Heterogenization of Molecular Water Oxidation Catalysts in Electrodes for (Photo)Electrochemical Water Oxidation

Abstract

Water oxidation is still one of the most important challenges to develop efficient artificial photosynthetic devices. In recent decades, the development and study of molecular complexes for water oxidation have allowed insight into the principles governing catalytic activity and the mechanism as well as establish ligand design guidelines to improve performance. However, their durability and long-term stability compromise the performance of molecular-based artificial photosynthetic devices. In this context, heterogenization of molecular water oxidation catalysts on electrode surfaces has emerged as a promising approach for efficient long-lasting water oxidation for artificial photosynthetic devices. This review covers the state of the art of strategies for the heterogenization of molecular water oxidation catalysts onto electrodes for (photo)electrochemical water oxidation. An overview and description of the main binding strategies are provided explaining the advantages of each strategy and their scope. Moreover, selected examples are discussed together with the the differences in activity and stability between the homogeneous and the heterogenized system when reported. Finally, the common design principles for efficient (photo)electrocatalytic performance summarized.

1. Introduction

1.1. Energy Challenge, Natural and Artificial Photosynthesis

In recent decades, the need to mitigate the increasing CO₂ emissions and transition to a sustainable society has increased in importance due to the high anthropogenic CO₂ emissions that have caused 1.1 °C of global warming. Indeed, at the current rate of energy consumption, this scenario is likely to reach 1.5 °C of global warming by 2050 [1]. In this line, the sustainable synthesis of fuels and chemicals using CO₂ and water as abundant primary feedstock, and the widespread and unlimited sunlight as driving energy, provides a potential feasible pathway toward a greener chemical industry and society [2,3], despite being very challenging at the same time [4,5].

In this context, nature, upon million years of evolution, has provided photosynthetic organisms with a very complex and elegant machinery that allows them to sustain their life by solely using sunlight, CO₂ and water to generate their fuel and source or energy (carbohydrates, producing only O₂ as byproduct) via so called natural photosynthesis [6,7]. During natural photosynthesis, solar energy is harnessed into the chemical bonds of carbohydrates such as glucose. This process takes place in the thylakoid membranes of chloroplasts and is possible due to the orchestration and synchronized cooperation of photosystem II (PSII) and photosystem I (PSI) together with other membrane protein complexes and carriers present in the chloroplasts [8,9]. During natural photosynthesis, light excitation of the chlorophyll pigments present in PSII triggers the oxidation of a tetranuclear manganese cluster which is the active center of the oxygen evolving complex (OEC), which upon four electron oxidations is able to oxidize the molecule of water to generate O₂ while releasing protons and electrons. These reductive equivalents are transferred to the reductive site by cytochrome b6f to PSI (electron transport chain or Z scheme), where they are separated to generate the reduced molecule NADPH by a ferredoxin-NADP⁺ reductase (FNR), and the proton gradient triggers the APT production by the ATP synthase. Finally, these reductive equivalents and energy molecules (NADPH and ATP, respectively) are used in the Calvin cycle to reduce CO₂ to carbohydrates (CO₂ fixation) (Scheme 1) [6,7]. Alternatively, in some organisms, hydrogenase enzymes use the released electron from the oxidation of water to reduce protons to hydrogen, which can be directly used as fuel. Despite the fact that the mechanism of photosynthesis is not fully unraveled, a key feature of natural photosynthesis is that it spatially arranges and orientates reaction centers and light harvesters along the thylakoid membrane of chloroplasts, separating the reduction form the oxidation side, minimizing cross-reactivity. Therefore, mimicking and understanding how natural photosynthesis operates and transfers primary energy vectors to harness solar energy into molecules are essential to develop efficient catalytic systems to produce solar fuels and chemicals [10,11,12].

In this regard, artificial photosynthesis has emerged as a forefront technology to harness solar energy into chemical bonds using available building blocks (CO₂, H₂O and N₂) as feedstock, mimicking natural photosynthesis but at higher efficiencies than natural systems (~1%) [3,13,14]. However, there are a few challenges to address for the fabrication of efficient artificial photosynthetic devices [5,15,16,17]. First, water oxidation [5,18,19] (WO) and CO₂ reduction [5,20,21,22,23] remain bottlenecks. Therefore, advances in mechanistic understanding, comprehension of the factors that control catalytic activity and selectivity [20,24,25], and catalyst development are needed to meet the requirements for practical use: (i) low overpotentials, (ii) high activities, (iii) inexpensive, non-toxic and robust systems [20], and (iv) for CO₂ reduction, specific product selectivity, especially beyond the 2-electron reduction [21,22,26,27]. Secondly, going beyond H₂ and C₁ products (CO, formate) is particularly desirable, since, so far, even with the use of semiconductor materials that are highly efficient in light utilization, only the simplest products derived from CO₂ reduction can be produced. Lastly, charge separation is also pursued to avoid energy losses due to cross-reactivity between the oxidation and reduction sites [17,24,25].

Therefore, the study of molecular systems for their application in artificial photosynthetic devices presents an advantage that is the atom-efficient catalysis with outstanding intrinsic activities and selectivity; as well as their easy tunability and study. However, to this end, important efforts must be directed into the efficient and effective heterogenization of molecular catalysts onto devices. Additionally, long-term stability is very important for the application and stability of the immobilization technique [5,28,29,30,31]. This review focuses on the description and general overview of the main strategies used for the heterogenization of molecular water oxidation catalysts onto electrodes for (photo)electrochemical water oxidation. Moreover, selected examples for each strategy are explained and compared to their homogeneous counterparts in terms of stability or catalytic activity.

1.2. Water Oxidation Reaction

Water oxidation is the core reaction of natural photosynthesis [32,33] and, as such, it is the central reaction for solar energy conversion and storage [2,3]. However, WO is a highly endergonic process at room temperature and pressure and it still remains one of the bottlenecks for efficient solar-to-fuel conversion due to the harsh oxidizing conditions and high overpotentials required for this transformation [5,34,35,36,37,38]. Development of molecular systems in recent decades has hugely contributed to the field of WO with systems outperforming the activity of the natural system [5,34,35,38,39]. Moreover, the study of molecular water oxidation catalysts has enabled a better understanding of the mechanism of water oxidation catalysis and the extraction of the principles governing catalysis and ligand design [37,39,40,41,42,43]. In this regard, the following considerations should be taken into account for rational design of water oxidation catalysts (WOCs): (i) the multiproton/electron nature of WO, (ii) the synergy between metal centers, (iii) the stabilization effect of the first coordination sphere and the ligand environment, (iv) substrate–catalyst interaction and (v) the thermodynamic stabilization of the protein environment in the O–O bond formation event. Water oxidation is a multistep redox process and, thus, redox-active compounds would be ideal platforms to assist this process by storing oxidizing equivalents and avoiding the formation of highly reactive radical species. In this line, metal oxides and coordination complexes based on transition metals (TM) open a wide range of redox possibilities. In addition, the accumulation of oxidizing power by the TM catalyst can very much benefit from the stepwise proton-coupled electron-transfer (PCET) processes, which lower the thermodynamic energy by avoiding highly charged intermediates. In the case of TM coordination complexes, the ligand environment can play a very important role in stabilizing the different oxidation states at the metal centers. Indeed, basic ligands can stabilize high-valent intermediates due to the significant electron density injection to the metal center. In the same group of stabilizing ligands, we can find anionic or redox non-innocent ligands. Anionic donor ligands constitute good scaffolds to achieve great stabilization of the high oxidation states, but they tend to hydrolyze easily. Further, redox non-innocent ligands have the advantage of facilitating the delocalization of the positive charge of a highly oxidized metal center, to enable the accumulation of oxidizing equivalents. Nevertheless, it is very important to find the right balance between stabilization and reactivity of the active species to prevent an overstabilization that hampers the O–O bond formation.

Catalysts based on biomimetic systems of the oxygen evolving complex (OEC), coordination complexes and organometallic complexes are the most suitable systems for the mechanistic study of water oxidation. These systems have shed some light on the mechanisms for WO and developing efficient and fast catalysts with rates comparable or even higher to that of PSII [35,38,44,45,46]. One of the most challenging factors in the development and study of those systems is the identification of the exact nature of catalytic activity. The strong oxidizing conditions required for the water oxidation can promote site oxidation reactions on the organic ligand scaffolds, leading to significant structural modifications and hindering further mechanistic investigations. Alternatively, the direct use of inorganic ligands, which are called polyoxometalates, has emerged as a competent possibility in recent years [47]. Polyoxometalates are discrete metal oxides of nanometric size at halfway between molecular complexes and metal oxides [48,49]. Nevertheless, they can also be hydrolyzed under catalytic conditions. Therefore, from a technological point of view, heterogeneous metal oxides have more potential as efficient and robust catalysts for industrial applications since they are more stable under catalytic conditions. Heterogeneous have been developed with the aim to create low-cost, efficient, and more long-lasting WOCs [50,51,52]. However, it is important to note that regarding mechanistic understanding, molecular catalysts present lots of advantages such as high catalytic activities, the possibility of rational design and accessible mechanistic investigations.

Among the reported mechanistic studies carried out on the reported systems for WO, two key steps have been identified in the WO reaction: a) the oxidation of the metal center, generating metal-oxo or metal-oxyl (M=O or M-O ^·) reactive species, and b) the O–O bond formation event, which is the rate-determining step (RDS) in most cases. Analogously to the OEC in PSII, two main mechanistic scenarios have been proposed for the O–O bond formation: (i) a water nucleophilic attack (WNA) over a highly electrophilic metal-oxo species (M=O), resulting in a 2e- reduction of the metal center (acid–base mechanism, Scheme 2, left) and (ii) the interaction of two metal-oxyl radical units (M-O·), resulting in a 1e- reduction per metal center (direct coupling (DC), Scheme 2, right). The WNA mechanism is favored with electron-withdrawing ligands, which make the metal-oxo more electrophilic and therefore more susceptible to nucleophilic attack of a water molecule. Moreover, the assistance of an internal or external proton acceptor, binding the water molecule, can facilitate proper orientation to the attack. On the other hand, the DC mechanism is favored by increasing the spin density on the oxygen atoms, generating radical species and facilitating (electronic and sterically) the dimeric interaction [19,34,38,53,54].

Altogether, since rational ligand and catalysts design toward catalytic activity and durability improvement is crucial to improve catalytic activity and durability, molecular WOCs are highly desirable. The development of highly active and efficient long-lasting WOCs would allow to efficiently couple the WO with reductive transformations (such as CO₂ or water reduction) in artificial photosynthetic devices. Therefore, an alternative approach to the utilization of homogeneous or heterogeneous systems is the immobilization (heterogenization) of molecular water oxidation catalysts (WOCs) together with photoredox catalysts or not on electrodes for (photo)electrochemical WO, which has attracted the interest of the community [28,30,31,55,56].

1.3. (Photo)Electrocatalysts for WO: Homogeneous vs. Heterogeneous and Heterogenization of Molecular Complexes

Molecular catalysts are synthetically versatile and the study of their mechanisms allows rationally modifying and designing ligand scaffolds towards higher activity and stability. Indeed, as we will see in the following lines, the study of molecular-based photoanodes using electrochemistry and their application in electrocatalyzed WO has allowed understanding their mechanism in an efficient manner [57,58,59], as well as the factors that need to be considered for the development of efficient, robust and highly active systems for their application in artificial photosynthetic devices. However, many catalysts deactivate in (photo)electrocatalytic conditions most likely via bimolecular pathways [30,31,58], diminishing the applicability of the plethora of molecular electrocatalytic systems.

In this context, heterogenization of homogeneous systems by anchoring molecular catalysts on electrode surfaces is a very powerful strategy that combines the high selectivity and activity of molecular systems with the robustness and stability of heterogeneous systems (

Table 1. Comparison of properties of homogeneous, heterogeneous and heterogenized catalysts.

Property	Homogeneous	Heterogenous	Heterogenized
Activity	High	Moderate	High
Selectivity	High	Low-Medium	High
Synthesis	Complex	Simple	Complex
Tuneability	High	Low	High
Efficiency	Moderate	High	High
Stability	Short term	Long-term	Elongated
Robustness	Low	High	High
Study	Established protocols, valuable information	Less explored	Established protocols, valuable information

) [50,60,61]. Molecular catalysts, however, are far less stable than traditional heterogeneous electrocatalysts, and therefore a method to easily replace anchored molecular catalysts that have degraded could make such electrosynthetic systems more attractive. Moreover, separation of the active centers upon immobilization onto electrodes also disfavors bimolecular decomposition pathways.

2. Approaches for the Development of WO (Photo)Anodes

Heterogenization of homogeneous systems by anchoring molecular catalysts on electrode surfaces is a very powerful strategy that combines the advantages of both homogeneous and heterogeneous systems. Therefore, these systems possess the high selectivity and activity of molecular systems with the robustness and stability of heterogeneous catalysts [28,30,31,55,60]. Moreover, surface-bound catalysts for electrocatalytic WO allow for easier recovery of the catalysts and the use of a wider variety of solvents due to the suppression of solubility issues in the bulk solution.

The most widespread anchoring strategies for the immobilization of molecular WOCs onto (photo)electrodes differ on the type of interaction used for the WOC immobilization, which can be covalent or non-covalent [30,31,55]. The choice of immobilization strategy depends on the type of electrode material and the tolerance of the molecular catalysts. Common immobilization strategies are summarized below (Figure 1):

−

Covalent anchoring,

−

Supramolecular interactions,

−

Drop-casting,

−

Physical confinement into porous materials such as metal organic frameworks (MOFs) or covalent organic frameworks (COFs),

−

Electropolymerization,

−

Atomic layer deposition (ALD), and

−

Layer-by-layer deposition.

2.1. Covalent Immobilization

Immobilization of molecular WOCs by ligand modification that allows further covalent functionalization of (photo)electrodes is the most extensively studied approach [30,55]. Despite the fact that the most common strategy involves the modification of the ligand backbone of the molecular WOCs for their subsequent attachment to the surface of the photoelectrode [62], depending on the type of WOC, semiconductor, metal oxide or conductive glass of the (photo)electrode, different strategies can be employed. For the covalent functionalization of the ligand backbones with surface anchoring groups, a similar approach to the one used in the design of light-harvesting dyes for dye-sensitized solar cells (DSSC) is followed [63,64]. Among the most common functional groups that allow further surface covalent anchoring by click chemistry, the carboxylate [65,66], phosphate [67,68,69], hydroxamate, and hydroxyl [70,71] functional groups (Figure 1a) can be immobilized onto a photoelectrode by an esterification reaction with its surface. This strategy has been used for the immobilization of several WOCs [72,73,74,75,76,77] on oxide semiconductors (see Section 3.1 for further selected examples) [62,65,72,78,79].

However, a decrease in catalytic activity is a common feature when this immobilization strategy is used, which can be attributed to changes in the electronic and energetic effects in the catalyst upon the introduction of anchoring functional groups to the ligand of the WOCs that modify their redox potentials and catalytic activity and also catalyst stability [30,62,80,81,82]; or the hydrolytic degradation of ester linkages in both acidic or basic conditions [62], resulting in detachment of the molecular catalysts from (photo)electrodes [5,62,80].

Therefore, to address these issues, several alternatives have been used. On the one hand, functional groups that are used as anchors have been introduced further from the active side by using long alkyl chains as spacers in the linkers [75,83,84]. This approach causes fewer electronic changes in the active side of WOCs as well as minimizes direct interactions with the (photo)electrode surface, preserving their catalytic activity and minimizing electron-hole recombination, respectively. Additionally, more chemically and thermally stable anchoring groups or multiple anchoring groups have been employed such as pyridine-2,6-bicarboxylic acid [76], 1-(3′-amino)-propylsilatrane [85], and pyridine-N-oxide 2-carboxylic acid [86], among others [63,77], at the expense of complex synthetic procedures and charge recombination at the WOC/(photo)electrode/electrolyte interfaces [39,62]. Alternatively, phosphonate surface binding sites that are more stable in acidic conditions, or metal-oxo bridging sites, have also been used as linkers, since they are more stable than ester groups [87].

An alternative approach to avoid changing the properties of the molecular WOCs is the modification of the surface of the (photo)electrodes for subsequent covalent attachment of the WOC [55,88,89,90]. In this regard, alkoxysilanes with specific end-functional groups (such as amines or pyridine derivatives) have been widely used for surface modification of several (photo)electrodes, such as TiO₂ [64,89,91,92], ZnO [93,94,95], and SnO₂ [96], due to their simple functionalization surface chemistry [88,90]. For instance, with the 3-aminoproyltriethoxysilane (APTES) moiety, amine groups can be readily displayed on photoelectrode surfaces by hydrolytic condensation, where WOCs can be immobilized at a later stage.

2.2. Supramolecular Host–Guest Interactions

This approach uses supramolecular chemistry to assemble WOCs into guests-electrode moieties forming large functional structures by self-assembly processes, offering a toolbox to develop WO (photo)anodes. Moreover, for WOCs that operate following a dinucelar mechanism (DC) to form the O–O bond, this strategy can enhance the reaction rate by approaching the two metal-oxo units or increasing the local concentration of active species due to the host–guest interactions [97,98], as has been previously observed for ligand systems that favor Π-Π-stacking interactions between two metal complexes [44,99], by self-assembly into micelles [100], or by encapsulation into mesoporous guests [101] or macrocycles [102].

In this context, most common methods rely on Π-Π-stacking, hydrogen bonding or coulombic interactions for the assembly of WOCs onto a guest that is deposited onto an electrode or directly functionalized on its surface. Alternatively, when the (photo)electrodes are hydrophobic, WOCs bearing hydrophobic polyaromatic functional groups can also be utilized as a surface anchoring group via hydrophobic interactions directly onto the electrode [30,55,103]. However, this latter approach would have limited stability, since these hydrophobic interactions are vulnerable to aggregation or disassembly in aqueous solutions under catalytic conditions [104,105].

Another advantage of this immobilization strategy is that, since the catalyst is generally immobilized onto a guest that is anchored on the electrode, there are fewer problems of charge recombination and it can also help stabilize reactive intermediates within the guest via supramolecular interactions [30,31,55,60,97,103].

2.3. Drop Casting

This is an immobilization strategy that consists of the physisorption of a solution of a WOC onto the (photo)electrode surface followed by evaporation of the solvent without requiring the introduction of functional groups in the molecular WOC to enable its anchoring. The disadvantage of this approach is that it is difficult to control the film thickness and achieve a uniform coating, so it can reproducibility issues may arise [106]. Moreover, it normally requires the coaddition of binders, such as Nafion [107], to facilitate the anchoring of the catalyst, which is costly and can hamper its applicability [108,109,110].

2.4. Immobilization via Physical Confinement

Another immobilization strategy that is very appealing to increase the density of active sites onto electrodes is via physical confinement of WOCs into porous hosts materials such as MOFs or COFs. This approach has the advantage of avoiding complex and tedious ligand modification or functionalization processes and thus, also avoiding changes in the inherent electrochemical and energetic parameters of the catalysts. This strategy allows for simple immobilization of molecular WOCs with or without other co-catalysts or photoredox catalysts in porous materials [111]. Another advantage is the host matrices can have their own functionality that complements the WOC, such as light harvesting, charge separation, passivation and stabilizing coating layers for the whole assembled (photo)anode. [111]

For this approach, the immobilization of catalysts in the host material can be achieved following three methods: (i) by mixing a preformed host material with a presynthesized WOC and allowing the impregnation of the catalyst onto the material; by (ii) a bottom-up assembly of a host material around the preformed WOC; and (iii) using an hybrid alternative that combines both methods [112].

Despite this method being suitable for avoiding tedious synthetic procedures and allows confinement of the catalysts in the material with negligible change in its properties, one challenge of this strategy is the choice of the host material pore size. On the one hand, the pore size of the host materials should be larger than the size of molecular WOCs, although this can lead to catalyst leaching from the host material and a decrease in catalytic activity. On the contrary, if the pores are similar in size to the WOC, blocking of the host material porous upon immobilization can also limit catalytic activity [30,55].

Among possible porous hosts materials, metal–organic frameworks (MOFs) have gained great attention due to their interesting properties originating from their flexibility in structure and synthetic methods, allowing a large degree of structural and chemical tunability [30,55,111,113,114,115].

2.5. Electrografting/Electropolymerization

This strategy is based on the immobilization of molecular electrocatalysts by in situ hybridization with functional groups that can undergo polymerization under applied potential on the surface of photoelectrodes. In this regard, following this strategy, molecular WOCs can be either electro-polymerized or -oligomerized themselves or, alternatively, molecular WOCs can be incorporated in situ during electropolymerization of conductive matrix polymers on the desired photoelectrode. This strategy enables the preparation of photoelectrodes for WO with long photostability and catalytic activity without significantly altering the physical and electrochemical properties of the immobilized catalysts within the polymer films. Therefore, this approach enables the screening of several catalysts onto (photo)electrodes. Moreover, with these methods, catalysts are deposited forming a uniform polymeric film, the thickness of which can be controlled [30,31,55]. This approach has been used for the multicomponent assembly of dyes and catalysts on metal oxide electrodes for applications in dye-sensitized photoelectrochemical cells (DSPECs) [116].

In this context, electrografting is commonly used for the functionalization of carbon-based electrodes since they are inexpensive and can be straightforwardly modified [117,118]. In this case, diazonium electrografting is used to attach molecular catalysts onto carbon electrodes forming covalent bonds [118].

2.6. Atomic Layer Deposition

This is a relatively new strategy for preparing spatially controlled, multicomponent films consisting of molecular light-absorbing chromophores and water oxidation catalysts on high-surface-area porous metal oxides. It was initially developed by Templeton and Meyer to prepare (photo)anodes in a more efficient manner and render more stable hybrids [119]. This approach consists of depositing a thin layer of inert aluminum oxide (Al₂O₃) onto the surface of a dye that is surface bound to an electrode, covering it, and followed by catalyst (WOC) binding to the new oxide surface. In a final step, catalyst surface binding is stabilized by a another ALD overlayer of Al₂O₃. The ALD assembly procedure bypasses synthetic difficulties arising from the preparation of phosphonic acid-derivatized, covalently linked assemblies.

2.7. Layer-by-Layer Assembly: Electrostatic Self-Assembly Non-Covalent Immobilization

The layer-by-layer (LbL) deposition strategy is a thin-film fabrication technique based on the alternate deposition of layers of oppositely charged species that was initially developed by J. J. Kirkland of DuPont [120]. Therefore, the polyelectrolyte self-assembly occurs by the alternate exposure of an electrode surface to solutions of oppositely charged polyelectrolytes or polyions [121,122]. Finally, electrostatic, non-covalent immobilization strategies can be employed for the immobilization of molecular electrocatalysts. Moreover, this strategy can be combined with other immobilization techniques or set ups for several applications and thus has been used for the fabrication of diverse functional hybrid materials for a manifold of applications in biomedicine, energy conversion and storage and catalysis [121,122,123]. Additionally, an advantage of LbL is that it is a general strategy that can be applied to any (photo)electrode surface regardless of their type of material and shape [122,123,124,125].

3. Selected Examples of WO (Photo)Anodes

In this section, selected examples of molecular catalysts immobilized onto (photo)electrodes using different strategies explained above are revised and the respective advantages of their molecular counterpart analyzed. In most reported systems, molecular WOCs are immobilized on semiconductors and metal oxides. Additionally, the coimmobilization of dyes together with molecular WOCs or molecular dyads or triads containing WOCs, photosensitizers and electron mediators enables light-driven photoelectrocatalyzed WO. These assemblies can be applied in artificial photosynthetic devices.

3.1. Covalent Anchoring of Molecular WOCs onto Electrodes

This strategy is very versatile since many groups can be incorporated to the ligand scaffolds of molecular WOCs to covalently anchor them to (photo)electrode surfaces [55,57,62]. Several molecular WOCs based on Ru, Fe, Co, Mn, Fe and Cu have been explored for (photo)electrochemical WO using this immobilization strategy onto electrode surfaces [39,43,56,60,61,62,126,127].

Among all examples, Li et al. reported a molecular Ir WOC covalently bonded to hematite photoanodes in a monolayer using an oxobridge between the Ir active sites and the Fe element of the photoanodes (Ir-O-Fe) forming a ultra-thin film with greatly increased photocatalytic activity for WO at acidic pH (WOC-1, Figure 2a), which was remarkable taking into account that there are still only a few heterogeneous OER electrocatalysts that maintain high catalytic activity and stability at acidic pH values, such as highly expensive IrO_x [128]. This heterogenized Ir-WOC exhibited much higher activity than its homogeneous counterpart and was active within a water-splitting device with stable activity for over 5 h at near 100% FE at acidic pH, which was the first example reported to complete unassisted water splitting by hematite-derived electrodes in acidic solutions. This was associated with the efficient transfer of charge carriers between the photoelectrode and the catalyst. However, the system needed further work to stabilize the WOC/hematite combination. At the same time, Hintermair, Crabtree and Brudvig reported the immobilization of the same system on metal oxides showing increased catalytic activity and stability [129]. The system was reported to produce over 100,000 TON O₂ per iridium atom after ~16 h of electrolysis, and XPS studies confirmed the stability of the system.

Another example using covalent immobilization of WOCs onto electrode surfaces was reported by the group of Sun. They reported a Ru-WOC-bearing 2,6-pyridinedicarboxylic acid (PDC) ligand that could be efficiently immobilized on hematite nanorod arrays via the carboxylic acid groups together with a Ru photoredox catalyst anchored via phosphonate functional groups [Ru(4,4′-(PO₃H₂)₂-bpy)(bpy)₂]²⁺ (4,4′-(PO₃H₂)₂-bpy = [2,2′-bipyridine]-4,4′-diylbis(phosphonic acid) and bpy = 2,2′-bipyridine) (WOC-2 and PC-1, respectively, Figure 2b). The system showed efficient solar WO at alkaline pHs with a steady photocurrent of 25 μA cm⁻² [76]. Moreover, the photoanode was stable at various pHs for a prolonged time for over 3 h due to the employment of strong PDC anchoring group. Using the same carboxylic acid as anchoring groups, L-Z. Wu et al. reported the immobilization of a Co-WOC bearing a porphyrin with carboxyl groups, CoTCPP (TCPP = meso-tetra(4-carboxyphenyl)porphyrin) on Al₂O₃-modified BiVO4 photoanodes for PEC water oxidation [130]. Under illumination of a 100 mW cm^–2 Xe lamp, the photoanode exhibited a 2-fold enhancement in photocurrent density at 1.23 V vs. RHE and nearly a 450 mV cathodic shift at 0.5 mA cm⁻² photocurrent density relative to bare BiVO₄ at pH = 6.8 in a PEC configuration, generating O₂ and H₂ with a faradic efficiency of 80% over 4 h. Moreover, the authors demonstrated that immobilized molecular CoTCPP catalyst greatly suppressed the hole–electron recombination on the surface of BiVO₄ semiconductor. Besides organometallic compounds, these approaches are also effective for the assembly and integration of other molecular catalysts [76,107,131].

Then, using silane-prefunctionalized electrodes and the same [Ru(4,4′-(PO₃H₂)₂-bpy)(bpy)₂]²⁺ as photoredox catalyst, the group of Sun reported a FTO/TiO₂ photoanode bearing the Ru photocatalyst immobilized via the phosphonate anchoring and a Ru-bda [132] derivative Ru-WOC [Ru(bda)(pic)(L)] (L = N-(3-(triethoxysilyl)propyl)isonicotinamide) anchored by the silane groups (PC-1 and WOC-3 Figure 2c) [91]. This system was active for visible light-driven water splitting at pH 6.8 evolving O₂ and H₂ at 0.2 V vs. NHE applied, generating a high photocurrent density > 1.7 mA cm⁻² after 10 s light irradiaiton, which at that time was higher than any PEC devices with molecular components reported in the literature. Other systems have been reported using silanes as anchoring groups in Ru-bda derivative WOCs by Meyer and coworkers showing high activity for photoelectrocatalyzed WO, but the system showed long-term loss of catalytic activity due to catalyst loss from the electrode surface by axial ligand dissociation in the high oxidation states of the WOC [96].

Meyer and Concepcion reported a new Ru-based polypyridylbased chromophore−catalyst assemblies, [Ru(4,4′-(PO₃H₂)₂-bpy)₂(4-Mebpy-4′-epic)-Ru(bda)(pic)]²⁺ (Mebpy-4′-epic = 4-(4-methylbipyridin-4′-yl-ethyl)-pyridine; bda = 2,2′-bipyridine-6,6′-dicarboxylate; pic = 4-picoline, PC-2 and WOC-4, respectively, Figure 2d) covalently anchored on a SnO₂/TiO₂ core shell electrode to be applied in a dye-sensitized photoelectrosynthesis cell (DSPEC) for solar water splitting with a Pt cathode. This was the first reported covalently linked chromophore–catalyst assembly incorporating a Ru-bda-derived WOC that performed both electrochemical and photoelectrochemical WO on oxide surfaces. The system was active for light-driven WO at pH 5.7 and under white illumination (100 mW cm⁻²) photocurrents of 0.85 mA cm⁻² were observed after 30 s under a 0.1 V vs. Ag/AgCl applied bias with a FE for O₂ production of 74% after 5 min irradiation [133]. The same group then reported the use of the same phosphate anchoring group to immobilize a Ru photoredox catalyst [Ru(4,4′-(PO₃H₂)₂-bpy)(bpy)₂]²⁺ (4,4′-(PO₃H₂)₂-bpy = [2,2′-bipyridine]-4,4′-diylbis(phosphonic acid) and bpy = 2,2′-bipyridine) and a derived Ru-bda [Ru(bda)Het₂] [133] (bda = 2,2′-bipyridine-6,6′-dicarboxylate) WOC with phosphonate anchoring groups onto MO_x/FTO electrodes to develop a photoanode for use in a DSPEC [75]. The system was active for photoelectrocatalyzed WO, generating a photocurrent of 1.4 mA cm⁻² at pH 7 with minimal loss of catalytic activity over 30 min, with an IPCE of 24.8% at 440 nm.

Another approach was reported by Reisner and coworkers, where they developed a dyad photocatalyst consisting of a diketopyrrolopyrrole chromophore (DPPdye, PC-3, Figure 2e) and Ru-based WOC (WOC-5) featuring a cyanoacrylic acid anchoring group for its immobilization onto a mesoporous TiO₂-derived FTO/TiO₂ electrode. The assembled dyad-sensitized photoanode was active for O₂ evolution using visible light (100 mW cm⁻², AM 1.5G, λ > 420 nm), producing a initial photocurrent of 140 mA cm⁻² at pH 5.6 under an applied potential of 0.2 V vs. NHE, with 44% FE with a TON of 2.3 for the WOC and 9.2 regarding the dye [73]. Under photoelectrochemical experiments, the photocurrent decreased during irradiation time, which was associated with catalyst detachment or decomposition, whilst the dyad remained adsorbed onto the electrode and the chromophore was intact.

3.2. Supramolecular Assembly of Molecular WOCs onto Electrodes

This strategy has gathered interest since it allows reversible immobilization of the WOCs onto guests that are anchored in the electrodes. One advantage of this approach is that it allows the detachment of the deployed WOC from the guest-electrode surface once catalytic activity decreases, as well as reattachment of a new batch of fresh catalyst to restart catalytic activity. Moreover, as seen in the previous section, this strategy generally avoids significant changes in the electrochemical and energetic properties of the WOCs [30,31,55,60,97,103].

In this line, the group of Reek mention a supramolecular encapsulation strategy to immobilize ruthenium complexes onto electrodes. Their approach involved the use of M₁₂L₂₄ nanospheres [134] (where M correspond to the corners which are Pt²⁺ or Pd²⁺, and L the bipyridyl building block) endohedrally functionalized with guanidinium that can strongly bind to other sulfonate bearing-catalysts via hydrogen bonding interactions. Therefore, they developed two Ru complexes derived from the [Ru(bda)Het₂] [132] WOC (bda = 2,2′-bipyridine-6,6′-dicarboxylate) and [Ru(phenda)(Het₂)] (phenda = [1,10] phenanthroline-2,9-dicarboxylic acid) [135] bearing sulfonate functionalized pyridine axial ligands for supramolecular attachment to the guanidinium nanospheres: [Ru(bda)(PySO₃TBA)₂] and [Ru(phenda)(PySO₃TBA)₂] based on the PySO₃TBA = pyridine-3-sulfonate, WOC-6 and WOC-7, respectively, Figure 3a). The encapsulation of dilute solutions of the catalysts (10⁻⁵ M) inside the guests lead to high local catalyst concentrations (up to 0.54 M). In the case of WOC-6, the increase in local concentration resulted in an enhanced water oxidation rate by two-orders of magnitude to 125 s⁻¹, 130-fold higher than that observed for the non-encapsulated catalyst [98]. This was attributed to a facilitated diffusion-controlled rate-limiting dinuclear coupling step, since for complex WOC-6 a dinuclear DC mechanism (radical oxo coupling) to form the O–O bond is also possible [53]. While for WOC-7, since it exclusively operates via WNA [135], the reaction rate for WO did not change by encapsulation of the catalyst. Electrolysis experiments at 1.3 V showed 3-fold more current that in the absence of nanosphere for WOC-6 with 96.3% Faradaic efficiency for O₂ solution using a rotating disk electrode. This example showed that great rate enhancements can be achieved by spatial organization of catalytically active species following the dinuclear mechanism [98]. Moreover, this strategy was proven successful to modulate reaction rates by changing local catalyst concentrations by supramolecular interactions and it enabled the study of electrocatalytic reaction mechanisms.

Another strategy was reported by Sun and coworkers, where they used supramolecular interactions to assemble a DSPEC for water splitting using poly(methylmethacrylate) (PMMA) as auxiliary binder and stabilizing material [136,137] and a Ru-WOC modified with long hydrocarbon chains via hydrophobic interactions. They used a phosphonate functionalized Ru-photoredox catalyst [Ru(4,4′-(PO₃H₂)₂-bpy)(bpy)₂]²⁺ (4,4′-(PO₃H₂)₂-bpy = [2,2′-bipyridine]-4,4′-diylbis(phosphonic acid) and bpy = 2,2′-bipyridine) and heterogenized it on a TiO₂/FTO electrode by the phosphate anchoring groups, and a Ru-bda [132] derivative WOC [Ru(bda)(pic)(Het)] (bda = 2,2′-bipyridine-6,6′-dicarboxylate, pic = 4-picoline) with one of the axial pyridine ligands modified with a C₁₂ hydrophobic alkyl chain, which was mixed with PMMA in solution and co-assembled with the dye-TiO₂/FTO electrode to generate the WO photoanode. The aliphatic chain was crucial to obtain stable light-driven water oxidation catalysis keeping the catalyst in the polymer matrix [138]. Upon irradiation, the photoanode generated a photocurrent density of 1.1 mA cm⁻² and an incident-photocurrent conversion efficiency (IPCE) of 9.5% at 0.2 V vs. NHE applied bias. Moreover, the catalytic system displayed good stability during photocatalytic experiments. Later, the group of Concepcion used a similar Ru-WOC functionalized with a C₁₀ alkyl chain in one of the axial pyridines [Ru(bda)(pic)(4-C10H21-py)] (WOC-8, Figure 3b) and combined it with an aliphatic C₉ chain functionalized Ru-based photoredox catalyst [Ru(4,4′-(PO₃H₂)₂-bpy)₂(4,4′-(C₉H1₉)₂-bpy)]²⁺ (PC-4, 4,4′-(PO₃H₂)₂-bpy = [2,2′-bipyridine]-4,4′-diylbis(phosphonic acid) and bpy = 2,2′-bipyridine) that also had phosphonate functional groups to allow anchoring onto electrode surfaces, preparing therefore, supramolecular assemblies of Ru-WOC and Ru dye in the hydrophobic interface and immobilized onto metal oxide-based electrodes. Moreover, in the same report, they also showed the functionalization of metal oxide electrode surfaces with self-assembled monolayers with long hydrophobic chains, which enabled the supramolecular immobilization of Ru-WOCs also functionalized with long alkyl chains, rendering more stable electrodes [103]. A follow-up paper of the same group used their previously reported WOC-8 and PC-4 assembly prepared by a self-assembled bilayer on a mesoporous SnO₂/TiO₂ core/shell electrode (FTO|SnO₂/TiO₂|PC4-WOC-8, Figure 3b). This photoanode was highly active for WO for over 3 h with 86% FE for O₂ production. Under 1 sun illumination, the photoanode reached a photocurrent density of 2.2 mA cm⁻² with an IPCE of 29% at 450 nm [139]. This demonstrated that hydrophobic interactions are suitable to in situ assemble and catalysts and chromophores on electrode surfaces.

Then, using a similar Ru WOC ([Ru(bda)((3-(pyridine-4-yloxy)propyl)phosphonic acid)₂], WOC-9) and photoredox catalyst ([Ru(bpy)(4,4′-(PO₃H₂)₂-bpy)₂](Cl)₂, PC-5, Figure 3c), and inspired by the supramolecular Zr-phosphonate chemistry developed by the group of Mallouk and used to prepare dye–catalyst assemblies using a layer-by-layer [141], the group of Meyer used this strategy to construct a preorganized multicomponent assembly consisting of electron acceptors, electron donors, photoredox catalysts and WOCs (MV²⁺ (methyl viologen), PC-5, Fe(II) donor ([Fe(2,2′:6′,2′′-terpyridyl-4-phenyl-phosphonic acid)](Cl)₂) and WOC-9, respectively, see Figure 3c), with the attempt to prepare a bioinspired artificial analogue of PSII [72]. The location of the components on the electrode also mimics the spatial arrangement of the key elements of PSII: conducting oxide electrode surface (Plastoquinone in PSII), MV²⁺ (pheophytin), PC-5 (P680 Dimer), Fe(II) donor (tyrosine) and WOC-9 (Mn cluster of the OEC), generating a nanoITO|-MV²⁺ -PC-5 -Fe(II)donor -WOC-9 photoelectrode assembly. In a three-electrode configuration photoelectrochemical cell, under 1 sun illumination (100 mW cm⁻², 400 nm filter cut off) for 30 s at pH 4.65 a photocurrent density of 250 μA cm⁻² were generated at 0.5 V vs. NHE applied bias. In continuous O₂ evolution experiments, the photocurrent decreased to 75 μA cm⁻² after the first 10 min of reaction and remained constant at 50 μA cm⁻² for over 1 h of irradiation with a FE for O₂ production of 67%. Further studies demonstrated that the system generated reductive equivalents at the electrode surface and oxidative equivalents at the WOC side a catalyst that persist for seconds in aqueous solutions, proving the ability of the system to mimic functional elements of PSII including charge separation and water oxidation. Moreover, steady-state illumination of the assembly with 440 nm light with an applied bias results in photoelectrochemical water oxidation with a per-photon absorbed efficiency of 2.3% after 10 s of photolysis [72].

More recently, the same group reported a related study describing a single molecular assembly electrode that duplicated the key components of PSII, consisting of a dyad formed by a similar Ru phosphonated polypyridyl photoredox catalyst, supramolecularly linked to an intermediate electron donor via the supramolecular Zr-phosphonate chemistry, and covalently linked to a molecular-based Ru-bda-derived WOC and protected and stabilized by a 4,5-difluoro-2,2-bis(trifluoromethyl)1,3-dioxole (AF) polymer on a SnO₂/TiO₂ core/shell electrode. This photoanode mimicked PSII in achieving sustained, light-driven WO catalysis. Under illumination and at pH 7 the photoanode produced O₂ with an efficiency of 83% and an IPCE value of 10.9% at its absorption maximum of 460 nm, with an applied bias of 0.6 V vs. NHE. This example highlighted the value of the tyrosine–histidine pair in PSII in achieving efficient water oxidation catalysis in artificial photosynthetic devices by the introduction of the intermediate electron donor between the photoredox catalyst and the WOC [142].

Using hydrophobic interactions as a strategy to incorporate Ru WOCs onto electrode surfaces to perform electrocatalyzed WO at neutral pH, Llobet and coworkers reported oligomeric Ru complexes that can be simply anchored on the surface of multiwalled carbon nanotubes via CH–Π interactions between the axial ligands bound to the Ru-WOCs and the hexagonal rings of the graphitic surface of the material, providing control of their molecular coverage. These hybrid molecular materials behaved as molecular anodes that catalyzed WO at pH 7 with high current densities [143]. With the same immobilization strategy, the same group reported a homogeneous Cu-WOC featuring a tetraanionic amidate ligand N₁,N₁′-(1,2-phenylene)bis(N₂-methyloxalamide) was derivatized with a pyrene moiety in the pyridine ligand were supported onto graphite electrodes through supramolecular Π-Π-stacking supramolecular interactions. This system showed electrocatalyzed WO activity at 538 mV vs. NHE overpotential, which was 160 mV lower than its homogeneous counterpart, and higher rate (540 s⁻¹). This enhancement of catalytic activity was attributed to the pyrene moiety and the electrode stacking interactions [144]. Additionally, Co-corrole complexes have also been immobilized for electrocatalyzed water splitting using the same immobilization strategy [145].

Finally, the group of Tilley applied a non-covalent chemistry approach to reversibly bind molecular electrocatalysts to electrode surfaces through host–guest complexation with surface-anchored bioinspired cyclodextrins, since allows for strong binding between the host and the guest and enables the flow of electrons between the electrode and the guest catalyst. The authors reported the use of two different molecular Ru and a Pt WOCs bearing a naphthalene moiety as substituent in the 2,2′:6′,2′’-terpyridine derivative ligand to allow assembly with the host: WOC-10, WOC-11 and WOC 12, respectively (Figure 3d). These systems showed a greater stability when immobilized onto the electrode surface than in solution, which was attributed to the large binding energies obtained by DFT calculations. Electrosynthesis in both organic and aqueous media was demonstrated on metal oxide electrodes, with stability on the order of hours. Moreover, the authors showed that the catalytic surfaces could be recycled by controlled release of the guest from the host cavities and the reabsorption of fresh WOC guest by tunning the conditions [140]. This study showed the suitability of multidentate supramolecular hosts as the anchoring moieties for molecular WOC guests to generate stable and electrocatalytically active surface-bound host–guest complexes in both aqueous and organic solvents for electrocatalytic WO.

3.3. Integration of Molecular WOCs onto Electrode Surfaces by Drop-Casting

The group of Sun reported the immobilization of a molecular Co₄O₄ cubane WOCs onto a bismuth vanadate (BiVO₄) electrode by drop-casting using Nafion as binder to develop different anodes for photoelectrochemically driven WO. They synthesized a family of Co₄O₄ cubane WOCs by changing the substituent groups on both the pyridine ligands and the alkoxy carboxylato bridging ligands (R₁ and R₂ in WOCs 13 to 21, respectively, Figure 4) to screen different photoanodes for WO. The authors could demonstrate that WO showed a ligand dependent activity that was tuned by modifying the substituent at the Co-WOC immobilized on the WOC/Nafion/BiVO₄ photoanode. Among the series, the cubane catalysts featuring long hydrophobic C₄ alkyl chains (to favor direct physisorption) in the alkoxy carboxylato bridging ligands (WOC-20 Figure 4) drop-casted onto BiVO₄ was found to be the most active of the series, yielding a photocurrent density for water oxidation of 5 mA cm⁻² at 1.23 V vs. RHE applied potential under the illumination of simulated sunlight (AM 1.5, 100 mWcm@2) in pH 9.3 borate buffer in a three-electrode configuration, together with a 1.84% solar-to-energy-conversion-a 6-fold enhancement over that of bare BiVO₄ [107]. Moreover, the obtained photocurrent was the highest photocurrent for undoped BiVO₄ photoanodes and comparable to or even better than those attained by the state-of-the-art metal oxide catalysts. However, the activity started decreasing after 1 h due to slow desorption of the Co-WOC from the electrode during long-term photolysis experiments at 0.7 V vs. RHE applied bias [107].

3.4. Utilization of Physical Confinement Strategies for the Generation of WO (Photo)Anodes

Due to the high porosity, crystallinity, large surface area, modular nature and synthetic versatility MOFs have been widely used for applications in catalysis. Moreover, they have received increasing interest for their use as hosts to stabilize molecular metal complexes for the production of fuels in the field of artificial photosynthesis. For instance, the group of W. Lin pioneered the utilization of MOF-based heterogeneous systems for chemically driven water oxidation upon the incorporation of iridium WOCs into the rigid and protective structure of MOF UiO-67 [146]. Moreover, there are some examples in which MOFs have been used to incorporate molecular WOCs for their application in (photo)electrocatalyzed WO when supported onto electrode surfaces [111,147,148,149,150], due to their stabilization effect over the metal complex. For instance, Ott and coworkers reported the incorporation of a [Ru(tpy)(dcbpy)(OH₂)](ClO₄)₂ (tpy = 2,2′:6′,2′′-terpyridine, dcbpy = 2,2′-bipyridine-5,5′-dicarboxylic acid) catalyst (WOC-22, Figure 5a) into a FTO-grown thin films of the UiO-67 (UiO = University of Oslo) MOF using a post-synthetic ligand exchange approach. The UiO-67-[Ru(tpy)(dcbpy)(OH₂)](ClO₄)₂@FTO system evolved O₂ at pH 8.4 at 1.5 V vs. Ag/AgCl applied potential with a FE of 82% and a current density of 11.5 μA cm⁻² [111]. The system was stable for hours under catalytic conditions, as analyzed by the persistence of the redox wave associated with the Ru^III/II couple in the CV indicating the integrity of the catalytic system, which was also supported by powder X-ray diffraction (PXRD) and scanning electron microscopy (SEM) images. Moreover, the catalytic current was stable over multiple scans.

Another system was reported by the group of X-M. Chen, which using post synthetic ion exchange approach to prepare a Co-WOC containing MOF ([Co₂(μ-OH)₂(bbta)] (H₂bbta = 1H,5H-benzo-(1,2-d:4,5-d′)bistriazole) that contained both open metal sites as well as hydroxide ligands (WOC-23, Figure 5b) and retained its original crystallinity in either acidic or basic solutions for over a week. This material showed an improved electrocatalytic activity for O₂ evolution obtaining 10.0 mA cm⁻² at 292 mV overpotential at pH 14 when coated on a glassy carbon electrode using Nafion as binder, which surpassed the activity of its inorganic Co(OH)₂ and Co₃O₄ analogues under similar conditions [147]. Isotopic labelling experiments confirmed the role of the hydroxide ligands in the WO reaction, which lowered the energy for the direct coupling pathway.

In the same line, using MOFs as host porous materials, Das and coworkers reported a hybrid metal–organic material [{Co₃(μ₃-OH)(BTB)₂(dpe)₂}^- {Co-(H₂O)₄(DMF)₂}_0.5]_n·nH₂O in which the mononuclear positively charged Co-WOC [Co-(H₂O)₄(DMF)₂]²⁺ (WOC-24, Figure 5c) is encapsulated in the host MOF resulting in a host–guest supramolecular system, where the negative charge of two {Co₃(μ₃-OH)(COO)₆}^- units (inactive for WO) are balanced by the molecular Co WOC-24 (active for WO). This results in the formation of a rigid environment around the active site. This system was coated onto a glassy carbon electrode to explore its catalytic activity towards WO, showing stability in basic media and a remarkable electrocatalytic activity for O₂ evolution with a TOF of 0.05 s⁻¹ at 390 mV vs. NHE overpotential at pH 13 with a current density of 1 mA cm⁻² [148]. Catalytic tests showed the maximum WO activity at pH 13 due to the coordination of hydroxo ligands (OH^-) to the mononuclear Co ion, which compensates the extra positive charge of the overall host–guest system. Additionally, mechanistic studies confirmed that only the [Co-(H₂O)₄(DMF)₂]²⁺ WOC-24 was responsible for the WO activity. Furthermore, stability tests by ICP, EDX and electrochemical studies discarded any leaching of the Co catalyst from the MOF into the electrolyte solution after 1000 catalytic scans, which reaffirms the stability provided by the physical confinement in the guest.

Later on, Wang and coworkers reported a Co-based WOC (WOC-25, Figure 5d) that consisted of a MOF where the redox function of the embedded cobalt active centers was modified by introducing imidazolate ligands as linkers to facilitate the proton transfer process involved in WO, generating a Co-containing zeolitic imidazolate framework (Co-ZIF) that was supported onto an FTO electrode to explore its electrochemical WO capability. This system showed WO activity at pH 7 with the onset of the catalytic current at 1.6 V vs. RHE (400 mV vs. RHE overpotential), producing 20 TON O2 after 3 h electrolysis with a TOF of 1.76·10⁻³ s⁻¹. The system remained stable over 25 h operation. Moreover, it was also electrocatalytically active for O₂ evolution in a wide range of pH values. Finally, DFT calculations of the system proved the thermodynamic feasibility of Co-ZIF to activate the water molecule via binding the OH group to the metal sites with low activation barriers, while the eliminated proton was accepted by the nearby benzimidazolate ligands, which was in agreement with the observed reactivity [149].

Kern and coworkers developed novel heterobimetallic MOFs by expanding a porphyrin network using 5,10,15,20-tetra(4-pyridyl)- 21H,23H-porphyrine (TPyP, Figure 5e) and rationally occupying the coordination sites by single iron or cobalt atoms in a specific manner, which was characterized by scanning tunneling microscopy (STM). That is, either with the Fe in the coordination center of the porphyrin while the Co were placed in the corners between four neighboring porphyrins, or by placing the Co atoms in the coordination center and the Fe in the corners (M1-TPyP-M2, WOC-26, Figure 5e) [150]. All the developed MOFs were active for electrochemically driven WO when supported on a gold electrode at pH 13. Moreover, catalytic activity increased dramatically depending on the coordination spheres of Fe and Co, with the heterobimetallic M1-TPyP-M2 MOFs displaying faster reaction kinetics and higher activity for O₂ evolution than the single Fe or Co TPyP porphyrins at 300 mV vs. Ag/AgCl overpotential, which is lower than the prototypical hangman porphyrins for WO (600 mV overpotential). The FeTPyP-Co MOF showed the highest electrocatalytic activity (3.43 nmol O₂ cm⁻²) and fastest kinetics (12.2 s⁻² TOF > 15-fold higher than the hangman porphyrin [151]). Mechanistic studies based on electrochemistry and preliminary X-ray absorption spectroscopy (XAS) showed the importance of the coordination environment of the metal centers and the synergistic effect of the two proximate metal centers, as well as suggesting that the pyridyl-Co moiety was a critical site for WO. This highly modular porphyrin-based MOF demonstrated the manifold synthetic possibilities and combinations to explore to develop more highly active heterogenized WOCs by just changing the metal ions and the meso-substituents in the porphyrin.

3.5. Heterogenization of WOCs by Electropolymerization

Among the first examples of electrografted WOCs, Lin et al. reported iridium-based WOC containing a pentamethylcyclopentadienyl (Cp*) ligand and 2,2′-bipyridine or 2-pehnylpyridine ligand with amine pendant groups ([Cp*IrCl(4-NH₂-bpy)]Cl, [Cp*IrCl(5-NH₂-bpy)]Cl and [Cp*IrCl(p-NH₂-ppy)]) that were immobilized onto a glassy carbon electrodes upon diazonium electrografting at −0.4 V vs. NHE applied potential for 4 min on a 1.13 cm² planar carbon electrode. The heterogenized systems showed TOFs up to 3.31 s⁻¹ when using catalyst [Cp*IrCl(p-NH₂-ppy)] and were proven to be more stable that its homogeneous counterpart and was active without loss of the covalently bound WOC for 3 h and then some catalysts was lost during the electrocatalysis, but this was only apparent after 10 days of electrolysis. This catalyst leaching from the electrode surface was responsible for the depletion of catalytic activity [118].

Then, Meyer and co-workers demonstrated that water-stable photoactive interfaces on TiO₂ electrodes could be prepared by reductive electropolymerization of catalysts onto TiO₂ immobilized dyes. The group reported a Ru dye containing vinyl groups in two of the 2,2′-bipyridine and a carboxylic acid groups in the other [Ru(vbpy)₂(dcb)](PF₆)₂ (vbpy = 4-methyl-4′-vinyl-2,2′-bipyridine and dcb = 2,2′-bipyridine-4,4′-dicarboxylic acid) immobilized onto a TiO₂ electrode via the carboxylate groups that was stabilized upon reductive electropolymerization of a [Ru(vbpy)₂(dppe)](PF₆)₂ (dppe = cis-1,2-Bis-(diphenylphosphino)ethylene) complex forming an immobilized TiO₂-[Ru(vbpy)₂(dcb)](PF₆)₂/poly-[Ru(vbpy)₂(dppe)](PF₆)₂ composite. The polymeric layer was transparent to visible light and stable in water that was photoactive and could last for up to 3 weeks [152]. Later on, the same group used this approach to test the stabilization of photosensitizers immobilized onto (photo)electrodes with phosphonate anchoring groups upon electropolymerization of a molecular Fe-based complex onto mesoporous TiO₂ electrodes previously functionalized with [Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂ (dvb = 5,5′-divinyl-2,2′-bipyridine, 4,4′-(PO₃H₂)₂-bpy = [2,2′-bipyridine]-4,4′-diylbis(phosphonic acid)) anchored to TiO₂ through the phosphonate at the 2,2′-bipyridine [153]. A [Fe(v-tpy)₂](PF₆)₂ (v-tpy = 4′-vinyl-2,2′:6′:2″-terpyridine, Fe-1 Figure 6a) complex was reductively electropolymerized onto TiO₂-[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂ generating an electropolymerized overlayer of TiO₂-[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂-poly-[Fe(v-tpy)₂](PF₆)₂ [153]. This resulted in a 30-fold enhancement of the photostability as compared to the surface-bound dye alone or immobilized via ALD. This example proved the generalization of electropolymerization to stabilize different types of anchored dyes onto electrode surfaces.

Later on, the same group proved the scope of this strategy and demonstrated that Ru-WOCs can be introduced onto those derivatized nTiO₂-, pFTO- or nITO-[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂ by reductive electropolymerization of the vinyl groups present in both the photoredox catalyst and the complexes, generating photostable and water stable photoanodes for WO [116,154]. First, they electropolymerized a [Ru(Mebimpy)-(dvb)(OH₂)]²⁺ (WOC-27, Mebimpy = 2,6-bis(1-methyl-1H-benzo[d]imidazole-2-yl)pyridine) WOC onto nTiO₂-, pFTO- or nITO-[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂ that showed a dramatic increase in photostability up to 16 h (Figure 6a) [116]. The electropolymerization of the Ru-WOC onto the dye-derivatized electrode stabilized and protected the dye from leaching, since the phosphate immobilized dye-oxide surface photoelectrode showed chromophore loss since the beginning of the irradiation time. Electrocatalyzed WO was maintained for more than 2 h using either nITO-poly-[Ru(Mebimpy)-(dvb)(OH₂)]²⁺ or nITO -[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂-poly-[Ru(Mebimpy)-(dvb)(OH₂)]²⁺ with 77% FE and 501 TON O₂ and 0.046 s⁻¹ TOF.

After that the same group reported a photocatalytically active photoanode for WO by reductive electropolymerization of a Ru-bda [132] derivative WOC [Ru(bda)(vpy)₂] (bda = 2,2′-bipyridine-6,6′-dicarboxylate, vpy = 4-vinylpyridine) on glassy carbon (GC) andnTiO₂ electrodes results in films containing semirigid polymer networks that produce sustained water oxidation at pH 7 as measured in real-time by use of a two-electrode cell [119,154]. Using a rotating disk electrode the GC-[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂-poly-[Ru(bda)(vpy)₂] produced O₂ with 62% FE at pH 7 with a catalytic onset at 0.95 V vs. SCE, which was consistent with the onset potential of WOC-28 (Figure 6b). With this configuration, a TOF of 8.5 s⁻¹ was obtained at 1.2 V vs. SCE applied bias (620 mV overpotential). Finally, the system was tested for photocatalyzed WO, using the nTiO₂-[Ru(dvb)₂(4,4′-(PO₃H₂)₂-bpy)]Cl₂-poly-[Ru(bda)(vpy)₂] photoanode in a DSPEC configuration using a Pt cathode for H₂ production at 0.2 V vs. SCE applied bias, producing O₂ with a FE of 8% under 1 sun illumination (100 mW cm⁻²) [154]. The study of this system showed that the coordination chemistry and proton-coupled electron-transfer properties (PCET) of complex WOC-28 are maintained upon electropolymerization. However, there is a change in the mechanism for the O–O bond formation from DC in solution to a single-site WNA in the film environment onto the electrodes. This example further reinforced the capacity of this immobilization strategy to heterogenized active WOCs with or without dyes onto electrodes for (photo)electrocatalyzed WO.

3.6. Use of Atomic Layer Deposition for the Fabrication of WO (Photo)Anodes

Templeton, Meyer and coworkers developed the atomic layer deposition (ALD) approach to prepare (photo)anodes in a more efficient manner and rendering more stable hybrids. They prepared a Ru(II) polypyridyl photoredox catalyst and WOC assembly on nanoparticle films of tin-doped indium oxide (nanoITO) and titanium dioxide (nanoTiO₂) for electrocatalytic and photoelectrocatalytic WO, respectively using ALD [119]. First, the photoredox catalyst [Ru(4,4′-(PO₃H₂)₂-bpy)(bpy)₂]²⁺ (4,4′-(PO₃H₂)₂-bpy = [2,2′-bipyridine]-4,4′-diylbis(phosphonic acid) and bpy = 2,2′-bipyridine, PC-1, Figure 7) was coated onto films of nITO and nTiO₂ generating nITO|PC-1 and nTiO₂|PC-1, after which overlayers of aluminum oxide (Al₂O₃) were deposited atop the nITO|PC-1 and nTiO₂|PC-1 by ALD to ensure the photostability of the photoredox catalyst. Then, a [Ru(Meipy)(4,4′-(PO₃H₂CH₂)₂-bpy)(OH₂)]²⁺ (Meipy = 2,6-bis(1-methyl-1H-benzo[d]imidazol-2-yl)pyridine and 4,4′-(PO₃H₂CH₂)₂-bpy = 4,4′-((HO)₂(O)P–CH₂)₂-2,2′-bipyridine) catalyst (WOC-29, Figure 7) was loaded from a methanol solution onto the nITO|PC-1|Al₂O₃ and nano TiO₂|-PC-1| Al₂O₃ and more layers of Al₂O₃ were deposited on top to ensure the electrochemical stability of the Ru WOC. The electrodes were characterized by CV, SEM and UV–vis spectroscopy. Currents higher than 60 μA cm⁻² and TOF O₂ 0.014 s⁻¹ (23% FE) were obtained for the nITO|PC-1|Al₂O₃|WOC-29|Al₂O₃ at 1.4 V vs. NHE applied potential, as quantified in a dual working electrode collector-generator cell [119]. However, the current decreased to 20 μA cm⁻² after 2 h most likely due to ligand decomposition in the Ru WOC at higher oxidation state. Photoelectrochemical studies with nTiO₂|PC-1|Al₂O₃|WOC-29|Al₂O₃ with a white light source at 0.64 V vs. NHE applied potential generated O₂ with at 16.8% FE after 6 h irradiation.

3.7. Use of Layer-by-Layer Deposition to Deposit WOCs onto Electrodes

The groups of Meyer and Schanze and coworkers reported for the first time the assembly of chromophores and WOCs by using the polyelectrolyte layer-by-layer deposition method [55,155]. Their system consists on a cationic polystyrene-based Ru polychromophore (PC-7) and a [Ru(tpy)(pyMebzi)(OH₂)]²⁺ (tpy = 2,2′:6′,2″-terpyridine and pyMebzi = 2-pyridyl-N-methylbenzimidazole) Ru-based catalyst (WOC-30), co-deposited with poly(acrylic acid) (PAA) as an inert polyanion onto planar indium tin oxide (ITO) substrates for electrochemical characterization and onto mesoporous electrodes consisting of a SnO₂/TiO₂ core/shell structure on top of fluorine doped tin oxide (FTO) for photoelectrochemical water oxidation in a dye-sensitized PEC (Figure 8) [155]. The resulting photoanodes were characterized by CV, UV–vis spectroscopy and SEM. Photoelectrolysis WO experiments were performed using a dual working electrode collector-generator cell [119] to allow for O₂ quantification, obtaining 22% FE for O₂ evolution at 0.44 V vs. NHE applied bias with 1.4 V vs. NHE as the onset for WO.

4. Conclusions and Perspective

Despite the challenges faced using molecular water oxidation catalysts in artificial photosynthetic devices, the heterogenization of molecular systems onto electrodes for (photo)electrocatalyzed water oxidation has several advantages over their respective homogeneous counterparts but also over heterogeneous systems. For instance, the heterogenization of molecular water oxidation catalysts allows for modular systems based on well-defined molecular WOCs that are easily tunable, while surpassing the long-term stability of their homogeneous counterparts. Moreover, these systems are easier to study and therefore to use to establish common designing principles after analysis of those catalytically active (photo)anodes for WO.

In this review we have seen several methods to immobilize WOCs and WOCs/photoredox catalysts onto different electrode surfaces. Some strategies are preferred over others as the initial properties of the molecular catalysts are preserved and due to the increased stability that they render to the final (photo)anodes. For instance, the use of electrografting or the more recent layer-by-layer deposition offer a more efficient and less synthetically tedious approach to the heterogenization of molecular water oxidation catalysts onto electrodes yielding more stable (photo)anodes. These strategies show those advantages over other approaches such as covalent immobilization which often requires complex synthetic derivatizations and can lead to dramatic changes in the electrochemical and physical properties of the molecular catalysts, which lead to changes in their catalytic performance. Despite the fact that supramolecular approaches can simplify synthetic derivatization procedure, the stability of these assemblies can be more susceptible to the catalytic conditions and hamper applications in artificial photosynthesis where synchronous reduction and oxidation reactions take place. It is worth mentioning that the revised immobilization strategies can also be applied to other catalysts beyond WO and there are several examples reported elsewhere.

To develop photoanodes for WO via heterogenization of molecular WOCs for practical applications, special attention needs to be put in the preservation of catalytic activity after immobilization onto electrode surfaces, as well as the stability of the assembly in the long term under (photo)catalytic conditions. Moreover, an overall improvement of the photo-driven catalytic activity is desired upon coimmobilization of dyes and WOCs onto electrodes for efficient and sustained WO. Despite some advances in the field of immobilization of molecular WOCs for artificial photosynthetic devices, future progress is needed in the development of (photo)anodes based on non-precious metal-based WOCs with high efficiency and catalytic activity at low overpotentials, long-term stability and durability.