Ni₂O₃–Au⁺ hybrid active sites on NiO_x@Au ensembles for low-temperature gas-phase oxidation of alcohols

Abstract

An interesting ensemble of NiO_x@Au (i.e. 20–30 nm gold particles partially covered with very small NiO_x segments) were clearly identified to be highly active for the low-temperature gas-phase oxidation of alcohols. On such active NiO_x@Au ensembles, large amounts of Ni₂O₃–Au⁺ hybrid active sites were defined, taking a step closer to identifying the low-temperature activity. By their nature, Ni₂O₃ specimens not only promote the formation of Au⁺ ions and stabilize them but also act as an oxygen supplier to transfer oxygen species onto the Au⁺ sites to react with alcohol.

---

Bulk gold is extremely inert for catalysis.^1a However, supported gold nanoparticles (Au NPs) have recently gained tremendous attention as an excellent catalyst and have blossomed in many fields.¹ Up to now, multiple factors, including Au particle size and chemical state,^2,3 the support properties⁴ (such as type,^4a morphology (shape)^4b–d and size^4e–g) and Au–support interactions,^2a,c,d have been found to contribute to the distinct catalytic activity of supported Au NPs. Among them, the small size of the Au NPs (<5 nm) has been considered as one dominant factor which contributes to higher catalytic activity, due to the size-related low coordination, unsaturated surface atoms which often act as active sites.^2a–d

However, some recent contributions reported just the opposite behavior in that large gold-based particles were also active in many reactions, such as Au–Pd core–shell particles (∼10 nm) for the liquid-phase selective oxidation of alcohols,^5a Au/SiO₂^5b with large Au particles (∼15 nm) and Au–Cu/SiO₂^5c with large AuCu alloy particles (∼25 nm) for the gas-phase selective oxidation of alcohols, Au/Al₂O₃ with Au particles (50–100 nm) for the dehydrogenation of amines to imines^5d and monolithic nanoporous gold for the gas-phase selective oxidative coupling of methanol to methyl formate.^5e These results are astonishing but unfortunately, the origin of this adverse behavior is not deeply discussed therein. Moreover, the catalytic activity of the large-sized Au particles could be further enhanced by depositing transition-metal oxides with low coverage onto a single gold crystal, such as CeO₂/Au(111)^6a and TiO₂/Au(111).^6a,b The oxide–Au interface has been proposed to explain the adverse behavior of CeO₂(TiO₂)/Au(111),^6a,b but the structure and nature of the active site at the oxide–Au interface is still unclear.

More recently, we have reported a microfibrous-structured gold catalyst (Au/Ni-fiber) which exhibited an extremely high activity and selectivity in the low-temperature gas-phase oxidation of alcohols,⁷ within which the formation of active NiO_x@Au ensembles is critical to promote the high catalytic activity. Despite the above advance, the origin of the catalytic performance over such ensembles, the finer structure and better understood nature of the active sites over the ensembles and the mechanism of alcohol oxidation over the active sites are still lacking or ambiguous. A clear understanding is of great importance for optimizing the current catalytic processes and rational design of novel catalysts. To this end, three issues will be addressed in the present research: (1) penetrate the indispensability of the nano-architecture of the NiO_x@Au ensembles contributing to higher catalytic activity; (2) reveal the activity originally from the specific hybrid active sites hidden on such NiO_x@Au ensembles; (3) describe how the hybrid active sites promote alcohol oxidation.

Generally, the activity of catalysts relies on the physical structure of their active ensembles. The base NiO_x@Au/Ni-fiber catalyst obtained via a gold galvanic deposition method (see the catalyst preparation in the ESI† for details) was checked by HRTEM, whereby small NiO_x segments were found to partially cover the large gold particle to form interesting NiO_x@Au ensembles (Fig. S1, ESI†). With these ensembles, the NiO_x@Au/Ni-fiber catalyst shows much higher low-temperature activity and TOF for the gas-phase oxidation of benzyl alcohol (conversion of 95% at 250 °C, Entry 1 in Table 1), compared with the Au/SiO₂ catalyst (conversion of 6% at 250 °C, Entry 2 in Table 1) and maintained an optimum equilibrium state for a long time in the reaction environment (Fig. S2, ESI†).

Table 1 The catalytic performance of various pre-activated catalysts for benzyl alcohol oxidation

Entry	Catalyst^a^,^b	DAu^d (nm)	DNiOx^d (nm)	Conv. (%)	Sel. (%)	TOF^e (h^−1)
The mass balance of 99.9%, with an experimental error of 0.2%, was achievable (see Note in Fig. S2 in the ESI for details).a For all the catalysts, the gold loading is 4.0 wt% and NiOx loading is 3.2 wt% if not otherwise specified. The catalysts were pre-activated by undergoing the gas-phase oxidation of benzyl alcohol at 380 °C for 1 h using a molar ratio of O2 to alcoholic hydroxyl (O2/hydroxyl) = 0.6 and weight hourly space velocity (WHSV) = 20 h^−1, and were then examined at 250 °C using O2/hydroxyl = 0.6 and WHSV = 20 h^−1.b The catalyst preparation is described in ESI.c The Au and NiOx loadings for the NiOx@Au/Ni-fiber catalyst were 3.8 wt% and 2.4 wt% (see characterization in the ESI).d The particle size was estimated from the XRD patterns using Scherrer's equation.e TOF (turnover frequency) was calculated based on the amount of surface Au atoms (see Table S1 in the ESI for details).
1	NiOx@Au/Ni-fiber^c	25	5	95	99	17097
2	Au/SiO2	27	—	6	99	1156
3	Au/Ni-fiber-W	25	—	31	99	5591
4	NiOx/Ni-fiber	—	5	6	99	—
5	NiOx@Au/Ni-fiber-W	25	6	93	99	16720
6	Au/Ti-fiber	30	—	4	99	839
7	NiOx/Ti-fiber	—	5	5	99	—
8	NiOx@Au/Ti-fiber	27	5	92	99	17907
9	Au/NiO–Ni-fiber	24	—	30	99	5155
10	NiOx@Au/NiO–Ni-fiber	24	—	93	99	16031

---

To penetrate the indispensability of such active NiO_x@Au ensembles for the low-temperature gas-phase oxidation of alcohol, a series of model catalysts were designed, prepared, (see the catalyst preparation in the ESI† for details), and then compared by testing the gas-phase oxidation of benzyl alcohol. For comparison, a Au/Ni-fiber-W catalyst with only Au particles was prepared by removing NiO_x segments from the NiO_x@Au ensembles (Fig. S3 and S4, ESI†), over which, not surprisingly, the benzyl alcohol conversion and TOF was markedly reduced (Entries 1 and 3 in Table 1). In addition, just NiO_x supported on Ni-fibers also delivered a very low benzyl alcohol conversion of 6% (Entry 4 in Table 1). We wondered what would happen with post-loading NiO_x onto the Au particles over the Au/Ni-fiber-W catalyst. After that, the regeneration of the active NiO_x@Au ensembles was observed clearly by TEM (Fig. S5A and S5B, ESI†), which showed that the large Au particles were partially covered with small NiO_x segments. As a result, benzyl alcohol conversion increased to 93% over the NiO_x@Au/Ni-fiber-W catalyst (Entry 5 in Table 1), almost as high as that over the NiO_x@Au/Ni-fiber catalyst (Entry 1 in Table 1).

Subsequently, the Ni-fibers were replaced by Ti-fibers to avoid in situ formation of the special NiO_x@Au ensembles, since a galvanic reaction cannot proceed between Ti-fibers and HAuCl₄. Analogously, only Au particles or NiO_x supported on the Ti-fibers induced a very low benzyl alcohol conversion of <5% (Entries 6 and 7 in Table 1). Excitingly, after post-introducing NiO_x artificially onto the Au/Ti-fiber catalyst, the conversion and TOF were remarkably promoted (Entry 8 in Table 1) along with the formation of the NiO_x@Au ensembles (Fig. S5C and S5D, ESI†). This again indicated that the NiO_x@Au ensembles are essential for the low-temperature gas-phase oxidation of alcohols.

Are Au@NiO ensembles (i.e. a NiO plate partially covered with 20–30 nm Au particles) still able to deliver a high catalytic activity? To answer this question, Au particles were supported onto NiO–Ni-fiber (obtained by calcining pure sintered Ni-fibers at 600 °C in air for 2 h) to form Au@NiO ensembles (see Discussion 1 in the ESI† for details). Unfortunately, over such ensembles, a benzyl alcohol conversion of only 30% was obtained with 99% selectivity (Entry 9 in Table 1, similar to the result for the Au/Ni-fiber-W catalyst with NiO_x removal (Entry 3 in Table 1). Likewise, by again post-adding NiO_x onto the Au/NiO–Ni-fiber catalyst to form NiO_x@Au ensembles on the NiO–Ni-fiber support, the benzyl alcohol conversion was sharply promoted to 93% and the TOF restored to a high level, similar to that over the NiO_x@Au/Ni-fiber catalyst (Entries 1 and 10 in Table 1).

Therefore, all the results undoubtedly demonstrate that, upon building the nano-architecture of the NiO_x@Au ensembles, the catalysts with such NiO_x@Au ensembles always show promising reactivity in the oxidation reaction, even when changing the support fibers or preparation methods. This has firmly proven the indispensability and popularity of the special active NiO_x@Au ensemble structure for the low-temperature gas-phase oxidation of alcohols by ruling out Au@NiO ensembles as well as only Au particles or NiO_x segments. For the metal–oxide catalysts, the metal–oxide interaction is a very important factor for their catalytic performance.⁸ Indeed, many studies have reported that Au–oxides interations also exist at their interface.^2a,c,d,9 However, the active sites over such NiO_x@Au ensembles and their corresponding catalytic nature are not yet clear.

To define the structure of the active sites over such ensembles, surface-sensitive X-ray photoelectron spectroscopy (XPS) was used to study the NiO_x@Au/Ni-fiber catalyst (i.e., Entry 1 catalyst in Table 1). Fig. 1 shows the XPS spectra in the O 1s, Au 4f and Ni 2p regions for several NiO_x@Au/Ni-fiber samples undergoing different reaction conditions. On the surface of the working catalyst sample (obtained after undergoing a reaction with benzyl alcohol in the presence of O₂ (O₂/hydroxyl = 0.6) at 250 °C for 0.5 h), apart from Au⁰ (Au 4f (7/2): 84.0 eV) and NiO (Ni 2p: 854.4 eV; O1s: 529.7 eV), not only were large amounts of Ni₂O₃ specimens (Ni 2p: 856.2 eV; O 1s: 531.8 eV) formed but considerable amounts of Au⁺ specimens were also clearly detected, with a Au 4f (7/2) peak at 84.8 eV (Fig. 1a–c). Note that O atoms adsorbed onto the Au⁺ sites generally exist in the form of OH groups that also provide an O 1s peak at approximately 532 eV (Fig. S6, ESI†).¹⁰ CeO₂ (or Fe₂O₃, TiO₂)–Au interfaces have been previously reported and proposed to generate cationic gold and activate molecular oxygen for oxidation reactions.^4e,f,6d,11

Fig. 1 The XPS spectra in the O 1s (A), Ni 2p (B) and Au 4f (C) regions of the surface of several catalysts. (a–c) The working NiO_x@Au/Ni-fiber catalyst, the same as the catalyst in Entry 1 in Table 1; (d–f) The spent catalyst, obtained by placing the working catalyst in a benzyl alcohol stream in the absence of O₂ for 0.5 h; (g–i) The spent catalyst, catalyzing benzyl alcohol oxidation by re-feeding O₂ again under the same conditions as in (A-a).

On the surface of the working catalyst, the Au⁺ surface content to the total amount of surface gold atoms is 31%, which was obtained by calculating the area of the corresponding XPS peaks (A_Au+/(A_Au0 + A_Au+) = 0.31), and the Ni₂O₃ surface content to the total amount of surface nickel oxides is 43% (A_Ni2O3/(A_NiO + A_Ni2O3) = 0.43) with a surface Au⁺ : Ni³⁺ peak area ratio of 1 : 45 (Fig. 1a–c, Table S2, ESI†). As the working catalyst underwent the reaction with benzyl alcohol in the absence of O₂ for 0.5 h (the resulting catalyst was called the spent catalyst), the benzyl alcohol conversion was decreased from 95% to only 0.7% (Table S3, Fig. S7, ESI†) while the Au⁺ surface content disappeared completely (Fig. 1f, Table S2, ESI†). Interestingly, the Ni₂O₃ surface content was decreased from 43% to 33% with the stoichiometric reduction of the O 1s peak for Ni₂O₃, while the NiO surface content was increased from 57% to 67%. Combined with the fact that the whole peak area for Ni remained unchanged, this observation definitely indicates the transformation of Ni₂O₃ to NiO when reacting without O₂ (Fig. 1d and e, Table S2, and the more detailed Discussion 2 in the ESI†). Subsequently, after the spent catalyst was reacted with benzyl alcohol by re-feeding O₂ at 250 °C, the Au⁺ and Ni₂O₃ surface content both recovered to previous levels, with a Au⁺ content of 33% and a Ni₂O₃ content of 42% (Fig. 1g–i, Table S2, ESI†), and the benzyl alcohol conversion returned to 93%.

X-ray absorption near-edge structure (XANES) is sensitive to the oxidation state of atoms on the surface of metallic particles,^12a–d because there is an intense whiteline in the compounds containing a metallic cation.^12a To further verify the high-low-high associate-changes of Au⁺ and Ni₂O₃, the above three catalysts were also investigated by XANES at the Au L₃ and Ni K edge, with the results shown in Fig. 2A. The working catalyst sample shows a similar XANES pattern with the reported mixture of Au⁺ complexes and Au⁰ particles (Fig. 2A-a),^12e definitely indicating the existence of Au⁺ ions. Interestingly, as the working catalyst underwent the oxidation reaction with only benzyl alcohol for 0.5 h in the absence of O₂, the intensity of the whiteline decreases and the resulting XANES pattern is similar to that of Au foil, indicating that the Au⁺ content decreases to almost zero (Fig. 2A-b and c). After re-feeding O₂ to the spent catalyst, the whiteline intensity increases to the previous level of the working catalyst, indicating that the Au⁺ ions are regenerated with a content close to the working catalyst level (Fig. 2A-d). Indeed, the same trend is also observed in the changes of the Au–O coordination (Fig. 2B). The working catalyst shows a Au–O coordination at a distance of 1.7 Å (Fig. 2B-a), a similar distance to that of other reported catalysts.^12f Over the spent catalyst, the Au–O coordination almost disappears (Fig. 2B-b and c), while after re-feeding O₂ into the reaction stream, the Au–O coordination likewise increases to the working catalyst level (Fig. 2B-d). Regarding Ni₂O₃, the existence of the Ni-fiber support leads to a very low Ni₂O₃/Ni⁰ molar ratio. This means the XANES technique is not as sensitive as the XPS method in tracking the changes of the Ni₂O₃ content in the above cases. Even so, the high-low-high changes of the whiteline intensity of the XANES pattern at the Ni K edge could still be clearly distinguished (Fig. S8, ESI†). Likewise, the absorption intensity at 8350 eV for the spent catalyst was slightly lower than that for the working catalyst and subsequently, could be revived to the previous level of the working catalyst by re-feeding O₂ into the reaction stream. As expected, similar evolution behaviors of XANES for both Au and Ni were also observed for the NiO_x@Au/Ti-fiber catalyst in the reaction processes with-without-with O₂ (Fig. S9, ESI†). Combined with the XPS results, it is reasonable to infer that the high-low-high change of the absorption intensity at 8350 eV can be assigned to the high-low-high change of the Ni₂O₃ content (see Discussion 3 in the ESI† for details).

Fig. 2 XANES of the Au L₃ edge (A) and the modulus of the fourier transform Au L₃-edge signal (B) of several catalysts. (a) The same was used sample as in Fig. 1A-a; (b) The same sample was used as in Fig. 1A-d; (c) Au foil; (d) The same sample was used as in Fig. 1A-g.

Along with the high-low-high surface content of Au⁺, the benzyl alcohol conversion clearly shows a high-low-high evolution behavior as well. Over the spent catalyst, the benzyl alcohol conversion decreased from 95% to 0.7%, with a significant decrease of the Au⁺ surface content from 31% to 1.7%. In this case, an associated reduction of Ni₂O₃ from 43% to 33% was observed. These results clearly indicate that the combination of Ni₂O₃ and NiO with the absence of Au⁺ (i.e., the spent catalyst) was inactive for the low-temperature gas-phase oxidation of alcohols by the direct-oxidation dehydrogenation route.¹³

On the surface of the NiO_x@Au ensembles, it is worth noting that the content change of Au⁺ in the reaction processes with-without-with O₂ was solely associated with a parallel change of Ni₂O₃ but not NiO (Fig. 1). This unique relationship between Ni₂O₃ and Au⁺ indicates a synergistic interaction between the Au particles and Ni₂O₃ specimens rather than NiO. By their nature, the Ni₂O₃ specimens are proposed to play a key role not only in promoting the formation of Au⁺ ions and stabilizing them, but also in acting as a surface-active-oxygen species supplier.

When looking at Table 1, the Au/Ni-fiber-W and Au/NiO–Ni-fiber catalysts (Entries 3 and 9 in Table 1) both delivered much lower benzyl alcohol conversions (∼30%) and TOFs (5100–5600 h⁻¹) than the NiO_x@Au/Ni-fiber catalyst (conversion of 95%, TOF of ∼17000 h⁻¹, Entry 1 in Table 1), but much higher values than the Au/Ti-fiber catalyst (conversion of 4%, TOF of 839 h⁻¹, Entry 6 in Table 1). Although the NiO_x segments were not presented on the Au particles for these two samples, the formation of a NiO_x circumference at the interface between the Au particles and fibrous support (Scheme S1, ESI†) was unavoidable during their preparation. This likely induced a similar synergistic effect as in the NiO_x@Au ensembles in order to form some Au⁺ with Ni₂O₃ specimens, as evidenced by XPS peaks at Au 4f (7/2) = 84.7 eV and Ni 2p = 856.2 eV (Fig. S10, ESI†). The amount of Au⁺–Ni₂O₃ active sites over these two samples, with a surface Au⁺ : Ni³⁺ peak area ratio of 1 : 43 (close to that (1 : 45) for NiO_x@Au/Ni-fiber), is only 35–37% of that over the NiO_x@Au/Ni-fiber catalyst (Tables S2 and S4). This might be the reason for the higher reactivity over both of them compared to the Au/Ti-fiber catalyst but a much lower reactivity than the NiO_x@Au/Ni-fiber catalyst (Table 1).

On the basis of the above investigation, a Ni₂O₃–Au⁺ hybrid active site was defined to be the genesis of the high low-temperature activity and the NiO_x@Au ensembles facilitated the generation of high density Ni₂O₃–Au⁺ sites (Au⁺ content of 31–33% and Ni₂O₃ content of 42–43%, Fig. 1) on the catalyst surface during oxidation. As shown in Fig. 3, the oxygen molecules can be abstracted by the oxygen vacancy of the Ni₂O₃ specimens to be stored as precursors to active oxygen species. Such abstracted oxygen atoms (O 1s B.E.: ca. 532 eV, Fig. 1) can be transferred onto the Au⁺ sites to become active oxygen species which react with alcohol while releasing an oxygen vacancy for the next catalytic cycle.

Fig. 3 The structure of the Ni₂O₃–Au⁺ hybrid active sites on the NiO_x@Au ensembles for the working catalyst and the reaction mechanism of benzyl alcohol oxidation over the active sites.

We thank the NSF of China (20973063, 21076083, 21273075), the MOST of China (2011CB201403), the Fundamental Research Funds for the Central Universities, the Shanghai Rising-Star Program (10QH1400800), the Shanghai Leading Academic Discipline Project (B409) and the Electron Spectroscope Center of the East China Normal University for the TEM measurements.

Notes and references

1. (a) M. Haruta, T. Kobayashi, H. Sano and N. Yamada, Chem. Lett., 1987, 405; (b) M. Haruta, N. Yamada, T. Kobayashi and S. Iijima, J. Catal., 1989, 115, 301; (c) G. J. Hutchings, J. Catal., 1985, 96, 292.
2. (a) G. C. Bond and D. T. Thompson, Catal. Rev. Sci. Eng., 1999, 41, 319; (b) M. S. Chen and D. G. Goodman, Science, 2004, 306, 252; (c) A. S. K. Hashmi and G. J. Hutchings, Angew. Chem., Int. Ed., 2006, 45, 7896; (d) B. K. Min and C. M. Friend, Chem. Rev., 2007, 107, 2709; (e) G. J. Hutchings, Dalton Trans., 2008, 5523.
3. J. C. Fierro-Gonzalez and B. C. Gates, Chem. Soc. Rev., 2008, 37, 2127.
4. (a) M. Comotti, W. C. Li, B. Spliethoff and F. Schüth, J. Am. Chem. Soc., 2006, 128, 917; (b) P. X. Huang, F. Wu, B. L. Zhu, X. P. Gao, H. Y. Zhu, T. Y. Yan, W. P. Huang, S. H. Wu and D. Y. Song, J. Phys. Chem. B, 2005, 109, 19169; (c) J. Li, N. Ta, Y. Li and W. J. Shen, Chin. J. Catal., 2008, 29, 823; (d) R. Si and M. Flytzani-Stephanopoulos, Angew. Chem., Int. Ed., 2008, 47, 2884; (e) J. Guzman, S. Carrettin and A. Corma, J. Am. Chem. Soc., 2005, 127, 3286; (f) A. Abad, P. Concepción, A. Corma and H. García, Angew. Chem., Int. Ed., 2005, 44, 4066; (g) A. A. Herzing, C. J. Kiely, A. F. Carley, P. Landon and G. J. Hutchings, Science, 2008, 321, 1331; (h) J. K. Edwards, B. Solsona, E. Ntainjua N, A. F. Carley, A. A. Herzing, C. J. Kiely and G. J. Hutchings, Science, 2009, 323, 1037.
5. (a) D. I. Enache, J. K. Edwards, P. Landon, B. Solsna-Espriu, A. F. Carely, A. A. Herzing, M. Watanabe, C. J. Kiely, D. W. Knight and G. J. Hutchings, Science, 2006, 311, 362; (b) S. Biella and M. Rossi, Chem. Commun., 2003, 378; (c) C. Della Pina, E. Falletta and M. Rossi, J. Catal., 2008, 260, 384; (d) B. Zhu, M. Lazar, B. G. Trewyna and R. J. Angelici, J. Catal., 2008, 260, 1; (e) A. Wittstock, V. Zielasek, J. Biener, C. M. Friend and M. Bäumer, Science, 2010, 327, 319.
6. (a) J. A. Rodriguez, S. Ma, P. Liu, J. Hrbek, J. Evans and M. Pérez, Science, 2007, 318, 1757; (b) T. Fujitani, I. Nakamura, T. Akita, M. Okumura and M. Haruta, Angew. Chem., Int. Ed., 2009, 48, 9515; (c) S. Shaikhutdinov, Y. N. Sun, Z. H. Qin, M. Lewandowski, E. Carrasco, M. Sterrer and H. J. Freund, J. Catal., 2009, 266, 359; (d) Q. Fu, W. X. Li, Y. Yao, H. Liu, H. Y. Su, D. Ma, X. K. Gu, L. Chen, Z. Wang, B. Wang and X. H. Bao, Science, 2010, 328, 1141.
7. G. F. Zhao, H. Y. Hu, M. M. Deng and Y. Lu, Chem. Commun., 2011, 47, 9642.
8. (a) S. J. Tauster, S. C. Fung and R. L. Garten, J. Am. Chem. Soc., 1978, 100, 170; (b) S. J. Tauster, S. C. Fung, R. T. K. Baker and J. A. Horsley, Science, 1981, 211, 1121.
9. (a) M. Baron, O. Bondarchuk, D. Stacchiola, S. Shaikhutdinov and H. J. Freund, J. Phys. Chem. C, 2009, 113, 6042; (b) T. Risse, S. Shaikhutdinov, N. Nilius, M. Sterrer and H. J. Freund, Acc. Chem. Res., 2008, 41, 949; (c) L. Kesavan, R. Tiruvalam, M. H. Ab Rahim, M. I. bin Saiman, D. I. Enache, R. L. Jenkins, N. Dimitratos, J. A. Lopez-Sanchez, S. H. Taylor, D. W. Knight, C. J. Kiely and G. J. Hutchings, Science, 2011, 311, 195.
10. L. C. Wang, L. He, Q. Liu, Y. M. Liu, M. Chen, Y. Cao, H. Y. He and K. N. Fan, Appl. Catal., A, 2008, 344, 150.
11. (a) G. C. Bond and D. T. Thompson, Gold Bull., 2000, 33, 41; (b) S. Y. Yu, H. Huang, H. B. Liu, Z. N. Chen, R. Zhang and M. Fujita, Angew. Chem., Int. Ed., 2003, 42, 686; (c) S. Carrettin, P. Concepcion, A. Corma, J. M. Lopez Mieto and V. F. Puntes, Angew. Chem., Int. Ed., 2004, 43, 2538; (d) D. Matthey, J. G. Wang, S. Wendt, J. Matthiesen, R. Schaub, E. Lægsgaard, B. Hammer and F. Besenbacher, Science, 2007, 315, 1692; (e) N. Weiher, A. M. Beesley, N. Tsapatsaris, L. Delannoy, C. Louis, J. A. van Bokhoven and S. L. M. Schroeder, J. Am. Chem. Soc., 2007, 129, 2240; (f) Q. Fu, H. Saltsburg and M. Flytzani-Stephanopoulos, Science, 2003, 301, 935; (g) A. Trovarelli, Catal. Rev. Sci. Eng., 1996, 38, 439.
12. (a) J. Guzman and B. C. Gates, J. Am. Chem. Soc., 2004, 126, 2672; (b) M. K. Oudenhuijzen, J. A. van Bokhoven, J. T. Miller, D. E. Ramaker and D. C. Koningsberger, J. Am. Chem. Soc., 2005, 127, 1530; (c) D. E. Ramaker, M. Teliska, Y. Zhang, A. Stakheev and D. C. Koningsberger, Phys. Chem. Chem. Phys., 2003, 5, 4492; (d) D. E. Ramaker and D. C. Koningsberger, Phys. Rev. Lett., 2002, 89, 139701; (e) J. A. van Bokhoven, C. Louis, J. T. Miller, M. Tromp, O. V. Safonova and P. Glatzel, Angew. Chem., Int. Ed., 2006, 45, 4651; (f) J. T. Miller, A. J. Kropf, Y. Zha, J. R. Regalbuto, L. Delannoy, C. Louis, E. Bus and J. A. van Bokhoven, J. Catal., 2006, 240, 222.
13. (a) T. Mallat and A. Baiker, Chem. Rev., 2004, 104, 3037; (b) R. A. Sheldon, I. Arends, G. J. Ten Brink and A. Dijksman, Acc. Chem. Res., 2002, 35, 774.

---

Footnote

† Electronic supplementary information (ESI) available: Catalyst preparation, testing and characterization details; additional XRD patterns, TEM images, XPS spectra, XANES spectra; data analysis and discussion; Fig. S1–S10; Tables S1–S4; Scheme S1; supplementary discussion 1 to 3. See DOI: 10.1039/c2cy20579c

---