DFT insights into the photocatalytic reduction of CO₂ to CO by Re(I) complexes: the crucial role of the triethanolamine “magic” sacrificial electron donor

Abstract

The reaction mechanism for the photocatalytic reduction of CO₂ to CO catalyzed by the [Re(en)(CO)₃Cl] complex in the presence of triethanolamine, R₃N (R = CH₂CH₂OH) abbreviated as TEOA, in DMF solution was studied in-depth with the aid of DFT computational protocols by calculating the geometric and free energy reaction profiles for several possible reaction pathways. The reaction pathways studied start with the “real” catalytic species [Re(en)(CO)₃], [Re(en)(CO)₃]⁻ and/or [Re(en)(CO)₂Cl]⁻ generated from the excited triplet T₁ state upon single and double reductive quenching by a TEOA sacrificial electron donor or photodissociation of a CO ligand. The first step in all the catalytic cycles investigated involves the capture of either CO₂ or the oxidized R₂NCH₂CH₂O˙ radical. In the latter case, the CO₂ molecule is captured (inserted) by the Re–OCH₂CH₂NR₂ bond forming stable intermediates. Next, successive protonations (TEOA also acts as a proton donor) lead to the release of CO either from the energy consuming 2e⁻ reduction of [Re(en)(CO)₄]⁺ or [Re(en)(CO)₂Cl]⁺ complexes in the CO₂ capture pathways or from the released unstable diprotonated [R₂NCH₂CH₂OC(OH)(OH)]⁺ species regenerating TEOA and the catalyst. The CO₂ insertion reaction pathway is the favorable pathway for the photocatalytic reduction of CO₂ → CO catalyzed by the [Re(en)(CO)₃Cl] complex in the presence of TEOA manifesting its crucial role as an electron and proton donor, capturing CO₂ and releasing CO.

---

Introduction

It is generally accepted that anthropogenic sources, such as burning of fossil fuels, are mainly responsible for the rise of atmospheric CO₂, which in turn accentuates the greenhouse effect.¹ Many studies have been devoted to how to tackle this problem and there are continuous ongoing efforts not only to capture CO₂ but also to catalytically transform it into fine chemicals.² Among the various strategies to achieve this goal, the most challenging and the very attractive one is to convert CO₂ efficiently into useful compounds using solar light as an energy source.³ Many transition metal-based complexes (Ni, Fe, Re, Cr, Ir, Mo, etc.) have been studied extensively for the homogeneous electrocatalytic and photocatalytic reduction of carbon dioxide.^4–16

Apaydin et al.⁷ reported an excellent comprehensive overview on the homogeneous and heterogeneous CO₂ reduction catalyzed by organic, organometallic and bioorganic systems. Mechanistic details of the CO₂ reduction processes undertaken by electrochemical, bioelectrochemical and photoelectrochemical approaches are thoroughly analyzed. In parallel to the electrochemical CO₂ reduction, a plethora of efforts were focused on the photocatalytic CO₂ reduction, which relied on systems comprising a light harvesting unit consisting of a photosensitizer (PS) compound and two catalytic sites.^8–14 In the oxidation site, a donor provides an electron e⁻ to the PS after its excitation to the triplet excited state (³MLCT) which is subsequently reductively quenched by the reduction site and finally, the e⁻ is transferred to CO₂. However, in many cases, the PS acts not only as a photosensitizer but also as a reduction site as well.

Among the mononuclear transition metal complexes developed so far for electro- and photocatalytic CO₂ reduction, polypyridyl transition metal complexes constitute the class of molecular catalysts employed in the reduction of CO₂ to CO. An excellent overview of the CO₂ reduction catalyzed by polypyridyl transition metal complexes has recently been published by Fontecave's group,⁸ presenting and thoroughly analyzing the proposed catalytic cycles for the CO₂ reduction by Re(bpy)CO₃(X) investigated by both experimental and computational approaches. Generally, mechanistic studies indicated that, initially a two electron reduction of the Re(bpy)CO₃(X) catalyst affords the anionic five-coordinated 17e⁻ [Re(bpy)(CO)₃]⁻ complex which captures CO₂ forming the [Re(bpy)(CO)₃(CO₂)]⁻ intermediate, which upon protonation yields the Re(bpy)(CO)₃(CO₂H) intermediate which upon second protonation and additional electron reduction is converted into [Re(bpy)(CO)₄]⁻. However, a question still remains concerning the release of the CO product from the [Re(bpy)(CO)₄]⁻ complex. More recently, Ishitani and co-workers¹¹ presented an excellent discussion of all the reaction mechanisms proposed for the photochemical CO₂ reduction catalyzed by Re(I) and Ru(II) complexes. The authors stated that no one from the numerous mechanisms proposed is a universal mechanism. This review article is recommended to the readers and researchers in the field. On the other hand, Cramer's group reported mechanistic details on the proton-dependent electrocatalytic reduction of CO₂ to CO by fac-Re(bpy)(CO)₃Cl using first principles quantum chemistry.¹³ The Cramer's group also compared the complete electrocatalytic cycles of CO₂ to CO reduction catalyzed by fac-Re(bpy)(CO)₃Cl and fac-Mn(bpy)(CO)₃Cl catalysts.¹⁴

The photo-induced reduction of CO₂ to CO in the acetonitrile/water/triethylamine solution in the presence of a [Ru(2,2′-bipyridine)₃]²⁺/Co₂⁺ system was first reported by Lehn and Ziessel.¹⁷ In subsequent publications, Lehn's group^18,19 showed that the most selective reduction catalysts to produce CO from CO₂ are Re(I) octahedral complexes with the general formula [Re(bpy)(CO)₃X] (X = Cl, Br). The proposed catalytic cycle for the photoreduction of CO₂, catalyzed by the (bpy)Re(CO)₃X complexes, comprises excitation to the ³MLCT state and reduction of the Re(I) catalyst using triethanolamine (TEOA) as an electron donor yielding the one electron reduced Re(I) complex (OER-species) that loses the X⁻ ligand forming an unstable 17e⁻ species. Next, the 17e⁻ Re(I) complex could capture CO₂via coordination to the rhenium metal center.^17–19 However, the mechanism of conversion of CO₂ to CO has not yet been fully understood and there are some points that are still under debate and need to be clarified.

The mechanism of the photoinduced reduction of CO₂ to CO in a TEOA/DMF/[ReBr(CO)₃(bpy)] system has been investigated by Kutal et al.^20,21 Reductive quenching of the photoexcited [ReBr(CO)₃(bpy)] complex by TEOA affords the reduced [Re′Br-(CO)₃(bpy⁻)] species which can be viewed as a Re’ center bound to a 2,2′-bipyridine radical anion. The TEOA˙⁺ generated can rapidly abstract a hydrogen atom from another TEOA molecule to produce the strong reducing radical, TEOA˙. Next, the 19e⁻ Re species activates CO₂ for reduction to CO, but the composition, structure, or subsequent reactivity of any intermediates formed is unknown.

Although several studies^22–29 point towards the existence of the 17e⁻ reactive species, the next step involving capturing CO₂ by the catalytic system is unclear. Kubiak and co-workers^22–25 investigated the catalytic activity of the Re(bpy)(CO)₃X complexes (bpy = 4,4′-dicarboxyl-2,2′-bipyridine, 2,2′-bipyridine, 4,4′-dimethyl-2,2′-bipyridine, 4,4′-di-tert-butyl-2,2′-bipyridine, and 4,4′-dimethoxy-2,2′-bipyridine) and found that Re(bipy-tBu)(CO)₃Cl has the most significant catalytic activity for the reduction of CO₂ to CO. DFT calculations showed that the geometry of the [Re(bipy-tBu)(CO)₃(CO₂)]⁻ doubly reduced species converges only upon inclusion of a cation (H, Li, Na, or K). Based on these studies, the authors proposed a catalytic cycle that involves a two-electron reduction of the [Re(bpy-R)(CO)₃X] (R = H, Me, tBu, X = halogen, OTf) complex which upon dissociation of the X ligand yields the anionic [Re(bpy-R)(CO)₃]⁻ species. Several [Re(bpy-R)(CO)₃]⁻ anions were isolated and fully characterized.²⁵ Next, the CO₂ molecule is coordinated to the Re metal center yielding a carboxylate [Re(bpy-R)(CO)₃(CO₂)]⁻ intermediate, which upon protonation forms the carboxylato [Re(bpy-R)(CO)₃(CO₂H)] complex. Further protonation of the [Re(bpy-R)(CO)₃(CO₂H)] complex releases H₂O and produces the cationic tetracarbonyl species. The release of CO from the [Re(bpy-R)(CO)₄]⁺ complex is catalysed by electron transfer.

On the other hand, Muckerman et al.²⁶ in a prominent theoretical study precluded direct coordination of the Re metal center and instead proposed that the catalytic cycle proceeds via the formation of a Re dimeric intermediate species. In the latter, CO₂ acts as a bridging ligand between the two Re metal centers, though the formation of the carboxylate dimeric species is entropically unfavorable. Ishitani et al.²⁷ proposed an alternative route by revealing the crucial role of TEOA in the CO₂ capture by the Re(I) catalyst. The authors proposed a catalytic cycle for the photocatalytic CO₂ reduction catalyzed by the [Re(bpy)(CO)₃(NCS)] complex that involves excitation of the S₀ ground state of the [Re(bpy)(CO)₃(NCS)] complex to the triplet ³MLCT state, which subsequently is reductively quenched by TEOA, giving the one electron reduced (OER) [Re(bpy)(CO)₃(NCS)]^−· species. The dissociation of the SCN⁻ ligand from the OER yields a 17e⁻ five-coordinated intermediate which reacts with CO₂ to give the CO₂ adduct(s). Further electron donation to the OER results in the release of CO and the [Re(bpy)(CO)₃]⁺ complex and regenerates the catalyst. Later Ishitani's group²⁸ performing spectrochemical and cyclic voltammetry measurements on adding TEOA to a solution of fac-[Re^I(bpy)(CO)₃(CH₃CN)]PF₆ dissolved in pure DMF provided evidence for the existence of the fac-[Re^I(bpy)(CO)₃(OCH₂CH₂NR₂)] (R = CH₂CH₂OH) intermediate in solution. A Re(I) intermediate with the CO₂ inserted between the Re(I) metal center and the TEOA anionic ligand was isolated and fully characterized.

The photocatalytic activity of transition metal-based catalysts towards the reduction of CO₂ to CO depends upon the structure of the complexes. Attempts to address this issue have already been made. For example, Kurz et al.²⁹ performed an experimental structure–activity study of [(diimine)(CO)₃ReX] complexes with respect to their ability to act as photocatalysts for the reduction of CO₂ to CO. Also, Ishida et al.³⁰ studied the effect of ligand L on the photocatalytic activity for CO₂ conversion in a series of trans-(Cl)-[Ru(L)(CO)₂Cl₂] complexes, revealing the significance of their reduction potential, E_p. The latter is modulated by the Lowest Unoccupied Molecular Orbital (LUMO) eigenvalue which in turn depends on the nature of L.

In this regard, the main goal that we pursued in this work was to scrutinize the mechanistic details of the photocatalytic conversion of CO₂ to CO by the octahedral [(en)(CO)₃ReCl] (en = ethylenediamine) complex selected as a model complex, employing DFT methods. Our efforts were directed towards the elucidation of the mechanism of the reaction step related to CO₂ capture and CO release in the reduction of CO₂ to CO catalyzed by Re(I) complexes in DMF/TEOA solutions.

Computational methods

All stationary points (reactants, transition states, and products) located on the potential energy surfaces (PES) were fully optimized at the wB97XD/Def2-TZVP level of theory as implemented in the Gaussian16 suite of programs.³¹ The wB97XD functional^32–35 containing empirical dispersion terms and long-range corrections provides good descriptions of the reaction profiles, including geometries, heats of reaction, and barrier heights.³⁶ Analytical frequencies were calculated at the same level of theory, and the nature of the stationary points was determined in each case according to the number of the negative eigenvalues of the Hessian matrix. Gibbs free energies were calculated at 298.150 Kelvin and 1 Atm. Solvent effects were accounted for by means of the Polarizable Continuum Model (PCM) with the integral equation formalism variant (IEF-PCM) being the default self-consistent reaction field (SCRF) method.³⁷ DMF was used as the solvent. Natural Bond Orbital (NBO) population analysis was performed using Weinhold's methodology.^38,39 Time-dependent density functional theory (TD-DFT)^40–42 calculations were performed on the equilibrium ground state geometries in DMF solution employing the wB97XD/Def2-TZVP/PCM computational protocol.

Results and discussion

Computation-based catalytic cycles

The photocatalytic CO₂ → CO transformation catalyzed by the selected [(en)(CO)₃Re(Cl)] model catalyst proceeds through three main reaction steps: (i) step A1 corresponds to the photo-excitation of the [(en)(CO)₃Re(Cl)] catalyst accompanied by one electron reduction of the triplet excited state T₁ by TEOA, the so-called “magic” sacrificial electron donor,²⁸ yielding very reactive coordinatively unsaturated, 17e⁻ intermediates that are the catalytically active species initiating step B1. (ii) Step B1 involves the interaction of the oxidized TEOA˙⁺ species formulated as [R₂NCH₂CH₂OH]˙⁺ (R = CH₂CH₂OH) with the Re metal center yielding [(en)(CO)₃Re(OCH₂CH₂NR₂)] intermediates capable of capturing CO₂ by insertion into the Re–O bond. Step B1 has thoroughly been investigated both experimentally and theoretically by Ishitani's group²⁸ employing the fac-[Re^I(bpy)(CO)₃(CH₃CN)]PF₆ catalyst. (iii) Step C starts with the [R₂NCH₂CH₂C(O)OH]˙⁺ species released upon protonation of the [(en)(CO)₃ReO(O)(COCH₂CH₂NR₂)] intermediate and its conversion upon further protonation to the desired CO molecule, H₂O and regenerating TEOA. Step C manifests the crucial role of TEOA in the photocatalytic reduction of CO₂ → CO, thus clarifying, the “magic” sacrificial electron donor features of TEOA.²⁸

The reaction steps for the photocatalytic reduction of CO₂ to CO catalyzed by the [(en)(CO)₃ReCl] (en = ethylenediamine) catalyst, calculated at the wB97XD/def2-TZVP level, in a DMF solution are depicted schematically in the ESI (Fig. S1†). The alternative reaction steps B1′ and B1′′ starting with the coordination of CO₂ to the Re metal center of ImA1 and/or ImA1′ (one electron reduced ImA1 species) intermediates yielding the [Re(en)(CO)₃(CO₂)] and/or [Re(en)(CO)₃(CO₂)]⁻ intermediates, which upon successive protonations afford the tetracarbonyl [Re(en)(CO)₄]²⁺ and [Re(en)(CO)₄]⁺ complexes respectively will be discussed separately and compared with step B1. We also investigated alternative reaction pathways starting with a five-coordinated [(en)(CO)₂ReCl]⁻ intermediate that resulted from the dissociation of one of the equatorial CO ligands of the excited T₁ triplet state (step A2) which will be discussed later on.

Reaction steps A1 and A2: photophysics of the [(en)(CO)₃ReCl] catalyst

Let us examine first steps A1 and A2 concerning the photo-excitation of the [(en)(CO)₃ReCl] catalyst. Since excitation of the [(en)(CO)₃ReCl] complex is a prerequisite in order to initiate the photocatalytic reduction of CO₂ to CO, we set out to study the photophysical properties by means of Time Dependent DFT (TDDFT) electronic structure calculations. The simulated absorption spectrum of the [(en)(CO)₃ReCl] complex calculated by the TDDFT/wB97XD/Def2-TZVP/PCM computational protocol shows absorption bands spanning between 200 and 300 nm with the most intense among them peaking at 211, 231 and 273 nm. The peak at 211 nm exhibits a complex nature and could be assigned as Metal to Ligand/Ligand to Ligand Charge Transfer (MLCT/LL′CT). The same assignment holds also true for the electronic transitions absorbing at 231 and 273 nm as well.

The equilibrium geometries of the [(en)(CO)₃ReCl] complex in the ground S₀ and excited triplet T₁ states along with the catalytically active five-coordinated [(en)(CO)₃Re] and [(en)(CO)₂ReCl] species formed in photo-physical steps A1 and A2 respectively are shown in Fig. 1. Calculations predicted that the fac-[(en)(CO)₂ReCl] isomer is more stable than the mer-[(en)(CO)₂ReCl] isomer in both the S₀ and T₁ states by 24.1 and 6.8 kcal mol⁻¹ respectively.

Fig. 1 Equilibrium geometries of the [(en)(CO)₃ReCl] complex in the ground S₀ and excited triplet T₁ states along with the catalytically active five-coordinated [(en)(CO)₃Re], ImA1 and [(en)(CO)₂ReCl], ImA2 species formed in photo-physical steps A1 and A2 respectively calculated at the wB97XD/Def2-TZVP level of theory in a DMF solution.

It can be seen that one electron transfer from TEOA to the T₁ state of [(en)(CO)₃ReCl] promotes the cleavage of either the Re–Cl or the Re–CO bond in the reduced 19e⁻ transient OER species formed, yielding the very reactive 17e⁻ [(en)(CO)₃Re], ImA1 and [(en)(CO)₂ReCl]⁻, ImA2 intermediates respectively. Noteworthily in T₁ one of the equatorial CO ligands forms a weaker Re–CO bond (Re–CO bond length of 2.062 Å) compared to the Re–CO bonds of the other two CO ligands in the equatorial and axial coordination sites with bond lengths of 1.907 and 1.932 Å respectively. Interestingly the Re–C–O bond angle of the weakly bonded CO ligand is 133.6°. Scheme 1 shows the 3D plots of the frontier molecular orbitals (FMOs) and the spin density distribution of T₁, along with the estimated Wiberg Bond Indices (WBOs) for the Re–Cl and Re–CO bonds in T₁ (the entry in the right side of Scheme 1).

Scheme 1 3D plots of the frontier molecular orbitals (FMOs), spin density distribution and Wiberg Bond Indices (WBI) of T₁, calculated at the wB97XD/Def2-TZVP level of theory.

It can be seen that the Single Occupied Molecular Orbital (SOMO) in T₁ is mainly localized on the weakly bonded CO ligand. The SOMO is composed of 19.1% Re spd, 45.2% C sp hybrid orbitals and 18.9% O p orbitals. The Lowest Unoccupied Molecular Orbital (LUMO) is composed of 25.8% Re spd and 24.0% C sp hybrid orbitals. The spin density is distributed on the Re–CO bond region, 1.078|e| on Re and 0.716 on the CO ligand. The estimated WBI(Re–CO) shows also their bond strengths, being 1.187 for the weaker bonded CO ligand and 1.720 and 1.576 for the CO ligands coordinated at the equatorial and axial coordination sites respectively.

Reaction step B1: CO₂ insertion

The geometric and free energy reaction profiles calculated for step B1 are shown in Fig. 2.

Fig. 2 Geometric and free energy, ΔG (in kcal mol⁻¹), reaction profiles for step B of the reduction of CO₂ to CO catalyzed by the [(en)(CO)₃ReCl] catalyst calculated at the wB97XD/Def2-TZVP level in the DMF solution. The sum of G_ImA1 + G_DMF at an infinite distance was considered to be 0.00.

Initially, a DMF molecule interacts with the very reactive 17e⁻ five coordinate intermediate ImA1, with the estimated interaction energy (IE) being −7.4 kcal mol⁻¹ forming a loose association through an endergonic process (ΔG = 4.5 kcal mol⁻¹). The ImA1⋯DMF interactions correspond to non-covalent interactions clearly shown in Scheme 2.

Scheme 2 Scatter graphs of the RDG vs. sign(λ₂)ρ(r) correlations (a) along with the 3D RDG isosurfaces (isosurface = 0.70 au) for the ImA1⋯DMF intermediate calculated at the wB97XD/Def2-TZVP level of theory.

Next, DMF is substituted by an OCH₂CH₂NR₂ ligand yielding an ImB1′ intermediate, in line with the experimental observations by Ishitani's group²⁸ that upon adding TEOA to a solution of the fac-[Re^I(bpy)(CO)₃(DMF)]⁺ complex (Im_DMF), a 1 : 2 thermodynamically equilibrated mixture of Im_DMF with the fac-[Re^I(bpy)(CO)₃(OCH₂CH₂NR₂)] (R = CH₂CH₂OH) (ImB) intermediate is obtained. The equilibrium between Im_DMF and ImB intermediates was further verified by IR spectroscopy.¹¹ The substitution reaction is exergonic (ΔG = −68.7 kcal mol⁻¹).

Natural Bond Orbital (NBO) population analysis revealed that the OCH₂CH₂NR₂ ligand is coordinated to the Re metal centre of the [(en)(CO)₃Re] catalytic species, ImA1, forming a relatively weak Re–O bond with an estimated WBI(Re–O) of 0.550. Note that both the Re metal centre and the O donor atom of the coordinated OCH₂CH₂NR₂ ligand acquire negative natural atomic charges of −0.592 and −0.779|e| respectively. The bonding σ(Re–O) NBO is constructed from the interaction of sp^2.69d^2.88 hybrid orbitals (40.8% p and 43.7% d character) of Re with an sp^1.82 hybrid (64.5% p-character) on the oxygen donor atom and is described as σ(Re–O) = 0.395h_Re + 0.919h_O. The occupancy or the σ(Re–O) NBO is 1.939|e|.

Next, a CO₂ molecule is inserted into the Re–OCH₂CH₂NR₂ bond transforming the OCH₂CH₂NR₂ to the OC(O)OCH₂CH₂NR₂ ligand yielding intermediate ImC1. The insertion process is slightly exergonic by only −3.3 kcal mol⁻¹ (Fig. 2). ImC1 could be subjected to protonations at the O atoms of the coordinated OC(O)OCH₂CH₂NR₂ ligand marked as O_a and O_b protonation sites (Fig. S1†). The natural atomic charges at the O_a and O_b atoms are −0.759 and −0.708|e| respectively. The protonation either at O_a or O_b of the coordinated [R₂N–CH₂CH₂O–C(O)O] ligand forces the rupture of the Re–O bond which yields ImD1 formulated as [R₂N–CH₂CH₂O–C(O)OH]⁺ and regenerates the five coordinated 17e⁻ImA1 catalytic species, thus closing the catalytic cycle. The protonation accompanied by ligand dissociation corresponds to a strongly exergonic process (ΔG = −222.6 kcal mol⁻¹).

Reaction step C

ImD is the species initiating reaction step C that delivers the desired CO product. The geometric and free energy reaction profiles calculated for step C are shown in Fig. 3. As in the case of ImC1 (vide supra), there are two possible protonation sites in ImD (marked as a and b in Fig. S1†). The natural atomic charges on the O_a and O_b atoms of ImD are −0.668 and −0.659|e| respectively. Thus, the protonation at the O_a site is the most favoured one yielding ImF which upon further protonation and one electron reduction delivers the CO product and regenerates TEOA. The ImD → ImF transformation is strongly exergonic (ΔG = −107.5 kcal mol⁻¹). ImF can also be obtained upon protonation at the O_b site in ImD forming a transition state TS (ν_i = −1888 cm⁻¹) surmounting an activation barrier of 41.1 kcal mol⁻¹ (Fig. 1). The normal coordinate vectors (arrows) of the vibrational modes corresponding to the imaginary frequency of TS show that the dominant motions reflect the structural changes leading to the formation of ImF. The relatively high activation barrier indicates that the ImD → TS → ImF is not a favoured pathway. Noteworthily reaction step C is a key step that manifests the crucial role of the TEOA sacrificial electron donor in the photocatalytic reduction of CO₂ to CO catalyzed by Re(I) complexes.

Fig. 3 Geometric and free energy, ΔG (in kcal mol⁻¹), reaction profiles for step C of the reduction of CO₂ to CO catalyzed by the [(en)(CO)₃ReCl] catalyst calculated at the wB97XD/Def2-TZVP level in the DMF solution.

Reaction steps B1′ and B1′′: CO₂ capture

Alternative reaction pathways (steps B1′ and B1′′) involving direct capture of CO₂ by the active ImA1′ catalytic species (step B1′) or its one electron reduced ImA1′′ species (step B1′′) yielding the [Re(en)(CO)₃(CO₂)] and [Re(en)(CO)₃(CO₂)]⁻ intermediates, respectively have also been explored by DFT methods. The proposed catalytic cycles for steps B1′ and B1′′ are shown in Fig. 4. The geometric and free energy reaction profiles calculated for steps B1′ and B1′′ are given in Fig. 5 and 6 respectively.

Fig. 4 Proposed catalytic cycles for steps B1′ and B1′′ calculated at the wB97XD/Def2-TZVP level in the DMF solution.

Fig. 5 Geometric and free energy, ΔG (in kcal mol⁻¹), reaction profiles for step B1′ of the reduction of CO₂ to CO catalyzed by the [(en)(CO)₃ReCl] catalyst calculated at the wB97XD/Def2-TZVP level in the DMF solution.

Fig. 6 Geometric and free energy, ΔG (in kcal mol⁻¹), reaction profiles for step B1′′ of the reduction of CO₂ to CO catalyzed by the [(en)(CO)₃ReCl] catalyst calculated at the wB97XD/Def2-TZVP level in the DMF solution.

After excitation of the S₀ ground state of the [Re(en)(CO)₃Cl] complex to the triplet T₁ state, the latter could undergo one and two electron reductions by TEOA accompanied by Cl⁻ dissociation affording the catalytically active 17e⁻ImA1′ and the anionic 18e⁻ImA1′′ intermediates which initiate reaction steps B1′ and B1′′ respectively (Fig. 4).

According to NBO population analysis, the Re metal centers in ImA1′ and ImA1′′ intermediates acquire negative natural atomic charges of −0.752 and −1.174|e| respectively. Noteworthily the Re metal center in ImA1′ and ImA1′′ is a nucleophilic center and therefore could not be attacked by a nucleophile. In this regard, direct coordination of CO₂ (CO₂ capture) to the Re metal center of ImA1′ and ImA1′′ intermediates seems to be highly unlikely in line with the experimental observations reported by Ishitani's group²⁸ that no reaction occurs when CO₂ is added to a solution of the fac-[Re^I(bpy)(CO)₃(DMF)]⁺ complex, while the capture of CO₂ was observed when bubbling CO₂ into a equilibrium mixture of fac-[Re^I(bpy)(CO)₃(DMF)]⁺/fac-[Re^I(bpy)(CO)₃(OCH₂CH₂NR₂)] (bpy = 2,2′-bipyridine) complexes in solution by ¹H and ¹³C NMR spectroscopy yielding the fac-[Re^I(bpy)(CO)₃{R₂N–CH₂CH₂O–COO}] complex which was isolated and fully characterized.

Scheme 3 shows the 3D plots of the frontier molecular orbitals (FMOs) and spin density distributions of ImA1′ and ImA1′′ catalytic species.

Scheme 3 3D plots of the frontier molecular orbitals (FMOs) and spin density distribution of ImA1′ and ImA1′′ catalytic species calculated at the wB97XD/Def2-TZVP level in the DMF solution.

The capture of CO₂ by ImA1′ in the reaction step B1′ yielding ImB1′ (Fig. 6) is slightly endergonic (3.1 kcal mol⁻¹), while the estimated interaction energy is 7.0 kcal mol⁻¹. According to the NBO population analysis, the CO₂ molecule is loosely associated with the Re metal centre of the ImA1′ catalytic species, with the Re–CO₂ bond having a WBI(Re–CO₂) of only 0.348. Note that the Re metal centre acquires a negative natural atomic charge of −0.808|e| and the C atom of the coordinated CO₂ a positive natural atomic charge of 0.755|e|. The side-on orientation of the CO₂ molecule in the ImB1′ intermediate indicates that non-covalent (dipole–dipole and electrostatic) interactions predominate in the coordination of CO₂ to the ImA1′ catalytic species. The weak covalent interactions are mirrored on the bonding σ(Re–CO₂) NBO which is constructed from the interaction of the sp^3.52d^2.74 hybrid orbitals (48.4% p and 37.7% d character) of Re with an sp^2.30 hybrid (69.6% p-character) on the carbon donor atom of the CO₂ ligand and described as σ(Re–CO₂) = 0.605h_Re + 0.797h_C. The occupancy or the σ(Re–CO₂) NBO is 0.947|e|.

Next the successive protonations of the coordinated CO₂ yielding ImC1′ and ImD1′ intermediates are exergonic processes, and the estimated ΔG values are −152.9 and −134.6 kcal mol⁻¹ respectively. ImD1′ releasing H₂O affords the dicationic [Re(en)(CO)₄]²⁺ complex ImE1′ which upon 2e⁻ reduction releases the desired CO product and regenerates the catalytically active ImA1′ species.

In reaction step B1′′ (Fig. 6), the capture of CO₂ by ImA1′′ yielding ImB1′′ is exergonic (−23.9 kcal mol⁻¹), indicative of the stronger non-covalent and covalent interactions between ImA1′′ and CO₂ (with an interaction energy of 23.9 kcal mol⁻¹). The NBO population analysis revealed that the CO₂ molecule is coordinated to the Re metal centre of the ImA1′′ catalytic species, forming a stronger nearly single Re–CO₂ bond with a WBI(Re–CO₂) of 0.755. The Re metal centre acquires a negative natural atomic charge of −1.016|e| and the C atom of the coordinated CO₂ a positive natural atomic charge of 0.631|e|. The covalent component of the Re–CO₂ bond is presented by the bonding σ(Re–CO₂) NBO which is constructed from the interaction of sp^2.83d^2.52 hybrid orbitals (44.5% p and 39.7% d character) of Re with an sp^1.62 hybrid (61.8% p-character) on the carbon donor atom of the CO₂ ligand and described as σ(Re–O) = 0.614h_Re + 0.789h_C. The occupancy or the σ(Re–O) NBO is 0.947|e|.

Successive protonations of the coordinated CO₂ yielding ImC1′′ and ImD1′′ are exergonic processes with estimated ΔG values of −194.2 and −154.4 kcal mol⁻¹ respectively. ImD1′′ releasing H₂O in an exergonic process (ΔG = −15.3 kcal mol⁻¹) affords the cationic [Re(en)(CO)₄]⁺ complex ImE′′ which upon 2e⁻ reduction releases the desired CO product and regenerates the catalytically active ImA1′′ species.

In summary reaction steps B1′ and B1′′ are less favoured than reaction step B1 due to higher energy consumption to achieve the release of the desired CO product from the cationic [Re(en)(CO)₄]⁺ complex by 2e⁻ reductions. Furthermore, the CO₂ insertion is more feasible than CO₂ capture illustrating the crucial role of the sacrificial electron donor TEOA in the photocatalytic reduction of CO₂ to CO catalyzed by the Re(en)(CO)₃Cl complex.

Reaction step B2

Reaction step B2 starts with the five-coordinated [(en)(CO)₂ReCl]⁻ intermediate that resulted from the dissociation of one of the equatorial CO ligands of the excited T₁ triplet state (step A2). It is important to note that Jean-Marie Lehn and co-workers¹⁹ showed, by a set of experiments concerning the carbonyl ligand exchange in [fac-Re(bpy)(CO)₃Cl], that reductive quenching of the excited state of the [Re(bpy)(CO)₃Cl] complex by tertiary amines gives an unstable five-coordinated [Re(bpy)(CO)₂Cl] species upon labilization of a CO ligand.

Two possible catalytic cycles following reaction steps B2′ and B2′′ have been thoroughly investigated. The reaction steps for the photocatalytic reduction of CO₂ to CO catalyzed by the five-coordinated [(en)(CO)₂ReCl]⁻, ImA2 intermediate are depicted schematically in Fig. 7.

Fig. 7 Proposed catalytic cycles for steps B2′ and B2′′ calculated at the wB97XD/Def2-TZVP level in the DMF solution.

Scheme 4 shows the 3D plots of the frontier molecular orbitals (FMOs) and spin density distribution of ImA2 and TEOA˙⁺ species.

Scheme 4 3D plots of the frontier molecular orbitals (FMOs) and spin density distribution of ImA2 and TEOA˙⁺ species calculated at the wB97XD/Def2-TZVP level of theory.

The SOMO of ImA2 is composed mainly of 84.8% Re sp_zd_z² hybrid orbitals, while the LUMO is composed of 72.0% Re pd hybrid orbitals and 8.3% 2p AOs of the carbon atom of the CO ligand in trans position to the chloride ligand. A spin density of 1.085|e| is localized on the Re central atom. According to the NBO population analysis, the Re metal centre acquires a negative natural atomic charge of −0.552|e|. The SOMO of TEOA˙⁺ is mainly localized on one of the CH₂CH₂OH alkyl groups (50.0%) mainly on the O atom (32.5%) and on the nitrogen atom (16.5%) of TEOA. Both the nature of the FMOs and the negative natural atomic charge on the Re metal centre could not support covalent interactions with the CO₂ ligand, thereby the capture of CO₂ by ImA2 might be supported mainly by non-covalent interactions. Indeed the interaction energy for the ImA2⋯CO₂ association is only 4.0 kcal mol⁻¹ calculated according to the wB97XD/Def2-TZVP/PCM computational protocol. However the nature of the FMOs of ImA2 and TEOA˙⁺ allows SOMO_ImA2–SOMO_TEOA˙⁺ interactions giving rise to the covalent component of the ImA2–TEOA bonding upon the formation of a bonding σ(Re–OCH₂CH₂NR₂) NBO.

Reaction step B2′: CO₂ insertion

The geometric and free energy reaction profiles calculated for step B2′ are shown in Fig. 8.

Fig. 8 Geometric and free energy, ΔG (in kcal mol⁻¹), reaction profiles for step B2′ of the reduction of CO₂ to CO catalyzed by the active [(en)(CO)₃Re]⁻ catalytic species, ImA2, calculated at the wB97XD/Def2-TZVP level in the DMF solution.

Perusal of Fig. 8 shows that an R₂N–CH₂CH₂O˙ radical is coordinated to the Re metal centre of the catalytically active ImA2 species which disposes of a vacant coordination site yielding ImB2′. The interaction of R₂N–CH₂CH₂O˙ with ImA2 is exergonic (ΔG = −176.4 kcal mol⁻¹). The NBO population analysis of ImB2′ revealed that the R₂N–CH₂CH₂O˙ species is coordinated to the Re metal centre of ImA2 forming a Re–OCCH₂CH₂NR₂ bond with a WBI(Re–O) of 0.604. The covalent component of the Re–O bond is described by the bonding σ(Re–O) NBO, with an occupancy of 1.929|e|. The σ(Re–O) NBO results from the interaction of the sp^2.89d^4.15 hybrid orbitals (35.8% p and 51.4% d character) of Re with an sp^2.16 hybrid (68.3% p-character) of the O donor atom of the R₂N–CH₂CH₂O˙ ligand and is described as σ(Re–O) = 0.399h_Re + 0.917h_O. The O atom of the coordinated R₂N–CH₂CH₂O ligand acquires a negative natural atomic charge of −0.759|e| indicative of an anionic R₂N–CH₂CH₂O⁻ ligand. The Re metal center in ImB2′ acquires also a negative natural atomic charge of −0.249|e|.

Next insertion of a CO₂ molecule into the Re–OCH₂CH₂NR₂ transforms the OCH₂CH₂NR₂ ligand to OC(O)OCH₂CH₂NR₂ yielding ImC2′. The insertion process is slightly exergonic (ΔG = −1.8 kcal mol⁻¹). The Re–OC(O)OCH₂CH₂NR₂ coordination bond is a relatively weak Re–O dative bond with a WBI(Re–O) of 0.485 described by the bonding σ(Re–O) NBO as shown in Scheme 5 with an occupancy of 1.934|e|.

Scheme 5 3D plot of the bonding σ(Re–O) NBO of ImC2′ calculated at the wB97XD/Def2-TZVP level of theory.

The σ(Re–O) NBO is constructed from the interaction of the sp^3.00d^3.39 hybrid orbitals (40.5% p and 45.7% d character) of Re with an sp^1.73 hybrid (63.3% p-character) of the O donor atom of the R₂N–CH₂CH₂OC(O)O ligand and is described as σ(Re–O) = 0.364h_Re + 0.931h_O. The carbonate O atoms of the coordinated R₂N–CH₂CH₂OC(O)O ligand acquiring negative natural atomic charges of −0.744, −0.666 and −0.526|e| are susceptible to protonation, particularly the carboxylate O atoms yielding the mono-ImD2′ and di-protonated ImE2_a′ or ImE2_b′ intermediates. ImE2_a′ and ImE2_b′ differ in the protonation sites of the coordinated R₂N–CH₂CH₂OC(O)O ligand. The successive protonations are exergonic, and the estimated ΔG values are −168.2 kcal mol⁻¹ and −127.5 (−110.6) kcal mol⁻¹. Both ImE2_a′ and ImE2_b′ intermediates are unstable transient species directly releasing the cationic di-protonated [R₂N–CH₂CH₂OC(OH)₂], ImE, species and regenerating the catalytically active ImA2 species. The ImE species participates in reaction step C (Fig. 1) that delivers the desired CO product.

Reaction step B2′′: CO₂ capture

The geometric and free energy reaction profiles calculated for step B2′′ are shown in Fig. 9. The capture of CO₂ by ImA2 in reaction step B2′′ yielding ImB2′′ is slightly exergonic (−4.0 kcal mol⁻¹) and it is achieved by the nucleophilic attack of the electrophilic carbon atom of CO₂ since the Re metal centre acquires a negative natural atomic charge of −0.507|e| and the C atom of the coordinated CO₂ a positive natural atomic charge of 0.737|e|. The NBO population analysis of ImB2′′ revealed that the CO₂ molecule is coordinated to the Re metal centre of the ImA2 catalytic species, forming a Re–CO₂ bond with a WBI(Re–CO₂) of 0.634. The covalent component of the Re–CO₂ bond is described by the bonding σ^α(Re–CO₂) and σ^β(Re–CO₂) NBOs. The σ^α(Re–CO₂) NBO, with an occupancy of 0.920|e|, results from the interaction of sp^2.47d^2.27 hybrid orbitals (42.9% p and 39.4% d character) of Re with an sp^1.79 hybrid (64.1% p-character) on the carbon donor atom of the CO₂ ligand and is described as σ(Re–CO₂) = 0.610h_Re + 0.792h_C. The σ^β(Re–CO₂) NBO, with an occupancy of 0.767|e|, results from the interaction of the sp^4.25d^3.94 hybrid orbitals (46.2% p and 42.8% d character) of Re with an sp^1.96 hybrid (66.2% p-character) on the carbon donor atom of the CO₂ ligand and is described as σ^β(Re–CO₂) = 0.679h_Re + 0.735h_C. The O atoms of the coordinated C(O)O ligand acquiring negative natural atomic charges of −0.646 and −0.746|e| are susceptible to protonation. The first and second protonations transforming the coordinated C(O)O ligand to C(O)OH and C(OH)OH ligands and yielding ImC2′′ and ImD2′′ intermediates respectively are strongly exergonic, with the estimated ΔG values being −184.6 and −149.5 kcal mol⁻¹. ImD2′′ is a transient species directly decomposed to give H₂O and ImE2′′ formulated as the [Re(en)(CO)₃Cl]⁺ complex which by 2e⁻ reduction releases the desired CO product and regenerates the starting ImA2 complex. The two electrons are provided by the TEOA sacrificial electron donor. Note that the proposed catalytic cycle for step B2′′ is exactly the same as a possible catalytic cycle for the photocatalytic reduction of CO₂ to CO catalyzed by the [Re(bpy)(CO)₂Cl] catalytic species generated from the loss of a CO ligand from the reduced form of the [Re(bpy)(CO)₃Cl] complex proposed by Lehn and co-workers.¹⁹

Fig. 9 Geometric and free energy, ΔG (in kcal mol⁻¹), reaction profiles for step B2′′ of the reduction of CO₂ to CO catalyzed by the active [(en)(CO)₃Re]⁻ catalytic species, ImA2, calculated at the wB97XD/Def2-TZVP level in the DMF solution.

In summary the free energy reaction profiles for reaction steps B2′ and B2′′ suggest that both the reaction pathways would proceed in parallel, but the CO₂ insertion pathway should be most favorable than the CO₂ capture pathway, since in the latter the CO₂ capture by the ImA2 catalytic species is difficult and the release of CO from the cationic [Re(en)(CO)₃Cl]⁺, ImE2′′ intermediate should be subject to 2e⁻ reduction, that is the energy consuming process. The CO₂ capture pathway is expected to be the most favorable reaction pathway in the electrochemical reduction of CO₂ to CO mediated by the [Re(en)(CO)₃Cl] catalyst in line with the catalytic cycle for the reduction of CO₂ to CO mediated by the [Re(bpy)(CO)₃Cl] catalyst proposed by Lehn and co-workers.¹⁹

Conclusions

This study reveals the crucial role of the TEOA sacrificial electron donor in the photocatalytic reduction of CO₂ → CO catalyzed by the [Re(en)(CO)₃Cl] complex. Despite its electron and proton donor abilities, TEOA can capture CO₂ upon coordination to the Re metal centre of the “real” catalytic species [Re(en)(CO)₃], [Re(en)(CO)₃]⁻ and/or [Re(en)(CO)₂Cl]⁻ generated from the T₁ state upon single and double reductive quenching by TEOA or photodissociation of a CO ligand. The oxidized R₂NCH₂CH₂O˙ radical of TEOA upon coordination with the Re metal centre of the catalytic species promotes the CO₂ insertion in the Re–OCH₂CH₂NR₂ bond forming stable intermediates that upon successive protonations yield unstable transient species that release CO and regenerate the catalytic species. In particular, in the CO₂ insertion pathways, CO is delivered by an unstable diprotonated [R₂NCH₂CH₂OC(OH)(OH)]⁺ species regenerating TEOA and the catalyst. The CO₂ insertion reaction pathway is the favourable pathway for the photocatalytic reduction of CO₂ → CO catalysed by the [Re(en)(CO)₃Cl] complex in the presence of TEOA manifesting its crucial role as an electron and proton donor, capturing CO₂ and releasing CO. The CO₂ capture reaction pathways involving direct capture of CO₂ by the catalytic species is less favorable because the direct capture of CO₂ is a difficult task and the release of CO from the [Re(en)(CO)₄]⁺ and [Re(en)(CO)₂Cl]⁺ intermediates formed upon successive protonation of the captured CO₂ ligand corresponds to energy consuming processes. This is why the CO₂ capture reaction pathways are favourable pathways in the electrochemical reduction of CO₂ to CO mediated by the [Re(en)(CO)₃Cl] catalyst in line with the catalytic cycle proposed by Lehn and co-workers.¹⁹ We believe that the present in-depth mechanistic investigations would obviously inspire the chemists towards new and more efficient, selective and stable catalytic systems for the CO₂ insertion step and easily releasing the reduction product of CO.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The work has been performed under the Project HPC-EUROPA3 (INFRAIA-2016-1-730897), with the support of the EC Research Innovation Action under the H2020 Programme; in particular, the author gratefully acknowledges the support of Prof. M. Bickelhaupt, Department of Chemistry, Vrije Universiteit Amsterdam and the computer resources and technical support provided by the SURFsara supercomputing centre.

Notes and references

1. M. Aresta and D. Dibenedetto, Dalton Trans., 2007, 2975–2992.
2. M. Cokoja, C. Bruckmeier, B. Rieger, F. Kühnand and W. A. Hermann, Angew. Chem., Int. Ed., 2011, 50, 8510–8537.
3. A. J. Morris, G. J. Meyer and E. Fujita, Acc. Chem. Res., 2009, 42, 1983–1994.
4. J. A. Keith, K. A. Grice, C. P. Kubiak and E. A. Carter, J. Am. Chem. Soc., 2013, 135, 15823–15829.
5. C. Riplinger, M. D. Sampson, A. M. Ritzmann, C. P. Kubiak and E. A. Carter, J. Am. Chem. Soc., 2014, 136, 16285–16298.
6. K. S. Rawat, A. Mahata, I. Choudhuri and B. Pathak, J. Phys. Chem. C, 2016, 120(16), 8821–8831.
7. D. H. Apaydin, S. Schlager, E. Portenkirchner, S. Niyaziand and N. S. Sariciftci, ChemPhysChem, 2017, 18, 3094–3116.
8. N. Elgrishi, M. B. Chambers, X. Wang and M. Fontecave, Chem. Soc. Rev., 2017, 46, 761–796.
9. K. E. Dalle, J. Warnan, J. J. Leung, B. Reuillard, I. S. Karmel and E. Reisner, Chem. Rev., 2019, 119, 2752–2875.
10. Y. Yamazaki, H. Takeda and O. Ishitani, J. Photochem. Photobiol., C, 2015, 25, 106–137.
11. Y. Kuramochi, O. Ishitani and H. Ishida, Coord. Chem. Rev., 2018, 373, 333–356.
12. H. Koizumi, H. Chiba, A. Sugihara, M. Iwamur, K. Nozaki and O. Ishitani, Chem. Sci., 2019, 10, 3080–3088.
13. Y. Asai, H. Katsuragi, K. Kita, T. Tsubomura and Y. Yamazaki, Dalton Trans., 2020, 49, 4277–4292.
14. I. Choudhuri and D. G. Truhlar, J. Phys. Chem. C, 2020, 124(16), 8504–8513.
15. S. Gong, J. Fan, V. Cecenc, C. Huang, Y. Min, Q. Xu and H. Li, Chem. Eng. J., 2021, 405, 126913–112627.
16. H. Yuan, B. Cheng, J. Lei, L. Jiang and Z. Han, Nat. Commun., 2021, 12, 1835–1844.
17. J.-M. Lehn and R. Ziessel, Proc. Natl. Acad. Sci. U. S. A., 1982, 79, 701–704.
18. J. Hawecker, J.-M. Lehn and R. Ziessel, J. Chem. Soc., Chem. Commun., 1983, 536–538.
19. J. Hawecker, J.-M. Lehn and R. Ziessel, Helv. Chim. Acta, 1986, 69, 1990–2012.
20. C. Kutal, M. A. Weber, G. Ferraudi and D. A. Geiger, Organometallics, 1985, 4, 2161–2166.
21. C. Kutal, M. A. Weber, G. Ferraudi and D. A. Geiger, Organometallics, 1987, 6, 553–557.
22. J. M. Smieja and C. P. Kubiak, Inorg. Chem., 2010, 49, 9283.
23. J. M. Smieja, E. E. Bensona, B. Kumara, K. A. Gricea, C. S. Seua, J. M. Miller, J. M. Mayer, P. Clifford and C. P. Kubiaka, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 15646–15650.
24. K. A. Grice, N. X. Gu, M. D. Sampson and C. P. Kubiak, Dalton Trans., 2013, 42, 8498–8503.
25. E. E. Benson, K. A. Grice, J. M. Smieja, P. Clifford and C. P. Kubiak, Polyhedron, 2013, 58, 229–234.
26. J. Agarwal, E. Fujita, H. F. Schaefer III and J. T. Muckerman, J. Am. Chem. Soc., 2012, 134, 5180–5186.
27. H. Takeda, K. Koike, H. Inoue and O. Ishitani, J. Am. Chem. Soc., 2008, 130, 2023–2031.
28. T. Morimoto, T. Nakajima, S. Sawa, R. Nakanishi, D. Imori and O. Ishitani, J. Am. Chem. Soc., 2013, 135, 16825.
29. P. Kurz, B. Probst, B. Spingler and R. Alberto, Eur. J. Inorg. Chem., 2006, 2966–2974.
30. Y. Kuramochi, K. Fukaya, M. Yoshida and H. Ishida, Chem. – Eur. J., 2015, 21, 10049–10060.
31. M. J. Frisch, et al., Gaussian 16, Revision C.01, Gaussian, Inc., Wallingford, CT, 2010.
32. S. Grimme, J. Comput. Chem., 2006, 27, 1787–1799.
33. D. Becke, J. Chem. Phys., 1997, 107, 8554–8560.
34. D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008, 10, 6615–6620.
35. Q. Wu and W. T. Yang, J. Chem. Phys., 2002, 116, 515–524.
36. Y. Minenkov, A. Singstad, O. Occhipinti and V. Jensen, Dalton Trans., 2012, 41, 5526–5541.
37. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3093.
38. A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988, 88, 899–926.
39. F. Weinhold, in The Encyclopedia of Computational Chemistry, ed. P. V. R. Schleyer, John Wiley & Sons, Chichester, U.K., 1998.
40. S. J. A. van Gisbergen, F. Kootstra, P. R. T. Schipper, O. V. Gritsenko, J. G. Snijders and E. J. J. Baerends, Phys. Rev. A, 1998, 57, 2556–2571.
41. C. Jamorski, M. E. Casida and D. R. Salahub, J. Chem. Phys., 1996, 104, 5134–5147.
42. R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett., 1996, 256, 454–464.

---

Footnote

† Electronic supplementary information (ESI) available: Reaction steps for the photocatalytic reduction of CO₂ to CO catalyzed by the [(en)(CO)₃ReCl] catalyst starting with the [(en)(CO)₃Re] intermediate resulted upon one electron reduction of the T₁ state by TEOA (Fig. S1). Cartesian coordinates of the reactants, intermediates and products involved in the catalytic cycles (Table S1). See DOI: 10.1039/d1dt02188e

---