Direct methylation of N-methylaniline with CO₂/H₂ catalyzed by gold nanoparticles supported on alumina

Abstract

Small gold nanoparticles (∼3 nm) loaded onto various supports have been prepared by a deposition–precipitation method and studied for direct methylation of N-methylaniline with CO₂/H₂. Among the catalysts examined, an acid–base bifunctional support γ-alumina supported gold catalyst (Au/γ-Al₂O₃) exhibits the best catalytic performance. Au/γ-Al₂O₃ catalysts with controlled mean Au particle sizes (1.8–8.3 nm) have also been successfully prepared by regulating the concentration of HAuCl₄ in solution, aging time, aging temperature and mole ratio of urea to gold in the process of deposition–precipitation with urea. The turnover frequency (TOF) values for direct methylation of N-methylaniline with CO₂/H₂ increase on decreasing the mean size of Au nanoparticles (from 8.3 to 1.8 nm), showing that methylation of N-methylaniline with CO₂/H₂ is a structure-sensitive reaction. A fast increase in TOF occurs when the mean Au particle size becomes smaller than 3 nm. Through TEM (transmission electron microscope), gold L₃-edge XAFS (X-ray absorption fine structure) and CO₂- and NH₃-TPD (temperature programmed desorption) analysis, we can conclude the Au particle sizes, oxidation state of the gold species and acid–base properties of the supports are responsible for the high catalytic activity of direct methylation of N-methylaniline with CO₂/H₂.

---

Introduction

Carbon dioxide (CO₂) is a greenhouse gas and the accumulation of a huge amount of CO₂ has caused significant and negative effects on the global environment.¹ It is crucial, therefore, to take effective measures to reduce CO₂ emission and utilize CO₂ resources. In this regard, carbon dioxide has been used as an inexpensive and nontoxic C₁ feedstock for producing commodity chemicals, fuels and various materials in the last few decades.^2–4 Owing to its thermodynamic stability, highly efficient activation of CO₂ and a strong thermodynamic driving force are required for efficient transformation of CO₂. In this respect, only a handful of industrial processes have been realized for CO₂ utilization, such as the production of urea, salicylic acid, and carbonates.⁵ Additionally, the catalytic hydrogenation of CO₂ to methanol, formic acid or its derivatives have been reported by using various catalysts.^6,7 Carbon dioxide can also be converted into cyclic carbonates,⁸ dimethyl carbonate (DMC),⁹ formamides¹⁰ and carboxylic acids,¹¹ activated by relevant reactive substrates such as epoxides, alcohols, amines and olefins respectively.

Methylamines, including dimethylamine, trimethylamine, methylamine and dimethylformamide (DMF), have been usually used as platform chemicals for preparing fertilizers, fungicides, synthetic leathers or polymers.¹² Hazardous alkylating agents, such as methyl iodide, dimethyl sulfate, and dimethyl carbonate, are traditionally preferred for the preparation of methylamines.¹³ Meanwhile, dimethylsulfoxide and formic acid have also been found to be the available carbon sources for methylation of amines by catalysts.¹⁴ Great efforts on the other hand have been made to develop new catalytic process for preparation of N-methylated compounds using CO₂ as a C₁ building block in the absence of the toxic alkylating agents. In early 2013, two catalytic systems were independently reported by two research groups for the methylation of amines with CO₂ and silanes by homogeneous catalysts, producing siloxane as a by-product.¹⁵ Shortly afterwards, a green method for methylation of amines with CO₂/H₂ has been reported using homogeneous Ru-based catalysts by the groups of Klankermayer and Beller, generating water as the only by-product.¹⁶ Lately, two new metal-free homogeneous catalysts for methylation of amines using CO₂ as a C₁ source, were developed by the groups of Dyson and Cantat.¹⁷ Despite the significant progress that has been made in the area of amine methylation using CO₂/H₂ with homogeneous systems, there are still few available relevant reports on the heterogeneously catalyzed methylation of amines with CO₂/H₂. Recently, N-methylation of amines with CO₂/H₂ was realized by heterogeneous catalysts such as CuAlO_x,¹⁸ Pt–MoO_x/TiO₂ (ref. 19) and Pd/CuZrO_x (ref. 20) respectively. These heterogeneous systems however are limited by their low turnover frequencies (TOF, <3.3 h⁻¹) and prolonged reaction times (24–48 h) for achieving the high product yields (>75%).

Catalysts based on supported gold nanoparticles (NPs) have been found to be versatile catalysts for many chemical transformations in various mild chemical processes, since Haruta and his associates discovered gold particles (<5 nm) supported on transition metal oxides were very active for CO oxidation at low temperature.²¹ It is generally accepted that the mean size of gold NPs, supports and preparation methods markedly influence the catalytic performances of supported gold catalysts.²² Bus et al. reported that small gold particles (size < 2 nm) loaded on γ-Al₂O₃ were active for selective hydrogenation of cinnamaldehyde to cinnamyl alcohol.²³ It was previously reported that both size and support of supported gold catalyst affected the reaction rates for the water–gas shift reaction.²⁴ Sun et al. showed that basic resin R201 supported Au catalyst displayed good performance in the one-pot synthesis of styrene carbonate from styrene and CO₂.²⁵ Recently, Liu et al. has reported the effective synthesis of benzimidazoles from 2-nitroanilines and CO₂/H₂ using supported gold catalyst.²⁶ Besides, as reported by the group of Manzoli, gold sites were essential for hydrogen dissociation on metal oxide supports, which could activate hydrogen.²⁷ These established knowledge provide inspiration for our current study of methylation of N-methylaniline using CO₂/H₂ with supported gold catalysts.

We previously reported that Au/Al₂O₃-VS (very small gold NPs loaded on γ-Al₂O₃) catalyzed direct methylation of aniline using CO₂/H₂ with total conversion of aniline and excellent selectivity towards N,N-dimethylaniline (92%) at 140 °C for 7 h.²⁸ To the best of our knowledge, it is the highest TOF value ever reported to date for methylation of aniline with CO₂/H₂ in this catalysis system based total gold atoms (26 h⁻¹). Small Au NPs (∼2 nm) supported on alumina is critical for high activity for methylation of amines using CO₂/H₂, in comparison with Au NPs supported on other supports.²⁸ We herein extend the previous work, and systematically study the effects of the mean size of Au NPs, gold species and acid–base properties of supports on catalytic activity of Au/Al₂O₃ catalyst for direct methylation of N-methylaniline with CO₂/H₂. We have also examined the mechanism of methylation of N-methylaniline in the presence of CO₂/H₂ catalyzed by Au/Al₂O₃ catalyst.

Experimental

Chemicals

Commercially available organic and inorganic compounds were used without further purification. Gold catalysts including 1 wt% Au/TiO₂ (catalogue number 79–0165), 1 wt% Au/ZnO (catalogue number 79–0170) were supplied by Mintek.

Synthesis of Al₂O₃, Mn₂O₃ and Co₃O₄ supports

Al₂O₃ powders were prepared by a conventional precipitation method. 45.0 g Al(NO₃)₃·9H₂O was dissolved in 500 mL deionized water at room temperature, the pH was adjusted to 9.0 by dropwise addition of NH₄OH (2.5 M). The resultant hydro gel was washed with deionized water until free of NO³⁻ ions. The precipitate was then dried at 110 °C overnight and calcined at 500 °C for 4 h in air. Then the particles were passed a 200-mesh sieve to get the qualities of particles with diameters of less than 0.074 millimeter. The crystal structure of Al₂O₃ was gamma phase (based on XRD analysis, see Fig. S1†).

Mn₂O₃ powders were prepared by calcination of MnCO₃ sample (AR) at 500 °C in flowing air (1 atm) for 4 h and then cooled.²⁹

Co₃O₄ powders were prepared by a conventional precipitation method. 10.0 g Co(NO₃)₂·6H₂O was dissolved in 200 mL deionized water at room temperature, the pH was adjusted to 9.0 by dropwise addition of NH₄OH (2.5 M). The resultant hydro gel was washed with deionized water and dried at 110 °C overnight. The product was calcined at 400 °C for 4 h in air and then cooled.

Catalyst synthesis

The theoretical Au loading for each catalyst used in this work was 1 wt%, unless otherwise stated. Au/SiO₂ was prepared according to sonication-aided impregnation method reported by Wang et al.³⁰ The calculated amount of SiO₂ (Degussa, Aerosil 380, specific surface area: 380 m² g⁻¹) was suspended in an aqueous solution of HAuCl₄, and the suspension was placed into the sonication bath with irradiation at 40 kHz and 200 W output power. After sonication, the water was removed by evaporation at 80 °C. The solid powder was finally reduced in H₂ atmosphere at 350 °C for 2 h. The deposition–precipitation with urea method was used for the preparation of Au/Al₂O₃, Au/MgO, Au/Co₃O₄, Au/Mn₂O₃ and Au/CeO₂ catalysts. In a typical procedure, urea was dissolved in HAuCl₄ solution at room temperature. The support was then added to this clear solution, and the temperature of the resulting slurry was increased gradually to a fixed temperature. The temperature was maintained for a certain time. After cooling to room temperature, the solid was recovered by filtering and washing several times with distilled water. The sample was dried under vacuum at 50 °C for 12 h, followed by reduction with a stream of 5 vol% H₂/Ar at 350 °C for 2 h. The standard conditions were as follows: [HAuCl₄] = 0.48 mmol L⁻¹; aging time = 6 h; aging temperature = 80 °C; mole ratio of urea to Au = 200. The concentration of gold by weight was confirmed by ICP-AES.

Characterization

Au loadings in all of the catalysts were measured by ICP-AES with an ICAP 6300 instrument when the sample was dissolved in aqua regia and diluted with distilled water to the proper concentration. The crystal structures of the prepared catalysts were characterized with powder X-ray diffraction (XRD) on a Bruker D8 Advance X-ray diffractometer using the Ni-filtered Cu Kα radiation source at 40 kV and 40 mA. TEM images for supported gold catalysts were taken with a Tecnai G2 electron microscope operating at 200 kV. Before being transferred into the TEM chamber, the samples dispersed with ethanol were deposited onto a carbon-coated copper grid and then quickly moved into the vacuum evaporator. The size distribution of the metal nanoclusters was determined by measuring about 200 random particles on the images. NH₃-TPD and CO₂-TPD were performed on an AutoChem 2950 HP instrument. Typically, the sample loaded in a quartz reactor was pretreated with high purity He at 300 °C for 1 h. After cooling the sample to 50 °C, NH₃ adsorption was performed by switching the He flow to a NH₃–He (10 vol% NH₃) gas mixture and then maintaining the sample at 50 °C for 1 h. The gas-phase (and/or weakly adsorbed) NH₃, was purged by high-purity He at the same temperature. NH₃-TPD was then performed in the He flow by raising the temperature to 700 °C at a rate of 10 °C min⁻¹. The desorbed NH₃ molecules were detected by using a MKS Cirrus 2 mass spectrometer at the signal of m/z 17. CO₂-TPD was performed by using a similar procedure. The X-ray absorption data at the Au L-edge of the samples were recorded at room temperature in transmission mode using ion chambers at beam line BL14W1 of the Shanghai Synchrotron Radiation Facility (SSRF), China. The synchrotron radiation was monochromatized with a Si (111) double crystal monochromator, and mirrors were used to eliminate higher harmonics. The incident and transmitted beam intensities were monitored using ionization chambers filled with pure N₂. During the measurement, the synchrotron was operated at energy of 3.5 GeV and a current of 250 mA. Data processing was performed using the program ATHENA and ARTEMIS. The X-ray absorption near-edge structure (XANES) was normalized with edge height and the first-order derivatives was taken to compare variation of absorption edge energies.

Catalytic testing

A mixture of solvent (10 mL), N-methylaniline (1.0 mmol), and supported Au catalysts (Au 2.7–5.0 μmol) were charged into a 50 mL Hastelloy-C high pressure Parr reactor. The reactor was then exchanged with H₂, followed by introducing 1.0 MPa CO₂ and 3.0 MPa H₂. The reaction was reacted at desired temperature for given reaction time under magnetic stirring. Subsequently, the autoclave was cooled to room temperature, and n-octane (1 mmol, internal standard) was added for quantitative analysis by GC-FID (Agilent 7820A).

Results and discussion

Solvent and support effects

In the course of identifying and optimizing key reaction parameters for N-methylaniline (1a) with CO₂/H₂, several solvents were investigated. The Au/γ-Al₂O₃ catalyst with very small Au NPs (2.0 nm) was used as the standard catalyst. The highest yield of N,N-dimethylaniline (1b) was obtained when cyclohexane was used as the reaction medium (Table 1, entry 1). In contrast, the reaction in hexane, toluene, 1,4-dioxane, CH₂Cl₂ and THF (Table 1, entries 2–6) gave moderate to good yield (24–61%) with excellent selectivity towards 1b (>90%), and in ethanol (Table 1, entry 7) obtained the lowest yield (2%) with poor selectivity towards 1b (18%). Based on above results, we empirically concluded that the use of apolar solvents for direct methylation of 1a with CO₂/H₂ gave better conversion than the use of polar solvents, which shows that the polarity of solvents has a remarkable influence on the conversion of 1a with CO₂/H₂ in the presence of Au/Al₂O₃ catalyst.

Table 1 Effects of different solvent and support on the synthesis of N,N-dimethylaniline with CO₂/H₂^a

Entry	Solvent	Catalyst	Au Size^b (nm)	Au content^c (wt%)	Conv.^d (%)	Sel.^e (%)
a Reaction conditions: 0.5 mol% Au, N-methylaniline (1 mmol), CO2 (1 MPa), H2 (3 MPa), solvent (10 mL), T = 140 °C, t = 5 h, n. r. = no reaction.b Obtained from TEM.c Measured by ICP-AES.d Conversation and selectivity were determined by GC with n-octane as the internal standard.e N-methylformanilide is the only other product.
1	Cyclohexane	Au/Al2O3	2.0	0.73	76	>99
2	Hexane	Au/Al2O3	2.0	0.73	61	>99
3	Toluene	Au/Al2O3	2.0	0.73	52	>99
4	1,4-Dioxane	Au/Al2O3	2.0	0.73	25	96
5	CH2Cl2	Au/Al2O3	2.0	0.73	34	>99
6	THF	Au/Al2O3	2.0	0.73	29	91
7	Ethanol	Au/Al2O3	2.0	0.73	12	18
8	Cyclohexane	Au/SiO2	3.7	0.97	1	>99
9	Cyclohexane	Au/α-Mn2O3	2.4	0.95	1	>99
10	Cyclohexane	Au/CeO2	3.7	0.74	3	>99
11	Cyclohexane	Au/Co3O4	4.4	0.96	7	42
12	Cyclohexane	Au/TiO2	2.8	1.00	15	>99
13	Cyclohexane	Au/MgO	2.6	0.91	n.r.	—
14	Cyclohexane	Au/ZnO	2.8	1.00	n.r.	—
15	Cyclohexane	Au/C	4.0	1.00	n.r.	—
16	Cyclohexane	Au powder	150	100	n.r.	—

---

Subsequently, the effects of some typical supports on catalytic activity were studied for methylation of 1a with CO₂/H₂. The mean size of Au NPs loaded on various supports at around 3 nm was successfully obtained by using the DPU method and sonication-aided impregnation techniques. The typical TEM micrographs and corresponding Au particle size distributions for these catalysts are shown in Fig. S2.† Of the different supports tested, Al₂O₃ (Table 1, entry 1) provided the best result in terms of high conversion (76%) and excellence selectivity (>99%). The use of SiO₂, α-Mn₂O₃, CeO₂ and Co₃O₄ resulted in low yields of 1b (<4%) under similar reaction conditions (Table 1, entries 8–11). Moderate yield of 1b (15%) could be obtained when Au/TiO₂ catalyst was used (Table 1, entry 12). However, 1b was not detected when MgO, ZnO and C were used as supports (Table 1, entries 13–15). In addition, the use of the Au⁰ powder, did not promote the reaction at all (Table 1, entry 16). It was reported that MgO has strong basic character and SiO₂ is an acidic to neutral oxide.³⁰ Al₂O₃ with moderate electronegativity can be considered to have both acidic and basic surface sites.³¹ These results may suggest that the acid–base properties of the support can affect the catalytic behavior of the supported Au catalysts for methylation of 1a with CO₂/H₂.

Next, the acid–base properties of the typical supported Au catalysts were examined by temperature-programmed desorption (TPD) of NH₃ and CO₂. As shown in Fig. 1, desorption peaks of NH₃ are observed from Au/Al₂O₃, Au/Mn₂O₃, Au/TiO₂, Au/ZnO, Au/CeO₂ and Au/SiO₂, which indicates that acidic sites exist on these catalysts. From the temperature and intensity of the NH₃ desorption peak, we can see that Au/Al₂O₃ possesses relatively strong acidity. No significant acidity, on the other hand, exists on Au/MgO, Au/C and Au/Co₃O₄, as no desorption peak of NH₃ can be detected from these catalysts. CO₂-TPD profiles in Fig. 2 suggest that Au/Al₂O₃ and Au/MgO both have relatively strong basic sites, whereas no basic sites can be observed on Au/TiO₂, Au/Co₃O₄, Au/C, Au/ZnO and Au/SiO₂. Weak desorption peaks of CO₂ are detected in Au/Mn₂O₃ and Au/CeO₂, which suggests weak basic sites exist on these catalysts. According to the NH₃-TPD and CO₂-TPD results, Au/Al₂O₃ possesses both relatively strong basicity and acidity, while Au/Co₃O₄ and Au/C have neither acidic nor basic sites. Therefore, we have observed significant support effects in the Au-catalyzed methylation of N-methylaniline with CO₂/H₂, and our studies suggest that the acid–base properties of the supports play important roles in determining both its activity and the selectivity. The strong basicity of supports may promote the adsorption of CO₂ and accelerate departure of product from the surface of catalysts. The strong acidity of supports may be important to anchor the amines (reactive molecules). So the strong acidity and basicity will benefit the methylation reaction.

Fig. 1 NH₃-TPD profiles for some typical supported Au catalysts: (a) Au/CeO₂, (b) Au/C, (c) Au/ZnO, (d) Au/Co₃O₄, (e) Au/Mn₂O₃, (f) Au/SiO₂, (g) Au/MgO, (h) Au/TiO₂, (i) Au/Al₂O₃.

Fig. 2 CO₂-TPD profiles for some typical supported Au catalysts: (a) Au/C, (b) Au/ZnO, (c) Au/Co₃O₄, (d) Au/CeO₂, (e) Au/Mn₂O₃, (f) Au/SiO₂, (g) Au/MgO, (h) Au/TiO₂, (i) Au/Al₂O₃.

In summary, acidity and basicity of supports in some typical supported gold catalysts were successfully detected in our experiment and had significant effects on conversion of 1a and selectivity towards 1b during direct methylation of 1a with CO₂/H₂. The Au/Al₂O₃ catalyst with relatively strong acidity and basicity contributes to high conversion of 1a and excellent selectivity towards 1b, whereas other catalysts with only strong basic (Au/MgO) or acidic to neutral sites (Au/SiO₂ and Au/C) are almost inactive. Catalysts with weak acid–base properties (Au/Mn₂O₃ and Au/ZnO) also have no significant activity. There are moderate activity and excellent selectivity when TiO₂ and CeO₂ with a few acidic sites are used as a support in Au/TiO₂ and Au/CeO₂ catalysts. Au/Co₃O₄ with neutral sites has very low activity and poor selectivity (<45%) towards 1b.

A further study to investigate the effect of reaction parameters was carried out in our experiment. As shown in Fig. 3, the methylation of 1a with CO₂ and H₂ can be carried out smoothly under relatively mild conditions although the reaction time becomes longer (18 h). It is noteworthy that the selectivity is constant (>99%) during the process, which suggests the reaction has high selectivity towards 1b. Besides, increasing the reaction temperature and pressure respectively can boost the conversion of 1a to a certain extent (see Fig. S3 and Fig. S4 in the ESI†). However, the conversion of 1a and selectivity towards 1b kept nearly the same when the total initial pressure and temperature were higher than 4 MPa and 140 °C respectively. Obviously, 1a could not be fully converted by increasing the total initial pressure or boosting reaction temperature under the present reaction conditions. Nevertheless, 1a could be fully converted with excellent selectivity towards 1b when reaction time was up to 18 h under the standard conditions. Therefore, total initial pressure and temperature both affected the yield of 1b when the reaction time was fixed at 5 h with the yield of 1b lower than 76%. In order to get 1a totally converted, prolonged reaction time is required under proper total initial pressure and at appropriate temperature.

Fig. 3 Effects of reaction time on catalytic activity of Au/Al₂O₃ catalyst for direct methylation of N-methylaniline with CO₂/H₂. Reaction condition: 0.5 mol% Au, 1.0 mmol N-methylaniline, cyclohexane 10 mL, CO₂ (1 MPa), H₂ (3 MPa), T = 140 °C, t = 5 h.

Preparation and characterization of Au/γ-Al₂O₃ with controlled mean Au particle sizes

There are still challenges to prepare supported Au catalysts with controlled Au sizes. It was found that aging temperature, aging time and the mole ratio of urea to Au used in DPU process played key roles to obtain controlled Au particle sizes. Louis et al. reported that the pH increased with time due to urea decomposition and most gold was deposited within the first hour in the preparation of Au/TiO₂ by DPU method.³² Urea begins to decompose when aging temperature is over 60 °C. Compared with Au/TiO₂ prepared by DPU, we found that by regulating aging temperature from 20 to 80 °C in the DPU process could markedly decrease the mean size of Au NPs from 8.3 to 2.0 nm of Au/Al₂O₃ catalysts (Table 2, entries 1–2 and 5). These results clarified that aging temperature used during the DPU process could be an important factor in determining the mean size of Au NPs supported on alumina. As is well known, the rates of nucleation and growth determine the size of a particle formed in the solution. Increasing the aging temperature from 20 to 80 °C can increase the nucleation rates and the pH values in solution because of urea decomposition. At 80 °C, when the mole ratio of urea to Au decreased to 50, the pH value could not reach at 8 as too little urea was added in solution. So a large Au particle size (5.7 nm) was formed (Table 2, entry 3). When aging time was 1 h instead of 6 h, large mean Au NPs size (3.6 nm) was obtained (Table 2, entry 4), indicating that less aging time caused the formation of large Au particle. Alternatively, prolonged aging time favored the formation of small Au particles in the DPU process.³² However, pH value was over 8 when urea added in solution was enough, which produced the smallest mean Au NPs size (1.8 nm) (Table 2, entry 6). Fig. 4 shows the TEM micrographs and the Au particle size distributions for these Au/Al₂O₃ samples. Grisel et al. reported the Au loading depended strongly on the final pH of the precipitation solution and a sharp decrease in Au loading was observed with pH value over 8.5.³³ Low Au loadings (yields: <50%) due to incomplete decomposition of urea in solution were obtained when aging temperatures were below 80 °C (Table 2, entries 1–2). Most Au was loaded on alumina (yield: >70%) when the mole ratio of urea to gold was less than or equal to 200 at 80 °C (Table 2, entries 3–5). Too much urea added in solution on the other hand caused low Au loading (yield: 40%) because of high pH value over 8.5 (Table 2, entry 6). Thus, regulating pH of the solution during DPU process contributed to the formation of small Au particle and could cause the alteration of Au loadings.

Table 2 Preparations by deposition–precipitation with urea: influence of aging temperature, aging time and the mole ratio of urea to Au on the mean Au NPs size^a

Entry	Aging temperature (°C)	Aging time (h)	Mole ratio of urea : Au (mol : mol)	Au loading^b (wt%)	Mean diameter^c (nm)
a If not mentioned, preparation conditions were: [HAuCl4] = 0.48 mmoL L^−1, aging time = 6 h, aging temperature = 80 °C, mole ratio of urea to Au = 200, theoretical Au loading = 1 wt%.b Measured by ICP-AES.c Obtained from TEM.
1	20	6	200	0.47	8.3
2	50	6	200	0.46	7.2
3	80	6	50	0.71	5.7
4	80	1	200	0.70	3.6
5	80	6	200	0.73	2.0
6	80	6	400	0.40	1.8

---

Fig. 4 TEM micrographs and corresponding Au NPs size distributions for the Au/Al₂O₃ catalysts with various mean sizes of Au NPs: (a) 1.8 nm, (b) 2.0 nm, (c) 3.6 nm, (d) 5.7 nm, (e) 7.2 nm, and (f) 8.3 nm.

Catalytic properties of Au/γ-Al₂O₃ catalysts with various Au NPs sizes for direct methylation of N-methylaniline with CO₂/H₂

The dependence of catalytic activity on the mean size of Au NPs in Au/Al₂O₃ catalysts prepared by DPU with various preparation parameters was shown in Table 3. The highest yield (83%) of 1b (Table 3, entry 1) was acquired with the smallest mean Au NPs size (1.8 nm). Small Au NPs (2.0–5.7 nm) could lead to moderate yields (61–76%) and high selectivity (>99%) towards 1b (Table 3, entries 2–7). However, on the other hand, low yields (<8%) of 1b (Table 3, entries 8–9) were obtained due to large mean size of Au NPs (>7.2 nm).

Table 3 Catalytic performance of Au/Al₂O₃ catalysts with various mean Au NPs size from 1.8 to 8.3 nm^a

Entry	Preparation conditions^b	Au loading (wt%)	Au size^c (nm)	Conv.^d (%)	Sel.^e (%)
a Reaction conditions: catalyst 0.14 g (Au 2.7–5.0 μmol), N-methylaniline (1 mmol), CO2 (1 MPa), H2 (3 MPa), cyclohexane (10 mL), T = 140 °C, t = 5 h.b If not mentioned, preparation conditions were: [HAuCl4] = 0.48 mmol L^−1, aging time = 6 h, aging temperature = 80 °C, mole ratio of urea to Au = 200.c Obtained from TEM.d Conversation and selectivity were determined by GC with n-octane as the internal standard.e N-Methylformanilide is the only other product.
1	Urea : Au = 400	0.40	1.8	83	>99
2	Standard condition	0.73	2.0	76	>99
3	[HAuCl4] = 0.12 mmol^−1 L	0.70	2.4	69	>99
4	Aging time = 16 h	0.65	2.4	70	>99
5	[HAuCl4] = 1.92 mmol^−1 L	0.60	2.6	63	>99
6	Aging time = 1 h	0.70	3.6	69	>99
7	Urea : Au = 50	0.70	5.7	61	>99
8	Aging temperature = 50 °C	0.46	7.2	8	86
9	Aging temperature = 20 °C	0.47	8.3	6	83

---

The conversion of 1a at the initial stages over Au/Al₂O₃ catalysts was recorded to prove the relations between mean Au NPs size and the intrinsic catalytic reactivity of Au sites. The time course of Au/Al₂O₃ catalysts was done with different mean Au NPs sizes in the range of 1.8 to 8.3 nm (see Fig. S5 in the ESI†). No conversion of 1a over Al₂O₃ (Fig. S5†) indicates that the Al₂O₃ species had no active sites for the methylation reaction with CO₂/H₂, which shows only the reduced gold species possess active sites for methylation of 1a with CO₂/H₂. The conversation of 1a increases almost linearly with reaction time over the initial 30 min in Fig. S5.† So we can calculate initial reaction rates by using the slope of the straight line in Fig. S5† for each catalyst. The results are shown in Table S1.† Obviously, the initial reaction rates decrease from 5.86 to 0.36 mmol h⁻¹ g⁻¹ (cat.) with increasing mean size of Au NPs from 1.8 to 8.3 nm.

Because the gold atoms contribute to the catalytic reactions, we have calculated the TOF (the moles of 1a converted per unit time at the initial stage per mole total gold atoms) using the initial reaction rate for each Au/Al₂O₃ catalyst. The results are shown in Fig. 5, suggesting the intrinsic activity per mole total gold atoms decreases with increasing the mean Au NPs size and the methylation of 1a with CO₂/H₂ is a structure-sensitive reaction.

Fig. 5 Dependence of TOF on the mean Au NPs size for the Au/Al₂O₃-catalyzed methylation of N-methylaniline with CO₂/H₂. Reaction conditions: catalyst 0.14 g (Au 2.7–5.0 μmol), 1.0 mmol N-methylaniline, cyclohexane 10 mL, CO₂ (1 MPa), H₂ (3 MPa), T = 140 °C.

The TOF values increase from 114 to 287 h⁻¹ with decreasing the mean size of Au NPs from 3.6 to 1.8 nm. When the mean sizes of Au NPs are larger than 5.7 nm, the TOF values are lower than 52 h⁻¹. The TOF value (about 300 h⁻¹) with the best catalyst (with 1.8 nm gold size) is much higher than those measured with other heterogeneous catalysts such as CuAlO_x (0.1 h⁻¹), Pt–MoO_x/TiO₂ (1.8 h⁻¹) and Pd/CuZrO_x (3.3 h⁻¹) for methylation of amines with CO₂/H₂.^18–20

Effects of reduction temperature of Au/Al₂O₃ catalysts on direct methylation of N-methylaniline with CO₂ and H₂

The as-prepared supported gold precursors are easily reduced to the zero state by thermal treatments with reducing gases or oxidant gases.³⁴ We previously demonstrated that the very small Au⁰ NPs supported on alumina treated with a stream of 5 vol% H₂/Ar at 350 °C for 2 h showed high catalytic activity for direct methylation of amines with CO₂/H₂.²⁸ In our experiment, the as-prepared Au/Al₂O₃ precursor of standard Au/Al₂O₃ catalyst was treated with a stream of 5 vol% H₂/Ar at 150, 250, 350 and 450 °C respectively for 2 h, at a heating rate of 5 °C min⁻¹ to obtain Au/Al₂O₃ samples. The samples were designated as Au/Al₂O₃-T, where T denoted the reduction temperature. With these catalysts, the catalytic activity was investigated for direct methylation of amines with CO₂/H₂. Table 4 shows that selective production of 1b in a 38% yield could be achieved when Au/Al₂O₃-150 catalyst was used (Table 4, entry 1). An increase in reduction temperature from 150 to 250 °C markedly increased the 1b yield from 38 to 53% (Table 4, entry 2). The highest yield of 1b was 76% as reduction temperature of the as-prepared Au/Al₂O₃ precursor was 350 °C (Table 4, entry 3). However, a further increase of reduction temperature from 350 to 450 °C led to a slight decrease in the desired product yield from 76 to 67% (Table 4, entries 3–4). The high reduction temperature may lead to decrease of the quantities of active hydroxyl group in alumina surface and a weak Au-support interaction in Au/Al₂O₃-450 catalyst, which may cause the slight decrease of the catalytic activity.³⁵ Notably, the selectivity towards 1b was almost the same, showing that the selectivity did not depend on reduction temperature of the as-prepared Au/Al₂O₃ precursor.

Table 4 Effects of reduction temperature on catalytic performance of Au/Al₂O₃ catalyst for direct methylation of N-methylaniline with CO₂/H₂^a

Entry	Reduction temperature (°C)	Au size^b (nm)	Conv.^c (%)	Sel. (%)
a Reaction conditions: 0.5 mol% Au, N-methylaniline (1 mmol), CO2 (1 MPa), H2 (3 MPa), cyclohexane 10 mL, T = 140 °C, t = 5 h.b Obtained from TEM.c Conversation and selectivity were determined by GC with n-octane as the internal standard.
1	150	7.1	38	>99
2	250	4.5	53	>99
3	350	2.0	76	>99
4	450	1.6	67	>99

---

The factor, which caused different catalytic activities of the Au/Al₂O₃ catalysts treated by different reduction temperature, could be the change of Au particle size and valence of Au species. To investigate the structure of Au species supported on alumina after the H₂ reduction, Au L₃-edge XAFS measurement of Au/Al₂O₃ samples and Au reference compounds was carried out. Fig. 6 shows Au L₃-edge X-ray absorption near-edge structures (XANES) spectra, which are known to be sensitive to the oxidation states of X-ray absorbing atoms. The XANES features of Au/Al₂O₃ samples (spectra c–f) after the H₂ reduction with various temperature from 150 to 450 °C are clearly different from that of Au₂O₃ (spectrum a) but rather similar to that of Au foil (spectrum g), indicating that gold species in these samples are in a reduced state.

Fig. 6 Au L₃ absorption edge XANES spectra: (a) Au₂O₃, (b) as prepared Au/Al₂O₃ precursor, Au/Al₂O₃ samples reduced at different temperatures: (c) 150, (d) 250, (e) 350, (f) 450 °C, and (g) Au foil.

Fig. 7 shows the Fourier transform (FT) of extended X-ray absorption fine structure (EXAFS) data of Au/Al₂O₃ samples reduced at different temperature (150, 250, 350, 450 °C), the Au/Al₂O₃ precursor, Au foil and Au₂O₃. The structural parameters derived from curve-fitting analysis are listed in Table 5. The Au–O coordination number of 3.8 was observed in pure Au₂O₃. The EXAFS of the as-prepared Au/Al₂O₃ precursor (spectrum b) showed the Au–O coordination number (CN) and the Au–Au CN were 1.7 and 3.0 respectively. These results, as reported previously by Grünert et al. for Au/TiO₂-MCM-48 catalyst,³⁶ confirmed that metallic gold species as a minority phase might exist in the as prepared Au/Al₂O₃ precursor due to auto reduction. After H₂ reduction at various temperature in the range of 150 to 450 °C for 2 h (spectra c–f), the Au–Au CN decreased from 10.7 to 5.6 and the Au–Au bond length decreased from 2.85 to 2.77 Å. It should be noted that there is a significant Au–O path contribution for Au/Al₂O₃-T, such as Au/Al₂O₃-450. However, this Au–O path has longer length (2.11 ± 0.06 Å) than the Au–O bond in Au₂O₃, implying a weak interaction, different from oxidic Au atoms by oxygen, between Au NPs and Al₂O₃ support. It implies that the oxygen of Au–O bond comes from Al₂O₃ support. This result can be further suggested from the big uncertainly (±0.06 Å) of the Au–O path length and the absence of Au oxidation determined by above XANES data. The Au–Au CN in these samples were much less than that in Au foil while close to that in Au/Fe–OH or Au/Fe–O catalysts due to the high proportion of surface atoms, suggesting that the small gold NPs supported on alumina were obtained.³⁷ TEM measurement in Fig. S6† showed that the mean gold NPs size remarkably decreased from 7.1 to 1.6 nm with increasing reduction temperature from 150 to 450 °C. So the TEM result is in good agreement with the XAFS conclusion.

Fig. 7 Au L₃ absorption edge EXAFS Fourier transforms: (a) Au₂O₃, (b) the Au/Al₂O₃ precursor, Au/Al₂O₃ samples reduced at different temperatures: (c) 150, (d) 250, (e) 350, (f) 450 °C, and (g) Au foil.

Table 5 Fitting parameters of the curve fitted k³-weighted EXAFS analysis of the Au/Al₂O₃-T samples, the Au/Al₂O₃ precursor, Au foil and Au₂O₃

Samples^a	Shell	N^b	R^c (Å)	σ^2^d (×10^−3 Å^2)	ΔE0 (eV)
a Reduction temperature of catalysts.b Coordination number.c Bond length.d Debye–Waller factor.
Au-foil	Au–Au	11.9 ± 0.5	2.86 ± 0.01	8.1 ± 0.3	3.9
Au/Al2O3-150	Au–O	0.5 ± 0.2	2.08 ± 0.03	3.0 ± 1.0	−7.0
	Au–Au	10.7 ± 0.9	2.85 ± 0.01	8.8 ± 0.6	3.8
Au/Al2O3-250	Au–O	0.5 ± 0.2	2.08 ± 0.03	3.0 ± 1.0	−6.9
	Au–Au	10.0 ± 1.2	2.83 ± 0.01	10.1 ± 1.2	3.5
Au/Al2O3-350	Au–O	0.8 ± 0.3	2.09 ± 0.03	3.0 ± 1.5	−3.4
	Au–Au	7.7 ± 0.8	2.82 ± 0.01	10.0 ± 1.3	4.3
Au/Al2O3-450	Au–O	0.9 ± 0.3	2.11 ± 0.06	3.5 ± 1.2	−2.0
	Au–Au	5.6 ± 1.4	2.77 ± 0.01	8.9 ± 1.5	0.5
As prepared precursor	Au–O	1.7 ± 0.4	2.02 ± 0.01	5.0 ± 3.0	11.8
	Au–Au	3.0 ± 1.0	2.83 ± 0.01	5.3 ± 2.0	1.8
Au2O3	Au–O	3.8 ± 0.3	1.99 ± 0.01	2.0 ± 0.7	11.0

---

Therefore, the EXAFS and XANES results showed that gold species in the Au/Al₂O₃ samples were fully reduced after H₂ treatment at different temperature (150, 250, 350, 450 °C) for 2 h. The mean Au NPs size decreased with increasing the reduction temperature from 150 to 450 °C.

Reaction mechanism

The group of Shi reported that using Pd/CuZrO_x under milder conditions or using Pd/ZnZrO_x during methylation with CO₂/H₂, N-formylation mainly processed instead of N-methylation, which indicates the N-formamide could be the intermediate.²⁰ Beller et al. also reported that formamide was the intermediate during methylation of amines with CO₂/H₂ catalyzed by Ru catalyst and methanol contributed insignificantly to the reaction.^15b In order to better understand the mechanism of the methylation reaction catalyzed by the Au/Al₂O₃ catalyst, control experiments were conducted to confirm the intermediates. Initially, the reduction of the N-methylformanilide (1c) with H₂ was investigated in the presence of Au/Al₂O₃ catalyst. Clearly, 1c was fully converted and 1b was the major products in 61% yield (Scheme 1). Aside from methylation of 1c reduced by H₂, 1c was simultaneously reduced by H₂ with Au/Al₂O₃ catalyst to produce 1a in 39% yield. Under standard reaction conditions, trace amount of 1c could be detected. Moreover, we found that increasing the mean size of Au NPs could not only decrease the conversion of 1a, but also decrease the selectivity towards 1b to a certain extent as shown in Fig. S7.† These results suggest that 1c could be the key intermediate and significantly contributed to methylation of 1a with CO₂/H₂ catalyzed by Au/Al₂O₃ catalyst.

Scheme 1 Control experiment.

Based on these results, a plausible mechanism is given in Scheme 2. Under our reaction conditions, the formamide 1c first produced due to Au-catalyzed reaction of 1a with CO₂/H₂. The formamide 1c could be quickly reduced to give 1b (major) or be decarbonylated into 1a (minor). Then, 1a continued to react with CO₂/H₂ catalyzed by Au/Al₂O₃ catalyst following a similar hydrogenation and decarbonylation process as described above to produce the corresponding products.

Scheme 2 Proposed reaction pathway of Au-catalyzed methylation of 1a with CO₂/H₂.

Conclusions

Au NPs with uniform mean sizes (∼3 nm) loaded onto various supports were successfully prepared. These supported Au catalysts were studied in detail to clarify the support and Au particle size effects in the direct methylation reaction of N-methylaniline with CO₂/H₂. Of the supports tested, Al₂O₃ with relatively strong acidic and basic surface sites showed the best catalytic performance. Other supports, such as SiO₂ with few acidic sites, Mn₂O₃ and ZnO with weak acid–base properties, Co₃O₄ and C with neither acidic site nor basic site and MgO with relatively strong basic sites, have little or low catalytic activity. There is moderate catalytic activity when Au/TiO₂ and Au/CeO₂ with a few acidic sites are used. Supports with relatively strong acid–base bifunctional sites may be necessary for high efficient methylation of N-methylaniline with CO₂/H₂.

By changing preparation parameters during the DPU process, we obtained the Al₂O₃ supported Au NPs with controlled mean sizes in the range of 1.8 to 8.3 nm. We proved that, during the DPU process, aging time, aging temperature and mole ratio of gold to urea were the main preparation parameters affecting the mean size of Au NPs supported on Al₂O₃.

The TOF values for direct methylation of N-methylaniline with CO₂/H₂ increased with decreasing the mean size of Au NPs from 8.3 to 1.8 nm and XAFS results demonstrated that only reduced gold NPs supported on alumina were responsible for high catalytic activity of Au/Al₂O₃ catalyst. The results suggested that the Au-catalyzed direct methylation of N-methylaniline with CO₂/H₂ was a structure-sensitive reaction. Then, we proposed that N-methylformanilide could be a key intermediate during methylation of N-methylaniline with CO₂/H₂.

Acknowledgements

This work was financially supported by the Program of International S&T Cooperation (Grant no. 2014DFG60230), the National Basic Research Program of China (2013CB933104), the National Natural Science Foundation of China (no. 21303246, 11275258, 91127001).

Notes and references

1. J. T. Houghton, G. J. Jenkins and J. J. Ephraums, Climate Change: The IPCC Scientific Assessment, Cambridge University Press, Cambridge, UK, 1990.
2. H. Arakawa, M. Aresta, J. N. Armor, M. A. Barteau, E. J. Beckman, A. T. Bell, J. E. Bercaw, C. Creutz, E. Dinjus, D. A. Dixon, K. Domen, D. L. DuBois, J. Eckert, E. Fujita, D. H. Gibson, W. A. Goddard, D. W. Goodman, J. Keller, G. J. Kubas, H. H. Kung, J. E. Lyons, L. E. Manzer, T. J. Marks, K. Morokuma, K. M. Nicholas, R. Periana, L. Que, J. R. Nielson, W. M. H. Sachtler, L. D. Schmidt, A. Sen, G. A. Somorjai, P. C. Stair, B. R. Stults and W. Tumas, Chem. Rev., 2001, 101, 953–996.
3. T. Sakakura, J. C. Choi and H. Yasuda, Chem. Rev., 2007, 107, 2365–2387.
4. (a) K. M. Y. Yu, I. Curcic, J. Gabriel and S. C. Edman Tsang, ChemSusChem, 2008, 1, 893–899; (b) J. Ma, N. N. Sun, X. L. Zhang, N. Zhao, F. K. Xiao, W. Wei and Y. H. Sun, Catal. Today, 2009, 148, 221–231; (c) W. Wang, S. P. Wang, X. B. Ma and J. L. Gong, Chem. Soc. Rev., 2011, 40, 3703–3727.
5. (a) G. Pagani, Hydrocarbon Process., Int. Ed., 1982, 61, 87–91; (b) S. Walendy, G. Francio, M. Hoelscher and W. Leitner, Exp. Green Sustainable Chem., 2009, 275–277; (c) S. G. Liang, H. Z. Liu, T. Jiang, J. L. Song, G. Y. Yang and B. X. Han, Chem. Commun., 2011, 47, 2131–2133; (d) M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner, P. Markewitz and T. Müller, ChemSusChem, 2011, 4, 1216–1240.
6. (a) S. N. Riduan, Y. Zhang and J. Y. Ying, Angew. Chem., Int. Ed., 2009, 48, 3322–3325; (b) A. E. Ashley, A. L. Thompson and D. OHare, Angew. Chem., Int. Ed., 2009, 48, 9839–9843; (c) S. Wesselbaum, T. vom Stein, J. Klankermayer and W. Leitner, Angew. Chem., Int. Ed., 2012, 51, 7499–7502.
7. K. Motokura, D. Kashiwame, A. Miyaji and T. Baba, Org. Lett., 2012, 14, 2642–2645.
8. Q. Su, J. Sun, J. Q. Wang, Z. F. Yang, W. Q. Cheng and S. J. Zhang, Catal. Sci. Technol., 2014, 4, 1556–1562.
9. Y. Yoshida, Y. Arai, S. Kado, K. Kunimori and K. Tomishige, Catal. Today, 2006, 115, 95–101.
10. L. Schmid, A. Canonica and A. Baiker, Appl. Catal., A, 2003, 255, 23–33.
11. T. G. Ostapowicz, M. Schmitz, M. Krystof, J. Klankermayer and W. Leitner, Angew. Chem., Int. Ed., 2013, 52, 12119–12123.
12. (a) A. Ricci, Modern Amination Methods, Wiley-VCH, Weinheim, 2000; (b) R. N. Salvatore, C. H. Yoon and K. W. Jung, Tetrahedron, 2001, 57, 7785; (c) Z. Rapport, The Chemistry of Anilines, Wiley Interscience, 2007.
13. (a) M. F. Ali, B. M. ElAli and J. G. Speight, Handbook of Industrial Chemistry-Organic Chemicals, McGraw-Hill, New York, 2005; (b) H.-J. Arpe and S. Hawkins, Industrial Organic Chemistry, Wiley-VCH, Weinheim, 1997; (c) A. Dhakshinamoorthy, M. Alvaro and H. Garcia, Appl. Catal., A, 2010, 378, 19–25.
14. (a) X. Jiang, C. Wang, Y. W. Wei, D. Xue, Z. T. Liu and J. L. Xiao, Chem. - Eur. J., 2014, 20, 58–63; (b) S. Savourey, G. Lefère, J. C. Berthet and T. Cantat, Chem. Commun., 2014, 50, 14033–14036; (c) I. Sorribes, K. Junge and M. Beller, Chem. - Eur. J., 2014, 20, 7878–7883.
15. (a) O. Jacquet, X. Frogneux, C. D. N. Gomes and T. Cantat, Chem. Sci., 2013, 4, 2127–2131; (b) Y. Li, X. Fang, K. Junge and M. Beller, Angew. Chem., Int. Ed., 2013, 52, 9568–9571.
16. (a) K. Beydoun, T. vom Stein, J. Klankermayer and W. Leitner, Angew. Chem., Int. Ed., 2013, 52, 9554–9557; (b) Y. Li, I. Sorribes, T. Yan, K. Junge and M. Beller, Angew. Chem., Int. Ed., 2013, 52, 12156–12160.
17. (a) S. Das, F. D. Bobbink, G. Laurenczy and P. J. Dyson, Angew. Chem., Int. Ed., 2014, 53, 12876–12879; (b) E. Blondiaux, J. Pouessel and T. Cantat, Angew. Chem., Int. Ed., 2014, 53, 12186–12190.
18. X. J. Cui, X. Dai, Y. Zhang, Y. Deng and F. Shi, Chem. Sci., 2014, 5, 649–655.
19. K. Kon, S. M. A. H. Siddiki, W. Onodera and K. I. Shimizu, Chem. - Eur. J., 2014, 20, 6264–6267.
20. X. Cui, Y. Zhang, Y. Q. Deng and F. Shi, Chem. Commun., 2014, 50, 13521–13524.
21. (a) M. Haruta, Catal. Today, 1997, 36, 153–166; (b) X. Zhang, H. Shi and B. Q. Xu, Angew. Chem., Int. Ed., 2005, 44, 7132–7135; (c) A. Corma, P. Serna and H. García, J. Am. Chem. Soc., 2007, 129, 6358–6359; (d) L. He, L. C. Wang, H. Sun, J. Ni, Y. Cao, H. Y. He and K. N. Fan, Angew. Chem., Int. Ed., 2009, 48, 9538–9541; (e) M. D. Hughes, Y.-J. Xu, P. Jenkins, P. McMorn, P. Landon, D. I. Enache, A. F. Carley, G. A. Attard, G. J. Hutchings, F. King, E. H. Stitt, P. Johnston, K. Griffin and C. J. Kiely, Nature, 2005, 437, 1132.
22. (a) U. Hartfelder, C. Kartusch, M. Makosch, M. Rovezzi, J. Sá and J. A. van Bokhoven, Catal. Sci. Technol., 2013, 3, 454–461; (b) K. R. Souza, A. F. F. de Lima, F. F. de Sousa and L. G. Appel, Appl. Catal., A, 2008, 340, 133–139; (c) R. Zanella, S. Giorgio, C.-H. Shin, C. R. Henry and C. Louis, J. Catal., 2004, 222, 357–367; (d) H.-S. Oh, J. H. Yang, C. K. Costello, Y. M. Wang, S. R. Bare, H. H. Kung and M. C. Kung, J. Catal., 2002, 210, 375–386; (e) A. Hugon, N. E. Kolli and C. Louis, J. Catal., 2010, 274, 239–250.
23. E. Bus, R. Prins and J. A. van Bokhoven, Catal. Commun., 2007, 8, 1397–1402.
24. M. Shekhar, J. Wang, W. S. Lee, W. D. Williams, S. M. Kim, E. A. Stach, J. T. Miller, W. N. Delgass and F. H. Ribeiro, J. Am. Chem. Soc., 2012, 134, 4700–4708.
25. D. Xiang, X. F. Liu, J. S. Sun, F. S. Xiao and J. M. Sun, Catal. Today, 2009, 148, 383–388.
26. L. D. Hao, Y. F. Zhao, B. Yu, H. Y. Zhang, H. J. Xu and Z. M. Liu, Green Chem., 2014, 16, 3039–3044.
27. M. Manzoli, C. Chiorino, F. Vindigni and F. Boccuzzi, Catal. Today, 2012, 181, 62–67.
28. X. L. Du, G. Tang, H. L. Bao, Z. Jiang, X. H. Zhong, D. S. Su and J. Q. Wang, ChemSusChem, 2015, 8, 3489–3496.
29. X. Q. Li, J. Xu, F. Wang, J. Gao, L. P. Zhou and G. Y. Yang, Catal. Lett., 2006, 108, 137–140.
30. W. H. Fang, J. S. Chen, Q. H. Zhang, W. P. Deng and Y. Wang, Chem. - Eur. J., 2011, 17, 1247–1256.
31. K. Shimizu, K. Sugino, K. Sawabe and A. Satsuma, Chem. - Eur. J., 2009, 15, 2341–2351.
32. R. Zanella, S. Giorgio, C. R. Henry and C. Louis, J. Phys. Chem. B, 2002, 106, 7634–7642.
33. R. J. H. Grisel, P. J. Kooyman and B. E. Nieuwenhuys, J. Catal., 2000, 191, 430–437.
34. G. C. Bond, C. Louis and D. T. Tompson, Catalysis by Gold, Imperial College Press, 2006.
35. A. Goguet, R. Burch, Y. Chen, C. Hardacre, P. Hu, R. W. Joyner, F. C. Meunier, B. S. Mun, D. Thompsett and D. Tibiletti, J. Phys. Chem. C, 2007, 111, 16927–16933.
36. M. W. E. van den Berg, A. de Toni, M. Bandyopadhyay, H. Gies and W. Grünert, Appl. Catal., A, 2011, 391, 268–280.
37. (a) Y. Guo, D. Gu, Z. Jin, P. P. Du, R. Si, J. Tao, W. Q. Xu, Y. Y. Huang, S. Senanayake, Q. S. Song, C. J. Jia and F. Schüth, Nanoscale, 2015, 7, 4920–4928; (b) D. P. Akolekar, G. Foran and S. K. Bhargava, J. Synchrotron Radiat., 2004, 11, 284–290.

---

Footnote

† Electronic supplementary information (ESI) available: Details containing figures of TEM, catalysis activity, XRD. See DOI: 10.1039/c5ra20991a

---