J. CHEM. SOC. FARADAY TRANS., 1994, 90(14), 2141-2146 Group Behaviour of SAP041 Molecular Sieves containing Various Metals (Mg, Zn, Mn or Cd, Ni, Cr) Jan Kornatowski" lnstitut fur Kristallographie und Mineralogie ,J-W.Goethe-Universitat, 60054 Frankfurt am Main, Germany and lnstytut Chemii, Uniwersytet M. Kopernika, 87-100 Torun, Poland Gerd Finger,? Karin Jancke and Jurgen Richter-Mendau Zentum fur Heterogene Kataluse, 12484 Berlin, Germany Dietrich Schultze Bundesanstalt fur Materialforschung und-prufung , 12484 Berlin, Germany Werner Joswig and Werner H. Baur lnstitut fur Kristallographie und Mineralogie, J-W. Goethe-Universitat, 60054 Frankfurt am Main, Germany Derivatives of SAPO-11 containing Mg, Zn, Mn, Cd, Ni and Cr as heteroatoms have been synthesized.These additions influence considerably both the properties and the morphology of the crystalline phases synthesized. The products noted as MeAPSO-11 were investigated with light microscopy, scanning electron microscopy (SEM), electron microprobe analysis (EPM), X-ray diffraction (XRD), temperature-programmed XRD, differential thermal analysis (DTA), thermogravimetry (TG), magic-angle spinning nuclear magnetic resonance (MAS NMR) and wet chemical analysis. The materials can be clearly divided into two groups: (1) Mg, Zn or Mn, or (2) Cd, Ni or Cr containing preparations. The results suggest that all metals are incorporated into the framework of SAPO-11 molecular sieve, though in rather different amounts. The content of heteroatoms decides the properties of the synthesized molecular sieves.Since 1982, when aluminophosphate molecular sieves were discovered,' several tens of structures and over two hundred compositions of this family have been reported, many of them by investigators from Union Carbide (now UOP).,v3 Numerous materials have been claimed on the basis of iso- morphous substitution of silicon and many other elements ('hetero-atoms') for phosphorus and/or aluminium in most structure types, as reported among others in ref. 2-4 and claimed in patents such as those given in ref. 5. The incorporation of Si and/or other heteroatoms into the framework of microporous aluminophosphates (AlPO,-TI) can result in the formation of acidic groups which can be catalytically active. It would be desirable if the properties could be controlled and fitted to a particular catalytic process by substitution of different heteroatoms in various amounts.In spite of the large number of materials ~ynthesized,~-~ conclusive evidence for incorporation of heteroatoms into the aluminophosphate molecular sieve frameworks is still not available in most cases. In particular, for AlPO,-ll and its derivatives (structure type code AEL6), the number of papers reporting isomorphous substitutions of metals is surprisingly low, e.g. ref. 7-13. For this reason we decided to synthesize the SAPO-11 molecular sieves containing Mg, Zn, Mn, Cd, Ni and Cr as heteroatoms. We noticed a considerable influ- ence of these heteroelements on the morphology of the resulting crystals.In testing for evidence for the incorpor- ation of these heteroatoms into the framework of SAPO-11, we discovered a surprising group behaviour of the synthe- sized molecular sieves. Experimental AEL-type molecular sieves were synthesized hydrothermally in PTFE-lined autoclaves in air-heated ovens at 463 K for t Present address: Markische Allee 84, 12681 Berlin, Germany. 36 h from gels of formal molar composition: (1 -42) A1203 . a Me"0 -P,05 + b Si02 -4.0DPA * 28OH,O * xH~SO, (1 -u) Al,03 * a Memo P20, * b SiO, 4.0DPA . 280H20 . xH2S04 with (I b molecular sieve AlPO,-ll - 0.1 SAPO-11 0.1 0.1 MeAPSO-11 Preparation of the gels was performed as for the AFI structure type materials,6 i.e.A1P04-5 l4 and SAPO-5," with the only difference that (a) di-n-propylamine (DPA) was used as templating agent instead of triethylamine, (b) for the MeAPSOs, Me" or Me'" sulfates were added before the DPA-H3P0,-H,0 solution was admixed. The sulfuric acid was used to adjust the pH value of the gels to 3.5 & 0.2. After a period of crystallization, the autoclaves were cooled and the resulting materials were decanted, filtered, washed, dried at 378 K, and sieved, if necessary. Calcination of the samples was performed in air for 12 h. The temperature was 873 K for the Mg, Zn and Mn preparations and 1023 K for the others, corresponding to the different thermal stability of the materials (cf. Results and Discussion). The crystals were examined by light microscopy, SEM, EPM, XRD, tempera- ture-programmed XRD (Guinier technique), DTA, TG, MAS NMR and wet chemical analysis.The Guinier XRD patterns were measured by an Enraf Nonius high-temperature X-ray camera within the tem-perature range 298-1273 K. The thermogravimetric investigations were performed by simultaneous TG-DTA measurements in flowing air or nitro- gen using the SETARAM thermobalance TAG 24. The 13C, 27Al, 29Si and 31P MAS NMR spectra were recorded on a Bruker MSL 400 spectrometer (details will be published in a separate paper16). Results and Discussion For all reaction batches, the experiments were successful in obtaining crystals of AEL-type molecular sieves. In most cases, the materials also contained some byproduct(s) which usually formed separate crystalline phases of different dimen- sions, habitus and specific density.Therefore, they could be completely or at least mostly separated by a simple decanta- tion treatment and/or by sieving. The X-ray patterns of the samples purified in this manner and used later for subsequent investigation are shown in Fig. 1. The appearance of the most frequently occurring bypro- duct is presented in Plate 1. It was identified as SAPO-31 by X-ray and 29Si MAS NMR methods." In spite of the fact that our procedure of synthesis was similar to that for growing large crystals of AFI-type material^,'^^' the present syntheses did not yield well formed monocrystals of AEL type. In every batch, the AEL phase was composed of growth aggregates, as shown in Plate 2.The only common feature in their appearance was a quasi-1 5 10 15 20 25 30 28/degrees Fig. 1 X-Ray powder patterns (Cu-Ka) of the investigated AEL samples: (a) AlPO,-11, (b) SAPO-11, (c) MgAPSO-11, (d) ZnAPSO-11, (e) CdAPSO-11, (f)MnAPSO-11, (9)NiAPSO-11, (h) CrAPSO-11 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 Content of metals (Me) and silicon in the investigated AEL samples a/(mol.%) blatoms U.C. -a sample Si Me Si Me 1(Si + Me) SAPO- 1 1 4.1 - 1.6 - 1.6 MgAPSO-11 ZnAPSO-11 2.6 3.5 5.7 5.7 1.0 1.4 2.3 2.3 3.3 3.7 CdAPSO-11 3.2 1.1 1.3 0.4 2.7 MnAPSO-11 2.9 3.4 1.2 1.4 2.6 NiAPSO- 1 1 4.7 1.0 1.9 0.4 2.3 CrAPSO- 1 1 3.8 0.4 1.5 0.1 1.6 u.c., unit cell.hexagonal symmetry of the external habitus of the aggregates. The differences in morphology must obviously be related to the presence of various cations in the reaction gels. The dimensions of the growth aggregates were usually smaller than 80 pm. The morphology of the aggregates can be divided clearly into two groups: more sphere-like crystals for the Cd, Ni and Cr preparations and more spindle-like elon- gated crystals for the Mg, Zn and Mn preparations (Plate 2). Qualitative electron microprobe analyses (EPM) showed for all products that the heteroelements, except for Mg, added to the reaction batches were present in the MeAPSO- 11 crystals in easily observable amounts. Wet chemical analyses supported these findings (Table 1) showing a con- siderable content of the heteroelements in the samples investi- gated, also in the case of Mg.This indicates that the Mg atoms were most likely located in the bulk of the crystalline aggregates and not in the surface layer analysed by EPM. The X-ray patterns of the samples containing various het- eroatoms differ characteristically in their peak intensities. For the AlP0,-11 and SAPO-11 samples, the peak at about 28 = 21" shows the highest intensity. The same holds true for the Cd-, Ni- and Cr-containing samples which have a rela- tively low content of heteroatoms (Table 1). In contrast, for the samples with the highest heteroatom contents, namely ZnAPSO-11, MgAPSO-11 and MnAPSO-11, the intensities of the peaks at about 28 = 23" and 21" are similar.On the basis of the temperature-programmed XRD pat- terns (Plate 3), the samples may again be divided with respect to their thermal behaviour (Table 2) into the same two groups that are apparent from the morphology and the inten- sities of the lines in the XRD powder patterns: (a) AlPO-11 and SAPO-11, as well as the Cd-, Ni- and Cr-containing samples, remain thermally stable up to about 1300 K. In the case of the Zn- and Mg-containing materials, the spectra show the beginning of a transformation into the tridymite structure at about 973 K and for Mn at about 1073 K. (b) An irreversible change of lattice parameters shown by a shift of Table 2 Thermal behaviour of the investigated AEL samples sample TchangcelK TStabblK AlPO,-ll ca. 523 >1300 SAPO- 1 1 ca.523 >1300 CrAPSO- 1 1 ca. 523 >1300 NiAPSO-11 ca. 523 >130 CdAPSO-11 ca. 523 >1300 MgAPSO-11 ca. 673 973 ZnAPSO-11 ca. 673 973 MnAPSO-11 ca. 723 1073 a Temperature at which change of lattice parameters occurs. * Temperature up to which the sampje is thermally stable. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Plate 1 Scanning electron micrograph of the main byproduct occurring in as-synthesized AEL phases: SAPO-31 J. Kornatowski et al. (Facing p. 2142) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 (61 Plate 2 Morphology of crystalline growth aggregates of the (Me)AEL phases synthesized (SEM): (a)SAPO-11, (b)CdAPSO-11, (c) NiAPSO-11, (d) CrAPSO-11, (e)AlP0,-11, (f)MnAPSO-11, (9)MgAPSO-11, (h)ZnAPSO-11 J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2Bldegrees Plate 3 Guinier-Lenne powder XRD pattern for the AEL materials investigated (*Pt-reflexes): (a)A1P04-11, SAPO-11 and MeAPSO-11 (Me: Ni, Cd, Cr); lattice parameters change at ca. 523 K, stable up to ca. 1300 K; (b) MnAPSO-11; lattice parameters change at ca. 723 K, transformation into tridymite at CQ.1073 K; (c) Mg- and Zn-APSO-11; lattice parameters change at ca. 673 K, transformation into tridymite at ca. 973 K J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2143 A I-..I *.I . I...I...l...I...I -40 -60 -80 -100 -120 '-140 6 Fig. 2 29Si MAS NMR spectra of the dehydrated calcined samples: (a) SAPO-11, (b) CrAPSO-11, (c) NiAPSO-11, (d)CdAPSO-11, (e) MgAPSO-11 and (f) ZnAPSO-11 the positions of XRD reflections is observed at about 523 K for the first group and at about 673 K for the second group of thermally less stable materials (Table 2).Similar effects were reported for AlPO,- 1 1, MgAPO-39 and MgSAPO-46.'8 The 29Si MAS NMR spectra indicated that the synthesized materials really contained Si within the framework, that is they were SAPO-11 (Fig. 2). This is shown by a narrow line at ca. -95 ppm, which can be assigned to isolated silicon atoms replacing phosphorus atom^.'^ This line dominates in the spectra measured for the group of SAPO-11 and Cr-, Ni- and Cd-containing samples. In the spectra of Mg- and Zn- containing samples, this line has an intensity similar to that of the second line at ca. -11 1 ppm, which is not related to isolated silicon atoms.16 Thus, the materials can be again divided into the same two groups based on their 29Si MAS NMR spectra.Thermal analyses of the synthesized materials were per- formed under streaming nitrogen (Fig. 3) and air (Fig. 4). Obviously, the calcination process proceeds in several stages (four under nitrogen, five under air), the DTA maxima of which are compiled in Table 3. Similar behaviour has been reported for A1P04-5 2o and metal-substituted A1P0,-5.21 After desorption of water (stage I), stages 11-IV reflect mainly an endothermal desorption and decomposition of the template molecules as observed under nitrogen. Under air, these processes are predominated by oxidation occurring in parallel and exothermal effects are observed.For AlPO,-11 itself, stage I1 is the only process of template removal which is in accordance with the TG measurements reported in ref. 18. Stage I11 remains an extremely weak effect for all studied preparations. The main process of the removal of the tem- plate proceeds under nitrogen in stages I1 and/or IV (Fig. 3). 373 573 773 973 1173 T/K B -24 t I 373 573 773 973 1173 TIK Fig. 3 Thermoanalytic curves of the AEL samples measured under flowing nitrogen atmosphere: A, DTA; B, TG. (a) AlP0,-11, (b) SAPO-11, (c) CrAPSO-11, (6) NiAPSO-11, (e) CdAPSO-11, (f) MgAPSO-11, (9)ZnAPSO-11, (h)MnAPSO-11.Table 3 Total mass loss and temperature of the DTA peaks for the AEL samples investigated (a) Nitrogen atmosphere (all four effects are endothermal) Tp, minK total mass sample I I1 111 IV loss (%) ~~ AlP0,-1 1 370 555 -10.8 SAPO-1 1 370 545 740 9.0 CrAPSO-1 1 370 -745 9.7 NiAPSO-1 1 370 560 745 7.0 CdAPSO-11 370 560 7 30 8.0 MgAPSO-1 1 370 -770 5.9 ZnAPSO-1 1 370 -755 5.7 MnAPSO-1 1 370 -? 760 7.3 (b) Air atmosphere (all effects, except for I, are exothermal) T,,maxK total mass sample I I1 I11 IV v loss (%) AlP0,-1 1 340 585 ---12.1 SAPO-1 1 340 560 685 735 775 10.1 CrAPSO-1 1 340 -685 735 795 11.8 NiAPSO-11 340 570 685 750 825 9.8 CdAPSO-11 340 545 -725 820 9.4 MgAPSO-11 340 --740 920 9.6 ZnAPSO-1 1 340 --740 870 10.3 MnAPSO-11 340 -685 740 860 8.8 A I 373 573 773 973 1173 1373 TIK B hs "f Y -3 -1 01 373 573 773 973 11'73 1373 TIK Fig.4 Thermoanalytic curves of the AEL samples measured under flowing air atmosphere. Labelling as for Fig. 3. Evidently, in stage IV, which is only slightly influenced by the heteroatoms present, a cracking and partial desorption/ oxidation of the template takes place. Certain amounts of these cracking products remain as coke, which can be burned out under air during stage V. This mechanism is supported by the experimental observation that samples calcined under air have a higher loss of mass than those calcined under nitrogen (Table 3). Note that stages IV and V are also reflec- ted in one or two corresponding steps in the TG curves [Fig.3(b) and qb)]depending on the atmosphere in which the cal- cination is performed. In contrast to stage IV under nitrogen, [Table 3(a) and Fig. 3A] the position, intensity and width of the exothermal effects in stage V under air [Table 3(b)and Fig. 4A] (burning out of the cracking products) are distinctly dependent on the heteroatoms present. In this process, SAPO-11 and CrAPSO- 11 are similar in their behaviour because a substitution of Crnl for A1 does not create any additional framework charge, i.e. apparently no new active centres. However, the effect in stage V is much stronger for CrSAPO-11 since stage I1 is not observed for this molecular sieve. This indicates that the low Cr"' content (Table 1) is suficient for a considerable change in the thermal stability of the template molecules.One might then expect the occurrence of interactions between the Cr and Si heteroatoms. The incorporation of Me" creates additional framework charges (additional active centres), which is followed by more remarkable changes of the exothermal peak V. Its broadening (Fig. 4) indicates the occurrence of differently bonded pro- ducts of template decomposition which cannot be burned out within a narrow and well defined range of temperatures. Their exothermal effects overlap forming a broader band. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Zn Mn Cd Ni Cr Si 760 I 1 I I > 4 .l '3I L 3 heteroatoms per U.C.Fig. 5 Dependence of the temperature of oxidative burning out of the template (stage V under air) on the content of heteroatoms in the Me, (a) (Me + Si).MeAPSO-11 samples. (0) Another clear change of the exothermal peak V is the shifting of its position with temperature. As is seen from Fig. 5, the temperature of oxidative burning out of the template increases linearly with the heteroatom content. This is true for both the Me heteroatoms and for the sum Me + Si. The parallel course of both lines (Fig. 5) indicates that Si plays no deciding role for the oxidation of the template. Consequently, the group of Mg-, Zn- and Mn-containing samples, which have the highest heteroatom content (Table l),shows: (a)the highest temperature of burning out of the template (Fig.5); (b)lack of stage I1 (Fig. 4); (c) the lowest thermal stability, as observed from temperature-programmed XRD. The total loss of mass (Table 3) decreases linearly with the number of heteroatoms per unit cell [Fig. 6(a)-(d)]. The most distinct effect is observed for the dependence of the mass loss on the sum Me + Si under a nitrogen atmosphere [Fig. 6(a)]. The same tendency, though weaker, holds for the Me hetero- atoms alone (without Si) [Fig. 6(b)]. Under an air atmo- sphere, such a tendency is only slightly pronounced for Me + Si [Fig. qc)] and practically no longer observed for Me heteroatoms alone [Fig. qd)].Such a dependence can be explained if one accepts that both Me and Si heteroatoms create charged centres which are able to bond to the template molecules: the higher the number of heteroatoms, the strong- er the bonding of the template and thus a clearly lower loss of mass under a nitrogen atmosphere (thermal decomposition of the template).Under an air atmosphere, the effect is weaker as the thermal processes are accompanied by oxida- tion reactions occurring in parallel. The above observations allow construction of the following interpretation of the thermogravimetric effects. Stage I1 at ca. 545-585 K reflects the removal of those template molecules that are simply occluded in the pores and able to diffuse freely during heat treatment. Such a situation occurs in 'pure' A1P04-11 and partially in the less substituted samples (owing to the appearance of stages IV/V and considerably weaker effects in stage 11). The other template molecules are bound to heterocentres.In order to leave the pore system, they have to be decomposed (stages I11 and IV) and/or oxidized (stage V). Therefore, the thermal effect of the decomposition under nitrogen is nearly temperature independent, as opposed to that of the oxidation reaction which shows a linear depen- dence on the content of bonding heterocentres. This linear dependence of the thermogravimetric effects on the number of heteroatoms can be accepted as a good indication of the incorporation of these heteroatoms into the framework. Otherwise, their influence would be random. It remains still an open question why the total mass loss differs between par- ticular samples under both nitrogen and air.A possible J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 l2 AlPOl210 jAlPO4-11 Cr i Si h ,\" 8-Y 8 6- :d 61 Ni Mn YgZn 24\ (a) , , , ~ 0 0 1 2 3 4 OJ 0 1 2 I (Si + Me) per U.C. I Me per U.C. Cr Zn. Mn Mg 0-l 0 1 2 (Si + Me) per U.C. Me per U.C. Fig. 6 Loss of mass at calcination as a function of the sum of heteroatoms (a), (c) or Me heteroatoms (b), (4under N, (a),(b) and air (c), (d) atmospheres explanation could be the bonding of the template molecules to the heterocentres as well. If such bonding is really able to hinder the diffusion of the template molecules, it should result in an immobilization of the template molecules during the synthesis process and, consequently, in a lower density of their packing.Conclusions All of the results presented show that the prepared materials differ with respect to both their morphology and properties and they can be clearly divided into two groups: (1) Mg, Zn, Mn or (2) Cd, Ni, Cr containing materials. The results show a strong influence of all heteroatoms even in the case of low contents of some of them. This indicates that the heteroatoms should not be only simply occluded within the pore system. None of the methods used offers by itself a definite proof for incorporation of the heteroatoms into the framework of SAPO-11. That holds true also for the elements which are commonly accepted as possible candidates for an incorpor- ation into microporous aluminophosphates, i.e.Mg, Zn and Mn. The group behaviour of the preparations investigated by us is likely to be due more to variations in the amount of particular heteroatoms present in the samples than to their physico-chemical nature. Thus, the nature of the metallic het- eroatoms and/or their compounds used for the syntheses seems to control the amount of incorporated heteroatoms. The content of heteroelements influences primarily the properties of the synthesized molecular sieves. The morphol- ogy, XRD and temperature-programmed XRD results, and especially the linear dependences of the thermal behaviour of the samples suggest that all of the metals studied can be incorporated into the framework of SAPO-11. The very low contents of some of the heteroatoms agree with the findings of Kevan and co-worker~~-~' and with our results for a group of MeAPO-3 1 materials2' The authors thank Dr.B. Zibrowius for MAS NMR mea-surements and stimulating discussions. The work was par- tially supported by the Bundesministerium fur Forschung und Technik, the Deutsche Forschungsgemeinschaft, and by the Polish Committee of Scientific Research (K.B.N.). References 1 S. T. Wilson, B. M. Lok and E. M. Flanigen, US.Pat., 4 3 10440, 1982. 2 E. M. Flanigen, B. M. Lok, R. L. Patton and S. T. Wilson, in New Developments in Zeolite Science and Technology, ed. Y. Murakami, A. Jijima and J. W. Ward, Stud. Surf. Sci. Catal. 28, Elsevier, Amsterdam, 1986, p.103. 3 E. M. Flanigen, R. L. Patton and S. T. Wilson, in Innooation in Zeolite Materials Science, ed. P. J. Grobet, W. J. Mortier, E. F. Vansant and G. Schulz-Ekloff, Stud. Surf. Sci. Catal. 37, Else-vier, Amsterdam, 1988, p. 13. 4 S. T. Wilson and E. M. Flanigen, in Zeolite Synthesis, ed. M. L. Occelli and H.E. Robson, ACS Symp. Ser. 398, Am. Chem. SOC., Washington, 1989, p. 329. 5 J. Kornatowski, M. Rozwadowski and G. Finger, Pol. Pat. Appl. 291 455, 1991; 291 459, 1991; 291 460, 1991; J. Kornatowski, M. Rozwadowski, G. Finger and W. H. Baur, Pol. Pat. Appl. 291456, 1991; 291458, 1991; B. M. Lok, C. A. Messina, R. L. Patton, C. F. Gajek, T. R. Cannan and E. M. Flanigen, U.S. Pat., 4440871, 1984; B. M. Lok, B. K. Marcus and E.M. Flani-gen, U.S. Pat., 4500651, 1985; C. A. Messina, B. M. Lok and E. M. Flanigen, U.S.Pat., 4 544 143,1985; S. T. Wilson and E. M. Flanigen, U.S. Pat., 4567029, 1986; L. J. Wright and N. B. Mile- stone, Eur. Pat. Appl., 141 662, 1985. 6 W. M. Meier and D. H. Olson, Zeolites, 1992, 12,20. 7 J. J. Pluth, J. W. Smith and J. W. Richardson Jr., J. Phys. Chem., 1988,92,2734. 8 B. Kraushaar-Czametzki, W. G. M. Hoogervorst, R. R. Andrea, C. A. Emels and W. H. J. Stork, J. Chem. SOC., Fatuday Trans., 1991,87, 891. 9 G. Brouet, X. Chen and L. Kevan, J. Phys. Chem.,1991, 95, 4928. 10 N. Azuma, Ch. W.Lee and L. Kevan, Prep. Am. Chem. SOC.,Div. Petr. Chem., 1993,38, 538. 2146 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 11 12 13 14 15 G.Brouet, X. Chen and L. Kevan, Proc. 9th ZZC, Montreal, 1992, ed. R. von Ballmoos, J. B. Higgins and M. M. J. Treacy, Butterworth-Heinemann, Boston, 1993, p. 489. M. P. J. Peeters, J. H. C. van Hooff, R. A. Sheldon, V. I. Zholo-benko, L. M. Kustov and V. B. Kazansky, in ref. 11, p. 651. S. M. Yang and S.Y. Liu, in ref. 11, p. 623. G. Finger, J. Richter-Mendau, M. Bulow and J. Kornatowski, Zeolites, 1991, 11,443. G. Finger, J. Kornatowski, J. Richter-Mendau, K. Jancke, M. Bulow and M. Rozwadowski, in Catalysis and Adsorption by Zeolites, ed. G. Ohlmann, H. Heifer and R. Fricke, Stud. Surf. 17 18 19 20 21 22 W. H. Baur, W. Joswig, D. Kassner, J. Kornatowski and G. Finger, Acta Crystallogr., 1994, in the press. A. Ojo and L. McCusker, Zeolites, 1991,11,460. C. S. Blackwell and R. L. Patton, J. Phys. Chem., 1988,92,3965. V. R. Choudhary and S. R. Sansare, J. Th.Anal., 1987, 32, 777. Ch. Minchev, V. Minkov, V. Penvchev, H. Weyda and H. Lechert, J. Them. Anal., 1991,37,171. G. Finger, J. Kornatowski, K. Jancke, R. Matschat, J. Richter-Mendau and W. H. Baur, to be published. 16 Sci. Catal. 65, Elsevier, Amsterdam, 1991, p. 501. B. Zibrowius, G. Finger, J. Kornatowski and E. Lomer, to be published. Paper 3/070G; Received 31st December, 1993