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/F198581FX005

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/F198581BX007

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

JOURNAL OF T H E CHEMICAL SOCIETY FARADAY TRANSACTIONS, PARTS 1 AND I 1 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-p hase 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 Transactions) 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, or 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 of 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 I500 words or word-equivalents. (9NOMENCLATURE AND SYMBOLISM 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 ‘ Systkme 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 1, 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.(ii)THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY GENERAL DISCUSSION NO. 79 (in conjunction with the Polymer Physics Group) Polymer Liquid Crystals University of Cambridge, 1-3 April 1985 The object of the meeting will be to discuss all aspects of the developing subject of polymeric liquid crystals. The hope is to bring together scientists from the fields of conventional polymer science and monQmeric liquid crystals who are active in this field. The discussion is aimed at understanding the following facets: (a) The chemical characteristics that give rise to polymer liquid crystalline behaviour. (b) The nature of the high local anisotropy of these systems and their structural organisation at the molecular, micron and macroscopic levels.(c) The physical properties and their industrial exploitation, with particular reference to the influence of external force fields such as flow, electric and magnetic fields. (d) The inter-relations of polymer liquid crystals with small-molecule mesophases, conventional flexible polymers and biopolymers which exhibit liquid-crystalline behaviour. The 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. 80 Physical Interactions and Energy Exchange at the Gas-Solid Interface McMaster University, Hamilton, Ontario, Canada, 23-25 July 1985 Organising Committee : Professor J. A. Morrison (Chairman) Dr M.L. Klein Professor G. Scoles Professor W. A. Steele Professor F. S. Stone Dr R. K. Thomas The discussion will be concerned with certain aspects of current research on the gas-solid interface: elastic, inelastic and dissipative scattering of atoms and molecules from crystal surfaces, and the structure and dynamics of physisorbed species, including overlayers. Emphasis will be placed on the themes of physical interactions and energy exchange rather than on molecular- beam technology or the phenomenology of phase transitions on overlayers. The interplay between theory and experiment will be stressed as they relate to the nature of atom and molecule surface interaction potentials, including many-body effects. The preliminary programme may be obtained from : Professor J.A. Morrison, Institute for Materials Research, McMaster University, Hamilton, Ontario, Canada L8S 4M1 or: Mrs Y. A. Fish, The Royal Society of Chemistry, Burlington House, London W1V OBN, U.K. (iii)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 THE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY ~ GENERAL Dl-SCUSSION NO. 81 The aim of the meeting is to discuss phase transitions at gashiquid, 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 preliminary programme may be obtained from : Mrs Y. A. Fish, The Royal Society of Chemistry, Burlington House, London W1 V OBN ~ Lipid Vesicles and Membranes ~ Loughborough University of Technology, 15-1 7 April 1986 ~ Full papers for publication in the Discussion Volume will be required by December 1985. 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. Contributions for consideration by the Organising Committee are invited and abstracts of about 300 words should be sent by 1 May 1985 to: Professor D.A. Haydon, Physiological Laboratory, Downing Street, Cambridge CB2 3EGTHE FARADAY DIVISION OF THE ROYAL SOCIETY OF CHEMISTRY GENERAL DISCUSSION NO. 82 Professor R. N. Dixon, Department of Theoretical Chemistry, University of Bristol, ~ Cantock's Close, Bristol BS8 1TS Dynamics of Molecular Photof ragmentation University of Bristol, 15-1 7 September 1986 Organising Committee : Professor R.N. Dixon (Chairman) Dr G. G. Balint-Kurti Dr M. S. Child Professor R. Donovan Professor J. P. Simons 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. Contributions for consideration by the Organising Committee are invited. Titles should be submitted as soon as possible, and abstracts of about 300 words by 30 September 1985, to:FARADAY DIVISION INFORMAL AND GROUP MEETINGS Theoretical Chemistry Group Group Meeting To be held at King's College, London on 6 March 1985 Further information from Dr G.Doggett, Department of Chemistry, University of York, York YO1 5DD Division Annual Congress: Solid State Chemistry To be held at the University of St Andrews on 25-28 March 1985 Further information from Professor P. A. H. Wyatt, Department of Chemistry, University of St Andrews, The Purdie Building, St Andrews KY16 9ST Electrochemistry Group Spring Informal Meeting To be held at Middlesex Polytechnic, London on 1-3 April 1985 Further information from Dr F. L. Tye, Middlesex Polytechnic, Bounds Green Road, London N11 2NQ Polymer Physics Group 6th Churchill Conference To be held at Churchill College, Cambridge on 1-4 April 1985 Further information from Professor I. M. Ward, Department of Physics, University of Leeds, Leeds LS2 9JT Statistical Mechanics and Thermodynamics Group Dense Fluids: Dynamic and Static Properties To be held at the University of Bristol on 10-1 1 April 1985 Further information from Dr D.J. Tildesley, Department of Chemistry, The University, Southampton SO9 5NH Neutron Scattering Group Small-angle Neutron Scattering from Organised Systems To be held at Imperial College, London on 17-1 8 April 1985 Further information from Dr R. W. Richards, Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow G1 1 XL Gas Kinetics Group with SERC Summer School in Gas Kinetics To be held at the University of Cambridge on 26 June to 3 July 1985 Further information from Dr I. W. M. Smith, Department of Chemistry, University Chemical Laboratory, Lensfield Road, Cambridge CB2 1 EP Industrial Physical Chemistry Group with the Food Chemistry Group Water Activity: A Credible Measure of Technological Performance and Physiological Via bi I i ty To be held at Girton College, Cambridge on 1-3 July 1985 Further information from Professor F.Franks, Department of Botany, Downing Street, Cambridge CB2 3EA Polymer Physics Group Biennial Conference To be held at the University of Reading on 11-1 3 September 1985 Further information from Professor Bassett, J. J. Thompson Physical Chemistry Laboratory, University of Reading, Whiteknights, Reading RG6 2AF Carbon Group Strength and Structure in Carbons and Graphites To be held at the University of Liverpool on 16-1 8 September 1985 Further information from The Meetings Officer, The Institute of Physics, 47 Belgrave Square, London SW1X 8QXSurface Reactivity and Catalysis Group with the Catalysis Section of the KNCV Mechanism and Structure in Heterogeneous Catalysis To be held at Noordwijkerhout, The Netherlands on 18-20 September 1985 Further information from: Dr R.Joyner, BP Research Centre, Chertsey Road, Sunbury on Thames TW16 7LN Industrial Physical Chemistry Group A Molecular Approach to Lubrication and Wear To be held at Girton College, Cambridge on 23-25 September 1985 Further information from Mr M. P. Dare-Edwards, Shell Research Ltd, Thornton Research Centre, Chester CH1 3SH Neutron Scattering Group jointly with the Materials Testing Group of the Institute of Physics Industrial Uses of Particle Beams To be held at the Institute of Physics, London on 26 September 1985 further information from Dr J.G. Booth, Department of Chemistry, University of Salford, Salford M5 4 W 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 30TH INTERNATIONAL CONGRESS OF PURE AND APPLIED CHEMISTRY Advances in Physical and Theoretical Chemistry Manchester, 9-1 3 September 1985 The Faraday Division is mounting the following symposia as part of the 30th IUPAC Congress: A. B. C. D. Reaction Dynamics in the Gas Phase and in Solution This symposium will examine the ways in which modern techniques allow detailed study of the dynamical motion of molecules which are undergoing chemical reaction or energy exchange. Micellar Systems The symposium will discuss various aspects of micellization, including size and shape factors, micellization in biological systems, chemical reactions in micellar systems, micelle structure and solubilization.Emphasis will also be given to modern techniques of examining micellar systems, including small-angle neutron scattering, neutron spin echo, photocorrelation spectroscopy, NM R and use of fluorescent probes. Surface Science of Solids The symposium will centre on recent advances in the study of kinetics and dynamics at surfaces and of phase transitions in adsorbate layers on single crystal surfaces. Both experimental and theoretical aspects will be reviewed with an emphasis on metal single crystal surfaces.New Electrochemical Sensors (in collaboration with the Electroanalytical Group of the Analytical Division) The symposium will cover such topics as the fundamentals of the subject, new gas sensors based on membrane electrodes and on ceramic oxides, the development of new ion- selective electrodes and the synthesis of new guest-host carriers, the development of CHEMFETS and other integrated devices together with the theory of the operation of such devices, and finally the development of biosensors including for instance enzyme electrodes, direct electron transfer to biological molecules and new potentiometric techniques for protein analysis. The second circular, giving details of all the symposia of the Congress and listing invited speakers may be obtained from: Dr J.F. Gibson, 30th IUPAC Congress, Royal Society of Chemistry, Burlington House, London W1 V OBN (vii)Arthur Adamson, Editor University of Southern California Arthur Hubbard, Associate Editor University of California at Santa Barbara This new journal published by the American Chemical Society fills the void existing in current literature available today-Langmuir’s broad coverage will bring together authoritative papers from all aspects of this major field of chemistry! Langmuir will include fundamental and applied papers covering ultra-high vacuum surface chemistry and spectroscopy, heterogeneous catalysis, all aspects of interface chemistry involving fluids, and disperse systems.Specifically, Langmuir will publish peer-reviewed research in r/ ‘Wet’ Surface Chemistry surface tension 0 spread monolayers 0 wetting and contact angle 0 adsorption from solution 0 nucleation and fundamental aspects of flotation, detergency, emulsions, foams, lubrication, etc. r/ Electrochemistry related to interfacial structure and processes r/ ‘UHV’ Surface Chemistry solid surfaces in ultra-high vacuum including surface structure, composition and spectroscopy fundamental papers in heterogeneous catalysis colloidal suspensions including aerosols 0 microemulsions 0 biological and polymeric colloids 0 and membrane systems 4 Disperse Systems In bimonthly issues of Langmuir, you will find experimental and theoretical original papers, letters to the editor, and book reviews, as well as some selected symposium collections. Papers having applied aspects will be included.And, published by the American Chemical Society, Langmuir will benefit from the Society’s vast international network of scientists and editorial resources. Note to Authors: Langmuir will not have page charges. Editorial Advisory Board N.R. Armstrong, Unrv. of Arizona 0 G.T. Barnes, Univ. of Oueensland. AUSTRALIA P. Biloen, Univ. of Pittsburgh 0 K.S. Birdi, Technical University of Denmark, DENMARK A.M Bond. Deakm University, AUSTRALlA B.V. Derjaguin, Academy of Science of USSR D.D. Eley, Univ. of Nottingham. ENGLAND G Ertl. Univ. of Munich, GERMANY 0 J . Fendler, Clarkson College of Technology T . Fort, Jr., California Polytechnic State Unw. G. Gaines. Jr.. General Nectric 0 W.A. Goddard, 111, California lnstitute of Technology R.S. Hansen. lowa State Univ. 0 J. Lyklerna. Agricultural Univ., THE NETHERLANDS R.J. Madix, Stanford Unrv. J.A. Mann. Jr., Case Western Reserve Univ. P. Mukerjee, Univ. of Wisconsin K.J. Mysels, Research Consulting A W . Neumann, Unw. of Toronto. CANADA R. Ottewill, Univ. of Bristol, ENGLAND G.D Parfitt, Carnegie-Mellon Univ. H . Reiss, Unrv of Californ/a at Los Angeles H.A. Resing, Naval Research Laboratory T . Rhodin. Cornell Univ. S Ross, Rensselaer Polytechnic Univ. J. Rouquerol, Centre de Thermodynamique et de Microcalorimetrie du CNRS. FRANCE R.L. Rowell, Univ. of Massachusefts 0 R. Rye, Sandia National Lab H. Seki, l8M K. Shinoda, Yokohama National Univ.. JAPAN G.A. Sornorjai, Univ. of California at 8erkley W A Steele, Pennsylvania State Unrv. q$J+ Subscription Information 1985 Foreign Rates (Includes Air Service) ACSMembers $ 56 (Personal Use) Nonmembers $308 January-February 1985 Volume 1 No 1 One Volume Per Year (SIX Issues) ISSN: 0743-7463 Cable Address. JIECHEM Telex. 440159 ACSPUI or 892582 ACS PUBS American Chemical Society 0 1155 Sixteenth St., N.W. 0 Washington, D.C. 20036 (viii)

ISSN:0300-9599    

DOI:10.1039/F198581FP017

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985, 81, 273-283 Structural Aspects of Certain Phase Transformations in Lyotropic Liquid-crystal Systems B Y RAYMOND M. WOOD* AND MALCOLM P. MCDONALD Departments of Physics and Chemistry, Sheffield City Polytechnic, Pond Street, Sheffield S1 1WB Received 20th December, 1983 Aqueous binary systems of potassium oleate and sodium dodecyl sulphate exhibit an hexagonal phase at low surfactant concentrations and appropriate temperatures. However, in the former system the phase is succeeded at higher surfactant concentrations by the so-called rectangular phase, while in the latter system a distorted hexagonal phase follows. X-ray diffraction results for these systems indicate that the rectangular phase is a consequence of axial growth of the hexagonal-phase surfactant cylinders with surfactant content at substantially constant radius.Water in the hexagonal phase occupies positions other than along and between cylinder lengths. Radial growth of cylinders leads to the distorted hexagonal structure. The cylinder growth and transformation behaviours are associated with characteristics of the respective molecular structures. When the composition of a single-phase crystalline substance is changed, the change in the size of the repeat unit of the structure is three dimensional. In contrast, a change in the composition of a single liquid-crystal phase results in an equivalent dimension change which may be one, two or three dimensional, according to the number and chemical form of the constituents. The importance of this latter range of behaviour lies in its relationship to the mechanism of the structural change which takes place when that phase undergoes a transformation and also to the subsequent phase structure.Thus a study of the way in which liquid-crystal phase structures change dimensions with composition is a prerequisite for an understanding of the physical basis of structural transformations in these substances. In this work, we have examined the effect on phase structure and transformation behaviour of two types of anionic amphiphilic molecule with different structures, namely sodium dodecyl sulphate (SDS) and potassium oleate (KO). Fig. 1 gives details of the published data for the liquid-crystal phases in the binary systems SDS + water and KO + water.In the case of the SDS+water system, Husson et a1.l claim that the hexagonal phase is followed at higher amphiphile concentrations by a complex hexagonal phase and finally a lamellar phase. The results of Bergeron2 do not contain any observation of a phase between the hexagonal and lamellar ones. For the KO + water system there is agreement that increasing the amphiphile content results in a set of phases which are hexagonal, rectangular, complex hexagonal and lamellar, in that order,lT although Ekwall et aL4 reported only a rectangular phase in the intermediate region. EXPERIMENTAL Parts of the binary systems SDS +water and KO + water have been examined by polarised light5 and low-angle X-ray diffraction. The SDS (specially pure) was obtained from B.D.H. and was recrystallised from ethanol, while the water was deionised and doubly distilled from alkaline 273274 SDS AND KO LIQUID-CRYSTAL PHASES I H u s s o n ef a l , 348 K 1 1 1 HJsson ei a l., 293 K B a l m b r a etal.,I 293 K I I I I I I I I I 0.2 0.3 0 .4 0 . 5 0.6 0.7 0.8 0.9 1.0 Fig. 1. Relevant phase structures and their compositions in the binary SDS+water and KO+water systems as observed by previous workers. -, Hexagonal; w, rectangular; \, complex hexagonal; /, lamellar; 0, two-phase. surfactant (mass fraction) potassium permanganate solution. Mixtures were obtained by repeatedly centrifuging the weighed constituents through a constriction in a sealed glass tube. Potassium oleate was prepared by dissolving redistilled oleic acid in ethanol and titrating the solution with potassium hydroxide to pH 9.The resulting solution was freeze-dried and the solid so produced was dried over phosphorus pentoxide. Mixtures of KO and water in selected proportions were prepared in the same manner as the SDS + water mixtures. X-Ray exposures were carried out using a Warhus low-angle camera with a specimen-to-film distance of 255 mm and filtered copper Ka radiation. For this purpose, samples were sealed either inside 0.5 mm Lindemann tubes or inside a disc-shaped cell of thickness 1 mm with Mylar windows. Sample temperatures were monitored during exposures (usually 16 h) by means of a thermocouple mounted in the sample heating chamber. Diffraction line positions on films were measured using a Hilger and Watts film measuring device with a precision of 0.05 mm.Specimens subjected to the diffraction technique were examined visually after exposure for deterioration and/or capsule failure. Any indication of crystal formation or capsule leakage led to rejection of the diffraction pattern. Since three-dimensional crystal structures generate diffracted beams of higher intensity than those of two-dimensionally ordered liquids, the absence of identifiable crystal diffractions was taken to indicate zero or insignificantly small quantities of crystalline SDS. RESULTS In association with Tiddy and coworkers we have published results for the SDS+ water system which show that the hexagonal structure becomes d i ~ t o r t e d . ~ This behaviour is to be compared with that indicated by the X-ray results for the KO + water system in fig.2, which shows that, in agreement with the other workers,'! we have obtained hexagonal, rectangular and complex hexagonal phases. There is only a narrow two-phase region between the hexagonal and rectangular single phases which contrasts with the broad region where both hexagonal and distorted hexagonal phases occurred in the SDS system. Such two-phase regions, being inherent parts of binary phase diagrams, must, by their absence, introduce doubt about the results of Husson et aI.,l which, as shown in fig. 1, contain no two-phase regions. In contrast, the phasesR. M. WOOD AND M. P. MCDONALD 275 1070 H 010 R 3 . 6 I I I I 1 0.30 0.34 0.38 0.42 0.46 0.50 0.54 0.58 0.62 0.66 0.70 KO (mass fraction) Fig. 2. X-Ray diffraction observations of phases in the KO+water binary system at room temperature.H, Hexagonal ; R, rectangular; H, complex hexagonal. and phase boundaries presented by Balmbra et are little different from those of the present work, and Balmbra et al. acknowledged the difficulty of establishing the precise locations of phase boundaries. The identification of the complex hexagonal phase in the work of Balmbra et al. rests on electron microscopy and a single diffraction line, designated as a reflection from (1 120) planes with a spacing of 5.60 nm. We have obtained a diffraction line from planes of spacing 5.70 nm at compositions similar to those of Balmbra et al. and accordingly have designated this a (1 120) reflection from a complex hexagonal structure. The effect of composition and temperature on the edge length of the rhomboid repeat unit of the hexagonal distribution of right circular SDS cylinders in the 'middle' phase of the SDS+water system is shown in fig.3. Plotted points are the means of values of the rhombus parameter calculated from two to four X-ray diffraction lines appearing on several films at each indicated composition and the probable error associated with each mean is presented as an error bar. Of interest are the two points which deviate from the best straight line for the results at 3 13 K. Their probable errors indicate major compositional errors or a real effect as the cause of the deviations. In view of the fact that the points closely follow the pattern of the deuterium n.m.r. results at this temperat~re,~ which was interpreted as indicating a change in the water environment, it is probable that they are similarly involved in an incipient transformation. Apart from the actual values for the size of the rhombus, the two main features of these results are the linear relationship between rhombus size and composition of the sample and the varying effect of temperature on the magnitute of this linear relationship, the slopes of the lines going through a minimum between 303 and 333 K.Fig. 4 shows the variation of the same parameter with composition for the potassium oleate + water system. Presented in fig. 4 are published results for the hexagonal phase at room temperature together with results for four compositions obtained in the present investigation. The data obtained from X-ray diffraction studies by Ekwall276 SDS AND KO LIQUID-CRYSTAL PHASES 5.2 5.0 4 .8 4 * 6 5.0 4 . 8 4 . 6 4 . 4 5 .O 4 . 8 G \ 4 .8 4 . 6 4 . 4 0.30 0.40 0.50 0.60 Fig. 3. Dependence of unit-cell dimensions on composition for the hexagonal phase in the SDS+water binary system at (a) 303, (b) 313, (c) 323 and (d) 333 K. SDS (mass fraction) et aI.,* Balmbra et aL3 and the present investigation are in good agreement, while the electron microscopy values of Eins6 follow a similar trend with composition, although slightly displaced to smaller rhombus dimensions. A significant difference between the hexagonal phases in the two systems, as shown by fig. 3 and 4, is the manner in which their basic dimension varies with composition, i.e. linear for SDS compared with a more complicated relationship for KO.It has been pointed out7 that the spacing, d, of the (1010) Bragg reflection for the hexagonal phase will be related to the mass fraction, c, of surfactant present by a negative half-power relationship (i.e. dlolo a c - O . ~ ) if the surfactant cylinder radius is invariant with composition. The relationship between surfactant cylinder dimensions and the dimensions of the rhomboid prism of the hexagonal phase is more fully given by (1) 7 t ~ ~ L ~ [ ( c - ~ - 1) v;' + 11 = L, a2 sin 60" when the partial specific volume of water is taken as unity. In eqn (l), L, and L, are the lengths of the surfactant cylinder and rhomboid prism, respectively, c is the mass fraction of surfactant which has partial specific volume v, and a is the rhombus edge length.When I,, and L, are assumed equal and v, is assumed to have the same value as that for water, the expression reduces to a GC c - O . ~ . Deviatiom from this relationship have been suggested7 to indicate a cylinder radius changing with surfactant concentration.R. M. WOOD AND M. P. MCDONALD 277 9 .o 8 . 6 8 . 2 7.8 7.4 7.0 6.6 2 \ ct 6.2 5 . 8 5 . 4 5.0 I I I I I I 0.20 0.30 0.40 0.50 Fig. 4. Dependence of unit-cell dimension on composition of hexagonal phase in the KO + water binary system at room temperature. 0, Balmbra et A, Ekwall et a1.;4 +, Eins;6 a, this work. KO (mass fraction) SDS +WATER The clearly linear variation of rhombus dimension with composition (fig. 3) indicates a marked deviation from the relationship a K c - O - ~ with the ensuing implication of a major dependence of cylinder radius on SDS content, The variation displayed in fig.3 will be expressed in the form A = A,-k(c-c,) (2) where A is the rhombus area at any SDS mass fraction c and A , is the rhombus area at an arbitrary mass fraction c, within the hexagonal-phase region. The gradient of the graph of area against mass fraction is k. An assumption of equality of specific volumes of surfactant and water leads to the equality of mass and volume fractions of surfactant (and of water). Since the water surrounding a cylinder will have approximately the same length as the cylinder itself, the fraction of rhombus area occupied by surfactant cylinder and by water will be the same as their volume fractions and hence their mass fractions, i.e.A , = CAR r2 = [(A, + kc,) c - kc2]/n. where A , = cylinder area = nr2. Eqn (1) now takes the form (3)278 1.90 1.85- 1.80- E 1.75 2 -a .- 2 1.70 4 -0 1.65 1.60 1.55 5. - - - - - SDS AND KO LIQUID-CRYSTAL PHASES 1.50 L I I I I 0.30 0 .GO 0 S O 0.60 0.70 SDS (mass fraction) Fig. 5. Plot of hexagonal-phase cylinder radius dependence on composition in SDS +water binary system derived from X-ray diffraction measurements of unit-cell dimensions. Arrows indicate minimum SDS contents at which the distorted hexagonal phase was observed. V, 333; a, 323; A, 313; 0, 303 K. It can be shown that on introducing the partial specific volumes of surfactant and water as v, and vH, respectively, eqn (3) becomes Using the Husson et al.l value for v,, namely 0.92 cm3 g-l for SDS and a mass fraction of 0.50, r is 2.2% smaller than the approximate value given by eqn (3), which permits the specific volume difference to be ignored. The data contained in fig.3 for the hexagonal phase of the SDS+water system at four temperatures have been analysed using eqn (3). Four resulting curves showing the variation of cylinder radius with composition appear in fig. 5 and each curve is marked to indicate the lowest composition at which the additional X-ray spacing of distorted hexagonal phase were observed in the present investigation. KO + WATER The non-linear relationship between size and composition shown by the data of fig. 4 for the hexagonal phase is well described for the Ekwall’s results4 by a oc c--O.~~, which is similar to the relationship for invariant cylinder radius, namely a cc c - O .~ . Of the data available, these results were selected for analysis because they most closely agreed with the measurements obtained in the present investigation. However, as will be apparent, it is departure from the above negative half-power relationship which is the vital factor in defining the growth behaviour of hexagonal-phase cylinders. The above dependence is in distinct contrast with that of the SDS hexagonal phase and implies a relatively small change of surfactant cylinder radius with composition.R. M. WOOD AND M. P. MCDONALD 279 However, if the water contribution to the rhombus area remains the same irrespective of whether a cylinder radius changes or not (i.e. the partial specific volume of the water is unaffected by the mixing), it would be necessary for the cylinder to contract in radius as the surfactant mass concentration increases in order to obtain an increase in the power of c from -0.5 to -0.44.Since the volume fraction of surfactant increases with its mass fraction, cylinder contraction is most unlikely and the cylinder radius probably remains essentially constant. The variation of rhombus edge length, a, with mass fraction, c, of KO may be described by the general equation a = mc-p (4) and eqn (I), on simplification by taking v, M 1, becomes L,/LR = (ca2/nr2) sin60" so that L,/LR = (m2 sin 6Oo/nr2) F 2 P . ( 5 ) Applying eqn ( 5 ) to the ideal situation of invariant cylinder radius and L, = LR leads to r = m(sin6O0/n)i = 0.525m.Hence eqn (4) can be interpreted as presenting the size of the invariant cylinder radius by the term m and the variation of the ratio LJL, by the index p . Ekwall's data appearing in fig. 4 can best be descnbed by the equation a = 4.176 c-0.443 obtained from a least-squares fit to the data in logarithmic form. This expression then provides a value of 2.19 nm for the surfactant cylinder radius. DISCUSSION The results presented here for the systems SDS+water and KO+water are of special interest for their indications of contrasting dependence of phase structure on composition. Although both systems contain a hexagonal phase followed at higher surfactant content by a non-hexagonal phase, these phases in the two systems have their own particular characteristics which appear to be determined by the lengths and structures of the hydrocarbon chains in the different molecules.Of striking contrast is the dependence of the hexagonal phase rhombus area for the two systems on surfactant content. In the case of the SDS+water system this dependence is a manifestation of a continuously increasing surfactant cylinder radius, as shown in fig. 5 where this parameter is plotted against composition for four temperatures. Note that at 313, 323 and 333 K the onset of a two-phase region was observed at almost identical values of the cylinder radius. It is possible that the two-phase region for the 303 K results would have been observed at lower SDS concentration had specimens containing between 0.50 and 0.55 mass fraction of SDS been examined.As it is, there is clear indication in fig. 5 that the hexagonal phase shows signs of changing its structure when the surfactant cylinder reaches a radius of ca. 1.77 nm. This value compares well with the length of an SDS molecule, 1.76 nm to the centre of the sulphur atom, obtained when the hydrocarbon chain is in the fully extended trans configuration. It is suggested that the hexagonal-phase structure in the SDS +water system is changed with increased surfactant content by the additional amphiphile molecules taking up positions radially within existing layers of the280 SDS AND KO LIQUID-CRYSTAL PHASES cylinders and so increasing their molecular content. Such a process would require increased cylinder circumference to maintain equilibrium separation of the polar head groups of the molecules and hence increased cylinder radius.A concomitant of this behaviour would be the ‘straightening’ of the SDS hydrocarbon chains to ensure that the available cylinder cross-section remained filled. If the additional SDS molecules were to adopt positions which resulted in longer cylinders, a different relationship between cylinder radius and rhombus dimensions would result from the requirement to maintain approximately equal lengths of cylinder and surrounding water. Fig. 3 and 5 reveal that, as pointed out earlier, temperature effects on the SDS hexagonal- phase dimensions are irregular. The results of fig. 3 show that up to a mass fraction of SDS of cu. 0.55 and a maximum temperature of 323 K, the rhombus dimensions decreased with temperature.This observation is in agreement with that of Luzzati et aZ.8 for all the systems which they reported on and for which the explanation was that the decrease in dimension follows from a contraction of cylinder radius because of the increased effect of higher temperatures on hydrocarbon-chain motion. The present results (fig. 5) show that for the lower SDS concentrations this explanation is probably correct but at higher concentrations the cylinder radius increases with temperature up to ca. 328 K and then decreases again. This behaviour is consistent with the restricted opportunity for chain flexing so leading to thermal expansion, which must occur with the proposed ‘ straighter’ chain. The chain-motion effect will then dominate at the higher temperatures.When an SDS hydrocarbon chain approaches its fully extended length, an increase in amphiphile concentration should result in little or no increase in cylinder radius by further chain ‘ straightening’. However, an increase in cylinder circumference can be obtained if the circular cross-section becomes elliptical so that the central core of a cylinder remains filled by the hydrocarbon groups. It is therefore tempting to speculate here that the continued increase in cylinder radius with amphiphile concentration of the distorted hexagonal phase is accompanied by a continued increase in eccentricity of the cylinder cross-section until eventually the ellipses are so eccentric as to be indistinguishable from the well known bilayers of the lamellar phase. Hendrikx and Charvolin, examining the SDS + decanol+ water ~ystem,~ obtained a phase which they identified as having a two-dimensional centred rectangular structure.They claim that the molecular aggregates at the lattice points are ‘ribbon shaped’ (i.e. cross-section similar to that of a lath) and incorrectly interpret their qualitative X-ray diffraction intensities as supporting evidence. As Oster and Riley have indicatedlO and Gale and Wood have demonstrated,ll the relative intensities of X-ray beams scattered by structures comprising cylinders at points of a two- dimensional net vary significantly as the ratio of cylinder-centre separation to cylinder diameter is changed. Taking up the concept of ribbon-shaped aggregates, Chidichimo et uZ.l2 studied part of the potassium [2H,]palmitate + potassium laurate + water system using deuterium n.m.r. While divergence from right-circular cylinders could be inferred from their results, it is clear that the maximum divergence was limited to an approximate ellipse of axial ratio 2: 1.Thus there is little basis for proposing cross-sections of great eccentricity. As has already been pointed out, the new phase appearing in the SDS+water system on the higher amphiphile concentration side of the hexagonal phase contains cylinders distributed on a two-dimensional lattice having a parallelogram as the repeat unit. Fig. 6 shows the variation of parallelogram acute angle with amphiphile concentration. These results may be interpreted as the outcome of a conflict between one set of forces attempting to retain the close-packed structure of a rhomboid arrangement of circles and another set of similar magnitude striving for the more openR.M. WOOD AND M. P. MCDONALD 28 1 1 I I I I 1 0.50 0.54 0.58 0.62 0.66 0.70 SDS (mass fraction) Fig. 6. Effect of composition on parallelogram acute angle in the distorted hexagonal phase structure of SDS + water binary system at 3 I3 K. t B 1711 SDS (mass fraction) Fig. 7. Effect of composition on the area of the repeat unit of the two-dimensional hexagonal and distorted hexagonal structures at 313 K. (-), Hexagonal phase; A, distorted hexagonal structure. 0.40 0.45 0.50 0.55 0.60 0.65 packing of a rectangle to fit the reduced symmetry possessed by an ellipse. In consequence, distorted hexagonal packing of the cylinders would be expected to form within a two-phase region extending over a relatively wide composition range until true hexagonal packing is unfavourable.Evidence in support of this proposal appears in fig. 7, which is taken from our previous paper and shows the two-phase region extending over ca. 0.1 mass fraction of SDS. As further evidence the n.m.r. data of the previous paper5 can be cited since, as might be expected from the above reasoning, this data displayed a continuous variation from hexagonal to distorted hexagonal structure, i.e. the transformation is not first order. In the KO + water system the cylinder radius, 2.19 nm, remains invariant with KO concentration (as calculated earlier for hexagonal phase) and agrees well with the length of a KO molecular, 2.25 nm, implying that for this molecule the hydrocarbon chain has a restricted range of conformations. The only factors which could be282 SDS AND KO LIQUID-CRYSTAL PHASES responsible for the lower flexibility compared with the apparently highly flexible state of the SDS molecule are the stronger intermolecular forces of alignment for the much longer oleyl chain and the nature of the head-group. A constant cylinder radius necessitates an increase in the relative total length of cylinder with increasing surfactant content since the monomer surfactant content of the water component will not increase.Eqn (1) states that for constant cylinder radius the rhombus area is related to the mass fraction of surfactant uia the partial specific volume of the surfactant and the relative total lengths of the cylinders and their surrounding water.However, the value of 0.96 for the partial specific volume of KO1 when used in that equation causes the index of c to move further away by a small amount from the empirical value of - 0.88 obtained from the data of Ekwall et aL4 than if the specific volume had been assumed to equal that of water. Adopting this last assumption, eqn (1) becomes equivalent to stating that the mass and volume fractions of surfactant are equal. This equality could be obtained for the data of Ekwall et al. plotted in fig. 4, which on conversion gives the relationship between cylinder rhombus areas as A , = c ~ . ~ ~ A ~ by relating the total length of the cylinders and their surrounding water by L, = c0.l2LR.For the approximate composition range of hexagonal phase in the potassium oleate +water system, 0.2-0.6 mass fraction, the ratio L,/LR changes from 0.82 to 0.94. Over this composition range the total length of surfactant cylinders increases with surfactant content relative to the length of the surrounding water until it approaches the water length at ca. 0.6 mass fraction of surfactant. This behaviour is consistent with the need for a constant-radius cylinder to grow in length with increased amphiphile content of the system. Note that for a constant cylinder radius and equal length of cylinder and surrounding water, the results of Ekwall et al. describe a situation where at each surfactant concentration there is more water present between cylinders than even equality of mass and volume fractions permits.Thus, greater surfactant contents require water to come from somewhere other than between the cylinders to maintain the equality of length of the cylinder and the surrounding water, constituting a most unlikely situation. The earlier argument that the flexibility of the hydrocarbon chains in SDS molecules and the resulting deformable amphiphile cylinders are responsible for a relatively wide two-phase region appears relevant in the present context of the transformation from a hexagonal to a rectangular phase in the KO+water system. Here the cylinders of constant radius containing apparently much less flexible molecules could be expected to exhibit a different type of transformation behaviour. This expectation is supported by the datssf fig.1 and 2, which show a relatively narrow two-phase region. In both systems, once the non-hexagonal phase has been established as the only phase present, the size of the two-dimensional lattice repeat unit decreases with increasing surfactant content because of loss of water from between the cylinders. Finally, the proposal of chain flexibility as an important factor in determining the mechanism of transformation from the hexagonal phase received some support from the work of Rendall et aZ.l3 These workers determined, by penetration experiments,14 the general types of intermediate phases occurring in sodium and potassium soaps and in sodium alkyl sulphates. Short hydrocarbon chains, i.e. C , to Cl0, appear to be associated with the so-called cubic phase; for intermediate chain lengths (Clo to C14) the phase is described as I,, which in some instances has been identified as the deformed hexagonal phase, and longer-chain systems exhibit a different phase, designated I,.The C,, sodium sulphate system did not contain an intermediate phase at all. Unfortunately, X-ray diffraction was not employed to ascertain the structures of these phases.R. M. WOOD AND M. P. MCDONALD 283 CONCLUSIONS The hexagonal phase in SDS + water becomes a deformed hexagonal phase at higher amphiphile contents. The hexagonal phase in KO +water becomes a rectangular phase at higher amphiphile contents. In the SDS system the cylinders of the hexagonal phase grow by increasing their cross-sectional area as the amphiphile content is increased.In the KO system the cylinders of the hexagonal phase remain substantially constant in cross-sectional area and grow with increased amphiphile concentration in an axial direction until the cylinders and their containing rhomboids are of the same length. The hexagonal phase in SDS +water appears to become a distorted hexagonal phase when the hydrocarbon chains of the molecules reach a fully extended configuration. The behaviour of the hexagonal phase in the two systems is determined by the length and flexibility of the molecule hydrocarbon chains, SDS molecules appearing to be much more flexible than KO molecules. F. Husson, H. Mustacchi and V. Luzzati, Acta Crystallogr., 1960, 13, 668. J. Bergeron, 1st Congr. Mondial Detergence et Pro& Tensio-Actif. (Paris, 1954), vol. 1, p. 24. R. R. Balmbra, D. A. B. Bucknall and J. S. Clunie, Mol. Cryst. Liq. Cryst., 1970, 11, 173. P. Ekwall, L. Mondell and K. Fontell, J. Colloid Interface Sci., 1969, 31, 508. I. Leigh, M. P. McDonald, R. M. Wood, G. J. T. Tiddy and M. A. Trevethan, J. Chem. Soc., Furaday Trans. 1, 1981, 77, 2867. S. Eins, Mol. Cryst. Liq. Cryst., 1970, 11, 119. P. Ekwall, in Advances in Liquid Crystals, ed. G. H. Brown (Academic Press, New York, 1975), vol. 1, p. 1. * V. Luzzati, H. Mustacchi, A. Skoulios and F. Husson, Acta Crystallogr., 1960, 13, 660. Y. Hendrikx and J. Charvolin, J. Phys. (Paris), 1981, 42, 1427. T. Gale and R. M. Wood, to be published. lo G. Oster and D. P. Riley, Acta Crystallogr., 1952, 5, 272. l2 G. Chidichimo, N. A. P. Vaz, Z. Yaniv and J. W. Doane, Phys. Rev. Lett., 1982, 49, 26. l 3 K. Rendall, G. J. T. Tiddy and M. Trevethan, J. Chem. SOC., Faraday Trans. I, 1983, 79, 637. l4 A. S. C. Lawrence, Liquid Crystals, ed. G. H. Brown (Gordon and Breach, London, 1969), vol. 1, p. 1. (PAPER 3/2240)

ISSN:0300-9599    

DOI:10.1039/F19858100273

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1985, 81, 285-298 Measurement of Coagulation Rate Constants using Number-fluctuation Spectroscopy BY J. G. RARITY* Royal Signals and Radar Establishment, St Andrews Road, Great Malvern, Worcestershire WR14 3PS AND K. J. RANDLE Royal Military College of Science, Shrivenham, Wiltshire SN6 8LA Received 9th January, 1984 The fluctuations in number and scattering power of particles in a small volume have been measured via the moments of the detected intensity distribution. Under the approximation that a k-fold aggregate scatters, on average, k times the intensity of a singlet, the evolution of the intensity moments during Smoluchowski aggregation are related to the aggregation half-life. In particular, the second moment increases linearly with time and the rate of increase leads to an estimate of coagulation half-life. Choosing the ‘angle of no change’ (the angle where the mean scattered intensity remains constant during coagulation) as the scattering angle ensures the validity of the approximations.Measurements made on rapidly coagulating polystyrene latex samples are compared with theory and with results obtained using other techniques. In conventional light-scattering and turbidity measurements light from an intense source (e.g. a laser) illuminates a large number of scatterers in a loosely defined scattering volume. During coagulation, changes in the scattered intensity reflect changes in the number and scattering cross-sections of particles in this volume. To obtain quantitative information from these measurements it is necessary to use accurate models for the coagulation kinetics and the aggregate scattering cross-sections.The second-order kinetics proposed by Von Smoluchowski’ have been verified by particle-counting techniques and in recent years a Rayleigh-Gans-Debye (R.G.D.) model for aggregate scattering has been Using this simple but accurate model for aggregate scattering, Lips et u L . ~ were able to measure Smoluchowski rate constants from high-angle light-scattering data. Later Lips and Willis developed a low-angle light-scattering technique5 that gave a direct measure of the rate constant. At low enough angles the aggregates scatter light in proportion to the square of their volume and the intensity changes linearly with time, i.e. I(t) = I(0) ( I + 2t/t;) (1.1) 8 - 0 where Z(t)e+o is the low-angle scattered intensity after time t and ti is the half life of the coagulation.This equation is analogous to the result obtained by Troelstra and Kruyt6 for the initial change in turbidity under the Rayleigh approximation. These static light-scattering techniques rely on large numbers of particles being seen at one time to provide adequate sampling of the aggregate distribution. In contrast, automatic particle-counting techniques7-* view aggregates singly and rely on a large throughput of particles to provide adequate statistics. Although these 285286 COAGULATION RATE CONSTANTS techniques provide more information on the aggregate distribution they are invasive. Samples are removed from the reaction vessel and quenched (by dilution) before counting.The technique described in this paper falls between these two approaches. Reducing .the size of the scattering volume by using high numerical aperture viewing optics in conjunction with a narrow slit can reduce the average number of particles in view to less than ten. On scanning the sample through the scattering volume the detected intensity fluctuates because of the changing number of scatters seen and, if they are polydisperse, because of their varying scattering cross-sections. Measurement of the second moment (variance) and higher moments of the intensity fluctuation distribution can thus yield information about the mean number of scatterers and the distribution of scattering cross-sections. Although it is quite easy to show that the number of particles present in a small volume follows Poisson statistic^,^ the scattered intensity distribution is, in the general case of arbitrary illumination profile and polydispersity, not easily accessible.However, the moments can be obtained using a generating-function approachlo or directly from general formulae.ll The moments approach also sidesteps the experi- mental problem of Poisson noise inherent in the photodetection process. Photon- counting techniques allow measurement of the photocount distribution (p.d.) and the factorial moments of the p.d. are directly proportional to the intensity moments. The first number fluctuation experiments carried out by Schaefer and Pusey12 verified the relations between the factorial moments of the p.d.and the mean number of monodisperse spheres in a Gaussian beam. Although other theoretical work on polydisperse systems has been publishedl39 l4 it is only recently1' that experimental verification has been attempted. It is thought that the work presented here and in ref. (1 1) represents the first complete test of moments theory for polydisperse systems. THEORY NUMBER FLUCTUATIONS In the typical number-fluctuation experiment (fig. 1) a small scattering volume is defined by the image of a detector slit along the beam waist of a finely focused laser beam. The use of high numerical aperture (n.a.) optics allows small volumes (10-7-10-10 cm3) to be defined and ensures a negligible contribution from the more familiar interference fluctuation signal. The Rayleigh resolution criteria ensures, however, that all intraparticle scattering remains coherent up to particle size 1-2 pm.Under these assumptions (negligible interparticle yet full intraparticle coherence) the first three normalised intensity moments can be written:l03 l1 where the angular brackets denote ensemble or time averages under of stationarity, p is a constant dependent on incident laser intensity the assumption and ( m ) is theJ. G. RARITY AND K. J. RANDLE 287 SCATTERING VOLVME , /' I I LASER I 1 'y IMAGING LENS I Fig. 1. Schematic diagram of the typical number-fluctuation experiment. In this work a photon-counting detector is placed behind the slit. mean number of particles present in the scattering volume. The a, are averages over the particle scattering cross-section distribution : rrn J a2np(a2) d(a2) (a2)n 1" a2p(a2) d(a2) (a2,> - 0 - a, = ~ (4) J O where a2 is the particle (intensity) scattering cross-section and p(a2) is the normalised probability of finding a particle with cross-section a2.The constant b, [in eqn (3)] is related to the exact form of the intensity profile across the scattering volume and is discussed in more detail in the Appendix. As the intensity cross-section of a focused laser beam at the beam waist is very nearly Gaussian and there will always be some diffraction blurring at the ends of the beam, the factor 6, lies between 4/3 and 1.54. Measurement of polydispersity information directly from the moments requires accurate prior knowledge of (m) and b,, information that is not normally available.In the case of aggregation (m) and b, are measured from the unaggregated latex and the changes in the moments caused by aggregation are studied. In practice photon-counting techniques are used. The number of photons detected in a finite sample time of duration Tis proportional to the intensity incident on the detector integrated over that sample time I,. The mean count-rate ( n ) is given bylo where q is known as the quantum efficiency of the detector and the normalised factorial moments dr) of the photocount distribution are equal to the normalised moments of integrated intensity : ( 6 ) (n> = 71 (IT) ( 5 ) (n(n- 1) ... ( n - r + 1)) - (PT) -~ n(r) = - ' When T is shorter than any number-fluctuation time-scales the intensity moments are given by eqn (1)-(3).However, ensuring that Tis much longer than any rotational time-scales associated with aggregate scattering allows one to use orientation averaged aggregate form factors in eqn (4). The subscript Tin intensity averages will be assumed from this point onwards.288 COAGULATION RATE CONSTANTS APPLICATION TO COAGULATING SYSTEMS The orientation-averaged scattering by a k-fold aggregate is well approximated by4 where I,(@) is the single-sphere Mie cross-section at angle 8 and is an R.G.D. form factor dependent on interparticle separation (rij) within the aggregate and the modulus of the scattering vector q. q is related to the scattering angle via 41cn sin (812) L O 9 = (9) where Lo is the free-space wavelength of the incident light and n is the refractive index of the suspending medium.Eqn (7) appears to be a good approximation for aggregate scattering even for a 2 1 .2,3 where a is the dimensionless size parameter of Mie theory a = 21cnr/Ao (10) and r is the radius of a singlet particle. The polydispersity factors (a2n) can be written where n, is the number density of aggregates of order k. It is clear that ( a 2 ) is the static structure factor fS(q)] of the system and the (a2n) can be regarded as higher-order structure factors. In general the (a2n) depend on the detailed aggregate distribution and aggregate morphologies. However, simplifications can be made at specific scattering angles. For low-order and linear aggregates the form factor Ak(q) is dominated by terms in sin (mqd,)/mqd,, where do is the centre-to-centre distance of particles at contact and rn is integer. Under the condition qd, = 1c ai(q) = k11(8) (12) for these aggregates. This approximation holds to within 10% for close-packed aggregates up to order k = 10.15 Using this formula in eqn ( 1 1 ) gives (a2n) = (kn)1r(8).(13) The polydispersity factors are directly proportional to the moments of the aggregate distribution. The Smoluchowski theory of coagulation kinetics predicts an aggregate distribution of the where E = t / z i is the extent of coagulation, the ratio of elapsed time t to coagulation half-life ti, and No is the original number density of singlet particles. No is related to the total number density N of aggregates in the system by289 J. G. RARITY AND K. J . RANDLE Combining eqn (1 l), (12) and (14) gives The sum is evaluated by successive differentiation of the geometric sum CF= rk, where r = E/ 1 + E.This yields (a2) = ( I + E ) Il(e) ( a 4 ) = (1 + E ) (1 + 2E) I,"(@ ( a 6 ) = (1 + E ) (1 + 6E+ 6E2) I;(@. These averages are continued up to (a12) in the Appendix. In the moments eqn (1)-(3), ( m ) represents the mean number of scattering entities (i.e. the mean number of aggregates) in the scattering volume. From eqn (1 5 ) it is clear that (mo> = (4 (1 + E ) (20) where (m,) is the original mean number of singlet particles present. With this relation and eqn (1 7)-( 19) we can write the normalised intensity moments as ( 0 = Kmo> I,(@ (21) (23) 3( 1 + 2E) b3( 1 + 6E+ 6E2) + -- ( I 3 ) - I + ( I > 3 (m0) (mO>2 From eqn (21) it is clear that under this approximation the mean intensity remains constant during coagulation.This scattering angle has been dubbed ' the angle of no change'.3 The Smoluchowski rate constant K is related to the half-life and the original number density No vial6 1 I tl = ~ KNo' The second moment can thus be written where V is the scattering volume. V can be calibrated using a monodisperse and unaggregated sample of known number density and measuring the second moment. With a knowledge of V an estimate of the rate constant can be obtained from the rate of change of the second moment with time. Eqn (21)-(23) and (25) form the basic results of this work. The following sections go on to discuss the corrections and limitations encountered in their practical application. Note that in the low-angle limit where ai(q -, 0) = k211(0) ( I ) = D (mo) Il(0) (1 + 2E) (26) (27) the mean intensity is given by as stated in eqn.(1.1). The relationship between eqn (1.1) and the second intensity moment [eqn (22)] is a direct manifestation of the Ornstein-Zernike relationship known in the theory of the liquid state.17290 COAGULATION RATE CONSTANTS When the scattering volume is much larger than any order correlation volume (the maximum aggregate size in this application) this relationship can be stated simply as:18,19 where Am is the instantaneous number fluctuation about the mean number (m,) and S(0) is the zero-angle value of the structure factor of the system. At the angle of no change one is essentially measuring the fluctuation in the number of singlet particles in the volume including those contained in aggregates; thus the second moment can be written? This relationship is discussed in more detail by Pusey,lg who used a number-fluctuation technique to study an interacting system in the high-angle limit.POLYDISPERSITY AND PREAGGREGATION Any polydispersity of the original sample would alter the measured initial moments. The assumption that particle size and incorporation into aggregates are not strongly correlated allows the separation of the (a2n) factors into (a2n> = (azn),. (a2n>, (31) where is the average over polydispersity. For low polydispersities (< 10%) this assumption is valid as it forms one of the basic approximations of Smoluchowski theory.16 is the nth-order average over the aggregate distribution and The equation for the second moment becomes altering the measuring gradient.Calibrating the volume with a sample of similar polydispersity will allow measurement of a2,/V and the problem is removed. If the sample is preaggregated the gradient should not be altered as long as the initial aggregate distribution approximates to that of Smoluchowski. However, it is vital to ensure that the volume is calibrated with an unaggregated sample. MIS ALIGNMENT If the assumption a m = kIl(8) (3 3) is in error the mean intensity [eqn (21)] will not remain constant. The fundamental component of A,(q) is the function sin qd,/qdo. If qdo = n - 6, where d/n < 0.1, a closer approximation for I(q) is 26(k- 1) n-6 ai(q) = k( 1 + (34)J.G. RARITY AND K. J. RANDLE 29 1 Analysis of the averages using this form for a; leads to corrected mean intensity and second moments of the form (0 = a<mo> I,(@ (1 + 6 2 E / 4 (35) (36) <I2) 1 + 2E[ 1 + 28( 1 + E)/n] (m0) (/)2 = + 1 (m0) 2Kz[ 1 + 26( 1 + E)/;n] V = 1+-+ (37) For small 6 a simple correction procedure involves estimating an uncorrected aggregation half-life ti from the second-moment measurements. The rate of change of the first moment with time is +26/nt;, from which an estimate of 28/71 can be obtained. This estimate can then be used in a correction factor for the measured value (38) of K : Em is a mean extent of the measured aggregation, which is taken as half the maximum extent seen, and Km is the measured rate constant.For corrections > 10% an iterative correction procedure would be more appropriate. Any change in the position of the angle of no change during the experiment (such as that seen by Giles and Lips)2o caused by the changing balance of particle separations should also be largely corrected for by this procedure. K = Km/[ 1 + 26( 1 + Em)/n]. EXPERIMENTAL ACCURACY For accurate estimates of the moments a representative sample of the aggregate distribution must pass through the scattering volume. Normally this involves scanning the sample through the scattering volume rather than waiting for Brownian diffusion to bring different particles in and out of view. Assuming linear translation perpendicular to the scattering plane it can be shownll that the error (a) in the determination of a,/(m) is given by where MTis the duration of the experiment, z, is the time taken to translate the sample a distance equivalent to the 1 /e2 radius of the incident beam and b4 is the higher-order average over the intensity profile of the beam.Writing the azn averages in terms of the aggregation extent E [eqn (17)-(19) and (A 7)] gives a(E) = 2(1+ 12E+ 12E2) 16(1 - J($d[d2+7!$(4(1+2E)+ (1+2E) 3( 1 + 2E) (40) where the maximum expected values of b3 = (4/3)1 and b4 = 2% (see Appendix) have been used. This expression diverges quite slowly as E is increased. Typically for MT/71i~, 900 and ( m , ) = 3 the error at E = 0 is only 4.25% , rising to 6% for E = 1 and 7.2% for E = 2. To ensure orientation-averaged aggregate scattering the sample time T must be much longer than the typical fluctuation time associated with rotational Brownian motion.For contacting doublets of 100 nm radius in aqueous suspension at 25 "C this is ca. 3 rns.,, T 2 30 ms is thus adequate until the number of slowly rotating long-chain292 COAGULATION RATE CONSTANTS MICROSCOPE OBJECTIVES C OR R E LATOR Fig. 2. Apparatus used in this experiment. The Datalab DL4000 system was used to evaluate the factorial moments from measurements of the photocount distribution and the Malvern correlator measured intensity autocorrelation functions from which the correlation time z,, used in the error analysis, could be estimated. P.M., photomultiplier; V.D.U., visual display unit. aggregates becomes significant in the later stages of the aggregation.Similarly a lower limit on z, is set by T < 2,/8 to ensure the number fluctuations themselves are not integrated out. z, = 0.27 s was used in this work and with experimental durations of ca. 420 s accuracies comparable to those quoted above can be achieved. To ensure that the aggregate distribution does not change appreciably during each 420 s measurement the coagulation half-life ti must again be appreciably longer. Working at number concentrations of lo8 particle ~ m - ~ ensures ti 2 3300 s when the rate constant K < 3 x m3 s-l. The second moment, being linear in E, is not distorted by using ti < 3000 s as long as a mean measuring time is taken. However, the third moment, being quadratic in E, would be slightly over-estimated. The expected statistical error in the measurement of the third moment is larger than that in the second.The formulae for estimated errors in the first and third moments are given in the Appendix, along with the values of related constants. EXPERIMENTAL APPARATUS A schematic diagram of the apparatus used in this work is shown in fig. 2. Light from a helium-neon (HeNe) laser is focused into the sample by a microscope objective. The image of the sample volume is transferred to a slit on the front end of a photon-counting photomultiplier tube via a x 3.7 microscope objective of numerical aperture 0.1. The photomultiplier is mounted on an XYZ micropositioning stage to allow alignment of the slit on the beamwaist (by maximising the second moment). The scattering cell of 5 x 2 x 30 mm internal dimensions was mounted in a thermostattable brass holder with viewing ports for incident and scattered light.This in turn was mounted on a motor-driven micropositioning stage to allow translation of the sample perpendicular to the scattering plane (the plane of the paper in fig. 2) at up to 4 mm min-I. The optics defined a scattering volume of ca. 2 x lo-* mm3, ideal for work with sample concentrations around los particle cmP3. A reflex viewer in the photomultiplier allowed direct visual inspection of the sample. Polarising optics were not thought necessary as depolarised scattering is small and does not change appreciably during c~agulation.~. 4.J. G. RARITY AND K. J. RANDLE 293 For aggregation measurements the detector was aligned at a scattering angle satisfying the condition qd, = nn, do being the electron-microscope-measured particle diameter and q the modulus of the scattering vector defined in eqn (9).Particles were chosen such that this condition was satisfied close to 90" so that distortion of the scattering volume was not large. The pulse train output from the photomultiplier electronics were fed through a ratemeter to a Datalab DL 4000B histogram analyser. This allowed collection of the photocount distribution in up to 4000 channels. An internal microprocessor was programmed to calculate the normalised factorial moments and the results were displayed on an attached visual display unit. MATERIALS AND METHODS Three polystyrene latices were used in this work, two commercial surfactant-stabilised latices (176D and 320P) and one university-produced charge-stabilised latex (195B).In general it was found that the charge-stabilised latex gave more reproducible results, and thus the bulk of the measurements were made on this latex. Electron microscope (e.m.) sizes and standard deviations, photon correlation spectroscopy (P.c.s.) size data and scattering angles where qd = nn are listed in table 1. Table 1. Details of the latices used, convenient scattering angles where qd, = nn and polydispersity factors calculated at these angles _ _ _ _ _ ~ - e.m. standard P.C.S. POlY- diameter deviation diameter scatter dispersity, latex supplier /nm /nm /nm angle/" aZpl1 stabilisation 176D Dow 176 k2.3 178 86 1.007 surfactant 320P Polyscience 320 k32 330 90 1.024 surfactant 195B Bristol University2, 195 k19 192 77 1.093 charge Polydispersity corrections [eqn (27)] estimated using the e.m.sizes and standard deviations and assuming full Mie scattering cross-sections have been calculated" and are also shown in table 1. The refractive index of the particles was assumed to be independent of size. This may not be the case for much smaller particles. The latices were suspended in 0.1 pm filtered double-distilled deionised (f.d.d.d.) water and coagulation was induced using solutions (in f.d.d.d. water) of AnalaR grade LaNO, (195B and 320P) or KC1 (176D). The pH of the suspensate was not rigorously controlled but always fell between 5 and 6 because of dissolved CO, in the f.d.d.d. water. Sample volume calibration was carried out using unfiltered exact dilutions of latex 195B. No correction for polydispersity was thus necessary for measurements on this latex but the 320P and 176D results were corrected accordingly.Before each aggregation experiment the sample cells were cleaned using concentrated nitric acid then soaked and thoroughly rinsed in f.d.d.d. water. The cells and stoppers were dried in filtered nitrogen and sealed. Rinsing and drying were carried out using a bent syringe inserted through the stopper (6.5 mm Subaseal) which was later used as both air vent and mixing paddle. Latex samples for aggregation were filtered directly into the precleaned dried cell through filters of pore size roughly twice the particle diameter. After determining the initial number density and mean scattered intensity a measured amount of salt solution was added and quickly mixed with the latex sample.The salt concentration in the final mix could be estimated either from the volume dilution or the reduction in the mean scattered intensity. These rigorous preparation procedures ensured reproducible dust-free ( > 0.2 pm in diameter) and preaggregation-free samples. As the presence of dust or large aggregates could seriously distort the photocount distribution, producing spuriously high estimates of moments, the samples were also checked by eye (using the reflex viewing system) before and often during experiments. If dust (scattering more than the particles) was seen the samples were discarded.294 COAGULATION RATE CONSTANTS 2 00 n 100 0 0 40 tlmin 80 1 . o l I 0 40 tlmin 00 Fig.3. Plots of the moments against time. Experimental points are denoted by crosses. (a) Mean count rate A (normalised to 100) against time, (6) second normalised factorial moment d2) against time and ( c ) third normalised factorial moment d3) against time. In (a) and (b) the full lines are straight-line fits to the data from which an estimate of the rate constant Kcan be made. This was used to predict the third moment [full line in (c)]. The errors (dotted) are estimated from eqn (A 5), (35) and (A 6). RESULTS Plots of the mean count rate and normalised factorial moments from a typical rapid aggregation, (195B 0.02 mol dm-, LaNO,) are shown in fig. 3. Least-squares straight- line fits to the mean count rate and second moment are shown as full lines and errors are estimated from eqn (A 5) and (40).The third moment is predicted from eqn (23) using the measured value of the half-life and an apparent b, of 1.61 estimated fromJ. G. RARITY AND K. J. RANDLE 295 I SMOLUCHOWSKI RATE I * t I o7 lo8 lo9 Nlparticle ~ 1 7 7 ~ ~ Fig. 4. Rapid coagulation rate constants (corrected to 25 "C) measured using number-fluctuation spectroscopy compared with single-particle-counting results. 0,195B in 0.02 mol dmd3 LaNO,; ., 320P in 0.02 mol dm-, LaNO,; A, 176D in 0.66 mol dm3 KCl. The full line is taken from the data of Hatton et al.' moments measurementsll and inclusive of polydispersity corrections. The error in the predicted third moment was estimated from eqn (A 6). The results appear to follow the theory quite well up to E = 2 (z z 70 min in fig.3). The high points on the third-moment plot reflect the passage of only one or possibly two anamolously large scatterers (dust or aggregates) through the sample volume, emphasising the sensitivity of the higher moments to contaminants. The rate of change of the mean count rate yields an estimate of the misalignment parameter of 6 = -0.013~. This indicates a slight deviation from the correct constant-intensity angle or a slow change of this angle during aggregation, but it is small enough for the correction procedure to be valid. The apparent volume for the 195B experiments was estimated as V/a,, = (2.03 k0.04) x cm3, leading to a value of the rate constant for this result (corrected to 25 "C) of K = (3.42k0.25) x 10l8 m3 s-l. The initial number density (No) could be estimated from the intercept of the fit to the second normalised moment bearing in mind the polydispersity correction listed in table 1.For this sample No = (1.43 kO.05) x lo8 particle ~ m - ~ . Rate constants measured from several experiments at different number densities are shown in fig. 4. The solid line is reproduced from the data of Hatton et aZ.,' who used an automatic particle-counting technique to determine rate constants over a wide range of initial number densities. Their data show a definite dependence of the rate constant on initial number density, which is supported by the data presented here. Low-angle5 and high-angle light-scattering4 measurements also lead to rate constants of this order but the dependence of rate constant on initial number density has not been directly investigated using these techniques, The theoretical Smoluchowski rate constant is also shown. The rate constant at infinite dilution is only ca.50% of the theoretical value because of hydrodynamic effects between particles at close The results of Giles and Lips20 suggests that the mean particle separation in296 COAGULATION RATE CONSTANTS aggregates, as measured by the position of the angle of no change, increases by between 5 and 10% per half-life during rapid electrolyte-induced coagulation. The measurements on the 195B latex are in qualitative agreement with this result. In all the experiments the mean intensity decreased during aggregation and values of n range from - 0 . 0 1 ~ to - 0 . 0 9 ~ . From the definition of 6 [eqn (34)] this implies misalignment of between 1 and 8" or an increase in mean particle separation of between 1 and 9% per half-life. As alignment was thought to be within 1" of the angle of no change the latter explanation is more probable.No better quantitative analysis of the mean count rate could be made as the incident intensity was not monitored. The importance of monitoring the incident intensity was not realised until the detailed results analysis was carried out. After two aggregation half-lives there was a definite tendency for the second and third moments to rise above theoretically predicted values. This, at first sight, is contrary to the expected effect from the falling mean intensity [see eqn (35) and (36)] and suggests non-Smoluchowski behaviour.Above E = 2,20% of the original singlet particles are contained in aggregates of order 8 and above. These particles may have large cross-sections for collision because of extended morphologies (suggested by the results of Giles and Lips20 and sedimentation effects. Thus all rate-constant measurements were made using data taken before E = 2. The change in measured rate constant with number density could be enhanced by this effect, as, with a short half-life, measurements are constrained to later stages of the aggregation. CONCLUSIONS This work has shown that number-fluctuation techniques can be used to study irreversible coagulation. The restrictions imposed by the time-scale of rapid coagulation obviously limit the technique to low number densities (< 5 x lo8 particle ~ m - ~ ) in its present form.The measurement range could be extended down to initial number densities around lo6 particle ~ m - ~ , although solvent scattering and dust effects would become greater. This would allow verification of the low-concentration asymptote of the measured coagulation rate constant. The major limitation in using higher initial number densities, the limited statistical accuracy in finite measurement time, could be alleviated by using a much higher scan speed and a shorter sample time with a smaller sample volume. The sample time would then be much shorter than the time-scale of aggregate rotation and a different, more morphology-dependent aggregate scattering approximation would be appropriate. This approach is being studied in more detail at present.Similarly, inadequate sampling of the aggregate distribution limits the amount of information obtainable from the third intensity moment in this study. In principle the third (and higher) intensity moments contain information about the higher moments of the aggregate distribution. When no a priori information about the aggregate distribution is available this information can be useful. Consequently, the technique shows potential in the study of reversibly coagulating systems and systems exhibiting gas-to-liquid (or solid) phase transitions.24$ 25 For such systems the rate of change of the overall aggregate distribution would be quite slow whereas locally the distribution could change quite rapidly, thus enhancing the accuracy of the measured higher intensity moments.J.G. RARITY AND K. J. RANDLE 297 APPENDIX The beam profile parameters b, are defined in ref. (1 1) as Vn-l [ bn(r)d3r and bear a strong resemblance to the polydispersity factors a,. b(r) is a parameter describing the intensity profile across the sampled volume. Two approximations for a sample volume limited by the image of a detector slit on the focused laser-beam waist are expressed in cylindrical polar coordinates as exp (- 2r2/a2) -az < z < +a, z > az,z < -oz b(r) = and where a is the l/ez intensity radius of the (Gaussian) beam and 20, is the length. The resulting b, are" b(r) = exp (- 2r2/a2) exp ( - 2z2/a3 2,/2n, hard-ended Gaussian [eqn (A 2)] (A 3) (A 4 4 (A 4b) In both these approximations b, = 1 and thus it is ignored in the moments equations.If the slit image on the beam is appreciably diffraction blurred one would expect the measured b, to lie between these two extremes. In practice it appears that the values lie closer to the Gaussian-ended Gaussian values, l1 possibly because of slight defocussing and imperfections in the optical elements. From eqn (A 4) b, should lie between 4/3 and (4/3): and b, between 2 and 2s. The error in any finite-time measure of the rth moment is related to the value of the 2rth and lower moments and the time dependence of the intensity fluctuations. The standard deviations in the measured mean intensity and third moments are given by bn={ (2,/2n)$ Gaussian-ended Gaussian [eqn (A 3)]. and = J( %) (92/2ai + (39.44 + 6.36b3 a4 - 11 .41b, a, a,) MT (m), (2/ 1565 a5 + 36b3 a3 U: - 4.60b4 a4 a, - 10.08b: a:) + (0.5576, a6+9b:a:a2-4.90b (m>5 + Clearly evaluation of a(d3)) requires knowledge of a, up to as.Continuing the evaluation section procedure detailed in the main text, eqn (2) leads to (A 7) (A 8) (A 9) = (1 +E)(I + i4~+36~2+24~3)1:(e) +P) = (1 +E)(I + ~ o E + 150~2+240~3+ 120~4)1:(e) (a1,) = (1 + E ) (1 + 62E+ 540E3 + 1560E3 + 1800E4 + 720E3) I:(@). Use of eqn (20) allows (d3)) to be expressed in terms of (m,) and E. In principle the full polydispersity correction [eqn (3 l)] should be used, but as the higher-order polydispersity factors are not large the use of the Gaussian-ended Gaussian values of b, [eqn (A 4b)l largely compensates for this. 11 FAR 1298 COAGULATION RATE CONSTANTS M. von Smoluchowski, Z. Phys. Chem., 1917, 92, 129. H. Benoit, R. Ullman, A. J. de Vries and C. Wippler, J. Chem. Phys., 1962, 59, 889. A. Lips, Thesis (CNAA, 1974). A. Lips, C. Smart and E. Willis, Trans. Faraday SOC., 1971, 67, 2979. A. Lips and E. Willis, J. Chem. SOC., Faraday Trans. 1, 1973, 69, 1226. S. A. Troelstra and H. R. Kruyt, Kolloid Chem. Beih., 1943, 54, 225. W. Hatton, P. McFadyen and A. L. Smith, J. Chem. Soc., Faraday Trans. I , 1974,70, 655. M. S. Bowen, M. L. Broide and R. J. Cohen, in Kinetics ofAggregation and Gelation, ed. F. Family and D. P. Landau (North Holland, Amsterdam, 1984), in press. S. Chandrasekhar, Rev. Mod. Phys., 1943, 15, I . (Plenum Press, New York, 1977), p. 45. lo P. N. Pusey, in Photon Correlation Spectroscopy and Velocimetry, ed. H . Z . Cummins and E. R. Pike l 1 J. G. Rarity and K. J. Randle, Opt. Acta, 1984, 31, 629. l 2 D. W. Schaefer and P. N. Pusey, Phys. Rev. Lett., 1972, 29, 843. l 3 R. Barakat and J. Blake, Phys. Rev. A , 1976, 13, 1122. l4 T. Yoshimura, H. Fukumaru and N. Wakabayashi, J. Opt. SOC. Am., 1982, 72, 780. l5 J. G. Rarity, Thesis (CNNA, 1984). l6 H. R. Kruyt, Colloid Science (Elsevier, Amsterdam, 1952), chap. VIII, p. 302. L. S. Ornstein and F. Zernike, Proc. Akad. Sci., Amsterdam, 1914, 17, 794 [reprinted in The Equilibrium Theory of Classical Fluids, ed. H . L. Frisch and J. Lebowitz (Benjamin, New York, 1964)l. l8 P. A. Egelstaff, An Introduction to the Liquid State (Academic Press, London, 1967), chap. 2, p. 12. l9 P. N . Pusey, J. Phys. A , 1979, 12, 1805. 2o D. Giles and A. Lips, J. Chem. SOC., Faraday Trans. 1, 1978, 74, 733. 21 J. G. Rarity and K. J. Randle, in Photon Correlation Techniques in Fluid Mechanics, E. 0. Shulz-DuBois (Springer-Verlag, Berlin, 1983), p. 359. 22 C. Young, Thesis (Bristol University, 1979). 23 L. A. Spielman, J. Colloid Interface Sci., 1970, 33, 562. 24 C. Cowell, R. Li-In-On and B. Vincent, J. Chem. SOC., Faraday Trans. I, 1978, 74, 337. 25 C. Cowell and B. Vincent, J . Colloid Interface Sci., 1982, 87, 518. (PAPER 4/037)

ISSN:0300-9599    

DOI:10.1039/F19858100285

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1985, 81, 299-310 Study of Small-port and Large-port Mordenite Modifications Part 2.-Ion-exchange Properties of Thermally Treated Ammonium Forms BY FRANCIS RAATZ,* EDOUARD FREUND AND CHRISTIAN MARCILLY Institut Frangais du PCtrole, B.P. 3 1 1, 92506 Rueil-Malmaison Cedex, France Received 6th February, 1984 The ion-exchange reactions 2H+ % Pt (NH,);+ H+ % Na+ and have been studied for H-mordenite obtained from its ammonium forms by a deep-bed-like calcination. Two kinds of procedures have been considered. The first corresponds to equilibrated acidic solutions starting from NaCl or Pt(NH,),Cl, ; the second corresponds to equilibrated alkaline solutions starting from NaOH. Under acidic pH conditions the exchange capacities of the HM forms are drastically reduced compared with that of the initial NH,-mordenite.This can be interpreted as being due to (i) a reduction in the number of exchange sites brought about by framework dealumination and (ii) occupancy of some of the remaining exchange sites by unexchangeable cationic AIV1 species formed during framework dealurnination. The exact nature of the AlV1 species is related to the calcination temperature, high temperatures favouring the formation of neutral hydroxyaluminium complexes. At high pH the decrease in apparent ion-exchange capacity is very limited. Reactions other than pure ion exchange, for instance the formation of sodium aluminates from the AIV* species present in the channels after the calcination step, are involved. In Part 1 of this series1 we studied in detail the preparation of the HM forms, from both large-port mordenite (LPM) and small-port mordenite (SPM), by calcination of the corresponding ammonium forms (NH,M).We demonstrated that in the course of this treatment, important structural modifications may occur even at relatively low temperatures, namely below 873 K. As expected, the main phenomenon which has been seen is the dealumination of the framework resulting from the action of water. More precisely, the numbers of tetracoordinated aluminium atoms (A1Iv) which are extracted from the crystalline skeleton depend strongly on the calcination procedures. This framework-dealumination process has been shown to be the direct cause of the pore unblocking of SPNH,M occurring in the course of calcination under deep-bed-like conditions.In contrast to the framework aluminium atoms, which are tetracoordinated (AP), the aluminium atoms located in extra-lattice positions are hexacoordinated (Alv1).2-5 The precise location of these AIV1 species has not yet been precisely defined. However, the AIV1/AI ratios in modified zeolites have already been estimated through A1 Kb X-ray fluorescence line shifts2* and more recently by means of 27Al magic-angle spinning n.m.r. spectro~copy.~~ This last method, combined with 29Si magic-angle spinning n.m.r.6 is potentially able to lead to a very detailed structural description of modified zeolites. Obviously, dealurnination of the aluminosilicate framework is likely to change profoundly the ion-exchange capacities of the HM forms, and indeed the ion- 299 11-2300 ION EXCHANGE IN MORDENITE exchange capacities of hydrogen forms resulting from the thermal treatment of NH,M and NH,Y decrease as the severity of the heat treatment inc~eases.~-~ In the case of modified HY, A1 atoms (most likely cationic AIV1 species) can be exchanged by Na+ cations in NaCP lo or in NaOH so1utions.lo? l1 In contrast, the treatment of modified HM in NaCl or NaOH solutions does not remove any aluminium from the solids.7$ lo In NaCl solutions the ion exchange Na,+, + HM f NaM + H$q is the main phenomenon7 but in alkaline media other reactions involving extra-lattice aluminium species have been proposed.In the present part of this series, we report a study of the ion-exchange properties of modified small-port and large-port hydrogen mordenites (SPHM, LPHM) for a large divalent cation [Pt(NH&+] and for a small monovalent cation (Na+).These SP- and LP-HM forms were prepared through a calcination of the original NH,M precursors under procedure 2 of Part 1 ,l namely deep-bed-like conditions. (These HM forms will be referred to in the following as HM ex NH,M forms). We focussed on this particular calcination procedure because it leads to stabilized products and causes an unblocking of SPM. We put a special emphasis on the modifications of the physico- chemical properties of the solids induced by treatment in solutions. In addition to the determination of the exchange levels through chemical analysis, the solids were characterized before and after the ion-exchange steps by hydrocarbon adsorption, X-ray diffraction and precise measurements of the A1 KB X-ray fluorescence line shifts.EXPERIMENTAL STARTING PRODUCTS Synthetic small-port NaM (Alite 150) was supplied by La Grande Paroisse (Montoir- de-Bretagne, France) and synthetic large-port NaM (Zeolon 100 Na) by Norton Co. (Akron, Ohio). The small-port and large-port NH,M forms were prepared by exchanging the NaM forms three times in 2 mol dm-3 aqueous NH,NO, at 373 K for 4 h to provide an exchange level of > 99.4%. PREPARATION OF THE HM FORMS The HM forms were prepared by the calcination of the LP- and SP-NH,M forms according to procedure 2 (deep-bed-like conditions). This procedure is characterized by a low flow of dry air and a high heating rate (for details see Part 1).ION EXCHANGE PLATINUM Procedure A (acidic medium): The HM ex NH,M forms were exchanged once for 4 h at 373 K in 0.5 mol dm-3 Pt(NH,),Cl, solutions containing two Pt(NH,);+ cations per aluminium atom present in the solids. SODIUM Na+ cations were introduced into the HM ex NH,M forms according to one of the following procedures. Procedure B (acidic medium): The HM forms were exchanged five times for 24 h at 293 K in 2 mol dm-3 NaCl solutions containing five Na+ cations per A1 atom present in the solids. Procedure C (alkaline medium): The HM forms were exchanged five times for 24 h at 293 K in 0.1 mol dmP3 NaOH solutions containing one Naf cation per A1 atom present in the solids. After the ion-exchange step the solids were thoroughly washed and dried at 373 K overnight.F.RAATZ, E. FREUND AND C. MARCILLY 301 CHARACTERIZATION OF THE SOLIDS The characterization techniques used in the present study have been described in detail in Part 1.' However, the technique we used to follow the modifications of the chemical state of aluminium, namely A1 Ks X-ray fluorescence, is a non-conventional one and needs further comment. The energy of the A1 Kp X-ray fluorescence line depends on the coordination of aluminium, and the measurement of the energy shift of this line allows the detection of the presence of hexacoordinated aluminium species (AlV1) in modified 3 9 l2 In the present study the experiments were carried out with a PW1410 Philips spectrometer. The chemical shifts of the A1 Ks line were determined in the following manner: AE/ev = { E[Al Ks(sample)] - E[Al Kp(standard)]). Either SPNaM or LPNaM was used as standard for tetracoordinated aluminium (AIIV).A complete description of the experimental procedure can be found in ref. (12) and (13). Since the publication of Part 1 we have been able to show that measurements of AE give a reliable quantitative estimate of the AIIV/A1 ratios in modified s01ids.l~ The following formula is closely followed throughout: AIV1/A1 M AEIeV or A1IV/A1 z 1-AE/eV. RESULTS Both LPM and SPM have been studied. However, most of their physicochemical properties are identical. Hence we only report here the results concerning LPM which are significantly different from those of SPM (namely the benzene adsorption capacities) or which are of specific interest [namely the ion-exchange capacity for the large cation Pt(NH,)t+]. Before presenting our results three general comments have to be made. First, only traces of silicon and aluminium were found in the solutions for all the experiments.Secondly the crystallinity of the solids was not affected by the treatments in either NaCl or NaOH solutions. Thirdly it was not possible to prevent a decrease in the pH in the course of the ion-exchange steps under procedures A[Pt(NH,),2+] and B (Na+ ex NaCl), except by using infinite volumes of solutions. However, the same exchange levels were obtained for solids which had been preliminarily neutralized by either aqueous or gaseous ammonia.', This shows that for procedures A and B the pH evolution is not of primary importance. PLATINUM PROCEDURE A The (Pt/2Al) ratios obtained for the original non-modified SPNH,M and LPNH,M forms are identical and equal to 0.58 (fig.1). Thus since SP- and LP-NH,M have the same ion-exchange capacity, the restrictions which are present in the main channels of SPNH,M are unable to block the diffusion of the large cation Pt(NH,)t+. However, the Pt/2A1 ratios are far below the maximum value which could be achieved if all the SP and LP cationic sites (especially those located in the side pockets) were accessible to platinum. This point will be discussed later. With regard to the modified solids, the ion-exchange capacities of both the LPHM and SPHM forms never reach the value typical of the original NH,M forms (Pt/2Al = 0.58), even for solids which have been treated at a temperature as low as 673 K (fig.1). More precisely, calcination in the temperature range 673-873 K, which does not lead to severe framework dealumination (cf. fig. 5 in Part 1 and fig. 3 later) and has nearly no effect on the crystallinity of the HM forms (cf. fig. 3 in Part l), causes a drastic decrease in the ion-exchange levels. The Pt/2A1 ratios relative to solids treated at temperatures > 873 K are nearly equal to zero.302 ION EXCHANGE IN MORDENITE 0.5 E 0.4 Tl K Fig. 1. Calcination of SPNH,M and LPNH,M under procedure 2 (deep-bed-like) followed by ion exchange under procedure A [Pt(NH,)i+]. The left-hand scale (Pt/2Al) and the right-hand scale (Pt/2A1)* represent the ion-exchange capacities assuming all the cationic sites and 5/8 of the cationic sites, respectively, of mordenite are accessible to platinum.The dashed line represents the theoretical maximum ion-exchange capacity assuming only 5/8 of mordenite exchange sites are accessible to platinum. 0.58 is the experimental value for the original SPNH,M. Fig. 2. Calcination Na/AI: 0, 1.01 Om9! 0.8 c . 0.3 0.2 I L 0.11 1 I I L 1 1 673 773 873 973 1073 1173 1273 T/K of SPNH,M under procedure 2 (deep-bed-like) followed by ion exchange. procedure B (Na+ ex NaCl) and u, procedure C (-Naf ex NaOH). SODIUM PROCEDURE B: Na+ ex NaCl Procedure B leads qualitatively to the same phenomena as those observed in the case of procedure A. The solid treated at 673 K exhibits the highest exchange level without reaching the value typical of the original NH,M forms (Na/Al = 1).The number of sodium cations which can be re-introduced into the HM forms decreasesF. RAATZ, E. FREUND AND C. MARCILLY 2800 m 2750- - 5 '2 - +J 3 0 5-' P 2700- 2650- 303 - ori@nal SPNaM -_ - __ -- - - - - - - - - - - - - - - - - -- - n ''\\A "\\\ *\ 0, 1 I I I I 10 I TI K Fig. 3. Calcination of SPNH,M under procedure 2 (deep-bed-like) followed by ion exchange. A1 Ks chemical shift (in eV): (---------- ) after the calcination step; 0, after ion exchange under procedure B (Na+ ex NaCl) and m, after ion exchange under procedure C (Na+ ex NaOH). very rapidly as the temperature used for the calcination step increases, and it is particularly low for solids treated above 873 K (fig. 2). Whatever the calcination temperature of the original NH,M forms, the treatments in NaCl solutions cause structural modifications to the solids.(i) The A1 Kp X-ray fluorescence line shifts decrease slightly when compared with those measured after the calcination step (fig. 3). (ii) The three unit-cell parameters tend to reach values which are closer to those of the original non-modified NaM forms. This is illustrated in fig. 4, which presents the evolution of the volume of the unit cell.304 9.0 8.0 7.0 3 6.0- 3 - 5.0- u” 4.0- 3.0~- 2.0 1.0- Y z? ION EXCHANGE IN MORDENITE - - - - NaOH w rn TIK Fig. 5. Calcination of SPNH,M under procedure 2 (deep-bed-like) followed by ion exchange. Benzene adsorption capacity (wt % ) : (- - - - - - - ) after the calcination step; 0, after ion exchange under procedure B (Na+ ex NaC1) and ., after ion exchange under procedure C (Na+ ex NaOH).9.0 TIK Fig. 6. Calcination of LPNH,M under procedure 2 (deep-bed-like) followed by ion exchange. Benzene adsorption capacity (wt % ) : (- - - - - - - ) after the calcination step; 0, after ion exchange under B (Na+ ex NaCl) and S, after ion exchange under procedure C (Na+ ex NaOH). Note that the introduction of Na+ cations into the SPHM form does not cause pore-blocking in the solids, which acquired LPM-like adsorption properties in the course of the deep-bed calcination step (fig. 5 ) . This agrees with the results of Part 1 which established that the pore blocking of SPM cannot be due to the particular location of the cations. However, ion exchange induces a decrease in the benzene adsorption capacities of both SPM and LPM (fig.5 and 6). In addition, it is worthF. RAATZ, E. FREUND AND C. MARCILLY 305 noting that LPM pretreated at a low temperature (673 K) adsorbs much less benzene after the ion-exchange step than a non-modified LPNaM form (6.8 wt % ). PROCEDURE c: Na+ ex NaOH In contrast to procedure B, procedure C leads to the following results. (i) High (Na/Al) ratios, namely > ca. 0.70, for all the HM forms obtained through calcination below 1073 K of the original &H,M form (fig. 2).The exchange levels only drop when the structure has been seriously damaged by heating above 1073 K (see Part 1, fig. 3). (ii) A strong decrease in the amount of AIV1 species formed during the calcination step, except in the case of amorphous solids resulting from a heat treatment at 1273 K (fig.3). (iii) A pore blocking of SPM (fig. 5 ) and LPM (fig. 6). The extent of this pore blocking increases with the amounts of AIV1 present in the zeolite at the end of the calcination step (fig. 6). Surprisingly, however, the increase in the volume of the unit cell induced by the successive treatments in alkaline solutions is the same as for procedure B (fig. 4). DISCUSSION According to Part 1 of this series' the calcination of NH,M under deep-bed-like conditions has the following effects. (i) Below 1073 K it causes a dealumination of the skeleton with the subsequent formation of AIV1 species which can be present either as cationic AIV1 species localized in framework cationic sites or as neutral hydroxyaluminium complexes located in the channels.(ii) Above 1073 K, in addition to the phenomena mentioned above, there is a destruction of the crystalline structure with the formation of amorphous AIV1. If one assumes that cations can only be reintroduced into HM ex NH,M forms by means of a true ion-exchange mechanism, the exchange levels which can be achieved are mainly determined by the fraction of framework AIIV atoms (i.e. of cationic sites) remaining in the solid after the calcination step. However, the cationic AIV1 species which are likely to be present in the HM forms are not exchangeable since only traces of aluminium are found in the NaCl or NaOH solutions. To calculate any theoretical ion-exchange capacity, the number of framework AllV atoms has to be corrected for both the amount and the electric charge of these unchangeable cationic AIV1 species.Hence the uncorrected A1IV/Al ratios which can be deduced from physical methods (such as 27Al magic-angle spinning n.m.r., A1 Kp X-ray fluorescence etc.) should always be higher than the experimental exchange capacities, i.e. Na/A1 ratios, when unexchangeable cationic AIV1 species are present. In the following we show that the above statement is obeyed for treatments in acidic media (procedures A and B) but that it does not hold for treatments in alkaline solutions (procedure C). PROCEDURES A AND B : ACIDIC pH To compare the results of procedures A[Pt(NH,)t+] and B(Na+ ex NaCl) we need first to know the theoretical exchange capacity of mordenite for Pt(NH3);+. In platinum-ammine complexes, the Pt-NH, distance is typically > 2.0 which means that Pt(NH3);+ is too large to enter mordenite side pockets easily.Thus, sites I, I1 and I11 in the nomenclature of ref. (1 5 ) are not likely to be occupied by this cation. According to this hypothesis, only 5 / 8 of mordenite cationic sites are accessible to Pt(NH,)i+. This leads to a theoretical Pt/2Al ratio of 0.63, which agrees fairly well with the experimental value of 0.58 obtained for both LP and SP non-modified Pt(NH,),-NH,M forms. This preliminary consideration shows that the results of procedures A and B can be compared quantitatively provided that, in the case of306 ION EXCHANGE IN MORDENITE 0.3 0.2 0.1 --c- 0.9 1.0 AIrV/Al Fig. 7. Calcination of SPNH,M under procedure 2 (deep-bed-like) followed by ion exchange under procedure B (Na+ ex NaCl).Correlation between the Na/Al and the A1IV/Al ratios. The A1IV/A1 ratios are measured before (0) and after (W) the ion-exchange step. The calcination temperature was varied between 673 and 1273 K. The dashed lines are the theoretical curves calculated assuming (a) no AIV1 species occupy framework cationic sites, (b) all AIV1 species have a + 1 positive charge per A1 atom and (c) all AIV1 species have a +2 positive charge per A1 atom. platinum, the Pt/2Al ratios are expressed assuming only 5 / 8 of the A1 atoms are possible cationic sites. With this new scale for the Pt/2Al ratios, namely (Pt/2Al)*, which is reported on the right-hand side of fig. 1, it appears that procedure B, i.e. Na+ ex NaCl (fig. 2), and procedure A, i.e.Pt(NH3)i+ (fig. l), give similar though not exactly identical exchange levels. Thus in the following, we will limit the discussion to the results relative to procedure B (Na+ ex NaCl). To analyse our data quantitatively, we have plotted in fig. 7 for the case of SPM the Na/Al ratios against the A1IV/A1 ratios deduced from A1 K X-ray fluorescence measured before and after treatment in NaCl solutions. These treatments have been shown to cause a small increase in the A1IV/Al ratios (fig. 3). Thus the experimental points corresponding to a same original HM ex NH,M form in the new coordinates of fig. 7 are not exactly identical; but they are close enough as to align on the same experimental curve. For the moment we cannot explain the increase in the A1IV/Al ratios occurring during the ion-exchange step.However, this increase is quite small and we can safely interpret the main features of our results without taking into account this phenomenon. The main point which is evident from fig. 7 is the following: whatever the calcination temperature of the original NH,M form, the Na/Al ratios achieved after the treatment in acidic media of the resulting HM forms are always lower than the A1IV/A1 ratios. This means that the experimental ion-exchange capacities, namely the Na/Al ratios, are lower than the theoretical ion-exchange capacities calculated assuming no cationic sites are blocked by AIV1 species [fig. 7 (a)]. This result, combined with (i) the decrease in the observed pH in the course of treatment in solution and (ii) the fact that only chemical shifts (identical results are obtained for LPM).The Al* dl /A1 ratios have beenF. RAATZ, E. FREUND AND C. MARCILLY 307 traces of aluminium are detected in the solutions, strongly supports the following two statements: (i) in acidic media HM + Naiq % NaM + H:q or HM +gPt(NH,);+ % [gPt(NH3)4] M + H+ is the main reaction which accounts for Na+ [or Pt(NH,)i+] consumption and (ii) a variable number of the extra-lattice AIV1 species formed during calcination of the original NH4M forms occupy framework cationic sites and are unexchangeable in slightly acidic media (of course they can be extracted by a strong acid attack). It is impossible to obtain directly from A1 Kg X-ray fluorescence chemical shifts information on the amounts or the positive charges of the cationic AIV1 species.However, an indirect route is to compare the experimental Na/Al against AlrV/A1 plot with theoretical curves corresponding to particular models. In that regard, in addition to curve (a) (no AIV1 species are assumed to occupy cationic sites), we have reported in fig. 7 two other theoretical curves corresponding to the following arbitrary limiting cases. (i) All AIV1 species have one positive charge per A1 atom and occupy framework cationic sites, i.e. Na/Al = 2(AlrV/A1) - 1 [fig. 7 (b)]. (ii) All AIV1 species have two positive charges per A1 atom and occupy framework cationic sites, i.e. Na/Al = 3(AlrV/A1) - 2. The reasonable agreement between the experimental curve and curves (b) and (c) at least indicates that a large number of the AIV1 species formed in the course of calcination under deep-bed-like conditions at low temperatures, i.e.below 873 K (right-hand region of fig. 7), are indeed cationic. Such a mechanism could explain why very low dealumination levels cause a large decrease in the ion-exchange capacities (it is difficult to furnish quantitative predictions about the exact positive charge of the AIV1 species since the kind of data we are dealing with cannot be unambiguously fitted to a particular model). Furthermore, it appears from fig. 7 that under a deep- bed-like calcination procedure, temperatures > 873 K are needed to favour the formation of neutral hydroxyaluminium complexes. The increase in the volume of the unit cell resulting from the treatment of the HM forms in NaCl solution can hardly be caused just by H: + Naiq exchange, particularly for solids with very low Na/Al ratios.The treatment in solution itself induces modifications of the zeolite framework. This point obviously needs further investigation before it can be explained. Another effect of the treatment in NaCl solutions concerns the adsorption properties, which even in the case of LPM are seriously affected. In Part 1, we demonstrated that the cations cannot be the cause of SPM pore blocking. Thus even if Na+ cations diminish the solid void volumes, another mechanism is involved. In agreement with Barrer and Klin~wski,~ we propose that an evolution in solution of some of the AIV1 species occurs; probably a slow hydrolysis followed by a polym- erization which restricts the available void volumes.Unfortunately, such a process cannot be detected by A1 Ks X-ray fluorescence line-shift measurements. PROCEDURE C : BASIC pH The treatment of HM ex NH,M forms in alkaline solution leads to very high Na/A1 ratios. However, both a strong decrease in the amount of AIV1 species and pore blocking occur. This indicates that reactions other than a pure ion exchange HM + Na& + NaM + H&308 ION EXCHANGE IN MORDENI’IE Al’” /A1 Fig. 8. Calcination of SPNH,M under procedure 2 (deep-bed-like) followed by ion exchange under procedure C (Na+ ex NaOH). Correlation between the Na/A1 and A1IV/A1 ratios. The A1IV/Al ratios are measured before (0) and after (m) the ion-exchange step. The calcination temperature was varied between 673 and 1273 K.The dashed line is the theoretical curve which corresponds to the limiting case Na/A1 = AIIV/A1. are involved. As before, in order to analyse our results, we have reported in fig. 8 the Na/A1 ratio against the A1IV/Al ratio in the case of SPM (identical results are obtained for LPM). The two experimental curves correspond to the A1IV/A1 ratios determined before and after the treatment in NaOH solutions. Two important points emerge from the curves plotted in fig. 8. (i) First, the Na/Al ratios achieved after treatment in NaOH solution are much higher than the A1IV/A1 ratios measured before this treatment (i.e. after the calcination step). This result means that the experimental Na/Al ratios are much higher than the theoretical exchange capacities of the HM ex NH,M forms calculated assuming no unexchangeable AIV1 species occupy cationic sites (fig.8, dashed line). (ii) Secondly, concerning the A1IV/Al ratios measured after the treatment in alkaline solution, the following relation is approximately followed : Na/Al= A1IV/A1. Thus, as a result of the NaOH treatment, the number of AIIV atoms not only increases at the expense of the AIV1 species but is then nearly identical to the amount of Na+ cations reintroduced in the solids. These results demonstrate that the Na contents which are achieved under procedure C are approximately equal to the sum of two terms: term I , the number of framework aluminium atoms (A1Iv) remaining in the HM forms after the calcination step corresponding to the maximal theoretical exchange capacity of the HM forms, and term 2, the number of AIV1 species (cationic or neutral) transformed into A P species in the course of the treatments in NaOH solutions.To account for these results we propose the following mechanism. Sodium hydroxide reacts with nearly all the cationic AIV1 species and possibly with a variable part of the neutral hydroxylaluminium species to form, inside the channels, sodiumF. RAATZ, E. FREUND AND C. MARCILLY 309 aluminates, compounds in which A1 is tetracoordinated. Starting from the postulated A13+, A1(OH)2+ and Al(0H); cationic species,l0 we may have, for example, A13+ + 4NaOH % NaAl(OH), + 3NaS where the resulting Na+ cations and the NaAl(OH), species are located in framework cationic sites and in the channels, respectively.Other reactions leading to the formation of NaAlO, may also occur. In contrast to the mechanism published earlier by Barrer and Klin~wski,~ our scheme is able to explain why Na/Al ratios much higher than the theoretical ion-exchange capacities of the HM forms are obtained. Another mechanism which has been proposed for hydrogen Y zeolites treated in NaOH solutions is the reinsertion, in the framework tetrahedral position, of AIV1 atoms.* However, from recent 29S magic-angle spinning n.m.r. experiments, Engelhardt et a2.l' concluded that such a process does not occur. Although we did not use n.m.r. spectroscopy in this study, some indirect experimental evidence supports the assumption that framework reinsertion plays a minor role in our system.The strong increase in the fraction of A1IV species observed in the course of the alkaline treatment never leads to a clear recrystallization process, as shown by the X-ray diffraction data. In fact, the changes in the unit-cell parameters (fig. 4) are the same as in the case of a treatment in NaCl solutions (procedure B). This result is surprising and needs further physical investigation in order to be explained. However, even if X-ray diffraction results cannot give a definitive answer, framework reinsertion of AIV1 should not in any case cause a blocking of the porosity. Thus the formation of sodium aluminates in the channels is likely to be the major reaction which accounts for the transformation of AIV1 into A P . On the basis of this single reaction, we are able to explain the two most important effects of procedure C : (i) Na/Al ratios higher than the theoretical ion-exchange capacity of the HM forms, with a final balance at the end of the NaOH treatments between the number of AIIV atoms and Na+ cations reinserted in the solid, and (ii) severe blocking of the porosity provided that AIV1 species are present in the HM forms.CONCLUSION The framework dealumination which occurs in the course of the calcination of NH,M forms causes a decrease in the number of cationic sites of the resulting HM forms. In the case of heat treatment under deep-bed-like conditions below 873 K, a large part of the extra-lattice AIV1 species which are then formed are likely to bear positive charges and to neutralize some of the remaining cationic sites.The number of cations which can be reintroduced in these modified HM ex NH,M forms strongly depends on the pH of the solutions in which the 'exchanges' are conducted. In acid solutions the cationic AIV1 species are not exchangeable and the Na/Al ratios which can be achieved are mainly determined by the replacement by Na+ of H+ located in framework cationic sites. In alkaline solutions reactions involving AIV1 species are predominant. Most of the cationic AIV1 species are transformed inside the mordenite channels into A P species and subsequently a complete blocking of the pores occurs. The most likely mechanism which could explain these phenomena is the formation of sodium aluminates in the pores.310 ION EXCHANGE IN MORDENITE F. Raatz, E. Freund and C. Marcilly, J. Chem. SOC., Faraday Trans. 1, 1983, 79, 2299. G. H. Kuhl, J. Phys. Chem. Solids, 1977, 38, 1259. R. L. Patton, E. M. Flanigen, L. G. Dowel1 and D. E. Passoja, Am. Chem. SOC. Symp. Ser., 1977, 40, 64. C. A. Fyfe, G. C. Gobbi, J. S. Hartman, J. Klinowski and J. M. Thomas, J. Phys. Chem., 1982, 86, 1247. J. B. Nagy, Z. Gabelica, G. Debras, G. Derouane, J. P. Gilson and P. A. Jacobs, Zeolites, 1984, 4, 133. R. E. Wasylishen and C. A. Fyfe, Annu. Rep. NMR Spectrosc., 1982, 12, 1 and references therein. ' R. M. Barrer and J. Klinowski, J. Chem. SOC., Faraday Trans. 1, 1975, 71, 690. D. W. Breck and G. W. Skeels, Proc. 6th Int. Congr. Catal., 1976, 2, 645. D. W. Breck and G. W. Skeels, 5th Int. Conf. Zeolites, Napoli, 1980, p. 335. lo G. H. Kuhl, Am. Chem. SOC. Symp. Ser., 1977, 40, 96. l 1 G. T. Kerr, J. Catal., 1969, 15, 200. l2 E. Freund, C. Marcilly, F. Raatz and G. Thomas, Rev. Inst. Fr. Petr., 1982, 37, 371. l 3 F. Raatz, Ph.D. Thesis (ENSPM, Technip, Paris, 1982). l4 M. Atoji, J. Richardson and R. Rundle, J. Am. Cham. SOC., 1957,79, 3017. l5 W. M. Mortier, J. J. Pluth and J. V. Smith, Naturul Zeolites: Occurrences, Properties and Use, ed. L. B. Sand and F. A. Mumpton (Pergamon Press, Oxford, 1978), p. 53. l6 G. Engelhardt, U. Lhose, M. Magi and E. Lippmaa, Stud. Surf. Sci. Catal., 1984, 18, 23. (PAPER 4/203)

ISSN:0300-9599    

DOI:10.1039/F19858100299

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985,81, 311-319 35Cl Nuclear Quadrupole Resonance Studies of Hydrogen Bonding in Solid Complexes of Chlorobenzoic Acids with Amines BY EUGENIUSZ GRECH,~ JERZY KALENIK AND LUCJAN SOBCZYK* Institute of Chemistry, University of Wroclaw, 50-383 Wroclaw, Poland Received 9th February, 1984 35Cl nuclear quadrupole resonance studies of hydrogen-bonded adducts containing o-, rn- and p-chlorobenzoic acid and 2,6-dichlorobenzoic acid have been carried out. Average frequencies are correlated with ApK, values in terms of the proton-transfer model. The results obtained for various proton donors are compared and discussed taking into account both intra- and inter-molecular effects. Numerous spectroscopic studies of the complexes of chlorobenzoic acids with amines carried out in solution1* have indicated the existence of a proton-transfer tautomeric equilibrium.Dipole-moment measurements3 in non-polar solvents have shown that in these systems an inversion region for ApK, [defined by pK,(BH+)- pK,(AH)], corresponding to a stepwise increase in the hydrogen-bond polarity, Ap, occurs and can be interpreted in terms of a shift of the proton-transfer equilibrium The proton-transfer equilibrium constant, KpT, may be related, according to the Huyskens and Zeegers-Huyskens to ApK, via the equation where and C' are constants. The constant ( expresses the coupling of hydrogen-bonded (HB) and proton-transfer (PT) states, whereas the constant C' depends on the environment and is related to the value of ApK, for which KPT = 1 (the inversion point).Infrared investigations of crystalline complexes of benzoic acids with pyridines5 have revealed the presence of such an inversion region in the solid state also. The purpose of this work was to investigate the influence of changes in the charge-density distribution of hydrogen bonds for crystalline complexes of chlorobenzoic acids with nitrogen bases of various strengths using the 35Cl n.q.r. technique. From earlier investigations on solid complexes with trichloroacetic pentachlorophenoP and 2,6-dichloro-4-nitropheno19 it follows that the complexation process is accompanied by a decrease in the average n.q.r. frequencies in relation to the average frequency of the pure proton donor. This means that the n.q.r. frequencies of nuclei in chlorine atoms which do not participate directly in hydrogen-bond formation can be applied successfully as a probe in order to examine the changes in charge-density distribution which take place in hydrogen bonds.t Permanent address : Institute of Fundamental Chemistry, Technical University, 7 1-065 Szczecin, Poland. 31 1312 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING Table 1. 35Cl quadrupole resonance frequencies and pK, values of bases for the 2,6-dichlorobenzoates no. of reson- ance PK," no. amine v*.q. r . /MHz lines v/MHz (amine) a b 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2,6-dichlorobenzoic acid potassium 3-cyanopyridine 4-cyanopyridine 2,6-dichlorobenzoate 4-form ylp yridine aniline quinoline isoquinoline 4-methylquinoline 2-methylp yridine 4-met hylp yridine 3 ,Cdimethylpyridine 4-methylmorpholine morpholine triethylenediamine 4-dimethylaminop yridine tributylamine ethylpiperidine trie t hy lamine piperidine dibutylamineC 35.622; 35.683; 36.01 8 ; 36.067 34.632; 34.812 35.508; 35.682 35.542; 35.662; 35.730; 35.785 35.345; 35.821 35.388; 35.774 34.971 ; 35.788 34.885; 35.093; 35.554; 36.058 35.259; 35.768 35.1 15; 35.331 34.981 ; 35.245 35.075 34.483; 35.401 34.694; 35.266 34.888; 35.004 34.862; 35.305 34.893 34.828; 35.109 34.669; 34.947 35.023 35.218; 35.274 4 2 2 4 2 2 2 4 2 2 2 1 2 2 2 2 1 2 2 1 2 35.85 34.72 35.60 35.68 35.58 35.58 35.38 35.40 35.51 35.22 35.1 1 35.08 34.94 34.98 34.95 35.08 34.89 34.97 34.81 35.02 35.25 1.59b - 1.35 1.86 4.53 4.6 1 4.93 5.40 5.59 5.94 6.03 6.48 7.38 8.49 8.82 9.61 9.93 10.45 10.75 11.20 11.25 Ref.(10). Ref. (1 1). This complex has not been used in further calculations. EXPERIMENTAL Solid complexes were prepared by crystallization from equimolar chlorobenzoic acid + amine solutions in acetonitrile. The complex compositions were determined from chlorine analysis. Measurements of the resonance frequencies were performed at liquid-nitrogen temperature (77 K) on an ISSh-1-13M pulse spectrometer. Since not all the complexes with a given chlorobenzoic acid exhibited the same number of resonance lines, the average resonance frequencies v ~ . ~ . ~ . , being arithmetic means of the all frequencies measured, were used in the correlations. RESULTS AND DISCUSSION The resonance frequencies of the 1 : 1 complexes with 2,6-dichlorobenzoic acid and with 0-, m- and p-chlorobenzoic acid, together with the frequencies for the pure acids and their salts (mainly those of potassium) are summarized in tables 1-4.Also given are the average frequencies v ~ . ~ . ~ . and the pK, values for amines as components of the hydrogen-bonded adducts investigated.E. GRECH, J. KALENIK AND L. SOBCZYK 313 Table 2. 35Cl quadrupole resonance frequencies and pK, values of bases for the o-chlorobenzoates no. amine no. of reson- ance PKa" vn. 9. r. /MHz lines o/MHz (armne) a b 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 o-chlorobenzoic acid ammonium o-chlorobenzoate" 3-c yanopyridine 4-c y anopyridine 4-acetylpyridine quinoline isoquinoline 4-met hylquinoline 3-rnethylisoquinoline 4-methylpyridine 3,4-dirnethylpyridine 2,6-dimethylpyridine 2-amino-4-methylpyridine triethylenediamine 4-aminop yridine 4-amino-2-met h y lquinolined 4-dimethylaminopyridine piperidine dibutylaminee 36.312 35.05; 35.18; 35.27 35.983; 361060 35.825 35.634; 36.114 35.587; 36.312 35.484 35.639; 35.998 35.876 35.697 35.675 35.505; 35.996 35.453 34.89 1 34.344 35.797 34.579 34.599; 34.719 34.59 1 3 2 1 2 2 1 2 1 1 1 2 1 1 1 1 1 2 1 36.31 35.17 36.02 35.83 35.87 35.95 35.48 35.82 35.88 35.70 35.68 35.75 35.45 34.89 34.34 35.80 34.58 34.66 34.59 2.92b - 1.35 1.86 3.51 4.93 5.40 5.59 5.64 6.03 6.48 6.64 7.38 8.82 9.12 9.42 9.61 11.20 11.25 a Ref.(10). of f 10 kHz. Ref. (12). Ref. (13). This complex has not been used in further caclulations. Since the resonance line was very broad the n.q.r. frequency was measured with an accuracy The dependence of v ~ ~ ~ .~ . on ApK, for 2,6-dichlorobenzoates (fig. 1) and for o-chlorobenzoates (fig. 2) may be interpreted in terms of a formal proton-transfer model (as previously*? g, such that Vn.q.r. = XHB VHB +XPT VPT (2) where xHB and xPT are the mole fractions of the complexes with no proton transfer and with full proton transfer, while vHB and v,, are average resonance frequencies of two boundary forms of the hydrogen bonding. (This means that the observed average resonance frequency v ~ . ~ . ~ . is a linear function of the proton-transfer fraction.) From the v ~ . ~ . ~ . and pK, values gathered in table 1 (for the 2,6-dichlorobenzoates) and in table 2 (for the o-chlorobenzoates), the parameters in eqn (1) and (2) have been estimated using the generalized least-squares method. Thus the values obtained for <, C', ijHB and vPT are 0.92, -3.86, 35.62 MHz and 34.95 MHz for the 2,6- dichlorobenzoates and 0.82, -4.00, 35.83 MHz and 34.95 MHz for the o-chlorobenzoates.The experimental points and the plots of v ~ . ~ . ~ . against ApK, calculated for the estimated values of c, C', vHB and vPT are presented in fig. 1 (2,6-dichlorobenzoates) and 2 (o-chlorobenzoates). The average frequencies of the pure acids and their salts are also shown. The scattering of experimental points visible in fig. 1 and 2 is presumably314 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING Table 3. 35Cl quadrupole resonance frequencies and pKa values of bases for the rn-chlorobenzoates no. no. of reson- ance PKaa amine vn .q .r . /MHz lines Y/MHz (amine) a b rn-chlorobenzoic acid potassium rn-chlorobenzoate" 4-cyanopyridine quinoline isoquinoline 4-methy lpyridine 3,4-dimethylpyridine 2-aminopyridine 2-amino-4-methylpyridine morpholine piperidine 35.232 34.74 34.710 34.242 34.607 34.767 34.985 34.589 35.116 34.61 7 34.51 3; 1 1 1 1 1 1 1 1 1 1 34.823 2 35.23 34.74 34.71 34.24 34.61 34.77 34.99 34.59 35.12 34.62 34.67 3.83b - 1.86 4.93 5.40 6.03 6.48 6.7 1 7.38 8.49 11.20 a Ref. (10). Ref. (12). " Ref. (13). Table 4. 35Cl quadrupole resonance frequencies and pK, values of bases for the p-chlorobenzoates no. amine no. of reson- ance PKaa vn. q. r. /MHz lines v/MHz (amine) a p-chlorobenzoic acid 34.673 b potassium 35.48 1 2-amino-4-methylpyridine 34.758 2 4-amino-2-met hylquinoline 34.302 3 piperidine 34.571 p-chlorobenzoate" 1 34.67 3.98b 1 35.48 - 1 34.76 7.38 1 34.30 9.42 1 34.57 11.20 a Ref.(10). Ref. (12). Ref. (13). due to contributions from lattice effects to the effective electric-field gradient on the quadrupole nuclei, which are not the same for all compounds. These contributions are dependent on the arrangement of molecules in the unit cell, which unfortunately is not known. The procedure of averaging the resonance frequencies allows us to eliminate some of the lattice effects. Some influence on the scattering of the experimental points may also be caused by differences in the conformations of the complexed chlorobenzoic acid molecules, these being characterized by a dihedral angle a between the planes of the carboxylic group and aromatic ring.The magnitude of the dihedral angle a is affected by two factors: (i) the interaction between the carboxylic group and substituents in ortho positions and (ii) the interaction between the neighbouring molecules (especially via hydrogen36.0 t 3 4.5- E. GRECH, J. KALENIK AND L. SOBCZYK -a 15 17 0 13 315 -b 0 2 4 6 8 10 A PK, 3 4 . 5 1 : : : : : : : : : : : : : ~ Fig. 1. Plot of pnn.s.r. against ApK, for the 2,6-dichlorobenzoates; for notation see table 1. bonds). The first of these factors determines the minimal value of a in the case when no specific interaction with the groups in ortho positions takes place. It can be easily calculated using the van der Waals radii.14 In the case of o-chlorobenzoic and 2,6-dichlorobenzoic acids the calculated values of a are 10 and 45", re~pective1y.l~ The second factor, connected with the arrangement of molecules in the crystalline lattice, leads to an additional enhancement of this angle.Thus in pure o-chlorobenzoic acid a is equal to 13.70,16 while e.g. in 2-chloro-5-nitrobenzoic acid it reaches 2 3 . O O . l ' Moreover, note that when there is only one substituent ortho to the carboxylic group the C(0)-OH group is arranged trans to that substituent.l6? l7 The phenomenon of a differentiation in the values of the dihedral angle a is particularly well illustrated for 2,6-dichlorobenzoates, where in some complexes both316 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING chlorine atoms are equivalent (one resonance line is observed), which means that the carboxylic group is perpendicular to the plane of the aromatic ring.However, in other 2,6-dichlorobenzoic acid complexes both chlorine atoms are not equivalent and two or four resonance lines are visible. This clearly suggests that the carboxylic group is not perpendicular to the plane of the aromatic ring. The particularly large deviations visible for point (1 9) in fig. 1 and for point (14) in fig. 2 are probably due to either steric effects (4-amino-2-methylquinoline-o- chlorobenzoate) or the formation of additional hydrogen bonds (dibutylamine- 2,6-dichlorobenzoate). Note that the calculated values of the parameters 5 and C' are similar for both acids (this has also been observed for phenol^).^ The position of the inversion point determined from eqn (1) (where the degree of proton transfer is 50 % ) in the ApK, scale is 4.2 for the 2,6-dichlorobenzoates, while for the o-chlorobenzoates it is 4.9.The above values are close to the value of 3.75 estimated from the i.r. investigations of crystalline complexes of benzoic acids with pyridine~.~ The values of the parameters vHB and v,, calculated for the 2,6-dichlorobenzoates and the o-chlorobenzoates differ from those corresponding to average resonance frequencies of the pure acids and their salts. The differences between vHB and the average resonance frequencies of the pure acids are probably caused by the fact that molecules of the pure acids in an associated form (most frequently dimersls) participate in 0-H- - -0 hydrogen bonding, differing greatly from the formation of O-H...N bonds in complexes without proton transfer.On the other hand, the discrepancies between v,, and the average resonance frequencies of the salts mainly result from the fact that, unlike the hydrogen-bonded proton-transfer complexes 0-. --H-N+, in salts one must deal with electric fields arising from ions which are a source of additional resonance-frequency shifts.lg The magnitudes and directions of these shifts depend on the arrangement of ions in the unit cell. Therefore the method of determining the degree of proton transfer in crystalline hydrogen-bonded complexes suggested by Chihara and NakamuraZ0 and based on the assumption that vHB is equal to the average resonance frequency of the pure acid and v,, is equal to the average resonance frequency of its salt, could lead to incorrect results.In the case where the Cl atoms are sufficiently far from the hydrogen bond the shifts in the average resonance frequencies caused by changes in the hydrogen-bond polarity are small in comparison with the shifts caused by the lattice effects, and no correlation between v ~ . ~ . ~ . and ApK, is observed. This is clearly visible in the m-chlorobenzoates, the results for which are shown in fig. 3. The scatter of the experimental points in fig. 3, more distinct than those in fig. 1 and 2, may be caused additionally by two possible conformations of the carboxylic group in relation to the chlorine atom in the meta position. Similarly no correlation can be expected forp-chlorobenzoic acid as a proton donor (see table 4). The resonance frequency of purep-chlorobenzoic acid is even lower than that of its potassium salt, which can be explained by the presence of electric fields arising from the ions19 which, when the difference between vHB and vPT is not high, can increase the resonance frequency of the salt to a value above that of the pure acid.Lynch et aZ.13 have tried to explain this anomaly in terms of the mesomeric effect in p-chlorobenzoic acid. If this were the case, the average resonance frequencies of complexes containing the ionic form of the hydrogen bond, where the carboxylic group is almost entirely ionized, should be closer to the resonance frequency of potassium p-chlorobenzoate than to that of the pure acid. It thus seems that the role of the mesomeric effect in p-chlorobenzoic acid has been overestimated by Lynch et aZ.13 Let us now try to compare the results of n.q.r.investigations for five series of theE. GRECH, J. KALENIK AND L. SOBCZYK 35.25 35.00- 5 5 34.75- I> 3 4.50 34.25.- 317 .- .- 0 1 7 0 5 0 0 4 0 0 8 3 6 2 0 -a -b 0 9 -2 0 2 4 6 8 34.00 A PKa Fig. 3. Plot of v ~ . ~ . ~ . against ApK, for the rn-chlorobenzoates; for notation see table 3. A PKa Fig. 4. Calculated xpT values for various series of hydrogen-bonded adducts plotted against ApK,; for notation see table 5. complexes containing trichloroacetic acid,’ pentachlorophenol,g 2,6-dichloro-4- nitr~phenol,~ 2,6-dichlorobenzoic acid (this work) and o-chlorobenzoic acid (this work). We describe the behaviour of all the complexes using a generalized scheme of the proton-transfer equilibrium model, although we have already stressed’ that a quantitative description of the resonance-frequency changes can also be expressed in terms of the Mulliken theory by using a and b coefficients, which are in turn connected to the contribution of non-polar and dative states.21 In fig.4 are presented plots expressing the dependence of xpT on ApK, for the series of complexes under consideration. The parameters of eqn (1) describing these curves318 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING Table 5. Values of < and C' in eqn ( 1 ) and ApK, values for which KPT = 1 for a series of complexes with various proton donors no. proton donor 1 trichloroacetic acid" 2 2,6-dichloro-4-nitrophenolb 3 pentachlorophenolc 4 2,6-dichlorobenzoic acidd 5 o-chlorobenzoic acidd 0.12 0.04 -0.4 0.75 - 0.62 0.8 0.76 -0.91 1.2 0.92 - 3.86 4.2 0.82 - 4.00 4.9 a Values of the parameters calculated from data taken from ref.(7). This work. Ref. (9). Ref. (8). are listed in table 5. The ApK, values at which 50% proton transfer takes place are also included. We note an unusual broadening (a small value of r) of the curve for the complexes of trichloroacetic acid. A similar broadening has been found when describing the dipole-moment dependences for complexes of carboxylic acids with trieth~lamine.~ Undoubtedly only the chemical properties of aliphatic carboxylic acids are of importance here since complexes of aromatic acids behave normally, a fact confirmed by the results obtained in this work. Attempts to explain this anomaly may seem premature since we have no information about the configuration of the complexes formed in the crystalline state.On the other hand, it seems easier to explain the large difference in the position of the inversion point (at which xPT z 1/2) for the phenol derivatives and benzoic acids. The essential factor here is the additional contribution of the mesomeric effect in phenol complexes, which does not play a prominent role in benzoic acid complexes. In conclusion, measurements of the n.q.r. frequencies of hydrogen-bonded complexes can serve as an useful indicator of the inversion region of donor-acceptor properties, where a stepwise charge rearrangement and proton-transfer equilibrium can be anticipated. Knowledge of this region seems to be important since hydrogen-bonded complexes from this region are characterized by an unusual potential for proton motion.We thank a referee for helpful remarks. 1 2 3 4 5 6 7 8 9 10 11 12 G. M. Barrow, J. Am. Chem. Soc., 1956, 78, 5802. J. Nasielski and E. Vander Donckt, Spectrochim. Acta, 1963, 19, 1989. L. Sobczyk and Z. Pawelka, J . Chem. Soc., Faraday Trans. I , 1974,70, 832. P. Huyskens and Th. Zeegers-Huyskens, J . Chim. Phys., 1964,61, 81. S. L. Johnson and K. A. Rumon, J . Phys. Chem., 1965,69, 74. 0. Kh. Poleshchuk, Yu. K. Maksyutin, 0. F. Sychev, K. K. Koshelev and I. G. Orlov, Zzuest. Akad. Nauk SSSR, Ser. Khim., 1975, 6, 143 1 . J. Pietrzak, B. Nogaj, Z. Dega-Szafran and M. Szafran, Acta Phys. Polon. A, 1977, 52, 779. E. Grech, J. Kalenik and L. Sobczyk, J. Chem. SOC., Faraday Trans. I , 1979, 75, 1587. E. Grech, J. Kalenik, Z. Malarski and L. Sobczyk, J. Chern. Soc., Faraday Trans. I , 1983, 79, 2005. D. D. Perrin, Dissociation Constants of Organic Bases in Aqueous Solution (Butterworths, London, 1965; supplement, 1972). J. F. Dippy, S. R. C. Hughes and L. G. Bray, J . Chem. SOC., 1959, 1717. G. Kortiim, W. Vogel and K. Andrussow, Dissociation Constants of Organic Acids in Aqueous Solution (Butterworths, London, 1961).E. GRECH, J. KALENIK AND L. SOBCZYK 319 l 3 R. J. Lynch, T. C. Waddington, T. A. OShea and J. A. S. Smith, J. Chem. SOC., Faraday Trans. 14' H. A. Staab, Einfiihrung in die theoretische organische Chemie (Verlag Chemie, Weinheim, 1959). l5 M. K. Chantooni Jr and I. M. Kolthoff, J . Phys. Chem., 1974,78, 839. l 6 G. Ferguson and G. A. Sim, Acta Crystallogr., 1961, 14, 1262. l7 G. Ferguson and G. A. Sim, J. Chem. SOC., 1962, 1767. L. Leiserowitz, Acta Crystallogr., Sect. B, 1976, 32, 775. lo C. H. Townes and B. P. Dailey, J. Chem. Phys.,. 1952, 20, 35. 2o H. Chihara and N. Nakamura, Bull. Chem. Soc. Jpn, 1971, 44, 1980. 21 R. S. Mulliken, J. Am. Chem. SOC., 1952, 74, 81 1. 1976,72, 1980. 2, (PAPER 4/235)

ISSN:0300-9599    

DOI:10.1039/F19858100311

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1985, 81, 321-333 X-Ray Scattering Structural Investigation of Pt and Pt-Sn Catalysts Supported on Nylon BY GIORGIO Cocco* AND STEFANO ENZO Facolta di Chimica Industriale dell’ Universita, Calle Larga S . Marta 2137, 30 123 Venezia, Italy AND SIGNORINO GALVAGNO AND ZBIGNIEW POLTARZEWSKI Istituto C.N.R., Trasformazione e Accumulo Energia, Via Salita S. Lucia sopra Contesse 39, 9801 3 Pistunina-Messina, Italy AND ROSARIO PIETROPAOLO Istituto di Chimica Industriale, Facolta di Scienze, Universita di Messina, Via dei Verdi, 98100 Messina, Italy Received 26th March, 1984 The physical properties of Pt and Pt-Sn/Nylon catalysts have been determined using X-ray scattering techniques (SAXS and WAXS). Two different structural conditions of Pt/Nylon were inferred from the characterization.A large amount of the supported metal phase, ca. 100 A in size, retains the Pt crystalline structure and appears to be embedded in the Nylon. A second fraction with a strong metal-ligand interaction is also present. Addition of tin to the Pt/Nylon system does not change the observed structural framework, although a Pt-Sn solid solution appears with a decrease in the size of the metallic phase. Noble metals supported on polyamides exhibit a surprisingly high selectivity for the partial hydrogenation of unsaturated compounds, e.g. in the hydrogenation of benzene, cyclohexene is selectively formed, in contrast to results obtained using conventional platinum or palladium catalysts, which only lead to cy~lohexane.l-~ In a recent study of the hydrogenation of propene it was discovered that the reactivity of platinum towards the C=C double bond is decreased by interaction with the polymeric support.Moreover, on Pt/Nylon, the higher order of reaction with respect to olefin indicated a weaker metal-olefin bond, suggesting the presence of electron- deficient platinum sites.5 It has also been found that the addition of Sn to Pt/Nylon causes a decrease in the catalytic activity and a further increase in reaction order with respect to propene. This suggests the presence on the bimetallic system of a larger number of electron-deficient platinum sites stabilized by interaction with tin. Note that Sn is known to form stable complexes with Pt11.6 In the last few years the Pt-Sn system has received special attention, mainly because of the better performance of Pt-Sn/Al,O, as a reforming catalyst.’-ll The main advantages are higher selectivity towards high-octane-number products and improved stability.To account for these peculiar features, different explanations have been forwarded : formation of Pt-Sn alloy particles,’ electronic interaction between Pt and SnI1 ions,8 influence of tin in the sintering of platinum crystallite~,~~-~~ suppression of surface carbiding,15 etc. It could be argued that the role of tin is still unclear and even Pt-Sn alloying remains an open question. It must also be noted that most of the information is available for unsupported Pt-Sn or Pt-Sn/A1,O3l6 and, with regard 32 1322 X-RAY STUDY OF Pt AND Pt-Sn CATALYSTS ON NYLON Table 1.Amounts of metal involved in the preparation of Pt-Sn/Nylon catalyst Pt Sn catalysta (wt % ) (wt ”/, ) RP9/ 1 1.07 0.062 RP7/3 1.12 0.16 RP3/2 0.99 0.53 RPl/l 1.31 0.87 RP 1 /O 0.94 - RP2/3 0.82 1.11 a The numbers describe the nominal Pt/Sn ratio. to this paper concerning Pt and Pt-Sn on Nylon, it is worth pointing out that the nature and composition of bimetallic clusters are strongly affected by the kind of support used. 17-19 It was therefore felt that knowledge of the microstructural properties is mandatory for further interpretative work as well as for the industrial exploitation of this system, which is still the ultimate goal of our research. Accordingly, we report here a detailed X-ray characterization of several Pt-Sn/Nylon catalysts over a broad range of compositions. In a forthcoming paper the results on selective hydrogenation reactions over Pt-Sn/Nylon will be presented.EXPERIMENTAL CATALYST PREPARATION Pt/Nylon and Pt-Sn/Nylon samples were prepared by impregnation of Nylon powder, under N,, with a C,H,OH solution of H2PtC1, or H,PtCl, and SnCl,, using appropriate concentra- tions of the metals. An excess of solvent was removed at room temperature by evaporation under a flow of N, and the resulting material was dried in vacuo at 313 K. Reduced samples were obtained by bubbling hydrogen into a slurry of catalyst in C,H,OH at 318 K. Reduction was accomplished in the liquid phase in order to use lower reduction temperatures: in the gas phase a temperature > 373 K was necessary. Nylon 66 powder, used as support, was obtained by grinding pellets of commercial Nylon 66 (SNIAMID) at liquid-nitrogen temperature. The surface area of the powder was < 1 m2 g-l.Chemical analyses of the catalysts, obtained by atomic absorption, are reported in table 1. X-RAY TECHNIQUE WIDE-ANGLE X-RAY SCATTERING (WAXS) The details of the approach used have been outlined in previous 21 We carried out two sets of measurements with wide-angle Bragg-Brentano geometry in the reflection mode, using the characteristic radiation lines of Mo and Cu, respectively. The former was employed to explore a wider portion of the reciprocal space, whereas a careful analysis of the broadening of the diffraction peaks was carried out on data taken with Cu Ka radiation. A Ni filter, Soller slits, a graphite focussing crystal as a monochromator and a proportional counter with pulse-amplitude discrimination were employed in this case.The angular ranges through the peaks were scanned stepwisely in the preset count mode, providing that at least 4 x lo4 counts were accumulated at each point. We observed noticeable line broadening in the patterns, caused by various sources, namely instrumentation, particle size and strain.22 Some of the peaks overlapped and, because of the low metal loading, intensities scattered by the organic support further smeared the line shapes.G . cocco et al. 323 Under these unfavourable conditions we devised a suitable procedure for background and overlapping-peak correction in order to perform a Fourier analysis of line profiles according to the well known Warren-Averbach method.23 First, scattering by support materials was subtracted from the pattern of the reduced catalysts.In doing this, care was taken to maintain constancy of experimental conditions and statistical counting.20 Subsequently, overlapping peaks were resolved by a curve-fitting routine based on a least-squares procedure: for this we used pseudo-Voigt (pV) or Pearson VII (PVII) functions, which have been demonstrated to be among the Provision was made for the program taking into account the a, a, splitting which may be a source of broadening. The separated peaks were subjected to the usual Warren-Averbach analysis capable of giving information on microstructural parameters such as mean size values (surface or volume weighted), percentage exposed and presence or absence of strain.In fact, after Stokes’ correctionz5 the resultant Fourier coefficients are the product of two terms, one depending on a form factor and the other on strain. While the former does not depend on the order of reflection which is examined, the latter does, so it is possible to evaluate and separate the two effects when two or more multiple-order series are available. The general approach has been detailed in ref. (21). The crystallite size-distribution function p,(D) may be obtained from size coefficients AS(L) according to the equationz6 where D is the representative size of a crystallite. From the size distribution of crystallites, normalized so that J; P,(D) dD = 1, average size dimensions may be obtained.SMALL-ANGLE X-RAY SCATTERING (SAXS) SAXS analysis was carried out with a Kratky ‘compact’ camera.27 The primary-beam intensity was determined by the moving-slit method.28 The approach used has been described else- ~ h e r e . , ~ - ~ l Briefly, smeared intensity I(h) (where h is the modulus of diffracting vector h = 47z sin O/A, with 8 the Bragg angle and A the wavelength) may be inverted, under particular assumptions to be satisfied in reality, through a Titchmarsh transform according to : cc jm [K-h3Z(h)] [2J0(hR) + (hR - 3/hR) J,(hR)] dh (2) where D,(R) is the volume distribution function of assumed spherical scattering heterogeneities whose diameter is R, K = lim, ~ h3Z(h) and J, and J1 are first-kind Bessel functions of order zero and one, respectively. Moreover, other representative parameters are available by directly analysing the experimental intensity curves.32 RESULTS A central issue in the characterization of supported bimetallic catalysts concerns alloy formation. For this, preliminary tests performed on the samples before reduction do not display crystalline forms, the scattering functions behaving similarly to the pure Nylon pattern (fig.1). The Mo Ka spectra relevant to the reduced catalysts (fig. 2) permit us to rule out the formation of tin-enriched Pt-Sn compounds. In this regard, the intermediate phases PtSn, Pt,Sn,, PtSn, and a further Sn-rich phase, namely Pt,Sn, or PtSn,, have been reported to exist with an hexagonal or orthorhombic ar~angement.,~ However, from the above patterns a f.c.c. structural arrangement is unequivocally ascertained even for samples with rapid disappearance of crystalline features, which is the case for higher tin loadings. The absence of crystalline Sn, which is tetragonal, may also be inferred from analysis of the angular intensity distribution.324 X-RAY STUDY OF Pt AND Pt-Sn CATALYSTS ON NYLON 45 4 0 2e 35 Fig.1. Diffraction patterns in 28range 34-48" using Cu Ka radiation for non-impregnated Nylon (upper curve) and after impregnation but prior to activation (lower curve). The similarity of the patterns suggests that, during impregnation, no modification of the structure of the support, on an atomic scale, occurs. A continuous line has been traced between the data points, obtained by smoothing the two curves with orthogonal polynomials.x c c P) c, c, .d c, ." ." c- 2 RP1/0 RP9/ 1 RP7/3 RP3/2 R P l / l 20 3 0 40 50 29 Fig. 2. Experimental diffracted intensities using Mo Ka radiation. Spectra display the essential feature of the f.c.c. crystalline system for the Pt/Nylon sample as well as for the Pt-Sn catalysts. An f.c.c. structure belongs to the remaining compound reported in the Pt-Sn equilibrium diagram, i.e. the Pt,Sn intermediate phase, for which a lattice parameter a, = 4.01 A has been To examine this possibility a careful evaluation of lattice parameters a, was therefore carried out using Cu Ka radiation. The usual plot of d.\/(h2 + k2 + P ) against cot 8hkl cos 8hkl is reported in fig. 3. Owing to the impossibility of precisely locating higher-order peaks for very dispersed samples, the first five reflections were considered for the sake of consistency. a, for sample RP 1 / O turns out to be in close agreement with that of pure Pt, whereas a systematic expansion up to 3.954 A was observed on increasing the percentage of tin; however, those values are far from the 4.01 A quoted above for Pt,Sn.Rather the lattice parameters obtained suggest the occurrence of a solid solution of Sn in Pt, although the upper a, value for the saturated solid solution of tin in Pt has been reported to be 3.933 A.35 Apart for the fact that the Pt-Sn solubility gap is still undefined, it seems pertinent here to recall that the quoted value refers to metal phases.G . cocco et al. I I I 1 I I 3.92 1 I I I I 1 325 0 1 2 3 cos 6hkl cot 8 h k l Fig.3. Plot of lattice parameters, a,, against cot 8 cos 0 for the samples studied. Only the first five reflections are reported owing to difficulties of precisely defining higher-order peak locations in patterns of more dispersed samples. Y L unreduced 32 36 40 44 48 28 RP2/3 after hydrogenation (dots, upper curve) and for support (continuous line). Care was taken in specimen preparation in order to observe the same quantity of substance. Resultant subtraction is shown by the lower curve. Further results on the microstructural parameters were obtained from peak profile analysis. As an example we present the raw data handling with reference to sample RP2/3, for which higher broadening was observed. The upper curves of fig. 4 refer to the experimental patterns across the (1 1 1) and (200) angular range for both reduced (data points) and unreduced catalysts (full line). Scattering from the latter was considered as a background to be removed from the former.Within the limits of this assumption, which requires an independent scattering process, the difference function (lower curve) would represent the interference phenomenon caused only by the metal f.c.c. phase. In a further step, the resultant overlapping peaks were resolved by326 X-RAY STUDY OF Pt AND Pt-Sn CATALYSTS ON NYLON 32 36 40 44 48 2e 73 77 81 85 89 2e Fig. 5. (a) Data encompassing (1 11) and (200) line profiles of sample RP2/3 (see lower curve of fig. 4) subjected to a least-squares fit in order to separate overlapping tails suitably. Residuals of the fit are shown in the upper part.(b) The above procedure when the (31 1) broadened line profile almost totally covers the (222) peak, whose determination is necessary for Warren- Averbach analysis. employing the described fitting routine, as essentially shown in fig. 5 (a). However, the complete Warren-Averbach method requires analysis of at least one more profile of the same family. As fig. 5(b) shows, the adopted procedure satisfactorily reduces and separates peak envelopes as in the present case, where noticeable broadening makes the (222) line profile appear as merely a hump of the (3 1 1 ) peak. This now allows us to pursue the analysis further. The Stokes-corrected Fourier coefficients for sample RP2/3 are reported in fig. 6. The close trend for the peak Fourier coefficients suggests that strain is absent or negligible, as confirmed in the In A(L) against h2 + k2 + 12 plot (inset). The other catalysts agree with this structural condition.Although to some extent unexpected, the absence of strain is in line with previous findings on Pd-Pt clusters,lg as well as other finely divided metal phases.20 To conclude the WAXS analysis, fig. 7 compares the crystallite volume distribution function for sample RP9/1 with that of the more dispersed RP2/3 sample. Note the absence of the ‘hook’ effect, which, if present, would have caused negative values at the beginning of the distribution curves.G . cocco et al. 327 1 .o vl +-I $ ._ 0 0.5 .g L 0.0 0 50 100 LlA Fig. 6. Stokes-corrected Fourier coefficients for (1 1 1) and (222) line profiles of Inset plot of In [ 1000 A(L)] against h2 + k2 + 1 for increasing L values ( L = n harmonic number and dhkl the interplanar spacing).sample R dhkl with .P2/3. n the Fig. 7. WAXS crystallite volume distribution functions for samples RP9/ 1 and RP2/3. Column 2 of table 2 collects the volume-weighted means, L,, determined from the WAXS crystallite size distribution for the samples studied The emerging structural pattern covers morphologically crystalline aspects. How- ever, the possible occurrence of X-ray amorphous aggregates might be suspected, even in view of the absence of crystalline Sn. No matter what the nature of the wide-angle diffraction patterns is, the above information may be deduced from SAXS, which only deals with the form factor, independently of the local atomic arrangement.Fig. 8 and 9 show the experimental SAXS intensities for some of the samples studied. Note in the log I against log h plot (fig. 8) that the scattering of sample RP1/0, less dispersed, approaches the forward direction (i.e. smaller h values), displaying an intensity magnitude higher than that of the more dispersed RPl/l catalyst, as expected from WAXS results. Moreover, it is interesting to note the peculiar behaviour and essentially different nature of the Nylon pattern (lower curve). While the monotonically descending trend observed for the catalysts is characteristic of particle-size effects, the maxima in the modulated curve arise from interference phenomena caused by preferential long spacing within the Nylon texture.Fig. 9 compares the experimental behaviour, properly normalized for absorption and primary-beam intensity, for328 X-RAY STUDY OF Pt AND Pt-Sn CATALYSTS ON NYLON Table 2. Structural parametersa from X-ray characterization WAXS SAXS catalyst code Ls L v GAXSb 0: 0: s::XSb RP9/ 1 50 80 56 53 96 53 RP7/3 26 42 108 31 46 90 RP3/2 30 38 93 31 41 90 RP1/1 25 35 112 26 40 108 RP2/3 20 28 140 23 32 122 RPl/O 52 112 54 59 120 48 a L, and L, stand for the average crystallite dimensions weighted according to surface or volume probability, respectively, and D, and D, are the same for SAXS distribution functions. Specific surface expressed in m2 g-l. 4 + 00 + 3 2 - 2 - 1 log h Fig. 8. SAXS intensities of some samples studied. Scattering of the support is also shown. 0 0 0.1 0.2 0.3 0.4 Fig.9. Experimental SAXS intensity for sample RPl / 1 and support as a plot of hl(h) against h. h1.A-lG. cocco et ul. 329 Fig. 10. Resultant 0 0.1 0.2 0.3 0.4 0.5 h/A-’ plots of h31(h) against h for sample RPl/ 1 after subtraction from data of fig. 9. _1_5 3 g2 6 1 0 0 40 80 120 160 RIA of support Fig. 11. Particle-size distribution functions of samples RP9/1 and RP2/3 after transformation of SAXS data according to eqn (2). sample RP1/1 and Nylon: for convenience the hl(h) function was used. The maximum which now appears in the catalyst curve may be evaluated in the frame- work of the ‘invariant’ theory,32 bearing in mind that s? h1(h) dh represents the scattering power of the diffracting particles. Moreover, the two curves look consistent, allowing subtraction of the matrix effect for the catalysts, as is generally our practice.Note that after subtracting the support scattering the course of the resultant intensity should conform to a constant limiting decay value, i.e. lim, ~ h31(h) = K . Actually, this asymptotic trend is observable in the usual h31(h) against h plot (fig. lo), even if a rising branch is observed at higher h values. This behaviour complements that of the remaining samples and will be examined later. In order to satisfy the convergence of the integral reported in eqn (2) and to obtain particle-size distribution functions, the outward behaviour was replaced by the true hP3 trend. Results for samples RPl /O and RP2/3 are reported in fig. 1 1. Comparison with the corresponding WAXS crystallite size distributions of fig.7 shows good agreement; this suggests that most of the scattering particles are coherent domains of diffraction. The most significant parameters of the structural characterization are collected in table 2. Fig. 12 also reports plots of the SAXS and WAXS average particle and 12 FAR 1330 100 5 : 7 5 - 2 5 5 0 - & .- [I: e, 2 5 X-RAY STUDY OF Pt AND Pt-Sn CATALYSTS ON NYLON D - X a D a x. fs - I I I I I I crystallite size values (volume weighted) against nominal tin loading (% ). It appears that the Sn content influences the final dispersion of the catalysts. A consistent structural framework was established by the two methods employed, which support each other. However, we wish to return to the anomalous trend observed in the h31(h) functions.Failure of the intensity tails to conform to h-3 at higher angles has generally been interpreted as electron-density fluctuations within the scattering phase or to concentration gradients at the phase boundary. However, the homogeneity inside the metallic phase has been previously proven; moreover, a well defined interphase is inferred from the presence of a range in which h3Z(h) behaves as forecast by the theory. The final rise in the h3Z(h) against h plot could therefore be assigned to a further complex structure at an atomic level. We believe this is a useful datum, since two different structural conditions of the active metal phase can be coherently surmised. DISCUSSION In comparison with Pt/A1,03, it has recently been shown that propene hydrogenation occurs on Pt/Nylon samples at a much lower rate and with a higher order of reaction with respect to the olefin.5 It is also known that Pt/Nylon does not chemisorb Confining our initial attention to Pt/Nylon, the results of our structural character- ization show that Pt atoms in the metallic state are arranged in the proper f.c.c.form with an average particle size of ca. 100 A. This rules out the possibility that changes in geometry and/or interaction of these relatively large metal particles with the support are responsible for the peculiar catalytic properties of Pt/Nylon. Therefore, the hypothesis may be advanced according to Teichner et aL3 that, during impregnation, a large fraction of metal is driven by a swelling effect of the solvent into the bulk of the polymer.Actually, as may be seen in fig. 13, the well defined interference signals in the Nylon (lower curve) appear smeared-out in the impregnated samples (upper curve), involving at this stage smearing-out of the initial spatial pseudo-periodicity and a strong perturbing effect in the underlying Nylon texture. Thus, after drying, the coalesced Pt fraction is buried and unavailable to gas-phase reactants. This alsoG. cocco et al. 33 1 - 1 RP 1/0 unreduced , 0 ‘ 1 0 01 0.2 0.3 0.4 h1A-l Fig. 13. SAXS pattern for non-impregnated (lower curve) Nylon and after impregnation (upper curve): the different behaviour indicates changes occurring within the Nylon in spite of the similarity to the WAXS pattern presented in fig. 1. agrees with experimental findings that it is possible to reduce the precursor in a slurry at a much lower temperature than in the gas phase.By considering that the metallic phase detected by X-ray scattering represents a very large amount of the platinum present in the sample, the unusual properties of Pt/Nylon must be ascribed to a small fraction of atoms strongly interacting with the amide groups, preserved from aggregation. On the basis of these considerations it may be envisaged that the catalytic properties are determined by the physical nature of the reactants (liquid or gaseous) and, for slurry reactions, by the swelling effect of the solvent. Hydrogenation reactions in the liquid phase on these catalysts have been performed and the results, reported e1sewhere,l4 may be explained with the above structural framework.The addition of Sn to Pt/Nylon has been found to decrease the rate of hydrogenolysis of propene and to increase the order of reaction with respect to the olefin. Furthermore, it has been observed that catalysts with a Sn content > 50% are ina~tive.~ The data reported in this paper show that the presence of tin influences the growth of platinum particles, which are smaller for higher Sn content. Stabilization against sintering has previously been reported.l2? l3 Burchll has also found an increase of chemisorbed hydrogen on Pt/Sn. However, as in the case of the unalloyed Pt, geometric factors cannot account for the catalytic behaviour of Pt-Sn toward olefin hydrogenation. C=C double-bond saturation is a structure-insensitive reaction, and a decrease in the metal particle size should lead to an increase in the number of active sites as a result of a larger surface area.Therefore the decreased catalytic activity in the higher-order reaction with respect to the olefin found in propene hydrogenation5 is probably related to an electronic interaction between platinum and tin. Formation of Pt-Sn alloys has been proposed by several authors and changes in catalytic performance have been attributed either to an ensemble or a ligand effe~t.’9~ The influence of tin has in particular been related to the donation of electrons into the 5d band, which tends to be filled, thus deactivating the active sites.’ As shown in the results, we did not detect any formation of alloyed compounds, although it is still possible that an electronic interaction occurs as a consequence of the insertion of Sn into the host Pt lattice to form the solid solution.However, the morphological structure pattern does not appear dissimilar for the Pt/Nylon and Pt-Sn/N ylon systems examined here, and therefore, by following the considerations previously reported for Pt/Nylon, we can say that in the gas phase the embedded Pt-Sn solid solution should have little influence on the reacting environment. 12-2332 X-RAY STUDY OF Pt AND Pt-Sn CATALYSTS ON NYLON Another model which has been proposed in the literature suggests an interaction between Pt and Sn ions bonded to the surface, which would result in electron withdrawal and the formation of electron-deficient Pt sites.* In such a case the weaker olefin-metal bond would be ascribed to a smaller back-donation into the n* antibonding orbital of the olefin. The absence of metallic tin in crystalline aggregates and the relatively small amount of this metal present within the Pt solid solution indicates that in our case Sn is still in a ionic form bonded to the support.This is also in accord with preliminary Mossbauer experiments, which have shown the presence of Sn1V.36 On the basis of the above considerations it may be suggested that in our system two types of interaction exist between Pt and Sn: the first occurs via the formation of a solid solution and the second through interaction with Sn ions. Although the former cannot be disregarded for catalytic purposes, the result would in any case be a decrease in the strength of the metal-olefin bond.The scheme reported here could account for the peculiar features of the systems investigated, and the approach followed appears to be a sound tool for relating microstructural features to the performance of Pt-Sn/Nylon catalysts. We thank the Consiglio Nazionale delle Ricerche (Piano Finalizzato Chimica Fine e Secondaria) for financial support. S. E. thanks W. Parrish for helpful discussions on the subect of X-ray profile fitting. D. P. Harrison and H. F. Rase, Ind. Eng. Chem. Fundam., 1967, 6, 161. P. Dini, D. Dones, S. Montelatici and N. Giordano, J. Catal., 1973, 30, 1. S. J. Teichner, C. Hoang-Van and M. Astier, in Metal-Support and Metal Additive Eflects in Catalysis, ed. B. Imelik (Elsevier, Amsterdam, 1982), p.121. S. Galvagno, P. Staiti, P. Antonucci, A. Giannetto and N. Giordano, React. Kinet. Catal. Lett., 1982, 21, 157. S. Galvagno, P. Staiti, P. Antonucci, A. Donato and R. Pietropaolo, J. Chem. Soc., Faraday Trans. I , 1983 79, 2605. J. F. Young, R. D. Gillard and G. Wilkinson, J. Chem. Soc., 1964, 5176. ’ J. Volter and U. Kurschner, Appl. Catal., 1983, 8, 167. R. Burch and A. J. Mitchell, Appl. Catal., 1983, 6, 121. F. M. Dautzenberg, J. N. Helle, P. Biloen and W. M. H. Sachtler, J . Catal., 1980, 63, 119. lo R. Burch and L. C. Garla, J. Catal., 1981, 71, 360. l 1 R. Burch, J. Catal., 1981, 71, 348. l2 H. Charcosset, R. Frety, G. Leclercq, B. Moraweck, L. Tourneyan and J. Vasland, React. Kinet. Catal. Lett., 1979, 10, 301. l 3 P. A. Zhdan, B.N. Kuznetsov, A. P. Shepelin, V. I. Kovelchuk and Yu, J. Yermakov, React. Kinet. Catal. Lett., 1981, 18, 267. l4 S. Galvagno, R. Pietropaolo and Z. Poltarzsewski, unpublished results. J. C. Rasser, W. H. Beindorff and J. J. F. Scholten, J. Catal., 1979, 59, 21 1. l6 B. Coq and F. Figueras, J. Catal., 1984, 85, 1975. l 7 J. W. Bassi, F. Garbassi, G. Vlaic, A. Marzi, G . R. Tauszik, G. Cocco, S. Galvagno and G. Parravano, l E S. Galvagno, J. Schwank, G. Parravano, F. Garbassi, A. Marzi and G . R. Tauszik, J. Catal., 1981, l9 G. Cocco, G. Carturan, S. Enzo and L. Schiffini, J. Catal., 1984, 85, 405. 21 G. Cocco, S. Enzo, F. Pinna and G. Strukul, J. Catal., 1983, 82, 160. 22 B. E. Warren, X-ray Diflraction (Addison Wesley, Reading, Mass., 1969). 23 B. E. Warren and B. L. Averbach, J. Appl. Phys., 1952, 23, 497. 24 R. A. Young and D. B. Wiles, J. Appl. Cryst., 1982, 15, 430. 26 A. Guinier, X-ray Dzffraction (W. H. Freeman, San Francisco, 1963). 27 H. Stabinger and 0. Kratky, Makromol. Chem., 1979, 180, 2995. 28 H. Stabinger and 0. Kratky, Makromol. Chem., 1978, 179, 1655. J . Catal., 1980, 64, 405. 69, 283. A. Benedetti, G. Cocco, S. Enzo, G. Piccaluga and L. Schiffini, J. Chim. Phys., 1981 78, 961. A. R. Stokes, Proc. Phys. SOC., 1948, 61, 382.G . cocco et al. 333 29 G. Cocco, L. Schiffini, G. Strukul and G. Carturan, J . Catal., 1980, 65, 348. 30 G. Cocco, L. Schiffini, L. Battezzati and A. Lucci, J . Non-cryst. Solids, 1983, 54, 301. 31 G. Cocco, L. Schiffini, S. Enzo and A. Benedetti, Phys. Status Solidi, 1982 69, 343. 32 A. Guinier and G. Fournet, Small-angle Scattering of X-Rays (Wiley, New York, 1955). 33 M. Hansen, Constitution of Binary Alloys (McGraw-Hill, New York, 1958). 34 K. Schubert and H. Pfisterer, Z . Metallkd., 1949, 40, 405. 35 K. Schubert and E. Jahn, 2. Metallkd., 1949,40, 399. 36 A. Pellerito, personal communication. (PAPER 4/491)

ISSN:0300-9599    

DOI:10.1039/F19858100321

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1985,81, 335-341 A Kinetic Study of the Formation and Dissociation of the Meisenheimer Complex Formed between 1,3,5Trinitrobenzene and the Hydroxide Ion in Micellar Dodecyltrimethylammonium Bromide Solution BY JOEL LELIEVRE* AND RENE GABORIAUD E.N.S.C.P., Laboratoire de Physicochimie des Solutions, 11 Rue Pierre et Marie Curie, 75231 Paris Cedex 05, France Received 29th March, 1984 The rates of formation and dissociation of the thermoadditioncomplex of 1,3,5-trinitrobenzene with OH- has been investigated in micellar solutions of dodecyltrimethylammonium bromide (DTABr) using a stopped-flow spectrometer. We have observed and measured three processes which proceed under different conditions of acidity: (a) formation of the r~ complex, (b) spontaneous (equilibrium) dissociation in less basic solutions and (c) dissociation via proton attack in acid solutions.These processes are influenced by the background KBr electrolyte. The interphase potential A$ = drn - &, depicted in the pseudophase model, allows us to explain the variations of the rate constants. The catalytic effect of cationic detergents, for example dodecyltrimethylammonium bromide solutions with water (DTABr), on nucleophilic aromatic substitution reactions has been studied by several workers.l-13 1 -Halogeno-2,4-dinitrobenzene compounds have often been used and reacted with the hydroxide ion, OH-, to form the corresponding phenate; the catalytic effect on such reactions is significant. In contrast, few investigations14-16 have been made on aromatic substitutions in which 0 addition complexes (or Meisenheimer complexes) are formed, e.g.when the aromatic substrate is a trinitrosubstituted benzene compound. The aim of this work is to investigate whether a cationic detergent (DTABr) has an active effect on the rate of formation of such complexes between 1,3,Strinitrobenzene (TNB) and the OH- ion. This aromatic substrate shows simple behaviour in sodium hydroxide solutions because there is no substitutable group (or leaving group) and the reaction is k , limited to k-1 TNB+OH- eTNBOH- (1) in which the 0 addition complex (TNBOH-) is stable in micellar solutions. We have already studied these reactions for several mixtures of water and methanol” and, depending upon the conditions of acidity used, we were able to describe several reaction processes.When the medium is strongly basic, the principal reaction is to give the o complex TNBOH- and only the rate constant k, may be determined experimentally. For a less basic range, the rate constant k-, is also accessible and corresponds to dissociation of the Meisenheimer complex, caused by thermal motion of solvent molecules. For strongly acid mixtures, dissociation occurs by direct attack and the reaction scheme is written as k,[H+l TNBOH--TNB + H,O. 3353 36 REACTION OF TRINITROBENZENE WITH OH - IN MICELLAR SOLUTION Fig. 1. Theoretical diagram showing the logarithm of the apparent constant A for the processes of ionization of TNB [reaction (I)] and dissociation of the TNBOH- complex [reaction (2)]. 3.= k,[OH-]+k-,+k,[H+]. The overall behaviour is summarized diagrammatically in fig. 1 . The effects of DTABr upon these three processes have now been investigated and this paper reports our results. EXPERIMENTAL DTABr detergent was used at a concentration of 2 x lo-' mol dm-3, higher than the critical micellar concentration (c.m.c. = I .56 x lop2 mol dm-3), and normally in the presence of an excess of background electrolyte in order to mask the secondary effects of the various buffer salts in solution, even though the background electrolyte is expected to affect the rates of reaction lo- 18. The rates of reaction were always high and a stopped-flow Durrum Gibson spectrometer was used with the measurement cells thermostatted at 20k0.1 "C, as were the two syringes for injection of reagents.The experimental kinetics curves were displayed on the screen of a Tektronix oscilloscope connected directly to the stopped syringe. However, the use of such a method with micellar solutions is difficult and tedious because many microbubbles are formed in the mixing jet during the injections of reagents. Thus, the different solutions had to be injected three or four times for each measurement in order to obtain reproducible results. RESULTS When TNB in micellar DTABr solutions is reacted with a range of concentrations of micellar sodium hydroxide solutions, we can calculate the rate constants k , and k-l, see fig. 2 and 3. On the other hand, the determination of the rate constant k , is experimentally more difficult. The reagents are mixed quickly in the mixing jet of the stopped-flow spectrometer, one being the preformed TNBOH- complex with an excess of OH- ions and the other being a weak acid solution chosen to obtain the desired final pH value.The two solutions of reagents were made up in such a way that, in every one, there were the same concentrations of surfactant (2 x mol dm-3) and background electrolyte (4 x lop2, 8 x lop2, 2 x 10-l or 4 x 10-1 mol dmP3 KBr). Such reagent conditions were necessary in order to ensure that there were no problems caused by viscosity differences in the mixingjet. The excess of salt enables not only the interphase potential but also the ionic surroundings of the micellar structure to be obtained. The preformed TNBOH- complex with an excess of OH- is stable in solutions of DTABr and KBr and the reaction observed after mixing with weak acid solutions corresponds to reaction (2). Indeed, in the measurementJ.LELIEVRE AND R. GABORIAUD 337 2 log [OH-] Fig. 2. Plot of the pseudo-first-order rate constant 2 against log [OH-]. [DTABr] = 2 x mol dmP3 and T = 20 "C. +, Without background electrolyte; 0, 4 x x , 8 x A, 2 x lo-' and .,4 x 10-' mol dmP3 KBr. t 2 1 5 7 9 11 13 PH Fig. 3. Plot of the pseudo-first-order rate constant 2 defined by kapp = A[OH-] against pH. Circles and crosses refer to KBr concentrations of 8 x and 4 x lo-' mol dm-3, respectively. [DTABr] = 2 x lop2 mol dmP3 and T = 20 "C. cell there occurs fast neutralization of the weak acid by the excess of OH- and formation of the buffer solution in situ, which sets the final pH value.Acetic acid and butylammonium, dibutylammonium or pyridinium salts were used in various proportions for obtaining the correct pH value over a range of ca. 6 pH units. The kinetic curves were recorded on the screen of the oscilloscope at a wavelength A = 460 nm, corresponding to the maximum absorbance of the CT complex. For reaction (2) we followed the decreasing absorbance at the same wavelength. In all experiments the concentration of TNB was very low ( 10-4-10-5 mol dm-3) and under338 REACTION OF TRINITROBENZENE WITH OH- IN MICELLAR SOLUTION Table 1. Variation of log mA (pseudo-first-order rate constant in s-l) with log [OH-] for different concentrations of KBr. CDTABr = 2 x lop2 mol dm-3, T = 20 "C; log A = log (k-.l + k,[OH-]) C,,,/mol dmp3 log [OH-] 0 4 x 10-2 8 x 10-2 2 x 10-1 4 x 10-1 - 0.7 - 1.0 - 1.3 - 1.6 - 2.0 -2.3 - 2.6 - 3.0 - 1.85 1.52 1.20 0.80 0.53 0.23 0.05 - 1.94 1.65 1.34 0.98 0.66 0.33 0.17 0.02 - 1.83 1.40 1.12 0.78 0.43 0.2 1 0.06 -0.03 1.60 1.20 0.88 0.54 0.29 0.07 - 0.02 -0.10 1.40 1 .oo 0.70 0.40 0.13 0.0 1 - 0.08 -0.10 such conditions the concentrations of OH- may be considered to be constant.Fig. 2 shows the variation of the pseudo-first-order rate constant, A, with log[OH-] (k = A/[OH-1; cf. caption to table 1) for solutions with and without KBr electrolyte (KBr = 4 x lo-,, 8 x lo-,, 2 x 10-1 and4 x 10-1 mol dm-3). All thedataare summarized in table 1, and fig. 2 leads to several conclusions. When the reaction occurs in the absence of electrolyte, the plot of log A against log [OH-] function is represented by a straight line whose slope is unity.The micellar catalysis is at a maximum compared with the results in the presence of KBr. On the other hand, when the reaction occurs in the presence of electrolyte, the rate constant is decreased in proportion to the increasing concentrations of KBr. For log [OH-] > - 2, the plots of log A against log[OH-] are represented by straight lines whose slopes are unity and for log [OH-] < - 2 these they tend towards a constant value corresponding to k1. The variations of the pseudo-first-order rate constants A with pH are illustrated (fig. 3) for solutions with two KBr concentrations, 8 x lop2 and 4 x mol dm-3. The form of the curves leads to several conclusions depending on the pH range.When pH > 11, the plot of log1 against pH is represented by a straight line whose slope is unity. The reaction is thus second order as expected. When 8 < pH < 11, log A is independent of pH. When 5 < pH < 8, as is seen in the third part of the kinetic diagram, the plot of logA against pH is a line whose slope is - 1, corresponding to the o complex dissociation reaction with the protons present in solution. In this case, we note that the proton reacts with the TNBOH- complex despite the repulsion of the positive micellar charges. Such a result has not previously been observed with cationic micelles, but the corresponding reaction with anionic surfactants (symmetrical trend) has been described for micellar dodecylsulphate in alkaline media.l3 When the concentration of the KBr electrolyte is increased, the rate constant k , invariably decreases.However, the k-, rate constant is modified very little by added salt. Finally, the dissociation rate, following reaction (2), is increased by adding KBr and we notice a comparable translation for the two lines of slope 1 and - 1 in the presence of salt. Thus the rate constants k , and k, are changed by almost the same absolute value. We were careful to check that this was not due to pH variation in the micellar solution.J. LELIEVRE AND R. GABORIAUD 339 DISCUSSION Two theories are able to show the logical development of the effects of ionic surroundings on micellar catalysis, one being based on the interphase potential of the two phases A# = #M - #w (& and dw are the potential of the micellar and bulk phases, respectively) and the other on ionic exchange on the micellar surface.These two models have been applied by different workers and from time to time they have been considered as mutually exclusive. However, we believe that they are compatible but that the interphase potential model is able to explain our experiments more simply. In this paper we have chosen to show the influence of the addition of a salt on the micellar catalysis of a model reaction [reactions (1) and (2)]. The presence of the salt enable us to set the interphase potential and for that reason we used different concentrations in order to experiment with several values of A#. It is known that the presence of salt generally involves a decrease in the rate of the reaction: either by competition between two anions (Br-, the common anion of the surfactant and the background electrolyte, and the OH- anion) or by decreasing the interphase potential and correspondingly the rate constant. The three processes of the reaction are influenced differently by increasing concentration of electrolyte: (a) a decrease of the rate constant k,, as has been described by several authors, (b) an unmodified value of the rate constant k-, and ( c ) an increase of the rate constant k,.A qualitative interpretation becomes obvious if we use the interphase potential expression. We have already 24 an expression for the interphase potential using the pseudophase model : where si and bi are the parameters of each species of ion which are determined experimentally for kinetic results, [ai] is the ionic concentration in the bulk phase and A& is a constant depending on amphiphiles, counter ion, solvent and temperature.Using eqn (i) we are able to foresee the effect of a variation of electrolyte concentration and we believe, in accord with other 26 that an excess of salt will give rise to an interphase potential difference. If we develop eqn (i), taking account of the anions present in solution, we use a cationic amphiphile (DTA+, Br-), a reactive anion OH- and a fixed excess of back- ground electrolyte (K+, &-), whose anion is common with the counter ion of the surfactant. Consequently eqn (i) may be written as Eqn (ii) explains the variation of A# with varying concentration of background electrolyte. We do not need the values of si and bi for explaining qualitatively our kinetic results, nevertheless we have dete~mined~~9 2 4 9 27 these coefficients using other kinetic results and also for shifts of the deprotonation equilibria. Thus, choosing the OH- ion as a reference sOH- = 1, we have obtained sBr- = 25.Consequently the micellar structure has a selectivity which is 25 times larger for Br- than the OH-. We are in agreement with the selectivity magnitude for these ions found by Bunton et aZ.10v21 The contribution of the [OH-] term in eqn (ii) is thus very small in com- parison with that for the Br- ion and eqn (ii) may therefore be simplified to340 REACTION OF TRINITROBENZENE WITH OH- IN MICELLAR SOLUTION The charge transfer between two phases involves the development of the transfer rate concurrently with Ab: Ad (iv) F logmk = constant-- 2.3RT where "k is the kinetic micellar rate constant (second order) and the constant is the sum of several constants: one being the value of the kinetic rate constant in water and the other the ratio of transfer activity coefficients of several species in solution.We mentioned previously that we have determined experimentally the pseudo- first-order rate constant mA, which can be reproduced by taking account of the expressions for log "k and Ad: log sBr-[Br-]. When the salt concentration increases, the interphase potential decreases and consequently the rate of the TNB ionization reaction decreases (slope + 1 in fig. 2 and 3). The thermal dissociation process (rate constant k - l ) cannot be modified by a A4 potential variation, as is observed.On the contrary, for the dissociation of the complex caused by proton attack (slope -1 in fig. 3), we obtain an inversion of reactivity, the rate constant k, increasing with the increasing electrolyte concentration. Again, a decrease in Ad causes a reduction in ionic repulsion between the proton and the micelles and reaction is promoted. This is a very interesting experimental example of an increase of rate constant obtained by increasing the electrolyte concentration, the opposite effect having always been described for cationic amphiphiles previously. The kinetic interpretations based on the ion-exchange model of the micellar surface22 are not able to explain such results without recourse to other assumptions such as the introduction of a local pH at the micellar surface;13 on the other hand our approach gives a plausible and direct interpretation.If the kinetic variations are due only to the variation in interphase potential that is determined by the change of electrolyte concentration in the bulk, then the acceleration of proton attack has to correspond exactly to the decrease of the TNB reaction with OH- ions, as observed. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 H. Chaimovich, A. Blanco, L. Chayet, L. M. Costa, P. M. Montero, C. A. Bunton and C. Paik, Tetrahedron, 1975, 31, 1139. M. J. Blandamer, G. H. Beatham, C. H. Branch and D. J. Reid, J. Chem. SOC., Faraday Trans. I, 1976, 72, 2139. C. A. Bunton and S.Diaz, J. Am. Chem. Soc., 1976, 98, 5663. C. A. Bunton, Pure Appl. Chem., 1977, 49, 969. M. Almgren and R. Rydholm, J. Phys. Chem., 1979,83, 360. F. Quina and H. Chaimovich, J. Phys. Chem., 1979, 83, 1844. H. Chaimovich, J. B. S. Bonhila, M. J. Politi and F. H. Quina, J. Phys. Chem., 1979, 83, 1851. C. A. Bunton, G. Cerichelli, Y. Ihara and L. Sepulveda, J. Am. Chem. SOC., 1979, 101, 2429. C. A. Bunton, L. S. Romsted and G. Savelli, J. Am. Chem. Soc., 1979, 101, 1253. C. A. Bunton, J. Frankson and L. S. Romsted, J. Phys. Chem., 1980, 84, 2607. C. A. Bunton, F. H. Hamed and L. S. Romsted, J. Phys. Chem., 1982, 86, 2103. F. Nome, A. F. Rubira, C. Franco and L. G. Ionescu, J. Phys. Chem., 1982, 86, 1881. F. H. Quina, M. J. Politi, 1. M. Cuccovia, S. M. Martins-Franchetti and H. Chaimovich, in Solution Behavior of Surfactants. Theoretical and Applied Aspects, ed. K. L. Mittal and E. J. Fendler (Plenum Press, New York, 1982), vol. 2, p. 1125. J. H. Fendler and E. J. Fendler, in Catalysis in Micellar and Macromolecular Systems (Academic Press, New York, 1975). J. H. Fendler, E. J. Fendler and L. M. Casilio, J. Chem. SOC. B, 1971, 1377. J. H. Fendler, E. J. Fendler and S. A. Chang, J . Am. Chem. SOC., 1973, 95, 3273. J. Lelievre, R. Gaboriaud and R. Schaal, Bull. SOC. Chim. Fr., 1971, 1246. N. Funasaki, J. Phys. Chem., 1979, 83, 1998. K. Shirahama, Bull. Chem. SOC. Jpn, 1976, 49, 2731.J. LELIEVRE AND R. GABORIAUD 34 1 Y. Miyashita and S. Hayano, J. Colloid Interface Sci., 1982, 86, 344. 21 C. A. Bunton, L. S. Romsted and L. Sepulveda, J. Phys. Chem., 1980, 84, 261 1. 22 H. Chaimovich, R. M. Valeixo, I. M. Cuccovia, D. Zanette and F. H. Quina, in Solution Behavior of Surfactants. Theoretical and Applied Aspects, ed. K. L. Mittal and E. J. Fendler (Plenum Press, New York, 1982), vol. 2, p. 949. 23 J. Lelievre, Thesis (Pans, 1982). 24 R. Gaboriaud, G. Charbit and F. Donon, Surfactants in Solution, ed. K. L. Mittal (Plenum Press, 25 G. Gunnarsson, B. Jonsson and H. Wennestrom, J. Phys. Chem., 1980,84, 31 14. 28 J. Framm, S. Diekmann and A. Hasse, Ber. Bunsenges. Phys. Chem., 1980,84, 566. 27 B. Rakotoson, Thesis 3" Cycle (Pans, 1983). New York, 1984), vol. 2, p. 1191. (PAPER 4/5 13)

ISSN:0300-9599    

DOI:10.1039/F19858100335

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1985, 81, 343-354 A Thermokinetic Foundation for Oscillatory Phenomena in Gaseous Organic Oxidations under Well Stirred Flowing Conditions BY JOHN F. GRIFFITHS,* STEPHEN M. HASKO, NIGEL K. SHAWt AND TOMAS TORREZ-MUJICA Department of Physical Chemistry, The University, Leeds LS2 9JT Received 13th April, 1984 An experimental and theoretical attack on the fundamentals of thermokinetic phenomena associated with the gaseous, non-isothermal oxidation of hydrocarbons and other organic substrates is described. Quantitative comparisons are made between numerical modelling and experimental measurements under well stirred flowing conditions. Two chemical systems are considered, involving reactions of methyl radicals. These are: (i) di-t-butyl peroxide decomposition in nitrogen and (ii) di-t-butyl peroxide decomposition in an excess of oxygen.Simplified kinetic mechanisms for each of these systems are described and numerical computations for non-isothermal reactions are discussed. Stationary states and two different types of oscillatory modes are predicted to exist within limited ranges of p , T, and composition, and these match experimental measurements quite satisfactorily. The integral role played by self-heating in thermokinetic oscillations is demonstrated and relationships to cool-flame phenomena are outlined. This paper is part of an experimental and theoretical programme concerned with the thermokinetic interpretation of non-isothermal oscillatory phenomena and multiple-stage ignitions associated with gaseous organic oxidations.It deals with reactions of methyl radicals in non-degenerately branched modes and aims to bridge the gap between thermal ignition on the one hand, for which exothermic reaction and a positive, Arrhenius-like temperature dependence of rate are the only chemical pre-requisites, and, on the other, thermokinetic ‘cool-flame ’ phenomena, in which kinetics involving realms of negative temperature dependence of the overall rate of reaction are strongly coupled to self-heating. The data presented here are drawn from experimental and numerical studies of two chemical reactions under well stirred flowing conditions [i.e. in a continuous, stirred-tank reactor (c.s.t.r.)]. Non-isothermal, but non-adiabatic, reactions in a c.s.t.r. may lead to stationary or oscillatory states, and even a multiplicity of them.They are equally amenable to straightforward experimental or numerical study : there is neither spatial variation within the reactor (as occurs in unstirred vessels) nor the total consumption of reactants (as occurs in closed systems). The systems considered are (i) the decomposition of di-t-butyl peroxide (DTBP) in nitrogen and (ii) the decomposition of DTBP in a substantial excess of oxygen. Each is very reactive from ca. 460 K up, at total pressures down to 50 Torr, and both stationary and oscillatory states are predicted and observed. The effects of changes in concentration of DTBP, the ( p , T,) conditions at which different reaction modes t Present address : Geometric Modelling Project, Department of Mechanical Engineering, The University, Leeds LS2 9JT.343344 THERMOKINETIC OSCILLATIONS IN A C.S.T.R. occur and the frequencies and amplitudes associated with oscillatory reactions are studied. DTBP is an efficient source of methyl radicals within the temperature range of interest (475-600 K). Studies of DTBP decomposition span many decades; they have yielded precise kineticll and thermochemical and they have also been used to validate aspects of thermal-explosion In addition, the decomposition has been used7 to test the performance of a Longwell type of c.s.t.r. The decomposition of DTBP in the presence of oxygen is less well characterised. It has been investigated in closed vessels8-10 and studied recently in the context of thermal explosions.11 It has been used a source of methylperoxy radicals in kinetic studies.12 In laminar flow, Williams et aZ.13 have utilised this oxidation to study the origins of cool-flame chemiluminescence (CH,O*), and Ballinger and Ryason14 investigated the stabilised ‘cool flame’ on a flat-flame burner.The mechanism for the isothermal decomposition of DTBP in the presence of oxygen has been subjected to a numerical study.15 The apparatus and experimental procedures adopted have been described in detail previously.l* Adaption of the existing stirred-flow apparatus to admit DTBP vapour is to be described fully in a subsequent paper. Briefly, the principal need is for its very carefully controlled generation and its ‘cold’ admission to the c.s.t.r. in order to prevent excessive preliminary decomposition.The experimental results discussed here have been obtained in a mechanically stirred, spherical, flow-through vessel (Pyrex glass, 0.5 dm3) under a mean residence time for the reactants of 3 s. Measurements have been confined so far to temperature changes within the reacting mixture using a very fine, silica-coated, Pt-Pt/ 13 % Rh thermocouple. Signals were recorded using either a chart or oscilloscope. THERMOKINETIC FOUNDATION TO NUMERICAL MODELS Numerical and analytical interpretations of behaviour in a c.s.t.r. are founded on a set of time-dependent, differential expressions representing the reactant temperature and concentration of each species. Stationary-state analysis and the application of theorems for stability-of-motion are very powerful techniques for qualitative identi- fication and location of the singularities for the thermokinetic system in parameter phase-space.l7.l8 Quantitative predictions require numerical integration of the set of differential expressions. The terms in each equation are determined by the mechanistic basis that is invoked and pre-ordained characteristics for heat and mass transport. One underlying premise in this work is to adopt the simplest kinetic schemes consistent with recognised features of low-temperature ‘ hydrocarbon’ oxidation. Even by using simplified schemes there is very satisfactory agreement between theory and experiment, but there are obvious areas for detailed improvement; the price that may be paid is the loss of the generality that is conveyed by simplicity.THE ENERGY-CONSERVATION EQUATION The rate of temperature change is expressed in each case in the form where c/J m-3 K-l is the volumetric heat capacity for reactor contents, T/K is the reactant temperature, TJK is the vessel temperature, hi/J mol-1 is the exothermicity of the jth elementary reaction (= -AH&), kj/s-l is the rate constant for the jth elementary reaction, li;./mol m-3 is the concentration term for the jth elementary reaction given by, e.g., [A]“[B]* . . . where a and b are orders with respect to speciesJ. F. GRIFFITHS, S. M. HASKO, N. K. SHAW AND T. TORREZ-MUJICA 345 A and B,f/m3 s-l is the volumetric flow rate, x/W m-, K-' is the Newtonian heat- transfer coefficient between the gas and the vessel walls and (S/V)/m-l is the surface-to-volume ratio.The first term on the right-hand side represents the summation of the heat-release rates from all of the elementary steps involved. We have taken constant enthalpies (AH%,) in the present study. The second term is the heat-loss rate from the system and is controlled by combined effects of (i) transport from the system by hot products and (ii) Newtonian heat loss via the walls of the vessel. The second part is overwhelmingly dominant when the mean residence time exceeds 1 s. Pre-heating of all reactants to the vessel temperature is assumed. The reference temperature T* may be to allow for entry of gases at a different temperature from that of the vessel, T,. T* is the temperature at the outflow to which non-reactive gases, entering a vessel at T,, would be raised when their inflow temperature is T, and is a weighted mean.Thus T* = (T,fc+ TxS)/(fc+xS)* (11) T, can be measured experimentally and its simultaneous measurement with T, and Ta at known flow rates and compositions is an established route to evaluation of the hea t-transfer coefficient. 2o Since the present work constitutes a pilot study, corrections for the variation of heat capacity with temperature or reactor contents, and of the heat-transfer coefficient x with pressure, have been omitted to avoid disproportionate consumption of C.P.U. time. The heat-transfer coefficient k) is known to vary with pressure, temperature and composition, and allowance for its change is possible.21,22 We have chosen to fix it at x = 15.8 W m-' K-l, consistent with measurements made both previously2'* 22 and during the course of this work.Nevertheless, there is an in-built correction for different rates of heat loss at different conditions because the volumetric heat capacity also plays a part [eqn (I)]. KINETIC AND THERMOCHEMISTRY FOR THE THERMAL DECOMPOSITION OF DTBP The mechanism, and kinetic and thermochemical data, for the thermal decompo- sition of DTBP are unequivocal,2*6 To be consistent with later requirements we represent it as a two-step process, as shown in scheme A. Scheme A A/s-' or E AH88 mol-1 m3 s-' /kJ mol-l /kJ mol-' (CH3)3COOC(CH,)3 + 2(CH3),C0 + 2CH3 ( 1 ) 6.3 x 1015 158 191 CH, + CH, + C,H, (2) 2.5x 107 0 -371 For reasons connected with the present computational method we regard the methyl radical recombination (2) to be at the high-pressure limit.This is also the case for reactions (3), (4) and (8) below. This kinetic foundation for DTBP decomposition leads to five, time-dependent, differential equations representing the rate of temperature change and the rate of concentration change for the four species DTBP, acetone, methyl and ethane. The initial concentration is solely that for DTBP, but allowance is made numerically for the flow-through and heat capacity of the inert carrier (N,) that is necessary experimentally.346 THERMOKINETIC OSCILLATIONS IN A C.S.T.R. KINETICS AND THERMOCHEMISTRY FOR THE DECOMPOSITION OF DTBP IN OXYGEN The minimum viable scheme consistent with present understanding of the low- temperature oxidation of methyl radicals involves 17 species in 14 reactions, as shown in scheme B.Scheme B A/s-l or E AH% mol-l m3 s-l /kJ mol-l /kJ mol-l DTBP -+ 2(CH,),CO + 2CH, CH, + CH, -+ C2H6 CH, + 0, -+ CH,O, CH,O, -+ CH, + 0, CH,O, + CH,O, -+ 2CH,O + 0, CH,O + CH,O -+ CH,OH + CH,O CH,O, + CH,O + CH,OOH + CHO CH,OOH -+ CH,O +OH CH,O + OH -+ CHO + H,O CH,OH +OH -+ CH,OH + H,O CHO+O,-+CO+HO, CH,O + 0, + CH,O + HO, CH,OH + 0, -+ CH,O + HO, HO, + HO, -+ H,O, + 0, (1) 6.3 x 1015 (2) 2.5 x 107 (3) 1 x 106 (4) I x 1014 (5) 1 x 106 (6) 1 x lo5 (7) 1 x 106 (8) 2 x 1015 (10) 2~ 107 (11) 5~ 107 (12) 7~ 107 (9) 6 x los (13) 1 x lo6 (14) 1 . 8 ~ los 158 0 0 115 0 0 43 170 0 7 8 1 1 7 0 191 -371 -115 115 - 26 - 338 - 12 184 - 135 - 106 - 128 - 102 - 60 - 177 Rate data are drawn from several source^^^-^^ without substantial amendment and enthalpies are calculated from heats of formation of the components in each elementary step.The most important features of this mechanism are (i) the secondary initiation mode (8) via methyl hydroperoxide and (ii) the ' thermokinetic switch' via the equilibrium CH3 + 0, * CH302 that diverts the main course of reaction away from the oxidation route as the temperature increases. To maintain simplicity we have purposely suppressed processes subsidiary to the fundamental kinetic framework; the main effect is to yield an overall stoichiometry and heat output that do not exactly match experiment. (3) (4) DEVELOPMENT AND TESTS OF THE NKSCHEM NUMERICAL PROGRAM The program in this study, NKSCHEM, has been developed for application to c.s.t.r.studies consistent with the principles set out above. The main numerical routine used is the NAG Library routine D ~ ~ Q B F exploiting Gear's method2' for numerical integration of first-order differential equations exhibiting considerable stiffness. In c.s.t.r. simulations the concentration of a particular species may become negative over a single step of the solver. An appropriate back-step is taken if this occurs and the size of the time step is reduced. In extremely stiff cases it is possible to proceed by setting the negative concentration to zero, corresponding to total depletion of the species concerned. Reactant pressures, flow rates and vessel temperatures, and the thermal, kinetic and stoichiometric parameters are specified in a single data file.The procedure is simple and maintains generality, enabling factors not explicit in a stoichiometric matrix toJ. F. GRIFFITHS, S. M. HASKO, N. K. SHAW AND T. TORREZ-MUJICA 347 be included, e.g. the presence of inert gases, which is sometimes necessary experi- mentally. If the reactants are set to zero initially and the internal temperature set at T,, the simulated evolution is equivalent to starting up an experiment at a vessel tempera- ture T, by displacement of an inert gas of the same volume heat capacity and pressure as those of the incoming reactants. Simulations of > 30 s experimental time from start-up to steady-state in the c.s.t.r. (10 residence times) rarely take > 10 s C.P.U. time. Simulations of oscillations for the same duration of experiment can take up to 150 s C.P.U. time, but are often completed more quickly.TEST OF NKSCHEM AND AN IMPORTANT CONSEQUENCE OF SELF-HEATING The same rates, extents of reaction and overall exothermicity will be observed in a stationary state under non-isothermal conditions at temperature T as under isothermal conditions where the vessel temperature is given by T, = T. The well characterised KINPAK isothermal program has thus been used to test the validity of results from the NKSCHEM non-isothermal program at the same reactant temperatures. K I N P A K ~ ~ is a program developed tc. predict isothermal reaction rates, extents and stoichiometries of chemical processes. In the present study KINPAK has been used to compute the extents of reaction achieved in the c.s.t.r.over the temperature range 450-600 K during isothermal reaction of DTBP in oxygen consistent with kinetic scheme B. Representative results displayed in fig. 1 refer to the composition DTBP (7.5 mol %)+02 at a total pressure of 80 Torr and a mean residence time of 3 s. For KINPAK the abscissa T/K signifies the vessel temperature TJK. For NKSCHEM the abscissa T/K = T, -+ AT, and the temperature excesses predicted using this program under non-isothermal conditions are also displayed in fig. 1. Over the initial part of the reactant temperature range, 450-490 K, the predicted extents of conversion (up to 30%) and product formation are exactly matched. Insofar that KINPAK is a well established program that does not have to cope with the non-linearities and stiffness associated with temperature change, the predictions using NKSCHEM are to be regarded as giving a reliable account of the input data.In the reactant temperature range 490-545 K no stationary state exists under the present non-isothermal conditions : oscillatory behaviour is predicted. Stationary states are established again beyond 545 K ; here the conversion of DTBP is virtually complete. Stationary states exist throughout the entire range of temperatures under isothermal conditions. COMPARISONS BETWEEN NUMERICAL AND EXPERIMENTAL RESULTS Although the present emphasis is directed principally to an experimental validation of numerical interpretations, from the thermokinetic point of view it also concerns the similarities and differences between DTBP decomposing in the presence or absence of oxygen.We consider numerical predictions and experimental measurements of stationary and oscillatory reactions for the compositions DTBP (10.0 and 16.5 mol?;) in N, and DTBP (5.0 mol%) in 0,, each at a total pressure of 80 mmHg* and mean residence time (tres) of 3 s in the vessel temperature range 460-540 K. NUMERICAL COMPUTATION OF TEMPERATURE EXCESS AND OSCILLATORY WAVEFORMS Fig. 2 shows the stationary-state temperature excesses that are predicted for DTBP decomposing in nitrogen. At the lowest concentration considered (10 mol % DTBP) * 1 mmHg = 13.5951 x 980.665 x lo-* Pa.348 r 50 -40 -30 M c Q - 20 - 10 THERMOKINETIC OSCILLATIONS IN A C.S.T.R. .- e.-.-.- #. T/K Fig. 1. Numerically computed fraction of reactant and extent of self-heating as a function of gas temperature for 7.5 mol % DTBP in an excess of oxygen.p = 80 torr, t,,, = 3 s. (-) Fraction of reactant remaining predicted by NKSCHEM and (---) that predicted by KINPAK for isothermal conditions; (- * - - -) extent of stationary-state self-heating from NKSCHEM. The open segments of (-) and (---.-) mark a region of oscillatory behaviour in non-isothermal circumstances. 20 - M Q 6 c- 10 = 0 I I I I I 1 1 460 480 500 520 540 560 5 a ~ 600 Ti K Fig. 2. Variations of the extent of self-heating with gas temperature for the exothermic decomposition of 16.5 mol ”/, DTBP in nitrogen (a), and 10 rnol % DTBP in nitrogen (5). p = 80 Torr, tres = 3 s. Solid lines (-) show computer-predicted values of AT, broken lines (---) and points (0, 16.5 mol x) ; A, 10 mol ”) mark experimentally measured values.For 16.5 moi % the open section [lines (a)] mark a realm of oscillatory behaviour. No oscillations are observed or predicted at 10 rnol DTBP. AKs rises smoothly and continuously throughout the range 480 < T,/K -= 540, reaching 14 K at 540 K. When the proportion of DTBP is increased to 16.5 mol % there is a smooth transition to oscillatory reaction at 501 K (T = 51 7 K). A stationary state is restored beyond 509 K ( T = 528 K ) , AGS rising slightly to 22.5 K by T, = 540 K. The oscillations occupy a very narrow realm, their amplitudes are small (see fig. 4 later) and their birth and death occur by a finite growth and decay, respectively, of amplitude as T, is raised. The threshold concentration of DTBP at which such oscillations are predicted is 13 rnol x .J.F. GRIFFITHS, S. M. HASKO, N . K. SHAW AND T. TORREZ-MUJICA 349 Ti K Fig. 3. Variations of the extent of self-heating with gas temperature for a mixture of 5 mol;( DTBP in oxygen. p = 80 Torr, t,,, = 3 s. The solid line (-----) shows the computed prediction; the broken line (- - -) and points (0) show experimentally measured values. The breaks in both curves mark the range of temperatures over which oscillatory behaviour is predicted or measured experimentally. When 5.0 mol2; DTBP reacts in an excess of oxygen the overall pattern is different from either of those described for decomposition in nitrogen (fig. 3). Stationary states exist at the lowest vessel temperatures and ATs rises as Ta is increased. There is then a marked discontinuity in the temperature excess and it heralds the birth of large-amplitude oscillations which exist over a range of vessel temperatures of ca.35 K . There is a gradual diminution in the amplitude and increase in the frequency when Td is raised (fig. 4); oscillations cease by convergence to a stationary state. The stationary temperature excess is close to its maximum at the upper limit for oscillations and it does not vary much when Ta is increased further. There is no oscillatory realm when the initial concentration of DTBP in oxygen is reduced to below 2.5 mol?;. EXPERIMENTAL1 .Y MEASURED STATIONARY TEMPERATURE EXCESSES AND OSCILLATORY REACTION MODE§ Measured stationary-state temperature excesses and their variation with vessel temperature during the decomposition of DTBP in nitrogen and in oxygen are given in fig.2 and 3, respectively. There is excellent agreement in kind with the predicted modes of behaviour at each of the conditions investigated, the quantitative stationary temperature excesses match satisfactorily, and the temperature ranges within which oscillations are predicted are close to those measured experimentally. The distinctions of the two types of oscillatory behaviour are to be seen in the temperature-time records displayed in fig. 5. Measured amplitudes of the oscillations are smaller than those predicted numerically. Oscillatory reaction is not found experimentally when the proportion of DTBP is reduced to below 13 mol yi in nitrogen and below 3 mol % in an excess of oxygen.350 THERMOKINETIC OSCILLATIONS IN A C.S.T.R.160 120 5L: 4 L 80 40 a 485 20 ~ 486 40 ~g G 1 100 1 t f s 508 - I 1 522 Fig. 4. Computed oscillatory temperature excesses for (a) 16.5 mol DTRP in nitrogen and (b) 5 mol % DTBP in oxygen at selected values of Ta. p = 80 Torr, Y,,, = 3 s. This comparison shows the contrasting entry to oscillatory behaviour by the presence of oxygen; the system enters the oscillatory realm via thermal criticality and a hard excitation. The decomposition in nitrogen (even at 16.5 mol :L DTBP) shows a soft excitation into the oscillatory realm. DISCUSSION The purpose of this paper is to lay the foundation for thennokinetic phenomena in organic oxidations. There are fundamental links to theories for stationary states, criticality and oscillatory instability in a gaseous, exothermic, first-order reaction under well stirred flowing conditions.We have not aimed at a definitive interpretation of methyl-radical oxidation at low temperatures; nevertheless, it is pertinent to makeJ. F. GRIFFITHS, S. M. HASKO, N. K. SHAW AND T. TORREZ-MUJICA 35 1 0 2 w a . . . . Y 476 ! 1 I I I I I I 1 I 0 2 4 6 a tlmin Fig. 5. Experimental records of oscillatory modes of reaction for (a) 16.5 mol % DTBP in nitrogen and (b) 5 mol % DTBP in oxygen. p = 80 Torr, t,,, = 3 s. Vessel temperatures are indicated. some comparisons between theory and experiment and to highlight some of the significant features. Associations to degenerately chain-branched reactions and cool flames are also noted. TEMPERATURE CHANGE IN A NON-ADIABATIC C.S.T.R.The agreement between the computed and experimentally measured compositions and vessel temperatures within which oscillations occur is very satisfactory. The discrepancies that exist between the magnitudes of the measured and calculated oscillatory-temperature excursions is attributed principally to our initial premise that352 THERMOKINETIC OSCILLATIONS IN A C.S.T.R. heat capacities do not change with temperature. The heat-transfer coefficient does not have a very significant effect on the amplitudes of oscillations; in them the rates of heat release are high and adiabatic temperature rises (ATad) may be reached. During stationary-state reaction heat transfer through the wall is the predominant heat-loss process and the ratio of the mean residence time to the Newtonian cooling time is important.(This ratio is expressed conveniently as xS/fc.) The highest possible stationary temperature excess that can be achieved is given 29 AT,,(max) = AT(ad)/( 1 +xS/fc) (111) and is 0.05 AT(ad) in the present experimental conditions. The adiabatic temperature excess [AT(ad)] is obtained from the quotient (overall exothermicity per mole DTBP) (inlet concentration of DTBP) volumetric heat capacity Since for the decomposition of DTBP diluted 10-fold by nitrogen AT(ad) = 310 K, AT,,(max) cannot exceed 15 K in our c.s.t.r. at a mean residence time of 3 s. Both experimental and numerical values are in good agreement with this. For decomposition of DTBP ( 5 mol % ) in oxygen the overall, simplified stoichiometryll derived from that measured by Williams et al.13 at 550 K is DTBP + 90, -+ 2(CH3),C0 + 1 .45CH30H + OSCO + 0.1 H, + 8 0 , from which AT(ad) = 806 K and so AT,,(max) = 40 K.The numerical value for AT,,(max) is roughly in accord with this but the experimental value is higher, probably because of different stoichiometry and exothermicity at higher temperatures than that specified in reaction (14). There are no other experimental data available to offer a more satisfactory comparison. AHg8 = -438 kJ mol-1 (14) OSCILLATORY REACTION MODES AND THE NATURE OF TRANSITIONS INTO THEM Sustained oscillatory reaction in a first-order, exothermic reaction under non- adiabatic conditions is possible when the criterion is ~ a t i s f i e d . ~ ~ The term on the left-hand side, the dimensionless adiabatic tem- perature excess (B), may be established experimentally from the minimum proportion of DTBP and the reactant temperature at which oscillations are found.For 13 mol % DTBP decomposing in nitrogen at 510 K, B = 26. B also equals 26 for 2.5 mol % DTBP decomposing in an excess of oxygen, in accord with the simplified stoichiometry of reaction (14). Each of these values corresponds to the limiting compositions found in the present study. The term on the right-hand side of eqn (IV) incorporates each of the characteristic time responses for the system: these are the residence time (t,,,), the Newtonian cooling time (tN = cV/xS) and chemical time [tch = 1 /k( T)]. For DTBP reacting at 5 10 K under a mean residence time of 3 s in our vessel they combine to yield the numerical value in broad agreement with the present results.Whether or not thermal criticality, and hence ‘hard excitation’ into the oscillatory reaction, occurs is determined by the condition B* = AT,,(max) E/RT: > 4+(RT/E). (V)J. F. GRIFFITHS, S. M. HASKO, N. K. SHAW AND T. TORREZ-MUJICA Since, from eqn (111) B* = B/(l + f c / x S ) 353 oscillations [guaranteed by eqn (IV)] may be entered continuously or discontinuously (see for example fig. 5) according to the ratio of the Newtonian cooling time to the mean residence time (= fc/xS). Changes of flow rate have been found to give these alternative transition^.,^ In the present study the different responses of 16.5 mol % DTBP in N, and 5 mol % DTBP in 0, are brought about by a change of exothermicity.Thus B* < 3 for 16.5 mol % DTBP in nitrogen at 510 K. The augmentation of heat output due to the presence of oxygen is sufficient to give B* > 4 even when the concentration of DTBP is much reduced. REACTION MECHANISMS AND THERMOKINETIC IMPLICATIONS The quantitative distinctions between the decomposition of DTBP in the presence of oxygen or nitrogen that are reported here may be accounted for solely on the basis of an enhanced exothermicity of reaction when oxygen is present. Chain branching does not have to be invoked. Although secondary initiation via methyl hydroperoxide decomposition is included, it does not lead to degenerate branching. A degenerately branched reaction would be brought about only if we admit generation of methyl hydroperoxide via H-atom abstraction by CH,O, from DTBP itself.We do not include this step because its activation energy is high ( E > 60 kJ mol-l) and test insertions show no measurable effect. In fact the predicted behaviour is the same qualitatively and not much modified quantitatively if kinetic scheme B is reduced to steps ( l t ( 6 ) . Further paring-down of the scheme by suppression of the dissociation step (4), CH,O, -+ CH,+O,, enhances temperature changes when T, is > ca. 530 K. This is, of course, a matter of degree and is due to enhancement of the overall exothermicity by forcing reaction in the presence of oxygen to the oxidation products [CH,O + CH,OH in the simplest scheme, reactions (1 t(6)]. When the equilibrium (3) (4) CH, + 0, + CH302 is established, C,H, may become a final product even in the presence of oxygen and the route to its formation releases less heat.The shift of equilibrium explains the slight fall in AGS at Ta > 530 K (fig. 2). Ballinger and Ryason14 reported the existence of a stabilised ‘cool flame’ of DTBP on a flat-flame burner. It is our belief that they stabilised a normal near-adiabatic combustion wave, not a ‘cool flame’ in the sense associated with hydrocarbon oxidation. The temperature was not reported, but with modest exothermicity (AH%,, = -480 kJ mol-l) and quite high product heat capacities it may remain sufficiently low that the only light emissions are those commonly associated with ‘cool flames’ of hydrocarbons (CH,O*); Williams et aZ.13 have established DTBP +O, as an excellent source of CH,O*.Fig. 1 demonstrates unequivocally that the existence of non-isothermal oscillatory phenomena in a c.s.t.r. is due to a subtle balance between the rates of heat release and heat loss via the vessel walls. We may take a step, in principle, to the realm of oscillatory cool flames. When reactions of the kind RH + CH,O, -+ CH,O,H + R ( E < 50 kJ mol-l) and RH+X-+R+HX3 54 THERMOKINETIC OSCILLATIONS IN A C.S.T.R. where RH is deemed to be an additional fuel with a labile hydrogen atom and, for mechanistic simplicity leading to R = CH,, are added to reactions (1x8) in kinetic scheme B, the predicted behaviour is significantly modified. Not only does a region of substantial negative temperature coefficient evolve in the overall dependence of reaction rate on temperature in isothermal reactions, but also, in non-isothermal circumstances, oscillatory ' cool flame ' phenomena are now predicted.Their major distinction from the oscillatory modes reported here is that not all of the fuel (RH) is consumed during the course of each cycle.16 Neither of these features is found if reaction (4) is suppressed. We thank Prof. P. Gray and Drs D. L. Baulch and A. J. Duke for helpful discussions, a referee for guidance, and ICI Corporate Laboratory for the use of KINPAK. We also thank the S.E.R.C. and British Gas for support and Interox for the provision of di-t - butyl peroxide. J. H. Raley, F. F. Rust and W. E. Vaughan, J. Am. Chem. SOC., 1948, 70, 88. D. H. Shaw and H. 0. Pritchard, Can.J. Chem., 1968,46,2721. G. Baker, J. H. Littlefair, R. Shaw and J. C. J. Thynne, J. Chem. Soc., 1965, 6970. D. Indritz, J. Stone and F. Williams, J. Chem. Eng. Data, 1978, 23, 6. 0. M. Egeiban, J. F. Griffiths, J. R. Mullins and S. K. Scott, 19th Symp. (Znt.) on Combustion (The Combustion Institute, Pittsburgh, 1983), p. 825. J. F. Griffiths and H. J. Singh, J. Chem. SOC., Faraday Trans. 1, 1982,78, 747. M. F. R. Mulcahy and D. J. Williams, Aust. J. Chem., 1961, 14, 534. D. E. Hoare and C. A. Wellington, 8th Symp. (Znt.) on Combustion (The Combustion Institute, Pittsburgh, 1962), p. 472. J. H. Raley, L. M. Porter, F. F. Rust and W. E. Vaughan, J. Am. Chem. SOC., 1951,73, 496. lo A. Hardacre, G. Skirrow and C. F. H. Tipper, Combust. Flame, 1965, 9, 53. l1 J. F. Griffiths and J. R. Mullins, Combust. Flame, 1984, 56, 135. l2 K. Selby and D. J. Waddington, J. Chem. SOC., Perkin Trans. 2, 1979, 1259. l3 F. W. Williams, R. S. Sheinson, D. J. Bogan, H. A. Rabitz and D. Indritz, presented at Eastern l4 P. R. Ballinger and P. R. Ryason, 13th Symp. (Znt.) on combustion (The Combustion Institute, l5 D. Indritz, H. A. Rabitz and F. W. Williams, J. Phys. Chem., 1977, 81, 2526. l6 P. Gray, J. F. Griffiths, S. M. Hasko and P. G. Lignola, Proc. R. SOC. London, Ser. A , 1981,374,313. l7 L. Salnikov, Dokl. Adad. Nauk SSSR, 1948, 60, 605; Zh. Fiz. Khim., 1949, 23, 258. l9 A. G. Merzhanov and V. G. Abramov, Chem. Eng. Sci., 1977,32,475. 2o V. Caprio, A. Insola and P. G. Lignola, 16th Symp. (Znt.) on Combustion (The Combustion Institute, 21 J. F. Griffiths, P. Gray and K. Kishore, Combust. Flame, 1974, 22, 197. 22 P. Gray, J. F. Griffiths and R. J. Moule, Faraday Symp. Chem. SOC. 1974, 9, 193. 23 S. W. Benson, Thermochemical Kinetics (Wiley-Interscience, New York, 2nd edn, 1976). 24 R. W. Walker, in Reaction Kinetics, ed. P. G. Ashmore (Spec. Period. Rep., The Chemical Society, 25 J. Phys. Chem. Ref. Data, 1982, 11. 26 J. A. Kerr and S. J. Moss, Handbook of Bimolecular and Termolecular Gas Reactions (CRC Press, 27 C. W. Gear, Numerical Initial Value Problems in Ordinary Diflerential Equations (Prentice Hall, 28 I. B. Parker and M. L. Harris (ICI Corporate Laboratory), unpublished work. States Section Meeting, The Combustion Institute, 1976. Pittsburgh, 1971), p. 271. B. F. Gray and C. H. Yang, J. Phys. Chem., 1965, 69, 2747. Pittsburgh, 1983), p. 1155. London, 1975), vol. 1, p. 161. Boca Raton, 1981). Englewood Cliffs, NJ, 197 1). P. Gray, J. F. Griffiths, S. M. Hasko and J. R. Mullins, 8th Znt. Symp. Chemical Reaction Engineering, 1984, 101. 3O K. Edwards, P. Gray, J. F. Griffiths and S. M. Hasko, to be published. (PAPER 4/6 13)

ISSN:0300-9599    

DOI:10.1039/F19858100343

出版商:RSC    

年代:1985

数据来源: RSC