Report

Solvent-Free Oxidation of Primary Alcohols to Aldehydes Using Au-Pd/TiO2 Catalysts

See allHide authors and affiliations

Science  20 Jan 2006:
Vol. 311, Issue 5759, pp. 362-365
DOI: 10.1126/science.1120560

Abstract

The oxidation of alcohols to aldehydes with O2 in place of stoichiometric oxygen donors is a crucial process for the synthesis of fine chemicals. However, the catalysts that have been identified so far are relatively inactive with primary alkyl alcohols. We showed that Au/Pd-TiO2 catalysts give very high turnover frequencies (up to 270,000 turnovers per hour) for the oxidation of alcohols, including primary alkyl alcohols. The addition of Au to Pd nanocrystals improved the overall selectivity and, using scanning transmission electron microscopy combined with x-ray photoelectron spectroscopy, we showed that the Au-Pd nanocrystals were made up of a Au-rich core with a Pd-rich shell, indicating that the Au electronically influences the catalytic properties of Pd.

Selective oxidation is important in the synthesis of fine chemicals and intermediates (1); and, in particular, the oxidation of primary alcohols to aldehydes is a fundamentally important laboratory and commercial procedure (18). Aldehydes are valuable both as intermediates and as high-value components for the perfume industry (1, 9, 10). Many oxidations of this type are carried out using stoichiometric oxygen donors such as chromate or permanganate, but these reagents are expensive and have serious toxicity issues associated with them (1, 9, 1114). In many cases, aldehydes are obtained only from activated alcohols in which the carbon bears a phenyl group, such as benzyl alcohol (7, 8). Sheldon and co-workers (3) have obtained good yields from a biphasic system for the catalytic conversion of pentan-1-ol to the aldehyde, but the acid is produced in the case of hexan-1-ol. Given these limitations, there is substantial interest in the development of heterogeneous catalysts that use either O2 or H2O2 as the oxidant (15). Au nanocrystals have been shown to be highly effective for the oxidation of alcohols with O2 in an aqueous base, in particular diols and triols; but under these conditions, the product is the corresponding monoacid, not the aldehyde (1619). Gold catalysts have, however, been found to be effective for the gas-phase oxidation of volatile alcohols to the corresponding aldehydes and ketones (20).

Most recently, two studies have shown that supported metal nanoparticles can be very effective catalysts for the oxidation of alcohols to aldehydes, using O2 under relatively mild conditions. Kaneda and co-workers (6) found that hydroxyapatite-supported Pd nanoclusters (Pd/HAP) give very high turnover frequencies (TOFs) for the oxidation of phenylethanol and benzyl alcohol but show limited activity for the oxidation of primary alkyl alcohols (such as octan-1-ol oxidation). Corma and co-workers (21) have shown that the addition of Au nanocrystals to CeO2 converts the oxide from a stoichiometric oxidant to a catalytic system, with TOFs similar to those obtained by Kaneda and co-workers (6).

Recently, we have shown that supported Au-Pd alloys are efficient catalysts for the direct synthesis of H2O2 from H2 oxidation by O2 at low temperatures (2225). In particular, Au-Pd/TiO2 catalysts were very selective for H2O2 synthesis. Hydroperoxy species are considered to be involved in this H2O2 formation process, and because hydroperoxy species are key reagents/intermediates in the oxidation of alcohols (1), we reasoned that these catalysts should also be effective for the oxidation of alcohols. We showed that TiO2-supported Au-Pd alloy nanocrystals give significantly enhanced activity for alcohol oxidation using a green chemistry approach with O2 under mild solvent-free conditions. When compared with monometallic supported Au (21) and Pd (6), the Au-Pd catalysts nanocrystals give TOFs that are enhanced by a factor of ∼25.

The TiO2-supported Au-Pd catalysts were initially investigated for the oxidation of benzyl alcohol at 373 K with O2 as oxidant in the absence of solvent (Fig. 1A). The Au-Pd/TiO2 catalysts were very active for this reaction, and the selectivity to benzaldehyde was ≥96%, with the only byproduct being benzyl benzoate. In contrast, Pd/TiO2 also produced toluene and benzene as byproducts, and a Au/TiO2 catalyst produced a significant amount of an acetal product. The selectivity of the Au/TiO2 catalyst for benzaldehyde decreased with the time on line, but further oxidation of the acetal byproduct led to an increase of the final selectivity in benzaldehyde. Carbon mass balances were 100%, and no carbon oxides were formed for Au-Pd/TiO2 or Au/TiO2 catalysts.

Fig. 1.

(A) Benzyl alcohol conversion and selectivity in benzaldehyde with the reaction time at 373 K and 0.1 MPa pO2. Squares, Au/TiO2; circles, Pd/TiO2; and triangles, Au-Pd/TiO2. Solid symbols indicate conversion, and open symbols indicate selectivity. (B) Au-Pd/TiO2 catalyzed reactions at 363 K, 0.1 MPa pO2, for cinnamyl alcohol (squares) and vanillyl alcohol (circles). Solid symbols indicate conversion, and open symbols indicate selectivity to the corresponding aldehydes.

The effect of adding Au to a Pd/TiO2 catalyst is apparent in these initial studies. Although the Pd/TiO2 catalyst has a high initial activity, and the addition of Au decreases the activity, the Au-Pd/TiO2 catalyst retained high selectivity to benzaldehyde at high conversion rates, a feature not observed with the supported pure-Au and pure-Pd catalysts. One of the key factors that must be considered for heterogeneous catalysts operating in three-phase systems is the possibility that active components can leach into the reaction mixture, thereby leading to catalyst deactivation or, in the worst case, to the formation of an active homogeneous catalyst. We have found that the catalysts calcined at 673 K are stable and do not leach Au or Pd into solution, a feature we have previously also observed in our studies on hydrogen peroxide synthesis (24, 25) (see supporting online material).

Because Al2O3 and Fe2O3 were also effective supports for the formation of H2O2 with Au-Pd nanocrystals, these catalysts were also investigated, but the TiO2-supported Au-Pd catalysts are preferred for the oxidation of benzyl alcohol (Table 1); it is likely that the more acidic nature of the Al2O3 and Fe2O3 supports led to enhanced byproduct formation.

Table 1.

Comparative data for benzyl alcohol oxidation and hydrogen peroxide synthesis. Results were obtained for the oxidation of benzyl alcohol after 0.5 hour and 8 hours of reaction and for H2O2 synthesis for 0.5 hour. The oxidation of benzyl alcohol was carried out at 373 K temperature, 0.2 MPa pO2, and 1500 rpm stirrer speed. The H2O2 synthesis was carried out under the conditions described in (26, 27). Productivities are quoted in units of moles of product per hour per kilogram of catalyst.

CatalystBenzyl alcohol oxidation
Conversion (%)Benzaldehyde selectivity (%)Benzaldehyde productivityView inline[mol/(hour/kgcat)]H2O2 productivity [mol/(hour/kgcat)]
0.5 hour8 hours0.5 hour8 hours
2.5% Au-2.5% Pd/Al2O3 2.6 83.3 90.5 86.6 174 23
2.5% Au-2.5% Pd/TiO2 3.7 74.5 95.2 91.6 165 64
2.5% Au-2.5% Pd/SiO2 3.6 35.7 97.3 88.0 76 80
2.5% Au-2.5% Pd/Fe2O3 3.6 63.4 74.9 66.4 102 16
2.5% Au-2.5% Pd/C 2.9 69.2 53.9 46.4 78 30
2.5% Au/TiO2 0.6 15.3 96.7 63.9 24 <2
2.5% Pd/TiO2 13.4 60.1 51.3 54.4 79 24

The Au-Pd/TiO2 catalysts were investigated with a range of substrates and conditions (Table 2). In the initial experiments conducted at 373 K (Table 2, entries 1 to 3), relatively high catalyst loadings were used, but selectivities >90% could still be achieved for the oxidation of benzyl alcohol to benzaldehyde. Subsequent experiments used much lower metal concentrations, and the effect of the pressure of O2 (pO2) (Table 2, entries 4 to 7) showed that the reaction was of zero order in O2. The selectivity to the aldehydes increased with pO2 [for conversions of benzyl alcohol, >70%, the selectivity to the aldehyde increased from 71% at low pO2 (Table 2, entry 4) to 86% at the higher pO2 (Table 2, entry 6)] and metal concentration, but selectivities of >80% were achieved readily under most conditions. These initial experiments were conducted in a closed autoclave with O2 at a constant reaction pressure, so that as the reaction proceeded and O2 was consumed, this oxygen was replenished. We have used air in place of O2 and have obtained the same initial TOFs, which demonstrates that in principle, air can be used for industrial applications. Increasing the temperature (Table 2, entries 4 and 8 to 10) increased the rate as expected (with activation energy EA = 45.8 kJ/mol), but the selectivity decreased to ∼60% at 433 K, and the byproducts included benzylbenzoate and toluene under these more forcing conditions. The Au-Pd/TiO2 catalysts were very active, and the best performance (Fig. 1A) was achieved at lower temperatures (≤373 K), higher pO2 (0.2 to 1.0 MPa), and higher catalyst loadings, but we stress that we have not attempted to optimize the performance at this time. Vanillyl alcohol and cinnamyl alcohol, two important aromatic unsaturated alcohols, were oxidized in the presence of solvents, and 100% specificity to the aldehydes was achieved in both cases at 363 K, further confirming the efficacy of these supported mixed-metal catalysts (Table 2, entries 13 and 14, and Fig. 1B).

Table 2.

Comparison of the catalytic activity for alcohol oxidation to the corresponding aldehyde. Catalyst is 2.5% Au-2.5% Pd/TiO2 unless noted otherwise; substrates oxidized without solvent unless specified; catalyst mass varied to give the metal concentrations indicated; and TOF was measured after first 0.5 hour of reaction. T, temperature.

EntryAlcoholReaction conditions[Metal] (10-5 mol/liter)TOF (/hour)
T (K)P (105 Pa)AuPd
1 Benzyl alcohol 373 2 63.5 118 607
2 Benzyl alcoholView inline 373 2 63.5 0 213
3 Benzyl alcoholView inline 373 2 0 118 2,200
4 Benzyl alcohol 373 1 2.1 3.9 6,190
5 Benzyl alcohol 373 2 2.1 3.9 6,440
6 Benzyl alcohol 373 5 2.1 3.9 6,190
7 Benzyl alcohol 373 10 2.1 3.9 5,950
8 Benzyl alcohol 383 1 2.1 3.9 14,270
9 Benzyl alcohol 393 1 2.1 3.9 26,400
10 Benzyl alcohol 433 1 2.1 3.9 86,500
11 1-Phenylethanol 433 1 1.8 3.2 269,000
12 3-Phenyl-1-propanol 433 1 2.1 3.9 2,356
13 Vanillyl alcoholView inline 363 1 21.6 40.6 10
14 Cinnamyl alcoholView inline 363 1 21.6 40.6 97
15 Octan-1-ol 433 1 2.5 4.7 2,000
16 Octan-2-ol 433 1 2.5 4.7 0
17 Octan-2-ol/octan-1-ol 433 1 2.1 3.9 0
18 Octan-3-ol 433 1 2.1 3.9 10,630
19 1-Octen-3-ol 433 1 2.1 3.9 12,600
20 Crotyl alcohol 433 5 2.1 3.9 12,600
21 Butan-1-ol 433 5 2.1 3.9 5,930
22 1,2-Butanediol 433 1 2.1 3.9 1,520
23 1,4-Butanediol 433 1 2.1 3.9 104,200
24 Benzyl alcoholView inline 433 1 2.1 3.9 12,500
25 Benzyl alcoholView inline 433 1 2.1 0 12,400
26 Benzyl alcoholView inline 433 1 0 3.9 24,800
27 Benzyl alcoholView inline 433 1 2.4 4.5 36,500
28 Benzyl alcoholView inline 433 1 0 3.6 37,600
29 1-PhenylethanolView inline 433 1 0 3.1 11,600
  • View inline* 2.5% Au/TiO2.

  • View inline 2.5% Pd/TiO2.

  • View inline 0.2 mol/liter in toluene as solvent.

  • View inline§ 0.2 mol/liter in water as solvent.

  • View inline 2.5% Au-2.5% Pd/HAP prepared by impregnation of HAP with HAuCl4·3H2O and PdCl2.

  • View inline 2.5% Au/HAP prepared by impregnation of HAP with HAuCl4·3H2O.

  • View inline# 2.5% Pd/HAP prepared by impregnation of HAP with PdCl2.

  • View inline** 2.5% Au-2.5% Pd/TiO2 prepared with the method of Kaneda (6) using TiO2 as support.

  • View inline†† 0.2% Pd/HAP prepared using the method of Kaneda (6) using HAP as support.

As noted above, both Kaneda and co-workers (6) and Corma and co-workers (21) have shown that supported Pd and Au monometallic catalysts are highly effective for the oxidation of 1-phenylethanol under solvent-free conditions at 433 K with a pO2 of 0.1 MPa. Under these conditions, the Pd/HAP and Au/CeO2 catalysts gave TOFs of 9800 and 12,500/hour for 1-phenylethanol, and we have replicated the results for the Pd/HAP catalyst (Table 2, entries 28 and 29). With our Au-Pd/TiO2 catalyst, we obtained a TOF of 269,000/hour. The Au-Pd/TiO2 catalyst is also effective for a range of straight-chain, benzylic, and unsaturated alcohols (Table 2, entries 12 to 23) and, in particular, for the oxidation of primary alcohols, such as butan-1-ol and octan-1-ol (Table 2, entries 15 and 21), and high TOFs are observed. This trend was also observed with 1,4-butanediol, for which the observed reactivity is significantly enhanced over that observed for butan-1-ol. This difference may be due to the interaction of this substrate with the active site (Table 2, entry 23), but 1,2-butanediol is much less active (Table 2, entry 22). Even at high reaction temperatures, octan-2-ol is inactive (Table 2, entries 16 and 17), which is in direct contrast with the Pd/HAP and Au/CeO2 catalysts, for which secondary alcohols are more reactive than primary alcohols. However, the effect of decreased reactivity appears to be limited to 2-alcohols, because octan-3-ol is very reactive (Table 2, entry 18). The addition of octan-2-ol to a reaction mixture leads to a total loss of activity with our catalyst (Table 2, entry 17), and the addition of octan-2-one in small amounts has a similar effect. These findings may indicate a specific interaction of these ketones with the Au-Pd catalyst. To show that TiO2 is a particularly effective support, we contrasted catalysts prepared using HAP and TiO2 as supports with the standard impregnation method we have used (Table 2, entries 10 and 25 to 27); it is clear that the TiO2 support gives improved activity. There is, however, considerable scope to improve and optimize the performance of these mixed metal Au-Pd catalysts in future studies.

The TiO2-supported Au-Pd catalyst was characterized using x-ray photoelectron spectroscopy (XPS) and scanning transmission electron microscopy (STEM). The XPS results (Fig. 2A) show that the surfaces of the metal nanoparticles in the calcined catalyst material are significantly enriched with Pd. The uncalcined sample shows XPS signals characteristic of both Au and Pd, but as we have stated earlier, these materials are not stable under the reaction conditions. Once calcined at 673 K, the catalysts are stable, and the signal for Au is significantly decreased. STEM analyses of individual metal nanoparticles in the calcined Au-Pd/TiO2 catalysts were carried out, such as those presented in the montage of annular dark-field (ADF) images and energy-dispersive x-ray (XEDS) maps in Fig. 2B. The Pd x-ray signal originates from a slightly larger spatial area than that of the corresponding Au x-ray signal. This effect is best illustrated in the color map that overlays the filtered Ti, Pd, and Au x-ray signals after the application of multivariate statistical analysis (26). In agreement with the XPS study, we deduce that Pd surface segregation occurs during calcination to produce alloy nanoparticles having a Pd-rich shell surrounding a Au-rich core.

Fig. 2.

(A) Au(4d) and Pd(3d) spectra for a 2.5 weight % Au–2.5 weight % Pd/TiO2 catalyst after different heat treatments: (a) uncalcined or (b) calcined at 673 K in air. (B) Montage showing the ADF-STEM image of a bimetallic particle and the corresponding MSA-processed STEM-XEDS maps of the Au-M2, Pd-Lα, O-Kα, and Ti-Kα signals. Also shown is a reconstructed MSA-filtered Au-Pd-Ti composition map (Ti, red; Au, blue; and Pd, green).

The catalytic data show that the introduction of Au to Pd improves selectivity, and we believe that the surface of the bimetallic nanoparticles will still contain some Au. Hence, we argue that the Au acts as an electronic promoter for Pd and that the active catalyst has a surface that is significantly enriched in Pd. Recent studies have started to provide insights into the nature of such effects. For example, Okazaki et al. (27) have shown, using a combination of experiment and theory, that the electronic structure of Au in Au/TiO2 catalysts is dependent on the particle size, and Goodman and co-workers (28), using model studies, have shown that Au can isolate Pd sites within bimetallic systems.

Supporting Online Material

www.sciencemag.org/cgi/content/full/311/5759/362/DC1

Materials and Methods

Fig. S1

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article