Solvent-Free Oxidation of Primary Carbon-Hydrogen Bonds in Toluene Using Au-Pd Alloy Nanoparticles

See allHide authors and affiliations

Science  14 Jan 2011:
Vol. 331, Issue 6014, pp. 195-199
DOI: 10.1126/science.1198458


Selective oxidation of primary carbon-hydrogen bonds with oxygen is of crucial importance for the sustainable exploitation of available feedstocks. To date, heterogeneous catalysts have either shown low activity and/or selectivity or have required activated oxygen donors. We report here that supported gold-palladium (Au-Pd) nanoparticles on carbon or TiO2 are active for the oxidation of the primary carbon-hydrogen bonds in toluene and related molecules, giving high selectivities to benzyl benzoate under mild solvent-free conditions. Differences between the catalytic activity of the Au-Pd nanoparticles on carbon and TiO2 supports are rationalized in terms of the particle/support wetting behavior and the availability of exposed corner/edge sites.

Selective oxidation of primary carbon-hydrogen bonds is of crucial importance in activating raw materials to form intermediates and final products for use in the chemical, pharmaceutical, and agricultural business sectors (1). One class of raw materials is alkyl aromatics; toluene, for example, the simplest member of this class, can be oxidized to benzyl alcohol, benzaldehyde, benzoic acid, and benzyl benzoate. These products are commercially significant as versatile intermediates in the manufacture of pharmaceuticals, dyes, solvents, perfumes, plasticizers, dyestuffs, preservatives, and flame retardants. Commercially, benzaldehyde is produced by the chlorination of toluene followed by saponification (2), and benzoic acid is produced by the liquid-phase cobalt-catalyzed reaction of toluene using oxygen at 165°C with acetic acid as solvent, but the conversion has to be limited to <15% to retain high selectivities (39). The use of halogens and acidic solvents makes these processes environmentally unfriendly. Vapor-phase oxidation has been considered, but the conversion must be limited to avoid overoxidation to CO2 and other byproducts (10). Attempts to overcome these problems have prompted investigation of the use of supercritical CO2 and ionic liquids, but these unfortunately resulted in low conversions (11, 12).

Often, heterogeneous catalysts are preferred over homogeneous catalysts, because these materials can be readily separated from the reaction mixture. Heterogeneous catalysts can also be readily used in flow reactors, facilitating the efficient production of materials using continuous processes. For the oxidation of toluene, there have been many attempts to find a suitable oxidation catalyst, and to date these have used copper and manganese (1315), cobalt (16), or chromium (17) catalysts, but all of these perform very poorly with turnover numbers (TONs: mole of product per mole of metal catalyst) of less than 100, even at temperatures in excess of 190°C (table S1) (18). There is clearly a need to develop heterogeneous catalysts for toluene oxidation that have greatly improved activity while retaining selectivity.

Recently, we have shown that Au-Pd alloy nanoparticles are very effective for the direct synthesis of hydrogen peroxide (19) and the oxidation of primary alcohols using oxygen (20). This catalyst operates by establishing a reactive hydroperoxy intermediate. Because these intermediates are known to be involved in the enzymatic oxidation of primary carbon-hydrogen bonds (21), we reasoned that it should be feasible for Au-Pd nanoparticles to be active for the oxidation of the primary carbon-hydrogen bonds in toluene. Here we show that Au-Pd alloy nanoparticles prepared by a sol immobilization technique can give significantly improved activity for the oxidation of toluene under mild solvent-free conditions. These catalysts have TONs that are a factor of ~30 greater than those of previous heterogeneous catalysts for this reaction and also display a remarkably high selectivity to benzyl benzoate.

We started by investigating the oxidation of toluene in an autoclave reactor with O2 in the absence of catalyst in order to determine the blank baseline rate. O2 is a di-radical and can initiate homogeneous oxidation processes at elevated temperatures and pressures. We found that such processes become substantial at 190°C under our conditions (fig. S1) (18). This important observation suggests that the earlier studies conducted at 190°C (table S1) (18) may not in fact have been heterogeneously catalyzed. In view of the potential role of O2 di-radicals, the maximum reaction temperature in our studies has been restricted to 160°C; at this temperature, the blank reaction in the absence of catalyst but in the presence of support is negligible for short reaction times (Table 1, entries 1, 3, and 4) and is very low at longer reaction times (Table 1, entry 2). We initially investigated Au-Pd/TiO2 catalysts, prepared by impregnation, because these had previously been shown to be very active for H2O2 synthesis (19) and alcohol oxidation (20); however, these catalysts were not found to be particularly active for toluene oxidation, although they did not produce any CO2 (table S2) (18).

Table 1

Comparison of catalytic activity for the oxidation of toluene in the absence of solvent with O2. Catalyst mass varied between 0.2 and 0.6 g to give a substrate/metal molar ratio of 6500. Toluene, 10 to 20 ml; stirring rate, 1500 rpm; TON calculated on the basis of the total metal. All metal catalysts were prepared using the sol immobilization method (18). T, temperature.

View this table:

The Au-Pd nanoparticles synthesized by the impregnation method can be relatively large (typically >6 nm) and have substantial compositional variations, which limits their reactivity. Therefore, in order to design more effective catalysts, we decided to investigate Au-Pd nanoparticles with smaller median particle sizes (2 to 5 nm) and a more controlled composition. These were prepared by sol immobilization of Au-Pd colloids, with carbon and TiO2 as supports (18), a method that afforded tightly controlled particle size distributions and composition. We investigated the effect of the Au-Pd molar ratio (Table 1, entries 5 to 12) using these sol-immobilized materials supported on carbon. For all catalysts, no CO2 formation was observed; the only products were benzyl alcohol, benzaldehyde, benzoic acid, and benzyl benzoate. By itself, Au was not active for this reaction, but the addition of Pd significantly enhanced the conversion, demonstrating a clear synergistic effect for the Au-Pd catalysts as compared with the monometallic species. A physical mixture of the separate Au/C and Pd/C catalysts showed no enhancement (Table 1, entry 13), highlighting the molecular-scale nature of the synergy. The observed synergy is due to electronic and morphological features because transmission electron microscopy (TEM) shows that the mean particle size of the nanoparticles decreases slightly on the addition of Au to Pd (table S3). When Au-rich catalysts were used, some benzyl alcohol was formed, but this was readily oxidized to the main product benzaldehyde, as this sequential oxidation is rapid at this temperature (20). We consider the initial oxidation of toluene to involve a surface hydroperoxy intermediate formed from the interaction of the metal with oxygen and toluene. However, as the fraction of Pd in the alloy was increased, the selectivity to benzyl benzoate became dominant. The optimum catalyst composition found was a 1:2 Au:Pd molar ratio (approximately 1:1 by weight). For this catalyst, the turnover frequency (TOF: mole of product formed per mole of metal per hour) of toluene oxidation after 7 hours of reaction was ~50. We also investigated this Au-Pd composition using a TiO2-supported material (Table 1, entry 14) and found that it was also active, but less so than the carbon-supported material, displaying a TOF of 20 hour−1 after 7 hours of reaction.

To demonstrate the general applicability of the AuPd/C catalyst, we oxidized 2-, 3-, and 4- methoxytoluene (Table 2), and 2-, 3-, and 4-nitrotoluene (table S4) (18). No CO2 was observed, and the reactivity trend (4-methoxy- ~ 2-methoxy- >3-methoxy ~ toluene > 2-nitro- > 3-nitro ~ 4-nitro) is indicative of the involvement of electron-deficient intermediate(s). Additional products formed were identified as a family of esters and C-C coupling products (table S5) (18). We have also investigated the reaction of xylenes; the catalysts are equally effective for the oxidation of this substrate and formed the aldehyde, acid, and esters as products, with the relative amounts being dependent on the conversion (table S6) (18), confirming the wider applicability of our catalysts.

Table 2

Comparison of the catalytic activity for oxidation of substituted toluenes in the absence of solvent with O2. Catalyst mass, 0.4 g; substrate, 20 ml; stirring rate, 1500 rpm; 160°C; 10 bar O2; TOF and TON calculated on the basis of the total metal. Au-Pd/C (1:1.85 Au:Pd molar ratio), 1 wt % total metal was prepared using the sol immobilization method (18).

View this table:

Given these initial promising results, we investigated the oxidation of toluene at longer reaction times (Table 1, entries 16 and 17). In this regime, much higher toluene conversions were attained and no CO2 was observed. Once again only four products were observed: benzyl alcohol, benzaldehyde, and benzoic acid in trace amounts and benzyl benzoate as the dominant product (~95%) for both catalysts. The carbon-supported catalyst was typically twice as active over this longer time scale, exhibiting a TON of 3300 with ~3150 mol of benzyl benzoate per mole of metal produced over the reaction period. We investigated the use of lower reaction temperatures (Table 1, entries 18 though 20) and observed some activity, albeit slight, even at 80°C. In a further set of experiments, we attempted to enhance the activity of the catalyst at 80°C by using t-butyl hydroperoxide as a co-oxidant (table S7) and found that the conversion increased appreciably when both the carbon- and TiO2-supported catalysts were used; TON values of 850 to 1200 can be achieved even under these mild conditions. However, using this hydroperoxide as a co-oxidant generated a different product profile. At long reaction times and high conversions, benzoic acid became the main product, together with appreciable amounts of benzaldehyde and benzyl alcohol; only negligible amounts of benzyl benzoate were formed. As with O2 as the sole oxidant, when using the hydroperoxide co-oxidant, the carbon-supported catalyst was about twice as active as the TiO2-supported Au-Pd nanoparticles.

The reaction profile over an initial 7-hour reaction period was investigated using a higher substrate-to-metal ratio at 160°C (fig. S2). This experiment showed that the toluene conversion increases linearly over this time scale and that the selectivity to benzyl benzoate increases as the selectivity to benzaldehyde decreases. The reaction profile over a longer time scale was also investigated using a lower substrate/metal molar ratio and showed that the conversion continued to increase steadily, fully depleting the toluene after 110 hours (Fig. 1), while the selectivity to benzyl benzoate also increased. Monometallic Au and Pd catalysts showed very low conversion (8%) over 110 hours (table S8) (18). The toluene conversion was also found to increase linearly with catalyst mass (fig. S3). Hence, with appropriate tuning of catalyst loading and reaction conditions, we are confident that complete and selective conversion of toluene to desirable products can be achieved (Table 1, entries 22 to 25 show results of an initial optimization). Furthermore, by increasing the reaction time at 140°C, we can obtain a 95% yield of benzyl benzoate (table S9) (18).

Fig. 1

Toluene conversion and selectivity to benzyl alcohol, benzaldehyde, benzoic acid, and benzyl benzoate: reaction temperature, 160°C (433 K); 0.1 MPa partial pressure of oxygen (PO2); 20 ml of toluene; 0.8 g of catalyst [1 weight % (wt %) Au-Pd/C prepared by sol immobilization with a 1:1.85 Au/Pd ratio]; toluene/metal molar ratio of 3250; and reaction time, 110 hours. Open circles indicate conversion, squares indicate selectivity to benzyl alcohol, diamonds indicate selectivity to benzaldehyde, triangles indicate selectivity to benzoic acid, and solid circles indicate selectivity to benzyl benzoate.

We consider that the high selectivity to benzyl benzoate 6 could result from four possible mechanisms: (i) coupling of the aldehyde 3 and the alcohol 2 to give the hemiacetal 5, followed by oxidation to the ester (Scheme 1); (ii) direct thermal (catalyzed) dehydrative esterification between acid 4 and alcohol 2 (Scheme 1); (iii) Cannizzaro reaction between alcohol 2 and aldehyde 3 via alkoxide 7 (Scheme 2); and (iv) Tishchenko coupling of two aldehydes 3 via “dimer” 8 directly to the ester 6 (Scheme 3).

Recently, Bäumer and co-workers (22) have shown that methyl formate is formed on an Au catalyst when methanol is oxidized with O2 under very dilute conditions; we consider that this precedent favors mechanism (i) in the present case. We performed several control experiments to rule out the other three possible pathways. Heating the oxidized intermediates with O2 in the absence of catalyst gave little ester (table S10), establishing the involvement of the catalyst in steps beyond the initial toluene-to-alcohol oxidation. Heating mixtures of alcohol 2 and acid 4 in the absence of O2, with and without catalyst, did not result in ester formation (tables S11 and S12), ruling out direct esterification (46; Scheme 1). Similarly, treating alcohol 2 with aldehyde 3 also failed to produce ester 6 (tables S11 and S12), ruling out the Cannizzaro mechanism (Scheme 2). Finally, we investigated reactions of the aldehyde 3 alone in the absence of O2, with and without catalyst, and again observed no ester 6 formation (tables S11 and S12), precluding the Tishchenko mechanism (Scheme 3). Hence we conclude that the high selectivities observed are consistent with the involvement of the hemiacetal 5 (Scheme 1).

To determine the origin of the differences in activity between the two catalysts, we have characterized the common starting sol (18) and the two sol-immobilized materials using EM. Figure S4 (18) shows a representative scanning TEM (STEM) high-angle annular dark field (HAADF) image of the starting colloidal particles deposited onto a carbon film. The nanoparticles have a mean particle size of 2.9 nm (fig. S5A) and were found to be homogeneous Au-Pd alloys. More than 80% of the particles were icosahedral or decahedral (multiply twinned), with the remainder being cub-octahedral or just single/double-twinned in character. They were also distinctly rounded in shape because they were still coated with protective polyvinyl alcohol ligands. After deposition of the colloids onto either activated amorphous carbon or TiO2, the sol-immobilized material was dried at 120°C for 3 hours. The low-temperature drying process in each case caused a very modest size increase in the Au-Pd particles, leading to HAADF-measured mean sizes for the TiO2- and carbon-supported catalysts of 3.9 and 3.7 nm, respectively (fig. S5 B and C).

The most important difference noted between the two samples relates to the morphology of the Au-Pd particles. In contrast to the starting colloid, many of the smaller particles were now found to be highly faceted, and primarily cub-octahedral (Fig. 2, A to D) or singly twinned (Fig. 2E) in character. Such particles preferentially exposed distinct {111}- and {200}-type facets. The larger particles (Fig. 2F) still tended to be multiply twinned and exclusively exposed {111}-type facet planes. In addition, the Au-Pd particles tended to form an extended flat interface structure with the crystalline TiO2 substrate (fig. S6), which could serve to improve particle adhesion and inhibit sintering at higher temperatures.

Fig. 2

Representative (A to F) STEM HAADF micrographs of Au-Pd nanoparticles in the Au-Pd/TiO2 sol-immobilized sample. Representative (G to L) STEM HAADF micrographs of Au-Pd nanoparticles in the Au-Pd/C sol-immobilized sample.

In contrast, the corresponding HAADF images (Fig. 2, G to I) from the carbon-supported samples show that the Au-Pd particles are more rounded and have a much lower ability to wet the amorphous carbon support. The distribution of Au-Pd particle morphologies on the carbon support was much closer to that of the starting colloid, with icosahedral and decahedral (fivefold-twinned) particles predominating (Fig. 2, G to J) and comparatively few single/double-twinned (Fig. 2K) and cub-octahedral (Fig. 2L) particles present. It is plausible that the strong interaction of the Au-Pd metal with the TiO2 support in the former case templates many of the smaller multiply twinned particles to restructure as cub-octahedral particles or simple single/double-twinned crystals during the drying step. X-ray photoelectron spectroscopy (XPS) analysis of the Au-Pd catalysts immobilized on carbon and TiO2 showed them to have a Pd/Au ratio of 2.1 and 2.2, respectively, and confirmed that in both cases, the Pd was predominantly in the metallic state (23).

The AuPd/C sample had approximately double the catalytic activity of the AuPd/TiO2 sample, despite having a very similar size distribution of Au-Pd particles. This suggests that simple metal surface area considerations are not dominating the catalytic activity in this instance, because the total numbers of exposed surface atoms are almost identical (24), and the TONs per surface-exposed atoms for the most-active catalysts are 1.03 × 104 for Au-Pd/C (Table 1, entry 16) and 0.56 × 104 for Au-Pd/TiO2 (Table 1, entry 17) (tables S14 and S15) (18). In addition, the similarity in surface composition, as evidenced by XPS, does not provide a clue as to the source of the activity difference. The Au-Pd/TiO2 catalyst does clearly have more support/particle periphery sites relative to the Au-Pd/C catalyst, by virtue of its flatter, better wetting interface. However, such periphery sites are probably not implicated in the catalytic process in this instance, because the Au-Pd/TiO2 catalyst displays the lower activity. It is possible that the disparity in catalytic activity may be related to subtle differences in the number of low coordination facet-edge and corner sites in the two cases. The flatter, more faceted, Au-Pd particles have fewer of these low coordination number sites, whereas the more irregular, rougher Au-Pd particles have substantially more of them. If these low coordination number corner and/or edge positions are implicated as active sites for toluene oxidation, then their relatively higher occurrence in the Au-Pd/C sample could account for the superior performance displayed by this catalyst. Another possible explanation could lie in the difference in the distribution of Au-Pd particle morphologies found in the two catalyst samples. The Au-Pd/C catalyst predominantly has multiply twinned (icosahedral and decahedral) particles, which tend to have {111} facet terminations. In comparison, the Au-Pd/TiO2 materials show an increased fraction of cub-octahedral and singly/doubly twinned particles, which exhibit mixed {100}/{111} facet terminations. Hence, the increasing proportion of {100}-type facets in the Au-Pd/TiO2 sample correlates with a lowering of the catalytic activity, and preparation strategies need to avoid them.

In a final set of experiments, we investigated the stability of the catalysts, because it is crucial to confirm that high-activity catalysts can be reused. With the Au-Pd/TiO2 catalyst, the reaction was stopped after 7 hours, and the catalyst was recovered by decantation. Identical conversion was obtained on reuse of the Au-Pd/TiO2 catalyst (Table 1, entries 14 and 15). For the Au-Pd/C catalyst, the reaction was stopped after 7 hours and the catalyst was allowed to settle. The liquid phase was then carefully removed by decantation and fresh toluene was added. No metal was observed to have leached into the liquid phase during reaction, and the decanted liquid showed no further reaction in the absence of catalyst (fig. S7 and table S16). The reaction was then allowed to proceed for a further 7 hours, and the whole process was repeated a further two times. The reaction profile obtained with the decantation experiments was identical to that obtained with the fresh catalyst (fig. S8). Detailed STEM characterization shows that there is minimal particle growth or morphology change for the Au-Pd/C catalysts when studied over an extended reaction period (Fig. 1), during which catalysts were recovered after 31 and 65 hours of reaction, followed by two reuse cycles of 7 hours (figs. S9 and S10). Therefore, it is clear that any sintering or structural modification of these highly active catalysts is minimal, and we consider them to be stable and reusable.

Supporting Online Material

Materials and Methods

Figs. S1 to S10

Tables S1 to S16


References and Notes

  1. See supporting on Science Online for detailed methods.
  2. We acknowledge the support of the Dow Chemical Company through the Dow Methane Challenge.
View Abstract

Navigate This Article