The Role of Geometry in Cobalt–Polypyridine Complexes in the Electrochemical Reduction of CO₂ Using UV-Vis Spectroelectrochemistry

Abstract

This work explores the effect of geometry and the presence of a site available for carbon dioxide coordination in molecular catalysis of CO₂ reduction for cobalt complexes using electrochemical and spectroelectrochemical studies. The octahedral complexes [Co^II(bztpen)Br]PF₆ and [Co^II(bpy)₃](BF₄)₂, along with the trigonal bipyramidal complex [Co^II(TPA)Cl]Cl, were selected for this study (where bztepen = N-benzyl-N,N′,N′-tris-(pyridine-2-ylmethyl)-ethylenediamine), TPA = tris (2-pyridimethyl)-amine, and bpy = 2′-2′- pyridine). DFT calculations were performed to predict the geometries of the complexes and to propose the sites at which electron transfer occurs. Among the studied compounds, [Co^II(bpy)₃(BF₄)₂] exhibited the highest catalytic rate constant for CO₂ reduction (k = 1.22 × 10² M⁻¹·s⁻¹) compared to [Co^II(bztpen)Br]PF₆ (k = 8.93 × 10¹ M⁻¹·s⁻¹). The trigonal bipyramidal complex [Co^II(TPA)Cl]Cl presented the lowest catalytic rate constant for CO₂ reduction (k = 1.47 × 10¹ M⁻¹·s⁻¹). UV-Vis spectroelectrochemical studies and DFT calculations suggested the formation of [Co^I(bztpen)(CO)]⁺ and [Co^I(tpa)(CO)]⁺ species, which are associated with catalyst deactivation and may account for the lower performance of CO₂ reduction.

1. Introduction

The greenhouse effect has contributed to a global temperature increase of 1.1 °C between 2010 and 2020 [1]. Within this context, CO₂ is identified as the primary agent associated with this effect. Data indicate that CO₂ emissions increased from 20 billion tons in 1990 to 41.6 billion tons in 2024 [2]. Since 2010, the atmospheric concentration of CO₂ has been increasing at an average annual rate of approximately 2 ppm, mainly attributed to coal combustion [3,4]. In response to this challenge, various strategies have been developed to mitigate CO₂ emissions that include chemical and physical absorption methods. However, they face significant limitations, including high energy demands for absorbent regeneration [5]. Consequently, catalysis for CO₂ conversion has emerged as a promising alternative. In particular, the transformation of CO₂ into liquid fuels, which can serve as feedstock for other industrial processes, has garnered considerable attention. This approach, often facilitated by metal complexes, is especially interesting due to its high efficiency [6,7].

Carbon dioxide is the product of the complete oxidation of carbon. As a small, stable, and inert molecule, it poses significant challenges for efficient activation. In its ground state, CO₂ exhibits a linear geometry and a standard Gibbs free energy of formation (ΔG^°_f) of −396 kJ/mol [8,9,10]. Consequently, CO₂ transformation into valuable products requires substantial energy consumption. In this context, CO₂ reduction involves multi-electron transfer processes that yield products such as formic acid, carbon monoxide, oxalic acid, methanol, or methane [7,11,12].

$$\mathrm{ }E°\left(V\right)vs\mathrm{ }NHE$$

$$C{O}_2+2H^{+}+2\mathrm{ }{e}^{-}\mathrm{ }\to HC{O}_{2}H\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }-0.61$$

$$C{O}_2+2H^{+}+2\mathrm{ }{e}^{-}\mathrm{ }\to CO+\mathrm{ }{H}_{2}O\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }-0.52$$

$$C{O}_2+4H^{+}+4\mathrm{ }{e}^{-}\mathrm{ }\to HCHO+\mathrm{ }{H}_{2}O\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }-0.48$$

$${CO}_2+6H^{+}+6\mathrm{ }{e}^{-}\mathrm{ }\to C{H}_{3}OH+\mathrm{ }{H}_{2}O\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }-0.38$$

$${CO}_2+8H^{+}+8\mathrm{ }{e}^{-}\mathrm{ }\to C{H}_4+\mathrm{ }{2H}_{2}O\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }-0.24$$

These reactions are particularly challenging due to involving multiple electron transfers and coupled chemical steps such as protonation [7,12]. An important parameter in these processes is the standard redox potential for the formation of the CO₂^•− radical anion, which is −1.97 V vs. SHE in DMF and −1.90 V vs. SHE in water. In both cases, the electrochemical process’s high overpotentials are required due to high reorganization energies [7,12,13]. The activation energy barrier for the formation of this chemical species can be decreased either using modified electrodes with semiconductor and transition metal electrocatalysts or using mercury electrodes [13,14,15]. However, these strategies are compromised by the mercury toxicity and lack of control of the chemical species responsible for the electrocatalytic reduction.

Molecular catalysis offers an alternative approach to promote CO₂ reduction, utilizing molecules as catalysts either in solution or immobilized on the electrode surface as a monolayer or multilayer [13]. A molecular catalyst is defined as an electron transfer agent with a redox potential value (E⁰) close to a specific target reaction. Additionally, it must exhibit a high homogeneous electron transfer rate constant, k_cat, for the target reaction. Furthermore, the heterogeneous electron transfer rate, k°, for the oxidation or reduction of the molecular catalyst must also be high. The characteristics already mentioned can be optimized using coordination compounds [7,11,15].

The initial step in CO₂ reduction in molecular catalysis using transition metal complexes involves coordination of CO₂ to the reduced metal center (Mⁿ⁺), forming an intermediate species M^m+L(CO₂). The most common coordination mode occurs through σ bonding between the metal center and the carbon atom. During this process, bond cleavage can lead to CO release, or alternatively, protonation of the reduced catalyst may form a metal hydride intermediate (H–Mⁿ⁺L). This species can follow two competing pathways: reaction with CO₂ to yield formate (HCO₂⁻) or reaction with H⁺ to produce H₂. The acidity of the medium critically influences the distribution of products such as carbon monoxide, formate, and hydrogen gas [14].

Several key factors must be considered in the design of transition metal-based molecular catalysts for CO₂ reduction. First, the stability of the complexes during preparation, isolation, and purification. Second, the presence of one or more vacant coordination sites for CO₂ binding [12].

Ternary coordination compounds employing macrocyclic, pyridinic, or phosphine ligands have been developed as molecular catalysts [14,15,16,17]. Among these, Fe⁽⁰⁾ porphyrins are the most extensively studied, demonstrating both efficiency and stability in selectively converting CO₂ to CO, particularly in the presence of Lewis or weak Brønsted acids [14]. Additionally, Ir and Pd pincer complexes in water–acetonitrile mixtures exhibit high catalytic activity for CO₂ reduction, achieving faradaic yields of up to 90%. Formate is the main product, although H₂ evolution due to water reduction also occurs [10].

Tetraazamacrocyclic cobalt and nickel complexes have achieved faradaic yields of up to 98% for CO₂ reduction to CO, with reduction potentials ranging from −1.3 to −1.6 V vs. SCE [14,15]. Re, Ru, Ir, Rh, Os, Pd, Mo, Cu, Co, Ni, Mn, and Fe complexes with similar ligands have been explored, yielding products such as oxalate, carbon monoxide, and occasionally, formic acid [18].

Molecular catalysts based on Re, Rh, Os, and Ru complexes with polypyridine ligands have also been reported for CO₂ reduction, where formic acid and formate are the principal products. Their effectiveness is attributed to the ability of these complexes to stabilize low oxidation states and store electrons in π orbitals [15]. Complexes with 1,10-phenanthroline (phen) and earth-abundant metals such as Co, Ni, Fe, and Cu have been proposed as promising low-cost molecular catalysts for CO₂ reduction in comparison with platinum group derivatives [19,20].

Cobalt has been extensively investigated as an efficient catalyst for CO₂ reduction in various forms, including single-atom catalysts, multi-metals, Co-based metal–organic frameworks (MOFs), cobalt oxides, and cobalt complexes, due to its cost-effectiveness, versatility, and high efficiency in producing carbon monoxide and formate [21]. Furthermore, the [Co(triphos)(bdt)]⁺ complex has been reported as an effective CO₂ reduction catalyst, with its geometry playing a key role in selectively producing formate as the main product [22].

In a recent work, we investigated, using cyclic voltammetry, the role of redox potential of a series of Co(II) octahedral complexes [Co(L)₃]²⁺ and [Co(L′)₂]²⁺, where L = 2,2′-bpyridine, 1,10-phenanthroline, 3,4,7,8-tetramethyl-1,10-phenanthroline, 5,6-dimethyl-1,10-phenanthroline, and 4,7-diphenyl-1,10-phenanthroline and L′ = terpyridine and 4-chloro-terpyridine in the mediated electrochemical reduction of CO₂ [23]. It was found that metal complexes with more negative redox potential enhance the catalytic performance for CO₂ reduction. However, no other geometries, the presence of a site available for CO₂ coordination, and the presence of intermediates using UV-Vis spectroelectrochemistry were explored. This technique combines in situ UV-Vis spectroscopy with potential step electrolysis to observe changes in the electronic structure of substances or the presence of new species that occur during a redox process. Such studies can provide information for the design of catalysts with abundant earth metals for electrochemical CO₂ reduction.

The objective of the presented work is to explore, for cobalt complexes, the effect of the complex geometry and the presence of a site available for carbon dioxide coordination in molecular catalysis of CO₂ using electrochemical and coupled UV-Vis spectroelectrochemical techniques that can give information about the presence of intermediates generated during the CO₂ electrochemical reduction. For this purpose, the octahedral complexes [Co^II(bztpen)Br]PF₆ and [Co^II(bpy)₃](BF₄)₂ and the trigonal bipyramidal complex [Co^II(TPA)Cl]Cl were selected (where bztepen = N-benzyl-N,N′,N′-tris-(pyridine-2-ylmethyl)-ethylenediamine and TPA = tris-(2-pyridymethyl)-amie)bpy = 2′-2′- pyridine). The chemical structures of the compounds [Co^II(bztpen)Br]PF₆, [Co^II(TPA)Cl]Cl, and [Co^II(bpy)₃](BF₄)₂ are shown in Figure 1. DFT calculations were also carried out to predict geometry in complexes and to propose sites where the electron transfer takes place.

2. Results and Discussion

2.1. Complexes’ Characterization

2.1.1. IR Spectroscopy

The IR spectra of the metal complexes [Co^II(Bztpen)Br]PF₆, [Co^II(tpa)Cl]Cl, and [Co^II(bpy)₃](BF₄)₂ exhibit characteristic vibrational frequencies of the ligands (Figure S1). The bond stretching ν(=C-H) is observed in the range of 3000–3100 cm⁻¹, while the overtone vibrations σ(=CH) appear between 1600 and 2000 cm⁻¹. The stretching bands ν(C=C) and ν(C=N) of the polypyridine rings are recorded in the 1420–1640 cm⁻¹ region. Additionally, out-of-plane bending vibrations of the aromatic C-H bonds are observed between 856 and 700 cm⁻¹. Intense signals at approximately 1061 cm⁻¹ and 831 cm⁻¹ correspond to the B-F and P-F stretching vibrations of the BF₄⁻ and PF₆⁻ counterions, respectively.

2.1.2. UV-Vis Spectroscopy

The electronic spectra of the metal complexes were recorded in acetonitrile solution. The [Co^II(bpy)₃](BF₄)₂ complex exhibits electronic transitions at ν₁ = ⁴T₁g(P)→⁴T₂g(F) around 910 nm and ν₂ = ⁴T₁g(P)→⁴T₁g(F) at 450 nm, resulting in a ν₁/ν₂ ratio of 2.02, typical for octahedral configurations [24,25]. The ν₃ transition is not observed due to overlap with the intense metal-to-ligand charge transfer (MLCT) band. For the complex [Co(Bztpen)Br]PF₆, only the electronic transition ν₂ = ⁴T₁g(P)→⁴T₁g(F), typical for octahedral geometry, was detected at 515 nm.

For the [Co^II(tpa)Cl]Cl complex, the electronic spectrum displays a broad absorption band around 276 nm, corresponding to the π→π* transition of the TPA ligand. Additionally, two absorption bands at 492 nm and 628 nm attributed to the d-d transitions are observed [7]. These spectral features are consistent with a five-coordinated cobalt (II) center adopting a trigonal bipyramidal (TBP) geometry and point group D_3h, in agreement with previous reports [26,27,28,29,30].

2.2. Cyclic Voltammetry Response of Metal Complexes

2.2.1. Cyclic Voltammetry of [Co^II(bztpen)Br]PF₆

Cyclic voltammetry experiments were carried out at a complex concentration of 1 mM in a MeCN solution with 0.1 M TBAPF₆ as the supporting electrolyte. Potential scan was initiated from the open circuit potential in a negative direction. All voltammograms at scan rates ranging from 50 to 1500 mV·s⁻¹ are presented in a normalized current representation (i/v^1/2). Figure 2a displays typical cyclic voltammograms for [Co(bztpen)Br]PF₆. At a scan rate of 100 mV·s⁻¹, one reduction peak (I_c) and one oxidation peak (I_a) were observed at E_p(_Ic) = −1.857 V and E_p(_Ia) = −1.757 V vs. Fc/Fc⁺, respectively, with a ΔE_p = 0.103 V. After completing the scan, additional redox signals appeared (II_a and II_c) at E_p(_IIa) = −0.026 V and E_p(_IIc) = −0.107 V vs. Fc/Fc⁺, with ΔE_p = 0.081 V. In both processes, ΔE_p > 0.060 V indicates a quasi-reversible electron transfer. Additionally, increasing the scan rate causes a shift in the peak potential toward more negative and more positive values, resulting in a large ΔE_p at higher scan rates consistent with electrochemical quasi-reversibility. According to the literature, the following mechanism is proposed [23,31,32]:

$${\left[{Co}^{II}\left(bztpen\right)Br\right]}^{+}+ 1e^{-} \to \left[{Co}^I\left(bztpen\right)Br\right] \mathrm{ }\mathrm{I}\mathrm{ }$$

$${ \left[{Co}^{II}\left(bztpen\right)Br\right]}^{+}\to \left[{Co}^{III}\left(bztpen\right)Br\right]+ 1e^{-} \mathrm{I}\mathrm{I}\mathrm{ }$$

2.2.2. Cyclic Voltammetry of [Co^II (tpa)Cl]Cl

For the complex [Co^II(tpa)Cl]Cl (Figure 2b), the cyclic voltammogram in normalized current representation at a scan rate of 100 mV·s⁻¹ exhibits three reduction peaks (I_c, II_c, and III_c) at E_pc(_Ic) = −1.808 V, E_pc(_IIc) = −2.028 V, and E_pc(_IIIc) = −2.303 V vs. Fc/Fc⁺, along with three corresponding oxidation peaks (I_a, II_a, and III_a) at E_pa(_Ia) = −1.702 V, E_pa(_IIa) = −1.932 V, and E_pa(_IIIa) = −2.213 V vs. Fc/Fc⁺. Additionally, two extra signals (I*_c and I*_a) are observed at E_pc(_Ic) = −1.12 V and E_pa(_Ia) = −0.718 V vs. Fc/Fc⁺.

The normalized current for the oxidation signals suggests the involvement of coupled chemical reactions, consistent with an electrochemical–chemical–electrochemical mechanism (ECE). The additional signals I*_c and I*_a are attributed to surface-adsorbed species on the electrode. Based on the literature, the following mechanism is proposed [23,31,32]:

$${ \left[{Co}^{II}\left(tpa\right)Cl\right]}^{+}+ 1e^{-} \to \left[{Co}^I\left(tpa\right)Cl\right] \mathrm{E}\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{I}\mathrm{ }\mathrm{ }$$

$$\left[{Co}^I\left(tpa\right)Cl\right] \to { \left[{Co}^I\left(tpa\right)\right]}^{+} + {Cl}^{-} C$$

$${ \left[{Co}^I\left(tpa\right)\right]}^{+}+1e^{-}\to \left[{Co}^0\left(tpa\right)\right] \mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{E}\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{I}\mathrm{I}$$

$$\left[{Co}^0\left(tpa\right)\right]+ 1e^{-} \to \left[{Co }^0\left({tpa}^{-}\right)\right] \mathrm{E}\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{I}\mathrm{I}\mathrm{I}\mathrm{ }$$

2.2.3. Cyclic Voltammetry of [Co^II(bpy)₃](BF₄)₂

For the complex [Co^II(bpy)₃](BF₄)₂, its electrochemical response (Figure 2c) at a scan rate of 100 mV·s⁻¹ displays three reduction peaks at E_pc(_Ic) = −1.38 V, E_pc(_IIc) = −1.995 V, and E_pc(_IIIc) = −0.098 V vs. Fc/Fc⁺, along with three corresponding oxidation peaks at E_pa(_Ia) = −1.322 V, E_pa(_IIa) = −1.928 V, and E_pa(_IIIa) = −0.01 V vs. Fc/Fc⁺. Process I is attributed to a reversible electron transfer, as indicated by ΔEp = 0.057 V. In contrast, processes II and III exhibit ΔEp > 0.060 V, suggesting quasi-reversible behavior. Based on the literature, the following mechanism is proposed [23,31,32]:

$${\left[{Co}^{II}{\left(bpy\right)}_3\right]}^{2+} \rightleftarrows {\left[{Co}^{III}{\left(bpy\right)}_3\right] }^{3+}+ 1e^{-} \mathrm{ }\mathrm{I}\mathrm{I}\mathrm{I}\mathrm{ }$$

$${\left[{Co}^{II}{\left(bpy\right)}_3\right]}^{2+}+ 1e^{-}\rightleftarrows {\left[{Co}^I{\left(bpy\right)}_3\right] }^{+} \mathrm{ }\mathrm{I}\mathrm{ }\mathrm{ }\mathrm{ }$$

$${ \left[{Co}^I{\left(bpy\right)}_3\right]}^{+} + 2e^{-}\rightleftarrows {\left[{Co}^I{\left(bpy\right)\left({bpy}^{-}\right)}_2\right] }^{-} \mathrm{ }\mathrm{ }\mathrm{I}\mathrm{I}\mathrm{ }$$

$${\left[{Co}^I{\left(bpy\right)\left({bpy}^{-}\right)}_2\right] }^{-}\to {\left[{Co}^I{\left({bpy}^{-}\right)}_2\right] }^{-}+\left(bpy\right)$$

2.3. Cyclic Voltammetry Response of Complexes in the Presence of CO₂

2.3.1. Cyclic Voltammetry of [Co^II(bztpen)Br]PF₆ in the presence of CO₂

After that, the electrochemical behavior of the cobalt complexes was presented, and their ability for electrochemical molecular catalysis for CO₂ reduction was evaluated. The voltammogram of [Co(bztpen)Br]PF₆ in the presence of CO₂ is shown in Figure 3a. At more negative potentials, where the reduction [Co^II(bztpen)Br]⁺ +1e⁻ → [Co^I(bztpen)Br] takes place, its E_pc value shifts to a more positive value (−1.814 V vs. Fc/Fc⁺), accompanied by a significant increase in current. Additionally, the corresponding oxidation peak (I_a) was not observed. This behavior can be associated with typical molecular catalysis of CO₂. A new reduction signal, I*c, which is associated with an adsorbed species, was observed. When the scan was inverted, a new oxidation signal, $I_a^{\mathrm{*}}$, also associated with adsorbed species, was detected. The anhydrous conditions allow us to discard catalytic activity for electrochemical proton reduction [33].

2.3.2. Cyclic Voltammetry of [Co^II (tpa)Cl]Cl in the presence of CO₂

For the [Co(tpa)Cl]Cl complex (Figure 3b), CO₂ molecular catalysis begins at a slightly more negative potential than the corresponding electron transfer [Co^II(tpa)Cl]⁺ +1e⁻ → [Co^I(tpa)Cl], recorded at E_pc(I_c′) = −1.865 V vs. Fc-Fc⁺. At more negative potentials, the voltammogram displays two additional reduction signals, E_pc(II_c′) = −2.202 and E_pc(III_c′) = −2.235 V vs. Fc-Fc⁺, indicating high molecular catalysis activity by highly reduced species, as reported in the literature [26,27,34]. No oxidation peaks are observed in the reverse scan.

2.3.3. Cyclic Voltammetry of [Co^II (bpy)₃(BF₄)₂] in the presence of CO₂

For [Co(bpy)₃(BF₄)₂] (Figure 3c), a catalytic current for CO₂ reduction is observed at −1.994 V vs. Fc/Fc⁺, corresponding to the reduction process [Co^I(bpy)₃]⁺ + 2e⁻ → [Co^I(bpy)(bpy⁻)₂]⁻. This behavior is consistent with previous reports [23,35].

A summary of the redox potentials of complexes and the onset potential related to CO₂ reduction is shown in

Table 1. Redox potential, electrochemical reaction, and onset potential for compounds studied in this work.

Compound	Redox Potential (V vs. Fc/Fc+)	Onset Potential (V vs. Fc/Fc+)
[CoII(bztpen)Br]PF6
${\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}\mathrm{I}}\left(\mathrm{b}\mathrm{z}\mathrm{t}\mathrm{p}\mathrm{e}\mathrm{n}\right)\mathrm{B}\mathrm{r}\right]}^{+}+1{\mathrm{e}}^{-}\to \left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}}\left(\mathrm{b}\mathrm{z}\mathrm{t}\mathrm{p}\mathrm{e}\mathrm{n}\right)\mathrm{B}\mathrm{r}\right]$	−1.807	−1.640
${\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}\mathrm{I}}\left(\mathrm{b}\mathrm{z}\mathrm{t}\mathrm{p}\mathrm{e}\mathrm{n}\right)\mathrm{B}\mathrm{r}\right]}^{+}\to \left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}\mathrm{I}\mathrm{I}}\left(\mathrm{b}\mathrm{z}\mathrm{t}\mathrm{p}\mathrm{e}\mathrm{n}\right)\mathrm{B}\mathrm{r}\right]+1{\mathrm{e}}^{-}$	−0.0655	n.o.
[CoII(tpa)Cl]Cl
${\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}\mathrm{I}}\left(\mathrm{t}\mathrm{p}\mathrm{a}\right)\mathrm{C}\mathrm{l}\right]}^{+}+1{\mathrm{e}}^{-}\to \left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}}\left(\mathrm{t}\mathrm{p}\mathrm{a}\right)\mathrm{C}\mathrm{l}\right]$	−1.755	−1.633
${\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}}\left(\mathrm{t}\mathrm{p}\mathrm{a}\right)\right]}^{+}+1{\mathrm{e}}^{-}\to \left[{\mathrm{C}\mathrm{o}}^0\left(\mathrm{t}\mathrm{p}\mathrm{a}\right)\right]$	−1.980	−1.951
$\left[{\mathrm{C}\mathrm{o}}^0\left(\mathrm{t}\mathrm{p}\mathrm{a}\right)\right]+1{\mathrm{e}}^{-}\to \left[{\mathrm{C}\mathrm{o}}^0\left({\mathrm{t}\mathrm{p}\mathrm{a}}^{-}\right)\right]$	−2.258	−2.255
[CoII(bpy)3](BF4)2
${\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}\mathrm{I}}{\left(\mathrm{b}\mathrm{p}\mathrm{y}\right)}_3\right]}^{2+}+1{\mathrm{e}}^{-}\rightleftarrows {\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}}{\left(\mathrm{b}\mathrm{p}\mathrm{y}\right)}_3\right]}^{+}$	−1.351	n.o.
${\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}}{\left(\mathrm{b}\mathrm{p}\mathrm{y}\right)}_3\right]}^{+}+2{\mathrm{e}}^{-}\rightleftarrows {\left[{\mathrm{C}\mathrm{o}}^{\mathrm{I}}{\left(\mathrm{b}\mathrm{p}\mathrm{y}\right)\left({\mathrm{b}\mathrm{p}\mathrm{y}}^{-}\right)}_2\right]}^{-}$	−1.961	−1.667

Data is reported at 100 mVs⁻¹ in MeCN solution with 0.1 M TBAPF₆ as the supporting electrolyte. A glassy carbon disk was used as the working electrode; n.o. = not observed.

, where the complex [Co(tpa)Cl]Cl presented the lowest value of overpotential for the CO₂ reduction in the first electron transfer. The highest overpotential value corresponds to the complex [Co(bpy)₃](BF₄)₂.

2.3.4. FOWA

Cyclic voltammetry experiments at variable scan rates were performed to evaluate the catalytic constant for electrochemical CO₂ reduction (Figure 4). For the [Co^II(bztpen)Br]PF₆ complex (Figure 4a), the voltammograms show that when the scan rate was increased, the E_pc(I_c″) shifted to negative values, with a decrease in the catalytic current. This behavior indicated that the molecular catalysis is decoupled at high scan rates. In the case of [Co^II(tpa)Cl]Cl (Figure 4b), an initial increase in the catalytic current is observed at 100 mV·s⁻¹. However, beyond 250 mV·s⁻¹, the redox processes involved in CO₂ molecular catalysis begin to decouple, resulting in a diminished catalytic current. For [Co^II(bpy)₃(BF₄)₂] (Figure 4c), increasing the scan rate causes a minor cathodic shift in the catalytic peak, but the current intensity remains relatively constant, suggesting a more stable catalytic behavior.

From the voltammograms recorded at variable scan rates, foot of the wave analysis (FOWA) [36,37] was carried out for the reduction processes I_c in the complex [Co^II(bztpen)Br]PF₆, for the process II_c in the complex [Co^II(bpy)₃(BF₄)₂, and for the process Ic in the complex [Co^II(tpa)Cl]Cl (see Figure 5a,c,e). The catalytic rate constant (k) was calculated using Equation (1) from the slope of the plot ${\displaystyle \frac{i_k}{i_p^{°}}}vs{\displaystyle \frac{1}{1+\mathit{exp}\left[\frac{F}{RT}\left(E-E°\frac{M{L}_{red}}{M{L}_{ox}}\right)\right]}}$, where i_k is the catalytic current and i_p° is the diffusion-controlled peak current in the absence of substrate:

$${\displaystyle \frac{i_k}{i_p^{°}}}={\displaystyle \frac{2.24\sqrt{\frac{RT}{F\nu }2k{C}_{{CO}_2}^0}}{1+exp\left[\frac{F}{RT}\left(E-E°\frac{M{L}_{red}}{M{L}_{ox}}\right)\right]}}$$

(1)

The shape of the FOWA curves for the [Co^II(bztpen)Br]PF₆ complex (Figure 5a) indicates that the catalytic reaction at high potential values presents the deactivation of the catalyst due to parallel reactions. For an ideal catalytic system, a linear response is expected. In contrast, the FOWA curves for [Co^II(bpy)₃(BF₄)₂] (Figure 5c) exhibit a response closer to the ideal even at lower scan rates, indicating a more efficient catalytic process under these conditions compared to [Co^II(bztpen)Br]PF₆. From the slope of the low-potential region of the FOWA, the catalytic rate constant k was calculated. The resulting values were 8.93 × 10¹ M⁻¹·s⁻¹ for [Co^II(bztpen)Br]PF₆ (Figure 5b) and 1.22 × 10² M⁻¹·s⁻¹ for [Co^II(bpy)₃(BF₄)₂] (Figure 5d). For [Co^II(tpa)Cl]Cl (Figure 5e), a catalytic rate constant k value was 1.47 × 10¹ M¹·s⁻¹ in the process Ic (Figure 5f), considering the lowest efficient catalytic process.

2.4. UV-Vis Spectroelectrochemical Response of the Complexes in the Presence and in the Absence of CO₂

2.4.1. UV-Vis Spectroelectrochemical Response of [Co^II(Bztpen)Br]PF₆

First, Uv-Vis spectroelectrochemical experiments were performed using a 1 mM solution of [Co^II(bztpen)Br]PF₆ in MeCN with 0.1 M TBAPF₆ in the presence of N₂. A gold mesh was used as the working electrode, and chronoamperometry measurements were performed by applying a constant potential of −1.95 V vs. Fc/Fc⁺. Simultaneously, UV-Vis spectra were recorded every 3 s during electrolysis at the electrode surface for a total duration of 3 min (Figure 6). The absorbance bands at 260, associated with the ligand [38], remained unchanged throughout the experiment, indicating that the electron transfer is located on the metal center (Figure S2).

In contrast, under a CO₂ atmosphere, a slight hypsochromic shift of the ligand band was observed, and a new absorption band is recorded at 400 nm (Figure 7 and Figure S3). These spectral changes are consistent with the formation of a new coordination species. The appearance of the band at 400 nm is attributed to Co^(I)L species, in agreement with previous reports [31]. This observation supports the formation of a [Co^I(Bztpen)CO] complex, yielded during the molecular catalysis of CO₂. Taking these results into consideration, the following mechanism is proposed:

$${\left[{Co}^{II}\left(bztpen\right)Br\right]}^{+}+ 1e^{-} \to \left[{Co}^I\left(bztpen\right)Br\right]$$

$$\left[{Co}^I\left(bztpen\right)Br\right]+ {CO}_2 \to {\left[{Co}^I\left(bztpen\right){CO}_2\right]}^{+}+{Br}^{-}$$

$${\left[{Co}^I\left(bztpen\right){CO}_2\right]}^{+}+ {CO}_2+2e^{-}\to {\left[{Co}^I\left(bztpen\right)CO\right]}^{+}+{C{O}_3}^{2-}$$

2.4.2. UV-Vis Spectroelectrochemical Response of [Co^II(tpa)Cl]Cl

Uv-Vis spectroelectrochemical experiments for [Co^II(tpa)Cl]Cl in MeCN containing 0.1 M TBAPF₆ were carried out at the potentials corresponding to signals IIc (–2.06 V vs. Fc/Fc⁺) and IIIc (–2.24 V vs. Fc/Fc⁺), both in the presence and absence of CO₂. Under N₂ atmosphere, the spectral response closely resembled that of the [Co^II(bztpen)Br]PF₆ complex (Figure 8, Figure 9, Figures S6 and S10), with no significant changes in the absorbance values of the ligand-associated electronic transition, irrespective of the applied potential.

In contrast, under a CO₂ atmosphere (Figure 10, Figure 11, Figures S7 and S11), the ligand signal exhibited a slight hypsochromic shift, and a new absorption band was detected at 400 nm, like the behavior observed for the [Co^II(bztpen)Br]PF₆ complex. The intensity of this band increased similarly at both applied potentials, indicating that the first reduced species plays a key role in controlling the molecular catalysis process of CO₂. Based on these findings, the following mechanism is proposed:

$${\left[{Co}^I\left(tpa\right)\right]}^{+}+{CO}_2 \to {\left[{Co}^I\left(tpa\right){CO}_2\right]}^{+}$$

$${\left[{Co}^I\left(tpa\right){CO}_2\right]}^{+}+ {CO}_2+2e^{-}\to {\left[{Co}^I\left(tpa\right)CO\right]}^{+}+{C{O}_3}^{2-}$$

2.4.3. UV-Vis Spectroelectrochemical Response of [Co^II(bpy)₃](BF₄)₂

Furthermore, the electrochemical behavior of the complex [Co^II(bpy)₃](BF₄)₂ was studied. Upon applying a potential of −1.55 V and −2.25 V vs. Fc/Fc⁺, a slight shift in the absorption band associated with the bipyridine ligand reduction was observed, moving from 290 to 300 [38]. Additionally, a new absorption band appeared at 380 nm, which is attributed to a metal-to-ligand charge (MLCT), and 600 nm, which is attributed to a d-d transition [39]. Similar spectral changes were observed at both potentials in the presence of CO₂ (see Figures S14–S23), suggesting that no CO-coordinated species are yielded during molecular catalysis. Based on these results, the following reaction is proposed for the CO₂ reduction:

$${\left[{Co}^I{\left(bpy\right)\left({bpy}^{-}\right)}_2\right]\mathrm{ }}^{-}+\mathrm{ }1e^{-}+\mathrm{ }2{CO}_2\to \left[{Co}^I{\left(bpy\right)}_2\left({bpy}^{-}\right)\right]+CO+{C{O}_3}^{2-}$$

2.4.4. Kinetics of the Formation of Intermediate Species

The spectroelectrochemical data indicate that the formation of new species and the degradation of the starting complex during electrolysis follow an exponential trend (Figures S4, S5, S8, S9, S12, S13, S19, S24 and S25) consistent with first-order kinetics [40]. Based on this observation, the data were fitted to a first-order kineti equation, and the kinetic rate constant (k) was extracted and summarized in

Table 2. Kinetic rate constants determined from the spectroelectrochemical response of the compounds.

Compound	Potential Applied (V vs. Fc/Fc+)	Wavelength (nm)	k (s−1)
[CoII(bztpen)Br]PF6	−1.95	400	0.026 ± 0.001
[CoII(tpa)Cl]Cl	−2.06	400	0.037 ± 0.001
[CoII(tpa)Cl]Cl	−2.24	400	0.015 ± 0.001
[CoII(bpy)3](BF4)2	−1.55	380	0.008 ± 0.002
[CoII(bpy)3](BF4)2	−2.25	380	0.0001 ± 0.002
[CoII(bztpen)Br]PF6	−1.95	260	n.o.
[CoII(tpa)Cl]Cl	−2.06	260	0.031 ± 0.002
[CoII(tpa)Cl]Cl	−2.24	260	0.021 ± 0.002
[CoII(bpy)3](BF4)2	−1.55	245	0.020 ± 0.002
[CoII(bpy)3](BF4)2	−2.25	245	n.o.

All experiments were conducted under identical conditions to ensure consistency and comparability with the results. n.o. not observed.

. The values obtained are in close agreement, indicating that the reactions proceed via a consistent mechanistic pathway. Furthermore, the values obtained at 400 nm are indicative of the catalytic reaction of CO₂ yielding by the species [Co^I(bztpen)(CO)] and [Co^I(tpa)(CO)] ⁺.

These results agree with the behavior of FOWA curves, where the deactivation of the catalyst for the complex [Co^II(bztpen)Br]PF₆ was proposed, which can be attributed to the formation of the chemical species [Co⁰(bztpen)(CO)]. The opposite is observed in FOWA curves for [Co^II(bpy)₃(BF₄)₂], where no deactivation was proposed due to coordination of CO.

2.5. DFT Calculations

DFT calculations were performed for the cationic complexes [Co(bztpen)Br]⁺, [Co(tpa)Cl]⁺, and the intermediate species [Co^I(bztpen)(CO)]⁺ and [Co^I(tpa)(CO)]⁺ to support the electrochemical findings (Figure 12a–d). The LUMO analysis shows that the electronic density is primarily localized on the cobalt center and the polypyridine ligands. This suggests that the first reduction process predominantly occurs at the metallic center.

By means of DFT-DT, the UV-Vis spectra were calculated for intermediate species electrogenerated during electrolysis in the presence of CO₂. The optimized chemicals [Co^I(bztpen)(CO)]⁺ and [Co^I(tpa)(CO)]⁺ (see Figure 12c,d) presented an electronic transition close to 400 nm, which is very similar to the signal recorded at spectroelectrochemical experiments in the presence of CO₂. The presence of Co(0) species must be neglected because electronic transitions are predicted to be around 600 nm (Figures S26 and S27). Finally, Mulliken charges of the [Co^II(bztpen)Br], [Co^I(bztpen)(CO)] ⁺, [Co^II(tpa)Cl], and [Co^I(tpa)(CO)]⁺ were estimated (Tables S1–S4) to identify the electronic charge distribution of the atoms within each molecule.

3. Materials and Methods

3.1. Reagents

All chemical reagents were of analytical grade and used as received from Aldrich Chemical Co. (St. Louis, MO, USA) and J.T. Baker (Radnor, PA, USA) without any further purification.

3.2. FTIR-IR and UV-Vis Measurements

IR characterization was conducted with an FTIR-IR Affinity-1S Shimadzu spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Electronic spectra were recorded on a Thermo Scientific Evolution Array spectrophotometer (Waltham, MA, USA) within a spectral range of 200–1100 nm.

3.3. Synthesis of Complexes

3.3.1. [Co(bztpen)Br]PF₆

A solution of Bztpen (1.005 g, 2.37 mmol in 10 mL of methanol) was added dropwise to a 10 mL methanolic solution of cobalt (II) bromide hydrate, CoBr₂·nH₂O (0.516 g, 2.37 mmol). The mixture was heated and stirred for 3 h. Subsequently, ammonium hexafluorophosphate (0.0815 g, 0.5 mmol) was added to the reaction mixture. A purple precipitate formed, which was filtered and washed with pure ethanol.

3.3.2. [Co(tpa)Cl]Cl

[Co(tpa)Cl]Cl was synthesized by dissolving 0.01 mmol CoCl₂·6H₂O in acetonitrile within a Schlenk system. After saturating the solution with N2, 0.01 mmol TPA was added, and the reaction mixture was stirred at room temperature for 1 h. The resulting products were washed with ethyl ether to remove any remaining ligand, then filtered and dried under vacuum.

3.3.3. [Co(bpy)₃](BF₄)₂

A 5 mL 0.06 M methanolic solution of 2,2-bpyridine was added dropwise to a 5 mL 0.02 M methanolic solution of [Co(H₂O)₆](BF₄)₂. The reaction mixture was stirred at room temperature for 1 h. Afterward, the solvent was removed by slow evaporation. The obtained product was washed with ethyl ether to remove any remaining ligand, filtered, and dried under vacuum.

3.4. Electrochemical Studies in the Absence and in the Presence of CO₂

The electrochemical experiments were conducted using a Biologic SP-300 potentiostat–galvanostat (Seyssinet-Pariset, France). The compounds were dissolved at a concentration of 1 mM in MeCN, with a supporting electrolyte of 0.1 M Tetrabutylammonium hexafluorophosphate (TBAPF₆). A typical three-electrode setup was employed, with a glassy carbon electrode as the working electrode, a platinum wire as the counter electrode, and a silver wire as the pseudo-reference electrode. All potentials were referenced to the Fc/Fc⁺ redox couple, in accordance with IUPAC standards [41]. Cyclic voltammetry was performed at scan rates of 50, 100, 250, 500, 750, 1000, and 1500 mV s⁻¹. Prior to use, the working electrode was polished with 0.5 µm α-alumina on a Buehler 740 cloth, rinsed with distilled water, and sonicated for approximately 130 s. The solutions were bubbled with N₂ or CO₂ before each experiment. Ohmic drop compensation was performed using positive feedback using Ru values measured by electrochemical impedance spectroscopy.

3.5. UV-Vis Spectroelectrochemical Studies in the Absence and in the Presence of CO₂

For the procedure, 1 mL of different complex solutions (1 mM [Co(tpa)Cl]Cl, 1 mM [CoII(bztpen)Br]PF₆, and 0.2 mM [Co(bpy)₃](BF₄)₂) in MeCN containing 0.1 M TBAPF₆, previously bubbled with N₂, was added into a spectroelectrochemical quartz cell with an optical path length of 0.7 mm. Then an optically transparent Au mesh, used as a working electrode, was placed in the quartz cell. A non-aqueous Ag/Ag⁺ reference electrode (0.1 M AgNO₃ in MeCN) and a Pt wire as the auxiliary electrode were employed. All potential values were referenced to the Fc/Fc⁺ redox couple, in accordance with IUPAC standards [41]. A Thermo Scientific Evolution Array spectrophotometer, which allows us to simultaneously measure absorbance values at a wavelength range from 190 to 1000 nm, coupled to a Bio-Logic SP-50 potentiostat–galvanostat (Seyssinet-Pariset, France), was used. UV-Vis spectra were recorded every 3 s for 3 min during the imposition of different potentials. The procedure was then repeated with the solution under a CO₂ atmosphere. The calibration of the spectrophotometer was carried out by the equipment supplier using a standard holmium oxide.

3.6. DFT Theoretical Studies

Density functional theory (DFT) calculations [42,43,44] were performed using Gaussian 16 [45]. Full geometry optimizations without symmetry constraints, frequency calculations, and UV-Vis spectra simulations were conducted using the CAM-B3LYP [46] density functional and the LanL2DZ [47,48]. The optimized geometries of local minima were confirmed by the absence of imaginary frequencies. The solvation model used was SMD with acetonitrile as solvent.

4. Conclusions

Electrochemical and spectroelectrochemical studies enabled the elucidation of the reaction mechanisms involved in the molecular catalysis of CO₂ complexes [Co^II(bztpen)Br]PF₆, [Co^II(TPA)Cl]Cl, and [Co^II(bpy)₃](BF₄)₂. The results suggest the formation of the chemical species [Co^I(bztpen)(CO)] and [Co^I(tpa)(CO)] during electrolysis in the presence of CO₂, and this was confirmed by DFT calculations. On the other hand, the [Co^II(bpy)₃](BF₄)₂ complex does not appear to form such intermediates. The trigonal bipyramidal [Co^II(tpa)Cl]Cl complex exhibits the lowest catalytic rate constant for CO₂ reduction. In contrast, the octahedral complexes [CoII(bztpen)Br]PF₆ and [Co(bpy)₃](BF₄)₂ exhibit higher catalytic efficiency, highlighting the influence of octahedral geometry on the catalytic behavior. Although their mechanisms differ, the CO₂ reduction mediated by [Co^II(bpy)₃](BF₄)₂ proceeds via an outer-sphere mechanism, whereas [Co^II(bztpen)Br]PF₆ operates through an inner-sphere mechanism. Key information is shown in Scheme 1.