Metal-Oxo Cluster Catalysts for Photocatalytic Water Splitting and Carbon Dioxide Reduction

Photocatalytic water splitting and carbon dioxide photoreduction are considered effective strategies for alleviating the energy crisis and environmental pollution. Polynuclear metal-oxo clusters possess excellent electron storage/release ability and unique catalytic properties via intermetallic synergy, which enables them with great potential in environmentally friendly photosynthesis. Importantly, metal-oxo clusters with precise structure can not only act as high-efficiency catalysts but also provide well-defined structural models for exploring structure–activity relationships. In this review, we systematically summarize recent progress in the catalytic application of polynuclear metal-oxo clusters, including polyoxometalate clusters, low-cost transition metal clusters, and metal-oxo-cluster-based metal–organic frameworks for water splitting and CO2 reduction. Furthermore, we discuss the challenges and solutions to the problems of polynuclear metal-oxo clusters in photocatalysis.


Introduction
The unrestrained consumption of fossil fuels has dramatically increased the concentration of carbon dioxide (CO 2 ) in the atmosphere and exacerbated the worldwide energy crisis and greenhouse effect [1][2][3]. Photocatalytic water splitting and CO 2 reduction are considered effective strategies for alleviating the energy crisis and excessive carbon dioxide emissions [4]. Reducing the reaction activation energy and improving the reaction rate is crucial for water splitting into hydrogen and carbon dioxide photoreduction. Photocatalytic overall water splitting driven by solar energy can simultaneously generate H 2 and O 2 , and H 2 is an ideal energy carrier to replace traditional fossil fuels because of its clean combustion characteristics and high energy density [5][6][7]. Photoreduction of CO 2 into value-added carbon-based products, including hydrocarbon fuels (e.g., CH 4 , C 2 H 4 , and C 2 H 6 ) or chemicals (e.g., HCOOH, CH 3 OH, and CH 3 COOH), can push the end products of fossil fuels back into the carbon cycle [8][9][10][11]. To date, a variety of inorganic semiconductors have been studied for efficient energy conversion [12], but their indistinct active sites and imprecise microenvironments with complicated structures have impeded the efforts to reveal mechanistic insights into the energy conversion process.
Molecular clusters composed of multimetal centers with adjustable composition are widely used as photocatalysts with excellent catalytic ability to further provide mechanistic insight at the molecular level because of their well-defined structure [13][14][15]. Among these cluster catalysts, polyoxometalates (POMs) are a large class of metal-oxo clusters composed of the earth-abundant elements V, Nb, Ta, Mo, and W, which represent a tremendous range of crystalline clusters with abundant physical and chemical properties [16][17][18]. Recently, POMs have been widely used in photocatalytic water splitting to produce hydrogen and oxygen, as well as carbon dioxide photoreduction. Moreover, various rare earth metal clusters have been widely investigated because of their good performance and extensively tunable catalytic properties in photocatalysis [19]. Recently, low-cost transition metals with unfilled valence 3d orbitals have been constructed into transition metal clusters with tens or hundreds of atoms, which may achieve excellent performance for photocatalysis via an interatomic coupling or synergistic excitation [20,21]. Recently, transition metal clusters have made great advances in photocatalysis. In this field, polynuclear metal-oxo clusters possess excellent electron storage/release ability and unique catalytic properties via intermetallic synergy, which enables them with great potential in environmentally friendly photosynthesis. Importantly, metal-oxo clusters with precise structure can not only act as high-efficiency catalysts but also provide well-defined structural models for exploring structure-activity relationships. This review summarizes the important application progress of cluster-based catalysts in energy conversion, such as water splitting and carbon dioxide reduction. Further, the challenges and solutions to the problems of polynuclear metaloxo clusters in catalysis are also discussed.

Photocatalytic Water Splitting for Hydrogen Production
Under the double pressure of the energy crisis and environmental pollution, sustainable clean energy must be developed to reduce fossil fuel use [22]. For a long time, researchers searched for effective, low-cost, and sustainable ways to promote water splitting into hydrogen [23]. Since 1972, when photocatalytic water splitting on TiO 2 electrodes was achieved [5], photocatalytic H 2 evolution has been considered a promising approach to overcoming the energy crisis and environmental issues. However, the current photocatalysts for water splitting have disadvantages, such as a low H 2 yield and the need for a precious metal as a cocatalyst or photosensitizer [24]. Therefore, developing a cheap and efficient photocatalyst for water splitting is desirable but remains very challenging.
Inspired by the photocatalytic H 2 production on TiO 2 in 1972, great progress has been achieved in recent decades by investigating metal oxide semiconductors (MOSs, such as TiO 2 , titanates, Ta 2 O 5 , and tantalates) as photocatalysts [25][26][27]. POMs, as an important subclass of metal-oxo clusters, possess inherent redox ability and semiconductor-like characteristics (Scheme 1).

POM Used as a Photocatalyst Directly
In 2011, Feng's group [28] 12 }based POMs can be further rationalized by the presence of distorted heptacoordinated Ta atoms as a TaO 7 pentagonal bipyramid. Furthermore, a series of high-nuclear spin nickel cluster-containing POMs were designed and synthesized with the lacunary POM as the coordination ligands. Introducing transition metal clusters improves their catalytic performance for hydrogen production under visible light irradiation with the assistance of noble metal photosensitizers [31,32]. Recently, Lv's group [33,34] introduced transition metals such as manganese and iron into the lacunary Keggin polyoxometalate to achieve photocatalytic H 2 evolution.

POM@photosensitizers Assembled and Used as Photocatalysts
However, traditional POMs often display light absorption in the ultraviolet region because of their large bandgap energy, which severely limits the improvement of the photocatalytic efficiency of the hydrogen evolution reaction (HER) [35][36][37]. Therefore, researchers have developed several strategies to broaden the spectral response range of catalysts. First, Lin were incorporated into a mesoporous Zr-based metal-organic framework (NU-1000), respectively. Under the optimized conditions, the resulting POM@MOF composites can effectively catalyze hydrogen production with superior long-term stability and reusability, and a hydrogen evolution rate of 3482 and 13,051 μmol/(g h) can be achieved for Ni 3 PW 10 @NU-1000 and Ni 3 P 2 W 16 @NU-1000, respectively.

Covalent Photosensitizer-POM Attached and Used as a Photocatalyst
Introducing a photosensitizer into the POM frameworks was considered an efficient method for obtaining visible light-active photocatalysts for water splitting. Zhang et al. [41] synthesized two W/Ta mixed-addendum nanoclusters grafted with TM-organic fragments, resulting in two UV, visible, and NIR light-active molecular clusters for water splitting. Carsten's group [42] reported a new route to link an iridium photosensitizer to an Anderson-type hydrogen-evolution catalyst. This covalent dyad catalyzes the visible-light-driven HER and shows superior HER activity compared with the noncovalent reference. Subsequently, they further used a fully integrated photochemical molecular dyad consisting of a Ru photosensitizer covalently attached to a Wells-Dawson POM, which can act as an electron storage and hydrogen evolving catalyst. The molecular photosensitizer-POM dyad shows excellent chemical and photochemical stability [43].

POM-Semiconductor Hybrids Used as a Photocatalyst
Compared with traditional organometallic or pure organic molecular photosensitizers, semiconductor nanocrystals materials have received extensive attention in recent years for solar energy harvesting [44]. More importantly, the semiconductor nanocrystal-based photosensitizers generally exhibit ultrahigh photostability and broad visible light absorption properties. In 2013, Zhang's group [45] successfully synthesized the molecular self-assembled CdS QD-POM-Au nanohybrids. The tricomponent nanohybrids exhibited photocatalytic activity for hydrogen evolution through water splitting via visible light irradiation. Recently, photocatalytic hydrogen evolution in pure water was realized by Ding's group [46]. CdS was used as the light-harvesting group, and CQDs were used as the electron acceptor and donor with [Ni 4 (H 2 O) 2 (PW 9 O 34 ) 2 ] 10− as the catalyst. Under visible light irradiation (λ = 420 nm), this catalytic system exhibits good water splitting activity, with a H 2 evolution rate of up to 145 μmol/(g cat ·h). Lv's group [47] coupled water-soluble CdSe QD lightabsorbers with Ni-substituted polyoxometalate (Ni-POM) catalysts and an AA electron donor. The present CdSe + POM catalytic system exhibits superior and robust hydrogen production activity. Fig. 1 Synthesis of the POM@photosensitizer system via chargeassisted self-assembly and photocatalytic water splitting for hydrogen production. Reproduced with permission from Ref. [24]. Copyright © 2015, American Chemical Society

Photocatalytic Water Oxidation
Water oxidation for oxygen production represents a multielectron transfer process with high oxidation potential, which is considered the bottleneck in overall water splitting [48]. As a result, considerable effort has been devoted to developing water oxidation catalysts (WOCs) in recent decades [49,50]. Among these heterogeneous and homogeneous catalysts, POMs represent a great subclass of WOCs that not only possess a well-defined structure but also can efficiently drive water oxidation.
I n 2 0 1 0 , H i l l e t a l . [ CaO 4 ] core of the oxygen-evolving center in photosystem II (PSII). These four compounds were the first POM-based cobalt-phosphate-cluster molecular catalysts for visible light-driven water oxidation; thus, they can serve as a functional model of the oxygen-evolving catalysts. A systematic study first revealed that heteroatom regulation can realize the regulation of photocatalytic performance for the water oxidation of these POM clusters. Subsequently, three high-nuclear spin nickel clusters, {Ni 12 [57], which opened the way to using non-expensive vanadium clusters for water oxidation in artificial photosynthesis. Then, Hill's group [58] reported a homogeneous carbon-free cobalt-based water oxidation catalyst based on redox-active V-centered POM ligands, [Co 4 (H 2 O) 2 (VW 9 O 34 ) 2 ] 10− . In 2018, Dolbecq and coworkers [59] immobilized the sandwich-type polyoxoanion [(PW 9 O 34 ) 2 Co 4 (H 2 O) 2 ] 10− in the hexagonal channels of a Zr IV -porphyrinic MOF-545 hybrid framework (Fig. 2). The composite material exhibits high photocatalytic activity and good stability for visible light-driven water oxidation.

Photocatalytic CO 2 Reduction
Global energy demands largely depend on the combustion of fossil fuels, including coal, petroleum, and natural gas [60]. As the main component of greenhouse gas, CO 2 is a key product during fossil fuel combustion. The immense emission of CO 2 has resulted in severe environmental issues, such as global warming and extreme weather [61]. As is well known, fossil fuels will remain a major energy source in the foreseeable future, and CO 2 conversion into fuels represents a straightforward strategy for solving the energy crisis and environmental problems [62]. Therefore, exploring cluster catalysts for CO 2 photoreduction has attracted wide attention, and great progress has been made in this field.
Su et al. [ 20 ] as a titanium source to anchor ultrafine TiO 2 -x nanoparticles on ultrathin carbon layers (C-TiO 2-x ), which were loaded on the g-C 3 N 4 matrix to construct the composite catalyst C-TiO 2-x @g-C 3 N 4 . The reported C-TiO 2-x @g-C 3 N 4 photocatalyst can efficiently reduce CO 2 to CO coupled with water oxidation via a two-electron/two-step pathway, and the CO yield can reach up to 12.30 mmol/g within 60 h of visible light irradiation.
Dolbecq et al. [69] constructed a hybrid POM@ MOF system by coimmobilizing a Keggin-type POM [PW 12 O 40 ] 3− and catalytically active unit Cp*Rh(bpydc) Cl 2 (bpydc = 2, 2′-bipyridine-5,5′-dicarboxylic acid) into a Zr(IV)-based metal-organic framework UiO-67 (Fig. 4). DFT calculations identified two possible locations of the POM in the octahedral cavities of the MOF, with the Cp*Rh complex pointing toward an empty pore. Photocatalytic experiments revealed that the (PW 12 , Cp*Rh)@UiO-67 composite can efficiently catalyze CO 2 reduction into formate and hydrogen. The formate production was much enhanced compared to that of the POM-free Cp*Rh@UiO-67 catalyst, which demonstrates the obvious influence of the POM on CO 2 photoreduction. Lan's group [70] reported two stable POM-grafted metalloporphyrin coordination frameworks

Low-Cost Transition Metal Clusters for Photocatalysis
Recently, increasing attention has been paid to high-nuclear spin metal clusters because of their structural aesthetics [71][72][73], intriguing properties, and potential technological applications. These metal-oxo clusters are widely used in photocatalysis for water splitting and CO 2 reduction because of their unusual chemical activity and catalytic features [74]. Artificial photosynthesis is an effective method for achieving sustainable development by directly converting solar energy into storable chemical fuels [75]. Thus far, the directed development of efficient WOCs is still a great challenge in synthetic and analytic chemistry [76]. The cubane-like {CaMn 4 O 5 } oxygen evolution active center in PSII has provided an excellent model for designing WOCs, but its structure and mechanism are still under investigation [77,78]. Cobalt-based clusters are considered economical and robust WOCs and have attracted wide attention in the past decades. In 2015, Patzke's group [79] II 3 Ln(OR) 4 } cubanes facilitated the design and synthesis of a biologically active WOC by combining Ln 3+ centers as a substitute for the redox-inactive Ca 2+ to combine the flexible aqua-/ acetate ligands (Fig. 6a) [80]. Moreover, they also presented a series of mixed Co/Ni-cubanes [Co II x Ni 4-x (dpy{OH}O) 4 (Oac) 2 (H 2 O) 2 ](ClO 4 ) 2 {Co x Ni 4-x O 4 -dpk} as the WOCs (Fig. 6b) [81]. Recently, Kong and coauthors [82] demonstrated a series of bio-inspired heterometallic cubane clusters LnCo 3 (Ln = Nd, Eu, and Ce), which can be considered synthetic analogs of the CaMn 4 O 5 cluster. They creatively anchored the LnCo 3 on phosphorus-doped graphitic carbon nitrides (PCNs). The combination of LnCo 3 clusters and PCN achieves efficient separation of photogenerated carriers and enables rapid production of H 2 and O 2 . Consequently, the resulting LnCo 3 @PCN composite catalysts show efficient overall water splitting without any sacrificial reagents.
In 2016, Lu et al. [83] reported a binuclear cobalt cluster that can be used as an efficient homogeneous photocatalyst for visible light-driven CO 2 reduction with the assistance of a [Ru(phen) 3 ] 2+ photosensitizer. The Co-Co synergistic catalytic effect in this binuclear cluster played an important role in improving its catalytic efficiency, which is supported by experimental and DFT results. Subsequently, they succeeded in obtaining a dual-core CoZn catalyst by replacing Co II with Zn II , which significantly improved the activity of photocatalytic reduction of CO 2 under lower light intensity [84]. The auxiliary Zn II site shows a strong binding affinity to OH − , which greatly promotes cleavage of O=C-OH to form C-OH to significantly improve the performance of CO 2 photoreduction. Further, Zhang et al. [85] explored a strong sensitizing photosensitizer of [Ru(Phen) 2 (3-pyrenyl-Phen)] 2+ to replace a [Ru(phen) 3 ] 2+ photosensitizer. These strong sensitizing photosensitizers can efficiently sensitize the binuclear cobalt cluster for efficient CO 2 -to-CO conversion to achieve a turnover number of 66,480.
Since the report of TiO 2 for water photolysis, semiconductor materials have attracted extensive attention in this field [5]. In recent years, loading metal nanoparticles on semiconductors has been found to efficiently modulate their electronic structure, thus improving photocatalytic performance [86]. However, the photocatalytic mechanism with multidispersed nanoparticles is difficult to understand at the atomic level. Metal-oxo clusters with clear crystal structures are ideal models for understanding the structure-activity relationship. Kong et al. [87] loaded the 3d-4f metal cluster of Ln 52 Ni 56 (Ln = Eu, Gd, Pr, Nd) on the surface of the CdS semiconductor to effectively improve the separation efficiency of photogenerated electrons and holes, thus improving the performance of photocatalytic water splitting. Some of the Ni 2+ in the Ln 52 Ni 56 cluster could be replaced by Cd 2+ during the cluster loading to form the Eu 52 Ni 56-x Cd x /CdS composite catalyst. Among these catalysts, multichannel electron transfer can confer higher photocatalytic performance to Eu 52 Ni 56 than other rare earth homologs, reaching a H 2 yield of 33,533 μmol/(h g).

Metal-Oxo Clusters in MOFs for Photocatalysis
Metal-organic framework materials are constructed from various metal cations and organic ligands, and they have rapidly grown into a well-known class of crystalline materials with porous structures [88]. These clusters can be used as connecting nodes to construct MOFs, which can not only exert the activity of clusters but also coordinate with photoactive ligands to improve their photocatalytic efficiency [89]. Kong et al. [90] demonstrated a stable metal-organic framework featuring dinuclear Eu(III) 2 clusters as connecting nodes and Ru(phen) 3 -derived ligands as linkers, which can be used for visible light-driven CO 2 reduction (Fig. 7a). Photoexcitation of the metalloligands initiates electron injection into the dinuclear {Eu(II)} 2 active sites to selectively reduce CO 2 to formate via a two-electron process. The electron transfer from Ru metalloligands to Eu(III) 2 catalytic centers is studied via transient absorption and theoretical calculations to reveal the photocatalytic mechanism.
In 2020, Lan et al. [91] synthesized a series of stable heterometallic Fe 2 M cluster-based MOFs, which were used as efficient photocatalysts to achieve artificial photosynthesis for coupling CO 2 reduction with H 2 O oxidation in the absence of an additional sacrificial agent. During visible light excitation, low-valence metal centers can accept electrons to reduce CO 2 , and the high-valence Fe center can be used for water oxidation (Fig. 7b).
A new strategy was explored by preincorporating metal precursors in the cavity of MOFs followed by in situ reduction to load nanoclusters into the MOF matrix [92,93]. In 2020, Lin's group [94] used low-intensity light to generate Cu I species in the cavities of a MOF to in situ activate a Cu II (H x PO 4 ) y @Ru-UiO catalyst for selective CO 2 hydrogenation to EtOH. Recently, Zhang et al. [95] synthesized a series of composite catalysts by integrating a Ru(bpy) 3 photosensitizer and single-metal catalysts of [bpy-CuCl 2 ] into a Eu-MOF platform. The copper catalytic active center and the assembly of the Ru(bpy) 3 photosensitizer play an important role in significantly improving their photocatalytic performance. The formate yield is up to 3040 μmol/g in 10 h with a selectivity of 99.7%. Systematic studies first revealed that the catalytic process is controlled by the Cu/X synergy to in-situ generate H-bonding between X and the CO 2 reduction intermediate in the photocatalytic process (Fig. 8). The selectivity control of HCOO − versus CO can be simply conducted by changing the coordination ligand to the Cu center. Subsequently, cobalt single-site and ultrafine CuPd nanocluster catalysts were integrated into a porphyrin-based MOF to construct the composite photocatalyst (Cu 1 Pd 2 ) z @ PCN-222(Co) (z = 1.3, 2.0, and 3.0 nm) [96]. Upon visible light irradiation, excited porphyrin can concurrently transfer electrons to Co single sites and CuPd nanoclusters, providing the possibility for coupling CO 2 photoreduction and Suzuki/Sonogashira reactions. Systematic investigations demonstrate that the close interaction of the CO 2 reduction center and carbonylation Suzuki coupling catalyst can promote the direct transfer of CO and CO* between these two catalytic active centers. The collaboration among different components in these composite catalysts highlights a new insight into developing a sustainable protocol for carbonylation reactions using the greenhouse gas CO 2 as a C 1 source (Fig. 9).

Conclusion
Currently, concern is growing that sustainable and clean energy should become an essential energy source for human life. In recent years, researchers have successfully Fig. 7 a Schematic of the light-induced dynamics in Eu-Ru(phen) 3 -MOF. Reproduced with permission from Ref. [90]. Copyright © 2018, Nature Publishing Group, b function of the heterometallic clusterbased organic framework for photocatalysis. Reproduced with permission from Ref. [91]. Copyright © 2020, Wiley-VCH investigated on various cluster photocatalysts for solar energy conversion. POMs and other metal-oxo clusters have exhibited excellent catalytic performance. These multimetal clusters have shown great potential in the field of photocatalysis because of their adjustable components, diverse structures, and multimetal synergistic catalysis with excellent catalytic properties. Some of them have been used as building units to assemble photosensitizers with excellent catalytic performance, which represents a popular research field. These cluster-based photocatalysts with well-defined structures have supplied typical models to show how to construct effective catalysts. However, the current metal clusters are mainly based on homogeneous catalysis with insufficient reusability and require additional photosensitizers and/or sacrificial agents to maintain the catalytic reaction cycle. The large crystal size of POM-based compounds needs to be ground to increase the specific surface area, and the limited light absorption ability and limited number of exposed active sites still need to be greatly improved. Although obstacles still need to be overcome, increasingly more studies have shown the advantages of POMs in the field of photocatalysis, and more effort is needed to explore efficient and robust cluster photocatalysts.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in Fig. 8 a Integration of a Ru(bpy) 3 photosensitizer and a single-metal catalyst of [bpy-CuCl 2 ] into a Eu-MOF platform for selectivity control of CO 2 photoreduction to HCOOversus CO, b comparison of production yield under different coordination conditions. Reproduced with permission from Ref. [95]. Copyright © 2021, American Chemical Society Fig. 9 (Cu 1 Pd 2 ) 1.3 @PCN-222(Co) photocatalyst for the solar-driven carbonylation Suzuki coupling reaction under CO 2 with multicomponent synergy. Reproduced with permission from Ref. [96]. Copyright © 2021, American Chemical Society the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.