Photoelectrochemical water oxidation at FTO|WO3@CuWO4 and FTO|WO3@CuWO4|BiVO4 heterojunction systems: An IMPS analysis

Electrochimica Acta 308 (2019) 317e327

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Photoelectrochemical water oxidation at FTOjWO3@CuWO4 and FTOjWO3@CuWO4jBiVO4 heterojunction systems: An IMPS analysis
rrez a, b, **, Essossimna Djatoubai b, Manuel Rodríguez-Pe
rez c, Ingrid Rodríguez-Gutie Jinzhan Su b, ***, Geonel Rodríguez-Gattorno a, Lionel Vayssieres b, Gerko Oskam a, *
n, 97310, Mexico Department of Applied Physics, CINVESTAV-IPN, Antigua Carretera a Progreso km 6, M
erida, Yucata International Research Center for Renewable Energy (IRCRE), State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi, China c
noma de Campeche, San Francisco de Campeche, Campeche, 24085, Mexico Facultad de Ingeniería, Universidad Auto a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 January 2019 Received in revised form 2 April 2019 Accepted 5 April 2019 Available online 6 April 2019

Small perturbation techniques such as intensity-modulated photocurrent spectroscopy (IMPS) have become an essential tool to unravel the complex, interrelated processes that govern the charge carrier dynamics in photoelectrochemically active materials for solar water splitting. We have fabricated CuWO4-based photoelectrodes by chemical modification of a WO3 nanorod array as a partially sacrificial template. The electrodes have been characterized by photoelectrochemical techniques including IMPS as a function of the annealing temperature, transforming WO3 either partially or completely to CuWO4. The optical properties illustrate the transformation with the absorbance onset moving from the typical onset for WO3 (about 2.5 eV) to that of CuWO4 (about 2.1 eV). In pure CuWO4 photoelectrodes bulk recombination dominates the photoelectrochemical performance, while for FTOjWO3@CuWO4 heterojunction photoelectrodes much larger charge separation and external quantum efficiencies are obtained. In addition, CuWO4 serves as a protective layer for the WO3 material that is not generally stable in neutral aqueous solutions. The FTOjWO3@CuWO4 heterojunction material was further explored as an underlayer substrate, using a thin BiVO4 film as overlayer. The advantages of this configuration include improved light harvesting as BiVO4 is a direct semiconductor with a much larger absorption coefficient than CuWO4, which is characterized by an indirect gap. It is found that the FTOjWO3@CuWO4 underlayer efficiently extracts the photogenerated electrons from the BiVO4 overlayer, hence, the inclusion of a second heterojunction plays an essential role in improving the charge separation and internal quantum efficiency, minimizing both bulk and surface recombination. © 2019 Elsevier Ltd. All rights reserved.

Keywords: Solar water splitting CuWO4-Based heterojunctions BiVO4 Intensity-modulated photocurrent spectroscopy Charge separation efficiency

1. Introduction The ever-increasing energy demand in the world and the negative environmental impact of fossil fuels has made the development of renewable fuels a priority in science and engineering research. Photoelectrochemical (PEC) water splitting is considered a promising method to produce hydrogen using water as raw material [1,2], combining the principles of photovoltaic

* Corresponding author. ** Corresponding author. Department of Applied Physics, CINVESTAV-IPN, Antigua
rida, Yucata
n, 97310, Mexico. Carretera a Progreso km 6, Me *** Corresponding author.
rrez), j.su@ E-mail addresses: [email protected] (I. Rodríguez-Gutie mail.xjtu.edu.cn (J. Su), [email protected] (G. Oskam). https://doi.org/10.1016/j.electacta.2019.04.030 0013-4686/© 2019 Elsevier Ltd. All rights reserved.

devices and electrolyzers in one system [3e5]. In a solar PEC system, a photoactive electrode that functions either as photocathode or photoanode is immersed into an aqueous electrolyte solution and irradiated with sunlight. Photon absorption generates electron-hole pairs and the minority carriers are transferred to the solution to produce either H2 or O2. A stable material with a suitable band gap capable of efficiently sustaining the oxygen (OER) and hydrogen (HER) evolution reactions is colloquially known as the Holy Grail of photoelectrochemistry. For the past few decades, metal oxides have gained increasing attention as inexpensive semiconducting materials for solar water splitting, also related to their inherently better stability; however, in aqueous solution and under illumination, stability remains a critical issue [6,7]. Tungsten trioxide (WO3) [8e10], with a band gap of about

318

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

2.6 eV and a sufficiently positive valence band edge [11], has been proposed as a promising n-type semiconductor material for water oxidation. However, WO3 only absorbs about 12% of the solar spectrum and it presents a relatively low stability in neutral solutions [12]. It has been shown that related tungstate materials possess better stability in electrolyte solutions with pH > 5 [13]. Copper tungstate (CuWO4) is a suitable option due to its favorable band gap of about 2.3 eV, stability at neutral pH, and correct valence band edge position [14,15]. However, although CuWO4 is characterized by a smaller band gap than WO3, the absorption coefficient close to the onset is small; in addition, the photoelectrochemical performance is limited partially due to the presence of empty Cu (3d2x2-y) orbitals that adversely affect the transport properties [16]. In order to improve the transport properties of the material, CuWO4 thin films have been prepared by different techniques including atomic layer deposition (ALD) [17], electrodeposition [14], impregnation [18] and spray pyrolysis [19], and oxygen evolution co-catalysts such as manganese carbodiimide [20], manganese phosphate [10] and gold [21] have been studied. Furthermore, the fabrication of heterojunction electrode configurations with CuO [22], WO3 [16,23,24] and BiVO4 [18,25] have been reported that result in an increase of the electrode photoresponse. Also, the influence of surface states at the CuWO4/electrolyte interface on the charge transfer processes has been investigated [26]. In this work, we report on the photoelectrochemical characterization of a heterojunction system based on a WO3 nanorod substrate partially or completely converted to CuWO4. The influence of the annealing temperature on the nanostructured heterojunction configuration and the influence on the charge carrier dynamics has been analyzed. In addition, the FTOjWO3@CuWO4 heterojunction has been used as a substrate for a thin overlayer of BiVO4, which has a much larger absorption coefficient than CuWO4. The focus of this work is on the application of intensitymodulated photocurrent spectroscopy (IMPS) in order to quantitatively determine the balance between charge transfer and recombination [27e29], and the effect of the heterojunction structure on the charge separation and external quantum efficiencies. 2. Experimental methods 2.1. Synthesis of the WO3 nanorod array WO3 nanorod array films have been synthesized following the methodology developed by Liu et al. [30]: 1 g ammonium paratungstate (H40N10O41W12$xH2O) was heated in the necessary volume of lactic acid until complete dissolution, resulting in a final concentration of 10 wt%. Then, 1 ml of the solution was diluted by a factor 10 with ethanol to obtain the seed layer precursor solution. Prior to deposition, the fluorine-doped tin oxide (FTO) coated glass substrates were cleaned by consecutive ultrasonication for 30 min in acetone, ethanol and deionized water, and then dried under N2 flow at room temperature. The precursor solution was deposited on the substrate via spin-coating at 3000 rpm for 15 s. The coated substrates were annealed at 500  C for 3 h in a furnace to obtain the WO3 seed layer. In order to further grow the WO3 nanorod array, the substrates with the seed layer were placed into a Teflon-lined hydrothermal reactor filled with a solution prepared by mixing 1.25 g of an ammonium paratungstate solution (10 wt%) with 8.75 g lactic acid in 20 g ethanol containing 0.1383 g reduced L-glutathione (C10H17N3O6S). The reactor was placed into a stainless-steel autoclave and heated at 200  C for 4 h. The obtained films were rinsed with ethanol and annealed in air at 550  C for 1 h.

2.2. Conversion of WO3 to CuWO4 Scheme 1 illustrates the fabrication route of the (partial) conversion of WO3 to CuWO4, and the subsequent deposition of a thin BiVO4 overlayer. The WO3 nanorod array was used as sacrificial template by treatment with a 0.05 M Cu(NO3)2 solution in acetic acid: 50 mL of the Cu(NO3)2 solution was drop-casted onto the surface of the WO3 nanorod array and dried in air at room temperature, resulting in the formation of a Cu(NO3)2 film at the interface. The procedure was repeated 3 times to deposit sufficient Cu(NO3)2 to be able to completely convert the WO3 array. Subsequently, the samples were annealed in air at different temperatures (450  C, 550  C, and 650  C), provoking a solid-state reaction between the WO3 substrate and CuO, obtained from Cu(NO3)2 upon heating, thus forming CuWO4. Depending on the thermal treatment, WO3 is either partially or completely converted to CuWO4. Finally, the excess CuO was removed by immersion in 0.5 M HCl for 30 min resulting in a heterojunction system denominated as follows: FTOjWO3@CuWO4(450  C), FTOjWO3@CuWO4(550  C), and FTOjCuWO4(650  C), respectively. 2.3. Preparation of the FTOjWO3@CuWO4jBiVO4 heterojunction In order to prepare FTOjWO3@CuWO4jBiVO4 heterojunction photoelectrodes, first the FTOjWO3@CuWO4 (550  C) heterojunction was synthesized as described in section 2.2. In order to deposit a thin BiVO4 layer on top, a 0.125 M BiVO4 precursor solution was prepared of 0.01 mol Bi(NO3)3,5H2O, 0.01 mol NH4VO3, 10 mL concentrated HNO3, and 30 mL deionized water, which was then mixed with 40 ml of a 50 g/L PVA (polyvinyl alcohol) solution. Subsequently, the BiVO4 precursor solution was deposited by spincoating and annealed in a rapid thermal annealing furnace at 500  C for 120 s. The procedure was repeated 3 times, and the final product was annealed at 500  C for 2 h in air. BiVO4 thin films were also directly deposited onto FTO; in this case, the procedure was repeated 6 times, and the total film thickness was twice that of the film of the heterojunction system. 2.4. Film characterization The crystal structure of the materials was determined by X-ray diffraction (XRD) using a X'pert PRO diffractometer (PANalytical, Cu-Ka irradiation). The surface morphology and film thickness of the photoelectrodes were analyzed using a JEOL JSM7800FE field emission scanning electron microscope and a FEI G2F30 transmission electron microscope (TEM). The optical properties were determined with a Cary 5000 UVeVis spectrometer from Agilent. Both transmittance and diffuse reflectance, using an integrating sphere and BaSO4 as reference, were measured. Because the films prepared in this work were relatively thin and exhibited a significant transmittance, the absorbance was calculated from the transmittance values. Photoelectrochemical measurements were carried out with a CH Instruments 760D potentiostat in an electrochemical cell with a three-electrode configuration, using the fabricated photoelectrodes as working electrode, Ag/AgCl (3 M NaCl) as reference electrode, and a Pt foil as counter electrode. A 0.1 M phosphate buffer (pH 7) was generally used as the electrolyte solution. Simulated sunlight (calibrated at 100 mW cm2 using a reference solar cell) was generated by a 500 W PerfectLight Xe lamp equipped with an AM 1.5G filter. The applied potential was converted into the reversible hydrogen electrode (RHE) potential scale using the following equation: E(RHE) ¼ EAg/AgClþE0Ag/AgCl þ 0.059·pH

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

319

Scheme 1. Schematic representation of the synthesis route corresponding to the preparation of FTOjWO3@CuWO4, FTOj CuWO4, and FTOjWO3@CuWO4jBiVO4 photoelectrodes.

where E0Ag/AgCl is 0.198 V at 25  C. The incident photon-to-charge carrier conversion efficiency (IPCE) measurements were conducted using the same cell configuration. Monochromatic light was generated using a 150 W Zolix Instruments LSP-X150 Xe lamp coupled with a Zolix Instruments Omni-l1805i spectrometer. The IPCE value was obtained by applying an external bias and was calculated according to the following formula:

 IPCE ¼

hc el

  Iph ðlÞ PðlÞ

(1)

where Iph(l) is the wavelength-dependent photocurrent density [A cm2], P(l) is the light intensity in [W cm2], l is the wavelength, and h, c, and e are Planck's constant, the speed of light in vacuum, and the elementary charge, respectively. IMPS measurements were performed using a Metrohm Autolab PGSTAT302 N/FRA2 set-up, using a 455 nm high-intensity blue LED for both the bias light intensity and the sinusoidally modulated light. The modulation frequency ranged from 10,000 Hze0.05 Hz, with a modulation amplitude of about 16% of the base light intensity; the linearity of the response was tested and confirmed using Lissajous plots. Normalization was performed by determining the number density of incident photons with a calibrated silicon photodiode, followed by a calculation of the corresponding maximum photocurrent assuming each incident photon generates one electron-hole pair. 3. Results and discussion 3.1. Characterization of WO3@CuWO4 films After hydrothermal synthesis of the WO3 nanorod array films, the samples presented a blueish color, which has previously been

related to the presence of (NH4)WO3 [30]. After annealing at 550  C, the layers turned to a yellowish color corresponding to stoichiometric WO3. After the conversion step using a Cu(NO3)2 solution, the impregnated films sintered at 550  C and 650  C turned to bright yellow in concordance with previous reports [16,31,32], while for films annealed at 450  C a light yellow coloration was observed. Fig. 1 shows the X-ray diffraction patterns and cross-sectional SEM images of the sintered FTOjWO3 and FTOjWO3@CuWO4 systems at different temperatures. Fig. 1aei illustrates that the WO3 nanorod film sintered at 550  C exhibits the characteristic plane reflections of monoclinic WO3. Note that the intensity of the (002) plane reflection is lower than half of the expected value indicating a considerable preferential orientation along the [001] direction, in accordance with the rod-like morphology, essentially orthogonal to the substrate surface. After impregnation with the Cu(NO3)2 solution followed by annealing at 450  C, weak reflections at 15.2 and 18.8 indicate a small fraction of CuWO4, thus forming a WO3@CuWO4 core-shell type heterojunction; these reflections correspond to the (010) and (100) planes of the triclinic crystal structure of the tungstate phase (JCPDS # 72-0616). After annealing at 550  C, a mixture of WO3 and CuWO4 is clearly present, indicating that the thickness of the CuWO4 film covering the WO3 nanorod array has increased. After annealing at 650  C, the XRD pattern indicates that complete crystallization of CuWO4 has been achieved. According to previous reports, the decomposition of Cu(NO3)2 occurs first at 170  C forming CuO [32,33], which then reacts with WO3 to produce CuWO4 at about 550  C [14]. A cross-sectional SEM image for a film annealed at 450  C is shown in Fig. 1b, where the 1D rod array structure is visible with a 960 nm thickness, suggesting that the morphology of the sacrificial template has not been appreciably changed at this temperature. For the film sintered at 550  C (Fig. 1c), the reduced elongation of the rods results in an apparently more compact film as result of the

320

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

Fig. 1. a) XRD patterns of the WO3 and WO3@CuWO4 films: i) WO3 film annealed at 550  C; ii), iii), iv) WO3@CuWO4 films prepared at annealing temperatures of 450  C, 550  C and 650  C. On the right, cross-sectional SEM images after removal of excess CuO: b) WO3@CuWO4 sintered at 450  C; c) WO3@CuWO4 sintered at 550  C; d) WO3@CuWO4 sintered at 650  C. e) HRTEM image of an individual WO3@CuWO4 nanorod sintered at 550  C.

volume contraction corresponding to the conversion of WO3 to CuWO4: the monoclinic WO3 unit cell has a volume of about 422.8 Å3, while CuWO4 is triclinic with a unit cell and a volume of about 132.1 Å3. Interestingly, the CuWO4 morphology observed is similar to the typical morphology of WO3 sintered at temperatures higher than 600  C, indicating an epitaxial growth mechanism. After sintering at 650  C, the CuWO4 grain size increases and the pinacoidal morphology is easily observable in particles larger than 100 nm. Fig. S1 in the Supporting Information shows the top views of the films, illustrating that the nanorod array configuration determines the morphology of the heterojunction films. The highresolution TEM image in Fig. 1e of particles annealed at 550  C shows a lattice spacing of 0.387 nm, that matches with the (110) inter-planar distance of triclinic CuWO4. However, as the triclinic structure shares several inter-planar distances with WO3, it is difficult to conclude from TEM if the system is phase-pure. The normalized UVeVis absorbance spectra of the WO3 and the CuWO4 array films obtained by annealing at different temperatures are shown in Fig. 2a. The absorbance onset wavelength gradually shifts from 480 nm for the WO3 nanorod array film to about 550 nm for the CuWO4 film obtained at 650  C. Fig. S2a shows the light harvesting efficiency (LHE) as obtained from the absorbance plots, illustrating that the LHE is very similar for the three systems. It can also be observed that LHE is relatively low in the visible region of the solar spectrum for the photoelectrodes prepared in this work, related to the low absorption coefficient of CuWO4 of about 4  103 cm1 at 455 nm [14].

3.2. Photoelectrochemistry and IMPS of WO3@CuWO4 heterojunctions Fig. 3a shows the current density e potential (I-V) curves for the FTOjWO3@CuWO4 photoelectrodes prepared at different annealing temperatures under simulated 1 sun illumination (100 mW cm2) in a phosphate buffer at pH 7. In the dark, the CuWO4 electrodes exhibit an (I-V) curve typical for an n-type semiconductor, with no appreciable anodic current between 0.5 V(RHE) and 1.8 V(RHE), while a forward cathodic current is observed at more negative potentials; at potentials more positive than 1.8 V a dark oxidation current is observed caused by either a breakdown mechanism or oxidation at the partially exposed FTO surface related to the porosity of the films. Under illumination, the anodic photocurrent onset for all films was observed at about 0.7 V(RHE), related to the water oxidation reaction [31,32,34,35]. Fig. S3 in the Supporting Information shows the IPCE plots recorded at 1.42 V(RHE) for the FTOjWO3@CuWO4 photoelectrodes prepared at different annealing temperatures under front-side illumination in the pH 7 buffer solution. In general, it can be concluded that the films prepared at 450  C and 550  C exhibit similar IPCE values, with much smaller values for the sample sintered at 650  C. Hence, although light absorption is very similar for the films annealed at 650  C, as shown in Fig. S2a, the photocurrent is much smaller suggesting that either a lower charge separation efficiency due to bulk recombination or surface recombination significantly affects the collection efficiency of the photoanode.

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

321

Fig. 2. a) Normalized UVeVis absorbance spectra of WO3 annealed at 550  C, and WO3@CuWO4 prepared by annealing at 450  C, 550  C and 650  C; b) Electrochemical Tauc plots for indirect optical transitions; the inset shows a photograph of samples.

Fig. 2b shows the corresponding electrochemical Tauc plots for the indirect optical transitions that dominate the response close to the band gap, using the IPCE instead of the absorption coefficient, thus combining optical and electrochemical properties of the films. The intercept on the energy axis can be interpreted as the effective band gap of the material, which may be different from the optical band gap depending on the optoelectronic properties of the material. For example, for CuBi2O4, the optical band gap is about 1.55 eV, while the effective, electrochemical band gap may be up to about 2.0 eV [36]. For all annealed photoelectrodes, an effective band gap of 2.1 eV is observed at photon energies close to the photocurrent onset energy, which is in good agreement with reported literature values for the optical band gap of CuWO4 [14,31,37]. For the films annealed at 450  C and 550  C, two regions can be observed: at higher photon energy, the curve is steeper and extrapolates to an effective band gap energy of 2.46 eV, which is closer to the optical band gap of WO3 of about 2.5e2.8 eV [8e11], demonstrating the presence of the heterojunction. In the range where absorption by CuWO4 dominates, the IPCE is largest for the films annealed at 450  C, followed by films annealed at 550  C, and the smallest IPCE value is observed for the film annealed at 650  C. The (I-V) curves in Fig. 3 corroborate these results, indicating that the photoelectrodes annealed at 450  C show better performance under both front and back-side illumination than the electrodes annealed at 550  C and 650  C, with a photocurrent of about 0.4 mA/cm2 at 1.5 V(RHE). This correlates with the observation that

Fig. 3. Current densityepotential curves of the FTOjWO3@CuWO4 photoanodes in a 0.1 M phosphate buffer solution at pH 7 measured under 1 sun, AM 1.5G at 100 mW cm2; a) comparison between the (I-V) curves under front-side and back-side illumination. Figure b shows the (I-V) curves for illumination with a 455 nm blue LED light source under back-side illumination. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

the material annealed at 450  C has a significant presence of WO3, which has been reported to exhibit a better charge collection efficiency than CuWO4 [9,38,39]. On the other hand, WO3 has a poor stability at pH 7, and it appears that even a small amount of CuWO4 is capable of preventing surface degradation. The electrodes annealed at 550  C and 650  C suffer from a large photocurrent difference under front and back-side illumination. This suggests that the collection efficiency is smaller for front-side illumination: although light absorption is better from the front side, the carriers are generated further from the FTO contact, and if the collection distance is smaller than the film thickness this results in a loss of photocurrent. The low photocurrent obtained for the samples annealed at 650  C, even under back-side illumination, indicates that the absence of a heterojunction strongly decreases the collection efficiency. Fig. S4 in the Supporting Information shows that the photoelectrochemical performance is very similar in neutral and slightly basic electrolyte solutions (pH 9), illustrating the promising stability of the system. Fig. 3b shows the (I-V) curves obtained under monochromatic illumination using the 455 nm

322

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

blue LED. The (I-V) curves show a better pronounced photocurrent plateau at potentials larger than 1.1 V(RHE), without change of the onset potential. Simulated sunlight contains photons of different energy, which have different penetration depth; if the collection efficiency is a function of applied potential, the (I-V) curve is expected to show a slow increase in photocurrent with increasing potential. As can be observed, the values of the photocurrent obtained are of a similar magnitude, indicating that the number of photons absorbed and electron - hole pairs generated is similar in both conditions. In order to better understand the effect of materials properties on the charge separation and collection dynamics for the different electrodes prepared, a detailed IMPS analysis has been performed as function of applied potential. The photoelectrodes prepared in this work have a complex configuration, and it is somewhat speculative to use a standard semiconductor/electrolyte interface description to interpret IMPS spectra. In addition, knowledge of the precise energetic positions of the conduction band minimum at the contact, heterojunction, and electrolyte solution interface is elusive but very important for the performance of the system; a recent review describes several methods to modify the position of the CBM of oxide semiconductors [40]. However, the processes can be described in a general way without the need of a specific, quantitative band model. Upon photon absorption, a valence band electron is excited to the conduction band, and photogenerated valence band holes may be separated from conduction band electrons across a potential difference within the heterojunction film, related to the energetic mismatch of the conduction and valence bands of the respective materials, or at the surface. At very high frequencies, the (RC) time constant of the photoelectrochemical cell prevents the observation of a modulated photocurrent, however, upon decreasing the frequency, the charging and discharging of the capacitor that represents the photoelectrode, results in a modulated photocurrent in the external circuit. At sufficiently low frequencies, the modulated photocurrent is in phase with the light intensity modulation, and a measure of the charge separation efficiency (CSE) is obtained. In general, photoelectrochemical charge transfer processes corresponding to water oxidation are much slower than charge separation within the photoelectrode, hence, these processes can be observed at lower frequencies. As a consequence, surface recombination, which involves the transport of a majority carrier back across the potential difference within the photoelectrode, may compete with charge transfer. In the IMPS spectrum, a measure of the charge separation efficiency (CSE) can be obtained at medium frequency, at which charge transfer and surface recombination cannot follow the modulation frequency. Upon applying lower modulation frequencies, the balance between charge transfer and recombination can be observed, and at frequencies going to zero, the external DC quantum efficiency (EQE) is obtained. Note that for the correct calculation of the CSE the light harvesting efficiency (LHE) needs to be taken into account: the medium-frequency crossing point on the H0 -axis is divided by LHE to obtain CSE. In a similar manner, the internal quantum efficiency (IQE) can be obtained by dividing EQE by LHE. Hence, the photoelectrochemical properties of the system are described in terms of the CSE and EQE, which may both be a function of the applied potential. Using this generalized approach, IMPS provides key information to design improved photoelectrochemical systems for solar water splitting. For the case where the RC time constant is significantly smaller, and the balance between effective surface recombination and charge transfer rates determines the internal quantum efficiency, the semicircle in the first quadrant can be analyzed using equation (2):

Scheme 2. Schematic representation of the IMPS spectrum for an n-type photoelectrode; the lower loop (in the fourth quadrant) is associated to the cell time constant, while the upper loop (in the first quadrant) is related to the competition between charge transfer and surface recombination.

00

H ¼ H0 þ i H ¼

  ~j jph ktr þ iu ph ¼ ~I j0 ktr þ krec þ iu

(2)

where H0 and H00 are the real and imaginary part of the complex transfer function; ~jph is the modulated photocurrent generated as a consequence of the modulated light intensity, ~I. Furthermore, jph and j0 are the amplitudes of the modulated photocurrent (electrons cm2 s1) and incident photon flux (photons cm2 s1), respectively, and u is the radial modulation frequency. In the IMPS spectrum, the processes described above lead to two semicircles: (i) in the high frequency range determined by the (RC) time constant of the photoelectrode, which appears in the 4th quadrant; and (ii) a semicircle in the first quadrant, corresponding to the balance between surface recombination and charge transfer. The upper loop moves from H ¼ CSE x LHE at medium frequency to H ¼ EQE at low frequency, with the frequency at the maximum given by the sum of the effective charge transfer (ktr) and surface recombination (krec) rate constants. Scheme 2 illustrates the corresponding IMPS spectrum. Note that both rate constants may depend on the applied potential, depending on the detailed impedance balance of the photoelectrode/electrolyte interface [29]. Fig. 4 shows the IMPS spectra as a function of applied potential for the samples annealed at 450  C, 550  C, and 650  C. In general, it can be observed that the crossing point on the H0 -axis moves to larger values for more positive potentials, indicating that the CSE increases with more positive potential. Hence, an internal potential difference is developed that helps charge separation for all three samples. The materials annealed at 450  C and 550  C show similar values for the value of (CSE x LHE), while the material annealed at 650  C shows a (CSE x LHE) value of a factor 10 lower. Taking into consideration that the value of LHE is essentially the same for the three systems (Fig. S2 in Supporting Information), this indicates that charge separation is not achieved efficiently in the CuWO4 photoelectrode, suggesting that bulk recombination prevents efficient separation. On the other hand, for the WO3@CuWO4 heterojunction samples, charge separation is much more efficient, which implies that the charge is separated across the heterojunction, thus decreasing bulk recombination.

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

323

smaller for the samples annealed at 650  C. On the other hand, the effective surface recombination rate constant is significantly smaller for the photoelectrode annealed at 450  C, consisting of the WO3@CuWO4 heterojunction with the largest fraction of WO3. Combined with the observation that the CSE for photoelectodes annealed at 450  C is similar to that for the samples annealed at 550  C, and a factor 10 larger than for the 650  C samples, these results are in agreement with the higher photocurrent observed for the heterojunction sample prepared at 450  C. Fig. 5 also illustrates that the potential dependence of the rate constants does not follow a well-defined trend, illustrating that it is mainly the charge separation efficiency that determines the photoelectrode efficiency rather than the balance of surface recombination and charge transfer. In summary, the IMPS results show that the FTOjWO3@CuWO4 heterojunction obtained by annealing at 450  C has the largest conversion efficiency, related to both a higher charge separation efficiency and a somewhat higher relative transfer efficiency at lower potentials. It can be concluded that although CuWO4 has an attractive band gap of about 2.1 eV, the low absorption coefficient related to the indirect nature of the optical transitions limits the observed photocurrent; in addition, it is necessary to implement a heterojunction configuration in order to achieve a high charge separation efficiency. 3.3. FTOjWO3@CuWO4jBiVO4 heterojunction

Fig. 4. IMPS spectra obtained for FTOjWO3@CuWO4 photoelectrodes annealed at: a) 450  C; b) 550  C; and c) 650  C. The experiments were performed using a 455 nm blue LED as the illumination source, with the modulation frequency ranging from 10,000 Hze0.05 Hz. The electrolyte solution consisted of a 0.1 M phosphate buffer at pH 7. Fig. S5 in the Supporting Information shows a zoom of the spectra in Figure c). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

For all samples, at potentials more negative than 0.6 V(RHE), the IMPS spectra essentially remain in the origin, indicating that charge separation is not achieved, i.e., all generated electron e hole pairs recombine in the bulk of the materials. Between 0.6 and 0.9 V(RHE), the IMPS spectra present two semicircles, the low-frequency semicircle in the first quadrant is associated to the competition between recombination and charge transfer to the electrolyte solution, while the high-frequency semicircle in the fourth quadrant is given by the cell time constant. For potentials more positive than 0.9 V(RHE), the semicircle in the first quadrant is no longer observed, suggesting that charge transfer is more efficient at more positive potentials [9,29,41e45]. Interestingly, for the material annealed at 450  C, at potentials more positive than 1.1 V(RHE) the semicircle in the fourth quadrant appears to be deformed. This is likely related to the significant presence and contribution of both phases, WO3 and CuWO4. At potentials more positive than about 1.3 V(RHE), the IMPS spectra stay essentially the same suggesting that the photocurrent plateau has been reached, which is in good agreement with the (I-V) curves in Fig. 3b. The effective rate constants for charge transfer and surface recombination are shown as a function of the applied potential in Fig. 5; the cell time constant is significantly smaller (i.e. at higher frequency) in the entire potential range allowing for the extraction of the rate constants. The effective charge transfer rate constant is similar for the samples annealed at 450  C and 550  C, but clearly

BiVO4 has been considered as a promising oxide material for solar water oxidation; the band gap of about 2.4 eV [46e49] permits the absorption of a considerable part of the solar spectrum, and the direct transitions translate in a large absorption coefficient of about 105 cm1 at 455 nm [50]. As for most semiconducting oxides, the integration of an oxygen evolution catalyst improves the photoelectrochemical efficiency [51e53], and strategies to further improve the performance consists in the formation of heterojunctions to enhance the charge carrier collection [25,54e56]. The intrinsically much better optical properties allow thin BiVO4 films to efficiently absorb incident light, unlike CuWO4 that has a much smaller absorption coefficient. Based on the results from the previous sections, we prepared an improved photoelectrode, using the FTOjWO3@CuWO4 (550  C) photoelectrode as substrate, and deposited a thin BiVO4 film by spin coating on top of the heterojunction substrate in order to improve the light harvesting efficiency, and to create a second heterojunction to optimize the charge separation efficiency. Fig. 6a and b shows the XRD patterns of the WO3@CuWO4 (550  C)jBiVO4 (100 nm) heterojunction, as compared to a spincoated BiVO4 film of about 200 nm. It can be concluded that the heterojunction material shows a pattern that is a simple sum of the patterns of both components. Fig. 6c shows SEM micrographs (top view and cross-sectional images) of the WO3@CuWO4 (550  C)j BiVO4 heterojunction film. A somewhat porous, thin BiVO4 layer (of about 100 nm) consisting of small and homogeneous particles is observed on the top of WO3@CuWO4 (550  C) film. The corresponding UVeVis absorbance spectra are shown in Fig. 6d, illustrating that the inclusion of the BiVO4 thin film significantly increases the absorbance of the photoelectrode, which translates in a higher LHE (see Fig. S2b). Fig. S6 in the Supporting Information shows the electrochemical Tauc plot for the 200 nm BiVO4 photoelectrode, illustrating that the effective band gap is about 2.56 eV, relatively close to the optical band gap of 2.4 eV. Fig. 7a shows the current density - potential curves for the FTOjWO3@CuWO4 (550  C), FTOjWO3@CuWO4 (550  C)jBiVO4 (100 nm), and FTOjBiVO4 (200 nm) photoelectrodes in a 0.1 M phosphate buffer (pH 7) under blue LED (455 nm) front-side

324

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

Fig. 5. a) Charge transfer (ktr) and, b) surface recombination rate constant (krec), as well as the frequency corresponding to the cell time constant (RC)1 as a function of the applied potential, corresponding to the results shown in Fig. 4.

Fig. 6. XRD patterns of: a) WO3@CuWO4 (550  C)jBiVO4 (100 nm); and b) FTOjBiVO4 (200 nm). The reference patterns correspond to CuWO4 (JCPDS # 72-0616) and BiVO4 (JCPDS #14-0688). Figure c) shows cross-sectional and top view SEM images of the photoelectrode, and d) shows the corresponding absorbance spectra.

illumination. Fig. 7 shows that the onset for the FTOjBiVO4 photoanode is at a potential about 500 mV more positive than for the FTOjWO3@CuWO4 (550  C)jBiVO4 double heterojunction electrode, which is similar as for the FTOjWO3@CuWO4 (550  C) photoelectrode. At about 2.0 V(RHE), the photocurrent is about 1.1 mA/ cm2 for the two photoelectrodes with BiVO4, while the photocurrent is about a factor 10 lower for the FTOjWO3@CuWO4 (550  C)

photoelectrode. To understand the role of each phase in the heterojunction photoelectrodes IMPS measurements were performed. Fig. 8 shows IMPS spectra as a function of the applied potential comparing the FTOjWO3@CuWO4 (550  C)jBiVO4 and FTOjBiVO4 photoelectrodes. For the FTOjBiVO4 photoelectrode, a first quadrant semicircle is observed from 0.6 to 1.5 V(RHE) that ends back in the origin at low

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

Fig. 7. a) Current density e potential curves for the FTOjWO3@CuWO4 (550  C), FTOjWO3@CuWO4 (550  C)jBiVO4 (100 nm), and FTOjBiVO4 (200 nm) photoelectrodes in a 0.1 M phosphate buffer solution at pH 7, illuminated with a 455 nm blue LED source from the electrolyte side of the cell. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

325

crossing point occurs at relatively small values. Upon applying more positive potentials, the low frequency intercept on the H0 -axis slowly shifts to larger values, indicating an increase in the EQE, which is reflected in the (I-V) curve. The IMPS spectra for the FTOjWO3@CuWO4 (550  C)jBiVO4 double heterojunction shows a different tendency. At potentials more negative than 1.0 V(RHE), an upper semicircle is observed that quickly becomes smaller when the applied potential becomes more positive, indicating that the balance between surface recombination and charge transfer is similar as was observed for the FTOjWO3@CuWO4 (550  C) photoelectrode. The value for the CSE increases significantly at potentials more positive than 1.0 V(RHE), indicating the influence of the heterojunctions. These results are highlighted in Fig. 9, where the charge separation efficiency, CSE, and internal quantum efficiency, IQE, are shown versus the applied potential. The value for the CSE is obtained by dividing the crossing point on the H0 -axis at medium frequency by the LHE (see Fig. S2 in the Supporting Information), while the IQE corresponds to the low-frequency intercept, i.e the EQE, divided by LHE. For both the FTOjWO3@CuWO4 (550  C) and the FTOjWO3@CuWO4 (550  C)jBiVO4 heterojunction photoelectrodes, CSE ¼ IQE for potentials more positive than about 1.0 V(RHE), indicating that charge transfer is faster than surface recombination, and only between 0.6 and 1.0 V(RHE) there is a slight influence of surface recombination. For the FTOjBiVO4 electrode CSE > IQE between 1.0 and 1.8 V(RHE), which shows that for this photoelectrode configuration surface recombination losses are important. Note that CSE for the FTOjBiVO4 system is smaller than for the FTOjWO3@CuWO4 (550  C)jBiVO4 double heterojunction, which indicates that although light is efficiently absorbed, bulk recombination prevents the observation of a high value for CSE. Interestingly, also the IQE is much larger for double heterojunction than for the two separate systems, indicating that charge separation across the heterojunction significantly improves collection, not only by improving the CSE but also by increasing the transfer efficiency by lowering the impact of surface recombination. One remaining disadvantage of the FTOjWO3@CuWO4 (550  C)jBiVO4 double heterojunction as compared to the FTOjWO3@CuWO4 (550  C) single heterojunction is that a more positive potential is needed to achieve the better charge separation and external quantum efficiency.

Fig. 8. IMPS spectra of: a) FTOjBiVO4; and b) FTOjWO3@CuWO4 (550  C)jBiVO4 double heterojunction photoelectrodes as a function of applied potential in a 0.1 M phosphate buffer (pH 7) under front-side blue LED illumination. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

frequencies (see Fig. S7 in Supporting Information for a zoom of the IMPS spectra). These results indicate that although charge separation is achieved, surface recombination dominates over hole transfer to the electrolyte solution, in agreement with previous reports [42,57]. It can also be concluded that the CSE for the FTOjBiVO4 photoelectrode is not large e the medium frequency

Fig. 9. Values for the CSE and IQE from the results shown in Fig. 8, using the values for the LHE shown in Fig. S2 in the Supporting Information comparing the FTOjWO3@CuWO4 (550  C)jBiVO4 double heterojunction to the FTOjWO3@CuWO4 (550  C) single heterojunction and the FTOjBiVO4 (200 nm) photoelectrode. Fig. S8 in the Supporting Information shows a zoom for the 0.6 Ve1.0 V(RHE) range.

326

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327

4. Conclusions In this work, WO3@CuWO4 heterojunction and pure CuWO4 films have been synthesized by chemical modification of a WO3 nanorod array as a partially sacrificial template, using a Cu(NO3)2 solution and subsequent annealing. The films annealed at 450  C and 550  C preserve the morphology of the sacrificial template, and XRD indicates that these films are composed of a WO3@CuWO4 heterojunction, where a larger fraction of the material corresponds to CuWO4 for the film obtained at 550  C. At 650  C, the material transforms completely into the CuWO4 phase, accompanied by a change in the grain size and morphology. The optical properties illustrate the transformation with the absorbance onset moving from the typical onset for WO3 to that of CuWO4. The photooxidation of water under front-side and back-side illumination demonstrate that FTOjCuWO4 photoelectrodes annealed at 650  C have a low photocurrent due to a low charge separation efficiency. These results indicate that bulk recombination dominates the photoelectrochemical performance. For the heterojunction photoelectrodes obtained at 450  C and 550  C, the current is higher, with similar charge separation efficiency. Hence, it can be concluded that the heterojunction structure of the FTOjCuWO4@WO3 photoelectrodes is beneficial in order to increase both the charge separation and external quantum efficiency. In addition, it is found that CuWO4 serves as a protective layer for the WO3 material, which is generally not stable at pH 7. The CuWO4@WO3 material obtained at 550  C was further explored as an underlayer substrate, using a thin spin-coated BiVO4 film as an overlayer. The advantages of this configuration include a better light harvesting, as BiVO4 is a direct semiconductor while CuWO4 is characterized by an indirect gap, and the inclusion of a second heterojunction. It is found that a BiVO4 photoelectrode alone suffers from a low charge separation efficiency due to bulk recombination, as well as a lower external quantum efficiency related to surface recombination. These two detrimental effects can be mitigated by using the FTOjWO3@CuWO4jBiVO4 configuration, which results in effective charge separation thus minimizing both recombination mechanisms. Acknowledgements The authors would like to acknowledge Dr. Antonio Zapien (City University of Hong Kong) for organizing and developing the NANOMXCN initiative that promotes collaboration between Mexico and China, and wish to thank him for his efforts to facilitate this collaborative work. IRG is grateful to CONACyT and LENERSE for financial support for a 10-month research placement at IRCRE, Xi'an Jiaotong University, where the fabrication and structural characterization of the heterojunction systems were performed. The authors gratefully acknowledge CONACYT, SENER and CICY for funding through the Renewable Energy Laboratory of South East Mexico (LENERSE; Project 254667; SP-4). The authors also thank the National Natural Science Foundation of China for funding under project number 51888103. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.electacta.2019.04.030. References [1] S. Jeong, J. Song, S. Lee, Photoelectrochemical device designs toward practical solar water splitting: a review on the recent progress of BiVO4 and BiFeO3 photoanodes, Appl. Sci. 8 (2018) 1388.

 _ G. Seniutinas, B. Sebeka, [2] J. Juodkazyte, I. Savickaja, T. Malinauskas, K. Badokas, K. Juodkazis, S. Juodkazis, Solar water splitting: efficiency discussion, Int. J. Hydrogen Energy 41 (2016) 11941e11948. [3] J.W. Ager, M.R. Shaner, K.A. Walczak, I.D. Sharp, S. Ardo, Experimental demonstrations of spontaneous, solar-driven photoelectrochemical water splitting, Energy Environ. Sci. 8 (2015) 2811e2824. [4] S.A. Bonke, M. Wiechen, D.R. MacFarlane, L. Spiccia, Renewable fuels from concentrated solar power: towards practical artificial photosynthesis, Energy Environ. Sci. 8 (2015) 2791e2796. [5] G. Peharz, F. Dimroth, U. Wittstadt, Solar hydrogen production by water splitting with a conversion efficiency of 18%, Int. J. Hydrogen Energy 32 (2007) 3248e3252. [6] P. Lianos, Review of recent trends in photoelectrocatalytic conversion of solar energy to electricity and hydrogen, Appl. Catal. B Environ. 210 (2017) 235e254. [7] B.S. Kalanoor, H. Seo, S.S. Kalanur, Recent developments in photoelectrochemical water-splitting using WO3/BiVO4 heterojunction photoanode: a review, Mater. Sci. Energy Technol. 1 (2018) 49e62. [8] H. Tong, Y. Jiang, Q. Zhang, J. Li, W. Jiang, D. Zhang, N. Li, L. Xia, Enhanced interfacial charge transfer on WO3 photoanode by molecular iridium catalyst, ChemSusChem 10 (2017) 3268e3275.
rez, I. Rodríguez-Gutie
rrez, A. Vega-Poot, R. García-Rodríguez, [9] M. Rodríguez-Pe G. Rodríguez-Gattorno, G. Oskam, Charge transfer and recombination kinetics at WO3 for photoelectrochemical water oxidation, Electrochim. Acta. 258 (2017) 900e908. [10] K.M. Nam, E.A. Cheon, W.J. Shin, A.J. Bard, Improved photoelectrochemical water oxidation by the WO3/CuWO4 composite with a manganese phosphate electrocatalyst, Langmuir 31 (2015) 10897e10903. [11] H. Wang, T. Lindgren, J. He, A. Hagfeldt, S.-E. Lindquist, Photoelectrochemistry of nanostructured WO3 thin film electrodes for water oxidation: mechanism of electron transport, J. Phys. Chem. B. 104 (2000) 5686e5696. [12] M. Anik, T. Cansizoglu, Dissolution kinetics of WO3 in acidic solutions, J. Appl. Electrochem. 36 (2006) 603e608. [13] T. Heumann, N. Stolica, The electrochemical behaviour of tungsten-II. The dissolution of tungsten in NaOH solutions, Electrochim. Acta. 16 (1971) 1635e1646. [14] J.E. Yourey, B.M. Bartlett, Electrochemical deposition and photoelectrochemistry of CuWO4, a promising photoanode for water oxidation, J. Mater. Chem. 21 (2011) 7651e7660. [15] J.E. Yourey, K.J. Pyper, J.B. Kurtz, B.M. Bartlett, Chemical stability of CuWO4 for photoelectrochemical water oxidation, J. Phys. Chem. C. 117 (2013) 8708e8718. [16] J.E. Yourey, J.B. Kurtz, B.M. Bartlett, Water oxidation on a CuWO4-WO3 composite electrode in the presence of [Fe(CN)6]3-: toward solar Z-scheme water splitting at zero bias, J. Phys. Chem. C. 116 (2012) 3200e3205. [17] Y. Gao, O. Zandi, T.W. Hamann, Atomic layer stack deposition-annealing synthesis of CuWO4, J. Mater. Chem. A. 4 (2016) 2826e2830. [18] W. Ye, F. Chen, F. Zhao, N. Han, Y. Li, CuWO4 nanoflake array-based singlejunction and heterojunction photoanodes for photoelectrochemical water oxidation, ACS Appl. Mater. Interfaces. 8 (2016) 9211e9217. [19] P.K. Pandey, N.S. Bhave, R.B. Kharat, Spray deposition process of polycrystalline thin films of CuWO4 and study on its photovoltaic electrochemical properties, Mater. Lett. 59 (2005) 3149e3155. [20] M. Davi, M. Mann, Z. Ma, F. Schrader, A. Drichel, S. Budnyk, A. Rokicinska, P. Kustrowski, R. Dronskowski, A. Slabon, An MnNCN-derived electrocatalyst for CuWO4 photoanodes, Langmuir 34 (2018) 3845e3852. [21] M. Valenti, D. Dolat, G. Biskos, A. Schmidt-Ott, W.A. Smith, Enhancement of the photoelectrochemical performance of CuWO4 thin films for solar water splitting by plasmonic nanoparticle functionalization, J. Phys. Chem. C. 119 (2015) 2096e2104. [22] A. Martínez-García, V.K. Vendra, S. Sunkara, P. Haldankar, J. Jasinski, M.K. Sunkara, Tungsten oxide-coated copper oxide nanowire arrays for enhanced activity and durability with photoelectrochemical water splitting, J. Mater. Chem. A. 1 (2013) 15235. [23] R. Salimi, A.A. Sabbagh Alvani, N. Naseri, S.F. Du, D. Poelman, Visible-enhanced photocatalytic performance of CuWO4/WO3 hetero-structures: incorporation of plasmonic Ag nanostructures, New J. Chem. 42 (2018) 11109e11116. [24] D. Wang, P.S. Bassi, H. Qi, X. Zhao, Gurudayal, L.H. Wong, R. Xu, T. Sritharan, Z. Chen, Improved charge separation in WO3/CuWO4 composite photoanodes for photoelectrochemical water oxidation, Materials (Basel) 9 (2016) 348. [25] S.K. Pilli, T.G. Deutsch, T.E. Furtak, L.D. Brown, J.A. Turner, A.M. Herring, BiVO4/ CuWO4 heterojunction photoanodes for efficient solar driven water oxidation, Phys. Chem. Chem. Phys. 15 (2013) 3273e3278. [26] Y. Gao, T.W. Hamann, Quantitative hole collection for photoelectrochemical water oxidation with CuWO4, Chem. Commun. 53 (2017) 1285e1288. [27] E.A. Ponomarev, L.M. Peter, A comparison of intensity modulated photocurrent spectroscopy and photoelectrochemical impedance spectroscopy in a study of photoelectrochemical hydrogen evolution at p-InP, J. Electroanal. Chem. 397 (1995) 45e52. [28] L.M. Peter, K.G.U. Wijayantha, A.A. Tahir, Kinetics of light-driven oxygen evolution at a-Fe2O3 electrodes, Faraday Discuss 155 (2012) 309e322. [29] E.A. Ponomarev, L.M. Peter, A generalized theory of intensity modulated photocurrent spectroscopy (IMPS), J. Electroanal. Chem. 396 (1995) 219e226. [30] Y. Liu, L. Zhao, J. Su, M. Li, L. Guo, Fabrication and properties of a branched (NH4)x WO3 nanowire array film and a porous WO3 nanorod array film, ACS

I. Rodríguez-Guti
errez et al. / Electrochimica Acta 308 (2019) 317e327 Appl. Mater. Interfaces 7 (2015) 3532e3538. [31] K.J. Pyper, J.E. Yourey, B.M. Bartlett, Reactivity of CuWO4 in photoelectrochemical water oxidation is dictated by a midgap electronic state, J. Phys. Chem. C. 117 (2013) 24726e24732. [32] D. Hu, P. Diao, D. Xu, M. Xia, Y. Gu, Q. Wu, C. Li, S. Yang, Copper(II) tungstate nanoflake array films: sacrificial template synthesis, hydrogen treatment, and their application as photoanodes in solar water splitting, Nanoscale 3 (2016) 5892e5901. [33] J. Zhang, H.W. Richardson, Copper compounds, Ullmann’s Encycl. Ind. Chem. (2006) 1e31. [34] T. Zhu, M.N. Chong, E.S. Chan, J.D. Ocon, Electrochemically-synthesized tungstate nanocomposites g-WO3/CuWO4 and g-WO3/NiWO4 thin films with improved band gap and photoactivity for solar-driven photoelectrochemical water oxidation, J. Alloy. Comp. 762 (2018) 90e97. [35] N. Gaillard, Y. Chang, A. Deangelis, S. Higgins, A. Braun, A nanocomposite photoelectrode made of 2.2 eV band gap copper tungstate (CuWO4) and multi-wall carbon nanotubes for solar-assisted water splitting, Int. J. Hydrogen Energy. 38 (2013) 3166e3176.
rrez, R. García-Rodríguez, M.J. Rodríguez-Pe
rez, A. Vega[36] I.G. Rodríguez-Gutie Poot, G. Rodriguez Gattorno, B.A. Parkinson, G. Oskam, Charge transfer and recombination dynamics at inkjet-printed CuBi2O4, J. Phys. Chem. C. 122 (2018) 27169e27179. [37] C.R. Lhermitte, B.M. Bartlett, Advancing the chemistry of CuWO4 for photoelectrochemical water oxidation, Acc. Chem. Res. 49 (2016) 1121e1129. [38] I. Grigioni, A. Corti, M.V. Dozzi, E. Selli, Photoactivity and stability of WO3/ BiVO4 photoanodes: effects of the contact Electrolyte and of Ni/Fe oxyhydroxide protection, J. Phys. Chem. C. 122 (2018) 13969e13978. [39] S. Sfaelou, L.C. Pop, O. Monfort, V. Dracopoulos, P. Lianos, Mesoporous WO3 photoanodes for hydrogen production by water splitting and PhotoFuelCell operation, Int. J. Hydrogen Energy. 41 (2016) 5902e5907. [40] L. Kavan, Conduction band engineering in semiconducting oxides (TiO2, SnO2): applications in perovskite photovoltaics and beyond, Catal. Today. 328 (2019) 50e56. [41] L.M. Peter, Dynamic aspects of semiconductor photoelectrochemistry, Chem. Rev. 90 (1990) 753e769. [42] C. Zach€ aus, F. Abdi, L.M. Peter, R. van de Krol, Photocurrent of BiVO4 is limited by surface recombination, not surface catalysis, Chem. Sci. 8 (2017) 4e7. [43] L.M. Peter, K.G. Upul Wijayantha, Photoelectrochemical water splitting at semiconductor electrodes: fundamental problems and new perspectives, ChemPhysChem 15 (2014) 1983e1995. [44] H. Cachet, E.M.M. Sutter, Kinetics of water oxidation at TiO2 nanotube Arrays at different pH domains investigated by electrochemical and light-modulated

327

impedance spectroscopy, J. Phys. Chem. C. 119 (2015) 25548e25558. [45] D. Klotz, D.S. Ellis, H. Dotan, A. Rothschild, Empirical in operando analysis of the charge carrier dynamics in hematite photoanodes by PEIS, IMPS and IMVS, Phys. Chem. Chem. Phys. 18 (2016) 23438e23457.
ndez, N. Russo, Recent advances in the BiVO4 photocatalyst [46] K. Tolod, S. Herna for sun-driven water oxidation: top-performing photoanodes and scale-up challenges, Catalysts 7 (2017) 13. [47] H. Li, X. Quan, S. Chen, H. Yu, Ferroelectric-enhanced Z-schematic electron transfer in BiVO4-BiFeO3-CuInS2 for efficient photocatalytic pollutant degradation, Appl. Catal. B Environ. 209 (2017) 591e599. [48] H.W. Jeong, T.H. Jeon, J.S. Jang, W. Choi, H. Park, Strategic modification of BiVO4 for improving photoelectrochemical water oxidation performance, J. Phys. Chem. C. 117 (2013) 9104e9112. [49] F.F. Abdi, T.J. Savenije, M.M. May, B. Dam, R. Van De Krol, The origin of slow carrier transport in BiVO4 thin film photoanodes: a time-resolved microwave conductivity study, J. Phys. Chem. Lett. 4 (2013) 2752e2757.
n-Llado
, M. Hettick, I.D. Sharp, Y. Lin, A. Javey, J.W. Ager, [50] L. Chen, E. Alarco Reactive sputtering of bismuth vanadate photoanodes for solar Water splitting, J. Phys. Chem. C. 117 (2013) 21635e21642. [51] J. Su, C. Liu, D. Liu, M. Li, J. Zhou, Enhanced photoelectrochemical performance of the BiVO4/Zn:BiVO4 homojunction for water oxidation, ChemCatChem (2016) 1e9. [52] R.P. Antony, P.S. Bassi, F.F. Abdi, S.Y. Chiam, Y. Ren, J. Barber, J.S.C. Loo, L.H. Wong, Electrospun Mo-BiVO4 for efficient photoelectrochemical water oxidation: direct evidence of improved hole diffusion length and charge separation, Electrochim. Acta. 211 (2016) 173e182. [53] F.F. Abdi, R. Van De Krol, Nature and light dependence of bulk recombination in Co-Pi-catalyzed BiVO4 photoanodes, J. Phys. Chem. C. 116 (2012) 9398e9404. [54] H. Jung, S.Y. Chae, C. Shin, B.K. Min, O.-S. Joo, Y.J. Hwang, Effect of the Si/TiO2/ BiVO4 heterojunction on the onset potential of photocurrents for solar water oxidation, ACS Appl. Mater. Interfaces. 7 (2015) 5788e5796. [55] J. Xie, C. Guo, P. Yang, X. Wang, D. Liu, C.M. Li, Bi-functional ferroelectric BiFeO3 passivated BiVO4 photoanode for efficient and stable solar water oxidation, Nano Energy 31 (2017) 28e36. [56] S. Xu, D. Fu, K. Song, L. Wang, Z. Yang, W. Yang, H. Hou, One-dimensional WO3/BiVO4 heterojunction photoanodes for efficient photoelectrochemical water splitting, Chem. Eng. J. 349 (2018) 368e375. [57] M. Antuch, P. Millet, A. Iwase, A. Kudo, The role of surface states during photocurrent switching: intensity modulated photocurrent spectroscopy analysis of BiVO4 photoelectrodes, Appl. Catal. B Environ. 237 (2018) 401e408.