Photocatalytic conversion of CO₂ to CO by Ru(II) and Os(II) octahedral complexes: a DFT/TDDFT study

Abstract

The reaction mechanisms of the photocatalytic reduction of CO₂ to CO catalyzed by [(en)M(CO)₃Cl] complexes (M = Ru, Os, en = ethylenediamine) in the presence of triethanolamine (TEOA), R₃N (R = –CH₂CH₂OH), in DCM and DMF solvents, were studied by means of DFT/TDDFT electronic structure calculations. The geometric and free energy reaction profiles for two possible reaction pathways were calculated. Both reaction pathways studied, start with the 17e⁻, catalytically active intermediate, [(en)M(CO)₃]˙⁺ generated from the first triplet excited state, T₁ upon reductive quenching by TEOA which acts as a sacrificial electron donor. In the first possible pathway, TEOA⁻ anion binds to the metal center of the catalytically active intermediate, [(en)M(CO)₃]˙⁺ followed by CO₂ insertion into the M-OCH₂CH₂NR₂ bond. The latter upon successive protonations releases a metal ‘free’ [R₂NCH₂CH₂OC(O)(OH)] intermediate which starts a new and final catalytic cycle, leading to the formation of CO and H₂O while regenarating TEOA. In the second possible pathway, the 17e⁻, catalytically active intermediate, [(en)M(CO)₃]˙⁺ captures CO₂ molecule, forming an η¹-CO₂ complex. Upon 2H⁺/2e⁻ successive protonations and reductions, CO product is obtained along with regenarating the catalytically active intermediate [(en)M(CO)₃]˙⁺. The nature of the proton donor affects the reaction profiles of both mechanisms. The nature of the solvent does not affect significantly the reaction mechanisms under study. Finally, since photoexcitation and T₁ reductive quenching are common to both pathways, we have srutinized the photophysical properties of the [(en)M(CO)₃Cl] complexes along with their T₁ excited states reduction potentials, . The [(en)M(CO)₃Cl] complexes absorb mainly in the UV region while the absolute are in the range 6.4–0.9 eV.

---

1. Introduction

The burning of fossil fuels is the main source of carbon dioxide, CO₂ released into the atmosphere by anthropogenic activity. Eighty per cent, 80% of the concentration of greenhouse gasses in the atmosphere is due to the human factor and more specifically to combustion for energy production. The energy sector, is to a large extent, based on the direct combustion of fuels, a process that in turn leads to large CO₂ emissions.¹ Carbon dioxide is one of the most important greenhouse gasses that contributes to the strengthening of this phenomenon on our planet. The main greenhouse gases are the following: carbon dioxide (CO₂), water vapor (H₂O), methane (CH₄), ozone (O₃), nitrogen dioxide (N₂O) and chlorofluorocarbons (CFC). Of these greenhouse gases, especially water vapor and carbon dioxide play a decisive role in regulating the temperature of the earth's surface and its atmosphere.² Many studies have been devoted to the way in which the problem of the continuous increase in atmospheric CO₂ can be resolved while at the same time there are continuous efforts and studies not only to bind CO₂ but at the same time to convert it into “fine chemicals”.³

Among various strategies to achieve this goal, the biggest and most attractive challenge is the efficient conversion of CO₂ into useful compounds using sunlight as an energy source.⁴ Many complexes with a transition metal element as metal center (such as Ni, Fe, Re, Cr, Ir, Mo etc.) have been extensively studied for the homogeneous electrocatalytic and photocatalytic reduction of carbon dioxide.^5–16

The majority of research has focused on the photocatalytic reduction of CO₂, which has been based on systems that include a light unit consisting of a photosensitizer, (PS) and two catalytic sites (a reduction site and an oxidation site) (Scheme 1).^{7–10,13–16}

Scheme 1 Photocatalytic CO₂ reduction by a system comprising a PS and two catalytic sites.

At the oxidation site the donor provides an e⁻ to the PS after its excitation to the triplet excited state (³MLCT) which is then reductively quenched at the reduction site and finally an e⁻ is transferred to CO₂. In many cases, however, PS acts not only as a PS but also as a reducing agent.¹⁷

So far, the majority of research efforts on the electro/photo-catalytic reduction of CO₂, has been put on 4d and 5d transition metal polypyridyl complexes as catalysts.¹³ Among them, Re and Ru polypyridyl complexes have attracted the major concern, being the subject of numerous studies. In pioneering works,^18–20 Lehn et al., studied the electro/photo-catalytic reduction of CO₂ to CO by Re(I) and Ru(II) bipyridyl complexes. Although, numerous mechanisms have been proposed for the photochemical reduction of CO₂ catalyzed by Re(I) and Ru(II) complexes, none of them could be regarded as universal.¹⁶ Among the various mechanisms proposed, the most widespread is based on the formation of either [Re(bpy)(CO)₃] or [Re(bpy)(CO)₃]˙⁻ very reactive, five-coordinated key intermediates, able to capture CO₂.^21,22 Another proposed mechanism, based upon experimental²³ as well as theoretical²⁴ studies, involves the formation of a Re(I) dimer species i.e. [Re(dmb)(CO)₃]₂(OCO), (dmb = 4,4′-dimethyl-2,2′-bipyridine) intermediate. In this dimeric Re(I) complex, from which CO is released, the CO₂ molecule bridges the two metal centres. Notice however, that the theoretical study²⁴ points out that formation of the Re(I) dimer is entropically unfavourable, while experiment²³ showed that this process is relatively slow. Nevertheless, the mechanism of CO₂ to CO reductions by Re(I) still remains elusive, especially the CO₂ addition to the metal centre.¹⁶ Ishitani et al.,²³ proposed that even the non-reduced [Re^I(N^N)(CO)₃]⁺ 17e⁻ species is able to capture CO₂ with the assistance of TEOA. Accordingly, they found that during the photocatalytic reduction of CO₂ by the fac-[Re^I(bpy)(Br)] complex, TEOA is coordinated to the metal centre of the non-reduced [Re^I(bpy)(CO)₃]⁺ 17e⁻ intermediate forming another intermediate namely the fac-[Re^I(bpy)(CO)₃(OCH₂CH₂NR₂)] (R = CH₂CH₂OH). In a subsequent step, CO₂ is inserted into the Re–O coordination bond of the fac-[Re^I(bpy)(CO)₃(OCH₂CH₂NR₂)] intermediate yielding the fac-[Re^I(bpy)(CO)₃(R₂NCH₂CH₂O-COO)] insertion product, which is proposed to be the key intermediate for the CO₂ photocatalytic reduction.²³

The other most studied catalysts for electro/photo-catalytic CO₂ conversion are the Ru(II) polypyridyl complexes. The light-induced reduction of CO₂ to CO in an acetonitrile/water/triethanolamine solution in the presence of the [Ru(2,2′-bipyridine)₃]²⁺/Co²⁺ system was first reported by Lehn and Ziessel.¹⁸ Since then, numerous studies have been devoted to the CO₂ reduction by Ru(II) catalysts. The most representative complexes studied are the cis-[Ru(bpy)₂(CO)X]ⁿ⁺, [Ru(tpy)(bpy)X]ⁿ⁺ and trans (Cl)-Ru(bpy)(CO)₂Cl₂ complexes as well as their derivatives.¹⁶

The earliest study of electrochemical CO₂ reduction by a Ru(II) complex was reported by Tanaka et al.,²⁵ which used a homogeneous [Ru(bpy)₂(CO)₂]²⁺ catalytic system to obtain a mixture of CO/formate from CO₂ upon electrolysis. The proposed mechanism, involves formation of an unsaturated five coordinate species which after reduction is attacked by CO₂ to form the [Ru(bpy)₂(CO)(η¹-CO₂)] intermediate (η¹-CO₂ adduct). Formation of the η¹-CO₂ adduct is thought to occur also in the CO₂ electrocatalytic reduction by the cis-[Ru(bpy)₂(CO)Cl]⁺ catalysts, which after reduction, loses the chloride ligand forming the unsaturated five coordinate species which in turn is attacked by CO₂.²⁶ An alternate combined η¹-CO₂ complex/CO₂ insertion (hydride) mechanism for CO₂ reduction has been proposed for Ru(II)-hydride catalysts.^16,27–29 Finally, Os(II) complexes are far less studied either as photosensitizers^30,31 or catalysts^32,33 for CO₂ conversion.

Taking into account the scarcity of theoretical studies of CO₂ to CO photocatalytic conversion by Ru or Os complexes we instigated to contact a theoretical study, by means of DFT/TDDFT electronic structure calculations, in order to delineate the mechanistic details of the CO₂ to CO photocatalytic conversion by simple Ru(II) and Os(II) octahedral model complexes of the general formula [(en)M(CO)₃Cl] (en = ethylenediamine, M = Ru, Os) in the presence of TEOA. The aim of our work is twofold: (a) to expand our previous molecular modeling, mechanistic investigation on the photocatalytic CO₂ to CO reduction by Re(I) complexes¹⁷ to cover also Ru(II) and Os(II) analogue complexes and (b) to theoretically delineate the possibility for the mechanism proposed by Ishitani's seminal work²³ on similar Re(I) bipy octahedral complexes to be valid also for Ru(II)/Os(II) complexes. Solvent effects on the reaction mechanisms of the CO₂ to CO conversion by [(en)M(CO)₃Cl] complexes, were also studied, upon conducting our calculations in the non polar DCM solvent as well as in the polar DMF solvent. Finally, we set out to explore the photophysical properties of these complexes as well as their excited state electrochemistry.

2. Computational details

The geometries of all species have been fully optimized, without symmetry constraints, using the 1997 hybrid functional of Perdew, Burke and Ernzerhof^34–39 as implemented in the program Gaussian16.⁴⁰ This functional uses 25% exchange and 75% weighting correlation and is denoted as PBE0. The Def2-TZVP basis set for all atoms was used for the geometry optimizations. The method used in the DFT calculations will be abbreviated PBE0/Def2-TZVP. All stable structures have been identified as energy minima with number of imaginary frequencies N_Imag = 0. In detail, the frequencies were calculated at the same level as in theory, and the nature of the fixed points was determined in each case according to the number of negative eigenvalues of the Hessian. The Gibbs free energy was calculated to be 298.15 K and 1 atm pressure. Solvent effects were calculated via the Polarizable Continuum Model (PCM) using the integral equation formalism variant (IEF-PCM), which is the default method (self-consistent reaction field (SCRF)).⁴¹ DCM and the more polar solvent DMF were used as solvents. Natural Bond Orbital (NBO) population analysis was performed using the Weinhold methodology.^42,43 Time-dependent density functional theory (TD-DFT) calculations^44–47 were performed on the ground-state, S₀ equilibrium geometries in DCM and DMF solvents using the PBE0/Def2-TZVP/PCM computational protocol, taking account the first 30 excited states.

3. Results and discussion

Based upon previous experimental as well as theoretical evidences¹⁶ we employed DFT electronic structure calculations in order to explore two possible reaction mechanisms for the photocatalytic CO₂ conversion by [(en)M(CO)₃Cl] octahedral complexes. The two possible mechanisms under investigation are: (a) The CO₂ to CO conversion, based on experimental studies by Ishitani et al.,²³ on Re(I) analogue complexes, where TEOA is thought to act, not only as a ‘sacrificial donor’, but also as a ligand as well, coordinated to the metal center of the 17e⁻ very reactive intermediate and (b) a more conventional mechanism where the 17e⁻ intermediate, after reduction, captures directly CO₂, to form the [(en)M(CO)₃(η¹-CO₂)] intermediate (η¹-CO₂ adduct).¹⁶ Let us start however, with the computational study of the initial step of the reaction, being common to both possible reaction mechanisms under consideration, namely the photo-excitation and subsequent reductive quenching processes of the starting photocatalyst (or pre-catalyst).

3.1 Initial step

The initial step of the photocatalytic CO₂ to CO conversion by the [(en)M(CO)₃Cl] complexes, is depicted schematically in Fig. 1.

Fig. 1 Initial step of the photocatalytic CO₂ to CO conversion by the [(en)M(CO)₃Cl] complexes.

In the initial step, the [(en)M(CO)₃Cl] complexes, upon irradiation, are excited in their first singlet excited state, S₁ which via Intersystem Crossing (ISC) yields their first triple excited states, T₁. Finally, the [(en)M(CO)₃Cl] complexes in their T₁ state, receive an e⁻ from TEOA forming the so called One Electron Reduced (OER), 19e⁻ complex, OER_M.

The optimized geometries of all species in DCM, involved in the initial step of the CO₂ → CO conversion by [(en)Ru(CO)₃Cl], 1 and [(en)Os(CO)₃Cl], 2 complexes are depicted in Fig. 2 while those in DMF are given in Fig. S1 (see ESI).† Perusal of Fig. 2 and ES1† reveals that the structural parameters of 1 and 2 in their S₀ ground states are not affected and practically remain the same in either the non polar DCM solvent or in the polar DMF solvent. The same holds true also for 1 and 2 in their T₁ states. Upon S₀ → T₁ excitation, the most striking structural change is observed for the axial CO ligand i.e. the Ru–CO bond is significantly elongated by 0.2–0.4 Å while the ∠Ru–C–O bond angle departs from linearity by up to 47°. Also, there is a small change of the Ru–CO bonds of the equatorial CO ligands in the range 0.01–0.07 Å while the ∠CO_ax–Ru–CO_ax and ∠CO_ax–Ru–CO_eq bond angles change is in the range 1.9–5.4°. The rest of the structural parameters around the coordination sphere exhibit only minor changes.

Fig. 2 Optimized geometries of 1 and 2 in their S₀ and T₁ states and respective OER species in DCM solvent at the PBE0/Def2-TZVP level.

On the other hand, the optimized geometry of the OER_M species in the two solvents, DCM and DMF remains practically unaltered. However, the structural parameters in the coordination sphere of OER_M species are significantly changed as compared to those found for the respective S₀ and T₁ states. Accordingly, the Ru–Cl bond in OER_M is elongated as compared to the respective bond in the S₀ and T₁ states. The same holds also for the two Ru–N bonds formed between the metal centers and the en ligand. In contrast, one Ru–CO_eq equatorial bond in OER_M is lengthen as compared to both the S₀ and T₁ states. This equatorial CO_eq ligand is also found to be significantly bended compared to S₀ and T₁ states. The other Ru–CO_eq equatorial bond in OER_M is shortened while retaining linearity. Finally, the Ru–CO_ax axial bond in OER_M is similar to that of S₀ state while both differ significantly compared to that in T₁ state. The bond angles around the coordination sphere in OER_M are similar to those found for the S₀ rather the T₁ state.

3.1.1 Photophysical properties. Since excitation of the initial photocatalyst/precatalyst, upon light irradiation, is a prerequisite in order to obtain the 17e⁻ catalytically active intermediate, we set out to calculate the respective absorption spectra. In Fig. 3 are depicted schematically the simulated absorption spectraof 1 and 2 in both DCM and DMF solvents.

Fig. 3 Simulated absorption spectra of 1 and 2 in DCM and DMF solvents at the PBE0/Def2-TZVP level.

In Table 1 are given the respective electronic transitions related to the simulated absorption spectra of 1 and 2. Inspection of Fig. 3 and Table 1 reveals that the two complexes in either DCM or DMF solvents, absorb in the UV region. Upon replacing the non-polar DCM solvent with the polar DMF solvent, has practically no effect on the absorption spectra which remain unaltered. The absorption spectra of 1, in either DCM or DMF solvents, exhibit one band with peak around 180 nm and a shoulder at around 250 nm. The band arises mainly from an electronic excitation at 179 nm which based upon the shapes of the MOs involved in the respective electronic excitation (Fig. S2†), could be assigned as MLCT/MC/IL. The same holds also true for the shoulder peaking around 250 nm.

Table 1 Principal singlet–singlet electronic transitions in the simulated absorption spectra, for 1 and 2 calculated in DCM and DMF solvents at the PBE0-GD3BJ/Def2-TZVP level of theory

Excitation (%composition)	E (eV)	λ (nm)	f
1_DCM
H−5 → L (32%), H−3 → L (23%), H−3 → L+2 (11%)	6.92	179	0.354
H−1 → L+1 (70%)	4.83	257	0.015
1_DMF
H−5 → L (43%), H−3 → L (15%)	6.96	178	0.373
H−1 → L+1 (73%)	4.86	255	0.014
2_DCM
H−4 → L+2 (12%), H−3 → L+2 (30%), H−2 → L+7 (13%)	7.32	169	0.060
H−5 → L+1 (15%), H−2 → L+6 (48%)	7.25	171	0.081
H−2 → L+4 (25%), H−2 → L+5 (10%), H−2 → L+6 (25%)	7.16	173	0.074
H−1 → L+7 (19%), H → L+7 (20%)	6.94	179	0.110
H−1 → L+2 (57%), H → L+2 (18%)	5.83	213	0.014
H−2 → L (36%), H−2 → L+1 (36%)	5.79	214	0.020
H−1 → L+1 (53%), H → L (30%)	5.13	242	0.043
2_DMF
H−4 → L+2 (15%), H−3 → L+2 (36%)	7.33	169	0.057
H−5 → L (12%), H−2 → L+6 (43%)	7.26	171	0.067
H−2 → L+4 (−23%), H−2 → L+6 (33%)	7.17	173	0.064
H−1 → L+7 (17%), H → L+7 (20%)	6.97	178	0.113
H−2 → L (14%), H−2 → L+1 (57%)	5.79	214	0.020
H−1 → L+1 (54%), H → L (22%)	5.16	240	0.042

---

The simulated absorption spectra of 2 in DCM and DMF solvents exhibit one high energy strong band, peaking around 170 nm and a lower energy, weak band at around 240 nm. In addition, there is a shoulder around 215 nm. Based upon the shapes of the MOs involved in the electronic excitations (Fig. S2†) the bands and the shoulder are assigned as MLCT/MC/IL.

Since the T₁ excited states of the complexes under study is of central importance in Step A, we set out to explore their properties in more detail. In Fig. 4 are depicted the spin density along with the SOMOs for the T₁ excited state of 1 and 2.

Fig. 4 (a) Spin density isosurfaces (0.005 au) along with atomic spin densities of T₁ state of complexes 1 and 2 in DCM solvent (numbers in parenthesis refer to DMF solvent) and (b) 3D contour plots of the relevant SOMOs at the PBE0-GD3BJ/Def2-TZVP level of theory.

Inspection of Fig. 4 reveals that the spin density in the T₁ state of 1 and 2 is mainly accumulated on the metal center and to a lesser degree on the axial chloride and carbonyl ligands. The spin density distribution resembles the 3D shapes of the SOMOs of 1 and 2 in the T₁ state.

3.1.2 Excited state redox potentials. Taken into account that formation of the 17e⁻ catalytically active intermediate is preceded by the reductive quenching of the first triplet excited state, T₁ of the pre-catalyst, we set out to calculate the redox potential for this excited state. The T₁ excited-state redox potentials were estimated upon employing the procedure followed by Hansen et al.⁴⁸ First, we calculated the ground-state reduction potentials, based upon the Born–Haber cycle shown in Scheme 2. Then, we calculated the excited state reduction potential, according to the Latimer diagram depicted also in Scheme 2.

Scheme 2 Born–Haber cycle employed in the calculation of the ground state reduction potentials, (above) and Latimer diagram employed for the calculation of T₁ excited state reduction potentials, (below).

The standard absolute ground-state reduction potential, is calculated by the following equation:

where F is the Faraday constant (23.061 kcal per volt gram equivalent) and Z is unity for one-electron redox processes. ΔG°(soln., redox) is obtained from the following equation:

The calculated Gibbs free energies appearing in Scheme 2 are given in Tables S1 and S2 of the ESI† along with 0–0 transition energies for the T₁ states, E_0–0 in DCM and DMF solvents. The E_0–0 are obtained as the differences of the zero point energy corrected total electronic energies of 1 and 2 in their S₀ and T₁ states in DCM and DMF solvents. The calculated T₁ excited state absolute reduction potentials, are 6.4 and 6.3 V for 1 in DCM and DMF respectively, while for 2 are 6.4 and 0.9 V in DCM and DMF respectively. Taking into account that SHE potential is −4.281 V,⁴⁹ the vs. SHE for 1 in DCM and DMF are 2.1 and 2.0 Volt respectively while for 2 in DCM and DMF are 2.0 and −3.4 V.

3.2 Catalytic cycle

After the photoexcitation and reductive quenching processes in the Initial step of the photocatalytic CO₂ to CO conversion, it follows the ‘actual’ catalytic cycle, starting with the five coordinated, 17e⁻ intermediate. We examined two possible mechanisms for the catalytic cycle. The first possible mechanism is based on the concept of the formation of a complex between the 17e⁻ catalytically active intermediate and TEOA, acting as a ligand. Henceforth, we will call this the TEOA complex mechanism. The second possible mechanism under investigation, is based on the direct CO₂ capture by the 17e⁻ catalytically active intermediate which henceforth we will call it η¹-CO₂ complex mechanism.

3.2.1 TEOA complex mechanism. This proposed mechanism is based on the formation of a complex between TEOA and the 17e⁻ five coordinated intermediate obtained from OER_M. This concept is thought^23,50 to be important for the CO₂ reduction in low concentrations e.g. in the air, and its utilization from dilute CO₂ sources. However, obtaining CO via this mechanism still remains elusive. Accordingly, we employed DFT electronic structure calculations to study in more depth the concept of the TEOA complex mechanism applying it for the Ru(II) and Os(II) catalysts under consideration. The proposed catalytic cycle is depicted in Fig. 5 and comprising two steps: (a) the TEOA complex formation step and (b) the CO release step.

Fig. 5 Proposed TEOA complex mechanism for CO₂ to CO conversion by 1 and 2.

3.2.1.1 TEOA complex formation step. The TEOA complex formation step starts with the 17e⁻ five coordinated, very reactive intermediates, ImA_M, obtained from OER_M, and considered to be the ‘true’ catalysts. Next, TEOA upon deprotonation, is coordinated, as an anionic ligand, TEOA⁻, to the metal center of ImA_M, forming the [(en)(CO)₃M(OCH₂CH₂NR₂)] intermediates, ImB_M. The CO₂ molecule could then be captured via an insertion reaction, where it is inserted into the M–O bond of ImB_M yielding the [(en)(CO)₃MO(O)COCH₂CH₂NR₂] key intermediates, ImC_M which are of significant importance.²³ The latter, upon protonation, dissociate to form the [R₂NCH₂CH₂OC(O)OH]⁺ species, ImD_M and giving back the ‘true’ catalysts, ImA_M. The free energy, ΔG profiles calculated for the TEOA complex formation cycle are depicted schematically in Fig. 6 and 7 for the Ru(II) complex 1 and Os(II) complex 2 respectively, in both DCM and DMF solvents.

Fig. 6 Free energy, ΔG (in kcal mol⁻¹), reaction profiles of TEOA complex cycle for 1 in DCM and DMF (numbers in parenthesis) solvents calculated at the PBE0/Def2-TZVP level.

Fig. 7 Free energy, ΔG (in kcal mol⁻¹), reaction profiles of TEOA complex cycle for 1 in DCM and DMF (numbers in parenthesis) solvents calculated at the PBE0/Def2-TZVP level.

Notice that, in the mechanistic profiles depicted in Fig. 6 and 7, the ‘magic’ TEOA sacrificial electron donor is thought of acting as the proton donor as well. Perusal of Fig. 6 and 7 reveals that the TEOA complex formation step is expected to be endergonic in the case where TEOA is thought to be the proton donor source. Initially, coordination of TEOA⁻ ligand to the metal center of either ImA_Ru or ImA_Os is exergonic in either DCM or DMF solvents. This process is slightly more favorable for Os(II) complex, 2 as compared to the Ru(II) complex, 1. In addition, the TEOA⁻ coordination to the metal centers is clearly more favorable in DCM rather in DMF solvent. Next, the CO₂ insertion into the M–O bond of ImB_M species proceeds via a transition state, TS₁_M, leading to intermediate ImC_M (Fig. 6 and 7). The calculated energy barrier for the formation of ImC_M from ImB_M is estimated to be in the range 36–42 kcal mol⁻¹ at the PBE0/Def2-TZVP level of theory. It should be noticed that, this process is slightly more favorable for Ru(II) complex, 1 rather for Os(II) complex, 2, since the energy barrier for the former is lower than for the latter. Finally, upon protonation of ImC_M, with TEOA as the proton source, we obtain the ImD intermediate and ImA_Mvia a somewhat endergonic process.

Taking into account that for similar Re(I) complexes the proton obtained from the deprotonated form of TEOA should be trapped by another TEOA molecule²³ we set out to explore this possibility also for the Ru(II) and Os(II) complexes under study. Inspection of Fig. 6 and 7 reveals that protonation with TEOA(H)⁺, as the proton source, is exergonic in contrast to the process where TEOA is used as a proton source.

The relatively weak Ru–O coordination bond, formed upon coordination of the R₂NCH₂CH₂O⁻ (R = CH₂CH₂OH) ligand to the Ru metal center of the [(en)Ru(CO)₃] intermediate, ImA_Ru, is reflected to the relatively small Wiberg Bond Indices (WBI) obtained from the NBO analysis of ImB_Ru, being equal to 0.582 and 0.581 in DCM and DMF solvents respectively.

In addition, NBO population analysis revealed that in DCM and DMF solvents both the Ru metal center and the O donor atom of the R₂NCH₂CH₂O⁻ ligand acquire negative natural atomic charges equal to −0.850 and −0.707|e| respectively, resulting in a repulsive electrostatic interaction. On the other hand, in both DCM and DMF solvents, the covalent component of the Ru–O coordination bond arises from a σ(Ru–O) bonding NBO, formed upon interaction of the sp^4.90d^4.24 hybrid orbitals (48.3% p and 41.7% d character) of Ru with the sp^2.60 hybrid orbital (72.2% p-character) of the oxygen donor atom of R₂NCH₂CH₂O⁻ ligand and is described as σ(Ru–O) = 0.429h_Ru + 0.903h_O. The occupation number of σ(Ru–O) NBO is 0.958|e|. The σ(Ru–O) NBO is depicted schematically in Scheme 3.

Scheme 3 3D surface of the σ(Ru–O) bonding NBO of ImB_Ru.

It follows the insertion of a CO₂ molecule into the Ru–O bond of ImB_Ru yielding ImC_Ruvia an exergonic process with ΔG equal to −31 to −21 kcal mol⁻¹ in DCM and DMF solvents respectively. Formation of ImC_Ru procceds via transition state TS₁_Ru, exhibiting a relatively low energy activation barrier calculated to be 8 and 19 kcal mol⁻¹ in DCM and DMF solvents respectively (Fig. 6). Subsequently, ImC_Ru could undergo protonation at three possible O atom sites of the R₂NCH₂CH₂O(O)CO ligand denoted as O_a, O_b and O_c (Fig. 5). The natural atomic charges on the O_a and O_b atoms are −0.719 and −0.649 respectively in DCM solvent and −0.724 and −0.648 respectively in DMF solvent. Protonation of either O_a or O_b of the R₂NCH₂CH₂OC(O)O ligand leads to cleavage of the Ru–O bond, yielding ImD and regenerating the initial 17e⁻ catalytic species ImA_Ru, thus closing the catalytic cycle (Fig. 5).

The third possible protonation site i.e. O_c acquires a natural atomic charge of only −0.519, thus being less prone to protonation as compared to the other two possible protonation sites, O_a and O_b of ImC_Ru. However, the protonation of the third possible site, O_c yields back ImA_Ru, CO₂ and TEOA.

Protonation of ImC_Ru, with TEOA proton donor, at O_a and O_b sites, leading to Ru–O bond cleavage to yield ImA_Ru and ImD, is estimated to be a slightly endergonic process with ΔG around to 12 kcal mol⁻¹ in both DCM and DMF solvents respectively (Fig. 6). In contrast, if TEOA(H)⁺ is the proton donor, this process is exergonic with ΔG equal to 77 and 62 kcal mol⁻¹ in DCM and DMF solvents respectively.

The calculated free energy reaction profile for 2 is quite similar with that calculated for 1 (Fig. 7). The anionic TEOA⁻ ligand interacts with the Os metal center of the highly active 17e⁻ intermediate ImA_Os, giving the neutral radical [(en)Os(CO)₃(TEOA)]˙ intermediate, ImB_Os. The estimated interaction energy (IE) of TEOA⁻ ligand with the Os metal center of ImB_Os is equal to −44.6 kcal mol⁻¹ in DCM solvent and −34.3 kcal mol⁻¹ in DMF solvent.

Obviously, TEOA⁻ interacts more strongly with Os rather than the Ru metal center of ImA_M. The conversion of ImA_Os to the ImB_Os intermediate is slightly more exergonic (ΔG = −33 and −22 kcal mol⁻¹ in DCM and DMF solvents respectively) as compared to the respective process calculated for the Ru counterparts.

NBO analysis indicates that the TEOA⁻ ligand coordinates to the Os metal center of ImA_Os, forming a relatively weak Os–O bond in the ImB_Os intermediate. The estimated WBI for Os–O bond is 0.528 in DCM solvent and 0.529 in DMF solvent. In DCM solvent both the Os metal center and the O donor atom of the coordinated TEOA⁻ ligand acquire negative natural atomic charges of −0.397 and −0.761|e| respectively. In DMF solvent, the respective charges are −0.400 and −0.757|e| respectively. The electrostatic interaction between TEOA⁻ ligand with Os metal center of ImB_Os is unfavorable similar to that found for the respective Ru intermediate, ImB_Ru.

Formation of ImC_Os for 2i.e. CO₂ insertion into the Os–O(TEOA) bond is calculated to be slightly more exergonic as compared to its Ru counterpart (Fig. 6 and 7). It follows protonation of ImC_Os at either O_a or O_b sites leading ultimately to the formation of ImA_Os and ImE intermediates (Fig. 7). The protonation sites O_a and O_b acquire natural atomic charges of −0.712 and −0.728|e| respectively in DCM solvent and −0.728 and −0.697|e| respectively in DMF solvent. Protonation of ImC_Os, with TEOA, accompanied by the cleavage of the Ru–O bond, is slightly endergonic, similar to that found for the protonation of the respective Ru intermediate, ImC_Ru. In contrast, protonation with TEOA(H)⁺ is exergonic just as for its Ru counterpart.

Notice that, as in the case of 1, there is a third protonation site at O_c (Fig. 5) being less probable however, due to the lower negative natural charge calculated for O_c, being equal to −0.516 and −0.508|e| in DCM and DMF solvents respectively. The protonation of the third possible site, O_c yields back ImA_Os, CO₂ and TEOA similar with the respective process being observed for 1 (vide supra). Finally, the optimized geometries of all species involved in the catalytic cycle of TEOA complex formation step (Fig. 5) along with selected structural parameters are given in Fig. S3 and S4 of the ESI.†

3.2.1.2 CO release step. The second step of the TEOA complex mechanism, starts with the [R₂NCH₂CH₂OC(O)OH]⁺ species, ImD, being the product of protonation of ImC_M during the preceding TEOA complex formation step (Fig. 5). The energetic profile calculated for the CO release step is depicted in Fig. 8. As for the TEOA complex formation step (vide supra), we examined two possible reaction mechanisms i.e. one where TEOA acts as the proton source and another where protonated TEOA, TEOA(H)⁺ is the proton source. There are three possible protonation sites in ImD which are denoted as O_a, O_b and O_c (Fig. 8). The natural atomic charges of the O_a, O_b and O_c atoms of ImD in DCM are −0.644, −0.680 and −0.476|e| respectively while in DMF are −0.648, −0.679 and −0.477|e| respectively. Thus, based upon the natural charges, the most probable protonation site is O_b followed by O_a and O_c. Protonation at O_b yields directly a protonated derivative of ImD namely intermediate ImF. The latter could also be formed, via a transition state TS, upon protonation at the second most probable site of ImDi.e. O_a (Fig. 8). Finally, protonation of ImF at site O_c (protonation at site O_a does not lead to products formation) yields CO and H₂O while regenerating TEOA.

Fig. 8 Free energy, ΔG (in kcal mol⁻¹), reaction profiles for CO release step in DCM and DMF (numbers in parenthesis) solvents, with TEOA and TEOA(H)⁺ as proton donors (blue and green lines respectively), calculated at the PBE0/Def2-TZVP level.

It should be stressed here, that the most facile route to products formation in the TEOA complex formation step is expected to proceed via protonation of ImD at site O_b, with TEOA(H)⁺ as the proton source since in this case direct formation of ImF exhibits a relatively low energy barrier around 20 kcal mol⁻¹. All the other paths examined i.e. protonation of ImD at either site O_b, with TEOA proton donor or at site O_a, with TEOA/TEOA(H)⁺ proton donors exhibit very high energetic barriers rendering them unfavorable.

The TEOA complex formation step yielding CO, H₂O and regenerating TEOA is calculated to be strongly exergonic for all routes examined with ΔG ranging from −98 to −268 kcal mol⁻¹.

The optimized geometries of all species involved in the catalytic cycle of CO release step (Fig. 8) are given in Fig. S5 of the ESI.†

3.2.2 η¹-CO₂ complex mechanism. The second possible mechanism, examined for the CO₂ electro/photocatalytic conversion by Ru(II)/Os(II) complexes under study, is based on the direct CO₂ capture by the 17e⁻ catalytically active intermediate. This mechanism is the most popular and widespread to people working in this area. The proposed catalytic cycle, involving the η¹-CO₂ complex formation, is depicted schematically in Fig. 9.

Fig. 9 Proposed η¹-CO₂ complex mechanism for CO₂ to CO conversion by 1 or 2.

It can be seen that, in the η¹-CO₂ complex mechanism, the five coordinated, 17e⁻ catalytically active intermediate, ImA_M is one electron reduced with TEOA to yield the five coordinated intermediate ImB′_M. This step has been proposed also to occur during the photocatalytic CO₂ to CO reduction by the trans-Ru(bpy)(CO)₂Cl₂ complex.⁵¹ Next, CO₂ is captured by the reactive ImB′_M species forming the η¹-CO₂ complex, ImC′_M which is considered to be a key intermediate. The CO₂ capture/activation step proceeds via transition state TS₁′_M. The formation of ImC′_M is followed by two successive protonations of ImC′_M with TEOA acting as the proton source, yielding subsequently ImD′_M and ImE′_M intermediates (Fig. 9). The ImE′_M intermediate loses a water molecule, producing the tetracarbonyl intermediate, ImF′_M. This process proceeds via transition state TS₂′_M. Finally, ImF′_M receives an e⁻ from TEOA forming the reduced ImG′_M intermediate which is then yields CO and the initial catalytic species ImA_M.

The optimized geometries of all the species participating in the catalytic cycles depicted in Fig. 9, along with selected structural parameters are given in Fig. S6 and S7† for Ru and Os species respectively in both DCM and DMF solvents. The calculated free energy, ΔG profiles for 1 and 2 in DCM and DMF solvents are given in Fig. 10 and 11 respectively. Inspection of Fig. 10 and 11 reveals that the one electron reduction of ImA_M with TEOA to yield ImB′_M is endergonic. Next, we have the formation of ImC′_Mvia transition state TS₁′_M with an energy barrier estimated to be in the range 48–50 kcal mol⁻¹. The protonation of ImC′_M to yield ImD′_M could proceed either with TEOA or with protonated TEOA(H)⁺. However, protonation with the former is predicted to be unfavourable while for the latter it is favourable. Subsequent protonation of ImD′_M, with TEOA (blue lines, Fig. 10 and 11) to yield ImE′_M as well as formation of the tetracarbonyl intermediates, ImF′_M and ImG′_Mvia transition state TS₂′_M are expected to be highly unfavourable exhibiting quite high energy barriers. Finally, formation of CO giving back also the initial intermediate ImA_M is an endergonic process. In contrast to TEOA, the path with TEOA(H)⁺ (green lines, Fig. 10 and 11) is expected to be more favourable.

Fig. 10 Free energy, ΔG (in kcal mol⁻¹), reaction profile for the CO₂ to CO conversion by 1, following the η¹-CO₂ complex mechanism, in DCM and DMF (numbers in parenthesis) solvents, with TEOA as proton donor (blue lines) and with TEOA(H)⁺ (green lines), calculated at the PBE0/Def2-TZVP level.

Fig. 11 Free energy, ΔG (in kcal mol⁻¹), reaction profile for the CO₂ to CO conversion by 2, following the η¹-CO₂ complex mechanism, in DCM and DMF (numbers in parenthesis) solvents, with TEOA as proton donor (blue lines) and with TEOA(H)⁺ (green lines), calculated at the PBE0/Def2-TZVP level.

Accordingly, protonation of ImD′_M, with TEOA(H)⁺ is almost barrierless, while formation of the tetracarbonyl intermediates proceeds viaTS₂′_M exhibiting a relatively low energy activation barrier. The path followed upon employing TEOA(H)⁺ is predicted to be slightly endergonic.

Since ImC′_M is of key importance for CO₂ to CO conversion we set out to analyse in more depth the nature of the M–CO₂ bond by means of NBO analysis method. Thus, the natural charges on Ru and C atoms of the Ru–CO₂ bond in ImC′_Ru are equal to −1.307 and 0.794|e| in DCM and −1.302 and 0.790|e| in DMF solvent. On the other hand, the natural charges on Os and C atoms of the Os–CO₂ bond in ImC′_Os are equal to −0.817 and 0.685|e| in DCM and −0.813 and 0.682|e| in DMF solvent. Obviously, there is favourable electrostatic attraction between the constituent atoms of the M–CO₂ bond. In addition, the M–CO₂ bond is expected to have a small covalent component as well, since the WBIs are 0.660 and 0.671 for the Ru–CO₂ bond in DCM and DMF solvents respectively while the respective WBIs for the Os–CO₂ bond are 0.672 and 0.679. The covalent nature of the M–CO₂ is demonstrated also by the existence of a bonding NBO (Scheme 4).

Scheme 4 3D surface of the σ(Ru–C) bonding NBO of ImC_Ru.

The above bonding NBO is formed upon interaction of the sp^3.28d^2.70 hybrid orbitals (47.0% p and 38.6% d character) of Ru with the sp^2.35 hybrid orbital (70.1% p-character) of the carbon atom of CO₂ and is described as σ(Ru–O) = 0.702h_Ru + 0.713h_C. The occupation number of σ(Ru–C) NBO is 1.752|e|.

4. Conclusions

The photocatalytic CO₂ to CO reduction by [(en)M(CO)₃Cl], Ru(II) and Os(II) octahedral complexes was studied by means of DFT calculations. In an initial step, the [(en)M(CO)₃Cl] complexes, upon irradiation, are excited in their T₁ state which upon one electron reduction with TEOA, are converted to the OER species. It follows the main catalytic cycle which starts with a 17e⁻ five coordinated species, obtained from OER upon losing its chloride ligand. Two possible mechanisms were examined for the main catalytic cycle: (a) the TEOA complex mechanism and (b) the η¹-CO₂ complex mechanism. At the beginning, we scrutinized the initial step of the photocatalytic process under study. The TDDFT simulated absorption spectra of the initial photocatalysts, [(en)M(CO)₃Cl] in DCM and DMF solvents, exhibit absorption bands mainly in the UV region. The polarity of the solvent does not affect the absorption spectra. In contrast, the calculated T₁ excited state reduction potentials, depend upon the solvent and reveal that 1 and 2 in DCM are the easiest to be reductively quenched.

The TEOA complex mechanism proposed for the main catalytic cycle, succeeding the Initial Step, is based on the formation of the complex [(en)M(CO)₃(OCH₂CH₂NR₂)] between the 17e⁻ five coordinated catalytically active intermediate with TEOA⁻ anion acting as O-donor ligand. A similar Re(I) complex as well as its CO₂ insertion product have already been experimentally observed.²³ The TEOA complex mechanism is thought to occur in two steps. In the first step, the catalytically active intermediate Im_A is regenerated and a metal ‘free’ intermediate, Im_D is also produced. The latter, starts the second step were via two possible pathways, produces CO, TEOA and H₂O. It is anticipated that the proton donor source should play a crucial role in the TEOA complex mechanism. Thus, if TEOA is the proton donor, the TEOA complex formation step is predicted to be slightly endergonic while if TEOA(H)⁺ is the proton donor, this step is strongly exergonic. The CO release step is exergonic using either TEOA or TEOA(H)⁺ proton donors. However, use of TEOA proton donor in the CO release step, results in very high activation energy barriers in contrast to TEOA(H)⁺ proton donor where the activation energy barriers are much lower. Therefore, this step is expected to be kinetically more favorable if the proton source is TEOA(H)⁺.

The η¹-CO₂ complex mechanism is based on earlier studies proposing a two electron – two proton process with TEOA acting as a sacrificial anode. This mechanism starts also with the 17e⁻ catalytically active intermediate which upon one electron reduction, captures CO₂ and forming a loosely bound η¹-CO₂ complex. Successive protonations and one electron reduction yields CO while regenerating the 17e⁻ catalytically active intermediate. The whole process is predicted to be endergonic though it has to surmount a significant activation barrier. The latter should be alleviated in the case of electrocatalytic CO₂ to CO reduction by the complexes under study. As is for the TEOA complex mechanism, the η¹-CO₂ complex mechanism is favored with TEOA(H)⁺ proton donor.

The geometric and the energetic profiles of both mechanisms are not significantly affected by the nature of the solvent. We believe that, the TEOA complex mechanism should be of importance in low CO₂ concentrations²³ while the η¹-CO₂ complex mechanism is a more general approach for CO₂ to CO electrocatalytic/photocatalytic conversion.

Overall, the present theoretical study, being a follow up of our previous work on Re(I) complexes,¹⁷ is an extension to cover also the photocatalytic CO₂ to CO conversion by similar Ru(II)/Os(II) simple model complexes, highlighting the role of the TEOA ‘magic’ sacrificial donor. Our molecular modeling works on Re(I), Ru(II) and Os(II) complexes proposes for the first time an alternative ‘unconventional’ mechanism which could trigger further experimental works aiming at fully delineating the photocatalytic CO₂ to CO reduction using TEOA.

Conflicts of interest

There are no conflicts to declare.

References

1. R. Quadrelli and S. Peterson, Energy Policy, 2007, 35, 5938–5952.
2. B. J. Mason, Contemp. Phys., 1989, 30, 417–432.
3. M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann and F. E. Kühn, Angew. Chem., Int. Ed., 2011, 50, 8510–8537.
4. A. J. Morris, G. J. Meyer and E. Fujita, Acc. Chem. Res., 2009, 42, 1983–1994.
5. J. A. Keith, K. A. Grice, C. P. Kubiak and E. A. Carter, J. Am. Chem. Soc., 2013, 135, 15823–15829.
6. C. Riplinger, M. D. Sampson, A. M. Ritzmann, C. P. Kubiak and E. A. Carter, J. Am. Chem. Soc., 2014, 136, 16285–16298.
7. H. Koizumi, H. Chiba, A. Sugihara, M. Iwamura, K. Nozaki and O. Ishitani, Chem. Sci., 2019, 10, 3080–3088.
8. Y. Asai, H. Katsuragi, K. Kita, T. Tsubomura and Y. Yamazaki, Dalton Trans., 2020, 49, 4277–4292.
9. I. Choudhuri and D. G. Truhlar, J. Phys. Chem. C, 2020, 124, 8504–8513.
10. Y. S. Gong, J. Fan, V. Cecenc and C. Huang, Chem. Eng. J., 2021, 405, 12613–12627.
11. K. S. Rawat, A. Mahata, I. Choudhuri and B. Pathak, J. Phys. Chem. C, 2016, 120, 8821–8831.
12. D. H. Apaydin, S. Schlager, E. Portenkirchner and N. S. Sariciftci, ChemPhysChem, 2017, 18, 3094–3116.
13. N. Elgrishi, M. B. Chambers, X. Wang and M. Fontecave, Chem. Soc. Rev., 2017, 46, 761–796.
14. K. E. Dalle, J. Warnan, J. J. Leung, B. Reuillard, I. S. Karmel and E. Reisner, Chem. Rev., 2019, 119, 2752–2875.
15. Y. Yamazaki, H. Takeda and O. Ishitani, J. Photochem. Photobiol., C, 2015, 25, 106–137.
16. Y. Kuramochi, O. Ishitani and H. Ishida, Coord. Chem. Rev., 2018, 373, 333–356.
17. A. C. Tsipis and A. A. Sarantou, Dalton Trans., 2021, 50, 14797–14809.
18. J.-M. Lehn and R. Ziessel, Proc. Natl. Acad. Sci. U. S. A., 1982, 79, 701–704.
19. J. Hawecker, J.-M. Lehn and R. Ziessel, J. Chem. Soc., Chem. Commun., 1983, 536–538.
20. J. Hawecker, J.-M. Lehn and R. Ziessel, Helv. Chim. Acta, 1986, 69, 1990–2012.
21. J. M. Smieja and C. P. Kubiak, Inorg. Chem., 2010, 49, 9283–9289.
22. H. Takeda, K. Koike, H. Inoue and O. Ishitani, J. Am. Chem. Soc., 2008, 130, 2023–2031.
23. T. Morimoto, T. Nakajima, S. Sawa, R. Nakanishi, D. Imori and O. Ishitani, J. Am. Chem. Soc., 2013, 135, 16825–16828.
24. J. Agarwal, E. Fujita, H. F. Schaefer and J. T. Muckerman, J. Am. Chem. Soc., 2012, 134, 5180–5186.
25. H. Ishida, K. Tanaka and T. Tanaka, Chem. Lett., 1985, 14, 405–406.
26. H. Ishida, K. Tanaka and T. Tanaka, Organometallics, 1987, 6, 181–186.
27. P. Kang, C. Cheng, Z. Chen, C. K. Schauer, T. J. Meyer and M. Brookhart, J. Am. Chem. Soc., 2012, 134, 5500–5503.
28. T. Matsubara and K. Hirao, Organometallics, 2001, 20, 5759–5768.
29. P. Voyame, K. E. Toghill, M. A. Méndez and H. H. Girault, Inorg. Chem., 2013, 52, 10949–10957.
30. M. Irikura, Y. Tamaki and O. Ishitani, Chem. Sci., 2021, 12, 13888–13896.
31. Y. Tamaki, K. Koike, T. Morimoto, Y. Yamazaki and O. Ishitani, Inorg. Chem., 2013, 52, 11902–11909.
32. C. E. Castillo, J. Armstrong, E. Laurila, L. Oresmaa, M. Haukka, J. Chauvin, S. Chardon-Noblat and A. Deronzier, ChemCatChem, 2016, 8, 2667–2677.
33. N. W. Kinzel, C. Werlé and W. Leitner, Angew. Chem., Int. Ed., 2021, 60, 11628–11686.
34. V. Vetere, C. Adamo and P. Maldivi, Chem. Phys. Lett., 2000, 325, 99–105.
35. C. Adamo and V. Barone, Theor. Chem. Acc., 2000, 105, 169–172.
36. C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158–6170.
37. M. Ernzerhof and G. E. Scuseria, J. Chem. Phys., 1999, 110, 5029–5036.
38. C. Adamo, G. E. Scuseria and V. Barone, J. Chem. Phys., 1999, 111, 2889–2899.
39. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
40. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox, Gaussian 16, Revision C.01, Gaussian, Inc., Wallingford CT, 2016.
41. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3093.
42. A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988, 88, 899–926.
43. F. Weinhold, Natural Bond Orbital Methods, in Encyclopedia of Computational Chemistry, ed. P. V. R. Schleyer, N. L. Allinger, T. Clark, J. Gasteiger, P. A. Kollman, H. F. Schaefer III and P. R. Schreiner, John Wiley and Sons, Chichester, UK, 1998, vol. 3, pp. 1792–1811.
44. S. J. A. van Gisbergen, F. Kootstra, P. R. T. Schipper, O. V. Gritsenko, J. G. Snijders and E. J. Baerends, Phys. Rev. A: At., Mol., Opt. Phys., 1998, 57, 2556–2571.
45. C. Jamorski, M. E. Casida and D. R. Salahub, J. Chem. Phys., 1996, 104, 5134–5147.
46. R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett., 1996, 256, 454–464.
47. C. Adamo and D. Jacquemin, Chem. Soc. Rev., 2013, 42, 845–856.
48. T. B. Demissie, K. Ruud and J. H. Hansen, Organometallics, 2015, 34, 4218–4228.
49. A. A. Isse and A. Gennaro, Absolute Potential of the Standard Hydrogen Electrode and the Problem of Interconversion of Potentials in Different Solvents, J. Phys. Chem. B, 2010, 114, 7894–7899.
50. T. Nakajima, Y. Tamaki, K. Ueno, E. Kato, T. Nishikawa, K. Ohkubo, Y. Yamazaki, T. Morimoto and O. Ishitani, J. Am. Chem. Soc., 2016, 138, 13818–13821.
51. Y. Kuramochi, J. Itabashi, K. Fukaya, A. Enomoto, M. Yoshida and H. Ishida, Chem. Sci., 2015, 6, 3063–3074.

---

Footnote

† Electronic supplementary information (ESI) available: 3D surfaces of the MOs involved in the most relevant electronic excitations in the simulated absorption spectra, Selected structural parameters of the various intermediates involved in step B of the catalytic cycle for the CO₂ to CO conversion, Gibbs free energies, G for all species involved in the Born–Haber cycle for redox potentials, parameters relevant to redox potentials, Cartesian coordinates and energetic data. See DOI: https://doi.org/10.1039/d4dt00125g

---