|
|
|
Mesoporous Silica Supported Solid Acid CatalystsS. Choi,(a) Y. Wang, and C. H. F. Peden Supported by the U.S. Department of Energy (DOE), Office of Science, Laboratory Technology Research Program. Currently, the chemical and petrochemical industries produce a wide range of organic compounds via alkylation and isomerization by homogeneous acid catalysts such as H2SO4, HF, and AlCl3 (Thomas 1992; Misono and Okuhara 1993; Corma and Martinez 1993). Although current homogeneous catalysts are efficient, as a result of their corrosive and toxic nature, these materials pose potential environmental hazards and present operational problems, including difficulty in separation, recovery, and reutilization, that lead to higher capital costs. Among many solid acid systems, heteropoly acids (HPA) with Keggin anion structures have received considerable attention due to their simple preparation and strong acidity (Misono 1987; Kozhevnikov 1995). Specifically, 12-tungstophosphoric acid (H3PW12O40), denoted as TPA hereinafter, is among the most extensively studied (Misono and Nojiri 1990; Corma 1995; Okuhara et al. 1996) since it possesses the highest Brönsted acidity (Misono et al. 1982). However, to date, low efficiency due to low surface area, rapid deactivation and relatively poor stability are some of the major problems associated with these TPAs in conventional bulk acid forms. Attempts to improve the efficiency of these materials have been made by supporting TPA on various high surface area supports (Kapustin et al. 1990; Pizzio et al. 1998). In part, because a strong interaction between TPA and a support material likely results in the decomposition of the Keggin structure, several reports in the literature have identified silica as a suitable support due to its intrinsic inertness (Kozhevnikov et al. 1995, 1996; Blasco et al. 1998; Kresge et al. 1994; Izumi et al. 1983; Moffat and Kasztelan 1988; Rocchiccioli-Deltcheff et al. 1990). Most recently, mesoporous silica (known as MCM-41), first developed by researchers at Mobil (Beck et al. 1992; Kresge et al. 1992), has been used to support TPA clusters to take advantage of the uniform pore size and highly ordered structures available in these unique materials (Kozhevnikov et al. 1995, 1996; Blasco et al. 1998; Kresge et al. 1994). Another method that could possibly enhance stability of the active clusters in solution is to prepare catalysts in the form of TPA salts (Okuhara et al. 1992, 1996; Soled et al. 1997). For example, partial substitution of Cs+ for protons render bulk TPAs with higher surface area (up to 150 m2/g compared to 5 m2/g) and improved thermal stability than their parent acids (Okuhara et al. 1996). In addition, TPA salts are known to be insoluble even in liquids as polar as water. Consequently, TPA salts should be better suited for practical applications that might involve polar reagents in harsh operating conditions. However, their small particle size (~mm) limit their application for use as catalysts in commercial fixed-bed or slurry-type reactors. An obvious solution as often applied in industrial practice is to support these TPA salts on a larger particle size (~mm) carrier. Unfortunately, preparation of these catalysts in an engineered form is challenging since direct aqueous impregnation is not feasible. For example, materials synthesized by Soled et al. (1997) consisted of thin internal rings of Cs-substituted TPA salt (so-called egg-white distribution) within the silica extrudate suggesting nonuniform dispersion of the active clusters on silica. The premise of our work (Wang et al. 1998, in press; Choi et al. in press), some of which is described here, is that dispersion of TPA can be manipulated by adopting appropriate grafting techniques (Figure 7.4). We have prepared a series of mesoporous silica supported TPA and Cs-TPA salts with highly dispersed and intact Keggin anions, and compared their structural and catalytic properties to conventionally prepared supported catalysts and their bulk counterparts. For our studies, MCM-41 type mesoporous silica with mono-dimensional pores of 18, 30, 50 and 100 Å were synthesized using a protocol reported elsewhere (Beck et al. 1992; Kresge et al. 1992; Wang et al. in press; Choi et al. in press). Preparation of highly dispersed TPA (Wang et al. 1998, in press; Choi et al. in press) and Cs-TPA (Wang et al. 1998; Choi et al. in press) on mesoporous silica, denoted as Csx-TPA/MS (x = Cs stoichiometry, MS = mesoporous silica) hereinafter, are described in detail elsewhere. For comparison, supported Cs-TPA samples using the two-step impregnation method reported by Soled et al. (1997) were also prepared. Bulk Cs-TPA materials were prepared by previously published (Okuhara et al. 1992) procedures.
Keggin structures of bulk and supported samples were examined using FT-IR spectroscopy and 31P NMR experiments. The physical properties of the catalysts were assessed with thermal analysis experiments, BET surface area and pore size distributions measurements, and transmission electron microscopy (TEM). The composition was also analyzed using the energy dispersive spectroscopy technique (EDS). Selected samples were examined for leaching of TPA by water after stirring vigorously for 2 hours at 50°C in a water bath. Thermal stabilities of the TPA/MS (x=0) catalysts were studied with in-situ x-ray diffraction. TPA/MS (x=0) catalysts were evaluated with a probe reaction, the alkylation of 4-t-butylphenol (TBP, Aldrich, 99%) with styrene (Aldrich, 99+%) (Kozhevnikov et al. 1992). The catalytic properties of the Csx-TPA/MS materials were evaluated using the liquid phase alkylation of 1,3,5-trimethylbenzene (mesitylene, Aldrich, 98%) with cyclohexene (Aldrich, 99%) (Soled et al. 1997; Okuhara et al. 1995). Specific experimental details are provided elsewhere (Wang et al. 1998, in press; Choi et al. in press).
1. Characterization and Reactivity of Supported TPA CatalystsIn collaboration with P. Kaviratna,(a) A. Kim, X. S. Li,(b) L.-Q. Wang,(b) B. Bunker,(c,d) and Prof. W. J. Thompson(c,e) (a) Postdoctoral Research Fellow. Mesoporous silicas with three uniform pore size distributions (18 Å, 30 Å, and 100 Å) were chosen as the supports for unsubstituted (i.e., x=0 in Csx-TPA/MS) TPA catalysts. For comparison purposes, a premium, amorphous Lyosil silica was also used as a catalyst support. Even at relatively high (70 wt%) TPA loadings, supported catalysts still have rather large surface areas (>150 m2/gm). It is noteworthy that the pore sizes of the 100 Å and Lyosil silica-supported TPA catalysts decrease as the TPA loading increases, while pore sizes of the 18 Å and 30 Å silica-supported TPA catalysts do not change with TPA loading. We hypothesize that the rather large-sized TPA clusters (12 Å) can readily clog the pores during solution impregnation when support pore sizes are smaller than 30 Å. In contrast, they likely have a better chance to uniformly coat the pores when the support pore size is sufficiently large as in the 100 Å and Lyosil silicas. Factors such as the surface properties of supports (Kozhevnikov 1995) and the pH of the sample preparation solution (McGarney and Moffat 1991) play important roles in affecting the Keggin structure of TPA. 31P NMR is the most convenient and revealing method for the assessment of the phosphorous environment in the phosphorous-containing HPA compounds, and, in turn, the stability of the Keggin structure of TPA. The spectra for the intact Keggin structure of TPA should consist of a single line at ~-15 ppm (Pope 1983). The presence of the second resonance at ~-13 ppm is assigned to lacunary (defect) anions, such as PW11O397- and P2W17O6110-, or unsaturated anions, such as P2W18O626- and P2W21O716- (Pope 1983), which are formed via decomposition of the Keggin structure of TPA. As evidenced by a small peak with the chemical shift at ~-13 ppm in the spectra of untreated MS-supported materials, impregnation of an untreated mesoporous silica leads to the partial decomposition of the Keggin-type structure of TPA (Wang et al. in press). The intact Keggin structure was preserved when TPA was supported on silica neutralized with 1N nitric at high TPA loadings although, again, decomposition was observed for lower loading levels. However, when methanol is used to impregnate acid-treated supports with TPA, the intact Keggin structure was retained even at a 10 wt% TPA loading. It should be noted that all catalysts with decomposed Keggin structures were found to be inactive in the probe reaction studied in this paper under the conditions investigated. Therefore, even with inert supports such as silica, neutralization of supports with acids and use of non-hydrolyzing polar solvents such as methanol are required to retain the Keggin structure when preparing supported TPA catalysts with a solution impregnation method. The effect of thermal treatment on the stability of silica supported heteropoly acid (HPA) catalysts has been studied previously (Misono 1987; Kozhevnikov 1995; Rocchiccioli-Deltcheff et al. 1990; Kasztelan et al. 1990) using TGA and/or DTA techniques, although both enhanced (Kasztelan et al. 1990) and unaffected (Rocchiccioli-Deltcheff et al. 1990) thermal stability of TPA by silica supports have been observed. Thus, we determined the thermal stability of TPA and supported TPA catalysts using in situ x-ray diffraction. Decomposition of TPA is believed to occur (Misono 1987) via: H3PW12O40 ® P2O5 + 12WO3 + 3/2 H2O (1) Because supported TPA catalysts are amorphous even at a 70 wt% loading, the decomposition of TPA was studied in this work by following the evolution of crystalline WO3. Crystalline rhombohedral WO3 was detected at about 510° C in the bulk TPA sample, which is consistent with the previously reported TPA decomposition temperature range (Misono 1987; Kozhevnikov 1995). With the 50 wt% TPA/30 Å SiO2 catalyst, WO3 diffraction lines appear at about 585°C, indicating that the thermal stability of TPA is enhanced by ~75°C when TPA is supported on 30 Å SiO2 (Wang et al. in press). The catalytic performance of supported TPA catalysts were evaluated using the probe reaction, alkylation of TBP by styrene. Both products, 2-(1-phenylethyl)-4-t-butylphenol (product 1) and 2,6-bis-(1-phenylethyl)-4-t-butylphenol (product 2), are bulky molecules. Thus, shape selectivity to the products provided by mesoporous silica can be studied. Selectivities at 90% conversion of TBP are summarized in Table 7.1. Compared with a bulk TPA material, supported TPA catalysts show at least a sixfold enhancement in activity due to more accessible protons when TPA is supported on large surface area silica (Wang et al. in press). TPA catalysts of 30 Å and 100 Å on silica exhibit about twice as high activity as that of Lyosil silica supported TPA catalysts, again due to the huge surface area with mesoporous silica. The higher activity observed with 100 Å supported TPA material over that of 30 Å supported TPA catalyst is ascribed to the better dispersion of TPA due to less TPA clogging in large pores (100 Å) during sample preparation. On the other hand, the 30 Å silica supported TPA catalyst showed much improved shape selectivity to monoalkylated product 1 (Table 7.1), which is attributed to a steric hindrance to dialkylated product 2 (about 20 Å; Kozhevnikov et al. 1995) by its uniform and compatible pore size (30 Å). 18 Å silica supported TPA showed very low activity compared with other supported TPA catalysts, likely due to the poor dispersion of TPA as described above, or to the severe pore diffusion resistance to the bulky reactants.
2. Characterization and Reactivity of Supported Cs-TPA CatalystsIn collaboration with Z. Nie,(a) D. Kambhampati,(b) and J. Liu (a) Environmental and Health Sciences Division. Primary structures of the supported, Cs-substituted (Csx-TPA/MS) catalysts were identified by comparing their FT-IR absorbance bands to those of bulk TPA, tungstophosphoric acid salt (Cs-TPA), and mesoporous silica (Wang et al. 1998; Choi et al. in press). Bulk TPA (H3PW12O40) and TPA salt (Cs2.5H0.5PW12O40) show the characteristic IR bands at ~1080 (P-O in the central tetrahedra), 984 (terminal W=O), 897 and 812 (W-O-W) cm-1 associated with the asymmetric vibrations in the Keggin polyanion. The same distinguishable features were observed for the 50 wt% TPA/MS and Cs-TPA/MS catalysts, indicating that the primary Keggin structure is preserved after supporting it onto mesoporous silica. Furthermore, 31P NMR results have confirmed that [PW12O40]3- were the only species present on the support as evidenced by a single 31P NMR peak at a chemical shift of ~-15 ppm referenced to a 0 ppm response from 85% H3PO4. Thus, the materials prepared in this study had intact polyanion structures on the silica surface. The dispersion of Cs-TPA on mesoporous silica can be inferred from the TEM results illustrated in Figure 7.5, and from EDS analysis (not shown). As mentioned above, direct impregnation using a Cs-TPA solution was not possible since Cs-TPA is not sufficiently soluble in any solvent. The material we prepared using a two-step method used previously by Soled et al. (1997) resulted in a segregated phase where Cs-TPA is not uniformly dispersed (Figure 7.5a). In contrast, the material newly prepared here (synthesis described in Wang et al. 1998) consists of uniformly dispersed Cs-TPA salt on mesoporous silica (Figure 7.5b).
An important potential benefit of supporting TPA on oxide supports (including MS) is enhanced stability for the TPA salts. In Section 3.1, we report that the thermal stability of HPA is enhanced by 75°C (to 585°C) when TPA is supported on mesoporous silica (Wang et al. in press). It is noteworthy that the thermal stability of the supported TPA is enhanced further even with a single Cs substitution. However, the enhancement in thermal stability for these catalysts comes at the expense of the catalytically active acid protons. Thus, the effect of Cs-substitution on catalytic activity must also be assessed (see below). The stability of the active species in solution has also been of concern for solid acids, specifically for the supported materials (Okuhara et al. 1996; Wang et al. in press) because of a likely weak interaction of TPA species with silica. This can result in significant leaching of TPA in the presence of a polar solvent. Thus, it is significant that we find that supported catalysts show the same trend as the bulk TPA materials where resistance to leaching is improved considerably with increasing Cs stoichiometry. Catalyst activities were evaluated using the alkylation of trimethylbenzene (mesitylene) by cyclohexene as a model reaction. The 50 wt% Cs2.5-TPA/MS material synthesized in this study (Wang et al. 1998; Choi et al. in press) was about five times as active as that made from a previously published method (Soled et al. 1997) and was also more active than a bulk Cs2.5-TPA material. The primary purpose for adopting the Cs = 2.5 for bulk materials is that this specific stoichiometry provides high surface area and mesoporous porosity giving rise to optimized activity (Okuhara et al. 1996). Considering the fact that a high surface area carrier with ordered structure is adopted here to support the TPA clusters and that more Cs substitution leads to a reduction of acidic protons, we have prepared and examined supported Cs-TPA catalysts with a lower Cs stoichiometry. In this way, Cs substitution levels, that are optimized for catalytic activity and stability to solvent leaching, can be determined. In Figure 7.6, activities of MS-supported 50 wt% Cs-TPA are plotted as a function of Cs stoichiometry.
Typically, the Cs-TPA catalysts prepared by us (Wang et al. 1998; Choi et al. in press) demonstrated superior catalytic properties to those prepared with the previously published method (Soled et al. 1997), regardless of the Cs stoichiometry. Even the supported TPA materials (without Cs substitution) were more active and therefore better dispersed than those prepared using the "conventional" (Soled et al. 1997) method. That is, the former materials were probably less susceptible to pore-clogging by poorly dispersed acid anion clusters. It is also noteworthy that supported Cs1-TPA catalysts are more active than their supported parent acids despite lower quantities of active, acid proton sites. This result suggests that the dispersion of Cs-TPA clusters is improved regardless of the preparation method. It is important to note that these measurements may underestimate the enhanced activity of the novel, improved catalysts prepared in this study relative to those prepared by the "conventional" method. This is because conversions for the most active materials were near 100% under the conditions tested here. ReferencesBeck, J. S., J. C. Vartuli, W. J. Roth, M. E. Leonowicz, and C. T. Kresge, J. Am. Chem. Soc. 114, 10,834 (1992). Blasco, T., A. Corma, A. Martínez, and P. Martínez-Escolano, J. Catal. 177, 306 (1998). Choi, S., Y. Wang, Z. Nie, J. Liu, and C. H. F. Peden, Catal. Today, in press. Corma, A., Chem. Rev. 95, 559 (1995). Corma, A., and A. Martínez, Catal. Rev.-Sci. Eng. 35, 483 (1993). Izumi, Y., R. Hasebe, and K. Urabe, J. Catal. 84, 402 (1983). Kapustin, G. I., T. R. Brueva, A. L. Klyachko, M. N. Timofeeva, S. M. Kulikov, and I. V. Kozhevnikov, Kinet. Katal. 31, 1017 (1990). Kasztelan, S., E. Payen, and J. B. Moffat, J. Catalysis 125, 45 (1990). Kozhevnikov, I. V., Catal. Rev.-Sci. Eng. 37(2), 311 (1995). Kozhevnikov, I. V., A. I. Tsyganok, M. N. Timofeeva, S. M. Kulikov, and V. N. Sidelnikov, React. Kinet. Catal. Lett. 46, 17 (1992). Kozhevnikov, I. V., A. Sinnema, R. J. J. Jansen, K. Pamin, and H. van Bekkum, Catal. Lett. 30, 241 (1995). Kozhevnikov, I. V., K. R. Kloetstra, A. Sinnema, H. W. Zandbergen, and H. van Bekkum, J. Mol. Catal. A: Chemical 114, 287 (1996). Kresge, C. T., M. E. Leonowicz, W. J. Roth, J. C. Vartuli, and J. S. Beck, Nature 359, 710 (1992). Kresge, C. T., D. S. Marler, G. S. Rav, and R. H. Rose, U.S. Patent 5366945 (1994). McGarney, G. B., and J. B. Moffat, J. Molecular Catal. 69, 137 (1991). Misono, M., Catal. Rev.-Sci. Eng. 29, 269 (1987). Misono, M., N. Mizuno, K. Katamura, A. Kasai, Y. Konishi, K. Sakata, T. Okuhara, and Y. Yoneda, Bull. Chem. Soc. Japan 55, 400 (1982). Misono, M., and N. Nojiri, Appl. Catal. 64(1), 1 (1990). Misono, M., and T. Okuhara, Chem Tech. 11, 23 (1993). Moffat, J. B., and S. Kasztelan, J. Catal. 109, 206 (1988). Okuhara, T., N. Mizuno, and M. Misono, Adv. in Catal., 41, 113 (1996). Okuhara, T., T. Nishimura, and M. Misono, Chem. Lett. 155 (1995). Okuhara, T., T. Nishimura, H. Watanabe, and M. Misono, J. Mol. Catal. 74, 247 (1992). Pizzio, L. R., C. V. Cáceres, and M. N. Blanco, Appl. Catal. A: General 167, 283 (1998). Pope, M. T., Heteropoly and Isopoly Oxometalates (Springer, Berlin, 1983). Rocchiccioli-Deltcheff, C., M. Amirouche, G. Herve, M. Fournier, M. Che, and J. M. Tatibouet, J. Catal. 126, 591 (1990). Soled, S., S. Miseo, G. McVicker, W. E. Gates, A. Gutierrez, and J. Paes, Catal. Today 36, 441 (1997). Thomas, J. M., Sci. Am. 266, 112 (1992). Wang, Y., S. Choi, and C. H. F. Peden, patent pending (November 1998). Wang, Y., A. Y. Kim, X. S. Li, L. Wang, C. H. F. Peden, and B. C. Bunker, in Shape-Selective Catalysis (American Chemical Society, Washington, D.C.) in press.
William R. Wiley Environmental Molecular Sciences Laboratory Feedback: webmaster@emsl.pnl.gov Revised: April 17, 2000 Security & Privacy PNNL-13147 |
