摘要:

Gas Kinetics Group and Division de Chimie-Physique de la Societe Francaise de Chimie 9th International Symposium on Gas Kinetics To be held in Bordeaux, France on 20-25 July 1986 Further information from Dr R. Lasclaux, Lab. Photophys. Photochim. MolBculaire, Universite de Bordeaux I, 33405 Talence Cedex, France Poiymer Physics Group Biologically Engineered Polymers To be held at Churchill College, Cambridge on 21-23 July 1986 Further information from Dr M. J. Miles, AFRC,Food Research Institute, Colney Lane, Norwich NR4 7UA Polymer Physics Group with the British Rheological Society Deformation in Solid Polymers To be held at the University of Leeds on 9-1 1 September 1986 Further information from Dr J. V. Champion, Department of Physics, City of London Polytechnic, 31 Jewry Street, London EC3N 2EY ~~_____________ ~~~~ Carbon Group Carbon Fibres- P ro pe rt i es and A p p I i cat i o ns To be held at the University of Salford on 1 5 1 7 September 1986 Further information from The Meetings Officer, The Institute of Physics, 47 Belgrave Square, London SW1 X 8QX ~ ~~~~~~~~ ~ Division with the Surface Reactivity and Catalysis Group-Autumn Meeting Promotion in Heterogeneous Catalysis To be held at the University of Bath on 23-25 September 1986 Further information from: Professor F.S. Stone, School of Chemistry, University of Bath, Bath BA2 7AY (viii)Gas Kinetics Group and Division de Chimie-Physique de la Societe Francaise de Chimie 9th International Symposium on Gas Kinetics To be held in Bordeaux, France on 20-25 July 1986 Further information from Dr R.Lasclaux, Lab. Photophys. Photochim. MolBculaire, Universite de Bordeaux I, 33405 Talence Cedex, France Poiymer Physics Group Biologically Engineered Polymers To be held at Churchill College, Cambridge on 21-23 July 1986 Further information from Dr M. J. Miles, AFRC,Food Research Institute, Colney Lane, Norwich NR4 7UA Polymer Physics Group with the British Rheological Society Deformation in Solid Polymers To be held at the University of Leeds on 9-1 1 September 1986 Further information from Dr J. V. Champion, Department of Physics, City of London Polytechnic, 31 Jewry Street, London EC3N 2EY ~~_____________ ~~~~ Carbon Group Carbon Fibres- P ro pe rt i es and A p p I i cat i o ns To be held at the University of Salford on 1 5 1 7 September 1986 Further information from The Meetings Officer, The Institute of Physics, 47 Belgrave Square, London SW1 X 8QX ~ ~~~~~~~~ ~ Division with the Surface Reactivity and Catalysis Group-Autumn Meeting Promotion in Heterogeneous Catalysis To be held at the University of Bath on 23-25 September 1986 Further information from: Professor F. S. Stone, School of Chemistry, University of Bath, Bath BA2 7AY (viii)

ISSN:0300-9599    

DOI:10.1039/F198581FX041

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

xxxij AUTHOR INDEX Singh, Km. S., 751 Sircar, S., 1527, 1541 Slade, R. C. T., 847 Smith, I. G., 1095 Snelling, C. M., 1761 Sobczyk, L., 311 Siiderberg, D., 17 15 Solar, S., 1101 Solar, W., 1101 Soma, M., 485 Somorjai, G. A., 1263 Somsen, G., 1015 Sorek, Y., 233 Souto, F. A., 2647 Spencer, S., 2357 Spichiger-Ulmann, M., 7 13 Spoto, G., 1283 Spotswood, T. M., 1623 Srivastava, R. D., 913 Stachurski, J., 1447, 2813 Staricco, E. H., 1303 Stock, T., 2257 Stockhausen, M., 397 Stokes, R. H., 1459 Stone, F. S., 1255 Strachan, A. N, 1761 Strohbusch, F., 2021 Stuckless, J. T., 597 Su, Z., 2293 Subrahmanyam, V. S., 1655 Sugimoto, N., 1441, 2959 Suminaka, M., 2287 Suprynowicz, Z., 553 Sutcliffe, L. H., 679, 1467, 1215 Suzanne, J., 2339 Suzuki, H., 3117 Swallow, A. J., 1225 Symons, M.C. R., 433, 565, 727, 2131, 2775, 1095, 1963, 242 1 Takagi, Y., 1901 Takahashi, Y., 3 117 Takeshita, H., 2805 Tamilarasan, R., 2763 Tamura, K., 2287 Tanaka, T., 1513 Taniewska-Osinska, S., 695, Tascon, J. M. D., 939, 2399 Taylor, M. J., 1863 Taylor, N., 2357 Tejuca, L. G., 939, 2399, 1203 Teller, R. G., 1693 Tempere, J-F., 1357 Teramoto, M., 2941 Theocharis, C . R., 857 Thomas, J. K., 735 Tielen, M., 2889, 3049 Tindwa, R. M., 545 Tissier, C., 3081 Toi, K., 2835 Tokuda, T., 2835 Torrez-Mujica, T., 343 Townsend, R. P., 1071, 173 1, Trasatti, S., 2995 Treiner, C., 2513 Trenwith, A. B., 745 Trifiro, F., 1003 Troncoso, G., 1631, 1637 Tseung, A. C. C., 1883 Tuck, J. J., 833 Turner, J. E., 1263 Uemoto, M., 2333 Uma, K., 2733 Valencia, E., 1631. 1637 Valigi, M., 813 Vallmark, T., 1389 Van Oort, M.J. M., 3059 Varma, M. K., 751 Vattis, D., 2043 Vecli, A., 433 Veseli, V., 2095 Vink, H., 1677, 1725 Vliers. D. P., 2009 Vukovid, Z., 1275 3081, 1913 3127 Waghorne, W. E., 2703 Ward, A. J., 2975 Watanabe, H., 1569 Waugh, K. C., 3073 Weckstrorn, K., 2947 Weinberg, N. N., 875 Weingartner, H., 1031 Wells, C. F.. 801, 1057, 1401, White, M. A., 3059 Williams, J. O., 271 1 Williams, P. A., 2635 Williams, P. B., 3067 Williams, R. T., 847 Wojcik, D., 1037 Wood, G. L., 265 Wood, R. M., 273 Woolf, L. A., 769, 2821 Wright, C. J., 2067 Wright, J. P., 1471 Wright, T. H., 1819 Wurie, A. T., 2605 Yadav, G. D., 161 Yadava, R. D., 751 Yamaguchi, M., 1513 Yamaguti, K., 1237 Yamasaki, S., 267 Yamashita, H., 2485 Yamatera, H., 127 Yelon, W., 1693 Yoshida, S., 1513, 2485 Yoshikawa, M., 2485 Zambonin, P.G.. 621 zdanov, S. P., 2541 Zecchina, A., 1283 Zelano, V., 2365 Zhan, R. Y., 2083 Zhao, Z., 185 Zhulin, V. M., 875 Zilnyk, A., 679, 1215 Zulauf, M., 2947 Zundel, G., 1425, 2375 1985. 2145, 2475, 3091xxxij AUTHOR INDEX Singh, Km. S., 751 Sircar, S., 1527, 1541 Slade, R. C. T., 847 Smith, I. G., 1095 Snelling, C. M., 1761 Sobczyk, L., 311 Siiderberg, D., 17 15 Solar, S., 1101 Solar, W., 1101 Soma, M., 485 Somorjai, G. A., 1263 Somsen, G., 1015 Sorek, Y., 233 Souto, F. A., 2647 Spencer, S., 2357 Spichiger-Ulmann, M., 7 13 Spoto, G., 1283 Spotswood, T. M., 1623 Srivastava, R. D., 913 Stachurski, J., 1447, 2813 Staricco, E. H., 1303 Stock, T., 2257 Stockhausen, M., 397 Stokes, R. H., 1459 Stone, F. S., 1255 Strachan, A.N, 1761 Strohbusch, F., 2021 Stuckless, J. T., 597 Su, Z., 2293 Subrahmanyam, V. S., 1655 Sugimoto, N., 1441, 2959 Suminaka, M., 2287 Suprynowicz, Z., 553 Sutcliffe, L. H., 679, 1467, 1215 Suzanne, J., 2339 Suzuki, H., 3117 Swallow, A. J., 1225 Symons, M. C. R., 433, 565, 727, 2131, 2775, 1095, 1963, 242 1 Takagi, Y., 1901 Takahashi, Y., 3 117 Takeshita, H., 2805 Tamilarasan, R., 2763 Tamura, K., 2287 Tanaka, T., 1513 Taniewska-Osinska, S., 695, Tascon, J. M. D., 939, 2399 Taylor, M. J., 1863 Taylor, N., 2357 Tejuca, L. G., 939, 2399, 1203 Teller, R. G., 1693 Tempere, J-F., 1357 Teramoto, M., 2941 Theocharis, C . R., 857 Thomas, J. K., 735 Tielen, M., 2889, 3049 Tindwa, R. M., 545 Tissier, C., 3081 Toi, K., 2835 Tokuda, T., 2835 Torrez-Mujica, T., 343 Townsend, R.P., 1071, 173 1, Trasatti, S., 2995 Treiner, C., 2513 Trenwith, A. B., 745 Trifiro, F., 1003 Troncoso, G., 1631, 1637 Tseung, A. C. C., 1883 Tuck, J. J., 833 Turner, J. E., 1263 Uemoto, M., 2333 Uma, K., 2733 Valencia, E., 1631. 1637 Valigi, M., 813 Vallmark, T., 1389 Van Oort, M. J. M., 3059 Varma, M. K., 751 Vattis, D., 2043 Vecli, A., 433 Veseli, V., 2095 Vink, H., 1677, 1725 Vliers. D. P., 2009 Vukovid, Z., 1275 3081, 1913 3127 Waghorne, W. E., 2703 Ward, A. J., 2975 Watanabe, H., 1569 Waugh, K. C., 3073 Weckstrorn, K., 2947 Weinberg, N. N., 875 Weingartner, H., 1031 Wells, C. F.. 801, 1057, 1401, White, M. A., 3059 Williams, J. O., 271 1 Williams, P. A., 2635 Williams, P. B., 3067 Williams, R. T., 847 Wojcik, D., 1037 Wood, G. L., 265 Wood, R. M., 273 Woolf, L. A., 769, 2821 Wright, C. J., 2067 Wright, J. P., 1471 Wright, T. H., 1819 Wurie, A. T., 2605 Yadav, G. D., 161 Yadava, R. D., 751 Yamaguchi, M., 1513 Yamaguti, K., 1237 Yamasaki, S., 267 Yamashita, H., 2485 Yamatera, H., 127 Yelon, W., 1693 Yoshida, S., 1513, 2485 Yoshikawa, M., 2485 Zambonin, P. G.. 621 zdanov, S. P., 2541 Zecchina, A., 1283 Zelano, V., 2365 Zhan, R. Y., 2083 Zhao, Z., 185 Zhulin, V. M., 875 Zilnyk, A., 679, 1215 Zulauf, M., 2947 Zundel, G., 1425, 2375 1985. 2145, 2475, 3091

ISSN:0300-9599    

DOI:10.1039/F198581BX043

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

JOURNAL OF THE CHEMICAL SOCIETY F,4RADAY TRANSACTIONS, PARTS I A N D 11 The Journal of the Chemical Society is published in six sections, of which five are termed Transactions; these are distinguished by their subject matter, as follows: Dalton Transactions (Inorganic Chemistry). All aspects of the chemistry of inorganic and organometallic compounds; including bioinorganic chemistry and solid-state inorganic chemistry; of their structures, properties, and reactions, including kinetics and mechanisms; new or improved experimental techniques and syntheses. Faraday Transactions I (Physical Chemistry). Radiation chemistry, gas-phase kinetics, electrochemistry (other than preparative), surface and interfacial chemistry, heterogeneous catalysis, physical properties of polymers and their solutions, and kinetics of polymerization, etc.Faraday Transactions II (Chemical Physics). Theoretical chemistry, especially valence and quantum theory, statistical mechanics, intermolecular forces, relaxation phenomena, spectroscopic studies (including i.r., e.s.r., n.m.r., and kinetic spec- troscopy, etc.) leading to assignments of quantum states, and fundamental theory. Studies of impurities in solid systems. Perkin Transactions I (Organic Chemistry). All aspects of synthetic and natural product organic, organometallic and bio-organic chemistry, including aliphatic, alicyclic, and aromatic systems (carbocyclic and heterocyclic). Perkin Transactions II (Physical Organic Chemistry). Kinetic and mechanistic studies of organic, organometallic and bio-organic reactions.The description and application of physicochemical, spectroscopic, and theoretical procedures to organic chemistry, including structure-activity relationships. Physical aspects of bio-organic chemistry and of organic compounds, including polymers and biopolymers. Authors are requested to indicate, at the time they submit a typescript, the journal for which it is intended. Should this seem unsuitable, the Editor will inform the author. The sixth section of the Journal of the Chemical Society is Chemical Communications, which is intended as a forum for preliminary accounts of original and significant work, in any area of chemistry that is likely to prove of wide general appeal or exceptional specialist interest. Such preliminary reports should be followed up eventually by full papers in other journals (e.g.the five Transaclions) providing detailed accounts of the work. NOTES I t has always been the policy of the Faraday Transactions that brevity should not be a factor influencing acceptability for publication. In addition however to full papers both sections carry at the end of each issue a section headed ‘Notes’, which are short self-contained accounts of experimental observations, results, OJ theory that will not require enlargement into ‘full’ papers. The Notes section is not used for preliminary communications. The layout of a Note is the same as that of a paper. Short summaries are required. The procedure for submission, administration, refereeing, editing and publication o f Notes is the same as for full papers.However, Notes are published more quickly than papers since their brevity facilitates processing at all stages. The Editors endeavour to meet authors’ wishes as to whether an article is a full paper or a Note, but since there is no sharp dividing line between the one and the other, either in terms of length or Character of content, the right Is retained to transfer overlong Notes to the full papers section. As a guide a Note should not exceed IS00 words or word-equivalents. (i)NOMENCLATURE AND SYMBOLISM 1 THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY Units and Symbols. The Symbols Committee of The Royal Society, of which The Royal Society of Chemistry is a participating member, has produced a set of recommendations in a pamphlet ‘Quantities, Units, and Symbols’ (1975) (copies of this pamphlet and further details can be obtained from the Manager, Journals, The Royal Society of Chemistry, Burlington House, London W 1 V OBN).These recommendations are applied by The Royal Society of Chemistry in all its publications. Their basis is the ‘ Systeme International d’Unites’ (SI). A more detailed treatment of units and symbols with specific application to chemistry is given in the IUPAC Manual of Symbols and Terminology for Physicochemical Quantities and Units (Pergamon, Oxford, 1979). Nomenclature. For many years the Society has actively encouraged the use of standard IUPAC nomenclature and symbolism in its publications as an aid to the accurate and unambiguous communication of chemical information between authors and readers.In order to encourage authors to use IUPAC nomenclature rules when drafting papers, attention is drawn to the following publications in which both the rules themselves and guidance on their use are given: Nomenclature of Organic Chemistry, Sections A , B, C, D, E, F, and H (Pergamon, Oxford, 1979 edn). Nomenclature of inorganic Chemistry (Butterworths, London, 197 I , now published by Pergamon). Biochemical Nomenclature and Related Documents (The Biochemical Society, London, 1978). A complete listing of all IUPAC nomenclature publications appears in the January issues of J. Chem. SOC., Faraday Transactions. It is recommended that where there are no IUPAC rules for the naming of particular compounds or authors find difficulty in applying the existing rules, they should seek the advice of the Society’s editorial staff.’ Marlow Medal and Prize Applications are invited for the award of the Marlow Medal for 1986 and Prize of f 100. The award will be open to any member of the Faraday Division of the Royal Society of Chemistry who, by the age of 32, had made in the judgement of the Council of the Faraday Division, the most meritorious contribution to physical chemistry or chemical physics. The award will be made on the basis of publications (not necessarily in the Transactions) on any subject normally published in J. Chem. SOC., Faraday Transactions / and I / , that carry a date of receipt for publication not later than the candidate’s 32nd birthday. Candidates should be members and under 34 on 1 st January 1986, the closing date for applications, which may be made either by the candidate himself or on his behalf by another member of the Society.Copies of the rules of the award and application forms may be obtained from: Mrs Y. A. Fish, The Royal Society of Chemistry, Burlington House, London W1V OBN (ii)THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY SYMPOSIUM NO. 20 Phase Transitions in Adsorbed Layers University of Oxford, 17-1 8 December 1985 Organising Committee : Professor J. S. Rowlinson (Chairman) Dr E. Dickinson Dr R. Evans Mrs Y. A. Fish Dr N. Parsonage Dr D. A. Young T’he aim of the meeting is to discuss phase transitions at gas/liquid, liquid/liquid and solid/fluid interfaces, and in other systems of constrained geometry or dimensionality less than three.Emphasis will be placed on molecularly simple systems, whereby liquid crystal interfaces and chemisorption phenomena are excluded. The final programme and application form may be obtained from: Mrs Y. A. Fish, The Royal Society of Chemistry, Burlington House, London W1V OBN THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY GENERAL DISCUSSION NO. 81 Lipid Vesicles and Membranes Loughborough University of Technology, 15-17 April 1986 Organising Committee: Professor D. A. Haydon (Chairman) Professor D. Chapman Mrs Y. A. Fish Dr M. J. Jaycock Dr I. G. Lyle Professor R. H. Ottewill Dr A. L. Smith Dr D. A. Young The aim of the meeting is to discuss the physical chemistry of lipid membranes and their interactions, in particular theoretical and spectroscopic studies, polymerised membranes, thermodynamics of bilayers and Iiposomes, mechanical properties, encapsulation and interaction forces between bilayers leading to fusion but excluding preparation and characterisation methodology.The preliminary programme may be obtained from : Mrs Y. A. Fish, The Royal Society of Chemistry, Burlington House, London W1 V OBN (iii)THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY GENERAL DISCUSSION NO. 82 Organising Committee : ~ Professor F. S. Stone (Chairman) ~ Dr R. Burch ~ Mrs Y. A. Fish Dr D. A. Young (Editor) Dr R. W. Joyner Professor J. Pritchard The symposium will form the Faraday Division Programme at the 1986 Autumn meeting of the Royal Society of Chemistry, however, it will be conducted as a discussion meeting, with pre-printed papers and subsequent publication, following the style of the traditional Faraday discussions and symposia. The role of promoters is of intrinsic interest as well as being important for many industrial processes.Promoters are used for three purposes, to improve catalyst activity, to increase selectivity for the desired reaction, and to prolong catalyst life at high activity and selectivity. There are current advances in both exprimental and theoretical aspects of promoter action, making this an opportune time for a Faraday symposium. Attention will be focussed on the role of promoters in enhancing activity and selectivity. Three areas will be highlighted - model studies using well-defined surfaces such as single crystals, characterization of promoter function in real catalysts, and theoretical aspects of promotion.The mechanisms of promoter action in metal, oxide and sulphide catalysts will be discussed. Contributions for consideration by the organising committee are invited and abstracts of about 300 words should be sent by 30th November 1985 to: I MrsY. A. Fish,The Royal Societyof Chemistry, Burlington House, London WlVOBN. Dynamics of Molecular Photof ragmentation University of Bristol, 15-1 7 September 1986 I Organising Committee: Professor R. N. Dixon (Chairman) Dr G. G. Balint-Kurti Dr M. S. Child Professor R. Donovan Professor J. P. Simons I Full texts of the accepted papers will be required by May 1986. The discussion will focus on the interaction of radiation with small molecules, molecular ions and complexes leading directly or indirectly to their dissociation. Emphasis will be given to contributions which trace the detailed dynamics of the photodissociation process.The aim will be to bring together theory and experiment and thereby stimulate important future work. Further information may be obtained from: Professor R. N. Dixon, Department of Theoretical Chemistry, University of Bristol, Cantock's Close, Bristol BS8 1TS THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY SY M POSl U M Promotion in HeterogeneousCatalysis University of Bath, 23-25 September 1986THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY SYMPOSIUM NO. 21 Interaction-induced Spectra in Dense Fluids and Disordered Solids University of Cambridge, 10-1 1 December 1986 Organising Committee : Professor A.D. Buckingham (Chairman) Dr R. M. Lynden-Bell Dr P. A. Madden Professor E. W. J. Mitchell Dr J. Yarwood Dr D. A. Young Mrs Y. A. Fish Whilst interaction-induced spectra have been studied in the gas phase for many years, their importance in the spectroscopy of condensed matter has been appreciated only relatively recently. At present a considerable number of studies of induced spectra are taking place in what are (nominally) widely separated fields of study. It is highly desirable to bring these communities together so that common issues can be identified and the progress of one field appreciated in another. Further information may be obtained from: Professor A. D. Buckingham, University Chemical Laboratory, Lensfield Road, Cambridge CB2 1 EW THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY GENERAL DISCUSSION NO.83 Brownian Motion University of Cambridge, 7-9 April 1987 Organising Committee Dr M. La1 (Chairman) Dr R. Ball Dr E. Dickinson Dr J. S. Higgins Dr P. N. Pusey Dr D. A. Young Mrs Y. A. Fish The aim of the meeting is to discuss new developments in the experimental and theoretical studies of Brownian motion of colloidal particles and macromolecules, with particular emphasis on the dynamics of aggregate formation and breakdown, computer simulation and many-body hydrodynamic interactions. Contributions for consideration by the Organising Committee are invited and abstracts of about 300 words should be sent by 15 June 1986 to: Dr M. Lal, Unilever Research, Port Sunlight Laboratory, Bebington, Wirral L63 3JW Full papers for publication in the Discussion volume will be required by December 1986JOURNAL OF CHEMICAL RESEARCH Papers dealing with physical chemistry/chemical physics which have appeared recently in J.Chern.Research, The Royal Society of Chemistry’s synopsis+microform journal, include the following : Solvent Effects on the Rotational Barriers of the N,N-Dimethylamides of 2- and 3-Furoic and 2- and 3-Thenoic Acids Gaetano Alberghina, Francesco Agatino Bottino, Salvatore Fisichella, and Caterina Arnone (1 985, Issue 4) A Partial Determination of the Stability Fields of Ferrierite and Zeolites ZSM-5, ZSM-48, and Nu-10 in the K,O-AI,O,-SO,-NH, [CH,],NH, System Abraham Araya and Barrie M.Lowe (1 985, Issue 6) The Level of Prochirality: the Analogy between Substitutional and Distortional Desymmetrization Amitai E. Halevi (1 985, Issue 6) Radical Cations of Di-, Tri-, and Tetra-bromoethane formed by Radiolysis: an Electron Spin Resonance Study Martyn C.R. Symons (1985, Issue 8) The Use of Deuterium N.m.r. Spectroscopy in Mechanistic Studies of Alkane-exchange Reactions on Supported Platinum and Rhodium Catalysts Ronald Brown, Charles Kemball, James A. Oliver, and Ian H. Sadler (1 985, Issue 9) Electron Spin Resonance Investigation of Environmental Effects in the Photosensitised Reaction of Uranyl Ion with Thioethers Hanna B. Ambroz and Terence J. Kemp (1 985 Issue 9) FeV,O, Spinel Solid solutions and Pierre Perrot (1 985, Issue 10) lonophore in Methanol Jean Juillard, Claude Tissier, and Georges Jeminet (1 985, Issue 10) Santiago Olivella, Miquel A.Pericas, Antoni Riera, and Albert Sole (1 985, Issue 10) The Iron-Vanadium-Oxygen System at 11 23,1273, and 1373 K. Part 2. activities in Fe,O,- Larbi Marhabi, Marie-Chantal Trinel-Dufour, Complexes of Sodium, Potassium, Magnesium, and Calcium Cations with the Lysocellin Is Singlet Cyclopentyne a True Minimum on the C,H, Potential-energy Hypersurface? FARADAY DIVISION INFORMAL AND GROUP MEETINGS Division-Endowed Lecture Symposium Molecular Spectroscopy and Dynamics (including the Faraday Lecture by A. Carrington) To be held at the Royal Institution, London on 10 December 1985 Further information from: Mrs Y. A. Fish, The Royal Society of Chemistry, Burlington House, London W1 V OBN ~~ ~ ~ Colloid and Interface Science Group with the Colloid and Surface Science Group of the SCI Interfacial Rheology To be held at Imperial College, London on 16 December 1985 Further information from Dr R.Aveyard, Department of Chemistry, The University, Hull HU6 7RX High Resolution Spectroscopy Group and Theoretical Chemistry Group Title to be Announced To be held at the University of York on 16-18 December 1985 Further information may be obtained from: Dr J. M. Hollas, Department of Chemistry, University of Reading, White knights, Reading RG6 2ADNeutron Scattering Group Time-resolved Scattering and Transition Kinetics To be held at Imperial College, London on 17 December 1985 Further information may be obtained from: Dr J. S. Higgins, Department of Chemical Engineering, Imperial College London SW7 2BY Molecular Beams Group with CCP6 Molecular Scattering-Theory and Experiment To be held at the University of Sussex on 19-21 March 1986 Further information from Dr A.Stace, School of Molecular Sciences, University of Sussex, Falmer, Brighton BN1 9QJ Electrochemistry Group Novel Techniques for the Study of Electrodes and their Reactions To be held at St. Catherine’s College, Oxford on 7-9 April 1986 Further information from: Dr S. P. Tyefield, CEGB, Rs Dept, Berkeley Nuclear Laboratories, Berkeley, Division-Annual Congress Structure and Reactivity of Gas Phase Ions To be held at the University of Warwick on 8-1 1 April 1986 Further information from: Professor K. R. Jennings, Department of Molecular Sciences, University of Warwick, Coventry CV4 7AL Gloucestershire GL13 9PB - Polymer Physics Group with the Statistical Mechanics and Thermodynamics Group Macromolecular Flexibility and Behaviour in Solution To be held at the University of Bristol on 16-1 8 April 1986 Further information from The Meetings Officer, The Institute of Physics, 47 Belgrave Square, London SW1 X 8QX Division with the Societe Frangaise de Chimie, Deutsche Bunsen Gesellschaft fur Ph ysikalische Chemie and Associazione ltaliana di Chimica Fisica Dynamics of Molecular Crystals To be held at Grenoble, France on 30 June to 4 July 1986 Further information from: Dr C.Troyanowsky, 10 rue Vauquelin, 75005 Paris, France Industrial Physical Chemistry Group Physical Chemistry of Water Soluble Polymers To be held at Girton College, Cambridge on 1-3 July 1986 Further information from Dr I .D. Robb, Unilever Research Laboratory, Port Sunlight, Bebington, Wirral L63 3JW Polymer Physics Group Biologically Engineered Polymers To be held at Churchill College, Cambridge on 21-23 July 1986 Further information from Dr M. J. Miles, AFRC Food Research Institute, Colney Lane, Norwich NR4 7UA Carbon Group Carbon Fibres-Properties and Applications To be held at the University of Salford on 1 5 1 7 September 1986 Further information from The Meetings Officer, The Institute of Physics, 47 Belgrave Square, London SW1 X 8QX Division with the Surface Reactivity and Catalysis Group-Autumn Meeting Promotion in Heterogeneous Catalysis To be held at the University of Bath on 23-25 September 1986 Further information from: Professor F.S. Stone, School of Chemistry, University of Bath, Bath BA2 7AY (vii)Faraday Discussions No. 77 Interfacial Kinetics in Solution This publication focuses attention on reactions involving liquid-gas, liquid-liquid and liquid-solid interfaces (excluding electrode kinetics as such). The subject encompasses processes of fundamental, industrial and environmental importance and includes such topics as the rate of dissolution of reactive gases, lanetics at liquid membranes, metal and solvent extraction, Marangoni effects, heterogeneous catalysis and photocatalysis in solution, and the kinetics of dissolution of minerals and drugs. Softcover 318pp ISBN 0 85186 628 X Price 5i31.50 ($59.50) RSC Members 26.50 No. 78 Radicals in Condensed Phases This publication is primarily concerned with the structure and reactions of ra&cals in liquds and solids. It brings together theoretical work on structure, environmental effects and reactivity with spectroscopic and mechanistic studies directly concerned with radicals. Fundamental aspects are stressed and particular attention is given to new developments including measurements at short time intervals, special solvent effects, and the effects of external fields. A special area of inclusion is electron gain and loss processes, including trapped and solvated electrons, electrochemical reactions, and specific electron-capture and electron-loss in low temperature systems. Photochemical charge-transfer processes are also included. Softcover 347pp ISBN 0 85186 618 2 Price $28.50 ($55.00) RSC Members $6.50 ORDERING Non-RSC Members should send their orders to: The Royal Society of Chemistry, Distribution Centre, Blackhorse Road, Letchworth, Herts SG6 lHN, UK. RSC Members should send their orders to: Assistant Membership Officer, The Royal Society of Chemistry, 30 Russell Square, London WClB 5DT. The Royal Society of Chemistry Burlington House Piccadilly London W 1V OBN (viii)

ISSN:0300-9599    

DOI:10.1039/F198581FP089

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1985, 81, 2569-2575 Reactions of Cobalt(II1) Compounds with some Free Radicals Derived from Uracil BY SUDHINDRA N. BHATTACHARYYA* AND PARIKSHIT C. MANDAL Nuclear Chemistry Division, Saha Institute of Nuclear Physics, Sector- 1, Block-' AF ', Bidhannagar, Calcutta 700 064, India Received 14th March, 1984 Reactions of Co"I complexes with the transient adducts of uracil (U), viz. U-, UH and UOH, have been studied by means of y-radiolysis of uracil in the presence of CoII'EDTA and Co'I'NTA (where EDTA is ethylenediamine tetra-acetate and NTA is nitrilotriacetate). The U-and UH radicals formed by the reactions of eiq and H with uracil reduce Co"' to Co" complexes. The UOH radicals, however, do not reduce the CoIII complexes although such reductions have been observed with Cu'I and FelI1 analogues as reported earlier. The absence of such electron transfer from UOH to CoII' could not be explained on the basis of the redox potential of the COI~~EDTA/CO'IEDTA couple, which was 0.38 V.However, this apparent anomaly has been explained on the basis of the individual structures of the respective complexes. When uracil in dilute aqueous solution is radiolysed, the water-derived radicals (eCq, H and OH) combine with the uracil molecule leading to the formation of the transient adductsl U-, UH and UOH. If metal ions, viz. FeIII and CuII, are present (either uncomplexed or complexed with aminopolycarboxylic acids) in the system, it has been reported earlier2, that such transients undergo electron-transfer process with the metal ions.CoIII constitutes another transition-metal species which has similar redox behaviour as that of FeIII and CuII. Incidentally, in connection with studies of the reactions of electron-affinic radiosensitizers with the radicals derived from pyrimidine bases it has been shown4 that the rate of electron transfer from the radicals to the electron-affinic compounds is dependent on their respective redox behaviour. It is known that the redox potentials of the relevant couples are FelI1/Ferl z 0.77 V,5 (CoIIIEDTA/Co*IEDTA z 0.38 V6 and CuI1/Cu1 z 0.17 V5 (EDTA and NTA repre- sent ethylenediamine tetra-acetate and nitrilotriacetate, respectively). Thus it is evident that the redox potential of the COIIIEDTA/CO~~EDTA couple lies between that of Fe1I1/FeI1 and Cur1/Cu1. It is therefore of interest to study whether the electron-transfer process also occurs with ColI1 ions.EXPERIMENTAL MATERIALS Uracil (Merck) was recrystallized three times from triply distilled water. [14Cz]uracil (14 mCi dm3 mmol-l) was obtained from BARC, Bombay. The Co'I' complexes, Co'IIEDTA and CoII'NTA, were prepared as reported earlier.'g All chemicals and solvents were of analytical- reagent grade. Deaeration was carried out by bubbling high-purity argon gas through the experimental solutions for ca. 30 min. Pure N,O was used throughout the investigation. IRRADIATION Irradiation was carried out with 6oCo y-rays and the dose rate was determined using a Fricke 2569 dosimeter.2570 REACTIONS OF CO"' COMPLEXES WITH URACIL ANALYSIS 2 x mol dmP3 uracil solution containing [14C,]uracil was radiolysed in the presence of the Co'II complexes, and the base degradation yield, G( - U), and the yields of the products were determined after separation of the different products by paper chromatography as described earlier.2* The reduction of the Co'" complexes was followed spectrophotometrically.RESULTS AND DISCUSSION The CoIII species, viz. CoIIIEDTA and CoIIINTA, have appreciable absorption in the range where uracil has maximum absorption, and hence the base decomposition of uracil in presence of the abovementioned metal complexes could not be measured by the direct measurement of loss of absorbance of uracil at the wavelength of its maximum absorption, Amax. However, the degradation yield and the yields of the different products formed in the radiolysis of uracil in the presence of cobalt ions could be determined after separating the products by paper chromatography as described earlier .2 ~ The radiolysis products were found to be similar to those obtained in the radiolysis of uracil in presence of other metal ions,2F3 e.g. CuII and FeIII. A typical plot of the formation of the various products arising from the degradation of the base as a function of absorbed dose is shown in fig. 1. The amounts of products formed were linear with dose, and hence the G values of each product were determined from the slope of the respective straight lines. However, the decomposition of the base was not linear at higher doses, and hence the value of G( - U) was determined from the initial slope of the line. The yields of the different products are shown in table 1, and table 2 shows the effect of varying the concentration of CoIIIEDTA on G(-U) and mol dmY3 uracil solution containing 5 x lop4 mol dm-3 CoIIIEDTA in the region 350-610 nm. As uracil has no absorption in this range, this absorbance is due to CoIIIEDTA, characterised by two peaks having A, at 380 and 535 nm, respectively.Fig. 2(b)-(4 represent the absorbances due to the same solution irradiated in an N,O-saturated medium to doses of 0.75 x 10l8, 1.5 x l0l8 and 3.0 x 10l8 eV ~ m - ~ , respectively. On irradiation the absorbances due to CoIIIEDTA at its absorption maximum decrease as the adsorbed dose increases. Hence, knowing the absorption coefficient, the decomposition yield of CoIIIEDTA, G(-CoIIIEDTA), was measured from the changes in absorbances at 535 nm.When the complex species was CoIIINTA, the decomposition of the complex was determined from the loss of absorbance at its absorption maximum, A,,, - 565 nm. Fig. 3 shows a typical plot for the decomposition of the CoIIINTA with absorbed dose. The decomposition of the complex was linear with dose and hence the G(-ColI1) values were determined from the slope of the straight line. The observed G values for the degradation of the ColI1 complexes are also included in table 1. Table 1 shows that when uracil is radiolysed in presence of CoI'I complexes in argon-saturated solutions, G( - U) z 1.1. Such low G( - U) values indicate that not all the primary radicals are involved in the radiolytic degradation of uracil.However, under the present experimental conditions, i.e. for the radiolysis of 2 x mol dm-3 uracil in the presence of 5 x mol dm-3 CoI1I complexes, the OH radicals react preferentially1* with uracil to give UOH. However, ca. 14.2 and ca. 57.7% of eiq will be scavenged9. lo by CoIIINTA and CoIIIEDTA, respectively. Nevertheless, this partial scavenging of eLq by the cobalt ions does not account for the large decrease in the G( - U) values. Hence it may be assumed that the U- radical, being a very good G( - CoI'IEDTA). Fig. 2(a) shows the absorption spectrum of 2 xS. N. BHATTACHARYYA AND P. C. MANDAL 257 1 6 h 6? - 4 9 ." x 1 0 0 1 2 3 4 Fig. 1. Radiolysis of 2 x 1 0-3 mol dm-3 uracil in the presence of CoIIIEDTA (5 x 1 0-4 mol dm-3) in N,O-saturated solution at neutral pH.@, Degradation of uracil; A, dimer; a, cis-uracil glycol; x , trans-uracil glycol; A, hydroxydihydrouracil. Table 1. Product yields in the radiolysis of 2 x lop3 mol dm-3 uracil in presence of various Co"I complexes (5 x mol dm-3) at neutral pH condition" G(products) A B C D G(dimer) G(cis-uracil glycol) G(trans-uracil glycol) G(hydroxydi hydrouracil) G(isobarbituric acid + dihydrouracil) G(dia1uric acid) G( - uracil) G( - CoI'I complex) 0.20 0.90 0.12 0.30 0.14 0.40 0.13 0.40 0.30 0.60 0.50 0.95 0.12 0.20 0.40 0.70 0.13 0.20 0.20 0.50 0.12 0.20 - - 1.10 2.50 1.20 2.90 2.70 0.70 3.30 1.10 a In the presence of (A) CoII'EDTA in argon-saturated solution, (B) CoIIIEDTA in N,O saturated solution, (C) CoI"NTA in argon-saturated solution and (D) Co"INTA in N,O saturated solution. Table 2.Effect of Co'IIEDTA Concentration on G( - U) and G( - Co'IIEDTA) in the radiolysis of 2 x lop3 mol dmp3 uracil in the presence of Co"IEDTA in deaerated solution at pH ca. 5.5 ~ ~ ~ ~~~ [Co"'EDTA] /lop4 mol dm--3 G( - U) G( - Co'I'EDTA) 3 5 8 10 1.1 1.1 1.2 1.1 3 .O 2.7 2.7 2.82572 REACTIONS OF cO"* COMPLEXES WITH URACIL 0.001 1 I I I 3 50 L1 0 L70 530 590 A/nm Fig. 2. Change in absorption spectrum on irradiation of N,O-saturated uracil (2x lop3 mol dmp3) solution in the presence of 5 x lop4 mol dmp3 CoIIIEDTA at neutral pH. (a) Unirradiated; (b), (c) and (4, irradiated to doses of 0.75 x 10ls, 1.5 x 10l8 and 3.0 x 10l8 eV ~ m - ~ , respectively dose/ 1 0-17 ev c n i ~ ~ Fig. 3. Decomposition of CoII'NTA with absorbed dose in the radiolysis of 2 x lop3 mol dmp3 uracil in the presence of 5 x mol dmp3 CoI'INTA in N,O-saturated solution at neutral PH.S.N. BHATTACHARYYA AND P. C. MANDAL 2573 electron donor,ll transfers an electron to the ColI1 species, as a result of which the metal ion is reduced and the uracil molecule is regenerated : U-+Co'II -+ COII+U. (1) (2) Likewise, the UH radical may also reduce the metal ion: UH + Co'II --+ CoII + UH+. However, the uracil carbonium ion, UH+, does not give rise to product but reverts back to uracil:2p3 H,O UH+ -U+H30+. ( 3 ) It thus follows that no degradation of uracil can be accounted for by the primary reducing radicals. This has been justified from a study of the radiolytic disappearance of CoIII species.If the U- and UH radicals undergo reactions (1)-(3), then CoII should constitute one of the products of radiolysis. This has been proved by the following experiment. When hydrogen peroxide is added to the irradiated solution the absorbance lost due to irradiation is recovered. Since the recovery of absorbance occurs in the absence of any further addition of the ligand, it may be concluded that the product formed constitutes only the CoII species having no ligand degradation. Further evidence in favour of occurrence of reactions (1) and (2) was obtained by following the reduction yield of ColI1 species in the presence of varying concentrations of uracil and CoIII complexes. If the fate of U- and UH is only to react with ColI1 by reactions (1) and (2), then the reduction yield of ColI1 arising from these reactions (together with those of e;ts and H) should be independent of uracil and ColI1 concentrations, and it has been found that G( -CoIr1) is independent of both uracil concentration (fig.4) and CoIII concentration (table 2). From the above discussion it may then be stated that the degradation of the base is achieved primarily due to the subsequent reactions of UOH. The question arises as to whether the UOH radicals undergo reaction with CoI" species by electron transfer or undergo reactions between themselves, as was observed in the absence of a metal ion.12 If the former were the case, as seen2. in the presence of CuII and FeIII, then the G( - U) value would have corresponded to a value of ca. 2.7. However, the observed yield is much lower than this.This reduced yield might have arisen from the disproportionation of the UOH radicals, as has been proposed earlier1, in the absence of a metal ion. It may then be argued that UOH does not undergo electron transfer with Co'II. Were this process effective, the products arising from UOH should depend on the concentrations of CoIII species. However, it is evident from table 2 that the G( -U) value does not change as the ColI1 concentration varies, indicating the non- involvement of UOH in these reactions. That the UOH adduct does not undergo electron transfer with the CorT1 species is further evident from a radiolysis study in N,O-saturated solution. When uracil (2 x mol dmP3 CoIII species in an N,O-saturated solution, the G( - U) value is double that observed in a deaerated solution.The observed G( - U) value of ca. 2.5-2.9 may be explained as being due to disproportionation of the UOH radicals, since the OH yield under these conditions should increase because of the reaction mol dm-3) is radiolysed in the presence of 5 x eiq+N,O -+ OH+OH-+N,. (4)2574 REACTIONS OF CO'" COMPLEXES WITH URACIL 1ci4 1 o - ~ lo-* [ uracil]/mol dm-3 Fig. 4. Effect of uracil concentration on the reduction yield of CoIII in the radiolysis of uracil in the presence of 5 x lop4 mol dm-3 Co"' complexes in Ar-saturated solutions at neutral pH. 0, CoIIIEDTA; a, CoIIINTA. From the observed yield of G( - CoIII) it is evident that eZq does not totally react with N,O. A consideration of the rate-constant data for eZq with N20,13 CoIIINTAlO and Co1I1EDTA9 indicates that G( -ColI1) arising from e;ts corresponds to ca.0.4 and ca. 0.7 for CoIIINTA and CoIIEDTA, respectively. Hence the expected G( - CoIII) values would correspond to 0.4+GH = 1.0 and 0.7+GH = 1.3 in the presence of CoIIINTA and CoIIIEDTA, respectively. The observed results for G( - CoIII) are close to those expected. The absence of electron transfer from UOH to ColI1 in its complex with EDTA, unlike that in case of CuI1 and FelI1, does not concur with its redox potential. This indicates that the redox potential is not the only factor responsible for determining the nature of the electron-transfer process. Differences in the rates of electron transfer in the case of metal complexes, despite the fact that they have similar redox potentials, have also been reported ea~1ier.l~ The mechanism of electron transfer has been thought of as occurring by tunnelling through the potential barrier. Evidently the tunnelling process will depend on the height and width of the potential barrier.The width will also depend on the distance of closest approach of the electron substrate, which will evidently depend on the structure of the metal complex concerned. Thus CoIIIEDTA is known to have a definite octahedral c o n f i g u r a t i ~ n , ~ ~ ~ whereas the corresponding FelI1 and CuII analogues are known to have pentagonal-bipyramida115 and distorted-octahedral15 structures, respectively. These structural considerations indicate that the electron-transfer process should be more difficult in the former case than in the latter two complexes. G. A. Infante, E. J. Fendler and J. H. Fendler, Radiat. Res. Rev., 1973, 4, 301. S. N. Bhattacharyya and P. C . Mandal, Znt. J . Radiat. Biol., 1983, 43, 141. S. N. Bhattacharyya and P. C . Mandal, J. Chem. Soc., Faraday Trans. I , 1983, 79, 2613. C . L. Greenstock, J . Chem. Educ., 1981,58, 156 Handbook of Chemistry, ed. N. A. Lange (McGraw Hill, New York, 1961), p. 1212. Encyclopedia of Electrochemistry of the Elements, ed. A, J. Bard (Marcel Dekker, New York, 1975), vol. 111, p. 53. ' F. P. Dwyer, E. C . Gyarfas and D. P. Mellor, J. Phys. Chem., 1955, 59, 296.S. N. BHATTACHARYYA AND P. C. MANDAL 2575 M. Mori, M. Shibata, E. Kyuno and V. Okubo, Bull. Chem. SOC. Jpn, 1958, 31, 940. M. Anbar, and P. Neta, Int. J. Appl. Radiat. hot., 1967, 18, 493. lo S. N. Bhattacharyya and E. V. Srisankar, Radiat. Res., 1977, 71, 325. l1 J. Cadet and R. Teoule, J. Radiat. Res., 1977, 18, 93. l2 S. N. Bhattacharyya and P. C. Mandal, J. Chem. SOC., Faraday Trans. 1, 1984,80, 1205. l3 J. P. Keent, Radiat. Res., 1964, 22, 1. l4 E. J. Hart and M. Anbar, The Hydrated Electron (Wiley-Interscience, New York, 1970), p. 189. l5 F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry (Wiley-Eastern, New Delhi, 1969), (a) p. 148; (b) p. 849; (c) p. 894. (PAPER 4/41 5)

ISSN:0300-9599    

DOI:10.1039/F19858102569

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. 1, 1985, 81, 2577-2591 Adsorption and Catalytic Properties of CosFe3-s04 Spinels Part 1 .-Preparation and Characterisation of Precursors to Ammonia-synthesis Catalysts BY RAJ R. RAJARAM AND PAUL A. SERMON* Department of Chemistry, Brunel University, Uxbridge, Middlesex UB8 3PH Received 25th April, 1984 Precipitation and dehydration has produced a series of cobalt-iron oxides Co,Fe,-,O, (0 < x < 3) whose surfaces at intermediate x show only marginal iron enrichment and which are useful precursors to ammonia synthesis catalysts. With increasing x activation energies for electrical conduction increase but surface areas estimated by adsorption of N,, H,O and n-hexane decrease (and are significantly different for three adsorbates, S,, < S,, < SHt0, reflecting increasing surface-adsorbate specific interactions). Even for n-hexane enthalpies of adsorption are stronger than expected from mere physical adsorption.The use of non-group VIII metals,l novel metal-support interactions,, metal oxycarbides and nitride^,^ ox~anions,~ intermetallic alloys5 and complexess has yet to affect the commercial catalysis of ammonia synthesis; the same is true for photo catalyst^^ and biologicala catalysts. Unusually, therefore, present commercial catalysts for ammonia synthesisg are not radically different from those of promoted iron developedlO in the first two decades of this century, although doping of the Fe,O, precursor with Co (Ni, Dy or Ce)ll has increased its Fe2+/Fe3+ ratio, decreased its reduction temperature and then increased its subsequent ability to chemisorb N, and synthesise ammonia" at lower pressures.However, our understanding is deficient and does not extend to the precise mode of electronic promotion12 or the extent of N-atom spillover. The present work was undertaken with samples of Co,Fe,-,O, of higher surface area than previously used. These were characterised prior to and after reduction with respect to their potential for ammonia synthesis and their ability to withstand O,, H,O, and CO poisoning and to compare the properties of these polycrystalline surfaces with those of single-crystal faces.14 In general, non-stoichiometry in metal oxides can lead to an oxygen excess or deficiency and p- or n-type semiconducting properties. Cations in p- and n-type semiconductors [e.g.cobalt(n) oxide and iron(m) oxide] have an accessible higher and lower oxidation state. The catalytic activity of such oxides is closely associated with their chemisorption properties, e.g. (i) chemisorption on cations (often more than one) or on anions, (ii) irreversible adsorption of H, or CO with water and carbon dioxide desorbed after reduction and (iii) oxygen adsorption that is more extensive on p- than n-type oxides. Cobalt iron oxides (Co,Fe,_,O, with 0 d x d 3) have the close-packed spinel structure of MgAl,O, with divalent cations in eight of the sixty-four tetrahedral holes surrounded by four 0,- ions and trivalent cations in sixteen of the thirty-two octahedral holes surrounded by six 0,- ions in the normal spinel. Eight tetrahedral sites can be occupied by trivalent cations with the remaining eight trivalent cations and eight divalent cations occupying the octahedral sites in a statistically disordered 25772578 Co,Fe,-,O, SPINELS manner in an inverse ~pine1.l~ Intermediate arrangements can also exist, and t and o subscripts will be used to denote ions at tetrahedral and octahedral positions.The saturation magnetism at room temperature and the high Curie temperature of MeFe,O, spinels (where Me is a divalent cation) have been explained16 in terms of the strong magnetic interactions between trivalent and divalent cations and weak negative interactions between ions of the same sublattice (divalent-divalent and trivalent-trivalent interactions). The resultant magnetic moment is the difference between the non-zero antiparallel magnetic moments of divalent cations and those of trivalent cations (AB,O,).Stoichiometric CoFe,O, contains only Co2+ and Fe3+, but the semiconducting properties of Co,Fe,-,O, prepared by the ceramic method have been investigated with 0.9 < x < 1.1 .17 When x < 1 the excess of ion as Fe2+ acts as an electron donor causing n-type semiconductivity, and when x > 1 Co3+ acts as an electron acceptor causing p-type semiconductivity. Variations in x of 0.05 at x = 1 are sufficient to cause electrical conductivity to change by orders of magnitude and decrease thermoelectric power from 800 to - 600 ,uV K-l.17 Thus p- and n-type samples can be prepared having similar compositions with the concentration of excess electrons and positive holes simply controlled by the concentration of Fe2+ and Co3+, respectively, without the necessity of introducing foreign ions into the lattice.The phase diagram for the Fe-Co-0 systemla suggests that very high oxygen pressures are required to prepare single-phase spinels x > 1 at 1573 K, but in contrast spinels with x < 1 require low-pressure preparations. Resultslg have shown that the rates of reactions (the decomposition of H,O, and N,O and the isotopic exchanges of carbon between CO, and CO and of hydrogen and deuterium between H, and D,) catalysed by cobalt ferrite Co,Fe,-,O, (0.9 < x < 1.1) under oxidising conditions had a maximum value (and those which occurred under reducing conditions a minimum value) and were explained in terms of the geometrical and electrical properties of the solid surfaces.Changes in the thermoelectric power of the cobalt ferrite during the adsorption of hydrogen and oxygen have been investigated20 as a function of catalyst composition; above 393 K hydrogen was adsorbed as an electron donor and oxygen as an acceptor. In each case the cobalt ferrite was prepared by the ceramic method rather than with a high surface area by coprecipitation.21 The present study was initiated to investigate the adsorptive and catalytic properties of iron-cobalt oxides of high surface area prepared by this coprecipi t ation met hod. +t EXPERIMENTAL MATERIALS AnalaR-grade cobalt chloride, ferric chloride and sodium hydroxide (B.D.H.) and laboratory- grade ferrous chloride (B.D.H.) were used for the preparation of spinels. n-Hexane (AnalaR, B.D.H.), N, (white spot, 99.999% purity, B.O.C.) and distilled H,O were used in the adsorption measurements.METHODS OF PREPARATION OF Co,Fe,-,O, The third of the available methods for the preparation of spinels was used here: (i) oxide calcination of the oxides at 1273-1 573 K, (ii) evaporation of metal nitrate solutions and decompositionsalcination at 573-673 K and (iii) coprecipitation of metal hydroxides from chloride solutions and decompositionxalcination at 373 K of the homogeneous precipitate to give a high-surface-area product : xCo2+C1, - 6H,O + (1 - x) Fe2+C1,. 6H,O + 2Fe3+C1, * 6H,O + 8NaOH = Co$+ Fef+, Feg+(OH), (gel) + 8NaCl+ 12H,O (0 d x d 1)R. R. RAJARAM AND P. A. SERMON 2579 xCo2+C1, - 6H20 + (3 - x) Fe3+C1, - 6H20 + (9 - x) NaOH Co2+ Fe;?, Fei+ (OH), = Co:+ Fet?, Fei+U, + 4H,O Coz+ Feif, (OH),-, + (x - 1) 0, (air) = Cof+Coi?, Fe;?, 0, + (5 + x) H,O (1 < x d 3).The solubility products of Co(OH),, Fe(OH), and Fe(OH), at 298 K are 6.3 x 8.0 x and 2.0 x a pH of at least 10 is needed for the complete coprecipitation of Co2+ and Fe2+, while Fe3+ begins to precipitate even in slightly acidic solutions of pH 4.22 Hence a pH of ca. 12 is required for production of the cobalt ferrite. The hydroxides should form a true solid solution, since the cationic radii in low spin-states are similar (Co2+ = 0.065 nm, Fe3+ = 0.055 nm).23 Seven samples of the iron-cobalt oxide series Co, Fe3-, 0, with x equal to 0.0, 0.6, 1.0, 1.5, 2.0, 2.4 and 3.0 were prepared.100 cm3 of a solution containing 0.15 mol dm-3 of metal ions as the metal chlorides in distilled deaerated water was rapidly added with stirring to 50 cm3 of deaerated freshly prepared 6 mol dmV3 NaOH solution at room temperature and the pH (which ranged from 8 for x = 0 to ca. 1 I for x = 3) was adjusted to 12.5 by addition of NaOH. The dark gel formed was dehydrated on a steam bath for 1 h. The reaction mixture became clear at 368 K and a black spinel powder precipitated and was then filtered, washed repeatedly with distilled water until free of alkali, dried (at 363 K for 16 h) and crushed in an agate mortar. = Co:+ Fegt, (OH),-, (gel) + (9 - x) NaCl + 6H20 (1 < x < 3) (0 < x < 1) METHODS OF CHARACTERISATION Total Co and Fe contents were estimated gravimetrically using the zinc oxide separation whilst the concentrations of Co3+ and Fe2+ were analysed according to the method devised by Smilten~.~, X-ray analyses of the oxide samples were obtained with iron-filtered Co Ka, radiation (A = 0.179 nm) using a powder camera (1 1.48 cm diameter) with 24 h exposure times and a microdensitometer or an X-ray diffractometer (Philips PW 1050 X-ray generator at 50 kV and 24 mA) scanning 0.5” 28 per min.Aluminium powder (B.D.H.) was used as reference. X-Ray photoelectron spectra of oxide samples mounted onto double-sided sticky tape were obtained using a Kratos ES 300 spectrometer with aluminium Ka radiation (1486.6 eV) below a residual pressure of 1.33 Pa. For a powder composed of perfect crystalline particles of minimum dimension d [given by KA/pp cos 8, where pp is the breadth at half maximum height of the X-ray diffraction profile due to particle size, K is a constant called the shape factor (0.893)25 and A is the wavelength of the incident radiation] and assuming Gaussian diffraction profiles, then to a first appro~imation~~ where Po is the observed line breadth, pi the intrinsic line breadth and P the line breadth due to instrumental factors (which can be determined by use of aluminium powder with zero particle-size broadening and strain broadening, Pp = /3, = 0).The average effective strain present in a spinel sample can also give rise to broadening p, (= 25 tan 8) of the powder diffraction lines. Thus Pi cos 8 = (KA/d) + 2( sin 8 and so pi cos 8 was plotted against sin 8; plots were linear and d was estimated from the intercept KA/d.Thermogravimetric (t.g.a.) and differential thermal analyses (d. t.a.) were performed simultaneously on a Stanton Redcroft STA780 with oxide samples (ca. 70 mg) and the reference (a-alumina) contained in Pt-Rh crucibles. Constant heating rates (8 K min-l) and gas flow rates (30 cm3 min-l) were used throughout. Samples were flushed with N, for 15 min and then flowing H, or N, during temperature programming to 873 K. Transmission electron microscopy was undertaken with oxide samples ultrasonically dispersed in acetone and then placed on carbon-coated copper grids using a Jeol Temscan 100 CX instrument at < 0.1 mPa. The electrical conductivity 0 of a semiconductor decreases exponentially with increasing reciprocal temperature26 [i.e.B exp ( -EO/2kT), where E, is the activation energy for conduction, k is Boltzmann’s constant and B is a constant] and so E, was calculated from the gradient of a plot of logloo against 1/T. A Wayne-Kerr conductivity bridge was used for measuring the conductivity ( W l cm-l) of the samples for which homogeneity and packing problems had been L? = E-p”) = Vp+PJ22580 Co,Fe,-,O, SPINELS overcome using a Teflon-insulated cell to contain samples (1 cm thickness, 1 cm2 area and 2 g weight) while under compression (2000 kg ern-,) at 293-333 K. Pore size distributions were measured using mercury (Carlo Erba 200) up to 197 mPa after outgassing (room temperature, < 1 Pa). Nitrogen adsorption measurements of surface area were undertaken on a semi-automatic Carlo Erba Sorptamatic (series 1800) instrument at 77 K.Water-vapour adsorption was measured gravimetrically (after buoyancy correction) using a quartz-spring balance of the McBain-Baker type in which the quartz-spring was maintained at 299.65 to f0.02 K. The pressure was measured by a transducer connected to a digital millivoltmeter. Outgassing to < 13.3 mPa was achieved ultimately with an oil diffusion pump. n-Hexane adsorption was followed in a flow chromatographic system using the retention-volume method. This involved the determination of the retention volume of a given concentration of vapour by switching a mixture of n-hexane and carrier gas (white spot N,, B.O.C.) into a clean column of adsorbent and measuring the amount retained.Nitrogen was purified by removing oxygen by passage through MnO, on celite catalyst,' and drying by passage through 5A molecular sieve held at 195 K. The flow rate of the gas stream was adjusted to 30 cm3 min-l by a Hoke needle valve. The purified carrier gas was then saturated with hexane at temperatures selected between 233 and 298 K and constant to kO.05 K using a Grant liquid kerosene thermostat. The flow rate of the hexane-saturated gas stream was adjusted to 30 cm3 min-l. Thermal-conductivity detectors (GOW Mac No 01106P) were used. The detector block was maintained at 298 K and was equilibrated with the incoming gas stream. After leaving the reference arm of the katharometer, the gas stream entered the Pyrex U-tube sample reactor, also held at 298 K.The sample bed produced no significant pressure gradient. A bypass valve permitted a reduction in the flow through the sample bed which was adjusted to 10 cm3 min-' and held constant throughout all experiments. After passing the sample, the gas stream was directed to the measuring arm of the katharometer. All dead-space volumes were kept to a minimum. The gas stream was saturated with different concentrations of n-hexane. Each sample (ca. 0.1 g) was placed in the sample tube, pretreated in situ with N, flowing (30 cm3 min-l), heated to 373 K and maintained at that temperature for 1 h. The sample was then cooled to 298 K in N,. The gas line to the sample was then closed and the N, stream was saturated with hexane at a fixed vapour pressure. The flow rate was readjusted to 30 cm3 min-'. At equilibrium the flow was split between the sample tube and the bypass line, the flow through the sample being maintained at 10 cm3 min-l.Adsorption was followed on the recorder chart until equilibrium was established and the sample valve was then closed. The temperature of the saturator was then increased slightly and the procedure for adsorption was repeated. The desorption curve was subsequently followed by subjecting the oxides to a N, stream saturated with hexane vapour at a relative pressure close to unity until equilibrium had been achieved. Points at successively lower hexane pressures on the desorption curve were then ascertained by successively decreasing the concentration of hexane in the gas stream and observing the amount desorbed.To determine the absolute amount of hexane adsorbed known volumes of hexane were added to the flow of pure N, through an injection port near the sample tube and a calibration curve of peak area against volume injected was plotted. p / p o was calculated from the vapour pressure of hexane at the bath temperature and the saturation pressure of hexane at 298 K (20.187 kPa). The cross-sectional area of hexane was assumed to be 0.54 nrn2.,* RESULTS AND CHARACTERISATIONS STRUCTURAL AND CHEMICAL ANALYSIS Bulk chemical analyses of Co,Fe,-,O, are given in table 1 ; all seven samples show only slight deviation from the intended composition and so the intended values of x 0, 0.6, 10, 1.5, 2.0, 2.4 and 3.0) are used to denote samples. The lattice parameter calculated (+O.OOl nm as a result of the substantial line broadening) decreases as x increases (ie.as cobalt replaces iron in the spinel lattice the unit cell becomes smaller), in agreement with previous data,21 and confirms that the iron-cobalt oxide spines have indeed been formed (see table 1 and fig. 1). TheTable 1. Compositions of samples of Co,Fe,-,O, prepared in this study P 7d intended X.r.d lattice composition, parameter, sample Fe (%> Co (%> X total Fe2+ Fe3+ total Co2+ Co3+ 0 (%) actual composition a/nm ( & 0.00 1) ~ 0.0 72.13 20.32 51.81 0 0 0 27.87 84.16% Co,Fe,O, + 15.84% Fe203 0.840 0.6 56.68 9.07 47.61 15.09 15.09 0 28.23 96.01 % Co,~,,Fe,~,,O,+ 3.99% Fe,O, 0.834 1 .o 47.61 0 47.61 25.12 25.12 0 27.27 100% Co,., Fe2.,04 0.833 1.5 34.75 0 34.75 36.67 24.27 12.40 28.58 100% Co,.,Fe,~,O, 0.827 2.0 26.56 0 26.56 46.46 25.58 20.88 26.98 99% Co,.,,Fe,~,,O, 0.823 2.4 1 I .68 0 11.68 62.53 25.60 36.93 25.79 85.85% Co,.,, Fe0.5604 + 14.15% Co,O, 0.8 I8 3.0 0 0 0 71.96 23.89 48.07 28.04 98% Co,.,Fe,O, 0.8082582 Co,Fe,-,O, SPINELS 1 1 I I 0 1 2 3 X Fig.1. Variation of spinel lattice parameter a with x in Co2Fe3--504: 0, values obtained here; 0, values obtained by Tseung et al. and Yate.21 Table 2. Binding energies of X.P.S. peaks observed for Co,Fe,-,O, samples corrected binding energy'/eV for x = binding- energy shift 0.0 0.6 1 .o 1.5 2.0 2.4 3.0 /ev Fe 2p; 723.9 725.0 724.8 725.0 725.2 725.3 - + 0.9- + 2.3 Fe 2pg 711.1 711.0 710.8 711.0 711.1 711.0 - +0.8-+ 1.1 c o 2p; - 794.4 794.1 794.4 794.6 794.7 795.0 +1.@-+ 1.9 co 2p: - 779.8 779.3 779.8 779.9 780.1 780.3 +0.7-+ 1.7 0 1s 530.0 530.0 530.0 530.0 530.0 530.0 530.0 - a Normalised to C 1s at 285 eV.separation of the observed X.P.S. C,, peak from its true binding energy of 285 eV gave a charging correction for all the other peaks. The difference between the corrected binding energies and their elemental positions29 indicates the oxidation state of that element (see table 2) in each oxide sample. Co 2p1/2 and Co 2p312 peaks (at cn. 794 and 780 eV) showed satellite peaks (at ca. 802 and 787 eV), but the Fe 2pll, and 2p,,,, peaks (at ca. 724 and 711 eV) only showed weak satellite peaks with Co,Fe,O,. The 0 Is peak was clearly visible at 530 eV and was of constant intensity for all the samples. Peak overlap made it difficult to differentiate the + 2 or + 3 cation oxidation states.The Co 2p3,, peaks were first deconvoluted from their shake-up satellite peaks. The peak areas (calculated from the full width at half height multiplied by height) were divided by a sensitivity factor, normalised to the 0 Is peak, and normalised to 100% to obtain the atomic proportions of each element. The atomic proportions divided by the atomic weight of each element yield the molar proportions. Fig. 2 shows the surface composition deduced from X.p. spectra as a function of the bulk composition determined by the wet analysis. The X.P.S. results show that there is only a slight ironR. R. RAJARAM AND P. A. SERMON 2583 0 2 4 (Co/Fe)b,,k Fig. 2. Surface and bulk Co/Fe ratios estimated by X.P.S.and chemical analysis for samples of differing x. The dotted line indicates the relationship expected in the absence of surface enrichment. enrichment at the surface of all the mixed iron-cobalt oxides compared to the bulk composition, in agreement with Parravano30 and theoretical prediction^.^^ D.t.a. and t.g.a. studies of the different iron cobalt oxides from 293 to 1073 K were made both in N, and in H,. Runs in N, showed only minimal effects (ca. 373 K) associated with loss of water or extraneous moisture. The weight losses observed for reduction in H, (27.69, 28.63, 29.12, 29-18, 27.14, 27.35 and 27.34% for x = 0, 0.6, 1 .O, 1.5, 2.0, 2.4 and 3.0) were only slightly greater than the theoretical weight losses associated with complete reduction (27.54, 27.43, 27.29, 27.1 1, 26.94, 26.76 and 26.59% for x = 0, 0.6, 1.0, 1.5, 2.0, 2.4 and 3.0) and it was therefore assumed that the d.t.a.-t.g.a.curves in fig. 3 represented complete sample reduction. At x = 0 a small exotherm is shown at 438 K (associated with reduction of Fe203 to Fe304) and an endotherm with a maximum at 873 K (associated with reduction of Fe304). At x = 0.6 exotherms and endotherms were at 673 and 848 K, respectively. At x = 1.0 the exotherm at 673 K became more intense while the endotherm appeared at a lower temperature (820 K). Reduction of Co,~,Fe,~,O, started at 473 K with two exotherms at 590 and 670 K, followed by an endotherm now at 770 K. Reduction profiles for x = 2.0 and 2.4 were similar to that of C O , ~ , F ~ , ~ , ~ ~ with the exotherms at ca.590 and 670 K becoming more intense and the endotherm less intense and broader and shifting to a lower temperature. Reduction of Co,O, started at 473 K with a broad exotherm at 625 K (associated with the reduction of C0,04) and 690 K (associated with the reduction of COO to Co). Therefore in spinels with low cobalt content (i.e. with x = 0.6 and 1.0) only the second exotherm in reduction of cobalt ions appears, probably because these spinels contain only Co2+. As expected, as cobalt replaces iron in the spinel with increasing x the endotherm shifts to lower temperature, indicating that the iron(IIrFiron(I1) ions are more easily reduced. It could be that cobalt is being reduced first to the zero-valent state and then promotes further reduction.The values of specific conductivity of different iron-cobalt oxides were measured with increasing temperature, and fig. 4 shows how the activation energy E, for2584 Co,Fe,-,O, SPINELS cj X cj -a P) 27 3 473 673 873 10 '3 0 - -10 E v) O m c.' - 2 0 - 30 27 3 17 6 6 73 873 1073 T/K Fig. 3. (a) D.t.a. and (b) t.g.a. profiles for different samples with different values of x in Co,Fe,-,O, in hydrogen: 0, 0; 0, 0.6; A, 1.0; A, 1.5; 0, 2.0; a, 2.4 and 0, 3.0. conduction varies with x. CooFe30, has the lowest activation energy for conduction (observed as 0.08 eV: the literature value is 0.05 eV). The activation for electrical conduction then increases (reaching a maximum value of 0.31 eV for CO,.~F~,~,O~), and above x = 1.5 the activation energy increases as the value of x increases (reaching a value of 0.53 eV for Co,.,Fe,O,, in agreement with earlier reports2'* 32).Co5Fe3-,04 shows normal semiconductor behaviour ; its conductivity can be related to the presence of different valency states of iron and cobalt. In this respect it belongs to the groups of controlled valency semiconductors.33 PARTICLE-SIZE ANALYSIS Samples of iron-cobalt oxides were first heated in air at 373 K for 5 h, since experiments carried out without this pretreatment gave irreproducible broadening, probably owing to the strain; 373 K was selected since this induced no significant sintering. Average crystallite sizes deduced from intercepts of plots of pi cos 8 against sin 8 are given in table 3 ; clearly as x increases the average crystallite size becomes larger.When the metal hydroxides are dehydrated at 373 K the particle size of the resultant oxides may depend on the number of OH- attached to each metal cation in the hydroxide lattice. The formation of one Fe304 molecule involves the expulsion of nine OH- from the hydroxide lattice, whilst only six OH- are expelled for the formation of CO304. Thus the resultant Fe,O, on this basis is expected to be moreR. R. RAJARAM AND P. A. SERMON 2585 0.6 / P / / / / / 0.4 - / / I ,d' % Lip 1 0.2 / / 0 / / / / / / / / / 0 / / I / 0 / 1 2 3 A Fig. 4. Activation energy for electrical conduction E, as a function of x in Co,Fe,-,O,. Table 3. Average crystallite sizes (dav), surface areas (SxR,,) and strain of Co,Fe,-,O, samples determined by X-ray diffraction line broadening av.intercept, 4, strain density S,,, X KLldav /nm /lo-, rad /g cm-, /m2 g-' 0.0 0.0 17 9.4 4.5 5.180 123.22 0.6 0.0 165 9.7 3.5 5.302 116.67 1 .o 0.0140 11.4 4.0 5.410 97.29 1.5 0.0 124 12.9 4.3 5.632 82.58 2.0 0.0 106 15.1 4.4 5.758 69.01 2.4 0.0095 16.8 4.5 5.890 60.64 3.0 0.0080 20.0 3.5 6.070 49.42 finely dispersed than Co,O,. It is possible to calculate the surface area S (m2 g-l) of the different iron-cobalt oxides from X-ray line broadening assuming that the samples are made of non-porous cubic or spherical particles of minimum dimensions d(nm) since S = 6 x 103/pd, where p (g cm-,) is the density of different iron-cobalt Table 3 gives the surface areas S thus calculated and shows that the measured densities are remarkably close to those of the bulk solids.SURFACE-AREA ANALYSIS AND PHYSICAL ADSORPTION Complete adsorption isotherms for N,, H,O and C,H,, were carried out on four of the iron-cobalt oxides (x = 0, 1, 2 and 3), but surface areas were calculated from adsorption points in the B.E.T. linear range on all of the iron-cobalt oxides. Isotherms2586 Co,Fe,-,O, SPINELS 0 0.3 0.6 0.9 PlPo Fig. 5. Extent of hexane adsorption qHC measured on different samples of Co,Fe,-,O, (where 0, A, 0 and 0 denote samples where x = 0, 1.0, 2.0 and 3.0, respectively) as a function of relative pressure at 298 K. Filled symbols denote desorption points. 0 0.3 0.6 0.9 PlPo Fig. 6. Extent of N, adsorption qNp measured on different samples of Co,Fe,-,O, (symbols as in fig. 5) as a function of relative pressures at 77 K.were always of type 11, with a type H3 (IUPAC) hysteresis loop.28 Fig. 5 shows the adsorption isotherms for hexane upon Co,Fe,-,O, with varying values of x at 298 K. Fig. 6 shows the full N, isotherms at 77 K. The similarity in the shape of the isotherms could be due to similarity in the physical nature of the surface and the pore shape. The closure of the hysteresis loop at p / p o w 0.4 of the four oxides is similar for both adsorbates. The water-vapour adsorption isotherms for the four oxides studied are shown in fig. 7. These are also similar in shape to hysteresis loops which now close at p / p o w 0.30, possibly due to slow chemical interaction between the adsorbate and the oxide, causing rehydroxylation of the surface. All isotherms suggest the presence of a wide range of pore sizes.An analysis of the different isotherms for the B.E.T. surface areas and the C values are given in table 4, which also gives the surface areas SXRD calculated from the average particle size using X-ray line broadening. Fig. 8 plots the B.E.T. surface areas of the oxides as a function of the composition for each of the adsorbates; as x increasesR. R. RAJARAM AND P. A. SERMON 2587 0.3 0.6 PIP0 0.9 Fig. 7. Extent of H,O adsorption qH20 measured on different samples of Co,Fe,-,O, (symbols as in fig. 5) as a function of relative pressure at 299 K. Table 4. Surface areas (m2 g-l) of Co,Fe,-,O, samples estimated by application of B.E.T. theory to adsorption of n-hexane, dinitrogen and water vapour ' and also X-ray diffraction line broadening 0.0 108 175 140 125 171 90 123 0.6 1 04 168 135 115 165 75 116 1 .o 96 1 40 127 121 142 80 97 1.5 82 158 110 112 138 60 82 2.0 53 120 87 110 122 65 69 2.4 42 115 77 85 90 73 60 3.0 33 95 64 75 85 56 49 (i.e.as cobalt replaces iron in the spinel) the surface area decreases (CooFe304 has the highest surface area and Co3Fe,0, has the lowest surface area). A similar pattern is obtained from X-ray line broadening, and surface areas thus derived correlate well. The apparent reason for this general trend in the surface areas of the oxides, S, with x is the particle size, which was shown to increase from CooFe304 to Co3Feo0,. However, the actual surface area estimated for each oxide varies depending on the adsorbate used. Thus SHIO > SNZ > SHC. It could be argued that such behaviour is a reflection of the cross-sectional areas a, of the adsorbates [i.e.H 2 0 (0.106 nm2) < N,(O.162 nm3) -c HC(0.541 nm2)]. SHzO/SN2 is approximately con- stant. However, there is good evidence28* 34 from B.E.T. analysis of physical adsorption on many oxides that water (with a dipole moment of 1.85 D) interacts more specifically than N, (or hexane) with surface hydroxyl groups and that hydroxylation processes lead to higher apparent surface areas for this polar adsorbate. Even water adsorption on Lewis-acid sites leads to rehydroxylation and surface H+.35 Measurements of the physical adsorption of N, on 77K on various samples after heating at various temperatures for 3 h in flowing argon (20 cm3 min-l) (before cooling in argon to2588 Co,Fe,-,O, SPINELS 1 1 I M N E t, \ 0 1 2 3 X Fig.8. Surface areas of samples of Co,Fe,-,O, with differing values of x measured by H,O (0), N, (0) and hexane (0) adsorption and X-ray line broadening (H). Table 5. Surface areas of Co,Fe,-,O, samples determined by N, adsorption at 77 K after pretreatment at various temperatures SN2/m2 g-' at T/K = X 323 373 473 573 673 773 873 0.0 139 138.5 137 136.2 118 81 69 0.6 135 135 133 132 109 75 55 1 .o 128 127 126 126 96 68 48 1.5 110 111 110 108 84 61 40 2.0 87 86 88 85 75 52 37 2.4 78 77 76 76 64 49 35 3.0 63 61 60 61 55 41 28 Table 6. Pore characteristics of Co,Fe,-,O, samples total pore volume, V % Vpresent in pores with * /cm3 g-l range of radii/nm X penetration 4-25 25-250 250-7500 0.0 0.6 1 .o 1.5 2.0 2.4 3.0 1.69 1.58 1.72 1.75 2.07 1.91 1 .so 53 40 7 30 45 25 27 54 19 29 60 11 15 67 18 11 72 17 9 70 21R.R. RAJARAM AND P. A. SERMON 2589 , I 1 I 0 1 2 3 p,clkPa Fig. 9. Enthalpies of hexane adsorption at 298 K as a function of partial pressure on samples of Co,Fe,-,O,. It is unlikely that the heat capacity of the samples varies sufficiently to modify the order of enthalpies (kJ mol-l) in table 7. (Symbols as in fig. 3.) ambient temperature) show (see table 5) that the surface areas of all samples decline as the temperature rises from 323 to 873 K, but particularly above 600 K. The extent of loss of surface due to sintering does not appear significantly affected by the value of x. POROSITY ANALYSIS An analysis of the average pore radii and pore volumes of the different iron-cobalt oxides is given in table 6.The results obtained clearly demonstrate that the range of pore sizes obtained varies between the different iron-cobalt oxides, although most of them exhibit pore sizes over the entire range of mercury porosimetry (i.e. > 3.75 nm). In general, as x increases (as cobalt replaces iron in the spinel) the total pore volume in the pore-radius range 3.75-25 nm tends to decrease, while the total pore volume in the pore-radius range 25-250 nm increases. This leads to a general decrease in surface areas, in agreement with measurements of physical adsorption. CONCLUSIONS Homogeneous iron-cobalt oxides with general formula Co,Fe,-,O, with 0 < x < 3 have been prepared successfully by coprecipitation and have a spinel structure of general formula (Al-yt Aye) (But B,-uo) 0, where A represents the divalent ion and B the trivalent ion.In normal and inverse spinels y = 0 and 1. The lattice parameter decreases and the average crystallite sizes increase as cobalt replaces iron. The increase in particle size has been associated with the number of OH groups lost during the2590 Co,Fe3-,O4 SPINELS Table 7. Extents (qHC) and enthalpies ((IHc) of adsorption of n-hexane at 298 K and 3.5 kPa on samples of Co,Fe,-,O, -QHC qHCa X /pmol g-l /J g-l /kJ mol-l 0.0 332.05 9.10 27.40 0.6 319.76 8.25 25.80 1 .o 295.16 11.30 38.28 1.5 252.13 12.80 50.77 2.0 162.96 7.05 43.26 2.4 129.13 6.10 47.24 3 .O 101.46 5.30 52.24 a Estimated from data in fig. 5 at 298 K and 3.5 kPa (i.e. p / p o = 0.17).dehydration process. Their surface composition (x # 3) shows slight surface enrichment by Fe. Heating the oxides in a N, atmosphere to 1073 K shows only a slight thermal effect associated with loss of extraneous moisture, indicating no phase change. Reduction in H, occurred at a lower temperature at higher cobalt concentrations. The specific conductivity B of the iron-cobalt oxides increases with temperature, demonstrating the semiconducting properties of these oxides. Hexane, nitrogen and water-vapour adsorption isotherms on these characterised iron-cobalt oxides were of type I1 with a type-B hysteresis loop. Their surface areas decrease as the cobalt concentration increases. Why the physical adsorption of water is so much more specific than hexane or why N, suggests larger than expected areas will be discussed in a future paper.It was of interest to determine whether even the enthalpies of hexane adsorption on Co,Fe3-,04 samples were within the range normally expected for physical adsorption (< 20 kJ mol-l). A Dupont 990 differential scanning calorimeter (precalibrated against the melting of indium in flowing N,) was used to measure the enthalpies of hexane adsorption on such samples at 290 K from flowing N, at varying hexane partial presses (0.5-17 kPa) using a hexane saturator at fixed but variable temperature. These results are shown in fig. 9. Dividing the amounts of heat liberated by the extents of hexane adsorbed indicated in fig. 5 and table 7 at 3.5 kPa hexane (p/p, = 0.17) gave enthalpies per mol adsorbate.These increased as x increased (as shown in table 7) from those expected for physical adsorption to a more specific and stronger interaction. These homogeneous oxides of varying Co/Fe ratio and unusually high surface area are particularly suitable to an investigation of the parameters affecting the chemisorption and catalytic properties of semiconductors and also parameters affecting ammonia-synthesis catalysts, the result of which will be reported. N. D. Spencer and G. A. Somorjai, J . Catal., 1982, 78, 142; G. Rambeau, A. Jorti, H. Amariglio, J . Catal., 1982,74, 1 10; Appl. Catal., 1982,3,273 ; G. Rambeau and H. Amariglio, J. Catal., 1981,72, I . * J. Santos and J. A. Dumesic, Stud. Surf. Sci. Catal., 1982, 11, 43; J. Santos, J. Philips and J. A. Dumesic, J.Catal., 1983,81, 147; A. Bossi, F. Garbassi, G. Petrini and L. Zandevighi, J. Chem. SOC., Faraday Trans. I, 1982, 78, 1029; European Patent 5853 1 (1 982). European Patent 530 18 (1 982). L. A. Chernysheva, V. Ya. Zabuga, N. S. Pivovarova and M. V. Tovbin, Vim. Kiiv. Univ., Ser. Khim., 1981, 22, 59.R. R. RAJARAM AND P. A. SERMON 259 1 W. E. Wallace, J. France and A. Shamsi, Rare Earth Modern Science and Technology, 1982, 3, 561 ; V. K. Yatsimirskii, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1982, 4, 13 1. J. Shen, Z. Wangand X. Huang, Sci. Sin., 1983,26B, 1 ; G. E. Bossard, A. T. George and R. K. Lester, Inorg. Chim. Acta, 1982, L227; US Patent 4309311; C. Shao, J. Li, L. Pau, X. Tang, M. Bei and H. Guo, Fund. Organomet. Chem. Proc. China-Jpn-US, 1982, 715; G.P. Pez, P. Apgar and R. K. Crissey, J. Am. Chem. Soc., 1982, 104, 482. ’ M. Schlavello, L. Rizzuti, A. Sclafani, I. Mayo, V. Angugliaro and P. L. Yue, Adv. Hydrogen Energy, 1982,3,821; F. Khan and P. Yue, Ind. Eng. Chem., Prod. Res. Dev., 1983,22,238; Japanese Patent 169126(1981);V. Angugliaro, A. Lauricella, L. Rizzuti; M. Shiavelloand A. Sclafani. Int. J. Hydrogen Energy, 1982, 7, 845; 851. Biotechnol. Lett., 1982, 4, 647. A. Kazusaka and I. Toyoshima, Z. Phys. Chem. ( Wiesbaden), 1981,128, 1 1 1 ; European Patent 60622 (1982); British Patent 2094169; USSR Patent 1009508 (1983); Netherlands Patent 82031 17 (1982). lo H. B. Leach, Appl. Ind. Catal., 1983, 2, 128; B. Timm, 8th int. Congr. Catal. (Berlin, 1984) (Verlag Chemie, Weinheim, 1985).l1 M. D. Dimitrov and R. Ruser. God. Vissh. Khimikotechnol. Inst., SoJia, 1980,26, 103; 11 1 ; 120; 126; R. Brown, M. E. Cooper and D. A. Whan, Appl. Catal., 1982, 3, 177; P. J. Smith, D. W. Taylor, D. A. Dowden, C. Kemball and D. T. Taylor, Appl. Catal., 1982, 3, 303; V. K. Yatsimirskii, V. N. Kovalenko and E. V. Ishchenko, Ukr. Khim. Zh., 1982,48, 614. G. Ertl, S. B. Lee and M. Weiss, Surf Sci., 1982, 114, 527. l3 G. Ertl, CRC. Crit. Rev. Solid State Muter. Sci.,1982, 349. l4 K. H. Huang, X. M. Zeng and J. T. Li, J. Catal., 1983,81, 259; G. Ertl, D. Prigge, R. Schloegl and M. Weiss, J. Catal., 1983, 79, 359; A. Baranski, J. M. Lagan, A. Pattek and A. Reizer, Appl. Catal., 1982, 3, 201; 207; G. N. Demina, L. A. Rudnitskii, A. M. Alekseev, P. D. Rabina, L. D. Kuznetsov and A. A. Bondareva, Kinet. Katal., 1982, 23, 451. l5 T. W. Barth and E. Posnjak, 2. Kristallogr., 1932, 82, 325. l6 L. Neel, Ann. Phys. Paris, 1948, 3, 137. l 7 G. H. Jonker, J. Phys. Chem. Solids, 1959, 9, 165. J. Robin and J. Benard, C. R. Acad. Sci., 1952, 234, 734; B. D. Roiter and A. E. Paladino, J. Am. Ceram. Soc., 1962, 45, 128. l9 G. Parravano, 4th Int. Congr. Catal., (Moscow 1968) (Akademiai Kiado, Budapest, 1971, p. 149). *O R. G. Squires and G. Parravano, J. Catal., 1963 2, 324. 21 A. C. C. Tseung and J. R. Goldstein, J. Muter. Sci., 1972, 7, 1383; J. R. Yates, B.Sc. Industrial 22 W. J. Schuele and V. D. Deetscreek, in Ultrafine Particles, ed. W. E. Kuhn (John Wiley, New York, 23 A. F. Wells, Structural Inorganic Chemistry (Oxford University Press, Oxford, 1984), p. 313. 24 J. Smiltens, J. Am. Chem. Soc., 1957, 79, 4881. 25 P. A. Sermon, J. Catal., 1972, 24, 467. 26 J. A. Becker and C. B. Green, Trans. Electrical Engineering, 1946, 65, 71 1. 27 C. R. McIllwrick and S. G. Phillips, J . Phys. E, 1973, 6, 1208. 28 S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Porosity (Academic Press, London 1982); A. A. Isinkyan and A. V. Kiselev, J. Phys. Chem., 1961, 65, 601; K. S. W. Sing, D. H. Everett, R. A. Haul, L. Moscou, R. A. Pierotti, J. Roquerol and T. Siemienicwika, Pure Appl. Chem., 1985, 57, 603. 29 D. Briggs, Handbook of X-Ray and Ultraviolet Photoelectron Spectroscopy (Heyden, London, 1978). :lo F. Garbassi, G. Petrini, L. Pozzi, G. Benedek and G. Parravano, Surf. Sci., 1977 68, 286. :I1 J. Haber and J. Stoch, Muter. Sci. Monogr., 1982, 10, 418. :I2 G. H. Jonker, J . Phys. Chem. Solids, 1959, 9, 165. A. R. West, Solid State Chemistry and its Applications (Wiley, Chichester, 1984), p. 515. J4 E. M. McCafferty and A. C. Zettlemoyer, J . Colloid Interface Sci., 1970, 34, 452; Discuss. Faraday SOC., 1971,52, 239; R. E. Day, G. D. Parfitt and J. Peacock, Discuss. Faraday Soc., 1971,52, 215. 35 S. R. Morrison, in Catalysis Science and Technology, ed. J. R. Anderson and M. Boudart (Springer, Berlin, 1982), chap. 4. Chemistry Project Report (City University, London, 1968). 1963), p. 224. (PAPER 4/674)

ISSN:0300-9599    

DOI:10.1039/F19858102577

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985, 81, 2593-2603 Adsorption and Catalytic Properties of CozFe3-z04 Spinels Part 2.-Hydrogen Chemisorption on Precursors to Ammonia Synthesis Catalysts BY RAJ R. RAJARAM AND PAUL A. SERMON* Department of Chemistry, Brunel University, Uxbridge, Middlesex UB8 3PH Received 25th April, 1984 At temperatures in the range 323-423 K H, adsorbs largely irreversibly onto the surfaces of Co,Fe,-,O, (0 < x < 3) spinels to an extent (0.91-4.76 molecules nm-,) which increases with x and temperature. Although the enthalpy of adsorption (60-210 kJ mol H,) is greater than bulk spinel reduction at these temperatures the latter process is not thought to be significant. Adsorption isotherms obey the Langmuir model well. Crystal-field considerations suggest sites responsible for adsorbing H, which are contrary to recent reports (J-P.Beaufils, Y. Barbaux and B. Saubat, J. Chem. Soc., Chem. Commun., 1982,1212). Cobalt additions improve the extent and enthalpy of the fast adsorption of hydrogen on these spinel precursors of non-promoted ammonia-syn thesis catalysts. Chemisorption of hydrogen on oxides, which is more complex than on transition metals, can be either reversible or irreversible (being thermally desorbed as water after adsorbent reduction). Studies by Garner1 have shown that the fraction of hydrogen reversibly chemisorbed at room temperature varies from oxide to oxide. While irreversible chemisorption of CO on oxide surfaces at room temperatures leaves the surfaces unsaturated with respect to oxygen, Garner1 suggested that hydrogen may be adsorbed on metal oxides with the simultaneous formation of surface hydroxide and a valence change at the adjacent nearby cation: $I2 + 02---MX+ + OH--M(5-1)'.Well characterised2 samples of Co,Fe,-,O, (0 d x d 3) were selected for this study of the nature of H, chemisorption on oxides because they could be prepared in a homogeneous and high-surface-area state with varying properties determined by the value of x and also because of their relevance to ammonia-synthesis catalysts., It is common for H, chemisorption to impede ammonia synthesis, which has N, chemi- sorption and dissociation as the rate-determining step. Conversely the influence of the electronic structure on the adsorptive properties of the iron-cobalt oxides Co,Fe,-,O, (0 d x < 3) could be elucidated from data for the chemisorption of hydrogen. The oxides prepared in the series not only have variable metallic cations but also a different distribution pattern between the divalent and trivalent cations in the tetrahedral and octahedral sites; such that Co,Fe,O, has an inverse spinel structure and Co,Fe,O, has a normal spinel structure.Dowden, has advocated correlation of catalytic and chemisorptive activity with the crystal-field stabilisation energy of the cation as a result of introducing an additional adsorbate ligand, and this approach has been used to interpret the results obtained. 85 2593 F A R 12594 HYDROGEN CHEMISORPTION ON Co2Fe3--504 EXPERIMENTAL MATERIALS Argon (99.98 % purity, B.O.C.) and hydrogen (white spot, 99.9% purity, B.O.C.) were further purified in adsorption measurements.H, adsorption measurements were expected to be extremely sensitive to 0, and H,O contaminants and so care was taken to ensure their removal. Ar was passed through a bed of 5A molecular sieve at 195 K to remove moisture and then through a bed of Mn0,-celite4 to remove traces of 0, (to < ppm) while H, was passed through a bed of 6% Pd-Al,03 to remove oxygen impurities and then a bed of 5A molecular sieve at 195 K to remove the moisture produced. Co,Fe3-,04 samples were prepared and characterised as described previously.2 FLOW METHODS OF MEASURING H, CHEMISORPTION A pulse-chromatographic system was used to study the interaction of hydrogen with the oxides in a manner similar to that used previou~ly.~ It consists essentially of two identical reactors through which flow the same gas stream at an equal rate.The sample was held in a silica reactor in a furnace at the adsorption temperature (k 5 K) or was programmed at 4-10 K min-l. The purified Ar stream could then flow through the adsorption cell or bypass it, and hydrogen pulses could be introduced via a constant- but changeable-volume gas-sampling valve. Since Ar and H, were of substantially different thermal conductivity (0.0162 and 0.1684 W m-l K-l, respectively) the concentrations of the latter could be determined with a hot wire or thermal-conductivity detector. The purified Ar carrier gas was split into two separate streams whose flow rates were carefully maintained equal and constant.One of the gas streams then flowed through a sample valve with an H, injector of selectable volume and thence to the sample reactor. The other gas stream was admitted directly to the reference reactor. Leaving the reactors, the gas streams were admitted to a gas chromatograph fitted with two stainless-steel columns with a detector current of 200 mA. The dead-space volume of the adsorption reactor and the entire volume between the injection point and the detector was kept to a minimum (ca. 18 cm3). Exit-gas flow rates were monitored using soap-film flow meters. Oxide samples (ca. 0.2 g) formed beds of < 5 mm thickness on a sinter in the silica reactor. Bed temperatures were measured directly on a chromel-aluminel thermocouple. SELECTION OF PULSE SIZE AND CARRIER-GAS FLOW RATE Clearly the flow rate of the carrier gas will be important in defining the contact time between the sample and the reactive gas.The effect of the Ar flow rate on the detector output was investigated by studying the H, uptake at 373 K by Co3Fe,04 using different flow rates and an H, pulse volume of 0.2 cm3. The sample was pretreated by the standard method described later using a fresh sample of the oxide for each flow rate. Successive pulses were admitted until peaks of identical heights were obtained. The time between pulses was maintained at 2 min. In general, the first pulses of reactive gas were adsorbed totally, then peaks of intermediate height were observed, followed by peaks of constant height. The volumes of H, consumed were then obtained by adding fractions not detected at the exits of the reactor.The volume of H, gas, VH,, which is taken up is given per gram of the oxide at 101.325 kPa and at the temperature of the experiment as: n 1000 VH2/cm3 g-' = C ( U , - U l ) / K l = ~ m i = 1 where n is the number of peaks, U, is the area of the first peak, U , is the area of constant peak height, Kl is the constant for hydrogen calibration and m is the mass of oxide used in mg. Dividing by the specific surface area for B.E.T. measurements gives the extent of adsorption per unit area. The amount of hydrogen, qH2, (calculated from VH2) taken up was found to be independent of the argon carrier-gas flow rates (up to 30 cm3 min-l) for Co,Fe,-,O, (x = 3), which previously has been shown2 to have the lowest specific surface area; thereafter qHP decreased with increasing flow rate and decreasing time of contact with the oxide.The linearity of the thermal-conductivity detector for H, pulses was good at carrier-gasR. R. RAJARAM AND P. A. SERMON 2595 flow rates of 20 and 30 cm3 min-l when the reactor contained no adsorbent. Therefore in the present study an Ar carrier-gas flow rate of 20 cm3 min-' and a pulse volume of 0.2 cm3 were selected. SELECTION OF CONDITIONS OF SAMPLE PRETREATMENT It has been shown previously2 that none of the samples of Co5Fe3--504 sinters in flowing Ar at temperatures up to 573 K in a period of 3 h. Therefore in all studies reported here samples (ca. 0.2 g) of each oxide were first pretreated in flowing Ar at 573 K for 2 h and then brought to the temperature at which H, uptake was to be considered. Adsorption and consumption of H, were studied at temperatures ranging from 323 to 673 K using the above method and are now described.VOLUMETRIC METHOD OF MEASUREMENT OF H, CHEMISORPTION The extent of hydrogen chemisorption, qH2) was also measured in the volumetric vacuum system described previously6 for sample Co,Fe,-,O, (x = 1.5) as a function of hydrogen pressure at 373 K. The sample was first outgassed to 1.3 kPa at 573 K for 2 h before the temperature was reduced to 373 K and adsorption measurements commenced. D.S.C. MEASUREMENT OF ENTHALPIES OF H, CHEMISORPTION Reactive gas streams were introduced at 30+2 cm3 min-* to a Dupont 990 differential scanning calorimeter after purification as described for the flow-adsorption system, except that beds of 5A molecular sieve were held at ambient temperature rather than 195 K.Samples (25-35 mg) of oxides were placed in A1 pans (an empty A1 pan being used as the reference). The d.s.c. cell was flushed with Ar and brought to 573 K for 2 h. The sample temperature was then reduced to the temperature at which H, adsorption was to be followed (in the range 323-423 K). Exotherms were monitored at constant temperature, flow rate and pressure (101 kPa) of white-spot H,. Exothenn areas were calibrated against endotherms for indium melting at 429.6 K under the same conditions of gas flow. Calibration coefficients varied depending upon whether Ar, N, or H, was used. RESULTS EFFECT OF TEMPERATURE The results from pulse-chromatographic methods for Co,Fe,-,O, (x = 1.5) will be given as an example of the observations made for each of the oxide samples.Three different thermal regions could be differentiated depending upon the strength of H,-spinel interaction. At 323-423 K hydrogen was irreversibly chemisorbed on CO,~,F~,~,O,; peaks for the first few pulses of hydrogen were absent, then peaks of intermediate areas and finally constant area were observed. So, after a certain amount of hydrogen had been adsorbed, further pulses of hydrogen were not consumed. Fig. 1 shows that the amount of H, adsorbed per unit area of Co,Fe,_,O, (x = 1.5) increases linearly with temperature (323-423 K). In order to establish the irreversible nature of this H,-spinel interaction at these temperatures, this oxide was maintained for 1 h at the adsorption temperature in a stream of flowing argon after completion of the adsorption measurements; at the end of this period hydrogen was once more introduced.There was no reproduction of the adsorption processes at 232 and 373 K, since an insignificant amount of H, was consumed in this second cycle. However, at 423 K ca. 3% of the amount of H, originally adsorbed ( i e . 0.41 cm3 8-l of Co,.,Fe,.,O,) was taken up in this second cycle. It is probable that this arises from a slight desorption of H, (although no negative desorption peaks were observed), simultaneous migration of hydrogen into the oxide and oxygen to the surface or the consumption of hydrogen held in the first cycle by reduction. However, thermal analysis indicates2 that such reduction starts only above 473 K.At 448-523 K the interaction between pulses of hydrogen and the oxide Co,~,Fe,~,O, produces different 85-22596 HYDROGEN CHEMISORPTION ON Co,Fe,-,O, I I I 32 3 373 423 T/K Fig. 1. Dependence of the maximum extent of H, adsorption on a unit area of Co,Fe,-,O, (x = 1.5) on temperature determined by pulse-chromatographic methods, where total spinel areas were estimated by B.E.T. analysis of N, adsorption at 77 K. X Fig. 2. Maximum extent of H, adsorption on a unit surface area of different samples of Co,Fe,-,O, at 0 , 3 2 3 ; a, 373 and ,423 K determined by pulse-chromatographic methods, where total surface areas were estimated by B.E.T. analysis of N, adsorption at 77 K.R. R. RAJARAM AND P. A. SERMON 2597 Table 1. Maximum extent of H, adsorption determined by pulse chromatography no.of H, molecules adsorbed per nm2a X 323 K 373 K 423 K 0.0 0.9 1 1.57 2.56 0.6 1.06 1.93 2.77 1 .o 1.18 2.02 2.98 1.5 1.72 2.83 3.86 2.0 2.29 3.47 4.07 2.4 2.53 3.68 4.37 3.0 2.95 3.92 4.76 a Surface areas were estimated by N, adsorption at 77 K on each oxide after thermal treatment in Ar at each temperature. types of response; first pulses were completely consumed and then a diffuse peak for water appeared which could not be quantified. It is evident that surface reduction had started at the three temperatures studied at 448-523 K, while at the three temperatures studied at 548-723 K the bulk reduction of Co,~,Fe,~,O, appeared to be occurring with initial H, pulses totally consumed followed by the appearance of diffuse water peaks.These temperatures of reduction agree well with those from thermal analysis., Thus reduction above 448 K (surface and bulk) prevents quantitative analysis of H, adsorption on this and the other samples of Co,Fe3-,04; the final uptake of H, is probably controlled by the rate of regeneration of oxygen atoms on the surface, which is in turn governed by the diffusion of oxygen from the crystal lattice to the surface of the oxide. In general, results obtained for other samples of Co,Fe,-,O, of different x were similar to those described above (except for Co,Fe,O, where the diffuse peaks for water started appearing only above 498 K, in agreement with t.g.a.,). Therefore H, consumption on all samples of Co,Fe,-,O, spinels was restricted to the temperature range 323-423 K, where irreversibility is not thought to suggest bulk reduction.EFFECT OF SPINEL COMPOSITION For all the oxides studied in the Co,Fe,-,O, series the amount of hydrogen chemisorbed was observed to increase with temperature. However, at each temperature the amount of hydrogen taken up varied with the value of x (see fig. 2). At each of the three temperatures studied the extent of hydrogen adsorption increases with increasing cobalt/iron ratio, with the least uptake at x = 0 and the highest uptake at x = 3. Table 1 gives the number of H, molecules adsorbed per nm2 of each oxide at the three temperatures which have been calculated using the specific surface areas of each sample at each temperature calculated from B.E.T.analysis of N, adsorption at 77 K.,j VOLUMETRIC AND CHROMATOGRAPHIC RESULTS The validity of the data obtained for H, chemisorption by the pulse-chromatographic system was cross-checked by comparing data obtained for hydrogen chemisorption on the pressure-volumetric system for the oxide Co,~,Fe,~,O, at 373 K. Under these conditions the rate of H, chemisorption was very fast. Fig. 3 illustrates the qH2/pH,2598 HYDROGEN CHEMISORPTION ON Co2Fe3--504 N E 0.6 d I - 0 % 5 0.4 . h X cr '" s 0.2 0 PH,/kPa Fig. 3. H, adsorption isotherm at 373 K on Co,Fe,-,O, (x = 1.5). adsorption isotherm thus obtained. The Langmuir equation8 for non-dissociative adsorption is qH2 = bPHp tg=------- qH,(max) (l -k bpHz) or PH,/qHz = (I/bH,(rnax)) -k (PH,/qH,(max)) where b is the ratio of adsorption to desorption rate constants.Extrapolation of adsorption data at 0.66-3.3 kPa to zero pressure may be assumed to give 'monolayer' capacity qH,(max) from the zero-pressure intercept. The isothermal data clearly obey the Langmuir equation (as shown by the linear Langmuir plots obtained) despite the likelihood that the spinel surface was notlenergetically hopogeneous and varies from other Langmuir assumptions. A plot of pk,/qH, against p h , was equally linear. From the gradient of the linear plot, the monolayer capacity was calculated as 5.33 pmol m-2, whereas a value of 5.10 pmol m-2 is obtained by extrapolation of the linear portion of the adsorption isotherm in fig. 3 to zero equilibrium pressure. These values agree reasonably well with the values of 4.7pmol m+ obtained by the pulse-chromatographic system.There is thus agreement and consistency betweenR. R. RAJARAM AND P. A. SERMON 2599 I h x- 100 ; t I 50t 1 1 I 0 1 2 3 X Fig. 4. Variation of enthalpies of H, adsorption with x in Co,Fe,-,O, and temperature (symbols as fig. 2). pulse-chromatographic and volumetric measurements of H, adsorption on these samples under conditions where reduction is not significant. ENTHALPY OF H, ADSORPTION 'The enthalpy of H, adsorption on samples of Co,Fe,-,O, was measured with a differential scanning calorimeter at temperatures (323-423 K) where spinel reduction was not observable. Enthalpies were determined (after indium calibration) as J g-l oxide and were converted to kJ mol-1 H, adsorbed by dividing by the corresponding maximum extent of H, adsorption on each oxide at each temperature determined by pulse-chromatographic methods. Fig.4 shows the variation of enthalpy of H, adsorption with x and temperature assuming that there was no major change in the heat capacity of the spinels with the value of x. The enthalpy of adsorption of hydrogen on all the iron-cobalt oxides increased with an increase in temperature. At any given temperature studied the enthalpy of hydrogen adsorption is approximately constant for the oxides between Co,Fe,O, to CoFe,O, and then increases with increasing cobalt content, to a lesser extent at 323 K. At 323 K enthalpy values for hydrogen adsorption are between -64 and -74 kJ mol-l, while at 423 K the values are - 180 to -209 kJ mol-l.Although the heat of complete reduction of Co,O,, calculated from the heat of formation of H,O and c0304, is - 15.63 k 2.5 kJ mol H;' (a smaller value than that observed by d.s.c. for the adsorption process) and that for reduction of Fe,O, is endothermic (1 27.8 kJ r n ~ l - l ) , ~ it is concluded that the processes observed here were mainly chemisorp tion.2600 HYDROGEN CHEMISORPTION ON Co5Fe3-,O4 1 ---------- -I- +--L ---- - ------------- I I t - - A ------ - - - - - - - - - 2q 1 I l o w x I h i g h x I r--------- I H2 ' I I L--I-------I-- ------ ---I Fig. 5. Possible modes of non-dissociative (a) or dissociative-heterolytic (b) adsorption of H, on cobalt cations in the latter involves the formation of an OH group. In addition, Fe cation sites will also be involved.Thermal rearrangement of cations may occur over the spinel sites at the temperatures used here. Table 2. Gain in crystal-field stabilisation energy (ACFSE/A,) upon some transformations of configurationa ~ ~~ ACFSE/A, electron states oc ta hedrai- octahedral- of tetrahedral- pentagonal square cation octahedral bipyramidal pyramidal high-field states d5 (Fe3+) 1 .oo -0.17 - 0.09 d 7 (Co2+) 1.20 -0.53 0.1 1 d 5 (Fe3+) 0 0 0 d 7 (Co2+) 0.20 0.26 0.11 ds (Co3+, Fe2+) 1.60 -0.85 - 0.40 low-field states d6 (Co3+, Fez+) 0.10 0.13 0.06 a The data of Dowden and Wells15 differ significantly from those above. DISCUSSION The extent and enthalpy of isothermal hydrogen chemisorption on the iron-cobalt oxides increase with increasing cobalt/iron ratio and with temperature (in the range 323-423 K).In this temperature interval hydrogen is irreversibly chemisorbed on the oxides. Chemisorption on oxides can be explained with reference to either their semiconductor or electronic properties. Boundary-layer theory suggests that n-type semiconductors (e.g. Co,Fe,-,O, when x = 0) show higher affinity for hydrogen than p-type oxides (e.g. Co2Fe3--504 when x = 3), but it should be remembered that the oxides used here were stoichiometric and had not been heated in oxygen. The conductivity of zinc oxide and nickel oxidelo below 373 K is unaffected by hydrogen chemisorption, and it has therefore been suggested that this process is not uia transfer of a single electron. Nevertheless, H, adsorption does affect the magnetic260 1 R.R. RAJARAM AND P. A. SERMON properties of Fe,03 and Fe304, and this may be evidence in favour of adsorption by electron transfer.ll9 l2 The extent of hydrogen chemisorption increases with increasing x and cobalt/iron ratio in the spinels, which can be explained by the ease of this electron-transfer process13 and changes in crystal-field stabilisation energy. Previously3J4 the catalytic activity of the first-row transition-metal oxides in H,-D, exchange and cyclohexane dehydrogenation has shown a maximum at 3dsP7 (and 3d3), which Dowden and Wellsf5 explained in terms of changes in the crystal-field stabilisation energy. Dowden suggested that H, adsorbs uia intermediate (b) in fig. 5 , which is similar to those considered in studieP of the isotopic adsorption of H, and D, on MgO.However, it has recently been suggested17 that Co3+ (3d6) in tetrahedral sites of strong crystal field is more active in H, chemisorption (and H,-D, exchange) than Co2+ and receives an electron from the adsorbate, converting it into Co2+. However, there seems to be clear evidence that CO adsorbs onto Co2+ sites18 and that CO adsorption poisons the COO, surface towards subsequent H, chemisorption and this could be taken to suggest that H, also needs to chemisorb on the same sites as CO. Therefore the evidence is equivocal. Fig. 5 shows the different cobalt adsorption sites (Octahedral and tetrahedral) on the surface of Co,Fe3-,04 samples, and table 2 indicates the gains in crystal-field stabilisation energies on changing from one configuration to another under conditions of high and low field.First, it is clear that given a choice between either tetrahedral or octahedral sites those having the greatest preference for octahedral sites are Co2+ > Co3+, Fe2+ > Fe3+ under low-field con- ditions (i.e. Co2+ will have the greatest preference for octahedral sites) or Co3+, Fe2+ > Co2+ > Fe3+ under high field conditions. Thus one traditional method of con- sidering lattice ion positions within the 32 octahedral and 64 tetrahedral sites within the spinel unit cell, has been19 to calculate octahedral site preference energies (OSPE) from differences in the octahedral and tetrahedral crystal-field stabilisation energies. Such OPSE calculations for cations in transition metal oxides show a decrease in the sequence d7 Co2+ (30.9 kJ mol-l) > d6 Co3+, Fe2+ (16.7 kJ mol-l) > d5 Fe3+ (0 kJ mol-l); this corresponds to a weak field condition in table 2.However, the true CFSE for spinels is not known and in addition Goodenoughlg has noted that even if there is ordering of ferrous and ferric ions in the octahedral sites of the inverse spinel Fe304 below 120 K, at higher temperatures a random arrangement prevails. Further, the OPSE are changed in strong fields (i.e. Co3+, Fe2+ > Fe3+ > Co2+). Although crystal field splittings caused by 0,- ion ligands are smaller for 3d metals than those for 4d and 5d metals, precise values of the splitting for spinels are not well known. These uncertainties mean that it is not possible to predict precisely the distribution and location of the different spinel cations in the lattice sites.Never- theless, it is often assumed that, under normal conditions, Co3+ and Fe2+ have preference for octahedral sites in the normal CO304 and inverse Fe304, i.e. a relatively high field is involved. CONCLUSIONS The total numbers of octahedral and tetrahedral sites (between which there may be some interconversion20) per unit surface area are likely to increase only slightly as x increases, as a result of the lattice parameter decreasing by 4% ,, and cannot explain the entire increase in the adsorption capacity of these samples as x increases. However, the occupancy of these sites depends upon the value of x (i.e. for x < 1 Co is found2 to be present predominantly as Co2+ but as predominantly Co3+ at higher x) and the local field strength felt.Thus if it is assumed that high field conditions prevail within these spinels (irrespective of the value of x) then the CFSE values in table 2 suggest2602 HYDROGEN CHEMISORPTION ON Co,Fe,-,O, that when x = 0 Fe3+ in the tetrahedral sites will preferentially convert to square pyramidal by H, adsorption (and that Fe3+ in octahedral sites cannot permit this adsorption without first losing an 0,- ligand and in so doing reverting to Fez+). Equally, when x = 3 within the normal CO,~,, Co2+ in tetrahedral sites can adsorb H, and convert to a square pyramidal form with substantial gains in CFSE; but this is not the case for Co3+. At high and low values of x under these conditions the most reactive adsorption sites may therefore be Fe3+ and Co2+ in tetrahedral sites, respectively.Table 2 shows that the latter tetrahedrally coordinated cation gains the most CFSE and this may explain why the enthalpy of H, adsorption increases with x. However, these sites may involve heterolytic adsorption of H, and non-dissociative adsorption of H,, respectively. Previo~slyl~ it has been suggested that Co3+ in tetrahedral sites is active in H, adsorption upon high surface area samples similar to those used here. In the light of the above results this does not seem very likely. Nevertheless, CFSE considerations remain uncertain and may not dominate. Thus it would be possible at x > 1 for H, adsorption to be predominantly on Co3+ sites (with a smaller CFSE gain, see table 2) and heterolytic-dissociativeZ1 [see fig.5(6)] and irreversible. In addition, the prevailing field may be weaker or thermal repopu- lation of cations over tetrahedral sites and octahedral sites may occur,1g in which case a wider range of adsorption sites may be involved. Nevertheless, bulk reduction under the conditions studied here is not thought to be significant. If such homogeneous spinels represent reasonable precursors to non-promo ted ammonia-syn thesis catalysts then hydrogen chemisorption is very fast thereon, and increases in strength and extent of hydrogen chemisorption by adding cobalt may have a beneficial effect on ammonia synthesis. 22 W. E. Garner, J. Chem. SOC., 1947, 1239. R. R. Rajaram and P. A. Sermon, J. Chem. SOC., 1985, 81, 2569. D. A. Dowden, N. McKenzie and B.M. Trapnell, Adv. Catal., 1957, 9, 65. C. R. McIllwrick and S. G. Phillips, J. Phys. E, 1973, 6, 1208. T. Paryjczak and J. Rynkowski, Pol. J. Chem., 1980, 54, 1931. A. R. Berzins, M. S. W. Vong, P. A. Sermon and A. T. Wurie, Adsorption Science and Technology, 1984, 1, 51. S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Porosity (Academic Press, New York, 1982). I. Langmuir, J. Am. Chem. SOC., 1918,40, 1361. M. A. Richard, S. L. Soled, R. A. Fiato and R. A. DeRites, Mater. Res. Bull., 1983, 18, 829. lo F. S. Stone, Adv. Catal., 1962, 13, 2. l1 S. Umeda, Gifu Daigaku Kogakubu Kenkyo Hokoku, 1980,30, 97. I3 P. A. Sermon, unpublished data. l4 G. Dixon, D. Nicholls and H. Steiner, in Proc. 3rd Znt. Congr. Catal. (Elsevier, Amsterdam, 1964), l5 D. A. Dowden and D. Wells, Act. 2me Congr. Catal. Int. (1961), vol. 2, p. 1489; D. A. Dowden in l6 T. Ito, T. Murakami and T. Tokuda, J. Chem. SOC., Faraday Trans. I , 1983, 79, 913. l7 J. F. Bailly-Lacresse, M. Guilbert, A. Dhuysser, F. Moriamez and J. P. Beaufils, Bull. SOC. Chim. Fr., 1969, 1073; J-P. Beaufils and Y. Barbaux, J. Appl. Crystallogr., 1982, 15, 301; J-P. Beaufils, Y. Barbaux and B. Saubat, J. Chem. SOC., Chem. Commun., 1982, 1212. R. G. Squires and G. Parravano, J. Catal., 1963, 2, 324. vol. 1, p. 815. Chemisorption and Catalysis, ed. P. Hepple (Institute of Petroleum, 1970), p. 1. la R. R. Rajaram and P. A. Sermon, J. Chem. Soc., Faraday Trans. 1 , to be published. l9 J. K. Burdett, G. D. Price and S. L. Price, J. Am. Chem. SOC., 1982,104,92; J. B. Goodenough, Proc. R. SOC. London, Ser. A, 1984,393,215; A. F. Wells, Structural Inorganic Chemistry (Clarendon Press, Oxford, 1984), p. 595; A. R. Wells, Solid State Chemistry and its Applications (Wiley, Chichester, 1984); N. N. Greenwood, Ionic Crystals, Lattice Defects and Non-stoichiometry (Butterworths, Seven- oaks, 1968); J. D. Dunitz and L. E. Orgel, Adu. Inorg. Radiochem., 1960, 2, 1 ; J. D. Dunitz and L. E. Orgel, J. Phys. Chem. Solids, 1957, 3, 318.R. R. RAJARAM AND P. A. SERMON 2603 2o S. Siegel, J. Catal., 1973, 30, 139. ** Z. Knor, in Catalysis Science and Technology, ed. J . R. Anderson and M. Boudart (Springer, Berlin, 1982), vol. 13, chap. 5. ** M. D. Dimitrov and R. Ruser. God. Vissh. Khimikotekhnol. inst., Sofia, 1980,26, 103; 1 1 1 ; 120; 126; R. Brown, M. E. Cooper and D. A. Whan, Appl. Catal., 1982, 3, 177; P. J. Smith, D. W. Taylor, D. A. Dowden, C. Kemball and D. T. Taylor, Appl. Catal., 1982, 3, 303; V. K. Yatsimiriskii, V. N. Kovalenko and E. V. Ishchenko Ukr. Khim. Zh., 1982, 48, 614. (PAPER 4/675)

ISSN:0300-9599    

DOI:10.1039/F19858102593

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I, 1985, 81, 2605-2626 Aspects of Temperature-programmed Analysis of some Gas-Solid Reactions Part 1 .-Dispersion Effects in Temperature-programmed Bulk Reduction and Temperat ure-programmed Desorp tion BY NIGEL D. HOYLE, PAUL H. NEWBATT, KEITH ROLLINS, PAUL A. SERMON* AND ALPHA T. WURIE Department of Chemistry, Brunel University, Uxbridge UB8 3PH Received 9th May, 1984 The reduction of hexachloroplatinic acid and the desorption of adsorbed hydrogen have been followed by temperature-programmed methods. Both the temperature-programmed bulk reduction (t.p.b.r.) and the temperature-programmed desorption (t.p.d.) profiles reveal several maxima at temperatures which vary with the ultimate Pt dispersion and the support used. Reductions commenced at subambient temperatures and good separation of all three t.p.b.r.peaks was only achieved on starting t.p.b.r. at least as low as 223 K. T.p.b.r. temperatures of maximum reduction rates for samples supported on porous silica decreased with increasing dispersion but were higher for the supported rather than the unsupported state. The maximum quantities of hydrogen consumed in t.p.b.r. exceed that expected for complete reduction in some silica-supported samples, but the contribution of the high-temperature y-peak (possibly associated with spillover) increased with increasing Pt dispersion. T.p.d. liberated less gaseous hydrogen (41-84% ) from silica-supported samples than the total quantity of preadsorbed hydrogen, but the percentage increased (and the temperatures at which desorption rates were a maximum decreased) as the Pt dispersion increased. Alumina- and titania-supported samples produced more complex t.p.d.and t.p.b.r. profiles than those of silica. It is argued that it is not justified to interpret these or other equivalent temperature-programmed profiles in terms of processes occurring on chemically distinct and distinguishable solid or surface states. The activation energy for diffusion is critically important and diffusion may be too slow in t.p.b.r. and too fast in t.p.d. to allow temperature-programmed profiles to arise from processes occur- ring to such distinguishable and identifiable species. In t.p.d. surface residence times are also important in defining the fraction of adsorbate subsequently desorbed.Temperature-programmed methods of analysing gas-solid reactions,' which are receiving increasing use and attention, include temperature-programmed desorption (t .p.d.) of a species preadsorbed at a lower temperature, temperature-programmed surface reaction, (t.p.s.r., including titration) and temperature-programmed bulk reaction (t.p.b.r., mainly reduction). In each case the rate of reaction between gaseous and solid states is measured as a function of the temperature of the solid sample (and time) as this temperature is raised linearly with increasing time. Maximum rates of consumption of gaseous reactant or production of gaseous products or maximum rates of heat liberation are observed at different temperatures T,. Here we intentionally restrict our consideration of these methods to platinum-based catalysts.Assuming that desorption is not masked by diffusion, decomposition, reaction or readsorption, t.p.d. peak areas and temperatures have been used2 to reveal the populations and desorption activation energies of different adsorbed states. However, C2H, adsorbed on silica-supported Pt3 may be desorbed as CH, at 369 K and as C,H, at 498 K after decomposition and self-hydrogenation, and the four t.p.d. peaks seen3 26052606 TEMPERATURE-PROGRAMMED REACTIONS for but-1-ene adsorbed on Pt have been associated with adsorbate fragmentation before desorption. Therefore the diffusivity and reactivity of the adsorbate during temperature programming may need to be considered. Desorption of spiltover adsorbate may also occur.4* High inert-gas flow rates are thought to minimise readsorption and diffusion,l* 2 * while shape-index analysis has been applied' to consider the role of diffusion limitation (both interphase and intraparticle) and whether this is rate determir~ing.~ Extensive t.p.d.of hydrogen preadsorbed on Pt29 3 9 9 9 lo has revealed the extent, activation energy and temperature of desorption; thus two or more hydrogen t.p.d. peaks have been observed,2T4 indicating different sites and modes of hydrogen attachment on Pt (a, and p,) with those appearing at lower temperatues being the more weakly adsorbed. l1 Nevertheless, it is clear that there is a dynamic equilibrium between thep, +p2 and also H, (gaseous) on Pt( 1 1 1),l1 which must result in part from relatively rapid surface diffusion between the p1 and p, states.Extending the association of t.p.d. peaks with distinguishable adsorbed states, other workers have correlated t.p..d. peak areas with catalytic activity where the catalysed reactions use the same adsorbed l1 Increasing the dispersion of Pt on silica may decrease the temperature at which hydrogen is de~orbed,~ where the activation energy for hydrogen desorption decreases from 50 to 34 kJ mol-l as the average Pt particle size decreases from 4.3 to 1.6 nm.9 The initial coverage of the adsorbate also affects the t.p.d. profile.12 T.p.s.r. has been applied13 to the reaction of hydrogen preadsorbed on alumina- supported Pt with CO, where maximum rates have been related to different modes of CO chemisorption. Temperature-programmed titration of hydrogen preadsorbed on Pt with a gaseous alkene is also reported14 to give maxima at similar relative values of T, to t.p.d.of hydrogen on the same samples. Temperature-programmed methanation and alkene oxidation and chlorination have also been discussed.13~ l5 T.p.s.r. maxima have been related to the number and activity of surface sites measured by continuous isothermal methods. T.p.b.r. was introduced16 to separate and analyse the reducible content of a solid from the temperatures at which the rate of reduction (measured by the rate of hydrogen consumption) was a maximum during linear temperature programming. 9 l7 Values of T, and areas of t.p.b.r. peaks can be used to identify and quantify these reducible components with greater accuracy than under isothermal conditions.T.p.b.r. of both bimetallic and monometallic Group VIIIB metal catalysts has been considered. Increasing the dispersion of oxide-supported Pt salts may change T, observed on t.p.b.r.l69 l7 but comparisons may be complicated by the use of different preparations and supports. Reductive and oxidative t.p.b.r. have also locatedlg Group IB metal particles in zeolites. Two t.p.b.r. peaks for reduction of CuO on silicala, l9 have suggested separate reduction of a copper silicate and dispersed CuO, while broad t.p.b.r. peaks are exhibitedz0 for reduction of CuII in zeolites. Never- theless, at least t.p.b.r. can indicate average changes in solid oxidation state. Despite uncertainties, temperature-programmed methods are convenient and rapid for analysis of solids,l particularly when they are too highly dispersed for other physical methods.However, they must be of most value when applied in conjunction with other methods quantitatively (rather than merely as qualitative fingerprints) to determine the number and nature of the solid or surface species involved in the processes related to each observed peak. Each temperature-programmed method considers the rate of desorption or reaction r between the solid or surface S and the gaseous G reactants which have orders of s and g :N . D. HOYLE, P. H. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2607 rtpd = exp (- &,,IRT) rtpsr = A N s ~ X P (-Etpsr/RT) rtpbr = exp ( - % p b ~ - / ~ ~ ) * r is measured as a function of temperature (defined by the programming rate 2 which equalsdT/dt) andanalysis time t.Tequals T, whendrldt = 2 d[S]/dT = d[G]/dT = 0 and these values can be converted into activation energies17t21 in the unlikely event that the pre-exponent term A is independent of temperature and the orders s and g are unity, since then: 2 In Tm-ln Z+ln [GI, = (Ej/RTm)-(ln A-ln Ej+ln R) and plotting (2 In T, -In Z+ln [GI,) against the reciprocal of T, gives the relevant activation energy Ej. Experimentally, values of T, are determined for different values of 2 and this is attractively simple, although errors introduced by changing thermal lag at different Z must be insignificant. However, s and g are unlikely to be unity, and indeed g has been found2, to approach zero for t.p.b.r. of V,O, in H,.In addition, mass transfer is known2, to induce large errors in estimates of A from t.p.d. 'To analyse the validity of temperature-programmed methods, two of these (t.p.b.r. and hydrogen t.p.d.) have been applied to the analysis of oxide-supported Pt and its precursor chlorometallic acid where their dispersion is intentionally varied, but 2 and other operating conditions have been held constant. EXPERIMENTAL MATERIALS All silica- (F, G, I, K and L), alumina- (M) and titania- (N) supported catalysts were prepared and characterized by chemisorption and physical methods, as described previously. 23 Samples L, M and N used non-porous supports. Only sample K was reduced before any use and all other samples had been prepared using hexachloroplatinic acid (Specpure, Johnson Matthey).All samples (except K) were analysed prior to reduction by t.p.b.r. and after reduction samples were analysed by t.p.d. of H,. In addition t.p.b.r. was also applied to the unsupported chlorometallic acid, although its hygroscopicity made the determination of the exact weight used difficult. EXPERIMENTAL METHODS In t.p.b.r. (reduction) the reactive gas stream (6% H2+94% N, at 101 kPa) or N, was passed through a Pt/A1,03 catalyst to remove 0, and then dried with molecular sieve at 195 K before entering the reference arm of a katharometer. This then passed (40k0.02 cm3 min-l) to the silica reactor containing the catalyst sample uia a two-way injection valve and a solid- CO, + acetone trap and thence to the other arm of the katharometer. The two-way injection valve (loop volume 0.81 cm3) was used to introduce H, calibrant to the detector. The sample of the catalyst (whose weight varied slightly with the Pt loading) within the reactor was held in a furnace which could linearly raise (Stanton Redcroft programmer LVP/CClO/R) the sample temperature (measured to k 0.01 K using a chromel-alumel thermocouple encased in thin-walled silica in contact with the bed).Results below ambient temperature could not be obtained with 2 strictly constant. The reactor was held in a CO, +acetone thermostat at 195 K or in an ice + salt + water slurry at ca. 258 K. After thermal equilibration, the gaseous reductant mixture was introduced and the sample reactor was allowed to warm to ambient temperature before being transferred to the furnace for linear programming to higher temperatures.In t.p.d. experiments the sample (ca. 1 g) of the catalyst in the silica reactor was exposed to flowing N,, 6% H, in N,, H, or 6% 0, in N, during pretreatment by one of the methods A, B or C as follows. A: (a) N, at 423 K for 30 min, (b) 6 kPa H, at 423 K for 30 min, (c) heating to 673 K at 10 K min-' in N, and holding at 673 K for 1 h, ( d ) cooling to 293-303 K in N,, (e) adsorbing H, at 6 kPa H, at room temperature for 30 min, (f) N, flushing with N, at room2608 TEMPERATURE-PROGRAMMED REACTIONS temperature for 15 min and (g) heating at 7 K min-l to 673-873 K in flowing N, (1 5 cm3 min-l) for t.p.d. analysis. B: (a) as A (a), (b) heating to 673 K in 6 kPa 0, and holding at 673 K for 30 min, (c) flushing with N, at 673 K for 15 min, (d) cooling in N, to 423 K, (e) re-reducing at 6 kPa H, and 423 K for 30 min and (f) as in A ( c t ( g ) .C: (a) heating in N, to 423 K for 30 min, (b) heating in 101 kPa H, to 673 K and holding at this temperature for 30 min, (c) flushing with N, at 673 K for 1 h, (d) cooling in N, to 293-303 K and (e) adsorbing H, at 101 kPa H, for 30 min. During desorption of preadsorbed H,, dry and oxygen-free N, flowed (1 5 cm3 min-l, 101 kPa) through the sample in the silica reactor within the programmable furnace and thence to the katharometer. Thus the rate of release of hydrogen to the inert gas stream as a function of the temperature (293-873 K) of the solid sample was measured. Peaks for 0.81 cm3 H, were used for calibration purposes. RESULTS T.P.B.R.RESULTS FOR UNSUPPORTED HEXACHLOROPLATINIC ACID Fig. 1 (a) shows the t.p.b.r. profile for a sample (0.06 g) of unsupported hexachloro- platinic acid in flowing 6% H, in N, as its temperature was raised at a linear rate from 293 to 673 K. Overlapping maxima in the rates of hydrogen consumption (one at 370 K denoted a and the other at 385 K denoted a) are shown suggesting that the unsupported acid may undergo at least a two-step reduction, which after neglecting hydrogen available within the acid could arbitrarily be ascribed to: PtCI,+PtCI,-+Pt. Significant a-/3 overlap prevents the exact value of w being ascertained from t.p.b.r. profiles. However, the total amount of hydrogen consumed in a +/3 (1.94 mol H, per mol Pt) corresponds to 97% of that predicted for complete reduction. Therefore, bearing in mind the uncertainties in the technique for this sample (estimated to be + 10 % in the light of its hygroscopicity), reduction of the unsupported acid is probably complete in < 5 min at this temperature.Interestingly, the rate d#/dt is relatively constant with time at intermediate q5 and at low temperature. T.P.B.R. RESULTS FOR SUPPORTED HEXACHLOROPLATINIC ACID T.p.b.r. profiles and results obtained at 298-693 K for the reduction of hexachloro- platinic acid supported upon silica (sample F, which had an average Pt particle size of 4.27 nm) and yielding 6% Pt after reduction are shown as runs I and I1 in table 1 and fig. 2. These show three peaks of maximum rates of hydrogen consumption (a, and y ) at < 373, 373-523 and > 523 K and illustrate good reproducibility.The average quantities of hydrogen consumed in each peak in these runs are shown (see table 1) to be 5,53 and 7% of the consumption predicted for complete reduction. Since these total only 65% , either reduction is incomplete at 693 K or it is complete but has commenced at low temperatures before the detector had been equilibrated (i.e. the a peak has been underestimated). T.p.b.r. run I11 in fig. 2 was therefore initiated at 223 K to throw light upon this point. Analysis (see table 1) shows that the a peak now accounts for as much as 39% of the quantity of H, expected to be consumed in complete reduction. When this is added to the 59 % theoretical uptake corresponding to the /3 peak, the total indicates that complete reduction is essentially possible within the processes denoted a and /3 and also shows the importance of starting t.p.b.r.at sufficiently low temperatures. However, the y peak also contributes a further 1 1 % of the theoretical uptake and since this is probably outside experimental error it may correspond to spillover reaction of adsorbate with silica as the support at ca. 673 K. Comparing the t.p. b.r. results for the unsupported hexachloroplatinic acid and the silica-supported acid, it could be argued that the support eases the a-phase reduction (lowering its T,) but hinders the P-phase step (raising its Tm). Thus PtCl, (the productN. D. HOYLE, P. w. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2609 1 I 1 1 273 373 473 573 T/K 1.0 8 c' .- U 0 B 0.5 C 0 u rcl .- U 2 0 tls Fig.1. Results of (a) temperature-programmed bulk reduction of unsupported hexachloroplatinic acid and (b) thermogravimetric analysis given as the fractional reduction (b as a function of time at different isothermal temperatures (shown on curves, K) for the same reduction process. of step a and reactant in step p) may interact with the silica support more strongly than the chlorometallic acid itself. T.p.b.r. results for silica-supported sample G (of higher ultimate Pt dispersion and lower average Pt particle size than sample F) are shown as runs IV-VII (which have decreasing starting temperatures) in table 1 and in part in fig. 2. In run IV at 303-673 K the low-temperature a peak is only observed as a shoulder on theppeak (both contributing only 67% of the hydrogen consumption expected for complete reduction).The high-temperature y peak then contributes a further 15%. Experiment V, undertaken starting at the lower temperature of 273 K, shows an a peak (1 5 % theoretical consumption) separate from the p ( 5 1 % theoretical consumption) peak. Experiment VI started at an even lower temperature (223 K) and the a peak now contributes as much as 46% of the total theoretical consumption, while the area of the p peak (corresponding to 50% theoretical hydrogen consumption) is similar to that in run V. Thus in this experiment the quantity of hydrogen consumed under peaks a and p corresponds essentially to complete reduction (96% ) within experimental error at 523 K (the maximum temperature).Run VII repeated run VI but continued to a higher temperature (723 K) and both were found to be in reasonable agreement. Unfortunately, incomplete separation of a and /3 peaks obscures any deduction as to whether these (when measured from sufficiently low starting temperatures) correspond to two steps with PtCl, as the intermediate, where w is ca. 2. The a and p t.p.b.r. peaks are at temperatures considerably higher than for the unsupported acid but lower than for sample F; this might suggest that the ultimate Pt dispersion is important (with T, decreasing as the particle size decreases). Two experiments were performed on the silica-supported sample I (which had an even higher ultimate Pt dispersion and lower eventual average Pt particle size than either F or G).Runs VIII and IX were undertaken at 303-773 and 223-773 K, respect- ively. The former did not separate a and p peaks satisfactorily and it is noteworthyTable 1. T.p.r. results for supported hexachloroplatinic acid ~~~ ~~ amount of H, consumed/cm3 gpt (% reduction) run weight/g reduction dpt/nma range Tm, Tm.8 Tm. y a /I Y T/K sample sample after temp. I I1 I11 IV V VI VII VIII IX X XI XI1 0.08 1 0.067 0.057 0.095 0.117 0.127 0.1 18 0.148 0.126 298-693 223-693 303473 6% Pt/SiO, (F) 3.59 273-523 223-523 223-723 303-773 223-773 3% Pt/SiO,(L) 14.92 223-773 3% Pt/Al,O, (M) 3.38 255-723 { 3% Pt/SiO, (G) 1% Pt/SiO, (I) 1.50 { 3% Pt/TiO, (N) 9.25 223473 344 355 317 283 298 299 30 1 310 - - 283-303 261-291 463 460 46 1 434 43 5 434-6 434 440 41 8-440 437-9 45 1 3 8 3 4 653-673 - 673 673 653 - 673 699 693-723 631-701 593 > 553 11.39 (5%) 12.60 (5.5%) 90.34 (39%) 35.55 (15%) 105.79 (46%) 100.60 (44%) 115.80 (50%) 91.89 (40%) 6.70 (3%) 208.00 (9 1 % ) - - 126.82 (55%) 116.28 (51 %) 136.70 (59%) 154.20 (67%) 117.56 (51%) 114.03 (50%) 108.05 (47%) 135.77 (58%) 109.4 (48%) 112.19 (49%) 223.60 (97%) 24.99 (13%) 19.42 13.77 (6%) 8.4%) 25.44(11%) 34.18 (15%) 40.18 (18%) 151.50 (64%) 174.00 (76%) 66.00 (29%) 18.36 (8%) 43.47 (19%) a Taken from ref.(23).273 473 673 I I rtpr / XI XI1 I I I I 273 323 373 473 T/K Fig. 2. Results of temperature-programmed bulk reduction of oxide-supported hexachloro- platinic acid. Runs I, I1 and I11 relate to silica-supported sample F. Runs VII and IX relate to silica-supported samples G and I. Runs XI and XI1 relate to alumina-and titania-supported samples M and N.The conditions of pretreatment of all samples in all runs are described in the text and in table 1. The values of d (shown on curves, nm) indicate the average Pt particle sizes estimated23 after reduction.2612 TEMPERATURE-PROGRAMMED REACTIONS that the high-temperature y peak is much greater (accounting for 64% of the theoreti- cal hydrogen consumption corresponding to complete reduction) than expected. Run IX, starting at a lower reduction temperature, did separate the a peak and found that it corresponds to 50% of the total theoretical consumption of hydrogen, while the peak contributed a further 48%. This again suggests the two-step reduction process where w in intermediate PtCl, is 2 between the a and p peaks in t.p.b.r.In run IX reduction therefore appears complete after the a peak (at 523 K) and it is therefore surprising that the y peak contributes a further 74% of the theoretical consumption. The total hydrogen consumption in run IX on sample I therefore sig- nificantly exceeds that expected for complete reduction. It is clear that the contribu- tion of the high-temperature y peak increases as the Pt dispersion increases and this may relate to spillover onto the support and could contribute to the excess hydrogen consumption. Nevertheless, the precise cause of this excess H, consumption remains uncertain. However, at this highest Pt dispersion the t.p.b.r. peaks have continued to appear at lower temperatures. One run (X) was also carried out on sample L involving the non-porous silica support at 223-773 K.This showed well separated a, p and y peaks, which were at relatively high temperature for this dispersion, reflecting the role of the support characteristics. T.p.b.r. runs XI and XI1 at 255-723 K and 223-673 K with alumina-supported and titania-supported samples denoted M and N (see table 1 and fig. 2) showed a total hydrogen consumption equivalent to approximately complete reduction, but the profiles are not an extension of those seen for the silica-supported samples and at high temperature may include support-adsorbate interactionsz4 and spillover. T.P.D. RESULTS FOR SUPPORTED Pt Sample F (with the lowest Pt dispersion on silica) after pretreatment A produced the t.p.d. profile shown as run I in fig.3 and table 2. The p and y maxima are clearly seen, the latter being significantly smaller. When this sample was pretreated by method B no y peak above 673 K was seen, but the areas of the a peaks after both pretreatments was approximately the same. The y peak may again be associated with hydrogen spillover onto the supportz4 and here it appears specific to some pretreatment methods and conditions and possibly the original reduction conditions (up to 673 K at 101 kPa H, for 1 h for this sample). Silica-supported sample G (lower Pt particle size than F) was subjected to each of the three pretreatments (see table 2). After pretreatment A (also shown as run I1 in fig. 3) the t.p.d. B peak was at a lower temperature than that for the lower dispersion sample F, but no y peak was seen.This was also true after pretreatment C , but after pretreatment B the high-temperature y peak was observed. Thus the t.p.d. method used affects the results obtained. Silica-supported sample K exhibited the highest Pt dispersion and its repeated hydrogen t.p.d. profiles are in good agreement, as shown in table 2 and fig. 3 (run 111). After each of the pretreatments the size of the peak was approximately constant but the y peak is seen clearly only after pretreatment B (as noted for other samples). However, its peaks are at lower temperatures than for any of the other samples, which again reflects its greater Pt dispersion. Extending the time of treatment with hydrogen using pretreatment A for this sample had no discernable effect on its t.p.d.profile, which indicates the suitability of the conditions chosen here for pretreatment. T.p.d. results for Pt supported upon the non-porous silica (sample L) gave a p peak at a lower temperature and a y peak at a higher temperature than similar dispersions on the porous silica support.N. D. HOYLE, P. H. NEWBATT, K . ROLLINS, P. A. SERMON AND A. T. WURIE 2613 rtpd 4.27 nm \ 1.09 nm 273 47 3 673 87 3 TI K Fig. 3. Temperature-programmed desorption of hydrogen from oxide-supported Pt. Run I relates to silica-supported Pt sample F after pretreatment A. Run I1 relates to silica-supported Pt sample G after pretreatment A. Run I11 relates to silica-supported Pt sample K after pretreatments A (-), B (---) and C (-.-.-). Run IV relates to alumina-supported Pt sample M after pretreatment A.Run V relates to titania-supported Pt sample N after pretreatment A. The values of d (shown on curves, nm) indicate the average sizes of Pt particles estimated23 for each sample after reduction.Table 2. T.p.d. results for supported Pt at 293-873 K after reduction at 673 K and H, adsorption during pretreatments A, B or C (see text) amount of H, adsorbed in ref. (25) sample method of in a+P+y amount of H, desorbed/cm3 gpt and now desorbed sample d,,/nm" weight/g pretreatment T,,JK T,, ?/K a P Y peaks (%> 6% Pt/SiO, (F) 4.27 0.995 A 3% Pt/SiO, (G) 3.59 1.110 A 6% Pt/SiO, (K) 1.09 0.305 A 3% Pt/SiO, (L) 14.92 1.04 A 3% Pt/A1,O, (MI 3.38 1.16 A B B C 0.507 B 0.300 C B B 433453 423-453 423453 443-473 423-448 423-443 423-443 393423 333-623 variable 773 - - 763 - 773 - 773 - 823 423,548 - 393,573 5.50 2.30 5.65 - - 12.01 - 9.67 ~ - 11.19 - - - - 40.12 3.57 42.49 7.86 48.97- - 3.30 4.59 - 6.23 - 17.49- 18.87- - ~~ 56.0 40.5 72.0 58.0 67.1 72.7 83.8 81.5 _ _ _ ~ ~ a Taken from ref.(23).N. D. HOYLE, P. H. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2615 As with t.p.b.r., the results for alumina- and titania-supported Pt samples for hydrogen t.p.d. do not follow simply from those for silica-supported samples. Multiple broad overlapping maxima are seen at temperatures between 300 and 800 K and in addition to desorption it is likely that support-adsorbate interaction^^^ are much more significant than in the case of silica supports. DISCUSSION The t.p.b.r. profiles for hexachloroplatinic acid and hydrogen t.p.d.profiles for platinum indicate that reduction and desorption are two- or three-step processes with maximum rates appearing at temperatures affected by the ultimate Pt dispersion, the nature and porosity of the support and the pretreatment and starting conditions used. In t.p.b.r. of samples with porous silica supports, peaks move to lower temperatures (and show better separation) as the ultimate Pt dispersion increases. This may also have been seen here and previouslyg with hydrogen t.p.d. on the silica-supported Pt surfaces produced by t.p.b.r. The problem is whether we can rationalise this affect of dispersion with the apparent initial ease of reduction and hydrogen desorption. Comparison with earlier studies may be complicated by different preparative methods, supports5 and techniques; indeed even when using exactly the same catalyst quantiative comparisons are often difficult because of significant differences in the experimental techniques used.Thus often absolute percentages of reduction or desorption are not quoted. Note that here the extent of hydrogen consumed in t.p.b.r. exceeds that expected for complete reduction and that in t.p.d. the percentage of adsorbed hydrogen desorbed varies for the silica supports (from 41 to 84%) as the Pt dispersion increases; spillover may have been included. Results for alumina- and titania- supported samples from t.p.b.r. and t.p.d. are rather more complicated than with silica supports. Perhaps it is also more relevant to ask whether it is justifiable to associate these peaks of maximum rate with reduction or desorption of distinct and distinguishable states (and to derive activation energies associated with their specific reduction or desorption by measuring T, as a function of heating rate 2) and to estimate the concentrations of these species from the integrated t.p.b.r.and t.p.d. peaks or whether these may be rather more complicated. Consider in the following theoretical treatments that at some intermediate time and temperature during a t.p.d., t.p.s.r. or t.p.b.r. experiment the solid contains two solid or surface species S, and S2 which differ in the free energy involved in their desorption, reaction and reduction. The likelihood that these will undergo independent desorption, reaction or reduction rather than undergo dynamic equilibration via diffusional exchange will depend on the relative rates of desorption, reaction and reduction and S,-S, diffusion.These will in turn be determined by the respective activation energies for these processes. Diffusional activation energies may well define the significance and separation of t.p.d., t.p.s.r. and t.p.b.r. peaks. THEORETICAL ANALYSIS TREATMENT OF T.P.D. Experimental measurementl’y 25 of rates of adsorption Y, and desorption Yd are given by Ya = k,pfle) exp (-Ea/RT) and Yd = kdf’(8) exp (-&/kT) as a function of temperature and have been used to measure the relevant activation energies E, and Ed. For hydrogen adsorbed dissociatively upon Pt,fl8) andf’(8) are2616 TEMPERATURE-PROGRAMMED REACTIONS / \ \ adsorption potential surface Fig.of 1 solid I 4. Schematic diagram of the activation energies involved in the desorption and reactions states S , and S,. The energy E varies periodically across the uniform adsorbent surface. (1 -O), and 02, respectively, where 0 is the fractional coverage of hydrogen. If we assume that S, and S, are energetically distinct chemisorbed states in fig. 4 then they will have distinct values of Ea and E d . Assume that a weakly held molecule of hydrogen designated =# is the critical intermediate between the chemisorbed states (S, and S,) and the gaseous state, then absolute rate theories26 lead to rates of isothermal desorption from these sites of rd(S1) = (kT/h) [Q( 4+ )/Q(sAl [SiI exP [ - Ed(Sl)/k TI r d ( S 2 ) = (kT/h) [Q( 4+ )/Q(s&l [S2I exP [ - Ed(S2)/kTl and where Q is the relevant partition function and we neglect differences between Q( # ) and Q +, the partial pressure of gaseous hydrogen and the power to which the concentrations [S,] and [S,] should be raised to allow for the dissociative nature adsorption.Isothermal rates of hydrogen desorption from Pt have been evaluated to be 0.001-N. D. HOYLE, P. H. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2617 0.100 cm2 atom-l s-l,l1? 25 and may be transposed into t.p.d. rates per unit area per degree (dn/dT), which may be summed for individualj sites to give the total rate of desorption in t.p.d. dn/dT(total) = dn/dT(S,)+dn/dT(S,)+. . . +dn/dT(Sj) This will pass through maxima as any dn/dT for each individual adsorbate species passes through a maximum independent of the nature of +k and its partition function. However, models of adsorption-desorption26 have considered diffusion, and the effects of S,-S, diffusion on the above desorption rates must now be considered. The rate of surface diffusion on Pt can be described by a surface flux of atoms or molecules per cm2 per second, N’,,’ as do dx N’ = -bD-Cma, where x is the distance normal to the diffusion interface between the S, and S, domains, C,,, is the excess concentration of hydrogen at the surface in the higher concentration domain and D is the surface diffusion coefficient.D and Ediffn at 333-373 K and 0-0.5 kPa for hydrogen on Pt are cm2 s-l and 23.9 kJ mol-1 27 and are almost independent of pressure, but D does vary with temperature according to D = Do exp (- Ediffn/kT) = Do exp (- cQ/RT) where Do is va2 and v is the frequency of jumping between available adsorption sites separated by distance a, c is a proportionality constant and Q is the differential heat of adsorption (42-75 kJ mol-l at 8 = 0.42-0.32).Therefore surface diffusion is lower as Q and Ediffn increase. Thus N’ is given by d8 dx N’ = va2bCma, - exp ( - Ediffn/kT). However, in t,p.d. the more relevant flux is N”, which has units of atoms or molecules per cm2 per K and is given by d8 dx N” = va2bCmax - exp ( - Ediffn/kT) (1 / Z ) . Diffusion processes are very fast and have relatively low activation energies.28 Thus Ediffn between two S, sites is given by (Emax- Emin)s, where E varies periodically between these across the surface,11725 as shown in fig.4. The activation energy for diffusion between two S, sites is given by (Emax-E,in)s2. However, the more important diffusion in randomisation is between S, and S, domains where S, states are more tightly bound and closer to the surface than S, states (see fig. 4) and the activation energy for this process is given by (Emax)s,. Therefore the activation energies for each of these diffusion processes (inter- and intra-domain diffusion) are always smaller than the activation energies of desorption (and adsorption). For example, Edinfn/Ea for hydrogen on Pt is as small as 0.08,29 and so Ediffn is thus an order of magnitude less than that for desorption. Thus hydrogen on Pt at ambient temperature is very mobile11’25 and S, and S, must be in dynamic equilibrium during the t.p.d.experiment. Inevitably then t.p.d. experiments must involve some randomisation of adsorbed states prior to desorption. However, surface diffusion becomes more likely as Emax - Emin decreases or T2618 TEMPERATURE-PROGRAMMED REACTIONS increases during t.p.d. experiments and kT only increases from 0.83 to 4.16 to 8.31 kJ mol-1 at 100, 500 and 1000 K. Bearing in mind the values of Edesn (37.66 kJ mol-l)ll and Ediffn (18.83 kJ rn01-l)~~ found for hydrogen on Pt it seems unlikely that adsorbed species of average energy (in vibrational degrees of freedom) will either surface diffuse or desorb, this being restricted to those of significantly above average energy (and probably from S, rather than the more tightly bound S, states).Thus is seems possible that t.p.d. experiments are only seeing a fraction of the weakly bound and high-energy adsorbed species. S, and S, can also be differentiated in terms of the average lifetime z that an adsorbed species will remain on the adsorbent surface before desorption into the inert-gas stream. This will depend upon the enthalpy of adsorption on these sites Q(S,) and Q(S,) and the temperature: where zo is the proportionality constant, taken to be 5 x s , ~ O and for hydrogen on Pt enthalpies of -23 to - 113 kJ mol-1 have been observed. Fig. 5 shows how the average residence time varies with temperature for these values of Q and for S, and S,, where the enthalpies of adsorption are assumed to be -60 and - 150 kJ mol-l. When z < 100 s or z > 1000 s (particularly at flushing temperatures) then t.p.d.is unlikely to observe the desorption. Fig. 5 suggests that in reality t.p.d. is likely to miss a significant fraction of weakly bound hydrogen states (and also some very tightly bound ones) on Pt. Fig. 5 also suggests that t.p.d. peaks should be extremely sharp (k 10 K); the fact that they are not must be a reflection of the variation of Q with coverage (as t.p.d. proceeds) and adsorbate randomisation. H: Pt ratios are known3, to change with d but in this laboratory differential scanning calorimetry has indicated that Q for hydrogen chemisorption on silica-supported platinum at 5-6 kPa H, and 293 K decreased from 91.99 to 60.60 kJ mol-1 H, as d for Pt decreased from 3.59 to 1.09 nm. This is in agreement with the trends in the values of Tm (and hence rd) seen in t.p.d.profiles in fig. 3 and decreases in desorption activation energy with decreasing d for Pt, as seen previo~sly.~ However, the above theoretical analysis and table 2 suggest that only a fraction of the H, initially adsorbed on Pt is subsequently detected by t.p.d. and there is self-consistency between theory and experiment despite the fact that the quantity of adsorbed hydrogen ignores the extent of additional hydrogen adsorbed and spilt over onto the support between the zero-pressure intercept (used in chemisorption measurement~,~) and the conditions of adsorption prior to t.p.d. The low fraction of adsorbed H, observed to be desorbed could be the result of either too slow or too fast a desorption process.TREATMENT OF T.P.S.R. T.p.s.r. titrations attempt to overcome these problems by reacting the adsorbate on different sites at different rates (and to different products) rather than by differential thermal desorption. T.p.s.r. tends to operate at lower temperatures than t.p.d. However, the success of the relatively new t.p.s.r. in extracting separate surface species has yet to be truly ascertained. Nevertheless, in principle the lower the temperature of operation the greater the likelihood of success in this respect. T.p.s.r. may well proceed via the same weakly held # state as t.p.d. TREATMENT OF T.P.B.R. This type of analysis can now be extended to t.p.b.r. (reduction), in which case S, and S , are the reducible sites located in the bulk solid reacting with the gaseous hydrogen reductant.For t.p.b.r. of hexachlorplatinic acid in the absence of hydrolysis,N. D. HOYLE, P. H. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2619 1 O'O m o \ 10-10 2 I I I 1 0 400 600 80 0 T/K Fig. 5. Average residence times of adsorbate before desorption in t.p.d. for various enthalpies of adsorption (shown on curves, kJ mol-l ads) relevent to hydrogen adsorption on Pt. S, and S, may differ in terms of the local C1:Pt ratio and the extent of interaction with any support surface. Whether S, and S, can be treated separately in their reduction by hydrogen or not will depend upon the activation energies of reduction and those of C1- diffusion between sites or domains of S, and S,. Hexachloroplatinic acid is best represented as (H,O);(PtCl,)..yH,O, with 2 < y < 6, containing octa- hedral PtC1;- anions.32 This decomposes in the unsupported state above 431 K to PtCl,, then PtCl,, then PtCl, and finally Pt,32 but it is reduced at a lower temperature of 373 K (in agreement with fig.1) with the liberation of HCl. Thus many of the t.p.r. peaks for silica- and alumina-supported hexachloroplatinic acid occur at temperatures comparable with those at which the acid would decompose. Decomposition suggests rapid diffusion of C1- at these temperatures. If t.p.b.r. is to provide the true activation energies of the reduction steps of homogeneous samples then the sample must have a homogeneous composition between t.p.b.r. peaks. Whether this can be achieved will depend upon T, Z , the relevant diffusion coefficient D and the average size of the reducible particles d.Let us assume that the reducible particles reduce from one nucleation face at a rate determined by diffusion at the chloride/metal interface rather than gas-phase diff~sion.,~ Then the parameter P can define the square of the percentage of the side of the particle across which the reduction interface proceeds per K in t.p.b.r. in terms of Do x 6 x lo5 x exp ( - E / R T ) (Zd2) where Do is the diffusion coefficient at absolute zero in cm2 s-l and E is the activation energy for diffusion-reduction. An equivalent relationship can be derived for cubic particles undergoing reduction simultaneously from all faces. The migration of C1-2620 TEMPERATURE-PROGRAMMED REACTIONS I l 4 6 8 tlmin Fig.6. Phase-front movement in the reduction of hexachloroplatinic acid to zero-valent Pt in plate 1 plotted as a function of time for different crystals. ions in several metal chlorides has been and Do and E have been found to be in the range 10-5-10-13 cm2 s-l and 97-193 kJ mol-l. Analysis of t.g.a. traces obtained here as In (dbldt) against l / T (where 4 is the fractional extent of reduction) indicate that the unsupported hexachloroplatinic acid is reduced with an activation energy which increases as 4 increases (20.8, 29.9, 59.9, 66.5 and 55.4 kJ mol-1 for 4 = 0.1,0.2,0.5,0.7 and 0.85 at 319-362 K and 101 kPa H2). It has also been possible to calculate the overall diffusion coefficient in the reduction of chloroplatinic acid from the movement of the acid/metal phase front under isothermal conditions as a function of time.Measurements of the movement of the metal/acid phase front were carried out in a manner similar to that used previously35 in an optical microscope using a sealed cell (see plate 1, which reveals transparent chloride particles converting to the black Pt phase in a manner not dependent on the direction of H, flow, and which shows the acid/metal interface) at 101 kPa H, and 292 K. Fig. 6 shows that the square of the phase-front movement per unit time (which defines the diffusion coefficient in the direction of the movement of the phase front) and that this is similar [i.e. D = (6, 6.2, 3.3, 5.0, 1.8, 6.6. 2.1 and 8.3) x lo-* cm2 s-,] for different acid crystals. There was no apparent effect of particle size on the time of nucleation or subsequent reduction.At this point it must be recognised that the exact C1: Pt ratio was found by microprobe analysis to vary with crystal and it is not possible to define2: P c r m 2 n ? 3 3 h 3 8' Plate 1. Progressive reduction of crystals of hexachloroplatinic acid or lower chlorides at 292 K in 101 kPa flowing hydrogen after (a) 4.45, (b) 5.00, ( c ) 5.10, ( d ) 5.20, (e) 5.30 and cf) 5.45 min, where transparent chloride phase is converted into the black Pt phase (confirmed by X.P.S.) from one nucleating edge (in a direction independent of the direction of H, flow) and progresses across the crystals. "tr s n h) o\ h) 0 U WN. D. HOYLE, P. H. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2621 20 10 4 c M 0 -10 -20 ( I i / f / I I I I 200 400 600 800 T/K Fig.7. Plot of In P calculated at d = 100 nm and 2 = 5 K rnin-l at constant Do (+, 0, A and cm2 s-l; 0, 0, A and 0 : 10-l2 cm2 s-l) where the activation energy decreases from 180 (0, H) to 120 (A, A) to 60 (0, 0 ) to 20 (0, +) kJ mol-l as a function of reduction temperature in t.p.b.r. : the composition of each crystal precisely for this apparently very complex chloro- metallic acid and its lower chlorides. In addition it is unlikely that reduction is thermally neutral. For these reasons only the orders of magnitude of activation energy derived experimentally have been used here to understand the parameters important in t.p.b.r. This requires the parameter P to be > ca. 1 and In P > 0. Fig. 7-10 show that P and In P increases as E, T and Do increase and as 2 and d decrease.Therefore if from the above measurements for the reduction of hexachloro- platinic acid Do is taken to be lo-* cm2 s-l, E as 60 kJ mol-1 and 2 has a commonly accepted value for t.p.b.r. ( 5 K min-l, although values as high as 30 K min-l have been discussed by Parkyns et aZ.;15 these values are much lower than those in flash de~orption~~) then In P only becomes significant in fig. 10 at a temperature which decreases (from > 1000 to < 300 K) as the particle size decreases (from 100 pm to 1 nm) (as observed experimentally in fig. 2). Therefore if the size of reducible particles is too large, the heating rate is too high or the diffusion coefficient and activation energy2622 TEMPERATURE-PROGRAMMED REACTIONS 2c 10 4 G - 0 -10 -2 c 20 10 0 % 9 - -1 0 -20 I I I I 200 400 600 800 T/K I I f 1 I 200 400 600 80 0 T/K Fig.8 and 9. For legends see facing page.N. D. HOYLE, P. H. NEWBATT, K. ROLLINS, P. A. SERMON AND A. T. WURIE 2623 Fig. 10. Plot of In P calculated at Do = lo-* cm2 s-l, 2 = 5 K min-' and E = 60 kJ mol-' as a function of reduction temperature in t.p.b.r. as the particle size of the reducible particle d increases (0, I nm; 0, 10 nm; 0, 100 nm; a, 1 pm; A, 10 pm; A, 100 pm). between S , and S , are too small, then t.p.b.r. will proceed upon heterogeneous shrinking-sphere (or cube) particles and the reverse condition is required for reduction of homogeneous particles with the resulting t.p.b.r. peaks being meaningful. The optimum heating rate 2 for successful t.p.b.r.can only be deduced from a knowledge of Do, E and d for the particular system to be studied, and all too frequently this is not considered. It is suggested that this should be undertaken in future. Certainly the broad acceptance of a rate of heating of 5 K min-l as being experimentally convenient is not reasonable. If these conditions are not fulfilled then t.p.b.r. may not satisfactorily resolve reduction steps and the observed temperature dependence of T, may not reveal activation energies of reduction but rather the increasing time required for diffusional homogenizing of the solid sample. Fig. 8. Plot of In P calculated at d = 100 nm, 2 = 5 K min-' and E = 20 kJ mol-l as a function of reduction temperature in t.p.b.r.at different values of Do (0, 0, 0, lo-*; B, A, cm2 s-l). Fig. 9. Plot of In P calculated at d = 100 nm, Do = cm2 s-l and E = 20 kJ mo1-l as a function of reduction temperature in t.p.b.r. at different rates of heating 2 (0, 1 ; x , 5; min; +, 10; +, 20 K min-l). Similar trends with 2 are seen for E = 60 kJ per mol (0, 1; 0, 20 K min-l), 120 kJ mol-l (A, 1 ; A, 20 K min-I) and 180 kJ mol-l (0, 1 ; ., 20 K min-l).2624 TEMPERATURE-PROGRAMMED REACTIONS -113 k i 5 0 80 0 200 400 600 800 TIK Fig. 11. (a) T.p.d. profiles predicted by the residence times in fig. 5 when the enthalpy of adsorption is - 60, - 1 13 or - 150 kJ mol-l. These are consistent with those in fig. 3 bearing in mind the decrease in the adsorption enthalpy with decreasing average Pt particle size d found here experimentally.(b) T.p.b.r. profiles predicted from fig. 7 when the activation energy of diffusion in the solid reducible phase is 20, 60, 120 or 180 kJ mol-l (and d = 100 nm, 2 = 5 K min-l and Do = lo-* cm2 -l). These profiles would relate to those found for the reduction of different oxides or chlorides etc. (c) T.p.b.r. profiles predicted from fig. 10 when the average particle size d is 100, 10 or 1 nm (and 2 = 5 K min-', Do = lo-* cm2 s-' and E = 60 kJ mol-l). These are consistent with those in fig. 2 (i.e. decreasing d from 10 to 1 nm causes T, to decrease by 50 K). CONCLUSIONS The above conclusions concerning the relative magnitudes of activation energies of diffusion and desorption are also true for a wide range of adsorbates and ads~rbents,~~ i.e.Ediffn/Ea are often 0.1-0.7. Therefore equal care should be taken in the interpretation of other t.p.d. results. Questions as to how well t.p.d. peaks reflect separate binding states on a surface, when distortion and artefacts may be caused by diffusion and interconversion of adsorbate states, have been raised previo~sly.~~ Diffusion in pores can also modify t.p.d. but the influence of precursor adsorption states + on the positions and breadths of t.p.d. peaks was neglected for a significant period.39 The heating rate in t.p.b.r. must be selected in the light of the values of the coefficient and activation energy of bulk diffusion, Do and E, and the average size of reducible particles, d. If this is undertaken then t.p.b.r.N. D. HOYLE, P. H. NEWBATT, K.ROLLINS, P. A. SERMON AND A. T. WURIE 2625 peaks can be associated with differentiable processes. Previously, Monti and Baiker2' defined a parameter which equalled F([S]/[G]), where I; was the flow rate of the gaseous reactant G, which they used to define the optimum operating conditions, i.e. at 0.1 c Z/K s-l < 0.3 this parameter needed to be 55-140 s mol-l, but this takes no account of the importance of the dispersion of the solid. The observed effects of dispersion on temperatures at which maximum rates of reduction occur in t.p.b.r. can be explained for the first time by the analysis used here. Bearing these points in mind, temperature-programmed techniques can be extremely useful quantitative analytical tools for a variety of gas-solid reactions.Fig. 1 1 shows t.p.b.r. and t.p.d. profiles predicted from the theoretical analysis; they are consistent with those found experi- mentally for hexachloroplatinic acid and Pt supported on porous silica. Using the analogy with gas chromatography, the extent of resolution of any two t.p.b.r. or t.p.d. peaks (twice the separation between their maxima divided by the sum of their breadths at the baseline) seen here and previously was never > 1.0. This indicates the improvements in the resolution of temperature-programmed methods that may be achievable in the future. We thank the S.E.R.C., Johnson Matthey, I.C.I. (Mond Division) and B.O.C. for support of N. H., K. R. and P. N., and the Sierra Leone Government for support of A.T.W. 1 2 3 4 5 6 7 8 9 10 I1 12 13 14 15 16 17 18 19 20 2 1 Z! 2 23 :! 4 J.L. Falconer and J. A. Schwarz, Catal. Rev. Sci. Eng., 1983, 25, 141. R. J. Cvetanovic and Y. Amenomiya, Adv. Catal., 1967,17,103; Catal. Rev., 1972,6,21; S. Tsuchiya, Y. Amenomiya and R. J. Cvetanovic, J. Catal., 1970, 19, 245. R. Komers, Y. Amenomiya and R. J. Cvetanovic, J. Catal., 1969, 15, 293; S. Tsuchiya and N. Yos- hioka, J. Catal., 1984, 87, 144. P. G. Menon and G. F. Froment, J. Catal., 1979,59, 138. J. R. Anderson, K. Foger and R. J. Breakespeare, J. Catal., 1979,57, 458. A. Brenner and D. A. Hucul, J. Catal., 1979, 56, 134. E. E. Ibok and D. F. Ollis, J. Catal., 1980, 66, 391; D. M. Jones and G. L. Griffin, J. Catal., 1983, 80,40. B. Halpern and J. E. Germain, J. Catal., 1975, 37, 44. Y. Takasu, M. Teramoto and Y. Matsuda, J.Chem. SOC., Chem. Commun., 1983, 1329. P. G. Menon, R. P. de Pauw and G. F. Froment, Znd. Eng. Chem., Prod. Res. Dev., 1979, 18, 110. K. Christmann, G. Ertl and T. Pignet, Surf. Sci., 1976, 54, 365; 1976, 60, 365. P. I . Lee and J. A. Schwarz, J. Catal., 1982, 73, 272. K. Fujimoto, M. Kameyama and T. Kunugi, J. Catal., 1980, 61, 7. M. S. W. Vong and P. A. Sermon, J. Chem. SOC., Chem. Commun., 1983, 660. N. Parkyns, I. J. Kitchener, C. Komodromos and D. Bradshaw, unpublished results; R. Prins unpublished results; R. M. Lambert and R. B. Grant, unpublished results; K. Rollins and P. A. Sermon, unpublished results. J. W. Jenkins, B. D. McNicol and S. D. Robertson, Chemtech., 1977, 316; S. D. Robertson, B. D. McNicol, J. H. de Baas, S. C. Kloet and J. W. Jenkins, J.Catal., 1975, 37, 424; H. Lieske, G. Leitz, H. Spindler and J. Votler, J. Catal., 1983, 81, 8. N. W. Hurst, S. J. Gentry, A. Jones and B. D. McNicol, Catal. Rev., 1982, 24, 233. S. J. Gentry and P. T. Walsh, J. Chem. SOC., Faraday Trans. I , 1982, 78, 1515. P. A. Jacobs, J. P. Linart, H. Nijs, J. B. Uytterhoeven and H. K. Beyer, J. Chem. Soc., Faraday Trans. I, 1977, 73, 1745. S. J. Gentry, N. W. Hurst and A. Jones, J. Chem. SOC., Faraday Trans. 1 , 1979,75, 1688. D. A. M. Monti, and A. Baiker, J , Catal., 1983, 83, 323. H. Bosch, J. Kip, J. G. Van Ommen and P. J. Gellings, J. Chem. SOC., Faraday Trans. I , 1984, 80, 2479; J. S. Rieck and A. T. Bell, J. Catal., 1984, 85, 143. A. R. Berzins, M. S. W. Vong, P. A. Sermon and A. T. Wurie. Ad. Sci. Tech.. 1984. 1. 51. P. A. Sermon and G. C. BoGd, Catal. Rev., 1973, 8, 21 1; 'R. T. K. Baker, E. B. Prestidge and R. L. Garten, J. Catal., 1979,56, 390; F. M. Dautzenberg and H. B. M. Wolters, J. Catal., 1978,51, 26; S . J. Tauster, S. C. Fung and R. L. Garten, J. Am. Chem. Sac., 1978,100, 170; R. T. K. Baker, J . Catal., 1980, 63, 523. 25 K. Christmann, Bull. SOC. Chim. Belg., 1979, 88, 519, (1979); Chem. Eng. News, 1983, 24. 86 FAR 12626 TEMPERATURE-PROGRAMMED REACTIONS 26 R. D. Findlay and C. A. Ward, J. Chem. Phys., 1982,76, 5624; 5615; L. A. Jonas, Carbon, 1978,16, 27 K. J. Sladek, E. R. Gilliand and R. F. Baddour, Znd. Eng. Chem. Fundam., 1974, 13, 100. 28 D. C. Torney and H. M. McConnell, Proc. R. SOC. London, Ser. A, 1983, 387, 147. 29 R. Lewis and R. Gomer, Sur- Sci., 1969, 17, 333. 30 J. R. Dacey, Znd. Eng. Chem., 1965, 57, 26. 31 G. C. Bond, 4th Znt. Congr. Catal., (Akademiai Kiado, Budapest, 1971) p. 266; A. P. Karnaukhov, Kinet. Katal., 1971, 12, 1520; 0. M. Poltorak, V. S. Boronin and A. N. Mitrofanova, 4th Znt. Congr. Catal. (Akademiai Kiado, Budapest, 1971) p. 276; P. W. Selwood, 3rd Int. Congr. Catal., 1965, 2, 1976; T. Kubo, H. Arai, H. Tominaga and T. Kunugi, Bull. Chem. SOC. Jpn, 1972,45, 607. 32 V. U. Greher and A. Schmidt, Z. Anorg. Allg. Chem., 1978,444,97; A. E. Schweizer and G. T. Kerr, Znorg. Chem., 1978, 17, 2326; A. T. Hubbard and F. C. Anson, Anal. Chem., 1966, 38, 1887; G. A. Korchinskii and N. P. Krasilenko, Zh. Fiz. Khim., 1976, 50, 1587. 33 A. J. Hick1 and R. W. Heckel, Metal. Trans. A, 1975, 6, 431; S. R. Shatlynski, J. P. Hirth and R. A. Rapp, Acta Metal., 1971, 24, 1071. 34 Y. Sensui, Bull. Chem. Soc. Jpn, 1972, 45, 359; 2677. 35 T. A. Clarke and J. M. Thomas, J. Chem. SOC., 1969, 2227; 2230. 36 E. Habenschaden and J. Kuppers, Surf. Sci., 1984, 138, L 147. 37 G. Ehrlich and D. 0. Hayward, Discuss. Faraday SOC., 1966, 41, 102. 38 E. Tronconi and P. Forzatti, J . Catal., 1985, 93, 197 39 R. Gorte and L. D. Schmidt, Surf. Sci., 1978, 76, 559; A Cassuto and D. A. King, Surf. Sci., 115. 1981, 102, 388. (PAPER 4/745)

ISSN:0300-9599    

DOI:10.1039/F19858102605

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1985, 81, 2627-2634 Chemistry of 1281 in Alkaline Aqueous Solutions of Sodium Iodate Activated by Thermal Neutrons BY RAM B. SHARMA* AND SHUDDHODAN P. MISHRA Department of Chemistry, Banaras Hindu University, Varanasi-221005, India Received 25th June, 1984 The retention of lZ8I in crystalline sodium iodate at room temperature irradiation is found to be 67.8%, which is quite different from the value of 56.0% observed on irradiation at liquid-nitrogen temperature. Radiative-neutron capture in frozen alkaline aqueous solution of the target irradiated at liquid-nitrogen temperature yields a retention value of 35.1 % compared with a value of 23.9% at room temperature for the same solution. Dilution of alkaline aqueous solution with additives (i.e.KC1, NaCl, KOAc and NaOAc) prior to neutron irradiation in varying amounts results in an increase of retention in general; also the yield of radioperiodate increases almost reaching a plateau at a molar ratio of additive to target of ca. 1 : 1. The results are explained in the light of a model which invokes the oxidizing-reducing nature of the intermediates produced during neutron irradiation. The chemical effects following thermal-neutron capture in inorganic systems have been extensively studied for several years, but most of these investigations have concerned the behaviour of hot halogen atoms in crystalline-phase irradiation. The results of such studies show clearly that the ultimate chemical fate of atoms undergoing nuclear transformation in halogen oxyanions is determined by complex interactions of nuclear, thermal, radiation and composition effects.Earlier reports on thermal-annealing experiments in NaIO, show that the retention of recoil 1281 increases on heating the irradiated Furthermore, the occurrence of thermal annealing reactions at room temperature and even to -80 "C has been reported in solid iodate targets.6+ So far retention in iodates and periodates in the solution phase has not received much attention, although the hot-atom chemistry of such systems appears an obvious extension of investigations in this direction. The study of hot-atom reactions in the solution phase also seem to offer certain prospects for gaining a more thorough insite into the mechanism of processes taking place in the radical nest.Results of (n, y ) reactions for iodates and periodates in aqueous solution are few in comparison with those in crystalline solids.8-10 Cleary et aL2 first showed that the radioiodate fraction produced by neutron irradiation of a neutral aqueous solution of iodate varies with the addition of oxidizing and reducing agents. They tried to explain their results on the basis of a ligand-loss hypothesis. Also, other studies on the (n, y ) irradiation of neutral iodate systems present a lower limit of retention that can be achieved in the neutral aqueous phase by the choice of a suitable reducing agent, and their explanation for the observed retention and yields are far from Arnikar et aZ.8* found that the retention from LiIO, and HIO, in aqueous solution increases with concentration. However, they were unable to find any radioperiodate in the targets studied. Recently some results for iodates and periodates t Present address : Department of Chemistry, The Johns Hopkins University, Baltimore, Maryland 21218, U.S.A.2627 86-22628 CHEMISTRY OF lZsI IN NaIO, SOLUTIONS have been published by the present authors.11-13 This work was undertaken as a reinvestigation of the fate of recoil iodine-128 atoms in an NaIO, target, in both the solid and solution phases (particularly in alkaline aqueous media), and to examine the influence of additives, i.e. KCl, NaCl, KOAc and NaOAc, with a view to comparing the effect of various chemical and physical processes on the observed data of recoil-1281 stabilization. EXPERIMENTAL THERMAL-NEUTRON IRRADIATION OF SOLIDS 300 mg of crystalline AnalaR NaIO, was irradiated by thermal neutrons from a 300 mCi (Ra-Be) neutron source having an integrated flux of 3.2 x lo6 neutron cm-2 s-l.The neutron source was surrounded by a cylindrical block of paraffin for thermalizing fast neutrons. The concomitant y-dose associated with the neutron source was ca. 172 rad h-' as estimated by Frick dosimetry. THERMAL-NEUTRON IRRADIATION OF SOLUTIONS Solutions of NaIO, (0.076 mol dm-,) were made by dissolving 0.300 g of target material in 20.0 cm3 of 0.1 mol dmW3 NaOH solution at a pH of ca. 12.0. For the irradiation of these solutions a hollow cylindrical paraffin block of ca. 3 cm thickness was made inside a beaker. The neutron source was placed in the inner hollow portion within the beaker.The beaker and neutron source were then surrounded by a second beaker of capacity 400cm3. Under this experimental set-up thermal-neutral irradiation of solutions of NaIO, along with the additives was performed by keeping the solution in the second beaker. THERMAL-NEUTRON IRRADIATION OF FROZEN ALKALINE AQUEOUS SOLUTION The alkaline aqueous solutions of the target were frozen drop by drop in a soda-glass test tube kept at liquid-nitrogen temperature and then dipped in liquid nitrogen contained in a Dewar together with the neutron source. The three stable radioactive products, i.e. iodide, iodate and periodate, were separated by fractional-precipitation and solvent-extraction mefhods.l4 The activity was corrected for time-decay and density following the method of Aten.l5 Throughout the experiment irradiations was performed for 3 h. RESULTS AND DISCUSSION Retention and yields expressed in % for anhydrous sodium iodate in the crystalline phase and in alkaline aqueous solution for room-temperature and liquid-nitrogen- temperature irradiation are given in table 1. Results from this work and from earlier investigations have revealed that thermal-neutron capture in crystalline NaIO, at room temperature produces a significantly different pattern of chemical consequences from those observed in chlorates and bromates. For example, a much larger fraction of radioactive halogen is retained by the target 10; ion than C10; and BrO; ions, where retentions in the range ca. 5-30% 16-21 have been observed for similar irradiations.The present retention values, 67.8 and 35.0%, for the crystalline solid and for the frozen solution are in good agreement with the results of Cleary et al. (67.0%)2 and Ambe and Saito (36.0 & 1 % ),4 respectively, which confirms the accuracy and reproducibility of the present results. The values of retention and yields reported are the average of several independent determinations with a reproducibility of The (n, y ) reaction was brought about in the NaIO, target in the presence of the additives KC1, NaCl, KOAc and NaOAc in the solution phase prior to neutron irradiation of an alkaline aqueous solution of the sodium iodate, and their consequent effects on the retention and yields during irradiation are presented in fig. 1 and 2. The 1 % .R.B. SHARMA AND S. P. MISHRA 2629 Table 1. Distribution of recoil lZ6I activity in forms of iodide, iodate and periodate ions produced during thermal-neutron irradiation of NaIO, (target NaIO,) activity distribution (%) state of sample concentration during irradiation T/"C /mol dmP3 10, I- 10, - crystalline solid 25.0 67.8 27.9 4.3 crystalline solid - 196.0 - 56.0 42.0 2.0 alkaline aqueous 25 0.076 23.9 71.7 4.4 frozen alkaline - 196.0 0.076 35.1 56.7 8.2 aqueous solution solution 40 t I 10,- I I I I I I 2 3 4 5 0' [additive]/[ NaIO, ] Fig. 1. Distribution of recoil iodine-128 activity in the form of I-, 10; and 10; ions produced during thermal-neutron irradiation of NaIO, in alkaline aqueous media containing (--O--) NaCl and (--.--) NaOAc additives. major observations are summarized below.(i) The retention and yields found in the alkaline aqueous phase are quite different from those observed in the crystalline solid. (ii) The retention at room temperature is 67.8%, while the retention measured by thermal-neutron irradiation of an alkaline aqueous solution was found to be 24.0%. Also, the retention in the crystalline phase under room-temperature irradiation is higher than the value observed at liquid-nitrogen temperatures. (iii) A significant yield of radioperiodate from alkaline aqueous solutions has been observed. (iv) The activity distribution depends on the nature of the additives. On increasing the concentration of an additive, there is a sharp increase in retention of 4-7%, as is seen in fig. 1 and 2. The yield of radioiodide decreases from 72% to ca.52 % , while the corresponding yield of radioperiodate increases from 4% to ca. 18 % .2630 CHEMISTRY OF lZ81 IN NaIO, SOLUTIONS 70 '% = I I I I I I 0' I 2 3 4 5 [additive]/[NaIO,] Fig. 2. Distribution of recoil iodine-128 activity in the form of I-, 10; and 10; ions produced during thermal-neutron irradiation of NaIO, in alkaline aqueous media containing (-O-) KCl and (--.--) KOAc additives. In a separate set of experiments it was observed that on changing the concentration of targets and additives while keeping the concentration ratio the same, no appreciable change in retention and yields was observed. In a series of parallel studies, the neutron-irradiated solid NaIO, was dissolved in an aqueous solution containing additives of the same concentrations as were used in the case of the alkaline aqueous phase; again no change in retention and yields was observed. The present experiment and those of Cleary et aL2 indicate that a large fraction of lZ81 atoms is initially in the zero oxidate state, except in the case of internal conversion.Therefore, the first reaction the lZ*I atom undergoes after recoil is the insertion of H,O, forming lZ8I- or lZ8IO-. Various experiments show that the radioiodate and radioperiodate are influenced by the additives.1°-12 In the present case, for NaIO, the retentions and yields of iodate and periodate are increased while those of radioiodide decrease, in line with the findings of Bera and ShuklalO for a KIO, target. This is probably due to a stepwise oxidation reaction by oxidizing species formed either from local radiolysis or from additives, as discussed later.The effect of the duration of the neutron irradiation on the value of the retention can be accounted for by either of two extreme phenomena.22 First it is assumed that the initial reaction of the recoil atom is with an electron or hole trapped the pray interaction which produces it. Thus recoil iodine +exciton + I* (slow) (1) followed by I* -+ *IO, (fast). (2) Also, upon exposure of crystalline iodates to energetic radiation, the 10; ion undergoes excitation and ionization leading to rupture of the iodine-oxygen bonds.R. B. SHARMA AND S. P. MISHRA 263 1 A number of reactions can be envisaged, but many of these are excluded by considerations of stability, electron affinity etc.of the fragments. The primary reactions which are most important in the room-temperature irradiation of iodates are the following: IO;+n -+*IO; -+ *IO;+O (3) 4 *IO-+20 (4) -+ *I-+30. ( 5 ) The existence of the above species in the iodates has been assumed, but not confirmed. However, the existence of Br-, BrO-, BrO; etc. in bromates, in both solid and solution phases, has been confirmed by chemical In cases where free radicals are formed in the primary processes, 10; -+ 1 0 , + e- or *IO, + e- (6) these will capture free electrons, forming *IO;.27 Electrons may also be trapped at anion vacancies, forming 10;- as 10; + e -+ IO:-. (7) The reactive intermediate 10;- formed may undergo dehydration to yield 10, and OH-: H2O IO;---+ 10, +OH-. Also, the oxygen atom will remain trapped in the neighbourhood of the sister fragment and react with IO;, forming 10;.The formation of 10; may also occur as follows: *IO, + 10, -+ *IO, + 10. (9) In the radiolysis of pure aqueous solutions of iodates it has been demonstrated2** 2y that various transients like 1 0 , 10, and 10, are formed owing to the interaction of primary radiolytic species, viz. elq, H, OH etc. The following sequence of reactions is envisaged : 10; + e;ls -+ IJv (10) I'V + IIV -+ 10, + 10, (1 1) 10, + 10, -+ 10, + IO- (12) 1 0 , + 10- -+ I- + 10, (13) 10- + 10- + I- + 10, (14) IO,+OH -+ Iv* (15) (16) IV' + IV' -+ 10, + 10,. These reactions explain the degradation of iodate ions via various steps. The formation of various radioiodine fragments in higher oxidation states also depends upon the availability of OH radicals, as expected from both local radiolysis and internal conversion.Some H,O, may also be formed, and according to Haissinsky et aZ.,O H,O, (OH +OH), if produced, will reduce the intermediate radioiodite ion as 10; + H,O, -+ 10- + H,O + 0,. (17)2632 CHEMISTRY OF 1281 IN NaIO, SOLUTIONS In comparison with the solid-state case, the prpbability of recombination is much less pronounced in solution because of the possibility of a redox reaction occurring in the aqueous phase. Therefore the factors that account for the thermalization of recoil atoms may be considered as follows: (i) with an increase in the concentrations of the additives, the ion sphere gradually decreases, (ii) with an increase in the masses of the additives, fewer collisions are required for the thermalization of recoil atoms, (iii) the water molecules in the second, third and higher hydration spheres surrounding a target ion are replaced by the additives and in addition water molecules occupy a smaller volume, (iv) the structure of the solution orients towards a close-packed one and finally (v) the mean distance between the target ions may form a hot reaction zone within a small cage.The behaviour of iodate in solution is more complex. In the aqueous phase the iodate ion has been considered as unhydrated. However, Nightingale et aL31 provided evidence, by comparing the limiting equivalent conductances and viscosity coefficients for the iodate and periodate ions, to indicate the presence of an interaction between iodate ions and water molecules.Water molecules maintain a tetrahedral ~keleton;,~~ 33 the interionic distances of the crystal lattice structure of the solute are assumed to be fixed in solution. According to Fajans et aZ.,34 most of the ions studied here fit into the water structure. The apparent volumes at infinite dilution for Na+, K+, C1- and OH- are 1.7, 8.4, 18.0 and 4.8 cm3 per equiv., respectively, while that of the water molecule is 18.0 cm3 per equiv. The ions having almost equal volumes can replace water molecules in the tetrahedral skeleton, and ionic forces will cause the inter- penetrating water molecules surrounding the ion in the first and second spheres to approach each other and occupy a smaller volume than in pure water.However, the effect of ions on water molecules not immediately in the hydration sphere, i.e. ‘free’ water, is to shift the thermodynamic equilibrium between the tetrahedral and close-packed water structure towards the latter.35 Obviously the ions with largest volumes (25.1 cm3 per equiv for 10; and 40.5 cm3 per equiv. for CH,CO;) will deform the water structure from a regular arrangement. From the above discussion it can be concluded that the probability of the recoil species encountering an additive increases with increasing concentration of the latter. At lower concentrations of additives the total number of added ions with respect to each target ion is four. However, in the case of periodate this number is raised to ca. 114 for KCl, 146 for NaCl and 104 for the acetates, respectively. The hydration numbers are 13 (Na+), 7 (K+), 5 (Cl-) and 11 (OAc-).The volume occupied by the ion atmosphere decreases with the increasing concentration and the water molecules will occupy a smaller volume than in pure water. The main ionic radii with respect to the total concentration decreases from 7.6 A to ca. 3.7 A. Therefore, the number of collisions required for the recoil atom to come to thermal equilibrium will diminish because of the higher mass of the additives. When cations are involved in the recoil reaction, the affinity of the cations for electrons (for example) may affect the oxidation state of the recoil atoms. Salts of different cations have different crystal structures except for the case of isomorphs.This difference in crystal structure is possibly the origin of the cation effect. have concluded that the higher the temperature or the concentration of the dissolved electrolytes, the more looser will be the structure. In the frozen alkaline aqueous state the introduction of 10; and NaOH forms lattice defects in the ice crystals; this may be one of the causes of the oxidation process during the (n, 7) reaction. Normally 12 water molecules constitute a rhombohedra1 of pure ice and the single crystal is found to be the tetragonal with a = 6.27 A and c = 5.79 A. This gives a qualitative idea of on the extent of packing Bernal andR. B. SHARMA AND S. P. MISHRA 2633 in the recombination process. In addition, hypothetical metastable species must be considered to explain the variation in retention.The reaction that lZ8IO- undergoes to form iodate is the oxidation reaction. The oxidizing species is the OH radical formed by the radiolysis of water. The OH radicals formed are concentrated around the 10 A region, while the H atoms diffuse away to ca. 150 A. High internal conversion in lZ81 is known to 39 and the emitted X-ray and highly positively charged iodine atoms are also expected to produce OH radicals. In presence of chloride, the increase in retention can also be explicable in the light of the formation of oxidizing radicals such as OH and C1,. Also, it is known that the C1- ion is a OH radical scavenger;** as the concentration of C1- increases the OH radicals are scavenged. The formation of C1; ions by irradiation of the solution cannot be accounted for by the theory of diffusion for 42 In the light of above model the Cl- ions present in large concentrations in the iodate/chloride target may react with H,O+ forming C1, as H,O+ + C1- C1+ H,O c1- 3- Cl -+ Cl,.The chloride and acetate additives may also be oxidized by OH radicals to hypochlorites and peracetates, which are strong oxidizing agents. At a higher concentration of acetate the deviation in activity distribution may be due to the association of acetate ions, which on reaction with OH probably form other compounds (e.g. acetone) rather than the peracetate Also, as the additive concentration increases the distance between the two ions With the increase of additives the dimensions of the ‘hot’ volume will tend towards a more compact form, thus increasing the probability of more high-energy atoms being trapped within this small volume Thanks are due to the referees for their valuable comments and suggestions and to Miss Tina Maxian for her graphic work.Chemical Eflects of Nuclear Transformations in Inorganic Systems. G. Harbottle and A. G. Maddock (North-Holland, Amsterdam, 1979). R. E. Cleary, W. H. Hamill and R. R. Williams, J. Am. Chem. Soc., 1952,744675. A. V. Bellido, Radiochim. Acta, 1967, 7, 122. F. Ambe and N. Saito, Radiochim. Acta, 1970, 13, 105. B. M. Shukla, R. N. Singh and S. P. Mishra, Proc. Chem. Symp. Bombay, Department of Atomic Energy (1970). R. N. Singh and B. M. Shukla, Radiochim. Acta, 1980, 27, 135. H. J. Arnikar, V. G. Dedgaonkar and K. K.Shrestha, J. Univ. Poona, Sci. Technol., 1970, 38, 169. H. J. Arnikar, V. G. Dedgaonkar and K. K. Shrestha, J. Univ. Poona, Sci. Technol., 1970,38, 177. ’ R. N. Singh and B. M. Shukla, Radiochim. Acta, 1980, 27, 11. lo R. K. Bera and B. M. Shukla, Radiochim. Acta, 1978, 25, 27. l1 S. P. Mishra, R. Tripathi and R. B. Sharma, Indian J. Chem., 1983, 22, 514. l 2 S. P. Mishra, R. B. Sharma and R. Tripathi, Indian J . Chem., 1983, 22, 790. l 3 R. B. Sharma and S. P. Mishra, Znorg. Chim. Acta, 1984, 86, 151. l4 G. E. Boyd and Q. V. Larson, J. Am. Chem. SOC., 1969,91,4639. l6 A. H. W. Aten Jr, Nucleonics, 1950, 6, 68. l6 N. K. Aras, B. Khan and C. D. Coryell, J. Znorg. Nucl. Chem., 1965, 27, 527. l7 H. Seiler and M. Seiler, Helv. Chem. Acta, 1967, 50, 2477. l9 J. S. Bube and S.R. Mohanty, Radiochim. Acta, 1971, 16, 46. 2o J. W. Chase and G. E. Boyd, ASTM Spec. Tech. Publ., 1966, 400, 17. 21 G. E. Boyd and T. G. Ward Jr, J. Chem. Phys., 1964,68, 3809. 22 Y. C. Lin and D. R. Wiles, Radiochim. Acta, 1970, 13, 43. 23 T. Anderson, J. R. Byberg and K. J. Olsen, J. Phys. Chem., 1967,71, 4129. ** G. E. Boyd and Q. V. Larson, J. Am. Chem. SOC., 1968,90, 254. C. W. Owens and F. S. Rowland, J. Znorg. Nucl. Chem., 1962, 24, 133.CHEMISTRY OF '''1 IN NaIO, SOLUTIONS L. C. Brown, G. M. Begun and G. E. Boyd, J. Am. Chem. SOC., 1969, 91, 2250. S. Radhakrishna and A. M. Karguppikar, J. Phys. SOC. Jpn, 1973,35, 578. A. V. Bellido and D. R. Wiles, Radiochim. Acta, 1969, 12, 94. J. K. Thomas, S. Gordon and E. J. Hart, J. Phys. Chem., 1964, 68, 1524. M. Anbar and E. J. Hart, Adv. Chem. Ser., 1968, 81, 79. M. Haissinsky, J. Jove and W. Szymansky, J. Chim. Phys., 1964, 61, 572. E. R. Nightingale Jr and R. F. Bench, J. Phys. Chem., 1959, 63, 1777. J. S. Rosen, NASA Tech. Note, Nasa-Tnd-4297, 1968, p. 17. A. K. Lyashchenko, Zh Strukt. Khim., 1968, 9, 781. K. Fajans and 0. Johnson, J. Am. Chem. SOC., 1942, 64, 668. I. G. Mikhailou and Yu. P. Syrnikov, Zh. Strukt. Khim., 1960, 1, 12. J. D. Bernal and R. H. Fowler, J. Chem. Phys., 1933, 1, 515. Kamb Barclay, Acta Crystallugr., 1964, 17, 1437. R. G. Kortling, J. D'Auria, C. H. W. Jones and T. Isenhour, Nucl. Phys., 1969, 38, 392. L. A. Schaller, J. Kern and B. Michaund, Nucl. Phys., 1971, 165, 415. J. Pucheault and C. Ferradine, J. Phys. Chem., 1980 3, 83. S. Khorana and W. H. Hammill, J . Phys. Chem., 1975, 75, 308. T. Sawat and W. Hamill, J. Phys. Chem., 1970, 74, 3914. M. W. Garrison, W. Bennett, S. Cole, H. R. Haymond and M. W. Boyd, J. Am. Chem. Suc., 1955, 77, 2720. R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworths, London, 1955), p. 16. 25 26 27 28 29 30 31 32 33 34 35 36 37 3R 39 40 41 42 43 44 (PAPER 4/1089)

ISSN:0300-9599    

DOI:10.1039/F19858102627

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1985, 81, 2635-2646 Effect of Sodium Polyacrylate on the Properties of Dilute and Concentrated BaSO, Dispersions BY PETER A. WILLIAMS* AND RAYMOND HARROP Research Division, The North East Wales Institute of Higher Education, Deeside, Clwyd CH5 4BR AND IAN D. ROBB Unilever Research Laboratory, Port Sunlight, Merseyside L63 3JN Received 13th September, 1984 The adsorption of sodium polyacrylate on BaSO, particles has been studied and the adsorption characteristics have been related to the properties of both dilute and concentrated dispersions. The extent of particle aggregation has been monitored by turbidity or viscosity measurements for dilute and concentrated dispersions, respectively. For dilute dispersions, the results show that in the absence of electrolyte, aggregation is prevented when the extent of adsorption is much less than plateau coverage.The e.s.r. technique reveals that the polyanions are adsorbed in ‘trains’ under these conditions and, since the zeta potential is sufficiently high, charge stabilisation predominates. For concentrated dispersions the same mechanism is effective once a critical polymer concentration has been reached, this concentration being greater than that for dilute dispersions but still less than plateau coverage. This is possibly due to the requirement of a more uniform charge distribution. In the presence of electrolyte the amount of polymer necessary to bring about stabilisation corresponds closely to the amount of polymer adsorbed at the end of the high-affinity region of the adsorption isotherm for both dilute and concentrated dispersions. Under these conditions the adsorbed polymer adopts a more extended configuration with loops and tails protruding out into solution, thus enabling steric repulsive forces to come into operation.Recent publications on the adsorption of polyelectrolytes onto colloidal particles have not only been concerned with the behaviour and configuration of the adsorbed molecules but also their effect on the colloidal stability of the particle~.l-~ These studies have dealt with dilute dispersions where the solid phase volume has been < 1 % . Under these conditions flocculation can be modelled by the particles approaching from large separations with the probability of collisions being approximately determined by classical DLVO considerations4 and the kinetics being interpreted according to modified Smoluchowski theorie~.~? However, for more concentrated systems, where the dispersed phase volume is > lo%, the distance between particles may be sufficiently small so that fluctuations in the surface charge cannot be averaged or considered as smeared out.In addition, irregularly shaped particles at close average separations may well have points or edges in contact markedly altering the rheological behaviour of the total dispersion. The aim of this work has been to study the adsorption of sodium polyacrylate onto BaSO, crystals and to relate the adsorption characteristics to the properties of both dilute and concentrated dispersions. 263 52636 EFFECTS OF SODIUM POLYACRYLATE ON BaSO, EXPERIMENTAL MATERIALS Commercial grade BaSO, was washed five times by decantation with hot distilled water, freeze dried and oven dried at 105 "C for 12 h.The particle size was determined using electron microscopy and the size distribution was shown to be positively skewed with a mode at 0.25 pm. The surface area was found to be 2.9 m2 g-l by nitrogen adsorption. ESCA showed the surface to contain the elements Ba, S, 0, Ca, and P in the ratio 1 : 1.05:4.34:0.30:0.05. The polymers used in this work were three commercial samples of sodium polyacrylate (SPA) having quoted relative molecular masses of 5000, 20000 and 230 000. These were dialysed overnight against distilled water using an Amicon hollow-fibre dialysing unit and freeze dried before use.AnalaR sodium citrate (B.D.H.) was used without further purification. Acridine orange (Sigma) was twice recrystallised as the free base from ethanol +water by dropwise addition of NaOH (0.1 mol dm-3).7 The precipitate was collected and dried under vacuum at 70 "C for 4 h. The free base was then converted to the hydrochloride by adding an exact amount of ethanolic HC1 and the neutralised acridine orange was recovered by evaporating off the ethanol and drying under vacuum at 80 "C for 4 h. METHODS VISCOSITY OF CONCENTRATED DISPERSIONS Dispersions of 50% w/w BaSO, (equivalent to a volume fraction of 0.18) were prepared by adding 2 cm3 polymer or citrate solution at pH 7 to 2 g BaSO, and sonicating for a few minutes using a Rapidus 300 instrument (Ultrasonics Ltd) at 300 W and 29 kHz to disperse all of the aggregates.In the case where the supernatant was analysed for residual SPA, double the above quantities were used. Further dispersions were prepared using polymer and citrate in the presence of 0.5 mol dm-3 NaCl at pH 7. The viscosities of these dispersions were determined at 25 "C at a range of shear rates using a Wells-Brookfield cone and plate viscometer fitted with the 1.565" cone and taking 1 cm3 samples. STABILITY OF DILUTE DISPERSIONS Dispersions 0.05% w/v were prepared by adding 0.05 g BaSO, to 100 cm3 SPA solution at various concentrations in water or in 0.5 mol dm-3 NaCl at pH 7. The samples were sonicated to ensure complete dispersion. In this work the stability was assessed from the wavelength dependence of the turbidity, as outlined previously.8 The dispersions were allowed to stand for 18 h then gently inverted to resuspend any sedimented particles and the absorbance measured over the range 550-800 nm at intervals of 50 nm using a Hewlett Packard 8541A diode array spectrophotometer.The plot of log absorbance against log wavelength was found to be linear in this range and the slope of the line, n, was obtained by linear regression. The greater the value of n the smaller the particle size. ADSORPTION ISOTHERMS 10 cm3 SPA solution at various concentrations in water or in 0.5 rnol dm-3 NaCl at pH 7 were added to 1 g BaSO, and the dispersion sonicated for 2 min, tumbled overnight at room temperature and centrifuged at 1300 g. The supernatant was analysed for SPA by utilizing the fact that acridine orange binds quantitatively to polyanions with a reduction in fluorescence intensity directly proportional to the concentration of the p~lyanion.~ The method adopted here was to add 0.25 cm3 of the supernatant to 10 cm? of 1.5 x mol dm-3 acridine orange solution.The solution was excited at 400 nm and the emission at 530 nm recorded using a Perkin-Elmer MPF43A spectrofluorimeter. The concentration of SPA was found from calibration curves obtained from known SPA concentrations. ADSORBED POLYMER CONFIGURATION The adsorbed polymer configuration was studied using the e.s.r. technique. The SPA was spin-labelled using 4-amino-2,2,6,6-tetramethylpiperidinyloxy (4-amino-TEMPO). Details on the spin labelling of the polymer samples and the interpretation of the e.s.r.spectra haveP. A. WILLIAMS, R. HARROP AND 1. D. ROBB 2637 n n U O 0 1 I I I rr 0 - 0 v 1 I I I 50 100 150 200 equilibrium polymer concentration/mg dm-3 Fig. 1. Isotherms for the adsorption of SPA [(a) 5000, (b) 20000 and ( c ) 230000] on BaSO,. 0, from water at pH 7 and e, from 0.5 mol dmF3 NaCl at pH 7. been given previously.l0. l1 The procedure adopted was as follows. Spin-labelled SPA (230000) was adsorbed from water at pH 7 (10 cm3) onto BaSO, (1 g) to the point corresponding to the start of the plateau region of the adsorption isotherm. Unadsorbed polymer was removed by washing with water and the adsorbed-polymer configuration was established by measuring the e.s.r. spectrum of the dispersion on a Jeol JES MEIX e.s.r.spectrometer using a flat cell suitable for aqueous solutions. The same dispersion was tumbled overnight in the presence of unlabelled SPA (230000) at 5 % w/w concentration and the adsorbed polymer configuration assessed without further washing. RESULTS The isotherms obtained for the adsorption of SPA onto BaSO, from water and 0.5 mol dm-3 NaCl are given in fig. 1. For the three SPA samples studied the amount adsorbed is seen to be greater for adsorption from 0.5 mol dm-3 than from water, as has been found previous1y.l Whereas the adsorption capacity for all three SPA samples was similar for adsorption from 0.5 mol dm-3 NaCI, the lower-molecular-mass2638 EFFECTS OF SODIUM POLYACRYLATE ON BaSO, v) m a x E 1 w .c( x .C. m I E -0 1000 - 0 500 - 3 0 I *ot 0 0.5 1.0 1.5 mg polymer per g BaS04 Fig.2. (a) Stability of dilute BaSO, dispersions (0.05% w/v) at pH 7; plot of n = d log absorbance/d log ,I as a function of added SPA (5000). (b) Viscosity of concentrated BaSO, dispersions (50% w/w) at a shear rate of 1.15 s-l as a function of added SPA (5000). (c) Concentration of unadsorbed polymer as a function of SPA (5000) added for the same dispersions as in (b). sample adsorbed to a greater extent from water, suggesting that the BaSO, may be slightly porous. The colloid stabilities of the dilute dispersions are compared with the viscosities of the corresponding concentrated dispersions and the amount of polymer adsorbed in fig. 2(a), 3(a) and 4(a). The viscosities of the 50% w/w dispersion were found to decrease with increasing shear rate and the results obtained at a shear rate of 1.15 s-l for dispersions prepared using SPA in the absence of electrolyte are given in fig.2(b), 3(b) and 5 (a) and in the presence of electrolyte in fig. 4(b). The viscosities are seen to decrease sharply once a critical amount of polymer is added and this critical concentration is compared with the amount of SPA adsorbed in fig. 2(c), 3(c) and 5(b).n P. A. WILLIAMS, R. HARROP AND I. D. ROBB 2639 E a 50t - 5 'I a :i I I I 0 0.5 1 .o 1.5 mg polymer per g BaSO, Fig. 3 (a) Stability of dilute BaSO, dispersions (0.05% w/v) at pH 7; plot of n = d log absorbance/d log 1 as a function of added SPA (230000). (b) Viscosity of concentrated BaSO, dispersions (50% w/w) at a shear rate of 1.15 s-' as a function of added SPA (230000).(c) Concentration of unadsorbed polymer as a function of SPA (230000) added for the same dispersions as in (b). Further addition of SPA to give concentrations greater than the critical amount results in an increase in the dispersion viscosity, which is much more apparent for the high-molecular-mass polymer (see fig. 6). The relative viscosity of the dispersions [ie. q(dispersion)/q(liquid phase)], however, does not increase after the sharp initial decrease. The e.s.r. spectra obtained for SPA adsorbed from dilute solution and also after tumbling in the presence of a concentrated polymer solution are given in fig. 7. Spectrum (a) is a typical broad immobile-type spectrum indicating that the SPA adsorbs with a large number of segments in trains, whereas spectrum (b), obtained2640 0.7 0.E n Of 0.4 0.: 0.2 150C rn (d PI E -s 1ooc Y .4 0 .$ 5oc EFFECTS OF SODIUM POLYACRYLATE ON BaSO, I I I 0 1.0 2.0 3.0 4.0 mg polymer per g BaS04 Fig.4. (a) Stability of dilute BaSO, dispersions (0.05% w/v) in the presence of 0.5 mof dm-3 NaCl at pH 7; plot of n = d log absorbance/d log A as a function of added SPA (0, 5000 and a, 230000). (b) Viscosity of concentrated BaSO, dispersions (50% w/w) at a shear rate of 1.15 s-' in the presence of 0.5 mol dmP3 NaCl as a function of added SPA (0,5000; 0,20000 and 0, 230000). after simply tumbling the dispersion in 5% unlabelled polymer, is a motionally narrowed mobile-type spectrum and indicates that the SPA has now rearranged on the surface such that the majority of segments are protruding out into solution in the form of loops and tails.It was also found that the supernatant contained traces of labelled SPA that had desorbed from the surface. The effect of sodium citrate on the viscosity of concentrated BaSO, dispersions is illustrated in fig. 8. In the absence of electrolyte the dispersion viscosity decreases rapidly at very low citrate concentrations but tends to increase again at high concentrations (0.25 mol dm-3). With electrolyte present the decrease in viscosity is far less pronounced and the minimum viscosity obtained is an order of magnitude greater than for viscosities obtained in pure water.P. A. WILLIAMS, R. HARROP AND I. D. ROBB 1500 v) m PI E z 1000- - 2641 0 0.5 1.0 1.5 mg polymer per g BaSO, Fig.5. (a) Viscosity of concentrated BaSO, dispersions (50% w/w) at a shear rate of 1.15 s-I as a function of added SPA (20000). (b) Concentration of unadsorbed polymer as a function of SPA (20000) added for the same dispersions as in (a). DISCUSSION Dilute dispersions of BaSO, can be readily stabilised in the absence of electrolyte by the addition of SPA, as shown by the turbidity data presented in fig. 2(a) and 3 (a). For the 5000 molecular-mass sample colloid stability occurs at approximately one-tenth of the adsorption capacity, whereas for the 230 000 molecular-mass sample one-third of the adsorption capacity is required. It was previously shown that adsorption enhances the surface charge and that the molecules adsorb in trains on the BaSO, surface,l* l3 so that stabilisation is brought about through a charge mechanism rather than a steric one.The reason for the lower stabilising efficiency of the high- molecular-mass polymer is not clear, although it may be related to the actual distribution of adsorbed molecules between and on the particles. It is calculated for the 230000 molecular mass SPA that at the onset of stability there are ca. 30 molecules adsorbed per particle and that these occupy < 5% of the available surface. The average charge on the particles will increase with increasing adsorbed polymer, although particularly for irregularly shaped crystals this increased charge is probably not uniformly distributed over the entire surface. As indicated by Gregory,12 adsorption of polymers, especially larger ones, may occur in patches leaving some parts of the crystal surface with a deficiency of polymer and thus a lower charge.With strong interactions between the polymer segments and the crystal, translation across the surface by diffusion would be slow. Smaller polymer molecules may thus be able2642 EFFECTS OF SODIUM POLYACRYLATE ON BaSO, 1 I 1 100 150 200 mg polymer per g BaSO, Fig. 6 (a) Viscosity of BaSO, dispersions (50% w/w) as a function of SPA (0, 5000 and 0, 230000) added at a shear rate of 1 I .5 s-l. (b) Relative viscosity [q(dispersion)/q(liquid phase)] of concentrated BaSO, dispersions as a function of added SPA (230000) at a shear rate of 11.5 s-l. to achieve a more uniform coverage of the surface and hence colloid stability would be attained at lower surface coverage.In the presence of high concentrations of electrolyte (0.5 mol dm-3 NaC1) the SPA adopts a more extended configuration on the surface1? l3 and, since the electrostatic repulsions are screened under these conditions, it is likely that stabilisation results predominantly from steric repulsive forces. However, the turbidity data presented in fig. 4(a), together with previous results,l shows that the stability achieved in the presence of electrolyte is less than that achieved in its absence. Note that stability in the presence of 0.5 mol dm-3 NaCl occurs at polymer concentrations correspondingP. A. WILLIAMS, R. HARROP AND I. D. ROBB 2643 1; Fig. 7. E.s.r. spectra of SPA (230000) adsorbed onto BaSO, (a) from water at pH 7 and (b) after tumbling the same dispersion 18 h in the presence of 5% w/v unlabelled SPA (230 000).to the end of the high-affinity region of the adsorption isotherm. This is in agreement with previous work reported on sterically stabilised systems2? and lends further support to the theory of Cohen Stuart et aL1* suggesting that the start of the shoulder on adsorption isotherms corresponds to saturation coverage for one polymer fraction and that the shoulder itself is due to the polydispersed nature of the sample. It is surprising that the low-molecular-mass polymer (5000) is more efficient than the high-molecular-mass polymer (230 000) in conferring steric stability, but this may be a result of its lower tendency to induce polymer bridging. The properties of concentrated BaSO, dispersions are also greatly affected by the addition of SPA [fig.2(b), 3(b) and 5(a)]. In the absence of electrolyte the critical polymer concentration (c.P.c.) required to bring about the sudden viscosity decrease is greater than the concentration required for stabilisation of the dilute dispersions and corresponds to ca. 70% of the adsorption capacity (taken at the end of the high-affinity region on the isotherm) for the 5000 molecular mass polymer and > 90% for the 230000 molecular-mass polymer. The results are shown in table 1. The sharp decrease in viscosity for the concentrated dispersions occurs when the particle-particle interactions change from attraction to repulsion. Clearly a higher surface coverage of polymer is required for stabilization in concentrated dispersions than in dilute ones.There is no firm explanation for this, although it is possible that the closer distance of approach of particles in concentrated dispersions reduces the2644 i 50 2 EFFECTS OF SODIUM POLYACRYLATE ON BaSO, t 0 0 25 50 75 100 125 Fig. 8. Viscosity of concentrated BaSO, dispersions (50% w/w) as a function of sodium citrate added, at a shear rate of 11.5 s-1: 0, in water at pH 7 and 0, in 0.5 mol dmV3 NaCl at pH 7. mg sodium citrate per g BaS04 Table 1. Adsorption capacity of BaSO, adsorption from 0.5 mol dm-3 NaCl at pH 7 adsorption from water at pH 7 adsorption capacity adsorption capacity/mg 8-l /mg 8-l molecular 10% w/w 50% w/w c.p.c./ 10% w/w c.p.c/ mass of SPA dispersion dispersion mg g-l dispersion mg g-' 5 000 1.0 +o.1 0.90 0.65 1.8 f 0.10 1.6 20 000 0.75 k 0.05 0.90 0.65 1.65 +_ 0.05 1.55 230 000 0.85 0.85 0.80 1.75 & 0.05 1.6 effect of smearing-out patches of charges on the particle surface. In concentrated dispersions, even with 50% of the particle surface covered with polymer, and after sonication, the high viscosity of the dispersion indicates that many of the particles are in contact, giving an extended network of particles. The same particles in a dilute system after sonication would not flocculate, as the electrostatic repulsion between the particles would be sufficient to maintain their (kinetic) stability. The need to have high surface coverage of polymer is probably increased in these systems of irregularly shaped particles compared with that in systems of spheres.Contacts at points and edges of crystals would be less influenced by charges on adjacent flat surfaces than contact between spheres carrying the same charge. In the presence of electrolyte the critical polymer concentration required to bring about the viscosity decrease is between 1.55 and 1.60 mg g-l for the three SPA samplesP. A. WILLIAMS, R. HARROP AND I. D. ROBB 2645 [fig. 4(b)], closely corresponding to the end of the high-affinity region of the adsorption isotherms shown in fig. 1 (see table 1). Under these conditions e.s.r. results 1v l3 indicate extended polymer configurations giving steric stability to the particles. The onset of steric stability at the end of the high-affinity part of the isotherm rather than at the plateau coverage may be explained as follows.If additional adsorption above the end of the high-affinity part of the isotherm is a result of fractionation as described by Cohen Stuart et a1.,14 then this fractionation does not take place in the timescale of the approach of colloid particles. Indeed fractionation on adsorption takes minutes to hours whereas the approach of colloidal particles will occur in less than a second. The increase in viscosity of the dispersions at high polymer concentrations, particularly for those using the high-molecular mass polymer, is due to the unadsorbed polymer increasing the viscosity of the liquid phase since the relative viscosity does not increase with increasing polymer concentration (see fig. 6). Confirmation of the role of electrostatic and steric forces in controlling the viscosity of concentrated dispersions was gained by preparing samples in the presence of sodium citrate, a molecule too small to give any steric stability.In the absence of electrolyte it was found that the dispersion viscosity decreased at low citrate concentrations (see fig. 8), but that at higher concentrations (> 0.25 mol dm-3 citrate) the viscosity started to increase considerably. The initial viscosity reduction is probably brought about by electrostatic repulsions, as was the case for SPA, but the consequent increase in viscosity at higher citrate concentrations is likely to be as a result of the increasing ionic strength of the solute itself leading to flocculation of the BaSO,. The fact that 5000 molecular-mass SPA does not show a similar viscosity increase at higher polymer concentrations is explained by the e.s.r.results (see fig. 7), which indicate that the polymer molecules rearrange on the surface and adopt a more extended configuration, thus enabling steric repulsions to operate. In the presence of 0.5 mol dm-3 NaCl the influence of sodium citrate on the dispersion viscosity was much less dramatic, indicating as above that the charge mechanism alone cannot completely prevent the particles from aggregating under these conditions. CONCLUSIONS SPA reduces the viscosity of concentrated dispersions of BaSO, in the presence and absence of electrolyte. In the absence of electrolyte more polymer is needed to ' stabilise ' concentrated dispersions compared with dilute dispersions.The reason for this may be that in concentrated dispersions the closer distance of approach reduces the smearing out of patchiness in the surface charge arising from non-uniform polymer adsorption. At high ionic strength the polymer provides steric stability at the end of the high-affinity portion of the isotherm, as would be expected if further adsorption were a result of fractionation on adsorption. The increase in the viscosity of the dis- persions at higher polymer concentrations is due to an increase in the viscosity of the liquid phase due to the polymer itself. P. A. Williams, R. Harrop, G. 0. Phillips, I. D. Robb and G. Pass, in The Efect of Polymers on Dispersion Properties, ed. Th. F. Tadros (Academic Press, London, 1982), p. 361. P. A. Williams, R. Harrop and I. D. Robb, J . Chem. SOC., Faraday Trans. I, 1984, 80, 403. P. A. Williams, R. Harrop and I. D. Robb, J . Colloid Interface Sci., 1984, 102, 548. P. G. Cummins, A. L. Smith, E. J. Staples and L. Thompson, in Solid-Liquid Separation, ed. J. Gregory (Ellis Horwood, Chichester, 1984), p. 161. M. von Smoluchowski, 2. Phys. Chem., 1918, 92, 1292646 EFFECTS OF SODIUM POLYACRYLATE ON BaSO, L. A. Spielman, J. Colloid Interface Sci., 1970, 33, 562. A. L. Stone and D. F. Bradley, J. Am. Chem. Soc., 1961,83, 3672. J. A. Long, D. W. J. Osmond and B. Vincent, J. Colloid Interface Sci., 1973, 42, 545. G. P. Diakun., H. E. Edwards, J. C. Allen., G. 0. Phillips and R. B. Cundall., Anal. Biochem., 1979, 94, 378. M. C. Cafe and I. D. Robb, Polymer, 1979, 20, 513. lo K. K. Fox, I. D. Robb and R. Smith, J. Chem. Soc., Faraday Trans. 1, 1974,70, 1186. l2 J. Gregory, J. Colloid Interface Sci., 1976, 55, 35. l3 M. C. Cafe and I. D. Robb, J. Colloid Interface Sci., 1982, 86, 41 1. l4 M. A. Cohen Stuart, J. M. H. M. Scheutjens and G. J. Fleer, J. Polym. Sci., Polym. Phys. Ed., 1980, 18, 559. (PAPER 4/ 158 1)

ISSN:0300-9599    

DOI:10.1039/F19858102635

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985, 81, 2647-2658 The Rotating Optical Disc-Ring Electrode Part 1 .-Collection of a Stable Photoproduct BY W. JOHN ALBERY* AND PHILIP N. BARTLETT Department of Chemistry, Imperial College, London SW7 2AY AND ANNA M. LITHGOW, JORGE RIEFKOHL, LORRAINE ROMERO AND FERNANDO A. SOUTO Department of Chemistry and Center for Energy and Environment Research, University of Puerto Rico, Mayaguez, Puerto Rico 00708, U.S.A. Received 14th September, 1984 The rotating optical disc-ring electrode consists of a transparent disc surrounded by a concentric ring electrode. Light is shone through the central disc to drive a photoredox system. The current from the photogenerated product is measured downstream on the ring electrode. The convective diffusion equation for this system is solved and analytical solutions are presented for the collection of a stable product.The theory takes into account the bleaching of the solution. Experiments performed on the iron-thionine system are found to be in good agreement with the theory. A useful electrode for the investigation of photoelectrochemical systems is the ‘ rotating photoelectrode’ developed by Johnson and co~orkers.~-~ This electrode consists of a central quartz disc through which light is shone into the electrolyte solution; the disc is surrounded by a concentric ring electrode, and the whole assembly is rotated. Such a ring-disc arrangement is similar to a conventional ring-disc electrode. The downstream ring electrode detects products and intermediates generated photochemicaly by the light shining through the disc.For instance, the electrode has been used to study the photopinacolisation of benzophenone in strongly alkaline alcohol-water mixtures.2 Again there are similarities to the optical rotating-disc electrode (ORDE), where light is shone through a semitransparent disc electrode and the photochemical intermediates and products are detected on the disc electr~de.~-~ Since the term ‘rotating photoelectrode’ does not discriminate between these two types of‘ electrode, we prefer to call the ring-disc system a ‘rotating optical disc-ring electrode ’ (RODRE). The advantage of using rotating-disc hydrodynamics is that one can often calculate how much of the electrochemically or photochemically generated product will be detected by the ring electrode.Hitherto these calculations for the RODRE2 have relied on the simulation method of Feldbergg as developed by Prater and Bardlo for ring-disc systems. In this paper we present an analytical theory for the collection of a stable photoproduct at the RODRE. Experimental results for the iron-thionine system are presented and shown to be in good agreement with the theory. 26472648 RING-DISC ELECTRODES THEORETICAL DISC ZONE We consider the following reaction scheme : disc zone : hv A + B ring electrode : B & ne + products. We assume that in the zone of the disc there is no significant radial variation in the irradiance of the incoming light. Then in the zone of the disc we have the following convective diffusion equation for the generation of B : a2b ab TI, D --u, -+- exp ax2 ax X& where b is the concentration of B, D is its diffusion coefficient, x measures distance normal to the electrode, u, is the velocity in the x direction, I, is the irradiance at the disc surface, cp is the quantum efficiency of the photoredox reaction and X&, the absorption length,5 is given by XE = (&a)-l.(2) In eqn (2) a is the concentration of A and E is its natural extinction coefficient. The absorption length describes the distance over which the light is absorbed in a Beer-Lambert profile. Eqn (1) is the same for the ORDE, but whereas for the ORDE the boundary condition for b on the disc surface is that b = 0, For the RODRE the boundary condition is that at x = 0, (abiax), = 0. (3) The interplay of the three lengths, the absorption length, the diffusion length, X,, and the generation length, XG, is much the same, where X,,, the thickness of the diffusion layer, is given by the Levich equation1' and XG describes the distance A diffuses in the irradiance I,: XG = ( D / ~ I , E ) ~ / ~ .(4) For most systems X& + XD, and we can ignore the exponential term in eqn (1). We can also write a+b = am ( 5 ) where a, is the bulk concentration of A. We solve eqn (1) with boundary condition (3) and the fact that b + 0 as x -+ GO by substitution of eqn ( 5 ) followed by splitting the solution into one part with the diffusion term inside the diffusion layer and a second part with the convective term outside the diffusion layer; the two solutions are joined at the edge of the diffusion layer.This procedure had been shown to work well for the ORDE.5 As shown in Appendix 1, we obtain the following result for the concentration of B on the disc surface : wherew. I. ALBERY et al. 2649 0.0 2 .o 4.0 6.0 Y Fig. 1. Plot of eqn (6) showing the surface concentration of B on the disc electrode as a function of the bleaching parameter y defined in eqn (7). diffusion -- layer 0 9 9 0.OL I I 1 0.0 ID 2 .o 3.0 X l X D Fig. 2. Concentration profile of B for the unbleached case ( y -+ 0) calculated from eqn (A 2) and (A 4). Inspection of eqn (6) shows that when X , >$XG ( y > 1) the solution inside the diffusion layer becomes bleached. The species A cannot survive its passage across the diffusion layer, and the concentration of B then becomes equal to the bulk concentration of A.On the other hand, when X , < X , ( y < 1) then we find that b, = a , y = a, vI0 &Xf,/D. (8) Under these conditions only a small fraction of A is converted into B. The right-hand side of eqn (8) shows the balance between the generation of B from A, vIoe, and the2650 RING-DISC ELECTRODES Table 1. Boundary conditions at x = 0 disc gap ring r < rl rl < r < r2 r2 < r < r3 loss of B by convective dilution DIPD, where1l.l2 XD = 1 .288(D/C)1/3 = 0.643 W-1/2~1/6D1/3 (9) Wis the rotation speed in Hz and v is the kinematic viscosity. Fig. 1 shows a plot of eqn ( 6 ) and the two regions for the bleached and unbleached cases. In fig. 2 we show the concentration profile for B from the disc surface across the diffusion layer and out into the solution when the system is unbleached.GAP ZONE Outside the disc zone the convective diffusion equation for B must include the term for radial convection, but not the photogeneration term : Following the usual ring-disc procedure this equation is transformed with x = (r/r1)(C/D)1/3x and 5, = ~ 4 ~ ) ~ - 11/3 to give Following our usual procedurel37 l4 we now split b into three contributions, bD, bG and b,, corresponding to the disc, gap and ring zones, respectively: The term b, is zero in the zone of the disc and the term b, is zero in both the zone of the disc and the zone of the gap. The boundary conditions at x = 0 for the different contributions are given in table 1. However, the differential equation for b , in the disc zone contains the photo- generation term. Hence b, does not obey eqn ( 1 1 ) ; inclusion of the photogeneration term and transformation with the ring-disc variables gives b = bD-bG-bR.( 1 2) where Y ”) = exp ( y / 2 ) cosh ( ~ l / ~ ) - 1w. J. ALBERY et al. 265 1 and b, is the concentration of B on the disc surface, given by eqn (6). In the unbleached case when y is smallfly) = 1. On the other hand, in the bleached casefly) --+ 0. In order for b to obey eqn (1 1) in the gap and ring zones it is necessary for bG to obey the following differential equation : a2bG abG =x---0.60bofly)(l +3<1)-2/3. ax2 at, In addition to the boundary condition in table 1, we have the two further boundary Taking the Laplace transform of eqn (15) with respect to 5, we obtain15 conditions that b, = 0 both at t, = 0 and as x + GO.where and r is the incomplete gamma function.16 By writing [ = sy3x w = s2/3 and G 1 /ndsl) eqn (16) becomes a 2 W p p - cw = - 1 in. With the boundary conditions the solution to this differential equation is17 w = Gi([)+3-lI2Ai([). On the disc surface [ = 0, and the value of w is17 wo = 2Ai(0)/2/3 = 0.41. Substitution of this value in eqn (17) and (18) gives an expression for 6G,o, which describes the variation of 6 , on the gap surface: For s, > 3 the incomplete gamma function can be expanded18 in terms of the form s;%, which can then be inverted term by term to give The condition s, > 3 corresponds to an assumption that, compared to the radius of the disc, the electrode has a thin ring and a thin gap.As discussed below, the treatment fails when [ ( r a / r ~ ) ~ - 13 2 1 /3. RING ZONE In the zone of the ring electrode b, obeys eqn (1 1) with 5, replaced by c2, where r2 = (r3 - r3/3r;. The usual Laplace transformation gives2652 RINGDISC ELECTRODES The ring current, i,, is given by where t;l = (r: - r 3 / 3 r : . From the boundary conditions in table 1 (ab/Q), = -(ab,/Q),, and the integral in eqn (21) is found by inverting 1/s2 times the Laplace transform of the gradient given in eqn (20). The value of bR,, on the right-hand side of eqn (20) is given by the boundary condition in table 1 : tR, 0 = bob2 - &, 0 (23 1 where b, is given by eqn (6) and &, , is the Laplace transform of bG, , [eqn (19)] with respect to r2. In eqn (19) tl is replaced by t2+ti, where Using the convolution integral we find that r; = (r;/r;-- 1 ) / 3 .( 2 4 ) where I&, q) is the incomplete beta function.l9 (20)-(23) and substitution for XD from eqn (9), gives Application of this lemma to each of the terms in eqn (19), together with eqn THE BLEACHED CASE We start by considering the case where the irradiance is large or the rotation speed is low, and the solution close to the disc electrode becomes bleached. Under these conditions, as discussed above, y is large [eqn (7)], fly) tends to zero [eqn (14)] and b, = a,. In eqn (25) we need only consider the first term in the large parentheses. In our previous workz0 we showed that the ratio of the limiting current at the ring electrode to that at the disc electrode was given by p/3, where = 35;.Inspection then shows that the ring current is identical to that which would be observed for the limiting current from the reaction of A. This is not surprising, since the light is converting all the A into B upstream of the ring electrode. Under these conditions the current will be proportional to the square root of the rotation speed. THE UNBLEACHED CASE At the other extreme the solution may be hardly bleached at all, Then y is small, and fly) + 1 in eqn (25). Substitution of eqn (8) for b, in eqn ( 2 5 ) gives I iR I = nFzr; @I,, @= l(XD, w-l/X,) W-li2M (26) where @ describes the fraction of light transmitted by the neutral density filter andw. J. ALBERY et al. 2653 Table 2. Values of M for common radius ratios ~ ~ ~ ~ ~~~~ r3/r2 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.12 1.14 1.16 1.18 1.20 1.22 1.24 0.147 0.192 0.23 1 0.268 0.302 0.334 0.364 0.393 , 0.42 1 0.475 0.526 0.574 0.621 0.667 0.71 1 0.754 0.147 0.192 0.232 0.269 0.303 0.335 0.366 0.396 0.424 0.478 0.529 0.579 0.626 0.672 0.717 0.148 0.193 0.233 0.270 0.305 0.337 0.368 0.398 0.427 0.482 0.533 0.583 0.63 1 0.678 0.148 0.194 0.234 0.272 0.307 0.339 0.371 0.401 0.430 0.485 0.538 0.588 0.637 0.684 0.149 0.195 0.236 0.273 0.309 0.342 0.373 0.404 0.433 0.489 0.542 0.593 0.642 0.150 0.196 0.237 0.275 0.31 1 0.344 0.376 0.407 0.436 0.493 0.547 0.598 0.648 0.151 0.198 0.239 0.277 0.3 13 0.347 0.379 0.410 0.440 0.497 0.551 0.603 0.152 0.199 0.241 0.279 0.3 15 0.350 0.382 0.413 0.444 0.501 0.556 0.609 0.I53 0.200 0.243 0.282 0.3 18 0.352 0.385 0.417 0.447 0.506 0.561 0.154 0.202 0.245 0.284 0.320 0.355 0.388 0.420 0.45 1 0.510 0.566 W is the rotation speed. In eqn (26) M, the detection efficiency, is a function only of the radii and is given by The ring current is M times the disc current that would be observed using an ORDE under these conditions. The fraction of incoming photons collected is given by X,/X,; the photons have to be absorbed in the diffusion layer for the product to have a chance of reaching the electrode. The geometric factor M is similar to the collection efficiency for a ring-disc electrode. Eqn (26) predicts that the ring current will vary with W-1/2; if the rotation speed is slower, the diffusion layer becomes thicker, more product reaches the electrode and the current is larger.Values of M calculated from eqn (27) for common radius ratios are collected in table 2. The theory was derived for the thin-ring-thin-gap geometry. The series cease to converge when and this provides a limit to the results in table 2. An interesting feature of the results in table 2 is that M does not vary much with the thickness of the gap given by r 2 / r 1 . This is in marked contrast to collection at an ordinary ringdisc electrode where the collection efficiency, N , decreases significantly as the width of the gap increases. We consider that the reason for the difference is that for the ring-disc electrode material is only generated on the electrode, and the bulk of the solution is an efficient sink for this material.For the RODRE the material is generated well out into the solution; only the small fraction of the material photogenerated in the diffusion layer finds the ring electrode; however, the rest of the material flowing past means that the outside edge of the diffusion layer is not such an efficient sink.2654 RING-DISC ELECTRODES When the system is neither completely bleached nor completely unbleached we write where b, is given by eqn (6), fly) by eqn (14), values of M are in table 2 and EXPERIMENTAL APPARATUS The apparatus for these experiments is similar to the d.c. apparatus described previ~usly.~ The photoelectrode, model AFDPOUQTPT, was supplied by Pine Instruments and has been described in detail by J0hnson.l For experiments with high light intensity the electrode supplied by the manufacturers was modified by inserting into the hollow electrode cavity a 12 in.quartz rod to act as a waveguide. The light was focussed onto the top of the rod and a layer of glycerol was used to make optical contact between the rod and the electrode disc. This arrangement increased the irradiance tenfold. The light source was the 180-W tungsten-halogen lamp of the side-illumination accessory of an Aminco DW2a spectrophotometer. This accessory is fitted with a rear parabolic mirror and associated optics, which were used for rough collimation of the light beam or to focus it onto the top of the quartz-rod waveguide. The excitation wavelength was selected with a three-cavity interference bandpass filter of 600 nm nominal wavelength supplied by Ditric Optics.The irradiance was varied using a set of neutral density filters, supplied by the same company, with values of @ of 0.75,0.53,0.37 and 0.14. The electrode assembly was rotated by an ASRPD Pine Instruments variable-speed rotator and an ASR motor controller. The jacketted cell and the reference (SCE) and counter- (Pt) electrodes were supplied by Astra Scientific Co. The potential of the rotating electrode was controlled by a M173 PAR potentiostat/galvanostat and the currents converted to voltages with a M 176 PAR current follower. The cell was thermostatted at 25 "C with a Haake low-temperature circulating bath. CHEMICALS AND SOLUTIONS All water was deionized, doubly distilled and passed through a four-bowl Milli-Q Super-Q millipore water-purification system.Acids were of Ultrex reagent grade supplied by the Baker Chemical Co. Thionine was supplied by Allied Chemical Corp. as the hydrochloride and purified by a series of recrystallisations from dilute HCl and aqueous n-butyl alcohol. The material was recrystallised until it gave a single spot on a thin-layer chromatogram (Analytech 0.25 mm normal phase silica gel, 10 x 20 cm plates, n-propyl alcohol-concentrated ammonia, 2: 1 v/v). The identity of the recrystallised thionine was confirmed at the high-field n.m.r. spectroscopy facility of Purdue University (470 MHz, [2H,]DMSO). Concentrations of thionine were found from the electronic absorption spectra of solutions diluted to ca. 1 pmol dm-3 (A,,, = 598 nm, E = 5.4 x lo4 dm3 mol-1 cm-l, 50 mmol dm-3 H,SO,).The experiments were carried out with 15-70 pmol dmb3 thionine in 50 mmol dmP3 aqueous H,SO, ; the concentration of the quencher, iron(rr1) ammonium sulphate (Fluka Chemical Co.), was > 5 mmol dm-3. This concentration of quencher is enough to trap all the triplet thionine. With no added FelI1 the maximum concentration of photogenerated FeIII is 140 pmol dm-3. Using the known value of the rate constant for the reaction of leucothionine and Fe"18 we estimate the lifetime of leucothionine to be > 10 s. Therefore, the loss of leucothionine in its passage from the disc to the ring is insignificant. Deoxygenation of the thionine-ironfm) solutions was effected by bubbling ultra-high-purity argon supplied by General Gases. MEASUREMENT OF IRRADIANCE AND PHOTOCURRENTS The flux of photons passing through the disc electrode was measured with a Spectra-Physics laser power meter (model 404, 450-900nm).The light was shone through the transparent electrode onto the collector. The irradiance, I,/mol ern+ s-l, was calculated from I, = F'/L hv A ,w. J. ALBERY et al. 2655 0.0 Fig. 3. Ring currents plotted according to eqn (26) against @/ W1/2 where @ is the transmittance of the neutral density filler and W is the rotation speed of the electrode. Values of W in Hz are as follows: a, 1; ., 2; A? 4; +, 6; V, 8; 0, 12; 0, 16; A, 25; x , 36 and V? 49. The values of @ were 1.0, 0.73, 0.53, 0.37 and 0.13. were F' is the power-meter reading in W and A , is the area of the collector exposed to the light (0.143 cm2).The electrode radii were measured with a travelling microscope and were found to be in good agreement with the manufacturers' specifications (rl = 0.501 cm, r2 = 0.507 cm and Y, = 0.625 cm). The ring currents were recorded on a bright Pt electrode (cleaned with concentrated HNO, and concentrated H2S04) using an 8120 Bascom-Turner digital recorder. They are the average of 500 readings with a sampling interval of 50-100 ms per point. The Pt ring was held at the potential of zero dark current (280-300mV us SCE). At this potential we find that a thionine-coated electrode reduces leucothionine but that there is negligible current from the photogenerated FelI1. RESULTS AND DISCUSSION We start with a set of results obtained using a relatively low irradiance where the RODRE did not have the waveguide inserted.Under these conditions there is no bleaching and so the ring current should be described by eqn (26). Fig. 3 shows a set of 50 results obtained at 5 different irradiances and 10 different rotation speeds. In accordance with eqn (26) the ring current has been plotted against OW-;. A good straight line is obtained. In comparing the gradient of the line in fig. 3 with that calculated from eqn (26) there is a problem with the value of I, to be used. As found by Johnson and coworkers,2 the irradiance is not uniform under the disc. Typical results for our electrode show that the irradiance at the edge of the disc is some 30% less than that at the centre. This result is similar to that found by Johnsod and coworkers.2 The ring electrode is more affected by the situation on the outside edge2656 RING-DISC ELECTRODES Table 3.Calculation of gradient in fig. 3 rl = 0.501 cm q~ = 0.27 lo, + = 0.73 nmol cm-2 s-l D = 5.8 cm2 Ms-l am = 70 pmol dm-3 E = 0.129 pmol-l dm3 cm-' M = 0.742 gradient [eqn (26)] = 1.10 pA Hz'/~ gradient (fig. 3) = 1.26 pA Hz112 -1 I I I -3 -2 -1 log [ (@/W'i2)/Hz-'i2 1 0 Fig. 4. Ring currents plotted according to eqn (29) against log (@/ W ) , where from eqn (3 1) y a @/ W for four different concentrations of thionine. At the higher irradiances and lower rotation speeds the solution becomes bleached. The thionine concentrations in pmol dm-3 were as follows: D, 65; .,45; 0, 29 and 0, 15. of the disc than by that at the centre. Therefore, we have used the value of the irradiance found at the outside of the disc.In table 3 we compare the calculated value of the gradient from eqn (26) with that found from fig. 3. The values agree to within 7%. This is reasonable agreement given the uncertainties about the uniformity of the irradiance. Next we turn to the experiments with the waveguide inserted into the RODRE. The irradiance is now large enough for the solution to be partly bleached, and we havew. J. ALBERY et al. 2657 to use the full form of eqn (25) as rearranged in eqn (28). Experiments were carried out at 4 different concentrations of thionine, 5 different irradiances and 7 different rotation speeds. The results of the 140 experiments are plotted in fig. 4. Using eqn (6) and (14) we have further rearranged eqn (28) to give where and from eqn (7) The function g(y) depends only on the geometry and the dimensionless parameter y, which describes the balance between bleaching and convective dilution.When y is small, corresponding to no bleaching, g(y) x M ; on the other hand when y is large for the bleached case g(y) x P/”y. Experimental data are plotted according to eqn (29) in fig. 4, where from eqn (31) y varies with @/W. Good agreement is found with theoretical curves calculated from eqn (29). In particular, the vertical separation between the curves for the different concentrations is correctly described by the log a, term, and as required the bleaching becomes significant for each concentration at the same value of y. We conclude that these experiments show that our theoretical description of the RODRE is correct.The experimental part of this work was sponsored by the U.S. Department of Energy, Office of Basic Energy Sciences under contract no. DE-AS05-82ER12088. We thank the U.S. National Science Foundation, the S.E.R.C. and the Atlantic Richfield Foundation for further financial support, and we acknowledge the assistance of the Biological Magnetic Resonance Facility of Purdue University. APPENDIX In this appendix we solve eqn (1) for the concentration of B in the zone of the disc. We first assume that X, is sufficiently large for the exponential term to equal unity throughout. We normalise the distance x with the Levich diffusion length,” X,, defined in eqn (9): cy = X/X,. v, = ex2 In the convective term5 where C = 8.03 W/2v-1/2.We also write Then eqn (1) becomes u = b/a,. where y is defined in eqn (7). u + 0 as cy -, 00 gives Outside the diffusion layer, ty > 1, we ignore the diffusion term in eqn (A 1); integration with (A 2) In (1 - u) x - y/21y. In particular, at the edge of the diffusion layer where cy = 1, u1 x l-exp(-y/2) 872658 RING-DISC ELECTRODES Inside the diffusion layer, ty < 1, we ignore the convective term in eqn (A 1); integration, with the boundary condition au/i3x = 0 at ty = 0, gives 1 - u = (1 - ul) cash (y1/3 ty)/cosh (y”‘). (A 4) Substitution of eqn (A 3) in eqn (A 4) gives the result at y = 0 for uo = bo/a, in eqn (6). 1 2 3 4 J 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 D. C. Johnson and E. W. Resnic, Anal. Chem., 1972, 44, 637. J. R. Lubbers, E. W. Resnick, P. R. Gaines and D. C. Johnson, Anal. Chem., 1974,46, 865. P. R. Gaines, V. E. Peacock and D. C. Johnson, Anal. Chem., 1975,47, 1393. W. J. Albery, M. D. Archer, N. J. Field and A. D. Turner, Faraday Discuss. Chem. Soc., 1973, 56, 28. W. J. Albery, M. D. Archer and R. G. Egdell, J. Electroanal. Chem., 1977, 82, 199. W. J. Albery, W. R. Bowen, F. S. Fisher and A. D. Turner, J. Electroanal. Chem., 1980, 107, 1. W. J. Albery, W. R. Bowen, F. S. Fisher and A. D. Turner, J. Electroanal. Chem., 1980, 107, 11. W. J. Albery, P. N. Bartlett, W. R. Bowen, F. S. Fisher and A. W. Foulds, J. Electroanal. Chem., 1980, 107, 23. S. W. Feldberg, in Electroanalytical Chemistry, ed. A. J. Bard (Marcel Dekker, New York, 1969), vol. 3. K. B. Prater and A. J. Bard, J. Electrochem. Soc., 1970, 117, 207. V. G. Levich, Physicochemical Hydrodynamics (Prentice Hall, Englewood Cliffs, N.J., 1962), W. J. Albery and M. L. Hitchman, Ring-Disc Electrodes (Clarendon Press, Oxford, 1971), p. 15. W. J. Albery, B. A. Coles and A. M. Couper, J. Electroanal. Chem., 1975, 65, 901. W. J. Albery, R. G. Compton and A. R. Hillman, J. Chem. Soc., Faraday Trans. I , 1978,74, 1007. G. E. Roberts and H. Kaufman, Tables of Laplace Transforms (W. B. Saunders, New York, 1966), p. 22. M. Abramowitz and I. Stegun, Handbook of Mathematical Functions (Dover, New York, 1965), p. 260. M. Abramowitz and I. Stegun, Handbook of Mathematical Functions (Dover, New York, 1965), p. 448. M. Abramowitz and I. Stegun, Handbook of Mathematical Functions (Dover, New York, 1965), p. 263. M. Abramowitz and I. Stegun, Handbook of Mathematical Functions (Dover, New York, 1965), p. 944. W. J. Albery, S. Bruckenstein and D. T. Napp, Trans. Faraday Soc., 1966, 62, 1932. pp. 63-69. (PAPER 4/ 1590)

ISSN:0300-9599    

DOI:10.1039/F19858102647

出版商:RSC    

年代:1985

数据来源: RSC