Recent Progress on Engineering Highly Efficient Porous Semiconductor Photocatalysts Derived from Metal–Organic Frameworks

Highlights In this review, we survey the recent developments in the fabrication of metal–organic framework (MOF)-derived porous semiconductor photocatalysts toward four kinds of energy-/environment-related reactions. A comprehensive summary of highly efficient MOF-derived photocatalysts, particularly porous metal oxides and metal sulfides, and their heterostructures are provided. Enhanced photocatalytic performance achieved with MOF-derived porous heterostructures as the photocatalyst is discussed in detail.

Porous micro-nanostructures can offer highly accessible surfaces and rich pores, which favor the exposure of numerous active sites in reactions, shorten the transfer distance to the pore surface for photoexcited carriers, and provide unrestricted diffusion of substrates and products, leading to their excellent performances in photocatalysis [27][28][29][30][31][32][33]. In addition, fabrication of semiconductor-based heterojunctions, including semiconductor-metal heterojunction and semiconductor-semiconductor heterojunction, is another useful strategy for enhancing the photocatalytic activity [10,[34][35][36][37][38][39][40][41][42]. Proper heterojunctions can tune the band gap, encourage the separation and migration of photogenerated electron-hole pairs, and enhance the efficiency of light utilization. In the last few years, much progress has been made in the design of photocatalysts with porous structures and heterostructures; however, the rational design of photocatalysts is hard because of complicated processes in traditional synthesis. Therefore, efficient and easily preparable photocatalysts with beneficial structural features are desired.
Recently, MOFs, well known as porous coordination polymers consisting of metal nodes and organic ligands, have attracted much attention and shown great potential for various applications, including photocatalysis, due to their fine-tuned structures, high specific surface areas, controlled pore structures, and various components [43][44][45][46][47][48][49]. Moreover, via the well-designed modification of MOFbased materials, not only the reactant adsorption and light absorption but also the charge separation and reactant activation can be largely promoted, leading to enhanced photocatalytic performances [50][51][52][53]. However, the poor stability and poor electronic conductivity of MOFs hinder their usage in the photocatalytic field. As an alternative, recent studies have found that MOFs can serve as precursors for porous semiconductor materials, including porous metal oxides, carbon materials, and metal sulfides, and their heterostructures, through the facile thermal treatment or sulfidation process [54][55][56][57][58][59][60]. Via the controlled derivation of MOFs in certain conditions, the as-synthesized products can maintain some of the initial structural features of parent MOFs or promote electrical conductivity, while maintaining the open diffusion channels and ensuring the monodispersion of metal centers, making them very promising in photocatalysis. More interestingly, porous heterostructures or solid solutions can be rationally derived from MOFbased hybrids or multimetallic MOFs, and the procedure has the following merits: (1) the versatility of MOFs in metal nodes and ligands endows enough choices for fabricating heterostructures or solid solutions, and the band gap of the obtained derivatives can be easily tuned by altering the metal or component ratios; (2) the in situ synthesis with MOFs as the precursor prevents the poor structural stability and weak coupling between the individual components of a heterostructure and solid solution; and (3) the in situ uniform distribution of metal nodes in multimetallic MOF precursors at the molecular level can contribute to an increase in the active sites, which can effectively participate in photocatalytic reactions. Therefore, MOFs can act as ideal precursors for rationally designed photocatalysts with enhanced performances.
In the past few years, a lot of progress has been made on porous photocatalysts, including porous metal oxides, porous metal sulfides, and porous carbon, and their heterostructures, derived from monometallic MOFs, multimetallic MOFs, or MOF-based hybrids for water-splitting reactions, pollutant degradation, CO 2 reduction, and organic synthesis. A timeline showing the breakthrough in the fabrication of highly efficient MOF-derived photocatalysts is shown in Fig. 1. With features of a porous structure and/or heterostructure, 1 3 which are beneficial for increasing light utilization efficiency and promoting the separation and migration of photoinduced electron-hole pairs, these MOF-derived photocatalysts have acted well in enhancing photocatalytic performances. However, a specialized discussion on the progress achieved in the photocatalytic application of MOF-derived porous materials is very rare. Herein, we endeavor to give a comprehensive summary of the progress in four parts, i.e., (1) photocatalytic water splitting, including the photocatalytic hydrogen evolution reaction and oxygen evolution reaction; (2) photocatalytic degradation of pollutants, particularly dye pollutants; (3) photocatalytic CO 2 reduction to CO or hydrocarbons; and (4) photocatalytic organic reactions. A brief summary and outlook are included in the final section.

General Methods
Firstly, we give a brief summary of the synthetic strategies used to fabricate MOF-derived photocatalysts, including porous metal oxides, porous metal sulfides, and carbon materials. Depending on the different types of MOF-derived photocatalysts, the synthetic strategies can be summarized, as given in Table 1

Photocatalytic Water Splitting
Hydrogen, which can be renewably produced from a variety of (non-fossil) feedstocks, is a globally accepted clean energy carrier. Water splitting, which involves two half-reactions, i.e., hydrogen evolution reaction (HER) and oxygen evolution reaction (OER), has been considered as an attractive route to sustainable H 2 generation [75]. Photocatalytic water splitting into H 2 and O 2 is a typical uphill reaction with a positive Gibbs energy change (ΔG = + 237.13 kJ mol −1 ), which requires a photocatalyst to trigger the reaction and convert solar energy to storable hydrogen energy [76]. Thermodynamically, the conduction band (CB) of the photocatalyst must be located more negative than the H + /H 2 energy level (− 0.41 eV vs NHE at pH 7) and the valence band (VB) of the photocatalyst must be located more positive than the O 2 /H 2 O energy level (+ 0.82 eV vs NHE at pH 7) [77]. However, the overall water splitting requires higher photon energy than 1.23 eV due to the large overpotential caused by the charge transfer process and interaction between catalysts, reactants, and products. Therefore, the development of an overall water-splitting system remains a great challenge, and photocatalytic water splitting is generally studied separately.
To date, many efficient porous photocatalysts derived from MOFs have been developed for HER; however, more progress needs to be made on OER (Table 2).

Photocatalytic HER
In the hydrogen evolution half-reaction system, the other half-reaction is replaced by the oxidation of an appropriate sacrificial reductant, such as methanol, ethanol, triethanolamine, triethylamine, ascorbic acid, lactic acid, and Na 2 S/ Na 2 SO 3 pairs [77]. Similar to that for other photocatalytic systems, the design principles for the HER photocatalysts involve a suitable band gap, enhanced charge transfer efficiency, and numerous active sites. MOFs serve as ideal precursors for highly efficient HER photocatalysts because of their ability to provide various metal ions (as options), facilitate the doping of C or N to MOF-derived semiconductors, and give porous structures with high surface areas.

MOF-Derived Porous Metal Oxides
Metal oxide nanostructures, one of the most important semiconductor nanomaterials, have attracted much attention as photocatalysts for water splitting. TiO 2 has been considered the most interesting photocatalyst for water splitting, especially for HER, due to its suitable band positions, low cost, low toxicity, high stability, and n-type semiconducting nature [78,79]  As it is an n-type semiconductor, TiO 2 can be combined with p-type semiconductors to fabricate p-n heterojunctions with enhanced photocatalytic performance due to more effective charge separation; rapid charge transfer to the surface of catalyst; longer lifetime of the charge carriers; and separation of locally incompatible reduction and oxidation reactions in nanospace [40,41]. The versatile metal nodes of MOFs facilitate the facile and rational design of TiO 2 -based p-n heterojunctions. For example, Mondal et al. reported an improved Co 3 O 4 /TiO 2 photocatalytic system with p-n heterojunction derived from several newly developed Co-MOFs [84]. As mentioned, the obtained p-n heterojunction consisted of spinel Co 3 O 4 and anatase TiO 2 . With an optimized Co loading of 2 wt%, the Co 3 O 4 /TiO 2 could deliver hydrogen at a high rate of 7000 µmol (g h) −1 under UV-Vis light due to the synergistic effect of the formed small heterojunction and cocatalytic role of Co 3 O 4 , which facilitated interfacial charge transfer and electron-hole separation.
Other metal oxide (In 2 O 3 , ZnO, Fe 2 O 3 , etc.)-based porous photocatalysts with high activities for hydrogen evolution have also been easily prepared with MOFs as templates and/or precursors [67,[85][86][87]. For instance, an In 2 O 3 -based photocatalyst, namely PHIC, with highly improved activity for hydrogen evolution was fabricated via a facile thermal decomposition of In-MIL-68 template by our group (Fig. 3) [86]. This PHIC catalyst with a hollow hexagonal micro-rod shape was assembled using In 2 O 3 @carbon core-shell nanoparticles. Due to the synergistic effect of efficient separation of photogenerated electron-hole pairs caused by the carbon Reprinted with permission from Ref. [66] coating, enhanced optical absorption attributed to hollow characters, and improved accessibility rendered by the porous structure, PHIC could exhibit a photocatalytic activity comparable to that of the Pt/In 2 O 3 photocatalyst toward hydrogen evolution with an extremely high production rate of 2,700,000 µmol (g h) −1 under solar-simulated light. In addition, ZnO, a typical semiconductor with absorption in the UV region, has been optimized as a better photocatalyst by compounding it with other materials, such as noble metal nanoparticles (NPs) and other metal oxides. With Zn-MOF/ metal NP hybrids as the precursor, ZnO/metal NP heteromaterials with porous structures could be easily obtained [67]. Wang group fabricated the Au/ZnO NP photocatalyst with yellow fluorescent GSH-Au nanoclusters (NCs)/ ZIF-8 NPs as the precursor, and this photocatalyst could extend the absorption of ZnO to the visible-light region. The obtained An/ZnO NP photocatalyst could achieve a hydrogen generation rate of ~ 29 µmol (g h) −1 under visible-light illumination.

MOF-Derived Porous Metal Sulfides
Cadmium sulfide (CdS) is another promising photocatalyst for solar-driven hydrogen evolution due to its visible-light response (E g = 2.4 eV) and conduction band located at a suitable energy level. However, there are two drawbacks for the CdS photocatalyst, i.e., easy recombination of photogenerated hole-electron pairs and high photocorrosion in aqueous media, which restrict its practical application. In order to enhance the photocatalytic activity of CdS, various strategies have been introduced, including increasing specific surface area and modifying CdS with cocatalysts or incorporating other materials to form solid solutions, each of which could be handily achieved with the MOF-derivation strategy. Notably, CdS materials with high surface areas as photocatalysts may still suffer from poor stability, but they can demonstrate improved activities by promoting the separation of photoinduced hole-electron pairs. For example, Xiao and Jiang synthesized a hierarchically porous CdS (HP-CdS) utilizing thermally stable MOFs as hard templates to afford porous CdO, and subsequently, CdS by a nanocasting method ( Fig. 4)   1 µm   [86] surface area of 119 m 2 g −1 , which facilitated the effective inhibition of photogenerated electron-hole recombination, leading to a higher photocatalytic activity for HER than nano-CdS and bulk CdS under visible light (> 380 nm). In addition, CdS with complex hollow nanostructures, which provides multiple structural advantages for the photocatalytic reaction, such as enhanced light absorption, improved separation efficiency of the photoinduced charge carriers, and increased specific surface area, could be obtained with MOFs as templates or precursors. Wang [73]. With an optimized loading content of CdS (30%) on the ZCO surface, CdS/ZCO achieved a high H 2 production rate of about 3978.6 μmol (g h) −1 , which was attributed to the synergistic effect, i.e., the efficient charge separation and transfer between the phase boundary of CdS and ZCO. Similarly, Co-MOF-derived onion slice-type hollow-structured Co 4 S 3 was developed and decorated with CdS nanoparticles for photocatalytic hydrogen production by Kim's group [70]. The optimized Co 4 S 3 / CdS material led to an enhanced rate of H 2 generation of 12,360 μmol (g h) −1 under simulated solar light irradiation. The low density, hollow interior, and shell permeability of the onion-type composite helped in accelerating the charge separation and transfer in photocatalytic reactions. They also prepared MOF-derived Ni 2 P nanoparticles as the cocatalyst of CdS, and the Ni 2 P/CdS heterostructure exhibited great improvement in performance during photocatalytic HER due to the decreased rate of charge carrier recombination [89]. Moreover, a family of photocatalysts (NiS/Zn x Cd 1−x S) for HER has been developed by decorating CdS with NiS as a cocatalyst and simultaneously incorporating Zn to form solid solutions [72]. As shown in 1 3 form the NiS/Zn x Cd 1−x S series via solvothermal sulfidation and thermal annealing. By adjusting the doping metal concentration in the MOFs, the chemical compositions and band gaps of the heterojunctions were fine-tuned, leading to an optimized HER rate of up to 16,780 μmol (g h) −1 with NiS/ Zn 0.5 Cd 0.5 S as the photocatalyst under visible-light irradiation. In-depth DFT calculations revealed the importance of NiS in accelerating the water dissociation kinetics, which was crucial for photocatalytic HER. Notably, a visible-light catalytic system with a HER rate of up to 17,360 μmol (g h) −1 over a Co-Zn 0.5 Cd 0.5 S solid solution catalyst was developed via a sulfidation process of Co/Zn-ZIF in the presence of Cd 2+ [90]. The doping of Co evenly around the skeleton of the porous Zn 0.5 Cd 0.5 S solid solution played an important role in improving the photocatalytic activity, as compared to that of Zn 0.5 Cd 0.5 S, which displayed a HER rate of 12,130 μmol (g h) −1 under the same reaction condition.
As per the reports mentioned above, utilizing MOFs as templates or precursors to construct heterojunctions between a semiconductor material and another semiconductor material or a noble metal material is an effect way to improve the performance of semiconductors in photocatalytic HER by water splitting. Accordingly, Li et al. reported a series of Zn/ Co-ZIF derivative/Pt photocatalytic systems, which exhibited high performance toward the photocatalytic HER under UV-Vis light (shown in Fig. 6) [65]. The synthesis processes involved the oxidation, sulfurization, or phosphorization of ZnCo-ZIF, and the subsequent photochemical doping of Pt nanoparticles, leading to Pt-ZnO-Co 3 O 4 , Pt-ZnS-CoS, and Pt-Zn 3 P 2 -CoP, respectively. The porous framework skeleton of the ZnCo-ZIF derivatives significantly enhanced the light utilization and simultaneously afforded abundant exposed catalytic active sites; the suitable band matching and strong electron coupling in the heterojunctions facilitated efficient electron-hole separation and transportation; the distribution of Pt nanoparticles on the porous structure offered enough redox active sites. These beneficial features were revealed to be responsible for the highly enhanced performances of the Zn/Co-ZIF derivative/Pt photocatalysts. This bimetallic MOF-directed fabrication strategy reported in this paper provided a new perspective to construct synergetic photocatalysts with excellent photocatalytic performances for water-splitting applications.

MOF-Derived Porous Carbon Materials
In general, progress on the photocatalytic hydrogen evolution from water over MOFs-derived photocatalysts have been made mainly around metal-involving materials, like metal oxides, metal sulfides, or their composites, which are commonly used in the photocatalytic field. As alternatives, metal-free photocatalysts, such as carbon-based materials, demonstrating great performance are highly desired, due to their tunable molecular structures, abundance, and high chemical stability. Meanwhile, facile synthesis of metal-free photocatalysts is still highly challenging. Cheng and coworkers have developed a facile synthesis method for N-doped graphene carbon using a well-designed ZIF-8-template ( Fig. 7) [58]. Under an Ar atmosphere at different temperatures, ZIF-8 were calcined to derive N-doped graphene (ZNG) analogs, which retained the polyhedron structure of the parent ZIF-8 particles and had nitrogen contents of 9-15 wt%; the contents of various N types in the materials were fine-tuned on the basis of the calcination temperatures. Among them, the product obtained at 1000 °C exhibited the best performance toward photocatalytic HER because of the highest content of graphitic nitrogen, which preserved the high mobility of the charge carriers and further affected the hydrogen evolution rate of the photocatalyst.

Photocatalytic Water Oxidation
Water oxidation to O 2 is regarded as the bottleneck of solardriven water splitting, ascribed to the intrinsic difficulty in multiple-electron transfer and sluggish kinetics of the subsequent oxygen evolution [97,98]. Although some noble metals and their oxides, such as IrO 2 and RuO 2 , have attracted much attention as efficient catalysts for water oxidation, their high cost and low abundance have impeded their commercial utilization. Therefore, first-row transition metal oxides and their derivatives, including cobalt oxides, iron oxides, nickel oxides, and manganese oxides, have been explored as water oxidation catalysts due to advantages with regard to economy and stability [99][100][101][102][103][104][105]. In the water oxidation reaction system (i.e., OER), the reduction of sacrificial oxidants, such as AgNO 3 and Na 2 S 2 O 8 , is normally used to replace the half-reaction of HER [77]. According to investigations, MOF-derived water oxidation photocatalysts, particularly cobalt oxide-based systems, are environmentally benign, thermally stable, inexpensive, and demonstrate high OER activity both in the electrochemical and in the photochemical fields. The Lu group developed a facile approach for the preparation of a porous cobalt oxide-carbon hybrid as a water oxidation catalyst by carbonizing nanocrystals of ZIF-67 in an inert atmosphere and subsequently air-calcining them [92]. Among the various CoO x /C hybrids obtained at different calcination temperatures, 700-CoO x /C acted as the best catalyst. The photocatalytic activity of 700-CoO x /C was accessed in the [Ru(bpy) 3 ] 2+ -S 2 O 8 2− system under visible light in a sodium phosphate buffer (pH 8.5) and a maximum turnover frequency (TOF) of up to 0.039 ± 0.03 s −1 per cobalt atom was estimated, which was among the highest TOFs for water oxidation with cobalt oxide-based photocatalysts. Furthermore, by combining ZIF-67 with suitable materials as precursors, adjustable cobalt oxide-based composites as photocatalysts for water oxidation could be fabricated. For example, the derivatives obtained by loading a single Keggin-type polyoxometalate (POM) cluster into each confined space of ZIF-67, i.e., POM@Co 3 O 4 composites (CWn, n depended on the added amount of POM) doped with highly dispersive molecular metal-oxo clusters, exhibited significantly improved photocatalytic activity in water oxidation compared to the pure MOF-derived nanostructure (as shown in Fig. 8) [93]. In the molecular cluster@oxide system, POMs accept and release electrons, thereby improving the separation of light-induced electrons and holes, which leads to higher catalytic activity with increasing POM concentration in the composite materials. Ding's group synthesized a series of Co 3 O 4 /CuO hollow polyhedral nanocages (HPNC) using ZIF-67/Cu hydroxide (HD) polyhedrons with various Co/Cu molar ratios as sacrificial templates [68]. With an optimized Co/Cu molar ratio, Co 3 O 4 /CuO-3 HPNCs afforded a high TOF of 4.9 × 10 −3 s −1 per metal atom and performed well in the stability test.
In addition, Prussian blue analogue (PBA), a type of crystalline MOF built from divalent and trivalent metal ions (such as Fe, Co, and Ni) linked by cyanide ligands, has also been chosen as the precursor to fabricate water oxidation photocatalysts, such as Fe/Co In summary, tremendous efforts have been devoted to develop efficient photocatalysts toward water splitting with MOFs or their hybrids as the precursors, and great progress has been made in this regard. However, the currently developed photocatalytic systems based on MOF-derived porous structures only involve one isolated half-reaction, and sacrificial reagents are needed to replace the other half-reaction. More studies have been performed on hydrogen evolution reaction than oxygen evolution reaction due to the intrinsic difficulty associated with OER in multiple-electron transfer and sluggish kinetics. However, the practical application of water splitting demands the production of stoichiometric amounts of H 2 and O 2 driven by sunlight without any dodecahedra (inset is the SAED pattern recorded for the whole particle). f HRTEM image of CW-2 hollow dodecahedra (inset is the SAED pattern recorded from the whole particle). g TOF and h transient photocurrent response of Co 3 O 4 and CW-n under visible-light irradiation (from left to right: Co 3 O 4 , CW-1, CW-2, and CW-3). Reprinted with permission from Ref. [93] sacrificial reagents. Therefore, photocatalytic overall watersplitting systems with high efficiency are highly desired. Moreover, the porous structures of photocatalysts can influence the attachment of gas bubbles on the surface of the catalysts, which can potentially block the active sites and prohibit mass transportation under strong gas evolution conditions, leading to bubble overpotential [106]. As far as we know, no investigation on the influence of pore structure of MOFs-derived photocatalysts on overpotential for watersplitting reaction, especially OER, has been performed, and therefore, further studies are needed.

Photocatalytic CO 2 Reduction
The global climate issues caused by the rapidly increasing CO 2 emissions have accelerated studies on photocatalytic CO 2 reduction to CO or hydrocarbons, which can not only decrease the CO 2 level in the atmosphere, facilitating environmental protection, but also generate materials for chemical industry or energy storage. Nowadays, the photocatalytic CO 2 reduction process is attracting growing attention because of its usage of solar power as the primary energy source. Photocatalytic CO 2 reduction involves the following steps: light harvesting, separation, transfer of photogenerated charge carriers, as well as adsorption and conversion of CO 2 [107]. Therefore, in addition to the impacts of light utilization, charge transfer efficiency, and active sites, the adsorption of CO 2 is another important factor to expedite the CO 2 reduction process in a photocatalytic system. The CO 2 reduction products include HCOOH, CO, HCHO, CH 3 OH, and CH 4 . MOF-derived materials are considered highly potential photocatalysts for CO 2 reduction due to their advantages such as structural and compositional variety and high surface areas. Some excellent works have been reported on MOF-derived materials as CO 2 reduction photocatalysts (Table 3).
Ye's group developed a MIL-101-derived Fe@C catalyst, consisting of an iron core (< 10 nm) and ultrathin (1-3 layers) carbon layers, for the photocatalytic reduction of CO 2 to CO [108]. The obtained Fe@C photocatalysts could produce 2196.17 µmol CO under broadband light irradiation for 120 min, which was better than that reported for most of the catalysts. Through a thorough investigation, they found that the intense adsorption of visible light and infrared radiation induced a thermal effect, which helped to drive the reaction, and UV-light-induced iron local surface-plasmon resonances also played a significant role in activating the nonpolar CO 2 molecules. Furthermore, DFT calculations revealed that the ultrathin layers of carbon shells on the Fe nanoparticles dramatically promoted desorption of the produced CO from the catalyst surface, thereby increasing the CO selectivity (99.76%). Another study on the photocatalytic CO 2 reduction to CO with MOF-derived materials as catalysts was reported by Lou's group; they rationally fabricated hierarchical In 2 S 3 -CdIn 2 S 4 heterostructured nanotubes via sequential ion exchange reactions with MIL-68 as the initial precursor ( Fig. 9) [74]. Benefiting from the unique structural and compositional features, such as nanosized interfacial contacts between In 2 S 3 and CdIn 2 S 4 nanospecies, reduced diffusion length for charge carrier separation and migration, large surface area for CO 2 adsorption and concentration, and rich catalytically active sites for surface redox reactions, the obtained hierarchical In 2 S 3 -CdIn 2 S 4 nanotubes manifested an optimized CO generation rate under visible light. Very recently, a three-component heterojunction C-Cu 2−x S@g-C 3 N 4 photocatalyst was found to be active toward the reduction of CO 2 to CO. With MOF-199 as the precursor, hollow tubular Cu 2−x S with carbon coating was successfully    Reprinted with the permission from Ref. [74] 1 3 fabricated, and further, different amounts of the material were loaded on g-C 3 N 4 to form a series of C-Cu 2−x S@g-C 3 N 4 photocatalysts. Under visible-light irradiation and water vapor condition, the optimized C-Cu 2−x S@g-C 3 N 4 with 0.71 wt% of C-Cu 2−x S exhibited a high reactivity of 1062.6 μmol g −1 and selectivity of 97% [109]. Photocatalytic reduction of CO 2 to hydrocarbons by water vapor with proper catalysts has also been taken into account. Fischer and coworkers have fabricated an Au NP/TiO 2 (GNP/TiO 2 ) photocatalyst active for the reduction of CO 2 to CH 4 [110]. After deposition of preformed and surfactant-stabilized gold nanoparticles on the surface of NH 2 -MIL-125, GNP/NH 2 -MIL-125 was transformed to GNP/TiO 2 through a thermal treatment; the product contained rutile TiO 2 and possessed the morphology of NH 2 -MIL- 125. Compared with the TiO 2 sample obtained via the pyrolysis of NH 2 -MIL-125 without GNP, P25, and commercial Au/TiO 2 , the GNP/ TiO 2 photocatalyst delivered a significantly higher yield of CH 4 due to the presence of GNP. Metal oxide composites obtained with MOFs as precursors were also utilized as catalysts for the photocatalytic reduction of CO 2 to hydrocarbons by Zhang's group [64]. They fabricated a series of ZnO/ NiO porous hollow spheres with sheet-like subunits through the calcination of bimetallic Zn/Ni-MOFs with different ratios of Zn/Ni. The p-n heterojunctions formed by p-type ZnO and n-type NiO, and the hollow character enhanced the performance of ZnO/NiO composites by facilitating the separation of the photogenerated hole-electron carriers. As a result, the optimized ZnO/NiO composites (ZN-30) acted three times better than pure ZnO in the photocatalytic reduction of CO 2 to CH 3 OH.
To date, despite the high potential in sustainability and energy storage, only few studies have been focused on the photocatalytic reduction of CO 2 with MOF-derived materials as catalysts. Therefore, more progress should be made on developing more efficient photocatalysts that can enhance the utilization of solar energy and perform better, in terms of both the activity and selectivity.

Photocatalytic Organic Reaction
The photocatalytic synthesis of organic compounds is a promising approach due to its mild, clean and atom efficiency methodologies, compared with the methods involving high temperature and pressure, and the current industrial synthetic strategies, which generate harmful by-products [112]. Photoinduced charge carriers transferred from semiconductor nanomaterials to organic molecules can catalytically trigger a variety of organic redox reactions. Recently, functionalized MOFs have served as interesting photocatalysts for various organic redox reactions due to their unique, tailorable, and highly porous characteristics. However, the poor thermal stability of MOFs hampers their commercialized utilization [51]. On the other hand, MOF-derived semiconductors, which can inherit the porous structure and diversity in structure from parent MOFs, as photocatalysts, have the advantages of high surface areas, high stability, and controllable band gaps, which can match the different HOMO-LUMO positions of the organic molecules. The photocatalytic mechanism of organic redox reaction shares similar process with other photocatalytic systems, i.e., light adsorption, separation and transfer of photoinduced charge carriers, and the subsequent redox reactions on the surfaceactive sites of the photocatalysts. In Table 4, some examples of MOF-derived porous structures exhibiting high performance in visible-light-driven photoredox catalysis are given.
Long-lived photogenerated carriers play an important role in improving the activity of photocatalysts. Multiple strategies, including structural engineering, semiconductor compositing, doping and so on, have been developed to enhance the transfer and separation efficiency of electrons and holes. The facile method with MOFs as the templates and precursors has been proven ideal for engineering photocatalysts with suitable structures and compositions. Our group have reported N-doped Cu 2 O@N-C NP catalysts derived from a N-rich NTU-105, which share several favorable features for photocatalysis: (1) a porous C matrix substrate, which facilitated uniform distribution of the small Cu 2 O nanoparticles by stabilizing them and preventing their aggregation and (2) high conductivity, attributed to the N-doped Cu 2 O and C substrate, which facilitated electron and hole transfer and separation [63]. The N-doped Cu 2 O@N-C NPs exhibited excellent performance in cross-dehydrogenative coupling (CDC) reactions, owning to the long-lived holes, whose existence was revealed by femtosecond transient absorption spectroscopy. By utilizing NH 2 -MIL-125 as a hard template, N-doped TiO 2 @N-C with butterfly structure (TNPC) was also fabricated as a highly efficient photocatalyst for visiblelight-induced amine oxidation [113]. Through an in-depth investigation with femtosecond transient absorption spectroscopy and DFT calculation, the number of actively longlived holes was found to be in the order TiO 2 < N-doped TiO 2 < TNPC, leading to significantly enhanced photocatalytic activity of TNPC. Moreover, MOFs have been realized as the templates for the construction of multilevel structures of composites, which can improve the utilization efficiency of light through the multiple reflections, thereby enhancing the activity of photocatalytic organic reactions. Combined with the structural features of N doping and C compositing, Cu 2 O nanoparticles anchored on an N-doped porous carbon yolk-shell cuboctahedral (CNPC) framework were fabricated with a benzimidazole-modified Cu-btc MOF as the precursor, which possessed a multilevel structure at the same time ( Fig. 10) [114]. Benefiting from the structural characteristics, the obtained CNPC nanoparticles, as a photocatalyst, provided several favorable features: prolonged lifetime of photogenerated electrons and holes, multiple reflection of light by the yolk-shell structure, and improved stability and dispersibility of Cu 2 O NPs, leading to excellent performance in CDC reactions under visible-light illumination.
In general, research on photocatalytic organic reactions over MOF-derived catalysts is still in its early stage, and more progress, such as development of more efficient photocatalytic systems from MOFs and broadening their applications in photoredox catalysis, needs to be made.

Photocatalytic Pollutant Degradation
Organic pollutants in water, as an overwhelming problem in environmental chemistry, have raised many concerns. Photodegradation of organic pollutants is considered an ideal strategy to solve the issue, which usually involves O 2 as the oxidizing agent to degrade organic pollutants to CO 2 , water, and other inorganic species. Among the various photocatalytic systems, MOF-derived materials, such as porous metal oxides, metal sulfides, and their composites (mostly combined with carbon materials) can serve as highly potential photocatalysts due to the merits of high stability, excellent optical absorption/mass transfer, and improved electron-hole separation. In this part, we will summarize the progress in MOF-derived photocatalysts for the degradation of pollutants, which are mostly normal dye pollutants, such as methylene blue (MB), methyl orange (MO), and rhodamine B (RhB) (Fig. 11), and organic pollutants, for example, nitrobenzene. In Table 5, the performances of MOF-derived photocatalysts for pollutant degradation are presented.

MOF-Derived Porous Metal Oxides
As the most commonly used metal oxides in photocatalysis, TiO 2 and ZnO show high activities for pollutant photodegradation. With proper MOFs as the precursors, TiO 2 -and ZnO-based photocatalysts with highly improved performances could be obtained via a simple thermal treatment. For example, MIL-125, one of the earliest reported Ticontaining MOFs, is a good choice to fabricate TiO 2 -based photocatalysts, and according to reports, tunable products could be obtained by changing the calcining atmosphere. Pan's group achieved cake-like porous TiO 2 particles with the mixed anatase/rutile (A/R TiO 2 ) phase via the pyrolysis of MIL-125 under air atmosphere; the as-prepared materials served as photocatalysts for nitrobenzene degradation [116]. Due to the reduced electron-hole pair recombination, the obtained A/R TiO 2 showed better catalytic activity compared with pure rutile or anatase TiO 2 . Zhao and coworkers found that a series of TiO x /C composites could be obtained by the thermal treatment of MIL-125 under an Ar atmosphere; these composites exhibited high catalytic activities toward the photodegradation of MB under UV light [117]. In the series of TiO x /C composites, the product achieved at 1000 °C acted as the best photocatalyst due to its Reprinted with the permission from Ref. [114] high surface area, reduced Ti 3 O 5 composition, and conductive carbon matrix. In addition, Zhao and coworkers also reported that the heterometallic MOF (ZTOF-1) containing Zn 2+ and Ti 4+ could be used as the precursor to fabricate TiO x -based catalysts, whereby Zn could be removed via vaporization at 1000 °C, leading to the TiO x /C composite [118]. The obtained TiO x /C contained the extra pores formed in the process of vaporization of Zn, and these were readily accessible to MB, leading to a considerable increase in the   . g-i are reprinted with permission from Ref. [120]. j-n are reprinted with permission from Ref. [62] photocatalytic activity compared with that of the pyrolytic products obtained below 1000 °C. ZIF-8 or its hybrids have been commonly utilized as the precursors to fabricate ZnO-based materials as catalysts for pollutant photodegradation [67,115,119]. In the past few years, various ZnO-based photocatalytic systems, including porous ZnO, porous C-or/and N-doped ZnO, ZnO/ carbon materials composites, and other ZnO-based composites, were fabricated (Fig. 12). With pure ZIF-8 as the template and precursor, which contains ligands with nitrogen, i.e., 2-methylimidazole, the resulting products could be tuned as porous ZnO, porous C-doped ZnO, and porous C,N-doped ZnO by changing the calcination conditions, including calcination atmosphere, calcination step, and calcination temperature [61,62,96], while, with the mixture of ZIF-8 and urea as the precursors, N-doped ZnO could be obtained [120]. A photocatalytic investigation on these catalysts proved that they all performed well in dye degradation; particularly, porous C-doped ZnO reported by Zhang and coworkers and porous N-doped ZnO reported by Yao's group exhibited visible-light photodegradation of RhB due to the narrower band gap of C-and N-doped ZnO [62,120]. When combined with carbon materials such as three-dimensional graphene networks (3DGN), ZIF-8-based hybrids could be transformed to the corresponding ZnObased hybrids [121]. Zhang's group developed a method to synthesize the ZnO/3DGN composite via a two-step annealing process with ZIF-8/3DGN as the precursor, whereby ZIF-8/3DGN was fabricated by the direct synthesis of ZIF-8 on 3DGN. The obtained ZnO/3DGN exhibited high activity in MB photodegradation as well as durability. Zhu and coworkers incorporated RGO to ZIF-8-derived ZnO via the microwave-assisted method to fabricate ZnO/RGO hybrids [122]. Attributed to the synergistic effect of enhanced light absorption and suppression of charge carrier recombination resulting from the interaction between ZnO and RGO, the ZnO/RGO composites were able to show higher activity for the visible-light degradation of MB than pure ZnO. In the series of composites with different amounts of RGO, the as-prepared composite with 1.5 wt% of RGO showed an optimal photocatalytic performance.
In addition to TiO 2 and ZnO, other metal oxide-based photocatalysts for the pollutant degradation have also been derived from the corresponding MOF materials [123][124][125]. In 2007, Madras and Natarajan et al. reported the photodegradation of four organic pollutants, including RhB, Rhodamine Blue (RBL), Orange G (OG), and Remazol Brilliant Blue R (RBBR), with a bimetallic Gd/Co-MOFderived nanosized GdCoO 3 as the catalyst [126]. At different calcination temperatures, GdCoO 3 with different particle sizes could be obtained. Compared with P25, the as-synthesized ~ 3 nm GdCoO 3 particles showed higher rates of degradation rates for the four pollutants. Wang's group has fabricated In 2 O 3 /Co 3 O 4 -palygorskite (PAL) composite photocatalysts for the degradation of MB and tetracycline (TC) from a Co/In-MOF/PAL hybrid [127]. In the presence of PAL, the photocatalytic activities were improved due to their higher adsorption of the substrates and the electrostatic interactions between photogenerated charge carriers and negatively charged PAL. Two-dimensional (2D) porous N-doped Cu 2 O/C composites derived from square-shaped Cu-MOF nanoplatelets as a photocatalyst for MO degradation were reported by Ma and coworkers [128]. The high surface area of the annealed Cu 2 O/C nanoplatelets led to a high degradation rate of 2.5 mg min −1 g −1 for MO pollution under visible light. Moreover, magnetic photocatalysts, such as γ-Fe 2 O 3 /C, Co/graphene materials, Fe 3 O 4 @C/Cu, and Fe 3 O 4 @CuO, for the degradation of pollutants with the advantage of easy separability were fabricated for the degradation of pollutants from the corresponding MOFs or MOF-based composites; these photocatalysts acted well in terms of both activity and reusability [129][130][131].

MOF-Derived Porous Metal Sulfides
Various metal sulfides with tunable structures and morphologies have been fabricated via the sulfurization process or thermal treatment of MOFs. For instance, Li and coworkers developed a series of M x S y @C (M = Co, Zn, Cd, Ni, and Cu) through one-pot sulfurization of five MOFs with the identical formula of [M(pa)(bib)] ∞ [132]. Among them, pure ZnS and CdS were subsequently obtained by the combustion of ZnS@C and CdS@C (to remove carbon) and utilized as moderately active photocatalysts for degradation of MB under visible-light illumination. Batabyal and coworkers developed a one-pot method, i.e., thermolysis of single-or dual-source coordination polymer precursors in the presence of different surfactants, to synthesize nanocrystals of In 2 S 3 , ZnIn 2 S 4 , and CdIn 2 S 4 [133]. The synthesized In 2 S 3 nanocrystals showed efficient photocatalytic degradation of MB under UV light irradiation, and the photocatalytic activities of the ternary chalcogenides of In 2 S 3 were considerably enhanced.
In addition to MOF-derived metal oxides, metal sulfides, and their composites, some other kinds of MOF-derived materials have been reported as highly efficient photocatalysts for the degradation of organic pollutants, such as MOFderived metal phosphide composites, MOF-derived carbonmodified g-C 3 N 4 , semi-transformed MOFs hybrids, and so on [134][135][136][137]. Although there are numerous papers focused on the photodegradation of organic pollutants with porous nanostructures derived from MOFs as photocatalysts, the gap between laboratory research and industrial application is still huge due to the following issues: (1) most of the MOF-derived photocatalysts show limited absorption of visible light and (2) the degradation mechanism is not thoroughly explored, and the degradation products are not clear. Therefore, more in-depth research on MOF-derived photocatalysts for the degradation of organic pollutants should be carried out.

Summary and Outlook
In this review, we have discussed the improvements in MOFderived porous structures serving as photocatalysts for various energy-/environment-related reactions, including water splitting, pollutant degradation, CO 2 reduction, and organic reactions, achieved in the last few years. Owing to the beneficial structural features, such as versatility in components and well-defined pore structures, MOFs have been considered as the ideal precursors of porous semiconductor photocatalysts, including porous metal oxides and porous metal sulfides, and their heterostructures. The derived porous photocatalysts exhibit enhanced performances toward various reactions due to their high accessible surface areas and rich pore structures. Moreover, porous heterostructures or solid solution photocatalysts can be rationally designed from multimetallic MOFs as the precursors. With tuned band gap, increased active sites, and increased efficiency of photogenerated charge carrier separation and migration, the photocatalytic activities of MOF-derived porous heterostructures and solid solutions have been further enhanced.
Despite the great progress achieved in the field of MOFderived photocatalyst development, there are some pending issues that need to be solved. Transformation processes always involve high-temperature reactions and lack precise control over the pore structure and active sites, which, however, will have great influence on the photocatalytic performances. Therefore, more efforts should be devoted to the controlled synthesis of MOF-derived materials with desired pore structures as well as the active sites, which also call forth in situ techniques to track the transformation process of MOFs. Moreover, MOF-derived semiconductor photocatalysts suffer from low solar-to-chemical-energy conversion efficiency because only a few photocatalytic systems can mainly utilize the energy of the UV region, which is only 5% of the solar spectrum. The development of visible or near-infrared (NIR) light-driven photocatalytic systems with fast kinetics is still in active demand. Fabrication of composites with heterojunctions has shown its superiority in promoting charge separation, yet the intrinsic mechanism is far from well understood. An in-depth investigation should be carried out; the combination of experimental investigation and the corresponding theoretical calculation seems to be an effective method.
Finally, the other issues and their possible recommendations are listed below toward the specific reactions. For photocatalytic water splitting, the currently developed photocatalysts derived from MOFs are aimed at only one isolated half-reaction, and sacrificial reagents are added to replace the other half-reaction. Yet the goal to produce hydrogen at a cost and scale that are comparable with fossil fuels should be realized by rational design of highly efficient water-splitting photocatalysts that can split pure water into stoichiometric amounts of H 2 and O 2 , driven by sunlight without using any sacrificial reagents. Thus, MOFs-derived photocatalysts for overall water splitting are highly desired. For photocatalytic CO 2 reduction, the study is in its infancy, and the lower efficiency and selectivity of MOF-derived photocatalysts make it difficult to meet the demands of industrial applications, and therefore, further efforts are needed to make a breakthrough in the fabrication of photocatalysts. From the viewpoints of design principles, the increased adsorption of CO 2 can improve the reduction efficiency of CO 2 . The exposure of special surface sites that can decrease the energy barrier of the reduction process or the enrichment of CO 2 molecules around the active sites can help to improve the photocatalytic performance. For photocatalytic organic redox reaction, hole scavengers or electron scavengers are usually used, and designing appropriate tandem reactions can help to avoid their use. With regard to photocatalytic pollutant degradation, research is mainly focused on the degradation of the dye pollutants and some organic pollutants by photogenerated reactive transient species of O 2 . Owing to the non-selective property of the reactive species, the conclusion about the activity of photocatalysts is not accurate. In-depth research on MOF-derived photocatalysts with thoroughly explored degradation mechanism and clear degradation products should be carried out.
In this review, we give a comprehensive survey of MOFderived porous semiconductor structures as very promising photocatalysts toward various reactions. Given the beneficial features of porous structure and well-defined heterostructures, we believe that MOF-derived porous semiconductor structures will present a bright future toward photocatalysis.