J. CHEM. SOC. FARADAY TRANS., 1994, 90(13), 1987-1991 X-Ray Photoelectron Spectroscopy, Temperature-programmed Desorption and Temperature-programmed Reduction Study of LaNiO, and La,NiO, +6 Catalysts for Methanol Oxidation Jacques Choisnet Centre de Recherche sur la Metiere Divisee, Unite Mixte CNRS, Universite d 'Orleans-Crystallochimie,Faculte des Sciences , Universite d 'Orleans, F-45067 Orleans Cedex 2, France Nevena Abadzhieva, Plamen Stefanov and Dimitar Klissurskif Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, Sofia 1040,Bulgaria Jean Marc Bassat Centre de Recherche sur la Physique des Hautes Temperatures, CNRS, ID Avenue de la Recherche Scientific, F-45071 Orleans Cedex 2,France Vicente Rives* Departamento de Qulinica lnorganica, Universidad de Salamanca , Facultad de Farmacia, 37007Salamanca, Spain Lev Minchev Institute of Kinetics and Catalysis , Bulgarian Academy of Sciences, Sofia 1040,Bulgaria An X-ray photoelectron spectroscopy (XPS), temperature-programmed desorption (TPD) and reduction (TPR) study of LaNiO, and La,Ni04+d perovskites has been carried out.The existence of at least two different forms of oxygen in these compounds is shown by both oxygen reactivity (TPD and TPR) and XPS characterisation. XP spectra have also revealed a surface enrichment in lanthanum and oxygen. TPR profiles have shown a reduction of LaNiO, through the formation of La,Ni,O, as an intermediate compound. Above 450 "C,LaNiO, and intergrowth nickelates (La,NiO, and La2Ni04+&) undergo a final reduction to metallic Ni and La,O,.Between 200 and 400°C all three compounds exhibit a high catalytic activity in the total catalytic oxidation of methanol. ABO erovskite mixed oxides are a subject of intensive stud;'! Their most outstanding feature is their capacity for partial substitution of A and B sites, which results in many structurally similar compounds. Moreover, the trend towards producing an intergrowth of ABO, perovskite layers with A0 rock-salt layers has resulted in new mixed oxides, such as (ABO,),AO (p = 1-3) phases,, which show a lower dimen- sionality of structure compared with the '3D' character of the so-called 'ReO,' octahedral network. This is another reason for studying their different physico-chemical properties.In addition, the discovery of high-temperature superconduc- tivity in La,CuO, +d phases has stimulated intensive studies of the structure and physical properties of oxides with K,NiF,-type structure. It is now well known that La,NiO,+, exists over a broad range of oxygen its structural, electric and mag- netic properties are very sensitive to the amount of non-st oichiome tric oxygen present .6-9 Perovskite-type oxides are known to be efficient catalysts for a large number of chemical reactions.'*-'' Most of them are excellent catalysts for complete oxidation.2~'0~"~'7 Their behaviour in catalytic partial oxidation reactions has also been investigated.', Oxidation of methanol is a very conve- nient test reaction for characterisation of the redox, as well as the acid-base, properties of oxide-type catalyst^.'^.'^ In view of this, an XPS, TPD and TPR study of LaNiO, and La,NiO,+, (6 < 0.16) has been performed in parallel with a study of their behaviour towards methanol oxidation.Mixed nickelates are of particular interest, as the presence of Ni3+ f Also at: Departamento de Quimica Inorganica, Universidad de Salamanca, Facultad de Farmacia, 37007-Salamanca, Spain. and their oxygen non-stoichiometry could be related to their catalytic properties and oxygen reactivity. Experimental La,NiO, was obtained from a 1 :1 stoichiometric mixture of La,O, and NiO, heated in air in the temperature range 930-1130°C. Two successive annealings at 1130°C for 15 h, with intermediate re-grinding, were necessary to complete the solid-state reaction. In order to achieve a maximum value for 6 in La,NiO,+&, a modified sol-gel method,,' using lanthanum and nickel nitrates as precursors, was followed. The gel was steadily heated in air up to 900°C.The resulting powder was further annealed in an oxygen flow at 1150 "C for 10 h. LaNiO, was synthesized by coprecipitation of lanthanum and nickel hydroxides.21 The precipitate was steadily cal- cined in oxygen up to 950"C, with a final annealing at this temperature for 15 h. The phase purity of the specimens was monitored by X-ray diffraction in a Siemens D500 instrument, using Cu-Ka radi-ation. Specific surface areas, as determined by the BET method, were in the range 1-2.5 mz g-'.TPD of oxygen was carried out in the temperature range 25-80O0C, at a heating rate of 25 "C min- in an He flow (60 ml min- I). TPR experiments were performed using a Micromeritics TPR/TPD 2900 apparatus, at a heating rate of 10°C min-', using a 5 vol.% H,-Ar mixture and calibrating the instru- ment by checking hydrogen consumption during CuO reduction. The amount of sample, heating rate and gas flow were chosen to ensure a good resolution of component peaks. XP spectra of fresh and tested samples were recorded with a VG Escalab I1 instrument, using Mg-Ko! radiation (1253.67 eV). Finally, the heterogeneous catalytic oxidation of meth- anol was performed by a flow method with a 4 ml min-' flow rate of the reaction mixture, 4 vol.% methanol in air.Analysis of the reaction products was performed using the hydrogensulfite (for formaldehyde) and chromatographic methods.22 Results and Discussion Structural Peculiarities The fully oxidized, Ni3 +-containing LaNiO, sample, exhibits, as previously ~hown,~',~~ a slightly distorted perovskite-like rhombohedra1 cell, i.e. a = ap,/2 = 5.395 8, and a = 60" (a, is the parameter of the simple cubic perovskite-type unit cell). The thermal stability of this perovskite depends on the tem- perature and surrounding atmosphere during calcination. The stability is high in air, as LaNiO, undergoes decomposi- tion only above 1130 "C, according to the following irrevers- ible reaction: 2LaNi0, + La,NiO, + NiO + 30, At moderate temperatures (200-430 "C) in a hydrogen flow, a progressive elimination of oxygen takes place, leading to the oxygen-deficient LaNiO, -phases, and further to the strongly reduced defect perovskite, LaNiO, .24 Consequently, catalytic tests on LaNiO, have been performed at tem-peratures below 450 "C in an oxidizing atmosphere.La,NiO, is another example of the 1 : 1 intergrowth of ABO, perovskite and A0 rock-salt-type layers, Fig. 1. The La,NiO, +, structure consists of p-type-doped NiO, layers alternating with rock-salt-type La,O, +,layers of variable oxygen content, the sequence along the c axis being Ni0,-La,O, +, . When prepared in air or in an oxygen flow, La,NiO,+, shows the K,NiF,-type tetragonal unit cell.Only slight variations of the cell parameters are observed, depend- ing on the preparation conditions: the value of c is somewhat larger when the oxygen in excess reaches a value close to 0.18 (cmaX= 12.71 A, vs. 12.67 8, for air-prepared La,NiO, 25,26). Several structural studies have elucidated the role of the oxygen con tent. 7-29 It has been demonstrated that the a ~a 0 Ni 00 o exc. 0 ""b p 4 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 occurrence of a tetragonal unit cell depends to some extent on Ni2++Ni3+ ~xidation.~' The air-prepared La,NiO, usually contains some non-stoichiometric oxygen. The nature and exact location of the extra oxygen atoms have produced some controversies. The existence of superoxide (02-)or per- oxide (0,2-)anions has been proposed. Stoichiometric La,NiO, (6 = 0) can be obtained only after annealing in an argon or nitrogen flow, i.e.in an inert atmo- sphere. Its crystal lattice is orthorhombic, a x b w upJ2, c = 12.547 A. Moreover, the La : Ni ratio can tolerate some variation (La3 vacancies in the La0 layers).26 Finally, the + actual oxygen content is controlled not only by the prep- aration conditions (oxygen, air or inert atmosphere), but also by the La : Ni stoichiometry. XPS Characterisation A study of the photoelectron spectra of 'fresh' samples and those used in methanol oxidation was carried out, in order to obtain information on their surface composition. The La 3d core-level spectra show split lines with maxima at 835.3 f0.2 and 838.1 f0.3 eV for fresh LaNiO,, air-prepared La,NiO, and La,NiO,+, .This agrees with values reported previ~usly.~' The position and shape of the peak remain unchanged after the catalytic experiments. The 0 1s peaks are broad and asymmetric for the three samples. This result is consistent with the presence of more than one type of oxygen species in the surface layer. This is clearly concluded from the peak deconvolution of the 0 1s spectra of samples LaNiO, and La,NiO,+, (Fig. 2 and 3, respectively), which points to the existence of three peaks. In view of the possible presence of surface impurities, the attribution of the high- energy (532 eV) peak to oxygen in the sample is not definite. On the basis of previous the low-energy peak can be ascribed to a lattice 0,-anion, whereas the other two peaks probably originate from chemisorbed 0-and adsorbed OH groups.There is a slight, but significant differ- ence between the binding energies (Ebs) of lattice oxygen in LaNiO, and La,NiO,.,, (528.1 and 529.1 eV, respectively); this can be related, at least in a qualitative way, to the non- equivalence of the lattice oxygens in the perovskite and inter- growth structures. 526 528 530 532 534 E,IeV Fig. 1 1 : 1 intergrowth of ABO, perovskite (P) and A0 rock salt Fig. 2 Deconvolution analysis (. ...) of the 0 Is signal of sample (RS) type layers in La,NiO, LaNiO, ('used' sample) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I L I 526 528 530 532 534 E,IeV Fig.3 Deconvolution analysis of the 0 1s signal of sample La,Ni04+d (‘used’ sample) The atomic surface concentration was determined in terms of the La :(La + Ni) and 0 : (La + Ni) ratios, Table 1. As a typical result, a systematic enrichment in La and 0 surface concentrations with respect to the bulk values is observed whichever oxide (perovskite or intergrowth) is considered. The La enrichment is not influenced by the catalytic process, whereas the 0 enrichment (in the intergrowth-type oxides only) varies slightly and reaches approximately the same value, 2.24, for air-prepared La,NiO, and La,NiO,+ 6. Such an observation is in agreement with previous results obtained for similar compound^,^ where significant differences between the bulk and surface concentration have been reported.A tentative hypothesis based on the structural pecu- liarities of the perovskite family can be proposed. The overall structure of any perovskite-like phase is a combination of ‘BO,’ and ‘AO’ atomic planes, i.e. ‘NiO,’ and ‘Lao’ in the nickelates studied here. Consequently, a modification in La and 0 concentrations can be related to some changes in the distribution of the ‘NiO,’ and ‘Lao’ planes. In this way, the La and 0 enrichment of the surface observed here is likely to be due to a preferential distribution of ‘Lao’ planes in the layers close to the surface. Obviously, this is a qualitative assumption, but it is consistent with the available crys- tallochemical data for perovskite-like phases.Finally, it can be assumed that surface oxygen in these compounds is bonded in at least two different ways, which probably correspond to two different forms of adsorbed oxygen. Generally speaking, the existence of more than one form of oxygen is characteristic also for some oxides which exhibit catalytic activity for complete oxidation of organic compounds, namely c0304, NiO and Cr,O, .36 Moreover, it is well known that the low-temperature adsorption of oxygen Table 1 Surface atomic concentrations in LaNiO,, La,NiO, and La,NiO,+, samples, as determined by XPS LaNiO La,Ni04 La,NiO, + ratio fresh used fresh used fresh used La :(La + Ni) 0 : (La + Ni) 0.71 1.82 0.71 1.80 0.81 2.24 0.81 2.40 0.81 2.22 0.80 2.31 results in ‘weakly bound’ forms, for which the bonding energy value is characteristic of chemisorbed oxygen and consequently shows a fairly high reactivity. Reactivity of Oxygen (TPD and TPR) The TPD curves of oxygen from fresh LaNiO,, La,NiO, and La,NiO,+, are shown in Fig.4. The desorption peaks are rather broad, especially for LaNiO, . Desorption maxima are recorded at 280-320 “C and 450 “C for air-prepared ‘La,NiO,’, and at 400°C (this peak being well defined) and 480°C for La,NiO,+,. LaNiO, shows a desorption maximum at 320-380°C. From these results, it can be assumed that two different forms of oxygen (at least) exist on the catalyst surface, in agreement with the XPS results men- tioned above. Fig. 5 illustrates the oxygen desorption from the same cata- lysts after use in the catalytic tests, i.e.in methanol oxidation. It clearly shows a lower amount of evolved oxygen. More- over, the TPD traces for the intergrowth samples are very similar. There is a desorption maximum at ca. 450°C. A well defined desorption maximum has not been observed, however, for LaNiO, . Note that after the catalytic tests, the desorption maxima at lower temperatures, which are ascribed to the most weakly bound oxygen, practically disappear. This is indicative of a strong modification of the catalysts during the catalytic reac- tion. Evidently, the stationary state of all catalysts studied is noticeably different from the initial one. Undoubtedly, part of 100 300 500 700 T/“C Fig.4 TPD profiles of ‘fresh’ (a) La,NiO,, (b) LaNiO,,, and (c) LaNiO, samples 100 300 500 700 TIcC Fig. 5 TPD profiles of ‘used’ (a) La,NiO,, (b) LaNiO,,, and (c) LaNiO, samples /-\ .........-............. ... -_.i-.'' ..-.............. I801 I I I a I I I I I I 100 200 300 400 500 600 700 TI"C Fig. 6 TPR profiles of 'fresh' (a) LaNiO,, (b) La,NiO, and (c) La,NiO, +d samples the most reactive oxygen is lost during the catalytic oxidation of methanol. A similar effect has been observed in our pre- vious studies on the catalytic oxidation of methanol on superconducting YBa,Cu,O, -x (orthorhombic phase), which can also be considered to be a defective perovskite-type com- pound.,, LaNiO, and intergrowth nickelates have shown noticeably different behaviours during the TPR experiments.~~Studies carried out by Wachowski et ~1 on. reduced LaMeO, (Me = Fe, Co, Ni) showed that all three perovskite- like oxides do not reduce directly to Me and La,O,, but form intermediate oxygen-deficient structures. According to Crespin et ~l.,~'reduction of LaNiO, occurs at low tem- perature (300 "C) according to the reaction: 2LaNi0, + H, -+La,Ni,O, + H,O Our TPR results (Fig. 6) are in full agreement with this state- ment. To make the comparison easier, the TPR traces have been referred to one unit of the reducible metal, i.e. nickel. For LaNiO, two different peaks are clearly detected, with maxima at 385 and 511 "C. Hydrogen consumption for the first peak corresponds to 0.50 mol H, (mol catalyst)-', i.e.one electron (mol catalyst)-'. For the second peak, the ratio is 0.99 mol H, (mol catalyst)-', corresponding to two elec- trons (mol catalyst)-'. So, reduction yields La,O, and metal- lic Ni. The first peak, corresponding to the one-electron reduction process, would then correspond to the formation of La,Ni,O,. The TPR curves of both La,Ni04+d and air-prepared La,Ni04 show weak maxima at ca. 380"C, which can be ascribed to removal of non-stoichiometric oxygen, and corre- spond to 0.085 and 0.105 mol H, (mol catalyst)-', respec-tively. Reduction at higher temperatures is practically identical in both cases, with broad maxima at 610-660°C (undoubtedly formed by several overlapping peaks in the case of sample La,NiO,), with areas corresponding to consump- tion of 1.043 and 0.933 rnol H, per mol of the corresponding nickelate.These values indicate a practically complete trans- formation of both samples to La,O, and metallic Ni. Catalysis of Methanol Oxidation The conversion degrees of CH,OH to CO, CO,, HCHO or other by-products were determined after attaining a regime corresponding to a steady state of the catalyst. CO and H, were not found in the reaction products. As can be seen from the data summarized in Table 2, the main reaction product for all of the catalysts tested here was CO,. Over the whole temperature range studied, LaNiO, , La,NiO, and J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Oxidation of methanol on LaNiO,, La,NiO, and La,NiO, ox ygen-containing products catalyst temperature/"c methanol conversion (Yo) HCHO CO, LaNiO, 200 40.3 0.5 39.8 300 51.4 1.6 49.8 370 85.9 2.5 83.4 400 89.0 4.5 84.5 450 90.0 5.0 85.0 La,NiO, 200 33.8 1.5 32.3 300 64.6 2.0 62.6 400 72.0 2.0 70.0 450 75.3 2.0 73.3 La2Ni0,+d 200 70.1 0.0 70.1 300 80.4 0.3 80.1 350 90.8 0.5 90.3 370 97.0 0.6 96.4 400 99.8 0.7 99.1 La2Ni0,+d behave as highly efficient catalysts for the com- plete oxidation of methanol.During our previous studies on the mechanism of ammonia oxidation over oxide-type catalyst^,^' we have found a distinct correlation between the binding energy of oxygen in oxide-type catalysts and their activity in complete oxidation. Boreskov et have shown that for a large series of complete oxidation reactions a simple relationship holds between the activation energy for the reaction (E,) and the binding energy of oxygen (Eb) in the surface layer of the cata- lyst : E, = I?,fbEb where b is a constant.The TPD curves for oxygen desorption from LaNiO,, La,NiO, and La,NiO,+, clearly show the existence of easily desorbable (i.e.weakly bound), highly reactive oxygen. Our further studies on methanol heterogeneous catalytic oxidation have shownI8 that selectivity to formaldehyde or CO, formation, i.e. to partial or complete oxidation, largely depends on the binding energy of oxygen in the surface layer of the catalysts and on the surface acid-base properties.The TPD and TPR studies of LaNiO,, La,NiO, and La2Ni04+6 are indicative, as already mentioned, of the exis- tence of weakly bound (and therefore highly reactive) oxygen in all three compounds. This is essential for catalytic activity during complete oxidation reactions, and is fully consistent with the results reported in the present study. Note that the catalysts studied show a high activity with respect to total oxidation, even at temperatures as low as 200 "C. Under the reaction conditions chosen, a similarity in the catalytic behaviour of all three nickelates has been observed, despite the structural peculiarities and different oxygen stoichiometries of the initial 'fresh' compounds. On the basis of the TPD and TPR data for the fresh and tested catalysts, it can be tentatively assumed that surface alterations seem likely to happen during the catalytic reac- tion, resulting not only in surface reconstruction, but also in changes in the surface stoichiometry.This could be the reason for a similarity in the catalytic behaviour of LaNiO, , La,NiO, and La,NiO, + The methanol oxidation reaction is very sensitive to the nature of the surface sites present in oxide-type catalysts. According to Wachs et al. surface redox sites (capable of being reduced and oxidized) primarily form formaldehyde, as J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1991 well as methyl formate and dimethyl ether; surface acid sites, Lewis as well as Brsnsted, facilitate the formation of dimethyl ether, and surface basic sites yield CO and CO, as the reac- tion products.” The compounds studied exhibit a high activ- ity with respect to methanol oxidation to carbon dioxide.This pathway of methanol oxidation could be related to the 6 7 8 9 C. N. R. Rao, P. Ganguly, K. K. Sin& and R. A. Mohan Rao, J. Solid State Chem., 1988, 72, 14. R. R. Shartman and J. M. Honig, Muter. Res. Bull., 1989, 24, 1375. T. Freltoft, D. J. Buttrey, G. Aeppli, D. Vakmin and G. Shirane, Phys. Rev. B, 1991,44, 5046. J. D. Jorgensen, B. Dabrowski, S. Pei, D. R. Richards and D. G. predominantly basic character of the catalysts. In this context, the results of Ai and Suzuki should also be noted.40 These authors have found an acceleration in the total oxida- tion of phenol with increasing surface basicity for a large series of complex oxide catalysts.According to Sei~ama,~’the catalytic activity of per- 10 11 12 13 14 15 Hinks, Phys. Rev. B, 1989,49,2187. T. Seiyama, Catal. Rev.-Sci. Eng., 1992, 34, 281. B. Viswanathan, in ref. 1, p. 271. K. Ichimura, Y. Inoue and I. Yasumori, in ref. 1, p. 301. T. Shimizu, in ref. 1, p. 289. J. L. G. Fierro, Catal. Rev.-Sci. Eng., 1992,34,321. T. R. N. Kutty and M. Avudaithai, in ref. 1, p. 307. ovskites in total oxidation reactions is mainly dependent on component B oxides, and the activity sequence is similar to those of single B oxides. So, the activity of the La-Ni per-ovskites in the complete oxidation of methanol should re-semble, to some extent, the activity of nickel@) oxide.It is well known that NiO catalyses the complete oxidation of 16 17 18 19 D. B. Hibbert, in ref. 1, p. 325. R. J. H. Voorhoeve, in Advanced Materials in Catalysis, ed. J. J. Burton and R. L. Garten, Academic Press, New York, 1977, p. 129. D. G. Klissurski, Proc. 4th Znt. Congr. on Catalysis, Moscow, 1968, Akademiai Kiado, Budapest, 1977, vol. 1, p. 477. I. E. Wachs, G. Deo, M. A. Vuurman, H. Hu, D. S. Kim and methanol to carbon dioxide in the temperature range studied here. 20 J-M. Jehng, J. Mol. Catal., 1993, 82,443. M. Crespin, J. M. Bassat, P. Odier, P. Mouron and J. Choisnet, J. Solid State Chem., 1990,84, 165. 21 M. Crespin, P. Levitz and L. Gatineau, J. Chem. SOC., Faraday Conclusions 22 Trans.2, 1983,79, 1181. D. G. Klissurski, J. Pesheva, Y. Dimitriev, N. Abadzhieva and L. XPS studies of LaNiO,, La,Ni04 and La2Ni04+d have Minchev, in New Developments in Selective Oxidation, ed. G. shown a noticeable difference between their surface and bulk Centi and F. Trifiro, Elsevier, Amsterdam, 1990, p. 287. compositions. XP spectra have evidenced a surface enrich- ment in lanthanum and oxygen. The existence of at least two different forms of desorbable oxygen has been demonstrated by both TPD and XPS studies. A significant difference in the stoichiometry and oxygen 23 24 25 A. Wold, B. Post and E. Banks, J. Am. Chem. Soc., 1957, 79, 491 1. P. Levitz, M. Crespin and L. Gatineau, J. Chem. SOC., Faraday Trans. 2, 1983,79, 1195. J. Choisnet, J.M. Bassat, H. Pillier, P. Odier and M. Leblanc, Solid State Commun., 1988,66, 1245. reactivity between the initial ‘fresh’ and the tested catalysts has been observed. This is indicative of surface alterations 26 D. J. Buttrey, P. Ganguly, J. M. Honig, C. N. R. Rao, R. R. Shartman and C. N. Subbana, J. Solid State Chem., 1988, 74, during the catalytic reaction, until a steady state of the cata- lyst is reached. LaNiO, and the intergrowth nickelates (La,NiO, and 27 28 233. P. Odier, Y. Nigara and J. Coutures, J. Solid State Chem., 1985, 56,32. P. Odier, M. Leblanc and J. Choisnet, Muter. Res. Bull., 1986, La,Ni04 +a) have shown a high efficiency for complete oxida- 21, 787. tion of methanol. This can be related to the relatively low binding energies of oxygen in their surface layers, i.e.a high reactivity of surface oxygen. 29 30 P. Odier, J. M. Bassat, M. Crespin and J. Choisnet, Jpn. J. Appl. Phys., Ser. 2, 1989, 129. Y. Uwamino, T. Ishizuka and H. Yamatera, J. Electron Spec- trosc. Relate. Phenom., 1989, 34. D.K. acknowledges a sabbatical grant from Ministerio de Educacion y Ciencia (Madrid, Spain, ref. SAB92-0302 and 31 32 33 J. L. G. Fierro, J. Catal., 1984,87, 126. J. L. G. Fierro and L. G. Tejuca, Appl. Su$. Sci., 1987,27,453. L. G. Tejuca and J. L. G. Fierro, Thermochim. Acta, 1989, 147, SAB94-0020). 34 361. J. L. G. Fierro, in ref. 1, p. 195. 35 K. Tabata, I. Matsumoto and S. Koshiki, J. Muter. Sci., 1987, References 22, 1882. 1 J. Twu and P. K. Gallagher, in Properties and Applications of Perooskite-type Oxides, ed. L. G. Tejuca and J. L. G. Fierro, Marcel Dekker, New York, 1992, p. 1. L. G. Tejuca, J. L. G. Fierro and J. M. D. Tascon, in Advances in Catalysis, ed. D. D. Eley, H. Pines and P. B. Weisz, Academic Press, New York, 1989, vol. 36, p. 237. A. F. Wells, Structural Inorganic Chemistry, Clarendon Press, Oxford, 1984, p. 602. B. Dabrowski, J. D. Jorgensen, D. G. Hinks, S. Pei, D. R. Richards, H. B. Vanfleet and D. L. Decker, Physica C, 1989,99, 2 3 4 36 37 38 39 40 41 B. Halpern and J. E. Germain, J. Catal., 1975, 37,44. L. Wachowski, S. Zielinski and A. Burewicz, Acta Chim. Acad. Sci. Hung., 1981, 106,217. St. Kynev, D. Klissurski and E. Vateva, Commun. Znst. Phys., Bulg. Acad. Sci., 1962,9, 57. G. K. Boreskov, V. V. Popovskii and V. A. Sazonov, in Scientific Bases for Prediction of Catalytic Action, Proc. 4th Int. Congr. on Catalysis, Nauka, Moscow, 1970, p. 343 (in Russian). M. Ai and S. Suzuki, Bull. Jpn. Petrol. fnst., 1974, 16, 118. T. Seiyama, in ref. 1, p. 215. 162. D. E. Rice and D. J. Buttrey, J. Solid State Chem., 1993, 105, 197.5 Paper 3/07433A; Received 17th December, 1993