1. Introduction

Photocatalysis is a potential technology for green energy production and environmental remediation because it can directly use solar energy [1-8]. However, photocatalytic activity remains low due to rapid charge carriers recombination. For efficient separation of charge carriers, photoelectrocatalysis is an excellent method combining the advantages of photocatalysis and electrocatalysis and has been successfully applied in many fields such as hydrogen production, oxygen production, carbon dioxide reduction, inorganic ion reduction, and organic compound oxidation [9-16]. In this method, a semiconductor is illuminated to generate charge carriers while being biased by a gradient potential to suppress charge carriers recombination. Since photoelectrocatalytic water splitting using a TiO₂ electrode was reported in 1972 [17], enormous efforts have been made to develop highly efficient photoelectrocatalytic materials for solving the growing energy shortage and environmental remediation problems [18, 19]. So far, many review articles have been published discussing various topics related to photoelectrocatalysis, which mainly focus on typical photocatalytic materials, for example, TiO₂ [20, 21], WO₃ [22], NiO [23], and graphenes [24]. However, as far as we know, there is no specific review article on the application of metal-organic frameworks (MOFs) materials in photoelectrocatalysis.

Emerging MOFs have attracted extensive attention in the fields of sensors, drug delivery, gas separation, storage, and catalysis, due to their large surface area, porous structure, and tunable organic linkers [25-28]. Among catalysis, using MOFs for photoelectrocatalysis is a fascinating work in the future, as MOFs can promote substrate/product transport and active site exposure, and shorten the charge carriers transfer distance.

Therefore, in this review, we focus on recent advances in photoelectrocatalysis based on MOFs, including the representative MOFs and their applications. Finally, the future perspectives of MOFs-based photoelectrocatalysis are analyzed.

2. Fundamentals of photoelectrocatalysis

In photocatalysis process, when a photocatalyst (semiconductor) is illuminated with photons exceeding its bandgap energy, the electrons in the valence band (VB) are excited to conduction band (CB), generating negative sites in CB (e_CB^-) and leaving holes in VB (h_VB⁺) [29, 30]. The e_CB^- and h_VB⁺ can react with water or CO₂, as shown in the following equations (Eqs. 1–14) [31-34]:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

The main loss of photocatalytic efficiency is caused by e_CB^- directly combined with h_VB⁺ or ^·OH, as shown below (Eqs. 15 and 16):

(15)

(16)

Photoelectrocatalytic technology can improve photocatalytic efficiency by adding electrochemical techniques to the photocatalytic process.

In general, photoelectrocatalytic reactions are carried out in photoelectrochemical cells [35, 36]. These cells are based on interfacial behavior between a photocatalyst and an electrolyte, as illustrated in Fig. 1. A photocatalyst on the photoanode electrode is excited under light illumination, and the electrons in VB are transferred to CB, which produces electron-hole. The electrons are then conducted by an external circuit under bias potential to the cathode electrode where they participate in the reduction reaction, for example, water splitting for hydrogen generation or conversion of carbon dioxide. The photogenerated holes undergo an oxidation reaction at the photoanode electrode, such as water splitting for oxygen generation or oxidation of organic substances [11, 37].

3. Photoelectrocatalysis on MOFs photoelectrodes

A lot of studies have been reported on the catalytic performance of MOFs. MOFs can be used as a novel photoelectrocatalyst due to ordered structure and large specific surface areas. Through modulating their metal clusters or organic linking groups, combining with other metal or semiconductor, MOFs can improve charge separation efficiency, resulting in excellent photoelectrocatalytic properties [38]. In addition, MOFs can be used as sacrificial templates for the manufacture of MOFs derivatives such as porous metal materials [39, 40] and carbon materials [41] for photoelectrocatalytic applications. This review describes various MOFs-based materials as photoelectrocatalysts for hot spot applications, such as CO₂ conversion and water splitting.

3.1. Ti-based MOFs

Ti-based MOFs have been extensively studied for photocatalysis and photoelectrocatalysis. However, the applications of single Tibased MOFs in photoelectrocatalysis are still limited due to the rapid recombination of photogenerated electron-holes. To overcome the disadvantages, constructing heterostructure by coupling with other semiconductors is a promising approach. As a typical semiconductor, TiO₂ is considered as one of the most promising photocatalysts because of its low costs, good biocompatibility, structure stability [42, 43]. Recently, Ti-MOF (NH₂-MIL-125(Ti)) modified TiO₂ nanotube array (TNTAs) hybrid electrode was employed as an anode for photoelectrocatalytic water splitting H₂ production [38]. In this electrode, Ti-MOF nanoparticles were about 300 nm in size and were mainly present in the pore of TNTAs. The TNTAs@Ti-MOF hybrid possessed excellent light absorption properties and could promote charge separation, which was beneficial for enhanced photoelectrocatalytic activity. As a result, when TNTAs@Ti-MOF was used as the photoanode, the hydrogen production rate was up to 132.86 μmol h^-1 cm^-2 irradiated by a Xenon lamp (300 W) with the light intensity of 336 mW/cm². As shown in Fig. 2, under illumination, Ti-MOF and TNTAs were excited. Then, the photogenerated electrons on the LUMO of TiMOF were transferred to the CB of TNTAs, and the photogenerated holes in the VB of TNTAs were transported to the HOMO of Ti-MOF. Subsequently, under bias, the photogenerated electrons and holes were effectively separated to the cathode (Pt electrode) and the anode, respectively, and the electrons on the Pt electrode reduced water to produce H₂.

Similarly, Lee et al. uniformly coated the NH₂-MIL-125(Ti) layer on vertically ordered TiO₂ nanorods (NRs) by hydrothermal reaction for photoelectrocatalysis under simulated AM 1.5 G solar irradiation (100 mW/cm²) [44]. TiO₂ NRs were grown in-situ on an fluorine-doped tin oxide (FTO) glass substrate. The diameter and length of the TiO₂ NRs could be varied by changing the hydrothermal conditions. The photocurrent density of TiO₂ NRs was only 0.6 mA/cm² at 1.23 V vs. RHE, while that of NH₂-MIL-125 (Ti)/TiO₂ NRs was 1.63 mA/cm². In addition, the incident photon conversion efficiency of NH₂-MIL-125(Ti)/TiO₂ NRs was greatly enhanced at 340 nm, suggesting that water oxidation was promoted. The enhancement of photoelectrocatalytic activity was attributed to ⅰ) uniform NH₂-MIL-125(Ti) coating on TiO₂ NRs; ⅱ) suitable bandgap of NH₂-MIL-125(Ti); ⅲ) improved light absorption and charge transport; ⅳ) NH₂-MIL-125(Ti) and TiO₂ type Ⅱ bandgap arrangement.

Apart from the MOFs themselves, MOFs derivatives were also reported as photoelectrocatalysts. Ti-based MOF (Sn⁴⁺ exchanged MIL-125(Ti))-derived dual-phase TiO₂ was coupled with layered MoS₂ as a photoanode for photoelectrochemical water splitting under simulated AM 1.5 G solar irradiation (100 mW/cm²) [45]. The MOFs-derived dual-phase TiO₂ hybrid (denoted as DT) was prepared by adding SnCl₄ to a precursor solution containing titanium isopropoxide and terephthalic acid, followed by hydrothermal treatment. This DT was added to a solution containing thiourea and sodium molybdate, and then hydrothermally reacted at 200 ℃ for 24 h to obtain the MoS₂/MOFs-derived TiO₂ hybrid (denoted as MDT). DT inherited the pore structure of the pristine MIL-125(Ti), which could prolong the light propagation length and promote carrier transport. After incorporation of MoS₂, the visible light absorption and water oxidation kinetics of MDT were significantly improved. Furthermore, the heterojunction formed by the dual-phase TiO₂ and the layered MoS₂ suppressed the recombination of charge carriers, and greatly enhanced charge transfer. Therefore, the photocurrent density of MDT was 1.2 mA/cm² at 1.23 V vs. RHE, which was 2 times that of the TiO₂ photoanode.

Fe₂O₃ is a thermodynamically stable iron oxide, and acts as an n-type semiconductor with a narrow bandgap, which has been popular in recent years. However, the low electronic mobility and low electrical conductivity of Fe₂O₃ restrain its widespread applications in photoelectrocatalysis [46]. It has been confirmed that the photoelectrocatalytic (PEC) efficiency of Fe₂O₃ can be greatly improved by incorporating Ti into Fe₂O₃ [47]. Li et al. combined Fe₂O₃ nanorod arrays with NH₂-MIL-125(Ti) and then calcined to obtain Ti_xFe_1-xO_y shell/Fe₂O₃ core nanorod arrays for photoelectrocatalytic reaction under simulated AM 1.5 G solar irradiation (100 mW/cm²) [48]. Its photoelectrochemical performance was studied by photocurrent density measurement. Gas chromatographic analysis confirmed the production of oxygen and hydrogen at the anode and cathode, respectively. The Ti_xFe_1-xO_y/Fe₂O₃ electrode with optimum Ti_xFe_1-xO_y shell thickness showed the photocurrent density of 0.724 mA/cm² at 1.23 V (vs. RHE), which was 26.7 times of that obtained by Fe₂O₃ electrode. Additionally, the current density retention of the Ti_xFe_1-xO_y/Fe₂O₃ reached 98.9% under simulated sunlight for 5 h, showing excellent stability. In this photoelectrocatalytic process, the Ti_xFe_1-xO_y shell derived from NH₂-MIL-125(Ti) played a key role. In this Ti_xFe_1-xO_y shell, the substitution of Ti in Fe₂O₃ created structural defects. These structural defects could capture photogenerated holes, which facilitated charge transfer and improved charge separation in the electrode, thereby significantly enhancing the photoelectrochemical activity of the Ti_xFe_1-xO_y/Fe₂O₃ nanorod.

3.2. Zn-based MOFs

Zn-based MOFs combined with semiconductor materials have proven to be a promising method for photoelectrocatalysis. Among semiconductor materials, ZnO has attracted a lot of research attention due to its nontoxicity, low cost, and excellent electron mobility [49]. However, the slow kinetics at the ZnO/electrolyte interface results in a relatively low charge separation efficiency. To solve the above disadvantage of ZnO, surface modification strategies have been developed [50]. The use of Zn-based MOFs as surface layers to modify ZnO has been poorly investigated. Liu et al. coated Zn-based MOF (ZIF-8) as a passivation layer on a Nifoam (NF) supported ZnO nanorod array (NA) photoanode, which showed significantly enhanced photoelectrocatalytic water splitting performance under simulated AM 1.5 G solar irradiation (100 mW/cm²) [51]. As shown in Fig. 3, ZnO NR was prepared by the hydrothermal method on NF (ZnO/NF). Then, using the formed ZnO NA as a self-sacrificial template and precursor, the ZIF-8 cap layer was grown in-situ on ZnO NA to obtain core-shell structure (ZIF-8/ZnO/NF). Finally, Ni(OH)₂/ZIF-8/ZnO/NF photoelectrode was obtained by surface adsorption and in-situ reaction with Ni(OH)₂ to treat ZIF-8/ZnO/NF. The activity of water splitting could be enhanced by optimizing the thickness of ZIF-8 and adding cocatalyst Ni(OH)₂. The photocurrent density of the optimized Ni(OH)₂/ZIF-8/ZnO/NF photoanode was 1.95 mA/cm², which was twice that of ZnO/NF. And the current conversion efficiency of incident photons was 40.05%, which was also twice that of ZnO/NF. The synergistic effect of surface passivation and cocatalyst was beneficial to extend charge lifetime, boost charge transfer, and reduce overpotential of water oxidation, thereby enhancing photoelectrocatalytic water splitting activity.

ZIF-8 can also decorate Ti/TiO₂ nanotubes for CO₂ photoelectrocatalytic reduction under 125 W mercury vapor lamp irradiation [52]. The photocurrent and the value of the solution saturated with CO₂ were significantly increased after coating the Ti/TiO₂ electrode with ZIF-8. Moreover, spectroscopic and voltammetric measurements confirmed that CO₂ formed stable carbamates on the surface of ZIF-8. The photoelectrocatalytic reactor was shown in Fig. 4. At E_app of +0.1 V, in 0.1 mol/L Na₂SO₄, CO₂ was reduced to form 0.7 mmol/Lmethanol and 10 mmol/L ethanol.

In addition, ZIF-8 derivatives have also been reported to be successfully used in photoelectrocatalytic reactions. Zou et al. prepared ZIF-8-derived carbon-doped ZnO (C-ZnO) with porous structure by two-step calcination of ZIF-8 for photoelectrocatalytic water splitting irradiated by a 300 W Xenon lamp (> 420 nm) [53]. ZIF-8 was synthesized by dropping an aqueous solution of zinc acetate into an ammonia solution of 2-methylimidazole, followed by stirring for 4 h. The prepared ZIF-8 was first calcined at 350 ℃ and then at 400 ℃ to obtain C-ZnO (labeled as ZnO (C350-400)). This two-step calcination method not only successfully doped C into the ZnO crystal lattice, but also maintained the cubic morphology of the ZnO with porous structure compared with the one-step calcination. Compared to ZnO (C450), the absorption edge of ZnO (C350-400) showed a redshift, indicating that the twostep calcined sample had better light absorption than the directly calcined sample. And the photoluminescence (PL) intensity of the two-step calcined sample was lower than that of the directly calcined sample, demonstrating that the former had higher charge separation efficiency. Finally, in the photoelectrochemical water splitting experiments, ZnO (C350-400) exhibited the highest activity with a photocurrent density of 83 μA/cm² at 1.23 V vs. RHE, which was 4 times higher than that of C450.

Besides, Zn-based MOF ([Zn(bpcda)(bdc)]_n, bpcda = N, N'-bis (pyridin-4-ylmethylene) cyclohexane-1, 4-diamine, bdc = benzenedicarboxylic acid) prepared by solvothermal method also exhibited significant photoelectrochemical catalytic activity under simulated AM 1.5 G solar irradiation (100 mW/cm²) [54]. In this Zn-based MOF, the metal center was attached to the bdc anion to generate one-dimensional chains that were connected by bipyridyl bpcda to construct a three-dimensional framework. Characterization demonstrated that [Zn(bpcda)(bdc)]_n not only absorbed a long-range of light but also exhibited multiple LUMO/HOMO levels. The presence of metal centers and two different ligands resulted in these multiple absorption bands, promoting photogenerated electronhole separation, thereby increasing the oxygen evolution reaction (OER) activity of [Zn(bpcda)(bdc)]_n. The maximum photocurrent density of [Zn(bpcda)(bdc)]_n anode exceeded 31 μA/cm² at 0.9 V (Ag/AgCl/Sat. KCl). In addition, the [Zn(bpcda)(bdc)]_n anode exhibited reproducibility throughout the experimental time (more than 12 h), indicating high photoresponse stability and negligible photocorrosion.

3.3. Co-based MOFs

Co-based MOF is also considered to be a promising material for photoelectrocatalysis, and it has been extensively studied. For instance, Co-MOF (ZIF-67) was coated on noble metal sensitized semiconductor (ZnO@M, M = Au, Pt, and Ag) for photoelectrochemical water splitting irradiated by a 150 W Xenon lamp (100 mW/cm²) [55]. The prepared ZnO@Au@ZIF-67 had the following advantages: (ⅰ) The narrow bandgap ZIF-67 could enhance effective light response, and its porous structure hardly blocked the light absorption of ZnO@Au. Compared with ZnO and ZnO@Au, the light absorption intensity of ZnO@Au@ZIF-67 increased significantly from the ultraviolet to the visible range. (ⅱ) The photogenerated electrons in the ZIF-67 shell were rapidly transferred to ZnO@Au because the well-contacted ZIF-67/ZnO@Au interface could accelerate charge transfer. Compared with ZnO, the photoluminescence intensity of ZnO@Au was decreased, indicating that Au was favorable for suppressing electron and hole recombination. In addition, after ZIF-67 growth, the photoluminescence intensity was further reduced, demonstrating that ZnO@Au@ZIF-67 can improve the transfer of photoelectron-hole. (ⅲ) The ZIF-67 could serve as a protecting layer to avoid aggregation or leaching of ZnO@Au, which improved the stability of the photoanode. As a result, the photocurrent density of ZnO@Au@ZIF-67 was 1.93 mA/cm², and the onset potential of ZnO@Au@ZIF-67 displayed a cathodic shift from 0.05 V to -0.1 V. Moreover, this strategy could be extended to the preparation of ZnO@Ag@ZIF-67 and ZnO@Pt@ZIF-67, which also had high photoelectrochemical water splitting activity.

Recently, ZIF-67 was in-situ grown on Fe₂O₃ for photoelectrocatalytic OER irradiated by a 300 W Xenon lamp (100 mW/cm²) [56]. In the preparation process, a Fe₂O₃ film was prepared on FTO glass by a simple hydrothermal method. To prepare a Co-modified Fe₂O₃ film, the synthesized bare Fe₂O₃ film was spin-coated with a Co(NO₃)₂ aqueous solution, and then heated. Subsequently, this Co-modified Fe₂O₃ film was immersed in a 2-methylimidazole solution for a period of time, washed with water, and then dried to obtain an in-situ Co-MOF modified Fe₂O₃ photoanode (Fe₂O₃@- Co@MIm). Compared with bare Fe₂O₃, the V_onset of Fe₂O₃@Co@- MIm showed a significant cathodic shift (180 mV), indicating that surface treatment with Co-MOF enhanced OER. This enhancement was due to the fact that the porous Co-MOF was rich in Co active sites, which can rapidly transport electrons as a redox medium. In addition, the photocurrent density of optimized Fe₂O₃@Co@MIm showed a significant increase to 2.0 mA/cm² at 1.23 V vs. RHE. However, an excessively thick Co-MOF layer weakened the interaction between the Fe₂O₃ and the electrolyte, thereby suppressing hole transfer and resulting in lower photocurrent density and a higher onset potential.

Co-based MOF (ZIF-9) was also used by Zhao et al. to modify Co₃O₄ nanowire to construct a biomimetic photoelectrocatalytic interface for photoelectrocatalysis under simulated AM 1.5 G solar irradiation (100 mW/cm²) [57]. Experiments revealed that CO₂ can be well adsorbed on the ZIF-9 modified Co₃O₄ nanowire. Moreover, on this hybrid interface, the CO₂ surface concentration increased by 3.44 times compared to that on Co₃O₄ nanowire. Density functional theory (DFT) showed that ZIF-9 has the ability to activate CO₂ molecules by combining Co atoms with O atoms of CO₂, leading to a positive shift in the CO₂ reduction onset potential on ZIF-9-Co₃O₄ hybrid compared to Co₃O₄ nanowire. The photoelectrocatalytic conversion of CO₂ to formate was 72.3 μmol L^-1 cm^-2 h^-1, and the liquid product selectivity was close to 100% at a low overpotential (290 mV).

Apart from ZIF-67 and ZIF-9, Co₂(benzimidazole)₄ was also used to modify BiVO₄ photoanode for OER under simulated AM 1.5 G solar irradiation (100 mW/cm²) [58]. Ultrathin CoO sheets were used as the metal source for in-situ preparation of small size CoMOF nanoparticle on BiVO₄. The photocurrent density of Co-MOF modified BiVO₄ photoanode was 3.1 mA/cm² at 1.23 V vs. RHE, which was much better than that of BiVO₄ photoanode (1.2 mA/cm²). In addition, after the incorporation of Co-MOF, a significant cathodic shift occurred at the onset potential of Co-MOF modified BiVO₄, indicating that Co-MOF had OER activity. Highly dispersed and small-sized Co-MOF acted as a cocatalyst to rapidly transfer the photogenerated holes in BiVO₄ and enhance the surface reaction kinetics of photoelectrochemical OER. Moreover, Co-MOF with abundant metal sites and porosity displayed high charge separation efficiency (83%), also resulting in boosted OER activity of the Co-MOF modified BiVO₄ photoanode. Recently, another Co-based MOF, (Co-MIm) nanosheets were grown on BiVO₄ for photoelectrocatalytic overall water splitting [59]. The photocurrent density of BiVO₄@Co-MIm at 1.23 V vs. RHE was 3.16 mA/cm², which was 2.4 times that of BiVO₄ photoanode. CoMIm coating caused a 230 mV cathode shift of onset potential, demonstrating that Co-MIm had excellent water oxidation properties. Studies also confirmed that BiVO₄@Co-MIm photoanodes had high charge separation efficiency compared with BiVO₄ photoanodes. As shown in Fig. 5, the photoelectrocatalytic process of the BiVO₄@Co-MIm photoanode was carried out under AM 1.5 G irradiation (100 mW/cm²) using an H-type cell. The oxygen and hydrogen produced after 3 h were 79 μmol and 161 μmol, respectively, which were in accordance with Faraday efficiency of 92.3% and 91.8%, respectively. During this process, BiVO₄ was excited and produced photogenerated electron-hole pairs under illumination. Co³⁺ of Co-MIm was first oxidized by holes to Co⁴⁺. Then, Co⁴⁺ was reduced to Co²⁺ in the water while generating oxygen. Subsequently, Co²⁺ was oxidized by holes to Co³⁺. At the same time, electrons were transferred to the cathode for hydrogen production.

Co-MOF derivatives were also reported as photoelectrocatalysts. The ZIF-67-derived Co₃C/TiO₂ heterostructure was constructed by the hydrothermal method for photoelectrocatalytic water oxidation irradiated by a 150 W Xenon lamp [60]. First, a layer of TiO₂ was coated on the Cu₂O template by hydrothermal. Then, the synthesized TiO₂/Cu₂O was added to a solution of Co(NO₃)₂ and 2-methylimidazole. Under electrostatic attraction, ZIF- 67 was deposited on TiO₂/Cu₂O. Finally, the synthesized ZIF-67/TiO₂/Cu₂O was calcined under argon to obtain ZIF-67-derived Co₃C/TiO₂ nanocage. ZIF-67-derived Co₃C could enhance light capture and promote water oxidation kinetics. At the same time, the type Ⅱ heterojunction formed by TiO₂ and Co₃C enhanced the separation of photogenerated electron-hole. Moreover, as a cocatalyst, ZIF-67-derived Co₃C could oxidize OH^- to O₂ to fast consume the trapped photogenerated holes, thereby significantly suppressing the recombination of surface carriers. As a result, the ZIF-67-derived Co₃C/TiO₂ photoanode significantly increased the photoelectrocatalytic water oxidation activity. The optimized ZIF- 67-derived Co₃C/TiO₂ anode had a photocurrent density of 2.6 mA/cm² and a charge separation efficiency of 92.6%, which was 201% and 152% higher than that of TiO₂, respectively. Besides, ZIF-67 derived Co₃O₄ coated TiO₂ nanorod arrays on Si were constructed as photoanode for efficient photoelectrochemical water oxidation under simulated AM 1.5 G solar irradiation (100 mW/cm²) [61]. This Co₃O₄/TiO₂/Si ternary heterojunction could promote carrier separation. The CB and VB potentials of Si were higher than that of TiO₂, and the photogenerated electrons from the CB of TiO₂ were rapidly transferred to the VB of Si, thereby effectively separating the electrons from TiO₂ and the holes from Si. In addition, because of the large specific surface area and porous channel structure, MOFs-derived Co₃O₄ significantly enhanced the photoresponse and water oxidation kinetics of Co₃O₄/TiO₂/Si. Particularly, the onset potential of TiO₂/FTO was 0.48 V vs. RHE, and the photocurrent density of TiO₂/FTO was 0.83 mA/cm² at 1.23 V vs. RHE under illumination. When TiO₂/FTO was coupled with Co₃O₄, the onset potential shifted to 0.43 V (vs. RHE), and the photocurrent density increased to 2.00 mA/cm² at 1.23 V vs. RHE. After incorporation of Si, the synergistic effect of Co₃O₄/TiO₂/Si anode resulted in a photocurrent density of 2.71 mA/cm² and the light conversion efficiency was greatly increased to 0.54%.

Very recently, Gimenez et al. developed a simple strategy to modify BiVO₄-based photoanode with a porous CoO_x cocatalyst for photoelectrocatalytic water splitting irradiated by a 300 W Xenon lamp (100 mW/cm²) [62]. The surface of BiVO₄ was first modified with benzene-1, 3, 5-tricarboxylic acid, and then ZIF-67 was assembled by simply mixing 2-methylimidazole and cobalt nitrate in methanol. By this method, uniform growth and strong adhesion of MOF particles were produced on the surface of BiVO₄ (labeled as BiVO₄-ZIF-67). Subsequently, the MOF particles were converted to porous CoO_x cocatalysts (labeled as BiVO₄-CoO_x) by calcining the BiVO₄-ZIF-67 sample in air. Benefited from the high surface area, this porous CoO_x cocatalyst exposed sufficient catalytically active sites for OER. In addition, MOF-converted CoO_x cocatalyst significantly enhanced light scattering and bandgap absorption on BiVO₄, which facilitated light trapping. Therefore, water oxidation kinetics could be significantly accelerated. Photoelectrochemical characterization of the BiVO₄-CoO_x photoanode showed a significant increase in catalytic performance compared with the BiVO₄ electrode (about 4 times higher photocurrent and 300 mV cathode offset of the catalytic onset potential).

3.4. Fe-based MOFs

Fe-based MOFs attract a lot of attention because they have strong light absorption and iron is an earth-abundant element. Jiao et al. used Fe-based MOF (MIL-100(Fe)) to modify nanoporous Fe/W co-doped BiVO₄ photoelectrodes for photoelectrochemical water splitting irradiated by a 300 W Xenon lamp (100 mW/cm²) [63]. The distribution of MIL-100(Fe) on the Fe/W codoped BiVO₄ photoanode was confirmed by scanning electron microscopy (SEM) images. Since the bandgap of MIL-100(Fe) was about 2 eV, the absorption edge was greatly red-shifted on the Fe/W co-doped BiVO₄ film after MIL-100(Fe) decoration. The photocurrent density of MIL-100(Fe) modified Fe/W co-doped BiVO₄ was 2.76 mA/cm², which was 14 times that of BiVO₄. Here, MIL-100(Fe) was used as a cocatalyst because of its low photoelectrochemical performance, small amount, and the significant effect of E_fb on Fe/W co-doped BiVO₄. In this case, electrons and holes generated from Fe/W co-doping BiVO₄ can be rapidly separated in the interface of MIL-100(Fe) and Fe/W co-doped BiVO₄. The electrons were transported to the counter electrode under bias for reduction reaction, and the holes are rapidly migrated to MIL-100(Fe) for the oxidation reaction. Additional, MIL-100(Fe) possessed a porous structure that was advantageous for its contact with the electrolyte, which may also contribute to an increase in photoelectrochemical activity.

MIL-53(Fe) was used as a cocatalyst to efficiently promote hole transport, significantly improving the photoelectrocatalytic activity of Mo-doped BiVO₄ photoanodes for water oxidation under simulated AM 1.5 G solar irradiation (100 mW/cm²) [64]. Modoped BiVO₄-MIL-53(Fe) (FMBV) photoanode was prepared by simply coating a Mo-doped BiVO₄ film on FTO and then spincoating a MIL-53(Fe) co-catalyst. By doping Mo₆⁺, the charge transport properties of BiVO₄ could be enhanced, thereby improving the water oxidation efficiency of Mo-doped BiVO₄. Moreover, in the presence of MIL-53(Fe), not only the charge carrier density was increased but also the photogenerated electron-hole separation was accelerated. Specifically, after the spin coating of MIL-53(Fe), FMBV showed a significantly improved photoresponse. The photocurrent density of FMBV was 2.2 mA/cm² at 1.23 V vs. RHE, while that of pure BiVO₄ was only 0.57 mA/cm². A possible mechanism was that the VB electrons of Mo-doped BiVO₄ were excited to CB under illumination and then transferred to the FTO. At the same time, photogenerated holes were transported to MIL-53(Fe). MIL-53(Fe) could rapidly transfer the holes from Modoped BiVO₄ to the electrolyte due to the three-dimensional framework structure of MIL-53(Fe). The junction formed between MIL-53(Fe) and Mo-doped BiVO₄ further suppressed photogenerated electrons and photogenerated holes recombination, thereby improving the photoelectrocatalytic performance.

Recently, Chen et al. prepared an ultrathin Fe-based MOF (NH₂- MIL-101(Fe)) film coated Fe₂O₃ nanorod array anode for effective photoelectrocatalytic water oxidation [65]. First, a Ti-doped Fe₂O₃ nanorod array was vertically grown on FTO glass. After modification with polyvinyl pyrrolidone (PVP), the NH₂-MIL-101(Fe) shell was prepared by solvothermal. In this preparation process, Ti dopants were intended to increase the conductivity of Fe₂O₃. Besides, the PVP was used as a bridging agent because it could be easily adsorbed on Fe₂O₃ and then provided a nucleation site for the growth of MOF. Under simulated AM 1.5 G solar irradiation (100 mW/cm²), the photoanode of the Ti-doped Fe₂O₃ nanorods had a photocurrent density of 0.98 mA/cm² at 1.23 V vs. RHE. After modification by NH₂-MIL-101(Fe), the photocurrent density was raised to 2.27 mA/cm². And the Ti-doped Fe₂O₃/NH₂-MIL-101(Fe) core/shell anode displayed the highest photoelectrocatalytic water oxidation performance. Since the Fe₂O₃ photoanode was greatly affected by poor conductivity and short hole diffusion length, the charge separation efficiency of Ti-doped Fe₂O₃/NH₂-MIL-101(Fe) electrodes was enhanced because of the type Ⅱ heterostructure formed. Under light irradiation, the photogenerated electrons were transferred from the LUMO of NH₂-MIL-101(Fe) to the CB of Tidoped Fe₂O₃, and the photogenerated holes migrated inversely to the HOMO of NH₂-MIL-101(Fe), promoting rapid separation of photogenerated electrons and holes. In addition, efficient charge separation would cause more photogenerated holes to accumulate on the photocatalyst surface, resulting in superior oxidizing power.

3.5. Cu-based MOFs

Cu-based MOFs have been explored in depth for photoelectrocatalysis because Copper is an earth-abundant element and relatively low in cost. As a typical Cu-based MOF, HKUST-1 (Cu₃(BTC)₂, BTC = benzene-1, 3, 5-tricarboxylate)-reduced graphite oxide (rGO)/Pt-Cu foam was prepared as a cathode catalyst for photoelectrochemical reduction of CO₂ irradiated by a 300 W Xenon lamp [66]. The bonding of rGO to Cu₃(BTC)₂ not only produced more active sites but also promoted the faster transfer of electrons on the catalyst surface. The CO₂ conversion rate reached 3256 nmol h^-1 cm^-2, which was much higher than pure Cu₃(BTC)₂ as a cathode catalyst. When -2 V voltage was applied to this photoelectrochemical system, the liquid product had a selectivity to alcohol of up to 94.6%. The Pt nanoparticles on the Cu foam and rGO in the catalyst first transferred electrons to CO₂ to produce C1 intermediate. Due to the good selectivity of Pt for CO production, more CO was produced than other C1 organics. Then, the porous structure of Cu₃(BTC)₂ promoted C—C bonding to produce alcohol.

In addition, HKUST-1 was also reported to be coated on Cu₂O cathode to inhibit photocorrosion and provide reactive sites for the reduction of CO₂ under simulated AM 1.5 G solar irradiation (100 mW/cm²) [67]. This photocathode was formed by depositing Cu₂O film on indium tin oxide (ITO) glass and then growing Cu₃(BTC)₂ by a solution-phase reaction using Cu₂O as a copper source. Various characterizations verified that the Cu₃(BTC)₂ coating not only does not change the bandgap of Cu₂O but also significantly promotes photogenerated charge separation and transport in Cu₂O. Photoelectrocatalytic reduction of CO₂ was completed in a threeelectrode quartz photoelectrocatalytic (PEC) cell containing tetrabutylammonium hexafluorophosphate (TBAPF₆) in CO₂ saturated acetonitrile as an electrolyte, as shown in Fig. 6. The ¹³C isotope labeling experiment confirmed that only CO was produced by reduction, and there were no liquid hydrocarbons and other gaseous products. In the absence of light, the CO yield increased with the applied potential, but the yield of Cu₃(BTC)₂/Cu₂O was 2 times higher than that of Cu₂O. When the PEC cell was irradiated with visible light, the yield of Cu₃(BTC)₂/Cu₂O was significantly improved about 4 times as compared with that of Cu₂O.

3.6. Mixed metal-based MOFs

Mixed metal-based MOFs were defined by Chen et al. [68] in 2011 and have been explored for efficient photoelectrocatalysis. For example, Cui et al. fabricated a ternary semiconductor that coated ZIF-8/67 on ZnO nanorod arrays (NAs), which exhibited high-efficiency photoelectrocatalytic water splitting activity under simulated AM 1.5 G solar irradiation (100 mW/cm²) [69]. ZnO-NAs were synthesized on FTO, then added with Co(NO₃)₂ and 2- methylimidazole for hydrothermal reaction to obtain ZnO@ZIF-8/67-NAs. ZnO@ZIF-8/67-NAs exhibited a strong visible light response, while the absorption edge of ZnO was located near 390 nm due to its wide bandgap. In addition, the one-dimensional structure of ZnO NAs was also beneficial to the rapid transmission of electrons. Under visible light irradiation, the photogenerated electrons generated by ZnO@ZIF-8/67-NAs were transported to the Pt electrode by an external circuit to generate hydrogen, and the photogenerated holes reacted with water to generate oxygen. As a result, the photocurrent density of the ZnO@ZIF-8/67-NAs anode at 1.23 V vs. RHE was 0.11 mA/cm², while that of the ZnO-NAs was only 0.012 mA/cm².

Similarly, mixed metal-based MOF (ZnNi-MOF) was successfully coated on the surface of the ZnO nanowire array for photoelectrochemical reactions under simulated AM 1.5 G solar irradiation [70]. First, ZnO nanowire arrays were synthesized by the hydrothermal method. Then, 1, 4-phthalic acid was added to form MOF-5. Here, ZnO not only provided Zn²⁺ ions for synthesizing MOF-5 but also served as a template. Finally, part of the Zn²⁺ was replaced by Ni²⁺ through ion exchange, thereby obtaining a ZnNi-MOF@ZnO nanowire array. In the dark conditions, the current densities of the three photoanodes (ZnO, MOF-5@ZnO, and ZnNi-MOF@ZnO nanowire arrays) were very low and all about 0.04 mA/cm² at 1.2 V vs. RHE. Under illumination, the photocurrent densities of ZnO, MOF-5@ZnO, and ZnNi-MOF@ZnO at 1.20 V vs. RHE were 0.6, 0.88, and 1.40 mA/cm², respectively, showing a considerable enhancement in photocurrent. The possible mechanism of photoelectrocatalysis of ZnNi-MOF@ZnO was shown in Fig. 7. The photogenerated electrons generated in ZnO by illumination moved to the substrate and then to the counter electrode for producing hydrogen. At the same time, the photogenerated holes generated in ZnO moved to the ZnNi-MOF, and then oxidize Ni²⁺ to Ni³⁺/Ni⁴⁺. Subsequently, the Ni³⁺/Ni⁴⁺ interacted with nearby water to produce oxygen.

Besides, mixed metal-based MOF derivatives can also act as highly efficient photoelectrocatalyst due to their unique structure and the synergistic effect of two uniformly distributed metals. An example of a mixed metal-based MOF derivative was reported by Dou's group [71]. In this article, mixed metal-based MOF (ZIF-8/67) was coated onto ZnO and then sulfurized to prepare ZnO@ZnS/CoS for photoelectrocatalytic water oxidation irradiated by a 300 W Xenon lamp (100 mW/cm²). Due to the framework structure of the original ZIFs, the surface area of ZnO@ZnS/CoS was 191.8 m²/g, which was much larger than that of ZnO powder (15.1 m²/g). And the pore size was mainly distributed between 2 nm and 5 nm, indicating that ZnO@ZnS/CoS possessed a hierarchical microporous and mesoporous structure. For ZnO and ZnO@ZnS/CoS, the fluorescence lifetime values were measured to be 3.03 and 7.06 ns, respectively, showing that the addition of ZIFs derivatives favored charge transfer and surface reactions. The onset potential of ZnO@ZnS/CoS showed a significant negative shift from -0.20 V to -0.35 V, and the photocurrent density was 2.46 mA/cm² at 0.6 V vs. SCE, which was 5.1 times that of ZnO. The improved photoelectrocatalytic activity was due to the unique structure imparting rich exposure sites and efficient light capture to ZnO@ZnS/CoS photoanode. In addition, the cocatalyst CoS could accept photogenerated holes to accelerate rapid hole transport and water oxidation reaction, while photogenerated electrons were quickly guided to the counter electrode through the ZnO rod, effectively suppressing electron-hole recombination.

3.7. Comparison of MOFs photoelectrodes with other photoelectrodes

Table 1 summarizes the photoelectrocatalysis based on MOFs materials in this review, and Table S1 (Supporting information) lists recently reported photocatalysis based on other heterogeneous materials. It can be seen that MOFs materials exhibit comparable or better performance in photoelectrocatalytic water splitting and CO₂ conversion compared to other materials.

4. Conclusions and perspectives

In summary, recent studies have demonstrated that photoelectrocatalysis has great potential for application in many fields including CO₂ conversion, water splitting for hydrogen evolution reaction (HER), and OER. Photoelectrocatalysis, coupled with heterogeneous photocatalysis and electrocatalysis, possesses the advantages of both methods. Various researches have shown that photoelectrocatalysis has better catalytic performance under similar conditions than photocatalysis and electrolysis. Since the anode material is a key component in the photoelectrocatalytic device, enormous efforts have been made to develop photoelectrocatalytic materials that achieve excellent catalytic performance under mild conditions. Among these photoelectrocatalysts, the emerging MOFs have attracted much attention because of their porosity, large surface area, and tunable structure. MOFs can form heterojunctions with other semiconductor materials and can be used as co-catalysts for photoelectrocatalytic processes. Moreover, MOFs derivatives can also be used to construct photoelectrodes due to their inherited porous structures from the pristine MOFs. In photocatalytic processes, MOFs photoelectrodes play triple roles: (ⅰ) promoting charge separation and transfer of electrons to the active sites, (ⅱ) improving photoelectrode stability and protecting photoelectrode from photocorrosion, and (ⅲ) providing more active sites to catalyze CO₂ conversion and water splitting. There is no doubt that the photoelectrodes containing various metal-based MOF materials exhibit high activity.

However, in order to finally realize the practical application of this technology, there are still a lot of challenges for MOFs in photoelectrocatalysis. First, the bandgap of MOFs determines the range of light absorption. For MOFs with a narrower bandgap, a stronger visible light response can be expected. Therefore, it is feasible to decrease the bandgap of MOFs by various methods such as modifying MOFs and developing new MOFs. In addition, the transmission speed of photogenerated electrons and photogenerated holes is also critical to the photoelectrode material. Constructing a rational heterostructure consisting of MOFs with other semiconductor materials is essential to promote charge separation. Moreover, the stability of the photoelectrode material must be high for long-term photoelectrocatalytic reactions. Therefore, because the stability of MOFs is usually limited, improving its stability is also an important research direction. Apart from these, the mechanism of photoelectrocatalytic reactions has not yet been fully elucidated, and more research efforts are needed. Combining experimental results with theoretical calculation (e.g., density functional theory calculation) to explore photoelectrocatalytic reaction mechanisms is a promising approach. Although there is still a lot of work to be done, MOFs are still good candidates for photoelectrocatalytic reactions, and the prospects for this field are bright.

Declaration of competing interest

The authors declare that no conflict of interest exits in the submission of this manuscript, and the manuscript is approved by all authors for publication.

Acknowledgments

This work was supported by the Natural Science Foundation of Jiangsu Province of China (Nos. BK20160430, BK20181070), the National Natural Science Foundation of China (No. 51808250), the Project Funded by China Postdoctoral Science Foundation (Nos. 2016M591757, 2017M610336), and the Jiangsu Planned Projects for Postdoctoral Research Funds of China (No. 1601179C).

Appendix A. Supplementary data

Supplementarymaterial related to this article can befound, in the online version, at doi: https://doi.org/10.1016/j.cclet.2019.12.015.