Heterogeneous Catalysis and Fine Chemicals IV H.U. Blaser, A. Baiker and R. Prins (editors) © 1997 Elsevier Science B.V. All rights reserved.
Influence of the acidity and of the pore structure of zeolites on the alkylation of toluene by 1-heptene. P. Magnoux, A. Mourran, S. Bernard and M. Guisnet URA. CNRS 350, Catalyse en Chimie Organique, University de Poitiers, 40 avenue duRecteur Pineau - 86022 Poitiers cedex France. Summary The alkylation of toluene with 1-heptene was used as a model reaction for the synthesis of long-chain linear alkylbenzenes which are precursors of biodegradable surfactants. The effect of the pore structure and of the acidity of large pore zeolites : HFAU (framework Si/Al ratio from 4 to 100), HMOR (Si/Al from 10 to 80), HMAZ (Si/Al=10), HBEA (Si/Al=10) and of an average pore size zeoUte, HMFI (Si/Al=40) on their catalytic properties was determined in Kquid phase at OO^'C with a toluene/heptene molar ratio of 3. With the large pore zeolites the main reactions are alkene double bond shift and toluene alkylation which occur through a consecutive scheme. Some of the alkylation products, mainly triheptyltoluenes, remain trapped in the zeolite pores. With HMFI, alkylation products are only found in the zeolite pores, because of the impossibility of desorbing these bulky products from the narrow pores of this zeolite. The activity of large pore zeolites depends on their acidity but also on the ease of desorption of alkylates from their pores. In particular mesopores created during dealumination which facilitate the product desorption have a positive role on the zeolite activity. Thus, HFAU with a Si/Al ratio of 30 which has a relatively high acidity and which contains mesopores is the more active catalyst. Highly dealimiinated (hence mesoporous) HMOR samples are also active. Moreover, their selectivity to 2-phenylalkanes which are the most biodegradable isomers is much higher than that of the HFAU samples. By comparison of the compositions of the heptene mixture and of the monoheptyltoluene mixture in the liquid phase and in the zeolite pores, this shape selective preference can be attributed to transition state control. INTRODUCTION Linear mono C10-C13 alkylbenzenes (LAB) which are used in the production of biodegradable surfactants are produced industrially by benzene alkylation with linear alkenes, HF or AICI3 being used as catalysts (1,2). Because their degradabilityis very high, the 2-phenylalkanes are preferred to the other isomers (3). A great effort is now being made in order to substitute soUd adds (in particular zeolites) for these polluting, corrosive, industrial catalysts (4-9). That is why it is important to understand how the alkylating activity, stability and selectivity of zeolite catalysts change as a function of their physico-chemical characteristics (pore structure, acidity). The effect of these characteristics is investigated here on a model reaction, the alkylation of toluene with 1-heptene. The interest which this
reaction presents is due to the ease in the quantitative analysis of the complex reaction mixture : heptene isomers, all the monoheptyltoluenes, biheptyltoluenes, triheptyltoluenes, etc. This model reaction is carried out over large pore zeolites : HFAU (framework Si/Al ratio of 4,16, 30,100) which wiUbe called HFAU4, 16, 30 and 100, HMOR (Si/Al=10, 20, 40, 80), HMAZIO (Si/Al=10), HBEAIO (Si/Al=10) and over an average pore size zeolite: HMFI40 (Si/Al=40). RESULTS AND DISCUSSION 1. Physico-chemical characteristics of the zeolite samples The main characteristics of the zeolite samples are given in table 1. Their unit cell formula was drawn from their elemental analysis and from the nimiber of framework aluminium atoms per unit cell (NAI) estimated from the relationship between the wavenumber of IR structure bands and NAI- Among the zeolite samples, only HFAU4 and 100 and HBEAIO contained a large amount of extraframework aluminium species. Table 1 Characteristics of the zeoUte samples : unit cell formula, number of extraframework aluminium atoms ( N E F A L ) , volume (cm^^-l) of micropores (V^) and of mesopores (VM) and acidity : nimiber (lO^Og-l) of protonic acid sites estimated from the imit cell formula (NH"^) and of sites retaining ammonia adsorbed above lOO'^C (NiooX Zeolite HFAU4 HFAU16 HFAU30 HFAUIOO HMORIO HMOR40 HMOR80 HMAZIO HBEAIO HMFI40
Unit cell formula Na0.4Al39.2Sil52.8O384 Nao.3Alll.3Sil80.70384 Nao.l5Al6.2Sil85.80384 Nao.5All.9Sil90.l0384 Nao.05Al4.4Si43.6O96 Nao.05All.lSi46.9O96 Nao.OlAlo.6Si47.4O96 Nao.03Al3.3Si32.7O72 Nao.2Al3.9Si60.lOl28 Nao.oiAl2.lSi93.90i92
9.6 2.4 0.1 10.0 2.1 0 0 0 1.5 0
0.298 0.295 0.282 0.249 0.195 0.210 0.210 0.170 0.240 0.175
0.056 0.140 0.193 0.161 0.030 0.070 0.085 0.105 0.540 0.000
Nioo 20.3 10.2 5.7 3.9 3.2 1.8 0.85 2.7 9.3 6.3 2.2 2.0 1.2 0.8 9.2 6.0 8.4 7.1 2.4 2.5 NH^
Nitrogen adsorption shows that all the zeolites except HFAU4, HMORIO, HMAZIO and HMFI40 have in addition to micropores a significant mesopore volume. The very large mesopore volume of HBEAIO is due to intercrystalline voids resulting from the agglomeration of the very small crystallites of this zeolite (10). The number of protonic acid sites estimated from the unit cell formulas was compared to the nimiber of sites on which ammonia remained adsorbed above lOO^^C and SOO^'C. In the FAU and MOR series, the number of ammonia molecules
109 which remained adsorbed decreased with the theoretical number of protonic acid sites. However, with HFAU 100, the number of ammonia molecules detected was greater than expected from the unit cell formula, which coxild be attributed to the presence of a large amount of extraframework Al species. Furthermore the lower N A I the greater Sie proportion of strong acid sites (retaining ammonia adsorbed above 300°C). 2. Alkylationof toluene with 1-heptene With all the zeolites 1-heptene isomerizes into 2 and 3-heptenes, monoheptyltoluenes are formed by alkylation of toluene with heptenes, biheptyltoluenes by alkylation of monoheptyltoluenes. Non desorbed products are also found. Neither skeletal isomerization and dimerization of heptenes nor formation of C i 4 alkyl toluene are observed. 2-Heptyltoluene (Mi), 3-heptyltoluene (M2) and 4-heptyltoluene (M3) can be separated by GC : three peaks of M i and M2 and two peaks of M3 are observed, which correspond to ortho, meta and para isomers. Most likely the two major peaks correspond to the ortho and para isomers which can be expected with electrophiUc aromatic substitution. Double bond shift and toluene alkylation involve n-heptyl carbenium ions as intermediates. By considering the toluene alkylation mechanism it can be concluded that Ml results from toluene alkylation with 1- or 2-heptenes, M2 from alkylation of toluene with 2- or 3-heptenes and M3 only from alkylation with 3-heptenes (Figure 1).
c-c-c~c-c-c-c + I-C7
O ^ -'
Figure 1 : Formation of monoheptyltoluenes Only a small amount (<10%) of the non desorbed products are soluble in methylene chloride when the zeoUte is treated by soxhlet. However, after dissolution of the zeolite in a hydrofluoric acid solution and extraction with methylene chloride all of the products were recovered. This shows that most of the non desorbed products are located inside the zeoUte pores. With all the zeolites these products are constituted of mono, bi and triheptyltoluenes.
2.1. Activity and stabiliiy of the zeolite samples. The initial rates of 1-heptene isomerization and of alkylation were estimated from curves : the conversion of 1-heptene into isomers and of n-heptenes into alkylation products plotted as a function of time. To obtain accurate values the experiments were carried out over three different weights of catalysts. The initial rates obtained with the series of FAU zeolites are plotted as a function of nH+, i.e. the theoretical number of acid sites estimated from the unit cell formula (Figure 2). With both reactions the most active catalyst is HFAU30 (nH+=:3.2 1020g^l). The shape of the curve represented in figure 2 is the one predicted by the topologic model of Barthomeuf (11-12). For low values of nH+ (hence of NAI the number of framework aluminium atoms per unit cell) the A104" tetrahedra have no close Al neighbour atom in the second layer of T atoms and the acidity of the protonic sites is maximal. Above a certain value of NAI, the aluminium atoms are no longer isolated and the strength of the corresponding protonic sites, hence their activity is lower. However, the value of NAI for which this situation occurs in the case of FAU zeolites was estimated to be equal to 28, which corresponds to a framework Si/Al ratio of 5.8 i.e. a Si/Al ratio much lower than the one corresponding to the maximum activities. Moreover, the alkylating activity of HFAU4 is abnormally low. All this indicates that acidity is not the only parameter determining the HFAU activity. i
^Ao (10"^mol h''' g"'')
1-/ X. /
MAZ BE)^^s^ T| ^^1 1 10 + 20 nH (10
Figure 2 : Initial rates R of isomerization (I) and of alkylation (A) as a function of the theoretical number of protonic sites (nH+).
sites g )
Figure 3 : Initial activities (Ao) of toluene alkylation as a function of the theoretical number of protonic sites (nH+).
The same occurs with HMOR zeolites : the maximum activity is found for a Si/Al ratio of 80 (Figure 3) while the aluminium atoms become isolated for Si/A1^9.4 (11, 12). Moreover the alkylating activity of HMORIO, the non dealxmiinated sample, is very low, much lower than that of a HFAU zeolite with the same number of acid sites. This is also the case for the non dealuminated HMAZ and HBEA samples. However, figure 3 shows that the greater the Si/Al ratio of the HMOR samples (hence the more pronounced their degree of
dealumination) the smaller the difference in activity between the HMOR and the HFAU samples. This can be explained by the presence of mesopores created by dealumination. Indeed these mesopores sJlow a quasi tridimensional diffusion of the reactant molecules (13) favouring the desorptionof the alkylates. The activity of non dealuminated HMAZ and HBEA samples is also very low, and with HMFI, an average pore size zeolite, alkylation occurs with the alkylation products remaining trapped inside the pores (13wt% of non desorbed mono and biheptyltoluenes after 10 hours reaction) thus no products appear in the liquid phase. Limitations in the desorptionof alkylation products from non dealuminated large pore zeolites could also be responsible for their low activity. In agreement witfithis, the lower the amount of non desorbed products retained in the HFAU or HMOR zeolites at 100% conversion of heptene into alkylation products, the faster the alkylation (Table 2). The increase in the alkylation activity caused by dealumination can be related to the creation of mesopores which would favour the desorption of the alkylation products. However, with this hypothesis a h i ^ alkylating activity should be found with the HBEAIO sample whose crystallite size is very small (around 200A). This is not the case, most likely because the extraframework aluminium species present in this sample Umit the desorption of alkylation products. In agreement with this proposal the amount of products retained in the pores of HBEAIO after complete conversion of heptenes into alkylation products is approximately 10 times greater than that retained in the pores of HFAU30 (Table 2). Table 2 Initial rates of alkylation of toluene, A (10'^ mol.h"l.g"l) and percentage of non desorbed products, C(wt%) at complete conversion of alkenes into alkylation products. Zeolite HFAU HFAU HFAU HFAU
4 16 30 100
A 20 350 600 40
C 23.0 8.1 2.3 9.6
Zeolite HMOR HMOR HMOR HBEA
10 20 80 10
A 7 27 120 13
C 4.4 2.8 2.2 23.5
2.2. Product distribution As shown in figures 4 for HFAU30, double bond shift and toluene alkylation occur through consecutive schemes with all the zeoUte samples: 1- Heptene ,^ ^ 2- Heptenes .^ ^ 3- Heptenes Toluene ^ ^ = = ^ Monoheptyltoluenes (M) ^
However the product distribution depends on the zeoUte. Thus, with HFAU30, biheptyltoluenes appear at a heptene conversion of 30%, 40% with HBEA, 60% withHMORSO and 80% with HMAZ. Whatever the heptene conversion, M l in the monoheptyltoluene fraction is more favoured with HMOR80 than with the other zeoKtes.
20 f 20
Figure 4a : Percentages of 2-heptenes Figure 4b : Yields in mono (M) and (2-C7=)and3-heptenes(3-C7=)inthe biheptyltoluenes (B) as a function of heptene mixture as a function of the ^^P^^^ conversion into alkylates (X). percentage of isomerized 1-heptene (Xi):Xi = 2C7= + 3C7=. Figure 5 illustrates the change in the monoheptyltoluene distribution on the more active faiyasite sample (HFAU30) as a function of n heptene conversion. At low conversion, Ml is formed with a high selectivity (>90% at zero conversion). The percentages of M2 and M3 increase with conversion at the expense of Ml, which can be related to the increase in the amount of 2-heptenes and 3-heptenes in the heptene mixture (Table 3). 80 T
00 -1 •
5^ 40 I 40 20
X (% heptenes)
Figure 5 : Distribution of monoheptyltoluenes on HFAU30 as a fimction of heptenes conversion into alkylation products (X).
•—•- I I I 29
• _ 1
"T"- ^ 80
X (% heptenes)
Figure 6 : Distribution of monoheptyltoluenes on HMOR80 as a function of heptenes conversion into alkylation products (X).
By considering the relative distributions of heptenes and monoheptyltoluenes (Table 3) with respect to the alkylation scheme (Figure 1) it can be concluded t h a t the monoheptyltoluene distribution is determined by the heptene distribution. There is no shape selective effect: steric constraints on certain alkylation steps or limitations of the desorption of the bulkier monoheptyltoluenes. The absence of limitations in the monoheptyltoluenes (M) desorption is confirmed by the very low amount of M in the non desorbed products (5%) and by the fact that the distribution of these monoheptyltoluenes is identical to that found in the desorbed products. Table 3 Comparison of the compositions of the mixtures of heptenes and of monoalkylation products at 40 and 90% conversion (X) of heptenes into alkylation products. X(%)
1-C7=:71% ; 2-C7=:24% ; 3-C7=:5% Mi:70% ; M2:25% ; M3:5%
1-C7=:50% ; 2-C7=:45% ; 3-C7=:5% Mi:88% ; M2:10% ; M3:2%
1-C7=:50% ; 2-C7=:40% ; 3-C7=:10% 1-C7=:30%; 2-C7=:60% ; 3-C7=:10% Mi:60% ; M2:30% ; M3:10% Mi:85% ; M2:10% ; M3:5%
Whatever the n-heptene conversion (Figure 6) the percentage of M l in the monoheptyltoluene fi*action found with the more active mordenite sample (HMOR80) is higher than the one found with HFAU30 (Figure 5). This observation is unexpected when the heptene distributions are compared at identical conversions (Table 3). Therefore this high selectivity to Ml can only be explained by the shape selectivity of mordenite. This selectivity could be partiy due to difficulties in the desorption of the more bulky M2 and M3 products. Indeed non desorbed products contain a larger amount of M products with HMOR80 than with HFAU30 (56% against 5%). However this diffusion control plays only a limited role, for non desorbed monoheptyltoluenes have practically the same composition as the desorbed monoheptyltoluenes (85% of M l instead of 90%). Therefore the high selectivity to M i is most likely due to steric constraints affecting the alkylation of toluene into M2 and M3 in the mordenite pores. EXPERIMENTAL The zeoUte samples were suppUedfi^om PQ (HFAU, HBEA, HMFI), fi-om Zeocat (HMOR) and fix)m Elf Aquitaine (HMAZ). These samples were characterized by IR spectroscopy (structure bands), by nitrogen adsorption at 77 K and by NH^ stepwise TPD. Alkylation of toluene with heptene was carried out in liquid phase under the following operating conditions : 90''C, toluene/alkene molar ratio of 3, 0.07 to 0.5 g powdered zeoUte with a stirring rate of 200 RPM. Before introduction into the toluene/alkene mixture the zeoUte samples were activated under dry air
flow at 500^*0 for 12 hours. Small samples of the reaction mixture (approximately 50 jil) were taken at various reaction times, diluted with methylene chloride and analj^ed by gas chromatography on a 25 m column of fused silica (DB5). The reaction products were identified by gas chromatography/mass spectrometry coupling. The experimental methods used to recover and to analyze the non desorbed products are described in ref. 14. CONCLUSION The activity of zeolites for alkylation of toluene with 1-heptene depends on their acidity and on the ease of desorption of alkylates. Thus average pore size zeolites such as HMFI and monodimensional large pore zeolites are practically inactive. Extraframework aluminium species which Hmit the product desorption have a negative effect, while mesopores (in particular with monodimensional zeolites such as HMOR) have a positive effect. HMOR zeolites are more selective for 2phenylalkanes than HFAU zeolites. This is mainly due to shape selective preference via transition state control. REFERENCES 1. P.R. Pujado, in "Handbook of Petroleum Refining Process" (R.A. Meyers, ed.) McGraw-Hm,1986. 2. J.L.G. de Almeida, M. Dufaux, Y. Ben Taarit and C. Naccache, J. Am. Oil Chem. Soc, 71 (1994) 675. 3. R.J. Larson, T.M. Rothgeb, R.J. Shimp, T.E. Ward and R.M. Ventullo, J. Am. Oil Chem. Soc, 70 (1993) 645. 4. J.L.G. de Almeida, M. Dufaux, Y. Ben Taarit and C. Naccache, Appl. Catal. A : General 114 (1994) 141. 5. S. Sivasanker, A. Thangar^], J. Catal. 138 (1992) 386. 6. S. Sivasanker, A. Thangar^', R.A. Abulia and P. Ratnasamy, in L. Guczi et al. (Eds.), Stud. Surf. Sci. Catal. vol. 75, Elsevier, Amsterdam(1993) p.397. 7. A. Mourran, P. Magnoux and M. Guisnet, J. Chim. Phys., 92 (1995) 1394. 8. L.B. Zinner, K. Zinner, M. Ishige and A.S. Araujo, J. Alloys and Compounds, 193 (1993) 65. 9. P. Magnoux, A. Mourran, S. Bernard and M. Guisnet. in J.Weitkamp and B. Liicke (Eds.), Proceedings of the D.G.M.K. Conference "Catalysis on SoUd Adds and Bases". Berlin (1996) p.49. 10. C. Coutanceau, J.M. Da Silva, M.F. Alvarez, F.R. Ribeiro and M. Guisnet, J. Chim. Phys., to be published. 11. D. Barthomeuf, Mater. Chem. Phys., 17 (1987) 49. 12. D. Barthomeuf, in J.W. Ward (Eds.), Stud. Surf. Sci. Catal. vol. 38 Elsevier, Amsterdam (1987) p.l77. 13. N.S. Gnep, Ph. Roger, P. Cartraud, M. Guisnet, B. Juquin and Ch. Hamon, C. R. Acad. Sci., 309 (1989) 1743. 14. M. Guisnet and P. Magnoux, Appl. Catal., 54 (1989) 1.