Reaction Mechanism and Activity of WO3-Catalyzed

Mar 30, 2009 - Efficient Visible Light Active CaFe2O4/WO3 Based Composite Photocatalysts: Effect of Interfacial Modification. Zhifu Liu , Zhi-Gang Zha...
0 downloads 0 Views 354KB Size
Download PDF
6602

J. Phys. Chem. C 2009, 113, 6602–6609

Reaction Mechanism and Activity of WO3-Catalyzed Photodegradation of Organic Substances Promoted by a CuO Cocatalyst Takeo Arai,† Masumi Horiguchi,†,‡ Masatoshi Yanagida,† Takahiro Gunji,†,‡ Hideki Sugihara,† and Kazuhiro Sayama*,† Energy Technology Research Institute, National Institute of AdVanced Industrial Science and Technology (AIST), Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan, and Graduate School of Science and Technology, Tokyo UniVersity of Science, 2641 Yamasaki, Noda, Chiba 278-8510, Japan ReceiVed: December 17, 2008; ReVised Manuscript ReceiVed: March 2, 2009

WO3 is known as a visible-light-responsive photocatalyst; however, WO3 has low activity for the photodegradation of organic substances in the absence of cocatalysts. The activity of WO3 was significantly improved by the loading of CuO as a cocatalyst, and this catalyst system can be used practically because it is inexpensive and easy to prepare. We investigated the properties of a CuO/WO3 catalyst by means of XPS measurements, electrochemical measurements, and a layered model reaction, and we proposed a reaction mechanism of CuO/WO3 in which CuO promotes oxygen reduction for WO3-catalyzed photodegradation by the Cu(II)/Cu(I) redox reaction. Furthermore, the oxygen reduction reaction over CuO was improved by light irradiation of CuO. The optimized CuO/WO3 photocatalyst showed higher activity for the photodegradation of acetaldehyde than did N-TiO2 and TiO2 under fluorescent-lamp irradiation simulating that used in indoor applications. Introduction Photocatalysis with sunlight or indoor light has the potential to be a beneficial technology for the degradation of harmful volatile organic compounds (VOCs) such as formaldehyde, acetaldehyde, and toluene. For indoor applications, the fraction of UV light available is very low; therefore, visible-lightresponsive photocatalysts are needed instead of UV-responsive photocatalysts such as TiO2. In recent years, many semiconductor photocatalysts, particularly those based on TiO2, such as M-TiO2 [M ) N, C, S, metals], have been intensively investigated.1-5 In the case of the TiO2-based photocatalysts, the band structure of TiO2 is modified by the introduction of midgap levels from other orbitals of anions such as N2p to achieve both visible light absorption and a sufficiently negative conduction band (CB) level for oxygen reduction. However, so far, the ability of these catalysts to completely oxidize organic compounds to CO2 under visible light has not been satisfactory. This lack of conversion efficiency might be related to the low oxidation ability (more negative potential) of the hole on the N2p orbital of N-TiO2 compared with that of the O2p orbital of pure TiO2. In contrast, WO3 is a visible-light-responsive n-type semiconductor and is a known photocatalyst for oxygen generation with sacrificial reagents such as Ag+.6-9 The valence band potential of WO3 is almost the same as that of TiO2;10 therefore, the oxidative ability of hole on the WO3 valence band is considered to be almost the same as that on TiO2. However, reports on the degradation of organic substances over WO3 have been limited until recently,11-24 and the activity of WO3 without cocatalyst is very low compared with that of TiO2.16-18,20,21 One important difference between WO3 and TiO2 is their conduction band bottom potentials. The conduction band level of WO3 * Towhomcorrespondenceshouldbeaddressed.E-mail:[email protected]. Phone: +81-(0)29-861-4760. Fax: +81-(0)29-861-4760. † National Institute of Advanced Industrial Science and Technology. ‡ Tokyo University of Science.

(+0.5 V vs NHE)25 is more positive than that of TiO2 (around -0.2 to -0.4 vs NHE),26 suggesting that reduction of oxygen by the electron on the conduction band of WO3 is more difficult than that of TiO2. Recently, WO3-catalyzed photodegradation of organic substances was found to be promoted by the use of suitable cocatalysts, and the development of WO3 photocatalysts has attracted attention.15-24 We have found Pd24 and various Cu compounds such as CuO, CuBi2O4, and copper ions16,17,20 to be effective cocatalysts over WO3 for the complete photodegradation of organic substances under visible light. Similar cocatalytic effects on WO3 have been reported by Irie et al. for the photodegradation of 2-propanol over Cu(II)-grafted WO3.22 Cu loading also enhances the photodegradation of acetaldehyde and acetic acid over N-TiO2, as reported by Morikawa et al.27,28 It is reported that WO3 loaded with nanoparticulate Pt exhibits high photocatalytic activity for the decomposition of organic compounds in both liquid and gas phases.21 If these cocatalysts can facilitate the multiple-electron reduction of oxygen (O2/ H2O2 ) +0.68 V vs NHE; O2/H2O ) +1.23 V vs NHE), then the electron on the WO3 conduction band could reduce oxygen effectively. The multiple-electron reduction of oxygen over Pt and Pd electrodes has been reported;29 however, Pt and Pd cocatalysts are expensive and resource-constrained. So, in this study, we focused our attention on the CuO cocatalyst, since CuO is inexpensive and easy to prepare. We attempted to clarify the reaction mechanism of CuO/WO3 by measuring the valence change using XPS and electrochemical properties of CuO. We also evaluated the apparent quantum efficiency of CuO/WO3 in the photodegradation of various organic substances and compared CuO/WO3 photodegradation activity with those of N-doped TiO2 and TiO2 under fluorescent irradiation simulating that used in indoor applications. Photodegradation of acetaldehyde by CuO/WO3 was also performed in a flow reactor.

10.1021/jp8111342 CCC: $40.75  2009 American Chemical Society Published on Web 03/30/2009

WO3-Catalyzed Photodegradation of Organic Substances Experimental Methods WO3 powders were obtained from various chemical companies (Kanto Chemical Co., Inc., Kojundo Chemical Laboratory Co., Ltd., and Wako Pure Chemical Industries, Ltd.). Commercially available transition metal oxides (TiO2, V2O5, Cr2O3, MnO2, Fe2O3, CoO, NiO, CuO, and ZnO) were used as cocatalysts. Each cocatalyst powder (2 wt %) was thoroughly mixed with WO3 powder in a mortar. In addition, WO3 loaded with 0.1 wt % CuO was prepared by an impregnation method: Cu(NO3)2 aqueous solution (0.02 mol/L, 1 mL) was mixed with WO3 powder (ca. 1.6 g) and the mixture was dried on a hot plate and then fired at 300 °C for 30 min in air. Thermogravimetry-differential thermal analysis measurement confirmed that Cu(NO3)2 was converted to CuO at 300 °C. N-TiO2, prepared by NH3 treatment of TiO2 (Ishihara Sangyo Kaisha, Ltd., ST01),as reported by Irokawa et al.,2 and TiO2 (Degussa Co., Ltd., P-25) were used as the reference photocatalysts. Acetaldehyde, formaldehyde, acetic acid, and formic acid were used in the photocatalytic degradation reaction. Photocatalytic degradation of acetaldehyde and the other organic substances proceeded in a borosilicate glass vial (4.4 mL) with a septum cap for sampling. An amount of photocatalyst sufficient to completely cover the bottom of the vial was placed into the vial, the vial was sealed with the cap, and then the organic substances were introduced into the vial by means of a syringe. Acetaldehyde was introduced as a gas (40 µL), and formaldehyde (1 mol/L, 2 µL), acetic acid (0.5 mol/L, 2 µL), and formic acid (1 mol/L, 2 µL) were introduced as aqueous solutions. The bottom of the vial (area: 1.1 cm2) was irradiated with light from a xenon lamp (PerkinElmer Co.,Ltd., Cermax LX-300) or a neutral white fluorescent lamp (Panasonic Electric Works Co., Ltd., FPL13EX-N). The full arc of the xenon lamp was used for UV and visible light irradiation. The intensity at 400-600 nm was adjusted to 1 sun (air mass 1.5) with use of a spectroradiometer (Yamashita Denso Corporation, YSR-1100M). A UV cutoff filter (Hoya Corporation, L-42, λ > 400 nm) and a bandpass filter (AGC Techno Glass Co., Ltd., UV-D35, λ ≈ 350 nm) were used with the xenon lamp for visible light irradiation and for UV light irradiation, respectively. In the evaluation of apparent quantum efficiency, a xenon lamp (Ushio Inc., UXL-500-D-O) equipped with a monochromator was used as a light source to obtain monochromatic light at λ ) 400 nm. The intensity of the monochromatic light was ca. 29 µW. The organic substances introduced in the reactor were gaseous acetaldehyde (40 µL), liquid acetic acid (2 µL), and formaldehyde solution (16 wt % in aqueous solution, 2 µL). Photocatalytic degradation of acetaldehyde was also carried out in a flow-type reactor consisting of a stainless-steel cell with an acrylic plate and a window made of Pyrex glass. A glass plate slide was coated with CuO/WO3 photocatalyst by means of the doctor blade method. Acetaldehyde gas was diluted to ca. 2 ppm with humid air (ca. 50% humidity). The flow rate was adjusted to 50 mL/min. A neutral white fluorescent lamp was used as the light source, and the illuminance on the photocatalyst film (ca. 60 cm2) was adjusted to ca. 3000 lx. The concentrations of acetaldehyde and CO2 generated in the photodegradation reactions were measured by means of a gas chromatograph (Shimadzu Corporation, GC-2014) with a flame ionization detector equipped with a methanizer. To determine the intermediates generated in photodegradation of acetaldehyde, the WO3 powders after 1 h of reaction were rinsed with 1 mL of deionized water and the rinsing was filtered and analyzed by high-performance liquid chromatography (Tosoh Corporation, TSK-GEL SCX, 7.8 mm i.d. × 30 cm)

J. Phys. Chem. C, Vol. 113, No. 16, 2009 6603 with an aqueous phosphate solution as an eluent (ca. 0.02 M), and detection was done with a UV (λ ) 210 nm) detector. The change in the bulk structure of the CuO cocatalyst was evaluated by means of an X-ray diffractometer (XRD, Mac Science Ltd., MX-Labo). The valence change of copper on the surface of the CuO cocatalyst before and after reaction was evaluated by means of X-ray photoelectron spectroscopy (XPS, Ulvac-Phi, Inc., XPS-1800). An Al anode with a monochromator was used to significantly reduce the background signal. The binding energy was referenced to the C-1s peak at 285.2 eV. To investigate the behaviors of CuO for the oxygen-reduction reaction, electrochemical measurements were performed. The effect of oxygen gas on a WO3 electrode coated with CuO was examined under air or bubbling nitrogen. A porous WO3 electrode was prepared from a precursor solution of tungstic acid in polyethylene glycol, as reported by Augustynski et al.30 The CuO film electrode was prepared from a Cu(NO3)2 aqueous solution (0.1 mol/L). Conductive glass (F-doped SnO2, 10 Ω/cm2) was spin-coated with the precursor solutions and fired at 550 °C for 30 min. A WO3 electrode coated with CuO (CuO/ WO3 electrode) was prepared by an impregnation method: the WO3 electrode was coated with Cu(NO3)2 solution and fired at 300 °C for 30 min in air. Electrochemical measurements were conducted with a potentiostat (BAS Inc., model 630a) and a Pyrex glass cell. A Pt wire and an Ag/AgCl electrode (0.21 V vs NHE, pH 0) were used as the counter and reference electrodes, respectively. A Na2SO4 aqueous solution (0.1 mol/ L, pH 5.8) was used as the electrolyte. The effect of light irradiation on CuO was also examined by comparing the photocurrent observed when a WO3 electrode was irradiated with light and when a CuO electrode was irradiated with light or not irradiated. The WO3 electrode was prepared as mentioned above. The CuO electrode was prepared from an enhanced metal organic decomposition (EMOD) solution (Symetrix Corporation). The solution was coated on conductive glass by means of a spin coater, and then the coated glass was fired at 550 °C in air for 30 min. Electrochemical measurements were conducted in a two-electrode system without external bias by using a two-compartment cell separated with a Nafion membrane. WO3 and CuO electrodes were used for the working and counter electrodes, respectively. They were connected to each other without any external bias. A mixed solution of Na2SO4 (0.1 mol/L, 20 mL, pH 5.8) and ethanol (5 mL) was used for the working electrode compartment. Na2SO4 solution (25 mL) was used as a supporting electrolyte for the counter electrode compartment, and the solution was bubbled with air before measurement. Two xenon lamps were used to irradiate the WO3 and CuO electrodes, respectively. Results and Discussion 1. Specific Cocatalyst Effect of CuO on WO3-Catalyzed Photodegradation. Various low-cost transition metal oxide cocatalysts were mixed with WO3 (Kanto Chemical Co., Inc., surface area ca. 5 m2 g-1), and the activities of these mixtures were compared with the activities of bare WO3 and TiO2. Figure 1 shows the amount of CO2 generated during photocatalytic degradation of acetaldehyde over various photocatalysts under visible light irradiation for 3 and 6 h. After 3 h, the amount of CO2 produced over WO3 without any cocatalyst was about half the amount of CO2 expected (3.2 µmol) from complete oxidation of the introduced acetaldehyde (1.6 µmol). The amount of CO2 did not increase even after 6 h of irradiation (Figure 2i), suggesting that stable intermediates were produced and accumulated on the bare WO3. Only when CuO was used as a

6604 J. Phys. Chem. C, Vol. 113, No. 16, 2009

Figure 1. Amounts of CO2 produced by photocatalytic degradation of acetaldehyde gas (ca. 1.6 µmol) over TiO2, WO3, or WO3 mixed with various cocatalysts (2 wt %) under visible light irradiation (λ > 400 nm). Complete oxidation of acetaldehyde corresponds to ca. 3.2 µmol of CO2.

Figure 2. Time courses of (i) CO2 formation and (ii) CH3CHO consumption during photocatalytic degradation of acetaldehyde gas (ca. 1.6 µmol) over CuO/WO3, WO3, or CuO (ca. 150 mg) under visible light irradiation (λ > 400 nm).

cocatalyst was the total amount of CO2 higher than the total amount produced over bare WO3. The total amount of CO2 generated over CuO/WO3 after 6 h of irradiation was ca. 3.2 µmol, suggesting that acetaldehyde was completely oxidized to CO2. The activity of CuO/WO3 was obviously higher than that of TiO2 under visible light irradiation. The other transition metal oxides produced less CO2 than did bare WO3 at 3 h, and the CO2 concentration was hardly increased when the irradiation time was increased from 3 to 6 h. These results suggest that

Arai et al. the loading of most of transition metal oxidessespecially V2O5, MnO2, Fe2O3, or CoOson WO3 deactivated the WO3 catalyst. The effects of other cocatalysts, mainly noble metals and Cu compounds (Cu, Ru, Rh, Pd, Ag, Ir, Pt, Au, PdO, Cu2O, Ag2O, RuO2, CuCl2, Cu(NO3)2, CuSO4), were also investigated. All Cu compounds, as well as Pt, Pd, and PdO, showed the promotion effect for WO3-catalyzed photodegradation. Of these effective Cu cocatalysts, CuO is the most practical because it is stable and easily prepared compared with the other Cu compounds. 2. Photocatalytic Activity of CuO/WO3 in the Complete Oxidation of Acetaldehyde and Its Intermediates into CO2. Photodegradation of acetaldehyde and its intermediates with CuO/WO3 was performed under visible light conditions. The photocatalytic ability to completely oxidize organic compounds into CO2 is a very important characteristic for the application of environmental purification, because there is a possibility that some partially oxidized byproduct may be more stable and more harmful than the parent organic compounds. Therefore, we checked the complete oxidation of organic substances by CO2 generation. The time courses of both the CO2 formation and the acetaldehyde consumption over CuO/WO3 under visible light irradiation were compared with the respective time courses over WO3 and CuO (Figure 2, panels i and ii). The initial rate of CO2 generation on bare WO3 was high; however, CO2 generation became saturated long before photooxidation was complete (Figure 2i), even though almost all the acetaldehyde in the gas phase disappeared (Figure 2ii). The amount of oxygen in the reactor was sufficient for complete oxidation of acetaldehyde. The rate of CO2 generation was much slower than the rate of acetaldehyde consumption, suggesting that stable intermediates were formed over WO3 during the photoreaction. Acetic acid, formaldehyde, and formic acid were detected as intermediates by high-performance liquid chromatography in the rinsing water of the bare WO3 powder during the photoreaction. We concluded that the high initial rate of CO2 generation on bare WO3 was caused by direct photodegradation of acetaldehyde to formaldehyde or formic acid with CO2 evolution, and that the rate became saturated because the amount of acetaldehyde decreased and only these persistent intermediates remained. Next, we compared the activities of CuO/WO3 for the photodegradation of these stable intermediates, acetic acid, formaldehyde, and formic acid, with those of WO3 and CuO (Figure 3, panels i through iii). The activity of bare WO3 for photodegradation of these intermediates was very low even at the initial period. Therefore, we clearly concluded that the presence of these intermediates on bare WO3 was the reason why acetaldehyde was not completely oxidized to CO2 on bare WO3. In contrast, CuO/WO3 showed much higher activity for the degradation of these intermediates than did WO3 or CuO alone, and CuO/WO3 oxidized them to CO2 completely, which explains the complete oxidation of acetaldehyde over CuO/WO3. CuO is a p-type semiconductor,31 and there is a possibility that it may have acted solely as the photocatalyst; however, CO2 formation was not observed, whereas the acetaldehyde in the gas phase decreased gradually over CuO alone (150 mg). We confirmed that acetaldehyde was converted to acetic acid, which was detected in the rinsing water of CuO after the photoreaction. Formation of acetic acid over CuO was also observed even under dark conditions. CuO/WO3 afforded complete oxidation of acetaldehyde (ca. 3.2 µmol as CO2), whereas WO3 and CuO did not. This result suggests that the combination of WO3 and CuO was important for photocatalytic activity with respect to complete oxidation of acetaldehyde into

WO3-Catalyzed Photodegradation of Organic Substances

J. Phys. Chem. C, Vol. 113, No. 16, 2009 6605

Figure 4. Ultraviolet-visible spectra of CuO(2 wt %)/WO3, WO3, CuO, and N-TiO2. The absorbance spectrum of the fluorescent lamp is also shown (gray line).

Figure 3. Time courses of CO2 formation during photocatalytic degradation of (i) acetic acid (ca. 1 µmol), (ii) formaldehyde (ca. 2 µmol), and (iii) formic acid (ca. 2 µmol) over CuO/WO3, WO3, or CuO under visible light irradiation (λ > 400 nm).

CO2. The threshold wavelength of WO3 was ca. 460 nm (Figure 4), and that of CuO was over 900 nm. The spectrum of CuO(2 wt %)/WO3 was almost similar to that of WO3 alone, but weak and wide absorption was present in the whole visible light wavelength region owing to the small amount of CuO present. When CuO/WO3 was irradiated with filtered light (λ > 500 nm), under which WO3 could not be excited, CO2 generation was negligible; therefore, there was no photosensitization effect by CuO. In Figure 2i, the initial rate of CO2 generation over CuO/ WO3 was slow compared with that over WO3. In the highperformance liquid chromatography of the rinsing water of the CuO/WO3 powder during the photoreaction before the complete oxidation, it was confirmed that acetic acid was detected as an

intermediate. This result suggests that acetic acid formation from acetaldehyde, which did not generate CO2, occurred mainly during the initial photodegradation of acetaldehyde over CuO/ WO3, as was observed over CuO. CO2 was generated in the photodegradation of acetic acid over CuO/WO3, and this generation proceeded concurrently with the formation of acetic acid from acetaldehyde. The rate of CO2 generation in the photodegradation of acetic acid to CO2 on CuO/WO3 was considered to be slow compared with the rate of direct CO2 generation in the photodegradation of acetaldehyde to formaldehyde, formic acid, or both on bare WO3. The CO2 generation process through acetic acid is considered to be the reason for the slow initial CO2 generation rate on CuO/WO3 compared with bare WO3 in the photodegradation of acetaldehyde. However, we note that CuO/WO3 was advantageous in that it afforded complete oxidation of organic substances to CO2, whereas bare WO3 did not. 3. Valence Change of Copper on the Surface of the CuO Cocatalyst. The turnover number of CuO was calculated for a long-term photoreaction with use of CuO(0.1 wt %)/WO3 prepared by the impregnation method. The total number of holes (or electrons) used for the photodegradation of acetaldehyde to CO2 toward the number of atoms in Cu content of CuO/WO3 was more than 30. Furthermore, CO2 generation was hardly observed during the photodegradation of acetaldehyde over CuO/WO3 in the absence of oxygen, suggesting clearly that the oxygen molecules in the gas phase were used for the reaction. In the case of the WO3-catalyzed photodegradation of organic substances using Cu(II) ions in wet condition, we observed the redox cycle of Cu(II)/Cu(I) ions by a colorimetric reaction using a 2,9-dimethyl-4,7-diphenyl-1,10-phenanthrolinedisulfonic acid disodium salt (bathocuproinedisulfonic acid disodium salt) and an electrochemical measurement, and confirmed that the Cu(II) ion was reduced to the Cu(I) ion by photoexcited electrons from WO3 and that the Cu(I) ion was oxidized to the Cu(II) ion in the presence of oxygen, as previously reported.20 Here, to clarify the reaction mechanism of the CuO cocatalyst loaded on WO3 photocatalyst, the change in bulk structure and the valence change of copper on the surface of CuO before and after reaction were evaluated by means of XRD and XPS, respectively. From the XRD measurements, the XRD pattern of monoclinic CuO was not changed after the photoreaction, suggesting that there was no change in the bulk structure of CuO. In XPS measurements, the Cu(II) was characterized by high-intensity shakeup satellites (ca. 943 and 963 eV) at higher binding energy than the main 2p3/2 (ca. 934 eV) and 2p1/2 (ca. 953 eV) peaks. These shakeup satellites are not observed in the spectra of Cu(0) and

6606 J. Phys. Chem. C, Vol. 113, No. 16, 2009

Arai et al.

Figure 5. XPS spectra of Cu on CuO/WO3 before and after photodegradation. WO3 mixed with CuO at 5 wt % was used as the sample to increase the intensity of the Cu-2p signal. Shake-up satellites of Cu(II) are marked with solid circles (b). The intensity was the integrated value from the first to the fourth cycle measurements with smoothing treatment. The sample was exposed to gaseous acetaldehyde and irradiated with a Xe lamp through the introduction chamber with a pyrex glass window under air (b) and N2 (c) atomosphere.

Figure 6. The measurement time dependence of XPS spectra of Cu on as-prepared CuO/WO3. WO3 mixed with CuO at 5 wt % was used as the sample to increase the intensity of the Cu-2p signal. Shake-up satellites of Cu(II) are marked with solid circles (b). Measurement condition: Each cycle was integrated in 20 times scan; one step ) 0.15 eV, 40 ms; without smoothing treatment.

Cu(I), and the presence of Cu(I) was distinguished by the positive shift of the binding energy (negative shift of kinetic energy) related to the Cu LMM Auger electron.27,32 In our XPS measurements of CuO on semiconductors, we found that the measurement time was an important factor for obtaining real information regarding the valence of the cocatalyst. Figure 5 shows the Cu-2p XPS spectra of CuO/WO3. The shape of the spectrum of CuO/SiO2 (Figure 5e) was standard as Cu(II). The peaks of 2p3/2 (934 eV) and 2p1/2 (953 eV), and their shakeup satellite peaks (b) at 943 and 963 eV, corresponding to Cu(II), were observed. The ratio of the peak height of the 2p3/2 to its satellite was ca. 3. On the other hand, in the case of the as-prepared CuO/WO3 (Figure 5a), the shoulder peak at 934 eV and its satellite at 943 eV corresponding to Cu(II) were observed, but a new peak (4) at 932 eV corresponding to reduced cupper, Cu(I) or Cu(0), was also observed. This reduced cupper peak was not observed on CuO/SiO2 (Figure 5e) or on CuO powder alone. It was found that, as shown in Figure 6, the peak intensity of the reduced Cu observed in as-prepared CuO/WO3 increased with long XPS measurement time. The same phenomenon was observed in the case of CuO on TiO2 as a semiconductor, though it was not observed in the case of CuO on SiO2 as an insulator. These results clearly suggest that WO3 was photoexcited by X-ray irradiation during the measurement and the photoexcited electrons in WO3 reduced the surface of the CuO cocatalyst. Therefore, a short measurement time should be used to estimate the real valence of CuO on semiconductors. However, the peak of reduced Cu was observed on the as-prepared CuO/WO3 even at the first cycle (Figure 6a). It is difficult to shorten the measurement time further, because the spectrum was very noisy. So, in the case of Figure 5, the integration of the measurement cycles was conducted from the first to the fourth cycle. The spectra of the as-prepared CuO/WO3 (Figure 5a) and the CuO/WO3 after the photodegradation of acetaldehyde in air

(Figure 5b) were similar. However, after photodegradation proceeded in a nitrogen atmosphere (Figure 5c), the shakeup satellite peaks and the shoulder peak at 934 eV of Cu(II) disappeared completely, suggesting conversion of the cocatalyst surface from Cu(II) to Cu(I) or Cu(0). Furthermore, the shakeup satellite peaks reappeared after aeration of the sample (Figure 5d), suggesting the reproduction of Cu(II) from the reduced Cu on the surface of the cocatalyst. The results in panels c and d of Figure 5 were repeated cyclically. From the Auger measurement of Cu KLL in CuO/WO3 after the photodegradation in N2, a positive shift of the Cu(I) peak at 571 eV, compared to the peak of CuO at 569 eV, was observed, suggesting that Cu(II) was reduced to Cu(I) by the light irradiation without air. In the case of Cu ions in wet condition, the redox cycle of Cu(II)/Cu(I) ions accelerated the oxygen reduction process, and Cu(0) was not observed at all.20 Cu(II) is reduced to Cu(I) with aldehydes in the Fehling’s test. From these results, we concluded that the surface of the CuO cocatalyst on WO3 was reduced from Cu(II) to Cu(I) mainly by photoexcited electrons from WO3 photocatalyst and that Cu(I) was easily oxidized to Cu(II) by oxygen in air. Cu(I) could not be oxidized in the nitrogen atmosphere; therefore, Cu(I) was observed in Figure 5c. 4. Oxygen-Reduction Property of the CuO Cocatalyst under Dark and Light Conditions. From the results of XPS measurement, we concluded that the surface of the CuO cocatalyst was reduced by the photoexcited electrons in WO3 and reoxidized by oxygen in air. In other words, the CuO cocatalyst promoted the oxygen reduction reaction on the WO3 photocatalyst. To confirm the oxygen-reduction property of the CuO cocatalyst, electrochemical measurements were also performed. The current-potential relationships on the CuO/WO3, WO3, and CuO electrodes under dark conditions with air or bubbling nitrogen were measured by a three-electrode system in a Pyrex

WO3-Catalyzed Photodegradation of Organic Substances

J. Phys. Chem. C, Vol. 113, No. 16, 2009 6607

Figure 7. Current-potential relationship observed at CuO/WO3, WO3, and CuO electrodes under dark conditions with air or nitrogen bubbling.

glass cell (Figure 7). Oxygen-reduction activities were evaluated by the difference in cathodic current under air or bubbling nitrogen conditions. The cathodic currents at the CuO on conductive glass electrode with (Figure 7b) and without (Figure 7a) oxygen were very low. Cathodic current at the WO3 electrode without oxygen (Figure 7c) was observed at less than ca. +0.5 V (vs NHE), which is almost the same as the conduction band potential of WO3.25 This current was probably due to the reduction of WO3 itself. The increase in cathodic current at the WO3 electrode caused by the introduction of air (Figure 7d) was also not so large, suggesting that oxygen reduction on the WO3 electrode was negligible. In contrast, the cathodic current in the presence of air on the CuO/WO3 electrode (Figure 7f) was obviously larger than those on the WO3 and CuO electrodes (Figure 7b,d). These results suggest a presence of drastic synergistic effect of CuO on WO3. Considering the various oxygen reduction potentials and the conduction band potential of WO3 (+0.5 V vs NHE),25 it is surmised that the electron on WO3 conduction band would not easily reduce oxygen via the single-electron reaction of oxygen (O2/O2- ) -0.56 V vs NHE; O2/HO2 ) -0.13 V vs NHE). On the other hand, in the case of the CuO/WO3 electrode (Figure 7f), the threshold of the cathodic current by the O2 reduction was ca. +0.64 V vs NHE. Therefore, we concluded that the multipleelectron reduction of oxygen with a more positive redox potential (O2/H2O2, ) +0.68 V vs NHE; O2/H2O ) +1.23 V vs NHE) might take place on the CuO cocatalyst on the WO3 electrode, similar to the two-electron reduction process over phenanthroline-copper(I)33 or the four-electron reduction process over a Cu electrode34 and copper complexes.35 From all results mentioned above, we concluded that the CuO cocatalyst on WO3 photocatalyst powder probably assisted in the reduction of oxygen significantly via a multiple-electron reaction in dark conditions. To investigate the effect of light irradiation on CuO during the photodegradation reaction, a chronoamperometry measurement was performed with a two-electrode system, using a twocompartment cell separated with a Nafion membrane (Figure 8i). A WO3 electrode in the working electrode compartment and a CuO electrode in the counter electrode compartment were connected to each other without any external bias, simulating the charge transfer between WO3 and CuO in the CuO/WO3 photocatalyst. Na2SO4 aqueous solution was used as an electrolyte and ethanol was added as an organic substance only in the working electrode compartment. The effect of light irradiation on CuO was evaluated by comparing the photocurrent

Figure 8. Schematic of chronoamperometry measurements performed in a two-electrode system using a two-compartment cell separated with a Nafion membrane (i) and difference in photocurrent observed at a photoexcited WO3 electrode connected with a CuO electrode under light/dark conditions during the photodegradation of ethanol (ii).

observed when the WO3 electrode was irradiated with light and the CuO electrode was irradiated with light or not irradiated (Figure 8ii). No current was observed under dark conditions for both electrodes (Figure 8ii, line a). Current caused by the electron transfer from the WO3 to the CuO electrode was observed when the WO3 electrode (Figure 8ii, line b) or both electrodes (Figure 8ii, line c) were irradiated with light. The current observed when both electrodes were irradiated was drastically higher than that observed when only the WO3 electrode was irradiated. This result suggested that the photoexcited electrons in the WO3 electrode transferred to the CuO electrode even in dark conditions; however, this electron transfer was significantly improved by inducing light irradiation at the CuO electrode. We concluded that the cocatalyst effect of CuO for oxygen reduction was improved by light irradiation. However, this effect is difficult to confirm in photocatalytic reactions using the CuO/WO3 powder system, because the light that excites WO3 also can excite CuO, and it is difficult to irradiate them separately. Therefore, we tried to confirm the effect of light irradiation on CuO by means of a layered model reaction. Three types of layered samples made from photocatalystpowderswereprepared(Figure9i):(a)[WO3],(b)[WO3]-[WO3]-[CuO/ WO3], and (c) [WO3]-[SiO2]-[CuO/SiO2]. CuO cocatalyst powder was in contact with WO3 in sample (b). In contrast, CuO in sample c was isolated from WO3 by SiO2. Photodegradation of formic acid, which is the most simple organic compound, was performed by using the three types of layered samples, and the amounts of CO2 generated in samples b and c after irradiation for 1 h were normalized to the amount observed in sample a (Figure 9ii). When ultraviolet (UV) light (λ ≈ 350 nm) was applied from the bottom of the samples, sample b showed higher activity than sample a, suggesting a presence of

6608 J. Phys. Chem. C, Vol. 113, No. 16, 2009

Figure 9. Schematic view of samples used in the layered reaction model (i) and relative activity evaluated by the amount of CO2 generated after irradiation for 1 h (ii). Values were normalized by the amount of CO2 observed in sample a. The thickness of the WO3 layer was ca. 0.5 mm.

long distance cocatalyst effect of CuO. We confirmed that UV light was totally absorbed by WO3 in the bottom layer and that CuO/WO3 in the top layer was not irradiated with UV light. Therefore, CuO acted as a cocatalyst in dark conditions. In addition, improvement of photocatalytic activity was not observed in sample c, in which CuO was isolated from WO3, suggesting that contact between WO3 and CuO is essential for enhanced photocatalytic activity. When both UV and visible light were applied from the bottom of the samples, the activity of sample b was also higher than that of sample a and the ratio of improvement was higher compared with that observed under only UV light irradiation. In this case, the CuO/WO3 in the top layer was irradiated with light that was not absorbed by the WO3 layer (ca. λ > 460 nm). These results suggest that the cocatalyst effect of CuO can be improved by light irradiation compared with under dark conditions, whereas CuO can act as a cocatalyst even in dark conditions. 5. Optimization of CuO/WO3 and Its Activity for Photodegradation of Organic Substances under FluorescentLamp Irradiation in a Flow-Type Reactor. To improve the activity of CuO/WO3, commercially available WO3 powders obtained from various companies (Kanto Chemical Co., Inc., Kojundo Chemical Laboratory Co., Ltd., Wako Pure Chemical Industries, Ltd.) and various homemade WO3 powders were mixed with CuO obtained from Wako Pure Chemical Industries, Ltd. (CuO(Wako); surface area ca. 1.4 m2 g-1) at 2 wt % CuO

Arai et al. fraction. Their activities were evaluated by the CO2 evolution rate for the photodegradation of acetaldehyde. WO3 powder (WO3(Wako); surface area ca. 5 m2 g-1) showed the highest activity among all WO3 powders; therefore, this powder was selected to be the standard WO3 photocatalyst for further experiments. Various methods of loading CuO, such as the impregnation method and the mixing of various CuO powders, were also investigated to improve the activity of CuO/WO3. The maximum activity among all CuO/WO3 catalysts was observed when CuO nanopowder obtained from Sigma-Aldrich Co. (CuO(Aldrich); surface area ca. 29 m2 g-1) was mixed with WO3 powder at 0.05 wt % CuO fraction. The optimized weight fraction of CuO(Aldrich) was small compared with that of CuO(Wako). The aggregated particle size of CuO(Aldrich) confirmed in scanning electron microscope observation was smaller than that of CuO(Wako). Therefore, we concluded that the amount of CuO particles contacting the WO3 was important and was optimized at 0.05 wt % CuO fraction when the particles were small. In contrast, most of the CuO in the large particles might not have been used for the reaction. This optimized CuO(Aldrich)/WO3(Wako) (referred to herein as op-CuO/WO3) was used to evaluate the apparent quantum efficiency (QEa) of the photodegradation of various organic substances and the activity under fluorescent-lamp irradiation simulating that used in indoor applications. The QEa of CO2 generation over opCuO/WO3 at λ ) 400 nm was evaluated for the photodegradation of acetaldehyde, formaldehyde, and acetic acid, and the calculated values of QEa were ca. 3.2%, 2.3%, and 6.3%, respectively. It is reported that the QEa of Pt (1 wt %) and Pd (0.1 wt %) deposited on WO3 in the photodegradation of acetic acid is ca. 10% and 40%, respectively.21,24 The QEa value of op-CuO/WO3 in the photodegradation of acetic acid seemed low compared with those of noble metal deposited on WO3. However, it is difficult to compare the QEa values simply, because the efficiency depends on the reaction conditions, such as the light intensity and the concentration of organic substances. Moreover, formaldehyde and formic acid can be generated as intermediates in the photodegradation of acetic acid, and the degradation of these intermediates to CO2 can proceed over Pt and Pd cocatalysts even under dark conditions, though this degradation cannot proceed over a CuO cocatalyst. Therefore, the QEa value in the presence of Pt and Pd cocatalyst might be estimated to be higher than that observed in the presence of a CuO cocatalyst. In addition, the QEa values of op-CuO/WO3 were evaluated under different condition where formaldehyde aqueous solution and acetic acid were introduced as liquid while gaseous acetaldehyde was used. The high concentration of acetic acid over the photocatalyst might be contributed to the high value of QEa for the acetic acid degradation. The activity of op-CuO/WO3 for the photodegradation of acetaldehyde was evaluated under fluorescent-lamp irradiation and was compared with that of TiO2 and N-TiO2 (Figure 10). The op-CuO/WO3 showed higher activity than TiO2 and N-TiO2, especially during the latter half of the reaction period. The difference between the activities of op-CuO/WO3 and TiO2based photocatalysts was caused by the difference in their activities with respect to intermediates such as acetic acid. Photodegradation of acetaldehyde was also performed in the flow-type reactor under fluorescent-lamp irradiation (Table 1). In the case of N-TiO2, the acetaldehyde removal ratio (ARR) was 76%, and the ratio of complete oxidation of acetaldehyde to CO2 (RCO) was only 13%; these values suggest that the acetaldehyde was not completely oxidized. In contrast, acetaldehyde was almost completely oxidized to CO2 over op-CuO/

WO3-Catalyzed Photodegradation of Organic Substances

J. Phys. Chem. C, Vol. 113, No. 16, 2009 6609 candidate for a photocatalyst capable of decomposing harmful organic substances and an alternative to Ti-based photocatalysts. Acknowledgment. . This study was supported by the Project to Create Photocatalyst Industry for Recycling-oriented Society from the New Energy and Industrial Technology Development Organization (NEDO) of Japan. References and Notes

Figure 10. Time courses of CO2 formation during photocatalytic degradation of acetaldehyde (ca. 450 ppm) under fluorescent-lamp irradiation (ca. 6000 lx).

TABLE 1: Removal of Acetaldehyde from Flowing Gas (5 ppm, 50 mL/min) over op-CuO/WO3 and N-TiO2 under Fluorescent-Lamp Irradiation (ca. 3000 lux)a photocatalyst

ARR [%]

RCO [%]

RCOA [mg m-2 h-1

op-CuO/WO3 N-TiO2

100 76

100 13

4.5 0.4

a ARR: acetaldehyde removal ratio. RCO: ratio of complete oxidation. RCOA: rate of complete oxidation of acetaldehyde.

WO3 (ARR ) ca. 100%; RCO ) ca. 100%), and the rate of complete oxidation of acetaldehyde (RCOA) was calculated to be ca. 4.5 mg m-2 h-1. Few reports have confirmed complete oxidation in a flow-type reactor; therefore, it is very meaningful that we observed stoichiometric CO2 generation in this complete oxidation. Conclusions WO3-catalyzed photodegradation of various organic substances was improved by the addition of a CuO cocatalyst. Complete oxidation of acetaldehyde was achieved over CuO/ WO3, whereas bare WO3 could not completely oxidize acetaldehyde. The reaction mechanism can be explained by the improvement of oxygen reduction reaction over the CuO cocatalyst in the WO3-catalyzed photodegradation of the organic substance. The photoexcited electrons in WO3 were transferred to CuO, and the surface of the CuO was reduced from Cu(II) to Cu(I). The reduced surface was reoxidized from Cu(I) to Cu(II) by oxygen in the air. These proposed reaction mechanisms were confirmed by means of XPS and electrochemical measurements. We concluded that the multiple-electron reduction of oxygen with a more positive redox potential than that observed for the single-electron reduction of oxygen might take place on the CuO cocatalyst. The cocatalyst effect of CuO was promoted by light irradiation. The activities of optimized CuO/WO3 for the photodegradation of acetaldehyde under fluorescent-lamp irradiation were higher than those of TiO2 and N-TiO2. The complete oxidation of acetaldehyde to CO2 over op-CuO/WO3 under fluorescentlamp irradiation was also observed in a flow-type reactor, whereas N-TiO2 could not achieve such oxidation. CuO/WO3 that can work under fluorescent-lamp irradiation simulating that used in indoor applications will be a great

(1) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269–271. (2) Irokawa, Y.; Morikawa, T.; Aoki, K.; Kosaka, S.; Ohwaki, T.; Taga, Y. Phys. Chem. Chem. Phys. 2006, 8, 1116–1121. (3) Umebayashi, T.; Yamaki, T.; Tanaka, S.; Asai, K. Chem. Lett. 2003, 32, 330–331. (4) Ohno, T.; Mitsui, T.; Matsumura, M. Chem. Lett. 2003, 32, 364– 365. (5) Irie, H.; Watanabe, Y.; Hashimoto, K. Chem. Lett. 2003, 32, 772– 773. (6) Darwent, J. R.; Mills, A. J. Chem. Soc., Faraday Trans. 2 1982, 78, 359–367. (7) Erbs, W.; Desilvestro, J.; Borgarello, E.; Gra¨tzel, M. J. Phys. Chem. 1984, 88, 4001–4006. (8) Bamwenda, G. R.; Sayama, K.; Arakawa, H. J. Photochem. Photobiol. A 1999, 122, 175–183. (9) Sayama, K.; Mukasa, K.; Abe, R.; Abe, Y.; Arakawa, H. J. Photochem. Photobiol. A 2002, 148, 71–77. (10) Scaife, D. E. Sol. Energy 1980, 25, 41–54. (11) Vinodgopal, K.; Bedja, I.; Hotchandani, S.; Kamat, P. V. Langmuir 1994, 10, 1767–1771. (12) Sclafani, S.; Palmisano, L.; Marcı´, G.; Venezia, A. M. Sol. Energy Mater. Sol. Cells 1998, 51, 203–219. (13) Di Paola, A.; Palmisano, L.; Augugliaro, V. Catal. Today 2000, 58, 141–149. (14) Matsuoka, H. Jpn. Kokai Tokkyo Koho 2001, P2001-38217a. (15) Kim, K. G.; Jeong, E. D.; Borse, P. H.; Jeon, S.; Yong, K.; Lee, J. S.; Li, W.; Oh, S. H. Appl. Phys. Lett. 2006, 89, 064103. (16) Arai, T.; Yanagida, M.; Konishi, Y.; Iwasaki, Y.; Sugihara, H.; Sayama, K. J. Phys. Chem. C 2007, 111, 7574–7577. (17) Arai, T.; Yanagida, M.; Konishi, Y.; Iwasaki, Y.; Sugihara, H.; Sayama, K. Catal. Commun. 2008, 9, 1254–1258. (18) Arai, T.; Yanagida, M.; Konishi, Y.; Sugihara, H.; Sayama, K. Electrochemistry 2008, 76, 128–131. (19) Chen, D.; Ye, J. AdV. Funct. Mater. 2008, 18, 1922–1928. (20) Arai, T.; Yanagida, M.; Konishi, Y.; Ikura, A.; Iwasaki, Y.; Sugihara, H.; Sayama, K. Appl. Catal., B 2008, 84, 42–47. (21) Abe, R.; Takami, H.; Murakami, N.; Ohtani, B. J. Am. Chem. Soc. 2008, 130, 7780–7781. (22) Irie, H.; Miura, S.; Kamiya, K.; Hashimoto, K. Chem. Phys. Lett. 2008, 457, 202–205. (23) Kim, Y.; Irie, H.; Hashimoto, K. Appl. Phys. Lett. 2008, 92, 182107. (24) Arai, T.; Horiguchi, M.; Yanagida, M.; Gunji, T.; Sugihara, H.; Sayama, K. Chem. Commun. 2008, 5565–5567. (25) Maruska, H. P.; Ghosh, A. K. Sol. Energy 1978, 20, 443–458. (26) Kavan, L.; Gra¨tzel, M.; Gilbert, S. E.; Klemenz, C.; Scheel, H. J. J. Am. Chem. Soc. 1996, 118, 6716–6723. (27) Morikawa, T.; Irokawa, Y.; Ohwaki, T. Appl. Catal., A 2006, 314, 123–127. (28) Morikawa, T.; Ohwaki, T.; Suzuki, K.; Moribe, S.; Tero-Kubota, S. Appl. Catal., B 2008, 83, 56–62. (29) Pletcher, D.; Sotiropoulos, S. J. Electroanal. Chem. 1993, 356, 109– 119. (30) Santato, C.; Odziemkowski, M.; Ulmann, M.; Augustynski, J. J. Am. Chem. Soc. 2001, 123, 10639–10649. (31) Hardee, K. L.; Bard, A. J. J. Electrochem. Soc. 1977, 124, 215– 224. (32) Poulston, S.; Parlett, P. M.; Stone, P.; Bowker, M. Surf. Interface Anal. 1996, 24, 811–820. (33) Goldstein, S.; Czapski, G.; Eldik, R.; Cohen, H.; Meyerstein, D. J. Phys. Chem. 1991, 95, 1282–1285. (34) King, F.; Litke, C. D.; Tang, Y. J. Electrocanal. Chem. 1995, 384, 105–113. (35) Shleev, S.; Tkac, J.; Christenson, A.; Ruzgas, T.; Yaropolov, A. I.; Whittaker, J. W.; Gorton, L. Biosens. Bioelectron. 2005, 20, 2517–2554.

JP8111342