Bi(NO₃)₃·5H₂O (Sinopharm Chemical Reagent Co., Ltd., 99.0%), potassium iodide, ethylene glycol are purchased from chemical reagent co, Ltd. P-benzoquinone (Tianjin Institute of Fine Chemicals, 99.0%). Ni(NO₃)₂·6H₂O (Sinopharm Chemical Reagent Co., Ltd., 99.0%), hexamethylenetetramine (HMTA, Chengdu Cologne Chemicals Co., Ltd., 99.0%), Anhydrous ethanol was purchased from Sinopharm Chemical Reagent Co., Ltd. All aqueous solutions were prepared with deionized water.

The preparation of BiVO₄ was synthesized with reference to previous reported work. Systematically, 50 ml of 0.4 M KI solution was first adjusted to pH 1.7 with 1 M HNO₃, and then 5 mmol Bi(NO₃)₃·5H₂O was added with rapid stirring until dissolved, resulting in an orange-red mixed solution. Then 20 ml of ethanol containing 4.6 mmol of 1,4-benzoquinone was slowly added dropwise to the above solution and stirred for several tens of minutes. Next, BiOI nanosheets were synthesized in a three-electrode system by electrodeposition. Among them, the platinum electrode was used as the counter electrode, the clean FTO glass was used as the working electrode, and the Ag/AgCl (3.5 M KCl) electrode was used as the reference electrode. Cyclic voltammetry (CV) was used for electrodeposition, and the resulting membrane was rinsed with distilled water to obtain a clean BiOI membrane. Immediately after, 150 μl of 0.2 M vanadyl acetylacetonate (VO(acac)₂) DMSO solution was dropped onto the above BiOI nanosheets. Calcined at 450°C for 2 h at a ramp rate of 2°C/min. The cooled membrane was washed with 1 M NaOH solution to remove excess V₂O₅ from the BiVO₄ electrode.

The BiVO₄/NiO photoanode was prepared by a hydrothermal method. The configuration takes 60 ml of solution in which the volume ratio of deionized water and ethanol solution is 2:1. After ultrasonically mixed uniformly, 1.5 mmol Ni(NO₃)₂·6H₂O and 6 mmol HMTA were added thereto, and stirred until dissolved. Then 20 ml of the above mixture was added to a PTFE-lined stainless steel autoclave (100 ml). And the as-prepared BiVO₄ photoanode was placed obliquely with the conductive side facing upwards, heated at 90°C for 4 h. The photoanode was washed three times with water and ethanol and dried at 60°C for 12 h. Finally, the BiVO₄/NiO photoanode was obtained after calcination in air with a heating rate of 2°C/min at 300°C for 2 h.

The microscopic morphology of the samples was characterized by scanning electron microscopy (SEM, JSM-6701E). The crystal structure of the as-prepared photoanode was measured by X-ray diffraction (XRD) tests on the X-ray diffractometer (D/MAX-2200/PC). UV-Vis diffuse reflectance spectroscopy was used to investigate the response of the prepared photoelectrode to visible light on a UV-3100 spectrometer.

The photoelectrochemical tests of the as-prepared photoanodes were carried out on a CHI 660D electrochemical workstation. A three-electrode system was used, in which a platinum sheet, Ag/AgCl (3.5 M KCl), and the samples were the counter electrode, the reference electrode and the working electrode, respectively. The electrolyte is 0.5 M Na₂SO₄ solution (pH = 6.86). All tests were performed with FTO backside irradiation at room temperature. The scan rate for linear sweep voltammetry (LSV) was 10 mV s⁻¹. The light intensity was calibrated to 100 mW/cm² with an optical power meter. And the incident photon current efficiency (IPCE) was measured using a 300 W xenon lamp with a monochromator in 0.5 M Na₂SO₄ electrolyte at 1.23 V vs. RHE. The photogenerated photocurrent ( J a b s ) undergo two major losses of charge carriers recombination in bulk and at interface. Hence the measured photocurrent during water oxidation ( J H 2 O ) is expressed as follows: J H 2 O = J a b s × η i n j × η s e p where J_abs is obtained by integrating the distribution of solar power density P ( λ ) with light absorption α ( λ ) of the photoanode as equation: J a b s = e ∫ 0 λ a b s α ( λ ) P ( λ ) h υ d λ where α ( λ ) = 1 − 10 − A , A is the absorbance according to the UV-vis spectrum, P ( λ ) = the distribution of solar power density.

The photocurrent during sodium sulfite oxidation was measured ( J S O 3 2 − ) in order to calculate charge separation efficiencies. Since the interface charge separation efficiency of sodium sulfite oxidation is almost 100% ( η sep  ,   S O 3 2 − = 1 ) , the charge separation efficiencies can be calculated as follows: η i n j = J S O 3 2 − / J a b s   η s e p = J H 2 O / J S O 3 2 −  

The morphology and elemental compositions of the synthesized BiVO₄, BiVO₄/NiO composite photoanode were studied with SEM and energy dispersive spectroscopy (EDS) in Figures 1A–H. Figure 1A exhibits the SEM image of nanoporous BiVO₄ film. Figure 1B shows the SEM image of BiVO₄/NiO. NiO nanosheets with size distribution about 5–10 nm are fairly continuous and uniformly loaded on the surface of BiVO₄ in large area. The elemental composition and content of BiVO₄/NiO anodes were further investigated by EDS. Figure 1C shows the EDS pattern of BiVO₄/NiO photoanode, the weight percentage of elements present in the BiVO₄/NiO photoanode composites are 39.4%, 31.1%, 18.6% and 10.09% for Bi, O, V and Ni, respectively. No other elements or impurities are found. The elemental composition and distribution in the BiVO₄/NiO photoanode were further observed by EDS elemental mapping (Figures 1D–H ). The results clarify that the elements of Bi, O, V and Ni are present and uniformly distributed in the BiVO₄/NiO photoanode.

The optical absorption properties of pure BiVO₄ and BiVO₄/NiO films were investigated by UV-Vis diffuse reflectance spectroscopy. It can be seen that BiVO₄ and BiVO₄/NiO show good light absorption properties around 500 nm (Figure 2A), corresponding to a band gap of 2.4 eV. Notably, the BiVO₄/NiO sample shows almost the same absorption edge as bare BiVO₄ due to the blocking by the thicker BiVO₄ layer, indicating that coating of NiO almost rarely affects light absorption of BiVO₄/NiO. But the BiVO₄ photocathode loaded with NiO co-catalyst shows the relatively low light absorption capacity. The reason is primarily ascribed to the poor optical transparency of the nickel based co-catalyst. (Zhou et al., 2020)

The XRD pattern characterization reveals that the product is composed of three kinds of materials with distinct crystal structures. Figure 2B displays the XRD patterns of BiVO₄ and BiVO₄/NiO nanocomposites. The XRD diffraction peaks of BiVO₄ and BiVO₄/NiO nanocomposites are completely consistent with monoclinic BiVO₄ (JCPDS No. 14-0688) and tetragonal SnO₂ (JCPDS No. 46-1088) derived from FTO substrates. No other impurity phases were detected. The appearance of characteristic diffraction peaks at 2θ = 43.3° is corresponding to the (200) crystal plane of the cubic phase NiO, which can be concluded that the composite sample has been successfully prepared.

In order to explore the effect of supported NiO on the PEC performance, the photoelectrochemical water splitting performance of BiVO₄ photoanode modified by NiO cocatalyst was studied by electrochemical workstation. Linear sweep voltammetry (LSV) curves reflect the water oxidation properties of BiVO₄/NiO and pure BiVO₄. As shown in Figure 3A, the photocurrent density of unmodified BiVO₄ at 1.23 V vs. RHE is 0.9 mA/cm² and the onset potential is about 0.58 V vs. RHE, which is due to its unique structure and specific crystal orientation, leading to rapid transfer and separation of photogenerated carriers. After loading the NiO cocatalyst on BiVO₄, the photoelectrode showed significant enhancement at all potentials, obtaining a photocurrent of 1.2 mA/cm² at 1.23 V vs. RHE, which was higher than that of pristine BiVO₄. In particular, BiVO₄/NiO obtained a more negative onset potential compared to pure BiVO₄, with a negative shift of about 0.35 V (Figure 3B). The increased photocurrent density and negatively shifted onset potential clearly indicate that the addition of NiO cocatalyst is a feasible way to enhance the water oxidation capacity of BiVO₄ photoanode. Figure 3C shows the chopped photocurrent density−voltage (J−V) curves of BiVO₄/NiO and pure BiVO₄. All the photoanodes show an obvious “photo-switching” effect with fast response. Clearly, the BiVO₄/NiO photoanode exhibits a much better PEC performance than BiVO₄ film.

The electron-hole pair recombination and charge generation kinetics of photoanode in PEC water oxidation process can be analyzed by EIS. The photoanode was measured at 1.23 V vs. RHE at AM 1.5G (100 mW/cm²), and the frequency range of the Nyquist plot was from 100 kHz to 0.1 Hz. The results of impedance spectra are useful for analyzing electrochemical surface reactions. The charge transfer resistance of the photoanode surface is estimated from the small semicircle in the Nyquist diagram, and the smaller the radius, the more effective the separation of charges. In addition, the EIS test result of dark reaction condition (Figure 4B) is consistent with that under light conditions in Figure 4A. The BiVO₄/NiO photoanode exhibits the highest charge transportation, suggesting good charge separation ability.

To further explore the charge recombination at the BiVO₄ and BiVO₄/NiO photoanode interfaces, the charge separation and injection efficiencies were tested in Figures 4B,C. Charge separation efficiency is an important parameter to evaluate the proportion of carriers reaching the electrode surface/electrolyte interface to participate in water oxidation (Wang et al., 2017b). Therefore, the constant charge separation efficiency is shown in Figure 3D, ascribed to the photocurrent density of the BiVO₄/NiO photoanode at an applied potential of 0.6 V vs. RHE when Na₂SO₃ was added to the electrolyte. The separation efficiency of the pure BiVO₄ photoanode increases with the applied potential, especially it can reach about 60% at 1.23 V vs. RHE. However, the BiVO₄/NiO nanostructured array photoanode exhibits a charge separation efficiency of 90% at 1.23 V vs. RHE. From this point of view, the cocatalyst NiO supported on BiVO₄ can significantly improve the charge separation efficiency and facilitate the flow of charge carriers to the electrode surface/electrolyte interface to participate in water oxidation.

The photogenerated holes generated on the surface of the photoanode participate in the water oxidation reaction or recombine with electrons. The charge injection efficiency, defined as the fraction of those holes at the photoanode and electrolyte interface that is used for water oxidation reactions, can be improved by reducing surface recombination or accelerating hole transfer kinetics. (Zhou et al., 2020) As shown in Figure 4D, the charge injection efficiency of the pure BiVO₄ photoanode reaches 28% 1.23 V vs. RHE. While the charge injection efficiency of the BiVO₄/NiO photoanode increases to 35% in the potential range of 1.23 V vs. RHE.

The quantum efficiencies of BiVO₄, BiVO₄/NiO photoanodes were determined by measuring incident photon current efficiency (IPCE) and absorbed photon current efficiency (APCE). (Wang et al., 2017b) The calculation of IPCE can refer to the following equation: IPCE = ( J × 1240 ) / ( P × λ ) where J is the current density (mA/cm²) measured at each specific wavelength, λ is the wavelength of the incident light (nm), and P is the power density of the incident light (mW/cm²). As shown in Figure 5A, BiVO₄/NiO nanocomposites exhibit slightly higher IPCE values in the 450–500 nm range compared to bare BiVO₄. The PEC performance mainly depends on the light-harvesting efficiency, charge separation efficiency, and collection yield. Since the light-harvesting efficiency is almost unchanged after the modification of the NiO cocatalyst, the enhancement of IPCE may be due to the fast separation of charge carriers and the accelerated water oxidation kinetics of the reaction, resulting in the enhanced photocurrent. The IPCE results are consistent with the above J-V measurements.

To obtain the absorbed photon-current efficiency, the APCE value of the photoanode was measured at 0.6 V vs. RHE, as shown in Figure 5B. The APCE value of BiVO₄/NiO photoanode is significantly higher than that of BiVO₄ from 380 to 500 nm, which is consistent with the overall PEC performance.

Figure 6A shows the Mott Schottky barrier of BiVO₄ and BiVO₄/NiO. The capacitance-voltage curve is usually used to analyze the reasons for the enhancement of semiconductor performance. In order to better study the reasons for the increase of photocurrent after supporting the cocatalyst. Therefore, the capacitance-voltage curves of BiVO₄ and BiVO₄/NiO electrodes were measured under dark reaction conditions, and the x-axis tangent was made on the curves of BiVO₄ and BiVO₄/NiO. The tangent was positive, indicating that BiVO₄ and BiVO₄/NiO are both n-type semiconductors. The smaller slope of the composite electrode, the greater the carrier density. Thus, the BiVO₄/NiO electrode has the largest carrier density. The increased carrier density causes the conductivity of BiVO₄ to increase and ultimately increases its photocurrent density. Fluorescence spectroscopy (PL) can be used to effectively analyze the separation and recombination effects of photogenerated carriers. As shown in Figure 6B, it can be observed that the peak intensity of the BiVO₄/NiO electrode is weaker than that of the pure BiVO₄ electrode, which indicates that after NiO is loaded on the BiVO₄ surface, the recombination rate of photo-generated electrons and holes becomes slower and the charge separation efficiency is improved.

Stability test is an important index parameter to evaluate the effect of photoelectric catalyst and whether it has application value. The stability test of the BiVO₄/NiO photoanode was analyzed under continuous irradiation under AM 1.5G. As shown in Figure 7, it can be observed that under the continuous irradiation of 3 h, the photocurrent density of the NiO/BiVO₄ electrode is not significantly attenuated, indicating that the stability of the BiVO₄ photoanode can be improved after loading NiO.