H. Chon, S.-K. lhm and Y.S. Uh (Editors) Progress in Zeolite and Microporous Materials Studies in Surface Science and Catalysis, Vol. 105 1997 Elsevier Science B.V.
Effect of Ga on the h y d r o g e n transfer activity of zeolites w i t h the offretite s t r u c t u r e Pei-shing E. Dai, C. Mark Tsang, and Randal H. Petty Texaco R & D Department P. O. Box 1608 Port Arthur, Texas 77641, USA M. Somervell, B. Williamson, and M. L. Occelli Georgia Tech Research Institute Zeolites and Clays Research Program Atlanta, Georgia 30332-0827, USA Temperature-programmed NH3 desorption (TPD) results, together with propylene conversion experiments, have shown that introduction of Ga atoms into the aluminosilicate framework of H-offretites generates a second population of strong acid sites responsible for thermodesorption of NH3 in the 400 to 600~ range. These H,Ga,Al-offretites have a hydrogen transfer index (HTI) that is significantly higher than either of the parent crystals containing only A1 or Ga. In the absence of A1, the pure H,Ga-offretite HTI is reduced by almost 50%, indicating that these materials should be studied as possible fluid cracking catalyst (FCC) additives for olefins generation. 1.
Among the complex network of reactions involved in FCC, hydrogen transfer is critical to the distribution of paraffinic and olefinic products. Propylene conversion can be used as a probe reaction to test the hydrogen transfer activity over ZSM-5, an important additive in FCC preparation (1-9). In a companion paper (10), the hydrogen transfer activity of ZSM-5 during conversion of propylene to paraffins and olefins was investigated in detail. It was observed that as the silica-to-alumina ratios (SAR) of the zeolite increase, more products are preserved as olefins. Lower temperatures and lower space velocities were found to favor hydrogen transfer reactions, giving paraffins as the major products. In the present paper, we explore the hydrogen transfer activities of Gacontaining H-offretites, using propylene conversion as the probe reaction. The
982 objective of this work is to determine the effects of isomorphous substitution of A1 by Ga on hydrogen transfer activity. We report herein for the first time the synergistic effects that framework (and extraframework) A1 and Ga have on the acid strength and hydrogen transfer activity of H-offretites. 2.
Crystalline molecular sieves with the offretite structures have been synthesized at 95~ from hydrogels containing both gallia and alumina (11,12). After a calcination step to remove the template used during synthesis, the crystals have been NH4-exchanged and activated according to a procedure described in details elsewhere (12). Temperature-programmed desorption (TPD) of ammonia was used to characterize the acidity of these offretites. Propylene conversion reactions were carried out in a fixed-bed reactor. A primary grade blend of 5 mol % propylene and 95 mol % nitrogen from Linde was used as the feed. Propylene conversion activities of H-offretites were measured under standard conditions, specifically, catalyst load: 0.5 g; reaction temperature: 400~ WHSV: 0.19 g C3-/g catalyst (11). 3.
Results and Discussion
X-ray diffractograms (XRD) of the offretites before and after NH4-exchange and calcination have shown that in their H-form these types of crystals give patterns containing variations in peak position, peak broadening and intensity that are consistent with changes in sample crystallinity and unit cell composition shown in Table 1 (11). Losses in surface area exhibited by the H,Ga-offretite and by the Ga-rich H,Ga,Al-offretite is probably the result of lattice degradation suffered by these two samples during thermal treatments. Table 1 Crystal's composition after H4-exchanl~e and calcination* Sample Catalyst A B C D E
[H, Ga]-Off [H, Al, Ga]-Off [H, Am,Ga]-Off [H, Al, Ga]-Off [H, All-Off
3.17 3.47 2.87 2.75 2.93
0.57 0.35 0.57 0.40 0.62
0.005 0.005 0.005 0.005 0.120
(A102) (GaO2) (SiO2) SA(m2/g) -3.27 3.13 3.01 3.68
3.75 0.56 0.32 0.15 --
14.25 14.17 14.55 14.84 14.32
Based on 18T (A1, Si) atoms per unit cell and on chemical analysis by atomic absorption (11).
322 350 525 503 548
5o 4 "
Temperature (C) Figure 1. Ammonia TPD profiles for several H-offretites. The symbols: (N), (A), (V), and ( 9 refer to H, Ga-offretite; H,Ga,Al-offretites with initial Ga/Ga+A1 ratio of 0.04 and 0.08, and to H,Al-offretite respectively. The symbol (.) refers to HZSM-5 with SAR=30.
984 Table 2 Total Acidity of Ga-containing offretites measured by temperature program desorption of NH3. The crystals compositon is given in Table 1. Sample
Total Acidity micromole NH3/g
t~ 3 -
~b ~c ~d
I 50 to 320 C
320 to 450 C 450 to 540 C Total Acidity
Acidity Distribution Figure 2. Acidity distribution from ammonia thermodesorption. The letters a, b, c, and d refers to H, Ga-offretite; H,Ga,Al-offretite with initial Ga/Ga+A1 ratio of 0.04 and 0.08, and to H, Al-offretite respectively.
985 Infrared spectra of chemisorbed pyridine contain bands characteristic of Bronstedt acid sites (B, at 1,540 cmd) and of Lewis acid sites (L, at 1,452 cmd), (11). Thermodesorption of pyridine from B-acid sites is more facile than from L-sites. In the Al-containing samples, pyridine is retained on both B and L sites, even after degassing in vacuo at 500~ (12). By contrast, at 500~ pyridine is removed from B-sites in the H, Ga-offretite, indicating that these B-sites are weaker than in the M-analog (17). Heat flow microcalorimetry results (12) at 150~ have indicated that Ga insertion into the offretite framework affects both acid site (B+L) distribution and strength. In fact, heats of NH3 chemisorption at 150~ show that the populations of sites present in H, Al-offretites are reduced and that in H, A1,Ga-offretites a new population of strong sites with strength in the 170 to 185 kJ/mol range appear at low (<100 ~tm/g) coverage (12). In contrast to what was observed in H,A1offretites, after sorbing 100 ~tm/g, acid site strength monotonically decreases with coverage in all the Ga containing crystals examined (12). Microcalorimetry results are consistent with the TPD data in Figures 1, 2. A1 and Ga offretites exhibit only one major desorption region in the 100-300~ range. Addition of Ga to the aluminosilicate framework, cause the appearance of a second desorption region in the 400 to 600~ temperature range attributed to the sites with strength near 180 kJ/mol that appears after Ga insertion (see Figure 1). Two desorption sites can be seen also in a sample of H-ZSM-5 with SAR=30. Variations in total acidity (expressed as micromole NH3 desorbed) are given in Figure 2 and Table 2. Propylene conversion data, product distributions and hydrogen transfer indices of the offretite crystals under study are summarized in Table 3. A comparison of H,Ga with H, Al-offretite indicates that H,Ga-offretite is less active and produces much less methane and isobutane than its A1 analog. The higher yields of methane may be due to the greater extent of protolytic cracking of C6 and C9 hydrocarbons since these crystals contain a larger population of stronger acid sites (12). H,Ga-offretite shows a much lower hydrogen transfer index (HTI) than H, Al-offretite (0.46 vs. 0.84), Table 3. The lower activity and HTI of H, Ga-offretite is consistent with its lower density of sites with strength near 175 kJ/mol (12). For H, A1,Ga-offretite crystals with initial Ga/Ga+A1 ratios of 0.04, 0.08 and 0.13 (11), (and with composition, after NH4 exchange and calcination, given in Table 1) propylene conversion levels are not affected by Ga concentrations but are higher than those obtained with either H, A1 or H,Ga-offretite crystals. Ga insertion into the offretite aluminoslicate framework decreases i-C4 formation and (at low levels) minimizes CH4 generation (see Table 3).
986 Table 3 Hydrogen transfer index (HTI) of Ga-containing offretites. The HTI is defined as the ratio of C4 paraffins and the total C4 produced. C3=Cony. Wt%
A. [H, Ga]-Off
B. [H, Al, Ga]-Off
C. [H, Al, Ga]-Off
D. [H, AI, Ga]-Off
Corma et al. (13,14) suggested that in heptane cracking the ratio of isobutane to the total C4 fraction (iC4p/X C4) is controlled by the structure and by the pore dimension of the zeolites used. Isobutane, which is the main component observed in the C4 fraction, is formed from the cracking of the bulkiest isomer, (i.e., 2,2'-dimethylpentane) and to a lesser extent from the 2,4-dimethyl pentane via hydrogen transfer of isobutene formed from cracking reactions. Wielers et al. (15) suggested that the ratio of isobutane to n-butane (iC4p/nC4p) from n-hexane cracking decreases with the zeolite pore diameter. As shown in Table 3, the ratio of (iC4p/nC4p) decreases with increasing Ga content. Thus, it appears that H,Ga,Al-offretite has an effective intermediate pore diameter between H,Ga- and H,Al-offretite owing to extra framework Ga formation and lattice degradation suffered during the NH4-exchange procedure used (see Table 1). The presence of extraframework metal atoms has been observed in NMR results (11). Lukyanov (16) proposed a simple method for the quantitive characterization of hydrogen transfer activity of ZSM-5 and Y zeolites based on the rate of isobutane formation during n-hexane cracking at 400~ HY (SAR=4.44.8) was found to be six-fold more active than HZSM-5 (SAR=34). Based on Lukyanov's proposal, H,Al-offretite and H,Ga,Al-offretite have similar hydrogen transfer activities, and they are more active than HZSM-5 with SAR=30 (16). Furthermore, since the yield of isobutane decreases with increasing Ga contents in H,Ga,Al-offretite, this data tend to suggest that hydrogen transfer activity decreases with increasing Ga levels in the aluminosilicate framework, Table 3.
987 In summary, the three H,Ga,Al-offretite samples have HTI of about 0.930.96, which are higher than the HTI of either H,Ga or H,Al-offretites (see Table 3). The Ga content does not seem to affect the HTI of H,Ga,Al-offretite samples. H,Ga,Al-offretites and HZSM-5 samples with SAR of 30 have similar HTI. This unusually high HTI of H,Ga, Al-offretites might be associated with the small population of acid sites with strength in the 170 to 185 kJ/mol range. The low HTI of H,Ga-offretites suggest that these crystals should be studied as possible FCC additives to increase olefin yields.
2. 3. .
9. 10. 11. 12. 13. 14. 15. 16.
R. J. Madon, J. Catal., 129, (1991), 275. J. S. Buchanan, Appl. Catal., 74, (1991), 83. G. Anders, I. Burkhardt, U. Illgen, I. W. Schulz and J. Scheve, Appl. Catal. 62, (1990), 271. M. F. Elia, E. Iglesias, A. Martinez and M. A. Perez Pascual, Appl. Catal., 73, (1991), 195. L. Nalbandian, A. A. Lemonidou, and I. A. Vasalos, Appl. Catal., 105, (1993), 107. F. N. Guerzoni and J. Abbot, Catalysis Letters, 16, (1992), 53. S. J. Miller and C. R. Hsieh, ACS Petro. Chem. Div. Symposium Preprints, 35, No. 4, (1990), 685. S. Bessell and D. Seddon, J. Catal., 105, (1987), 270. P. Yarlagadda, C. R. F. Lund and E. Ruckenstein, Appl. Catal., 62, (1990), 125. Dai, P. S. E., C. M. Tsang, R. H. Petty, and M. L. Occelli, J. Mol. Catal (Submitted). Occelli, M. L., Eckert, M., Iyer, P. S., and Ritz, P., in "Synthesis of microporous Materials", Marcel Dekker (In Press). Occelli, M. L., Eckert, M. Hudalla, C., Auroux, A., Ritz, P., and Iyer, P. S., Microporous Solids, (Submitted). A. Corma, in "Zeolites: Facts, Figures, Future", (P.A. Jacobs and R. A. van Santen, Eds.), p. 49, 1989, Elsevier, Amsterdam. A. Corma, V. Fornes, J. B. Monton and A. V. Orchilles, J. Catal., 107 (1987) 288. J. Biswas and I. E. Maxwell, in "Zeolites: Facts, Figures, Future", (P.A. Jacobs and R. A. van Santen, Eds.), p. 1263, 1989, Elsevier, Amsterdam. A. D. Lukyanov, J. Catal. 145, 54 (1994).