Recently, Au-Ag bimetallic catalysts have attracted growing attention because of the remarkable synergistic effects exhibited by these material during various oxidative reactions, such as CO oxidation over a Au-Ag alloy catalyst supported on mesoporous aluminosilicate [1-4], the liquid phase aerobic oxidation of glucose by PVP-protected Agcore/Aushellbimetallic nanoparticles (NPs) [5], and the oxidative coupling of CH3OH or CO oxidation by nanoporous gold synthesized by the selective leaching of silver from a Au-Ag alloy in HNO3[6-9].
Zhang et al. [5] has proposed that the synergistic effect demonstrated by Au-Ag bimetallic catalysts can be attributed to charge transfer from Ag to Au, similar to that which occurs between Pd and Au in Au-Pd bimetallic nanoparticles having a “crown jewel” structure that are highly active during the aerobic oxidation of glucose [10]. Conversely, Liu et al. [11] have suggested that, because Ag interacts much more strongly with O2 than with Au, the boundaries between Au and AgOx zones (generated during pretreatment with O2 or under oxidative reaction conditions) play a key role in promoting the oxidative catalytic activity. Evidence for this theory is found in the case of TiO2/Au(111) [12], Fe2O3/Au(111) [13] and CeO2-x/Au films [14]. From these discussions, it is evident that determining whether Ag exists as an oxide or in the form of a metallic alloy on the surface of the Au may be the key to obtaining additional insights into the origin of the synergistic effect in Au-Ag bimetallic catalysts.
Iizuka et al. [15] previously measured the activity for CO oxidation of seven samples of unsupported Au powder synthesized by evaporating high purity metallic Au ( > 99.99%) under an inert gas. The surfaces of these samples had been purposely contaminated with various trace amounts of Ag (such that these specimens are hereafter referred to Ag/Au powder) using a closed recirculation reaction system. They calculated the kCO2 values (that is, the rate constant for CO oxidation per unit surface area) of these samples and plotted these values against the surface atomic concentrations of Ag to determine the extent to which surface Ag atoms take part in the oxidation. A strong correlation between kCO2andtheAg surface concentration was found [15]. Since these Ag/Au powders were processed under 1 atm of circulating oxygen at temperatures between 423 and 453 K before exhibiting any activity for CO oxidation, it is believed that their catalytic activity resulted from the so-called “contact interface effect” between Au and AgOx in the surface layer [15].
Thereafter, Iizuka et al. [16] studied the CO oxidation mechanism, using a representative active sample of Ag/Au powder: Ag/Au-b, with a BET specific surface area of 1.89 m2/g, a bulk Ag content determined by inductively coupled plasma (ICP) of 0.110 wt%, and a surface Ag content determined by X-ray photoelectron spectroscopy (XPS) of 10.8 atomic%. Kinetic measurements, including catalytic CO oxidation over Ag/Au-b and the reduction and/or the oxidation of the sample surface by CO and/or O2 were also performed at low pressures. Prior to these low pressure reaction rate measurements, the Ag/Au-b was subjected to one of two pretreatments, as follows.
(1) Reaction with a stoichiometric mixture of CO and O2 at an initial pressure of 5.3 kPa at 313 K for 1 h, followed by exposure to a vacuum of approximately 1.3 × 10-3 Pa, denoted as Ag/Au having steady state surfaces.
(2) Oxidation of the above Ag/Au in circulating O2 at 13.3 kPa and 393 K for 2 h, followed by cooling to room temperature and subsequent exposure to a vacuum of approximately 1.3 × 10-3 Pa at 313 K, denoted as Ag/Au having oxidized surfaces.
In these trials, the reactor was loaded with 1.89 g of Ag/Au-b and the amount of surface Ag estimated from the BET surface area and the XPS Ag atomic content, assuming that the Ag atoms were present solely on the upper surface layer, was 8.92 μmol. This amount corresponds to 198 μL of an ideal gas at standard temperature and pressure, and to a pressure of 25.5 Pa based on the pressure of CO in the reactor. During each test, the pressure decrease accompanying the catalytic CO oxidation or the reduction and/or oxidation of the catalyst surface was compared with the amount of CO2 produced [16]. The kinetic behavior of the Ag/Au-b with the steady state surface indicated that the CO oxidation proceeded via the coadsorption of CO and O2 to form (CO3)* followed by the decomposition of this species, promoted by interactions with CO and O2, to produce CO2 at the contact interface between surface Ag and the Au support [16].
In trials in which 17 Pa CO was introduced to the Ag/Au-b having oxidized surfaces at 313 K for 60 min, the CO pressure decreased rapidly over the initial 7 min while the pressure dropped below 4 Pa. Following this phase, the CO pressure fell more slowly for the remaining 53 min of the trial while the pressure continued to decrease to 1.5 Pa. The initial pressure decrease of 13 Pa corresponded to one half of the amount of the surface Ag of the sample (4.5 μmol). In this same test, 5.4 μmol of CO were consumed and 5.6 μmol of CO2 were produced. The initial rapid decrease in the CO pressure and the production of CO2 corresponding to the amount of CO consumed suggested the presence of reactive O species on the surfaces of the Ag/Au-b having oxidized surfaces prior to CO exposure. The fact that the amount of CO2 produced was close to half the quantity of surface Ag further indicated that reactive O species may be adsorbed on the surface Ag atoms at a ratio of about 1 to 2 [16].
In other experiments, a reaction mixture of CO and O2with the composition CO/O2=2/1 at an initial pressure of approximately 11 Pa was introduced twice in succession to a sample of Ag/Au-b having oxidized surfaces. In this case, only the reduction of reactive adsorbed O with the CO in the reaction mixture to produce CO2 proceeded during the first dose of the reaction mixture, without incorporation of the O2 in the gas phase. The participation of the O2 in the CO2 production process, as well as its incorporation into the catalyst surface, was found to begin gradually as the reduction of reactive adsorbed O on the Ag proceeded during exposure to the second dose [16]. This result indicated that AgOx species tended not to be present as Ag2O in the reaction mixture on the surface of Au [16].
On the basis of the above work, the present study prepared a material composed of a Au catalyst supported on Ag2O to confirm the activity enhancement as a result of the contact between Ag2O and the Au surface, using a deposition-precipitation method. This paper describes the CO oxidation activity of this Au/Ag2O catalyst as determined using a closed recirculation reaction system, and compares this activity to that of Ag2O. In the second part of this report, the kinetic behavior of Au/Ag2O for CO oxidation is compared with that of the above-mentioned Ag/Au-b having oxidized surfaces. Finally, the kinetic behavior of Ag/Au-b having steady state surfaces during CO oxidation as well as the structure and composition of active sites that tend to enhance the catalytic activity of this material are discussed [17].
Silver oxide (Ag2O, Wako special grade, 99.9% purity, 0.93 m2/g) purchased from Wako Pure Chemicals Industries, Ltd. was used both as a catalyst and as a support for Au NPs. Au/Ag2O was prepared using a deposition-precipitation method in which the Ag2O powder is suspended in a HAuCl4 solution adjusted to pH=8 and heated at 343 K. The nominal Au loading in these specimens was 0.90 wt%. Changes in the Au concentration in the suspension were followed by colorimetry, employing a SnCl2 solution as a reducing reagent to form a Au colloid. Soon after adding the support, the concentration of Au was confirmed to drop to below the detection limit of the colorimetric assay, after which the suspension was stirred for 1 h at the same temperature. The mixture was subsequently decanted to separate the precipitate from the supernatant and the precipitate was washed with deionized water several times then heated in air overnight at 393 K. The Ag/Au-b powder reexamined in the present work was synthesized by evaporating high purity metallic Au ( > 99.99%) in an inert gas, as described in previous publications [15, 16].
A closed recirculation reaction system constructed of glass (base pressure of 1.3 × 10-3 Pa) previously used for the kinetic study of CO oxidation over Ag/Au-b powder was also used in the present investigation of CO oxidation over Au/Ag2O and Ag2O [16]. The details of this system and the associated experimental procedures are described elsewhere [16]. The pretreatment conditions and the procedures used in the present kinetics measurements for Au/Ag2O and Ag2O were also the same as those employed for the kinetic study performed with Ag/Au-b powder. Briefly, the sample in the reactor was heated under vacuum at 453 K for 1 h and then pretreated under 1 atm of circulating O2 for 3 h, then cooled to room temperature.
After removing O2 from the reaction vessel, already at the reaction temperature, a reaction mixture of CO and O2 (CO/O2=2/1) at an initial pressure of approximately 4.0 kPa was introduced into the circulation line (the volume of which was known) and the resulting CO2 was condensed in a trap cooled by liquid nitrogen. In this manner, the reaction rate was derived from the decrease in the total pressure. A typical reaction rate trial ran for 60 min, following which the remaining reaction mixture was completely removed from the circulation lines. The amount of reaction mixture consumed was calculated from the pressure decrease and the CO2 collected in the trap was evaporated at 195 K to determine the amount volumetrically.
When low pressure rate measurements were performed in the case of Ag/Au-b, the reaction system was changed from circulatory to static. The pressure changes accompanying catalytic CO oxidation over Ag/Au-b or the reduction and/or the oxidation of Ag/Au-b with CO and/or O2were followed using a capacitance type manometer (MKS Instrument Inc., Baratron 627B) [16].
After completing the kinetics trials with the recirculation apparatus, the Au/Ag2O was subjected to a series of characterizations, with the results listed in Table 1. The specific surface areas of the catalyst samples were determined in a static adsorption apparatus constructed of glass, using the BET method with N2 as the adsorbate at liquid nitrogen temperature. The bulk Au content of the Au/Ag2O was determined by spectrochemical emission analysis combined with ICP (Shimadzu Techno-Research, Inc., ICPS-8000). The mean particle diameters of the Au NPs were found to be 4.4±2.9 nm based on the analysis of 24 particles in a transmission electron microscope (TEM) image of Au NPs supported on Ag2O (JEOL, JEM-2010/SP). The chemical state of the Au on the surface of the Au/Ag2O was ascertained by XPS (JEOL, JPS-9010MC/SP), using the Mg Kα line at 1365 eV as the X-ray source. The Au 4f7/2 binding energy was determined to be 84.1 eV after calibration relative to the C 1s peak at 284 eV. Comparing this value with those reported in the literature [18, 19], we concluded that the Au had been reduced to Au0on the Ag2O support.
The results of a series of characterizations of the Ag/Au-b powder and some physico-chemical properties calculated based on the results of these data are summarized in Table 1 of ref. [16].
Fig. 1 presents first order plots of the pressure decrease of the reaction mixture of CO and O2 introduced over Au/Ag2O or Ag2O at room temperature. Both catalysts were pretreated under 1 atm of circulating oxygen at 423 K prior to acquiring these data. The plot for Ag2O is seen to be relatively linear, while the Au/Ag2O data exhibit a higher slope compared with that of Ag2O over the initial 20 min, followed by a break to a lower slope approximately equal to that of the Ag2O. The initial higher slope of the Au/Ag2O data is attributed to the Au NPs dispersed on the Ag2O.
The amounts of reaction mixture consumed and CO2 produced after exposure of the reaction mixture to the Au/Ag2O and Ag2O for 1 h are presented in Table 2 together with the expected amounts of CO2 calculated on the assumption that the mixture was consumed according to the reaction 2CO + O2 → 2CO2.
These data show that the amounts of CO2 generated by both the Au/Ag2O and Ag2O are in good agreement with the amounts of reaction mixture consumed but do not agree as well with the expected CO2 quantities calculated according to the catalyzed oxidation reaction equation. This result clearly indicates that the pressure decreases observed in Fig. 1 was not a result of catalytic oxidation but rather the reduction of Ag2O by CO in the reaction mixture, over both the Au/Ag2O and Ag2O. That is, oxidation to restore the reduced Ag2O via the incorporation of O2 did not proceed under the O2 pressure in the mixture of CO and O2having a stoichiometric composition. It is worth noting that the initial higher slopes seen in the first order plots of the Au/Ag2O data could be reproduced only when the catalyst was reprocessed under 1 atm of circulating oxygen at room temperature.
We subsequently exposed both Au/Ag2O and Ag2O to pure CO using almost the same initial pressure to that of the reaction mixture, and compared the resulting CO decrease rates with that of CO in the reaction mixture. The results are shown in Fig. 2, and Table 2 also presents the amounts of CO consumed and CO2 produced in these trials. The quantities of CO consumed are in good agreement with the CO2 generation over both the Au/Ag2O and Ag2O. It is notable that, in Fig. 2, the slope of the first order plot for the decrease in CO pressure over Au/Ag2O is much higher than that generated by Ag2O. That is, the reduction rate of the Au/Ag2O in contact with pure CO was almost six times that of the Ag2O, at least over the initial 20 min. This clearly shows the remarkable ability of the Au NPs to promote the reduction of the surface of the Ag2O support. It is likely that CO molecules adsorbed on the Au NPs migrate to the Ag2O support surface and then create oxygen vacancies according to the Eq. (1)
A similar promotional effect of Au NPs on the reduction of the lattice O of a support oxide has also been reported for the catalytic oxidation of CO over Au/TiO2 based on assessments using in situ electrical conductance measurement [20].
Fig. 2 demonstrates that the rates of reduction over both Au/Ag2O and Ag2O under pure CO are much faster than those obtained when exposed to the CO in the reaction mixture. One possibility is that the reduction of the surface lattice O of the Ag2O by CO is significantly hindered by the presence of O2 in the gas phase, such that the promotional effect of the Au NPs on the reduction of Au/Ag2O is counteracted. In this case, O2 molecules interfere with the reduction of surface lattice O by CO, as expressed by Eq. (1)
Next the reduction process of Ag2O in both the Au/Ag2O and pure Ag2O samples by CO in the reaction mixture was investigated by oxidizing samples in circulating oxygen at 1 atm and 423 K to restore them to their initial oxidized states. The reaction mixture was subsequently introduced to the Au/Ag2O or Ag2O at temperatures between 253 and 313 K to examine the effect of temperature on the rate of reduction by CO in the reaction mixture. The rate constants for the reduction of Au/Ag2O and Ag2O by CO in the reaction mixture, kCO, were calculated from the rates of pressure decrease. In the case of the Au/Ag2O, the rate was determined after the slope break (Fig. 1). Fig. 3 presents Arrhenius plots of kCO, in which the plots of the Au/Ag2O and Ag2O data are seen to coincide and to have almost the same slope. Since the Au NPs supported on the Ag2O do not affect the value of kCO nor the slope of the Arrhenius plot, the reduction of the surface lattice O of these materials by CO in the presence of gaseous O2 presumably occurs on the Ag2O surface apart from the interface with the Au NPs, and must be controlled by some process in which O2 molecules retard the creation of O vacancies.
It is worth noting here some results that were obtained by simple calculations employing the experimental results presented in Fig. 1. These calculations were based on a model in which hemispherical Au NPs with a mean particle diameter of 4.4 nm are dispersed on the Ag2O surface. These results were as follows.
(1) A total of 2.4 μmol of oxide ions were present at the interfaces between Au NPs and Ag2O.
(2) A total of 36.5 μmol of lattice oxide ions were present on the surface monolayer of the Au/Ag2O.
(3) A total of 48.6 μmol of lattice oxide ions were present on the surface monolayer of the Ag2O.
(4) A total of 62 μmol of the reaction mixture (equivalent to the amount of CO) was consumed before the inflection in the Au/Ag2O plot.
From the above, we can make the following conclusions.
(1) The reduction of lattice O on the Au/Ag2O by the CO in the reaction mixture prior to the inflection point far exceeded the amount available at the interfaces around Au NPs and evidently proceeded to the second subsurface layer.
(2) Au NPs appear to promote the reduction of Au/Ag2O by CO in the reaction mixture only in the case of oxide ions on or very near to the upper surface layer and do not affect the reduction of subsurface layers.
The reduction of Ag2O by CO in the reaction mixture proceeded at the same rate observed in the case of the Au/Ag2O after the inflection point seen in Fig. 1. The results presented in Table 2 show that O vacancies were generated in proportion to the progress of the reduction and remained on or near the upper surface layer of the Ag2O. In the absence of O2, both the reduction rate of the Ag2O and the amount of CO2 generated were approximately 8.4 times greater than those obtained when using CO in the reaction mixture, as can be seen from the results for the reduction of Ag2O by pure CO in Fig. 2 and Table 2. It is clear that the presence of O2 in the reaction mixture retards the rate at which O vacancies are produced by the reduction of lattice oxide ions by CO on or near the upper surface layer of the Ag2O where the O vacancies remain.
Catalytic CO oxidation over many reducible metal oxides can be explained by the Mars-van Krevelen mechanism, in which lattice oxide ions are reduced by CO to produce CO2 molecules, and the O vacancies thus formed are replenished by the dissociative adsorption of O2. In other versions, the adsorption of O2 occurs at O vacancies adjacent to lattice oxide ions. In the case of Ag2O, however, adsorbed O2 cannot dissociate and oxidize O vacancies. The fact that the presence of O2 remarkably retards the rate of the production of O vacancies indicates that the adsorption of O2at O vacancies is much stronger than the adsorption of CO on adjacent lattice oxide ions, and inhibits the reduction of the oxide ions by CO to produce CO2 and oxygen vacancies [21].
Here it is helpful to reconsider the origin of the synergistic effect exhibited by Au-Ag bimetallic catalysts for CO oxidation, based on the experimental results obtained with Au/Ag2O and a review of work performed previously with Ag/Au-b [16].
Some experimental results obtained with Ag/Au having oxidized surfaces have already been described briefly in the Introduction, and Table 2 summarizes these data for comparison with the results obtained from Au/Ag2O [16].
The interaction of CO at 17 Pa, which corresponds to 67% of the total amount of surface Ag, with Ag/Au having oxidized surfaces at 313 K for 60 min was found to result in a rapid decrease in the CO pressure over an initial period of 7 min, with a continued gradual decrease up to 60 min [16]. Likewise, in the case of the reduction of Au/Ag2O and Ag2O by CO, the amount of CO consumed was in good agreement with the quantity of CO2 produced (Table 2).
The initial rapid decrease in the CO pressure ceases at approximately the point at which the pressure change corresponds to one half of the amount of surface Ag atoms. Judging from this result, it can be proposed that Ag species on the surface of the Au powder must be oxidized to be nearly all in the +1 state following the pretreatment under 1 atm circulating oxygen at 393 K for 1 h, and that almost all of these Ag species can be reduced to the 0 state by CO at an initial pressure of 17 Pa at 313 K. Judging from the experimental results showing that the reduction rate of Au/Ag2O by CO is about six times that of Ag2O (Fig. 2), the reduction of AgOx species on Ag/Au having oxidized surfaces by CO is evidently greatly enhanced by the presence of the bulk Au substrate.
Table 2 also shows the amounts of reaction mixture consumed and CO2 produced in the experiment during which a reaction mixture with an initial pressure of 11 Pa was introduced successively to Ag/Au-b having oxidized surfaces at 313 K for 60 min [16]. It can be seen that the total amount of CO contained in the reaction mixture over the two exposures (14.7 Pa) was somewhat greater than half the amount of Ag on the surface of the Au powder (13 Pa, see Introduction) [16]. During the first exposure, the pressure of the reaction mixture decreased rapidly over the initial 10 min, although this decrease ceased after 30 min [16]. The amount of CO2 produced (2.56 μmol) was somewhat larger than the amount of reaction mixture consumed (2.23 μmol, Table 2). If the consumption of the reaction mixture had been solely a result of catalytic oxidation with O2, the amount of CO2 produced should have been 1.49 μmol, equal to two-thirds of the reaction mixture consumed (Table 2). This result indicates clearly that only the reduction of reactive O species on the Ag/Au-b having oxidized surfaces by reaction with the CO in the reaction mixture proceeded, without any incorporation of the O2 in the gas phase. Therefore, almost half the reactive O species on the Ag/Au-b having oxidized surfaces were removed as CO2 in the first exposure [16].
Throughout the second exposure, the pressure of the reaction mixture decreased more slowly compared with the first exposure, and continued to decrease until 60 min [16]. Table 2 gives the amount of reaction mixture consumed in the second exposure (2.44 μmol), which again almost coincided with the amount of CO2produced (2.19 μmol). Although the amount of CO2 generated was much greater than two-thirds of the reaction mixture consumed (1.63 μmol), this amount was somewhat less than the quantity of reaction mixture consumed. This result differs from those obtained in the first exposure, indicating that gaseous O2 was incorporated into the CO2 production process due to the catalytic oxidation reaction of CO and O2 [16]. Accordingly, a significant proportion of the CO2 production during the second exposure was still a result of the reduction of O species in the AgOx initially formed on the Ag/Au-b having oxidized surfaces via reaction with CO in the reaction mixture. However, the catalytic oxidation between CO and O2 began to contribute to the CO2 production process at those surfaces where the majority of the reactive O species had been removed by reaction with CO [16].
Comparing the kinetic results obtained using Au/Ag2O with those generated by the Ag/Au having oxidized surfaces described above, it can be seen that both catalysts exhibit similar behavior in conjunction with the reaction mixture. That is, only the reduction of O species by CO proceeds, without incorporation of O2in the gas phase to restore the active O sites. As such, the reduced AgOx species cannot dissociate adsorbed O2 molecules. This might be ascribed to the very weak bond strength between Ag and O, which is the lowest among all the metal oxides, including Fe2O3, TiO2 and CeO2 [22]. With regard to the origin of the synergistic effect shown by Au-Ag bimetallic catalysts, the possibility that the perimeters between AgOx regions and the bulk Au surface function as important active sites during CO oxidation appears unlikely.
Let us review here the experimental results reported for Ag/Au-b having steady state surfaces [16]. The steady state surfaces of Ag/Au-b were supplied by catalytic oxidation using a stoichiometric reaction mixture of CO and O2 with an initial pressure of 5.3 kPa at 313 K for 1 h, followed by exposure to a vacuum of approximately 1.3 × 10-3 Pa [16]. It is worth to note that the amount of CO2 produced while establishing the steady state surface was calculated to be 32 times greater than the number of surface Ag atoms and is in good agreement with two-thirds of the amount of the reaction mixture consumed [16]. The Ag/Au-b having steady state surfaces was subsequently exposed to CO and/or O2 at an initial pressure of 11 Pa [16], and Table 2 presents the amounts of CO and/or O2 consumed and of CO2produced during the exposure to CO and/or O2.
In the case of CO exposure, the first order plot of the pressure decrease data tends to overlap with the data for the catalytic CO oxidation with O2 over the initial 20 min [16]. However, the CO pressure decrease almost ceases after the slope break at 20 min. The amount of CO consumed during the initial 20 min was 0.41 μmol, equivalent to 4.6% of the surface Ag, suggesting that only a small fraction of these surface Ag atoms participated in the catalytic oxidation of CO with O2[16]. The total amount of CO consumed after the exposure for 60 min was 0.54 μmol, whereas the amount of CO2 produced was found to be 0.79 μmol, which is 46% greater than the amount of CO consumed (Table 2) [16]. In contrast, the O2 pressure did not decrease at all during the 60 min trial, although 0.21 μmol of CO2 was produced [16]. Several observations can be made here. First, the CO pressure decreased at the same rate as that of a reaction mixture of CO and O2until the pressure decrease reached a value corresponding to 4.6% of the surface Ag atoms. Second, the amount of CO2 produced during CO exposure was 46% greater than the amount of CO consumed. Last, the O2 exposure generated a small amount of CO2 without exhibiting a decrease in pressure. On this basis, a mechanism has been proposed for CO oxidation with O2 at the steady state surfaces of Ag/Au-b, in which the reaction proceeds at the interfaces between surface Ag and the Au support by repeated co-adsorption of CO and O2 to form COO2, followed by decomposition due to interactions with CO and/or O2 to yield CO2[16]. It should be noted that, in this mechanism, the bond dissociation of O2 molecules is effected through the formation and decomposition of the COO2 complex.
It is evident from the data in Table 2 that both the amounts of CO consumed and of CO2 produced in the experiment using 11 Pa of CO and the Ag/Au-b having steady state surfaces were approximately 10% of the amounts obtained in the trial using CO at the same pressure in conjunction with Ag/Au-b having oxidized surfaces. Furthermore, according to the proposed reaction mechanism, the production of CO2 by CO exposure over the Ag/Au-b having steady state surfaces can be assumed to have resulted from decomposition of COO2 intermediates. Accordingly, we can state that no AgOx species were present on the Ag/Au-b having steady state surfaces; almost all Ag atoms would have been in the 0 state.
Following the reduction of AgOx species to Ag(0) during CO oxidation over the Ag/Au-b having steady state surfaces, it is probable that metallic Ag(0) zones formed on the Au support surface on which the catalytic CO oxidation took place [23]. To examine this possibility, unsupported Ag powder (1.03 m2/g) manufactured by evaporating high purity Ag metal in an inert gas was obtained from the Vacuum Metallurgical Co., Ltd. The catalytic activity for CO oxidation of this material was determined using the same equipment and procedure employed in the measurement of the Ag/Au-b powder [16], and the rate constant for CO oxidation per unit surface area of Ag, kCO2, was calculated. The value of kCO2 over the Ag/Au-b was found to be 30 times greater than that on unsupported Ag powder at 253 K and four times that at room temperature. The apparent activation energies were 7 kJ/mol for the Ag/Au-b and 28 kJ/mol for the unsupported Ag powder. The significant reduction in the activation energy would be expected to greatly increase the kCO2 value for Ag/Au-b below room temperature. These results indicate that the CO oxidation mechanism in the case of Ag/Au-b powder differs from that over unsupported Ag powder owing to the contribution from the Au support. As well, it appears that Ag(0) interacts chemically with the surrounding Au atoms on the support surface of the Ag/Au-b powder. In fact, recent investigations of the chemical environments of Ag atoms in Ag/Au-b using Ag K-edge extended X-ray absorption fine structure (EXAFS) have shown that the Ag atoms are dispersed atomically within two upper surface layers on the Au powder through the formation of Ag-Au bonds and by maintaining Ag-Ag bonds slightly. The EXAFS data also show the absence of a peak attributable to Ag-O bonds [17]. These results demonstrate that Ag atoms dispersed on the surface of Ag/Au-b bond not with O but with surrounding Au and Ag atoms to form a Ag-Au alloy [17]
Zhang et al. [5] reported that PVP-protected Agcore/Aushellbimetallic NPs exhibited enhanced catalytic activity during the liquid phase aerobic oxidation of glucose. They proposed that the formation of Au-Ag bonds induced charge transfer from Ag to Au and ascribed the higher activity of the Ag/Au NPs for glucose oxidation to the presence of negatively charged Au atoms [5]. Chang et al. [24] performed a theoretical investigation of CO oxidation on unsupported icosahedral Au25Ag30 nanoclusters. The reaction path they proposed involves an OOCO intermediate formed through the coadsorption of CO on Au sites and O2 on Ag sites [24], which agrees with the mechanism we propose for Ag/Au-b [16]. Furthermore, they reported that the co-adsorption energy required to produce the COO2 complex from CO and O2 on the Au25Ag30 alloy clusters is larger than the values calculated for monometallic Au55 and Ag55 clusters [24], which should lead to a higher reactivity for CO oxidation.
The combination of Ag and Au can therefore be concluded to enhance the catalytic activity of Ag/Au-b for CO oxidation with O2 by lowering the activation barrier to the formation and decomposition of the COO2 and this may explain the synergistic effect seen in other Ag-Au bimetallic catalysts.
This work investigated the enhanced catalytic activity for CO oxidation exhibited by Ag-contaminated Au powder after an oxidizing pretreatment. To determine if this effect can be explained by the generation of a contact interface between AgOx and the Au support, the catalytic activity for CO oxidation over Au/Ag2O was compared with that over Ag2O. Differences between Au/Ag2O and Ag2O in the rate of reaction mixture pressure decrease were absent in the steady state. In the case of both catalysts, the pressure decrease was attributed to the reduction of surface lattice O on Ag2O by CO in the reaction mixture. The kinetic results obtained with Au/Ag2O were similar to those observed when working with Ag/Au having oxidized surfaces, indicating that it is unlikely that the interfaces between AgOx regions and the bulk Au surface function as important active sites on Ag/Au-b. AgOx species in Ag/Au-b having oxidized surfaces produced by an oxidizing pretreatment was completely reduced to the 0 state under steady state CO oxidation using a reaction mixture composed of CO and O2 with a composition of CO/O2=2/1. A comparison of the kCO2 and activation energy values between Ag/Au-b and unsupported Ag powder indicated that Ag atoms on the surface layer of the Ag/Au-b powder form a Ag-Au alloy. This theory agrees with the results of recent investigations concerning the chemical environment of Ag atoms in Ag/Au-b by means of Ag K-edge EXAFS.
The authors hope to express their deep thanks to Professor Masatake Haruta of Tokyo Metropolitan University for his kind discussion which gave us a hint to start this study. They also hope to express their deep thanks to Dr. Tetsuo Honma and Dr. Hiroshi Oji of Japan Synchrotron Radiation Research Institute for their kind help in the EXAFS measurement performed at BL14B2 of Spring-8.
2016,
Vol. 37 Issue (10): 1712-1720 
PDF
2016,
Vol. 37 Issue (10): 1712-1720 PDF