J. Chem. SOC., Faraday Trans. I , 1985,81, 69-82 Adsorption and Conductivity Studies in Oxychlorination Catalysis Part 4.-Effect of Adsorption on the Conductivity of Copper(1) Chloride Films BY PETER G. HALL, RICHARD A. H A N N , ~ PHILIP HEATON f AND DAVID R. ROSSEINSKY* Department of Chemistry, University of Exeter, Stocker Road, Exeter EX4 4QD Received 12th March, 1984 Thin copper(1) chloride films on glass have been prepared, and the effects of adsorbed nitrogen, oxygen and ethylene on conductivity have been studied using special electrode configurations to maximise the effects. Compositions have been examined by secondary-ion mass spectrometry (SIMS) and X-ray photoelectron spectroscopy (XPS), porosity by scanning electron micrography (SEM) and net reaction by microreactor studies.The conductor is electronic, the 0.53 eV activation energy agreeing with literature values. Adsorbed ethylene gives an irreversible decrease in conductivity, oxygen a reversible one, while nitrogen, physisorbed, is ineffective. The constancy of activation energy despite the adsorption of gases implies that adsorption results mainly in fewer charge carriers rather than a change in mechanism. From the conductivity results oxygen is reversibly chemisorbed, at higher temperatures possibly becoming included within the lattice. In ~tudiesl-~ of oxychlorination catalysis, interactions of absorbed ethylene and other gases with copper(1, 11) chloride and similar systems have been examined. This work is now extended to the effects of adsorption on the conductivity of copper(1) chloride in film form.The active-site formulation of adsorption proposed by Taylor4 and elaboratored by Kevan5 and Boudart6 has led on to the view that catalytic and semiconductor properties are intimately related.:-" Concomitantly, ideas of charge transfer in catalysisl2* l3 have been examined1"-': with but tenuous experimental support until re~ent1y.l~- l9 Adsorption can change the work function of crystalline semiconductors20 and the consequent alteration of the contiguous space-charge region is expected to affect the magnitude of the surface conductivity.21- 22 The only marginal effect of ethylene adsorption on the conductivity of bulk powder compactions of copper(1, 11) ~hlorides,~ has led us now to examine adsorption effects on thin films of CuCl, where the surface/bulk ratio is greatly enhanced.Adsorbates were nitrogen, ethylene, water and oxygen, and their effects on the d.c. conductivity and its time and temperature dependence were observed. Characterisation of the solid was effected by scanning electron microscopy while chemical composition was established by SIMS and XPS. The interaction of CuCl with C1, to give CuIl-doped material was studied by Harrison and ~ o w o r k e r s . ~ ~ - ~ ~ The temperature dependence of CuCl conducti- 2 7 , 29--36 has been variously depicted as either curved or bilinear Arrhenius behaviour. Some theoretical studies have been attem~ted.~' 42 The dependence on potential and pressure has been studied ;43 high pressure introduces metallic c o n d u ~ t i o n ~ ~ ~ 44-49 and the possibility of s~perconduction.~~~ 51 Single crystals t Present address : Electronics Group, ICI, Runcorn.1 Present address: Johnson Matthey Research Centre, Blounts Court, Sonning Common, Reading. 6970 OXYCHLORINATION CATALYSIS have29v44v 4 5 9 5 2 9 5 3 conductivities o = 2 x to lo-* C2-l cm-l at ca. 300 K, which have been compared with compaction values.3o CuCl as solid ele~trolyte~~ (or doped semiconductor) has been examined.2g+ 55-59 There is scope for debateso* as to whether band conduction or hopping predominates. For CuCl and doped CuCl no simple Cole-Cole dielectric dispersion characteristic of pure hoppings2 is observable, and the band form~lation~~ seems the more tenable. In mixed valence, e.g. CuII-doped CuCl, systems, the Fermi level, conductivity and adsorption properties will be determined by the valence compo~ition.~~ EXPERIMENTAL SAMPLE PREPARATION AND MOUNTING Copper(1) chloride, prepared as before,23 was purified using a vacuum depositer (Creative Vacuum Services, Worthing, Sussex).The white powder was placed in a molybdenum sample boat and brought up to deposition temperature at CQ. mmHg,* so that it just began to deposit on glass slides placed above it. The evaporating/heating current was removed and a solid plug of CuCl was left, free of volatile impurities. A frame supporting four clean microscope slides was placed ca. 3 in. above the plug of CuCl in a sample boat. The ends of the slides were masked and the chamber was evacuated to 8 x mmHg. Current was passed through the sample boat via a Variac until visible melting began.The evaporating current was then kept constant to prevent vaporisation of any impurities with higher melting points. A white film was allowed to form slowly on the slides, until opaqueness indicated a continuous layer. For a completely opaque layer depositions from several plugs of CuCl were required. Gold electrodes were then deposited on the surface, using a mask which left a strip (1 mm or 2.5 cm wide) of copper(r) chloride free of gold [fig. 1 a)]. An alternative electrode configuration was also used where platinum was deposited on a microscope slide first. Then a discontinuity was scored across the middle of the slide, before copper(1) chloride was deposited over the platinum [fig. l(b)]. Coated slides were kept under vacuum, rapidly transferred to a nitrogen-filled dry box, dried over P20,, and mounted in a cell with 4 painted leads of gold wire.The evacuated mmHg) cell was immersed in a water thermostat. Alternatively, for high temperatures, the sample side was transferred to a stainless-steel cell where it was mounted with silver paint onto gold wire. The stainless-steel shell was closed by means of a rubber or metal O-ring and evacuated to lod5 mmHg. A close-fitting furnace allowed the sample to be thermostatted over the range 25-400 "C, as monitored by a chrome-alumel thermocouple positioned inside the cell. CONDUCTIVITY MEASUREMENTS AND GAS DOSING Direct-current measurements were taken using either a polarograph (Bruker, model EMS) with an ( X , Y) chart recorder, or a voltage supply (DEB Electronics Ltd) with an ammeter (Keithley Instruments 610C electrometer), giving currents generally of the order of A.Between readings the sample was shorted to avoid any build up of charge. A potential range of k 3 V was commonly used. Known pressures of gas (nitrogen, oxygen or ethylene) were connected to the high-temperature conductivity cell, to a pressure of 0.57 atmt (chosen so as not to exceed the seal specification). For dosing with H20 a small glass tube was filled with glass wool and ca. 2 cm3 of added distilled water was degassed by a freeze-thaw cycle under vacuum, repeated twice. The sample was then opened to the manifold and subject to the saturated vapour pressure of water at ca. 17 "C. SECONDARY-ION MASS SPECTROMETRY AND OTHER TECHNIQUES SIMS was performed by Dr J.Myatt at the ICI Research Laboratories using a Riber SIMS instrument. The mass spectrometer gave both a negative-ion and a positive-ion spectrum. CuCl * 1 mmHg x 133.322387 Pa. t 1 atm = 101 325 Pa.P. G. HALL, R. A. HANN, P. HEATON copper (I) chloride microscope slide AND D . R. ROSSEINSKY copper (1) chloride < d i scon t inu i t y microscope platinum sIide (6) 71 Fig. 1. (a) Gold/copper(I) chloride substrate/gold configuration. (b) Substrate Pt/superposed CuCl/substrate Pt configuration. was deposited on 12 mm diameter glass discs to provide samples for this technique. It was possible to vary the sample temperature and dose with a gas (oxygen) within the machine. All spectra were necessarily recorded under ultra-high vacuum (ca.1 0-lo Torr).* X-ray photoelectron spectroscopy (XPS) was performed at the Analytical Laboratory, ICI on a Kratos ES200 instrument. Samples of CuCl on glass rectangles of size 2 cm x 0.5 cm were analysed and exposed to oxygen at various temperatures within the machine. Scanning electron microscopy was performed on a Cambridge S600 microscope. The microreactor used in Part 33 was used to study the bulk reaction of CuCl with 0,. Dry helium was the carrier gas and temperature thermostatting was effected by a chromatography oven. Conventional mass spectra were obtained on a VG Micromass MM16F spectrometer. RESULTS AND DISCUSSION Using the low-temperature glass cell the conductance, R-l, of a film of CuCl with platinum electrodes [fig.l(b)] was measured from the time of mounting and evacuation. The result is shown in fig. 2, with an initial strong decrease in R-l with time, but stabilisation to within an order of magnitude after several days. A similar decrease was observed for the alternative [fig. 1 (a)] configuration. This decrease was attributed to annealing rather than an outgassing process. After 1 month the CuCl layer was shown by electron micrography of a step where part of a layer had broken away from on top of another, exposing a layer cross-section, to be a very ordered microcrystalline array. The deposited layer was 5-6 pm thick, and commonly total thicknesses were estimated to lie between 1 and 20 pm. Such film, being on top of the electrodes, was not very robust and became partially detached from the metal on first exposure to vacuum.Therefore in subsequent experiments use of the configuration shown in fig. 1 (b) was favoured and care was taken to leave the sample for several (2-3) days before conducting experiments. The linearity or otherwise of the voltage against current plot produced by the chart recorder indicates the deviation from Ohm's law of a sample, depending on the vacuum. Initial vacua of only ca. lop4 mmHg resulted in non-Ohmic behaviour. At high voltages (> 10 V) the white sample blackened and the resistance suddenly decreased by several orders of magnitude, similar observation~~~ being attributed to a phase change. However, a perfectly Ohmic response was obtained for vacua better than mmHg at potentials + 3 v. At near-to-laboratory temperatures with a 15 min equilibration time at each temperature, values of E, = 0.46 eV for ascending temperature (1 3-39 "C) and 0.64 eV for descending temperature were obtained.The discrepancy is not unexpected as many solids are ' structure-sensitive' in this temperature range owing to imperfect structure * 1 Torr = 101 325/760 Pa.72 - 3 r OXYCHLORINATION CATALYSIS - 9 I 1 I I I 13 26 39 52 tlh Fig. 2. Change of log(conductance) with time for configuration of fig. 1 (b). or lack of Extending the temperature range with an icebath gave a linear graph of In R-l against 1/T [see fig. 3(a)], with E, = 0.53 eV, between the extremes. The high-temperature cell allowed measurements over a wider range of temperature. 45 min for temperature stabilisation and 15 min for equilibration were allowed.Some deterioration of vacuum from the usual mmHg occurred at > 200 "C, but no effect on E, was discovered. This was later traced to a nitrogen-producing organic impurity which fortunately played no further discernible role. From fig. 3(b) an Ea value of ca. 0.5 eV is obtained and compared (table 1) with literature values, falling well within the upper group. When two activation energies for conduction are reported in the literature, that with E, z 1 eV above 180 "C is attributed to ionic conduction (cationic conduction by a Frenkel defect mechanism), while that with Ea z 0.5 eV below 140 "C is attributed to hole conduction, allowing27, 30 an estimate of the percentage ionic conduction at any temperature. No upturn in E, above 180 "C is seen.Harrison and PrasadZ7 report that the electrode material caused the d.c. values above 190 "C to fall by 40 and 80% below the a.c. value for Ag and Pt electrodes, respectively, and doubtless our Au electrodes also suppress ionic carriers and the consequent upturn. Thus the electronic conductivity of a thin film of CuCl can be studied by restricting the temperature range to below 140 "C to minimise the ionic contribution, the former being potentially more sensitive to gas effects than ionic conductivity. Using the configuration of fig. l(b), no apparent effect on the resistance of the sample was observed on exposure to oxygen, nitrogen, helium or ethylene. Prolonged exposure was complicated by the obscuring of small changes by the background decrease due to aging.The effect of adsorption of water vapour on R-l is shown in table 2. During exposure the film physically disintegrated. There was slight reversibility in that when the sample was re-evacuated after 2 h, readings steadied at 29% above the original R-l value. Further studies with H,O were not attempted because of the destruction of the film. Since the micrographs had indicated very little porosity in the copper(1) chloride, gaseous penetration to the main conducting pathways between the electrodes isP. G. HALL, R. A. HANN, P. HEATON AND D . R. ROSSEINSKY 73 -21.3 -22.0 - - I C 7 -22.7 1 s M 3 - 2 3 . 1 3.3 3.5 103KIT - 13 -16 - x .. 5 -19 M - - 22 1 1 1 I 1.6 2.2 2.8 3 . 4 103K/T Fig. 3. (a) Activation energy at low (ambient) temperature.(b) Activation energy up to 280 "C: 0, ascending temperature ; 0, ascending temperature (same sample). Table 1. Activation energy for electronic conduction in CuCl activation energy for electronic temperature conduction/eV range/"C ref. 0.53 < 280 this work 0.51-0.59 < 140 27 0.54 90- 130 55 0.51 < 225 at 20 kbar 29 0.39 < 220 44 0.37 < 280 30 doubtful and therefore the electrode configuration of fig. 1 (a) was used subsequently with the low-temperature glass cell. In a preliminary experiment, using an electrode separation of 2.5 cm, the results in table 3 were obtained. These results all come from the same slide and are therefore subject to possible error due to a previously unreversed effect. Accuracy in measuring R-l was difficult ( f 10%) because the relatively wide electrode separation gave low current, and the experiment was repeated using new slides for each gas, with electrode separations reduced to 1 mm.Fig. 4 and 5 show the effects of 0, and C,H, on the conductance. The addition and evacuation of N, had no discernible effect. The aging behaviour of the film characterised before74 OXYCHLORINATION CATALYSIS Table 2. Effect of water vapour on conductance of CuCl duration of exposure to change in water vapour/ h R-' (%) 0 0.5 1 2 0 15 23 35 Table 3. Preliminary observations of effects of gases on conductance of CuCl change in R-l on gas exposure to gas change in R-l on re-evacuation helium none none ethylene decrease of 30-50% after 4 h none oxygen nitrogen none none decrease of 20% after 4 h ca.0.4 of decrease reversed 0 t l h Fig. 4. Effect of oxygen on the conductance of CuCl film : (A) oxygen added ; (B) re-evacuation. admission of the gas serves to indicate the value to which the gas should return on re-evacuation. Some scatter in results is attributed to experimental error in the current measurement, due to the need for high sensitivity. In the case of oxygen, a decrease of ca. 40% over 7.25 h is apparently totally reversible, rather than only partially as earlier indicated in table 3 when no account was taken of aging. The slow rate of change of R-l with time confirms the possibility that diffusion through the film is slow. Ethylene (fig. 5 ) causes an irreversible decrease of ca. 25%. The adsorption results for ethylene (Part 2)2 indicate it will chemisorb onto anyP.G. HALL, R. A. HANN, P. HEATON AND D . R. ROSSEINSKY 75 0 \ A 0 20 40 60 ti h Fig. 5. Effect of ethylene on CuCl conductance: (A) ethylene added; (B) re-evacuation. impurity Cu2+ sites, in localised adsorption. The 25% decrease suggests that the Cu2+ impurity is responsible for a large proportion of the conduction mechanism. CuCl has been cla~sified~~y 39 as a p-type semiconductor with acceptor levels lying just above the valence band. If these acceptor impurity levels are filled by electrons from an adsorbate molecule, no electrons can be excited from the valence band and no positive holes are left as carriers, hence conductivity decreases. Wolkenstein13 has defined ‘ weak ’ and ‘strong’ chemisorption; only in the latter case is an effect on conductivity to be expected.Nitrogen is also included in the category of localised adsorption but bonding is so weak (van der Waals only) that conduction is not altered. Ionosorption is a possibility for adsorbate bonding when an electron from the conductor band or hole from the valence band becomes captured or injected by the surface species, but calculation^^^ show that pure inosorption is likely to occur only for 0, adsorbate. Considerable work has been carried out and reviewedl31 66-68 on the adsorption of oxygen on semiconducting oxides but not chlorides. Ionosorption is generally irreversible, being associated with chemisorption, and causes an increase in conductivity for p-type semiconductors. Because of the adsorption data3 of Part 3 (qo = - 7 kJ mol-l on CuC1, qo = - 16 kJ mol-1 on CuCl,, A; H = 6.82 kJ mol-l) and the observed decrease in conductivity, ionosorption is not suspected.The mechanism by which 0, decreases R-l for CuCl is unclear: while the reversibility does not preclude localised bonding, adsorption possibly being both weak yet specific to particular sites, incorporation of the adsorbate into the chemical lattice seems likely. The slow change of R-l with time is consistent with such absorption of oxygen. As a further investigation, the effect of the gases on the activation energy of conduction E, was established. Using the high-temperature cell, a fresh sample was76 OXYCHLORINATION CATALYSIS Table 4. Activation energies on exposure to gases (eV). condition (gas) vacuum exposed re-evacuated oxygen 0.65 & 0.06 0.57 0.02 0.62 ethylene 0.61 0.65 0.65 nitrogen 0.60 f 0.02 0.61 0.66 a & indicates mean of T-increasing and T-decreasing runs, the error limits representing observed differences. mounted and outgassed for ca.3 days to allow aging effects to stabilise. Then Ea was determined by ascending and descending temperature runs. A gas was then admitted at room temperature and left on the sample for ca. 12 h, and the activation energy again established by ascending and descending temperature runs. Finally the sample was re-evacuated and outgassed at room temperature for > 12 h to a vacuum of better than mmHg before a single ascending temperature run. 1 h was allowed for equilibrium at each reading. The results are shown in fig.6 and table 4. For oxygen and ethylene R-l again decreased on admission of the gas, while for nitrogen it remained constant. For oxygen there appears to be no increase in R-l on re-evacuation, although we omitted to take readings immediately after re-evacuation or to characterise the aging effect. No substantial change was observed in E, for any gas, and so decreases in conduction must be attributed to a removal of carriers rather than a change in mechnism. The reaction of oxygen with copper(1) chloride was further investigated by a range of techniques, starting with SIMS. Spectra were recorded for a copper(1) chloride sample before any oxygen was added in order to gauge sample purity. The negative-ion spectrum [see fig. 7(a)] shows large peaks for [Cll-, [ClJ, [CuClI- and [CuCl,]- with very few other peaks.This technique does not give information on the valence state of the copper ions. The positive-ion spectrum was considerably more complex [see fig. 7(c)], because of identifiable impurities including Na+, Al+, K+, [CuOH]+ and traces of a nitrogen-containing contaminant. A sample was exposed to ca. 1 0-1 Torr of oxygen at room temperature and at 150 "C for 1 h. The negative-ion spectrum was unchanged except for a [CuClCNI- peak. A positive-ion spectrum was not taken, as the entities which indicate a reaction has taken place, e.g. [CuOCl], would probably produce negative rather than positive ions. Samples were subjected to high vacuum (ca. 10-lo Torr) when mass spectra were taken, thus removing any weakly absorbed material. The sample was then heated for 30 min at 300 "C under ca.10-1 Torr of oxygen and outgassed overnight before spectra were taken. The positive-ion spectrum [see fig. 7 (c)] indicated that the organic impurity had largely been removed. There were strong sodium impurity levels due to diffusion from the soda-glass microscope slide and the high ion yield of alkali metals generally. The negative-ion spectrum [see fig. 7(b)] showed peaks for [CuO,]-, [CuClCNI-, [CuOCl]-, [CuOI-, [CNI- and [CNOI-; it is concluded that at 300 "C a reaction has occurred incorporating oxygen and CN- into the surface. At lower temperatures the inclusion of oxygen within the surface cannot be discounted, as a vacuum removes reversibly held species. The XPS studies indicated that copper@) chloride samples were virtually free of copper(r1) chloride (fig.8). The prepared samples remained stable to atmospheric oxidation over 2 h. Quantitative comparison of the copper and chloride ion peak areas0 pu P. G. HALL, R. A. HANN, P. HEATON AND D. R. ROSSEINSKY ( I - W I - x) 801 - - ? m - 9 0 Fr M . m 0 + ,? N I I I I I h Y - m 7778 I OXY CHLORINATION CATALYSIS CUCli CuCN' CUCI- I i I I 3x104 -f K* c 1- I 3 Ao3 I ! lo3 I 106 I 3 x 1 0 ~ I Fig. 7. Secondary-ion mass spectra of CuCl: (a) negative-ion spectrum of untreated showing absence of impurity, (b) negative-ion spectrum after preheating in 0, at 300 "C and (c) positive- ion spectrum after preheating in 0, at 300 "C.P. G. HALL, R. A. HANN, P. HEATON AND D. R. ROSSEINSKY 2 5 c M 0 0 2 20 rc.n E 5 1 5 - % - 3 a en 1- g l o - 5 L. 0 79 -- - 234 2 50 266 2 82 energyIeV Fig. 8. CuCl and CuCl, (reference) X-ray photoelectron spectra. 0 80 160 240 T/"C Fig. 9. Extent of reaction of 0, with CuCl from microreactor studies. Flow rate : lJ,0.32 cm3 s-l ; 0, 0.25 cm3 s-' (corrected to 0.32 cm3 s-l). shows a deficiency of chlorine assuming the 1 : I stoichiometry of CuC1. A simple explanation is offered by the presence of some zero-valent copper metal on the surface, as has been reported69 for samples outgassed at 350 "C. Another plausible possibility is that some of the copper(1) is present as a species other than chloride, e.g. the oxide. However, intrinsic differences between sample and standards could introduce errors of 10-2004 in inferred stoichiometries.80 OXYCHLORINATION CATALYSIS Using a microreactor to study the bulk reaction of copper(1) chloride with oxygen, injections of oxygen of constant size (25 x lops dm3) were made over ca.6 g of solid. Assuming negligible adsorption the eluted peak area should be constant (see dashed line in fig. 9). However, as the temperature was increased, the amount of oxygen emerging from the solid decreased. It appears that even at room temperature some small extent of reaction is occurring. Above 200 "C eluted peaks showed tailing, suggesting an additional process above this temperature. The colour of the white solid appeared to darken, and therefore the formation of an oxychloride was suspected. Two samples, copper(1) chloride and copper(1) chloride exposed to a continuous stream of oxygen at 250 "C for ca.40 h, which caused a substantial grey discolouration, were subjected to mass-spectrometry analysis. The spectra were taken with a sample-probe temperature of 250 "C. At 450 "C the vapour of CuCl has been shown to contain substantial amounts of the cyclic trimer CU,C~,~~, 71 and also a t e t ~ a m e r . ~ ~ A later mass-spectrometry study of the vapour in equilibrium with solid CuCl(280-430 "C) has indicated comparable concentrations of Cu3Cl, and Cu,Cl, molecules and a smaller concentration of CU,C~,.~~ Peaks of associated species up to Cu,Cli only were observed by us, with characteristic isotope-splitting patterns. The spectrum of sample (ii) was identical to that of (i) and therefore it was concluded that mass spectrometry is too insensitive to examine the possible oxychloride, probably because the temperature is too low to volatilise the grey-black impurity.Previous attempts to characterise this oxygen/CuCl product have used X-ray diffracti~n,~~ electron spin resonance spectroscopy and mass spe~trometry.~~ The original proposition of an oxychloride being formed at high temperatures was made by The situation was initially complicated by reports of Cu,OCl, formation from copper(I1) 76 The work of Allen and Clarks9 established that CuCl, does not react with oxygen while CuCl forms CuO.CuC1, (or CuO/CuCl,). There are also reported studies carried out in solvents at lower temperatures (5-1 10 "C) of the reversible uptake of oxygen by Cu177 and a change in valency from CuI to CUI*.~* CONCLUSIONS The value of the activation energy of conduction (0.53 eV) obtained for the prepared thin films of copper(1) chloride indicated that the electronic conduction of 'pure' copper@) chloride was being measured, despite the nitrogen-containing impurity of probably organic origin and an apparent non-stoichiometry in the Cu to C1 ratio as shown by XPS.Owing to the non-porous nature of the film, the electrode configuration was important if the effects of gas on conductivity were to be observed. The absence of an effect at room temperature and above with helium or nitrogen was as expected for gases which are only physically adsorbed. The behaviour of R-l with ethylene is consistent with irreversible chemisorption. The site to which C,H, attaches is responsible for a large proportion of the charge carriers. Adsorption results indicate that ethylene chemisorbs to Cu2+.Therefore Cur* appears to be present as an impurity site and these acceptor impurity levels participate in the electronic conduction mechanism. The reversible behaviour of oxygen in decreasing conductivity was not expected for a p-type semiconductor. Absorption is tentatively suggested as being responsible since microreactor studies indicated that incorporation of oxygen into the solid was occurring. Its characterisation by SIMS failed, possibly because of the high vacuum used in detection. The colour change and SIMS results indicate that at higher temperatures (> 200 "C) an irreversible uptake of oxygen occurs. This could explain why the oxygen effect was not reversible after temperature cycling.As no significant change in E, was observed for any gas, decreases in conduction were attributed toP. G. HALL, R. A. HA", P. HEATON AND D. R. ROSSEINSKY 81 removal of charge carriers by donation of electrons from the adsorbate into the acceptor impurity levels, rather than a change in mechanism. We thank the S.E.R.C. and Imperial Chemical Industries p.1.c. for the CASE award under which the work was performed, and the latter for access to analytical and spectral services arranged by Dr J. Wolstenholme. We also thank both referees for their helpful comments. P. G. Hall, P. Heaton and D. R. Rosseinsky, J . Chem. Soc., Faraday Trans. I , 1984, 80, 2777. P. G. Hall, P. Heaton and D. R. Rosseinsky, J .Chem. Soc., Faraday Trans. I , 1984, 80, 2785. P. G. Hall, P. Heaton and D. R. Rosseinsky, J . Chem. Soc., Faraday Trans. I , 1984, 80, 3059. H. S. Taylor, J . Chem. Phys., 1927, 15, 624. T. Kwan, Adv. Catal., 1954, 6, 67. M. Boudart, J . Am. Chem. Soc., 1952,74, 1531. P. Aigrain and C. Dugas, Z . Elektrochem., 1952, 56, 363. K. Hauffe and H. J. Engel, Z . Elektrochem., 1952, 56, 366. P. B. Weisz, J . Chem. Phys., 1952, 20, 1483. lo P. B. Weisz, J . Chem. Phys., 1953, 21, 1531. l 1 S. R. Morrison, Adv. Catul., 1955, 7, 259. l2 T. E. Madey, J. T. Yates, D. R. Sandstrom and R. J. H. Voorhoeve, in Treatise on Solid State Chemistry, ed. N. B. Hannay (Plenum Press, New York, 1976) vol. 6B, p. 59. l 3 Th. Wolkenstein, The Electronic Theory of Catalysis on Semiconductors (Pergamon Press, Oxford, 1963); Adu.Catal., 1960, 12, 189. l4 P. J. Fensham, Q. Rev. Chem. Soc., 1957, 11, 227. l5 K. J. Miller, J . Chem. Ed., 1971, 48, 582. l6 A. Clark, The Theory of Adsorption and Catalysis (Academic Press, London, 1970). l7 J. J. Herrmann, J. Portefaix, M. Forissier, F. Figueras and P. Pichat, J . Chem. Soc., Faraday Trans. l8 M. Takata and H. Yanagida, Yogyo Kyokai Shi, 1979,87, 13. l9 N. V. Suntsov, V. I. Arkharov, V. N. Konev and A. G. Miloslavskii, Dokl. Akad. Nauk S.S.S.R., *O A. R. Plummer in The Electrochemistry of Semiconductors, ed. P. J. Holmes (Academic Press, London, 21 A. Many, Y. Goldstein and N. B. Grover, Semiconductor Surfaces (North-Holland, Amsterdam, 22 D. R. Frankl, Electrical Properties of Semiconductor Surfaces (Pergamon Press, Oxford, 1967).23 P. G. Hall, M. Parsley, D. R. Rosseinsky, R. A. Hann and K. C. Waugh, J . Chem. SOC., Faraday 24 L. G. Harrison and C. F. Ng, Trans. Faraday Soc., 1971, 78, 1787. 25 L. G. Harrison and C. F. Ng, Trans. Furaday Soc., 1971, 67, 1801. 26 L. G. Harrison and C. F. Ng, Trans. Furaday Soc., 1971, 67, 1810. 27 L. G. Harrison and H. Prasad, Trans. Faraday Soc., 1974, 70, 471. 28 C. F. Ng and K. S. Leung, J . Catal., 1981, 67, 410. 29 R. S. Bradley, D. C. Munro and P. N. Spencer, Trans. Faraday Soc., 1969, 65, 1912. 30 L. G. Maidanovskaya, I. A. Kirovskaya and G. L. Lobanova, Izv. Akad. Nauk SSSR, Neorg. 31 Y. W. Hsueh and R. W. Christy, J . Chem. Phys., 1963, 39, 3519. 32 J. B. Wagner and C. Wagner, J. Chem. Phys., 1957, 26, 1597. 33 C. Tubandt, Handb.Experimentalphys., 1932, 12, 383. 34 A. K. Shukala, S. Chattopadhyay and C. N. Rao, Cryst. Lattice Defects, 1976, 6, 191. 35 K. D. Becker, G. W. Herzog, D. Kanne, H. Richtering and E. Stadler, Ber. Bunges. Phys. Chem., 36 R. E. Malpas, Ph. D. Thesis (Exeter University, 1977). 37 K. S. Song, J . Phys. (Paris), 1967, 28, 195. 38 K. S. Song, J . Phys. Chem. Solids, 1967, 28, 2003. 39 A. B. Kunz, R. S. Weidman, J. Boettger and G. Cochran, Int. J . Quantum Chem., Quantum Chem. 40 L. Kleinman and K. Mednick, Phys. Reti. B, 1979, 20, 2487. 41 P. C. de Mello, M. Hehenberger, S. Larsson and M. Zerner, J . Am. Chem. Sac., 1980, 102, 1278. 4 2 H. A. Cerdeira, Solid State Commun., 1980, 35, 109. I, 1979, 75, 1346. 1979, 245, 1372. 1962). 1965). Trans. I , 1983, 79, 343.Muter., 1967, 3, 936. 1970, 74, 527. Symp., 1980, 14, 585. 4 FAR 182 OXYCHLORINATION CATALYSIS 43 C. Divakar, M. Mohan and A. K. Singh, Solid State Commun., 1980, 34, 385. 44 C. W. Chu, S. Early, T. H. Geballe, A. Rusakov and R. E. Schwall, J. Phys. C, 1975,8, 1241. 45 C. W. Chu, A. P. Rusakov, S. Huang, S. Early, T. H. Geballe and C. Y. Huang, Phys. Rev. By 1976, 18, 2116. 46 G. J. Piermarhi, F. A. Mauer, S. Block, A. Jayaraman, T. H. Geballe and G. W. Hull Jr, Solid State Commun., 1979,32,275. 47 G. J. Piermarini, F. A. Mauer, S. Block, A. Jayaraman, T. H. Geballe and G. W. Hull Jr, High Pressure Sci. Technol., Proc. 7th Znt. AIRAPT Con$ (National Bureau of Standards, Washington D.C., 1980), vol. 1, p. 395. 48 C. W. Chu, S. Early, T. H. Geballe and C.Y. Huang, J. Less-Common Metals, 1978, 62, 463. 49 B. Batlogg, J. P. Remeika and R. G. Maines, Solid State Commun., 1981, 38, 83. 50 G. C. Vezzdi and J. Bera, Phys. Rev. B, 1981, 23, 3022. 52 A. V. Joshi and L. B. Wagner Jr, J. Electrochem. Soc., 1975, 122, 1071. 53 J. Rivera, L. A. Murray and P. A. Hoss, J. Cryst. Growth, 1967, 1, 17. 54 A. V. Joshi, in Fast Zon Transport in Solid, ed. W. van Goo1 (North-Holland, Amsterdam, 1973). 55 T. Matsui and J. B. Wagner Jr, J. Electrochem. Soc., 1977, 124, 610; 941 ; 937. 58 S. Geller, J. R. Akridge and S. A. Wilber, Phys. Rev. B, 1979, 19, 5396. 57 T. Takahashi, 0. Yamamoto, S. Yamada and S. Hayashi, J. Electrochem. Soc., 1979, 126, 1654. 58 T. Takahashi, 0. Yamamoto and A. Sawai, J. Appl. Electrochem., 1978,8, 161. 5g T. Jow and J. B. Wagner Jr., J. Electrochem. Soc., 1979, 126, 1963. 6o N. F. Mott, Nuovo Chim., 1958, 10, 7, suppl. 312. 61 I. G. Austin and N. F. Mott, Science, 1970, 168, 71. 62 D. R. Rosseinsky and J. S. Tonge, J. Chem. Soc., Faraday Trans. 1, 1982,78, 3595. 63 P. Day, Endeavour, 1970, 29, 45. 64 N. F. Mott and R. W. Gurney, Electronic Processes in Ionic Crystals (Dover, New York, 1964). 65 M. Green and M. J. Lee, in Solid State Surface Science, ed. M. Green (Marcel1 Dekker, New York, 68 S. J. Thomson and G. Webb, Heterogeneous Catalysis (Oliver and Boyd, Edinburgh, 1968), p. 59. 67 F. S. Stone, A h . Catal., 1962, 13, 1. 68 D. 0. Hayward and B. M. W. Trapnell, Chemisorption (Buttenvorths, London, 2nd edn, 1964), p. 261. 6e J. A. Allen and A. J. Clark, J. Appl. Chem., 1966, 16, 327. 70 L. Brewer and N. L. Lofgren, J. Am. Chem. Soc., 1950,72, 3038. 71 C. H. Wong and V. Shomaker, J. Phys. Chem., 1957,61, 358. 72 H. M. Rosenstock, J. R. Sites, R. J. Walton and R. Baldock, J. Chem. Phys., 1955, 23, 2442. 73 Guido, G. Balucci, G. Gigli and M. Spoliti, J. Chem. Phys., 1971, 55, 4566. 74 €3. Deacon, Brit. Pat. 1948 (1866). 75 A. E. Korvezee, Reel. Trav. Chim., 1931,50, 1085. Von K. Jellinek and A. Rudat, 2. Anorg. Allg. Chem., 1926, 155, 73. 77 J. E. Bulkowski, P. L. Burk, M-F. Ludmann and J. A. Osborn, J. Chem. Soc., Chem. Commun., 1977, 498. G. Davies, M. F. El-Shazly, D. R. Kozlowski, C. E. Kramer, M. W. Rupich and R. W. Slaven, in Inorganic Compounds with Unusual Properties, ed. R. B. King, Adv. Chem. Ser. no. 173 (American Chemical Society, Washington D.C., 1979), vol. 2, p. 178 and references therein. 1969), vol. 1. (PAPER 4/404)