摘要:

THE THIRD ANNUAL GENERAL MEETING OF THE FARADAY DIVISION of the Chemical Society was held at 9.00 a.m., on 11 th September 1974, in the Physical Chemistry Lecture Theatre, The University of Cambridge, with Professor T. M. Sugden, M.A., Sc.D., F.R.S., in the Chair. 1 Minutes The Minutes of the Second Annual General Meeting of the Faraday Division, which had been circulated previously, were taken as read and confirmed. 2 Election of Council The Members of Council of the Faraday Division of The Chemical Society elected to take office from 9 April 1975 were as follows: President Prof. T. M. Sugden, M.A., Sc.D., F.R.S. Vice-presidents who have held the ofice of President SIR FREDERICK DAINTON, MA., ScD., F.R.S. PROF. C . E. H. BAWN, C.B.E., PH.D., F.R.S. PROF. G. GEE, C.B.E., M.Sc., Sc.D., F.R.S.PROF. J. W. LINNETT, M.A., D.PHIL., F.R.S. PROF. SIR GEORGE PORTER M.A., Sc.D., F.R.I.C., F.R.S. Vice-presidents PROF. D. H. EVERETT, M.B.E., M.A., D.Sc. PROF. P. GRAY, M.A., Sc.D. PROF. J. N. MURRELL, PH.D. PROF. W. C. PRICE, Sc.D., F.INsT.P., F.R.S. PROF. J. S . ROWLINSON, M.A., D.PHIL., F.R.I.C., DR. H. A. SKINNER, B.A., B.Sc., D.PHIL. PROF. F. C . TOMPKINS, DSc., F.R.I.C., F.R.S. F.R.S. Ordinary Members of Council PROF. A. D. BUCKINGHAM, M.A., PH.D., PROF. M. MAGAT, D.Sc., D.PHIL. PROF. I. M. MILLS, D.PHIL. PROF. MANSEL DAVIES, Sc.D. PROF. A. M. NORTH, D.Sc., F.R.S.E., F.R.I.C. DR. D. N. HAGUE, M.A., PHD. DR. R. PARSONS, D.Sc., F.R.I.C. PROF. N. B. H. JONATHAN, PH.D. DR. B. A. PETHICA, D.Sc. PROF. DR, 5. LYKLEMA F.R.I.C., F.R.A.C.I. Honorary Treasurer PROF.P. GRAY, M.A., Sc.D. Honorary Secretary PROF. F. C . TOMPKMS, D.Sc., F.R.I.C., F.R.S. The President thanked Professor Allen and Professor Caldin, the retiring members of Council, for their services. 3 Annual Report During 1973 the Faraday Division was very active both in continuing the traditional Faraday Discussions and Symposia and also representing physical chemistry and chemical physics interests in the activities of The Chemical Society as a whole. Two General Discussions were held, " Molecular Beam Scattering " at University College, London and " Intermediates in Electro- chemical Reactions " at Oxford; the proceedings being published as Faraday Discussions of The Chemical Society Nos. 55 and 56. Two Symposia were also held during the year; the first, " Fogs and Smokes ", which formed part of the Annual Congress of The Chemical Society at Swansea, and the second, " High Temperature Studies in Chemistry " at the Royal Institution, London : both will appear in print as Symposia Nos. 7 and 8 respectively.In addition, two informal discussions took place, one on " Physical Methods of Studying Surface Adsorbed Molecules " arranged in 28THE THIRD ANNUAL GENERAL MEETING OF THE FARADAY DIVISION of the Chemical Society was held at 9.00 a.m., on 11 th September 1974, in the Physical Chemistry Lecture Theatre, The University of Cambridge, with Professor T. M. Sugden, M.A., Sc.D., F.R.S., in the Chair. 1 Minutes The Minutes of the Second Annual General Meeting of the Faraday Division, which had been circulated previously, were taken as read and confirmed.2 Election of Council The Members of Council of the Faraday Division of The Chemical Society elected to take office from 9 April 1975 were as follows: President Prof. T. M. Sugden, M.A., Sc.D., F.R.S. Vice-presidents who have held the ofice of President SIR FREDERICK DAINTON, MA., ScD., F.R.S. PROF. C . E. H. BAWN, C.B.E., PH.D., F.R.S. PROF. G. GEE, C.B.E., M.Sc., Sc.D., F.R.S. PROF. J. W. LINNETT, M.A., D.PHIL., F.R.S. PROF. SIR GEORGE PORTER M.A., Sc.D., F.R.I.C., F.R.S. Vice-presidents PROF. D. H. EVERETT, M.B.E., M.A., D.Sc. PROF. P. GRAY, M.A., Sc.D. PROF. J. N. MURRELL, PH.D. PROF. W. C. PRICE, Sc.D., F.INsT.P., F.R.S. PROF. J. S . ROWLINSON, M.A., D.PHIL., F.R.I.C., DR. H. A. SKINNER, B.A., B.Sc., D.PHIL.PROF. F. C . TOMPKINS, DSc., F.R.I.C., F.R.S. F.R.S. Ordinary Members of Council PROF. A. D. BUCKINGHAM, M.A., PH.D., PROF. M. MAGAT, D.Sc., D.PHIL. PROF. I. M. MILLS, D.PHIL. PROF. MANSEL DAVIES, Sc.D. PROF. A. M. NORTH, D.Sc., F.R.S.E., F.R.I.C. DR. D. N. HAGUE, M.A., PHD. DR. R. PARSONS, D.Sc., F.R.I.C. PROF. N. B. H. JONATHAN, PH.D. DR. B. A. PETHICA, D.Sc. PROF. DR, 5. LYKLEMA F.R.I.C., F.R.A.C.I. Honorary Treasurer PROF. P. GRAY, M.A., Sc.D. Honorary Secretary PROF. F. C . TOMPKMS, D.Sc., F.R.I.C., F.R.S. The President thanked Professor Allen and Professor Caldin, the retiring members of Council, for their services. 3 Annual Report During 1973 the Faraday Division was very active both in continuing the traditional Faraday Discussions and Symposia and also representing physical chemistry and chemical physics interests in the activities of The Chemical Society as a whole. Two General Discussions were held, " Molecular Beam Scattering " at University College, London and " Intermediates in Electro- chemical Reactions " at Oxford; the proceedings being published as Faraday Discussions of The Chemical Society Nos. 55 and 56. Two Symposia were also held during the year; the first, " Fogs and Smokes ", which formed part of the Annual Congress of The Chemical Society at Swansea, and the second, " High Temperature Studies in Chemistry " at the Royal Institution, London : both will appear in print as Symposia Nos. 7 and 8 respectively. In addition, two informal discussions took place, one on " Physical Methods of Studying Surface Adsorbed Molecules " arranged in 28

ISSN:0300-9599    

DOI:10.1039/F197470FX001

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Kinetics of Electrode Reactions in Liquid Ammonia Part 2.-Fe1"/FeI1 and C O ~ ~ ' / C O ~ ~ Redox Couples BY OLIVER R. BROWN * AND SEAN A. THORNTON Electrochemistry Research Laboratories, Department of Physical Chemistry, University of Newcastle upon Tyne NEl 7RU Received 19th June, 1973 Polarisation data has been obtained for the CoII/CoIII and FeII/FeIII couples at rotating disc electrodes of platinum, gold and vitreous carbon at - 30°C in ammonia acid solution. In addition, metal wires and carbon rod have been used to examine the low level behaviour of the cobalt system and to record charge measurements in the oxidation of the anode surfaces. The iron couple is reversible on carbon but is inhibited by anodic films on gold and platinum. Apparent standard rate constants for the CoII/CoIII reaction are affected only slightly by the electrode material, but the oxidation branch is noticeably inhibited by anodic film formation. The mechanism of inhibition by anodic films is discussed.FeIIl is unstable in acid ammonia. Electrode kinetic studies of redox reactions have not been made previously in liquid ammonia although polarographic waves for metal ion reductions have been analysed. Similar studies have been carried out in ammoniates at higher tempera- t u r e ~ . ~ ' ~ The reduction of Co"' to ColI in such media was shown to be polaro- graphically irreversible.2* 5 * The chemistry of Fe"' ammines does not appear to have been reported previously. In Part 1 we described the anodic oxidation of a platinum surface in acid am- monia solutions and the effect on the kinetics of hydrogen oxidation.Now we describe the effect of surface oxidation on two redox reactions and extend the study to gold and vitreous carbon anodes. EXPERIMENTAL All measurements were made in gently boiling ammonia solutions at - 30& 1°C. Cells were those used previously.' Working electrodes were of vitreous carbon (Vitreous Carbons Ltd.), gold and platinum (Johnson-Matthey). Rotating disc electrodes (r.d.e.) of these three materials were prepared by machining truncated cones, half angle lo", smaller diameter 4 mm, which were push-fitted into matching teflon holders. Stationary electrodes of glassy carbon were prepared by sealing a rod, diameter 3 mm, into Pyrex whereas those of platinum and gold were made by sealing the metal wires into Pyrex and soda glass respectively.All noble metal electrodes were immersed in warm concentrated HCl prior to use and carbon electrodes were treated with 1 : 1 HN03 : H2S04 cleaning mixture. Gold and platinum r,d.e. were polished with y-alumina but those of carbon were brought to a glassy finish on very fine emery paper lubricated with ethanol because alumina particles were found to embed in the carbon surface. Counter electrodes were platinum sheets, approximately 2.5 cm square. As before,' the reference electrode was Pb/O.l N Pb(NO& and all potentials are quoted relative to it. Bulk conversions of CoIII to CoI1 were carried out at the platinum sheet electrode using as subsidiary electrode a small area of glassy carbon at which, owing to mass transfer control, the prevailing reaction was nitrogen evolution.FelI solutions were obtained in precise 140. R . BROWN AND S. A. THORNTON 15 concentration by the anodic dissolution at controlled current, of pure iron (Johnson- Matthey) in situ. In order that this dissolution should proceed at an acceptable rate without promoting passivation of the iron it was found necessary first to reduce cathodically the iron electrode for one minute by evolving hydrogen from the surface. A large vitreous carbon rod electrode (area 12 cm2) was employed to convert FelI to Fe"I at controlled current density. Tn these bulk oxidations of Fe to FeII and FeII to FelI1 platinum microcathodes (area 4x cm2) were used in order that hydrogen evolution should predominate over- whelmingly over the reverse of the anode process.Except when otherwise stated, the background electrolyte was 1 mol dmF3 NH4N03 prepared as described previously.' However, in the FelI/FelI1 experiments the ammonia was first carefully condensed over sodium and the resulting blue solution redistilled. Hexa- amminocobalt(TI1) trinitrate was prepared in the usual way and was added to the working compartment in preweighed tubes. Potentiostats, function generator and integrator were designed and built in this laboratory. Current against voltage curves were recorded on a Bryans 26,000 X- Y recorder and transients on a TOA EPR-2T chart recorder. 2 0 N I E 3 IC - B .- .= { OO RESULTS THE cO1l'/cO" SYSTEM On rotating disc electrodes of all three materials well-defined waves were obtained for the reduction of CO~~I, using 1 mo1dm-3 NH4N03 as supporting electrolyte.Limiting currents varied linearly with u)* (co = angular rotational velocity of the electrode) although a small intercept current remained at u) = 0, probably attributable to natural convection occurring in the gently boiling solution (fig. 1). Waves were analysed by plotting currents in the form log (id - i ) / i against E (fig. 2). In addition the relation log [(id-i)2/iJ against E was tested for the glassy carbon results; the superior linearity of the former plot suggested that the reaction order was unity rather than two. This order was confirmed by the independence of E+ on the concentration of ColI1. Parameters describing the reduction waves are tabulated 5.0 0.016 ELECTRODE REACTIONS I N LIQUID AMMONIA (table 1).These data were obtained at potential scan rates of 0.01 V s-l on electrodes which, prior to the scan, had been subjected to a hydrogen evolution current of 1 mA cm-2 for some 2 min in order to remove anodic films from the surface. 0.2 0.1 0 -0.1 EIV FIG. 2.-Polarisation data for CoIII reduction at rotating disc electrodes. The points 0, 0 and + refer to a 21 mol m-3 solution at platinum, an 11 mol m-3 solution at gold, and a 13.5 mol m-3 solution at glassy carbon. TABLE 1 .-CoiIJ REDUCTION WAVES iiidc Tafel slope/ E+ mV A cm mol- 1 mV ca tilode (at 0 = 150rads-I)/ (at o = 150rads-l)/ gold + 54 780 92 platinum + 80 804 71 carbon - 75 800 112 After a substantial fraction of the ColI1 in solution had been converted in a bulk reduction to CoII, polarisation data was collected at low current densities on stationary electrodes for both cathodic and anodic branches.In order to minimise hysteresis between scans in the anodic and cathodic directions, low potential sweep rates Mean currents are presented in the form of Butler-Volmer plots (fig. 3) and the reaction parameters are presented in table 2. Apparent standard rate constants are calculated in the usual way according to the relation io = nFkoa&Ea&II. a is the cathodic transfer coefficient, obtained from the Tafel slope (= 2.303 RT/ctnF). Although activity coefficients of ionic species in ammonia deviate even further from unity than in aqueous media,g we have in this work identified activities with con- centrations.V s-l) were used.0. R. BROWN AND S. A . THORNTON 17 The anodic branch of the polarisation curve was measured by cyclic linear potential scans in sections at higher current density values, using for each section scan rates sufficiently small to produce a tolerably small hysteresis. Results for platinum and gold are shown in fig. 4. IOU 1C N I 1 .c Eo % \ 5 h fs 2 0. I 4 v 0.0 0.0 0 I I I -0. I 0 0. I IlV FIG. 3.-Butler-Volmer plots for the CoII/CoIII system at platinum (O), gold (0) and glassy carbon (+) electrodes. The concentrations of CoIII and CoII are as given in table 2. TABLE 2 . - C o I I / C o ~ ~ ~ REACTION AT STATIONARY ELECTRODES Tafel slope ko(apparent)/ i0lAcm-2 at EelrnV cm s-1 [CoIIIl I [CO'II I electrode Ee/mV mol m-3 mol m-3 EolV carbon 362 12.6 5.4 0.345 6 .5 ~ lo-' 65 6 . 8 ~ gold 368 14.4 6.1 0.35 1.03 x 62 1.01 x platinum 344 14.3 6.6 0.33 4 . 9 ~ 10-7 79 4.g5 x 10-7 Oxidation of the platinum electrode surface for two minutes has a corresponding effect upon the E+ value for the subsequent reduction of ColI1. Results in table 3 show how progressively severe surface oxidation diminishes the electrode activity towards the redox process. These reduction waves were recorded at 0.03 V s-l immediately after the surface oxidation treatment.18 ELECTRODE REACTIONS IN LIQUID AMMONIA The effect of ionic strength on the kinetics of ColI1 reduction was investigated by increasing the concentration of ammonium nitrate from 0.01 to 0.1 and then to 1.0 mol dnr3 ; in each case the Co"' concentration was 3 mol m-3.The concentration of NH,NO, in the reference compartment was 1.0 mol dm-3 throughout. On vitreous carbon electrodes the reaction rate appeared to be independent of ionic strength but on gold and platinum electrodes the potential shift values (dE/d log CNH4N03)i of - 0.04 and - 0.05 V respectively were obtained. too N E 0 2 1 0 2 t t o 441 3 I I I L 1.5 0.7 0.9 EIV FIG. 4.-Tafel plots for CoII oxidation at platinum (0) and gold (0) electrodes. The concentrationi of CoII and CoIII are as in table 2. TABLE 3.-EFFECT OF PLATINUM OXIDATION UPON REDUCTION OF CO"' Pretreatment E$at w = 150 rad s-l)/ cathodic potential/V mV Tafel slope/mV 0.6 + 45 75 0.8 - 50 95 1 .o -255 125 THE Fe1I1/Fet1 SYSTEM The addition of ammonia to an aqueous solution of Ferrl precipitates the hydrous oxide.l0 Therefore it was necessary to exclude water rigorously from the liquid ammonia electrolyte when examining the electrochemical behaviour of this system.Iron dissolved quantitatively in ammonium nitrate solution to form Fe" nitrate. This was proved easily by a comparison of the weight loss from the anode and the charge passed during the dissolution. On glassy carbon Fe" gave well-formed oxidation waves only after the anode had been oxidised at +2.0 V for several minutes. Otherwise diffusion limited plateau currents were not obtained. This effect arose presumably from deactivation of large0. R . BROWN AND S . A . THORNTON 19 areas of the carbon surface. The phenomenon was not investigated further. Acti- vated carbon surfaces gave stationary electrode voltammograms (fig.5) characteristic of a simple, essentially reversible one electron transfer system. The rotating disc polarisation curves for a 7.8 mol m-3 solution of Fe" were analysed by plotting log [(id-i)/i] against E (fig. 6). Linear plots were obtained with slopes 51 mV (at o = 150 rad s-l) and 49 mV (w = 37.5 rad s-l) ; the small deviations from the theoretical N I 4 E \ *-l -0.2 - -0.1 - I I I 0.6 0.8 LO EIV FIG. 5.-Cyclic voltammogram at a stationary glassy carbon disc in an 11 mol m-3 FeII solution taken at a scan rate of 0.03 V s-'. EIV FIG. 6.-Polarisation data for Fell oxication at a glassy carbon rotating disc electrode in a 7.5 mol m-3 FeII solution taken at 150 (+) and 37.5 (0) rad s-l.20 6.0 LO 20 0- ELECTRODE RBACTIONS IN LIQUID AMMONIA - - - k 0.7 0.9 I ..I EIV FIG.7.-Cyclic voltammograms taken at 0.03 V s-' at a gold anode rotating at 150 rad s-l int 7.1 mol m-3 FeIl solution. The broken line indicated the return scan when the anodic limit:; restricted to +0.9 V. I i I 0.0 0.9 EIV FIG. 8.-Polarisation data for FeII oxidation at a gold rotating disc electrode in a 7.1 mol m-3 FeII solution. The points (a), (b), (c), and (d) are obtained after holding the electrode at +0.7, +0.95, + 1 .O and + 1.2 V respectively.0. R . BROWN AND S . A . THORNTON 21 value (47 mV at - 30°C) may indicate a small degree of irreversibility but the reaction was considered too rapid to obtain a value for the rate constant on carbon. Assuming equal diffusion and activity coefficients for Fe" and Fe"', the standard potential was equated to the E& (0.833 k0.003 V).A charge of 180 C was passed in an attempt to oxidise an appreciable fraction of the bulk Fe" to Fe"'. Thus a 30 % conversion should have been effected on the 0.55 dm3 solution (initially 1 I mol M - ~ in Fe") but the diffusion limited cathodic current, measured subsequently at a rotating carbon disc electrode, corresponded to a much smaller conversion and this current feu rapidly with time indicating a half life for FeIrL of approximately 3 min. The decay was not simply an oxidation of the solvent to nitrogen with simultaneous regeneration of Fe", because the diffusion limited Fe" oxidation current did in fact decrease as a result of the conversion, although only about half as much as expected.Addition of water to the anhydrous system accelerated the decomposition of Fe"'. FIG. 9.-Polarisation data for FeII oxidation at a platinum rotating disc electrode in a 7.0 mol m-3 FeII solution taken after 3 min at + 0.9 V (0) and + 1.3 V (@). The broken line is an extrapolation of the linear region to the half wave potential. Anodic oxidation of Fe" was examined also on gold and platinum rotating disc electrodes. On these metals it was impossible to scan up to the plateau potentials of the wave without the surface incurring severe oxidation, manifested as a con- siderable hysteresis, the currents on the cathodic (return) scan being much smaller than on the anodic scan (fig. 7). However, if the anodic limit potential for cyclic linear voltammetry was restricted to 0.9 V then the hysteresis on gold was essentially obviated (fig.7) and the polarisation curve, when analysed, showed almost reversible behaviour (E* = 0.848 V, slope 52 mV) illustrated in fig. 8. In the same diagram axe included the polarisation curves obtained with cathodic-going scans on the gold electrode following pre-polarisation for 3 min at various anodic potentials. Reversible22 ELECTRODE REACTIONS I N LIQUID AMMONIA behaviour could be regained by reactivating the gold electrode by a brief excursion below 0 mV. On platinum, inhibition of Felt oxidation by anodisation of the metal surface was more severe than on gold. Thus a platinum anode pretreated by polarisation at 0.9 V gave, on scanning in a cathodic direction, the polarisation curve shown in fig.9, indicating irreversible behaviour. DIRECT MEASUREMENTS OF ELECTRODE SURFACE OXIDATION Previously we described coulometric transient experiments in which smooth platinum surfaces were oxidised by a potential step and subsequently reduced at the end of the pulse. We now report analogous results obtained using gold wire and vitreous carbon disc electrodes (fig. 10). -0.4 V for gold and -0.7 V for carbon. Initial cathodic base potentials were FIG. 10.-Charge measurements on (a) a gold wire and (b) a glassy carbon disc in 1 mol dm-3 ammonium nitrate solution showing anodic (+) and cathodic (0) charges. DISCUSSION The reaction Fe"/Fe"' was found to be almost completely reversible on vitreous carbon and on reduced gold electrodes. This is not surprising, for no major re- arrangement of ammonia ligand molecules is necessary in converting one outer orbital sp3d2 complex ion to another.Fe3+(d5, sp3d2) + e + Fe2+ (d6, sp3d2). The corresponding reaction in aqueous acid solution is well known to be quasi reversible. Any marked deviation from reversibility of the FelI/FelIt couple in ammonia, such as that shown by platinum, must arise from very extensive inhibition of the electron transfer process. A gold surface prepolarised at 0.9 V where the oxidation0. R . BROWN AND S. A . THORNTON 23 charge is 0.4 mC cm-2 (fig. lO(a)) shows reversible behaviour. This oxidation charge corresponds to one or two monolayers of oxidised material. Fe" oxidation is decidedly irreversible when the oxidation charge is 1 mC cm-2 (at 1.2 V) and it exhibits a high Tafel slope associated with an electrode reaction at an inhibited surface.Although a surface oxidation charge of 1 mC cm-2 is reached on vitreous carbon at a much less positive potential (0.7 V), it is unlikely that full surface coverage is approached. The reason is that vitreous carbon is a microporous material l4 presenting a rough surface on an atomic scale; the carbon electrode used in these studies exhibited a low-frequency capacitance value at cathodic potentials of 0.4 mF cm-2 contrasting with 5 0 ~ F c m - ~ of the gold electrode. We conclude that the carbon surface is not extremely oxidised ( < 0.4 mC real cm-2) even at 1.5 V. Hence the reversible behaviour of the FeI1/Fe1I1 couple on vitreous carbon corresponds with the activity expected of an uninhibited surface. A platinum surface acquires an oxidation charge of 1 mC cm-2 before an electrode potential of 0.7 V is reached (ref.(7) fig. 5). As the oxidation of Fe" proceeds only under more anodic conditions, the irreversible behaviour observed with platinum is not unexpected. Even so, the polarisation behaviour at less positive potentials (fig. 9) does not diverge very far from that obtained on carbon. This behaviour contrasts markedly with the severe inhibition observed for hydrogen evolution and oxidation on platinum.' The difference must arise because the redox process is a low overlap reaction whereas the hydrogen electrode reaction proceeds through chemisorbed hydrogen atom intermediates. Previous workers have reported the formation of yellow films and yellow solutions at platinum anodes in ammonia.1s Carbon was not observed to dissolve and gold was heavily oxidised.In the present work we have not attempted to establish the chemical identities of the adsorbed layers on Pt and Au anodes. Herlem et all6 examined the anodic behaviour of silver and mercury in ammonia ; in acid solutions simple ions were formed but in neutral and basic media, amides, imides and nitrides were proposed as oxidation products. Also the salts Hg,NX where X = ClO,, bromide or iodide were found. A recent study of titanium oxidation l7 indicated the formation of an insoluble anodic film composed of Ti(NH2)4 mixed with a small amount of ammonium salt ; these are supposed to arise from the reaction of ammonia with the initially formed titanium halide.It seems probable that the films formed on gold and platinum in the present study are amides, imides or nitrides. Clearly their affect upon the kinetics of redox reactions shows that they are poor electronic con- ductors and the failure of the films to thicken suggests that they are also poor ionic conductors. Inhibition by surface oxide has been reported for the redox system FeIr/FeIrl in aqueous acid so1ution.18 The effect was more severe on platinum than on gold and was ascribed to an increase in the work function of the metal owing to the dipolar nature of the metal-oxygen bond. In terms of the theories of electron transfer reactions at electrodes 19* 2o the presence of anodic films can reduce the electron transfer rate either by (1) affecting Wand W* (respectively the work required to bring the product and reactant ions from the bulk of solution to the reaction site) or (2) reducing K the transmission coefficient which depends upon the magnitude of A the quantum splitting between the upper and lower states arising from the interaction of the degenerate reactant and product systems in the transition state.Hale has discussed 21 the dependence of A upon xp, the distance separating the ion from the metal electrode. He suggested that A decreases by an order of magnitude for each additional 0.7-1.0 A increase in xp. However, the reaction is likely to be completely24 ELECTRODE REACTIONS IN LIQUID AMMONIA adiabatic ( K = 1) until xp increases beyond 15 A. A surface oxidation charge density of 1 mC cm-2, above which platinum manifests inhibition towards the Fe"/Fe"' redox system in ammonia, corresponds to a film thickness somewhat lower than this value.It is possible therefore that effect (1) plays a part in the observed inhibition. This will arise if the point of zero charge (P.z.c.) on the oxidised surface occurs at less anodic potentials than that of the bare metal. Cobalt(II1) reduction takes place at electrode potentials considerably more negative than the Eo for the FeIr/FerIJ system. Even the Eo for the Corl/Corrl reaction is 0.5 V more cathodic than that of iron. Therefore the extent of oxidation of the anode surface, and consequently the effects of inhibition are less under the conditions used to study Co"' reduction.Nevertheless the rate constant of the Corl/Collr couple is very small (< cm s-l) and the Tafel plots are non linear (especially the anodic branches). The low rate of homogeneous exchange between ColI and ColI1 ammines and also the irreversible nature of the CO~~(NH,),/CO~~~(NH,), electrode process in aqueous solutions are well established. The high reorganisation energy for this process is clearly related to the fact that Corl ammine is an outer orbital complex whereas Co"' ammine uses the 3d orbitals for octahedral hybridisation. Thus extensive ligand rearrangement is necessary to achieve the transition state. The cathodic Tafel slope on platinum lies in the range 7 1-8 1 mV at potentials more cathodic than 0.4 V at which point the surface oxidation charge is 0.3 mC cm-2 but the kinetics are retarded at more anodic potentials (fig.3). This result suggests that monolayer coverage by amido radicals (or an equivalent partial coverage by imide or nitride) does not affect the electron transfer rate. Even allowing for the high roughness factor, carbon electrodes commence to oxidise at quite cathodic potentials (fig. 10) and this is manifested by curvature of Tafel plots beyond 0.2 V (fig. 3). On gold, appreciable variation in the transfer coefficient, a, with potential takes place where the surface oxidation is less than 20 pC cm-2. This phenomenon clearly cannot be ascribed to inhibition of the electron transfer process by a surface layer. If, on the other hand, variation in a were to be interpreted in terms of a corresponding variation in p, the symmetry factor for the reaction,22 then similar behaviour should be observed on carbon and platinum electrodes.Short of postulating a mechanism involving ion-electrode overlap, the only explanation for kinetics which differ with change of electrode material lies in terms of the variation of P.Z.C. from one substance to another. The polarisation data recorded for gold (fig. 3) can be explained qualitatively if the P.Z.C. potential for that metal is supposed to be more positive than 0.3 V where the curvature becomes significant. Systematic variation of ionic strength by alteration of the ammonium nitrate supporting electrolyte concentration corro- borates this indication of the position of the P.Z.C. potential of gold. Following this argument, the p.2.c.of platinum should occur at a still more positive potential because curvature in the Tafel plots appears only at potentials positive to 0.4 V and also the shift in the cathodic potential with ionic strength was most pronounced on platinum. Only one previous measurement of the polarisation behaviour of the system CoIr/CorJ1 has been made in liquid ammonia at low temperatures.2 Despite a value of 0.058 V for the slope of the wave analysis plot, which approaches the theoretical value 0.047 V for a reversible one electron transfer process, Laitinen concluded that the polarographic reduction of ColIr nitrate in unbuffered ammonia solution was irreversible because he was unable to observe the reverse process at a dropping mercury electrode. Laitinen determined the E+ as 0.04 V relative to the Pb/O.l N Pb(NO& reference electrode, A comparison of this value with our E+ values obtained at rotating disc electrodes must allow for Our results clearly support that view.0.R . BROWN AND S . A . THORNTON 25 this difference in mass transfer regimes. The mean Nernst diffusion layer thickness in the polarographic experiment can be estimated as approximately cm whereas the value obtained for the r.d.e. from fig. 1 using the Levich equation is 1.08 x cm at cu = 150 rad s-l. The kinematic viscosity has been taken as 3.74 x cm2 s-l at -30°C. Now the variation in Et with 6, the diffusion layer thickness, is given for an irreversible wave by aF(E%- E")/RT = In (nk,S/D). Therefore under the hydrodynamic conditions in our experiments the E4 at mercury would be moved negative by some 0.05 to 0.06 V (accepting Laitinen's value for a) to between - 0.01 and - 0.02 V.This value lies between the half wave potentials obtained at carbon and gold cathodes. Co" oxidation is clearly inhibited by film formation on the metal anodes (fig. 4). Thcre it is seen that the various sections of the polarisation curves do not join and the anodic transfer coefficients are small. The progressive oxidation of the anode surfaces at increasingly positive potentials accounts for these observations. It will be noted that inhibition of CoIr oxidation by gold oxidation is noticeable even at 0.6 V whereas kinetics of the Ferr/FeIrl system on gold were not affected until the potential exceeded 0.9 V. This is no paradox; the iron couple being rapid, requires extensive inhibition before it ceases to be fully reversible.Several workers have studied 23-26 the cobalt ammine couple in aqueous media. This is possible in the presence of excess ammonia because of the large stability constants of these complexes. Laitinen 2 5 found anodic limiting currents much smaller than the expected diffusion controlled values. He explained them in terms of a kinetic wave : slow - e Co" (outer orbital) --+ Co" (inner orbital) + Corr' (inner orbital). Three major objections exist to this proposal. First, the Co(NH,);+ ion cannot exist in the d2sp3 configuration. The reaction order in CoII(NH& was not unity over the whole concentration range. Finally the authors were unable to explain the hysteresis effects whereby currents considerably larger than the steady state limiting value were obtained on the anodic-going scan. Bartelt 26 explained this phenomenon as inhibition of the anode surface by adsorption of hydrolysis products of Co"' which also affected the exchange currents at higher CorIr concentrations.After extrapolation to eliminate these effects the standard rate constant obtained at 25°C (ito = 1.33 x cm s-l) was compared, using Marcus' theory 27 with the homogeneous isotopic exchange rate constants obtained by various workers.28* 29 Reasonable agreement was obtained 2 5 with one of Traube's values.28 If the temperature difference (55°C) between the present work and that of Bartelt alone accounts for the difference in rate constants, then a value for the mean heat of activation can be calculated from the relation Thus the mean activation energy for the cathodic and anodic reactions of the Corr/Colrl couple on platinum is AH: = 86.7 kJ mol-l.Several attempts have been made previously to relate the electrode potential scale in ammonia to that in water. Jolly 30 calculated the displacement between the scales by assuming that the complexing constant for an ion in aqueous solution is a quantita- tive measure of the free energy change for transferring an ion from ammonia to26 ELECTRODE REACTIONS I N LIQUID AMMONIA water. Pleskov assumed that the solvation energy of the rubidium ion is the same in both solvents.31 An dternative to the Rb/Rb+ couple which might be used in a similar manner is the CO~I(NH,),/CO~~~(NH,), redox system.This cannot be carried out accurately with the Eo value determined in the present work because it applies to a different temperature from that used by previous workers in their deternii- nation of the Eo in aqueous solution.32 If the temperature difference is ignored, then AEo = Eo(s.h.e., aq)-E,(s.h.e., NH3) = 0.6V. This value agrees remarkably well with that calculated by Jolly 30 as the free energy of transfer of the proton between the solvents. We express our gratitude to Mr. S . Clarke for his assistance in carrying out several of the experiments. One of us (S. A. T.) thanks S.R.C. for a research studentship. H. A. Laitinen and C. E. Shoemaker, J. Anzer. Chem. SOC., 1950,72,4975. W. B. Schaap, R. F. Conley and F. C.Schmidt, And. Chem., 1961,33,498. W. Hubicki and M. Dabkowska, Anal. Chem., 1961,33,90. T. C. Ichniowski and A. F. Clifford, J. Inorg. Nuclear Chem., 1961,22, 133. G. W. Leonard and D. E. Sellars, J. Electrochem. SOC., 1955, 102,96. J. Bjerrum and J. P. McReynolds, Inorg. Synth., 1946,2,218. J. J. Lagowski and G. A. Moczygemba in Chemistry of Non Aqueous Solvents, ed. J. J. Lagowski (Academic Press, New York, 1967), vol. 11. l o F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry (Interscience, New York, 2nd edn., 1967), p. 860. ref. (10) pp. 856 and 681. ' A. D. McElroy and H. A. Laitinen, J. Phys. Chem., 1953,57,564. ' 0. R. Brown and S. A. Thornton, J. C. S. Furaduy I, 1973, 69,1568. I2 D. H. Angel1 and T. Dickinson, J. Electrounal. Chem., 1972, 35, 55. I3 D. Gilroy and B. E. Conway, J. Phys. Chem., 1965,69,1259. l4 H. W. Davidson, Nuclear Eng., 1962,7, 159. l 5 A. K. Vijh, J. Electrochem. SOC., 1972, 119, 861. l 6 J-J. Minet, M. Herlem and A. Thiibault, J. Electrounal. Chem., 1971, 30,203. l7 A. Comte, M. Renaud, D. Deroo and M. Rigaud, Compt. rend. C, 1971, 273, 1465. J. P. Hoare, Electrochim. Actu, 1972, 17, 1907. R. A. Marcus, Canad. J. Chem., 1959,37,155. 'O V. G. Levich, Adv. Electrochem. Electrochem. Eng., 1966,4, 249. 21 J. M. Hale in Reactions of Molecules at Electrodes, ed. N. S . Hush (Interscience, London, 1971), '' R. A. Marcus, Ann. Rev. Phys. Chem., 1964,15, 155. 23 H. Bartelt, Electrochim. Acta, 1971, 16, 307. 24 L. N. Klatt and W. J. Blaedel, Anal. Chem., 1967,39, 1065. 2 5 H. A. Laitinen and P. Kivalo, J. Amer. Chem. Soc., 1953, 75, 2198. 26 H. Bartelt and S. Landazury, J. Electroarzal. Chem., 1969, 22, 105. 27 R. A. Marcus, J. Chem. Phys., 1965, 43, 679. 2 8 J. F. Endicott and H. Traube, J. Amer. Chem. SOC., 1964,86, 1686. 29 N. S. Birader, D. R. Stranks and M. S. Vaidya, Trans. Faraduy SOC., 1962,58,2421. 30 W. L. Jolly, J. Chem. Ed., 1956, 33, 512. 31 V. A. Pleskov, Fortschr. Chem. U.S.S. R. 1947, 16, 254. 37 J. Bjerrum, Metal Ammine Formation in Aqueous Solution (Haase, Copenhagen, 1941), p. 250 p. 229.

ISSN:0300-9599    

DOI:10.1039/F19747000014

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Activity Coefficients of Hydrochloric Acid in Mixtures with Potassium Chloride in Methanol+ Water Solvents from Electro- motive Force Measurements at 298.15 M BY CHAN CHEE-YAN" AND KEAN H. KHOO Department of Chemistry, University of Malaya, Kuala Lumpur, Malaysia Received 2nd July, 1973 E.m.f. measurements of the cell, PtlHz I HCl(rnl), KCl(m,), methanol(X), waterlAgClIAg, have been made at 298.15 K for solvent compositions X = 0, 10,30,40 and 45 % (w/w) of methanol, the total ionic strength of each solution being 1 mol kg-'. The logarithm of the mean ionic molal activity coefficient of hydrochloric acid was found to vary linearly with potassium chloride molality for the systems in which the solvent mixture contained 0, 10 and 30 % methanol. For methanol composition above 30 %, the activity coefficient of the acid was a quadratic function of the salt concentration.Secondary medium effects are discussed in relation to differences in specific ionic interactions for the different electrolyte mixtures in water and in the mixed solvents. It has been shown 1 * from precise emf. measurements that for the system (HCl, NaCl, methanol, water) at constant ionic strength of 1 mol kg-1 the mean ionic molal activity coefficient, y of hydrochloric acid varies linearly with sodium chloride molal concentration according to eqn (1) at each solvent composition, ranging from 0 to 60 % (w/w) of methanol, at constant temperature : (1) where 1 and 2 refer to the x i d and the salt respectively. 7; is the mean ionic molal activity coefficient of the acid at the same total ionic strength but in the absence of the salt, and uI2 is an empirical parameter.A significant feature found for this system is that the parameter a I 2 is the same for each solvent mixture at constant temper- ature. It would be reasonable in this case, according to H ~ n e d , ~ to assume that the parameters in eqn (2), which represents the variation of the activity coefficient of the salt, would also be independent of solvent composition : Furthermore, this would lead to eqn (3).2* log y1 = log y;-a,,nz, log y 2 = log y;-a2,ln1 -pzlm:. (2) where m is the constant total ionic strength. Therefore, if the activity coefficient of one of the electrolytes is known in the methanol+water mixture, that of the second electrolyte may be calculated using eqn (3) from the activity coefficient ratio determined in water alone as the solvent. Harned reported that the values of the activity coefficients of lithium chloride and sodium chloride in methanol + water mixtures calculated by this method were in good 2728 ACTIVITY COEFFICIENTS OF HC1 AND NaCl agreement with experimental results but those for potassium chloride were not.Very few e.m.f. data for systems involving mixtures of electrolytes in mixed solvents are available for testing the general applicability of Harned's method. The present work gives the reason for the non-applicability of Harned's computation for t hc systems involving hydrochloric acid, potassium chloride, methanol, and water. The influence of methanol+ water solvent structure on the parameters attributed to specific ionic interactions is also discussed.EXPERIMENTAL E.m.f. measurements were made using a Pye Universal potentiometer with a Pye Scalamp galvanometer as a null detector. The experimental cell used is a slight modification of the one used by Kh00.~ The experimental technique used was a combination of that used before and that used by Dowries.' Both silver-silver chloride and hydrogen electrodes were used in pairs and were similar to those used by Downes. They were prepared, stored and used in the same manner as before.4* The bias potentials of the silver-silver chloride electrodes used in the measurements were & 5 x V, or less, with reference to an aged reference silver-silver chloride electrode stored in 0.1 mol kg-1 hydrochloric acid aqueous solution. Each reported e.m.f.reading given in table 1 was the mean of four readings obtained from a pair of silver- silver chloride electrodes and a pair of hydrogen electrodes. The four readings of each set at any given time were within +3x The aqueous hydrochloric acid used was the constant-boiling mixture doubly-distilled from the A.R. grade acid. A.R. potassium chloride was recrystallised before use. Anhydrous methanol was obtained from A.R. methanol by distillation as before.6 Distilled water obtained from an all-glass still was re-distilled over alkaline potassium permanganate solution and the water used in the experiments was obtained by re-distillation of this conden- sate. Hydrogen and nitrogen purification were as before except that instead of using copper filings heated at 450°C for the removal of traces of oxygen, we used a commercial catalyst, labelled BASF BTS catalyst, heated at 140°C.The BTS catalyst proved to be just as efficient. Corrections of measured e.m.f. readings to a hydrogen partial pressure of 1 atm (101 325 N m-2) were made using data interpolated from the vapour pressure corrections at different temperatures tabulated by Akerlof et al.' for methanol+ water mixtures with an electrolyte ionic strength of 1 mol kg-l. The temperature was maintained at 298.15 Kk0.01 K. SolubiIity limitations of potassium chloride in the methanol+ water mixtures restricted the range of methanol composition to 0-45 % in this work. V of each other. RESULTS AND DISCUSSION The e.m.f. data are given in table 1, where Eis the e.1n.f.of the cell, PtlH21HCl(rnl), KCl(rn2), methanol(X), waterIAgClIAg, corrected to a hydrogen partial pressure of 1 atm, and m, +m, = 1 mol kg-l. The best-fit values of the parameters, a I 2 and Pl2, and of E" for each methanol composition, A', are computed by the method of least squares using eqn (4). Assuming for each mixture that we have log Y1 = 1% Y ; - % 2 ~ v - - P 1 2 4 , Y = ES2klog m, = 2ka,,m2 +2k/?,,m% + E" (4) where E" = (E&IArfC1-2k logy;), k = 0.059 158,' and EGIAgCl is the standard electrode potential in the given solvent. Table 1 gives values of AY, where A Y = Y(expt.) - Y(ca1c.). The values of Y(ca1c.) were from eqn (4) and the best-fit valuesC . - Y . CHAN AND K. H. KHOO 29 of a12, Pl2 and E". Values of E&iAgCI for each solvent mixture given in table 2 were derived from the best-fit values of E" using y'; data obtained elsewhere.s These agree favourably with those reported by Maya Paabo, Bates and Robinson.8 TABLE 1 .-E.m.f.DATA FOR THE SYSlEM (HCl(ml), KCl(mz), METHANOL(X), WATER) AT 298.15 K ; ml+rn2 = 1 mol kg-1 1 o5a ri 105a ri 1 0 5 ~ Y/ EIV Y(expt.)/V V ~ ~ 1 2 1 inolkg-1 E/V Y(expt.)/V V E/V Y(expt.)/V V X = O % x= 10% X = 3 0 % 0 0.1 0.3 0.4 0.5 0.7 0.8 0.9 0 0.1 0.3 0.4 0.5 0.7 0.8 0.9 0.233 20 0.236 53 0.244 27 0.248 83 0.254 23 0.268 54 0.279 71 0.298 13 0.233 20 2 0.233 82 0 0.235 11 0 0.23642 3 0.238 36 4 0.238 97 1 0.23571 -4 0.237 61 -6 0.227 90 0,227 90 -3 0.21998 0.223 43 0.239 13 0.229 97 0 0.231 46 0.243 83 0.23071 6 0.23620 0.249 15 0.231 34 2 0.241 65 0.274 71 0.233 36 0 0.267 58 0.293 20 0.234 04 0 0.286 15 0.263 56 0.232 63 -5 0.256 39 0.216 66 0.220 18 0.228 07 0.232 91 0.238 34 0.253 12 0.264 51 0.283 22 0.216 66 -4 0.21475 0.21475 2 0.21747 5 0.218 19 0.21548 -1 0.21891 -3 0.22624 0.21707 -1 0.22053 -2 0.23654 0.218 73 -1 0.222 19 -6 0.251 45 0.220 51 3 0.22406 1 0.281 44 0.222 28 -2 0.219 79 6 0.223 16 2 0.21998 3 0.222 30 0 0.22308 0 0.22546 3 0.226 23 1 0.220 72 -2 0.223 84 -3 0.22699 - 1 TABLE 2.-vALUES OF THE PARAMETERS, CClz and /?I*, AND EiglAgC1 Xi:< crlz/mol 1 kg P12/mol-2 kg2 E"IV Y F 8 E 2 g InscllV (0.222 34) 10 0.0573+0.0005 0 0.227 93 f O.OO0 03 0.783 0.215 4 (0.215 5) (0.203 1) 0.009 92 0.216 70 0.683 0.197 1 (0.196 8) 0.214 73 0.668 0.194 0 (0.194 1) 0 0.0543+0.0004 0 0.233 1 8+ 0.000 03 0.809$- 0.001 0.222 29 30 0.0662+ 0.0002 0 0.219 95+0.000 01 0.715 0.202 7 40 0.0601 45 0.0636 0.008 28 ft EAglAgcl values in parentheses are from ref.(8). HARNED'S METHOD Values of the parameters, a1 and p1 2 , varied significantly with solvent composition (see table 2). This clearly indicates that Harned's method as discussed previously cannot be applied to the system (HCI, KCI, methanol, water) as the critical assumption of constancy in cxl2 and Pl2 for the various solvent mixtures is not valid in this case. From tables 1 and 2 it is seen that the logarithm of the activity coefficient of hydrochloric acid is a linear function of potassium chloride molality for X = 0, 10 and 30 % methanol, but above 30 % methanol it is a quadratic function of the salt30 A C T I V I T Y COEFFICIENTS OF HCl AND NaCl molality. Akerlof et al.' have shown that the linear equation for hydrochloric acid in the system (HCl, NaCl, methanol, water) at constant temperature fitted their experi- mental data over the whole range of X = 0-60 % methanol.In view of this fact, the deviations from linearity as observed for the system (HCl, KCl, methanol, water) for methanol composition above 30 % could possibly indicate that at the higher methanol compositions, ion association between potassium and chloride ioiis may be just significant but is slight enough to be accounted for by the inclusion of the Dl2n?; term as in eqn (4) where no association is assumed. SECONDARY MEDIUM EFFECTS The secondary medium effect for a given electrolyte in a given electrolyte-mixed solvent system is defined by the quotient J / ~ ? , where the subscripts w, s denote water and the mixed solvent, respectively. We have shown that for hydrochloric acid at constant ionic strength in the systems studied, 1% wY1 = 1% wY;-wa12f1l2 ( 5 ) and We can then write X/ %(w/w) methanol FIG.1.-Plots of -A(or12, Pl2, ~ n ~ ) ~ , ~ against solvent composition X . A, m2 = 0 mol kg-' ; B, m2 = 0.3 rnol kg-' ; C , m2 = 0.5 mol kg-' ; D, uz2 = 0.7 mol kg-' ; E, m2 = 0.9 mol kg-' ; F, m2 = 1 mol kg-', trace concentration of HCl.C . - Y . CHAN AND K. H. KHOO 31 The sum of the terms on the right in eqn (7) gives the difference between the second- ary medium effect for hydrochloric acid in the presence of a given concentration of potassium chloride at constant total ionic strength and that for the acid alone at the same total ionic strength for the same solvent.This quantity, denoted by A(alz, Pl2, m2)w,s, would be some function (though uncertain) of the specific ionic interactions ''9 of the two electrolytes in water and in the mixed solvent. Its variation with solvent composition for different potassium chloride concentrations, m2, are given by the plots in fig. 1. The plots show marked inflexions in the region of methanol composi- tion between X = 25 and 40 %. Solvent structure changes with composition in mixed solvent systems and effects of such changes on certain properties (e.g. acid-base behaviour, ionic transport) of single electrolytes dissolved in alcohol+water mixtures have been known l 2 9 l3 to result in observed extrema in the variation of these properties with solvent composition.The extrema are considered to occur in the region of composition in which the order- disorder relationships 12* l3 in the structure of the mixed solvent are undergoing the most pronounced changes. Similar solvent structural effects are manifested by the inflexions in the plots in fig. 1. Ionic interactions have been conveniently, but arbitrarily, separated into non- specific ionic interactions of the type considered in the well-known Debye-Huckel theory for very dilute solutions and specific ionic interactions including ion-solvent interactions different from the Debye-Huckel type.l' Thus in eqn (8) which gives the activity coefficient for hydrochloric acid in a mixed so1vent,14 the term logyyH is assumed to account for non-specific ionic interactions and the term bTm for specific ionic interactions.log 7 1" = - Am'/( 1 + 4.3Bmf) - log (1 + 0.002mm) + bim = log yYH+b$,2 (GI where A , B are the Debye-Huckel constants on the molal scale, M the mean molecular weight of the solvent and bi is the empirical specific ionic interaction parameter for hydrochloric acid. From eqn (7) and (8), we can write where m = rn, +m2 = 1 mol kg-l. Thus the secondary medium effect for hydro- chloric acid is given by the sum of the contributions of the effects of change in solvent composition on the Debye-Huckel non-specific ionic interactions and the specific ionic interactions in the electrolyte mixture. Values of for the various methanol compositions have been computed using data on the values of A , B, M , and 7; reported el~ewhere,'~ interpolating data where necessary.Fig. 2 shows that no marked characteristics are observed for the variation of log(wyyH/sy:H) with solvent composition, but for - A(bi, a12, Pl2, m2)w,s versus solvent composition the plots show distinct maxima in the same region of solvent composition as for the plots in fig. 1 where inflexions occurred. The positions of the maxima appear to be shifted in the direction of higher methanol concentration with increasing concentration of potassium ions. Thus it appears that solvent structural effects are reflected in the parameters used to account for specific ionic interactions but not in the commonly used deb ye-Huckel term whose variation with solvent composition is considered as merely a variation with solvent dielectric constant and density.32 ACTIVITY COEFFICIENTS OF HCl AND NaCl 0 10 2 0 3 0 4 0 5 0 X] %(w/w) methanol FIG.2.-A-G, plots of -A@;, a12, Pl2, m2)w,sagainst solvent composition X . A, m2 = 0 mol kg-' ; B, mz = 0.1 mol kg-' ; C, m2 = 0.3 mol kg-l ; D, m2 = 0.5 rnol kg-' ; E, m2 = 0.7 rnol kg-' ; F, m2 = 0.9 mol kg-I ; G, m2 = 1 mol kg-', trace concentration of HCI. 0- . -, plot of Iog(,yya/ SyyH) against solvent composition X. We thank Mr. W. W. Koh and Mr. S. H. Yeow for their assistance. G. Akerlof, J. W. Teare and H. Turck, J. Amer. Chetn. Soc., 1937, 59, 1916. H. S. Harned and R. A. Robinson, Multicomponent Electrolyte Solutions (Pergamon, Oxford, 1968), p. 96. €3. S. Harned, J. Phys. Chem., 1962, 66, 589. C. J. Downes, J. Phys. Chem., 1970, 74, 2153. D. Feakins and K. H. Khoo, J. Chem. SOC. A , 1970, 361. 1969), p. 469. Maya Paabo, R. G. Bates and R. A. Robinson, Anal. Chem., 1965,37,463. ref. (7), p. 354. York, 1967), p. 515. ref. (2), p. 10. and Francis, London, 1968), p. 59. 4K. H. Khoo, J. Chem. SOC. A, 1971, 1177. ' R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworths, London, 2nd edn., l o H. S. Harned and B. B. Owen, The Physical Chemistry of Electrolytic Solutions (Reinhold, New l2 R. G. Bates, Hydrogen-Bonded Solvent Systems, ed. A. K. Covington and P. Jones (Taylor l 3 F. Franks and D. J. G. Ives, Quart. Reu., 1966,20, 1. l4 ref. (lo), p. 472. l5 R. G. Bates and R. A. Robinson, Chemical Physics of Ionic Solutions, ed. B. E. Conway and R. G. Barradas (Wiley, New York, 1966), p. 211.

ISSN:0300-9599    

DOI:10.1039/F19747000027

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Gas Phase Addition of HI to Ketene and the Kinetics of Decomposition of the Acetyl Radical BY LAJOS SZIROVICZA AND ROBIN WALSH * Department of Chemistry, The University, Whiteknights, Reading RG6 2AD Received 11 th May, 1973 The addition of HI to ketene results in the rapid formation of acetyl iodide at temperatures 498-525 K. The subsequent slower reaction between HI and acetyl iodide, which produces methane and acetaldehyde has been investigated at low conversions (c 10 %), at various initial pressures at 507 K. The results are shown to be consistent with the mechanism Most experiments were done in the presence of the inert additives N2 or cyclo-C4Fs. (M)+CH3CO + CH3+CO+(M) (5) CH3CO + HI + CH3CHO + I (3) CHs+HI + CH4+I. (6) The data at 507 K yield kF/k3 = (2.4k0.3) x mol dm-3 and Z/S = 65+ 8 dm3 mol-', where I and S refer to the intercept and slope of a Lindemann plot for the unimolecular step (5).An in- dependent estimate of k3 is used in conjunction with theoretical calculations on the acetyl decomposition reaction to show that, The limits cover all reasonable uncertainties, both experimental and theoretical. This result disposes of previous reports of a " low " A factor for this reaction. log (@/s-') = (13.3 +0.5)-(91.2+ 7.5 kJ m01-~)/2.303 RT. The first order rate constant for the decomposition of the acetyl radical (CH3CO) is known to be pressure dependent under experimentally accessible conditions.' * This behaviour is characteristic of a unimolecular reaction in its " fall-off " region. In their study of the photolysis of acetone in the presence of HI, O'Neal and Benson (hereafter referred to as OB) obtained, by an extrapolation technique to infinite pressure log (km/s-') = 10.3-62.8 kJ mol-'/2.303 RT.In a separate study, using a different system, Kerr and Calvert (hereafter refmed to as KC) concluded that their results were consistent with this expression. The A factor of 1010.3 s-l is considerably lower than that for propionyl de- composition and Frey and VinaII have carried out theoretical calculations which show that the observed pressure dependence of these rate constants is more consistent with an A factor - 1012e5 s-l. In view of these inconsistencies we decided to attempt a reinvestigation of this system. In a previous publication it was noted that the addition reactions of HI to unsaturates provided a means of generating radicals and studying their unimolecular reactions by a competitive method.This technique is applied here to the case of HI addition to ketene. 'r present address : Institute of General and Physical Chemistry of the University of Szeged Szeged 1, Hungary. 1-2 3334 ADDITION OF HI TO KETENE EXPERIMENTAL APPARATUS This was similar to that described in an earlier paper.6 MATERIALS HI vapour was prepared from a solution (Fisons) by dehydration with P205. The gas was passed through a trap at - 78°C and collected at - 196°C. It was stored at room temperature in a blackened bulb. KETENE was prepared by the pyrolysis of acetic anhydride by the method of Jenkin~.~ The gas was passed through 2 traps at -78°C and collected at - 196°C. It was stored at room temperature (at pressures < 100 Torr*) in a blackened bulb.V.P.C. analysis showed the presence of small quantities of C2H4 (-0.1 %) impurity, but the absence (<0.01 %) of CH4 and CH3CH0 or other impurities. CH4 was obtained from Cambrian Chemicals. CH3CH0 was obtained from Fisons. N2 (white spot) was supplied by British Oxygen Co. PERFLUOROCYCLOBUTANE (c-C4F8) was obtained from Matheson. It was >98 % pure and contained no impurity which interfered with reaction product analysis by V.P.C. PROCEDURE Prior to a kinetic run, HI and ketene (CH,CO) were both degassed at - 196°C. CH2C0 was shared into the reaction vessel at a known pressure. HI was then added and the total pressure monitored by the pressure transducer.Inert gas (N, or c-C4FC) was added 3-4 min after HI. After a suitable time the run was ended by expansion of the contents of the reaction vessel through a tube containing solid glycine to remove HI and into a sample pipette. The contents of the sample pipette were then taken for V.P.C. analysis. The sampling system, pipette and sample inlet valve and tubes on the V.P.C. were all heated to - 60°C to avoid condensation or adsorption (other than of HI). Separate tests showed that no losses of either CH4 or CH3CH0 occurred during the sampling operation and that, furthermore, these products were not generated from others amongst the reaction products during sampling or analysis. ANALYSIS AND PRODUCT IDENTIFICATION Routine analysis by V.P.C. of the contents of the sample pipette was performed on a 2.5 mx 3 mm Porapak Q column heated to 100°C with N2 carrier gas at an inlet pressure of 10 p.s.i.This was used to separate CH4 and CH3CH0. The flame detector was calibrated with carefully made up mixtures of these products in the presence of known quantities of the inert gases. This was important since, in the presence of c-C4F8, the detector response for CH3CH0 depended to some extent on the quantity of c-C,F8 (which eluted before it). Product analyses were, therefore, always carried out in the presence of a fixed quantity of c - C ~ F ~ . CO, which was expected amongst the products, was not analysed since it gave no signal on the flame detector. All analyses were performed in duplicate and the results averaged. The major product (>90 %) of addition of HI to CH2C0 at 488 K was shown to be acetyl iodide CH3COI by comparison of the n.m.r.spectrum (z = 7.04), and i.r. spectrum (v for >C=O stretch = 1757 an-') with those of an authentic sample made by reaction of CH3COCI and HI. Attempts at V.P.C. analysis (on a ppG column) of CH3COI were un- successful and the only (unidentified) impurity (7+3 %) in the reaction product n.m.r. spectrum has an absorption at z = 2.7 (singlet). In particular CH31 and iodoacetaldehyde. CH21CH0, were clearly absent (< 0.5 %). * 1 Tom = 133.3 N m-".L. SZIROVICZA AND R . WALSH 35 RESULTS PRELIMINARY EXPERIMENTS When HI and CH2C0 were mixed at 488 K an immediate rapid pressure decrease occurred. It was complete in <2 min (pressures up to 100 Ton of each reactant).The product of this reaction was identified as CH3COI (see experimental section). Although the initial pressures of the reacting mixtures were impossible to record, calculations from apparatus sharing ratios showed the pressure changes to correspond within f 5 % to the completion of the reaction ; HI + CH2C0 -+ CH3COI. No further pressure changes were recorded for up to 30 min. of excess HI, corresponds closely with that of Benson and O’Neal temperature range 495-539 K, investigated the kinetics of the reaction ; This reaction system, therefore, after the initial rapid addition and in the presence who, in the CH,COI+HI + CH,CHO+12. observed the formation of CH,CHO, but found no CH4. We have also obtained CH,CHO, but in addition our detailed analytical evidence shows that at 506.7 K during the first 10 % conversion to CH,CHO, small quantities of CH4 (from 1-5 % of the CH3CHO) are produced.The accepted mechanism ’* of formation of CH,CHO is the following : Benson and O’Neal M+I2 +21+M I + CHSCOI + CH3CO + 12 CH3C0 + HI e CH3CH0 +I. The formation of CH4 suggests in addition ; (M) + CH3C0 -+ CH, + CO + (M) (5) CH3+HI -+ CH4+I CH3+12 -+ CH31+I. Supporting evidence for the occurrefice of these added steps was obtained from the kinetic tests described in the next section. KINETIC TESTS OF THE MECHANISM In a series of experiments at 506.7 K, the ratio R = [CH,]/[CH,CHO] was measured as a function of time at three different initial pressures of HI and in the presence of excess N2 to a constant total pressure of -340 Torr.The results are plotted in fig. 1. They support the proposed mechanism in two ways. First the decrease in the ratio R, with time, at fixed initial conditions, is consistent with the competition between steps (6) and (7). At the start of the reaction in the virtual absence of 12, all CH3 is scavenged by HI in step (6), but as the reaction proceeds and I2 is formed, it competes for CH3 in step (7) thus reducing the CH4 yield. This effect is most marked for the lowest HI concentration, since at a given conversion of CH3COI to CH3CH0 the highest conversion of HI to I2 occurs. The effective competition of I2 with HI for CH3 even during these early stages of reaction (80 min corresponds to II 10 % conversion of CH,COI) is consistent with the known ratio l o of rate constants, k6/k, 2: 0.14.36 ADDITION OF HI TO KETENE I 2 0 4 0 6 0 8 0 time /min FIG.1.-The time dependence of [CH,]/[CH,CHO] (=R) : 0, HI = 4 Torr ; A, HI = 10 Torr ; 0 HI = 15 Torr. N2 added to total pressure - 340 Torr. Temperature = 507 K. [H1lo /Torr FIG. 2.-The effect of HI on [CH,CHO],/[CH410 (=&I). [CH,CO], - 4 Tom. N2 added to total pressure - 340 Torr. Temperature = 507 K.L . SZIROVICZA AND R . WALSH 37 Secondly the ratio R, is largest at low HI concentrations, where the CH3C0 decomposition step (5) competes most favourably with the CH3C0 scavenging step (3). At the outset of the reaction, steps (4) and (7) are unimportant and a stationary state treatment of steps (3), (5) and (6) yields d[CH3 CHO] /d[CH4] = k3 [HI]/k5. Since d[CH,CHO]/d[CH,] = l/Ro, a plot of l / R o against [HI] should be linear. Fig.2 shows such a plot where Ro values are taken as the intercepts of the lines (assumed linear) shown in fig. 1. Similar plots to those shown in fig. 1 and 2 were obtained in the presence of c-C4F8 (at fixed total pressure). This latter inert gas tended to give higher yields of CH4 (larger values of R,) in support of the fact that k , is in its pressure dependent region, and cyclo-C4F8 is a more efficient energy transfer agent. Because Ro is directly proportional to k,, the effect of variation of pressure of inert gas on Ro offers a direct test of the pressure dependence of k,. This test is described in the next section. PRESSURE DEPENDENCE OF THE ACETYL DECOMPOSITION In a series of experiments at 506.7 K, with [CH3CO1], = 4.3 Torr, HI = 10.0 torr, the ratio R was measured after 25 min for a number of different pressures of added c-C4F8.The conditions of these experiments were chosen to keep the total con- version to CH,+CH,CHO small (in practice between 3 and 5 %), but also to allow sufficient time for reaction that the delay in addition of cyclo-C,F8 leads to only a small overall timing uncertainty. The R values obtained were increased by a factor of 1.125 to convert them to R,. This correction corresponds with the time dependence of R observed previously (fig. 1). The results shown in table 1 indicate a variation of a factor of 10 in the ratio Ro for an 18 fold variation in total pressure. The rate constant k5 is clearly pressure dependent.TABLE THE PRESSURE DEPENDENCE OF Ro AT 507 K Ro - [CH410 pressure/Torr effective a pressure/Torr [CH 3CHOlo 720 713.5 0.0450 634 627.5 0.041 6 504 497.5 0.0371 464.5 458 0.0356 395.5 3 89 0.0286 358.5 352 0.0275 198 191.5 0.01 80 168 161.5 0.01 66 89.5 83 0.00885 65 58.5 0.0073 41 34.5 0.0044 a See text ; b Ro = k5/k3[HIl0 : [HIIo = constant = 10 Torr. To compare this data with previous work and also with theoretical calculations described later, a plot was made of l/R,[HI] against [concentration]-l. Since the composition of mixtures in our experiments was variable, the concentrations have been obtained not from the total pressures, but rather from adjusted pressures expressed in terms of c-C,F,. These have been worked out by allowance for the38 ADDITION OF HI TO KETENE different collisional efficiency, I i , diameter, oI, and reduced mass, pi, of each gas in the mixture according to ; The values for I * , of and pi used were obtained either directly or by analogy from the compilation of Rabinovitch et Q Z .~ ~ and are shown in table 2. The effect of this adjustment is fairly small except for the lowest pressure where HI and CH3COI make up 30 % of the mixture. TABLE 2.-PARAMETERS FOR GAS COLLISIONS WITH CHjCO gas 1 OlA O’lA p/a.rn.u. CH3COI 1 .o 5.1 4.8 34.3 HI 0.6 3.5 4.0 32.2 C~CIO-C,F* 1 .o 5.7 5.1 35.4 a u’ = $(u+ UCH~CO) where UCH~CO = 4.5 A. 80 - 4 7 0 - 6 0 - Q > 5 0 - P qo 40- X 3 0 - .-( “E X Y 1 H I o, 20- I 100 200 300 400 500 600 700 6 0 0 9 0 0 1000 [MI-’ /dm3 mol-’ FIG. 3.-Lindemann-type plot for product formation at 507 K : 0, upper line, this work ; lower line OB.Fig. 3 shows a least-mean-squares linear plot of our data and also includes the line obtained by OB’ for the same product function at 508.7 K. Although the inter- cepts are similar, there is clearly a substantial disagreement on slopes. The data are fitted reasonably well by straight lines, as predicted by the original Lindemann treatment of unimolecular reactions. Although modern RRKM theory l2 in prin- ciple predicts a curve for such a plot, this curve can, as is shown later, approximate closely to a straight line. In these circumstances such a plot can be usefully employed as a criterion of “ f a l l - ~ f f ” . ~ For this plot the intercept and slope correspond closely to k3/kg and k3/k; where and k; are high pressure (first order) and low pressure (second order) limiting rate constants respectively. The ratio intercept/ slope ( = I / S ) approximates to kz/ky and is independent of k3.Thus I/S is a para- meter of the acetyl decomposition alone (regardless of the inadequacies of theL. SZIROVICZA AND R . WALSH 39 Lindemann treatment). Our data at 506.7 K gave I / S = 65.4f7.9 dm3 mol-'. In addition? Z = (4.77k0.57) x lo4 dm3 mol-l. The theoretical calculations show this to be a 15 % overestimate of k 3 / k y and hence k y / k , = (2.41 f0.34) x lo-, mol dm-3. DISCUSSION Our values for k y / k , are in satisfactory agreement with those of OB1 but our quoted Z/S values for the Lindemann plot differ considerably from theirs (252.4 dm3 mol-' at 508.7 K).We do not understand the reason for this difference but note that the discrepancy is greatest at low pressures, with OB finding a greater extent of acetyl decomposition. A possible explanation is that the acetyl radical carries over some internal energy when formed photochemically in OB's system. This could not occur with our purely thermal method of generation. Further support for the validity of our results is obtained from the A factor derived from theoretical consideration of the pressure dependence of k,. Following the approach adopted by Vinall and Frey we have employed the Forst procedure l 3 to calculate k5 as a function of pressure for various values of A? and Ey and followed this by subjecting a plot of k; against (concentration)-l to a least squares analysis in our experimental (adjusted) pressure range, in order to obtain Z/S values for com- parison with experiment. The Forst procedure calculates the specific rate constant for decomposition as a function of energy, k(E), where N(E) represents the density of states at energy E, for the reactant, and A" and E" are the high pressure Arrhenius parameters.k(E) = 0 for E < E*. The important feature of eqn (A) is that it enables unimolecular fall-off curves to be calculated without consideration of the structure of the transition state. Only the reactant species requires a vibrational assignment. It should be noted, however, that while k(E) calculated by this method gives the correct average, (k(E)), for thermal systems, it is not the k(E) as normally defined in RRKM theory.13 Using the same assignment for acetyl as Vinall and Frey we obtained the Z/S values shown in table 3.k(E) = A"N(E-E")/N(E) (E 3 E") (A) TABLE 3.-THEORETICAL VALUES OF I / s AT 506.7 K FOR ACETYL DECOMPOSITION 17 19 21 25 23 11.5 537.2 912.8 1487 2346 3609 5445 12.1 149.7 260.7 433.6 692.7 1069 1606 12.7 40.07 71.42 121.8 199.4 315.0 481.6 13.3 10.36 18.76 32.60 54.54 88.15 138.0 13.9 2.626 4.791 8.411 14.26 23.44 37.39 14.5 0.6614 1.209 2.131 3.633 6.016 9.670 Collision parameters in table 2. Acetyl vibrational assignment, vlcm-': 3024, 2996, 2967, 1743, 1441(2), 1352, 1122,919, 867, 509,128. I/S in dm3 mol-' : ( I / S ) obs = (65.4f 7.9) dm3 mol-I. Bath gas is cyclo-C4F8. The plots from which these I / S values were obtained? in all cases had correlation coefficients of better than 0.9994 and the deviations from linearity were very slight over the observed (concentration)-l range, thus fully justifying the linear fit of the experimental data.The intercepts were found to be less than 1-20 % higher than the values of (ky)-'. The calculations were repeated for reasonable alterations in collision diameter, collision efficiency and vibrational assignment, but the resulting40 ADDITION OF HI TO KETENE Z/S values were well within a factor of two of those shown for a given A? and E," From the data of table 3, pairs of A: and EY for which I / S equalled the observed figure (65.4 dm3 mol-l) were extracted (by interpolation). The locus of these values is shown in fig. 4. This graph also shows the locus of AT and E," values required to fit k? = 103e9 s-l.This latter value is obtained from the experimental figure for k?/k3 at 506.7 K and the best available estimate for k3, viz : log (k3/s-l) = 9.20 - 6.3 kJ m01-'/2.303 RT. The intersection of these two lines in fig. 4 then permits a unique estimate of the high pressure Arrhenius parameters for the acetyl decomposition log (kY/s-l) = (13.3*0.5)-(91.217.5 W mol-')/2.303 RT. The error limits quoted here correspond to the shaded area in the figure which was obtained simply by assuming the maximum likely error in either Z/S or k? would be a factor of 2. FIG. 4.-Determination of Am and Em for acetyl decomposition. A, line giving correct ka, at 507 K ; B, line giving correct I / S value at 507 K. This approach to obtaining the high pressure limiting Arrhenius data for uni- molecular reactions in their pressure dependent regions is clearly superior to that of extrapolation of Lindemann plots combined with the traditional Arrhenius plot of the intercepts, which are subject to large error and can usually only be obtained over a restricted temperature range.of acetyl decomposi- tion and obtained log (A?/+) - 12.5 (cf. 10.3 observed) at each of four temperatures at which data was available. These values clearly improve upon the unreasonably low original figures, but our new results remove all remaining inconsistency with transition state theory expectations for bond breaking reactions (viz. A > kT/h = 1013e1 s-l). Furthermore, they show the expected similarity with the propionyl decomposition for which Kerr and Lloyd Vinall and Frey applied the method to the earlier studies obtained log (ka/s-') = 13.32-61.5 kcal m01-~/2.303 RT.L .SZIROVICZA AND R . WALSH 41 Despite the now reasonable A factor, and the agreement in absolute magnitude of k? between this work and that of OB1 there remain discrepancies in the absolute magnitude of k y implied by our Arrhenius Parameters and values obtained by other workers.2* 4* 14* l 5 Table 4 illustrates the differences. TABLE 4.-cOMPARISON OF k r VALUES WITH PREVIOUS DETERMINATIONS reference temp./K k? Is- 1 k?(corr.)/s-1 a kF(caIc.)/s-l b 14 298 2.5 0.4 0.0023 (0.01 3) 15 298 13 2.0 0.0023 (0.013) 2 338 94 20 0.16 (0.74) 2 d 298 - 0.007 0.0023 (0.01 3) 4 325.7 23.4 2.0 0.047 (0.24) a See text ; b calculated from our Arrhenius equation, maximum values in parentheses ; c calculated by OBI ; d data obtained from reverse addition of CH3 + CO.k? calculated via estimated equilibrium constant. Most of the previous work is based on comparative studies where the comparison reaction was either 2 CH3CO -+ (CH,CO), (8) (9) or The rate constants ks and kg have not been directly measured and assumptions about their magnitudes had to be made. Recent studies on alkylradical recombination 6* have shown that recombination rate constants are much smaller than was hitherto thought. The new figures for these are much more consistent with rate constants for the reverse bond-breaking processes for hydrocarbons l8 and the currently accepted thermodynamic data on free radi~a1s.l~ We have therefore taken the rate data 2o for biacetyl decomposition and used it in conjunction with the thermodynamic data (see Appendix) to estimate k8 = 108m4 dm3 mokl s-l.Assuming the geometric mean rule then kg = mol-1 s-l . These values, assumed temperature independent, were used to recalculate the figures for kg, which are also shown in table 4. When these corrected figures are compared with our own the discrepancies are somewhat reduced and in the case of KC’s study of CH3 addition to CO there is implied agree- ment within experimental error. Even so the same workers’ direct data on acetyl decomposition remains in disagreement with ours by two orders of magnitude, thus there is inconsistency between KC’s forward and reverse rate constants and currently accepted thermodynamic data (see Appendix and ref.(19)). Whatever the cause of the discrepancies in k y values KC’s study, unlike the others in table 4, is free from possibility that the CH3C0 generated contained any internal photochemically- derived excess energy. The remaining differences, which we cannot account for, are still such that if our results are correct, previous workers should not have seen the purely thermal de- composition of acetyl at lower temperatures. If their results are correct, then acetyl should have predominantly decomposed our system. There is clearly room for further investigation of acetyl decomposition, particularly at room temperature. CH3CO + CH3 -+ CH3COCH3. We thank H. M. Frey and I. C. Vinall for useful discussion and R.A. Smith for help with programming.42 ADDITION OF HI TO KETENE APPENDIX CALCULATION OF THE RECOMBINATION RATE CONSTANT FOR ACETYL RADICALS This was done via k, = k*-/Keq for the reaction f r CH3COCOCH3 + 2 CH3CO where kf = 7.6 x giving k, = 108.41 cm3 mol-1 s-l. s-l at 730 K20 and Keq = 2.95 x moI dm-3 at 730 K (see below) Keq was estimated from the following thermodynamic data. compound AH,O/kJ mol-1* So/J K-1 mol-1* CHSCOCOCHS -329 (I 360 CHSCO -24.3 269 * At 298 K and 1 atm standard state ; a most recent estimate b obtained by bond additivity ;22 Cref. (23); dcalculated from structure and vibrational assignment used in this work for fall-off calculation (see also ref. (4)). Cg corrections were neglected in estimating Keq at 730 K. This calculation updates that of Benson and O’Nea1.24 H.E. O’Neal and S. W. Benson, J. Chem. Phys., 1962,36,2196. J. A. Kerr and J. G. Calvert, J. Phys. Chem., 1965, 69, 1022. J. A. Kerr and A. C. Lloyd, Trans. Faraday SOC., 1967, 63, 2480. H. M. Frey and I. C. Vinall, Int. J. Chem. Kinetics, 1973, 5, 523. P. J. Gorton and R. Walsh, Chem. Comm., 1973. R. Walsh, Trans. Faraday SOC., 1971, 67,2085. H. E. O’Neal and S. W. Benson, J. Chem. Phys., 1962,37,540. D. M. Golden and S . W. Benson, Chem. Rev., 1969,69,125. lo M. C. Flowers and S. W. Benson, J. Chem. Phys., 1963,38,882. S . C . Chan, B. S. Rabinovitch, J. T. Bryant, L. D. Spicer, T. Fujimoto, Y. N. Lin and S. P. Pavlou, J. Phys. Chem., 1970,74, 3160. l2 see, for example, P. J. Robinson and K. A. Holbrook, Unirnolecular Reactions (Wiley-Inter- science, New York, 1972), chap. 4. l 3 W. Forst, J. Phys. Chem., 1972, 76, 342. l4 D. S. Herr and W. A. Noyes, J. Arner. Chem. Soc., 1940,62,2052. l 5 J. J. Howland and W. A. Noyes, J. Amer. Chem. SOC., 1941,63, 3404 ; 1944,66,974. ‘A. D. Jenkins, J. Chem. SOC., 1952,2563. R. Hiatt and S. W. Benson, J. Amer. Chem. Soc., 1972,94,25 ; Int. J. Chem. Kinetics, 1972,4, 151. l7 P. D. Pacey and J. H. Purnell, Int. J. Chem. Kinetics, 1972,4, 657. l 8 W. Tsang, Int. J. Chem. Kinetics, 1970,2, 311 and references cited therein. l9 S . W. Benson, Thermochemical Kinetics (Wiley, New York, 1968), p. 204 ; H. E. O’Neal and 2o K. J. Hole and M. F. R. Mulcahy, J. Phys. Chem., 1969,73, 177. 21 S. W. Benson, F. R. Cruickshank, D. M. Golden, G. R. Haugen, H. E. O’Neal, A. S. Rodgers, 22 S. W. Benson and J. H. Buss, J. Chem. Phys., 1958,29,546. 23 J. A. Devore and H. E. O”ea1, J. Phys. Chem., 1969,73,2644. 24 S. W. Benson and H. E. O’Neal, Kinetic Data on Gas Phase Unirnolecular Reactions, NSRDS- S. W. Benson, Int. J. Chem. Kinetics, 1969, 1, 221. R. Shaw and R. Walsh, Chem. Rev., 1969,69,279. NBS 21 (U.S. Govt. Printing Office, Washington D.C., 1970), p. 424.

ISSN:0300-9599    

DOI:10.1039/F19747000033

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Gas-phase Unimolecular Pyrolyses of cis- and trans-2,3- Dimethyloxetan BY KENNETH A. HOLBROOK AND ROBERT A. SCOTT Department of Chemistry, University of Hull, Hull HU6 7RX Received 18th June, 1973 The gas phase pyrolyses of both cis and trans isomers of 2,3-dimethyloxetan have been studied at temperatures from 415483°C and at initial pressures from 2 to 32 Torr. Both isomers undergo two modes of decomposition producing either propene and acetaldehyde or but-2-ene (cis and trans) and formaldehyde. The cis-trans isomerisation of the starting material is relatively unimportant. The reactions all appear to be unimolecular processes in their first-order regions and the following rate expressions were obtained : cis isomer+C3H6+CH3CH0 : logl0(kl/s-l) = (15.70f0.22)-(63 216f621) K/4.576T cis isomer+C4H8 +CH20 : loglo(kZ/s-l) = (15.24f0.25)-(62 493 If: 715) K/4.576T trans isomer+C3H6+ CH3CH0 : l0glO(k3/s-’) = (15.91 f0.25)- (64 652f 664) K/4.576T trans isomer+C4H8+CHz0 : 1og10(k4/s-1) = (15.49&0.26)-(63 676f745) K/4.576T.Possible mechanisms involving 1 +biradical intermediates are discussed. The pyrolysis of oxetans is important since these compounds occur as intermediates in the oxidation of hydr0carbons.l Relatively few fundamental studies have been made of these pyrolyses apart from those of oxetan itself which decomposes primarily by a unimolecular mechanism to give ethylene and formaldehyde,2 and of 3,3- dimethyloxetan which similarly yields isobutene and f~rmaldehyde.~ It therefore seemed of interest to study the pyrolysis of an unsymmetrical oxetan in some detail.The pyrolysis of 2,3-dimethyloxetan was chosen, since this is capable of undergoing the following concurrent unimolecular reactions : CH3CHzCHZ + CHJCHO ’QH3 0 CH3 trans + CH20 This paper describes the pyrolysis of both cis- and trans-isomers of the starting material and the results enable the rate constants kl to k, to be derived. The cis-trans isomerisation of the reactant was negligibly slow under the chosen conditions. 4344 ? PYROLYSIS OF DIMETHYLOXETAN EXPERIMENTAL MATERIALS 2,3-Dimethyloxetan was prepared by ring-closure of the chloroester obtained from the reaction of acetyl chloride with 2-methylbutane-l,3-diol. The method differed from that of Searles et aL4 in that the ring closure involved slow addition of the chloroester (1 mol) to a rapidly stirred mixture of potassium hydroxide (2 mol) in ethylene glycol at 120°C and removal of the volatile oxetan product as it was formed. The isomers of 2,3-dimethyloxetan so obtained were later shown to be in the cis : trans ratio of 4.4 : 1.After distillation from sodium metal the cis and trans isomers of 2,3-dimethyloxetan were each obtained in ca. 95 % purity by distillation through a 1 m Nester-Faust spinning band column. The isomers were then separately purified by preparative g.1.c. employing a 6 mx 10 mm i.d. glass column packed with 20 % di-isodecyl phthalate on Celite at 90°C. (cis b.p. 86°C (760mmHg), trans b.p. 84°C (760 mmHg)). The cis and trans isomers were shown to be chromatographic- ally pure (2 99.4 %) on the following columns : 5 ft Porapak Q (SO/lOO) at 150"C, 5 ft 30 % squalane on Celite at 70°C and 2 m 10 % polypropylene glycol on Celite at 50°C.Infra-red and n.m.r. spectroscopic studies revealed no impurity. cis- and tvans-2,3-dimethyloxetan were unambiguously identified by proton and I3C n.m.r.5 Propene, nitric oxide, cis- and trans-but-2-ene were obtained from Matheson Co. Ltd., and had stated purities of at least 99 %. This was confirmed chromatographically. Acet- aldehyde (May and Baker Ltd.) was doubly distilled and dried before use. G.1.c. analysis showed no impurity. Formaldehyde was obtained from paraformaldehyde by the method of Spence and Wild and stored at liquid air temperature. APPARATUS A N D PROCEDURE The reaction vessels used were cylindrical Pyrex vessels of length 20 cm and internal radius 4 cm.The unpacked vessel had a surface/volume ratio = 0.94 cm-l, and the packed reaction vessel contained Pyrex tubes to give a surface/volume ratio of 10.6cm-l. A conventional high-vacuum static apparatus with greaseless stopcocks was used. The furnace temperature was controlled to +O.l"C and the maximum variation along the reaction vessel length was kO.75"C. The dead-volume was approximately 2 % of the reaction vessel volume and corrections were applied in the calculation of rate constants. At the end of a run, the reaction mixture was expanded down a heated line into a 1 cm3 g.1.c. sample loop placed within the column oven of a Pye 104 gas-liquid chromatograph. The columns used were 2 m 10 % polypropylene glycol on Chromosorb W at 50°C (for oxetan and acetaldehyde analysis) and 4 m 13 % bis-2-methoxyethyl adipate 7 % di-2- ethylhexyl sebacate on Chromosorb P at 40°C (for analysis of propene and butenes). All analyses were performed in triplicate.Analyses of calibration mixtures of reaction products were then accurate to +2 % of the amount of each component. Formaldehyde was analysed by the method of Rayner and Jephcott.' The formaldehyde was condensed into slightly acidified water and determined by a colorimetric procedure using acetone and Schiffs reagent. RESULTS At temperatures between 415 and 483°C the decompositions of cis- and trans-2,3- dimethyloxetan are adequately described by the concurrent reactions (1) to (4). This is confirmed by the equivalence of the pressure change and the amounts of propene + but-2-ene determined by g.1.c.starting from either isomer (table I). Further confir- mation is provided by the equivalence of the amounts of propene and acetaldehyde formed and by the amounts of but-Zene and formaldehyde formed (table 2). No products other than propene, but-2-ene, acetaldehyde and formaldehyde were de- tected. In some runs which were allowed to go to completion, the final pressure was found to be very close to twice the the initial pressure.K . A . HOLBROOK AND R. A . SCOTT 45 For each isomer the rate constants for the concurrent paths producing propene ( facetaldehyde) and but-2-ene (+formaldehyde) were obtained from overall rate constants determined from the first-order log plots and the ratios of propenelbut-2-ene determined chromatographically at 15-25 % decomposition.Thus for the cis isomer : kl = (propene)/(propene + but-2-ene)k' k2 = (but-2-ene)/(propene + but-2-ene)k' k3 = (propene)/(propene + but-2-ene)k" k4 = (but-2-ene)/(propene + but-2-ene)k" where k' and k" are overall rate constants determined from pressure change. The rate of reaction determined by pressure change was found to be first-order up to 30 % decomposition (fig. 1). The small amount of curvature after this extent of ,reaction is probably due to the occurrence of some cis-trans isomerisation of the reactant. In general, this was very slow under the conditions reported here and it was estimated and for the trans isomer : TABLE 1 .-COMPARISON OF AP AND (PROPENE+ BUT-2-ENE) FORMATION Po/Torr 16.03 18.16 16.05 16.30 16.59 8.73 14.83 32.50 6.37 13.79 :is-isomer at 4.66.0"C propene + but-2-eneI APITorr Torr 5.42 5.83 4.99 5.76 4.86 2.59 4.76 11.36 2.21 5.23 5.40 5.86 5.09 5.65 4.64 2.58 4.68 11.10 2.19* 5.177 1 Po/Torr 7.44 7.51 8.29 7.69 7.30 7.89 8 .oo 8.09 24.33 frans-isomer at 466.2"C APITorr 2.32 2.36 2.43 2.29 2.16 2.21 1.36 1.42 3.65 propene+ but-Zene/ Torr 2.25 2.26 2.41 2.20 2.19 2.16 1.30 1.35" 3.61 * 25 % NO added ; 9.8 % NO added.TABLE EQUIVALENCE OF PROPENE TO ACETALDEHYDE AND BUT-2-ENE TO FORMALDEHYDE cis-isomer at 466.0"C trans-isomer at 446.5OC Po/Torr C3H&H3CH0 Po/Torr C~HG/CH~CHO 32.50 1.03 4.22 1.03 18.54 1.03 4.25 1.02 16.08 1.02 4.20 1.03 14.83 1.02 4.06 1.03 8.73 1.02 2.24 1.03 6.37 1.01 2.39 1.03 cis-isomer at 457.6"C trans-isomer at 446.5"C Po/Tarr C4Hs/Torr CH20/Torr Po/Torr CH20IAP C4HsIA.P 32.31 3.73 3.52 2.24 0.42 0.427 18.70 2.46 2.32 4.25 0.41 0.435 18.56 2.36 2.22 4.06 0.42 0.43 8 18.53 2.99 2.96 2.33 0.4 1 0.432 8.53 1.90 1.85 4.13 0.42 0.43 146 PYROLYSIS OF DIMETHYLOSETAN that at 15 % decomposition the extent of cis to trans isomerisation is less than 2 %, and trans to cis isomerisation is less than 1 %.The first-order rate-constants obtained were independent of initial pressure in the range 2-32 Torr and were unaffected by the addition of up to 50 % of the radical-inhibitor nitric oxide. Virtually no change occurred in the values of the rate-constants on changing the surfacelvolume ratio of the vessel by a factor of over ten (table 3). 1.7 Y, Y .” I 5 z a +a $ 8 --.1.6 0 0 4 - 1.5 0 IOG 2 0 0 3 0 0 time/s FIG. 1.-First-order log plots. A, cis and B, trans isomcr. TABLE 3 .-EFFECT OF INITIAL PRESSURE, ADDED NITRIC OXIDE AND SURFACE/VOLUME RATIO cis-isomer 446.4 8 unpacked none 4.81 3.09 1.72 446.4 5 unpacked 14-50 % 4.84 3.1 1 1.73 466.0 lo* unpacked none 16.2 10.4 5.82 466.0 5 packed none 16.5 9.70 6.80 NO added 104 k’ls-1 104 /cl/s-l lo4 k2lS-l T/”C no. of runs vessel 466.0 4 unpacked 10-50 % 16.4 10.5 5.90 trans-isomer 7’PC no. of’ runs vessel NO addcd 1otk‘’/s 1 Io4k3/S-1 1O4k4/S 446.5 10-1 unpacked none 3.34 1.90 1 . 4 4 446.5 6 unpacked 13-50 ”/, 3.39 1.94 I .45 466.2 7 unpacked none 10.8 6.19 4.58 466.2 5 packed none 10.9 6.05 4.80 * Runs with initial pressures 2.4-32.5 Torr ; runs with initial pressures 2.0-24.0 Torr.K.A . HOLBROOK AND R. A. SCOTT 47 TABLE 4.-MEAN FIRST-ORDER RATE CONSTANTS FOR THE PYROLYSIS OF 2-3-DIMETHYLOXETAN rqT 483.2 475.6 473.3 466.0 457.6 446.4 435.1 418.8 T/"C 476.5 466.2 458.0 446.5 436.8 424.8 415.7 no. of runs 8 6 7 9 6 7 7 6 no. of runs 5 6 5 7 5 5 5 cis-isomer IOS k'ls-1 432 288 242 162 97.1 48.1 23.8 9.20 trans-isomer 105 k"/s-1 205 108 65.0 33.4 18.3 8.42 4.37 105 lills- 1 272 187 1 60 104 63.1 30.9 15.2 5.78 105 k31s-l 117 61.9 37.7 19.0 10.4 4.75 2.42 105 kz/s 161 .O 103 86.1 58.2 35.2 17.6 8.70 3.42 105 k41s-l 89.0 45.8 27.3 14.6 7.90 3.62 1.96 The effect of teinperature upon the rate constants is given in table 4. The rate constants were fitted to least squares Arrhenius plots with the result that the rate constants are expressed by the equations 10glo(kl/S-l) = (15.70*0.22)-(63 216+621) K/4.576T lOglo(k2/s-l) = (15.24f0.25)-(62 493 +715) K/4.576T lOglo(k3/s-l) = (15.91 kO.25)-(64 652+664) K/4.576T 10g10(k4/S-1) = (1 5.49 & 0.26) - (63 676 * 745) K/4.576T. Error limits quoted are the 95 % confidence limits.The product distribution, expressed by the rztio of propeiie to but-2-enc, was found to vary very little with temperature; this variation is reflected in the difference in activation energies between reactions (1) and (2) for the cis starting material and B 8 C 4 4 0 46c) 480 TIT FIG. 2.Variation of (cis-but-2-ene/tbut-2-ene) with temperature A, from cis-2,3-dimethyl- oxetan ; B, thermal equilibrium ratio ; C, from frans-2,3-dimethyloxetan.48 PYROLYSIS OF DIMETHYLOXETAN between reactions (3) and (4) for the trans starting material.The ratios of cis- to trans-but-2-ene obtained were also very little dependent on temperature as is shown in fig. 2. DISCUSSION From the data presented above, it is clearly established that both the cis and trans isomers of 2,3-dimethyloxetan undergo parallel ring-cleavage reactions to give either propene and acetaldehyde (reactions (1) and (3)) or but-2-ene and formaldehyde (reactions (2) and (4)). All of these reactions appear to be unimolecular, homogene- ous ht-order processes under the experimental conditions used and to be unaffected by additions of nitric oxide. and of 3,3-dimethyloxetan have shown that these also are mainly unimolecular processes and we have confirmed this for oxetan in a recent re-examination of the pyrolysis at low pressures.* Ring-cleavage of 2,3-dimethyloxetan can occur either by a concerted process involving the simultaneous rupture of two ring bonds or by rupture of a single ring bond initially to lead to the formation of an intermediate 1,4-biradical.The former process would be expected to produce complete retention of configuration for the product but-2-ene i.e., cis-2,3-dimethyloxetan would produce entirely cis-but-2-ene and trans-2,3-dimethyloxetan would produce entirely trans-but-2-ene. Our results (fig. 2) show partial retention of configuration i.e., cis-oxetan-+68 % cis-butene and trans-oxetan-76 % trans-butene and in neither case are these the equilibrium proportions of cis- and trans-b~t-2-em.~ It was independently shown that the rate of cis-trans isomerisation of but-2-ene is negligible under these conditions.These facts are consistent with the formation of a l74-biradical intermediate capable of existing for a few rotations before undergoing reaction to the cleavage products. The observation of some cis-trans isomerisation of the reactant, although this is slow compared with ring cleavage, is also evidence for the biradical mechanism. In addition, concerted mechanisms would be forbidden for the ring-cleavage reactions from considerations of conservation of orbital symmetry.1o The magnitude of the A-factors and the resulting calculated entropies of activation are consistent with processes involving ring opening via a biradical. It is instructive to compare the results for the pyrolysis of cis- and trans-2,3- dimethyloxetan with those for the pyrolysis of cis- and trans- 1,2-dimethylcyclobutane obtained by Gerberich and Waiters.'' Previous work on the pyrolyses of oxetan The comparable reactions for l,2-dimethylcyclobutane are : trans and the kinetic data are compared with those from this work in table 5.In both systems the cis-isomer decomposes faster than the trans-isomer which could be due to the release of stereochemical repulsion between the cis-methyl groups. The Arrhenius parameters for all eight reactions in this table are very similar, taking into account experimental errors, and the small differences between them for the 2,3-dimethyloxetan decomposition are not too significant.K. A . HOLBROOK AND R . A. SCOTT 49 1 TABLE 5 cis- ly2-dimethylcyc1obutane cis-2,3 -&met hyloxet an 60.4 15.5 (1) 63.2 15.7 15.2 reaction Elkcal mol-1 log10(A/s-9 reaction E/kcal mol-1 lol?lo(~/s-9 62.5 (5) (6) 63 .O 15.6 (2) ks/k6(43O0C) = +[propene]/[butene] = 5.2 kl/k2(4300C) = [propene]/[butene] = 1.7 trans- 1 ,Zdimet hylcyclo butane trans-2,3 -dime t hyloxe t an 61.6 15.4 (3) 64.7 15.9 15.5 63.7 (7) (8) 63.4 15.5 (4) k7/ks(43O0C) = *[propene]/[butene] = 3.5 k3/k4(4300C) = [propene]/[butene] = 1.3 A better comparison is obtained by considering the rate constant ratios at a given temperature for the alternative paths available to a given isomer.From table 5 it is seen that whereas both for dimethylcyclobutane and dimethyloxetan, the production of propene is always favoured over that of butene, the preference is less marked in the di methyloxe tan case.A possible explanation for this is obtained if the various 1,4-biradical inter- mediates which may be formed are considered in detail. From cis- 1 ,2-dimethylcyclobutane, three possible biradicals may be formed by C-C bond fission i.e., 4 3 CiS \ Evidence that the initial ring-cleavage is rate-determining is provided by the fact that decomposition is faster than the cis-trans isomerisation of the reactant which must imply that a biradical, if formed, has a lower energy path for decomposition .than for recyclisation. Of these biradicals, (c) which has both electrons located on secondary carbon atoms is the most favoured and the observed activation energies for the forma- tion of propene and butene of 60.4 and 63.0kcalmol-l respectively reflect the probable bond dissociation energies for C( 1)-C(2) and C(2)-C(3).(Bond dissocia- tion energies for the open-chain analogues 2,3-dimethylbutane and 2-methylbutane are 78 +2 and 80+2 kcal mol-1 respectively 12). From cis-2,3-dimethyloxetan, there are four possible biradicals, (d) and (e) from C-C fission and (f) and (9) from C-0 fission.50 PYROLYSIS OF DIMETHYLOXETAN CHj CH3 C3H6fCH3CH0 _.__) C,H,+CH,O -> C,H,+CH,O -> C3H6 + CHJCHO (9) Estimates based on group-additivity calculations and known heats of formation of alkoxy radicals l 2 show that in an unstrained ring the dissociation energies of the C-C and C-0 bonds in this compound are close to 80+ 1 kcal mol-l. It is therefore probable that all four biradicals (4-(g) should be considered and the 'observed smaller difference in activation energies between the reactions producing propene and butene in the case of the 2,3-&methyloxetan decomposition is reasonable. These arguments account qualitatively for the fact that (kl/k2) .c (k,/k,) for the cis-isomers and (k3/k4) < (k,/k,) for the trans-isomers as is shown in table 5. P. Barat, C. F. Cullis and R. T. Pollard, 13th Int. Combustion Symposium, Combustion Institute, 1971, p. 179. D. A. Bittker and W. D. Walters, J. Amer. Chem. Soc., 1955, 77, 1429. G. F. Cohw and W. D. Walters, J. Phys. Chem., 1967,71,2326. S . Searles Jr., K. A. Pollart and F. Block, J. Amer. Chem. Soc., 1957, 79, 952. D. F. Ewing, K. A. Holbrook and R. A. Scott, to be published. R. Spence and W. Wild, J. Chem. Soc., 1935, I, 338. A. C. Rayner and C. M. Jephcott, Anal. Chern., 1961, 33, 627. K. .4. Holbrook and R. A. Scott, to be published. J. L,. Holmes and L. S. M. Ruo, J. Chem. Soc. A , 1969, 1924. H. R. Gerberich and W. D. Walters, J. Amer. Chem. Soc., 1961,83,3935, 4884. J. A. Kerr, Chem. Rev., 1966, 465. lo R. Hoffmann and R. B. Woodward, Angew. Chem. In?. Edn., 1969,8,781.

ISSN:0300-9599    

DOI:10.1039/F19747000043

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Primary Yields in the r-Radiolysis of Gaseous Aliphatic Amines BY MOHAMMAD A. SAMI AND DONALD SMITHIES * Department of Physical Chemistry, University of Leeds, Leeds LS2 9JT Received 22nd June, 1973 Yields of Hz and CH4 from methyl-, dimethyl-, trimethyl- and ethyl-amine vapours y-irradiated at room temperature have been measured. The effects of dose, pressure, electron and free radical scavengers on these yields have been studied. From the decreases in G(H2) in the presence of scavengers values for G H ~ , Ge, GH and relative rate constants for some reactions of H atoms have been calculated. Except for methylamine the values of G, obtained in this way are lower than those obtained from measurements of the energy of ion pair formation, W. Ethylenediamine is a product from radiolysis of methylamhe but dimethylhydrazine is not.Following investigations of the radiation chemistry of ammonia 1* and liquid methylamine a study of the radiolysis of some aliphatic amines in the gas phase has been made. EXPERIMENTAL MATERIALS Methylamine, dimethylamine, trimethylamine and ethylamine (Matheson Co.) were purified by contact with sodium hydroxide pellets and then sodium in a cardice-acetone bath until a blue solution was obtained. Ethylene and propene (Phillips research grade), isobutene, butadiene, sulphur hexafluoride and perfluorocylobutane (Matheson Co.) and nitrous oxide (British Oxygen Co.) were purified by condensing at 77 K and pumping. A middle fraction was taken from the residue. Carbon tetrachloride (May and Baker, A.R.) was degassed and fractionated on the vacuum line.IRRADIATION PROCEDURE The irradiation cells were spherical Pyrex bulbs of volume about 1 dm3. A 6 o C ~ pray source of about 7 kCi was used. Dose rates were measured using the nitrous oxide dosimeter for which G(N2) = 10.0 was assumed. The dose rates in the amines were calm- lated by multiplying by the relative stopping powers SN~O : & e ~ ~ z : SMeZNH : SMeSN : SE~NH~ = 1 : 0.869 : 1.256 : 1.643 : 1.256. ANALYSES Hydrogen, nitrogen and methane were determined by vapour-phase chromatography. The same technique was used in an attempt to identify organic products from radiolysis of methylamine. A 1.5 mx 3 mm column of Poropak Q coated with 20 % w/w polyethylen- imine was used with nitrogen as the carrier gas and flame ionization detection. RESULTS Hydrogen and methane are produced when methylamine, dimethylamine, tri- methylamine or ethylamine is irradiated.Ethane, G < 0.02, was not detected on irradiation of any of these amines. No attempt was made to detect ammonia. 5152 RADIOLYSIS OF AMINES Vapour phase chromatography indicated that ethylenediamine was a product from radiolysis of methylamine ; NN'-dimethylhydrazine was not detected. An un- identified compound with a retention time much longer than either of these compounds or methylamine or dimethylamine was also formed. HYDROGEN YIELDS The values of G(H,) from y-irradiated gaseous mines at 25°C are shown in table 1. These yields are independent of dose in the range 4-9 x lo1* eV g-l and of the pressure in the range (1.3-9.3) x lo4 N m-2 (100-700 mmHg).The agreement between our data and those of others is good for methylamine and dimethylamine but less good for trimethylamine. TABLE HYDROGEN YIELDS FROM 7-IRRADIATED AMINES IN THE GAS PHASE AT 25°C amine addi tivc conccntration/mol % G(Hd AG(H2) = Ge 1001 w 25 MeNH2 none 12.5 3- 0.2 13.1 a SF6 0.6-4.5 8.2f0.1 4.3k0.2 CCI, 1 .O-7.5 8.5f0.2 4.0& 0.3 cycl o - C4F8 0.2-2.0 8.3k0.2 4.2f 0.3 mean 4.2+ 0.3 4.23f 0.07 Me,NH none 11.5+0.5 12.1 a SF6 0.2-0.7 8.1k0.2 3.4k0.5 cc14 0.3-2.3 7.7f0.1 3.8f0.5 mean 3.6k0.5 4.37f0.10 Me3N none 9.5k0.4 7.8 a SF6 0.5-0.9 5.5h0.1 3.9k0.4 cc14 1.6-5.0 5.8+0.1 3.6k0.4 mean 3.8A0.4 4.39k0.08 EtNHz none 9.5k0.2 SF6 0.5-1.1 6.0+ 0.2 3.5k0.3 CCl, 0.2-2.2 6.1k0.1 3.4f 0.2 cyclo-C,F, 0.5-2.2 6.1k0.2 3.410.3 Calculated taking W/eV (ion pair)-' = 23.6 for MeNH2, 22.9 for MezNH and 22.8 mean 3.4k0.3 3.98k0.06 a ref.(6). for Me3N. HYDROGEN YIELDS I N THE PRESENCE OF ADDITIVES The technique of measuring changes in the yield of a product of radiolysis in the presence of an additive which reacts efficiently with one or more reactive species (electrons, ions, radicals) and using this change and an assumed mechanism to calculate yields of the reactive species is well known in radiation chemistry. Values of G(H,) in the presence of some well-known electron scavengers are shown in table 1. In each case G(H,) is independent of concentration of additive in the range indicated. Yields are calculated on the basis of energy absorbed by the amine. The effects of isobutene and butadiene on G(H,) from methylamine are shown in fig.1 in the form of" competition plots '' (AG(H,)-l against [amine]/[olefin] where AG(H2) = G(H,)"-G(H,) and G(H,)" is the hydrogen yield in pure amine). Slopes andM. A . SAM1 AND D. SMITHIES 53 intercepts were calculated using a weighted least-mean squares procedure. Similar linear plots were obtained with ethylene and propylene and with other amines. These data are discussed later. [MeNH2] /[olefin] 0 12.5 2 5 0 375 5 0 0 0 . 5 0 . 4 7 0.3 n s $ 0 . 2 0. I 0 0 5 00 1000 1 5 0 0 2 0 0 0 [MeNH2]/[olefin] FIG. 1 .-Effects of isobutene, butadiene and butadiene + SF6 on G(H,) fromyirradiated methylamine. 0, C ~ H S ; 0, C4H6 ;A, C4H6+ 1.6 mol % SF6. Data for isobutene are displaced upwards by 0.1 unit.2.0 // ~ ~~ - 0 I 2 3 4 5 6 7 amine pressure/102 mmHg A, Me3N ; 0, EtNH2 ; filled points show data from ref. (6). FIG. 2.-Dependence of G(CH4) from y-irradiated amines on pressure. 0, MeNH2 ; 0, Me2NH ; TABLE 2.-EFFECT OF ADDITIVES ON G(CH4) FROM ETHYLAMINE additive 102 [addit ive]/[ethylamine] G(CH4) none 2.42+ 0.30 SF6 4.65 2.45+ 0.04 SF6 9.55 2.66f 0.04 C2H4 1.86 1.93 + 0.04 C2H4 3.70 1.60f0.03 Amine pressure 700 mmHg.54 RADIOLYSIS OF AMINES METHANE YIELDS The dependence of G(CH4) on pressure of amine is shown in fig. 2. In all cases G(CH,) is independent of dose up to at least 1 x 1019 eV g-l. There is fair agreement between our data and those of others.6 The effects of sulphur hexafluoride and of ethylene on G(CH4) from ethylamine are shown in table 2.DISCUSSION Ions, electrons and excited molecules are primary products of the action of radia- tion on the mines. Mass spectrometric studies indicate that the main ions produced by electron impact on methylamine at pressures around mmHg are CH3NHz, CH2NH;, and CH2N+. These ions all react rapidly (k N 6 x 1011 dm3 mol-l s-l) with methylamine 7-9 to form CH3NH'; and at a pressure of 0.3 mmHg this is virtually the only ion detected. CH2NH2 or CH3NH are produced in reaction (1) with about equal probability.1° Similar, although slightly slower, reactions of the main ions in the mass spectra of dimethylamine and trimethylamine to give amine- H+ ions have been observed except that there was no evidence for reaction of the Me2NCHz ion.' There is no evidence of formation of negative ions.The most likely neutralization reactions are 4 and 5. CH3NH: + CH3NH2 + CH3NHi + CH2NH2 (CH3NH) CH2NH; + CH3NH2 -+ CH3NHi + CH3N CH2N+ + CH3NH2 -+ CH3NHi + HCN CH3NH: +e -+ CH3NH2 +H CH3NH'; + e + CH4N + H2 (1) (2) (3) (4) ( 5 ) Studies of the photochemistry of methylamine vapour 11* l 2 indicate primary processes (6)-(9) may occur. CH3NHz + H+CH3NH (6) 4 H + CH2NH2 (7) -+ CH3 +NH, (8) + H2 +CH3N (9) The free radicals formed either from excited molecules or in ion-molecule reactions may react with methylamine, or with one another, (10)-(17). H + CH3NH2 -+ H2 + CH2NH2 (CH3NH) NH2 + CH3NH2 -+ NH3 + CHzNH2 (CH3NH) (CH3NW CH3 + CH3NH2 4 CH4 + CHzNHz CH3NH + CH3NH2 -+ CH3NH2 + CHzNH2 2CH2NH2 + (CH2NH2)2 2CH3NH 4 (CHSNH), 2CH3 -+ C2H6 2NH2 -+ N2H4 Reaction (10) is consistent with the reduction of G(H2) by known H atom scavengers.M.A . SAM1 AND D. SMITHIES 55 Abstraction of H from both the methyl and the amino groups is expected. On the basis of known rate constants for reactions of methyl radicals l 3 - I 5 (12) is expected to occur to the exclusion of (16) under the conditions used in this work. This is consis- tent with methane but not ethane being a product. Production of ethylenediamine is consistent with the occurrence of (14). The absence of dimethylhydrazine suggests that CH3NH radicals react via (13) rather than (15). When methylamine glass was irradiated at 77 K an e.s.r. spectrum attributable to CH3NH was observed l6 but this was replaced by a spectrum attributable to CH2NH2 as the temperature was raised.Oxidising radicals e.g. OH,17 CH3018 are known to react with SCN- to form (SCN); . When solutions of potassium thiocyanate in methylamine were pulse irradiated the characteristic spectrum of (SCN), with Amax at 490 nm was observed. By assuming that reactions (13), (18) and (19) occur and that l9 &490 = 760 rnol-' as in aqueous solution GCH3NH = 2.9k0.3 and kl8/kl3 = 27rfi3 in liquid methyl- amine at 25°C were obtained. CH3NH + SCN- + CH3NH-+ SCN SCN+SCN- -+ (SCN); For the other amines even less is known about the reactions of the ions, excited molecules and free radicals than is known for methylamine. However, all the data obtained in this work can be understood on the basis of reactions analogous to those written above for methylamine. YIELDS OF PRIMARY PRODUCTS On the basis of the proposed mechanism the hydrogen yield in a pure amine G(H2)" is given by G(H2)" = GH2 + GH+ G,.In the presence of an electron scavenger which eliminates reactions (4) and ( 5 ) and gives an anion which does not yield H or H2 on neutralization G(H2) = GH2+GH. Values of G, calculated from AG(H2) are shown in table 1. In the presence of an olefin H atoms may add to or abstract H from the olefin, (20) and (21), as well as abstract H from the amine, (10). H+ol+ Hol (20) H + 01 + product + H2 (21) On this basis and assuming that all electrons are converted to H atoms on neutraliza- tion G(H2) = GH2+(GH+Ge)(flo+f21) wheref,, andfil are the fractions of H atoms reacting via (10) and (21) respectively. Hence AG(H2)-' = (G(H2)'-G(H2))-' = (C,+G,)-'( where [MI is the concentration of amine.Hence any of the " competition plots " has intercept = (GH+G,)-l(l +kzl/kzo) and slope = (GH+G,)-' k10/k20. Values of k2,/kZ0 were taken 2o as zero for ethylene, 0.063 for propene, 0.025 for isobutene and 0.024 for butadiene. These values and the data from the competition plots were used to obtain the values of GH+G,, GH2 = G(H2)'-(GH+Ge) and k1,/k2, shown in table 3. From the known values of G(HJ in radiolysis of the olefins 21-24 it can be shown that any contribution to G(H2) in this work from direct action on the olefin is within the limits of experimental error. There is no clear trend in primary yields with change in structure of the amines. While our data for methyl- amine agree well with those of Takamuku and Sakurai6 agreement is poor for trimethylamine.While the value of G, determined by measuring the change in G(H,) in the presence of an electron scavenger agrees well with that from measurement56 RADIOLYSIS OF AMINES of the energy required to form an ion pair, W, in the case of methylamine poor agree- ment is found for the other amines. This may be understood if in the amines other than methylamine a neutralisation reaction occurs which does not lead to formation of H or Ha, e.g. (22) RNHS+e -+ R+NH3. (22) TABLE 3.-G~z, GH+G, AND RELATIVE RATE CONSTANTS FOR REACTIONS OF H ATOMS FROM 7-IRRADIATION OF AMINES IN THE PRESENCE OF OLEFINS amine olefin GH+ Ge MeNH2 C2H4 10.1 k 0.3 C3H6 10.2+ 0.1 iso-C4H8 10.8 * 0.1 butadiene 10.4+0.1 mean 10.45 + 0.25 butadiene+ SFs 5.88k0.06 10.6 GfIz k2 olk 1 o 2.4+ 0.3 99+ 8 2.3 + 0.2 130+ 10 1.7k 0.2 283k 5 2.1k0.2 2.2+ 0.2 2.3k0.1 450+ 30 2.5 Me2NH iSO-C& 9.5+ 0.1 2.0+ 0.5 60*2 butadiene 10.2+ 0.1 1.3 + 0.5 mean 9.85$- 0.35 1.7+ 0.5 Me3N iso-C4H, 8.6k0.8 0.9k 0.9 30+ 4 4.6 3.2 EtNH2 iSO-C4& 8.7k0.1 0.8k0.2 37k 1 butadiene 8.95k0.23 0.6k0.3 mean 8.7k0.1 0.7k0.3 a GH ; b ref.(6). METHANE YIELDS In all the amines studied methane was a product and ethane was not. In ethyl- amine G(CH4) was not reduced by addition of an electron scavenger, SF6, but was reduced by a radical scavenger, ethylene (table 2). This is consistent with methane being formed via CH, radicals (e.g. reaction (12)) and not as a result of a neutralisation reaction. There are insufficient data to enable calculation of GCHs and the ratio of rate constants k(CH, + C,H,)/K(CH, + MeNH,).The absence of ethane in radio- lysis of ethylamine indicates that C-C rather than C-N bond fission occurs. Combination of two methyl radicals, (16), cannot occur in any of the amines studied since ethane is never a product. This is as expected at the intensities used from the known rate constants for reactions of methyl radicals with amines.13-15 The decrease in G(CH4) at low pressures is consistent with reaction between methyl and another radical rather than with amine; the second radical must be other than methyl or H since no ethane is formed and G(H,) is independent of pressure. RATE CONSTANTS Relative rate constants at 298 K for abstraction of H from amines by H atoms and for H atom addition to olefins are shown in table 3.Data on rate constants for H atom reactions have been reviewed and assessed recently.26 Our data for addition of H to olefins agree with the recommended values ; rate constants for reactions of H atoms with amines have not been reported previously. This is consistent with all the scavengeable hydrogen arising via reactions of thermal rather than hot H atoms.M . A . SAM1 A N D D. SMITHIES 57 Apparently different values for k(H + butadiene)/k(H + MeNH,) found in the presence and in the absence of SF6 can be interpreted in terms of reaction of electrons with butadiene; it is known that the reaction of aquated electrons with butadiene is fast. F. S. Dainton, T. Skwarski, D. Smithies and E. Wezranowski, Trans. Faraday SOC., 1964, 60, 1068.J. A. Eyre and D. Smithies, Trans. Faraday Soc., 1970,66,2199. D. Smithies and A. J. Whitworth, J. Chem. SOC. A., 1969, 1987. D. W. Huyton and T. W. Woodward, Rad. Res. Rev., 1970, 2,205. P. W. Jones and H. D. Gasser, J. Chem. SOC. B., 1971, 1873. S. Takamuku and H. Sakurai, Bull. Chem. SOC. Japan, 1965,38,791. E. G. Jones and A. G. Harrison, Canad. J. Chem., 1967,45,3119. T. J. Zielinska and H. Wincel, Nukleonika, 1970, 15, 343. ' M. S. B. Munson, J. Phys. Chem., 1966,70,2034. lo C. J. Ogle, Ph.D. Thesis (University of Leeds, 1968). l 1 J. V. Michael and W. A. Noyes, jun., J. Amer. Chem. SOC., 1963, 85, 1228. l2 J. J. Magenheimer, R. E. Varnerin and R. B. Timmons, J. Phys. Chem., 1969,73,3904. l 3 P. Gray and J. C. J. Thynne, Trans. Faraday SOC., 1963,59,2275. l4 P. Gray, A. Jones and J. C. J. Thynne, Trans. Faraday SOC., 1965,61,474. l5 P. Gray and A. Jones, Trans. Faraday SOC., 1966, 62, 112. l6 P. Wardman and D. R. Smith, Canad. J. Chem., 1971,49,1869. '' G. E. Adams, J. W. Boag and B. D. Michael, Trans. Faraday SOC., 1965,61,1674. l8 D. H. Ellison, G. A. Salmon and F. Wilkinson, Proc, Roy. SOC. A., 1972,328,23. l9 J. H. Baxendale, P. L. T. Bevan and D. A. Stott, Trans. Faraday SOC., 1968,64,2389. 2o G. R. Wooley and R. J. Cvetanovic, J. Chem. Phys., 1969,50,4697. 21 R. A. Back, T. W. Woodward and K. A. McLauchlan, Canad. J. Chem., 1962,40,1380. 22 F. W. Lampe, Rad. Res., 1959,10, 691. 23 G. J. Collin and J. A. Herman, Canad. J. Chem., 1967, 45, 3097. 24 K. Hamanoue, J. Okamoto, 0. Tokunaga and A. Danno, Bull. Chem. SOC. Japan, 1972, 45, 2 5 R. H. Leblanc and J. A. Herman, J. Chim. phys., 1966,63,1055. 26 J. A. Kerr and M. J. Parsonage, Evaluated Kinetic Data on Gas Phase Addition Reactions. Reactions of Atoms and Radicals with Alkene, Alkyne and Aromatic Compounds (Butter- worth, London, 1972), p. 357. 1306.

ISSN:0300-9599    

DOI:10.1039/F19747000051

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Photoconductivity and Crystal Structure of Organic Molecular Complexes BY VERA M. VINCENT AND JOHN D. WRIGHT* University Chemical Laboratory, University of Kent at Canterbury, Canterbury, Kent CT2 7NH Received 25th April, 1973 The steady state semiconductivity and photoconductivity of a series of single crystals of 24 niolecuIar complexes, in which the electron donor (D) and acceptor (A) molecules were varied systematically, has been studied. Resistivity trends, semiconduction activation energies and Seebeck coefficients suggest extrinsic semiconduction due to A- ions. Spectral response of photoconduction is little affected by impurities, and is discussed in terms of crystal electronic states and interaction between charge transfer and D or A singlet excitations. The magnitude of photoconduction varies markedly within the series.Independent parameters relating to photoconduction quantum yield are derived from a consideration of the steady state conditions and from measured photoconduction activation energies. Both parameters show the same trends, which are interpreted in terms of two structural requirements for efficient photoconduction : (a) a-orbitals of adjacent D or A molecules (but not both) should overlap appreciably, and (b) the orientation of D with respect to A should not be ideal for overlap of orbitals involved in the charge transfer transition leading to photoconduction. These requirements may be used for prediction of photoconduction efficiency in molecular complexes. The photoconductivity of a limited number of molecular complexes has been studied by several workers 1-3 but the interpretation of the observed properties has been hindered by the absence of comparable data for a series of complexes in which both the electron donor and acceptor are varied systematically.In particular, the factors determining the efficiency of photoconduction in these materials have not been investigated. Furthermore, crystal structures are now known for many complexes of this type,4* and the influence of structure on photoconductivity may therefore be investigated. We present here the results of an extensive study of photoconductivity in 135 single crystals of a series of 24 complexes, under similar experimental conditions, with the object of establishing the role of various factors, particularly crystal structure, in photoconductivity.EXPERIMENTAL The electron donors (durene, hexamethylbenzene (HMB), naphthalene, acenaphthene, anthracene, chrysene, pyrene, coronene and tetrabenzonaphthalene (TBN)) and acceptors (7,7,8,8-tetracyanoquinodimethane (TCNQ), tetracyanoethylene (TCNE), 1,2,4,5-tetra- cyanobenzene (TCNB), chloranil and fluoranil) were purified by the methods indicated in table 1. Crystals of the complexes were obtained by slow evaporation at room temperature of solutions of the mixed components in a suitable solvent. Solvents used included chloro- form, acetone, ether, 2-methoxyethanol and glacial acetic acid. All solvents were either A.R. grade or carefully dried and distilled. The crystals were generally good quality needles or plates, of millimetre dimensions, though some donor-acceptor combinations did not produce crystalline complexes, either because of reaction between the components (e.g.anthracene- TCNE) or as a result of unfavourable solubility relationships in the solvents used (e.g. naphthalene-TCNQ). Elemental analysis confirmed that all the complexes were of 1 : 1 stoichiometry, except for the HMB+ TCNE system, where crystals grown from ether were of 58v . M . VINCENT AND J . D. WRIGHT 59 1 : 1 stoichiometry while those from chloroform were 2 : 1. (The 2 : 1 complex was used in conductivity measurements in this case). For conductivity measurements, a d.c. two-probe technique was used. No attempt was made to use a guard ring on the small crystals studied, but the resistivity and activation energy of semiconduction showed no dependence on sample dimensions when these were varied over an order of magnitude, as expected for bulk conduction.Unless otherwise stated, the voltage was applied along the needle axis of needle shaped crystals. Except for the HMB-TCNE complex, no chemical reaction was observed at the interface between the crystal and the dispersion of silver in methyl isobutyl ketone, used to mount the crystal between the copper electrodes. The HMl3-TCNE crystals discoloured slightly at this interface, but this did not lead to any obvious effects on the conductivity. Measurements were generally carried out with the sample in uacuo, although in dry air the photoconductivity was unchanged and the semiconductivity increased marginally.Complexes involving the components acenaphthene, TABLE 1 .-PREPARATION DETAILS AND MEAN SPECIFIC RESISTIVITIES (IN R m) (WITH OBSERVED RANGE IN PARENTHESES) TCNQ a E.A. = 2.88 eV HMB A i.p. = 7.80 eV 2 . 4 ~ lo8 (0.2-6.8) acenapthene B 7.65 eV anthracene d 7.39 eV chrysene c* 7.75 eV pyrene c* 7.47 eV perylene 7.15 eV coronene 7.64 eV TBN 7.58 eV 9 . 4 ~ 107 (5-15) A 5.3x lo8 (0.4- 14) A (1.9-4.7) (0.4-3.2) A 3 . 6 ~ lo8 (0.4- 12) A (0.4-10) (c) 2 . 9 ~ 104 ( d ) 1 . 5 ~ lo6 (i) 2 . 7 ~ lo1' (ii) 1 . 4 ~ lo9 (0.35-6.4) A (1 -2.9) A 1.9x lo6 (1.6-2.2) 2.1 x 107 TCNE chloranit fluoranil 2.89 eV 2.59 eV 2.48 eV A, c A A 2 . 4 ~ 10'' 4 . 4 ~ 10" 1 . 3 ~ 1013 ( 1 . 0 4 ) (1.1-9.9) (0.6-3.2) B 1.8x 10l2 (1-2.8) A 1.3 x 10'' (0.14-3.4) A A A 1 . 0 ~ 109 1 . 6 ~ 1013 3 .2 ~ 1011 (0.2-1 .S) (0.9-2.7) A A A 1 . 6 ~ 10 7 . 7 ~ 10" 5 . 8 ~ 1011 (0.3-4.3) (0.04-22) (2.2-13) TCNB 2.30 eV D (0.1-3.7) 1 . 6 ~ 1013 D 3.8 x 1OI2 (0.4- 10) Purification of starting materials : * twice entrainer sublimed in Pyrex tube under argon (667 Pa pressure) ; b sublimed under a low pressure of helium ; C commercial product used directly ; * recrystallised, chromatographed and zone refined ; e recrystallised, twice chromatographed. Solvents : A, chloroform ; B, acetone ; C, ether ; D, glacial acetic acid. (For perylene-TCNQ- (i) and (ii) refer to measurements perpendicular and paralle 'to the needle axis of the crystals, respect- ively).60 PHOTOCONDUCTIVITY OF MOLECULAR COMPLEXES TCNE or fluoranil tended to decompose in vacuum by loss of these components.This was prevented successfully by carrying out measurements in dry air at atmospheric pressure, saturated with the vapour of the component in question. The apparatus used was that described in a previous paper.6 Several crystals of each complex were studied, and room temperature resistivity, current against voltage characteristics, spectral response of photo- current at constant incident light intensity (approximately 10l8 photon m-2 &), and depend- ence of photocurrent on light intensity at selected wavelengths were measured. In addition, semiconduction and photoconduction activation energies, dependence of photocurrent on direction of polarisation of incident light, and Seebeck coefficients were measured for selected complexes. The Seebeck coefficient determinations were made with the cold electrode at room temperature, with temperature differences (AT) of up to 70 K.The observed voltage depended linearly on AT in this range. Visible spectra of the solid complexes were measured by the diffuse reflectance te~hnique,~ using samples diluted by grinding with KBr, on a Pye Unicam SP500 spectrophotometer, with SP540 diffuse reflectance attachment. Crystal structure projections were calculated using a program (MOLAX) kindly made available by Dr. C. K. Prout of Oxford University, utilising subroutines from the NAGF library of the Nottingham University Algorithms Group, modified for the University of Kent ICL 41 30 computer. RESULTS AND DISCUSSION SEMICONDUCTIVITY The majority of the crystals studied exhibited ohmic behaviour over the range of applied fields used (< lo5 V m-I).The only exceptions were chrysene-TCNQ, where linear plots of log(app1ied voltage) against log(current) had slopes varying from 1.1 to 1.4, and perylene-chloranil, with slopes of 0.9 to 1.8. For HMB- and acenaphthene- chloranil, HMB-, chrysene- and pyrene-fluoranil, durene- and pyrene-TCNB and HMB-TCNE, the combination of high resistivity and small sample size made deter- mination of current against voltage curves impossible. Table 1 lists the room temperature resistivities of the complexes studied, obtained as averages from several crystals of each complex. Resistivity depends more strongly on acceptor electron affinity than on donor ionisation potential, with high electron affinity corresponding to lower resistivity.The temperature dependence of conductivity was studied for several of the complexes, and semiconduction activation energies ( E ) were determined using the standard expression CJ = oo exp( - E/kT). These results, together with those of other workers,l* * * *-Io are collected in table 2. The activation energies are all small, and those for TCNQ complexes are close to the values of 0.42-0.45, and 0.56-0.63 eV found for surface and bulk extrinsic conduction in TCNQ itself.6 This, with the resistivity data previously mentioned, suggests extrinsic semiconduction in the complexes, possibly due to impurities of the type M+ acceptor-, where M is a metal such as sodium. Such impurities are known to occur in TCNQ itself,6 and their formation would be most favoured with acceptors of high electron affinity, consistent with observed resistivity trends.With such impurities, the sign of the majority charge carrier should be negative. Seebeck coefficients were determined for a range of the FIG. 1 .-Photoconduction response curves (with diffuse reflectance spectra - - -). (For pyrene- TCNQ, solid and broken lines refer to the complex prepared from pyrene e and C, respectively, see table 1. For perylene-TCNQ, solid and broken lines refer to needle and plate crystals, respectively. The yield of the TBN-TCNQ complex was too small for determination of its reflectance spectrum), (a) pyrene-TCNQ, (6) chrysene-TCNQ, (c) perylene-TCNQ, (d) TBN-TCNQ, (e) HMB-TCNQ, (f) anthracene-TCNQ, (9) perylene-TCNE, (h) pyrene-TCNE, (i) perylene-chloranil, ( j ) pyrene- chloranil, (k) perylene-fluoranil, (1) chrysene-fluoranil.V .M. VINCENT AND J . D . WRIGHT 61 _ * _ _ _ - - _ _ - - - - - - _ 4 0 0 500 600 700 803 900 1033 400 500 600 700 aoo 900 1000 w avelengt h/nm FIG. 1 .-See caption opposite62 PHOTOCONDUCTIVITY OF MOLECULAR COMPLEXES complexes in order to check this, and the values, given in table 3 confirm that the majority carrier is negative as required by the model. Recent e.s.r. studies on the carbazole complexes of TCNQ and 2,3-dichloro-5,6-dicyano- 1,4-benzoquinone support this interpretation. TABLE 2.-sEMICONDUCTION ACTIVATION ENERGIES ( E ) complex EIeV complex EIeV acenaphthene-TCNQ 0.45 HMB-TCNE 0.58 lo anthracene-TCNQ 0.75,8 0.49 p yrene-TCNE 0.82 ' 1 l o chrysene-TCNQ 0.39 per y lene-TCNE 0.72 lo pyrene-TCNQ 0.73,2 0.59 pyrene-chloranil 0.73 peryleneTCNQ 0.64,8 0.46 perylene-chloranil 0.73 coronene-TCNQ 0.60 perylene-fluoranil 0.73 TABLE 3.-sEEBECK COEFFICIENTS (mv K-') HMB-TCNQ - 1.2 chrysene-TCNQ - 0.5 acenap ht hene-TCNQ - 1.6 perylene-chloranil - 0.3 ant hracene-TCNQ -0.5 perylene-fluoranil - 1.4 perylene-TCNQ - 0.0(4) p yrene-TCNQ - 1.2 PHOTOCONDUCTIVITY The spectral response of photoconduction, for all complexes whose photocon- duction was large enough for this to be measured, is shown in fig. 1, together with the diffuse reflectance spectra of the complexes.(The photocurrent scale in these dia- grams is not identical for all crystals of a given complex, and has been selected to permit comparison of the shapes of the curves rather than relative magnitudes of photocurrents.) The main features of the curves are generally similar for all crystals of a given complex, although there are minor variations in detail.This is in marked contrast to the behaviour of crystals of the electron acceptor TCNQ,6 and suggests either the absence of impurity effects, or similar impurity concentrations in all crystals of a given complex. The latter is unlikely, since the crystals studied were frequently from several different batches, using starting materials of differing purity. The following three observations show that the onset of recombinztion of charge carriers in regions of strong optical absorption is not controlling the shapes of the curves. First, the photoconduction response curves for some complexes (e.g.pyrene-TCNQ) rescmble the diffuse reflectance spectra. Second, the dependence of photocurrent on light intensity shows no sharp variations at different wavelengths across the response region, whereas such variations would be expected for a sudden onset of carrier recombination. Third, the shapes of the spectral response curves for photoconduction in the pyrene and perylene complexes of TCNQ are not affected when the direction of polarisation of the incident light is changed from parallel to perpendicular to the needle axis of the crystal. Since the charge transfer bands of the complex and the singlet transitions of the component molecules are strongly polarised, this change affects the penetration depth of the light into the sample, and.hence the density of excited states or excitons. Irrespective of the mechaiiism for charge carrier generation from excited state species, this changes the charge cai-rier density and hence the rate of carrier recombination. As the photoconduction response curves are not changed under these conditions, recombination cannot be responsible for the observed shapes. The curves are of two types, those which closely follow the absorption spectrumV. M. VINCENT AND J . D. WRIGHT 63 of the complex (e.g. pyrene-TCNQ) and those showing maxima in a region between the charge transfer band and the singlet transitions of the component molecules. In only one case (anthracene-TCNB) was any photocond.uction observed on the low energy side of the charge transfer absorption maximum.Previous workers have reported a small response in this low energy region for similar complexes, but this diminished when well-developed crystals were used. Its absence in our samples may therefore be an indication that they are relatively free from defects. The components of some of the complexes are also known to photoconduct in the region corresponding to the high energy side of the charge transfer absorption maximum (e.g. TCNQ,6 perylene 12). The observed photoconduction of the com- plexes in this region is not simply due to the components, however, since : (a) the wavelength of maximum photoconduction in the perylene complex varies as the acceptor molecule is varied, and corresponds to lower energies than that of the lowest energy photoconduction maximum in pure perylene ; (b) the photoconduction of perylene in this region is strongly dependent on the direction of polarisation of the incident light with respect to the crystal axes, whereas this is not so for the perylene- TCNQ complex ; (c) the photoconduction of TCNQ is strongly dependent on sample purity, whereas this is not so for the TCNQ complexes.Peak photoconduction can occur on the high energy side of the charge transfer absorption maximum if the conduction state is of higher energy than the charge transfer excited state, and/or the process of separation of the charges in the excited state ion pair to give charge carriers has a high activation energy. The energy gap A between the separated charge carrier state and the charge transfer excited state is determined by the parameters outlined in fig. 2.A = E-hv,, (1) E = i.p.-e.a.-P,+-PA,+AWf (2) (3) (4) hv,, = i.p. -e.a. - A , + R, - E + Re,,- PDfA- +AW, ... A = A2-R2 + E - Re,, +(P,+A- -P,+ -PA-). Hence A is determined by the balance between a number of opposing factors, and may have a very small magnitude in favourable cases, although prediction of its value is difficult on account of the number of factors involved. However, A is not the only factor determining the wavelength of maximum photoconduction, as the hopping process by which charges move through a crystal of this type requires activationenergy, and this must be provided, irrespective of the value of A, if carrier generation is to occur. In cases where both of these energy requirements are met at the wavelength of maximum absorbance, the photoconduction response may follow the absorption spectrum closely.In other cases, the required energy may be provided by excitation to higher vibrational levels of the charge transfer excited state. However, it may be significant that appreciable photoconduction in regions to high energy of the charge transfer absorption maximum is seen only in complexes where donor and/or acceptor provide singlet transitions which overlap with the charge transfer absorption band. In such cases, a mechanism involving interaction of charge transfer and singlet excited states may be important. Fig. 3 outlines a possible mechanism, invoiving charge transfer between the excited singlet and charge-transfer states to give the initial charge separation.Further charge separation may utilise the energy released by the processes shown in fig. 3. Such a mechanism would require overlap between adjacent donor (or acceptor) molecules, but the excitations need not be on adjacent molecules initially since the singlet (Frenkel) exciton is mobile in the lattice and could readily diffuse to the less mobile charge transfer (Wannier) exciton.64 D', A-- t.p-e.a, PHOTOCONDUCTIVITY OF MOLECULAR COMPLEXES :A 2 hVct sol1 d) complex E - ground state FIG. 2.-Energy level diagram for photoconduction in crystalline molecular complexes. (A Wf = energy of formation of complex DA in solid phase. A l and A2 are attractive forces, other than charge-transfer resonance forces, in ground and excited state species, respectively.R1 and R2 are the corresponding repulsive forces. E = exchange energy in excited state. Res, and Res, are perturbations of ground and excited stated states due to charge transfer resonance. PDA, PD+A-, PD+ and PA- are the interaction energies between DA, D+A-, D+ and A-, respectively, and the crystal lattice.) - (lowest vacant orbital)+ + - - + / + + +- +(highest filled orbital) + D* + D++D+ + D A*+A--+A-+A (A*D)(A-D+)+(A-D)(AD+) 4-4- 1- i.e. (D*A)(D+A-)+(D+A)(DA-) FIG. 3.-D+A- ion pair separation by interaction with donor or acceptor singlet excited states. While these considerations rationalise the observed photoconduction spectral response curves. they do not explain why some of the complexes show little or no photoconduction. For example, anthracene-TCNQ is a very poor photoconductor, although its charge transfer band does overlap with the singlet transition of TCNQ.The magnitudes of steady state photocurrents depend on the balance between the rates of carrier generation and loss processes. Provided the incident light is totally absorbed within the crystal, these magnitudes axe observed by ourselves and others l 3 to be independent of the polarisation direction of the light with respect to the crystal axes. This suggests that carrier generation is a single photon process not involving the crystal surface. Carrier loss may occur by three main processes : (i) loss to the electrodes, (ii) loss by bimolecular recombination, and (iii) loss at impurity sites acting as recombination centres. The balance between these processes depends on the ratio of the transit time ( t ) of the carriers through the crystal to the carrier lifetime (7).V.M. VINCENT AND J . D. WRIGHT 65 Lifetime is limited by the rates of processes (ii) and (iii), while transit time depends on mobility and voltage gradient. With high applied voltages t becomes shorter than z, and saturation is observed, but this simplifying condition cannot be achieved with these low mobility materials in the present conditions owing to the onset of carrier injection from the electrodes before the necessary voltage gradient is reached. An alternative approximate approach has therefore been adopted, using the observation that photocurrent depends on light intensity (L) to a power ( I ) between 0.5 and 1.0 for all crystals.The value of I varies slightly for different crystals of a given complex, but is never uniformly close to 0.5 or 1.0. Since carrier generation involves a single photon, this means that unimolecular and bimolecular carrier losses are comparable under the conditions used. Thus, in the absence of impurities acting as recombina- tion centres, the total carrier loss rate is approximately twice the observed rate of carrier loss to the electrodes. The dominant impurities in the crystals used are the M+ acceptor- centres referred to previously and by other workers." The effects of such centres on recombination rates in crystals of the electron acceptor TCNQ have been discussed elsewhere.6 For the low carrier concentrations found here, their effect on recombination rates is small, and may even be to reduce recombination slightly by localising positive holes, the majority ca.rrier in photoconduction &en being negative.(This is also required if the photocurrent carriers are not to recombine rapidly with thermally generated carriers, which are known to be negative from the Seebeck coefficient data.) Further confirmation of the validity of this approach is provided by applying a simple collision model for carrier recombination to the steady state data. For a carrier density N m-3 in a crystal of length I, thickness t and width w, the recombination rate is where d = distance of approach below which carriers spontaneously drift together and recombine, ,u = mobility and F = applied field gradient. Now, N = i/wtepF, where i is the steady state photocurrent and e is the electron charge.Hence, if the electrode loss (i/e) and recombination loss terms are approxi- mately equal, d21p r 2*Fwte/nli. Calculations show that this is true, using the literature values of d 2: 10 nrn l4 and P 0 Thus, in the steady state, PLA 21 2ile where P is a parameter approximating to the quantum yield, L = photon flux per unit area, A = area of the illuminated face of the crystal. Since the dependence of photocurrent on light intensity is relatively insensitive to the applied field gradient, whereas i is linearly dependent on this in the ohmic region, the values of P quoted in table 4 are based on photocurrent values i corrected to a standard field gradient of 5 x lo4 V m-l assuming ohmic behaviour. The true form of the intensity dependence of photoconduction is quadratic, so the observed values of the power law I cannot be used simply to reduce errors introduced by this procedure.Table 4 gives values of P for the complexes studied. The range. of values quoted is that for at least 5 crystals of each complex and provides some estimate of the com- bined effects of experimental error and the approximations involved in deriving P. An independent estimate of the quantum yield of photoconduction may be obtained from the activation energy Eph associated with photoconduction? Eph 2-%d2pFN2wtl s-' > 10-7 m2 V-1 s-l 15-17 1-366 PHOTOCONDUCTIVITY OF MOLECULAR COMPLEXES may contain contributions from activation energies associated with (i) production of carriers from the excited state (ii) charge transport and (iii) trapping of charge carriers.However, (ii) and (iii) will also appear as contributors to the semiconduction activationenergy. Comparison of the semiconduction and photoconductionactivation energies (tables 2 and 5 ) shows that they are similar in magnitude, and (ii) and (iii) must therefore be very s m d contributors to Eph if unacceptably low values of the activation energy for thermal carrier generation are to be avoided. The quantum yield of photoconduction is therefore approximately exp( - Eph/kT). Table 5 gives values of this quantity deduced from the activation energies determined in this work. TABLE 4.-MEAN VALUES (WITH OBSERVED RANGES IN PARENTHESES) OF THE PHOTOCONDUCTION HMB acenaph thene anthracene chr y sene pyrene perylene (i) (ii) coronene TBN TCNQ 3 .2 ~ 10-5 (0.8-7.5) 1.1 x 10-4 (0.1-2.6) 5.3 x 10-9 (3.9-7.2) 3.4x lo-' (0.140) 5 . o ~ 10-5 (0.6-1 6) 5.8 x 10-5 (1.2-9.9) 2.0x (0.3 -4.5) 6.1 x 10-5 2.1 x 10-3 (0.2-14) (0.5-3.7) PARAMETER P TCNE chloranil * * * 4 . 6 ~ 10-5 1.5 x 10-7 (0.6-17) (0.1-3.1) (2.0-7.8) (0.3-5.5) 4 . 2 ~ 2 . 6 ~ fluoranil * 4.0~ lo-" (1.3-8.8) * 2 . 3 ~ (0.1-8.2) TCNB 3 . o ~ 10-7 (1.8-3.8) * * Indicates photoconduction below the limits of detectability. (For perylene-TCNQ, (i) and (ii) refer to measurements along and perpendicular to the needle axis of the crystals, respectively). TABLE 5 .-PHOTOCONDUCTION ACTIVATION ENERGIES (&h) (with values of the quantum yield parameter P, see text, for comparison) Eph/eV exp(-Eph/kT) at 293 K P acenaphthene-TCNQ 0.35 0 .8 ~ 1.1 x 10-4 anthracene-TCNQ 0.59 0 . 6 ~ 5.3 x 10-9 chrysene-TCNQ 0.18 0.8 x 10-3 3.4x 10-1 pery lene-TCNQ 0.23 10.1 x 10-5 5 . 8 ~ 10-5 pyrene-TCNE 0.20 3 3 . 6 ~ 10-5 4 . 6 ~ 10-5 anthracene-TCNB 0.26 3 . o ~ 10-5 3 . o ~ 10-7 chr ysene-fluoranil 0.32 2 . 8 ~ 4 . 0 ~ The agreement between these values and P values is variable, but similar trends are shown, suggesting that large differences in either parameter in different complexes are significant and that the general trends are worthy of analysis. The majority of the complexes showing measurable photoconduction have P values between The notable exceptions are the chrysene and tetra- benzonaphthaleno complexes of TCNQ, which are in the range =. and the and lo-'.V. M .VINCENT AND J . D. WRIGHT 67 anthracene-TCNQ complex, which is a very poor photoconductor (P N 5 x lo-”). The P values do not correlate with trends in ionisation potential or electron affinity, as can be seen from the data presented in tables 1 and 4, nor do they show any correlation with the position of the photoconduction response with respect to the charge transfer absorption spectrum. The efficiency of photoconduction must therefore be discussed in terms of the kinetics of the various competing processes by which the excited charge transfer (D+A-) state can dissociate or decay, and not merely in thermodynamic terms. hvct k2 DA ~f D’A- -+ . . . D’. . . A“. k i To maximise the quantum yield of photoconduction the rate constant k2 should be as large as possible while the rate constant k l should be minimised.Decay to the ground state becomes less likely as the overlap between the relevant orbitals of D and A decreases, and the chief factors affecting this overlap are the relative orientations, and separation, of D and A. k, depends on the relative heights of the energy barriers to migration of positive and negative charge through the lattice. If these barriers are high for both charges, the ion pair will neither separate nor move through the lattice. If they are low, the ion pair may move but there will be no particular tendency for charge separation. However, if the barrier to migration is high for one charge, but low for the other, charge separation is facilitated in the presence of an applied voltage. The barrier heights are determined by the energy levels of the charges on the initial ion pair in the crystal environment and by the energy of the transition state through which the charge passes in migrating to a neighbouring molecule.Since the charges will migrate most readily between like molecules (i.e. positive charge between neighbouring donors and negative charge between neighbouring zcceptors) the energy of this transition state will be lowest when there is good overlap of D with D or A with A. Thus there are two structural conditions which lead to high quantum yield for photo- conduction in molecular complexes : (a) n-orbitals of adjacent donor molecules or adjacent acceptor molecules (but not both) should overlap appreciably in the crystal, and (b) the relative orientation of donor and acceptor should be such as to produce poor overlap of the n-orbitals involved in the charge transfer transition involved in photoconduction. Fig.4 shows projections of nearest neighbour donor molecules onto the plane of the donor molecule, and similarly for acceptor molecules, for the crystal structures of some of the complexes studied here. Comparison of these projections for the anthracene-TCNQ and chrysene-TCNQ complexes shows that the acceptor network is virtually identical in the two structures, whereas the donor molecules approach much more closely in the case of chrysene than for anthracene. In both cases, acceptor- acceptor overlap is very slight. Thus, structural condition (a) above is satisfied for chrysene-TCNQ, but not for anthracene-TCNQ, and the P value for the chrysene complex is a factor of los greater than that for the anthracene complex.Other complexes where the D-D overlap is greater than the A-A overlap are perylene-TCNE, pyrene-TCNE and pcrylene-fluoranil (see fig. 4) and these are all moderate photoconductors, factors of 1 03-lo4 better than anthracene-TCNQ, as are the pyrene and perylene complexes of TCNQ, where the A-A overlap is greater than D-D. Anthraccne-TCNQ is not the only example of a poor photoconductor where there is little D-D or A-A overlap; this is also the case for hexamethylbenzenel chloranill, whose photocoilduction is not measurable. First, the perylene-TCNE, pyrene-TCNE and perylcne-fluoranil complexes show the required differential overlap, yet they are Two anomalies remain, however.68 PHOTOCONDUCTIVITY OF MOLECULAR COMPLEXES chrysene-TCNQ / w)ct( ant hracene-TCNQ -3.07 Q 3.07 Q >$ perylene-TCNE perylene-fluoranil p yrene-TCNQ perylene-TCNQ )---J-( / w-< HMB-TCNQ HMB-chloranil FIG.4.-ProJections of nearest neighbour donor (or acceptor) molecules onto the plane of the donor (or acceptor) molecule.V. M. VINCENT AND J . D. WRIGHT 69 considcrably poorer photoconductors than chrysene-TCNQ. Fig. 5 shows the pro- jection of the acceptor molecule onto the plane of the donor molecule for these complexes, together with the symmetries of the orbitals involved in the charge transfer transition occurring at the point of maximum photoconduction. It can be seen that the overlap of these orbitals is far from ideal in the chrysene complex, closer to ideal in pyrene-TCNE and almost ideal in perylene-fluoranil and perylene-TCNE. Thus the second structural condition for high quantum yield may be operating in these complexes, so that the complex with poorest D-A overlap is the best photoconductor.FIG. 5.-Overlap diagrams, with orbital symmetries, for (a) chrysene-TCNQ, (b) pyrene-TCNE, (c) perylene-TCNE and ( d ) perylene-fluoranil complexes. The second anomaly is the hexamethylbenzene-TCNQ complex. This has poor D-D and A-A overlap, yet shows appreciable photoconduction. The planes of the two nearest TCNQ molecules in this complex are only 1.1 A apart, so there is an appreciable void in the lattice in the region between closest -C(CN)2 groups of adjacent molecules. This void is large enough to accommodate a small cation, such as sodium, which is known to occur as an impurity in TCNQ itself.None of the other structures in fig. 4 provides this possibility, and although the semiconduction of these complexes suggests that similzir impurities occur in all of them, the actual site of the sodium ion must be an irregular one (e.g. dislocation). The MMB-TCNQ complex, however, can contain Na+ ions at regular sites in the lattice, and the presence of these ions, with the associated TCNQ-, will affect the energy levels of the charges on the excited state ion pair HMB+TCNQ-, producing the differential in barrier heights required to facilitate separation of the ion pairs. A similar'example of this chemical (as opposed to structural) trapping is the influence of bis-8-hydroxyquino- linato-Cu"-chloranil doped into a matrix of the corresponding Pdrr complex.The structural conditions for efficient generation of charge carriers could also be relevant to the discussion of carrier mobility in these complexes. Charge may migrate either (i) over networks of molecules of a single type (D or A), or (ii) along the70 PHOTOCONDUCTIVITY OF MOLECULAR COMPLEXES stacks of alternating D and A molecules which are always present in crystals of this type. The overlap of molecules involved in the migration steps is greater for (ii) than for (i), but (ii) is less favourable energetically, as, for example, negative charge must move from acceptor molecules to donor molecules in this process. If (ii) is the preferred pathway, mobility should be greatest along the direction of the stacks, but if (i) is preferred, it will be greatest approximately perpendicular to the stacks.We have studied the anisotropy of semiconductivity and photoconductivity in the perylene- TCNQ complex to distinguish between these alternatives. Perylene-TCNQ was chosen as it may be crystallised in both needle and plate forms, with the stacking direction approximately along the needle axis and perpendicular to the plane of the plate. The resistivities observed (table 1) were two orders of magnitude greater perpendicular to the stacking axis than along it. The crystals of both shapes were grown from identical starting materials and solvents, so this is a reflection of anisotropy of mobility in the sense required by pathway (ii) above rather than a spurious effect due to differing impurity concentrations.This is supported by values of d 2 / p deduced from steady state photoconductivity data, which are two orders of magnitude smaller for measurements along the stacking axis than for those perpendicular to it. However, the ratio of the mean P values for six crystals of each type is only 30, compared to a ratio of 200 in the resistivities (tables 1 and 4). The discrepancy suggests that the quantum yield for photoconduction is greatest when the voltage is applied perpendicular to the stacking axis, partially compensating for the lower mobility. This is consistent with the model proposed above for carrier generation, since the initial charge movement as the excited state ion pair separates is perpendicular to the stacking axis, and is thus enhanced when the applied field is in this direction.CONCLUSION Our results show a strong relationship between crystal atructure and photocon- ductivity of molecular complexes, and provide for the first timc a basis for predicting the relative magnitudes of photoconductivity in these materials. The TBN-TCNQ complex was predicted on this basis to be a good photoconductor, because of the similar sizes and shapes of the TBN and chrysene molecules, and it is encouraging that this is almost as good a photoconductor as the chrysene-TCNQ complex. The work emphasises the value of studies of a range of complexes, varying both D and A systematically. Further structural and electrical work is in progress, in the light of the results of this extensive preliminary survey. We wish to thank the Royal Society for an equipment grant, the Plessey Company and the S.R.C. for a C.A.P.S. Studentship for one of us (V. M. V.), and Mr. P. J. Munnoch for assistance with computer programming. l H. Akamatu and H. Kuroda, J. Chem. Phys., 1963,39,3364. H. Kokado, K. Hasegawa and W. G. Schneider, Canad. J. Chem., 1964,42,1084. C. K. Prout, R. J. P. Williams and J. D. Wright, J. Chem. SOC. A, 1966, 747. B. Mayoh and C. K. Prout, J.C.S. Faraday ZZ, 1972, 68, 1072. P. J. Munnoch and J. D. Wright, to be published. R. J. Hurditch, V. M. Vincent and J. D. Wright, J.C.S. Faraday I , 1972, 68,465. 1969). R. J. Hurditch, personal communication. 0. V. Kolninov, Doklady Akad. Nauk S.S.S.R., 1969, 189, 353. M. J. Hove, B. M. Hoffman and R. J, Loyd, J. Phys. Chem., 1972,76,1849. ' G. Kortum, Reflectance Spectroscopy; Principles, Methods, Applications (Springer, Berlin, lo H. Kuroda, K. Yoshihara and H. Akamatu, Bull. Chem. SOC. Japan, 1962, 35, 1604.V. M . VINCENT AND J. D . WRIGHT 71 B. J. Mulder, Rec. Trav. chim., 1965, 84, 713. l 3 H. Kuroda, personal communication. l4 F. Gutmann and L. E. Lyons, Organic Semiconductors (Wiley, New York, 1967), chap. 6. l 5 H. Kuroda, K. Yoshihara and H. Akamatu, Bull. Chem. SOC. Japan, 1963, 36, 1365. l6 M. C. Tobin and D. P. Spitzer, J. Chem. Phys., 1965, 42, 3652. l7 M. M. Sokolova and L. D. Rozenshtein, Soviet Phys.-Solid State, 1971, 12, 2465. l 8 M. Akiyama and Y. Utsugi, J. Phys. SOC. Japan, 1969, 27, 886.

ISSN:0300-9599    

DOI:10.1039/F19747000058

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Reaction of Halogens with Oxide Surfaces BY JOHN F. J. KIBBLEWHITE AND ANTHONY J. TENCH* Chemistry Division, A.E.R.E., Harwell Received 30th May, 1973 Charge transfer reactions at oxide surfaces are discussed for the halogens, chlorine, bromine and iodine as strong electron acceptors. Adsorption of chlorine and bromine on MgO prepared in vucuo is characterised by the appearance of a band at 430nm in the diffuse reflectance spectrum and an e x . signal with g-factors of 2.0099 and 2.0020. These features are destroyed on heating and oxygen is evolved. Chlorine and bromine react with the lattice oxygen ions to give halide ions corresponding to about 20 and 10 % respectively of the surface MgO ion pairs ; no evidence of the formation of molecular X i ions was obtained. These results can be understood in terms of the initial formation of charge deficient groups of oxygen ions, such as 0;- and subsequent release of oxygen.The reaction of bromine with MgO would not be expected purely on a consideration of the bulk heats of formation, and indicates an appreciable concentration of surface oxygen ions with less than 5-fold co-ordination. Iodine, as expected from the low heat of formation of the iodide, does not displace oxygen ions on the surface but some iodide ions are formed. All the halogens are able to displace oxygen adsorbed as 0, from the surface indicating that this ion is only weakly bonded to the surface. The adsorption of oxygen has been studied on a wide variety of oxides. Both insulating and non-stoichiometric semi-conducting oxides have been investigated using a variety of techniques; in particular, the paramagnetic oxygen species that can be formed on such oxides have been studied in some detail using electron spin resonance indicating that 0; is formed on many 0~ides.l'~ By analogy we might expect that the halogens would react to form similar molecular ions, and could be used to investigate the charge transfer properties of oxides.Ions such as C1, have been well characterised in the alkali halides 5-8 and are formed by low temperature irradia- tion; they have also been identified on SiOa and zeolite lo surfaces where the adsorbed halogens have been y- or photo-irradiated at low temperature. Very little work has been reported on the adsorption of halogens on oxides; the adsorption of chlorine on ZnO has been studied recently and a new e.s.r.signal with a g-factor of 2.015 has been reported, there appears to be no evidence for the formation of the molecular ion.ll The adsorption of iodine on ZnO is reported to show a signal with a g-factor of 2.0068 which has been ascribed to These results are difficult to interpret because the e.s.r. evidence is contradictory and does not give a clear indica- tion of the species involved. In this work the adsorption of halogens has been studied on MgO. Although there is no indication of the formation of stable molecular ions, the results indicate that a number of changes occur on the surface of the oxide, suggesting that charge transfer reactions occur which involve oxide ions of the lattice.' EXPERIMENTAL The samples of magnesium oxide used were prepared by thermal decomposition of the carbonate in silica tubes as described previously,14 and finally heated at 1250 K for 15 h at a pressure of less than 1.33 x N m-2.Specpure chlorine and oxygen, and A.R. broming and iodine were used; the halogens were purified prior to use by several freeze-pumpine cycles and by distillation in uacuo. The amount of halide ion formed during the adsorption 72J . F. J . KIBBLEWHITE AND A . J . TEkCH 73 was determined spectrophotometrically using Hg(CNS)2 and ferric alum following the method of Swain.15 The Cary 14 spectrophotometer was calibrated using standard solutions of sodium chloride, bromide and iodide. Blanks carried out on magnesium oxide prepared in vacuo indicated that the chloride content was < 10l8 ion g-l before the addition of chlorine.The specific surface areas of the samples were obtained using a nitrogen sorptometer. The adsorption of gases on to the MgO samples was conducted using a standard gas handling line. Quantitative measurements of the amount of oxygen evolved were obtained by, first, measuring the pressure-volume characteristics and, second, passing the gas products to a gas chromatograph. The products were pumped through a liquid nitrogen trap, to remove excess chlorine, and into a Toepler pump which was then used to transfer the samples to a carrier-gas loop of a Perkin-Elmer F11 chromatograph, which incorporated a 2 metre molecular sieve 5A column. The e.s.r. spectra were obtained at 77 K on a Varian 4502 spectrometer operating at 9.2 GHz with 100 kHz modulation.In all the e.s.r. spectra shown in the figures the magnetic field increases from left to right and first derivative spectra were recorded. Spin concentra- tions were calculated by direct double integration using an on-line computer (LABCOM). The g factors were measured relative to diphenylpicrylhydrazyl taken as 2.0036. Diffuse reflectance spectra were recorded on a Cary 14 spectrophotometer using MgC03 as a reference ; the method has been described previously.16 The analogue output from the e.s.r. spectrometer and reflectance spectrophotometer was fed directly into the LABCOM l7 " on-line " computing system. This system accepts input from several instruments at the same time, the data is digitised using CAMAC l8 modular units and fed into a PDP8/I computer ; the data in the computer are displayed on an oscillo- scope.The data can be stored and processed, using a normal teletype mode of interaction with the system. The e.s.r. data were processed using a regression analysis to determine the base line and the spectrum integrated between chosen points. The integrated spectrum was displayed, and, if acceptable, a second integration was carried out to determine the spin concentration in comparison with the standard. The great advantage of this method is that any number of integrations can be carried out (each taking only a few seconds) on the same data to cover different ranges of magnetic field and this gives a much better assessment of the contribution from the wings of the line.For the reflectance work the spectrometer output was converted to give a scale linear in energy and processed using the Kubelka-Munk l9 function to give an absorbance scale, and the data were then analysed using a Du Pont Model 310 curve resolver. RESULTS REACTION WITH CHLORINE ADSORPTION AND DESORPTION Chlorine was found to be adsorbed on the surface of the MgO samples at 300 K in a fast process which was only partially reversible after 15 min evacuation at this temperature ; analysis of the desorbed gases showed that only chlorine was present and no oxygen was evolved. The samples turned a beige colour which remained after evacuation at 300 K. Analysis of the evolved gases after heating in vacuo at successively higher temperatures indicated that oxygen was the only component present apart from chlorine.Separate experiments on several samples showed that the desorbed oxygen approached a limiting value at 673 K and only minor quantities were evolved above this temperature (fig. 1). A similar heating experiment on a MgO sample that had no adsorbed chlorine gave a negligible desorption of oxygen con- firming that no adsorbed oxygen is present on the surface of the sample following preparation at 1273 K and subsequent sealing off (table 1). Analysis of the MgO samples subsequent to the heating experiments showed that a high level of chloride ion was present in comparison to an untreated sample. For a series of samples ranging in specific surface area (SSA) from 70 to 90 m2 g-' the74 HALOGEN REACTIONS ON MgO surface chloride coverage, expressed as chloride ions/surface MgO pair, varied be- tween 0.16-0.21 (table 1).No oxygen was evolved if the samples were kept at 300 K but colorimetric analysis showed a high chloride content corresponding to a surface coverage of 0.16. Heating the samples with adsorbed chlorine to 673 K produced a slight decrease in the specific surface area of the sample (table l), suggesting that sintering of the oxide is promoted by the presence of chloride ions. 0 temperature/K FIG. 1.-Oxygen liberated as a function of temperature after adsorption of chlorine on magnesium oxide at 300 K. The effect of excess chlorine on the analytical procedure was investigated using an aliquot of chlorine gas, identical to the amount of chlorine evolved from the sample during heating.The result showed that the chlorine contribution to the final chloride concentration was < 3 %, which is regarded as insignificant. TABLE 1 .-EVOLUTION OF OXYGEN FROM HALOGEN TREATED MgO total oxygen evolved I sample %:'i molecule g-1 M958/C12 72 4 . 6 ~ 1019 M961/C12 79 4 . 7 ~ 1019 M1008/C12 92 4 . 4 ~ 1019 M1007-f 104 < 1 . 0 ~ 1 0 ~ ~ M936/CI2 70 <4.0x lo1' maximum temperature / K 673 673 673 673 300 halide ions coveragelion ratio (surface MgO pair)-1 coverage/ion g-1 halide/@ 0.21 1 . 7 ~ 10'' 3.7 0.20 1 . 8 ~ lo2' 3.8 0.16 1 . 7 ~ 10'' 3.9 0.16 1 . 3 ~ lozo - < 1 . 7 ~ 10" - - M500/Br2 54 1.6 x 1019 873 0.1 1 0 . 6 9 ~ 1020 4.2 M5Ol /I2 67 0 . 0 4 ~ lOI9 1073 0.01 0 . 0 9 ~ lozo 22.0 * measured after thermal treatment ; 7 pure MgO which was prepared simultaneously with M1008 ; no chlorine adsorbed.REFLECTANCE MEASUREMENTS The beige colour of the sample was investigated more closely by recording the diffuse reflectance spectrum of the sample before and after chlorine treatment (fig. 2) over the range 200-800 nm. The MgO prepared in vacuo has a significant absorptionJ. F. J. KIBBLEWHITE AND A . J . TENCH 75 at 250nm and at shorter wavelengths there is some evidence of a fluorescence in agreement with previous w0rk.l The adsorption of chlorine destroys this fluoresc- ence and leads to new absorption bands at 210, 260 and 430nm and a shoulder at 350 nm ; the broad band at 260 nm is easily removed on pumping and is presumably associated with weakly adsorbed chlorine. Gaseous chlorine shows bands at 330-340 and 215 nm in CC1, 2 0 , 21 and 330 nm in the gas phase,22 which compare favourably with the bands at 350 and 210 nm observed when adsorbing chlorine on MgO.The band at 430nm is probably responsible for the beige colour because of the tail extending to long wavelengths. Comparison with the reflectance spectrum of de- hydrated MgC12 (fig. 2(e)) shows no evidence of the band at 430 nm and so it is not associated with a chloride lattice. 10 600 800 200 400 wavelength /nm FIG. 2.-Diffuse reflectance spectra of (a) MgO in U ~ C U O ; (b) MgO with excess dorine at 300 K ; (c) MgO with excess chlorine pumped off; (d) MgO/C12 after heating to 673 K; (e) magnesium chloride ; (f) (Oi)s on MgO. In a separate series of experiments a sample was heated and the diffuse reflectance spectra recorded at various temperatures. The output was recorded directly in digital form in the computer as described in the Experimental section.The reflectance data were then analysed using the Kubelka-Munk equation l9 : where R is the reflectance, k the absorption coefficient and S the scattering coefficient, and the wavelength scale was transformed to a scale linear in energy. The spectrum of chlorine adsorbed on MgO after pumping off the excess chlorine was analysed using Gaussian profiles (fig. 3) and bands at 430,345,290,245 and 220 identified. The band (1 - R)2/2R = kS76 HALOGEN REACTIONS ON MgO at 345 nm is rapidly removed on pumping and therefore associated with physisorbed chlorine. The band at 245 nm also decreases on pumping and a similar effect has been observed when physisorbed oxygen is removed from the surface; in this latter case it has been attributed by the author to interaction of the adsorbed species with surface states of the Mg0.23 The band at 220 nm is not significantly affected by heating (fig.2(b), (c) and (d)) but it is difficult to resolve this band precisely because it was recorded at the far end of the wavelength range of the instrument. The origin of the band at 290 nm is uncertain, but it could originate from an oxychloride species of the oxide, or more probably it is associated with surface states modified by the presence of the chloride ion. The curve resolved spectrum of pure MgO showed only a band at 250 nm. wavelength/nm energy/eV FIG.3.-Curve resolved reflectance spectrum of chlorine treated MgO with excess chlorine pumped Off. 300 400 500 600 (4 temperature/K reflectance peak at 430 nm ; (b) e.s.r. spin concentration. FIG. 4.-Normalised isochronal annealing curves of the chlorine treated MgO: (a) area underJ . F . J . KIBBLEWHITE AND A . J . TENCH 77 The remaining band at 430 nm is of particular interest since it seems to be associa- ted with the interaction of chlorine with the surface to form a strongly chemisorbed species. Heating at elevated temperatures causes a reduction of intensity of the band and the isochronal annealing curve (fig. 4(a)) shows the same type of pattern as the evolution of oxygen from the surface. The reflectance spectrum of (Ol)s on MgO (prepared by reaction of oxygen with the F:(H) centre 14* 24 was also recorded for reference purposes (fig.2cf)). E.S.R. MEASUREMENTS Magnesium oxide prepared in vacuo shows no significant e.s.r. signals before the addition of halogens, apart from a trace of Cr3+ indicating an impurity level of < 1 pg g-l. A new signal (fig. 5) with g-factors of 2.0099 and 2.0020 appears immed- iately when the chlorine is adsorbed and increases in intensity at room temperature to reach a concentration of 2 x 1017 spin g-1 after about 20 min. Identification of the species giving rise to the signal is difficult because it shows no hyperfine structure which indicates the absence of chlorine containing species such as C1;.l0* 2 5 The intensity of the signal is increased by the addition of small pressures of oxygen but no broadening of the lines is observed when 1.33 kN m-2 (10 Torr) of oxygen is added. U 2.0099 2.0020 FIG.5.-E.s.r. spectrum of chlorine adsorbed on magnesium oxide at 300 K. The g-factors are similar to those reported for 0, but no line corresponding to a gzz value is observed at high resolution. The g,, values for are a sensitive function of crystal field and could be smeared out by surface heterogeneities. No evidence for this was obtained either from inspection of the integrated spectrum (displayed using LABCOM) or by successive double integration over different field ranges. Area measurements of the first derivative signal using a planimeter show that the area of the positive and negative signal segments are equal. This confirms that the complete signal has been observed and no peaks corresponding to broadened gzz factors have been ignored.The signal was also unaltered when treated with 1.46 kN m-2 (1 1 Torr)78 HALOGEN REACTIONS ON MgO of hydrogen at 300 K, thus indicating that the signal does not arise from a species such as 0- which is known to react with hydrogen.26* 27 The absence of exchange when the sample was left in contact with oxygen enriched to 58 atom percent 1 7 0 indicates that the species is not (O;)s on the oxide surface, since this is known to exchange readily with gas phase 1702.28 The intensity of the e.s.r. signal decreases with increasing temperature (fig. 4(b)) but does not show a correlation with the de- crease of intensity of the reflectance peak at 440 nm (fig.4(a)). REACTION WITH BROMINE AND IODINE \TJl?on the bromine treated samples were heated oxygen was evolved and subsequent analysis showed that bromide ions were present. Heating samples with adsorbed iodine produced no significant oxygen release, even at temperatures as high as 1100 K ; the iodide ion concentration was also low (table 1). The reflectance spectra of the bromine and iodine treated samples were less resolved than those of the chlorine samples due to the difficulty of removing the excess halogens which have high optical absorption. However, curve resolution of the computer treated spectra suggests that after heating the bromine treated samples to 473 K bands are present at 360, 280, 235 and 215 nm. No bands could be resolved in the iodine system at 300 K or higher temperatures.E.s.r. measurements on samples with adsorbed bromine and iodine produced spectra similar to the chlorine induced signal shown in fig. 5 ; with g-factors of 2.0104 and 2.0023 for the bromine sample, and 2.0091 and 2.0020 for the iodine treated sample. Bromine containing radicals such as BrF are known to exist under certain conditions on Mg0,29 but are unstable at room temperature. The possible formation of these species was investigated by freezing bromine at the top of the sample tube with liquid nitrogen and allowing it to adsorb onto the sample in the microwave cavity. The sample became a creamy yellow but no evidence for molecular ions was observed and an e.s.r. signal similar to that shown in fig. 5 was formed. Iodine behaved similarly and the signal intensity increased slowly on standing at room temperature and could be completely destroyed at 400 K.TABLE 2.-HEATS Ot I'OKMATION AH OF SOLID OXIDE AND HALlDES chloride bromide iodide - -- oxitie AH! AH! AH! AH1 metal kcal rnol-1 AH/eV kcal rnol-1 AH/eV kcd rnol-1 AH/eV kcal moI-1 AHIeV in3gnesium (2f) 143 6.20 153 6.63 125 5.42 87 3.77 aluminium (3+) 391-401 17.20 168 7.30 126 5.46 75 3.25 silicon (4+) 216-218 9.41 liq - liq - 45 1.96 zinc (2+) 83 3.60 99 4.30 79 3.40 50 2.15 COMPETITIVE CHARGE TRANSFER When oxygen is added to a MgO sample containing F:(H) centres 14* 24 the blue colour disappears and (O;)s is formed.' Similarly, the reaction of chlorine with F2(H) centres might be expected to form (Cly)s. However, adsorption of chlorine was found to destroy the characteristic signal and blue colour, form an e.s.r.signal similar to that shown in fig. 5 and produce the beige colour also observed from unirradiated MgO treated with chlorine. The reaction of chlorine and iodine with (O;)s on MgO was also studied. The e.s.r. signal from (O;)s (fig. 6(a)) has a number of lines close to a gzz factor of 2.07, these corresponding to various different sites on the surface. The adsorption of aJ . F . J . KIBBLEWHITE A N D A . J . TENCH 79 halogen destroys these signals and gives a signal identical to that obtained by halogen adsorption on unirradiated MgO (fig. 6(b)). Quantitative adsorption/desorption measurements showed that when 6.9 x lo1' molecules oxygen per gram were adsorbed by the sample to form (OY)~, 6.9 x lo1' molecules oxygen per gram were evolved on reaction with chlorine at room temperature.This shows that (O;)s is destroyed completely by reaction with chlorine. No Cl; formation was observed and further reaction of the sample with oxygen did not reform 0; ; adsorption of iodine gave the same result. I I 1 2.0726 2.0078 2.0012 U 2.0099 2.0020 FIG. 6.-E.s.r. spectra of : (a) (Oi)s on magnesium oxide ; (6) after adding chlorine 'to (O& DISCUSSION Previous studies have shown that oxides can donate electrons to electron acceptor molecules adsorbed on the surface and this has led to the suggestion l 2 that halogens can react with oxide surfaces containing readily available electrons to form adsorbed molecular halides which can be observed by e.s.r. spectroscopy.Earlier work on MgO indicates that electron donation can occur to electron acceptor molecules such as nitrobenzene 30 and trinitrobenzene 31 adsorbed on the surface. However, our results on MgO show no evidence for the stabilisation of the molecular halide ions X, during the adsorption process even when readily available electrons are trapped at the surface. Although the simple analogy between the 0: and X; ions does not appear to hold, there is clear evidence that the halogens do react strongly with the surface in a more complex process. The presence of chloride ions after adsorption of chlorine indicates that a charge transfer reaction has taken place between the adsorbed chlorine and the MgO lattice.80 HALOGEN REACTIONS ON MgO Similar charge transfer reactions have been reported when methyl iodide is adsorbed on The number of chloride ions formed corresponds to some 20 % of the surface oxide ions, which indicates far reaching changes in the surface topography and is much higher than the number of adsorbed nitrobenzene ions formed 30 on a similar surface (- 1 %).The formation of a band in the reflectance spectrum and an e.s.r. signal indicate that at least one new species has been formed. The evolution of oxygen gas on thermal treatment is accompanied by a decrease in intensity of both the optical band and the e.s.r. signal. This suggests that the new species formed must contain oxygen, and therefore arise from what were originally oxide ions of the lattice. The possibility of intrinsic or extrinsic point defects acting as sources of electrons can be excluded since the high concentrations required are more than an order of magni- tude higher than the impurity content and several orders of magnitude higher than the intrinsic vacancy concentration. It is now necessary to discuss the results in more detail, and clearly the origin of the reflectance band at 430 nm is of considerable importance.Possible species are surface oxychloride compounds, such as hypochlorite and chlorates ; however, these species are expected to absorb light in the 200-300 nm region and not at 430 nm, for example, sodium hypochlorite absorbs at 295 nm. Test reactions with lead nitrate for hypochlorite and o-tolidine for chlorate 33 gave no indication of the presence of such species.These arguments suggest that the band does not arise from chlorine containing species and is therefore probably associated with lattice oxygen ions in some unknown form. This idea is supported strongly by the observation that MgO samples irradiated in the presence of oxygen give a band at 430 nm 1 3 9 29 very similar to that obtained when chlorine is adsorbed. A comparison of the number of chloride ions formed to the number of oxygen gas molecules evolved during heating to 673 K gives a ratio of 3.8 which is close to the correct stoichiometry for the reaction of the chlorine with oxide ions of the lattice 2CI2 +202-(lattice)-+4C1- + O,(gaseous). The quantity of oxygen evolved is two orders of magnitude larger than the spin concentration of the e.s.r. signal and it seems fairly clear that the e.s.r.signal relates only to a small fraction of the total process occurring on the surface. Absolute measure- ments on the concentration of species corresponding to the reflectance peak are not possible at present, since no information is available on extinction coefficients, but the thermal stability broadly corresponds to the inverse of the oxygen evolution curve. From this comparison we have assumed that the reflectance peak at 430 nm is likely to be associated with the precursor of the evolved oxygen and thus the site that has donated electrons to the adsorbed chlorine. No exact correlation with the oxygen desorbed from the surface is expected since the oxygen may be held adsorbed on the surface for a while after the precursor has been destroyed.The oxygen precursor and the band at 430nm probably arise from an oxygen complex formed by abstraction of electrons from surface and subsurface 02- ions of the lattice. A number of such species are possible, for example 0-, 0 2 , O;-, and 0; ; however, some of these can be eliminated quite easily. The peak at 430 nm is consistent with 07 279 34 but not 0; ; however, the e.s.r. evidence is not consistent with the presence of either 0; or 0-. The peroxide ion 0;- is diamagnetic and would be expected to release gaseous oxygen easily on heating; however, peroxide ions usually absorb at shorter wavelengths, for example at 250 nm for BaO,. The evidence put forward in these arguments appears to be somewhat contra- dictory, and the most satisfactory model for the donor site appears to be a site in- volving two or more oxygen nuclei that have lost part of their negative charge.J .F. J . KIBBLEWHITE AND A . J . TENCH 81 Alternatively, this could be regarded as the trapping of holes on neighbouring lattice oxide ions and one possible form of this model is shown in fig. 7. Such a pair of trapped holes would form a species which is intermediate between two 0- ions and a peroxy ion OZ-. C I- C I' / / MgZt COY' Mg2' 02- Mg2+ J 02- MgZt 02- Mg2' 02- FIG. 7.--.Formation of an 0;- complex such as 0;- by electron abstraction from lattice 02- ions to form surface chloride ions. In the light of these ideas it is interesting to compare the results on the reactivity of the surface with various halogens. The data in table 1 shows that chlorine and to a lesser extent bromine are capable of replacing oxide ions of the lattice with the evolu- tion of oxygen, this is not true for iodine, although some iodide ions are formed. This is not easily understood in terms of the electron affinity values 35 from the gas phase or bond energies,36 since the energy available from the reaction : Xz +2e+2X- is 5.0 eV for C12, 5.3 eV for Br, and 4.9 eV for I2 ; which are very similar values, although we have neglected important coulombic and polarisation interactions with the charged lattice ions.However, a comparison of either the heats of formation or free energy values for the oxide and halides 37 is more useful. Since this is not an equilibrium process the heats of formation have been used (table 2) and in practice the free energy values also lead to the conclusions outlined below.From these data we can see that chlorine would be expected to liberate oxygen from MgO. The effect of bromine is somewhat surprising, and must indicate a significantly lower stabilisation of the surface oxide ions compared to the bulk. It is reassuring that iodine does not liberate oxygen and this clearly places a lower limit on the stability of the surface oxide ions. However, it is likely that some of them are able to donate one electron even to iodine and this accounts for the significant quanti- ties of iodide formed. This is a different type of electron transfer where lattice oxygen is not replaced and probably corresponds to the same situation as has been observed previously for nitrobenzene 30 and similar electron acceptors adsorbed on the surface.The concentration of the iodide ion formed is similar to that observed for the negative ions of the organic electron acceptors and is much lower than for the chloride ion. These results allow some conclusions to be drawn about the nature of the surface oxide ions. The liberation of gaseous oxygen from the surface by bromine indicates that some surface MgO ion pairs must be destabilized with respect to the normal bulk material by at least 18 kcal mol-1 (0.8 eV). The origin of this effect lies in the changed coulombic and polarisation effects present at the surface. Theoretical calculations show that the coulombic and polarisation terms act in opposite ways and the net effect is to reduce the net stabilization of the ion pairs at the surface by an amount 389 3982 HALOGEN REACTIONS ON MgO corresponding to 5.6 kcal niol-l (0.25 eV) for a (100) surface.This alone is not sufficient to account for the higher reactivity of the surface ; although the values of the surface energy show considerable variation 38 it seems unlikely that they are in error by more than a factor of two and some experimental results 38 suggest a figure of about 7.4 kcal mol-l (0.33 eV). However, it is unlikely that the surface of these pow- ders can be regarded as a good (100) plane since defects such as steps, kinks, edges will be present in addition to the possibility of other crystal planes of lower symmetry than The net effect of these sites and low index surfaces is to provide surface ions of lower symmetry and therefore of greater surface energy.Clearly, these are the surface features that are most likely to be decreased on high temperature treatment but it is unlikely that they have been eliminated completely. Calculations on the ratio (7) of surface to bulk Madelung potential for NaCl type structure 41 indicate considerable variation as the surface co-ordination decreases, for a (100) surface with five-fold co-ordination y = 0.96, for a (110) surface with four-fold co-ordination y = 0.86, and for a (210) surface with three-fold co-ordination, y decreases to 0.60. The magnitude of the effects on the surface energy when both coulombic and polarisation effects are present does not appear to be known in detail.The electrostatic part of the surface energy is proportional to 1 -y, which would give an increase in the surface energy of a factor of 3.5 on going from five-fold to four-fold co-ordination. This is in good agreement with the values calculated by Shuttle- worth 42 for NaF where polarisation terms were included but are fairly small. For MgO the polarisation terms will be more important and will result in a smaller increase in surface energy ; however, it seems likely that a sufficient change in surface energy will be obtained in sites or surface planes of lower co-ordination to permit reaction of bromine with the surface with the elimination of oxygen. Chlorine on the other hand can react with a wider range of surface sites and planes and twice the coverage is obtained.From these figures it seems likely that as many as 10 % of the oxygen ions exist in sites of lower than five-fold co-ordination at the surface. This is not unreasonable since even assuming perfect cubes with (100) faces then 160 particles (specific surface area = 100 m2 g-’) will have - 3 % of surface ions in lower than five-fold co-ordination. In practice, the crystal surface will contain steps and similar defects together with a proportion of lower symmetry planes. The surface sites involved with the formation of iodide ions may be of even lower symmetry; this could correspond to situations where the oxygen ions have three-fold co-ordination on the surface (for example corner sites), and such ions could readily donate an electron. The absence of an observable e.s.r.signal from 0- can be explained in terms of a delocalisation of the electrons over several oxygen ions. The use of very powerful electron acceptors such as the halogens to explore different types of electron transfer reactions with oxide surfaces opens up a new way of exploring the surface energies of crystals and measuring the reactivity of surface ions. This approach will allow the comparison of the effects of different methods of preparation on the surface reactivity. Exploratory work on other oxides indicates that oxygen is not released when chlorine is adsorbed on N,03 and SOz, and this can be predicted from a comparison of the relevant heats of formation ; on the other hand, chlorine does react with ZnO to release oxygen,29 and this is expected from a comparison of the heats of formation of ZnO and ZnC1, (table 2).This indicates that surface studies involving the halogens must be treated with caution since in some cases reaction with the surface is of import- ance. Lastly, the desorption of (Ol)s from the surface as gaseous oxygen when halogens are adsorbed at room temperature confirms the low electron affinity of the adsororbJ . F. J . KIBBLEWHITE AND A . J . TENCH 83 oxygen on the surface. This could provide a useful way of exploring adsorbed species on a surface that has been exposed to chemical treatments and contains adsorbed ions about which little is known. A. J. Tench and P. Holroyd, Chem. Comm., 1968,471. A. J. Tench and T. Lawson, Chem. Phys. Letters, 1971,8,177. C. Naccache, P. Meriaudeau, M.Che and A. J. Tench, Truns. Faraduy SOC., 1971, 67, 506. P. Meriaudeau, C. Naccache and A. J. Tench, J. Catalysis, 1971, 21,208. D. L. Griscom, P. C. Taylor and P. J. Bray, J. Chem. Phys., 1969,50,977. E. Boesman and D. Schoemaker, J. Chem. Phys., 1962,37,671. J. Roncin, Chem. Phys. Letters, 1968, 49, 2876. J. A. R. Coope, C. L. Gardiner, C. A. McDowell and A. I. Pelamn, Mol. Phys., 1971,21,1043. R. D. Iyengar, V. V. Subba Rao and A. C. Zettlemoyer, Surface Sci., 1969,13,251. ’ E. B. Zvi, R. A. Beaudet and W. K. Wilmarth, Chem. Phys. Letters, 1969,51,4166. lo J. E. Bennett, B. Mile and B. Ward, J. Clzem. Phys., 1968, 49, 5556. I2 S. Larach and J. Turkevich, Surface Sci., 1970, 20, 192. l3 A. J. Tench and J. F. J. Kibblewhite, Chem. Phys. Letters, 1972, 14, 220. R. L. Nelson, A. J. Tench and B. J. Harnisworth, Trans. Faraduy SOC., 1967,63, 1427. I 5 J. S. Swain, Chem. Ind. (London), 1956, 418. l 6 R. L. Nelson, J. W. Hale and B. J. Harmsworth, Truns. Fmuday SOC., 1971, 67,1164. l7 A. M. Deane, C. Kenward and A. J. Tench, A.E.R.E. Report 7020. l8 CAMAC. l9 G. Kortuin, Reflexion Speltroscopie (Springer, Berlin, 1969). 2o M. Anbar and I. Dostrovsky, J. Chem. SOC., 1954, I, 1105. 21 F. W. Czech, R. J. Fuchs and H. F. Antczak, Anal. Chem., 1961, 33,705. 22 C. W, Weber and 0. H. Howard, Anal. Chem., 1963,35,1002. 23 Pi. L. Nelson and J. W. Hale, Disc. Furaduy SOC., 1971,52, OOO. 24 A. J. Tench and R. L. Nelson, J. Colloid Interface Sci., 1968, 26, 364. 2 5 A. J. Tench and J. F. J. Kibblewhite, J. Chem. SOC. A, 1971, 2282. 26 A. J. Tench and T. Lawson, Chem. Phys. Letters, 1970,7,459. ’’ A. J. Tench, T. Lawson and J. F. J. Kibblewhite, J.C.S. Faraday I, 1972, 68, 1169. 28 A. J. Tench and M. Che, to be published. 29 A. J. Tench and J. F. J. Kibblewhite, to be published. 30 A. J. Tench and R. L. Nelson, Trans. Faraday SOC., 1967, 63,2254. 31 M. Che, C. Naccache and B. Imelik, J. Cutulysis, 1972,24, 328. 32 J. Cunningham and A. L. Penny, J. Phys. Chem., 1972,76, 2353. 33 P. Urone and E. Bonde, Anal. Chem., 1960,32,1666. 34 A. J. Tench, J.C.S. Furuday I, 1972, 68, 1181. j 5 H. S. W. Massey, Electronic and Ionic Impact Phenomena (Oxford, 1969), vol. 2, p. 1062. 36 NSRDS-NBS 31, January 1970. 37 NBS Technical Note-Series 270-3 and 6, January 1968. 38 G. C. Bernon and K. S. Yun, The Solid-Gus Intet$uce, ed. E. A. Flood (Arnold, London, 1967), 39 P. J. Anderson and A. Scholz, Truns. Furaduy SOC., 1968,64,2973. 40 M. Boudart, A. Delbouille, E. G. Derouane, V. Indovina and A. B. Walters, J. Amer. Chem. 41 J. D. Levine and P. Mark, Phys. Rev., 1966, 144, 751. 42 R. Shuttleworth, Proc. Phys. SOC. A, 1949, 62, 167. A Modular Instrumental System for Data Handling, Euratom report, EUR 41W- 1969. vol. 1, p. 203. SOC., 1972, 94, 6622.

ISSN:0300-9599    

DOI:10.1039/F19747000072

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Low Pressure Hysteresis in the Sorption of Carbon Tetrachloride Vapour on Polymer Carbons BY BRIAN MCENANEY School of Materials Science, Bath University, Bath BA2 7AY Received April 1973 Adsorption-desorption isotherms for carbon tetrachloride vapour at 20.0"C on microporous cellulose and polyacrylonitrile carbons exhibit low-pressure hysteresis which is reduced or eliminated by widening of pores which accompanies steam-activation. Comparison of sorption of carbon tetrachloride on unactivated carbons heat-treated to 900 and 2700°C shows that the adsorbate is largely confined to external surfaces and macropores and excluded from the major part of the micro- pores by molecular sieve action. The bulk of adsorbate retained by unactivated carbons on desorp- tion to P/Po = 0.0 was adsorbed at P/Po > 0.9 indicating a pressure-threshold effect for low pressure hysteresis ; about 5 % of the micropore volume originally inaccessible to carbon tetrachloride is penetrated by the expansion-intercalation process in the case of cellulose (900°C) unactivated carbon.A model for this process is proposed based on localised fracture of the carbons. Application of the theory of adsorption-extension of Flood and Heyding shows that stresses induced in volume elements of the carbons are commensurate with or greater than measured bulk fracture strengths for carbons. Low-pressure hysteresis in sorption of organic vapours by porous carbons is usually attributed to intercalation of adsorbate molecules in micropores leading to changes in the structure of the adsorbent which are either irreversible or can only be reversed with difficulty by processes such as annealing under vacuum.Bailey et aZ.l have recently discussed this phenomenon in general thermodynanzic terms. The adsorbate plus adsorbent is considered as a two-component system and, at some point in the course of the sorption isotherm when conditions are thermodynamically favourable, an irreversible transition from the unperturbed to the perturbed (ex- panded) state occurs accompanied by intercalation of the adsorbate molecules. This paper reports measurements of sorption of carbon tetrachloride vapour at 20°C on series of steam-activated carbons prepared from cellulose and polyacrylonitrile which exhibit low-pressure hysteresis. A mechanism is proposed for the expansion- intercalation process based on the theory of adsorption-extension which is believed to be consistent with the general thermodynamic description of the phenomenon by Bailey et aZ.l EXPERIMENTAL Cellulose (Whatman's ashless powder for chromatographic purposes) and polyacrylo- nitrile powder (a laboratory sample from I.C.I.Limited) were compressed into one inch diameter pellets and carbonised in oxygen-free nitrogen at 900°C. Steam-activation was also carried out at this temperature; details of the preparation of active carbons have been reported elsewhere.2 The amount of porosity developed in carbons by reaction with steam is related to the extent of reaction. The degree of activation is, therefore, commonly expressed as the burn-off, % BO, i.e.the amount of carbon removed by the reaction expressed as a percentage. Alternatively, the degree of activation may be expressed by the comple- mentary term, the activation yield, % AY, i.e. the amount of carbon remaining after reaction, 84B. MCENANEY 85 where % BO = 100- % AY ; % BO is employed in this paper. Sorption isotherms on carbon particles of 20-40 BSS mesh were determined gravimetrically at 20.0"C in a flow system based on the method of Davie~.~ The carbon tetrachloride sorbate was prepared by fractional distillation of a commercial sample over phosphorus pentoxide ; the refractive index of the distilled sample was nio = 1.4606 in good agreement with the value given by Timmerman~.~ Desorption of carbon tetrachloride at PIPo = 0.3 to 0.0 resulted in an initial rapid weight loss followed by very slow desorption which continued for several days.It was, therefore, not possible to determine the equilibrium amount of carbon tetrachloride sorbed after desorption at PIPo = 0.3 to 0.0 in a practicable period of time. The value reported is the amount of carbon tetrachloride remaining after desorption for 24 h. Compressive fracture strengths were determined by the method of Ki~ling.~ Since the statistical variation of compressive strength for such carbons is quite large, the results are expressed as the mean and standard deviation of 25 determinations. It can be seen that for both cellulose and polyacrylonitrile carbons, compressive fracture strength decreases with increasing activation (table 1). TABLE 1 .-COMPRESSIVE FRACTURE STRENGTHS OF POLYMER CARBONS cellulose carbons ai/(MN m-2) polyacrylonitrile carbons m/(MN m-2) 25 14.8 (k3.8) 12 12.9 (k2.9) 45 10.7 (k3.0) 30 10.1 (k2.7) 60 9.8 (k2.9) 48 8.8 (k2.5) 75 8.7 (k2.6) 83 7.5 (k2.3) RESULTS Adsorption-desorption isotherms for activated cellulose and polyacrylonitrile carbons are shown in fig.1 and 2 respectively. The results for cellulose carbons are similar to those reported by Cadenhead and Everett for adsorption of benzene 1.2 1 .c n I M 3 M 0.8 9 W \ 04 0 *4 60% BO 45 % BO 25 % BO FIG. 1.-Sorption of carbon tetrachloride vapour on activated cellulose carbons at 20.0"C. 0, adsorption ; 0, desorption .86 LOW PRESSURE ADSORPTION HYSTERESIS vapour on coconut shell activated carbons. The occurrence of low-pressure hysteresis in these carbons appears to require the presence of micropores or micropore entrances which are similar in size to the adsorbate molecule.The low-pressure hysteresis found in the 25 % BO cellulose carbon may, therefore, be attributed to intercalation of the carbon tetrachloride molecule (diameter ca. 8 A) in commensurate micropores which molecular probe measurements have shown to be present. The molecular probe measurements also show that the principal effect of steam activation is to increase the mean micropore width to values greater than 12& the upper limit in the study. 48 % BO 30%BO I I I I 1 1 0-2 0.4 0-6 0.8 FIG. 2.-Sorption of carbon tetrachloride vapour on activated polyacrylonitrile carbons at 20.0"C. 0, adsorption ; e, desorption. PIP0 BO/ % FIG.3.-Adsorptive capacities for water vapour and carbon tetrachloride vapour at 20.0"C on cellulose (A) and polyacrylonitrile (B) carbons 0, HzO ; a, CCL.B. MCENANEY s7 The elimination of low pressure hysteresis in the 45 % BO and 60 BO cellulose carbons may be associated with this process. The closed hysteresis loops in the range PIPo = 0.35 to 1.0 which are found for the highly activated carbons may be attributed to reversible capillary condensation in transitional pores (meso-pores) which have been developed by the activation process. The results for activated polyacrylonitrile carbons (fig. 2) are similar to those for activated cellulose carbons except that, although low pressure hysteresis is reduced by activation, it is not entirely eliminated.This suggests that widening of micropores by steam-activation of polyacrylonitrile carbons does not occur to the same extent as in activation of cellulose carbons. This view is supported by a comparison of adsorptive capacities of the two activated series at PIPo = 1 .O for carbon tetrachloride and water vapour (fig. 3). In harmony with the previously reported density measurements,2 these results show evidence of molecular sieve action between these adsorbates in unactivated cellulose carbon, which is eliminated by activation beyond 25 % BO. Molecular sieve action is also found in the case of unactivated polyacrylonitrile carbon (fig. 3) but this effect, although reduced by activation, is not entirely eliminated until ca. 80 % BO. I' I I I I 24 f PIP0 FIG. 4.-Sorption of carbon tetrachloride vapour on unactivated cellulose carbons at 20.0"C.Open points, adsorption; closed points, desorption. 0, 0, heat treated to 900°C; A, A, heat treated to 2700°C. Sorption of carbon tetrachloride on unactivated cellulose and polyacrylonitrile carbons is of particular interest since comparison of adsorptive capacities for carbon tetrachloride and water vapour (fig. 3) indicates that the major part of the micropore volume is inaccessible to the larger molecule. In fig. 4 and 5 sorption on unactivated cellulose and polyacrylonitrile (900OC) carbons is compared to sorption on unactivated carbons which have been heat-treated to 2700°C in a graphite-resistance furnace. The effect of such heat-treatment has been shown by a variety of measurements to be a reduction in open microporosity and the development of a significant amount of closed microporosity in both carbons.For both cellulose and polyacrylonitrile carbons it can be seen that the amounts of carbon tetrachloride vapour adsorbed in the range PIPo = 0.0 to 0.9 on 900 and88 LOW PRESSURE ADSORPTION HYSTERESIS 2700°C carbons are similar (fig. 4 and 5). This result together with the observations that carbon tetrachloride is largely excluded from micropores in the case of the 900°C unactivated carbons (fig. 3) and that heat-treatment to 2700°C reduces micropore sizes,7 suggests that adsorption on unactivated carbons in the range PIP, = 0.0 to 0.9 is occurring on the surfaces of macropores and the external surface of the carbons. If this analysis is correct, then application of the B.E.T.equation to adsorption in the range P/Po = 0.0 to 0.6 gives surface areas available to carbon tetrachloride of 4.8 m2 g-l for unactivated cellulose carbons and 2.5 m2 g-1 for unactivated poly- acrylonitrile carbons. These surface areas have been obtained using a cross-sectional area for the carbon tetrachloride molecule of 32 A2, which was calculated assuming two-dimensional close-packing of the sorbate. I 1 I I I I I 1 0 *2 0 '4 0.6 0.8 PiPo FIG. 5.-Sorption of carbon tetrachloride vapour on unactivated polyacrylonitrile carbon at 20.0"C. Open points, adsorption ; closed points, desorption. 0, a, heat treated to 900°C ; A, A, heat treated to 2700°C. Fig. 4 and 5 show that for both cellulose and polyacrylonitrile carbons the amount of carbon tetrachloride retained by the 2700°C carbon after desorption to PIP, = 0.0 is less than that retained by the 900°C carbon.This effect may be associated with the reduction in the volume of micropores originally inaccessible to carbon tetrachloride produced by heat-treatment to 2700°C. For example, the micropore volume which is inaccessible to carbon tetrachloride has been shown to be 0.22ml/g for the 900°C cellulose carbon (ref. 2, table V) and 0.03 ml/g for the 2700°C cellulose carbon (ref. 7, table VI). From fig. 4, the amounts of carbon tetrachloride retained after desorption to PIPo = 0.0 are 0.01 1 ml/g (900°C carbon) and 0.004 ml/g (2700°C carbon). Thus, in the case of the 900°C unactivated cellulose carbon, about 5 % of the micropore volume originally inaccessible to carbon tetrachloride vapour is irreversibly penetrated by the intercalation process.A further feature of sorption on unactivated carbons is that the major part of the adsorbate retained after completion of the adsorption-desorption cycle was originally adsorbed at P/po>o.9, i.e. in large macropores and on the external surface of the carbons. Thus, adsorbate which is retained at P/P, = 0.0 in micropores which haveB. MCENANEY 89 been made accessible by expansion of the adsorbent was originally adsorbed in much wider pores. This observation also suggests that the process which caused low pressure hysteresis was initiated at high relative pressures ; this deduction would be consistent with the occurrence of pressure-threshold effects reported by other workers for low-pressure adsorption hysteresis.’ *- DISCUSSION Bailey et a2.l have given a general description of low-pressure adsorption hysteresis in which the phenomenon is associated with irreversible transitions from unperturbed to perturbed (expanded) states within domains in the solid adsorbent.They have suggested that such transitions may occur when local regions of the solid are strained beyond their elastic limit. In this paper the latter view of the process is developed by considering the application of the theory of adsorption-extension of Flood and Heyding.lo” If the carbons are considered as brittle materials containing micro- cracks (pores) of different sizes, then it may be proposed that the stresses induced in the solid by adsorption eventually cause localised fracture thus changing the structure of the adsorbent. Low-pressure hysteresis then occurs because micropore entrances which were previously inaccessible to adsorbate molecules are revealed and/or entrances to micropores filled with adsorbate are blocked by the localised fracture process. It has been shown above that only a small fraction ( 5 %) of the available closed porosity is penetrated by carbon tetrachloride in unactivated cellulose carbon indicating that the process may be confined to a relatively small proportion of the micropore volume.The stresses which cause low-pressure hysteresis are essentially those which cause dimensional changes in porous solids during adsorption. The lowering of the surface free energy of a solid which accompanies adsorption of a gas at pressure, P, causes dilational stresses.According to Flood and Heyding lo-’ the resultant strain, dV/Vt, is given by where K is the compressibility of the adsorbent, 4 = Va/Vc where Va and V, are adsorbate and adsorbent volumes respectively and K, is a structure factor. (Since carbons are not homogeneous solids, the distribution of stresses and strains throughout a carbon depends upon structure.) dV/V, = K $ K , ~ P (1) In the absence of directional stresses in the adsorbate CXP is given by : pa = spp”dP = aP (2) 0 P g where pa and pg are the densities of adsorbate and free gas respectively and pa is termed the mean volumetric stress intensity of the adsorbate. While dilation is the normal consequence of adsorption on a free surface, con- traction may occur in the case of porous adsorbents.If capillary condensation occurs in meso-(transitional) or macro-pores, contractional stresses may result from concave meniscus effects in the condensed adsorbate. In micropores (width < 10 A) contrac- tion has been attributed l1 to bridging of pores by adsorbate molecules. Since micropores fill at low relative pressures, dimensional changes in microporous carbons are characterised by an initial shrinkage followed by expansion at higher relative pressures. Thus at intermediate relative pressures there is presumably a complex distribution of compressive and dilational stresses in a microporous carbon. Adsorption of gases in microporous carbons is well described by the Polanyi potential theory of adsorption particularly if the adsorption temperature is much less90 LOW PRESSURE ADSORPTION HYSTERESIS than the critical temperature of the adsorbate.Adsorption of carbon tetrachloride on activated cellulose and polyacrylonitrile carbons obeys the Dubinin-Kaduskevich equation l4 : where the adsorption potential E = RTlog P,/P, V = volume adsorbed at P/Po, Vo is the micropore volume and p is a constant. Fig. 6 and 7 show typical isotherins plotted according to the linear form of eqn ( 3 ) from which V, may be estimated assuming that pa = p l , the density of the liquid adsorbate. Assuming that the potential theory of adsorption is appropriate and with the additional assumption of ideal behaviour by the free adsorbate gas, eqn (2) becomes V = V, exp( - fie2) (3) pa = (RTlog P)/V (4) where Vis the molar volume of the liquid-like adsorbate.Flood and Heyding lo have shown that for a volume element, Vt = V,+ V,, of the adsorbate-adsorbent system or where pc is the mean volumetric stress intensity of the adsorbent in the volume element. In this context, the volume element of Flood and Heyding l o may be equated to the domain of Bailey et a2.I V,P,+ VCP, = (Va+ VCP P c = P(1 + 4 > - 4 P a ( 5 ) I I I 1 0.4 0.8 1.2 1- log* POiP 60 7; BO 45 % BO 25 % BO FIG. 6.-Adsorption of carbon tetrachloride vapour on activated cellulose carbons at 20.0"C plotted according to the Dubinin-Radushkevich equation.B . MCENANEY 91 . 30% BO 12% BO log2 P,lP FIG. 7.-Adsorption of carbon tetrachloride vapour on activated polyacrylonitrile carbons at 20.0"C plotted according to the Dubinin-Radushkevich equation.The relatively small increases in adsorption of carbon tetrachloride at P/Po > 0.3 on the activated carbons (fig. 1 and 2) support the view that the majority of micropores are filled in these carbons at P/Po<o.3. Accordingly for microporous volume elements of carbons, 4 is effectively constant in the range P/Po = 0.3 to 1.0 and may be estimated from # = VOpHc where pHe is the helium density of the carbon corrected for adsorption.2* The values of 4 so calculated must be regarded as typical, since, in practice, a range of values of 4 would be expected in the various domains or volume elements of the solid adsorbent. If eqn (4) defines pa then for adsorption of carbon tetrachloride at 20°C in the range P/Po = 0.3 to 1.0 (P = 3.5 to 10.6 kN m-2) pa%P and thus from eqn ( 5 ) : Values of pc estimated from eqn (6) for adsorption of carbon tetrachloride on cellulose and polyacrylonitrile activated carbons are given in table 2, where it can be seen that pc increases with activation, reflecting the increase in 4.Externally applied compressive stresses, CJ, in porous carbons may be separated into stresses transmitted through adsorbate, CJ,, and adsorbent, nc, by an analogue of eqn (5). If it is assumed that stresses transmitted through adsorbate may be neglected under the conditions of the compression test on the carbons, then compressive fracture strengths, cf, (table 1) may be corrected for the presence of porosity by pc 21 -(+ RTlog P ) p .(6) c c = q ( l + 4). (7)92 LOW PRESSURE ADSORPTION HYSTERESIS Values of oc for the activated carbons are also included in table 2, where it can be seen that although of decreases with increasing activation (table 1) values of o - ~ are inde- pendent of the degree of activation within experimental uncertainty. TABLE 2.-CALCULATED STRESSES IN POLYMER CARBONS volume -pcKMNm-2) BO/( %) ratio, d* n,/(MN in-2) (PIP0 = 0.3) (PIP,, ~ 1.0) cellulose carbons 25 0.59 23.6 I22 138 45 1.04 21.7 214 243 75 1.79 24.3 3 69 41 9 60 1.44 23.9 297 337 pol yacryloni tri le carbons 12 0.07 13.8 14.3 16.1 30 0.14 11.5 28.2 32.0 83 0.77 13.3 157 179 48 0.21 10.7 42.3 47.9 *PHe = 1.9 8 ~ ~ l i - ' . ~ As previously mentioned, compressive stresses resulting from bridging of micro- pores may occur in carbons in addition to dilational stresses such as those calculated from eqn (6).Flood has shown l 2 that where directional stresses occur in an adsor- bate such as those resulting from bridging of micropores eqn (2) must be modified to and, applying the same assumptions as before, eqn (6) is modified to pc 21 - ( 4 RTlog P)/V+6$ (9) where, in this case, $ is the contribution to pa due to bridging potentials. It is expected that the Contribution of $ to pc will increase in the range P/P, = 0.0 to 0.3 as micropores fill but make a constant contribution to pc at higher relative pressures. The fact that volume contractions for microporous carbons are found at low adsorbate pressures suggests that $ > (RT log P ) / B in the early stages of adsorp- tion, and assuming this to be so, an estimate of @ may be obtained from measured volume contractions.Dacey and Evans l 3 have recently reported maximum volume contractions dV/Y, = for adsorption of benzene, methanol and water vapour on microporous Saran carbons which they attribute to bridging in micropores. For small volume contractions dV/V, = ti$' where $' is the bulk compressive stress of the adsorbate-adsorbent system due to bridging potentials and, using a value of the compressibility (K = m2 N-l) intermediate between those for single crystal graphite and diamond l5 a value of @' = lo8 N mA2 is obtained. The compressive stress calculated above and the dilational stresses estimated from eqn (6) may be compared with the values of 0, for cellulose and polyacrylonitrile carbons (table 2) and with the range of bulk fracture strengths found for carbons and graphites in general (from lo6 N m-2 in tension to lo8 N m-2 in compre~sion).l~-'~ It can be seen that the calculated stresses are either greater than or approximately equal to the bulk fracture strengths. Thus the proposition that low-pressure adsorp- tion hysteresis in microporous carbons is caused by localised fracture is consistentB.MCENANEY 93 with the finding that calculated compressive and dilational stresses induced in volume elements or domains in carbons by adsorption are commensurate with or greater than stresses which are found to cause bulk fracture. In additional support of this view, bulk fracture resulting from adsorption on poly(viny1idene chloride) carbon pellets which exhibit low-pressure hysteresis has been reported. Although localised fracture may occur as a result of adsorption, it does not follow that bulk fracture will always occur since the work expended in localised fracture in volume elements or domains within the carbon may reduce the stresses induced by adsorption to a value below the minimuin fracture stress for further crack propagation.Additionally, it may be necessary for adsorption-induced stresses in adjacent domains to act co-operatively before bulk fracture can occur. It has been noted that the extent of low-pressure hysteresis decreases with increas- ing activation for cellulose and polyacrylonitrile carbons (fig. 1 and 2) although, with apparent lack of correlation, the calculated values of adsorption-induced stress, pc, increase with activation (table 2). This is because low-pressure hysteresis is only observed if localised fracture reveals microporosity which was previously inaccessible to the adsorbate molecule and, as previously noted for cellulose carbons, microporosity which was inaccessible to the carbon tetrachloride molecule is progressively reduced by activation.2 CONCLUSIONS It is concluded that the theory of adsorption-extension of Flood and Heyding offers an interpretation of low-pressure adsorption hysteresis. Application of the theory to sorption of carbon tetrachloride in microporous polymer carbons has shown that calculated stresses in carbons induced by adsorption are commensurate with experimental bulk fracture strengths.The theory appears capable of accounting for pressure and temperature threshold effects associated with low-pressure hysteresis, since, from eqn (6), the adsorption-induced stress, pc, is directly ,proportional to temperature and the logarithm of pressure. The apparently anomalous adsorption effects which Bailey et al.' have attributed to collapse of pores due to compressive stresses may be understood qualitatively in terms of eqn (9) when $%(RTlogP)/F Further work is necessary to determine the extent to which the theory of adsorption- extension can be quantitatively applied to low-pressure adsorption hysteresis. For example, further theoretical work is required to describe the development of compres- sive stresses due to bridging of micropores by adsorbate molecules and experimental work is required to test the ability of the theory to accurately describe presssure and temperature threshold effects.NOTATION dV/Vt volumetric strain induced by Pa,pc K compressibility of the adsorbent Va, Vc, Vt volume of adsorbate, adsorbent adsorption (m2 N-l) and adsorbate+ adsorbent res- pa, pe, pl pectively (ml g-l) = VJV, PHe a dimensionless structure factor P, Po vapour pressure and saturated vapour pressure of the adsorbate, respectively (N m-2) a 4 Ks for porous adsorbents v V mean volumetric stress intensity of the adsorbate and adsorbent, respectively induced by adsorp- tion (N m-') density of adsorbate, gas and liquid respectively (g ml-l) helium density of the adsorbent (g ml-l) molar volunie of the adsorbate (ml mol-l) volume of micropores filled at PIP0 (ml g-9 = Pa/P94 LOW PRESSURE ADSORPTION HYSTERESIS N 0 TAT I 0 N-continued vo micropore volume of the adsorb- oar oC, 0 volumetric stress intensity of ent (ml g-l) adsorbate, adsorbent and adsorb- constant of the Dubinin-Radush- ate+ adsorbent respectively due kevich equation (mol" J-") to externally applied stresses contribution to pa due to bridging compressive fracture stress of the potentials in micropores (N m-") adsorbent (N m-").bulk compressive stress due to bridging potentials (N m-") P $ $' & adsorption potential (J moI-') (N m-2) of A. Bailey, D. A. Cadenhead, D. H. Davies, D. H. Everett and A. J. Miles, Trans. Faraday Soc., 1971, 67, 231. J. J . Kipling and B. McEnaney, Proceedings of the 2nd Conferelwe on Industrial Carbon and Graphite (SOC. Chem. Ind., London, 1966), p. 380. R. G. Davies, Chem. and Ind., 1952, 160. J. Timmermans, Physico-chemical Constants of Pure Organic Compounds (Elsevier, London, 1965), vol. 2, p. 185. J. J. Kipling, Fuel, 1950, 29, 42. D. A. Cadenhead and D. H. Everett, Proceedings of the 1st Conference on Industrial Carbon and Graphite (SOC. Chem. Ind., London, 1958), p. 272. ' J. J. Kipling, J. N. Sherwood, P. V. Shooter and N. R. Thomson, Carbon, 1964, 1, 321. * H . L. McDermott and J. C Arnell, Canad J. Chem., 1955,33,913. H. L. McDermott and B. E. Lawton, Canad. J. Chem., 1956,34,769. A. E. Flood, Canad. J. Chem., 1957,35,48. l o A. E. Flood and R. P. Heyding, Canad. J. Chem., 1954,32,660. l 2 A. E. Flood, Proceedings of the 4th Conference on Carbon (Pergamon, New York, 1960), p. 3. l3 J. R. Dacey and M. J. B. Evans, Carbon, 1971,9, 579. I4 M . M. Dubinin, Chemistry andPhysics of Carbon, ed. P. L. Walker Jr. (Arnold, London, 1966), vol. 2, p. 51. W. N. Reynolds, Physical Properties of Graphite (Elsevier, London, 1968), p. 35. l6 R. E. Nightingale, Nuclear Graphite (Academic Press, New York, 1962), p. 150. C. L. Mantell, Carbon and Graphite Handbook (Interscience, New York 1968 ), p. 24.

ISSN:0300-9599    

DOI:10.1039/F19747000084

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Light Scattering from the Free Surface of Water BY D. LANGEVIK Ecole Normale SupQieure Laboratoire de Physique, 24 Rue Lhomond, Pa.ris(Se), France Received 21st June, 1973 The spectrum of light scattered by thermally excited surface waves on water has been investigated and the value of the surface tension and, after careful analysis of instrumental broadening, the value of the viscosity determined. The results are in good agreement with those obtained by more con- ventional techniques. In contrast to a recent report by McQueen and Lundstrom (ref. (1)) who also performed a light scattering experiment, an anomalously high viscosity was not found. These authors invoked the concept of a highly structured surface region, in order to interpret their data. This region probably does not exist.In recent years, light scattering techniques have been used to study liquid inter- faces. Several theoretical and experimental papers have already been published on this ~ubject.l-~ The mechanism of scattering may be described as follows: thermal motion produces small asperities (about lOA high) on the liquid interface. The interface is submitted to a capillary restoring force and its motion is damped by viscosity (one can neglect gravity forces). Spatial and temporal evolution of the fluctuations can be described on a macroscopic scale by the laws of hydro- dynamics. Consider a surface vibration mode, or surface phonon, with a given wave vector q. The scattering process is simply : Scattered light is concentrated in a well defined direction, simply related to q by expressing momentum conservation in the surface plane, during the scattering process.Our experimental procedure, which involves heterodyne spectroscopy techniques, has already been de~cribed.~ We are presently interested in the study of water covered by monomolecular films and have computed the theoretical spectrum of the light scattered by such films.5 This spectrum depends on the surface tension, on the viscosity of pure water and on four viscoelastic parameters of the film : surface pressure, compression modulus and their two associated viscosities. We report here preliminary experiments on light scattering from the free surface of pure water. In this case the theoretical spectrum depends only on two parameters : the surface tension CT and the viscosity 71 of the liquid.However, the experimental spectra are broadened by instrumental effects resulting from the finite angular resolution of the optical system. The main contribution to the instrumental broaden- ing is the divergence of the laser beam wising from diffraction. The shape and the width of the experimental spectra differ slightly from the theoretical predictions, when using conventional values of viscosity.6 By studying another liquid, ethyl alcohol, and assuming that the viscosity is equal to its conventional value, we found that the instrumental function can be satisfactorily described by a Voigt function.’ The Voigt functions belong to a family of functions depending on one parameter a. They incident photon surface phonon 3 scattered photon.9596 LIGHT SCATTERING AT LIQUID SURFACES include Gaussian (a = 0) and Lorentzian (a = co) functions as special cases. For the instrumental function, we found a = 0.3 (fig. 1, curve 1). Using this instrumental function, we deduced from the line-shape and from the width of the experimental spectra for water, values of the viscosity in good agreement with the conventional value, and of the instrumental line-width, in good agreement with estimated values deduced from the angular divergence of the laser beam arising from diffraction. - 3 - 2 - 1 LJ - vsolALJ1 FIG. 1.-Instrumental functions: (1) Voigt function of parameter a1 = 0.3; (2) square pulse; (3) diffraction function by a circular aperture. Several workers have already studied light scattering by the free surface of water.Mann et aL8 studied mechanically excited surface waves because their signal to noise ratio was not sufficiently good to detect thermally excited waves. They observed two peaks in their experimental spectra and interpreted these strange results using an incorrect dispersion equation, the roots of which are not complex conjugate. More recently, McQueen and Lundstrom studied thermally excited surface waves on water with an experimental arrangement similar to ours. They found the same value of surface tension, though with a higher uncertainty (three times higher than ours), but a different viscosity value : rj = 4 cP. They used an instrumental function having the shape of a square pulse (fig. 1 , curve 2). We have used this second instrumental function also, and have deduced, from our experimental spectra on water, viscosity values lower than 4 CP but higher than 1 cP, which decrease when the scattering angle increases.This clearly indicates that the instrumental broadening is not well corrected for by their method. McQueen and Lundstrom invoked the concept of a highly ordered water surface region in order to interpret their data. It follows that the exis- tence of such a surface zone now appears unlikely. Accordingly this concept is re- j ected. In the following, we describe the experimental conditions, discuss the procedure used for the instrumental broadening correction, and present our experimental results for water at 21°C.D. LANGEVIN 97 EXPERIMENTAL We did not observe any difference between the spectra obtained using com- mercially available bidistilled water and extremely pure water tridistilled in a quartz apparatus.Light from a helium-neon laser (approximatively 60 mW output power) is incident from above the liquid surface at an angle of 3". We select with a diaphragm light scattered by fluctu- ations having a well defined wave vector q. This diaphragm has the shape of a circular annulus.* If 6 is the angle between scattered and reflected beams one has, k, being the wave vector of the incident light : The experimental equipment is described else~here.~ The light scattered by the fluctuations is collected on a photomultiplier tube which also receives light scattered elastically by the windows of the liquid container which thus fulfils the role of a local oscillator for optical heterodyning. In such conditions, it can be shown that the photocurrent power spectrum reflects exactly the power spectrum for the fluctuations of the vertical displacement of the free surface.The photocurrent is amplified, squared, frequency-analysed and recorded. The theoretical spectrum for the free surface of a simple liquid is given by :2 1 D(S) = 0 being the dispersion equation for surface waves : D(S) = y+(1 +s>'-JiTzs and introducing the following parameters : P , z o = - y = - OP 4q2q 2w2 v is the frequency, p the density, 0 the surface tension and q the viscosity; J is the square root determination having a positive real part. When the liquid viscosity is low, as in the case of water, the parameter y is much greater than unity and the roots of the dispersion equation, which are complex conjugates, can be approximated by S , = s;* = iyf [ 1+0 (31 a - [ 1 - $i +o($)] - Pg(v) takes significant values only when v z yh/2nzO. spectrum can be approximated by For these frequencies the 1 * It can be shown that for small incidence angles, the scattering is symmetric around the reflected beam and that a circular annulus selects light scattered with a wave vector q constant in modulus (its direction varies, but the spectrum does not depend on this direction).For higher incidence angles, this is no longer valid. For example, with an incidence angle of 45" and a circular annulus, one detects a finite range of wave vectors : 1-498 LIGHT SCATTERING AT LIQUID SURFACES The spectrum is Lorentzian, centred at the frequency v, = ISq1/2nzo and of half width Avq = -Re Sq/271z0.It follows The scattered intensity varies as q-2 : kT 271 P,(v) dv = -2. s aq One is thus restricted to small scattering angles, i.e., to s.mal1 wave vectors q. In the present experiment 153 cm-l < q < 722 cm-1 which corresponds to 5' < 8 < 25'. For larger scattering angles, the signal to noise ratio becomes very smdl and the interpretation of the spectra is difficult. In order to reduce mechanical vibrations, the optics and the sample were placed on a heavy (one ton) table, which rested on five rubber tyres. INSTRUMENTAL BROADENING CORRECTION When surface waves are not strongly damped (as is the case for water) experimental spectra are broadened by important instrumental effects.Several causes contribute to the instrumental width. (a) The finite aperture of the diaphragm, which therefore selects wave vectors inside a finite range qo & Aq, qo corresponding to the mean radius of the diaphragm. In our experiment, typically Aq/qo - 10 %. The experimental spectrum is then If Aq/qo is small enough, q/vq which is proportional to q-+ varies little with q within the range A4. If we neglect also the variation of Ava with q, the observed spectrum is simply the convolution of Pq(v) with an instrumental function having the shape of a square pulse (fig. 1, curve 2), of half width : vqo. Av, = - - 3 A4 2 40 We can easily compute P(v) with this approximation : when The half width of P(v) is then : AV = J(Av," +AvT). (b) The divergence of the laser beam due to diffraction is also a cause of imperfect definition of q.In our experimental set-up, the reflected beam is focused at the diaphragm centre, and has at this point a diameter roughly equal to the diaphragm aperture. Taking into account the diffraction effect only, the instrumental function would be close to the diffraction function for a spherical aperture (fig. 1, curve 3).D. LANGEVIN 99 (c) Mechanical vibrations cause a small random displacement of the reflected beam, and therefore add a small contribution to Aq. This effect modifies the instrumental width AvI, and probably also the shape of the instruniental function. As the results will show, this effect is relatively small. (d) Finally, the electronic system used for detection might contribute to the instrumental broadening.But in the present experiment this contribution is negligible. These different points were tested using experimental spectra obtained by light scattering from the free surface of ethyl alcohol, We choose this liquid since the damping conditions of surface waves were similar to the water case. Moreover, intermolecular forces and therefore surface zone organisation (if this zone exists) are expected to be different. We first fitted experimental spectra for ethyl alcohol with eqn (6). The viscosities we then deduced from eqn (4) and (7) were higher than the values obtained by conventional methods (y = 1.18 CP at 21°C). Moreover they decreased when q increased. We used Voigt f~nctions,~ which have interesting properties for the present problem : they form a one parameter dependent family of functions, including Gaussian and Lorentzian functions, and the convolution of a Lorentzian function with any Voigt function is another Voigt function.The experimental spectrum. is the convolution of the Lorentzian theoretical spectrum (eqn (2)) with the instrumental Voigt function : it is therefore a Voigt function having a shape characterized by the parameter a. We fitted every experimental spectrum with a Voigt function characterized by a parameter a. Assuming that Ava has the value obtained from eqn (4) using the conventional value of viscosity, we deduced the parameter ai for the instrumental Voigt function and the half width Av,, from the following set of equations :7 We then tried to find a different but still simple instrumental function.Tables for functions c(aI) and P(Av/AvI, al) are given in ref. (7). The instrumental function is in each case close to a Voigt function having a, = 0.30. This function is represented in fig. 1 (curve 1). It is closer to the diffraction function than to the square pulse. This indicates that the instrumental broadening arises mainly from diffraction. In the following, when dealing with the interpretation of data corresponding to the free surface of water, we assume that aI is still equal to 0.30 and allow Avl to vary in order to take into account the effect of mechanical vibrations. The parameter a, which characterizes the shape of an experimental spectrum, depends on q since instru- mental broadening is more important at small q.As q increases, the experimental spectra become closer to Lorentzian spectra, so a increases. We see also that there is a certain dispersion of a values deduced from different spectra for a given q : this comes from the random broadening effect of vibrations. Finally, from the width of the experimental spectrum Av and its shape characterized by the parameter a, we deduce Avq and A+ using eqn (8) and (9). RESULTS WATER AT21'C Measurements of the position of the peak of the spectra v,, and of their half widths Av are presented in table 1, and in fig. 2 using a logarithmic scale.c 0 0 q/cm- 1 153 240 304 366 469 615 772 vq/Hz 2540 2560 5030 5000 7200 7230 7300 9610 9600 9400 9460 9480 13600 13700 13400 13500 20200 28400 Av/Hz 319 362 500 500 880 770 660 727 710 796 853 810 1210 1210 1170 1060 1820 2100 Y 1220 786 614 509 397 308 242 TABLE WATER AT 21°C a/p/(c.g.s.units) (q/p)aa/(c.g.s. units) a 71.5 72.8 72.4 72 73.2 73.5 74.9 74.6 72.8 71.3 72.4 72.6 71.2 72.3 69.4 70.6 69.5 69.7 4.68 5.32 3 3.1 3.33 2.91 2.5 1.9 1.86 2.09 2.23 2.12 1.95 1.95 1.88 1.71 1.72 1.27 0.5 0.5 0.8 0.8 0.75 0.8 0.85 1.3 1.2 1.25 1.1 1.05 1.1 1.3 282 320 373 373 676 575 482 422 434 475 545 534 750 618 2Aq/cm- 1 22.8 25.6 23.8 23.7 38 32.2 26.8 21.3 22.2 24.8 28 27.4 35 28.6 Avq/HZ 56.4 54 186 186 304 288 275 422 3 89 452 43 6 400 600 618 E: q/p/(c.g.s. units) 0 3: 0.83 cl 0.94 1.11 > w 0 cl 1.11 H 1.15 m z 1.09 E 1.04 0 1.10 > 1.02 cl 1.18 z 1.14 rQ 5 1.04 UD. LANGEVIN 101 Taking the values obtained by conventional techniques at 21°C : (r = 72.6 dyn/ cm, q = 0.981 cP, p = 0.998 g/cni3, we computed the values of the parameter Y = 0p/4y2q.Using these y values,* we deduced from v, values and eqn (3), the quantity alp, and from Av values and eqn (4), a quantity ( r , ~ I p ) ~ ~ which would be equal to q / p if there were no instrumental broadening : Av = Av,. We obtain : o/p = (72$.2) c.g.S. units which agrees with conventional values. On the other hand ( ~ / p ) ~ ~ differs from the conventional value; it decreases rapidly as q increases and approaches 1 cgs. unit for high q. q/cnl-' FIG. 2.-Results of measurements of peak frequencies vq and half widths Av for spectra recorded on water at 21"C( +). Values of corrected half widths Av,. (0) Straight lines are the theoretical curves for vq and Avq obtained from conventional values of o, 11 and p.We fitted all the spectra with Voigt functions. Values of a axe shown in table 1 together with values of AvQ and AvI obtained from eqn (8) and (9); Aq values were deduced from AvI using eqn (5) and q / p values from Av4 values using eqn (4). For a given q, there is a lasge scatter in Av, a, and AvI values (precision of Av values is 10 Hz and of a values k0.05) this arises from the random effect of vibrations. We obtain u/p = (l.OS+O.lO) c.g.s. units which agrees with conventional values. Notice that the precision for q/p (10 %) is lower than for alp (3 %) because instrumental broadening is very important and * This procedure is justified so long as these values of y are used to compute relatively small corrections.102 LIGHT SCATTERING AT LIQUID SURFACES q/p depends in it crucial way on the spectral shape : frequency peak measurements are more precise than spectral shape measurements. Moreover the instrumental function is probably affected by vibrations and aI may not be rigorously constant.On the other hand we see from table 1 that 2Aq = 30+8 cm-1 and deduce from eqn (1) that the corresponding angular divergence of the laser beam is A0 = 2Aq/ko N 3 x rad, approximately equal to the divergence of the beam caused by diffraction.* Experimental broadening could not be corrected for several large wave vectors (see table 1) because the signal to noise ratio was low, and differences between a Voigt function and a Lorentzian become small when a increases.Fig. 3 and 4 represent typical spectra corresponding to wave vectors q = 240 cm-l and q = 469 cm-l. v/kHz FIG. 3.-Experimental spectrum of light scattered by the free surface of water, recorded at 21°C for a wave vector q = 240 cm-'. The solid line is the result of a best fit performed with a Voigt function having parameter a = 0.8. The dashed curve is a Lorentz curve. The dotted curve is the result of a fit with the function of eqn (6) where Avq has been computed using T i p = 0.983 c.g.s. units. The corresponding Voigt functions which fit these best have values of a = 0.8 and a = 1.3 respectively. The dashed line curve represents the Lorentzian (a = 03) of identical half width Av. The second spectrum (fig. 4) is closer to the Lorentzian curve : experimental broadening is 1ower.t * The laser beam is diffracted by the lenses and the mirrors of the optical set-up.The main angular divergence is produced by the first lens where the beam has its smallest diameter d, which is about 2.5 mm. The corresponding angular divergence is : Aediff = 1.22 Ald - 3 x rad. t Experimental spectra are not symmetrical with respect to the frequency v = vq because the theoretical spectrum is slightly different from a Lorentzian spectrum (eqn (2)), especially near zero frequency : it has values about four times higher at zero frequency. The differences become negligible only for v > v4- 2Av. To avoid difficult computations we restricted ourselves to this frequency range.D. LANGEVIN 103 Lorentzian function 0 13 15 v/kHz FIG.4.-Spectrum of light scattered by the free surface of water at 21°C for q = 469 cm-l. solid line is the result of a best fit performed with a Voigt function having parameter a = 1.3. The The dashed curve is a Lorentz curve. We have also used the instrumental broadening correction explained in the preceding paragraph and fitted the curve of fig. 3 (q = 240 cm-l) with eqn (6). The best fit corresponds to AvJAv, = 1.2. Using eqn (7) and (4) one can deduce y/p = 1.92 c.g.s. units. Note that this y/p value is higher than ( ~ / p ) ~ ~ for larger wave vectors, and a fortiori higher than y / p values obtained after experimental broadening correction for these wave vectors. On the other hand, if we assume that q/p is equal to the conventional value, we obtain for the spectrum of fig.2: AvI = 3 Av4 = 492 Hz almost equal to the true half-width Av. The corresponding spectrum is shown in fig. 2 (dotted line); as can be seen, the corresponding fit is not satisfactory. McQueen and Lundstrom found, for wave vectors in the range : 353 < q < 707 cm-l that a constant viscosity q = 4 CP can account for their measurements. They invoked, therefore, the existence of a highly structured surface region. If this were true, we should similarly have obtained a value = 4cP had we adopted their method of correction. But we found different values, which decrease with q toward the conventional value q = 1 cP. This trend was observed for all our samples ; instrumental broadening is therefore not well corrected for by this method. Since completion of this work, McQueen and Lundstrom have indicated in a per- sonal communication that they have become aware of the probability that their estimates of the instrumental corrections are incorrect.At the present time they do not believe a highly structured surface region is necessary to explain their data. Av4 = 320 Hz CONCLUSION The spectrum of light scattered by thermally excited surface waves on water has been examined. We found the value of surface tension and, after a careful instrumen- tal correction, the value of the viscosity. These results agree well with those obtained by more conventional techniques. We used a broadening correction method, involving Voigt functions which are a generalisation of Gaussian and Lorentzian104 LIGHT SCATTERING AT LIQUID SURFACES functions.This method has already given satisfactory results with ethyl alcohol. We believe that the reason why it is appropriate is that the main contribution to instrumental broadening is the diffraction of the laser beam. The correction procedure of McQueen and Lundstrom leads to estimates of viscosity which, in our case, vary with scattering angle and which are greater than conventional values. These are however lower than estimates of McQueen and Lundstrom, which apparently do not depend on scattering angle. Such discrepancies clearly show that this type of broadening correction is not adequate. We conclude that it is not necessary to invoke the concept of a high viscosity surface zone to interpret the experimental results. D. McQueen and I. Lundstrom, J.C.S. Faraday I, 1973,69,694. M. A. Bouchiat and J. Meunier, J. Physique, 1971, 32, 561. M. A. Bouchiat and J. Meunier in Polarisation Mufigre et Rayonnernent, p. 121 Volume Jubilaire en l’honneur #Alfred Kastler, ed. by Soci6tk FranGaise de Physique, P.U.F. Paris (1969). For a review of some recent experimental work on light scattering by liquid interfaces, see for example, Supplement au J. Physique, 33. Fasc 2-3 C1(1972), especially the following papers : J. Zollweg, G. Hawkins, I. W. Smith, M. Giglio and G. Benedek, p. C1-135 ; M. A. Bouchiat and J. Meunier, (21-141 ; E. S. Wu and W. W. Webb, C1-149 ; D. Langevin and M. A. Bouchiat, C1-77. For a description of the experimental equipment see, for instance, the first reference in (3). D. Langevin and M. A. Bouchiat, Compt. rend., 1971,272, 1422. Handbook of Chemistry and Physics (Chemical Rubber publishing company, Cleveland, Ohio, 45th edn., 1964.) ’ H. W. Leidecker, Jr. and J. T. Lamacchia, J. Acoust. Soc. Amer., 1967, 43, 143. * J. A. Mann, J. F. Baret, F. J. Dechow and R. S. Hansen, J. Colloid Interface Sci., 1971,37,14. J. Meunier, Thesis (Paris, 1971).

ISSN:0300-9599    

DOI:10.1039/F19747000095

出版商:RSC    

年代:1974

数据来源: RSC