Electrochemical/Photochemical CO2 Reduction Catalyzed by Transition Metal Complexes

Conversion of CO2 into useful chemicals is attractive as a solution of the fossil fuel shortage and the global warming problems. Reduction of CO2 into carbon monoxide (CO) and formic acid (HCOOH) is also important for obtaining the materials in organic syntheses. There are a lot of studies on the catalysts for electrochemical/photochemical CO2 reduction. Especially, transition metal complexes have actively researched as the molecular catalysts for CO2 reduction. In this chapter, the electrochemical/photochemical CO2 reduction catalyzed by cis-[Ru(bpy)2(CO)2] (bpy: 2,2′-bipyridine) and trans(Cl)-[Ru(bpy) (CO)2Cl2] is described as a representative example.


Introduction
Utilization of CO 2 becomes more and more important with increasing CO 2 emission which causes the global warming and the ocean acidification problems [1,2]. The huge CO 2 emission also relates on depletion of fossil fuels. The conversion of CO 2 into useful fuels and chemicals is very urgent to solve the abovementioned problems. The use of biomass instead of fossil fuels is actively researched and partly undertaken [3]. In many chemical laboratories, fixation of CO 2 into organic compounds by organometallic catalysts is vigorously studied [4].
Reduction of CO 2 with electrons is an attractive chemical conversion to obtain the useful products for fuels and chemical materials. It is so simple that it can be applied to photocatalyses which supply electrons from electron donors such as water. The equilibrium potentials (E 0' V vs. SHE at pH 7) for CO 2 reduction are listed in Figure 1 [5,6]; they are thermodynamic values and tend to positively shift with increasing the numbers of electrons participated. One-electron reduction of CO 2 requires very high energy. Furthermore, the product, CO 2 anion radical (CO 2 − •), is difficult to give useful organic chemicals because it is a very strong reducing reagent to reduce other molecules and recover CO 2 . Thus, the CO 2 reductions with multielectrons are desired; however, their reactions are generally difficult even in the electrochemical reduction. A reason is that the intermediates would release from the surface of the electrode as the products before accepting further electrons. To achieve CO 2 reduction with more than two electrons, the catalysts which allow to lower the activation energies are required. In other words, the catalysts can undergo the CO 2 reduction at the potentials closed to the equilibrium ones. The two-electron reduction of CO 2 produces carbon monoxide (CO) and formic acid (HCOOH). The equilibrium potentials are more negative than the proton reduction to afford H 2 . Therefore, the catalysts which can selectively reduce CO 2 rather than H + are also desired. Both CO and HCOOH are useful chemicals: CO can be converted into liquid hydrocarbons by using the Fischer-Tropsch reaction [7], and HCOOH which can be readily converted to H 2 is a safe storage material for H 2 [8].
In this chapter, the electrochemical CO 2 reduction catalyzed by the ruthenium complexes as the examples is described. The reduction products are CO and formic acid, while the nickel and rhenium complexes selectively yield CO. Discussion for the catalytic mechanisms is introduced particularly for the factors determining the product selectivity. In the next section, the photocatalytic CO 2 reduction assisting by the photosensitizers is described. The reaction procedures, the principles for selecting the photosensitizers and the electron donors, and the photocatalytic mechanisms are summarized. Furthermore, application of the homogeneous catalytic systems to heterogeneous catalyses, which is practically advantageous in the viewpoints of separation of the catalysts from the reactants and the products, is described. In the final section, the artificial photosynthetic systems, which would be realized by utilizing the molecular catalysts, are prospected.

Electrochemical CO 2 reduction
The representative molecular catalysts based on ruthenium complexes are cis-[Ru(bpy) 2 (CO) 2 ] 2+ (bpy: 2,2′-bipyridine), trans(Cl)-[Ru(bpy)(CO) 2 Cl 2 ] and the derivatives (Figure 3). They have the bipyridyl ligand which would act as an electron reservoir. The efficient catalysts have the carbonyl ligand, which would draw electrons. It tends to lower the reduction potentials of the metal complexes as well as to lower the overpotentials for the CO 2 reduction.

Electrochemical analysis
Electrochemical analyses (e.g., cyclic voltammetric measurements) are recommended to know the electrochemical properties of the molecular catalysts. The analyses do not only teach us the reduction potentials of the metal complexes but also show whether the complexes can react with CO 2 or not. Figure 4 shows the cyclic voltammograms (CVs) of cis-[Ru(bpy) 2 (CO) 2 ] 2+ in CH 3 CN or CH 3 CN/H 2 O (9:1). The Ag-Ag + (CH 3 CN) reference electrode (0.10 M Tetrabutylammonium perchlorate (TBAP) /0.01 M AgNO 3 in CH 3 CN) is used; the potential (0.00 V vs. Ag/AgNO 3 (CH 3 CN)) corresponds to −0.09 V vs. Fc/Fc + in CH 3 CN. The CV of cis-[Ru(bpy) 2 (CO) 2 ] 2+ in CH 3 CN under Ar shows an irreversible reduction wave at −1.3 V vs. Ag-Ag + (CH 3 CN) as shown in Figure 4 (black line). The irreversible reduction suggests that the one-electron reduction accompanies with a chemical reaction followed by further one-electron reduction. Such a reaction mechanism is called as electrochemical-chemical-electrochemical (ECE) one. The CV under CO 2 is a little different from that under Ar, suggesting that the reduced species react with CO 2 (Figure 4, blue line). In CH 3 CN/H 2 O (9:1), the CV exhibits a strong cathodic current under CO 2 (Figure 4, red line), which corresponds to the catalytic reduction of CO 2 in the presence of a proton source such as water. The catalytic reduction currents can be analyzed to estimate the efficiency of the catalyst [20,81]; however, it should be noted that the cathodic currents do not always exhibit the catalytic CO 2 reduction [82]. The electrolyses of the metal complexes under CO 2 should be carried out to confirm the catalytic efficiency.

Electrolysis
A typical electrolysis cell is shown in Figure 5. The cell for reduction (the side of the working electrode) is separated from the cell for oxidation (the counter electrode) with a membrane such as Nafion. A glassy carbon or a Pt plate is used for the electrodes. The metal complex is dissolved in the reaction solution and acts as the homogenous catalyst. CO 2 is bubbled   with a needle through the septum before electrolysis. Electrochemical CO 2 is carried out in batch mode. Reduction of CO 2 occurs on the working electrode at the electrochemical cell. Sampling of the gaseous and liquid phases is performed by a syringe through the septum. The gaseous products (CO and H 2 ) are analyzed by gas chromatography. The liquid product, HCOOH, is analyzed by electrophoresis, ion chromatography or gas chromatography.
The electrolysis is carried out by the controlled potential method, where the potential is determined from the electrochemical analysis (e.g., CVs). The chronopotentiometry, in which the current is constant during the electrolysis, is important for the industrial use. However, the results in the constant potential lead to elucidate the catalyses because the electrolysis potential relates on the catalytic species. Thus, almost all the scientific researches adopt the controlled potential electrolyses.

Electrocatalytic CO 2 reduction by cis-[Ru(bpy) 2 (CO) 2 ] 2+
The ruthenium complexes are used as the homogeneous catalysts by dissolving in the reaction solution. The electrolysis of the CO 2 -saturated H 2 O/DMF (1:1) solution of cis-[Ru(bpy) 2 (CO) 2 ] 2+ was carried out at −1.50 V vs. SCE with an Hg pool as the working electrode ( Figure 6) [62]. The catalyst could selectively reduce CO 2 to afford CO and HCOOH, while H 2 , the reduction product of water, scarcely evolved. As the reaction proceeded, the speed for CO production got slow, but HCOOH production became fast. It was interpreted as the result of the decreasing the proton concentration ([H + ]) in the reaction solution by consumption of the proton during the reduction. Actually, the reactions in the buffered solution with H 3 PO 4 -NaOH exhibited that the production speeds of the CO 2 reduction were unchanged during the reactions. It was the decisive result that HCOOH selectively produced when phenol with the high pKa (ca. 9.95) was used as the proton source. These results suggest that there is an acid-base equilibrium between two intermediates in which one is for CO production and another for HCOOH. Thus, the mechanism involving the equilibrium among the carbonyl complex [Ru(bpy) 2 (CO) 2 ] 2+ , the carboxylic acid complex [Ru(bpy) 2 (CO)(C(O)OH)] + and the CO 2 adduct complex [Ru(bpy) 2 (CO) (CO 2 )] was proposed for the catalytic CO 2 reduction (Figure 7, left cycle) [58,62]. All the complexes were isolated, and the crystal structures were characterized [83]. In the mechanism, [Ru(bpy) 2 (CO) 2 ] 2+ is reduced to yield the coordinated unsaturated species [Ru(bpy) 2 (CO)] with evolving CO. The five coordinated complex reacts with CO 2 to afford the η 1 -CO 2 adduct complex, [Ru(bpy) 2 (CO)(CO 2 )], in which CO 2 coordinates to the metal center at the carbon atom. The electronic structure of the CO 2 bound complex still remains unknown. In the original report [62], it is drawn as [Ru(bpy) 2 (CO)(COO − )] + in which as electron localizes on the CO 2 ligand. In Figure 7, it is drawn as [Ru(bpy) 2 (CO)(CO 2 )] 0 which is the resonance structure of [Ru(bpy) 2 (CO)(COO − )] + . The CO 2 adduct complex is protonated to give the carboxylic acid complex [Ru(bpy) 2 (CO)(C(O) OH)] + and further protonated to recover the carbonyl complex [Ru(bpy) 2 (CO) 2 ] 2+ . The carboxylic acid complex could be reduced to yield HCOOH, and the carbonyl complex to produce CO. The proposed idea reasonably elucidates the experimental results that the catalytic CO 2 reduction gives CO and HCOOH under protic and less protic conditions, respectively. This idea is also supported by the result that the ruthenium complex derivatives give the CO/HCOO − selectivity depending on the different equilibrium constants [58]. The carbonyl complex reacts with dimethylamine to afford the carbamoyl complex [Ru(bpy) 2 (CO)(C(O)N(CH 3 ) 2 )] + , and the electrochemical CO 2 reduction in the presence of dimethylamine produces N,N-dimethylformamide (DMF) [84]. It is also an evidence which the carbonyl complex would exist in the catalysis.   [85]. In the conversion, CO 2 is inserted into the Ru-H bond. The formate complex can release formate ion (HCOO − ) and is considered to be an intermediate for HCOO − production. Based on the results, the hydride mechanism is proposed (Figure 7, right cycle). In the mechanism, the coordinated unsaturated species [Ru(bpy) 2 (CO)] does not react with CO 2 but a proton to yield the hydride complex. The hydride mechanism reasonably explains the CO 2 reduction to produce HCOO − . However, it has a couple of problems [16]. One is that the mechanism is difficult to elucidate the CO production. Production of HCOO − may occur through the hydride mechanism, while CO may produce through the M-CO 2 adduct mechanism. In this case, the product selectivity (CO/HCOO − ) should be controlled by the reactivity difference between CO and H + with the coordinated unsaturated complex. Under the protic conditions, the selectivity of HCOO − production should be enhanced; however, the selectivity of the catalyses gives the opposite tendency. Thus, the pH in the solution or the pKa value of the proton source dependence on the electrochemical CO 2 reduction cannot be explained. Another is that the ruthenium catalyst does not evolve H 2 so much in the CO 2 reduction. It suggests that the catalyst intermediate strongly binds with CO 2 rather than H + .
Nevertheless, the hydride mechanism is supported by many researchers. It is because there are many research works on the CO 2 insertion into Metal-H bonds to afford the corresponding metal formate complexes. On the other hand, the research works on the carboxylic acid complex are fewer, and no mechanical pathways of HCOO − production from the carboxylic acid complex are not understood on the molecular levels.

Electrocatalytic CO 2 reduction by trans(cl)-[Ru(bpy)(CO) 2 Cl 2 ]
Trans(Cl)-[Ru(bpy)(CO) 2 Cl 2 ] is known to be an efficient catalyst for electrochemical CO 2 reduction [58]. The catalytic activity and the product selectivity are similar as these of cis-[Ru(bpy) 2 (CO) 2 ] 2+ . Reduction of trans(Cl)-[Ru(bpy)(CO) 2 Cl 2 ] induces to release Cl − ion to afford the coordinated unsaturated complex. This complex is considered to an intermediate which can bind with CO 2 ; however, it induces polymerization in the absence of CO 2 as shown in Figure 8 [86,87]. The polymer with Ru(0)-Ru(0) bonds is also an efficient electrocatalyst for CO 2 reduction [50,54]. The complex is electrochemically reduced to polymerize on the cathode electrode.
The electrode modified with the polymer is moved to another electric cell, and it works in the presence of CO 2 as the active electrode for electrochemical CO 2 reduction. Researches to make the modified electrode stable have been actively done: introduction of pyrrole groups to the bipyridyl ligand also yields pyrrole polymers to stabilize the ruthenium polymer.
The catalytic reaction mechanisms are also unknown but are considered similar as these of

cis-[Ru(bpy) 2 (CO) 2 ] 2+ . Machan et al. reported the electrochemical CO 2 reduction catalyzed by trans(Cl)-[Ru(6Mesbpy)(CO) 2 Cl 2 ]
, which has two bulky groups at 6,6′-positions in 2,2′-bipyridine [35]. The complex does not polymerize because of the steric hindrance. They discussed the reaction mechanisms based on the hydride mechanism (Figure 9). The precursor complex is at the center of the scheme. It is reduced with releasing Cl − ion to yield the coordinated unsaturated species, which does not bind with CO 2 but H + to afford the hydride complex. The hydride complex reacts with CO 2 to yield the formate complex, which is reduced to produce HCOO − with recovering the original complex. However, the catalyst mainly produces CO not HCOO − . In the mechanism, the formate complex converts to the carboxylic acid complex and then the carbonyl complex by dehydration. The conversion of the formate complex to the carbonyl complex via the carboxylic acid complex is not known, and therefore further researches are expected.
Homogeneous catalysts are advantageous from the viewpoints of elucidating the catalytic reaction mechanisms compared to heterogeneous ones because the homogenous catalysts can be examined by using many spectroscopic techniques. Nevertheless, the mechanisms of the electrochemical CO 2 reduction catalyzed by the ruthenium complexes still remain unknown. There may be potentially many intermediates and pathways in the catalyses, and they depend on the reaction conditions and the subtle difference among the catalyst structures [16,37].

Photochemical CO 2 reduction
In the preceding section, the electrocatalytic activities of the ruthenium complexes are introduced. The electrocatalyst can be utilized in photocatalytic systems by combining with a photosensitizer (PS). Figure 10 shows a schematic drawing of the photocatalytic system, in which the excited PS (PS*) receives an electron from an electron donor to afford the one-electron reduced PS (PS − ). The PS − is the more powerful reagent than PS*, and it can inject an electron to the electrocatalyst. The catalyst can work similarly as the electroreduction occurs. In this section, the photocatalytic CO 2 reduction by the ruthenium complexes is expounded.

Photosensitizer and sacrificial electron donors
The most common photosensitizer used in photocatalytic CO 2 reduction is [Ru(bpy) 3 ] 2+ and the derivatives. Figure 11 shows the absorption and emission spectra of [Ru(bpy) 3 ] 2+ in acetonitrile. The complex exhibits an absorption band at 400-500 nm, which is assignable to metal-to-ligand charge transfer (MLCT). When excited at the band, the emission at the longer wavelengths is observed. The emission is not fluorescence but room-temperature phosphorescence, which is sensitive to O 2 . Therefore, the emission spectrum should be carefully measured under deaerated conditions [88]. The lifetime of the excited state of [Ru(bpy) 3 ] 2+ is 1.10 μs in acetonitrile [89,90]. The quantum yield has been recently reevaluated as 0.095 in acetonitrile [91]. The oxidation potential (corresponding to the reducing ability) of the excited state (PS*) is −0.81 V vs. SCE (CH 3 CN), while this of the one-electron reduced species (PS − ) is −1.33 V. As the electrochemical Figure 10. A schematic drawing of photocatalytic reduction by combining a photosensitizer (PS) with an electrocatalyst (Cat.). CO 2 reduction catalyzed by the ruthenium complexes proceeds under electrolysis at −1.30 V vs. SCE, it requires the reducing ability of PS − . In general, the CO 2 reduction requires higher energy than H 2 production by reduction of H 2 O, and therefore, the photocatalytic CO 2 reduction does not utilize the excited state but the one-electron reduced species.
To generate the one-electron reduced species PS − , the electron donors can reductively quench the excited state of the photosensitizer. As the reduction potential of the excited state of [Ru(bpy) 3 ] 2+ is +0.77 V vs. SCE (CH 3 CN), the electron donors which can be oxidized at less positive potentials than +0.77 V. Figure 12 shows the examples of the electron donors which are actually used in photocatalytic CO 2 reduction [16,92]. Ascorbate ion (AscH − ) can be used in aqueous solution, but amines (triethylamine (TEA) and triethanolamine (TEOA)) cannot work in the presence of water because they are protonated to afford the ammonium ions which cannot give an electron. 1-Benzyl-1,4-dihydronicotineamide (BNAH) is a model compound of NADH in nature. NADH is a two-electron donor and is oxidized to yield NAD + . However, the model compound BNAH cannot give two electrons in the oxidation by the excited state of [Ru(bpy) 3 ] 2+ but provides one electron to afford the dimer BNA 2 . 1,3-Dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH) and the derivatives (e.g., BI(OH)H), which have much stronger reducing power than BNAH, have been recently utilized in photocatalytic CO 2 reduction. BIH provides two electrons to yield the oxidation product BI + .
These electron donors are called the sacrificial reagents because the one-electron oxidized species occur chemical changes or decompose so as to prevent back electron transfer. They are useful in order to investigate the reductive half reaction. However, from the viewpoint of the energy balance, the reduction-oxidation (redox) systems in which water is oxidized and CO 2 is reduced are desired.

Photocatalytic CO 2 reduction
Our group have investigated the photochemical CO 2 reduction by the system consisting of trans(Cl)-[Ru(bpy)(CO) 2 Cl 2 ], [Ru(bpy) 3 ] 2+ and BNAH as the catalyst, the photosensitizer, and the electron donor, respectively (Figure 13). The catalysis had been carried out in N,Ndimethylformamide (DMF)/water [57,59]. However, it was indicated that hydration of DMF affording formate became a serious problem in quantifying formate [93]. We proposed the use of N,N-dimethylacetamide (DMA), of which the dehydration does not produce formate but acetate, instead of DMF [39]. Although the photocatalysis strongly depends on the solvent system, the reaction proceeds in DMA/water similarly as in DMF/water.  The catalytic reaction proceeds by receiving electrons from the photochemically driven electron relay system. For two-electron reduction of CO 2 to CO or HCOOH, the electron relay cycle has to go round two times when the catalytic cycle turns one time. The electron source is not an electrode, but the reaction had been supposed to proceed according to the same mechanism as in electrochemical reduction. However, it has been recently indicated that in some cases, the reaction mechanisms of the photochemical CO 2 reduction are likely different from the electrochemical one [16]. For example, unusual catalyst concentration dependence on the product selectivity (CO/HCOO − ) in the photocatalysis has been observed: at high catalyst concentration the selectivity of HCOO − increases [37]. To elucidate the peculiar catalyst concentration effect, the mechanisms as shown in the right cycle in Figure 14 are proposed. At the high concentration of the catalyst, the reduced catalyst forms a dimer, which is proposed to selectively afford HCOO − . The dimer of the complex is similar as the intermediate of polymerization, but it is not detected in the photocatalytic system because the absorption spectrum cannot be conformed due to the overlapped absorption of [Ru(bpy) 3 ] 2+ . Alternatively, the photocatalytic CO 2 reduction by trans(Cl)-[Ru(6Mesbpy)(CO) 2 Cl 2 ] which does not dimerize because of the steric hindrance of the ligand has been examined. The ruthenium complex selectively produces CO in the photochemical CO 2 reduction, and it demonstrates that the dimerization of the catalyst relates on the HCOO − production. It is suggested that the catalyst concentration dependence is not observed in a DMA/ethanol solution. It indicates that HCOO − also produces in the cycle consisting of mono-nuclear ruthenium complexes as proposed for the electrocatalytic CO 2 reduction (Figure 14, left cycle) [31]. The photochemical CO 2 reduction catalyzed by trans(Cl)-[Ru(2,2′-bipyridine)(CO) 2 Cl 2 ] bearing two methyl groups at 4,4′-or 5,5′-positions in the ligand has been recently reported [64].
As the catalytic activities of these complexes at low catalyst concentrations are almost the same, the intrinsic activities are considered to be identical. However, the catalytic activities of these complexes are different at high catalyst concentration, where the rate-determining step is not in the catalytic cycle but in the electron relay cycle: the ruthenium complex with dimethyl groups at 5,5′-positions in the 2,2′-bipyridyl ligand is higher than that at 4,4′-positions. The efficiency of the back-electron transfer from the reduced catalyst to the photosensitizer is lower, or the cage escape yield for the sensitizer-catalyst complex is higher in the 5,5′-dimethyl complex than in the 4,4′-complex.
These phenomena have not been observed in electrochemical CO 2 reduction. It is probably because that the homogenous photocatalytic CO 2 reduction contains the diffusion process of the electron relay between the photosensitizer and the catalyst. The speed of the electron supply also sometimes affects the reaction mechanisms [16,37].

Application to heterogeneous catalysts
Heterogeneous catalysts are industrially important because they are useful for separating the starting materials and the products from the catalyst and can be recovered and reused. The molecular catalysts can be utilized to develop the heterogeneous catalysts. For photocatalysts of CO 2 reduction, combining the molecular catalysts with semiconductor [32,94,95], metalorganic frameworks (MOFs) [96,97] or periodic mesoporous organosilicas (PMOs) [98][99][100][101]  are actively researched. We have also developed a novel PMO consisting of 2,2′-bipyridyl framework by introducing two different ruthenium complexes as a photosensitizing site (Ru(PS)) and a catalytic site (Ru(Cat)) as shown in Figure 15 [99]. Photochemical CO 2 reduction by the PMO catalyst has catalytically produced CO and formate. The product selectivity (CO/formate) becomes large with increasing the ratio of Ru(PS) to Ru(Cat) (x/y). The photocatalysts can be recycled at least three times without losing the catalytic activity, demonstrating that the Ru(PS) and Ru(Cat) units are strongly immobilized on the BPy-PMO framework.

Future prospects
The molecular catalysts are applicable to various photocatalytic systems. Ultimately, our goal is to construct an artificial photosynthetic system. An example is shown in Figure 16.
In the system, the electrons are not supplied from the sacrificial electron donor but from water which is the same as in natural photosynthetic system. As the CO 2 reduction requires a high potential, two photosensitizing systems would be combined as the Z-scheme mechanism in the natural photosynthesis. In order to realize the artificial photosynthesis, we have to overcome some problems. One is to perform these reactions (water oxidation, photo-induced electron transfer and CO 2 reduction, etc.) under the similar conditions or in the separated circumstances. Another is to match the velocities among the reactions; even if the efficient catalyst for CO 2 reduction was obtained, the speeds for the water oxidation and the electron supply have to match with that of CO 2 reduction. There would be many other problems to construct the artificial photosynthesis. However, the real system which can efficiently work has already existed in nature. We will realize it with a lot of ideas to overcome many problems one by one.
[42] Tamaki Y, Morimoto T, Koike K, Ishitani O. Photocatalytic CO 2 reduction with high turnover frequency and selectivity of formic acid formation using Ru(II) multinuclear complexes. Proceedings of the National Academy of Sciences of the United States of America. 2012;109:15673-15678. DOI: 10.1073/pnas.1118336109