摘要:

THE ROYAL SOCIETY OF CHEMISTRY Journal of the Chemical Society Fa rada y Transact ions Scientific Editor Dr. Peter J. Sarre Department of Chemistry University of Nottingham University Park Nottingham NG7 2RD, UK Faraday Editorial Board Prof. I.W. M. Smith (Birmingham) (Chairman) Prof. M. N. R. Ashfold (Bristol) Dr. B. E. Hayden (Southampton) Dr. D. C. Clary (Cambridge) Prof. A. R. Hillman (Leicester) Dr. L. R. Fisher (Bristol) Prof. J. Holzwarth (Berlin) Prof. H. M. Frey (Reading) Dr. P. J. Sarre (Nottingham) Dr. R. K. Thomas (Oxford) Editorial Manager and Secretary to Faraday Editorial Board Dr. Robert J. Parker The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF, UK Senior Assistant Editors: Mrs.S. Shah, Dr. R. A. Whitelock Assistant Editor: Mrs. C. J. Seeley Editorial Secretary: Mrs. J. E. Gibbs International Advisory Editorial Board R. S. Berry (Chicago) Y. Marcus (Jerusalem) A. M. Bradshaw (Berlin) B. J. Orr (North Ryde) A. Carrington (Southampton) R. H. Ottewill (Bristol) M. Che (Paris) R. Parsons (Southampton) M. S. Child (Oxford) S. L. Price (London) B. E. Conway (Ottawa) F. Rondelez (Paris) G. R. Fleming (Chicago) J. P. Simons (Oxford) R. Freeman (Cambridge) S. Stolte (Amsterdam) H. L. Friedman (Stony Brook) J. Troe (Gottingen) H. lnokuchi (Okazaki) J. Wolfe (Kensington, NSW) J. N. lsraelachvili (Santa Barbara) C. Zannoni (Bologna) M. L. Klein (Philadelphia) A. Zecchina (Turin) R. A. Marcus (Pasadena) C. Zhang (Dalian) Journal of the Chemical Society, Faraday Transactions (ISSN 0956-5000) is published twice monthly by The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF.UK. All orders accompanied with payment should be sent directly to The Royal Society of Chemistry, Turpin Distribution Services Ltd., Black- horse Road, Letchworth, Herts. SG6 IHN, UK. NB Turpin Distribution Services Ltd., dis- tributors, is wholly owned by the Royal Society of Chemistry. 1994 Annual subscription rate EC f744.00, Rest of World f800.00, USA $1400.00, Canada f840 (excl. GST). Customers should make payments by cheque in sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank. Second class postage is paid at Rahway, NJ.Airfreight and mailing in the USA by Mercury Airfreight International Ltd. Inc., 2323 Randolph Avenue, Avenel, NJ 07001, USA and at additional mailing offices. USA Postmaster: send address changes to Journal of the Chemical Society, Faraday Trans- actions, c/o Mercury Airfreight International Ltd. Inc., 2323 Randolph Avenue, Avenel, NJ 07001. All despatches outside the UK by consolidated Airfreight. PRINTED IN THE UK. @ The Royal Society of Chemistry, 1994. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, photographic, recording, or otherwise, without the prior permission of the publishers. Advertisement sales: tel.+44(0)71-287-3091; fax. +44(0)71-494-1134. INFORMATION FOR AUTHORS The Royal Society of Chemistry welcomes submission of manuscripts intended for pub- lication in two forms, Research papers and Faraday Communications. These should describe original work of high quality in the sciences lying between chemistry, physics and biology, and particularly in the areas of physical chemistry, biophysical chemistry and chemical physics. Research Papers Full papers contain original scientific work which has not been published previously. However, work which has appeared in print in a short form such as a Faraday Communi- cation is normally acceptable. Four copies including a top copy with figures etc. should be sent to The Editor, Faraday Transactions, at the Editorial Office in Cambridge.Authors may, if they wish, suggest the names (with addresses) of up to three possible referees. Faraday Communications Faraday Communications contain novel scientific work in short form and of such importance that rapid publication is war-ranted. The total length is rigorously restricted to two pages of the double-column A4 format. For a Communication consisting entirely of text and ten references, with no figures, equations or tables, this cor- responds to approximately 1600 words plus an abstract of up to 40 words. Submission of a Faraday Communication can be made either to The Editor, Faraday Transactions, at the Editorial Office in Cam- bridge or via a member of the International Advisory Editorial Board, who will arrange for the manuscript to be reviewed. In the latter the top copy of the manuscript including any figures etc., together with the name of the person through whom the Com- munication is being submitted, should be sent simultaneously to the Editor at the Cambridge address.Proofs of Communications are not normally sent to authors unless this is specifically requested. Fa rada y Resea rc h Articles Faraday Research Articles are occasional invited articles which are published follow- ing review. They are designed to be topical articles of interest to a wide range of research scientists in the areas of Physical Chemistry, Biophysical Chemistry and Chemical Physics. Full details of the form of manuscripts for Articles and Faraday Communications, con-ditions for acceptance etc. are given in issue number one of Faraday Transactions, published in January of each year, or may be obtained from the Editorial Manager. There is no page charge for papers published in Faraday Transactions. Fifty reprints are supplied free of charge. Dr. P. J. Sarre, Scientific Editor. Tel.: Nottingham (0602) 51 3465 (24 hours) E-Mail (JANET): PCZPSF@UK.AC.NOTT.VAX Fax: (0602) 51 3466 Telex: 37346 UNINOT G Dr. R. J. Parker, Editorial Manager. Tel. : Cambridge (0223) 420066 E-Mail (INTERNET): RSCl @RSC.ORG (For access from JANET use RSC 1 RSC.0RG@ UK.AC. N SF NET-R ELAY) Fax : (0223) 423623 or 420247 Telex: 81 8293 ROYAL G

ISSN:0956-5000    

DOI:10.1039/FT99490FX037

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

FARADAY DIVISION INFORMAL AND GROUP MEETINGS Statistical Mechanics and Thermodynamics Group Cellular Automata and their Applications to Molecular Fluids To be held at the University of Manchester on 19 and 20 July 1994 Further information from Dr A. Masters, Department of Chemistry, University of Manchester, Manchester M13 9PL Division Autumn Meeting: Reactions and Mechanisms for Fine Chemicals To be held at the University of Glasgow on 6-9 September 1994 Further information from Dr J. F. Gibson, The Royal Society of Chemistry, Burlington House, London WlV OBN Gas Kinetics Group 13th International Symposium on Gas Kinetics To be held at University College, Dublin on 11-15 September 1994 Further information from Dr H. Sidebottom, Department of Chemistry, University College, Dublin Electrochemistry Group with the SCI ELECTROCHEM 94 To be held in Edinburgh on 12-16 September I994 Further information from Professor D.E. Williams, Department of Chemistry, University College London, 20 Gordon Street, London WClH OAJ Biophysical chemistry Group with the Industrial Division Biotechnology Group Peptide + Water = Protein To be held at University College, London on 19 September 1994 Further information from Professor J. L. Finney, Department of Physics and Astronomy, University College London, Gower Street, London WClE 6BT British Carbon Group Applications of Microporous Carbons To be held at the University of Leeds on 28 and 29 September 1994 Further information from Professor B. Rand, Department of Chemistry, The University, Leeds LS2 9JT Theoretical Chemistry Group with CCPI Electronic Structure: From Molecules to Enzymes To be held at University College London on 30 November 1994 Further information from Dr P.J. Knowles, School of Chemistry, University of Sussex, Falmer, Brighton BN1 9QJ Division Annual Congress: Lasers in Chemistry To be held at Heriot Watt University, Edinburgh on 1&13 April 1995 Further information from Dr J. F. Gibson, The Royal Society of Chemistry, Burlington House, London W1V OBN Division Joint Meeting with the Division de Chimie Physique de la Societe' Francaise de Chimie, Deutsche Bunsen Gesellschaft fur Physikalische Chemie and Associazione Italiana di Chimica Fisica Fast Elementary Processes in Molecular Systems To be held at the UniversitC De Lille, France on 16-30 June 1995 Further information from Dr C.Troyanowsky, Division de Chimie Physique, Laboratoire de Chimie Physique, 11 rue Pierre et Marie Curie, 75005 Paris, France British Carbon Group Carbon '96 To be held at the University of Newcastle upon Tyne on 7-12 July 1996 Further information from Dr K. M. Thomas, Northern Carson Research Laboratories, The University, Newcastle upon Tyne NE1 7RU Editors: R E Hester, University of York, UK R M Harrison, University of 1Sirmingham, UK A new series fo tackle importan t environmental issues. . . In response to the rapid rowth of interest in the environment and t7le acute need for concise, authoritative and up-to-date reviews of topical issues, the Royal Society of Chemistry is launching Issues in Environmental Science and Technology.This new series will publish review articles on to ics of global concern, written b worlcpexperts in their specialized fie1Js. It will present a multidisciplinary approach to pollution and environmental science and in addition to covering the chemistry of environmental processes, will focus on broader issues such as economic, legal and political considerations. Issues in Environmental Science and Technology will review the effects on human and non-human biota of man-made substances and will provide assessments of the possible practical solutions to perceived environmental problems, . . ......including the worldwide ettorts currently underway to establish ‘Clean Tech no Iogies’. Who will be I Issues in Envi, Science and ?eChnOlOgyC ‘Issues’ will prove invaluable for scientists and engineers in industry, public service, consultancy and academic institutions, who wish to keep up-to-date on topical subjects in this emotive field. It will also be essential reading for students taking specialized courses in environmental chemistry and will provide excellent supplementary reference material for general science courses. Each issue will address a different theme and will contain approximately six articles, most of which will be specially commissioned by the editors. The first two issues will cover: Waste Incinl2 and the Envirt ISBN 0 85404 200 8 April 7994 ISBN 0 85404 205 9 Price f75.00 Pri Also Available ISSN 1350-7583 on s~bscriprion... EC f25.00 USA $47.00 Published twice yearly from 1994 Canada f28.00+ GST Rest of World f27.00 1 ROYAL To order please contact: SOCIETY OF Turpin Distribution Services Ltd, Blackhorse Road, Letchworth CHEMISTRY Herts SG6 1HN, United Kingdom Tel: -1-44 (0)462 672555. Fax: 44 (0)462 480947 RSC Members should order from: Membership Administration, Royal Socie of ChemistryThomas Graham House, Science Park, M%on Road I II Information Cambrid e CB4 4WF, United Kingdom Services Tel: +44 70) 223 420066. Fax: +44 (0)223 423623 0956-5000(1994110:1-4

ISSN:0956-5000    

DOI:10.1039/FT99490BX039

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

ISSN 0956-5000 JCFTEV(10) 1365-1466 (1994) JOURNAL OF THE CHEMICAL SOCIETY Faraday Transactions Physical Chemistry & Chemical Physics CONTENTS 1365 Rotational spectrum of the gas-phase dimer OC. * .BrCl S. Blanco, A. C. Legon and J. C. Thorn 1373 Proton transfer to the fluorine atom in fluorobenzene. Temperature and pressure dependence R. S. Mason, A. J. Parry and D. M. P. Milton 1381 Virial theorem decomposition as a tool for comparing and improving potential-energy surfaces : Ground-state Li, A. A. C. C. Pais, R. F. Nalewajski and A. J. C. Varandas 1391 Pulse radiolytically induced redox and alkylation processes of C,, D. M. Guldi, H. Hungerbiihler, M. Wilhelm and K-D. Asrnus 1397 Electronic transitions in metallocenes by resonance Raman scattering.Part 1.-Analysis of the ferrocene spectrum in the visible region M. L. Ceccarani, P. Sassi and R. S. Cataliotti 1405 Rotational dynamics in liquid water: A simulation study of librational motions I. M. Svishchev and P. G. Kusalik 141 1 Nature of hydrogen bonds formed by phenol derivatives and N,N-dimethylaniline in aprotic solvents : Low-temperature NMR studies M. Ilczyszyn 1415 Comparison of DNA duplexes with and without 04-methylthymine : Nanosecond molecular dynamics simulations L. Cruzeiro-Hansson and J. M. Goodfellow 1429 Kinetics of reduction of hexacyanoferrate(II1) by thiosulfate ions mediated by ruthenium dioxide hydrate A. Mills, X. Li and G. Meadows 1435 Unsteady state, non-isothermal dissolution of a solid particle in liquid J-P. Hsu and B-T.Liu 1441 Electrical properties of pure vanadium phosphate phases and of VPO catalysts used in the partial oxidation of n-butane to maleic anhydride F. Rouvet, J-M. Herrrnann and J-C. Volta 1449 [(q2-C2H4)Os(CO),] as a vibrational model for type I’ ethene chemisorbed as a metallacyclopropane on metal surfaces C. E. Anson, N. Sheppard, D. B. Powell, B. R. Bender and J. R. Norton 1455 Characterization of several y-alumina-supported nickel catalysts and activity for selective hydrogenation of hexanedinitrile F. Medina, P. Salagre, J-E. Sueiras and J-L. Garcia Fierro 1461 Raman spectroscopic study of the Pt-CeO, interaction in the Pt/Al,O,-CeO, catalyst M. S. Brogan, T. J. Dines and J. A. Cairns Note: Where an asterisk appears against the name of one or more of the authors, it is included with the authors’ approval to indicate that correspondence may be addressed to this person.COPIES OF CITED ARTICLES The Royal Society of Chemistry Library can usually supply copies of cited articles. For further details contact: The Library, Royal Society of Chemistry, Burlington House, Piccadilly, London W1V OBN, UK Tel: +44 (0)71-437 8656 Fax: +44 (0)71-287 9798 Telecom Gold 84: BUR210 Electronic Mailbox (Internet) LIBRARY @RSC.ORG. If the material is not available from the Society’s Library, the staff will be pleased to advise on its availability from other sources. Please note that copies are not available from the RSC at Thomas Graham House, Cambridge. F Since April 1st 1994 all authors submitting work for publication in Royal Society of Chemistry journals have been required to sign an exclusive copyright licence, to formalise the agreement with the Society.The form is reproduced overleaf, and may be photocopied; it will HHalso be reproduced in future, as part of Instruction for Authors, in the January issues of the journals. All future submissions of papers for publication should be accompanied by a completed form, without which publication cannot proceed. Paper no (inserted by office): ROYAL SOCIETY OF CHEMISTRY EXCLUSIVE COPY RIGHT LICENCE Authors submitting manuscripts for publication in Royal Society of Chemistry Journals are requested to read the notes below and to enclose with the manuscript a copy of this form, duly completed.Please type, or use BLOCK CAPITALS. Journal to which the manuscript is submitted: Name of Author: Address: Title of Contribution: 2 To be completed if the author(s) is(are) the owner(s) of copyright in the Contribution In consideration of the publication in a Royal Society of Chemistry Journal of the above Contribution, I hereby assign to the Royal Society of Chemistry an Exclusive Licence in respect of the copyright in the Contribution for the full legal term of copyright throughout the world, in all formats, and through any medium of communication. Signed (on behalf of hidherself and of all the authors of the Contribution) Date 3 To be completed if the author(s) is(are) not the owner(s) of copyright in the Contribution In consideration of the publication in a Royal Society of Chemistry Journal of the above Contribution, I, as the authorised representative of the employer of the author(s) of the Contribution, hereby assign to the Royal Society of Chemistry, the publishers, an Exclusive Licence in respect of the copyright in the Contribution for the full legal term of copyright throughout the world in all formats and through any medium of communication, subject to reservation to of the right to reproduce the Contribution at any time for internal purposes. All intellectual property rights other than copyright are reserved.Signed Date Name Employer CONDITIONS OF PUBLICATION To be completed if you are the owner of copyright in the Contribution I warrant to the Royal Society of Chemistry that the Contribution is my original work, has not been published before, that I have obtained all necessary permissions for the reproduction as part of the Contribution of copyright works (including artistic works, eg photographs, charts, maps, etc) not owned by me, that the Contribution contains no illegal statements and does not infringe any rights of others, and agree to indemnify the Royal Society of Chemistry against any claims in respect of the above warranties.Signed Date To be completed if you are not the owner of copyright in the Contribution I, as the authorised representative of the employer of the author of the Contribution, warrant to the Royal Society of Chemistry that all necessary permissions have been obtained for the reproduction as part of the contribution of copyright works (including artistic works, eg photographs, charts, maps, etc) not owned by the employer, that the Contribution contains no illegal statements and does not infringe any rights of others, and agree to indemnify the Royal Society of Chemistry against any claims in respect of the above warranties.Signed Date Name Employer Copyright: notes for contributors 1 The Society’s policy is to acquire an Exclusive Licence in respect of the copyright in all contributions. There are two reasons for this: (a) control of copyright by the publisher ensures maximum protection against piratical infringement anywhere in the world; (b) it also ensures that re uests by third arties to reprint a contribution, or part of it, are handled efficiently in accordance with our general policy which encourages dissemination of knowle%ge inside the lamework of copyright.2 We will not withhold permission for any reasonable request from you to publish the whole or any art of your Contribution in connection with any other work by you, provided the usual acknowledgements are given regarding copyright notice and reference to first pubyication by us. 3 The Royal Society of Chemistry is a si natory of the STM Guidelines on Permissions, whereb the Society has agreed to a standard rocedure for granting permission for the reproduction of copyright mated (especially figures and similar illustrations) from the iocie s ublications b other bonafrPde publishers, without charge Fd without reference to the onginal authors, except where he quantity of material to be reproduced is judge%o ge unreasonably large.A standard condition of such pemssion is that the original authors and the source are acknowledged. Accordingly, the Society will not normally refer to you requests for permission to reproduce figures, tables etc from your contribution, and you are requested to forward to the Society any requests for rmssion which you receive from other authors or their publishers in respect of this contribution. The Society believes that this procedure protects Authors’ interests wit out adding an excessive amount of bureaucracy. 4 The Journal mandates the Copyright Licensing Agency in the UK, which offers centralised licensing arrangements for photocopying in the UK and has reciprocalarrangements with licensing agencies in other countnes. 5 The Contribution may be stored by electronic (including digital) means by us and then transmitted to meet individual requests.6 You hereby agree to the journal making the necessary arrangements to include the Contribution in a document delivery service. 7 If you are a US Government employee and the Contribution was made in that capacity the assignment applies only to the extent allowable by US law. 8 If ou are an employee of the British Government then HMSO will, as standard practice, grant a non-exclusive licence to publish this paper in the journal in any form or meJa provided British crown copyright is reserved. 9 If the Contribution is not accepted for publication by the Society, this Exclusive Licence is null and void.

ISSN:0956-5000    

DOI:10.1039/FT99490FP093

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

Cumulative Author Index 1994 Aas,N., 1015 Afanasiev, P., 193 Aikawa, M., 911 Aitken, C. G., 935 Burdisso, M., 1077 Busca, G., 1161,1293 Butt, M. D., 727 Byatt-Smith, J. G., 493 El Baghdadi, A., 1313 Elisei, F., 279 Elliot, A. J., 831, 837 Engberts, J. B. F. N., 727 Herzog, B., 403 Heyes, D. M., 1133 Higgins, S., 459 Hindermann, J-P., 501 Lei, G-D., 233 Lerner, B. A,, 233 Leslie, M., 641 Li, J., 39 Akanuma, K., 1171 Cabaleiro, M. C., 845 Enomoto, N., 1279 Hirst, D. M., 517 Li, P., 605 Akolekar, D. B., 1041 Albery, W. J., 11 15 Caceres, M., 12 17 Caceres Alonso, M., 553 Eustaquio-Rincon, R., Ewins, C.. 969 113 Hiyane, I., 973 Hoekstra, D., 727 Li, X., Li, Y., 1429 947 Aldaz, A., 609 Cairns, J. A., 1461 Fantola Lazzarini, A. L., Holmberg, B., 559 Liang, Y., 1271 Alfimov, M.V., 109 Calado, J. C. G., 649 423 Holz, M., 849 Lin, J., 355 Al-Ghefaili, K. M., 383, Caldararu, H., 213 Fausto. R., 689 Hoshino, H., 479 Lincoln, S. F., 739 1047 Calvente, J. J., 575 Favaro, G., 279, 333 Hosoi, K., 349 Lindblom, G., 305 Ah, V., 579, 583 Calvo, E. J., 987 Feliu, J. M., 609 HSU, J-P., 1435 Liu, B-T., 1435 Aliev, A. E., 1323 Allegrini, P., 333 Allen, N. S., 83 Al Rawi, J. M. A., Amorim da Costa, A. M., 845 689 Camacho, J. J., 23 Cameron, B. R., 935 Campa, M. C., 207 Campos, A., 339 Canosa-Mas, C. E., 1197, 1205 Filimonov, I. N., 219, 227 Flint, C. D., 1357 Fogden, A., 263 Fornes, V., 213 Fracheboud, J-M., 1197, 1205 Hungerbiihler, H., 1391 Hutchings, G. J., 203 Hutton, R.S., 345 Iizuka, Y., 1301, 1307 Ikawa, S-i., 103 Ikonnikov, I. A,, 219 Liu, C-W., 39 Liu, X., 249 Loginov, A. Yu., 219, 227 Lohse, U., 1033 Longdon, P. J., 3 15 Lorenzelli, V., 1293 Amoskov, V. M., 889 Ando, M., 1011 Andrews, S. J., 1003 Anson, C. E., 1449 Capobianco, J. A,, 755 Caragheorgheopol, A., 213 Carlile, C. J., 1149 Carlsen, L., 941 Franck, R., 667,675 Freeman, N. J., 751 Frety, R., 773 Frey, J. G., 17, 8 17 Ilczyszyn, M., 1411 Indovina, V., 207 Inoue, Y., 797, 815 Ishiga, F., 979 Lunelli, B., 137 Ma, J., 1351 Mabuchi, M., 899 Lu, J-X., 39 Aragno, A,, 787 Arai, S., 1307 Aramaki, K., 321 Carvill, B. T., 233 Castaiio, R., 1227 Castro, S., 1217 Frostemark, F., 559 Fujiwara, Y., 1183 Gandolfi, R., 1077 Ishigure, K., 93, 591 Isoda, T., 869 Ito, O., 571 Machado, V.G., 865 Mackie, J. C., 541 Mackintosh, J. G., 1121 Aravindakumar, C. T., 597 Catalina, F., 83 Gans, P., 315 Iwasaki, K., 121 Macpherson, A. N., 1065 Asai, Y., 797 Cataliotti, R. S., 1397 Gao, Y., 803 Jacobs, W. P. J. H., 1191 Madariaga, J. M., 1227 Ashfold, M. N. R., 1357 Cavasino, F. P., 3 11 Garcia, R., 339 Jakubov, T., 783 Maeda, T., 899 Asmus, K-D., 1391 Ceccarani, M. L., 1397 Garcia Fierro, J-L., 1455 Jameel, A. T., 625 Maestre, A., 575 Avila, V., Baba, T., Badri, A,, 69 187 1023 Chang, T-h., 1157 Charlesworth, P., 1073 Chen, J-S., 429, 71 7 Garcia-Paiieda, E., Gautam, P., 697 Geantet, C., 193 575 Janchen, J., 1033 Jayakumar, R., 161 Jayasooriya, U. A., 1265 Maginn, S.J., 1003 Mahy, J. W. G.. Maity, D. K., 703 327. 1363 Bagatti, M., 1077 Chen, Y-H., 617 Gengembre, L., 895 Jenneskens, L. W., 327, Makarova, M. A,, 383, Ball, M. C., 997 Ball, S. M., 523 Baonza, V. G., 553 Cheng, A., 253 Cheng, C. P., 1157 Cherqaoui, D., 97 Gil, A. M., 1099 Gil, F. P. S. C., 689 Gilchrist, J., 1149 Jennings, B. J., 55 Jiang, D-z., 1351 1363 Maksymiuk, K., 745 Malatesta, V., 333 1047 Baonza, V. G., 1217 Barbaux, Y., 895 Chesta, C. A., Chevalier, S., 69 667, 675 Gill, D. S., Gill, J. B., 579, 583 315 Jiang, P-Y., Jiang, P. Y., 591 93 Malcolm, B. R., 493 Mallon, D., 83 Barthomeuf, D., 667,675 Chmiel, G., 1153 Goede, S. J., 327, 1363 Jobic, H., 1191 Mandal, A. B., 161 Basini, L., 787 Cho, T., 103 Gomez, C.M., 339 Johansson, L. B.-A., 305 Marcheselli, L., 859 Bassoli, M., 363 Christensen, P., 459 GonGalves da Silva, A. M., Johari, G. P., 883, 1143 Marchetti, A., 859, 1089 Battaglini, F., 987 Climent, M. A., 609 649 John, S. A,, 1241 Mariani, M., 423 Bauer, C., 51 7 Bell, A. J., 17, 817 Belton, P. S., 1099 Bender, B. R., 1449 Coates, J. H., 739 Colmenares, C. A., 1285 Cordischi, D., 207 Corma, A., 213 Goodfellow, J. M., 1415 Gouder, T. H., 1285 Gray, P. G., 369 Green, W. A., 83 Joseph, E. M., 387 Joshi, P. N., 387 Kagawa, S., 349 Kakuta, N., 1279 Martins, A., 143 Maruya, K-i., 9 1 1 Masetti, F., 333 Mason, R. S., 1373 Bendig, J., 287 Cormier, G., 755 Grein, F., 683 Kaler, E. W., 471 Massucci, M., 445 Bengtsson, L.A., 559 Corradini, F., 859, 1089 Grieser, F., 1251 Kalugin, 0.N., 297 Matijevic, E., 167 Benko, J., 855 Benniston, A. C., 953 Corrales, T., 83 Cosa, J. J., 69 Grifith, W. P., Grimshaw, J., 1105 75 Karge, H. G., 1329 Kato, R., 763 Matsuda, J., 321 Matsumura, Y., 1177 Bensalern, A., 653 Cottier, D., 1003 Grzybowska, B., 895 Katsumura, Y., 93, 591 May, I. P., 751 Berces, T., 41 1 Coudurier, G., 193 Guelton, M., 895 Kaur,T., 579 Mazzucato, U., 333 Bergeret, G., 773 Courcot, D., 895 Guillaume, F., 1313 Kawashima, T., 127 McGilvery, D., 1055 Beutel, T., 1335 Crawford, M. J., 817 Guldi, D. M., 1391 Keil, M., 403 Mchedlov-Petrossyan, N. O., Beyer, H. K., 1329 Cruzeiro-Hansson, L., 1415 Gulliya, K.S., 953 Kemball, C., 659 629 Bickelhaupt, F., 327, 1363 Biczok, L., 41 1 Cullis, P. M., Curtis, J. M., 727 239 Hachey, M., 683 Haeberlein, M., 263 Kessel, D., 1073 Kida, I., 103 McNaughton, D., 1055 Meadows, G., 1429 Biggs, P., Binet, C., 1197, 1205 1023 Dang, N-T., 875 Danil de Namor, A. F., 845 Hall, D. I., 517 Hall, G., 1 Kiennemann, A,, Kim, J-H., 377 501 Medforth, C. J., 1073 Medina, F., 1455 Black, S. N., 1003 Blackett, P. M., 845 Blanco, S., 1365 Blandamer, M. J., 727 Das, T. N., 963 Dasannacharya, B. A,, Davey, R. J., 1003 Davidson, K., 879 1149 Hallbrucker, A., 293 Halpern, A., 721 Hamnett, A., 459 Hancock, G., 523 Kimura, M., 1355 King, F., 203 Kirschner, J., 403 Kita, H., 803 Melrose, J. R., 1133 Merga, G., 597 Meunier, F., 369 Mezyk, S.P., 831 Blower, C., 919,931 Demeter, A., 41 1 Handa, H., 187 Klein, M. L., 253 Mills, A,, 1429 Boddenberg, B., 1345 Boggis, S. A,, 17 Dempsey, P., 1003 Demri, D., 501 Hann,K., Hao, L., 733 133, 1223 Kleshchevnikova, V. N., 629 Milton, D. M. P., Min, E-z., 1351 1373 Borge, G., 1227 Borisenko, V. N., 109 Boutonnet-Kizling, M., Derrick, P. J., 239 Dewing, J., 1047 Diagne, C., 501 Harada, S., 869 Haraoka, T., 911 Harland, P. W., 935 Knozinger, H., Kobayashi, A,, Kobayashi, H., 1335 763 763 Misono, M., 1183 Mitchell, P. J., 1133 Mittal, J. P., 597, 703, 71 1, Bowker, M., 1023 1015 Dickinson, E., 173 Dines, T. J., 1461 Harper, R. J., Harriman, A,, 659 697,953 Kobayashi, T., 1011 Kondo, Y., 121 Miyake, Y., 825 979 Bozon-Verduraz, F., 653 Bradley, C.D., 239 Bradshaw, A. M., 403 Doblhofer, K., 745 Domen, K., 91 1 Dossi, C., 1335 Harris, K. D. M., 1313, Harrison, N. J., 55 1323 Kossanyi, J., 41 1 Kurrat, R., 587 Kusalik, P. G., 1405 Mizuno, N., 1183 Mizushima, T., 1279 Moffat, J. B., 1177 Braun, B. M., 849 Doughty, A., 541 Haruta, M., 1011 Kuwamoto, T., 121 Mohan, H., 597,703 Breysse, M., 193 Briggs, B., 727 Brocklehurst, B., 271 Douglas, C. B., Dunmur, D. A,, Dunstan, D. E., 471 1357 1261 Hashimoto, K., 1177 Hashino, T., 899 Hattori, H., 803 Laachir, A., 773 tajtar, L., 1153 Lambert, J-F., 667,675 Monk, P. M. S., 1127 Mordi, R. C., 1323 Moriguichi, I., 349 Brogan, M. S., 1461 Brown, N. M. D., 1357 Duxbury, G., 1357 Dwyer, J., 383, 1047 Haymet, A.D. J., 1245 Heal, M. R., 523 Lamotte, J., 1029 Langan, J. R., 75 Morikawa, A,, 377 Morioka, Y., 1279 Brown, R. G., 59 Brown, S. E., 739 Bruna, P. J., 683 Dyke, J. M., 17 Eastoe, J., 487 Easton, C. J., 739 Healy, T. W., 1251 Heenan, R. K., 487 Helmer, M., 31, 395 Lavalley, J-C., 1023, 1029 Lavanchy, A., 783 Lizar, K., 1329 Morokurna, M., 377 Morrison, C. A,, 755 Mount, A. R., 11 15, 1121 Brzezinski, B., 843, 1095 Ebitani, K., 377 Herein, D.. 403 Lazzarini, E., 423 Muir, A. V. G., 459 Buckley, A. M., 1003 Buemi, G., 121 1 Egsgaard, H., El-Atawy, S., 941 879 Herod, A. A., 1357 Herrmann. J-M., 1441 Leaist, D. G., Legon, A. C., 133, 1223 I365 Mukherjee, T., 71 1 Mukhopadhyay, R., 1149 1 Nagaishi, R., 93, 591 Nagaoka, H., 349 Naito, S., 899, 1355 Naito, T., 763 Nalewajski, R.F., 1381 Navaratnam, S., 83 Neoh, K. G., 355 Nerukh, D. A., 297 Nicholson, D., 181 Nickel, U., 617 Ninomiya, J., 103 Nishihara, H., 321 Nogami, T., 763 Nonaka, O., 121 Norton, J. R., 1449 Nuiiez, J., 1217 Nuiiez Delgado, J., 553 Nyholm, L., 149 Occhiuzzi, M., 207,905 Ohji, N., 1279 Ohtsu, K., 127 Okamura, A., 803 Olazabal, M. A., 1227 Olejnik, J., 1095 Oliveri, G., 363 Onishi, T., 91 1 Ono, Y., 187 Oiadd, G., 305 Ortica, F., 279 Oswal, S. L., 1083 Ota, K-i., 155 Otlejkina, E. G., 297 Otsuka, K., 451 Ottavi, G., 333 Ouellette, D. C., 837 Owari, T., 979 Ozutsumi, K., 127 Padley, M. B., 203 Pais, A. A. C. C., 1381 Pal, H., 711 Pal-Borbely, G., 1329 Palleschi, A., 435 Paradisi, C., 137 Pardo, A., 23 Parry, A.J., 1373 Parsons, B. J., 83 Patel, S. G., 1083 Pathmanathan, K., 1143 Patrykiejew, A., 1153 Paul, D. K., 1271 Pavanaja, U. B., 825 Pedulli, G. F., 137 Peeters, M. P. J., 1033 Peng, W., 605 Pepe, F., 905 Pereira, C. M., 143 Perez, J. M., 609 Perrichon, V., 773 Sapre, A. V., 825 Sarre, P. J., 517 Sassi, P., 1397 Takahashi, K., 155 Takasawa, A., 91 1 Tamaura, Y., 1171 Villemin, D., 97 Visscher, P. B., 11 33 Vlietstra, E. J., 327, 1363 Peter, L. M., 149 Petrov, N. Kh., 109 Pispisa, B., 435 Pivnenko, N. S., 297 Plane, J. M. C., Plowman, R., 1003 Porcar, I., 339 Potter, C. A. S., 59 Powell, D. B., 1449 Poyato, J. M. L., 23 Prenosil, J. E., 587 Previtali, C. M., 69 Pringle, T.J., 1015 Priyadarsini, K. I., 963 Pryamitsyn, V. A,, 889 Psaro, R., 1335 31, 395 Sato, K., 797 Saur, O., 1029 Sbriziolo, C., 31 1 Schedel-Niedrig, Th., 403 Schlogl, R., 403 Schnabel, W., 287 Scremin, M., 865 Seddon, B. J., 605 Seidel, A., 1345 Sellen, D. B., 1357 Shahid, G., 507, 5 13 Shallcross, D. E., 1197, Sharma, A., 625 Shaw, N., 17,817 Sheil, M. M., 239 1205 Tamura, K-i., 533 Tanaka, I., 349 Tanigaki, H., 1307 Taravillo, M., 1217 Tassi, L., 859, 1089 Tateno, A., 763 Tatham, A., 1099 Taylor, A., 1003 Taylor, M.G., 641 Teixeira-Dias, J. J. C., Teo, W. K., 355 Teramoto, M., 979 Teraoka, Y., 349 689 Thompson, K. M., 1105 Thompson, N. E., 1047 Thorn, J. C., 1365 Vollarova, O., 855 Vollmer, F., 59 Volta, J-C., 1161, 1441 Vyunnik, I. N., 297 Wales, D.J., 1061 Wang, C. F., 605 Wang, J., 1245 Watanabe, H., 571 Waters, M., 727 Wayne, R. P., 1197,1205 Weckstrom, K., 733 Weingartner, H., 849 Weir, D. J., 751 Werner, H., 403 Whitaker, B. J., 1 White, L. R., 1251 Rabold, A., 843 Ramaraj, R., 1241 Rama Rao, K. V. S., Ramis, G., 1293 Ramsden, J. J., 587 Rao, B. S. M., 597 Rastelli, A., 1077 Rehani, S. K., 583 Rettig, W., 59 Rey, F., 213 Rezende, M. C., 865 Rhodes, N. P., 809 Ricchiardi, G., 1161 Richter, R., 17 Robertson, E. G., 1055 Rocha, M., 143 Rochester, C. H., 203 Rodes, A., 609 Rofia, S., 137 Rosenholm, J. B., 733 Rosmus, P., 517 Rosseinsky, D. R., 1127 Rossi, P. F., 363 Rout, J. E., 1003 Rouvet, F., 1441 Rudham, R., 809 Ryde, N., 167 Sacco, A., 849 Sachtler, W. M. H., 233, Saitoh, T., 479 Salagre, P., 1455 Salmon, G.A., 75 Sam, D. S. H., 1161 Sanada, M., 1307 Sano, T., 869 825 1335 Shen, J-p., 1351 Sheppard, N., 507,513, Sherwood, P. M. A., 1271 Shiao, J-C., 429 Shihara, Y., 549 Shiralkar, V. P., 387 Shishido, T., 803 Shizuka, H., 533 Siders, P., 973 Silva, C. J., 143 Silva, F., 143 Simkiss, K., 641 Singh, J., 579, 583 Singh, R., 583 Smart, S. P., 1313 Smith, K. M., 1073 Smith, T. D., 919,931 Soares, V. A. M., 649 Sokotowski, S., 11 53 Soria, V., 339 Spiro, M., 617, 1105 Stanley, D. R., 1003 Stewart, B., 969 Stoeckli, F., 783 Sueiras, J-E., 1455 Sun, L. M., 369 Sun, T., 1351 Suquet, H., 667,675 Surov, Y. N., 297 Suzuki, T., 549 Svishchev, I. M., 1405 Tabata, M., 1171 Tabrizchi, M., 17 Tagliazucchi, M., 859, 1089 Takagi, T., 121 1449 Timms, A. W., 83 Timney, J.A., 459 Togawa, T., 1 17 1 Tomkinson, J., 1149 Tosi, G., 859, 1089 Touret, O., 773 Tournayan, L., 773 Trau, M., 1251 Trejo, A., 113 Treviiio, H., 1335 Truscott, T. G., 1065, 1073 Tsuchiyama, T., 1355 Tsuji, H., 803 Tsuji, M., 1171 Tsunashima, S., 549 Tsunetoshi, J., 1307 Tung, C-H., 947 Turco Liveri, M. L., 311 Turco Liveri, V., 311 Turner, P. H., 1065 Udagawa, T., 763 Ueno, A., 1279 Ugo,R., 1335 Umemoto, H., 549 Unayama, S-i., 549 Upadhyaya, H. P., 825 Valat, P., 411 Valls, M. J., 609 van Hooff, J. H. C., van Santen, R. A., 1191 van Wolput, J. H. M. C., Varandas, A. J. C., 1381 Vedrine, J. C., 193 Venanzi, M., 435 Villamagna, F., 47 1033 1033 Whitehead, M. A., 47 Wikander, G., 305 Wilde, C. P., 1233 Wilhelm, M., 1391 Williams, D. E., 345 Wilpert, A,, 287 Wintgens, V., 411 Woermann, D., 875 Wohlers, M., 403 Wolthuizen, J.P., 1033 Wormald, C. J., 445 Xin, Q., 973 Yagci, Y., 287 Yamaji, M., 533 Yamamoto, M., 899, 1355 Yamanaka, I., 451 Yamasaki, M., 869 Yamauchi, N., 1307 Yanes, C., 575 Yang, Z-Q., 947 Yano,H., 869 Yasuda, H., 1183 Yeh, C-t., 1157 Yoshitake, H., 155 Yotsuyanagi, T., 93,479 Young, R. N., 271 Zanotto, S. P., 865 Zhang, M., 1233 Zhang, X., 605 Zhang, Z. C., 1335 Zholobenko, V. L., 233, Zhong, G. M., 369 Ziolek, M., 1029 Zubarev, V. E., 721 Zundel, G., 843,1095 1047 The following papers were accepted for publication between 1st and 3 1st March 1994: General thermodynamic analysis of the dissolution of non-polar molecules into water. Origin of hydrophobicity M. Costas, B. Kronberg and R.Silveston Dissolution of amorphous ahminosilicate zeolite precursors in alkaline solutions. Part 2.-Mechanism of dissolution B. Subotik, T. Antonik and A. &mek Two members of the ABC-D6R family of zeolites: Zeolite Phi and Linde D K. P. Lillerud, R. Szostak and A. Long Pitzer model parameters for sparingly soluble salts from solubility measurements: Thallium(1) chloride in aqueous solutions of ammonium chloride, rubidium chloride and caesium chloride at 298.15 K K. H. Khoo, K. R. Fernando and L-H. Lim Irreversible thermodynamic coupling between heat and matter fluxes across a gadliquid interface S. C. Doney Micellar aggregates of sodium glycocholate and taurocholate and their interaction complexes with bilirubin-IXa. Structural models and crystal structure E.Giglio, M. D'Alagni, L. Galantini, E. Gavuzzo and L. Scaramuzza Hydration of polar interfaces. A generalized mean-field model G. Cevc and S. Kirchner Photoinduced electron transfer in a-helical poly(L-lysine) carrying randomly distributed donor-acceptor pairs. A kinetic and conformational statistics investigation B. Pispisa, M. Venanzi and A. Palleschi Hydrogenation behaviour over Si0,-supported lanthanide-palladium bimetallic catalysts with considerable hydrogen uptake H. Imamura, K. Igawa, Y. Kasuga, Y. Sakata and S. Tsuchiya A6 initio quantum chemistry study of the gas-phase reaction of CIO with HO, D. M. Hirst and D. Buttar Glass transition of liquid-crystalline 4-alkoxyphenyl and 4-cyanophenyl 4-(2,4-dialkoxybenzoyloxy)benzoates S.Takenaka and H. Yamasu Quartz crystal microbalance study of the adsorption of ions onto gold from non-aqueous solvents A. P.Abbott, D. C. Loveday and A. R. Hillman Studies of silver electronucleation onto carbon microelectrodes J. Sousa, S. Pons and M. Fleischmann Mechanism of the atmospheric oxidation of 1,l. 1,2-tetrafluoroethane (HFC 134a) 0. V. Rattigan, D. M. Rowley, 0. Wild, R. L. Jones and R. A. Cox Use of a neural network to determine the normal boiling points of acyclic ethers, peroxides, acetals and their sulfur analogues D. Villemin, D. Cherqaoui, A. Mesbah, J-M. Cense and V. Kvasnicka Host-guest complexes of cucurbituril with the 4-methylbenzylammonium ion, alkali-metal cations and NH,' W. Knoche, R.Hoffmann, C. Fenn and H-J.Buschmann X-Ray photoelectron, temperature-programmed desorption and temperature-programmed reduction study of LaNiO, and La,NiO,+& catalysts for methanol oxidation V. Rives, J. Choisnet, N. Abadzhieva, J. M. Bassat, P. Stefanov, L. Minchev and D. Klissurski Addition of manganese to iron catalysts supported on silicalite-1 and its effect on the CO hydrogenation reaction D. K. Chakrabarty, G. Ravichandran and D. Das Characterisation of iron/titanium oxide photocatalysts. Part 2.-Surface studies R. I. Bickley, T. Gonzalez-Carreiio, A. R.Gonzalez-ElipC, G. Munuera and L. Palmisano Mechanistic aspects of biological redox reactions involving NADH. Part 5.-AM 1 transition state studies for the pyruvate-L-lactate interconversion in L-lactate dehydrogenase J.E. Gready and S. Ranganathan Group behaviour of SAPO-11 molecular sieves containing various metals (Mg, Zn, Mn or Cd, Ni, Cr) J. Kornatowski, G. Finger, K. Jancke, J. Richter-Mendau, D. Schultze, W. Joswig and W. H. Baur Configuration interaction study of the 02-C2H4 exciplex: Collision-induced probabilities of spin-forbidden radiative and non-radiative transitions H. Agren, B. F. Minaev and V. V. Kukueva Monte Carlo simulations of a single polyoxyethylene C,,E, chain headgroup fixed on a bilayer surface ir? water Y. C. Kong, D. Nicholson, N. G. Parsonage and L. Thompson Kinetics of self-replicating micelles J. Billingham and P. V. Coveney ... 111 Mechanistic study of sec-butyl alcohol dehydration on zeolite H-ZSM-5 and amorphous aluminosilicate J.M. Thomas, M. A. Makarova, C. Williams and K. I. Zamaraev Raman band shifts of y-Bi,MoO, and a-Bi,Mo,O,, exchanged with '*O tracer and active sites for reoxidation T. Ono and N. Ogata Transition vector symmetry and the internal pseudo-rotation and inversion paths of ClF,+ D. J. Wales and R. M. Minyaev Gradient-line reaction paths for 1,2 H shift reactions in phosphinonitrene and formaldehyde, and H, elimination from formaldehyde D. J. Wales and R. M. Minyaev Ionic partial molar volumes in non-aqueous solvents Y.Marcus, G. Hefter and T-S. Pang Self-diffusion and viscoelasticity of dense hard-sphere colloids D. M. Heyes and P. J. Mitchell Normal and anomalous positronium states in ionic and molecular solids investigated via magnetic field effects G.DuplPtre, T. Goworek and A. Badia Dynamics calculations and isotopic effect in 0 + OH(D)-+O, + H(D) at low energies A. J. C. Varandas, J. M. C. Marques and W. Wang Grand Canonical Monte Car10 study of Lennard-Jones mixtures in slit pores. Part 3.-Mixtures of two molecular fluids: Ethane and propane R. F. Cracknell and D. Nicholson Electroanalytical/X-ray photoelectron spectroscopy investigation on glucose oxidase adsorbed on platinum C. Malitesta, G. E. De Benedetto and C. G. Zambonin Heterogeneous catalysis in solution. Part 27.-Reaction between titanium(II1) and triiodide ions catalysed by platinum M. Spiro and S. Xiao Non-linear optical properties of organic molecules. Part 13.-Calculation of the structure and frequency- dependent hyperpolarisability of a blue azothiophene dye J.0. Morley Non-linear optical properties of organic molecules. Part 14.-Calculations of the structure, electronic properties and hyperpolarisabilities of cyclopentadienylpyridines J. 0. Morley Collisional behaviour with Ar of the A doublets of CH(X * n)N"=15 produced in the two-photon dissociation of CH,CO at 279.3 nm G. Hancock, S. M. Ball and M. R. Heal Catalytic combustion of methane: Copper oxide supported on high-specific-area spinels synthesized by a sol-gel process M. Primet and N. Guilhaume Effects of hydrogen and deuterium concentration on measurements of the solubility and diffusivity of hydrogen isotopes in yttrium S. Naito, T. Maeda, M. Yamamoto and M. Mabuchi Ultra-low-temperature kinetics of neutral-neutral reactions: Rate constants for the reactions of OH radicals with butenes between 295 and 23 K I.R.Sims, P. Bocherel, A. Defrance, D. Travers, B. R. Rowe and I. W. M. Smith Redox properties of ubiquinone (UQl0) adsorbed on a mercury electrode D. J. Schiffrin and G. J. Gordillo Micellisation and gelation of triblock copolymers of ethylene oxide and E-caprolactone, CL,E,CL,, in aqueous solution D. Attwood, L. Martini, J. H. Collett, C. V. Nicholas, S. Tanodekaew, N-J. Deng, F. Heatley and C. Booth Differential scanning microcalorimetric study of sodium di-n-dodecylphosphate vesicles in aqueous solution M. J. Blandamer, B. Briggs, P. M. Cullis, J. B. F. N. Engberts and D. Hoekstra EPRENDOR characterization of radicals produced in the photopolymerization of a dimethacrylate monomer E.Selli, C. Oliva and G. Termigone Very large thermal separations for polyelectrolytes in salt solutions D. G. Leaist and L. Hao Pulse radiolytically induced redox and alkylation processes of C7,) K-D. Asmus, D. M. Guldi, H. Hungerbuhler and M. Wilhelm iv THE ROYAL SOCIETY OF CHEMISTRY, FARADAY DIVISION, GENERAL DISCUSSION 98 Polymers at Surfaces and Interfaces University of Bristol, 12-14 September 1994 Organising Committee: Professor Sir Sam Edwards (Chairman) Dr R. Buscall Professor R. H. Ottewill Dr T. Cosgrove Professor J. S. Higgins Dr R. W. Richards Dr R. A. L. Jones New experimental methods and new theoretical and computational techniques have recently led to great progress in understanding the difficult but technologically important problems associated with the conformation of polymer molecules at surfaces and interfaces.The purpose of this Discussion is to bring together experimentalists and theoreticians working towards a molecular understanding of polymers at surfaces and interactions to survey the progress in the area to date and to indicate future directions of research. The meeting will attempt to bring a unified approach to the problem, encompassing problems of the structure of surfaces and interfaces in polymer melts, the conformation of polymers at solifliquid and 1iquidAiquid interfaces, and extensions towards more complicated biological systems. The preliminary programme may be obtained from Mrs Angela Fish, The Royal Society of Chemistry, Burlington House, Piccadilly, London W1 V OBN.THE ROYAL SOCIETY OF CHEMISTRY, FARADAY DIVISION, GENERAL DISCUSSION 99 Vibrational Optical Activity: from Fundamentals to Biological Applications University of Glasgow, 19-21 December 1994 Organising Committee Professor L. D. Barron (Chairman) Dr A. F. Drake Dr D. L. Andrews Professor R. E. Hester Professor A. D. Buckingham Traditional optical activity measurements such as CD are confined to the visible and near-ultraviolet spectral regions where they provide stereochemical information on chiral molecules via polarized electronic transitions. Thanks to prompting from theory and new developments in instrumentation, optical measurements are now being made in the vibrational spectrum using both infrared and Raman methods.Studies over the past decade on a large range of chiral molecules, from small organics to biological macromolecules, have demonstrated that vibrational optical activity opens up a whole new world of fundamental studies and practical applications undreamt of in the realm of conventional electronic optical activity. The meeting seeks to bring together experimentalists and theoreticians to discuss the current and future experimental possibilities and the development of theories, including ab initio computational methods, which can relate the observations to stereochemical details. The increasing importance now being attached to molecular chirality and solution conformation in the life sciences should also encourage the partipation of biomolecular scientists.The preliminary programme may be obtained from Mrs Angela Fish, The Royal Society of Chemistry, Burlington House, London WlV OBH. V THE ROYAL SOCIETY OF CHEMISTRY, FARADAY DIVISION, GENERAL DISCUSSION 100 Atmospheric Chemistry: Measurements, Mechanisms and Models University of East Anglia, Norwich, 19-21 April 1995 Organising Committee: Professor 1. W. M. Smith and Dr J. R. Sodeau (Co-chairmen) Dr R. A. Cox Dr J. C. Plane Dr J. Pyle Professor F. Taylor The priority now given by national governments to the study of atmospheric science confirms that our understanding of global climate and compositional changes depends upon measurements in both the laboratory and the field.The data obtained by the experimentalists are then applied by modellers who provide the most significant input into legislative controls on pollution matters. However there have been few opportunities for laboratory and field workers along with the modelling community to attend an “interdisciplinary” discussion in which overall progress in our understanding of specific atmospheric problems is assessed. The object of this discussion is to bring together the researchers in the diverse disciplines that make up atmospheric chemistry so that their individual results and conclusions can be communicated to each other. Some of the key issues to be discussed will include: ozone balances in the atmosphere; heterogeneous processes; the interaction of chemistry and dynamics in determining atmospheric composition and change.Particular reference will be made to the input of data to global models from the use of satellite, airborne and ground-based instrumentation. Contributions are invited for consideration by the Organising Committee covering topics within the area of chemistry, dynamics and modelling in the lower and upper atmosphere. Abstracts of about 300 words should be submitted by 31 May 1994 to: Professor I. W.M. Smith OR Dr R. J. Sodeau School of Chemistry School of Chemical Sciences University of Birmingham University of East Angliu Edgbaston, BirminRham Norwich B15 2TT. UK NR4 7TJ, UK Full papers for publication in the Discussion volume will be required by December 1994.THE ROYAL SOCIETY OF CHEMISTRY, FARADAY DIVISION, GENERAL DISCUSSION 101 Gels Paris, France, 6-8 September 1995 Organising Committee: Dr J. W. Goodwin (Chairman) Dr R. Audebert Dr R. Buscall Professor M. Djabourov Dr A. M. Howe Professor J. Livage Professor J. Lyklema Professor S. B. Ross-Murphy During the last few years there has been an increase in both theoretical and experimental work on gels as new techniques have been applied to a wide range of gelling systems. Typical of these are gels formed from polymers by both physical and chemical interactions as well as gels formed by inorganic and surfactant systems. The meeting will deal with the structure and dynamics of gels with the latter heading covering both swelling and rheological behaviour. Mixed systems such as polymer/surfactant and polymer/particle gels will also be discussed. The Discussion will bring together experimentalists and theoreticians interested in different types of gelling systems and encourage them to interact and assess the current scene and provide a benchmark for future developments. Contributions are invited for consideration by the Organising Committee. Titles and abstracts of about 300 words should be submitted by 30 September 1994 to: Dr J. W. Goodwin, School of Chemistry, University of Bristol, Cuntock’s Close, Bristol, BS8 1 TS. UK Full papers for publication in the Faraday General Discussion 101 volume will be required by May 1995. vi

ISSN:0956-5000    

DOI:10.1039/FT99490BP097

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(10), 1365-1371 Rotational Spectrum of the Gas-phase Dimer OC-.BrCI Susana Blanco,? A. C. Legon* and Joanna C. Thorn Department of Chemistry, University of Exeter, Stocker Road, Exeter, UK EX4 400 ~~~ The ground-state rotational spectra of the six isotopomers '601 2C. . .*' 16012C. . -79Br37CI,'601 3C. . .79Br35CI and 1601 2C* .79Br35CI, '601 e8' 2C* -*'Br37CI, 601 3C* Br35CI of a dimer formed by carbon monoxide and bromine monochloride have been observed by pulsed-nozzle, Fourier-transform microwave spectroscopy. The rotational constant B, , the centrifugal distortion constant D, and the halogen nuclear quadru- pole coupling constants x,,(X), where X = CI or Br, have been interpreted to give the properties of the weakly bound complex.It was found that the molecule is linear and has the arrangement OC. .BrCI, with the Br atom forming the intermolecular bond to carbon. Values of the rs type for the distances r(C- . .Br) and r(BrCI) have been derived and the limitations imposed on these by the small coordinatea,, are considered. The effect of bond shrinkage resulting from isotopic substitution at Br indicates that the rs value of a,, is underestimated and it is concluded that there is probably only a small increase in r(BrCI) on formation of the complex. An interpreta-tion of the x,,(X) values indicates that the electric charge distribution in BrCl is also only slightly perturbed. i.e. with the axis of the Series 9 (General Valve 1. Introduction de~ign,~ Complexes B.* -X, and B. ..XY formed by halogen, X, , or Corporation) solenoid valve perpendicular to the axis of the interhalogen, XY, molecules with electron donor molecules, Fabry-Perot cavity. The gas mixture, held at a stagnation pressure of ca. 3 atmB, are usually referred to as being of the charge-transfer in a stainless-steel reservoir, was pulsed from the solenoid type.' In these complexes, the halogen molecule is generally valve at a rate of ca. 2 Hz and had the following approximate accepted to function as an electron acceptor. Recently, we composition : 2% chlorine (Aldrich), 2% bromine (Aldrich), have been conducting investigation^^-^ of the rotational 4% carbon monoxide (Argo International) and the remainder spectra of a series of molecules B. sC1, with the aim not only argon.The standard equilibrium constant' for the reaction of characterising the isolated complexes but also of estab- Br, + C1, = 2BrCl at T = 300 K is sufficiently large (Ke xlishing how molecular properties derived from these spectra 5) that the heteronuclear species is the predominant halogen vary as B is systematically varied. Among molecular proper- in the gas mixture before expansion. ties that are available from the rotational spectrum of a The spectrum attributed to the four abundant isotopomers complex B. . C1, are the molecular symmetry, the intermolec- of (CO, BrC1) exhibited a rich nuclear quadrupole hyperfine ular distance r(B. -Cl), the intermolecular stretching force structure arising from the presence of a pair of I = 3 nuclei in constant k,, and the electric charge redistribution that each species.Individual hyperfine components had a fullattends formation of the complex. The main conclusions from width at half maximum of ca. 16 kHz (see Fig. 1) and their the axially symmetric B...Cl, (B = HCN, PH, and NH,) so frequencies were measured with an estimated precision of 2far are that any electric charge redistribution on kHz. The accuracy of frequency calibration of the spectrom- forming B. * C1, is small and that the strength of binding (as eter was checked to be better than 0.5 kHz at 12 GHz by measured by k,) decreases in the order NH, > HCN > PH, . measuring the J = 1 t0 transition of '6012C32S, which has The consensus of evidence' from a wide variety of investi- a frequency of 12162.979ql) MHz according to the accurate gations of B...XY and B...X, in various phases is that the spectroscopic constants B,, D, and H, published by Dub- relative electron acceptor strengths of C1, ,Br, and BrCl with rulle et aL9respect to a given donor B lie in the order BrCl > Br, > C1, .The isotopomers (l3CI6O, 79Br35C1) and (13Ct60,We have accordingly begun a series of investigations of s1Br35C1) were observed by using a sample of carbon mon- Be * SBrCl that parallel those of Be. C1, with the aim of com- oxide containing 99 at.% of 13C supplied by Amersham paring the properties of C1, and BrCl as electron acceptors in International. The limited amount of sample available and isolated gas-phase complexes.We report here the ground- the relatively small natural abundance of 7Cl precluded state rotational spectra of six isotopomers of a weakly bound measurement of transitions of (l3CI6O, 79Br37C1) and complex of carbon monoxide and bromine monochloride, as (' ,C160, ''Br37C1).observed by the pulsed-nozzle, Fourier-transform microwave technique. To our knowledge, this is the first study of a species B...BrCl in the gas phase. Analysis of the spectro- 3. Results scopic constants thereby determined allows a detailed charac- 3.1 Spectroscopic Constantsterisation of this species in isolation. Each isotopomer of (CO, BrCl) investigated exhibited a rota- 2. Experimental tional spectrum characteristic of a linear molecule. Each J + 1 tJ transition was split into a large number of hyper- The ground-state rotational spectra of the six isotopomers fine components as a result of the nuclear quadrupole coup- (l2Cl60,79Br35C1), (l2Ct60, "Br3'C1), ("Ct60, 79Br37C1), ling involving the I = 3 nuclei of the two halogen atoms.An (l2Cl60, 81Br37C1), (13C160, 79Br35C1) and (13C160, additional complexity was conferred on the spectra because 81Br35C1) were observed by using a pulsed-nozzle, Fourier- of the overlap of hyperfine structure for the isotopomers (CO, transform microwave spectr~rneter~.~ of the Balle-Flygare 79Br35C1) and (CO, "Br3'C1) in a given J + 1 eJ transition. A similar problem arose for the pair (CO, 79Br37C1) and (CO, t Permanent address: Departamento de Quimica-Fisica, Facultad 81Br37C1).In both cases the effect resulted from the very de Ciencias, Prado de la Magdalena, E-47005, Valladolid, Spain. small frequency shift that accompanies isotopic substitution L 9428.80 9428.60 9428.40 frequency/M Hz Fig. 1 Frequency-domain recording of two nuclear quadrupole hyperfine components belonging to J = 5 t4 transitions of the iso- topomers 16012C.. .79Br35C1 and 1601*C. * .81Br35C1.Adjacent dots are spaced by 3.90625 kHz and have been joined by straight lines. The Doppler doubling effect in each component is just discernible. The stick diagram indicates the calculated relative intensities but the observed relative intensities are affected by the exact tuning of the high-Q Fabry-Perot cavity. The assignment of quantum numbers to each component can be made by reference to Table 1.at Br. These observations establish that Br lies close to the dimer centre of mass. Observed frequencies for the J = 4 + 3, 5 t4 and 6 t5 transitions of the isotopomers (l2Cl60, 79Br35C1) and (l2Cl60, 81Br35C1) are given in Table 1. Fewer transitions of the less abundant species (l2Cl60, 79Br37C1) and (l2Cl60, 81Br37C1) were measured, as shown in Table 2. The observed frequencies of (13C160, 79Br35C1) and (13C160, 81Br35C1) are set out in Table 3. Some difficulty in the mea- surement of the J = 5 t4 transitions for the 13C iso- topomers was caused by overlap of these transitions with the J = 1t0 transition of the monomer species 79Br35C1, fre- quencies for which are in ref.10. For each isotopomer investigated, observed hyperfine fre- quencies were fitted in an iterative non-linear least-squares analysis to give the rotational constant B,, the centrifugal distortion constant D,, the halogen nuclear quadrupole coupling constants Xuu(Br) and xuu(C1), and the bromine spin- rotation coupling constant Mbb(Br). The form of the Hamilto- nian operator used was = HR + HQ(Br) + HQ(Cl) + HSR(Br) (1) The term HR is the familiar energy operator for a semi-rigid linear molecule. HR = B, J2 -D,J4 (2) while HQ(x,and HSR(X)describe the energy of interaction of the halogen nuclear electric quadrupole and magnetic dipole moments Q(X)and p(X) = gxpNIxwith the electric field gra- J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 dient V(X) and the magnetic field strength H,, respectively, at the nucleus X. The operators have the forms HQ(x)= -do@): v(x) (3) and HSR(X)= -1, * M(X)* J (4) where M(X)is the halogen spin-rotation coupling tensor. The matrix of H was set up in the coupled basis ZBr + Icl = I, I + J = F and diagonalized in blocks of F. The elements of HQ(,) and HsRin the coupled basis have been given in a con- venient form by Keenan et a/." and Read and Flygare,12 respectively. The residuals Av = vObs-vCalc from the final cycle of the least-squares fits are included, as appropriate, in Tables 1, 2 and 3 while the resulting set of spectroscopic constants is shown in Table 4. For a linear molecule, the only determin- able components of the hyperfine coupling tensors are x,,(X) = -eQVJX) and Mbb(X), where a is the molecular symmetry axis.It was found that hfbb(C1) was too small to be determined and hence this constant was set to zero in the fitting procedure. For the isotopomers l60l2C.* .79Br37C1 and 16012C.--81Br37Cl, hfbb(Br) was fixed at the value for the corresponding 35Cl species. We note from Table 4 that the standard deviation, ui, of the fit for each isotopomer is approximately the estimated accuracy of frequency measure- ment, confirming the appropriateness of the Hamiltonian used to fit the spectrum. Where ambiguity in assignment existed because of overlap of components, lines were omitted from the fit, as indicated in the footnotes to Tables 1 and 3.3.2 Geometry of (CO,BrCI) The nature of the ground-state rotational spectrum indicates that (CO, BrCl) is a linear molecule. The changes in the rota- tional constants on isotopic substitution establish unam-biguously that the nuclei lie in the order OC.e-BrC1. The analysis of the rotational constants to give the internuclear distances is based on a model, illustrated in Fig. 2, in which the CO and BrCl execute the angular oscillations 8 and 4, respectively, about the monomer mass centres. If the isotopomer l60l2C- .79Br35C1 is taken as the parent species, it is possible to determine coordinates a,, aBr and a,, of the rs type with the aid of the model shown in Fig. 2. It has been shown13 that the coordinate a, of an atom i in a weakly bound linear complex, like OC.* .BrCl, is given by (a?> = a)/2&) (5)(Azb/ps)-(Azr(sin2 where a is the oscillation angle of the subunit in which the isotopic substitution is made (a = 8 or 4, as defined in Fig. 2), AZb and A12 are the accompanying changes in the moment of inertia of the dimer and the monomer for isotopic substitut- ion at i and ps = M Arn/(M + Am). Eqn. (5) was derived under the assumption that the atoms in the linear molecule each execute a circular motion in a plane perpendicular to the line between the two subunit mass centres and hence is an C Br Fig. 2 Model used for the discussion of the OC...BrCl geometry. The OC and BrCI subunits execute the angular oscillations 8 and 4, respectively. Each atom describes a circle in the plane perpendicular to the line r,, between the mass centres.J. CHEM.SOC. FARADAY TRANS., 1994,VOL. 90 Table 1 Observed and calculated transition frequencies of 16012C. ."Br3'C1 and l60l2C. * .81Br35C1 16012C.. .79Br35C1 16012C.. .81Br35Cl J' 4-J" I' F 4-I" F" Vob$MHz AvlkHf VobJMHz AvlkHz" 44-3 0 44-0 3 753 7.8 104 -0.5 7537.9842 1.6 154-14 7538.5400 0.8 7538.8253 1.2 2 54-24 7539.3738 -0.3 7539.5359 -1.0 144-13 7540.3501 2.9 7540.53 16 -4.6b 2 64-2 5 7540.3708 -1.4 7540.53 16 1.3 3 74-3 6 7541.1688 -0.2 7541.3 18 1 -0.1 3 64-3 5 7541.6940 0.7 7541.9168 0.9 3 44-3 3 7560.8898 0.1 7557.4958 0.3 3 54-3 4 7565.9292 1.o 7562.4233 0.4 2 24-2 1 7565.99 16 -0.6 --2 44-2 3 7567.7196 0.2 7563.8749 0.0 54-4 2 54-3 5 9225.2461 -0.9 9259.2776 -0.2 3 64-3 6 9369.6511 0.4 9378.0686 0.4 2 44-2 4 9423.3129 -0.3 9420.7100 0.3 3 54-3 5 9425.5865 -2.6 9423.3846 -2.2 0 54-04 9427.1889 0.1 9426.8683 0.4 164-1 5 9427.9911 1.2 9427.7499 -5.3' 2 64-2 5 9428.1045 -0.2 9427.7749 4.1b 154-14 9428.6784 -0.9 9428.3629 -0.5 144-1 3 9428.8755 -6.3' 9428.5621 4.5' 2 74-2 6 9428.8983 2.9' 9428.5621 -1.9 3 84-3 7 9429.3550 1.3 9429.01 49 -1.5 3 74-3 6 9429.65 18 0.2 9429.3550 -0.7 3 34-3 2 9440.4891 0.2 9437.8719 -0.7 3 44-3 3 9441.0037 1.6 9438.5668 0.3 3 5+3 4 9441.5428 -0.6 9439.0810 0.5 324-3 1 9442.1992 -1.3 9439.5467 -0.6 2 44-2 3 9442.8873 -0.2 9440.2660 1.1 2 34-2 2 9443.2340 0.1 9440.7325 -1.3 3 64-3 5 9444.0222 0.2 9441.4925 0.6 2 54-24 9444.7604 -0.6 9442.1338 1.3 3 44-3 4 9446.8422 -0.2 9444.2570 -1.2 3 34-2 3 945 1.3630 0.9 9448.7074 -0.1 144-04 9448.2945 -0.8 9448.0110 -1.0 3 24-3 2 9462.3528 -1.5 9459.7945 0.3 3 74-3 7 11266.5435 2.3 64-5 0 64-0 5 11315.3252 -0.3 11314.6281 0.3 2 74-2 6 11315.8977 -0.2 11315.1885 -0.9 174-1 6 11316.0297 0.7 11315.4204 0.8 164-1 5 --11315.5615 -4.v 154-14 --11315.5944 4.2' 2 84-2 7 11316.5194 1.2 11315.8137 1.o 3 84-3 7 11316.9907 -0.1 11316.3071 0.2 344-3 3 11324.41 57 2.1 11322.2046 -0.1 3 54-3 4 11324.5904 -1.4 11322.4808 1.2 3 64-3 5 11325.1665 0.3 3 34-3 2 11325.2638 -1.8 2 44-2 3 11325.7868 -0.3 11323.6324 -2.1 2 54-2 4 11325.9607 -1.7 11323.7470 1.1 3 74-3 6 11326.5420 -0.1 11324.3347 2.2 2 64-2 5 11326.9258 0.9 11324.7078 -3.0 394-3 8 11316.8059 -0.1 11316.0983 1.9 a Av = vobs -vCalc.' Omitted from fit, see text.attempt to correct for the contribution of the intermolecular (7)bending modes. andWhen the rotational constants of the dimers from Table 4 and the mon~mers'~*'~ from Table 5 are used with the angles a,, -aBr= r(BrC1)cos +,, (8) O,, = 10" and +,, = 5" (these choices will be justified later), where r and r' are the distances of the C and Br atoms from the resulting coordinates ai are those given in column I of their respective subunit mass centres and are available from Table 6. Since isotopic substitution of oxygen was not carried the bond distances in Table 5 on the assumption ofout, the value for a, was obtained by using the other a, in the unchanged monomer geometries.A comparison of r(C0)andfirst moment equation Cimiai = 0. The distances rij are also r(BrC1) in Table 6 with the corresponding quantities for the given in Table 6 and have been calculated from the expres- monomers (Table 5) suggests that the CO bond has short- sions ened by ca. 0.1 8, while the BrCl bond has lengthened by 0.036 8, on formation of the complex. However, such a con- r(C. .Br) = I a, 1 -r( 1 -cos O,,) clusion needs to be treated cautiously, as the following argu- ments illustrate. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Observed and calculated transition frequencies of 16012C.. .79Br37C1and 16012C.. .81Br37C1 44-3 2 6+2 5 7392.5737 0.5 7393.0555 -0.9 3 763 6 7393.2104 -1.5 739 3.6909 0.6 3 64-3 5 7393.5688 1.o 7394.1267 0.3 54-4 0 54-04 9242.7778 0.8 9242.8829 1.8 164-1 5 9243.3541 3.0 2 64-2 5 9243.5036 -4.9 9243.5844 -0.3 154-14 9243.9467 -1.5 9244.0321 -3.7 144-1 3 9244.1243 4.0 2 74-2 6 9244.1243 -3.1 9244.2028 -2.6 3 84-3 7 9244.4967 1.5 9244.5 72 1 2.3 3 74-3 6 9244.6975 0.3 9244.8 197 2.4 3 34-3 2 9256.3059 0.7 9254.0963 0.4 3 44-3 3 9256.5666 -1.3 9254.5409 -0.2 3 54-3 4 9257.0208 -3.3 9254.9839 2.4 324-3 1 9257.6820 5.0 9255.4414 -1.0 2 44-2 3 9258.2067 -1.4 2 34-2 2 9258.3797 1.6 9256.3059 -2.9 3 64-3 5 9259.0093 -0.2 9256.9390 0.8 2 54-24 9259.6792 -1.2 9257.4638 0.7 " Av = 'obs -vcalc* -.79Br35C1and 16013C*.Table 3 Observed and calculated transition frequencies of 16013C. .'lBr3'C1 ~~ 54-4 2 64-2 5 9293.2343 5.5' 154-1 8 9293.7997 -3.1 9293.261 3 -1.1 2 74-2 6 9294.02 12 2.1 929 3.4640 0.6 3 84-3 7 9294.4755 -1.5 9293.9155 -0.2 3 74-3 6 9294.77 17 0.7 9294.2543 0.8 3 34-3 4 9302.7670 -0.8 3 44-3 6 9306.1084 -1.9 9303.4569 1.1 3 54-3 8 9306.6526 2.0 9303.9700 0.7 3 24-3 2 9304.4396 -3.9 2 44-2 6 -9305.1610 1.4 3 64-3 5 9309.1280 -0.7 9306.3830 1.5 2 54-24 9309.8767 2.4 64-5 0 64-0 5 1 1 153.4846 1 1.7' 1 1 152.5092 2.3 2 74-2 6 1 1 153.069 1 -0.3 174-1 6 1 1 154.1809 5.1' 11 153.2986 -0.7 164-1 5 1 1 154.4233 11.8 11 153.4430 -2.4 154-14 --11 153.4686 -0.6 2 84-2 7 1 1 154.6616 -6.0 1 1 153.6946 1.8 3 84-3 7 1 1 155.1368 -0.2 1 1 154.1865 0.3 3 4+3 3 11 162.5762 20.8' 3 54-3 4 --11 160.3567 3.3 3 64-3 5 11 163.3044 0.2 1 1 160.9026 0.4 2 44-2 3 --11 161.5052 -3.4 2 54-24 11164.1013 -1.6 11161.6233 0.6 3 74-3 6 1 1 164.6779 -1.8 11 162.2060 -0.3 2 64-2 5 1 1 165.0656 0.4 11 162.5762 -11.8' 3 94-3 8 11 154.9523 -2.8 1 1 153.9756 -0.9 " Av = vobs -vCalc.'Omitted from fit, see text.Table 4 Ground-state spectroscopic constants of six isotopomers of OC. * .BrCl 12~160.. .79~~35~1 943.34024 (6) 0.61 1 (1) 875.835 (5) -97.615 (3) -3.1 (1) 1.1 12C160.. .81Br35C1 943.24082 (7) 0.611 (1) 731.700 (5) -97.619 (3) -3.6 (2) 1.2 12~160.. .79~~37C1 924.8632 (4) 0.581 (7) 876.13 (5) -76.96 (3) -3.1' 2.7 12~160.. .81~~37~1 924.8057 (3) 0.591 (5) 732.05 (4) -76.98 (2) -3.6' 2.1 13c160.. .79Br35C1 929.8514 (4) 0.599 (5) 876.0 (2) -97.54 (7) -4.4 (1.1) 4.5 13C160...81Br35C1 929.7300 (2) 0.602 (2) 731.79 (8) -97.63 (3) -3.5 (5) 1.9 a Standard deviation of the fit. 'Fixed at value for 3sCl species. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 5 Properties of the monomers CO"and BrClb ~ ~~ monomer B,/MHz Xo(Br)/MHz xo(CI)/MHz rJA 12C'60 57635.9687 (26) --1.129 49' l3CI6O 55101.0205 (122) --12C180 54891.4239 (122) --79Br35C1 4559.3827 (1) 875.309 (1) -102.450 (2) 2.13738' 'lBr3'C1 4524.8598 (1) 731.223 (1) -102.451 (2) 79Br37C1 4388.9109 (1) 875.304 (1) -80.740 (2) 81Br37CI 4354.3855 (1) 731.219 (1) -80.740 (2) Ref. 14. Ref. 10. Calculated from the rotational constants. Laurie and Herschbach" have shown, for example, that isotopic substitution of 13C in CO, causes a shortening of 5 x 8, in the average CO bond length.It is known that this type of shrinkage leads to a serious underestimate of principal-axis coordinates of atoms that lie close to the centre of mass, as for example when "N substitution is used to obtain the N coordinates in nitric acid. By assuming that the bonds involving N shrink by ca. 8 x lo-' A on "N substi-tution, Cox and Riveros16 demonstrated that a sensible value for aNcould be obtained. It is of interest to recalculate aBrin OC--.BrCl by assuming that the C-..Br and BrCl bonds shrink by 5 x lo-' 8, and 1 x 8,, respectively, when 'lBr is substituted for 79Br, the second of which is suggested by the change of this magnitude that occurs in ro for free BrCl on isotopic substitution. lo The contribution AZFrr that such a shrinkage makes to the change AZb in 1, on 81Br sub- stitution can be calculated as follows.First, using unchanged monomer geometries (Table 5), the distance between the CO and BrCl subunits is adjusted to find the value of r(C...Br) that reproduces the observed B, for 16012C-. -79Br35C1.This model is then used to calculate 1, for 6012C. ." Br3'C1. Next, the distances r(C...Br) and r(BrCI) are allowed to shrink by 5 x lo-' 8, and 1 x A,respectively, and the modified value Zb for l60l2C. .'lBr3'C1 recalculated. The correction AZYr = 1; -1, is then added to If', for l60l2C.* .81Br35C1 and the resulting value used to recalcu- late aBrfrom eqn. (5). The result is shown in column I1 of Table 6, in which are also included corrected values of a,, and a, calculated in a similar manner shrinkage^'^.'^ of 5 x A,1 x 8, and Sbxy,O"? in r(CO), r(C.* .Br) and r(BrCl), respectively. We note that aBr is substantially increased from 0.164 to 0.183 8, while the cor- rections to a,, and a, are very small. Also included in column I1 of Table 6 are a,, obtained from the first moment condi- tion and, thence, r(C0) and r(BrC1) recalculated using the corrected a coordinates in eqn. (7) and (8) with O,, = lo" and 4av= 5". We find that these quantities now differ from their Table 6 r,-type coordinates in l60l2C...79Br3'C1 coordinate aJA or bond length r,/A I" IIb 111' 2.3295 2.3304 2.3304a,, a,, 0.1644 0.1810 0.1838 a, -2.7918 -2.7936 -2.7936 a0 -3.8095 -3.8922 -3.9050 r(C0) 1.0334 1.1 156 1.1295 r(C.. .Br) 2.9439 2.9623 2.9650 r(BrC1) 2.1734 2.1576 2.1548 ~~~~~~~ r,-type coordinates and bond lengths calculated using eqn. (5)-(8). The value of a, is estimated from Z, mi a, = 0. r,-type coordinates calculated from eqn. (5) after correction for bond shrinkage on iso- topic substitution (see text for discussion). a, is estimated from Zirniai= 0. The bond lengths were calculated using eqn. (6)-(8). 'In this set of coordinates, acl and a, were calculated as in column 11. a, was estimated using la,( -I a,/ = r(C0)cos O,,, where 8 = 10" and r(C0) is assumed unchanged from free CO. aBrwas then obtained from the first moment condition. free molecule values by only 0.0139 and 0.0202 A,respec-tively. We conclude that the changes in r(C0) and r(BrC1) implied by the uncorrected r,-type coordinates ai are prob- ably too large since the application of only small shrinkages of the expected magnitude is suficient to bring these quan- tities near to the free monomer values.That the correction to aBris of reasonable magnitude is suggested by the following argument based on an alternative estimation of aBrvia the first moment equation Xi mi a, = 0. If the very strong bond in CO is taken as unchanged in length on formation of OC. * eBrC1, then the coordinate a, of oxygen can be accurately estimated from the corrected a, of column 11, Table 6 using eqn. (7). When O,, is again assumed as lo", the result for the coordinate a, is that shown in column I11 of Table 6.This value of a, taken with the corrected a, and a,, coordinates in the first moment equation allows the value of aBrshown in column 111 of Table 6 to be obtained. The result is in excellent agreement with that estimated using the shrinkage correction. The set of bond lengths determined in this alternative approach is shown in column I11 of Table 6. A similar approach to that described above can be applied to data existing in the literature' for OC.. C1,. In this case, all four atoms were isotopically substituted. Application of eqn. (5) without shrinkage corrections leads to a, = -3.8492 8,, a, = -2.7375 8,, acli= 0.3473 A and a,,, = 2.3829 8, (i = inner, o = outer). The small aclicoordinate also can be estimated alternatively, by the first moment condition, to be 0.3857 8,.When a shrinkage correction of 1 x lop4 8, is assumed for r(C-.C1) and 5 x lo-' 8, for r(CIC1) on 37Cl substitution at Cli, the above procedure leads to a corrected value of the rs coordinate acli= 0.396 A,in excellent agree- ment with that obtained using the conditions Cimiai = 0. We then find a,,o -acli = 1.9868 A,which implies r(CICI) = 1.994 8, via eqn. (8) if 4,, z 5 and therefore negligible change in the Cl-Cl bond length on dimer formation. In the same manner, we note that when these shrinkage corrections are applied to determine acli in the much more strongly bound dimer H,N.. CI,, the corrected value of r(ClC1) = 2.0096 8, is obtained, corresponding to an increase of 0.018 A in the CI, bond length on dimer formation." When, on the other hand, uncorrected r,-type coordinates a,, are employed, the distance r(ClC1) appears to shorten on for- mation of H,N..C1,. This result seems unlikely on chemical grounds but can be understood when it is recognised that aCli is too small in magnitude when uncorrected and that Cli lies on the opposite side of the dimer centre of mass from C1, in H,N. * C1,. We also note that the order of the intermolecular binding strength, as measured by k, (see Section 3.4), is OC. .Cl, < OC.. .BrCI < H,N.. C1, (the k, values are 3.6, 6.3 and 12.7 N m-', respecti~eIy~*~*'~) but that the order of lengthening of the halogen bond on dimer formation is OC.. C1, > OC. . aBrC1 > H,N. -C1, when uncorrected coordinates a, are used (see above). The application of the shrinkage corrections as described earlier, however, leads to lengthenings of ca. 0.003, 0.017 and 0.018 8, along this series. This revised order, while dependent on some (reasonable) assumptions, does appear more acceptable chemically. Another method of obtaining the distance r(C.--Br) is based on the model shown in Fig. 2 but assumes unchanged monomer geometries and attempts to account for the contri- bution of the intermolecular bonding modes to the zero-point moments of inertia through the 1, z (Ibb) = p(r,?m) + ilto(l f (cos2 8)) + llBrCl7 b (l + (cos2 6)) (9) In eqn. (9), ZFo and IFrc1are the ground-state moments of inertia of the monomers, available from the B, values in 1370 Table 5 via B, = h/87r21,, r,, is the distance between the monomer mass centres, and p = mC0mBrC1/(mCo+ mBrCI).When 8,, = COS-~(COS~8)1/2= 10" and 4," = COS-~(COS~ t9)1/2 = 5" are used in eqn. (9), the values of (r~,,J1" given in Table 7 result. A value of r(C. . .Br) is then available via r(C. * .Br) = (r&)1/2 -r' -r (10) where r and r' are the distances of the C atom of CO and the Br atom of BrCl from their respective monomer mass centres. In fact, (r:m)1/2 and r(C...Br) so estimated are relatively insensitive to the assumed values of 8 and 4, as indicated by the quoted uncertainties which arise from assumed errors of +2" in 4 and +5" in 8. We note that r(C...Br) so deter-mined is in only reasonable agreement with that obtained from corrected r,-type coordinates (column I11 of Table 6).3.3 Interpretationof the Halogen Hyperfine Coupling Constants The halogen nuclear quadrupole coupling constants x,,(X), where X = Br and C1, are only slightly changed from those" of free BrC1, which are collected in Table 5. If electric charge redistribution between CO and BrCl is ignored, three factors act to change the x,,(X) on formation of the complex. First, there is the additional electric field gradient (efg) at the nucleus X that arises from the presence of the CO electric charge distribution. The efg at X is, moreover, enhanced by the response of the BrCl electronic charge distribution to the field and its derivatives arising from CO.This enhancement is likely to be more serious for Br than C1. The result is that, in the equilibrium linear geometry, the magnitude of x,,(Br) should be increased while that of x,,(Cl) should be decreased. The second factor is the angular oscillation of the BrCl subunit of the type depicted in Fig. 2. If the effects of the electric charge distribution of CO were ignored, the angular oscillation would act to reduce both free molecule values xo(X) by the same factor, namely xaa(X) = +x0(X)(3 COS' 4 -1) (11) The third effect is the change in xo(X) that results from lengthening of the BrCl bond on formation of OC...BrCl. Even though the extent of any lengthening is uncertain, the effect can be ignored, since (ax,,Jdr)Gr is only 8.5 MHz and -2.7 MHz for 79Br and 35Cl, when 6r is as large as 0.1 A.1o We note from Tables 4 and 5 that x,,(~'B~) is almost unchanged on formation of the complex.This arises because the reduction in magnitude through the BrCl angular oscil- lation is presumably equal to the increase brought about by the presence of the CO electric charge distribution. The latter effect is less serious for the C1 nucleus because of its greater distance from the CO subunit but in this case acts in the same direction as the angular oscillation, i.e. both effects tend to reduce the magnitude of ~,,(~~cl). inHence use of ~,,(~~cl)The investigation of the ground-state rotational spectra of six J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 eqn. (11) leads to the upper limit of = COS-~(COS~ 4)'12 = 10.2". The value of 4avafter correction for the efg due to CO is likely to be considerably smaller and we have chosen 5 &-2". If 4,, = 5" is assumed, we can make a very rough estimate of x,,(~~B~) in the equilibrium linear arrange- and ~,,(~~cl) ment by using eqn. (11) and the xo(X) of Table 5. The resulting values are ca. 886 and -99 MHz, respectively. It is of interest that according to the Townes-Dailey model of BrCl, the coupling constants x~(~~B~)and x,(~~C~) correspond2' to a ca. 10% contribution of the valence bond structure Br+Cl- to the description of the free molecule BrCl. The estimated equilibrium values of x,,('~B~) and xa,(35Cl) in OC--.BrCl suggest, on the same basis, only an additional ca.3% contribution of the ionic structure on for- mation of the complex, even if all the change in efg at the halogen atoms were attributed to this effect. On these grounds, it seems unlikely that charge transfer, either within BrCl or from CO to BrCl, is of appreciable magnitude when OC. .BrCl is formed. The projection formula analogous to eqn. (1 1) for the spin- rotation coupling constant kfbb(Br) is22 kfbb(Br)= (B0/2B3(1 4-COS2 +)M,(Br) (12) where Br is the monomer rotational constant and M,(Br) is the spin-rotation coupling constant" of BrCl. Eqn. (12) leads to the predictions kfbb(79Br) = -3.3 kHz and kfbb(81Br) = --3.5 kHz for 16012C. .79Br35C1 and 16012C. .81Br35C1, in excellent agreement with the observed values of -3.1(1) and -3.6( 1) kHz, respectively.Finally, in the absence of nuclear hyperfine constants associated with the CO subunit it is not possible to estimate O,,. We have chosen the value 10 & 5" in view of the smaller mass than BrCl and the e~timate'~ of 8,, = 11.7" for CO in the somewhat more weakly bound OC. * .HCN. 3.4 Intermolecular Stretching Force Constant k, It is customary to use the quadratic force constant k, associ-ated with the intermolecular bond in complexes such as OC. * eBrC1 to estimate the strength of the interaction. In the quadratic approximation and when the subunits are assumed rigid, Millen23 showed that the centrifugal distortion con-stant D,is conveniently related to k, via the expression k, = (16n2B;p/D,)(1 -B,/B;O -B,/BfrC1) (13) where p is as defined earlier and Bzo and Brc' are ground- state rotational constants of the monomers.The values of k, calculated for the various isotopomers using eqn. (1 3) are included in Table 7. The appropriate B, values are in Tables 4 and 5. 4. Conclusion Table 7 Values of (r~m)1~2.r(C. ..Br) and k, for various iso- topomers of OC-. .BrCI isotopomer (rz,,,)1/2/8ia r(C. * .Br)/dib k,fN m-' ' I6O1*C.. .79Br35Cl 4.306 (2) 3.004 (2) 6.29 (1)' 6O"C. . . Br35CI 4.295 (2) 3.004 (2) 6.30 (1)l60l2C-. .79Br37CI 4.332 (2) 3.005 (2) 6.22 (6) 16012C...81Br37C1 4.320 (2) 3.005 (2) 6.18 (6)16013C.. .79Br35C1 4.283 (2) 3.004 (2) 6.34 (5)I6Ol3C.. .81Br35C1 4.272 (2) 3.004 (2) 6.31 (2) Calculated using 8," = 10 (5)O and 4av= 5 (2)" in eqn.(9). Cal-culated using (r:m)1/2 in eqn. (10) with r and r' estimated from the rs geometries of the monomers from Table 5. Calculated using D, in eqn. (1 3). isotopomers of the weakly bound complex OC. * .BrCl reported here leads to the conclusion that the species is linear with the nuclei in the order indicated. The positive end of the BrCl electric dipole moment therefore appears to align with the negative end (at carbon) of the CO electric dipole moment. A comparison of k, = 3.6 N m-l for OC..C12 [recalculated from the D, value in ref. 17 and eqn. (13)] with k, = 6.3 N m-' for the analogue OC. -.BrCl indicates that the latter species is the more strongly bound. This conclusion is also consistent with simple electrostatic considerations, since in OC.-Cl, the interaction would be of the dipole- induced dipole or the dipole+quadrupole type while in OC.. .BrCl both components have a permanent electric dipole moment. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1371 A detailed consideration of the r,-type coordinates in 3 A. C. Legon and J. C. Thorn, J. Chem. SOC., Faraday Trans., OC..-BrCl and especially the effects of bond shrinkage on the small Br coordinate indicates that the increase in the BrCl bond length on formation of the dimer is smaller than sug- gested by the raw rs coordinates and that the increase is rela- tively small. A consideration of the changes in the nuclear 4 5 6 1993,89,4157. A. C. Legon, D.G. Lister and J. C. Thorn, J. Chem. SOC.,Chem. Commun., in the press. A. C. Legon, Annu. Rev. Phys. Chem., 1983,34,275. C. A. Rego, R. C. Batten and A. C. Legon, J. Chem. Phys., 1988, 89, 696. quadrupole coupling constants X,,(Br) and x,,(Cl) on forma- tion of OC. nBrC1 reveals that the extent of charge transfer either from CO to BrCl or from Br to Cl is rather small. This result is not inconsistent with the apparently small changes in monomer geometry and the two taken together indicate that a simple electrostatic model might well be appropriate to 7 8 9 10 11 T. J. Balle and W. H. Flygare, Rev. Sci. Instrum., 1981, 52, 33. T. Moeller, Inorganic Chemistry, Wiley, New York, 1952, p. 449. A. Dubrulle, J. Demaison, J. Burie and D. Boucher, Z. Natur-forsch., Teil A, 1981,35, 471.A. C. Legon and J. C. Thorn, Chem. Phys. Lett., 1993,215,554. M. R. Keenan, D. B. Wozniak and W. H. Flygare, J. Chem. Phys., 1981, 75, 631. describe the species OC. .BrCl. 12 13 W. G. Read and W. H. Flygare, J. Chem. Phys., 1982,76,2238. A. Haynes and A. C. Legon, J. Mol. Struct., 1988,189,153. We thank the SERC for support of this work through a research grant, the Ruth King Trust of the University of Exeter for a studentship in support of J.C.T. and Dr. Peter 14 15 F. J. Lovas and P. H. Krupenie, J. Phys. Chem. Ref: Data, 1974, 3, 245. V. W. Laurie and D. R. Herschbach, J. Chem. Phys., 1962, 37, 1687. Cox for helpful discussions about rs coordinates and the 16 A. P. Cox and J. M. Riveros, J. Chem. Phys., 1965,42,3106. effect of bond shrinkage. S.B. gratefully acknowledges a grant from the Vicerrectorado de Investigacion of the Universidad de Valladolid and the Direccion General de Investigacion Cientifica y Tkcnica (DGICYT) of the MEC. 17 18 19 W. Jager, Y. Xu and M. C. L. Gerry, J. Phys. Chem., 1993, 97, 3685. A. C. Legon, D. G. Lister and J. C. Thorn, unpublished observa- tions. G. T. Fraser, K. R. Leopold and W. Klemperer, J. Chem. Phys., 1984,80, 1423. 20 A. C. Legon and L. C. Willoughby, Chem. Phys., 1984,85,443. References 21 22 B. P. Dailey and C. H. Townes, J. Chem. Phys., 1955,23, 118. M. R. Keenan, T. K. Minton, A. C. Legon, T. J. Balle and W. H. 1 A. J. Downs and C. J. Adams, in Comprehensive Inorganic Chem- Flygare, Proc. Natl. Acad. Sci. USA, 1980,77, 5583. istry, ed. A. F. Trotman-Dickenson, Pergamon, New York, 23 D. J. Millen, Can. J. Chem., 1985, 63, 1477. 1973, vol. 11, ch. 26. A. C. Legon and H. E. Warner, J. Chem. Phys., 1993,98,3827.2 Paper 4/003201; Received 18th January, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949001365

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(10), 1373-1380 I373 Proton Transfer to the Fluorine Atom in Fluorobenzene Temperature and Pressure Dependence Rod S. Mason,* Alyn J. Parry? and Dafydd M. P. Milton Department of Chemistry, University College of Swansea, Singleton Park, Swansea, UK SA2 8PP Proton transfer from CH,' to the fluorine atom substituent of fluorobenzene has been studied as a function of temperature and pressure using a high-pressure ion source and collision-induced dissociation (CID) mass spec-trometry. The behaviour at low temperatures examined (<350 K) is similar to that expected of a 'locking' ion-dipole mechanism. However, the third-order pressure and overall T-' .46 temperature dependences are consistent with a mechanism involving the formation of an intermediate complex.This complex has been iso- lated from a benzene/fluorobenzene mixture in methane at ca. 200 K and was identified as having the C,H,F-H+-CH, structure, in a potential-energy well depth of 10-11 kJ mol-' according to theory. Although decomposition directly from the complex is very likely, under the conditions investigated thermal decomposition from the F-protonated isomer does not appear to be a significant factor in determining its population relative to the ring-protonated species. Proton transfer is the simplest class of truly chemical reac- tions, it is also one of most important, often being the initi- ation step in acid-catalysed reactions of a much more complex nature. Ultimately the site of protonation on a poly- atomic molecule is determined by thermodynamic factors, but during the initial encounter kinetic factors may intervene.Proton transfer to fluorobenzene is a case in point, which in the gas phase can be studied directly using tandem mass spectrometry to monitor the different protomeric species during the reaction. In general, gas-phase proton transfer to the halogeno- benzenes has been widely studied over the years.'-8 In the case of fluorobenzene Lau and Kebarle' originally showed that the thermodynamically favoured site of protonation is on the ring on the carbon-4 position, opposite the F atom. peaks. Very recently Schwarz and co-workers' ' have report- ed an underlying third component in the peak of the metasta- ble ions, which they have ascribed to the eviction of HF by 1,l elimination from the ring-protonated species, the broadest component being caused by the 1,2 elimination. We had pre- sented very similar data previ~usly,'~~'~ but believe that it is difficult to distinguish this third component from that due to direct HF ejection from the F-protonated species.However, direct proton attachment to the fluorine atom is seen only at relatively low reaction temperatures. It was proposed' that the negative temperature dependence of this reaction might have two origins: (1) a locked ionaipole com- ponent and (2) reaction uia formation of a proton-bound complex. A third possibility has been very recently suggested by Harrison and co-~orkers'~ who studied deuteron transfer Further studies on equilibrated systems f~llowed.~.~to a However, studies of the chemical ionisation spectra of these molecule^,^^^ particularly by Harrison and co-workers,6 indi- cated that the proton may attach directly to the halogen atom rather than the ring.Cacace and Speranza7 reached similar conclusions from their pulse radiolysis experiments. When the reaction is allowed to equilibrate the site of protonation is decided purely by thermodynamic factors, the proton ending up on the part of the molecule which has the highest proton affinity (PA). However, the site on the mol- ecule to which the proton becomes attached in the first instance will be determined by both kinetic and energetic factors. This was observed' in our preliminary experiments on the reaction between CH,' and C6H,F, using tandem mass spectrometry to show that at low temperatures the proton, attracted by the dipole, is observed to transfer directly onto the F atom, even though it was shown that its PA is 200 kJ mol-' less than the aromatic ring, the thermo- dynamically preferred site.This was confirmed using deuterium labelling experi- ment~.~.''It was also shown that the ejection of HF due to direct bond cleavage during high-energy CID of the proto- nated molecule is accompanied by low kinetic energy release. A second higher energy process is a the loss of HF accompa- nied by high kinetic energy release. In mass-analysed ion kinetic energy spectroscopy (MIKES) these are easily distin- guished by the widths and shapes of the two component Present address: Unilever Research, Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, UK L63 3JW.series of mono-, di- and tri-fluorinated benzene mol- ecules, using low-energy CID mass spectrometry to dis-tinguish the isomers. They observed very similar behaviour but believe that at the higher temperatures the F-protonated species is depleted, relative to the ring, because of its thermal instability. In this paper we report more fully on the temperature- dependence work previously presented in only a preliminary or informal manner. To resolve the mechanism more fully the reaction kinetics have been studied as a function of total and partial pressures.Further experimental, high-energy CID, and theoretical studies show the existence of the proposed intermediate complex. Experimental Some experiments were carried out using a reverse-geometry MIKES instrument described previously.' Most were carried out using a specially constructed 'high-pressure' ion-molecule reaction source attached to an updated and modi- fied MS9 mass spectrometer, described in detail elsewhere.', The gaseous reaction mixture was premixed in a heated gas storage bulb by injection of the liquid fluorobenzene directly into an atmosphere of methane. The gas was then allowed to flow continuously into the ion-molecule reactor, maintained at a constant pressure by an automatic pressure control valve and pressure transducer arrangement.The temperature could be varied from -100 to +400 "C. Reactant and product ions drifting through the ion exit aperture were accelerated to produce a 6 keV ion beam. The different structures of protonated halogenobenzene were then monitored by tandem mass spectrometry using the technique of B/E linked scanning.I6 N, gas was added to the first field- free region of the mass spectrometer until the primary ion beam passing through it was attenuated by ca. 30% [the pressure at the ion gauge in the analyser rising to (1.8-2.0) x Torr]. Daughter ions resulting from the CID were then detected by slowly scanning the electrostatic analyser voltage (E) and magnetic field (B)simultaneously, in direct proportion, and recording the resultant ion beam signal impinging on the single-channel electron multiplier detector using a single-ion counting and D/A amplifier system.The constant B/E scan law was achieved using purpose-built electronics by which the electrostatic analyser (ESA) reference voltage was derived from the magnet's Hall probe voltage. The 99% pure fluorobenzene and perdeuteriated fluoro- benzene (Aldrich) were used without further purification. Any impurities present did not show up in the chemical ionisation spectra and therefore did not interfere with the reactions under study. The methane (BOC) was of research grade (99.992%). Results and Discussion Identification of Isomeric Protonated Structures from the CID Spectra In methane as the ionising gas a number of protonating agents are formed in the reaction mixture, but mainly CH,+ and C,H5+.It was shown previously' that the proton affinity of the F atom is too low to be protonated by anything except CH, + . The reaction with fluorobenzene [reaction (l)] leads to protonation either on the ring or on the F atom. X X As already reported,8 the two different forms can be unam- biguously identified from their high-energy CID spectra. The F-protonated form gives rise to a sharp peak at mass (mH+ -20) due to the expulsion of HF from the molecule, where mH+ = 97 and represents the protonated molecular mass, whereas the ring protonated form gives rise to the loss of H, ,(mH+ -2). A fraction of the HF loss is due to a much higher energy process whereby collision-induced migration of the proton occurs from the ring over to the fluorine atom prior to HF expulsion.When this happens the HF is expelled with a significant amount of excess kinetic energy. This is easily identified in the MIKE spectrum as a low-intensity component, significantly broader than the narrow 'direct loss' peak, and relatively insensitive to changes in source con- ditions and is shown in Fig. 1. Whilst the broad component is caused by 1,2 elimination, there appears to be an underlying narrow component (but broader than that due to the direct bond cleavage), also due to a high-energy process, which Schwarz et ul." have ascribed to 1,l elimination oia the ips0 ring-protonated form. J. CHEM. SOC.FARADAY TRANS., 1994, VOL. 90 4625 4725 4825 3098 3174 ion beam energy/eV ion beam energy/e\l Fig. 1 MIKES/CID spectrum of HF peak lost from C6HsFH+ formed (a)in CH, (protonating agent CH,+) at 428 K; and (b) in H, (protonating agent H3+) at 273 K" not occur. However, the contribution from this source does not appear to be large and under CID conditions is not dis- tinguishable from the base of the narrow F-protonated com- ponent. As previously reported,8 the energy of H3+, the protonating agent, appears to be too high to protonate the F atom directly and efficiently. This is probably because the reaction would lead directly to dissociative proton transfer, but also if the energy is greater than the barrier height for ring/substituent intramolecular transfer, the proton can readily migrate to the ring.Schwarz et al." have calculated the barrier to be 263 kJ mol-' above the most stable ring- protonated form (see Fig. 2). The energy of CH,+,on the other hand, which readily protonates the F atom under the 400 1 263 1 i -CH 30H 2+ This component can be observed most readily in the MIKES Fig. 2 Potential-energy profile showing isomers of C,H,FH+, their metastable ion spectrum of C,H,F protonated in an H, CID products and the relative energies of CH,+, H,+, C,H,+, atmosphere (see Fig. 1) where direct HF loss probably does CH,OH,+ and the complex C,H,FHCH,+ J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 right conditions, is almost resonant with the F-protonated form.It is interesting that it is possible to observe metastable ions for these processes using methane as the protonating gas," because there is insufficient energy in the protonation reaction to overcome either the rearrangment barrier or the dissociation threshold. We too have observed" a small degree of metastable ion decomposition similar to that published by Schwarz et a!. When methanol, which has a proton affinity near-resonant with the ring, is used no meta-stable ions were observed. On the other hand, the H,-generated metastable ion spectrum was so intense as to be very little different from the CID spectrum. It is probable that the methane-generated metastable ion spectrum arises from collisional activation of the already internally energetic ions, to levels above the dissociation limit, in the accelerating region of the ion source as they pass through the escaping source gas.It can be assumed that the overall CID collision cross-sections for the two isomers are equal, since, apart from the proton, they are very similar. However, the differential cross-sections for H, and HF loss are unlikely to be equal, favour-ing the latter (by ca. 2 : 1 according to expectations of relative rates based on ion-dipole theory, see below) because of the large differences in energy required for the two processes (see Fig. 2). Although not the true value, changes in the corrected [mH+ -20/[mH+ -21 ratio must nevertheless be represen-tative of changes in the population ratio of fluorine to ring-protonated isomers.Contribution from '3CC,HSF+ to the 197 -HF1/[97 -H, I Ratio A potential problem is the contribution to the (m + 1)+ parent ion of the 13C isotope peak of the molecular ion, m+, which is 6% of the intensity of m+. Therefore under source pressure conditions when [m + 13 : [m] M 1, the contribu-tion to m/z = 97 (C6H5FH++ l3CCSHsF) from the I3C isotope is 6%. Comparative measurements from the CID spectra, under identical conditions when [m + 13 : [m] = 1, shows that the peaks due to loss of HF (ie.m -20) and H, (m -2) in the molecular ion spectrum are 10 and 20% respectively of the same peaks in the CID spectrum of mH+. At the lowest pressures used in this study m + 1 : m was mea-sured to be 0.2. The 13C contribution to the CID peaks caused by loss of HF and H,is calculated to be a maximum of 3 and 6%, respectively, under these conditions (i.e.in the range where direct F atom protonation is significant). Iso-topic interference is therefore not significant in this region, although it may cause inaccuracies when the F atom proto-nation is low, i.e. at pressures ~0.02Torr and >0.12 Torr, depending on the temperature. B/ELinked Scan Spectra Dramatic changes in the CID spectra have already been reported when the source temperature is When experiments were conducted at a constant temperature (0 "C) on the MIKES instrument, varying the pressure also showed dramatic effects. This is shown in Fig. 3(a).The population of F-protonated species increases sharply with pressure, before reaching a maximum and then decaying.Unfortunately we were unable to measure the pressure directly in this instru-ment. Similar experiments were, however, carried out on the forward geometry instrument equipped with a controlled high-pressure source and inlet system. Because of the ESA/ magnet geometry it was necessary to use B/E linked scanning to record the CID mass spectra. Fig. 4(a) shows the B/E linked scan CID spectrum of the protonated mixture. This 1375 \ I II I I I I I \\ \ I \ \ I II J I....I....I....( 0 5 10 15 20 source housing pressure/l 0-6Torr 0 2-t, I \ \i \ \ I \ \ \ \ \ \ \ \ \ \ \ \ .-: I' I cOl.'.'.~.~'~'~' 0 2 4 6 8 10 12 14 16 18 20 pressure/l 0-2Torr Fig.3 Dependence on source pressure of [HF loss] : [H, loss] ratio in the CID mass spectrum of C,H,FH+ (protonated in CH,); (a) MIKES data obtained at 275 K, (b) B/E linked scan data obtained at 253 K; 0,[20]/[22]; +, [21]/[22]; 0,[23]/[22]. method does not show the same velocity discrimination effects as the MIKES technique and the B/E measured HF loss peak therefore contains both ring and substituent proto-nated components, hitherto distinguished by their different peak widths. Nevertheless the pressure dependence was reproduced as shown in Fig. 3(b),except that, instead of decaying close to zero the ratio of peak heights eventually reaches a constant baseline value because the high-energy component is largely independent of the source pressure (it is more usually the case in tandem mass spectrometry that the spectra of high-energy ~' 5 10 15 20 mass lost r5 10 15 20 L, mass lost Fig.4 CID mas spectra (B/E linked scan) of (a)C6HsFHf and (b) C6DsFH+, both protonated in CH, at source pressures of 0.07 and 0.045 Torr and 253 and 258 K, respectively processes are relatively insensitive to the source conditions "). The peaks corresponding to neutral losses at 21, 22, and 23 (corresponding to the consecutive losses HF + H, HF + H, and HF + H, + H) are also obviously high-energy processes and their relative intensities are observed to be independent of pressure as also shown in Fig. 3(b).The high-energy contri- bution to the HF loss peak in the B/E linked scan spectrum can therefore be estimated by reference to one of its neigh- bouring peaks and subtracted from the overall peak intensity.The resulting peak height then corresponds to the signal resulting from direct fluorine protonation. This corrected signal was used to follow the changes in the ratio of ring- to F-protonated species with the source conditions. The linked scan CID spectrum of C,D,FH+, formed in reaction (2), was also recorded under conditions where proto- nation directly on the F atom was thought to be most favoured [see Fig. 4(b)], and has been reported in a prelimi- nary form before. 2,' CH,+ + C,D,F +CH, + C,D,FH+ Comparing with Fig. 4(a)it is evident that the peaks at 21, 22 and 23 have been shifted by 1, 2 and 3 mass units, respec- tively, signifying the involvement of D atoms from the ring (via high-energy processes) with the peak at 20 remaining, without any scrambling, confirming that protonation has occurred on the F atom.The small peak at m -21 is due to DF loss after migration of the D from the ring. This shows that >85% of HF loss under these reaction conditions arises J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 from direct bond cleavage of the F-protonated form. This is in good agreement with the data in Fig. 3 and the recent experiments by both Harrison and co-~orkers'~ and Schwarz and co-workers' ' which confirm the direct proto- nation on the substituent. The loss of the H, equivalent gives rise to a scrambled spectrum, as expected, but showing an isotope effect in favour of HD over D, by a factor of 2.5.Kinetic Mechanism, assuming Bimolecular Protonation In addition to the competing proton-transfer reactions (ring or F atom) the ions of interest are removed from the plasma by rapid (ambipolar) diffusion to the walls of the ion source, which acts to maintain the steady state. Also, it is probable that some of the F-protonated species will be removed by proton transfer to fluorobenzene to form the preferred ring- protonated species, .its importance increasing with partial pressure of the halogenobenzene. This is shown to be broadly true in Fig. 5 where the overall pressure (in the region where the signal is decaying) remained constant, whilst the fluorobenzene :methane mixing ratio was increased.In addition there is evidence, from Harrison and co-workers' chemical ionisation studies6 of these types of system, for dissociative proton transfer, which has been attributed to direct protonation at the halogen substituent, leading to direct loss of HF. This is not surprising in view of the small bond dissociation energy which is estimated? to be ca. 28 kJ mol-',but could be higher or lower dependent on the exact proton afinity of the F atom, which we have estimated' as being in the region of 590 kJ mol- (see Fig. 2). If this occurs by thermal decomposition of the otherwise stable F-protonated molecule (as opposed to direct dissociative proton transfer, which will be invisible to the experiments, see below) then this removal process will also contribute to the observed steady-state concentrations and is expected to be 01 1 1 1 0 0.005 0.01 0.015 0.02 mixing ratio (x) Fig.5 Variation of [HF loss] : [H, loss] ratio in the CID (MIKE) spectrum of C6H5FH+ with mixing ratio x = [C,H,F]/[CH,], at constant temperature (275 K) and pressure (in the region 0.01-0.1 Torr) t The energies were calculated using the heats of formation data of S. G. Lias et a/.'* J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 directly dependent upon pressure (assuming conventional 'Lindeman' kinetics) at the low pressures employed in these experiments. The possible reactions are summarised in a generalised way in the following reaction scheme, assuming bimolecular protonation for both isomers.CH,' + RX+CH,+ RXH' (a) CH,' + RX-+CH,+ H'RX (b)? RXH' + RX-+ RX + H'RX (4 RXH'( + M) + R' + HX (4 RXH' -+ lost by diffusion (e)H'RX R and X denote the aromatic ring and the halogen substit- uent and RXH' and H'RX represent substituent- and ring- protonated species, respectively. M represents a gas molecule here. It is assumed that since the two isomers are very similar in size and overall structure that their diffusion coefficients are equal. This helps to simplify the kinetics without affecting the result in any significant way. Diffusion rates are, of course, inversely dependent on pressure. In addition, during an experiment, the fluorobenzene :methane mixing ratio, x, was kept constant whilst total and partial pressures were varied.At constant pressure, 'steady-state' analysis yields the ratio for the concentration of ring-(R) to substituent-(X) proto- nated species as : [H'RX] k, (k,[RX] + k,. + ke) +-k,[RX] (1)-=-[RXH'] k, ke ke where k, etc. represent the appropriate rate constants and kd = &,3CM1.Hence, if at sufficiently low pressures k,[RX] 6ke and ke/(kd*+&e) -+ l, then [loss of HX] [RXH'] k [H'RX] k,[loss of H,] xpz2 (1 7 Temperature Dependence Fig. 6 shows the [HF loss] :[H, loss] ratio for fluorobenzene protonated in methane, obtained in the earlier MIKES experiments,* under conditions of constant mixing ratio and a source pressure in the region close to optimum F-protonation (i.e.0.02-0.06 Torr ; see below) but varying tem- perature. According to eqn. (1') the ratio at constant relatively low pressure is proportional to the rate coefficients of the com- peting proton-transfer paths. Under optimum conditions therefore, the changing peak height ratio as a function of temperature reflects their relative temperature dependences. Exothermic proton-transfer reactions are usually largely independent of temperature." The inverse temperature dependence of this ratio therefore at first seemed surprising, since the proton affinity of the F atom site is significantly less than the ring. However, there are three possible mechanisms which could contribute to this effect.Note that there will probably be a contribution to the ring protonation from C,H,+ [E,,(C,H,) = 680 kJ mol-'1. This will have some effect on the [HF] : [H2] ratios in favour of the ring- protonated species, noticeable particularly at the higher pressures. Over the pressure range of interest the [CH,'] :[C,H5+] ratio (x 1) remained constant. It was ignored here in the interests of simplicity since the gross features of the behaviour are not affected. 260 340 420 500 580 T/K Fig. 6 Variation of [HF loss] : [H2 loss] peak ratio with source temperature in the CID (MIKE) spectra of C,H,FH+ (produced in CH, reactant gas) showing T-'.,, dependence (-) over the 350-610 K range, at a fixed source pressure (in the region 0.04-0.06 Torr); (---) shows the expected variation according to the ADO theory ;(-.-) shows the variation expected for an Arrhenius based function for the dissociation of the F-protonated species, for which the energy of activation = 28 kJ mol- The 'locking ion-dipole' is a well known" mechanism in gas-phase ion chemistry in which, at low temperatures, the positive charge of the incoming ion (CH,') 'locks on' to the rotating negative end (F) of the dipolar neutral molecule (C,H,F) from a relatively large distance.In the present case this guides the proton transfer from the methane molecule onto the fluorine atom before it has chance to experience the attraction, at closer quarters, from the aromatic ring. At higher temperatures, however, the rotational energy of the molecule becomes large enough to overcome the locking ten- dency of the ion-dipole interaction, exposing the proton to the energy surface leading to ring protonation.Normal exothermic proton transfer is largely temperature independent, usually occurring at the Langevin collision rate. The magnitude of the temperature dependence of the ion- dipole effect is simply calculated using the averaged dipole orientation (ADO) theory, which has the formz0 k -= 1 + PT"2 = 1 + 6.8T1',k, where k,-is the Langevin collision rate and P = ~p~1.4l(ank)-'/~,with p (electric dipole moment) = 1.6 Dt2' and a (polarisability) = 11 A3 (the measured value for benzene21 and here assumed to be the same for fluoro- benzene, though, in fact the value will be slightly higher), is Boltzmann's constant and c z 0.21 is the dipole locking con- stant.' If this effect acts in favour of F atom protonation then it must act in the opposite sense against protonation on the ring, since it would constitute a rotational barrier to the incoming proton, making the relative rate constant for ring protonation less than k, by the same fraction that proto- 1 D z 3.33564 x C m.nation on the F atom is increased. Thus when comparing the two, as is done in these experiments, the ratio is expected to change according (1 + 6.8T0.5)2.The expected variation is also shown in Fig. 6, with the theoretical curve normalised to the experimental data at 273 K by a factor of cu. 2. (This implies that the differential cross-section for direct HF loss during CID is twice that for H, loss, see above.) Although the ion4ipole mechanism must operate during the initial encounter, the temperature effect is clearly much too small to account for the observation, except at the lower tem-peratures. The second possibility arises if thermal decomposition [reaction (d)] and hence kd, becomes significant compared with k, (diffusive loss) at the higher temperatures.If this were the case a very much sharper exponential variation might be expected than is observed, unless the energy of activation is <5 kJ mol-'. With such a small value it is very doubtful that the F-protonated species would ever be observed in the first place except at very much lower temperatures.The indica- tions are that the enthalpy of the reaction C6H5FH+-+C6H,+ + HF is in the region of 28 kJ mol-', which, if equated to the energy of activation, gives an Arrhenius (exponential) tem- perature profile which is very much steeper than that observed as shown in Fig. 6. The temperature profile is similar to the behaviour exhibited by other reactions,,, exhibiting a negative tem- perature dependence and thought to be due to the formation of an intermediate complex. Harrison et a!. postulated some time ago23 the formation of a long-lived, though weakly proton-bound, complex between CH,+ and neutral fluoro- benzene. Such reactions are characterized by a third-order pressure dependence and this might help explain the dra- matic increase in fluorine protonation observed with increas- ing pressure below 6 x lop2Torr.Pressure Dependence The [HF loss] : [H, loss] ratios for fluorobenzene protonated in methane, corrected for the HF loss contribution caused by isomerisation of the ring protonated form, are given in Fig. 7 from two experiments at different fluorobenzene : methane mixing ratios. These show the variation of the [HF loss] : [H2 loss] peak ratios with pressure at a fixed tem- perature of 275 K. If the gas density is represented by [MI, then the mixing ratio x = [RX]/[M] and the pressure-independent diffusion rate coefficient is is given by k, = k,./[M]. Substitution of these terms into eqn. (I) and rearrangement gives the overall pressure variation expected according to reaction scheme A, still assuming straightfor- ward bimolecular proton transfer, as : where a, fland y are constants. Although there is undoubtedly an inverse quadratic depen- dence on pressure at higher pressures, this does not have the very distinctive feature, under nearly all the conditions inves- tigated, of a sharp increase in the F-protonated isomer with increasing pressure at the lower pressures, before reaching a maximum and then decaying at the higher pressures (see Fig.7). However, this behaviour can be accommodated by J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 pressure/l 0 -Torr Fig. 7 [HF loss] : [H2 loss] dependence on source methane pres- sure in the CID (B/E) linked scan) mass spectrum of C,H,FH+, when the source temperature is constant at 273-278 K; and at differ- ent mixing ratios where x = 0.0043 (*), 0.00425 (A)and 0.0085 (0).The (best fit) model function (0.5 [MI-' + 0.233x[M] + 4.65x[M]*)-is shown by the curves (-) and (---) corresponding to x = 0.0043 and 0.0085 respectively. assuming a pressure-dependent first step in scheme A as in (a') instead of the strictly bimolecular reaction (a). CH,++RX+M+CH,+RXH+ (a') Steady-state analysis then leads to the following result: [RXH'] kb kb &x+k" [M]+?x[M]2---{mkr, (k,, i'+[H'RX] k,) k,. ---+ bx[M] + CX[M]~ (m))-This does have the correct form and can be fitted by the func- tion as shown by the two lines drawn through the points in Fig. 7, for two different mixing ratios. It should be emphasised that the data shown were obtained under the most favourable conditions.At the rela- tively low pressures required for these experiments the main ion beam current from the ion source was small; this is com- pounded by 1OOO: 1 reduction in the secondary ion beam current resulting from the CID process, and the correction J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 procedure. There is, therefore, a large degree of scatter, for example demonstrated by the x = 0.0043 data obtained on two different occasions. Nevertheless, a good fit could not be obtained by assuming any other kinetically sensible function. For example, if the thermal decomposition reaction (d) was more important than proton exchange (c), the functional dependence would be closer to l/(a/[M]) + bx[M]).This is, therefore, evidence consistent with the proposed mechanism, involving ion-molecule complex formation in the initial stage of the reaction. The fit is worst at higher pressures when the ratio has decayed, but this is not surprising in view of the assumptions and approximations of the reaction scheme. A significant degree of decomposition of the F-protonated species is not ruled out because this could occur as disso- ciative proton transfer directly from the complex. The energy of the complex is 210 kJ mol-' below the energy of the separated C,H, and FH. It seems unlikely that significant + decomposition will occur via a route which involves col- lisional relaxation towards the ground state of the fluorine- protonated molecule followed by collisional reactivation back again to excited levels.As far as the kinetics is concerned it is much more likely that collisionally activated decomposition from the complex will be in competition with collisional relaxation to the stable protonated molecule. This situation is modelled in reactions (f)-(h). CH,+ + RX complex (f)k-f complex + M 2RXH' + CH, (g) khcomplex + M -R' + HX + CH, (h) Steady-state analysis yields the rate of formation of [RXH'] as : For third-order kinetics (k, + k,[M]) + k-, and the overall rate constant for this reaction is therefore z k, k,/k which is -f equivalent to the rate constant k,. above. In this mechanism therefore the decomposition is largely invisible to the mea- surements.The principle mechanism of removal of the F-protonated isomer, at low temperatures at least, seems to be proton exchange with neutral fluorobenzene itself to give the ring- protonated form. The Proton-bound Complex between Fluorobenzene and CH, + Previous indirect evidence forwarded by Harrison et aL2, and the pressure dependence above indicate the formation of an intermediate complex. H H H -loss of 16 H ,of 173%H H The strong dipole along the C-F bond would cause the CH,+ to align itself towards the F atom giving a species of mass 113, similar to the structure shown (I),determined using MO theory calculation.? Such a species has been observed in the chemical ionisation spectrum, at very low abundance, at low temperatures in previous experiments.* Using the present apparatus it proved difficult to isolate this ion in sufficient quantity to carry out further experiments except when the source was cooled to about 193 K (-80OC) during experi- ments using a benzene : fluorobenzene mixture in methane.The spectrum is recorded in Fig. 8(a), which is dominated by a number of other charge-bound and proton-bound clusters which have been previously reported. The small peak at m/z = 113 is also clearly visible. It was possible to perform a B/E linked scan spectrum of this species (although 70% beam attenuation was required to achieve the necessary sensitivity), the relevant portion of which is shown in Fig.8(b). This shows the most significant peaks at m -16, m -17 and m -36 (m = 113), corre-sponding to losses of CH,, CH, (CH, + H) and CH,F (CH, + HF), respectively. We believe this to be very strong evi- dence that the observed m/z = 113 ions are indeed the complex sought after with the structure anticipated. Semi-empirical MO theory calculations using Dewar's AM1 method24 located a minimum which is just 10-11 kJ 40 Q, .-w 10 0 10 15 20 25 30 35 mass lost Fig. 8 (a) Chemical ionisation spectrum of a C6H,-C6H5F mixture in methane at T = 193 K and pressure = 0.79 Torr, and (b) the CID (B/E linked scan) mass spectrum of m/z = 113 ~~ ~ ~ t The calculations were carried out using MNDO (semi-empirical)loss of 36 and Gaussian (ab initio) Quantum Chemistry Program Exchange I packages at the MND0//4-31G level.mol-’ below the CHSf/C,H5F dissociation threshold. The theoretical structure is shown as structure I, with a short F-H bond and long CH,-HFC,H, expected of a simple ion-molecule complex. The bond lengths shown are in A units. All other bond lengths and angles are close to conven- tional values. The methane-H-F-C part of the molecule is in a plane almost at right angles to the ring. Conclusions The pressure dependence studies support the mechanism of fluorine protonation by protonated methane via a proton- bound complex, but the formation of the fluorine-protonated species is still a relatively low-pressure phenomenon. In our experiments the principle mechanism of removal is thought to be by proton transfer to the neutral fluorobenzene to yield the more thermodynamically favoured ring-protonated species.The occurrence of thermal decomposition of the fluorine-protonated isomer, as suggested by Harrison et a!.; cannot be ruled out as a possible reason why these species are not observed at high temperatures. However, it is an unlikely explanation of the present results, because the tem- perature dependence would be expected to be very much greater than is observed. It is more likely that when it does occur it does so directly from the excited complex (effectively, this would be a direct dissociative proton-transfer mechanism) rather than going through collisional quenching of the newly formed protonated fluorobenzene followed by collisional reactivation.This process would therefore be invis- ible to the present experiments. Although a ‘locking’ ion-dipole mechanism must apply during the initial approach of ion and molecule, calculation shows that the temperature dependence of this phenomenon is larger than predicted by classical theory, except at the lowest temperatures studied. The T-1.5 dependence above 350 K is very similar to other third-order reactions involving intermediate complex formation. A proton-bound complex between methane and fluoro- benzene has been invoked on numerous occasions previously. We believe the evidence of this work gives conclusive proof of its existence, though in view of the difficulty of its observation it is thought to be weakly bound.MO theory calculations give a well depth of 10-1 1 kJ mol- ’. We thank Mass Spectrometry Services Ltd. for a studentship (D.M.P.M.) and the SERC for an earmarked studentship (A.J.P.), and The British Mass Spectrometry Society for a small equipment grant. We are also grateful to Prof. F. Harris and Dr. G. Brenton of the Mass Spectrometry J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Research Unit at UCS for allowing us the use of their equipment for the MIKES work. References 1 B. S. Frieser, R. L. Woodin, J. L. Beauchamp, J. Am. Chem. SOC., 1975,97,6893. 2 Y. K. Lau and P. Kebarle, J. Am. Chem. SOC.,1976,98,7452. 3 K. G. Hartmann and S. G. Lias, Int. J. Mass Spectrom.Ion Pro- cesses, 1978, 28, 213. 4 D. K. Bohme, J. A. Stone, R. S. Mason, R. S. Stradling and K. R. Jennings, Int. J. Mass Spectrom. lon Processes, 1981,37, 283. 5 D. P. Martinsen and S. E. Buttrill, Org. Mass Spectrom., 1976, 11, 5236. 6 W. G. Liauw and A. G. Harrison, Org. Mass Spectrom., 1981, 16, 388. 7 F. Cacace and M. Speranza, J. Am. Chem. SOC., 1976,98,7299. 8 R. S. Mason, D. M. P. Milton and F. M. Harris, J. Chem. SOC., Chem. Commun., 1987, 1453. 9 loth International Symposium on Gas Kinetics, University College of Swansea, 1988; Ilth lnternational Mass Spectrometry Conference, Bordeaux, 1988. 10 D. M. P. Milton, PhD Thesis, University College of Swansea, 1991. 11 J. Hrusak, D. Schroder, T. Weiske and H. Schwarz, J. Am. Chem. SOC., 1993,115,2015. 12 12th International Mass Spectrometry Conference, Amsterdam, 1991, p. 352. 13 A. J. Parry, PhD Thesis, University College of Swansea, 1991. 14 M. Tkaczyk and A. G. Harrison, Int. J. Mass Spectrom. Ion Pro- cesses, 1994, in press. 15 R. S. Mason and A. J. Parry, J. Chem. SOC., Faraday Trans., 1992,88,3331. 16 A. P. Bruins, K. R. Jennings and S. Evans, Int. J. Mass Spec- trom. Ion Phys., 1978,26,395. 17 P. J. Todd and F. W. McLaffery, in Tandem Mass Spectrometry, ed. F. W. McLafferty, J. Wiley and Sons, New York, 1983, ch. 7, p. 149. 18 S. G. Lias, J. E. Bartmess, J. F. Liebmann, J. L. Holmes, R. D. Levin and W. G. Mallard, J. Phys. Chem. Ref: Data, 1988, 17. 19 T. Su and M. T. Bowers, in Gas Phase Ion Chemistry, ed. M. T. Bowers, Wiley, New York, 1979, vol. 1, ch. 3, p. 84. 20 D. R. Lide, CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Boca Raton, 73rd edn., 1992-93, p. 9-42. 21 R. J. W. LeFevre, in Physical Methods of Organic Chemistry, vol. 1, Part 111, ed. A. Weissberger and B. W. Possiter, Wiley, New York, 1972. 22 M. Meot-Near, in Gas-Phase Ion Chemistry, ed. M. T. Bowers, 1979, vol. 1, ch. 6, p. 198. 23 A. G. Harrison and P-H. Lin, Can. J. Chem., 1975,53, 1314. 24 M. J. F. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart, J. Am. Chem. SOC., 1985,107,3902; QCPE package. Paper 3/06914A ;Received 19th November, 1993

ISSN:0956-5000    

DOI:10.1039/FT9949001373

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, %lo), 1381-1390 Virial Theorem Decomposition as a Tool for Comparing and Improving Potential-energy Surfaces :Ground-state Li, A. A. C. C. Pais, R. F. Nalewajskit and A. J. C. Varandas* Departamento de Quimica Universidade de Coimbra P-3049 Coimbra Codex, Portugal The virial theorem has been used to decompose existing potential-energy surfaces for the ground state of Li, into the electronic, kinetic and potential components, and to improve some of the physical attributes to the most accurate potential-energy surface for this system. The behaviour of these components and the total energy for a C,, uniform-scaling cut of the various surfaces is compared. The individual contributions of both two- and three-body energy terms are also discussed.Specific results are presented for the Varandas-Morais and Thompson et a/. potential-energy functions, in addition to the recent double many-body expansion (DMBE) potential-energy surface, this being the only existing Li, surface for which the normalization of the kinetic field, implied by the virial theorem, is satisfied. By slightly modifying adjustable parameters of the latter surface, an improved DMBE surface has been obtained with qualitatively correct trends in the kinetic and potential energy profiles. The implications for the ideal general form of a potential-energy surface are also assessed. 1. Introduction Virial theorem decomposition analysis is known to be a sen- sitive probe to details of potential-energy surfaces (PES), and hence may be envisaged as a useful tool for discriminating between potential-energy functions with similar topo-graphical features.lV2 It can also be used to identify physically unjustified trends in potential and kinetic components of the Born-Oppenheimer potentials.In this work, we use the virial theorem to compare three potential-energy for ground-state Li, . Being derivative properties, the kinetic and potential components' of the PES are expected to be more sensitive to subtle topographical details, allowing each func- tion to be inspected at a more resolved level while ensuring a more lucid separation of common general physical trends from spurious features. The present analysis follows previous work' (paper I), and focuses on the two-dimensional representation of the fixed- angle atom-exchange surface. The included angle is therefore fixed at the value of the corresponding equilibrium Li, struc- ture.This kind of representation is especially useful because it provides a simple picture of the surface from the united (two- or three-) atom limit up to the dissociation channels, includ- ing the equilibrium geometry. We will discuss the total energy and its kinetic and poten- tial components as well as the corresponding two- and three- body contributions. An appropriate minimum-energy path will be determined for each function, and the various energy cuts along this path presented and rationalized. The uni- formly scaled path, ranging from the united atom to the com- plete fragmentation limit will also be investigated.This kind of path is particularly useful for testing the integral virial theorem restrictions. 2. Basic Theory Within the virial the electronic kinetic T(R)and potential energy U(R)components of the Born-Oppenheimer PES V(R)= U(R)+ T(R) (1) t Permanent address: K. Guminski Department of Theoretical Chemistry, Jagiellonian University, R. Ingardena 3, 30-060 Cracow, Poland. are further-related by$ 2T+U+V"=O (2) where V" denotes the virial of forces acting on the nuclei of the system for each fixed position (3) and R = {Rl, R2,...} is a collective variable grouping all internuclear distances. Using the analytical form of the PES for the system, eqn. (1) and eqn.(2) then allow the calculation of its components through the expressions T=-V-V" (4) U=2V+V (5) For cases where the total energy results from a summation of terms relating to the n-body building blocks of the whole system, the above equations can be extended to each such n-body contribution to the total energy.' In particular, within the many-body or double many-body expansion formalisms (MBE" or DMBE' '-13), 7= 1Vbi'(Rab) + V%(Rab Rbc 3 Rat) + ' * ab abc = V2)(R)+ V3)(R)+ . . * C V(")(R) (6) n= 2 the result is' 0 = 27-W + U(n)+ y(n) (7) p)= -V(n)-yen) (8) U(n) = 2v(n)+ y(n) (9) where V"(n)= R(aV(")/dR), T = T(") and U = In=Vn).2 If we now write the virial theorem relations for the uniform scaling through the total energy and its kinetic and potential parts using a given reference configuration (Ro), Rs = sR,; s E (0, CC) (10) $ Whenever convenient, the explicit dependence on R will be omitted for simplicity.1382 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 the functions V, U and T become dependent on a single Note, however, that the potential curve for the lowest triplet scaling factor s with s = 0 corresponding to the united atom state of Li, shows R-' behaviour as R +0, i.e. the surface limit, and s -,co representing the separated atoms limit. For has a pole when any pair of atoms coalesce. example, the kinetic component assumes the form The potential-energy function of Thompson et ~1.~ (TILTM) consists of an LEPS form augmented with three- dVT(s)= -V(s)-s -body energy terms, which have been fitted to corrected ab ds initio energies on both the lower and upper sheets of the ground doublet surface of Li,.As for the triplet curve of the (121 VM surface, the two-body energy curves show here an effec- ,Ids tive Coulomb-type behaviour as the united atom is which, upon integration, leads to approached. In this case only data corresponding to the two innermost RKR points of the singlet curve have been used to extrapolate the singlet and triplet potential-energy functions cT(s)ds = -sV(s)I," (13) into the region of the collapsed diatomic limit. As a result, both curves are expected (see later) to present significant Thus, since in the limit s -,0 we have V x constant + z.deviations from the correct coulombic dependence as R +0. ZiZi/Rj, the above integral virial constraint becomes the It should also be pointed out that the overall input data used familiar normalization criterion of the kinetic field to calibrate these curves are basically the same as those used to parametrize the diatomic potentials of the other Li, sur-faces considered in the present work.However, an important difference is that the intermediate region of the Li, curves has where Zi and Zg denote the nuclear charges of atoms A and been here represented seminumerically using cubic splines. B separated by distance Rj; for further details, the reader is Finally, we observe that the TILTM function allows analyti- referred to ref. 1 and 14, and references therein.cal continuation from the lower to the upper Riemann sheet but fails to describe the correct long-range behaviour at the 3. Potential-energy Functions asymptotic atom-diatom dissociation channels. The most recently proposed function for the Li, system' is The simplest of the three potential-energy functions that will of the DMBE type. 11-' Accordingly, it describes accurately be considered in the present work is a generalized LEPS-type the long-range behaviour of every possible dissociation function due to Varandas and Morais3 (VM). This belongs to channel besides fragmentation into atoms. The virial theorem a series of potential-energy functions developed both for restrictions are also satisfied by the corresponding two-body homonuclear alkali-metal trimers, and heteronuclear ones,' energy curves, which have been represented using the which have been widely used in molecular dynamics studies EHFACE2U2O (extended Hartree-Fock approximate corre- (see ref. 16-18 for recent examples).The VM function repro- lation energy) model; in the acronym, the digit stands for duces the experimental dissociation energy of Li, '' by means diatomics while the letter U implies that the model is valid of a parametrized triplet curve, which rules out the possibility also at the united atom limit of the collapsed diatomic. More- of using data on the lower Riemann sheet of the ground-state over, the Li, DMBE function conforms with existing three- Li, surface to predict the upper Riemann sheet. Owing to body ab initio data at the valence regions of the PES, while pair-additive behaviour at large separations, the VM function predicting the atomization energy of Li, within the error also fails to reproduce the correct atom-diatom asymptotic limits of the reported" experimental value.In addition, behaviour. In addition, the associated two-body energy dynamic studies of the Li + Li2(u) exchange reaction using curves do not comply with the virial theorem restriction^,'^ this Li, DMBE PES have shown a good agreement with the although they correctly describe dissociation into atoms. experimentally determined behaviour of similar systems, Table 1 Geometries (characteristic bond length, R, and angle), binding energies relative to the atomdiatom asymptote, energies relative to that of the equilibrium obtuse-angled structure (2B2), AE, and frequencies (asymmetric stretching, bending and symmetric stretching) of some critical points on the lower PES of Li, potential symmetry ~~ R 1% angle /degrees energy /kJ mol-' AE /kJ mol-' as wavenumber"/cm -bend ss 5.4 60 64.8 5.9 - - - 5.2 180 57.8 11.1 368 39i 224 5.7 52 69.2 1.5 161i 235 363 5.2 74 70.7 0 139 209 343 5.44 60 40.2 5.2 - - - 5.12 180 24.6 20.8 364 75i 206 5.72 52.5 44.9 0.5 1OOi 252 367 5.25 70.1 45.4 0 130 190 349 5.46 60 49.5 5.4 - - - 5.59 180 33.3 21.6 264 34 186 5.73 52.8 54.2 0.7 124i 266 398 5.27 70.0 54.9 0 150 202 382 5.47 60 49.5 5.2 - - - 5.52 180 27.4 27.3 265 26 197 5.72 52.8 54.1 0.7 125 270 380 5.27 69.9 54.8 0 151 207 368 For 'Li, .Ref. 3. 'Fit number 1 of ref. 4. Fit IA of ref. 5. DMBE 111, this work. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 namely a marked increase in the reactive cross-section with increasing vibrational quantum number u. A new DMBE-type surface (labelled DMBE 111) will be proposed and analysed in Section 5; it represents a minor modification of the above DMBE (labelled DMBE I in the original paper') function. Table 1 gathers the characteristics of all the PESs discussed in the present work at important critical points. 4. Fixed-angle Atom-exchange Cuts Fig. 1 shows contour plots of the total energy for the lower Riemann sheet of the VM, TILTM and DMBE Li, surfaces in the fixed-angle atom-exchange representation.In paper I, we also examined plots for atom-fixed-diatomic cuts of the DMBE PES. Because such plots were not deemed essential for the purposes of the present comparison, they have been omitted here for brevity. Fig. 1 shows that the three functions have similar overall topographical features, except for a slight maximum in the Li + Li, dissociation valley that is present in the TILTM function [Fig. l(b)] but absent on both the VM and DMBE surfaces. Note also that there are differences in the well depth associated with the Li, equilibrium struc- ture as seen from Table 1 that summarizes the most relevant attributes of some critical points in the three PESs.The virial of forces acting on the nuclei is compared in Fig. 2. A common overall pattern such as that observed for the total energy is now hardly seen. However, in the strong inter- action region, there is a marked similarity between the VM and TILTM functions, which both exhibit a maximum in R, = R, z8u,. On the other hand, the DMBE PES shows this single maximum resolved into a complex structure com- posed of three maxima. This may be attributed to the depen- dence of the long-range Li-Li, dispersion coefficients (C, ,C, and Clo) on the interatomic diatomic coordinate. Clearly, these additional features on the DMBE plots are more likely to be present in regions of the molecular configuration space where these coefficients approach their maximum values.Fig. 3 and 4 show the potential and kinetic components of the total energy. The behaviour of these components for diatomic molecules is well known. At equilibrium geometry, where the forces acting on nuclei vanish, the virial theorem predicts a decreasing trend in U and an increasing trend in T, with decreasing (uniformly scaled) distances. At long dis- tances, close to the separated atoms dissociation limit, the theorem also predicts a decreasing trend in T and an increas- ing trend in U with uniformly decreasing distances. Such general trends are found for both the kinetic and potential- energy components of all Li, functions examined in the present work. Specifically, the components of the VM surface exhibit the smoothest behaviour of all three functions, while those of the TILTM function show some artificial kinks due to the use of one-dimensional cubic splines to represent the intermediate region of the diatomic potential-energy curves.These peculiarities are clearly reflected by the shape of the contours at regions where R, or R, approach 3.5 or which correspond to the points where the splines join physi- cally motivated functions for the singlet curve. The DMBE function maintains in the kinetic and potential components (and this comes as no surprise) features that stem from the complicated structure identified in the virial (as discussed in detail in ref. 2). Such features are not present in the two-body contributions and hence are due to the three- body energy terms, which are compared for the case of the electronic kinetic energy in Fig.5. Again, the generalized LEPS-type VM function is the smoothest of the three, while both the TILTM and DMBE potentials deviate in a more complicated way from a simple sum of two-body curves. As 1 t t I r c t c tt c 7 c t a< 81 t Fig. 1 Total energy V for the fixed-angle atom-exchange cut of (a) the Varandas-Morais, (6) Thompson et al. and (c) DMBE functions for the ground state of the Li, system. In (a)the included angle (a)is 74", while in (6) and (c) a = 70". The initial contour and contour separation are -0.059 E, and 0.0025 E,, respectively. The dashed lines display the approximate minimum-energy path.1384 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 16 12 4 nW0 4 8 12 16 Rho 16 12 0 38 et 4 0 0 4 8 12 16 R,bo 16 16 12 12 0 0 %8 %8 et R A 4 n 0 4 8 12 16 R,la0 Fig. 2 Representation of the virial, "Y, for the corresponding dia- Fig. 3 Representation of the potential-energy component U for the grams of Fig. l. The contours start at -0.1 E, with a separation of corresponding diagrams of Fig. 1. Theinitial contours are -0.1 E, 0.01 E, for the solid lines and -2.0 Eh separated by 0.1 Eh C0.2 E, in (solid lines) and -2.0 E, (dashed lines) with separations of 0.01 E, (a)]for the dashed lines. and 0.1 E, ,respectively. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 16 12 12 4 4 0 0 4 8 12 16 n I,J 0 4 8 12 16 R,la0 R,lao r 12 4 0 0 4 8 12 16 R, la0 c 0 4 8 12 16 0 4 8 12 16 R,b, R,la0 Fig.4 Same as in Fig. 3, but for the electronic kinetic component, Fig. 5 Representation of the three-body kinetic energy, F3’,for the T.The initial contour value and contour separation are -0.1 Ehand corresponding diagrams of Fig. 1. Solid contours start at -0.01 Eh 0.005 Eh for the solid lines and 0.1 Eh and 0.1 E, for the dashed lines, and are separated by 0.01 Ehand dashed contours start at -1.0 Eh respectively. and are separated by 0.05 E, . might be anticipated from the corresponding calibration pro- cedure~,~,~they show a similar pattern in the central region for values of the internuclear distances over the range 8 < R,, R,/a, < 12, while significant differences are manifest at regions where any pair of atoms tends to collapse.This may be partly attributed to the use of different polynomial forms (these are based in hyperspherical and D3hsymmetry coordi- nates, respectively), which behave differently as one distance approaches zero. Note that the area enclosed by the contours labelled 2 in Fig. 5(c), for R, or R, > 12a,, corresponds to regions of the atom-diatom dissociation channel where the Li-Liz dispersion coefficients approach their maximum values. As already noted, this feature of the DMBE surface is absent in the other two Li, functions, which may be attrib- i 50 r -50 -_ -100 -150 0 5 10 15 20 cia, 100 50 -100 I-150 I J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 uted to their failure to describe properly such atomdatom asymptotic channels. Superimposed on the contour plots of Fig. 1 are the approximate minimum-energy paths, the pro- files of which are shown in Fig. 6.Clearly, a common feature to all functions is the occurrence of a maximum in the kinetic energy at the equilibrium structure, which is counterbalanced by a minimum in the potential component. However, there are also notable differences as we move away from the equi- librium structure. In particular, the TILTM function shows an additional maximum (minimum) in the kinetic (potential) component as one moves outwards in the atomdiatom dis- sociation channels.These additional extrema are probably due to the slight maximum found in the total energy as the atom separates from the interacting diatomic. Although such 50 L 100 1 0 'L I-.., 1 0 5 10 15 20 clao Fig. 6 Minimum-energy path profiles for the corresponding diagrams of Fig. 1. The solid line represents the total energy, and the short- and long-dash lines the kinetic and potential components, respectively. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Numerical definition of the polynomial coefficients for the DMBE I11 Li, PES of the present work; see eqn. (34) and (39) of ref. cooo/E, = -7.4164076 x cloo/a~'= -4.1102270 x c200/a;2 = 6.215621 3 x co10/a;2 = 8.3844444 x c300/a;3 = 7.821 7005 x 10-~ cll0/aO3= -1.6460804 x c001/u03= 5.231 5723 x lo-' d/a;' = 0.5 k = 0.413 503 27 d/a,3 = -0.1 a general pattern can be observed in Fig.qc) for the DMBE function, the extrema are in this case less pronounced and clearly not related to the existence of a barrier for disso- ciation, which is absent in this case. Fig. 7 shows, for each PES, the profile calculated for the 500 400 300 200 100G E.-.roP c -100 I I -2001 I/ -4001 \ , ,/ , , , I ~-500 ' 0 1 2 3 s uniform scaling cut using the equilibrium geometry as the reference configuration. As anticipated, these profiles have a general appearance similar to those obtained for diatomics, especially in Fig. 7(4, (b).Again, the DMBE curves show a slightly more complicated structure, exhibiting undulations that are absent in the VM and TILTM cases.Such features can probably be ascribed to the interplay of the various three-body terms, which tend to restrain the two-body behav- iour.2 Finally, we have tested the integral virial relation of eqn. (14) for each function. The results confirm that they all exhibit poles at the united atom limit, while yielding effective nuclear charges of 2 = 1.2, 47.4 and 3.0 e for the VM, TILTM and DMBE functions, respectively. Of course, the first two values are meaningless in what concerns the nuclear charge of the Li atom, while the DMBE surface emerges as the only function with the properly normalized kinetic field. 400 0 1 2 3 s I500 s Fig. 7 Uniformly scaled profiles generated by the equilibrium geometry for each of the diagrams presented in Fig.1. Lines as in Fig. 6. 5. A further Analysis of the Kinetic and Potential Profiles A comparison between (a)-(c) of Fig. 7 reveals that the DMBE PES has distinctly less monotonic behaviour of its kinetic and potential components in comparison with the corresponding profiles generated by the other this applies to both DMBE I and DMBE I1 Li, PESs (notation as in ref. 5). In particular, the short-range plateau in the repulsive part of the kinetic-energy profile, and the associated feature in the potential-energy profile at the same geometries, are somewhat intriguing. However, one could expect some slowing down of the increase in the kinetic energy at such short distances due to the expected shell reconstruction as the united-atom regime is entered.Similar behaviour has been observed for diatomic molecules involving heavy atoms.20 This effect can be interpreted as a manifestation of the Pauli forces starting to populate the united-atom outer shells at such short distances. Clearly, this effect may be responsible for a temporary relaxation of the atomic electronic densities on top of the ongoing overall contraction due to the extra attractive potential of other atoms. It would be interesting to see how sensitive this feature is to a small variation in the adjustable parameters of the DMBE functional form. This is shown in Fig. 8 and the numerical values of the polynomial coefficients for the modified DMBE potential (DMBE 111)are given in Table 2.Although the short-range plateau of the kinetic-energy profile is now transformed into a much less pronounced shoulder, some short-range trend discontinuity in the T and V profiles remains; Table 3 shows that the high quality of the original DMBE surface is generally preserved so ' , .--I I I I / I -300 1; ,/I'/ -400 p,,, I -500 S Fig. 8 Uniformly scaled profiles generated by the equilibrium refer- ence geometry for the DMBE 111 Li, PES of the present work. The surface modifications include only a slight adjustment of so from so = 0.61507 a, to a somewhat larger value (so = 0.675 a,), and a decrease by 50% of the original value (0.20) of the qn exponent in ?,, damping function (see ref.13). Note that f,,basically controls the rate of convergence of the three-body dynamical correlation term to its asymptotic behaviour: the smaller it is the longer is the distance at which such asymptotic atomdiatom behaviour is reached. In the present case, the spherically averaged component of the Li-Li, potential agrees with the asymptotic form to within 10% for atom- diatom separations 220 a,. The open circles indicate corrected5 ab initio energies,24 while the lines are as in Fig. 6. \-J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 3 Unweighted rms errors (in kJ mol-') for PESs fitted to ab initio data" TILTM~ DMBE-I' DMBE HId ~~ lower sheete 1.09 0.75 0.55 upper sheetf 4.14 1.11 " Only the ab initio data points referring to ref.23 and 24, have been considered for the calculation of the rms error. Fit number 1 of ref. 4. Fit IA of ref. 5. This work. Value referring to all ab initio energies used in the fit. Predicted by analytic continuation. in DMBE 111. We note that the above discontinuity is missing on both the remaining PESs compared in this study; this is because the kinetic-energy normalization is not satis- fied. As a result, these PESs either produce repulsive walls that are too hard (TILTM) or too soft (VM) in the inner part of the uniform-scaling cut, corresponding to united-atom effective nuclear charges that are too high or too low, respec- tively. The DMBE curves in Fig. 7 also show two stationary points in the outer part of the T and V profiles.This reflects a slight discontinuity in the polynomial interpolative region between the ab initio covered geometries and the long-range behaviour. Although not visible within the scale of the figure (and most likely, of little practical consequence for molecular dynamics), this feature has been removed in Fig. 8 with the resulting (T,V) profiles of the DMBE I11 surface now resem- bling diatomic curves. As also seen in Fig. 8 and in Table 3, the very few ab initio points existing close to the uniform scaling section are reproduced by all DMBE functions. Topo- graphical details concerning this new DMBE PES and its virial partitioning components are shown in Fig. 9-14. Also shown in Fig.14 are the relevant profiles along the minimum-energy path for the bond-stretching contour diagram of Fig. 9. Although the new DMBE PES shows topographical attributes similar to those of the previous DMBE I PES (see Fig. 1 and related properties in Table l),it is now devoid of some features of the previous surface that were considered, in the absence of accurate relevant ab initio information, as physically unjustifiable. This subtle improve- ment is clearly seen in the virial partitioning maps of Fig. 10-14. The fixed-angle maps now exhibit smooth two-body- t 0 0 4 8 12 16 Wa, Fig. 9 Total energy I'for the fixed-'angle atom-exchange cut of the DMBE 111 PES of the present work. Contours and dashed line as in Fig. 1. J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 a 0 0 4 8 12 16 R,la0 Fig. 10 Representation of the virial, Y,for Fig. 9. Contours as in Fig. 2. 16 12 %8 a 4 0 4 8 12 16 R,la, Fig. 11 Representation of the potential-energy component U for Fig. 9. Contours as in Fig. 3. 0 4 8 12 16 RIbO Fig. 12 Same as in Fig. 11, but for the electronic kinetic component, T.Contours as in Fig. 4. 16 \ 12 0 %8 a 4 0 0 4 8 12 16 R,la0 Fig. 13 Representation of the three-body kinetic energy, F3),for Fig. 9. Contours as in Fig. 5. 100 50 0 LLlr E--.>. P Q, -50 -100 -150 I,I,III 0 5 10 15 20 clao Fig. 14 Minimum-energy path profiles for Fig. 9. Lines as in Fig. 6. type behaviour.We recommend these plots as the first qualit- atively reliable (T, U)surfaces (profiles) in the angle-constrained three-body system exhibiting a minimum in the interaction region. Thus, they can be used to comment on possible shortcomings of the virial partitioning maps for the VM and TILTM PESs. Fig. 5 and Fig. 13 show that the three-body energy has been modified only slightly in the DMBE I11 surface relative to the DMBE I potential. We also observe from the virial plots [Fig. 2(a), (b) and Fig. lo] that the VM virial map shows an unphysical minimum close to the united-atom region [(R,,R,) < 4a,], most likely due to improper normal- ization of the diatomic potential-energy curves. In turn, the TILTM virial surface descends too steeply in comparison with the corresponding trend observed in Fig.10. Regarding the potential component plots, they indicate that the TILTM Usurface lacks the very deep minimum close to the united atom, (Rl, R2)< 2a,, as a result of the unphysical shape of the associated virial in this region. This minimum is seen on the remaining U surfaces, although that on the VM plot is much too shallow as a result of the underestimated effective nuclear charges. Both these observations are consistent with our kinetic-energy normalization findings. As for the VM and TILTM T surfaces, the above artifacts of the V and U plots are reflected in the corresponding kinetic-energy surfaces. Specifically, the VM T map exhibits an unphysical short- range maximum [(R, , R2)< 2 a,], while the TILTM T map shows a hard-sphere-type repulsive wall at short distances.The observed locations of the kinetic-energy minimum are similar on all PESs, being placed at about R, = R2z 7a,. Finally, the MEP profiles of the VM surface are similar to the DMBE-I11 profiles of Fig. 14 and in agreement with qualitative predictions from the virial theorem.' Both plots are devoid of the artifacts present in the corresponding pro- files of Fig. 6. 6. Conclusion The analysis of the VM, TILTM and DMBE PESs for the ground state of the Li, system has shown that the electronic, kinetic and potential components of otherwise similar func- tions for the total energy may differ significantly. The criteria suggested by Kuntz2' and later elaborated by others12.22 for the ideal general form of a PES should therefore be com- plemented if the virial theorem is to be taken into account. Accordingly, such an ideal form should represent the true PES accurately in interaction regions for which experimental or non-empirical theoretical data are available" (including the intermediate and long-range regions associated with the various asymptotic channels' 2, and describe correctly the virial and the electronic kinetic and potential components.Moreover, it should behave in a physically reasonable manner in those parts of the interaction region for which no data are available, and so should its components. Complementary to paper I, the present analysis allows some additional conclusions. Intuitively, we may say that the simpler the potential form the smoother the behaviour of its components, as illustrated by the VM surface. However, the potential-energy function has to have sufficient complexity to accommodate all physically justified features of the kinetic and potential components.In fact, such components may act as a 'magnifying glass' to identify any subtle (physical or unphysical) behaviour not noticeable in the total energy. We have demonstrated how this virial partitioning analysis can be used to improve physical features of the already accurate DMBE PES, without sacrificing its low rms error relative to the fitted ab initio energies. This approach may be especially sensitive to unphysical kinks due to use of local terms in the representation of the PES, as it is the case for the TILTM potential.J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Finally, we would like to stress that the virial partitioning represents an extremely demanding test for analytic PESs. We would advise it especially in the case of tailor-made sur- faces (i.e. built from functions valid only for limited regions of the whole configuration space) for which local trends have to be compromised. It would be interesting to see how such deli- cate features of the potential-energy function affects molecu- lar dynamics. This work was supported by the Junta Nacional de Investigaqgo Cientifica e Tecnologica, Portugal. R.F.N. thanks the State Committee for Scientific Research in Poland, and the Commission of the European Communities (COST Action D3 programme) for a travel grant.References 1 A. J. C. Varandas and R. F. Nalewajski, Chem. Phys. Lett., 1993, 205,253. 2 A. A. C. C. Pais, R. F. Nalewajski and A. J. C. Varandas, J. Chem. SOC.,Faraday Trans., 1993,89,3885. 3 A. J. C. Varandas and V. M. F. Morais, Mol. Phys., 1982, 47, 1241. 4 T. C. Thompson, G. Izmirlian Jr., S. J. Lemon, D. G. Truhlar and C. Alden Mead, J. Chem. Phys., 1985,82,5597. 5 A. J. C. Varandas and A. A. C. C. Pais, J. Chem. SOC., Faraday Trans., 1993,89, 1511. 6 V. Fock, 2.Physik, 1930, 63, 855. 7 P-0. Lowdin, J. Mol. Spectrosc., 1959,3, 46. 8 P-0. Lowdin, Adv. Chem. Phys., 1959,2,207. 9 B. Nelander, J. Chem. Phys., 1969,51,469.10 J. N. Murrell, S. Carter, S. C. Farantos, P. Huxley and A. J. C. Varandas, Molecular Potential Energy Functions, Wiley, Chi- Chester, 1984. 11 A. J. C. Varandas, Adv. Chem. Phys., 1988,74,255. 12 A. J. C. Varandas, in Trends in Atomic and Molecular Physics, ed. M. Yaiiez, Universidad Autonoma de Madrid, Madrid, 1990, p. 113. 13 A. J. C. Varandas, Chem. Phys. Lett., 1992,194,333. 14 R. F. Nalewajski, Chem. Phys. Lett., 1978,54, 502. 15 A. J. C. Varandas, V. M. F. Morais and A. A. C. C. Pais, Mol. Phys., 1986,58, 285. 16 V. M. F. Morais and A. J. C. Varandas, J. Chem. SOC.,Faraday Trans. 2, 1987,83,2247. 17 V. M. F. Morais and A. J. C. Varandas, J. Chem. SOC., Faraday Trans. 2, 1989, 85, 1. 18 H-G. Rubahn and N. Sathyamurthy, Mol. Phys., 1993,78, 1047. 19 C. H. Wu, J. Chem. Phys., 1976,65, 3181. 20 A. J. C. Varandas and J. D. Silva, J. Chem. SOC.,Faraday Trans., 1992,88,941. 21 P. J. Kuntz, in Dynamics of Molecular Collisions, ed. W. H. Miller, Plenum Press, New York, 1976, p. 53. 22 J. S. Wright and S. K. Gray, J. Chem. Phys., 1978,69,67. 23 W. H. Gerber and E. Schumacher, J. Chem. Phys., 1978, 69, 1692. 24 W. H. Gerber, Ph.D. Thesis, University of Bern, 1980. Paper 3/07073E; Received 29th November, 1993

ISSN:0956-5000    

DOI:10.1039/FT9949001381

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(10), 1391-1396 Pulse Radiolytically Induced Redox and Alkylation Processes of C,, Dirk M. Guldi, Hartmut Hungerbuhler, Martin Wilhelmt and Klaus-Dieter Asmus* Hahn-Meitner-lnstitut Berlin, Bereich Physikalische Chemie , Abteilung Strahlenchemie Glienicker Str. 100, 14109 Berlin, Germany Redox and alkylation (radical addition) reactions with C70 have been investigated by means of radiation chemi- cal methods, particularly time-resolved pulse radiolysis measurements. All experiments were conducted in solutions (propan-2-01, 2-methylpropan-2-01 and 1,2-dichIoroethane) at room temperature. The differential spec- trum obtained upon one-electron reduction of C70 by (CH,),C(OH)' (k = 8 x lo8 dm3 mol-' s-') and attributable to the formation of C,o*- shows a relatively weak but still distinct absorption band in the IR at 880 nm.One- electron oxidation of C70 by the cation of 1,2-dichIoroethane (1,2-DCE'+) yields C70" radical cations (k b 2 x 10" dm3 mol-' s-') with an even weaker but still noticeable band at 925 nm. Compared with the IR-absorption bands of the analogue &*-and CG0'+ (which exhibit relatively strong and characteristic bands at 1080 and 980 nm, respectively) those attributable to the C7,-derived species are considerably less intense but not absent as might have been deduced from earlier experimental and theoretical studies. IR absorptions (with maxima around 960 nm) have also been observed for (C70-CH,CH,CI)' and (C,,-CH,C(CH3),0H)' which are formed upon addition of 'CH,CH,CI (k = 1.9 x lo9 dm3 mol-' s-') and 'CH,C(CH3),0H (k = 1.8 x lo9 dm3 mol-' s-'), respectively, to C70.The absolute rate constants are very similar to those measured for the corre- sponding radical reactions with Ce0. In conclusion, C70 exhibits rather similar features to those of c60 with respect to radical and one-electron redox initiated reactions with, however, some noticeable differences. Since the first successful synthesis of c60 by Kratschmer et some in the IR region, attributable to C70'-, C70'+ and C70 al.' many interesting properties have been unveiled in numer- radical adducts. In addition, absolute rate constants will be ous studies. The most essential physico-chemical properties of reported for the reaction of C70 with a number of radicals in this so called buckminsterfullerene have been summarized in solution.several review Redox chemistry studies, con-ducted particularly with electrochemical methods,' '-' have revealed, for example, the ability of c60 to accommodate up Experimentalto six electrons in subsequent reversible reduction steps' while oxidation appeared to be restricted to a one-electron C70 was purchased from Fluka Chemika; its purity was con- process (yielding the fullerene radical cation).' ' A further sig-trolled by high-performance liquid chromatography (HPLC). nificant feature of fullerenes appears to be their capability of The c60 content (as major impurity) amounted to <5?40 and adding radicals to their double-bond system, as shown es- did not adversely affect the results.This became apparent by pecially by the electron paramagnetic resonance (EPR) the lack of any of the characteristic absorption features experiments of Krusic and co-workers.' 6-1 * A very suitable known for the C,,-deriVed radical species. The solvents technique for the direct study of all these three types of reac- [propan-2-01, 2-methylpropan-2-01 and 1,2-dichloroethane (1, In a recent detailed study we 2-DCE)] were generally redistilled prior to use. Solutionstions is pulse radioly~is.'~-~~ reported, for example, on absolute rate constants for the reac- containing (1-5) x lo-' mol dm-3 C70 in the desired tion of c60 with a variety of free radicals and on character- medium were always freshly prepared for each set of experi- istic absorption features, particularly in the IR regi~n.'~.~~ ments.Concentrations of C70 were measured by UV-VIS The majority of the studies on fullerenes have been devoted spectrophotometry by comparison with authentic samples of to the most stable and symmetrical fullerene, namely c60, known concentration. Prior to irradiation the samples were while much fewer investigations deal with its higher and less purged with either N,-N,O mixtures or pure N, for ca. 1 h symmetric analogue, C70.24 Many properties of C70 are per dm3 of solution. N,O was employed in order to scavenge expected to be similar to those of C,,, particularly with the solvated electrons, generated as a result of irradiation,26 respect to redox reactions.However, some differences appear thus preventing them from direct reaction with C70. to exist for the optical absorption spectra of the transients Pulse radiolysis experiments were performed by utilizing derived from these two most prominent fullerenes. For either 500 ns pulses of 1.55 MeV electrons from a Van de example, while c60'-and c6,'+ have been shown, both Graaff accelerator facility, or 50 ns pulses of 15 MeV elec- experimentally as well as by theoretical calculations, to exhibit trons from a linear electron accelerator (LINAC). Basic distinct optical absorptions in the IR (at 1080 and 980 nm, details of the equipment and the analysis of data have been Dosimetry was based on the oxida- respectively), no such features have been detected so far for described previo~sly.~~.~~ the C70 analogues in, for example, low-temperature matrix tion of SCN- to (SCN),'- which in aqueous, N,O-saturated y-irradiation experiment^,,^ nor did they emerge from earlier solutions takes place with G x 6.(G denotes the number of species per 100 eV, or the approximate pmol dm-3 concen- CNDO/S calculation^.^^ In the present study, relying on sen- sitive pulse radiolysis experiments, we demonstrate that there tration per 10 J absorbed energy.) The radical concentration are nevertheless some absorption characteristics, including generated per pulse typically amounts to (1-3) x mol dm-for all the systems investigated in this study. All experi-ments were carried out at room temperature, 22 k2°C. t Attached to Photochemistry Department at Hahn-Meitner-Insti- Experimental error limits are estimated to be 10% unless tut.specifically noted. Results and Discussion Reduction of C,, One-electron reduction of C70 was achieved by reaction of this substrate with (CH,),C(OH)' radicals in N,O-saturated propan-2-01 solutions. In such a system the solvated electron and the parent radical cation formed as primary species upon irradiation of the solvent (CH,),CH(OH) -+esol-+ (CH,),CH(OH)'+ (1) practically all convert to (CH,),C(OH)' radicals in the course of the following subsequent reactions: esol-+ N,O --* N, + 0'-(2) (CH,),CH(OH)'+ + (CH,),CH(OH) -+ (CH3)2C(OH)' + Hsol+ (3) 0'-+ Hml+-+ 'OH (4) O*-/'OH + (CH,),CH(OH) -+ (CH,),C(OH)' + OH-/H,O (5) The strongly reducing (CH,),C(OH)' radicals (Eo= -1.39 V us.NHE2') are then expected to transfer an electron to the fullerene (CH,),C(OH)' + C70 +C70*-+ (CH3)ZCO + Hf (6) in analogy with the well established corresponding reaction sequence with c60.19,,0 Fig. l(a) shows the differential spectrum obtained upon pulse irradiation of an N,O-saturated, ca. 2 x lo-' mol dmP3 C70 solution in propan-2-01 within the 275-600 nm range. At most wavelengths a positive change in absorption is found, indicating a higher absorptivity of the radical anion than of the C70 itself. Bleaching is observed, on the other hand, at around 360 and 480 nm, i.e. at wavelengths where C70 exhibits noticeable ground-state absorption bands and C70'- absorbs comparatively less.The IR part of the spec- trum (700-1 100 nm), recorded using a different experimental set-up, is displayed in Fig. l(b). In this wavelength range the change in optical absorption is always positive. A reasonably distinct band is noted at 880 nm which may help to identify any one-electron reduction process of C70. Earlier theoretical calculations have not indicated any distinct spectral charac- teristics of C70.- in the IR region.25 This statement relates, however, to a comparison with the analogue radical anion obtained upon reduction of C60. The 880 nm band of C70*- is indeed considerably weaker in intensity than the well resolved IR band (at 1080 nm) of C60*- and this is, at least from a quantitative point of view, in accord with the theoreti- cal considerations.Absorption bands of C70'- in the far-IR (1184 and 1386 nm) have, in fact, just recently been published., Similar experiments were carried out with N,-saturated solutions and qualitatively gave the same spectra. In this case, the solvated electrons reduced the fullerene directly via rather than first being converted to hydroxyl radicals and, subsequently, to (CH,),C(OH)'. Kinetic traces recorded at 860 nm (build up) and 380 nm (bleaching) for N,O-saturated solutions of ca. 1.8 x lo-' mol dm-3 C70 are displayed in Fig. 2(a) and (b),respectively. At both wavelengths the changes in absorption occur exponen- tially with an average half-life of t',, = 15 ps and are attri- buted to the reduction of C70 by (CH,),C(OH)' radicals J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 900 n .-E 600 C $ 300 v 30 -300 -600 250 300 350 400 450 500 550 600 wavelengt h/nm 600 (b1 -500 400 --300 200 -I I II 700 800 900 1000 1100 wavelength/nm Fig. 1 (a) UV-VIS spectrum and (b) IR spectrum of the C70*-radical anion obtained ca. 100 ps after pulse irradiation of a 2 x mol dm-3 C,, solution in propan-2-01 [reaction (6)]. A plot of the first-order rate constants, k = In 2/t1,,, extracted from this section as a function of [C,,] is shown in Fig. 2(c). From the slope of the straight line a bimolecular rate constant of 8 x 10' dm3 mol-' s-' is derived. The rate constant from these concentration-dependent experiments is somewhat higher compared with the value of 5 x 10' dm3 mol- s-measured earlier for the corresponding reduction of c60.19,20 However, owing to the lack of reliable molar absorption coefficients of C70 in propan-2-01, which prevents an accurate determination of the actual C70 concentration, our k6 value may just be good to within a factor of 2.This and all other rate constants measured in the present study are listed in Table 1. In solutions saturated with N, rather than N20 an addi- tional strong and very short-lived absorption signal (tl,, d 1 ps) is seen in the IR region immediately after the pulse. It is attributed to the solvated electron, the decay of which allows an estimate of the rate constant for reaction (7) of k, 2 4 x 10" dm3 mol-' s-' (with the same considerations apply- ing with respect to the error limit as above).It can also be seen that reaction (7) contributes about 20-30% to the longer-lived change in the optical absorption at this wave- length. The second slower process again describes the reduction of C70 by the propan-2-01 radical and occurs at the same rate as in N,O-saturated solutions. Practically the same results were obtained when the reduction of C70 was carried out in a solvent mixture of J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1 - IyL , , , , , , 100 ps , , 9 14.0 -212.0 10.0 8.0 1 , . I J 0.0 2.0 4.0 6.0 8.0 10.0 12.0 [C,,]/106 mol dm-3 Fig. 2 Absorption-time traces recorded at (a) 860 and (b) 380 nm for pulse-irradiated N,O-saturated solutions of 1.8 x mol dm-, C7, in propan-2-01; and (c) plot of k = In 2/t1,, us.[C,,] for the reduction of C,, by (CH,),CO'-/(CH,),C(OH)' in propan-2-01 toluene (80 vol.%), acetone (10 vol.%) and propan-2-01 (10 vol.%). As has been demonstrated earlier for c60, reduction in such a system occurs mainly via formation of the electron adduct to acetone, (CH,),CO'-, and its protonated form, (CH,),C(OH)'.' 9320 1393 Oxidation of C,, One-electron oxidation of C70 was achieved by irradiating a solution of C70 in 1,2-DCE. The solvent radical cations gen- erated via 1,2-DCE +esol-+ (l,ZDCE)'+ (8) are known to be strong enough oxidants to remove an elec- tron from fullerenes, as shown in our previous studies on c60 .19v20 The C70" radical cations resulting from (1,2-DCE)'+ + C70 4C70*++ 1,2-DCE (9) also give rise to changes in the absorbance.An absorption- time trace recorded at 980 nm from an N,-purged solution of 4 x lop5mol dmP3 C70 is displayed in the insert of Fig. 3. It shows an immediate increase followed by a slight decrease on the early ps timescale, before the absorption again starts to increase. The latter has nothing to do with the actual one- electron oxidation process and is attributable to a radical adduct, as will be demonstrated and discussed in the final chapter below. The initial, short-lived absorption can be assigned, however, to the C70'+ radical cation. The entire IR spectrum, composed of the initial absorption changes, is dis- played as the main curve in Fig.3. It exhibits two distinct bands, a narrow one at 830 nm and a somewhat broader one which peaks around 930 nm. By comparison, the analogue C60'+ species also shows a characteristic IR band, although somewhat red-shifted at 980 nm.19,20 As in the case of the respective radical anions (see above), the absolute absorption intensities are considerably lower (ca. 7 times) for the C70- than for the C6,-derived radical cation. Differential absorp- tion changes were also observed in the UV-VIS spectra. They are, however, mainly due to a radical adduct (see later). 140 120 100 h VI+-.-5 80 4 2. 60B 40 20 0 750 800 850 900 950 1000 1050 wavelength/nm Fig.3 IR spectrum of the C7,'+ radical cation obtained imme- diately after pulse irradiation of a 4 x mol dmP3 C7, solution in 1,2-DCE. Insert: Absorption-time trace at 980 nm. Table 1 rate constant" radical /dm3 mol-'~-~ 34x loio (CH3)2C(0H)' (8.0 f2.0) x 10' (CH Cl-CH ,Cl)' + >2 x loio 'CH,CH,Cl (1.9 k 0.5) x lo9 'CH2C(CH3)2(0H) (1.8 f0.5) x lo9 (6.0 4.0) x lo9GO'+ (C,,-CH,CH,Cl)' (2.8 f0.5) x 10' Rate constants for the reactions of various radicals with C70 solvent tYpe.Ofreaction transient product ~ propan-2-01 propan-2-01 reduction reduction CH,Cl-CH,Cl oxidation CH,ClCH,Cl addition CH,ClCH,Cl CH 3C(CH3)2(0H) addition addition CH,ClCH,Cl addition " Evaluation of rate constants was based on pseudo-first-order kinetics.Since some of the C7, concentrations applied were close to the radical concentrations per pulse, the actual rate constants derived therefrom would thus represent lower limits. The build up of the C7,'+ absorption could not be time- resolved further. From the insert in Fig. 3 it is apparent, however, that this process is completed within 61 ps. Accordingly, a lower limit of k, 2 x 10" dm3 mol-' s-l can be given for the rate constant of reaction (9). In the case of C60'+ the decay of the radical cation absorp- tion became more rapid with increasing fullerene concentra- tion. 19,20 Unfortunately, the overall weak absorption changes, together with the influence of the subsequent process of adduct formation (see below), prevents an unam- biguous proof of such a process in the case of C70.It may be anticipated though that the decay of C70" proceeds via the same mechanism as the decay of C60*+, i.e. under formation of a dimer radical cation complex: c70'+ + c70 (c70)2'+ (10) From the kinetic trace in the insert of Fig. 3 a half-life of 3 (& 2) ps may be extrapolated for this reaction, yielding a rate constant of k,, z 6(f4) x lo9 dm3 mo1-ls-l. Radical Addition to C,, 1,2-DCE Solutions The kinetic trace recorded at 980 nm in the pulse-irradiated 1,2-DCE solution (Fig. 3) indicates that the one-electron oxi- dation process is accompanied by a further, slower reaction. Completion of the latter process, recorded at 960 nm from a 2.5 x lop5 mol dm-3 solution of C70, is shown on an extended timescale in the insert of Fig.4(b). In analogy with the c60 this slowly formed absorption is attrib- uted to an addition of 'CH,CH,Cl (formed from the solvent J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 by dissociative electron attachment) to the fullerene : + 1,2-DCE -+ 'CH,CH,Cl+ C1-Any direct involvement of the C,,'+ radical cation in this longer-lived species can be excluded. This becomes evident, for example, from the presence of a very similar, slowly formed absorption upon pulse irradiation of C70 solutions in 2-methylpropan-2-01. In this solvent, only a radical adduct but no cation is formed (further details will be described below). Furthermore, in a photochemical oxidation study of c60, where the only radical species formed was the c60'+ radical cation, no such long-lived IR species could be dete~ted.~'Considering the similarity between the fullerenes with respect to their redox and radical reactions this would also suggest that the C70" radical cation is not a precursor of the slowly formed follow-up transient.The IR spectrum of (C7,-CH,CH,C1)', present ca. 150 ps after the pulse, is displayed in Fig. 4(b). Its most characteristic feature is a relatively broad band with a maximum at 960 nm and two shoulders at around 900 and lo00 nm. By compari- son, the analogous (C60-CH2CH2Cl)' also showed an IR band with a maximum at 900 nm." A plot of the first-order rate constants, k = In 2/t,,,, for the formation of the 960 nm band as a function of [C,,] serves for the determination of the bimolecular rate constant for the underlying process [reaction (12)].From the slope of the straight line, a rate constant of k,, = 1.9 (k0.5) x lo9 dm3 mol-' s-l is calculated for the addition process, which is of the same order of magnitude as those for a large number of radical additions to c60.'9720 Upon extension to the ms timescale the monitored absorption in the IR region exhibits further changes. Owing to the lack of strong absorption fea- tures for the (C7,-CH2CH2C1)' adduct in the IR region, a 1200 , . ry].~ detailed analysis will, however, be restricted to the UV-VIS part of the spectrum (see below). 3 30!I/----600-300 250 300 350 400 450 500 550 wavelengthlnm 800 850 900 950 1000 1050 wavelengthlnm Fig.4 (a) IR spectrum of the (C,,-CH,CH,Cl)' adduct radical obtained ca. 150 ps after pulse irradiation of a 2.5 x mol dmp3 C,, solution and (b) UV-VIS spectrum of the (C,,-CH,CH,Cl)' adduct radical obtained ca. 150 ps after pulse irradiation of a 2.0 x mol dmP3 C,, solution in 1,ZDCE. Insert: Absorption- time trace at (a) 330 and (b) 960nm. Absorption changes resulting from the 'CH,CH,Cl radical addition to C70 are also observed in the UV-VIS spectra. The spectrum recorded about 200 ps after the pulse is shown in Fig. qa). Various regions with an increase in absorption are apparent (particularly around 300, 350 and 410 nm). On the other hand, a net bleaching is observed at around 330, 380 and 480 nm, where the C70 ground-state absorption exhibits maxima and the transient products absorb compara- tively less.The kinetic picture in the UV-VIS region is, however, somewhat more complex than that in the IR. A bleaching trace recorded from a 2 x lo-' mol dmp3 C7, solution at 330 nm, for example, is shown in the insert of Fig. qa). An initial fast step (ca. 50-100 ps) is seen to be followed by a process which is considerably slower (extending over almost 1 ms) than the secondary increase at 960 nm [ca. 100 ps in the insert of Fig. 4(b)]. At the C,, concentration of this experiment the primary radical addition occurs with L,,~= 30 ps while the secondary process exhibits a half-life of about 100 ps.In other words, reaction (12) will contribute only to the earlier part of the signal (i.e.up to about 100 ps). (The strongly disturbing Cerencov signal prevents an unam-biguous resolution at short times.) The slow increase occurs more or less exponentially, and at an increasingly faster rate, with increasing C70 concentration. Furthermore, it parallels a decrease of the 960 nm absorption at longer times. Both observations suggest a consecutive reac- tion of the primary radical adduct with another fullerene molecule, possibly a first step for an oligomerisation : J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 The rate constant for this process, as evaluated from the kinetic trace at 380 nm, amounts to k13 = 2.8 x lo8 dm3 mol-' s-'. Note, however, that the half-life of this secondary reaction does not depend on the C70 concentration alone, but also decreases with increasing dose and thus contains some second-order component, possibly attributable to a radical-radical dimerization.Increasing the dose by a factor of three decreases the half-life from 230 to 80 ps at constant C70 concentration. The above rate constant, k,, ,is therefore presumably associated with a rather large error limit. Corre- sponding observations, with respect to secondary reactions as well as to the associated rate constants, have been made for the analogous c60 system.20 The kinetic trace at 380 nm, a wavelength where bleaching is also observed, is displayed in the insert of Fig. 5. Two sec- tions can be distinguished.The initial step in the kinetic trace (up to ca. 200 ps) occurs exponentially and becomes pro- portionally more rapid with increasing C70 concentration. A plot of the first-order rate constants, k = In 2/t,,,, extracted from this section as a function of [C,,] is shown in Fig. 5. From the slope of the straight line a bimolecular rate con- stant of 1.1 x lo9 dm3 mol-' s-' is derived. It is practically identical with k,, obtained for the primary radical adduct formation from the IR measurements. This section of the 380 nm bleaching is accordingly assigned to reaction (12). At still longer times an additional bleaching process occurs (not shown in the figure) and can be associated with the consecutive reaction of the primary radical adduct [reaction (13)].Its yield, in terms of absolute signal change, is rather small, however, and may just be an indica-tion that the absorptivities of (C,,-CH,CH,Cl)' and (C70-CH,CH,C1)-(C70)' do not differ very much at this wavelength. 2-methylpropan-2-01 Solutions Corroborating results for the radical addition processes are provided by experiments with C70 solutions in N,O-saturated 2-methylpropan-2-01. In this solvent the P-hydroxy radical, 'CH,C(CH3),0H, is practically the only reactive species available for reaction with the fullerene [generated in analogy to reactions (1)-(5)]. For the discussion of the fol- lowing results it is necessary to recognize that P-hydroxy rad- icals, in general, are practically redox-inert. In the case of c60, for example, *CH,C(CH,),OH yielded no c60'+ or C60*-, but only the adduct radi~al.'~.~' The same appears to apply for C70.In the IR, the transient differential absorption spectrum, displayed in Fig. 6(b), closely resembles that shown in Fig. 6.0 I I 5.0 1 / I 2.01.o t/ I I 0.0' " ' I ' ' ' I ' ' ' ' 0.0 1.0 2.0 3.0 4.0 5.0 6.0 [C,,]/105 mol dm-3 Fig. 5 Kinetic analysis of the formation of the (C,,-CH,CH,Cl)' adduct radical absorption at 380 nm in terms of k = In 2/t,,, us. [C,,]. Insert: Absorption-time trace at 380 nm recorded upon pulse irradiation of a 4 x mol dm-3 C,, solution in 1,2-DCE. 000 750 500 250 0 -250 250 300 350 400 450 500 550 600 wavelength/nm 800 850 900 950 1000 1050 wavelengt h/nm Fig.6 (a) IR spectrum and (b) UV-VIS spectrum of the (C,,-CH,C(CH,),OH)' adduct radical, obtained ca. 100 p after pulse irradiation of a 2.7 x loe5 mol dm-3 C,, solution in 2- methylpropan-2-01. Insert: Absorption-time trace at (a) 280 and (b) 980 nm. 4(b)for the (C,o-CH,CH,Cl)' adduct radical. It also exhibits a maximum at 960 nm with shoulders at 900 and lo00 nm. The UV-VIS part of the differential spectrum, shown in Fig. 6(a),is also reminiscent of that of (C,,-CH,CH,Cl)'. Accord-ingly, the transient spectrum observed in the 2-methyl-propan-2-01 solutions is attributed to the radical adduct formed in the reaction *CH,C(CH,),OH + C70 + [C7o-CH,C(CH3),OH]' (14) The kinetic traces, shown in the inserts of Fig. 6(a) and (b), have been recorded from an N,O-saturated, 2.7 x lo-' mol dmP3 solution of C70 at 280 and 960 nm, respectively.From the build up of the transient absorptions (note the two differ- ent timescales) a bimolecular rate constant of k,, = 1.8 x lo9 dm3 mol-' s-l is calculated for the addition process. The trace at 280 nm shows not only the formation but also the decay of [C,,-CH,C(CH3),0H]'. As in the case of (C7,-CH,CH,C1)', the decay is accelerated upon increasing the dose and fullerene concentration. This suggests second- ary processes in analogy with reaction (13). From 11,, % 320 ps for the almost exponential decay of the 280 nm signal, a bimolecular rate constant of 8 x lo7 dm3 mol- s-l is derived, attributable to a possible [C70-CH,C(CH3)20H]-[C70]' dirner radical formation.However, a competing possible radical-radical dimerisation introduces a relatively large error limit. Conclusion In conclusion, C70 undergoes practically the same types of radical reactions as c60, i.e. it is prone to reduction, oxida- tion and radical addition processes. Furthermore, the associ- ated reaction kinetics and mechanistic features resemble each other very closely. Optical measurements also revealed simi- larities, particularly with respect to the seemingly character- istic bands in the near-IR region. However, from a 1396 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 quantitative point of view, the C,,-related IR bands appear to be considerably weaker than those for the corresponding C,, species.15 16 17 D. Dubois, K. M. Kadish, S. Flanagan and L. J. Wilson, J. Am. Chem. SOC., 1991,113,7773. J. R. Morton, K. F. Preston, P. J. Krusic, S. A. Hill and E. Was- serman, J. Phys. Chem., 1992, %, 3576; 5454. P. J. Krusic, E. Wasserman, B. A. Parkinson, B. Malone, E. R. References Holler, J. R. Keizer and K. F. Preston, J. Am. Chem. SOC.,1991, 113,6274. 1 2 3 W. Kratschmer, L. D. Lamb, K. Fostiropoulos and D. R. Huff- mann, Nature (London), 1990,347, 354. H. W. Kroto, Science, 1988,242, 1139. H. W. Kroto, A. W. Allafs and S. P. Balm, Chem. Rev., 1991,91, 1213. 18 19 20 P. J. Krusic, D. C. Roe, E. Johnston, J. R. Morton and K. F. Preston, J. Phys. Chem., 1993,97, 1736. D. M. Guldi, H. Hungerbuhler, E. Janata and K-D. Asmus, J. Chem. SOC.,Chem. Commun., 1993,84.D. M. Guldi, H. Hungerbuhler, E. Janata and K-D. Asmus, J. 4 5 6 H. W. Kroto, Angew. Chem., 1992, 104, 113; Angew. Chem., Int. Ed. Engl., 1992,31, 111. J. F. Stoddart, Angew. Chem., 1991, 103, 71; Angew. Chem., Int. Ed. Engl., 1991,30, 70. F. Diederich and R. L. Whetten, Angew. Chem., 1991, 103, 695; Angew. Chem., Int. Ed. Engl., 1991,30, 678. 21 22 23 Phys. Chem., 1993,97,11258. N. M. Dimitrijevic, Chem. Phys. Lett., 1992, 194, 457. N. M. Dimitrijevic, P. V. Kamat and R. W. Fessenden, J. Phys. Chem., 1993,97,615. H. Hou, H. Luo, Z. Liu, D. Mao, Q. Qin, Z. Lian, S. Yao, W. Wang and N. Lin, Chem. Phys. Lett., 1993,203,555. 7 8 H. Hopf, Angew. Chem., 1991, 103, 1137; Angew. Chem., Znt. Ed. Engl., 1991,30, 11 17. H. W. Kroto, J. Chem. SOC., Faraday Trans., 1991,87,2871. 24 25 Online buckyball reference data, e-mail: ‘bucky@soll.lrsm.upenn.edu’. T. Kato, T. Kodama, T. Shida, T. Nakagawa, Y. Matsui, S. 9 10 R. E. Smalley, Science, 1991,31, 22. R. F. Curl and R. E. Smalley, Spektr. Wiss., 1991, 12, 88. Suzuki, H. Shiromaru, K. Yamauchi and Y. Achiba, Chem. Phys. Lett., 1991, 180,446. 11 C. Jehoulet, A. J. Bard and F. Wudl J. Am. Chem. SOC., 1991, 26 K-D. Asmus, Methods Enzymol., 1984,105, 167. 12 13 113, 5456. D. Dubois, K. M. Kadish, S. Flanagan, R. E. Haufler, L. P. F. Chibante and L. J. Wilson, J. Am. Chem. SOC.,1991,113,4364. Q. Xie, E. Perez-Corder0 and L. Echegoyen, J. Am. Chem. SOC., 1992,114,3978. 27 28 29 30 E. Janata, Radiat. Phys. Chem., 1992,40,437. P. Wardman, Phys. Chem. Re5 Data, 1989,18,1637. H. Hase and Y. Miyatake, Chem. Phys. Lett., 1993,215,141. S. Nonell, J. W. Arbogast and C. S. Foote, J. Phys. Chem., 1992, 96,4169, 14 R. E. Haufler, J. Conceicao, L. P. F. Chibante, Y. Chai, N. E. Byrne, S. Flanagan, M. M. Haley, S. C. O’Brien, C. Pan, Z. Xiao, W. E. Billups, M. A. Ciufloini, R. H. Hauge, J. L. Margrave, L. J. Wilson, R. F. Curl and R. E. Smalley, J. Phys. Chem., 1990, 94, 8634. Paper 3/07306H ; Received 10th December, 1993

ISSN:0956-5000    

DOI:10.1039/FT9949001391

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(10), 1397-1403 Electronic Transitions in Metallocenes by Resonance Raman Scattering Part 1.-Analysis of the Ferrocene Spectrum in the Visible Region Maria Letizia Ceccarani, Paola Sassi and Rosario Sergio Cataliotti Dipartimento di Chimica, Laboratorio di Chimica Fisica , Universita di Perugia, Via Eke di Sotto 8, 1-06100,Perugia, Italy Resonance Raman (RR) spectra of ferrocene in three different solvents have been measured and the excitation profiles for two totally symmetric molecular vibrations derived. These latter have also been reproduced by calculations together with the visible absorption spectrum. The model we used in these calculations, based on the non-adiabatic vibronic-coupling approach, is presented and discussed.As compared with previous publications, the results lead to a slightly different assignment of the electronic transitions in the low-energy region of the visible absorption. Ferrocene is considered the prototype molecule among the sandwich organometallic compounds. Its electronic eigen- states have been subjected to extensive theoretical studies and MO calculations of the energy levels have led to various dif- ferent ordering in the HOMO-LUMO pattern.'-14 This fact produces important consequences for the correct assignment of the dipole-allowed electronic transitions in the low-energy region of the electronic absorption spectrum and different assignments have indeed been proposed for the visible bands with wavenumber 25000 to 18000. As long ago as 1978, Gordon and WarrenI4 questioned previous assign-ments."-13 Although several papers have been published in the last few years on the theoretical treatment of the elec- tronic states of ferrocene and other metallocenes,' 5-25 no new interpretation of the low-energy transitions in the visible spectra of this molecule has been proposed since Gordon and Warren's The RR effect has been widely demonstrated to be a powerful tool for investigating the excited electronic states of molecules, since only a few vibrational modes, vibronically coupled with a particular electronic transition, may be strongly intensified by the resonance phenomenon when these motions lead to the intermediate state being displaced in the equilibrium nuclear configuration, with respect to the ground state.In these conditions, particularly noticeable during totally symmetric molecular vibrations, a considerable number of the molecular vibrations of a large molcule may be spectroscopically silent whilst those involving the chromo- phoric part will be enhanced. Metallocenes are appealing molecules from this point of view in that MO electronic wavefunctions are obtained by combining the pn wavefunctions of the cyclopentadiene (Cp) rings with the d-orbital wavefunctions of the metal atom. Therefore, the mixed character of the resulting electronic states may belong, to a greater or lesser extent, to one of the two partners. How much this situation is present in the energy levels producing an allowed electronic transition can, in principle, be discovered by simulating the RR effect and by following intensity variations (excitation profiles, hereafter referred to as REPs) of different vibrational motions which are essentially motions of the Cp rings or motions involving the metal-ligand bonds. In particular, in the ferrocene electronic spectrum recorded in a solution of non-polar solvents, there are present, in the visible region, a medium intensity band at 22 700 cm- and a shoulder around 19400 cm-'.The first has been assigned to a forbidden d-d transition localized on the iron atom and seems to be due to a degenerate transition because, at liquid nitrogen temperature, it splits into two absorptions at 24000 and 21 800 ~rn-'.'~,~~Because of its low intensity, the shoul- der has been considered as a singlet-triplet transition.'*I2 It appears quite surprising that a forbidden transition, which mainly involves atomic d orbitals of iron, could give a medium-intensity absorption in the low-energy region of the electronic spectrum of ferrocene. Therefore, with the aim of obtaining more detailed information on the nature of the two above-mentioned electronic transitions of this molecule, and in the framework of an extensive study we are presently con- ducting on the electronic spectra of various metall~cenes,~' we have carried out an RR study of the compound dissolved in benzene, carbon disulfide and cyclohexane. The analysis considers two totally symmetric motions of the molecule, namely the Fe-Cp stretching (04)and the Cp ring-breathing (~0~).We shall present and discuss also the model we used to reproduce the electronic absorption and the experimental REPs.This is, to the best of our knowledge, the first analysis of the RR spectrum of a metallocene, the first Raman study reported by Lippincot and Nelson on ferrocene28 and those by other author^,^^-^^ refer to the ground electronic state of the molecule. Experimental RR spectra of organometallic compounds are not easy to measure because of their poor stability under laser irradia- tion with concomitant energy absorption. In particular, fer- rocene is quite rapidly oxidized to the ferricinium ion and the process is certainly speeded up by light and by the presence of a solvent.Much care was therefore employed in handling the ferrocene solution and conducting the RR measurements. Ferrocene was a high-purity (>98% Fe) sample from Fluka, which was stored at low temperature and used without further purification. Cyclohexane, carbon disulfide and benzene were spectroscopic grade solvents from Carlo Erba, Italy, and were subjected to prolonged flushing with dry nitrogen to remove any dissolved oxygen. mol dm-3 solutions of ferrocene were prepared in a dry box under an inert atmosphere, obtained by nitrogen saturation, and stored in the dark at below room temperature. All trans- fers from the mother solutions to the measuring couvettes were carried out under nitrogen and using de-aerated flasks.The visible absorption spectra were recorded by a Cary 17, UV-VIS-NIR spectrometer and a Perkin-Elmer model A5 UV-VIS system. The Raman spectra were measured with a Jobin-Yvon model HG 2s double monochromator, having 1 m focal length holographic gratings and photon-counting detection. The exciting source was a Coherent model Innova 400/10 Argon ion laser, used in a single line excitation mode in the range 19440-21 840 cm-’. At lower energies, use was made of a Coumarin 6 dye laser (Spectra Physics model 375) which allows one to get more experimental points near the maximum of the 306 cm-l mode excitation profile. The power focused on the samples was always less than 150 mW, to avoid sample decomposition; use was also made of a Jobin-Yvon spinning cell, to dissipate the impinging laser energy.Exciting radiation was sent near the extreme edge of the cell to minimize the reabsorption of the scattered Raman signals by the absorbing samples. Correction for the amount of the reabsorption was made at each exciting wavelength, by considering the absorbance values extracted from the visible absorption spectra of solutions having the same concentra- tion. The scattered Raman photons were detected by a ther-moelectrically cooled Hamamatsu model 943XX phototube which, through a photon-counting acquisition board system, is computer controlled with the Jobin-Yvon ‘Prisma’ package, which allows automatic handling of the experiment. For each excitation wavelength, 10 acquisitions were carried out and results then averaged.We changed the solu- tion under laser irradiation in the measuring cell each time the laser-line frequency was changed. Results Survey absorption spectra of ferrocene in the visible region, dissolved in the three solvents are presented in Fig. 1. It is possible to note the presence of a shoulder around 19500 cm-’ in the benzene solution spectrum; the shape of the absortion spectrum in carbon disulfide is quite different from those in the other two solvents, but this is only an artifact due to mismatching in the optical path of cuvettes and strong absorption of the relevant solvent in this region. Despite this, the vibrational frequencies of the Raman lines and the shape of the relevant REPS are not influenced by their different environments in the three solvents used.Fig. 2 shows the visible spectrum of ferrocene, recorded at room temperature and at 140 K in methylcyclohexane. Whereas the room-temperature spectrum has the shape of a Gaussian symmetric band peaking at 22700 cm-’ (we have observed the same situation in the other solvents we used), the low-temperature I I 20 000 25 000 ’ 30 000 wavenumber/cm- Fig. 1 Electronic absorption spectra of ferrocene (1.25 x mol dm -3, in :(-) benzene, (-* -) cyclohexane and (---) carbon dis- ulfide J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I 1 I I I 1.o 0.8 Q1 0.6 m X 0.4 m 0.2 0.0 17500 20000 22500 25000 wavenum ber/cm -’ Fig.2 Visible absorption spectrum of ferrocene in methyl-cyclohexane in the region 17500 to 27500 cm-’; (-) room tem- perature; (---) 140K spectrum looks like an asymmetric absorption in which the unique maximum at 22 700 cm-’ seems to be resolved into an intense absorption at 24 000 cm-’and a shoulder around 21800 cm-’. Such a situation, which is similar to that observed by previous authors,’3.26 is assumed to result from a separation of the degenerate El, state into two sublevels. Fig. 3-5 show the excitation profiles of the two totally sym- metric vibrational motions of the molecule which showed the I ----._____------.. ....-_.-._-.-____..._ 0.02 --1 -~p~_-l_ I --L-~ ~~ -! 15000 17500 20000 22500 25000 27500 30000 wavenu mber/cm -’ Fig.3 Excitation profiles of the (W) 306 cm-’ and (A) 1105 cm-’ Raman lines of ferrocene 2.5 x mol dm-3 in benzene. The fitting curves are those deriving from best-fit analysis. The calculated visible absorption spectrum, in rescaled absorbance values, is also shown. 0.38 h 0.29 3 4 v >. c.-;0.2c 4-.-0.11 15000 17500 20000 22.500 25000 27500 30000 wavenumber/cm-’ Fig. 4 As Fig. 3 but in cyclohexane solution .: J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 II 15000 17500 20000 22500 25000 27500 30000 wavenumber/cm-' Fig. 5 As Fig. 3 but in carbon disulfide solution greatest intensity amplification by tuning the laser excitation energy. In Table 1 we collect all the spectral data of impor- tance for the discussion of the results.The results of our experiments clearly indicate that the two totally symmetric motions of ferrocene that we followed through their RR excitation profiles behave in a completely different way. The intensity amplification of the Cp ring- breathing mode at 1105 cm-' is reached when the excitation is tuned over the tail of the medium-intensity absorption peaking at 22 700 cm- in the visible spectrum of the mol- ecule. The mode intensity drops dramatically by exciting the Raman effect with the intense lines of the Ar' laser at 20500 cm-' and 19440 cm-'. The behaviour of the Fe-Cp sym-metric stretching at 306 cm-' is reversed, in that we do not observe amplification near the absorption at 22 700 cm- ', but the intensity of the excitation profiles in the three sol- vents used is enhanced when we irradiate the sample near 19400 cm- ',where the weak singlet-triplet transition of the molecule is foreseen to occur.The orthogonal nature of the normal coordinates of the two motions could also give some indication of the polariza- tion direction of the dipole-transition vector associated with the relevant vibronically coupled electronic transitions. In fact, the Fe-Cp motion is directed along the 2 axis of the molecule (coincident with the five-fold axis), whilst the breathing motions of the cyclopentadiene ring are parallel to the XY plane of the molecule, where the iron atom lies. If one applies group theoretical rules of the direct product to the relevant wavefunctions, whose nature is described by correct MO calculations, it is possible to derive a correct assignment of these low-frequency transitions, on the basis of a true D5d symmetry assumed for the molecule in the staggered configu- ration, or considering distortion of this symmetry by effects due to solute-solvent interactions.These aspects will be treated in the Discussion section. We have tried to model our experimental results by means of appropriate calculations, to see if either the visible absorp- tion spectrum or the excitation profiles could be reproduced to some extent. In doing so, we carried out a mono-mode calculation of the REPs of the two totally symmetric motions; this choice was made because of their orthogonal directions of vibration and also their connection to molecular vibrations having different localizations of oscillating bonds in the compound. Moreover, they are separated by ca.lo3 wavenumbers so that vibrational coupling also seems to be irrelevant from a mechanical point of view. Integrated band areas (corrected by the intensity of the solvent lines, used as an internal standard, and by the v4 factors) were used as experimental points to obtain the REPs in Fig. 3-5. A description of the model and details of the equations used for our model calculations can be found in the following section and in other papers from this labor- The calculated intensities were scaled by an appropriate factor to simulate experimental points.The Model In the description on the DSd symmetry the spin-allowed electronic transition at ca. 22 700 cm- has been essentially regarded as the Elg+Alg transition between the HOMO (So) and LUMO (S,) states of ferrocene. The low-energy onset of the absorption spectrum is thought to be due to a spin-forbidden electronic transition from So to a triplet state of the molecule (T,) connected, by spin-orbit coupling, with the higher singlet state (S,) which governs the electronic tran- sition around 30800 cm-' and has a medium-intensity tran- sition moment. Even if the T, tSo transition is symmetry forbidden, luminescence spectra' show evidence of intensity borrowing from the absorption at 30800 cm-', in that the phosphorescence from TIincreases its yield by irradiating at this last frequency, and completely disappears when lower exciting frequencies are used.For the description of the absorption and Raman-scattering cross-sections, we treated the o3and o4 modes separately, essentially because of their different behaviour with respect to the resonant enhancement and because of the different nature of the transitions to which each mode is coupled. In the calculations we handled the intermediate resonant state in a different manner. For the ring-breathing mode at 1105 cm-', whose REP follows the shape of the electronic absorption at 22700 cm-', we used two excited electronic states, assuming that solvent-induced interactions can par-tially remove the degeneration of the excited state as does low temperature.This fact seems to be possible in that the experimental absorption shows a tendency to be resolved into two bands at low temperature (see Fig. 2 and ref. 13 and 26). Table 1 Spectral data of the visible and Raman spectra of fetrocene solutions Raman visible wj (ring breathing) w4 (CpFe stretch) solvent wavenumber /cm - absorbance" wavenumber /cm - intensity /counts s - dep. rat.' wavenumber /cm - intensity /counts s-l dep. rat.c cyclohexane benzene 22753 0.123 0.127 sh. 1105 1104 3251 3106 ca. 0 ca. 0 305 305 2326 2510 0.183 0.185 carbon disulfide 22 831 0.104 1105 1840 ca. 0 306 1764 0.161 a Absorbance value at 1.25 x lop3mol dmP3. Measured with 514.5nm exciting line, power 1 W.'Depolarization ratio. The Fe-Cp stretching at 306 cm-', having an REP with a maximum around 19 400 cm- ', was modelled by considering interactions with only one excited electronic state, assumed to be a triplet configurational state, where spin-orbit coup-ling with the allowed transition at 30 800 cm- ' permits inten- sity borrowing. o3Vibration at 1105 cm-' In order to reproduce the experimental REP of the 1105 cm-' Raman line of ferrocene, it was necessary to consider the 22700 cm-' transition to be split into two absorptions arising from the ground state, correlated with two near-lying levels subject to vibronic coupling. The absorption spectrum in this region is structureless and quite broad; this could suggest that it has contributions from several vibrations. However, when moving from 19400 cm-' to the blue, the only resonance-enhanced Raman line is 0,;moreover, the symmetry-forbidden transition at 19 400 cm- ' has such a low intensity that it could not possibly be the cause of the broadening of the 22 700 cm-band.Such a situation could be described using a non-adiabatic or an adiabatic basis set; the choice would depend on the relative order of magnitude of the adiabatic and non-adiabatic coupling terms. Considering that the origin of the two states is from a doubly degenerate manifold, we preferred to adopt a non-adiabatic basis set, applying the model of Gregory and co-~orkers~~*~' in order to solve the vibronic- coupling problem by a numerical method.The electronic states involved in the Raman process, peaked at 21 800 and 24000 cm-', are described in zero- order approximation by the non-adiabatic wavefunctions @: and a):, respectively. The non-adiabatic potentials are har- monic with their adiabatic operator linear in Q: E,O(Q) = Eo + B, Q + in.' Q2 (W unm = 42 (2) where Q is the normal coordinate relative to the 0,mode and Q, and R, are the frequencies of the same mode in the rnth and nth state, respectively; we assume that R, = R,. E, = E,(O) -E,(O) where we set E,(O) = 0 for convenience; A is the vibronic coupling parameter; B, and B, are the dis- placement parameters in the rnth and nth state, respectively; U,, is the term of adiabatic coupling. The strength of vibro-nic coupling is given in terms of the ratio J = A/Eo (A/E, in ref.36). Even if it would seem more reasonable to refer to a pseudo-Jahn-Teller system, we believe that this case is at the limit of intermediate coupling. This assumption is justified by the fact that the energy gap between the two states is quite high (on the same order of a vibrational quantum); moreover, the lack of dispersion of the depolarization ratio suggests that no high values of the vibronic-coupling parameter should be used. The lowering of symmetry could be due to interactions of the single molecule with its microenvironment; the resolution of the band itself is solvent dependent. We can also assume that this kind of interaction is such that the perturbation is not in the energy levels only (through the A value of the damping constant) but also in the geometry of the molecule.Many authors claimed that, even if the equilibrium configu- ration of the molecule is that with staggered rings (D5& the barrier to internal rotation is only 0.9 0.3 kcal mol-' for the free molecule.38 This means that, even at room tem-perature, a significant number of molecules can show any instantaneous conformation among eclipsed, staggered or J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 intermediate, thus leading to a distribution of D,,, D,, and D ,~ymrnetries.~2,3*593 In this two-state approach, the wavefunction describing the excited states taking part in the Raman event has the follow- ing expression : Y = + a):<: (3) where the vibrational components C,,, and c, are solutions of the following Schrodinger equation : (4) In such a situation the scattering tensor has the form: (5) Excitation profile O(v) of the 1105 cm-' mode can be calculated with the use of the standard f~rmula,~' where, for brevity, only the dependence on the invariant terms is re- ported as : O(v) -loco+ 7c2 (6) The aij are defined by the equation4' (7) where i,fand e are the wavefunctions of the initial, final and intermediate states involved in the Raman process, 2zv1 is the incident photon energy, re is the damping constant resulting from the sum of homogeneous and inhomogeneous broadening, M, and M, are the components of the electric dipole operator. o,Vibration at 306 cm-' Since the excitation profile of this mode shows a resonance enhancement when the excitation wavelength increases, we previously suggested a vibronic coupling of the 306 cm-' vibration with only the 19 400 cm- ' electronic transition.Consequently, the scattering tensor of m4 should have only one diagonal element different from zero, which leads to a depolarization ratio of 0.33. If we look at the analogous vibration, o4, in the nickelocene molecule, the depolarization ratio is exactly 0.33.27*41 However, in these measurements on ferrocene, for this mode we observed p x 0.2. Even if the main reason for this low value of pis most probably from the contribution arising from two electronic transitions with per- pendicular polarization, we want to give a better description of the behaviour of the 306 cm- vibration which would take into account the intensity borrowing of the triplet (T,) elec- tronic state from the singlet (S,) state.Even though the spectroscopic properties of this molecule have been extensively studied, no clear evidence has been found for the presence of the 19400 cm-' transition in the visible absorption spectra of ferrocene; this has led to a lack of information about the nature of the transition itself. We would like to hypothesize a symmetry assignment of the T, state, taking into account the role of this excited level as the intermediate state of the RR event. The numerical method we adopted to reproduce the excita- tion profile of the 306 cm-' motion is essentially the same as that used for w3, even though the interpretation of the physics of the scattering process involving o4is quite differ- ent.The non-adiabatic approach is still valid but its use refers to the description of a weak vibronic-coupling system. In fact, J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 the high value of the electronic energy gap, rather than the vibronic-coupling parameter which has been assumed to have the same value as the 1105 cm-' mode, is responsible for the strength of interaction between the electronic excited states, through the value of the J parameter (A/Eo). In this case @: and 4); are the non-adiabatic wavefunctions [see eqn. (3)], in zero-order approximation, describing T, and higher singlet (S,) states respectively, and Q stands for the normal coordinate of the 306 cm-' vibration [see eqn.(la), (lb) and (2)]. Discussion Assuming that ferrocene retains its D,, symmetry in solution, as in the solid state, one should observe several vibrational motions in the Raman effect, since the a,, , el, and e,, modes give rise to polarizability tensor variations. If, however, inter- actions with the solvent or increased internal freedom of motion produces symmetry lowering, then even more Raman lines would be observed in its spectrum. In conditions of res- onance with electronic transitions only two Raman lines are seen, these being the a,, modes of the Cp ring-breathing and the Fe-Cp stretching.Vibrational Behaviour of Ferrocene in the RR Effect The most recent Raman spectra reported for ferrocene are those of Hyams3, and of Sorai et both concerning single-crystal measurements. After the work carried out by Bodenheimer et who put the earlier assignments into ~rder,,~.~~the 1105 cm- ' and the 306 cm- ' Raman lines are definitely considered a,, motions and assigned as referred to many times above. Since our measurements are performed in solution, all molecular orientations are of course probable, but there is no doubt that the coupling of the nuclear motions with the elec- tronic motions, as happens in the RR effect, should produce a polarization of the transition moment in a definite direction, which is the same for both types of motion.There is experimental evidence that the low-intensity visible absorption around 19400 cm-', appearing as a shoul- der of the 22700 cm-' band, can be considered a spin-forbidden (singlet-triplet) transition. This is, in fact, depressed when triplet-state quenchers, such as I, or other systems, are present in the same environment as ferrocene.12 On the other hand, it has been reported13 that, just in this region, the high- resolution electronic spectrum shows a vibrational band system typical of a progression involving a vibrational quantum of ca. 300 wavenumbers. Our results clearly show that the symmetric Fe-Cp stretching at 306 cm-' is reson- ance enhanced in the region of the shoulder at 19400 cm-' so that a strong vibronic coupling seems to be active with reasonably high displacement parameters.The fact that the depolarization ratio of the 306 cm-' mode is different from the value which is expected for a mode of a,, type, being ca. 0.18, is in our opinion a consequence of the fact that this mode is vibrationally mixed with two elec- tronic transitions with perpendicular polarizations, which contribute to the scattering in this region.28 The behaviour of the ring-breathing mode at 1105 cm-' follows, instead, the shape of the visible absorption at 22700 cm-' (at least in the region in which our measurements were carried out), as indicated by our REPs. The lowest-energy spin-allowed electronic transition, assuming the MO scheme by Gordon and Warren,14 should be El,+ A,,, where the states involved have a definite molecular nature.The strong intensification of the 1105 cm- mode in the Franck-Condon region, different from the 0-0 or 0-1 transitions, indicates that appreciable displacement parameters should be taken into consideration, as our calcu- lations also demonstrate, and means that the excited electron involved is not non-bonding in nature and is not localized in the d-orbitals of iron as reported in ref. 11, 12 and 14. The polarizability variation due to this mode produces appre- ciable isotropic tensor element perturbations and only small contributions from the off-diagonal elements. This produces very low values of the depolarization ratio, showing that the vibronic transition moment has a symmetry which does not modify the totally symmetric nature of the nuclear oscil- lations with the two polarizations present in the 22700 cm-' transition.Electronic Transitions of the Ferrocene Visible Spectrum As described in the previous sub-section, we consider the shoulder at 19400 cm-' to be due to a spin-forbidden tran- sition from the ground electronic state of A,, symmetry to a triplet state which is connected to that producing the 30800 cm-band, but with a different spin configuration. The near 22 700 cm- ' transition is also a singlet-singlet transition. It is commonly observed, especially in large molecules, that when two different spin configurations are described, the triplet state usually has lower zero-point energy than that of the singlet state (Hund's rule), and therefore, if the spin forbidden transition takes place, its frequency will be lower than that of the singlet-singlet allowed transition.This is the situation observed here and we must then envisage the mechanisms producing such an increase of intensity for the spin-forbidden transition. These are in our opinion the following: (i) the proximity, in terms of energy, of the spin-allowed transition, with the consequence of remarkable momentum transfer as regards the electric-dipole variation ; (ii) interaction with the solvent (most experimental data so far available have sug- gested that the emitting state can correspond to a forbidden state, when solvent-induced interactions determine mecha- nisms of symmetry lowering); (iii) vibronic coupling with vibrations of the same symmetry and polarization direction, such as the Fe-Cp stretching mode at 306 cm-',whose REPs show pronounced enhancement in this spectral range.Of these three mechanisms, we consider the second to be the most important, especially in benzene solution where the solvent-molecule interactions can easily take place via pn electronic interactions between rings. The solvent-induced effect on symmetry considered here is based on the assump- tion that the electronic wavefunctions mi, @:, . . . etc. of the isolated molecule can be strongly mixed by environmental interaction. The result of such a mixing will be that the elec- tronic states of the molecule in solution can be approximated by @; = xi Cji@o,where the coefficients of the summation reflect the extent of the solvent-molecule interaction.In other words, when the molecule is placed in an anisotropic environment, one should expect that the solvent will modify the electronic structure of the two states correlated by the electronic transition, so that the spin-forbidden transition can borrow intensity from the allowed transition through the solvent-molecule interactions. For the description of the El, t A,, transition and its pos- sible correct assignment, we put forward the following con- siderations : By comparing the intensities of the visible absorption bands at 22700 cm-' and 30800 cm-', it is difficult to accept the idea that the 22 700 cm-' band can be assigned to a d-d forbidden transition localized on the iron atom.14 In fact, its transition dipole moment is at least three times higher than that of the 30800 cm-' band; as a consequence it is quite intense with respect to what we observe in the 1402 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Parameters used in the calculation 0.0711 I I 1 I I parameter 306 cm - mode 1105 cm - mode AEn 11500 lo00 nv) Bb 0.6' 0.4' 0.4.:0.054 Ire 800 700 3 Af 500 500 Cig 20" 10" Ih 0.5 0.5 I/Eoi 41 ca. 0.55 a Energy difference/cm -'. Displacement parameter. 'Parameter referred to the lower electronic excited state. 'Parameter referred to the upper electronic excited state.Homogeneous broadenindcm- '. __.-J Inhomogeneous broadening/cm -Angle between transition 0.0201-. ''I I I I I dipole moments. Vibronic-coupling parameter. J parameter giving 15000 17500 20000 22500 25000 27500 30000 the strength of vibronic coupling. other 3d" metall~cenes.'~ We therefore attribute the 22 700 cm-' band to the E,,(LUMO) tA,,(HOMO) transition where, due to solvent-induced (or also temperature-dependent26) symmetry lowering, the doubly degenerate excited state can be separated into two states with little differ- ence in energy level. It is difficult, without any other supporting experimental data, to make any assumptions about the nature of the two resulting states, apart from their energy separation.From our experimental results, it seems, however, possible to affirm, with some confidence, that the 22 700 cm- ' transition is pol- arized in the XY plane so that the molecular symmetry lowering should lead to the maintenance of at least a plane as symmetry element. In fact, vibronic coupling solely with the ring-breathing mode, demonstrated by the fact that the REPs of the 2 polarized Fe-Cp stretching peak essentially over the shoulder at 19400 cm-', lends support to a transition in which the electric dipole variation has components in the XY plane. This situation has been confirmed through our model cal- culations, which lead both to a reproduction of the shape of the visible absorption spectrum and a good fit of the REPs of the two totally symmetric vibrational motions of ferrocene (see Fig.3-5). In Table 2 the parameters used to calculate excitation pro- files of co3 and w4 motion are listed. The o3mode is assumed to be displaced in both the first and the second excited states, with the same displacement parameter, so that the difference between the two potential-energy surfaces is only in the value of the equilibrium configuration energy. With an energy gap established around 2200 cm-' (as seems to be shown from the low-temperature resolution of the absorption band at 22700 cm-I), we were unable to obtain simultaneously a good fit of both the absorption spec- trum and the REP of 03.Since our main goal was to fit the experimental data obtained at room temperature, we con-sider lo00 cm-' to be a reasonable value for the energy gap E, between the two sublevels connected with the separation of the El, state.Once we fixed this E, figure, the calculations were performed by using different values of the vibronic- coupling parameter in the situation described, as a case of intermediate coupling. As shown in Fig. 6, the greatest number of experimental points are fitted with the curve in which assumes the value of 0.5, thus giving a A/E, value ~0.55(E, is given in !2 units). This trial and error procedure led to the best fit adopting high values of the damping constant, for both the homoge- neous contribution (r)and the inhomogeneous contribution (A), when 2 is fixed at 0.5. In Fig. 7, the results of our calculations made by varying the A value thus causing solvent-induced broadening of the wavenumber/cm-' Fig.6 REP of the o3mode (1105 cm-I) calculated using different values of vibronic-coupling parameter; (-) 1 = 0.5; (. . . .)A = 0.7; (---) 1 = 1.0. For the other parameter see Table 2. (A) Experimental points. 0.071 n$ 0.054 3 4 v > c.-0.037 c .-c 0.02c 15000 17500 20000 22500 25000 27 500 30000 wavenumber/cm-' Fig. 7 As Fig. 6 but the three fitting curves are obtained using dif- ferent values of the inhomogeneous broadening; (-) A = 500 cm-'; (. . . .) A = 300cm-'; (---) A = 100 cm-' REP are shown. Even if the differences between 500 and 300 cm-' used for the A parameter in the fitting curves are not very remarkable we can definitely see that a value of A = 100 cm-' gives the worst results.This fact is not surprising if one considers that a consistent amount of broadening in the REP should derive from inhomogeneous effects when the distribu- tion of different solute-solvent configurations can be rela- tively high, thus shortening the excited-state lifetime. In Fig. 8 the fitting curves for the w4 mode reproduced with the model described earlier are shown. Here the A value was fixed at 500 cm-' as for the o3mode and the calcu- lations were performed with a different set of A. values. Again A. = 0.5 gives the best results but, in this case (A./E,,) -g 1, as it should be in the case of weak Herzberg-Teller coupling, according to the interpretation we adopted to describe the intensity borrowing of the 19 400 cm- electronic transition from that at 30 800 cm-'.Concluding Remarks The novel aspect of the present measurements is represented by the unusual behaviour of the REPs for the two vibrational motions, which have shown resonance enhancement by tuning the excitation wavelength in different parts of the elec- tronic absorption spectrum of ferrocene. It is, in fact, com- monly observed that, when resonance occurs, the Raman lines of a molecule are intensified in the same way, even if with different shapes and cross-sections. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0.062 0.020 I 15000 17500 20000 2250025000 27500 30000 wavenumber/cm -Fig. 8 As Fig.6 for the o4mode (306 cm-'). (m) Experimental points. Our interpretation has been based on the assumption that the two electronic transitions of the low-energy visible absorption of ferrocene, viz the spin-forbidden transition at 19400 cm-' and the spin-allowed transition at 22 700 cm -*, have orthogonal polarization of their transition moments. In the former, this vector is directed along the 2 axis and there- fore strong coupling with the Fe-Cp in phase stretching can occur, while the latter transition seems to have components of this vector in the XY plane, where the ring-breathing mode is polarized. With this interpretation, we assign the 19400 cm-' shoulder to the singlet-triplet transition and the 22 700 cm -' medium-intensity band to a singlet-singlet E,, tA,, transition of the visible spectrum of ferrocene. The simple model we suggest for the interpretation of REPS of ferrocene does not describe the interactions between solvent and molecule very accurately; however, the good agreement with our experimental results seems to confirm the symmetry lowering suggested by other authors.32 References W.Mofitt, J. Am. Chem. SOC., 1954,76, 3386. J. D. Dunitz and L. E. Orgel, J. Chem. Phys., 1955, 23, 954. A. D. Liehr and C. J. Ballhausen, Acta Chem. Scand., 1957, 11, 207. F. A. Matsen, J. Am. Chem. SOC., 1959,81,2023. R. E. Robertson and H. M. McConnell, J. Phys. Chem., 1960,64, 70. E. M. Shustorovich and M. E. Dyatkina, J. Struct. Chem. USSR, 1960, I, 98.J. P. Dahl and C. J. Ballhausen, Mat. Fys. Medd. Kgl. Danske Vid. Selsk, 1961, 33, 5. R. D. Fischer, Theoret. Chim. Acta, 1963, 1,418. I403 9 (a) A. T. Armstrong, F. Smith, E. Elder and S. P. McGlynn, J. Chem. Phys., 1967, 46, 4321; (b)A. T. Armstrong, D. G. Carroll and S. P. McGlynn, J. Chem. Phys., 1967,47, 1104. 10 J. H. Schachtschneider, R. Prins and P. Ros, Inorg. Chim. Acta, 1967, 1, 462. 11 R. Prins and J. D. W. van Voorst, J. Chem. Phys., 1968,49,4665. 12 D. R. Scott and R. S. Becker, J. Chem. Phys., 1961,35,576. 13 Y. S. Sohn, D. N. Hendrickson and H. P. Gray, J. Am. Chem. Soc., 1971, 93, 3603. 14 K. R. Gordon and K. D. Warren, Inorg. Chem., 1978,17,987. 15 C. Cauletti, J. C. Green, M. R. Kelly, P. Powell, J. van Tiborg, J.Robbins and J. Smart, J. Electron Spectrosc. Relat. Phenom., 1980, 19, 327. 16 J. C. Green, Struct. Bonding (Berlin), 1981, 43, 37. 17 M. C. Bohm, Z. Naturforsch A, 1982,37, 1193. 18 T. Vondrak, J. Organomet. Chem., 1986,306, 89. 19 D. C. Driscol, P. A. Dowben, N. M. Boag, M. Grade and S. Barfuss, J. Chem. Phys., 1986,85,4802. 20 N. Rosch and H. Jorg, J. Chem. Phys., 1986,84, 10. 21 D. L. Lichtemberg and A. S. Copenhaver, J. Chem. Phys., 1989, 91, 663. 22 G. Cooper, J. C. Green and M. P. Payne, Mof. Phys., 1988, 63, 1031. 23 G. A. von Wald and J. W. Taylor, J. Electron Spectrosc. Relat. Phenom., 1988,47,3 15. 24 D. O'Hare, J. C. Green, T. P. Chadwick and J. S. Miller, Organometallics, 1988, 7, 1335. 25 M. Ohno and W. von Niessen, Chem. Phys., 1991, 158, 1. 26 L. D. Dave, D. F. Evans and G. Wilkinson, ref. in ref. 12. 27 P. Sassi, M. L. Ceccarani and R. S. Cataliotti, Mof. Phys., sub-mitted. 28 E. R. Lippincott and R. D. Nelson, Spectrochim. Acta, 1958, 10, 307. 29 J. Bodenheimer, E. Loewenthal and W. Low, Chem. Phys. Lett., 1969,3, 7 15. 30 R. T. Bailey, Spectrochim. Acta, Part A, 197 1, 27, 199. 31 I. J. Hyams, Chem. Phys. Lett., 1973, 18, 399. 32 M. Sorai, S. Murakawa, K. Ogasahara and H. Suge, Chem. Phys. Lett., 1980, 76,510. 33 R. S. Cataliotti, S. M. Murgia, G. Paliani, A. Poletti and M. Z. Zgierski, J. Raman Spectrosc., 1985, 16, 258. 34 R. S. Cataliotti, G. Paliani, G. Dellepiane, S. Fuso, S. Destri, L. Piseri and R. Tubino, J. Chem. Phys., 1985,82,2223. 35 P. Sassi and R. S. Cataliotti, Mol. Phys., 1992,77,937. 36 A. R. Gregory, W. H. Henneker, W. Siebrand and M. Z. Zgierski, J. Chem. Phys., 1976, 65, 2071. 37 A. R. Gregory, W. H. Henneker and W. Siebrand, J. Chem. Phys., 1977, 67, 3 175. 38 A. Halland, Top. Curr. Chem., 1975,53, 1. 39 0.S. Mortensen and S. Hassing, Ado. Infrared Raman Spectrosc., 1961,6, 429. 40 H. C. Longuet-Higgins, Adv. Spectrosc., 1961, 2,429. 41 K. Yokoyama, S. Kobinata and S. Maeda, Bull. Chem. SOC.Jpn., 1976,49,2182. 42 H. Stammreich, quoted by H. P. Fritz, in: Ado. Organomet. Chem., 1964,1,239. Paper 3/06705J; Received 8th November, 1993

ISSN:0956-5000    

DOI:10.1039/FT9949001397

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(10), 1405-1409 Rotational Dynamics in Liquid Water: A Simulation Study of Librational Motions lgor M. Svishchev" and Peter G. Kusalik Department of Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J3 The rotational dynamics in liquid water have been studied. The power spectra of the single-molecule orienta- tional autocorrelation functions (ACF) have been calculated in molecular dynamics simulations with the SPC/E potential and have been used to characterize various rotational motions of water molecules. Both ordinary and heavy water have been examined at temperatures of -10 and 25 "C. For liquid H,O at 25 "C the power spectra of the second-order (Raman) orientational ACFs contain three intense bands, centred at ca.500, ca. 560 and ca. 670 cm-', and a less intense high-frequency shoulder at ca. 820 cm-'. Two intense librational bands with maxima at ca. 570 and ca. 650cm-' are present in the power spectrum of the single-dipole orientational ACF for the same system. The average temperature coefficient for the librational frequencies of SPC/E water is found to be about -0.65 cm-' K-', which agrees well with experimental estimates. A well resolved rototranslational band centred at ca. 55 cm-' is observed in the low-frequency region of the power spectra of the single-molecule orientational ACFs. This band is relatively insensitive to temperature variations and shows no isotopic effect. 1. Introduction The dynamical structure in liquid water is often characterized as a fluctuating network of hydrogen (H-) bonded molecules'*'in which specific molecular motions, such as restricted translations and rotations (librations), give rise to a remarkably rich vibrational spectrum in the THz region.'-'' Many practically important collective phenomena in aqueous media (acoustic' '*17 etc.) involve these (low-frequency on a spectroscopic timescale) vibrational motions, and hence they have received a great deal of attention in experimental (Raman and IR)3-'0 and theoretical (computer simu-lation)"-16 studies.The low-frequency Raman yield three vibra- tional bands of librational origin centred at ca. 720, ca. 550 and cu. 450 cm- ' and two major translational bands centred at ca. 190 and ca.60 cm-'. It is now well documented that the translational band at CQ. 190 cm-' (usually referred to as an 00 stretching mode of water pentamers) shows a pro- In principle, we can also use computer simulations to study these low-frequency molecular vibrations in liquid water structure and a number of such investigations have been carried out. Impey et a!.'' have examined MCY water at several state points. In this work they calculated the power spectra arising from single-particle orientational autocorrela- tion functions"*22 and found a single broad librational peak at ca. 400-450 cm-'. More recently, extensive simulations with the TIP4P p~tential'~''~,'~ have been carried out. Again, only a single librational mode has been found at a frequency of ca. 550 cm-'.However, these results also indi- cate the presence of a low-lying dumped resonance mode at ca. 20 cm- ' in the rotational spectra of TIP4P water. It has been suggested in the work of Bertolini and Tani" that this mode may be due to rototranslational 'interactions'. Clearly, many questions still remain concerning both the structure and the physical origin of these low-lying vibra- tional modes. For example, it remains unclear as to how nounced increase in intensity with decreased temperat~re~.~ many (behaviour which is consistent with an increase in intermolec- ular H-bonding), while the band at ca. 60 cm-' (believed to be induced by 000 bending of three H-bonded molecules) is relatively insensitive to temperature variation^.'.^ In several investigations with supercooled water a weak translational band at cu. 260 cm-' has also been rep~rted.~ The interpre- tation of this band is that it arises from splitting of the major stretching mode into two components due to the difference in force constants for the vibrational motions parallel and per- pendicular to the molecule's dipole axis.In recent studies an additional relaxation mode has been detected' which appears as a weak shoulder on the elastic Rayleigh peak with a char- acteristic Raman frequency of ca. 8 cm-' at 25 "C. The struc- tural mechanism responsible for this relaxation process has yet to be clarified; it has been suggested that large rotational motions of water molecules which control the lifetime of H- bonded pentamers give rise to this peak in Raman spectra.' Data from the infrared studies generally support the Raman data in mapping librational and translational modes for liquid water.The major feature of the far-infrared spec- trum of water9.10,18,19.21 is an extremely intense broad band centred at ca. 700 cm-which overlaps at the low-frequency end with a much less intense band with a maximum at ca. 200 cm-'. There also exists some evidence for additional contributions at low frequencies to the IR absorption from the 550 and 55 cm-' band^.^*'^*'' librational modes are actually active in a non-polarizable water model. These questions can be most readily addressed through computer simulation. In this paper we present the rotational spectra for the SPC/E model of liquid water in the 10-lo00 cm- ' region in which the features due to various molecular motions are clearly resolved.These spectra have been obtained via Fourier transforms of the single-molecule orientational ACFs. We have also calculated the power spectra of the principal components of the angular velocity ACF. In order to resolve the complex spectral struc- ture in the frequency region under consideration, large storage arrays for these ACFs (thereby providing high- resolution spectra) have been exploited in rather long simula- tion runs (with production periods of ca. 800 ps). In this work two temperatures, -10 and 25"C, have been examined and both ordinary and heavy water have been studied.The remainder of this paper is organized as follows. In Section 2 we outline the methodological details of our simu- lations, and then in Section 3 we present spectra and discuss their origin. Finally, our conclusions are given in Section 4. 2. Simulation Details In this article we report results from molecular dynamics simulations of liquid water performed with the SPC/E pair- p~tential.'~The principal frame is defined in Fig. 1. As in our previous work with SPC/E water,24 where the spatial dis- tributions of molecules in the local frame were examined, we J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 tz +0.4238 e A HOH = 109.47 t A Fig. I Principal frame coordinates and geometry of the SPC/E model have taken the Z-axis to be along the dipole moment and the XZ plane to be the molecular plane.For heavy water the interaction parameters and the corresponding molecular geometry of the SPC/E model were assumed to be unchanged except for the increase of the hydrogen mass. Our MD simulations have been carried out with samples of 256 molecules at experimental densities at constant tem- peratures of -10 (H,O) and 25 "C(H20 and D20).In order to ensure that our spectra did not contain artificial peaks due to the periodic boundaries, test calculations were conducted with 108 particles at 25°C and no significant changes in the spectral structure were detected. In our calculations isother- mal conditions were maintained by means of a Gaussian thermo~tat.~~We have utilized a truncated octahedral geometry for the simulation and periodic boundary conditions (Ewald Our implementation of the Ewald method is described in detail in ref.24. A fourth-order Gear algorithm29 with a time step of 1.25 fs was used to inte- grate the isokinetic equations of motion and the rotational degrees of freedom were represented using qua tern ion^.^^ At each temperature the system was allowed to equilibrate for ca. 100 ps and averages were collected over the subsequent 800 ps. Single-molecule orientational autocorrelation functions have been employed to calculate the spectral functions: (1) where In eqn. (1) and (2),k is a positive integer, e, (a= x, y, z) are the unit vectors of the principal frame, o is the wavenumber defined as v/c (v is the frequency and c is the speed of light), and P, denotes a Legendre polynomial of order k.The Fourier transforms were carried out after the principal relax- ation modes (the long-time exponential decays) have been subtracted from the autocorrelation functions C,,,(t). We remark that the Fourier transforms of the second-order orientational ACFs C,,,(t) can be used to approximate an allowed Raman spectrum for our system, while the transform of the ACF Cl, ,(t) represents the single-dipole contribution to its allowed IR spectrum. The statistical errors in our simulated spectra were esti- mated using the procedure outlined by Allen and Tildesley;27 they are <0.5%. Large storage arrays, spanning at least two13 time-steps (ca.10 ps), have been used to accumulate ck,.(t) during simulation runs, since smaller arrays, while requiring less memory, would have resulted in a much lower resolution. 3. Simulation Results and Discussion Our major simulation results, the Fourier transforms of the single-particle orientational autocorrelation functions of SPC/E water, are presented in Fig. 2 and 3 for ordinary water and Fig. 4 and 5 for heavy water. The transforms of the second-order (Raman) orientational ACFs C2,,(t) and C2, .(t) possess very similar structure (both band shapes and posi- tions of spectral maxima) and hence for reasons of clarity we have plotted only the X and 2 contributions. The Fourier transforms of the first-order single-dipole ACF Cl, ,(t) for ordinary and heavy water are plotted, respectively, in Fig.3 and 5. To illustrate the temperature dependence in these spectra, data for ordinary water at -10 and 25 "C have been directly compared in Fig. 2 and 3, in which the upper curves are always the low-temperature result. The spectral resolution achieved can be clearly seen in Fig. 2(b), 3(b), 4(b) and 5(b),where the discrete points in frequency are explicitly shown. 5 A h -3 v N 200 300 400 500 600 700 800 900 1000 o/cm -' 0.3 B/ 0.2 h -3 v N 0.1 0.0 -~ 0 50 100 ' 150 2 w/cm - Fig. 2 Fourier transforms of the second-order (Raman) orientation- a1 ACFs C,,,(t) (a) and C,,Jt) (b) for ordinary SPC/E water in the 10-lo00 cm-' frequency region.A, High-frequency end; B, low-frequency end. The solid and the dotted lines are, respectively, the results at 25 and -10"C. In B only I,. ,(w) is shown. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 A1 ~ 0.0 200 ' 300 400 500 600 700 ' 800 900 '1000 o/cm-I0.00 I 0 50 100 ' 150 200 o/cm-' Fig. 3 Fourier transforms of the single-dipole orientational ACF C,..(t) for ordinary SPC/E water in the 10-loo0 cm-' frequency region. (a) High-frequency and (b)low-frequency end. The solid and the dotted lines are as in Fig. 2. We begin our analysis with the power spectra of the second-order orientational ACFs for liquid H20 and focus first on the high-frequency end. It can be seen in Fig. 2(a)that while the Fourier transform of C2,,(t) has a single librational peak centred at ca.500 cm-' (at 25"C), the Fourier trans- form of C,,.(t) has a rather complex structure with major maxima at 560 and 670 cm-'. We point out that all these peaks have been detected, although at slightly different fre- quencies, in experimental Raman st~dies.~~~-'~~~ In addition to these major maxima, a high-frequency shoulder at ca. 820 cm-' (at 25 "C)is also evident in the spectrum of C2,z(r). In order to relate these spectral features to specific molecu- lar motions in as direct a way as possible we have also con- sidered the power spectrum of the angular velocity autocorrelation function. In Fig. 6(a) the Fourier transforms of the three principal components of the angular velocity ACF are shown (at frequencies 100-lo00 cm-').We observe that the rotational oscillations of water molecules about their Y and 2 axes, as reflected by peaks at 500 cm-' in the power spectra of the Y and Z components of the angular velocity ACF [see Fig. qa)], give rise to a nearly symmetric maximum in the Z2,x(0) [see Fig. 2(a)].At the same time we can see that the more rapid rotations about the X-axis essen- tially do not contribute to the Z2,x(co). This fact can be explained on the basis of the cumulant expansion of the Ck,x(r),in which the leading order dependence does not contain terms due to (mx(t)cox(0)).'l We also find that, unlike I \ \\ \/ \ \ 0.0 ' 1 100 200 300 400 500 600 700 800 olcm- 0.50 B 30.25-1 I0.004 0 50 100 150 200 o/cm -Fig.4 Fourier transforms of the second-order (Raman) orientation- a1 ACFs C2,,(t) (a) and C2,,(t) (b)for heavy SPC/E water at 25 "Cin the 10-lo00 cm-' frequency region. A, High-frequency and B, low-frequency end. In B only 12.,(a)is shown. molecular rotations about the Y and 2 principal axes, rota- tional motion about the X axis appears more complex; the power spectrum of (cox(t)cox(0)) is noticeably skewed toward higher frequencies. This seems to indicate that the fine libra- tional structure in the spectra of the second-order (Raman) orientational ACFs reflects different types of oscillatory motions primarily about the principal X axis (given our defi- nition of the local frame).The librational spectrum of the single-dipole ACF Cl,,(t) for liquid H20 is shown in Fig. 3(a)and, as we might expect, its shape appears to be quite similar to that of the second- order ACF C2,z(t).Both major librational peaks, centred at ca. 570 and ca. 650 cm-', are present, although the high- frequency peak clearly displayed in 12,,(m)is not resolved in Z1,,(co). We recall that the major difference between experi- mental IR and Raman spectra at these frequencies arises largely from the fact that only the single-dipole ACF C,,,(t) contributes to the IR spectrum, while the Raman spectrum has contributions from all second-order orientational ACFs.' 1*22*30 Correspondingly, the 500 cm-' band which dominates the Raman spectrum is absent in the experimental IR data.In order to illustrate how temperature affects the libra- tional motions of molecules in SPC/E water, simulation results at -10 and 25" C have been compared directly in Fig. 2(a) and 3(a). From these data we have then estimated the average temperature coefficient, A(omaX,,ib,(H20))/AT,for 1408 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2 h \ 0.15 0.10. h 3 W c \ 0.05 0.004 I I 0 50 100 150 2"J o/cm -Fig. 5 Fourier transforms of the single-dipole orientational ACF C,,,(t)for heavy SPC/E water at 25°C in the 10-lo00 cm-' fre-quency regon. (a)High-frequency and (b)low-frequency end. the librational frequencies of SPC/E water. The value obtained by averaging over the four observed maxima in 12,Jm)and 12Jm) and the two major maxima in Il,z(m)is ca.-0.65 cm-' K-' . The available experimental estimate from IR data18 is -0.7 cm-' K-' and we conclude that we have successfully reproduced both the sign and the magni- tude of the average temperature coefficient for librational fre- quencies. As we might expect, the librational frequencies for SPC/E water exhibit a pronounced isotopic effect. The power spectra of the second-order orientational ACFs for liquid D20 at 25°C are given in Fig. 4(a) and the corresponding spectrum of the single-dipole ACF is plotted in Fig. 5(a).The average isotopic ratio for frequency maxima, (urnax,lib(D20)) /(mrnax+lib(HZO)),obtained from the librational peaks in these spectra appears to be ca.1.38, reproducing the ideal isotopic ratio for the water molecule. We now shift our focus to the low-frequency end (<200 cm-') in our spectra. As mentioned above, the experimental data indicate the presence of at least two resonance bands in this region centred around 190 and 60 cm-' whose intensity is largely determined by the many-body interaction-induced polarization effects.' 2*14-16 Our simulation results in this fre- quency range from the explicitly non-polarizable SPC/E model are shown in Fig. 2(b)-3(b) (for H20) and Fig. 4(b)-5(b) (for D,O). They clearly indicate the presence of a low- lying band centred at ca. 55 cm-' (25 "C). It can also be seen 0' '100 200 300 400 500 600 700 ' 800 ' 900 ' 1000 o/cm -"I 10 40 70 160 o/cm -' Fig.6 Power spectra of the angular velocity ACF of ordinary SCF/E water at 25 "C.A, High-frequency and B, low-frequency end. (a), (b) and (c) represent the X, Y and Z principal components, respectively. that this band is well resolved from the more intense libra- tional peaks. At the same time, we observe that the 190 cm-' band does not appear active in the power spectra of the single-molecule ACFs for SPC/E water. The low-frequency power spectra of the components of the angular velocity ACF for SPC/E water at 25 "C are shown in Fig. qb) and it can be seen that they also exhibit shallow maxima at ca. 65 cm-', most evident in the X and 2 com-ponents. Given the fact that a peak appears in the power spectrum of the linear-velocity ACF at approximately the same frequency, it becomes apparent that the band at ca.55 cm-' in the spectra of the single-molecule orientational ACFs arises from a coupled rotational-translational motion. This claim is strongly supported by our recent work3' which explicitly examines rototranslational motion in liquid water. As we might expect, the rototranslational mode at ca. 55 cm-is rather insensitive to isotopic substitution [compare Fig. 2(b)and 4(b)or Fig. 3(b)and 5(b)].We also find that the temperature coefficient, A(w,,,)/AT, for this mode is very J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 small, ca. -0.1 cm-' K-', which agrees with experimental observations. Another feature of this low-frequency rototranslational band (exhibited in our model spectra at ca.55 cm-') is that its intensity is rather insensitive to changes in temperature (this becomes clearer after the contribution due to the sym- metric librational peak at ca. 500 cm-' has been removed from the spectral data in Fig. 2), which is again consistent with experimental observation^.^.' A remarkable increase in intensity with decreased temperature is well documented in the experimental Raman spectra for another low-lying band,3b.7 at 190 cm-', and in early work it was correlated with the degree of association of H20 molecules in tetra- hedrally H-bonded fragments in liquid water Later, this temperature variation in the experimental Raman spectra was interpreted as a collective interaction-induced polarization effect which accompanies immediate structural changes in real ~ater.~~,'~,~~ It h as also been argued that the 60 cm-' band of the experimental spectra originates from an induced polarization effect.22 Our results obtained with the non-polarizable SPC/E model, however, indicate that the slow rotational dynamics of individual molecules activates the 55 cm-' mode.This mode, as shown in ref. 12-14, is enhanced in the experimental Raman and IR spectra by many-body polarization contributions. At the same time, the absence of the 190 cm-' band in our simulation data suggests that it is the interaction-induced polarization effect that activates this band in the experimental spectra and governs its temperature behaviour. Clearly, the simulations with polarizable water models are needed to understand more fully the influence of temperature on the low-frequency region in the optical spectra of liquid water.Sciortino and C~rongiu~~ have recently calculated the low-frequency optical spectra for hexagonal ice using the pol- arizable NCC potential and have compared them with the spectra obtained for the TIP4P model. Their results indicate that the polarizable model is far better at reproducing the positions of experimental bands in the spectra of the ice I,. 4. Conclusions In this article we have reported results from MD simulations of liquid water performed with the SPC/E potential. We have analysed the Fourier transforms of the single-molecule orien- tational autocorrelation functions focussing upon the fine structure of the molecular librational modes.These modes are widely used as probes of the vibrational dynamics in liquid water, yet their previous statistical-mechanical analysis has been limited because of insufficient spectral resolution in simulation data. Both ordinary and heavy SPC/E water have been studied at temperatures of -10 and 25 "C. Three major librational modes centred at ca. 500, ca. 560 and ca. 670 cm-' and a less intense high-frequency shoulder at ca. 820 cm-' have been identified in the spectra of the second-order (Raman) orientational ACFs (liquid H20 at 25 "C). Two intense librational bands with maxima at ca. 570 and ca. 650 cm-' have been found in the spectrum of the single-dipole ACF (liquid H20 and 25°C).Analysing the principal components of the angular velocity ACF we have clarified that the complex librational structure in these spectra is largely associated with the rotational dynamics of water molecules about the principal X axis. We have also found that the average temperature coefficient for the libra- tional frequencies of SPC/E water agrees well with available experimental estimates. One of the most interesting features of our simulation data was the presence of a well resolved rototranslational mode centred at ca. 55 cm- ' which is relatively insensitive to tem- perature variations. This mode has been well characterized in experimental Raman studies. We have also found that another characteristic low-lying mode, active in experimental spectra at ca.190 cm-',is absent in the spectra of the single- molecule orientational ACFs for SPC/E water. This observa- tion supports the claim that its origin in experimental optical spectra is due to interaction-induced polarization effects accompanying structural transformations in liquid water. We are grateful for the financial support of the Natural Sciences and Engineering Research Council of Canada. References 1 G. E. Walrafen, in Water-A Comprehensive Treatise, ed. F. Franks, Plenum Press, New York, 1972, vol. 1. 2 M. G. Sceats, M. Stavola and S. A. Rice, J. Chem. Phys., 1979, 70, 3927. 3 (a) G. E. Walrafen, J. Chem. Phys., 1964, 40, 3249; (6) G. E. Walrafen, M.R. Fisher, M. S. Hokmabadi and W-H. Yang, J. Chem. Phys., 1986,856970. 4 D. Eisenberg and W. Kauzmann, The Structure and Properties of Water,Oxford University Press, Oxford, 1969. 5 K. Mizoguchi, Y. Hori and Y. Tominaga, J. Chem. Phys., 1992, 97, 1961. 6 Y. Yeh, J. H. Bilgram and W. Kanzig, J. Chem. Phys., 1982, 77, 23 17. 7 S. Krishnamurthy, R. Bansil and J. Wiafe-Akenten, J. Chem. Phys., 1983, 79, 5862. 8 V. Mazzacurati, A. Nucara, M. A. Ricci, R. Ruocco and G. Signorelli,J. Chem. Phys., 1990,93, 7767. 9 0. A. Simpson, B. L. Bean and S. Perkowitz, J. Opt. SOC. Am., 1979,69, 1723. 10 P. Schiebener, J. Straub, J. M. H. Levelt Sengers and J. S. Gal-lagher, J. Phys. Chem. Ref: Data, 1990, 19,677. 11 R. W. Impey, P. A. Madden and I.R. McDonald, Mol. Phys., 1982,46, 513. 12 P. A. Madden and R. W. Impey, Chem. Phys. Lett., 1986, 123, 502. 13 R. Frattini, M. Sampoli, M. A. Ricci and G. Ruocco, Chem. Phys. Lett., 1987, 141, 297. 14 M. A. Ricci, G. Ruocco and M. Sampoli, Mol. Phys., 1989, 67, 19. 15 D. Bertolini and A. Tani, Mol. Phys., 1992, 75, 1065. 16 V. Mazzacurati, M. A. Ricci, G. Ruocco and M. Sampoli, Chem. Phys. Lett., 1989, 159, 383. 17 S. Sastry, F. Sciortino and H. E. Stanley, J. Chem. Phys., 1991, 95, 7775. 18 D. A. Draegert, N. W. B. Stone, B. Curnutte and D. Williams, J. Opt. SOC. Am., 1966,56,64. 19 C. W. Robertson, B. Curnutte and D. Williams, Mol. Phys., 1972, 20, 183. 20 J. B. Hasted, Aqueous Dielectrics, Chapman and Hall, London, 1973. 21 J. B. Hasted, S. K. Husain and F. A. M. Frescura and J. R. Birch, Chem. Phys. Lett., 1985, 159, 622. 22 P. A. Madden, in The Physics and Chemistry of Aqueous Ionic Solutions, ed. M-C. Belissinet and G. W. Neilson, Reidel, Dor- drecht, 1987, p. 181. 23 H. J C. Berendsen, J. R. Grigera and T. P. Straatsma, J. Phys. Chem., 1987,91,6269. 24 1. M. Svishchev and P. G. Kusalik, J. Chem. Phys., 1993, 99, 3049. 25 D. J. Evans and J. P. Morriss, Statistical Mechanics of Nonequi-librium Liquids, Academic Press, San Diego, 1990. 26 D. Fincham and D. M. Heyes, Adv. Chem. Phys., 1985,63,493. 27 M. P. Allen and D. J. Tildesley, Computer Simulations of Liquids, Oxford University Press, Oxford, 1987. 28 P. G. Kusalik, J. Chem. Phys., 1990,93, 3520. 29 D. J. Evans and J. P. Morriss, Comput. Phys. Rep., 1984, 1, 297. 30 P. A. Madden, in Liquids, Freezing and Glass Transition, ed. J. P. Hansen, D. Levesque and J. Jinn-Justin, North Holland, Amsterdam, 1991,p. 549. 31 I. M. Svishchev and P. G. Kusalik, Chem. Phys. Lett., 1993, 215, 596. 32 F. Sciortino and G. Corongiu, Mol. Phys., 1993, 79, 547. Paper 3/07030C; Received 10th December, 1993

ISSN:0956-5000    

DOI:10.1039/FT9949001405

出版商:RSC    

年代:1994

数据来源: RSC