摘要:

THE ROYAL SOCIETY OF CHEMISTRY Journal of the Chemical Society Faraday Transactions Scientific Editor Editorial Manager Prof. A. Robert Hillman Dr. Robert J. Parker Department of Chemistry The Royal Society of Chemistry University of Leicester Thomas Graham House University Road Science Park Leicester LEI 7RH, UK Milton Road Cambridge CB4 4WF, UK Staff Editor: Dr. R. A. Whitelock Senior Assistant Editor: Mrs. S. Shah Assistant Editors: Dr. G. F. McCann, Miss J. C. Thorn Editorial Secretary: Mrs. J. E. Gibbs Faraday Editorial Board Prof. M. N. R. Ashfold (Bristol) (Chairman) Dr. J. A. Beswick (Paris) Prof. A. R. Hillman (Leicester) Dr. D. C. Clary (Cambridge) Prof. J. Holzwarth (Berlin) Dr. L. R. Fisher (Bristol) Dr. D. Langevin (Paris) Dr.B. E. Hayden (Southampton) Dr. S. K. Scott (Leeds) Prof. J. S. Higgins (London) Dr. R. K. Thomas (Oxford) Dr. R. J. Parker (RSC, Cambridge) (Secretary) International Advisory Editorial Board R. S. Berry (Chicago) R. A. Marcus (Pasadena) A. M. Bradshaw (Berlin) Y.Marcus (Jerusalem) A. Carrington (Southampton) B. J. Orr (North Ryde) G. Cevc (Munich) 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) D. K. Russell (Auckland) R. Freeman (Cambridge) J. P. Simons (Oxford) H. L. Friedman (Stony Brook) S. Stoke (Amsterdam) H. H. J. Girault (Lausanne) J. Troe (Gottingen) H. lnokuchi (Okazaki) J. Wolfe (Kensington, NSW) J.N. lsraelachvili (Santa Barbara) C. Zannoni (Bologna) M. L. Klein (Philadelphia) R. N. Zare (Stanford) A. C. Legon (Exeter) ' A. Zecchina (Turin) 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 1 HN, 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 51400.00, 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 ;ubmission of manuscripts intended for pub- ication in two forms, Research papers and -araday Communications. These should jescribe original work of high quality in the sciences lying between chemistry, physics md biology, and particularly in the areas of 2hysical chemistry, biophysical Chemistry snd chemical physics. Research Papers Full papers contain original scientific work Nhich has not been published previously. However, work which has appeared in print n a short form such as a Faraday Communi- :ation is normally acceptable.Four copies ncluding a top copy with figures etc. should 38 sent to The Editor, Faraday Transactions, 3t the Editorial Office in Cambridge. Authors nay, if they wish, suggest the names (with sddresses) of up to three possible referees. Faraday Communications Faraday Communications contain novel Scientific work in short form and of such mportance that rapid publication is war-,anted. The total length is rigorously ,estricted to two pages of the double-:olumn A4 format. For a Communication zonsisting entirely of text and ten references, with no figures, equations or tables, this cor- ,esponds to approximately 1600 words plus 3n abstract of up to 40 words.Submission of a Faraday Communication :an 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 case, 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. Faraday Research 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. Prof. A. R. Hillman, Scientific Editor. Tel. : Leicester (01 16) 2525226 (24 hours) E-Mail (JANET):ARH7@UK.AC.LElCESTER Fax: (01 16) 2525227 Dr. R. J. Parker, Editorial Manager. Tel.: Cambridge (0223) 420066 E-Mail (INTERNET): RSCI (GRSC.ORG (For access from JANET use RSCI %,RSC.ORG@UK.AC.NSF NET- RELAY) Fax: (0223) 426017

ISSN:0956-5000    

DOI:10.1039/FT99490FX093

出版商:RSC    

年代:1994

数据来源: RSC

ISSN:0956-5000    

DOI:10.1039/FT99490BX095

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

ISSN 0956-5000 JCFTEV(24) 3591-3733 (1994) JOURNAL OF THE CHEMICAL SOCIETY Faraday Transactions Physical Chemistry & Chemical Physics CONTENTS 3591 Resonance-enhanced multiphoton ionisation spectroscopy of thiirane R. A. Morgan, P. Puyuelo, J. D. Howe, M. N. R. Ashfold, W. J. Buma, J. B. Milan and C. A. de Lange 3601 Equilibrium geometry and vibrational frequencies of the 1:2 van der Waals complexes between methyl chloride and hydrogen chloride W. A. Herrebout and B. J. van der Veken 3609 Pressure and temperature dependence of the rate constants for the association reaction of OH radicals with NO between 301 and 23 K P. Sharkey, I. R. Sims, I. W. M. Smith, P. Bocherel and B. R. Rowe 3617 Reaction of atomic oxygen with some simple alkenes. Part 1.-Low pressure studies on reactions involving ethene, propene and (Q-but-2-ene C.Anastasi and M. G. Sanderson 3625 Reaction of atomic oxygen with some simple alkenes. Part 2.-Reaction pathways involving ethene, propene and (E)-but-2-ene at atmospheric pressure. C. Anastasi, M. G. Sanderson, P. Pagsberg and A. Sillesen 3633 Excimer-laser photofragmentation of boron trichloride and its implications for laser vapour deposition and doping processes S. Georgiou, E. Raptakis, X. Xing, E. Hontzopoulosand Y. P. Vlahoyannis 3639 Kinetics of disproportionation of hypoiodous acid V. W. Truesdale, C. Canosa-Mas and G. W. Luther 111 3645 Thermodynamic behaviour of liquid p-xylene near freezing S. Castro, M. Taravillo, V. G. Baonza, M. Caceres and J. Nhiiez 365 1 Pivalic acid as combined buffer and scavenger for studies of cloud water chemistry with pulse radiolysis T.Nauser and R. E. Buhler 3657 Apparently viscoelastic responses in a quartz crystal microbalance study of an electrodeposited non-aqueous colloid film J. S. Graham and D. R. Rosseinsky 3663 Hydrogen adsorption-desorption and oxide formation-reduction on polycrystalline platinum in unbuffered aqueous solutions D. Pletcher and S. Sotiropoulos 3669 Relation between the surface states of oxide films at Rh electrodes and kinetics of the oxygen evolution reaction G. Jerkiewicz and J. J. Borodzinski 3671 Synthesis, photochemistry and cross-linking of visibly sensitised photopolymers of PVA based on (E)-2-(4-formylstyryI)-3,4-dimethylthiazolium methylsulfate I.C. Barker, N. S. Allen, M. Edge, J. A. Sperry and R. J. Batten 3685 Growth of a well oriented layer of p-nitroaniline on the (100) plane of a p-cyanoaniline single crystal, studied by polarized absorption and second-harmonic generation H. Kobayashi, T. Ehara and M. Kotani 3689 Water adsorption in carbons described by the Dubinin-Astakhov and Dubinin-Serpinski equations F. Stoeckli, L. Currit, A. Laederach and T. A. Centeno 3693 Samarium oxide catalyst; Formation, characterization and activity towards propan-2-01 decomposition. An IR spectroscopic study G. A. M. Hussein 3699 In situ FTIR study of CO-H, reactions over Rh/TiO, catalysts at high pressure and temperature J. A. Chudek, M. W. McQuire, G. W. McQuire and C. H. Rochester 3711 Ternary V-Ti-Si catalysts and their behaviour in the CO + NO reaction M.Galan-Fereres, R. Mariscal, L. J. Alemany, J. L. G. Fierro and J. A. Anderson 3719 "A1 and ,'Si solid-state NMR studies of dealuminated mordenite J. Barras, J. Klinowski and D. W. McComb 3725 Dissolution of amorphous aluminosilicate zeolite precursors in alkaline solutions. Part 3.-Influence of temperature on the dissolution process T. Antonic, A. &hek and B. Subotic ~~ FARADAY COMMUNICATIONS 3729 Carbon-halogen second-order quadrupolar and indirect spin-spin coupling effects in high-resolution 3C NMR spectra of halobenzenes A. E. Aliev, K. D. M. Harris, P. J. Barrie and S. Camus 3731 High sensitivity in reaction cross-section measurements by optical methods: The Ca(’D,) + C2H,0H +CaOH* + C,H, reaction M.Garay, M. Esteban, E. Verdasco and A. Gonzalez Ureia ~~ ~ ~~ ~~ CORRIGENDUM 3733 Corrigendum to reactivity of zeolite hydroxy groups toward a-donor bases. H-D exchange with 3-methylpentane C. J. A. Mota, R. L. Martins, L. Nogueira and W. B. Kover 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 WlV 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. I

ISSN:0956-5000    

DOI:10.1039/FT99490FP262

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3591-3599 Resonance-enhanced Multiphoton lonisation Spectroscopy of Thiirane Ross A. Morgan, Pilar Puyuelo,? Jonathan D. Howe and Michael N. R. Ashfold School of Chemistry, University of Bristol, Bristol, UK, BS8 1TS Wybren Jan Buma, Jolanda B. Milan and Cornelis A. de Lange Laboratory for Physical Chemistry, University of Amsterdam, Nieuwe Achtergracht 127, 10I8 WS Amsterdam, The Netherlands We present a new and more extensive interpretation of the vertical electronic spectrum of the thiirane molecule based on (i) analysis of observed two-and three-photon resonance enhancements in the wavelength-resolved multiphoton ionisation spectrum of jet-cooled thiirane and (ii) measurements of the kinetic energies of the photo- electrons accompanying the multiphoton ionisation process.Six Rydberg series are identified, including the four series identified in previous one-photon vacuum ultraviolet absorption studies of this molecule. Series assign- ments are offered based on the measured quantum defects, their relative two- and three-photon transition strengths and the dependence (if any) of the two-photon transition intensities upon changing from linearly to circularly polarised photoexcitation. Our analysis suggests that, as in H,S, many of the Rydberg orbitals have hybrid I character; it also points to the need for some revision of all previous interpretations of the electronic spectrum of the thiirane molecule. Multiphoton excitation methods are increasingly finding use as a means of providing new spectroscopic and dynamical information about the excited electronic states of small and medium-sized molecular species.' This paper demonstrates the way in which, through use of a range of multiphoton spectroscopy methods, it has been possible to obtain the first comprehensive and internally consistent interpretation of the vertical electronic spectrum of the three-membered hetero- cycle thiirane (ethylene sulfide, H2mH2). Several experimental2p5 and theoretical6 studies of the vertical elec- tronic spectrum of thiirane have been reported previously ; most interpretations to date have been based on comparison with the prototypical dihydride, hydrogen ~ulfide.~Such comparisons have been justified by the fact that, in both mol- ecules, the highest occupied molecular orbital in the respec- tive ground electronic states is largely non-bonding, being well approximated as the 3p, orbital of atomic sulfur.The one-photon absorption spectra of both molecules exhibit a region of continuous absorption in the near UV followed, at shorter wavelengths, by numerous sharper vibronic features which have been assigned in terms of Rydberg series converg- ing to the ground-state ion. Extrapolation of these various series has yielded values for the respective first ionisation potentials of H2S8l9 and thiirane5 (84428 & 10 and 73090 & 60 cm-', respectively) in good accord with those determined both by conventional He1 photoelectron and, more recently in the case of H2S, by two-photon non-resonant zero-kinetic-energy (ZEKE) photo- electron spectroscopy.' Unambiguous assignment of the structured Rydberg fea- tures in the electronic absorption spectrum of thiirane is com- plicated by the fact that spectral congestion, almost certainly compounded by predissociation broadening, precludes obser- vation of any rotational fine structure. Previous experimental analyses2p5 of the Rydberg structure exhibited by this mol- ecule have thus been based largely on determination of the quantum defects, 6, of the various features, and analogy with the then-existing interpretations of the vertical electronic spectrum of H,S.The validity of the latter strategy must be t Present address : Departamento de Quimica, Universidad de la Rioja, c/ Obispo Bustamante 3, Logroiio, Spain.questioned, however, since both recent two-photon resonant multiphoton ionisation (MPI) spectroscopy experiments involving the H,S m01ecule'~ and recent ab initio calcu-lations of the vertical electronic spectra of both H2S15 and thiirane6 have indicated the need for some revision of the traditional assignments. In particular, the recent and theoretical' studies have served to high- e~perimental'~ light the 'atomic-like' nature of the Rydberg transitions in H2S and the propensity for odd(even) changes in the quantum number in electronic transitions brought about by the absorption of an odd(even) number of photons. Given these concerns, and the presumed parallels between H,S and thiirane, it seemed worthwhile to record multiphoton excita- tion spectra of thiirane in order to try to gain further insight into the nature and energetic ordering of the excited elec- tronic states in this molecule.Here we report the results of a systematic study of the two-and three-photon resonance-enhanced multiphoton ion- isation (REMPI) spectra of thiirane, and a small portion of the corresponding one-photon spectrum. The multiphoton spectra both show a wealth of resolved vibronic structure, much of which can be arranged into well defined Rydberg series converging to the ground-state ionisation limit. The interpretation of individual vibronic features appearing in these spectra has been greatly aided by concurrent measure- ments of the kinetic energies of the photoelectrons accom-panying the MPI process, whilst series assignments have been guided by quantum defect considerations, by the relative intensities of the series members in the two- and three-photon excitation spectra and, in appropriate cases, by the observed change in the two-photon band intensities upon switching from linearly to circularly polarised photoexcitation.Experimental The results reported herein were obtained using two comple- mentary apparatus. Mass-resolved MPI spectra of thiirane were recorded in Bristol by expanding the pure gas (Aldrich, 98%, backing pressure ca. 1 atm) through a pulsed nozzle into the source region of a home-built time-of-flight (TOF) mass ~pectrometer.'~~~~ Two-photon REMPI spectra at exci- tation wavelengths shorter than 335 nm were recorded using the frequency-doubled output of an Nd :YAG pumped dye laser (Spectron 803 plus SL4000 together with an Inrad ‘autotracker’ system employing a selection of KDP or /3-barium borate doubling crystals).The longer wavelength portion of the two-photon excitation spectrum, and the cor- responding three-photon REMPI spectrum, were recorded using an excimer pumped dye laser (Lambda Physik EMG 201 and FL 2002). In all cases the output radiation was focussed (fl = 200-300 mm) into the source region of the TOF mass spectrometer so as to interrogate the gas pulse at its point of maximum flux. Under normal operating condi- tions both laser outputs were linearly polarised, with their electric vector aligned perpendicular to the TOF axis.When necessary, these outputs could be circularly polarised by insertion of an appropriately oriented Fresnel rhomb. Ions formed in the source region were subjected to two stages of acceleration before entering a field-free drift region and ulti- mate detection by a channel electron multiplier. TOF spectra were obtained by feeding the amplified output from this multiplier into a fast digital oscilloscope (LeCroy 9400, 125 MHz bandwidth) and thence, uia an IEEE interface, to a 386 PC. MPI spectra associated with the formation of a particu- lar ion mass were obtained by scanning the laser wavelength and collecting that fraction of the total ion signal that fell within the user-selected narrow time-window spanning the appropriate TOF.Wavelength calibration, in the visible, was achieved by simultaneously recording the optogalvanic spec- trum of neon excited in a hollow cathode discharge. REMPI photoelectron spectra (PES) were recorded in Amsterdam using a ‘magnetic bottle’ spectrometer the design 5 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 and operation of which have also been detailed pre-viously.’7*18 Briefly, light from an excimer pumped dye laser system (Lumonics HyperEx-460 and HyperDye-500), doubled in frequency if necessary using a frequency doubling unit (Lumonics Hypertrak-1000), is focussed (fl = 25mm) into the ionisation region of the electron spectrometer where it inter- cepts an effusive beam of pure thiirane vapour.Approx- imately 50% of the photoelectrons resulting from each laser pulse are extracted into the spectrometer and their kinetic energies determined via their measured times of arrival at a pair of microchannel plates situated at the end of the 500 mm flight tube. The signal from the microchannel plates is pre- amplified and passed to a transient digitiser, which is inter- faced to a PC to enable further data manipulation and analysis. A PE spectrum is constructed by increasing the retarding voltage on a grid in the flight tube in a stepwise fashion, and transforming only the high-resolution part of the TOF spectrum. In this way it was possible to achieve 15 meV (FWHM) resolution at all kinetic energies in the present experiments.The measured electron kinetic energies were placed on an absolute scale by doping the thiirane sample with xenon and measuring, simultaneously, the kinetic ener- gies of electrons resulting from well documented REMPI transitions terminating on the two spin-orbit states of the Xe+ ion. Results and Discussion Fig. 1 displays a composite three-photon REMPI spectrum of thiirane (i.e. involving resonances in the wavenumber 6 7 I I II I i I I I I I I I I I I I I I I I I I I I 55 000 60 000 65000 three-photon wavenumber/cm -’ 70 000 I I I I I I I I I I I I I 540 520 500 480 460 440 420 excitation wavelength/nm Fig. 1 3 + 1 MPI spectrum of thiirane over the wavelength range 540-417 nm recorded using linearly polarised light and monitoring only those ions with times-of-flight appropriate to m/z 45 and/or 60.The displayed spectrum is a composite, obtained by splicing together spectra recorded using five different dyes. As discussed in the text, we have not taken any particular steps to ensure correct normalisation of the relative intensities of features appearing within the tuning range of any one dye, or between one dye tuning curve and the next. Arrows indicate the wavelengths at which we switched from one dye to another. Members of the Rydberg series that appear clearly as three-photon resonances are indicated via the combs superimposed over the spectrum.The indicated principal quantum numbers, n, are consistent with the dominant n’ll t3b, assignments offered in Table 1.Ei = 73 090 cm-’ J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 3593 I1 5 I 6 I 7 I 8 I 9 10 I 1 4 I11 I Iv 3 I V 3 I 4 I 5 I 6 55 000 60 000 65 000 two-photon wavenumber/cm -' 70 000 ~ 380 360 340 320 300 280 excitation waveleng t h/n rn Fig. 2 2 + 1 MPI spectrum of thiirane over the wavelength range 390-278 nm recorded using linearly polarised light and monitoring only those ions with times-of-flight appropriate to m/z 59 and/or 60. As in Fig. 1 the displayed spectrum is a composite, and the same comments apply. Members of the Rydberg series that appear clearly as two-photon resonances are indicated via the combs superimposed over the spectrum.range 55 300-72 OOO cm-') recorded using linearly polarised laser radiation in the wavelength range 540-417 nm. Fig. 2 shows the corresponding linearly polarised two-photon REMPI spectrum, recorded using excitation wavelengths in the range 390-278 nm (corresponding two-photon wavenum- ber range 51 500-72 000 cm-'). Clearly, the two show signifi- cant differences. At this point, we comment that we are here interested primarily in the peak positions and caution against reading too much into the relative intensities of the various peaks in these spectra. As in any REMPI spectrum, these spectra will tend to be dominated by transitions involving the less predissociated excited states; however, simply by increas- ing the laser intensities it is possible to enhance the relative showing of transitions involving more short-lived excited state^.'^'^ The present spectra were recorded using a number of different dyes; we have not taken any particular care about normalising the laser power either between dyes, or within the tuning range of any one dye.Ion fragmentation is a further factor affecting the relative peak intensities. The spectra shown in Fig. 1 and 2 were obtained by monitoring the wavelength dependence of one or more specific features in the TOF spectrum, e.g. the parent ion peak. However, as Fig. 3 shows, the relative showing of the various m/z peaks observed at any one wavelength varies with the incident pulse energy.The relative peak intensities also vary with excitation wavelength, with fragmentation becoming increasingly preva- lent as we scan to shorter wavelengths. Checks were made to confirm that, for any given set of experimental conditions (i.e. laser polarisation, intensity etc.),the excitation spectra for all of the major molecular ion peaks exhibit the same pattern of band positions though, inevitably, there are some variations in relative peak intensities. We reserve discussion of these fragmentation Datterns until later in this section. I I I I I r 10 20 30 40 50 60 mlz Fig. 3 Ion TOF spectra obtained following REMPI at 329.2 nm using laser pulse energies of (a)3.5 and (b) 1.5 mJ. These show clearly the increased fragmentation that results when using higher incident- lieht intensities.3 594 REMPI-PES In common with previous workers, we shall attempt to assign the observed features as members of Rydberg series converg- ing to the first ionisation potential. Tokue et al.' derived a value Ei = 73090 f60 cm-' by extrapolating one of the more prominent series in the one-photon absorption spec- trum of thiirane. We determine an identical value for Ei from measurements of the photoelectron kinetic energies arising in the REMPI of thiirane. Fig. 4 shows three illustrative REMPI-PES. The first [Fig. 4 (a)],obtained following exci- tation of thiirane at 497.1 nm, is dominated by a single peak with a maximum at 0.912 eV (7360 cm- '). This spectrum can be understood most simply by assuming that the electrons arise as a result of a three-photon resonance-enhanced, four- photon ionisation process.All of the measured REMPI-PES are consistent with ionisation occurring via an n + 1 process. In this particular case the spectral simplicity suggests that the final one-photon ionisation from the resonance-enhancing Rydberg level involves a Au = 0 transition. Furthermore, our inability to discern any faster electrons in this PES encour- ages the assumption that the parent ion is formed in its vibra- tionless level and thus that the resonance enhancement involves an electronic origin transition. Given this assump- tion, the measured photoelectron kinetic energy implies a 0.2 0.4 0.6 0.8 1.o (b) 1.4 1.6 1.8 2.0 2.2 2.4 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 value for the first adiabatic ionisation potential of thiirane identical to that reported previously by Tokue et al.' The photoelectron spectra shown in Fig. 4(b) and (c), obtained following excitation at 328.9 nm and 323.3 nm, respectively, provide further examples of multiphoton excita- tions in which the final ionisation step shows a strong pro- pensity for being Franck-Condon diagonal. In both cases the resonance enhancement occurs at the two-photon level, and the electron kinetic energies are consistent with ion formation involving absorption of a total of three photons. In the former PES, the measured photoelectron kinetic energy of the major peak C2.246 eV (18 120 cm-')I indicates that the parent ion is formed predominantly in its ground vibrational state; this, in turn, suggests that the resonance enhancement involves a vibrationless Rydberg level.However, in the latter PES [Fig. 4(c)], the peak associated with formation of the vibrationless ion [at a kinetic energy of 2.440 eV (19680 cm-')I is weak. The dominant peak occurs at lower kinetic energy, indicating that most of the ions are formed with ca. 130 meV (1050 cm-') of internal energy. This separa- tion is reminiscent of the energy spacings between the short progression of features observed in the conventional one-photon PES of thiirane"." and, as in these earlier studies, we attribute it to excitation of the v4 mode (a totally sym- metric mode involving concerted CH, wagging and ring breathing motions) of the parent ion.Franck-Condon con-siderations then point to the likelihood that the excited vib- ronic level of the neutral providing the resonance enhancement at the two-photon energy carries one quantum of v4. This conclusion gains added weight once we recognise that the two-photon resonances that give rise to the two PES shown in Fig 4(b) and (c) are themselves separated by essen- tially the same interval, 1040 cm-.'. The availability of PES such as these, and the insight they provide concerning the vibronic character of the resonance- enhancing level, proved to be absolutely crucial to the analysis of the REMPI spectra shown in Fig. 1 and 2. As a final example of the way in which such PES can aid in spec- tral assignment, consider the triplet of strong three-photon resonances observed at excitation wavelengths ca.500 nm (Fig. 1). We have already seen from the PES shown in Fig. 4(a) that the central feature (A,,, = 497.1 nm) can be assigned to an origin transition. What should we make of the two neighbouring features, which are both obvious as three-photon resonances at A,,, = 501.0 nm and 495.8 nm but barely discernible as two-photon resonances? Neither obvi- ously fits as part of a Rydberg series, yet the apparent strength of the three-photon resonances might seem to miti- gate against their assignment in terms of transitions involving vibrationally excited levels of a Rydberg state derived from excitation of what is, essentially, a non-bonding electron.However, the associated REMPI-PES [Fig. 5(a) and 5(b)] show that this, indeed, is how they should be viewed. Consider first the higher-frequency peak observed at an ILexcitation wavelength of 495.8 nm (corresponding three- Ijl.li...jl....l....I ..,,,., 1.6 1.8 2.0 2.2 2.4 2.6 photoelectron kinetic energy/eV Fig. 4 MPI-PES taken following excitation at: (a) 497. 1 nm (60358 cm- ') where we excite a three-photon resonance involving the origin level of the state we identify as the first member of Series V; (b) 328.9 nm (60805 cm-l), for which the process is resonance enhanced at the two-photon energy by the origin level of the first member of our Series 11; (c) 323.3 nm (61 845 cm-I). The kinetic energy of the dominant peak in this spectrum is consistent with the partner ion carrying ca.1050 cm-' of internal energy. Thus we deduce that the ion and, uia Franck-Condon arguments, the resonance-enhancing level at the two-photon energy both involve one quantum of the v4 vibration. photon wavenumber 37 = 60 504 cm- '): The 3 + 1 REMPI-PES [Fig. 5(b)] is dominated by two features, one corre- sponding to the formation of vibrationless ions, the other to ions with some 420 meV (3150 cm-') of internal energy. We note that this three-photon resonance occurs similarly shifted (by ca. 3020 cm- ') from a demonstrable origin transition at 57481 cm-'. Built upon the faster of the two strong PES features is a progression of peaks separated by ca.141 meV (1120 cm-'), each of which appears with a satellite shifted some 67 meV (540 cm-l) to higher energy. Guided by the earlier ab initio calculations,' we associated these intervals with excitation of, respectively, the v3 and v5 modes of the ion. The first of these two totally symmetric vibrations J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 mi I il I1h I.j./l...,,,l LL,,,,,, 0.2 0.4 0.6 0.8 1.o I..,.,....,...,J,I..,... 0.2 0.4 0.6 0.8 1.o I.IIII.ILII,~/II,,,IIL 0.4 0.6 0.8 1.o 1.2 photoelectron kinetic energy/eV Fig. 5 MPI-PES taken following excitation at (a)501.0 nm (59 883 cm-'),(b) 495.8 nm (60504 cm-')and (c) 488.9 nm (61 362 cm-'). The REMPI-PES reveal that, in each case, the resonance-enhancing level at the three-photon energy should be viewed as a vibrationally excited level of the Series IV member whose origin appears at 57 48 lcm -'.involves a concerted C-C stretching and CH, waggingmotion, whilst the latter, low-frequency, mode is associated with motion of the S atom along the C, axis perpendicular to the C-C bond. The simplest interpretation of these observa- tions is that the 60504 cm-' resonance is associated with a three-photon transition to a level which, in a zero-order description, involves ca. 3020 cm- ' of vibrational energy (at least part of which is concentrated in the v3 and v5 modes). Because of its accidental near-degeneracy, this zero-order state II/, mixes with, and gains transition intensity from, the Rydberg origin state (zero-order wavefunction II/R) observed as a particularly strong three-photon resonance at 35 = 60 358 cm- '.A similar explanation most likely accounts for the lower-frequency resonance observed at 35 = 59883 cm-', ca. 2400 cm-' above the 57481 cm-' Rydberg origin. The REMPI-PES in this case [Fig. 5(a) also shows two dominant peaks; again, the faster feature is attributable to the formation of vibrationless ions whilst the slower elec- trons must rise in association with ions with ca. 337 meV (2720 cm- ') of internal energy. Once again, the REMPI-PES also displays a progression in the 141 meV (1 120 cm-') mode of the ion and peaks associated with excitation of one quantum of the 67 meV (540 cm-') vibration.The pattern of vibronically induced resonances is repeated ca. 1050 cm- ' to higher energy in the 3 + 1 REMPI spectrum; these we attrib- ute to the same pattern of three interacting states, with the transition strength now provided by the three-photon reson- ance involving the 4' level of the Rydberg state whose origin appears at 60358 cm-'. As expected, the REMPI-PES associated with each of these three resonances display a pattern of features very similar to that shown by their respec- tive counterparts at lower energy, each shifted in energy by an amount corresponding to one quantum of v4 in the ion, and each showing one (or two) higher-kinetic-energy peaks associated with ion formation in its ground vibrational state, and with one quantum in the low-frequency v5 mode.Fig. 5(c) provides one illustrative example. In concluding this section, we should comment that since the actual eigenstates associated with each of these three interacting levels have a wavefunction of the general form + then the REMPI-PES obtained following excitation via what, in a zero-order description, we would regard as the Rydberg origin level should, in fact, be contaminated by slower elec- trons associated with vertical one-photon ionisation of the II/, component of the wavefunction. Comparison of the REMPI- PES displayed in Fig. 4(a) and (b) show that, at least qualit- atively, this is the case. Assigning the Multiphoton Resonances Armed with this insight into the vibronic character of the various resonance-enhancing levels we are now in a position to offer a global assignment of the excited electronic states of the thiirane molecule.Band maxima (5) measured in the present work are listed in Table 1. In the case of proposed Rydberg origins we also list the associated effective quantum numbers, n*, calculated using the relationship 5 = Ei -R/(n*)' (1) , where Ei = 73090 cm-', R is the Rydberg constant (109 737 cm-') and n* = (n-a), with n being the principal quantum number and 6 the quantum defect. For complete- ness, we also list the major peaks identified in the one-photon absorption studies of Basco and Morse3 and Tokue et al.' Before embarking on a detailed discussion of these results it is instructive to consider the range of excited states we might expect.Ground-state thiirane has C,, symmetry. Adopting the convention that z corresponds to the C, axis and that the heavy atom ring lies in the yz plane, then the ground-state electronic configuration may be written ...(2bl)2(7a,)2(la,)2(8al)2(4b,)2(3bl)2; 'A, (2) in which the highest occupied 3b1 orbital is largely comprised of the non-bonding sulfur 3p, orbital, directed perpendicular to the plane containing the ring of heavy atoms.' Promoting one 3b, electron to higher nl Rydberg orbitals can be expected to give rise to one 's' (I = 0) series involving states of 'B, electronic symmetry and n 2 4, three 'p' series (of 'B1, 'A, and 'A, symmetries, respectively, all with n 2 4), five 'd' series (two of 'B, symmetry together with one each of 'A,, 'A, and 'B, symmetry, all with n 2 3), seven 'f' series (two each transforming as 'A1, 'A, and 'B1, the other of 'B, sym-metry, all with n 2 4) etc.The quotation marks are included as a reminder that configuration mixing may render 1 an approximate label, whilst not invalidating the overall state count. Analogy with atomic sulfur and with H,S suggests that 'pure' 1 series will exhibit quantum defects ca. 2.0 (s series), ca. 1.6 (p series) and CQ. 0 (for d, f and higher l f~nctions).~,~.'~Such expectations led Tokue et al.' to propose the identification of one s (6 = 1.91), one p (6 = 1.66) and two d Rydberg series (6 = 0.35 and 0.07, respectively) in their one-photon absorption spectrum of thiirane.Further, given that the second (adiabatic) ionisation potential of thii- fane (associated with removal of a 4b, electron) lies at ca. 11.35 eV,".'2 it is unlikely that any features associated with J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 Wavenumbers, effective quantum numbers (n*) and proposed assignments of the thiirane resonances observed in the present multi- photon study. The 1 function we presume to be dominant in each of the Series I-VI is indicated (between quotation marks). ;/cm - n* vibronic assignment one-pho ton VUV" Series I (n's'a, t3b,) (2.06) 47 137 62 294 3.19 63 345 4; 64 396 66 898 4.2 1 4; 69 065 5.22 Series I11 (n'p'b, t3bJ 51 986 2.28 52029 (52010) 52 510 5; 53015 4; (53 087) (3.31) 63 119 67 235 4.33 67 308 69 202 5.31 69 286 70 360 6.34 70 428 (70 505) Series V (n'd'b, t3bJ 60 358 2.94 60 507 61 370 4; 62 400 4;65 986 3.93 66 01 1 (65 970) 67 038 4: 68 543 4.91 68 592 (68 535) 69 929 5.89 69 984 (69 913) a Ref.5, value in parentheses, ref. 3. Rydberg states belonging to series converging to this second limit should appear at energies below ca. 68000 cm-' in the vertical excitation spectrum. We now turn to discuss the various Rydberg series in turn. The lowest-energy feature identified in the multiphoton spectra appears as a two-photon resonance at 51986 cm-' (Fig. 2). We did not search for any three-photon resonance associated with this excited state since, at the relevant excita- tion wavelengths (ca.584 nm) ionisation would require absorption of a further two photons; interpretation of any spectrum so obtained would thus be complicated by possible resonance enhancements at the energy corresponding to four absorbed photons. This same feature appears strongly (E,,, z 6 x cm2) in one-photon two weaker features at higher energy have been assigned in terms of transitions to excited vibronic levels involving, respectively, one and two quanta of v4. Basco and Morse3 originally pro- posed that this feature be viewed as the first member of the nst 3b1 series but, given its quantum defect (6 x 1.72, assuming n = 4) and the fact that dichroism measurements4 are consistent with the excited state having A, symmetry, assignment in terms of electron promotion to the 4pb, Ryberg orbital seems more reasonable. Recent ub initio calculations6 support this conclusion, though the intense showing of this transition in one-photon absorption must cast doubt on the purity of the 1 = 1 description.Evidence for the predissociative nature of this excited state is provided by vibronic ii/cm -n* assignment Series I1 (n'p'a, c3b,) 47 660 2.08 60 805 2.99 61 587 11; 61 845 4;62 627 4: 11; 66 356 4.04 67 409 4;68 777 5.04 70010 5.97 70 826 6.96 71 358 7.95 71 730 8.98 71 985 9.96 Series IV (n'd'a, t3bJ 57481 2.65 58 200 11; 58 540 4;59 884 (+2403 cm-') 60 504 (+ 3023 cm-') 64 789 3.64 68 080 4.68 69 593 5.60 71 221 7.66 Series VI (n'd'b, c3b,) 61 316 3.05 62 392 4: 66 506 4.08 67 556 4;68 854 5.09 69 910 4;70 129 6.09 70 899 7.08 71 409 8.07 one-photon VUV" 57 508 (57 490) (58 535) 64821 (64771) 67 939 (68 032) 69604 (69592) 71 159 (71 308) 61 286 66 520 68 856 (69913) 70 112 photofragment translational spectroscopy experiments,20 which show C-H bond fission to be the dominant primary decomposition route following photoexcitation at 193 nm.S('D) atom elimination is a minor (ca. 10%) competing frag- mentation pathway at this wavelength; measurements of a positive recoil anisotropy parameter (/3 M + 1) for these S atom products provides further confirmation of the A, excited-state symmetry.20.21 A n umber of higher-frequency features exhibiting similar 6 values are grouped together as 'Series 111' in Table 1.These are clearly evident (though with diminishing intensity) in the one-photon spectra reported previously3*' but none show particularly clearly in the REMPI spectra, perhaps indicative of a short excited-state lifetime. The next feature apparent in the two-photon REMPI spec- trum is a weak, diffuse maximum centred around 54800 cm-'. This band has no reported counterpart in the one-photon spectrum. Given its quantum defect (6 z 1.55, assuming n = 4) and the recent ab initio results,6 it is tempt- ing to speculate that this resonance may be associated with the one-photon dipole-forbidden 'A2 6 'A, transition arising as a result of electron promotion to the 4pb2 orbital.If correct, the diffuseness of the observed resonance would suggest a significant antibonding valence contribution to the excited b, orbital. We were unable to recognise any higher- energy features that could obviously be arranged into a series with this 54 800 cm-band as first member. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 The feature centred at 57481 cm-', which we see both as a two- and a three-photon resonance, also shows strongly in one-photon ab~orption.~-~ Electric dichroism measurements are consistent with the excited state having 'B, symmetry, and assignment in terms of the electronic promotion 3da, +-3b1 has been propo~ed.~The quantum defect (6 x0.35, assuming n = 3) is uncharacteristically large for a pure 1 = 2 function; this, and the fact that the resonance shows in one-, two- and three-photon absorption, indicates that the excited a, orbital probably has hybrid 1 character.Similar p/d mixing has been invoked previously to account for the relative strengths of Rydberg transitions observed in the multiphoton spectrum of H2S and D2S.14Higher members of this series (labelled Series IV in Table 1) show strongly in one-photon ab~orption~-~and are clearly evident as three-photon reson- ances in Fig. 1. However, inspection of Fig. 2 reveals that the n > 3 members of Series IV do not appear as two-photon resonances, perhaps indicating that, for this series at least, the extent of 1 mixing decreases rapidly with increasing n. Returning to the first (n = 3) member, we have already seen from their associated REMPI-PES [e.g.Fig. 5(u) and (b)] that several of the resonances appearing immediately to higher energy in both the two- and three-photon excitation spectra can be associated with vibrationally excited levels of this Rydberg state. These resonances are also listed in Table 1. Series V, which dominates the spectrum shown in Fig. 1, also contributes two-photon resonance enhancements (Fig. 2) and has been identified previously in one-photon ab~orption~-~,though the earliest analysis3 failed to associate the first (60358 cm-') feature with this series. Short (one- member) vibronic progressions, involving vk built on this first origin and on the next series member (at 65986 cm-') are clearly evident in Fig.1;the n = 6 member of Series IV over- laps the expected position for the 4; transition associated with the third member. Both Basco and Morse3 and Tokue et aL5 assign this series in terms of the excitation nd +-3b, which, given the observed quantum defect (6 x0.06, assuming that the 60358 cm-' member has n = 3), seems reasonable. However, the recent ab initio study6 advocated the alternative plausible assignment, namely that Series V in Table 1 be associated with the promotions ns +-3b1. In this interpretation, the 60 358 cm- ' feature would involve excita- tion to the 5s orbital; the implied quantum defect (6 x 2.06) is also reasonable. The present multiphoton studies provide a means of distinguishing between these two alternative assign- ments since, as Fig.6 shows, we find that the relative inten- sities of the two-photon resonances we associate with Series V members decrease when we use circularly polarised light to record the spectrum. This is a characteristic diagnostic for a two-photon transition linking two states of the same sym- metry, since only for such totally symmetric transitions can the scalar [T:(A)] part of the two-photon transition moment (the component that is forbidden in circularly polarised excitation) have non-zero amplitude. Thus we conclude that the excited states involved in Series V have 'A, electronic symmetry, and that the populated Rydberg orbitals have b, symmetry. This, in turn, provides strong support for the orig- inal experimental assignment of Series V in terms of the orbital promotion ndb, t3b,.The fact that these states also appear as two-photon resonances, however, suggests a degree of d/p mixing in the Rydberg orbitals. Our assignment of the Series IV and V members as having, respectively, 'B ,and 'A, electronic symmetry allows further speculation on the nature of the vibronic interaction linking the states responsible for the three strong resonances appear- ing ca. 60 300 cm- ' in the 3 + 1 REMPI spectrum (Fig. 1). The PES [Fig. 4(4, 5(a)and (b)] led to the conclusion that the interacting zero-order states comprised the Series V n = 3 3 4 5 v1 1>7----7-3 4 v4' 5 60000 62000 64000 66000 68000 two-photon wavenumber/crn- ' Fig.6 2 + 1 MPI spectrum of thiirane over the wavelength range 333-289 nm recorded using (a) linearly and (b)circularly polarised light and monitoring only those ions with times-of-flight appropriate to m/z 59 and 60. As in Fig. 2 the displayed spectrum is a composite obtained using more than one dye, and the same comments apply. Two-photon resonance enhancements brought about by vibronic levels of A, symmetry are highlighted by the combs superimposed over the spectrum. origin level (of 'A, electronic symmetry) and two vibrationally excited levels of the n = 3 member of Series IV. Given that this latter state is deduced to have 'B, electronic symmetry, vibronic interaction requires that the 'vibration' (be it a fun- damental or a combination mode) has the overall vibrational symmetry b,.The form of the associated REMPI-PES [Fig. 5(a) and (b)]points to activity in the v3 and v5 modes of the ion (both of which are totally symmetric modes). Ener- getic considerations, together with the requirement that the overall combination has b, vibrational symmetry, then suggest the combinations 3l5' 11 and 31521 1 ' as plausible assignments for the vibrational contributions to the zero-order states contributing to the 59 883 and 60 504 cm- ' res-onances. Vibronic interaction with the near-resonant 'A, electronic origin accounts for the appearance of resonances involving the non-totally symmetric v1 ,mode. The REMPI-PES studies allow us to identify three further origin transitions in the energy region associated with near- integer n* values.Two appear in the present multiphoton study exclusively as two-photon resonances, the third only as a three-photon resonance enhancement. We shall discuss these, and their associated series, in order of increasing energy. The first, which for reasons that become apparent later we shall label Series 11, makes its first contribution to our two-photon REMPI spectrum through the feature centred at 60805 cm-' [see Fig. 2 and the associated REMPI-PES displayed in Fig. 4(b)].Higher members of this series can be identified immediately to lower frequency of the more intense Series VI origins (see below). As Fig. 6 illus- trates, the relative showing of our Series I1 origin transitions are enhanced upon changing to circularly polarised excita- tion.Their quantum defects (6 x 0, if we were to assume that the 60 805 cm- feature has n = 3) might suggest an assign- ment in terms of an nd t3b, orbital promotion, but such an analysis runs contrary to any expected propensities based on atomic selection rules. The interpretation that follows is based on unravelling the vibronic structure that appears on the high-frequency side of the 60805 cm-' member. The REMPI-PES studies confirm that the two-photon resonance at 61 845 cm-' is its associated 4; vibronic transition. More revealingly, the REMPI-PES studies indicate that the two very sharp peaks at 61 587 and 62627 cm-' are also due to vibrationally excited levels of the state whose origin appears at 60805 cm-'.The two very sharp peaks are separated by 1040 cm- '(a typical wavenumber for the v4 vibration which we see to be active in many of the Rydberg absorptions of thiirane), but the energy gap between the first of these two features and its electronic origin is only 782 cm-'. Such a separation matches none of the totally symmetric vibrational modes of ground-state thiirane22 or its ~ation;~ once again, we invoke the involvement of the v,,(b,) vibration (824 cm-' in the ground state neutral22) and assign the two sharp peaks as the 11; and 4A11; vibronic bands, respectively. Given that these excited levels have A, vibronic symmetry (from the fact that these two-photon resonances 'disappear' when excited using circularly polarised light), it therefore follows that the state responsible for the 60805 cm-' resonance has 'B, elec- tronic symmetry.Once again, we invoke vibronic mixing, in of the excited state of 'A, electronic symmetry that provides two-photon resonance enhancement at 61 316 cm-(and 62392 cm-I), as the source of the two-photon transition strength to these particular vibronic levels. We now return to the question of the electronic configu- ration from which this 'B, Rydberg state is derived. As we have seen, simple quantum defect considerations suggest assignment of these resonances in terms of the orbital pro- motion 3da1 + 3b1 or %a, + 3b1, whereas the observed two- photon transition strength is only compatible with the anticipated A1 propensity rule if we invoke significant p char- acter in the Rydberg orbital.The ab initio calculations of Carnell and Peyerimhop lend some support to this latter proposal. These workers6 reported calculated term values for the three members of the '4p-complex' spanning an energy spread of some 6400 cm-', with the 'B, member lowest in energy (close, in fact, to the 'B, state derived from the 4sa, +3b1 excitation) and exhibiting a quantum defect, 6 x 1.90. Indeed, so confident are Carnell and Peyerimhoftd of their theoretical result that they go so far as to offer a reassignment of the poorly resolved vibronic structure seen in the earlier near-UV absorption studies of thiirane.2*5 To test this conclusion further, we searched for ions and/or photo- electrons arising as a result of two-photon ionisation, resonance-enhanced at the one-photon energy by the predict- ed 'B,(4pa1) Rydberg state.Fig. 7 displays the spectrum obtained by monitoring the yield of parent ions as a function of excitation wavelength over the range 212.8-208.0 nm. Strikingly, we see clear evidence for a feature at 47 660 cm- hitherto associated with transition to the 5' component of the 'B,(4sa1) state, but not the corresponding origin band (which appears in one-photon absorption at ca. 47 137 cm-' with substantially greater intensity). This can be understood if the higher-energy feature has a longer lifetime (and thus a higher ionisation probability). Such would be unusual (though not without precedent23) if the two excited levels were associated with the same electronic state, but would be readily believable if the two levels were actually associated (and 4') levels 0' this case with the accidentally near-resonant J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1,,,,I,,,Il,/lll/l//I,/,,//I/Il 47 000 47 200 47 400 47 600 47 800 48 000 48 200 one-photon waven umber/cm -' 213 212 21 1 210 209 208 excitation wavelength/nm Fig. 7 1 + 1 MPI spectrum of thiirane over the wavelength range 212.8-208.0 nm recorded using linearly polarised light and monitor- ing those ions with times-of-flight appropriate to m/z 60. with two different electronic states. Precise measurement of the kinetic energies of the photoelectrons resulting from 1 + 1 MPI via the 47660 cm-' feature proved challenging: the REMPI-PES is found to be dominated by a single peak whose energy is consistent with formation of ground-state ions.Franck-Condon considerations thus point to the 47660 cm-' feature being another origin band, in accord with the theoretical conclusions of Carnell and Peyerimhoff.6 Guided by these calculations, we propose that this origin be associated with the electronic promotion 4pa, + 3b,, but rec- ognise the inevitablity of core-induced mixing between the close lying 4sa, and 4pa, orbitals. Hence the assignment of this feature as the first member of Series I1 in Table 1. The excited states which we collect together as Series VI appear as two-photon resonances (Fig.2) but are barely identifiable in the three-photon REMPI spectrum (Fig. 1). This is a somewhat surprising observation, since these same excited states were identified in the earlier one-photon absorption study of Tokue et aL5 Quantum defect consider- ations led these workers' to propose that this series derives from the orbital promotion ma, +3b, but the spectra shown in Fig. 6 clearly refute this assignment. Once again, the obser- vation that the relative intensities of the two-photon excita- tions to this set of excited states are greatly diminished when carried out using circularly polarised radiation indicates that the excited states have A, symmetry. This rules out assign- ment in terms of the electronic promotion nsa, + 3b1 (since the resulting states would necessarily have B, electronic symmetry); rather, we propose that, as with Series V dis-cussed earlier, the excited b, orbital has p/d hybrid character, possibly augmented by f Rydberg functions once n > 3.Such configuration mixing is implicit in the orbital descriptions given in Table 1, where we assign two series in terms of the promotion n'd'b, t3b1 notwithstanding the fact that there is only one d orbital of b, symmetry. Compared with higher members of this series, the first member appears both anom- alously weak and broad. We take this as an indication that predissociation of this excited state is particularly efficient. Vibronic coupling between this first series member and acci- dentally near-resonant vibrational levels of the 60 805 cm- member of Series I1 has been discussed previously.The final series identified in this work, which for eventual consistency we label as Series I, appears as a sequence of three-photon resonances; none of the members we identify were reported in the earlier one-photon Given J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 their quantum defects (6 M 1.81 if the 62294 cm-' resonance involves an excited state with n = 9,and the absence of these features in the two-photon REMPI spectrum, we favour assignment of this series in terms of the orbital promotion nsa, + 3b1. The first (n= 4) member of this series falls outside the range of wavelengths scanned in the present work, but has been identified previously' via its one-photon absorption, ca.47 137 cm-'. Its quantum defect (6 M 1.94) is thus a little larger than that of the higher series members, but such an observation is not uncommon for low-lying Rydberg states and is generally attributed to valence dilution of the Rydberg function. If, as is usually the case, this valence dilu- tion imparted antibonding character to the excited orbital then this suggestion would be consistent with the non-observation of the 4sa1 +-3b1 resonance in the 1 + 1 MPI spectrum shown in Fig. 7. Mechanism of Fragment Ion Formation We have been able to interpret all the observed photoelec- tron kinetic energy peaks in terms of REMPI resulting in the formation of just the parent ion. This, and the fact that the excitation spectra for all the molecular fragment ions show the same wavelength dependence as that of the parent ion, indicates that the observed fragment ions arise as a result of photoexcitation and subsequent fragmentation of the parent ion itself.Recalling the electronic configuration of thiirane (2) we see that the ground-state ion will have 'B, symmetry. The first two excited electronic states of the ion arise as a result of removing an electron from, respectively, the 4b, and 8a1 orbitals; conventional photoelectron spectroscopy' '9' 'shows the resulting 'B, and 'A, ion states to have vertical excita- tion energies (relative to the ground state) of ca. 2.3 and 2.8 eV. Excitation from the ground state to the first of these will be electric-dipole forbidden, but the 'A, +-X 'B, tran-sition should give strong one-photon absorption for wave- lengths less than ca.480 nm. Photoionisation studies24 indicate that the absorption of one such photon will provide more than sufficient energy to cause C-H bond fission (leading to formation of C,H,S+ fragment ions) and CH, elimination (i.e.production of HCS fragment ions). Smaller + fragment ions, which appear most notably in TOF spectra recorded at higher laser intensities, could arise via multi-photon excitation of the parent ion or following further photon absorption by the larger fragment ions (especially C2H3S+ which is never that dominant in our ion TOF spectra). Conclusions This study of the vertical electronic spectrum of the thiirane molecule provides further graphic demonstration of the way in which the many complementary facets of REMPI spectros- copy can be used to unravel details of the electronic spectros- copy of small and medium-sized gas-phase molecules, even in cases where spectral congestion and/or the short excited-state lifetimes preclude observation and analysis of any rotational fine structure.Members of six Rydberg series of thiirane have been identified via the two- 'and/or three-photon resonance enhancements they provide to the mass-selected multiphoton ionisation spectrum of a jet-cooled sample of the molecule. Measurements of the kinetic energies of the photoelectrons resulting from the MPI process provide a means of determin- ing the vibronic character of the resonance enhancing level, e.g.of distinguishing Rydberg origins from those involving vibrationally excited intermediate levels. The REMPI-PES studies also yield a value for the first adiabatic ionisation 3 599 potential of thiirane, E, = 73090 & 60 cm-', in excellent accord with that deduced from earlier analysis of the vacuum-UV absorption spectrum.' Series assignments are based on the measured quantum defects, their relative two- and three-photon transition strengths and the dependence (if any) of the two-photon transition intensities upon changing from linearly to circularly polarised photoexcitation. The present analysis suggests that, as in H,S"*'' many of the Rydberg orbitals have hybrid 1 character; it also suggests the need for some revision of all previous interpretations, both experimental'-' and theoretical,6 of the electronic spectrum of the thiirane molecule. The Bristol group is grateful to the SERC for their support of this work and for studentships (to R.A.M.and J.D.H.) and to Mr K. N. Rosser for all his practical help and encour-agement. M.N.R.A. is grateful to the Ciba Fellowship Trust for their support of this collaboration, whilst P.P. is grateful to the MEC of the Spanish Government for the award of a research fellowship. The Amsterdam group is happy to acknowledge the Netherlands Organisation for Scientific Research (NWO) for equipment grants and financial support. References 1 M. N. R. Ashford and J.D. Howe, Annu. Rev. Phys. Chem., 1994, 45, 57. 2 L. B. Clark and W. T. Simpson, J. Chem. Phys., 1965,43,3666. 3 N. Basco and R. D. Morse, Chem. Phys. Lett., 1973,20,404. 4 D. D. Altenloh and B. R. Russell, Chem. Phys. Lett., 1981, 77, 217. 5 I. Tokue, A. Hiraya and K. Shobotake, J. Chem. Phys., 1989,91, 2808. 6 M. Carnell and S. D. Peyerimhoff, Chem. Phys. Lett., 1993, 212, 654. 7 M. B. Robin, Higher Excited States of Polyatomic Molecules, Academeic Press, New York, 1974, vol. 1; 1985, vol. 3. 8 H. Masuko, Y. Morioka, M. Nakamura, E. Ishiguro and M. Sasanuma, Can. J. Phys., 1979,745. 9 C. A. Mayhew, J-P. Connerade, M. A. Baig, M. N. R. Ashfold, J. M. Bayley, R. N. Dixon and J. D. Prince, J. Chem. SOC., Faraday Trans. 2, 1987,83,417. 10 D. C. Frost, F. G. Herring, A. Katrib and C. A. McDowell, Chem. Phys. Lett., 1973, 20, 401. 11 L. Karlsson, L. Mattsson, R. Jadrny, T. Bergmark and K. Sieg-bahn, Phys. Scr. 1976, 13, 229. 12 D. H. Aue, H. M. Webb, W. R. Davidson, M. Vidal, M. T. Bowers, H. Goodwhite, L. E. Vertal, J. E. Douglas, P. A. Kollman and G. L. Kenyon, J. Am. Chem. SOC., 1980,102,5151. 13 I. Fisher, A. Lochschmidt, A. Strobel, G. Niedner-Schatteburg, K. Miiller-Dethlefs and V. E. Bondybey, J. Chem. Phys., 1993, 98, 3592. 14 M. N. R. Ashfold, W. S. Hartree, A. V.Salvato, B. Tutcher and A. Walker, J. Chem. SOC., Faraday Trans., 1990, $6,2027. 15 V. Galasso, J. Phys. B., 1989,22, 2241. 16 M. N. R. Ashfold, A. D. Couch, R. N. Dixon and B. Tutcher, J. Phys. Chem., 1988,92, 5327. 17 B. G. Koenders, D. M. Wieringa, K. E. Drabe and C. A. de Lange, Chem. Phys. 1987,118, 113. 18 N. P. L. Wales, E. de Beer, N. P. C. Westwood, W. J. Buma, C. A. de Lange and M. C. van Hemert, J. Chem. Phys., 1994, 100, 7984. 19 M. N. R. Ashfold, Mol. Phys., 1986,58, 1. 20 P. Felder, E. A. J. Wannenmacher, I. Wiedmer and J. R. Huber, J. Phys. Chem., 1992,%, 4470. 21 H. L. Kim, S. Satyapal, P. Brewer and R. Bersohn, J. Chem. Phys., 1989, 91, 1047. 22 W. D. Allen, J. E. Bertie, M. V. Falk, B. A. Hess Jr., G. B. Mast, D. A. Othen, L. J. Schaad and H. F. Schaefer 111, J. Chem. Phys., 1986,84,4211, and references therein. 23 See, e.g., M. N. R. Ashfold, C. L. Bennett and R. N. Dixon, Chem. Phys. 1985,93,293. 24 J. J. Butler and T. Baer, Org. Mass Spectrorn., 1983, 18, 248. Paper 41042941;Received 14th July, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003591

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3601-3607 Equilibrium Geometry and Vibrational Frequencies of the 1 :2 van der Waals Complexes between Methyl Chloride and Hydrogen Chloride W. A. Herrebout and B. J. van der Veken* Department of Chemistry, Universitair Centrum Antwerpen, Groenenborgerlaan 171,B-2020 Antwerpen, Belgium The equilibrium geometry, relative stability and vibrational frequencies for the 1 : 2 van der Waals complexes between CH,CI and HCI have been investigated using MP2/6-31 + G** ab initio calculations. The results are used to interpret the IR spectrum of the 1 : 2 complex observed in solutions of CD,CI-HCI mixtures dissolved in liquefied argon, and show that in these solutions the chain complex CD,CI . HCI .HCI is formed. The mid- and far-IR spectra of several solutions in liquefied noble gases containing both an alkyl chloride [CD,Cl, C,D,Cl, (CD,),CDCl, etc.] and hydrogen chloride have been 9'described.In all spectra, evidence was found for the occurrence of a 1 : 1 species RCl -HCl in which the HC1 mol- ecule is hydrogen bonded to the chlorine atom of the RC1 molecule. In several spectra, evidence was also found for the existence of another species, the stoichiometry of which was determined to be (RCl).(HCI), and which was expected to have a chain structure. As can be seen in Fig. l(a) in such a complex, further described as RCl .HCl * HCl, the second HCl molecule is hydrogen bonded to the chlorine atom of the first HCl molecule. However, by analogy with the results obtained for H,O * (HCl), ,, dimethylether -(HCl), and dimethyl sulfide (HCl), ,, a bifurcated structure can also be proposed, in which both HCl molecules are hydrogen bonded to the chlorine atom of the RCl molecule.Such a structure is shown in Fig. l(b). Note that for all the solutions described before,'., only one band of the 1 : 2 species was observed in the vHCl region, while both geometries are expected to give rise to two bands. Furthermore, no absorption bands of the 1 : 2 species were observed in the vccl region. Thus, it is clear that in the pre- vious studies'V2 a rather incomplete description of the spectra of the 1 : 2 complex has been given. As little is known about the 1 : 2 complex between methyl chloride and hydrogen chloride, in this study it was decided to carry out ab initio calculations of the equilibrium geometry, the relative stability and the vibrational fre-quencies for both the bifurcated and the chain structure.As reliable information on weakly bound complexes can only be obtained at the post-Hartree-Fock le~el,~?~ the present calcu- lations were performed at the MP2/6-31 + G** level. Also, to complete the body of experimental data, in this study new IR experiments were carried out, studying solu- tions in liquefied argon of CD,Cl-HCI and C,D,Cl-HCl mixtures at higher concentrations than those used before. 'v2 These experiments were interpreted using the results from the ab initio calculations. (a1 (b1 .H-CI ,H-CI , ,H-CI' , R-Cl: .R--Ci *H-CI Fig. 1 Possible equilibrium geometries for the (CH,Cl). (HCl), van der Waals complex: (a) chain structure, (b) bifurcated structure Experimental Computational Details For both the chain structure and the bifurcated structure the MP2/6-31 + G** equilibrium geometry was calculated using the GAUSSIAN 92 program,* as implemented on an IBM RS/6000 workstation. For all calculations, the correlation energy was calculated using all molecular orbitals, while the Berny geometry optimization was used with the tight con- vergence criteria. Furthermore, no restrictions due to a pos- sible symmetry of the species were imposed. For both structures, the vibrational frequencies and the corresponding IR intensities were calculated using standard harmonic force fields.The latter were obtained by calculating the numeric second derivatives of the energy with respect to the Cartesian coordinates using the analytically determined first derivatives. Synthesis and Spectroscopy The sample of CD,C1 was synthesized by mixing CD,OD (Janssen Chimica 16.635.48) and a small amount of PCl, at room temperature. C,D,Cl was synthesized by adding an amount of C,D,I (MSD Isotopes, MD-211) to an excess of dry, fresh AgCl. After 48 h, the reaction mixture containing both C,D,CI and C2D,I was collected by bulb-to-bulb distil- lation. The hydrogen chloride was made in small amounts by hydrolysing PCl, with water, and was purified afterwards by pumping the reaction mixture through a propan-2-01 slush. All compounds were purified on a low-pressure, low-temperature fractionation column.The argon used in this study has a stated purity of 99.9999%, and was used without further purification. All spectra were recorded using a Bruker 113v Fourier- transform spectrometer, equipped with a Globar source, Ge/KBr beamsplitter and an LN, cooled broadband MCT detector. The interferograms of the mid-IR spectra, recorded at a resolution of 0.5 cm-', were averaged over 200 scans, Happ Genzel-apodized and Fourier-transformed using a zero-filling factor of 4.A detailed description of the liquid noble-gas setup was given in a previous study' and will not be repeated here. Results and Discussion Equilibrium Geometry CH,Cl.HC1. HCl A starting geometry for CH,Cl. HCI .HC1 was obtained by combining the structural parameters of CH,Cl* HCl and Fig. 2 MP2/6-31 +G** equilibrium geometry of CH,Cl. HCl .HC1 HCl HCl." At the MP2/6-31 +G** level, these complexes are planar, i.e. the dihedral angle zX-Cl...H-CIis equal to 0". Therefore, for all calculations on CH3C1-HC1- HCl, the initial value for the dihedral angles zR-CI...H-CIand z~-~~...~-~~was set at 0".For a complete description of the relative orientation of the three molecules, the value of the dihedral angle zcl.. .H-Cl.. .H must also be given. Because at the outset no information was available on this angle, several initial geometries with a different value for zCI...H-CI...Hwere refined using standard convergence criteria.All optimizations resulted in the same equilibrium geometry, which was subse- quently refined using the tight convergence criteria. The equi- librium geometry obtained is shown in Fig. 2, and its structural parameters are given in Table 1. As can be seen in Fig. 2, the equilibrium geometry of CH,Cl* HCl HCl is cyclic. This structure suggests an inter- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 action between H(4) and Cl(9). Such an interaction was observed in (CH,),CO * HC1 ''and in CH3C1 HC1,' and its occurrence in CH,C1 * HCl HC1 is therefore not surprising. As described before,8 during the formation of RC1-HCl, the structural parameters of the methyl group in CH,Cl are hardly affected. In contrast, Table 1 shows that when a second HC1 molecule is added, the distortion of the methyl group strongly increases.This proves the existence of a weak interaction C-H. .C1-H. In Table 1, the structural parameters of CH,Cl.HCl and HCI-HCl and of the monomers are also given. The data show that the hydrogen-bond lengths Cl(2). * aH(6) and Cl(7)-+ sH(8) in CH3Cl* HCl HCl are both shorter than the corresponding bond in CH,Cl.HCI. As a decrease in bond length reflects an increase in bond strength, these differences neatly demonstrate the cooperative effect" in the 1 :2 species. This effect also explains the increased H-C1 bond lengths in the 1 :2 complex compared with the corresponding bonds in CH3C1 -HCl and HC1. HCl. During the formation of CH3C1.HCl, the C-Cl bond lengthens by 0.0036 A, indicating a weakening of the bond.In agreement with this, the C-C1 stretch of the complex is observed at a lower frequency than that of the monomer.'.' Adding a second HCI to form the chain complex, the C-Cl bond length increases by another 0.0027 A. Thus, the C-Cl stretch of the 1 :2 complex must again be shifted to lower frequencies, by an amount of the same magnitude as that of the first shift. Hence, this stretching mode is expected to be observed separately in the vibrational spectra. It is clear from Table 1 that the dihedral angles zR-.H -c1 and zH-cl.. .H -c1 have values significantly different from 0. Consequently, the chain complex is not planar. In this respect, the 1 :2 complex obviously differs from CH,Cl- HCl.Table 1 MP2/6-31 +G** structural parameters of CH,Cl. HCl, HCl *HCI, CH,Cl. HC1. HCl and CH,Cl(. HCI), monomer species 1 :1 species 1 :2 species internal coordinate" CH,Clb HClb HCl .HCl' CH,Cl. HClb CH,Cl. HC1* HCld CH,CI(- HCl),d ~~ ~ bond length/A rCC( 1 )-C1(2)1 ~[Ic(1)-~(3)1 ~CC(1)- ~(411 m1)-~(5)1 rm-c1(7)1 r[~(8)- ~~911 r[C1(2)-..H(6)] r[C1(7). ..H(8)] 1.7761 1.0836 1.0836 1.0836 1.2689 1.2689 1.7807 1.0832 1.0835 1.0835 2.4634 1.2700 1.2746 2.5926 1.2714 1.7834 1.0830 1.0827 1.0834 2.4102 1.2778 2.5348 1.2739 1.7853 1.0833 1.0831 1.0831 2.4957 1.2733 bond angle/degrees L [H(3)-C(l)-C1(2)] L [H(4)-C(l)-C1(2)] L [H(5)-C( l)-C1(2)] L[C(l)-C1(2)...H(6) 108.9 108.9 108.9 108.5 108.8 108.8 93.4 108.14 108.83 108.69 100.77 108.71 108.40 108.40 92.68 L[C1(2)..-H(6)-C1(7) 166.0 167.92 167.27 L [H(6)-C1(7)* **H(8)] 105.0 90.29 L[C1(7). ..H(S)-C1(9)] 174.6 162.93 z[H( 4) -C( 1 )-C1( 2) -H( 3)] z[H(S)-C( l)-Cl(2)-H(3)] 120.0 -120.0 119.9 -119.9 120.06 -119.64 120.08 -120.08 z[H(3)-C( 1)-C1(2). .-H(6)] 180.0 186.20 64.36 ~[C(l)-C1(2). .-H(6)-C1(7)] 0.0 22.93 -1.60 ~(Cl(2)e..H(6)-Cl(7). .*H(8)] 15.89 z[H(6)- C1( 7). ..H( 8)- C1( 9)] 0.0 -12.00 dipole moment/D' 2.04 1.46 2.57 2.78 1.98 2.37 E/Eh BSSE/ h -499.396666 --460.218345 -920.4401 13 -959.620436 -0.001 704 -0.002477 -1419.844747 -0.005320 -1419.843449 -0.005013 -0.001719 -0.002921 -0.00607 1 -0.005080 AE/kJ mol-' -4.51 -7.65 -15.94 -13.36 Atom numbering as defined in Fig.4. Taken from ref. 8. 'Taken from ref. 10. This study. 1 D (Debye) z 3.335 64 x C m. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 CH,Cl(. HCl), In order to find theoretical support for the possible occurrence of a bifurcated 1 :2 complex MP2/6-31 + G** calculations were also carried out. The resulting equilibrium geometry is shown in Fig. 3 and the structural parameters are given in Table 1. From Table 1 it can be seen that on going from CH,Cl.HCl to CH,Cl(.HCl),, the H-Cl bond length decreases from 1.2746 to 1.2733 A. At the same time, the C1. .H hydrogen-bond length increases from 2.4634 to 2.4957 A. This shows that when a second HCl molecule is attached to the chlorine atom of CH,Cl, the hydrogen bond loses some of its strength.Also, the addition of a second HC1 causes the C-Cl bond length to increase further from 1.7807 to 1.7853 A. The resultant C-Cl distance in CH,Cl(. HCl), is somewhat larger than that in the linear complex. Finally, when forming the bifurcated complex from monomers, the structural parameters of the methyl group are hardly affected, showing that in this complex, in contrast to the linear one, there is no direct interaction between the methyl grouping and the HC1 molecules. Relative Stability It has been demonstrated before' that the contribution to the energy of a van der Waals complex due to the basis set super- position error, AEBSSE, is of the same order of magnitude as the difference between the energy of the complex and the energies of the monomers.In order to take this into account, the energy of complexation, AE, is defined as: AECH3CI.HCI.HCI = ECH~CI.HCl.HCI -ECH3Cl -2EHCl -AEBSSE (1) Using the full counterpoise correction method,' the value for AEBSSE, was estimated to be -0.005320 E, (Hartree). Using this value, and the energies of the species involved, from the above equation the complexation energy AE for the chain complex is found to be -0.006071 E, (-15.94 kJ mol-'). The influence of the cooperative effect on the structural parameters has been noted above, and it is of interest to see what the impact of this effect is on the energy of the complex. The stabilisation due to the cooperative effect, AEcoop,can be estimated from the expression: AEcoop = AECH3Cl.HCl.HCl -"CH3CL.HCl -AEHCl.HCl (2) Using this equation, the cooperative effect, AEcoop,was calcu- lated to be -0.001413 Eh (-3.71 kJ mol-I).The complexation energy of the bifurcated complex is derived from an expression similar to the one for the chain structure: -AECH3CI(.HC1)2 -ECH3Cl(.HC1)2 -ECH3Cl -2EHCl -AEBSSE (3) In an analogous way AEBs,, for the complex was estimated to be -0.005013 E,. With the above expression, the complex- ation energy of CH,Cl(.HCl), is then found to be equal to -0.005080 E, (-13.36 kJ mol-I). Comparison with the result for the chain structure shows that the latter is the more stable of the two. The energy difference between them is 2.58 kJ mol-l. If this calculated energy difference is believed to be an acceptable measure of the free enthalpy difference between the two types of complexes, in an environment in which the chain structure is present in a reasonable concentration, the bifurcated complex will be present in a much lower, but in principle measurable, concentration.By considering that both hydrogen bonds in CH,Cl(.HCl), are weaker than that in CH,Cl.HCl, the complexation energy of the former species is expected to be smaller than twice the value of the complexation energy of CH,Cl, thus showing an anti-cooperative effect. Using the complexation energy of CH,C1 * HCl obtained bef~re,~ the energy of CH,Cl(. HCl), calculated in this study confirms this anti-cooperative effect. Vibrational Spectra For both types of complexes a vibrational analysis was made in terms of a symmetry force field. The chain complex has no elements of symmetry, and no subdivision of its normal coor- dinates based on symmetry can be made.The point-group symmetry of CH,Cl(.HCl), is C,, and the normal coordi- nates divide into 12 having A' and 9 having A" symmetry. Symmetry coordinates were constructed from the internal coordinates defined in Fig. 4 for the chain complex, and in Fig. 5 for the bifurcated complex. The symmetry coordinates are given in Tables 2 and 3, respectively. These coordinates Fig. 4 Internal coordinates for CH,Cl. HCl * HCl Ir12 Fig. 3 MP2/6-31 + G** equilibrium geometry of CH,Cl(- HCl), Fig. 5 Internal coordinates for CH,Cl(* HCl), J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Symmetry coordinates for the CH3Cl. HCl .HCI van der Table 4 MP2/6-31 + G** vibrational frequencies, IR intensities Waals complex and potential energy distributions for the CH3C1. HC1. HCl van der Waals complex approximate description symmetry coordinate" IR intensity CH, symmetric stretch sl = '13 + '14 + '15 C/cm -/km mol -P.E.D." CH, asymmetric stretch S, = 2r15-~13-'14 CH, asymmetric stretch s3 = '13 -'14 3299.5 0.5 C-Cl stretch s4 = '12 3289.7 3.0 CH, asymmetric deformation s5 = 2P34 -P35 -P45 3 170.6 15.9 CH, asymmetric deformation s6 = 835 -845 3056.6 195.4 CH, symmetric deformation s7 = 834 + P35 + P45 2998.7 309.7 1531.7 6.7-a32 -'42 -'52 CH, rocking S8 = 20152-'32 -'42 1527.1 7.0 1464.2 17.6CH, rocking s9 = '32 -'42 C1.. .H stretch s10 = '26 1084.7 3.3 H-Cl stretch st, = '67 1079.5 2.2 C-CI. . .H in-plane deformation s12 = e16 766.5 27.2 C1.. .H-Cl in-plane deformation 't3 = '27 434.5 38.1 CH, torsion st4 = '512 357.0 50.6 C-CI.. .H-CI torsion st, = '26 275.7 62.0 C1.. .H stretch '16 = '78 263.6 31.3 H-Cl stretch '17 = '89 107.6 9.7 C-CI. . .H in-plane deformation '18= '68 90.8 1.4 C1.. -H-Cl in-plane deformation '19 = '79 80.8 3.9 C1.. .H-Cl. . .H torsion '20 = '67 61.3 1.o H-Cl.. .H-Cl '2, = '78 29.3 1.1 17.1 2.1 Not normalised. " Symmetry coordinates defined in Table 2. were used to transform the Cartesian force field by GAUSS- IAN 92 into a symmetry force field.Applying Wilson's FG Table 4 shows that CH,Cl- HCl * HC1 is characterized by matrix method14 then led to the vibrational frequencies and two well separated fundamentals in the vHCl region, at 2998.7 potential-energy distributions. These are collected in Tables 4 and 3056.2 cm- ',respectively. The C-Cl stretching funda- and 5. mental of the complex is predicted to appear at 766.5 cm- ', From the vibrational eigenvectors, and using the Cartesian which is lower than the frequencies of 772.0 and 781.9 cm-' dipole derivatives produced by GAUSSIAN 92, the IR inten-calculated for CH,Cl. HC1 and CH,CI, respectively.' For the sities of the fundamental transitions were calculated. The methyl vibrations small frequency shifts have been calculated intensities have been included in Tables 4 and 5.on going from CH,Cl to CH,CI.HCl.' It can be seen in Table 4 that a similar effect is predicted for CH,Cl. HCl * HCl. For example, the antisymmetric CH, Table 3 Symmetry coordinates for the CH3C1(. HCl), van der stretches of the latter are calculated at 3299.5 and 3289.7 Waals complex approximate description symmetry coordinate" Table 5 MP2/6-31 + G** vibrational frequencies, IR intensities and potential-energy distributions for the CH3CI(. HCl), van der A' Waals complex CH, asymmetric stretch CH, symmetric stretch IR intensity CH, asymmetric deformation i/cm -/km mol-' P.E.D." CH, symmetric deformation A' CH, rocking 3298.8 1.02 C-Cl stretch 3170.1 13.71 H-CI stretch, in-phase 3062.1 63.5 C1...H-Cl in-plane deformation, 1524.0 8.4 in-phase 1462.7 12.6 C-CI. . .H-Cl torsion, in-plane '9 = '26 + '28 1079.3 2.3 C1.. .H stretch, in-phase '10 = '26 + '28 762.9 27.1 C-Cl. . .H in-plane deformation, '11 = '16 + '18 321.7 79.5 in-phase 269.6 20.3 H.. .C1.. . H bending s12 = x 78.6 0.6 28.2 4.4A" 11.1 0.7CH, asymmetric stretch '13 = '15 -'14 CH, asymmetric deformation = 834 -A"'14 835 CH, rocking 3295.1 1.5'15 = '32 -'42 H-Cl stretch, out-of-phase 3056.1 324.8'16 = '67 -'89 CI-. .H-Cl in-plane deformation, 1522.0 2.4'17 = '27 -'29 out-of-phase 1081.5 2.7 C-Cl.. .H-C1 torsion, out-of-plane S,, = T~~ -zZ8 303.9 22.9 C1.. .H stretch, out-of-phase '19 = '26 -'28 299.3 62.4 C-Cl. . .H in-plane deformation 110.0 8.1'20 = '16 -'tR out-of-phase 63.2 45.4 CH, torsion s2, = '12 45.4 5.9 " Not normalised." Symmetry coordinates defined in Table 3. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 cm-', while at the same level they are calculated at 3281.9 cm-' for CH,Cl. ' The corresponding fundamentals for CH,CI. HCl were calculated at 3290.7 and at 3287.2 cm- '. Thus, the frequency splitting of these modes in the chain complex, 9.8 cm-', is much larger than that in the 1 : 1 complex. This again illustrates the increased perturbation of the methyl group in CH,CI. HCl * HCl as a consequence of the cyclic nature of the structure. The H-CI stretches of CH,CI(.HCI), appear as a sym- metric and an antisymmetric combination. The asymmetric stretch is calculated to be at 3056.1 cm-', while the sym- metric combination is calculated to be at a slightly higher frequency, at 3062.1 cm-', with an intensity which is less than 20% that of the antisymmetric stretch.The vccl funda-mental of CH,Cl(- HCl), is calculated to be at 762.9 cm- ',i.e. at a frequency somewhat lower than the 766.5 cm-' obtained for CH,Cl. HCl HC1. This is in agreement with the longer C-C1 bond length in the bifurcated complex. Small fre- quency shifts were also calculated for the vibrational modes localized in the methyl group. The shifts for the bifurcated structure vary between 1.3 and 2.2 cm-', which is substan- tially less than some of the shifts calculated for the chain structure. Comparison with Experimental Results No direct structural study of the 1 : 2 complex between methyl chloride and hydrogen chloride has been published.Therefore, the identification of the 1 : 2 complex observed in liquefied argon at the present stage has to rely on a compari- son with ab initio data. The experimental values for the enthalpies of complexation of the 1 : 1 and 1 : 2 complexes have been published.' These data show a cooperative, and not an anti-cooperative effect. As it is unlikely that the thermal and solvent influences that are included in the enthalpy differences differ strongly between the 1 : 1 and the 1 : 2 complex, the observed coo- perative effect strongly suggests that in liquefied argon the chain complex is formed. Supplementary evidence for the chain structure of the observed 1 : 2 complex can be obtained from a detailed com- parison of the calculated and the experimental IR spectra.Even when the absolute values of the calculated vibrational frequencies are not as accurate as desired, experience suggests for instance, from conformational analysis, that the relative positions of the calculated frequencies in general are reliable. The situation with calculated IR intensities is somewhat less favourable, as even relative intensities sometimes differ from the experimental values by an order of magnitude, for reasons not always understood. Consequently, in the analysis below emphasis will be placed on the comparison of fre- quencies. From the discussion above it is clear that the ab initio cal-culations predict a doublet in the H-Cl stretching region for each of the complexes, with substantial differences in fre- quency splitting and relative intensity of the components.In Fig. 6, the H-Cl stretching region of the experimental IR spectra of two different CD,Cl-HCl mixtures in liquefied argon are compared with the spectrum of a solution contain- ing only HCl. A previous analysis of the spectra showed that only one band in the H-Cl stretching region could be attrib- uted to a 1 : 2 complex.'*2 Therefore, considerations based on the above differences cannot be exploited to identify the nature of the experimentally observed complex. Comparison of the vibrational frequencies obtained here with those calcu- lated for the 1 : 1 complex,' shows that both components of the vHCl doublet of the bifurcated complex are calculated at a higher frequency than the vHCl of the 1 : 1 complex, while for 3605 I 1 I I I 1 :1 3000 2800 2600 i/cm-Fig.6 H-C1 stretching region of the IR spectra of CD,Cl-HCl mixtures dissolved in liquefied argon. (a) 1.5 x lo-, mol dm-, CD,CI and 50.0 x lo-, mol dmP3 HCI, 95 K; (b) 1.5 x lo-, mol dn1-3 CD,Cl and 2 x lo-, mol dm-, HCI, 95 K; (c) 2.0 x lop3 mol dm-3 HCI, 96 K. The bands marked D and T are due to the HCI dimer and trimer, respectively. The bands marked with 1 : 1 and 1 : 2 are assigned to the CD,Cl. HCl and CD,Cl. (HCI), complexes, respectively. the chain complex one component is predicted on the high- frequency side, and the other on the low-frequency side of the CH,Cl.HCl frequency. In the experimental spectra, the H-Cl stretch due to a 1 : 2 complex in all cases was observed at a frequency lower than the HC1 stretch of the 1 : 1 complex.'*2 This suggests that the observed band is due to the complex having the chain structure. This, of course, is in agreement with the conclusion drawn from the cooperative effect. The ab initio calculations predict that the chain structure must have a second absorption band, with a slightly weaker intensity, in the HC1 stretching region. In the search for an experimental band due to the second HCl stretching, the fre- quency predicted for this vibration cannot be used without rescaling, as becomes clear from a comparison of the predict- ed and observed frequencies of the H-Cl stretches.The H-C1 stretches predicted at the same level for CH,Cl* HCl and for KCl are 3040 and 3120 cm-', respectively, and the experimental values are 2765 and 2869 cm-'. The ratio of experimental to calculated frequency for both molecules pro- duces a frequency scaling factor of ca. 0.91. This factor leads to a frequency of 2681 cm-' for the low-frequency H-Cl stretch of the 1 : 2 complex, which is experimentally observed at 2713 cm- '. This discrepancy suggests that the frequency of the other H-C1 stretch calculated with this scaling factor, 2788 cm- presumably is also not very accurate. Therefore, another estimate of the frequency of this mode was made in the following way.Recently, the HCl stretching frequencies of several €3. HCl and B .HaC1. HbC1 complexes observed in solid matrices have been analysed." In that study, the observed frequencies were expressed through their difference, A?, from the frequency of monomer HCl in the same environ- ment. It was found that AVa and A;,, the shifts of vHScI and vHbCl,respectively, can be expressed as a second-degree poly- nomial in AFS, the frequency shift observed for the 1 : 1 complex B.HC1. As the environmental influence of a solid argon matrix will not be too different from that of an argon solution, it may be hoped that the results of the matrices will J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I I I I 1 RCI*HCI*HCI 1 'I I I I.RCI I RCI*HCI 710 700 690 680 670 ilcm-Fig. 7 Mid-IR spectra of CD,Cl-HCl mixtures dissolved in lique- fied argon. (a) 1.5 x mol dm-3 CD,Cl and 50.0 x lo-, mol dm-, HCl, 108 K; (b)1.5 x lop3rnol dm-, CD,Cl and 2.0 x lo-, mol dm-3 HCl, 90 K; (c) 1.5 x lo-' mol dm-, CD,Cl and 2.0 x mol dmP3 HCl, 105 K. be applicable to the solutions studied here. With the poly- nomials obtained in ref. 15, and using the value of 104 cm-' for A?, observed for CH,Cl.HCl,' A?a and Aqb for CH,Cl. HCl HCl are calculated to be 152.5 and 55.0 cm-'. This brings the predicted frequencies at 2716.5 and 2814 cm-'. The former agrees quite well with the experimentally observed value of 2713 cm-', and it is assumed that the value of the other frequency has a similar accuracy.Note that the frequency predicted by this method is substantially higher than the rescaled ab initio value. In an attempt to detect experimentally the high-frequency H-Cl stretch of the 1 :2 complex, new HCl-CD,CI solu-tions in liquefied argon were investigated. The goal of these experiments was to increase the concentrations of the com- plexes. This can be achieved by increasing the concentrations of the monomers. Unfortunately, owing to the limited solu-bility of CD,Cl, the concentration of this monomer could not be increased compared with the previous study.' Hydrogen chloride, on the other hand, is quite soluble, and in this study concentrations of up to 50.0 x lo-, mol dm-, have been used. The region of the spectra in the vicinity of the frequency of 2814 cm-' was carefully checked, but no bands attribut- able to the 1 :2 complex could be identified.One reason for this is the complexity of the spectra in this region of solutions containing considerable concentrations of HCl, due to the presence of dimers, trimers and higher oligomers of HCl itself.16 This complexity is obvious from Fig. 6(a). Also, it is likely that, in agreement with the ab initio predictions, Table 4, the intensity of the high-frequency vHCl is somewhat lower than that of the low-frequency stretching, so that in the spectra of the solutions the band escapes detection. 1 I I I I 1 I I I!II I RCI*HCI*HCI I I* 1 I1 I RCI I I RCI*HCI 630 620 610 600 i/cm-' Fig. 8 Mid-IR spectra of C,D,Cl-HCl mixtures dissolved in lique- fied argon.(a) 2.0 x loT3mol dm-3 C,D,Cl and 45.0 x lo-, mol dmP3 HCl, 103 K; (b)2.0 x mol dm-, C,D,Cl and 3.5 x mol dmP3 HCl, 104 K. In addition to the band observed in the vHCl region, the complex CD,Cl -HCl also gives rise to a well separated iso- topic doublet in the vccl region. Although a similar doublet was calculated for the 1 : 2 species, in a previous study' no such doublet was observed owing to the relatively low con- centration of CD,Cl* HCl -HCl and to the medium IR inten- sity of this fundamental. Therefore, the IR spectra recorded in this study were also carefully checked in the C-C1 stretching region. In Fig. 7, this region of a solution containing 2.0 x lo-, mol dm-, CD,Cl and 50.0 x lo-, mol dm-, HCl, recorded at 108 K [Fig.7(a)], is compared with that of a solution containing 1.5 x lo-, mol dm-, CD,Cl and 2.0 x lo-, mol dmP3 HC1, recorded at 90 K [Fig. 7(b)] and at 105 K [Fig. 7(c)]. On the low-frequency side of the monomer bands at 697.5 and 691 cm-', two bands can be observed at 689 and 683 cm-' [Fig. 7(b)]. These have been assigned to the CD,Cl. HCl complex.' Next to these bands, in the region between 677 and 688 cm-' several bands of solid CD,Cl can be observed. In the spectrum of the more concentrated solution, Fig. 7(a), apart from the bands due to CD,C1 and CD,Cl- HCl, two weak bands are detected at 685 and 679 cm-', while in this spectrum no bands due to solid CD,C1 are observed. The 685 and 679 cm-I bands therefore are assigned to the C-C1 stretching fundamentals of the 1 : 2 complex. The ab initio calculations above show that on going from CH,Cl to CH,Cl * HCl, the C-,'Cl stretching funda- mental undergoes a red shift of 9.9 cm-'. In contrast, the red J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 shift induced by the addition of a second HC1 molecule is predicted to be only 5.5 cm-'. Taking into account an average scaling factor for stretching frequencies, these values agree quite well with the experimental shifts of 8.5 and 4.0 cm-',respectively, deduced from the data above. The behaviour of C,H,Cl dissolved in liquefied argon with respect to the formation of 1 : 1 and 1 : 2 complexes with HCl was found to be similar to that of methyl chloride.2 There- fore, it is reasonable to expect that in the IR spectra a vccl doublet due to the 1 :2 complex must be present.As such a doublet was not observed in the previous study, in this study the spectra of C,D,Cl-HCl mixtures dissolved in liquefied argon were re-investigated at higher concentrations of HCl. In Fig. 8, the C-Cl stretching region of the mid-IR spectrum of a liquefied argon solution containing 2.0 x low3mol dm-3 C,D,Cl and 3.5 x lop3mol dmP3 HCl is compared with that of a solution containing 2.0 x lop3 mol dmP3 C,D,Cl and 45.0 x lop3 mol dmP3 HCl, recorded at the same temperature. In the former spectrum, Fig. 8(b), four bands belonging to either C2D,Cl or C,D,Cl.HCl are observed at 622, 617, 613, 608 cm-'.When a large excess of HCl is added to the solution [Fig. 8(a)], in agreement with the behaviour of the CD3Cl solutions, two new weak bands appear at 610 and 605 cm-'. By analogy with the results described above, these bands are assigned to the C-35Cl and the C-37Cl stretching fundamental in the 1 : 2 species C2D,C1 * HCl * HCl. Conclusions In this study, MP2/6-31 + G** calculations have been made on two different structures of the complex formed between methyl chloride and two hydrogen chloride molecules. The chain structure CH3C1-HCl.HCl is found to be cyclic, and is more stable, by 2.58 kJ mol-', than the bifur- cated structure CH3Cl(. HCl), . The experimentally observed cooperative effect in the enthalpy difference for the 1 : 2 complex formed in liquefied argon shows that the complex must have the chain structure.By comparing the pattern observed in the H-Cl stretching region of the IR spectra of CD3C1-HCl mixtures dissolved in liquefied argon with the ab initio force field calculations, it is concluded that the observed band is due to the complex with the chain structure. At higher concentrations of HC1, in the C-C1 stretching region of the IR spectra of CD3Cl-HC1 mixtures in liquefied 3 607 argon, the isotopic vccl doublet due to the 1 :2 complex is detected. Its shift from the monomer frequency is in agree- ment with the ab initio calculations for the chain complex. Finally, in the IR spectra of C,D,Cl-HCl mixtures dis- solved in liquefied argon, the vccl isotopic doublet of the 1 : 2 complex has been observed.The Nationaal Fonds voor Wetenschappelijk Onderzoek (NFWO) (Belgium) is thanked for financial help towards the spectroscopic equipment used in this study and for a grant to W.A.H. References 1 W. A. Herrebout and B. J. van der Veken, J. Phys. Chem., 1993, 97,10622. 2 W. A. Herrebout and B. J. van der Veken, J. Phys. Chem., 1994, 98, 2836. 3 A. Schriver, B. Silvi, D. Maillard and J. P. Perchard, J. Phys. Chem., 1977,81,2095. 4 A. Schriver, A. Louteillier, A. Burneau and J. P. Perchard, J. Mol. Struct., 1982,95, 37. 5 G. Maes and M. Graindourze, J. Mol. Spectrosc., 1985, 113,410. 6 S. Scheiner, in Theoretical Treatment of Large Molecules and Their Interactions, ed. Z. B. Maksic, Springer-Verlag, Berlin, 1991. 7 S. Scheiner, in Reviews in Computational Chemistry II, ed. K. B. Lipkowitz and D. B. Boyd, VCH, New York, 1990. 8 GAUSSIAN 92, Revision E.3, M. J. Frish, G. W. Trucks, M. Head-Gordon, P. M. W. Gill, M. W. Wong, J. B. Foresman, B. G. Johnson, H. B. Schlegel, M. A. Robb, E. S. Replogle, T. Gom- perts, J. L. Andres, K. Raghavachari, J. S. Binkley, C. Gonzalez, R. L. Martin, D. J. Fox, D. J. Defrees, J. Baker, J. J. P. Stewart and J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1992, 9 W. A. Herrebout, B. J. van der Veken and J. R. Durig, Theo-chem., accepted for publication. 10 W. A. Herrebout and B. J. van der Veken, unpublished results. 11 X. K. Zhang, E. G. Lewars, R. E. March and J. M. Parnis, J. Phys. Chem., 1993,97,4320. 12 S. Scheiner, J. Mol. Struct. (Theochem.), 1989, 202, 177. 13 S. B. Boys and F. B. Bernardi, Mol. Phys., 1970,19,553. 14 E. B. Wilson Jr., J. C. Decius and P. C. Cross, Molecular Vibra- tions, McGraw-Hill, New York, 1955. 15 Th. Zeegers-Huyskens, J. Mol. Struct., 1993,297, 149. 16 B. J. van der Veken and F. R. de Munck, J. Chem. Phys., 1992, 97, 3060. Paper 4/04153D; Received 7th July, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003601

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3609-3616 Pressure and Temperature Dependence of the Rate Constants for the Association Reaction of OH Radicals with NO between 301 and 23 K Paul Sharkey, Ian R. Sims and Ian W. M. Smith School of Chemistry, The University of Birmingham, Edgbaston, Birmingham, UK B 15 2TT Pascal Bocherel and Bertrand R. Rowe Departement de Physique Atomique et Moleculaire, U. A. 1203 du C.N.R.S., Campus de Beaulieu, Universite de Rennes I,35042 Rennes Cedex, France The pulsed laser photolysis (PLP), time-resolved laser-induced fluorescence (LIF) technique has been used to study the association reaction OH + NO( +M) -+ HONO(+M) at low and ultra-low temperatures. Using a cryoge- nically cooled cell, rate constants have been determined at temperatures down to 80 K and with M = N, over a range of total gas densities.The PLP-LIF method has also been implemented in the ultra-cold environment provided by the gas flow in a CRESU (Cinetique de Reaction en Ecoulement Supersonique Uniforme) apparatus yielding rate constants at temperatures as low as 23 K. At 52 K, rate constants have been measured at total gas densities from 5.1 x 10l6 to 8.2 x lo'' molecule ~m-~. The results reported here are the first for any neutral- neutral radical recombination at such low temperatures. The pressure and temperature dependence of the rate constants observed in the cryogenically cooled cell are fitted to obtain rate constants in the limits of low and high pressure given by the expressions: k"(T)[N,] = 8.9 x 10-31(T/298)-2*1[NZ.cm3 molecule-' s-' k"(T) z 5.4 x 10-"(T/298)-0.' cm3 molecule-' s-' The rate constants determined at ultra-low temperatures in the CRESU apparatus are used to derive values of limiting low-pressure rate constants for the combinations of temperature and third-body gas (Ar, N, and He) employed in these experiments. The rate constants for M = Ar at temperatures between 295 and 23 K are fitted well by the following expression for the limiting low-pressure rate constant: k"(T)[Ar] = 4.5 x 10-31(T/298)-2.6[Ar] cm3 molecule-' s-' The limiting low-pressure rate constants deduced from the experiments are compared with values calculated using the methodology of Troe (J. Chem.Phys., 1977, 66, 4758). The results are in fair agreement. Association reactions in which two free radicals combine con- stitute one of the major categories of gas-phase reaction and are important in a variety of environments, notably the earth's atmosphere.' Their kinetics, of course, depend on the total pressure, and the nature of the third-body gas, as well as on temperature.In some cases, the temperature dependence of the rate constants can be examined over a wide range by studying association at low temperatures and the reverse dis- sociation at high temperatures, applying the principle of detailed balance to coordinate the two sets of results. Reactions between two different unstable free radicals are difficult to study experimentally since both radicals must be produced in situ and because the absolute concentration of at least one of the radicals in the reaction zone must be known if the reaction rate constant is to be determined accurately. For this reason, and because of their importance in atmo- spheric chemistry, the reactions of OH radicals with the 'stable' free radicals NO and NO,, 0 + NO( +M) +HONO( +M), A,H(298 K) = -208.8 kJ mol-' (1) 0+ NO,(+M)-+HONO,( +M), A,H(298 K) = -207.6 kJ mol-' (2) have been studied extensively.' They serve as prototypical examples of association reactions between small free radicals whose kinetics at room temperature are between the high and low pressure limits in the experimentally convenient range of total pressure between 1 Torr and 1 atm.At the lower end of this pressure range, both discharge- flow and pulsed photolysis experiments can provide rate con- stants for these reactions.There have been extensive measurements of both kinds which provide accurate values for the rate constants (k") in the limit of low total pressure. Evaluating these data recently, Atkinson et al. le recommend the following expression for the limiting low-pressure rate constant for the association of OH and NO between 200 and 300 K:t k:(T)[N2] = 7.4x 10-31(T/300)-2.4[N,]cm3 molecule-' s-' Obtaining accurate values for the rate constant (k") for association in the limit of high pressure has proved more problematical.' In order to extend the range of the experi- mental data base, to provide a more comprehensive test of models of the pressure 'fall-off, and in particular to obtain a better estimate of k"', Robertshaw and Smith2 carried out measurements of the rate of reaction (2) at total pressures up to ca.10 atm. More recently, Troe and co-workers3 have spectacularly increased the total pressure range for which kinetic data for reactions (1) and (2)are available by carrying out room-temperature measurements of the rate in pressures of He up to ca. 140 atm. At the higher end of this range, the rate constants became independent of pressure yielding : ky(295 K) = 3.2 x lo-" cm3 molecule-'s-' ~ ~ t Although ref, l(e) states that the preferred value of k'; was not changed from that given in the 1984 evaluation,'c the index on the T/300 term is -2.4 in ref.l(e) rather than the value of -2.6 given in ref. l(c). This value is in excellent agreement with that of 3.8 x lo-" cm3 molecule-' s-l inferred by Smith and William~.~ They measured the rate constant for vibrational relaxation of OH(u = 1) by NO and arg~ed~-~ that this would correspond to the rate constant for formation of an energised and strong- ly bound HONO complex and that the rate of formation of this complex, which is also the first step in the association reaction, would be essentially independent of the vibrational state of the OH radical. Here, we report the results of experiments which extend the data base for .reaction (1) in a different way: namely, to much lower temperatures than in any previous study.Until very recently, there have been only three reports of kinetic studies at temperatures below 270 K. With their temperature ranges, these are the investigations of: Anastasi and Smith (233- 505 K),7 Anderson et a!. (230-450 K),8 and Lovejoy et al. (249-299).9 It is worth noting that none of these studies actually reached 200 K, the lowest temperature for which the expression for k"(T)given by Atkinson et a/.,'@) and repro- duced above, is claimed to hold. The practice of estimating rate constants by extrapolating from data obtained at higher temperatures must be viewed with caution." The results reported here include measurements down to what we refer to as 'low' temperatures (down to 80 K) and 'ultra-low' tem- peratures (below 80 K, and in this case down to 23 K).These two sets of measurements both employ the well established PLP-LIF technique, in which PLP of HNO, or H202 is used to generate OH radicals whose subsequent kinetic decays are observed by time-resolved LIF. The first set of experiments, at low temperatures, were carried out in a cryo- genically cooled tubular reactor that had been used in pre- vious low-temperature kinetic studies of reactions of and CN.12-14 In the second set of experiments, rate constants are measured at ultra-low temperatures in a CRESU appar- atus, in which gas cooling is achieved by expansion through a Laval nozzle. We have recently reported4-17 the results of a number of kinetic studies at ultra-low temperatures on reac- tions between neutral species using the CRESU technique but the current measurements on reaction (1) are the first on a pressure-dependent association reaction.As this paper was being written, Atkinson and Smith18 reported new measurements on the association of OH and NO at temperatures down to 90 K in experiments using a pulsed Laval nozzle. These experiments differ from our own in two major ways. First, and as already indicated, the nozzle incorporates a pulsed valve resulting in a considerable saving in pumping capacity. Secondly, the hydroxyl radicals are created in a discharge upstream of the throat of the nozzle and are observed by LIF using a dye laser which crosses the gas flow downstream of the nozzle. Changes in the concentra- tion of OH are observed as the concentraion of NO in the gas mixture is varied (and possibly as the distance between the exit of the nozzle and the observation point is altered).Atkinson and Smith's results are compared with our own later in this paper. Experimental The experimental methods used to achieve low and ultra-low temperatures in the present series of experiments have both been described previously.' ',l Consequently, we give only a brief, but self-contained, description of the apparatus and procedures for obtaining rate constants for reaction (1). In both sets of experiments, OH radicals were generated by laser pulsed photolysis, using the output at 266 nm from a frequency-quadrupled Nd : YAG laser. In the low-temperature experiments in cryogenically cooled cells, the radical precursor was HNO, ,whereas in the CRESU experi- J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 ments we used H202. In both cases, the OH radicals were observed by off-resonance LIF; that is, fluorescence was excited using a tunable dye laser tuned to a line band of the A2X+-X2n system of OH at ca. 282 nm and observed at ca. 310 nm in the (0, 0) and (1, 1) bands of the same band system. The cryogenically cooled cell was the same as that used in previous experiments on the reactions of OH with HC1, CH, and C,H6 lo and of OH with CO." Low temperatures were achieved by passing various cryogens through the inner jacket which surrounded the tubular reaction cell. A second outer jacket was evacuated to provide good thermal isula- tion.The temperatures which were achieved for different con- ditions of cooling were measured in separate, subsidiary experiments."." The beams from the photolysis and probe lasers counterpropagated along the axis of the reaction cell and LIF signals from OH were observed at right angles to the direction of the laser beams using a lens to gather the emission and a narrow-band interference filter centred at 312 nm to reduce spurious signals from scattered laser light and window fluorescence. The principles of the CRESU technique, its adaptation to the measurement of rate constants for reactions between neutral species by the PLP-LIF technique, and its applica- tion to reactions of the OH radical, have been described in '9'detail elsewhere.' Briefly, expansion through an axisym- metric Laval nozzle generates a supersonic flow of gas in which the Mach number, the temperature, the density of the gas, the mole fraction of the reagent in excess (here NO), and the velocity of the gas stream are constant along the axis of the flow.The temperature in the flow can be calculated from the Mach number and the reservoir temperature" and checked by observing the relative intensities of rotational lines within the band of an LIF spectrum.'5b Different nozzles must be fabricated to generate uniform flow for a given carrier gas, at a given density and temperature. This, and the lower and upper limits to the densities which can be used, constitute more of a handicap to rate measurements for pressure-dependent association reactions than for pressure- independent metatheses.In the CRESU experiments, the laser beams from the pho- tolysis and probe lasers were combined and co-propagated along the axis of the supersonic gas flow. Fluorescence from OH was gathered by an optically fast telescope-mirror com- bination mounted within the main vacuum chamber and directed onto a photomultiplier tube through an interference filter centred at 310 nm, (FWHM 10 nm). In both sets of experiments, the rate of loss of OH was observed by scanning the delay time between the pulses from the photolysis and probe lasers using a commercial delay generator. The signals from the photomultiplier were accu- mulated, processed and analysed by procedures which we have used and described1°-16 before.In every case, the varia- tion of the LIF signals with time was well fitted by a single exponential decay (using a standard non-linear least-squares algorithm), yielding a pseudo-first-order rate constant (k1J for removal of OH under the particular conditions in that experiment : i.e. concentration of NO, temperature and total pressure of the chosen diluent gas. Results and Discussion Rate Constants obtained from Measurements in the Cryogenically Cooled Cell between 301 and 79 K The kinetic experiments performed in the cryogenically cooled cell and the results obtained from these measurements are summarised in Table 1. At each selected total pressure J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 361 1 Table 1 Second-order rate constants for the association of OH and NO in the presence of N, obtained in the cryogenically cooled cell at temperatures between 298 and 80 K ~ ~ number of [NO]/10' k2&O-experiments molecule cm -cm3 molecule-' s-' 30 1 3.24 10 4.6-60.5 0.26 f0.01" 294 4.93 8 6.8-27.9 0.32 f0.02 296 8.15 8 3.5 -23.2 0.59 f0.01 298 17.3 12 2.7-23.8 1.09 f0.03 298 32.4 8 0.9-5.3 2.03 f0.06 297 63.3 7 0.6-7.8 3.40 f0.1 1 294 98.5 8 0.9-3.8 4.4 f0.5 294 164.0 8 0.7-2.1 7.3 f0.8 295 245.0 (0.40)b 8 0.2-0.8 22.1 f1.1 216 2.24 8 2.7-13.6 0.47 f0.02 216 3.13 8 3.5-16.1 0.64 f0.02 216 4.17 9 2.1-12.3 0.81 f0.03 216 6.70 9 2.3-10.9 1.25 f0.03 216 11.2 8 1.0-8.2 1.55 f0.06 216 17.9 8 0.7-4.5 2.91 f0.17 216 29.1 8 1.3-3.1 3.96 f0.22 216 44.7 7 0.2-1.2 7.2 f0.8 216 62.0 (0.20)b 8 0.3-1.2 9.7 f1.8 178 1.90 7 3.7- 10.1 0.71 f0.02 178 2.7 1 8 1.5-9.4 0.79 f0.03 178 3.80 7 2.0-8.4 0.97 f0.04 178 5.42 10 1.8-1 0.1 1.49 f0.06 178 8.14 8 1.2-4.8 1.91 0.09 178 13.6 8 0.6-4.2 2.67 f0.15 178 16.3 7 1.0-2.4 3.83 f0.23 178 21.7 (O.lO)b 6 0.3 -2.3 4.17 f0.26 138 1.40 8 3.3-10.0 0.82 f0.05 138 1.40 7 1.6- 10.1 0.63 f0.02 138 2.10 8 2.0- 13.6 0.79 f0.05 138 3.50 7 1.3-6.7 1.18 0.1 138 4.90 7 1.3-4.4 2.43 f0.19 138 4.90 8 1.9-8.6 1.63 f0.07 138 5.60 7 1 .O-4.1 1.70 f0.11 138 7.00 7 1.9-5.4 1.57 f0.32 138 7.00 8 1.2-4.7 1.81 f0.13 138 10.5 9 0.5-3.3 2.34 f0.25 138 14.0 (0.1 l)b 4 1.7-3.5 2.4 f0.85 80.1 1.23 6 0.7-4.2 2.3 f0.3 79.3 2.42 7 0.3-2.4 4.1 f0.3 80.5 2.41 10 0.3-2.3 2.9 f0.2 80.3 2.42 5 0.3-1.2 3.1 f0.35 80.2 3.62 7 0.4-1.3 4.3 f0.2 80.5 3.62 8 0.4-2.1 2.8 f0.3 80.5 3.62 7 0.4-1.8 3.5 f0.45 80.0 4.82 7 0.5-1.2 5.2 f0.6 80.0 6.03 (0.14)b 9 0.5-2.1 4.6 f0.7 a Quoted errors are single standard deviations in the gradients of the plots of k,,, us.[NO]. [M]/[M]1,2 where [MI,,, = k"(T)/k"(T)and the values of the limiting low-and high-pressure rate constants are calculated from eqn. (VIa) and (VIb). and temperature, several values of klst were measured at dif- association with NO. The principal contributions to such loss ferent concentrations of NO. A second-order rate constant under the conditions of our experiments are likely to have (k2,,J for these conditions of temperature and total pressure been reaction with HNO, and diffusion of OH from the was then determined by plotting the values of kist obtained in region probed by the dye laser.The rates of both processes, individual experiments us. [NO]. and that of any second-order loss of OH, depend on the con- Examples of the plots of klst us. [NO] from experiments in centration of HNO,, the total pressure, and the size of the the cryogenically cooled cell are given in Fig. I. These two observation region in a particular series of experiments, so plots have been chosen to represent the best and worst of the the intercepts on the plots of k,,, us. [NO] vary. However, those obtained in this part of our work.The gradients of such these conditions were kept constant for each individual straight-line plots yield the second-order rate constants for experiment within a given series, so that the same association of OH with NO which are listed in the last 'background' rate applied to each measurement and the gra- column of Table 1. The errors which are quoted are single dient was a true measure of kist us. [NO]. The initial concen- standard deviations in the gradients of the plots of kist us. tration of OH formed by photolysis of HNO, was too small "01. and the reaction of OH with HN02 too slow for the second- The intercepts of these plots at zero [NO] correspond to ary reaction between OH and HNO, to have a significant the rate of loss of OH radicals by processes other than their influence on the derived values of kZnd.in mz> r 15 .-v) m 10 > 5 L .I U I [N0]/1015 molecule ~rn-~ Fig. 1 Examples of the variation of pseudo-first-order rate con-stants for loss of OH with the concentration of NO measured in the cryogenically cooled cell: (a) [N2] = 2.73 x 1017 molecule ~rn-~, T = 178 K and (b)[Nz] = 6.1 x lOI7molecule cm3, T = 80 K All the experiments in the cryogenically cooled cell were carried out with N, as the diluent, third-body gas, as it has been the most popular choice in previous experiments at higher temperatures,' partly because of the possible atmo- spheric importance of the reaction. Particularly below room temperature, the range of total pressure in which reliable rate constants could be obtained in our experiments was limited: first, by the necessity for the gas flow to be slow enough to allow appropriate cooling but fast enough to minimise loss of HNO, by condensation on the cold wall of the reaction cell" and secondly, by the increased quenching of the fluo- rescence from OH at higher pressures of N, .A cursory examination of the results given in Table 1 shows that the second-order rate constants are strongly dependent on the total pressure but fall short of being pro- portional to [N2]. This result implies that rate constants have been obtained towards the low-pressure end of the 'fall- off region,,' a conclusion which is entirely consistent with previous measurement^.'.^-^ Given this fact, it was clearly necessary to construct fall-off curves to extract information about the values of the rate constants in the limits of low (k") and high (k") total pressure.Furthermore, because the range of total pressure covered in the experiments at any single temperature was limited, it was decided to adopt the 'global' fitting procedure proposed by Keifer et aL2' and based on detailed work by Troe2, concerned with the form of fall-off curves for radical association reactions. This global fitting procedure allows one to include, in the same fit, data obtained at different temperatures, by assuming, reasonably, that the temperature dependence of the limiting low-and high-pressure rate constants can be expressed by relationships of similar form : k"(T)= A"(T/298)"' (14 and J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 Fall-off curves are constructed according to the method developed by Troe2, in which the ratio k,,d(T)/k"(T)is written as the product of three terms: FLH ,Fsc and Fwc. k,,,(T)/k"(T) = F~~F~~F~~ The first of the terms on the right-hand side would be the only one present in the case where the simple Lindemann- Hinshelwood mechanism was operative. It can be written as FLH = {k"(T)CMI/k"(T))/{1 + k"(T)CMl/k"(T)) (111~) = (CMI/CM11,2)/(1 + CMI/CMI1,2) (IIIb) where [MI1/, = k"(T)/k"(T)is the total gas density at which the value of k2"d would reach half its limiting high-pressure value according to the Lindemann-Hinshelwood model.The factors Fsc and Fwc allow for broadening of the fall- off curves, beyond that predicted by the Lindemann-Hinshelwood expression, because of: (a) the energy spread in the energised collision complexes initially formed in collisions between the two radical reagents, and (b) the effect of 'weak' collisions, which arises because most third-body gases do not remove sufficient energy on every collision with an energised collision complex to stabilise it and prevent subsequent disso- ciation. In Troe's method,23 Fsc and Fwc both depend on [M]/[M]1/2. In addition, Fsc depends on two 'Kassel parameters' S, and B,, whereas Fwc depends on S, and p, the collisional efficiency of the particular third-body gas. The equations relating Fsc and Fwc to S,, B, and fl are given in ref.22. In the global fitting procedure, recommended by Keifer et aL21 and used by us to analyse our results, S, is 'pegged' to the value of nooby the equation S, = n" + 2 (IV) ~ by the equa- whilst B, is related to B' = (E + ~(E,)E,,)/~,T, tion BK = B'(SK -l)/(s -1) (V) where s is calculated22 from the numbers of oscillators, inter- nal rotors and active external rotors in the molecule, c0 and E,~are the dissociation and zero-point energies of the mol- ecule (HONO), and a(c0) is the Whitten-Rabinovit~h~~,~~ factor. The non-linear least-squares fitting program incorpor-ated a sub-routine which, given values of [MI, k"(T)and k"(T), returned values of k2,,d. The values of A", no, A" and n" were varied to minimise x2, the sum of the squares of the deviations between the experimental rate constants and the corresponding calculated values.Care had to be exercised in the initial choice of trial values of A", no, A" and n", since these parameters are strongly correlated and the program was prone to finding 'local minima' in x2 if the starting values of A", no, A" and n" were far from their best values. The initial estimates were based on the recommendations of Baulch et a!.' Fig. 2 shows two sets of experimental data with the fitted fall-off curves. The expressions for k"(T)and k"( T) resulting from this global fitting procedure are k"(T)[N,] = 8.9 x x (T/298)-2.1[N,] cm3 molecule-' s-' (VIa) k"(T) x 5.4 x lo-'' x (T)/298)-'.' cm3 molecule-' s-' (VIb) The expression of k"(T) must be viewed as approximate in view of the fact that all the experimental data were obtained at total gas densities far from those required to reach the J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 L h c -1, 2.0 7 I a,-30 -% 1.5 E m5 ; 1.0 -0 F1 CN s 0.5 -UJ -10 Of I Rate Constants obtained from Measurements in the CRESU Apparatus between 295 and 23 K The kinetic measurements performed in the CRESU appar--atus and the results obtained from these measurements are summarised in Table 2. For each set of conditions, a second- order rate constant was obtained by the same procedure as described in the previous section. An example of the plots of -kist us.[No] is displayed in Fig. 3. The number of measurements carried out in the CRESU apparatus was smaller than the number carried out in the -cryogenically cooled cell and only at one temperature (ca. 52.7 K) were several measurements carried out at different concentrations of diluent. As before, the measurements were -carried out towards the low-pressure end of the fall-off region. Estimates of [M]/[M] based on the expressions (VIa) and (VIb) for k"(T)and k"(T) derived from the low- temperature experiments and with due allowance for the I 15 10 r v) m 0 0 1 2 3 [N0]/1 Oi5 molecule ~m-~ I 1 17.0 17.5 18.0 18.5 log([N,]/molecule cm-3) Fig. 2 Examples of the variation of second-order rate constants for association of OH with NO with concentration of the N, 'third-body' gas: (a) T = 296 K and (b) T = 178 K. The experimental results are compared with 'fall-off curves calculated using the 'global' fitting procedure described in the text.[N0]/10'4 molecule ~m-~ high-pressure limit. The expressions for k"(T) and k"(T) Fig. 3 Examples of the variation of pseudo-first-order rate con-allow values of [M]/[M]l,F to be calculated for all our stants for loss of OH with the concentration of NO measured in the experiments. Those for the highest pressure used at each tem- CRESU apparatus: (a) [He] = 4.73 x 1OI6 molecule cmP3, T = 23.0 perature are recorded in Table 1. K and (b) [Ar] = 8.21 x 1017molecule mP3,T = 52.4 K. Table 2 Second-order rate constants for the association of OH and NO in the CRESU apparatus at temperatures between 295 and 23 K [M]/1017 number of [NO]/lO" k,,,J10-l2 molecule cm - experiments molecule cm - cm3 molecule-' s-' 29 5 295 6.78 (Ar) 9.42 (Ar) 7 7 7.5-29.0 7.2-28.1 0.30 f0.02" 0.31 k0.02 295 52.1 52.3 0.51 (Ar) 1.03 (Ar) 11.7 (Ar) (~0.02)~ 6 7 7 5.6-32.8 1.16-2.80 1.91-6.82 0.46 f0.05 1.63 k0.15 2.97 k0.24 52.8 2.02 (Ar) 8 0.39-3.08 5.18 f0.24 53.9 52.4 4.01 (Ar) 8.21 (Ar) ( < 0.44) 10 8 0.25-2.74 0.12-0.97 7.70 f0.24 14.8 k0.9 48.0 44.0 30.5 0.274 (N2) (0.02) 0.290 (Ar) (<0.02) 0.189 (Ar) (<0.03) 7 5 10 0.12-0.80 0.58-2.68 0.20-1.85 1.69 0.16 1.33 k0.30 2.14 k0.20 27.0 0.454 (Ar) (<0.09) 8 0.27-2.08 8.90 _+ 0.55 23.0 0.473 (He) (0.13) 8 0.43-3.27 3.50 +_ 0.23 " Quoted errors are single standard deviations in the gradients of the plots of k,,, us.[NO]. [M]/[M],,2 where = k"(T)/k"(T)and the values of the limiting low- and high-pressure rate constants are calculated from eqn. (VIa) and (VIb). lower collisional efficiencies of Ar and He compared with N, , are given in Table 2. For these reasons, the strategy adopted to treat the results from the CRESU apparatus was different from that used to analyse the data obtained using the cryoge- nically cooled cell. Our aim was to determine a value of k"(T)for each com- bination of third-body gas and temperature employed in the CRESU experiments. To this end, eqn. (11) and (IIIa) were used to relate k"( T) to k,,d( T): The values of k"(T)[M]/k"(T) in the case of M = N, were estimated using the expressions for k"(T)[M] and k"(T) given in eqn.(VIa) and (VIb); those for M = Ar and M = He were estimated assuming that k"( T)[He] : k"(T)[Ar] : k"(T)[N,] = 0.5 : 0.6 : l.O1" and was independent of tem-perature. The values of Fsc and Fwcwere estimated by the methods outlined above and described fully by Tree," assuming /3,, Ar = 0.20 and /3,, He = 0.10. Table 3 lists the estimated values of (1 + k"(T)[M]/k"(T)}, FSCand Fwc appropriate to the conditions of each experi- ment performed in the CRESU apparatus. These correction factors are all relatively small, so the errors introduced into the estimates of k"(T)by uncertainties in these factors should also be small.The last column of Table 3 lists the derived values of k"(T) for each combination of temperature and third-body gas employed in the CRESU experiments. Fig. 4 compares the experimental values of kind obtained at several concentrations of Ar at ca. 52.7 K with the calculated fall-off curve for M = Ar at this temperature. The agreement is satis- factory. Limiting Low-pressure Rate Constants and Comparisons with Theory Because of the restricted range of total pressure covered in both sets of experiments, the information derived about the limiting high-pressure rate constant for association of OH and NO and its dependence on temperature is only approx- imate. All that can be said is that the value which we have derived at room temperature is consistent with those derived from recent measurements at much higher pressures3 and deduced from measurements of the rate of relaxation of OH(v = 1) by NO.4 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 :;-3/:I/, , , , , ,\ ?--. CN -Y 0 2 4 6 8 [Ar]/lO" molecule ~m-~ Fig. 4 Second-order rate constants for association of OH with NO measured at ca. 52.5 K in the CRESU apparatus plotted against the concentration of Ar. The experimental results are compared with a calculated 'fall-off curve (see text). On the other hand, our determinations of the limiting low- pressure rate constants should be accurate and reliable and they cover a very wide range of temperature, especially as such rate constants are found to vary as 7'"'and our mea- surements cover over an order-of-magnitude variation in T.The expression given in eqn. (VIa) for k"(T)and resulting from the analysis of experiments performed between 301 and 79 K is similar to that recommended by Baulch et al.' for a more restricted range of temperature. These two expressions are compared diagrammatically in Fig. 5. Fig. 5 also displays the values of k"(T)deduced from our experiments at ultra-low temperatures in the CRESU appar- atus. Those obtained with M = Ar at five temperatures from 295 to 27 K are all well matched by the expression: k"(T)[Ar] = 4.5 x x (T/298)-'e6[Ar] cm3 molecule-' s-' The single value of k"(T)measured with M = N2 at 48 K is somewhat higher than that predicted by applying the expres- sion given in eqn.(VIa). This observation, together with the expression for k"(T)given by Baulch et al.,lesuggests that the application of the global fitting procedure to the results Table 3 Limiting low-pressure rate constants for the association of OH and NO at ultra-low temperatures deduced from experiments in the CRESU apparatus ~~ 295 Ar 0.30k0.02" 1.007 0.965 0.925 0.050 295 Ar 0.31 f0.02 1.0095 0.96 0.92 0.038 295 Ar 0.46f0.05 1.012 0.96 0.915 0.045 0.045 52.1 Ar 1.63f0.15 1.017 0.945 0.935 3.7 52.3 Ar 2.97 k0.24 1.033 0.93 0.92 3.5 52.8 Ar 5.18f0.24 1.064 0.905 0.89 3.4 53.9 Ar 7.70& 0.24 1.128 0.87 0.87 2.9 52.4 Ar 14.8 f0.9 1.262 0.82 0.815 3.4 3.4 48.0 N2 1.69 & 0.16 1.018 0.94 0.93 7.2 44.0 Ar 1.33 f0.30 1.013 0.95 0.94 5.2 30.5 Ar 1.48 f0.14 1.018 0.94 0.935 9.1 27.0 Ar 8.90f0.55 1.055 0.91 0.885 25.7 23.0 He 3.50f0.23 1.065 0.90 0.935 9.4 a Errors correspond to &to, where t is the appropriate value of the Student's t-distribution for the 95% point.J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1 \ a IF 10 100 TIK Fig. 5 Variation of the limiting low-pressure rate constants, k"(T), for association of OH with NO at room temperature and below. The solid line represents the expression obtained for k"(T)with M = N, from measurements in the cooled cell, and the points are the values derived from experiments in the CRESU apparatus with M = N, (O),Ar (0)and He (0).The dotted line represents the expression for k"(T) with M = N, recommended by Atkinson et d.le and the dashed line joins values of k"(T)estimated by the method of TroeZ4 for M = N,, assuming p = 1.obtained in the cryogenically cooled cell may have given a value of no which is slightly too small. Our results for M = N, are also in fair agreement with the low-pressure rate con- stants recently deduced by Atkinson and Smithla from their experiments using a pulsed Lava1 nozzle to generate tem- peratures down to 90 K. However, this agreement should be viewed with some caution, as they assumed that their mea- surements were carried out in the true low-pressure regime so they made no correction for fall-off effects.The analysis of such effects which we have described above indicates that Atkinson and Smith's values of ko(T)should be increased by between 5 and 14%. Values of the rate constants for radical-radical association reactions in the limit of low pressure can be estimated using the semi-empirical methods proposed by Tr~e.,~These methods are straightforward and now quite standard, and they will not be described here in detail. We have employed them chiefly to examine whether, for the association of OH with NO, theory would predict a depedence of ko(T)on T" over the complete range of temperatures covered in our experiments. In Troe's method24 one begins by estimating the density of harmonic vibrational states associated with the product mol- ecule at its dissociation limit and uses this quantity to esti- mate the low pressure rate constant [k"(T)] for the HON0,7,25 and OH.27 The rovibronic partition func- tions for OH and NO were evaluated by direct summation at low temperatures.The results of the calculations with T = 20,40, 80, 160 and 300 K and assuming strong collisions with N, (i.e.p, = 1) are shown in Fig. 5. It can be seen that the calculated rate constants approximately follow a T"' dependence with no * 2.3 at temperatures down to ca. 80 K, and that the difference between the calculated and experi- mental values are reasonably consistent with our earlier assumption of p, = 0.3 for M = N,. At lower temperatures, our application of Troe's method suggests a slower increase in ko(T)as the temperature is lowered than is found experi- mentally, at least for M = Ar for this reaction.In considering this comparison, two points should be kept in mind. First, Troe's equations for computing values of ko(T) are based on a number of assumptions and approximations which may not be valid at the very low temperatures achieved in the CRESU apparatus. Secondly, we have assumed that /?, for N, is inde- pendent of temperature. If, on the other hand, the average energy transferred in collisions between energised HONO adducts and N, molecules remains constant, then p, would increase to lower temperatures improving the linearity of the plot of calculated values of In ko(T)us.ln(T/K). Although there is no compelling theoretical reason why it should be so, our experimental results indicate that rate constants for col- lisionally stabilised radical-radical association reactions in the limit of low pressure can be estimated, with a fair degree of confidence, using expressions of the form ko(T)= A"T"' with values of A" and no estimated from experiments at higher temperatures. One of the major stimuli to our work on gas-phase reac- tions between neutral species at ultra-low temperatures is the wish to provide accurate rate constants, either directly or indirectly by testing theoretical predictions, for use in chemi- cal models of interstellar clouds. Of course, even in 'dense' interstellar clouds, the total gas density is far too low for association via collisional stabilisation to occur.However, great interest has focused on the possible importance of radi- ative ion-molecule ass~ciation,~~ in which some fraction of the energised adducts formed by combination of an ion and a neutral molecule would lose energy by emitting infrared photons. It is widely acknowledged that the rate of any such process is closely related to that for the corresponding col- lisionally induced association, whose rate can often be deter- mined in laboratory experiments, since the equilibrium concentration of energised adducts is determined by exactly the same factors in the two cases, and the rates of collisional stabilisation, on the one hand, and radiative stabilisation, on the other, can be estimated with a fair degree of confidence. The radiative association of free radicals, resulting from collisions on the electronic ground-state potential-energy surface, has received less attention, although both Smith28 and Barker,' have estimated rate constants for a number of radical-radical systems.The importance of the present work, in this respect, is not that the association of OH and NO is of any direct importance in interstellar clouds (even if formed, HONO would rapidly photodissociate, just as it does in the earth's atmosphere), but our new results at ultra-low tem- association of two non-rotating radicals. Correction factors are then applied to this result to allow for the neglect of peratures confirm that the methods used to estimate rate con- anharmonicity, the dependence of the internal state density stants for low-pressure association do appear to work well on energy, and for the effects of external rotations and, if down to the extremely low temperatures which characterise interstellar clouds. present, internal rotations.The method has been sho~n~,~~ to predict values of ko(T)within a factor of ca. 2 for a wide range of association reactions involving small free radicals (and the reverse dissociations) at room temperature and above. In applying this theoretical procedure to reaction (l), we used spectroscopic and structural information about Summary Rate constants for the pressure-dependent association reac- tion between OH radicals and NO have been measured over a wide range of low (down to 80 K) and ultra-low (down to 23 K) temperatures in two separate programmes of experi- ments using PLP to generate OH radicals and time-resolved LIF to monitor the kinetics of their removal.The results of the set of experiments performed in a cryogenically cooled cell at temperatures between 301 and 79 K and with N, as the diluent gas have been fitted to find temperature-dependent expressions for the rate constants in the limits of low and high pressure, although the accuracy of the latter is low since the experiments are limited to relatively low pres- sures. The rate constants measured at ultra-low temperatures in the CRESU apparatus, which are the first kinetic experi- ments on any radical-radical association reaction below the temperature of liquid N, , are also obtained fairly close to the low-pressure limit.They are corrected for the effects of fall-off to yield rate constants in the limit of low pressure for the combinations of temperature and third-body gas employed in these experiments. All the limiting low-pressure rate con-stants have been compared with values predicted according to the methodology due to Tr~e.~~ The agreement between experiment and theory is only satisfactory down to 80 K. It is suggested that values of the rate constants for collisional and radiative association at low pressures and very low tem- peratures can be reasonably estimated by extrapolation from higher temperatures assuming a Tn0dependence on tem-perature.We acknowledge funding from the SERC and from the CEC under the Science Plan (Contract No. SC*CT89-0261). The experiments carried out in the CRESU apparatus in Rennes were also supported by the GDR’s ‘Physicochemie des Mole- cules Interstellaires ’ and ‘ Dynamique des Reactions Moleculaires’ programmes. The lasers for these experiments were borrowed from the SERC Laser Loan Pool at the Rutherford-Appleton Laboratory, for which we express thanks. We should also like to thank Samantha Nimmo for carrying out the global fitting computations. References (a) D. L. Baulch, R. A. Cox, R. F. Hampson Jr., J. A. Kerr, J. Troe and R. T. Watson, J. Phys. Chem. Ref: Data, 1980, 9, 295; (b) D.L. Baulch, R. A. Cox, P. J. Crutzen, R. F. Hampson Jr., J. A. Kerr, J. Troe and R. T. Watson, J. Phys. Chem. Ref: Data, 1982, 11, 327; (c) D. L. Baulch, R. A. Cox, P. J. Crutzen, R. F. Hampson Jr., J. A. Kerr, J. Troe and R. T. Watson, J. Phys. Chem. Ref: Data, 1984, 13, 1259; (d) R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson Jr., J. A. Kerr and J. Troe, J. Phys. Chem. Ref: Data, 1989, 18, 881; (e) R. Atkinson, D. L. Balch, R. A. Cox, R. F. Hampson Jr., J. A. Kerr and J. Troe, J. Phys. Chem. Ref: Data, 1992,21, 1125. J. S. Robertshaw and I. W. M. Smith, J. Phys. Chem., 1982, 86, 785. H. Hippler, R. Forster, M. J. Frost, A. Schlepegrell and J. Troe, results cited in ref. l(e) and reported at the XXth Informal Con- ference on Photochemistry, Atlanta, USA, 1992. J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 4 I. W. M. Smith and M. D. Williams, J. Chem. SOC., Faraday Trans.2, 1985,81, 1849. 5 M. J. Howard and I. W. M. Smith, Prog. React. Kinet., 1983, 12, 55. 6 I. W. M. Smith, J. Chem. SOC., Faraday Trans., 1991,87,2271. 7 C. Anastasi and I. W. M. Smith, J. Chem. SOC., Faraday Trans. 2, 1978, 74, 1056. 8 J. G. Anderson, J. J. Margitan and F. Kaufman, J. Chem. Phys., 1974,60,3310. 9 E. R. Lovejoy, T. P. Murells, A. R. Ravishankara and C. J. Howard, J. Phys. Chem., 1990,94, 2386. 10 P. Sharkey and I. W. M. Smith, J. Chem. SOC., Faraday Trans., 1993,89,631. 11 (a)M. J. Frost, P. Sharkey and I. W. M. Smith, Faraday Discuss. Chem. SOC., 1991, 91, 305; (b) M.J. Frost, P. Sharkey and I. W. M. Smith, J. Phys. Chem., 1993,97, 12254. 12 I. R. Sims and I. W. M. Smith, Chem. Phys. Lett., 1988,151,481. 13 I. R. Sims and I. W. M. Smith, J. Chem. SOC., Faraday Trans., 1993,89, 1. 14 I. R. Sims, J-L. Queffelec, A. Defrance, D. Travers, B. R. Rowe, L. Herbert, J. Karthauser and I. W. M. Smith, Chem. Phys. Lett., 1993,211,461. 15 (a)I. R. Sims, J-L. Queffelec, A. Defrance, C. Rebrion-Rowe, D. Travers, B. R. Rowe and I. W. M. Smith, J. Chem. Phys., 1992, 97, 8798; (b) I. R. Sims, J-L. Queffelec, A. Defrance, C. Rebrion- Rowe, D. Travers, P. Bocherel, B. R. Rowe and I. W. M. Smith, J. Chem. Phys., 1994,100,4229. 16 I. R. Sims, P. Bocherel, A. Defrance, D. Travers, B. R. Rowe and I. W. M. Smith, J. Chem. SOC., Faraday Trans., 1994,90,1473. 17 I. R. Sims, I. W. M. Smith, D. C. Clary, P. Bocherel and B. R. Rowe, J. Chem. Phys., 1994,101,1748. 18 D. B. Atkinson and M. A. Smith, J. Phys. Chem., 1994,98,5797. 19 (a) B. R. Rowe, G. Dupeyrat, J. B. Marquette and P. Gaucherel, J. Chem. Phys., 1984, 80, 4915; (b) B. R. Rowe and J. B. Mar- quette, Int. J. Mass Spectrom. Ion Processes, 1987, 80, 239; (c)c. Rebrion, J. B. Marquette and B. R. Rowe, J. Chem. Phys., 1989, 91, 6142. 20 I. W. M. Smith, Kinetics and Dynamics of Elementary Gas Reac- tions, Butterworth, London, 1980. 21 M. Keiffer, M. J. Pilling and M. J. C. Smith, J. Phys. Chem., 1987,91,6028. 22 J. Troe, J. Phys. Chem., 1979,83, 114. 23 G. Z. Whitten and B. S. Rabinovitch, J. Chem. Phys., 1963, 38, 2466. 24 J. Troe, J. Chem. Phys., 1977,66,4758. 25 (a)I. M. Mills, personal communication; (b) G. Herzberg, Elec-tronic Spectra of Polyatomic Molecules, Van Nostrand Reinhold, Toronto, 1966. 26 G. Herzberg and K. P. Huber, Molecular Spectra and Molecular Structue: Constants of Diatomic Molecules, Van Nostrand Rein- hold, Toronto, 1979. 27 See, e.g. (a) D. R. Bates and E. Herbst, in Rate Coeficients in Astrochemistry, ed. T. J. Millar and D. A. Williams, Kluwer, Dordrecht, 1989, p. 17; (b) I. W. M. Smith, Astrophys. J., 1989, 347, 282; (c) D. Smith, Int. J. Mass. Spectrom. Ion Processes, 1993, 129, 1. 28 I. W. M. Smith, Chem. Phys., 1989,131,391. 29 J. R. Barker, J. Phys. Chem., 1992, %, 7361. Paper 41049511; Received 12th August, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003609

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3617-3624 Reaction of Atomic Oxygen with Some Simple Alkenes Part 1.-Low-pressure Studies on Reactions involving Ethene, Propene and (0-But-2-ene Christopher Anastasi" and Michael G. Sanderson Department of Chemistry, University of York, Heslington, York, UK YOI 5DD The reaction of atomic oxygen O(3P) with ethene, propene and (E)-but-2-ene has been studied between 10 and 50 mbar total pressure using the alkene as both reactant and bath-gas. NO, was used as the source of atomic oxygen, and the decay of NO,, which is sensitive to the radical products from the reaction of 0 atoms with alkenes, was followed by laser-induced fluorescene. Branching ratios for total radical production were obtained for the propene and (€)-but-2-ene systems.The radical branching ratio for the propene system was observed to decrease with pressure by a factor of about two over the range measured, failing from 0.64 at 10 mbar to 0.36 at 40 mbar total pressure. The radical branching ratio for the (€)-but-2-ene system also decreased by a factor of two, falling from a value of 0.22 at 10 mbar to 0.11 at 30 mbar. The results for the ethene system were inconclu- sive, in that only a range for the branching ratios for production of CH, + HCO and H + CH,CHO could be obtained. These findings were confirmed by extensive modelling of the results. The reaction of atomic oxygen with alkenes is an important reaction in smog-chamber experiments and combustion studies. There are many published studies on this subject at high and low pressures which have been reviewed recently.'--4 These studies include the rate constant for the overall reac- tion and the branching ratios for the various product chan- nels although there is still considerable uncertainty in the latter.There is a large volume of work on the ethene system at low and high pressures, but much less for propene and (E)- but-2-ene. No data exist for any of the reactions carried out at between 20 and 50 mbar, which is the pressure regime investigated in the present work. Although the rate constants for the overall reactions of atomic oxygen with ethene, propene and (E)-but-2-ene are known quite well, as indicated ab~ve,~,~ there is still a large amount of uncertainty in the branching ratios for the product channels (see ref.1 and 2, and references therein). The identity of the major products from the reaction of 0 atoms with ethene and propene has been established,, but not, so far, for (E)-bu t-2-ene. 10 mW He-Cd laser quartz reaction cell It was the object of the present work to measure the branching ratio for total radical production from the reaction of atomic oxygen with ethene, propene and (E)-but-2-ene. Experimental Apparatus A schematic of the experimental apparatus is given in Fig. 1. A conventional mercury-free vacuum line with greaseless Teflon stopcocks was used for gas handling and mixing. Gases were transferred from the cylinders and stored in blackened bulbs prior to use. Mixtures were made in a black- ened vessel containing a magnetic stirrer.The gases were agi- tated for 30 min, and left for a further 30 min to ensure complete mixing had occurred. Samples of a mixture were then transferred to a quartz reaction cell. The cell possesses mutually perpendicular quartz windows at its centre, which allow the entry and exit of the laser beam, and the escape of the fluorescence light from the excited NO,. light' photolysis lamp todiode pre-ioniser Fig. 1 Diagram of the experimental apparatus NO, was photolysed by a Phillips Blacklight Lamp which emits radiation between 300 and 430 nm, with a broad maximum at ca. 350 nm. Both the cell and lamp were enclosed inside a blackened light-proof box. A cooling fan was also present to maintain the temperature at a reasonably constant level.A rise of 2 K could not be prevented, but the average temperature in all runs was 298 & 1 K. Apparatus for Laser-induced Fluorescence The decay of NO, was monitored by laser-induced fluores- cence (LIF) at 441.6 nm using an He-Cd Laser.' The laser beam is chopped by a rotating wheel (Scitec Instruments) at 500 Hz and passed through a series of baffles before reaching the cell to reduce the amount of scattered laser light. The fluorescence signal is passed through a 590 nm cut-on filter (Schott OG-590) and collected by a photomultiplier (EM1 9815QB) at right angles to the laser beam. The output of the photomultiplier is passed through a resistor to a lock-in amplifier (EG & G Princeton Applied Research Model M5207).The chopper also provides a reference signal to the lock-in amplifier, and thus allows the signals pulsed at 500 Hz to be selected from the background noise, improving the signal-to-noise ratio. The output of the lock-in amplifier was recorded by a Ventura Servoscribe chart recorder. A curve was fitted to the decay trace by eye, and readings taken at regular intervals. The subsequent analysis is described later. At least two experiments were performed at each pressure and sometimes as many as seven times for each alkene. Materials NO, (99.5% purity) was supplied by Fisons, and was purified by freeze-pumpthaw cycles, twice at 77 K using liquid nitro- gen, and twice at 195 K using an C0,-acetone slush bath. The purification was performed in a blackened storage bulb equipped with a finger.All other gases were used as supplied with no further purification. Ethene (CP grade, 99.9% purity) and (E)-but-2-ene (>95% purity; the other 5% is the cis isomer) were supplied by Matheson, propene (99.5% purity) by Aldrich and helium (99.995% purity) by BOC. Results 0.5 mbar of NO, was photolysed in 10-50 mbar of ethene, 10-40 mbar of propene and 10-30 mbar of (E)-but-2-ene. Experiments at higher pressures could not be performed owing to severe quenching of the fluorescence signal. The decay of NO, in helium, ethene, propene and (E)-but-Zene is compared in Fig. 2; the decay rates of NO, in helium and (E)-but-2-ene, and in ethene and propene, are similar at 10 mbar.It is also obvious from these results that the decay rates are faster in the ethene and propene systems. This sug- gests that radicals are a larger percentage of the products from the reaction of 0 atoms with ethene and propene than for (E)-but-2-ene. If the alkene is in sufficient excess, the reaction between atomic oxygen and NO, may be ignored, and the decay of NO, will be first order during the first part of the reaction time. Therefore, plotting ln(Zf)o -ln(If)t against time should yield a straight line, where and are the fluorescence signals at the start of the reaction and after t seconds, and are proportional to the concentration of NO,. This relationship is shown in Fig.3 and 4. The decay of NO, is indeed first order over the initial part of the reaction time. Simple kinetic schemes describing the chemistry may be constructed and solved to obtain the branching ratio for radical production from the initial part of the curve where the decay of NOz is first order. Detailed chemical models were also used to fit the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0.5 0.4 = 0.2 150 300 450 600 '750 900 1050 1200 1350 ti me/s Fig. 2 Comparison of the decay rates of NO, in 50 mbar helium (O), and 10 mbar of ethene (O),propene (A) and (E)-but-Zene (+). The initial partial pressure of NO, was 0.5 mbar in all cases. whole curve in attempts to confirm the results from the simple analyses and these are discussed in more detail later.Photolysis Rate of NO, :Experimental Determination The photolysis rate of NO, was found by photolysing NO, in helium bath-gas at 50 mbar. At this low pressure, only two A'5.014.5 0 150 300 450 600 750 900 1050 1200 1350 time/s Fig. 3 First order decay plot for NO, in 50 mbar helium (0)and 10 mbar propene (A). The initial partial pressure of NO, was 0.5 mbar in each case. ::q 0 0 0z 5.04.5i4.0 Y + 5 3.54 -2.5 fi 2.0 = 1.5 1.o 0.5 0.0 time/s Fig. 4 First order decay plot for NO, in 10 mbar ethene (0)and 10 mbar (E)-but-Zene (+). The initial partial pressure of NO, was 0.5 mbar in each case. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 reactions are important: khv NO,+hv -NO+O kzO+N02 -NO+O, (1) (2) The decay of NO, is given by: -d"ozl = kh,[NO,] + k'[O][NO,]dt The concentration of 0 atoms is maintained at a low level due to their rapid removal uia reaction with NO,. The rate of change of the 0 atom concentration is low, and if a steady- state analysis is performed on the concentration of 0 atoms, an expression for [O] may be found.Insertion of this expres- sion into eqn. (I) and integrating gives: ln[NOJ0 -ln[NO,], = 2k,,t (11) The decay of NO, will therefore be first order until the con- centrations of NO and 0, are sufficiently high that the reac- tion between them and the 0 atoms becomes significant. Hence, plotting the left-hand side of eqn. (11) against time should yield a straight line whose gradient is 2k,,.This linearity is shown in Fig. 3. Six such experiments were per- formed, giving k,, = (1.70 k0.04) x s-', the error being la. Photolysis Rate of NO, :Theoretical Determination The photolysis rate constant for NO, may also be estimated theoretically, which provides a useful check on the experi- mental figure and also allows photolysis rates for other species, such as nitrites, which are difficult to measure experi- mentally to be calculated. The strength of the 0-NO bond6 is 305 kJ mol- ',which is the energy possessed by a photon of wavelength 391 nm. Photolysis of NO, will occur up to ca. 422 nm, as the energy deficiency is made up by the vibrational and rotational energy of the NO, molecules.' The Beer-Lambert law, given in eqn.(111) is assumed to be valid for light absorption by (and hence photolysis of) NO, in this instance: A = acd (111) where a is the absorption cross-section (in cm2 molecule- '), c is the concentration of NO, and d is the pathlength. The pathlength, d, was estimated as 2.04 cmt for the cell used in these experiments. Not all the photons emitted will reach the cell. Assuming a uniform radial output for the lamp, the pro- portion of photons which reach the cell is given by 8/27r, where 8 = 2 arcsin(r/C),r is the radius of the cell (1.3 cm) and I is the distance between the centres of the lamp and the cell (4.8 cm). Hence 8/2n = 0.087. The number, n,, of photons available for photolysis at a wavelength, A,is given by n, = 2.04 x 0.0870, 4, yi (IV) where 4, is the quantum yield for photolysis and yn is the number of photons emitted per second.The above expression gives the photolysis rate over the entire cell. To convert this figure to a photolysis rate constant, it is necessary to divide by the volume of the cell, 242 cm3. The rate of loss of NO, is The area of a circle = nr2 = 2rp, where r is the radius and p is the average length of a chord in the circle. The quantity p is therefore the average pathlength for photons passing through the cell. Hence p = +nr. the photolysis rate, -d"O,I = k,,[NO,] = $ [NO,]dt where k,, is the photolysis rate constant and V the volume of the cell. Thus, for any wavelength A, k,, = nJV. The actual photolysis rate constant is found by summing n, over all wavelengths emitted by the dark lamp, 4 4301-= 7khv nA 300 k,, was evaluated by summing values of n, over 5 nm inter- vals.Averaged values of CT and 4 between 300 and 410 nm were calculated from the review by Atkinson et UL,~and between 410 and 430 nm from Schneider et aL9 Values of yn for the lamp were calculated from the manufacturer's spe-cifications.$ The photolysis rate constant for NO, is esti- mated as 1.55 x s-'. This is in good agreement with the experimental value of 1.70 x s-'. Photolysis rate con- stants for other species for use in the modelling studies may be estimated by calculating the photolysis rate constant as shown above, and multiplying by the ratio of the experimen- tal and theoretical photolysis rate constants for NO,, which is 1.097.Reaction of 0 Atoms with Ethene The products of the reaction of atomic oxygen with ethene will be CH, + HCO, H + CH,CHO, and possibly H, + CH,C0.4 The scheme below describes the decay of NO, over the first-order region. khvNO, + hv -NO + 0 k30 + C2H4 ____* CH3 + HCO k40 + C2H4 -H + CHZCHO 0 + C,H, ksH, + CH,CO k6CH, + NO, -CH30 + NO kiCH30 + NO, -CH30N0, keHCO + NO, -H + CO, + NO k9H+NO,-OH+NO OH + C,H, HOC,H4 HOC,H4 + NO, AHOC2H40+ NO HOC2H4 + NO, --%HOC,H,NO, HOC,H40 + NO, -2HOC2H40N0, CH,CHO + NO, aHNO, + CH,CO Note that other work described by uslo suggests that the pro- ducts of reaction (3) may be CH, + H + CO and not CH, + HCO.The effect of this will be discussed below. Other pos- sible products from the reaction between methyl radicals and NO, [reaction (6)] are CH3N0, and CH,ONO. The ther- mochemical arguments of Gray' ' indicate that the formation of the nitrite, CH30N0, almost certainly does not occur. Yamada et a/." have studied this reaction at 295 K and 1.05 1The Philips Lighting Catalogue (1988) gives the number of W of energy emitted per 10 nm for the 40 W Blacklight lamp, and states that the wattage per 10 nm for the 20 W lamp may be obtained by multiplying the 40 W lamp wattage by 0.4. Using this proportion, a conversion factor of 0.3 was calculated for the 15 W lamp used here. The output per 5 nm was assumed to be half the output per 10 nm.Torr, and the only product detected was NO. CH,NO, was not seen. The co-product of NO, CH,O, was also not detected although these authors expected this.', Biggs et ul.' also studied this reaction between 1 and 10 Torr at 298 K. The formation of CH,NO, and/or CH,ONO will require third-body stabilisation, and so the rate constant should exhibit a pressure dependence if one or both of these two products are formed. However, no such pressure dependence was observed, and Biggs et al.' concluded that the only pro- ducts of reaction (6) were CH,O + NO. These authors also performed a quantum RRK calculation on this reaction. Their results indicate that at zero pressure, 97.5% of the pro- ducts are CH,O + NO. This yield only drops to 72.3% at 10 atm.From the above evidence, the only products expected from reaction (6) are CH,O + NO over the pressure range used in the present work. The products of reaction (8), HCO + NO,, may be HONO + CO as well as H + CO, + NO. The mechanism for the similar reaction CH,CO + NO, has been shown to produce only CH,CO, and NO.', It may be inferred that reaction (8) above may also proceed in this manner,', in which case the products are HCO, (which rapidly decom- poses to H + CO,) and NO. The decay of NO, is given by the expression: Steady-state analyses may be applied to the concentrations of the radicals, and the resulting expressions for the concentra- tions inserted into eqn. (VII). Integration then yields : (VIII) If the left-hand side of eqn.(VIII) is plotted against time, a straight line is obtained for experimental points in the first half of the reaction time (see Fig. 4). The gradient m of the straight line is given by: Rearranging this expression gives 4k, + (3 + kll )k4 m--1= kll + k12 kh, k, + k, + k, (X) The ratio kll/(kll + k,,) = 0.75 for the ethyl radical,16 and it is likely to be very similar for the 2-hydroxyethyl radical. The individual values of k, and k,, and hence the branching ratios for the production of CH, + HCO and H + CH,CHO, cannot be found. However, upper limits for the branching ratios for radical production may be calculated by assuming either k, or k, are equal to zero. The average upper limits to the branching ratios obtained between 10 and 50 mbar are given in Table I, and suggest there is no correlation J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 Average branching ratios for radical production for the reaction 0 + C,H, upper limit for branching ratio pressurejmbar CH, + HCO H + CH,CHO 10.0 0.68 20.0 0.56 0.61 30.0 0.76 0.8 1 40.0 0.76 0.81 50.0 0.61 0.72 0.64 between the upper limits and pressure; it is possible to con- clude that the branching ratio for CH, + HCO is d0.76 and that for H + CH,CHO is d0.81 between 10 and 50 mbar. The upper limit for the production of CH, + HCO, calcu-lated assuming the products of HCO + NO, [reaction (S)] are HONO + CO instead of HCQ, + NO, is larger than 1.0 at 30 and 40 mbar total pressure.Clearly, it is not possible to have a branching ratio larger than 1.0 and this suggests that, if HONO and CO are formed, the assumption that k, is equal to zero is erroneous, at least at these pressures, but also indicates that production of CH, + HCO and H + CH,CHO from the reaction of 0 atoms with ethene must be occurring to some extent. If the products of reaction (3) are CH, + H + CO (see above), the upper limit for production of H + CH,CHO is unchanged, whereas the upper limit for CH, + H + CO is lower than that derived for CH, + HCO, being 0.64 as opposed to 0.76. The upper limits in Table 1 are less than 1.0 for both radical channels, and so both these channels must be occurring. Hence, the branching ratios for CH, + HCO and H + CH,CHO must also be greater than 0.19 and 0.24, respectively (or 0.19 and 0.36 if the co-products of the methyl radical are H + CO instead of HCO).Reaction of 0 Atoms with Propene The products of the reaction of 0 atoms with propene will be CH, + CH,CHO and H + CH,CHCH0.2 Arutyunov et al." have shown that the alternative H atom producing channel, H + CH,COCH,, does not occur. A scheme to describe the initial decay of NO, is given below: kh\NO, + hv _____+NO+O (1) kisO+C3H, ___ CH, + CH,CHO (15) kl6O+C,H6 ___-+ H + CH,CHCHO (16) ki?0 + C3H, ~ C,H,O (17) CH, + NO, ?--+ CH,O + NO (6) ~CH,O+NO, ki + CH,UNO, (7) CH,CHO + NO, A+HNO, + CH,CO (18) H+NO, ks OH + NO (9)~ + k19OH 4-C3H6 ___+ HOC3H6 (19) kzo ~HOC,H6 + NO, + HOC,H60 + NO, (20) k2 I HOC,H,+NO, + HOC,H6N0, (21)~ HOC,H,O fast RCHOH + R'CHO (22)~ + k23 ~RCHOH + NO, + RCHO + HONO (23) CH,CHCHO + NO, 2-+ HNO, +CH,CHCO (24) J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 The species RCHOH and R'CHO may be CH,CHOH + CH,O or CH,OH + CH,CHO. Their identity is not impor- tant as both CH,OH and CH,CHOH react in the same manner with NO,.' The decay of NO, is given by eqn. (XI): + k7[CH,OI"O,I + k,8CCH,CHOI"O,I + k9CHI"OZI + (k20 + k21)[HoC3H61[No21 + k,,[RC H OH] [NO,] + k,,[CH,CHCHO][NO,] (XI) Once again, a steady-state analysis of the radical concentra- tions will give expressions for those concentrations. If these expressions are inserted into eqn (XI), integration will give eqn.(XII). ln[NO,], -ln[NO,], = khv (XII) Plotting the left-hand side of eqn. (XII) against time yields a straight line over the first half of the reaction time, as shown in Fig. 3. As before, the gradient m of the line is equal to the right-hand side of eqn. (XII). However, the individual values of k15 and k16 cannot be found from this equation. If it is assumed that k,, = 0 then rearrangement of eqn. (XII) gives an expression for the branching ratio for radical production, (k15 + k16)/(k15 + k16 + k17)* This will affect the results, as the ratio k20/(k20+ k21) may be the same for C,H7 , which is 0.69.19 The ratio k2"/(kzO+ k21) may be lower than 0.69 for the species HOC3H6. This species represents CH,CHCH,OH and CH,CH(OH)CH, .The former is made 65% of the time, and the latter, 35%,20 and is likely to form more stable products than a primary radical. The radical branching ratios obtained for the propene system are summarised in Table 2. There is a clear pressure dependence for the radical branching ratio for 0 + C3H,, which falls from 0.64 at 10 mbar to 0.36 at 40 mbar. Reaction of 0 Atoms with (E)-But-Zene The products of the reaction of 0 atoms with (E)-but-2-ene will consist of a combination of CH, + C,H5C0, C,H, + CH3C0, and C3H7 + HCO.,l HCO radicals are apparently not produced from the reaction of 0 atoms with propene,' and so are unlikely to be made in this system. To produce C,H, + HCO considerable rearrangement of the initial biradical is also required, and the production of these two species has therefore been assumed not to occur.H atoms are not produced because the loss of methyl radicals is energeti- cally more favourable.' Table 2 Average branching ratios for radical production for the reaction 0 + C,H, pressure/mbar branching ratio 10.0 0.64 f0.04 20.0 0.60 f0.08 23.4 0.54 0.01 30.0 0.41 f0.02 40.0 0.36 f0.07 3.621 In the scheme below, R, = C2H5, R, = C2H5C0 and R, = CH, if the initial products of the 0 + C4H8 reaction are CH, + C2H5C0, but R, = C,H,, R, = CH,CO and R, = CH, if the initial products are C2H5 + CH,CO. For the product channel CH, + C,H5C0, the radicals produced from 0 + C4H8 are R, + R,, and the products from the reaction of R, and NO, are R, and NO.However, the steady-state equations for the radical concentrations are not changed. A simple, chemical scheme for the decay of NO, in this case is: NO,+hvLNO+O k25 0 + C4H8 -Rl + R2 k260 + C4H8 -C4H80 k21R, + NO, R,O + NO R1 + NO, k2t( RlNO, k29R,O + NO, -R'ONO, k30R, + NO, --+R3 + NO kbCH, + NO, --CH30 + NO kJCH,O + NO, -CH,0N02 The decay of NO, is given by eqn. (XIV). As before, a steady-state analysis of the radical concentra- tions may be performed. Inserting the resulting expressions into eqn. (XIV) and integrating gives eqn. (XV) Once again, plotting the left-hand side of eqn. (XV) against time yields a straight line over the initial part of the reaction time, as shown in Fig.4.The gradient rn is equal to the right- hand side of eqn. (XV), from which the branching ratio for radical production, k25/(k25 + k,,), may be calculated. The ratio k27/(kz7 + k28) = 0.75 if R, = C2H516 or 0.69 if R, = C,H7 .19 The branching ratios for radical production are summarised in Table 3, and are independent of the value of k27/(k27 + k28). When one compares the results from the 10 and 20 mbar data (0.22 and 0.12, respectively), there appears to be a pressure dependence for the radical branching ratio. The error in the point at 30 mbar is large, but confirms this finding. The radical products of the reaction between 0 atoms and (E)-but-2-ene reported by Ferrieri and Wolf" (see above) are Table 3 Average branching ratios for radical production for the reaction 0 + C,H, press ure/mba r branching ratio 10.0 0.22 k0.03 20.0 0.12 f0.04 30.0 0.11 f0.08 0 100 200 300 400 500 600 700 800 900 time/s Fig.5 Model fit to experimental data (0)for the ethene system. Initial partial pressures were: 0.5 mbar NO,, 9.5 mbar C,H4. (-*--) 70% H+CH2CHO; (-) 100% H+CH2CHO; (---) 10%stable products, 90%radicals. generally an alkyl radical and an acyl radical. Acyl radicals, however, are not produced from the reaction of 0 atoms with ethene and propene.2 Ferrieri and Wolf2' used 'hot' 0 atoms, and this may be the reason for acyl radical formation. t m5 z2 1 0 0.1 time/s Fig. 6 Model fit to experimental data (0)for the propene system.Initial partial pressures were: 0.5 mbar NO,, 9.5 mbar C,H,. (---) 60% radicals; (-----) 70% radicals; (---) 80% radicals; (-) 90% radicals. 1 10-5 9--$ 8-2 7-5 6-;5-z 4-173-O 2-t 1-0 120 240 360 480 600 720 840 960 1080 1200 time/s Fig. 7 Model fit to experimental data (0)for the (E)-but-Zene system. Initial partial pressures were: 0.5 mbar NO,, 9.5 mbar C4H, . (-. . . -) 10% radicals; (----) 20% radicals; (. .-..) 25% 30% radicals; (-.-.-) 35% radicals; (---) 40% radicals; (. . -) 50% radicals. (-) radicals; J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Hunziker,, has reported the detection of the methylvinyloxyl radical, CH,cHCHO, as a product from the reaction of 0 atoms with propene and (E)-but-Zene.This species is formed by displacement of a methyl group from the biradical initially formed. By examination of the structure of the intermediate biradical, no other vinyloxyl-like radical products are pos- sible, and H atom elimination does not occur.17 If the radical products from the reaction of 0 atoms with (E)-but-2-ene are assumed to be CH, + CH,CHCHO only, and that the methylvinyloxyl radical reacts in the same way as the vinyl- oxyl radical with NO, ,23 the total radical branching ratios at 10, 20 and 30 mbar are 0.35 f0.04, 0.20 0.08 and 0.18 0.12, respectively. These branching ratios are ca. 50% larger than those in Table 3, but the values at 20 and 30 mbar are essentially identical to the branching ratio recom- mended by Atkinson and Lloyd24 of 0.20 atm.Modelling Studies Detailed chemical models were developed to predict the whole of the NO, decay curve, and not just the first order region. Detailed descriptions of the chemical mechanisms are given by Sanderson,' and need not be repeated here. Briefly, reactions of the initially formed radicals with NO and 0, ,as well as NO,, were included in the analysis, as were sub- sequent reactions of the peroxyl and alkoxyl radicals formed. Species such as alkyl nitrates were assumed to be inert. Rate constants for the majority of the radical reactions formed in the ethene system have been reported in the literature; the remainder were estimated by comparison with similar reac- tions with known rate constants.More estimations of rate constants had to be made for the radicals produced in the propene and (E)-but-2-ene systems. The fate of certain species is not clear. For example, nitroso compounds, such as C,H,NO, are formed in large yields during the second part of the NO, decay curve, where the concentration of NO, is low, and the decay is no longer first order. These species may isomerise to the corresponding oximes, which are potentially reactive species, due to the presence of a C=N double bond, and an OH group. Clearly the use of the entire curve is beneficial, but this is balanced by an increased complexity of the models used over the simple method of analysis presented earlier; there is then a larger uncertainty in the uniqueness of the detailed models.Typical sizes of the models were between 40 and 70 species, and 60 and 100 reactions. It was possible to reduce the size of the models, by up to 40%, by excluding certain reactions based on a systematic sensitivity analysis. For example, the reaction of alkyl radicals with 0, could be ignored because reaction with NO and NOz dominates. Despite the uncer- tainties mentioned above, the models were found to fit the data well at all pressures. Sample fits at 10 mbar are shown in Fig. 5-7. Fig. 5 shows that the ethene model was not very sensitive to changes in the radical branching ratios for the initial reac- tion, 0 + C2H4, of less than 10%.No correlation with pres- sure was observed for the radical branching ratios for the two product channels CH, + HCO and H + CH,CHO, in agree- ment with the experimental results.The propene and (E)-but-2-ene models were more sensitive to changes in the radical branching ratio, as shown in Fig. 6 and 7. The propene model confirmed the results of the simple analysis of a pressure dependent radical branching ratio, which fell from 0.8 at 10 mbar to ca. 0.3 at 40 mbar. Simi- larly, the radical branching ratio for the (E)-but-2-ene model fell from 0.25-0.30 at 10 mbar to 0.10-0.15 at 30 mbar. The modelling studies confirm the results of the simple analysis, J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 of a pressure-independent radical branching ratio for 0 + C2H4, and a decrease by a factor of roughly two for the radical branching ratios for 0 + C,H, and 0 + C,H,, which is in good agreement with the decrease by a factor of about two observed experimentally. Discussion 0 + Ethene It has not been possible to establish whether there is a corre- lation with pressure for the upper limits for the two channels CH, + HCO and H + CH,CHO.The only conclusion is that the branching ratios for CH,HCO and H + CH,CHO lie within the ranges 0.19-0.76 and 0.24-0.81, respectively, between 10 and 50 mbar (or 0.19-0.64 and 0.36-0.81 for CH, + H + CO and H + CH,CHO). These ranges are consistent with the recent results of Knyazev et aL2 who measured the branching ratio for the channel H + CH,CHO at low pres- sure. The branching ratio was observed to fall from 0.60 at 8.7 mbar to 0.49 at 16.1 mbar.The lack of correlation with pressure for the branching ratios of CH, + HCO and H + CH,CHO implies that the total radical branching ratio is also independent of pressure. This result is in agreement with the recommendation of Baulch et aL4 0 + Propene Table 2 shows that the branching ratio for radical production is dependent on pressure, falling from 0.64 at 10 mbar to 0.36 at 40 mbar. This result is in qualitative agreement with the results of Hunziker et a/.,,, who observed a decrease in the yield of the vinyloxyl radical from 0.29 at 53 mbar to 0.19 at ca. 1 bar. The mechanism for the reaction of 0 atoms with ethene proposed by Knyazev et al., predicts a decrease in the yield of radicals and an increase in stable products as the pressure is increased.The products formed by rearrangement of the biradical have excess energy, and as the pressure is increased, stabilisation rather than decomposition of these products occurs. This effect will be more pronounced for propene, because the stable products, which are likely to be propanal and 1,2-epo~ypropane,~ possess more bonds to distribute excess energy, and are thus less likely to fragment than the potential stable products of the ethene system, ethanal and epoxyethane. The increase in the proportion of stable products with increasing pressure observed here is con- sistent with the mechanism of Knyazev et al., The radical yields obtained here are likely to be a slight overestimate, however, as the reaction of the hydroxypropyl radical with NOz was assumed to produce only the nitro compound [see reactions (20) and (21) of the propene scheme], whereas some formation of HOC3H60 + NO will almost certainly occur.Koda et measured the yield of the vinyloxyl radical as 0.29 & 0.15 at 0.04 mbar, the same value obtained by Hunzi- ker et aLZ6at 53 mbar. Knyazev et al., measured the yield of H atoms from this reaction as 0.46 f0.11 at both 1.2 and 14.9 mbar. Combining the results of Hunziker et Koda et ~1.~’and Knyazev et al.’ gives a total radical yield of 0.75 rt 0.19 at 10 mbar; the result obtained here, 0.64 0.07, is in reasonable agreement with this value. The value obtained here may also be a slight underestimate, because at 10 mbar, ca.10% of the 0 atoms will react with the NO, during the first minute of the reaction. 0 + (Q-But-2-ene The branching ratio for radical production appears to decrease by a factor of two between 10 and 20 mbar, and is independent of pressure above 20 mbar. Using the products of Ferrieri and Wolf,21 the branching ratio at 30 mbar, 0.11, is about half that recommended by Atkinson and Lloyd,24 0.20, at atmospheric pressure. Given that the latter figure is only approximate, and the errors in the results obtained here, they are actually in fair agreement. However, if the products are CH, and CH,CHCH0,22 the branching ratio falls from 0.35 at 10 mbar to 0.20 at 20 mbar, and is essentially the same at 30 mbar, 0.18.This is in excellent agreement with the recommendation of Atkinson and Lloyd24 at atmospheric pressure. Further work is required to establish exactly the products of the reaction of 0 atoms with (E)-but-2-ene. However, a preliminary spectrum of the products from 0 atoms with (E)-but-2-ene at atmospheric pressure” shows a clear sharp peak at 216.4 nm, which may be attributed to the methyl radical. Mechanism for O(3P)+ Alkenes The mechanism of Knyazev et a1.’ predicts that as the pres- sure increases, so should the degree of stable product forma- tion. For ethene, the reaction with 0 atoms produces an excited biradical in the triplet state. At very low pressures, where virtually no collisions occur, only decomposition back to the initial reactants or H + CH,CHO may occur.’ As the pressure increases, conversion of the triplet biradical to the singlet state occurs via intermolecular collisions.This singlet biradical has more decomposition routes available than the triplet state, and may produce other products, for example CH, + HCO and (CH,),O as well as H + CH,CHO. Hence, as the pressure is increased, the yield of H + CH’CHO should fall. This is observed experimentally.2.26 Stable product formation is very minor below atmospheric pressure. The reaction of 0 atoms with propene and (E)-but-2-ene will presumably operate via an analagous mechanism, except that stable product formation will occur at lower pressures. The intermediate biradicals produced from these two species will be larger and may be less likely to dissociate.The frac- tion of radical products in the propene system is observed to fall in the present work with increasing pressure (see Table 2); a fall in the yield of CH,CHO was observed by Hunziker et This effect should be more pronounced for (E)-but-2-ene, and is indeed observed in the present work. Hunziker et found the branching ratio for production of the vinyloxyl radical from the reaction of 0 atoms with but-1-ene to be almost independent of pressure between 53 and loo0 mbar. To summarise, the pressure dependence of the radical branching ratio is explained by competition between the decomposition of the triplet biradical, and conversion of this species to a singlet state which may rearrange to a stable species instead of decomposing to radical products.The authors would like to thank John Busby for his help in setting up the equipment, the Rutherford Appleton Labor- atory for the loan of the laser, and Agnes Vehovsky and Alfred Anston for translating ref. 17. M.G.S. would like to thank the NERC for a research grant. References 1 R. J. Cventanovic and D. L. Singleton, Rev. Chem. Intermed., 1984,5, 183. 2 V. D. Knyazev, V. S. Arutyunov and V. I. Vedeneev, Int. J. Chem. Kinet., 1992,24, 545. 3 R. J. Cventanovic, J. Phys. Chem. ReJ Data, 16,251. 4 D. L. Baulch, C. J. Cobos, R. A. Cox, C. Esser, P. Frank, T. Just, J. A. Kerr. M. J. Pilling, J. Troe, R. W. Walker and J. Warnatz, J. Phys. Chem. Re& Data, 1992,21,411.3624 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 5 6 7 A. W. Tucker, M. Birnbaum and C. L. Fincher, Appl. Opt., 1975, 14, 1418. CRC Handbook of Chemistry and Physics, ed. D. R. Lide, Chemi- cal Rubber Company Press Inc., Boca Raton, 1991-1992. H. Okabe, Photochemistry of Small Molecules, Wiley, New York, 1978. 17 18 19 20 V. S. Arutyunov, V. I. Vedeneev and V. D. Knyazev, Khim. Fiz., 1990,9, 1383 (in Russian). F. L. Nesbitt, W. A. Payne and L. J. Steif, J. Phys. Chem., 1989, 93, 5158. S. Jaffe and E. Wan, Environ. Sci. Technol., 1974,8, 1024. R. Atkinson, Chem. Rev., 1986,86,69. 8 9 10 11 12 13 14 15 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson Jr., J. A. Kerr and J. Troe, J. Phys. Chem. Ref: Data, 1989, 18, 881. W. Schneider, G. K. Moortgat, G. S. Tyndall and J. P. Burrows, J. Photochem. Photobiol. A, 1987,40, 195. C. Anastasi, M. G. Sanderson, P. Pagsberg and A. Sillesen, J. Chem. SOC.,Faraday Trans., 1994,90,3625. P. Gray, Trans. Faraday SOC., 1955,51,1367. F. Yamada, I. R. Slagle and D. Gutman, Chem. Phys. Lett., 1981, 83,409. P. Biggs, C. E. Canosa-Mas, J-M. Fracheboud, A. D. Parr, D. E. Shallcross, R. P. Wayne and F. Caralp, J. Chem. SOC.,Faraday Trans., 1993, 89, 4163. I. R. Slagle and D. Gutman, J. Am. Chem. SOC., 1982,104,4741. R. S. Timonen, E. Ratajczak and D. Gutman, J. Phys. Chem., 21 22 23 24 25 26 27 28 R. A. Ferrieri and A. P. Wolf, J. Phys. Chem., 1992,96,4747, H. E. Hunziker, personal communication to Cventanovic and Singleton (ref. 1). K. I. Barnhard, A. Santiago, M. He, F. Asmar and B. R. Weiner, Chem. Phys. Lett., 1991, 178, 150. R. Atkinson and A. C. Lloyd, J. Phys. Chem. Ref: Data, 1984,13, 315. M. G. Sanderson, D. Phil. Thesis, University of York, UK. H. E. Hunziker, H. Kneppe and H. R. Wendt, J. Photochem., 1981,17, 377. R. J. Cventanovic, Can. J. Chem., 1958,36,623. S. Koda, Y. Endo, S. Tsuchiya and E. Hirota, J. Phys. Chem., 1991,95,1241. 1988,92,651. 16 J-Y. Park and D. Gutman, J. Phys. Chem., 1983,87, 1844. Paper 4/042638; Received 12th July, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003617

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3625-3631 Reaction of Atomic Oxygen with some Simple Alkenes Part 2.-Reaction Pathways involving Ethene, Propene and (€)-But-Zene at Atmospheric Pressure Christopher Anastasi* and Michael G. Sanderson Department of Chemistry, University of York, Heslington, York, UK, YO1 5DD Palle Pagsberg and Alfred Sillesen Chemical Reactivity Section, Environmental Science and Technology Department, Riso National Laboratory, DK-4000 Roskilde, Denmark Results from a study of the reaction of atomic oxygen with ethene and propene at atmospheric pressure are reported. Pulse radiolysis of CO, has been used to generate ground-state atomic oxygen in the presence of the alkenes. The yield of methyl radicals from the reaction of 0 with ethene, propene and (€)-but-2-ene was mea- sured as 0.70 & 0.05, 0.15 f0.01 and 0.25 f0.08, respectively.The yield of the vinyloxyl radical, CH,CHO, from the reaction of 0 atoms with propene was found to be 0.12 f0.03, in good agreement with the methyl radical yield. An upper limit for other radical-producing channels for this reaction of 0.21 f0.07 was derived. The formyl radical (HCO) was also searched for in the ethene and propene systems but was not detected in either, which placed an upper limit on its yield of 0.03 in both systems. The work presented here complements the studies on the reaction of atomic oxygen with alkenes at low pressure.' The reaction of atomic oxygen with alkenes at atmospheric pres- sure is important for smog-chamber m~delling.~.~ Under almost all ambient conditions, the reaction of atomic oxygen with molecular oxygen to form ozone will be the only signifi- cant loss route, except in exhaust plumes where the NO, levels are high.4 High NO, levels are also used in smog chamber studies.Here, to ensure that sufficient products are made to allow quantitative measurements of their yields, levels of NO, NO, and the hydrocarbon in question are sometimes higher than ambient levels.' The concentration of the alkene and NO, may be sufficiently high that reaction between the alkene and atomic oxygen becomes important. The product from such reactions is a biradical, which rapidly decomposes to either radicals or stable products; radical pro- ducts will increase the reactivity of the system significantly.2 Clearly, the proportion and nature of radicals produced is important in, for example, the validation of chemical schemes describing the degradation of alkenes in smog-chambers.This knowledge is then used in atmospheric models. There is little data on the yields of the radical products at atmospheric pressure, especially for the reaction of 0 atoms with propene and (E)-but-2-ene.' The possible products from the reaction of 0 atoms with ethene are given below: 0 + C2H4 + CH3 + HCO (14 --+ H + CH2CH0 (1b) --+ H, + CH2C0 (14 .+CH, + CH,O (14 --+ OH + C2H3 (14 +(CH2)ZO (Sf1 + CH,CHO Previous work,5'6 has shown that channels (lu)and (lb) account for almost 100% of the products.Channel (lc) accounts for no more than 5%,6 and the occurrence of channel (Id) is uncertain. Channel (le) is a very minor route7 (k,,/k, = 0.0033) at ambient temperature, and channels (Sf) and (lg) are only important at pressures greater than 1 atrn.'" More recently, Pagsberg et al." have obtained a lower limit for the yield of CH, radicals from the reaction of 0 atoms with ethene of 0.70 at atmospheric pressure. The possible products from the reaction of 0 atoms with propene are given below: 0 + C,H, 4CH, + CH,CHO (24 + H + CH,CHCHO (24 .+C,H5 + HCO (24 -,CH,COCH, (24 +CH,CH,CHO (24 +CH,CH(O)CH, (2f) Some results'." on this system suggest that channels (2a) and (2b)have branching ratios of ca. 0.20 and 0.10, respectively, and the rest of the products are stable, being principally 1,2-epoxypropane and pr~panal.'~~' haveArutyunov et ~1.'~ shown that the alternative H-atom-producing channel, H + CH,COCH,, does not occur.Atkinson and Lloyd,', however, recommended a branching ratio of 0.2 for channels (2a) and (2c), and did not include an H-atom-producing channel. The possible products from the reaction of 0 atoms with (E)-but-2-ene are given below : 0 + C4H8-,CH, + CH,CHCHO (34 + CH, + C2H,C0 --+ C2H5+ CH,CO .+C,H7 + HCO +(CH,),CHCHO + CH,COC2Hs +(CH,CH),O (34 (34 (34 (34 (3f) (39) The detection of the methylvinyloxyl radical, CH,CHCHO [reaction (3u)lhas been reported by Hunziker "from (Z)-but- 2-ene. Atkinson and Lloyd', have recommended a yield of 0.20 for reaction (3a).However, Ferrieri and Wolf,', using 'hot' 0 atoms, reported the products of reaction (3) to be pairs of alkyl and acyl radicals, as shown in reactions (3b)- (34.The stable products [reactions (3e)-(3g)] were detected in the work by Cventanovic.12 In the present work an attempt was made to measure the yield of the primary radical products methyl (CH,) and formyl (HCO) from the ethene, propene and (E)-but-2-ene systems, and also the yield of vinyloxyl radicals (CH,CHO) from the propene system. Experimental Pulse radiolysis coupled with UV absorption was used to generate the species necessary to initiate the chemistry and monitor the initial products of the reactions.The decay traces consist of a very rapid rise in the absorption signal which then decays to a plateau. The spectral analysis used only the prompt signal since the absorption at longer times may have been corrupted by secondary products of the radical reac- tions. A sample decay trace, due to production and loss of the vinyloxyl radical is given in Fig. 1. The signal-to-noise ratio reflects the low concentration of the vinyloxyl radical, and also the fact that its absorption cross-section is low. The signal-to-noise ratio is much improved when monitoring absorbance owing to the methyl radical produced from radiolysis of C0,-alkene mixtures, because the absorption cross-section of the methyl radical is much larger. The pulse radiolysis equipment has been described in detail bef~re.'~,~*Briefly, a 1 dm3 stainless steel cell based on the design of Gordon et a!." was filled with the appropriate gases using an all-stainless-steel vacuum line.The gas mix- tures consisted of up to 10 mbar of the alkene made up to ca. 1010 mbar with either CO, or SF, depending on the atomic species required to initiate the chemistry. The pressures of the gases were measured by an MKS Baratron 170 series absol- ute electronic membrane manometer with a resolution of lop2mbar. Gas mixtures were irradiated with electrons by charging a Febetron 705B Field Emission Accelerator to 30 kV, and then discharging it. The discharge produces a 30 ns pulse of 2 MeV electrons which irradiate the cell. The electrons cause the CO, to decompose to CO and ground-state oxygen atoms which may then be exploited in reactions involving this species.lllllllllll(IIIIJI ! I .--30 0 180ti me/ps Fig. 1 Sample decay trace from the radiolysis of an SF,-CH,CHO mixture. The rnaximum absorbance was 0.033 (base 10) at 348.3 nm using a pathlength of 120 cm. The signal-to-noise ratio was greater when monitoring the methyl radical from C0,-alkene systems. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 CO, + 2 MeV e- --+ [CO,]* -+CO + 0 (4) Excited oxygen atoms, O('D), are not produced.20 Also, irra- diation of SF, yields F atoms, SF, + 2 MeV e- + [sI'6]* -+ SF, + 2F (5) which are used to generate radicals of interest as shown below. The relatively high electron capture potential of CO, and SF,, and the low concentrations of reactants ensures the latter are not irradiated directly.The build-up and decay of an absorbing species is moni- tored by UV absorption. The source of the analysing UV light is a 150 W Xe lamp. A pulsing device synchronised with the electron pulse from the febetron increases the lamp inten- sity up to 50 times.21 Optical feedback is also incorporated so that the pulse is of uniform intensity over a few ms. The light from the Xe lamp is reflected through the cell several times by a pair of mirrors in the cell, improving the signal-to- noise ratio.,, The wavelength of interest was selected via a McPherson 1 m scanning monochromator. The monochromator was cali- brated in the UV region by using the mercury emission lines in the 250-370 nm region, and in the visible region using the emission lines at 614.30 and 616.36 nm from a neon Pen-Ray Rare Gas Lamp (Ultra Violet Products).The intensity of light at the selected wavelength was measured by two types of detector. First, the signals from a Hamamatsu R955 photo- multiplier were stored by a Biomation 8100 digitiser and transferred to a DEC PDP-11/23 minicomputer for sub- sequent analysis. The minimum sampling time was 20 p,and the maximum 1 ms. The formation of the products is very rapid (see Fig. l), and no loss of product occurs on this time- scale. Secondly, an EG & G 1420 solid-state detector (diode array) was used, which allowed the absorbance over a 14 nm range to be monitored at the same time, the centre of the range being controlled by the monochromator used above.The diode array was controlled by a computer uia an EG & G model 1461 detector interface, and signals from the detec- tor were transferred directly to the computer. A mechanical shutter was placed over the entry port of the monochromator to reduce stray light. The shutter was only open when the diode array itself was operating. The opening of the shutter could be delayed, and the exposure time of the diode array could also be varied to allow the optimum signal-to-noise ratio to be found. The diode array thus records the average absorbance over the exposure time. This offsets the lower sensitivity of this detector relative to the photomultiplier.All absorbances quoted are base 10. Materials Commercially available high-purity gases were used as sup- plied with no further purification. CH, was supplied by Dansk Ilt, CO, by AGA, SF, by Ausimont, and C,H,, C3H, and (E)-but-Zene by Matheson. CH,CHO was supplied by Aldrich, and was thoroughly degassed before use. Results and Discussion Calibration of Yield of Atomic Oxygen The yield of atomic oxygen was found by radiolysing a mixture of 960 mbar CO, and 40 mbar of 0,. The atomic oxygen produced reacts with excess oxygen to produce ozone, 0 + OZ(+M)+O,(+M) (6) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 where M represents the bath gas (CO,). The final concentra- tion of ozone is equal to the initial yield of atomic oxygen, and may be found by converting the absorbance to a concen- tration using a well established value for the absorption cross-~ection~~at 253.7 nm.960 mbar CO, gave an initial yield of 0 atoms of (7.30 & 0.14) x lo1, ~m-~,the error is lo. If the initial concentration of a radical product from the reaction of 0 atoms with the alkenes is known, the branching ratio for that radical may be found by dividing the radical concentration by the 0-atom concentration. Calibration of Yield of Atomic Fluorine F atoms are used to generate formyl radicals from HCHO for spectral calibration purposes, as will be shown later. The yield of F atoms from the radiolysis of SF, was found by radiolysing an atmosphere of SF, containing a few mbar of CH,.,, The F atoms produced react very rapidly with the excess CH, producing HF and CH, radicals, the latter having a strong and well characterised absorbance at 216.4 nm.SF, + 2 MeV e- .+SF, + 2F (5) F + CH, --+ CH, + HF (7) 2CH3(+M) .+C,H,( +M) (8) The maximum absorbance of CH, radicals may be converted to a concentration using, once again, a well established value for the absorption cross-~ection.~~ The concentration of CH, radicals is thus the initial yield of F atoms, which is (2.34 & 0.04) x 1015 cmP3. The error (1 a) reflects small uncertainties in the absorbance. Yield of the Methyl Radical (CH,) from Q + C2H4 As mentioned earlier, Pagsberg et al." have obtained a lower limit of 0.70 for the yield of methyl radicals from the reaction of 0 atoms with ethene at atmospheric pressure. In the present work, a mixture of 10 mbar CzH4 in lo00 mbar of CO, was radiolysed, and the absorbance at and around 216.4 nm was monitored.By comparison of the yield of methyl radicals with the initial yield of 0 atoms, the relative yield of methyl radicals was found from the measured ratio, [CH,],,JIO]o, which is 0.70 & 0.05, consistent with the pre- viously derived lower limit." Yield of the Formyl Radical (HCQ) from 0 + C2H4 The formyl radical (HCO) is the most likely partner of the methyl radical in reaction (la). Only one previous study is known where the formyl radical was detected directly as a product from the reaction of 0 atoms with ethene" and this was achieved by monitoring the product HCO using absorb- ance of light at 563.2 nm.For the present work it was necessary to establish the absorption cross-sections of HCO in the UV region. To achieve this, a glass tube containing paraformaldehyde was attached to the vacuum line, and heated with a hot-air gun. When heated, the paraformaldehyde decomposes to meth- anal, HCHO. A mixture consisting of 4 mbar of HCHO and 995 mbar SF, was radiolysed, and a signal was observed which increased very quickly before decaying via second-order kinetics, indicating a species which is self-reacting. These observations are attributed to the following reactions : SF, + 2 MeV e- .+ SF, + 2F (5) F+HCHO-+HF+HCO (9) 2HC0 -+ products (10) The absorption cross-section at 230 nm was calculated rela- tive to the methyl radical at 216.4 nm, i.e.a(HC0, 230 nm)/a(CH,, 216.4 nm) = 0.215 f0.009. Using the CH, cross- section,' of 4.1 x cm2 molecule-1, the cross-section of HCO at 230 nm is calculated to be (8.8 f0.4) x cm2 molecule- '. To identify the products of the reaction of 0 atoms with ethene, a mixture of 10 mbar of ethene in 1000 mbar of CO, was radiolysed and the absorbance due to the products was monitored between 215 and 260 nm with the photomultiplier. This spectrum is given in Fig. 2(a). A diode array detector centred at 220 nm yields a similar spectrum as shown in Fig. 2(b). Both these figures clearly show the distinctive absorb- ance of the methyl radical and a broad featureless absorption spectrum due to other products, of which HCO may be one.The yield of HCO, if produced, should be the same as that for the methyl radical, 0.70 [see reaction (la)]. Also shown in Fig. 2(a) is the predicted absorbance due to HCO. The HCO spectrum of Pagsberg et ~1.'~was used, calibrated using a(HC0, 230 nm) = 8.8 x 10-l8 cm2 molecule- as described above. Clearly, the shape of the product spectrum is not com- patible with the expected HCO absorbance under these con- ditions. This indicates that HCO may not be produced in our system, and to test this further an attempt to monitor the formyl radical uia its absorption at 614.6 nm26 was made. Once again, it was necessary to establish the absorption 0.40 1(a) 0 210 220 230 2io 250 260 wavelengt h/n m nui v g 200-C e 100-01 800 600 400 200 0 diode number Fig.2 (a)Product spectrum recorded with the photomultiplier from the radiolysis of 10 mbar ethene in lo00 mbar CO, at 298 K. Circles mark experimental points. The solid line is a spline curve fit to the data. The dotted line shows the predicted absorbance of the HCO radical if it were made in the same yield as the CH, radical, 0.70. (b) Product spectrum recorded with the diode array centred at 220 nm from radiolysis of 10 mbar ethene in lo00 mbar CO, at 298 K. The gate opening was delayed by 5ps, and the diodes were exposed for 10 ps. The slitwidth was 1 mm. The wavelength range recorded is 213-227 nm. 3628 cross-section at this wavelength and thus to determine the minimum detection limit (this is discussed later).The feature was recorded with the diode array using the SF6-HCH0 method described above, and is given in Fig. 3. The absorp- tion cross-section of HCO at 614.6 nm was calculated using the method described above as (1.6 0.2) x lo-'' cm2 molecule- ', where the error reflects uncertainties in the absorbance measurement and to a lesser extent the F atom yield. Baggot et who were studying the reaction HCO + HCO -+ HCHO + CO (10) obtained k,o/0614.59 = (3.2 k 0.8) x 10' cm s-'. They deter- mined 0614.59 to be 2.3 x cm2 molecule-1 in separate experiments, from which k,, = 7.5 x lo-" cm3 molecule-' s-'. However, the recent evaluation by Baulch et aL6 recom-mends klo = 5.0 x lo-" cm3 molecule-' s-'.Using the above ratio for kl()/0614.s9 and the recommended value for k,,'j yields 0614.59 = (1.5 f 0.4) x cm2 molecule-', in excellent agreement with the value determined in this work, (1.6 & 0.2) x cm2 molecule-'. An attempt was made to measure the HCO yield in the C0,-C,H4 system with the diode array using the same setup as was used to record the spectrum in Fig. 3, and also with the photomultiplier, as this detector is more sensitive than the diode array. A mixture of 20 mbar ethene and 975 mbar CO, was made and radiolysed, using a pathlength of 120 cm. No absorption at 614.6 nm was seen with either detector. 975 mbar of CO, produce 7.41 x 1014 atoms cmP3 of oxygen.Using the 70% yield of CH, derived in the present work, the expected absorbance due to HCO at this wavelength which should have been observed is 0.043. For the experimental set-up used here, this absorbance is sufficiently high to be easily distinguished from the background noise with both detectors. With the photomultiplier, assuming a signal-to-noise ratio of 1 : 1, the smallest absorbance which can be dis- tinguished from the background noise is 0.002. This means that, if HCO was produced, its concentration must be less than 2.4 x lo', molecules cm-,, and hence the branching ratio for production of HCO is d0.03. Yield of CH, Radicals from 0 + C,H6 There is some, albeit limited, information to suggest that the methyl radical, CH,, is produced with the vinyloxyl radical from the reaction of atomic oxygen with propene,' 0 + C,H6 +CH, + CH,CHO (24 A mixture containing 5 mbar of propene in 1000 mbar CO, was radiolysed, and the maximum absorbance was measured 0.10 -0.08 w.-C 0.06 '10 kI v 8 0.04 Km 0$ 0.62 n mI J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 between 213 and 219 nm at ca. 1 nm intervals using the photomultiplier. The spectrum is given in Fig. 4. The shape of this spectrum confirms that the absorbance was due to CH, , and the magnitude allows the concentration of CH, radicals to be calculated. As the initial concentration of 0 atoms is also known, the branching ratio for CH, production may be found as already described, yielding a branching ratio of 0.15 k 0.01, where the error reflects uncertainties in the absorbance readings and the 0 atom concentration.Yield of the Formyl Radical from 0 + C,H6 The formyl radical is thought to be produced from the reac- tion of atomic oxygen with propenel, as shown below: O 4-C3H6 +C2H5 + HCO (24 The branching ratio for this channel has been recommended by Atkinson and Lloyd', to be 0.20 at atmospheric pressure. A mixture of 5 mbar propene and 1005 mbar CO, was radiolysed. No absorption that could be attributed to HCO was seen even at the longest possible pathlength, 120 cm. Once again, the concentration of HCO must be less than 2.4 x lo', molecules cm-,, and so the upper limit for the branching ratio for HCO in this system is 0.03, the same value as that obtained for the ethene system.Reaction of F Atoms with CH,CHO: Yields of CH,CO and CH,CHO The UV absorption spectrum of the vinyloxyl radical published by Hunziker et d." between 280 and 360 nm con- sists of a broad continuum up to ca. 330 nm, and then four distinct absorption bands between 330 and 360 nm. In order to measure the yield of CH,CHO from the reaction of 0 atoms with propene, the absorption cross-sections of this species were needed. The reaction of F atoms with CH,CHO will produce both acetyl radicals, CH,CO, and vinyloxyl radicals, CH,CHO, as shown below: F + CH,CHO -+ CH,CO + HF (34 F + CH,CHO +CH,CHO + HF (3b) The aldehydic C-H bond is weaker than the methyl C-H bonds, and so the yield of vinyloxyl radicals is likely to be less than the statistical value of 0.75.Bartels and -0.30-w.-C 3 ujn 2 0.25-a Cm 4;0.20-m 0. 51 , I I I0.00.-, I----17--800 600 400 200 213 214 215 216 217 218 219 220 diode number wavelength/nrn Fig. 3 Spectrum of the HCO radical recorded with the diode array Fig. 4 Spectrum recorded from radiolysis of 5 mbar propene in centred at 614.6 nm following radiolysis of 4 mbar HCHO in lo00 1000 mbar CO, at 298 K, showing absorbance due to the methyl mbar of SF,. The gate opening was delayed by 0.2 ps, and the diodes radical. Circles mark experimental points. The solid line is a spline were exposed for 2 ps. The slitwidth was 0.5 mm. curve fit to the data. J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 HoyermannZ7 have measured the branching ratio for the reaction of F with CD,CHO as shown below: F + CD,CHO --+ CD,CO + HF (114 F + CD,CHO +CD,CHO + DF (W They measured the yields of HF and DF, and obtained a ratio of HF :DF of 1.0 & 0.3. On this basis the yield of vinyl-oxyl from reaction (3) may be expected to lie between 0.50 and 0.75. The acetyl radical does not obsorb, or absorbs very weakly indeed, above 280 nm," and so will not interfere with absorption by the vinyloxyl radical over the wavelength of interest, 330-355 nm. A mixture of 5 mbar of CH,CHO and 1000 mbar SF, was radiolysed. The absorption spectrum of the products was recorded between 215 and 250 nm, where the acetyl radical absorbs (Fig.5), and between 330 and 355 nm where the vinyloxyl radical absorbs (Fig. 6). Using previously derived absorption cross-sections for CH,CO, the concentration of this species may be found. This spectrum has been recorded by Adachi et Parkes" and Pagsberg et a1." The spec- trum of Adachi et ~1.'~differs considerably from the other two both in shape and size of the absorption cross-sections. At 215 nm, ParkesZ9 and Pagsberg et ~1.'~obtained 1.4 x and 1.34 x cm2 molecule-', respectively, whereas Adachi et obtained 3.8 x cm2 molecule-'. In view of the agreement between the values of the absorption cross-section at 215 nm of ParkesZ9 and Pags- berg et a1.,18and the complicated method used by Adachi et aLZ8to calibrate their results, a value of 1.34 x cm2 molecule-' at 215 nm was used in this work to calibrate the CH,CO spectrum.' As the initial F atom yield is also known from the cali- bration work described earlier, the concentration of the vinyl- oxyl radical may be found by subtracting the acetyl radical concentration from the concentration of F atoms; by measur- ing the absorption due to the vinyloxyl radical over the range 330-355 nm, the absorption cross-sections for this species may then be calculated. The average concentration of CH,CO was measured as (1.42 f0.10) x 1015 molecules ern-,.As the F atom yield is known, (2.34 & 0.04) x lo", the concentration of the vinyl- oxyl radical is (9.2 & 1.1) x 1014 molecules cm-,, and the yield from reaction (3) is 0.39 0.05.This is close to the value of 0.50 & 0.15 obtained by Bartels and H~yermann.'~ The UV absorption spectrum (Fig. 6) of the vinyloxyl radical, recorded in the present work between 330 and 355 nm, may 3629 c& 100-3 0 -80 N 6 60 I 0 F1 5 40.-w ?:20 gC .g 0 332 336 340 344 348 352 356 za wavelength/nm Fig. 6 Absorption spectrum of the vinyloxyl radical, CH,CHO, at 298 K generated from tadiolysis of an SF,-CH,CHO mixture. Circles mark experimental points. The solid line is a spline curve fit to the data. now be calibrated as the concentration of the vinyloxyl radical is known. Although the vinyloxyl radical is a weak absorber (the maximum absorption cross-section, at 337.3 nm, is 9.4 x cm2 molecule-'), it should be made in sufficient quantities to be detected in the present work.The shape of the vinyloxyl radical spectrum (see Fig. 6) is in good agreement with the spectrum of Hunziker et al.," but the absorption cross-sections obtained here are on average 2.7 times larger. The reason for this discrepancy is not clear at this time. Yield of the Vinyloxyl Radical (CH,CHO) from 0 + C,H, The vinyloxyl radical is produced in the propene system as shown below: 0 + C3H, -P CH, + CH,CHO (24 Mixtures of 5 mbar C,H, and 1005 mbar CO, were radiolysed, and the absorption spectrum (Fig. 7) of the pro- ducts was recorded between 335 and 338 nm, and 346 and 349 nm using a pathlength of 120 cm. Another peak was observed in this work and by Hunziker et al." between 338 and 346 nm, but the absorption here would not be expected 10-z 9-.-+ C 3 uj 8-n (El m4 0.6 Q '-v v 0.41 als 0.5 E 6-9 92 0.3 4a 5-a 0.2 0.0 1O.' 334 336 338 340 342 344 346 348 350 215 220 225 230 235 240 245 250 wavelength/nm wavelength/nm Fig.7 Spectrum recorded from radiolysis of 5 mbar propene in Fig. 5 Absorption spectrum of the acetyl radical, CH,CO, at 298 lo00 mbar CO, at 298 K. Circles mark experimental points. The K. Circles mark experimental points. The solid line is a spline curve solid line is a spline curve fit to the data. The peak at 337.3 nm was fit to the data. assigned to the vinyloxyl radical. J. CHEM. SOC.FARADAY TRANS., 1994, VOL. 90 to be visible above the noise. For this reason the spectrum shown in Fig. 7 has a gap in this spectral region. A sharp peak with a maximum absorbance at 337.3 nm of 0.01 was observed in the first wavelength range, with a 'shoulder' at a slightly shorter wavelength, but no peak was seen in the second range. The peak at 337.3 nm was assigned to the vinyloxyl radical as the shape and position coincide with a peak observed in the vinyloxyl radical spectrum (see Fig. 6). If the shoulder is extended underneath the vinyloxyl radical peak, the absorbance due to the vinyloxyl radical is 0.004 only. Calculating the concentration of CH,CHO (using the absorption cross-sections obtained in the present work), and dividing by the initial 0 atom concentration means that the branching ratio for CH,CHO is 0.12 f 0.03, in good agreement with the measured yield of methyl radicals, 0.15 f 0.01.The error is larger for the vinyloxyl radical yield than the methyl yield due to greater uncertainties in the absorbance readings. The species responsible for the 'shoulder' at 336 nm is not known at this time. It is almost certainly not due to HCHO, C,H,CHO, CH,COCH, or 1,2-epoxypropane because the absorption cross-sections of these species at this wavelength are too low. The shoulder may be due to the methylvinyloxyl radical, CH,CHCHO; the UV absorption spectrum of this species has been recorded by Hunziker and Wendt but has not been published.' The monochromator slitwidth was 1.0 mm, equivalent to a bandwidth of 0.8 mm.Using a smaller slitwidth would have yielded a smaller bandwidth and thus possibly a more pro- nounced feature at 337.3 nm, but would have resulted in a poorer signal-to-noise ratio. The experiment above was repeated using the diode array detector. Owing to the lower sensitivity of this detector, the signal-to-noise ratio was very poor, but a peak at 337.3 nm, which is attributed to the vinyloxyl radical, was observed with a signal-to-noise ratio of less than 1 : 1. Products from the Reaction 0 + (E)-But-Zene A spectrum of the products from the reaction 0 + (E)-but-2-ene was recorded with the photomultiplier and is given in Fig. 8. From the shape of the feature centred at 216.4 nm, it is clear that the methyl radical is produced in this system.From the height of the absorbance due to the methyl radical, the yield of this species is 0.25 +_ 0.08. The large error reflects the 0.24 i"i E 0.201 I I uj .2! 0.15 0.00 +- I I ! I 1 210 220 230 240 250 260 wavelength/nm Fig. 8 Spectrum recorded from radiolysis of 10 mbar (E)-but-Zene in 1OOO mbar CO, at 298 K. Circles mark experimental points. The solid line is a spline curve fit to the data. uncertainties in establishing a baseline for the absorbance peak. Other products are clearly present in this system, and the absorption spectrum must be due to either the stable pro- ducts or radical species. An attempt was made to fit the rest of the product spectrum obtained from this system using a combination of channels (3b)-(3d), as the absorption cross- sections of the products have been measured.Although the basic shape of the fitted spectrum was similar to the experi- mental one when a combination of channels (3b), (3c)and (36) was used, the absorbance at 216.4 nm was 50% too low. However, it was assumed that the absorption spectrum of C2H,C0 was identical to that of CH,CO; if it is weaker, then the overall quality of fit would be markedly improved. Further analysis of this system must await an accurate deter- mination of the absorption cross-section of C2H ,CO. Discussion The methyl radical has been detected in the ethene system as expected from the earlier work of Pagsberg et al." with a yield of 0.70 k 0.05.This implies that the yield of the vinyl- oxyl radical is 0.30 f. 0.05. Hunziker et al." measured the yield of the vinyloxyl radical between 53 and 1000 mbar as 0.36 0.04, which is independent of pressure. Smalley et obtained a yield for this species of 0.27 f 0.05 and also inde- pendent of pressure between 67 and 133 mbar. If the results of Smalley et aL30 are also valid for atmospheric pressure, the average of these two values is 0.32 k 0.06, in good agreement with our value of 0.30 f 0.05. The methyl radical was also observed in the propene system with a yield of 0.15 f. 0.01. The vinyloxyl radical yield in the propene system was measured as 0.12 k 0.03, in good agreement with the methyl radical yield, but lower than the vinyloxyl radical yield of Hunziker et al.," who obtained a value of 0.19.The methyl radical yield obtained in the 0 + (E)-but-2-ene system, 0.25 f 0.08, is in reasonable agree- ment with the yield recommended by Atkinson and Lloyd', of 0.20 for the methyl radical, and the results obtained in Part 1 of this paper' when it was assumed the only products from the reaction 0 + C,H, are CH, + CH,CHCHO; in this latter case, the yield of the methyl radical was 0.20 and 0.18 at 20 and 30 mbar, respectively. The product absorption spectra for the ethene and (E)-but- 2-ene systems clearly show that other species must be present. Other potential products from the ethene system are the ethyl and vinyloxyl radicals. The ethyl radical may be formed by reaction of product H atoms produced with the excess ethene.However, the absorption cross-sections of this species are too low to reproduce the spectra of Fig. 2(a) and 2(b).The only other species made which could be responsible is the vinyloxyl radical. The absorption spectrum of this species has not been measured below 280 nm. Similarly, the absorption spectra of methyl ethyl ketone, 2,3-epoxybutane and 2-methylpropanal which may be made in the reaction of 0 atoms with (E)-but-Zene are either too weak or the wrong shape to fit this spectrum. By analogy with the ethene system, it is proposed that the absorbance observed may be due to the methylvinyloxyl radical. The formyl radical was not detected in the ethene or propene systems.It has, however, been observed previously by Hunziker et a!." in the ethene system using absorption of visible light at 563.2 nm. The yield of HCO in the ethene system was measured by these authors as 0.52 0.06 at atmospheric pressure, which is considerably lower than might have been expected based on the methyl radical branching ratio obtained in the present work. Using the yield of Hunzi- ker et d.,' the theoretical absorbance due to HCO under the conditions used here would be 0.032, easily visible above the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 3631 noise. The lack of HCO absorbance suggests that, under the conditions employed here, the methyl radical producing channel may proceed mainly via the reaction 3 4 B.J. Finlayson-Pitts and J. N. Pitts Jr., Atmospheric Chemistry. Fundamentals and Experimental Techniques, Wiley, New York, 1986, ch. 6 and 7. W. P. L. Carter, Atmos. Environ. A, 1990, 24,481. 0 + C2H4 + [O -C,H,]* -+ CH3 + H + CO (la') and to a very much lesser extent CH, + HCO. Reaction (la') is 51.9 kJ mol-' exothermic and the products of reaction (la') may be formed directly from decomposition of the ener- gised biradical, as shown above, or via decomposition of the vinyloxyl radi~al.~'.~, As HCO was not detected in either the ethene or propene systems, the upper limit on channels (la) 5 6 V. D. Knyazev, V. S. Arutyunov and V. I. Vedeneev, Int. J. Chem. Kinet., 1992,24, 545. D. L. Baulch, C. J. Cobos, R. A. Cox, C. Esser, P. Frank, T. Just, J. A. Kerr, M.J. Pilling, J. Troe, R. W. Walker and J. Warnatz, J. Phys. Chem. Ref. Data, 1992, 21,41 i. K. Mahmud, P. Marshall and A. Fontijn, J. Phys. Chem., 1987, 91, 1568. R. J. Cventanovic and D. L. Singleton, Rev. Chem. Intermed., 1984, 5, 183. and (2c) is 0.03. More work is needed to establish detailed UV absorbance spectra of the products from the reaction of 0 with alkenes, and also of the potential products, such as CH,CHO and CH,CHCHO. However, in an atmospheric context it is not relevant if the co-product of the methyl radical is H or HCO, as both of these species are rapidly 10 11 U. Bley, P. Dransfeld, B. Himmer, M. Koch, F. Temps and H. G. Wagner, 22nd Symposium (International) on Combustion (Proc.),The Combustion Institute, Pittsburgh, 1988, p. 997.P. Pagsberg, E. Ratajczak and A. Sillesen, in Research in Chemi-cal Kinetics, ed. R. G. Compton and G. Hancock, Elsevier, New York, 1993, vol. 1, 65. H. E. Hunziker, H. Kneppe and H. R. Wendt, J. Photochem., oxidised to HO, . From the work at low pressure,' a yield for total radical production of 0.36 & 0.07 at 40 mbar total pressure was obtained for the reaction of 0 atoms with propene. Using the methyl radical yield obtained in the present work, 0.15 0.01, an upper limit for all other radical-producing 12 13 14 15 1981, 17, 377. R. J. Cventanovic, Can. J. Chem., 1958, 36, 623. R. Atkinson and A. C. Lloyd, J. Phys. Chem. Ref. Data., 1984, 13, 315. V. S. Arutyunov, V. I. Vedeneev and V. D. Knyazev, Khim. Fizika,1990,9, 1383 (in Russian). H. E. Hunziker, Personal communication to R.J. Cventanovic routes of 0.21 & 0.07 at atmospheric pressure may be derived (e.y. H + CH,CHCHO). As shown above, the intensity of the methyl radical absorb- ance at 216.4 nm in Fig. 8 allowed the yield of methyl rad- icals from the reaction of 0 atoms with (E)-but-2-ene to be calculated as 0.25 & 0.08. Again, it is suggested that the rest 16 17 18 and D. L. Singleton (ref. 8). R. A. Ferrieri and A. P. Wolf, J. Phys. Chem., 1992,%, 4747. P. Pagsberg, J. Erikson and H. C. Christansen, J. Phys. Chem., 1979, 83, 582. P. Pagsberg, 0.J. Nielsen and C. Anastasi, in Advances in Spec- troscopy, ed. R. E. Hester and R.J. H. Clark, 1994, vol. 24, in the press. of the product spectrum in Fig. 2(a) and 2(h),and Fig. 8 may be due to CH,CHO and CH,CHCHO, respectively.However, the absorption cross-section of CH,CHCHO at 230 nm would have to be 1.15 x cm2 molecule-' to fit the observed spectrum assuming a yield of 0.25, which is much larger than the absorption cross-sections of many other 19 20 21 22 23 S. Gordon, W. Mulac and P. Nangia, J. Phys. Chem., 1971, 75, 2057. M. C. Sauer Jr., in Advances in Radiation Chemistry, Wiley-Interscience, New York, 1976, vol. 5, 97. T. Hviid and S. 0.Nielsen, Rev. Sci. Instrum., 1972, 43, 1198. J. U. White, J. Opt. SOC.Am., 1942,32, 285. J. Barnes and K. Mauersberger, J. Geophys. Res. D, 1987, 92, alkyl radicals. Once again, more work is needed to establish the cross-sections of the vinyloxyl, methylvinyloxyl and propanoyl radicals between 210 and 280 nm. 24 25 14861. P. Pagsberg, J. Munk, A. Sillesen and C. Anastasi, Chem. Phys. Lett., 1988, 146, 375. M. T. Macpherson, M. J. Pilling and M. J. C. Smith, J. Phys. The authors are indebted to K. B. Hansen for his efforts in setting up and maintaining the pulse radiolysis equipment, 26 Chem., 1985,89,2268. J. E. Baggott, H. M. Frey, P. D. Lightfoot and R. Walsh, Chem. Phys. Lett., 1986, 132, 225. and Agnes Vehovsky and Alfred Anston for translating ref. 14. M.G.S. would like to thank the SERC for a research grant. 27 28 M. Bartels and K. Hoyermann, An. Asoc. Quim. Argent., 1985, 73, 253. H. Adachi, N. Basco and D. G. L. James, Znt. J. Chem. Kinet., 1981,13, 1251. 29 D. A. Parkes, Chem. Phys. Lett., 1981,77, 527. References 30 J. F. Smalley, F. L. Nesbitt and R. B. Klemm, J. Phys. Chem., 1986,90,491. 1 C. Anastasi and M. G. Sanderson, J. Chem. Soc., Faraday 31 D. J. Donaldson and J. J. Sloan, J. Chem. Phys., 1982,77,4777. Trans., 1994,90, 3617. 32 D. J. Donaldson and J. J. Sloan, Can. J. Chem., 1983,61, 906. 2 Photochemical Smog and Ozone Reactions: Advances in Chem- istry Series 113, ed. R. F. Gould, American Chemical Society, Washington DC, 1972. Paper 4/04264F; Received 12th July, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003625

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3633-3638 Excimer-laser Photofragmentation of Boron Trichloride and its Implications for Laser Vapour Deposition and Doping Processes S. Georgiou,* E. Raptakis, X. Xing, E. Hontzopoulos and Y. P. Vlahoyannis Foundation of Research and Technology-Hellas , Institute of Electronic Structure and Laser, PO Box 1527,lraklion , Crete, Greece The photolysis of BCI, at the excimer-laser wavelengths of 193 and 248 nm has been investigated by mass spectroscopic and fluorescence techniques. At 193 nm, the BCI (A 'II), excited B (B*) and BCI,+ (x = 0, . .. , 3) species are detected, whereas at 248 nrn, BCI (A 'II),B*, CI, (2311,) and BCI,' (x = 1, 2) are detected. The pressure and fluence dependences of the signals are determined. At 193 nm, BCI, is proposed to undergo one- and two-photon excitations, the former resulting in BCI, and the latter in BCI (A 'n) and BCI,' (x = 3, 2).In the case of the 248 nm excitation, the observed emitting species are shown to derive from secondary reactions that take place following two-photon photoexcitation/fragmentation of BCI, . Implications of the observations herein for the vapour and doping processing by laser of BCI, are discussed. In particular, for processing at 248 nm, the observed photoinduced reactivity may be significant and should be considered in addition to any pyrolytic mechanisrn. The growing importance of laser chemical vapour deposition techniques (LCVD) has necessitated a re-examination of the photochemistry of several molecules with the purpose of opti- mizing the corresponding applications. In several cases, examples of rich photochemistry have been uncovered, which are important both for their technological relevance, but also in their own right (e.g.for the information they provide for the photophysics of the corresponding compounds). BCl, is a molecule of considerable technological interest, as it is used extensively in the plasma etching of aluminium, silicon, etc.' Recently, excimer lasers have been used to induce fragmentation of the molecule for the deposition of boron2,, or boron-alloy coating^,^ and for gas immersion laser d~ping.~.~ Several interesting observations were report- ed, which are worth detailed examination. Further interest in BCl, derives from the fact that it is a high-symmetry, 'closed- shell' planar molecule which, upon electron attachment or excitation to higher electronic states, is expected to undergo significant structural changes resulting in non-statistical photodissociation dynamics.Despite this multifaceted interest, very little is known about the photophysics of the molecule in the ultraviolet and visible (UV-VIS) region. Electron-impact e~citation,~ negative-ion formation tech- niques,* photoelectron spectroscopy9 and more recently, syn- chrotron radiation in the 45-190 nm have been used to investigate the electronic structure of BCl, . Of more relevance to the photophysics of the molecule in the UV-VIS spectral region is the measurement of its absorption spec- tr~rn,~~,'~*'~but unfortunately, the results of the three studies are contradictory.Recently, Slaoui et a1." showed that, at 193 nm, BCl, undergoes one- and two-photon excitations and measured the corresponding cross-sections. The products of the excitations were not, however, identified. This lack of information on the photophysics of BCl, has prevented elucidation and assessment of the importance of the observations made in the excimer-laser processing of the molecule. Generally, there is a lack of mechanistic under- standing of the processes. At 193 nm, the boron doping process has been disc~ssed~.~ in terms of a photolytic mecha- nism, although in other cases,4 photodissociation of the mol- ecule was not considered to be significant. In contrast, for processes employing longer wavelengths, a pyrolytic mecha- nism has generally been ass~med,~,~ since the molecule does not absorb at 1 > 220 nm.However, the possibility of contri-butions from photolytic processes through multiphoton excitation/fragmentation of the molecule cannot be excluded. A comparative study of the boronization of metallic sub- strates induced by laser radiation at 193 and 248 nm was performed16 recently and differences were noted between the two cases. The observations, however, could not be explained satisfactorily on the basis of the simple photolytic us. pyroly-tic scheme. These considerations prompted us to study the gas-phase photolysis of the compound at the corresponding wave-lengths. In the present paper fluorescence and mass spectro- scopic techniques are used to establish the fragmentation patterns of BC1, at 193 and 248 nm.At both wavelengths, excitation of BCl, results in various neutral and ionic species (summarized in Table 1). At 193 nm, we confirm the report by Slaoui et al." and identify the products of the multi- photon excitations. At 248 nm, BCl, is indicated to undergo a two-photon excitation/fragmentation which at high pres- sures results in a complicated, but efficient, reactivity sequence that may contribute to LCVD processes in parallel with any pyrolytic mechanism. Experimental The fluorescence experiments were conducted in a stainless- steel flow cell which was equipped with laser entrance and exit ports, as well as with observation ports at right angles to the laser direction. The base pressure of the cell was <2 x low4 Torr (measured with a Penning gauge), and during experiments, BCl, pressures up to 50 Torr were estab- lished.The emission from the excited fragments was focused (f= 10 cm) onto a 0.25 m Jarrell-Ash scanning grating monochromator and detected with a photomultiplier tube Table 1 Species observed in the photoexcitation of BCl, wavelength/nm mass spectroscopy fluorescence 193 BCl,+ (X = 0, ..., 3) BC1 (A 'n)B (2s)'(2p)', 'D B (2~)'(3s)~,2S 248 BCl-+ (X = 1, 2) BCI (A 'n). I c12 (29-4)B (2s)'(2p)', 2D B (2s)'(2p)'(3s)', 4P B (2s)'(3s)', 'S (PMT). The PMT signal was amplified (EG&G Ortec fast amplifier), sent into a boxcar averager and output to a recorder.For recording the time profile of individual lines, the amplified PMT output was digitized by a LeCroy TR8828C transient recorder, averaged and stored in a micro- computer for analysis. The ionization experiments employed a quadrupole mass spectrometer (Elttec), housed in a stainless-steel cross. The cell was evacuated by a turbo-molecular pump to a base pressure of lop7Torr. The compound was introduced via an effusive inlet and maintained at a constant pressure of about Torr. The ions, produced by focusing the laser beam in the source region of the spectrometer, were mass selected and detected by a secondary electron multiplier (§EM).The SEM output was amplified, sent to a boxcar averager and fed into a microprocessor. Both fluorescence and ionization experiments used the output of a Lambda Physik LPX 150 excimer laser, oper- ating either on ArF at 193 nm (nominally 17 ns pulse duration) or on KrF at 248 nm (nominally 29 ns pulse duration). In the fluorescence experiments, the laser light was focused in the cell with a lens of 10 cm length. For the mass spectroscopic experiments, lenses of various focal lengths were used for reasons to be explained in the next section. Focal areas were estimated from the beam image on ther- mosensitive paper. BCl, (Aldrich) with a stated purity of 99.95% was used as received. Absorption spectra (PBC130.5-15 Torr in a 10 cm = long cell) were measured on a Perkin-Elmer UV-VI§ spectrometer. Results Absorption Spectrum The absorption spectrum of BCl, exhibits a diffuse structure- less band with a maximum at 208 nm and a FWHM of 6 nm. Both the band area and the peak height increase linearly with pressure (correlation coefficient, r = 0.994), showing that the band is not due to dimer or polymer formation.Plots of the peak intensity at 208 nm vs. pressure yield an absorption coefficient of 0.965 atm-' cm-' at room temperature. The position of the band, as well as the value of the absorption coefficient, support the study by Maria et all3 Another absorption feature appears at 195 nm and extends to wavelengths shorter than 190 nm (the limit of our instrument) with a value of 0.561 atm- cm-for the absorp- tion coefficient at 193 nm.This low value is consistent with the observed feature being the tail of the absorption band centred at 173 nm that has been observed previously" and assigned to an CII;+"(TC* tn)tran~ition.'~ Photofragmentation Spectra 193 nm Excitation Upon irradiation of flowing BCl, with 193 nm excimer light, the fluorescence spectrum illustrated in Fig. l(a) is obtained. The spectrum displays sharp peaks at 209 and 250 nm, and a well defined vibrational progression in the 260-280 nm region with peaks at 266, 272 and 278 nm. Of these features, the lines at 209 and 250 nm can be assigned, respectively, to the (2~)'(2p)~, 'D-(2~)'(2p)', 'Po and (2s)'(3s)', '§--(2s)'(2p)', 2P, atomic boron transitions." The resolution of our instru- ment was not sufficient to analyse the spin-orbit splitting of the peaks.The vibrational peaks at 266,272 and 278 nm cor- respond to the BCI (A 'n-X 'Z) electronic transition with Av = -1, 0 and + 1, respectively, and are in excellent agree- ment with previously observed and calculated spectra.18 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 -I 220 240 260 280 C 0 wavelengt h/n m Fig. 1 Fluorescence spectra recorded upon irradiation of flowing BCl, gas with (a)ArF and (b) KrF laser light. (a)The line near 290 nm is due to scattering of the ArF laser. (b)Note the weak structure in the 254-264 nm region and attributable to the C1, (2 311,-1 311,) transition. This structure is better defined and resolved in the second diffraction order of the spectrum.Similarly, the second-order spec- trum was used to resolve the 249.75 nm line of the B atom from the much stronger laser line in (b).In the 300-600 nm spectral region, no new intense features (other than second- or higher diffraction orders of the features already specified) are detectable, and for this reason, the relevant spectra are not depicted here. Note that no emission was detected into the 300-600 nm region, where according to the synchrotron experiment^,'^ we should observe intense emission if a three-or higher order-photon excitation of BCl, were important. For nominal laser energies ranging from 5 to 28 mJ (corresponding to laser densities of ca. 0.3-0.8 J cm-'), the intensities of all emission bands increase with approximately the square (slope of 1.9 & 0.3) of the laser energy.All lines are visible at pressures as low as 0.1 Torr and initially their intensities increase essentially linearly [Fig. 2(a)] with increasing pressure, providing strong evidence that the observed species constitute primary (nascent) photo- products of the excitation process.7 At pressures higher than 5 Torr, the graphs of fluorescence intensity us. pressure become flat. Note that the 'flattening' of the graphs occurs at almost the same pressure at which Foulon et al.' noted a t This is further demonstrated by the fact that all emissions are detectable even upon addition of an excess of He buffer gas. Their integrated intensities show an inverse linear dependence on the He pressure (of a typical Stern-Volmer type).J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 60 48 20 0 0 5 10 15 I 20 , 25 BCI, pressure/Torr Fig. 2 Pressure dependence of the intensities of the B* and BCl* emissions observed at (a) 193 nm (Elaser= 12 mJ) and (b) 248 nm (E!,,,, = 65 mJ) irradiation of BCl, . Intensities represent peak heights, but essentially identical results are obtained if peak areas are used instead. Standard deviation (n = 3) of the data points is ca. 10-15% of their y values. (a)For reasons of clarity, the data for the 209 nm line, 0,is plotted on a different y axis (indicated on the right-hand side of the plot). The linear fitting of the data up to the point that 'saturation' becomes evident gives slopes of 1.1 f0.3 for the 209 nm B line, 1.4 & 0.2 for the 250 nm B line, 0,and 0.8 f0.2 for the BCl emission, w (measured at 272 nm).(b) Least-squares fitting of the rising part of the curves indicates a limiting fourth-order dependence of the depicted emissions on the BCl, pressure. 0,C1, emission. change in the efficiency of the doping of Si in the correspond- ing ArF-induced doping experiments. The authors suggested that this effect was due to a change in the mechanism of the BCl, gas-phase photolysis. However, in this higher pressure regime, we could not detect any new emitting species indica- tive of such a change. In the lack of any evidence to the con- trary, we suppose that the 'flattening' of the curves is due rather to the operation of quenching processes.Quenching processes may still account for the observations by Foulon et a!.' (for instance, if one of the excited species is responsible for the doping process), but the exact mechanisms are not clear at present. We further investigated the fragmentation mechanism by mass-selective detection of the ionic photoproducts. At the lowest laser densities examined herein (ca. 50-100 mJ cm-2), BCI,+ and BC12+ are the only ions observed within the signal : noise (S :N) ratio of our set-up. Because the laser power dependence of the ion signals could not be determined accurately (because of low S: N ratio), we decided to compare, instead, spectra obtained with different focal length 3635 lenses.Under tighter focusing (at a laser density of ca. 0.5 J cm-2), fragmentation is extensive and the spectrum consists of strong ion signals due to all the possible boron-containing species (BC1,+, where x = 3, .. . ,0) and some C1+. The inten- sity of the various isotope components for each species estab- lish the independence of the multiphoton ionization mechanism from the fragment isotopic composition. Note that under the conditions of these experiments, ionization of authentic (neat) chlorine is observed. Thus, the absence of C1, + signal in the corresponding examination of BCl, defi- nitely rules out molecular chlorine formation at 193 nm under coilisionless conditions. Overall, these results establish that at 193 nm, under irra- diation conditions typical of LCVD processing, BCl, under- goes extensive fragmentaion and ionization.As indicated above, the fluorescence results suggest that a three- or higher-order multiphoton process does not take place. Indeed, all observed species can be fully accounted for in terms of two mechanisms, namely a two-photon excitation of the parent molecule resulting in ionization and/or fragmentation, and a parallel pathway initiated by the one-photon fragmentation of the molecule to BCl, (Fig. 3). We consider first the two-photon excitation process, where information on the assessed states is available from the pre- vious synchrotron '' and photoionization experiment^.^ Spe-cifically, upon excitation of the molecule near A = 193/2 nm, BCl (A 'II) (u' = 0) is the only species detected in fluorescence studies.' On the other hand, mass-resolved photoionization measurements7 indicate that the Rydberg state assessed in the 94.5--87.5 8, region autoionizes to produce parent mol- ecule ions and BC12+ fragments, in parallel.Thus, in agree- ment with these reports, the two-photon excitation at 193 nm can account for the observation of BCl,+ and BC12+ in our mass ionization experiments, and of BC1 (A 'II) (u' = 0) in the fluorescence ones. However, the two-photon process is not consistent with the observation of the higher vibrational levels of BC1 (A 'II) indicated in Fig. l(a) [i.e. the available energy does not suffice for the vibrational excitation of the BCl (A'II) fragment], nor can it explain the fact that the BC13+: BC12+ intensity ratio is fluence dependent.These observations indicate the operation of a second, parallel pathway initiated by the one-photon excitation of the BCl, precursor. Fragmentation at this wavelength is expected since the available energy far exceeds the bond dissociation energy of the precursor molecule (Table 2). In support of the dissociative nature of the assessed state, no emission from this state is observed in the fluorescence experiments reported here and by Lee and co-workers." As for the nature of the x lc I BC13 BCI2 BCI B Fig. 3 Schematic energy diagram for BCI, and its fragments and ions depicting the excitation/fragmentation steps that are proposed for 193 nm excitation.The diagram is based on the data of Table 2. Vertical arrows represent excitation steps, whereas dashed arrows represent fragmentation steps. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Thermodynamic and spectroscopic data for BCl, and its photofragments process BCl," BC1," BC1" B Cl,b BC1,' dissociation 4.61' 3.55 5.61 - 2.48 0.66 5.65d ionization excited statese 11.64 6.0 7.2 8.7 9.5 10.5 7.7 (A) 2.56 (B) 5.19 (C) 5.70 10.21 ('IT) 4.56 8.30 (4P)3.58 (2S) 4.96 ('D) 5.93 11.50 (2 311g)7.19 (A) 0.55 (B) 1.02 (C) 2.58 (D) 3.68 (E) 6.10 " The values are generally taken from ref. 10 and 11, unless otherwise stated. All values are given in eV. Values from ref.18. Energy for rupture of BCl, to BCl, and CI. Energy for dissociation of BCl, to BCl and C1, . For the BCl, , B and C1, species, only states relevant to this work are tabulated. photoproduct, it can be identified with some certainty as BC1, (X). Energetically, the only other conceivable alternative is formation of BCl (x)through molecular elimination of C1, (Table 2). However, the failure to detect C1,+ in the mass spectrometer provides strong evidence against this possibility. The BCl, fragment that is produced is not detected in the fluorescence experiments, because the one-photon excitation places it high up in the C state (0.7 eV or" excess energy, see Table 1) above its dissociation threshold, and the excited molecule quickly dissociates.On the other hand, it does contribute to the BC1,' signal that is observed in the mass spectrometer through a two-photon excitation process. This process accounts for the increase in the BCl,+ : BCl,' inten-sity with laser fluence. Furthermore, the two-photon excita- tion deposits sufficient energy for the formation of the higher vibrational levels of BCl (A 'n) 2x 193nmBC1, -BC1 (A 'n); U' > 0 Subsequently, the BCl (A'II) state can absorb an addi-tional laser photon and ionize or dissociate to ground-state C1 and atomic boron in the (2s)'(2p),, 2D and (2~)~(3s)', 2S states. The available energy is barely enough to populate the ,D state, which is consistent with the low intensity of the 209 nm line as compared with the 250 nm one [Fig.l(a)]. Finally, the B produced, in either ground or excited states, can be ionized, respectively, by two-or one-photon processes, thereby accounting for the strong B+ signal in the mass spectra. Some B+ may also derive from excitation and disso- ciation of the BCl' ions. In conclusion, our experimental results are in full accord with the previous ob~ervation'~ that at 193 nm, BCl, under- goes efficient one- and two-photon excitations. Of these, the one-photon process is strongly indicated to result in BCl,(X) production, where the two-photon one results in BCl, + and BCl,+, and formation of BCl (A'II) (u' = 0). The primary neutral photofragments undergo further excitations, produc- ing the species reported in Table 1.248 nm Excitation In contrast to the situation at 193 nm, the molecule does not absorb radiation at 248 nm. This has been the reason for assuming the absence of photolytic activity in the corre-sponding LVCD processes. Nevertheless, upon 248 nm irra- diation (at a laser density of ca. 0.5-1.5 J ern-,) of flowing compound, a relatively intense fluorescence spectrum [Fig. l(b)] is recorded with a somewhat richer structure than that at 193 nm. Specifically, in addition to the peaks observed upon 193 nm irradiation, we observe the (2~)'(2p)'(3s)', 4P-(2s)'(2p)2, 4P, boron emis~ionl~,~~ at 206.5 nm, and a weak, structured band extending from 254 to 264 nm and centred at 258 nm, which can be assigned to the well studied21 C1, (2 '&-l ,nu)transition. Importantly, no emis- sion is observed in the 400-500 nm region, where various excited states of BC1, and BC13+ are known to fluoresce." Logarithmic plots of the intensity of each line us.nominal laser energy before focusing yielded reasonably good straight lines (r = 0.98) with slopes ranging from 3.8 to 5.8, suggesting at least a four-photon process for the photodissociation and subsequent fragment emission. Importantly, only the BCl (A 'II-X 'X) emission is appre- ciable at pressures lower than 10 Torr. At higher pressures, all emission intensities increase sharply [Fig. 2(b)] with increasing pressure up to about 40 Torr, at which point pro- cesses evidently limit the emission intensity. The sharp pres- sure dependence seen in Fig.2(b) establishes that secondary reactions provide the major mechanism for the generation of the observed species and, in particular, for the formation of the Cl; and B* ones. This conclusion is further supported by the fact that upon addition of He buffer, all emissions dimin- ish drastically and eventually disappear, with the BCl* one being the least affected (i.e. just detectable at the highest employed pressures). As discussed later, the observed reacti- vity pattern has interesting implications for the LVCD pro- cessing of BCl, with a KrF laser. Turning next to the mass spectroscopic examination, only a weak BC1,' signal is observed at low fluences (ca. 0.3 J ern-,). No BC13+ ion could be observed under any irradia- tion conditions.When the laser fluence is increased (0.6-1.0 J cm-,), a BCl' ion is detected along with a very weak B+ signal. The very weak B+ signal is consistent with the pre- vious observation that the B* excited species are produced mainly by collisional processes at high pressures. As for the molecular chlorine, no ion signal is detected in the mass spec- trometer. However, under identical irradiation conditions, ionization of authentic (neat) chlorine is not observed either. Thus, this failure to observe C1,+ from BCl,' in the mass spectrometer is not conclusive. In summarizing the observations, the photophysics of BCl, at 248 nm is dominated by secondary reactions that take place following the multiphoton excitation/fragmentation of the molecule.Despite the complexity of the involved pro- cesses, the presented data suffice to establish the nature of the primary photoexcitation step. Specifically, since three 248 nm photons suffice to ionize the molecule, the failure to observe parent ion formation in the mass spectrometer sug- gests that only two photons are involved. Similarly, if a three- or higher-order process were involved, we should have observed fluorescence from various excited states of BCl, and BC13+ (not seen). Taken together, these results strongly support the idea that the primary photoexcitation step entails a two-photon absorption process. Indeed, at 124 nm, BCl, exhibits the maximum of a rela- tively sharp electronic band." This feature was tentatively assigned to an E" transition, in which case, a two-photon excitation would be symmetry forbidden.We, instead, ascribe J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 the band to an E' state,t which is both one- and two-photon allowed under D,, symmetry. On the failure to detect parent ion signal in the ionization experiments, it appears that the assessed state is strongly dissociative so that fragmentation competes effectively with up-pumping. We advance the following stepwise excitation/frag-mentation scheme to account for the ionic pattern observed in the mass spectrometer: BC1,'t BCl,/ (?) hv (" Maximum number of photons required for the indicated process, as required if BCl, is produced in the X state.) This scheme is based on the fact that for BC1, in the X state, one KrF photon does not suffice for excitation to the B state (at 5.19 eV, see Table 2), while any BCl, produced1 in the A or B states will probably absorb additional photon(s) and ionize or dissociate.Thus, the failure to observe BCl: emission in the fluorescence experiments is understandable. As for BCl', its generation from BC1, (X),requiring at least three photons, may occur in a direct or stepwise manner. We favour a stepwise mechanism involving the intermediate for- mation of BCl, because in such a way, one can envision the formation of the intermediates that are necessary for pro- moting the secondary reactions that take place at higher pressures. Nevertheless, we clearly stress the tentative charac- ter of the proposed steps in the above scheme.The above scheme accounts sufficiently for the ionic pattern recorded in the mass spectrometer under collisionless conditions. It is not, however, evident which intermediate(s) is (are) responsible for the generation of the fluorescing species detected in Fig. 2(b). In this respect, it is important to note the very similar dependence on pressure that is exhibited by the emission of the observed species. Given the overall simi- larity of the curves in Fig. 2(b), it appears that in fact, all three species (excited Cl,, BCl and B) derive from a common precursor, plausibly vibrationally 'hot' BCl, or BCl. In the analysis of the observed reactivity, we believe that a most significant mechanistic clue is provided by the observa- tion of C1, (2,H,).Molecular chlorine is a common product of BCl, reactivity and in almost all cases, it is formed through halogen abstraction steps. In fact condensation and halogen-abstraction reactions dominate boron chemistry.22 These reactions are highly exothermic and can, in principle, ~~ ~~ ~~~ ~~ t One of the referees has brought to our attention ref. 12, where the character of the excited states of BCl, is discussed in great detail. The discussion therein confirms our assignment. $ Energetically, the two 248 nm photon excitation can result in the formation of BC1, in the A and fi states and indeed fluorescence from these states was detected upon synchrotron irradiation (albeit with a relatively small cross-section) by Lee and co-workers" and very recently by Creasey et al." In the present work, at low pressures, weak signals were indeed detected that could be indicative of the formation of these states.However, at the higher pressures employed for most experiments, these signals remained generally weak and usually difficult to identify unambiguously. Probably once formed, these states absorb additional photon(s) to ionize and/or dissociate or alternatively they participate in the observed reactivity sequence. provide the necessary energy for producing the electronically excited species observed in Fig. l(b). Several plausible reac- tions could be written yielding the observed products, but in view of the limited literature data, this approach is too ambiguous.At any rate, though its mechanism cannot be specified at present, the observed reactivity process has important conse- quences for the laser vapour processing of BCl,. In particu- lar, increasing the pressure of the precursor may lead to higher deposition rates, since this results in enhanced pro- duction of unsaturated species (B and BC1) through removal of ligands in the form of the stable diatomic Cl,. This situ- ation would contrast sharply with the laser vapour pro- cessing of metal hexa~arbonyls~~ where, at high pressures, recombination reactions between the photolysis products [i.e. M(CO),, M = metal, and CO] become dominant with detri- mental effects on the deposition process (lower deposition rates, and higher carbon impurity concentrations in the film). Laser vapour processing of BCl, at 248 nm has been examinedI6 over a rather limited pressure range of the pre- cursor and in the presence of an excess of buffer gas.The deposition resulted in the formation of scattered 'islands', which were rich in C1 impurities. It appears that the use of excess buffer gas and limited precursor pressures restricted further film growth, exactly as suggested above. Further studies of the pressure dependence of the laser vapour and doping of BCl, at 248 nm are warranted and appear most promising for the mechanistic information that they can provide. Conclusions The UV laser photolysis of BCl, at 193 and 248 nm was studied by fluorescence and mass spectroscopic techniques.At 193 nm, we confirm the previous report that BCl, undergoes one- and two-photon processes. Strong evidence has been provided for the nature of the products of the two processes. Furthermore, the pressure dependence of the observed species was studied and correlated with laser-doping results. At 248 nm, we propose that BCl, undergoes a two-photon excitation which, under collisionless conditions, is followed by a stepwise excitation/fragmentation to BCl'. At higher pressures, secondary reactions dominate, resulting in the for- mation of C1, (Z311g), BCl (A'II) and excited B* species. Implications of these observations for laser vapour pro- cessing were discussed. Overall the photodissociation of BCl, appears to be complex and may hide a number of interesting phenomena.Experiments for the study of the BCl, photophysics in the presence of surfaces are currently in progress at our labor- atory. Support by the Human Capital and Mobility Programme (Contract No. ERBCHRXCT930157) is gratefully acknow- ledged. References R. H. Burton, R. A. Gottscho and G. Smolinsky, Dry Etching for Microelectronics, ed. R. A. Powell, Elsevier, New York, 1984, and references therein. K. J. Schmatjko, G. Endres and H. Durchhole, LIM-5 Proc., 1988, 145. H. Guo and Q. Lou, Appl. Surf: Sci., 1993,68,2090. J. Elders, D. Bebelaar and J. D. W. van Voorst, Appl. Su$ Sci., 1990,46,215. F. Foulon, A. Slaoui, E. Fogarassay, R. Stuck, C.Fuchs and P. Siffert, Appl. Surf Sci., 1989, 36,384. 3638 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 6 T. F. Deutch, D. J. Ehrlich, D. D. Rathman, D. J. Silversmith 17 C. E. Moore, Atomic Energy Levels, Natl. Bur. Stand. NSRDS- 7 8 9 and R. M. Osgood, Appl. Phys. Lett., 1981,39, 825. V. H. Dibeler and J. A. Walker, Inorg. Chem., 1969,8,50. J. A. Stockdale, D. R. Nelson, F. J. Davis and R.N. Compton, J. Chem. Phys., 1972,56,3336. P. J. Bassett and D. R. Lloyd, J. Chem. SOC. A, 1971, 1551. 18 19 NBS 35,1971. K. P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure 1V. Constants of Diatomic Molecules, Van Nostrand Reinhold, New York, 1979. W. K. Bischell, L. E. Jusinki, M. N. Spencer, D. J. Eckstrom, J. 10 M. Suto, C. Ye, J. C. Han and L. C. Lee, J. Chem. Phys., 1988, 89,6653. 20 Opt. SOC. Amer. B, 1985,2, 877. R. A. Roig and G. Tondello, J. Phys. B,1976,9,2373. 11 L. C. Lee, J. C. Han and M. Suto, J. Chem. Phys., 1989,91,2036. 21 T. Shinzawa, A. Tokunaga, T. Ishiwata and I. Tanaka, J. Chem. 12 13 J. C. Creasey, P. A. Hatherly, I. R. Lambert and R. P. Tuckett, Mol. Phys., 1993,19, 413. H. J. Maria, J. R.McDonald and S. P. McGlynn, J. Am. Chem. 22 Phys., 1985,83, 5407. A. F. Cotton and G. Wilkinson, Advanced lnorganic Chemistry, A Comprehensive Text, John Wiley, New York, 1979, 4th edn., SOC.,1973,95, 1050. 299-303. 14 A. A. Planckaert, P. Sauvageau and C. Sandorfy, Chem. Phys. 23 N. Okada, Y. Katsumura and K. Ishigure, Appl. Phys. A, 1993, Lett., 1973, 20, 170. 56, 138. 15 A. Slaoui, F. Foulon, C. Fuchs, E. Fogarassy and P. Siffert, Appl. Phys. A, 1990,50, 317. 16 M. Georgopoulos, G. S. Fu, E. Hontzopoulos and C. Fotakis, NATO AS1 Ser. B, 1989,69,198. Paper 4/03967J; Received 30th June, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003633

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

J. CHEM. SOC. FARADAY TRANS., 1994, 90(24), 3639-3643 Kinetics of Disproportionation of Hypoiodous Acid Victor W. Truesdale" 36Ladycroft Park, Ble wbury, Oxon, UK OX1 1 9QW Carlos Canosa-Mas Physical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, UK OX13QZ George W. Luther 111 College of Marine Studies, University of Dela ware, Pilotto wn Road, Le wes , DE 19958,USA The disproportionation of hypoiodous acid into iodide and iodate has been studied at pHs between 7 and 13 and without added iodide. The second-order behaviour of the reaction velocity with respect to total molecular iodine (1I,), well established for highly alkaline conditions, has been found to hold at these lower pHs provided the solutions are not brown with iodine. It is shown that the kinetics can be modelled as a second-order reaction in [HOl] influenced by two opposing pre-equilibria involving I, and HOI, and HOI and lo-, respectively.The equi- librium between I, and I,-is shown to be of low significance when the background iodide concentration is less than about mol I-'. The model predicts that in highly alkaline solutions additional reactions involving I,OH-are necessary to comply with the first-order behaviour with respect to iodide reported in historical rate data. The reliance of the model on iodine speciation also obviates the need for a reciprocal dependence of the reaction rate upon OH-concentration in alkaline solutions. The sluggishness of the reaction at pHs less than about 8 (brown solutions) is consistent with either the rate-limiting step changing to one involving lo,-, or the retarding effect of iodide generated during the reaction.The disproportionation of iodine (hypoiodous acid) into iodate and iodide has been of interest in en~ironmentall-~ as well as inorganic and physical4 chemistry. It seemed possible that molecular iodine generated naturally at the sea surface by the photochemical oxidation of iodide3 could produce iodate at a seawater pH of 8.2. As a result, dispro-portionation offered a possible non-biological mechanism for re-cycling iodide to iodate in the oceans. Moreover, since molecular iodine is volatile, means have been sought to retain its radioisotopes in solution during rupture of nuclear reactor vessels.One way to achieve this' is to make the reactor solu- tions alkaline so that the disproportionation reaction pro- duces iodate. Despite these diverse interests and almost a century's study6 the kinetics of the reaction are still not known clearly. Classically, the reaction has been studied in highly alkaline solutions using iodine solutions containing iodide to prevent loss of molecular iodine by v~latilisation.~~~*~ More recently, however, Thomas et owing to their interest in nuclear reactor environments, studied the reaction at pHs between 6 and 10, without iodide. Their work provides a very significant change in emphasis by demonstrating that the high-pH studies are not necessarily the most representative. We were particularly interested in the peak in disproportionation rate which Thomas et al.' described at about pH 9, as we already knew from speciation studies of molecular iodine in sea water that the concentrations of HOI and 1,OH -also peak in that region.',* It seemed, therefore, that there were good reasons for believing that the universal kinetic mechanism might well require the adoption of HOI as the main reactive species.In this paper we report measurements of the rate of dispro- portionation at several pHs and present a model in which the system is depicted as two conflicting equilibria coupled to the second-order disproportionation of hypoiodous acid. This model has been developed as a prelude to working in greater detail with a fuller set of elementary reactions in a more com- plete kinetic model.Nevertheless, this current model, although cruder than the one we hope to develop, has the distinct advantage of allowing the changes in chemistry across the entire pH range to be visualised more clearly than was hitherto possible. Experimental A spectrophotometric versiong~'' of the classical iodine titra- tion has been used throughout. The apparent iodine content (1I, = I' + 1') was determined by placing 0.400mi of each reaction mixture into 40.0 ml of 0.15 mol 1-' sodium dihy- drogen phosphate buffer at pH 6.7, with an excess of iodide (3.0 mi of 0.60 mol 1-l potassium iodide) and measuring its absorbance at 285 nm. A Pye-Unicam SP-500spectropho-tometer was used equipped with 10 cm quartz cuvettes.AnalaR reagents were employed throughout. Dispro-portionation was carried out in 50 mi, glass-stoppered mea- suring cylinders equipped with Pyrex stoppers. Mixtures of 0.15 mol 1-' sodium dihydrogen phosphate and 0.15 mol 1-' sodium hydroxide, after Thomas et a/.,' were made up to 262 pmol 1-1 by addition of a saturated solution of molecular iodine in distilled water. Experiments were conducted under ordinary laboratory fluorescent lighting ; apart from the routine shielding of the potassium iodide solution from any light no other precautions were deemed necessary. The kinetic modelling was performed using a spreadsheet. Results Reaction at pHs between 7 and 12 In alkali solution the reaction kinetics, in terms of 1I,, have been shown to fit the classical reciprocal plot used for iden- tifying a second-order reacti~n.~*~,~'*~~ The measurement of apparent iodine concentration encompasses both I' and I', but not I", and so the observation can be expressed as : where u is the reaction velocity.Eqn. (1) is not strictly a mechanistic statement since it considers only the total 'molecular iodine'. We find that this behaviour extends down to pH 8.90. Good linear plots were obtained (Table 1) with correlation coefficients close to 1.0 for over 95% of the reac- tion span. At the highest pH the reaction was only followed to 73% completion, but otherwise all points were included in 3640 Table 1 Fit of experimental data to reciprocal concentration us.time plots; experiments conducted at 25 "C no. of span of gradient" correlation pH points reaction (%) /Imol-' min-' coefficient 12.06 30 100 23.2 rf: 0.2 0.9984 11.10 21 95 596 k 9 0.9962 10.57 25 95 1930 f40 0.9943 10.00 25 98 6320 k80 0.9981 8.90 25 98 6060 f 180 0.9898 7.99 27 81' 1450 rf: 30 0.9960 C -C--C7.18 28 " Standard errors are given. * The reaction came to steady state; the plot offers only a crude estimate of the rate constant. Second-order plot not appropriate. the plot. At pHs of 7.99 and 7.18 the reaction approached a steady-state condition corresponding to 87 and 54% com- pletion, respectively. Thomas et a1.' first reported this effect, and in separate experiments we have confirmed their finding that it allows the solutions at pHs between 6 and 8 to retain their brown colour, sometimes for many days or even weeks at 25 "C.As expected, therefore, the reciprocal concentration us.time plot for reaction at pH 7.18 was not linear. However, an estimate of the rate was obtained for pH 7.99 using the steady-state position as infinite absorbance, from which a good linear plot was obtained. We also found, as reported elsewhere,' a good linear plot for artificial sea water,13 dosed with one tenth of the amount of molecular iodine (27 pmol 1-') at pH 8.2 and a salinity of 35%. The reaction was studied for over 36 h and regression yielded a correlation coefficient of 0.9991 for 40 points covering 94% of the observed reaction. A plot showing the variation of the experi- mentally determined second-order rate constant with pH in the phosphate buffer mixtures is presented in Fig. 1.Effect of Iodide on Reaction Rate The results presented in Table 2 show that, whereas the addi- tion of small amounts of iodide (comparable to the concen- trations of iodide generated during the course of the reaction) accelerates the reaction slightly at pH 11.3, at pH 8.1 such an addition slows the reaction markedly. These experiments were conducted by adding up to 0.8 ml of either 0.88 or 0.088 mol I-sodium iodide solution (and any compensating 10 I -1I I i-brown i colourless i P:I. -I !/ \ I'I \ a2 i c)20' 9I I I 1 J6 7 8 10 11 12 13 PH Fig. 1 Variation with pH of the predicted and observed rate con- stant for disproportionation of hypoiodous acid.The predicted curve assumes an iodide concentration of mol 1-' and k of 9 x lo3 mol-' min-'. Whereas to the right of the vertical line the reaction goes to completion to give a colourless solution, to the left, the reac- tion mixtures remain brown with iodine for ccjnsiderable periods. The lower broken line depicts qualitatively the diminution of the overall rate by the steady-state condition. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Effect of added iodide upon the amount of residual I, in the reaction mixture after 21 min of reaction at 25 "C" absorbanceb concentration of added iodide/mmol 1-' pH 8.12 rf: 0.03 pH 11.33 k0.02 ~~ 0.00 0.348 0.530 0.44 0.618 0.513 0.88 0.795 0.499 1.76 0.85 1 0.48 1 2.64 0.878 0.479 3.52 0.900 0.479 4.40 0.904 - 8.80 0.922 - 17.60 0.925 - " Initial and final absorbances are 0.925 and 0.00, respectively.* At 825 nm, 10 cm pathlengths. amount of pure water to maintain constant volume) to 20.0 ml of the appropriate phosphate buffer solution, before addi- tion of 5 ml of saturated iodine. Modelling Approach From Fig. 1, and earlier studies of the reaction state at set times of 1 and 40 min,' the maximum rate appears at a pH of ca. 9.5. As this is also the pH at which the HOI species pre- dominates in the system, our initial assumption was that HOI has a central role in the mechanism.' It suggested that the reaction might be determined by a simple second-order reaction in [HOI], u=--dCC I21 --k[HOIl2dt and the equilibria: I, + H20sHOT + Hf + HOIeH' + IO-I, + I-e13-Of course, the approach demands that the equilibria control- ling the speciation of iodine are established much faster than the disproportionation reaction.The hypoiodite equilibrium is a proton-base reaction and will therefore be very fast, and as Eigen and Kustin14 showed, the I,-HOI equilibrium takes a fraction of a second to adjust. In contrast, the dispro- portionation reaction is much slower, taking at least tens of minutes to reach completion at 25 "C. Rearranging eqn. (3)-(5) for the equilibria to provide [I2], [IO-] and [I3-] respectively, and substituting these into the mass-continuity equation, CI2l + c13-I + WOIl + c01-1 = cc I21 (6) gives (7) which can also be written as, J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 where a represents the terms within the main bracket of eqn. (7) and is constant for a specific combination of iodide and hydrogen ion concentrations. Substitution of this into eqn. (2) then provides an equation equivalent to eqn. (l),but incorporating the mechanism, dCC I21 -k/{l+ K3CH 'I [I -3' -= dt Kl This explains why the simple test for a second-order reaction, i.e. plotting reciprocal concentration us. time, has worked for all workers since Schwicker;6 [II,] substitutes for [HOI] at constant iodide concentration and pH. Comparing the coeffi- cients of eqn.(1) and (9) leads to, where lob, is the predicted value of kobs. A plot of lobsis presented in Fig. 1, where K, and K, were taken as 5.44 x l5 and 2.3 x lo-" mol 1-', respec-tively. Fitting of the predicted values to the experimental ones was accomplished by trial and error alterations in the parameter k and the iodide concentration; in essence, whereas k allows the predicted curve to be expanded verti- cally by a constant factor, [I-] changes its shape. A reason- able fit of Fig. 1 was obtained with a k of 9 x lo3 1 mol-' min-' and an iodide concentration of low4rnol I-'. Discussion The above results provide a rate equation and mechanism for the disproportionation reaction which are numerically con- sistent across a much larger part of the available pH range than has hitherto been possible.Thus, all earlier studies which provide a mechanism'2*'6 have been conducted at single or restricted pH settings. Also, it is important to appre- ciate that this study is unusual in the degree to which the link between the concentrations of individual species and the measured variable, C I,, has been exploited. Some earlier statements5 of the rate equation made in terms of concentra- tions of individual species were really merely speculations, since this link was not established. Thomas et d5provided something of an exception to this in that they recognised the need for equations like eqn. (lo), but were unfortunately forced to restrict their application because of difficulties in measuring the rate constants due to divergences from simple second-order kinetics.As is demonstrated below, by mathe- matically predicting the effects of variation in iodide and hydrogen ion concentration this study provides, for the first time, a semi-quantitative appreciation of the entire system. Effect of Iodide Concentration on Reaction Rate The model predicts that as the iodide concentration is increased the peak in the plot of rate constant vs. pH sharp- ens and loses height (Fig. 2). The sharpening occurs because the right-hand side of the peak, being determined by the hypoiodite equilibrium which is independent of the iodide concentration, is fixed. Fig. 2 also shows that the 1,-13-equi-librium has an insignificant effect upon the predicted rate constant with less than 0.0001 mol 1-iodide in the reaction mixture.In contrast, with iodide concentrations of 0.001 mol 1-and above this equilibrium increasingly reduces the rate as production of I,-becomes more competitive. Note, however, that the attenuation of the peak with increasing iodide concentrations (Fig. 2) does not mean that under some 3641 10 I 7 c I .-C E PH Fig. 2 Effect of ten-fold increases in iodide concentration on the disproportionation rate. The continuous curves represent the system without the I,--I,-equilibrium; the left curve represents an iodide concentration of 10- '. The broken curves show how inclusion of the additional equilibrium affects the predicted peak. For the two left- hand curves, broken and continuous curves superimpose.conditions it will not be seen. Any experimentalist approach- ing the system using constant iodide concentrations will find a peak because, by choosing a temperature which provides a measurable reaction rate at some pH, they will in effect adjust the scale of Fig. 2 to their own particular purposes. The model predicts that the addition of iodide to reaction mixtures at constant pH below that of the observed peak-rate will reduce the rate of reaction by shifting equilibrium (I) to the left, as indeed is consistent with the results of Table 2. Fig. 3 shows the variation in predicted rate obtained by the generation of up to 4.35 x mol 1-' iodide at pHs of 7-11 in solutions to which no background iodide has been added.This is the maximum amount of iodide generated by the complete disproportionation of the 262 mol 1-' of I, used. Of course, the actual range of iodide concentrations that needs to be considered varies with pH, as the lower the pH of the solution the less is the initial hydrolysis of I, to HOI and hence, the lower the effective initial iodide concen- tration for disproportionation. Although we believe that the fit of the model and experi- mental points shown in Fig. 1 can be taken as reasonably representative of the overall behaviour of the system, we also recognise that the results in Fig. 3 mean that it cannot be interpreted too precisely. The fact that the iodide concentra- tions required for the fit in Fig.1 are close to those existing in solution, and that larger or smaller ones do not give anything like as good a fit (Fig. 2), is encouraging. However, the pre- dicted decrease in the rate constant with reaction progression at pH 8 would be expected to disrupt seriously the linearity r 10 I II .-C E -mI 0 I U," '0 100 200 300 400 J L iodide concentration/pmol I-' Fig. 3 Variation in the predicted rate constant, lobs,with increase in iodide concentration at unit pHs between 7 and 10. In each case the continuous line depicts the conditions of the reaction mixture between the end of I, hydrolysis and completion of the dispro- portionation reaction, i.e. 435 pmol 1-of iodide. of the graphs of reciprocal I, concentration us.time since integration of eqn. (1) assumes its constancy. Fortunately, this effect applies only at the two lowest of our pH settings and therefore it need not deter us from our aim of gaining a broad understanding of this system’s kinetics. We intend to study the effect in detail later using a more complete set of elementary reaction steps, particularly to establish whether, as the results in Table 2 suggest, slowing by in situ generated iodide could be responsible for the extremely slow, steady- state condition that exists at pHs lower than 8.’ At pHs above and some distance away from the peak rate, e.g. a pH of 12, the model predicts that iodide should have no effect whatsoever. The influence of the I,-HOI equilibrium in the continuity equation is negligible at that point [eqn.(7)] and only the HOI-10- equilibrium, which has no iodide dependence, influences the rate constant. This prediction con- tradicts many earlier observation^^^'^^' which show that addition of iodide increases the rate in highly alkaline solu- tions. Indeed, some previous workers have included iodide in the rate equation, the apparent rate constant of which has the been given the mathematical kobs = k’ + k”[I-]/[OH-] (11) where k’ and k” are constants. By including the relevant terms for [I,OH-] in the above model, and substituting one of the [HOI] terms of eqn. (2) by [I,OH-I, it has been found that the equation analogues to eqn. (lo), predicts first-order behaviour with respect to iodide.(Substitution with 10-has no similar effect.) This was accomplished by introducing the equilibrium, I, + H,O=I,OH-+ H+ (IV) whose equilibrium constant, K, ,from Palmer and Lietzke’ is calculated to be 1.41 x lo-’’. This study therefore verifies the earlier conclusion of Wren et that the 1,OH- species is involved in the reaction in highly alkaline media. More- over, it shows that there need be no conflict between the observation of first-order behaviour with respect to iodide in alkaline media7*’ and the different behaviour brought about by the I,-HOI equilibrium [eqn. (I) and (3)] at the lower pHs. This is because there are two separate reactions occurring at highly alkaline pHs [eqn. (1l)], and the residual reaction due to k’ of eqn.(11) is the reaction we have studied at lower pHs [eqn. (l)], where the second reaction, i.e. k”, is insignificant. The existence of the two reactions also explains the slight discrepancy between the behaviour toward iodide of the experimental and model-predicted rate constants. Thus, whereas the experimental results show a slight acceleration by iodide at pH 11.33 (Table 2) the model predicts a slight retardation (Fig. 3) at similar pH. We suspect that the changeover in significance of the two reactions occurs in the region of pH 11, and that the arbitrary choice of pH 11.33 for our investigation biased the results slightly toward the ‘alkaline’ reaction. Effect of pH Variation on Reaction Rate The model predicts a decrease in disproportionation rate with increase of pH above the peak rate.The lack of know- ledge of previous workers about the pH-wide speciation changes described here led them to the incorrect belief that the rate constant varies inversely with [OH-] [eqn. (ll)]. In any case, re-working of the data presented by Forster7 and Li and White16 reveals that the inverse behaviour of eqn. (11) offers only a crude approximation to their experimental results; there is too much curvature in any plot of rate con- stant us. reciprocal of hydroxide concentration to justify its use. Evidently, Li and White’s16 plot of rate constant us. the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 quotient of iodide and hydroxide concentrations [eqn. (1 l)] was influenced more by the iodide variation than by that of the hydroxide concentration.Although eqn. (10) cannot reproduce Forster’s7 results on [OH -3 variation, analogous equations obtained by substitut- ing [OI-] in place of one or both of the [HOI] terms of eqn. (2) do (Table 3). Therefore, there is some justification for adding a second term to our rate equation [eqn. (2)], e.g. dCCdt‘,I --kl[HOI]2 -k,[HO1][01-] (12) as indeed Li and White16 did for quite different reasons. [Despite the increased complexity of eqn. (12), as is shown below, it can still be factorised to the form of eqn. (1) through the use of equilibrium (11) between HOI and 10-1. While such expansions can provide the required [OH-] variation, like eqn. (lo), they do not elicit the first-order behaviour with respect to iodide at high pHs.However, as above, substitut- ion of C1,OH-I for [HOI], this time in the second term of eqn. (12), will achieve this. In consequence, our current view is that the full rate equation is best expressed as, dCCdt12’ --k1[HOIl2 -k,[I,OH-][OI-] (13) Interchangeability of Iodine Species within the Model By substituting one iodine species for another, as above, it is useful to appreciate that, for a given iodide concentration, each combination has a characteristic position on the pH scale (Fig. 4), and that all species are in direct proportion to each other for a given combination of iodide and hydrogen Table 3 Fitting of the expanded models to the data of Forster’ for variation of the rate constant with [OH-] at high pHs rate constant predicted by expanded model (25 “C) I moi-’ min-’ Forster’s rate constant calculated [HOI][OI-] [OI-][Ol-] /lop31 mol-’ min-l pH added added 125 13.70 122 123 24 1 13.41 242 193 485 13.10 503 477 C‘E 10 HOI + HOI I2OH -+ I2OH -I20H 7HOI Ip+ 10-m HOI+OI--“6 7 8 9 10 11 12 13 14 PH Fig.4 Rate constant predicted by the model for various com-binations of iodine species at a constant iodide concentration of lop4 mol 1-’. The species contributing to each peak are given on the figure. Taken from left to right the k values are (lo31 mol-’ min-’): 18; 9, 13.3, 0.345, 2.42; 0.20, 0.76; 0.76, respectively, for the com- binations of species given on the figure from top to bottom.J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 ion concentrations. Thus, the continuity equation [eqn. (6)], as a linear combination of the relevant iodine concentrations, gives, PCI21 = cc I21 4wH-I = cC I21 (17) where, for a set of combination of pH and iodide concentra- tion, B, y, 6 and E are constants analogous to a of eqn. (8). Moreover, under these same circumstances, Kl&=-KJ-1 a Thus, for example, the above substitution of [I,OH-] for [HOI] in the development of eqn. (13) requires the use of eqn. (17)rather than eqn. (8). However, as the relationship between E and a [eqn. (21)] does not involve [H’], the position of its peak is unaffected (Fig. 4). Conversely the [I-] term in eqn. (21), which is not present in eqn.(9,introduces the required first-order behaviour with respect to iodide. Moreover, in fitting predicted curves to actual ones, as in Fig. 1, differences produced by the K,, K, and [I-] terms of eqn. (21) are com- pensated for precisely by changes in the apparent rate con- stant, in this case, k, of eqn. (13). In this way some combinations can provide exactly the same pH distribution as do some others (Fig. 4). Whichever of the individual iodine species is used in the rate equation, the latter can always be expressed in the same mathematical form as eqn. (l), i.e. retaining the overall second-order behaviour. Thus, for eqn. (13), application of eqn. (8), (15) and (17)gives, Further Comment on the Slow Reaction below pH 8 The aptness of Forster’s’ classification of disproportionation studies into those concerning brown and those with colour- less solutions, established so long ago, is fully appreciated here (Fig. 1).The satisfactory treatment of the kinetics at pHs lower than 8.0 requires that the steady-state condition, which allows for longer lifetimes of I,, be addressed.The model, with and without the 12-13- equilibrium (Fig. 2), shows that it is not the 12-13-equilibrium which causes the increased stability of added molecular iodine, as might well have been thought from the prevalent belief that iodide stabilises I, solutions ;without the addition of further iodide, the concen- tration of the latter is too low to produce significant concen- trations of I,-.As iodate remains the truly thermodynamically predomi- nant species in the system until superseded by I, below about pH 5.5,8 it is possible, as suggested above, that it is the increasing presence of internally generated iodide that gives I, its longer lifetime at pHs below 8. Equally, however, this effect could be due to the increasing influence of an equi- librium between HOI and the next highest oxidised form, probably IO,-. Kinetically, this means that at lower pHs there could be an increasing influence of the back reaction, 10,- +2H+ +I- -+ HOI +HOI (V) with a shift to another rate-limiting reaction, 10,-+HOI +103- +H+ +I- (VI) In principle, rate equations for this new system could be derived in the manner used above.However, in practice the presence of the two HOI terms in eqn. (6) makes the sub- sequent algebra complex. Once again then, we intend to pursue this problem instead by the kinetic approach using individual elementary steps. Nevertheless, qualitative addi- tion of the back reaction (or indeed the slowing by the inter- nally generated iodide) is equivalent to accelerating the decrease of the overall rate as the pH decreases, as depicted by the broken line in Fig. 1. Special thanks go to Peter Foster and Sylvia East of the Department of Ocean Sciences, University of North Wales, for providing V.W.T. with bench space and support by which to do the practical work of this study. G.L.’s contribution was supported by a grant OCE-9217245 from U.S. National Science Foundation. References 1 V.W.Truesdale, G.W. Luther I11 and C. Canosa-Mas, in prep- aration. K. Sugawara and K. Terada, Nature (London), 1958,182,250. Y. Miyake and S. Tsunogai, J. Geophys. Res., 1963,68,3989. K. J. Morgan, Quart. Rev., 1954, 8. 123. T. R. Thomas, D. T. Pence and R. A. Hasty, J. Inorg. Nucl. Chem., 1980,42, 183. 6 A. Schwicker, Z. Phys. Chem., 1985,16,303. 7 E. L. C. Forster, J. Phys. Chem., 1903,7, 640. 8 D. A. Palmer and M. H. Lietzke, Radiochim. Acta, 1982,31, 37. 9 V.W.Truesdale and C. P. Spencer, Marine Chem., 1974,2,33. 10 V.W. Truesdale, Marine Chem., 1978,6, 253. 11 A. Skrabal, Z. Electrochem., 1934,40,232. 12 J. C. Wren, J. Paquette, S. Sunder and B. L. Ford, Can. J. Chem., 1986,64,2284. 13 J. Lyman and R. H. Fleming, J. Marine Res., 1940,3, 134. 14 M. Eigen and K. Kustin, J. Am. Chem. SOC.,1962,84, 1355. 15 J. D. Burger and H. A. Liebhafsky, Anal. Chem., 1973,600. 16 C. H.Li and C. F. White, J. Am. Chem. Soc., 1943,65,335. Paper 4/04198D; Received 1 lth July, 1994

ISSN:0956-5000    

DOI:10.1039/FT9949003639

出版商:RSC    

年代:1994

数据来源: RSC