Influence of Surface Modification of Ti− SBA15 Catalysts on the

Department of Chemistry, University of California, Berkeley, Berkeley, California ... Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkele...
1 downloads 0 Views 160KB Size
9576

Langmuir 2005, 21, 9576-9583

Influence of Surface Modification of Ti-SBA15 Catalysts on the Epoxidation Mechanism for Cyclohexene with Aqueous Hydrogen Peroxide Richard L. Brutchey,†,‡ Daniel A. Ruddy,†,‡ Lars K. Andersen,§ and T. Don Tilley*,†,‡ Department of Chemistry, University of California, Berkeley, Berkeley, California 94720-1460, and the Chemical Sciences and Physical Biosciences Divisions, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, California 94720 Received May 2, 2005. In Final Form: August 11, 2005 The thermolytic molecular precursor method was used to introduce site-isolated Ti(IV) centers onto the surface of a mesoporous SBA15 support. The resulting surface Si-OH/Ti-OH sites of the Ti-SBA15 catalysts were modified with a series of (N,N-dimethylamino)trialkylsilanes, Me2N-SiMe2(R) (where R ) Me, nBu, or nOc). Compared with the unmodified catalysts, the surface-modified catalysts are more active in the oxidation of cyclohexene with H2O2 and exhibit a significantly higher selectivity (up to 58%) for cyclohexene oxide formation (vs allylic oxidation products). In situ Fourier transform infrared (FTIR) and diffuse reflectance UV visible (DRUV-vis) spectroscopies were used to probe this phenomenon, and it was determined that active sites with capped titanol centers, (SiOsurface)3Ti(OSiR3), likely undergo Ti-OOH formation upon addition of H2O2 in a manner analogous to that for active sites of the type (SiOsurface)3TiOH. On the basis of the observation of similar Ti-OOH intermediates for both species, the electron-withdrawing effects on the Ti(IV) active site, resulting from the surface modification, are likely responsible for the observed increase in selectivity.

Introduction Numerous Ti(IV)/SiO2 materials have been studied for the selective catalytic oxidation of hydrocarbons since the development of a heterogeneous silica-supported Ti(IV) epoxidation catalyst by Shell in the 1970s.1 Perhaps the most studied Ti(IV)/SiO2 catalyst is titanosilicalite-1 (TS1), an efficient oxidation catalyst for a range of organic substrates. Notably, TS1 is an excellent olefin epoxidation catalyst that utilizes aqueous H2O2 as the oxidant.2-5 Epoxides are important synthetic intermediates for fine chemical production, and high-tonnage products such as propylene oxide are used in polyether polymer synthesis. TS1 is limited, however, to reactants and products that can diffuse in and out of the relatively small 0.6 nm diameter channels.6 Consequently, titanium-containing mesoporous catalysts (e.g., Ti-MCM41 and Ti-SBA15) that accommodate larger reactants and products were developed.7-13 †

University of California. Chemical Sciences Division, Lawrence Berkeley National Laboratory. § Physical Biosciences Division, Lawrence Berkeley National Laboratory. ‡

(1) Sheldon, R. A. J. Mol. Catal. 1980, 7, 107. (2) Taramaso, M.; Perego, G.; Notari, B. U.S. Patent 4410501, 1983. (3) Clerici, M. G.; Bellusi, G.; Romano, U. J. Catal. 1991, 129, 159. (4) Clerici, M. G.; Ingallina, P. J. Catal. 1993, 140, 71. (5) Notari, B. Adv. Catal. 1996, 41, 253. (6) Corma, A.; Camblor, M. A.; Esteve, P.; Martinez, A.; Pe´rezPariente, J. J. Catal. 1994, 145, 151. (7) Luan, Z.; Maes, E. M.; van der Heide, P. A.; Zhao, D.; Czernuszewicz, R. S.; Kevan, L. Chem. Mater. 1999, 11, 3680. (8) Wu, P.; Tatsumi, T.; Komatsu, T.; Yashima, T. Chem. Mater. 2002, 14, 1657. (9) Chiker, F.; Launay, F.; Nogier, J. P.; Bonardet, J. L. Green Chem. 2003, 5, 318. (10) Blasco, T.; Corma, A.; Navarro, M. T.; Pe´rez-Pariente, J. J. Catal. 1995, 156, 65. (11) Berlini, C.; Ferraris, G.; Guidotti, M.; Moretti, G.; Psaro, R.; Ravasio, N. Microporous Mesoporous Mater. 2001, 44-45, 595. (12) Widenmeyer, M.; Grasser, S.; Ko¨hler, K.; Anwander, R. Microporous Mesoporous Mater. 2001, 44-45, 327.

The thermolytic molecular precursor (TMP) route is a useful and proven method for grafting low-weight loadings of a metal onto a silica surface to give stabilized, siteisolated active sites.14-20 We have previously developed several Ti-SBA15 catalysts, prepared via grafting oxygenrich molecular precursors onto silica to give isolated, surface-bound Ti(IV) centers after calcination (eq 1).14,15

These catalysts are highly active (up to 100% conversion after 2 h) and selective (90-100%) in the catalytic epoxidation of cyclohexene using either tert-butyl hydroperoxide (TBHP) or cumene hydroperoxide (CHP) as oxidants. The catalytic activity is drastically lower, however, when aqueous H2O2 is used as the oxidant, presumably because the surface-bound Ti(IV) sites are hydrophilic and deactivated by water coordination. In particular, water is thought to hinder formation of the (13) Notestein, J. M.; Iglesia, E.; Katz, A. J. Am. Chem. Soc. 2004, 126, 16478. (14) Jarupatrakorn, J.; Tilley, T. D. J. Am. Chem. Soc. 2002, 124, 8380. (15) Brutchey, R. L.; Mork, B. V.; Sirbuly, D. J.; Yang, P.; Tilley, T. D. J. Mol. Catal. A 2005, 238, 1. (16) Nozaki, C.; Lugmair, C. G.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2002, 124, 13194. (17) Fujdala, K. L.; Drake, I. J.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2004, 126, 10864. (18) Drake, I. J.; Fujdala, K. L.; Baxamusa, S.; Bell, A. T.; Tilley, T. D. J. Phys. Chem. B 2004, 108, 18421. (19) Drake, I. J.; Fujdala, K. L.; Bell, A. T.; Tilley, T. D. J. Catal. 2005, 229, 538. (20) Brutchey, R. L.; Lugmair, C. G.; Schebaum, L. O.; Tilley, T. D. J. Catal. 2005, 229, 72.

10.1021/la051182j CCC: $30.25 © 2005 American Chemical Society Published on Web 09/16/2005

Modified Ti-SBA15 Catalysts for Epoxidation

titanium-hydroperoxide complex by competitively binding to the metal center.1 Nonetheless, the use of dilute aqueous H2O2 is still of interest because it is an inexpensive and environmentally benign oxidizing agent. Commercial epoxidation reactions often employ peroxyacid oxidants that produce organic acid waste as a side product, whereas H2O2 degrades to H2O. For surface-bound active sites, the chemical properties of the surface play an important role in determining catalyst activity and selectivity. Surface properties may, for example, influence the rate of substrate adsorption or product desorption, a concept well established in the mechanism of enzyme catalysis.21 Several reports over the past decade have described surface modification of Ti(IV)/SiO2 catalysts with alkylsilane coupling agents to render hydrophobic surfaces. The reagents employed for this purpose include silyl chlorides,22-24 silylalkoxides,25-29 silylamides,30-32 silylimidazoles,33 and silylacetamides.34,35 The first examples of surface-modified Ti(IV)/SiO2 catalysts were microporous xerogels with alkylsilyl groups incorporated via sol-gel synthesis.26,28 Of particular interest, however, are mesoporous Ti(IV)/SiO2 catalysts (vide supra) modified by various alkylsilanes to yield surface-bound, hydrophobic alkyl groups. These mesoporous catalysts accommodate bulky substrates such as cyclohexene,23,33 R-pinene,25 1-octene,34 and cyclooctene34 and utilize aqueous hydrogen peroxide as the oxidant. Tatsumi and co-workers first reported a surface-modified Ti-MCM41 catalyst that was 21 times more active in turnover number (13.3% conversion after 3 h) than the unmodified catalyst for the oxidation of cyclohexene with aqueous H2O2.23 Similarly, Rhee and Bu reported that surface modification of mesoporous Ti-MCM41 catalysts led to a 7.3 times increase in turnover number (35% conversion after 4 h) and a 3 times increase in selectivity (40% for epoxide) for the oxidation of cyclohexene with aqueous H2O2 as compared to the unmodified catalysts.33 It has been implied that the increased activity and selectivity achieved with surface-modified epoxidation catalysts stem from the hydrophobic nature of the surface. The observed increase in selectivity for epoxidation (which occurs by an oxygen-transfer mechanism) over allylic oxidation (which occurs by a radical mechanism) is not sufficiently explained, however, by the hydrophilicity/ hydrophobicity of the surface.36 Additionally, not a lot of research has gone into understanding this observation from a mechanistic standpoint. (21) Andwander, R. Chem. Mater. 2001, 13, 4419. (22) Zhao, X. S.; Lu, G. Q.; Hu, X. Microporous Mesoporous Mater. 2000, 41, 37. (23) Tatsumi, T.; Koyano, K. A.; Igarashi, N. Chem. Commun. 1998, 325. (24) Nur, H.; Ikeda, S.; Ohtani, B. J. Catal. 2001, 204, 402. (25) Kapoor, M. P.; Bhumik, A.; Inagaki, S.; Kuraoka, K.; Yazawa, T. J. Mater. Chem. 2002, 12, 3078. (26) Kochkar, H.; Figueras, F. J. Catal. 1997, 171, 420. (27) Deng, Y.; Maier, W. F. J. Catal. 2001, 199, 115. (28) Klein, S.; Maier, W. F. Angew. Chem., Intl. Ed. Engl. 1996, 35, 2230. (29) Mu¨ller, C. A.; Deck, R.; Mallat, T.; Baiker, A. Top. Catal. 2000, 11/12, 369. (30) Pen˜a, M. L.; Dellarocca, V.; Rey, F.; Corma, A.; Coluccia, S.; Marchese, L. Microporous Mesoporous Mater. 2001, 44-45, 345. (31) Corma, A.; Domine, M.; Gaona, J. A.; Jorda´, J. L.; Navarro, M. T.; Rey, F.; Pe´rez-Pariente, J.; Tsuji, J.; McCulloch, B.; Nemeth, L. T. Chem. Commun. 1998, 2211. (32) Corma, A.; Garcı´a, H.; Navarro, M. T.; Palomares, E. J.; Rey, F. Chem. Mater. 2000, 12, 3068. (33) Rhee, H.-K.; Bu, J. Catal. Lett. 2000, 66, 245. (34) D’Amore, M. B.; Schwarz, S. Chem. Commun. 1999, 121. (35) Bu, J.; Rhee, H.-K. Catal. Lett. 2000, 65, 141. (36) Holmes, S. A.; Quignard, F.; Choplin, A.; Teissier, R.; Kervennal, J. J. Catal. 1998, 176, 182.

Langmuir, Vol. 21, No. 21, 2005 9577

This report describes a systematic study of surface modifications for two Ti-SBA15 catalysts developed by the TMP method via silylation of surface OH groups with three silyl amides of varying chain length, Me2NSiMe2(R) (R ) methyl, n-butyl, or n-octyl). We set out to understand mechanistically how surface modification affects the epoxidation of cyclohexene by Ti(IV) active sites with aqueous H2O2 to offer an explanation for the observed difference in catalytic selectivity. Results and Discussion Surface Modification and Characterization of Ti-SBA15 Catalysts. The mesoporous silica SBA-15 was prepared according to the literature procedure.38 The Ti-SBA15 catalysts were prepared as previously described by grafting either Ti[OSi(OtBu)3]4 (1) or [(tBuO)2Ti{µ-O2Si[OSi(OtBu)3]2}]2 (2) onto the silica surface to yield Ti1SBA15 or Ti2SBA15, respectively, after calcination to 200-300 °C under nitrogen.14,15 The Ti1SBA15 and Ti2SBA15 materials are spectroscopically very similar (by UV-vis and photoluminescence spectroscopies) and are thought to possess site-isolated, 4-coordinate Ti(IV) centers on the surface. Because of the nature of the molecular precursor, however, the Ti2SBA15 catalyst is believed to have an active site with two isolated Ti(IV) centers in close proximity to one another on the surface. Titanium analyses by inductively coupled plasma atomic emission spectroscopy determined that the catalysts possessed Ti wt % values of 0.22% for Ti1SBA15 and 0.17% for Ti2SBA15. After synthesis and drying, the Ti-SBA15 catalysts were treated with excess (N,N-dimethylamino)trialkylsilane, Me2N-SiMe2(R) (R ) Me, nBu, or nOc), in hexane at room temperature to silylate surface silanol and titanol sites (eq 2). The (N,N-dimethylamino)trialkylsilanes were

chosen because (a) they are stronger silylating agents than the analogous silyl chlorides and alkoxides, (b) they are monofunctional bonding reagents that will produce a specific surface structure, and (c) their use is accompanied by elimination of a readily removed elimination product (HNMe2).39,40 The surface-modified catalysts were filtered and dried in vacuo at 120 °C to yield the corresponding RcapTi-SBA15 catalysts. The initial concentration of surface Brønsted acid sites, [OH], was determined by quantification of the amount of toluene evolved after reaction of the unmodified materials with Mg(CH2Ph)2‚2THF.41,42 The degree of surface functionalization for each catalyst after modification was determined by combustion analysis and was corroborated by thermogravimetric analysis (TGA) (Table 1). On the (37) Coles, M. P.; Lugmair, C. G.; Terry, K. W.; Tilley, T. D. Chem. Mater. 2000, 12, 122. (38) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc. 1998, 120, 6024. (39) Gun’ko, V. M.; Sheeran, D. J.; Augustine, S. M.; Blitz, J. P. J. Colloid Interface Sci. 2002, 249, 123. (40) Cao, C.; Fadeev, A. Y.; McCarthy, T. J. Langmuir 2001, 17, 757. (41) Fujdala, K. L.; Tilley, T. D. J. Am. Chem. Soc. 2001, 123, 10133. (42) Brutchey, R. L.; Goldberger, J. E.; Koffas, T. S.; Tilley, T. D. Chem. Mater. 2003, 15, 1040.

9578

Langmuir, Vol. 21, No. 21, 2005

Brutchey et al.

Table 1. Surface Modification of Ti-SBA15 Catalysts

material

surface coverage (%)

Ti1SBA15 MecapTi1SBA15 BucapTi1SBA15 OccapTi1SBA15

(1.2 OH/nm2)a 74 73 63

0 0.89 0.88 0.76

625 420 340 270

3.3 3.2 2.8 2.8

0.80 0.53 0.46 0.39

Ti2SBA15 MecapTi2SBA15 BucapTi2SBA15 OccapTi2SBA15

(1.4 OH/nm2)a 69 64 58

0 0.97 0.90 0.82

630 380 330 265

3.3 2.9 2.8 2.8

0.70 0.48 0.43 0.38

a

-SiMe2R SBET rp Vp (nm-2) (m2 g-1) (nm) (cc g-1)

Concentration of surface OH sites before functionalization.

Figure 2. DRUV-vis spectra of BucapTi2SBA15 (1) and Ti2SBA15 (2) taken in ambient conditions. Table 2. Thermal Analysis (to 150 °C) of Ti-SBA15 Catalysts Stored in a Humid Environmenta for 24 h at 25 °C

Figure 1. Nitrogen adsorption-desorption isotherms for Ti2SBA15 (9) and BucapTi2SBA15 ([). Pore size distributions calculated from the adsorption isotherm branch are shown as insets for Ti2SBA15 (--) and BucapTi2SBA15 (- - -).

basis of these results, the surface functionalization ranged between ca. 60 and 75% of the available OH sites. Generally, as the alkyl groups on the silylating agent become larger, the degree of surface functionalization decreases slightly (e.g., 74% for MecapTi1SBA15 and 63% for OccapTi1SBA15). On the basis of initial surface OH coverage, there were, on average, 0.76-0.97 trialkylsilyl groups per square nanometer. The extent of grafting can also be quantified by solution 1H NMR spectroscopy. The amount of (N,N-dimethylamino)trimethylsilane depleted from solution as it grafts to the surface of Ti1SBA15 was quantified to give a calculated surface functionalization of 71% of OH sites, within 4% of the value calculated by combustion analysis and TGA. As an indication of the high reactivity of the (N,N-dimethylamino)trialkylsilanes, modification of SBA-15 with Cl-SiMe2Pr under identical conditions resulted in functionalization of only 32% of the available OH sites. Nitrogen porosimetry was used to evaluate the pore structures and surface areas of the surface-functionalized materials (Table 1). The adsorption-desorption data correspond to type IV isotherms (Figure 1), which are characteristic of mesoporous SBA-15 type materials.14,15,38 The unmodified Ti-SBA15 catalysts were found to have high Brunauer-Emmet-Teller (BET) surface areas of ca. 630 m2 g-1 (pore volume of 0.70-0.80 cm3 g-1).43 Upon surface modification, the surface area and pore volume decreased depending on the size of the silylating agent. The MecapTi2SBA15 catalyst had a BET surface area of 380 m2 g-1 (pore volume of 0.48 cm3 g-1), whereas the OccapTi2SBA15 catalyst had a lower surface area of 265 m2 g-1 (pore volume of 0.38 cm3 g-1). The range of pore radii for the modified materials, as calculated from the nitrogen adsorption isotherm, was observed to be narrow and slightly lower than that for the unmodified samples. Furthermore, the mesostructured architecture of the Ti-SBA15 catalysts was maintained after surface modification, as observed by transmission electron microscopy

material

wt % H2O adsorbed

desorption T (°C)b

Ti1SBA15 MecapTi1SBA15 BucapTi1SBA15 OccapTi1SBA15

31.7 2.65 2.76 2.38

72 47 51 46

a Samples were stored in a sealed container with a saturated H2O atmosphere. b Measured minimum of endothermic transition observed for H2O loss by DSC.

(TEM) and by retention of the low-angle (100) reflection in the powder X-ray diffraction patterns. Previous reports have described the characterization of Ti-SBA15 catalysts by diffuse reflectance UV visible (DRUV-vis) spectroscopy, which provides information regarding the structure of the Ti(IV) sites.14,15 It is generally accepted that absorption maxima (λmax) of 210-250 nm are indicative of the ligand-to-metal charge transfer (LMCT) band for site-isolated, 4-coordinate Ti(IV) sites.44 Typically, the LMCT band shifts toward lower energy values of 240-270 nm as water coordinates to the site-isolated Ti(IV) sites to give (SiO)4-nTi(OH)n(OH2)x species.15,44 As illustrated in Figure 2, there is little difference between the UV-vis spectra for the unmodified and modified Ti-SBA15 catalysts, taken in ambient conditions, with λmax values of 245 and 250 nm, respectively. This, along with the extremely low titanium weight loading, is consistent with site-isolated Ti(IV) centers and suggests that the coordination number of the Ti(IV) species is not affected by surface modification. Hydrophobicity of Ti-SBA15 Catalysts. To analyze the hygroscopic nature of the materials, they were kept in a sealed container with a saturated H2O atmosphere for 24 h and then directly examined by TGA (Table 2).30 The unmodified material exhibited a wt % H2O loss of 32% below 150 °C. The modified materials, however, exhibited very little H2O loss (2.4-2.8%) in this temperature range. The water desorption temperature, determined by differential scanning calorimetry (DSC), also decreased upon surface modification (Figure 3). The desorption temperature (measured as the minimum of the endothermic transitions corresponding to H2O loss by DSC) decreased from 72 °C for the unmodified material to 46-52 °C for the surface-modified materials. The H2O adsorption isotherms for BucapTi2SBA15 and Ti2SBA15 are shown in Figure 4. The adsorption volume (43) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area, and Porosity, 2nd ed.; Academic Press: London, 1982. (44) Marchese, L.; Maschmeyer, T.; Gianotti, E.; Coluccia, S.; Thomas, J. M. J. Phys. Chem. B 1997, 101, 8836.

Modified Ti-SBA15 Catalysts for Epoxidation

Langmuir, Vol. 21, No. 21, 2005 9579

Figure 3. DSC traces with minima of endothermic transitions corresponding to H2O desorption marked (+) at 72 °C (Ti1SBA15), 51 °C (BucapTi1SBA15), 47 °C (MecapTi1SBA15), and 46 °C (OccapTi1SBA15). Figure 6. FTIR difference spectrum before and after loading H2O2 onto Ti2SBA15(a), and difference spectrum upon 355 nm photolysis for 10 min (b).

Figure 4. H2O adsorption isotherms for Ti2SBA15 ([) and BucapTi2SBA15 (9).

Figure 5. DRUV-vis difference spectrum of the LMCT absorption for Ti2SBA15/H2O2 obtained for the dehydrated Ti2SBA15 pellet before and after addition of H2O2.

for the surface-modified material led to an order of magnitude less water adsorption at high relative pressures (P/Po), further confirming the hydrophobicity of the modified samples.45 In Situ DRUV-Vis, FTIR, and Photochemical Investigations. Upon addition of one drop of aqueous H2O2 (30%) to a neat pellet of Ti2SBA15, a strong yellow color appeared. The DRUV-vis difference spectrum obtained for the dehydrated pellet before and after addition of H2O2 (followed by dehydration under vacuum) is given in Figure 5. Clearly, the yellow color is due to the broad absorption band ranging from the UV (λmax ≈ 350 nm) to 550 nm. No band was observed upon addition of neat H2O (45) Sing, K. S. W. Pure Appl. Chem. 1985, 57, 603.

to Ti2SBA15 following the same procedure, or upon addition of H2O2 to titanium-free SBA-15. The band is thus associated with both Ti(IV) and H2O2 being present in the sample, which is a strong indication that titaniumhydroperoxo species (Ti-OOH) are formed in Ti2SBA15 upon addition of H2O2.46,47 The broad band observed at λmax ≈ 350 nm is thus assigned to the LMCT transition for Ti-OOH.47 Irradiation of a H2O2-loaded Ti2SBA15 pellet with 355 nm light from a pulsed laser (10 Hz) at an incident energy of 8.6 mJ cm-2 pulse-1 resulted in almost complete bleaching of the yellow band in 10 min by photodecomposition of Ti-OOH to Ti-OH and 1/2O2. This is in good agreement with the photolysis rate of Ti-OOH species observed in TS147 and thus supports assignment of the UV-vis band to Ti-OOH species formed in Ti2SBA15. Additionally, Fourier transform infrared (FTIR) spectroscopy was used to identify the OH stretch of the Ti-OOH moieties in Ti2SBA15. Upon addition of one drop of aqueous H2O2 onto a dehydrated sample of Ti2SBA15 followed by prolonged evacuation (12 h at 10 mTorr) to remove physisorbed H2O2 and H2O, a residual broad band centered at ca. 3460 cm-1 was observed by FTIR spectroscopy (Figure 6a). Irradiating the sample with 355 nm laser light (incident energy of 8.6 mJ cm-2 pulse-1) for 10 min resulted in partial absorbance loss of this IR band. Specifically, a negative difference band centered at 3350 cm-1 was observed to appearat the same rate as that of the disappearance of the LMCT band observed in the UV-vis experiments (Figure 6b). Thus, most likely, the observed difference band centered at 3350 cm-1 is due to Ti-OOH hydroperoxo species formed in Ti2SBA15 upon addition of H2O2. This is in good agreement with the observation by Lin and Frei of a similar IR band in H2O2loaded TS1, which they assigned to a H-bonded OH stretch originating from a Ti-OOH hydroperoxo species.47 In this system, however, the band at 3460 cm-1 formed upon addition of H2O2 to Ti2SBA15 could only be partially removed upon irradiation, which indicates that the band originates only partially from Ti-OOH hydroperoxo species. The residual band remaining after irradiation is thus probably due to hydrogen-bonded OH species (e.g., Si-OH groups) formed upon addition of aqueous H2O2 to (46) Fujiwara, M.; Wessel, H.; Hyung-Suh, P.; Roesky, H. W. Tetrahedron 2002, 58, 239. (47) Lin, W.; Frei, H. J. Am. Chem. Soc. 2002, 124, 9292.

9580

Langmuir, Vol. 21, No. 21, 2005

Brutchey et al.

Table 3. Ti-SBA15-Catalyzed Oxidation of Cyclohexene with H2O2a,b catalyst

selectivity for cyclohexene oxide (%)

selectivity for cyclohexenol (%)

selectivity for cyclohexenone (%)

total yield based on H2O2 (%)

Ti2SBA15 Ti2SBA15c MecapTi2SBA15 MecapTi2SBA15c BucapTi2SBA15 OccapTi2SBA15

11.6 >98.0 29.4 >98.0 44.5 50.9

37.7 0 27.2 0 22.2 17.2

50.7 0 43.4 0 33.3 31.9

13.2 67.0 23.1 59.0 24.3 20.8

19.2 30.3 20.0 58.2

30.8 22.3 27.0 15.4

50.0 47.4 53.0 26.4

5.8 9.4 15.3 6.3

Ti1SBA15 MecapTi1SBA15 BucapTi1SBA15 OccapTi1SBA15

a At 65 °C with a reaction time of 2 h. b Substrate/catalyst (wt) ) 57.9 and substrate/oxidant (mol) ) 4.5; CH CN solvent. c A CHP oxidant 3 and a toluene solvent.

Ti2SBA15.48 This band cannot be assigned to Ti-OOH hydroperoxo moieties because the UV-vis LMCT band is no longer present after irradiation. Additionally, the negative band formed at 3745 cm-1 upon addition of aqueous H2O2 to Ti2SBA15 is due to a partial loss in free silanol groups, which thereby become hydrogen bonded and contribute to the broad peak around 3450 cm-1 (Figure 6a).48 Similarly, the MecapTi2SBA15 catalyst also forms Ti-OOH moieties upon addition of H2O2 (by observation of the LMCT band in UV-vis and the ν(OH) band at 3400 cm-1) which photodecompose after irradiation at 355 nm, as seen by the disappearance of the LMCT band and the evolution of a negative difference band at 3350 cm-1. This suggests that both samples form similar Ti-OOH hydroperoxo moieties in the presence of H2O2. Catalytic Oxidation of Cyclohexene. To compare the activities for the heterogeneous catalysts described above, results were standardized with respect to moles of Ti(IV) (see Table 3). In control experiments with a catalyst and no oxidant or with an oxidant and no catalyst, no cyclohexene oxidation products were observed by gas chromatography (GC) analysis. To test for leaching, catalysts were stirred in acetonitrile and cyclohexene for 1 h at 65 °C before addition of H2O2. The solution was then hot-filtered after 5 min and stirred for 1 h at 65 °C. There was no significant oxidation of cyclohexene after the hot filtration, suggesting that leaching of catalytically active, soluble titanium species was limited under these experimental conditions. Additionally, the reaction mixture was hot-filtered after 2 h and extracted with 0.10 M HCl. No titanium was detected by inductively coupled plasma atomic emission spectroscopy (ICPAES) analysis of the aqueous phase within the detection limit (250 °C. The resultant RuncapTi-SBA15 materials had comparable surface coverages (ca. 65-70%) and similar hydrophobicities (H2O loss by TGA). The selectivities for cyclohexene oxide formation using the MeuncapTi2SBA15 and BuuncapTi2SBA15 catalysts modified in this way were 10.4 and 11.9%, respectively. This is essentially identical to the selectivity achieved using the unmodified (49) Fraile, J. M.; Garcı´a, J. I.; Mayoral, J. A.; Vispe, E. J. Catal. 2000, 189, 40.

Figure 8. Effects of surface modification of Ti2SBA15 on cyclohexene oxidation selectivities after 2 h at 65 °C: (9) cyclohexenone/(cyclohexenol + cyclohexenone).

Ti2SBA15 catalyst with H2O2 (11.6%) and less selective than the corresponding silylated (SiOsurface)3Ti(OSiR3) catalysts (30-45% selectivity). This suggests that selective cyclohexene oxide formation results not only from the hydrophobic nature of the surface but also from the presence of a capped titanol active site, (SiOsurface)3Ti(OSiR3). Further silylation of the (SiOsurface)3TiOH centers of BuuncapTi2SBA15 yields a catalyst that is 66% more selective for cyclohexene epoxidation with H2O2 than the catalyst with only uncapped Ti-OH sites. On the basis of recent theoretical predictions for the epoxidation mechanism of olefins with Ti(IV)/SiO2 catalysts, an electron-poor Ti(IV) center would increase the electrophilicity of the activated oxygen in the Ti-OOH hydroperoxy complex, thereby making the catalyst a more efficient oxygen-transfer agent.50-52 The Ti(IV) center of a (SiOsurface)3Ti(OSiR3) active site would be more electropositive than the Ti(IV) center of a (SiOsurface)3TiOH active site, which may explain why the capped titanol is more selective for cyclohexene oxide formation (vs allylic oxidation). This hypothesis is consistent with the high selectivity and activity achieved with uncalcined catalysts produced by grafting 1,14 2,15 and Ti[OSiPh3]453 onto silica surfaces for the epoxidation of cyclohexene with organic peroxides. The structure of the Ti(IV) active sites for these systems was postulated to be a 4-coordinate (SiOsurface)4-nTi(OSiR3)n geometry. This also corroborates a recent interpretation of catalytic and spectroscopic data by Ohtani and co-workers for 1-octene epoxidation with aqueous hydrogen peroxide.54 On the basis of X-ray absorption data, they propose the active site structure for a selective epoxidation catalyst to be a capped titanol, i.e., (SiOsurface)3Ti(OSiR3), with the uncapped titanol being less selective. For all types of Ti-SBA15 catalysts, use of an organic peroxide such as cumene hydroperoxide (CHP), rather than H2O2, in toluene at 65 °C gave much higher yields and selectivities for epoxidation, as has been previously observed.14,15 Use of CHP as an oxidant gave a 59% yield of cyclohexene oxide (MecapTi2SBA15 catalyst, 2 h), with a selectivity of 98%. With unmodified Ti2SBA15, a higher yield of cyclohexene oxide (67%) was obtained, but with (50) Sinclair, P. E.; Catlow, R. A. J. Phys. Chem. B 1999, 103, 1084. (51) Wells, D. H.; Delgass, W. N.; Thomson, K. T. J. Am. Chem. Soc. 2004, 126, 2956. (52) Urakawa, A.; Bu¨rgi, T.; Skrabal, P.; Bangerter, F.; Baiker, A. J. Phys. Chem. B 2005, 109, 2212. (53) Attfield, M. P.; Sankar, G.; Thomas, J. M. Catal. Lett. 2000, 70, 155. (54) Ikeue, K.; Ikeda, S.; Watanabe, A.; Ohtani, B. Phys. Chem. Chem. Phys. 2004, 6, 2523.

9582

Langmuir, Vol. 21, No. 21, 2005

the same selectivity. It appears that surface modification has no beneficial effect on cyclohexene oxidation using CHP as the oxidant in nonpolar media (toluene). Baiker and co-workers have previously observed this effect as well.28

Brutchey et al. Scheme 2. Proposed Mechanism for Cyclohexene Epoxidation with H2O2 and RcapTiSBA15 Catalysts

Conclusions Modification of TMP-derived Ti-SBA15 catalysts with alkylsilyl groups yielded hydrophobic surfaces that made the catalysts tolerant to aqueous H2O2. For the Ti2SBA15 catalysts, the yield of 13.2% was increased by a factor of 1.8 over 2 h (24.3% yield of oxidation products) upon surface modification. Similarly, for the Ti1SBA15 catalysts, surface modification resulted in a 2.6 times increase in yield over 2 h (15.3% yield of oxidation products). The observed enhancement in activity is similar to that reported by Rhee and Bu (2.6 times increase in yield over 4 h)33 as well as by Tatsumi and co-workers (18.5 times increase in yield over 3 h)23 using surface-modified Ti-MCM41 catalysts for cyclohexene epoxidation with H2O2. The highest selectivity for cyclohexene oxide formation with a surface-modified catalyst, previously reported by Rhee and Bu, was 40% with the ring-opened product, 1,2-cyclohexanediol, also being formed with 17% selectivity. Here, the cyclohexene oxide selectivity achieved by surface modification of the TMP-derived catalyst is greater than that previously reported for mesoporous Ti(IV)/SiO2 catalysts with aqueous H2O2. Silylation of the Ti-SBA15 catalysts with -OSiMe2Oc groups yields a catalyst that is almost 60% selective for cyclohexene oxide formation, with no 1,2-cyclohexanediol formation. The formation of 1,2-cyclohexanediol results from the acidcatalyzed ring opening of cyclohexene oxide by water. The hydrophobic surface properties repel water from the active site, however, and the number of acidic Si-OH/Ti-OH sites are reduced by silylation. Thus, the lack of 1,2cyclohexanediol formation is consistent with the surface properties of the modified catalysts.6 A recent report by Ohtani and co-workers addressed the epoxidation of 1-octene with a surface-modified Ti(IV)/SiO2 catalyst and aqueous H2O2.54 They suggest that the most selective active sites for epoxide formation have the capped titanol structure (SiOsurface)3Ti(OSiR3) on the basis of X-ray absorption data. This is consistent with our results on the selectivity of cyclohexene oxide formation using hydrophobic catalysts with capped and uncapped titanol groups. We observed that catalysts with uncapped Ti-OH groups were less selective for epoxide formation regardless of whether the neighboring Si-OH groups were silylated or not. Thus, the hydrophobic surface does not directly cause an increase in selectivity for epoxidation. Furthermore, it was demonstrated by in situ FTIR and DRUV-vis spectroscopies that both the capped and uncapped Ti-OH sites likely react to form Ti-OOH upon addition of H2O2. If both (SiOsurface)3TiOH and (SiOsurface)3Ti(OSiR3) active sites go through a hydroperoxy intermediate (instead of different structured intermediates such as Ti-OO-Si for the surface-modified catalyst),47 it seems reasonable that the electronic effects of surface modification may cause the difference in catalytic selectivity. The Ti(IV) in a (SiOsurface)3Ti(OSiR3) active site is more electron deficient than that in a (SiOsurface)3TiOH site, which may render an electrophilic oxygen-transfer reaction from the Ti-OOH complex to the olefin more active (Scheme 2), as suggested by theoretical predictions.50-52 A more active oxygentransfer reaction would increase the amount of epoxide formed relative to allylic oxidation, which was demon-

strated to go through a radical mechanism. Once oxygen transfer is complete and the epoxide product is formed, the remaining titanol and coordinated silanol likely condense to release H2O, as has been experimentally observed for TS1.47 Experimental Section General Procedures. All manipulations were conducted under a nitrogen atmosphere using standard Schlenk techniques or in a Vacuum Atmospheres drybox, unless otherwise noted. Acetonitrile (99.9%) was purchased from Riedel-de Hae¨n and used without further purification. Cyclohexene was purchased from Aldrich and distilled prior to use. CHP (80%) and H2O2 (30%) were purchased from Aldrich and used as received. (N,NDimethylamino)trimethylsilane, (N,N-dimethylamino)butyldimethylsilane, and (N,N-dimethylamino)octyldimethylsilane were purchased from Gelest, Inc. and used without further purification. Ti[OSi(OtBu)3]4,37 [(tBuO)2Ti{µ-O2Si[OSi(OtBu)3]2}]2,15 and SBA1538 were prepared as reported in the literature. Ti-SBA15 Preparation. The SBA-15 was dried at 130 °C in vacuo for 15 h and thereafter handled under a nitrogen atmosphere. A 0.5 g sample of SBA-15 was suspended in pentane (25 mL). A pentane solution (30 mL) of Ti[OSi(OtBu)3]4 or [(tBuO)2Ti{µ-O2Si[OSi(OtBu)3]2}]2 was prepared, the concentration of which depended on the desired titanium loading. The pentane solution of the titanium siloxide complex was then added to the stirred suspension of SBA-15 (25 °C). The resulting mixture was stirred for 20 h and then filtered and washed with pentane (3 × 20 mL). The grafted material was dried for 2-3 h in vacuo before calcination to 300 °C (10 °C min-1) under flowing nitrogen. Preparation of Surface-Modified Ti-SBA15 Materials. To a suspension of 0.250 g of the Ti-SBA15 catalyst in 20 mL of hexane was added a 10 mL solution of RMe2Si(NMe2)2 in hexane (5 equiv based on surface [OH]) at room temperature via cannula. The suspension was stirred at room temperature for 20 h under flowing nitrogen. The solid was then filtered on a Bu¨chner funnel and rinsed with hexane (3 × 10 mL). The solid was dried in vacuo at 120 °C for 12 h, and the resulting white solid was stored in a drybox. Catalysis Procedure. A sample of catalyst (ca. 0.035 g) was added to a 50 mL round-bottom flask that was fitted with a reflux condenser and a septum. Acetonitrile (5.0 mL) and cyclohexene (2.5 mL) were added by syringe through the septum under a flow of nitrogen. Toluene (23 µL) was added as an internal standard. The mixture was allowed to equilibrate at the reaction temperature of 65 °C for 10 min. H2O2 (0.62 mL) was added by syringe to the rapidly stirring solution. Aliquots (ca. 0.08 mL) were removed from the reaction mixture by syringe after 5, 30, 60, 90, and 120 min and then filtered and cooled. The filtrate was analyzed by gas chromatography (GC), and assignments were made by comparison with authentic samples analyzed under the same conditions.

Modified Ti-SBA15 Catalysts for Epoxidation Characterization. Elemental analyses were performed by the College of Chemistry microanalytical laboratory at the University of California, Berkeley, or by Galbraith Laboratories. Powder X-ray diffraction (PXRD) experiments were performed on a Siemens D5000 X-ray diffractometer using Cu KR radiation. DRUV-vis spectra were acquired using a Perkin-Elmer Lambda-9 spectrophotometer equipped with a 60 mm integrating sphere (ratio of apertures to sphere surface ) ca. 8%), a slit width of 4 nm, and a collection speed of 120 nm min-1. Samples were run using BaSO4 as the reference material. Nitrogen adsorption isotherms were performed on a Quantachrome Autosorb 1 surface area analyzer, and samples were outgassed at 120 °C for at least 15 h prior to measurement. Water adsorption measurements were performed by Quantachrome, Inc. on a Hydrosorb 1000 water vapor sorption analyzer. Thermal analyses were performed on a TA Instruments SDT 2960 Integrated TGA/ DSC analyzer with a heating rate of 10 °C min-1 under a flow of nitrogen or oxygen. Calcinations were performed using a Lindberg 1200 °C three-zone furnace with a heating rate of 10 °C min-1 under a flow of nitrogen, and the temperature was held constant for 4 h. GC analyses were performed with an HP 6890 GC system using a methylsiloxane capillary (50.0 m × 320 µm × 1.05 µm nominal), and integration was performed relative to toluene. In situ DRUV-vis and FTIR measurements were performed by mounting a self-sustaining pellet of TiSBA15 (10 mg, 10 mm diameter) in a stainless steel cell equipped with quartz

Langmuir, Vol. 21, No. 21, 2005 9583 (UV-vis) or KCl windows (FTIR). Hydrogen peroxide (30%) was loaded by placing one drop onto the mounted pellet and evacuating the cell (10 mTorr) for 12 h. Photolysis at 355 nm was conducted with emission of a pulsed Coherent Nd:YAG laser, model Infinity, operated at 10 Hz. The beam intensity was 8.6 mJ cm-2 pulse-1. FTIR spectra were recorded at 2 cm-1 resolution using a Bruker IFS 66v/S spectrometer. DRUV-vis spectra were recorded on a Shimadzu UV-2100 spectrometer equipped with a Shimadzu ISR260 integrating sphere.

Acknowledgment. This work was supported by the Director, Office of Energy Research, Office of Basic Energy Sciences, Chemical Sciences Division, of the U.S. Department of Energy under Contract DE-AC03-76SF00098. R.L.B. thanks the National Science Foundation for support with an NSF Graduate Fellowship. We are extremely grateful to H. Frei at Lawrence Berkeley National Laboratory for invaluable discussions regarding the in situ FTIR/DRUV-vis experiments and for use of his spectrometers. We also thank A. M. Stacy at the University of California, Berkeley, for use of instrumentation (PXRD, DRUV-vis). LA051182J