Photocatalytic Activity of (Ga1-x Zn x)(N1-x O x) for Visible-Light

Feb 12, 2008 - The photocatalytic activity of (Ga1-xZnx)(N1-xOx), a solid solution of GaN and ZnO, for H2 and O2 evolution in the presence of methanol...
0 downloads 0 Views 144KB Size
J. Phys. Chem. C 2008, 112, 3447-3452

3447

Photocatalytic Activity of (Ga1-xZnx)(N1-xOx) for Visible-Light-Driven H2 and O2 Evolution in the Presence of Sacrificial Reagents Kazuhiko Maeda,†,§ Hiroshi Hashiguchi,† Hideaki Masuda,† Ryu Abe,‡ and Kazunari Domen*,† Department of Chemical System Engineering, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, and Catalysis Research Center, Hokkaido UniVersity, Sapporo 001-0021, Japan ReceiVed: NoVember 10, 2007; In Final Form: December 6, 2007

The photocatalytic activity of (Ga1-xZnx)(N1-xOx), a solid solution of GaN and ZnO, for H2 and O2 evolution in the presence of methanol and silver nitrate as sacrificial reagents under visible light (λ > 420 nm) is investigated in detail. (Ga1-xZnx)(N1-xOx) evolves H2 from an aqueous methanol solution when loaded with nanoparticulate Rh2-yCryO3 as a cocatalyst, and evolves O2 from an aqueous silver nitrate solution without a cocatalyst. Structural analyses indicate that the H2 evolution activity is strongly dependent on the crystallinity and composition of the catalyst, while the rate of O2 evolution is proportionally related to the specific surface area. The activity for H2 evolution from methanol solution is of the same order as for overall water splitting, but is an order of magnitude lower than that for O2 evolution from silver nitrate solution. The results of photocatalytic reactions and photoelectrochemical measurements suggest that the rate-determining step for overall water splitting using (Ga1-xZnx)(N1-xOx) is the H2 evolution process.

1. Introduction Overall water splitting using a heterogeneous photocatalyst has been studied extensively as a potential method to supply hydrogen from sunlight and water.1 As it is generally difficult to achieve overall water splitting by one-step photoexcitation under visible light due to the uphill nature of the reaction, sacrificial reagents such as methanol or silver nitrate, which act as a hole or electron scavenger, have been used to evaluate the photocatalytic activity of a material.1a,d When a given material exhibits activities independently for H2 and O2 evolution (half reactions) in the presence of a suitable sacrificial reagent, the material can be said to meet the thermodynamic requirements for photocatalytic overall water splitting; the top of the valence band must be located at a more positive level than the water oxidation potential, and the bottom of the conduction band must be located at a more negative level than the water reduction potential. The band structure of such a photocatalyst and the redox potentials for relevant reactions are shown in Scheme 1. Our group has reported that (oxy)nitrides containing d0 transition metal cations or d10 typical metal cations are photocatalytically active for H2 and O2 evolution in the presence of sacrificial reagents (i.e., half reactions) under visible light.1a High quantum efficiencies of over 30% have been achieved for O2 evolution from an aqueous silver nitrate (AgNO3) solution.2,3 However, while most of the (oxy)nitride materials examined to date exhibit visible-light photocatalytic activity for such half reactions in the presence of a sacrificial reagent,1a there are only two (oxy)nitrides that are active for the full overall water splitting reaction under visible light: (Ga1-xZnx)(N1-xOx)3,4 and (Zn1+xGe)(N2Ox).5 As half reactions are carried out as test reactions for overall water splitting, comparison of the activities * Corresponding author. Tel.: +81-3-5841-1148. Fax: +81-3-58418838. E-mail: [email protected]. † The University of Tokyo. ‡ Hokkaido University. § Research fellow of the Japan Society of Promotion Science (JSPS).

SCHEME 1: Schematic Illustration of Band Structure of Photocatalyst and Redox Reactions at pH 0 vs NHE

for the half reactions with those for overall water splitting is of interest and is expected to provide useful information for the development of efficient catalysts. Such a comparison has been made for a number of ultraviolet-driven metal-oxide photocatalysts, and the relationship between the structural characteristics of the material and photocatalytic activity has been reported.6-8 However, no comparative study has yet been made for visiblelight-responsive (oxy)nitride photocatalysts. In a previous report, the photocatalytic activities of (Ga1-xZnx) (N1-xOx) for H2 or O2 evolution in the presence of sacrificial reagents under visible light were described briefly.4e In the present study, a systematic analysis of the dependence of the activities of (Ga1-xZnx)(N1-xOx) for H2 or O2 evolution in the presence of sacrificial reagents on structural properties is performed, and the results are compared to the data for overall water splitting. 2. Experimental Section 2.1. Preparation of Catalysts. The (Ga1-xZnx)(N1-xOx) solid solution was prepared according to the method described

10.1021/jp710758q CCC: $40.75 © 2008 American Chemical Society Published on Web 02/12/2008

3448 J. Phys. Chem. C, Vol. 112, No. 9, 2008

Maeda et al.

TABLE 1: Preparation Conditions for (Ga1-xZnx)(N1-xOx) Samples

TABLE 2: Specific Surface Area, Atomic Composition, and Band Gap Energy of (Ga1-xZnx)(N1-xOx) Samples

nitridation condition sample

starting materiala

Zn/Ga molar ratio

(a) (b) (c) (d) (e) (f) (g)

β-Ga2O3 + ZnO (K) β-Ga2O3 + ZnO (K) β-Ga2O3 + ZnO (K) β-Ga2O3 + ZnO (K) β-Ga2O3 + ZnO (K) β-Ga2O3 + ZnO (K) β-Ga2O3 + ZnO (W)

0.5 1.5 2.5 2.5 2.5 1 1

temp/K

time/h

1123 1123 1123 1123 1123 1123 1223

10 10 10 20 30 15 0.25

a

ZnO (K) and ZnO (W) indicate ZnO purchased from Kanto Chemicals and Wako Pure Chemicals, respectively.

previously4a,b from a physical mixture of β-Ga2O3 (99.9%, High Purity Chemicals, Japan) and ZnO (99%, Kanto Chemicals, Japan; 97%, Wako Pure Chemicals, Japan). The starting mixture was heated to the specified temperature (1123-1223 K) at a rate of 10 K min-1, and then maintained at that temperature for 0.25-30 h under NH3 flow (250 mL min-1). The molar ratio of zinc to gallium in the starting material was varied among samples from 0.5 to 2.5. After nitridation, the sample was cooled to room temperature under NH3 flow. Detailed preparation conditions are summarized in Table 1. The as-prepared (Ga1-xZnx)(N1-xOx) catalyst was loaded with nanoparticulate Rh-Cr mixed oxide (Rh2-yCryO3) as a cocatalyst for H2 evolution according to the method described previously.3,4d-f Briefly, 0.3-0.4 g of (Ga1-xZnx)(N1-xOx) powder and 3-4 mL of distilled water containing the appropriate amounts of Na3RhCl6‚2H2O and Cr(NO3)3‚9H2O were placed into an evaporating dish over a water bath. The suspension was stirred using a glass rod to complete evaporation, and the resulting powder was collected and heated in air at 623 K for 1 h to convert the rhodium and chromium species to Rh2-yCryO3.4f Rhodium and chromium were each loaded at a concentration of 2 wt % (metallic content). 2.2. Characterization of Catalysts. The prepared samples were studied by powder X-ray diffraction (XRD; RINTUltimaIII, Rigaku; Cu KR), scanning electron microscopy (SEM; S-4700, Hitachi), energy dispersive X-ray spectroscopy (EDX; Emax-7000, Horiba), and ultraviolet-visible diffuse reflectance spectroscopy (UV-vis DRS; V-560, Jasco). The Brunauer, Emmett, Teller (BET) surface area was measured using a BELSORP-mini instrument (BEL Japan) at liquid nitrogen temperature. 2.3. Photocatalytic Reactions. Reactions were carried out in a Pyrex top-irradiation reaction vessel connected to a glass closed gas circulation system. Photoreduction of H+ to H2 was performed by dispersing 0.1 g of the Rh2-yCryO3-loaded catalyst powder in an aqueous methanol solution (MeOH, 10 vol %, 100 mL) as the sacrificial reagent without pH control (pH 6-7), while photooxidation of H2O to O2 (O2 evolution) was conductedbydispersing0.1goftheas-prepared(Ga1-xZnx)(N1-xOx) powder in an aqueous silver nitrate solution (AgNO3, 0.01 M, 100 mL). The reactant solution was evacuated several times to remove air completely prior to irradiation under a 300 W xenon lamp fitted with a cutoff filter and a water filter. The temperature of the reactant solution was maintained at room temperature by a flow of cooling water during the reaction. Overall water splitting reactions were carried out in a similar manner using 0.1 g of the Rh2-yCryO3-loaded catalyst in an aqueous sulfuric acid solution (H2SO4, 100 mL) adjusted to pH 4.5. The peak catalytic performance of this material for overall water splitting

sample

specific surface area/m2 g-1

Zn/Ga atomic ratioa

band gap energyb/eV

(a) (b) (c) (d) (e)

5.4 8.3 8.5 3.4 2.7

0.28 0.25 0.17 0.09 0.06

2.63 2.66 2.68 2.71 2.77

a Estimated from EDX measurements. b Estimated from the onset of DRS.

has been shown previously to occur at pH 4.5.4e The evolved gases were analyzed by gas chromatography. 2.4. Photoelectrochemical Measurements. A porous (Ga1-xZnx)(N1-xOx) electrode was also prepared by pasting a viscous slurry onto conducting glass according to a method described previously.9 A mixture of 0.1 g of the as-prepared (Ga1-xZnx)(N1-xOx) powder (sample (g)), 10 µL of acetylacetone (Kanto Chemicals), 10 µL of TritonX (Aldrich, USA), and 200 µL of distilled water was ground in an agate mortar for preparation of the viscous slurry. The slurry was then pasted on fluorine-doped tin-oxide (FTO) glass slides (12 Ω sq-1, transparency 80%, thickness 1 mm; Asahi Glass, Japan) to prepare a 1 × 4 cm2 electrode, and the sample was calcined in air at 623 K for 1 h. Measurements were performed using a conventional electrochemical cell made of Pyrex glass with a platinum wire as a counter electrode and an Ag/AgCl reference electrode under potentiostat control (HZ-5000, Hokuto Denko, Japan). Currentvoltage curves were measured in an aqueous sodium sulfate solution (Na2SO4, 0.1 M, 100 mL) as a supporting electrolyte with and without an additive (10 vol % methanol or 0.01 M NaI). The electrolyte solution was purged with argon prior to measurements and was maintained at room temperature by a flow of cooling water during measurements. A 300 W xenon lamp fitted with a cutoff filter was used as a light source for visible irradiation (λ > 420 nm). 3. Results and Discussion 3.1. Physicochemical Properties of (Ga1-xZnx)(N1-xOx). The effects of preparation parameters on the physicochemical properties of (Ga1-xZnx)(N1-xOx) have been discussed in our previous papers4b,c and are reported briefly in the present study as Supporting Information. The specific surface area, atomic composition, and band gap energy of the as-prepared samples are listed in Table 2. XRD measurements (Figure S1) confirm that all samples exhibit a single hexagonal wurtzite phase similar to that of GaN and ZnO and that the (101) diffraction peak shifts successively to higher angles with increasing nitridation time and increasing Zn/Ga molar ratio of the starting materials. The samples are therefore not simple physical mixtures of GaN and ZnO but solid solutions of the two components. The peak shift also shows that the Zn/Ga atomic ratio in the final samples decreases with increasing nitridation time and increasing Zn/ Ga molar ratio of the starting materials, attributable to the effect of the larger ionic radius of Zn2+ (0.74 Å) as compared to Ga3+ (0.61 Å).10 It has been also confirmed by previous elemental analyses that the ratios of Ga to N and Zn to O are close to 1 and that the nitrogen and oxygen concentrations increase with the gallium and zinc concentrations.4a The degree of crystallization, indicated by the diffraction peak intensity, increases with starting zinc content and nitridation time, as can be confirmed from SEM observations (Figure S2). The band gap energy increases with decreasing Zn/Ga atomic ratio in the final

Photocatalytic Activity of (Ga1-xZnx)(N1-xOx)

Figure 1. Photocatalytic activities of samples (a)-(e) for overall water splitting under visible light (λ > 420 nm) from water adjusted to pH 4.5 with H2SO4 (100 mL). Reaction conditions: catalyst, 0.1 g; light source, xenon lamp (300 W) with cutoff filter; reaction vessel, Pyrex top-irradiation type.

Figure 2. Time courses of (A) H2 evolution from an aqueous methanol solution (10 vol %, 100 mL) and (B) O2 evolution from an aqueous silver nitrate solution (0.01 M, 100 mL) on sample (b) under visible light (λ > 420 nm). Reaction conditions: catalyst, 0.1 g; light source, xenon lamp (300 W) with cutoff filter; reaction vessel, Pyrex topirradiation type.

product, as discussed previously,4b,c causing the absorption edge to shift to longer wavelengths (Figure S3). The crystallinity thus improves from sample (a) to (e), that is, with increasing nitridation time and Zn/Ga ratio of the starting material, while the Zn/Ga atomic ratio in the final product decreases. The band gap energy also increases in the order (a)-(e). The observed relationship between preparation parameters on the physicochemical properties of (Ga1-xZnx)(N1-xOx) in the present study is consistent with the results of the previous studies.4b,c 3.2. Photocatalytic Activities for Overall Water Splitting. Figure 1 shows the steady rates of H2 and O2 evolution in overall water splitting under visible light (λ > 420 nm) using samples (a)-(e) (see Table 2) loaded with the Rh2-yCryO3 cocatalyst. All samples except for sample (e) produced H2 and O2 steadily and stoichiometrically upon irradiation with visible light. The activity increases markedly from sample (a) to (c) (i.e., with decreasing Zn/Ga ratio), yet decreases sharply from sample (c) to (e) (i.e., with increasing nitridation time). 3.3. H2 Evolution from an Aqueous Methanol Solution. As reported previously, (Ga1-xZnx)(N1-xOx) exhibits little activity for the photoreduction of water even in the presence of methanol as a sacrificial electron donor.4e However, the catalyst becomes active for the reaction when loaded with nanoparticulate Rh2-yCryO3 as a cocatalyst. These results indicate that bare (Ga1-xZnx)(N1-xOx) does not host active sites for H2 evolution and that all H2 evolution takes place in association with Rh2-yCryO3.4e Figure 2A shows the time course of H2 evolution from an aqueous methanol solution under visible light (λ > 420 nm) using sample (b) after loading with Rh2-yCryO3. H2 evolution occurs steadily without noticeable degradation. No O2 evolution occurs during the reaction due to the irreversible consumption of methanol by photogenerated holes in the valence

J. Phys. Chem. C, Vol. 112, No. 9, 2008 3449

Figure 3. Photocatalytic activities of samples (a)-(e) for (A) H2 evolution from an aqueous methanol solution (10 vol %, 100 mL) and (B) O2 evolution from an aqueous silver nitrate solution (0.01 M, 100 mL) under visible light (λ > 420 nm). Reaction conditions: catalyst, 0.1 g; light source, xenon lamp (300 W) with cutoff filter; reaction vessel, Pyrex top-irradiation type.

band of the photocatalyst.11 H2 evolution using methanol over other oxynitride photocatalysts such as TaON is typically accompanied by the evolution of a small amount of N2,2 attributed to the oxidation of N3- species near the catalyst surface by valence-band holes.1a Interestingly, no such N2 evolution occurs over Rh2-yCryO3-loaded (Ga1-xZnx)(N1-xOx) catalysts under the present reaction conditions, demonstrating the enhanced stability of the catalyst material. The corresponding photocatalytic activities of the present samples for H2 evolution are shown in Figure 3A. Although overall water splitting (stoichiometric H2 and O2 evolution) was not achieved using sample (e) (Figure 1), all samples (including (e)) produced H2 steadily from an aqueous methanol solution. The rate of H2 evolution increased from sample (a) to (c) and declined from sample (c) to (e), consistent with the activities observed for overall water splitting, although the H2 evolution activities are slightly higher than those achieved in overall water splitting. 3.4. O2 Evolution from an Aqueous Silver Nitrate Solution. The time course of O2 evolution using sample (b) under visible light (λ > 420 nm) is shown in Figure 2B. In contrast to the stable gas evolution observed for overall water splitting and H2 evolution using methanol, the activity for O2 evolution using silver nitrate decreased gradually with reaction time. Such a decrease in O2 evolution over time is primarily attributable to the deposition of metallic silver on the catalyst surface, which blocks light absorption and obstructs active sites. The pH of the reactant solution also decreases with progress of the reaction due to the production of nitric acid according to the following reaction:

AgNO3 + e- f Ag + NO3- (conduction band)

(1)

2H2O + 4h+ f O2 + 4H+ (valence band)

(2)

The rate of O2 evolution by TiO2 from an aqueous silver nitrate solution has previously been reported to decrease due to decrease in pH of the reactant solution.12 As the pH decreases, surface hydroxyl groups on TiO2 undergo protonation, causing the surface to become positively charged and thus inhibiting the adsorption of silver cations on the surface. As O2 evolution from an aqueous silver nitrate solution is initiated by electron injection from the conduction band of the photocatalyst to the silver cations adsorbed on the photocatalyst surface, the inhibition of Ag+ adsorption is also considered to contribute to the progressive decrease in O2 evolution activity. Previous studies using X-ray photoelectron spectroscopy have indicated that the

3450 J. Phys. Chem. C, Vol. 112, No. 9, 2008 (Ga1-xZnx)(N1-xOx) surface possesses hydroxyl groups,4b similar to the case expected for the TiO2 system. Therefore, the decrease in O2 evolution with reaction time is considered to be due to both the deposition of metallic silver and the decrease in pH of the reactant solution. Figure 3B shows the corresponding rates of O2 evolution using silver nitrate under visible light (λ > 420 nm). The rates were determined in the early stage of the reaction. In all cases, appreciable O2 evolution occurs even without the cocatalyst, and the activities are an order higher than those achieved in overall water splitting. In contrast to the case for H2 evolution using methanol and overall water splitting, sample (b) exhibits the highest activity for O2 evolution. A small amount of N2 was detected during O2 evolution, indicating that a proportion of the valence-band holes reacted with N3- species near the catalyst surface without promoting water oxidation. However, the N2 evolution did not continue through the reaction period. This difference in N2 evolution behavior between the H2 and O2 evolution reactions can be explained in terms of the number of photogenerated carriers available for reaction. In the O2 evolution reaction, silver cations adsorbed on the (Ga1-xZnx)(N1-xOx) surface are irreversibly reduced by photogenerated electrons in the conduction band, elevating the concentration of valence-band holes near the surface. As the concentration of photogenerated holes increases, so does the probability of reactions between such holes and N3species. 3.5. Dependence of Photocatalytic Activity on Physicochemical Properties. The crystallinity, specific surface area, atomic composition, and band gap energy appear to be the main factors affecting the photocatalytic activity of (Ga1-xZnx)(N1-xOx). It has previously been reported that the photocatalytic activity of (Ga1-xZnx)(N1-xOx) for overall water splitting is dependent on the crystallinity and Zn/Ga atomic ratio of the final material, but largely independent of the specific surface area.4b,c The present results are in good agreement with the previous reports in that the increase in activity from sample (a) to (c) is attributable to an increase in crystallinity (as identified by XRD and SEM, see Figures S1 and S2), while the decrease in activity from sample (c) to (e) is associated with a reduction in Zn/Ga atomic ratio (Table 2). An increase in crystallinity reduces the density of defects that can act as recombination centers between photogenerated electrons and holes, thereby contributing to an enhancement in activity. A decrease in Zn/Ga atomic ratio (in other words, O/N ratio), on the other hand, which occurs at the surface of the catalyst rather than in the bulk, produces defect sites (zinc and/or oxygen defects) that act as recombination centers between photogenerated electrons and holes, thus reducing activity.4b,c In H2 evolution from an aqueous methanol solution, the activities of the present samples vary in essentially the same order as in overall water splitting, although H2 evolution activities using methanol and water are slightly different from each other. As reported previously, the rates of H2 and O2 evolution in overall water splitting by Rh2-yCryO3-loaded (Ga1-xZnx)(N1-xOx) at pH 4.5 are similar to those at pH 6.3 in the initial stage of the reaction, although pH of the reactant solution greatly affects the long-term stability of the catalyst.4e Thus, H2 evolution activities from an aqueous methanol solution and water (pH 4.5 adjusted by H2SO4) can be compared to each other. The H2 evolution activity improves with increasing crystallinity from sample (a) (3.2 µmol h-1) to (c) (35 µmol h-1), and declines with decreasing Zn/Ga atomic ratio from sample (c) (35 µmol h-1) to (e) (1.5 µmol h-1). The degree of

Maeda et al.

Figure 4. Rate of O2 evolution over samples (a)-(g) under visible light (λ > 420 nm) in an aqueous silver nitrate solution (0.01 M, 100 mL) as a function of specific surface area. Reaction conditions: catalyst, 0.1 g; light source, xenon lamp (300 W) with cutoff filter; reaction vessel, Pyrex top-irradiation type.

activity increase and decrease is smaller than that for overall water splitting, suggesting that low crystallinity and low Zn/ Ga atomic ratio have a less negative effect on H2 evolution activity using methanol than in overall water splitting activity. In O2 evolution using silver nitrate as an electron accepter, all catalysts except for sample (c) displayed activities an order higher than those achieved in overall water splitting. Sample (b) exhibited the highest activity for this reaction among the samples examined. Whereas the activities for overall water splitting and H2 evolution from an aqueous methanol solution are strongly dependent on the crystallinity and Zn/Ga atomic ratio in the final product, it appears that the O2 evolution activity is relatively robust with respect to such factors. For example, the activity for H2 evolution achieved using sample (c) is an order higher than that for sample (a), while the activity of sample (c) for O2 evolution is only 1.3 times higher than that for sample (a). The positive effect on activity by increasing crystallinity is thus less prominent for O2 evolution using silver nitrate. Furthermore, while the activity of sample (d) for overall water splitting is higher than that of sample (e), both samples exhibit almost the same activities for O2 evolution from an aqueous silver nitrate solution. Increases in the density of zinc and/or oxygen defects and the band gap energy (i.e., decrease in number of photons that can be absorbed by the material) therefore appear to have almost no detrimental effect on O2 evolution activity. Samples (a) and (d) display almost the same H2 evolution activities from water and an aqueous methanol solution, yet the activity of sample (a) for O2 evolution from an aqueous silver nitrate solution is 1.8 times higher than that of sample (d). Considering the observation that the O2 evolution activity is largely insensitive to crystallinity, the final Zn/Ga atomic ratio, and band gap energy, it appears that the difference in activity between samples (a) and (d) is attributable to the effect of specific surface area. Figure 4 shows a plot of the O2 evolution activities of (Ga1-xZnx)(N1-xOx) as a function of specific surface area of the catalyst. A nearly linear relationship between the rate of O2 evolution and the surface area is apparent, indicating that the activity of (Ga1-xZnx)(N1-xOx) for O2 evolution from an aqueous silver nitrate solution is predominantly dependent on the specific surface area of the material. Samples (f) and (g) also satisfy this relationship,13 with the O2 evolution activity being nearly proportional to the specific surface area regardless of the starting composition. A higher specific surface area results in a higher density of active catalytic sites for gas evolution, thereby contributing to an increase in activity. Thus, as far as has been investigated in the present study, it can be concluded that the photocatalytic activity of (Ga1-xZnx)(N1-xOx) for O2 evolution from an aqueous silver nitrate solution is dependent

Photocatalytic Activity of (Ga1-xZnx)(N1-xOx)

J. Phys. Chem. C, Vol. 112, No. 9, 2008 3451 TABLE 3: Rate of H2 Evolution and Photocurrent under Visible Light (λ > 420 nm) in Various Solutions entry

electron donor

steady rate of H2 evolutiona/µmol h-1

photocurrentb/µA

1 2 3

H2O MeOH (10 vol %) NaI (0.01 M)

3.2 5.5 5.5

0.5 0.5 28.4

a Catalyst, 0.1 g (sample (f)); aqueous methanol solution, 10 vol %, 100 mL; light source, xenon lamp (300 W) with cutoff filter; reaction vessel, Pyrex top-irradiation type. b Applied potential, +0.4 V vs Ag/ AgCl; working electrode, sample (g); aqueous electrolyte solution, 100 mL (pH 4.5); light source, xenon lamp (300 W) with cutoff filter.

Figure 5. Current-voltage curves for porous (Ga1-xZnx)(N1-xOx) (sample (g)) electrodes under intermittent visible irradiation (λ > 420 nm) in (a) 0.1 M Na2SO4 (pH 4.5), (b) 0.1 M Na2SO4 (pH 4.5) with methanol (10 vol %), and (c) 0.1 M Na2SO4 (pH 4.5) with 0.01 M NaI. Scan rate: 10 mV s-1.

on the specific surface area of the material, and largely independent of crystallinity and composition. 3.6. Reason for Low H2 Evolution Activity. A noticeable feature in a series of the present photocatalytic reactions using (Ga1-xZnx)(N1-xOx) is the relatively low activity for H2 evolution from an aqueous methanol solution, as compared to that in overall water splitting. This result is in contrast to the performance of metal-oxide photocatalysts, which typically realize substantially higher H2 evolution activities from an aqueous methanol solution than from pure water.14,15 For example, the rate of H2 evolution from an aqueous methanol solution using Ni-loaded K4Nb6O17 is approximately 17 times higher than that from pure water.15 As photocatalytic H2 evolution using methanol is initiated by the oxidation of methanol by photogenerated holes in the valence band of the photocatalyst, the H2 evolution rate is expected to be limited by the rate of methanol oxidation. To examine whether methanol oxidation is the rate-limiting process, photoelectrochemical measurements were carried out using a porous (Ga1-xZnx)(N1-xOx) film prepared as a photoanode. Figure 5 shows current-voltage curves for the porous (Ga1-xZnx)(N1-xOx) electrode under intermittent irradiation with visible light (λ > 420 nm) in (a) 0.1 M Na2SO4 (pH 4.5), (b) 0.1 M Na2SO4 (pH 4.5) with methanol (10 vol %), and (c) 0.1 M Na2SO4 (pH 4.5) with 0.01 M NaI. The (Ga1-xZnx)(N1-xOx) electrode generated an anodic photocurrent associated with water oxidation upon irradiation with visible light, although the absolute value of the photocurrent was at most 0.5 µA (curve a, Figure 5, inset). This result indicates that (Ga1-xZnx)(N1-xOx) has n-type semiconducting property. The addition of methanol, which is thermodynamically more oxidizable than water, to the electrolyte solution does not result in any appreciable increase in photocurrent, indicating that the activity for methanol oxidation on the surface of the (Ga1-xZnx)(N1-xOx) electrode is of the same order as that for water oxidation. This behavior differs markedly from that observed for metal-oxide-based photoelectrochemical cells,16 but is consistent with reports for TaON photoelectrodes.17 The addition of NaI to the electrolyte, on the other hand, results in a clear enhancement of anodic photocurrent due to the oxidation of I- anions on the

(Ga1-xZnx)(N1-xOx) surface, demonstrating that the photooxidation of I- anions occurs more readily on (Ga1-xZnx)(N1-xOx) than the photooxidation of water or methanol. It is of interest that the order of increase in photocurrent among the various electrolyte solutions does not correspond to the order of increasing oxidation potential of the electrolytes. For example, methanol, which has the lowest oxidation potential, is oxidized less efficiently than any of the other electrolytes. Detailed analysisofthephotoelectrochemicalpropertiesof(Ga1-xZnx)(N1-xOx) and the improvement of photocurrent will be presented in future work. At this moment, one can expect that the rate of H2 evolution on Rh2-yCryO3-loaded (Ga1-xZnx)(N1-xOx) catalyst is much improved when I- anions are used as a sacrificial electron donor, because of the increased number of conduction band electrons resulting from the efficient consumption of I- anions by the valence band holes, as indicated by the results of photoelectrochemical measurements (Figure 5). Table 3 lists the photocatalytic activities for H2 evolution and the photocurrent at an applied bias of +0.4 V vs Ag/AgCl under visible light (λ > 420 nm) for various aqueous solutions. The activity for H2 evolution from an aqueous NaI solution (entry 3) is only approximately 2 times higher than that from water (entry 1) and is the same as that from an aqueous methanol solution (entry 2), whereas the photocurrent in the NaI-containing solution is about 60 times higher than that in the Na2SO4 solution with or without methanol. IO3- anions produced as a result of the oxidation of I- anions have been reported to prevent the reduction of H+ (i.e., H2 evolution) by reducing ahead of H+.18 This behavior is exemplified by the gradual decrease in the rate of H2 evolution on platinum-loaded anatase-type TiO218a and TaON powder18b from an aqueous NaI solution as the reaction progresses. In the present case, however, the rate of H2 evolution on Rh2-yCryO3-loaded (Ga1-xZnx)(N1-xOx) from an aqueous NaI solution did not decrease with reaction time (Supporting Information, Figure S4). Furthermore, preliminary experiments have indicated that the intentional addition of IO3- anions into the reactant solution does not result in a decrease in activity. IO3- anions in the reactant solution thus appear to have no negative impact on H2 evolution by Rh2-yCryO3-loaded (Ga1-xZnx)(N1-xOx), allowing IO3- anions to be ruled out as a cause of the inhibition of H+ reduction to H2. It can be concluded that the low activity of Rh2-yCryO3-loaded (Ga1-xZnx)(N1-xOx) for H2 evolution is attributable to inefficient electron transfer from (Ga1-xZnx)(N1-xOx) to the loaded Rh2-yCryO3 nanoparticles, and/or slow H+ reduction on the Rh2-yCryO3 surface. This behavior is considered one of the most serious drawbacks to overcome for the construction of an efficient overall water splitting system using the (Ga1-xZnx)(N1-xOx) photocatalyst. The relatively low activity for H2 evolution as compared to O2 evolution is a common characteristic of all of the oxynitride-type photocatalysts examined to

3452 J. Phys. Chem. C, Vol. 112, No. 9, 2008 date.1a It is therefore expected that improvements in H2 evolution activity will contribute substantially to the achievement of overall water splitting using (oxy)nitride-based photocatalysts. 4. Conclusion Visible-light-induced photocatalytic reactions using (Ga1-xZnx)(N1-xOx) in the presence of sacrificial reagents were examined. The activities for overall water splitting and H2 evolution from an aqueous methanol solution were found to be dependent on the crystallinity and composition of (Ga1-xZnx)(N1-xOx), whereas the activity for O2 evolution from an aqueous silver nitrate solution was found to be proportional to the specific surface areaofthematerial.Therelativelylowactivityof(Ga1-xZnx)(N1-xOx) for H2 evolution was attributed primarily to the inefficient water reduction process, which limits the efficiency of overall water splitting. Acknowledgment. This work was supported by the Solution Oriented Research for Science and Technology (SORST) program of the Japan Science and Technology Corporation (JST). Acknowledgment is also extended to the 21st Century Center of Excellence (COE) and the Research and Development in a New Interdisciplinary Field Based on Nanotechnology and Materials Science programs of the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan. K.M. gratefully acknowledges the support of a Japan Society for the Promotion of Science (JSPS) Fellowship. Supporting Information Available: Powder XRD patterns, SEM images, and UV-vis diffuse reflectance spectra for samples (a)-(e) (see Table 1), and a time course of H2 evolution from an aqueous NaI solution using sample (f) loaded with Rh2-yCryO3 under visible light (λ > 420 nm). This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) (a) Maeda, K.; Domen, K. J. Phys. Chem. C 2007, 111, 7851. (b) Hoertz, P. G.; Mallouk, T. E. Inorg. Chem. 2005, 44, 6828. (c) Lee, J. S. Catal. SurV. Asia 2005, 9, 217. (d) Kudo, A.; Kato, H.; Tsuji, I. Chem. Lett. 2004, 33, 1534.

Maeda et al. (2) Hitoki, G.; Takata, T.; Kondo, J. N.; Hara, M.; Kobayashi, H.; Domen, K. Chem. Commun. 2002, 1698. (3) Maeda, K.; Teramura, K.; Lu, D.; Takata, T.; Saito, N.; Inoue, Y.; Domen, K. Nature 2006, 440, 295. (4) (a) Maeda, K.; Takata, T.; Hara, M.; Saito, N.; Inoue, Y.; Kobayashi, H.; Domen, K. J. Am. Chem. Soc. 2005, 127, 8286. (b) Maeda, K.; Teramura, K.; Takata, T.; Hara, M.; Saito, N.; Toda, K.; Inoue, Y.; Kobayashi, H.; Domen, K. J. Phys. Chem. B 2005, 109, 20504. (c) Sun, X.; Maeda, K.; Le Faucheur, M.; Teramura, K.; Domen, K. Appl. Catal., A 2007, 327, 114. (d) Maeda, K.; Teramura, K.; Saito, N.; Inoue, Y.; Domen, K. J. Catal. 2006, 243, 303. (e) Maeda, K.; Teramura, K.; Masuda, H.; Takata, T.; Saito, N.; Inoue, Y.; Domen, K. J. Phys. Chem. B 2006, 110, 13107. (f) Maeda, K.; Teramura, K.; Lu, D.; Takata, T.; Saito, N.; Inoue, Y.; Domen, K. J. Phys. Chem. B 2006, 110, 13753. (g) Maeda, K.; Teramura, K.; Lu, D.; Saito, N.; Inoue, Y.; Domen, K. Angew. Chem., Int. Ed. 2006, 45, 7806. (h) Maeda, K.; Teramura, K.; Lu, D.; Saito, N.; Inoue, Y.; Domen, K. J. Phys. Chem. C 2007, 111, 7554. (5) (a) Lee, Y.; Terashima, H.; Shimodaira, Y.; Teramura, K.; Hara, M.; Kobayashi, H.; Domen, K.; Yashima, M. J. Phys. Chem. C 2007, 111, 1042. (b) Lee, Y.; Teramura, K.; Hara, M.; Domen, K. Chem. Mater. 2007, 19, 2120. (6) Kudo, A.; Tanaka, A.; Domen, K.; Onishi, T. J. Catal. 1988, 111, 296. (7) Machida, M.; Yabunaka, J.; Kijima, T.; Matsushita, S.; Arai, M. Int. J. Inorg. Mater. 2001, 3, 545. (8) Machida, M.; Murakami, S.; Kijima, T.; Matsushita, S.; Arai, M. J. Phys. Chem. B 2001, 105, 3289. (9) Abe, R.; Takata, T.; Sugihara, H.; Domen, K. Chem. Lett. 2005, 34, 1162. (10) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751. (11) Kudo, A. Catal. SurV. Asia 2003, 7, 31. (12) Ohtani, B.; Okugawa, Y.; Nishimoto, S.; Kagiya, T. J. Phys. Chem. 1987, 91, 3550. (13) Atomic compositions and band gap energies of sample (f), Zn/Ga ) 0.13, 2.68 eV; and sample (g), Zn/Ga ) 0.78, 2.5 eV. (14) Kato, H.; Asakura, K.; Kudo, A. J. Am. Chem. Soc. 2003, 125, 3082. (15) Kudo, A.; Tanaka, A.; Domen, K.; Maruya, K.; Aika, K.; Onishi, T. J. Catal. 1988, 111, 67. (16) (a) Nosaka, Y.; Sasaki, H.; Norimatsu, K.; Miyama, H. Chem. Phys. Lett. 1984, 105, 456. (b) Ohno, T.; Izumi, S.; Fujihara, K.; Masaki, Y.; Matsumura, M. J. Phys. Chem. B 2000, 104, 6801. (c) Sayama, K.; Nomura, A.; Arai, T.; Sugita, T.; Abe, R.; Yanagida, M.; Oi, T.; Iwasaki, Y.; Abe, Y.; Sugihara, H. J. Phys. Chem. B 2006, 110, 11352. (17) Nakamura, R.; Tanaka, T.; Nakato, Y. J. Phys. Chem. B 2005, 109, 8920. (18) (a) Abe, R.; Sayama, K.; Sugihara, H. J. Phys. Chem. B 2005, 109, 16052. (b) Abe, R.; Takata, T.; Sugihara, H.; Domen, K. Chem. Commun. 2005, 3829.