Abstract

Au/BiVO4 visible-light-driven photocatalysts were synthesized by coprecipitation method in the presence of sodium dodecyl benzene sulfonate (SDBS) as a dispersant. Physical characterization of the obtained materials was carried out by X-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDX), UV-Vis diffuse reflectance spectroscopy (DRS) and Brunauer, and Emmett and Teller (BET) specific surface area measurement. Photocatalytic performances of the as-prepared Au/BiVO4 have also been evaluated via mineralizations of oxalic acid and malonic acid under visible light irradiation. XRD and SEM results indicated that Au/BiVO4 photocatalysts were of almost spherical particles with scheelite-monoclinic phase. Photocatalytic results showed that all Au/BiVO4 samples exhibited higher oxalic acid mineralization rate than that of pure BiVO4, probably due to a decrease of BiVO4 band gap energy and the presence of surface plasmon absorption upon loading BiVO4 with Au as evidenced from UV-Vis DRS results. The nominal Au loading amount of 0.25 mol% provided the highest pseudo-first-order rate constant of 0.0487 min−1 and 0.0082 min−1 for degradations of oxalic acid (C2) and malonic acid (C3), respectively. By considering structures of the two acids, lower pseudo-first-order rate constantly obtained in the case of malonic acid degradation was likely due to an increased complexity of the degradation mechanism of the longer chain acid.

1. Introduction

In the past few years, interest has been paid to research on water remediation with the application of an ideal “green” technology known as semiconductor photocatalysis. It has been widely accepted that this process successfully combines the principle of heterogeneous catalysis with a utilization of solar energy. By using this photocatalytic process, degradation of a wide range of organic pollutants into harmless carbon dioxide and water is made possible. Titanium dioxide, a well-known UV-light-active photocatalyst, has demonstrated an outstanding photocatalytic performance on degradation of various organic compounds [13]. However, with its wide band gap energy of 3.2 eV, the application of TiO2 is limited to UV light region which accounts for only 4% of the whole solar energy [4]. Therefore, extensive research has currently been devoted to the development of visible-light-driven catalyst in order to effectively utilize the vast majority of the solar energy [46]. Bismuth vanadate (BiVO4) has long been recognized for its ferroelasticity [7] and its application as a nontoxic and bright yellow pigment [8]. It has also been used as a gas sensing semiconductor, solid-state electrolyte, and cathode material in solid oxide fuel cells, and has recently been proved to be an active visible-light-responsive photocatalyst for water splitting [9] and organic pollutant decomposition [10, 11]. Since BiVO4 is stable and neutral in water without altering the solution pH, its application as photocatalyst for environmental treatment is extensively investigated. The photocatalytic property of BiVO4 is strongly dependent on its morphology and crystalline form [1214]. Generally, synthetic BiVO4 was found to exist in three crystalline phases including scheelite-monoclinic type, scheelite-tetragonal type, and zircon-tetragonal type [15]. Among the three polymorphs, the scheelite-monoclinic structure with band gap energy of 2.4 eV is reported to possess the highest photocatalytic activity [4, 13, 16]. Therefore, many synthesis methods have been focused on the selective preparation of scheelite-monoclinic BiVO4 photocatalyst.

Different synthetic routes have previously been employed to prepare the scheelite-monoclinic BiVO4 such as traditional solid-state reaction [17] and hydrothermal method [18]. However, these strategies have encountered similar problem in which the obtained BiVO4 possessed very low surface area, normally in the range of less than 10 m2 g−1 and, as a consequence, low photocatalytic performance has usually been attained. Apart from the low surface area of BiVO4, difficult separation of photogenerated electron-hole pair was also reported to be one of the main reasons accounting for its poor photocatalytic efficiency [12, 19]. However, by loading BiVO4 with only small amount of metals such as Pt [11], Au [12] Pd [19], Ag, Co, and Ni [20], enhanced photocatalytic activity was achieved possibly due to the metals acting as electron traps, thus promoting electron-hole separation and the interfacial charge-transfer process from catalyst to adsorbed substrate [12, 19, 20]. However, there are few reports on the development of Au/BiVO4 composite to affect photocatalysis under visible-light irradiation. Recently, Cao et al. [21] reported that the Au/BiVO4 composite showed superior visible-light activities in decomposing methyl orange dye. However, Long et al. [22] synthesized Au/BiVO4 composite photocatalysts and found that the photocatalysts exhibited enhanced visible-light photocatalytic activities on degradation of phenol. Since dicarboxylic acids are generally observed as intermediate products in the degradation pathways of various organic pollutants in real wastewaters [2325], the influence of mutual interactions on the photocatalytic conversion process needs to be investigated. However, there has been no report that demonstrates the simultaneous detoxification of the dicarboxylic acids by Au/BiVO4 composite. Herein, we report the preparation and photocatalytic performance of a visible-light-driven Au/BiVO4 catalyst. BiVO4 with the scheelite-monoclinic structure was prepared by surfactant-assisted coprecipitation method and then subsequently impregnated with HAuCl4 solution to finally obtain Au/BiVO4. Photocatalytic performances of the as-prepared scheelite-monoclinic Au/BiVO4 samples were evaluated through the mineralizations of oxalic acid and malonic acid under visible light irradiation.

2. Experimental

2.1. Synthesis of Au/BiVO4

Bismuth nitrate pentahydrate (Bi(NO3)3·5H2O, 99%, Ajax) and ammonium vanadate (NH3VO4, 99%, Carlo) were used as bismuth and vanadium precursors. Sodium dodecylbenzene sulfonate (C18H29SO3Na (SDBS), 98%, Aldrich), an anionic surfactant, was employed as a dispersant in this study. All chemicals were used as received without further purification. Firstly, 0.125 M each of Bi and V were separately prepared by dissolving Bi(NO3)3·5H2O in 4.0 M nitric acid solution and NH3VO4 in 4.0 M ammonia solution. The as-prepared bismuth nitrate solution was then mixed with 0.1 M SDBS in ethanol. To this mixture, the vanadium precursor solution was slowly added and the solution was kept under stirring for 30 min. Then 4.0 M ammonia solution used as a precipitant was added drop-wise until pH 7 was attained. The resultant precipitate was washed with deionized water, centrifuged, and dried at 60°C for 12 h. The dried powder was then calcined at 400°C for 2 h to obtain BiVO4 sample. The as-prepared BiVO4 sample was then impregnated with aqueous solution of gold chloride (HAuCl4·2H2O, ≤48% Aldrich) containing the nominal gold amounts of 0.10, 0.25, 0.50, 0.75, and 1.00 mol%. Then the impregnated powder was dried at 60°C for 3 h and subsequently calcined at 350°C for 2 h to finally obtain Au/BiVO4.

2.2. Characterization

Powder X-ray diffraction (XRD) measurement was performed on an X-ray diffractometer (JEOL, JDX-3530) using Cu K radiation ( Å) and scanning from 10° to 75°. Specific surface area (SSA) of the particles was measured on Beckman Coulter SA 3100 according to the Brunauer-Emmett-Teller (BET) method. UV-Vis diffuse reflectance spectra were obtained on a UV-Vis spectrometer (PerkinElmer, Lambda 650S) using MgO as a reference and were converted to absorbance by Kubelka-Munk method [26]. Particle morphology and chemical composition were also investigated on a scanning electron microscope (Hitachi, S3400N) equipped with an energy dispersive X-ray spectrophotometer (Oxford, ISIS300).

2.3. Photocatalytic Activity

Visible light photocatalytic activities of the as-prepared Au/BiVO4 powders were evaluated by the mineralization of oxalic acid (0.208 mM) and malonic acid (0.139 mM) in aqueous solution at ambient temperature and pressure. The photocatalytic studies were performed using a 100 mL spiral photoreactor equipped with a fluorescent lamp (Davis 33 cool white, 18 W, intensity of 4.39 mW/m2), filtered with double-layer of Rosco E-colour UV filter to remove any UV component (), in the middle of the reactor. Typically, 50 mL of 1.0 g L−1 catalyst suspension was prepared by dispersing the predetermined amount of catalyst in deionized water with ultrasonic probe for 15 min. The suspension pH was then adjusted to using 1 M perchloric acid solution before charging into the spiral reactor. Prior to catalytic testing, adsorbed carbon contaminants on the catalyst surface were firstly removed by illuminating the catalyst suspension with UV light for 1 h. Photocatalytic mineralizations of oxalic acid and malonic acid were then carried out by injecting 100 μL of organic compound solution containing 500 μg of carbon. Adsorption/desorption equilibrium of the organic substrates on the catalyst surface was attained by circulating the suspension for 30 min under the dark. Then the system was irradiated and the photocatalytic reaction was initiated. Dissolved carbon dioxide (CO2) in water, generated from dicarboxylic acid, could be detected by online conductivity meter (Eutech 5000) as described by Abdullah et al. [27].

3. Results and Discussion

3.1. Physical Property of Au/BiVO4

Figure 1 illustrates XRD diffraction patterns of pure BiVO4 and Au/BiVO4 with different Au loading amounts. The XRD patterns revealed sharp peaks, indicating high crystallinity of the obtained particles. XRD patterns of all samples presented similar profiles and the diffraction peaks matched well with scheelite-monoclinic BiVO4 (JCPDS file no. 14-0688). However, these obtained samples are not well-distorted scheelite-monoclinic BiVO4 because there is no peak at 15° of 2θ and the peaks at 18.5°, 35° and 46° of 2θ are not well split [28].

Results from XRD suggested that the scheelite-monoclinic BiVO4, although with less distortion, was selectively prepared by using SDBS-assisted coprecipitation method. Upon loading pure BiVO4 with Au, no significant change in diffraction patterns was observed which suggested that loading with Au up to 1.0 mol% did not affect the crystal structure of scheelite-monoclinic BiVO4. No other peak due to Au metal was found possib because the Au loading amount was small and high dispersion of Au was obtained as supported by EDX spectra in Figures 2(d) and 2(f) [29]. BET specific surface area (SSA) of the resulted BiVO4 was about 23 m2 g. Upon loading the obtained BiVO4 with Au from 0.1–1.0 mol%, SSA of the sample was gradually increased from 23 m2 g−1 to 31 m2 g−1.

Morphologies of pure BiVO4, 0.25 mol% Au/BiVO4, and 1.0 mol% Au/BiVO4 as well as the presence of Au on the surface of Au/BiVO4 samples were studied by SEM and EDX as illustrated in Figure 2. As seen from the SEM images in Figures 2(a), 2(c), and 2(e), pure BiVO4, 0.25 mol% Au/BiVO4, and 1.0 mol% Au/BiVO4 were formed in an almost irregular morphology with the particle size in the range of 300–800 nm. No significant change in terms of BiVO4 particle size was observed upon loading pure BiVO4 with 0.25 or 1.0 mol% of Au. The existence of Au in 0.25 mol% Au/BiVO4 sample was not clearly observed from the EDX spectrum (Figure 2(e)); however, the peak due to Au metal was clearly seen from the nominal 1.0 mol% Au/BiVO4 as shown in Figure 2(f). The atomic percentage of each element is given in Table 1. The amount of Au in both samples was less than the loading amount since the Au nanoparticles were not evenly distributed on the BiVO4 support; therefore, the selected areas had less amounts than the actual amounts.

3.2. Optical Absorption Behavior

The inset of Figure 3(a) indicated that pure BiVO4 had no absorption in the region of 550–800 nm; however, all Au/BiVO4 samples showed enhanced absorption in this region upon increasing Au content. The absorption in this range can be ascribed to the surface plasmon resonance (SPR) of Au nanoparticles which is attributed to a collective of conduction electrons in response to optical excitation [30]. In addition, the SPR peak was slightly shifted to longer wavelength as increasing Au content, possibly due to an increase of Au particle size [31]. Optical absorption near the band edge and represent, a constant, absorption coefficient and the incident photon energy, respectively [4, 32]. The value depends on the characteristics of the electron transition in a semiconductor. Since the electron transition in BiVO4 is a direct transition, the value of [4, 19, 32, 33]. The band gaps were estimated () by using the intercept of the tangent to the -axis as illustrated in Figure 3(b). The estimated band gap energy of pure BiVO4 was 2.53 eV. However, upon loading BiVO4 with the nominal gold amount of 0.1−1.0 mol%, a small shift towards lower band gap energy in the range of 2.48–2.50 eV was observed, probably due to a charge-transfer transition between Au and BiVO4 [12].

3.3. Photocatalytic Property

Photocatalytic activities of scheelite-monoclinic BiVO4 and Au/BiVO4 were evaluated by studying the mineralizations of two model organic compounds which were oxalic acid and malonic acid in aqueous solution under visible light illumination. Photocatalytic mineralizations of oxalic acid and malonic acid over the as-prepared Au/BiVO4 as a function of irradiation time are shown in Figures 4(a) and 4(b), respectively.

Results from Figure 4(a) also indicated that a remarkable photocatalytic performance was obtained from 0.25 mol% Au/BiVO4. Further loading of Au more than 0.25 mol% resulted in a decreased photocatalytic activity. Therefore, an optimum Au loading amount in this study is 0.25 mol%. The existence of an optimum dopant concentration was previously explained by Zhang et al. [34]. Therein, the author proposed that, at low dopant concentration, metal ion dopant can act as a trap for both electron and hole which then leads to a lengthening in the lifetime of the generated charge carriers and thus resulting in enhanced photocatalytic efficiency. However, at high dopant concentration, the charge trapping is high and as such, the charge carrier recombination through quantum tunneling is highly possible [34, 35]. Therefore, there exists an optimum dopant concentration. A comparison between Figures 4(a) and 4(b) indicated that longer irradiation time was required for malonic acid to attain complete mineralization. This is probably due to the longer hydrocarbon chain length of malonic acid compared with that of oxalic acid.

Photocatalytic mineralization rates of oxalic acid and malonic acid over 0.25 mol% Au/BiVO4 were investigated and presented in Figure 4(b). The obtained results were found to fit well with Langmuir-Hinshelwood (LH) kinetics as evidenced by high correlation coefficient values () of 0.99 and 0.94 for degradations of oxalic acid and malonic acid, respectively.

The LH kinetic expression is given by [36, 37] where represents the initial mineralization rate of organic substrate, is the concentration of the substrate at an illumination time , and and are the mineralization rate constant and the adsorption coefficient of the reactant, respectively. Integration of (1) yields (2): where is the initial concentration of the organic substrate and in the concentration of the substrate at time . When is very small, (2) can be reduced to (3), where is the initial apparent rate constant of a pseudo-first-order reaction. By plotting versus as shown in Figure 4(b), values for photocatalytic mineralizations of oxalic acid over pure BiVO4, 0.25 mol% Au/BiVO4 and malonic acid over 0.25 mol% Au/BiVO4 can be obtained from slopes of the graphs. Good regression coefficients observed in this work indicated that the kinetics of dicarboxylic acid mineralizations followed a simplified Langmuir-Hinshelwood rate equation (3) with the pseudo-first order rate constants for oxalic acid degradation over pure BiVO4 and 0.25 mol% Au/BiVO4 of 0.0188 min−1 and 0.0487 min−1, respectively, as shown in Figure 5. The of 0.25 mol% Au/BiVO4 was more than twice higher compared to pure BiVO4. The of oxalic degradation was also found to be higher than that of malonic acid ( = 0.0082 min−1). Generally, photocatalytic mineralization rate is governed by both structural and functional properties of the target molecule.

In this work, the two acids used as the targeted substrates are dicarboxylic acid; therefore, functional characteristic could not be the main reason explaining the difference in rate constants. However, by considering the structures of these acids, oxalic acid (C2) with shorter carbon chain length than malonic acid (C3) could provide higher apparent rate constant. Our finding is in good agreement with that of Denny et al. [38]. Therein, a decrease in 50% mineralization rate with increasing carbon chain length was observed and the authors ascribed this behavior to the increased complexity of the degradation mechanism of the longer chain hydrocarbons. Therefore, the decrease in observed in our study was likely attributed to an increase in carbon chain length of the targeted molecule. However, other factors including an increased steric hindrance effect, a decrease of the positive charge on the carbon of carboxyl group as increasing carbon chain length, thus making hydroxyl radical attraction more difficult may also have some contribution to the obtained apparent rate constant [38, 39].

4. Conclusion

Significantly improved photocatalytic efficiency was observed upon loading pure monoclinic-scheelite BiVO4 with low concentrations of gold dopant (0.1–1.0 mol%). Maximum apparent rate constant observed from 0.25 mol% Au/BiVO4 was found to be more than twice higher compared to pure BiVO4. However, loading Au more than 0.25 mol% was detrimental to the photocatalytic activity since excess Au atoms may act as charge recombination centers, resulting in a decrease of charge carrier lifetime and low photocatalytic performance. The initial apparent pseudo-first-order rate constants of 0.25 mol% Au/BiVO4 were found to be 0.0487 and 0.0082 min−1 for the degradations of oxalic acid and malonic acid, respectively. By considering structures of the two acids, lower pseudo-first order rate constant obtained in the case of malonic acid degradation was likely due to an increased complexity of the degradation mechanism of the longer chain acid. However, the fact that an enhanced steric hindrance effect and a decrease of the positive charge on the carbon of carboxyl group may affect the observed rate constantly could not be neglected.

Acknowledgments

The authors gratefully acknowledge the financial support from Thailand Graduate Institute of Science and Technology (TGIST) and The National Science and Technology Development Agency (NSTDA) to Kanlaya Pingmuang; the Research, Development and Engineering (RD&E) fund through the National Nanotechnology Center (NANOTEC), NSTDA, Thailand (P-11-00982), for Chiang Mai University to Burapat Inceesungvorn; the National Research University Project under Thailand’s Office of the Higher Education Commission, the Graduate School and Faculty of Science (Chiang Mai University) and the NANOTEC, NSTDA, Ministry of Science and Technology, Thailand, through its program of Center of Excellence Network. The authors would like to thank Mr. Ekkapong Kuntarak for his help in spectroscopy measurements.