1 Introduction

Since the groundbreaking work of Fujishima and Honda was published in 1972, photoelectrochemical (PEC) water splitting using semiconductor photoanodes materials to produce hydrogen and oxygen, is widely regarded as a promising route to settle the energy crisis worldwide in the future [1,2,3,4,5,6]. To date, a considerable number of photoanode materials, such as TiO2 [7,8,9,10], WO3 [11,12,13,14], α-Fe2O3 [15,16,17,18], Ta3N4 [19,20,21,22,23], and BiVO4 [24,25,26] and so on, have been extensively developed by scientists. Among these candidates, BiVO4, as an n-type semiconductor, is identified as a promising photoanode material for PEC water splitting, on account of its appropriate valance band level, visible-light response, excellent physicochemical stability, nontoxic, and low-cost. Unfortunately, bare BiVO4 photoanodes always suffer from sluggish charge mobility, and fast surface charge recombination at the interface of the photoanode [27]. Therefore, tremendous effective strategies have been adopted, including surface modification [25], together with other semiconductors [28], and element doping [29], to alleviate the bottleneck of bare BiVO4 photoanode. Specially, surface modification using cocatalysts, made up of earth-abundant elements, were considered as a convenient, low-cost and reproducible pathway to handled above-mentioned problems [24, 30,31,32,33]. For example, our group recently demonstrated that CoP nanoparticles can act as a noble metal-free cocatalyst on α-Fe2O3 photoanode to improve photoelectrochemical solar-driven water splitting for the first time [15]. Ding and coworkers reported the use of Co–borate as an efficient cocatalyst by electrodeposited onto the surface of BiVO4 photoanode, which exhibits high PEC stability and performance [31]. Yu and cooperators found that the use of ferrihydrite as hole-storage layer to reduce the interface recombination for BiVO4 PEC, resulting in a remarkable photocurrent density of 4.78 mA/cm2 at as low as 0.6 V vs RHE [34]. Jason and collaborators fabricated FeOOH/BiVO4 photoanode by a simple electro-deposition and heat treatment process, which exhibitied significantly improved photocurrent and stability for photoelectrochemical water splitting [35].

More recently, FexP (x = 1 or 2), which is made of earth-abundant elements, can be used as a good electrocatalyst for H2 production due to its enhancing the electronic conductivity [36,37,38]. Meanwhile, it was also found that Fe2P were utilized as an active cocatalyst to enhance photocatalytic hydrogen production [39, 40]. However, to the best of our knowledges, the use of FeP as a cocatalyst on BiVO4 photoanodes using the drop casting method for PEC water splittingr has rarely been received prior investigation.

Motivated by above mentioned analysis, in this study, pure BiVO4 photoanodes with nanoporous morphology were initially synthesized to effectively increase the bulk charge separation. Furthermore, for the first time, we employ low-cost FeP nanoparticles as an effective cocatalyst to modify the surface of nanoporous BiVO4 photoanodes, leading to the highly efficient suppression of surface recombination for PEC water splitting. The resulting FeP/BiVO4 photoanode achieved a improved photocurrent density of ~ 3.05 mA/cm2 at 1.23 V vs RHE under AM 1.5G illumination. Meanwhile, the surface charge recombination could be effectively suppressed by decorating photoanode with earth-abundant cocatalyst FeP nanoparticles.

2 Experimental Section

2.1 Materials

Chemicals (analytical grade) in this experiments, mainly including bismuth nitrate pentahydrate [Bi(NO3)3·5H2O, 99.0% purity], potassium iodide (KI, 99.0% purity), p-benzoquinone (C6H4O2, 99.0% purity), vanadium(IV)oxy acetylacetonate [VO(acac)2, 98.0% purity], Iron(III) chloride(FeCl3, 99.0% purity), sodium hypophosphite monohydrate (NaH2PO2·H2O, 98.0% purity), and dimethyl sulfoxide (DMSO, 99.0% purity), were all purchased from Aldrich or Aladdin Reagent Co., Ltd. (China) and used without specially purification.

2.2 Preparation of FeP/BiVO4 Photoanodes

BiVO4 photoanodes samples were prepared according to the original literature developed by Kim and Chio [27]. Typically, ~ 3.32 g KI was added to ~ 50 mL of distilled water. The solution with the pH was adjusted to ~ 1.70 by s dilute HNO3, then ~ 970.14 mg Bi(NO3)3·5H2O was dissolved in the KI solution, which was mixed with ~ 20 mL of absolute ethanol solution containing ~ 497.26 mg p-benzoquinone. The BiOI film (1 cm × 1 cm) was obtained by electrodeposition at −0.1 V vs Ag/AgCl. The resulting film was rinsed with absolute ethanol carefully and dried at room temperature. ~ 75 µL of DMSO solution containing ~ 0.20 M VO(acac)2 was dropped uniformly on the BiOI film (1 cm × 1 cm), which was annealed at 450 οC for ~ 2 h in air. Then, the as-annealed BiVO4 electrodes were immerged in 1 M NaOH solution for ~ 20 min to remove the excessive V2O5, rinsed with deionized water and dried in air at room temperature.

FeP nanoparticles were synthesized by phosphorlysis of Fe3O4 prepared nanoparticles using a hydrothermal method [41]. In a typical synthesis procedure, ~ 1.35 g FeCl3 0.6H2O, ~ 0.72 g urea and ~ 25 mL of ethylene glycol (EG) were vigorously magnetic stirred at room temperature. Then, the mixtures were transferred into a ~ 50 mL autoclave for hydrothermal reaction at 200 °C for 3 h in a static state. The resulting black product was collected by centrifugation, washed with deionized water and absolute alcohol for three times, dried at 60 °C in vacuum overnight. Subsequently, ~ 0.10 g of the as-prepared Fe3O4 and ~ 0.50 g of NaH2PO2·H2O were mixed together and grounded. Then, the mixtures were calcined at 300 °C for 2 h under N2 atmosphere. The obtained black product was sufficiently washed with deionized water and dried at 80 °C ~ 24 h.

For comparison, Fe2P nanoparticles were prepared by phosphorlysis of FeCl3. In a typical synthesis procedure, ~ 1.0 g FeCl3 0.6H2O and ~ 3.0 g of NaH2PO2·H2O were mixed together and grounded. Then, the mixture was calcined at 300 °C for 2 h under N2 atmosphere, washed with deionized water and dried at 60 °C overnight.

 ~ 1 mg of FeP nanoparticles were ultrasonic dispersed in ~ 1 mL of absolute ethanol for ~ 15 min.Then a certain amount of the FeP suspension was drop-casted onto the surface of BiVO4 electrodes (1 cm × 1 cm), and dried at room temperature. Meanwhile, for comparison, Fe2P nanoparticles suspension was used as a cocatalyst replacing FeP to prepare Fe2P/BiVO4 photoanode composite under the same conditions.

2.3 Characterization

Power X-ray diffraction (XRD) patterns were obtained from X-ray diffraction (XRD, D/max–TTR III) using graphite monochromatized Cu Kα radiation of 1.54178 Å at a scanning rate of 5/min. The accelerating voltage and applied current were 40 kV and 200 mA, respectively. The scanning electron microscopy (SEM) images and energy-dispersed X-ray (EDX) analysis were performed using a JSM-6700F field emission scanning electron microscopy (FE–SEM). Transmission electron microscopy (TEM) images and high resolution transmission electron microscopy (HRTEM) images were obtained from a JEM-2010 electron microscopy, operated at an acceleration voltage of 200 kV. The UV–vis absorption spectra of the samples were carried out on a SOLID 3700 UV–vis spectrometer. X-ray photoelectron spectroscopy (XPS) were acquired with a Thermo ESCALAB 250 X-ray photoelectron spectroscopy instrument.

2.4 Photoelectrochemical Measurements

Photoelectrochemical performances of photoanodes were measured in a typical three-electrode configuration with an Ag/AgCl (~ 4 M KCl solution) reference electrode and a platinum foil counter electrode. The simulated solar illumination was obtained from a 300 W Xenon arc lamp equipped with an AM 1.5G filter, and the power intensity of the incident light reached at the surface of the working electrodes was carefully calibrated to 100 mW/cm2. 1 M potassium borate buffer solution was used as the electrolyte for photoelectrochemical measurements. For sulfite oxidation, an additional 0.2 M Na2SO3 solution was added into the electrolyte as a hole scavenger. Back-side illumination with exposed area of 1 cm2 (1 cm × 1 cm) was carried out in a glass cell with ~ 40 mL of electrolyte. Photocurrent-potential curves were obtained on an electrochemical workstation (Model CHI660E, purchased from Shanghai Chen Hua Instrument Co., Ltd.) with a scan rate of 10 mV/s. The recorded potential versus Ag/AgCl was converted against RHE according to the Nernst equation \(E_{RHE}=E_{Ag/AgCl}+0.059\times pH+0.197\).

Applied bias photon-to-current efficiency (ABPE) can be calculated using the following equation: \(ABPE=((1.23-V_{RHE})\times\frac{J_{light}-J_{dark}}{P_{light}})\times100\%\), where \(V_{RHE}\) represents the applied potential vs RHE, \(J_{light}\) and \(J_{dark}\) are the measured photocurrent and dark current density (mA/cm2), respectively. \(P_{light}\) (100 mW/cm2) is the power density of AM 1.5 G.

Electrochemical impedance spectra (EIS) were measured in 1 M potassium borate (pH = 9) by applying a voltage amplitude of 5 mV with the frequency range from 100 kHz to 0.01 Hz under the open-circuit potential and AM 1.5G illumination. The Mott–Schottky plots were obtained in 1 M potassium borate (pH = 9) at a preset frequency of 1000 Hz using an AC amplitude of 10 mV in the dark.

3 Results and Discussion

3.1 XRD Analysis

Pristine BiVO4 electrodes were synthesized by electrodeposition and annealing in air described in previous literature. The FeP nanoparticles were prepared from Fe3O4 via a phosphorization process and loaded onto the nanoporous BiVO4 electrodes by a drop-casting method.The crystalline phases composition of pristine BiVO4, pristine FeP nanoparticle along with the FeP/BiVO4 composites were analyzed by X-ray diffraction. Figure 1a shows the XRD patterns of Fe3O4 sample and the corresponding FeP. After phosphorization, the diffraction peaks located at 22.9°, 30.8°, 32.7°, 34.5°, 37.2°, 45.5°, 46.3°, 47.1°, 48.3° and 50.4° were observed, which were indexed to the (101), (002), (112), (011), (200), (112), (200) (202), (211) and (103) planes of orthorhombic FeP (PDF#78–1443), respectively. Meanwhile, the crystal texture of the as-prepared BiOI and BiVO4 photoanodes were further confirmed by XRD patterns (Fig. 1b). As shown in Fig. 1b, all the peaks can be indexed to tetragonal BiOI (PDF#10–0445) and monoclinic BiVO4 (PDF#14–0688), respectively [27]. However, no peak attributed to FeP was observed from the XRD pattern of FeP/BiVO4 composites, which may be due to its low loading amount of FeP sample.

Fig. 1
figure 1

XRD patterns of Fe3O4, FeP, BiOI, BiVO4 together with FeP/BiVO4

Full size image

3.2 Microstructures and Compositions Analysis

The microstructures of the photoelectrodes composites were investigated using SEM and TEM characterization. Figure 2a displays the BiOI nanoplates with a thickness of ∼20 nm were possessed with sufficient voids.Meanwhile, as shown in Fig. 2b, c, the nanoporous network of BiVO4 photoelectrodes were confirmed by the top-view and side-view SEM analysis respectively. Compared with pure BiVO4, Fig. 2d unambiguously reveal FeP nanoparticles, marked as the red circle, were successfully deposited on BiVO4 surface.In order to furtherly confirm the existence of FeP nanoparticles, as seen in Fig. 2f, the EDX analysis of FeP/BiVO4 reveals the presence of Co, P, Bi, V and O elements, implying the FeP nanoparticles were successfully loaded onto the BiVO4 photoelectrode. Meanwhile, TEM analysis was used to investigate the FeP/BiVO4 composite. As shown in Fig. 2e, it is also found that FeP nanoparticles were deposited on BiVO4 surface.

Fig. 2
figure 2

Compositions and morphologies of BiOI, BiVO4 together with FeP/BiVO4 composites

Full size image

3.3 XPS Analysis

With aim to further investigate the valence state and chemical composition of FeP/BiVO4 photoanode, the XPS spectra of FeP/BiVO4 were examined. The XPS survey spectrum of FeP/BiVO4 showed the sample surface presence of Fe, P, Bi, C, V and O elements (Fig. 3a), as well as the O element is attributed to surface partly oxidation of FeP, and the C element derived from the reference. Furthermore, the high resolution XPS spectrum of Bi 4f (~ 159.4 eV and ~ 164.5 eV) and V2p (~ 516.8 eV and ~ 524.5 eV) were well matched to the reported values of monoclinic BiVO4 (Fig. 3b, c).As shown in Fig. 3d, the two peaks at ~ 130.5 eV and ~ 129.6 eV, which can be assigned to the low binding energy of P 2p1/2 and P 2p3/2 in FeP, respectively [31]. Meanwhile, the spectrum of P 2p also exhibits a peak with high binding energy of ~ 133.8 eV, which may be due to partial oxidation of P element, such as phosphate. Furthermore, as shown in Fig. 3e, the two peaks at ~ 709.1 eV and ~ 717.5 eV likely corresponding to the Fe 2p3/2 and Fe 2p1/2 peaks in FeP, respectively [42]. The other two peaks at binding energies of ~ 714 eV and ~ 724.5 eV, which arises from the iron oxide [43]. Hence, these above results further reveal that FeP cocatalyst nanoparticles were successfully deposited on surface of the worm-like nanoporous BiVO4.

Fig. 3
figure 3

a XPS survey spectra and high-resolution XPS spectra of. b Bi 4f. c V2p. dP 2p, and f Fe 2p of the FeP/BiVO4 sample

Full size image

3.4 UV–Vis Analysis

In order to gain the light harvesting properties of pritine BiVO4 and FeP/BiVO4 hybrid, the UV–vis diffuse reflectance spectroscopy technique were investigated. As demonstrated in Fig. 4a, the UV–vis absorption spectra of BiVO4 and FeP/BiVO4 exhibited the identical absorption edge at around 525 nm, which are consistent with the optical band gap energy of ~ 2.47 eV (Fig. 4b), manifesting that FeP cocatalyst nanoparticles were not doped into the monoclinic BiVO4 crystal lattice to without change its band gap.

Fig. 4
figure 4

a The optical absorption spectroscopy of the bare BiVO4 and FeP/BiVO4; b Tauc plot of bare BiVO4 and FeP/BiVO4

Full size image

3.5 Photoelectrochemical Water Oxidation

All the PEC performances of photoanodes in this study were acquired according to back-side illumination, which was due to the poor electron transport property and long electron transfer distance of the relative thick photoanode under front-side illumination. As exhibited in Fig. 5b. The pristine BiVO4 exhibited a photocurrent density of ~ 1.88 mA/cm2 at 1.23 V vs. RHE and the cathodic shift (∼430 mV) in the onset potential. The maximum photocurrent density of FeP/BiVO4 photoanode hybrid for PEC solar water oxidation could reached ~ 3.05 mA/cm2 under optimal amount of FeP nanoparticles(Fig. 5b), which is ~ 1.6 times higher than that of the unmodified pritine BiVO4.

Fig. 5
figure 5

Photoelectrochemical performances of photoanodes under differen conditions

Full size image

So as to identify the surface reaction kinetics of BiVO4-based photoanodes, 0.2 M Na2SO3 solution was served as a hole scavenger. As shown in Fig. 5c, an obvious distinction of BiVO4 between Na2SO3 solution and water oxidation under dark condition. The poor water oxidation kinetic of BiVO4 photoanode, resulting in the serious surface charge recombination. However, after the loading of FeP nanoparticles, a dramatic cathodic shift of onset potential for water oxidation compared to the curve of bare BiVO4 was observed, indicating the FeP is an active dark electrocatalyst for oxygen evolution. Meanwhile, Fig. 5d displays an identical PEC sulfite oxidation property of BiVO4 and FeP/BiVO4, indicating the similar charge separation efficiency in the bulk upon loading of FeP nanoparticles. Thus, the significant suppression of the surface recombination by loading of FeP nanoparticles could improve PEC property of solar-driven water oxidation. For comparation, the Fe2P nanoparticles were used as cocatalyst, which exhibits a lower photocurrent density than that of FeP nanoparticles, highlighting FeP nanoparticles is an active cocatalyst for solar water splitting (Fig. 5e).

As demonstrated in Fig. 5f, before FeP modification, the ηsurface of BiVO4 photoanode is ~ 40% at 1.23 V vs. RHE, indicating more than half of the photogenerated holes are captured by charge recombination at the surface of pristine BiVO4 photoanode. Interesting phenomenon is the ηsurface of FeP/BiVO4 could reached ~ 57% at 1.23 V vs RHE, which was much higher than that of the pure BiVO4 photoanode. Base on above analysis, FeP nanoparticles could efficiently decrease the surface recombination losses at the surface of BiVO4 photoanodes.

3.6 Photoelectrochemical Mechanism

To further investigated the interface charge transport behavior of the photoanodes charge carries, the photocurrent transients of BiVO4 and FeP/BiVO4 photoanodes were recorded in Fig. 6a. During each light on–off cycle process, incisive photocurrent transient spike was observed for bare BiVO4 photoanode, indicating partly surface charge recombination. Meanwhile, the maximum photocurrent density of FeP/BiVO4 for PEC water oxidation could reached ~ 2.85 mA/cm2 at 1.23 V vs RHE. The Mott–Schottky (MS) curves of the photoelectrodes were shown in Fig. 6b. Compared with bare BiVO4, the FeP/BiVO4 composites exhibited the identical flat band potential. Thus, the dramatically enhanced PEC properties of FeP/BiVO4 hybrid may be attributed to improve surface charge transfer efficiency. Furthermore, the electrochemical impedance spectroscopy (EIS) of BiVO4 and FeP/BiVO4 were performed. As shown in Fig. 6c, compared with naked BiVO4 photoanode, a much smaller semidiameter of the semicircle was observed for FeP/BiVO4 under open-circle potential and AM 1.5G illumination, indicating a relatively favorable interfacial charge transfer ability of FeP/BiVO4. in accordance with the quantification analysis results of the surface charge separation efficiency.

Fig. 6
figure 6

a The photocurrent transients of BiVO4 and FeP/BiVO4. b Mott–Schottky curves of BiVO4 and FeP/BiVO4 c EIS curves for BiVO4 and FeP/BiVO4

Full size image

The maximum applied bias photon-to-current efficiency (ABPE) for FeP/BiVO4 was calculated to be ~ 0.53% at 0.6 V vs RHE (Fig. 7a), which was much higher than that of the unmodified BiVO4. Moreover, the PEC stability of the FeP/BiVO4 photoanode was conducted in 1 M potassium borate solution at 0.6 V vs RHE. As shown in Fig. 7b, the photocurrent density gradually decreased during the consecutive light illumination in ~ 30 min, probably on account of the slowly drop of FeP nanoparticles into solution or the intrinsic instability of the FeP/BiVO4 photoanode.However, the photocurrent density gradually increased after ~ 30 min light illumination, which was probably resulted from the other active substance (For example FeOOH).

Fig. 7
figure 7

a ABPE curves for BiVO4 and FeP/BiVO4. b Stability test carried out for FeP/BiVO4 at a constant potential of 1.23 V vs RHE

Full size image

4 Conclusion

In summary, decoration of BiVO4 photoanode surface with a novel FeP nanoparticles has achieved a improving photocurrent density for PEC water splitting for the first time, which is ~ 1.6 times higher than that of the bare BiVO4 photoanode. The highly efficient suppression of surface recombination through FeP modification was also certified by PEC measurements using 0.2 M Na2SO3 solution as a hole scavenger. Meanwhile, the maximum ABPE of FeP/BiVO4 photoanodes was ~ 0.53% at 0.6 V vs RHE. Furthermore, the FeP/BiVO4 photoanode was prepared by inexpensive noble-metal-free materials, will provide insights on the design and development of efficient water splitting BiVO4-based photoanodes.