Active Centres on NaH-Y Zeolite in But4 -ene Transformations BY JAN GALUSZKA," ANDRZEJ BARA~SKI AND STANISLAW C~CKIEWICZ 30-060 Cracow, Poland Institute of Chemistry, Jagellonian University, Krupnicza 41, Received 17th November, 1976 The kinetics of but-1-ene transformations on zeolites NaH-Y with varying degrees of Na+ ion exchange (0-82 %) have been investigated at constant pressure and also by temperature-programmed desorption combined with gas liquid chromatography. A correlation between the degree of cation exchange and localization of the Bronsted acid centres was noted and an assignment of the active centres involved in the but-1-ene transformations is proposed. Hydrogen sodium zeolites obtained from Na-Y parent specimens by ion exchange are useful for various catalytic processes, such as transformations of but-1-ene as studied by the temperature programmed desorption TPD meth0d.l The experimental data from that study were used as a basis for the following scheme of consecutive transformations of but-1-ene on NaH-Y zeolite.but-1-ene(gas) -+ adsorption -+ isomerization -+ polymerization dismutation cracking " further reactions ". (1) - - called for brevity 4 \ dehydrogenation The aim of the present paper is mainly to identify the active sites involved in butene transformations. The distribution of potentially active Na+ and H+ ions among the possible accessible positions in the crystal lattice of zeolite NaH-Y is still an open question despite many studies.2 However, a comparison of reaction rates for catalysts con- taining different amounts of different accessible sites, combined with a temperature programmed desorption study, should throw some light on the location of the active centres in the zeolite crystal lattice.EXPERIMENTAL MATERIALS Na-Y zeolite of (Si02/A1203 molar ratio 5.22) was kindly supplied by the Institute of Industrial Chemistry, Warsaw. Samples of NH4Na-Y zeolite were obtained by Na+/NHz ion exchange followed by pelletizing without binder. For the various samples prepared, degrees of exchange 0 ; 9.5 ; 12 ; 23 ; 43 ; 59 ; 68 ; 77 and 82 % were determined by flame photometry. The pellets of NH4Na-Y were stored over saturated aqueous NaCl at room temperature. Water content in the zeolite estimated derivatographically amounts to - 21 %. But-1-ene of purity 99.8 mol % supplied by Fluka AG contained isobutane, isobutene and butadiene as commercially listed impurities.Commercial cylinder nitrogen was used as carrier gas after deoxidation over BTS catalyst (B.A.S.F., FRG) and passage through silica gel and liquid nitrogen-cooled traps. 146J. GALUSZKA, A. B A R A ~ S K I AND s. C~CKIEWICZ 147 APPARATUS The TPD apparatus and the manostat system were as described previously. PROCEDURE In order to avoid possible deactivation fresh zeolite pellets (5 mm in diameter, 3.5-4 mm thick, 0.1 g average weight) was used in each experiment. NH4Na-Y zeolite was decomposed in situ in a TPD reactor at 430°C for 4 h in vacuo ( p < 1.3 x N m-2 = Torr). NaH-Y zeolite samples, obtained after calcination, were kept in a reactor in vacuo for 24 h and then again outgassed for 1 h at 430°C.The reactor was then cooled down to room temperature, when sorption of but-1-ene from gas phase was carried out at constant pressure 693-746f 7 N m2 (5.2-5.6kO.05 Torr). The kinetic curves of butene consumption were automatically recorded. After the kinetic run, but-1-ene remaining in the gas phase was removed from the reactor by nitrogen flushing for 15 min. It was proved additionally that the removal of gaseous but-1-ene by outgassing for 1 h using a diffusion pump introduces no change in the TPD spectra. Accordingly, the simpler flushing procedure only was used. ,4fter removing the gas phase, temperature programmed desorption was followed up to 430°C using a heating rate of 25f 1 deg min-'. The stream of carrier gas containing the desorbed species was monitored by a Gow Mac thermal conductivity cell, after which condensable products were frozen out in a liquid nitro- gen-cooled trap.These products, again after evaporation, were transferred by a syringe into a gas chromatograph. The packing of the chromatograph column consisted of chromo- sorb W (80-100 mesh size) covered with 15 % of 2,4-dimethylsulpholane. RESULTS THE KINETIC CURVES Fig. 1 shows the kinetic curves obtained for the system NaH-Y zeolite + but-1-ene using preparations with various degrees of exchange of Na+ ions for protons. Fig. 1 6.5 t 36 of cc!ions exchanged min degree of ion exchange as shown in the figure. FIG. 1.--Kinetic curves for the system NaH-Y zeolite+ but-1-ene. Zeolites amples of varied148 ACTIVE CENTRES ON ZEOLITES 1 I 1.0; I 0 .5 1 FIG. 2.-Relation between % of cations exchanged the amount of but-l-ene consumed by the samples after (a) 5 and (6) 100 min and the degree of ion exchange. contains many intercrossing curves, making discussion difficult, so we direct attention to the simpler plots in fig. 2, derived from fig 1, which show the dependence of the amount of but-1-ene consumed upon the degree of ion exchange for two arbitrarily chosen times : 5 and 100 min. The shape of the plots is thelsame for any time between t 120 I temperature/"C FIG. 3 . T P D evolution spectra as they depend upon the exchangelextent1of zeolite NaH-Y samples.J. GALUSZKA, A. B A R A ~ S K I AND s. C ~ C K I E W I C Z 149 2 and 100min. The results at times < 2min are uncertain, for experimental reasons.From a comparison of fig. 1 and fig. 2 it can then be seen that the depend- ence presented in fig. 2 is valid for conditions close to equilibrium. It follows that diffusion phenomena do not affect the shape of the plots in fig. 2. We conclude that the shape of these plots illustrates an intrinsic property of the investigated system. The sequence of kinetic curves and the results presented in fig. 2 show the existence of three characteristic intervals related to the degree of ion exchange. From the SiOz/Al,03 molar ratio it was deduced that there are 53.2 sodium cations per unit cell. So interval I from 0 to 12 % is equivalent to 6.4 Na+ cations per unit cell : interval I1 from 12 to 68 % is equivalent to 29.8 cations per unit cell and interval 111 above 68 % is equivalent to 17.0 cations, assuming 100 % of exchange as the upper limit. TPD SPECTRA The TPD spectra are presented in fig.3, where a pronounced dependence of the size and shape of particular peaks upon the extent of exchange is easily seen. Trends in the temperature for the maximum of peak I can also be seen from this figure. The change of the trend is particularly marked at the end of interval I. The qualitative composition of the desorbate presented previously was confirmed by chromatographic analyses of desorption products. Especially important is the fact that in the temperature range of peak I only but-l-ene was found for unexchanged zeolite, whereas from the all other samples three butene isomers, and only these species, are desorbed in the same temperature range.DISCUSSION The sequence of consecutive reactions, given in the introduction, is taken as the starting point for the following discussion. It was inferred from chromatographic analyses of desorbed products after TPD experiments that, on unexchanged zeolite, only adsorption of but-l-ene takes place ; on 9.5 and 12 % exchanged samples isomerization also proceeds, whereas on samples exchanged to a higher extent further butene reactions additionally occur. These analyses agree with data from our previous paper.l It follows that at least three different types of active sites should be considered : adsorption sites (A), isomerization sites (B) and sites responsible for further butene reactions (C). Unexchanged zeolite contains only A sites; 9.5 and 12 % exchanged samples contain A and B sites; samples exchanged to greater extent contain A, B and C sites.The maximum at 68 % (fig. 2) may suggest the existence of a fourth type of active site. W'e consider first the adsorption step of the sequence (1). But-l-ene is adsorbed at the highest rate on unexchanged zeolite (fig. 1 and 2). When the degree of exchange increases, the amount of desorbed but-l-ene decreases as evidenced by chromato- graphic analyses. In addition, peak I, and hence the total amount of butene isomers, also decreases (fig. 3). Since Bronsted centres are absent from unexchanged zeolite, their amount increasing with the degree of exchange, it seems unlikely that they are responsible for the adsorption of but-l-ene.For the same reason, those Lewis centres which are produced from Bronsted centres via dehydroxylation can be ex- cluded, especially as the samples were not heated above 430°C.6* ' Possible species obtained after the adsorption step of the sequence (1) which must be considered are physically adsorbed but- l-ene and chemisorbed molecules on Na+-sites. Arguments against physical sorption are : (i) i.r. spectra of but-l-ene adsorbed on unexchanged zeolite are similar when measured at room temperature150 ACTIVE CENTRES ON ZEOLITES and 120°C.5 (ii) TPD spectra do not change if, instead of nitrogen flushing, butene initially sorbed is outgassed by diffusion pump (see Experimental section). From the data given in fig. 1 and the composition of the sample it is easy to calcu- late that there are about 33 molecules of but-1-ene adsorbed for each unit cell of zeolite, i.e., 4 for each super cage or 0.6 for each sodium ion.It is plausible that adsorbed butene molecules interact first of all with easily accessible Na+ cations located in the super cages. The literature indicates that a majority of exchangeable Na+ ions are located inside the super cages and that their number amounts to 30-40 cations per unit ce11.*-12 These numbers are in adequate agreement with the number 33 found in our present case. The importance of exchangeable cations in the surface chemistry of zeolites has already been emphasised.5. 3* l4 We, therefore, advanced the hypothesis that Na+ cations are active sites during adsorption of but-l-ene. We consider now the isomerization step of the sequence (1).Isomerization of but-1-ene proceeds on the exchanged samples; up to 12 % exchange only isomer- ization, additional to adsorption, of but-1-ene takes place. Exchanged samples differ from unexchanged ones in containing Bronsted sites. It is therefore reasonable to assume that B and C are Bronsted centres. In order to distinguish between the two kinds of Bronsted sites we focus attention on the shape of the plots given in fig. 2, where a sudden change of the catalytic properties of zeolite is seen to occur at the end of interval I, above 12 % exchange, equivalent to 6.4Naf per unit cell. It seems that the initially introduced H+ ions are responsible for isomerization of butenes (B sites). We consider now the further catalytic reactions of butene which proceed on zeolites exchanged to an extent > 12 %.The amount of but-1-ene consumed increases almost proportionally with the degree of exchange in interval I1 from 12 to 68 %, after which it falls (fig. 2). Sodium cations exchanged within interval I1 amount to 29.8 per unit cell. This number is in good agreement with 32 accessible sites at position SII suggested originally by Breck,’ confirmed experimentally by Eulenberger et al. O and since accepted by many authors.12* 13’ 15-18 The intensity of the 3650 cm-l i.r. band of OH groups in zeolites increases only up to - 70 % exchange, as can be seen from experimental plots given in ref. (14) and (19). It follows that for higher degrees of exchange the groups do not increase in number. Since these Bronsted centres are easily accessible 2o and are catalytically active in various reactions,5* 12* 21 it is assumed that they are identical with centres C.Olson and Dempsey 22 reported a single-crystal X-ray diffraction study of hydrogen faujasite from which they conclude that the O1 oxygen atoms (the bridging oxygen atoms of the hexagonal prism) are the most favourable sites for formation of hydroxyl groups having a 3650 cm-l i.r. band, Other authors 17* 18* 2 3 agreed with this suggestion. Taking into consideration the fact that in each hexagonal prism there are two O1 oxygen atoms which could be associated with 24 there will be 32 sites per unit cell for hydroxyl groups here discussed. This number agrees well with our results.The 17.0 sodium cations not yet discussed belong to the interval above 68 % exchange. During the process of exchange of the last 20-25 Na+ cations by protons 12* 1 4 9 25 the i.r. band at 3550 cm-l appears. It is postulated that the protons responsible for this band lie inside hexagunal prisms and, therefore, are not easily accessible.12* 21* 22 Uytterhoeven et aZ.17 point out that the classification of ion-exchange sites given by Breck and X-ray data published by Eulenberger et aZ.1° agree on “ the location of some cations up to a maximum of 16 per unit cell in the hexagonal prisms ”. These sites will be discussed in more detail later.J. GALUSZKA, A. B A R A ~ S K I AND s. C ~ C K I E W I C Z 151 This point of view accords with other experimental data.g* 26 Thus a number of authors agree that there are 16 or so cations per unit cell having essentially different properties in that they are not easily acce~sible.~~ 11* l6 This inaccessibility and the decrease of the number of Na+ cations postulated as active sites for adsorption may explain the decrease of the relevant parts of the plots in fig. 2 especially since the number of centres C responsible for further catalytic processes attains maximum value at the end of interval 11, so their density is constant within interval 111.Another possible explanation of this trend, namely the poisoning of the sample by products of the further butene reactions, was excluded experimentally. It was found that repetition of a typical experiment on NaH-Y samples which have been used once (they turn grey) yields the same kinetic curve and the same TPD spectrum.The present discussion is based on the comparison of our kinetic data with in- formation available in the literature. It follows from the data that 53.2 cations in the unit cell can be divided into three groups, namely 6.4, 29.8 and 17.0 cations for the intervals I, I1 and I11 respectively. The assignment for groups of 29.8 and 17.0 cations from intervals I1 and I11 has already been discussed. We now return to the 6.4 cations from interval I which are related in some way to sites B. These cations are easy to exchange. They are close in number to the 8 occupied positions in the SIII positions as suggested by Breck.* Perhaps B sites have this origin. We turn now to the TPD spectra presented in fig.3. Special attention will be given to peak I representing but-1-ene for NaY and all three butenes for NaAH-Y samples. It is seen from fig. 3 that peak I decreases with increasing exchange. As a result of the decreasing amount of desorbed butene isomers the temperature of the maximum of peak I increases from 120-136°C as the exchange increases from 0 to 12 % (within interval I), which agrees well with the commonly observed shift in the TPD spectra of a heterogeneous s ~ r f a c e . ~ However, from 12-68 % exchange (interval 11) decreasing amounts of desorbed species cause a decrease in the tem- perature of maximum of peak I from 136to 77°C. Such a trend (opposite to that observed in interval I) may be explained by the assumption that the more strongly bonded of the adsorbed butene isomers are consumed by further catalytic reactions of sequence (1).It is seen from fig. 3 that the decrease of peak I is further enhanced at higher exchange. Near 80 % exchange the peak I is almost invisible. It seems that for a sufficiently high exchange it will vanish completely. If this is true, no adsorbed but-1-ene on Na+ centres will be present on the sample. It follows that the first step of sequence (1) will not occur. As a result, zeolite should be inactive for butene-1 transformations provided that the consecutive sequence is valid for these highly exchanged samples. All the main conclusions concerning the active centres are summarized briefly by the following enlargement of sequence (1) already given in the introduction.Nas+111? J OiH Na+ H + but-1 -ene(gas) --4 but- 1 -ene(ads) --+ isomerization -----+ polymerization 0iH dismutation ----+ further reactions dehydrogenation. We are grateful to the referees for helpful comments and for assistance with the phrasing of some sentences.152 ACTIVE CENTRES ON ZEOLITES A. BaraAski, S. Cgckiewicz and J. Galuszka, Bull. Acad. polon. Sci., S6r. Sci. chim., 1976, 24, 645. D. J. V. Smith, Adv. Chem. Series, 1970,101, 171. R. J. CvetanoviC and Y. Amenomiya, Adv. Catalysis, 1967, 17, 103. S. Cgckiewicz and J. Galuszka, Inz. Chem., 1975, 5, 643. A. Bieladski, J. Datka, A. Drelinkiewicz and A, Mdecka, Bull. Acad. polon. Sci., Sir. Sci. chim., 1976, 24, 137. J. Datka, Bull. Acad. polon. Sci., Sir. Sci. chim., 1974, 22, 975. ' P. D. Hopkins, J. Catalysis, 1968, 12, 325. D. W. Breck, J. Chem. Ed., 1964,41,678. R. Beaumont, D. Barthomeuf and Y. Trambouze, Adv. Chem. Series, 1971, 102,327. lo G. R. Eulenberger, D. P. Shoemaker and J. G. Keil, J. Phys. Chem., 1967,71,1812. l1 H. S. Sherry, J. Phys. Chem., 1966,70,1258. l2 J. W. Ward and R. C. Hansford, J. Catalysis, 1969, 13, 364. l3 J. A. Rabo and M. L. Poutsma, Adv. Chem. Series, 1971, 102, 284. l4 V. Bosa&ek, V. Patzelowa, C. Hybl and Z. Tvaruikova, J. Catalysis, 1975, 36, 371. l5 J. W. Ward, J. Colloid Interface Sci., 1968, 28, 269. l6 K. Tsutsumi and H. Takahashi, J. Phys. Chem., 1970,74,2710. l7 J. B. Uytterhoeven, P. .lacobs, K. Makay and R. Schoonheydt, J. Phys. Chenz., 1968,72, 1768. l 8 T. R. Hughes and H. M. White, J. Phys. Chem., 1967,71,2192. l9 A. Bieladski and J. Datka, Bull. Acad. polon. Sci., Sbr. Sci. chim., 1974, 22, 341. 2o L. Moscou, Adv. Chem. Series, 1971, 102, 337. 21 P. E. Eberly, Jr., J. Phys. Chem., 1967, 71, 1717. 22 D. H. Olson and E. Dempsey, J. Catalysis, 1969, 13,221. 23 R. A. Schoonheydt and J. B. Uytterhoeven, J. Catalysis, 1970, 19, 55. 24 E. Dempsey, J. Catalysis, 1975, 39, 155. 25 J. W. Ward, J. Phys. Chem., 1969,73,2086. 26 P. Gallezot and B. Imelik, J. Chim. phys., 1977, 68, 816. . (PAPER 6/2121)