Improved Photoelectrochemical Performance of BiVO 4 for Water Oxidation Enabled by the Integration of the Ni@NiO Core–Shell Structure

: The development of highly efﬁcient and stable photoelectrode materials is of signiﬁcant importance for the conversion of solar energy into chemical fuels. Herein, a novel Ni@NiO/BiVO 4 photoanode is designed and prepared for efﬁcient water splitting by the deposition of Ni particles on the surface of BiVO 4 with subsequent thermal treatment. The integration of the Ni@NiO core–shell structure can efﬁciently passivate the surface states and accelerate the oxygen evolution kinetics along with the in situ-generated NiOOH, consequently contributing to the signiﬁcantly improved charge separation efﬁciency. The resulting Ni@NiO/BiVO 4 photoelectrode enabled a photocurrent density of 2.6 mA/cm 2 with a surface charge separation efﬁciency of nearly 80% at the potential of 1.23 V RHE —much better than the unmodiﬁed BiVO 4 (1.8 mA/cm 2 , 64%).


Introduction
In order to address the global energy crisis and environmental problems caused by the excessive use of fossil fuels, a number of researchers have been engaged with the energy conversion of solar energy to chemical fuels through photoelectrochemical cells [1]. In recent decades, metal oxide semiconductors have attracted considerable attention as photoelectrode materials due to their favorable valence band positions for water oxidation reactions and have been extensively studied [2], including TiO 2 , Fe 2 O 3 , WO 3 and BiVO 4 [3]. Among them, monoclinic scheelite-phase BiVO 4 has emerged as a promising photoanode for water oxidation as a result of its narrow bandgap and appropriate band edge positions [4]. However, the photoelectrochemical performance of BiVO 4 is still far from the theoretical value (7.5 mA cm −2 ) [5] ascribed to the severe charge recombination, slow oxygen evolution reaction kinetics and poor stability.
Up to the present time, plenty of strategies have been used to optimize the efficiency of BiVO 4 for water splitting. In particular, transition metal-based promoters have been explored to promote the reaction kinetics on the surface due to their high-efficiency OER activity. For instance, FeCoO x has been demonstrated as a co-catalyst on the surface of a planar BVO 4 photoanode to effectively reduce the water oxidation barrier between the photoanode and the electrolyte [6]. Similarly, metal-organic frameworks on the surface of BiVO 4 can provide active sites for OER to improve the water oxidation capacity of the resulting photoanode [7,8]. Additionally, a thin layer of in situ-formed FeNi oxyhydroxide on BiVO 4 can effectively prevent the precipitation of V 5+ through the formation of the V-O-Ni bond at the interface, thereby improving the photoelectrochemical stability of the resultant BiVO 4 photoanode [4]. However, the electron-hole recombination centers at the interface of the heterostructures may be introduced and detrimentally affect the performance of BiVO 4 [9].
Promisingly, inserting an electron mediator, such as Au [10], Pt [11], Ni [12], etc., in the middle of the heterojunctions can effectively improve the bulk charge separation. Among them, Ni and its oxides are attractive materials both as passivation layers and oxygen evolution catalysts, and have been widely studied for photoelectrochemical energy conversion [13,14]. Impressively, Lee et al. constructed a metal-insulator-semiconductor structure of NiO x /Ni/n-Si for efficient water splitting [15]. The nickel islands with a high-work function act as nano-emitters to efficiently collect the charges in the metalinsulator-semiconductor (MIS) structure, with NiO x expanding the Schottky barrier [16]. Furthermore, Malara et al. found that NiO x could be further oxidized to form highvalent Ni oxides during alkaline OER, contributing to the improved photoelectrochemical process [17]. Inspired by this, we propose a new strategy to improve the performance of BiVO 4 for photoelectrochemical water splitting. A layer of Ni particles can be deposited on the surface of BiVO 4 and then a thin NiO layer is in situ formed simply by heat treatment. The formed Ni@NiO core-shell structure can be used as a highly active OER co-catalyst and can passivate the surface states of BiVO 4 to improve the efficiency for water oxidation.

Results and Discussion
As shown in Figure 1a, the BiVO 4 electrode was prepared according to a previously reported method [18], and the Ni@NiO/BiVO 4 photoanode was obtained by spin-coating Ni nanoparticles that were prepared by a solution method on the surface of BiVO 4 followed by annealing in air. The BiVO 4 and Ni@NiO/BiVO 4 electrodes both show a porous structure composed of worm-like nanoparticles (Figure 1b,c), consistent with the literature [19]. Compared with the bare BiVO 4 , more fine nanoparticles appeared on the surface of the Ni@NiO/BiVO 4 photoanode. However, no additional diffraction peaks were observed ( Figure S1) in the X-ray diffraction (XRD) spectra, except for the peaks assigned to the monoclinic phase of pristine BiVO 4 (No. 14-0688)-probably due to the low content and small future size [4]. To figure out the phase composition of Ni species, X-ray diffraction (XRD) was conducted directly on the freshly prepared Ni nanoparticles as well as the particles after annealing at 400 • C. The results reveal that Ni nanoparticles display a characteristic peak at 44.5 • corresponding to the (111) crystal facet of cubic Ni (No.04-0850, Figure S2). Additionally, several weak characteristic peaks were observed at 37.2 • , 43.2 • and 62.8 • , indexed to the (101), (012) and (110) crystal facet of hexagonal NiO (No.  after annealing, thus, proving the formation of the Ni@NiO core-shell structure [20]; the peaks assigned to Ni were significantly intensified due to the improved crystallinity and the grown particle size caused by agglomeration during the annealing process. Furthermore, the TEM and high-resolution TEM (HRTEM) characterization on BiVO 4 and Ni@NiO/BiVO 4 were conducted, and the results are shown in Figures S3 and 1d. Clearly, some core-shell nanoparticles were observed on the surface of BiVO 4 , consistent with the observation of SEM. In addition, the heterostructure formed between the core-shell structure and BiVO 4 was confirmed with the three observed areas assigned to BiVO 4 , Ni and NiO, respectively, along with the results of XRD, SEM and XPS [21][22][23]. Specifically, the observed d-spacing of 0.461 nm corresponds to the (011) crystal plane of BiVO 4 , the d-spacing of 0.203 nm on the inner particles of the core-shell structure can be assigned to the (111) plane of the Ni nanoparticles, and the d-spacing of 0.205 nm on the shell of the core-shell structure corresponds to the (012) crystal plane of NiO, which are also consistent with the XRD results.
XPS was conducted to further prove the heterojunction formed by the core-shell structure and BiVO 4 and study the chemical states of the elements. In the high-resolution XPS spectrum of the Bi element (Figure 2a), the peak binding energies around 158.74 and 164.04 eV could be assigned to Bi 4f 7/2 and Bi 4f 5/2 of Bi 3+ in the pure BiVO 4 photoanode [24]. Furthermore, the V 5+ in the products was disclosed by the V 2p peak centers at around 516.40 and 523.80 eV (Figure 2c) in the BiVO 4 photoanode [25]. In the fine spectrum of Ni 2p, the Ni 2p 3/2 peaks centered at~856 eV and~852 eV were attributed to Ni 2+ and Ni 0 , respectively, which proved the existence of the Ni/NiO core-shell structure [26][27][28][29]. Only a trace of Ni 0 species was found in the 2p peak of Ni as a result of the detection depth of XPS-which is relatively shallow-and the presence of the annealed oxide layer ( Figure S3d) [30]. Compared with pure BiVO 4 , the positions of Bi 4f, O 1s, and V 2p exhibit a positive shift after being surface-loaded with a Ni@NiO core-shell structure, indicating that the electron density of BiVO 4 decreases, which can be ascribed to the surface binding interaction between the BiVO 4 and Ni@NiO core-shell structures, thus, inducing the electron transfer from BiVO 4 to Ni@NiO core-shell structures [31]. For instance, the Bi 4f 7/2 peak of Ni@NiO/BiVO 4 was shifted by 0.37 eV toward the higher-binding energy, which is close to that of the treated BiVO 4 with HCl [25], suggesting a stronger interaction between BiVO 4 and Ni@NiO with the formation of heterojunction [23]. XPS was conducted to further prove the heterojunction formed by the core-shell structure and BiVO4 and study the chemical states of the elements. In the high-resolution XPS spectrum of the Bi element (Figure 2a), the peak binding energies around 158.74 and 164.04 eV could be assigned to Bi 4f7/2 and Bi 4f5/2 of Bi 3+ in the pure BiVO4 photoanode [24]. Furthermore, the V 5+ in the products was disclosed by the V 2p peak centers at around 516.40 and 523.80 eV (Figure 2c) in the BiVO4 photoanode [25]. In the fine spectrum of Ni 2p, the Ni 2p3/2 peaks centered at ~856 eV and ~852 eV were attributed to Ni 2+ and Ni 0 , respectively, which proved the existence of the Ni/NiO core-shell structure [26][27][28][29]. Only a trace of Ni 0 species was found in the 2p peak of Ni as a result of the detection depth of XPS-which is relatively shallow-and the presence of the annealed oxide layer ( Figure  S3d) [30]. Compared with pure BiVO4, the positions of Bi 4f, O 1s, and V 2p exhibit a positive shift after being surface-loaded with a Ni@NiO core-shell structure, indicating that the electron density of BiVO4 decreases, which can be ascribed to the surface binding interaction between the BiVO4 and Ni@NiO core-shell structures, thus, inducing the electron transfer from BiVO4 to Ni@NiO core-shell structures [31]. For instance, the Bi 4f7/2 peak of Ni@NiO/BiVO4 was shifted by 0.37 eV toward the higher-binding energy, which is close to that of the treated BiVO4 with HCl [25], suggesting a stronger interaction between BiVO4 and Ni@NiO with the formation of heterojunction [23]. The optical properties of BiVO4 and Ni@NiO/BiVO4 films were measured by UV-vis-NIR absorption spectroscopy. In the UV-vis-NIR spectra, both the absorption edges of the obtained photoanodes are located at approximately 500 nm; an increase in the optical absorption intensity of Ni@NiO/BiVO4 films can be clearly observed over 500 nm, which The optical properties of BiVO 4 and Ni@NiO/BiVO 4 films were measured by UV-vis-NIR absorption spectroscopy. In the UV-vis-NIR spectra, both the absorption edges of the obtained photoanodes are located at approximately 500 nm; an increase in the optical absorption intensity of Ni@NiO/BiVO 4 films can be clearly observed over 500 nm, which may be attributed to the surface plasmon resonance effect of metallic Ni nanoparticles [32].
The photoelectrochemical performance of the prepared BiVO 4 and Ni@NiO/BiVO 4 electrodes was evaluated in a typical three-electrode cell (Figure 3b-f). Generally, the spin-coating cycle and annealing temperature show critical effect on the PEC performance ( Figure S4) due to their modulation of the coverage, crystallinity and agglomeration of Ni@NiO on BiVO 4 . As displayed in Figure 3b, the optimized Ni@NiO/BiVO 4 photoelectrode exhibits a significantly improved photocurrent in contrast to the bare BiVO 4 , where a photocurrent of 2.6 mA/cm 2 at 1.23 V RHE is obtained, much higher than that of BiVO 4 (1.8 mA/cm 2 ). No significant change in onset potential was observed, this was likely due to the V 4+ /V 5+ redox peak near the onset potential [33]. Assuming that the Faraday efficiency is 100%, the applied bias photocurrent efficiency (ABPE) of BiVO 4 and Ni@NiO/BiVO 4 photoanodes was calculated from the LSV curve to evaluate the solar energy conversion efficiency in the PEC process for water splitting (Figure 3c). The maximum ABPE of the bare BiVO 4 photoanode is 0.48% at 0.80 V, while the maximum ABPE of the Ni@NiO/BiVO 4 photoanode is largely enhanced to about 0.73% at 0.77 V vs. RHE, suggesting significant efficiency improvement for solar water splitting compared with the BiVO 4 photoanode ( Table S1). The PEC stability of BiVO 4 and Ni@NiO/BiVO 4 were studied at 1.23 V RHE with/without chopping illumination (Figure 3d,e). Both electrodes show good optical switching characteristics and a fast response, and apparently the Ni@NiO/BiVO 4 photoanodes exhibited slightly enhanced stability compared with the bare BiVO 4 [34]. After the operation for 60 min, the photocurrent of the heterostructure photoanode is still 2.2 times that of BiVO 4 with the maintained percentage of the initial current enhanced from 49.31% to 62.64%, implying the superiority of the heterojunction structure. Clearly, the surface charge separation efficiency of Ni@NiO/BiVO 4 is significantly higher than that of the bare BiVO 4 photoanodes (Figure 3f), suggesting that the formation of the Ni@NiO core-shell structure on the surface can accelerate surface charge transfer, promote water oxidation kinetics and contribute to the improved PEC performance [35].
In order to understand the work mechanism of the heterostructure photoanode well, the catalytic chemistry on the surface and the transportation and separation of the photo-generated charge carriers were investigated. When the electrochemical OER activity without illumination was evaluated within a wider potential range (Figure 4a), the Ni@NiO/BiVO 4 photoanode exhibited a lower overpotential and higher water oxidation current than the bare BiVO 4 photoanode, suggesting the higher OER activity on the surface of Ni@NiO/BiVO 4 . Furthermore, the charge-transfer rate at the photoanode/electrolyte interface is related to the electrochemically active surface area (ECSA), and a low ECSA may result in the accumulation of photo-generated holes that reach the surface, detrimentally promoting their recombination with the photo-generated electrons [36][37][38]. As shown in Figure 4b, the relationship between the capacitive currents and the scanning rates of the two photoanodes was calculated from their cyclic voltammogram ( Figure S5), where the slope is proportional to the ECSA of the photoanode. Interestingly, the ECSA of the Ni@NiO/BiVO 4 photoanode is 36% higher than that of the pure BiVO 4 , indicating that the integration of Ni@NiO core-shells brings more active sites [39] for the oxygen evolution reaction and simultaneously implies that the enhanced intrinsic catalytic activity of the modified BiVO 4 is also an important factor for the enhanced performance. To further verify the critical role of the surface Ni@NiO core-shell in enhancing the photoelectrochemical performance of BiVO 4 , the behavior of the photo-generated charge carrier was studied by the open circuit photovoltage (OCP, Figures 4c and S6). The degree of band bending is determined by the built-in potential, minority carrier accumulation, and charge recombination in the photoanode/electrolyte junction. Generally, larger band bending at the photoanode/electrolyte interface is more favorable for the photo-generated electron-hole separation. Basically, a more positive OCV value in the dark and a more negative OCV light under illumination imply larger band bending [40]. Compared with bare BiVO 4 , the energy band bending of Ni@NiO/BiVO 4 is significantly improved with a more positive OCV dark , indicating that the Ni@NiO core-shell can effectively passivate the surface states of BiVO 4 [41]. The lifetime decay process of the photo-generated charge carrier when the light is turned off can also be used to study the properties of the semiconductor/electrolyte interface. The accumulation of charge carriers due to band bending enables a rapid recombination of photo-generated electrons and holes in the dark [42]. Clearly, the photoanode of Ni@NiO/BiVO 4 exhibits a faster decay (Figure 4d), indicating the rapid separation of photo-generated electrons and holes at the semiconductor/electrolyte interface under illumination. In order to understand the work mechanism of the heterostructure photoanode well, the catalytic chemistry on the surface and the transportation and separation of the photogenerated charge carriers were investigated. When the electrochemical OER activity without illumination was evaluated within a wider potential range (Figure 4a), the Ni@NiO/BiVO4 photoanode exhibited a lower overpotential and higher water oxidation current than the bare BiVO4 photoanode, suggesting the higher OER activity on the surface of Ni@NiO/BiVO4. Furthermore, the charge-transfer rate at the photoanode/electrolyte interface is related to the electrochemically active surface area (ECSA), and a low ECSA may result in the accumulation of photo-generated holes that reach the surface, detrimentally promoting their recombination with the photo-generated electrons [36][37][38]. As shown in Figure 4b, the relationship between the capacitive currents and the scanning rates of the two photoanodes was calculated from their cyclic voltammogram ( Figure S5), where the slope is proportional to the ECSA of the photoanode. Interestingly, the ECSA of the Ni@NiO/BiVO4 photoanode is 36% higher than that of the pure BiVO4, indicating that the integration of Ni@NiO core-shells brings more active sites [39] for the oxygen evolution reaction and simultaneously implies that the enhanced intrinsic catalytic activity of the modified BiVO4 is also an important factor for the enhanced performance. To further verify the critical role of the surface Ni@NiO core-shell in enhancing the photoelectrochemical performance of BiVO4, the behavior of the photo-generated charge carrier was studied by the open circuit photovoltage (OCP, Figures 4c and S6). The degree of band bending is determined by the built-in potential, minority carrier accumulation, and charge recombination in the photoanode/electrolyte junction. Generally, larger band bending at the photoanode/electrolyte interface is more favorable for the photo-generated electron-hole separation. Basically, a more positive OCV value in the dark and a more To further study the charge-transfer process, EIS measurements were carried out ( Figure 4e) and the results were fitted using the corresponding equivalent circuit in the inset. The model consists of R s and two RC circuits connected in series. R s refers to the ohmic resistance generated by the electrolyte, electrical contacts and wires; R tr represents the bulk resistance; and R ct corresponds to the charge-transfer resistance at the photoelectrode/electrolyte interface with CPE representing the ideal capacitor for the corresponding process [43]. Compared with the bare BiVO 4 (744.6 Ω), R tr of the Ni@NiO/BiVO 4 sample was reduced to 519.7 Ω. It was deduced that the decrease in R tr was due to the close contact between the Ni@NiO core-shell and BiVO 4 through annealing as well as the Schottky junction formed between the Ni@NiO core-shell structure and BiVO 4 [12,15]. In addition, the R ct of Ni@NiO/BiVO 4 is significantly reduced from 668.6 Ω of bare BiVO 4 to 226.5 Ω, indicating that Ni@NiO is a good co-catalyst that can accelerate the slow oxygen evolution kinetics on the BiVO 4 surface. This result is also consistent with the result of surface separation efficiency. As shown in Figure 4f, there is a clear intensity difference in the peaks around 493 nm related to the radiative recombination of holes in the O 2p band and electrons in the V 3d band, which is related to the electron-hole recombination in the bulk [44,45]. While for wavelengths greater than 520 nm, BiVO 4 does not show intrinsic light absorption, and the emission peaks may be related to the surface states of BiVO 4 [46]. Clearly, the Ni@NiO/BiVO 4 photoanode exhibited a reduced PL peak in less than 520 nm, indicating that the deposition of Ni@NiO core-shells can significantly reduce the bulk recombination of photo-generated electron-hole pairs due to the Schottky junction formed between the Ni@NiO core-shell and BiVO 4 [12,15]. In addition, Ni@NiO/BiVO 4 exhibited a smaller PL intensity in the range greater than 520 nm, which was due to the Ni@NiO core-shell structure effectively passivating the surface states, which further promoted the charge separation; this is also consistent with the above results. a rapid recombination of photo-generated electrons and holes in the dark [42]. Clearly, the photoanode of Ni@NiO/BiVO4 exhibits a faster decay (Figure 4d), indicating the rapid separation of photo-generated electrons and holes at the semiconductor/electrolyte interface under illumination. To further study the charge-transfer process, EIS measurements were carried out ( Figure 4e) and the results were fitted using the corresponding equivalent circuit in the inset. The model consists of Rs and two RC circuits connected in series. Rs refers to the ohmic resistance generated by the electrolyte, electrical contacts and wires; Rtr represents the bulk resistance; and Rct corresponds to the charge-transfer resistance at the photoelectrode/electrolyte interface with CPE representing the ideal capacitor for the corresponding process [43]. Compared with the bare BiVO4 (744.6 Ω), Rtr of the Ni@NiO/BiVO4 sample was reduced to 519.7 Ω. It was deduced that the decrease in Rtr was due to the close contact between the Ni@NiO core-shell and BiVO4 through annealing as well as the Schottky junction formed between the Ni@NiO core-shell structure and BiVO4 [12,15]. In addition, the Rct of Ni@NiO/BiVO4 is significantly reduced from 668.6 Ω of bare BiVO4 to 226.5 Ω, indicating that Ni@NiO is a good co-catalyst that can accelerate the slow oxygen evolution kinetics on the BiVO4 surface. This result is also consistent with the result of surface separation efficiency. As shown in Figure 4f, there is a clear intensity difference in the peaks around 493 nm related to the radiative recombination of holes in the O 2p band and electrons in the V 3d band, which is related to the electron-hole recombination in the bulk [44,45]. While for wavelengths greater than 520 nm, BiVO4 does not show intrinsic light absorption, and the emission peaks may be related to the surface states of BiVO4 [46]. Clearly, the Ni@NiO/BiVO4 photoanode exhibited a reduced PL peak in less than 520 nm, indicating that the deposition of Ni@NiO core-shells can significantly reduce the bulk recombination of photo-generated electron-hole pairs due to the Schottky junction formed between the Ni@NiO core-shell and BiVO4 [12,15]. In addition, Ni@NiO/BiVO4 exhibited a smaller PL intensity in the range greater than 520 nm, which was due to the Ni@NiO Furthermore, after the PEC operation for 10 mins, the role of the Ni@NiO core-shell becomes more obvious (Figure S7a-c). To explore the reason for this change, we performed XPS measurements on the samples. It was found that the Ni 3+ peak appeared in the spectrum of Ni@NiO/BiVO 4 instead of the Ni 0 peak. This result indicates that the Ni species was oxidized during the PEC process and the Ni core was gradually oxidized to Ni oxides with the in situ formation of NiOOH on the surface, consistent with the previously reported results [15,47]. The generated NiOOH is also an excellent oxygen evolution catalyst and enables the accelerated oxygen evolution kinetics on the surface of the Ni@NiO/BiVO 4 photoanode [30], thus, contributing to the improved PEC performance.
Summarily, the integration of the Ni@NiO core shell on the surface of BiVO 4 plays an important role in improving the PEC performance ( Figure 5). Usually, the poor oxygen evolution kinetics and surface states of BiVO 4 result in the serious recombination of the photo-generated holes and electrons on the surface of BiVO 4 . The introduction of the Ni@NiO core-shell structure on the surface of BiVO 4 can form a heterojunction to promote the charge separation in the bulk; the photo-generated holes can be transferred from the valence band of BiVO 4 to Ni@NiO for the water oxidation reaction, promoting the chargetransfer process. On the other hand, the Ni@NiO core-shell structure on the surface can not only improve the oxygen evolution kinetics of BiVO 4 , but also passivate the surface state, thereby significantly improving the charge separation efficiency at the BiVO 4 /electrolyte interface (Figure 5b) [48]. Additionally, the conversion of the Ni@NiO core-shell into NiOOH during the PEC process enables the acceleration of the photoelectrochemical reaction process.

Preparation of the Worm-Like BiVO 4 Film
The BiVO 4 film on FTO was prepared according to a modified method reported by Choi [18]. Typically, the Bi(NO 3 ) 3 solution (0.04 M) was prepared by dissolving Bi(NO 3 ) 3 ·5H 2 O in 50 mL of the KI (0.4 M) solution with the pH adjusted to 1.7 by HNO 3 . Afterwards, 20 mL of the p-benzoquinone (0.23 M) solution in absolute ethanol was slowly added into the as-prepared Bi(NO 3 ) 3 solution, and then the resultant solution was vigorously stirred for 40 min at room temperature. The precursor of BiOI nanosheets was prepared via electrodeposition in a typical three-electrode cell. FTO (3 × 1 cm 2 ) serves as the working electrode (WE) with a saturated Ag/AgCl as the reference electrode (RE) and platinum foil (1.5 × 1.5 cm 2 ) as the counter electrode (CE). Cathodic deposition was conducted potentiostatically at −0.147 V vs. SCE at room temperature for 420 s. The obtained orange BiOI precursor was rinsed with DI water to remove the excess electrolyte on the surface. Next, 0.11 mL of the dimethyl sulfoxide (DMSO) solution containing vanadyl acetylacetonate (VO(acac) 2 ) (0.2 M) was placed on the BiOI film (1 × 2 cm 2 ), and the film was annealed in a muffle furnace at 450 • C (ramping rate = 2 • C min −1 ) for 2 h. The BiVO 4 film electrode was obtained intact on the FTO with the color changing from orange to yellow after removing the excess V 2 O 5 by soaking it in the NaOH solution (1 M) for 30 min with mild stirring.

Preparation of Ni Nanoparticle
The nickel nanoparticles were synthesized according to a previously reported method with a slight modification [49]. Typically, 3.07 g of SDS, 1.3 mL of OA and 0.31 g of Ni(NO 3 ) 2 ·6H 2 O were sequentially added to 107 mL of DI water whilst stirring vigorously for 30 min to obtain a uniform solution. Afterwards, 0.05 g of NaBH 4 was added to the above solution whilst stirring for 1 h to obtain Ni particles.

Preparation of Ni@NiO/BiVO 4 Photoanode
The freshly prepared Ni particles were spin-coated on the freshly prepared BiVO 4 film with a rotation speed of 2000 rpm/s and a spin-coated time of 15 s. Repeated different cycles can be applied to obtain different samples of Ni/BiVO 4 . Finally, the Ni/BiVO 4 samples were annealed in a muffle furnace at 400 • C for 1 h.

Materials Characterization
The X-ray diffraction patterns of all the electrodes were recorded on a Rigaku X-ray diffractometer D/Max-2200/PC (Nishiko Corporation of Japan, Tokyo, Japan) equipped with Cu-Kα radiation (40 kV, 20 mA). The optical properties of all the samples were estimated by a UV-vis-IR spectrophotometer (Cary 5000) (Agilent Technologies, Palo Alto, CA, USA) equipped with an integration sphere unit using a diffuse reflection method. The structure and morphology of all the samples were studied using a JSM6701E field emission scanning electron microscope (JEOL, Tokyo, Japan) and transmission electron microscopy Talos F200x (Thermo Fisher Scientific, Waltham, MA, USA). XPS was recorded on a Thermo Scientific K-Alpha photoelectron spectrometer (Thermo Fisher Scientific). The photoluminescence (PL) spectra were obtained on a F-4500 FL Spectrophotometer (Hitachi, Tokyo, Japan) under a laser excitation of 400 nm.

Photoelectrochemical Measurements
The PEC performance of all the samples was evaluated by an electrochemical workstation (CHI 660E) in a typical three-electrode cell containing a 0.5 M phosphate-buffered saline (pH = 9). The bare BiVO 4 and Ni@NiO/BiVO 4 electrodes served as the working electrode; the saturated Ag/AgCl electrode and platinum foil acted as the RE and CE, respectively. A xenon lamp simulating sunlight AM 1.5G (100 mW/cm 2 ) was used as the light source. The linear sweep voltammetry curves were recorded within a potential range from −0.75 to +0.6 V (vs. Ag/AgCl) at a scan rate of 10 mV s −1 . For all the samples, the illumination was from the backside of the fluorine-doped tin oxide (FTO) glass with an area of 1 cm 2 . All the potentials were calibrated to the RHE according to the below equations: Electrochemical impedance spectroscopy (EIS) was carried out by applying an AC voltage amplitude of 10 mV over a frequency range of 10 5 -0.01 Hz at an open circuit potential under AM 1.5G illumination.
The applied bias photon-to-current efficiency (ABPE) for water splitting was calculated from the J-V curve according to the following equation: where I is the photocurrent density at a specific potential, V bias is the applied bias between WE and RHE and P light is the incident illumination power density (100 mW cm −2 ).

Conclusions
An Ni@NiO/BiVO 4 three-component heterostructure photoanode was designed and prepared for efficient water splitting through the deposition of Ni particles on the surface of BiVO 4 with subsequent thermal treatment. The PEC performance and the working mechanism were investigated, and the results indicate that the formation of the heterostructure can efficiently promote the charge separation. More importantly, the integration of the Ni@NiO structure enables the passivation of the surface states and accelerates the oxygen evolution kinetics on the surface of BiVO 4 , consequently leading to a significantly enhanced PEC performance. The optimized Ni@NiO/BiVO 4 photoanode delivered a photocurrent density of 2.6 mA/cm 2 and a surface charge separation efficiency of nearly 80% at the potential of 1.23 V RHE , which is much better than the bare BiVO 4 (1.8 mA/cm 2 , 64%). Additionally, the Ni@NiO/BiVO 4 photoanode also exhibits good photoelectrochemical stability; the photocurrent density is 2.2 times that of BiVO 4 after a 60 min long-term operation.