Charge Transport Enhancement in BiVO₄ Photoanode for Efficient Solar Water Oxidation

Abstract

Photoelectrochemical (PEC) water splitting in a pH-neutral electrolyte has attracted more and more attention in the field of sustainable energy. Bismuth vanadate (BiVO₄) is a highly promising photoanode material for PEC water splitting. Additionally, cobaltous phosphate (CoPi) is a material that can be synthesized from Earth’s rich materials and operates stably in pH-neutral conditions. Herein, we propose a strategy to enhance the charge transport ability and improve PEC performance by electrodepositing the in situ synthesis of a CoPi layer on the BiVO₄. With the CoPi co-catalyst, the water oxidation reaction can be accelerated and charge recombination centers are effectively passivated on BiVO₄. The BiVO₄/CoPi photoanode shows a significantly enhanced photocurrent density (J_ph) and applied bias photon-to-current efficiency (ABPE), which are 1.8 and 3.2 times higher than those of a single BiVO₄ layer, respectively. Finally, the FTO/BiVO₄/CoPi photoanode displays a photocurrent density of 1.39 mA cm⁻² at 1.23 V_RHE, an onset potential (V_on) of 0.30 V_RHE, and an ABPE of 0.45%, paving a potential path for future hydrogen evolution by solar-driven water splitting.

1. Introduction

The production of hydrogen through solar-driven water splitting typically stores solar energy in chemical bonds, which is a potential strategy for overcoming shortages of energy and related global environmental problems [1,2,3,4,5]. The characteristics of photoanode materials decide the solar-to-hydrogen (STH) efficiency of a photoelectrochemical (PEC) device, because the kinetic demand of oxygen evolution reaction (OER) in the PEC water-splitting procedure is higher than the hydrogen evolution reaction (HER) [6,7,8]. To date, various neutral electrolytes have been studied to function as PEC photoanodes, such as TiO₂ [9,10], Fe₂O₃ [11], WO₃ [12], and BiVO₄ [13,14,15,16,17,18]. Among them, the monoclinic bismuth vanadate (BiVO₄) has received significant attention because of its favorable band-edge positions, its moderate band gap energy (E_g, 2.4–2.5 eV) [19], low cost, etc. However, the surface charge recombination and severe bulk defect delay the OER kinetics limits PEC performance [20]. BiVO₄ photoanodes have been the subject of numerous studies aimed at enhancing their PEC performance, including doping [21], building heterojunction [22], crystal facet or morphology engineering [23], oxygen vacancies (O_v) introduction, and the surface modification of oxygen evolution catalysts (OECs), etc. [24,25].

The surface modification of oxygen evolution catalysts is a more promising method for enhancing the PEC performance of BiVO₄ photoanodes than its other counterparts, due to the effective passivation of the surface charge recombination centers and improving the interfacial OER kinetics for PEC water oxidation. For instance, J. Hu et al. successfully synthesized iron oxhydroxide (FeOOH) with different crystalline phases (α-, β-, and δ-) through a regulated solvothermal pathway, where the electrocatalytic OER activity of β-FeOOH was highest [26]. Carbon quantum dots (CQDs) are also able to remarkably improve electrocatalytic OER activity owing to the increased O_v density; at the same time, they can form heterojunctions with semiconductors, thereby effectively promoting charge separation and transport [27]. Moreover, L. Wu et al. synthesized the nickel boride (NiB) layer by adjusting the composition of the neutral electrolyte. The borates with B-O bonds become promoters of catalyst activity by accelerating proton-coupled electron transfer and interact with Ni²⁺ ions to inhibit charge recombination on BiVO₄ surface, thereby reporting a J_ph of 6.0 mA cm⁻² at 1.23 V_RHE [14]. Cobaltous phosphate (CoPi) is an effective electrocatalyst for water oxidation and was first reported by G. Nocera and W. Kanan, who also demonstrated that hydrogen phosphate ions are proton acceptors in oxygen production reactions under neutral pH conditions [28]. Moreover, J. Durrant et al. discussed the oxidation degree of the CoPi catalyst on the BiVO₄ photoanode under simulated sunlight irradiation and determined the appropriate degree of catalyst oxidation to drive substantial water oxidation. Additionally, the relative kinetics of water oxidation on the surfaces of electrocatalyst and semiconductor and the kinetics of holes transfer to electrocatalysts were discussed for the first time [18].

In this work, a BiVO₄ thin film was initially fabricated, which is inherently beneficial for charge transport through the adjustment the electrodeposition time of the BiOI precursor film and the dropping of excessive vanadium source solution for annealing in a muffle furnace. Then, the cobaltous phosphate (CoPi) layer was electrodeposited on optimal BiVO₄ electrodes while the composition of neutral electrolyte was regulated by adding cobaltous nitrate and phosphate species. Finally, the FTO/BiVO₄/CoPi photoanode was successfully prepared. Combining the electrocatalytic and photoelectric technologies, the CoPi for OER catalytic activity was investigated in detail and notable results were obtained, e.g., a J_ph of 1.39 mA cm⁻² (at 1.23 V_RHE) and an undoped ABPE of 0.45% under AM 1.5 G illumination.

2. Materials and Methods

2.1. Preparation of BiOI Precursor Film

The BiOI precursor film was electrodeposited on an FTO glass substrate with an effective area of 1.5 cm × 2 cm in a three-electrode system, where a Pt-foil was used as the counter electrode and an Ag/AgCl with saturated KCl solution was used as the reference electrode. The BiOI precursor deposition solution was fabricated by dissolving lactic acid (0.03 M), KI (0.4 M), and Bi(NO₃)₃·5H₂O (0.015 M) in deionized water (100 mL), with 1,4-Benzoquinone (0.046 M) in ethanol (40 mL) solution. The pH of the mixed solution was adjusted to 3.7 by adding 0.1 M nitric acid aqueous solution after stirring for 20 min. Initially, a 60 s deposition was conducted at −0.40 V_Ag/AgCl to prevent the falling off of BiOI films from the surface of the FTO substrate. After the initial deposition, the BiOI film was obtained at a constant voltage of −0.25 V_Ag/AgCl with different deposition durations, then rinsed thoroughly with deionized water and dried in a drying oven.

2.2. Preparation of BiVO₄ Electrode

The vanadium source solution was prepared by dissolving VO(acac)₂ (0.5 M) in dimethylsulfoxide (10 mL). The as-prepared BiOI precursor film was dropped into the superfluous vanadium source solution. The electrode was then shifted to a muffle furnace and annealed for about 12 h. The heating rate was 3 °C/min to 120 °C, 0.67 °C/min to 280 °C, and 1.41 °C/min to 450 °C, and then held at 450 °C for 1 h. All the annealing processes ended with furnace cooling. After annealing, the electrodes were immersed in a 1.0 M NaOH solution for 15 min with gentle stirring to wash off the V₂O₅ on the BiVO₄ surface. The prepared BiVO₄ electrodes were rinsed thoroughly with deionized water and dried in a drying oven.

2.3. Photo-Assisted Electrodeposition of CoPi Cocatalyst

The CoPi cocatalyst was electrodeposited on the BiVO₄ electrode under AM 1.5 G simulated sunlight by using the three-electrode system containing the solutions of NaH₂PO₄ (0.1 M), Na₂HPO₄ (0.1 M), Co(NO₃)₂·6H₂O (0.001 M), and deionized water (100 mL). Similarly, an Ag/AgCl with a saturated KCl solution was used as the reference electrode and a Pt-foil as the counter electrode. The deposition voltage and time were −0.40 V_Ag/AgCl and 90 s, respectively.

2.4. Characterizations

The crystallinity and structure of BiVO₄ films were examined by X-ray diffraction (XRD, Ultima-iv with Cu/K_α radiation). Surface morphologies were observed via scanning electron microscope (SEM, Germany Zeiss SUPRA 55). The transmittance and absorption of BiVO₄ films were measured via a Shimadzu UV-3600 spectrophotometer. Ultraviolet photoelectron spectroscopy (UPS) was performed using a PHI 5000 VersaProbe with an energy value of a He I source with 21.22 eV. The PEC performance was obtained by using the CHI 660E electrochemical workstation under a three-electrode system, where an Ag/AgCl with saturated KCl solution was used as the reference electrode, with Pt-foil as the counter electrode and the BiVO₄ photoanode as the working electrode. Photoelectrochemical impedance spectroscopy (PEIS) was assessed under light conditions, and its frequency ranged from 10⁻¹ Hz to 10⁵ Hz. Mott–Schottky (M-S) measurements were used to calculate the flat band potential (E_fb) of the BiVO₄ films and the analysis of defects in junctions formed at the interface of semiconductor electrolytes.

3. Results and Discussion

BiVO₄ film with different thicknesses can be prepared by adjusting the electrodeposition time of the BiOI precursor film, i.e., the BiOI-1 precursor film deposited at 300 s, BiOI-2 at 330 s, and BiOI-3 at 360 s, respectively, and then dropping excessive vanadium source solution to anneal in muffle furnace. The resulting BiVO₄ films were labeled as BiVO₄-1, BiVO₄-2, and BiVO₄-3, and the corresponding thicknesses were 358.5 nm, 424.3 nm, and 485.6 nm, respectively (Figure S1). The X-ray diffraction (XRD) patterns of BiVO₄ film with three different thicknesses are shown in Figure 1a. The existence of three major diffraction peaks (011), (–121), and (040) and the standard monoclinic BiVO₄ (JCPDS Card No.14-0688) peaks without any extra peaks, confirm the high crystallinity and pureness of the as-fabricated BiVO₄ thin films [1]. Figure 1b shows that the smaller FWHM (full width at half maxima) values of different diffraction peaks demonstrate larger crystal grains in the BiVO₄-2 film. The grain size can also be calculated according to the Scherrer formula [29]:

$$D=\frac{K\lambda }{\beta cos\theta }$$

(1)

where K is the Scherrer constant, λ is the wavelength of the X-ray sources (0.15406 nm), D is crystallite size (nm), β is the FWHM value, and θ is the Bragg angle at peak position. Figure 1c displays the proportion diagram of grain size distribution, and the BiVO₄-2 obviously accounts for a large proportion in areas with a large grain size distribution. In addition, the average grain sizes of BiVO₄-1, BiVO₄-2, and BiVO₄-3 are 15.54, 16.28, and 15.51 nm, respectively. The surface scanning electron microscopy (SEM) images of the BiVO₄-1, BiVO₄-2, and BiVO₄-3 films are shown in Figure 1d–f. It can be seen that Figure 1e depicts larger crystalline grains compared to those of BiVO₄-1 (Figure 1d). The BiVO₄-1 film consists of many small grains with obvious voids on their surfaces, which are not conducive to charge transport, possibly due to the increase in charge recombination centers on the BiVO₄-1 film surface. On the other hand, a quasi-uniform BiVO₄-2 film with large grains can be obtained by increasing the deposition time of the BiOI precursor film, which is directly related to its effective thermodynamic/kinetic growth under a sufficient annealing temperature. However, the BiVO₄-3 film shows a stacked structure and more voids than the BiVO₄-2 film as the deposition time of the precursor film increases (Figure 1f). Therefore, highly compact and quasi-uniform BiVO₄-2 films are better for photo-assisted electrodeposition with a CoPi catalyst.

The energy band gap (E_g) of the BiVO₄ semiconductor was gained from the transmission spectra in the wavelength range of 300 nm to 1500 nm, as shown in Figure 2a. The transmittance value of the BiVO₄-2 film is lower than that of the BiVO₄-1 and BiVO₄-3 films, indicating that it has a higher absorbance. The FTO was used as a substrate, and the E_g value was calculated using the equations below [30]:

$$\alpha =\mathit{ln}\left(1/T\right)/d$$

(2)

$$\alpha hv={C(hv-E_g)}^n$$

(3)

where n is an index equal to 0.5 in a direct band-gap semiconductor, d is the thickness of the BiVO₄ film, T is the transmission, hν is the energy of a photon, α is the absorption coefficient, and C is a constant. The BiVO₄ films with three different thicknesses display a similar E_g value of 2.41 eV (Figure 2b). The energy band information of the BiVO₄ film was calculated via ultraviolet photoelectron spectroscopy (UPS), as we can see in Figure 2c. According to the secondary electron cut-off (SEC) edge and valence band (V_B) position, the work function (Φ) of BiVO₄ was calculated as 5.57 eV [31]; meanwhile, the E_F and E_V are determined as −5.57 eV and −7.23 eV, respectively (Supplementary Materials, Note S1). In combination with its optical band gap (E_g), at a value of 2.41 eV, the conduction band (E_C, vs. vacuum) of the BiVO₄ semiconductor can be obtained, and its value is –4.82 eV.

The photocurrent density potential (J-V) curves of pure BiVO₄ photoanodes with different thicknesses corresponding to the differing BiVO₄/CoPi photoanodes are shown in Figure 2d. It can be seen that the BiVO₄-2 photoanodes are significantly superior to other BiVO₄ photoanodes, indicating that the BiVO₄ film with the fewest charge recombination centers was obtained by adjusting the electrodeposition time of the BiOI precursor film and the annealing temperature. In addition, we can also see that the J_ph and the fill factor of all BiVO₄ photoanodes increased after CoPi catalyst surface modification. Of these, the BiVO₄-2/CoPi photoanode has a maximum J_ph value of 1.39 mA cm⁻². Therefore, BiVO₄-2 with optimal thickness was selected in this work in order to study the effect of CoPi catalyst on OER occurrence and PEC performance. The potential relative to the Ag/AgCl reference electrode (V_Ag/AgCl) can be converted into V_RHE using the Nernst equation [32]:

$$V_{RHE}=V_{\mathrm{A}\mathrm{g}/\mathrm{A}\mathrm{g}\mathrm{C}\mathrm{l}}+0.059\times PH+0.198$$

(4)

while ABPE was obtained from the J-V response of the photoanodes according to the following equation [33]:

$$\mathrm{A}\mathrm{B}\mathrm{P}\mathrm{E}\left(\%\right)=J_{\mathrm{p}\mathrm{h}}\times \left(V_{RHE}-V_{{\mathrm{H}}_2\mathrm{O}/{\mathrm{O}}_2}\right)/P_{sun}\times 100\%$$

(5)

where V_RHE is the potential relative of a relative hydrogen electrode (RHE), $V_{{\mathrm{H}}_2\mathrm{O}/{\mathrm{O}}_2}$ is the oxidation potential for oxygen (1.23 V_RHE), and P_sun is the simulated sunlight intensity (100 mW/cm²). The electrocatalytic OER activities of BiVO₄ and BiVO₄/CoPi electrodes were measured by linear sweep voltammetry (LSV) tests in 0.2 M Na₂HPO₄/NaH₂PO₄ solution (pH = 6.5) under dark conditions. Figure 3a depicts the LSV curves with a 100% iR drop compensation and a 0.1 mV s⁻¹ scanning rate of the OER catalysts. In general, a catalyst’s OER activity is typically evaluated based on its overpotential at a current density of 10 mA cm⁻² [27]. Therefore, the BiVO₄/CoPi displays the overpotential value of 0.99 V_RHE, which is smaller than that of BiVO₄ (1.87 V_RHE) at 10 mA cm⁻² and the Tafel slope of BiVO₄/CoPi (106.0 mV dec⁻¹) is also obviously lower than that of BiVO₄ (265.7 mV dec⁻¹), indicating that the CoPi catalyst is able to drive reactions at lower overpotentials and play an excellent role in BiVO₄ surface modification (Figure 3b,c). Moreover, the chopped J-V curves are relevant to photoanodes in the dark and continuous illumination within a lesser potential range (<1.5 V_RHE) and are shown in Figure 3d. The dark current density is very low (~0 mA cm⁻² at 1.23 V_RHE); hence, under AM 1.5 G simulated sunlight illumination, the photocurrent density of BiVO₄ and BiVO₄/CoPi photoanodes increases sharply, confirming the rapid photo-generated carrier generation, separation, and transport, without the need for excess drive potentials. The J-V curves based on BiVO₄ and BiVO₄/CoPi photoanodes are displayed in Figure 3e. Specifically, the J_ph values at 1.23 V_RHE are determined to be 0.75 mA cm⁻² and 1.39 mA cm⁻² in sequence. Furthermore, Figure 3f exhibits the as-calculated ABPE and the BiVO₄/CoPi photoanode (0.45%) is about ~3 times higher than the BiVO₄ photoanode (0.14%), indicating a simultaneous upgrade in PEC performance.

The charge transport kinetics were explored according to the electrochemical impedance test. Figure 4a,b exhibits the photoelectrochemical impedance spectroscopy (PEIS) characterization of the BiVO₄ and BiVO₄/CoPi photoanodes under illumination. The equivalent circuit is shown in Figure 4a; R_W can be attributed to the resistance of the electrolyte solution, and R_ct and C_sc represent the charge transport resistance and capacitance in the space charge region of the electrode/electrolyte interface. The fitted results are shown in

Table 1. Summary of the PEIS fitted parameters for FTO/BiVO₄ and FTO/BiVO₄/CoPi photoanodes.

Photoanodes	RW (Ω)	Rct (Ω)	Csc (F)
FTO/BiVO4	18.8	690.3	5.03 × 10−4
FTO/BiVO4/CoPi	18.22	137.9	5.20 × 10−4

, the chi-square values of the BiVO₄ and BiVO₄/CoPi photoanodes are both less than 0.02, and we can also see that the raw data and the fitted results match well (Figure 4a). Similar R_W values (18.2–18.8 Ω) indicate the stability of the test environment, and the remarkably smaller R_ct value of the BiVO₄/CoPi photoanode (137.9 Ω) implies a more effective charge transport at the electrode/electrolyte interface. The larger C_sc value of the BiVO₄/CoPi photoanode (5.20 × 10⁻⁴ F) indicates that the ability of the charge collection is strengthened. Generally, the PEIS-derived low-frequency region (10⁻¹~10¹ Hz) normally represents the mass transfer reactions in the electrode/electrolyte interface [34]. As shown in Figure 4b, the reduction of interface impedance (|Z|) also proves that the CoPi catalyst accelerates the mass transfer process. Moreover, the peaks in the Bode diagrams for the BiVO₄ photoanode are located at low frequencies (10⁻¹~10¹ Hz), while the BiVO₄/CoPi photoanode is located between 10¹ and 10² Hz (insert in Figure 4b). The rapid response of the BiVO₄/CoPi photoanode to frequency indicates that charge transport and mass transfer simultaneously accelerated. The Mott–Schottky (M-S) measurement was also introduced in order to study the junction formed at the semiconductor–electrolyte interface in reaction to the applied potential (V_Ag/AgCl, i.e., the aforementioned E_appl). Figure 4c reveals that the 1/C² increases with the potential V_Ag/AgCl in the presence of the space charge region (SCR), indicating n-type properties for the BiVO₄ semiconductor. Moreover, the conduction band position of BiVO₄ semiconductor relative to normal hydrogen electrode (NHE) can be obtained from the flat band potential (E_fb), donor density (N_D), the effective density of states functions in the conduction band (N_c), and the effective mass of the electron ($m_n^{*}$), according to the following equations [35,36]:

$$\frac{1}{C_{sc}^2}=\frac{2}{e{A}^2\epsilon {\epsilon }_0{N}_D}(E_{appl}-E_{fb}-\frac{k_{B}T}{e})$$

(6)

$$E_c(vs.NHE)=E_{fb}+kTln\left(\frac{N_c}{N_D}\right)$$

(7)

$$N_c=2(\frac{2\pi m_n^{*}kT}{h^2}{)}^{\frac{3}{2}}$$

(8)

where C_sc is the SCR capacitance in the semiconductor–electrolyte interface, A is the active device area, ε₀ is the permittivity of vacuum, ε is the relative dielectric coefficient, k_B is Boltzmann constant, e is the unit charge, and T represents the temperature. The E_fb value and N_D value of the BiVO₄ are 0.38 V_NHE and 1.92 × 10²¹ cm⁻³, respectively. At room temperature (T = 25 °C), the vacuum energy level can be converted to V_NHE by the following formula [37,38]:

$$Energy(vs.vacuum)={-eE}_{appl}\left(vs.NHE\right)-4.44\mathrm{e}\mathrm{V}$$

(9)

The calculated E_C value range is –4.82 ± 0.02 eV, using the above equations, which is consistent with the E_C of –4.82 eV obtained from the UPS measurement. According to the E_g value (2.41 eV) calculated by transmission, the valance band (E_V, vs. NHE) is 2.79 V_NHE, which thermodynamically supports the occurrence of water oxidation with oxygen production by solar-driven water splitting. In contrast, the E_fb value and N_D value of the BiVO₄/CoPi quasi-semiconductor are 0.40 V_NHE and 8.87 × 10²⁰ cm⁻³, respectively (Figure 4d). The lower N_D value demonstrates that the CoPi catalyst passivates defects (e.g., charge recombination centers) on the BiVO₄ film surface, reducing electron–hole recombination during charge transport to the electrode/electrolyte interface, thereby improving photocurrent density. We can also clearly observe that the CoPi catalyst effectively reduces the defect pinholes on the BiVO₄ surface through a top-view SEM image of the BiVO₄/CoPi photoanode (Figure S2). The increase in E_fb of 0.02 V_NHE also indicates that the BiVO₄/CoPi photoanode is more beneficial for water oxidation in thermodynamics (Figure S3).

The surface charge transfer efficiency (η_tran) and bulk charge separation efficiency (η_sep) of the BiVO₄ photoanode were further investigated in order to determine the reasons for the significantly improved photocurrent density after surface modification with CoPi catalyst. The integrated photocurrent density (J_abs) of BiVO₄ and BiVO₄/CoPi photoanodes can be obtained according to wavelength-dependent light harvesting efficiency (LHE) and the standard AM 1.5 G solar spectrum, utilizing the following formulas [1]:

$$J_{abs}={\int }_{300}^{{\lambda }_e}\frac{\lambda }{1240}{·N}_{ph}(\lambda )·LHE(\lambda )d\lambda$$

(10)

$$LHE=1-{10}^{-A\left(\lambda \right)}$$

(11)

where J_abs is the integrated photocurrent density, λ_e is the absorption cut-off wavelength that is linked to the band gap, N_ph (λ) is the photo flux, and A (λ) is the wavelength-dependent absorption, covering wavelengths from 350 to 800 nm (Figure 5a). The λ_e values for BiVO₄ and BiVO₄/CoPi were determined to be at 514 nm and 520 nm, suggesting CoPi can also effectively broaden and heighten the LHE range of the BiVO₄ photoanode (Figure 5b), and giving the J_abs values of 6.29 mA cm⁻² and 6.68 mA cm⁻², respectively (Figure 5c). Moreover, the transient photocurrent response spectra of the BiVO₄ and BiVO₄/CoPi photoanodes are shown in Figure 5d, while the η_tran can be obtained through the measured photocurrents associated with the “light off” state and the “light on” state, according to the following formulas [32]:

$${\eta }_{tran}=\frac{J_{ss}}{J_{inst}}$$

(12)

where J_ss is the photocurrent density in steady state and J_inst signifies the instantaneous photocurrent density. Accordingly, the η_tran value of the BiVO₄/CoPi photoanode (82.1%) has increased in comparison to the BiVO₄ photoanode (76.0%), demonstrating that the CoPi catalyst accelerates holes transfer to the electrode/electrolyte interface and then oxidizes water. Moreover, we can observe in Figure 5d that the BiVO₄/CoPi photoanode has a spike peak in the “light off” state compared with the BiVO₄ photoanode. The negative current transient suggests that there is significant back electron/hole recombination after “light off” [39], reiterating that the BiVO₄/CoPi photoanode has a larger capacitance value (Table 1), which demonstrates that the CoPi catalyst can delay charge recombination and promote the OER’s continuous progress. Additionally, the η_sep can be calculated using following equation [32]:

$${\eta }_{sep}=\frac{J_{ph}}{J_{abs}\times {\eta }_{tran}}$$

(13)

The BiVO₄ and BiVO₄/CoPi photoanodes’ calculated η_sep values are 15.7%, and 25.4%, respectively. The obvious improvement of the η_sep value indicates that the CoPi catalyst is also conducive to promoting the rapid separation of electron–hole pairs in the BiVO₄ body.

4. Conclusions

In summary, a highly compact and quasi-uniform BiVO₄ film was obtained at a suitable electrodeposition time and annealing temperature. After the successful electrodeposition of the CoPi catalyst on the optimized BiVO₄ electrode, the FTO/BiVO₄/CoPi photoanodes were fabricated, and their PEC performances were systematically investigated. Due to the surface modification of the CoPi catalyst, i.e., passivating charge recombination centers on the BiVO₄ surface and promoting the separation of electron–hole pairs in the BiVO₄ body, the interface impedance (|Z|) of the mass transfer process was decreased, while significantly enhancing the η_tran value of 82.1% and η_sep value of 25.4%. The J_ph and ABPE of the BiVO₄ photoanode were increased by about ~2 times and ~3 times, respectively, demonstrating that the CoPi catalyst can accelerate holes transfer from the BiVO₄ semiconductor to the catalyst in order to be competitive with water oxidation by holes in the semiconductor and to improve PEC performance. The results of this study may open up the possibility for the future design and construction of extremely effective photoanodes for solar-driven water splitting.