Shape selectivity in Y-zeolites

Shape selectivity in Y-zeolites

Applied Catalysis, 58 (1990) 105-117 105 Elsevier Science Publishers B.V., Amsterdam-- Printed in The Netherlands Shape Selectivity in Y-Zeolites C...

702KB Sizes 0 Downloads 12 Views

Applied Catalysis, 58 (1990) 105-117

105

Elsevier Science Publishers B.V., Amsterdam-- Printed in The Netherlands

Shape Selectivity in Y-Zeolites Catalytic Cracking of Decalin-Isomers in Fixed Bed Micro Reactors H.B. MOSTAD*, T.U. RIIS and O.H. ELLESTAD Department of Hydrocarbon Process Chemistry, Center for Industrial Research, Box 124, Blindern, 0314 Oslo 3 (Norway)

(Received 4 July 1989, revised manuscript received 12 October 1989)

ABSTRACT

The shape selectivity in zeolites for the catalytic cracking of the structural isomeric di-naphthenes, cis- and trans-decalin, has been studied. Catalytic cracking of the pure cis- and transdecalin isomers and a decalin-mixtureover ZSM-5, offretite, mordenite, two Y-zeolites and an amorphous silica-aluminacatalyst for reference,has been performedin a fixedbed micro reactor. The conversion increasedwith larger pore size. Only the Y-zeolites showedselective cracking of the cis-decalin isomer. The ratio of the bfalculatedrate constants for catalytic crackingof cis- and trans-decalin in LZ-Y82exceeded i05,~hile the same ratio was about 4 for the amorphoussilicaalumina catalyst. This may be e~ala~ed7by the slightly smaller moleculardimensionsand more flexible structure of cis-decalin comparedto trans-decalin.

INTRODUCTION Catalytic cracking of alkane hydrocarbons in the presence of zeolite catalysts has been extensively studied. Much less work has been reported on the conversion of naphthenes, in spite of the fact t h a t this group of hydrocarbons represents another major class of petroleum constituents in the feedstock used for catalytic cracking. Potential feedstocks to catalytic crackers consist frequently of more t h a n 25% naphthenes. In addition there are considerable amounts of naphthene-aromatics. Weitkamp and coworkers have investigated hydrocracking of several mononaphthenic compounds over bifunctional catalysts [1-3]. T h e y have recommended hydrocracking of a Clo-naphthene, preferentially butylcyclohexane or pentylcyclopentane, as a potential test reaction for estimating the effective pore width of zeolites with u n k n o w n structures and for calculating the spaciousness index (SI) [4 ]. Model alkanes with at least eight carbon atoms should

106

be employed to allow for the most favorable ionic pathways of hydrocracking in large pore zeolites [5]. Ernst and Weitkamp have shown that this critical carbon number is even higher for naphthenes and for mechanistic studies they recommend model naphthenes with at least ten carbon atoms [3]. Jacobs and coworkers have investigated the shape selectivity effects in bifunctional zeolites from studies of product distributions from a Clo mono-naphthene, cyclodecane [6 ]. It was reported that the c/s-decalin isomer was formed in significantly more abundant amounts than the thermodynamic more stable trans-decalin [6,7 ]. There are only a few publications where decalin has been considered as a model compound for catalytic cracking. In 1970 Nace compared the cracking of various naphthenes over silica-alumina and a REHX-zeolite [8]. For silicaalumina, the cracking rate increased progressively with molecular size of the 1- to 4-ring naphthenes. By contrast, the cracking rate went through a maximum for the 2-ring naphthene (1,3,5 tri-methyl-decalin ) when the REHX catalyst was used. This could be explained only by the occurrence of diffusional limitations for the 3 and 4 fused ring reactants [8]. The product distribution from hydrotreated coal liquids has been studied, using among others decalin as a model compound for catalytic cracking [9]. The initial cracking reaction of this compound was found to be ring-opening, yielding alkylcyclo-hexenes. Hernandez et al. have investigated the reactions of decalin over mordenite, faujasite, offretite and ZSM-5 [10]. They observed an exclusion of decalin from H-ZSM-5 and dehydrogenation in mordenite and faujasite. Offretite was the only catalyst which gave appreciable cracking of decalin. Rabo observed occlusion of alkali ions into the sodalite cages of an Y-zeolite, explained by cleavage of the framework bonds by interference of the penetrating ions [ 11 ]. He proposed that in catalytic cracking reactions, carbenium ions may account for the allowance of larger molecules into the zeolite pores by a similar mechanism. The scope of this work has been to study the reactant shape selectivity of the structural isomeric di-naphthenes, c/s- and trans-decalin, in pure zeolite systems. These results are part of a detailed study of the product distribution, including olefins, from catalytic cracking of model compounds [12]. Preliminary experiments on catalytic cracking of decalin indicated a discrimination between cracking of c/s- and trans-decalin. A more detailed study was of interest since the calculated size of c/s- and trans-decalin is in the critical range compared to the pore dimensions in the zeolites, but with very different molecular conformational characteristics. EXPERIMENTAL

Five different pure zeolite systems and one silica-alumina support have been studied. H-mordenite, LZ-Y82- and LZ-Y62- extrudates were obtained from

107

Ventron GmbH, while H-offretite [ 13 ] and ZSM-5 [ 14 ] were synthesized. The amorphous silica-alumina support (86% SiO2, 13% A1203) was obtained from Davison. The catalysts were crushed and sieved to a particle size of 70-325 mesh. Y-82 is a low sodium grade ultrastabilized molecular sieve, while Y-62 is a HY-type catalyst. The Y-zeolites were hydrothermally treated in 100% steam (25 ml H20/h ) for 18 h at variable temperatures (Table 1 ). The silica-alumina support was calcined in air for 18 h at 550 ° C. The surfaces given in Table 1 were measured by nitrogen adsorption (BET). The reduction in surface area from the steam treatment is due to the partial destruction of zeolite crystallinity, which creates mesopores in the zeolite particles.

Hydrocarbons A mixture of cis/trans-decalin (decahydronaphthalene) and the pure isomers were used as model compounds for catalytic cracking. The cis/trans mixture was 60%: 40% (purity > 98%, EGA- Chemie). Pure c/s- and trans-decalin isomers were purchased from TCI-Tokyo Kasei (purity=99% and 98% respectively).

Reactors and detection systems Fixed bed pulse reactor The reactor was connected directly to a gas chromatograph-mass spectrometer. The mass spectrometric analysis of the reactor effluent from catalytic cracking of decalin gave individual identification of compounds with carbon numbers from C3 to C13. The details concerning the construction of the fixed bed pulse reactor and the quantification of the products from decalin have been described elsewhere [15]. The reactor temperature varied between 330 and 464°C and the catalyst quantity varied between 0.10 and 0.25 g. Blank runs were carried out at the appropriate reaction temperatures to determine the extent of thermal cracking. TABLE 1 Specific surface areas of the catalyst-systems Catalyst Y-82 Y-62A Y-62B Silica-alumina

Specific surface area

Steaming temperature

(m2g-~)

(°C)

384 301 126 343

730 500 670

108

Micro activity test (MAT) reactor With the standard MAT reactor, the different reactor parameters could be more precisely controlled [16]. The catalyst and feed quantity in this reactor were 4.0 g and 1.3 g respectively. Both the ASTM-standard temperature of 482 ° C and 420 ° C were used in the reactor. The purpose behind our work was among others to make a detailed identification of the product composition, including olefins, from catalytic cracking of model compounds. A continuous flow reactor with hydrogen could therefore not be used. The MAT reactor is the most accepted laboratory reactor for correlation of bench scale data in fluid catalytic cracking processes. A comparison of the product distribution from the pulse reactor and the MAT reactor revealed qualitative agreement [12]. A nalyses The product distribution in the liquid fraction (C4-C12) from the MAT reactor was determined by mass spectrometry. The compounds were identified by matching their mass spectra with a NBS spectrum library (National Bureau of Standards). The standard deviation was estimated to 1 area percent. The compounds were in addition identified in a PIONA-library [ 17 ] and classified as paraffins, iso-paraffins, olefins, naphthenes and aromatics. The gas chromatographs were a Finnigan 9610 and a Varian 3400. The mass-spectrometers were Finnigan MAT 4000 and 8200. The GC column was a cross-linked Methyl Silicone Gum, 50 m × 0 . 2 mm I.D., 0.5/zm film thickness (Hewlett Packard PONA-column). Injection of 0.8/zl model compound was done through a split injector and the carrier gas was helium. The temperature programme was - 2 5 ° C (2 rain.), 2°C/min-~70°C, 2.5°C/min-.150°C, 10°C/ min--250°C and 250°C (10 rain). The mass-spectrometric conditions were electron impact at 70 eV and 250 ° C. RESULTS AND DISCUSSION

A mixture of cis/trans-decalin (60:40) was cracked over four different zeolites in a fixed bed pulse reactor. The composition of unreacted feed from the Y-zeolite was found to be significantly different from the composition of unreacted feed from mordenite, offretite and ZSM-5 (Table 2). From thermodynamics, trans-decalin is the more stable isomer and the equilibrium composition of c/s- and trans-decalin at 460°C is 14% and 86%, respectively [18]. Equilibrium is not reached in fixed bed pulse reactors. However, because of the high reproducibility (standard deviation below 0.5% ) of product composition in these experiments, it was possible to make a relative comparison of the compositions in the reactor effluent from the four zeolites. Only the products from the Y-zeolite contained a relative amount of transdecalin above the equilibrium composition. The observed decalin composition in Table 2 was confirmed within + 3% by

109 TABLE 2 The composition of the decalin fraction after catalytic cracking of cis/trans-decalin (60:40) in a fixed bed pulse reactor at 460 °C Zeolite

Conversion

Composition a

(%) Y-82 H-mordenite H-offretite ZSM-5

90 60 33 21

% cis-decalin

% trans-decalin

3 54 54 54

97 46 46 46

aComposition of the decalin fraction (%) = 100% -conversion (%).

in the

reactor

effluent: the

decalin

fraction

cracking experiments over Y-82 and offretite in the MAT reactor [16] at different space velocities. Blank runs (absence of catalyst) showed that thermal cracking was negligible (below 0.5%) in the range of 390°C-490°C and the isomerization from c/s- to trans-decalin increased only slightly with increasing temperature. The stability of the catalyst activity was investigated by following the extent of decalin conversion with pulse number. ZSM-5 and the ultrastabilized Yzeolite were found to have the most stable activities, while mordenite showed fast deactivation. The zeolite activity order, based on conversion of the feed compound, demonstrates an increasing conversion of decalin with increasing pore size. This is in agreement with the activity order observed by Jacobs and Tielen in the conversion of cyclodecane over the same zeolites [6 ]. Note in Table 2 that only Y-82 has been deactivated by hydrothermal treatment. In spite of this, the conversion of decalin with Y-82 was substantially higher than with any of the other zeolites. The decalin composition in the reactor effluent from mordenite (6.7 X 7 ~,), offretite (6.4 A) and ZSM-5 (5.4X5.6 ~,) was the same and also remarkably constant at different conversions for the same catalyst. Compared to previous results from the catalytic cracking of cyclodecane in a flow reactor over the same zeolites [6,7], our results were somewhat different. Jacobs and co-workers observed that the thermodynamic less stable c/s-decalin isomer was most abundantly formed and the relative amount of c/s-decalin increased with decreasing pore size of the zeolites investigated [6,7]. Thus, c/s-decalin was able to escape also from small pore zeolites. Furthermore, Hernandez et al. [10] observed conversion of decalin in the offretite catalyst (2.8-9.0 wt.-% ), when this compound was hydrocracked in a continuous flow reactor at 450-550 °C, i.e. the decalin molecule was able to pass through the pore openings in the offretite catalyst. Decalin was, however, excluded from H-ZSM-5 [10].

110

Catalytic cracking of the pure c/s-decalin isomer over the amorphous silicaalumina catalyst and Y-82/Y-62A is compared in Table 3. The experiments were carried out both in the fixed bed pulse reactor and in the MAT reactor. The Y-62A catalyst showed slightly less cis-decalin selectivity compared to Y-82. Measurements of the catalyst surfaces showed lower values for the former (Table 1), most likely due to the formation of mesopores caused by the hydrothermal treatment. Although influencing the values of the rate constants, the qualitative picture of the shape selectivity remains. This was further confirmed from the data of the cis/trans ratio of unreacted decalin from the amorphous silica-alumina and the Y-zeolites (Table 3). Differences in the product composition from cracking of cis-decalin over Y82 and the amorphous silica-alumina catalyst were revealed by PIONA-analysis (paraffins, iso-paraffins, olefins, naphthenes, aromatics) of the gasoline fractions (C4-C12). More primary cracking products, i.e. more olefins and less aromatics were produced from the Si-A1 catalyst [19]. The observed differences in product composition from the two catalyst systems indicate a conversion of c/s-decalin on the internal surface of the Y-zeolite. Cracking reactions on the external surface could be expected to give a product composition more similar to the one from the Si-AI catalyst. Table 4 shows the average product compositions of residual decalin from several experiments. The data demonstrate a higher conversion of pure c/sdecalin in both reactors, suggesting a larger cracking reaction rate for this isomer. The results also indicate a fast isomerization from c/s- to trans-decalin, TABLE 3 The composition of cis/trans-decalin after catalytic cracking of pure cis-decalin ( > 99% ) Reactor

Temperature (°C)

Catalyst

Conversion (%)a

Compositionb

cis (%)

trans (%)

MAT

482 482 482 420

Si-A1 Y-62A Y-82 Y-82

71 99 96 93

90 5 3 4

10 95 97 96

Pulse

464 330c 330c

Si-A1 Y-62A Y-82

95 89 98

87 10 0

13 90 100

~Conversion does not include isomerization from cis-decalin to trans-decalin and standard deviation (S.D.) was estimated from parallel experiments to 1% conversion. bComposition of the decalin fraction in the reactor effluent: the decalin fraction (%) = 100% -conversion (%) CTo diminish conversion with the Y-zeolites, the fixed bed pulse reactor temperature had to be reduced from 464°C to 330°C.

111 TABLE 4 The composition of cis/trans-decalin after catalytic cracking of different decalin isomers over the Y-82 catalyst. Reactor

Temperature ( ° C )

Feed

Conversion (%)

Composition cis ( % )

trans ( % )

MAT

482 482 482

cis cis/trans a trans

96 86 82

3 1 1

97 99 99

Pulse

392 392 392

cis cis/trans a trans

99 97 89

0 0 0

100 100 100

Pulse

460

cis/trans ~

84 b

6

94

% i s / t r a n s = 60 : 40.

bDeactivated catalyst through repeated injections of the reactant.

while the opposite reaction proceeded slowly: when pure cis-decalin was used as feed, high yields of the trans isomer (99-100%) were found in the unreacted decalin, whereas feeding trans-decalin, resulted in only minor amounts (01% ) of the cis isomer in the decalin fraction. Equilibrium composition of decalin in the temperature range investigated (392-482°C) contains 85-88% trans-decalin [18]. The amount of trans-decalin in the reactor effluent was above the equilibrium values in all the experiments listed in Table 4. Thus, it seems to be a selectivity for cis-decalin in the Y-zeolite. Fig. 1 shows the yields of unreacted decalin isomers in the reactor effluent in the conversion area of 60-100%. The pore discrimination of trans-decalin may explain the different shapes of the cis- and trans-decalin curves and consequently the selective conversion of c/s-decalin. Most of the starting 60% of the c/s-decalin molecules in the feed mixture were converted before the reaction of the last 27% of the trans-decalin molecules. The reaction rates were further investigated by construction of a simplified three lump model as illustrated in Fig. 2. The rate constants for the isomerization of cis- and trans-decalin are kl and k_ 1, while the rate constants for the conversion reactions of cis-decalin and trans-decalin are summarized in k3 and k2 (Fig. 2). The composition of c/s- and trans-decalin in the reactor effluent will depend on the composition of cis- and trans-decalin in the feed, the equilibrium constant for the isomerization reaction between the two isomers; K= kl/ k 1, as well as the two reaction rate constants k2 and k3. The cracking reaction rate for cis-decalin will depend on the concentration of c/s-decalin in the feed and the reaction rates for all the reactions taking place in the catalyst bed.

112

3O

20 ~~rans-Decalin

0|

m

,

,

60

, ,

70

~r-----~-.~.-~ , - -

80

Conversion(%)

_

_

,

90

100

Fig. 1. Yields of unreacted cis- and trans-decalin in the reactor effluent from catalytic cracking of cis/trans-decalin (60:40) over Y-62B (m) and Y-82 ([2) at different conversion levels.

cis-DECALIN ]• k~ k.~ _1 -Itrans'DECALIN k3 [ PRODUCTS 1. k2 Fig. 2. Three lump model for catalytic cracking of decalin, products ( % ) = conversion (%).

Besides cracking, these reaction include dehydrogenation, alkylation and isomerization to other Clo-products. The products from these conversion reactions contribute to the term "products" in Fig. 2. From the temperature dependency of the equilibrium constant (K) for the isomerization reaction of the decalin isomers [ 18,20 ], it can be calculated that the rate constants for the isomerization from c/s- to trans-decalin (kl) will be 5.7 to 10.0 times larger than the reverse reaction (k_ 1) with the applied reactor temperatures between 330 and 482 ° C. A computer program, STELLA [21 ] has been used to simulate the model in Fig. 2, assuming all reactions to be of first order and residence time to be 0.1 s. Using this program, it is possible to determine the set of relative rate constants which best fits the experimental data for the different catalyst systems. The rate constant kl was chosen and k_ calculated from the equilibrium constant (K). The rate constants for the conversion to other products, k2 and k3, were adjusted together with k~ to fit the experimental values of the conversion and composition of c/s- and trans-decalin in the reactor effluent. The best estimate for a set of the relative rate constants for the Y-zeolite and the amorphous silica-alumina catalyst is given in Table 5.

113 TABLE5 Relativerateconstants~rthecatal~iccrackingofthecis-andtrans-decalinisomers

Catalyst

kl

k_l a

k2

k3

Reaction

cis--, trans

trans--,cis

trans-,prod,

cis--*prod.

Y-zeolite Amorphous Si-A1

100 1

18 0.2

0.01 4

700 15

ak , = k l / K TABLE 6 Catalytic cracking of decalin-isomers over Y-82 in the MAT reactor at 482 °C Feed

Conversion

Composition

Values

(%) cis (%)

trans (%)

cis (99%) cis (99%)

97 96

2 3

98 97

Calculated Experimental

cis/trans = 60: 40 cis / trans = 60: 40

90 86

2 1

98 99

Calculated Experimental

trans (98%) trans (98%)

79 82

2 1

98 99

Calculated Experimental

The Y-zeolite and the amorphous silica-alumina catalyst showed different c/s-decalin selectivity. This is clearly illustrated by the two sets of rate constants in Table 5. The ratio of the rate constants for catalytic cracking of c/sdecalin and trans-decalin (k3/k2) in the Y-zeolite bed exceeded l0 b, while the same ratio was about 4 for the amorphous silica-alumina. This is as expected from a catalyst system with no cis-/trans-decalin shape selectivity. The rate constants for the cis/trans-decalin isomerization reaction over the Y-zeolite were about 100 times larger than the rate constants for the same reaction over the silica-alumina catalyst, demonstrating the larger activity of the former catalyst system. The relative rate constants in Table 5 have been used to calculate the conversion and the composition of c/s- and trans-decalin in the reactor effluent within the selected time-interval (t=0.1 s). The calculated and the experimental values compared in Table 6 and 7 for the Y-zeolite and the amorphous silica-alumina respectively, are in acceptable agreement. The simultaneous agreement in the conversion and cis/trans-decalin composition was obtained by minimizing the sum of quadratic differences between calculated and experimental values.

114 TABLE 7 Catalytic cracking of decalin-isomers over amorphous silica-alumina in the MAT - reactor at 482°C. Feed

Conversion

(%)

Composition

Values

cis (%)

trans (%)

cis (99%) cis (99%)

76 71

82 89

18 11

Calculated Experimental

c i s / t r a n s = 60: 40 cis/trans = 60: 40

59 57

30 25

70 75

Calculated Experimental

trans (98%) trans (98%)

34 32

2 1

98 99

Calculated Experimental

The calculation of the molecular dimensions of the two model compounds in Fig. 3 was based on relevant molecular fragments included in crystal structures from X-ray data [22,23]. Van der Waals radii were taken to be 1.8 ,~ and 1.4 A for C and H, respectively. When the two rings, in decalin are joined through two equatorial-type bonds, trans-decalin results, whereas an axial-equatorial union gives c/s-decalin (Fig. 3 ). The trans isomer is more stable than the c/s isomer, largely because of the relatively unfavourable nonbonded interactions within the concave area of c/sdecalin [19]. The two compounds cannot interconvert unless C-C or C-H bonds are broken. Trans-decalin is a relatively rigid molecule and, unlike cyclohexane, the two rings cannot flip from one chair to another. The conformation of c/s-decalin is relatively flexible, and a simultaneous inversion of both rings occurs fairly easily [20]. The Y-zeolites has a pore size of 7.4 it and the diameter of the almost spherical supercage is about 12 it [24]. The selective conversion of c/s-decalin can in principle be explained by the smaller molecular dimensions, the less rigid structure of this isomer or from a possible more bulky transition state of trans-decalin during the cracking reaction. It is however not probable that cracking of trans-decalin will be hindered by the internal space in the supercages of the Y-zeolite. Moreover, PIONAanalysis of the gasoline fractions revealed close agreement in the normalized product compositions from c/s- and trans-decalin thus suggesting similar reaction routes [ 12 ]. Both c/s- and trans-decalin are in the same range as the critical size of molecules which can enter the Y-zeolite pores and gain access to the internal catalyst sites (Fig. 3). The most reasonable explanation for the apparent easier access of c/s-decalin to the internal catalyst surface, is the smaller dimensions

115

A

B

Fig. 3. Structure of the decalin isomers: (A) cis-decalin, 5.6×6.7X8.8 ~.; (B) trans-decalin: 5.2 × 7.6 X 10.0 A. The molecular axis of inertia are indicated.

of this isomer and even more important: its flexible conformation compared to trans-decalin. There is a significant difference in conformational mobility for the cis-and trans-decalin isomers. Chair-chair conformation of both isomers are energetic most favourable. Trans-decalin can not undergo any inversion process, since one ring can not be diaxially linked to the other, and the conformation is rigidly defined. The conversion of each ring from chair to boat is, however, not impeded. It is the further conversion from the boat to an alternative chair that is blocked. In contrast, c/s-decalin may undergo a coordinated inversion in both rings via very flexible intermediate conformations with both rings in the boat form [25]. Consequently there is a rapid interconversion of the chair forms. The barrier to chair inversion is in fact much lower than in cyclohexane, possibly owing to the three gauche-butane interactions in cisdecalin, which are not present in trans-decalin [26].

116

Our hypothesis on c/s-decalin selectivity is further supported by the observations of Jacobs and co-workers during catalytic cracking of cyclodecane [6,7 ]. They concluded that the thermodynamic less stable c/s-decalin isomer was the most abundantly formed isomer from cyclodecane over the Y-zeolite. This deviation in cis/trans-decalin composition from the thermodynamic equilibrium may be explained by product shape selectivity. The easier escape of c/s-decalin from the internal zeolite surface compared to trans-decalin is in agreement with our observations of selective c/s-decalin conversion because of the slightly smaller molecular dimensions and conformational characteristics of this isomer. A possible effect of the hydrocarbon reactants on the zeolite pore size during the catalytic cracking process has been considered by Rabo [ 11 ]. The carbenium ions, which are presumably strongly attached to the intracrystalline zeolite surface, may play a role of unknown extent in affecting the framework bond cleavage. Accordingly, Rabo suggested that the pore size of the Y-zeolite, as defined by low-temperature adsorption experiments, is probable only the lower limit of the effective pore size of this molecular sieve in the catalytic cracking process. If the carbenium ion intermediates from c/s- and trans-decalin were able to affect bond cleavage and thereby related temporary pore size enlargement, a discrimination between c/s- and trans-decalin would not be likely to occur because the differences in molecular dimensions are small (Fig. 3 ). The observed selective cracking of c/s-decalin in the Y-zeolites consequently indicates that temporary ring opening does not occur in our experiments, and the suggested influence by carbenium ions seems to be negligible. CONCLUSION

Shape selectivity of the c/s- and trans-decalin isomers in the Y-zeolite have been observed. The pore discrimination of trans-decalin and consequently the selective conversion of c/s-decalin can be explained by the slightly smaller molecular dimensions and more flexible conformation of this isomer. ACKNOWLEDGEMENTS

The authors want to express their gratitude to I.M. Dahl for valuable suggestions and helpful discussions. They also want to thank ,~. Raknes and I. Balk for skillful technical assistance with the reactor experiments and K. Urdal for valuable advices in mass spectrometry. Financial support of this work by the Norwegian Council for Scientific and Industrial Research (NTNF) is gratefully acknowledged.

117 REFERENCES 1 J. Weitkamp, P.A. Jacobs and S. Ernst, in P.A. Jacobs, N.I. Jaeger, P. Jiru, V.B. Kazansky and G. Schulz-Ekloff (Editors), Structure and Reactivity of Modified Zeolites, Proc. Int. Conf., Prague, July 9-13, 1984, Vol. 18, Elsevier, Amsterdam, 1984, p. 279. 2 J. Weitkamp, S. Ernst and H.G. Karge, ErdS1 Kohle, 37 (1984) 457. 3 S. Ernst and J. Weitkamp, Acta Phys. Chim., 31 {1985) 457. 4 J. Weitkamp, S. Ernst and R. Kumar, Appl. Catal., 27 (1986) 207. 5 J. Weitkamp, Ind. Eng. Chem., Prod. Res. Dev., 21 (1982) 550. 6 P.A. Jacobs and M. Tielen in Proceedings of the 8th International Congress on Catalysis., Berlin, July 2-6, 1984, Vol. IV, Verlag Chemie, Weinheim, 1984, p. 357. 7 P.A. Jacobs and J.A. Martens, in Y. Murakami, A. Iijima and J.W. Ward (Editors), New Developments in Zeolite Science and Technology, Studies in Surface Science and Catalysis, Vol. 28, Elsevier and Kodansha, Amsterdam, Tokyo, 1986, p. 23. 7 D.M. Nace, Ind. Eng. Chem. Prod. Res. Dev., 2 (1970) 203. 9 S. Swaminathan, Diss. Abstr. Int. B, 41, (8), 3117, Univ. Microfilms Int., Order No 8103900, 1981. 10 F. Hernandez, L. Moudafi, F. Fajula and F. Figueras, in Proceedings of the 8th. International Congress on Catalysis, Berlin, July 2-6, 1984, Vol. II, Verlag Chemie, Weinheim, 1984,p. 447. 11 J.A. Rabo, in Proceedings of the 6th International Zeolite Conference, Reno, U.S.A., July 10-15, 1983, Butterworths, London, 1984, p. 41 12 H.B. Mostad, O.H. Ellestad and T.U. Riis, Appl. Catal., submitted for publication. 13 K.P. Lillerud and J.H. Raeder, Zeolites, 6 (1986) 474. 14 C.J. Plank, E.J. Rosinski and A.B. Schwarz, British Patent 1 402 981 (1973). 15 H.B. Mostad, O.H. Ellestad and T.U. Riis, in P. Longevialle (Editor), Advances in Mass Spectrometry, Proc. of the llth Int. Mass Spec. Conf., Bordeaux, Aug. 29-Sept. 2, 1988, Heyden, London, llB, 1989, p. 1664. 16 ASTM D 3907-80, Fluid Cracking Catalysts by Micro Activity Test. 17 K. Urdal and S. Sporstol, in P. Longevialle (Editor), Advances in Mass Spectrometry, proc. of the 1lth Int. Mass Spec. Conf., Bordeaux, Aug. 29-Sept. 2, 1988, Heyden, London, 11B, 1989, p. 1670. 18 D.R. Stull, E.F. Westrum Jr. and G.C. Sinke, (Editors), The Chemical Thermodynamics of Organic Compounds, Wiley, New York, 1969, p. 402. 19 H.B. Mostad, O.H. Ellestad and T.U. Riis, in preparation. 20 J.D. Roberts and M.C. Caserio, (Editors), Basic Principles of Organic Chemistry, Sc. Ed., Benjamin, Menlo Park, CA, 1977, p. 480 21 B. Richmond, STELLA/forthe Macintosh Computer, High-Performance Systems, Lyme, NH. 22 PG. Jones, O. Kennard, S. Chandrasekhar and A.J. kirby, Acta Cryst. B34 (1978) 2947. 23 A. Dubourg, R. Roques and E. Guy, Acta Cryst. B35 (1979) 2938. 24 D.W. Breck, Zeolite Molecular Sieves, Wiley, New York, 1969. 25 J. Dale, Stereochemistry and Conformational Analysis, Verlag Chemie, Weinheim, 1978, p. 217. 26 E.L. Eliel, N.L. Allinger, S.J. Angyal and G.A. Morrison, (Editors), Conformational Analysis, Wiley New York, 1965, p. 226.