Compared with traditional porous materials (such as zeolites, clays or mesoporous silica), MOF-based materials hold novel properties and particular advantages, and can be rationally designed for the use in hydrogenation reactions. 1) MOFs possess multiple and tunable active species (nodes, linkers, and pores). Moreover, various catalytic sites can be encapsulated into MOFs to develop MOFs-based catalysts for diverse hydrogenation reactions. 2) The tunable pore structure of MOF can efficiently change the diffusion of reactants to the active sites, and then tune the activity and selectivity of the MOFs-based catalysts.[2] 3) The well-defined structure of MOFs provides great chance to design active sites at an atomic level and to understand the reaction mechanism at the molecular level.[70] 4) The high specific surface area of MOFs can increase the concentration of active sites and enhance the adsorption of substrates.

Photocatalytic hydrogenation of organics is considered to be an environment-friendly process for the production of industrial chemicals as it can directly convert solar energy into chemical energy.[71] The unique properties of MOFs make them the ideal alternatives for heterogeneous photohydrogenation.[5, 72]Scheme 2 illustrates the photoexcitation and the charge transfer process in MOFs. Regarding the above analysis on the mechanism of photocatalysis, the main causes of low efficiency for photohydrogenation are limited visible light adsorption, fast photogenerated electron–hole recombination and high reaction barriers.[25, 73] Many efforts have been devoted to enhance the efficiency for photocatalytic hydrogenation over MOFs-based catalysts. For instance, through doping metal nodes, functionalizing ligands, loading metal NPs and constructing heterogeneous structures, the light adsorption capacity and charge separation efficiency can be effectively improved.[22, 25] In the following sections, we will discuss in detail the design of MOFs-based catalysts to improve the efficiency of photohydrogenation and the recent advances on photohydrogenation of organic chemicals and CO2 catalyzed by MOFs-based composites.

4.1 Photohydrogenation of Organic Chemicals

MOFs hold great potential for exploring new prospective catalysts to realize high performance toward photohydrogenation of unsaturated organic substrate. Under light driven, MOFs-based photocatalysts can result in charge separation to induce direct reduction reaction by photoexcited electrons, which provides a green and new synthesis route for fine products.[14] Moreover, the interconnected channels facilitate the accessibility of high-density active sites and the transport of substrate/product. Then, the interaction between reactant and photogenerated electrons can be strengthened to increase catalytic activity. Up to date, several MOFs-based photocatalytic systems have been known to be capable of photoconverting organic chemicals.[74]Table 2 summarizes the representatively MOFs-based photocatalysts developed by different strategies for the photocatalytic hydrogenation of organic compounds.

Table 2. MOF-based photocatalysts developed by different strategies for the photocatalytic hydrogenation of organic chemicals Strategies Catalysts Light source Hydrogen source Substrate Active species Ref. Modified metal nodes or linkers Hf/ZrUiO-66 UV–vis iPrOHa Nitroaromatic e− [75] Ni-porphyrin MOF Visible light NaBH4 Nitroaromatics BH4− [76] Zn-MOF Visible light N2H4 Nitroaromatics e− [15] Loading of plasmonic metal NPs Pd@ZIF-8 Visible light H2 Olefin Hadb [77] Au@Pd@ZIF-8 Visible light H2 Alkynes Had b [78] Au/PtAu@UiO-66-NH2 Visible light H2 Cinnamaldehyde Hadb [78] Deposition of noble metals Pt or Au/ NH2-MIL-101(Fe) Visible light HCOOH Aromatic aldehyde H• [79] Pt@UiO-66 Full spectrum H2O Nitrobenzene Hadb [80] Pt/Ti-MOF-NH2 Visible light H2O Nitrobenzene H• [81] Construct heterojunction Ag/MIL-125(Ti)/g-C3N4 Visible light iPrOHa Nitrocompounds e− [82] CdS@MIL-68(Fe) Visible light H2O 4-Nitroaniline e−; ∙CO2− [49] Ce-doped UiO-66/graphene Visible light iPrOHa Nitroaromatic compounds e−; H• [63] 4.1.1 Single MOFs Catalysts

MOFs with active sites locating on metal node or organic linker can act as the photocatalyst for some reduction reactions. Nevertheless, most MOFs possess limited light harvest ability and suffer from high photoinduced charge carriers transfer energy. Accordingly, it is necessary to tailor the band structure of MOFs by the design of organic linker with suitable length, geometry, and functional group. Besides, the metal identity of inorganic node also affects MOFs property.

For the reduction of nitrobenzene (NB), the law of reaction thermodynamics should be followed: ECB >EC6H5NO2/C6H5NH2 + ER (ER is the overpotential for nitrobenzene reduction). The more negative potential of CB can provide a stronger driving force for the reduction process. UiO-66 (Zr6O4(OH)4(BDC)12), as a typical MOF, possesses tunable structure, excellent thermal stability and low energy level of CB (−0.6 V, vs NHE at pH 7.0), which is more negative than the reduction potential of NB (−0.486 V vs NHE). Elkin[75] compared a series of MOFs in UiO-66 family with diverse linkers and metals for photocatalytic reduction of nitro-aromatics, and disclosed the roles of both the metal and the linker on photocatalytic process. The Hf-based MOF outperformed the Zr analogues. Moreover, the impacts of linker identities on increasing the activity of MOF are in the following order: pyridine (py) > 2-aminoterephthalic acid (BDC-NH2) > BDC. The enhanced performance of Hf-py was ascribed to the altered LMCT energy and ability to conduct protons, which is critical to proton-coupled electron transfer redox reactions. Other groups also reported related research results with respect to the MOF photocatalyst. Jiang et al. showed that NH2 can effectively suppress the recombination of photogenerated charge carriers.[24] Li et al. demonstrated that the doping of Ce or Ti into the skeleton of Zr-based MOFs dictated light adsorption, band structure and LMCT efficiency.[23]

4.1.2 Coupling MOFs with Carbon Materials

Combing MOFs with carbon materials is another strategy to improve the catalytic performance toward hydrogenation reaction. Previously, multifunctional carbon materials including graphene (GR) and carbon nanotubes (CNT) have been employed as building blocks for photocatalyst carrier/promoter due to their superior conductivity, excellent mobility of charge carriers and abundant active sites. Inspired by these, Yang and co-workers prepared GR/Ce-UiO-66 core–shell hybrids with adequate interfacial contact by a solvent thermal method[63] The order of photoreduction rate of nitrobenzene (NB) was as follows: GR/Ce-UiO66(10) > GR/Ce-UiO66(5) > GR/Ce-UiO66(15) > Ce-UiO66 (the number in bracket refers to the GR amount in the composite). The introduction of appropriate amount of GR can remarkably improve the transfer of photoinduced electron from Ce-UiO66 and increase the separation efficiency of photogenerated carriers. In especial, GR provides sufficient adsorption sites for NB molecules, which is beneficial to the receiving of photogenerated electrons by NB in the subsequent process to initiate reduction reaction (Figure 6).

a) UV–vis diffuse reflectance spectra and b) photoluminescence spectra of the samples. c) The rate constant (k) of different catalysts for photocatalytic nitrobenzene reduction. d) Possible mechanism for the photocatalytic reduction of nitrobenzene over GR/Ce-UiO. Reproduced with permission.[63] Copyright 2017, Elsevier. 4.1.3 Introducing Plasmonic Metal into MOFs

Plasmonic metal nanostructures can be used in photocatalysis due to their strong interactions with electromagnetic radiation through an excitation of localized surface plasmon resonance (LSPR).[83] MOF is an ideal candidate for the stabilization of plasmonic metal nanoparticles (MNPs) due to its high porosity and tunable pore size. Incorporating plasmonic NPs into MOFs is a promising way to enhance the catalytic performance of hydrogenation. In the charge-carrier-driven reaction, external light is used to excite charge carriers on the metal surface. Few electrons with higher energy can be transferred to the lowest nonoccupied orbit of the adsorbed species, resulting in the activation of chemical bond and chemical transformation. In particular, electron-driven reaction can potentially target certain reaction pathway, which is unselective in purely thermal reaction. In addition, the photothermal effect converting light into heat has been generated on plasmonic metal with local heating of the lattice, and the reactivity of metal can be promoted by tuning light intensity.[61] So far, many researchers reported the successful applications of plasmonic NPs-MOFs based materials in the field of hydrogenation.[84] Two main strategies to improve catalytic performance of MOFs-based materials by introducing plasmonic NPs are as follows. 1) Immobilizing plasmonic NPs into MOF, tuning their size, geometry, and location for regulating catalytic performance; 2) integrating a second metal together with the plasmonic NPs into MOFs to form an efficient catalyst due to the synergistic effect.

A plasmonic Pd nanocubes (NCs) @ZIF-8 composite has been rationally fabricated for selective hydrogenation at room temperature under 1 atm H2 and light irradiation.[77] The Pd NCs, acting as active sites, have well-defined structure and maintain high dispersion with the size of 17 ± 3 nm in ZIF-8. This composite showed a plasmonic band covering the UV-to-visible spectral range, which reveals the photon adsorption for inducing high temperature to drive the hydrogenation of olefins. Meanwhile, the ZIF-8 shell offers the following advantages: 1) the pore structure of ZIF-8 is beneficial to the transportation of reactant/product and sieving different molecules with specific size for tunable selectivity. 2) It serves as hydrogen reservoir to accelerate the reaction. The reaction efficiency under 100 mW cm−2 full spectrum irradiation was much higher than that upon heating at 50 °C, exhibiting the great potential of photothermal effect in the field of catalysis. However, when light is coupled with plasmonic nanomaterial, the hot electron effect and the photothermal effect often work together in the research system, and it is difficult to study a certain type of effect separately (Figure 7). Xiong's group fabricated Au@Pd nanorods, in which the hot electrons excited on Au can directly migrate to Pd, forming electron-rich Pd surface for the hydrogenation of styrene.[78] Both experiments and density functional theory (DFT) calculations revealed the negative effect of hot electrons: the presence of hot electrons did not benefit the hydrogenation reaction due to the strong binding of hydrogen to metal sites. Furthermore, they encapsulated Au@Pd nanorods into ZIF-8 photocatalytic system for semihydrogenation of alkynes.[78] The ZIF-8 shell hinders the diffusion of heat produced in the plasmonic cores into solution, and maintains a high local temperature for reaction. Thereby, the catalytic activity under light was still 2.5 times that in dark condition even in the presence of mischievous hot electrons. On the basis of the preliminary work, Xiong's group chose Zr-based MOF UiO-66-NH2 with larger pore size and higher stability to confine Au NRs-Pt/Au cores for selective hydrogenation of cinnamaldehyde.[78] Thanks to the photothermal effect of Au NRs-Pt/Au, the reactant conversion reached nearly 100% after 4 h under light, while it was less than 60% in the dark. It is worth noting that cinnamaldehyde preferred to bind to metal sites via the parallel mode under light, which causes the improvement of C═C hydrogenation. Then, the selectivity toward C═O hydrogenation was suppressed. Both the confinement effect of porous UiO-66-NH2 and electronic structure regulated by light irradiation lead to the higher proportion of C═C hydrogenation product.

a) H2 sorption isotherms for Pd NCs@ZIF-8 and Pd NCs at 298K; b) UV/vis adsorption spectra for ZIF-8, Pd NCs and Pd NCs@ZIF-8; c) the yield of the hydrogenation of 1-hexene with 1 atm H2 over Pd NCs@ZIF-8 under full-spectrum irradiation with different light intensities at room temperature or upon heating at different temperatures; d) TEM images of Pd NCs@ZIF-8. Reproduced with permission.[77] Copyright 2016, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim. 4.1.4 Construction Schottky Junction between Noble Metal Nanoparticles and MOFs

Some nanometal-loaded MOFs are expected to afford high photocatalytic performance of hydrogenation because of the synergic effect between metal NPs and MOFs, in which high distribution of metal NPs, fast molecular transportation, enhanced light utilization and effective photogenerated charge separation are guaranteed.[85] The photogenerated electrons can be migrated from MOFs to MNPs which serve as the reservoir of electrons. Rationally loading metal NPs in semiconductor materials had been widely applied to promote the separation efficiency of the photogenerated electrons and holes. Because Schottky barriers can be formed between the semiconductor and metal NPs.[86] Moreover, in the terms of the direction of electron migration, there exists a competitive relationship between the hot electron effect and the Schottky junction. In the MNPs/MOF composite, if the energetic electrons excited from MOFs can efficiently transfer to MNPs, the catalytic activity of MNP can be enhanced significantly.[34] Inspired by this, Dong and co-workers[79] reported visible-light-induced selective transfer hydrogenation of aromatic aldehyde to alcohol catalyzed by Pd/MIL-101(Fe)-NH2 with triethylamine as electron donor and HCOOH as proton source. The catalyst of Pd/MIL-101(Fe)-NH2 was prepared by the in situ photodeposition of a Pd salt (MIL-101(Fe)-NH3)+•1/2(PdCl4)2−), possessing highly dispersed and uniform Pd NPs with a mean size around 1.8 nm. Theoretical research confirmed the dual functions of amine group: stabilizing Pd NPs and promoting the electron density of the Pd under light irradiation. Over the Pd/MIL-101(Fe)-NH2 photocatalyst, the conversion of benzaldehyde was 100% which is much higher than that in the dark (37%). It suggests that the high activity mainly results from the light driving (Figure 8). Ma et al.[80] synthesized UiO-66-NH2 with regulated structural defects, and investigated the effect of these defects on photocatalytic H2 production. The H2 production rate displayed a volcano-type trend with incremental levels of defects. Notably, as a preliminary inquiry, the tandem reaction of photocatalytic H2 production and nitrobenzene hydrogenation can be promoted without additional hydrogen source, which provides valuable inspiration of engineering MOF-based photocatalysts for organic hydrogenation transformation.

a) Photoinduced transfer hydrogenation of aromatic aldehydes by using Pd/MIL-101(Fe)-NH2; b) synthesis of Pd/MIL-101(Fe)-NH2. Reproduced with permission.[79] Copyright 2018, American Chemical Society. 4.1.5 Coupling MOFswith Semiconductor

The single-component semiconductor materials suffer from the low solar energy utilization and the easy photoelectron–hole recombination, resulting in limited photocatalytic performance. Constructing heterojunction by coupling two semiconductors with well-matched band gap structure is feasible to improve catalytic performance. MOFs is undoubtedly an appreciate alternatives to build heterojunction composite due to its advantages as follows: 1) the facile tuning of band level and light adsorption. 2) highly porous structure with the capability of loading more active sites. A variety of MOFs–semiconductor heterojunctions have been developed and applied in the field of photocatalytic hydrogenation. For instance, a heterostructured bi-semiconductor material with well-defined interface was fabricated by coupling Ag (electron-conduction bridge) and g-C3N4/MIL-125(Ti) for photoreduction nitrocompounds.[82] Both g-C3N4 and Ag NPs adsorb the visible light, the photogenerated electrons directionally migrate to Ti4+ of MIL-125(Ti) due to the close interfacial connections among MIL-125(Ti), Ag NPs and g-C3N4. Ti3+ acted as the active species for the reduction of nitrocompounds due to its strong reducing ability (−1.37V vs SHE). This hydrogenation reduction process avoided using unsafe reduction agent and provided a sustainable and green route for organic reduction. The great potential of MOFs as the photocatalyst for more organic transformation was highlighted. Notably, high selectivity toward various aromatic nitro compounds by this photocatalytic hydrogenation process was observed (Figure 9). CdS has been investigated widely in photocatalytic hydrogenation of 4-nitroaniline (4-NA) to p-phenylenediamine (4-PDA) due to its visible light response and appropriate reductive power. Nevertheless, the low separation efficiency of photogenerated electron–hole pairs and poor stability hinder the application of pure CdS. Liang et al. encapsulated CdS into MIL-68(Fe) with intimate interfacial contact using a photodeposition method.[49] MIL-68(Fe) not only served as a host to stabilize CdS, but also trapped photogenerated electrons of CdS by forming well-matched band structure. Trapping experiments and ESR studies revealed that HCO2NH4 not only served as a hole scavenger but also produced active reducing species (•CO2−). This work made full use of the redox capacity of electrons and holes, providing a highly efficient method for hydrogenation reduction and offering new insight into the design of MOF–semiconductor photocatalyst (Figure 10). Recently, the researches on MOFs@MOFs hybrids have emerged, and these materials showed high potential in the field of photocatalytic organic transformation due to their unique properties superior to their original counterparts (Figure 11). The core–shell UiO-66-NH2@MIL-101(Fe) designed by Liu et al. exhibited increased light-adsorption ability and high charge separation efficiency due to the formation of type II heterojunction with well-matched band energy.[87] Kitagawa et al. reported a MIL-101(Cr)@MIL-125-NH2(Ti) heterostructure with improved photocatalytic performance for the reduction of CrVI.[88] We believe that Z-scheme MOF photocatalytic system has considerable potential for photocatalytic hydrogenation and there is still a broad scope for the development in this research area.

a) TEM images of MIL-125/Ag/C3N4(12); b) UV–vis DRS of the samples; c) EIS Nyquist plots of the samples under dark conditions; d) the rate constant (k) of different catalysts. Reproduced with permission.[82] Copyright 2017, Elsevier. a) Photocatalytic selective reduction of 4-NA to 4-PDA over different samples; b) transient photocurrent response of pure CdS and 4-CdS-MIL68 in 0.2 m Na2SO4 aqueous solution under irradiation of visible light (λ ≥ 420 nm). Reproduced with permission.[49] Copyright 2017, Elsevier. Schematic illustration for the synthetic process of the heterostructures. (a) Reproduced with permission.[87] Copyright 2019, Royal Society of Chemistry. (b) Reproduced with permission.[88] Copyright 2017, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim. 4.2 Photoreduction of CO2 Using MOFs-Based Photocatalysts

At present, a large number of CO2 emissions caused a series of environmental problems. Therefore, effectively reducing the content of CO2 in the atmosphere, and developing environmentally friendly and renewable new energy sources has become the focus of attention. To this end, a variety of methods for CO2 conversion has been developed, such as photocatalytic reduction, electrochemical reduction and biological conversion. Among these technologies, photocatalytic reduction of CO2 is considered as a promising technology to obtain chemical or fuel like CH4, CH3OH, or HCOOH due to its clean and environmentally friendly characteristics.[89] In the semiconductor photocatalytic system, photogenerated electrons play the role of reducing active components. Upon light irradiation, the photoexcited electrons were transferred from VB to CB, and subsequently, the photogenerated electrons migrate from CB to catalytic active sites to activate the adsorbed CO2 and start reduction reaction. The prerequisite for the reaction to occur is that the CB potential of semiconductor must be negative than the redox potential of CO2 reduction for the formation of specific final product (Scheme 3). The final product distribution largely depends on the energy level of CB as well as the reduction potential of the product.[90] Due to the high reduction potential of single-electron reduction of CO2 to CO2− (−1.90 V), another feasible reduction route is to convert CO2 to carbonaceous with the aid of multiple protons and electrons. The reduction potential of CO2 into CO, methane, methanol and formic acid are −3.50, −3.79, −3.65, and −3.65 eV, respectively.[25] The activity and selectivity are also affected by other factors, including charge separation efficiency, properties of catalytic center and pore structure.

CO2 photoreduction process over MOF and redox potential of corresponding products.

Based on this background, many MOFs-based complexes have been demonstrated to be effective photocatalysts[91] for CO2 conversion, such as ZIF-8,[92] ZIF-67,[93] UiO-66,[94] MOF-74,[95] MIL-125,,[96] HKUST-1,[97] and so on. Their tailored band structure, abundant reaction active sites and superior ability to capture and activate CO2 qualify them as excellent catalysts for photoreduction of CO2.The activities of those recently developed presentative MOFs-based catalysts are compared in Table 3.

Table 3. Photoreduction CO2 activity of recently developed presentative MOFs-based catalysts Strategies MOF Light source Reaction condition Photocatalytic activity Ref. Single MOF MIL-125(Ti)-NH2 Visible light MeCN/TEOA 0.814 µmol h−1 for HCOOH [98] UiO-66-NH2 Visible light MeCN/TEOA 1.32 µmol h−1 for HCOOH [99] Ti-doped NH2-UiO-66 Solar light Photovoltaic devices 65.35 µA cm−2 [100] NH2-MIL-101(Fe) Visible light MeCN/TEOA 22.25 µmol h−1 for HCOOH [101] NNU-31 Visible light H2O 26.3 µmol g−1 h−1 for HCOOH [102] PCN-222 Visible light MeCN/TEOA 3 µmol h−1 for HCOOH [103] Zn/PMOF Visible light H2O 2.6 µmol h−1 for CH4 [104] PCN-601 Visible light H2O 6.0 µmol g−1 h−1for CO; 10.1 µmol g−1 h−1 for CH4 [105] Loading of metal NPs Au/PPF-3 Visible light MECN/C2H5OH 42.7 µmol g−1 h−1 [106] Ag/MOF-101(Cr) Visible light acetone/TEOA 427.5 µmol g−1 h−1 for CH4; 808.2 µmol g−1 h−1 for CO [107] Pt/NH2-MIL-125(Ti) Visible light MeCN/TEOA 32.4 µmol g−1 h−1 for HCOOH [108] Au@UiO-68-NHC Full-spectrum MeCN/MeOH