Enhancement of Epoxidation Efficiencies for Ta-SBA15 Catalysts. The

Apr 25, 2011 - The Influence of Modification with –EMe3 (E = Si, Ge, Sn) Groups ...... Pascal Guillo , T. Don Tilley , Michael J. Janik , and Robert...
0 downloads 0 Views 1MB Size
ARTICLE pubs.acs.org/Langmuir

Enhancement of Epoxidation Efficiencies for Ta-SBA15 Catalysts. The Influence of Modification with EMe3 (E = Si, Ge, Sn) Groups Paul J. Cordeiro†,§ and T. Don Tilley*,‡,§ †

Department of Chemical and Biomolecular Engineering Department of Chemistry University of California, Berkeley, Berkeley, California 94720, United States § Chemical Sciences Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, California 94720, United States ‡

bS Supporting Information ABSTRACT: Site-isolated Ta(V) centers were introduced onto the surface of a mesoporous SBA-15 support via the thermolytic molecular precursor method. After thermal treatment under oxygen, the resulting Si OH and Ta OH sites of TaSBA15-O2 were modified with a series of trimethyl group 14 species, Me3E , by treatment with Me3E NMe2 (E = Si, Ge, Sn) reagents. The resulting surface-modified catalysts (Me3E)capTaSBA15 exhibit a significantly increased rate of cyclohexene epoxidation with H2O2 as an oxidant, and provided a decreased amount of allylic oxidation products with respect to the unmodified material, TaSBA15-O2. The rate of nonproductive H2O2 decomposition, as monitored via 1H NMR spectroscopy, significantly decreased after the surface modification. The structure of the TaSBA15 catalysts and potential Ta(V) epoxidation intermediates (formed upon treatment of Ta(V) materials with H2O2) were probed using UV visible absorbance and diffuse-reflectance UV visible spectroscopy. A Ta(V)(η2-O2) intermediate species is proposed for the TaSBA15-O2, (Me3Si)capTaSBA15, and (Me3Ge)capTaSBA15 catalysts, while intermediate species for the (Me3Sn)capTaSBA15 catalysts could not be characterized.

’ INTRODUCTION In recent years, numerous supported transition metal oxide catalysts have been studied for direct functionalization of inexpensive and readily available alkanes and alkenes into more synthetically useful molecules such as aldehydes, ketones, and epoxides.1 5 Epoxides are important chemical intermediates in the production of fine chemicals, pharmaceuticals, and pesticides, and products such as propylene oxide are used in polyether polymer synthesis.6 8 Large-scale industrial production of epoxides in the liquid phase often involves chlorinated solvents or coreactants, and produces equimolar amounts of coproducts that can be of limited industrial use, are difficult to recycle, and/or are environmentally harmful. Thus, catalysts that efficiently utilize green oxidants such as aqueous hydrogen peroxide (H2O2) are more desirable, since the byproduct of the catalytic reaction (H2O) is environmentally benign. A well-established method for the preparation of isolated, single-site transition metal catalysts is based on the thermolytic molecular precursor (TMP) method,9 which employs metal containing species rich in silicon and oxygen (in the form of siloxy and alkoxy ligands) to deliver isolated active sites onto mesoporous supports under mild conditions. These precursor molecules react with surface silanol (Si OH) groups to eliminate silanol or alcohol molecules and anchor the metal center to the surface. The resulting materials may then be treated in oxygen r 2011 American Chemical Society

at mild temperatures ( Sn > Si. More recently, Liu et al. reported enhanced activity in the epoxidation of propylene with H2 and O2 using a modified microporous TS-1 framework incorporating Ge(IV). Theoretical calculations suggested that the presence of germanium in close proximity to titanium enhanced the electrophilicity of the active metal peroxo species due to an apparent electron-withdrawing effect of the germanium.33 Heterogeneous group 5 metal catalysts are known to be active for a variety of oxidation reactions.38 50 However, few studies have focused on supported tantalum(V) materials as alkene epoxidation catalysts.14,39,51 54 Previous work in our laboratory explored surface modification of supported transition metal catalysts for the oxidation of alkenes with aqueous H2O2.52,55,56 Catalysts modified with OSiR3 groups were found to be highly selective in the epoxidation of cyclohexene, possibly due to the hydrogen bonding interaction of a trialkylsilanol group with the proposed metal-peroxo or metal-hydroperoxo intermediate (Scheme 1). Along these lines, hydrogen bonding interactions involving metal-peroxo species have been calculated to be critical

for oxidation reactions in fluorinated solvents and in metal-based oxidation reactions.57,58 Thus, substituent effects on such hydrogen-bonding interactions are worthy of further examination. In this study, previously described site-isolated Ta(V) centers were introduced onto the surface of mesoporous SBA15 via the thermolytic molecular precursor method.37,52,56 The resulting materials were characterized by diffuse reflectance UV visible spectroscopy, thermogravimetric analysis, and hydroxyl group titration to probe the local environment of the Ta(V) species. After treatment in oxygen, the resulting Si OH and Ta OH sites were modified with a series of group 14 trimethyl-amines to introduce EMe3 (E = Si, Ge, Sn) groups onto the catalyst surface. The performance of these catalysts in the epoxidation of cyclohexene with aqueous hydrogen peroxide was also evaluated.

’ RESULTS Synthesis and Characterization of Materials. Mesoporous SBA-15 silica was synthesized according to the literature procedure.59 TaSBA15 was prepared by methods previously described, by grafting Ta(OiPr)2[OSi(OtBu)3]3 (1) onto the silica surface to yield TaSBA15.14 This material was treated in oxygen at 300 °C to generate TaSBA15-O2, which presumably contains Ta OH sites suitable for functionalization, though these were not detectable by infrared spectroscopy (Scheme 2).52 After synthesis and drying, TaSBA15-O2 was treated with an excess of (N,N-dimethylamino)trimethylsilane, -germane, or -tin, in hexanes at room temperature, to titrate surface Si OH and Ta OH sites in a manner analogous to that previously reported for (RMe2Si)capMSBA15 (R = Me, Bu, Oc; M = Ti, Ta) catalysts (Scheme 2).37,52,55 The resulting materials were filtered 6296

dx.doi.org/10.1021/la200090u |Langmuir 2011, 27, 6295–6304

Langmuir

ARTICLE

Scheme 2. Synthesis of TaSBA15-O2, and Surface Modification of TaSBA15-O2 to Yield (Me3E)capTaSBA15 (E = Si, Ge, Sn)

Table 1. Elemental Analysis, Surface Modification, and Nitrogen Physisorption Data for Dispersed Ta(V) Materials (E = Si, Ge, Sn) Ta content a

material

[wt %]

SBA15

-

TaSBA15-O2 (Me3Si)capTaSBA15

1.61 1.37

Ta content 2

[nm ]

E content

Me3E coverage

a

2 b

Me3E coverage 2 c

rpd

SBET 2

1

[wt %]

[nm ]

[nm ]

[m g ]

[nm]

-

-

-

-

800

3.3

0.07 0.06

-e

1.4

1.7

610 400

3.3 2.8

(Me3Ge)capTaSBA15

1.32

0.06

(Me3Sn)capTaSBA15

1.25

0.06

9.19 13.9

1.3

1.5

380

2.8

1.2

1.3

280

2.8

a

Determined by inductively coupled plasma (ICP) methods. b Determined by carbon elemental analysis and titration with Mg(CH2Ph)2 3 2THF. Determined by thermogravimetric analysis. d Pore radius determined from BJH adsorption. e Not determined due to the large Si content in the mesoporous support SBA15. c

via cannula and dried under vacuum at 120 °C to yield the corresponding (Me3E)capTaSBA15 (E = Si, Ge, Sn) catalysts. Solution-state 1H NMR spectra of the filtrate exhibited methyl proton resonances for HNMe2, and infrared spectra of solid samples contained new vibrations corresponding to CH3 groups (Supporting Information, Figure S1). The amount of tantalum present in each sample was determined by inductively coupled plasma optical emission spectroscopy, with a maximum tantalum content of 1.61 wt % Ta found for TaSBA15-O2 (Table 1). Elemental analysis of the hexanes solution after filtration contained no tantalum (by ICP methods). The degree of surface functionalization was determined from elemental analysis, thermogravimetric analysis (TGA), and differences in surface hydroxyl group coverages. The concentration of surface hydroxyl groups was determined by quantification of the amount of toluene evolved after reaction of the catalyst samples with Mg(CH2Ph)2 3 2THF via 1H NMR spectroscopy.60 All surface-modified catalysts exhibited roughly equivalent incorporations of capping groups, of between 1.2 and 1.4 groups per square nanometer (Table 1).

The surface areas and pore structures of the materials were evaluated using nitrogen porosimetry. All samples displayed N2 adsorption desorption data consistent with type IV isotherms (Supporting Information, Figure S2), with narrow pore size distributions characteristic of SBA15 type materials.59 The unmodified TaSBA15-O2 catalyst was found to have a BrunauerEmmet-Teller (BET) surface area of 610 m2 g 1. Upon surface modification, the surface area decreased with the size of the capping group. (Me3Si)capTaSBA15 had a surface area of 400 m2 g 1, and (Me3Ge)capTaSBA15 had a surface area of 380 m2 g 1, while (Me3Sn)capTaSBA15 had a surface area of 280 m2 g 1. The pore size distribution for all modified (Me3E)capTaSBA15 catalysts was found to be smaller than that of the parent TaSBA15-O2 (Table 1). The well-ordered mesostructures of the TaSBA15 catalysts were preserved, as indicated by transmission electron microscopy (TEM) and by retention of the low angle reflections in the small-angle X-ray scattering (SAXS) patterns (Supporting Information, Figure S3). Figure S1 presents the FTIR spectra of tantalum-containing materials, with unmodified SBA15 and germanium-modified 6297

dx.doi.org/10.1021/la200090u |Langmuir 2011, 27, 6295–6304

Langmuir SBA15, (Me3Ge)capSBA15, shown for comparison. The M O M stretching region of all spectra is dominated by bands assigned to Si O Si stretching modes.14 The band normally assigned to the Ta O Si asymmetric stretch is not observed for all tantalum-containing materials, presumably because of the low tantalum content in the samples, as has been previously observed.14 The spectrum of (Me3Ge)capTaSBA15 exhibits a new band at ca. 1025 cm 1, which is not present in tantalum-containing samples without germanium (TaSBA15-O2) or in germaniumcontaining samples without tantalum ((Me3Ge)capSBA15). Thus, this band is associated with both tantalum and germanium being present in the sample, and its origin could be due to a Ta O Ge group. Diffuse reflectance UV visible (DRUV vis) spectroscopy was used to examine the local structure of the supported Ta(V) catalysts. Although this technique alone cannot unequivocally determine the structure of the Ta(V) species, the molecular precursor used in the TMP method can be used as a spectroscopic model for the grafted materials. The UV visible spectrum of the molecular precursor (1) exhibits a ligand-to-metal charge transfer band (OfTa) that is centered at 220 nm and extends to 250 nm in the solution state (Supporting Information, Figure S4, trace a). In the solid state, the absorption band (for 1) broadens and shifts to 230 nm and extends to 280 nm (Supporting Information, Figure S4, trace b). This observation has been rationalized as being due to a difference in structure for the molecule in the solution and solid states.14,56 The tantalum centers exist in a distorted octahedral environment in the solid state, and possibly in a pentacoordinate environment in solution.14 As illustrated in Figure 1, the DRUV vis spectra of TaSBA15-O2, (Me3Si)capTaSBA15, and (Me3Ge)capTaSBA15 are dominated by an LMCT band centered near 210 nm, similar to that of the molecular precursor (1) in the solution state. The spectrum for (Me3Ge)capTaSBA15 (Figure 1B, trace a) displays a broader LMCT band and an additional band centered at 310 nm, which is also present in a sample of SBA15 treated with Me3Ge NMe2 (Supporting Information, Figure S5). The spectrum for (Me3Sn)capTaSBA15 is dominated by an absorption band centered at 204 nm due to the presence of tin in the sample. This band overlaps with the OfTa LMCT band that would normally be visible for this sample, preventing its identification. Elemental analysis, however, confirms that tantalum is still present in the sample. It should be noted that pure SBA15 functionalized with EMe3 groups (E = Si, Ge, Sn) has absorbance bands between 205 and 270 nm (Supporting Information, Figure S5), attributable to the EMe3 (E = Si, Ge, Sn) group. As expected, these bands are retained in the catalytic tantalum-containing materials. Hydrophobicity of Surface-Modified Materials. To study the hygroscopic nature of the materials, samples were placed in a sealed container with a saturated water atmosphere for 48 h, followed by thermal gravimetric analysis and differential scanning calorimetry analysis (Table 2, Supporting Information, Figure S6). Mass losses below 150 °C are generally assumed to be physisorbed water, since dehydrated samples do not show significant mass loss until temperatures above 250 °C (as observed by thermal gravimetric analysis),61 corresponding to the loss of trimethylsilyl, -germyl, or -stannyl groups. Elemental analysis of samples treated with anhydrous media or aqueous H2O2 demonstrated that no tantalum, silicon, germanium, or tin is removed by these treatments. The water desorption temperature was taken as the minimum of the endothermic transition corresponding to water loss. The unmodified material

ARTICLE

Figure 1. (A) DRUV vis spectra for (a) TaSBA15-O2 and (b) (Me3Si)capTaSBA15. (B) DRUV vis spectra for (a) (Me3Ge)capTaSBA15 and (b) (Me3Sn)capTaSBA15.

Table 2. Thermal Analysis (to 150 °C) of Dispersed Ta(V) Catalysts after Hydration for 48 ha material

H2O content

desorption temperature

(wt %)a

(°C)b

SBA15

39.8

65

TaSBA15-O2

34.3

86

(Me3Si)capTaSBA15

1.7

45

(Me3Ge)capTaSBA15

20.4

68

(Me3Sn)capTaSBA15

29.9

86

a

Samples were stored in a sealed container with a saturated water environment. b The minimum of endothermic transition for water loss determined by DSC.

exhibited a mass loss of 34.3 wt %, with a desorption temperature of 86 °C, corresponding to physisorbed water. The (Me3Si)capTaSBA15 sample exhibited little mass loss below 150 °C (1.7 wt %), and significantly decreased water desorption temperature (45 °C), confirming the hydrophobic nature of the material after modification. The (Me3Ge)capTaSBA15 sample 6298

dx.doi.org/10.1021/la200090u |Langmuir 2011, 27, 6295–6304

Langmuir

ARTICLE

Table 3. Percent H2O2 Remaining As a Function of Time for Dispersed Ta(V) Catalystsa H2O2 remaining material

2 h [%]

SBA15

>99

6 h [%]a >99

9 h [%]a >99

78.8

47.7

26.6b

(Me3Si)capTaSBA15 (Me3Ge)capTaSBA15

>99 97.3

99.1 92.8

98.1 89.0

(Me3Sn)capTaSBA15

90.6

78.0

66.7b

TaSBA15-O2

a

a

1

Determined by H NMR spectroscopy with CH2Cl2 internal standard as a reference. b Taken at 9.5 h reaction time.

was found to be much less hydrophobic than (Me3Si)capTaSBA15, with 20.4 wt % loss below 150 °C, but with a significantly lower water desorption temperature (68 °C) than TaSBA15-O2. The (Me3Sn)capTaSBA15 sample displayed a large loss in mass (29.9 wt %), and a water desorption temperature of 86 °C, nearly identical to that of the unmodified TaSBA15-O2 sample. Hydrogen Peroxide Decomposition Reactions. The modified (Me3E)capTaSBA15 (E = Si, Ge, Sn) materials are active epoxidation catalysts with hydrogen peroxide as an oxidant. Since several transition metals (e.g., V, Ti, Mn, Fe, Cu) are known to decompose hydrogen peroxide into water and oxygen gas at elevated temperatures, determination of the efficiency of oxidant conversion during epoxidation is of great interest.62 Solution state 1H NMR spectroscopy can be employed to monitor low concentrations of hydrogen peroxide. Reactions involving no catalyst, bare SBA15, or surface modified (Me3E)capSBA15 (E = Si, Ge, Sn) displayed little decomposition of hydrogen peroxide (