The development of clean renewable energy and cost-effective large-scale energy storage conversion system has become one of the research hotspots [1-5]. Solar energy itself can be directly converted into chemical energy with photoelectrochemical (PEC), such as water splitting to form hydrogen (H₂), making it one of the most ideal and environmentally friendly energy production routes in the future [6,7]. So far, there are many semiconductor materials as photocatalysts for water oxidation, for instance, WO₃ [8,9], BiVO₄ [10], Fe₂O₃ [11], SrTiO₃ [12], among which BiVO₄ has attracted much attention because of its unique advantages. BiVO₄ has a suitable band gap of 2.4 eV and is an n-type photocatalytic material, which can absorb a large amount of visible light with a favorable band edge position [13-15]. Nevertheless, severe charge recombination in the bulk BiVO₄ films usually results in a relatively low charge separation and transfer efficiency. To tackle the aforementioned drawbacks, an effective strategy based on surface modification have been developed to retard back electron/hole recombination for improved photocurrent densities [16,17]. For example, the inhibition of carriers' recombination of BiVO₄ by the surface passivation effect of ferrihydrite results in an increased PEC activity of BiVO₄ [18]. Li et al. demonstrated that ZIF-8 overlayer acts as surface passivation layer, which improves the hole utilization of ZnO [19]. However, the synthesis methods of most surface passivation materials are complicated, the production time is long, and a vacuum atmosphere is required. Therefore, a simple synthesis route is desired for preparing the surface passivation photoanodes.

Metal-organic frameworks (MOFs), as a novel family of 2D materials, are considered to be an important candidate in the field of PEC water oxidation due to their high carrier mobility, simple preparation, low cost and easy low-temperature film formation [20]. MOF has been proven to be an effective modified layer for ZnO [19,21], BiVO₄ [22,23], and Fe₂O₃ [24-26] photoanodes. However, as far as we know, MOFs have not been adequately studied as passivation layers for PEC water oxidation.

Herein, the 2D Co-MOF nanosheets (Co(bpdc)(H₂O)₄ (bpdc = biphenyl-4, 4′-dicarboxylic acid), Co-MOF) were decorated on the surface of BiVO₄ photoanode via a hydrothermal method. Meanwhile, with the introduction of Co-MOF, an active PEC water splitting reaction site is realized and the PEC water-splitting activity of BiVO₄ is further enhanced, which reveals that the photocurrent density can reach 6.0 mA/cm² at 1.23 V vs. RHE. At last, the excellent PEC performance of Co-MOF/BiVO₄ film is attributed to the introduction Co-MOF nanosheets passivation layer, which improves the charge separation efficiency in the pure BiVO₄ film.

The synthesis of Co-MOF/BiVO₄ consists of the following three steps: firstly, BiVO₄ films were synthesized by the previously reported method (Support information) [27]. Co-MOF/BiVO₄ electrode is synthesized by in situ loading Co-MOF nanosheets on BiVO₄ using a hydrothermal method [28]. In order to prepare Co-MOF layer, the 30 mL mixed solution contained 26 mL DMF, 2 mL water and 2 mL ethanol was prepared. Subsequently, 0.05 mmol CoCl₂·6H₂O and 0.05 mmol biphenyl-4, 4′-dicarboxylic acid (bpdc) were added to the above solution under stirring. Then the above solution was transferred to a stainless steel autoclave and reacted for 8 h at 120℃. A scaffold was placed in the autoclave to allow the BiVO₄ electrode to stand upright in the autoclave. The dark pink precipitate was washed three times with distilled water and ethanol and dried overnight at 60℃. The steps of the fabrication of a Co-MOF/BiVO₄ are schematically presented in Scheme 1.

Fig. 1 provides images of the synthesized photoanodes taken with a scanning electron microscope (SEM). The pristine BiOI are comprised of uniformly deposited nanosheets that are 30–40 nm thick (Fig. 1a). According to Figs. 1b and c, the original BiVO₄ film has a thickness of about 2 µm and consist of small connecting ellipsoid particles. As shown in Fig. 1d, the Co-MOF nanosheets were vertically distributed throughout the whole BiVO₄ substrate, revealing that the Co-MOF nanosheets have been successfully introduced. In general, the resulting 2D Co-MOF nanosheets are associated with more active metal sites, which promotes more contact between the sample and oxygen, resulting in better PEC activity [29]. The corresponding analysis of the mapping image shows an even distribution of the elements Co, Bi, V and O elements in the Co-MOF/BiVO₄ sample (Figs. 1e–h). The microstructure of Co-MOF/BiVO₄ photoanode was revealed by the TEM. The image in Fig. 1i clearly shown that Co-MOF nanosheets are in close contact with BiVO₄. As presented in Fig. 1j, an interplanar spacing of 0.30 nm was identified, consistent with (121) lattice plane of BiVO₄. The interlayer lattice spacing of Co-MOF is observed to be 0.24 nm in Fig. 1j, which is consistent with previous reports [30].

The crystal structure patterns of BiVO₄ and Co-MOF/BiVO₄ photoanode are shown in Fig. 2a. In the case of BiVO₄ film, all peaks were indexed to the monoclinic scheelite BiVO₄ (JCPDS No. 14-0688) [31]. The diffraction peaks at 26.6°, 37.9°, 51.7°, 65.9° corresponded to tetragonal phase SnO₂ (JCPDS No. 46-1088) [32]. The XRD peaks of Co-MOF/BiVO₄ electrode are similar to that of BiVO₄ film. However, no diffraction peaks assigned to Co-MOF can be observed, due to the small quantity of Co-MOF on the BiVO₄ surface. The structural components of the as-prepared BiVO₄, Co-MOF/BiVO₄ and MOF powder were further analyze via Fourier transform infrared spectroscopy (FT-IR). This peak is the result of the stretching vibration of C=O which is mostly caused by the presence of C=O functional groups within the bpdc bond in Fig. 2b. The antisymmetric stretching vibration peak of the carboxyl group in the ligand appeared at 1589 cm⁻¹ and the peak of the symmetrical stretching vibration occurred at 1367 cm⁻¹ [33]. The absorption peak with a wavelength of 1180 cm⁻¹ can be caused by the expansion and contraction vibration of C-H [34]. Furthermore, the bending vibration of C=C−H bond of benzene ring was observed at 796 cm⁻¹. Compared with pure MOF powder and BiVO₄, the FT-IR of Co-MOF/BiVO₄ not only showed the characteristic vibration peak of MOF ligand, but also the absorption peak of BiVO₄, which proved the successful modification of Co-MOF nanosheets.

Valence and surface chemical elements of BiVO₄ and Co-MOF/BiVO₄ electrode were detected by XPS. The measured spectra in Fig. S1 (Supporting information) indicated that the Bi, V, O and a small amount of Co elements are present in Co-MOF/BiVO₄ sample. In Fig. 2c, the peak at 781.3 eV is associated with Co 2p_3/2 and 797.3 eV with Co 2p_1/2, respectively [35]. Besides, the peaks of 786.4 and 802.8 eV are the signals of Co satellite peaks. In Fig. 2d, both the 159.2 and 164.5 eV peaks were observed to be associated with Bi 4f_7/2 and Bi 4f_5/2, respectively, which were identified as Bi³⁺ in Co-MOF/BiVO₄. Accordingly, the binding peaks were at 516.8 eV and 524.4 eV for V 2p_3/2 and V 2p_1/2, respectively. (Fig. 2e). In O 1s mode from Fig. 2f, the peaks of 531.7 and 532.4 eV correspond to lattice oxygen and hydroxyl groups [36]. In conclusion, the XPS results indicated that the Co-MOF nanosheet were deposited on the BiVO₄ anode through in situ hydrothermal way.

As-fabricated anodes are shown in Fig. S2 (Supporting information) as diffuse reflectance spectra (UV-vis DRS). The BiVO₄ film has shown extended light absorption up to 500 nm, which is attributable to inherent bandgap energy of BiVO₄. After introduction of Co-MOF nanosheets on BiVO₄ surface, the light absorption band edge of Co-MOF/BiVO₄ anode exhibited a slight red shift compared to pure BiVO₄. The band gap of BiVO₄ electrode and Co-MOF/BiVO₄ samples were evaluated from their UV-vis DRS using the Kubelka-Munk formula in Fig. S3 (Supporting information). It is obvious that the band gap value of Co-MOF/BiVO₄ sample (2.38 eV) is narrower than that of BiVO₄ (2.43 eV), implying that Co-MOF nanosheets can slightly affect the band gap value of BiVO₄.

To fundamentally explain the excellent PEC water performance of Co-MOF/BiVO₄ sample, linear sweep voltammograms (LSV) were carried out. In order to exclude the effect of FTO on the PEC properties of other samples, the LSV test of FTO was carried out under the same conditions. As shown in the Fig. S4 (Supporting information), the photocurrent density of the FTO is almost zero, indicating that the FTO acts as a support material in the composite. To determine the optimal thickness of 2D Co-MOF nanosheets, the different concentrations of CoCl₂·6H₂O and bpdc (0.2, 0.1, 0.05 and 0.025 mmol) have been investigated. As displayed in Fig. S5 (Supporting information), the optimal loading of Co-MOF nanosheets is 0.05 mmol. As shown in Fig. 3a, the photocurrent density of BiVO₄ (1.5 mA/cm²) at 1.23 V vs. RHE is low. However, the highest photocurrent density of 6.0 mA/cm² at 1.23 V vs. RHE is obtained for Co-MOF/BiVO₄ sample, obviously higher than pure BiVO₄, as a result of the improvement of charge carries transfer. It is reported that the higher the negative overpotential, the higher the PEC water oxidation capacity [37]. Co-MOF/BiVO₄ photoanode delivered a higher current density at the same potential and obtained a water splitting initial potential of ca. 0.18 V vs. RHE, negatively shifted relative to BiVO₄, indicating suppression of back carries recombination and improved carries migration kinetics [17]. This observation is consistent with that reported for Al₂O₃-passivated hematite photoanodes [37]. Interestingly, Co-MOF/BiVO₄ is superior to the other samples as far as we concerned (Table S1 in Supporting information). Fig. 3b compares the LSV measurements of Co-MOF/BiVO₄ photoanode and BiVO₄ film without illumination. Notably, after introducing Co-MOF nanosheets, the LSV curve of Co-MOF/BiVO₄ moved significantly to the negative potential direction, indicating that reduction of the reaction energy barriers is to be achieved. Besides, the chopped light irradiation of all photoanodes were determined at 0.6 V vs. Ag/AgCl. The photocurrent value of Co-MOF/BiVO₄ is higher than that of BiVO₄ in Fig. 3c, which means that the transmission of photogenerated electrons in Co-MOF/BiVO₄ is more unimpeded. To investigate the interfacial charge separation behavior of all electrodes, a study of electrochemical impedance spectroscopy was performed, as illustrated in Fig. 3d. It is easy to distinguish two different semicircles at low frequencies and high frequencies. The latter is a characteristic of the event impedance in the bulk of the film, while the former is a characteristic of the impedance generated by the surface process [38,39]. Obviously, the semicircle of Co-MOF/BiVO₄ at low frequencies is much smaller than that of BiVO₄, indicating that Co-MOF nanosheets passivate the surface state of BiVO₄ then reduce electron-hole recombination in the surface state.

At 1.23 V vs. RHE, the PEC H₂ evolution activity on BiVO₄ alone and Co-MOF/BiVO₄ composite were evaluated under xenon arc lamp illumination (Figs. 3e and f). Using bare BiVO₄, development of H₂ and O₂ was very low, with 12 µmol h⁻¹ cm⁻² and 6 µmol h⁻¹ cm⁻², respectively, due to rapid electron and hole recombination (Fig. 3e). When a small amount of Co-MOF nanosheets is modified on the BiVO₄ surface in Fig. 3f, it shows a very high PEC H₂ and O₂ evolution activity of 233 µmol/cm² and 115 µmol/cm² and passed charge during the decomposition of water (Fig. S6 in Supporting information), which fully illustrates that the use of Co-MOF nanosheets to modify BiVO₄ is a very effective mean to improve the BiVO₄ PEC hydrogen production activity. The ratio of evolved O₂ and H₂ is equivalent to stoichiometric ratio of 1:2. Moreover, H₂ and O₂ have a Faraday efficiency greater than 90%. The Co-MOF/BiVO₄ electrode shows good reproducibility in Fig. S7 (Supporting information) with a photocurrent density of 4.8 mA/cm² at 1.23 V vs. RHE after PEC water oxidation reaction. It is worth noting that the Co-MOF/BiVO₄ photoanode exhibited excellent photocurrent and structural stability throughout the PEC water oxidation reaction test (Figs. S8–S10 in Supporting information), indicating that the introduction of the Co-MOF layer not only significantly increased the photocurrent density but also effectively contributed to the PEC stability of the BiVO₄ photoanode. The XRD pattern of the Co-MOF/BiVO₄ film shows a high match with fresh XRD and no other diffraction peaks appears, demonstrating the good stability of the Co-MOF/BiVO₄ photoanode (Fig. S11 in Supporting information).

The incident photon-to-current efficiency (IPCE) is also performed to measure the Co-MOF/BiVO₄ and BiVO₄ photoanodes PEC performance. From Fig. 4a, the Co-MOF/BiVO₄ anode achieved the highest IPCE value in the entire spectrum. Especially at 440 nm, the IPCE value was significantly increased to 93%, which is 6.2 times that of BiVO₄ (15% at 430 nm). According to Fig. 4b, the BiVO₄ photoanode has a low applied bias photon to current efficiency (ABPE) value. After modification of the BiVO₄ photoelectrode with the Co-MOF nanosheets, the ABPE is estimated as 2.12% at 0.64 V vs. RHE, corresponding to an eight-fold boost over for BiVO₄ (0.26% at 0.92 V vs. RHE). In order to reveal the mechanism of action by quantitative analysis, the charge injection efficiency (η_surface) of the synthetic film is provided in Fig. S12 (Supporting information). Less than 32% of the charge injection efficiency occurs at the bare BiVO₄ anode in the overall potential window over the entire potential range, implying that the pores generated by photons to participate in the oxidation reaction of water partially recombine with electrons instead of reaching the surface of the photoanode. However, the charge injection efficiency of Co-MOF/BiVO₄ photoelectrode showed significant improvement over the entire potential window, with the maximum value reaching up to 77.0% max at 1.23 V vs. RHE, indicating the addition of Co-MOF nanosheets greatly inhibited electron-hole recombination in BiVO₄. To further explore PEC behavior, the separation efficiency of the charges (η_bulk) was also studied in Fig. 4c. The charge separation efficiency of the Co-MOF/BiVO₄ photoanode was 96%, which was approximately about 1.62 times that of the BiVO₄ photoanode (59%), indicating that the electron-hole recombination was more severe in the original BiVO₄, while the Co-MOF mainly reduced the occurrence of this phenomenon. This result further demonstrated that the introduction of passivation layer of Co-MOF nanosheet can achieve efficient charge separation. To further investigate the recombination characteristics of photogenerated carriers, we tested the photoluminescence spectrum (PL) of the sample in Fig. S13 (Supporting information). The emission peak position of the Co-MOF/BiVO₄ anode is consistent with that of BiVO₄ (508 nm, Fig. S14 in Supporting information), but the emission peak intensity is reduced. Therefore, the introduction of Co-MOF nanosheets improves the photogenerated charges separation rate of BiVO₄ and further promotes the improvement of PEC water decomposition performance of BiVO₄ anode. To get a sense the critical role of charge separation efficiency in the ultimate PEC performance, time-resolved PL (TRPL) attenuation measurements of BVO₄ and Co-MOF/BiVO₄ sample were performed. Significantly, the decreased fast carrier lifetime (τ₁) of Co-MOF/BiVO₄ implied a more effective surface charge transfer compared with that of pure BiVO₄ (inset Fig. 4d), which is consistent with previous reports [37,40]. Besides, the free exciton lifetime (τ₂) in Co-MOF/BiVO₄ anode is increased relative to BiVO₄, indicating that the introduction of Co-MOF nanosheets can effectively promote the interfacial photogenerated carries separation of BiVO₄.

To a better comprehension, the existence mechanism of BiVO₄ and Co-MOF nanosheets, the Mott-Schottky of as-prepared anodes is shown in the Fig. 4e. In general, a smaller slope of graph in Fig. 4e indicates a higher density of free carriers and a higher conductivity [41]. In our case, Co-MOF/BiVO₄ has a lower slope than BiVO₄, indicating that Co-MOF nanosheets passivates the surface of BiVO₄ and thus reduces carriers' recombination. The addition of Co-MOF nanosheets leads to an increase in the conductivity of BiVO₄ and further improves the PEC water decomposition performance of BiVO₄, which corresponds to the result in Fig. 3d. A function of time of open circuit voltage (V_oc) has been used to analyze the electronic recombination kinetics of the prepared photoanode in 1 mol/L borate buffer electrolyte under sunlight irradiation. Fig. 4f illustrates that Co-MOF/BiVO₄'s V_oc decay rate is lower than bare BiVO₄'s. Therefore, the lower V_oc decay rate of Co-MOF/BiVO₄ provides a greater driving force for charge separation.

According to CV cycle diagram of Co-MOF/BiVO₄ sample in Fig. S15 (Supporting information), there is a reduction peak at ca. 1.6 V vs. RHE [42]. XPS was used to further analyze the changes in the surface structure and chemical state of Co-MOF/BiVO₄ after PEC water splitting testing. The Co³⁺ can be observed obviously in Fig. S16 (Supporting information), indicating that the Co²⁺ becomes Co³⁺ oxidized via holes of BiVO₄ in part in the Co-MOF nanosheets [43]. The above experimental results indicate that Co²⁺/Co³⁺ are constantly transformed into each other in Co-MOF/BiVO₄ electrode. Considering the experimental results and literature reports, we discussed the possible reaction mechanism of Co-MOF/BiVO₄ photoanode PEC water splitting in the Fig. 5. Under irradiation, the photogenerated electrons will transfer to the Co-MOF/BiVO₄ surface, and further captured by the Pt electrode through the external circuit, thereby participating in the reduction of water to produce H₂. While the photogenerated holes tend to move to the surface of BiVO₄ and captured by the Co-MOF nanosheets. Subsequently, the Co²⁺ is oxidized to Co³⁺ by holes. Afterwards, the oxygen is produced in the process of unstable Co³⁺ is reduced to Co²⁺ by electrons to oxidize the interfacial H₂O. Moreover, the continuous conversion of Co²⁺ and Co³⁺ to each other provides a steady stream of active sites for water oxidation [44,45]. These processes realize the effective transfer and separation of photogenerated charges of BiVO₄, which is conducive to PEC water spitting. In short, the reason for this enhanced PEC performance can be associated with the synergistic effect of Co-MOF nanosheets and BiVO₄.

In summary, a novel passivation layer, namely 2D Co-MOF nanosheets, was in situ deposited on the BiVO₄ photoanode through a unique strategy of hydrothermal reaction to inhibit recombination rate of carriers. With a current density of 6.0 mA/cm², the resultant Co-MOF/BiVO₄ photoanode displayed significant improvements in PEC oxidation over the original BiVO₄ film at 1.23 V vs. RHE under AM 1.5G irradiation. Based on the optimized Co-MOF/BiVO₄ photoanode, 96% of charge separation efficiency is achieved under AM 1.5G irradiation at 1.23 V vs. RHE, indicating that the water oxidation is carried out efficiently with surface holes to improve charge separation efficiency. Moreover, the Co-MOF/BiVO₄ film displayed higher H₂ and O₂ evolution rates of 233 µmol/cm² and 115 µmol/cm² for PEC water spitting, about 20 times that of pure BiVO₄ film. The outstanding PEC performance of Co-MOF/BiVO₄ photoanode may be attributed that the passage of photoexcited charge carriers of BiVO₄ is effectively inhibited by the passivation layer of 2D Co-MOF nanosheets.

The authors declare no conflict of interest.

This work was financially supported by the National Natural Science Foundation of China (No. 52173277), the Innovative Research Team for Science and Technology of Shaanxi Province (No. 2022TD-04), the Fundamental Research Funds for the Central Universities of Chang'an University (Nos. 300102299304, 300102291403), the Natural Science Basic Research Fund of Shaanxi Province (No. 2020JZ-20).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.108007.