i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 Available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/he Ag grid induced photocurrent enhancement in WO3 photoanodes and their scale-up performance toward photoelectrochemical H2 generation Won Jae Lee*, Pravin S. Shinde, Guen Ho Go, Easwaramoorthi Ramasamy Nanohybrid Energy Conversion Devices Research Center, Korea Electrotechnology Research Institute (KERI), Changwon 641-120, Republic of Korea article info abstract Article history: The hydrogen generation from photoelectrochemical (PEC) water splitting under visible Received 10 December 2010 light was investigated using large area tungsten oxide (WO3) photoanodes. The photo- Accepted 3 February 2011 anodes for PEC hydrogen generation were prepared by screen printing WO3 films having Available online 5 March 2011 typical active areas of 0.36, 4.8 and 130 cm2 onto the conducting fluorine-doped tin oxide (FTO) substrates with and without embedded inter-connected Ag grid lines. TiO2 based Keywords: dye-sensitized solar cell was also fabricated to provide the required external bias to the WO3 photoanodes for water splitting. The structural and morphological properties of the WO3 Scale-up films were studied before scaling up the area of photoanodes. The screen printed WO3 film PEC water splitting sintered at 500  C for 30 min crystallized in a monoclinic crystal structure, which is the Dye-sensitized solar cell most useful phase for water splitting. Such WO3 film revealed nanocrystalline and porous Solar-to-hydrogen morphology with grain size of w70e90 nm. WO3 photoanode coated on Ag grid embedded conversion efficiency FTO substrate exhibited almost two-fold degree of photocurrent density enhancement than that on bare FTO substrate under 1 SUN illumination in 0.5 M H2SO4 electrolyte. With such enhancement, the calculated solar-to-hydrogen conversion efficiencies under 1 SUN were 3.24% and w2% at 1.23 V for small (0.36 cm2) and large (4.8 cm2) area WO3 photoanodes, respectively. The rate of hydrogen generation for large area photoanode (130.56 cm2) was 3 mL/min. Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. 1. Introduction Photoelectrochemical (PEC) hydrogen (H2) production involving photovoltaic electrolysis of water under sunlight illumination has been investigated as a potential means of a clean, environmentally friendly, large scale fuel production and lot of review articles and research papers can be found in the literature [1e6]. The solar powered PEC H2 production system is schematically illustrated in Fig. 1(a). It consists of two components mainly a PEC cell for producing hydrogen at counter electrode and oxygen at semiconductor electrode; and a dye-sensitized solar cell (DSSC) bias power for supplying the bias to the PEC cell. Fig. 1b shows the energetic of hydrogen generation from PEC water splitting. More details on energetic of semiconductor anode for H2 generation can be found in the literature [7]. Some of the prime materials requirements for efficient visible light water splitting are: i) the semiconductor band gap should be in the range of 1.5e3.2 eV considering the hydrogen generation potential to be 1.23 eV so that it could absorb a significant portion of the visible part of the solar * Corresponding author. E-mail address: wjlee@keri.re.kr (W.J. Lee). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.02.013 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 5263 Fig. 1 e (a) Schematic diagram for PEC H2 production system consisting of a semiconductor photoanode (where O2 is produced), a counter electrode (where H2 is produced) and a DSSC for external bias; (b) Energy-level diagram for a PEC cell with zero electrical bias. Two half-cell reactions occurring at the photoanode and cathode to generate O2 and H2 are shown. O2/H2O and H2/H2O are redox potentials for generation of O2 and H2. Vfb is the flat band potential and Uox is the built-in oxygen over-potential. Conditions for efficient water splitting: Ec < Ered(H2O/H2) and Ev > Eox(O2/H2O). spectrum (380e830 nm); ii) the conduction band edge (Ec) position of semiconductor should be at a more negative potential than the reduction potential of water (Ered) while the valence band edge (Ev) position to be at a more positive potential than the oxidation reaction (Eox) i.e. energetic of semiconductor should meet the conditions: Ec < Ered(H2O/H2) and Ev > Eox(O2/H2O); iii) the photocatalyst should have a sufficiently negative flat band potential so as to have sufficient over-potential to drive the water splitting reaction at a reasonable rate; and finally iv) semiconductor should be stable against photo-corrosion in aqueous environment [8,9]. Meeting these criteria, few photocatalysts such as TiO2, Fe2O3, WO3, BiVO4 etc. have been investigated for water splitting to produce H2 under light illumination [10e16]. Among these, Fe2O3 is a narrow band gap (w2.2 eV) material capable of absorbing w40% of solar energy (up to 600 nm) which is sufficiently enough for H2 production mostly in aqueous solutions (pH>3). However, poor electron mobility (0.01e0.1 cm2 V 1s 1) consequently leads to rapid electronhole recombination and a very low hole transport due to short hole diffusion length (2e4 nm) [17]. In other words, it has insufficient negative flat band potential. Despite having good catalytic activity and stability over wide pH range, TiO2 is generally limited by too large band gap (w3.2 eV), which fail to absorb a significant fraction of visible light (below 385 nm), resulting in poor conversion efficiencies under terrestrial conditions. WO3 is again a wide band gap material with slightly less band gap (2.7e2.8 eV) that can absorb reasonable part of solar spectrum. WO3 with high hole diffusion length (w150 nm) as compared with Fe2O3 (w2 nm) and TiO2 (w20 nm) can be a potential material for water splitting as immediate recombination of electron-hole pairs could be averted inside the semiconductor. It is stable in strong acidic (pH<4) solutions (e.g. aq. H2SO4, H3PO4), however, shows low chemical stability in alkaline solution (e.g. aq. NaOH, KOH). The semiconductor based PEC water splitting system involving a combination of photovoltaic cells and semiconductor liquid junctions (PV/SCLJ) approach is found to be more efficient for H2 production [18]. In this approach, the required input energy both for splitting of water and generation of external bias is provided by means of solar light. Swiss research group [19] in an attempt proposed the tandem cell system with a solar-to-hydrogen (STH) conversion efficiency up to 7% using DSSC bias of w1.5 V. This so-called STH 5264 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 efficiency is sufficiently high in contrast to the theoretical maximum STH efficiency for any material (only 20%) [20]; and is therefore encouraging for further research. Thermodynamically, the STH conversion efficiency of WO3-based devices remains less than 5% [20]. Arakawa et al. [21] reported solar hydrogen production with STH efficiency of 2.8% from a tandem cell system consisting of metal oxide semiconductor films of area 1  1 cm2 and DSSC. A STH efficiency of 3.6% is observed for 2.5 mm thick WO3 photoanode in 3 M H2SO4 and Pt cathode using a PEC/PV ‘‘tandem cell’’ configuration [2]. The, current PEC performance status of 4% STH has been achieved in multijunction configurations using WO3 PEC interfaces [22]. However, all the previous attempts toward H2 production using WO3 photoanode are limited only to small electrode area (<1 cm2) at laboratory scale. There are no studies of hydrogen production using large area photoanodes. Since, photoanode is a key element in PEC water splitting system, it is important to witness its scale-up performance in a development for full realization of hydrogen in the energy market. Our group is mainly working on portable, scale-up processes for energy harvesting nanocrystalline materials to meet the energy requirements for their practical applications [23e25]. Transparent conducting oxides (TCOs) such as fluorinedoped tin oxide (FTO) are the commonly used substrates to make photoanodes for PEC water splitting systems, solar cells etc. However, relatively high sheet resistance (10e15 U/,) of TCO substrates and its further increase upon high temperature annealing [26] has delayed the entry of large area PEC cells into the commercial market. PEC performance of scaled up electrodes is lower than that of tiny electrodes, since a carrier loss occurs in resistive TCO substrate. Here we demonstrate a simple method to reduce resistive loss and improve the current collection of photogenerated charge carriers via strip type semiconductor layers coated between inter-connected metal grids embedded TCO substrates. Silver has been a material of choice for grid line application in solar cells for enhancing the current collection because of its high conductivity and low dark current [27]. To the best of authors’ knowledge, nobody has employed Ag grid embedded large area TCO substrates for PEC water splitting to generate hydrogen. In the present communication, we have successfully demonstrated the scale-up process for WO3 photoanodes prepared on Ag grid embedded FTO substrates by screen printing method with remarkable improvement in photocurrent and hence hydrogen generation. PEC H2 generation has been achieved by applying required external bias to the photoanodes using array of TiO2 based DSSC module. The novelty of our work is that we have fabricated photoanodes with photoactive area > 1 cm2 and have employed metal grids to overcome the internal resistance in TCO substrates for efficient current collection in scaled up photoanodes. 2. 2.1. were also used to fabricate electrodes for DSSC module. The FTO substrates were cut into required dimension and successively cleaned using acetone (for 10 min), ethanol (for 10 min) and deionized water (for 10 min) in ultrasonic cleaner. FTO substrates of different sizes mainly 1  1.2, 1  10 and 15  15 cm2 were used after drying under nitrogen gas. Silver current collector grid lines (dimension, width  height: 0.6 mm  10 mm) were printed on FTO substrates of different sizes using screen printable silver paste by a semi-automatic screen printing machine (Automax). After drying at room temperature for 30 min, Ag embedded FTO substrates were heat treated at 180  C for 10 min. The distance between the Ag metal grid lines was set to 8 mm (see Fig. 2). Two point measurements show that after heat treatment at 180  C, resistance of 5 cm length, silver grid line was around 1 U. 2.2. Preparation of WO3 photoanodes First, nanocrystalline WO3 powder was prepared by mixing 70% (21 g) polyethylene glycol (PEG) with a little water and 30 wt.% (9 g) ammonium metatungstate (AMT) powder. The mixture was grinded using agate mortar for 1 h to make it homogenous. This mixture was calcined at 500  C for 1 h to remove the volatile components, yielding a yellow WO3 powder. After grinding smoothly, 30 wt.% of such WO3 powder was slowly mixed with a blend of 65 wt.% a-terpinol and 5 wt.% ethyl cellulose to obtain a screen printable WO3 paste. A 20 mm thick screen having 200 mesh per inch was used to obtain approximately 5 mm thick WO3 film. Initially small size Experimental Substrates and silver metal grids The conducting fluorine-doped tin oxide (FTO) coated glass (TEC15, thickness: 2.3 mm, w80% transmittance, Rs 10 U sq 1) substrates were used to make photoanodes. Such substrates Fig. 2 e Schematic diagram of (a) small area 0.6 3 0.6 cm2; scaled up (b) 0.6 3 8 cm2; and (c) 9.6 3 13.6 cm2 WO3 photoanodes with Ag grid embedded FTO as substrate; and (d) schematic representation showing coating of WO3 on Ag grid embedded FTO substrate and Ag capsulation using epoxy. i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 (square type) WO3 having an active area of 0.6  0.6 cm2 was coated on 1  1.2 cm2 size FTO substrate. One time screen printing of WO3 paste on to the FTO substrate and subsequent sintering at 500  C for 30 min resulted in w5 mm thick mesoporous layer of WO3. By repeating the screen printing and drying steps 5 times, an optimal film thickness of 25 mm (5 screen printed layers) was accomplished gradually. The optimization was based on the photocurrent performance of the different photoanodes and is described in the results and discussion section. Screen printing was performed in a cleanroom environment to avoid any sources of contamination on fresh films. We had however difficulty in coating multiple layers of WO3 between Ag grid lines embedded on FTO substrates using screen printing. Therefore, to have a good understanding and comparison of the results on PEC hydrogen generation, the scale-up of WO3 electrodes on Ag grid embedded FTO substrates was carried out with 5 mm thick WO3 films. The 5 mm thick WO3 layers with active area of 0.6  0.8 cm2 and 9.6  13.6 cm2 were coated on FTO substrates of size 1  10 cm2 and 15  15 cm2, respectively. In case of 15  15 cm2 photoanode, there are 12 stripe type 5 mm WO3 layers of dimension 0.6  13.6 cm2 between Ag grid lines embedded on FTO substrates giving rise a total active area of 130.56 cm2. Note that the WO3 films are not touching the Ag grid lines. Ag current collecting grids were further encapsulated using non-conductive and non-corrosive epoxy resin in order to protect them from the electrolyte. Providing adequate protection to the Ag grid is an important issue considering the mass production and long-term stability of the water splitting system. A scheme of the grid design and dimensions of WO3 photoanodes toward their scale-up process is shown in Fig. 2 (aed). In addition, a platinum wire (0.1 cm thick, w8 cm long) was used as a counter electrode for the hydrogen production. A calibrated cylindrical tube from a syringe (of 3 mL capacity) surrounding the Pt wire was used to quantify the collected hydrogen gas. 2.3. 3. 5265 Results and discussion Fig. 3 shows the X-ray diffraction pattern of screen printed WO3 film sintered at 500  C for 30 min. XRD analysis of characteristic diffraction peaks of WO3 using standard diffraction data (JCPDS 01-072-0677) reveals the polycrystalline nature of tungsten trioxide (WO3, a-phase) with monoclinic crystal structure. Such structure closely resembles to that of prototype structure of ReO3 consisting of WO6 octahedra linked together at the corners to produce a highly symmetrical threedimensional network with cubic symmetry [28]. The two characteristic triplets of WO3 consisting (002), (020), (200) and (022), ( 202), (202) reflections are seen in the diffraction angle range of 22.5e24.7 and 33e34.5 , respectively. The faces (200), (020), and (002) in the monoclinic phase of WO3 are proven experimentally to give best results for photo-oxidation of water [29,30]. Also, the crystallized monoclinic phase of WO3 is advantageous for the H2 production since it is more useful to oxidize the water and organic species under visible light [31e33] and is quite stable at room temperature. Using a wellknown Scherrer equation, average crystallite size of WO3 is calculated to be 10e15 nm suggesting the nanocrystalline nature of the formed material. Fig. 4 shows FESEM images of thick WO3 film coated on FTO substrate. The image reveals nanocrystalline and porous morphology and size of nanocrystalline grains is in the range of 70e90 nm. Inset of Fig. 4 shows the cross-sectional FESEM image of WO3 photoanode at low magnification (1,000X) estimating a film thickness of w25 mm. Number of voids interconnecting the WO3 nanoparticles, ranging in size from 1 to 4 mm, can be seen in the FESEM image. This suggests the presence of highly porous network in WO3. Binding such Characterization and PEC H2 production system The photelectrodes were characterized by using X-ray diffractometer (Phillips, Model 3234) and field emission scanning electron microscopy (FESEM, Hitachi S4800) for structural and surface morphological features of WO3. The solar light PEC water splitting system for H2 production consisted of a WO3 photoanode placed inside a glass container filled with 0.5 M of aqueous H2SO4 electrolyte. Pt wire (0.1 m thick, 8 cm long) assembled within a calibrated cylindrical syringe, served as hydrogen evolving cathode (counter electrode). The saturated calomel electrode (SCE) and/or a Ag/AgCl (saturated with 3 M KCl) electrode were used as reference electrodes. Required external bias voltage for splitting of water was provided by means of a DSSC module. The details about preparation of DSSC module are given elsewhere [24]. The currentevoltage (IeV) characteristics of the PEC cell and DSSC were respectively measured using Perkin Elmer Potentiostat/Galvanostat (Model 2273) and Keithley digital source meter (Model 2400) in both dark and light illumination. Abet Technologies Sun 2000 solar simulator (1000 W Xe source) was used as the source of illumination and the intensity of incident light was fixed to 1 SUN (AM 1.5G, 100 mW cm 2) using a reference solar cell (PVM-259). Fig. 3 e X-ray diffraction pattern of screen printed WO3 film deposited on FTO substrate and sintered at 500  C for 30 min. XRD reveals monoclinic crystal structure as evidenced from comparison with standard diffraction data # 01-072-0677. 5266 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 Fig. 4 e FESEM images of thick WO3 film coated on FTO substrate exhibiting nanocrystalline and porous morphology (Magnification 50,000X). Inset shows the cross-sectional view of WO3 photoanode estimating a film thickness of 25 mm (Magnification 10003). porous morphology with the crystalline structure induces some strain of the dislocation that could enhance the chemical reactivity [34]. The rate of water splitting reaction on the WO3 surface can be closely connected with the density of spaceecharge-separated electron-hole pairs. This can be increased by enhancing electron-trapping active sites on the WO3 surfaces by engineering the microstructure of the WO3. The morphology, surface area, surface acidity, and defect structure are the important factors affecting the water splitting [35]. Upon illumination, the photogenerated minority carriers (holes) in case of compact thin films have a short traversing distance to reach the water/WO3 interface. On the other hand, with increasing film thickness for such films raises majority carrier (e ) diffusion path length to the back contact (TCO) causing recombination of carriers and hence decreasing the photocurrent. Nevertheless, the high photocurrent observed in our case for film thicknesses up to 25 mm result from the large surface area of the inter-linked nanoporous geometry of WO3. The large, well connected voids in WO3 film allow the electrolyte to diffuse freely. Similar kind of behavior is shown by highly porous 22 mm thick TiO2 films [36]. Fig. 5 shows the photocurrent density curves as a function of applied potential for various WO3 photoanodes under simulated 1 SUN illumination in 0.5 M H2SO4 electrolyte using a conventional three-electrode configuration. Fig. 5a shows the J-V curves vs. SCE reference electrode under dark and light illumination recorded for WO3 photoanodes (active area 0.6  0.6 cm2) of different thicknesses. The dark current remained low until the redox potential was reached for oxygen evolution in 0.5 M H2SO4 solution (1.5 V vs. SCE). For 5 mm thick WO3 (Fig. 5a), the onset of photocurrent starts after 0.35 V, which increases slowly and attains a saturation level of 2.28 mA/cm2 at 1.4 V. The plateau photocurrent region (average of 2.28 mA/cm2) extends up to 1.8 V. Photocurrent shoots up gradually to infinity beyond applied voltage of 1.8 V. With increase in film thickness (increasing screen printing layers), the plateau photocurrent increases slightly to 2.42 mA/cm2; however appearing at lower applied potentials (1.2 V) than that in case of 5 mm thick WO3 films. A shoulder appears near the photocurrent onset potential at around 0.25 V vs. SCE. A similar kind of behavior is reported by others [37] in which appearance of such shoulder has been attributed to the tendency of formic acid molecules to be specifically adsorbed on the WO3. Although there is not much increase in plateau photocurrent with increase in film thickness, the onset potential shifts toward lower applied potential. This is advantageous since PEC water splitting to generate hydrogen can be achieved at much lower applied potentials using such WO3. Inset of Fig. 5a shows the variation of photocurrent densities measured at 1.23 V vs. SCE as a function of film thickness (different screen printed layers) for various WO3 photoanodes (0.36  0.36 cm2). It is seen that the photocurrent saturates gradually for 25 mm WO3 films. A similar photocurrent increment was also noticed in thick nanoporous TiO2 with increasing film thickness until it reaches to its maximum optimized value of 22 mm [36]. Hence, photocurrent of a semiconductor photoanode is determined by the crystallinity, the porous structures, the contact between the particles and the thickness of the film. Fig. 5b shows variation of photocurrent density for WO3 photoanode (5 mm thick, 0.6  8 cm2 active area) vs. SCE with and without use of Ag grid embedded FTO substrate (1  10 cm2). In the absence of Ag grid, photocurrent density of 0.64 mA/cm2 is observed at 1.23 V vs. SCE for WO3 photoanode. Using Ag grid embedded FTO substrate, WO3 photoanode shows remarkable enhancement in photocurrent (almost 2-fold increment) illustrating clearly the influence of Ag grid. The noticed photocurrent enhancement can be attributed to the use of inter-connected current collecting Ag grids by virtue of which the photogenerated electrons are easily collected after being transported from the WO3 to the FTO substrate. Such current collecting grids also minimize the voltage drop across the highly resistive TCO substrates, thereby increasing the overall photocurrent. In other words, incorporation of Ag grids on FTO substrate lowers the sheet resistance of overall FTO substrate and provides more effective carrier transfer leading to efficient current collection than that in the case of highly resistant grid-free FTO substrate. Surface plasmon resonance (SPR) is believed to be associated with photocurrent enhancement in semiconductors when noble metals such Ag, Au are incorporated [38]. In the present study, however, SPR phenomenon due to Ag grids is ruled out as Ag grids are encapsulated using non-conducting epoxy in order to protect from corrosion and are nowhere under the influence of light. To see comparatively the effect of up-scaling the photoanode active area on the PEC performance, the current density-voltage curves are recorded (see Fig. 5c) for WO3 photoanodes having different active areas such as 0.6  0.6 cm2 (square type) and 0.6  8 cm2 (stripe type) and 9.6  13.6 cm2 (stripe type WO3 layers in the Ag grid lines with a total active area of 130.56 cm2). It is obvious from figure that the photocurrent density measured at 1.23 V decreases from 2.63 mA/cm2 to 1.5 mA/cm2 to 1.18 mA/cm2 with scale-up of WO3 photoanode active area from 0.36 cm2 to 4.8 cm2 to 130.56 cm2. The observed relative decrease in photocurrent i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 5267 Fig. 5 e IeV characteristics of WO3 photoanodes of different sizes in 0.5 M H2SO4 electrolyte under 1 SUN illumination: (a) Plot of photocurrent density as a function of applied potential for WO3 photoanodes (active area 0.6 3 0.6 cm2) of different thicknesses. Inset shows the variation of photocurrent densities measured at 1.23 V vs. SCE as a function of film thickness (different screen printed layers); (b) Variation of photocurrent density for 0.6 3 8 cm2 WO3 photoanode (5 mm) vs. SCE with and without use of Ag grid embedded on FTO; (c) Effect of scale-up process on photocurrent of WO3 (5 mm thick) on Ag grid embedded FTO substrate. density, while increasing the photoanode active area from 0.36 to 4.8 cm2 (w13 times scale-up) and from 4.8 to 130.56 cm2 (w27 times scale-up) is 43.0% and 21.3%, respectively. Photocurrent density decreased considerably for small change in active area for tiny TiO2 photoanodes [39], in that almost 50% relative decrement in photocurrent density (from 5 to 2.55 mA/cm2) is observed when active area increased from 0.21 cm2 to 0.72 cm2. Such a behavior of photocurrent (decrement with increase in area of electrode) is known to be due to increase in surface states, originating mostly from grain boundaries/surface defects, thereby creating recombination centre for charge carriers [39]. Although, the photocurrent density observed for scale-up WO3 photoanode (130.56 cm2) is relatively lower than that for tiny photoanode (0.36 cm2), such photocurrent can sufficiently split the water to generate hydrogen. This can further be increased by modifying the substrate surface (large surface area) as well modifying the morphology of WO3 material. Ideally, the photocurrent should increase with increase in photoanode area. However, decreased photocurrent with scaling up of photoanode could be due to defect based recombination centers in the photoanode. We assume that scale-up of photoanode might have led to disordered structure or defects. For small area photoanode, due to lower defect density, the photogenerated charge carriers would possess lower recombination rate as compared to the large area photoanode. Fig. 6 shows a load-line analysis (power matching design) for a PEC water splitting system based on scale-up WO3 photoanode (130.56 cm2 active area) and a DSSC (front-side series, 3 cells connected). To see the power matching design for supplying the required bias to the water splitting system, photocurrent-voltage response curves of WO3 photoanode (130.56 cm2 active area) in 0.5 M H2SO4 electrolyte vs. SCE (curves 1 and 2) and power output voltage characteristics of TiO2 based DSSC (curve 3) under dark and 1 SUN illumination are superimposed. A DSSC module fabricated in this study can generate a voltage of about 2.1 V at open-circuit condition. However, when connected to the load (here, PEC cell), this 5268 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 hSTH ¼ Fig. 6 e Load-line analysis (power matching design) for a PEC water splitting system based on scale-up WO3 photoanode (130.56 cm2 active area) and a DSSC module (front-side series, 3 cell connected). Shown are the superimposed IeV response curves of WO3 photoanode in 0.5 M H2SO4 electrolyte vs. SCE (curves 1 and 2) and TiO2 based DSSC (curve 3) under 1 SUN illumination, indicating an operation point of 153 mA at 1.4 V. value will decrease. Crossing of superimposed IeV curves at 1.4 V under 1 SUN illumination indicate an operation point of the coupled system, meaning that a bias of 1.4 V (153 mA) from DSSC will extract maximum photocurrent from PEC cell. This suggests that DSSC can effectively be coupled with water splitting system to supply the required external potential bias to the PEC cell for hydrogen generation. The effective solar-to-hydrogen (STH) conversion efficiency of the PEC water splitting system is of prime importance for the economic evaluation. The STH conversion efficiency is determined using relation (1) [40e42], 1:23  J  100% Ip (1) where J is the current density (mA cm 2), Ip is the incident power intensity (1.5 AM in the present case i.e. 100 mW cm 2). The measured short circuit current densities (J ) such as 2.63 and 1.5 mA cm 2 at 1.23 V vs. Ag/AgCl for small size (0.36 cm2) and single stripe size (4.8 cm2) WO3 photoanodes linearly translates to STH conversion efficiencies of 3.24% and w2%, respectively. The obtained STH efficiencies are quite better since the theoretically maximum possible value for monoclinic WO3 is only 5% [20]. Although, these STH efficiencies are slightly less that the reported one for 1 cm2 WO3 electrode [21], we report the highest ever STH efficiency using scale-up WO3 photoanodes. These efficiencies can further be expected to increase with structural and morphological improvements in the WO3 photocatalyst. To examine the hydrogen generation under solar light, the PEC cell with photoanode area of 130.56 cm2 is biased using a DSSC module. Fig. 7 shows the experimental arrangement of actual PEC set-up involving 130.56 cm2 WO3 photoanode, platinum counter electrode and gas collection arrangement (calibrated cylindrical tube from a syringe) immersed in 0.5 M H2SO4 electrolyte. The volume of hydrogen gas evolved was determined from downward displacement of the electrolyte in the cylindrical tube. The hydrogen production rate was determined for an applied bias from DSSC module under 1 SUN illumination. About 1 mL of H2 gas was generated in 20 s using scale-up of WO3 (130.56 cm2) photoanode. The rate of hydrogen generation was calculated to be 3 mL/min. The rate of hydrogen generation mainly depends on the charge carrier generation and their combination with hydrogen ions at the counter electrode and hence on the photocurrent density in the PEC cell. More work is underway so as to improve the hydrogen generation rate while scaling up the photoanodes. One has to accept that hydrogen generation scheme will not come into practice with tiny electrodes and necessitates large Fig. 7 e Experimental arrangement of actual PEC set-up involving 130.56 cm2 WO3 photoanode, platinum counter electrode and gas collection arrangement (calibrated cylindrical syringe) immersed in 0.5 M H2SO4 electrolyte. 1 mL of H2 was generated in 20 s using scaled up WO3 (130.56 cm2) photoanode. i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 scale synthesis of photoanodes at the cost of decreased performance. Both scale-up of electrodes and conversion efficiency are however important for commercialization. Therefore, a compromise between efficiency and photoanode area should be established. A further study is needed in this respect to simultaneously achieve the required target so as to have efficient realization of photoelectrochemical water splitting. The present work provides a direction for future development of solar powered photocatalysis towards hydrogen generation. [5] [6] [7] [8] 4. Conclusions We have successfully demonstrated the scale-up of screen printed WO3 photoanodes for H2 generation under simulated sunlight. A new idea of embedding Ag grid lines on FTO substrate is introduced for efficient water splitting using WO3. The additional bias required for photoelectrochemical water splitting is fulfilled by using DSSC module. The WO3 coated on Ag grid embedded FTO showed almost two-fold degree of photocurrent density enhancement as compared to that on bare FTO substrate under 1 SUN illumination in 0.5 M H2SO4 electrolyte. The solar-to-hydrogen (STH) conversion efficiencies calculated at 1.23 V for small (0.36 cm2) and large (4.8 cm2) area WO3 photoanodes prepared using Ag grid embedded FTO substrates are 3.24% and w2%, respectively. Although the present STH efficiency in large area photoanode (4.8 cm2) is less than that for tiny electrode (area <1 cm2), an efficiency of w2% is the highest ever reported value using large area WO3 photoanodes. [9] [10] [11] [12] [13] [14] [15] Acknowledgments This work was performed for Hydrogen Energy R&D center, one of the 21st century frontier R&D program, funded by Ministry of Education, Science and Technology of Korea. [16] [17] Appendix A. Supplementary material [18] Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ijhydene.2011.02.013. [19] references [1] Bak T, Nowotny J, Rekas M, Sorrell CC. Photo-electrochemical hydrogen generation from water using solar energy: materials-related aspects. Int J Hydrogen Energy 2002;27(10): 991e1022. [2] Alexander BD, Kulesza PJ, Rutkowska I, Solarska R,  ski J. Metal oxide photoanodes for solar hydrogen Augustyn production. J Mater Chem 2008;18:2298e303. [3] Grätzel M. Mesoscopic solar cells for electricity and hydrogen production from sunlight. Chem Lett 2005;34(1):8e13. [4] Weinhardt L, Blum M, Bar M, Heske C, Cole B, Marsen B, et al. Electronic surface level positions of WO3 thin films for [20] [21] 5269 photoelectrochemical hydrogen production. J Phys Chem C 2008;112(8):3078e82.  ski J. Nanostructured thin-film WO3 Alexander BD, Augustyn photoanodes for solar water and sea-water splitting. In: Vayssieres L, editor. On solar hydrogen and nanotechnology. Chichester, UK: John Wiley & Sons, Ltd; 2010. p. 333e47. Linsebigler AL, Lu G, Yates Jr JT. Photocatalysis on TiO2 surfaces: principles, mechanisms, and selected results. Chem Rev 1995;95(3):735e58. Butler MA. Photoelectrolysis and physical properties of the semiconducting electrode WO2. J Appl Phys 1977;48(5): 1914e20. Honda K, Fujishima A. Electrochemical photolysis of water at a semiconductor electrode. Nature 1972;238(5358):37e8. Chakrapani V, Thangala J, Sunkara MK. WO3 and W2N nanowire arrays for photoelectrochemical hydrogen production. Int J Hydrogen Energy 2009;34(22):9050e9. Asahi R, Morikawa T, Ohwaki T, Aoki K, Taga Y. Visible-light photocatalysis in nitrogen-doped titanium oxides. Science 2001;293(5528):269e71. Khan SUM, Al-Shahny M, Ingler Jr WB. Efficient photochemical water splitting by a chemically modified nTiO2. Science 2002;297(5590):2243e5. Ohno T, Akiyoshi M, Umebayashi T, Asai K, Mitsui T, Matsumura M. Preparation of S-doped TiO2 photocatalysts and their photocatalytic activities under visible light. Appl Catal A 2004;265(1):115e21. Miller EL, Marsen B, Cole B, Lum M. Low-temperature reactively-sputtered tungsten oxide films for solar-powered water splitting applications. Electrochem Solid-State Lett 2006;9(7):G248e50. Kudo A, Omori K, Kato H. 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. J Am Chem Soc 1999;121(49):11459e67. Sayama K, Nomura A, Arai T, Sugita T, Abe R, Yanagida M, et al. Photoelectrochemical decomposition of water into H2 and O2 on porous BiVO4 thin-film electrodes under visible light and significant effect of Ag ion treatment. J Phys Chem 2006;110(23):11352e60. Duret A, Grätzel M. Visible light-induced water oxidation on mesoscopic a-Fe2O3 films made by ultrasonic spray pyrolysis. J Phys Chem 2005;109(36):17184e91. More GK, Prakasam HE, Varghese OK, Shankar K, Grimes CA. Vertically oriented Ti Fe O nanotube array films: toward a useful material architecture for solar spectrum water photoelectrolysis. Nano Lett 2007;7(8):2356e64. Currao A. Photoelectrochemical water splitting. Chimia 2007; 61(12):815e9.  ski J, Calzaferri G, Courvoisier JC, Grätzel M. Augustyn Photoelectrochemical hydrogen production; state of the art with special reference to IEA’s hydrogen program. Proceedings of the world hydrogen energy conference, 11th; 1996, June 23-28; Stuttgart, Germany. In: Veziroglu TN, Winter CJ, Baselt JP, Kreysa G, editors. Hydrogen Energy Progress XI, vol. 3; 1996. p. 2379e87. Chen Z, Jaramillo TF, Deutsch TG, Kleiman-Shwarsctein A, Forman AJ, Gaillard N, et al. Review: accelerating materials development for photoelectrochemical hydrogen production: standards for methods, definitions, and reporting protocols. J Mater Res 2010;25(1):3e16. Arakawa H, Shiraishi C, Tatemoto M, Kishida H, Usui D, Suma A, et al. Solar hydrogen production by tandem cell system composed of metal oxide semiconductor film photoelectrode and dye-sensitized solar cell. 2007, Aug 27; San Diego, CA, USA. Proc. SPIE. In: Jinghua Guo, editor. Solar hydrogen and nanotechnology II, vol. 6650; 2007. p. 665003e12. 5270 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 2 6 2 e5 2 7 0 [22] Miller EL, Gaillard N, Kaneshiro J, DeAngelis A, Garland R. Progress in new semiconductor materials classes for solar photoelectrolysis. Int J Energy Res 2010;34(14):1215e22. [23] Ramasamy E, Lee WJ, Lee DY, Song JS. Portable, parallel grid dye-sensitized solar cell module prepared by screen printing. J Power Sources 2007;165(1):446e9. [24] Lee WJ, Ramasamy E, Lee DY, Song JS. Dye-sensitized solar cells: scale up and currentevoltage characterization. Sol Energy Mater Sol Cells 2007;91(18):1676e80. [25] Lee WJ, Ramasamy E, Lee DY. Effect of electrode geometry on the photovoltaic performance of dye-sensitized solar cells. Sol Energy Mater Sol Cells 2009;93(8):1448e51. [26] Onoda K, Ngamsinlapasathian S, Fujieda T, Yoshikawa S. The superiority of Ti plate as the substrate of dye-sensitized solar cells. Sol Energy Mater Sol Cells 2007;91(13):1176e81. [27] Tulloch GE. Light and energyddye solar cells for the 21st century. J Photochem Photobiol A 2004;164(1e3):209e19. [28] Berger O, Fischer WJ, Melev V. Tungsten-oxide thin films as novel materials with high sensitivity and selectivity to NO2, O3 and H2S. Part I: preparation and microstructural characterization of the tungsten-oxide thin films. J Mater Sci Mater Electron 2004;15(7):463e82.  ski J. [29] Santato C, Odziemkowski M, Ulmann M, Augustyn Crystallographically oriented mesoporous WO3 films: synthesis, characterization, and applications. J Am Chem Soc 2001;123(43):10639e49. [30] Valdés Á, Kroes G-J. First principles study of the photooxidation of water on tungsten trioxide (WO3). J Chem Phys 2009;130(11):114701e9. [31] Bamwenda GR, Arakawa H. The visible light induced photocatalytic activity of tungsten trioxide powders. Appl Catal A 2001;210(1e2):181e91. [32] Solarska R, Santato C, Jorand-Sartoretti C, Ulmann M,  ski J. Photoelectrolytic oxidation of organic species Augustyn at mesoporous tungsten trioxide film electrodes under [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] visible light illumination. J Appl Electrochem 2005;35(7e8): 715e21.  ski J. Photoelectrochemical Santato C, Ulmann M, Augustyn properties of nanostructured tungsten trioxide films. J Phys Chem B 2001;105(5):936e40. Enesca A, Duta A, Schoonman J. Study of photoactivity of tungsten trioxide (WO3) for water splitting. Thin Solid Films 2007;515(16):6371e4. Kwon YT, Song KY, Lee WI, Choi GJ, Doz YR. Photocatalytic behavior of WO3-loaded TiO2 in an oxidation reaction. J Catal 2000;191(1):192e9. Liu M, Snapp NdL, Park H. Water photolysis with a crosslinked titanium dioxide nanowire anode. Chem Sci; 2011. doi: 10.1039/c0sc00321b. Advance article. Monllor-Satoca D, Borja L, Rodes A, Gámez R, Salvador P. Photoelectrochemical behavior of nanostructured WO3 thinfilm electrodes: the oxidation of formic acid. Chem Phys Chem 2006;7(12):2540e51.  ski J. Silver nanoparticle Solarska R, Królikowska A, Augustyn induced photocurrent enhancement at WO3 photoanodes. Angew Chem Int Ed 2010;49(43):7980e3. Mishra PR, Shukla PK, Singh AK, Srivastava ON. Investigation and optimization of nanostructured TiO2 photoelectrode in regard to hydrogen production through photoelectrochemical process. Int J Hydrogen Energy 2003;28 (10):1089e94. Khaselev O, Turner JA. A monolithic photovoltaicphotoelectrochemical device for hydrogen production via water splitting. Science 1998;280(5362):425e7. Miller EL, Rocheleau RE, Deng XM. Design considerations for a hybrid amorphous silicon/photoelectrochemical multijunction cell for hydrogen production. Int J Hydrogen Energy 2003;28(6):615e23. Scaife DE. Oxide semiconductors in photoelectrochemical conversion of solar energy. Sol Energy 1980;25(1):41e54.