J. MATER. CHEM., 1994, 4(1), 125-131 Microcalorimetric Study of the Acidity of Tungstic Heteropolyanions Frederic Lefebvre, Feng Xian Liu-Cai and Aline Auroux lnstitut de Recherches sur la Catalyse, Laboratoire Propre du CNRS conventionne a I’Universite Claude-Bernard Lyon I, 2 avenue Albert Einstein, 69626 Villeurbanne Cedex, France The number and strength of the acid centres of tungstic heteropolyacids have been determined by absorption calorimetry of ammonia. The initial heats are in the order H3PW,20,0 >H,SiWl,O,O >H6P2W21071(H20)3>H,p2W1&62, varying from 200 to 155 kJ mol-’. An increase in the number of protons in Keggin heteropolyanions decreases the acidic strength. Moreover, the influence of the activation temperature on the acidity has been studied and confirms that there is a drastic modification of the solid at ca.350°C.The heat capacities of the heteropolyanions and of the corresponding ammonium salts, the thermokinetic parameters of ammonia absorption and the heat capacities of the acids and ammonium salts have been related to the porosity of the various samples. Heteropolyanions are complexes used in acid and/or redox catalysis whose general formula may be represented by (X,M,0,)4-(x<rn) where M is usually molybdenum or tungsten in their highest oxidation states, X is a heteroatom (P, Si, . . .) and q is the anion charge. When the counter-cations are protons, they are called heteropolyacids. In the formula, these exchangeable protons are placed before the heteroatoms as in simple oxoanions.Owing to their high number and diversity, classification, nomenclature and formulation of these complexes cause problems. For this reason, a classification based on structural relationships and acidic properties should be of great interest. They are used industrially mainly in the reaction of hydration of alkenes (acid catalysis) and in the synthesis of methacrolein or isobutyric acid (redox catalysis). The first experiments concerning the activity of heteropolyanions in catalysis were performed in Japan’,2 but recently more and more studies devoted entirely to this field developed owing to the industrial challenge of finding alternatives for the usual acid catalysts (H2S04,HF) for pollution reasons.3p5 The purpose of this work is to have a better understanding of the acidic properties of heteropolyacids. Indeed, because of the considerable number of available polyanion structures, it is possible to vary their acidic properties as well as their redox ones.Molybdic polyacids, which can be easily reduced, display mainly redox properties, whereas tungstic acids are often stronger. This study concerns tungstic compounds whose acidic properties are not simultaneously linked to redox character. The aim is to measure their acidity using ammonia absorption microcalorimetry and also to determine the specific heats of the initial acids and the corresponding ammonium salts. Four tungstic compounds were studied (see Table 1). Two of them, H3PW,2040 and H4SiW12040, display a Keggin Table 1 Physical characteristics of the heteropolyanions formula molecular weight/g density/ g cmP3 a BET surface area/m2g - 2882 6.4 5 2933 3.3 130 2880 2948 3.7 - 5 3.3 5104 6.5 3.5 5206 4.3 - 4372 5.6 4 4474 5.8 - structure6 [Fig.l(a)] with, respectively, 3 and 4 protons for the same structure, allowing the study of the charge effect. These compounds display a tetrahedral symmetry based on a central X04 tetrahedron, surrounded by twelve M06 octahedra arranged in four groups M3013 of three edge- bridged octahedra. The two other compounds under study, H6P2W18062 and &P,W2,0,, (H20)3, are phosphotungstic heteropolyacids but they present structures that are different from the previous ones.[P2w&62]6-, with a ‘Dawson’ heteropolyanion structure6 is represented in Fig. l(b) and can be described in terms of two PW,03, groups (obtained by removing three W06 octahedra from a Keggin anion) connected by six terminal oxygen atoms. [P2W21071(H20)3]6-[Fig. 1(c)] is coniposed of two PW903, units linked by three WO,(H,O) octahedra. All the compounds can be prepared by acidification of aqueous solutions of the simple oxoanions and the required heteroatoms, for example, according to the reaction pattern: 12(WO,)’-+(HPO,)’-+23H+-+(PW12040)3-+ 12H,O The equilibrium constants and the formation rates are suffi- ciently high to allow the crystallization of the salts of the polyanions starting from stoichiometric mixtures of the compounds at room temperature.6 Table 1 shows the chemical formulae of the compounds and of the corresponding ammonium salts, the molecular weight of the anhydrous compounds, their density relative to water (measured in n-hexane) and their BET surface area.Their acidity was determined using ammonia absorption microcalor- imetry, since this is one of the most reliable methods for measuring the number and strength of acid sites of a ~atalyst;~ we discuss these results in terms of the macroscopic and microscopic properties of the samples. In particular, we shall consider their porosity. For this purpose, the samples are separated in two groups, those with a low BET surface area (< 10m2g-’) and those with a high BET surface area (> 100 m2 g-’).The former will be referred to as ‘non-porous’ even if they have some meso- or macro-pores, but which represent only a small part of the total surface area. The latter, referred to as ‘microporous’ compounds, have micropores (diameter < 1.5nm) which represent most of the surface area. These micropores are arguably zeolitic channels (see later). Some of the calorimetric data have been published in preliminary form’ but were not then related to porosity and particle size. Experimental Measured by using a pycnometer; bmeasured by N2 absorption after The solids were purchased by the Laboratoire de Chimie des evacuation of the sample during 2 h at 150 “C. Metaux de Transition (University of Paris VI). Thev were J.MATER. CHEM., 1994, VOL. 4 Fig. 1 Structure of various tungstic heteropolyanions: (a) structure of the Keggin heteropolyanion [PW,,0,,]3-; (h) structure of the Dawson heteropolyanion [P2W18062]6-;(c) structure of [P2W21071( H20)3]6- characterized by infrared and UV-VIS spectroscopies, polar- ography on mercury drop, X-ray diffraction and chemical analysis. The amount of water of hydration was determined by thermogravimetry. All compounds were used without further purification. The microcalorimetric study has been performed using a heat-flow microcalorimeter (HT from Setaram). The heteropolyacid sample was outgassed under vacuum (150"C, 1.33mPa) prior to any absorption of ammonia in the calorimetric cell maintained at 150°C. A 150°C temperature has been chosen in order to avoid too much physisorption and thus to favour the accessibility of the probe to the active acid sites of the solid.The experiment consists in admission of successive doses of reactive gas (NH,) on to the catalyst and waiting for the thermal equilibrium after each pulse. The pressure evolution is measured with a differential Barocel gauge (Datametrics) until a residual equi- librium pressure of ca. 100 Pa is reached. The evaluation of the absorbed volume and of the evolved heat for each dose allows us to plot the differential heat curve as well as the integral heat and the differential entropy as a function of the number of sites. Ammonia from Air Liquide (purity >99.9%) has been purified by successive freeze-thaw cycles after drying on sodium chips. The specific heats of the acids and the corresponding ammonium salts were determined using a differential scanning calorimeter (Setaram TG-DSC 11l),by a discontinuous scan- ning temperature programme, the temperature varying by 5°C steps and the heat consumed being measured for each increment.This method allows the sample to return to thermal equilibrium after each increment but needs two runs with and without a sample so as to measure the specific heat with accuracy. Results and Discussion Fig. 2 represents the differential heats of NH, absorption as a function of the sorbed volume for the four samples under study. It appears from these results that the heteropolyacids can be separated into two groups: those compounds showing a curve with a constant differential heat (H3PW1204, and H6P2W18062)and those for which it is continuously decreas- ing [H&W210,1(H20), and H,SiW,,O,,].However, in all cases, the initial heats of absorption are much higher than the values reported in the literature for acidic catalysts such as zeolites or simple oxides.' The total volume of sorbed ammonia can be deduced from these results, being 1004, 899, 1257 and 996 pmol g- (hydrated catalyst) for H,PWl2O4,, H4SiW,,040, H6P2W18062 and H6P2W2 I 071(H20)3, re-spectively. If these values are calculated as the number of NH, molecules for one heteropolyacid, the following values are obtained: 2.89 (H,PW,,O,,), 2.36 (H4SiW120,,), 5.0 (H6P2wl8o6,) and 4.65 [H6P2W210,1(H20),].These data show that most of the protons of the solids have reacted. So the reaction: H,(HPA)+xNH,+(NH,),(HPA) can be considered as complete in a first approximation and, as a consequence, it appears more reliable to speak of ammonia absorption instead of adsorption for these compounds. Most of our results are in disagreement with previous studies on the acidity of heteropolyacids. Indeed, studies using Hammett indicators showed that there is a distribution of the acidic strength for heteropolyacids with the Keggin struc- ture.l0,l' Our results disagree also with those of a recent paper where the reaction of H,PW120,, with ammonia was studied by microcalorimetry,12 but the activation conditions were not the same.In order to understand why our results are different we have undertaken a more detailed study of the absorption of NH, by H3PW1204,. This heteropolyacid was chosen for many reasons: (i) it is, to our knowledge, the only compound for which other microcalorimetric experiments were per-formed; (ii) many data on its acidic properties have been reported; (iii) we have previously studied its thermal behaviour and so we know its temperature range of ~tability;'~ (iv) the structure of its ammonium salt is isomorphous to that of H3PW12040-6H,0'4~'5and it was found that it has a porous structure like zeolites.16 The curves showing the variation of the differential heat of absorption of ammonia as a function of the pretreatment temperature of H3PW1204, are shown in Fig.3. When the solid is treated under vacuum at 150 or 250°C, the curves are quite similar and do not show a very important modification of the acidity of the proton. However, when the catalyst is evacuated at 400'C, the curve is very different with an initial heat of absorption of 140 kJ mol-' (200-210kJ mol-' when the pretreatment is performed at 150 "C) and the heat of absorption is continuously decreasing with the sorbed amount of ammonia. This difference can be explained easily if we take into account the fact that at temperatures below 300 "C the H3PWI20,, polyacid is in its J. MATER. CHEM., 1994, VOL. 4 -6 200 400 600 800 1000 1200 amount NHJ~O-~mol g-‘ Fig.2 Differential heat of NH, absorption uersus the ammonia uptake for samples: 0 H3PW120,,; 0 H4SiW12040;* H6P2W1,062; H6P2W21071(H20)3 0 .I I 1 I . I u m .I 0 200 400 600 800 1000 1200 amount NH~IO-~ mol g-’ Fig. 3 Influence of the outgassing temperature on the acidity of H,PW,,O,,: 0,150; 0,250; I?, 400 ”C anhydrous form while at 400°C its transformation into the anhydride form is ~omplete,’~this latter phase being meta-stable and leading slowly to the constitutive oxides. This thermal evolution of the polyacid can be described by the following equations: H3PW1204,,, 6H20-+6H20+H3PWI2O4,,(anhydrous acid) nH3PWl2O4,+3nH20 +[PW12038.5]n (anhydride phase) [PW12038.5]n+$nP205+ 12nW03(oxides) The anhydride phase can be described as Keggin units linked by W-0-W bridges. Its formation has been deduced from thermogravimetry and BET surface area measurements. By reaction with water, it transforms into the polyacid.Thus, we can conclude that, in the anhydrous form of the H3PWI2O4,,polyacid, all protons display the same acidity, which is very high and corresponds to super-acidic centres, as has been demonstrated by cataly~is,””~while the anhydride (or the melt of the constitutive oxides) phase is less acidic with a heterogeneous distribution of sites and strength. The results of Kapustin et all2 can be interpreted in terms of an intermediate structure between the anhydrous and the anhy-dride phases as their pretreatment temperature is 300’C but with an evacuation time of 50 h (as the transformation of the anhydrous acid into the anhydride phase is a kinetic phenom-enon, the longer the time of evacuation of the sample the more complete will be the dehydration).It seems more difficult to explain the differences between the present results and the J.MATER. CHEM., 1994, VOL. 4 studies of the acidity by the method of the indicators'0*'' which showed that there is a heterogeneous distribution of the acidic strength. However, from a chemical point of view, our results seem more realistic, as it is difficult to obtain a distribution of proton strengths and so of chemical environ- ments in a crystalline phase where the atoms are located in defined positions.The only case should be a delocalization of protons in the solid, but it is not reasonable to think that such a distribution could induce chemical environments so different that weak and strong acidic centres could be created. Moreover, it seems from literature data that using coloured indicators in aromatic hydrocarbons solutions is not the most reliable method for the characterization of the acidity of solid catalysts. For example, in the present case, while the base used for the titration of the protons goes inside the polyacid grain, the indicator stays at its surface and the observed result is probably correlated to the acidity of the surface, not of the bulk. It can also be pointed out that before these studies with Hammett indicators, it was generally admitted that all acid sites of the solid polyacids displayed the same strength.We can now analyse the results obtained with the other heteropolyacids. In a first step, we can compare the initial heats of absorption of ammonia (in kJ mol-' of NH,), which give a measure of the acidity of the compounds. The values are: 196 kJ mol-' for H,PW,2040; 185 kJ mol-' for H4SiWl2O4,; 164 kJ mol-' for H,P2W2,O7,(H2O),; 156 kJ mol-' for H6P2W1gO,,. These values show that the order of acid strength is: H3PW12040>H4SiW12040> H6P2W21071(H20),>H6P2W18062, in agreement with pre- vious results which prove that Keggin compounds are much more acidic than Dawson ones and that increasing the number of protons in the Keggin heteropolyacid decreases the acid ~trength.'~*".'~ These values are also in agreement with the results of Tzumi et al.who found, by ammonia TPD, that silica-supported tungstic heteropolyacids show the acidity order: H3PW,2040>H4SiW,20,0.20 They agree also with recent results on the acidity of several solids which showed that H3PW12040is more acidic than the H-ZSM-5 zeolite but less than SO~-/Zr02.21 Let us now discuss the shape of the curves in Fig. 2. The difference between H,PW12040 and H6PzW18062 on the one hand and H4SiW12040 and H6P,W21071(H20), on the other could be related to a different stability of the heteropolyacids, a partial degradation occurring for the last two, as observed above for H3PW1204, as a function of the temperature of prior activation (Fig.3). However, differential thermal analysis showed that the stabilities of these compounds decrease in the order: H,PW12040 =H6P2W2,07,(H20), > H4SiW12040>H6P2W18062, and, thus, such an explanation is not appropriate. In order to explain these results, it is neces- sary to study the decreasing slopes of the calorimetric peaks for each pulse of ammonia. Indeed, the decrease of the heat can be approximated by an exponential function AH =AHo exp(-t/z) where t is the time and z the thermokinetic parameter. The value of z gives an indication of the speed of the sorption reaction and also in the case of heteropolyacids it allows a study of the diffusion of ammonia through the crystal. Fig.4 shows the evolution of the thermokinetic parameter for the four experiments presented in Fig. 2 as a function of the sorbed volume. For H3PW12040. the thermokinetic parameter does not increase with the amount of ammonia. This result is in agreement with the studies of Moffat et al. who showed that the ammonium salt of the phosphotungstic acid has a porous str~cture.'~*'~-~~ Thus, it is easy to understand that ammonia can diffuse very easily through the solid: it reacts first with the protons near the surface forming (NH4),PW1204,, which has a porous structure allowing NH, to react more easily with the acidic centres inside the solid. For H,SiW,,O,,, the same authors reported a similar stru~ture,'"'~ but the thermokinetic parameter shows that there is a drastic increase with the amount of sorbed ammonia, the diffusion through the crystal occurring more and more slowly.In order to elucidate this point, we undertook the study of the ammonium salt of H4SiW12040. We prepared it by the same way as Moffat (precipitation from an aqueous solution of H,SiW120,, by addition of ammonium carbonate, solid 1) and measured its BET surface area. This compound was then dissolved in hot water and solid 2 was obtained after precipi- tation by ammonium chloride. The BET surface area of solid 1 was ca. 120 m' g-' in agreement with the results of Moffat et al. and decreased to ca. 50 m2 g-' for solid 2. In addition, keeping solid 1 in a desiccator under ammonia pressure (by contacting it with NH3 vapour arising from some drops of a concentrated ammonia solution) led to solid 3 which had a BET surface area of 3.3 m' g-'.The infrared spectra of solids 1-3 are very similar to those reported for H4SiW12040 and its salts." Indeed, if a degradation of the anion occurred amount NH3/1o4 mol g-' Fig. 4 Thermokinetic parameter (in seconds) as a function of the ammonia uptake for: 0,H,PWl,O,O; a, H,SiW,,O,,,; Ir, H,P,W,,O,,; 0, H6P,W210,1(H,O), J. MATER. CHEM., 1994, VOL. 4 during the treatment by ammonium carbonate and/or ammonia, the first degradation product should be SiWl10;;, whose infrared spectrum is very different from that of SiW120:,.26 For example, the v(Si-0) band, found at 1020cm-' for SiW,,O& is shifted to lo00 cm-' in SiW,,O&.As a consequence, we can assume that the polyanionic struc- ture remains intact in compounds 1-3. Fig. 5 shows the X-ray diffraction spectra of solids 1 and 3, of K,SiWl,O,, and (NH4)3PW12040. lead to ident- Solid 1 and (NH4)3PW12040 ical XRD patterns showing that their structures should be quite similar, in agreement with their high BET surface areas. On the other hand, solid 3 and the potassium salt of the silicotungstic acid give patterns very different from the pre- vious ones. It must also be pointed out that infrared experi- ments showed that the heteropolyanion structure had been retained in these two compounds. Unfortunately, it was not possible to record the X-ray diffraction pattern of the sample of H4SiW,2040 which has been submitted to react with ammonia in the calorimetric cell but we can assume that its spectrum should be quite similar to that of solid 3. The XRD spectra of solid 1 and (NH4)3PW,2040 are also similar to those of H3PW12040.6H20 [Fig.5(e)]whose struc- ture has been refined by Brown et Only small variations of position arise from the smaller width of the peaks of the hexaaquo acid and from the replacement of (H502)+ by NH;. Intensity variations can be due to the different cations and/or to preferential orientations. These authors pointed out also that many alkaline salts of heteropolyanions showed similar XRD patterns and they gave the examples of the caesium salts Cs3PW1204,, Cs,HSiW,,O,, and CS,H,BW,~O,~. It is well known that CS~PW,,O~~ has a 2oool I1000 1 1 50011 I1 2000~ 1 1 20 40 60 80 2tYdegrees Fig.5 X-Ray diffraction patterns of: (a) solid 1 [(NH,),HS~W120,,~xH,0(x=5-6)]; (b) solid 3 [(NH4),SiW1204,-5 H201; (4 WLM"12040.8H20; (4 K4SiW12040*6H20;(4H3PW,,040~6H20 porous structure as (NH4)3PW,,0,,,16~22-24and so the follow- ing interpretation can be proposed. The alkaline salts of heteropolyanions having an XRD pattern similar to that of H3PW1,O4,-6H2O have a microporous structure and corre- spond to salts with only three alkaline cations [such as (NH4)3HSiWl,0,0]. The XRD patterns of solid 3 and of K4SiW12040 correspond to compounds with four alkaline cations and with no microporosity, as it can be shown by the BET surface-area measurements (K4SiW12040 has a BET surface area lower than 2m2g-').It is surprising to obtain (NH4)4SiW,,0,0 by the reaction of ammonia at room tem- perature, only the triammonium salt should be obtained by reaction at 150"C during the microcalorimetric experiment. However, the experimental conditions are completely dii€erent, time and ammonia pressure being some orders of magnitude higher in the first case. In addition, the presence of water in the reaction leading to solid 3 probably has a non-negligible role. Indeed, if it is well known that the structure of the porous water-insoluble heteropoly salts is not dependent on their water content (as water is only physisorbed), very probably there is initially reaction of water with the protons of the solid, leading to the formation of the highly mobile (H30)+ species which can then react with ammonia at the surface of the grain.Water can then be considered as a catalyst of the reaction. However, two questions remain. First, H3PW120,0~6H,0 should be microporous, but experimentally its BET surface area is found to be very small. This could be due to slightly different positions of the cations in the structure or to a blocking effect of residual water. This problem is, however, under study.27 Secondly, McMonagle and Moffat report16 the preparation of a non-porous ammonium 12-phosphotungstate salt by reaction of the acid with ammonia at high temperature. However, this result was not reported in their subsequent papers on this subject.As the preparation method was not well described, in particular the pretreatment of the polyacid and the quantity of ammonia introduced in the cell, it is difficult to discuss this point. Indeed, the thermokinetic param- eter increases with the activation temperature of 1Zphosphotungstic acid showing that there is a blocking effect due to the surface degradation of the poljanion. Moreover, if a high quantity of ammonia was introduced, it is also possible that this base reacts not only with the protons but also with the polyanionic species, leading also to a blocking effect due to a modification of the surface of the grains. In these two cases, the X-ray diffraction patterns should not be modified (in agreement with the experience) but the BET should not show microporosity.The shape of the microcalorimetric curves for the silico- tungstic acid can now be explained easily. Indeed, there is initially formation of the triammonium salt which is micro- porous and during the course of the experiment this salt reacts with ammonia leading to the formation of the tetra- ammonium salt. But this latter compound is not porous and ammonia diffuses through it very slowly leading to (i) an increase in the thermokinetic parameter and (ii) an apparent non-stoichiometric reaction, as it proceeds more and more slowly. Thus, if one waits sufficiently long, one will obtain an almost stoichiometric reaction, leading to nearly complete neutralization, with essentially no decrease of the initial heat of absorption with an increasing quantity of ammonia admit- ted.However, the decrease of the reaction rate with time makes the evaluation of the heats difficult, because it is not easy to integrate quantitatively broad signals. This explains the apparent decrease of the absorption heats (Fig. 2). It has been proposed by some authors that the crystals of alkaline salts of phosphotungstic acid are non-porous by themselves, the microporosity being in fact the space between J. MATER. CHEM., 1994, VOL. 4 mY' 100 120 140 160 180 200 7°C Fig. 6 Heat capacities C, as a function of temperature for acidic heteropolyanions (continuous line) and for their corresponding ammonium salts (dashed line): 0,H,PW,,O,,; 0,H,SiW,20,,; *,H6PZW18062;0,H6P2WZ1071(H20)3 ultrafine non-porous particles which were detected by electron microsc~py.~,~~This idea is also consistent with the structure determined by Brown et a2.I4 However, it seems to us that such an explanation is not reasonable for the following reasons.(i) If we assume that the microporosity is due to the interparticle space, such a phenomenon should also be observed for highly dispersed oxides such as Aerosil but it is not; (ii) the ultrafine particles have a diameter corresponding to less than 10 unit cells and so the X-ray diffraction pattern should be very broad, while in some cases we obtained very nice spectra allowing a structure determination by Rietveld refinement; (iii) finally we studied the p-xylene adsorption on the potassium salt of phosphotungstic acid and we found the same curve than for the ZSM-5 zeolite, in agreement with an intrinsic microporosity.As a consequence, it is necessary to assume that the alkaline salts of phosphotungstic acid are bidispersed systems with ultrafine particles and big crystals. This duality, which was previously observed for supported heteropolyacid~,~~*~~is responsible of the broadening of the foot of the X-ray diffraction peaks.27 As no data are available on the ammonium salts of H6P2W&6, and H,P,W,,O,,( one can only propose that a similar behaviour as above occurs for these compounds, the shapes of the curves being related to the porosity of the material. However, it is worth noticing the presence of two maxima in Fig.4 for H,P,W,,O,,(H,O)3 which could be explained by the formation of a mixed intermediate salt. Fig. 6 shows the evolution of the specific heats as a function of the temperature for the acidic and ammonium compounds of the four polyanions which were studied here. The determi- nation has been performed between 100 and 2OO0C, the temperature increasing by increments of 5 "C.Differences are more or less important between the acid and the correspond- ing ammonium salt, depending on the sample. For example, H3PW,,0,, presents a decrease of C, when going from the acid to the ammonium salt. This result is in good agreement with the fact that its density decreases after absorption of ammonia (Table 1).Indeed, the ammonium salt structure is more microporous and organized than the correspond- ing acid. In the same way, a decrease of C, for [P,W,,07,(H,0)3]6-is observed when going from the acid to the ammonium salt, corresponding to a decreasing density. On the contrary, the two other polyanions, whose density is only slightly modified by ammonia absorption, lead to a significant increase in C, when going from the acid to the salt, in agreement with the absence of microporosity and the lower organization of this latter. Conclusions The calorimetric study of the acidity of tungstic heteropoly- acids by ammonia absorption has shown that these are very acidic compounds displaying high absorption heats, which accounts for their use as substitution catalysts for sulfuric acid.The acidity varies between compounds, which so giving inorganic solid compounds of different acidities but with sites of the same strength. Diffusion phenomena are important for some of the compounds and are related to the structure of the heteropolyacids and their ammonium salts. F.X. L.-C. thanks the CNRS for financial support. References Y. Onoe, Kagaku Kogyo, 1975,26,355. Y. Oda and K. Uchida, Jpn. Petrol. Inst., 1977, 20, 1054. I. V. Kozhevnikov, Russ. Chem. Rev., 1987,56.811. M. Misono, Catal. Rev.-Sci. Eng., 1987,29, 269. I. V. Kozhevnikov and K. I. Matveev, Appl. Catal., 1983,5135. M. T. Pope, Heteropoly and Isopolyoxometalates (Inorganic Chemistry Concepts, Vol. 8), Springer Verlag, New-York, 1983. A.Auroux, J. C. Vedrine and P. C. Gravelle, Stud. Surf. Sci. Catal., 1982, 10, 305. A. Auroux, F. X. Cai and F. Lefebvre, Calorim. Anal. Therm., 1990,20-21,103. J. MATER. CHEM., 1994, VOL. 4 9 A. Auroux and J. C. Vedrine, Stud. Surf. Sci. Catal., 1985,20,311. 20 Y. Izumi, R. Hasebe and K. Urabe, J. Catal., 1983,84,402. 10 A. K. Ghosh and J. B. Moffat, J. Catal., 1986,101,238. 21 T. Okuhara, T. Nishimura, H. Watanabe and M. Misono. J. Mol. 11 B. Viswanathan, M. J. Omana and T. K. Varadajan, Znd. J. Chem., Catal., 1992,74,247. 1988,27A, 674. 22 J. B. Moffat, J. Mol. Catal., 1989,52, 162. 12 G. I. Kapustin, T. R. Brueva, A. L. Klyachko, M. N. Timofeeva, S. M. Kulikov and I. V. Kozhevnikov, Kinet. Catal., 1990,31,896. 23 D. B. Taylor, J. B. McMonagle and J. B. Moffat, J. Colloid Interface Sci.,1985, 108, 278. 13 14 15 A. Oulmekki, D. Olivier, G. Herve and F. Lefebvre, C. R. Acad. Sci. Paris,Sbie ZI, 1990,311, 619. G. M. Brown, M. -R. Noe-Spirlet, W. R. Busing and H. A. Levy, Acta Crystallogr. B, 1977,33, 1038. P. Pascal, Nouveau Traitk de Chimie Minkrule, Volume XIV, 24 25 26 J. B. Moffat, J. B. McMonagle and D. B. Taylor, Solid Statr Ionics, 1988,26, 101. C. Rocchiccioli-Deltcheff, R. Thouvenot and R. Franck, Spectrochim. Acta Part A, 1976,32, 587. C. Rocchicioli-Deltcheff and R. Thouvenot, J. Chein. Rex 16 17 18 Masson, Paris, 1959, p. 964. J. B. McMonagle and J. B. Moffat, J. Colloid Interface Sci., 1984, 101,479. K. Nowinska, J. Chem. Soc., Chem. Commun., 1990,44. A. Oulmekki and F. Lefebvre, React. Kinet. Catal. Lett., 1992, 27 28 29 30 Synopses, 1977,46. F. Lefebvre and B. F. Mentzen, unpublished results, 1992. N. Mizuno and M. Misono, Chem. Lett., 1987,967. F. Lefebvre, J. Chem. Soc., Chem. Commun., 1992,756. A. Oulmekki and F. Lefebvre, React. Kinet. Catal. Lett., 1992, 48, 601. 48,607. 19 B. Viswanathan, M. J. Omana and T. K. Varadajan, Catal. Lett., 1989,3, 217. Paper 3/03391K; Received 14th June, 1993