaction between the peroxy radical and alkane CnH2n+l02'+ CnH2n+2

-+

CnH2n+102H+ CnH2n+l'

is proposed to be the main propagation step. The decomposition of hydroperoxide results in the formation of olefins and oxygenates A similar mechanism was proposed by Otsuka et al. for the oxidation of propane on B-P mixed oxides.85Results of llB NMR showed that trigonal B 0 3 species being present in both crystalline and amorphous phases are active in the dehydrogenation of propane to propene. The water formed during the reaction causes, however, the loss of these species most probably as boric acid.84Due to this fact boriacontaining catalysts cannot be considered as promising for the ODP reaction. Another point related to these catalysts is the formation of a wide spectrum of

178

Catalysis

different oxygenates which are not always desired in the propene production; those products are most likely due to gas-phase radical reactions. 4.3 Catalytic Materials Containing Noble Metals. - Huff and Schmidt published results on the conversion of propane in the presence of oxygen over noble metal coated ceramic foam monoliths at T > 1173 K and contact times of about 5-10 ms.86 The reactor was operated autothermally and the heat generated by the reaction was sufficient to sustain reactions after ignition. The propene selectivities did not exceed 30%; ethylene was formed with selectivities up to ca. 40%. Total olefin (C2H4, C3H6, C4Hg) selectivities between 55 and 60% were reported for propane conversion > 90% at 1283 K. A heterogeneous reaction pathway was proposed by Huff and Schmidt;86 this scheme is, however, debatable as indicated below. Beretta et al. investigated the role of a Pt/A120, catalyst in an annular reactor as well as an autothermal ODP operation using a Pt c a t a l y ~ t .It~ ~ ? ~ ~ was shown that up to 773 K in the presence of the catalyst only products of combustion were produced; above this temperature olelins were also formed in large amounts. The experimental data confirmed a homogeneous formation of propylene and ethylene and that a Pt catalyst serves primarily as an ignitoreg7 Olefin yields of 50% were obtained in an autothermal mode when operating at sufficiently high flow rates (>0.5 1 (STP) min-') which allowed the ignition. Propene selectivities did not, however, exceed 26%. Compared to the oxidative pyrolysis the use of a catalyst accelerated the ignition and resulted in shorter contact times.88 4.4 Reactor Concepts and Modes of Operation. - Besides the search for new catalytic formulations and mechanistic studies, different reaction engineering concepts aimed at improving the ODP performance have been examined. Some studies dealing with non-conventional reactor types and modes of operation are summarized below. The use of membrane reactors was proposed by several author^.^^-^' Capannelli et al.90 reported the performance of a membrane reactor for ODP in comparison to packed-bed and monolith-like reactors. No significant improvement of selectivity was achieved when using a membrane instead of a monolith. Recently, Alfonso et al. studied the ODP reaction using a reactor with a membrane wall with a layer of a V/A1203 ~ a t a l y s t .The ~ ' best results were achieved with a reactor configuration in which the membrane served as an oxygen distributor. A propene yield of 12% (S = 44%) was obtained at 823 K. Increasing the reaction temperature to 873 K resulted in a change from a kinetic to a mass-transport controlled regime. The maximum propene yield was 15% ( S = 35%) at 873 K. The potential of the autothermal processes for ODP was analysed by Buyevskaya et aLa based on experimental data and on the results of Beretta et aLg8and Huff and Schmidtg6For the ODP over REO-based catalysts used as ignitors, olefins yields of 28-44% ( S = 56-59%) were achieved at contact times of ca. 60 ms.2,82Higher selectivities to C2H4 (ca. 36%) compared to C3H6 (ca. 20%) were observed over the whole range of reaction conditions for all the

5: Oxidative Functionalization of Ethane and Propane

179

catalysts. The lower propene selectivity was due to C-C bond cleavage. The propene yield amounted to approx. 10% at conversions of ca. 50%; a yield of 17% was obtained at 77% of conversion. Propene yields reported by Beretta et al? over a Pt/yA1203 catalyst or by Huff and Schmidt over a Pt/monolith86 did not exceed significantly the values obtained on REO-based catalysts; selectivities towards propene were of about 20%. Beretta et al. confirmed that a Pt catalyst acts as an ignitor, similar to REO-based catalysts." Contrary to ODE, in which true dehydrogenation occurs, C-C bond scission takes place in the autothermal operation of ODP leading to the formation of ethylene and methane in significant amounts in addition to the COX formed by total oxidation. Thus, an autothermal ODP in the presence of a catalyst (which implies catalytic ignition and subsequent gas-phase reactions) cannot be considered as a promising alternative to the existing technologies for the production of propene even when operating in an energy-efficient manner, since the selectivities at the same level of conversion of 60% are lower than those in catalytic non-oxidative dehydrogenation. Catalytic autothermal operation can, however, be of interest when considering the cumulative yields to propene and ethylene. Periodic operation, i.e. feeding the reactants (propane and oxygen) periodically was considered as a possibility for improving the propene selectivity on catalysts operating via redox mechanism. Dehydrogenation of propane occurs in this case with participation of lattice oxygen of the catalyst which in most cases acts more selectively than oxygen species formed on the catalyst surface in the presence of gas-phase oxygen. Switching to oxygen-containing gas leads to catalyst regeneration, i.e. providing the required lattice oxygen. Del Rosso et al. reported on ODP in a periodically operating flow reactor with a nickel molybdate catalyst .92 Although propene selectivities > 80% were obtained propane conversions did not exceed 3%. Formation of C02 was observed even for periods longer than 30 s. It was estimated that only ca. 1% of a monolayer of lattice oxygen was involved in the product formation and reaction was limited by oxygen diffusion. Furthermore, a deactivation of P-NiMo04 due to deep reduction during the periodic operation takes place.92*93 A concept of dehydrogenation with selective hydrogen oxidation in separate reactors or zones using two different catalysts A and B for the following two reactions: catalyst A: C3H8 --+ C3H6 + Hz; catalyst B: H2 + 0.502 --+ H2O was proposed (see, e.g. refs. 94, 95). An overall process can be realized as a sequence of dehydrogenation selective hydrogen combustion --+ dehydrogenation steps in separate reactors or zones. For dehydrogenation (catalyst-A), noble metals, e.g. Pt, were used. For the second step (catalyst-B), A203 supported Cs-Pt-(Sn) or Bi-Pt-(Sn) as well as catalysts containing Bi, In, W, Mo have been reported to act selectively for the combustion of hydrogen ( S > 90%) without oxidation of the olefin produced. High propene yields, also in excess of equilibrium, can be achieved. Moreover, the process provides also the possibility of a better heat balance compared to the conventional dehydrogenation. --+

Catalysis

180

5

Partial Oxidation of Ethane and Propane to Oxygenates

The partial oxidation of ethane and propane to oxygenates suffers generally in a similar manner as oxidative dehydrogenation from the increased reactivity of the desired products for consecutive total oxidation steps as compared to the alkane feed. In most cases, increased alkane conversion leads to decreasing product selectivity in favour of COX.The main task of present research is to find catalysts or modes of operation designed to reduce consecutive total oxidation of the desired oxygenates derived from ethane (acetic acid) and from propane (acrolein and acrylic acid). In the following, results on the partial oxidation of ethane and propane are discussed; the emphasis is on recent studies. 5.1 Ethane to Acetic Acid. - The formation of acetic acid via catalytic gas-phase oxidation of ethane was first reported by Thorsteinson et al. as a by-product in the oxidative dehydrogenation of ethane over unsupported Mo-V-Nb-O catalysts (600 K; 2 MPa).23Acetic acid selectivity of 26 % was observed at low ethane conversion (5 %). During the last decade much research activity has been focused on direct catalytic synthesis of acetic acid from ethane. BP chemical^,^^,^^ The Standard Oil C ~ m p a n y ,Rhone-Poulenc,lOO~lO1 ~~?~~ former Hoechst C ~ m p a n y l ~and ~-'~~ Saudi Basic Industries Corporationlo5 claim catalysts and processes. In general, the reaction was carried out at elevated pressures and in an excess of ethane at temperatures between 573 and 623 K. Adding water increased the selectivity towards acetic acid. The conversion of ethane and the formation of the target product per pass are limited by the concentration of oxygen in the feed gas. Contrary to other reactions for functionalization of ethane and propane the production of acetic acid from ethane reveals selectivities > 70-80%, which are appropriate for its commercial realization. The relative low ethane conversions (< 15%) per pass can be tolerated since using ethane-oxygen mixtures allows recycling of unreacted ethane without problems as compared to N2 dilution when air would be used. An ethane-to-acetic acid process appears to be a promising technology of alkane functionalization and could be put into practice in the next few years as an alternative for the existing methanol carbonylation technology; demonstration plant activities have been recently reported by SABIC. 5.1.1 Catalysts. Most of the reported catalysts are based on a Mo-V-Nb

oxides. In a patent issued to BP in 1993, a Mo-Re-V-Nb-0 catalyst was described, on which a selectivity of 78 YO(&H6 = 14 %) was obtained at 550 K and a total pressure of 2.8 MPa when water vapour was present.97 In the patents issued to the former Hoechst company the following mixed oxide systems were claimed: Mo,PdbRe,XdY,Ox (X = Cr, Mn, Nb, B, Ta, Ti, V and/or W; Y = Bi, Ce, Co, Te, Fe, Li, K, Na, Rb, Be, Mg, Ca, Sr, Ba, Ni, P, Pb, Sb, Si, Sn, T1 and/or U; a = 1; b > 0; c >O, d=0.05-2; e=0-3)lo2 and

5: Oxidative Functionalization of Ethane and Propane

181

Mo,PdbX,YdO, (X= Cr, Mn, Nb, Ta, Ti, V, Te and/or W; Y = B, Al, Ga, In, Pt, Zn, Cd, Bi, Ce, Co, Cu, Rh, Ir, Au, Ag, Fe, Ru, Os, K, Na, Rb, Cs, Mg, Ca, Sr, Ba, Zr, Hf, Ni, P, Pb, Sb, Si, Sn, T1 and/or U; a = 1; b > O ; c > O , d = 0-2). lo3 The authors proposed that adding palladium enhances the selectivity towards acetic acid of Mo-V-Nb catalysts ( S = 80 YO,&2H6 = 10 YO) and of Mo-Re-V-Nb oxide catalysts ( S = 91 %, &2H6 = 4 0/,).1029103 A Mo-V-Nb-0 catalyst modified with small amounts of P, B, Hf, Te and/ or As was claimed by Karim et al. lo5 A maximum yield of acetic acid of 26.6% (S=49.9%) was obtained on Mo2.5V1.0Nb0.32P0.0420x at 533 K, P = 1.38 MPa and space velocity 1100 h- using ethane/air (15/85) reaction mixture. On a catalyst consisting of the oxides of W, V, Re, Nb, Sb and Ca, a high yield of 10.9% was reported by Kitson at 550 K and 2.8 MPa when adding water to the reaction mixture.97 Other types of catalysts such as V/Ti02, V-P/TiO2, V-P-Mo/Ti02 were also proposed100~101-106*107 but their performance was rather poor compared to the above results on doped Mo-V-Nb and W-V catalysts. Vanadium-containing heteropoly acids (H3+iPM012-iViO40; i = 1,2,3) supported on Ti02 claimed by Bordes et al. allowed 70 % selectivity at ethane conversion of 8%.'O0 On Mo-V-P-MoKi02, selectivity to acetic acid amounted to 21% at ethane conversion of 3.2% at 275 "C and 0.7 MPa.lo7 Hydrothermal synthesis of Mo-V-M-0 (M = Al, Fe, Cr and Ti) complex metal oxide catalysts for partial oxidation of ethane was reported by Ueda et al. Io8 XRD patterns of the samples revealed two sharp diffraction peaks at 28 ca. 22 and 45" and very broad diffraction peaks appeared at 8, 11 and 27". No definite conclusion on the phase composition was derived. The authors speculated that all the samples were at least biphasic. For the phase giving the set of the narrow diffraction lines, the layer-type structure was assumed. The hydrothermally synthesized solid materials were tested for the production of ethene and acetic acid from direct ethane oxidation at 613 K and T = 1.2 g s ml- using C2Hd02/H20/N2= 3/1/2/4 feed gas. A maximal yield of acetic acid catalyst. of 1.7% ( S = 1I S%) was obtained on a Mo6V2AllTi0.70x Catalytic performance of a MolV0.~~Nb0.12Pd0.00050x catalyst was investigated by Linke et al. at temperatures between 500 and 580 K and elevated pressures between 1.3 and 1.6 MPa with and without water addition to the feed gas.lo9The conversion of ethane and product yields as a function of water partial pressure at reactor inlet are given in Figure 8. Catalytic data from publications since 1995 are given in Table 5.

'

5.1.2 Fundamentals and Kinetics. Ruth et al. studied the catalytic performance

of Mo-V-Nb oxides and of the individual phases therein which are an essential part of (nearly) all catalysts used.112The crystalline phases found were MOgV9040, Mo3Nb201 and MOO,; furthermore, an amorphous part was identified of the approximate composition Mo84V13Nb2Ox to which the selectivity towards acetic acid is ascribed based on a comparison with the catalytic performance of the pure phases. Tessier et al. studied pure VPO and titania-supported VPO, and VO,. lo6 They reported that polyvanadates are the

Catalysis

182

5 4

s 3 \

>x

2

1

0

Figure 8

B YC2H4

0 :o

0.1

0.2

0:3

Conversion of ethane and product yields as a function of water partial pressure at reactor inlet ( T = 519 K, z,,,RTp = 19.6.103 kg s .m-’, meat = 6.0 g, p c 2 ~ 6= 0.6 MPa, Po2 = 0.12 MPa, P ~ 1 . MPa) 5 (Reproduced by permission from J. Catal., 2002,205, 16)

active species responsible for the formation of acetic acid. Roy et al. reported for titania supported V-P-MwO catalysts that a different MoN ratio does not change the selectivity towards acetic acid at iso-conversion and concluded that only vanadium sites participate in acetic acid formation.lo7 The role of molybdenum was assumed to influence the electronic density around the vanadium sites. A detailed study on the reaction mechanism over a MolV0.25Nbo.12Pd0.00050~ catalyst was carried out by Linke et al. 109a~1lo It was found that a change in the reaction pathways occurs with temperature. While at low temperature the formation of acetic acid is a consecutive reaction with ethylene as intermediate, at high temperature ethylene and acetic acid are mainly formed in parallel reactions. Formation of carbon dioxide occurs from ethane, ethylene and acetic acid. Addition of water strongly accelerates the oxidation of ethylene to acetic acid forming HO centres which are supposed to be involved in the oxidation of ethylene to acetic acid (heterogeneous Wacker oxidation); the respective reaction scheme is presented in Figure 9. Pulse experiments applying the TAP reactor revealed that oxygen is not adsorbed on this catalyst and thus, only lattice oxygen is responsible for both selective and unselective oxidation steps. Within Linke’s et al. studies also kinetic modeling was lo The tubular reactor used for providing the integral kinetic data was modelled as an isothermally operating pseudohomogeneous one-dimensional plug-flow reactor. Two different models taking into account surface processes such as catalyst reduction by ethane or ethylene and reoxidation by oxygen and surface hydroxylation were suggested. The

Catalyst

Table 5

H20 20 20 0 0 20 10 0 20

O2 10 10 17 17 8 17 5 8

C2Hb 30 30 15 15 40 62 85 40

Feed gas vol. Yo

T

K 613 613 533 533 553 548 498 539

ptot.

MPa 0.1 0.11 1.38 1.38 1.5 0.7 0.1 1.6

Catalytic results on the partial oxidation of ethane to acetic acid

4.2 14.7 53.3 64.4 8 3 0.5 10 93.

n.g.

21

n.g. n.g n.g.

18.4 61.3

0 2

10.7 11.5 49.9 38.2 90 22.5 73 76.8

Acetic acid

55.8 59.6 10.5 26.3 2 17 8 7.6

C2H4

YO

YO c2H6

Selectivity

Conversion

108 108 105 105 102 107 106 110

Rex

3

i2

fa

3

%

g

2

% !i;.

s

$.

!sCZI

2.

Fg

9

Catalysis

184

H

OH

I

M,

I

0 M, 0

t

M,

+ 112 0,

0 M,

0

1 fH2-0

Ml

0

M2 0

*

H CH2

H

\ \ M,

0 M,

0

H

\

M,

\

0 M,

0

Catalytic cycle of the oxidation of ethylene to acetaldehyde andlor acetic acid (heterogeneous Wacker oxidation); oxygen vacancies are indicated by the open square."O

Figure 9

models were discriminated on the basis of experimental data. For the superior kinetic model two different catalytic centres were assumed, i.e. one for the oxidative dehydrogenation of ethane and the other one for the heterogeneous Wacker oxidation of ethylene to acetic acid. The reaction steps for the optimal model were as follows: Ethane activation (formation of adsorbed ethylene)

(1) C2H6 + 2-0 + Z-CzH4 + H20 Acetic acid formation

(2) Z-C2H4

+0 2

--+

Z + CH3COOH

(3) C2H4 + X-OHOH + 0.5 0 Catalyst reoxida t ion (4) 0.5 0

2

+ z + 2-0

(5) 0.5 0 2 + X + X-0

2 -+

CH3COOH + X + H20

5: Oxidative Functionalization of Ethane and Propane

185

Ad- ldesorption of ethylene and water (6) C2H4 + Z

* Z-C2H4

(7) H20 + Z + Z-OH2 Formation of Wacker-like centre (8) X-O + H 2 0 + X-OHOH Unselective reaction steps (9) C2H6 + Z-0 + 302 --+ [ * ] 4 2C02 + 3H20 + Z

(10) C2H4 + Z-O + 2.502

--+

[ * ] -+ 2C02 + 2H20 + Z

(11) CH3COOH+Z-O+ 1 . 5 0 2 -+ [ * ] --+ 2C02+2H20+Z Kinetic constants and activation energies of each reaction were estimated. The analysis of the kinetic results leads to the conclusion that ethane activation is the rate-determining step for oxidizing ethane and that the formation of the Wacker centre by water adsorption is rate determining for converting ethylene to acetic acid. For a comprehensive characterization of the MolVo.2sNbo.12Pd0.ms0x catalyst various techniques (SEMIEDX, XRD, TGIDTA, TPD with ethane) were app1ied.logaSEM/EDX analysis showed that the catalyst is a heterogeneous mixture. Different morphologies and an inhomogeneous distribution of the constituent elements were observed. The catalyst was found by XRD to consist of crystalline Moo3 and M05O14 and a large amorphous fraction containing Nb and V. The role of the various metal oxides is discussed elsewhere.logs 5.1.3 Reactor Concepts. Higher selectivities and yields respectively depend on further catalyst development but also on reaction engineering improvements in carrying out the reaction. From patents it appears that for an ethane-to-acetic or a fluidized-bed acid process either a (multitubular) fixed-bed96~97~102~103~10s r e ~ t o r ~ ~may - ~be~used. * ~ For ~ ~better , ~ ~isothermicity ~ and for minimizing any danger of explosion, the process is mostly carried out at high ethane-tooxygen ratios. Thus, ethane conversion and yield of acetic acid are limited by oxygen concentration in the reaction mixture. The use of oxygen-rich conditions was claimed by Karim et al. lo5 The authors used a feed gas consisting of 15% ethane and 85% air allowing to achieve high ethane conversions (60-65%), selectivity towards acetic acid amounted, however, only to 30-38%. In the method claimed by Benkalowycz et al.98 ethane and oxygen are separately introduced in the reactor zone containing the fluidized catalyst particles. After cooling the reaction gas and separation of acetic acid part of the ethane was returned in the feed gas. The use of fluidized-bed reactors suffers in general from back mixing which favours consecutive reactions; in the present case back mixing leads to oxidation of the intermediate ethylene to acetic acid but also to its further total oxidation depending on temperature.

186

Catalysis

The mixing of feed gas with the inert reaction product C02 allowed to use higher oxygen concentrations and to increase ethane conversion. The addition of water which has a positive effect on the formation of acetic acid leads to diluted acetic acid. In order to increase the yield of a concentrated product a method for producing acetic acid from ethane and oxygen using a fixed bed catalyst which includes a reactor cascade with a distributed oxygen feed between the reactors was claimed by Borchert et a1.'04 The reaction gas from the first reactor is mixed with oxygen and passed in the second reactor without condensation of water or acetic acid. This reactor concept provided high conversions, highly pure acetic acid and decreased amounts of gas in the recycle. The further advantage of this method is that the carbon dioxide was almost the only by-product, the formation of CO and ethylene could be neglected. As catalyst the compositions claimed in earlier patent^'^^*'^^ as well as W-containing mixed oxides can be used. Selectivity towards acetic acid was claimed as 2 60 mol% at the degrees of ethane conversion of 4-6% in each reactor. As follows from the given example the concentration of acetic acid in aqueous solution amounted to 45 wt.%. The operation of a fixed-bed and fluidized-bed reactor for the partial oxidation of ethane to acetic acid was simulated by Linke et al. The kinetic model used for reactor simulation' lo is described above (Section 5.1.2). For the fixed-bed reactor, the influence of inlet temperature was investigated in a temperature range from 500 to 525 K and at a total pressure of 16 bar. The simulation indicated serious problems in temperature control. It was found that the temperature maximum appears in the front part of the reactor. For an assumed reactor diameter of 0.025 m and a gas velocity of 0.045 m s- stable reactor operation without runaway can be maintained up to an inlet temperature of the feed of 520 K. The highest yield of acetic acid was obtained for a gas inlet temperature between 533 and 550K; in this situation the safety of reactor operation could be improved by a higher dilution of the catalyst bed and by a decrease of pressure, reactant concentration, reactor diameter and/or feed inlet temperature. All of these efforts would, however, limit the space time yield of acetic acid. An alternative might be the application of a fluidized-bed reactor due to the opportunities of efficient heat removal. 5.2 Propane to Acrylic Acid and Acrolein. - Catalytic results on the direct oxidation of propane to acrylic acid and acrolein are summarized in Table 6.

5.2.1 Direct Oxidation of Propane to Acrylic Acid. In 1986 Ai reported the partial oxidation of propane to acrylic acid over V205-P205-based catalysts using reaction mixtures with a very high oxygen content (OJC3Hs >41).'13 Selectivities of about 60% were obtained for propane conversion below 10%; with increasing conversion selectivity decreased due to consecutive oxidation of acrylic acid to COX.Selectivity could be improved by feeding water vapour along with the propane-oxygen feed. The maximum yield of acrylic acid amounted to 10.5% (S = 25%) when using feed gas consisting of 0.54% propane in oxygen (catalyst: TeIPN = 0.15/1.15/1). l 3

'

H20 16 67 47 23 30 0 23 20 45 63

0 2

83 32 10 73 14 40 15 10 13 11 20 30 30 18 18 21

41 45 80 22 19 47 18 8 34 19 4 14 10 10 8.4 36.6 10

653 633 653 573 673 653 613 613 663 673 633 803 803 733 733 833/753 703

6 23" 2a 4" 3 4 3" 3.6 1.3 3" 1.8" 1 1 0.36 0.36 1.7/0.9

a

s; n.g. - not given; Trace.

Y O

K

gsml-'

X ( C3H,)

T

z

V1Zro.5P1.50xafter 50 h H3PMo12040 treated with pyridine Mo6V3Te10xafter grinding BiMo 12V5Nb0.5SbKOx 0 50 Mo-Sn-O Mg7Bi5Mo 1 2 o x 60 0 Ca7Bi5M0120x 60 0 NiMo04-0.5Mo03 + 2wt.Y0Te2Mo07+2wt.Y0P2O5 15 0 NiMo044.5Mo03+lwt.% Te2Mo07 + 2wt.Yo P2O5 15 11 Layer 1: B203/A1203+ Zn7Bi~Mo12OX 7 0 Layer 2: Ca7Bi5M0120x MolC00.950~followed by M O ~ ~ W ~ B ~ ~ C O ~ . ~ F ~ ~80S ~ I 20 . ~ I ( O0 . O ~ O 3" ~

~ ~ 2 . 5 ~ ~ 0 . 0 8 ~ 1 . 2 6 ~ ~ ~ ~ 1 1 0 4 0

0.54 1 3.3 3.2 2 30 3.2 20 6.3 26

C3

Feed gas /kPa

Catalytic results on the partial oxidation of propane to acrolein (Ac.) and acrylic acid ( A A )

V-P-Te-0 V205-p205 Mo lVO.3Te0.23NbO.12ox V-P4/TiO2-Si02 v-P-0

Catalyst

Table 6

59

20 14 16 18 20

n.g.

0.4 1

n.g n.g. n.g. tr. n.g. n.g.

14

25 25 0

-

3

5 21 18 24 7.0 24 27 2.6

n.g. Tr.C

7 5

n.g. n.g.

0

n.g.

26 17 61 39 26 27 81 28.5 46 29 48 -

n.gb 0

CjH,

AA

/"/a

Ac.

Selectivity

132

113 114 126 115 117 119 123 121 25 127 125 74 74 133 133 74

ReJ

%

7

$

9

g

trj

q

g.

2

$

'$ 2

2 rp

8

?l

188

Catalysis

No significant progress for V-P-0-based catalysts has been published; Ai reported in 1998 a slightly improved maximal acrylic acid yield of 7.5% (S=17%).l14Acidic acid was formed under all conditions as an important side product with a maximal yield of 7.5%. Han et al. reported improved performance of dispersed vanadium phosphorus oxide when using a reaction mixture with a high excess of oxygen ( 0 2 / C3H8= 23); V-P-0 supported on titania-silica xerogel resulted in an acrylicacid yield of 8.5% (S=39%) at 573 K.'15 Gribot et al. suggested that the catalytically crucial phases of the V-P-0 catalytic material is the VOP04 phase (mainly 6) associated with poorly crystallized (V0)2P207, which was observed in the selective catalysts. l6 The catalytic performance of these phases was ascribed to their Brarnsted- and Lewis-acid sites. A comprehensive study of V-P-O catalysts was performed by Quandt.' l 7 Different preparation and activation methods have been tested, supplemented by catalyst characterization applying various techniques (e.g. 31PNMR, XPS, XRD, REM/STEM). The highest propane conversions and selectivities to acrylic acid were obtained on catalysts consisting of (VO)ZP2O7 phase in which all vanadium was in 4+ oxidation state. Propane conversions varied from 10 to 50% for differently activated samples. The selectivity was not significantly affected by the activation procedure. At a conversion of 22%, the selectivities to acrylic acid, acetic acid and propene amounted to 24-27, 12-17 and 7-10%, respectively. The in situ activation processes were affected by the rest of solvent from catalyst preparation as well as by the water content in the feed gas. The time-on-stream behaviour of the V-P-0 precursor during in situ activation is shown in Figure 10. All V-P-O catalysts suffered from deactivation caused by a loss of phosphorus due to the presence of water in the feed gas. Addition of different promoters (Co, Mo, Nb) influenced the activation process but did not result in an increased yield of acrylic acid. Another type of catalyst active in the direct oxidation of propane to acrylic acid is heteropoly compounds. A yield of 9% was obtained by Centi and Trifiro on H ~ P V ~ M O ~ OMizuno O ~ O . et ~ ~al.~ reported a yield of 13% when using C S ~ . ~ F ~ O . ~ ~ H1 0~4 0. as ~ ~a Pcatalyst V M Oat~ 653 K.' l 9 The maintenance of the Keggin structure during the reaction in the range from 573 to 673 K was determined from IR and XRD measurements. Oxidation of propane over heteropolymolybdophosphoric acids which were treated in pyridine during catalyst preparation was investigated by Ueda and coworkers.120,121 A yield of acrylic acid of 2.1% ( S = 28.5%) was obtained at 613 K on a (PyH)3PMo12040catalyst which had been heated in a flow of N2 at 693 K before the catalytic reaction.121The FT-IR study revealed the generation of Lewis acid sites on the surface along with the formation of the primary oxygen-deficient Keggin structure. The use of heteropolyacids or polyoxoanions comprising at least nine atoms of a first framework (Mo, W or V or their combinations) and one to three atoms of a second framework (Zn or a transition metal different from the first framework metal) was claimed by Lyons et al. (Sun Company, Inc.) for the production of unsaturated carboxylic acid from alkanes.122

5: Oxidative Functionalization of Ethane and Propane

b

0

1

I

1

I

20

40

60

80

I

189

I

I

100 120 140 160

tAlh

Figure 10

Time-on-stream behaviour of the V-P-0 catalyst precursor during in situ activation (CjHSlairlH20 = 2/68/30, z = 3 g s mi-', T = 673 K )

An increase of research activities in propane oxidation to acrylic acids has been observed recently. Ueda and Oshihara studied the partial oxidation of propane over the hydrothermally synthesized Mo-V-M-O (M = Sb or Te) catalysts.25Similar to the systems reported by Ueda et al. for ethane oxidationlo* the above mixed oxides revealed two sharp X-ray diffraction peaks at 22 and 45" and characteristic diffractions below 10". An interesting finding in this paper concerned a significant increase in activity due to grinding the samples after treatment at 873 K in N2. For Mo6V3Te10,catalyst, the yield of acrylic acid increased due to grinding from 4.7 to 15.7% coincided with increasing selectivity from 36.4 to 45.9%. Unfortunately, neither definite explanation of this effect nor data on the stability of the ground samples were given.25 For V-Zr-P-O catalysts, a yield of acrylic acid of 14.8% at high selectivity of 8 1% has been reported by Han et al. 123 By the addition of Zr, the yield and selectivity to acrylic acid increased significantly in comparison to a V-P-0 catalyst resulting in a yield of 11.2% (S = 48%) after 50 h. As the ratio of Zr to V was changed from 0.125 to 0.5, the selectivity increased from 70 to 81% and the yield from 13.5 to 14.8%, respectively. With increase of the Zr to V ratio towards 1.0 both, activity and selectivity started to decrease. The authors

190

Catalysis

performed a stability test over a period of 100 h and showed that a V1Zro.5P1.50x catalyst revealed a stable performance with a yield of ca. 14.5%; this is contrary to a V-P-0 catalyst on which deactivation took place leading to a decrease of the yield of acrylic acid from 15 to 8%. Industrial interest in the direct propane oxidation to acrylic acid appears to be high as suggested by numerous patent applications (see, e.g. refs. 50, 122, 124-27 and 137). A high yield of acrylic acid of 48% (S=60.5%) was claimed in a Mitsubishi patent over a Mo-V-Te-Nb-0 catalyst.126As was pointed out by Grasselli, efforts of several laboratories to reproduce this yield were not successful.4 Experiments carried out by the present authors' group were also not successful; it is believed that this is probably due to some missing information on catalyst preparation not revealed in the patent; this was confirmed in a recent patent137 by Rohm & Haas in which results were obtained similar to those of Mitsubishi. The yields claimed in other patents did not exceed 16%. These yields appear to be not high enough to consider a direct conversion of propane to acrylic acid as a real alternative to the respective well-established two-step process via propene to acrolein. Some new concepts for achieving high yields and selectivities can be, however, expected in the near future, in particular taking into account the recent results of Ueda and O ~ h i h a r aas~well ~ as those of Han et al. 123

5.2.2 Direct Oxidation of Propane to Acrolein. This reaction appears to be an even more difficult task than propane to acrylic acid due to much lower stability of acrolein than acrylic acid under reaction conditions required for propane activation. Nevertheless, direct oxidation of propane to acrolein would be an interesting alternative to propene oxidation if high yields and selectivities towards acrolein will be achieved. The data published indicate, however, that the reaction still leads to significantly lower selectivities and yields of the target product than the propane-to-acrylic acid reaction. A maximum acrolein yield of 13% as reported by Kim et al. was achieved using a Bi-Mo-V-Ag-0 catalyst under conditions where, however, primary gas-phase dehydrogenation of propane occurred.128 The results of See1 showed that Bi-Mo-0 based catalysts do not exhibit any appreciable activity for the required primary oxidative dehydrogenation of propane. 129 The best data reported up to 1997 are summarized by Baerns et aZ.;74the acrolein selectivities as a function of propane conversion are given in Figure 11. The selectivities did not exceed 60% even at low propane conversions. Total oxidation as a non-selective side reaction of propane as well as the consecutive oxidation of acrolein contributes to the loss in acrolein selectivity. Yields reported are generally below 3-5%. 130 Some new catalytic systems and modes of operation have been introduced during the last three years but no significant improvements of yields were, however, In the patent of Tenten et aZ., a process of propane oxidation with molecular oxygen using propane-rich (>70%) feed gas was ~1aimed.l~' The use of Bi0.85V0.54Mo0,45Ag0.01Ox,VOPO4 and Bi8PMO12Ox catalysts resulted, however, only in a maximal acrolein selectivity of 9%.

5: Oxidative Functionalization of Ethane and Propane

S(C,H,O)

191

/%

I

.loo 80

60 40

20 0 0

Figure 11

10 20

30

40

50

60

70

Dependence of acrolein selectivity on propane conversionfrom different studies according to rej 130 (A - with contribution of gas-phase dehydrogenation of propane according to re$ 128) 74 (Reproduced by permission from Catal. Today, 1997,33, 8 5 )

In a study of Baerns et al.,74 the catalyst selection for the partial oxidation of propane to acrolein was based on the widely accepted assumption that propene is an intermediate product in the overall process. According to this fact a suitable catalyst composition should consist of two different active sites for propane dehydrogenation and for oxygen insertion in propene. Me,O,supported Ago.01Bi0.85V0.54M00.450x (Me,O, = y-AlzO3, Ti02, Si02) and Me7Bi5M0120x(Me = Mg, Ca, Zn) catalysts were tested and characterized using XPS, XRD, ESR, IR and transient experiments. For MexO,-containing catalysts, it was found that in spite of the presence of active phases for oxidative dehydrogenation of propane (dispersed VO, species) and oxygen insertion in propene (scheelite structure) almost no formation of acrolein was observed. Thus, the catalyst-design approach combining two active sites still has severe shortcomings; the two active phases do not act cooperatively to any large extent. For Me7Bi~Mo120, (Me = Mg, Ca, Zn) catalysts, it was determined that different types of lattice oxygen are involved in propane dehydrogenation to propene and in consecutive reactions towards acrolein and While propane was not converted over pure bismuth molybdates, modification by introducing a third cation resulted in more active catalysts for propane oxidation. The performance of these Me-Bi-Mo-O catalysts depended on the third added cation. Oxidative conversion of propane over Me7Bi5Mol2 0 , catalysts was studied in a wide range of propane and oxygen partial pressures. Propane partial pressure significantly affected propane conversion over a Ca7Bi5Mo120, catalyst but did not influence the conversion for Mg7Bi5Mo120x.For easily reducible Ca-Bi-Mo-0 the increase in acrolein formation

192

Catalysis

was observed only at a low oxygen content (10 kPa) under conditions of complete oxygen conversion. In the case of the Mg-containing catalyst the suppression of total oxidation on using a reaction mixture with higher oxygen content (30 kPa) can be assigned to more strongly bound forms of lattice oxygen. Following the approach of combining materials being active in the oxidative dehydrogenation of propane and in the partial oxidation of propene, a twolayer fixed catalyst bed was used where each layer was operated at an optimal temperature (so-called structured catalyst bed); this procedure led expectedly to an improvement of the acrolein yield.74Applying such type of reactor (1st layer: B203(30 wt.0/o)/A1203/Zn7Bi5M0120x, T = 833 K; 2nd layer: Ca7BisMo120x,T = 753 K) an acrolein yield of 7.4% (S=20%) was obtained without contribution of any gas-phase propane dehydrogenation. Catalytic gas-phase oxidation of propane with molecular oxygen at 573-773 K in a fixed catalyst bed consisting of layers of two different catalysts was claimed by Jachow et al. 132 Using a Mo12W2BilCo5.5Fe3Sil.61(0.080x catalyst placed after a MolC00.950~catalyst and a feed gas consisting of 80% propane and 20% oxygen an acrolein yield of 5.9% (S=59%) was achieved; additionally, the yield of acrylic acid amounted to 1.4%. The reactivity of Ni-Mo-Te-P-0 systems was studied by Kaddouri et al. for the direct oxidation of propane to acrolein and acrylic acid.133It was found that the addition of Te and P improved the yields and selectivities of acrolein and acrylic acid. The best results in terms of total selectivity (acrolein and acrylic acid) were obtained with NiMo044.5Mo03+ 2wt.%Te2 M0O7+2wt.%P2O5 without the addition of water vapour at 733 K; the selectivity was 41.3% at a propane conversion of 9.5%; only trace amounts of ethylene and acetaldehyde were observed. For improving the catalytic performance and limiting the oxidative degradation of the oxygenate products, the effect of water vapour was also investigated. The addition of water resulted in a decrease of propane conversion which coincided with an increasing selectivity to propene and in some cases to acrolein. At degrees of conversion of ca. 12% the addition of water vapour resulted in an increase in acrylic-acid selectivity. Kaddouri et proposed that the reaction pathway on doped Ni-Mo-0 catalysts differs from that on V-P-0 catalysts reported by The scheme includes the formation of the allylic intermediate from propylene formed in the first dehydrogenation step; the presence of water in the feed gas did not influence this step. 6

Other Reactions

A new approach for the oxidation of light alkanes catalysed by NO, in the gas phase at pressures less than 1 bar was reported by Otsuka et al. 134 It was found that addition of NO to the mixtures of alkane and O2 enhanced the yield and selectivities towards oxygenates. The reaction was proposed to occur via the formation of NO2 which initiates the reaction mostly via C-C bond fission

5: Oxidative Functionalization of Ethane and Propane

193

resulting in the predominant production of C1 oxygenates. In order to avoid the formation of nitroalkanes, reaction temperatures above 873 K have to be applied. For the oxidation of ethane in the presence of NO, the main selective product was HCHO; its selectivity did not, however, exceed 20%. The maximum HCHO yield was 8%. In propane oxidation, HCHO, and CH3CH0 were formed with maximum selectivities of ca. 50%, the highest oxygenates yields were 10-12%. Whether this method might eventually lead to any process scheme that could be transferred to practice appears unlikely at present. The partial oxidation of ethane to ethanol and acetaldehyde in the H2/02 fuel cell at 343 K in the presence of Fe2+or Cu2+ions was reported by Kuzmin et al.135 Selectivities to C2-oxygenates of 95-98% were obtained, but no data on the degree of conversion are given. Ethane activation with participation of OH or H 0 2 radicals was assumed; high selectivity towards C2 oxygenates was explained in terms of their efficient separation through the porous membrane electrode.

7

Summary and Outlook

All of the direct oxidation reactions described and the corresponding processes are still in an exploratory state. At present, only the partial oxidation of ethane to acetic acid appears to be close to commercial practice. The ODE carried out in an autothermal mode, combining heterogeneous catalytic and non-catalytic gas-phase reactions, has some potential if selectivity and yield of ethylene are further increased. The outlook for ODP and for acrolein and acrylic acid from propane is less favourable; new concepts for catalyst design are necessary. From a fundamental point of view still more knowledge is needed on cooperative effects of active phases and on identification of factors determining the selective pathway.

8

Acknowledgments

Financial support by the German Federal Ministry for Education and Research and the Department of Science of the State of Berlin as well as by Fonds der Chemischen Industrie is appreciated.

9 1 2 3 4 5 6

References H.H. Kung, Adv. Catal., 1994,40, 1. M. Baerns, O.V. Buyevskaya, Catal. Today, 1998,45, 13. E.A. Mamedov, V. Cortes Corberan, Appl. Catal. A:, 1995,127, 1. R.K. Grasselli, Catal. Today, 1999,49, 141. S . Albonetti, F. Cavani, F. Trifiro, Catal. Rev. Sci. Eng., 1996,38,413. D. Wolf, O.V. Buyevskaya, M. Baerns, Appl. Catal. A., 2000,200,63.

194 7

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

28 29 30 31 32 33 34 35 36

37 38 39 40 41

42 43

Catalysis

O.V. Buyevskaya, A. Briickner, E.V. Kondratenko, D. Wolf, M. Baerns, Catal. Today, 2001,67,369. E.R.S. Winter, J. Chem. SOC.( A ) , 1968,2889. R.X. Valenzuela, J.L.G. Fierro, V. Cortes Corberan, E.A. Mamedov, Catal. Lett. 1996,40, 223. J. Le Bars, A. Auroux, M. Forissier, J.C. Vedrine, J. Catal., 1996,162,250. G. Colorio, J.C. Vedrine, A. Auroux, B. Bonnetot, Appl. Catal., 1996,137, 55. E. Morales, J.H. Lunsford, J. Catal. 1989,118,255. R. Burch, E.M. Crabb, Appl. Catal. A , 1993,97,49. A. Wanzek, PhD Thesis, Ruhr-University Bochum, 1991. M. Huff, L.D. Schmidt, J. Phys. Chem., 1993,97,11815. C. Yokoyama, S.S. Bharadwaj, L.D. Schmidt, Catal. Lett., 1996,38, 181. S.J. Conway, D.J. Wang, J.H. Lunsford, Appl. Catal., 1991,79, L1. H.X. Dai, Y.W. Liu, C.F. Ng, C.T. Au, J. Catal. 1999,187,5920. S . Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki, Catal. Lett., 1999, 59,173. H.X. Dai, C.F. Ng, C.T. Au, J. Catal., 2000,193,65. H.X. Dai, C.F. Ng, C.T. Au, Catal. Lett., 1999,57, 115. H.X. Dai, C.F. Ng, C.T. Au, J. Catal., 2000,189, 52. E.M. Thornsteinson, T.P. Wilson, F.G. Young, P.H. Kasai, J. Catal., 1978, 52, 116. F.A. Al-Sherehy, A.M. Adris, M.A. Soliman, R. Hughes, Chem. Eng. Sci.,1998, 53, 3965. W. Ueda, K. Oshihara, Appl. Catal. A , 2000,200, 135. F. Cavani, M. Koutyrev, F. Trifiro, Catal. Today, 1996,28, 319. J. Santamaria-Gonzalez, M. Martinez-Lara, M.A. Bafiares, M.V. MartinezHuerta, E. Rodrigues-Castellon, J.L.G. Fierro, A. Jimenez-Lopez, J. Catal., 1999,181,280. A. Kadduouri, R. Anouchinsky, C. Mazzochia, L.M. Madeira, M.F. Portela, Catal. Today, 1998,40,201. Z. Dang, J. Gu, J. Lin, D. Yang, Catal. Lett., 1998, 54,129. O.J. Velle, A. Andersen, K.-J. Jens, Catal. Today, 1990, 6, 567. L. Ji, J. Liu, Chem. Commun., 1996, 1203. V.R. Choudhary, B.S. Uphade, S.A.R. Mulla, Angew. Chem. Int. Ed. Engl., 1996, 34,665. O.V. Buyevskaya, M. Baerns, DGMK-Tagungsbericht, 1998,9803, 139. M. Baerns, O.V. Buyevskaya, DE 19734541A, 1997. S.A.R. Mulla, O.V. Buyevskaya, M. Baerns, J. Catal., 2001,197,43. J. Coronas, M. Menendez, J. Santamaria, Ind. Eng. Chem. Res., 1995, 34, 4229. E.V. Kondratenko, O.V. Buyevskaya, M. Baerns, J. Mol. Catal. A: Chem., 2000, 158, 199. S.S. Bharadwaj, L.D. Schmidt, J. Catal., 1995,155,403. S.S. Bharadwaj, C. Yokoyma, L.D. Schmidt, Appl. Catal. A , 1996,140,73. M.C. Huff, I.P. Androulakis, J.H. Sinfelt, S.C. Reyes, J. Catal., 2000, 191,46. C.T. Au, K.D. Chen, H.X. Dai, Y.W. Liu, C.F. Ng, Appl. Catal. A , 1999, 177, 185. K. Nakagawa, M. Okamura, N. Ikenaga, T. Suzuki, T. Kobayashi, Chem. Commun., 1998,1025. Y. Liu, J. Xue, X. Liu, R. Hou, S . Li, Stud. Surf: Sci. Catal., 1998, 119, 593.

5: Oxidative Functionalization of Ethane and Propane 44

45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

75

76 77 78 79

195

L. Xu, L. Lin, Q. Wang, L. Yan, D. Wang, W. Liu, Stud. Surf. Sci. Catal., 1998, 119,605. S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki, Chem. Lett., 1999, 569. F. Buonomo, D. Sanfilippo, F Trifiro, in Handbook of Heterogeneous Catalysis, ed. Knozinger and Weitkamp, VCH, 1998, Ch. 4.3, p. 2140. T. Blasko, J.M. Lorez Nieto, Appl. Catal. A , 1997,157, 117. F. Cavani, F. Trifiro, Appl. Catal., 1995,133,219. G. Centi, S. Perathoner, F. Trifiro, A. Aboukais, C.F. Aissi, M. Guelton, J. Phys. Chem., 1992,96,2617. H.H. Kung, M.C. Kung, Appl. Catal. A: General, 1997,157, 105. D.L. Stern, R. Grasselli, J. Catal., 1997, 167,550. D. Sam, V. Soenen, J.C. Volta, J. Catal., 1990, 123,417. L.E. Cadus, M.C. Abello, M.F. Gomez, J.B. Rivarola, Ind. Eng. Chem. Res., 1996,35, 14. L.E. Cadus, M.F. Gomez, M.C. Abello, Catal. Lett., 1997,43,229. Y.S. Yoon, W. Ueda, Y. Moro-oka, Catal. Lett., 1995,35, 57. K.H. Lee, Y.S. Yoon, W. Ueda, Y. Moro-oka, Catal. Lett., 1997,46,267. D.L. Stern, J.M. Michaels, L. DeCaul, R. Grasselli, Appl. Catal. A , 1997,153, 21. P. Viparelli, P. Ciambelli, L. Lisi, G. Ruoppolo, G. Russo, J.C. Volta, Appl. Catal. A , 1999,184,291. V. Ermini, E. Finocchio, S. Sechi, G. Busca, S. Rossigni, Appl. Catal. A: General, 2000, 198, 67. R.B. Watson, U.S. Ozkan, J. Catal., 2000, 191, 12. M.C. Abello, M.F. Gomez, L.E. Cadus, Catal. Lett., 1998,53, 185. A. Kaddori, C. Mazzocchia, E. Tempesti, Appl. Catal. A: General, 1998,169, L3. S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki, Chem. Lett., 1999, 25. O.V. Buyevskaya, D. Wolf, M. Baerns, Catal. Today, 2000,62,91. M. Baerns, O.V. Buyevskaya, P. Claus, U. Rodemerk, D. Wolf, DE 198 43 242 A, 1998. G. Centi, Appl. Catal. A , 1996,147,267. I.E. Wachs, B.M. Weckhuysen, Appl. Catal. A , 1997,157,67. A. Khodakov, B. Olthof, A.T. Bell, E. Iglesia, J. Catal., 1999, 181,205. K. Chen, E. Iglesia, A.T. Bell, J. Catal., 2000,192, 197. D. Creaser, B. Andersson, R.R. Hudgins, P.L. Silveston, Appl. Catal. A: General, 1999,187,147. A. Pantazidis, A. Burrows, C.J. Kiely, C. Mirodatos, J. Catal., 1998, 177, 325. A. Burrow, C.J. Kiely, J. Perregaard, P.E. H~Jlund-Nielsen,G. Vorbeck, J.J. Calvino, C. Lopez-Cartes, Catal. Lett., 1999,57, 121. M. Okamoto, L. Luo, J.A. Labinger, M.E. Davis, J. Catal. 2000,192, 128. M. Baerns, O.V. Buyevskaya, M. Kubik, G. Maiti, 0. Ovsitser, 0. Seel, Catal. Today, 1997,33,85. J.E. Miller, N.B. Jackson, L. Evans, A.G. Sault, M.M. Gonsales, Catal. Lett., 1999,58, 147. Y.S. Yoon, K. Suzuki, S. Hayakawa, T. Shishido, K. Takehira, Catal. Lett., 1999,59, 165. D.L. Stern, R. Grasselli, J. Catal., 1997,167, 560. J.N. Michaels, D.L. Stern, R.K. Grasselli, Catal. Lett., 1996,42, 139. J.N. Michaels, D.L. Stern, R.K. Grasselli, Catal. Lett., 1996,42, 135.

196

Catalysis

80 81 82 83 84

W. Zhang, X. Zhou, D. Tang, H. Wan, K. Tsai, Catal. Lett., 1994,23 103. C.-T. Au, W.-D. Zhang, J. Chem. SOC.Faraday Trans., 1997,93,1195. O.V. Buyevskaya, M. Baerns, Catal. Today, 1998,42,315. O.V. Buyevskaya, M. Kubik, M. Baerns, ACS Symp. Ser., 1996,638,155. O.V. Buyevskaya, D. Miiller, I. Pitsch, M. Baerns, Stud. Surf:Sci. Catal. 1998, 119, 671. K. Otsuka, Y. Uragami, T. Komatsu, M. Hatano, Stud. Surf:Sci. Catal., 1991, 61, 15. M. Huff, L.D. Schmidt, J. Catal., 1994,149, 127. A. Beretta, L. Piovesan, P. Forzatti, J. Catal., 1999,184,455. A. Beretta, P. Forzatti, E. Ranzi, J. Catal., 1999,184,469. A. Pantazidis, J.A. Dalmon, C. Mirodatos, Catal. Today, 1995,25,403. G . Capannelli, E. Garosini, F. Cavani, 0. Ponticelli, F. Trifiro, Chern. Eng. Sci., 1996,51,1817. M.J. Alfonso, A. Julbe, D. Farrusseng, M. Mentndez, J. Santamaria, J. Chern. Eng. Sci., 1999,54, 1265. R. Del Rosso, A. Kaddouri, R. Anouchinsky, C. Mazzocchia, P. Gronchi, P. Centola, J. Mol. Cal. A: Chemical, 1998,135, 181. R. Del Rosso, A. Kaddouri, D. Fumagalli, C. Mazzocchia, P. Gronchi, P. Centola, Catal. Lett., 1998,55,93. T. Imai, R.J. Schmidt, US 4 886 928,1989. P.A. Agaskar, R.K. Grasselli, J.M. Michaels, P.T. Reischman, D.L. Stern, J.G. Tsikoyiannis, US 5 563 314, 1996. C. Hallett, BP Chemicals, EP 0 480 594 A2, 1991. M. Kitson, BP Chemicals Ltd., US 5,210,293, 1993. N.C. Benkalowycz, P.R. Blum, D.R. Wagner, Standard Oil Company, US 5,300, 684,1994. P.R. Blum, M.A. Perera, Standard Oil Company, USP 5 300 682,1994. E. Bordes, L. Tessier, M. Gubelmann-Bonneau, Rhone-Poulenc Chimie, EP 683 153 Al, 1995. A. Aubry, A.-M. Le Govic, M. Gubelmann-Bonneau, Rhone-Poulenc Chimie, EP 719 756 Al, 1996. H. Borchert, U. Dingerdissen, J. Weiguny, DE 196 20 542 Al. 1996. H. Borchert, U. Dingerdissen, DE 196 30 832 Al, 1996. H. Borchert, U. Dingerdissen, R. Roesky, Hoechst Research & Technology Deutschland GmbH & CO. KG, WO 98/47851,1998. K. Karim, M.H. Al-Hazmi, E. Mamedov, Saudi Basic Industries Corporation, WO 99/13980, 1999. L. Tessier, E. Bordes, M. Gubelmann-Bonneau, Catal. Today, 1995,24,335. M. Roy, M. Gubelmann-Bonneau, H. Ponceblanc, J.-C. Volta, Catal. Lett., 1996,42,93. W. Ueda, N.F. Chen, K. Oshihara, Chem. Cornrn., 1999,517. (a) D. Linke, D. Wolf, M. Baerns, 0. Timpe, R. Schlogl, S. ZeyB, U. Dingerdissen, J. Catal., 2002, 205, 16; (b) D. Linke, D. Wolf, M. Baerns, S. ZeyD, U. Dingerdissen, J. Catal., 2002,205,32. D. Linke, PhD Thesis, Technical University of Berlin, 2001. D. Linke, D. Wolf, M. Baerns, S. ZeyD, U. Dingerdissen, L. Mleczko, Chern. Eng. Sci., 2002,57, 39. (a) K. Ruth, R. Kieffer, R. Burch, J. Catal., 1998, 175, 16; (b) K. Ruth, R. Kieffer, R. Burch, J. Catal., 1998,175,27.

85

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109

110 111 112

5: Oxidative Functionalization of Ethane and Propane 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

131 132 133 134 135 136 137

197

M. Ai, J. Catal., 1986,101, 389. M. Ai, Catal. Today, 1998,42,297. Y. Han, H. Wang, H. Cheng, R. Jin, J. Deng, New J. Chem., 1998,1175. N. Gribot-Perrin, J.-C. Volta, A. Burrows, C. Kiely, M. Gubelmann-Bonneau, Stud. Surf:Sci., 1996, 101, 1205. T. Quandt, PhD Thesis, Ruhr University Bochum, 1999. G. Centi, F. Trifiro, Catal. Sci. Technol., 1991,1,225. N. Mizuno, M. Tateishi, M. Iwamoto, Appl. Catal. A: General, 1995,128, L165. W. Ueda, Y. Suzuki, Chem. Lett., 1995,541. W. Li, K. Oshihara, W. Ueda, Appl. Catal. A: General, 1999,182,357. J.E. Lyons, P.E. Ellis Jr., S. Karmakar, S.N. Shaikh, EP 0 771 781 Al, 1996. Y.-F. Han, H.-M. Wang, H. Cheng, J.-F. Deng, Chem. Commun., 1999,521. M. Takahashi, S. To, T. Hirose, JP 10118491, 1996. G. Blanchard, G. Ferre, Rhone-Poulenc Chimie, EP 0 609 122 Al, 1994. T. Ushikubo, H. Nakamura, Y. Koyasu, S. Wajiki, Mitsubishi Kasei Corporation, EP 0 608 838 A2,1994. J.P. Bartek, A.M. Ebner, J.R. Brazdil, US 5 198 580, 1993. Y.-C. Kim, W. Ueda, Y. Moro-oka, Appl. Catal., 1991, 70, 175; Catal. Today, 1992, 13, 673. 0. Seel, PhD Thesis, Ruhr-Universitat Bochum, 1995. (a) Y. Takita, H. Yamashita, K. Moritaka, Chem. Lett., 1989, 1733; (b) I. Matsuura, N. Kimura, Stud. Surf: Sci. Catal., 1994, 82, 271; (c) J. Barrault, L. Magaud, Stud. Surf: Sci. Catal., 1994, 82, 305; (d) C. Mazzochia, C. Aboumrad, C. Diagne, E. Tempesti, C. Herrmann, G. Thomas, Catal. Lett., 1991, 181; (e) Y.C. Kim, W. Ueda, Y. Moro-oka, Chem. Lett., 1989, 531; (f) W.C. Conner, S. Soled, A. Signorelli, Stud. Surf:Sci. Catal., 1980, 7, 1224; (8) N. Giordano, J.C.J. Bart, P. Vitarelli, S. Cavallaro, Oxid. Commun., 1984, 7 , 99; (h) Y.C. Kim, W. Ueda, Y. Moro-oka, Appl. Catal., 1991,70,175. A. Tenten, T. Proll, H.-P. Schildberg, BASF, DE 196 22 331 Al, 1996. H. Jachow, A. Tenten, S. Unverricht, H. Arnold, BASF, DE 197 46 210 Al, 1997. A. Kaddouri, C. Mazzocchia, E. Tempesti, Appl. Catal. A: General, 1999, 180, 271. K. Otsuka, R. Takahashi, I. Yamanaka, J. Catal., 1999,185, 182. A.O. Kuzmin, E.R. Savinova, V.N. Parmon, React. Kinet. Catal. Lett., 1999, 66, 351. Ullmann’s Encyclopedia of Industrial Chemistry, 6th Edn., 2000 Electronic Release, Ch. 5.1.3. M. Lin, M.W. Linsen, Rohm and Haas Comp., US 6 180 825,2001.

6 Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation BY WATARU UEDA AND SUI WEN LIN

1

Introduction

Interest in the potential use of alkanes as sources of the corresponding alkenes or their derivatives is increasing year by year. A great deal of research has already been done to achieve selective functionalization of light alkanes, although it is well known that these hydrocarbons are unreactive because there are no lone pairs of electrons, no empty orbitals and little polarity of the C-H bonds. Among various functionalization methods, catalytic selective oxidation has received much a t t e n t i ~ n . ”The ~ oxidation of light alkanes is usually accompanied by many difficulties as a result of the low reactivity of the reactant. The low reactivity normally requires more severe reaction conditions than in the case of olefin oxidation, so that the subsequent oxidation of the partially oxidized products unavoidably takes place simultaneously, resulting in low selectivity to the desirable product^.^ This situation is most significant in developing heterogeneous oxidation catalyst^.^ This is also true for some olefin catalytic oxidations, even though some processes have been commercialized. For instance, catalytic oxidation of propylene to propylene oxide with molecular oxygen is a highly difficult r e a c t i ~ nOxidation .~ of benzene to phenol also has the same problems.6 It is believed that, to optimize these processes selectively, precise control of molecular oxygen activation and reactant activation is necessary. At the same time catalysts must be designed not to be active for further oxidation of desired products. One of the ways to overcome this situation is to develop a new system of solid catalysts where catalytic functions controlling multielementary steps of oxidations are well-organized in the catalyst structure or on the catalyst ~ u r f a c e . ~ The most well known case in this respect is the selective oxidation of n-butane to maleic anhydride, which has been successfully established using crystalline V-P-0 complex oxide catalysts.8 It is well known that no other mixed oxides show performance comparable to (VO)2P2O7. It is also well established that the crystal face (100) is selective for the formation of maleic anhydride, while the side faces such as (110) are active for non-selective o ~ i d a t i o nThis . ~ clearly means that each crystal face displays a surface crystal field which strongly controls the transformation of the reactant to a given Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 198

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

199

product, although the oxide catalysts are metastable under reductionoxidation conditions and have to be maintained by bulk properties. The other case is the selective oxidation of ethane to ethene and/or acetic acid.12 The most prominent catalyst for this reaction is that of Thorsteinson et al. l 3 They found that molybdenum-vanadium-based oxide catalysts with a layer structure with d-spacing at 4.0 were highly active for ethane oxidative dehydrogenation at temperatures as low as 300 "C. Particularly, the catalyst containing niobium showed the highest activity and almost 100% selectivity to ethene. The catalyst was also active for acetic acid formation in the oxidation of ethane at low temperatures.13 The basic methodology of catalyst design for acetic acid formation from ethane in this catalyst system seems to be simple.14-1 Since the molybdenum-vanadium oxide phase having the XRD diffraction of d-spacing at 4.0 is extremely selective for ethene formation, the main design question is how to introduce a catalytic functional component to promote further immediate oxidation of ethene to acetic acid without diminishing the principle activity of the molybdenum-vanadium oxide catalyst. Obviously Pd and Sn-Mo-0,17 which are well-known catalysts for Wacker oxidation and oxidative dehydration, respectively, play this role in the alkane oxidation. In fact, these functional components clearly accelerated the formation of acetic acid to some extent,1416but the yield is still very low, so that undoubtedly the contribution of these coexisting functional components is still not sufficient. For the next stage of the catalyst design, very close and regular arrangement of each catalytic functional component in the level of crystal structure or in atomic level seems to be highly important. l8 If the well-arranged components is isolated s t r ~ c t u r a l l yby ~ ~other components and can complete one cycle of catalytic selective oxidation, the resulting catalyst would be highly selective. In this sense, the Mo-V-Te-NbO oxide system seems to be the most successful example. This catalyst was disclosed by Mitsubishi Chemicals in their patents as an extremely active and selective catalyst for the ammoxidation of propane to acrylonitrile20-22and the oxidation to acrylic acid.23 The Te component, which is often involved directly or indirectly in the selective hydrocarbon oxidation,24is incorporated in the isolated active site in a suitable way to promote selective oxidation of propane. Along these same lines, the choice of heteroply corn pound^^^-^^ and crystalline V-P-08-' as a catalyst for alkane oxidation is quite natural. Above all, it should be emphasized that artificial or self organization of catalytic components in a particular structure during catalyst preparation is highly important in creating new types of active sites which are inevitably necessary for highly difficult selective oxidation. However, it is not easy to prepare these catalyst materials by normal catalyst preparation methods, which often result in the formation of complicated multi-phasic materials.29 The preparation of the above Mo-V-Nb-O and Mo-V-Te-NbO catalysts is not exceptional; it appears that very careful controls should be made in the whole preparation procedure.30Although it is impossible to avoid the effect of

A

A

Catalysis

200

unknown factors completely, inorganic synthesis, such as sol-gel methods, microemulsion methods, hydrothermal methods etc., should be applied in many cases to generate the local structural arrangement of elements in complex metal oxide catalysts instead of conventional dry up methods. For local bond structure control of active sites, precise elemental substitution or introduction into active sites has to be done while monitoring structure and catalytic activity. In this chapter, we focus on the remarkable catalytic functions of halogens in halogen-containing solid-state catalysts active for the selective oxidation of light alkanes. It is very well known that halogens markedly enhance catalytic activity and selectivity for selective oxidations. Halogens are indispensable components of catalysts for some oxidation processes which are commercially utilized, although halogen causes problems in reaction system and reaction facilities. We review here recent developments of halogen-containing solidstate catalysts, particularly for selective oxidations, and also show the possibility of catalyst tuning in the case of layered metal chloride oxides active for light alkane oxidations. 2

Gas-phase Selective Oxidation in Halogen-containingSystems

Typical examples of halogen-containing catalyst systems active for hydrocarbon oxidations in gas-phase are listed in Table 1. As mentioned above, various types of halogen-containing solid-state catalysts and halogen-added catalytic systems are used for many kinds of reactions. Particularly, the catalysts are frequently used or tested for selective oxidation of hydrocarbons. The main reason is that only a small amount of addition of halogen in either solid catalysts or in the reactants results in a remarkable selectivity increase to a desired oxidation product. In many cases, particularly alkane oxidations, catalytic activity greatly increases, although halogens, if present in large amounts, damage reaction facilities severely. 2.1 Epoxidation. - The most well known reaction is the epoxidation of ethylene to ethylene oxide over alumina supported silver catalysts.62Metallic silver is the key catalytic element in this reaction but the system always needs the addition of organic chloride into the reactant feed in order to increase the selectivity to ethylene oxide. It is believed that the added chloride compounds decomposes on the silver metal surface and the chlorine on the silver surface is self-organized into a surface structure that affects the adsorption and activation of molecular oxygen on silver atoms.63Several possible roles of chlorine in enhancing ethylene epoxidation have been suggested so far. One is that chlorine blocks specific surface sites, specifically positively-charged silver surface, at which total oxidation takes A second hypothesis is that enhanced selectivity can be attributed to high surface oxygen concentration since surface chlorine roughens the s ~ r f a c e . Self-organization, ~~.~~ therefore, seems to occur along with surface restr~cturing.~~ Electronic effects have also

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

Table 1

20 1

Various gas-phase selective oxidations in halogen-containing catalyst systems

React ion

Epoxidation

Oxidation

Catalyst system

Re$

Propylene to propylene oxide Butadiene to epoxybutene

Ag/A1203+ C2H4C12 Ag-Na-Cs-CYAlzO3 Ag/NaCl Ag-CsCVA1203

32 33 34 35

Anthracene to anthraquinone

V20&sC1

36

Ethylene to ethylene oxide

Wacker oxidation Ethylene to acetoaldehyde Ethylene to vinyl acetate

PdClz+CuCl2/activecatbon 37 PdCl2+CuCl2/activecatbon 38

Oxycarbonylation Methanol to methyl carbonate Pd-Cl/C

+ HCl

39 40

Oxycyanation

Ethylene to acrylonitrile

Pd-V-C~/A1203 + HC1

Ammoxidation

Propane to acrylonitrile

Bi-M&

Methane to ethylene

LiCYMnOx LiCl-MgO LiC1-NiO NaBi304C12 Ln203+CCL NaC1-ZrO2 BiOCYLiC03/Mg0

42 43 44 45 46 47 48

LiCl-MgO KSr2Bi304Cls SrFe03- xC1x LiCYsulfated Zr02 SrC12/Ln03 SmOF Ln2O3+CCL CeF3/Ce0, LiCYMnO, Pd-KBr/A120 3

49 50 51 52 53 54 55 56 57 58

KBr/SiO2

59

BiOCl

60

CuC12-KCYA1203

61

Oxidative dimerization

Oxidative Ethane to ethylene dehydrogenation

Propane to propylene Ethyl benzene to stylene Oxidative methylation

Methane + acetonitrile to acrylonitrile

Oxidative cracking Butane to C2,C3 Oxychlorination

Ethylene to ethylene chloride

+ organic chloride 41

been proposed in which surface chlorine affects the ethylene oxide-surface bond by modifying an electronic state of atomic oxygen species.71 The gas phase epoxidation of ethylene using molecular oxygen to produce ethylene oxide is one of the most successful examples of heterogeneous catalysis to date. In addition to this, the Eastman Chemical Company recently commercialized the Oxirane process to produce e p o ~ y b u t e n eIn .~~ this process alkali metal chloride promoted silver catalysts were used, They found that the addition of an optimum level of CsCl increases activity and selectivity. It appears that metallic silver and CsCl create an active site together by

202

Catalysis

organization, which promotes proper activation of molecular oxygen and prompt desorption of the epoxy ~ l e f i n . ~ ~ While the importance of olefin epoxidation is apparent, efforts to selectively epoxidize higher olefins to their corresponding epoxides using molecular oxygen have been unsuccessful.72 In particular the epoxidation of olefins having allylic C-H was highly difficult because of the lower dissociation energy of this bond than that of the vinylic C-H bond. Nevertheless, there is one interesting topic in catalytic epoxidation in which the remarkable function of halogen for enhancing epoxide selectivity is applied in the development of catalysts; that is epoxidation of propylene using silver catalyst supported alkali chloride.34 Recently, the ARC0 Chemical Company reported that at 25% conversion of propylene, 45% selectivity to propylene oxide was achieved over supported silver catalyst containing inorganic chlorides and other promoters.73 Chlorine is undoubtedly involving in this system and therefore further precise control of chlorine state in the catalyst will bring about a new stage of gasphase epoxidation of olefins with molecular oxygen.

Redox System. - Metal halides are frequently used in liquid phase oxidation, for instance PdCl2-CuC12 in the Wacker process and CoBr2 in the Amoco process. For many years, heterogenation of these catalysts has been attempted by using conventional inorganic inert supports to avoid corrosive halogen. However, the earliest examples listed in Table 1 for Wacker type oxidations contained chlorine because chlorine was needed to assist redox coupling between Pd and Cu on an active carbon s ~ p p o r t . Current ~~?~~ heterogeneous oxyanionization catalysts based on Pd are constructed without halogens except oxychlorination catalysts. For oxycarbonylation of alcohol to alkyl carbonates, Pd-Cl supported on active carbon is known to be effective.39 It is proposed in this system that chlorine is an important ligand for Pd, stabilizing an oxidized state of Pd and then controlling CO activation. 2.2

2.3 Gas-phase Halogen Promoters. - The next important example utilizing halogen as catalyst is the ammoxidation of propane to acrylonitrile in the presence of gas-phase halide compounds as promoters, such as methyl bromide and hydrogen bromide. Solid catalysts used for the halogen co-feed process are those widely used for propene ammoxidation, such as multicomponent bismuth molybdate and uranium antimonate catalysts.7676 The addition of small amounts (< 1 mol%) of the halogen promoter in the reaction feed increases not only the conversion of propane but also the selectivity to a~rylonitrile,~' resulting in high acrylonitirle yield compared with the various ammoxidation processes using non-halogen catalyst systems. A key role of the halogen compound in the feed is to promote the radical reactions of propane to propene in the gas phase, since the dehydrogenation of propane to propene takes place in the presence of halogen promoter even without the catalysts. The solid catalyst in this process plays the role of selectively oxidizing the propene formed by the gas-phase radical reaction to acrylonitrile. In other words, the process is a combined system of gas-phase reaction promoted by 377

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

203

halogen and surface reaction over oxide catalyst. However, Burylin has reported that halogen added in the feed changes the surface of the metal oxide catalyst, facilitating dehydrogenation of propane over the surface and formation of an allylic intermediate, promoting the reaction of ammonia over the surface to the products, and thus resulting in an increase in the conversion and ~electivity.~~ It seems also to be important that the halogen modifies the catalyst properties. A logical explanation of this concept is a catalyst that contains metal halide as a catalyst c~nstituent.~’Although the catalytic performance is quite poor at present, this catalyst system will be more important if the activity is increased and its stability is improved by structural design of the solid. The effect of adding a halogen promoter is significant and brings about relatively higher yield of acrylonitrile, and so commercialization of the propane ammoxidation process using halogen promoters had been planned. However, the plan was abandoned even after extensive research because of the low space time yield, which is strongly dependent on the dehydrogenation of propane in the gas phase, and due to severe damage to the reactor with halogen during the reaction. 2.4 Oxidative Reaction of Alkanes. - In relation to the alkane selective oxidation, there are many examples of halogen-containing catalysts, most of which were tested for oxidative dehydrogenation of light alkanes. This is simply because the oxidation of alkanes starts from dissociation of highly stable C-H bond and the catalyst does not need properties for oxygen insertion to oxidized intermediates on the surface. For C-H dissociation, catalysts must contain sites having a strong hydrogen abstracting ability. Coordinatively unsaturated metal cation sites, strongly basic oxygen species and hydrogen acceptor like halogens and sulfur can provide this catalytic function.80*81 Halogens as hydrogen acceptor are particularly effective for oxidative alkane dehydrogenation. The pronounced effect of the halogen is easily observed by the addition of halogen into the system, so that a variety of solid-state catalysts containing halogens have been tested for oxidative reaction of alkane^.^^.^'.^^ The basic chemistry of the oxidative dehydrogenation with halogen is as follows: If gaseous halogen (X2) is added in a reactant feed containing a light alkane and oxygen, the halogen abstracts hydrogen from alkane at high temperature via halogen radical pathway without catalytic materials to form olefin and hydrogen halide (equation 1):

CnHzn+2 + X2

MO MX2

+ 2HX + $02

-+

CnH2, + 2HX

(1)

-+

MX2

+ H20

(2)

+

MO

+ X2

(3)

CnH2

+ H20

If metal oxides of alkali, alkali earth, rare earth or some transition metals are loaded in the gas-phase reaction zone, the oxides readily react with the

204

Catalysis

hydrogen halide to form metal halide and water (equation 2). Immediately after this reaction, the metal halide reacts with molecular oxygen to regenerate the metal oxide and halogen (equation 3). The net reaction is the oxidative dehydrogenation of alkane with molecular oxygen, so that a catalytic amount of the halogen compound is enough to achieve these three sequential reactions. In order to prevent complete oxidation with molecular oxygen activated over the metal oxide surface, the metal oxides having poor oxygen activation ability, like those mentioned above, are normally effective for the reaction. The addition of halogen sources into the reaction system is done not only by the use of gaseous compounds containing halogen but also by introducing metal halide compounds into the metal oxides. The latter are halogen-containing metal oxides or halide oxide catalysts which are variously reported to be a highly efficient catalyst for alkane oxidation processes listed in Table 1. In the solid catalyst system, the halogen is not necessarily involved in a gas-phase reaction after liberation from the oxide surface into the gas-phase. Although it is believed that the halogen mostly takes part in the catalytic oxidation cycle while remaining on the metal oxide surface, the chemistry of the halogencontaining system in metal oxides is not so simple. More details will be given in the following section. 3

Catalytic Properties of Halogen-containingMetal Oxides

3.1 Oxidative Coupling of Methane. - After the pioneering works of Keller and Bhasin in 1982,82oxidative coupling of methane to C2 hydrocarbons has received the attention of scientists and has been shown to occur on many different types of solid catalysts at high temperature^.^^.^^ During continuing research on developing highly effective catalysts, Otsuka et al. reported for the first time that several LiCl/transition metal oxides were particularly active for the reaction and characteristically produced ethylene with higher yield than As described in the preceding section, halogen is known to effect ethane.44.85*86 the dehydrogenation of alkanes like ethane in the gas phase, so that it was suspected that homogeneous reactions may be responsible for the large ethylene/ethane ratios. Even though the possibility of the homogeneous reaction in addition to very short lifetimes of this type of catalyst were negative factors, various halogen-containing catalysts were extensively examined because of their high catalytic efficiency. Soon after Otsuka’s work, we reported that complex bismuth halide oxide catalysts having layered structures also showed high catalytic performance in the oxidative coupling of methane.45*87*88 The details of the catalysts and their catalytic performance design will be given in the next section, but briefly LiCa2Bi30&, X3 catalysts, in which the cation-oxygen sheets alternate with triple chlorine layers which contain Ca2+ion in their interstices, showed about 20% yield of C2 compounds with ethylene/ethane ratio of 25, which is extremely high compared with non-halogen catalysts.88Similarly, Khan and Ruckenstein found BiOCYLi2C03/Mg0materials to be active and selective.48

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

205

Neither BiOCl nor BiOCUMgO was very effective, but the Li-added catalyst to BiOCUMgO gave 18% conversion of methane and 83% selectivity to C2 in which the ethylenelethane ratio was 2.9. Analyses of the catalyst suggested that Li stabilizes the chlorine on the surface. These results clearly demonstrate high catalytic performance of alkali-modified bismuth chloride oxides for the oxidative coupling of methane. The effectiveness of metal chloride oxides as catalysts was also supported in the halogen-containing solid catalyst studied by Burch et al.42They concluded that the improved C2 selectivity results from the formation of a manganese chloride oxide on the surface of alkali chlorideadded Mn02 catalysts. There are also many examples of simple halogen-containing oxides active for the methane oxidative coupling. The most prominent catalysts for the reaction are Li+-MgO-Cl- and Na+-Zr02-C1-. The Li+-MgO-C1 - catalyst prepared by the sol-gel method was found to be highly stable and active for the reaction at 640°C, yielding up to 20% with ethylene/ethane ratios of ca. 5.43 Similarly, it was found that the Na+-Zr02-C1- catalyst prepared by the solgel method showed 14% conversion, 77% selectivity and ethylene/ethane ratios of ca. 3.3 at 750 0C.47 As described above, the influence of halogen in the oxidation catalyst systems is prominent, but the precise role of halogen is not fully established yet. One of the significant features of the halogen-containing catalysts was the high ethylene/ethane ratio. With oxide catalysts, the major C2 product is usually ethane, whereas with chlorine-containingcatalysts substantial amounts of ethene were produced. To explain this point, the mechanism of ethene formation over chloride-containing catalysts has been described in terms of a purely gas-phase ethane dehydrogenation brought about by CH3Cl or HCl released from the surface of the catalyst during the methane coupling reaction. However, Burch et al. demonstrated that this is not correct.89Although some CH3Cl is detected in the gas phase, this compound was proven to have no influence on the conversion of ethane to ethene in the gas phase. On the other hand HCl significantly promoted the conversion of ethane to ethylene in a gasphase process, However, in the case of HCl, 1.5% HCl concentration in the feed was required to obtain a better ethylene yield, which means that HCl is not involved in the gas-phase reaction because such a high concentration of HCl is never achieved under normal reaction conditions using chlorinecontaining catalysts. More recent workg0 by Burch et al. using a transient technique showed that there is no detectable contribution of gas-phase chlorine at concentrations of ca. 25 ppm, which is a more plausible concentration. They concluded that an important role of chlorine is to produce new sites that are particularly effective in the activation of ethane. Many other data supported this c o n c l ~ s i o n . ~Lambert ~ - ~ ~ and co-workers have also concluded from the study on the positive role of chlorine in Mn203that the greater C2 selectivity should be attributed either to site modification or the formation of active chlorine radicals on the surface.95 Lunsford et al. reported other points on the role of halogen,96on the basis of the research in which the role of C1- ions was evaluated in detail by

Catalysis

206

P(%)O-*, Torr

0

10

20

P(%)Oo6, Torr Figure 1

Order of CH4 oxidation with respect to 0 2 over the Li+-MgO (square) and Li* -MgO-Cl- (circle) catalysts at 650 "Cwithout (open) and with addition (filled) of 6 Torr of CO,

independently varying the concentration of Li+ and C1- ions in LilMgO catalysts. With the addition of C1- ions, the intrinsic rate of methane activation decreases, but the catalysts are less subject to poisoning by C 0 2 (Figure 1). Moreover, Cl ions inhibits extensive sintering of the catalysts. As a result, the Li+-MgO-Cl - catalyst gave greater methane conversions than the conventional LilMgO catalysts under normal operating conditions. 3.2 Oxidative Dehydrogenation of Ethane. - Oxidative dehydrogenation of alkanes with oxygen over solid catalysts often results in the formation of carbon oxides which are thermodynamically much more favored, decreasing the selectivity toward olefins which are normally more reactive than the starting alkanes. Therefore, this is a limit to the potential use of oxidative dehydrogenation. However, the oxidative dehydrogenation of ethane to ethylene is exceptional because the stability of ethylene against further oxidation is extremely high as ethylene has only the vinylic C-H bond, which is very strong, so that it is relatively easy to attain high selectivity to ethylene in the oxidative dehydrogenation of ethane compared with other alkanes.97 In fact, there have been many reports on the oxidative dehydrogenation of ethane and a large number of catalysts have shown high selectivity to ethylene. Metal oxide-based catalysts developed so far can be classified into two types. One is generally molybdeum- or vanadium-based oxides and metal oxide

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

207

promoted Mo-V-0 catalysts, and the other is alkali or alkali-earth metal oxide-containing metal oxides.lm The former are active at lower temperature, while the working temperature for the latter catalysts is relatively high. Furthermore, the latter catalysts can be modified with halogens to be more selective. As a typical example of the second group, a lithium-promoted magnesium oxide catalyst and a Li+-MgO-Cl- catalyst have been reported by are highly efficient for the reaction. Lunsford and c o - w ~ r k e r slol . ~ ~They , Particularly, Li+-MgO-Cl- has high selectivity to ethylene and it was demonstrated that the addition of chloride ions to a Li+-MgO catalyst significantly increased the yield of ethylene. This promotional effect of chloride ions coincides well with the observation that ethene/ethane product ratio obtained in the oxidative coupling of methane was much higher than those obtained over a chloride-free Li+-MgO catalyst at comparable levels of methane oxidative c o n ~ e r s i o n . " ~Moreover, *~~ as pointed out by Burch and Crabb, lo2 among various oxide based catalysts that they considered for the oxidative dehydrogenation of ethane, only Li+-MgO-Cl- exhibited ethene yields which exceeded those attained in a non-catalytic reaction under optimum conditions. One of the more significant features of the Li+-MgO-Cl- catalyst is that the activity is sufficiently high, so that the catalyst can be used at the reaction temperature below 650°C. This means that the loss of chlorine from the catalyst during the catalysis should be slow because of the lower reaction temperature, and that the contribution of the homogeneous gas-phase reaction of ethane and the secondary homogeneous oxidation of the produced ethene, assisted by gas-phase halogen, can be minimized, if it exists. For this catalyst system, it is possible to be more active without sacrificing selectivity by the addition of lanthanum oxides, so that the oxidative dehydrogenation of ethane can be carried out at the temperatures as low as 570 OC.lo3 Other examples of supported alkali halide catalysts include LiC1-promoted Zr02 and Nd203-doped LiChulfated lo4 The latter showed remarkable activity for the oxidative dehydrogenation of ethane, giving a 93% ethane conversion and about 83% selectivity to ethylene at 650 "C. Rare-earth metal oxides are also used as the supports for metal halides. Au et al. studied this type of catalyst extensively, specifically MX2 (M = Sr, Ba; X = F, C1, Br)/Ln203 (Ln = La, Nd, Sm, Ho, Metal halide doping is highly effective in promoting the catalysts and can significantly enhance both ethane conversion and ethylene selectivity. The activity enhancement may result from the introduction of chloride ions into the oxides, which prevents C02 poisoning of surface active sits constructed with metal and oxygen anion. The most striking point is that the halogen-containing catalysts are inherently inactive for the oxidation of ethylene, which is the reason for the high ethylene selectivity of the catalysts. This might be explained by the decreasing of the reactivity of adsorbed oxygen species by the action of chlorine on surface and bulk properties. In relation to the oxygen activation, Au's group has reported interesting results.51 They investigated catalytic performance of perovskitetype SrFe03 -o.19 and SrFe03-0.382C10.443and reported that the chloride ion implanted analogue showed a higher ethane conversion of ca. 90% with ca. 13920998399

Y).539105-113

208

Table2 Catalyst

SrFe03-6 SrFeO3-&1,

Catalysis

Oxidative dehydrogenation of ethane to ethylene at 680°C over SrFeO3-6 and SrFe03-o catalysts Compositions 6 Fe4+lFe Cl (“Yo) (“Yo)

62.1 68.0

-

7.81

0.190 0.382

0

-

0.443

Ethane Ethylene Yield conversion selectivity of ethylene (%) (”/.I (“Yo) 54.8 89.9

50.5 73.0

27.7 62.2

70% selectivity to ethylene at 680°C (Table 2). The inclusion of chloride ions in the SrFe03-d can change the concentration of oxygen vacancies and the oxidation state of Fe, which causes the nature of adsorbed oxygen species to be less reactive and suitable for the selective formation of ethylene. It has been demonstrated that the activity of an oxide catalyst for the oxidative dehydrogenation of ethane can be preferentially increased by treating it with organic chloride. For example, when a Co-Zr-P-Na-K oxide catalyst was treated with CH3C1, an ethene selectivity of 85% at 55% ethane conversion was obtained at 675”C, compared with 74% selectivity to ethene at 32% conversion of ethane on the same catalyst without the chloride treatment.l14 Sugiyama et al. reported the oxidative dehydrogenation of ethane in the presence and absence of tetrachloromethane (TCM) over various rare-earth oxides.55They observed that the conversion and the selectivity to ethylene on Ce02 were distinctly increased by the introduction of TCM in the feed. XPS measurements provided a evidence for the formation of CeOCl surface compounds by the introduction of C1 from gas to surface, suggesting that CeOCl is responsible for the enhanced catalytic performance. Chloride oxides like LaOCl and BiOCl were found to be extremely active and selective for the reaction after modification with alkali elements.50 Details of this catalyst group will be shown in Section 4. Fluorine-containing catalysts are also reported in addition to chlorinecontaining materials. Basically fluorine does not have hydrogen accepting ability, so that the introduction of fluorine into metal oxides has effects mainly on the catalytic function of molecular oxygen activation. The fact that the fluoride ion has nearly the same ionic radius as the oxide ion supports this conclusion. This is completely different from the chlorine-containingcatalysts. BaF2-LaOF catalysts were investigated for the oxidative dehydrogenation of ethane? It was found that BaF2-LaOF was more effective than LaOF for the catalytic conversion of ethane to ethene. A selectivity of 74% for ethene was obtained at 55% ethane conversion over 8 mol% BaF2-LaOF compared with a selectivity of 58% for ethene at 44% ethane conversion over LaOF under the same reaction conditions: reaction temperature 660 “C, C2H6:02 = 67.7:32.3and a feed gas flow rate of 90 ml min- l. Several papers related to the promotion effect of alkaline earth metal fluorides supported on a number of metal oxide catalysts, e.g. Ce02,116J17

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

209

Tho2,''* La20311' and Zr02,120 for the selective oxidation of methane and/or ethane have been published. The addition of alkaline-earth metal fluoride has apparently improved the oxidation selectivity of the catalysts. It has been suggested that the presence of alkaline earth metal fluorides in the Ce02/BaF2, Ce02/CaF2, Ti02/BaF2 and Th02/LaF3 catalysts facilitated the activation of molecular oxygen. Because of ionic exchange or substitution in the lattice, active centers are generated and the adsorption of O2 on these centers would result in the production of 0-, 0 2 2 - , and 02- species, which might be responsible for the oxidation. Au et al.lo5 reported that promotion of rhombohedra1 SmOF by SrF2 or BaF2 led to significant gain in ethane conversion but little change in ethene selectivity in the oxidative dehydrogenation of ethane. At 893 K, ethane conversion and ethene selectivity over SmOF were 38.3% and 71.2%, respectively. With 10% SrF2/SmOF and 20% BaF2/SmOF catalysts, ethane conversion became 69.7% and 68.4%, respectively, while ethene selectivity was 68%. It is natural to think that active oxygen species generated by the adsorption of molecular oxygen on fluorine-containing metal oxide catalysts determine the catalytic activity and the distribution of the products, because fluorine cannot directly activate light alkanes via radical process as chlorine or bromine can do. Therefore, 0- is believed to be responsible for the activation and selective oxidation of ethane. The role of fluorine in the catalyst is not clear at present, but undoubtedly metal oxides are modified with fluorine to be able to induce the formation of a particular type of oxygen species and hence be capable of fulfilling its specific role in a catalytic reaction. This seems to be a completely different point from chlorine-containing catalysts.

3.3 Oxidative Dehydrogenation of C3+.- Compared with extensive research on propane oxidative dehydrogenation using metal oxide catalsyts, there have been few studies with respect to halogen-containing catalysts. Since propylene, which is the product of propane oxidation, has allylic hydrogen of low C-H dissociation energy,97it is natural that propane oxidative dehydrogenation is more difficult than ethane oxidation in which the ethylene formed is highly stable against further oxidation. In fact, our unpublished data125 revealed that cracking products like ethylene were formed in large amounts, along with propylene over complex layered chloride oxide catalysts. Recently, a BiOCl catalyst was reported to show high performance for oxidative cracking of n-butane to lower olefimm BiOCl has particularly high olefin selectivity and suppresses deep oxidation. High activity for oxidative dehydrogenation and low activity for complete oxidation, which are characteristic properties of the halogen-containing catalysts, are shown in the oxidative dehydrogenation of C3+.However, ethylene and butadiene are major products at high conversion conditions for n-butane oxidation and propylene and n-butenes become minor products, so that the product selectivity is determined by the products' stabilities or reactivities. This means that even the catalysts containing halogens cannot suppress C-C bond fission of C3+ compounds to carbon oxides under oxidative conditions. This is also true for metal oxide catalysts on 12~573122-124

Catalysis

210

0 Mo-Mg-0 0 0

I

10

I

20

I

I

I

I

30 40 5 0 6 0 7 0

Conversion /% Figure 2

Oxidative dehydrogenation of propane

which propylene selectivity decreases markedly with an increase of propane conversion. 124 There is, however, a different catalyst, which is CeF3 added to Ce02 oxides. Wan et al. reported56 that this catalyst showed exceptionally high yields of propylene in the oxidative dehydrogenation of propane, as can be seen in Figure 2. Ce02 itself is very poor for propylene formation, catalysing mostly the cracking reaction to methane, ethylene and ultimately carbon oxides. The oxidation over CeF3 was negligible. However, when the two materials were combined, the resulting catalyst possesses a surprisingly high activity for the reaction. The catalyst achieved 8 1% selectivity to propylene at 41%conversion. Wan et al. concluded that fluoride ions introduced in the lattice by the ionexchange with lattice oxygen can isolate active oxygen species on the surface and modify the reactivity of this oxygen species, which result in high selectivity to propylene and even high propane conversion. Further advances in this catalyst system are highly desirable. 4

Structural Design of Layered Metal Chloride Oxide Catalysts

Often, a structural metal halide oxide is formed in halogen-containing solidstate catalysts after heat-treatments or during catalytic The structural metal halide oxide phase is stable under catalytic oxidation conditions, so that the formation of this type of compound can be easily observed by XRD. For example, Au et al. recognized the formation of a chloride oxide phase of rare-earth elements in metal chloride doped rare-earth oxide catalysts.126Even though the compounds are unstable and not detectable by XRD, it can normally be assumed that a quasi-crystalline phase of metal halide oxide is formed on the surface of the catalysts, or that the halogen on the surface strongly resembles the structure of a metal halide oxide. Because the surface

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

211

halogen, either in a structural form or in a non-structural form, plays an important role in catalytic oxidation, fundamental studies using crystallized metal halide oxides are useful for understanding the catalytic effect of halogen in halogen-containing catalysts. At the same time, such a fundamental approach will allow the design of catalysts because these materials are usually structurally simple compounds. In this section, structural information about layered metal halide oxides is summarized first and then the catalytic role of the halogen in alkane selective oxidation is discussed from data obtained in experiments on structural and electronic modification of halogen in layer structures. 4.1 Structural Background. - Metal halide oxides are of interest since they exhibit a variety of interesting properties, for example luminescence 127 and selective oxidation ~ a t a l y s i s . ~Recently ~ - ' ~ ~ interest has intensified because of the reports of high-temperature superconductivity in S ~ ~ C U O and ~ F ~ ' ~ ~ (Ca,Na)2Cu02C12.130 Many of these structures are based on layered intergrowths of oxide- and halide-containing blocks, which may be stacked in varying sequences to produce structures with increasing complexity. Here, the structures of metal halide oxide compounds will be introduced for the sake of understanding catalytic properties from a structural point of view. It is now more than fifty years since Sillen and co-workers synthesized and studied the complex bismuth halide oxides for the first time.131-140 These materials were the parents of a whole family of phases with a layered structure. Naturally, this family has subsequently been greatly extended, both in the elemental composition of the substances and in the configuration of the structural elements. The metal halide oxides are a large group of layered compounds containing cation-oxygen layers alternating with single or multiple thickness sheets of halide ions. Various structure models are summarized in Figure 3. Most of the structures in this figure are based on compounds of bismuth chloride oxides. Some details of these structures have not been finally determined, but the principle of the construction of these phases has been reliably established. Their principal structural characteristic, according to Sillen, is the presence of planar metal-oxygen layers separated by a planar layer (or layers) of halogen, parallel to one of the coordinate planes. It is very convenient to describe these structures of metal halide oxides in terms of the packing of spheres. They are based on planar square networks formed by close-packed oxygen ions. Metal ions lie in the depressions between four oxygen ions, and are situated above and below the oxygen network. The pattern of the metal-oxygen mosaic is a square unit, whose composition is made up of two oxygen atoms and two metal atoms, [M202]. An interesting feature is that [M202]layers correspond exactly to the tetragonal form of PbO (and SnO),14c142but in this case the space between these layers contains no 'filler'. This homologous series is considered as the 'zeroth' term (XO).The large halogen ions are situated in layers between the metal-oxygen layers, the cavities between four metal ions containing one halogen ion. This phase is

Catalysis

212

(c) :Metal

0 :o

0c1 :

Figure 3

Structure of thephases: (a) XI (LiBi3OdCl); (b) X, (BiOCl); (c) Xj (CdI.IBi1.602C13)

called the Sillen phase. The halide layers may also contain additional cations in their interstices. Large cations such as Pb2+ and Ca2’ often occur in place of Bi3+and play the same structural role as that ion. Within this family, one type of structure differs from another only in the thickness of the halogen fragment, which may be one-, two-, or three-layer (Figure 3). They are denoted by the symbols XI, X2, X3, where the subscript indicates the number of intermediate halogen layers. The XI type, exemplified by LiBi304C12,is shown in Figure 3(a). It consists essentially of layers of [Bi2O2I2+between which are regularly intercalated layers of halide ions. The X2 structure is that of BiOCl [Figure 3(b)] and the salient features of X3 are shown in Figure 3(c). The essential characteristics of these structural phases is that the ‘anionic’ layers are monoatomic, continuous, and planar and are classified in Table 3. The halogen ions (X: F, C1, Br and I) exist between the identical [M202] sheets and assume the same structures. Thus the symbols XI, X2 and X3 acquire a purely crystallographic significance. On the other hand, the metal-oxygen square unit will be regular only when the atoms M are identical or when their

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

Table 3

213

Structural parameters of metal halide oxides

Formula of compound

Space group

Unit cell parameter, C a

PbO Pb0.92Ti0.08O1.08

P4lnmm P4lnmm

XO 3.978 3.966

BaBi02Cl BaBi02Br CdBi02Br NaBi304C12 NaBi304Br2 NaBi30412 LiBi304C12 LiBi304Br2 LiBi30412 PbBi02C1 PbBI02Br PbBiO2I

I4lmmm I4lmmm I4lmmm I4lmmm I4lmmm Mlmmm I4lmmm I4lmmm Mlmmm I4lmmm Mlmmm I4lmmm

X1 4.019 4.080 3.944 3.878 3.925 3.991 3.841 3.876 3.942 3.956 3.989 4.053

BiOCl LaOCl CeOCl PrOCl NdOCl PmOCl SmOCl EuOCl GdOCl TbOCl DyOCl HoOCl ErOCl

P4lnmm P4lnmm P4lnmm P4lnmm P4lnmm P4lnmm P4lnmm P41nmm P4lnmm P4lnmm P4lnmm P41nmm P4lnmm

3.883 4.1 19 4.080 4.051 4.018 4.020 3.982 3.965 3.950 3.927 3.91 1 3.893 3.880

A

A

Interatomic, M-40 M-4X M-X

5.026 4.977

2.31 2.3 1

12.98 13.27 12.02 12.14 12.55 13.27 12.03 12.48 13.19 12.63 12.80 13.52

2.33 2.36 2.28 2.24 2.27 2.30 2.24 2.26 2.29 2.27 2.29 2.35

3.52 3.57 3.44 3.34 3.43 3.58 3.29 3.37 3.51 3.47 3.50 3.61

7.348 6.883 6.831 6.810 6.782 6.740 6.721 6.695 6.672 6.645 6.620 6.602 6.580

2.32 2.39

3.05 3.18

3.49 3.14

2.30 2.29

3.1 1 3.1 1

3.09 3.08

2.25

3.05

3.04

x2

x 3

tetragonal

3.897

21.69

positions are occupied randomly by different kinds of metal atoms, In general, however, partial deformation with a consequent lowering of the symmetry of the unit occurs. In the structure of the compounds X1 shown in Figure 3(a), each of the two nearest [M202]layers are displaced by a half-diagonal to one another, and the second layer is in an equivalent position relative to any selected layer. The unit cell therefore contains two such layers. Each metal atom is surrounded by four oxygen atoms and four X (halogen) neighbors. Table 3 shows that in the phases X1 the [M202] layer generally contains metal atoms of different kinds.

214

Catalysis

With rare exceptions, they are arranged randomly in the lattice, so that they can be denoted simply as M without further specification. The single ‘elementary’ square not only retains its regularity but also shows little change in size (parameter a in Table 3) from one substance to another. The phases BaBi02X show some tendency towards ordering, but it is not detected by X-ray diffraction. In lead antimony halide oxides, however, the cation distribution is strictly ordered, and the [M202]layers are constructed by the pairing. This leads to some deformation of the unit square. In general, however, the metal halide oxides have a high structural flexibility, so that not only the large Ba2+ but also the small Li+ ion (R Li+=O.68 can be ion (R Ba2+= 1.38 included without significant modification. In the structure of the compounds X2, which is shown in Figure 3(b), all the [M202]layers are crystallographically identical, so that the unit cell contains one such layer, and this explains the smaller value of their unit cell parameter c compared with the phase X1 (Table 3). The parameter a (for the tetragonal compounds), however, is close to the ‘ideal’ value. Each metal atom in the X2 lattice is surrounded by four oxygen neighbors and by four nearest halogen atoms. The structure of the complex X3 is shown in Figure 3(c). Here, within the triple halogen layer, there are positive ions such as Ca2+ or Cd2+, which evidently stabilize the wide anionic fragment. Recurrent inter growth^'^^ at the sub-unit-cell level are possible in the family of bismuth chloride oxides, as witnessed by phases designated X1X3 or X1XIX3 in Figure 4(a) and Figure 4(b), respectively. The occurrence of such intergrowths is identified through the magnitude of the observed unit cell (z-direction) dimension, which is 6.065 A for XIo(c= 12.13 7.37 for X2, 10.845 for X3 and 13.435, 16.91 and 18.215 A for X1X2, X1X3 and X2X3 structures, respectively. Structures designated by X1X1X2,X1X2X2 and X1X1X3have c = 19.50, 20.805 and 24.28 respectively, and X2X2X3, X1X3X3andX2X2X3 have c = 25.585, 27.755 and 29.06 In above series, Harris et aZ.l4 have elucidated the structure of at least one member of this series. This is the structure of the compound Cs2BilO Ca6Cl12016, in which there is a layered portion of the well known CsCl structure (containing only partial occupancy of the Cs sites) flanked on either side by modified bismuth chloride oxide layers of XI type, within which the Bi3+and Ca2+ions occupy the same type of site. There are reverse structures to the above phase^;'^^-'^^ that is, the compositions in these structures are regulated by the thicknesses of the metal oxygen layer sheets with an invariant single halide layer. Typical structures of bismuth compounds in this group are illustrated in Figure 5. The XI type structure model is shown as a reference. This layered structure consists of the single chlorine sheet and multiple bismuth-oxygen sheets having a general composition, [Bi,O,,, - 1),2]+.The thickness of the bismuth-oxygen sheet is changeable. Thus a highly multiple metal oxygen sheet is possible. The structure of the multiple bismuth-oxygen sheet itself is fluorite-related. The most popular phases in this group are Bi24C110031,the so-called Arppe’s

A)

A),

A,

A.

A)

A

A

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

215

00

0c1

-

CL?

Figure 4

Elevation views of (a) the XiX3 and (b) XlXlX3 structures

Figure 5

Schematic representation of four distinct types of bismuth chloride oxide: (a) BiOCI; (b) Bi24031ClIo;(c)Bi304C1;(d) Bi12017C12; (e) Bi203

Catalysis

216

phase, and Bi3C104;the former is a chlorine deficient and oxygen sufficient analogue of X2 and in the proposed structure a part of the chloride ion sheet originally presented in the X2 type is replaced by an oxide layer, forming zigzag staking sheets [Figure 5 (b)], and the latter is closely related to the Arppe structure but without the zigzag staking sheet; that is, the structure is a simple layered structure consisting of the single chlorine sheet and double metal oxygen sheet having a composition: [Bi304]+[Figure 5 (c)]. 146 Another example is Bil2Cl2OI7,whose structure is closely related to that of Bi3C104;both have a single chlorine sheet but different thicknesses of oxide sheet. If we take the X1 type consisting of single chlorine and oxide sheets as a reference, the thickness of the metal oxygen sheet along the c direction of the sub-unit cell increases stepwise by about 2.9 with increasing thickness of the sheet; the thickness is calculated to be n = 5 from the observed co value (35.7 A). The ideal structure of this sample is shown in Figure 5(d). In this sample, however, a superstructure was observed from the SAED patterns and HREM images, 2-fold along the [1101 and 16-fold along the [-1101 direction of the subunit cell. Therefore, the real unit cell will be orthorhombic with a = 5.76, b x 45.3 and c x 35.7 Although the basic structure is shown in Figure 5(d), more structural investigations are obviously needed to confirm the structure. The last sample in Figure 5(e) is Bi2O3, which has a fluorite structure and no chlorine sheet exists. We can consider this sample to be an extreme of bismuth chloride oxide catalysts. A [Bi202I2+ layer also occurs in a recurrent fashion in the Aurivillius phases148where the interleaving components are perovskitic and of general formula [A, - 1Bn03n+1]2-where n represents the thickness of the perovskitic layer. A compound which belongs to the simplest member of the large family of the Aurivillius phases is [Bi2O2]y04]. This structural compound can be formed by elements of both Sillen and Aurivillius phases.149Some of the simplest examples (with n = 1 in the notation used above for the perovskitic component) are PbBi3W08Cl and Bi4NbO&1 and have layer sequences given by: [PbBiO2l2+[C1] -[Bi202]2'[w04]2- and [Bi202]2'[Cl]-[Bi202]2+[Nb04]3- . These compounds are the first members of the so-called bipox (bismuthperovskitic-oxyhalide) family of structures. Mixed Aurivillius-Sillen structures consisting of recurrent chloride oxides and multiply condensed perovskitic layers (n = 2, 3, . . . in the component [An- 1Bn03,+1]), as well as those in which components of double halogen layers (as in BiOCl itself) are recurrently interspaced with pervoskitic layers, may also be possible, and have indeed been synthesized. The compound PbBi3Re08C12, reported by Ackerman, is one such compound. As related examples of metal halide oxide, there are superconducting solid materials like SrCu02F2+6 with a maximum Tc of 46 K.129 In a superconducting fluoride oxide, fluorine plays a dominant structural role, rather than merely being an electronic dopant as in L ~ ~ C and U ~ Nd2Cu04-,Fy. 15' The fluoride oxide Sr2Cu02F2 was demonstrated to be able to support superconductivity; in other words, superconductivity in Sr2Cu02F2 has been induced by incorporating interstitial fluorine. The evidence that

A

A.

~

F

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

Table 4

217

Catalytic performance of complex bismuth chloride oxides having different structures in the oxidative coupling of methane at 993 K

Catalyst

Structural % of Conversion1 Selectivity Yield of Ethylenel type CI sheet a % methane to Czof c2ethane compoundsl % compoundsl %

LiCa2Bi304C16 NaCa2Bi304C16 BiOCl

x3

75

41.7 33.8

46.5 43.2

19.4 14.6

25 35

x 2

67

29.8

35.3

10.6

13

LiCaBi304C14 NaCaBi304C14

XlX3

67

21.9 16.0

67.0 64.9

14.7 10.4

4 2

Li3Ca2Bi9O12Cll0 XIXIX3 63 Na3Ca2Bi9012Cllo

17.7 15.3

66.8 63.8

11.8 9.8

LiBi304C12 LiBi304C12

15.6 13.2

62.9 56.6

9.8 7.2

3 2 2 2

a

XI

50

Single metal oxide sheet was counted as one sheet.

fluorine avoids the equatorial sites clearly demonstrates the potential role of the fluoride oxides for creating new high temperature copper oxide superconductors. The partial replacement of the apical F- anions by 02-could provide an alternative means of anionic chemical control of the electron carrier density, in analogous fashion to Sr and Ba substitutions for La in La2Cu04. The F site preference can promote interchange of 0 and F ions to create the structural features necessary to support superconductivity.

4.2 Layer Structure-Activity Relationship. - As described above, the bismuth halide oxides constitute a large family of layered compounds, all of which crystallize into structures consisting of cation-oxygen (anion) layers associated with the tetragonal PbO structure, alternating with single or multiple sheets of halide ions. In addition, the halide layers may also accommodate additional cations in their interstices. Therefore one can expect to control catalytic performance of the bismuth chloride oxides for alkane oxidations by designing the layered structures. A wide variety of bismuth chloride oxide catalysts has been synthesized and tested for oxidative coupling of methane.87,88.128.152 Table 4 summarizes the catalytic activity and selectivity to C2 compounds along with structural information for the catalysts. The performance of these catalysts was measured as follows. First, the following four types of catalysts, XI,Xs,X1X3,and XlXIX3, were tested. These four types of chloride oxides exhibit high activity for the formation of C2 products in the oxidation of methane.87.88.128-152 The X 3 type showed the highest activity. The catalytic activity of this type of solid is, however, unstable, decaying quickly with reaction time. After prolonged reaction, the X3phase changes into X1X1X3 and the resulting activity is the type catalyst when freshly prepared. On the other same as that of the X1XlX3

218

Catalysis

hand, the X1, X1X3 and X1X3 types of catalyst showed highly stable activity for the conversion of methane. Clearly, the existence of the X1 type layer unit in the chloride oxide structure contributes to the high catalytic stability, even though the catalytic activity is significantly impaired. The catalytic performances of the X1X3 and X1X1X3 types of solid were superior among the various structural types. Their all-round catalytic performance seems to originate both from the activity of the X3 unit and from the intrinsic stability of the X1 unit. The same trend is observed on the sub-set of both Li and Naanalogues.128No difference between the activities of these two sets was found. This indicates that the influence of the chloride ion dominates over that of the alkali ion. The above results reveal that the X3 layer unit in the structure largely governs the catalytic activity and that the oxide layer seems to be less influential. It was, however, observed that Bi2O3 showed poor conversion and selectivity while BiOCl was highly selective, which means that the metal oxide layer has some catalytic function in the methane oxidation. Secondly, the catalytic performance was compared for the bismuth chloride oxides having multiple metal oxygen sheets in the oxidative coupling of methane. As can be seen in Table 5, the catalysts were in the following orders in terms of the activity (methane conversion, selectivity, and ethylene/ethane ratio), 1539154

> Bi2O3, BiOCl > Bi24031C110> Bi3O4C1> Bi12017C12 the chlorine content (or in the reverse of the thickness of oxide sheet), > Bi3O4C1> Bi12017C12 > Bi2O3, BiOCl> Bi24031Cl10 and the stability, BiOCl< Bi24031C110 < Bi3O4C1= Bi12017C12 = Bi2O3. The observed parallelism between the orders of the thickness of oxide sheet and the stability is quite natural, because structural deformations in which chlorine is liberated from the lattice are inevitably easier for a sample having thinner oxide sheets. In other words, the bond strength between chlorine anion and metal oxygen sheet might be weaker when the catalyst consists of thinner oxide sheets. We therefore believe that such weakly bound chlorine in the layered lattice is highly reactive for activating methane. An apparent relationship can be seen between the orders of the thickness of oxide sheet and the catalytic activity; the higher the thickness the lower the catalytic activity. Naturally, the absolute content of chlorine in the catalyst must directly correlate with the activity in this case. This does not exclude involvement of the oxide sheet in the course of methane oxidation. There still exists a possibility that the lattice oxygen close to the chlorine anion is active. However, since the selectivity decreased with increasing thickness of the oxide sheet without apparent changes in the conversion (Bi3O4C1 > Bil2OI7Cl2> Bi2O3, as shown in Table 9, the oxide sheet is also active for methane oxidation but should be less selective, therefore diluting the effect of the selective oxidation site of the chlorine sheet.

5

00

1

1

1

Bi304Cl

Bi12017C12 Bi203

2

fresh fresh

fresh used fresh

1,2

Bi24031c l 10

fresh used

1

BiOCl

2

Thickness Thickness of CI sheet of M e - 0 sheet

BiI2Ol7Cl2 Bi2O3

Bi24031C110 Bi3O4C1 Bi3O4C1 6.7

8.4

19.1 16.3 8.6 18.0 22.7

90.6 20.6 40.5 33.8 17.0

Conversion/% methane 0 2

29.8 BiOCl Bi24031C110 + Bi304C1 11.7

Observed phase

5

9

39.6 18.4

35.3 55.5 43.4 48.9 45.6 3.3 1.2

10.5 6.5 8.3 8.0 3.9

0.6 0.4

13 2.3 2.7 2.0 1.o

\o

c

N

6' 5

g

9

9m

s

b

k 2.

5 B

9 2

95

g

3

Selectivity to Yield of Ethylenel E C2-compoundsl% C2-compounds/% ethane

Catalytic performance of bismuth chloride oxides having different structures in the oxidative coupling of methane

Catalyst

Table 5

220

Catalysis

As a consequence, the highly multiplied chlorine layer is responsible for effective activation of methane to yield ethylene, while the multiplied metal oxide sheet can stabilize the structures. This clear structure-activity relationship indicates that catalytic methane oxidation can take place both over the metal oxygen sheet and over the chlorine sheet, the former mainly promoting nonselective oxidation and the latter effectively catalysing the oxidative coupling of methane. This leaves little doubt that methyl radical production is favored by the solid catalyst. 155~156Since gas-phase chlorine species, if any, which are liberated from the catalysts, do not contribute to the reaction,9c94 the key to the effectiveness of the metal chloride oxide catalysts may be linked to the state of surface chlorine anions which are supported by the coexistence of a bismuth-oxygen sheet in the structures. Both sheets seem to be interconnected by electron transfer at the reaction temperature. 4.3 Structural Tuning in Local Chlorine Environment. - One of the most important properties of the metal chloride oxide catalysts is their low activity for total oxidation. Its inability to perform complete oxidation seems to be as important as its ability to perform the desired oxidative coupling and dehydrogenation reaction. The high yield of ethylene and high ethylene to ethane ratio may be related to their high catalytic activity for the oxidative dehydrogenation of ethane to ethylene. Based on this activity, the local environment of chlorine in the layer structure of bismuth chloride oxide catalysts can be tuned by structural modifications to achieve high yields of ethylene in the oxidative dehydrogenation of ethane. Various layered complex metal chloride oxides have been reported as catalysts for the oxidative dehydrogenation of ethane to ethene with molecular 0xygen.~O7'~~ It was found that a BiOCl catalyst having X2 type structure showed high activity and selectivity in the oxidative dehydrogenation of ethane but was unstable. On the other hand, LiBi304C12 catalyst having a single chlorine layer between metal-oxygen sheets (XI type) displayed an extremely high durability for the reaction, but showed only a moderate catalytic performance. It can therefore be expected that the X2 type phase could be stabilized by structural combination with a highly stable XI type unit without diminishing the high activity of the X2 type catalyst. For this purpose, X1X2 type structural materials were selected. In fact, SrBi304C13catalyst, as a typical X1X2 type structural material (space group I4/mmm with lattice parameters of a = 3.9370(1), c = 27.0177(7)), showed a good catalytic performance and a good structural stability in the oxidative dehydrogenation of ethane. 57 The XIX2 structure with the formula Cl/Bi-O-Bi(Sr)/C12/Bi(Sr j O - B E 1 is shown in Figure 6. It can be seen that a [M202In+(M=Bi or Sr) layer is present. This cationic layer is balanced by the anionic chlorine layers. In the metal-oxygen layers oxygen is in a tetrahedral coordination to metal ion M, while in the chloride layers chlorine is in a six-fold coordination. As a result, a square of four oxygen atoms is located above the metal ions and a square of four chlorides below, as shown in Figure 6. The double chlorine layers are held by the van der Waals bond. In general, strontium ions can occupy both the

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

221

L

:Bi

0

:Bi, Sr

0

:O

0 c1 :

b

Figure 6

The structure of SrBi30&13 along [OlO]

Bi(1) and Bi(2) sites in the metal-oxygen sheets by replacing bismuth ions with strontium ions. However, from the structural Rietveld analysis, it has been revealed that between two Bi(1) and Bi(2) sites, the former Bi in the metaloxygen sheet facing the single chlorine sheet and the later facing the double chlorine sheet, strontium ions occupy the Bi(2) site only, and do not occupy the Bi( 1) site. This structural information clearly supports that a stronger bonding of C1 with Sr than with Bi stabilizes the double chlorine layer under catalytic ethane oxidation conditions, resulting in better catalytic performance in the oxidative dehydrogenation of ethane. By tuning the catalyst (modification of the electronic state of chlorine in the double layers by means of element substitution), the XIX2 type catalysts became more active for the oxidative dehydrogenation of ethane. 158 This was demonstrated by the use of lanthanum-substituted complex bismuth chloride oxide catalysts having the XlX2structure, as shown in Figure 7. The general composition of the lanthanum-substituted complex bismuth chloride oxide catalysts is given by Bi3-nLanSrC1304, the nominal components used in the preparation of the catalysts. All the catalysts exhibited good performance for the formation of ethene as shown in Figure 7. The Bi3SrC1304catalyst showed a high selectivity to ethene, but the conversion of ethane was low,

Catalysis

222 100

r

20

0 0.00

0.25

0.50

0.75

1.oo

Lanthanum content /n Figure 7

The oxidative dehydrogenation of ethane over SrBi3-,La,O4C13 catalysts at 640 "C: A conv. 0 2 ; conv. C2H6; 0 sel. C2H4 0 sel. CO;A sel. CO2

about 10%. The introduction of lanthanum ion (n = 0.25) gave 15% conversion of ethane. When n >0.25, the conversion of ethane increased markedly. The Bi2.25L%.75SrC1304catalyst is about four times more active than the Bi3SrC1304 catalyst for the oxidation of ethane and a maximum ethane conversion of 53% was observed over the Bi2.25L~.75SrC1304 catalyst. However, the selectivity to ethene slightly decreased with the increase of lanthanum content. For example, the selectivity decreased to 81% (n = 0.75) from 96% (n = 0). Since the surface area of the catalysts (less than 0.5 m2 g-I (BET method)) was not changed significantly by the replacement, it is apparent that the surface area of the catalysts is not a important factor for the increased conversion. For the oxidative dehydrogenation of ethane, the increased catalytic activity can be attributed to the structural deformation of Bi3-,La,SrC1304 catalysts and the enhanced oxidation ability of the cationoxygen sheet by the introduction of lanthanum ions. It was confirmed by X-ray diffraction that the catalysts remained in the XlXz type structure during the replacement with lanthanum ion to the degree of n 5 0.75. By increasing the lanthanum content, the value a along the sheet in the sub-unit cell increased from 3.937 ( n = 0 ) to 3.973 (n=0.75). Although the change in a was very small because the ionic radius of lanthanum ion was slightly larger than that of bismuth, it is clear that the lanthanum ion invaded the cation-oxygen sheet by substitution. On the other hand, the value of c representing the thickness of the layer structure decreased significantly from 27.017 to 26.849 Such a preferential variation in the sub-unit cell parameters is very interesting. The contraction of the thickness of the layer structure indicates that the interaction between the metal-oxygen sheet and the chlorine sheet becomes stronger.

A

A

A.

A

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

223

50000

40000 ;>s

c.,

-

*g

8 c

30000 2

m

H

lo000

0 I I

20

-

.

I

I

40

.

. I

.

.

I

I

60

I

80

I

I

I

100

1; 0

28ldeg. Figure 8

Final Rietveld for SrBi~.2sLao,7sCl~Oq catalyst, X-ray data (CuKa)

To explain these changes, the structure of the Bi2+25La0,75SrC1304catalyst was analysed by the Rietveld method. First, the powder pattern of Bi2.25L%.~~SrC130~ was indexed to be the same as the Bi3SrC1304catalyst, giving a primitive tetragonal cell. The reflection condition hkl was h + k + I = 2n, and physical tests on non-centrosymmetry were negative, so that the centrosymmetric space group Z4/mmm was chosen as the space group of the Bi3SrC1304 catalyst. Final Rietveld plots for the X-ray refinement are given in Figure 8. The structural model of Bi2.25L~.75SrC1304 is similar to that of Bi3SrC1304, which is shown in Figure 6. Lanthanum ions occupied the site of bismuth ion impartially, and the anticipated Sillen-type [M202](M=Bi, La or Sr) layers can be seen. The metal-oxygen sheets were separated by single chlorine layers and double chlorine layers alternately. The selected bond lengths are listed in Table 6, which indicates that the interatomic distances of M-O and M-Cl in the Bi2.25L~.75SrC1304 catalyst are longer than those in the Bi3SrC1304 catalyst; this can be attributed to the larger ionic radius of lanthanum ion. In the XIX2 structure, metal ions lie in the depressions between four oxygen ions. It is understandable that introducing the lanthanum ion can result in a little change in size of the Bi3-nLa,SrC1304 catalysts (lattice parameter a increase), so the value of a along the sheet in the sub-unit cell is slightly increased by the introduction of lanthanum ion. The position of the chlorine of the double chlorine layer is decided by the space coordinate of C1( 1) (O,O,z). According to the structural symmetry relationships, the perpendicular distance between the double chlorine layer can be calculated using the space coordinate of Cl(1) and the value of c . The space coordinate z of Cl(1) in the structure of the Bi3SrC1304catalyst is 0.205, and that in the structure of the Bi2.25La0.75SrC1304 catalyst is 0.210. The calculated

224

Table 6

Catalysis

Interatomic distance I

A Bi(l)-O Bi( 1)-C1(2) Bi(2)-O Bi(2)-C1( 1)

2.18 3.39 2.36 3.08

M: Bi, La or Sr

a

Figure 9

The local structure of SrBi2.25La0.&l304

catalyst

result showed that the distance in the double chlorine layer in the Bi2.25which is smaller than that of the Bi3SrCI3O4 Lao.75SrC1304catalyst is 2.10 catalyst (2.43 It is clear that the thickness of the double chlorine layer becomes thinner by the replacement. The metal ions in the Bi3SrCI3O4catalyst show 8-coordination: a square of four oxygens is located above the metal ion and a square of four chlorines below. The double chlorine layers are connected by van der Waals bonds. On the other hand, if the bismuth site is replaced by a lanthanum ion, the lanthanum ion can interact with the second chloride layer in the double chlorine sheet as shown in Figure 9 because lanthanum ion can exist in a 9-coordination structure. In fact, the distance between the metal ion (Bi/La) and the second chloride layer in the double chloride sheet was measured to be 3.40 (the distance of the dotted line) in the structure of the Bi2.25La0.75SrC1304 catalyst, as indicated in Figure 9, and the Bi-Cl distance in the structure of Bi3SrC1304. The increased was measured to be 3.77 coordination number is the reason that the thickness of the double chloride layers becomes thinner. However, the unit cell of the X1X2 structure contains two double chlorine layers. So if the contraction of the c axis (BBi2.25Lao.75SrC1304) was attributed to the variation of the thickness of the double chlorine sheets only, the contraction of the c axis (Bi2.25Lao.75SrC1304)would be 0.66 but in this catalyst, only a 0.17 contraction was observed as indicated in the result of

A,

A).

A

A

A,

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

1

204

1

1

202

1

1

200

1

1

198

1

1

1

196

1

194

1

1

192

225

1

190

Binding Energy / eV

Figure

The ClZPXPS spectra of (a) SrBiClJ04and (b) SrBi2,Z5La0.&1304

lattice parameters. The reason for this is very simple; the insertion of lanthanum ion into the metal-oxygen sheet increases the value a of the subunit cell because of its larger ionic radius, and therefore the thickness of the metal-oxygen sheet is also increased. As a consequence, the thickness of the c axis of the lanthanum-substituted X1X2 layered complex bismuth chloride oxide catalysts decreased at the double chlorine sheet and the interaction between the oxide and double chlorine sheet became stronger. The change of chlorine by the introduction of lanthanum was also clearly revealed by the changes of C12p binding energies of XPS. The C12p spectra of the Bi3SrC1304and Bi2.25LaO.~~SrC1304 catalysts are shown in Figure 10. These spectra show a doublet 28312 and 2p112with a split of 1.6 eV. The C12p peaks of Bi3SrC1304 were at 197.7eV (Cl 2~312)and 199.3 eV (C1 2p1/2). The C1 2p peaks of Bi2.25La0.75SrC1304appeared as broad peaks shifted to lower binding energy, 197.1 eV (C12p312) and 198.7 eV (C12p1/2), respectively, indicating that the chlorine ions were perturbed by the presence of the lanthanum ion in the metal-oxygen sheet of the Bi3SrC1304catalyst. With the increase of lanthanum content, binding energies of C12p decreased. According to the result of the structural analysis, lanthanum ion and bismuth ion occupied the same sites; the coordination number of metal ion facing the double chlorine sheet changed from 8 (40+4Cl ) to 9 ( 4 0 + 5C1), and the contraction of the double chlorine sheet made the distance between th? chlorine ions in the double chlorine sheet decrease from 3.7 (n = 0) to 3.5 A (n = 0.75). Therefore, the interaction between the double chlorine sheet and the metal (lanthanum)-oxygen sheet decreases the binding energy of the electron of chlorine ions. On the other hand, the single chlorine sheet has no obvious changes, since there is no significant structural change in the environment of the chlorine. Therefore, after the replacement the spectra of C1 2p became

A

226

Catalysis

broader. From the Rietveld and XPS analysis, it is suggested that the chlorine ion existing in the Bi3-nLanSrC1304catalysts can release the electron more easily than that in the Bi3SrC1304 catalyst under catalytic oxidation conditions. This is the reason why the introduction of lanthanum ions promotes the catalytic activity remarkably. The lattice chloride ion can release electrons to the metal-oxygen sheets to become a radical chlorine by the thermal process. The radical chlorine thus formed would be capable of activating ethane to ethyl radicals. The ethyl radicals are immediately converted into ethene. From this point of view, the activation of chlorine is very important for oxidative dehydrogenation of ethane on the chloride-containing surface of metal chloride oxide-based catalysts.

4.4 Structural Tuning of the Metal Oxide Layer. - As shown in Figure 7, the selectivity to ethene decreased slightly with the replacement. Since it was observed that the selectivity to ethene is almost independent of the conversion of ethane, the decrease of selectivity to ethene can be attributed to the lanthanum ions in the metal-oxygen sheets. The destructive oxidation reaction was dominant over pure LaOCl catalyst, suggesting that the lanthanum ion in the metal-oxygen sheets can form activated oxygen species for the destructive oxidation of ethane. It was explained earlier that the activation of chlorine ion is promoted by the introduction of lanthanum ion into the metal-oxygen sheet, but the existence of lanthanum ions in the metal-oxygen sheet would also result in the complete oxidation of ethane at the same time. In this section, structural tuning of the metal oxide layer is summarized. For this purpose, the chloride oxides compounds BiM2O4C1(M=La, Nd or Y), which were discovered by Milne et U Z . , ’ ~ ~ were used. Bi2LaO4C1has been characterized by a combination of X-ray and neutron powder diffraction: space group P4/mmm, a = 3.95471(3) c = 9.1275(3) The structure consists of alternating layers of ‘triple fluorite’, [M304]’ and halide, C1-. Within the fluorite layers there is perfect ordering of the cations, -Bi-O2-La-O2-Bi, such that lanthanum ion is coordinated only by oxide layers and Bi by both oxide and chloride layers. This phase represents the simplest example of an intergrowth structure containing the [M304] block layer, which is found in several of the copper oxide superconductors. The structure also suggests the existence of a new family of layered chloride oxides parallel to the well known Sillen phases, [M202]Xn, with triple [M304] fluorite layers replacing the normal double [M202] layers. Similar to Bi2La04C1,Bi$3m04Cl can be prepared by reaction of stoichiometric quantities of BiOCl, Sm2O3 and Bi2O3. The powder pattern was indexed, Fnd gave a primitive tetragonal cell with approximate dimensions a = 3.89 A, c = 8.97 A. After Rietveld refinement of this model a reasonable profile fit was obtained. Final convergence was achieved at Rwp = 13.18% and Rp = 9.72%. The final refined structure is shown in Figure 11. Different from the Sillen-type [Bi202] layers, it can be seen that the [Bi2Sm04]layer is present, which is regarded as the triple fluorite unit. This cationic layer is balanced by the simplest single anionic layer, C1-, hence the structure has the layer

A,

A.

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

227

:Bi

:Sm

0 :o 0:CI

a

Figure 11

Structural view of Bi,Sm04Cl catalyst

sequence Cl-[Bi2Sm04]-Cl- along the crystallographic c axis. As the structural result of the BizLaO4Cl compound, Sm ion is also coordinated only by oxide layers and Bi by both oxide and chloride layers. The catalytic performances of Bi2MO4C1(M: Bi, La, Sm, Y, Nd) for the oxidative dehydrogenation of ethane are as follows: the Bi3O4C1 catalyst gave the highest selectivity to ethene; ethane conversion and selectivity to ethene were 14.7% and 88%, respectively. The Bi2MO4C1 (M=La, Sm, Y, Nd) catalysts gave higher catalytic activities, but lower selectivities to ethene. For the BizLa04C1catalyst, ethane conversion and selectivity to ethene were 21.8% and 74.6%, respectively. For the Bi2Y04Cl catalyst, the corresponding values were 83.7% and 19.4%. These data obviously show that the catalytic performances were strongly affected by the metal ions M in the [Bi2M04]+ sheets, although chlorine is a key catalytic element for the reaction. Accordingly, the results reveal that the oxidative dehydrogenation of ethane is controlled both by the metal-oxygen sheets and by the chloride sheets, the former mainly promoting the activation of molecular oxygen and the latter catalysing the oxidative dehydrogenation of ethane. In order to tune the catalytic function of the metal-oxygen sheet more, the lanthanum-substituted bismuth chloride oxides catalysts Bi3- .La,O4Cl were investigated for the oxidative dehydrogenation of ethane. 125 The catalytic performance of Bi3-,La,04C1 is shown in Figure 12. The Bi304Cl catalyst showed a good catalytic performance for the formation of ethene, but the conversion was low, about 14.7% conversion of ethane. On increasing the

Catalysis

228

I

0.00 0.25

I

I

I

1

0.50

0.75

1.00

1.25

I

1.50

La Content (n)

Figure 12

The catalytic performance of Bi3-,La,04Cl catalyst for the oxidative dehydrogenation of ethane: reaction temperature 640 "C;catalyst weight 2 g; totalflow rate 50 ml min-I (C2Hb:O2:Nz= 1:4:15); 0 conversion of ethane; selectivity to ethene

amount of lanthanum ion to n = 1, the ethane conversion slowly increased to 21.8% and the selectivity to ethene slightly decreased to 74.6% (n = 1) from 88% (n = 0). When n > 1, the conversion of ethane increased markedly. The Bi1.5La1.504C1catalyst is about three times more active than the Bi3O4C1 catalyst for the oxidative of ethane. On the other hand, the selectivity to ethene decreased markedly, to 43%(n = 1.5) from 74.6% (n = 1). The variation in the catalytic performance is not directly proportional to the lanthanum content. For lanthanum content n c 1, lanthanum ion replaced the Bi site coordinated by oxide layers only, and the catalytic activity showed no obvious change. For lanthanum content n > 1, the lanthanum ions occupied the sites of bismuth ions which are coordinated by both oxide and chloride layers, and then the catalytic activity increased significantly. On the other hand, the selectivity to ethene decreased markedly with the increase of lanthanum (n > 1). Since it was confirmed that the selectivity to ethene is almost independent of the conversion of ethane, the decreased selectivity is attributed to the intrinsic change of the catalyst. As described in the preceding section, the promotion effect of lanthanum also was obtained for the lanthanum-substituted XlX2 type complex bismuth chloride oxides. The catalytic activity increased markedly and the selectivity only slightly decreased with the increase in lanthanum content. On the other hand, for the lanthanum-substituted Bi3O4C1 having triple fluorite, the Bi1.5La1.504Cl catalyst was about three times more active than the Bi3O4C1 catalyst for the oxidative dehydrogenation of ethane, but the selectivity to ethene decreased markedly to 43% (Bil.~Lal.~O~Cl) from 88% (Bi304C1). It is, therefore, clear that the lanthanum ion invaded the [M202] and [M304]cationoxygen layers, which resulted in the different effect on the selectivity to ethene.

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

229

Only when the bismuth ion in the [M304] cation-oxygen sheet was replaced by lanthanum ion did the sudden change in the selectivity to ethene occur, but not in the [M202] sheet. From the simple structural consideration, this can be attributed to the formation of three-dimensional La-O-La bonding in [Bi304 layers with the replacement of lanthanum (n > I), which is in accordance with the fact that the destructive oxidation reaction of ethane was the main reaction over pure La203 catalysts.

4.5 Halogen-involved Alkane Activation Mechanism. - On the basis of all catalytic and structural data, the following reaction mechanism for alkane oxidation can be deduced for the metal halide oxide catalysts and halogencontaining oxide catalysts. The oxidative activation of ethane occurs first through a direct interaction with the chloride-containing surface of the metal halide oxide-based catalysts. The lattice chloride ion releases electrons to the metal-oxygen sheets to become a radical state of chlorine by a thermal process. The radical chlorine thus formed would be capable of activating ethane to ethyl radical on the surface, and the ethyl radical formed is readily converted on the same site to ethylene which is readily desorbed into the gas phase. The abstracted hydrogens from ethane remain on the chlorine site for a while but migrate in a protonic state to the metal oxide layer to be oxidized to water. During this process, surface chlorine can regain its original electronic state and the metal oxide layer is reduced. In the final stage the reduced metal oxide layer is reoxidized with molecular oxygen and then the catalytic cycle is completed. Since in the lanthanum-substituted XlXz type catalyst system catalytic activity of the oxidative dehydrogenation of ethane was well controlled by modifying the chlorine state in the structure through elemental substitution, the rate-determining step of the reaction is undoubtedly the abstraction of hydrogen from ethane by surface chlorine sites. The electron transfer process from the lattice chloride ion to the metal-oxygen sheets to form a radical state of chlorine seems to be in thermal equilibrium. The ethylene selectivity change depends on the thickness of the metal oxide layer and is explained as follows. At the final stage of the catalytic cycle the reduced metal oxide layer must be reoxidized by molecular oxygen. However, stable intermediate oxygen anion species can often be formed on the metal oxide layer during the four-electron reduction of molecular oxygen. Since some of these oxygen species are extremely active for promoting complete oxidation, high selectivity to ethylene cannot be achieved when such oxygen species are stabilized on the metal oxide surface. This is the reason why lower selectivity was observed on the bismuth chloride catalyst having multiple oxide layers. In addition, steric hindrance of chlorine, having a larger ionic radius than the oxygen anion, may have an effect on the formation of oxygen species when the metal oxide layer is thin. In the case of fluorine-containing oxide catalysts, surface fluorine or structural fluoride ion can directly control the reactivity of surface oxygen species, so that the catalyst becomes active and selective for alkane oxidation. Analogously, the stability of the metal chloride oxide catalysts can be

230

Catalysis

explained as follows. In the above reaction mechanism, the abstracted hydrogen remains on the chlorine sites for some time. Because this state can be regarded as an adsorbed hydrogen chloride, there must be a possibility for this adsorbed species to desorb before the hydrogen in the protonic state migrates to the oxide layer. Apparently the desorption causes the liberation of the structural chlorine into the gas phase. To prevent the liberation of halogen, strong bonding between the halogen layers and the metal oxide layers is desirable but, as observed in the structure-activity relationship, the oxidation activity will be sacrificed when the bond is strong. The bond between halogen and metal cation is highly ionic, resulting in suppression of the formation of the radical state of the halogen. Although the above explanation is based on the layer structure, the site-isolation theorylg works positively to form layered metal halide oxides, which might be the main reason for the extremely high effectiveness of the halogen-containing oxide catalysts in the alkane selective oxidations. 5

Conclusions

In this chapter, a brief summary of studies on halogen-containing metal oxides and on structural metal halide oxide catalysts active for selective oxidation of various hydrocarbons, particularly light alkanes, is presented. There are many types of active catalysts for selective light alkane oxidations but most of them are not satisfactory for practical application due to their still low catalytic effectiveness, even though complex catalysts have been examined. For overcoming these difficulties we emphasize that artificial or self organization of catalytic components, in particular structure during catalyst preparation, is highly important to create new types of active sites which are inevitably necessary for highly difficult selective oxidation. Although it is impossible to avoid the effect of unknown factors completely, inorganic synthesis should be used to control the local structural arrangement of elements in the preparation of complex metal oxide catalysts instead of conventional dry methods. It was also pointed out that for controlling the local bond structure of active sites, precise elemental substitution, or introduction of active sites, should be done by monitoring structure and catalytic activity. For this purpose, metal halide oxide materials are most suitable because the halogen located in a structure significantly enhances the catalytic activity and selectivity for selective oxidations. It was demonstrated that halogens are an indispensable component of catalysts for some oxidation processes which are commercially utilized, even though halogens cause problems in reaction systems and reaction facilities. Furthermore this chapter has described a successful example of tuning catalyst activity and selectivity by modifying the structural circumstances of the halogen on the basis of solid-state chemistry. Despite the results shown here, more work is needed to demonstrate clearly the possibilities of activity tuning of solid catalysts. We strongly believe that this type of approach will become more important for catalysts which require multifunctionality.

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

23 1

Acknowledgment

6

We are pleased to acknowledge the support for this work by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan. Very stimulating discussions with Prof. J.M. Thomas (The Royal Institution of Great Britain) are gratefully acknowledged.

7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

References Y. Moro-oka and W. Ueda, Catalysis, Vol. 11, ed. J.J. Spivey, Royal Society of Chemistry, Cambridge, 1994, p. 223. F. Cavani and F. Trifiro, Catalysis, Vol. 11, ed. J.J. Spivey, Royal Society of Chemistry, Cambridge, 1994, p. 246. B.K. Hodnett, Petrochemie, 1998, 114,256. R.K. Grasselli, Catal. Today, 1999,49, 141. T. Hayashi, K. Tanaka and M. Haruta, J. Catal., 1998,178, 566. I. Yamanaka, M. Katagiri, S. Takenaka and K. Tatsuka, Stud. Surf: Sci. Catal., 2000,130,809. G. Ferey, J. Solid. State Chem., 2000, 152, 37. B.K. Hodnett, Catal. Rev.-Sci. Eng., 1985,27, 373. T. Okuhara, K. Inumaru and M. Misono, ACS Symp. Ser., 1993,523,156. G. Koyano. T. Okuhara and M. Misono, J. Am. Chem. Soc., 1998,120,767. E. Bordes and P. Courtine, Topics Catal., 2000, 11/12, 61. K. Ruth, R. Kieffer and R. Burch, J. Catal., 1998,175, 16. E.M. Thornsteinson, T.P. Wilson, F.G. Young and P.H. Kasai, J. Catal., 1978, 52, 116. Mitsubishi Chemicals, JP 10-17523, 1998. Hoechst, WO 97-44299, 1997. UCC, JP 64-85945,1988. S. Tan, Y. Moro-oka and A. Ozaki, J. Catal., 1970, 17, 132. J.M. Thomas, R. Raja, G. Sankar and R.C. Bell, Nature, 1999,398,227. J.L. Callahan and R.K. Grasselli, AIChE J., 1963,9, 755. Mitsubishi Chemicals, USP 5 472 925, 1995. T. Ushukubo, K. Oshima, A. Kayou, M. Vaarkamp and M. Hatano, J. Catal., 1997,169,394. T. Ushukubo, K. Oshima, A. Kayou, and M. Hatano, Stud Surf: Sci. Catal., 1997,112,473. Mitsubishi Chemicals, JP 10-36311, 1998. R.K. Grasselli, Appl. Catal., 1990,57, 149. T. Okuhara, N. Misuno and M. Misono, Adv. Catal., 1995,41, 113. W. Ueda and Y. Suzuki, Chem. Lett., 1995,541. W. Ueda, Y. Suzuki, W. Lee and S. Imaoka, Proc. 11th Intern Congr. Catal., Baltimore, 1996, Stud. Surf Sci. Catal., 1996, 101B, 1065. N. Mizuno, T. Tateishi and M. Iwamoto, Appl. Catal., A , 1994,118, L1. W. Ueda and K. Oshihara, Appl. Catal., 2000,200,135. H. Watanabe and Y. Koyasu, Appl. Catal., 2000,194,479. M.M. Bhasin, Catal. Lett., 1999,59, 1 X.E. Verkios, F.P Stein and R.W. Coughlin, Catal. Rev.-Sci. Eng., 1980,22, 197.

232 33

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

Catalysis

A. Ayame, T. Kimura, M. Yamaguchi, H. Miura, N. Tekeno, H. Kanoh and I. Toyoshima, J. Catal., 1983,79,233. G. Lu and X. Zuo, Catal. Lett., 1999,58,67. J.R. Monnier, Stud. Surf: Sci. Catal., 1997,110, 135. T. Mtsuura, Yugoukyou, 1981,39,423. Kurare, JP 40-3969, 1961. D.M. Fenton, J. Org. Chem., 1972,37,2034. Ube Kosan, JP 9-1 10807, 1997. N. Kominami, Bull. Jpn. Petrol. Inst., 1971,13, 114. W.R. Knox, K.M. Taylor and G.M. Tullman, USP 3 833 638,1974. R. Burch, G.D. Squire and S.C. Tsang, Appl. Catal., 1989,46,69. P.G. Hinson, A. Clearfield and J.H. Lunsford, J. Chem. Soc., Chem. Commun., 1991,1430. K. Otsuka, Q . Lin, M. Hatano and A. Morikawa, Chem. Lett., 1986,467. W. Ueda and J.M. Thomas, J. Chem. SOC.Chem Commun.,1988,1148. S . Sugiyama, Y. Matsumura and J.B. Moffat, J. Catal., 1993,139,338. E. Ruckenstein and A.Z. Khan, Catal. Lett., 1993,18,27. A.Z. Khan and E. Ruckenstein, Catal. Lett., 1992,13,95. E. Morales and J.H. Lunsford, J. Catal., 1989,118,255. W. Ueda, S.W. Lin and I. Tohmoto, Catal. Lett., 1997,44,541. H.X. Dai, C.F. Ng and C.T. Au, Catal. Lett. , 1999,57, 1 15. S.B. Wang, K. Murata, T. Hayakawa, S. Hamakawa and K. Suzuki, Catal. Lett., 1999,59, 173. C.T. Au, K.D. Chen, H.X. Dai, Y.W. Liu and C.F. Ng, Appl. Catal., 1999, 177, 185. C.T. Au, X.P. Zhou and H.L. Wan, Catal. Lett. 1996,40, 101. S. Sugiyama, K. Sogabe, T. Miyamoto, H. Hayashi and J.B. Moffat, Catal. Lett., 1996,42, 127. W. Zang, X. Zhou, D. Tang, H. Wan and K. Tsai, Catal. Lett., 1994,23, 103. Q . Ge, B. Zhaorigetu, C. Yu, W. Li and H. Xu, Catal. Lett., 2000,68, 59. Mitsubishi Petrochemicals, JP 53-44525, 1978. T. Hattori, Kagakukeizai, 1983, 12. N. Kitajima, K. Matano, M. Saito, T. Oikawa, T. Konishi, H. Yasuda, T. Sat0 and Y. Yoshimura, Sekiyu Gakkaishi, 2000,43,89. R.V. Carrubba, Ind. Eng. Chem. Process Des. Dec., 1970,9,414. R.A van Santen and H.P.C.E. Kuipers, Adv. Catal., 1987,35,265. P.A. Kilty, N.C. Rol, W.M.H. Sachtler, Proc. 5th Intern. Congr. Catal., Palm Beach, 1972, Vol. 2, p. 929. C.T. Cambell, J. Catal., 1985,94,436. C.T. Cambell, J. Catal., 1986,99,28. E.A. Carter and W.A. Goddard, J. Catal., 1988,112,80. E.A. Carter and W.A. Goddard, Surf: Sci., 1989,209,243. K. Wu, D. Wang, X. Wei, Y. Cao and X. Guo, J. Catal., 1993,140,370. D. Wang, K. Wu, Y. Cao, X. Wei and X. Guo, Stud. Surf. Sci. Catal., 1993, 75, 1579. E.R. Frank and R.J. Hamers, J. Catal., 1997,172,406. K.A. Jorgensen and R. Hoffmann, J. Phys. Chem., 1990,94,3046. J.F. Roth, CHEMTECH, 1991,357. B. Cooker, A.M. Gaffney, J.D. Jewson, and W.H. Onimus, USP 5 780 657, 1998.

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115

233

Y. Moro-oka and W. Ueda, Adv. Catal., 1994,40,233. R.K. Grasselli and J.D. Burrington, Adv. Catal. 1981,30, 133. Y. Sasaki, Appl. Catal., 2000,194-195,497. R.K. Grasselli and R.C. Miller, USP 4 010 188, 1977. S.Yu. Burylin, Z.G. Osipova and V.D. Sokolovskii,Kinet. Katal., 1983,24,639. N. Giordano, J.C.J. Bart, P. Vitarelli and S. Cavallaro, Oxidn. Commun., 1984,7, 99. M. Chono, Bull. Jpn. Petrol. Inst., 1973,16, 333. M. Akimoto, K. Yamagami and E. Echigoya, J. Catal., 1984,86,205. G.E. Keller and M.M. Bhasin, J. Catal., 1982,73,9. G.J. Hutchings, Chem. SOC.Rev., 1989,18,251. Y. Amenomiya, Catal. Rev. -Sci. Eng., 1990,32, 163. K. Otsuka, Q. Lin and A. Morikawa, Inorg. Chim. Acta, 1986,118, L23. T. Komatsu, T. Amaya and K. Otsuka, Catal. Lett., 1989,3, 317. J.M. Thomas, W. Ueda, J. Williams and K.D.M. Harris, Faraday Discuss. Chem. SOC.,1989,87, 33. W. Ueda, T. Isozaki, Y. Morikawa and J.M. Thomas, Chem. Lett., 1989,2103. R. Burch, E.M. Crabb, G.D. Squire and S.C. Tsang, Catal. Lett. , 1989,2,249. R. Burch, S. Chalker and P. Loader, Proc. 10th Intern. Congr. Catal. Budapest, 1993, p.174. R. Burch, S. Chalker, G.D. Squire and S.C. Tsang, J. Chem. SOC.,Faraday Trans., 1990,86, 1607. T. Ohno and J.B. Moffatt, Catal. Lett., 1991,9,23. S.J. Conway and J.H. Lunsford, J. Catal., 1991,131, 513. R. Burch, S. Chalker, J.M. Thomas, W. Ueda and P. Loader, Appl. Catal., 1992, 82, 77. G.D. Moggridge, T. Rayment and R.M. Lambert, J. Catal., 1992,134,242. J.H. Lunsford, P.G. Hinson, M.P. Rosynek, C. Shi, M. Xu and X. Yang, J. Catal., 1994,147,301. H. H. Kung, Adv. Catal., 1994,40, 1. T.J. Yang and J.H. Lunsford, J. Catal., 1980,63,505. A. Erodhelyi and F. Solymosi, J. Catal., 1990,123, 31. R . Burch and R. Swarnakar, Appl. Catal., 1990,70,129. S.J. Conway and J.H. Lunsford, J. Catal., 1991,131, 513. R. Burch and E.M. Crabb, Appl. Catal., 1993,97,49. S.J. Conway, D.J. Wang and J.H. Lunsford, Appl. Catal. 1991,79, LA1. S.B. Wang, K. Murata, T. Hayakawa, S. Hamakawa and K. Suzuki, Catal. Lett., 1999,62, 191. C.T. Au and X.P. Zhou, J. Chem. SOC.Faraday Trans., 1997,93,485. C.T. Au and X.P. Zhou, Catal. Lett., 1996,40, 101. J.Z. Luo and H.L. Wan, Appl. Catal., 1997,158, 137. C.T. Au, H. He, S.Y. Lai and C.F. Ng, J. Catal., 1996,159,280. C.T. Au, Y.W. Liu and C.F. Ng, J. Catal., 1997,171,231. C.T. Au, X.P. Zhou, Y.W. Liu, W.J. Ji and C.F. Ng, J. Catal., 1998,174, 153. C.T. Au, Y.W. Liu and C.F. Ng, J. Catal., 1998,176,365. H.X. Dai, Y.W. Liu, C.F. Ngand C.T. Au, J. Catal., 1999,187, 59. H.X. Dai, C.F. Ng and C.T. Au, Appl. Catal., 2000,202, 1. A.D. Eastman, J.P. Guillory, C.F. Cook and J.B. Kimble, USP 4 497 971, 1985. X.P. Zhou, Z.S. Chao, J.Z. Luo, H.L. Wan and K.R. Tsai, Appl. Catal., 1995, 133,263.

234

Catalysis

116 X.P. Zhou, S.Q. Zhou, S.J. Wang, J.X. Cai, W.Z. Weng, H.L. Wan and K.R. Tsai, Chem. Res. Chin. Univ., 1993,9, 264. 117 X.P. Zhou, Z.S. Chao, W.Z. Weng, W.D. Zhang, H.L. Wan and K.R. Tsai, Catal. Lett., 1994,29, 177. 118 Z.S. Chao, X.P. Zhou, S.J. Wang, H.L. Wang and K.R. Tsai, J. Nut. Gas Chem., 1995,4,202. 119 D.M. Dai, X.P. Zhou, 2.S Chao, S.J. Wang, H.L. Wan and K.R. Tsai, J. Nut. Gas Chem., 1995,4,45. 120 X.P. Zhou, W.D. Zhang, H.L. Wan and K.R. Tsai, Catal. Lett., 1993,21,113. 121 M.A. Chaar, D. Pate1 and H.H. Kung, J. Catal., 1988, 109,463. 122 K. Hguyen and H.H. Kung, J. Catal., 1990,122,415. 123 Y.S. Yoon, N. Fujikawa, W. Ueda and Y. Moro-oka, Catal. Today, 1995, 24, 327. 124 M.C. Abello, M.F. Gomez and L.E. Cadus, Ind. Eng. Chem. Res., 1996,35,2137. 125 S.W. Lin, Thesis, Tokyo Institute of Technology, 1998. 126 H.X. Dai, C.F.Ng and C.T. Au, J. Catal., in press. 127 G. Blasse, J. Sytsma and L.H. Brixner, Chem. Phys. Lett., 1989,155,64. 128 W. Ueda and J.M. Thomas, Proc. 9th Int. Congr. Catal., Vol. 2, eds. J.M. Phillips and M. Ternan (Chem. Inst. of Canada, Ottawa, 1988), p. 960. 129 M. Al-Mamouri, P.P. Edwards, C. Greaves and M. Slaski, Nature (London), 1994,369,382. 130 Z. Hiroi, N. Kobayashi and M. Takano, Nature (London), 1997,371, 139. 131 L.G. Sillen, 2. Anorg. Allg. Chem., 1941,242,41. 132 L.G. Sillen, 2. Anorg. Allg. Chem., 1941,246, 115. 133 L.G. Sillen, 2. Anorg. Allg. Chem., 1941,246,331. 134 L.G. Sillen, Svensk. Kem. Tidskr., 1941,53, 39. 135 L.G. Sillen and B. Aurivillius, 2. Kristallogr., 1939,101,483. 136 L.G. Sillen and A.S. Gjoerling-Husberg, 2. Anorg. Allg. Chem., 1941,248, 121. 137 L.G. Sillen and A.S. Gjoerling-Husberg, 2. Anorg. Allg. Chem., 1941,248, 135. 138 L.G. Sillen and L. Melander, 2. Kristallogr., 1941,103,420. 139 L.G. Sillen and E. Jornstad, Z. Anorg. Allg. Chem., 1942,250, 367. 140 L.G. Sillen, Naturwissenschaften, 1942,22,318. 141 P.R. Garnier, J. Morean and J.R. Gavarri, Muter. Res. Bull., 1990,25,979. 142 A. Deschanver, J. Gallay, J.M. Hunout, M.Th. Thiault and C. Victor, C.R. Seances, Acad, Sci., 1970,27OC, 696. 143 C.N.R. Rao and J.M. Thomas, Acc. Chem. Res., 1985,18,113. 144 K.D.M. Harris, W. Ueda, J.M. Thomas and G.W. Smith, Angew. Chem. Int. Ed., 1988,27,1364. 145 B.Z. Nurgaliev, B.A. Popvkin and S.Yu. Stefanovich, Russ. J. Inor. Chem., 1983, 28, 1252. 146 B. Aurivillius, Chem. Scr., 1984,24, 125. 147 H. Kodama, S. Horiuchi and A. Watanabe, J. Solid State Chem., 1988,75,279. 148 B. Aurivillius, Arkiv. Kemi, 1950,463,499. 149 J.F. Ackerman, J. Solid State Chem., 1986,62,92. 150 Z. Hiroi, M. Takano, M. Asuma and Y. Takeda, Nature, 1993,364,315. 151 B. Tissue, K.M. Cilillo, J.C. Wright and D.C. Daeumling, Solid State Commun., 1988,65, 51. 152 W. Ueda and J.M. Thomas, J. Chem. SOC.Chem. Commun., 1988,1148. 153 W. Ueda, F. Sakyu, T. Isozaki, Y. Morikawa and J.M Thomas, Catal. Lett., 1991,10, 83.

6: Metal Halide Oxide Catalysts Active for Alkane Selective Oxidation

235

W. Ueda, T. Isozaki, F. Sakyu, S. Nishiyama and Y. Morikawa, Bull. Chem. SOC. Jpn., 1996,69,485. 155 K.D. Campbell, H. Zhang and J.H. Lunsford, J. Phys. Chem., 1988,92,750. 156 G . Hutchings, M.S. Scurrell and J.R. Woodhouse, J. Chem. Soc., Chem. Commun., 1988,253. 157 S.W. Lin, Y.C. Kim and W. Ueda, Bull. Chem. SOC.Jpn., 1998,71,1089. 158 S.W. Lin and W. Ueda, Bull. Chem. SOC.Jpn., 1999,72,997. 159 C.J. Milne, P. Lightfoot, J.D. Jorgensen and S. Short, J. Mater. Chem., 1995, 5 , 1419.

154

7 Selective Catalytic Reduction of Nitrogen Oxides by Ammonia BY SUNG-WON HAM AND IN-SIK NAM

1

Introduction

Selective catalytic reduction (SCR) of nitrogen oxides (NOx) by NH3 is the commercially proven technology for removing NOx from stationary sources, including utility and industrial boilers. Indeed, the emission of NOx, in both the flue gases from stationary combustion sources and the exhaust streams from mobile combustion sources has become an environmental problem of great concern in industrialized countries throughout the world. NOx poses a significant health hazard as an air pollutant’ and also plays a key role in photochemical reactions and the formation of smog. Still further, it may contribute to the precipitation of ‘acid rain’, particularly in the eastern United States, southeastern Canada, western and northern Europe and northeastern Asia. It is generally estimated that over 50% of the total NOx emissions in the advanced countries are from stationary sources, primarily utility and industrial boilers.’ For the level of control of NOx from these sources anticipated to be required by future emission regulations, it is widely believed that selective catalytic reduction (SCR) using ammonia offers the greatest promise for technical and economic success. The reduction of NOx by NH3 is selective in the presence of a catalyst and this process is termed ‘selective catalytic reduction’. When CO and H2 are employed as reductants, the process becomes nonselective catalytic reduction. It requires larger amounts of reductants than that for the stoichiometry of the NOx reduction reaction converting NOx with CO or H2 to harmless N2 and water on the surface of the catalyst. The term ‘selective’ means that the ammonia is not wasted by reactions other than the reduction of NO. The catalyst enables NH3 to selectively react with NOx. To attain the given selectivity, the catalyst should maintain the proper operating temperature, which must be high enough to provide useful NOx removal activity, but low enough to avoid the oxidation of ammonia. The technology as typically practiced removes 80-90% of the NOx, which is a higher conversion than that obtainable by any other commercially proven technology. In the past 20 years, growing public awareness of air pollution has led to a proliferation of studies related to selective catalytic reduction of NOx by NH3. The backbone of SCR technology research is the development of active Catalysis, Volume 16 0The Royal Society of Chemistry, 2002 236

7: SeIective Catalytic Reduction of Nitrogen Oxides by Ammonia

237

catalysts for the removal of NOx. A variety of SCR catalysts including noble metal^,^-^ transition metal oxide~,~-lO mixed metal oxides,’ and z e ~ l i t e ’ ~ - ~ ~ have been proposed for the reduction of NO with NH3. So far, V2O5 supported on Ti02 is known to be the most effective and widely employed commercial SCR catalyst due to its high NO removal activity and resistance to sulfur compounds commonly contained in the flue gas. On the basis of German experience, the SCR process installed at power plants increases the cost of electricity by 5%, while providing 80% NOx removal efficiency.26Approximately 50% of the ‘levelized cost’, i.e. the sum of operating and capital costs of SCR, is catalyst-related.26Therefore, a strong incentive still exists to improve the performance of SCR catalysts. The purpose of the present review is to elucidate the SCR system, primarily focusing on catalyst and process characteristics. It will give direction and challenge for the development of a ‘better’ SCR catalyst. Two distinctive commercial SCR catalysts, V205- and zeolite-based catalysts, are extensively examined for further development of SCR catalysts. The acidity and the structure of active components of the catalyst are illustrated as the important characteristics determining the NO removal activity of SCR catalyst. Attempts to improve the conventional V205-Ti02 catalyst by adding additives and altering the preparation method of Ti02 support are also included. Pillared clay, which is a zeolitic material, is a newly proposed SCR catalyst in the present paper. A catalyst employed in a commercial SCR process must possess high activity and selectivity, since the volume of the flue gas to be treated is extraordinarily large. Besides these requirements for SCR catalysts, any catalyst chosen for commercial use must be highly resistant to poisoning by sulfur compounds, mainly S02, which is commonly contained in the flue gas as well as NOx. The deactivation of an SCR catalyst is primarily due to the blocking and filling of catalyst pores caused by the formation of sulfate compounds on the catalyst The deactivation phenomena by SO2 is an important feature in developing an SCR catalyst commercially applicable to actual flue gas containing S02. SO3 formed by the oxidation of SO2 plays a crucial role in the formation of deactivating agents. Therefore, the oxidation capability of SO2 to SO3 is also a critical characteristic of SCR catalysts in determining catalyst life. For commercial application, the SCR system requires a low pressure drop reactor due to the severe pressure drop over the reactor containing the catalyst, creating the heavy operating costs of the p r o c e ~ s . ~A~ honeycomb * ~ * ~ ~ ~ type reactor is commonly employed as a commercial SCR reactor consisting of SCR catalyst. It offers several advantages such as low pressure-drop, high geometric surface area and short diffusion distance over the conventional fixed bed reactor packed with pellet and bead type catalyst. The modeling of the reaction kinetics and the reactor system is also included for the design of the commercial SCR reactor subject to the catalyst deactivation by SO2. Finally, the recent development of SCR by using hydrocarbons as reductants to substitute NH3 is examined as an alternative technology to NH3-SCR.

238

2

Catalysis

Reactions Involved

SCR technology has been proposed for the reduction of NOx by reducing agents such as NH3, CO, H2 and hydrocarbon~.'9~~$~~ NH3, in particular, shows high activity and selectivity for the reduction of NOx. Three categories of the overall reactions for selective catalytic reduction (SCR) of NO by NH3 are considered in the present chapter:

6 N 0 + 4NH3

-+

5N2+ 6H20

(1)

4NO + 4NH3+ O2 -+ 4N2+ 6H20

(2)

6 N 0 2 + 8NH3 -+ 7N2 + 12H20

(3)

2 N 0 2 + 4NH3+ O2 -+ 3N2 + 6H20

(4)

Since NO (nitric oxide) is the primary component of NOx emitted from combustion sources, reactions (1) and (2) are the main reactions which occur during the course of SCR reaction by NH3. Moreover, since the two main reactions (1) and (2) maintain the distinctive stoichiometric ratio of NH3 to NO, which of the two reactions occurs should be closely examined to avoid NH3 slip, thereby causing another air pollution problem.32Two other categories of side reactions can occur along with the above main reactions in the SCR process. The most undesired side reaction is the oxidation reaction of NH3 which becomes important at high reaction temperature^:^

4NH3 + 302 + 2N2 + 6H20

(5)

2NH3 + 2 0 2 + N20+ 3H20

(6)

4NH3 + 5 0 2 + 4 N 0 + 6H20

(7) Since these NH3 oxidation reactions uselessly consume NH3, which should be utilized for the reduction of NO, the overall NOx conversion decreases at high reaction temperature and consequently shows a bell-shaped activity curve with the increase of reaction temperatures, This bell-shaped activity curve is typical for most SCR They even produce NO or N20 which should be eliminated in the main reaction. The other undesired side reaction may be the oxidation reaction of SO2: SO2 +1/202 -+ SO3 + 98 kJ/ mol-'

(8) SO3 produced by the oxidation of SO2 can react with NH3 and water, and then form ammonium sulfate, ammonium bisulfate and other ammonium salts including (NH4)S207 etc. according to the following reactions:

NH3+ SO3+ H 2 0 + NH4HS04 2NH3+ SO3+ H 2 0 + (NH4)2S04

(9)

(10) The formation of these salts is an equilibrium process largely depending on the concentration of NH3, SO3 and reaction temperature. Since these ammonium salts can reduce the catalytic performance by plugging the catalyst pores and

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

239

cause plugging or corrosion problems downstream of the SCR reactor, the concentration of SO3 should be maintained as low as possible to reduce the possibility of these ammonium salts forming. More details on SO2 oxidation will be discussed later.

3

SCR Catalysts

3.1 VzOsbased Catalysts. - The V2O5 catalyst supported on Ti02 for the selective reduction of NO by NH3 has been widely employed as a commercial catalyst due to its high NOx removal activity.33Another reason for employing V2O5/TiO2 as an SCR catalyst may be chemical properties of titania support for the present catalytic system. Since SO2 is also contained in the flue gas from power plants as well as NO, it can be easily oxidized to SO3 to form sulfate on the catalyst surface. TiOz, however, can be weakly and reversibly sulfated under the operating conditions of the SCR reaction. Furthermore, the stability of sulfates on the surface of Ti02 is weaker than that on any other metal oxide supports, including A1203 and Zr02. Thus, Ti02-based commercial catalyst can be partially and reversibly sulfated under SCR reaction conditions containing S02, and the sulfation of the catalyst even enhances NO removal activity.34 In addition, Ti02 can contain highly dispersed surface vanadia species generating isolated vanadyl species and polymeric vanadate species, which are known as the active reaction sites for SCR r e a c t i ~ n . ~ ~ - ~ ~ For most commercial SCR catalysts, it has been observed that a monolayer of vanadium and tungsten oxide is formed on the surface of Ti02.38 The amount of vanadium loading varies with respect to the operating conditions of the SCR process. However, V205is generally loaded in the range 1-2 wt.% of the catalyst. Vanadia is the active reaction site mainly responsible for the NO removal activity of the catalyst as well as for the oxidation of SO2 to SO3 when SO2 exists in the flue gas along with NO. It has been reported that W03 is loaded in a range of content up to 10% to improve catalytic activity and thermal stability and to suppress the oxidation of S02.39 Promoted V2O5 catalyst supported on Ti02 has been commonly employed for commercial chemical processes including the oxidation of hydrocarbons and reduction of NOx by NH3. However, the effect of additives on the catalytic performance of the promoted catalyst has been superficially understood without systematic studies of the catalytic characteristics including the state of metal on the catalyst surface. While some of the additives reduced the activity of the catalyst, others enhanced the catalytic activity, selectivity and life. In addition, the amount of additive is often critical on the catalyst surface to prepare a 'better' ~atalyst.~' The typical additives commonly employed for supported vanadia catalyst are tungsten, niobium, molybdenum, silicon, potassium and phosphorus. The additives do not improve the redox capability of the catalyst but create additional acid sites on the catalyst surface.41Similar to the vanadia species, these promoters preferentially interact with the surface of titania to form two-

240

Catalysis

dimensional surface metal oxide species with vanadia. Based on a recent report,42 these additives indirectly or directly contact the surface vanadia species laterally. The lateral interactions may reduce the effect of the ratio of polymerized to isolated surface vanadia species on the catalytic performance, although vanadia sites can also be occupied by a promoter. The influence of promoters on the reaction varies with respect to the number of sites involved in the reaction mechanism. For a simple reaction requiring one surface vanadia site, the role of the promoter may be negligible, since they do not generate the redox sites and the reaction only proceeds on the surface vanadia site. However, the acidic characteristics of the promoters can produce a variety of by-product^.^^ For a reaction requiring two adjacent surface sites, the improvement of reaction rate and selectivity has been observed by the presence of promoters on the catalyst surface. Van Hengstum et aZ.44reported that the effect of phosphorus and potassium additives on V205/Ti02 catalyst totally depended upon the type of the reactants. They observed that the additives rarely contributed to enhancing the catalytic activity and yield. Similar results have been also observed for catalyst containing a high content of vanadium oxide for the oxidation of o-xylene. The addition of phosphorus to V205/Ti02 catalyst promoted side reactions since it increased the surface acidity of the catalyst. Potassium altered the nature of the active reaction site, creating the possible formation of an amorphous alloy of vanadium oxide with potassium. Zhu and Anderson45 studied the effect of potassium and phosphorus additives on the surface of V205/Ti02catalyst and observed that the reaction rate rapidly decreased when the additives were loaded. The addition of potassium can also suppress the side reaction on the acidic site and enhance the combustion reaction. The addition of phosphorus, however, improved the selectivity of the desired products by the primary reaction and deteriorated the selectivity of the products. The characterization of the phase of the surface vanadium oxide for potassium and phosphorus-doped V2O5/TiO2 catalyst by XPS, XRD and infrared spectroscopy indicated that the additives and vanadium oxide are highly dispersed on the catalyst surface.46In particular, the formation of small quantities of KV03 with potassium was detected by XRD. An excess amount of potassium bonded to vanadium and resulted in a significant decrease in surface area and the transformation of Ti02 from anatase to rutile. The excessive addition of phosphorus created agglomeration of the phosphorusvanadium-oxygen phase and formation of vanadium phosphate compound identified by XRD. Bond and Tahifl7 investigated the influence of potassium and phosphorus for V205/Ti02 catalyst on the oxidation reaction. The addition of potassium exhibited little effect on the catalyst reducibility or activity. Phosphorus is generally known to decrease the catalyst reducibility and activity. Vuurman et aZ.48 examined V205-W03/Ti02 catalyst by using Raman spectroscopy under ambient and dehydrated conditions as shown in Figure 1. They reported that the structure of the surface vanadium oxide species can be seriously altered by the addition of tungsten oxide, which changes the surface

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

24 1

3% 1%

O C

wo

loo0

-

.

so0

Raman Shift (em')

11% V,O,-X% WOJTLO, dehydrated X%

wo, %

u l

1%

0%

I

I

loo0

I

1

I

I

800

Raman Shift (cm-I)

Figure 1

Raman spectra of 1% V,05 catalysts supported on Ti02 with respect to the loadings of tungsten up to 1009: (a) ambient and (b) dehydrated condition^^^

acidity of titania under ambient conditions. It has been confirmed that the vanadium and tungsten oxides on the surface of Ti02 exist as separate species and that no influence of the second metal oxide by the loadings of tungsten species is observed under dehydrated conditions. The sequence of the impregnation of the metal oxides did not affect the nature of the surface vanadium and tungsten oxides species. Ramis et aZ.49 also examined V205-W03/Ti02 catalyst by infrared spectroscopy under dehydrated conditions and showed the

Catalysis

242 0

TlTANlA SUPPORT

M,O,

= WO,

or NbO, or SiO,

(a)

TlTANlA SUPPORT

(b) 0

w

w

0

u

w

TITANIA SUPPORT

Figure 2

w

w

pox

Schematics of the interactions of promoters with the surface vanadium species on titania: (a) non- interacting promoters, (b) interacting promoters and (c) interactingpromoters

presence of V=O and W=O species on the catalyst surface in the identical position as in the single metal oxide system for V205/Ti02 and WO3/TiO2. They suggested that the surface vanadium and tungsten oxides do not strongly interact with each other. Deo and Wachsso reported that the promoters on vanadia-supported titania catalyst are classified into two types, non-interacting (W03,Nb2O5 and Si02) and interacting (K20, P2O5) promoters with respect to the surface vanadium oxide phase, which have been extensively studied by Raman (Figure 2). Non-interacting promoters simply coordinate to the support without direct contact with the surface vanadium. However, the introduction of interacting species on the supported catalyst significantly modified the phase of the surface vanadium to improve the activity and selectivity of the catalytic system. Chen and Yang5' have examined the effect of W 0 3 on the V205/Ti02

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

243

catalyst. They observed that the addition of W 0 3 enhanced the activity of SCR reaction and significantly improved the tolerance of W03-V205/Ti02 catalyst to alkali metals also contained in the flue gas. An increase of Brarnsted acidity was also observed upon the addition of W 0 3 to V205/Ti02catalyst; it is generally regarded as the active reaction site for SCR reaction. Thus, it can be concluded that additives to the V205/Ti02catalyst for the reaction of NO by NH3 could affect the catalytic performance by altering the surface phases of V205.There have been studies on the correlation of the structure of the surface vanadium oxide phase with the catalytic activity of the supported vanadium oxide catalysts containing a variety of additive^.^^.^^ They may provide a guideline for the molecular design of the supported vanadia catalyst. The number of active reaction sites is determined by the coverage of vanadia existing in the form of a two-dimensional layer. The specific activity of the active reaction sites is not a function of the coverage of vanadia on the catalyst surface, but a strong function of the number of isolated surface vanadia sites on the catalyst surface. However, the catalytic activity for the SCR reaction requires more than two adjacent reaction sites and totally depends on the coverage of the catalyst surface by vanadia. The addition of the promoters to V2O5 catalyst may create a variety of active reaction sites on the catalyst surface. 3.2 Zeolite Catalysts. - Due to their peculiar characteristics such as ionexchange capability, acidic properties, higher surface area with small pore size and high thermal stability, transition metal ion exchanged zeolite catalysts have been extensively examined and reported to be quite effective for SCR reactions. 12-25 Among many transition metal exchanged zeolites, Cu-exchanged zeolite has attracted interest because of its unique catalytic activity, which widens the operating temperature window, as well as its high performance for SCR reactions. The zeolite itself also showed a significant activity for NO reduction with NH3, to some extent due to the presence of Brransted acid sites where NH3 can adsorb. The addition of copper to zeolite greatly improved catalytic reduction of NO by NH3 by increasing the adsorption sites for NH3 forming coppercomplex, [ C U ( N H ~ ) ~ ]This ~ + . ~complex ~ was confirmed to be a reaction intermediate under SCR reaction conditions over copper-exchanged zeolites by X-ray absorption spectro~copy,~~ and electron paramagnetic resonance s p e c t r o ~ c o p yFor . ~ ~other transition metal exchanged zeolites, the formation of a [rnetal(NH3)(NO),] complex was also observed under steady-state reaction conditions by dynamic IR studies.I5 The catalytic activity of copper-exchanged Y zeolite showed a bell-shaped temperature dependence in the absence of oxygen with the maximum activity appearing at ca. 110-120 OC.13914At higher reaction temperatures, the reaction mechanism includes a redox step where the oxidation rate of the Cu+ complex is much slower than that of Cu2+ reduction. Williamson and Lunsford13 proposed a reaction mechanism that the breaking of N-H bond is the ratelimiting step at lower temperatures, while the reoxidation of Cu+ to Cu2+is the 14~17718?54-57

244

Catalysis

rate-limiting step at higher temperatures. Mizumoto et al. 21 also proposed a similar reaction mechanism over CuY zeolite based upon the alteration of the rate-limiting step with respect to reaction temperatures. In the presence of oxygen, copper-exchanged zeolites also show a bellshaped activity with respect to the reaction temperatures. But, in this case, the reason for maximum activity is simply due to ammonia oxidation reactions becoming important at higher reaction temperatures. In general, the operating window of the reaction temperature is determined by the oxidation activity of the catalyst. To broaden the window of operating temperature for NOx removal, Medros et al. l 8 have developed a dual catalyst bed containing copper-exchanged and non-exchanged mordenite catalysts. 3.3 Newly Proposed Catalysts. - Titania as a popular support for SCR catalysts still suffers from limited surface area and pore structure, lack of abrasion resistance and poor mechanical strength. The anatase type of titania has poor thermal stability. The stability of a catalyst is of vital importance, since the catalyst life mainly depends on the thermal stability of the catalyst. In order to overcome these drawbacks of titania, an attempt to improve the physical properties of titania has been recently carried out in two ways. One is in the synthesis of titania by the alteration of the preparation method. The other is the development of a mixed oxide material including titania as a minor or major component of the catalyst. The method for the preparation of Ti02 generally involves varying the synthetic technique for titania supports or altering the choice of the precursor containing Ti02. For the synthesis of titania, the most common technique may be the hydrolysis of TiC14 and titanium alkoxides or the precipitation of Ti(S04)2.26*59y60 The sol-gel method was also employed.61-63 Recently, Ciambelli et aLa suggested the pyrolysis of titanium alkoxides by COZ laser for the preparation of monocrystalline Ti02. However, these modifications of titania cannot be used to produce commercial Ti02 for use as a catalyst support due to the limitation of the control of physicochemical properties and the high cost of the manufacturing process. A mixed oxide support containing titania as a major or minor component reveals the peculiar support characteristics including the superior thermal stability over a single titanium oxide. Shikada et al.65 observed the beneficial effect of titania added to silica for the SCR reaction. Reddy et al.66 have proposed that TiOz-Si02 is a good support for vanadia-based catalysts for NOx reduction with NH3, since the composite binary oxide support can combine both the mechanical properties of silica and the chemical characteristics of titania. Clarebout et al.67 and Matralis et al.68 have reported that vanadia supported on mixed Ti02-A1203 carriers also exhibits significant NO removal activity, strong sulfur tolerance, superior mechanical strength and thermal stability compared to the conventional V205/Ti02catalyst. Figure 3 shows NO removal activity over vanadia catalysts supported on mixed Ti02-Al203 with respect to the content of alumina in the catalyst. Compared to vanadia catalyst supported on pure titania, significant enhancement of SCR

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

0'

245

I

0

20

40

60

80

100

% Wt AI,O,

Figure 3

NO removal activity of V20~ITiO,A1203 catalysts with respect to the content of alumina on the supports: 0 at 150 "C, at 200 "C, o at 250 0C68

activity has been observed over V205/TiO2-Al2O3 catalyst as revealed in Figure 3. Although the introduction of alumina into the matrix of titania increases the amount of Brarnsted acidity on the support which results in enhancement of the SCR activity of the catalyst, alumina introduced beyond a certain level in the support, rather weakens the interaction of titania with vanadia, which decreases the SCR activity. Therefore, there is an optimal content of alumina in mixed Ti02-A1203 support for SCR catalysts as shown in Figure 3. The PILCs (pillared interlayered clays) intercalated by titania have recently been proposed as a composite support for SCR catalyst to replace titania.69-70 PILCs, commonly called a bidimensional zeolite, consist of porous structures where the pore size varies from 0.2 to > 2 nm larger than that of zeolites (0.2-0.8 nm). The surface acidity is as strong as that of Y-zeolite. Na+ ions existing in the interlayer of the original clay can be easily exchanged by the cations of A1 hydroxyl cluster. When it is calcined at 30O-50O0C, the cations become pillars in the form of oxide releasing protons. An acidic and porous pillared clay containing a surface area of 200-500 m2 g-' is thus formed. In addition, the basal spacing, do01 of clay increased to 18-19 from the 12 A of the original space after pillaring and retained this space even after the calcination of PILC at 400 "C.The high space of the basal line simply indicates the formation of thermally stable pillared structures, and the tetrahedrally substituted clay possesses strong and dense Brarnsted acid sites on its surface. As the pillars open the layers of the clay, the accessibility of the reacting molecules to the interlamellar catalytic sites increases, creating a high catalytic activity. Simultaneously, the pillar can exert a shape-selective role that controls the diffusion rates of reactants and products or the formation of the reaction intermediates. Thus, pillared clay is also regarded as a shape-selective catalyst.

A

Catalysis

246

It was found that if the cationic species existing as the pillar itself were catalytically active, the PILC would also be catalytically active. Most of the reactions using PILC as a catalyst take advantage of its characteristics including the porous structure and the surface a ~ i d i t y .PILCs ~ ~ , ~containing ~ high surface area can also be used as a support for metal or metal oxide catalyst to achieve homogeneous dispersion of metals on the surface of the support, increased adsorption area of the reactants and high shape-selectivity. The studies on PILC have been mainly carried out to confirm its capability as a novel zeolite-like material. PILCs containing a variety of metal oxides as a pillar may exhibit a variety of applications to a chemical process. However, few studies on the use of PILCs have been reported so far. In particular, the performance of an SCR catalyst is significantly influenced by the morphology and surface acidity of the catalyst. Thus, PILCs containing peculiar physicochemical properties including pore structure and acidity may be promising as a SCR catalyst itself or its support. In fact, Yang and cow o r k e r ~ have ~ ~ . examined ~~ the application of PILC to SCR reactions. They suggested that the delaminated Fe-PILC and iron and chrome oxides supported on Ti-PILC exhibit superior NO removal activity to the conventional Ti02 based catalysts. Further developments of Ti-PILCs as a support for vanadia-based SCR catalysts to overcome the drawbacks of titania have been reported.69570'74-76 The high deNOx performance of V205/Ti-PILC has been recently observed. Long and Yang74 and Chae et have reported the high deNOx performance of V205/Ti-PILC compared to that of commercial V205-W03/Ti02. Figure 4 shows NO removal activity for different supports, including a commercial SCR catalyst. The NO removal activity of vanadia catalyst on Ti02-based supports is superior to that on the other materials aZ.69770775976

1.0 0

z b

0

'S Q)

0

-

0.6 0.4 -

0.8

0

+Commercial 0.0

L

150

200 250 300 350 400 450

500

Temperature ("C)

Figure 4

NO removal activity of V2OsITi-PILC, V20,1Ti02 and V205- WOjl Ti0 2Catalysts 75 ,76 6977Q9

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

247

including alumina, zirconia and bentonite. Among vanadia catalysts on titania-based supports, V205Ki-PILC catalyst exhibits the highest NO removal activity. Moreover, without any catalyst promoters such as W 0 3 and Mo03, V205/Ti-PILC catalyst reveals superior NO removal activity to a commercial V205-WOJI'i02 catalyst, particularly in the range of reaction temperature lower than 350 "C. In particular, Chae et al. have examined the dependence of the NO removal activity on the physicochemical characteristics of Ti-PILC such as the surface acidity and morphology of PILC. They have also confirmed that the deactivation of Ti-PILC catalyst by SO2 can be improved due to its peculiar pore structure. This may be another feature of PILC regarded as a promising SCR catalyst. It is hoped that Ti-PILC based catalyst may be utilized as a commercial SCR catalyst substituting for the current Ti02 based catalyst in the near future. 69975976

4

Catalyst Poisoning and Deactivation

A distinction can be made between the catalytic process for environmental control and that for industrial production. In a manufacturing process, temperature, pressure, concentration and all other operating conditions should be optimized in order to attain the maximum performance of the catalyst and to produce the desirable product with the highest efficiency. On the other hand, the catalyst in an emission control process has to be employed under the conditions of the effluent gas to be treated. It is, therefore, of vital importance that the catalyst should be matched to the process conditions. In addition to a high activity and selectivity of environmental catalysts summarized in Figure 5, resistance to thermal shock and abrason is also required. The catalyst should be tolerant to other compounds in the flue gas. An SCR catalyst should also contain the above-mentioned characteristics for its commercial application. Since the cost of the catalyst is a major element in SCR processes, it is important to design the catalyst and reactor to maintain High Selectivity

Long Life

High Strength Thermal Shock Resistance Low pressure Drop

Figure 5

Requirementfor a pollution control catalyst

248

Catalysis

catalyst activity as long as possible. A common engineering goal is to establish a catalyst life of 3-5, or even 10 years, depending on the severity of the service. The anticipated deactivation mechanisms of SCR catalysts include: 0

0

0 0

Fouling caused by deposition of fly ash or ammonium bisulfate on the external and internal surfaces of the catalyst containing the major portion of the pores Chemical deactivation: reaction of alkali or alkaline earth oxides, or arsenic compounds with the catalytically active reaction sites on the catalyst surface. Sulfur oxides are the primary cause of deactivation for SCR processes Hydrothermal stability Abrasion by fly ash

4.1 Sulfur Tolerance of SCR Catalysts. - It is well known that sulfur compounds deactivate the catalyst for the reduction of NOx with NH3. The deactivation of SCR catalysts is primarily due to the blocking and filling of catalyst pores caused by the formation of sulfate compounds on catalyst support or active components from the oxidation of The cause of the deactivation of SCR catalysts by the deposition of sulfur compounds also depends on the catalyst supports. The sulfur tolerance of the catalyst supports is as follows:7o S02.9224327378779

Ti02 = Si02 > a-A1203> q-Al203 > y-A1203

(1 1)

Alumina is a good support for an SCR catalyst. However, it is readily deactivated by the formation of aluminum sulfate [A12(SO4)3], resulting in pore blocking which reduces the performance of the ~ a t a l y s t . ~ Table 1 shows the effect of sulfur on the lifetime of SCR catalysts reported in the literature. V2OJTiO2 and W03/Ti02 catalysts seem to be the most suitable SCR catalysts. This is simply attributed to the resistance of Ti02 against sulfur compounds. Saur et aZ.84claimed that the resistance of Ti02 to SO2 or SO3 is ascribed to the relatively low decomposition temperature of the sulfates formed on the surface of Ti02 during the course of the reaction. Many inve~tigators~~-~O also reported the pore filling and blocking phenomena by ammonium salts such as (NH&SO4 and NH4HS04 formed by the reaction among SO3, H20 and NH3. Matsuda et aZ.85observed the deposition of ammonium bisulfate on Ti02-based catalysts even at a higher temperature than expected based upon the thermodynamic characteristics of the sulfate. They explained these phenomena by the capillary condensation theory in the micropores of the catalyst. The formation of ammonium sulfate and ammonium bisulfate is an equilibrium process and depends largely on the concentration of NH3, SO3,water and reaction t e m p e r a t ~ r e . ~ ~ The supported noble metal catalysts including Pt, Pd, and Ru were extensively studied in an early stage of the development of SCR catalysts and were confirmed to be effective for SCR r e a ~ t i o n .These ~ ~ . ~catalysts, ~ however, were easily deactivated by SO2 in the NO-NH3-S02 system. Sulfur exists in the form of sulfide on the catalyst surface as confirmed by Auger spectro-

a

real flue-gas conditions.

-

1000 700 700 700 1000 300-400 300 20-80 1000

-

30-40 2-10 0.1-3

-

-

-

so3

so2

Life of SCR catalysts

V20s/Ti02 V205/Ti02-Si02 V205/Si02 V20s/A1203 W03/Ti02 W03/Fe203 W03/Ti02 CuO/A1203 Zeolite

Catalyst

Table 1

Gases

150 300 300 300 150 200 230-300 65 500

NO

1 1.3 1.3 1.3 1 1.1 1 0.9 1

NHjINO 6600a 60 000 30 000 30 000 2700" 44000" 10 000" 3200-10 800" 6000

(h-9

Space velocity

500 25 17 8 208 9 280 145 17

(days)

Time on stream

93-85 80-60 95-85 100-20 90-90 90-65 >90 70-80 >95

(%)

350 230 230 230 400 380 380 340-370 373

("C )

Change of NO Temperature conversion

83 16

;;

27 80 80 80 81

Refs.

3

$

%

&g*

@

$

sis

n

$. e

2

7

250

Catalysis

s c ~ p yThe . ~ ~addition of 0.5% oxygen to the feed gas stream almost restored the initial activity of the catalysts when the feed of SO2 terminated. The regeneration of the deactivated catalysts is mainly due to the removal of sulfur in the form of metal sulfide by 02.It cannot be excluded that 0 2 can form sulfates with SO2, a primary deactivation precursor for SCR reaction. For zeolite catalysts, Kiovsky et al. l 6 reported that NC-300 catalyst (hydrogen mordenite) has shown a high performance for NO removal without any decline in its activity at 370°C for over 400h, although the activity has significantly declined at reaction temperatures lower than 370 "C in the presence of 2000 ppm of SO2 in the feed gas stream. However, they did not discuss the reasons for the activity loss depending on reaction temperatures in the presence of S02. On the other hand, there are a few reports that SO2 in the feed gas promotes the SCR activity at high reaction temperatures where NH3 oxidation reaction is prevailing. When SO2 and SO3 exist in a flue gas, they can react with oxygen to form S042-, which may be adsorbed on the catalyst surface and combined with the chemical component of the catalyst. Okazaki et al.94studied the effect of S042- on the activity for NO reduction with NH3 from the viewpoint of surface acidity. The activity of CuOx and MnOx supported on Ti02 significantly decreased by the addition of a small amount of S042- due to the formation of CuS04 and MnS04. In contrast, for MoOx and VOx supported on Ti02, its activity for NO reduction was improved, probably due to the increase of the surface acidity. The addition of S042- enhances the NO removal activity by facilitating the adsorption of NH3 and the formation of NH4+ on the catalyst surface. It is believed, however, that the enhancement of SCR activity by the formation of S042- on the catalyst surface, is a transient behavior. The catalyst will eventually lose its activity as the sulfur content increases on the catalyst surface. reported that the addition of SO2 in the feed stream Chen et increased the SCR activity of V205/Ti02 catalyst. They also reported that the sulfation reaction forming S042- by the reaction of SO2 and O2 on the surface of Ti02 became important as the temperature increased above 400°C where the activity of V2O5/TiO2 catalyst declines at this temperature. They explained that the reason for the promotion of SCR activity by SO2 is the formation of surface S042- that would improve the surface acidity, especially increasing the Brransted acid site on the catalyst surface. With another type of catalyst, including perovskite-related structure and latielite catalysts, the promoting effect of SO2 was also r e p ~ r t e d . It ~ ~should . ~ ~ be noted that the promoting effect of SO2 was mainly attributed to the suppression of NH3 oxidation reaction at higher reaction temperatures where the NH3 oxidation reaction predominantly occurs. aZ.51%95996

SO2 Oxidation. - A commercial SCR catalyst requires high activity for NO removal in the wide range of the reaction temperature, high thermal stability and low SO2 oxidation activity. In particular, sulfur trioxide (SO3) produced by the oxidation of SO2 along with a small amount of SO3 originally

4.2

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

25 1

existing in the flue gas is known to react with ammonia and water to form NH3-S03 compounds such as ammonium sulfate [(NH4)2S04]and ammonium bisulfate (NH4HS04) according to reactions (9) and (10). These ammonium salts may deposit on the catalyst surface, causing catalyst deactivation, and on the equipment downstream of the SCR reactor, creating corrosion and plugging problems. Therefore, one of the important features for SCR catalyst to be minimized is the oxidation capability of the catalyst for the conversion of SO2 in the flue gas to SO3. It is so important that industrial specifications for SCR process typically include upper limits for both NH3 slip and SO3 concentration downstream of the SCR reactor. The outlet concentration of SO3 corresponding to the conversion of SO2 during the course of the reaction is in the range 1-2% of the feed concentration of SO2 to the reactor. The chemical equilibrium of the reaction forming SO3 shifts to the lower temperatures of the reaction, since it is an exothermic reaction (reaction 8). The equilibrium constant decreases approximately by two orders of magnitude when the temperature is increased from 250 to 400°C. Nevertheless, the oxidation of SO2 to SO3 is thermodynamically favored over the entire reaction temperature range typically operated for an SCR reaction. While a great many scientific and technical reports on SCR technology have been published, a study addressing the oxidation of SO2 to SO3 over SCR systems is rare in the literature. This is probably due to the difficulty in the systematic measurement of the concentration of SO3 downstream of reactor, since the SO3 concentration is limited to only a few percent of the feed concentration of S02. The concentration of SO3 cannot be analysed by a conventional GC and/or IR detector. Therefore, SO3 in the downstream of the reactor is absorbed into a solution of 80% isopropyl alcohol in deionized ~ a t e r . For ~ ~ the , ~ ~measurement of SO3 concentration, an aliquot of the absorbing solution is titrated with 0.0 1N barium perchlorate [Ba(C104)2.3H20] by using a thorin indicator. The most important aspect determining the activity for SO2 oxidation is the composition of the catalyst. An SCR catalyst is also an oxidation catalyst and can easily oxidize SO2 to SO3 during the course of the reaction. This is especially true for V2O5 catalyst, one of the most widely used catalysts for commercial SCR and oxidation processes. Although V205/Ti02 catalyst is commonly employed as the commercial catalyst for the reduction of NO by NH3, the content of vanadium loadings should be adjusted to suppress the oxidation reaction of SO2.The activity for the SO2 oxidation reaction increases with the increase of vanadium content of V205/Ti02 catalyst. Attempts had been made to reduce the oxidation of SO2 to SO3 over V205/Ti02 catalyst by adding a second metal to the catalyst. Morikawa et a2.97,98 reported that SO2 oxidation reaction could be retarded by decreasing the surface excess oxygen on the catalyst, in other words, by adding the oxides, which are composed of cations in low valence states of metals such as Ge02 and ZnO. They did not observe the formation of ZnS04 on the catalyst surface. The promoted V205/Ti02catalyst reveals higher oxidation activity of SO2 to SO3 than the catalyst without promoters. Sazonova et aZ.,99 Dunn

252

Catalysis

et al. loo and Choo'O' have similarly discussed the role of the promoters for the oxidation of SO2. There is still controversy on the effect of promoter on the oxidation of SO2 to SO3, which is critical for the sulfur tolerance of an SCR catalyst. Besides altering the chemical composition of the catalyst, the activity of SO2 oxidation reaction could also be minimized by modifying the pore structure of the catalyst.lo2 The results indicate that substantial reconfiguration of the pore structure of the catalyst could enhance NOx removal activity by about so%, while simultaneously improving the catalyst deactivation and thereby extending catalyst life. It should be noted that the undesired SO2 oxidation reaction is not enhanced by the modification of the pore structure configuration. However, this can minimize the catalyst deactivation by pore blocking and plugging and eventually extend the catalyst life.9y913lo3

5

Reaction Mechanism and Kinetics

During the past few decades, much effort has been focused on the development of active catalysts for the removal of NO. A variety of SCR catalysts including noble metals, transition metal oxides, mixed metal oxides and zeolite have been proposed for the reduction of NO with NH3. Among them, vanadium oxide supported on Ti02 is the best catalyst for the reduction of nitric oxide from stationary sources due to its high activity at relatively low reaction temperatures (< 400 "C) and its high resistance to the catalyst deactivation by SO2. The reaction mechanism for SCR systems has been discussed since the 1970s, mostly on V2O5/TiO2 catalysts. However, the mechanism of the reaction over commercial SCR catalysts consisting of V2O5 and/or W 0 3 or Moo3 on Ti02 has not been completely clarified yet. Two major mechanisms have been proposed in the literature. Takagi et al.'04*105 suggested that NO-NH3 reaction in the presence of oxygen over bulk V2O5 and V205/A1203 proceeds via adsorbed species of NO2 and NH4+ which react on the catalyst surface to form nitrogen and water through Langmuir-Hinshelwood mechanism. However, Inomata et al. argued that the oxidation of NO to NO2 can easily occur at room temperature with high partial pressures of NO and 02, but it can hardly occur under dilute gas composition, characteristic of the SCR process. They proposed the following mechanism for the reduction of NO and ammonia over V205-based catalysts in the presence of oxygen under dilute gas conditions: (i) ammonia is first strongly adsorbed adjacent to V5+=0as NH4+(ad), (ii) then bulk NO reacts with NH4+(ad) according to Eley-Rideal mechanism to form nitrogen, water and V-OH, (iii) V-OH species are eventually reoxidized to V5+=0 by either gaseous oxygen or bulk oxygen from V5+=0 specie^.^,'^^ They claimed that V=O species are active reaction sites for the SCR reaction and the activity of catalyst is proportional to its density on the catalyst s u r f a ~ e . ' ~ ~ - ~ ~ ~ Recently, however, Bronsted acid sites of V-OH groups have been suggested as reaction sites for SCR r e a c t i ~ n . ~ 'This * ~ ~is* quite ~ ~ consistent with IR

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

253

results that the NH4+, i.e. NH3 chemisorbed on Brarnsted sites, is the primary reacting species for the SCR reaction system. The other is the Langmuir-Hinshelwood mechanism in which an oxidationreduction reaction scheme has been proposed for the V2O5/TiO2 catalyst, where the catalyst is reduced by NH3 and reoxidized by NO or 02.110-113 Noble metals such as Pt, Pd, Ru and Rh are also known to be effective for 14-1 l7 SCR reaction with NH3 by the Langmuir-Hinshelwood me~hanism.~-~.' Otto et al. l 4 proposed a mechanism involving the dissociative adsorption of NH3, leading to the formation (NH2)ads and (H)ads and the associative adsorption of NO. They suggested that the reaction of these adsorbed species through the Langmuir-Hinshelwood mechanism leads to the formation of N2, N 2 0 and H 2 0 as follows:

where S stands for an active reaction site. They also observed that the ratedetermining step is dissociation of NH3 by the isotopic study. Meier and Gut4 reported that the reaction of NO with NH3 in the presence of oxygen over alumina-supported platinum catalyst was well described by a dual site Langmuir-Hinshelwood mechanism involving associative adsorption of ammonia and nitric oxide on the identical reaction sites and associative adsorption of oxygen on the distinctive reaction sites. The dissociative adsorption of ammonia as proposed by Otto et al. l5 could not be confirmed. Many investigator^^.^^'^'^^'^^,^ l8 developed a variety of SCR systems, commonly pointing out that the presence of oxygen in the feed gas stream plays an important role for the rate of reduction of NO with NH3. Markvart and Pour reported the increase of reaction rate when oxygen was present in the reactor system.2 The adsorbed NH3 dissociates to H and NH2 fragments on the surface of platinum. Oxygen reacts with one of these fragments; thus the rate of the reaction of the remaining fragment with nitric oxide is also enhanced. They concluded that oxygen in the feed accelerated the SCR reaction by enhancing the dissociation of the ammonia, increasing the amount of NH3 adsorption on the catalyst surface. From IR studies, Takagi et al. lo4 observed that NO reacts readily with NH3 in the presence of oxygen via adsorbed NO and adsorbed NH4+ on the surface of V2O5 catalyst. On the other hand, Inomata et al. proposed that the gaseous bulk NO reacts with the adsorbed NH3, i.e. NH4 adsorbed adjacent to V5+=0 by Eley-Rideal mechanism. In order to understand the role of oxygen for zeolitic catalysts, an SCR reaction was carried out in the absence of 0 2 with either only NO or NO2 using N2 as a carrier gas. Kiovsky et a l l 6 suggested that the role of oxygen is the oxidation of NO to NO2, which is a more reactive species than NO, based on the result that the complete reduction of NO2 existing in the feed gas stream can be easily achieved, whereas no reduction of NO is observed. Amidiris et aZ.19 also reported that NO removal activity for iron-Y zeolite l49l

Catalysis

254

was enhanced by about two orders of magnitude in the presence of oxygen. In this case, the rate-limiting step appears to be the reoxidation of iron in the absence of oxygen. There were several more studies for the effect of oxygen in the literature, although the function of oxygen varies with the catalytic and experimental systems employed.6-8*1

6

Characteristicsof an SCR Catalyst

6.1 Surface Acidity. - Acidity is one of the most important characteristics of a catalyst, which may make it catalytically effective. l9 Particularly in the SCR system, NH3, which is a common probe molecule for the measurement of catalyst acidity, is one of the major reactants. It may be of interest to correlate the catalytic activity of the SCR reaction with the acidity of the catalyst. To investigate the strength? number and the structure of acid sites on the catalyst surface, a variety of methods for acidity measurement can be utilized, such as titration, calorimetry and infrared spectroscopy.120 All these methods have their own advantages as a way to examine the acidity, quantitatively and qualitatively. A restriction that has been frequently encountered concerns the size of the base molecule, as well as its strength. From such a viewpoint, temperature programmed desorption (TPD) of NH3 may be a fair method for measuring the acidity of the catalyst surface.58Although this method strongly depends on the experimental conditions, including the rate of temperature increase and equilibrium conditions? which makes comparisons difficult, it is still considered a good method for examining the amount and the distribution of acid sites on the catalyst surface. NH3 TPD will provide information on catalyst acidity including the number of acid sites and the distribution of acid strength as well as the adsorption property of NH3 on the catalyst surface. There have been many efforts to correlate the catalytic activity of the SCR reaction with the surface acidity of the catalyst. Choi et al. 58 observed that unexchanged mordenite catalyst, H mordenite (HM), exhibited substantial NOx removal activity for a certain region of reaction temperatures. The NO removal activity of NaHM prepared by ionexchanging sodium ions (Na+) in mordenite by protons (H+) increased with the increase of the level of H+ ion exchange. The number of acid sites was also enhanced with the content of H+ ion exchanged, which indicates that the acid sites acted as active reaction sites where NH3 was adsorbed in the SCR system. It has been well known that the number of Brsnsted acid sites increases when sulfur exists on the catalyst surface, and this also enhances NO removal activity by NH334935.Figure 6 shows the NH3-TPD profiles of three distinctive Ti02 supports with and without sulfur. ST-500 and P25S-500 containing sulfur adsorbed more NH3 on the catalyst surface than P25-500 without As shown in Figure 7, both catalysts containing sulfur exhibit higher NO removal activity than P25 without sulfur. This indicates the acid sites generated from the sulfate on the surface of Ti02 play an important role in enhancing NO removal activity.

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

255

J

I

0

100

, 200

300

400

500

600

Temperature ("C)

Figure 6

NH3 TPD profiles of Ti02 with and without sulfur 35

100

80 60 40

20

0

200

250

300

350

400

450

500

Temperature ("C)

Figure 7

Catalytic activity of NO reduction of Ti02. Reaction conditions: Space velocity = I00 000 hr - NHjINO = I . 0, NO 500 ppm, NH3 500 ppm, 0, 5% 35

',

M o O ~ - S O ~ ~ - / Tcatalyst ~ O ~ also shows higher NO removal activity than Mo03/Ti02 without sulfur. The addition of 0.5% S042- on Mo03/Ti02 catalyst significantly increased the conversion of NO from 58 to 84% at 300 "C due to the surface acidity enhanced by S042- on the surface of Ti02. Chen and Yangs1 also reported that S042-/Ti02 catalyst shows superior NO removal activity by NH3 at a reaction temperature of 300 "C. 6.2 State of the Vanadia Phase. - The structure of vanadia species on the surface of the supported vanadia catalysts greatly affects the catalytic perfor-

Catalysis

256

C Thin overlayer

B Isolated species

A Crystallite

v,o, /wt% 0

5

10

15

20

CVD

Figure 8

Models of vanadia overlayers on supports and the regions of the vanadia loading levels in the presence of the overlayers 127

mance on the NO removal activity. Thus, extensive studies have been performed on the characteristics of vanadia species existing on a variety of supports. Based on IR spectroscopy, Busca et al.121j122 showed that the monomeric vanadyl species are mainly formed on the surface of anatase type titania at low loadings of vanadia. The structure of vanadyl species on the catalyst surface was also examined by Went et al. and Wachs et al.125,126 using Raman spectroscopy. They have observed that for V2O5/TiO2 catalyst, the monomeric species formed predominantly at the low loadings of vanadia are converted to the polymeric vanadates as the content of vanadia increases. The crystallites of V205 are formed when the amount of vanadia exceeds the capacity of the surface area of Ti02 for the formation of the monolayer vanadate. have proposed that the structure of vanadia depends on Inumaru et the support and may be controlled by the method of vanadia impregnation. Figure 8 shows that vanadia on Si02predominantly exists in a crystalline form even at the low loadings of metal, and vanadia on MgO easily forms a complex compound of Mg3V205 on its surface, whereas vanadia on A1203 is fairly stable even at high loadings of vanadia. However, the chemical vapor deposition (CVD) of vanadia onto a support stabilizes the structure of vanadia species even at the high metal content of the catalyst. In addition, Went et al.'23J24have also compared the effect of support on the structure of vanadia by using Raman spectroscopy. Figure 9 shows Raman spectra of vanadia catalysts supported on Si02, Ti02 and A1203 with respect to the 1239124

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

200

400

600

Lloo

lMlOl100

200

400

600

600

loo0

1200

Wavmumkr (cm-')

Wmmnumkr (cm')

Figure 9

257

Raman spectra of vanadia catalysts supported on (A) Si02, (B) Ti02, and ( C )A1203 calcined at 773 K with respect to the contents of vanadia: (a) 9.8%, (b) 5.6%, 4.2% and 6.7% for (A), (B) and (C), respectively, (c) 1.4%, and (d) 0096 124

amount of vanadia loadings on the catalyst surface. Vanadia species on Si02 exist in a crystalline form which contradicts the formation of polyvanadate as observed by Inumaru et al. 127 They also reported that the dispersion of vanadia on the catalyst surface improves in the following order: Si02c A1203 Ti02. Eckert and Wachs128have investigated the local environments of the layer in VOx on Ti02 and A1203 using solid-state 51V NMR spectroscopy. They concluded that under ambient conditions, the low surface coverage of vanadia mainly forms the four-fold coordination of VOx on the surface of the support, while at high surface coverage it transformed to the six-coordinated species. The chemical nature of the support has also affected the state of VOx, either six- or four-coordinated species. The formation of six-coordinated VOx with respect to the support is in the following order: A1203 Ti02 (rutile) <Ti02 (P25 from Degussa) < Ti02 (anatase). These results are also confirmed by 51V NMR spectra of vanadia catalysts supported on Ti02 and A1203with respect to the amount of vanadia on the catalyst surface. Furthermore, Kozlowski et aZ.129have shown that the structure of the surface vanadates on Ti02 and A1203 differs from that of the bulk V205 crystal as confirmed by EXAFS. In addition, XANES spectra are also useful for the examination of the structural symmetry of vanadia, since XANES spectra originate from the transition of a core electron to the ground state. In particular, the features of K-edge XANES spectra for VOx significantly vary with respect to the geometry and valence of the vanadium existing on the catalyst surface.

-

-

258

Catalysis

Table 2

Mathematical models for honeycomb reactor

Model

Condition

Heat conduction in solid

Internal mass transfer in solid

Radiation

Refs.

Distributed Distributed Lumped Lumped Lumped Lumped Lumped Distributed Lumped Distributed Lumped Lumped Lumped

Steady Steady Transient Steady Transient Steady Steady Steady Steady Steady Transient Steady Steady

Yes No No No Yes Yes Yes Yes Yes Yes Yes No No

No No No No No No No No No Yes No Yes Yes

No No No No No Yes No No No No No No No

147 148 146 145 149 150 151 152 153 133 132 137 138

7

Reactor Modeling and Process Characteristics

7.1 Low Pressure-drop Reactor. - Since a honeycomb-type monolith reactor has unique advantages over a packed-bed reactor, such as low pressure drop and high geometric surface areas per unit volume, it has been widely employed in catalytic pollution control processes.28~1307131 Many studies have been conducted on the design and modeling of the honeycomb reactor for automotive applications. 132-136 However, the design and modeling of the honeycomb reactor for the SCR process has recently attracted a great deal of interest. 102~133-144Mathematical modeling of the honeycomb reactor has been done mainly for its a ~ t o m o b i l e 'and ~ ~ non-automobile applications. 146 The models developed in the literature are listed in Table 2. H e g e d ~ s lpresented ~~ the simplest model with complete external mass transfer limitations. The simplified balance equations could be solved analytically. He also showed the experimental measurements of temperature profile with respect to reactor length, but made no comparison of experimental observation with the predictions by the model. Heck et a1.146have developed one-dimensional and two-dimensional models of the honeycomb reactor for automobiles. They found that the light-off point is close to the inlet of the reactor with sinusoidal channels and that it increases with respect to the configuration of channels in the following order: triangular, square, and circular channels. Large Nusselt number and small channel size will lead to uniform heat transfer throughout the reactor and consequently will delay the light-off of the reactor. The disadvantage of the delay will be partly compensated by a better mass transfer and higher conversion in the channel after ignition. When the Lewis number is greater than 1, the wall temperature may increase to a value much higher than the adiabatic flame temperature as experimentally verified by Hegedus. 145The simulation by the one-dimensional

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

259

model with asymptotic Nusselt numbers was in good agreement with that of the two-dimensional model. Heynderickx et al. 155 modeled a honeycomb reactor for high-severity thermal processes. Silicon carbide was chosen as the material of the reactor because of its high thermal conductivity and resistance to thermal shocks. A detailed kinetic model including 1054 reactions was extensively developed. Both convective and radiative heat transfer between a wall and the fluid as well as transverse and longitudinal conductive heat exchange inside the wall were included for the model development. Heat exchange by natural conduction and convection was also considered. A good agreement between the results from simulations and experiments was observed. An overall simulated conversion of feed was found to be higher in the honeycomb reactor (99.1%) than in the other type of reactor (from 91.3 to 92.4%). For deNOx application, studies have been mainly done for an extruded honeycomb reactor with V205/Ti02catalyst. They include pore diffusion and/ or external mass transfer resistance for the development of the kinetic model. Buzabowski and Yang137reported the theoretical and experimental results for SCR of NO by NH3 in the extruded honeycomb reactor under the operating conditions of power plants. The model is a first order reaction with respect to NO and independent of the concentrations of all other species - NH3, 02,N2, H20, SO2 etc. A simple analytic solution is given by which NO conversion is expressed as an explicit function of reactor space velocity and other reaction parameters. Beekman and Hegeduslo2 developed a mathematical model to describe the reaction of NO with NH3 over the internal surfaces of a honeycomb reactor for power plants. In their study, SCR activity can be substantially improved by the configuration of pore structure. A new catalytic system developed on the basis of the restructuring of the catalyst pore network was shown to exhibit a 50% increase of NO conversion in the laboratory. Svachula et al. lS6 developed a one-dimensional model based upon Eley-Rideal mechanism assuming SCR reaction between strongly adsorbed ammonia and gaseous or weakly bonded NO on the catalyst surface. They included both external and internal transport of the reactants for the model development. The effects of the reaction variables and the feed compositions including SO2 and NO conversion were also considered on the performance of the honeycomb reactor. Tronconi and Forzatti140 compared one- and two-dimensional models of honeycomb reactors. They examined the two models for circular, square and triangular geometry as well as for first-order and Rideal kinetics. For the square geometry, practically no differences could be observed between the two models. For the other geometries, the lumped model slightly underpredicted NO conversion in circular channels, but slightly overestimated it in triangular channels. In their study, the prediction of the reactor performance by a lumped parameter model agrees well with the solution by a far more rigorous twodimensional model. In particular, the agreement between two models was excellent for the square channel of the honeycomb, which is mainly employed as the cell type of SCR honeycomb reactor.

260

Catalysis

Chae et al. 147 developed the honeycomb reactor model directly employing the kinetic parameters obtained from the kinetic study over a packed-bed flow reactor. The model could also predict the effects of catalytic wall thickness on the honeycomb reactor and the pore structure of the catalyst on the NO removal activity and NH3 slip, regardless of the types of the honeycomb, washcoated or extruded. The study also identified that the diffusion resistance in the honeycomb reactor plays a critical role in the design of the commercialscale SCR reactor despite the relatively thin catalyst layer of the reactor. Moreover, the diffusion effect was more significant for a zeolite catalyst primarily containing micropores than for a V205-based catalyst primarily containing mesopores. The lumped parameter model of honeycomb reactor developed by Chae et al. 147 can be described by equations (14)-(20): Material balances for bulk flow within a channel become

with the following initial conditions: Ch0 = C&o,oand CkH3= Cf;H,,o at x = 0

At any axial position x of the honeycomb reactor, a material balance over the catalyst layer of thickness dy yields

with boundary conditions: CNO= Ci0 and CNH~ = CgH3 at y = R A material balance at axial position x of the honeycomb reactor over the external gas film yields

d Sh = B( 1 + 0.095 L Pe)0.45 where B is a shape factor of the honeycomb channel.

(20)

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia 2.41 -

0.0

2.4

.4

.8

1.2

1.6

Rsecta D i m s i i m(IT#

(4 Figure 10

2.0

2.4

26 1

' 1

0.0

.4

.8

1.2

1.6

2.0

2.4

Re~ctaDirr~mirn

(b)

Contour map of NO conversion in the cross-sectional area at the exit of the reactor in the duct:'47 (a) without guide vane (average NO conversion: 0.83) and (b) with guide vane (average NO conversion: 0.93). Reaction conditions: Space velocity = 10 000 hr - I, NHjINO = 1 .O, Temperature = 400 C

The flow pattern and the NH3 distribution in the commercial-scale honeycomb reactor are also important for a high performance of NO removal. Good distribution of the flow by the guide vanes installed in the reactor can improve the NO removal activity by more than 10% of NO conversion as shown in Figure 10.

7.2 SCR Process Characteristics.- Generally, the deNOx process is classified into clean-gas, high-dust and low-dust systems for utility applications illustrated in Figure 11. A fourth possibility is that the catalyst may be coated on the surface of the heating element of air preheater to save space and energy consumption. The choice of the installation location of the system should be determined by the consideration of catalyst life and economics. Clean-gas application has many advantages such as: (i) relatively short period of installation time, (ii) long catalyst life and (iii) less catalyst. Despite these advantages, it has not been popular due to the considerable amount of energy required for reheating the flue gas to the temperature required for SCR reaction. Another reason may be insufficient space available between economizer and air preheater in a utility boiler. Therefore, the SCR process is commonly designed for high- and low-dust systems. A high-dust system places the SCR reactor before the electrostatic precipitator (ESP) and SOXscrubber. The advantages are that hot flue gas is available and expensive reheating of flue gas can be avoided. The disadvantage may be that the catalyst life can be reduced by fly ash, SO2 and metals contained in the flue gas as well as NOx. In a low-dust system with the system placed after ESP, it has been observed that the significantly reduced level of fly ash and catalyst poisons leads to the high

262

Catalysis

I

(a)

B

370

* APH

-

150

-

1 50

ESP

-

FGD -

I (b)

B

370

SCR

-

370

APH

150 '

-

55

370

-

SCR

290

,100

ESP

150

-

FGD

1 00

NY

,100

(c)

370

Hot ESP

370

SCR

370

e

APH

FGD

150

Figure 11

100

Arrangement of SCR processes: (a) Clean-gas, (b) High-dust, and (c) Lowdust SCR systems. FGD (Flue-Gas Desulfurization), B (Boiler), APH (Air Pre-Heater), ESP (Electrostatic Precipitator), GGH (Gas-Gas Heatexchanger), and H (Heater). The numbers in the Figure denote the temperature ("C ) of the feed gas stream

geometric surface area of the reactor. On the other hand, reheating of the entire flue gas is required to the temperature where the catalyst is sufficiently active since the temperature of the gas stream has been cooled-down by ESP. Note that the energy requirement for a low-dust system is less than that for a clean-gas system. Since the reaction temperature of the SCR process is also important for catalyst activity and life, the location of the SCR system should be based upon the design of a commercial SCR process. However, the development of strong sulfur and dust-tolerant catalyst may always relieve the constraints of reactor installation for a commercial SCR process. 8

Alternative Technology to NH3-SCR

The SCR system with NH3 as a reductant still has disadvantages such as NH3slip by unreacted or excess NH3 and the high cost of facilities and operation. There is a strong demand for the development of a new type of catalytic system for the removal of NO from both stationary and mobile sources. Although the catalytic decomposition of NO is the ultimate goal for the development of NOx removal technology, the alternative reductants may be a good attempt to achieve the final target. Hydrocarbons could be used as an attractive and alternative reductant for NH3 in SCR technology. The selective catalytic reduction of NO by hydrocarbons has been already investigated for

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

263

metal oxides and supported-noble metal catalysts.158- 162 None of the catalytic systems proposed so far, however, has been attractive for commercial applications, since they were mainly developed in an oxygen-free stream andor excess concentrations of NO and hydrocarbons to avoid the immediate oxidation of hydrocarbons in the feed gas stream. In the early 1990s, the selective catalytic reduction of NO by hydrocarbon over zeolite catalyst was reported by Iwamoto et al. 163 and Held et al. 164 They observed the high deNOx efficiency of copper ion-exchanged ZSM-5 catalyst under dilute conditions of NO and hydrocarbon in the presence of oxygen similar to a practical exhaust stream containing NOx. A number of catalysts including transition metal ionexchanged zeolite^,'^^-'^^ H-form zeolite^,'^^-^'^ supported noble metals,174 supported metals,168y174-177 metal oxides,168-177*178 solid acid168'174-179 and perovskites16*were extensively investigated to develop a new concept of SCR technology employing hydrocarbons as a reductant. NO removal activity totally depends on the types of catalyst and reductant, as well as the reactor operating conditions. The metal ion-exchanged zeolites including ZSM-5, mordenite, ferrierite, L- and Y-zeolites reveal much higher deNOx efficiency than any other type of catalyst. ZSM-5 and mordenite especially exhibit the best removal activity of NO, regardless of the catalyst and operating variables as discussed. This suggests that a unique catalytic system can be developed to reduce NO in a flue gas containing O2 by the new concept of SCR technology employing hydrocarbons as a reductant. A variety of hydrocarbon^,'^^-^^^ CO and H21809181 as well as alcoh o l ~ ~have ~ been ~ *examined ~ ~ ~as -a reductant ~ ~ ~ for the reduction of NO. Iwamoto and Hamada181 have classified the reductant into two categories: selective (C2H4, C3&, C3H8 and C4H8) and non-selective (H2, c o , CH4 and C2H6). Truex et al. 185 also reported a similar classification for the reductant. It should be noted that the categories might only be valid for Cu-ZSM-5 catalyst. As an example, CH4and C2H6 were non-selective for the reduction of NO in the presence of O2 for Cu-ZSM-5 and A1203catalysts. However, these are selective reductants for Co-ZSM-5,1g6 Ga-ZSM-5187,188and H-type zeolite cata1 y ~ t as~well ~ as ~ for ~ Pt/A1203,I9l 3 ~ ~ ~ Li-promoted Mg0182and Pd/ZSM-5192 catalysts. Alkenes have also been employed as reductants for the selective catalytic reduction of NO due to their rapid activation during the course of reaction because of their double bond. A conclusive reaction mechanism for the removal of NO by hydrocarbons has not been reported in the literature, even though many proposals have been made.193*194 A better understanding of the reaction mechanism might not be an urgent topic due to the weak tolerance of the catalyst for water and SO2 commonly contained in the flue gas as well as NO. This causes serious activity loss of NO reduction by hydrocarbons in a wet stream. It may be that a reaction mechanism is also involved for the deactivation of NO conversion by H 2 0 and S02. Not only is the high deNOx activity of the catalyst important for its commercial application, but its tolerance to the composition of the flue gas containing NOx as well as H20 and SO2 must be carefully considered from the viewpoint of catalyst life. Based upon the previous reports, SO2 in the feed gas stream exhibited a moderate

264

Catalysis

poisoning effect,195-199 while water vapor showed serious detrimental effect on NO removal activity even in trace The poisoning by SO2 for SCR catalyst with hydrocarbon is further complicated by the deactivation behavior with water. For Co-ZSM-5 catalyst, Li and Armor217 reported that the cause of the deactivation by SO2 was simply due to the adsorption of SO2 on Co2+ sites. Kim et aL218 have extensively studied the deactivation of mordenite-type zeolite catalyst in the presence of S02. They observed that the sulfur species, which are deactivation agents, existed in the form of sulfate as confirmed by XPS and Raman. The sulfate species deposited on the deactivated catalyst did not significantly alter the chemical environment of the metal ions on the surface of zeolite confirmed by Cu K-edge absorption spectra. The water tolerance of SCR catalyst by hydrocarbons is the primary subject to be resolved for the commercial application of the technology. The causes of the deactivation are generally known as follows: (i) the inhibition of NO adsorption by the competitive adsorption of NO and water on adsorption sites and (ii) the formation of copper oxides from copper ions on the catalyst surface. The former can be easily anticipated by the reversibility of the catalytic activities of Cu-ZSM-5 (157% of Cu ions),200Co-ZSM-5(70% of Co ions)201 and Cu-Mordenite (47% of Cu ions)218in a wet stream. On the other hand, the latter is still controversial. The hydrothermal sintering of the catalyst is the main cause of the formation of metal oxides on the catalyst surface. Chung et al.214 strongly insisted that the hydrothermal stability of zeolite catalyst became stronger as the Si/Al ratio of the catalyst increased. Overcoming the catalyst deactivation by water is the key for the successful development of SCR technology by hydrocarbon replacing the conventional NH3 based technology. 9

Conclusions and Future Prospects

SCR by NH3 is currently the best available control technology for reducing NOx emissions from stationary sources. The interest in the subject is reflected by the large number of studies including research papers and patents recently issued throughout the world. The focus of SCR studies is the development of ‘better’ catalysts. Among them, vanadium oxide supported on Ti02 is generally regarded as one of the best and widely commercialized catalysts for the removal of NOx due to its high activity at relatively low reaction temperatures (<400°C) and its high resistance to poisoning by S02. However, since approximately 50% of the ‘levelized cost’, i.e. the sum of operating and capital costs of SCR, is catalyst-related, there is still a strong incentive for the improvement of the performance of an SCR catalyst. A great deal of work is still under way for the development of a ‘better’ SCR catalyst from the viewpoint of catalytic performance and cost, since the technology suffers from several detrimental effects such as NH3 slip to atmosphere, the temperature dependence of NO conversion and catalyst deactivation by sulfur compounds.

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

265

Recently, there have been many efforts to replace NH3 with other reductants including urea and hydrocarbons. In particular, hydrocarbons could be considered as an attractive and alternative reductant to substitute NH3 in SCR technology. The variety of zeolite-based catalysts such as ZSM-5, mordenite and ferrierite etc. reveals high NO removal activity in the absence of water vapor. However, a significant level of water vapor in actual exhaust streams may block and inhibit effective NOx reduction. More efforts are needed to develop a reasonably active SCR catalyst with hydrocarbon as reductant. Ultimately, the catalytic decomposition of NOx into its elements, N2 and 02, may be the most desirable technology for NOx removal. Again, the development of the catalyst is the greatest challenge for scientists and engineers in the area of environmental catalysis. Although our understanding of the surface characteristics, the reactivity, and the deactivation of SCR catalyst along with the reactor design through the present review has increased significantly, the intriguing results summarized here reveal that much is yet to be learned for the development of a ‘better’ SCR catalyst. Indications are strong that a better knowledge of SCR catalysts combined with a more complex understanding of the surface chemistry involved in the catalyst deactivation by SO2 will lead to improved SCR systems in the future.

10 1 2 3 4 5 6 7 8 9 10 11 12 13 14

15 16 17

References H. Bosch and F. Janssen, Catal. Today, 1988,2(4), 369. M. Markvart and V. Pour, J. Catal., 1967,7, 279. G.L. Bauerle, S.C. Wu and K. Nobe, Ind Eng. Chern. Prod. Res. Dev., 1975, 14, 123. H. Meier and G. Gut, Chem. Eng. Sci., 1978,33, 123. M. Inomata, A. Miyamoto and Y. Murakami, J. Catal., 1980,62, 140. A. Kato, M. Matsuda, F. Nakajima, M. Imanari and Y. Watanabe, J. Phys. Chem., 1981,85,1710. A. Kato, M. Matsuda, F. Nakajima, H. Kuroda and N. Tsuneo, J. Phys. Chem., 1981,85,4099. J. Blanco, G. de la Banda, P. Avila and F. Melo, J. Phys. Chem., 1986,90,4789. I.S. Nam, J.W. Eldridge and J.R.Kittrell, Ind. Eng. Chem. Prod. Rex Dev., 1986, 25, 192. H.G. Lintz and T. Turek, Appl. Catal. B: Gen., 1992,85, 13. W.C. Wong and K. Nobe, Ind. Eng. Chem. Prod. Res. Dev., 1986,25,179. T. Seiyama, T. Arakawa, T. Matsuda, N. Yamazoe and Y. Takita, Chem. Lett., 1975,781. W.B. Williamson and J.H. Lunsford, J. Phys. Chem., 1976,80,2664. T. Seiyama, T. Arakawa, T. Matsuda, Y. Takita and N. Yamazoe, J. Catal., 1977,48,1. T. Iizuka and J.H. Lunsford, J. Am. Chem. SOC.,1978,100,6106. J.R.Kiovsky, P.B. Kotadia and C.T. Lim, Ind. Eng. Chem. Prod. Res. Dev., 1980, 19,218. I.S. Nam, J.W. Eldridge and J.R. Kittrell, Stud Surf: Sci. Catal., 1988, 38, 589.

266

18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 33 34

35 36

37 38

39 40 41 42 43 44 45 46 47 48 49 50

Catalysis

F.G. Medros, J.W. Eldridge and J.R. Kittrell, Ind Eng. Chem. Res., 1989, 28, 1171. M.D. Amiridis, F. Puglisi, J.A. Dumesic, W.S. Millman and N.Y. Topsoe, J. Catal., 1993, 142, 572. K. Otto and M. Shelef, J. Phys. Chem., 1972,76,37. M. Mizumoto, N. Yamazoe and T. Seiyama, J. Catal., 1978,55, 119. M. Kiyomiya and M. Kawai, Atrnos. Environ., 1979, 13, 599. M.D. Oates and J.H. Lunsford, J. Mol. Catal., 1980,9,91. E.Y. Choi, I.S. Nam, Y.G. Kim, J.S. Chung, J.S. Lee and M. Nomura, J. Mol. Catal., 1991,69,247. S.W. Ham, I.S. Nam and Y.G. Kim, Korean J. Chem. Eng., 2000,17,318. F.P. Boer, L.L. Hegedus, T.R. Gouker and K.P. Zak, CHEMTECH, 1990, May, 312. H. Yoshida, K. Takahashi, Y. Sekiya, S. Morikawa and S. Kurita, Proc. 8th Int. Cong. on Catal., 1984, Vol. 3, p. 649. S. Irandoust and B. Anderson, Catal. Rev. Sci. Eng., 1988,30,341. A. Cybulski and J.A. Moulijon, Catal. Rev. Sci. Eng., 1994,36, 179. V.I. Parvulescu, P. Grange and B. Delmon, Catal. Today, 1998,46,233. L.L. Sloss, A.K. Hjalmarsson, H.N. Soud, L.M. Campbell, D.K. Stone, G.S. Shareef, T. Emmel, M. Maidobi, C.D. Livengood and J. Markussen, Nitrogen Oxides Control Technology Fact Book, Noyes Data Corp., NJ, USA, 1992. I.S. Nam, J. Catal., 1989, 119,269. I. Nova, L. Lietti, L. Casagrande, L. DallAcqua, E. Giamello and P. Forzatti, Appl. Catal. B: Environ., 1998, 17, 245. S. Okazaki, M. Kumasaka, J. Yoshida, K. Kosaka and K. Tanabe, Ind. Eng. Chem. Prod. Res. Dev., 1981,20, 301. S.T. Choo, Y.G. Lee, I.S. Nam, S.W. Ham and J.B. Lee, Appl. Catal. A: Gen., 2000,200,177. S.T. Choo, J.Y. Koh, I.S. Nam, S.W. Ham and J.B. Lee, AIChE Meeting at Los Angeles, USA, Nov. 13-17,2000. S.T. Choo, J.Y. Koh, I.S. Nam, S.W. Ham, J.B. Lee and K.Y. Kim, Hwahak Konghak (Korean), 2000,38,597. Y. Kosaki, A. Miyamoto and Y. Murkami, Bull. Chem. SOC. Jpn., 1979, 52, 617. M. Imanari and Y. Watanabe, T. Seiyama and K. Tanabe (Eds), Proc. 7th Intern. Congr. Catal., Tokyo, 30 June-4 July 1980, Tokyo, 1980. F. Janssen, F.M. van den Kerkhof, H. Bosch and J.R.H. Ross, (a) J. Phys. Chem., 1987,1,5921;(b) J. Phys. Chem., 1987,91,6683. R.A. Rajadhyaksha and H. Knozinger, Appl. Catal., 1989,51,81. M.D. Amiridis, R.V. Duevel and I.E. Wachs, Appl. Catal. B: Environ., 1999, 20, 111. R.Y. Weng and J.F. Lee, Appl. Catal., 1993,105,41. A.J. van Hengstum, J. Pranger, J.G. van O m e n and P.J. Gellings, Appl. Catal., 1984, 11, 317. J. Zhu and S.L.T. Andersson, J. Chem. SOC.Faraday Trans. I., 1989,85,3629. G. Deo and I.E. Wachs, J. Catal., 1994,146,323. G.C. Bond and S.F. Tahir, Catal. Today, 1991,10,393. M.A. Vuurman, A.M. Hirt and I.E. Wachs, J. Phys. Chem., 1991,95,9928. G . Ramis, G. Busca and P. Forzatti, Appl. Catal. B: Environ., 1992,1, L1. G. Deo and I.E. Wachs, J. Catal., 1994,146,335.

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

71 72 73 74 75 76 77 78 79

80 81 82

83 84

267

J.P. Chen and R.T. Yang, Appl. Catal. A: Gen., 1992,80,135. G. Deo and I.E. Wachs, J. Phys. Chem., 1994,95,5889. G. Deo and I.E. Wachs, J. Catal., 1991,129,307. I.S. Nam, 1st Korea-Japan Joint Symposium on Catalysis, Seoul, Korea, 1987. C. Hwang, I.S. Nam and Y.G. Kim, Hwahak Konghak (Korean), 1990,28,104. U.C. Hwang, S.W. Ham, I.S. Nam and Y.G. Kim, Catal. Sci. Technol., 1991, 1, 165. S.T. Choo, D.J. Koh, I.S. Nam and Y.G. Kim, Catal. Today, 1997,38, 181. E.Y. Choi, I.S. Nam and Y.G. Kim, J. Catal., 1996,161, 597. P. Wanthoz, M. Ruwet, T. Machej and P. Grange, Appl. Catal., 1991,69, 149. H. Matralis, S. Fiasse, R. Castillo, Ph. Bastians, M. Ruwet, P. Grange and B. Delmon, Catal. Today, 1993,17, 141. M. Schneider, M. Maciejewski, S. Tschudin, A. Wokaun and A. Baiker, J. Catal., 1994, 149, 326. U. Scharf, M. Schneider, A. Baiker and A. Wokaun, J. Catal., 1994,149, 344. M. Schneider and A. Baiker, J. Mater. Chem., 1992,2, 587. P. Ciambelli, G. Bagnasco, L. Lisi, M. Turco, G. Chiarello, M. Musci, M. Notaro, D. Robba and P. Ghetti, Appl. Catal. B: Environ., 1992, 1,61. T. Shikada, K. Fujimoto, T. Kunugi, H. Tominaga, S. Kaneko and Y. Kubo, Ind. Eng. Chem. Prod. Res. Dev., 198 1,20,91. B.M. Reddy, I. Ganesh, and E.P. Reddy, J. Phys. Chem. B, 1997,101,1769. G. Clarebout, M. Ruwet, H. Matralis and P. Grange, Appl. Catal., 1991,76, L9. H. Matralis, S. Fiasse, R. Castillo, Ph. Bastians, M. Ruwet, P. Grange and B. Delmon, Catal. Today, 1993,17, 141. H.J. Chae, I.S. Nam, Y.G. Kim, H.S. Yang, H.C. Choi, and S.L. Song, Stud. Surf: Sci. Catal., 1999,125, 595. H.J. Chae, I.S. Nam, H.S. Yang, S.L. Song and I.D. Hur, J. Chem. Eng. Jpn, 2001,34,148. F. Figueras, Catal. Rev. Sci. Eng., 1988,30,457. S . Cheng, Catal. Today, 1999,49,303. L.S. Cheng, R.T. Yang, and N. Chen, J. Catal., 1996,164,70. R.Q. Long, and R.T. Yang, Appl. Catal. B: Environ., 2000,24, 13. H.J. Chae, I.S. Nam, H.S. Yang, S.L. Song and I.D. Hur, Hwahak Konghak (Korean), 2000,38,783. H.J. Chae, PhD Thesis, POSTECH, Korea, 2001. F. Nakajima, Catal. Today, 1991,10, 1 . I.S. Nam, J.W. Eldridge and J.R. Kittrell, Joint Spring Symposium of the New York-New England Catalysis Societies, New Haven, Connecticut, 1983. B.W. Soh, J.B. Lee, I.S. Nam and Y.G. Kim, Stud. Surj Sci. Catal., 1999, 126, 389. T. Shikada, K. Fujimoto, T. Kunugi and H. Tominaga, J. Chem. Tech. Biotech., 1983,33A, 446. S . Morikawa, K. Takahashi, J. Mogi and S . Kurita, Bull. Chem. SOC.Jpn., 1982, 55, 2254. M. Imanari, Y. Yatanabe, S. Matsuda and F. Nakajima, Proc. 7th Int. Cong. on Catalysis, 1981, p. 84 1 . H. Fukuzawa and Y. Ishihara, Report of Central Res. Inst. Elec. Power Ind., July, 1979. 0 .Saur, M. Bensitel, A.B. Mohammed Saad, J.C. Lavalley, C.P. Tripp and B.A. Morrow, J. Catal., 1986,99, 104.

268

Catalysis

85

S. Matsuda, T. Kamo, A. Kato and F. Nakajima, Ind. Eng. Chem. Prod. Res. Dev., 1982,21,48. K. Kusakabe, M. Kashima, S. Morooka and Y. Kato, Fuel, 1988,67,714. K. Kusakabe, H. Kawamura, H.J. Kim and S. Morooka, Fuel, 1990,69,917. S.W. Ham, H. Choi, I.S. Nam and Y.G. Kim, Stud. Surf: Sci. Catal., 1991, 68, 573. S.W. Ham, H. Choi, I.S. Nam and Y.G. Kim, Catal. Today, 1992,11,611. S.W. Ham, H. Choi, I.S. Nam and Y.G. Kim, Stud. Surf: Sci. Catal., 1994, 88, 441. S.W. Ham, H. Choi, I.S. Nam and Y.G. Kim, Ind Eng. Chem. Res., 1995, 34, 1616. J. Tsai, P.K. Agrawal, J.M. Foley, J.R. Katzer and W.H. Manogue, J. Catal., 1980,61, 192. J. Tsai, P.K. Agrawal, D.R. Sullivan, J.R. Katzer and W.H. Manogue, J. Catal., 1980,61,204. S . Okazaki, M. Kumasaka, J. Yoshida and K. Kosaka, Ind. Eng. Chem. Prod. Res. Dev., 1981,20,231. J.P. Chen and R.T. Yang, J. Catal., 1990,125,411. J.P. Chen and R.T. Yang, J. Catal., 1993,139,277. S . Morikawa, H. Yoshida, K. Takahashi and S. Kurita, Chem. Lett., 1981,251. S . Morikawa, K. Takahashi, H. Yoshida and S. Kurita, Proc. 8th Int. Cong. On Catalysis, 1984, Vol. 3, p. 661. N. Sazonova, L. Tsykoza, A. Simakov, G. Barannik and Z. Ismagilov, React. Kinet. Catal. Lett., 1994,52, 101. J.P. Dunn, P.R. Koppula, H.G. Stenger and I.E. Wachs, Appl. Catal. B: Environ., 1998,19, 103. S.T. Choo, PhD Thesis, POSTECH, Korea, 2001. J.W. Beeckman and L.L. Hegedus, Ind. Eng. Chem. Res., 1991,30,969. J. Svachula, L.J. Alemany, N. Ferlazzo, P. Forzatti, E. Tronconi and F. Bregani, Ind. Eng. Chem. Res., 1993,32, 826. M. Takagi, T. Kawai, M. Soma, T. Onishi and K. Tamaru, J. Phys. Chem., 1976, 80,430. M. Takagi, T. Kawai, M. Soma, T. Onishi and K. Tamaru, J. Catal., 1977, 50, 441. M. Inomata, K. Mori, A. Miyamoto, T. Ui and Y. Murakami, J. Phys. Chem., 1983,87, 754. S.C. Wood, Chem. Eng. Prog., 1994, Jan., 32. A. Miyamoto, Y. Yamazaki, M. Inomata and Y. Murakami, J. Phys. Chem., 198 1,85,2366. M. Inomata, A. Miyamoto and Y. Murakami, J. Phys. Chem., 1983,87,761. H. Bosch, F. Janssen, F. van den Kerkhof, J. Oldenziel, J.G. van O m e n and R.H. Ross, Appl. Catal,, 1986,25,239. G. Ramis, G. Busca, F. Bregani and P. Forzatti, Appl. Catal., 1992,64,259. G. Ramis, G. Busca, V. Lorenzelli and P. Forzatti, Appl. Catal., 64, 1992,243. G. Ramis, G. Busca and P. Forzatti, Appl. Catal. B : Environ., 1992,1, L9. K. Otto, M. Shelef and T. Kummer, J. Phys. Chem., 1970,74,2690. K. Otto, M. Shelef and T. Kummer, J. Phys. Chem., 1971,75,875. T. Katona, L. Guczi and G.A. Somorjai, J. Catal., 1991,132,440. T. Katona, L. Guczi and G.A. Somorjai, J. Catal., 1992,135,434. S.W. Ham, H. Choi, I.S. Nam and Y.G. Kim, Catal. Lett., 1996,42, 35.

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105

106 107 108 109 110 111 112 113 114 115 116 117 118

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144

145 146 147 148 149 150 151 152 153 154 155

269

D. Barthomeuf, Catalysis by Zeolites, ed. B. Imelik et al., Elsevier, Amsterdam, 1980, p. 55. K. Tanabe, Solid Acids and Bases, Kodansha, Tokyo, Academic Press, New York, 1970. C. Cristiani, P. Forzatti and G. Busca, J. Catal., 1989, 116,586. G. Busca and G. Mater. Chem. Phys., 1988,19, 157. G.T. Went, L.J. Leu and A.T. Bell, J. Catal., 1992,134,479. G.T. Went, S.T. Oyama and A.T. Bell, J. Phys. Chem., 1990,94,4240. T. Machej, J. Haber, A.M. Turek and I.E. Wachs, Appl. Catal., 1991,70, 115 . M.D. Amiridis, I.E. Wachs, G. Deo, J.M. Jehng and D.S. Kim, J. Catal., 1996, 161,247. K. Inumaru, M. Misono and T. Okuhara, Appl. Catal. A: Gen., 1997,149,133. H. Eckert and I.E. Wachs, J. Phys. Chem., 1989,93,6796. R. Kozlowski, R.F. Perrifer and J.M. Thomas, J. Phys. Chem., 1983,87,5176. R. Prasad, L.A. Kennedy and E. Ruckenstein, Catal. Rev.-Sci. Eng., 1984,26, 1 . L.D. Pfefferle and W.C. Pfefferle, Catal. Rev. -Sci. Eng., 1988,29,219. L.C. Young and B.A. Finlayson, AIChE J., 1976,22,331. K. Zygourakis and R. Aris, Chem. Eng. Sci., 1983,38,733. L.C. Young and B.A. Finlayson, AIChE J., 1979,25,192. S.H. Oh and J.C. Cavendish, Ind. Eng. Chem. Prod. Res. Dev., 1983,22, 509. S.H. Oh, E.J. Bissett and P.A. Battiston, Ind. Eng. Chem. Res., 1983,32, 1560. M.A. Buzanowski and R.T. Yang, Ind. Eng. Chem. Res., 1990,29,2074. E. Tronconi, P. Forzatti, J.R. Gomez Martin and S. Malloggi, Chem. Eng. Sci., 1992,47,2401. H. Choi, S.W. Ham, I.S. Nam and Y.G. Kim, Ind. Eng. Chem. Res., 1996, 35, 106. E. Tronconi and P. Forzatti, AIChE J , 1992,38,201. E. Tronconi, A. Beretta, A.S. Elmi, P. Forzatti, S. Malloggi and A. Baldacci, Chem. Eng. Sci.,1994,49,4277. E. Tronconi, A. Cavanna and P. Forzatti, Ind Eng. Chem. Res., 1998,37,2341. E. Tronconi, L. Lietti, P. Forzatti and S. Malloggi, Chem. Eng. Sci., 1996, 51, 2965. A. Beretta, C. Orsenigo, N. Ferlazzo, E. Tronconi and P. Forzatti, Ind. Eng. Chem. Res., 1998,37,2623. L.L. Hegedus, AIChE J., 1975,21,849. R.H. Heck, J. Wei and J.R. Katzer, AIChE J., 1976,22,477. H.J. Chae, S.T. Choo, H. Choi, I.S. Nam, H.S. Yang and S.R. Song, Ind. Eng. Chem. Res., 2000,39, 1159. B. Bernauer, A. Simecek and J. Vosolsobe, Collect. Czech. Chem. Commun., 1982,47,2087. B.K. Harrison and W.R. Ernst, Combust. Sci. Technol., 1978,19, 31. S.H. Oh and J.C. Cavendish, Ind. Eng. Chem. Prod. Res. Dev., 1982,21,29. S.M. Senkan, L.B. Evans and J.B. Howard, Ind. Eng. Chem. Prod. Res. Dev., 1979,18, 125. J. Sinkule and V. Hlavacek, Chem. Eng. Sci., 1978,33, 839. H.R. Stock and A. Lowe, Chem. Eng. Sci., 1983,38,839. J. Votruba, 0. Mikus, K. Nguen, V. Hlavacek and J. Skrivanek, Chem. Eng. Sci., 1978,30,201. G.J. Heynderickx, G.F. Froment, P.S. Broutin, C.R. Busson and Weill, AIChE J., 1991,37, 1354.

270

Catalysis

156 J. Svachula, L.J. Alemany, N. Ferlazzo, P. Forzatti, E. Tronconi and F. Bregani, Ind. Eng. Chem. Res., 1993,32, 1053. 157 R.J. Ayen and M.S. Peters, Ind. Eng. Chem. Proc. Des. Dev., 1962,1, 205. 158 A. Sotoodehnia-Korrahi and K. Nobe, Ind. Eng. Chem. Proc. Des. Dev.,1970, 9, 326. 159 J.W. Ault and R.J. Ayen, AIChE J, 1971, 17,265. 160 Y.-H. Hu and J.W. Hightower, ACS Div. Petro. Chem., 1976,21,841. 161 Y. Murakami, K. Hayashi, K. Yasuda, T. Ito, T. Minami and A. Miyamoto, Nippon Kagaku Kaishi, 1977, 173. 162 J.R. Hardee and J.W. Hightower, J. Catal., 1984,86, 137. 163 M. Iwamoto, Proceedings of Meeting on Catalytic Technology for Removal of Nitrogen Monoxide, Tokyo, Japan, 1990, p. 17. 164 W. Held, A. Konig, T. Richter and L. Puppe, SAEpaper No. 900496,1990. 165 M. Iwamoto, H. Yahiro, S. Shundo, Y. Yu-u and N. Mizuno, Appl. Catal., 1991, 69, L15. 166 S. Sato, Y. Yu-u, H. Yahiro, N. Mizuno and M. Iwamoto, Appl. Catal., 1991,70, L1. 167 Y. Teraoka, H. Ogawa, H. Furukawa and S. Kagawa, Catal. Lett., 1992,12, 361. 168 S . Sato, H. Hirabayashi, H. Yahiro, N. Mizuno and M. Iwamoto, Catal. Lett., 1992,12, 193. 169 M. Kishida, T. Tachi, H. Yamashita and H. Miyadera, Shokubai, 1992,34, 148. 170 M.H. Kim, I S . Nam and Y.G. Kim, 7th Japan-China-USA Symposium on Catalysis, Tokyo, Japan, July 26-28, 1995. 171 H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and M. Tabata, Appl. Catal., 1990, 64, L1. 172 H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and M. Tabata, Appl. Catal., 1991, 70, L15. 173 M.H. Kim, I.S. Nam, Y.G. Kim, J.K. Lee and D.K. Park, Hwahak Konghak (Korean), 1994,32,402. 174. H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and M. Tabata, Appl. Catal., 1991, 75, L1. 175 Y. Torikai, H. Yahiro, N. Mizuno and M. Iwamoto, Catal. Lett., 1991,9,91. 176 H. Hosose, H. Yahiro, N. Mizuno and M. Iwamoto, Chem. Lett., 1991,1859. 177 Y. Kintaichi, H. Hamada, M. Tabata, M. Sasaki and T. Ito, Catal. Lett., 1990,6, 239. 178 H. Hamada, Catal. Today, 1994,22,21. 179 H. Hamada, T. Yoshinari, M. Tabata, Y. Kintaichi, M. Sasaki and T. Ito, Shokubai, 1992,34,147. 180 Y. Yu-u, Y. Torikai, S. Sato, H. Hosose, H. Yahiro, N. Mizuno and M. Iwamoto, Shokubai, 1991,33,61. 181 M. Iwamoto and H. Hamada, Catal. Today, 1991,10,57. 182 H. Hamada, T. Yoshinari, M. Tabata, M. Sasaki, and T. Ito, Shokubai, 1992,33, 59. 183 M.H. Kim, I.S. Nam and Y.G. Kim, Catal. Today, 1998,44,57. 184 H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and T. Yoshinari, Appl. Catal. A: Gen., 1992,88, L1. 185 T.J. Truex, R.A. Searles and D.C. Sun, Platinum Metals Rev., 1992,36,2. 186 Y. Li and J.N. Armor, Appl. Catal. B: Environ., 1992,1, L31. 187 K. Yogo and E. Kikuchi, Intern. Forum on Environ. Catal. '93, Tokyo, Japan, Feb 4-5,1993.

7: Selective Catalytic Reduction of Nitrogen Oxides by Ammonia

188 189 190 191 192 193 194 195 196

197 198 199 200 201 202 203 204 205 206 207 208 209 210 21 1 212 213 214 215 216 21 7 218

27 1

T. Tabata, M. Kokitsu and 0.Okada, Catal. Lett., 1994,25, 393. K. Yogo, M. Ihara, M. Umeno, I. Terasaki, H. Watanabe and E. Kikuchi, Shokubai, 1993,35,126. K. Yogo, M. Umeno, H. Watanabe and E. Kikuchi, Catal. Lett., 1993,19, 131. M.C. Demicheli, L.C. Hoang, J.C. Menezo and J. Barbier, Appl. Catal. A: Gen., 1993,97, L l l . C.J. Loughran and D.E. Resasco, Appl. Catal. B: Environ., 1995,5, 351. M.H. Kim, I.S. Nam and Y.G. Kim, Korean J. Chem. Eng., 1999,16, 139. M.H. Kim, I.S. Nam and Y.G. Kim, Chem. Commun., 1998,1771. M. Iwamoto, H. Yahiro, S. Shundo, Y. Yu-u and N. Mizuno, Appl. Catal., 1991, 69, L15. F. Suganuma, M. Tabata, K. Miyamoto, M. Kawatsuki, T. Yoshinari, H. Tsuchida, H. Hamada, Y. Kintaichi, M. Sasaki and T. Ito, Shokubai, 1993, 35, 114. G. Mabilon and D. Durand, Catal. Today, 1993,17,285. T. Tabata, M. Kokitsu, 0. Okada, T. Nakayama, T. Yasumatsu and H. Sakane, Stud. Surf:Sci. Catal., 1994,88,409. M.H. Kim, I.S. Nam and Y.G. Kim, Stud. Surf:Sci. Catal., 1997, 111,213. M. Iwamoto and N. Mizuno, J. Auto. Eng., 1993,207,23. Y. Li, P.J.Battavio and J.N. Armor, J. Catal., 1993,142, 561. Miyadera, Appl. Catal. B: Environ., 1993,2, 199. R. Gopalakrishnan, P.R. Stafford, J.E. Davidson, W.C. Hecker and C.H. Bartholomew, Appl. Catal. B: Environ., 1993,2, 165. Y. Li and J.N. Armor, Appl. Catal. B: Environ., 1993,2,239. Y. Li and J.N. Armor, Appl. Catal. B: Environ., 1993,3, L1. Y. Li and J.N. Armor, J. Catal., 1994, 145, 1. K.A. Bethke, D. Alt and M.C. Kung, Catal. Lett., 1994,25, 37. K.C.C. Kharas, H.J. Robota and D.J. Liu, Appl. Catal. B: Environ., 1993,2,225. M.H. Kim, I.S. Nam and Y.G. Kim, 1st World Congress on Environmental Catalysis - For a Better World and Life, Pisa, Italy, May 1-5, 1995. M.H. Kim, I.S. NamandY.G. Kim, Stud Surf:Sci. Catal., 1997,105, 1493. M.H. Kim, I.S. NamandY.G. Kim, Appl. Catal. B: Environ., 1997,12, 125. S.H. Oh, S.Y. Chung, M.H. Kim, I.S. Nam and Y.G. Kim, 12th International Zeolite Conference, Baltimore, July 5- 10, 1998. S.Y. Chung, M.H. Kim, I.S. Nam and Y.G. Kim, 2nd World Congress on Environmental Catalysis, Miami, Nov. 15-20, 1998. S.Y. Chung, S.H. Oh, M.H. Kim, I.S. Nam and Y.G. Kim, Catal. Today, 1999, 54,521. S.Y. Chung, B.S. Kim, S.B. Hong, I.S. Nam and Y.G. Kim, Stud. Sut$ Sci. Catal., 2000,130, 15 1 1. M.H. Kim, I.S. Nam and Y.G. Kim, Appl. Catal. B: Environ., 1995,6,297. Y. Li and J.N. Armor, Appl. Catal. B: Environ., 1995,5, L257. M.H. Kim, I.S. NamandY.G. Kim, J. Catal., 1998, 179,350.

8 Recent Advances in Selective Conversion of Polycyclic Hydrocarbons into Specialty Chemicals over Zeolites BY CHUNSHAN SONG

1

Introduction

Heterogeneous catalysis plays an important role in hydrocarbon processing, chemicals synthesis, and environmental protection. According to a recent study conducted jointly by several professional societies and industrial associations summarized in the ‘Technology Vision 2020’ report, catalysis is used in making over 60% of the products and accounts for 90% of the manufacturing process in the US chemical industry that generates 7,000 different products. 1*2 Governmental studies also indicate that catalytic technologies contributed some 20-30% to gross domestic product in both the UK and the USA.3 This includes not only chemical industry but also other manufacturing sectors such as the petroleum refining industry for making clean transportation fuels, and automobile industry where catalysts are used for control of emissions from exhaust gases. Zeolite catalysis occupies a unique position in manufacturing industries for environmentally benign chemical processing, because it can provide specific, desired pathways leading to desired products in hydrocarbon processing and chemicals synthesis. Shape-selective zeolite catalysis was first reported in 1960 by Weisz and co-workers using zeolites A and X.4 Venuto and co-worker$ first reported their work on selective alkylation of benzene and alkylbenzenes using crystalline aluminosilicates, and subsequently they published a seminal review in 1968 on zeolite catalysis in synthetic organic chemistry.6 Chen and Garwood7 published the first review on industrial applications of shapeselective catalysis which includes many original and important contributions by researchers at Mobil, including the authors, since the 1960s. Since these pioneering studies, shape-selective catalysis has made great strides, and has led to the development of many industrial processes. The remarkable progress in the fundamental chemistry8-l 7 and industrial a p p l i ~ a t i o n s ~ - ~of J ~zeolite J~J~ catalysis in the past decades has been well documented. Faujasites and MFItype metallosilicates have been widely used as catalysts for petroleum processing and petrochemicals production in commercial manufacturing p l a n t ~ . ~ J ~ * l ~ Selective conversion of acyclic and monocyclic compounds has been studied Catalysis, Volume 16 0The Royal Society of Chemistry, 2002

272

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

273

extensively using various zeolites, as summarized in many review^.^-'^ However, until the late 1980s/early 1990s, little attention had been paid to selective catalytic conversion of polycyclic aromatic hydrocarbons (PAH). This article is concerned with zeolite catalysis for making organic chemicals from polycyclic hydrocarbons. Venuto' and Chen et al. 1 3 9 1 7 have published comprehensive reviews of zeolite catalysis for organic reactions. It is clear from these reviews that zeolites have great potential as catalysts for organic synthesis. However, compared to the successful use of zeolites in petroleum processing and petrochemicals production, their application for specialty chemicals, particularly multi-ring compounds, has been very limited. The needs for research on selective PAH conversion have been discussed recently.18-20 Recent years have witnessed significant growth of existing aromatic polymer materials and rapid development of advanced aromatic polymer materials such as engineering plastics, polyester resins and fibers, polyimides, and liquid crystalline polymers (LCPs). Scheme 1 shows some of these aromatic polymer materials. 1 8 9 1 9 The advanced polymer materials of interest include thermoplastic polyethylene naphthalate (PEN), polybutylene naphthalate (PBN), thermotropic LCPs and heat-resistant polymers, in addition to well-known polyethylene terephthalate (PET), polybutylene terephthalate (PBT), polycarbonate (PC), and polyphenylene oxide (PPO). All these polymers require a one- to four-ring aromatic monomer with a specific structure (more linear ones in most cases). The key to the large-volume application of these materials (as resins, films, fibers and engineering plastics) is the development of highly efficient catalytic processes for making the desired monomers. Scheme 2 shows structures of the important monomers for aromatic polymer materials. Therefore, selective PAH conversion research can be applied to making chemicals from one- to four-ring compounds that are rich in some industrial process streams (by-product tars from coal carbonization or gasification, oil refinery streams etc.). Many high-value aromatic chemicals and important monomers from two- to three-ring polycyclic compounds for advanced aromatic polymer materials have been discussed in some recent reviews. This chapter is a review of recent studies which demonstrate that valueadded organic chemicals can be synthesized from two-ring and three-ring polycyclic hydrocarbons by selective conversion over certain 12-MR zeolite catalysts or zeolite-supported metal catalysts. Specifically, the review will focus on the new developments in the following catalytic reactions over large-pore zeolite catalysts: (1) ring-shift isomerization of phenanthrene derivatives to anthracene derivatives; (2) shape-selective alkylation of naphthalene for synthesizing 2,6-dialkylnaphthalene; (3) shape-selective alkylation of biphenyl for synthesizing 4,4'-dialkylbiphenyl; (4) conformational isomerization of cisdecahydronaphthalene into trans-decahydronaphthalene; (5) shape-selective hydrogenation of naphthalene into either cis- or trans-decalin and (6) regioselective hydrogenation of 1-naphthol and quinoline for synthesizing specialty fuel additives. The products of such selective reactions are intermediates for making specialty chemicals, monomers of advanced polymer materials such as 18919

Catalysis

274 Thermoplastic Polyesters

Polyethylene Terephthalate (PET) / Dacron, Mylar 0

Polyethylene Naphthalate (PEN) / Teijin, BP Ammo

Polybutylene Terephthalate (PBT) / Valox, Celanex 0

Polybutylene Naphthalate (PBN) I Teijin's PBN Resin

Polycarbonate (PC)/GEs Lexan, Dow's Calibre

Polyphenylene Oxide (PPO) / GE's PPO, Noryl w3

Thermotropic Polyester LCPs BP Ammo's Xydar; Sumitomo's Ekonol

Hoechst Celanese's Vectra High-Temp Heat-Resistant Polymers 0

n

Polyimide / Du Pont's Kapton

Polybenzimidazole/ Celanese's Celazole

Scheme 1

Structures of some important aromatic polymer materials

high-performance polyesters (including polyethylene naphthalate) and liquid crystalline polymers and components of advanced thermally stable aviation jet fuels for high-Mach aircraft, The possible reaction mechanisms, pathways and factors affecting the selectivity will also be discussed. The main purpose of this review is to demonstrate the usefulness of zeolite catalysts for selective synthesis of polycyclic specialty chemicals. This review

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

275

OH

0

0

HO-!e!!-OH 0

HO-C

Scheme 2

0

HOO -H !-

HO

Some aromatic chemicals and monomers ofpolymer materials

describes some new developments on environmentally-benign synthesis of useful chemicals from conversion of polycyclic hydrocarbons over large-pore zeolites including mordenites and Y-zeolites (and medium-pore molecular sieves). The results given here are selected examples showing the level of selectivity and conversion. 2

Selective Synthesis of Polycyclic Specialty Chemicals

Selective catalysis with polycyclic hydrocarbons is in general more difficult than with monocyclic hydrocarbons for two primary reasons. First, the former has a larger number of reactive sites on the ring system, and in many cases the more reactive sites lead to undesired products. Therefore, control of selectivity requires tailoring both the catalysts and reaction conditions such that the more easily formed but undesired products can be minimized. As pointed out clearly by Weisz,8 the unique role and purpose of a catalyst is to provide selectivity to direct the chemical transformation along a very specific, desired path. The selective feature of catalysis, whereby reactions can be directed onto desired pathways, is not simple with polycyclic compounds because the so-called shape-selective reactions can still give two or more products. For example, there are three isomers of dialkylbenzene (ortho, meta and para) while shapeselective alkylation of alkylbenzene can lead to the para-isomer as the desired product. On the other hand, there are ten isomers of dialkylnaphthalene, and the so-called P,P-selective reaction can still give three products (see below). Second, formation of carbonaceous substances can take place during acid-

276

Catalysis

catalysed reactions of polycyclic compounds over zeolites, which can passivate acid sites and block the pore channels. Our attention on chemicals has focused on shape-selective catalytic synthesis of value-added chemicals from polycyclic aromatic compounds that are rich in coal liquids and some refinery streams. We are studying ring-shift isomerization of phenanthrene derivatives to anthracene derivatives,21-23 shape-selective alkylation of naphthalene, 18,24-29 shape-selective alkylation of b i p h e n ~ l , ~ ~ ? ~ ~ and selective conformational isomerization of cis-de~ahydronaphthalene,~~~~~ hydrogenation of two-ring aromatics such as n a ~ h t h a l e n e as , ~ described ~~~~ below. 2.1 Ring-shift Isomerization. - Phenanthrene and its derivatives are rich in various coal-derived liquids such as coal tars, but their industrial use is still very limited. On the other hand, anthracene and its derivatives have found wide industrial application^^^^^^ We have found that some mordenite and ionexchanged Y zeolite catalysts selectively promote the transformation of symoctahydrophenanthrene (sym-OHP) to sym-octahydroanthracene (symOHAn) under mild c ~ n d i t i o n s . ~Scheme l - ~ ~ 3 shows such a conversion that we call ring-shift isomerization.21The term ring-shift was also subsequently used by Chen et al. in their review.13This reaction is in distinct contrast to the wellknown ring-contraction isomerization, which results in methylindane-type products. Such an isomerization is known to occur over Lewis acid catalysts such as A1Cl3 and MoCl3I8 but a zeolite-catalysed reaction process would be more environmentally friendly.

-003 7

sym-OHP

Scheme 3

sym-OHAn

Ring-sh@ isomerization

Figure 1 shows the relationship between sym-OHP conversion and selectivity to sym-OHAn over HML8 catalyst.23 Table 1 shows some representative results.23The properties of the Y-zeolite and mordenite catalysts in Table 1 are described elsewhere.21The selectivity and activity of the catalysts also depend on the reaction conditions, as shown in Figure 1 and Table 1. Under mild conditions, some zeolites show more than 90% selectivity to sym-OHAn with high conversion of sym-OHP. Figure 2 illustrates the changes in selectivity as a function of conversion over HML8 catalyst.23Many catalysts have been tested in our laboratory for this reaction. In general, when conversion is over 50%, selectivity to ring-shift decreases rapidly with further increasing conversion, as shown in Figure 2. This is because equilibrium conversion of sym-OHP to sym-OHAn is about 5 0 ' 3 0 . Further ~~ increase in conversion is caused mainly by sequential reactions which involve ring-contraction, ring-opening and cracking. Detailed results

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons -f- 200

277

'C

+250 "C 100

90 n

80

$

60 50

0

2

4

6

8

10

12

Time (h)

Figure 1

Temporal plot of sym-OHP conversion and selectivity towards sym-OHA plus 1,2,3,4-THA over HML8 at three different temperatures (200,250, and 300 "C)under N2

Table 1

Ring-Shqt isomerization of sym-OHP over zeolite catalysts in the presence of mesitylene solvent

Catalyst ID

Run conditions

HY LaHY NiHY HML8 HML8 HML8 HML8 HM30A a

250°C-1 h 250°C-1 h 250°C-1 h 250 "C-O.5 h 250 "C-2 h 200 "C-O.5 h 300 "C- 0.5 h 250°C-1 h

(%I

Yield (% of OHP) symTHAn THP OHAn"

72.7 73.4 51.5 50.3 51.9 13.1 55.4 55.1

22.1 21.8 46.9 48.8 49.3 14.7 42.7 42.9

OHP Conv.

3.4 3.6 1.2 1.o 1.3 0.1 2.1 2.1

5.3 5.5 4.5 1.7 1.7 1.4 3.7 2.5

% Selectivity

Ring-shVt 31.1 30.7 87.9 93.1 91.9 91.2 75.6 76.2

The feed originally contained 91% sym-OHP, 2.9% sym-OHAn and 6.1% others.

have been published recently for various HY and metal-ion exchanged Y-zeolites and several different mordenites at different temperature^.^^ A possible mechanism was proposed23 to explain the conversion of symOHP to sym-OHAn on zeolite, as shown in a modified version in Scheme 4. The first requirement would be the appropriate adsorption of sym-OHP on the catalyst surface. After the adsorption, the reaction is likely initiated by the protonation of the central aromatic ring in sym-OHP on Brarnsted acid sites. However, the protonated intermediate could lead to several different products

278 1.o

Y

+

M

M*

c.

0.8 -

-$ P l

. I

+

0 0

0.6 -

.B

0

250 "C (HY)

A 250 "C (N1Hy) A 250°C(LaHY)

a 0.2

250 "C (HM20A) 250 "C (HM30A)

-

0.4

250 OC (HMLI)

300 "C (HML8)

@O

h c)

. I

200 "C (HMLS)

+ .+

M

I

I

20OoC(HM30A)

I

Figure 2

-

ZO-H +

THP

H sym-OHP

+-

Cracking Products

/

I

sym-OHAn

Scheme 4

ZO-

H

THAn

Reaction pathwaysfor ring-sh$t of sym-OHP

due to ring-opening cracking, alkyl chain isomerization and subsequent cracking. In fact, the cracking reactions occur extensively on H-Y at 2 250 "C, as can be seen from Table 2. Therefore, the second step is dependent upon whether or not the positive charge is stabilized. For the ring-shift isomerization, the cationic intermediate should be stabilized by nearby anionic sites. However, the reaction mechanism for the ring-shift isomerization has not yet been clearly established. Since zeolite acidity is associated with the aluminum ions, a certain level of density of acid sites is required for effective ring-shift isomerization. From the above considerations, good zeolite catalysts should possess appropriate pore structure and the desired density and strength of the acidic sites for ring-shift isomerization of sym-OHP. This suggests that highly Al-deficient zeolite may not be suitable for this reaction. The fact that a hydrogen mordenite HML8 displayed higher selectivity than many others examined at 250 "C in both aliphatic and aromatic solvents suggests that it has characteristics closer to those required for the reaction at 250 "C.

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table 2

279

Isopropylation of naphthalene with isopropanol over zeolites at 250 "C (ref. 24)

Catalyst ID a

Structure Type

Naph. Conv. (%)

Yield (% of Naph.) IPN MIPN

% Isomer Select. % 2-IPN % 2,6-DIPN

HY HM30A

Y Zeolite Mordenite

38.2 27.6

90.4 84.1

33.8 88.6

a

9.5 15.9

16.0 67.0

The alkylation was conducted in the presence of mesitylene solvent.

This zeolite-catalysed ring-shift isomerization could provide a cheap route to anthracene and its derivatives, which are valuable chemicals, from phenanthrene that is rich in liquids from coal. Possible uses of sym-OHAn include the manufacturing of anthracene (for dyestuffs), anthraquinone (pulping agent), and pyromellitic dianhydride (the monomer for polyimides such as Du Pont's Kapton).18 In addition to the above-mentioned study, we conducted the isomerization tests under H2 pressure to further explore the effect of operating conditions. In our recent work on conformational isomerization of cis-decalin to transdecalin, the presence of H2 was found to promote the isomerization substanti all^.^^ However, when the reaction of sym-OHP was carried out using the zeolite-supported metal catalyst under H2 pressure, hydrogenation of symOHP to perhydrophenanthrene was significant. Simultaneous hydrogenation and ring-shift isomerization of phenanthrene was also explored in our laboratory using zeolite-supported metal catalysts at 200, 250 and 300°C. The results show that we can perform the simultaneous hydrogenation and isomerization of phenanthrene at 250-300 "C, but the hydrogenation also leads to the formation of unsym-OHP (172,3,4,4a,9,10,lOaOHP) and the selectivity is not very high. Further study is needed.

2.2 Shape-selective Akylation of Naphthalene. - Until recently, only limited attention has been paid to shape-selective alkylation of two-ring aromatics such as naphthalene and biphenyl. Due to the demand for monomers for making advanced polymer materials such as PEN and PBN, 2,6-dialkyl substituted naphthalene (2,6-DAN) is needed now for making the monomers for PEN, PBN and LCPs. In some refinery streams such as LCOs and in tars or liquids derived from coal, naphthalene and its derivatives are major components. Shape-selective alkylation over molecular sieve catalysts can produce 2,6-DAN. There are ten possible DAN isomers. The P,P-selective alkylation over molecular sieve catalysts (Scheme 5) can produce 2-alkylnaphthalene and 2,6-, 2,7- and 2,3-DAN. The key challenge is to obtain 2,6-DAN with high selectivity, which means increasing the ratio of 2,6/2,7-DAN. Several recent papers described shape-selective isopropylation. Katayama and coworkers34 reported preferential formation of 2,6-diisopropylnaphthalene (2,6-DIPN)

Catalysis

280 a

P

Tz

P

m

J.

+ R-OH

mR R-Rm: R m A R

2-AN Scheme 5

2,6-DAN

2,7-DAN

2,3-DAN

Alkylation of naphthalene

using mordenite. On the other hand, Moreau and co-workers have observed that 2,6- and 2,7-disubstituted products are formed in equal yields with a ratio of around 1 when using mordenite and Y-zeolite catalysts for isopropylation with isopropyl bromide and for cyclohexylation with cyclohexyl bromide.35 The results from our laboratory show that by using partially dealuminated mordenite catalysts, regioselective alkylation of naphthalene can be achieved with over 65% selectivity to 2,6-DIPN by using isopropan01~~ with 2,6-DIPN/ ~ ~the - ~alkylating ~ agent with 2,7-DIPN ratio of about 3 or using p r ~ p y l e n e as 2,6-DIPN/2,7-DIPN ratio of > 2. Compared to parent mordenites, the partially dealuminated proton-form mordenites are more effective as shape-selective catalysts for isopropylation of naphthalene. The results on the effects of dealumination have been reported r e ~ e n t l y . ~ ~It- has ~ ~ been - ~ ~indicated y ~ ~ that 2,6-DIPN is slightly smaller than 2,7-DIPN.l4 Horsley and c o - ~ o r k e r have s~~ shown by computer simulation that the diffusion of 2,6-DIPN inside mordenite pore channel is easier than that of 2,7-DIPN, accounting at least partially for the observed selectivity of mordenite for 2,6-DIPN and higher 2,6-DIPN/ 2,7-DIPN ratio. Schmitz and Song25327328 have reported some simple and effective methods for enhancing the shape selectivity by using water and dealuminated mordenite. Table 2 shows some typical results for naphthalene alkylation with isopropanol as the alkylating agent in the presence of mesitylene solvent.24 Sugi and c o - ~ o r k e r have s ~ ~ reported detailed results on the effect of Si02/ A1203ratio of mordenite on naphthalene isopropylation with propylene. They analysed not only the bulk of the reaction products, but also the products trapped in the pores of mordenite after the isopropylation. The results reveal that inside the mordenite pore channel 2,6-DIPN was formed in a much higher selectivity than 2,7-DIPN, but the external surface sites contribute more to the non-selective reactions as well as coke formation. The same group also reported39 that the impregnation of cerium is an effective method for the deactivation of external acid sites of hydrogen mordenite, which leads to improved selectivity to 2,6-DIPN, up to 70%. Xenon adsorption and 129Xe NMR suggest that pore sizes of the hydrogen-mordenite with and without

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

28 1

impregnated ceria are very similar and thus ceria exists on external particle surface.39 The increase in selectivity to 2,6-DIPN upon ceria loading by impregnation and calcinations of precursor salt was attributed to the deactivation of external acid sites; such deactivation is ascribed to the amphoteric property39of ceria which can lead to neutralization of acid sites. The reactivity of 2,6-DIPN on Ce-loaded HM was shown to be lower than that on the parent HM under isopropylation condition^.^^ Schmitz and Song26have shown that the reactivity of 2,6-DIPN is substantially lower on dealuminated mordenite than on parent mordenite. It should be noted that the type of alkylating agents and added modifiers can also significantly affect the selectivity and activity of the mordenite catalysts. Our early work21 on naphthalene alkylation demonstrated that high selectivity to 2,6-DIPN can be achieved using isopropanol as alkylating agent, but our later work on isopropylation using propylene led to relatively lower selectivity and lower 2,6-DIPN/2,7-DIPN ratio.25*26 This suggested that alkylating agent has a role to play, and using isopropanol as alkylating agent may give higher selectivity than using propylene under comparable conditions. In the course of verifying the difference between the alkylating agents, we found a simple and effective method for enhancing the shape selectivity for alkylation with propylene using dealuminated mordenite together with a proper amount of added water or alcohol, as can be seen from comparison of Figures 3 and 4.27

2,7-DIPN

other-DIPN

c. O

I

I

'

'

I

I

'

I

'

I

'

I

-

0 0

20

40

60

80

100

H,O/Cat (mmovg) Figure 3

Influence of added water on naphthalene alkylation with propylene over HM38 catalyst at 200 C ( r e J 2 7 )

Catalysis

282 2-MIPN 2,6-DII" c

s 401

2,7-DIPN other-DIPN

MIPN

B

.a ;

:

40

2a0E 2

-.

. o-,-

hl,

A

A

-

9

DIPN TrIPN+

:;:?+J+ 6

Figure 4

20

Influence of added water on naphthalene alkylation with propylene over HM74 catalyst at 200 "C (ref: 27)

The product distribution shown in Figures 3 and 4 is relative distribution among the products, not the absolute yield based on the reactant. As shown in Figure 3, with propylene as the alkylating agent, selectivity for P-substitution of naphthalene increased when water was added to HM38 (Si02/A1203ratio = 38), and reached a maximum with a water-to-catalyst mass ratio of 0.80 (ca. 45 mmol water g-cat-'). Figure 4 reveals very interesting trends in isopropylation activity and selectivity of the dealuminated mordenite HM74 (Si02/A1203ratio = 74), together with added water. Small amounts of added water cause the catalyst activity to decrease, but a minimum is reached at about 20 mmol/g added water, and more added water results in higher conversion. A very large amount of added water always decreased activity. Reproducibility of the trends shown in Figure 4 was confirmed. The following model was proposed27 to explain the interesting trends observed in Figure 4 with HM74 catalyst: (1) In the range where H20/ Cat = 0-20 mmol g-', primarily, adsorption of water occurs on Brarnsted acid sites, as shown in Scheme 6, particularly those on the catalyst's external surface. Naphthalene conversion and non-selective alkylation reactions decrease due to passivation of catalyst sites (as also observed for the HM38 catalyst). Compared to HM38, HM74 is substantially more silica-rich and therefore more hydrophobic. The remaining catalytic activity at H20/Cat ratio of around 20 mmol g-' is probably due to the active sites inside the pore

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

283

7

H/o--..

H+

I

Type 11 Neutral Complex

Scheme 6

Interactions of water molecules with Brmsted acid sites on zeolite surface

O\

/O\

s1

~l/o\si/o\si+

H+

I

Scheme 7

Al / O

H+

I

A model for illustrating dissociation of water on Lewis sites to form Brmsted sites

channel. At lower H,O/Cat range (lower partial pressure of steam), it is more difficult for water to diffuse into the pore channel of HM74 as compared to that of HM38. (2) At higher partial pressures of water (H20/Cat=20-60 mmol g- *), water diffuses into the micropores and adsorbs dissociatively on strong Lewis sites, converting them to Brarnsted acid sites, as illustrated in Scheme 7. Naphthalene conversion increases in this range because more Brarnsted acid centers are generated, having appropriate acid strength to supply protons for the formation of the carbocation intermediate involved in the alkylation. Increased activity in this range may also be partially attributed to reduced coke formation upon water addition (see below). (3) Water added beyond H2OICat = 60 mmol g-' decreases the number of acid sites available for catalysing the alkylation reaction, either by adsorption on Brarnsted acid sites or by capillary condensation. Figure 5 shows the effect of water addition on coke formation. Among the results in Figures 3-5, of general importance was the increase in P,P-selectivity to 2,6-DIPN and 2,7-DIPN, and decrease in coke formation (Figure 5) with added water.27Without added water, P,P-selectivities were 56% and 77% for HM38 and HM74, respectively. Selectivities to the highly desired 2,6-DIPN

Catalysis

284

300

400

500

600

700

800

Temp ("C) Figure 5

TGA-derived weight loss proJles during oxidative removal of carbonaceous deposits on used HM74 catalysts from runs with the following added H2Ol Cat ratios (mmolg-I): ( a ) 0; ( b ) 3.7; ( c ) 14.4; ( d ) 34.8; (e) 60.1; (f)110.2 (ref 27)

isomer were 37% and 54%, respectively. With added water, p, p-selectivities increased to 90% and %YO,and 2,6-DIPN selectivities increased to 60% and 70% for HM38 and HM74, respectively. Figure 5 clearly shows that adding water decreased coke formation on HM74 catalyst during naphthalene isopropylation with p r ~ p y l e n e .The ~ ~ observed selectivity changes with water addition are consistent with the consideration for the passivation of external acid sites on both HM38 and HM74 by water, since further increasing water beyond the H,O/Cat wt ratio of 20 does not cause any significant changes in selectivity with HM38 and HM74 catalysts. There may exist other effects of H 2 0 that also contribute to the observed selectivity change. The inhibiting effect of water on coke formation on the catalyst increased with increasing amount of water over the whole range of H20/Cat ratios, regardless of its effect on catalytic activity or selectivity. This desirable effect of H 2 0 may be ascribed to several possible contributing factors, including deactivation of the sites responsible for coke formation, inhibition of the condensation of aromatic species on the surface, and promotion of the removal of coke precursor from the surface. While a better fundamental understanding requires further study, the above considerations are consistent with the following experimental observations from other studies. Sugi and co-workers3*have reported detailed results on the effect of Si02/A1203 ratio of mordenite on naphthalene isopropylation with propylene. They analysed not only the bulk of the reaction products, but also the products trapped in the pores of mordenite after the isopropylation. The results reveal that inside the mordenite pore channel 2,6DIPN was formed in a much higher selectivity than 2,7-DIPN, but the external surface sites contribute more to the non-selective reactions as well as coke formation. The same group also reported that the impregnation of cerium is an

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table3

285

Isopropylation of naphthalene at 300°C and cracking of 1,3,5triisopropylbenzene at 350 O C over H-mordenites and cerium impregnated mordenite (refs. 38, 39)

Catalyst

HM (10) HM(23) HM (128) Ce(30)HM(128)

Naph. conversion (%)

Isopropy lation DIPN Selectivity to Yield (%) 2,6-DIPN

Cracking TIPB conversion

44.2 92.4 85.6 74.4

8.4 46.6 43.5 36.3

43.7 52.3 35.9 0.6

16.8 46.5 54.1 70.8

Isopropylation: 1 g catalyst, 200 mmol naphthalene, 8 kg cm-2 propylene, 300 "C,4 h.

effective method for the deactivation of external acid sites of hydrogen mordenite, which leads to improved selectivity to 2,6-DIPN, up to 70%, as shown in Table 3.39 As will be discussed later in Section 4, the reactivity of 2,6DIPN has been shown to be substantially lower on dealuminated mordenite than on the parent m~rdenite.~' The selective methylation of naphthalene has also been reported in literat ~ r e . ~ ' .Fraenkel ~' and co-workers4' first published on selective naphthalene methylation over ZSM-5 type catalysts in 1986, but 2,6- and 2,7-isomers were not separated. Komatsu and Yashima4' recently reported on the selective formation of 2,6-dimethylnaphthalene (2,6-DMN) from methylation of 2methylnaphthalene with methanol on HZSM-5 and metallosilicates with MFI structure. They demonstrated that isomorphous substitution of A1 by other elements such as B and Fe and deactivation of external surface (by using basic nitrogen compound) can increase the selectivity to 2,6-DMN. They concluded that in order to obtain 2,6-DMN in high selectivity, it is effective to weaken the acid strength while keeping the pore dimension of the MFI structure constant (or, wider, if possible), which can be achieved by using Fe-MFI as a catalyst. Based on the literature, it is difficult for methylation of naphthalene over medium-pore zeolites to reach the same level of selectivity achieved in isopropylation to 2,6-DIPN with mordenite catalysts. The results in Tables 4 and 5 probably represent the level of highest selectivity in 2-MN methylation that has been achieved and reported4' in open literature so far using HZSMS and metallosilicates with MFI structure. When a bulky basic nitrogen compound is used to poison the external acid sites41or a catalytically inert compound is introduced to the external surface, the p-selective or shape-selective methylation reactions inside the MFI pore channel become the only dominant reactions, as can be seen from comparing the data for Sb-MFI and HZSM-5 in Table 5, and from comparing the data for HZSM-5 in Table 5 with those in Table 4. It remains to be clarified whether the higher 2,6-DMN/2,7-DMN ratio with Fe-MFI (as compared with HZSM-5) is due to its lower acid strength or other factors.

286

Table 4

Catalysis

Methylation of 2-methylnaphthalene with methanol on MFImetallosilicate (re$ 41)

Catalyst

HZSM-5

YO2-MN Conv. 1-MN yield, mol% DMN yield, mol% DMN fraction:! 2,62,72,31,21,31,61,72,6-/2,7-DMN Tm,, (NH3-TPD1, K 10.3, min"

Ga-MFI

Fe-MFI

B-MFI

Sb-MFI

25.1 13.9 9.9

16.9 6.0 9.7

13.4 1.1 11.0

4.5 0.5 3.7

4.5 0.1 3.9

26.2 21.6 5.9 5.5 4.8 17.4 13.9

34.5 27.4 7.5 2.1 4.9 12.6 9.5

49.2 28.4 9.9 2.9 1.8 4.8 2.9

47.5 31.0 7.7 4.7 2.2 3.7 3.2

51.4 36.6 7.8 3.0 0.5 0.4 0.5

1.21 565

1.26 530

1.73 520

1.53 510

1.40 490

0.8

0.5

1.o

5.3

29.5

Conditions: T=723 K, W/F = 2.48 (HZSM-5, Ga-MFI, Fe-MFI), 20.9 (B-MFI) or 11.9 g h mol-I (Sb-MFI). Data obtained at 30 min time on stream. a Larger values of t0.3 indicate increase in the pore tortuosity, probably by the deposited boron oxide or antimony oxide.

Table 5

Effect of 2,4-dimethylquinoline addition on methylation of 2-methylnaphthalene with methanol on MFI-metallosilicates ( r e j 41)

Catalyst

YO2-MN Conv. 1-MN yield, mol% DMN yield, mol% DMN fraction:! 2,62,72,31,21,31,61,72,6-/2,7-DMN

HZSM-5

Ga-MFI

Fe-MFI

B-MFI

Sb-MFI

9.4 0 8.6

9.6 0 8.6

10.5 0 9.5

3.4 0 3.1

3.8 0 3.6

46.0 37.7 10.3 5.2 0.6 0.1 0.2

49.5 36.9 10.7 1.7 0.8 0.1 0.2 1.34

56.0 31.4 10.9 0.9 0.6 0.1 0.2

55.5 35.0 7.9 0.9 0.4 0 0.3

52.5 37.7 7.3 2.1 0.3 0 0.2

1.22

1.78

1.59

1.39

Conditions: T = 723 K, W/F = 2.48 (HZSM-5, Ga-MFI, Fe-MFI), 20.9 (B-MFI) or 11.9 g h mol- (Sb-MFI). Data obtained at 30 min time on stream.

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

4,4'-DAB

3,4'-DAB

287

3,3'-DIPB

Scheme 8

Biphenyl alkylation

Table 6

Isopropylation of biphenyl with propylene over zeolites at 250 "C

Catalyst IDa

Conv. moPA

Product, moPA MIPB DIPB TrIPB

HM14 HM21 HM38 HM71B HM230

49 60 71 46 23

74

DIPB isomer selectivity, moPA 333,44,4Other ~

a

64 54

62 67

25 33 42 37 32

0.5 1.5 3.0 1.1 0.5

3.9 3.0 2.7 1.3 0.9

~~~~

17 15 13 11 10

66 72 72 83 87

13.8 10.8 12.2 4.9 2.5

Trailing digit indicates the Si02/A1203molar ratio of the mordenite catalysts.

2.3 Shape-selective Alkylation of Biphenyl. - Biphenyl and its derivatives are present in some refinery streams and in coal-derived liquids, although at concentrations lower than those of naphthalene derivatives. Shape-selective alkylation of biphenyl can produce 4,4'-dialkyl substituted biphenyl (4,4'DAB), the starting material for monomer of some LCP materials represented by Xydar. Partially dealuminated proton-form mordenite can be used as shape-selective catalyst for isopropylation of biphenyl (Scheme 8). Lee and G a r ~ e first s~~ published the effect of dealumination of mordenite on selective biphenyl isopropylation in 1989. They have demonstrated the beneficial effect of dealumination for selective formation of 4,4'-diisopropylbiphenyl (4,4'-DIPB). Sugi and c o - w ~ r k e r s have ~~?~ carried out a series of studies on biphenyl isopropylation over mordenites. They have reported on the influence of p r ~ p y l e n pressure, e~~ effects of Si02/A1203ratio of mordenites, on shape-selectivity and coke deposition46 and impact of cerium exchange47of sodium mordenite. It was shown in a report by Schmitz and Song29 that dealumination of some commercial mordenites by acid treatment first increases then decreases their activity, but increases their selectivity toward 4,4'-DIPB in isopropylation with propylene. Table 6 shows some typical results for selective biphenyl isopropylation over dealuminated m ~ r d e n i t e s .It~ ~was observed that dealumination of some commercial mordenites by acid (HCl or HN03) treatment first increases and then decreases their activity; the acid treatment generally increases their selectivity toward 4,4'-DIPB in isopropylation with propylene,25329v30and

288

Table 7

Catalysis

Effect of water addition on biphenyl isopropylation using propylene over HM38 at 200 "Cfor 2 h (reJ30)

HZOICat Conv. mmolg-I mot% 0 20.4 65.6

138.1

Table8

52 51

45 15

MIPB isomer DIPB isomer Product distribution, selectivity, moPA selectivity, moPA mo PA MIPB DIPB TrIPB Other 2343,3'- 3,4'- 4,4- other 72 67 68 83

27 30 31 15

1.1 1.6 0.9 0.5

0.7 0.8 0.4 0.9

7.2 5.8 1.3 1.4

23 23 19 13

70 71 80 85

2.2 2.2 1.6 2.4

14 14 13 15

75 77 82 77

9.1 6.8 4.3 6.1

Activity and selectivity of HA4 and tributyl phosphite treated mordenitefor biphenyl isopropylation at 250 "C (ref 48)

Catalyst (SiO2lAl203)

HM (10)

HA4 (20)

P Treated P-HM(I0)

P Treated P-HM(20)

BP Conversion (YO) Yield (mol%) IPB DIPB TIPB

40.7

34.7

31.8

31.0

32.4 7.9 0.4

27.8 6.8 0.1

25.0 6.8 0

24.5 6.5 0

4-IPB in IPB (YO) 4,4-DIPB in DIPB (YO)

63 57

70 78

76 82

76 79

Conditions: 250°C, 1 h, 1 g catalyst, 50 mmol biphenyl, 50 mmol propylene, 40 mL decalin

changes in mordenite structure were followed systematically by A1 and Si NMR.30More recently, we have found that addition of water to dealuminated mordenite is a simple method to inhibit deactivation of the partially dealuminated mordenite catalysts without losing activity and sele~tivity,~~ as shown in Table 7. Kikuchi and co-workers have shown that biphenyl can be selectively alkylated not only with m ~ r d e n i t ebut ~ ~ also with SAPO-1149catalysts. In both cases, deactivation of external surfaces was shown to increase the selectivity to 4,4'-DIPB. In their report on mordenite, they demonstrated that deactivation of external acid sites on mordenite by treatment with tributylphosphite is effective for improving selectivity to 4,4'-DIPB, as shown in Table 8.

2.4 Conformational Isomerization. - Commercial decalins obtained from naphthalene hydrogenation are almost equimolar mixtures of cis-decalin and trans-decalin (DeHN). In the course of studying sym-OHP isomerization using decalin as solvent, we accidently found that cis-decalin isomerizes to transdecalin over ion-exchanged Y zeolite and mordenite catalysts21 at low temperatures (250"C), as shown in Scheme 9. This reaction would require a temperature of above 400 "C in the absence of a catalyst.50

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons H

H

cis-DeHN

Scheme 9

289

trans-DeHN

Conformational isomerization of cis-DeHN

0.9 1

1

+

3 +

508 K (LaHY) 508 K 0 523 K 523 K 0 538 K (LaHY) 538 K (HY) 548 K (LaHY) 548 K (HY)

u.9 '

0.0

Figure 6

0.2 0.4 0.6 0.8 cis-Decalin conversion

1.0

trans-Decalin selectivity vs cis-decalin conversion plots for LaHY and H Y catalysts at four different temperatures

Figure 6 and Table 9 show some recent results from our lab~ratory.~' The catalytic reactions were mainly conducted at 200-250 "C for 0.15-8 h under an initial pressure of 0.79 MPa N2 or H2 using six catalysts: a hydrogen Y zeolite, a lanthanum ion-exchanged Y zeolite, a hydrogen mordenite, and three noble metal loaded mordenites. Selected results are given in Table 9. Pt- and Pdloaded mordenites displayed the highest selectivity towards trans-DeHN (nearly loo%), with a trans-DeHNlcis-DeHN ratio of about 13 under H2 at 200 "C;however, they are less effective under N2. Pre-reduction of Pt/HM30A could improve its catalytic effectiveness in a N2 atmosphere. These results reveal a molecular H2-promoted isomerization reaction of cisdecalin (cis-DeHN). HY, LaHY, and HM30A are not as effective as Pt- and Pd-loaded mordenites, and the catalyst with the lowest acidity (HM30A) displayed the lowest activity as well as selectivity to trans-DeHN. The activity for cis-DeHN conversion decreases in the following order: Pt/HM30A > Pd/ HM30A > Pt/HM20A > HY > LaHY > HM30A.31An overall kinetic model for the catalytic reaction was proposed and empirical equations capable of predicting reaction conversion and product yield were presented. The theoretical equilibrium compositions of trans-DeHN and cis-DeHN at several temperatures were calculated and correlated well with the experimental results over effective catalysts such as Pt- and Pd-loaded mordenites. In addition, a lower reaction temperature is found to be thermodynamically favorable. In considering the possible reaction mechanisms, this isomerization may

290

Table 9

Catalysis

Conformational isornerization of cis-decalin over zeolite catalysts at 200-250 O Cfor 2 h under 0.79 MPa N2 or H2 ~~

Catalysta

~~~~~

Reaction conditions

% trans-

% cis-

Decalin

Decalin

Decalin

Feed

48.34

50.62

1.04

0.95

LaHY LaHY HM30A

200 "C - N2 250 "C- N2 250 "C- N2

48.29 65.15 53.02

50.73 18.82 3 1.60

0.98 16.03 15.38

0.95 3.9 1.7

Pt/HM30A PdHM30A

200 "C- N2 200 "C- N2

62.40 54.50

33.46 42.43

4.14 3.07

1.9 1.3

Pt/HM30A PdHM30A

200 "C- H2 200 "C- H2

92.34 92.31

7.25 7.23

0.41 0.46

12.7 12.8

a

% Other products

trans-Dlcis-D ratio

Pt and Pd loading was 6 wt.%.

or

Catalyst

Catalyst

H

H

cis-DeHN

A199-Octalin Or A99 10-Octalin

trans-DeHN

Scheme 10 Proposed reaction pathway for isomerization of cis-DeHN into trans-DeHN

proceed through a two-step process involving dehydrogenation of cis-DeHN to form an olefinic intermediate (possibly A1.9-octalin, or A9.10-octalin,or both) followed by addition of hydrogen to this olefinic intermediate to give trans-DeHN, as illustrated in Scheme 10. It is speculated that the isomerization of cis-DeHN to trans-DeHN starts with dehydrogenation occurring on the metal surface to form olefinic intermediates, for example, A' *9-octalin.The olefinic compounds later are hydrogenated to trans-DeHN. Since the dehydrogenation and hydrogenation steps occur on the metal and involve hydrogen, pre-reduction of the catalyst and the presence of H2 could facilitate the isomerization reaction. The experimental results also suggests that the noble metals loaded on mordenites strongly promote the isomerization by catalysing both the initial dehydrogenation of cis-DeHN to form the olefinic intermediates and the subsequent hydrogen addition to the olefinic intermediate to form trans-DeHN. This consideration is supported by the following facts: (1) these catalysts became much less effective under N2 for isomerization but enhanced dehydrogenation of cis-DeHN; (2) the pre-reduction improved the catalytic

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

29 1

activity for isomerization under N2; and (3) these catalysts can drive the reaction to equilibrium quickly and highly selectively under H2 atmosphere. The presence of acid sites is beneficial to the conformational isomerization under mild conditions, since this reaction was observed initially with acidic zeolite alone.21The role of acid sites and their interactions with metal species remain to be clarified. Possible contributions by acid sites may involve (A) abstraction of hydride by a proton to promote the formation of A1*9-ocatalin or A9*'O-octa1in, (B) the subsequent protonation of the double bond followed by hydride addition and (C) interaction with metal sites which make the catalyst more effective. The interaction between the surface acid sites and metal species and possible hydrogen spillover may play a role in the catalytic isomerization. These factors need to be verified by further study. Now it is possible to achieve over 90% conversion of cis-decalin with 95% selectivity to trans-decalin with some zeolites at 200 OCe31The trans-decalin isomer has substantially higher thermal stability at temperatures above 400 0C.50 Possible applications of this process are high-temperature heattransfer fluids and advanced thermally stable jet fuels, which can be used both as heat sinks and as fuels for high-Mach a i r ~ r a f t . ~ O - ~ ~

2.5 Selective Hydrogenation of Naphthalene. - Complete hydrogenation of naphthalene in conventional processes produces mixtures of cis- and transdecalin. Our work on selective naphthalene hydrogenation is motivated from an accidental finding2* on zeolite-catalysed isomerization of cis-decalin and from the need to tailor the formation of desired isomers from two-ring compounds. Our previous studies on naphthalene hydrogenation showed that certain catalysts show higher selectivity towards cis-decalin or trans-decalin. More recently, we found that mordenite and Y zeolite-supported Pt and Pd catalysts can selectively promote the formation of cis-decalin or transd e ~ a l i nas , ~shown ~ in Scheme 11. Table 10 shows some representative results.32Catalytic selectivity was found to depend on both the metal and the zeolite. Consequently, the catalyst composition could be tailored for selective production of either cis-DeHN or trans-DeHN. For example, Pt/HY showed especially high selectivity (80%) for cis-DeHN and, unlike all other catalysts tested, did not promote the isomerization of cis-DeHN. On the other hand, equilibrium between the DeHN isomers was achieved with Pd/HM21, giving ca. 93% trans-DeHN at 200°C. In

Scheme 11 Selective hydrogenation of naphthalene

Catalysis

292

Table 10

Selective hydrogenation of naphthalene over zeolite-supported Pt and Pd catalysts in tridecane solvent ~~~~~~

~

Catalyst a

Conditions

% cis-Decalin

% trans-Decalin

translcis ratio

Pt/HY Pt/HM38

200°C-1 h 200°C-1 h

82.3 29.9

15.1 70.1

0.2 2.3

PdIHY PdlHM38

200°C-1 h 200°C-1 h

27.1 18.4

72.9 81.6

4.4

a

2.7

Pt and Pd loading was 6 wt.%.

general, Pd catalysts showed higher initial selectivity for trans-DeHN than Pt catalysts. Also, the isomerization of cis-DeHN to trans-DeHN was more rapid on Pd catalysts. Selectivity for trans-DeHN appeared to increase with the fraction of weak acid sites on the zeolite (measured gravimetrically, using TPD of n-butylamine). No correlation between metal crystallite sizes (determined from XRD line-broadening) and DeHN isomer selectivity was found. Now we can produce cis-decalin with over 80% selectivity (or over 80% trans-decalin) at 100% conversion by using some zeolite-supported catalysts at 200 0C.32cis-Decalin may have potential industrial application as the starting material for making sebacic acid. Sebacic acid can be used for manufacturing Nylon 6,lO and softeners. There is also an industrial need for selective production of tetralin, a chemical intermediate and a hydrogen-donor, from naphthalene. Partial passivation of some zeolite-supported noble metal catalysts by sulfur can make them highly selective for the production of tetralin during metal-catalysed hydrogenation of naphthalene at low temperatures.33 In addition, recent work has shown that the selectivity in hydrogenation of polyaromatic compounds such as naphthalene and phenanthrene depends on the type of metal and the type of supports as well as the metal-support combinations and reaction 2.6 Regio-selective Hydrogenation of Quinoline and Naphthol. - More recently we have begun to explore regio-selective hydrogenation of heteroatomcontaining aromatic compounds. Examples of such compounds are 1-naphthol and quinoline shown in Scheme 12. Partial hydrogenation of 1 -naphthol can give 1,2,3,4-tetrahydro-l-naphthol(THNol-1) and 5,6,7,8-tetrahydro-1naphthol (THNol-2). Under fuel hydrotreating conditions, hydrogenolysis of the C-0 bond can also take place. It is of interest to see whether we can selectively produce THNL- 1. Similarly, for partial hydrogenation of quinoline, either 1,2,3,4-tetrahydroquionoline(THQ-1) or 5,6,7,8-tetrahydroquionoline (THQ-2) can be produced. It is interesting to clarify how one of the two isomers can be produced selectively under practically useful conditions for catalytic processing. The hydroaromatic products of regio-selective hydrogenation have some unique applications. One example is their potential as hydrogen

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons OH

OH

THNol-1

THNol-2

THQ-1

THQ-2

OH

293

Scheme 12 Regio-selective hydrogenation

donors or radical scavengers for stabilizing fuels at high temperature^.^^^^^ Preliminary experimental work has been reported on regio-selective hydrogenations7 over various metal catalysts supported on zeolites, alumina and titania as well as activated carbon. As in the case of naphthalene hydrogenation described above, the type of metal and support were found to be important for achieving regio-selectivity (for example, to THNol- 1).

3

Effects of Mordenite Dealumination on Naphthalene Isopropylation

Isopropylation of naphthalene over mordenite catalysts has received more attention than other alkylation reactions of naphthalene in the l i t e r a t ~ r e . ~ ~ * ~ * , ~ ~ For mono-substitution, the equilibrium yields depend on the type of alkyl groups. As reported by Olah and Olah,60 among the two mono-alkylated (2-AN and 1-AN) isomers from AlC13-catalysed alkylation at 25”C, the equilibrium yield of 2-alkylnaphthalene increases in the order of methyl (75.5%) c ethyl (90.5%) < isopropyl(98.5%) and tert-butyl (100%). This earlier study by Olah did not include dialkylation. However, the selective dialkylation of naphthalene is much more difficult. There have been some reports in the open literature on methylation of naphthalene,40*41961.62 and oxidation of 2,6DMN is advantageous over that of 2,6-DIPN. However, in an extensive review by Tanabe and H ~ e l d e r i c hthat ~ ~ also covered naphthalene alkylation at some length, it was concluded that ‘a shape selective effect is not expected in the methylation of naphthalene because of the small size difference of the isomers, particularly of 2,6- and 2,7-isomers’. This is considered to be the major reason why many researchers worldwide are studying isopropylation. The so-called P,P-selective isopropylation of 2-AN can produce three different isomers, 2,6-, 2,7- and 2,3-DIPN7 with all of them originating from 2-AN. It is difficult to selectively produce 2,6-DIPN, because 2,7-DIPN is also a major product of Ij-selective reaction (Scheme 5). The challenge is how to enhance selective formation of 2,6-DIPN against 2,7-DIPN, and to increase the ratio of 2,6-DIPN to 2,7-DIPN. Hydrogenmordenite (HM) catalysts have been shown to be effective for the shapeselective isopropylation of naphthalene to produce 2-isopropylnaphthalene (2-IPN) and P,P-diisopropylnaphthalene which includes both 2,6- and 2,7-

294

Table 11

Catalysis

Properties of the starting mordenite samplesa

Porosity, cm3g-' Mordenite Si02 A1203 Na20 Si021A1203 Surface area, n? g-I ID wt.% wt.% wt.% ratio Total Microb Mesoc Total Microb Mesoc

NaM14 85.0 10.1 6.24 14.3 NH4M21 86.4 6.94 0.02 21.1 NH4M38 88.1 4.44 0.07 37.5 a

Data as reported by the !upplier. Mesopore refers to 20-600A pores.

466 606 512

456 536 429

10 70 82

0.31 0.32 0.30

Micropore refers to <20

0.17 0.21 0.17

0.14 0.11 0.13

A (diameter) pores.

DIPN. Importance of dealumination of mordenite has been shown by Lee et al.42 for shape-selective biphenyl isopropylation. Mordenites have excellent thermal stability and acid resistance. Consequently, a significant fraction of the framework aluminum can be removed by acid leaching without altering the basic structure. At the onset of this work, it was not clear how the dealumination conditions affect the selectivity and activity of mordenites for naphthalene isopropylation. Sugi and c o - ~ o r k e r shave ~ ~ examined the effect of Si/Al ratio of mordenites on naphthalene alkylation using samples from commercial source. In this section, we will review recent studies on the effects of dealumination of mordenite on shape-selective isopropylation of naphthalene. The effects of dealumination of three different mordenites on alkylation of naphthalene (and biphenyl) have been r e p ~ r t e d . ~ ~ -The ~O following discussion focuses on the effects of dealumination of mordenites (mainly by acid leaching) on their catalytic performance for shape-selective isopropylation to produce 2,6-DIPN. The influence of reaction conditions is also discussed. 18926927

3.1 Characterization of the Mordenite Catalysts. - Two different samples of mordenites, NaM14 (CBV IOA) and NH4M38 (CBV 30A), were obtained from PQ Corporation as powders with average particle size of 9-10 pm. Their properties are listed in Table 11. NaM14 (6.24 wt % Na20) was treated with 1 M NH4Cl at reflux, washed until free of chloride, and dried to generate the ammonium form (NH4M14, 0.19 wt YONa20). The acid forms, HM14 and HM38, were generated by calcination in air at 465 "C for 4 h. Dealumination was accomplished by stirring the hydrogen mordenites in aqueous hydrochloric or nitric acid at reflux temperature.26Time and acid concentration were varied to control the extent of aluminum removal. HM230 was prepared according to a procedure described by Lee et al.42 for extensive aluminum removal. Accordingly, NaM14 was first treated at reflux with 1 M HC1 to generate the HM54 sample. In the second step, HM54 was calcined at 700°C and treated with 6 M HNO3, followed by final calcination at 700°C. All other catalysts were calcined at 465 "C for 5.5 h. Table 12 lists the detailed conditions for preparing the dealuminated mordenites along with their chemical composition. The last two or three digits in the catalyst name give the Si02/A1203

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table 12

Dealurnination conditions and ICP-ES analysis of dealurninated rnordenites

Mordenite Start ID with

Stir at Reflux in

Duration Calcine SiO,, A1203, Na20, Si0,1Al,03 h T,"C wt.% wt.% wt.% Ratio 6.24 0.19 0.15 0.15 <0.01 cO.01 <0.01 <0.01

14.3 13.1 53.6 53.6 61.9 70.0 71 .O 74.7

90.2 97.7

1.95 0.14 1.85 <0.01

78.5 89.6

465 700

95.6 97.4

1.51 CO.01 107.4 0.73 <0.01 226.4

N a M 3 8 - used as supplied 3MHCl 2 HM38 3MHC1 24 HM38

465 465 465

88.1 91.8 91.5

0.07 37.5 4.44 1.67 <0.01 93.3 1.15 CO.01 135.0

NH4M21 - used as supplied 3MHC1 24 HM2 1

465 465

86.4 92.0

6.94 1.94

NaM14 HM14 HM54 HM54-A HM62 HM70 HM7 1 HM74

NaM14 NaM14 NaM14 NaM14 HM14 HM14 HM14 HM14

- used

as supplied 1 MNH4Cl 24 1 MHCl 24 1 MHCl 24 1MHCl 2 5 1MHC1 1 MHCl 10 1 MHCl 24

465 465 465 700" 465 465 465 465

85.0 83.5 90.9 90.9 91.6 89.5 92.4 95.5

HM79 HM90

HM14 HM14

3MHCl 3MHC1

2 24

465 465

HM110 HM230

6MHN03 HM54 HM54-A 6 M H N 0 3

24 24

HM38 HM93 HM140 HM21 HM8 1 a

295

10.1 10.8 2.88 2.88 2.51 2.14 2.21 2.17

0.02 0.06

21.1 80.5

Not available (sample used for preparing HM230).

Table 13 Surface area and pore volume analysis of dealuminated rnordenites H- Mord. Catalysts

Si02lAl20, Ratio

Surface Area ( d g - ' ) Total Micro Meso

HM38 HM62 HM70 HM7 1 HM74 HM90 HM110 HM230

37.5 61.9 70.0 71.0 74.7 89.6 107.4 226.4

512 504

556 572 583 540 539 498

429 41 3 395 497 509 47 1 480 437

82 91 161 75 74 69 59 60

Pore Volume (cm3 g - ' ) Total Micro Meso 0.29 0.25 0.28 0.35 0.39 0.3 1 0.36 0.34

0.17 0.16 0.18 0.19 0.20 0.19 0.18 0.17

0.13 0.09 0.10 0.13 0.15 0.12 0.14 0.14

molar ratio. Table 13 shows the physicochemical characteristics of the dealuminated samples. The Si02/A1203ratio in Tables 12 and 13 is molar ratio determined by bulk chemical analysis of zeolites using ICP-ES. The Si02/ A1203 ratio increased upon acid leaching under any of the dealumination treatment conditions employed in this work. However, the removal of some aluminum species was relatively fast and the remaining A1 species was more

Catalysis

296

I . . . . . . . . . . . . . . , . . . . . . . 120 80 40 0 -40

-80

6 (PPm)

Figure 7 (a) 27AlMAS NMR spectra of dealuminated hydrogen mordenites. The trailing digit in the name of the samples indicates their Si02lA1203 ratios Si( 1Al) \

Si(0Al) I

I . . . . . . . . . . . . . . . . . . .

-50

-75

-100

-125

-150

6 (PPm)

Figure 700) 29Si MAS NMR spectra of dealuminated hydrogen mordenites. The trailing digit in the name of the samples indicates their SiO2lAl,O3 ratios

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table 14

297

Contents of tetrahedral aluminum in framework and in unit cell of mo rdenites S i 0 , l A l ~ Obased ~ on

Catalyst

% AlT"

AldAl$

Aldunit cellC Chem. anal.

HM14 HM38 HM62 HM7 1 HM74 HM230

77 86 95 97 98 99

3 6 21 28 48 173

5.1 2.1 1.4 1.3 1.2 0.4

13.1 37.5 61.9 71 .O 74.7 226.4

AlT from N M R

16.9 43.5 64.9 73.5 76.3 227.7

Percentage of the tetrahedral A1 peak among the total area of A1 peaks. Ratio of the tetrahedral-to-octahedral A1 peak areas. Number of aluminum ions per unit cell.

a

difficult to extract. As shown in Table 12, the decrease in A1 content with increasing time of acid leaching was faster in the first few hours. For example, the treatment of HM14 in 1 M HCl for 2 h increased the Si02/A1203 ratio from 13.1 to 61.9, but increasing the acid leaching time from 2 to 5 h only increased the ratio to 71.O. The increase in leaching time from 5 to 10 h resulted in essentially no change in the Si02/A1203 ratio with the error range of chemical analysis, and further increasing the leaching time to 24 h only caused a slight increase in the Si02/A1203 ratio to 74.7. A much higher degree of dealumination was achieved only by combining the acid leaching with a higher calcination temperature (700 vs. 465 "C). Figures 7a and 7b show the 27Al NMR and 29Si NMR spectra of the representative samples of parent and dealuminated mordenites, respectively. The 27Al MAS NMR spectra shows a progressive decrease of octahedral A1 peak (Alo) around -3.5 ppm and a gradual increase in sharpness of the tetrahedral A1 peak (AIT) around 52 ppm, and the 29SiNMR spectra show the corresponding changes, with increasing degree of dealumination. The results of 27Al NMR indicated that HCl leaching removed octahedral A1 species. The 29Si NMR spectra in Figure 7b revealed that dealumination reduced the Si species that are bound to two A1 atoms through 0 bridge but had little impact on the Si species that are bound to only one A1 atom. Table 14 shows the contents of tetrahedral aluminum in framework and in unit cell of mordenites. The framework Si02/A1203ratio determined from 27Al NMR is generally larger than the Si02/A1203ratio from bulk chemical analysis, but the difference becomes smaller with increasing degree of dealumination, as can be seen from Table 14. Figure 8 shows the XRD patterns of the dealuminated mordenites. We have also performed a detailed XRD crystallographic calculation, and derived the unit cell dimensions shown in Table 15. These results revealed a reduction of unit cell dimension, particularly the unit cell volume, of mordenite upon dealumination. From the results of porosity and surface area analysis, de-

Catalysis

298

I

I

5

10

1

*

20

15

25

30

35

28

Figure 8

Powder X R D of dealuminated hydrogen mordenites. The trailing digit in the sample name indicates the SiOzlA1203 ratio. The * mark indicates the peak due to silicon internal standard

Table 15 Lattice constants determined from XRD of dealuminated rnordenitesa Lattice constants, 2 b b c

H-Mord. catalyst

Si02lA1203 Ratio

a

HM14 HM54 HM7 1 HM74 HMllO HM230

13.1 53.6 71.0 74.7 107.4 226.4

18.16 k 0.02 18.05 k 0.01 18.06 k 0.02 18.09 f 0.03 18.07 f 0.02 18.06 f 0.02

20.31 k 0.01 20.16 k 0.01 20.16 f 0.02 20.22 f 0.03 20.23 k 0.02 20.22 k 0.02

7.49 k 0.01 7.45 f 0.01 7.44 f 0.01 7.46 k 0.01 7.46 k 0.01 7.46 k 0.01

Unit cell vol., 23

2764 f 3 2710 f 3 2708 f 3 2730 k 4 2730 2 4 2724 k 4

A,

a Radiation CuKa1, 1.540598 Ni filter. Error estimates are three times the standard errors determined in the unit-cell refinement regression analyses.

alumination changed mesopore volume considerably but did not alter micropore volume to any significant extent. Table 16 shows the acidity and acid strength distribution for the mordenites that have also been studied by NMR and XRD. Clearly, dealumination treatment generally decreased the acidity and the number of strong acid sites (n-BA desorbed at 340-500°C). Reduction in number of acid sites can be

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table 16

299

Acidity of dealuminated hydrogen mordenite catalysts measured by nBA TPD

Sample

Weak" (100-240 " C )

Acidity (mmolla) Intermediate Strong (240-340 "C) (340-500 "C)

HM 14

0.15

0.08

0.92

1.15

HM38

0.52

0.14

0.60

1.26

HM47 HM47b

0.59 0.65

0.18 0.18

0.49 0.51

1.26 1.34

HM62

0.60

0.16

0.44

1.20

HM7 1

0.55

0.21

0.42

1.18

HM74 HM230 HM230b HM230"

0.51 0.43 0.53 0.52

0.17 0.13 0.12 0.16

0.46 0.15 0.21 0.20

1.14 0.7 1 0.86 0.88

Total (100-500 "C)

Including physically adsorbed n-BA. Repeated TPD experiments on same n-BA adsorbed sample. One more repeated TPD experiment on same n-BA adsorbed sample.

a

partially responsible for enhanced selectivity, and this is consistent with a theoretical, computational analysis using molecular orbital package (MOPAC). As discussed later in this chapter, the 6-position in 2-IPN has higher frontier electron density in HOMO compared to that of 7-position, thus the former is more reactive for electrophilic substitution. In this context, reduced acidity can lead to better selectivity to 2,6-DIPN. However, this is not the only factor that affects the selectivity.

3.2 Selectivity and Activity of the Catalysts. - On the basis of quantitative GC analysis, the changes in the mon-, di- and polyalkylated naphthalene product with dealumination were observed, Peaks for P, P-substituted DIPN isomers increase, while peaks due to other DIPN isomers and tri- and tetraisopropylnaphthalenes (TrIPN+) diminish with dealumination. Products marked as C,-substituted naphthalenes (n = 5, 7, 8 or 9) give the characteristic mass spectral parent ions, but the fragmentation patterns indicate that these products are not solely isopropyl-substituted. The yield of these alkylnaphthalenes is not significantly decreased by dealumination. Greater than 98% of the products are isopropylated naphthalenes. Side reactions result in small amounts of alkylnaphthalenes that are not solely isopropyl-substituted. Mass balances are greater than 96% in all cases, with material losses being primarily attributed to carbonaceous deposits on the catalyst. Figure 9 shows the distribution of mono-, di-, and polyisopropylated (TrIPN+) products, and Figure 10 illustrates the isomer selectivity as a

300

Catalysis

Table 17

Effect of rnordenite dealumination on naphthalene isopropylation using propylene at 200 "Cfor 2 h

H-Mord. Catalyst ID

Naph. Conv. mol%

Isomer Selectivity, moPA Prod Distn, mol% MIPN DIPN MIPN DIPN TrIPN+ 22,6- 2,7- Other 2,6/2,7

HM14 HM54 HM62 HM70 HM7 1 HM74 HM79 HM90 HM110 HM230

76 43 78 82 74 47 69 36 15 41

64 76 62 53 61 76 66 80 84 74

33 23 34 41 37 23 30 19 15 25

3 1 4 6 2 1 4 1 1 1

60 68 54 58 64 71 59 70 83 74

HM38" HM93" HM140"

73 84 38

61 49 78

34

44 21

5 7 1

95

5

0

58 62 73 54

NaM14b

5.27

51 55 39 53 61 58

19 24 16 20 22 25 21 24 30 25

48 26 54 36 27 20 40 24 9 17

1.76 2.11 1.78 2.17 2.29 2.24 1.86 2.21 2.05 2.32

39 48 56

19 22 25

42 31 19

1.99 2.21 2.20

43

21

36

2.05

33 50 29

44

~

~~

a Derived from CBV30A (HM38). All the other samples derived from CBVlOA (HM14). A baseline test with sodium-form mordenite NaM14 (CBVlOA).

80

60

40

20

0 10

Figure 9

30

50

70 Naphthalene Conv (%)

90

Distribution of mono-, di-, and tri- and polyisopropylated products as a function of naphthalene conversion (for isopropylation with propylene using the standard reactor charge at 200 "Cfor 2 h) over HMl4-derived (filled symbols) and HM38-derived (open symbols) H-mordenite catalysts

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

70 -

50

\

30 1

Fl A

2,7-DIF"

-

A

10

I

I

30

50

70

90

Naphthalene Conv (%)

Figure 10

Isomer selectivity as a function of naphthalene conversion (for isopropylation with propylene using the standard reactor charge at 200 "Cfor 2 h) over HA414-derived (jlled symbols) and HM38-derived (open symbols) Hmordenite catalysts

function of naphthalene conversion. It is clear that yields of DIPN products generally increase at the expense of MIPN products with increasing conversion. Table 17 shows the isomer distribution for isopropylation over HM14 derived and HM38-derived catalysts as a function of Si02/A1203ratio. With HM 14-derived samples, dealumination can increase P,P'-selectivity to 2,6DIPN from about 30-60%, but the effects of Si02/A1203ratio on the product yields or on shape selectivity are not straightforward, as further discussed below. This trend can be seen clearly from Figures 11 and 12, where the yields of products and selectivity are plotted as a function of Si02/Al2O3ratio of the H-mordenite catalysts. A baseline test was conducted with NaM14, the precursor to HM14. As shown in Table 17, NaM 14 has very limited activity, as expected. However, the NaM 14 sample does display catalytic selectivity to 2,6-DIPN. It appears that there are internally located acid sites inside the channel of this sodium-form mordenite which promote shape-selective isopropylation reaction. Dealumination affects activity as well as selectivity. In particular, the changes between HM70, HM71, and HM74 are significant, although their differences in Si02/A1203ratio (Table 13) are very small. For HM38-derived samples, the P-selectivity and p, P'-selectivity appear to increase with increasing Si02/A1203ratio of the mordenites up to about 93. Further increasing the degree of dealumination (increasing Si02/A1203ratio up to 140) decreased the activity. The selectivity was also compared at similar conversion levels (see below). The changes in catalytic activity for naphthalene conversion between HM70, HM71 and HM74 are significant, but their selectivity values in terms of 2,6-/

Catalysis

302 80

-

60

-

40

-

20

0

50

100

150

200

250

Si02/Al203 (molar)

Figure 11

Distribution of mono-, di-, and tri- andpolyisopropylated products as a function of Si02lA1203 ratio of HM-14 derived (filled symbols) and HM38derived (open symbols) H-mordenite catalysts for naphthalene isopropylation (with propylene at 200 " Cfor 2 h using the standard reactor charge) 90 1

70

-

50

-

30

0

50

100

150

200

250

SiOZ/Al2O3(molar)

Figure 12

Isomer selectivity vs. Si021A1203ratio for naphthalene isopropylation with propylene (using the standard reactor charge at 200 "Cfor 2 h ) over HM14derived (filled symbols) and HM38-derived (open symbols) H-mordenite catalysts

2,7-DIPN ratio are similar to each other (2.17, 2.29, and 2.24, respectively). It can be concluded that there are no direct linear correlations between the SiO,! A1203molar ratio (from chemical analysis) and the product distribution and selectivity. This may be a reflection of a key structural factor that it is not the number of acid sites but the number of accessible catalytic sites that controls

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

303

catalytic activity. There may be active sites that are accessible to the probe molecules but not to the multi-ring reactant molecules. Although the SiOJ A1203molar ratio is directly proportional to the total number of acid sites, this ratio does not provide information on the location and distribution of acid sites. Sugi and c o - ~ o r k e r sexamined ~~ a series of H-mordenites with different Si02/A1203 ratio for isopropylation of naphthalene. They indicated that aluminum concentrations at intracrystalline and external surfaces of Hmordenite are not directly related to catalyst performances. However, dealuminated H-mordenite with a Si02/A1203ratio higher than 30 exhibited high catalytic activity and high selectivity for 2,6-DIPN. They attributed the enhancement of catalyst performances (with the increase of Si02/A1203ratio) to the suppression of coke deposition and the increase of shape-selective catalysis in the pores because the dealumination caused the decrease of acid density and strength. They also reported that low selectivity for 2,6-DIPN is due to non-selective catalysis at external acid sites which are active in spite of severe coke deposition. Analysis of naphthalene derivatives encapsulated in the pores showed that 2,6-DIPN was formed selectively in the pores over all Hmordenites because of the minimum steric requirement at the transition state composed of substrates and acid sites.38 The same research group reported that ceria-modification is an effective method for the deactivation of external acid sites of H-mordenite (HM).39964The selectivity of 2,6-DIPN in the isopropylation of naphthalene (NP) was enhanced by the modification with 10-50 wt.% of cerium without significant decrease of catalytic activity: the highest selectivity for 2,6-DIPN was up to 70% over HM (Si02/A1203= 128) with 30 wt.% of cerium.64 3.3 StructureSelectivity Relationships. - Detailed analysis of XRD data for dealuminated mordenite samples revealed a trend of decrease in unit cell volume upon dealumination, as shown in Table 15. The unit cell volume decrease revealed by XRD data suggested a slight decrease in size of the micropores. This is in distinct contrast to an earlier conclusion that HCl leaching of mordenite may increase channel size.65 On the other hand, N2 adsorption data indicate an increase in mesopore volume upon proper dealumination, and the increase in mesopore volume appears to be parallel with increase in selectivity to 2-IPN and 2,6-DIPN, as shown in Figure 13 and Figure 14. However, it was not clear how an increase in mesopores can cause the increase in shape selectivity. The increased mesopore volume enhances diffusion of molecules, but does not account for increased 2,6-DIPN/2,7-DIPN ratio. Other changes may accompany the pore volume change due to dealumination. The relationship between isomer selectivity and unit cell volume (Figure 15, top) and unit cell dimension (Figure 15, bottom) shows that 2,6DIPN selectivity is more sensitive to the changes in unit cell volume or unit cell dimension. Apparently, the formation of 2,6-DIPN is more sensitive to the changes in unit cell parameters, but such changes are not directly proportional to the degree of dealumination or the Si02/A1203 ratio, as can be seen from

Catalysis

304 100 0.38

Total Pore Volume '-0

0.33

0.28

0.23

0.18

Q

m b a

20

0.13

0.08

3

Mesopore Volume

0

0

50

100

150

200

SiQ/&O, (Molar Ratio)

Figure 13

Pore volume (left axis) and isomer selectivity (right axis) vs. Si02IA1203 ratio of the H-mordenite catalysts for isopropylation with propylene using the standard reactor charge at 200 "Cfor 2 h

Figure 15. Another important observation is that the changes in the 2,6-DIPN/ 2,7-DIPN ratio parallel the changes in the unit cell volume. The reduced unit cell dimension (implying reduced pore size) could improve P, P'-selectivity and 2,6-DIPN/2,7-DIPN ratio. 2,6-DIPN has a more linear structure and a slightly smaller critical diameter than 2,7-DIPN.24-66y67 3.4 Effect of Reaction Conditions and Catalyst Loading. - The effect of reaction temperature and residence time and the influence of naphthalene/ catalyst mass ratio have also been examined under comparable conditions for several cata1y st s. Table 18 shows the effects of reaction temperature and residence time for isopropylation with propylene over HM 110 and HM230 catalysts. Increasing temperature increased naphthalene conversion and improved the yields of DIPN. Interestingly, the 2,6-DIPN/2,7-DIPN ratio also increased with increasing temperature from 200 to 275°C over these two catalysts. Increasing the residence time over both catalysts resulted in increases in both conversion and yield of DIPN. However, residence time change had no positive impact on the 2,6-DIPN/2,7-DIPN ratio. Table 19 compares the product distribution and selectivity at similar conversions for many dealuminated mordenite catalysts used in the isopropylation reactions under various conditions. Raw data are given in this Table for comparison at several conversion levels, where the corresponding reaction

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons 85 I

0

305

HMllO

' o 2-MIPN 70 n

8

W

-

. 1A

2,7-DIPN

55-

e

0

m

8s1 40:

A

U

A

A

10 ! 0.24

I

I

I

0.28

0.32

0.36

0.40

~ o t aPore l Volume (cm3/g)

1'

85 I

I

- .-_

-

! o 2-MIPN

cln

I

_

0

HMllO

HM230

2,6-DIPN

HM90

1)HM74

e

W . I

0

m

8 400

bz

l

10

!

0.08

o

A

I

I

I

0.10

0.12

0.14

0.16

Mesopore Volume (cm3/g) Figure 14

Isomer selectivity vs. totalpore volume (top) and mesopore volume (bottom) of HMIQ-derived and HM38-derived H-mordenite catalysts for naphthalene isopropylation with propylene (using the standard reactor charge at 200 C for 2 h)

conditions are also shown. There is a general trend that proper dealumination can enhance the selectivity to 2,6-DIPN and increase the 2,6-DIPN/2,7-DIPN ratio. However, not all dealumination treatments lead to good catalysts. It is also clear from Table 10 that neither the activity nor the selectivity is a simple linear function of the catalyst Si02/A1203 ratio. This is true even when the tetrahedral framework Si02/A1203ratio (Table 4) is considered. This is due in part to the difference in the geometric distribution and nature of the catalytically active sites.

Catalysis

306 s

o

OHM110

o

2-MIPN 2,6-DIPb

\

AHMlIO

10 ! 2700

I

I

I

2720

2740

2760

Unit Cell Volume 85

(A3)

'L

0 HM230

HM74

10

20.15

20.19

20.23

b Figure 15

0

20.27

20.31

(4

Isomer selectivity vs. unit cell volume (top) and unit cell dimension (bottom) of HMl4-derived and HM38-derived H-mordenite catalysts for naphthalene isopropylation with propylene (using the standard reactor charge at 200 C for 2 h)

The effect of the catalyst loading on the conversion and isomer selectivity is also a concern. Table 20 presents the data on the influence of naphthalene/ catalyst mass ratio under comparable conditions for HM38 catalyst at 200 "C for 2 h. The run for lower NapWCat ratio (0.62) was done in the 30 cm3 reactor, while the tests with higher NapWCat ratios (2.0-29.9) were conducted in a larger batch reactor with about 80 cm3 volume. Increasing NapWCat ratio decreased the conversion, but increased the yield of MIPN products. The initial increase in the NapWCat ratio range of 2-10 appears to be beneficial to

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table 18

T, " C

307

Effect of temperature on isopropylation over HA4110 and HM230 catalysts Isomer selectivity, mol% Prod. Distn., mol% MIPN DIPN Conv.,% MIPN DIPN TrIPN+ 22,6- 2,7- Other 2,6/2,7

HMllO catalyst, 2 h runs 200 15 85 250 28 79 275 46 71

14 21 28

0.4 0.6 1.o

83 86 87

61 65 66

30 26 25

9 8 9

2.05 2.48 2.67

HA41 I0 catalyst, 4 h runs 200 19 83 250 39 75 275 58 65

16 25 33

0.4 0.7 1.3

83 86 86

61 65 65

29 26 25

9 8 10

2.09 2.48 2.64

HM230 catalyst, 2 h runs 200 41 74 250 63 59 275 76 49

25 39 47

1.1 2.3 4.0

74 78 79

58 62 62

25 24 24

17 14 14

2.32 2.57 2.58

HM230 catalyst, 4 h runs 200 49 69 50 250 75 275 84 41

29 46 54

1.2 4.3 4.4

73 75 78

59 60 62

25 24 24

17 16 14

2.37 2.51 2.54

the 2,6-DIPN/2,7-DIPN ratio, but further increase in the ratio between 10 and 30 had no effect or somewhat negative impact on the 2,6-DIPN/2,7-DIPN ratio. These results suggest that the results obtained in micro-reactors can be reproduced in large reactors with respect to selectivity trends, although the absolute yields of products or conversion levels depends on the NapWCat ratio. 3.5 Effects of Mordenite Dealumination on Further Conversion of 2, 6-DIPN. There are two requirements for achieving high catalytic selectivity for 2,6DIPN: the selective formation of 2,6-DIPN, and the prevention of further reaction or conversion of 2,6-DIPN once it is formed. Catalyst sites on the external surface of mordenite can promote non-selective reactions which decrease the 2,6-DIPN selectivity. Examples of such reactions are isomerization, transalkylation, and continued alkylation leading to tri- and tetraalkylnaphthalenes. An assessment of 2,6-DIPN reactivity on the catalyst would be useful in understanding factors that control product selectivity. Therefore, experiments were performed to examine the reactivity of 2,6-DIPN on HM at 2OO0C, and how the reactivity changes with HM aluminum content and propylene pressure. 2,6-DIPN reactivity was measured on three HM catalysts (Si02/A1203ratio 14, 38 and 74), under propylene pressure, in batch reactor tests at 200°C.26The propylene pressure was varied to give propylene to 2,6DIPN mole ratios of 4.0, 1.O and 0.5.

-

308

Catalysis

Table 19

Comparison of selectivity at similar level of conversion over dealuminated mordenite catalysts Product distribution, molar basis

Mord. Cat. ID

T,

HM14 HM38 HM62 HM70 HM71 HM93 HM230 HM230

200 200 200 200 200 200 250 275

2.0 2.0 2.0 2.0 2.0 2.0 4.0 2.0

76.1 72.8 78.3 82.4 74.4 84.46 74.7 76.4

63.7 61.1 61.8 53.1 60.5 49.1 49.9 48.6

32.7 34.4 34.3 41.2 37.2 44.2 45.7 47.4

3.6 4.5 3.9 5.7 2.3 6.7 4.43 4.0

60.4 58.0 53.6 58.5 63.9 62.4 75.1 78.8

33.4 38.7 29.3 43.9 51 .O 47.7 60.4 61.8

19.0 19.5 16.5 20.2 22.2 21.6 24.0 24.0

47.6 41.8 54.2 35.9 26.8 30.7 15.6 14.2

1.76 1.99 1.78 2.17 2.29 2.21 2.51 2.58

HM38 HM54 HM74 HMllO HMllO HM230 HM230

200 200 200 275 275 200 200

0.5 2.0 2.0 2.0 4.0 2.0 4.0

51.9 43.2 46.8 46.2 57.8 41.4 49.1

72.1 76.1 76.3 70.9 65.4 73.5 69.5

25.6 22.7 23.1 28.1 33.3 25.4 29.4

2.3 1.2 0.6 1.0 1.3 1.1 1.1

60.7 67.9 71 .O 86.5 86.5 73.7 72.9

39.3 50.1 55.2 66.2 65.4 58.1 58.5

20.2 23.8 24.7 24.8 24.8 25.0 24.7

40.5 26.1 20.2 9.0 9.8 16.9 16.8

1.95 2.11 2.24 2.67 2.64 2.32 2.37

HM74 HM81 HM90 HMllO HM140

200 200 200 250 200

4.0 2.0 2.0 4.0 2.0

38.5 34.6 35.9 38.8 37.9

76.3 79.2 79.6 74.8 78.6

23.1 20.0 19.5 24.5 21.0

0.6 0.8 0.9 0.7 0.4

73.6 73.4 69.5 85.6 73.1

56.4 57.1 52.6 65.2 56.0

25.7 25.5 23.8 26.3 25.4

17.9 17.4 23.6 8.5 18.6

2.19 2.24 2.21 2.48 2.20

time%NAP % % % %2%2,6- %2,7- %other 2,6/2,7 "C h conv MIPN DIPN TRIPN+ in MIPNin DIPN in DIPN in DIPN ratio

Table 20

Effect of naphthalenelcatalyst mass ratio in isopropylation over HM38 ~~

NIP

Conv.,%

Isomer selectivity, mol% Prod. distn., mol% MIPN DIPN MIPN DIPN TrIPN+ 22,6- 2,7- Other 2,6/2,7

0.62 2.0 9.9 14.8 19.8 29.9

88 89 83 69 62 49

51 39 52 68 70 79

46 50 42 30 28 21

3.4 11.0 5.7 2.0 2.5 0.8

90 84 59 62 60 62

52 45 41 43 40 42

28 26 19 20 20 21

20 29 40 37 40 37

1.85 1.75 2.13 2.13 2.02 1.94

"Naphthalene to HM38 catalyst mass ratio in the reactor charge

Tables 21-23 show the results for reaction of 2,6-DIPN over the three HM catalysts, at the three different propylene12$-DIPN ratios. 2,6-DIPN conversions can be positive or negative (see Table 11) depending upon whether the reaction results in a net decrease or net increase in the amount of 2,6-DIPN. The concentration (in mol%) of each component is given along with the

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

Table 21

309

Reaction of 2,6-DIPN over HM14 catalysts under propylation conditions a

Propylene to 2,6-DIPN mole ratio % 2,B-DIPN conversion

4.0

1.0

0.5

27.1

25.4

23.9

Pdt. distribution (mol%, change) 2-MIPN 1-MIPN 2,6-DIPN 2,7-DIPN other-DIPN TrIPN TeIPN RNAP

0.0 (-2.4) 0.0 (-1.1) 69.6 (-25.9) 0.1 (-0.3) (0.0) 0.3 26.1 (25.9) 3.7 (3.7) 0.3 (0.1)

0.1 (-2.3) 0.0 (-1.1) 71.2 (-24.3) 0.3 (-0.1) 1.9 (1.6) 23.4 (23.2) 2.8 (2.8) 0.3 (0.1)

0.4 (-2.0) 0.0 (-1.1) 72.6 (-22.9) 1.1 (0.7) 4.6 (4.3) 19.1 (18.9) 1.9 (1.9) 0.3 (0.1)

Reaction at 200 "C for 2 h. Propylene pressure was varied for adjusting propylene/2,6DIPN ratio.

a

Table 22

Reaction of 2,6-DIPN over HM38 catalysts under propylation conditions a

Propylene to 2,6-DIPN mole ratio % 2,6-DIPN conversion

4.0

1.0

0.5

25.0

22.9

12.5

Pdt. Distribution (mol%, change) 2-MIPN 1-MIPN 2,6-DIPN 2,7-DIPN other-DIPN TrIPN TeIPN RNAP

0.1 (-2.3) 0.1 (-1.0) 71.6 (-23.9) 0.1 (-0.3) (0.0) 0.3 24.9 (24.7) 2.7 (2.7) (0.0) 0.2

2.0 (-0.4) 0.2 (-0.9) 73.6 (-21.8) 0.2 (-0.2) 1.2 (0.8) 19.6 (19.4) 1.9 (1.9) 1.2 (1.1)

0.1 (-2.3) 0.0 (-1.1) 83.5 (-11.9) 0.2 (-0.2) 1.2 (0.9) 13.4 (13.3) 1.2 (1.2) 0.3 (0.1)

a

Reaction at 200 "C for 2 h. Propylene pressure varied.

change in its concentration vs. the starting material. On HM14 and HM38 catalysts, 2,6-DIPN conversions were 12-27%, mostly to higher alkylates. It should be restated that only P-substituted isopropylnaphthalenes can be formed within the HM micropores. Reactions leading to a-substituted isopropylnaphthalenes must occur on external surface catalyst sites. Some isomerization of 2,6-DIPN occurred at the lowest propylene pressure, but was suppressed by increasing the propylene pressure. 2,6-DIPN is nearly unreactive on dealuminated mordenite (HM74). These results reveal that the high selectivity toward 2,6-DIPN in the

3 10

Table 23

Catalysis

Reaction of 2,6-DIPN over HM74 catalysts under propylation conditions"

Propylene to 2,6-DIPN mole ratio % 2,6-DIPN conversion Pdt. Distribution (mol%, change) 2-MIPN 1-MIPN 2,6-DIPN 2,7-DIPN other-DIPN TrIPN TeIPN RNAP a

4.0

0.8

0.0 0.0 94.7 0.1 0.1 4.4 0.5 0.2

(- 2.4) (- 1.1)

(-0.7) (-0.3) (- 0.2) (4.2) (0.5) (0.1)

1.0 -1.3

0.0 0.0 96.7 0.1 0.1 2.6 0.3 0.3

(-2.4) (-1.1)

(1.2) (-0.3) (-0.2) (2.4) (0.3) (0.1)

0.5 - 1.5

0.0 (-2.4) 0.0 (-1.1) 96.9 (1.4) 0.1 (-0.3) 0.2 (-0.1) 2.3 (2.2) 0.2 (0.2) 0.2 (0.1)

Reaction at 200 "Cfor 2 h. Propylene pressure varied.

isopropylation of naphthalene is partly attributable to the very low degree of further conversion of 2,6-DIPN once it is formed on dealuminated mordenites. On regular H-mordenites further alkylation and isomerization of 2,6-DIPN can take place on the acidic sites on the external surface, but the dealumination treatment can remove such external acidic sites as well as internal acid sites thus minimize the undesirable secondary reactions. 3.6 Computational Analysis of Shape-Selective Isopropylation. - Based on the experimental results alone, it was not clear whether the differentiation between the 2,6-DIPN and 2,7-DIPN isomers was caused by their differences in molecular dimensions or in electronic properties. In order to clarify this aspect, a computational analysis of the molecular dimensions and frontier electron densityf,(E) was carried out using MOPAC program for naphthalene, IPN and DIPN66as well as methyl- and ethyl-naphthalene isomers.67 Scheme 13 shows the definition of molecular dimension and energy-minimized conformations of 2,6- and 2,7-DIPN. Table 24 shows the molecular dimensions of IPN and DIPN isomers.66The calculation shows that 2,6-DIPN has a slightly smaller critical diameter than 2,7-DIPN. Table 25 shows the frontier electron density f,(E) for naphthalene and isopropylnaphthalene isomers. The f,(E) value for electrophilic substitution reaction represents the density of electrons in the highest occupied molecular orbital (HOMO). According to the frontier molecular orbital theory, the most reactive position (the carbon atom on which electrophilic attack occurs most likely) has the highest frontier electron d e n ~ i t y .More ~ ~ ? importantly, ~~ position 6 in 2-IPN has a higher f,(E) value than that of position 7 in 2-IPN, as shown in Table 25. This suggests that during 2-IPN isopropylation inside a mordenite channel, the formation of 2,6-DIPN is electronicallymore favored than that of 2,7-DIPN.

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

311

Scheme 13 Illustration of critical diameter, length, thickness and cylinder diameter of 2,6-DIPN (top, A ) and the steric structures including space-filling model of 2,bDIPN (middle) and 2,7-DIPN (bottom) in their more stable conformation A

312

Table 24

Catalysis

Molecular dimensions of naphthalene and alkylnaphthalenes Steric Heat of conform formation kcalmol-'

Molecule

Naphthalene 1-1PN 2-IPN 2,6-DIPN 2,7-DIPN

40.79 23.07 24.14 22.1 1 22.20 3.44 3.61 3.44 3.65 3.20

A B A B A B A B

2,3-DIPN

Critical a (iameter A

Molecular Molecular Cylinderb lfngth thickness (iameter A A A

7.19 7.90 8.28 7.21 7.53 7.2 1 7.53 7.26 7.76 8.97

8.95 9.99 9.52 11.03 10.94 13.14 12.65 13.14 11.99 10.71

A,

A

Van der Waals radius for H =1.10 for C ~ 1 . 5 5 cross-section circle of a molecular conformation

a

Table 25

Frontier electron density isopropylnaphthalene isomers

fr(E)

3.10 6.52 6.52 6.52 6.52 6.52 6.52 6.52 6.52 6.03

7.19 8.19 8.55 7.21 7.84 7.21 8.1 1 7.26 8.05 8.97

Diameter for a largest possible

of

naphthalene

and

Molecule

Carbon I

Carbon 2

Carbon 5

Carbon 6

Carbon 7

Naphthalene 1-IPN (A) 2-IPN (A) 2,6-DIPN (A) 2,7-DIPN (A)

0.347 0.373 0.373 0.350 0.366

0.153 0.181 0.193 0.2 17 0.159

0.347 0.31 1 0.329 0.349 0.325

0.153 0.137 0.177 0.217 0.135

0.152 0.143 0.122 0.082 0.159

Therefore, for p-selective isopropylation of 2-IPN inside mordenite channel, formation of 2,6-DIPN is favored more than that of 2,7-DIPN, because position 6 has a higher frontier electron density and thus higher reactivity towards electrophilic substitution. Scheme 14 shows the energetics of the intermediates formed upon isopropyl cation attack on 2-IPN. The formation of 2,6-DIPN is favored electronically more than those of 2,7-DIPN and 2,3DIPN. This may be termed as spatially-restricted electronic transition-state selectivity affected by frontier electron dens it^,^^.^^ which is different from the well-known transition-state selectivity (one of the three types of shape-selective reactions, discussed in refs. 10, 13 and 17). The fact that 2,6-DIPN has a more linear structure and a slightly smaller critical diameter (as compared to 2,7DIPN) makes it easier to diffuse out. 3.7 Factors Affecting Shape-selective Naphthalene Alkylation. - Many batch tests have been conducted on naphthalene isopropylation. On the basis of the experimental results, the basic factors that affect the selectivity of the 18926-30

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

313

The weakest bond in the intermediate

[2-IPN + c-d-C] Intemrediate (&position) AH = 184.94 W m o l Most stable

[2-IPN + c-e-c] Intermediate (7-position)

AH = 186.28 kcaymOl

[2-IPN + c-d-C] Intumediate (3-position) AH 186.74 kcaVmol Lest stable

-

Scheme 14 Energetics of cationic intermediates formed by attack of an isopropyl cation at the 6-, 7- and3-positions of 2-IPN leading to 2,6-DIPN (top), 2,7-DIPN (middle) and 2,3-DIPN (bottom)

isopropylation reactions include pore structure, pore size, and surface acid characteristics of the catalysts, type of alkylating agents and reaction conditions. The variables that contribute to the changes in the catalysts and surface interactions include degree of dealumination, method of A1 removal (acid, calcination temperature etc.), nature of alkylating agent, type and amount of additives, and catalystheactants ratio, in addition to temperature, pressure, and molar ratio of reactants. The results in Figure 15 clearly show that there exist optimum conditions. The complex effects of mordenite treatments are probably due to the changes, that remain to be clarified, in the distribution, location, density and strength of the acid sites in mordenites that are active for

314

Catalysis

the alkylation reactions. Recently a review by Sugi and K ~ b o t aand ~ ~more report^^^-^^ on alkylation of naphthalene have been published. Among the many possible DIPN and IPN isomers, the selective production of 2,6- and 2,7-DIPN is probably controlled by the transition-state shape selectivity, and by the product-shape ~ e l e c t i v i t y .The ~ ~ ?enhanced ~~ selective formation of 2,6-DIPN against 2,7-DIPN by proper dealumination treatment may be attributed to decreased acidity, a slight shrinkage of unit cell dimension and spatially-restricted electronic transition-state ~ e l e c t i v i t y . ~ ~ . ~ ~ Mordenite is not the only type of zeolite that can be used for naphthalene isopropylation. It has been reported that 0-zeolite can also be used for isopropylation of n a ~ h t h a l e n e .The ~ ~ performance of different zeolites for naphthalene alkylation with isopropanol at 200 "C depend on their structure. Chu and Chen70 observed that the selectivity to 2,6-DIPN decreased in the order of H-mordenite > H-Beta > USY > H-ZSM-5. Earlier study has shown that for naphthalene alkylation at 250°C using isopropanol over mordenite and Y zeolite catalysts, H-M catalysts (Si02/ A1203ratio = 17-35) display high selectivity to 2,6-DIPN and high 2,6-DIPN/ 2,7-DIPN ratio (up to 3), whereas H-Y (Si02/A1203r a t i o d ) shows high activity but no selectivity to 2,6-DIPN. Higher 2,6-DIPN/2,7-DIPN ratio (up to 3) was achieved with isopropanol than with propylene as the alkylating agent over the same catalyst under comparable ~ o n d i t i o n s . ' ~The ~ ~ type ~ 7 ~of ~ solvent (decalin and mesitylene) and the isopropanolhaphthalene ratio also affect both the conversion and the selectivity of H-M catalysts at 250"C.24 Recent studies by Colon and co-workers7' revealed the occurrence of transalkylation during liquid-phase isopropylation of naphthalene with isopropanol over a HY zeolite (Si02/A1203ratio=4) in the liquid phase with decalin as solvent at higher temperature, 350 "C. In a related study, He et aZ.75conducted liquid-phase alkylation of naphthalene with isopropanol over zeolite H-beta and observed some unexpected compounds which were confirmed to be cyclized products from naphthalene derivatives. Moreau and c o - w ~ r k e r sexamined ~ ~ . ~ ~ the effects of the size of alkyl groups for naphthalene alkylation over H-Y zeolites at 200 "C using isopropyl bromide and cyclohexyl bromide. The change from isopropyl to cyclohexyl bromide increased the 0,P-selectivity to 2,6- and 2,7-isomers, although the ratios of 2,6-isomer/2,7-isomer were close to 1 in both isopropylation and cyclohexylation over H-Y zeolites as well as in the case of isopropylation over H-M.35975These results are interesting but in contrast with those from other laboratories in that 2,6-DIPN and 2,7-DIPN were formed in nearly the same yields with isopropyl bromide as the alkylating agent, although many other laboratories have observed higher 2,6-DIPN/2,7-DIPN ratio for alkylation using either isopropanol or propylene over various H-mordenite catal y s t ~It is . possible ~ ~ ~ that~this~ difference ~ ~ ~was~ due ~ mainly ~ to the nature of the alkylating agent, and the by-product from alkyl bromide is an acid (HBr), which also affects the reaction that can decrease the selectivity to 2,6-DIPN against 2,7-DIPN, as suggested by a computational a n a l y ~ i s . ~ ~ ~ ~ ~ An interesting feature of cyclohexylation observed by Moreau and co-

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

315

worker^^^.^^ is that 2,6-dicyclohexylnaphthaleneis easily separated from the reaction mixture. The same group has also studied liquid-phase alkylation of naphthalene with t-butanol using HY and H-beta zeolite^.^^^^^ Over both series of zeolites, 2-(t-butyl)naphthalene (2-TBN) was observed as the only monoalkylated product. They found that t-butylation of naphthalene gives high selectivity to the 2,6-isomer under relatively mild reaction conditions, and the desired product (2,6-DTBN) can be easily separated from the reaction mixtures by cry~tallization.~~ Up to 84% 2,6-di(t-butyl)naphthalene (2,6-DTBN) selectivity with P,P-selectivity (2,6- + 2,7-) of 98-99% and 2,6-/2,7- ratios from 5.6 to 5.9 were obtained on silica-rich HY zeolites HY zeolites under mild reaction conditions (160 “C, 2h).78 The increase in selectivity with the increasing bulkiness can be understood based on the configurational diffusion limitation that was first proposed by Weisz.8 4

Considerations on Environmentally Friendly Synthesis of Chemicals over Zeolites

Zeolite catalysis is promising for environmentally benign synthesis of polycyclic chemicals. Homogeneous catalysis and Lewis acid catalysts are still widely used in synthesis of fine chemicals and pharmaceuticals. Table 26 shows ~ - ~ l is defined as the mass the E factor concept proposed by S h e l d ~ n , ~ which ratio of waste to desired product. The use of stoichiometric agents and inorganic salts and separation of catalysts and products as well as multi-step synthesis increase the amounts of waste by-products and decreases the efficiency, particularly in the synthesis of fine chemicals and pharmaceuticals, as shown by Sheldon’s E factor in Table 26.80 It is highly desirable to replace some of the high E-factor processes by ‘green’ or ‘benign’ catalytic processes using heterogeneous catalysts. Environmentally benign processing should become a priority for considering future routes for synthesis of organic chemicals including 2,6-DAN. Environmentally benign processing includes several aspects. First, the process should use more environmentally friendly feedstocks and reagents including reactants, catalysts, and solvents. Therefore, the choice of catalysts in the future may exclude those used in the process that are corrosive or involve formation of wasteful materials (including waste water) or by-products or formation and release of toxic substances. There are also recent reports on functionalization of alkylnaphthalene using HF/BF3 as shown in Scheme 1582and on alkylation of naphthalene using various corrosive acid catalysts, such as BF3-H3P0473 and AlC13.74It follows naturally that the selective alkylation over zeolite-based solid acid catalysts should be superior over those with corrosive acid catalysts such as A1Cl3, BF3, and BF3-H3P04.The advantage can be viewed in terms of the E-factor c ~ n c e p t . ~ ~ - ~ l Second, the process should be more efficient and less energy-intensive, and thus it will consume less energy and less resources that are non-renewable. Currently there is a commercial process for production of 2,6-NDCA related

316

Catalysis

Table 26

The E factor and the need for waste reduction in chemical processes (ref 80)

Industry segment

Product tonnage

kg-by-productlkg-product

Oil refining Bulk chemicals Fine chemicals Pharmaceuticals

106-108 104-106 102-1 04 101-1 0 3

c 0.1 1-5 5-50 25-100

mR+HF/BF3

CO + R-CH=CH2

R-CH2-C II

0

Scheme 15 HFIBFj catalysed functionalization of naphthalene

products based on a multi-step process for synthesizing 2,6-dimethylnaphthalene using xylene and b ~ t a d i e n e . ~It~may > ~ ~be* ~ desirable ~ to replace the processes employing corrosive catalysts such as HF/BF3 or multistep synthesis with more environmentally friendly and efficient processes. For example, shape-selective alkylation of naphthalene using zeolites would be more advantageous than the process shown in Scheme 15 for making the intermediates for 2,6-NDCA; oxidation of 2,6-DIPN to produce 2,6-NDCA has been tested at a pilot In this context, zeolite-catalysed selective alkylation may be superior over the multi-stage processes for synthesis of 2,6-DAN based on p-xylene or o-xylene that have been discussed in a recent review by Tanabe and H ~ e l d e r i c hThe . ~ ~future process for 2,6-DAN should be more energy-efficient and environmentally benign. Third, the production process should be as selective as possible so that the by-products can be minimized. In this context the minimization of the products that are also formed from shape-selective reaction such as 2,7-DAN and 2,3DAN present a challenge. Moreover, shape-selective methylation, though still very difficult, may be superior to shape-selective isopropylation, in terms of the oxidation reaction and the final product desired. This can be viewed in terms of atom economy, a concept that was first proposed by Trost,86 or the atom utilization efficiency defined by Sheldon" that can be calculated by dividing the molecular weight of the desired product by the sum of the molecular weights of all substances produced in the stoichiometric equation. Excellent reviews on environmentally benign catalytic processing and green chemistry have been published recently by Sheldon and Downing,81 by Armor,87 and Anastas and Williamson.88 In addition, it should be noted that disproportionation, and transalkylation of polyaromatics such as biphenyl and naphthalene with alkylbenzene, can also be used to obtain 4,4'-dialkylbiphenyl and 2,6-dialkylnaphthalene, respectively. Disproportionation and isomerization of alkylbenzenes have

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

317

been studied exten~ively,~*' but such reactions are less well known for alkylated polyaromatics. Kikuchi and ~ o - w o r k e r examined s ~ ~ ~ ~ ~disproportionation of methylnaphthalene. They observed higher selectivity of H-mordenite for formation of 2,6-DMN against 2,7-DMN, and this selectivity appeared to increase with increasing SiO2/Al203ratio of m o r d e n i t e ~2,7-DMN .~~ and 2,6DMN are in different groups of DMNs, and can not be isomerized into each other; 2,6-, 1,6- and 1,5-DMNs are in one group among which the isomers are interconvertible, but 2,7-, 1,7- and 1,s-DMNs are in another group.90 Interestingly, Pu and Inui9' recently reported that 2,7-DMN can be isomerized into 2,6-DMN over HZSM-5 and a reduction in crystallite size and decrease in external acid sites improve the catalytic activity and shape selectivity. It was suggested that the selectivity for 2,6-DMN versus 2,7-DMN is favored by stronger acid sites during methylene di~proportionation.~~ On the other hand, Komatsu et aL41 reported that in methylation of 2-methylnaphthalene with methanol, the selectivity to 2,6-DMN increased with decreasing acid strength of MFI catalyst, and that weaker acid sites prefer to produce 2,6-DMN over 2,7-DMN. In addition, Neuber et aZ.92reported an IR study related to isomerization and disproportionation of methylnaphthalene over Y, ZSM-5, and ZSM-12. Takeuchi and co-workers reported on transalkylation of biphenyl with triethylbenzene over Y - ~ e o l i t e .Recently, ~~.~~ synthesis of 4,4'-dimethylbiphenylbiphenyl has been achieved by shape-selective methylation of biphenyl with methanol under mild conditions using modified ZSM-5 type catalyst^^^-^^ or using silylated ZSM-5 cataly~t,~' which is different from earlier work using zeolites.99 Alternative reactions for synthesizing dimethylbiphenyl isomers includes coupling of toluene in strong acidslm or coupling of chlorotoluene.lol Shape-selective alkylation of polyaromatics and isomerization as well as transalkylation of alkylated polyaromatics using environmentally friendly solid acid catalysts warrant future research and development. 1~13917

5

Conclusions

Selective catalytic conversion of PAHs can be very useful for making polycyclic specialty chemicals that have potential industrial applications. Because of the presence of two or more isomers of polycyclic compounds, the challenge is to direct the reaction along a very specific pathway to the desired isomer. Substantial progress has been made in applied fundamental research in the past several years. The recent progress in selective catalysis has made it possible to synthesize some specialty polycyclic chemicals through one of the following reactions: (A) ring-shift isomerization of phenanthrene derivatives to anthracene derivatives; (B) shape-selective alkylation of naphthalene for synthesizing 2,6-dialkylnaphthalene; (C) shape-selective alkylation of biphenyl for synthesizing 4,4'dialkylbiphenyl; (D) conformational isomerization of cis-decahydronaphthalene into trans-decahydronaphthalene;(E) shape-selective hydrogenation of

318

Catalysis

naphthalene into either cis- or trans-decalin, and (F) regio-selective hydrogenation of 1-naphthol and quinoline for synthesizing specialty fuel additives. The effects of dealumination of mordenite on the structural and acidic characteristics and on the shape selectivity and activity were examined by physicochemical analysis, TPD, solid-state 27Aland 29SiMAS NMR, XRD, as well as catalytic alkylation reactions. Dealumination removes octahedral A1 species as well as tetrahedral A1 species, decreases the unit cell dimensions and reduces the number of strong acid sites in mordenites. Dealumination by the method discussed here can improve selectivity to 2,6-DIPN from 33 to 61% and significantly increases the 2,6/2,7 ratio. Improved selectivity to 2,6-DIPN upon proper dealumination was attributed to decrease in mordenite acidity, reduction in unit cell dimension and removal of some strong acid sites from external particle surface. However, neither the change in selectivity nor that in activity is a simple function of dealumination degree or Si02/A1203 ratio. Minor differences in the apparent framework Si02/Al2O3 ratio can result in major differences in activity or selectivity. There exist optimum conditions for dealumination as well as optimum reaction conditions for achieving higher selectivity to 2,6-DIPN. In addition to experimental research, computational analysis has also begun to shed light on selective conversion of polycyclic compounds. On the basis of computational and experimental results, a new concept called ‘spatially restricted electronic transition-state selectivity’ has been proposed, which can rationalize the observed enhancement in selectivity towards 2,6-DAN against 2,7-DAN in response to the modifications of catalysts and changes in reaction conditions. In most cases, further improvement in catalyst selectivity and stability is needed, and many fundamental questions concerning mechanistic aspects remain to be answered by future study. 6

Acknowledgments

The author is very grateful to Prof. Harold H. Schobert of PSU for his encouragement, support and many helpful discussions, and to Prof. Paul B. Weisz of PSU, and the late Dr. Werner 0. Haag of Mobil for their encouragement and helpful discussions. The author is very grateful to his former and current co-workers, particularly Drs. Andrew D. Schmitz, Wei-Chuan Lai, Shawn D. Lin, K. Madhusudan Reddy and Jian-Ping Shen for many helpful discussions. The author also wishes to thank Dr. Michael Ford of Air Products & Chemicals Inc. and Prof. John W. Larsen of Lehigh University for helpful discussions on the possible mechanism of sym-OHP ring-shift isomerization, and Dr. Juan M. Garces of Dow Chemical Company and Prof. Yoshihiro Sugi of Gifu University for general discussions on PAH conversion over zeolites. Various portions of our research were supported through funding or donations of special samples from the US Department of EnergyNational Energy Technology Laboratory, US Air Forcemright Laboratories, Air Products and Chemicals Inc., and PQ Co.

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons

7 1

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

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

319

References Special Report. Technology Vision2020: The U.S. Chemical Industry, Washington DC, 1996, prepared by American Chemical Society (ACS), American Institute of Chemical Engineers (AIChE), Chemical Manufacturers Association (CMA), Council for Chemical Research (CCR), and Synthetic Organic Chemical Manufacturers Association (SOCMA), 9 1 pp. Office of Technology Policy, U.S. Department of Commerce, Chemtech, 1996,26, 46. I.E. Maxwell, Cattech, 1997, 1, 5. (a) P.B. Weisz and V.J. Frilette, J. Phys. Chem., 1960, 64, 382; (b) P.B. Weisz, V.J. Frilette, R.W. Maatman and E.B. Mower, J. Catal., 1962, 1,307. (a) P.B. Venuto, L.A. Hamilton, P.S. Landis and J.J. Wise. J. Catal., 1966,4, 81; (b) P.B. Venuto, L.A. Hamilton and P.S. Landis. J. Catal., 1966,4,484. P.B. Venuto and P.S. Landis, Adv. Catal., 1968,18,259. N.Y. Chen and W.E. Garwood, Catal. Rev. - Sci. Eng., 1986,28, 185. P.B. Weisz, Pure Appl. Chem., 1980,52,2091. P.B. Venuto, Micropor. Mater., 1994,2,297. S.M. Csicsery, Stud Surf:Sci. Catal., 1995,94, 1. W.O. Haag, Stud Surf: Sci. Catal., 1994,84, 1375. J. Weitkamp, U. WeiD, and S. Ernst, Stud SurJ Sci. Catal., 1995,94,363. N.Y. Chen, W.E. Garwood, and F.G. Dwyer, Shape-Selective Catalysis in Industrial Applications. 2nd Edition, Marcel Dekker, New York, 1996,282 pp. J.E. Naber, K.P. de Jong,, W.H.J. Stork, H.P.C.E. Kuipers and M.F.M. Post, Stud. Surf:Sci. Catal., 1994,84,2197. M.E. Davis, Acc. Chem. Res., 1993,26, 111; S.L. Suib, A. C.S. Symp. Ser., 1993,517, 1. Shape-Selective Catalysis. Chemicals Synthesis and Hydrocarbon Processing, ed. C. Song, J.M. Garces, and Y. Sugi, ACS Symposium Series No. 738, American Chemical Society, Washington DC, 1999,432 pp. C. Song and H.H. Schobert, Fuel Process. Technol., 1993,34,157. (a) C. Song and H.H. Schobert, Chem. Znd., 1996,7, 253; (b) C. Song and H.H. Schobert, Fuel, 1996,75, 724. Y. Sugi and M. Toba, Catal. Today, 1994,19, 187. C. Song and K. Moffatt, Micropor. Mater., 1994,2,459. C. Song, US Patent, No. 5 523 505, 1996. W.-C. Lai, C. Song, A. van Duin and J.W. de Leeuw, Catal. Today, 1996, 31, 145. C. Song and S. Kirby, Micropor. Mater., 1994,2,467. A. Schmitz and C. Song, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1994,39,986. A. Schmitz and C. Song, Catal. Today, 1996,31,19. A. Schmitz and C. Song, Catal. Lett., 1996,31, 59. C. Song and A. Schmitz, Extended Abstracts, Program, I I th Znternat. Congress on Catalysis, Baltimore, Maryland, June 30-July 5, 1996, Po-164. A. Schmitz and C. Song, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1995,40,918. C. Song and A. Schmitz, 15th North Am. Catal. SOC.Meet., Technical Program, Chicago, May 18-22, 1997, p. 159. W.-C. Lai and C. Song, Catal. Today, 1996, 1996,31, 171. A. Schmitz, G. Bowers and C. Song, Catal. Today, 1996,31,45. C. Song and A. Schmitz, Energy Fuels, 1997,11,656.

320 34 35 36

37 38

39

40 41 42 43

44 45 46 47 48 49 50

51 52 53 54 55 56 57 58

59 60 61

Catalysis

A. Katayama, M. Toba, G. Takeuchi, F. Mizukami, S. Niwa and S. Mitamura, J. Chem. SOC.Chem. Commun., 1991,39. P. Moreau, A. Finiels, P. Geneste and P. Solofo, J. Catal., 1992,136,487. J.D. Fellmann, J. Saxton, P.R. Wentrcek, E.G. Derouane and P. Massioni, US Patent, No. 5026942, 1991. J.A. Horsley, J.D. Fellmann, E.G. Derouane and C.M. Freeman, J. Catal., 1994, 147,231. J.-H. Kim, Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, X. Tu and M. Matsumoto, Micropor. Mater., 1995,5, 113. J.-H. Kim, Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, X. Tu, M. Matsumoto, S. Nakata, A. Kato, G. Seo and C. Pak, Appl. Catal. A: Gen., 1995,131, 15. D. Fraenkel, M. Cherniavsky, B. Ittah and M. Levy, J. Catal., 1986,101,273. T. Komatsu, Y. Aoki, S. Namba and T. Yashima, Stud. SurJ Sci. Catal., 1994, 84, 1821. G.S. Lee, J.J. Maj, S.C. Rocke and J.M. Garces, Catal. Lett. , 1989,2,243. Y. Sugi, T. Matsuzaki, T. Hanaoka, H. Takeuchi, H. Arakawa, T. Tokono and G. Takeuchi, Shokubai, 1989,31,373. Y. Sugi, T. Matsuzaki, T. Hanaoka, T. Takeuchi, T. Tokono and G. Takeuchi, Stud. Surf: Sci. Catal., 1991,60, 303. Y. Sugi, X.-L. Tu, T. Matsuzaki, T. Hanaoka, Y. Kubota, J.-H. Kim, M. Matsumoto, K. Nakajima and A. Igarashi, Catal. Today, 1996,31, 3. Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, J.-H. Kim, X.-L. Tu and M. Matsumoto, Catal. Lett., 1994,26, 181. Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, J.-H. Kim, X.-L. Tu and M. Matsumoto, Catal. Lett., 1994,27, 315. T. Matsuda, T. Urata, U. Saito and E. Kikuchi, Appl. Catal. A: Gen., 1995, 131, 215. T. Matsuda, T. Kimura, E. Herawati, C. Kobayashi and E. Kikuchi, Appl. Catal. A: Gen., 1996, 136, 19. W.-C. Lai, C. Song, H.H. Schobert and R. Arumugam, Am. Chem. SOC.Div. Fuel Chem. Prepr., 1992,37, 1671. T. Edwards, W.E. Harrison and H.H. Schobert, 33rd AIAAIASMEIASMEISAE Joint Propulsion Conference and Exhibit, July 6-9, 1997, Seattle, WAYPaper no. AIAA 97-2848. C. Song, S. Eser, H.H. Schobert and P.G. Hatcher, Energy Fuels, 1993,7,234. S.D. Lin and C. Song, Catal. Today, 1996,31,93. K.M. Reddy and C. Song, Catal. Today, 1996,31,137. K.M. Reddy and C. Song, Mater. Res. SOC.Sum. Proc. Ser., 1997,454, 125. (a) C. Song, W.-C. Lai and H.H. Schobert, Ind. Eng. Chem. Res., 1994, 33, 548; (b) E. Yoon, L. Selvaraj, C. Song, J. Stallman and M.M. Coleman, Energy Fuels, 1996,10, 806. (a) J. Shao and C. Song, Catal. Today, 2001, 65, 59; (b) J. Shao, G. Holtzer and C. Song, Am. Chem. SOC.Div. Petroleum Chem. Prepr. ,2000,45, 18. Y. Sugi and Y. Kubota, Catalysis, Vol. 13, ed. J.J. Spivey, Royal Society of Chemistry, Cambridge, 1997, p. 55. C. Song, Stud. Surf Sci. Catal., 1998,113, 163. G.A. Olah and J.A. Olah, J. Am. Chem. SOC.,1976,98,1839. (a) J. Weitkamp and M. Neuber, Stud. Surf: Sci. Catal., 1991, 60, 291; (b) M. Niwa, M. Endo and Y. Murakami, Res. Chem. Intermed., 1995,21,127.

8: Recent Advances in Selective Conversion of Polycyclic Hydrocarbons 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101

32 1

T. Inui and S.-B. Pu, Appl. Catal. A: Gen., 1996,146,305. K. Tanabe and W.F. Hoelderich, Appl. Catal. A: Gen., 1999,181, 399. Y. Sugi, K. Nakajima, S. Tawada, J.-H. Kim, T. Hanaoka, T. Matsuzaki, Y. Kubota and K. Kunimori, Stud. Surf: Sci. Catal., 1999, 125, 359. C.N. Satterfield, Heterogeneous Catalysis in Industrial Practice, McGraw-Hill, New York, 1991, p. 232. C. Song, X. Ma, A.D. Schmitz and H.H. Schobert, Appl. Catal. A: Gen., 1999, 182, 175. C. Song, X. Ma and H.H. Schobert, Am. Chem. SOC.Sym. Ser., 1999,738,305. J.-P. Shen, J. Ma, D.-Z. Jiang and E.-Z. Min, Chem. Res. Chin. Univ., 1993, 14, 845. S.-J. Chu and Y.W. Chen, Ind. Eng. Chem. Res., 1994,33, 31 12. S.-J. ChuandY.W. Chen, Appl. Catal. A: Gen., 1995,123, 51. G. Colon, I. Ferino, E. Rombi, E. Selli, L. Forni, P. Magnoux and M. Guisnet, Appl. Catal. A: Gen., 1998,168,81. R. Brzozowski and W. Tecza, Appl. Catal. A: Gen., 1998,166,21. G. Calik and T.H. Ozdamar, Chem. Eng. J. Biochem. Eng. J., 1994,53, 173. A.A. Khalaf, M.S.I.T. Makki and R.A. Kabli, Rev. Roumaine Chim., 1995, 40, 67. P. Moreau, A. Finiels, P. Geneste, J. Joffre, F. Moreau and J. Solofo, Catal. Today, 1996,31, 1 1 . C.Q. He, Z.M. Liu, F. Fajula and P. Moreau, Chem. Commun., 1998,18,1999. Z.M. Liu, P. Moreau and F . Fajula, Chem. Commun., 1996,23,2653. Z.M. Liu, P. Moreau and F . Fajula, Appl. Catal. A: Gen., 1997,159, 305. R.A. Sheldon, Chemtech, 1994,24,38. R.A. Sheldon, Chem. Ind., 1997, January 6, 12. R.A. Sheldon and R.S. Downing, Appl. Catal. A: Gen., 1999, 189, 163. News-NDCA, Petrotech (Japan), 1989,12,513. News-PEN, Hydrocarbon Process., 1993,5,33. News-DMNDC, Modern Plastics, 1994, May, 69. News-DIPN, Chem. Eng., 1993,11,17. B.M. Trost, Science, 1991,254, 1471. J. Armor, Appl. Catal. A: Gen., 1999, 189, 153. P. Anastas and T. Williamson, Green Chemistry, Theory and Practice, Oxford University Press, Oxford, 1998. T. Matsuda, K. Yogo, Y. Mogi and E. Kikuchi, Chem. Lett., 1990,1085. E. Kikuchi, Y. Mogi and T. Matsuda, Collect. Czech. Chem. Commun., 1992, 57, 909. S.-B. Pu and T. Inui, Zeolites, 1996,17, 334. M . Neuber, H.G. Karge and J. Weitkamp. Catal. Today, 1988,3,11-22. G. Takeuchi, Y. Shimoura and T. Hara. Appl. Catal. A: Gen., 1996,137,87. G. Takeuchi and Y . Shimoura, Catal. Surveys Jpn., 1998,2,77. J.-P. Shen, L. Sun and C. Song, Catal. Lett., 2000,65, 147. J.-P. Shen, L. Sun and C. Song, Stud. Surf: Sci. Catal., 2000,130,3023. J.P. Shen, L. Sun and C. Song, Stud. Surf: Sci. Catal., 2001, 135, 151. S. Dubuis, R. Doepper and A. Renken, Stud. Surf: Sci. Catal., 1999, 122,359. C. Brechtelsbauer and G. Emig, Appl. Catal. A: Gen., 1997, 161, 79. B.Q. Xu, D. Sood, A.V. Iretskii and M.G. White, J. Catal., 1999,187, 358. D.M. Roberge and W.F. Hoelderich, Appl. Catal. A: Gen., 2000,194,341.