Breakthrough in the large area photoanode fabrication process: high concentration precursor solution with solvent mixing and one step spin coating for high PEC performance of BiVO₄

Abstract

Advancements in the efficiency of hydrogen production via photoelectrochemical water splitting have led to a focus on scalability, ease of fabrication, and cost-effectiveness. Yet, challenges such as prolonged processing times and complex procedures persist, demanding innovative breakthroughs. Our study presents a streamlined method for fabricating high-efficiency BiVO₄ films for PEC application, employing one-step spin coating with high-concentration precursor solutions, making it apt for large-scale deployment. Optimizing solvent and precursor ratios led to a notable photocurrent density of 5.03 mA cm⁻² at 1.23 V_RHE. This enhancement could be attributed to an increase in light absorption owing to an increase in the (040) crystal plane and Mie scattering, optimized film thickness, large grain size and decreased surface dangling bonds, resulting in enhanced carrier density and improved carrier transfer and transport. This approach enabled the cost-effective production of large-area BiVO₄ photoanodes, which effectively generated high current in a PEC-PV system through self-driven solar water splitting. Our study highlights a pathway towards commercial-scale solar-powered hydrogen production technologies.

---

Introduction

Photoelectrochemical (PEC) water splitting has attracted intense research interest as a promising technology for sustainable and eco-friendly hydrogen production.^1,2 Semiconductor metal oxides have received considerable attention as photoanode materials due to their low cost, excellent photoelectrochemical stability, optimal band edge position and appropriate bandgaps.^3,4 Bismuth vanadate (BiVO₄), in particular, is distinguished by its superior visible light absorption, advantageous valence band edge conducive to hydrogen production, and a theoretical photocurrent density up to 7.5 mA cm⁻², while maintaining chemical stability in nearly neutral aqueous solutions.^5–7 Despite its potential, BiVO₄ has challenges such as a short hole diffusion length, poor electron transport properties, and rapid electron–hole recombination.^8–10 To address these challenges, methods like nanostructure fabrication,^11–13 facet engineering,^14–16 and the deposition of electrocatalysts^15,16 have been utilized, resulting in significant PEC performance enhancements. Notably, Kim et al. synthesized BiVO₄ with a porous structure formed by small particles using a two-step method including electrodeposition, achieving good PEC performance due to excellent charge transport properties.¹¹ Han et al. demonstrated that BiVO₄ grown in a preferred orientation exhibited superior charge transport and excellent current density.¹⁵ Jeong et al. fabricated multifaceted BiVO₄ through a crystal reconstruction process to enhance charge transport and thereby improve current density.¹⁶ Wang et al. improved efficiency and stability by incorporating more oxygen vacancies into BiVO₄via sulfur oxidation and applying NiFeO_x on the surface for reducing charge recombination.¹⁷ These advancements, however, have been mostly realized in lab-scale, small-area electrodes, which, despite their ease of fabrication and suitability for experimentation, fall short of the large-scale hydrogen production needs.^12,13,18 The demand for large-area photoelectrodes, which are capable of generating more hydrogen at lower costs, remains unmet due to challenges like reduced PEC performance in larger electrodes, attributed to non-uniform metal oxide films over large area, increased ohmic resistance and charge recombination, and the non-linear diffusion of reactants (OH⁻) in the electrolyte, underscoring the need for further research to bridge the gap to commercial application viability.^19,20 To address the challenges associated with large-area photoanodes, several innovative studies have been conducted. Huang et al. demonstrated a 54.32 cm² large and homogeneous BiVO₄ photoanode fabricated through the electrodeposition method with a tilt-electrode-configuration, and the modified thermal-chemical conversion method for achieving uniform film morphology and high photocurrent density.²¹ Ahmet et al. fabricated a 50 cm² large BiVO₄ photoanode using spray pyrolysis, one of the metal–organic deposition (MOD) methods, and achieved high efficiency in a PEC-PV configuration.²² Qayum et al. produced a uniform 25 cm² large BiVO₄ photoanode employing spray coating and in situ combustion methods, demonstrating a notable photocurrent density of 2.05 mA cm⁻² at 1.23 V_RHE and superior photostability.²³ Despite these advances, challenges still remain: electrodeposition methods with a tilt-electrode-configuration, though effective in creating uniform electrodes, rely on toxic materials and demand simplification for better productivity; MOD-derived BiVO₄ necessitates numerous coatings to attain optimal film thickness, prolonging production times; and the spray coating process, complicated by numerous process variables, requires extended spray and drying durations, thereby extending overall production time. Overcoming these challenges necessitates an innovative breakthrough of a simplified process that enables the creation of optimally thick films in one step, thereby cutting down on both time and material expenses. A high-concentration precursor solution is pivotal for this purpose, playing a critical role in the efficient production of large-area electrodes and paving the way for the commercial viability of hydrogen production technologies.

In this study, we present a straightforward approach to fabricate high-efficiency BiVO₄ films for PEC applications, utilizing a one-step spin coating process with high-concentration precursor solutions, facilitating large-scale deployment. The high-concentration precursor solution was prepared by solvent mixing, showing good dissolution of precursors, and used for one-step spin coating, which is a rapid and straightforward process for mass production of large area photoanodes. We optimized the solvent and precursor mixing ratios to enhance the PEC performance of BiVO₄ films. The films' morphology, crystallography, and optical and electrochemical properties were analyzed to investigate the mechanism of their superior PEC performance. Furthermore, the BiVO₄ films were subjected to an optimization process to deposit oxygen evolution cocatalysts, significantly enhancing the efficiency of PEC water splitting. Using this streamlined fabrication process, large-area BiVO₄ photoanodes measuring 144 cm² were prepared cost-effectively and with uniformity. Four such units were assembled to create an ultra large-scale 576 cm² photoanode, and its performance was evaluated under natural sunlight when integrated into a PEC-PV system, demonstrating its capability for efficient hydrogen production.

Results and discussion

BiVO₄ (BVO) photoanodes were fabricated as illustrated in Fig. 1a (for detailed information see the Experimental section) utilizing a high-concentration precursor solution achieved by mixing acetylacetone and acetic acid (collectively referred to as AA) with dimethyl sulfoxide (DMSO) and a one-step spin coating process. While DMSO's application in BVO film precursor solutions has been infrequent, AA is typically used at a lower concentration of ∼100 mM. In this study, we prepared high concentration precursor solutions of 600 mM, adjusting the DMSO contents in the mixed solvent to 0 (AAD0), 30 (AAD30), 50 (AAD50), 70 (AAD70), and 100% (AAD100) as shown in Fig. S1.†

Fig. 1 (a) Schematic showing fabrication of BVO photoanodes using high concentration precursor solution and a one-step spin coating process. (b) Cross-sectional SEM images of BVO photoanodes formed by one-step spin coating with precursor solutions of varying DMSO contents of 0 (AAD0), 30 (AAD30), 50 (AAD50), 70 (AAD70) and 100% (AAD100) and (c) high-resolution TEM images of AAD0 BVO and AAD50 BVO. (d) XRD patterns of AAD0, AAD30, AAD50, AAD70, and AAD100 BVO.

The cross-sectional scanning electron microscope (SEM) image (Fig. 1b) of BVO films showed that they were composed of aggregated nanoparticles. Top view SEM images (Fig. S2†) showed that all BVO photoanodes except for AAD50 BVO appeared as porous films. The particle size was measured by analysing the top view and cross-sectional SEM images (Fig. S3 and S4†). When the DMSO content in the mixed solvent was between 0% and 30%, films were formed with nanoparticles agglomerated and stacked in multiple layers, featuring average heights of 55 and 60 nm. As the DMSO content reached 50% and 70%, the structure of the films transitioned to a single layer of larger nanoparticles instead of multiple layers of small ones. This change led to a decrease in film thickness and the average heights of the nanoparticles dramatically increased to 180 nm and 132 nm, respectively. The width of AAD50 nanoparticles measured from the top view SEM images was 349 nm. For AAD100, where the DMSO content was 100%, both the size and thickness of the nanoparticles decreased, and it can be observed from Fig. S5† that the uniformity of the coating also worsened. The images of the fabricated BVO photoanodes with different solvent mixing ratios showed a uniform yellow BVO film across all except AAD0 and AAD100 in Fig. S5,† where films appeared light yellow or non-uniform yellow, respectively. Specifically, at AAD100, insufficient dissolution of precursors in the solvent led to their removal by a syringe filter during application, resulting in a diluted and uneven coating. After preparing the precursor solutions and aging them over time, it was observed that precipitation occurred more with higher DMSO content as displayed in Fig. S6.† This result shows that the dissolution is best at 50% DMSO content and that increasing DMSO content accelerates precipitation. Consequently, when the DMSO content was high, the precipitates did not pass through a syringe filter, resulting in the formation of thinner films. Furthermore, high-resolution transmission electron microscopy (HRTEM) images showed the difference in surface atomic arrangement between AAD0 and AAD50 (Fig. 1c and S7†). While the measured lattice spacing in AAD50 (0.43 nm) matched the (011) plane of monoclinic BiVO₄,²⁴ AAD0 showed spacings of 0.31 nm and 0.66 nm, corresponding to the (231) plane of Bi₄(V₂O₁₁) and the (004) plane of Bi₂(CO₃)O₂, respectively. X-ray diffraction (XRD) analysis was performed to identify the crystal phases of the BVO film at different contents of DMSO, as shown in Fig. 1d. In all photoanodes except the AAD0 photoanode, the diffraction peaks at 2θ values were observed at 15.15°, 18.70°, 18.96°, 28.92°, and 30.56°, which corresponded to the (020), (110), (011), (121), and (040) planes of monoclinic BiVO₄ (ICDD 01-074-4894), respectively.²⁵ Among them, AAD50 BVO exhibited high crystallinity as observed from the excellent separation and well-defined peaks. In addition, the intensity ratio of the (040) to (121) peaks in AAD50 BVO was the highest compared to that of the other photoanodes (Table S1†). It has been reported that a prominent (040) plane indicates delayed charge recombination and improved charge transport efficiency.^14,26 It can be expected that the crystal structure of AAD50 BVO could contribute to the facilitation of charge transport. Interestingly, for AAD0, diffraction peaks were observed at 10.58°, 28.37°, 31.84°, and 54.63°, corresponding to the (020), (231), (400), and (631) planes of Bi₄(V₂O₁₁), respectively. Also, peaks at 12.93°, 23.83°, 30.28°, and 32.72° were identified, corresponding to the (004), (112), (116), and (020) planes of Bi(CO₃)O₂, respectively.^27,28 Peaks associated with tetragonal, monoclinic, or orthorhombic BiVO₄ were not observed at all. In the case of AAD0, the precursor solution was prepared solely with AA solvent without DMSO. Additionally, the average grain sizes of the BVO film from the precursor solutions of varying DMSO contents and Bi/V molar ratios were calculated by using the Scherrer equation through X-ray diffraction (XRD) as shown in Fig. S8.† AAD50 BVO showed the largest grain size, and it can be expected that this largest grain size could contribute to enhancing charge transport.¹⁶ Typically, when it was made using only AA solvent, the concentration of the precursor was prepared to be as low as 50–100 mM and it has been reported that monoclinic BiVO₄ is formed.²⁹ However, in this study, with a high concentration of 600 mM, it was found that Bi₄(V₂O₁₁) and Bi₂(CO₃)O₂ other than BiVO₄ are synthesized. In the absence of DMSO, the dissolution of VO(acac)₂ was insufficient, resulting in the formation of compounds with structures of Bi₄(V₂O₁₁) and Bi₂(CO₃)O₂, which have a lower vanadium content.

To investigate the effect of DMSO content changes on PEC performance, linear sweep voltammetry (LSV) curves of the photoanode were obtained using a three-electrode system under AM 1.5 G 1-sun illumination (100 mW cm⁻²) with a 0.5 M PBS electrolyte containing Na₂SO₃. As shown in Fig. 2a, it is observed that the photocurrent density of the photoanode increases with increasing DMSO content, reaching its peak at a 50% concentration. However, the photocurrent density decreases when the DMSO content exceeds 50%. AAD50 BVO exhibited a notable photocurrent density of 5.03 mA cm⁻² at 1.23 V_RHE, indicating a 1.32-fold enhancement when compared to AAD100 BVO. Comparing these results with previously reported BVO fabricated through a one-step coating of a high-concentration precursor solution based solely on DMSO (J: 4.8 mA cm⁻² at 1.23 V_RHE),³⁰ it is evident that the AAD50 BVO fabricated in this paper exhibits a higher current density. This result highlights that combining DMSO with AA as a solvent in the preparation of the high concentration precursor solution is an effective strategy. On the other hand, the AAD0 photoanode exhibited a negligible photocurrent density, suggesting an absence of photoactivity due to the formation of different BVO compositions such as Bi₄(V₂O₁₁) and Bi₂(CO₃)O₂.

Fig. 2 PEC properties of BVO photoanodes dependent on DMSO solvent contents in BVO precursor solution. (a) Linear sweep voltammetry (LSV) curves, (b) UV-vis absorbance spectra, (c) incident photon-to-current conversion efficiency (IPCE), (d) Nyquist plots from electrochemical impedance spectroscopy (EIS) data, and (e) Mott–Schottky (M–S) plot. (f) Schematic illustration of the OER at AAD50 BVO and AAD100 BVO under illumination.

Optical absorption spectroscopy, incident photon-to-current efficiency (IPCE), electrochemical impedance spectroscopy (EIS), and Mott–Schottky (M–S) analyses were conducted on the photoanodes to further investigate the results from LSV curves. On examining the optical absorbance, it was observed that the absorbance increased with the DMSO content up to 50%, and after exceeding a 50% concentration of DMSO, the absorbance sequentially decreased (Fig. 2b). It is notable that thinner AAD50 BVO demonstrates more light absorption than thicker AAD30 BVO. The exceptional light absorption of AAD50 BVO can be attributed to the Mie scattering effect and well-defined monoclinic crystalline structure. Mie scattering occurs with particles of a size similar to the wavelength of the incident light, and it can enhance light absorption by maintaining the direction of the incident light along the material. Particles with a size of 380 nm tend to induce stronger Mie scattering in the visible light range. Nanoparticles of the AAD50 BVO film possess a size within the 350–450 nm wavelength range, which is suitable for effectively inducing Mie scattering, thereby increasing light absorption in this range.³¹ In contrast, nanoparticles of AAD30 BVO, which are relatively smaller, are unsuitable for activating this scattering mechanism. Second, the AAD50 BVO film featured a better defined crystal structure than AAD30, which is positive for light absorption.³² IPCE was measured under 1-sun AM 1.5 G illumination at 1.0 V_RHE using 0.5 M PBS with Na₂SO₃ electrolyte. According to Fig. 2c, AAD50 BVO exhibits excellent IPCE values across all wavelengths, showing a maximum IPCE value of 68% at 440 nm. The results calculated for LHE and APCE using optical absorption and IPCE are depicted in Fig. S9 and S10.† AAD50 BVO showed high APCE values, which indicated efficient charge transfer and separation. The kinetics of charge transfer in BVO, as influenced by variations in DMSO content, were investigated using EIS, and the Nyquist plots derived from the EIS data were fitted to the equivalent circuit model depicted in the inset of Fig. 2d. In this model, R_s represents the electrolyte resistance as a series resistance, Q denotes the constant phase element (CPE) at the electrode/electrolyte interface, and R_ct is the resistance associated with charge transfer at the electrode/electrolyte interface. As shown in Fig. 2d, as the DMSO content increases, the size of the semicircle decreases, with AAD50 BVO showing the smallest semicircle. A trend of increasing semicircle size was observed when the DMSO content exceeded 50%. R_s, representing the resistance of the electrolyte, was similar across all photoelectrodes. However, a comparison of R_ct (Table S2†) showed that it was the lowest in the order of AAD50, AAD70, AAD 100, AAD30, and AAD0 BVO, indicating that lower R_ct values correspond to superior kinetics of charge transfer. The cause of the low R_ct was analyzed by examining the surface composition and chemical state of AAD BVO through XPS as shown in Fig. S11.† When DMSO was mixed with AA, dangling oxygen species were present at 532.0 eV, but it was at its lowest intensity in AAD50. Reduction of dangling oxygen species, which exists in an unstable state on the surface, can decrease the energy barrier and facilitate charge transfer through the interface, resulting in reduced electric resistance.^33,34 Additionally, the hydrophilicity of the photoanode was measured by using the contact angle between BVO and the electrolyte solution, which can be seen in Fig. S12.† Among AAD50, 70, and 100 BVO, which formed a single layered film of nanoparticles, AAD50 BVO exhibited the lowest contact angle. This indicates the rapid transfer of holes (h⁺) at the BVO-electrolyte interface.³⁵

To evaluate the carrier density upon solution mixing, Mott–Schottky (M–S) analysis was performed under dark conditions using 0.5 M PBS with Na₂SO₃ electrolyte (Fig. 2e). The flat-band potential and carrier density calculated from the M–S plots are shown in Table S3.† The slopes of the M–S plots for the BVO photoanode were all positive, indicating that BVO is an n-type semiconductor. AAD50 BVO showed the lowest slope, indicating a high carrier density and suggesting efficient electron–hole separation and suppression of recombination.

In summary, when AA and DMSO are mixed for use, this combination of solvents demonstrates an enhanced capability to dissolve a greater amount of precursors. AAD50 BVO, prepared by mixing AA and DMSO solvents in equal proportions (50%), exhibited a more pronounced (040) plane in its crystal structure, higher light absorption, larger grain size, lower charge transfer resistance and higher carrier density due to excellent electron–hole separation. These characteristics contributed to the improved PEC performance of the AAD50 BVO photoelectrode, as schematically illustrated in Fig. 2f. It is confirmed that mixing AA and DMSO, rather than using them individually, is a good strategy for realizing high-performance BVO photoanodes.

To investigate the effect of changes in the Bi/V molar ratio on PEC performance, BVO films were fabricated by adjusting the concentration of VO(acac)₂ while keeping Bi(NO₃)₃ constant. First, we optimized the concentration of Bi(NO₃)₃ in the precursor solution, varying it from 100 to 800 mM, and found that 600 mM yielded the highest photocurrent density (Fig. S13†).

Next, the Bi/V molar ratio was altered to 0.7 (0.7 BVO) and 0.9 (0.9 BVO), where V > Bi, and 1.1 (1.1 BVO), where V < Bi. The results are shown in Fig. 3. The SEM images indicated that the size of the nanoparticles in the BVO films varied with the Bi/V ratio (Fig. 3a). For 0.7 and 0.9 BVO, a single layer was formed and the nanoparticle size was approximately 180 nm, whereas for 1.1 BVO, similar to when the DMSO content is less than 50%, it consists of multiple layers of small nanoparticles about 60 nm in size which aggregated to form the film. The film thickness of all samples was approximately 180 nm regardless of the Bi/V molar ratio. As the Bi/V molar ratio increases, the morphology of BVO resembles that observed with a low DMSO solvent concentration. When DMSO is insufficient, vanadium fails to dissolve adequately, leading to a higher Bi/V ratio. Consequently, as the amount of vanadium diminishes, nanoparticles increasingly bind with Bi at a higher ratio, resulting in small particles to form aggregated structures.

Fig. 3 Results of fabrication of BiVO₄ photoanodes from different precursor solutions with varying Bi/V molar ratios of 0.7, 0.9 and 1.1. (a) SEM images, (b) XRD patterns, (c) LSV curves, (d) UV-vis absorbance spectra, (e) IPCE and (f) Nyquist plots from EIS data.

XRD analysis was performed to investigate the crystalline phase changes in the BVO films (Fig. 3b). In 0.7 and 0.9 BVO, only the monoclinic BiVO₄ phase (ICDD 01-074-4894) was observed. However, in 1.1 BVO, in addition to the monoclinic phase, peaks appeared at 2θ values of 18.28°, 24.38°, 30.51°, and 32.67°, which correspond to the (110), (200), (211), and (121) planes of tetragonal BiVO₄ (ICDD 00-014-0133). This indicates that at a Bi/V molar ratio of 1.1, with a lower concentration of VO(acac)₂, both tetragonal and monoclinic BiVO₄ were formed. According to previous studies, the crystalline phase of BiVO₄ is closely related to the amount of vanadium, and a lower tendency to induce the formation of tetragonal BiVO₄.³⁰

The LSV measurements for each sample were conducted in 0.5 M PBS electrolyte containing 1.0 M Na₂SO₃. As shown in Fig. 3c, when the Bi/V molar ratio was lower at 0.7 and 0.9, the photocurrent density of the photoanode was more than two-times higher compared to that at 1.1. To analyze this result, optical absorbance spectroscopy, IPCE, EIS, and M–S analyses were performed. As indicated in Fig. 3d, a decrease in the Bi/V molar ratio showed a trend of increased light absorption across the entire wavelength range, with 1.1 BVO exhibiting relatively lower light absorption at all wavelengths. The lower light absorption in 1.1 BVO is attributed to the lower surface roughness and lower absorption due to the reduced multi-scattering effect of light (Fig. S14†).

IPCE measurement results (Fig. 3e) demonstrated that 0.9 BVO exhibited superior IPCE values across all wavelengths when compared to photoanodes with other Bi/V ratios, notably achieving a significant improvement over 1.1 BVO at wavelengths above 450 nm, and even surpassing 0.7 BVO despite its greater light absorption.

The surface chemical composition of samples with Bi/V molar ratios of 0.7, 0.9, and 1.1 was analysed using XPS, as shown in Fig. S15.† The XPS analysis revealed the characteristic peak of oxygen vacancies (Ov) at 530.3 eV in all three samples. The intensity changes among all samples were within 10% and oxygen vacancies seem to have a negligible effect on the superior photocurrent density of 0.9 BVO.

Tauc plot analysis, shown in Fig. S16,† was used to measure the bandgaps for 0.7, 0.9, and 1.1 BVO, found to be 2.60 eV, 2.59 eV, and 2.66 eV, respectively. This suggests that the presence of the tetragonal structure in 1.1 BVO contributes to its relatively larger bandgap, a result of the structural distinction between monoclinic and tetragonal phases.

The kinetics of charge transfer of BVO according to the Bi/V ratio was investigated through EIS analysis, where the obtained Nyquist plots were fitted with an equivalent circuit model (Fig. 3f). The R_ct value of 0.9 BVO was lower than that for 0.7 and 1.1 BVO (Table S4†), indicating that 0.9 BVO has an enhanced charge transfer ability compared to 0.7 and 1.1 BVO. The M–S analysis displayed changes in carrier density according to the Bi/V molar ratio, with 0.9 BVO showing the highest carrier density (Fig. S17†). The calculated carrier density, presented in Table S5,† was 2.66 times higher for 0.9 BVO than for 1.1 BVO, indicating better electron–hole separation. These characteristics are considered to be influenced by the crystal phase difference of BiVO₄, consistent with reports that the tetragonal phase of BiVO₄ hinders charge transport and increases carrier recombination.

To promote the oxygen evolution kinetics, NiFeOOH oxygen evolution cocatalysts (OEC) were introduced to AAD50 BVO to make a NiFeOOH/BVO photoanode via the photo assisted LSV method, and the PEC water splitting performance of NiFeOOH/BVO was evaluated in 0.5 M PBS electrolyte without a hole scavenger. The morphological characteristics of NiFeOOH/BVO were observed through SEM analysis (Fig. S18†). NiFeOOH particles, several tens of nanometers in size, were observed on the surface of BVO nanoparticles, and EDX elemental mapping images revealed the particles to be composed of iron (Fe) and nickel (Ni). The attachment of NiFeOOH was optimized by controlling the number of LSV sweeps, with the best PEC performance observed at 30 cycles. The optimized NiFeOOH/BVO achieved a photocurrent density of 4.53 mA cm⁻² at 1.23 V_RHE, indicating a significant improvement in PEC water splitting performance compared to BVO without the OEC (Fig. 4a). Furthermore, the chemical state of NiFeOOH was confirmed via XPS (Fig. S19†) confirming the successful deposition of NiFeOOH on the BVO surface.^35–37 The OER LSV measurement of BVO photoanodes fabricated by adjusting the DMSO contents and Bi/V molar ratio confirmed that AAD50 BVO shows the best PEC performance both in sulfite oxidation and water oxidation (Fig. S20†) To investigate the effect of NiFeOOH on the charge transfer characteristics of the BVO photoanode, the LSV measurements were performed on AAD50 BVO with and without NiFeOOH photoanodes in an electrolyte containing Na₂SO₃ as a hole scavenger, and the charge transfer efficiency curves were obtained as shown in Fig. 4b according to the equation:

η_transfer = J^OER/J^SOR

where J^OER represents the photocurrent densities for the PEC water oxygen evolution reaction (OER), and J^SOR corresponds to the photocurrent densities for the PEC sulfite oxidation reaction (SOR), respectively.

Fig. 4 The PEC performance of the NiFeOOH/BVO photoanode for water oxidation. (a) LSV curves of NiFeOOH/BVO compared with BVO, (b) charge transfer efficiency of NiFeOOH/BVO compared with BVO, (c) ABPE curves, and (d) oxygen and hydrogen evolution of NiFeOOH/BVO photoanodes.

The NiFeOOH/BVO photoanode demonstrated a charge transfer efficiency of 84.6% at 1.23 V_RHE, indicating excellent OER performance (Fig. 4b). Fig. S21† compares the charge separation and transfer efficiencies of BVO photoanodes with varying DMSO solvent contents and Bi/V molar ratios, confirming that AAD50 BVO (Bi/V = 0.9) exhibits superior charge separation efficiency and transfer efficiency at 1.23 V_RHE. The applied bias photon-to-current efficiency (ABPE) was calculated to be 1.44% at 0.65 V_RHE (Fig. 4c). Consequently, the integration of NiFeOOH with the BVO photoanode significantly enhanced PEC performance by facilitating the transfer of photogenerated holes to the interface between the photoanode and the electrolyte.

To confirm the H₂ production and Faraday efficiency (FE), a J–t test of NiFeOOH/BVO was conducted under AM 1.5 G 1-sun illumination at 1.23 V_RHE in a 0.5 M PBS electrolyte and the amounts of H₂ and O₂ generated were measured by gas chromatography. Fig. 4d and S22† show the amounts of H₂ and O₂ generated, and photocurrent density over time, respectively. The molar ratio of H₂ to O₂ was 2 : 1, which matches the stoichiometric ratio of water splitting. The Faraday efficiency calculated at 1.23 V_RHE was between 84 and 97%. Therefore, it was confirmed that the majority of the photogenerated charge was utilized for water splitting.

To demonstrate the effectiveness of the rapid, straightforward, and high-quality fabrication of large-area BiVO₄ photoanodes by the one-step spin coating method using high-concentration solutions proposed in this study, we fabricated large photoanodes having 25 and 144 cm² active area, with substrate areas of 49, and 196 cm², respectively. The maximum substrate size we can fabricate in our lab is currently 196 cm², limited by our spin coater's capacity. We evaluated their PEC performance by comparison with a small photoanode of 0.08 cm² active area, whose substrate size is 6.25 cm². The precursor solution having AA and DMSO solvents mixed at a 50 : 50 ratio was used. The photocurrent density of the large area NiFeOOH/BVO photoanode was measured by the LSV method in PBS electrolyte (0.5 M) under AM 1.5 G illumination. Fig. 5a shows that a uniform and dark yellow color BVO film was formed on large-area FTO glass. As the active area of the photoanode increased from 0.08 cm² to 144 cm², the photocurrent density gradually decreased from 4.53 mA cm⁻² to 1.50 mA cm⁻² at 1.23 V_RHE. This reduction in photocurrent density can be ascribed to an increase in electrical resistance, nonlinear diffusion of the reactants, and increased charge recombination, as previously reported by Yao. et al..²⁰ Nevertheless, the BVO photoanode of 144 cm² active area represents the largest single photoanode ever reported to date and exhibits the highest photocurrent density to the best of our knowledge. To also evaluate the uniformity of PEC performance in large-area photoanodes, photocurrent densities were measured at nine different points under a 9 cm² illumination area on the BVO photoanode of 144 cm² active area. Their average was 2.76 mA cm⁻² and the deviation according to positions was as low as ±3% (Fig. S23†). This high uniformity of PEC performance is attributed to the uniform morphology and thickness of the BVO film over the whole area (Fig. S24†). Also, the long-term stability of the 144 cm² BVO photoanode in PBS electrolyte containing hole scavengers at 1.0 V_RHE was measured as shown in Fig. S25.† Photocurrent density was maintained for 12 h, showing stable performance while it decreased to 65% of the initial performance after another 12 h. The proposed fabrication technique has been confirmed to be highly effective for fabricating large-area photoanodes with uniform and high PEC performance.

Fig. 5 (a) Image showing the BVO photoanodes with different substrate areas of 6.25 cm², 49 cm² and 196 cm². (b) LSV curves of NiFeOOH/BVO of different substrate and active areas of 6.25 cm², 49 cm² and 196 cm², and 0.08 cm², 25 cm², and 144 cm² respectively. PEC measurements were performed for water oxidation under AM 1.5 G illumination with PBS electrolyte without Na₂SO₃.

To evaluate the unassisted water splitting performance of a large-area NiFeOOH/BVO photoanode with an active area of 144 cm², a PEC-PV tandem system was implemented. In this configuration, the I–V curves of the large-area NiFeOOH/BVO photoanode and the PV cell were obtained independently (Fig. 6a). A single PV cell and a double PV cell, where two single PV cells are electrically connected in parallel and placed side-by-side without overlap, displayed short-circuit currents of 194 mA and 403 mA, respectively. Both configurations achieved an identical open-circuit voltage (V_oc) of 1.75 V. The PV-PEC tandem system operates at the points where the I–V curves of the PEC cell and PV cell intersect. These intersection points for the PEC-single PV cell and PEC-double PV cell were at 1.57 V, 158 mA and 1.68 V, 172 mA, respectively.

Fig. 6 (a) I–V curve of the PV solar cell under AM 1.5 G illumination and the LSV curve for water oxidation of the 144 cm² NiFeOOH/BVO photoanode when they were integrated into the PEC-PV tandem system. (b) Current over time for water oxidation by the assembled PEC-PV system, which has a total area of 576 cm², and consists of four NiFeOOH/BVO photoanodes, each measuring 144 cm², under real solar light irradiation of 0.8 sun (measurement on March 1, 2024, at 14:00).

We constructed an ultra large-scale PEC-PV system as shown in Fig. S26† and conducted a field test. The PEC performance of this system was evaluated under actual sunlight irradiation of 0.8 sun while being immersed in 0.5 M PBS electrolyte. The 576 cm² PEC-PV system generated a photocurrent of 448 mA through self-driven solar water splitting, sustaining this performance over 60 min of continuous operation. Vigorous gas bubbling was observed at the Pt electrode, as depicted in Video S1† and an accompanying video. These results highlight the robust and effective photocurrent generation capability of the large-area 576 cm² PEC-PV system, emphasizing its significant potential for environment-friendly solar hydrogen production.

Conclusion

In this study, we achieved a major breakthrough in the fabrication of large-area, high-efficiency BiVO₄ photoanodes by creating a high-concentration precursor solution via solvent mixing and applying an efficient one-step coating technique. The standout PEC performance was observed in BVO films prepared with a 50% precursor solution, using an equal mix of AA and DMSO. This performance can be attributed to the optimized thickness, large grain size, monoclinic structure, and the superiority of specific crystal planes of the BVO film. These factors contribute to enhanced light absorption, improved electrolyte interface, a higher carrier density, and better charge transfer. Additionally, adjusting the bismuth to vanadium molar ratio, particularly to 0.9, further optimized PEC performance.

Large-area photoanodes were fabricated via this simplified and expedited production process, showing extensive surface uniformity and superior PEC performance. When integrated into a PEC-PV tandem system, these photoanodes attained high photocurrent and efficient hydrogen production under natural sunlight. This streamlined technique for developing high performance and sizable BiVO₄ photoanodes holds great promise for enhancing commercial feasibility of solar-powered hydrogen generation, contributing significantly to the development of sustainable energy technologies.

Experimental section

Preparation of high-concentration BiVO₄ precursor solution

Acetylacetone (≥99%, Sigma-Aldrich, USA) and acetic acid (≥99.7%, Sigma-Aldrich, USA) were fixed as a 1 : 0.12 ratio and denoted as AA. Then, various solvents were prepared by adjusting the content of dimethyl sulfoxide (DMSO, 99.9%, Sigma-Aldrich, USA) and AA. The following names were assigned to the DMSO contents of 0, 30, 50, 70, and 100%: DMSO 0%, AA 100% (AAD0); DMSO 30%, AA 70% (AAD30); DMSO 50%, AA 50% (AAD50); DMSO 70%, AA 30% (AAD70); DMSO 100%, AA 0% (AAD100). In the prepared solvent, bismuth nitrate pentahydrate (Bi(NO₃)₃·5H₂O, 2910 mg, 98%, Sigma-Aldrich, USA) and vanadyl acetylacetonate (VO(acac)₂, 1767 mg, 98%, Sigma-Aldrich, USA) were dissolved in 10 ml of AAD solvent and stirred in a sand bath at 60 °C for 4 h.

Fabrication of the BiVO₄ film

Fluorine-doped tin oxide (FTO) glass substrates (TEC8, Sigma-Aldrich, USA) were prepared into three different sizes (2.5 × 2.5 cm², 7 × 7 cm², and 14 × 14 cm²). The FTO glass substrates were ultrasonically cleaned using acetone, deionized (DI) water, and isopropyl alcohol (IPA) sequentially for 10 min each. After cleaning, the prepared high concentration BiVO₄ precursor solution was dropped on the FTO glass substrate and kept for 1 min. The high concentration BiVO₄ precursor solution on the FTO glass substrate was spin-coated at 2000 rpm for 30 s and baked on a hot plate at 100 °C for 10 min and 200 °C for 10 min. The coated FTO glass substrate was then annealed in a muffle furnace at 450 °C for 2 h.

Deposition of the NiFeOOH co-catalyst on the BiVO₄ film

NiFeOOH oxygen evolution cocatalysts (OECs) were deposited on BiVO₄ films using photo-assisted electrodeposition (PED) under AM 1.5 G illumination. 180 mg of FeSO₄·7H₂O (≥99%, Sigma-Aldrich, USA) and 60 mg of NiSO₄·6H₂O (99%, Sigma-Aldrich, USA) were dissolved in 200 ml of 0.5 M sodium bicarbonate (NaHCO₃) aqueous solution for 2 h, followed by N₂ purging. The prepared BiVO₄ films were plasma treated and immersed in the electrodeposition solution for 5 min under dark conditions. Subsequently, 25 cycles of linear sweep voltammetry (LSV) were conducted at a scan rate of 200 mV s⁻¹ within a bias range of −0.2 V to 0.8 V. Afterwards, a pretreatment was carried out for 5 s at −0.3 V under 1 sun illumination, followed by 25 to 35 cycles of LSV at a scan rate of 50 mV s⁻¹ within a bias range of −0.3 V to 0.5 V.

Configuration of the PV-PEC system

An ultra-large PEC-PV system was prepared for the field test as shown in Fig. S26.† It consists of an ultra-large photoanode, a large bath, PV cells, and a photocathode. To prepare the ultra large BiVO4 photoanode, four large-area BiVO₄ photoanodes, each with an active area of 144 cm² were connected side-by-side without overlap, forming a total area of 576 cm². Each 144 cm² photoanode was electrically connected in parallel and linked to individual PV double cells. The large PMMA bath, equipped with a front quartz window, was designed to fully immerse the 576 cm² photoanode in PBS electrolyte. The photoanode was positioned at the front of the bath (facing the sunlight), with a Pt cathode serving as the counter electrode at the backside of the bath. Four PV double cells were attached outside the bath, behind the Pt electrode. This arrangement allowed light to pass through the photoanode before illuminating the PV cells. The field test of the ultra large-scale PEC-PV system was conducted on March 1, 2024, at 14:00 under 0.8 sun conditions.

Characterization

The morphology of the fabricated samples was investigated using a field-emission scanning electron microscope (FESEM, Hitachi S-4800, Japan). Transmission electron microscopy (TEM) images were obtained with a JEM-ARM200F (JEOL, Japan). Atomic force microscopy (AFM) was performed using an XE-100 (Park system, Korea). The crystallographic structure investigation of the samples was performed using X-ray diffraction (XRD, SmartLab, Rigaku, Japan). Cu Kα radiation (k = 0.15405 nm) was used to obtain diffraction patterns for surfaces ranging from 10° to 70° at a scan rate of 2°/min. The grain size was calculated using the Scherrer equation.
where D represents the grain size, k indicates the Scherrer constant, β is the FWHM in radians, λ denotes the X-ray wavelength and θ represents the peak position.

The optical absorption spectra were analyzed using an ultraviolet-visible-near infrared (UV-vis-NIR) spectrophotometer (Cary 5000, Agilent Inc., USA). The surface chemical state of the photoanode was analyzed by X-ray photoelectron spectroscopy (XPS, K-alpha, UK). All XPS spectra were corrected using the C 1s peak with a binding energy of 284.8 eV, and a fitting curve for deconvolution was determined using a Gaussian peak function.

The PEC performance evaluations were conducted using a potentiostat (BioLogic, VSP 300) equipped with a three-electrode system, using a working electrode, a reference electrode of Ag/AgCl/saturated KCl (3 M) and Pt mesh as the counter electrode, under AM 1.5 G 1-sun illumination (100 mW cm⁻²) generated by a 300 W xenon lamp (Newport, USA). The conversion of the applied potential to the reversible hydrogen electrode (RHE) scale was performed using the following equation:

V_RHE = V_Ag/AgCl + 0.059 × pH + 0.197 V

where V_Ag/AgCl represents the measured potential, and 0.197 V corresponds to the standard potential of the Ag/AgCl electrode at a temperature of 25 °C. Linear sweep voltammetry (LSV) was performed at a scan rate of 10 mV s⁻¹ utilizing back illumination, and an aqueous solution consisting of 0.5 M PBS with or without a hole scavenger of 1.0 M Na₂SO₃ was used as the electrolyte for the oxygen evolution and sulfite oxidation reactions, respectively.

The photocurrent density measured in a water oxidation reaction can be expressed as follows:

J^OER = J_max × η_abs × η_sep × η_trans = J_abs × η_sep × η_trans

where J_max denotes the maximum theoretical photocurrent density, J_abs is the maximum photocurrent density derived from light absorption, η_sep represents the bulk charge separation efficiency, and η_trans indicates charge transfer efficiency.

J _abs can be determined using the following equation.

P_abs(λ) = P₀ × LHE

LHE = 1 − 10^−A

η_sep = J^SOR/J_abs

P _abs represents power of light absorption in photoanodes, λ is the wavelength of this monochromatic light, and ‘A’ is the absorbance at the incident wavelength.

IPCE measurements were performed at 1.0 V vs. RHE using a Cornerstone monochromator (Newport, USA) integrated with a 300 W xenon lamp. The intensity of the monochromatic light was calibrated with a silicon photodiode (SP71650, Newport, USA). The equations for calculating IPCE, LHE, and APCE were as follows.

where j_ph is the photocurrent density and P_mono is the intensity of the incident monochromatic light.

Electrochemical impedance spectroscopy (EIS) was performed in the frequency range of 0.1 Hz to 100 kHz at a signal amplitude of 10 mV and potential of 1.0 V vs. RHE.

Mott–Schottky (M–S) analysis was performed under dark conditions using a 0.5 M PBS electrolyte containing 1.0 M Na₂SO₃, with the voltage set within the range of 0 to 1.2 V vs. RHE at a frequency of 100 kHz. This setup facilitated the calculation of carrier density (N_d) and flat band potential (V_FB) using the Mott–Schottky equation:

where, C represents the capacitance of the space charge region; ε₀ is the permittivity of free space; ε denotes the relative permittivity of BiVO₄ (68); A refers to the active area; e represents the electron charge; ‘V’ is the applied potential; k_B indicates the Boltzmann constant; and T is the temperature.

The equations for calculating the applied bias photon-to-current efficiency (ABPE) were as follows:

where J_ph is the photocurrent density indicated in LSV curves, V_b represents the applied bias vs. RHE (V), and P_total is 100 mW cm⁻².

Hydrogen evolution and oxygen evolution of the photoanode were evaluated within a sealed PEC cell at 1.23 V vs. RHE under AM 1.5 G 1sun illumination conditions. The amount of gases generated during this process was measured at every 10 min using a gas chromatography system (Younglin, YL6500 GC).

Data availability

The data supporting this article have been included as part of the ESI† and Video.

Author contributions

Hoyoung Lee: conceptualization, methodology, investigation, data curation, formal analysis, writing – original draft, writing—review & editing. Gil-Seong Kang: conceptualization, methodology. Hanyi Lim: methodology, formal analysis. Hyobin Han: methodology, formal analysis. Tae Woo Kim: conceptualization, investigation, validation, funding acquisition. Jun-Hyuk Choi: investigation. Dae-Geun Choi: investigation. Joo-Yun Jung: investigation, funding acquisition. Jong Hyeok Park: writing—review & editing, validation, supervision, project administration. Jihye Lee: writing—review & editing, validation, supervision, project administration, funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by The Basic Research Program of the Korea Institute of Machinery and Materials (NK248B), the National Research Council of Science & Technology (NST) grant funded by the Korean Government (MSIT) (No. CAP20034-200), the Technology Innovation Program (No. 20019400) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea), the project for Industry-University-Research Institute Platform cooperation R&D funded by the Korea Ministry of SMEs and Startups in 2022 (S3311782), and the Alchemist Project grant funded by the Korea Evaluation Institute of Industrial Technology (KEIT) & the Korea Government (MOTIE) (Project Number: 1415179744 and 20019169).

Notes and references

1. M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q. Mi, E. A. Santori and N. S. Lewis, Chem. Rev., 2010, 11, 6446.
2. J. Qi, W. Zhang and R. Cao, Adv. Energy Mater., 2018, 8, 1701620.
3. C. Xia, H. Wang, J. K. Kim and J. Wang, Adv. Funct. Mater., 2021, 31, 2008247.
4. Y. Yang, S. Niu, D. Han, T. Liu, G. Wang and Y. Li, Adv. Energy Mater., 2017, 7, 1700555.
5. S. Feng, T. Wang, B. Liu, C. Hu, L. Li, Z.-J. Zhao and J. Gong, Angew. Chem., Int. Ed., 2020, 59, 2044.
6. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520.
7. J. H. Kim and J. S. Lee, Adv. Mater., 2019, 31, 1806938.
8. H. L. Tan, R. Amal and Y. H. Ng, J. Mater. Chem. A, 2017, 5, 16498.
9. J. A. Seabold, K. Zhu and N. R. Neale, Phys. Chem. Chem. Phys., 2014, 16, 1121.
10. Y. Park, K. J. Mc Donald and K.-S. Choi, Chem. Soc. Rev., 2013, 42, 2321.
11. T. W. Kim and K.-S. Choi, Science, 2014, 343, 990.
12. K. J. McDonald and K.-S. Choi, Energy Environ. Sci., 2012, 5, 8553.
13. Y. Kuang, Q. Jia, H. Nishiyama, T. Yamada, A. Kudo and K. Domen, Adv. Energy Mater., 2016, 6, 1501645.
14. Y. J. Jeong, S. W. Hwan, S. Chaikasetsin, H. S. Han and I. S. Cho, Chem. Eng. J., 2022, 435, 135183.
15. H. S. Han, S. Shin, D. H. Kim, I. J. Park, J. S. Kim, P.-S. Huang, J.-K. Lee, I. S. Cho and X. Zheng, Energy Environ. Sci., 2018, 11, 1299.
16. Y. J. Jeong, D. H. Seo, J. h. Baek, M. J. Kang, B. N. Kim, S. k. Kim, X. Zheng and I. S. Cho, Adv. Funct. Mater., 2022, 32, 2208196.
17. S. Wang, P. Chen, Y. Bai, J.-H. Yun, G. Liu and L. Wang, Adv. Mater., 2018, 30, 1800486.
18. F. F. Abdi, L. Han, A. H. Smets, M. Zeman, B. Dam and R. van de Krol, Nat. Commun., 2013, 4, 2195.
19. S. Hernμndez, G. Gerardi, K. Bejtka, A. Fina and N. Russo, Appl. Catal., B, 2016, 190, 66.
20. X. Yao, D. Wang, X. Zhao, S. Ma, P. S. Bassi, G. Yang, W. Chen, Z. Chen and T. Sritharan, Energy Technol., 2018, 6, 100.
21. M. Huang, W. Lei, M. Wang, S. Zhao, C. Li, M. Wang and H. Zhu, J. Mater. Chem. A, 2020, 8, 3845.
22. I. Y. Ahmet, Y. Ma, J.-W. Jang, T. Henschel, B. Stannowski, T. Lopes, A. Vilanova, A. Mendes, F. F. Abdi and R. van de Krol, Sustainable Energy Fuels, 2019, 3, 2366.
23. A. Qayum, M. Guo, J. Wei, S. Dong, X. Jiao, D. Chen and T. Wang, J. Mater. Chem. A, 2020, 8, 10989.
24. C. Li, M. Chen, Y. Xie, J. Jian, H. Wang and L. Jia, J. Power Sources, 2022, 528, 231242.
25. Y. Nakabayashi, M. Nishikawa, N. Saito, C. Terashima and A. Fujishima, J. Phys. Chem. C, 2017, 121, 25624.
26. C. W. Kim, Y. S. Son, M. J. Kang, D. Y. Kim and Y. S. Kang, Adv. Energy Mater., 2016, 6, 1501754.
27. G. Mairesse, P. Roussel, R. N. Vannier, M. Anne, G. Nowogrocki and C. Pirovano, Solid State Sci., 2003, 5, 851.
28. C. Greaves and S. K. Blower, Mater. Res. Bull., 1988, 23, 1001.
29. J. K. Kim, X. Shi, M. J. Jeong, J. Park, H. S. Han, S. H. Kim, Y. Guo, T. F. Heinz, S. Fan, C.-L. Lee, J. H. Park and X. Zheng, Adv. Energy Mater., 2018, 8, 1701765.
30. N. Yang, S. Zhang, Y. Xiao, Y. Qi, Y. Bao, P. Xu, S. Jin and F. Zhang, Angew. Chem., 2023, e202308729.
31. C. Lin, C. Dong, S. Kim, Y. Lu, Y. Wang, Z. Yu, Y. Gu, Z. Gu, D. K. Lee, K. Zhang and J. H. Park, Adv. Mater., 2023, 35, 2209955.
32. M. B. Vishlaghi, A. Kahraman and S. Kaya, J. Phys. Chem. C, 2020, 124, 1337.
33. T. Li, J. He, B. PeÇa and C. P. Berlinguette, Angew. Chem., Int. Ed., 2016, 55, 1769.
34. Y. Yang, S. Wan, S. Li, R. Wang, M. Ou, B. Liud and Q. Zhong, J. Mater. Chem. A, 2023, 11, 1756.
35. S. G. Shim, J. Tan, H. Lee, J. Park, J. Yun, Y. S. Park, K. Kim, J. Lee and J. Moon, Chem. Eng. J., 2022, 430, 133061.
36. J. Cui, M. Daboczi, M. Regue, Y.-C. Chin, K. Pagano, J. Zhang, M. A. Isaacs, G. Kerherve, A. Mornto, J. West, S. Gimenez, J.-S. Kim and S. Eslava, Adv. Funct. Mater., 2022, 32, 2207136.
37. D. Chen, Z. Liu and S. Zhang, Appl. Catal., B, 2020, 265, 118580.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta03349c

---