Reticular-Chemistry-Inspired Supramolecule Design as a Tool to Achieve Efficient Photocatalysts for CO₂ Reduction

Abstract

Photocatalytic CO₂ reduction into C1 products is one of the most trending research subjects of current times as sustainable energy generation is the utmost need of the hour. In this review, we have tried to comprehensively summarize the potential of supramolecule-based photocatalysts for CO₂ reduction into C1 compounds. At the outset, we have thrown light on the inert nature of gaseous CO₂ and the various challenges researchers are facing in its reduction. The evolution of photocatalysts used for CO₂ reduction, from heterogeneous catalysis to supramolecule-based molecular catalysis, and subsequent semiconductor–supramolecule hybrid catalysis has been thoroughly discussed. Since CO₂ is thermodynamically a very stable molecule, a huge reduction potential is required to undergo its one- or multielectron reduction. For this reason, various supramolecule photocatalysts were designed involving a photosensitizer unit and a catalyst unit connected by a linker. Later on, solid semiconductor support was also introduced in this supramolecule system to achieve enhanced durability, structural compactness, enhanced charge mobility, and extra overpotential for CO₂ reduction. Reticular chemistry is seen to play a pivotal role as it allows bringing all of the positive features together from various components of this hybrid semiconductor–supramolecule photocatalyst system. Thus, here in this review, we have discussed the selection and role of various components, viz. the photosensitizer component, the catalyst component, the linker, the semiconductor support, the anchoring ligands, and the peripheral ligands for the design of highly performing CO₂ reduction photocatalysts. The selection and role of various sacrificial electron donors have also been highlighted. This review is aimed to help researchers reach an understanding that may translate into the development of excellent CO₂ reduction photocatalysts that are operational under visible light and possess superior activity, efficiency, and selectivity.

1. Introduction

The incessant demand for energy, unabated deterioration of fossil fuel reserves, and global warming are three serious challenges mankind is facing at present.¹ It is believed that the continuous release of CO₂ into the atmosphere through anthropogenic activities, generally by burning of fossil fuels, affects the global temperature with drastic consequences.² In this direction, the sequestration and subsequent chemical processing of CO₂ have become a global focus so as to reduce its emission and accumulation. Various methods have been tested by researchers so far, which include: (1) sequestration and storage of CO₂ under the earth and ocean;^3,4 (2) chemical processing of CO₂, and (3) biological fixation of CO₂.^5,6

However, there are various limitations associated with the above-mentioned procedures concerning viable applications. Geological dumping is not considered an efficient method as it may lead to wastage of resources due to leakages of CO₂ gas. CO₂ dissolved in oceans may acidify the waters, which will affect the marine ecosystem. Further, the mineralization of CO₂ could be very costly and a potential hazard to the environment.^3,4 Biological CO₂ fixation is a feasible method, but the uneven distribution of the global forest area and deforestation are limiting factors.

Hence, the current focus is mainly on the research related to the chemical processing of CO₂. However, the thermodynamic stability and limited reactivity of CO₂ make its sequestration and chemical processing a daunting task. CO₂ has a very high standard Gibbs free energy of −394.39 kJ mol^–1,⁴ indicating that it is a kind of inert molecule. Thus, activating this molecule requires overcoming the kinetic inertia and the thermodynamics energy barrier. Scientists have applied high temperature and^7,8 high pressure^9,10 and employed various catalysts,^11,12 including coordinating activation,^13,14 acid–base synergism-based activation,^15,16 photoelectrochemical activation,¹⁷⁻²⁰ enzymatic activation,²¹ and plasma-based activation.²² However, these procedures did not produce viable results. More recently, photocatalytic CO₂ reduction is being foreseen as the most effective method for CO₂ reduction due to the easy availability of sunlight, ambient reaction conditions, and simpler processing.^4,23,24 However, there are various obstacles to the development of optimized systems for photocatalytic reduction of CO₂ into C1 compounds, which will be discussed below in detail.

1.1. Bottlenecks

Photocatalytic reactions are determined in terms of three phenomena that actually control the economic viability of the reaction, i.e., efficiency, activity, and selectivity. The interplay among these three plays a pivotal role in photocatalytic CO₂ reduction into value-added carbon products (Figure 1). The efficiency depends on the catalyst’s optical and electronic properties, and it determines the energy cost of the reaction. While optical properties could be optimized by band-gap engineering, charge carrier mobility is an intrinsic material property. The optical property determines the quantum efficiency, while the electronic property influences the faradic efficiency. The activity is a function of the concentration of active sites, and it governs the turnover and yield of the photocatalytic reaction. Thus, the photocatalyst morphology, dimension, and surface area are pivotal to achieve the desired activity. The productivity and purity of a product are governed by selectivity, which is the function of the catalyst’s chemical nature. The selective binding efficiency and the redox properties of the photocatalyst are key in determining the selectivity of the desired product. Albeit a huge number of photocatalysts have been developed by researchers, the central bottleneck is that all three phenomena could not be combined together in one material. Nature has evolved perfect catalysts to fix CO₂ with remarkable activity and selectivity. However, the efficiency is limited since a vast source of light is applied to reduce a small concentration of CO₂ in the atmosphere at a particular time. Hence, a lot of energy goes unused.

Figure 1.

Interplay of activity, efficiency, and selectivity of photocatalysts for reduction of CO₂ into C1 compounds.

Metal and metal oxide heterogeneous catalysts are the most primitive class of photocatalyst materials synthesized.^25,26 These materials exhibit enhanced efficiency and activity subject to optimizing their morphological, optical, and electronic properties.^27,28 However, using heterogeneous inorganic catalysts, it is very difficult to control all three phenomena at the molecular level rationally.²⁹ As a step forward, homogeneous molecular catalysts came to the scene. These catalysts can be tailored toward enhanced selectivity and better activity by tailoring the organic ligands. However, the activity could be a limitation.^26,30 The main causes of the limited activity could be the catalyst deactivation and solubility issues of the catalyst. Furthermore, molecular catalysts require multiple components, which hinder the activity due to the randomness issues in the solution. Hence, if we could combine the multiple components with the desired features in an ordered manner, a thought arises that they can perform synergistically.³¹ This prospect of achieving all three aspects is made possible by exploring reticular chemistry.²⁶ Hence, researchers are trying to combine different types of components to design a photocatalyst with synergistic properties. In this direction, molecular catalysts have gained greater attention due to the wide scope of molecular engineering potential in them.²⁶

1.2. Why Supramolecules?

One of the most serious limitations of photocatalytic CO₂ reduction is that it requires a high reduction potential to activate CO₂ molecules. The photogenerated electrons should have an extra potential (overpotential) to carry on the efficient photoreduction of CO₂.³² The primary activation of CO₂ to CO₂^– requires a standard potential of −1.90V vs NHE at pH 7, which is very high (eq 1).^33,34 The required potentials (pH = 7) to achieve various solar fuels from CO₂ reduction are presented in eqs 1–8.^24,32,35

1

However, multiproton- and/or multielectron-mediated reduction of CO₂ requires a smaller potential^30,34,36 by overcoming high thermodynamic thresholds, which originate from a single-electron CO₂ reduction to form CO₂^–. It can be inferred here that methane and methanol formation from CO₂ reduction requires a smaller reduction potential. However, these reactions require more electrons. Hence, these reactions are kinetically controlled where methane and methanol production is much difficult with respect to carbon monoxide, formic acid, and formaldehyde, which require a smaller number of electrons.³⁷ The reactions that require 2–8 electrons and protons for the reduction reactions are very much difficult to occur.³²

Inorganic photocatalysts were not found to be effective for multielectron processes due to their complicated electrochemical surface and lower charge mobility. Hence, it is imperative to explore molecular and supramolecular photocatalysts with a superior visible-light absorption coefficient, generating sufficient electrons, and have a high overpotential, which could improve the reaction efficiency.³⁸ Supramolecules are seen as highly attractive photocatalysts as they exhibit good selectivity, activity, and efficiency. More importantly, the modifications in the catalyst structure can be easily engineered to suit the desired reaction through synthetic means.³⁹

2. Results and Discussions

2.1. Evolution of Supramolecule Photocatalysts

In view of the above introduction, the transition from molecular catalysis to development of supramolecule photocatalysts is discussed below.

2.1.1. Two-Component Systems

The most primitive molecular systems as photocatalysts for reduction of CO₂ are rhenium(I)-based molecular systems.^1,40−42 Lehn and co-workers (1983) presented the first report regarding the fabrication of a Re-based complex—fac-Re(bpy)(CO)₃X (bpy = 2,2′-bipyridine; X = Cl^– or Br^–)—and investigated its photocatalysis for CO₂ reduction in a mixed solution of triethanolamine (TEOA) and N,N-dimethylformamide (DMF).^43,44 They obtained a good selectivity and efficiency for CO production (Φ_CO = 0.14; X = Cl^–). However, these complexes exhibit stability issues, resulting in smaller turnover numbers (TON_CO = 27; X = Cl^–) and small CO production, and they also possess a weak visible-light absorption coefficient. To address these limitations, mechanistic investigations were performed using laser flash photolysis techniques.⁴⁵⁻⁴⁷ It was observed that the lowest excited state of the rhenium(I) complex, i.e., the triplet metal-to-ligand-charge-transfer state (³MLCT), undergoes reductive quenching by TEOA without undergoing oxidative quenching by CO₂. This leads to the generation of one electron-reduced species (OERS) of the rhenium(I) complex. Since the removal of the Cl^– ligand associated with the OERS is a delayed process, a 17-electron species is being observed. Fujita et al. reported that [Re(4,4′-Me₂bpy)(CO)₃] can take up CO₂ despite its reduced and coordinatively unsaturated nature. However, this reaction was found to be very slow as the extra electron got mainly accumulated on the bpy ligand moiety, represented as Re^I(4,4′- Me₂bpy^•–)(CO)₃. This electronic structure must be more feasible with respect to Re⁰(4,4′ Me₂bpy)(CO)₃. Hence, it was inferred that the slow formation of the former might be the reason for the smaller activity of rhenium(I) complexes for the photocatalytic CO₂ reduction.⁴⁸ Ishitani and co-workers reported that using only one rhenium(I) complex inhibits the activity. They cite the reason that the OERS generated in the process must involve two opposite characteristics: one is the stability for behaving as an electron donor, and another is the instability for producing the “CO₂ adduct”.

To address these limitations, Ishitani and co-workers proposed the two-component-based supramolecular photocatalytic systems. These photochemical systems comprise a redox photosensitizer complex that actually facilitates the one-electron transfer and a catalyst component that modulates the one-electron transfer to the multielectron reduction of CO₂, requiring a smaller negative reduction potential. Ishitani and co-workers used the fac-[Re{4,4′-(MeO)₂bpy}-(CO)₃{P(OEt)₃}]⁺ complex as the photosensitizer component and fac-[Re(bpy)(CO)₃(MeCN)]⁺ was taken as the catalyst component. The OERS generated by the photosensitizer complex was found to be highly durable as its P(OEt)₃ ligands exhibited an enhanced π-accepting property and high reduction power owing to the presence of electron-releasing methoxy groups. They found that a ratio of 25:1 photosensitizer to catalyst exhibited a significant CO₂ reduction efficiency with respect to the single-component systems. A quantum efficiency of Φ_CO = 0.59 was obtained while irradiating at λ_ex = 365 nm.⁴⁹ A more efficient ring-shaped rhenium(I) trinuclear complex was developed by these researchers (Figure 2), in which the bidentate phosphine ligands were used to connect between Re components. This rhenium ring exhibits good properties as a redox photosensitizer in terms of enhanced visible-light absorption, lifetime of the excited state, and stability of the OERS. These properties were attributed to the effective π–π interaction among the phenyl moieties on phosphine ligands and the diimine.¹

Figure 2.

Ring-shaped rhenium(I) trinuclear complex where the constituent complexes are connected by bidentate phosphine ligands; reprinted with permission from ref (1). Copyright [2015] [American Chemical Society].

Facilitating an effective electron transfer between the photosensitizer component and the catalyst component requires collisions among the OERS of the photosensitizer unit and the catalyst unit during the process.^50,51 To address this limitation, researchers thought of a bridging ligand that could connect the photosensitizer component to the catalyst component. This idea led to the supramolecular photocatalyst systems, where different functions are performed by different component units in one molecule.⁵² These supramolecular photocatalysts are considered more feasible systems as the OERS species of the photosensitizer and the catalyst component do not require to collide as being bridged to each other.¹

2.2. Selection of Photosensitizers

The selection of photosensitizers in supramolecule photocatalysts is a very important issue. Two modes of reaction mechanisms are possible for the photoreduction of CO₂. One is the reductive quenching (RQ) and the other is an oxidative quenching (OQ) mechanism. The OQ mechanism involves the transfer of an electron from the excited state of the photosensitizer (*PS) to the catalyst unit producing an OEOS pertaining to the photosensitizer component (PS⁺) and an OERS on the catalyst component (Cat^–). Subsequently, the sacrificial electron donor (D) donates an electron to the PS⁺. On the other hand, in the case of the RQ mechanism, a sacrificial electron donor first reduces the excited photosensitizer *PS, producing an OERS (PS^–), which then transfers the electron to the catalyst unit where the reduction of CO₂ finally takes place.

9

The OQ mechanism has the disadvantage that there is the possibility of a simultaneous fast intramolecular backward electron transfer where the oxidized photosensitizer unit may readily take the electron back. Therefore, the RQ mechanism has been responsible for the enhanced activity in various high-performance supramolecular photocatalysts.⁵³ The photosensitizers should have a larger light absorption coefficient with respect to the reductant and the catalyst component for efficient CO₂ reduction. The photosensitizer component should also have a larger emission lifetime, which provides ample time for an efficient reductive quenching process. The photosensitizer in its excited state should have a good oxidation capability so as to accept an electron from the reductant, and the OERS produced after the electron transfer must have an enhanced stability.⁵⁴ For efficient electron transfer, the reduction potential of PS* should be comparable to or higher than the oxidation potential of the sacrificial donor/reductant.^54,55

In view of the above discussion, researchers have investigated several classes of photosensitizers to optimize the activity, efficiency, and selectivity. Here, we will mention a few of the most commonly used photosensitizers. The selection of photosensitizer complexes and the challenges thereof have been discussed in Table 1.

Table 1. List of Common Photosensitizers that Have Been Employed with Various Catalyst Components for Photocatalytic CO₂ Reduction^k.

Solvents: a = Acetonitrile.

b = DMF.

c = Dichloromethane.

d = Methanol.

e = Toluene.

f = Acetonitrile/dioxane (1:1 (v/v)).

g = Tetrahydrofuran.

h = Benzene.

i = E_p (irreversible wave).

^jP-P = Diphenylphosphinoethane.

^kThe table has been reproduced with permission from ref (54). Copyright [2015] [Elsevier. B. V.].

2.2.1. Ru(II) Diimine Complexes

Photosensitizer complexes based on ruthenium are the most widely used in photocatalytic CO₂ reduction reactions. These types of complexes, viz. [Ru^II(N^∧N)₃]ⁿ⁺, containing a diimine ligand like 2,2′-bipyridine (bpy), exhibit powerful visible-light absorption owing to the singlet metal-to-ligand charge transfer (¹MLCT) band of absorption.^56,57 On excitation of the photosensitizer, a triplet excited-state ³MLCT is produced involving the rapid and efficient intersystem crossing.⁵⁸ These Ru(II) complexes are photochemically stable and have a relatively longer lifetime of ³MLCT. However, rapid ligand substitution may occur under acidic pH. Albeit various [Ru^II(N^∧N)₃]ⁿ⁺ kinds of complexes produce stable OERS when in the dark, it may undergo decomposition on photoexcitation, producing [Ru(N^∧N)₂L₂]^m+ types of complexes, where L may be the solvent containing Cl^– or CO.^55,59−61

2.2.2. Re(I) Carbonyl Diimine Complexes

Like Ru(II) complexes, rhenium-based photosensitizer complexes exhibit relatively enhanced visible-light absorption, longer excited-state lifetime (τ_em = 5.4 μs), and enhanced OERS stability owing to the effective π–π overlap between the diimine and phosphine ligands containing the phenyl moieties. Among the various complexes of this kind, fac-[Re(N^∧N)(CO)₃(PR₃)]⁺ type of complexes has been considered as superior photosensitizers.^49,62−64 The OERS generated with these photosensitizers possesses a very high quantum yield (fac-[Re{4,4′-(MeO)₂bpy}(CO)₃{P(OEt)₃}]⁺ exhibited Φ_OERS = 1.6 when triethanolamine was used as the reductant).⁴⁹ Similarly, complexes like [Re(N^∧N)(CO)₂(PAr₃)₂]⁺ have been reported as suitable photosensitizers owing to their longer-wavelength ¹MLCT absorption band, unlike complexes with nonaromatic ligands.⁶⁵⁻⁶⁸ This leads to enhanced emission lifetimes with enhanced oxidation capability of the ³MLCT excited states.

2.2.3. Os(II) Diimine Complexes

In view of the various limitations of Ru-based photosensitizers, like limited photostability and a narrow band of absorption, researchers are looking for alternative photosensitizer complexes. Os(II) diimine complexes, viz. [Os^II(N^∧N)₃]ⁿ⁺, possess enhanced visible-light absorption up to longer wavelengths (∼700 nm).^69,70 This property originates as Os-based photosensitizer complexes do not allow a singlet to triplet (³MLCT) absorption band around 500–700 nm due to the larger heavy-atom effect. Although these complexes exhibit higher negative reduction potentials in their excited states, the shorter emission lifetimes pose a limitation requiring strong reductants in the photocatalytic CO₂ reduction processes.^24,54

2.2.4. Ir(III) Complexes

Cyclometalated Ir(III) complexes were also explored as photosensitizers for CO₂ reduction photocatalysts although they exhibit poor absorption in the lower wavelength area of the visible-light spectrum. One example is the fac-[Ir(ppy)₃] complex, where ppy = 2-phenylpyridine,^50,71−76 which was exploited owing to its good photostability and the enhanced oxidation power possessed by its excited states. These properties exhibited by the [Ir(ppy)₂(bpy)]⁺ and Ir(ppy)₃ complexes were attributed to their large ligand field stabilization by ppy ligands due to the significant σ-donor ability, which leads to the high photochemical stability and a long lifetime of the excited states in these complexes.⁷⁷⁻⁷⁹ For [Ir(C^∧N)₂(N^∧N)]⁺, where C^∧N = cyclometalated ligands, viz. ppy, the HOMOs comprise the d orbitals associated with the Ir atom and the π-orbitals are located on the cyclometalated ligands. The π* orbitals of the N^∧N ligands comprise the LUMOs of these types of Ir complexes. The LUMO obtained from the π* orbitals of the C^∧N ligands results in larger lifetimes in their excited states and hence the reduction power.⁵⁴ However, the relatively weak absorption in visible light is a limitation. Kuramochi and co-workers⁷⁹ used a 1-phenylisoquinoline (piq)-based Ir complex, [Ir(piq)₂(dmb)]⁺, where dmb = 4,4′-dimethyl-2,2′-bipyridine (Figure 3 right).⁸⁰ This complex exhibits a relatively larger absorption at wavelengths higher than 450 nm and also shows a longer lifetime of the excited states with respect to that of [Ru(dmb)₃]²⁺.²⁴

Figure 3.

Structures and abbreviations of the iridium(III) complexes; reprinted with permission from ref (79). Copyright [2016] [American Chemical Society].

Chang and co-workers (2013) demonstrated that an Ir-based photosensitizer could introduce a high selectivity and superior activity in nickel complexes containing N-heterocyclic carbene-amine as ligands for the conversion of CO₂ to CO and reached a TON of up to 98 000.^50,81

2.2.5. Chlorophylls and Metal Porphyrins

Chlorophylls^82,83 and metal porphyrins⁸⁴⁻⁸⁷ are the two large classes of photosensitizers that include the naturally occurring chlorophylls. Such complexes exhibit two strong visible-light absorption bands. The first is the Soret band, which involves the transition to the S2 excited state, ε_{(Soret band)} > 105 M^–1 cm^–1, while the Q-band involves the transition to the S1 excited state, ε_(Q band) > 104 M^–1 cm^–1. Pd^II-based porphyrin complexes are characterized by enhanced excited-state lifetimes owing to the rapid intersystem crossing and hence act as efficient photosensitizers like Ru diimine complexes.⁸⁶ Although Zn²⁺-based porphyrins exhibit quite short emission lifetimes, the internal conversion taking place from the S2 excited state to the S1 excited state is a delayed process,⁸⁸ which enhances their reduction power. Hence, Zn-porphyrins are also under wide focus as photosensitizers for photocatalytic CO₂ reduction.

2.2.6. Organic Photosensitizers

Various organic compounds that contain superior visible-light absorption and also exhibit relatively long lifetimes of their excited states were also developed as photosensitizers. For example, phenazine (Phen) has a very stable triplet excited state, and triethylamine (TEA) can reductively quench it to give Phen^•–.⁸⁹ The as-generated Phen^•– may react with a proton to produce PhenH^•, which can eventually provide a hydride toward the catalyst component, such as a Co-cyclam-based complex.⁹⁰ Likewise, oligo(p-phenylenes) (OPPs) can yield OPP^•– on getting reductively quenched by TEA.⁹¹ OPP^•– exhibits a very strong reduction power (E_1/2 = −2.45 V vs SCE), which is very favorable for the photoreduction of CO₂.⁵⁴

A copper purpurin-based complex, which exhibits a reduction potential more negative by 540 mV than its organic dye moiety, was reported by Yuan and co-workers.⁹² They employed a copper purpurin photosensitizer along with an iron porphyrin complex as the catalyst component. They achieved a remarkable TON of over 16 100 CO₂ to CO conversion with high selectivity (95%) using 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole as the sacrificial reductant. This is the best performance reported for noble-metal-free homogeneous systems.⁹²

Yamashita and co-workers⁹³ developed a photocatalyst for CO₂ reduction by intercalating the [Re(bpy)(CO)₃Cl] complex into the layers of Zr(HPO₄)₂·H₂O (a-ZrP), and rhodamine B was used as the photosensitizer, as shown in Figure 4. Rhodamine B (RhB) is a stable and very fluorescent molecule and has a strong absorption in the visible range of the solar spectrum.⁹⁴ Hence, it can be used as a promising photosensitizer owing to the ease of electron transfer from its LUMO to the catalyst component.

Figure 4.

(a) Interlayer structures of a-ZrP (Zr(HPO₄)₂·H₂O) complexes, (b) rhodamine B molecule, and (c) [Re(bpy)(CO)₃Cl] complex. Reprinted with permission from ref (93). Copyright [2015] [Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim].

The attractive performance of RhB as a metal-free photosensitizer has opened the doors for investigating various organic-molecule-based photosensitizers for photocatalytic CO₂ reduction systems.⁹³

2.3. Supramolecules for CO₂ Reduction

Ishitani and co-workers demonstrated the photocatalytic CO₂ reduction properties of supramolecule photocatalysts on the basis of the following traits: (1) selectivity (Γ); (2) quantum yield (Φ); (3) turnover number (TON); and (4) turnover frequency (TOF).⁵³ In view of the various developments in terms of the above traits, different classes of supramolecular photocatalysts were designed for CO₂ reduction.

The various supramolecule photocatalyst systems for CO₂ reduction have been discussed next.

2.3.1. Ru(II)–Ni(II)- and Ru(II)–Co(III)-Based Systems

Ruthenium-based supramolecule photocatalysts with non-noble metals were very scarce until the 1990s. Kimura and co-workers (1992) reported the first such bimetallic system for photocatalytic CO₂ reduction.¹⁰² This supramolecule photocatalyst consisted of a [Ru(phen)₃]²⁺ component as the photosensitizer, where phen = 1,10-phenanthroline, and a [Ni(cyclam)]²⁺ complex as the catalyst component, where cyclam = 1,4,8,11-tetraazacyclotetradecane. In the presence of ascorbic acid as a reductant, the photocatalyst showed good selectivity (Γ_CO = 72%) for CO generation with respect to H₂ evolution. Tamaki and Ishitani (2017)⁵³ summarized the photocatalytic activities exhibited by these photocatalysts, as shown in Table 2. When mixed systems were used, the activity was far less (Table 2: entry 2; Γ_CO = 36%). However, S1 (Chart 1) is not considered an efficient photocatalyst as its TON for CO generation is below unity. Also, while using a pyridinium cation as a bridging unit between the photosensitizer unit and the catalyst unit (Chart 1; S2), improvement in the photocatalytic performance was again not achieved (Table 2: entry 3; TON_CO < 1), and the TON = 2 was even much smaller than obtained from the mixed systems involving the [Ru(bpy)₃]²⁺ moiety, where bpy = 2,2′-bipyridine and the [Ni(6-((N-benzylpyridin-4-yl)methyl)-1,4,8,11-tetraazacyclotetradecane)]³⁺ complex (Table 2: entry 4).⁹⁵⁻⁹⁷ Ni-based supramolecule complexes are hence not considered good photocatalysts for CO₂ reduction owing to their weak absorption in the higher wavelength visible spectrum and decreased excited-state lifetimes of the ruthenium photosensitizer component.⁹⁹⁻¹⁰³ The various chemical structures and designations of the Ru(II)–Ni(II) supramolecules have been presented in the chart below.

Table 2. Summary of the Photocatalytic Performance by Ru(II)–Ni(II)- and Ru(II)–Co(III)-Based Systems^a.

entry	photocatalyst	reductant	product formed	selectivity (%)	TON	reference
1	S1	ascorbic acid	CO	72	<1	(102)
2	[Ru(Phen)3]2+ + [Ni(cyclam)]2+	ascorbic acid	CO	36	<1	(102)
3	S2	ascorbic acid	CO	11	<1	(95)
4	[Ru(bpy)3]2+ + Ni moiety (as in S2)	ascorbic acid	CO	09	2	(95)
5	S3	TEOA	CO	50	2	(105)
6	[Ru(bpy)3]2+ + [Ni(bpy)3]2+	TEOA	CO	79	2	(105)
7	S4	TEOA	CO	73	3	(105)
8	S5	TEOA	CO	79	5	(105)
9	[Ru(bpy)3]2+ + [Co(bpy)3]3+	TEOA	CO	35	9	(105)

^aReproduced with permission from ref (53). Copyright [2017] [American Chemical Society].

Chart 1. Ru(II)–Ni(II) and Ru(II)–Co(III)-Based Supramolecule Complexes: Various Chemical Structures and Designations.

Komatsuzaki and co-workers (1999) also reported the synthesis of supramolecule photocatalysts, where the [Ru(bpy)₂(phen)]²⁺ component was taken as a photosensitizer and a Co(III) or Ni(II) polypyridyl component was employed as a catalyst component.¹⁰⁴ Such bimetallic complexes generated both CO and H₂ when irradiated under 400–750 nm using TEOA as a sacrificial electron donor. RuNi₃ (Table 2: entry 5) generated an equal amount of CO and H₂ with the turnover number TON_CO = TON_H2 = 2. On the contrary, the mixed system involving [Ru(bpy)₃]²⁺ as a photosensitizer and [Ni(bpy)₃]²⁺ as the catalyst component selectively evolves CO (Table 2: entry 6; TON_CO = 2; TON_H2 < 1).^53,54

Further, the bimetallic supramolecules based on Ru(II)–Co(III), viz. S4 (Table 2: entry 7; TON_CO = 3; TON_H2 = 1) and S5 (Table 2: entry 8; TON_CO = 5; TON_H2 = 1), exhibited a smaller activity with respect to a mixed system involving their constituents, viz. the [Ru(bpy)₃]²⁺ complex and the [Co(bpy)₃]³⁺ complex (Table 2: entry 9; TON_CO = 9; TON_H2 = 16).⁵³

Wang and co-workers (2017)¹⁰⁶ reported new dinuclear Ru–Co complexes for CO₂ photoreduction (Chart 2).⁴⁴ The Ru–Co complexes (Chart 2; S6–S10) are considered the most efficient non-noble-metal-based supramolecular systems for photocatalytic CO₂ reduction into both CO and H₂ reported so far. The complexes designated S6, S8, and S10 in Chart 2 actually exhibited good results pertaining to their TOFs and stability.^106,107

Chart 2. Dinuclear Complexes of Ru and Co.

Apart from the above systems, Fabry and co-workers (2020) reported the Ru(II)–Mn(I) supramolecular photocatalysts for the predominant production of HCOOH by photocatalytic reduction of CO₂. [MnBr(CO)₃(BL)] was taken as the catalyst component, and [Ru(dmb)₂(BL)]²⁺ was employed as the photosensitizer, where dmb = 4,4′-dimethyl-2,2′-bipyridine and BL is the bridging ligand. An equimolar ratio of the redox photosensitizer and catalyst components generated a supramolecule, which predominantly produced HCOOH, while the corresponding mixed system of individual units had a far lower activity.¹⁰⁸

2.3.2. Ru(II)–Re(I) Systems

Ishitani and co-workers reported various Ru(II)–Re(I)-based complexes for photocatalytic CO₂ reduction, which are presented in Chart 3. The photocatalytic activities achieved by such systems have been presented in Table 3.⁵³ The foremost successful Ru–Re supramolecule photocatalyst reported by Ishitani and co-workers is S12.⁵² This complex comprised a [Ru(N^∧N)₃]²⁺-based photosensitizer component and a fac-Re(N^∧N)(CO)₃Cl complex as a catalyst component.^43,44 The two units are bridged by a −CH₂CH(OH)CH₂– chain through two 4-methyl-bpy moieties. S12 exhibited high stability, selectivity, and efficiency (Table 3: entry 10; Φ_CO = 0.12; TON_CO = 170) for photocatalytic CO₂ reduction when visible light was used in the presence of a sacrificial electron donor called 1-benzyl-1,4-dihydronicotinamide (BNAH). These results were much better against a 1:1 mixture of the constituent mononuclear complexes (Table 3: entry 11; Φ_CO = 0.062; TON_CO = 101). Contrary to this, S13 (Table 3: entry 12) and S14 (Table 3: entry 13), involving bpy or ((CF₃)₂bpy), i.e., 4,4-bis(trifluoromethyl)-2,2′-bipyridine on the Ru component as the peripheral ligand in place of ⁴dmb, exhibit smaller photocatalytic performance with respect to S12 and even smaller when compared with the mixed systems involving [Ru(⁴dmb)₃]²⁺ and fac-Re(⁴dmb)(CO)₃Cl.

Chart 3. Various Chemical Structures and Designations of Ru(II)–Re(I)-Based Supramolecule Complexes.

Table 3. Summary of the Photocatalytic Performance by Various Ru(II)–Re(I)-Based Supramolecules Involving One or More Photosensitizers and/or Catalyst Components^a.

entry	photocatalyst	reductant	product	selectivity (%)	TON	reference
10	S12	BNAH	CO	96	170	(52)
11	[Ru(4dmb)3]2+ + fac-Re(4dmb)(CO)3Cl	BNAH	CO/HCOOH	changes from CO to HCOOH	101	(52, 79)
12	S13	BNAH	CO		50	(52)
13	S14	BNAH	CO		03	(52)
14	S15	BNAH	CO		14	(52)
15	S16	BNAH	CO		28	(52)
16	S17	BNAH	CO		240	(52)
17	S18	BNAH	CO	97	232	(111)
18	S19	BNAH	CO	96	97	(111)
19	S20	BNAH	CO	97	180	(79, 112)
20	S21	BNAH	CO		120	(112)
21	S22	BNAH	CO		120	(112)
22	S23	BNAH	CO	94	207	(98)
23		BIH	CO	>99	3029	(113)
24	S24	BNAH	CO	91	144	(98)
25	S25	BNAH	CO	73	27	(98)
26	S26	BNAH	CO	98	233	(53)
27		BIH	CO	>99	2915	(53)
28	S27	BNAH	CO	95	253	(114)
29	S28	BNAH	CO			(114)
30		BIH	CO	>99	>1000	(115)
31	S29	BNAH	CO		123	(116)
32	S30	BNAH	CO		204	(116)
33	S31	sodium ascorbate	HCOOH	81	25	(117)
34	S32	BI(CO2H)H	CO	81	130	(118)
35	S33	BNAH	CO		315	(109)
36	S34	BNAH	CO		50	(109)
37	S35	BNAH	CO		110	(110)
38	S36	BNAH	CO		190	(110)
39	S37	BNAH	CO	69	283	(119)
40	S38	BNAH	CO	75	313	(119)

^aReproduced with permission from ref (53). Copyright [2017] [American Chemical Society].

The lower performance was attributed to the fact that the intramolecular electron transfer from the OERS of the Ru component to the Re component was ineffective due to its endergonic nature (e.g., S14 has E^1/2 red (Ru) = −1.23 V and E^1/2 red (Re) = −1.76 V vs Ag/AgNO₃). However, S12 involved the reduction of Ru and Re components at almost equivalent potential (E^1/2 red = −1.77 V) using one electron.

Various photocatalysts involving more than one photosensitizer and/or catalyst components have been introduced by Ishitani et al. (Table 3: S17),⁵² Rieger et al. (Table 3: S33),¹⁰⁹ and Furue et al. (Table 3: S35, S36).¹¹⁰ The supramolecule photocatalysts containing more than one catalyst unit, viz. S17 (Table 3: entry 16; TON_CO = 240), S33 (Table 3: entry 35; TON_CO = 315), and S36 (Table 3: entry 38; TON_CO = 190), presented a better photocatalytic performance with respect to S12 (Table 3: entry 10; TON_CO = 170). On the contrary, S35, containing two photosensitizer components and one catalyst component, exhibited a lower performance (Table 3: entry 37; TON_CO = 110) against S12.

Ishitani and co-workers investigated that the OERS of the photosensitizer, [Ru(⁴dmb)₃]²⁺, undergoes photochemical degradation, while the Re(I)-based catalyst component is quite stable and retains its original chemical structure when applied for a photocatalytic reduction reaction (Scheme 1).¹²⁰

Scheme 1. Solvent-Based Deactivation of the [Ru(⁴dmb)₃]²⁺ Photosensitizer on Accepting the Electron from the Sacrificial Reductant.

It is interesting to observe that the product obtained on decomposition, viz. [Ru-(⁴dmb)₂(solvent)₂]²⁺, acts as a catalyst component and undergoes a CO₂ reduction reaction along with the [Ru(⁴dmb)₃]²⁺ photosensitizer component to evolve HCOOH as the major product.^53,79

Stanley and co-workers (2021) tried to address the stability issue by embedding these molecular systems in a metal−organic framework (MOF). They chose a benchmark catalyst component, fac-ReBr(CO)₃(4,4′-dcbpy) (dcbpy = dicarboxy-2,2′-bipyridine), and photosensitizer Ru(bpy)₂(5,5′-dcbpy)Cl₂ (bpy = 2,2′-bipyridine) and entrapped them inside the framework of an inexpensive and nontoxic MOF, MIL-101-NH₂(Al), via noncovalent host–guest interactions. The photocatalyst exhibited enhanced stability for photocatalytic CO₂ reduction, and selective evolution of CO was achieved for 1.5–40 h, which was a much better performance with respect to its free molecular component systems.¹²¹

Cancelliere and co-workers (2020)¹²² developed a novel tris-chelating polypyridine ligand, which was employed as a bridging ligand for producing different supramolecular photocatalysts. The central bridging ligand is a phenylene ring containing three ethylene chains at its 1, 3, and 5 positions, which connect to the 2,2′-bipyridine moieties of the Ru photosensitizer or Re catalyst units (Chart 4). As such, various combinations of photosensitizer and catalyst components were taken to develop various supramolecule photocatalysts. These trinuclear systems could selectively produce CO (up to 97%) by photocatalytic reduction of CO₂ with good efficiency while BIH was used as a reductant. They exhibited excellent durability with the largest TONs for CO₂ reduction employing these supramolecular photocatalysts in homogeneous solutions.¹²²

Chart 4. Various Chemical Structures and Designations of Ru(II)–Re(I)-Based Supramolecule Complexes Having a Central Phenylene Ring Bridging Ligand Containing Three Ethylene Chains at its 1, 3, and 5 Positions, which Are Connected to the 2,2′-Bipyridine Moieties of the Ru Photosensitizer or Re Catalyst Units¹²².

Ishitani and co-workers (2016) reported new Ru(II)–Re(I)-based supramolecule photocatalysts involving cyclic bridging ligands (Chart 5). It was expected that significant electronic interaction between the Ru and Re components would be acquired as the two diimine moieties are connected by two ethylene chains. However, the number of bridging ligands and the extended conjugation were found to determine the extent of electronic interaction.

Chart 5. Structure and Abbreviations of Various Bridging Ligands and a Representative Ru(II)–Re(I)-Based Supramolecule Photocatalyst Based on These Ligands.

Ohkubo and co-workers reported the synthesis of two new Ru(II)–Re(I)-based supramolecular photocatalysts, viz. [Ru(BL1)Re(CO)₃Cl]²⁺ and [Ru(BL1)Re(CO)₂{P(p-F-C₆H₄)₃}₂]³⁺, and their photocatalytic activities were compared with respect to the BL2 containing supramolecular photocatalysts, viz. [Ru(BL2)Re(CO)₃Cl]²⁺ and [Ru(BL2) Re(CO)₂{P(p-F-C₆H₄)₃}₂]³⁺. It was found that CO was the main product of CO₂ reduction.¹¹⁶

A new biscarbonyl-complex-based supramolecule photocatalyst was reported by Tamaki and Ishitani et al., involving cis,trans-[Re(N^∧N)(CO)₂(PR₃)₂]⁺ (where R = p-Fph, Ph, OEt) as a catalyst component. The authors suggest that these supramolecules with tri(p-fluorophenyl)phosphine ligands, Ru–Re(FPh), exhibited high efficiency, selectivity, and durability (Chart 6).^65,66,98

Chart 6. Ru(II)–Re(I)-Based Supramolecule Complexes with Various Substituents on the Ligands.

2.3.3. Ir(III)–Re(I)-Based Systems

The representative structure and designation of the Ir(III)–Re(I)-based complex are presented in Chart 7.

Chart 7. Representative Structure of the Ir(III)–Re(I)-Based Complex.

When two 1-phenylisoquinoline (piq) ligands are used as a bridge between the Ir(III)-based photosensitizer component and the Re-based catalyst component, an enhanced light absorption below 560 nm was obtained with better emission lifetime (τ_em = 3.0 μs). CO was the main product obtained by it on reduction of CO₂ using BNAH as a sacrificial electron donor with selectivity Γ_CO > 99%.⁷⁹ This photocatalyst also exhibited superior stability and efficiency (Φ_CO = 0.21, TON_CO = 130).⁵³

The photosensitizer can also play a dual role in some cases. It can act as a light-harvesting unit and also a catalytic site for CO₂ reduction. The complexes of the kind, viz. fac-[Re(bpy)(CO)₃Br]⁴³ and [Ir(tpy)(bpy)Cl]¹²³ (where tpy= 2,2′:6′,2″-terpyridine), were developed by the Lehn group and the Ishitani group, respectively. The supramolecules so obtained have been explored for the selective photocatalytic CO₂ reduction to CO achieving TONs close to 50. Subsequently, Beller and co-workers⁸¹ used [Ir(ppy)₂(bpy)]PF₆, where ppy = 2-(pyridine-2-yl)benzene-1-ide, as the photosensitizer to develop a new photocatalyst for the selective photocatalytic reduction of CO₂ to obtain HCOOH.^124−128 These supramolecule photocatalyst systems exhibit good visible-light absorption and long lifetimes of their excited states.^81,129,130

2.3.4. Os(II)–Re(I) Systems

Os(II)-based complexes, viz. [Os(N^∧N)₃]²⁺, act as efficient photosensitizers owing to the strong visible-light absorption up to λ_abs < 730 nm, and the reducing power of their OERS was found to be equivalent to that of [Ru(N^∧N)₃]²⁺-based photosensitizers. For example, a photocatalyst of this kind was developed when the [(⁵dmb)₂Os(⁴dmb)]²⁺ complex (where ⁵dmb = 5,5′-dimethyl-2,2′- bipyridine) was bridged to a cis,trans-[Re(N^∧N)- (CO)₂{P(p-X-C₆H₄)₃}₂]⁺ catalyst component (where X = F, Cl) through an ethylene chain. This photocatalyst was found to selectively produce CO when exploited in a photocatalytic CO₂ reduction reaction, using BIH as a sacrificial reductant and irradiation at λ_ex > 620 nm (Chart 8).⁶⁹

Chart 8. Os(II)–Re(I)-Based Supramolecule Complexes with Various Substituents on Their Ligands.

It has been reported that the photocatalytic performance gets influenced by the introduction of phosphine ligands in the Re component, and addition of Cl on the phosphine moiety as in S48 showed a better photocatalytic performance (Φ_CO = 0.12, TON_CO = 1138, TOF_CO = 3.3 min^–1) as compared to that of S47, which contains the F substituent in the phosphine moiety (Φ_CO = 0.10, TON_CO = 762, TOF_CO = 1.6 min^–1). The relatively high photocatalytic activity was also observed with S49, where the catalyst component contains fac-[Re(N^∧N)(CO)₃(PPh₃)]⁺(TON_CO = 973).⁵³

Ishitani and co-workers¹³¹ reported for the first time a trinuclear complex based on the Os(II)–Re(I)–Ru(II) supramolecular photocatalyst (Chart 9). These complexes show superior absorption in a wide range of visible light and act as stable photocatalysts for CO₂ reduction. These supramolecule photocatalysts showed the highest TON for CO generation (TON_CO = 4347) among the various Re(I) catalyst component-based supramolecular photocatalysts.¹³¹

Chart 9. Trinuclear Os(II)–Re(I)–Ru(II) Supramolecular Complexes and Constituent Fragments.

2.3.5. Binuclear Ru(II)-Based Systems

The various chemical structures and designations of binuclear Ru(II) supramolecule complexes are presented in Chart 10. Tamaki and Ishitani (2017) summarized their photocatalytic reduction performances as given in Table 4.⁵³ Tanaka and co-workers originally reported that [Ru(N^∧N)₂(CO)₂]²⁺-based supramolecule photocatalysts generate HCOOH with good selectivity under basic pH.^132−134

Chart 10. Ru(II)–Ru(II)-Based Photocatalyst Systems.

Table 4. Summary of Performance by Ru(II)–Ru(II) Systems^a.

entry	photocatalyst	reductant	product	selectivity (%)	TON	reference
41	S52	BNAH	HCOOH	90	315	(135)
42	S53	BNAH	HCOOH	91	562	(135)
43	S53	MeO-BNAH	HCOOH	89	671	(135)
44	S53	BI(OH)H	HCOOH	87	2766	(136)
45	S53	BIH	HCOOH	72	641	(136)
46	S54	BNAH	HCOOH	77	353	(135)
47	S55	BNAH	HCOOH	70	234	(135)
48	S56	BNAH	HCOOH	91	337	(53)
49	S57	BNAH	CO	98	40	(53)

^aReproduced with permission from ref (53). Copyright [2017] [American Chemical Society].

Various other Ru-complex-based systems, viz. [Ru(⁴dmb)₃–m(BL)_m]²⁺, where m = 1, 2, 3 and BL = 1,2-bis(4′- methyl-[2,2′]bipyridin-4-yl)-ethane, act as photosensitizers, which bridge to cis-[Ru(⁴dmb)_2–n(BL)_n(CO)₂]²⁺ (n = 1, 2) catalyst components through a nonconjugated bridging ligand. The as-developed photocatalysts could successfully reduce CO₂ to HCOOH when irradiated in the presence of BNAH.¹³⁵ The ratio of photosensitizer to catalyst components influences the photocatalytic activity. For example, S53 (Table 4: entry 42; Γ_HCOOH = 91%) and S52 (Table 4: entry 41; Γ_HCOOH = 90%) exhibited better selectivity in the photocatalytic reduction reaction, whereas S54 (Table 4: entry 46; Γ_HCOOH = 77%) and S55 (Table 4: entry 47; Γ_HCOOH = 70%) deactivated readily due to modification in the catalyst components.^137,138

When a higher number of catalyst components are used in one photocatalyst, viz. S54 and S55, it facilitates the degradation of the catalyst components as the photosensitizer is unable to transfer sufficient electrons to it. However, when the ratio of photosensitizers is increased, viz. S52 and S53, it facilitates CO₂ reduction and suppresses deactivation. Interestingly, the Ru(II)–Re(I)-based supramolecular photocatalysts exhibited higher durability when multiple catalyst components were used.^52,109,110

Various multinuclear Ru-based complexes containing cis,trans-Ru(BL)(CO)₂Cl₂^139,140 (S56 and S57) have been reported. Various ratios of photosensitizer to the catalyst component were developed. S56, having the ratio of 1:1, majorly evolved HCOOH (Table 4: entry 48; TON_HCOOH = 337; TON_CO = 22, TON_H2 = 12), whereas S57 containing three catalyst components (1:3 ratio) selectively evolved CO (Table 4: entry 49; TON_HCOOH < 1, TON_CO = 40, TON_H2 < 1). The selectivity of S57 toward CO formation and not HCOOH was attributed to the polymerization of the catalyst components during the photocatalytic reaction.^141−144 The generation of CO instead was attributed to the Ru oligomer involving the Ru photosensitizer component generated from S57 (Scheme 2).⁵³

Scheme 2. Polymerization of the Catalyst Components during the Photocatalytic Reaction.

2.3.6. Porphyrin–Re(I) and Ru(II)-Based Systems

There are fewer reports regarding dinuclear supramolecule complexes based on a porphyrin component and a Re(I) bipyridine tricarbonyl component.^{84,86,87,145−147} In most of these complexes, an aliphatic amide bond connects the bpy ligand of the Re(I) component to one phenyl group in the meso-position of the porphyrin component.^84,86,87 Many other groups have reported the use of a phenanthroline moiety to combine a porphyrin component so as to facilitate the interaction between two different metal centers;^148−150 however, their photocatalytic activity was not investigated. Tamaki and Ishtani (2017) summarized the photocatalytic performance by porphyrin–Re(I) systems as shown in Table 5.⁵³ A Zn(II)-based porphyrin photosensitizer (S58) was developed by Inoue et al., which could photocatalytically reduce CO₂ to CO using visible light for irradiation at λ_ex = 428 nm, and triethylamine (TEA) was used as the sacrificial electron donor (Table 5: entry 50; Φ_CO = 0.0064).⁸⁴ Perutz and co-workers reported the synthesis of a supramolecule photocatalyst (S59) involving a Pd(II) tetraphenylporphyrin-based photosensitizer and a fac-[Re(N^∧N)(CO)₃(3-picoline)]⁺ derivative-based catalyst component, and the two are connected through an amide linkage.⁸⁶ This photocatalyst assembly also reduced CO₂ to CO when visible-light irradiation was used in the presence of TEA as the sacrificial electron donor (Table 5: entry 51; TON_CO = 2). However, their photocatalytic activity was found to be lower than the mixed system containing its constituent components (Table 5: entry 52; TON_CO = 3). When Zn(II) tetraphenylporphyrin derivative was used as the photosensitizer unit in place of Pd, again, a lower photocatalytic efficiency was achieved with respect to their mixed system (Table 5: S60; entry 53; TON_CO = 14; S61; entry 54; TON_CO = 32 and mixed system: entry 56; TON_CO = 103).⁸⁷ The photocatalytic efficiency could be enhanced by introducing a methylene group on the −NHCO– linkage and the Re component (Table 5: S62; entry 55; TON_CO = 332) owing to their larger reducing power with respect to S60 and S61 (S62: E^1/2 red = −1.68 V vs Fc/Fc⁺; S60: E^1/2 red = −1.44 V; S61: E^1/2 red = −1.42 V).^53,82,151 More recently, Kuramochi and co-workers (2020) developed a dyad, ZnP-phen = Re, where a zinc-based porphyrin photosensitizer has been coupled with a fac-Re(phen)(CO)₃Br catalyst component via its meso-position. This system exhibited an excellent durability, achieved a healthy TON_CO of 1300, and selectivity for CO > 99.9%, which was attributed to the strong interaction between the porphyrin and the Re component where the electron reservoir on the porphyrin readily transfers to the Re component and causes hydrogenation of the porphyrin structure.¹⁵² The various structures and designations of porphyrin–Re(I)-based supramolecule photocatalysts, along with their photocatalytic performances, are presented in Chart 11 and Table 5, respectively.

Table 5. Summary of Performances by Porphyrin–Re(I)-Based Supramolecule Photocatalysts^a.

entry	photocatalyst	reductant	product	ϕproduct	TON	reference
50	S58	TEA	CO	0.0064		(84)
51	S59	TEA	CO		2	(86)
52	Pd(POR) + fac-[Re(bpy)(CO)3(3-picolene)]+	TEA	CO		3	(86)
53	S60	TEOA	CO		14	(87)
54	S61	TEOA	CO		32	(87)
55	S62	TEOA	CO		332	(151)
56	Zn(POR) + fac-[Re(bpy)(CO)3(3-picolene)]+	TEOA	CO		103	(87)
57	S63	TEA	CO		13	(153)
58	S64	TEA	CO		<1	(153)
59	S65	TEA	CO		<1	(153)
60	S66	TEA	CO		<1	(153)
61	S67	TEA	CO		<1	(153)
62	S68	TEA	CO		<1	(153)
63	S69	BIH	CO		18	(82)

^aReproduced with permission from ref (53). Copyright [2017] [American Chemical Society].

Chart 11. Various Structures and Designations of Porphyrin–Re(I)-Based Supramolecule Complexes.

Various mononuclear porphyrin metal complexes, viz. M-TPP, containing a redox-active metal such as iron or cobalt, have also been reported to undergo a photocatalytic reduction of CO₂.^154−157 Fujita et al. and Neta et al. reported the syntheses of FeCl-TPP and Co-TPP. They evolved CO as the reduction product with TON_CO = 80 for Co-TPP and TON_CO = 65 for FeCl-TPP; both were irradiated at λ > 320 nm for 200 h.^156,157

Schwalbe and co-workers¹⁵³ reported the development of M-Por and M-Por-Ru compounds, where M = H₂, Cu, Pd, Co, and FeCl evolved from the phenanthroline extended tetramesityl porphyrin ligand (H₂-Por). Here, the Ru-based component is the [Ru(tbbpy)₂]²⁺ complex, where tbbpy = 4,4′-di-tert-butyl-2,2′-bipyridine.^158,159 Among the various developed supramolecules, FeCl-1 and FeCl-1-Ru exhibited the best photocatalytic performance, even higher than the earlier known photocatalysts like Co-TPP and FeCl-TPP. They, for the first time, developed a new type of M-1-Re complex with Re = Re(CO)₃Cl, where a Re(I) tricarbonyl component is connected through a delocalized π-system linker (Chart 12).¹⁵³

Chart 12. Structures of M-Por and M-Por-Re Compounds where M = H₂, Cu, Pd, Co, Zn, and Fe.

Takeda and co-workers (2016)¹⁶⁰ reported the design of a novel, durable, and efficient photocatalyst for CO₂ reduction based on some nonprecious-metal complexes. They developed a Cu^I diimine complex as the photosensitizer and an Fe^II diimine moiety as the catalyst component (Chart 13).

Chart 13. Various Structures and Designations of Cu^I Photosensitizers (a) and Fe^II-Based Catalyst Component (b).

Thus, various reports on photocatalytic CO₂ reduction by supramolecule complexes based on Co,^161−165 Ni,^50,166,167 Fe,^160,168,169 and Mn^170−172 have been reported in the literature. However, the overall efficiency and selectivity of most of these photocatalytic systems are not satisfactory.

Leung and co-workers (2012) have reported that a Co-based complex [Co(qpy)(OH₂)₂]²⁺ (where qpy = 2,2′:6′,2″:6″,2‴-quaterpyridine) can prove to be a superior photocatalyst for water oxidation and CO₂ reduction.¹⁷³ Guo et al. (2016) also supported such reports. Interestingly, these photocatalysts enhance their photocatalytic performance when a cheap organic dye purpurin is used as a photosensitizer (Chart 14).¹⁷⁴

Chart 14. Cobalt-Based Porphyrin-Type Catalyst and Its Fe-Based Analogue.

2.3.7. Polyoxometalate-Based Supramolecules

The polyoxometalates (POMs) are combined with specific organic- or inorganic-type molecules through noncovalent interactions to develop a new class of supramolecules, which can be exploited for photocatalytic CO₂ reduction reactions. These systems comprise POMs along with metal–organic frameworks, cationic metal complexes, organic molecular networks, or an assembly of various POM units. The POM-based supramolecular hybrids require compatibility in the dimensions of the constituent components so as to deliver the desired activity and stability.¹⁷⁵

Das and co-workers (2018) developed a complex polyoxometalate Na₄₈[H_XMo₃₆₈O₁₀₃₂(H₂O)₂₄₀(So₄)₄₈] ca. 1000 H₂O designated as Mo₃₆₈, which was exploited for the photocatalytic reduction of CO₂ to HCOOH. They achieved a high TON of 27 666 with a TOF of 4611 h^–1 and a quantum efficiency of 0.6%. This photocatalyst oxidizes the H₂O in the medium to provide the electrons for the reduction of CO₂. The superior activity was attributed to various Mo subunits in the cluster, and no external photosensitizer or reductant was used, which makes this reaction environmentally friendly.¹⁷⁶

Huang and co-workers (2020) developed two structurally analogous porphyrin-POM coordination frameworks, where [PMo₈^VMo₄^VIO₃₅(OH)₅Zn₄]₂[Zn-TCPP][2H₂O]·xGuest (NNU-13) and [PMo₈^VMo₄^VIO₃₅(OH)₅Zn₄]₂[Zn-TCPP][2H₂O]·yGuest (NNU-14) are assembled with porphyrin {ϵ-PMo₈^VMo₄^VIO₄₀Zn₄}, designated as (Zn-ϵ-Keggin).¹⁷⁷ Here, PMo₁₂ is surrounded by four Zn^II ion nodes and a photosensitive tetrakis(4-carboxylphenyl) porphyrin derivative (H₂TCPP) behaving as bridging ligands. These systems showed good stability in aqueous solutions for photocatalytic CO₂ reduction in the absence of an external photosensitizer or a cocatalyst. These systems showed a high selectivity for CH₄ formation (96.6% for NNU-13 and 96.2% for NNU-14) on reduction of CO₂. The superior photocatalytic activity was attributed to the strong reduction potential of the Zn-ϵ-Keggin moiety and enhanced charge-transfer properties of the TCPP linker.^178,179

Haviv and co-workers (2017) demonstrated the photoreductive capability of an acidic polyoxometalate, H₅PW₂^VW₁₀O₄₀, linked to a dirhenium molecular complex. This system exhibited a superior performance for the selective production of CO from photocatalytic reduction of CO₂.¹⁸⁰

Researchers envisioned that by combining a chiral organic moiety and a POM in a MOF, the chemical conversion of CO₂ can be performed.^181,182 Han and coworkers introduced Keggin-type [ZnW₁₂O₄₀]^6– anions, NH₂-bipyridine, zinc(II) ions, pyrrolidine-2-yl-imidazole, NH₂-functionalized bridging links, and asymmetric organocatalytic groups within a single MOF. They revealed that the Lewis acid zinc centers could synergistically interact with CO₂ to undergo its assimilation.¹⁸³

Another MOF-entrapped POM, TBA₅[P₂Mo₁₆^VMo₈^VIO₇₁(OH)₉Zn₈(L)₄] (NNU-29), was reported by Li and co-workers (2019).¹⁸⁴ These systems exhibited photocatalytic CO₂ reduction for the production of HCOO^– (220 μmol g^–1 h^–1) and selectivity of 97.9% in an aqueous medium after 16 h.

Some other POM-based hybrid catalyst systems such as Na₆[Co(H₂O)₂(H₂tib)]₂{Co[Mo₆O₁₅(HPO₄)₄]₂}·5H₂O (1), Na₃[Co(H₂O)₃][Co₂(bib)] (H₂bib)_2.5{HCo[Mo₆O₁₄(OH)(HPO₄)₄]₂}·4H₂O (2), and (H₂bpp)₃ {HNa[Mo₆O₁₂(OH)₃(HPO₄)₃(H₂PO₄)]₂}·6H₂O (3) have been reported by Du and co-workers (2020).¹⁸⁵ Here, bib = 1,4-bis(1-H-imidazol-4-yl) benzene, tib = 1,3,5-tris(1-imidazolyl) benzene, and bpp = 1,3-bi(4-pyridyl) propane. These POMs were integrated with the [Ru(bpy)₃]Cl₂ complex to develop novel photocatalytic systems, which exhibited good photocatalytic activity for CO₂ reduction reactions. The first two systems (1 and 2) yielded CO (322.7 and 281.81 nmol) with selectivities of 96.3 and 96.4%, respectively. The superior performance was attributed to the fact that {P₄Mo₆}-based POMs could readily oxidize TEOA and Ru(bpy) acted as the catalytic component for the reduction of CO₂ (Figure 5).¹⁸⁵

Figure 5.

Possible regulation mechanism for photocatalytic CO₂ reduction by 1 or 2/Ru(bpy). Reprinted with permission from ref (185). Copyright [2020] [Elsevier B.V.].

In view of the above discussions, POM-based supramolecular assemblies have found new dimensions for the photocatalytic assimilation of CO₂. Albeit these systems are quite complex, extensive research can unravel their vast potential for reductive reactions.

2.4. Selection of the Bridging Ligand

The bridging ligand has an important role in the overall performance of the supramolecule photocatalysts.⁶³ It has been generally found that the introduction of conjugation in the bridging ligand is detrimental to the photocatalytic performance of a supramolecule photocatalyst.¹⁸⁶ This difference should occur for the reason that there must be a difference in their E^1/2 red.⁵³ The various supramolecule photocatalysts with different bridging ligands are presented in Chart 15. The UV–vis spectra of photocatalysts S78 and S81 are presented in Figure 6, which also differ only in terms of the bridging ligand.

Chart 15. Multinuclear Ru(II)–Re(I)-Based Supramolecule Photocatalysts with Different Bridging Ligands⁵²a.

^a The Chart has been reproduced with permission from ref (1). Copyright [2015] [American Chemical Society].

Figure 6.

UV–visible spectra of various metal complexes dissolved in acetonitrile: (A) S88 (dashed line), S89 (dot-dashed line), and S78 (solid line); (B) S90 (dot-dashed line) and S81 (solid line). Reprinted with permission from ref (52). Copyright [2005] [American Chemical Society].

As is evident from Figure 6, the complex S78 has a very identical spectrum to that of its mixed constituents (Figure 6a). However, the complex S81 has a bit different spectrum from that of the mixed system from its constituents (Figure 6b). It follows that in the case of complex S81, there is an effective interaction between the two diimine moieties in the bridging ligand due to conjugation. It was found that both complexes act as photocatalysts in the CO₂ reduction reaction. However, unexpectedly, complex S81 had a much lower activity than S78 because of conjugation in the bridging ligand. Complex S78 exhibited TON_CO = 170, which was much higher than that of S81. They found that complex S78 had a reduction potential of −1.10 V, which was very much lower than that of complex S78.

They revealed that the presence of conjugation in the bridging ligand actually deepens the π*-orbital energy in the diimine moiety of the Re component, which lowers its reduction potential.¹

Koike and Ishitani¹⁸⁷ for the first time employed time-resolved infrared (TRIR) spectroscopy to three Ru(II)–Re(I)-based supramolecules, viz. RuRe(X), where X = FPh, Ph, and OEt (Chart 16), in the presence and absence of a reductant BNAH (Scheme 3).

Chart 16. Binuclear Complexes Based on Ru(II)–Re(I) with Varying Substituents on the Ligands and Structures of BNAH and BIH Reductants.

Scheme 3. Oxidative and Reductive Quenching Pathways for the Photochemical Generation of OERS from the Ru(II)–Re(I)-Based System.

They found that while using a saturated linker between them, there is a rapid intramolecular electron transfer between the excited states and OERS of the Ru component to the Re-based catalyst component (k_ET > 2 × 10⁷ s^–1).⁵³ This observation discouraged the use of conjugated linkers in supramolecule photocatalysts.

Kato and co-workers (2015) also reported the development of supramolecule photocatalysts based on Ru(II)–Re(I) assemblies connected through a saturated chain, viz. Ru(CH₂CH₂)Re.⁵² They achieved a good quantum yield and turnover number for CO generation, 0.15 and 207, respectively, when BNAH was used as a reductant.⁹⁸ This was the best-performing photocatalyst of the time.¹¹³ They also introduced −CH₂OCH₂– or −CH₂SCH₂– as bridging linkers. The −CH₂OCH₂– chain containing the supramolecule photocatalyst exhibited the highest activity when BNAH was used as the reductant.¹¹⁴ Cheung and co-workers (2019) added the amide groups on the Ru- and Re-based photosensitizer and catalyst components, respectively, to ensure hydrogen bonding between the two moieties.¹⁸⁸ They revealed that hydrogen bonding enhanced the performance of the supramolecule photocatalyst for CO₂ reduction with respect to the individual components. This photocatalyst exhibited almost thrice better turnover number (TON_CO = 100 ± 4) and quantum yield (Φ_CO = 23.3 ± 0.8%) for CO production with respect to the unsubstituted Ru and Re components (TON_CO = 28 ± 4 and Φ_CO = 7 ± 1%) in acetonitrile (MeCN) using BIH as the reductant. The better performance was attributed to the enhanced intramolecular charge transfer in the supramolecule system and the prevention of dimerization in the individual components.¹⁸⁸

Wang and co-workers (2021) investigated the role of coordinative interactions between the photosensitizer and the catalyst components.¹⁸⁹ They revealed that the coordinative interaction between the iridium-based photosensitizer and the cobalt phthalocyanine-based catalyst component potentially enhanced the electron transfer and resulted in superior quantum efficiency (10.2 ± 0.5% at 450 nm) for CO production. It exhibited almost 100% selectivity for CO, and a turnover number of 391 ± 7 was achieved, which is 4 times better than a comparative system with no coordinate interactions.¹⁸⁹

From the above discussion, it can be concluded that there are a few technicalities in choosing a linker. (1) The electron on the excited photosensitizer component should be located on the bridging ligand so as to ensure effective electron mobility to the catalyst component. (2) A saturated or nonconjugated linker must be tried.⁵³ (3) The peripheral ligands of the photosensitizer component should have equal or higher π* orbital energy with respect to that of the bridging ligand so as to ensure smooth electron mobility from the photosensitizer OERS to the catalyst component. (4) The bridging ligand must have sufficient π* orbital energy to reinforce CO₂ reduction.⁵⁴

2.5. Length of the Alkyl Chain

The photocatalytic activity of supramolecule photocatalysts for CO₂ reduction is well influenced by the alkyl chain length in the bridging ligand.¹¹² The Ru(II)–Re(I) supramolecule complex achieved the highest photocatalytic performance when an ethylene chain (−CH₂CH₂−) was used as the alkyl chain in the linker. On increasing the carbon number of the alkyl chain, the photocatalytic performance diminished. The weaker transfer of electrons along Ru and Re components on using the ethylene chain as the linker modulated the oxidation potential of the Ru component. This rapidly quenches the ³MLCT excited state involving the sacrificial donor BNAH. As such, the reduction potential of the Re component went much more toward the negative—E^red = −1.72 V vs Ag/AgNO₃—so as to undergo a photocatalytic reduction reaction (Chart 17).^1,98

Chart 17. Ru(II)–Re(I) Supramolecule Complex with Different Alkyl Chain Lengths.

Yamazaki and co-workers (2019) investigated the rate constants of intramolecular charge transfer between OERS of the Ru moiety and the Re catalyst unit in Ru–Re supramolecular photocatalysts containing different bridging ligands.¹⁹⁰ They revealed that the rate of charge transfer is the function of the alkyl chain length in the bridging ligand and the number of chains. A linear relationship was obtained between the logarithm of the rate constants and the chain length of the bridging ligands. The gradual slope demonstrates a small decay coefficient factor (β) of 0.74 Å^–1. However, the introduction of two ethylene chains as bridging ligands results in an enhanced electron transfer with respect to the complex containing only one ethylene chain. It was found that the electron transfer in lower chain bridging ligands was even faster than the reductive quenching process happening on the photocatalyst, thus avoiding the rate-determining step and being of great importance for a higher activity.¹⁹⁰

2.6. Peripheral Ligand

Ishitani and co-workers (2015)⁵⁴ studied the role of peripheral anionic ligands in Re(I)-based complexes for the photocatalytic reduction of CO₂ using fac-Re(bpy)(CO)₃X, where X = Cl^–, CN^–, and SCN^–.⁴⁹ They found that the complex with the SCN^– ligand exhibited a better performance (Φ_CO = 0.20 and TON_CO = 30) with respect to that of the Cl^–-containing complex, and the CN^–-containing complex was quite ineffective for CO₂ reduction. The zero activity of the CN^–-containing complex is because the OERS does not release the CN^– ligand. However, the OERS of the SCN^–-containing complex was much higher than that of the Cl^–-containing complex, which provided a better electron source for photocatalytic CO₂ in the SCN^–-containing complex.¹

Triethylphosphite was also introduced as a peripheral ligand (S98) in place of Cl^–, which enhanced the photocatalytic performance (Φ_CO = 0.16, TON_CO = 232). On the other hand, use of a pyridine ligand (S99) diminished the performance of the photocatalyst (TON_CO = 97).¹¹¹ It was revealed that (S98) got rapidly modified into a complex possessing the −OC(O)OC₂H₄N(C₂H₄OH)₂ moiety as a ligand, generated by the reaction between deprotonated TEOA at the Re center and CO₂.^111,115,191 This assembly was actually responsible for the photocatalytic reduction of CO₂. In (S99), the free pyridine in the medium facilitated the deterioration of the photocatalyst (Charts 18 and 19).⁵³

Chart 18. Ru(II)–Re(I)-Based Supramolecule Photocatalysts with Different Peripheral Ligands.

Chart 19. Representative Structure of Ru(II)–Re(I)-Based Supramolecular Photocatalyst.

Gholamkhaas and co-workers (2005) reported the influence of −CF₃, −H, and −CH₃ as substituents on the peripheral ligands for the photocatalytic reduction of CO₂.⁵² The photocatalytic performances of the different possible photocatalysts have been summarized by Ishitani and co-workers (2015), as shown in Table 6.⁵⁴

Table 6. Summary of the Photocatalytic Performance by Ru(II)–Re(I)-Based Supramolecules Containing Different Substituents on the Peripheral Ligands^a.

entry	Complex	R	X	L	L′	n	TON	ϕCO
64	Ru(CF3)–X–ReCl	CF3	CH2CH(OH)CH2	CO	Cl	2	3
65	Ru(H)–X–ReCl	H	CH2CH(OH)CH2	CO	Cl	2	50
66	Ru(Me)–X–ReCl	CH3	CH2CH(OH)CH2	CO	Cl	2	170	0.12

^aReproduced with permission from ref (54). Copyright [2015] [Elsevier. B. V.].

They found that Ru(Me)–X–ReCl (entry 66) exhibited much better TON_CO and Φ_CO in comparison to (Ru(CF₃)–X–ReCl) and (Ru(H)–X–ReCl). This phenomenon was attributed to the better reduction potential in OERS of the former than the latter, governed by the nature of the substituent.^1,54,111 Morimoto and co-workers (2013) demonstrated the role of triethanolamine (TEOA) as the peripheral ligand. They added TEOA to a rhenium(I) tricarbonyl diimine complex containing N,N-dimethylformamide (DMF) as one ligand. The DMF ligand gets replaced by TEOA, which then captures a CO₂ molecule. This helps assimilate even the lower concentrations of CO₂, as in air.¹⁹¹ Similarly, Kamogawa and co-workers (2021) reported the role of the triethanolamine ligand on the Re-based catalyst unit in a Ru-C2-Re-based supramolecule photocatalyst. They demonstrated that the triethanolamine ligand on the catalyst moiety reacts with CO₂ to produce a carbonate ester complex, which then subsequently gets reduced to CO. This phenomenon was found beneficial in the sense that even lower concentrations of CO₂ (10%) could get reduced efficiently.¹⁹²

2.7. Selection of the Semiconductor Component

There are various visible-light-responsive semiconductor photocatalysts exploited for photocatalytic oxygen evolution from H₂O.^193−196 However, these photocatalysts have very low activity and selectivity for photocatalytic CO₂ reduction.^197−199 Researchers suggest if they could hybridize a supramolecule photocatalyst with a semiconductor component containing a strong oxidation power, a new and efficient photocatalyst for CO₂ reduction can be achieved with water as the reductant. The semiconductor–supramolecule hybrid system can be advantageous as the photosensitizer and catalyst components of the supramolecule can be brought in proximity to each other on the semiconductor substrate. In these systems, the excited electrons in the conduction band of the semiconductor component can pump electrons on the photosensitizer unit of the supramolecular component. The photosensitizer component will eventually transfer the electrons to the catalyst component with a strong reduction power. The schematic representation of the whole process is provided in Figure 7.

Figure 7.

Schematic representation showing the electron transport in a semiconductor–supramolecule hybrid for CO₂ reduction.

2.7.1. Metal Oxides

Various metal oxide-based semiconductors, possessing a higher conduction band potential, have been employed to integrate with the supramolecular photocatalysts through the anchoring groups. Various researchers processed these metal oxide supports by doping with metallic and nonmetallic atoms to optimize the light absorption and reduction potential characteristics. Some reports demonstrate the use of a heterostructure formed between a metal oxide and some other organic or inorganic species, which enhances the charge-transfer properties and mitigates charge recombination. Sato and co-workers used N-doped Ta₂O₅, a p-type semiconductor (E_CB = −1.3 V, E_VB = 2.7 V, E_Donor-Level = 1.1 V vs NHE), as the substrate for the supramolecule involving [Ru{4,4′-(COOH)₂bpy}₂(CO)₂]²⁺ (E_p red = −0.8 V vs NHE) as the photosensitizer.²⁰⁰ Using TEOA as the sacrificial electron donor, HCOOH was again obtained as the major product with TON_HCOOH = 89 and selectivity more than 75%, along with H₂ and CO. A carbon nitride (C₃N₄)-based semiconductor–supramolecule hybrid was also developed involving cis,trans-Ru{4,4′-(PO₃H₂)₂bpy}(CO)₂Cl₂ (E_p red = −1.4 V vs Ag/AgNO₃). Using TEOA as the sacrificial reductant, TON_HCOOH > 1000 and apparent Φ_HCOOH = 5.7% were observed using irradiation at λ_ex = 400 nm.^201−203

TiO₂ was used as a semiconductor support involving a dye-based photosensitizer and a Re(I)-type complex as the catalyst component so as to develop an effective photocatalyst for CO₂ reduction (TON_CO > 435).²⁰⁴

A NiO electrode was used to develop a photocathode that combines with the Zn(II) porphyrin–Re(I) component through carboxyl anchor groups to develop a CO₂ reduction photocatalyst.²⁰⁵ Ru(II)–Re(I)-based supramolecules have also been hybridized with the NiO electrode through methylphosphonate groups (Figure 8), and TON_CO = 32 was achieved.^205−207 Similarly, a p-type InP photocathode was combined with the Ru(II) polymer to develop a CO₂ reduction photocatalyst. This photocathode hybrid photocatalyst was also combined with a TiO₂ photoanode containing Pt particles as cocatalysts. HCOOH and O₂ with 0.03% of the solar conversion efficiency were achieved.²⁰⁸ The photocatalytic efficiency was enhanced when a reduced SrTiO₃ photoanode was used in place of TiO₂ (0.14%).²⁰⁶ Amorphous silicon–germanium and Ru polymer catalyst and IrO_x were brought together to achieve a better solar-energy conversion efficiency (4.6%).²⁰⁹

Figure 8.

NiO-Ru(II)–Re(I) hybrid supramolecule photocatalyst anchored by methylphosphonate groups; reprinted with permission from ref (210). Copyright [2019] [American Chemical Society].

Faustino and co-workers (2018) reported the use of hexaniobate nanoscrolls (K_xH(_4–x)Nb₆O₁₇) in place of TiO₂ as a semiconductor substrate. The hexaniobates are bestowed with a better conduction band reduction potential (E_CB⁰ ≈ −0.75 V vs NHE)^211,212 with respect to the TiO₂ substrate (E_CB⁰ ≈ −0.3 V vs NHE),²¹² which enhances their reducing capability. Further, their scrolled lamellar morphology can introduce better electron transfer by facilitating a vectorial charge transfer along with the individual layers. This phenomenon is seen to suppress the ohmic losses occurring along the grain boundaries.²¹³ This semiconductor was made hybrid with two different Re(I) complexes. The insights obtained from the results highlight its role and that of the anchor group on the photocatalytic performance of immobilized Re(I) complexes.

Wang and co-workers (2012) investigated various supramolecule photocatalysts based on porphyrins of Zn(II), Cu(II), and Co(II) as photosensitizers. They integrated the Ru(II) polypyridyl/complexes with these porphyrins and were supported by modified TiO₂ nanotubes. They demonstrated the role of metal-ion coordination and peripheral ligands in the porphyrins on the efficiency of the photosensitization when they were impregnated on a TiO₂ nanotube surface. It was revealed that the Zn(II) porphyrin-based supramolecule complex showed the best efficiency for CO₂ reduction to methanol in an aqueous medium.

Kumagai and co-workers developed a Ru(II)–Re(I)/CuGaO₂ photocathode where CuGaO₂ works as a p-type semiconductor electrode and the Ru(II)–Re(I) complex acts as a photocatalyst for CO₂ reduction.^214,215 The synthesized Ru(II)–Re(I)/CuGaO₂ photocatalysts provided an onset potential for CO₂ reduction 0.4 V more positive with respect to that of the Ru(II)–Re(I)/NiO electrode. Ru(II)–Re(I)/CuGaO₂ and a CoO_x/TaON photoanode were used to develop a photoelectrochemical cell that could behave as a CO₂ reduction photocatalyst, and water was taken as a reductant.²¹⁶

Saito and co-workers (2020) demonstrated the photocatalytic reduction of CO₂ to CO using a Ru–Re-based supramolecule catalyst supported on insulating Al₂O₃ through phosphate linkers. They found that the Al₂O₃ support actually brought the photosensitizer and catalyst component in proximity, hence enhancing the durability of the resultant supramolecule through covalent bridging of the molecular units. The higher density of the individual components reduces the TOF as the sacrificial reductant is hindered from approaching the catalyst.²¹⁷

2.7.2. Metal Oxynitrides

Metal oxynitrides have a good visible-light absorption characteristic due to their small band gap and have a favorable conduction band potential for reduction reactions. Researchers have tried to optimize these properties by doping of metal ions or heterostructure formation with metal oxynitrides.

Sahara and Ishitani (2015) reported a Ag/TaON-based semiconductor–supramolecule hybrid where [Ru(4,4′-Me₂bpy)(N^∧N){4,4′-(H₂O₃PCH₂)₂bpy}]²⁺ is the photosensitizer and the cis,trans-Ru(N^∧N)-(CO)₂Cl₂ moiety is the catalyst component. Methylphosphonate groups were used as anchors to bring the connection between the Ag/TaON component and the supramolecule component.²¹⁸ This hybrid catalyst was found to produce HCOOH as the major product along with CO, HCHO, and H₂. The supramolecule photocatalyst without Ag/TaON did not reduce the CO₂ in methanol. Similarly, Ag/TaON was not able to reduce CO₂ without the supramolecule component but evolves H₂.¹ When CaTaO₂N was used instead of TaON to produce the hybrid photocatalyst, high selectivity for HCOOH production (>99%) was achieved in the dimethylacetamide–triethanolamine mixed solution.²¹⁹ Hence, metal oxynitrides were found to have more selectivity for the formation of HCOOH from CO₂.

2.7.3. Metal Chalcogenides

Metal chalcogenides, viz. metal sulfides and metal selenides, have also been employed for making hybrids with supramolecule-based CO₂-reduction photocatalysts.^220−222 Various metal chalcogenides, which were used for HER reactions, such as CdS, Ni-doped ZnS, (AgIn)_0.22Zn_1.56S₂, and (CuGa)_0.8Zn_0.4S₂, have now been employed as semiconductor supports for complex catalysts such as trans(Cl)-[Ru{4,4′-X₂-2,2′-bipyridine}(CO)₂Cl₂] (X = H, CH₃, COOH, PO₃H₂, or CH₂PO₃H₂), for visible-light-driven CO₂ reduction to obtain HCOOH using acetonitrile as the solvent and TEOA as the reductant.²²³ Some chalcogenide quantum dots have also been used with supramolecules in aqueous mediums.^220,222 For example, CdS quantum dots were combined with nickel terpyridine complex photocatalysts, which exhibited photocatalytic reduction of CO₂ to produce CO in 0.1 M aqueous TEOA solution, where a TON_CO of ≈20 and a good selectivity of ≈90% for CO production were obtained.^220,224 ZnSe quantum dots combined with a Ni(cyclam) catalyst through a phosphonate anchoring group showed a good CO₂ reduction activity to produce CO, having TON_CO higher than 120.²²¹ When these ZnSe quantum dots were surface-modified with 2-(dimethylamino)ethanethiol, the selectivity for CO was enhanced while suppressing the HER reaction. A TON of 280 was obtained with 33% selectivity for CO production after visible-light irradiation for 20h. Another hybrid catalyst was developed when CuInS₂/ZnS (core/shell) quantum dots combined with meso-tetraphenylporphyrin iron(III) chloride to undergo photocatalytic reduction of CO₂ to CO. A TON higher than 50 (on 40 h irradiation) and selectivity of 84% for CO were obtained in dimethyl sulfoxide (DMSO) with traces of H₂O.²²⁵ When a negatively charged CuInS₂/ZnS quantum dot colloidal solution was combined with oppositely charged trimethylamino-functionalized iron tetraphenylporphyrin complexes, the TON for photoreduction of CO₂ to CO was enhanced to 450 (on 30 h irradiation) and a selectivity of 99% was achieved in an aqueous solution of 15 mM TEOA and 5 mM KCl.²²²

From the above discussion, the metal chalcogenide support can be exploited for enhancing the selectivity for CO production from photocatalytic CO₂ reduction.²²⁶

2.7.4. Carbon-Based Materials

Various metal-free semiconductor materials have been used as supports for supramolecule photocatalysts, viz. graphene, carbon nitride, carbon nanotubes, carbon nanodots, etc. These materials have been found to be potential candidates due to their low band gap, high visible-light absorption, and enhanced charge transfer, and their functionalized surface can efficiently integrate with the supramolecule moiety via the anchoring groups, leading to superior charge transfer throughout the system.

Wada and co-workers (2017) developed a hybrid photocatalyst involving a binuclear Ru(II)–Re(I) complex and mesoporous graphitic carbon nitride (mpg-C₃N₄). This photocatalyst could selectively generate CO by CO₂ reduction.²²⁷ When the Ru(II)–Re(I) moiety is immobilized on the TiO₂/NS-C₃N₄ composite (NS = nanosheet), a better photocatalyst is achieved as the TiO₂/NS-C₃N₄ substrate provides a Z-scheme mechanism. This was possible because the conduction band maximum of TiO₂ is more negative with respect to the reduction potential attained by the excited state of the Ru unit in Ru(II)–Re(I)²²⁸ and hence readily transfers the electrons to the photosensitizer component. Hence, the Ru(II)–Re(I)/TiO₂/NS-C₃N₄ hybrid photocatalyst showed an enhanced performance four times better than Ru(II)–Re(I)/NS-C₃N₄, selectively producing CO with TON_CO = 73.²²⁹ Chen and co-workers (2020) investigated the fabrication of a lanthanum–nitrogen charge-transfer bridge on g-C₃N₄ as the active site for photocatalytic CO₂ reduction. They revealed that the bridge between La–N was responsible for the higher CO production (92 μmol g^–1 h^–1) with an attractive CO selectivity of 80.3%, which was superior to most of the g-C₃N₄-based photocatalysts. This performance was attributed to the formation of a charge-transfer channel between the La–N atoms through p–d orbital hybridization.²³⁰ Thus, engineering with C₃N₄-based materials has been explored to enhance the selectivity for CO production in photocatalytic CO₂ reduction reaction.

2.7.5. Silica-Based Materials

Periodic mesoporous organosilica (PMO) has been widely used as a light-harvesting (LH) material. The silica framework with various organic groups introduced into it acts as a light-absorbing antenna. These materials have also been hybridized with metal complexes so as to function as photocatalysts.^228,231 As a first example, PMOs containing biphenyl groups in their framework were used to immobilize the Re(I) complex so as to achieve a photocatalyst.²³¹ Another photocatalyst used PMO where acridone or methylacridone moieties have been introduced in the framework, and a Ru(II)–Re(I) supramolecule complex is anchored with it through methylphosphonate groups (Figure 9).²²⁸ These acridone groups harvest light energy, which can transfer to the Ru photosensitizer component of the supramolecular photocatalyst and then reduce CO₂ to selectively produce CO while using BIH as a sacrificial donor (TON_CO = 635). The methylacridone hybrid system exhibited a photocatalytic capability almost tenfold larger than that of the same supramolecule system immobilized on an ordinary mesoporous silica gel (MCM-41).

Figure 9.

PMO having acridone groups in its framework and the Ru(II)–Re(I) supramolecule photocatalyst immobilized on it. Reprinted with permission from ref (54). Copyright [2015] [Elsevier Ltd].

Various other hybrid systems involving Re(I) tricarbonyl complexes immobilized over zeolites,^232,233 silica nanoparticles,²³⁴ mesoporous silica,^232,235,236 MXenes,²³⁷ and clays²³⁸ were also applied for photocatalytic CO₂ reduction. Metal–organic frameworks (MOFs) and other porous coordination polymers have also been blended with Ru(II)–Re(I) supramolecule photocatalysts.^239−241 More recently, perovskite materials are also considered as efficient supports for supramolecule photocatalysts for CO₂ reduction reactions. Sheng and co-workers (2020) demonstrated the performance of bismuth-based stable perovskite quantum dots, Cs₃Bi₂X₉ (where X = Cl, Br, I), for photocatalytic reduction of CO₂ to selectively produce CO. These materials showed a superior performance for CO₂ conversion and produced 134.76 μmol g^–1 CO with a good selectivity of 98.7% under solar light.²⁴²

2.8. Role of Anchoring Groups

The anchoring groups must have appropriate terminal functionality through which they can attach to the semiconductor support on one side and to the photosensitizer on the other side. Various linkers having terminal −COOH, −PO₃H₂, −SO₃H, pyridine, −OH, −SH, etc. have been investigated. For example, mesoporous graphitic carbon nitride (mpg-C₃N₄) was coupled with trans(Cl)-[Ru{4,4′-X₂-2,2′-bipyridine}(CO)₂Cl₂]-based complexes via X = COOH, PO₃H₂, CH₂PO₃H₂, H, or CH₃.^243,244 The electron-donating groups (such as CH₃, OCH₃, or CH₂PO₃H₂) in the linker were found to enhance E_red of the complex, while the electron-withdrawing groups (such as COOH or PO₃H₂) reduce the E_red. Thus, we can tune and optimize the reduction potential of the semiconductor–supramolecule hybrid catalyst by changing the functionality of the anchoring linkers. This way, the selectivity for a desired product can also be optimized. A phosphonate anchoring group on the [Ru(X₂bpy)₂(CO)₂]²⁺ complex has been found to be more selective for formate (HCOOH) production than an anchoring group with carboxylate.²⁴⁵ While using methylphosphonate (CH₂PO₃H₂) as the anchoring group for coupling Ru–Re complexes with the mpg-C₃N₄ semiconductor, the selectivity for HCOOH production can be enhanced in the photocatalytic CO₂ reduction.²²⁶

Ishitani and co-workers (2015) investigated the role of phosphonate groups used to anchor a Ru-based complex with carbon nitride (C₃N₄).²⁴⁶ It was observed that the anchoring group having no methylene spacers connecting between the phosphonic acid groups and bpy showed a better photocatalytic performance. This hybrid system achieved TON_HCOOH = 1000 with a quantum yield of 5.7% when irradiated at λ = 400 nm (Chart 20).²⁴⁶

Chart 20. Ru(II)–Ru(II) Supramolecule Photocatalyst with Methylphosphonate Groups as Anchoring Groups to Attach a Solid Substrate.

Kang and co-workers developed a hybrid system in which a [ReL(bpy)(CO)₃] (where L = 3-picoline or Br^–) complex was immobilized on the TiO₂ substrate through phosphonate groups.²⁴⁷ They revealed that the [ReL(CO)₃] unit (L = 3-picoline, Br) is coordinately bonded to a bpy ligand through phosphonate anchor groups along with the 4- and 4′-positions, firmly bound to TiO₂. This heterogeneous photocatalyst system (TON_CO = 52) outperformed the homogeneous molecular photocatalysts, viz. [ReCl(bpy)(CO)₃] (TON_CO = 30). Thus, the selection and modification of the anchoring ligands significantly determine the activity and product selectivity in photocatalytic CO₂ reduction.

2.9. Role of Reductants

Most of the photosensitizers based on transition-metal complexes require a sacrificial donor to undergo reductive quenching where an electron is taken by the OERS of the photosensitizer from the reductant/sacrificial donor. Hence, the selection of the reductant has a great influence on the photocatalytic activities of the photocatalyst systems. The reductive quenching process gives rise to OERS of the photosensitizer (PS^•–) and the OEOS of the sacrificial donor (D^•+). The durability of D^•+ highly influences the performance of the photocatalyst as the back-electron transfer from PS^•– to D^•+ is detrimental to the photocatalytic CO₂ reduction reaction. Hence, D^•+ should readily get degraded via deprotonation so as to mitigate its oxidation capability. The final oxidation products of the sacrificial donor must not react with the PS*, PS^•–, or Cat^•– species so as to suppress the photocatalytic activity.

From the above discussion, various sacrificial electron donors for supramolecular photocatalysts for CO₂ reduction have been employed, viz. BIH, BNAH, ascorbic acid, TEOA, isopropanol, Na₂S, and Na₂SO₃.²⁴⁸ The following classes of sacrificial reductants are being extensively used for photocatalytic CO₂ reduction using semiconductor–supramolecule hybrid systems.

2.9.1. NAD(P)H Model-Based Reductants

1-Benzyl-1,4-dihydronicotinamide (BNAH) is the main NAD(P)H-type compound. It has been applied in various photocatalytic reduction reactions as the sacrificial electron donor for CO₂ reduction. These types of compounds exhibit a higher reducing power with respect to triethylamine (TEA; E_oxp = 0.96 V vs SCE) and triethanolamine (TEOA; E_oxp = 0.80 V).⁴⁶ BNAH is reported to have a high reduction potential—E°(BNAH/BNAH^·+) = 0.57 V²⁴⁹—which facilitates the reduction of the excited [Ru(⁴dmb)₃]²⁺-type photosensitizers. Thus, BNAH can actually undergo reductive quenching of the excited photosensitizer component of supramolecule photocatalysts. A higher quenching fraction (η_q) causes the value of quantum efficiency to be higher and vice versa. A better sacrificial donor, viz. 1-(4-methoxybenzyl)-1,4-dihydronicotinamide (MeO-BNAH, E°ox = 0.50 V),²⁴⁹ enhances the efficacy of the reductive quenching process and eventually improves the photocatalytic performance.¹³⁵

It has been investigated that the OEOS of BNAH (BNAH^•+) deprotonates quite easily to give BNA^• under basic conditions.¹³⁵ On the other hand, TEOA does not undergo quenching of the excited state but can efficiently behave as a base so as to take a proton from BNAH^•+ present in the reaction medium. Two types of reactions can occur with BNA^• generated in the reactions; one is the formation BNA dimers, and in another process, BNA^• loses an electron to the medium and gets converted to BNA⁺. Hence, BNAH can behave as a single-electron donor or a two-electron donor. The BNA dimers pose a problem for photocatalytic reduction reactions since they are stronger electron donors (E°ox = 0.26 V vs SCE)²⁵⁰ with respect to BNAH. Hence, the quenching process of the dimers competes with that of BNAH. The OEOS values of the dimers are stable, and hence, the back-electron transfer preferentially occurs from the reduced photosensitizer to the OEOS of the dimer.^98,135 This reductive quenching and subsequent backward electron transfer are detrimental to the efficacy of photocatalytic reactions involving BNA dimers (Scheme 4).

Scheme 4. Mechanism of Sacrificial Reduction by BNAH and BIH Reductants.

2.9.2. Dihydrobenzoimidazole-Based Reductants

1,3-Dimethyl-2-phenyl-2,3-dihydro-1H-benzoimidazole (BIH) and BI(OH)H were found to be possessing a much better reducing power (BIH: E^1/2 ox = 0.33 V; BI(OH)H: E^1/2 ox = 0.31 V vs SCE) with respect to that of BNAH.^251−255 BIH undergoes oxidation quite differently from that of BNAH. It has been investigated that BIH gives two electrons on a photon excitation of the photocatalyst sequentially through electron transfer, deprotonation, and subsequent electron transfer. Since the OEOS and a deprotonated moiety of BIH (BI^•) possess a very high reducing power (E_p ox = −2.06 V vs Fc⁺/Fc),²⁵¹ it easily donates an electron to the catalyst component in the photocatalytic CO₂ reduction reaction. When BIH is used as a sacrificial donor in photocatalytic reactions, TEOA can also be used as a base so as to enhance the photocatalytic performance of supramolecular photocatalyst systems. BIH can also act as a base along with being a sacrificial agent. However, the protonated BIH (BIHH⁺) is not capable of reductive quenching and exhibits a lower activity when TEOA is not used.

BI(OH)H also acts as a better reductant than BNAH.¹³⁶ BI(OH)H also donates two electrons. In a photocatalytic CO₂ reduction reaction, a BI(OH)H molecule gives two electrons and two protons, which are taken by CO₂ so as to obtain BI(OH)H (BI(O^–)⁺) and HCOOH. Hence, BI(OH)H has better selectivity for HCOOH production than BIH. Thus, BNAH can provide one electron and one proton, BIH supplies two electrons and one proton, while BI(OH)H offers two electrons and two protons.

3. Conclusions

In this work, a comprehensive review has been provided on supramolecule catalyst systems for photoreduction of CO₂. The review has focused on the various challenges associated with CO₂ reduction and the ways forward. The evolution of various photocatalyst systems for CO₂ reduction has been thoroughly discussed. It can be understood from the literature that supramolecule photocatalysts have opened new ways of thinking and strategy in this field. Supramolecules have an advantage in that we can engineer with the variety of its constituent moieties so as get the desired performance in terms of activity, efficiency, or selectivity. The main motive of researchers is the efficient electron transfer from the photosensitizer unit to the catalyst unit and, also, the enhanced extra overpotential acquired by the catalyst component. The back-electron transfer from the OERS of the photosensitizer is also an issue. Researchers suggest the introduction of hybrid semiconductor–supramolecule photocatalysts. The semiconductor support can pump the electrons on the photosensitizer, facilitating multiple electron transfer to the catalyst component with a high overpotential. The semiconductor support can hence indirectly suppress the back-electron transfer on the photosensitizer. This review throws light on the selection of various components of the hybrid semiconductor–supramolecule photocatalyst. Researchers suggest that the semiconductor component must have a low band gap and visible-light active with good charge-transfer capabilities. It is also important that the conduction band potential of the semiconductor support must be higher than that of the photosensitizer unit so as to allow the electron transfer. The photosensitizer must have a strong visible-light absorption and should readily go into the excited state. The excited state must readily get the electron from the sacrificial reductant so as to undergo reductive quenching. The OERS of the PS obtained on reductive quenching should have a strong reducing power and should readily transfer the electron on the catalyst component. This is possible when the OERS has a reduction potential equal to or higher than the catalyst unit. The oxidized sacrificial donor should decompose through a process so as to suppress the back-electron transfer. This can be done by the deprotonation process through which it can lose its oxidation power. It is to be mentioned that the oxidation product of the sacrificial donor must not interfere with the progress of the photocatalytic reaction. The linker between the photosensitizer and the catalyst unit is also an important factor that determines the activity, efficiency, and selectivity of the CO₂ reduction products. Generally, aliphatic linkers without conjugation were found to show better performance. Some bridging linkers were also used to achieve better selectivity for a product. The length of the linker was found to influence the performance, and it was recommended to use smaller-chain-length linkers. Similarly, the anchoring groups used to attach the supramolecule on the semiconductor substrate must have a good affinity with the substrate. They should augment the charge transfer from the semiconductor to the photosensitizer unit. Finally, the catalyst unit has the main role in terms of performance and selectivity. The distribution of active sites, the overpotential, and the durability of the catalyst unit determine the quantum efficiency and turnover frequency of a product. The various peripheral ligands attached to the catalyst component are crucial for inculcating such properties in the catalyst unit.

To dream about the practical utility of artificial photosynthesis, a lot is yet to be done in terms of a deeper understanding of the subject. The fundamental mechanisms involved in photocatalytic CO₂ reduction must be clearer and fool-proof so that efficient photocatalysts can be designed.

The authors declare no competing financial interest.