Enhanced photoelectrochemical water oxidation over a surface-hydroxylated BiVO₄ photoanode: advantageous charge separation and water dissociation

Abstract

Bismuth vanadate (BiVO₄) is one of the most effective photoanode materials for photoelectrochemical water splitting, but its reaction rate is greatly limited by the poor separation efficiency of photo-generated charges and sluggish kinetics. In this work, a surface hydroxylation strategy was carried out to overcome the above-mentioned shortcomings. A simple post-synthetic NaOH immersion method was used to successfully hydroxylate the BiVO₄ photoanode surface without affecting its bulk properties. The modified surface hydroxyl group not only promotes the photo-generated charge separation efficiency of hydroxylated BiVO₄ but also tailors its local electronic structure to form high-valence Bi^(3+x)+ states. The Bi^(3+x)+ species serving as the active site is advantageous for water dissociation, thus decreasing the free energy barrier of the potential-determining step for the boosted overall water oxidation kinetics. As a result, the best hydroxylated BiVO₄ photoanode achieves a higher photocurrent density of 1.14 mA cm⁻² at 1.23 V vs. RHE under visible light irradiation, which is 4.75 times that of bare BiVO₄ (0.24 mA cm⁻²), and moreover, a cathodic shift of 190 mV at 1 mA cm⁻² is obtained. Our work affords a practical method with strong operability to improve the intrinsic photoelectrochemical performance of photoanodes by the efficient surface hydroxylation strategy.

---

Introduction

The environmental pollution and the impending energy shortage caused by massive consumption of fossil fuels have serious impacts on human survival.^1,2 It is an urgent requirement to develop renewable sources of green energy. Since Fujishima and Honda described the process of photoelectrochemical (PEC) water splitting to generate hydrogen (H₂) and oxygen (O₂) in 1972,³ harnessing solar energy in the form of green chemical fuel H₂via PEC water splitting has been recognized to be one of the most promising methods.^4,5 Generally, the PEC performance for H₂ generation is determined by a water oxidation half reaction due to its four-charge process.⁶ Therefore, the search for efficient and cheap photoanode materials has led scientists to explore various kinds of semiconductors.^7–10 Among them, monoclinic bismuth vanadate (BiVO₄) has been studied as a promising visible-light-driven photoanode for solar energy conversion, because of its low cost, narrow band gap (∼2.4 eV) and suitable band-gap structure for water oxidation.^11,12 However, its severe photo-generated charge recombination and sluggish oxidation kinetics have hampered its further application in water splitting.¹³

It is well known that PEC water oxidation is a multi-step and four-charge surface reaction process as follows:

H₂O + * + h⁺ → OH* + H⁺ (1)

OH* + h⁺ → O* + H⁺ (2)

O* + H₂O + h⁺ → OOH* + H⁺ (3)

OOH* + h⁺ → *O₂ + H⁺ (4)

*O₂ → * + O₂ (5)

where * acts as the active site, and *OH, *O and *OOH are the adsorbed oxygen intermediates.^6,14 Therefore, its reaction rate is closely related to the efficient separation and utilization of photo-generated charges, which can be influenced by modifying the electronic structure of the active site on the surface of photoanodes.^15,16 Moreover, many studies have shown that the speed of such surface reaction kinetics is largely determined by the adsorption, decomposition, and conversion of raw and intermediate reactants, which is also connected with the surface property of photoanodes.¹⁷ Therefore, developing an effective surface strategy for the BiVO₄ photoanode to enhance its water oxidation activity is of great significance.

Recently, surface hydroxylation has been put forward as an efficient surface modifying strategy for enhancing the water oxidation activity of catalysts. For example, Li et al. reported that hole transfer can be facilitated by enriching the α-Fe₂O₃ surface with hydroxyl groups via a sonication step, which had no change in the bulk properties of photoanode materials.¹⁸ Pei et al. found that with the hydroxyl group modification, the electrocatalytic activity of graphene-supported single atomic catalysts (Co and Fe) can be improved by the stable adsorption of intermediates (OH*, O* and OOH*) via a density functional theory study.¹⁹ Furthermore, this modification strategy avoids the problems of blocking light or creating additional interfacial traps that exist in an overcoated layer.^20,21 More importantly, as for BiVO₄ semiconductors, the hydroxyl group engineering can directly oxidize [Bi–O]⁻ or [V–O]⁻ to [Bi–O] or [V–O]˙, which is beneficial for the initiation step for water photooxidation.²² In addition, as a changeable valence metal element, Bi has an electronic structure that could be easily tailored under the action of surface groups and then enhance the reactivity towards water adsorption and oxygen-related reactions.²³ Therefore, surface hydroxylation is a suitable strategy for the promotion of the water oxidation property of BiVO₄, and it is necessary to expound the function of surface hydroxylation. However, there is no relevant report.

Hence, we first in situ synthesized a BiVO₄ film on a metal Bi substrate and then directly soaked it in a NaOH solution to form the surface-hydroxylated BiVO₄ photoanode. The hydroxylated BiVO₄ photoanode obtained by controlling the appropriate reaction conditions exhibited superior PEC water oxidation activity and stability under visible light irradiation, indicating that this facile treatment method is practically feasible. Via the structural, morphological and optical characterizations, photoelectrochemical measurements and density functional theory (DFT) calculations, the connotation of the surface hydroxylation strategy to improve the reaction rate of BiVO₄ was further clarified. Based on the explored essential relationship between surface behaviour and water oxidation activity, the PEC reaction mechanism of the hydroxylated BiVO₄ photoanode was finally proposed, which would provide valuable reference for the subsequent research on water oxidation.

Experimental

Reagents and chemicals

Ammonium metavanadate (NH₄VO₃) was purchased from Aladdin Company. Sodium hydroxide (NaOH) and nitric acid (HNO₃) were purchased from Tianjin Kermel Chemical Reagent Co., Ltd. All reagents were of analytical grade and used without further purification. Bi substrates were purchased from Wochang Metal Products Co., Ltd and used after further treatment.

Preparation of surface-hydroxylated BiVO₄ photoanodes

I. Treatment of Bi substrates. A metal Bi substrate (1.9 cm × 1.0 cm × 0.2 cm) was repeatedly polished using a piece of abrasive paper until the oxide layer on the surface was removed, washed several times with distilled water, dried naturally at room temperature, and then kept for later use.

II. Preparation of bare BiVO₄ photoanodes. A bare BiVO₄ photoanode was in situ synthesized on a metal Bi substrate by a facile hydrothermal process according to our previous work.²⁴ First, 2 mmol NH₄VO₃ was dissolved into 9 mL of 2 M HNO₃, and 3.6 mmol NaOH was dissolved into 25 mL of deionized water at room temperature. After mixing, they were slowly poured into a 50 mL Teflon-lined autoclave with a pretreated Bi substrate. The autoclave was sealed and kept at 160 °C for 90 min, and then naturally cooled down to room temperature. The bare BiVO₄ photoanode was obtained after washing several times with deionized water and drying naturally at room temperature.

III. NaOH treatment of the bare BiVO₄ photoanode. The surface hydroxylation treatment of the bare BiVO₄ photoanode was performed by simple NaOH solution immersion under certainly controlled conditions. The resulting sample was washed with deionized water and dried overnight under atmosphere, and then the surface-hydroxylated BiVO₄ photoanode was obtained. To explore the optimal treatment condition, the bare BiVO₄ photoanodes were immersed in different NaOH solutions (0.5/1.0/2.0/4.0 M) for different time periods (10/30/60/120 minutes), and the obtained surface-hydroxylated BiVO₄ photoanodes were named BiVO₄-X–Y (X refers to the concentration of the NaOH solution and Y refers to the immersion time).

Characterizations and photoelectrochemical tests

X-Ray diffraction (XRD) images of the samples were recorded using a Rigaku D/max 2700 diffractometer with Cu Kα (λ = 1.5418 Å) radiation operating at 40 kV and 30 mA. Scanning electron microscopic (SEM) images were acquired using a JEOL JSM-7001F at an accelerating voltage of 10 kV. Fourier transform infrared (FTIR) spectra were recorded using a Shimadzu IRAffinity-1 FTIR spectrometer in the frequency range of 4000–500 cm⁻¹. Raman spectra were recorded from 50 to 1200 cm⁻¹ using a Renishaw inVia microscopic spectrometer. X-Ray photoemission spectroscopy (XPS) and valence band (VB)-XPS were performed using a VG MultiLab 2000 system with an Mg Kα X-ray source operating at 20 kV. Photoluminescence (PL) spectra were recorded using a FLS980 Series of Fluorescence Spectrometer.

Photoelectrochemical tests were performed using an Admiral Squidstat Plus electrochemical system (America) with a three-electrode glass cell in a 0.1 M Na₂SO₄ electrolyte solution (pH ≈ 6.5). A 300 W Xe lamp (PLS-SXE 300 +, Perfect Light, China) equipped with a 420 nm filter (λ > 420 nm) was used as the simulated visible light source with a light intensity of 120 mW cm⁻². The reference electrode was a standard calomel electrode (SCE), the counter electrode was a Pt wire and the working electrode was the prepared BiVO₄ photoanode. All the potential vs. RHE data presented in this study were calculated according to the Nernst equation: E_RHE = E_SCE + 0.0591 pH + 0.24 V. Linear scan voltammetries (LSVs) were performed at a scan rate of 10 mV s⁻¹. Cyclic voltammetric (CV) curves were recorded from 0.45 to 0.8 V vs. RHE at scan rates of 10, 20, 30, 40 and 50 mV s⁻¹. Chrono-amperometric (I–t) measurements were conducted at 1.23 V vs. RHE, and the generated H₂ and O₂ gases were analyzed using a gas chromatograph (Panna A91 Plus, 5 Å molecular sieve column and Ar carrier). Mott–Schottky curves were obtained at a fixed frequency of 2 kHz. Electrochemical impedance spectroscopy (EIS) measurements were performed in the frequency range of 0.1 Hz to 10⁵ Hz at an amplitude of 10 mV.

DFT calculations

All calculations were based on DFT (density functional theory) using the CASTEP (Cambridge sequential total energy package) module in Materials Studio.²⁵ The generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) functional was used for the exchange–correlation interaction.²⁶ In all calculations, a DFT+D empirical correction was taken to describe the van der Waals interaction.²⁷ The plane wave cut-off energy was set to 450 eV. The Monkhorst-Pack k-point grids (2 × 2 × 1) were utilized to approximate the Brillouin-zone integrations. Total energy, force convergence threshold and maximum displacement were less than 10⁻⁵ eV per atom, 0.02 eV Å⁻¹ and 5.0 × 10⁻⁴ Å, respectively. The calculation model of bare BiVO₄ was built based on the BiVO₄ (010) surface. The BiVO₄ (010) slab was modeled by a 2 × 2 × 1 supercell, and its size is as follows: a = 10.2869 Å, b = 10.2777 Å. For surface-hydroxylated BiVO₄, the one hydroxy group was bonded to a Bi atom on the BiVO₄ (010) slab. A vacuum layer thickness of 12 Å was set to avoid the influence between neighbouring layers in the slab model. The reaction Gibbs free energy (ΔG) for each step was calculated by the expression: ΔG = ΔE + ΔE_ZPE − TΔS, where ΔE, ΔE_ZPE, and ΔS represent the total energy, zero-point energy, and entropy, respectively.

Results and discussion

In this research, the surface-hydroxylated BiVO₄ film was successfully developed on the metal Bi substrate (Fig. S1, ESI†) by a facile hydrothermal-NaOH immersion technique. The schematic diagram of its preparation process is shown in Fig. 1a. The morphology, structure and composition of the as-prepared samples were studied by various characterizations. The XRD analysis characterizes the crystalline structure of the prepared BiVO₄ samples. The detectable patterns of bare BiVO₄ and hydroxylated BiVO₄-X–Y samples are both well consistent with that of monoclinic BiVO₄ (JCPDS No. 14-0688), proving the successful synthesis of BiVO₄ photoanodes (Fig. S2, ESI†).²⁸ The SEM and TEM images show that BiVO₄ and BiVO₄-1–30 photoanodes have the same dendritic morphology with a thickness of 50–52 μm (Fig. 1b–f). The EDS mapping of BiVO₄-1–30 verifies the uniformly distributed Bi, V and O spots (Fig. 1g–i). In addition, via N₂ adsorption–desorption isotherms test, the similar specific surface area and pore average diameter are found for BiVO₄ and BiVO₄-1–30 (Fig. S3, ESI†). These phenomena manifest that the post-synthetic NaOH immersion treatment is effective without affecting the bulk properties of the BiVO₄ photoanode.

Fig. 1 (a) Schematic diagram of the preparation process of surface hydroxylated BiVO₄ photoanode. (b) Top-view and (c) cross-sectional SEM images of bare BiVO₄, (d) top-view and (e) cross-sectional SEM images, (f) TEM image, and EDX mappings of (g) Bi, (h) V and (i) O elements of BiVO₄-1–30 photoanode.

The FTIR characterization was used to investigate the surface composition of BiVO₄ and BiVO₄-1–30 samples. As shown in Fig. 2a, the broad regions peaked at 723 and 502 cm⁻¹ could be attributed to the vibrations of V–O and Bi–O bonds, respectively. The peak at 1608 cm⁻¹ is assigned to H–O–H of the adsorbed H₂O in samples.²⁹ Specially, it is clearly found that BiVO₄-1–30 has a broader and stronger spectrum than that of the bare BiVO₄ sample around the 3445 cm⁻¹ region (purple labeled in Fig. 2a), which is ascribed to the stretching vibrations of O–H groups chemisorbed on the surface of catalyst.^30,31 This result provides evidence for the BiVO₄-X–Y photoanode with a hydroxylated surface obtained by the treatment of the NaOH solution.

Fig. 2 (a) FTIR and (b) Raman spectra of bare BiVO₄ and BiVO₄-1–30 photoanodes.

The Raman spectroscopy further offered an intuitive observation of the chemical bonding vibrations in BiVO₄ and BiVO₄-1–30 samples. Fig. 2b displays that the two presented samples have the same vibrational bands around 126.7 cm⁻¹, 212.2 cm⁻¹, 330.3 cm⁻¹, 364.2 cm⁻¹, 711.8 cm⁻¹ and 825.2 cm⁻¹.^32,33 However, the peak associated with the vibrational mode of the Bi–O bond shows difference, where it is at 54.8 cm⁻¹ for BiVO₄ but shifts to 57.3 cm⁻¹ for BiVO₄-1–30 (inset of Fig. 2b, purple labeled).³⁴ One possible explanation for this change is that OH⁻ anions are inclined to adsorb at Bi³⁺ cations with the electrostatic adsorption effect to form a hydroxylated surface; hence, the vibration of the Bi–O bond in BiVO₄-1–30 is influenced.^35,36

Furthermore, the detailed information on the elemental composition and chemical valence states of the samples are obtained by XPS. As shown in Fig. 3a, the full-scale XPS spectra reveal the presence of Bi 4f, O 1s and V 2p. The C 1s signal peak at a binding energy of 284.6 eV is applied to calibrate the relevant XPS peak position. In Fig. 3b, the peak at around 524.13 eV is the characteristic peak of Na Auger,³⁷ which is due to the residual Na element in the preparation process of the BiVO₄ and surface-hydroxylated BiVO₄ photoanodes.³¹ The identical V 2p high resolution spectra were found for BiVO₄ and BiVO₄-1–30 photoanodes, indicating the unchanged bonding environment of the V⁵⁺ element with a hydroxylated surface.³⁸ For the high-resolution Bi 4f spectra, the binding energies of Bi 4f_7/2 and Bi 4f_5/2 in bare BiVO₄ are located at 159 eV and 164.3 eV, respectively; however, the Bi 4f peaks of BiVO₄-1–30 exhibit higher binding energies shifting to 159.31 eV and 164.61 eV, respectively (Fig. 3c).³⁹ It means that the higher valence Bi^(3+x)+ state was obtained due to the strong electronic interaction between the strongly electronegative OH⁻ group and Bi³⁺ site.⁴⁰ Concomitantly, a negative shift for O 1s peaks in BiVO₄-1–30 is revealed (Fig. 3d).³² These demonstrations obviously confirm that the surface-hydroxylated BiVO₄ photoanode was successfully prepared, and the modified hydroxyl groups could effectively regulate the surface states of BiVO₄ photoanodes for a beneficial water oxidation reaction.²³

Fig. 3 (a) XPS spectra, high-resolution XPS spectra for (b) V 2p, (c) Bi 4f, and (d) O 1s of bare BiVO₄ and BiVO₄-1–30 photoanodes.

The linear sweep voltammetry (LSV) method was employed to investigate the PEC water oxidation performance of the prepared BiVO₄ photoanodes in the three-electrode cell. As shown in Fig. 4a and b, the current densities of both bare BiVO₄ and BiVO₄-1–30 photoanodes are ignorable in the darkness, illustrating that the surface hydroxyl group has no obvious effect without the photo-generated charges. Under the visible light illumination, all of the prepared BiVO₄ photoanodes display water oxidation activity, and the activities are greatly increased with the treatment of NaOH. The minimum overpotential (at 1 mA cm⁻²) was obtained with 1 M NaOH treatment for 30 min (BiVO₄-1–30 sample), which decreases by ∼190 mV with respect to bare BiVO₄ (Fig. 4a). The current density of BiVO₄-1–30 attains 1.14 mA cm⁻² at 1.23 V vs. RHE, being ∼4.75 times than that of bare BiVO₄ achieving 0.24 mA cm⁻² (Fig. 4b). For the applied bias photon-to-current efficiency (ABPE) estimation, eqn (E1) (ESI†) was applied. As shown in Fig. 4c, the maximum ABPE of the BiVO₄-1–30 photoanode is ∼0.049% at 1.09 V vs. RHE; however, the maximum ABPE of bare BiVO₄ is only ∼0.014% at 1.05 V vs. RHE. Furthermore, Fig. S4 (ESI†) exhibits that the maximum ABPE of the BiVO₄-1–30 photoanode (∼0.045% at 1.09 V) is also higher than that of bare BiVO₄ (∼0.009% at 1.12 V) in the two-electrode configuration, which reflects that the system with the BiVO₄-1–30 photoanode is with better overall water splitting performance.

Fig. 4 LSV curves of the as-prepared samples (a) treated with different concentrations of NaOH solution for 30 min and (b) immersed in 1 M NaOH solution for different durations, (c) ABPE and (d) the photocurrent stability curves of bare BiVO₄ and BiVO₄-1–30 photoanodes in the three-electrode cell.

The PEC water oxidation stability of the samples was evaluated by chronoamperometry at 1.23 V vs. RHE. It shows that BiVO₄-1–30 keeps around 1.08 mA cm⁻² after testing for 5 h, demonstrating its excellent PEC stability (Fig. 4d). Importantly, the photocurrent of BiVO₄-1–30 is always higher than that of the bare BiVO₄ sample. Meanwhile, the H₂ and O₂ gases generated under the same conditions were also measured. As shown in Fig. S5 (ESI†), the total amount of H₂ and O₂ gases after 5 h was measured to be 117.69 and 58.85 μmol, respectively, by using the BiVO₄-1–30 photoanode, which are higher than that measured by the bare BiVO₄ one as 27.66 and 13.83 μmol, respectively,. Taken together, the results reveal that the surface hydroxylation strategy by a simple NaOH treatment could greatly improve the PEC water oxidation property of BiVO₄ photoanodes, hence enhancing the overall water splitting performance.

The UV-vis diffuse reflectance spectra (DRS) display that both bare BiVO₄ and BiVO₄-1–30 photoanodes have excellent absorption in the visible light region (Fig. S6, ESI†). To evaluate the solar energy conversion capability of prepared photoanodes, the incident photon-to-current conversion efficiency (IPCE) was performed from the region of 420–700 nm excitation wavelengths at 1.23 V vs. RHE. The IPCE value was figured out using eqn (E2) (ESI†). The BiVO₄-1–30 reveals the highest IPCE of 16% at 420 nm, whereas the IPCE value of bare BiVO₄ is only around 3.69% at 420 nm (Fig. 5a). The factors of light harvesting efficiency (LHE), bulk charge separation (η_sep) and surface injection (η_inj) efficiencies are closely linked to IPCE.⁴¹ According to the DRS data, the LHE was calculated using eqn (E3) (ESI†). It can be found the two samples display similar light harvesting ability (Fig. 5b), stating that the surface hydroxylation has no effect on LHE. Fig. S7 (ESI†) shows the LSV curves, which were measured in the electrolytes of 0.1 M Na₂SO₄ with/without 0.1 M Na₂SO₃ (hole scavenger). Based on the LSV curves, the η_sep and η_inj were estimated using eqn (E6) and (E7) (ESI†), respectively. It is well known that η_sep is to test the migration capacity of holes in the photoanode phase to the photoanode/electrolyte interface, and η_inj reveals the efficacy of holes on the photoanode/electrolyte interface to participate in the water oxidation process.⁴² As Fig. 5c shows, η_sep of bare BiVO₄ at 1.23 V vs. RHE is only 16.45%, while η_sep of BiVO₄-1–30 increases to 28.7%. Moreover, compared with bare BiVO₄, η_inj of BiVO₄-1–30 is greatly improved. The η_inj value of BiVO₄-1–30 at 1.23 V vs. RHE has reached 53.16%, which is much higher than that of bare BiVO₄ which is 19.22% (Fig. 5d). The enhanced η_sep and η_inj values confirm the significantly promoted separation and utilization efficiency of photo-generated holes by surface hydroxylation, hence enhancing the PEC water oxidation activity of the BiVO₄-1–30 photoanode.

Fig. 5 (a) IPCE plots, (b) light harvesting efficiency (LHE), (c) the charge separation efficiency (η_sep) and (d) charge injection efficiency (η_inj) of bare BiVO₄ and BiVO₄-1–30 photoanodes.

In order to further explore the reason for the increased PEC water oxidation properties of the hydroxylated BiVO₄ photoanode, the electrochemical impedance spectroscopy (EIS) was investigated, in which the arc radius reflects the charge transfer resistance.⁴³ Compared with bare BiVO₄ photoanode, the arc radius of BiVO₄-1–30 is smaller, which indicates that the hydroxylated surface reduces the charge transfer resistance for the improved water oxidation kinetics (Fig. 6a).³⁸ Furthermore, the transient photocurrent response (PC) and photoluminescence (PL) curves were applied to investigate the separation condition of photo-generated carrier. The relatively higher PC spikes (Fig. 6b) and lower PL curve (Fig. 6c) of BiVO₄-1–30 than those of bare BiVO₄ indicate that the hydroxylated surface can assist faster charge separation, effectively reduce the recombination of carriers, thus making more holes to participate in the water oxidation reaction.⁴⁴ At last, the cyclic voltammograms (CV) (Fig. S8, ESI†) were applied to investigate the electrochemically active surface area plots (ECSA). As Fig. 6d shows, the BiVO₄-1–30 sample displays a bigger linear slope of 3.59 mF cm⁻² than that of bare BiVO₄ of 1.6 mF cm⁻², which reflects the capacitance and is proportional to the ECSA.⁴⁵ In other words, BiVO₄-1–30 possesses an increased ECSA, so that it could provide more active sites for PEC water oxidation to improve the activity.

Fig. 6 (a) EIS, (b) photocurrent response, (c) PL spectra and (d) charging current densities recorded at 0.625 V_RHE at different scan rates of bare BiVO₄ and BiVO₄-1–30 photoanodes.

The performance of photoanodes is closely related to its energy band structure. Therefore, it was studied to explore the effect of the hydroxylated surface on the photo-induced carrier mechanism. The DRS, valence band XPS (VB-XPS) and Mott–Schottky test were performed to determine the band gaps, valence band (VB) and conduction band (CB) positions. Plotted by the DRS spectra, Fig. 7a exhibits the band gap energy values of 2.38 and 2.4 eV for bare BiVO₄ and BiVO₄-1–30 photoanodes, respectively. The VB-XPS spectra (Fig. 7b) show that the energy gaps are 1.56 and 1.72 eV for bare BiVO₄ and BiVO₄-1–30, respectively, which represent the difference in values between the VB and Fermi level (E_f).⁴⁶ The value of flat band potential (E_FB) is nearly equal to that of E_f.⁴⁷ Calculated by Mott–Schottky plots, the E_FB values of two samples are 0.20 and 0.15 V vs. SCE (Fig. 7c), i.e. 0.42 and 0.37 V vs. NHE (normal hydrogen electrode), respectively.⁴⁶ Consequently, the VB positions are equal to 1.98 and 2.09 V, respectively. Furthermore, based on the band gap energies, the CB positions are −0.40 and −0.31 V of bare BiVO₄ and BiVO₄-1–30 photoanodes, respectively. The diagram of the band energy structure is shown in Fig. 7d. It is directly observed that BiVO₄-1–30 with the hydroxylated surface possesses a positive shift of VB position. The more positive position of VB is advantageous for the generation of photo-induced holes with more powerful oxidizing ability.⁴⁸ Moreover, the flat potential value decreases from 0.44 V to 0.39 V vs. NHE after the surface hydroxylation. The negative shift of the flat potential is regarded to increase the band bending, thus promote the interfacial hole transfer of the hydroxylated BiVO₄ photoanodes.⁴⁹

Fig. 7 (a) The optical band gap estimation, (b) VB-XPS spectra, (c) Mott–Schottky plots, and (d) schematic band structure diagram of bare BiVO₄ and BiVO₄-1–30 photoanodes.

Density functional theory (DFT) calculations were carried out to deeply study the reason for the enhanced water oxidation activity of hydroxylated BiVO₄ photoanodes. The optimized models are shown in Fig. S9 (ESI†). It can be seen that the hydroxyl group tends to adsorb at the Bi cation site, which is consistent with the change of Bi in Raman and XPS results. The charge density difference of hydroxylated BiVO₄ with regard to BiVO₄ was simulated. It clearly reveals that the Bi atom donates its electrons mainly to nearby O atoms and the hydroxyl group via orbital hybridization (Fig. 8a). Namely, the electrons have an active tendency to transfer from BiVO₄ to the surface hydroxyl group, and thus, holes could efficaciously separate from them on the hydroxylated BiVO₄ photoanode surface for water oxidation reactions.⁵⁰ Meanwhile, electron-deficient Bi, i.e. the high-valence Bi^(3+x)+ state, can be obtained, which is reported to be beneficial for water adsorption and oxygen-relevant reactions towards the boosted water oxidation kinetics.^23,51 Therefore, the relevant Gibbs free energy for each step in the overall water oxidation progress of bare BiVO₄ and hydroxylated BiVO₄ photoanodes was calculated (Fig. 8b). It was obviously that the first step (* + H₂O + h⁺ → *OH + H⁺) is the potential-determining step (PDS) for bare BiVO₄ with the biggest overpotential of 2.5 eV. However, with the modification of the hydroxyl group for hydroxylated BiVO₄, the energy barrier of this step is decreased to 1.26 eV, and the PDS changes to the second step (*OH + h⁺ → *O + H⁺) with an overpotential of 1.42 eV. Notably, the free energy change in each step for hydroxylated BiVO₄ is close to the ideal 1.23 eV, being more beneficial for proceeding the water oxidation process.⁵²

Fig. 8 (a) The simulated electron density distribution between BiVO₄ and the adsorbed hydroxyl group (the yellow and blue areas represent electron accumulation and deficiency, respectively), and (b) Gibbs free energy diagram for the overall water oxidation process of bare BiVO₄ and hydroxylated BiVO₄ photoanodes.

In detail, the first step (PDS: * + H₂O + h⁺ → *OH + H⁺) is the catalytic dissociation of the H₂O molecule with the generation of *OH. Hence, the reactant H₂O behavior on the photoanode surface was considered. As shown in Fig. 9, the H₂O molecule has a lower energy barrier of adsorption on hydroxylated BiVO₄ (−0.05 eV) than that on bare BiVO₄ (0.05 eV), indicating more favoured H₂O adsorption (* + H₂O → *H₂O). Hydroxylated BiVO₄ also exhibits a quite low barrier of 0.45 eV for *H₂O dissociation (*H₂O → *OH + *H); in contrast, a higher energy barrier is 0.61 eV on bare BiVO₄. These show the much easier dissociation of H₂O at the high-valence Bi^(3+x)+ active site, leading to a faster *OH supply. Meanwhile, the *H desorption (*OH + *H → *OH + H⁺) plays a very important role in the water dissociation process, in which the rapid *H desorption would shift the equilibrium toward water dissociation, thus the reaction rate can be increased.²² It can be found that hydroxylated BiVO₄ model exhibits a much lower energy barrier of 0.86 eV for *H desorption than that of bare BiVO₄ (1.84 eV), further proceeding the water dissociation kinetics.

Fig. 9 Gibbs free energy diagram of the detailed steps for *OH generation (insets of corresponding atomic configurations: including H₂O adsorption (*H₂O), *H₂O dissociation (*OH + *H), and *H desorption (*OH + H⁺)) of bare BiVO₄ and hydroxylated BiVO₄ photoanodes.

Based on the above-mentioned findings, a possible PEC water oxidation mechanism of the prepared BiVO₄ photoanode is proposed, as displayed in Fig. 10. When the BiVO₄ photoanode is excited by visible light, the photo-generated electrons (e⁻) in bulk migrate to the Pt cathode through the external circuit to reduce H⁺ to H₂, and the holes (h⁺) move to the photoanode/electrolyte reaction interface to oxidize water to O₂. Due to the sluggish kinetics of PDS step (* + H₂O + h⁺ → *OH + H⁺) and the massive recombination of e⁻–h⁺ on bare BiVO₄, the overall water oxidation process is hard to take place, while for the surface-hydroxylated BiVO₄ photoanode, the water oxidation process becomes easy. Combined with various characterization studies, the great change is ascribed to the hydroxylated surface modification. The hydroxyl group is considered as a modification ligand on the surface of BiVO₄, which plays the key roles in (i) modifying the electronic configuration of Bi sites to attain a high-valence Bi^(3+x)+ state toward boosted water dissociation, thus decreasing the PDS energy barrier for the enhanced overall water oxidation kinetics and (ii) promoting efficient e⁻ separation to itself, so more h⁺ could take part in the water oxidation process rather than recombining with e⁻. Taken together, the surface-hydroxylated BiVO₄ photoanode displays the enhanced PEC water oxidation activity, which even attains the level of most reported BiVO₄-based composite photoanodes (Table S1, ESI†). Meanwhile, the applied post-synthetic NaOH immersion way for surface hydroxylation is more practically feasible and cost-effective than other methods (Table S2, ESI†). These advantages will be more meaningful for practical PEC water oxidation applications.

Fig. 10 The illustration of proposed PEC water oxidation reaction mechanism of bare BiVO₄ and hydroxylated BiVO₄ photoanodes.

Conclusions

A surface-hydroxylated BiVO₄ photoanode has been successfully prepared by a hydrothermal and NaOH post-treatment method. The optimal hydroxylated BiVO₄ photoanode displayed a 4.75 times higher photocurrent density and a cathodic shift of 190 mV than those of the bare BiVO₄ photoanode under visible light irradiation. The greatly enhanced activity verified the high efficiency of surface hydroxylation strategy in improving the PEC water oxidation property of BiVO₄. The modified hydroxyl group not only plays a role in enhancing the electron–hole separation efficiency, but also leads to high-valence Bi^(3+x)+ states for promoting the water oxidation kinetics. This work affords a practical surface hydroxylation method with strong operability and high efficiency, which would provide a valuable reference for the subsequent research on PEC water oxidation.

Author contributions

Yaru Li: methodology, data curation, writing – original draft; Fangxia Xie: validation; Zijun Sun: software; Zhuobin Yu: investigation, funding acquisition; Jianxin Liu: resources, funding acquisition; Xiaochao Zhang: supervision, funding acquisition; Yawen Wang: conceptualization; Yunfang Wang: formal analysis; Rui Li: writing – review & editing, funding acquisition; Caimei Fan: project administration, funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work is financially supported by the National Natural Science Foundation of China (No. 22008167, 21978187, 21978196), Natural Science Foundation for Young Scientists of Shanxi Province (No. 201901D211100, 20210302124336, 201901D211058).

Notes and references

1. L. Jiang, Q. Qin, Y. Wang, Y. Su, L. Xia, S. Lin, W. Yao, Q. Wu, Y. Min and Q. Xu, J. Colloid Interface Sci., 2022, 619, 257–266.
2. N. Mughal, A. Arif, V. Jain, S. Chupradit, M. Shabbir, C. Ramos-Meza and R. Zhanbayev, Energy Strategy Rev., 2022, 39, 100745.
3. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
4. B. Trześniewski, I. Digdaya, T. Nagaki, S. Ravishankar, I. Herraiz-Cardona, D. Vermaas, A. Longo, S. Gimenez and W. Smith, Energy Environ. Sci., 2017, 10, 1517–1529.
5. Y. Yang, C. Zhang, C. Lai, G. Zeng, D. Huang, M. Cheng, J. Wang, F. Chen, C. Zhou and W. Xiong, Adv. Colloid Interface Sci., 2018, 254, 76–93.
6. B. Kang, M. Hussain, X. Cheng, C. Peng and Z. Wang, J. Colloid Interface Sci., 2022, 626, 146–155.
7. C. Lhermitte and B. Bartlett, Acc. Chem. Res., 2016, 49, 1121–1129.
8. L. Xia, J. Bai, J. Li, Q. Zeng, L. Li and B. Zhou, Appl. Catal., B, 2017, 204, 127–133.
9. G. Zheng, J. Wang, H. Liu, V. Murugadoss, G. Zu, H. Che, C. Lai, H. Li, T. Ding, Q. Gao and Z. Guo, Nanoscale, 2019, 11, 18968–18994.
10. X. Li, J. Wang, J. Xia, Y. Fang, Y. Hou, X. Fu, M. Shalom and X. Wang, ChemSusChem, 2022, 15, e202200330.
11. Y. Park, K. McDonald and K. Choi, Chem. Soc. Rev., 2013, 42, 2321–2337.
12. T. Soltani, A. Tayyebi and B. Lee, Sol. Energy Mater. Sol. Cells, 2018, 185, 325–332.
13. P. Hajra, S. Kundu, A. Maityc and C. Bhattacharya, Chem. Eng. J., 2019, 374, 1221–1230.
14. S. Anantharaj and S. Noda, Small, 2020, 16, 1905779.
15. Y. Li, S. Li, Y. Wu, J. Zhang, Y. Yang, H. Mao, Y. Zhang and X. Song, Appl. Surf. Sci., 2022, 578, 151914.
16. K. Song, F. He, E. Zhou, L. Wang, H. Hou and W. Yang, J. Energy Chem., 2022, 68, 49–59.
17. J. Zhang, Q. Zhang and X. Feng, Adv. Mater., 2019, 31, 1808167.
18. C. Tang, B. Sun, M. Li, J. Zhang, X. Fan, F. Gao, Y. Tong, L. Dong and Y. Li, J. Mater. Chem. A, 2019, 7, 8050.
19. X. Zhao, X. Liu, B. Huang, P. Wang and Y. Pei, J. Mater. Chem. A, 2019, 7, 24583–24593.
20. J. Kim, D. Youn, K. Kang and J. Lee, Angew. Chem., Int. Ed., 2016, 55, 10854–10858.
21. W. Kim, T. Tachikawa, D. Satoca, H. Kim, T. Majima and W. Choi, Energy Environ. Sci., 2013, 6, 3732–3739.
22. M. Long and W. Cai, Chin. J. Catal., 2008, 29, 881–883.
23. L. Cao, Q. Luo, W. Liu, Y. Lin, X. Liu, Y. Cao, W. Zhang, Y. Wu, J. Yang, T. Yao and S. Wei, Nat. Catal., 2019, 2, 134–141.
24. Y. Li, X. Zhang, J. Liu, Y. Wang, Y. Wang, R. Li and C. Fan, Catal. Commun., 2020, 144, 106071.
25. D. Zhong, T. Li, D. Wang, L. Li, J. Wang, G. Hao, G. Liu, Q. Zhao and J. Li, Nano Res., 2022, 15, 162–169.
26. X. Gao, Y. Zhou, S. Liu, Z. Cheng, Y. Tan and Z. Shen, Appl. Surf. Sci., 2020, 502, 144155.
27. A. Tkatchenko and M. Scheffler, Phys. Rev. Lett., 2009, 102, 073005.
28. S. Hernández, S. Thalluri, A. Sacco, S. Bensaid, G. Saracco and N. Russo, Appl. Catal., A, 2015, 504, 266–271.
29. S. Chen, M. Yuan, W. Feng, W. Liu, W. Zhang, H. Xu, X. Zheng, G. Shen, C. Guo and L. Wang, Water Res., 2020, 185, 116220.
30. M. Janus, M. Inagaki, B. Tryba, M. Toyoda and A. Morawski, Appl. Catal., B, 2006, 63, 272–276.
31. L. Nie, J. Yu, X. Li, B. Cheng, G. Liu and M. Jaroniec, Environ. Sci. Technol., 2013, 47, 2777–2783.
32. R. Mirabal-Rojas, O. Depablos-Rivera, S. Thalluri, J. Medina, M. Bizarro, J. Perez-Alvarez, S. Rodil and A. Zeinert, Appl. Phys. A: Mater. Sci. Process., 2016, 122, 325.
33. T. Dabodiya, P. Selvarasu and A. Murugan, Inorg. Chem., 2019, 58, 5096–5110.
34. G. Zhu, S. Li, J. Gao, F. Zhang, C. Liu, Q. Wang and M. Hojamberdiev, Appl. Surf. Sci., 2019, 493, 913–925.
35. X. Ren, G. Cui, L. Chen, F. Xie, Q. Wei, Z. Tian and X. Sun, Chem. Commun., 2018, 54, 8474–8477.
36. N. Kharche, M. Hybertsen and J. Muckerman, Phys. Chem. Chem. Phys., 2014, 16, 12057–12066.
37. H. Nesbitt, G. Bancroft, G. Henderson, R. Ho, K. Dalby, Y. Huang and Z. Yan, J. Non-Cryst. Solids, 2011, 357, 170–180.
38. X. Zhong, H. He, J. Du, Q. Ren, J. Huang, Y. Tang, J. Wang, L. Yang, F. Dong, L. Bian and Y. Zhou, Electrochim. Acta, 2019, 304, 301–311.
39. R. Li, F. Xie, J. Liu, Y. Wang, Y. Wang, X. Zhang and C. Fan, Dalton Trans., 2016, 45, 9182–9186.
40. Y. Bi, Z. Zhang, X. Huang and B. Zhang, Energy Environ. Sci., 2022, 15, 2867–2873.
41. S. Majumder, M. Gu and K. Kim, Appl. Surf. Sci., 2022, 574, 151562.
42. H. Bai, X. Li, Y. Zhao, W. Fan, Y. Liu, Y. Gao, D. Xu, J. Ding and W. Shi, Appl. Surf. Sci., 2021, 538, 148150.
43. H. Wu, L. Wu, S. Shen, Y. Liu, G. Cai, X. Wang, Y. Qiu, H. Zhong, Z. Xing, J. Tang, Z. Dai, C. Jiang and F. Ren, Int. J. Hydrogen Energy, 2020, 45, 9408–9415.
44. Z. Jia, R. Lv, L. Guo, J. Zhang, R. Li, J. Liu and C. Fan, Sep. Purif. Technol., 2021, 257, 117872.
45. G. Liu, K. Wang, X. Gao, D. He and J. Li, Electrochim. Acta, 2016, 211, 871–878.
46. N. Tian, Y. Zhang, X. Li, K. Xiao, X. Du, F. Dong, G. Waterhouse, T. Zhang and H. Huang, Nano Energy, 2017, 38, 72–81.
47. H. Li, X. Quan, S. Chen and H. Yu, Appl. Catal., B, 2017, 209, 591–599.
48. H. Zhang, W. Wei, F. Hou, W. Guo, Q. Zhang, T. Wang and A. Wei, Sep. Purif. Technol., 2021, 265, 118510.
49. X. Xiong, C. Zhang, X. Zhang, L. Fan, L. Zhou, Y. Chu, W. Huang, C. Wu, J. Li, X. Yang and D. Han, Electrochim. Acta, 2021, 389, 138795.
50. G. Zuo, Y. Wang, W. Teo, A. Xie, Y. Guo, Y. Dai, W. Zhou, D. Jana, Q. Xian, W. Dong and Y. Zhao, Angew. Chem., Int. Ed., 2020, 59, 11287–11292.
51. Y. Liu, X. Liu, X. Wang, H. Ning, T. Yang, J. Yu, A. Kumar, Y. Luo, H. Wang, L. Wang, J. Lee, A. Jadhav, H. Hu, M. Wu, M. Kim and H. Lee, ACS Nano, 2021, 15, 15017–15026.
52. J. Nørskov, J. Rossmeisl, A. Logadottir and L. Lindqvist, J. Phys. Chem. B, 2004, 108, 17886–17892.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2nj04470f

---