Photoanode characteristics of multi-layer composite BiVO4 thin film in a concentrated carbonate electrolyte solution for water splitting

Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

Contents lists available at SciVerse ScienceDirect

Journal of Photochemistry and Photobiology A: Chemistry journal homepage: www.elsevier.com/locate/jphotochem

Photoanode characteristics of multi-layer composite BiVO4 thin film in a concentrated carbonate electrolyte solution for water splitting Rie Saito, Yugo Miseki, Kazuhiro Sayama ∗ Energy Technology Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki, Japan

a r t i c l e

i n f o

Article history: Received 5 December 2012 Received in revised form 7 February 2013 Accepted 25 February 2013 Available online 14 March 2013 Keywords: Bismuth vanadate Tungsten oxide Tin oxide Photoelectrochemical cell Solar energy conversion

a b s t r a c t The improvement of the solar energy conversion efficiency is important for the solar hydrogen production using semiconductor photoelectrodes. In this paper, the photoelectrochemical properties of multi-layer composite photoelectrodes of thin film BiVO4 in various electrolyte solutions were investigated in detail. The improvement of photocurrent and the decrease of onset potential were observed on BiVO4 composite electrodes in carbonate electrolyte solution as well as bare BiVO4 electrode. The LHE (light-harvesting efficiency) and photocurrent were significantly improved by the light trapping structure of the double stacked photoelectrodes. The photocurrent was increased by insertion of an optimum SnO2 intermediate layer. The decrease of resistance at the BiVO4 composite electrodes was observed comparison with the bare BiVO4 . In the BiVO4 /SnO2 /WO3 photoelectrode, the highest IPCE (incident photon to current efficiency) was 53% at 420 nm. The H2 and O2 were evolved stoichiometrically. The maximum value of the applied bias photon-to-current efficiency (ABPE) was 1.35%. The reaction mechanism of carbonate anions, mainly affecting the BiVO4 layer, was discussed. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The solar splitting of water into H2 and O2 using a photoelectrochemical cell composed of an n-type semiconductor photoanode and H+ -reducing cathode has been widely investigated for application on solar energy conversion and storage [1–3]. Some nano-crystalline oxide semiconductor photoelectrodes with narrow band-gap energy (Eg ) [2,4,5], such as Fe2 O3 (Eg = 2.1 eV), WO3 (Eg = 2.7 eV) and BiVO4 (Eg = 2.4 eV) on a conductive glass substrate, are easily prepared by wet coating process and calcination under an air atmosphere. These photoelectrodes have significant advantages for the practical production of solar hydrogen, including simple preparation, H2 gas accumulation and large area production. These nano-crystalline photoelectrodes offer an improved photocurrent. Augstynski et al. reports the ca. 2.7 mA cm−2 (at 1.23 VRHE ) using the nano-crystalline WO3 photoelectrode [6]. Grätzel et al. reported that in the case of a modified Fe2 O3 film photoelectrode consisting of a perpendicularly oriented dendritic nanostructure [7–10], the best photocurrent reached ca. 3.2 mA cm−2 (at 1.23 VRHE ) [9]. Moreover, Pt-doped Fe2 O3 /Pt nanorod arrays on a gold substrate photoelectrode was developed by Park et al., and the photocurrent at 1.23 VRHE was very high (ca. 7.0 mA cm−2 ) [11]. On the other hand, the solar energy conversion efficiency (sun ) value is still low

∗ Corresponding author. Tel.: +81 29 861 4760; fax: +81 29 861 4760. E-mail address: [email protected] (K. Sayama). 1010-6030/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jphotochem.2013.02.019

and the improvement of sun poses the greatest challenge. Several equations have been used to calculate the value of sun [5,12,13], and a sun considering the energy loss resulting from the external bias in a two-electrode system, that is, applied bias photon-tocurrent efficiency (ABPE or ex sun ), is calculated by the following equation [5].



ABPE (%) =

Jopt × (1.23 − Eopt ) Int



× 100

(1)

where Jopt /mA cm−2 is the photocurrent density at Eopt ; Eopt /V is the applied voltage at the optimal operating conditions between the working and counter electrodes; Int/mW cm−2 is the intensity of incident solar light under A.M. 1.5, 1 Sun condition; and 1.23 V vs. reversible hydrogen electrode (RHE) is the standard electrode potential of H2 O. To improve ABPE, decreases of Eopt and onset potential, and an improvement of I–V curve shape are very important factors of the research as well as an increase in the photocurrent. The applied bias and the onset potential are influenced by the conduction band potential (ECB ) of semiconductors. In the case of the Fe2 O3 photoelectrodes, it was reported that large applied bias was needed to get high photocurrent due to the positive ECB . Recently, the BiVO4 photoelectrode has been attracting considerable attention. It is noteworthy that the ECB of BiVO4 (−0.4 V vs. NHE, pH = 7 [14]) is higher than that of WO3 , and is close to that of TiO2 . The photocurrent of BiVO4 was significantly increased by the addition of an under-layer coating of WO3 [14–16] or SnO2 [17,18]

52

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

on a conducting glass. However, the photocurrent of theses BiVO4 multi composite photoelectrodes is still low. Lee et al. reported that the photocurrent at 1.23 VRHE of heterojunction BiVO4 /WO3 electrodes reached ca. 1.4 mA cm−2 [14]. Moreover, various surface treated and modified BiVO4 photoelectrodes were developed [19–25] and Choi et al. reported that the photocurrent at 1.23 VRHE of FeOOH/BiVO4 photoelectrodes reached ca. 2.3 mA cm−2 [24]. We have investigated the improvement of the photoelectrochemical properties of bare BiVO4 photoelectrode in a carbonate electrolyte [26], as opposed to the sulfate electrolyte that has generally been used. Recently, we reported BiVO4 composite electrodes and the effect of introducing a SnO2 intermediate layer between the BiVO4 and WO3 layers (i.e., BiVO4 /SnO2 /WO3 photoelectrode) in a carbonate electrolyte [27]. The photocurrent and ABPE were found to increase by the insertion of a SnO2 intermediate layer and by carbonate ion effect. In this study of a multi-layer composite thin-film BiVO4 photoelectrode, we investigated the effects of the BiVO4 and SnO2 film thickness, photoelectrochemical properties on composite electrodes in various electrolyte solution, and light trapping structure on the photoelectrochemical and optical properties in detail. The photocurrent was prominently affected by the film thickness of the BiVO4 upper layer and the SnO2 intermediate layer. Regarding the effect of the light trapping structure, it was found that the incident photon-to-current conversion efficiency (IPCE) was improved in the visible light region. The reaction mechanism of effect of the carbonate anions was also discussed.

electrolyte [26]. The electrochemical impedance spectroscopy (EIS) used CIMPS-system 1 (ZAHNER/ZENNIUM, XPOT, Germany). The Mott–Schottky plots of the photoelectrode was analyzed by result of the impedance-potential. In the main case of the I–V measurement, the working electrode was irradiated through a black mask/aperture (0.31 cm−2 ) from the side of BiVO4 film (front side irradiation) using a solar simulator (JIS-A-class, SAN-EI ELECTRIC Co.) through a light chopper. Sometimes, photoanode properties were measured under the light irradiation from the side of a glass substrate (back side irradiation) for the comparative experiment. The light intensity of the solar simulator was calibrated to A.M. 1.5 (1 Sun, Supporting Information, Fig. S1) using a solar simulator spectroradiometer (SOMA Optics, Ltd.), immediately after calibration using a standard light source (certification body, Japan Electric Meters Inspection Corporation, JEMIC). The mismatch between the A.M. 1.5 (1 Sun) and the solar simulator spectra was precisely corrected by the spectral mismatch factor (MMF) method [29]. The IPCE was automatically measured by a quantum efficiency-IPCE system (EKO Instruments Co. Ltd.) with a Xe lamp (USHIO INC). The photon flux of the monochromatic light was measured by a Si photodiode detector (Hamamatsu Photonics K. K. and this Si photodiode detector was calibrated by National Metrology Institute of Japan, NMIJ). The IPCE was calculated from Eq. (2). IPCE =

1240 × photocurrent density (mA cm−2 ) wavelength (nm) × photon flux (mW cm−2 )

(2)

2. Experimental

The LHE (light-harvesting efficiency) of the electrode was calculated from transmittance (T) and reflectance (R) using an integrating sphere (Jasco, V-570, ISN-470) (Eq. (3)).

2.1. Photoelectrodes preparations

LHE = 1 − R − T

BiVO4 /WO3 and BiVO4 /SnO2 /WO3 photoelectrodes were prepared as follows. The precursor solutions of each oxide semiconductor were coated on a F-doped SnO2 (FTO) conductive glass substrate (surface resistance 10 /sq, Nippon Sheet Glass Co. Ltd.) using a spin coater (1000 rpm, 30 s) and then calcinated at 500 ◦ C for 30 min for each coating. Initially, a WO3 layer was coated on a FTO glass with subsequent multiple coatings of the SnO2 film (in the case of the inserted intermediate layer) and the BiVO4 upper layer. The precursor solution used the following. In the case of the WO3 , it was 1.4 M peroxotungstic acid, as reported previously [28]. In the case of the BiVO4 , bismuth oxide and vanadium oxide of EMOD (Enhanced Metal Organic Decomposition) materials provided by Symetrix Co., USA were mixed with Bi:V = 1:1 and then these were diluted with butyl acetate. The SnO2 precursor solution was EMOD materials of tin oxide provided by Symetrix Co. (USA) and diluted with xylene.

(3)

The amount of H2 and O2 evolved from the photoelectrochemical cell, which consisted of a photoanode and a Pt cathode, was investigated by a closed gas-circulating system with on-line gas chromatography (Shimadzu Co., GC-8A, TCD, 5A molecular sieves, Ar carrier). The characterization of samples was investigated by a scanning electron microscopy (SEM, Hitachi S-800, SE mode), transmission electron microscopy (TEM, Hitachi High Technologies HD-2700), X-ray fluorescence spectroscopy (XRF, Rigaku ZSXmini), and X-ray photoelectron spectroscopy (XPS, Ulvac-Phi XPS-1800). A stylustype step measuring instrument (Surfcorder ET-3000, Kosaka Laboratory Ltd.) was used to the evaluation of film thickness. The solar energy conversion efficiency has various derivations [5,12,13]. We used the ABPE by Eq. (1) [5]. The point, where size of photocurrent (at E) × potential (E) of obtained I–V curves became to the maximum, was chosen as the Eopt and Jopt . 3. Results and discussion

2.2. Photoelectrochemical measurements and characterizations The main photoelectrochemical properties were measured by an electrochemical analyzer (BAS. Inc. ALS660B). The I–V and the time dependence of the photocurrent measurements for the photoanode characteristics were assessed using a three-electrode cell with an Ag/AgCl reference electrode and a Pt coil counter electrode. The I–V curves of the forward and backward scans of potential should overlap in the steady state. The scan rate was slow (50 mV s−1 ). The gas measurement for water splitting and the I–V measurement for the ABPE calculation were performed using a two-electrode cell without a reference electrode. These electrodes were soaked in 80 mL of an electrolyte aqueous solution containing KHCO3 with CO2 gas bubbling in a Pyrex glass cell. Previously, we reported that the photocurrent of the BiVO4 photoelectrode was increased with bubbling CO2 gas into the carbonate

3.1. Morphology and optical properties of the BiVO4 /WO3 photoelectrode To investigate the effect of BiVO4 -upper layer film thickness on photoelectrochemical properties, we prepared multi-layer composite BiVO4 photoelectrodes with several different BiVO4 upper-layer film thicknesses maintaining the WO3 -under-layer film thickness (190 nm). The adjustment of the BiVO4 -upper-layer film thickness was carried out by changing the coating number of times of the BiVO4 precursor solution. The BiVO4 film thickness was ca. 60, 100, 150 and 200 nm. Fig. 1 shows the SEM photographs of the WO3 and the BiVO4 /WO3 photoelectrode surface. The nanocrystalline WO3 film surface was very smooth. With an increase in the film thickness of the BiVO4 , the BiVO4 particle grew and the interparticle crevice widths became narrower. The amounts of Bi

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

53

Fig. 1. SEM images of the (A) WO3 and the BiVO4 /WO3 surfaces on FTO conducting glass. The film thickness of BiVO4 (B) ca. 60 nm, (C) ca. 100 nm, and (D) ca. 150 nm.

and V of these electrodes were analyzed by XRF. These amounts increased keeping ca. Bi:V = 1:1 as the BiVO4 film thickness increase (Fig. S2). Fig. 2 shows the LHE spectra of the BiVO4 /WO3 film. The threshold of the LHE was approximately 510 nm, a value that was approximately consistent with Eg of BiVO4 (2.4 eV, ca. 516 nm). The increase of LHE was seen until the BiVO4 film thickness became ca. 150 nm, but when the film thickness was accompanied by an additional increase, the LHE reached saturation. It is considered that the LHE increase at over 510 nm is caused by the increase of absorption at the conducting glass by the light scattering through the semiconductor films. 3.2. I–V characteristics of the BiVO4 /WO3 photoanode The I–V characteristics were measured by a BiVO4 /WO3 photoelectrode as shown in Fig. 3(a). The current under dark was 10 ␮A cm−2 or less. There was an optimum condition on the BiVO4 thickness. We compared the photoanodes with various BiVO4 thickness in the front and back side irradiation, and it was found that the BiVO4 film with ca. 100 nm thickness and the front side irradiation were the optimum condition for the best photocurrent. As the BiVO4 film thickness increases and the surface morphology changes, the diffusion length distance of the excited electrons generated near the BiVO4 surface to FTO glass substrate increases in the case of the front side irradiation, so it is thought that the photocurrent was decreased. Using this BiVO4 /WO3 photoelectrode which had the highest photoelectrochemical properties, the IPCE in Fig. 3(b) was measured under +0.9, +1.2, and +1.8 V vs. RHE anodic bias. The IPCE value increased with anodic bias, and was ca. 36% (at 420 nm, +1.2 V). We calculated photon number of the irradiation light from a spectrum of the solar simulator and estimated the expected photocurrent in which this photon number was converted into the photoelectron with the IPCE of Fig. 3(b). In fact, the photocurrents obtained under the solar-simulated light corresponded to the photocurrent 1

LHE

0.8 0.6 0.4 0.2 0 300

400

500

600

Wavelength/nm Fig. 2. LHE spectra of BiVO4 /WO3 films. BiVO4 -upper-layer film thickness; — ca ca. 100 nm, · · · ca. 150 nm, and ca. 200 nm. 60 nm,

expected from the IPCE spectra within an approximate 10% error.

3.3. Effect of carbonate electrolyte solution on the BiVO4 /WO3 photoanode In our previous paper [26], we investigated the photocurrents in various electrolyte aqueous solutions only using bare BiVO4 electrode. Here, we compared the photoelectrochemical properties using BiVO4 /WO3 , bare BiVO4 and bare WO3 photoelectrodes. The photocurrent characteristics of the photoelectrode, pH and conductivity in various electrolyte aqueous solutions are shown in Table 1. The pH of those solutions was unified with the around neutral. Fig. 4 shows typical I–V curves of the effect of the electrolyte on three kinds of photoelectrodes ((a) BiVO4 /WO3 , (b) bare BiVO4 , and (c) bare WO3 ) in sulfate and carbonate aqueous solutions. The difference in the photocurrent in carbonate aqueous solution between K+ and Na+ cations was not large on the bi-layer composite BiVO4 /WO3 and bare WO3 photoelectrode as well as bare BiVO4 . The photocurrents of BiVO4 /WO3 were higher than those of bare WO3 and bare BiVO4 in all electrolyte solutions. In the BiVO4 /WO3 photoelectrode, the photocurrent in sulfate solution was worse than that in carbonate or phosphate solution. The pH change around the electrodes in non-buffered sulfate solution usually has a negative effect on the photoelectrochemical properties, but the difference between carbonate and phosphate solutions cannot be explained by the buffer effect. The conductivity depended on the concentration and anion of electrolyte solution, and a direct relationship between the conductivity and the photocurrent was not observed in bare WO3 and bare BiVO4 except the carbonate electrolyte. Solarska et al. investigated the photoanode reaction of nanocrystalline WO3 electrode for a water photoelectrolysis system under solar simulator illumination, and reported that the photocurrent depended on the concentration and conductivity of a special electrolyte aqueous solution (CH3 HSO3 ) [30]. In our case, the concentration/conductivity dependence in the bare WO3 photoanode reactions were not observed with sulfate and phosphate aqueous solutions under neutral pH. In the case of BiVO4 /WO3 , the photocurrents in 1.5 M solutions with higher conductivity were higher than those in 0.1 M solutions. The photocurrent of each photoelectrodes in various electrolyte solutions with the similar conductivity was compared, and the highest photocurrent was observed in highly concentrated carbonate solution. Additionally, in the case of the BiVO4 /WO3 and bare BiVO4 , the onset potentials in carbonate electrolyte solution were lower than those in other electrolyte solutions. Irrespective of the electrolyte used, the I–V characteristics were improved by the presence of a WO3 layer under the BiVO4 film as shown in Fig. 4. With respect to the bare BiVO4 and BiVO4 /WO3 photoelectrodes, the photocurrent in carbonate aqueous solution was higher than those in other

54

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

Fig. 3. (a) I–V characteristics of the BiVO4 /WO3 photoanode with the different BiVO4 film thickness. Symbol, —; ca. 60 nm, ; ca. 100 nm, ; ca. 150 nm. Electrolyte, KHCO3 (0.1 M) with CO2 bubbling. Light source, solar simulator (1 Sun, 100 mW cm−2 ). (b) IPCE spectra of the BiVO4 /WO3 photoanode with the ca. 100 nm ; +1.2 V vs. RHE, and ; +1.8 V vs. RHE. Electrolyte, KHCO3 (0.1 M) with CO2 bubbling. Light source; Xe lamp with a BiVO4 film thickness. Anodic bias, ; +0.9 V vs. RHE, monochromator.

Table 1 The photocurrent properties of various photoelectrodes in various electrolyte aqueous solutions. Photoelectrode

BiVO4 /WO3

Bare BiVO4

Bare WO3

Concentration/M

pH

Conductivity/mS cm−1

KHCO3 + CO2 NaHCO3 + CO2 Na2 SO4 KH2 PO4 + KOHa KHCO3 + CO2 Na2 SO4 KH2 PO4 + KOHa KHCO3 + CO2 NaHCO3 + CO2 Na2 SO4 KH2 PO4 + KOHa KHCO3 + CO2 Na2 SO4 KH2 PO4 + KOHa KHCO3 + CO2 NaHCO3 + CO2 Na2 SO4 KH2 PO4 + KOHa KHCO3 + CO2 Na2 SO4 KH2 PO4 + KOHa

0.1 0.1 0.1 0.1 1.5 S(ca. 1.5)b 1.5 0.1 0.1 0.1 0.1 1.5 S(ca. 1.5)b 1.5 0.1 0.1 0.1 0.1 1.5 S(ca. 1.5)b 1.5

6.9 6.8 6.2 6.4 7.9 6.1 6.3 6.9 6.8 6.2 6.4 7.9 6.1 6.3 6.9 6.8 6.2 6.4 7.9 6.1 6.3

7.7 6.5 14 12 78 72 75 7.7 6.5 14 12 78 72 75 7.7 6.5 14 12 78 72 75

Photocurrent/mA cm−2

Onset potential/VRHE

at l.23 VRHE

at 0.8 VRHE

1.80 1.84 1.17 1.59 2.08 1.27 1.82 0.64 0.58 0.17 0.14 1.01 0.09 0.11 0.32 0.33 0.33 0.35 0.41 0.30 0.35

0.98 1.11 0.08 0.32 1.58 0.11 0.24 0.07 0.13 0.03 0.03 0.18 0.01 0 0.09 0.11 0.04 0.10 0.14 0.05 0.09

0.43 0.43 0.74 0.45 0.33 0.70 0.57 0.45 0.50 0.70 0.82 0.53 0.90 0.94 0.62 0.62 0.78 0.66 0.66 0.78 0.74

pH was adjusted at around neutral. Saturated solution (S).

aqueous solutions. However, the positive effect of the photocurrent of the bare WO3 photoelectrode in the carbonate aqueous solution was marginal. Therefore, it is concluded in the case of the BiVO4 /WO3 composite photoelectrode, that the BiVO4 film part was predominantly affected by the carbonate solution rather than the WO3 part. In the pH around neutral, the zeta potential of WO3 particles [31] was negative than the zeta potential of BiVO4 particles [32]. It is thought that the adsorption behavior of carbonate anions on the WO3 surface might be different from that on the BiVO4 surface.

1.5

(a)

3

2 1 0 0.2

-1

0.4

0.6

0.8

1

1.2

Potential/V vs RHE

1.4

1.6

1.8

Current density/mAcm-2

Current density/mAcm-2

4

We previously reported that the TiO2 photoelectrode and various oxide photocatalysts including Pt–TiO2 were positively affected by HCO3 − and speculated that in the reaction mechanism, that is, the photooxidation of water was improved by the reaction of HCO3 − via an intermediate such as carbonate radicals or peroxocarbonates (HC2 O6 − , HCO4 − , etc.) [26,33]. It was reported that the peroxocarbonates were synthesized by the oxidation carbonate solutions with H2 O2 [34,35] and by the electrolytic oxidation of carbonate aqueous solution at low temperature [36], suggesting a possibility of oxidative production of carbonate anion to

0.6

(b)

1 0.5 0 0.2 -0.5

0.4

0.6

0.8

1

1.2

Potential/V vs RHE

1.4

1.6

1.8

Current density/mAcm-2

a b

Electrolyte

(c)

0.4 0.2

0 0.2

-0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Potential/V vs RHE

Fig. 4. I–V characteristics of the (a) BiVO4 /WO3 , (b) bare BiVO4 , and (c) bare WO3 photoanodes. Light source, solar simulator (1 Sun, 100 mW cm−2 ). Electrolyte aqueous ; 0.1 M KHCO3 , and ; 0.1 M Na2 SO4 . solution, —; 2.5 M KHCO3 ,

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

peroxocarbonates photoelectrochemically. The pH in our KHCO3 aqueous solution with CO2 bubbling was ca. 7–8, and most of the carbonate anions were present as HCO3 − at this pH, as calculated from the acid dissociation constant (pKa1 = 6.35, pKa2 = 10.3) in the carbonate salt aqueous solution. We speculated on the reaction mechanism of the HCO3 − effect as follows: Anodic reaction: 2HCO3 − + 4h+ → (intermediates) → 2CO2 + O2 + 2H+

(5)

CO2 gas solution in water: 2CO2 + 2H2 O ↔ 2HCO3 − + 2H+

(6)

Cathodic reaction: +

4H + 4e− → 2H2

(7)

Total: (5) + (6) + (7) 2H2 O + 4e− + 4h+ → 2H2 + O2

(8)

Recently, Dimitrijevic et al. reported that the generation of CO3 −• radical anions was observed on a TiO2 photocatalyst in a carbonate aqueous solution under UV illumination by electron paramagnetic resonance (EPR) spectroscopy and that the carbonate anions can act as a more efficient hole scavenger than water [37]. The O2 production process is very complicated, and the actual intermediate and these redox potentials are not clear now. However, if HCO3 − anions adsorbed on the BiVO4 surface also act as hole scavenger, the HCO3 − might be photooxidized to CO2 and O2 gases via some intermediates by multi-holes. One speculated reaction through peroxocarbonates intermediate adsorbed on semiconductor surface (HCO4 − (a)) is shown as an example below: 2HCO3 − + 2H2 O + 4h+ → (2HCO4 − (a) + 4H+ ) → 2CO2 + 2H2 O + O2 + 2H+

(9)

Various backward reactions easily occur on the way of O2 evolution from water on semiconductor surface without carbonate anions (Eq. (10)). O2 reduction by e− is also one of the backward reactions. H2 O  H2 O2 + 2H+  O2 + 4H+

(10)

On the other hand, in the latter of Eq. (9), non-invertible elementary processes with CO2 gas evolution exist at the decomposition of peroxocarbonates. Therefore, it is thought that photoelectrochemical properties were also improved by an enhancement of gases evolution. The generated CO2 gas is re-dissolved into the aqueous solution, and then becomes an HCO3 − anion (Eq. (6)), so that the CO2 /HCO3 − ion behaves like a catalyst for effective and irreversible O2 evolution on the BiVO4 film surface. 3.4. Effect of light trapping on the BiVO4 /WO3 photoelectrode When the BiVO4 film thickness was over 150 nm, the LHE of the BiVO4 /WO3 film was saturated, and the photocurrent of the BiVO4 /WO3 photoelectrode decreased. Therefore, we investigated the effects of light trapping to improve the LHE of the multi-layer composite BiVO4 thin film photoanode with an optimal BiVO4 film thickness. Fig. 5(a-1) and (b-1) shows the light trapping structures of single stack and double stacks in which a diffuser was positioned at the back of the glass substrate of the photoelectrode, respectively. The double stacks used the same two electrodes, which were connected in parallel. We called the positions of each photoelectrode “Front 1” or “Front 2” on the light source side, as shown in Fig. 5(b-1). Fig. 5(a-2) and (b-2) shows the I–V characteristics of the single and double stacked photoelectrodes, respectively. In the single stack

55

with an applied light trapping structure, the photocurrent at 1.23 V vs. RHE increased by 18% compared with the unmodified structure (Fig. 5(a-2)). In the double stacks with a light trapping structure, the LHE obtained was approximately 80% from 300 to 460 nm, and this structure was more effective in the long wavelength region in the LHE (Fig. 6(a)). We investigated IPCE and I–V properties of the photoelectrode at each position. The photoelectrode of “Front 2” absorbed the transmitted light though the photoelectrode of “Front 1”. In the photoelectrode of “Front 2,” the peak IPCE value was obtained at 460 nm (Fig. 6(b)) and the photocurrent was one-third of the photocurrent of the “Front 1” photoelectrode (Fig. 5(b2)). The photocurrent of each photoelectrode was summed and this value was almost coincident with the photocurrent of the double-stacked photoelectrode. It is suggested that the LHE and photoelectrochemical properties were easily improved by setting the double stacks of the photoelectrode with the light trapping structure using the diffuser on the back. 3.5. Effects of insertion of the SnO2 intermediate layer We previously reported that the photocurrent was increased by inserting a SnO2 film between the WO3 and the BiVO4 layers [27]. In this paper, we investigated the various photoelectrochemical properties of the multi-layer composite BiVO4 /SnO2 /WO3 photoelectrode on the preparation conditions of the SnO2 intermediate layer in greater detail. XPS analysis was carried out on three photoelectrodes (WO3 /FTO, SnO2 /WO3 /FTO and BiVO4 /SnO2 /WO3 /FTO). We confirmed that the SnO2 in the FTO glass substrate and the intermediate film was clearly distinguishable, because no SnO2 was detected from the WO3 /FTO (Table 2). Surface W atoms (47%) were detected on the SnO2 /WO3 /FTO photoelectrode; thus, considering the escape depth of XPS electrons, it would appear that most of the WO3 surface was deposited with a very thin SnO2 layer, or that WO3 was covered with SnO2 inhomogeneously. In the case of BiVO4 /SnO2 /WO3 /FTO, this multi-layer composite film had many crevices between BiVO4 particles (Fig. S3) similar to that observed for BiVO4 /WO3 /FTO in Fig. 1(b). Therefore, it was thought that the photoelectrons of WO3 and SnO2 were detected through the crevices from the interface between the BiVO4 and SnO2 /WO3 layers, and that the electrolyte solution and O2 gas can be diffused through the crevices. The ellipsoidal particle of BiVO4 with some grain size was observed by the surface SEM image when the SnO2 intermediate was present. The ratio of Sn/W by XPS in BiVO4 /SnO2 /WO3 /FTO was smaller than that in SnO2 /WO3 /FTO. It is suggested that BiVO4 was coated preferentially on the Snrich part of the surface over the SnO2 /WO3 /FTO electrode. The ratios of Bi/V in BiVO4 /SnO2 /WO3 /FTO and BiVO4 /WO3 /FTO were similar each other, but the ratios of (Bi + V)/(Sn + W) were different. As the number of deposited BiVO4 film on a SnO2 /WO3 /FTO photoelectrode increased, the atomic ratio of BiVO4 /(Sn + W) in BiVO4 /SnO2 /WO3 /FTO increased. These results suggest that the way of BiVO4 crystal growth on the under layer and the decrease Table 2 W, Sn, Bi, and V atomic composition in WO3 /FTO, SnO2 /WO3 /FTO, BiVO4 /WO3 /FTO and BiVO4 /SnO2 /WO3 /FTO films from XPS. The concentration of precursor solution of SnO2 intermediate layer; 5 mM. Photoelectrode

WO3 /FTO SnO2 /WO3 /FTO BiVO4 /WO3 /FTO BiVO4 /SnO2 /WO3 /FTO

Elements atomic (%) W

Sn

Bi

V

100 47 17 19

0 53 0 12

0 0 37 30

0 0 46 39

56

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

Fig. 5. Illustration of the light trapping structures of a single (a-1) and double stack (b-1). (A), mask; (B), BiVO4 /WO3 photoelectrode; (C), white plate. I–V characteristics of ; without white plate. the single- (a-2) and double-stacked (b-2) BiVO4 /WO3 photoelectrode in a 0.1 M KHCO3 aqueous solution. Symbol, (a-2) —; with white plate, ; only Front 1, ; only Front 2. (b-2) —; Front 1 and Front 2 combination,

1

(a)

(b)

70

60 0.8

50

IPCE (%)

LHE

0.6 0.4

0.2

40 30 20

10 0

0 300

400 500 Wavelength/ nm

600

300

400 500 Wavelength/ nm

600

Fig. 6. LHE spectra (a) and IPCE spectra (b) of the BiVO4 /WO3 photoelectrode under front illumination. IPCE was measured at +1.2 V vs. RHE in 0.1 M KHCO3 . Symbol, (a) single; — double stacks. (b) ♦ Front 1,  Front 2,  double stacks.

of porosity may one of the factors affected on the photoelectrochemical properties. In the I–V curves shown in Fig. 7, the photocurrent increased when the SnO2 intermediate layer was prepared from 5 mM or 1 mM concentration of precursor SnO2 solution. On the other hand, the photocurrent decreased when the SnO2 intermediate layer

Current density/mAcm-2

4

(c) (b)

3

(a)

2

(d)

1 0 0.4 -1

0.8

1.2 1.6 Potential/V vs RHE

2

Fig. 7. I–V characteristics of the BiVO4 /SnO2 /WO3 photoanode which was prepared with various precursor SnO2 solution. (a) Without an intermediate SnO2 layer, (b) 1 mM, (c) 5 mM and (d) 50 mM of precursor SnO2 solution. The electrolyte aqueous solution was 0.1 M KHCO3 .

was prepared from the 50 mM precursor SnO2 solution. The SnO2 layer thickness was regulated by the concentration of its precursor solution. It is concluded that the intermediate layer film thickness affected the photocurrent, and this improvement became apparent in the very thin SnO2 intermediate layer. We measured the photocurrent of the BiVO4 /SnO2 photoelectrode. The photocurrent at 1.23 and 0.8 V vs. RHE of the BiVO4 /SnO2 increased by about 1.6 times and 1.1 times compared to bare BiVO4 photoelectrode, respectively. However, this under layer effect by SnO2 was smaller than that by WO3 (the photocurrent increased by about 2.8 times and 14 times, respectively). Therefore, the design of the multi-layer BiVO4 composition film photoelectrode was important. The cross-section of BiVO4 /SnO2 /WO3 photoelectrode with the SnO2 intermediate layer prepared from 5 mM precursor SnO2 solution was observed by TEM (Fig. S3). In a cross-section image, the thickness of the SnO2 layer was not identified clearly. However, it is thought that the very thin SnO2 intermediate layer is existed because the SnO2 was confirmed by the XPS analysis of the multi-composite photoelectrode surfaces as mentioned above. The SnO2 layer thickness was roughly calculated at approximately 6 nm in case of the 50 mM precursor solution, assuming that the

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

57

Table 3 The photocurrent properties of BiVO4 /SnO2 /WO3 photoelectrode in various electrolyte aqueous solutions. Photoelectrode

BiVO4 /SnO2 /WO3

a

Electrolyte

KHCO3 + CO2 Na2 SO4 KHCO3 + CO2 Na2 SO4

Concentration/M

0.1 0.1 1.5 S(ca. 1.5)a

pH

6.9 6.2 7.9 6.1

Conductivity/mS cm−1

Photocurrent/mA cm−2

7.7 14 78 72

Onset potential/VRHE

at l.23 VRHE

at 0.8 VRHE

2.15 1.26 2.24 1.23

1.12 0.16 1.60 0.17

0.43 0.74 0.33 0.66

Saturated solution (S).

SnO2 covered whole WO3 surface uniformly. The IPCE spectrum of BiVO4 /SnO2 /WO3 photoelectrodes is shown in Fig. 8. The maximum IPCE value of the BiVO4 /SnO2 /WO3 photoelectrode in visible light region in high concentration KHCO3 (2.5 M) was 53% (at 420 nm, 1.23 V vs. RHE). Table 3 shows the photoelectrochemical properties of the BiVO4 /SnO2 /WO3 photoelectrode in carbonate and sulfate electrolyte aqueous solutions. The increase of the photocurrent was confirmed under the carbonate electrolyte. The onset potential in the carbonate electrolyte shifted 0.3 V to the negative direction in comparison with the sulfate electrolyte. We investigated the resistance components on various photoanodes by EIS measurement. The Nyquist plots with a semicircle were obtained (Fig. 9). They were distinguished between resistances with and without parallel electrical capacitance component. The resistances without capacitance, which were attributable to the conducting glass and electrolyte conductivity, were ca. 100  in all photoanodes. The resistances at the semicircle with parallel capacitance component were attributable to the semiconductor films, and were changed by the applied potential. We attracted

Fig. 8. IPCE spectrum of multi-layer composite thin film BiVO4 /SnO2 /WO3 photoanodes in 2.5 M KHCO3 aqueous solution. The applied bias was +1.23 V vs. RHE.

attention with the shape of the I–V curve of the multi-layer composite BiVO4 photoelectrode which showed the high photocurrent. The data in Fig. 9 were obtained at +0.9 V vs. RHE, because the resistance become smallest and the maximum ABPE were obtained at around the potential. The resistance at semicircle of the bare BiVO4 photoanodes were very large (13,000 ) compared with the composite photoanode. The semicircles of the composite photoanodes were flatly distorted, suggesting that a speculated equivalence circuit with resistances and capacitances are complicated. The resistance at semicircle and the total resistance of the BiVO4 /SnO2 /WO3 photoanodes (120  and 210 ) were smaller than those of the BiVO4 /WO3 (160  and 260 ). The total resistances were almost coincided with those in the I–V properties at +0.9 V vs. RHE. The capacitances of the BiVO4 /SnO2 /WO3 , BiVO4 /WO3 and BiVO4 electrodes, estimated by the nyquist plots of these photoelectrodes, were almost same level (7.6 × 10−5 , 2.4 × 10−5 and 4.5 × 10−5 F cm−2 , respectively). This suggests that there is not much difference in the carrier density of those photoelectrodes. Therefore, it is surmised that the great resistance improvement by the under layers was caused by other factors such as improvement of carrier mobility or decrease of interfacial barrier, rather than the increase of the carrier density. Next, the reasons of the photocurrent increase by the SnO2 intermediate layer were discussed on the basis of the position of the energy-band of the n-type semiconductor. Fig. 10 shows the speculated energy-band diagram of the BiVO4 /SnO2 /WO3 composite thin-film photoanode in the KHCO3 electrolyte aqueous solution (neutral pH). Under the irradiation from the side of BiVO4 film, photoinduced hole–electron pairs are produced mainly at the BiVO4 film surface. The potentials of the valence band top (EVB ) of WO3 and SnO2 semiconductor are more positive than that of BiVO4 , and especially the EVB of SnO2 is very positive. It was reported that the photoinduced hole could not transfer from BiVO4 to the under-layer

Potential/V vs NHE (pH=7)

e-

0

CB

hν O2

e-

CO2 HCO3-

+1.0

H 2O

+2.0 h+

+3.0

VB

FTO

BiVO4

WO3 SnO2

Fig. 9. Impedance spectra of the multi-layer composite BiVO4 and bare BiVO4 photoelectrode in KHCO3 (0.1 M) aqueous solution. Nyquist plot of the impedance measurements at +0.9 V vs. RHE with 5 mV of amplitude. The frequency was from BiVO4 /WO3 , bare BiVO4 . 1 MHz to 10 mHz. Symbol; BiVO4 /SnO2 /WO3 ,

Fig. 10. Speculated energy diagram of the BiVO4 /SnO2 /WO3 multi-composite thinfilm photoanode in a carbonate electrolyte solution (pH 7) for photoelectrochemical water splitting.

58

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

100

2

1

0 0

1

2

3

4

5

Time/h Fig. 11. Time course of the photocurrent of the BiVO4 /SnO2 /WO3 photoelectrode in 0.1 M KHCO3 with stirring under +1.2 V vs. RHE anodic bias.

semiconductor due to the deep under-layer EVB [16,18]. Therefore, it is thought that the hole was immediately consumed to oxidize H2 O and/or HCO3 − . The flat band potential (Efb ) values (V vs. NHE, pH = 7) of BiVO4 , SnO2 , and WO3 are −0.37 [14], +0.074 [38], and +0.32 [14,39], respectively. The Efb is close to the bottom of the ECB and the Fermi level (EF ) for an n-type semiconductor. It is considered that a cascade electron transfer from BiVO4 to WO3 through SnO2 and these conduction bands might occur. However, there is a possibility that an energy barrier was generated by contact with oxide semiconductors with different values of EF [40]. If there is an energy barrier between the BiVO4 and WO3 , it was thought that the energy barrier height can be reduced by the insertion of SnO2 , because the difference of Ef between the contacted semiconductors would become small. Therefore, it is inferred that an excited electron in BiVO4 layer became easy to move to WO3 film by the insertion of SnO2 intermediate layer. Moreover, it is thought that the electrons that were transferred to the ECB of WO3 would be recombined with the holes in the EVB of BiVO4 , but the presence of SnO2 intermediate layer might inhibit this unfavorable pathway. Therefore, the photoinduced electron transfer from BiVO4 to FTO glass would be facilitated. We tried to compared the onset potentials and the Efb s measured by Mott–Schottky plots of the various photoelectrodes in different solutions. We found as a tendency that the Efb differences in various aqueous solutions were not distinguished, and that the Efb s of the BiVO4 and multi-layer composite BiVO4 thin film electrodes were similar. The order of the Efb s of these electrodes was accorded with these onset potentials in a solution. But, the reason of the negative shift of the onset potential in carbonate solution is not clear by the Efb measurement now. We investigated the photostability of BiVO4 /SnO2 /WO3 photoelectrode in the KHCO3 electrolyte aqueous solution at +1.2 V vs. RHE anodic bias for 5 h (Fig. 11). The photocurrent was reduced by approximately 3.8% from the initial value after light irradiation. In the literature for the BiVO4 /WO3 [16] and BiVO4 /SnO2 [18] photoelectrodes, it appeared that the photocurrent of these photoelectrodes in an Na2 SO4 electrolyte aqueous solution under a constant anodic bias decreased by approximately 17% and 60%, respectively, within 1 min. The photocurrent of the BiVO4 layer/WO3 nanorod heterojunction photoelectrode in an Na2 SO4 aqueous solution decreased by 60% within ca. 40 min [15]. In comparison with these, it is thought that our multi-layer composite BiVO4 photoelectrode has photostability in KHCO3 (0.1 M) aqueous solution. The turnover number of electrons which flowed through the circuit to the BiVO4 unit was over 4000 in an irradiation time of 5 h. The quantity of BiVO4 of the multi-layer composite photoelectrode after 28 h photoanodic reaction did not decrease at all in comparison with before its reaction. Moreover, the BiVO4 /SnO2 /WO3 photoelectrode was soaked in 0.1 M KHCO3 under 0.1 MPa CO2 gas in a sealed cylindrical cell for ca. 5 days. The turnover number of electrons to the BiVO4 unit was greater than 2.5 × 104 in an irradiation time of 113 h.

Evolved gases/μmol

Current density/mAcm-2

3

80 60 40

20 0 0

2

4

6

8

Irradiation time/h Fig. 12. Time course of H2 and  O2 evolution using the BiVO4 /SnO2 /WO3 photoanode in a KHCO3 electrolyte aqueous solution under Xe lamp irradiation. The applied bias was +1.1 V to the Pt counter electrode. The dashed line is the amount of evolved gas that is calculated by the total passage of photoelectrons.

The measurement of H2 and O2 generated by the water splitting is very important. Fig. 12 shows the time course of evolved gases under light irradiation conditions using the multi-layer composite BiVO4 photoelectrode in two-electrode system with Pt counter electrode. Evolution of H2 and O2 at a stoichiometric ratio (H2 /O2 = 8.9/4.5 ␮mol h−1 ) was observed under light irradiation. Additionally, the increase of the gases was not obtained in the dark. The total amount of evolved H2 gas at the early period was almost coincident with that calculated by the total passage of photoelectrons. The H2 gas evolution at the latter period was lower than the expected value from the current, and it is mainly caused by the backward reaction such as the water production from H2 and O2 on the Pt counter electrode. 3.6. Applied bias photon-to-current efficiency (ABPE) of the photoelectrochemical cell composed of the BiVO4 /SnO2 /WO3 and the Pt electrode Finally, we calculated the ABPE. As discussed in Section 2, we used one-compartment cell with two electrode system for the calculation of the ABPE. Fig. 13 shows I–V curves of single and double stacks of the BiVO4 /SnO2 /WO3 photoelectrode in a highly concentrated carbonate electrolyte. The optimum potential position, when the red rectangular area should be maximized in Fig. 13 and Fig. S4, becomes the Eopt . Fig. 14 shows the dependence of ABPE on [1.23 – applied potential (E)] for the photoelectrochemical cell, and the position of the peak is ABPE at the Eopt . The high value of ABPE greater than 1% was obtained using the double stacks of the BiVO4 /SnO2 /WO3 photoelectrode with an optical confinement structure in a highly concentrated carbonate electrolyte. Using the 2.5 M KHCO3 electrolyte solution, the maximum ABPE in

Fig. 13. I–V characteristics of a photoelectrochemical cell composed of the BiVO4 /SnO2 /WO3 photoanode and Pt (H+ reducing) cathode in 2.5 M KHCO3 . Sweep direction, +1.6 → +0.3 V vs. C.E. Scan rate, 50 mV s−1 . Light source; solar simulator ; Double-stacked photoelectrode, ; single(1 Sun, 100 mW cm−2 ). stacked photoelectrode.

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

ABPE (%)

1.5

1 0.5 0 0

0.5

1

1.23-E/V vs C.E. Fig. 14. ABPE as a function of applied bias for the photoelectrochemical cell composed of the BiVO4 /SnO2 /WO3 photoelectrode and Pt electrode. Sweep direction (E), +1.6 → +0.3 V vs. C.E. Symbol, ♦; single-stacked photoelectrode in 0.1 M KHCO3 , ; single-stacked photoelectrode in 2.5 M KHCO3 , ; double-stacked photoelectrode in 2.5 M KHCO3 .

the single stack and double stacks of the photoelectrode reached 0.86% (at Eopt = 0.76 V) and 1.35% (at Eopt = 0.76 V), respectively. We draw a comparison between these ABPE values and the photocurrent characteristics of oxide photoanodes published in other articles. As for having to be careful, we calculated ABPE from the I–V curves of two electrodes system, however, a lot of the photocurrent characteristics have been reported in the I–V curves of three electrodes systems. Generally, in the two electrode system, an influence of the overvoltage for hydrogen generation at the counter electrode is included. However, in the I–V curves of three electrode system, a calculated ABPE were overestimated as compared with the I–V curves of two electrode system because the overvoltage for counter electrode is not included. Actually, we compared the I–V curves of two electrodes system with the I–V curves of three electrodes system as Fig. S4 shows. Those I–V curves were similar very much, however, the calculated ABPE by the I–V curve of three electrodes system were slightly higher than the I–V curve of two electrodes system. Therefore, we thought that if the overestimation in three electrodes system is recognized, the estimated ABPE by the I–V curves of photoelectrode published in other articles in three electrodes system could be compared with the our data in two electrode system. The estimated ABPE of WO3 film photoelectrode was 0.68% (Eopt = 0.87 VRHE ) from a paper of Augustynski et al. [6]. Recently, Solarska et al. reported the photoanode properties of a nanocrystalline WO3 photoelectrode in methane sulfonic acid electrolyte. According to the paper, ABPE was 0.8% (Eopt = 0.8 VRHE ) [30]. The estimated ABPE of ␣-Fe2 O3 /Pt nanorod photoelectrode was 1.1% (Eopt = 0.88 VRHE ) from a paper of Park et al. [11]. The estimated ABPE of FeOOH/BiVO4 photoelectrode was 0.79% (Eopt = 0.60 VRHE ) from a paper of Seabold and Choi [24]. The estimated ABPE of BiVO4 /WO3 nanorod heterojuction film photoelectrode was 0.7% (Eopt = 0.61 VRHE ) from a paper of Su et al. [15]. Thus, to the best of our knowledge, the 0.86% of the our BiVO4 /SnO2 /WO3 photoelectrode in highly carbonate aqueous solution value was the highest in the BiVO4 photoelectrodes among the report that we recognized, this value (0.86%) was the second-highest after the Fe2 O3 with precious metal (1.1%), and 1.35% value of double stacks of BiVO4 /SnO2 /WO3 photoelectrode was the highest among oxide photoelectrodes. 4. Conclusion In this paper, the photoelectrochemical properties of multilayer composite photoelectrode of BiVO4 thin film in carbonate electrolyte solution were investigated in detail. BiVO4 /WO3 photoelectrodes with several different BiVO4 upper-layer film thicknesses were prepared. With an increase in the BiVO4 film thickness, the BiVO4 particle grew and the interparticle crevice widths became narrower, and the LHE of this

59

photoelectrode increased remarkably until BiVO4 film thickness became ca. 150 nm. The photocurrent of the BiVO4 /WO3 composite photoelectrode in a carbonate electrolyte solution increased compared to the sulfate electrolyte solution that has generally been used, and it was found that HCO3 − anions mainly affecting the BiVO4 film layer, not the WO3 layer. In the visible region, the LHE and IPCE of the multi-layer composite photoelectrode of BiVO4 thin film were significantly improved by the light trapping structure of the double stacked photoelectrodes. The photocurrent of the BiVO4 /WO3 photoelectrode was increased by insertion of a very thin SnO2 intermediate, and the highest IPCE at 420 nm was 53% in 2.5 M KHCO3 aqueous solution. In the EIS measurement, the total resistance decreased remarkably by the inserting WO3 -under-layer and SnO2 intermediate layer in comparison with the bare BiVO4 . These decreases of total resistance would be led to the improvement of photocurrent. For the water splitting of the BiVO4 /SnO2 /WO3 photoelectrode, the reaction mechanism of effects of the inserted SnO2 intermediate layer and carbonate electrolyte were discussed; it was thought that the excited electron transfer was positively controlled by the contact with oxide semiconductors with different band structures, and that the carbonate anion behaved like a catalyst for irreversible O2 evolution on the BiVO4 film surface. The H2 and O2 were evolved stoichiometrically. In the double stacks BiVO4 /SnO2 /WO3 photoelectrode, the maximum value of ABPE was 1.35% in highly concentrated KHCO3 solution. To the best of our knowledge, this value was the highest among all reports on the oxide semiconductor photoanodes for water splitting. Moreover, the APCE (absorbed photon-to-current efficiency) of the BiVO4 /SnO2 /WO3 photoelectrode in 2.5 M KHCO3 was ca. 76%; i.e., this value shows the intrinsic quantum efficiency of the photocurrent generated from the excited electron. Therefore, further performance enhancement will be possible by modified preparation of photoelectrode with high charge separation properties. Acknowledgements The present work was partially supported by the Funding Program for Next Generation World-Leading Researchers (NEXT Program) from the Cabinet Office, Government of Japan. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. jphotochem.2013.02.019. References [1] A. Fujishima, K. Honda, Electrochemical photolysis of water at a semiconductor electrode, Nature 238 (1972) 37–38. [2] M. Grätzel, Photoelectrochemical cells, Nature 414 (2001) 338–344. [3] Y. Lin, G. Yuan, R. Liu, S. Zhou, S.W. Sheehan, D. Wnag, Semiconductor nanostructure-based photoelectrochemical water splitting: A brief review, Chemical Physics Letters 507 (2011) 209–215. [4] A. Kudo, K. Omori, H. Kato, A novel aqueous process for preparation of crystal form-controlled and highly crystalline BiVO4 powder from layered vanadates at room temperature and its photocatalytic and photophysical properties, Journal of the American Chemical Society 121 (1999) 11459–11467. [5] Z. Chen, T.F. Jaramillo, T.G. Deutsch, A. Kleiman-Shwarsctein, A.J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E.W. McFarland, K. Domen, E.L. Miller, J.A. Turner, H.N. Dinh, Accelerating materials development for photoelectrochemical hydrogen production: standards for methods, definitions, and reporting protocols, Journal of Materials Research 53 (2010) 3–16. [6] B.D. Alexander, P.J. Kulesza, I. Rutkowska, R. Solarska, J. Augustynski, Metal oxide photoanodes for solar hydrogen production, Journal of Materials Chemistry 18 (2008) 2298–2303. [7] I. Cesar, A. Kay, J.A.G. Martinez, M. Grätzel, Translucent thin film Fe2 O3 photoanodes for effient water splitting by sunlight: nanostructure-directing effect of Si-doping, Journal of the American Chemical Society 128 (2006) 4582–4583.

60

R. Saito et al. / Journal of Photochemistry and Photobiology A: Chemistry 258 (2013) 51–60

[8] A. Kay, I. Cesar, M. Grätzel, New benchmark for water photooxidation by nanostructured ␣-Fe2 O3 films, Journal of the American Chemical Society 128 (2006) 15714–15721. [9] S.D. Tilley, M. Cornuz, K. Sivula, M. Grätzel, Light-induced water splitting with hematite: improved nanostructure and iridium oxide catalysis, Angewandte Chemie International Edition 49 (2010) 6405–6408. [10] T. Hisatomi, F.L. Formal, M. Cornuz, J. Brillet, N. Tétreault, K. Sivula, M. Grätzel, Cathodic shift in onset potential of solar oxygen evolution on hematite by 13group oxide overlayers, Energy & Environmental Science 4 (2011) 2512–2515. [11] A. Mao, N.G. Park, G.Y. Han, J.H. Park, Controlled growth of vertically oriented hematite/Pt composite nanorod arrays: use for photoelectrochemical water splitting, Nanotechnology 22 (2011) 175703 (7 pp). [12] J. Brillet, M. Cornuz, F.L. Formal, J.H. Yum, M. Grätzel, K. Sivula, Examining architectures of photoanode–photovoltaic tandem cells for solar water splitting, Journal of Materials Research 25 (2010) 17–24. [13] O.K. Varghese, C.A. Grimes, Appropriate strategies for determining the photoconversion efficiency of water photoelectrolysis cells: A review with examples using titania nanotube array photoanodes, Solar Energy Materials & Solar Cells 92 (2008) 374–384. [14] S.J. Hong, S. Lee, J.S. Jang, J.S. Lee, Heterojunction BiVO4 /WO3 electrodes for enhanced photoactivity of water oxidation, Energy & Environmental Science 4 (2011) 1781–1787. [15] J. Su, L. Guo, N. Bao, C.A. Grimes, Nanostructured WO3 /BiVO4 heterojunction films for efficient photoelectrochemical water splitting, Nano Letters 11 (2011) 1928–1933. [16] P. Chatchai, Y. Murakami, S. Kishioka, A.Y. Nosaka, Y. Nosaka, Efficient photocatalytic activity of water oxidation over WO3 /BiVO4 composite under visible light irradiation, Electrochimica Acta 54 (2009) 1147–1152. [17] Y. Liang, T. Tsubota, L.P.A. Mooij, R. van de Krol, Highly improved quantum efficiencies for thin film BiVO4 photoanodes, Journal of Physical Chemistry C 115 (2011) 17594–17598. [18] P. Chatchai, Y. Murakami, S.-Y. Kishioka, A.Y. Nosaka, Y. Nosaka, FTO/SnO2 /BiVO4 composite photoelectrode for water oxidation under visible light irradiation, Electrochemical and Solid-State Letters 11 (2008) H160–H163. [19] K. Sayama, A. Nomura, T. Arai, T. Sugita, R. Abe, M. Yanagida, T. Oi, Y. Iwasaki, Y. Abe, H. Sugihara, Photoelectrochemical decomposition of water into H2 and O2 on porous BiVO4 thin-film electrodes under visible light and effect of Ag ion treatment, Journal of Physical Chemistry B 110 (2006) 11352–11360. [20] D.K. Zhong, S. Choi, D.R. Gamelin, Near-complete suppression of surface recombination in solar photoelectrolysis by Co–Pi catalyst-modified W:BiVO4 , Journal of the American Chemical Society 133 (2011) 18370–18377. [21] T.H. Jeon, W. Choi, H. Park, Cobalt–phosphate complexes catalyzed the photoelectrochemical water oxidation of BiVO4 electrodes, Physical Chemistry Chemical Physics 13 (2011) 21392–21401. [22] S.K. Pilli, T.E. Furtak, L.D. Brown, T.G. Deutsch, J.A. Turner, A.M. Herring, Cobalt–phosphate (Co–Pi) catalyst modified Mo-doped BiVO4 photoelectrodes for solar water oxidation, Energy Environmental Science 4 (2011) 5028–5034. [23] W. Luo, Z. Li, T. Yu, Z. Zou, Effects of surface electrochemical pretreatment on the photoelectrochemical performance of Mo-doped BiVO4 , Journal of Physical Chemistry C 116 (2012) 5076–5081. [24] J.A. Seabold, K.S. Choi, Efficient and stable photo-oxidation of water by a bithmuth vanadate photoanode coupled with an iron oxyhydroxide

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

oxygen evolution catalyst, Journal of the American Chemical Society 134 (2012) 2186–2192. Q. Jia, K. Iwashina, A. Kudo, Facile fabrication of an efficient BiVO4 thin film electrode for water splitting under visible light irradiation, Proceedings of the National Academy of Sciences of the United States of America 109 (2012) 11564–11569. K. Sayama, N. Wang, Y. Miseki, H. Kusama, N. Onozawa-Komatsuzaki, H. Sugihara, Effect of carbonate ions on the photooxidation of water over porous BiVO4 film photoelectrode under visible light, Chemistry Letters 39 (2011) 17–19. R. Saito, Y. Miseki, K. Sayama, Highly efficient photoelectrochemical water splitting using a thin film photoanode of BiVO4 /SnO2 /WO3 multi-composite in a carbonate electrolyte, Chemical Communications 48 (2012) 3833–3855. K. Sayama, H. Hayashi, T. Arai, M. Yanagida, T. Gunji, H. Sugihara, Highly active WO3 semiconductor photocatalyst prepared from amorphous peroxo-tungstic acid for the degradation of various organic compounds, Applied Catalysis B: Environmental 94 (2010) 150–157. H. Müllejans, A. Ioannides, R. Kenny, W. Zaaiman, H.A. Ossenbrink, E.D. Dunlop, Spectral mismatch in calibration of photovoltaic reference devices by global sunlight method, Measurement Science and Technology 16 (2005) 1250–1254. R. Solarska, R. Jurczakowski, J. Augustynski, A highly stable, efficient visiblelight driven water photoelectrolysis system using a nanocrystalline WO3 photoanode and a methane sulfonic acid electrolyte, Nanoscale 4 (2012) 1553–1556. J. Kim, C.W. Lee, W. Choi, Platinized WO3 as an environmental photocatalyst that generates OH radicals under visible light, Environmental Science and Technology 44 (2010) 6849–6854. R. Strobel, H.J. Metz, S.E. Pratsinis, Brilliant yellow transparent pure, and SiO2 coated BiVO4 nanoparticles made in flames, Chemistry of Materials 20 (2008) 3651–6346. K. Sayama, H. Arakawa, Effect of carbonate salt addition on the photocatalytic decomposition of liquid water over Pt–TiO2 catalyst, Journal of the Chemical Society, Faraday Transactions 93 (1997) 1647–1654. D.P. Jones, W.P. Griffith, Alkali-metal peroxocarbonates, M2 [CO3 ]·nH2 O2 , M2 [C2 O6 ], M[HCO4 ]·nH2 O, and Li2 [CO4 ]·H2 O, Journal of the Chemical Society, Dalton Transactions (1980) 2526–2532. C.A.S. Regino, D.E. Richardson, Bicarbonate-catalyzed hydrogen peroxide oxidation of cysteine and related thiols, Inorganica Chimica Acta 360 (2007) 3971–3977. A. Kh Mel’nikov, T.P. Frisova, A.N. Molokina, Preparation of pure potassium peroxodicarbonate and some of its properties, Russian Journal of Inorganic Chemistry 7 (1962) 637–640. N.M. Dimitrijevic, B.K. Vijayan, O.G. Poluektov, T. Rajh, K.A. Gray, H. He, P. Zapol, Role of water and carbonates in photocatalytic transformation of CO2 to CH4 on titania, Journal of the American Chemical Society 133 (2011) 3964–3971. J. Bandra, R.A.S.S. Ranasinghe, Thae effect of MgO coating on photocatalytic activity of SnO2 for the degradation of chlorophenol and textile colorants; the correlation between the photocatalytic activity and the negative shift of flatband potential of SnO2 , Applied Catalysis A-General 319 (2007) 58–63. M. Fujii, T. Kawai, S. Kawai, Photocatalytic activity and the energy levels of electrons in a semiconductor particle under irradiation, Chemical Physics Letters 106 (1984) 517–522. J.F. Wager, Transparent electronics: Schottky barrier and heterojunction considerations, Thin Solid Films 516 (2008) 1755–1764.