摘要:

Contents 3663 3669 3675 3683 3693 370 1 3709 3717 3725 3737 Normal and Abnormal Electron Spin Resonance Spectra of Low-spin Cobalt(r1) IN,]-Macrocyclic Complexes. A Means of Breaking the Co-C Bond in B12 Co-enzyme M. Green, J. Daniels and L. M. Engelhardt The Interaction between Superoxide Dismutase and Doxorubicin. An Electron Spin Resonance Approach V. Malatesta, F. Morazzoni, L. Pellicciari-Bollini and R. Scotti Biomolecular Dynamics and Electron Spin Resonance Spectra of Copper Complexes of Antitumour Agents in Solution. Part 2.-Rifamycins R. Basosi, R. Pogni, E. Tiezzi, W. E. Antholine and L. C. Moscinsky An Electron Spin Resonance Investigation of the Nature of the Complexes formed between Copper(I1) and Glycylhistidine D. B. McPhail and B. A. Goodman A Vibronic Coupling Approach for the Interpretation of the g-Value Temperature Dependence in Type-I Copper Proteins M.Bacci and S. Cannistr aro The Electron Spin Resonance Spectrum of Al[C,H,] in Hydrocarbon Matrices J. A. Howard, B. Mile, J. S. Tse and H. Morris N; and (CN); Spin-Lattice Relaxation in KCN Crystals H. J. Kalinowski and L. C. Scavarda do Carmo Single-crystal Proton ENDOR of the SO, Centre in y-Irradiated Sulphamic Acid N. M. Atherton, C. Oliva, E. J. Oliver and D. M. Wylie Single-crystal Electron Spin Resonance Studies on Radiation-produced Species in Ice 1,. Part 1.-The 0- Radicals Single-crystal Electron Spin Resonance Studies on Radiation-produced Species in Ice I,. Part 2.-The HO, Radicals J. Bednarek and A. Plonka J. Bednarek and A. PlonkaContents 3663 3669 3675 3683 3693 370 1 3709 3717 3725 3737 Normal and Abnormal Electron Spin Resonance Spectra of Low-spin Cobalt(r1) IN,]-Macrocyclic Complexes.A Means of Breaking the Co-C Bond in B12 Co-enzyme M. Green, J. Daniels and L. M. Engelhardt The Interaction between Superoxide Dismutase and Doxorubicin. An Electron Spin Resonance Approach V. Malatesta, F. Morazzoni, L. Pellicciari-Bollini and R. Scotti Biomolecular Dynamics and Electron Spin Resonance Spectra of Copper Complexes of Antitumour Agents in Solution. Part 2.-Rifamycins R. Basosi, R. Pogni, E. Tiezzi, W. E. Antholine and L. C. Moscinsky An Electron Spin Resonance Investigation of the Nature of the Complexes formed between Copper(I1) and Glycylhistidine D. B. McPhail and B. A. Goodman A Vibronic Coupling Approach for the Interpretation of the g-Value Temperature Dependence in Type-I Copper Proteins M. Bacci and S. Cannistr aro The Electron Spin Resonance Spectrum of Al[C,H,] in Hydrocarbon Matrices J. A. Howard, B. Mile, J. S. Tse and H. Morris N; and (CN); Spin-Lattice Relaxation in KCN Crystals H. J. Kalinowski and L. C. Scavarda do Carmo Single-crystal Proton ENDOR of the SO, Centre in y-Irradiated Sulphamic Acid N. M. Atherton, C. Oliva, E. J. Oliver and D. M. Wylie Single-crystal Electron Spin Resonance Studies on Radiation-produced Species in Ice 1,. Part 1.-The 0- Radicals Single-crystal Electron Spin Resonance Studies on Radiation-produced Species in Ice I,. Part 2.-The HO, Radicals J. Bednarek and A. Plonka J. Bednarek and A. Plonka

ISSN:0300-9599    

DOI:10.1039/F198783FX025

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

Electrochemistry Group Workshop on Electrochemical Techniques and Instruments To be held at the University of Warwick on 6-7 January 1988 Further information from Dr P. N. Bartlett, Department of Chemistry, University of Warwick, Coventry CV4 7AL Surface Reactivity and Catalysis Group with the Process Technology Group and the Institute of Chemical Engineers Opportunities for Innovation in the Application of Catalysis To be held at Queen Mary College, London on 6-7 January 1988 Further information from Professor J. Pritchard, Queen Mary College, London Division with the Institute of Mathematics and its Applications Mathematical Modelling of Semiconductor Devices and Processes To be held at the University of Loughborough on 7-8 January 1988 Further information from the Institute of Mathematics, Maitland House, Warrior Square, Southend-on-Sea SS1 2JY Division London Symposium: Modern Electrochemical Systems To be held at Imperial College, London on 12 January 1988 Further information from Mrs Y.A. Fish, Royal Society of Chemistry, Burlington House, London W1V OBN Polymer Physics Group with the 3Ps Group Plastics, Packaging and Printing To be held at the Institute of Physics, 47 Belgrave Square, London on 18 February 1988 Further information from Dr M. Richardson, National Physical Laboratory, Teddington, Middlesex l w 1 1 OLW Theoretical Chemistry Group Postgraduate Students’ Meeting To be held at University College, London on 2 March 1988 Further information from Dr G. Doggett, Department of Chemistry, University of York, York Colloid and Interface Science Group with The Society of Chemical Industry and British Radio frequency Spectroscopy Group Spectroscopy in Colloid Science To be held at the University of Bristol on 5-7 April 1988 Further information from Dr R. Buscall, ICI Corporate Colloid Science Group, PO Box 11, The Heath, Runcorn WA7 40E Annual Congress: Division with Electrochemistry Group Solid State Materials in Electrochemistry To be held at the University of Kent, Canterbury on 12-15 April 1988 Further information from Dr J.F. Gibson, Royal Society of Chemistry, Burlington House, London W1V OBN Electrochemistry Group with The Society of Chemical Industry Electrolytic Bubbles To be held at Imperial College, London on 31 May 1988 Further information from Professor W. J. Albery, Department of Chemistry, Imperial College of Science and Technology, South Kensington, London SW7 2AY Electrochemistry Group with The Society of Chemical Industry Chlorine Symposium To be held at the Tara Hotel, London on 1-3 June 1988 Further information from Dr S.P. Tyfield, Central Electricity Generating Board, Berkeley Nuclear Laboratories, Berkeley, Gloucestershire GLI 3 9BP Gas Kinetics Group Xth International Symposium on Gas Kinetics To be held at University College, Swansea on 24-29 July 1988 Further information from Dr G. Hancock, Physical Chemistry Laboratory, South Parks Road, Oxford OX1 302 (xiii)Electrochemistry Group Workshop on Electrochemical Techniques and Instruments To be held at the University of Warwick on 6-7 January 1988 Further information from Dr P.N. Bartlett, Department of Chemistry, University of Warwick, Coventry CV4 7AL Surface Reactivity and Catalysis Group with the Process Technology Group and the Institute of Chemical Engineers Opportunities for Innovation in the Application of Catalysis To be held at Queen Mary College, London on 6-7 January 1988 Further information from Professor J. Pritchard, Queen Mary College, London Division with the Institute of Mathematics and its Applications Mathematical Modelling of Semiconductor Devices and Processes To be held at the University of Loughborough on 7-8 January 1988 Further information from the Institute of Mathematics, Maitland House, Warrior Square, Southend-on-Sea SS1 2JY Division London Symposium: Modern Electrochemical Systems To be held at Imperial College, London on 12 January 1988 Further information from Mrs Y.A. Fish, Royal Society of Chemistry, Burlington House, London W1V OBN Polymer Physics Group with the 3Ps Group Plastics, Packaging and Printing To be held at the Institute of Physics, 47 Belgrave Square, London on 18 February 1988 Further information from Dr M. Richardson, National Physical Laboratory, Teddington, Middlesex l w 1 1 OLW Theoretical Chemistry Group Postgraduate Students’ Meeting To be held at University College, London on 2 March 1988 Further information from Dr G. Doggett, Department of Chemistry, University of York, York Colloid and Interface Science Group with The Society of Chemical Industry and British Radio frequency Spectroscopy Group Spectroscopy in Colloid Science To be held at the University of Bristol on 5-7 April 1988 Further information from Dr R.Buscall, ICI Corporate Colloid Science Group, PO Box 11, The Heath, Runcorn WA7 40E Annual Congress: Division with Electrochemistry Group Solid State Materials in Electrochemistry To be held at the University of Kent, Canterbury on 12-15 April 1988 Further information from Dr J. F. Gibson, Royal Society of Chemistry, Burlington House, London W1V OBN Electrochemistry Group with The Society of Chemical Industry Electrolytic Bubbles To be held at Imperial College, London on 31 May 1988 Further information from Professor W. J. Albery, Department of Chemistry, Imperial College of Science and Technology, South Kensington, London SW7 2AY Electrochemistry Group with The Society of Chemical Industry Chlorine Symposium To be held at the Tara Hotel, London on 1-3 June 1988 Further information from Dr S. P. Tyfield, Central Electricity Generating Board, Berkeley Nuclear Laboratories, Berkeley, Gloucestershire GLI 3 9BP Gas Kinetics Group Xth International Symposium on Gas Kinetics To be held at University College, Swansea on 24-29 July 1988 Further information from Dr G. Hancock, Physical Chemistry Laboratory, South Parks Road, Oxford OX1 302 (xiii)

ISSN:0300-9599    

DOI:10.1039/F198783BX027

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

A SYMPOSIUM ON Promotion in Heterogeneous Catalysis 23rd, 24th and 25th September 1986 A Symposium on Promotion in Heterogeneous Catalysis was held at the University of Bath on 23rd, 24th and 25th September, 1986, as part of the 1986 Autumn Meeting of the Royal Society of Chemistry. The President of the Faraday Division, Professor N. Sheppard, F.R.S., was in the chair: about 140 Members of the Faraday Division and visitors from overseas attended the meeting. Among the overseas visitors were: Dr A. G. Andersen, Norway Mr S. Astegger, Austria Prof. B. G. Baker, Australia Mr A. G. T. M. Bastein, The Netherlands Prof. A. T. Bell, U.S.A. Dr S. Bernal, Spain Miss Boulinguiez, France Dr L. Bruce, Australia Dr L. Carlsen, Denmark Prof. J. K. A. Clarke, Ireland Dr R. E. Copperthwaite, South Africa Mr H.Daigo, Japan Dr J. P. Damon, France Mr C. Dossi, Italy Prof. J-C. Duchet, France Prof. M. Folman, Israel Mr L. Furong, Norway Dr S. Gobolos, Hungary Dr A. R. Gonzalez-Elipe, Spain Prof. J. Grimbolot, France Dr G. L. Haller, U.S.A. Miss C. Harendt, Federal Republic of Mr K. Hatakeyama, Japan Prof. R. Haul, Federal Republic of Mr B. F. Hegarty, Ireland Mr J-P. Hindermann, France Miss M. J. Holgado, Spain Mr L. Hongyuan, The Netherlands Dr G. J. Hutchings, South Africa Mr M. Iwata, Japan Dr L. Jalowiecki, France Dr C. S. John, The Netherlands Dr A. Kampf, Israel Mr K. Kasano, Japan Germany Germany The organising committee comprised Prof. F. S. Stone (Chairman) Dr R. Burch Mrs Y. A. Fish Dr J. T. Klug, Israel Prof. H. Knozinger, Federal Republic of Mr T.Kondoh, Japan Dr M. J. Ledoux, France Dr J. A. Lercher, Austria Prof. D. R. Lloyd, Ireland Miss M-J. Luys, The Netherlands Dr P. Meriaudeau, France Dr K. Moller, Sweden Dr B. D. Murray, U.S.A. Mr S. Nakamura, Japan Dr J. A. Odriozola, Spain Prof. P. Pietropaolo, Italy Dr E. K. Poels, The Netherlands Dr. J. Pohl, Federal Republic of Germany Prof. V. Ponec, The Netherlands Prof. R. Prins, U.S.A. Mr P. B. Rasmussen, Denmark Prof. V. Rives, Spain Dr F. Roessler, Switzerland Mr H. Sato, Japan Prof. J. J. F. Scholten, The Netherlands Dr L. Schuster, Federal Republic of Mr T. Shimokawa, Japan Prof. F. Solymosi, Hungary Dr M. Sprague, Federal Republic of Mr N. Suyama, Japan Dr H. Topsse, Denmark Prof. F. Trifiro, Italy Prof. R. A. van Santen, The Netherlands Mr K.Yokomizo, Japan Mr D. L. Yun, The Netherlands Germany Germany Germany Dr R. W. Joyner Prof. J. Pritchard Dr D. A. Young 63 FAR 1A SYMPOSIUM ON Promotion in Heterogeneous Catalysis 23rd, 24th and 25th September 1986 A Symposium on Promotion in Heterogeneous Catalysis was held at the University of Bath on 23rd, 24th and 25th September, 1986, as part of the 1986 Autumn Meeting of the Royal Society of Chemistry. The President of the Faraday Division, Professor N. Sheppard, F.R.S., was in the chair: about 140 Members of the Faraday Division and visitors from overseas attended the meeting. Among the overseas visitors were: Dr A. G. Andersen, Norway Mr S. Astegger, Austria Prof. B. G. Baker, Australia Mr A. G. T. M. Bastein, The Netherlands Prof. A. T. Bell, U.S.A.Dr S. Bernal, Spain Miss Boulinguiez, France Dr L. Bruce, Australia Dr L. Carlsen, Denmark Prof. J. K. A. Clarke, Ireland Dr R. E. Copperthwaite, South Africa Mr H. Daigo, Japan Dr J. P. Damon, France Mr C. Dossi, Italy Prof. J-C. Duchet, France Prof. M. Folman, Israel Mr L. Furong, Norway Dr S. Gobolos, Hungary Dr A. R. Gonzalez-Elipe, Spain Prof. J. Grimbolot, France Dr G. L. Haller, U.S.A. Miss C. Harendt, Federal Republic of Mr K. Hatakeyama, Japan Prof. R. Haul, Federal Republic of Mr B. F. Hegarty, Ireland Mr J-P. Hindermann, France Miss M. J. Holgado, Spain Mr L. Hongyuan, The Netherlands Dr G. J. Hutchings, South Africa Mr M. Iwata, Japan Dr L. Jalowiecki, France Dr C. S. John, The Netherlands Dr A. Kampf, Israel Mr K. Kasano, Japan Germany Germany The organising committee comprised Prof.F. S. Stone (Chairman) Dr R. Burch Mrs Y. A. Fish Dr J. T. Klug, Israel Prof. H. Knozinger, Federal Republic of Mr T. Kondoh, Japan Dr M. J. Ledoux, France Dr J. A. Lercher, Austria Prof. D. R. Lloyd, Ireland Miss M-J. Luys, The Netherlands Dr P. Meriaudeau, France Dr K. Moller, Sweden Dr B. D. Murray, U.S.A. Mr S. Nakamura, Japan Dr J. A. Odriozola, Spain Prof. P. Pietropaolo, Italy Dr E. K. Poels, The Netherlands Dr. J. Pohl, Federal Republic of Germany Prof. V. Ponec, The Netherlands Prof. R. Prins, U.S.A. Mr P. B. Rasmussen, Denmark Prof. V. Rives, Spain Dr F. Roessler, Switzerland Mr H. Sato, Japan Prof. J. J. F. Scholten, The Netherlands Dr L. Schuster, Federal Republic of Mr T. Shimokawa, Japan Prof. F. Solymosi, Hungary Dr M.Sprague, Federal Republic of Mr N. Suyama, Japan Dr H. Topsse, Denmark Prof. F. Trifiro, Italy Prof. R. A. van Santen, The Netherlands Mr K. Yokomizo, Japan Mr D. L. Yun, The Netherlands Germany Germany Germany Dr R. W. Joyner Prof. J. Pritchard Dr D. A. Young 63 FAR 1ISSN 0300-9238 JCFTAR 83(7) 1893-2260 (1 987) 1893 1915 1935 1945 1963 1967 1975 200 1 201 5 203 5 2047 206 1 207 1 209 1 2103 2113 21 19 2129 2145 2157 2169 JOURNAL OF THE CHEMICAL SOCIETY Faraday Transactions I Physical Chemistry in Condensed Phases CONTENTS Introductory Lecture : Promotion in Heterogeneous Catalysis : Retrospect and Prospect S. J. Thomson Coordination of Carbon Monoxide to Transition-metal Surfaces R. A. van Santen Electrostatic Interactions and their Role in Coadsorption Phenomena S.Holloway, J. K. Nsrskov and N. D. Lang Catalytic Implications of Local Electronic Interactions between Carbon Monoxide and Coadsorbed Promoters on Nickel surfaces J. M. MacLaren, D. D. Vvedensky, J. B. Pendry and R. W. Joyner General Discussion Adsorption and Reaction on Strained-metal Overlayers D. W. Goodman and C. H. F. Peden A Comparison of the Effects of Cu and Au on the Surface Reactivity of Ru(0001) B. Sakakini, A. J. Swift, J. C. Vickerman, C. Harendt and K. Christmann Chemical and Electrostatic Interactions between Alkali Metals and Carbon Monoxide on Metal Single-crystal Surfaces D. Lackey and D. A. King Adsorption and Dissociation of CO, on a Potassium-promoted Rh( 1 1 1) Surface Alkali Metal, Chlorine and other Promoters in the Silver-catalysed Selective Oxidation of Ethylene R.B. Grant, C. A. J. Harbach, R. M. Lambert and S. A. Tan The Promotion of Surface-catalysed Reactions by Gaseous Additives. The Role of a Surface Oxygen Transient The Effects of Titania and Alumina Overlayers on the Hydrogenation of CO over Rhodium General Discussion Comparison of Hydrogenation and Hydrogenolysis on Unsupported and Silica- supported Rh-V,O, and Pt-V,O, Manipulation of the Selectivity of Rhodium by the Use of Supports and Promoters G. van der Lee, A. G. T. M. Bastein, J. van den Boogert, B. Schuller, H. Luo and V. Ponec Promotion of Platinum-based Catalysts for Methanol Synthesis from Syngas P. Mkriaudeau, K. Albano and C. Naccache Ethanol Promotion by the Addition of Cerium to Rhodium-Silica Catalysts A.Kiennemann, R. Breault, J-P. Hindermann and M. Laurin General Discussion The Role of Co in Sulphidised Co-Mo Hydrodesulphurisation Catalysts supported on Carbon and Alumina J. P. R. Vissers, V. €3. J. de Beer and R. Prins Inorganic Cluster Compounds as Models for the Structure of Active Sites in Promoted Hydrodesulphurization Catalysts H. Topsae, B. S. Clausen, N-Y. Topsse, E. Pedersen, W. Niemann, A. Miiller, H. Bogge and B. Lengeler General Discussion F. Solymosi and L. Bugyi C-T. Au and M. W. Roberts M. E. Levin, M. Salmeron, A. T. Bell and G. A. Somorjai Y-J. Lin, D. E. Resasco and G. L. Haller 63-22 175 Promotion of Nitrogen and Hydrogen Chemisorption and Ammonia Synthesis on Alumina-supported Hexagonal Tungsten Bronze K, WO, S. Stevenson and P. A. Sermon Promotion of Methanol Synthesis and the Water-gas Shift Reactions by Adsorbed Oxygen on Supported Copper Catalysts G. C. Chinchen, M. S. Spencer, K. C. Waugh and D. A. Whan The Promoting Role of Cr and K in Catalysts for High-pressure and High- temperature Methanol and Higher-alcohol Synthesis A. Riva, F. Trifiro, A. Vaccari, G. Busca, L. Mintchev, D. Sanfilippo and W. Manzatti 2227 Propan-2-01 Adsorption and Decomposition on Zinc Oxide promoted by Alkali Metal 2243 General Discussion 2259 Index of Names 2260 List of Posters 2193 2213 D. Chadwick and P. J. R. O’Malley

ISSN:0300-9599    

DOI:10.1039/F198783FP085

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF THE CHEMICAL SOCIETY 1065 1073 1083 1105 1109 1113 1121 1143 1157 1171 1177 1189 1205 1215 1229 1235 1253 Faraday Transactions II, lssue7,1987 Molecular and Chemical Physics For the benefit of readers of Faraday Transactions I, the contents list of Faraday Transactions 11, Issue 7 , is reproduced below. This issue contains papers given at a meeting of the Polar Solids Discussion Group on ‘Fundamentals of Uranium Oxide and other Actinide Oxides ’ held at Mansfield College, Oxford, in March 1986. Recent Problems and Progress in the Study of UO, and Mixed U0,-PuO, C. R. A. Catlow Crystallographic Studies of Anion-excess Uranium Oxides High-temperature Studies of UO, and Tho, using Neutron Scattering Techniques M. T. Hutchings A High-resolution Neutron Scattering Investigation of the Crystal Field Splittings of UO, R.Osborn, B. C. Boland, Z. A. Bowden, A. D. Taylor, M. A. Hackett, W. Hayes and M. T. Hutchings Inelastic Neutron Scattering Investigation of the Lattice Dynamics of Th0,and CeO, K. Clausen, W. Hayes, J. E. Macdonald, R. Osborne, P. G. Schnabel, M. T. Hutchings and A. Magerl A Neutron Powder Diffraction Study of U,-,Ce,O,-, A. D. Murray, C. R. A. Catlow and B. E. F. Fender Atomic Transport Properties in UO, and Mixed Oxides (U, P U ) ~ , H. Matzke Thermodynamic and Transport Studies of Mixed Oxides. The Ce0,-UO, System T. G. Stratton and H. L. Tuller Oxygen Diffusion in UO,, Tho, and PuO,. A Review G.E.Murch and C. R. A. Catlow Point-defect Calculations on UO, R. A. Jackson, C. R. A. Catlow and A. D. Murray Defects and Clusters in UO, and (U, P U ) ~ , Electronic Structure and Ground-state Properties of the Actinide Oxides P.J. Kelly and M. S. S. Brooks Recent Spectroscopic Studies of UO, An Investigation of Defect Structures in Single-crystal UO,,, by Optical Absorption Spectroscopy Infrared Dielectric Response of UO, Single-crystal Surfaces investigated by High-resolution Electron Energy-loss Spectroscopy P. A. Thiry, J-J. Pireaux, R. Caudano, J. R. Naecjele, J. Rebizant and J-C. Spirlet Equation of State of Uranium Oxide R. W. Ohse Specific Heat of UO,, Tho,, PuO, and the Mixed Oxides (Th,U,-,)O, and (Puo~2Uo~8)Ol~97 by Enthalpy Data Analysis J. Ralph B. T. M. Willis J. H. Harding J. Schoenes H. V. St A. Hubbard and T. R. Griffiths 11263 1273 The Potential Use of Uranium Oxides and Uranium-Bismuth Mixed Oxides in Catalysis H.Collette, V. Deremince-Mathieu, Z. Gabelica, J. B. Nagy, E. G. Derouane and J. J. Verbist Technological Problems and the Future of Research on the Basic Properties of Actinide Oxides J. R. Matthews 61 1059 61 1060 611421 611751 612214 612498 71008 71084 71129 7/ 140 71160 7/ 167 71261 71374 71485 71500 The following papers were accepted for publication in Faraday Transactions I during April 1987. Transport Phenomena in Concentrated Aqueous Solutions of Sodium-Caesium Polystyrene Sulphonates H. R. Corti Thermodynamic Properties of Concentrated Heteroionic Polyelectrolyte Solutions H. R. Corti The Second. Planar Virial Coefficient for Nitrogen, Oxygen and Carbon Monoxide adsorbed on Graphite Doping Effect of Sodium on y-Ray Irradiated NgO T .Matsuda, K. Yamada, Y. Shibata, H. Miura and K. Sugiyama Activity and Selectivity in Toluene Oxidation on Well Characterized Vanadium Oxide Catalysts K. Mori, A. Miyamoto and Y. Murakami A New Analytical Solution of the Quaternary Gibbs-Duhem Equation Z-C. Wang Empirical Analysis on the Constituent Terms of Transfer Enthalpies. Quaternary Ammonium Ions in Acetonitrile-Methanol Mixtures Y. Kondo, R. Uematsu, Y. Nakamura and S. Kusabayashi The Non-SMSI Behaviour of a new Ti0,-attached Rh Metal Catalyst and its Characterizations K. Asakura, Y. Iwasawa and H. Kuroda One-electron Reduction of 2,2’-Bipyrimidine in Aqueous Solution K. Barqawi, Ta-a1 Akashen, B. J. Parsons and P. C. Beaumont The Ferroelectric and Liquid-crystalline Properties of Some Chiral Alkyl 4-n Alkanoyloxybiphenyl-4’-Carboxylates J.W. Goddby, E. Chin, J. M. Geary, J. S. Pate1 and P. L. Finn Electrokinetices of Polyelectrolyte Solutions in Capillary Tubes H. Vink Temperature Dependence of Adiabatic Compressibility of Aqueous Solutions of Alkyltrimethylammonium Bromides R. Zielinski, S. Ikeda, H. Nomura and S. Kato Surface Structure and Reactivity of Vanadium Oxide Supported on TiO, : V,O,/TiO, (Rutile) Catalysts prepared by Hydrolysis F. Cavani, G. Centi, E. Forseti, F. Tifiro and G. Busca Anomalous Trapping of Carbon Monoxide by Carbon Dioxide on Magnesium Oxide M. Kobayashi, T. Kanno and Y. Konishi Electron Spin Resonance and Electron Spin Echo Modulation Studies of 5- Doxylstearic Acid and N,N,N’,N’-Tetramethylbenzidine Photoionization in Sodium Dodecylsuphate Micelles : Effects of 15-Crown-5 and 18-Crown-6 Ether Addition Rotating-disc Electrode Voltammetry.Wave Shape Analysis for DISPZ and EC, Processes L. Lajtar and S. Sokolowski P. Baglioni and L. Kevan R. G. Compton, D. Mason and P. R. Unwin 117/501 7/509 7/519 7/639 7/640 7/64 1 7/642 7/66 1 7/662 71663 71664 71665 71666 71727 71728 71729 7/730 The Reduction of Fluorescein in Aqueous Solution (at pH 6): A New DISP2 Reaction Absorption, MCD and MCPL Studies of Ru(bpy)r and Complexes with a Di(ethoxycarbony1)-substituted Bipyridine Ligand as a Probe of Rigid Environments E. Krausz Alkenyl Carbocation Formation from Propene in H-ZSM-5 I. Kiricsi and H. Forster Interpretation of Electron Spin-echo Modulation for Cupric Ion Complexes in Crystalline Powders Single Crystal Proton ENDOR of the SO, Centre in y-Irradiated Sulphamic Acid E.P.R.of Layered Chromium Silver Chalcogenides R. E. Benfield, P. P. Edwards and A. M. Stacy An Electron Spin Resonance Study of Isomerism in a Co(0) Acetylene 71- Complex The E.P.R. Spectra of CpCr(CO), and CpCr(CO),P(C,H,), in Single Crystals of their Mn Analogues J. R. Morton, K. F. Preston, N. S. Cooley, M. C. Baird, P. J. Krusic and S. J. McLain Binding of Paramagnetic Ions in Ionomers. Cu2+ and Ti+, in Nafion Membranes S. Schlick and M. G. Alonso-Amigo E.P.R. and Electronic Spectroscopy of Low-spin Manganese(I1) Complexes (C,R,) (CO), (L)Mn with L-Hydrazido( 1-), Arylamido, Anionic Nitrile and Purine-type Ligands Single-Crystal E.S.R.Studies of Radiation-produced Species in Ice I,. Part 1.-The 0-Radicals Studies of the Manganese Site of Photosystem I1 by Electron Paramagnetic Resonance Spectroscopy J. C. de Paula, W. F. Beck, A. Miller, R. B. Wilson and G. W. Brudvig Electron Capture by FeIII and FeO, Centres in Haemoglobin, and the Absence of Subsequent Electron Transfer from FeO, Centres to Fe"'. An Electron Spin Resonance Study MAGRES: A General Program for E.P.R., ENDOR and ESEEM. C. P. Keijers, E. J. Reijerse, P. Stam, M. F. Dumont and M. C. M. Gribnau An Electron Spin Resonance Study of complexes of Oxovanadium(1v) with Simple Dicarboxylic Acids An Electron Spin Resonance Investigation of the Nature of the Complexes Formed between Copper(I1) and Glycylhistidine D. B. McPhail and B.A. Goodman Oxovanadium( ~ v ) Complexes of Cysteine Characterized by Electron Spin Resonance Spectroscopy R. G. Compton, D. Mason and P. R. Unwin M. W. Anderson and L. Kevan N. M. Atherton, C. Oliva, E. J. Oliver and D. M. Wylie J. A. DeGray, Q. Meng and P. H. Rieger R. Gross and W. Kaim J. Bednarek and A. Plonka M. C. R. Symons and F. A. Taiwo D. B. McPhail and B. A. Goodman D. B. McPhail and B. A. Goodman ... 111Cumulative Author Index 1987 Agnel, J-P. L., 225 Akalay, I., 1137 Akitt, J. W., 1725 Albano, K., 21 13 Alberti, A., 91 Allen, G. C., 925, 1355 Andersen, A., 2140 Anderson, A. B., 463 Anderson, J. B. F., 913 Antholine, W. E., 151 Ardizzone, S., 1159 Atherton, N. M., 37, 941 Aun Tan, S., 2035 Avnir, D., 1685 Axelsen, V., 107 Baker, B. G., 2136 Baldini, G., 1609 Balon, M., 1029 Barratt, M.D., 135 Barrer, R. M., 779 Basosi, R., 151 Bastein, A. G. T. M., 2103, Bastl, Z., 51 1 Bateman, J. B., 841 Battesti, C. M., 225 Baussart, H., 171 1 Becker, K. A., 535 Bell, A. T., 2061, 2086, 2087, Bennett, J.E., 1805 Berclaz, T., 401 Berleur, F., 177 Berroa de Ponce, H., 1569 Berry, F. J., 615 Bertagnolli, H., 687 Berthelot, I., 231 Beyer, H. K., 511 Bianconi, A., 289 Bjorklund, R. B., 1507 Blandamer, M. J., 559, 865, Blyth, G., 751 Boerio-Goates, J., 1553 Bogge, H., 2157 Bond, G. C., 1963, 2071, 2088, 2129, 2130,2133,2138,2140 Borbely, G., 511 Boucher, E. A., 1269 Brandreth, B. J., 1835 Braquet, P., 177 Brazdil, J. F., 463 Breault, R., 21 19 Briscoe, B. I., 938 Bruce, J. M., 85 Au, C-T., 2047 2129 2088 1783 Brustolon, M., 69 Budil, D.E., 13 Bugyi, L., 2015 Bulow, M., 1843 Burch, R., 913, 2087, 2130, 2134, 2135, 2141, 2250 Burgess, J., 559, 865, 1783 Burggraaf, A. J., 1485 Burke, L. D., 299 Busca, G., 853, 1591, 2213 Buscall, R., 873 Cairns, J. A., 913 Carley, A. F., 351 Caro, J., 1843 Cassidy, J. F., 231 Celalyan-Berthier, A., 401 Chadwick, D., 2227, 2249, 2255 Chalker, P. R., 351 Chandra, H., 759 Chieux, P., 687 Chinchen, G. C., 2193 Chittofrati, A., 1 159 Christmann, K., 1975 Clark, B., 865 Clausen, B. S., 2157 Clifford, A. A., 751 Colin, A. C., 819 Coller, B. A. W., 645, 657 Coluccia, S., 477 Compostizo, A., 819 Compton, R. G., 1261 Conway, B. E., 1063 Corvaja, C., 57 Couillard, C., 125 Courbon, H., 697 Craven, J. B., 779 Crossland, W. A., 37 D’Alba, F., 267 Danil de Namor, A.F., 1569 Dash, A. C., 1307 Daverio, D., 705 Davies, M. J., 1347 Davoli, I., 289 Dawber, J. G., 771 de Beer, V. H. J., 2145 De Doncker, J., 125 De Laet, M., 125 De Ranter, C. J., 257 Declerck, P. J., 257 Delafosse, D., 1137 Delahanty, J. N., 135 Delobel, R., 171 1 Despeyroux, B. M., 2081, 2139, 2171, 2243, 2256 Di Lorenzo, S., 267 Chu, D-Y., 635 Diaz Peiia, M., 819 DimitrijeviC, N. M., 1193 Dodd, N. J. F., 85 Duarte, M. A., 2133 Ducret, F., 141 Dudikova, L., 51 1 Dusaucy, A-C., 125 Eicke, H-F., 1621 Elbing, E., 657, 645 Elders, J. M., 1725 Empis, J. M. A., 43 Endoh, A., 41 1 Engberts, J. B. F. N., 865 Evans, J. C., 43, 135 Fahim, R. B., 1601 Fan, G., 323 Fatome, M., 177 Fejes, P., 1109 Fletcher, P. D. I., 985, 1493 Flint, N. J., 167 Formaro, L., 1159 Formosinho, S. J., 431 Forrester, A.R., 21 1 Forster, H., 1109 Fraissard, J., 451 Freude, D., 1843 Freund, E., 1417 Fricke, R., 1041 Fujii, K., 675 Fujitsu, H., 1427 Galli, P., 853 Gampp, H., 1719 Garbowski, E., 1469 Garrido, J., 1081 Garrone, E., 1237 Gellings, P. J., 1485 Geoffroy, M., 401 Gervasini, A., 705 Gilbert, B. C., 77 Gilbert, R. G., 1449 Goates, J. R., 1553 Goates, S. R., 1553 Goffredi, M., 1437 Golding, P. D., 1203 Goodman, D. W., 1963, 1967, 2071, 2072, 2073, 2075, 2082, 2086, 2251, 1967 Gottschalk, F., 571 Gozzi, D., 289 Grampp, G., 161 Grant, R. B., 2035 Gratzel, M., 1101 Grauer, G. L., 1685 Grauer, Z., 1685 Gray, P., 751 Greci, L., 69AUTHOR INDEX Grieser, F., 591 Grigorian, K. R., 1189 Grimblot, J., 2170 Grossi, L., 77 Groves, G. S., 1119, 1281 Grzybkowski, W., 281, 1253 Guardado, P., 559 Guilleux, M-F., 11 37 Hada, H., 1559 Hagele, G., 1055 Hakin, A.W., 559, 865, 1783 Halawani, K. H., 1281 Hall, D. G., 967 Hall, M. V. M., 571 Haller, G. L., 1965, 2072, 2080, 2089, 2091, 2129, 2131, 2132, 2133, 2135, 2136, 2137, 2138, 2243 Halpern, A., 219 Hamada, K., 527 Harbach, C. A. J., 2035 Harendt, C., 1975 Harland, R. G., 1261 Harrer, W., 161 Harris, R. K., 1055 Hartland, G. V., 591 Hasegawa, A., 759 Hatayama, F., 675 Haul, R., 2083 Hayashi, K., 1795 Heatley, F., 517 Hemminga, M. A., 203 Henriksson, U., 1515 Herold, B. J., 43 Hertz, H. G., 687 Hidalgo, J., 1029 Higgins, J. S., 939 Hilfiker, R., 1621 Hinderman, J. P., 21 19, 2142, Hindermann, J. P., 21 19 Holden, J. G., 615 Holloway, S., 1935 Howe, A. M., 985, 1007 Howe, R.F., 813 Hudson, A., 91 Hunger, M., 1843 Hunter, R., 571 Hussein, F. H., 1631 Hutchings, G. J., 571 Ikeyama, N., 1427 Imamura, H., 743 Imanaka, T., 665 Ismail, H. M., 1601 Ito, T., 451 Iwaki, T., 943, 957 Iwamoto, E., 1641 Jackson, S. D., 905, 1835 Jaenicke, W., 161 Janes, R., 383 Joyner, R. W., 1945, 1965, 2074, 2085, 2138, 2250 Juszczyk, W., 1293 Kakuta, N., 1227 Kaneko, M., 1539 2143 Kanno, T., 721 Karger, J., 1843 Kariv-Miller, E., 1169 Karpinski, Z., 1293 Kato, C., 1851 Kawaguchi, T., 1579 Kazusaka, A., 1227 Kerr, C W., 85 Kiennemann, A., 2 1 19 King, D. A., 1966, 2001, 2079, 2080, 2081 Kira, A., 1539 Kiricsi, I., 1109 Kitaguchi, K., 1395 Kiwi, J., 1101 Klein, J., 1703 Knozinger, H., 2088, 2171 Kobayashi, J., 1395 Kobayashi, M., 721 Koda, S., 527 Kondo, Y., 1089 Konishi, Y., 721 Koopmans, H.J. A., 1485 Kordulis, C., 627 Korf, S. J., 1485 Korth, H-G., 95 Koutsoukos, P. G., 1477 Kowalak, S., 535 Kubelkova, L., 51 1 Kubokawa, Y., 675, 1761 Kumamaru, T., 1641 Kuroda, K., 1851 Kusabayashi, S., 1089 Kuzuya, M., 1579 La Ginestra, A,, 853 Lackey, D., 2001 Lajtar, L., 1405 Lambelet, P., 141 Lambert, R. M., 1963, 1964, 2035, 2082, 2083, 2084 Lamotte, J., 1417 Lang, N. D., 1935 Laschi, F., 1731 Laurin, M., 21 19 Lavagnino, S., 477 Lavalley, J-C., 1417 Lawin, P. B., 1169 Lawrence, S., 1347 Le Bras, M., 1711 Leaist, D. G., 829 Lecomte, C., 177 Lee, E. F. T., 1531 Lengeler, B., 2157 Lercher, J. A., 2080, 2255 Leroy, J-M., 171 1 Letellier, P., 1725 Levin, M. E., 2061 Lin, C. P., 13 Lin, Y-J., 2091 Linares-Solano, A., 108 1 Lindgren, M., 893, 1815 Lippens, B.C., Jr, 1485 Liu, R-L., 635 Liu, T., 1063 Loliger, J., 141 Lorenzelli, V., 853, 1591 Loretto, M. H., 615 Luckham, P. F., 1703 Lund, A., 893, 1815, 1869 Luo, H., 2103 Lycourghiotis, A., 627, 1179 Lynch, J., 1417 Lyons, C. J., 645 Lyons, M. E. G., 299 McAleer, J. F., 1323 McCarthy, S. J., 657 McDonald, J. A., 1007 Machin, W. D., 1203 MacLaren, J. M., 1945, 1965, McLauchlan, K. A., 29 Maestre, A., 1029 Maezawa, A., 665 Makela, R., 51 Manfredi, M., 1609 Maniero, A. L., 69, 57 Manzatti, W., 2213 Marchese, L., 477 Marcus, Y., 339 Mari, C. M., 705 Markarian, S. A., 1189 Martin Luengo, M. A., 1347, Martin-Martinez, J. M., 1081 Mashkovsky, A. A., 1879 Masiakowski, J. T., 893, 1869 Masliyah, J. H., 547 Matralis, H., 1179 Matsuura, H., 789 Maxwell, I.A., 1449 Mehandru, S. P., 463 Meriaudeau, P., 21 13, 2140 Merwin, L. H., 1055 Micic, 0. I., 1127 Mintchev, L., 2213 Miyahara, K., 1227 Miyata, H., 675, 1761, 1851 Mochida, I., 1427 Molina-Sabio, M., 1081 Monk, C. B., 425 Montagne, X., 1417 Morazzoni, F., 705 Morimoto, T., 943, 957 Moseley, P. T., 1323 Moyes, R. B., 905 Mozzanega, M-N., 697 Muller, A., 2157 Muiioz, M. A., 1029 Nabiullin, A. A., 1879 Naccache, C., 21 13 Naga6, M., 1739 Nagaoka, T., 1823 Nair, V., 487 Nakai, S., 1579 Nakajima, T., 13 15 Napper, D. H., 1449 Narayanan, S., 733 Narducci, D., 705 Nayak, R. C., 1307 Nazer, A. F. M., 1119 Nedeljkovic, J. M., 1127 1651AUTHOR INDEX Nenadovic, M. T., 1127 Niccolai, N., 1731 Niemann, W., 2157 Nishida, S., 1795 Nomura, H., 527 Nomura, M., 1227, 1779 Norris, J.0. W., 1323 Norris, J. R., 13 N~rskov, J . K., 1935 Nukui, K., 743 Nuttall, S., 559 O’Brien, A. B., 371 Odinokov, S. E., 1879 Ogura, K., 1823 Ohno, M., 1559 Ohno, T., 675 Ohshima, K., 789 Okabayashi, H., 789 Okamoto, Y.,. 665 Okuda, T., 1579 Okuhara, T., 1213 O’Malley, P. J. R., 2227 Ono, T., 675, 1761 Otsuka, K., 13 15 Ott, J. B., 1553 Pallas, N. R., 585 Parry, D. J., 77 Patrono, P., 853 Peden, C. H. F., 1967 Pedersen, E., 2157 Pedersen, J. A., 107 Pedulli, G. F., 91 Penar, J., 1405 Pendry, J. B., 1945 Perez-Tejeda, P., 1029 Pethica, B. A., 585 Pethrick, R. A., 938 Pichat, P., 697 Pielaszek, J., 1293 Pilarczyk, M., 281 Pizzini, S., 705 Poels, E. K., 2140 Pogni, R., 151 Pomonis, P., 627 Pomonis, P. J., 1363 Ponec, V., 1964, 1965,2071, 2072, 2074, 2083, 2103, 2136, 2138, 2139, 2244,2251 Primet, M., 1469 Prins, R., 2087, 2136, 2137, 2145, 2169, 2170, 2172 Priolisi O., 57 Pritchard, J., 1963, 2085, 2249 Prugnola, A., 1731 Puchalska, D., 1253 hrushotham, V., 21 1 Radulovic, S., 559 Raffi, J.J., 225 Rajaram, R. R., 2130 Ramaraj, R., 1539 Ramis, G., 1591 Rees, L. V. C., 1531, 1843 Resasco, D. E., 2091 Reyes, P. N., 1347 Richards, I). G., 2138 kchter-Mendau, J., 1843 Riley, B. W., 2140, 2253 Ritschl, F., 1041 Riva, A., 2213 Riviere, J. C., 351 Roberts, M. W., 351, 2047, 2084, 2085, 2086, 2248 Robinson, €3. H., 985, 1007 Rodriguez-Reinoso, F., 1081 Rollins, K., 1347 Roman, V., 177 Romgo, M. J., 43 Rooney, J. J., 2077, 2080, 2086, Rosseinsky, D. R., 231, 245 Rossi, C., 1731 Rowlands, C. C., 43, 135 Rubio, R.G., 819 Rudham, R., 1631 Sakai, T., 743, 1823 Sakakini, B., 1975 Salmeron, M., 2061 Sanchez, M., 1029 Sanfilippo, D., 22 13 Sangster, D. F., 657 Saraby-Reintjes, A., 271 Sato, T., 1559 Saucy, F., 141 Savoy, M-C., 141 Sayed, M. B., 1149, 1751, 1771 Scholten, J. J. F., 1966, 2073, Schuller, B., 2103 Seebode, J., 1109 Segal, M. G., 371 Segre, U., 69 Sermon, P. A., 1369, 1651, 1667, 2175, 2243, 2256, 1347 Seyedmonir, S., 813 Sheppard, N., 1966, 2075 Sidahmed, I. M., 439 Simonian, L. K., 1189 Smith, D. H., 1381 Soderman, O., 1515 Sokolowski, S., 1405 Solymosi, F., 2015, 2074, 2078, 2081, 2082, 2086, 2137, 2142, 2247 2089 2246, 2256, 2258 Somorjai, G. A., 2061 Spencer, M. S., 2193, 2245, Steenken, S., 113 Stevens, D. G., 29 Stevenson, S., 2175 Stone, F. S., 1237, 2080, 2084, Suda, Y., 1739 Sugahara, Y., 1851 Suppan, P., 495 Sustmann, R., 95 Suzuki, T., 1213 SvetliEiC, V., 1169 Swartz, H.M., 191 Swift, A. J., 1975 Symons, M. C. R., 1, 383, 759 2246, 2247, 2248, 2249, 2250 2254 Szostak, R., 487 Tabner, B. J., 167 Taga, K., 789 Takaishi, T., 41 1 Tan, W. K., 645 Tanaka, H., 1395 Tanaka, K., 1213, 1779, 1859 Tanaka, K-i., 1859 Tempere, J.-F., 1137 Tempest, P. A., 925 Theocharis, C. R., 1601 Thiery, C. L., 225 Thomas, T. L., 487 Thomson, S. J., 1893, 1964, 1965, 2083 Thurai, M., 841 Tilquin, B., 125 Tomellini, M., 289 Tonge, J. S., 231, 245 Toprakcioglu, C., 1703 Topsse, H., 2157, 2169, 2171 Topsore, N-Y., 2157 Torregrosa, R., 108 1 Toyoshima, I., 1213 Trabalzini, L., 151 Trifiro, F., 2213, 2246, 2251, 2254, 2255 Tsuchiya, S., 743 Tsuiki, H., 1395 Tsukamoto, K., 789 Turner, J. C. R., 937 Tyler, J. W., 925, 1355 Ueno, A., 1395 Ukisu, Y., 1227 Uma, K., 733 Unwin, P. R., 1261 Vaccari, A., 2213 Vachon, A., 177 van de Ven, T. G. M., 547 van den Boogert, J., 2103 van der Lee, G., 2103 van Santen, R. A, 1915, 1963, Varani, G., 1609 Vattis, D., 1179 Vickerman, J . C., 1975,2075 Vincent, P. B., 225 Vink, H., 801, 941 Vissers, J. P. R., 2145 Vong, M. S. W., 1369, 1667 Vordonis, L., 627 Vuolle, M., 51 Vvedensky, D. D.; 1945 Waddicor, J. I., 751 Waller, A. M., 1261 Waters, D. N., 1601 Waugh, K. C., 2193 Wells, C . F., 439, 939, 1 1 19, Wells, P. B., 905 Whan, D. A., 2193 White, L. R., 591, 873 Whyman, R., 905 Williams, D. E., 1323 Williams, 3. O., 323 1964, 2077, 2140, 2251 1281AUTHOR INDEX Williams, R. J. P., 1885 Wilson, H. R., 1885 Wilson, I. R., 645, 657 Winstanley, D., 1835 Wbjcik, D., 1253 Wune, A. T., 1651 Williams, W. J., 371 Xyla, A. G., 1477 Yamada, K., 743 Yamamoto, Y., 1641, 1795 Yarnasaki, S., 1 6 4 1 Yanagihara, Y., 1579 Yanai, Y., 1 6 4 1 Yanv, S., 1685 Yonezawa, Y., 1559 Yoshino, T., 1823 Yun, D. L., 2251 Zaki, M. I., 1601 Zhang, Q., 635 (vii)(vii)

ISSN:0300-9599    

DOI:10.1039/F198783BP089

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chern. SOC., Faraday Trans. I , 1987,83, 1893-1914 (Faraday Symposium 21) Introductory Lecture Promotion in Heterogeneous Catalysis : Retrospect and Prospect Samuel J. Thomson Chemistry Department, University of Glasgow, Glasgow GI2 8QQ The aim of this introductory paper is to survey current views on the mode of action of promoters in heterogeneous catalysis. Examples of current beliefs on mechanism of promotion have been taken from recent literature and these include support effects, reactant self-promotion, site creation, electronic and geometric effects, activation energies, reactant availability, orders of reaction, synergism, compound formation, bimetallics, particular reaction steps affected by promotion, redox effects and surface attachment of molecules to promoters. Several promising developments in theoretical studies of promotion, the use of Cheveral phases, single-crystal studies, molecular probes, microengineering and ion implantation point the way towards future developments and understanding. Seventy years ago in his book on catalysis Joblingl has a section on promotion: he says ‘it has recently been discovered that there are other substances which, when added in minute quantity to a catalyst, increase its activity. Thus practically all metallic catalysts become activated when certain oxides or compounds or other metals are distributed throughout them.In Haber’s synthetic ammonia manufacture, for instance, the iron, platinum, osmium or uranium employed as a catalyst is quickened to an enormous extent by the presence of traces of salts of other metals or, with certain exceptions, of the other metals themselves.Further research in connection with these promoters, as they are called, would be amply repaid.’ Scientific study of promoters2 began about 1920 and the advances since then would have delighted the great eighteenth century philosopher Hume. He recognised the distinction between observables - cause and effect - and the reasons for the e f f e ~ t . ~ The reasons are in the mind, and human imagination creates a diversity of reasons, conjecture and speculation. This conference will be about the observables and the controversies, for promotion has advanced on both fronts. The causes and effects have multiplied: more important has been the vigour of the argument and controversy, the life force of the subject.These have inspired elegant experiments and clever detective work in aid of reason and understanding. This lecture reflects the variety of promotional effects and offers examples of the reasons for these effects. The examples have been taken mainly from the 1985-86 literature. The word promotion does not occur frequently in the titles of papers on catalysis. Yet the majority involve promotion or modification of solids. The survey could only be made by reading papers regardless of their titles. It was obvious that the emphasis for promotion had shifted from acceleration to specificity, selectivity and lifetime of catalysts. 18931894 Introductory Lecture 5 1 I I 1 2 i 1 K (76) from KX or KOH Fig. 1. Effect of promoter concentration on metal exposure.(-)9-11% Ni on SO,:(- -)9-12% Ni on SiO2/Al20,. After Chai and Falc~ner.~ Relative not Absolute Promotion suggests improvement in process compared with some baseline condition. We should not think that the reference unpromoted reaction occurs on the perfectly clean surface of an unperturbed solid. Metal blacks are impure, metal films not ultra clean, metal particles may interact with supports. Single crystals may lack perfection, for LEED sees only that which is regular, and the electron spectroscopies must have limits on their sensitivities : there are multiple reports of foreign elements penetrating their surfaces to lie beneath them. Non-metallic catalysts similarly have deviations from perfection in both purity and stoichiometry.Thus promotion is a relative phenomenon. Academic and Industrial Promotion The discovery of an efficient promoter in laboratory experiments under mild conditions does not mean that it will meet the demanding conditions for an industrial catalyst. Aika et aZ.* have a good example in their work on NH, synthesis. Although a commercial doubly promoted Fe catalyst has an activity one tenth that of RuCsOH/A1,0, on a bulk-metal activity basis, this does not mean that Ru catalysts are better than Fe catalysts under commercial conditions. The reason is that the kinetics, which are functions of the pressures of the components, are not favourable for Ru, in that the order for hydrogen is negati~e;~ on iron it is positive.s Support as Promoter We have no reason to exclude the support from the concept of promotion.Solymosi has long been an exponent of the idea of electron transport to or from metal particles. In a recent paper' he and his co-workers have examined the variation of electric properties of TiO, as support on Rh particles in hydrogenation of CO and CO,. A further example occurs in the work of Morris et aZ.8 on Ru for the H, + CO reaction. The reaction, they say, is principally determined by the support for activity and selectivity.S. J . Thomson 1 1895 300 200 1 1 1 1 1 2 3 4 K promoter (%) Fig. 2. CO hydrogenation over (a) Ni/SiO, and (b) Ni on SiO,/Al,O,. After Chai and Falc~ner.~ Metal exposure and activity are a function of support and promoter. Chai and Falconer9 show curves giving percentage exposed metal us.concentration of promoter on two supports (fig. 1). There is a rise and fall in exposure and a variation between the supports. The curve shapes for activity for CO conversion illustrate a surprising feature of promotion. Fig. 2(a) shows an abrupt fall for Ni on SiO,, whereas fig. 2(b) shows a significant rise in activity then a fall. Simple electronic or geometric explanations are likely to be unsatisfactory. In support interactions the so-called SMSI has been the subject of many papers: a recent paper carries essential references to SMSI, its advantages and limitations.1° The perspective for SMSI can be found in a recent paper by Solymosi,ll who reminds us that it is essential from time to time to inspect the older literature. Support effects also exist for non-metallic systems.Mori et a1.12 give a good example with an additional clear statement on what is meant by increase in activity. Their work was on the oxidation of furan over V,O, catalysts in the following forms, unsupported, supported on TiO, or Al,O,. The rate of reaction was larger for the supported catalyst: this statement has to be analysed in a more incisive way for the results reveal that the unsupported catalyst has a higher activity per site than the supported catalysts. These latter have a larger number of sites but their specific activities per site are lower. Promotion in this case means more sites. A masterly account of supports and their effects, and the pitfalls in interpretation, has been written by Bond.', Molecular Interactions In promotion the possibility of molecule-molecule interaction between reactants should be raised.The example quoted is not catalytic, but the interactions are relevant for catalysis. Tsyganenko et al. l4 were interested in spectroscopic changes which occurred as CO was increasingly adsorbed on ZnO and other oxides. The downward shifts of 20-30 cm-l as O,, increased were ascribed to a chemical effect in which electrons on Zn2+ ions became less readily available when neighbouring sites were occupied by CO. They express the view that dynamic dipole coupling gives an upward shift as O,, increases, but that this effect is small compared with the chemical effect. The vco shift has also been calculated by Hush and Williams15, who found a downward shift, which was too small to account for the present results.Other evidence for intermolecular effects comes from changes of selectivity with pressure. Chuang et a1.16 found that over Rh/TiO, at 300 OC, 1 atm, with CO/H, = 2,1896 Introductory Lecture 98% (w/w) of product was hydrocarbon (C1+C2+), with some CH,CHO. At 10 atm 17% (w/w) of oxygenates was produced (MeOH, MeCHO, EtOH, acetone and EtOAc). The overall conclusion is therefore that molecule-molecule interaction does occur. In catalysts Roberts" and co-workers have become interested in such interactions. Mode of Action Several modes of promoter behaviour must be considered: these range from changes in area or exposure for supported metals, the effects of different levels of promoter concentration, the consistency of promotion effects for any one promoter, possible surface reconstructions and compound formation, electronic and geometric effects.The view i s often expressed that electronic interactions modify activation energies, whereas structural alterations leave Eapp unchanged but affect pre-exponential factors18. Thus a fundamental question is whether one or other, or both E and A , change on promotion: the answers vary. In the first example19 Au on Ru/SiO, reduced the Fischer-Tropsch (FT) activity of the catalyst. Au on Ru/MgO gave a maximum in activity then a decline. In both cases the major yields were of hydrocarbons, not oxygenates. The effects seem to be geometric, since Au affected the pre-exponential factor and not Eapp for methanation. CO adsorption on the bimetallic catalyst did not reveal infrared band shifts for spectra compared to those for Ru: evidently the Ru electronic structure was unchanged.Site density, and in particular its effect on hydrogen availability, has also been invoked by Morris et aLs to explain the activities in FT synthesis of Ru supported on SiO,, zeolite or TiO,. Ru on Cl-free magnesia was different. Here low activity resulted from an increase in activation energy for this support. Selectivity can change with promoter concentration which brings about changes in orders of reaction and activation energies. One might expect an electronic explanation. This turns out not to be so in the example which follows,2o for all changed in parallel and the explanation depended on site requirement. The paper illustrates not only these features but the necessary investigation of mechanism before the mode of action of a promoter can be understood.The catalytic oxidation of ethylene to the epoxide over Ag/a-Al,O, is promoted for selectivity by trace amounts of chlorinated hydrocarbons in the feed. Production of CO,+ H,O is inhibited. The financial dollar saving is ca. $ lOOM per year. This system has been investigated by Campbell and Koel using u.h.v., A.e.s., LEED, X.P.S. and t.d.s. on single crystals of Ag( 110) and Ag(ll1). They established the identity of rate and selectivity dependence on pressure and temperature for single-crystal and supported Ag. Thus single crystals were good model catalysts: there is an interesting and, as yet, unexplained difference between the single-crystal and supported-metal catalysts : on the former, rates per silver atom are 50 to 100 times higher than on the latter.It was established that chlorinated hydrocarbons behaved like atomic chlorine and that selectivity towards epoxide increased to 80 % with accompanying loss of activity. The major effect was between 8,, = 0.5 and O,, = 0.75. The question was whether this was an ensemble or an electronic effect, or both. The authors present the following model for reaction. 0, is adsorbed as O,(ads) with a charge of -2, and this is the precursor for O(ads). Epoxidation occurs through the reaction of ethylene and O,(ads) through an intermediate which may either produce ethylene oxide and an adsorbed O(ads) or the intennedi;,te can dissociate, where the fragments lead to formation of CO, and H,O.The O(ads) produced by the intermediate when epoxide is produced can also produce CO, and H,O. Thus even if the intermediate can be influenced to produce epoxide the remaining O(ads) will convert 1 in 7 molecules of ethylene to CO, and H,O, giving a theoretical maximum selectivity of ca. 80%. One might expect that since coverage by chlorine rapidly suppresses dissociativeS. J . Thomson 1897 0 chlorine coverage, 8,, 0.7 Fig. 3. Curve shapes for ethylene conversion to ethylene oxide (a) or CO,(b), and selectivity (c), over Ag( 1 10) as the surface is modified by Cl. After Campbell and Koel.,O oxygen adsorption such that it is negligible beyond 8 = 0.3, there would be an effect on selectivity. Very little effect was observed on rates or selectivity, indicating that molecular, not atomic, oxygen was the reactive species.Po, and activation energies for CO, + H,O production were measured as a function of OCl. Epoxidation occurred at 440-610 K in a ca. 100 Torr* reaction mixture giving 3% conversion. At 563 K turnover numbers for EtO and CO, fell from ca. 8.5 and 36, respectively, to ca. 0 as eCl rose from 0 to 0.7, whereas selectivity for EtO rose from 33 to 80% [SEtO = TONEto(TONEto TON,,,)], where TON indicates turnover number. Fig. 3 shows an example. After reaction the coverage regions for C1 were established as p(2 x I)-CI, e = 0.39 to e = 0.58 and 4 4 x 2 ) - ~ 1 e = 0.7. The reaction orders in ethylene pressure for EtO and CO, production both decreased as 8,, rose to 0.3, then they increased.Reaction orders in oxygen pressure for both EtO and CO, production followed the same trend in that both were steady for OCl = 0 to 0.15, then they rose at around 8 = 0.3 to a steady value. Activation energies for EtO and CO, production also changed in concert as 8,, increased. These results are surprising for changes in 8,, seemed to affect reaction orders and activation energies for EtO and C02. production in parallel. This leaves the significant question: what causes the change in selectivity? The authors postulate a common intermediate, formed from molecularly adsorbed oxygen and ethylene, which can break down to EtO or CO, + H,O. Since Cl on the surface decreases the heat of adsorption of O,(ads) and its coverage, increasing OCl should increase the order in pot as is observed.C1 increases the heat of adsorption of ethylene for 0 < 0 < 0.35, and the diminishing order in PEt is similarly explained. Starting at OCl = 0, however, there could be an increase in rate at low values of BCl, where ethylene adsorption and reaction are enhanced. This was observed in the results. Beyond 8,, 0.5 C1 decreases the heat of adsorption of ethylene, and the order in pEt increases. The parallel changes in activation energy and orders of reaction fit the idea of a common intermediate. This may be the key to understanding the selectivity. If the breakdown steps of the intermediate are considered then the different pathways to EtO or H,O + CO, have different site requirements. Formation of EtO does not involve bond * 1 Torr = 101 325/760 Pa.1898 Introductory Lecture breaking, whereas for CO, + H,O production five bonds per ethylene molecule must be broken.This probably requires a large ensemble of free Ag atoms. Increasing values for 9,, destroy these sites and lead to EtO production. Site Creation A quite different approach to understanding promoter action and also to improving it arises in the use of mixed catalysts of different phases. Such catalytic synergy between different phases in physical mixtures of MOO, and BiPO, has been established by Tascon et al.,l in the N-ethyl formamide dehydration 0 'H + H,O isonitrile nitrile provided molecular oxygen was present : activity and selectivity were high. A good example of hybrid catalysts which gives an indication of their potential is the Pd/SiO,-Y-zeolite system.Fujimoto et aL2, used such a catalyst in H, + CO conversion to increase the C, to C, alkane yield: Fischer-Tropsch catalysts have an inevitably high yield of CH,. C, yields on the hybrid catalyst were 5-16% (wt% in carbon base); C,, Compound formation is a possible explanation of some promotions: thus steam gasification of graphite to produce H,, CO and CO, at 800 K can be catalysed by alkali metal, which may be promoted by a transition-metal oxide.,, Carraza et al. suggest compound formation of K,NiO, and KFeO, as one possible explanation. Deliberate promotion through site creation manifests itself in the fluorine promotion of aluminas. An example is the fluorination by NH,F of y-A1,0,.24 Normally considered to have few Bronsted sites, these can be created as follows: 17-39% ; C,, 21-41 %.H giving rise to Lewis and Bronsted sites with strong acidity (pK d 1.5) at 2-4 g F per 100 g A1,0,. Such F-promoted y-Al,O, catalysts have been studied in cumene cracking by Boorman et al.25 Bond strength has been invoked to explain promotion in the PtRe(S) system, which has produced both a successful hydrocarbon conversion catalyst and a flood of papers on its mechanism of action: Shum et aLZ6 believe that, since PtRe catalysts behave differently from Pt + Re physically mixed catalysts, PtRe is an alloy system. They say that for a given size of ensemble of metal atoms, catalyst performance in cracking/hy- drogenolysis will improve as Re is substituted for Pt. This is because the Re-C bond is stronger than the Pt-C bond, and hence more than one metal carbon bond may be formed per molecule.The role of the S they see as one of inhibiting the reorganisation of hydrocarbonaceous fragments into that quality of carbon which most inhibits activity, viz. pseudographitic entities.S. J . Thomson 1899 40 80 120 160 time on stream/h Fig. 4. Extension of catalyst lifetime by promotion. After Hayes et a1.27 (-) Promoted, (---) unpromoted. Lifetime The attribute of extended lifetime of a catalyst by promoter action if it is successful will give the behaviour shown in fig. 4. Such behaviour has been observed by Hayes et al.,' In their study of the methanation reaction over Ni on y-Al,O, they investigated three classes of promoters. Ag and Pt were selected because of their hydrogen spillover characteristics,28 potassium for its inhibition of graphiti~ation,~~ and lanthanum with a noble metal for rate increase for hydrogenation of carbon.30 Of these only Pt was satisfactory as promoter for a 9 % Ni/Al,O, catalyst, as was Ru, in that they significantly improved and sustained catalytic activity.Ag, K and La all led to reduction of lifetimes. Of these only La gave an enhanced initial activity. Their evidence pointed to a mechanism in which the steps are 2Ni + CO(g) -+ Ni-C + Ni -0 Ni-C + H(ads) -+ Ni EC-H with further stepwise addition of hydrogen to produce CH,. Ni-0 + CO(g) -+ CO,(g) + Ni Ni-0 + CO,(g) -+ Ni -0-COT Ni-OH + bH, -+ Ni + H,O(g) and are accompanying reactions. The kinetics were best expressed by a power rate law: r = 8.79 exp (-9270/T)p'$i pijbS7 p-g$ with r in mol s-' g-l and p in bar.This corresponded to an activation energy of 78 kJ mol-l, lower than the normal value for Ni: this was ascribed to the promoter, Pt. The authors point out that the power rate law is the most reliable means of predicting rates within the experimental boundaries. Attempts to produce satisfactory kinetic expressions assuming one rate-determining step by the Langmuir-Hinshelwood Hougen- Watson approach failed.1900 Introductory Lecture Promoter Action on a Particular Step in a Reaction A paper which gives experimental evidence on the particular reaction step affected by a promoter is that of Grant and Lambert.,, In ethylene epoxidation over Ag ethylene oxide may be further oxidised to CO,+H,O. (The rate is lower than that for ethylene oxidation to CO,.) This process is inhibited by Cs.By using a single crystal of Ag exposing the (1 11) plane, making partial coverages of Cs metal and then studying C,H,O isomerisation to CH,CHO, they were able to establish the inhibiting effect of Cs on this reaction : it had been suspected that isomerisation preceded further oxidation of C,H,O. In a reaction study on the single crystal in a reactor Cs at 0.1 monolayer inhibited the rate of isomerisation by x 4, and thus its precise point of impact on the overall reaction has been established for temperatures < 410 K. The authors speculate about the mode of operation of the Cs. The Ag probably contains dissolved oxygen which can enhance C,H,O uptake. O(d) induces surface A&+ species which interact with the 0 of C,H,O.Although Cs slightly enhances the rate of adsorption of C,H,O it may assist the reaction C,H,O(g) --+ C2H4 + O(ads) the 0 in turn enhancing the number of Ag8+. Cs, however, has a strong negative effect on isomerisation. The reaction scheme examined was k. C2H40(g) C,H,O(ads) k-1 C,H,O 2 CH,CHO(ads) etc. Step (1) is ruled out as rate determining for it is enhanced by Cs and the rate is then given by (k, k,/k-,)pEO.. Cs inhibition may be explained by a scheme which draws widely on soft acid-base, frontier-orbital arguments. The case rests on Cs altering the HOMO level on the O(d). Chemical Form and Persistence of a Promoter A fundamental question concerns two aspects of the state of a promoter: is the chemical form important and is it likely to persist throughout a catalytic reaction? A clear answer to these questions by a molecular-beam method was given by Steinbach and Kra113, who were interested to know if Cl(ads) or C1 as ‘NiCl,’ on polycrystalline Ni had different consequences for the decomposition of CH,OH.They established that CH,OH at low pressure on Ni led to an associative adsorption, followed by formation of CH,O(ads), with dehydrogenation to CO and H, as final products. At higher pressure CH,OH dissociated via CH,(ads) and OH(ads) to produce CH,, H,O and C(ads). The associative mechanism was favoured by the presence of C(ads) or O(ads) and by an electron density for the metal close to the Fermi edge which was reduced compared with that for Ni.The dissociative mechanism was favoured by hydrogen enrichment of the surface, high surface roughness and an electron density like that of Ni. A surface with adsorbed C1 was poisoned for CH,OH decomposition since CH,OH could be recovered without reaction from such a surface. Ar sputtering removed Cl(ads) but left behind ‘NiC1,’ and an electron density at the Fermi edge which was unchanged: on this surface CH,OH began to decompose to produce CO(ads) which in turn dissociated at ca. 370 K to C(ads) and O(ads). The promotion effect has therefore been that of producing CH,O(ads) and CO(ads) by the associative route. However, in addi- tion CH,(ads) and O(ads) are produced by the dissociative route. This would not have been observed simultaneously on clean Ni. The ‘ NiCl,’ catalyst would more closely resemble a working catalyst than the Cl(ads) contaminated surface.S.J . Thomson 1901 A general observation made by the authors is that C-0 and C-C dissociation depends on both electronic and geometric factors. High electron density close to the Fermi edge is required but this is not exclusive. For Ni there is high electron density but for low index planes this is counteracted by the geometric factor. Only high index planes or highly disordered structures produce dissociative adsorption. CH30H decomposition is observed on rough surfaces generated by Ar sputtering without annealing. Perhaps ' NiCl, ' creates surface disorder; however, note that it does not disturb the Fermi edge density as do adsorbed 0, C1 or C. There is a further principle which emerges from this paper : it is that an inversion effect might occur in poisoning-promoting situations during the course of catalytic reactions and that coadsorption studies to examine poisons or promoters may not or cannot reveal the true state of a working catalyst.Fischer-Tropsch and Related Synthesis In Fischer-Tropsch synthesis over Fe catalysts using Fe,O, or K/Fe20, as starting materials, Dictor and Bell3, deduced from the observed induction period that the oxide was not the catalyst. Under working conditions activity and selectivity change little, yet the bulk phase of the catalyst undergoes substantial change:,, initially it can consist of 85% a-Fe, finally 65% Fe,O,, 25% FeC, and 8% aFe. This raises the question of the nature of the surface of such working catalysts.Laser-Raman spectroscopy by Zhang and S~hrader,~ has produced interesting answers. They used CO as a probe for their K,0/Fe,03/A120, catalyst and identified the surface as having FeO, Fe2+ and Fe3+ species. Thus they deduce that two active phases exist on the surface, Feo/FeC, and probably Fe,O,, the former active in hydrocarbon the latter active in oxygenate production. Infrared spectroscopy offers a key to understanding the role of a promoter. Peri and King36 used NO to study different K on Fe loadings. The antibonding n* orbitals of NO are capable of accepting electronic charge density from the metal chemisorption sites : this back-donation gives stronger Fe-N bonding and a corresponding decrease in the strength of the NO bond.The NO frequencies they observed did decrease with potassium loading for FeO and Fen+ sites. K20 they say is a basic species capable of electron donation to the metal. NO (or CO) may withdraw electrons from modified sites and be more strongly bonded and ultimately dissociated. Hydrogen, as an electron donor, will be inhibited from adsorption. Thus K,O additions will give catalysts in which the delicate balance of competitive hydrogen and CO adsorption will be altered. This accounts for the balance between the production of oxygenates from undissociated CO and the production of hydrocarbons from dissociated CO. Alkene production should be favoured when hydrogen surface coverage is diminished. Since CO frequency shifts figure largely in the promotion literature it is appropriate to point out that electric field effects offer an alternative explanation to back-bonding.15 The same picture of promotion emerges3' for the H,+CO reaction over Ru/A120, doped with K, P or B.K, from K,C03, influenced thermal desorption of adsorbed CO, and the infrared spectra confirmed stronger adsorption. The alkene/alkane ratio was increased and CH, production suppressed. P and B had opposite small effects. The authors claim that the spectroscopy reveals an electronic change in the Ru: CO adsorption was stronger because of back-donation into the Ru-C-0 system. They make the significant observation that K may increase the amount of inactive CO.,* The surface CO may influence the availability of H, with a consequent effect on alkene production.The idea of blocking of H sites has been tested in the next paper.39 Alkali-metal doping by K O of Mo(100) surfaces and polycrystalline foils at coverages of < 0.3 monolayer increases the rate of hydrogenation of CO and it enhances the alkene to alkane ratio.1902 Introductory Lecture Logan et al. account for this by suggesting that doping enhances CO dissociation into C and 0. They also invoke the back-bonding arg~menf.,~ Thus if C and 0 coverages are enhanced and hydrogenation is rate limiting then the yield of unsaturated species will be increased. The argument of the authors is sustained by the observation that S adsorption blocks sites and reduces reaction rate but it does enhance C,H, production as measured relative to CH, by a factor of 5. Caution is necessary in the interpretation of spectra as a basis for explaining promotion.Thus Deligianni et aL41 explored the Li activation of Pd/SiO, for methanol synthesis. Whilst they saw weekly adsorbed linear and bridged forms of CO (2075 and 1975 cm-l) and strongly held CO (2060 and 1900 cm-l), the addition of Li showed 'no obvious correlation of activity with any particular type of CO bonding'. There was a small effect on the bridged CO, but the authors say the observed six-fold increase in activity cannot be ascribed to this. They concluded that there could be an electronic effect or a structural effect or an effect of the Li on the support in which heterolytic splitting of H, is induced. Thus in all spectroscopic papers there should be investigation or at least comment on the relationship between the spectroscopy and catalysis.An ideal example arises in the work of Vannice and his collaborator^^^ on Pt supported on Al,O,, SiO, or Al,O,/SiO,. On some of their most active catalysts for CH, production there is no evidence, spectroscopically, for adsorption of CO. CO attachment to oxide-promoted catalysts has been approached both intuitively and spectroscopically. Sudhakar and Vannice,, studied the H, + CO reaction over a series of rare-earth supports. In their results it was found that Pd on CeO, gave a high yield of CH, (1 atm, 548 K), Pd on Eu,03 gave a low yield of CH,. The authors offer an explanation based on a metal-support model involving (I). Since they worked under reducing conditions there would be Ce3+ species present and 0 vacancies.Now Ce3+ can most easily be oxidised to Ce4+ and the C-0 bond broken with C being hydrogenated to CH,. The poorer methanation performance of Eu can be understood for Eu,O, is the most easily reduced oxide: i.e. its tendency to accept 0 is reduced and it is therefore poorer as a methanation catalyst. c-0 / ---- -. M"+-0 H,- Pd 0' It is therefore not unexpected that a promoter may shift the ratio of linear to bridge-bonded C0.44 Thus Lao, species believed to be present in La,O,-promoted Pd/SiO, decreases the bridged population. It also enhances CO fission and the rate of CH, formation from CO+H, compared with Pd/SiO,. This may be an example of nucleophilic attack on the oxygen end of the adsorbed CO by the Lao, species. There is a parallel here with the TiO, promotion of Group VIII metals proposed by Bracey and B ~ r c h ., ~ Sachtler et al.46 have proposed a model for which the starting point was promotion by alkali-metal ions, or early transition-metal ions such as Mn2+, for the formation of oxygenates on Rh. Rates increase, but selectivity does not change. In Mn2+Rh/Si0, enhanced metal dispersion was not the cause: indeed H, and CO chemisorptions were diminished on promoted catalysts. Their proposal is that the ions influence CO adsorption as shown in (11) where P is the promoter and B the counterion. 0S. J. Thomson 1903 Structure characterisation of certain organometallics support this 0 and C bonding of CO. Alkali-metal ions binding 0 enhance CO insertion reactions. Spectroscopy reveals a weaker CO bond when both 0 and C are bound: frequencies decrease by 100-300 cm-l compared with terminal (1 950-2 130 cm-l) or bridging ( 1 750- 1900 cm-l) CO species.CO on Mn2+Rh gives a band at 1530 cm-l, as does CO on stepped Ni [Ni5( 1 1 1) x (no)] or Fe (1 1 1). The authors note that if the oxide and promoter ions formed a separate phase the promotion would fail. They point out, however, that in SMSI on TiO,, Rh and Ni become partially covered by Tin+ (n -c 4) and 0,- ions. Whereas the expected effect of alkali metal might be to produce hydrocarbons on FT catalysts through C-0 fission, this is not always the case: alkali metals may enhance yields of oxygenates. Chuang et aL4' chose to experiment on Rh, for this metal produces alcohols and hydrocarbons, in the hope of unravelling the problem.The basicity of the support was known to be critical: MgO, very basic, gives 90% selectivity for alcohols: Al,O,, acidic, gives mainly CH,. SiO,, TiO, and CeO, are intermediate in behaviour. The alkali-metal species they used were Li, K and Cs from nitrates for Rh/TiO,. Alkali metal suppressed activities as far as rates were concerned, perhaps by interfering with hydrogen availability : selectivity towards oxygenates increased upon alkali-metal promotion, e.g. TiO, as support at 300 "C and 10 atm. Rh Li-Rh K-Rh Cs-Rh wt% CH, 49.9 46.6 37.7 20.6 total OX 17.7 22.3 46.5 46.1 (OX = MeOH, EtOH, MeCHO, or acetone). The expectation might have been enhancement of CO dissociation. The authors explain their results on the basis that CO insertions produced oxygenates from CH,(ads), which seems to leave intact the basic role previously described for the alkali metals.In what follows there is a fine illustration of double promotion and the many effects a promoter can have on a catalytic system. Changes occur in rate of reaction, selectivity, activation energy, bond strengths and order of reaction. For their study van den Berg et chose Rh on SiO,, the reaction CO+H,. Rh appears to be a metal whose behaviour can be influenced to a considerable extent by its surroundings: e.g. on MgO and ZnO it produces MeOH; on La,O,, EtOH; on SiO,, CH, and hydrocarbons. This sensitivity to environment made it a promising metal for promotion study using Mn+Mo as promoters. In summary, the findings were as follows.(1) After 24 h the activity in terms of gas conversion was ten times higher for MnMo/Rh/SiO, than for Rh/SiO,. (2) EtOH was the major product at the start of a 136 h run. CH,CO,H was the major product from 72 h on MnMo/Rh/SiO,: on Rh/SiO, MeOH is the major product with an increasing yield. (3) Physical mixtures of Mo/SiO,, Mn/SiO, and Rh/SiO, behaved like Rh/SiO,. (4) Promoter oxides seemed to be on the Rh surface, for ethane hydrogenolysis was inhibited. (5) Reduced catalysts (the promoted catalyst is more difficult to reduce) were extracted with acetyl acetone: Rh/SiO, yielded 4+ 1 % of total Rh as ions: the promoted catalyst 15+ 3%. The Rh ions were of RhI. (6) Infrared investigation showed that the promoted catalyst adsorbed 2 to 3 times less CO than Rh/SiO,.The CO frequencies were also shifted for ihey were lowered perhaps by a Rh-C-O/Mn or MO ion interaction. (7) Promoter and oxygen ions may cover part of the Rh metal and this overlayer may stabilise the Rh+ ions, which were extractable by AA. Kinetic information suggests a difference in the power rate law for promoted and non-promoted catalysts: Approximate values were r = k exp(-Ea/RT) Pf€&O- E X Y Rh/SiO, 26 0.9 - 0.4 Mn/Rh/SiO, 24 0.6 -0.5, -0.31904 Introductory Lecture The change in x is seen as significant in that H availability is of smaller consequence on the promoted catalyst. Over a period of time on a promoted catalyst the decrease in hydrogenating ability led to a change from CH,CH,OH production to CH,CO,H. The increased hydrogenation rate on promoted catalysts can be accounted for by the decreased heat of adsorption on Rh+, where back-donation of electrons by the ion to the 2n* orbitals of CO is limited compared with Rho.Thus a change in 8,, from 0.999 to 0.99 could result in a change in 8, from 0.001 to ca. 0.01, i.e. ten-fold increase, and a ten-fold increase in activity. The authors conclude that MnO and MOO, change the relative surface concentrations of CO and H. For selectivity the authors propose that: (a) hydrocarbons are formed by CO dissociation on Rho followed by H + C(a) reactions, (b) CH,OH is formed on Rh+ and (c) alkyl groups formed on Rho transfer to Rh+ to give a nucleophilic attack of alkyl on a Rh+-coordinated CO. Thus the CO/H ratio on the surface and the Rh0/Rh+l ratio determine the selectivity of the catalyst.C deposition on the Rho will also alter selectivity during the course of a reaction. Electronegativity Although the concept of average electronegativity of a compound is not accepted universally, the concept produces some interesting correlations. Aika et ~ 1 . ~ ~ have systematically studied the effect of alumina supports from various sources and alkali-metal hydroxides on Ru/Al,O, catalysts for NH, synthesis. If the average TON per exposed Ru atom for ammonia synthesis at 315 "C and 600 Torr is taken as 1.2 x lo4 then the average values are 6 x lo4 for KOH/Ru/Al,O,, 7 x lo4 for CsOH/Ru/Al,O, and 60 x lo4 for K/Ru/Al,O,. This latter value is for the metal K. The authors note that the activity sequence is related to the electron-donating tendencies of the promoters.The order of the average electronegativity of the compounds is 2.5 for Al,O,, 1.83 for KOH, 1.75 for CsOH and 0.8 for K. In which they examined promoted Ru/oxides as catalysts for NH, synthesis. The authors used supports which were themselves basic, and were interested in KO on basic supports as a possible superbase. The turnover number for Ru mounted on such catalysts plotted against the ele~tronegativity~l of the predominant promoter compound gave the result shown in fig. 5. The Ru electron binding energies follow the same trend in the graph, and they can be interpreted in terms of Ru receiving 'an electron' from the alkaline oxide: on this basis electron availability is lowest on Ru/Al,O,, highest on K on the oxides MgO, BeO, Al,O,, CaO or C.The authors state their view that the lower the electronegativity of the support or promoter the higher is the turnover number, and that electron donation [rom the support promotes N, dissociation. The superbase Ru-K/MgO is lo3 times more efficient than Ru powder in NH, synthesis. Activation-energy changes may not be sufficient to account for this difference, which may depend more on the stabilisation of an activated complex. Ertl et ~ 1 . ~ ~ have taken the view that K on Fe stabilises adsorbed N,. Electronegativity is also the subject of a paper by Aika etS. J . lo-* z E i P h 0” & c lo4 Thomson ’ 1905 1 2 3 electronegativity Fig. 5. Ammonia synthesis rates at 315 OC, 600 Torr, over Ru on a variety of oxides of different electronegativities. After Aikai et aL50 (1) Konoxide, (2) CsOH, (3) KOH, CaO, (4) MgO, (5) BeO, (6) AW,.Future Prospects Theoretical A few areas of particular promise have been selected for comment, the first of which illustrates a theoretical approach. It is not on promotion but it reveals how subtle that effect must be. Harrison and Chianelli53 have written a remarkably fine paper on the relative activities of transition-metal sulphides in hydrodesulphurisation (h.d.s.). The first-row elements in the Periodic Table show small activity compared with second- and third-row elements. Across the rows these latter show a volcano-shaped activity curve. A cluster calculation for MSr- gave the following valence-energy diagram. 3ega* - M(d)S(3p) component 2tzgn* 1 t 1 g =I 2ega - non-bonding M(d) component 11906 Introductory Lecture AE decreases on going from left to right in the Periodic Table.The M (d) component increases from left to right for a 0 bond, and increases from left to right for a 7c bond for second-row metals but not for first-row. A plot of h.d.s. activity us. the number of d electrons on the metal, corrected for valency, did not give a correlation. The number of d electrons in the highest occupied molecular orbital (HOMO) gave the first indication of a correlation with h.d.s.: the higher the number of d electrons, the higher the activity. The later 4d elements have the highest numbers of electrons. This is because the M - - - S interaction is weak for 3d with poorer orbital splitting and hence high-spin 2t2g and 3eg electrons. In contrast, for Ru and Rh only the 2t,, orbital is occupied in a low-spin configuration.Thus M . . . S splitting gives a high n(HOM0). The authors then consider M-B bond strength as a further factor in activity. Related to bond strength must be the quantity D where D = 2 0 , + 3 0 , : 2 and 3 are degeneracies, D, and D, are metal contributions to 0(2e,) and "(lt,,) orbitals. D us. h.d.s. does not give a correlation, but A , = nD does and gives an excellent predictor for activity: n is the number of electrons in the HOMO. The authors then refine this by considering the M-S relative covalent bond strength, B. They regard this as more important than ionic strength. B = n, D,+n, D, where these n are net- numbers of bonding electrons (bonding-antibonding).A plot of A , = nB, which takes account of the occupancy of the HOMO, n, and the relative measure of the covalent bond strength, B, gives excellent correlation. B takes account, for example, of the greater metal orbital mixing on the right of the Periodic Table and the greater number of antibonding electrons. The concepts in the Harris-Chianelli paper have been extended to promotion by Ledoux et ~ 1 . ~ ~ for Co promotion of Mo/S catalysts. Changing the Co/Mo ratio on their SO,-supported catalyst gave a volcano-shaped curve for thiophene h.d.s. activity. Co,S,, which has octahedral and tetrahedral Co was used as a model compound for their n.m.r. investigation. Very active tetrahedral sites were prominent at the height of the volcano which were associated with activity of a new distorted tetrahedral phase.The e, levels now lie below the t,, levels for the tetrahedral Co and the number of electrons in the HOMO is modified to 3 (?), whereas it is 1 in the 3e, level of octahedral Co. There is therefore an increase in activity from A , = nB with Co, now behaving more like Mo, present in a highly dispersed active phase. Promotion on metal surfaces has also been the subject of study. In a considerable break from the conventional approach to promotion Marks and Heine55 have, in an elegant fashion, drawn together a number of observations and ideas already in the literature and their own work, to produce a coherent view of promotion. They start from known modifications which occur on single-crystal surfaces through the addition of foreign atoms.For noble metals they then examine the pairwise tensions and pressures which arise from sp and d electrons. The balance between these forces gives a situation in which there is inherent instability at surfaces. The authors then quote examples of surface rearrangements which can be explained by surface stresses which favour close-packed surfaces : e.g. the (1 1 1 -like) overgrowth on noble metals and (1 1O)Au surfaces undergoing microfacetting to (1 11) surfaces. The key point is that structural changes are occurring because of electronic factors, and this points to a rde for electronic donors or acceptors as modifiers of catalysts. They suggest that alkali-metal modifiers on Ag( 110) induce reconstructions to ( 1 1 1) surfaces which are good for epoxidation, whereas the opener (1 10) or (100) surfaces favour reaction to CO, + H,O.The idea can be carried further to supported metals, where epitaxy in growth of small crystals may produce structures unlike those of the bulk and which have different electronic properties.S. J . Thomson 1907 Single Crystals Studies on single crystals are likely to reveal much in the future. Already Grant and LambertS6 claim the first single-crystal study of ethylene epoxidation and oxidation. They chose the (1 11) plane of Ag as a representative face. The reaction scheme is con- troversial over the contribution of paths (l), (2) and (3) (+ CO,+H,O 4 3 ) and there are various proposals over O,(ads) as the crucial species for C,H40 formation, or for O(ads) or O,(ads) giving both epoxidation and oxidation.Dissolved oxygen has also been considered. Selective removal by t.p.r. of O,(ads) from a surface covered by O,(ads) and O(ads) demonstrated that O(ads) was the active species. Intermediates such as CH,CHO, CH,CO,H and (CO,H), were detectable. When C2D4 was used yields of C,D,O/CD,CDO were enhanced 1.5 times, CD,CO,D 7 times. Dissolved 0 activated the surface and participated in C,H,O oxidation but not its formation. Grant and Lambert propose as a mechanism (a-CO, peak) C2H40(g) / (2) C,H,O(ads) <(ads) (3p O(ads)+C H 4 \ (4) CO,+H,O (p-CO, peak) where reaction (4) consists of C,H40 -+ CH,CHO + CH,CO,H -+ (CO,H), + CO, + H,O each step involving O(ads). When C,D, is oxidised the yield of the intermediate in reaction (4) increases: this suggests ' H '-transfer reactions hindered by D substitution and not C-C fission.Hindering the steps in reaction (4) results in increased probability of desorption of intermediates. This is in agreement with the batch-reactor observation that k , / k , = 0.5, a normal isotope effect, for CO, production. This suggests C-H cleavage as a crucial step in reactions (1) and (4). In the production of C,H40 there is an inverse isotope effect; k,!k, = 6.5. The reason offered is that D substitution favourably affects the branching ratios of both (1)/(2) and (3)/(4)- Promoters enter into the scheme as follows. In reactions (1) and (2) weakly acidic H atoms are attacked by basic O(ads). Assuming the C,H4 to be adsorbed on Ag(6+) one can write [The high charge on O(ads) favours total oxidation.]1908 Introductory Lecture Alternatively an electrophilic attack on O(ads) might be represented as + 0 0: &- 0 [Charge removal by Cl(a) or O(d) favours epoxide.] Competition between reactions (1) and (2), i.e.selectivity, is affected by the effective charge state of O(ads): thus total oxidation involves charge transfer from O(ads), and the higher the charge, the more favoured will be total oxidation. Epoxidation involves charger transfer to O(ads), a process favoured when O(d) competes for metal electrons with O(ads). Cl(ads) should likewise compete for electrons favouring reaction ( 2 ) over (1). Reaction ( 2 ) is not subject to an isotope effect, and therefore the isotope effect in the reaction should decrease as Cl coverage increases.Work on single crystals is now making significant advances where modifiers are added to surfaces. On such surfaces there is a fundamental question to be considered on the mode of action of the poison 0: promoter: is it blocking sites or is it in addition exerting an electronic effect on the surface? A combination of surface physics and theoretical approaches has provided answers. Goodman5’ has recently summarised the position. Kinetic studies revealed that the addition of S to Ni( 100) rapidly reduced H availability for methanation. The poisoning effect is non-linear, and in the early stages of the poisoning it appears that one S atom deactivates about ten Ni atoms. This might be caused by a long-range electronic effect or the destruction of ensembles.Substitution of S by P gave the answer. The shape of the poisoning curve was changed markedly at low coverage, as shown in fig. 6. It was thought that only four near-neighbour sites were blocked by each P atom. The conclusion was that the difference between the actions of S and P lay in their electronegativities. The effect of additives on H, adsorption has been shown5* by KO and Madix to be more than blocking of sites. They showed that the total amount of hydrogen which could be adsorbed on poisoned surfaces could be enhanced by H, predissociation to atoms. They used Mo( 100) partially precovered by 0 or C, both of which reduced adsorption when the surface was exposed to H,. However, when a carbided surface with 8, ~ 0 . 7 was exposed to atomic H, 0.3 of a monolayer was taken up, about a 5 x enhancement.They concluded that H, adsorption was an activated process. Theoretical work by Grimley and T ~ r r i n i ~ ~ and Einstein and SchriefferGo supports the concept of electronic interactions over several atomic spacings and recent calculations by Narrskov et aLG1 cast further light on electronic interactions. They examined electrostatic potential distributions around Na, Si and C1 adsorbed on an Mg surface where the metal was treated as a jellium. It emerged that the approach of H, to an unpoisoned surface would have a binding energy us. reaction coordinate diagram as shown in fig. 7. The diagram has an interesting molecular bound state which suggests surface migration for molecules which cannot overcome the activation barrier to dissociation.The barrier to dissociation can be lowered either by a partially filled d orbital at the Fermi level or by lowering of the metal work function. Taking CO as an example at a distance of 2.5 A from a Na atom the electric potential is ca. +0.3 to 0.5 V. This lowers the work function by these amounts. The calculation also depends on estimation through frequency changes of the electron density transferred to the CO, where electrons are accommodated in the 2n* antibonding orbital. They then calculate an overall stabilis- ation of the binding energy of CO as 0.5 to 0.3 eV.S. J . Thornson 1909 Fig. 6. The shape of poisoning curves for the methanation reaction when S or P are added to Ni( 100). After Goodman.57 I dissociated molecular reaction coordinate Fig.7. The approach of H, to a Mg(OO1) surface: binding energy against reaction coordinate. After Norskov et C1 atoms behave in the opposite sense, and thus the enhanced adsorption of CO on metal in the presence of Na, Cs or K can be understood as can the opposite effect for S, C1, P and C.62 The distance over which electronic effects operate has been studied theoretically by MacLaren et aE.63 On Ni(lO0) C adatoms affect nearest-neighbour sites whereas S also poisons next-nearest neighbours. There is another way in which surface physics stimulates advances in the understanding of catalysis and promotion. The extensive studies of surfaces of different metals and the behaviour of adsorbates upon them are stimulating work on coherent explanations.Such a theoretical account of chemisorption which fits experiment has been published in a review of exceptional merit by Shustor~vich.~~ The range of phenomena to be explained was extensive and daunting: the monotonic decrease of Q A , the heat of chemisorption, for atomic radicals of H, 0 and N from left to right across a transition series; the non-monotonic behaviour of QAB for small molecules such as CO and NO; preferred sites of chemisorption on flat surfaces, on top for molecules, hollows for atoms ; activation barriers for migration and dissociation where APAB for dissociation depends on atomic heats of adsorption and not the molecular heat; the A-B stretching frequencies vAB; modifier effects on vAB; work functions; coadsorption. The approach involved perturbation theory and bond-order conservation (BOC) in its Morse-potential version ' not among the beaten theoretical paths '.The conclusions from the perturbation theory were that variations in QA and Q A B could be explained. Most ad-molecules behaved as electrons acceptors: all adsorbates led to surface polarisation changes with the dipole directed into the bulk with a decrease in work1910 Introductory Lecture function: contrary to popular belief a molecule can be an acceptor and yet decrease 4. Also AB molecules adsorb upright on flat surfaces with coordination through the less electr onegative atom. When an adsorbate AB moves or dissociates on a surface the coordination M,-AB changes, the M,-AB distances change as does the total energy E. E can be related to the bond order x for the M-A, M-B and A-B bonds by x = exp [ -(r -- r,)/a] where ro and a are constants.The original Pauling and BEBO bond orders were thus improved upon. Then the Morse potential E(x) = - Q(x) = - Q, (2x-x2) may be used to calculate minimum energies. The potential has both attractive and repulsive terms and an energy minimum at x, = 1. In this equilibrium state ro and Q, are observables. The model then assumes the conservation of x = 1 for multiple M,-A interactions. Along a migration path up to dissociation bond order for a complex Mn-X, system is conserved and normalised to 1. At low coverages adatoms are predicted to be in hollow sites. For molecules AB chemisorption can be described by n where xAi. and xBi are M-A and M-B bond orders and xAB,, is the AB bond order for coordination to n M atoms.Analysis shows that on-top or bridge sites are typical, with vAB decreasing as n increases. Coordination via C in CO or N in NO is predicted. Dissociation of AB depends on its dissociation energy and Q A and QB, not on the molecular heat of chemisorption, Q A B . In the presence of an adatom, D, AB molecules may be pushed from their on-top or bridge sites into hollows. This changes the molecular adsorption heat Q A B and affects vAB. As OD increases Q A B drops and vAB increases. However, increased coordination in hollow sites lowers vAB. The conclusion is that vAB is no longer a reliable guide to information on a chemisorption site. In an example of an attractive D-AB interaction on a metal M, viz.K and CO on Ru(OOl), the model accounts for lowering of vco, increased CO-M, coordination and reorientation of CO parallel to the metal leading to the possibility of a K-CO complex. These are but a few extracts from many examples and applications of the model. The last topic under single crystals opens up a new strategy for understanding molecular interactions on surfaces where one of the adsorbed species may be a promoter. Roberts and his colleague^^^ have described their approach to quantitative identification and determination of species on surfaces through combined X.P.S. and U.P.S. investigation. One of the aims of their studies was to explore the extent to which reaction pathways can be influenced by a second, different molecule. They say that co-adsorbed molecules may follow thermodynamically unlikely pathways.In a typical example they show how oxygen from N20, NO or 0, might be expected to follow a reaction path as follows on a Mg(0001) surface: (1) (2) (3) O,(g) -+ O,(ads) + 0-(ads) --+ 02-(ads). Although ammonia is only weakly adsorbed on a Mg(0001) surface (it desorbs at 170 K) and although it is relatively inactive on an oxide surface at 295 K, when co-adsorbed with oxygen it behaves differently. Step (3) is intercepted by the NH, in such a way that 0-(ads)+NH, + NH,(ads)+OH(ads). Other examples are given which open a new channel of thinking on surface mechanisms.S. J . Thornson 191 1 Multiphase Solids These offer the possibility of producing an infinite variety of catalysts or supports. In, for example, Co/Mo catalysts for h.d.s.the role of Co has been examined by Ng and Gu1a1-i.~~ Raman and infrared spectroscopy revealed that Co suppresses the formation of bulk-like MOO, on a TiO, support: instead a polymeric molybdate monolayer is formed. At higher loadings of Mo layers of CoMoO, species are formed on top of the molybdate monolayer. An interesting reversal in thinkings6 about such catalysts is that the tenfold increase in h.d.s. activity when Co is added to MoO,/Al,O, is that Mo may be promoting an activity in Co. Multiphase systems appear also in the molybdate catalysts used in selective hydro- carbon oxidation. They may fall outwith the class of promoted catalysts but none the less it is deviation from perfection which produces working catalysts.Ozkan and SchradeP7 point out that pronounced changes in the behaviour of simple molybdates can be observed owing to incorporation of excess MOO,: thus the Mo/Ni, Mo/Co, Mo/Bi ratios are larger than those to be expected in simple molybdates, e.g. NiMoO,, and they are no longer simple one-phase oxides. Keulks et aZ.68 have reviewed the role of iron as a promoter in molybdate catalysts. M,Mo6S8, with a very wide range of guest atoms and concentrations, offer possibilities for studying promotion and the importance of the oxidation states of Mo. The importance of MoIV has already been established in h.d.s. The Chevral phases are pseudomolecular compounds based on Mo& clusters joined by Mo-Mo and MO-S bonds. They have a channelled structure with guest atoms in precise locations. The guests are conveniently classified as large (Ho, Pb, Sn), inter- mediate (Ag, In) and small (Cu, Fe, Ni, Co), and their activities and selectivities in dehydrosulphurisation follow that order.They are at least, if not more active than conventional catalysts and more selective in desulphurisation rather than hydrogenation. What is also of importance is the controlled value of the Mo oxidation number between +2 and +d, depending on concentration/valency of the guest. The test reaction at 400 "C was thiophene to C, hydrocarbons: hydrogenation was of but-1-ene. Chevral phase Mechanistic Probes Just as tracers have been used in studying catalytic mechanisms there is a parallel development in the technique of adding molecules as probes for mechanistic studies.This is illustrated in alkali-metal promotion of the CO : H, reaction over Rh/TiO,. Chuang et aL70 say that the role of the promoter is not understood in the production of oxygenated compounds. Their investigation involved addition of small amounts of ethylene to H, + CO. Over unpromoted and promoted Rh/TiO, CO conversions decreased in the order unpromoted. > Li > K > Co. The downward changes in rates were, for hydrocarbons x 55, for CH,CHO and C,H,OH x 22, and for CH,OH x 4. Thus there was enhanced selectivity for oxygenates. Addition of ethylene revealed interesting observations. Rates of formation of all products were increased, except for CH,. On promoted catalysts the rates for hydro- carbon formation dropped significantly but the rates for CH,CH,CHO little. The difference between oxygenated yields and hydrocarbons may be explained as follows.(i) CH,CH,CHO can be formed by CO insertion into an adsorbed C,H, species. (ii) CH,CHO is probably formed by CO insertion into M-CH,, which has to be derived from CO dissociation. The authors then pose questions on the alkali-metal promoters: do they block sites on the larger ensembles necessary for CO dissociation? This explanation is not satisfactory, since alkali metal enhances CO adsorption and dissociation although it does1912 Introductory Lecture reduce hydrogenation of olefins and surface C. The authors then examine the alternative chemical possibility. Since CO insertion to produce CH,CH,CHO is fairly insensitive to the nature of their catalysts they propose a single site mechanism for this insertion: this site cannot be a hydrogenation site for it was highly sensitive to the ‘promotion’.The authors note that alkalis may change the electron density on metal catalysts, but the alkali-metal cations themselves cannot donate electrons. Alkali metals may lower the average oxidation state of Rh but this runs counter to the idea that Rh ions are involved in oxygenate formation. In practice the authors saw only slight modification of the production of oxygenates. Microengineering There are two approaches which seem to offer promise. In addition to the Chevral phase approach there is another chemical method, that of using compounds as precursors for an active component and its promoter. For the H,+CO reaction addition of K to Fe and Ru catalysts is known to produce selectivity towards olefins.Preparation methods for highly dispersed catalysts have evolved to the use of compounds as precursors for the active metal and its activator. Thus McVicker and Vannice71 have used K,Fe(CO), on Al,O,, and Doi et aL7, have Na[Ru,H(CO),,] In the latter case, supported on Mgo and TiO,, the promoted catalyst exhibited a selectivity of 69 to 75% for formation of C, to C, alkenes. On SiO, the hydrocarbon yields were given by the Schulz-Flory prediction : the selectivity for olefins was 5%. The second approach of promise comes from a book entitled Surface Engineering.73 The subject is controlled ion implantation in solids. Penetration depth and number and type of ions can be controlled with surface properties adjusted independently of the bulk.Homogeneous dispersions can be produced as well as new metastable phases of, for example, Au/Ni, Au/Co and amorphous alloys where the constituents have different crystalline structures: e.g. f.c.c./b.c.c. for Ni/Mo, Al/Nb and Ni/Nb. Metastable systems can be induced to produce high yields of small precipitates in the 10-100 A range. Manipulation Entropy and Electric Fields In the search for increased rates of reactions and selectivities we should extend our horizons and examine two other branches of science in which quite different methods have been employed to manipulate reactions. The changes in rates brought about by enzyme catalysis may be of the order of lo9. In his recent book on enzymes Fersht7, places great emphasis on explanations based on considerations of entropy.The argument is that free molecules with many degrees of freedom in translation, rotation and vibration undergo most unfavourable entropy changes in the formation of transition states. When, in contrast, the reacting molecules are trapped on enzyme sites with greatly reduced freedom of motion then the exp (ASIR) term in the rate equation is reduced by many orders of magnitude. Equivalent concentrations to produce the same reaction rates rise by factors of lo9. This concept of trapping we may already be using in pores and in zeolites but we do not yet plan our activators around this idea. The second area in which significant changes in rates occur, by design, is in the application of electric fields to catalysis.The volume of papers on these effects is not large, yet there are papers in which important changes have been observed not only in rates of and yields76 but also in extents of ad~orption~~ and in equilibrium constants them~elves.~~ Thus it is strange that this experimental variable in catalysis has been so neglected.S. J . Thomson 1913 There could be not better statement to close this address and to give a theme to the conference than the words of Jobling in 1916: ‘catalysts appear to possess the capacity for invigoration .’ References 1 E. Jobling, Catalysis and its Industrial Applications (J. and A. Churchill, London, 1916), pp. 10 and 43. 2 N. K. Adam, The Physics and Chemistry of Surfaces (Oxford University Press, Oxford, 2nd edn, 1938), p.236. 3 P. Jones, in A Hotbed of Genius, The Scottish Enlightenment, 1730-1790, ed. D. Daiches, P. Jones and J. Jones (University Press, Edinburgh, 1986), p. 43. 4 K. Aika, A. Ohya, A. Ozaki, Y. Inoue and I. Yasumori, J. Catal., 1985,92, 305. 5 G. Rambeau and H. Amariglio, J. Catal., 1981, 72, 1. 6 A. Ozaki, H. S. Taylor and M. Boudart, Proc. R. SOC. London. Ser. A, 1960,258,47. 7 F. Solymosi, I. Tombacz and J. Koszta, J. Catal., 1985, 95, 578. 8 S. R. Morris, R. B. Moyes, P. B. Wells and R. Whyman, J. Catal., 1985, 96, 23. 9 G-Y. Chai and J. L. Falconer, J. Catal., 1985, 93, 152. 10 G. B. McVicker and J. J. Ziemiak, J. Catal., 1985, 95, 473. 11 F. Solymosi, J. Catal., 1985, 94, 581. 12 K. Mori, A. Miyamoto and Y. Murakami, J. Catal., 1985, 95, 482. 13 G.C. Bond, in Metal-Support and Metal-Additive Effects in Catalysis, ed. B. Imelik, C. Naccache, G. Coudurier, H. Praliaud, P. MQiaudeau, P. Gallezot, G. A. Martin and J. C. Vedrine, Studies in Surface Science and Catalysis, vol. 1 1 (Elsevier, Amsterdam, 1982), p. 1. 14 A. A. Tsyganenko, L. A. Denisenko, S. M. Zverev and V. N. Filimonov, J. Catal., 1985, 94, 10. 15 N. S. Hush and M. L. Williams, J. Mol. Spectrosc., 1974, 50, 349. 16 S. C. Chuang, J. G. Goodwin and I. Wender, J. Catal., 1985,95,435. 17 C. T. Au and M. W. Roberts, Nature (London), 1986, 319, 206: C. T. Au, A. F. Carley and M. W. 18 G. M. Schwab, Adv. Catal., 1978, 27, 1. 19 A. K. Datye and J. Schwank, J. Catal., 1985, 93, 256. 20 C. T. Campbell and B. E. Koel, J. Catal., 1985,92, 272. 21 J. M. D. Tascon, P.Grange and B. Delmon, J. Catal., 1986, 97, 287. 22 K. Fujimoto, H. Saima and H-0. Tominaga, J. Catal., 1985, 94, 16. 23 J. Carraza, W. T. Tysoe, H. Heinemann and G. A. Somorjai, J. Catal., 1985,%, 234. 24 A. Coma, V. Fornes and E. Ortego, J. Catal., 1985, 92, (2), 285. 25 P. M. Boorman, R. A. Kydd, Z. Sarbak and A. Somogyvari, J. Catal., 1985, 96, 115. 26 V. K. Shum, J. B. Butt and W. M. H. Sachtler, J. Catal., 1985, %, 371. 27 R. E. Hayes, W. J. Thomas and K. E. Hayes, J. Catal., 1985, 92, 312. 28 D. L. Trimm, Design of Industrial Catalysts (Elsevier, New York, 1979). 29 G. A. Somorjai, Catal. Rev., 1981, 23, 189. 30 T. Inui, T. Hagiwara and Y. Takagami, Fuel, 1982,61, 537. 31 R. B. Grant and R. M. Lambert, J. Catal., 1985, 93, 92. 32 F. Steinbach and R.Krall, J. Catal., 1985, 94, 142. 33 R. A. Dictor and A. T. Bell, J . Catal., 1986, 97, 121. 34 S. M. Loktev, A. N. Bashkirov, E. B. Slivinskii, L. I. Zvezdkina and Yu. B. Kagan, Kinet. Catal., 1973, 14, 175: S. M. Loktev, L. I. Makarenkova, E. V. Slivinskii and S. D. Entin, Kinet. Catal., 1972, 13, 933. Roberts Int. Rev. Phys. Chem., 1986, 5, 57. 35 H-B. Zhang and G. L. Schrader, J. Catal., 1985, 95, 325. 36 D. L. King and J. B. Peri, J. Catal., 1983, 79, 164. 37 T. Okuhara, H. Tamaru and M. Misono, J. Catal., 1985, 95, 41. 38 C. S. Kellner and A. T. Bell, J. Catal., 1981, 70, 418: T. Kimura, T. Okuhara, M. Misono and Y. Yoneda, Nippon Kagaku Kaishi, 1982, 162. 39 M. Logan, A. Gellman and G. A. Somorjai, J. Catal., 1985, 94, 60. 40 B. E. Nieuwenhays, Surf. Sci., 1981, 105, 505. 41 H. Delliganni, R. L. Mieville and J. B. Peri, J. Catal., 1985, 95, 465. 42 M. A. Vannice, C. C. Twu and S. H. Moon, J. Catal., 1983, 79, 70; M. A. Vannice and C. C. Twu, J. Chem. Phys., 1981, 75, 5944. 43 C. Sudhakar and M. A. Vannice, J. Catal., 1985,95, 227. 44 J. S. Rieck and A. T. Bell, J. Catal., 1985, 96, 88. 45 J. D. Bracey and R. Burch, J. Catal., 1984, 86, 384. 46 W. M. H. Sachtler, D. F. Shriver, W. B. Hollenberg and A. F. Lang, J. Catal., 1985, 92, 429.1914 Introductory Lecture 47 S. C. Chuang, J. G. Goodwin and I. Wender, J. Catal., 1985,95,435. 48 F. G. A. van den Berk, J. H. E. Glezer and W. M. H. Sachtler J. Catal., 1985, 93, 340. 49 K. Aika, K. Shimazaki, Y. Hattori, A. Ohya, S. Ohshima, K. Shirota and A. Ozaki, J . Catal., 1986,92 (2), 296. 50 K. Aika, A. Ohya, A. Ozaki, Y. Inoue and I. Yasumori, J. Catal., 1985,92, 305. 51 R. T. Sanderson, Science, 1951,114,670; J. Chem. Ed., 1952,29, 539; 1954,31,2; J. Am. Chem. SOC., 1952, 74, 272. 52 G. Ertl, S. B. Lee and M. Weiss, Surf. Sci., 1982, 114, 527. 53 S. Harris and R. R. Chianelli, J. Calal., 1984, 86, 400. 54 M. J. Ledoux, 0. Michaux, G. Agostini and P. Panissod, J. Catal., 1985, 96, 189. 55 L. D. Marks and V. Heine, J. Catal., 1985, 94, 570; Surf. Sci., 1986, 165, 65. 56 R. B. Grant and R. M. Lambert, J. Catal., 1985,92, 364. 57 D. W. Goodman, Acc. Chem. Res., 1984, 17, 194. 58 E. I. KO and R. J. Madix, Surf. Sci., 1981, 109, 221. 59 T. B. Grimley and M. Torrini, J. Phys. C, 1973,8, 868. 60 E. Einstein and J. R. Schrieffer, Phys. Rev. B, 1973, 7, 3629. 61 J. K. Nerrskov, S. Holloway and N. D. Lang, Surf Sci., 1984, 137, 65. 62 T. N. Rhodin and C. F. Brucher, Solid State Commun., 1977, 23, 275; M. Kiskinova and D. W. Goodman, Surf. Sci., 1981, 108,64; D. W. Goodman and M. Kiskinova, Surf. Sci., 1981, 105, L265. 63 J. M. McLaren, J. P. Pendry and R. W. Joyner, Surf. Sci., 1985, 165, L80. 64 E. Shustorovich, Surf. Sci. Rep., 1986, 6, 1. 65 K. Y. S. Ng and E. Gulari, J. Catal., 1985, 92, 340. 66 P. Arnoldy and J. A. Moulijn, J. Catal., 1985, 93, 38. 67 U. Ozkan and G. L. Schrader, J. Catal., 1985,95, 120; 137, 147. 68 G. W. Keulks, L. D. Krenzke and T. Nottermann, Ado. Catal., 1978, 27, 209. 69 K. F. McCarty, J. W. Anderegg and G. L. Schrader, J. Catal., 1985,93, 375. 70 S. C. Chuang, J. G. Goodwin and I. Wender, J. Catal., 1985,92,416. 71 G. B. McVicker and M. A. Vannice, J. Catal., 1980, 63, 25. 72 Y. Doi, H. Miyake, A. Yokota and K. Soga, J. Catal., 1985, 95, 293. 73 Surface Engineering, ed. R. Kossowsky and S. C. Singhal, Nato ASI Series, Series E, Applied Sciences 74 A. Fersht, Enzyme Structure and Mechanism (Freeman, Reading, MA, 2nd edn, 1985). 75 S. A. Hoenig and F. Tamjidi, J. Catal., 1973, 28, 200. 76 P. M. Stadnik and V. P. Fentisk, Kinet. Catal., 1961, 2, 509. 77 W. W. Lincoln and J. L. Olinger, AIChE. Symp. Ser., 1975,71,77: S . A. Hoenig and J. R. Lane, Surf. 78 K. P. Wisseroth and H. Braune, J. Phys. (Paris), 1977, 38, 1249. No. 85 (Martinus Nighoff, Dordrecht, 1984). Sci., 1968, 11, 163. Paper 6/1912; Received 26th September, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301893

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chern. SOC., Furuduy Trans. I, 1987,83, 1915-1934 (Faraday Symposium 21) Coordination of Carbon Monoxide to Transition-metal Surfaces Rutger A. van Santen Koninklijke/Shell-Laboratorium, Amsterdam (Shell Research B. V.), P.O. Box 3003, 1003 AA Amsterdam, The Netherlands The coordination of CO chemisorbed on a transition-metal surface is a sensitive function of the electronic structure of the surface metal atoms. The group orbital concept appears to provide a key to the understanding of the fundamental electronic features that determine the stability of adsorption complexes. This will be demonstrated using a simple quantum-chemical approach, which is the surface analogue of the Huckel molecular-orbital method. Analysis of chemisorption on the surface of an s-band lattice as a function of band occupation shows the following: multi-atom coordination is favoured at low electron band occupation and single-atom coordination at high electron band occupation for adsorbate orbitals of s-symmetry ; interaction with orbitals ofp-symmetry is only possible at bridging positions and increases with band filling; the effect of changes in surface geometry on chemisorption is a function of band occupation.Chemisorption of CO on platinum is discussed in detail. It is shown that CO prefers coordination to a single atom because of the relative large interaction of the CO 50 orbital with the highly occupied d-valence electron band. The increase in 2n*-occupation, deduced from i.r. studies, for CO adsorbed on a single-atom position with increased Pt surface-atom un- saturation is found to depend critically on the occupation of the Pt surface d-valence electron band.Coadsorption effects of potassium and sulphur are discussed. Sulphur coadsorption induces changes in the electronic structure that can be under- stood on the basis of changes in covalency. Low-coverage alkali-metal coadsorption has two effects on CO chemisorption. Direct interaction with adsorbed alkali metal occurs, resulting in very large decreases in CO frequency and indirect long-range effects occur, resulting in the case of Pt in a shift of CO from a single-atom to a bridge position. The latter effect is calculated to be due to the changed electrostatic potential at the metal surface. The geometry of the adsorption complex formed by chemisorption of an adsorbing molecule or atom on a metal surface plays a significant role in theories that attempt to explain changes in adsorption behaviour caused by promoting ions, adsorbed moderators or alloying metals, which are extensively used in heterogeneous cata1ysis.l Chemisorption of CO has been used widely to study the effect of changes in the electronic structure of the transition metals. Here we will be concerned with the electronic factors that govern the coordination of chemisorbed CO on Group VIII metal surfaces.After having discussed the general theory on a level comparable to that of the Huckel theory, which is often fruitfully used in different areas of chemistry, we will apply the concepts developed to three topics : single-atom versus bridge coordination of CO; the influence of the degree of Pt coordinative unsaturation on the stretching frequency of CO; and coadsorption effects.At low coverage, CO chemisorbs preferentially on one metal atom apical of the Pt(l1 1 ) 2 y and Rh(l1 1)4 surfaces, but it bridges between two 19151916 Coordination of CO atoms at the Ni( 1 1 1)53 and Pd( 1 1 1)’ faces. We will analyse the causes of these differences in terms of changes in surface valence electron density. Recent infrared adsorption experiments by Hayden and coworkerss-10 on CO chemi- sorbed on different faces of Pt show a definite trend of the CO stretching frequency with increased degree of coordinative unsaturation of the surface atoms. In terms of the classical bonding picture of CO proposed by Blyholder,ll the infrared experiments indicate increased back-donation of electrons from the metal into the unoccupied CO 2n* level.A theoretical analysis of these effects will be presented. This is not only of interest as a test of our theoretical model, but also because back-donation of electrons from the metal surface atoms into the unoccupied CO 2n* orbital is one of the factors determining CO dissociation, which is an important issue in CO chemisorption12 and ~ata1ysis.l~ Theoretical ~tudiesl~-~l show that occupation of unoccupied CO 2n* level as well as the strength of the resulting metal-carbon and metal-oxygen bonds play a role. Sulphur adsorption on the Ni( 11 1) face22-25 suppresses chemisorption of CO in the bridge position, but forces CO to be adsorbed on top of the Ni atoms, which behaviour agrees with that expected on the basis of the geometric arguments according to the ensemble concept.26 Also a new weakly bonded form of CO is observed, indicating a change in bonding due to the altered electronic structure around the so-called ligand effect.26 Chemisorption of alkali metal leads to chemical behaviour characteristic of the metal at high alkali-metal coverage;28 at low coverage new phenomena become a p p a ~ e n t .~ * ~ ~ The CO stretching frequency is found to change significantly. Two kinds of shifts are observed. At low CO coverage we observe decreases of the CO stretching frequency of the order of a few hundred wavenumbers, implying significant weakening of the CO bond strength. At the same time the occupied In orbital shifts considerably and the CO molecule tilts towards the ionized alkali-metal atom.At higher CO coverage smaller shifts of the CO stretching frequency are observed. These shifts are typically of the order found when CO changes coordination from single-atom to bridge coordination. Since direct interaction of CO with alkali metal is absent, we have to deal with long-range effects. Previously we demonstrated3’ that two effects play a role. Screening of the positive charge of the alkali-metal ion by metal valence electrons changes the metal surface electron distribution close to the Fermi level, which tends to favour bridge-coordinated CO. However, the long-range electro- static interaction between alkali-metal cation and CO leads to an increase in bond strength, as is experimentally observed.Recent experimental data from ion-scattering experiments confirm the size of the local potential induced by adsorbed alkali Theoretical work from Narrskov and ~ o w o r k e r s ~ ~ - ~ ~ and Wimmer et aZ.42 confirms the presence of a direct electrostatic effect between CO and alkali metal. and N e w n ’ ~ ~ ~ ? 47 Anderson-model calculations on one-dimensional arrays. The adsorbate and metal surface are considered within the linear combination of atomic orbitals (LCAO) or tight binding approximation, including only nearest-neighbouring inter- actions. We use a modified Anderson Hamiltonian, which includes explicitly changes in electrostatic interactions on the adsorbate and between adsorbate and neighbouring surface metal atoms.The semi-infinite face-centred cubic (f.c.c.) metal lattice will be reduced to a Cayley lattice48 using the Bethe approximation. Chemisorption to a transition-metal surface involves interaction with essentially two valence electron bands. The broad spvalence electron band, which contains approxi- mately one electron per atom for a Group VIII metal and the narrower d-valence electron band system, which has varying electron occupation. We have extended the Anderson model to compute the interaction with d electrons The method used is an extension of Grimley and Pisani’s embeddingR. A . van Santen 1917 as well as spelectrons using the tight-binding approximation. In this our method differs from other, semi-empirical treatments, which only account for interaction with the d-bandlg or describe the interaction with the s,p-band assuming free-electron behaviour40- 42 for the latter.An advantage of the second free-electron approach procedure is that it leads to significant simplifications. Effective medium t h e ~ r y ~ ~ - ~ l shows that interaction with the spband can now simply be calculated knowing the local s,p electron density. A disadvantage is that the interference effects due to the position of the surface metal atoms only appear in higher-order corrections to the theory. Since we are interested in coordination energy differences, which we will demonstrate to be a sensitive function of the group orbitalg9 electron density close to the Fermi level, we prefer the LCAO approach for all valence electrons. The method used by us should be considered an approximate iterative Extended Huckel method.The advantage of our method compared to the Extended Huckel method is that closed expressions can be derived for the total energy, which are readily evaluated. The Cayley lattice or Bethe approximation replaces the use of slab calculations. Method When we include one-centre electron-electron repulsion-energy integrals on the adsorbate and two-centre repulsion integrals between the adsorbate orbitals and their neighbouring surface atoms, the Hartree-Fock expression for the adsorption energy 5 0 v 51 ( E is within the metal valence electron-band energy). @ are the discrete roots of det (Oa- EY) outside the valence electron band. U$ds and Uiy are the one-centre repulsion integrals on the adsorbate and the two-centre repulsion integrals between adsorbate and metal surface atoms, respectively, ny and the Mulliken orbital occupation on the adsorbate orbital i and ey the bond order of the bond between adsorbate orbital i and metal surface orbital y ; 0 labels the spin.EF is the Fermi level of the metal and 4 the adsorbate Coulomb energy of orbital i. qa(E), the phase shift, only has a value different from zero as long as E is inside the valence electron band. N; is the orbital occupation before adsorption. The matrix 0" is given by: (2) = ----_ads_- 4 ---_______ [ %;. ads ~ 2 ~ ~ ~ a t t 1 * %ads is the Hamiltonian matrix of the adsorbate and %ads. latt are Hamiltonian matrix elements that couple with the metal lattice orbitals.1918 Coordination of CO Table 1.Values of uy and Zy for some different crystal structures crystal type aya zr simple cubic a, 9 body-centred cubic cc, 16 face-centred cubic am++4/? 16 "am+ is the Coulomb potential, /3 is the overlap energy integral of the metal atomic orbital. pi,, is the overlap energy integral between orbitals i and y. The metal lattice Hamiltonian is calculated with the Bethe lattice approximation, and the different types of metal valence electrons are assumed to form independent bands: ( 4 4 (4b) O,,tt(YY) = 5 - Ayy(E) Ayv(E> = Z s y p2 G y p ) E-a 1 Gyy(E) = J f - d[(E-- aJ2 - 4b,2]. (4c) 2by 2by The minus sign is used when E-ay/2by > 0, the plus sign when E-ay/2by < 0. The parameters ay and b, depend on the connectivity of the lattice formed by the atomic orbitals y.Zsy is the number of neighbouring metal atom orbitals not interacting with the adsorbate, = Zya, and ay and Zy are given in table 1. 2, is the number of bulk metal orbital neighbours minus 1. The indented Greens function Gyy(E)53 has a real and imaginary part for values of E inside the valence electron-band y, otherwise the value of G,,(E) is real. Since H;ds(ii) and H;ds.latt(iy) depend on the electron density, a self-consistent method is used and the solutions are found with the restricted and, if necessary, unrestricted Hartree-Fock method. The electron occupation ny and bond order ey are given by EF 1 fly= [ - - Im [0°-(E+ic)9]iy1dE If det (On - E 9 ) contains discrete roots outside the valence electron band, the contri- bution due to these poles has to be added to the integration over the valence electron energy bands in eqn (5).The electron occupation of the d-band increases moving from left to right in the periodic table; the spband occupancy is chosen as 1 electron per atom. For platinum, palladium and nickel, with a d-band occupation of 9 electrons per atom, the d-band can be considered to consist of two ~ub-bands,~~ uiz. a relatively narrow, completely filled dz2, d3c2-y2 band and a broader, partially filled degenerate dzy, dyz and dZz band (fig. 1). The electron occupation of each of these latter bands is If. In the spirit of the HOMO-LUMO concept interaction between adsorbate valence electrons andE Fig. disti z t 1. Band structure (schematic) of face-centred cubic metal : (a) Bulk valence-electron energy bution, (b) surface valence-electron energy distribution at (100) face, (c) surface d valence-electron energy distribution at (1 1 1) face.U 1.01 Fig. 2. Surface d-electron density of state: (a) dangling bonds of dxy, dyz, d,, orbitals; (b) Local density of states of symmetric (1) and antibonding combinations (2) of dzy, dyz and d,, orbitals, /3” is the overlap energy integral of lobes centred at different atoms. /3 = - 1, 2 = 4, /3” = - 0.5, 2, = 3 (all in eV). metal d-electrons will be considered only with the partially filled dzy, d,, and dzz sub-band system. At the (111) face each dzy, dyz and d,, orbital loses one of its four neighbours [fig. 2 (a)]. The degeneracy at the surface is lifted by interaction of these lobes, each from a different atom.We find one bonding and two antibonding contributions, as sketched in fig. 2(b). Parameters are chosen such that the d-bandwidth and level positions of CO are reproduced. The width of the s-band has been chosen as six times that of the d-band. The interaction parameters between adsorbate and metal surface have been chosen such that bond energies of the proper order of magnitude are calculated. The atomic orbital basis set is assumed to be orthonormal.1920 Coordination of CO General Theory of Adsorbate Coordination to Metal Surfaces As discussed in the introduction, at low coverage CO adsorbs preferentially to a single-atom position on Pt or Rh. This is contrary to our expectations, according to which a chemisorbed molecule would prefer such sites that the coordination number agrees with its van der Waals radius.The phenomenon is due to a quantum-mechanical effect occurring when chemisorption-interaction energies are small compared to the bond strengths of the metal-metal bonds. The bond energy is then dominated by the contribution due to the phase shift. If the combination of each adsorbate orbital to the total bond energy can be considered independent and the interaction between adsorbate and metal surface is weak, AE can be approximated by: AE x Z [k Sf: qf(E) dE- ((q - EF) Nf + 3 UtdS Nf Nf - UFts Nr Nf (3.1 a) i a a' j # i (3.1 b) i Labels the adsorbate molecular orbital and y the metal surface orbitals with which it interacts. Zir is orbital i coordination number, pZd,(EF) the local density of states (LDOS) at the Fermi level of a group orbital formed from orbitals y corresponding to coordination Ziy and the orbital symmetry of orbital i .f & is a factor depending on the relative energetic position of the adsorbate orbital level and the Fermi level. Hence a relationship is found with the LDOS at the Fermi level. Fig. 3 and 4 illustrate this relationship for an H atom adsorbed to an s-band of an f.c.c. crystal (ao - E , = - 9 eV; /3'/j3 = 1, U;fs = 12.6 eV). Fig. 3 shows the LDOS of group orbitals for three-, two- and single-atom coordination as a function of energy. At E = 1, there is one electron per atom in the s-band. For later use also the LDOS of the bridging antisymmetric 1 / 2 / 2 ( 4 , - b2) is shown, b1 and d2 being two s-atomic orbitals of neighbouring metal surface atoms.It should be noted that the maximum of the LDOS for group orbital 1 / ~ ' 3 ( 4 , + $ ~ + 4 ~ ) , dl, 42 and 43 being the metal surface atomic orbitals of surface metal atoms forming a triangle, occurs at the lowest energy, followed by that for 1/d2(41+42), 4, and the one for the antibonding combination. Calculated bond energies for single-atom and tri-coordination are shown in fig. 4 as a function of surface metal valence-electron band occupation, N,,. The curves calculated for single-atom coordination are found to show maximum values, as observed previo~sly.~~ At low electron band occupation, bonding is strongest for hydrogen atoms coordinated to metal atoms with the smallest number of metal neighbours, in agreement with the classical expectation.For higher electron band occupation numbers, the relative order of stabilities is inverted, agreeing with the observed faster decrease in LDOS for low surface metal atom coordination as compared to high surface metal atom coordination (fig. 5). Whereas initially tri-coordination is favoured over single-atom coordination, at high electron occupation numbers single-atom adsorption becomes favoured, in agreement with the behaviour of the corresponding LDOS curves. We may conclude that for s-type orbitals multi-coordination is favoured for low valence-electron occupation, but single-atom coordination becomes favoured at high electron occupation. p-Type orbital interactions become important at high valence- electron band occupations and for s-bands are only possible in bridging coordination sites. We have discussed 54 how these effects can be understood in terms of changes in the bond order between the surface metal atoms as a function of valence- electron band occupation.R.A . van Santen 1921 1 . 2 1 .o 0.8 3 0.6 0.4 0.2 0 n Q - Fig. 3. Group-orbital local density of states of the s-band [p(E)] at the (1 11) face of an f.c.c. crystal: 1, adsorbate orbital of o symmetry, three-coordinated; 2, adsorbate orbital of o symmetry, two-coordinated; 3, adsorbate orbital of o symmetry, single-atom coordinated; 4, adsorbate orbital of 71 symmetry, bridged. E, = 1, electron occupation: 1 electron per atom. - l u 6 U - 1 . 4 0.4 . 0.8 1.2 1.6 2 .o Nd Fig. 4. Interaction energy Eads as a function of N,, for surfaces of an f.c.c.crystal. Numbers in brackets indicate (Ziy, Z8). (-) Restricted Hartree-Fock solution, (---) unrestricted Hartree- Fock solution. a,,- Ei = -9, U,, = 12.6, /I’ = - 1, /I = - 1. 64 FAR 11922 0.8 0.6 n P s 0.4 0.2- 0. Coordination of CO - - - o q I 1 I 6% -d, 1 9 P Fig. 5. Surface local density of states of s metal electron band. Surface metal neighbours Z , = (a) 5 and (b) = 7. Adsorption of CO to the Platinum( 11 1) Face We will describe chemisorption with CO using the Chatt-Dewar model, as originally proposed by Blyholder' for chemisorption to transition-metal surfaces. On the (1 1 1) faces of most of the Group VIII metals CO adsorbs perpendicular to the metal surface, with the carbon atom directed towards the metal. The interaction with the CO molecule can be described as the sum of two contributions. The first is due to overlap of the symmetric highest doubly occupied 5 0 orbital located at the carbon atom.Since this interaction is accompanied by donation of electrons from the 50 orbital into empty metal surface orbitals, this term is called the donating term. The second term is due to overlap of the surface-electron density with the antisymmetric unoccupied 2n* orbitals. This is a back-donating term, since now electrons are transferred from the metal surface orbitals towards antibonding 2n* orbitals. Because of the different symmetries of the 50 and 2n* orbitals of CO, their contribution to the bond energy can be added independently within the tight-binding description of the lattice used in the studies presented here.Previo~sly,~~ we discussed chemisorption of CO to platinum, in which we added the contribution due to each metal valence-electron band independently; here we present results of calculations where the adsorbate orbital interacts with two valence-electron bands, viz. the s valence-electron band and the d valence-electron band, as described in the introduction. Calculated interaction energy diagrams for orbitals 50 and 2n* are shown in fig. 6 for single-atom and bridging positions. The figures give the corresponding group-orbital electron-density distributions before and after adsorption, the bond-order electron-density distributions [pi,@)] between group orbital and adsorbate orbital [crystal orbital overlap population (COOP) according to and the adsorbate electron-density distribution.The results can be interpreted in the manner of Sung and Hoffman21 and ourselves.37 Clearly, bonding and antibonding electron-density contri- butions can be distinguished. For the 50 orbital the bonding contribution is found to be mainly located on the adsorbate, whereas the antibonding contribution shifts to the metal, with a corresponding shift to higher energy of the metal electron density. Metal s- as well as d-electrons are involved. The interaction with the 2n* orbitals shows opposite behaviour. The corresponding metal group-orbital electron density shifts to lower energy, which gives the bonding contribution. The antibonding electron density is now mainly located around the 2n* level. The local electron densities of the adsorbate orbitals are found to broaden with increased coordination.R. A .van Santen 1923 0.25 0.50 0.75 1.00 1.25 r------ 4 t 0 Fig. 6. For 0 legend see page 1924. 64-21924 Coordination of CO 0.0 0.2 0.4 0.6 -12.5 E ( d ) Pr” (El’ 0.6 0.4 0.2 0.0 Pr” (El 0.6 0.4 0.2 0.0 -12.5 1 0 Fig. 6. Changes in electron energy distribution. p$E) is the group-orbital local density of states of symmetry 0 of surface metal orbitals interacting with adsorbate before adsorption, p;(E) is the group-orbital local density of states of symmetry n of surface metal orbitals interacting with adsorbate before adsorption, p;(E)’ is the group-orbital local density of states of symmetry 0 of surface metal orbitals interacting with adsorbate after adsorption, p;(E)’ is the group-orbital local density of states of symmetry n of surface metal orbitals interacting with adsorbate after adsorption, p,(E) is the bond-order energy density between 5 0 orbital and corresponding metal surface group orbitals, p,,(E) is the bond-order energy density between 2n* orbital and corresponding metal surface group orbitals, p,,(E) is the 5 0 adsorbate orbital local density of states, p,,(E) is the 2n* adsorbate orbital local density of states.In relation to curves 1, 2 and 3 solid lines refer to s-electron contributions and dashed lines to d-electron contributions. (a) Single-atom adsorption of CO, interaction between 5 0 orbital and metal surface; (b) bridged adsorption of CO, interaction between 50 orbital and metal surface ; (c) single-atom adsorption of CO, interaction between 2n* orbital and metal surface; (d) bridged adsorption of CO, interaction between 2n* orbital and metal surface.The parameters are the same as in table 2. In table 2 the calculated interaction energies are presented. As can be observed by comparison with the figures, a broadening of the local density of states does not necessarily correspond with an increased bonding energy contribution. The favoured interaction of the 50 orbital with the s- and d-band in the single-atom position stems from the increased group-orbital electron density of states at the Fermi level in the single-atom position, as discussed in the previous section and as can be seen in fig. 6. The question whether single-atom or bridge coordination is favoured is clearly the result of a subtle balance between the 50 single-atom favouring interaction and the 2n* orbital, which favours coordination in the bridging position.We have shown elsewhere3’ that the bonding contribution to the 50 component stems equally from s- and d-electron band contributions. Both favour single-atom coordination, for which the 2n* orbitals only interact with the metal d-band electrons. In bridge coordination there is a significant contribution due to interaction with metal valence s-electrons. For single-atom coordination, the 2n* orbitals of CO do not interact with the metal valence s-electron band, because of the differences in symmetries of the atomic orbitals involved. We have shown earlier that in the bridge position interaction of an antisymmetric orbital with the metal valence s-electron band becomes optimum at high valence-electron band occupation. Clearly, the single-atom position becomes less favoured if: (1) the interaction with the d-valence electron bands decreases; (2) electron back-donation from metal to adsorbate becomes strongly favoured over electron donation from adsorbate to metal; (3) the d-valence electron band significantly depletes, so that interaction with the 50 orbital, too, becomes more favourable for bridge coordination.This is illustratedR. A . van Santen 1925 Table 2. Bond-energy contributionsa (in eV) for CO absorbed to the Pt(ll1) face coordination contribution single-atom bridged 5 0 - 0.442 -0.337 2n* -0.690 - 0.727 Eads - 1.132 - 1.064 Table 3.Bond-energy contributionsa (in eV) for CO adsorbed to the Ni( 11 1) face coordination contribution single-atom bridged 5 0 -0.285 -0.190 2n* -0.395 - 0.805 Eads - 0.680 - 0.995 in table 3 for the chemisorption of Ni. The work function of the Ni( 1 11) surface is 5.1 eV, but that of the Pt( 1 1 1) surface is 5.6 eV. This enhances back-donation to the 27r* orbitals significantly, e.g. with otherwise the same parameters of Pt the interaction of a CO 27r* orbital with Ni increases from - 0.727 to - 2.866 eV (the interaction with the 5 0 orbital changes from -0.337 to -0.287 eV). In order to reproduce bond energies comparable to those found experimentally, the & values have to decrease significantly. On Ni, CO is found to favour bridge coordination because of its much weaker interaction with the d-band and low work-function. Because of the much smaller d bandwidth, the valence d-electron orbitals are less extended and have much less overlap with adsorbate orbitals.This provides an opportunity to at least give a qualitative explanation for the observations that CO favours the single-atom position of the Rh and Pt(ll1) faces, whereas on the Ni and Pd faces it favours bridge coordination. The geometric structure of all these faces is the same and the metal-metal distances are very similar. As can be seen in table 4, however, there is a significant difference in the d bandwidths of these metals. By implication, the interaction with the d-band is largest for Pt and Rh and is significantly less for Ni and Pd. The work-functions of Ni, Pd and Pt are larger than that of Rh, so that the relative stability of single-atom versus bridge-coordinated1926 Coordination of CO Table 4.Metal parameters d-band ionization work- half-widtha potential functiona* metal l e v l e v l e v Ni 1.82 7.6 5.15 Rh 3.80 7.46 4.98 Pd 2.93 8.33 5.12 Pt 4.09 9.0 5.65 a Ref. (14). Polycrystalline. CO is not simply related to a change in work-function as suggested by Whereas in the case of Ni and Pd the decrease in work function and the decreased interaction with the d valence-electron band enhance each other, in the case of Rh the interaction with the d valence-electron band forces CO to be adsorbed on single atoms. In the study of 0, coordination to transition-metal complexes Hoffman et ~ 1 . ~ ~ indicated the importance of including the interaction with the occupied 1n molecular orbitals to determine whether the diatomic molecule is bonded side-on or linearly. Mehandru and have suggested that the interaction with the In orbitals explains coordination of CO parallel with the Cr surface instead of perpendicular to it.We concluded earlier that interaction with an adsorbate orbital of p-symmetry in the bridge position is strongest at high metal valence-electron band occupation. This has been demonstrated explicitly for interaction with an s valence-electron band, but is also true for a molecule in a bridge position interacting with a d valence-electron band. In the bridge position, however, orbitals of s-symmetry have their optimum interaction at low to medium valence-electron band occupation.A In orbital of CO perpendicularly adsorbed has p-symmetry with respect to the metal surface. In parallel coordination it has s-symmetry. Clearly, in bridge coordination the interaction with the In CO orbitals at low to medium valence-electron band filling favours side-on parallel bonding. The symmetry of the interaction with the unoccupied 2n* orbitals is unchanged in parallel or perpendicular coordination. The interaction with the occupied 50 orbital localized on carbon decreases, but increases with the 40 orbital on oxygen. This analysis indicates that the different symmetry of the interaction with the In CO orbitals, in parallel and perpendicular coordination, tends to favour parallel coordination in the bridge position if the d valence-electron band occupation decreases.Bridge coordination will be the favoured coordinations as long as interaction with the metal s valence-electron band remains dominated by 27r* back-donation. Back-donation tends to increase moving from right to left across the transition series, in view of the decreasing work function. Face Dependence Hayden and coworkers8-10 have assigned i.r. frequencies of CO adsorbed in a single-atom manner to the (1 11) face and to corner and edge atoms on the Pt surface. The bands appear at 208 1, 2070 and 2063 cm-l, respectively. Since the change in adsorbate orbital occupation (An,,*) is inversely related to the frequency shift, we can calculate An,,+ and compare its relative size and sign with the shift observed.60~ 61 Since CO is adsorbed on a single atom, in our approximation the effect is only due to interaction of the 27P orbital with the d valence-electron band orbitals.At the (1 11) face the surface valence d orbitals have one neighbour less than in the bulk, two at the corner and three at the edges. Calculated results appear to be extremelyR. A . van Santen 1927 10 6 -4 -21 a I 1 I 1 9.0 9.5 10.0 nd Fig. 7. 2n* orbital population of CO adsorbed on a single atom, at the corner and at the edge of a Pt surface: 0, ratio of corner to face 2n* orbital population (%); 0, ratio of edge to face 2n* orbital population rA); nd is the total d-band electron occupation per metal atom; (-) theoretical, (- - -) experimental Pt. sensitive with respect to the d valence-electron band filling.This is shown in fig. 7, where the orbital occupation has been plotted as a function of the d valence-electron band occupation. Whereas for a d-electron occupation of 9 electrons the 2n* orbital population at the corner is calculated to be 4% higher than for edge adsorption, this difference is found to decrease with increasing d-orbital population. If nd > 9.6 the sequence changes. If nd = 9.8 we even find a 4% lower 2n* orbital population for corner coordination and 1 % lower 2n* orbital population for edge coordination with respect to face coordination. Experimentally, a 0.5 % increase in 2n* orbital population is found at the edge site and 1% increase in 2n* orbital population is found at the corner site compared to the population at the (1 11) face.In fig. 8 the dzy, yz, 5z local density of state is given for the (1 1 1) face and its edge. We observe an increase or decrease in the electron density at the Fermi level depending on whether the energy sampled is at the edge of the d-electron band or at the centre. This is due to the fact that the d-electron band narrows when the number of neighbouring atoms decreases. The increased 2n* orbital populations found experimentally clearly agree with a partially filled d valence-electron band. The low values found for the changes indicate a surface d valence-electron band occupation close to 9.5. If coordination changes from single-atom to bridge at the (1 1 1) face, an increased 2n* population of 10% is calculated, using this d valence-electron band occupation. Recently, Hollins and Pritchard62 reviewed i.r.studies of carbon monoxide chemisorbed on single atoms to different single-crystal faces of copper and observed a trend contrary to that found on platinum. The CO stretching frequency on Cu is found to increase with increasing degree of unsaturation of the metal-surface atoms, indicating a decrease in back-donation of metal electrons into the unoccupied CO 2n* orbital. This agrees with the explanation we propose for the trend found on platinum. In Cu the dvalence-electron1928 Coordination of CO 1.50 .oo n 5 Q I. 50 1.00 - 6 - 2 0 *2 E Fig. 8. dzv, dvz, d,, local density of states p(E) : (a) surface atom (1 1 1) face, (b) edge atom. band is filled and back-donation behaves according to the local density of states at the band edge of the Cu d valence-electron band edge.This decreases with increasing degree of coordinative unsaturation (fig. 7). Coadsorption of Potassium and Sulphur Attractive as well as repulsive interactions occur between chemisorbed molecules or atoms. The forces are attractive and short-range if complex or compound formation occurs. A familiar example is the effect of oxygen on the adsorption of CO, to a base metal face at low temperatures when carbonates are formed.s3 Similarly, it is found that high potassium coverages of copper faces lead to a potassium-CO complex, if exposed to co.28 Long-range attractive forces have been found to be responsible for the formation of ordered adsorbate layers. A good example is the growth of oxygen atom strings perpendicular to the grooves of the Ag(ll0) face.At low coverage the distance between the oxygen strings is large, while at higher coverage the distance between the strings decrease~.~~ This growth pattern can be understood on the basis of electrostatic interaction. Oxygen ions adsorbed in one groove will experience repulsive interactions, similar to parallel dipoles. Oxygen atoms adsorbed in a string where the positive silver atom and negatively charged oxygen atom alternate will attract each other, in a way similar to that of dipoles arranged in a row with their dipole moments in the same direction. So, depending on the relative orientation of the adsorbates, long-range electrostatic interactions may be attractive or repulsive. The presence of long-range electrostatic interactions on a metal surface with low surface coverage of potassium is responsible for the shift of CO from its single-atom position at the (1 11) of platinum to the bridge position, as we will discuss in more detail later.29 Interactions due to changes in the electronic structure at the metal surface induced upon adsorption have also been found to play a role.Such covalent interactions can beR. A . van Santen 1929 repulsive as well as attractive, as has been shown by K o u t e ~ k y , ~ ~ Grimleys69 67 and Einstein and Schrieffer.68 We have shown for a simple model system that the nature of this interaction depends strongly on valence-electron band filling.37 For a partially occupied s-electron valence band and single-atom adsorption of an H-type adsorbate we find a weakening of the bond orders between the atom to which the adsorbate bonds and the nearest metal-atom neighbours of this surface metal atom, an increase in the bond orders between nearest metal neighbours and next-nearest metal neighbours, a decrease between next-nearest neighbour atoms and next-next nearest neighbours, etc.The oscillations show an exponential decrease with distance [fig. 9 (a)]. The change in bond order is a measure of the change in bond strength. In the case of adsorption to a completely filled metal surface s valence-electron band, the bond orders between the surface metal atoms are found to increase. Again this change decreases exponentially with distance. The oscillating behaviour of the bond order with distance observed for a partially filled valence electron band will be reflected in small changes in the metal atom distances, which for a few cases have been observed e~perimentally.~~ This behaviour derives from the increased bandwidth of the local density of states of the electrons on the metal atom to which the adsorbing molecule or atom is chemisorbed.As a result, according to the discussion earlier, bonding to the neighbouring atoms decreases because of the corresponding decrease in the local density of states at the Fermi level. In their turn, the local density states of the next-nearest neighbours will change, etc. The LDOS at the metal atom involved in chemisorption broadens, since its number of neighbours increases. This will decrease the width of the LDOS of its neighbouring metal atoms, since they experience a loss of nearest neighbours, etc.Earlier we used as an example of the dominance of electrostatic forces the oxygen-covered Ag( 1 10) face. Oxygen-atom string formation is accompanied by sharing of silver atoms between two oxygen atoms. Covalency effects will oppose this sharing, because two oxygen atoms now compete for the same silver atom electrons. As follows from the discussion earlier, an effective loss of neighbours may affect single-atom-adsorbed molecules or atoms differently from bridge-coordinated atoms. Table Yo demonstrates this for chemisorption to a metal cluster where the electronic structure changes because of alloying. Replacing a transition metal by Au effectively decreases the number of Pt neighbours, because of the small overlap of the transition metal and gold d valence-electron orbitals. As expected, the bond strength of single- atom-coordinated H atoms increases, but that for bridge-coordinated H atoms decreases.It has been shown by Feibelman and Hamann71 for S adsorbed to the Rh(OO1) face and Wimmer et a1.42 that very little charge transfer occurs from the transition metal to S. Covalent bonding of sulphur to Ni is so strong that the Ni d-orbital becomes strongly antibonding. Experimental studies of the coadsorption of sulphur and CO on iron,72 platinum73 and palladi~m’~ single surfaces indicate the presence of repulsive interactions, leading for low sulphur coverage to a decrease in CO coverage with increasing sulphur coverage that is much faster than predicted on the basis of steric blocking of CO adsorption sites by sulphur.Also, a new bonding CO adsorption state is found. In this state CO is bonded very weakly to single atoms. Feibelman and Hamann7’ as well as Joyner et aZ.759 76 have shown that the change in local density of states extends over several atom distances measured from the sulphur adsorption site. CO coordinated to a single surface metal atom also coordinated directly to sulphur will become more weakly bonded because of a decrease in the LDOS at the Fermi level of s- and partially occupied valence d-electron bands. In addition, the short-range interaction between slightly negative S and CO may be unfavourable. Similarly, for CO adsorbed in bridging coordination, bonding of CO between the metal atoms nearest and1930 Coordination of CO 1 o-8 0 2 4 n Fig.9. Single-atom adsorption on s-band Cayley tree. (a) Change in bond order A P as a function of atom number n from adsorbate-bonded metal surface atom; metal electron band occupation per atom = 1, 2, = 7 , Z = 12. (b) Change in bond order A P as a function of atom number n from adsorbate-bonded metal surface atom; metal electron band occupation per atom = 2 , Z , = 7,5,3; Z = 12, a,, = E,; = /3’ = - 1, no electron repulsion included.R. A . van Santen 1931 Table 5. Differences in bond energy (in eV) of hydrogen atoms bonded to a bimetallic cluster and to a monometallic cluster [extended Huckel calculations, AE(bi) - AE(mono)] no. of single-atom electronsa coordination tri-coordination 0.20 0.12 0.12 0.02 - 0.02 -0.12 -0.31 -0.11 a Number of valence electrons per transition-metal atom.aH = - 10, ad (transition metal) = - 10.5, ad(Au) = - 15, as = -8.5, ap = -4.5, pd = 3, ps = 1.5, pp = 1.83, pH = 1 (p is the orbital exponent). next-nearest to the sulphur atom adsorption site will decrease if covalent effects dominate. This is due to the increase in the effective number of neighbours of the metal atoms sharing the adsorbed sulphur atom. This will lead to a decrease in surface coverage of CO with S much faster than expected on the basis of site blocking. We predict that the bond strength of CO adsorbed on a single transition-metal atom that is the next-nearest neighbour of a metal atom directly coordinated to sulphur may increase as long as direct or electrostatic interaction is absent.The experimentally observed decrease of covalent effects in the order of S, 0 and C are to be ascribed to corresponding decreases of the LDOS at the Fermi level in the same order. Walch and G ~ d d a r d ~ ~ t 78 compute a difference of 1 eV between bond strengths of S or 0 adsorbed in bridging coordination to the Ni( 100) face. Interestingly, the interaction of 0, with the reconstructed Pt(100) face is found to increase upon CO adsorption because the dense (1 11) face reorganizes back to the unreconstructed (100) face upon CO chemis~rption.~~ CO coordinated on single Pt atoms with 8 nearest-neighbour atoms is bonded more strongly than CO adsorbed on a Pt atom that has at least 9 metal neighbour atoms. Apparently this increase in bonding compensates for the loss of the average number of metal-metal atom surface bonds.This is an example of the effect of a favourable covalent interaction, effective because it induces a surface reconstruction, resulting in stronger adsorption. We can also consider this to be due to the weakening of the surface-surface metal bonds upon adsorption of CO. Addition of potassium to the surface also increases the number of neighbours of the surface atoms. As discussed, this tends to decrease the bond strength between surface and coadsorbate. If the difference in electronegativity is large compared to the covalent overlap energy, this decrease, however, will be small and the dominant effect will be the change in the surface electrostatic potential due to the charge on the alkali-metal atom.At low surface coverage (0, < 0.1) potassium atoms adsorbed to the metal atoms can be considered partially ionized, giving rise to a large electrostatic field.38 The resulting electrostatic field around the potassium ion will have two effects. It will induce an attractive electrostatic potential on the electrons of the surrounding Pt atoms as well as of the adsorbate molecules coordinated to the other metal-surface atoms. The effect of an attractive electrostatic potential on the platinum neighbour atom electrons of the potassium ion will enhance the electron density on these Pt atoms, shielding the positive potential of the K ion in the direction of the metal. As follows from the discussion earlier, this tends to increase the LDOS at the Fermi level of antisymmetric group-orbital combinations, but decreases those of the symmetric group-orbital combi-1932 Coordination of CO Fig.10. Bimetallic cluster: 0, transition-metal atom (not gold); 0, gold atom. Table 6. Potassium coadsorption on platinum:a bond energies of co A?'= -0.025 AV = -0.5 contribution single-atom bridged single-atom bridged Sa -0.379 -0.290 -0.330 -0.261 2n* - 0.773 - 0.848 - 0.849 - 0.955 Eads - 1.152 - 1.138 - 1.179 - 1.216 aSame parameters as in table 2, except: a5,,(K) = a5a+AV, %,(K) = %,+AV, G(K) = G + A V , &(K) = &+AV. nations of the Group VIII metals. The result will be a more favoured interaction with the 2n* orbitals of CO, but a decrease in the interaction with the 5 0 orbital of adsorbed CO molecules.We have demonstrated previo~sly~~ that this Ieads to favouring of bridge-coordination of CO with respect to single-atom adsorption of CO to platinum atoms of the (1 11) face covered with potassium (0, < 0.1). However, this effect does not increase the bond strength. Since the d-electron LDOS at the Fermi level decreases with increased electron occupation (see fig. 6), not only does the contribution to the bond strength due to the interaction with the CO 5 0 electrons decrease but also the interaction between metal d-valence electrons and the CO 2n* orbital. The resulting decrease in bond strength is in conflict with experimental results.29 Only when the direct electrostatic interaction between CO and ionized potassium is included do we observe not only a shift from the single-atom to the bridge position, but also an increase in the bond strength of CO.Note that CO still remains coordinated to platinum atoms only. Table 6 shows the effect of the electrostatic potential of an adsorbed potassium ion on the adsorption of CO. Parameters are the same as those used in table 2, except that the Coulomb potentials of Pt and CO have been made more attractive. Again it is observed that the CO molecule has to experience significant electrostatic stabilization to produce an increased chemisorption bond strength. These results agree with the presence of long-range attractive electrostatic effects on the platinum surface due to ionized potassium. report that a direct electro- static interaction between ionized alkali metal and CO has to be postulated to reproduce the experimental results.Calculations by Wimmer et al.42 show that the 2n* orbitals of CO become significantly lowered in the presence of potassium, demonstrating that the electrostatic interaction is indeed important. Ray and Andersonao showed that a lowering of the metal work-function results in increased bridge-coordination of CO. Lowering of the work-function ultimately derives from the positive charge of alkali metal adsorbed onto the metal surface. Narrskov and ~ o w o r k e r s ~ ~ - ~ ~ and White andR. A . van Santen 1933 Conclusions We have presented an analysis of the electronic parameters that determine the coordina- tion of carbon monoxide to a transition-metal surface. (i) The LDOS at the Fermi level of group orbitals plays a crucial role.(ii) We have found that an adsorbate s-type orbital prefers multi-coordination at low valence-electron band filling, but single-atom coordination for high valence-electron band filling. An opposite trend holds for adsorbate p-type orbitals. (iii) Increased unsaturation of surface atoms results in stronger chemisorption bonds, except when the metal Fermi level is at the edges of the valence-electron band. (iv) The preferred single-atom coordination of CO on the (1 1 1) faces of Pt and Rh is ascribed to their relatively high work-function in combination with the strong interaction with the highly occupied metal valence-electron d-bands. (v) There is a close balance between the interaction of the CO 2n* orbitals, which favour bridge coordination, and the CO 5 0 orbital, which favours single-atom coordination.(vi) Coadsorbates may influence CO chemisorption significantly. (vii) Long-range effects due to electrostatic, as well as covalent, interactions have been discussed. Covalent effects usually behave in a periodic manner as a function of distance and show an overall exponential decline. Electrostatic interactions are often important. Charges are shielded by the metal electrons, but cause significant external potentials that modify the position of adsorbate orbitals with respect to the metal Fermi level. I thank Mrs L. Claassens for her skilful programming assistance and R. Nada from the University of Turin for assistance with the calculations. References 1 See for instance E. K. Rideal, Concepts in Catalysis (Academic Press, New York, 1968).2 H. Froitzheim, H. Ibach and S. Lehwald, Appl. Phys., 1977, 13, 147. 3 A. M. Bar0 and H. Ibach, J. Chem. Phys., 1979,71,4812. 4 L. H. Dubois and G. A. Somorjai, Surf. Sci., 1980, 91, 514. 5 S. Anderson, Solid State Commun., 1976, 20, 229. 6 W. Erley, H. Wanger and H. Ibach, Surf. Sci., 1979, 80, 612. 7 A. Bradshaw and F. Hoffman, Surf. Sci., 1978, 72, 513. 8 R. G. Greenler, K. D. Burch, K. Kretrschmar, R. Klauser, A. M. Bradshaw and B. E. Hayden, Surf. 9 B. E. Hayden, K. Kretrschmar and A. P. Bradshaw, Surf: Sci., 1985, 155, 553. Sci., 1985, 152/153, 338. 10 B. E. Hayden, K. Kretrschmar, A. M. Bradshaw and R. G. Greenler, Surf. Sci., 1985, 149, 394. 11 G. Blyholder, J. Phys. Chem., 1964, 68, 2772. 12 G. Brodeis, T.N. Rhodin, C. Brucker, R. Benbow and Z . Hurych, Surf. Sci., 1976, 59, 593. 13 P. Biloen, W. M. H. Sachtler, Adv. Catal., 1981,30, 165. 14 W. Andreoni and C. M. Varma, Phys. Rev. B, 1981, 23, 437. 15 A. B. Anderson, J. Chern. Phys., 1976, 64, 4046. 16 J. N. Allison and W. A. Goddard 111, Surf. Sci., 1981, 110, L615. 17 G. Doyen and G. Ertl, Surf. Sci., 1977, 69, 157. 18 P. S. Bagus and K. Kermann, Phys. Rev. B, 1975,6, 4195. 19 J. W. Davenport, Phys. Rev. Lett., 1976, 36, 945. 20 D. W. Bullett and M. L. Cohen, J. Phys. C , 1977, 10, 2101. 21 S-S. Sung and R. Hoffman, J. Am. Chem. Soc., 1985, 107, 578. 22 W. Erley and H. Wagner, J. Catal., 1978, 53, 287. 23 D. W. Goodman and M. Kiskinova, Surf: Sci., 1981, 105, L265. 24 R. J. Madix, S. B. Lee and M. Thornburg, Surf.Sci., 1983, 133, L441. 25 J. L. Gland, R. J. Madix, R. W. McCabe and C. DiMaggio, Surf. Sci., 1984, 143, 46. 26 W. M. H. Sachtler, Vide, 1983, 164, 67. 27 M. Trenary, K. J. Uram and J. T. Yates, Surf. Sci., 1985, 157, 512. 28 D. Lackey, M. Surman, S. Jacobs, D. Gnder and D. A. King, Surf. Sci., 1985, 152/153, 513. 29 J. E. Crowell, E. L. Garfunkel and G. A. Somorjai, Surf. Sci., 1982, 121, 303. 30 M. Kishinova, G. Pirug and H. P. Bonzel, Surf. Sci., 1983, 133, 321. 31 H. S. Luftman, Y-M. Sun and J. M. White, Surf. Sci., 1984, 141, 82. 32 H. S. Luftman and J. M. White, Surf. Sci., 1984, 139, 369.1934 Coordination of CO 33 F. M. Hoffman, J. Hobek and R. A. dePaola, Chem. Phys. Lett., 1984, 106, 83. 34 L. J. Whitman and W. Ho, J. Chem. Phys., 1985,83,4808.35 W. Eberhardt, F. M. Hoffman, R. de Paola, D. Heskett, I. Strathy, E. W. Plummer and H. R. Moser, 36 R. A. de Paola, J. Hobek and F. M. Hoffman, J. Chem. Phys., 1985,82, 2484. 37 R. A. van Santen, in Proc. 8th Int. Congr. Catal. (Springer-Verlag, Berlin, 1984), vol. IV, p. 97. 38 J. J. C. Geerlings and J. Los, Phys. Lett. A , 1984, 102, 204. 39 J. K. Nerrskov, S. Holloway and N. D. Lang, Surf. Sci., 1984, 137, 65. 40 J. K. Nerrskov, Physica, 1984, 127B, 193. 41 N. D. Lang, S. Holloway and J. K. Nerrskov, Surf. Sci., 1985, 150, 24. 42 E. Wimmer, C. L. Fu and A. J. Freeman, Phys. Rev. Lett., 1985, 55, 2618. 43 T. B. Grimley and C. Pisani, J . Phys. C, 1974,7, 2831. 44 C. Pisani, Phys. Rev. B, 1978, 17, 3143. 45 R. A. van Santen and L. H. Toneman, Int. J. Quantum Chem., 1977, 12, suppl. 2, 83. 46 D. M. Newns, Phys. Rev. B, 1969, 178, 1123. 47 G. P. Muscat and D. M. Newns, Progr. Surf. Sci., 1978, 9, 1. 48 R. Haydock, V. Heine and M. G. Kelly, J. Phys. C, 1975, 8, 2591. 49 M. G. Kelley, Surf Sci., 1974, 43, 587. 50 J. Gallaway, P k s . Rev. B, 1971, 83, 2556. 51 M. Seel, G. Del Re and J. Ladik, J . Comput. Chem., 1982, 3, 451. 52 J. A. Pople and D. Beveridge, Approximate Molecular Orbital Theory (McGraw-Hill, New York, 53 D. Kalkstein and P. Soven, Surf. Sci., 1971, 26, 85. 54 R. A. van Santen, Reel. Trav. Chim. Pays-Bas, 1982, 101, 121. 55 F. Cyrot-Lackman, M. C. Desjonqueres and J. P. Gaspard, J. Phys. C, 1974, 7, 925. 56 J-Y. Saillard and R. Hoffman, J. Am. Chem. Soc., 1984, 106, 2006. 57 A. B. Anderson and Md. K. Awad, J. Am. Chem. Soc., 1985, 107, 7854. 58 R. Hoffman, M. M-L. Chen and D. L. Thorn, Inorg. Chem., 1977, 16, 503. 59 S. P. Mehandru and A. B. Anderson, unpublished results. 60 D. Post, Ph.D. Thesis (Free University, Amsterdam, 1981). 61 D. Post and E. J. Baerends, J. Chem. Phys., 1983, 78, 5663. 62 P. Hollins and J. Pritchard, Progr. Surf. Sci., 1985, 19, 275. 63 C. Backx, C. P. M. de Groot, P. Biloen and W. M. H. Sachtler, Surf. Sci., 1983, 128, 81. 64 H. A. Engelhardt and D. Menzel, Surf. Sci., 1976, 57, 591. 65 J. Koutecky, Trans. Faraday Soc., 1958, 54, 1038. 66 T. B. Grimley, Proc. Phys. Soc. (London), 1967, 92, 776. 67 T. B. Grimley and S. M. Walker, Surf. Sci., 1969, 14, 395. 68 T. L. Einstein and J. R. Schrieffer, Phys. Rev. B, 1973, 7, 3629. 69 J. E. Inglesfield, Progr. Surf. Sci., 1985, 20, 105. 70 R. A. van Santen and W. M. H. Sachtler, Surf. Sci., 1977,63, 358. 71 P. J. Feibelman and D. R. Hamann, Phys. Rev. Lett., 1984,52, 61. 72 J. Benziger and R. J. Madix, Surf. Sci., 1980, 94, 119. 73 E. L. Garfunkel, M. H. Farias and G. A. Somorjai, J. Am. Chem. Soc., 1985, 107, 349. 74 S. W. Jorgensen and R. J. Madix, Surf. Sci., 1985, 163, 19. 75 R. W. Joyner, J. B. Pendry, D. K. Saldin and S. R. Tennison, Surf. Sci., 1984, 138, 84. 76 R. W. Joyner, P. Meehan, J. M. Maclaven and J. B. Pendry, J. Appl. Catal., in press. 77 S. P. Walch and W. A. Goddard 111, Surf. Sci., 1978, 72, 645. 78 S. P. Walch and W. A. Goddard 111, Surf. Sci., 1978, 75, 609. 79 G. Ertl, Surf. Sci., 1985, 152/153, 328. 80 N. K. Ray and A. B. Anderson, Surf. Sci., 1983, 125, 803. Phys. Rev. Lett., 1985, 54, 1856. 1970). Paper 6/1081; Received 19th May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301915

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1987, 83, 1935-1943 (Faraday Symposium 21) Electrostatic Interactions and their Role in Coadsorption Phenomena Stephen Holloway* Donnan Laboratories, University of Liverpool, P.O. Box 147, Liverpool L69 36X Jens K. Nerskov-f. Haldor Topsoe Research Laboratories, Nymdevej, DK-2800 Lyngby, Denmark Norton D. Lang IBM Thomas J. Watson Research Center, Yorktown Heights, New York 10.598, U.S.A. A theoretical framework for comparing binding energies and activation energies for adsorption on surfaces is presented. It is applied to a microscopic description of the mechanisms underlying the promotion of the molecular adsorption process on metal surfaces by coadsorbed species. Estimates of the range of this effect, its magnitude and the important factors which differentiate promotional efficiency within a particular chemical group are discussed.It is shown that while correlating surface activity with the macroscopic work function may give useful trends, it is in fact the local derivative of the electrostatic potential with respect to the surface normal which governs the free energy difference when a molecule is adsorbed in the presence or absence of a promoter. During the last two decades the ‘surface-science’ approach has evolved to become a powerful and informative means for studying heterogeneous catalysis. By using ultra- high-vacuum (u.h.v.) tools, such as LEED, UPS and EELS, it is possible to characterize accurately both the surface and any adsorbates present.’ More recently it has become possible to study some dynamic aspects of surface processes using time-of-flight measurements coupled with laser- and electron-induced fluorescence.2 The philosophy behind such experimental (and theoretical) studies is that the catalytic process may be divided up into several elementary reaction steps : (i) adsorption, (ii) surface diffusion, (iii) molecular rearrangements and (iv) desorption of products. Clearly the relevance of u.h.v.measurements has to be established for each system being studied, as has already been done in some cases.3* Results for some model reactions, including the ammonia synthesis on iron and the hydrogenation of CO on nickel, illustrate this in particular. The theoretical approach to these problems involves the study of the electronic structure of the substrate, the adsorbate and the combined system, and the resulting energetics.Although detailed self-consistent chemisorption calculations are beginning to a ~ p e a r , ~ it will be a long time before such methods can be directly applied to ‘real’ catalytic systems. In the place of such detailed calculations, and indeed to supplement them once they appear, we need simple, physically sound schemes which reflect the wisdom obtained from the more ‘ first-principles ’ methods. Theoretical treatments of adsorbate-adsorbate interactions, starting with the work of Grimley and Walker6 and Einstein and Schrieffer,’ have concentrated mainly on the indirect interaction via the surface conduction electrons. For large adsorbateadsorbate separations this interaction is oscillatory in both magnitude and sign.This has been used to explain the occurrence of different open-ordered structures. More recently, Feibelman Also at NORDITA, Blegdamsvej 17, DK-2100, Copenhagen, Denmark. 19351936 Electrostatic Interactions and Coadsorption and Hamanna and Joyner et aL9 have calculated the change in the surface one-electron density of states due to the adsorption of an electropositive or electronegative atom. Such changes are assumed to affect the interaction of an adsorbing atom or molecule with the surface, and the lateral range of these changes is taken as a measure of the range of the interactions. The direct interaction due to an overlap between orbitals on different adsorbates has also been studied.1°9 l1 At very short distances this is dominated by the kinetic-energy cost of orthogonalizing the orbitals to one another.It is this, for instance, that constrains the fractional saturation coverage in a single layer to be of the order of unity. Finally, the direct electrostatic interaction has been treated.12-14 At intermediate distances of the order of a surface lattice constant (2-3 A), this interaction can give rise to substantial (0.1 eV) interaction energies, when both adsorbates in question induce electron transfer to or from the surface or have a large internal electron transfer. The effective medium theory15 provides a simple way of classifying these various contributions to the adsorbate-adsorbate interaction energy. Assuming that the influence of a nearby adsorbate A, on the adsorbate A, in question can be treated to first-order in the close vicinity of A, and vice versa close to A,, it can be shown within the local-density approximation that the energy of A, will be changed by the presence of A, by the amount15 6E = S#L-“)(r) n,(r) dr + 6 (r nu(&) E dc) .I -a (1) The first term in eqn (1) is simply the direct electrostatic interaction, with being the electrostatic potential induced by A,, and nu@) the charge density induced by A,. Here and throughout, the term ‘induced’ is taken to mean the difference between the surface with an adsorbate and the surface without an adsorbate. The integral is over the near-adsorbate (A,) region ‘a’ only, and the superscript (-a) on &b0 indicates that only charges outside region ‘a’ are to be included in calculating the electrostatic potential.In most cases of interest, where the direct overlap between the two interacting adsorbates is small, the exclusion of region ‘a’ in S&j-@ is of no importance. The second term in eqn (1) is the difference in the sum of one-electron energies with and without A, present, with na(c), the A,-induced density of states, as weighting function, and cf the Fermi energy. The difference must be evaluated taking the potential inside region ‘a’ to be unaltered (expect for a constant shift) by the introduction of A,. This term contains both the direct (orbital overlap) and indirect electronic adsorbate- adsorbate interactions. Note that to the extent that the effective medium result eqn (1) holds, it is the one-electron energy sum which determines the indirect and direct (orbital overlap) interaction energies.This has been used by Muscatls to study the interactions in ordered hydrogen overlayers on transition-metal surfaces. In the next section a brief description of the way in which eqn (1) may be applied to the study of the coadsorption of gasses and electropositive species is presented. Attention will be focused on particular classes of systems which are relevant to heterogeneous catalysis. The effects of such interactions will be presented in terms of modifications to the potential-energy (hyper) surfaces which ultimately govern both the dynamics and the kinetics of the reaction. Finally in the last section conclusions will be presented. Coadsorption Studies In terms of eqn (l), a coadsorbed atom can change the interaction energy of the adsorbing molecule with the surface in two ways.The coadsorbate-induced electrostatic potential 6#o will interact with the molecule. This is described by the first term. The second term describes both the direct interaction due to an overlap between the molecule and coadsorbate orbitals as well as the indirect one mediated through the surfaceS . Holloway, J . K . Nsrskov and N . D. Lang 1937 1 .o I 1 1 2 s" 1 (0 -0.5 - -1.0 - K I 1 I a -1.5 -2 0 2 4 6 z/bohr Fig. 1. The self-consistently calculated electrostatic potential due to an S and a K atom chemisorbed at the calculated equilibrium distance (1.9 and 4.0 bohr, respectively) outside a jellium (rs = 2) surface. The potential is that felt by an electron at a lateral distance of 5 bohr from the atom and it is shown as a function of distance outside the jellium edge, which is half an interlayer spacing outside the first metal layer.In order to give an idea of the range of the charge transfer to a chemisorbed molecule, note that a CO molecule at its equilibrium position outside an Ni(100) surface occupies the region from ca. 2 to 4 b o b in the figure. From ref. (13). 0.8 0.6 0.4 2 0.2 \ s" w 0.0 - 0.6 L -6 - 4 - 2 0 2 4 6 z/bohr Fig. 2. The electrostatic potential due to an 0 atom located at distances (a) 1.0, (b) -0.5 and (c) - 1 .O bohr outside a jellium (r, = 2) surface. The potential is shown as a function of distance from the surface 4.5 bohr away from the atom. Although the resulting dipole moments corresponding to these three situations are - 1.6, - 0.2 and 0.1 debye, respectively, the figure shows quite clearly that for coadsorbed molecules which donate charge to the substrate the overall change in binding energy of a coadsorbate due to the 0 will be similar in each case.This implies that although certain coverages of preadsorbed electronegative atoms may have positive or negative work-function changes,8 the effect on reactivity need not show opposite behaviour in the two limits. This is yet another manifestation of the fact that simple, dipolar representations of surface charge distri- butions are quite inadequate at the microscopic level.1938 Electrostatic Interactions and Coadsorption Table 1. Dipole moment, p, as a function of metal-oxygen distance, d (measured relative to the positive background edge), for a high density (r, = 2) substrate (the sign of the dipole moment is defined so that a negative moment corresponds to an increase in the substrate work function) distance from the jellium dipole edge/ bohr momen t/D - 1.0 - +0.1 - 0.5 - 0.2 0.0 -0.7 0.5 - 1.1 1 .o - 1.6 In what follows, only the first term will be considered. This is not, of course, meant to imply that the second term is unimportant but that the essential experimental obser- vations may be rationalized on the basis of the behaviour of the former.In fig. 1 the self-consistently screened electrostatic potential due to an electropositive and an electronegative atom chemisorbed on a jellium surface is shown.l39 l4 It should be stressed that while it is not intended to suggest that jellium would be a useful catalyst( !), it has the virtue that it is the simplest model of a metal which treats screening self-consistently, a feature of paramount importance when discussing electrostatic interactions.It is clear that a molecule like CO, which extracts electrons from the surface (into the antibonding 27r* orbital), will be stabilized by nearby electropositive coadsorbates and destabilized by electronegative coadsorbates. Molecules like NH, or H,O which, when adsorbed, have a large internal charge transfer towards the surface will react in an opposite way. It must be stressed that even in cases where electronegative coadsorbed atoms like 0 and S are chemisorbed so close to the metal that the induced dipole moment is small or even positive they will give rise to a destabilization of a molecule like CO.This is a reflection of the importance of a proper self-consistent description of the charge transfer. A simple dipole model of the charge transfer is thus not adequate in these cases.l3? l4 This is exemplified in fig. 2, where the self-consistently screened electrostatic potential for an oxygen atom is plotted for different values of the surface-oxygen bond distance. For the three geometries shown, the values of the total dipole moment are presented in table 1, indicating a reversal of the direction of the apparent charge transfer! Interestingly enough though, irrespective of the sign of the dipole moment (and for a monolayer, the macroscopic work-function) it will always cost energy to transfer charge from the metal to a neighbouring coadsorbate, since in each case the sign of the electricJieId (i.e.the slope of these curves) is the same in the relevant region outside the metal surface. Thus a molecule like CO which accepts charge from the metal will be destabilized in all cases, as shown in fig. 2. Extreme care must be taken in correlating changes in reactivity with trends in work-function measurements for electronegative adsorbates since dipolar fields are simply not a good enough representation of the self-consistent electrostatic potential in the immediate vicinity of the surface. If the energy required to reorient, e.g. an H20 molecule on the clean surface is smaller than the gain in electrostatic energy due to the interaction with a nearby alkali-metal atom, the alkali-metal atom could induce a reorientation of the molecule, and the total interaction binding energy could actually increase. Indications of such an effect for H20 on metal surfaces in the presence of coadsorbed alkali metals have been f0und.l' OrderingS .Holloway, J . K . Norskov and N . D. Lang 1939 n n cs K 0 2 4 6 0 10 alkali metal/mmol (g-cat1-l Fig. 3. Rate of ammonia synthesis on Ru-active carbon-alkali-metal catalyst as a function of alkali-metal ~ 0 n t e n t . l ~ and tilting of polar molecules in the presence of electropositive and electronegative coadsorbates is observed quite generally1* and can be understood in the same terms. In fig. 3 are shown some experimental data for the change in reactivity for the production of ammonia with increasing amounts of different alkali-metal additive.lg It shows quite clearly that the promoting effect increases with atomic number. In fig.4 are contour plots of the electrostatic potential induced by Li, Na and K at their calculated equilibrium positions outside a jellium (r, = 2) surface, and in fig. 5 we show the variation perpendicular to the surface at a typical nearest-neighbour distance ( 5 bohr) from the alkali-metal. The increasing depth of the electrostatic potential [and from eqn (1) the concomitant increase in binding] with atomic number is clearly seen. This is due to the fact that while the amount of charge transferred in each is similar, the distance over which it is transferred is greater for the heavier ones since their valence shells are larger.In order to quantify these arguments, the evaluation of eqn (1) requires both a knowledge of ddo as well as the induced density of states for the adsorbed CO molecule. As an example, for CO at the equilibrium position outside an Fe( 110) surface, work- function measurements and calculations suggest charge transfers between 0.2 and 0.6 l4 For an electrostatic potential like that shown for K in fig. 1, this corresponds to an interaction energy of between 0.1 and 0.3 eV, close to the experi- mentally determined one of 0.17 eV.4 The range of this interaction is given by the range of the induced electrostatic potentials. While at 4 A this could not be classified as being long-range, on certain crystal faces it can encompass up to third-nearest neighbours. To1940 Electrostatic Interact ions and Coadso rpt ion + Fig.4. Contour maps of the induced electrostatic potential for (a) Li, (b) Na and (c) K at their equilibrium distances outside a jellium (r, = 2) surface. The vertical line denotes the jellium edge and the crosses the atomic positions. The vacuum side is to the right. Contour values of f 2 , f 1, k0.5, k0.3 and kO.1 eV are shown within a sphere centred at the atomic position and having a radius of 7 bohr. From ref. (22).1940 Electrostatic Interact ions and Coadso rpt ion + Fig. 4. Contour maps of the induced electrostatic potential for (a) Li, (b) Na and (c) K at their equilibrium distances outside a jellium (r, = 2) surface. The vertical line denotes the jellium edge and the crosses the atomic positions. The vacuum side is to the right.Contour values of f 2 , f 1, k0.5, k0.3 and kO.1 eV are shown within a sphere centred at the atomic position and having a radius of 7 bohr. From ref. (22).1942 Electrostatic Interact ions and Coadsorption Fig. 7. Superposition of the self-consistent electrostatic potentials for isolated K and 0 outside a jellium (r, = 2) surface. This serves to illustrate that although in actual NH, catalysts both species are present, the promoting effect of potassium dominates over any poisoning due to the oxygen. illustrate this, fig. 6 shows schematically the potential-energy turning-surface? for the interaction of H, with an f.c.c. (100) surface. The diagram shows 9 unit cells for the surface which, in the absence of any impurity species, has an activation barrier for adsorption.Shown, however, is the situation when a preadsorbed alkali-metal atom is present. This has been modelled by superposing the induced electron density for an adsorbed K atom with overlapping metal atoms, and assuming a linear relationship between it and the repulsive energy experienced by the Clearly, although the site containing the alkali metal is still not active for the dissociation reaction, the four neighbouring sites all contain regions of reduced activation energy (shown in black). This has the implication that those molecules impinging upon this localized areal around the alkali metal could dissociatively adsorb, whereas those impacting elsewhere would be scattered back into the gas phase.20 The underlying principle to emerge from these calculations is that certain regions of the molecule-metal potential-energy surface are shifted in energy owing to the presence of a preadsorbed atom.This differential shifting will be most pronounced at larger distances from the surface since metallic screening is least efficient in this region. In terms of the overall kinetics for adsorption, these modifications can be described as a (de)stabilization of the molecular precursor. This picture of the promoting effect of K on N,/Fe(l 1 1) has recently been confirmed experimentally by Whitman et a1.21 It also forms the basis for the successful modelling of the NH, kinetics over K-promoted Fe catalysts by Stoltze and N ~ r s k o v . ~ ~ It is interesting to note that the electrostatic potential around the preadsorbed atom can also explain the decrease in vibrational frequency observed for CO on alkali- metal-dosed surfaces.23 The negative potential (as measured by a negative test charge) arising from the alkali metal gives rise to an effective downshift of the molecular CO levels, which results in an increased occupancy of the partly filled 2n* antibonding level.This in turn reduces the intramolecular vibrational frequency.24 Estimates of the magnitude of such shifts have been made and they are in the region 50-200cm-l, depending critically upon such factors as surface geometry and adsorbate coverage. 25 In those cases where molecule-surface binding energies are of the same order or even weaker t The classical turning surface is defined as the locus of heights above a surface at which a normally incident, monoenergetic beam of molecules would be scattered. 1 It would also be possible for molecules to access this region of space if trapping into a precursor state followed by surface diffusion were possible.S. Holloway, J .K . Nmskov and N . D . Lang 1943 than the ‘chemical ’ interaction between coadsorbed species, then surface complexing will occur.23c This has been reported for the coadsorption of K-CO/Cu( 1 10) where frequency shifts of 750 cm-l have been associated with the formation of a squarate complex upon the surface. Finally, on perhaps a more speculative note, it is suspected that on a working catalyst, for, e.g. NH, production, there may be 0 present in connection with the K promoter.We expect that while this will reduce the exact number of sites promoted, it will not destroy the entire effect. This is illustrated in fig. 7, which gives a superposition of the electrostatic potentials due to K and 0 placed at their respective equilibrium separations at a lateral distance corresponding to the near-neighbour separation on an Ni( 100) surface. It is clearly seen that the range and absolute magnitude of the K-induced electrostatic potential is so much larger than for 0 that the effect due to the latter is expected to be small. Conclusions In this paper we have discussed how electrostatic interactions between coadsorbed species can give rise to significant modifications in the system energetics. The promoting effect of electropositive atoms on the adsorption of molecules like 0,, CO and N, and the increasing effect through the alkali metals have been explained.In addition an expla- nation of the modification of the vibrational frequency shift for strongly adsorbing molecules has been presented. Quantitative estimates based upon the model appear to be in good agreement with recent experimental values as demonstrated by Uram et a1.26 in their extensive analysis of the vibrational properties of K and CO coadsorbed on N P . References 1 See, e.g. G. Ertl and J. Kuppers, in Low Energy Electrons and Surface Chemistry (Verlag Chimie, 2 See, e.g. B. Kasemo and B. I. Lundqvist, Europhys. News, 1982, 5, 9. 3 D. W. Goodman, R. D. Kelly, T. E. Madey and J. T. Yates, J.Catal., 1980, 63, 226. 4 G. Ertl, S. B. Lee and M. Weiss, Surf. Sci., 1982, 114, 527. 5 N. D. Lang, in Theory of the Inhomogeneous Electron Gus, ed. S . Lundquist and N. H. March (Plenum 6 T. B. Grimley and S. M. Walker, Surf. Sci., 1969, 14, 395. 7 T. L. Einstein and J. R. Schneffer, Phys. Rev. B, 1973,7, 3629. 8 P. J. Feibelman and D. R. Hamann, Phys. Rev. Lett., 1984,52, 61. 9 R. W. Joyner, J. B. Pendry, D. K. Saldin and S. R. Tennison, Surf. Sci., 1984, 138, 84. Weinheim, 1974). Press, New York, 1983), pp. 309-391. 10 Y. Muda and T. Hanawa, Jpn J. Appl. Phys., 1974, 13,930. 1 1 J. Benziger and R. J. Madix, Surf. Sci., 1980, 94, 1 19. 12 K. H. Lau and W. Kohn, Surf. Sci., 1977,65, 607. 13 J. K. Nsrskov, S. Holloway and N. D. Lang, Surf. Sci., 1984, 137, 65. 14 N. D. Lang, S. Holloway and J. K. Nsrskov, SurJ Sci., 1985, 150, 24. 15 J. K. Nsrskov, Phys. Rev. B, 1982, 26, 2875. 16 J. P. Muscat, Surf. Sci., 1981, 110, 85. 17 D. L. Doering, S. Semancik and T. E. Madey, Surf Sci., 1983, 133,49. 18 C. Benndorf and T. E. Madey, Chem. Phys. Lett., 1983, 101, 59. 19 A. Ozaki and K. Aika, in Catalysis, ed. J. R. Anderson and M. Boudart (Springer, Berlin, 1981), vol. 1, p. 87. 20 M. Karikorpi, S. Holloway, N. Henrikson and J. K. Nsrskov, Surf. Sci., 1987, 179, L41. 21 L. J. Whitman, C. E. Bartosch, W. Ho, G. Strasser and M. Grunze, Phys. Rev. Lett., 1986, 56, 1984. 22 P. Stoltze and J. K. Nsrskov, Phys. Rev. Lett., 1986, 55, 2502. 23 ( a ) J. E. Crowell, E. L. Garfunkel and G. A. Somorjai, Surf. Sci., 1982, 121, 303; (b) L. Wallden, Surf. Sci., 1983,134, L513; (c) D. Lackey, M. Surman, S. Jacobs, D. Grider and D. A. King, Surf. Sci., 1985, 1521153, 51 3. 24 S. Holloway and J. K. Nsrskov, J. Electroanal. Chem., 1984, 161, 193. 25 J. K. Nsrskov, S. Holloway and N. D. Lang, J. Vac. Sci. Technol., 1985, A3, 1668. 26 K. J. Uram, L. Ng and J. T. Yates Jr, Surf. Sci., 1987, in press. Paper 61 1863; Received 20th August, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301935

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1987,83, 1945-1962 (Faraday Symposium 21) Catalytic Implications of Local Electronic Interactions between Carbon Monoxide and Coadsorbed Promoters on Nickel Surfaces James M. MacLaren, Dimitri D. Vvedensky and John B. Pendry The Blackett Laboratory, Imperial College, London S W7 2BZ Richard W. Joyner B. P . Research Centre, Chertsey Road, Sunbury-on- Thames, Middlesex TW16 7LN Recent experimental and theoretical work has highlighted the electronic factors that influence the interactions between CO and coadsorbed species on transition-metal surfaces. Although the Blyholder model and modifica- tions thereof provide a useful framework for interpreting many of these results, the microscopic mechanisms by which coadsorbates either poison or promote catalytic reactions are still not well understood.The development of theoretical approaches that can account for established experimental data, while having the flexibility to explore possible reaction paths is therefore an important step in being able to design more efficient catalysts. We present multiple scattering-Xa and Green function density of states calculations for Ni( 100) and Ni( 1 1 1) with CO and coadsorbed poisons (e.g. sulphur) and promoters (e.g. lithium potassium) for the methanation reaction. For the promoted Ni( 100) surface we find that the polarisation of the alkali-metal charge towards the surface induces a decrease in the work function and a large electrostatic shift of the CO levels to greater binding energies, which increases the occupancy of the 2n* orbital.There is furthermore evidence of direct CO-promoter interaction in the In and 2n* levels, although the 5 0 appears to be largely unaffected by the alkali metal. The implications of these calculations for the promotion of the methanation reaction by nickel catalysts are discussed. For sulphur on Ni( 1 1 1) we find an oscillatory interaction, with some sites within the range of poisoning showing an enhanced local density of states. This is not observed on the (100) surface and is attributed to the absence of a CO adsorption site at the appropriate distance from the sulphur atom. The modification of catalysts by coadsorbed adatoms is a well known and important effect. Industrial catalysts often include small amounts of preadsorbed electropositive atoms on the surface to increase either the reaction rate or the selectivity towards the desired products.In contrast, adsorbates such as electronegative atoms, which may appear as impurities in the catalyst or reactants, tend to poison reactions and reduce the rate of product formation. In general, for reactions that involve a dissociation step, such as the methanation of CO, alkali metals increase and electronegative adsorbates decrease the molecular dissociation rate over many transition meta1s.l For example, on a silver surface chlorine acts to increase the product yield in the selective oxidation of ethylene to ethylene oxide,2 while in Fischer-Tropsch syntheses, chlorine is recognised to be a catalyst p ~ i s o n . ~ A reaction which has been studied extensively both on commercial and single-crystal catalysts is CO + 3H2 CH, + HZO.Ni 19451946 Electronic Interactions between CO and Promoters on Ni co co Ni gas phase adsorbed Fig. 1. Schematic representation of the Blyholder model of CO chemisorption, showing how the metal substrate facilitates transitions from the 50 to the 2n*. The open circles and small arrows indicate unoccupied and occupied electronic spin-orbitals, respectively. The hopping rates for 50 -+ Ef and Ef --+ 2n* are proportional to the density of metal states at the Fermi level, according to Fermi's Golden Rule. Many theoretical and experimental studies of chemisorbed CO on a variety of poisoned, promoted and clean surfaces have been performed using many u.h.v. analytical techniques such as LEED, HREELS, ARUPS and AES, thus making this reaction ideal to study.The influence of alkali metals on the methanation reaction is summarised as follows: (1) increasing the CO binding energy, as seen in the increase in the thermal desorption temperature of C0,4 (2) lowering of the CO stretch freq~ency,~ indicative of increased back-donation from the catalyst to the 2 ~ * , ~ (3) increasing the dissociation rate over many transition-metal catalysts7v and (4) slightly decreasing the methanation rate, and increasing the yield of higher weight hydrocarbons, consistent with increased surface carbide concentration.8 The influence of the electronegative poisons, on the other hand, can be summarised as (1) a decrease in the CO binding energy, as seen in the decrease in the thermal desorption temperature of C039 9 9 lo and (2) a dramatic decrease in the methanation rate, and the ability of the surface to chemisorb C0.39 lo l2 have indicated that there is a correlation between the electronic structure around the molecular adsorption site and catalytic activity.Specific emphasis has been placed on the Fermi-level density of states, which for CO adsorption can be easily illustrated within the Blyholder m0de1.l~ Chemisorption and dissociation are related to charge transfer from the 5 0 to the 2n* orbital. The influence of the metal catalyst, typically with a large density of states at the Fermi level, is to provide an easier route via a two-step model (fig. 1). In particular, in a single-electron picture, the electronic structure of the metal appears, simply through Fermi's Golden Rule, as the product of the local density of states at the Fermi level and the matrix elements for the transitions 5 0 -, Ef and Ef + 2n*.Increased hopping rates therefore decrease the lifetimes for these states, particularly for the 2n* because of its close proximity to the Fermi level. This results in a broad molecular 2n* resonance, accounting for the partial occupancy of this state for chemisorbed CO. More detailed model Hamiltonians such as the Newns-Anderson mode1,14 and generalisations for the two-level system of CO by Doyen and Ert1,15 also involve the Fermi level density of states in the broadening of the molecular levels, and hence in the occupancy of molecular resonances. The argument given above provides some justification for the role of electronic structure in catalysis.However, this is clearly an oversimplification, for although chemisorption, vibrational modes and hence energy exchange with the surface can be Previous theoreticalJ. M . MacLaren, D. D . Vvedensky, J. B. Pendry and R. W. Joyner 1947 understood in terms of perturbations in the local electronic structure, the lack of knowledge of the microscopic details of catalytic reaction pathways makes detailed comparison to changes observed in product-yield difficult to quantify. Bearing these theoretical and experimental observations in mind, we address our calculations to the influence of adsorbates on the surface electronic structure of nickel and on CO chemisorption. Other factors affecting catalytic activity have been suggested, such as the role of electrostatics on the promoted surfaces,lS charge mobilityl7? la and the role of the In level on both the clean and promoted l9 Electrostatic interactions and their influence on chemisorbed molecules has been considered in some detail by Lang, Holloway and Nsrskov using the effective medium theory of Nsrskov and Lang.20 Changes in catalytic activity as a consequence of coadsorbed atoms are correlated to changes in the electrostatic potential.In this model the action of promoters is solely to reduce and poisons to increase the electrostatic potential near to the equilibrium site for molecular adsorption. The influence of charge mobility has been touched upon by Heine and Marks,17 who propose that surfaces which reconstruct easily, such as Ni( loo), should also be catalytically active.The more mobile the charge at the Fermi level, the more easily the surface is able to respond to the presence of reaction species, both in static properties (e.g. chemisorption), and in the dynamics (e.g. the normal vibrational modes of the chemisorbed species, and its surface diffusion), a point which has also been stressed by Haydock.ls Edmonds and McCarrol121 have indeed observed that both poisons and promoters can induce surface reconstructions. Whether the role of the modifier is structural as well as electronic is not yet clear, although changes in local electronic structure may well cause local modifications to surface structure. Method The calculations presented here are based on two models.The first involves calculating the local electronic structure at the adsorption site in terms of the local density of states (LDOS), p(r, E), evaluated at a representative point near this site. The Green function G+(r, r, E ) is calculated for a cluster of ca. 30 atoms in a single-centre angular-momentum representation about the adsorption site, from which the LDOS is obtained via 1 p(r, E ) = -- Im G+(r, r, E ) 71 where a convenient representation of G+ is The calculations were checked for convergence in cluster size and in angular-momentum components. Since the calculation is made at complex energy, with the imaginary part of the constant potential of 1 eV, it is equivalent to calculating the LDOS for an extended surface via a band-structure calculation broadened by the appropriate amount.However, inclusion of an adsorptive part to the energy, the self-energy, models the lifetime effects in a solid which would be observed in a photoemission experiment for example. The calculation is non-self-consistent, and as such can only be applied to systems with small charge transfer, such as sulphur on nickel. A fuller discussion of this model and the choice of substrate and adsorbate potentials can be found in ref. (1 1). The second approach uses the multiple-scattering-Xa (MS-Xa) scheme of Johnson and Slater22 to solve self-consistently for the energy levels of a molecular cluster of atoms chosen to model the local environment of CO with the coadsorbates on Ni( 100) (fig. 2).23 The cluster is divided into three separate regions (see fig.3 for the free CO molecule as1948 Electronic Interactions between CO and Promoters on Ni coadsorba t e @ first-layer nickel 0 second-layer nickel n cross-section I cross-section I1 Fig. 2. Top view of the cluster used in the MS-Xa calculations for carbon monoxide coadsorbed with poisons and promoters on nickel (100) and the cross-sections used for displaying the wavefunctions in fig. 6, 10 and 12-14. I11 Fig. 3. The MS-Xa cluster for the free CO molecule illustrating the partitioning of space between intrasphere (I), intersphere (11) and extramolecular regions (111).J. M . MacLaren, D. D . Vvedensky, J. B. Pendry and R. W. Joyner 1949 an example): (1) an atomic region, inside the atomic muffin tins centred on the constituent atoms; (2) an intersphere region, between each atomic sphere and the surrounding outer sphere and (3) an extramolecular region, outside the sphere that surrounds the complex.The potential is constructed by spherically averaging in regions 1 and 3, and by volume-averaging in region 2. Contiguous muffin-tin spheres are used, although there is a small overlap between the lithium and the CO owing to the large sphere size for the lithium atoms. Thus solutions to the Schrodinger equation in regions 1 and 3 are derived by direct numerical integration of the radial equation, while between the atomic spheres the constant potential facilitates an analytic rapidly convergent multicentre partial-wave expansion for the wavefunction. The eigenstates are then found by matching the logarithmic derivative of the intersphere wavefunction to that of the radial solutions at the sphere boundaries.The inclusion of self-consistency allows us to study the influence of promoters upon CO chemisorption where charge transfer is expected to be important. To illustrate in detail the chemical bonding we show wavefunction contour plots of the salient orbitals in CO chemisorption for the two cross-sections displayed in fig. 2. Previous studies with the MS-Xa method by Balazs et aZ.24 established a correlation between the relative rates of certain reductive reactions and bond strengths, as evidenced by the occupancy of antibonding orbitals, Although we have not performed a total energy analysis of our systems, Danese and C ~ n n o l l y ~ ~ have demonstrated the feasibility of such calculations within the MS-Xa framework, even for diatomic molecules, by including non-muffin-tin corrections to the charge density. Moreover, Case et aZ.26 have used the MS-Xa wavefunctions for accurate calculations of a variety of one-electron properties, including dipole and quadrupole moments, diamagnetic susceptibility and nuclear quadrupole coupling constants. These results indicate that the MS-Xa wave- functions are a suitable basis for the analysis of CO-adsorbate interactions on transition- metal surfaces.Results and Discussions Sulphur as a Promoter? In this section we summarise the results and conclusions of a study of sulphur adsorbed on Ni( 1 1 l).27 Based on the model of CO chemisorption outlined in the introduction, an increase in the LDOS at the Fermi level should promote CO adsorption.Fig. 4 and 5 show the cluster used to investigate sulphur on Ni(ll1) and the results obtained for three sulphur configurations. While no reports exist of the promotion of CO adsorption by sulphur, it is interesting to compare our results with those of Trenary et aL2* These workers have shown that the influence of adsorbed sulphur on carbon monoxide adsorption on nickel (1 1 1) extends only to the ‘near’ and ‘far’ sites of fig. 4. As noted previously, there is a very strong suppression of the density of states at the Fermi level, by sulphur in the ‘near’ sites. Sulphur in the ‘remote’ sites has little influence, but an interesting result is observed for the ‘far’ sites. Here the density of states close to the Fermi level is enhanced, which within our model would imply promotion of CO adsorption. This finding is discussed in more detail el~ewhere,~~ but we note in passing that the CO-sulphur distance in this site is intermediate between that of the ‘near’ and ‘far’ sites previously examined on Ni( 100)’ and again indicates the short range over which adatoms act, in agreement with our overall conclusions. Trenary et aZ.have also observed a new, weakly held, ‘atop’ CO state which they designate CO*, and which is induced by the presence of sulphur. The CO stretch frequency is 2105 cm-l, compared to 1910 cm-l observed on the clean (1 11) surface. They suggest that, since the influence of sulphur is short-range, their CO* sites are the same as our ‘near’ sites, whilst the excluded sites are the same as the ‘far’ sites of fig.4. We believe that the situation is the reverse of that suggested by Trenary et aZ. The1950 Electronic Interactions between CO and Promoters on Ni 0 0 0 0 top-layer second- layer Ni remote S far S near S Ni( 1 1 1) cluster nnn Fig. 4. Top view of cluster used in calculating the local density of states (LDOS) for the central nickel adsorption site on the (1 11) face, showing the near, far and remote sulphur adsorption sites. The LDOS is calculated at a height of 1.23 A above the central nickel atom, with the sulphur placed at a height of 1.64 A above the top plane of nickel atoms. Fig. 5. The DOS for the cluster shown in fig. 3, for the clean surface (-), and ldr sulphur adsorbed in the near (-0-), far (---) and remote (-*-) hollow sites.near sites are very strongly poisoned, the LDOS as shown in fig. 4 being appreciably reduced by the presence of sulphur. Also sulphur atoms in the three-fold hollow sites would not permit occupancy of the nearby atop site and could be expected to push the CO away, at least to one of the two adjacent bridge sites. By contrast, the ‘far’ sites are not sterically hindered and our calculations show that they are close enough to the sulphur atoms to be influenced. Alkali-metal Promoters In this section we will discuss in detail the influence of the alkali metals on adsorbed CO on Ni(100), and make comparisons with some of the experimental and theoreticalFig. 6. Wavefunction contour plots the molecular and chemisorbed carbon monoxide levels 5 0 (a,b), 172 (c,dj and 2n* ( e , f ) , in cross- section I1 of fig.5. Contour values are 0.08 1,0.027,0.009,0.003 and 0.001. Solid and broken lines represent positive and negative values of the wavefunction, respectively. Atomic positions are marked by the small dark circles.1952 Electronic Interactions between CO and Promoters on Ni all levels > 5 % on central Ni +0.2 I - x d x -0.2 2 s \ -0.4 (b ) all levels > 5 % on central Ni t0.2 E€ -0.2 -0.4 -0.6 -0.0 (c) all levels > 5 % on central Ni = 2r' ~ -5a - -40 Fig. 7. MS-Xa orbital energy levels of A,(a) and E(n) symmetry for the entire cluster and those with appreciably weight on the central nickel atom for carbon monoxide adsorbed on a clean nickel (100) surface, and with coadsorbed sulphur and lithium in the geometry of fig.2. To facilitate direct comparison, the Fermi levels of all three systems have been aligned. (a) Ni/S/CO, (b) Ni/Li/CO and (c) Ni/CO. co Ni poison f 50 co Ni promoter Fig. 8. Schematic illustration of the influence of poisons and promoters on the Blyholder model of carbon monoxide chemisorption (cf. fig. 1). The open circles and small arrows indicate unoccupied and occupied electronic spin-orbitals, respectively. According to the indicated direction of charge flow, the poison decreases and the promoter increases the density of substrate states at the Fermi level. results for CO on promoted surfaces. All of the calculations in this section are based on the MS-Xa scheme for the cluster shown in fig.2. The adsorbates sulphur and lithium are placed in the four-fold hollow sites with the vertical spacing from the first nickel plane for sulphur of 1.3 A, as obtained by LEED measurements on a c(2 x 2) sulphur overlayer on Ni( As there are no data available for the lithium nickel bond length,J. M . MacLaren, D. D . Vvedensky, J. B. Pendry and R. W. Joyner 1953 we used a value obtained by scaling from the sodium-nickel bond length determined by LEED for sodium on Ni(100),30 which gives a nickel-lithium layer spacing of 1.86 A. Carbon monoxide is known, from LEED ~tudies,~' to sit on the atop site, with the nickekarbon bond length of 1.7 A and the CO bond length the same as in the gas phase, 1.15 A. While the gross features of the interaction between the CO and the coadsorbates are seen to be repulsive for the poison and attractive for the promoter, a closer examination of the electronic structure reveals the differences to be more subtle.Therefore we will briefly review the results for the system Ni/S/CO and Ni/CO and then contrast these with a more detailed look at Ni/Li/CO. A preliminary report of this work has appeared elsewhere. 23 To illustrate the CO-metal bonding for the clean surface we plot in fig. 6 the molecular orbitals 5n, In and 2n* for the gas-phase CO molecule and chemisorbed CO on Ni(100). The interaction is not strong, as can be seen from the small changes observed in the orbitals, and charge transfer occurs from the 5 0 and In to the metal and from the metal to the 2n*. As expected, adsorbed sulphur decreases the density of levels on the central nickel atom throughout the band.Fig. 7 plots the orbital energies of the A,(o) and E(n) irreducible representations of the C,, point group, and those orbitals with weight > 5% on the central nickel atom, which represents in a qualitative way the local density of states at the molecular adsorption site. The position of the 2n* relative to the Fermi level has changed only slightly and is still predominantly unoccupied. For both the clean and poisoned surfaces the 2n* is broadened owing to the mixing with the metal states. The cooperative effects of the 5 0 and the 2n*, mediated by the metal substrate, provide the chemisorption of CO to the surface, consistent with the Blyholder rnode1,l3 as shown schematically in fig.8. In fig. 9 we plot the 2n* and the 50 orbitals in cross-section (I) for the clean and poisoned surface. For the Ni/CO 50, charge is placed between the central nickel atom and the carbon atom, through orbital mixing between the 5 0 and an s-d orbital on this nickel atom. There is also weak bonding to the other nickel atoms in the cluster. The 2n* is a spatially diffuse orbital, which is antibonding between the CO and the central nickel atom, and to the other nickel atoms. Differences between the clean and poisoned 2n* are fairly small and are repulsive in nature. The influence of sulphur is seen to be destabilising and most evident in the 50. This is schematically shown in fig. 10, where the sulphur p orbitals draw charge from the CO-metal bond via the substrate to themselves, inducing a rehydridisation from s-p to p-d on the central nickel atom.This results in the additional nodal line in the associated contour plot in fig. 9. Similar behaviour, involving charge redistribution as related to rehydridisation, has also been noted recently in Green's-function calculations of the local density of states for sulphur on Ni( 1 1 l).27 This charge rearrangement evidently involves nickel atoms in the second layer owing to their proximity to the sulphur atoms in the hollow sites. The importance of the second layer in sulphur poisoning has also been noted in recent calculations of MacLaren et al.," where the absence of these atoms increases the strength and range of poisoning. The first study presented here for the promoted surface is for the system Ni/K/CO.The geometry used is slightly different to that used in the cluster shown in fig. 2, in that the potassium atoms are placed on the atop sites surrounding the CO molecule, with a nickel-potassium bond length of 3.14 A. This arrangement was chosen to facilitate a direct comparison to the FLAPW slab calculation of Wimmer et aZ.32 in fig. 11. We see that for fairly large changes in the electronic structure observed in the promoted system, the cluster chosen accurately reproduces the slab calculations. This point is important for three reasons: (1) the extended nature of the surface is not necessary to reproduce the gross features of chemisorption of CO on the promoted surface; (2) cluster 65 FAR 11954 Electronic Interactions between CO and Promoters on Ni (b' Fig.9. Wavefunction contour plots in cross-section I1 of fig. 5 for the Ni/CO and Ni/S/CO 5 0 (a, b) and 2n* (c, 6) chemisorbed levels. The contour values are the same as in fig. 6. calculations are particularly convenient for studying the molecular adsorption site be- cause of the flexibility afforded by not being limited to long range periodicity and ( 3 ) the MS-Xa scheme is numerically fast, and thus allows the study of low-symmetry clusters (e.g. in the final calculation discussed below, the symmetry of the cluster is lowered from C4v to CJ. The results for Ni/K/CO and Ni/CO are compared to those of Wimmer et al. by aligning the Fermi levels and marking the positions of the MS-Xa levels for the chemisorbed CO molecule against the partial density of states in the carbon muffin tin (fig.11). In both the clean and the promoted surfaces the position of the 40, 50 and In agree quite well with their positions in the partial density of states. The position and width of the broadened 2n* level seen in the slab calculation agrees well with the set of levels characteristic of the 2n* resonance seen in the MS-Xa results. The potassium- covered surface shifts all the chemisorbed levels to lower energy with respect to the Fermi level, owing to an electrostatic interaction caused by charge transfer from the potassium atoms to the substrate. All the metal states are raised, including the Fermi level, shown schematically in fig. 8. This effectively drives the molecular levels down with respect to the conduction states and in particular with respect to the Fermi level.Some evidence of this purely electrostatic effect is seen in the position of the CO 2n* resonance relativeJ . M . MacLaren, D . D . Vuedensky, J. B. Pendry and R. W. Joyner 1955 Fig. 10. Schematic illustration of the effects of fig. 9 showing (a) the charge flow and (b) the resulting rehybridisation induced by coadsorbed sulphur on the carbon monoxide 5a and the central nickel orbitals. energy/eV Fig. 11. Direct comparison of the MS-Xa orbital energies and the partial density of states in the carbon muffin tin calculated by Wimmer et al.32 for (a) Ni/CO and (b) Ni/K/CO. Long and short vertical bars indicate positions of MS-Xaenergy levels with weight w > lo%, and 1 % < w < 10% in the carbon sphere, respectively.After aligning the Fermi levels the positions and the widths of the FLAPW 40, 5t7, In and 2n* agree remarkably well with the MS-Xa calculation. 65-21956 Electronic Interactions between CO and Promoters on Ni Fig. 12. Wavefunction contour plots using the values of fig. 6 for (a) Ni/CO 5 0 in cross-section I, (b) Ni/Li/CO 5 0 in cross-section I, (c) Ni/CO 5 0 in cross-section I1 and (d) Ni/Li/CO 5 0 in cross-section 11. to the la, in the carbon Kedge NEXAFS results, which remains the same for clean and promoted surfaces on several metals.33 The systems Ni/CO and Ni/Li/CO will form the basis of the rest of the discussion of the promoters. Fig. 7 shows the MS-Xa levels, of the A,(a) and E(n) irreducible representations of the C,, point group, with weight > 5% on the central nickel atom.In contrast to the Ni/S/CO system, where there is depletion of levels near the Fermi level, lithium increases the density of levels near the Fermi level, since the partially occupied lithium s-p states are in this energy range. There is again an electrostatic interaction owing to the polarisation of the lithium states as in the Ni/K/CO system illustrated by fig. 8, and a direct interaction between the lithium atoms and the chemisorbed CO, causing the broadening of the 2n* level. The direct interaction can be seen more clearly from figs. 12, 13 and 14, which respectively show the contour plots of the 50, In and 2n* levels for these two systems. Fig. 12 shows cross-section (I) for the 5 0 for the systems Ni/Li/CO and Ni/CO.Compared with the clean Ni/CO. The lithium 2s orbitals mix and interact weakly with the 50, drawing charge towards the lithium atoms, since in cross-section (11) (also seen in fig. 12) there is relatively little change in the 50 uponJ. M . MacLaren, D . D. Vvedensky, J. B. Pendry and R. W. Joyner 1957 Fig. 13. Wavefunction contour plots using the values of fig. 6 for (a) Ni/CO la in cross-section I, (b) Ni/Li/CO la in cross-section I, (c) Ni/CO la in cross-section I1 and (6) Ni/Li/CO la in cross-section 11. lithium adsorption. The interaction is weak and localised around the lithium atoms, because of the large orbital electronegativity difference between the lithium 2s states and the 50. On the promoted surface there is a strong interaction between the 2n*, the Fermi-level states, and the lithium s-p-like states due to their proximity in energy, which leads to charge polarisation towards the metal.The 2n* is a spatially diffuse orbital in both clean and promoted cases and is antibonding between the carbon and oxygen and central nickel atoms, but weakly bonding to the other nickel atoms in the top layer. Any rearrangement of the charge in the 2n* must involve a compensating effect on the occupied orbitals due to screening. The effect is most pronounced on the l n (fig. 13), despite it being energetically as far away from these lithium states as the 50, owing to the constraint of orthogonality. In fig. 13 there is some mixing of the lithium s states and 2n* in this orbital, which results in charge transfer from the metal and the CO bond to the oxygen end of the molecule, and to the lithium atoms.This is turn results in much less weight of the orbital on nickel atoms, apart from the central nickel. The interaction is stronger and less localised than the influence of lithium on the 5 0 , as can been seen1958 Electronic Interactions between CO and Promoters on Ni Fig. 14. Wavefunction contour plots using the values of fig. 6 for (a) Ni/CO 2n* in cross-section I, (6) Ni/Li/CO 2z* in cross-section I, (c) Ni/CO 2n* in cross-section I1 and (d) Ni/Li/CO 2n* in cross-section 11. from cross-section (11). The orbital is still noticeably perturbed from the clean Ni/CO llt and showing the same general features as in section (I). Correspondingly, fig. 14 shows the 2n*, and the compensating effects with a mixing of lithium p states and some llt character.This results in charge transfer in the opposite sense to that observed in the ln, with charge drawn from the lithium and the 0 end of the CO molecule to CO bond and the metal CO region, shown schematically in fig. 15. The direct interaction of lithium and the rehybridisation induced has been discussed in some detail above. Our model for the interaction of the promoters is, therefore, a local one with electrostatic and direct chemical interactions, for the reasons discussed above both are important for partially occupying the 2n* through back-donation of charge to the region of space between the carbon and oxygen atoms. The combined role of the llt and the 2lt* in the promoted system has been studied in some detail, and shown to be more important than the clean surface, a point also raised by Plummer et aL59 l9 in the interpretation of their U.P.S.results for Cu/K/CO. In contrast, the sulphur atoms interact with the 5 0 level directly, destabilising molecular adsorption, and indirectly by depleting states near the Fermi level, thus inhibiting transitions to the 2n*. Even in theJ. M . MacLaren, D. D . Vvedensky, J. B. Pendry and R . W. Joyner 1959 I N 2n" Fig. 15. Schematic illustration of the effects of fig. 13 and 14 showing the influence of lithium on the n bonding. Arrows indicate the direction of charge flow from the promoter to the metal for the 2n* and a compensating flow from the substrate to the lithium in the 1n due to screening and orthogonalisation.case of the promoted CO, the degree of occupancy of the CO 27r* does not appear large enough to suggest molecular dissociation in the upright configuration. Since dissociation of CO, like that of N,, is an activated process, it is important to consider possible channels of energy exchange with the surface. For example the frustrated rotational mode can couple to surface phonons, a point we will address in the next section. An alternative reaction pathway would involve tipping the CO away from the vertical, as a route to CO dissociation. As results for the electronic structure for CO chemisorbed at various angles of tilt will be presented in detail elsewhere, we report here only the conclusions of that work.The chemisorbed levels of CO remain relatively unaltered in energy as the molecule is tipped to 45", from which we conclude that the CO chemisorption bond is spatially local, and that the charge density, particularly near the Fermi level, is fairly mobile for this type of perturbation. The energy barrier to CO tipping is therefore expected to be small, with the frustrated rotational mode having a comparatively low frequency and a large amplitude. The value for the frequency of 41 1 cm-l for CO bonded atop on Ni34 indeed confirms these observations. Calculations of the total energy as a function of this tilt angle were performed by Allison and G ~ d d a r d , ~ ~ who concluded that the minimum in the total energy was very shallow and centred for CO in the upright configuration.The influence of sulphur and lithium on this mode will also be the subject of a later paper. Discussion : Relation to Catalysis The above analysis reveals that the influence of the surface alkali metal on the electronic structure of chemisorbed carbon monoxide can be summarised as follows. (1) a shift to larger binding energy of the CO chemisorbed levels, owing to a decrease in the work1960 Electronic Interactions between CO and Promoters on Ni function caused by polarisation of charge from promoter towards the substrate; (2) a direct interaction of promoter s-p states with the C end of the 2n* and (3) a compensating interaction with the 0 end of the ln, resulting in a weakening of the carbon oxygen bond. As a result, the promoter makes dissociative adsorption easier on those metals where the surface carbide is This conclusion is supported by the study by Campbell and Goodman of methanation on single nickel crystals,8 where the coverage of the surface carbide was higher in the potassium-promoted than the unpromoted case, other things being equal.Since the surface carbide is closely related to the polymerising species, the increase in selectivity to longer-chain hydrocarbons than methane can be understood. It is less easy to account for the observed overall decrease in the rate of CO conversion. There is a significant reason for arguing that the weakening of the carbon-oxygen bond is only a partial explanation of the role of the promoter. It is important to recognise that carbon monoxide dissociation is not rate determining on the unpromoted nickel metal.This is shown by the lack of structure sensitivity of the reaction, which has been noted in single-crystal and supported-catalyst studies.8 This contrasts with clear evidence that CO dissociation is structure sensitive, occuring to a very limited extent on the close-packed (1 1 1) plane, and much more rapidly on stepped surfaces.3s These observations for carbon monoxide adsorption on nickel closely parallel those of Ertl and his colleagues on the interaction of alkali metals with nitrogen on the iron single-crystal planes. Dissociative adsorption of nitrogen on iron surfaces is an activated process,39 and Ertl et al. showed that the presence of alkali metals lowers this activation barrier significantly.Recent evidence suggests that, in this case also, the role of the promoter may be complex. Kinetic analysis by Bowker et al.40 has indicated that, under ammonia synthesis conditions, the barrier to nitrogen dissociation is insignificant compared to that involved in hydrogenation of the adsorbed nitrogenous species. The reduction in the barrier to dissociative adsorption is inadequate to explain the promotional influence of the alkali metals on ammonia synthesis on iron catalysts. Bowker et al. argue that the heat of adsorption of the reactive nitrogen adspecies is much lower than that observed in the static experiments of Ertl et al., and we may speculate that the promoter has some role in this decrease of the nitrogen binding energy at high coverage. An alternative analysis of the Ertl data has been proposed by Stolze and N~~rskov,~' but the situation remains to be resolved.There are additional reasons for considering that the role of the alkali metals in carbon monoxide adsorption and catalytic chemistry is complex. Evidence for the magnitude of the weakening of the carbon-oxygen bond by the promoter is provided by vibrational- spectroscopy studies. The extent of this weakening, however, is very variable, ranging from 120 cm-l on the Pt( 11 1)/K to over 700 cm-l on copper promoted by and results from an increased occupancy of the chemisorbed 2n* on the promoted surface. Using effective medium theory, Holloway and N I Z I ~ S ~ O V ~ ~ have shown that the electrostatic part of the promoter/CO interaction can account for a weakening of ca.120 cm-l. It is generally assumed that shifts of a greater magnitude indicate a tipping of the molecule away from the vertical or the change in the adsorption site. The formation of a squarate species, K+(CO),, has been suggested on copper surfaces.43 One possible role of the alkali-metal promotion in methanol synthesis could be to tip the adsorbed CO species towards the surface. This could be significant if methoxide, which is known to bind to the metal through the oxygen atom, is a reaction intermediate. It would be particularly interesting to understand the role of the promoter in methanol synthesis, since the promotional effect is much greater than the change noted in ammonia synthesis on iron catalysts or in methanati~n.~~ Our calculations show, however, that little change in the electronic state of the adsorbed CO molecule on tipping the CO away from the vertical plane.This is perhaps not surprising given the soft nature and large amplitude of the metal-CO wag mode, This mode will be important for the exchangeJ. M . MacLaren, D . D. Vvedensky, J. B. Pendry and R. W. Joyner 1961 of energy between the surface and the molecule since its energy is low enough to couple to the surface phonons. The anharmonicity in the CO metal potential allows coupling between this frustrated rotational mode and the CO stretch frequency, observable in the spectroscopic studies of Hoffman et al.46 and by Harris et aL4' in a temperature-dependent lineshape in the CO stretch mode. The coupling to the surface phonons is more effective for CO bonded to the bridge rather than the top sites, because of the lower frequency of this adsorption geometry.Other models for energy exchange between substrate and adsorbed molecules involve the creation of electron-hole pairs, a point discussed by Persson and Ryberg for CO on Ni( The electron-hole excitation will be most effective when the 2n* resonance is close to the Fermi as on the promoted surface, and consequently less important for clean and sulphur covered surfaces. The CO rotational mode neither shifts the position nor changes the occupancy of the 2n* level greatly, so energy-damping via surface phonons will be more likely than the creation of electron-hole pairs. The changes in bonding of the molecular carbon monoxide state brought about by the alkali-metal promoter, which have been the subject of the calculations presented here are insufficient to explain all of the changes noted in catalysis. Additional roles for the promoter may be to modify the stability of catalytic intermediates, to mediate the reactivity of chemisorbed hydrogen, or to change the surface diffusion parameters of reactants, intermediates or products. It is not clear which of these is of dominant importance, and we intend to explore some of these questions in subsequent work.and Persson and Persson for CO on Cu( J.M.M. thanks British Petroleum for their support. We thank Dr M. E. Eberhart and Prof. K. H. Johnson for making an updated version of the MS-Xa programs available to us, and Prof. K. Baberschke for helpful and stimulating discussions. Permission to publish this paper has been given by the British Petroleum Company PLC.References 1 G. Broden, T. N. Rhodin, C. Brucker, R. Benbow and Z. Hurych, Surf. Sci., 1976,59,593; H. P. Bonze1 2 M. Boudart, in Interactions on Metal Surfaces, ed. R. Gomer (Springer-Verlag, Berlin, 1975), p. 292. 3 M. P. Kiskinova and D. W. Goodman, Surf. Sci., 1981, 108, 64. 4 M. P. Kiskinova, Surf. Sci., 1981, 111, 584. 5 W. Eberhardt, F. M. Hoffman, R. dePaola, D. Heskett, I. Strathy, E. W. Plummer, H. R. Mosen, 6 J. Lee, C. P. Hanrahan, J. Arias, R. M. Martin and H. Metiu, Phys. Rev. Left., 1983, 51, 1803. 7 J. W. Erickson and P. J. Estrup, Surf. Sci., 1986, 167, 519. 8 C. T. Campbell and D. W. Goodman, Surf. Sci., 1981, 108, 519; D. W. Goodman, in Role of Poison and Promoters in CO Hydrogenation, in Proc.IUCCP Symp., Texas, 1984; D. W. Goodman, Acc. Chem. Res., 1984, 17, 194. 9 J. Benziger and R. J. Madix, Surf Sci., 1980, 94, 119; M. Kiskinova and D. W. Goodman, Surf. Sci., 1981,109, L555; J. L. Gland, R. J. Madix, R. W. McCabe and C. deMaggio, Surf. Sci., 1984,143,46; R. J. Madix, M. Thornburg and S. B. Lee, Surf. Sci., 1983, 135, L447. 10 E. L. Hardegree, P. Ho and J. M. White, Surf Sci., 1986,165,488; D. W. Goodman and K. Kiskinova, Surf Sci., 1978, 105, L265. 11 J. M. MacLaren, J. B. Pendry, R. W. Joyner and P. Meehan, Surf. Sci., in press. 12 P. J. Feibelman and D. R. Hamann, Phys. Rev. Lett., 1984,52,61; P. J. Feibelman and D. R. Hamann, 13 G. Blyholder, J. Phys. Chem., 1964, 68, 2772. 14 D.M. Newns, Phys. Rev., 1969, 178, 1123. 15 G. Doyen and G. Ertl, Surf. Sci., 1974, 43, 197. 16 J. K. Nsrskov, S. Holloway and N. D. Lang, Surf. Sci., 1984, 137, 65; N. D. Lang, S. Holloway and J. K. Nsrskov, Surf. Sci., 1985, 150, 24. 17 L. D. Marks and V. Heine, J . Catal., 1985, 94, 570. 18 R. Haydock, J . Phys. C, 1981, 14, 3807. and H. J. Krebbs, Surf. Sci., 1982, 117, 639; G. A. Somorjai, Surf. Sci., 1979, 89, 499. Phys. Rev. Lett., 1985, 54, 1856. Surf Sci., 1985, 149, 48.1962 Electronic Interactions between CO and Promoters on Ni 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 D. Heskett, E. W. Plummer and W. Eberhardt, Phys Rev. B, 1986, 33, 5171; D. Heskett, I. Strathy, E. W. Plummer, and R. A. de Paola, Phys.Rev. B, 1985,32, 6222. J. K. Nsrskov and N. D. Lang, Phys. Rev. B, 1980, 21, 2131. T. Edmonds and J. J. McCarroll, in Topics in Surface Chemistry, ed. E. Kay and P. S. Bagus (Plenum Press, New York, 1977), p. 261. J. C. Slater and K. H. Johnson, Phys. Rev. B, 1972, 5, 844; K. H. Johnson and F. C. Smith Jr, Phys. Rev. B, 1972,5, 831; K. H. Johnson, Adv. Quantum Chem., 1973,7, 143. J. M. MacLaren, D. D. Vvedensky, J. B. Pendry and R. W. Joyner, Surf. Sci., 1985, 162, 322. A. C. Balazs, K. H. Johnson and G. M. Whitesides, Inorg. Chem., 1982, 21, 2162. J. B. Danese and J. W. D. Connolly, J. Chem. Phys., 1974,61, 3061. D. A. Case, M. Cook and M. Karplus, J. Chem. Phys., 1980,73, 3294. J. M. MacLaren, J. B. Pendry and R. W. Joyner, Surf. Sci., in press. M. Trenary, K. J. Uran and J. T. Yates Jr, Surf. Sci., 1985, 157, 512. M. Van Hove and S. Y. Tong, J. Vac. Sci. Technol., 1975, 12, 230. J. E. Demuth, D. W. Jepsen and P. M. Marcus, J. Phys. C, 1975,8, L25. S. Anderson and J. B. Pendry, J. Phys. C, 1980, 13, 3294 E. Wimmer, C. L. Fu and A. J. Freeman, Phys. Rev. Lett., 1985,55, 2618. J. Stohr, F. Sette and A. L. Johnson, Phys. Rev. Lett., 1984, 53, 1685; F. Sette, J. Stohr, E. B. Kolun, D. J. Dwyer, J. L. Gland, J. L. Robbins and A. L. Johnson, Phys. Rev. Lett., 1985, 54, 935; K. Baberschke, personal communication. N. V. Richardson and A. M. Bradshaw, Surf. Sci., 1979,88, 132. J. N. Allison and W. A. Goddard 111, Surf. Sci., 1982, 115, 553. R. W. Joyner, J. Catal., 1977, 50, 176. M. A. Vanice, Catal. Rev., 1976, 14, 153. D. E. Eastman, J. E. Demuth and J. M. Baker, J. Vac. Sci. Technol., 1973, 11, 273; W. Erley and H. Wagner, Surf. Sci., 1978, 74, 273. G. Ertl, in Catalysis, Science and Technology, ed. J. R. Anderson and M. Boudart (Springer-Verlag, Berlin, 1983), vol. 4, p. 273. M. Bowker, I. B. Bowker and K. C. Waugh, Appl. Catal., 1985, 14, 101. P. Stolze and J. K. Norskov, Phys. Rev. Lett., 1986, 55, 2502. J. E. Crowell, E. L. Garfunkel and G. A. Somorjai, Surf. Sci., 1982, 121, 303. D. A. King, to be published. S. Holloway and J. K. Norskov, J. Electroanal. Chem., 1984, 148, 303. Y. Kikuzono, S. Kagami, S. Naito, T. Onishi and K. Tamaru, Faraday Discuss. Chem. SOC., 1982, 72, 135. F. M. Hoffman, N. J. Levinos, B. N. Perry and P. Rabinowitz, Phys. Rev. B, 1986, 33, 4309. C. B. Harris, R. M. Shelby and P. A. Cornelius, Phys. Rev. Lett., 1977, 38, 1415. B. N. J. Persson and R. Ryberg, Phys. Rev. Lett., 1985, 54, 2119. B, N. J. Persson and M. Persson, Solid State Commun., 1980, 36, 175. Paper 6/ 122 1 ; Received 16th June, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301945

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

GENERAL DISCUSSION Prof. G. C. Bond (Brunel University) (communicated): It is well known that almost all molecules, but especially carbon monoxide, can exist in a variety of chemisorbed states, the situation depending critically upon the nzture of the metallic surface, the surface coverage and the temperature, as well as upon surface impurities or promoters, and other factors. The degree of complexity is indicated not only by thermal desorption spectra but also by the occurrence of precursor states needed to interpret the kinetics of chemisorption. In catalysis it is certainly possible that relatively weakly adsorbed states, variously described as precursor states, Type C chemisorption etc., are the reactive ones. My question is: when will the techniques of theoretical chemistry be ready to address the problem of describing these weak intermediate states as well as the final strongly bound state, and in general of modelling the dynamics of the whole chemisorption process ? Prof.R. A. van Santen (Amsterdam, The Netherlands) replied: As far as calculations on electronic structure are concerned, calculations with varying degree of sophistication can be done and very encouraging quantitative agreement is found. Calculation of potential- energy curves at a metal surface for chemisorbed molecules of high quantitative accuracy has been demonstrated for a few selected cases. These developments are still ongoing. It is difficult to predict when such techniques will become refined enough so that computation of interaction energies of weak intermediate states can also be addressed.Especially since such problems will require an adequate approach to account for correlation energy effects. Prof. J. Pritchard (Queen Mary College, London) said: In the Blyholder model of CO chemisorption the interaction of the 50 orbital with the metal involves partial charge transfer to the metal and the n-interaction a charge transfer from the metal in a synergetic process in which both o and n components are bonding. Your calculations support this view and indicate that the o interaction is bonding. Recently, however, Bagus and Hermannl have used a cluster model to account for the properties of chemisorbed CO and conclude that the o-interaction is repulsive. Could you comment on the relationship between their model and yours? 1 E.g.P. S . Bagus and K. Hermann, Phys. Rev. B, 1986,33, 2987 Dr D. W. Goodman (Sandia, Albuquerque, NM) said: Given the nature of the inter- action of the o and 2n* levels of CO with the d-band of metals and the oftentimes poor job that extended Hiickel methods do in accurately placing those levels, could you comment on the reliability of these techniques to define quantitatively transition- metal-CO bonding. Dr R. M. Lambert (Cambridge Uniuersity) said: Recent theoretical studies by Bagus et a1.l indicate that the Blyholder model does not provide a correct description of the bonding of CO to copper. Our own MCSCF calculations on the 2E ground state of Cu-CO are in agreement with this: this is not a bound state. The first excited state (2n) is, however, bound with a Cu-CO bond strength of ca.100 kJ mol-l; the Cu 4pn orbital is responsible for metal --+ adsorbate charge transfer in this case. 1 C. W. Bauschlicher, P. S . Bagus, C . J. Nelin and B. 0. Roos, J . Chem. Phys., 1986, 85, 354. 19631964 General Discussion Prof. R. A. van Santen (KoninklijkelShell Laboratorium, Amsterdam, The Nether- lands) responded: In our paper the Blyholder model for chemisorption is used. For the energy values of the 5 0 and 2n* orbitals empirical data have been used to adjust them to the experimental value. Since the orbitals are considered to be orthogonal, Born repulsion effects as discussed in the paper by Bagus are not accounted for. However, the bonding picture of CO we derive is very similar to that discussed by Sung and Hoffman,’ using a related method, but including non-orthogonality of the atomic orbitals.1 S . S. Sung and R. Hoffman, J . Am. Chem. SOC., 1985,107, 578. Prof. S. J. Thomson (Glasgow University) said: I would like to air a general request to the theoretical chemists. Would they please address the problem of the constancy of the Fermi level on chemisorption? Prof. R. A. van Santen (KoninklijkelShell Laboratorium, Amsterdam, The Nether- lands) answered: The Fermi level EF is the energy of the highest occupied molecular orbital of the metal. This.can be defined with respect to the bottom of the electron energy band or the vacuum. In the work presented in my paper the Fermi level is defined with respect to the bottom of the valence electron band and therefore is a constant in the calculations. If one defines the Fermi level with respect to vacuum then its position depends on the moment of the surface dipole layer p.The workfunction measures the sum of this contribution and the Fermi level. This cannot be considered a constant. The effect is accounted for in our calculations on the effect of alkali-metal coadsorption by explicitly introducing electrostatic interactions. Prof. V. Ponec (University of Leiden, The Netherlands) added: The statement that ‘Fermi energy does not shift upon chemisorption’, i.e. upon formation of a surface dipole layer, is not correct generally; it is only correct when the zero of the (electro-) chemical potential of electrons is put at the bottom of the conductivity band. If the zero energy is defined as a state with an electron at an infinite distance from a metal, such a statement is not valid.Of course, we are free in choosing the zero energy where it is most convenient and for solid-state theory this would be the bottom of the conduction band. However, when discussing photoemission phenomena from adsorbed layers, work-function effects caused by chemisorption etc. the other choice is frequently better. Choosing a zero energy for a state of an electron in vacuum and far from the metal (‘vacuum level ’) has also another, not really negligible advantage. When doing that, one avoids the frequent, but incorrect statement which can be found in the literature, such that the Fermi level is (upon measurement) at a distance from the vacuum level which is equal to the work function of the sample (actually, it is work function and the contact potential imposed by the measuring circuit) etc.With a zero vacuum level one does not forget so easily that the Fermi energy is a total energy that also includes the electrostatic energy of the measured sample. Dr R. M. Lambert (Cambridge University) said: Adding alkali-metal promoters to a metal surface is equivalent to adding a chemical compound, i.e. some kind of counter-ion is presumably present. Dr Holloway’s calculations indicate that on grounds of size the electrostatic effect of oxygen surface species will be dominated by the effect of much larger potassium species. Presumably this will not always be so (e.g. in the case of carbonate, sulphate and so on). One might therefore expect to find a systematic effect of counter-ion size on promoter efficiency for a given alkali metal.Is this information already available in the general literature or in the patent literature?General Discussion 1965 Prof. S. J. Thomson (Glasgow University) pointed out that there could well be a collective viewpoint. Aika et aZ.l had shown that average electronegativity of promoter bases gave a good correlation with activity in ammonia synthesis over Ru where the Ru was supported on a range of bases. Ru binding energies followed the expected trend. 1 K. Aika, A. Ohya, A. Ozaki, Y. Inoue and I. Yasumori, J. Catal., 1985, 92, 305. Dr R. W. Joyner (BP Sunbury) said: The information which Dr Lambert requests, on the role of counter-ions in alkali-metal promotion, is available from the work of Ozaki et aZ.l and the patent literature.Alkali-metal chlorides, sulphates and sulphides act as poisons or are inert. Carbonates and nitrates, which will decompose to oxides under reducing conditions, are effective promoters. Similar promotion effects are also observed in the absence of a counter-ion, where the promoter is introduced as the metal, by sublimation; or where an azide is used, which decomposes directly to the metal. It is not unambiguous that these results should be interpreted under Holloway's framework, since it is unlikely, in a supported catalyst, that the alkali-metal coverage of the transition-metal surface will be independent of the counter-ion. Alkali-metal chlorides, for example, may well exist solely as crystallites on the support, so that neither promotion nor poisoning should be expected.1 A. Ozaki, Isotopic Studies of Heterogeneous Catalysis (Academic Press, New York, 1978). Prof. G. L. Haller (Yale Uniuersity) said: So much of the discussion of the effect of promoters on CO dissociation has been couched in terms of donation of electron density into the z* orbital with accompanying weakening of the C-0 bond. However, this potential-energy pathway appears to me to correlate with chemisorbed C and gas-phase 0, i.e. it is not the final state of interest, it is likely to have a high energy barrier and does not tell us anything about how the promoter lowers the barrier to CO dissociation. However, tilting of the CO to bring the oxygen close to the surface does appear to be along a reaction coordinate of interest and the effect of promoters on energy change to move along this reaction coordinate is of great interest.You report that the chemisorbed levels are not much altered and the barrier to tipping to 45" is small. My question is what happens at angles beyond 45", and is the barrier to tilting to acute angles lowered by a promoter like K? Dr J. M. MacLaren (Imperial College, London) replied : The calculations discussed have only considered tipping the CO molecule on clean Ni( 100) to angles up to 45". There must eventually be direct interactions involving both the C and 0 ends of one of the 2n* levels and the substrate, although little evidence of this was seen for angles up to 45" in the calculations. The promoter, which predominantly acts to lower the work function will bring the 2n* closer to the Fermi level leading to increased occupancy and increased surface reactivity. The reaction pathway to dissociated CO involves populating the antibonding 2z* level, thus the promoter will lower the energy barrier.In addition direct CO-K-substrate chemical interactions are also possible. Prof. V. Ponec (University of Leiden, The Netherlands) said: On another point, it is unavoidable that the theoretical studies start with the simplest case - the influence of the promoters on CO adsorption. There, where dissociation is the most essential step this might even be the ultimate information required. However, let us not forget that on the way from CO to other products, intermediates such as formyl, formate, methylene dioxy- or alkoxy-groups etc.can play a role and promoters can influence their rate of formation and/or their stability. Perhaps a theoretical study of these (potential) effects is already feasible.1966 General Discussion Prof. N. Sheppard (University of East Anglia) said : The effects of coadsorbed alkali metal on the frequencies of CO species adsorbed on metal surfaces are most interesting. The large decreases in CO stretching frequencies with increasing alkali-metal coadsorp- tion undoubtedly reflect substantial weakenings of the CO bonds of the surface species with corresponding implications for reactivity. However, on reading the wider practical catalytic literature, it seems to me that the promotion effect of potassium is often associated with the additional presence of potassium ions, K+, rather than uncharged potassium atoms.Clearly, electrons can be released much more readily from the latter than the former. Can we be talking of the same ‘potassium-promotion’ phenomenon? Prof. J. J. F. Scholten (Technical University of Delft, The Netherlands) said : From kinetic studies on ammonia synthesis with and without addition of a K 2 0 promoter1 it is known that K,O is a typical ‘high-pressure promoter’, i.e. the acceleration factor is higher the higher is the pressure of the N2-H, feed. Furthermore, it is known2 that potassium is a more effective promoter than potassium oxide. My suggestion is that under industrial conditions potassium oxide is partly reduced to potassium on the surface of iron.It is true that the equilibrium K 2 0 + H2 $2K + H20 is strongly to the left. Nevertheless, it can be shifted to the right by (a) increasing the hydrogen pressure, (b) applying a high space velocity by which the H20 pressure above the catalyst surface is strongly lowered, and (c) by the fact that K atoms are chemisorbed on the Fe(ll1) plane, where they fit perfectly into the holes in this plane, and the heat of chemisorption is gained, counterbalancing the negative heat of reaction of K20 reduction. 1 C. Bokhoven, C. van Heerden, R. Westrik and P. Zwietering, in Catalysis, ed. P. H. Emmett (Reinhold Publishing Corporation, New York, 1955), vol. 111, pp. 265-348, especially fig. 26 on p. 337. 2 A. Ozaki, K. Akai and Y. Morikawa, in Proc. 5th Int.Congr. Catal., ed. J. W. Hightower (North Holland Publishing Company, Amsterdam 1973), paper 90, especially p. 337. Prof. D. A. King (University of Liverpool) commented as follows: Holloway and coworkers’ have used effective medium theory to explain on an atomic level the action of poisons and promoters. The effect is related to the electrostatic potential around the coadsorbed atom (e.g. K) and the direction of the induced or intrinsic dipole around the molecule (e.g. CO) caused by filling of antibonding molecular orbitals during adsorption. Electropositive atoms like K usually act as promoters for the adsorption process, but sometimes as a poison for the overall reaction. Campbell and Goodman2 showed that K increases the CO dissociation rate on Ni, but since that step is not rate-determining this results in a build up of surface carbon which hinders adsorption of H,. K thus poisons the reaction. As Dr Holloway pointed out in his paper,3 for ammonia synthesis where the adsorption step is rate-determining, K acts as a promoter. The effect of oxygen, he argued, is to reduce the number of promoted sites, but since the absolute magnitude of the K-induced electrostatic potential is much larger than for oxygen, the effect due to the latter is expected to be small. 1 J. K. N~rrskov, S. Holloway and N. D. Lang, Surf. Sci., 1984, 137, 65. 2 C. J. Campbell and D. W. Goodman, Surf. Sci., 1982, 123, 41 3. 3 S. Holloway, J. K. Nmskov and N. D. Lang, J. Chem. SOC., Faraday Trans. I , 1987,83, 1935.

ISSN:0300-9599    

DOI:10.1039/F19878301963

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1987,83, 1967-1973 (Faraday Symposium 21) Adsorption and Reaction on Strained-metal Overlayers Cu/Ru(0001) D. W. Goodman* and C. H. F. Peden Sandia National Laboratories, Albuquerque, New Mexico 87185, U.S.A. These studies have addressed the adsorption of CO on very thin (submono- layer to multilayer) deposits of Cu on an Ru(0001) single crystal as well as the measurement of the high-pressure kinetics of the methanation, ethane hydrogenolysis and cyclohexane dehydrogenation reactions on this model bimetallic catalyst. Temperature-programmed desorption (t .p.d.) spectra of CO at submonolayer Cu coverages show a monotonic diminution of features due to CO adsorbed on Ru with a concomitant appearance of new t.p.d. peaks, which can be assigned to desorption from the top and edges of [l x 11 ‘strained-layer’ Cu islands. Importantly, these new CO features have significantly higher desorption temperatures than those observed on bulk Cu(lll), indicating a stabilization relative to bulk Cu of CO on monolayer Cu films supported on Ru.This behaviour can be correlated with LEED, ARUPS and work-function measurements, which indicate the unique character of the first layer of Cu on Ru(0001). For methanation and ethane hydrogenolysis the overall surface activity is found to decrease monotonically with decreasing Ru surface area upon Cu addition. For cyclohexane dehydrogenation, however, the Ru specific activity is observed to increase by ca. 40 at a copper coverage of 0.75 of a monolayer. The rate enhancement observed for submonolayer Cu deposits may relate to the modified activity of the strained Cu overlayer owing to its altered geometric and electronic properties.It has long been recognized that the addition of impurities to metal catalysts can produce large effects on the activity, selectivity and resistance to poisoning of the pure meta1.l For example, the catalytic properties of metals can be altered greatly by the addition of a second transition metal.2 A long-standing question regarding such bimetallic systems is the nature of the properties of the mixed-metal system which give rise to its enhanced catalytic performance relative to either of its individual metal components. These enhanced properties (improved stability, selectivity and/or activity) can be accounted for by one or more of several possibilities.First, the addition of one metal to a second may lead to an electronic modification of either or both of the metal constituents. This electronic perturbation can result from direct bonding (charge transfer) or from a structural modification induced by one metal upon the other. Secondly, a metal additive can promote a particular step in the reaction sequence and thus act synergistically with the host metal. Thirdly, the additive metal can serve to block the availability of certain active sites, or ensembles, prerequisite for a particular reaction step. If this ‘poisoned’ reaction step involves an undesirable reaction product, then the net effect is an enhanced overall selectivity. Further, the attenuation by this mechanism of a reaction step leading to undesirable surface contamination will promote catalyst activity and durability.The present studies are part of a continuing e f f ~ r t ~ - ~ to identify those properties of bimetallic systems which can be related to their superior catalytic properties. A pivotal question concerns the relative importance of ensemble (steric or local) uersus electronic (non-local or extended) effects in the modification of catalytic properties by surface 19671968 Adsorption and Reaction on Strained-metal Overlayers impurities. A complete understanding of surface impurity effects (including alloying) in catalysis is likely to include components of both electronic and ensemble effects, the relative importance of each to be determined for a given reaction and reaction conditions.Our research has been concerned with this assessment of the relative magnitude of these two effects in the modification of surface activity by alloying and surface impurities. Experiment a1 The studies to be discussed were carried out using a specialized apparatus described in ref. (10)--(12). This device consists of two distinct regions, a surface analysis chamber and a microcatalytic reactor. The custom-built reactor, contiguous with the surface analysis chamber, employs retraction bellows that support the metal single crystal and allows translation of the catalyst in vacuo from the reactor to the surface analysis region. Both regions are of ultra-high vacuum construction, bakeable and capable of ultimate pressures < 2 x 10-lo Torr.* Auger spectroscopy (AES) was used to characterize the sample before and after reaction.The single-crystal catalysts, ca. 1 cm diameter x 1 mm thick, were aligned within 1/2' of the desired orientation. Thermocouples were spot-welded to the edge of the crystal for temperature measurement. Details of sample mounting, cleaning procedures and preparation techniques are given in the references accompanying the relevant data. All reactants were initially of high purity ; however, further purification procedures were generally used to improve the gas quality. These typically included multiple distillations for condensables and/or cryogenic scrubbing using a low-conductance glass wool-packed trap at 80 K. The kinetic data presented were obtained under steady-state reaction conditions and at low conversions (typically < 5%).Products were measured by a gas chromatograph equipped with a flame ionization detector (fid.). Results and Discussion Interest in bimetallic catalysts has risen steadily over the years because of the commercial success of these systems. This success results from an enhanced ability to control the catalytic activity and selectivity by tailoring the catalysts l3 A key question in these investigations has been the relative roles of ensemble and electronic effects in defining the catalytic behaviour. 14-17 In gathering information to address this question it has been advantageous to simplify the problem by using models of a bimetallic catalyst, such as the deposition of metals on single-crystal substrates in the clean environment familiar to surface science.Many such model systems have been studied, but a particularly appealing combination is that of Cu on Ru. Cu is immiscible in Ru, which facilitates coverage determinations by t . ~ . d . ~ and circumvents the com- plication of determining the three-dimensional composition. The adsorption and growth of Cu films on the Ru(0001) surface have been studied3+* 18-26 by work-function measurements, LEED, AES and t.p.d. The results from recent s t ~ d i e s ~ - ~ indicate that for submonolayer depositions at 100 K the Cu grows in a highly dispersed mode, subsequently forming two-dimensional islands pseudo- morphic to the Ru(0001) substrate upon annealing to 300 K. Pseudomorphic growth of the copper indicates that the copper-copper bond distances are strained ca.6% beyond the equilibrium bond distances found for bulk copper. A comparison of CO desorption from R u , ~ from multilayer Cu (10 ML)t on Ru and 1 ML Cu on Ru is shown in fig. 1. The t.p.d. features of the 1 ML Cu (peaks at 160 * 1 Torr = 101 325/760 Pa. t ML = monolayer.D. W. Goodman and C. H. F. Peden 1969 Fig. 1. T.p.d. results for CO adsorbed to saturation levels on (a) clean Ru(0001), (b) on multilayer Cu and (c) on a monolayer Cu covered Ru(0001). and 240 K) on Ru are at temperatures intermediate between CO desorbing from Ru and bulk Cu. This suggests that the monolayer Cu is electronically perturbed and that this perturbation manifests itself in the bonding of CO. An increase in the desorption temperature relative to bulk Cu indicates a stabilization of the CO on the monolayer Cu, suggesting a coupling of the CO through the Cu to the Ru.The magnitude of the CO stabilization implies that the electronic modification of the Cu by the Ru is significant and should be observable with a band structure probe. Recent angularly resolved photoemission studies8 indeed show a unique interface state which is likely to be related to the altered CO bonding on Cu films intimate to Ru. Fig. 2 shows the results6 of CO chemisorption on the Cu/Ru(OOOl) system as a function of the Cu coverage. In each case the exposure corresponds to a saturation coverage of CO. Most apparent in fig. 2 is a monotonic decrease upon addition of Cu of the CO structure identified with Ru (peaks at 400 and 480 K) and an increase of the CO structure corresponding to Cu (peaks at 160 and 210 K).The buildup of a third feature at ca. 300 K (indicated by the dashed line) is assigned to CO desorbing from the edges of Cu islands. Integration of the 160, 210 and 300 K peaks provides information regarding island sizes, i.e. perimeter-to-island area ratios, at various Cu coverages. For example, at O,, = 0.66 the average island size is estimated to be ca. 50 A in diameter. This island size is consistent with an estimate of the two-dimensional island size corresponding to this coverage of 40-60 A derived from the width of the LEED beam profiles at this coverage of C U . ~ Model studies of the Cu/Ru(0001) catalyst have been carried out7 for methanation and hydrogenolysis reactions.These data suggest that copper merely serves as an inactive diluent, blocking sites on a one-to-one basis. Similar results have been found in analogous studies2’ introducing silver onto an Rh( 1 1 1) methanation catalyst. Sinfelt et aZ.28 have shown that copper in a Cu/Ru catalyst is confined to the surface of ruthenium. Results from the model catalysts discussed here then should be relevant to those on the corresponding supported bimetallic catalysts. Several such studies have1970 lo< zoo 300 400 500 6 I I I I Adsorption and Reaction on Strained-metal Overlayers C 4 Q 10 TIK Fig. 2. T.p.d. results corresponding to CO adsorbed to saturation levels on the clean Ru(0001) surface and from this same surface containing various coverages of Cu (monolayers).been carried out to investigate the addition of copper or other Group 1B metals on the rates of CO hydr~genationl~. 2 9 v 30 and ethane hydrogenoly~is~~ catalysed by ruthenium. In general, these studies show a marked fall in activity with addition of the Group 1B metal, suggesting a more profound effect of the Group 1B metal on ruthenium than implied from the model studies. A critical parameter in the supported studies is the measurement of the active ruthenium surface using hydrogen chemisorption techniques. Haller and 9 32 have recently suggested that hydrogen spillover during chemisorption may occur from ruthenium to copper, complicating the assessment of surface Ru atoms. Recent studies in our laboratory*. have shown directly that spillover from Ru to Cu can take place and must be considered in the hydrogen chemisorption measurements on supported metal catalysts.H, spillover would lead to a significant overestimation of the number of active ruthenium metal sites and thus to significant error in calculating ruthenium specific activity. If this is indeed the case, the results obtained on the supported catalysts, corrected for the overestimation of surface ruthenium, could become more comparable with the model data reported here. Finally, the activation energies observed on supported catalysts in various laboratories are generally unchanged by the addition of Group 1B meta1,l49 29-32 in agreement with the model studies. These arguments suggest that Ru specific rates for methanation and ethane hydro- genolysis on supported Cu/Ru catalysts may approximate those values found for pureD.W. Goodman and C. H. I;. Peden 10.0 1971 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Cu coverage (monolayers) Fig. 3. Relative rate of reaction us. surface Cu coverage on Ru(0001) for cyclohexane dehydrogenation to benzene. PT = 101 Torr, HJcyclohexane = 100, T = 650 K. Ru. As a consequence, the rates for cyclohexane dehydrogenation reaction on suppor- ted Cu/Ru, similarly corrected, must exceed those specific rates found for pure Ru. The uncorrected specific rates for the supported Cu/Ru system remain essentially unchanged upon addition of Cu to R u . ~ ~ An activity enhancement for cyclohexane dehydrogenation in the mixed Cu/Ru system relative to pure Ru is most surprising given that Cu is less active for this reaction than Ru.Fig. 3 shows the effect of addition of Cu to Ru on the rate of cyclohexane dehydrogenation to benzene. The overall rate of this reaction is seen to increase by ca. an order of magnitude at a copper coverage of 3/4 of a monolayer. This translates to a Ru specific rate enhancement of ca. 40. Above this coverage, the rate falls to an activity approximately equal to that of Cu-free Ru. The observed non-zero rates at the higher Cu coverages are believed to be caused by three-dimensional clustering of the Cu overlayers.6 Similar data have been obtained for this reaction on epitaxial and alloyed Au/Pt( 1 1 1) surfaces.34 The rate enhancement observed for sub-monolayer Cu deposits may relate to an enhanced activity of the strained Cu film for this reaction due to its altered geometric6 and electronic8 properties.Alternatively, a mechanism whereby the two metals cooper- atively catalyse different steps of the reaction may account for the activity promotion. For example, dissociative H, adsorption on bulk Cu is unfavourable due to an activation barrier of ca. 5 kcal m01-l.~~ In the combined Cu/Ru system, Ru may function as an atomic hydrogen source/sink via spillover to/from neighbouring Cu. A kinetically controlled spillover of H, from Ru to Cu, discussed above, is consistent with an observed optimum reaction rate at an intermediate Cu coverage. Finally, we note the differences between a Ru(0001) catalyst with or without added Cu with respect to attaining steady-state reaction rates. On the Cu-free surface (fig.4), an induction time of ca. 10 min is required to achieve steady-state a~tivity.~ During this time, production of benzene is quite low, while the hydrogenolysis to lower alkanes, primarily methane, is significantly higher than at steady-state. During this induction time the carbon level (as determined by AES) rises to a saturation value coincidental with the onset of steady-state reaction. This behaviour suggests that a carbonaceous layer on the metal surface effectively suppresses carbon<arbon bond scission, or hydrogenolysis on the Ru surface.1972 Adsorption and Reaction on Strained-metal Overlayers I I I I 1 5 10 15 20 25 reaction time/rnin Fig. 4. (a) Amount of benzene formed from the dehydrogenation of cyclohexane and (b) lower carbon number alkanes from cyclohexane hydrogenolysis on Ru(0001) as a function of time.( c ) Relative quantities of carbonaceous material present on the surface following reaction at T = 650 K, H,/C,H,, = 100, PT = 101 Torr. Cu addition lead to an enhanced rate of benzene production with little or no induction time, i.e. the initial rate of cyclohexane hydrogenolysis relative to the Cu-free surface is suppressed. Further, Cu reduces the relative carbon buildup on the surface during reaction. Thus, Cu may play a similar role as the carbonaceous layer in suppressing cyclohexane hydrogenolysis, while concurrently stabilizing those intermediates leading to the product benzene. In addition, copper may serve to weaken the chemisorption bond of benzene and thus limit self-poisoning by adsorbed product.This latter possibility has been proposed by Sachtler and S o m o ~ j a i ~ ~ to explain the role of Au in AuPt( 1 1 1) catalysts for this reaction. A weakening of benzene chemisorption satisfactorily accounts for our observation that the reaction changes from zero order in cyclohexane on Ru(0001) to approximately first order upon the addition of C U . ~ Summary These studies have shown that: (a) The structure of the first monolayer of Cu is pseudomorphic with respect to the Ru(000 1) substrate, whereas successive Cu layers grow epitaxially with a Cu( 1 1 1) structure. (b) The adsorptive properties of 1 monolayer of Cu on Ru(0001) toward CO is markedly different than CO on Ru(0001) or CO on Cu(ll1). (c) The addition of Cu to Ru(0001) results in a dramatic enhancement of the rate of cyclohexane dehydrogenation, despite the fact that Cu is much less active for this reaction than is Ru.( d ) The mechanism by which Cu alters the catalytic activity of Ru may involve one or more of several possibilities including an electronic modification of either or both of the catalyst metals or a synergism between the activities of the two metal components.D. W. Goodman and C . H . F. Peden 1973 We acknowledge with pleasure the partial support of this work by the Department of Energy, Office of Basic Energy Sciences, Division of Chemical Sciences, U.S.A. References 1 Metal-Support and Metal-Additive Eflects in Catalysts, ed. B. Imelik, C. Naccache, G. Coudurier, H. Praliaud, P. Mlriaudeau, P.Gallezot, G. A. Martin and J. C. Vedrine (Elsevier, Amsterdam, 1982). 2 J. H. Sinfelt, Bimetallic Catalysts: Discoveries, Concepts and Applications (John Wiley, New York, 1983). 3 J. T. Yates Jr, C. H. F. Peden and D. W. Goodman, J . Catal., 1985,94, 576. 4 D. W. Goodman, J. T. Yates Jr and C. H. F. Peden, Surf. Sci., 1986, 164,417. 5 D. W. Goodman and C. H. F. Peden, J . Catal., 1986, 95, 321. 6 J. E. Houston, C. H. F. Peden, D. S. Blair and D. W. Goodman, Surf. Sci., 1986, 167, 427. 7 C. H. F. Peden and D. W. Goodman, Ind. Eng. Chem. (Fundam.), 1986,25, 58. 8 J. E. Houston, C. H. F. Peden, P. J. Feibelman and D. R. Hamann, Phys. Rev. Lett., 1986, 56, 375. 9 C. H. F. Peden and D. W. Goodman, J . Catal., submitted for publication. 10 D. W. Goodman, R. D. Kelley, T.E. Madey and J. T. Yates Jr, J . Catal., 1980, 63, 226. 11 D. W. Goodman, J. Vac. Sci. Technol., 1982, 20, 522. 12 D. W. Goodman, Ace. Chem. Res., 1984, 17, 194. 13 G. M. Schwab, Discuss. Faraday Soc., 1950, 8, 166. 14 G. C. Bond and B. D. Turnham, J. Catal., 1976, 45, 128. 15 C. Betizeau, G. Leclercq, R. Maural, C. Bolivar, H. Charcosset, R. Frety and L. Tournayan, J . Catal., 16 S. Galvagno and G. Parravano, J . Catal., 1979, 57, 272. 17 H. C. de Jongste, V. Ponec and F. G. Gault, J . Catal., 1980, 63, 395. 18 K. Christmann, G. Ertl and H. Shimizu, J . Catal., 1980, 61, 397. 19 H. Shimizu, K. Christmann and G. Ertl, J . Catal., 1980, 61, 412. 20 J. C. Vickerman, K. Christmann and G. Ertl, J. Catal., 1981, 71, 175. 21 S. K. Shi, H. I. Lee and J. M. White, Surf. Sci., 1981, 102, 56. 22 L. Richter, S. D. Bader and M. B. Brodsky, J . Vac. Sci. Technol., 1981, 18, 578. 23 J. C. Vickerman and K. Christmann, Surf. Sci., 1982, 120, 1. 24 J. C. Vickerman, K. Christmann, G. Ertl, P. Heiman, F. J. Himpsel and D. E. Eastman, Surf. Sci., 25 S. D. Bader and L. Richter, J . Vac. Sci. Technol., 1983, 41, 1185. 26 C. Park, E. Bauer and H. Poppa, Surf. Sci., submitted for publication. 27 D. W. Goodman, in Heterogeneous Catalysis (Proc. IUCCP Conf.), (Texas University, 1984). 28 J. H. Sinfelt, G. H. Via and F. W. Lytle. Catal. Rev. Sci. Eng., 1984, 26, 81. 29 L. J. M. Luyten, M. van Grondelle and J. H. C. van Hooff, J. Phys. Chem., 1978, 82, 2000. 30 A. K. Datye and J. Schwank, J . Catal., 1985, 93, 256. 31 A. J. Rouco, G. L. Haller, J. A. Oliver and C. Kemball, J . Catal., 1983, 84, 297. 32 G. L. Haller, D. E. Resasco and J. Wang, J . Catal., 1983, 84, 477. 33 J. H. Sinfelt, J . Catal., 1973, 29, 308. 34 J. W. A. Sachtler and G. A. Somorjai, J . Catal., 1984, 89, 35. 35 M. Balooch, M. J. Cardillo, D. R. Miller and R. E. Stickney, Surf. Sci., 1975, 50, 263. 1976, 45, 179. 1983, 134, 367. . Paper 611080; Received 20th May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878301967

出版商:RSC    

年代:1987

数据来源: RSC