Nanoparticulate gold catalysts (hereafter denoted as NPGCs) have been attracting much attention over the past few decades [1-5]. NPGCs deposited on base metal oxides act as excellent catalsts, particularly for CO oxidation below room temperature. In order to obtain highly active NPGCs, the diameter of the gold nanoparticles (NPs) should be smaller than 5 nm, which increases the surface boundary between the Au NPs and the support [6, 7]. Significant effects of a metal oxide support on catalytic behavior have also been observed, and many kinds of gold-supported catalysts have been developed. A solid acid, which typically shows a low isoelectric point, is attractive as a support, but there have been only a few reports on gold-supported acidic oxides other than silica and zeolite supports [8-10].
Niobium oxide (Nb2O5 or hydrated niobium oxide, Nb2O5·nH2O) is widely used in catalysis and in electrochromic and photoelectrochemical devices [11, 12]. Nb2O5 has been used as a water-tolerant solid acid catalyst for many reactions such as alkylation, esterification, hydrolysis, dehydration and hydration [13, 14]. Nb2O5 also has potential applications as a support of NPGCs for oxidation of glycerol, which may lead to the development of new processes for glycerol upgrading and for oxidative condensation of furfural and alcohol, aimed at the development of new processes for renewable biomass. Sobczak et al. [15-17] reported Au (or Au-Cu) catalysts supported on Nb2O5 and Nb/MCF (silica) by modification of the niobium oxide surface by an amino group. Tong et al. [18] reported a simple Au/Nb2O5 catalyst; however, the mean diameter of the gold particles was about 9 nm. Recently, we have reported new crystalline niobium oxides (deformed orthorhombic) consisting of NbO6 octahedra, NbO7 and micropores based on the 7-membered ring in its structure with corner-sharing in the c-direction [19]. This deformed orthorhombic niobium oxide, denoted as Nb2O5(HT), was synthesized by a hydrothermal process, and catalytic activity based on its crystalline structure was revealed. An advantageous feature as a support is that Nb2O5(HT) has a large surface area exceeding 200 m2 g-1. Nb2O5 is classified as an n-type semiconductor and has a redox property, indicating that Nb2O5 has the possibility of showing high catalytic activity as a support of an NPGC if gold NPs are deposited on Nb2O5.
Gold catalysts are usually prepared by a deposition method in a liquid phase, such as the deposition precipitation (DP) method or deposition reduction (DR) method, in order to disperse nano-scale gold particles on the support. Control of the charged state of the metal oxide surface is essential for depositing gold NPs with a high degree of and uniform dispersion. However, it was difficult to prepare NPGCs in previous studies when the isoelectric point of the support was below pH 5. To obtain gold NPs on the support surface, electrostatic interactions between gold precursors and the support surface should be strong [20]. A positively charged ethylenediamine-gold complex could be attracted by supports that show a negative electrical charge in water.
The DP method, DP method with urea, DR method and one-pot method were used to prepare 1 wt% Au/Nb2O5 catalysts in this study to obtain NPGCs. The preparation conditions for the DP method and DR method were optimized by changing gold precursors and reagents for pH control of the precursor solution. The effects of Nb2O5 supports on catalytic activity of the Au/Nb2O5 catalysts obtained were examined through CO oxidation reaction.
Niobium oxide was prepared in a manner similar to that reported previously [19]. Layered-structure-type (deformed orthorhombic) niobium oxide was synthesized by a hydrothermal method from ammonium niobium oxalate (NH4[NbO(C2O4)2(H2O)2]·nH2O, Aldrich) and was denoted as Nb2O5(HT). The Nb precursor (6 mmol based on Nb) was dissolved in distilled water (40 mL) and then the solution was treated by hydrothermal synthesis for 1 d at 175 ℃. The resultant solid was separated and washed with distilled water by using filtration, and then the solid was dried at 80 ℃. The dried solid was crushed in an agate mortar and heat-treated for 2 h at 400 ℃ in air. Commercially available Nb2O5 (Wako) and Nb2O5·nH2O (Soekawa) were used for comparison after calcination at 400 ℃ in air for 2 h.
Gold catalysts were prepared by the DP method and DR method. Au(en)2Cl3 or HAuCl4 (0.0507 mmol) was used as a precursor and was dissolved in 51 mL of water followed by the addition of niobium oxide (1 g). The temperatures of the dispersed solutions were 50 ℃ for the DP method and room temperature for the DR method. The mixture was stirred for 1 h with pH adjusted to 9 by using 0.1 mol L-1 NaOH solution. In the DR method, NaBH4 was added to the mixture, and the suspension was washed with 3 L of deionized water using suction filtration and dried at 80 ℃. The dried catalyst was calcined at 300 ℃ for 4 h in air.
Preparation of Au/Nb2O5 by deposition precipitation with urea was carried out by adding urea (0.06 mol) to the same precursor solution as that used for the DP method without pH control and stirring at 90 ℃ for 4 h. In the one-pot method, the gold precursor and ammonium niobium oxalate (Au/Nb=1/19; Nb, 6 mmol) were dissolved in 40 mL of water, and then the solution was treated by hydrothermal synthesis for 1 d at 175 ℃.
Catalytic activity was tested through CO oxidation reaction. The catalyst (0.15 g) was set in a fixed bed reactor, and 1 vol% CO in air was flowed (50 mL min-1). H2O concentration in the flow gas was monitored by a dew-point meter, and the concentrations were 20-100 ppm at T1/2 in all experiments. Catalytic activities were compared by the reaction rate per all Au atoms (TOF, μmol(CO) μmol(Au)-1 s-1), which was obtained under the condition in which conversion of CO is below 15% for differential reactor assumption.
The catalysts were characterized by the following techniques. Powder X-ray diffraction (XRD) patterns were recorded with a diffractometer (RINT Ultima+, Rigaku) using Cu-Kα radiation (tube voltage: 40 kV, tube current: 20 mA). Diffractions were recorded in the range of 4°-60° at 5° min-1. Morphology was investigated by using a transmission electron microscope (TEM, JEM-2100F, JEOL) at 200 kV. The samples were dispersed in ethanol by ultrasonic treatment for several minutes, and drops of the suspension were placed on a copper grid for TEM observations.
Temperature-programmed desorption of ammonia (NH3-TPD) was used to measure oxide surface acidity. The experiment was carried out using an autochemisorption system (Bel Cat, BEL Japan). The experimental procedure was as follows. The catalyst (ca. 100 mg) was set on quartz wool and preheated under helium (50 mL min-1) at 300 ℃ for 1 h. Then ammonia was introduced at 100 ℃ for 30 min. The desorption profile from 100 to 700 ℃ was recorded with a mass spectrometer under a helium flow (50 mL min-1).
Different types of Nb2O5 supports were used to prepare 1 wt% Au/Nb2O5 catalysts, and their effects on catalytic activity for CO oxidation were examined. The layered-type Nb2O5(HT), hydrothermally synthesized Nb2O5, and commercially available Nb2O5 and Nb2O5·nH2O were used as supports. GNPCs were prepared by the DP method using Au(en)2Cl3 as a cationic gold precursor. The detailed preparation conditions will be discussed later. The DP method is the most conventional and convenient to deposit gold on supports and is particularly useful for basic metal oxides. Since Nb2O5 is a typical acidic support, preparation of highly dispersed Au/Nb2O5 catalysts by the DP method has not succeeded yet [15-18].
Fig. 1 shows XRD patterns of Nb2O5(HT) and Au/Nb2O5 catalysts. Table 1 summarizes the physical properties of the samples obtained and their catalytic activities. The XRD patterns of Nb2O5(HT) showed diffraction peaks at 2θ=22.7° and 46.2°, which were attributed to (001) and (002) planes of the layered-type structure of Nb2O5(HT), respectively, and the diffraction peak based on Nb2O5(HT) did not change after deposition of gold. The prepared Au/Nb2O5 catalysts (Fig. 1 (b-d)) showed diffraction peaks at 2θ=38.3° and 44.5° based on Au(111) and Au(200), respectively. The crystalline diameters calculated by the Scherrer equation were 4.0 nm for Au/Nb2O5(HT), 6.1 nm for Au/Nb2O5·nH2O and 23 nm for Au/Nb2O5. The values of BET surface area of niobium oxides were 208 m2 g-1 (Nb2O5(HT)), 19 m2 g-1 (Nb2O5·nH2O) and 5.8 m2 g-1 (Nb2O5). In previous studies, the values of BET surface area of Nb2O5 supports were small (85 m2 g-1 in [17], 7 m2 g-1 in[15, 16] and 1 m2 g-1 in[18]). These data suggested that the large surface area (208 m2/g) of Nb2O5(HT) was one of the key factors for the formation of NPs of gold dispersed on Nb2O5.
Fig. 2 shows the effects of reaction temperature on CO conversion over the catalysts obtained. The Nb2O5(HT) sample showed no activity for CO oxidation even at 250 ℃. In contrast, the activity of the Au/Nb2O5(HT) catalyst for CO oxidation was greatly enhanced, and the temperature for 50% CO conversion (T1/2) was 85 ℃. The catalytic activity of Au/Nb2O5(HT) was much higher than that of Au/SiO2, which was an acidic support. The catalytic activities of Au/Nb2O5·nH2O and Au/Nb2O5 were much lower than the catalytic activity of Au/Nb2O5(HT). The values of T1/2 were 171 ℃ for Au/Nb2O5·nH2O and 250 ℃ for Au/Nb2O5. The loading amount of gold of Au/Nb2O5(HT) was 0.81 wt% and that of Au/Nb2O5·nH2O was 0.88 wt%. It was difficult to deposit gold NPs for the commercial Nb2O5 support, and the amount of gold was 0.10 wt%. To obtain gold NPs on the support surface, electrostatic interactions between gold precursors and the support surface should be strong. The use of Au(en)2Cl3, which forms the Au(en)23+ species in the precursor solution. These results suggested that the interraction between the Au(en)23+ speciesand the surface of Nb2O5, which show a negative electrical charge in water, was different, even though the support has the same chemical composition (Nb2O5), and that the interaction has an important role in the deposition of NPs of gold on Nb2O5.
Deposition of gold in liquid phase by the DP, DR, DP method with urea (DP-urea method), and one-pot method was carried out by using Nb2O5(HT), and the catalytic activities were compared. Deposition of gold in a solid state, a method called the solid grinding (SG) method, was also carried out for comparison. The loading amount of gold was 1 wt% for all catalysts. Fig. 3 shows TEM images and the distribution of diameters of gold particles of Au/Nb2O5 prepared by the different deposition methods. The crystalline diameter of gold particles of the catalyst obtained agreed well with TEM observation. The mean diameters were 4.1 nm (standard deviation (SD), 1.1 nm) for Au/Nb2O5 prepared by the SG method, 4.9 nm (SD, 1.4 nm) for the catayast prepared by the DR method and 5.0 nm (SD, 2.4 nm) for the catayast prepared by the DP method.
Then the effects of deposition methods on catalytic activity for CO oxidation were investigated by Au/Nb2O5 prepared by the different deposition methods (Fig. 4 and Table 2). The values of TOF at 20 ℃ and T1/2 of those catalysts corresponded to the gold particle size, and the catalytic activity decreased in the order of SG > DR > DP > DP-urea > one-pot methods. The catalytic activity of Au/Nb2O5(HT) prepared by the SG method was the highest; however, the precursor of gold, Me2Au(acac), is currently too expensive to use widely. Therefore, improvement of deposition of gold in liquid phase or the development of an alternative precursor for the SG method will be needed for the deposition of gold NPs on Nb2O5.
Fig. 5 shows that time-on-stream change of catalytic activity for CO oxidation at 100 ℃ with Au/Nb2O5(HT) prepared by the DP and DR methods. In the case of the DP method, the preparation condition was the same as that in entry 1 in Table 2. The precursor used was Au(en)2Cl3 and pH value of the precursor solution was adjusted to 9 with NaOH solution. The sample obtained was dried under a vacuum condition at room temperature and then calcined at 300 ℃ for 4 h (see also next section and supporting information).
In the case of the DR method, the condition for preparation was the same as that in entry 5 in Table 2. The precursor used was Au(en)2Cl3 and pH value of the precursor solution was adjusted to 9 with NaOH solution, and then NaBH4 solution was added. The sample obtained was dried at 80 ℃ and then calcined at 300 ℃ for 4 h (see also next section and supporting information). The conversion of CO with the Au/Nb2O5(HT) catalyst prepared by the DR method was maintained above 98% and that of the catalyst prepared by DP the method was 70% at 100 ℃. Then the conversion of CO with both catalysts decreased with prolongation of reaction time. In-situ calcination treatment at 250 ℃ in air flow was carried out for 1 h after 20-h reaction. The catalytic activity was recovered by calcination treatment. Therefore, deactivation of the catalyst was not due to aggregation of gold NPs. Fig. 5(b) and (c) show the distribution of the diameters of gold particles of Au/Nb2O5(HT) catalysts prepared by the DR and DR methods after the CO oxidation. The diameters of gold particles of catalysts after the reaction were almost no change compared to that of the fresh catalysts (Fig. 3). The active lattice oxygen in the perimeter of gold NPs could be recovered by calcination in air.
Fig. 6 shows NH3-TPD spectra of Nb2O5(HT) and Au/Nb2O5(HT) catalysts prepared by the DP and DR methods. The temperatures of preheat-treatment under He (50 mL min-1) were 300 ℃ for the Nb2O5(HT) catalyst and 250 ℃ for the Au/Nb2O5(HT) catalyst. A broad desorption peak of NH3 was observed for Nb2O5(HT), and its peak temperature was about 230 ℃. The amount of absorbed NH3 was 0.48 mmol g-1. The shapes of desorption peaks for Au/Nb2O5(HT) catalysts prepared by the DP and DR methods were similar to those for Nb2O5(HT). The amounts of absorbed NH3 were 0.39 mmol g-1 for the Au/Nb2O5(HT) catalyst prepared by the DP method and 0.30 mmol g-1 for the catalyst prepared by the DR method. More than 64% of the acid sites were functioning to adsorb NH3 after the deposition of gold NPs. NH3-TPD mesurement showed that acid sites exsisted on the surface of the Au/Nb2O5 catalyst even after deposition of gold NPs.
Au/Nb2O5 catalysts were prepared by the DP and DR methods with changes in the preparation conditions. The preparation amount of gold was 1 wt% for all catalysts. Table 3 shows the effects of the conditions on the activity for CO oxidation and on the diameters of gold particles. Crystalline diameter of gold was calculated by using the Scherrer equation from the XRD pattern, and mean diameters of gold particles and their distribution were estimated by TEM observations.
In the DP method, entry 9 (Table 3) is a typical condition for preparation of a gold catalyst by using HAuCl4 as a precursor; however, this condition was not effective for the deposition of gold (Au loading 0.18 wt%). The mean diameter of the gold NPs was 7.5 nm with a large distribution (SD, 5.9 nm), and T1/2 was 262 ℃. Au(OH)4- species were formed from the HAuCl4 precursor in the precursor solution at pH 9 [21] and could not interact with niobium oxide, of which the isoelectric point of its surface is estimated to be 2.8 [22]. The use of Au(en)2Cl3, which forms the Au(en)23+ species in the precursor solution, enhanced the amount of gold loading and the catalytic activity (entry 1). The amount of gold for the catalyst obtained was 0.81 wt% and T1/2 was 85 ℃ with mean diameter of gold NPs of 5.0 nm as described above.
In the DR method, HAuCl4 was used as a precursor for entry 10 (Table 3) and T1/2 for the sample obtained was 110 ℃. The amount of gold for the catalyst obtained was 0.26 wt%; however, small gold particles (MD, 4.3 nm; SD, 1.1 nm) were deposited with a relatively large population. TOF based on all Au atoms at 20 ℃ was 0.010 s-1, which is higher than that of the catalyst prepared by using the Au(en)2Cl3 precursor (entry 5; TOF, 0.0051 s-1). TEM observation of Au/Nb2O5(HT) catalyst prepared by using the HAuCl4 precursor solution showed that gold NPs were deposited on the basal planes of the rod-type Nb2O5(HT) support (Fig. 7). The Nb2O5(HT) catalyst obtained by the hydrothermal method from ammonium niobium oxalate shows a layered-type structure, in which NbO6 octahedra were corner-sharing in the c-direction, and consists of NbO6 octahedra, NbO7 and a 7-membered ring in the a-b plane, indicating the formation of a high-dimensional structure [19]. Strong Brñnsted acidity based on the basal planes has been reported. Kundu et al. [23] suggested that the differences in surface energies owing to the polar character of the {0001} facets of ZnO resulted in a preferential reduction of Au on the basal planes. These kinds of structural differences would affect the interaction between Au(OH)4- species and the surface of Nb2O5, and Au3+ was reduced by NaBH4 to form NPs on the basal planes of the rod.
The effects of supports were compared by using a commercially available niobium catalyst (entries 11-14). The catalytic activities of gold supported on commercially available Nb2O5·nH2O (Au/Nb2O5·nH2O) and Nb2O5 (Au/Nb2O5) were less than the catalytic activity of Au/Nb2O5(HT). The value of T1/2 was 194 ℃ for Au/Nb2O5·nH2O and conversion was 29% at 250 ℃ for Au/Nb2O5by using Au(en)2Cl3 precursor. These activities are similar to those of the catalyst prepared by the DP method.
The deposition methods in the liquid phase are summarized schematically in Fig. 8. The use of the Au(en)2Cl3 precursor was effective for the deposition of gold NPs on Nb2O5(HT) by both the DP and DR methods. The use of the HAuCl4 precursor was not effective for the deposition of gold by the DP method; however, in the DR method, gold NPs were deposited on basal planes of rod-type Nb2O5(HT) particles. Because Brñnsted acid sites are located at the basal planes of Nb2O5(HT), utilization of the combination reaction of the a-b plane of Nb2O5(HT) with gold NPs will be expanded in the future to many reactions.
Layered-type Nb2O5(HT), hydrothermally synthesized Nb2O5, and commercially available Nb2O5 and Nb2O5·nH2O were used as supports for the preparation of 1 wt% Au/Nb2O5 catalysts. The Nb2O5(HT) sample was inactive for CO oxidation; however, the catalytic activity was significantly enhanced by the deposition of gold NPs. Moreover, the catalytic activity of Au/Nb2O5(HT) was much higher than these of the catalysts prepared from Nb2O5·nH2O and Nb2O5. The Nb2O5(HT) sample was effective as a support because of its large surface area.
NPGCs supported on Nb2O5(HT) were prepared by different deposition methods. The DP method, DR method, DP-urea method and one-pot method were used to prepare Au/Nb2O5 catalysts. Conditions for preparation were optimized in the DP and DR methods. The use of Au(en)2Cl3 is suitable as a gold precursor for Nb2O5 in both deposition methods. The mean diameters of gold particles were 5.0 nm (SD, 2.4 nm) for Au/Nb2O5(HT) prepared by the DP method and 4.9 nm (SD, 1.4 nm) for Au/Nb2O5(HT) prepared by the DR method in the optimized conditions. In the case of the DR method, gold nanoparticles (MD, 4.3 nm; SD, 1.1 nm) were deposited on the basal plane of Nb2O5(HT). The activity of gold NPs on the basal plane of Nb2O5(HT) was very high, but improvement is still needed to increase the amount of gold loading.
NH3-TPD mesurement showed that acid sites exsisted on the surface of the Au/Nb2O5 catalyst after deposition of gold NPs. Therefore, application of the Au/Nb2O5 catalyst can be expanded in the future to many reactions by utilization of the combination of acidic sites and oxidation reaction sites.
2016,
Vol. 37 Issue (10): 1694-1701 
PDF
2016,
Vol. 37 Issue (10): 1694-1701 PDF