Influence of the Bis‐Carbene Ligand on Manganese Catalysts for CO₂ Electroreduction

Abstract

First row transition metal complexes have attracted attention as abundant and affordable electrocatalysts for CO₂ reduction. Manganese complexes bearing bis‐N‐heterocyclic carbene ligands defining 6‐membered ring metallacycles have proven to reduce CO₂ to CO selectively at very high rates. Herein, we report the synthesis of manganese carbonyl complexes supported by a rigid ortho‐phenylene bridged bis‐N‐heterocyclic carbene ligand (ortho‐phenylene‐bis(N‐methylimidazol‐2‐ylidene), Ph‐bis‐mim), which defines a 7‐membered ring metallacycle. We performed a comparative study with the analogues complexes bearing an ethylene‐bis(N‐methylimidazol‐2‐ylidene) ligand (C₂H₄‐bis‐mim) and a methylene‐bis(N‐methylimidazol‐2‐ylidene) ligand (CH₂‐bis‐mim), and found that catalysts comprising a seven‐membered metallacycle retain similar selectivity and activity as those with six‐membered metallacycles, while reducing the overpotential by 120–190 mV. Our findings reveal general design principles for manganese bis‐N‐heterocyclic carbene electrocatalysts, which can guide further designs of affordable, fast and low overpotential catalysts for CO₂ electroreduction.

We present a comparative study on the effect of extending the size of the metallacycle, defined by bis‐N‐heterocyclic carbene ligands, from six to seven members on manganese carbonyl catalysts for CO₂ electroreduction. The complexes with larger metallacycles retain the high activity and selectivity typically observed in manganese bis‐carbene electrocatalyst with a smaller metallacycle, while reducing the overpotential of the reaction.

Introduction

Electrochemical reduction of carbon dioxide utilizing electricity from renewable sources is a promising strategy to produce sustainable fuels/chemicals from this highly accessible greenhouse gas, which is generated in large amounts due to industrial activity.[ ¹ , ² , ³ ] Product selectivity is one of the main issues of concern due to the complex nature of CO₂ electroreduction. The variety of products accessible at similar thermodynamic driving forces makes that the product selectivity relies mainly on the kinetic control of the process, and this can only be accomplished if well‐selected catalysts are used.

In CO₂ electroreduction, molecular catalysts are prone to mechanistic studies, enabling optimization of their catalytic performance by determining structure‐function relationships.[ ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ ] From the different types of transition metal catalysts used for this process, those based on Mn have gained special attention due to the abundance of this first row transition metal, ^[15] although it needs to be mentioned that some authors have also pointed out some drawbacks regarding the use of manganese, because the lack of efficient recycling methods may cause a risk in the supply chain of this metal. ^[16] Since the pioneering work of Chardon‐Noblat and Deronzier, ^[17] fac‐tris‐carbonyl bromide manganese complexes bearing 2,2’‐bipyridine ligands have been widely used as catalytic models to modify overpotential, selectivity and activity in CO₂ electroreduction.[ ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ ] Later, it was observed that the replacement of one pyridyl arm by a more strongly σ donating N‐heterocyclic carbene (NHC) ligand allowed to improve significantly the performance of these Mn(I) catalysts. In this regard, imidazolylidenes,[ ²⁸ , ²⁹ ] benzymidazolylidenes[ ³⁰ , ³¹ ] and triazolylidenes[ ³² , ³³ ] were utilized to build mixed pyridyl‐NHC ligands, whose related Mn complexes ((NHC−py)−Mn−Br, Figure 1) showed increased longevity or allowed low overpotential pathways, while keeping a high selectivity towards CO generation.

Figure 1.

Mn electrocatalysts for CO₂ reduction featuring NHC‐based ligands.

In 2018 Royo and Lloret‐Fillol reported a Mn complex featuring a methylene‐bridged bis‐NHC ( ^Me CH₂−Mn−Br, Figure 1). ^[34] This catalyst enabled selective CO₂ to CO electroreduction at remarkably high rates, which are among the fastest ever reported for a molecular electrocatalyst. The introduction of a bulky mesityl group as N‐substituent at the bis‐NHC ligand ( ^Mes CH₂−Mn−Br, Figure 1) by Duan, proved to maintain the high catalytic activity and allowed the isolation and characterization of singly and doubly reduced intermediates. ^[35] More recently Franco, Lloret‐Fillol and Luis decoded the role of protic acids in the catalytic process. ^[36] It was found that the mechanism and the selectivity are highly dependent on the concentration of protons in the reaction media. Low proton concentration favors selective CO formation, while high proton concentration favors Mn(I)‐hydride formation, which leads to formate production. Interestingly, Royo and co‐workers reported a Mn electrocatalyst based on a bis‐triazolylidene ligand (i—bitz—Mn−Br, Figure 1) which provided CO selectively, albeit with lower rate and Faradaic efficiency than the bis‐imidazolylidene counterparts. ^[32] Mn(I) catalysts with pincer‐(CNC) ligands were explored recently (CNC−Mn⁺ , Figure 1).[ ³⁷ , ³⁸ , ³⁹ , ⁴⁰ ] These catalysts displayed good activity and selectivity in CO₂ to CO electroreduction, most likely due to the high stability conferred by the pincer bis‐NHC ligand, in contrast with the polypyridine analogue bearing 2,2′:6′,2′′‐terpyridine which proved to be unstable after a few turnovers. ^[41]

Among the catalysts cited above, those featuring a bis‐chelating bis‐NHC ligand with a methylene bridge showed superior activity and efficiency than the NHC‐py, i‐bitz and CNC counterparts. From these previous results, it seems rather reasonable that there is an influence of the nature of the bis‐NHC ligand on the catalytic performance of the electrocatalysts. This influence may be related to the different geometrical constraints provided by the chelating ligand upon binding to the Mn(I) centre. For CH₂−Mn−Br type complexes, the coordination of the bidentate ligand defines a 6‐membered ring metallacyle in a “boat” conformation, with a bite angle close to the ideal 90° for octahedral complexes. This contrasts with the flat conformation of the 5‐membered ring formed by the rest of the ligands shown in Figure 1, which also display a smaller bite angle (typically lower than 80°).[ ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ ] All this strongly suggests that subtle modifications of the nature of bidentate bis‐NHC ligands will very likely give rise to improved catalytic systems.

Built on these grounds, we herein report the preparation of a new manganese complex supported by a rigid ortho‐phenylene‐bis(N‐methylimidazol‐2‐ylidene) (Ph‐bis‐mim), which extends the size of the metallacycle to seven members (Figure 2). Aiming to establish a correlation between the structural features and catalytic performance, we compared the catalytic activity of this compound with that of the related complex [MnBr(CO)₃(C₂H₄‐bis‐mim)][ ⁴⁷ , ⁴⁸ ] bearing a more flexible ethylene linker between the NHC rings ( ^Me C₂H₄−Mn−Br in Figure 2, C₂H₄‐bis‐mim is ethylene‐bis(N‐methylimidazol‐2‐ylidene)). As will be described below, we found that classic parameters in coordination chemistry, such as the bite angle, the flexibility of the bis‐chelating ligand and the size of the metallacycle, strongly influence the stability of the complexes and their performance in CO₂ electroreduction.

Figure 2.

Mn complexes focus of this work.

Results and Discussion

Complexes of the type [MnBr(bis‐NHC)(CO)₃], in which the bis‐NHC contains an ethylene linker, have recently shown to be effective catalysts for the hydrogenation of unsaturated substrates and the N‐formylation/N‐methylation of amines.[ ⁴⁷ , ⁴⁸ ] However, their performance in CO₂ electroreduction remains unexplored. We decided to utilize ^Me C₂H₄−Mn−Br for our studies due to its structural similarity to ^Me CH₂−Mn−Br. We also synthesized a new complex bearing a bis‐NHC with an ortho‐phenylene bridging group following the procedure shown in Scheme 1. Treatment of [Ph‐bis‐mim‐H₂][Br]₂ with Ag₂O formed the corresponding silver‐carbene complex, which was transmetallated in situ by addition of MnBr(CO)₅, affording ^Me Ph−Mn−Br in 73 % yield (Scheme 1). ^Me Ph−Mn−Br is stable in the presence of air and moisture for long periods of time in the solid state and in solution, while ^Me C₂H₄−Mn−Br slowly decomposed even in the solid state, in agreement with previously reported observations. ^[47] The ¹H NMR spectrum of the reaction product showed a mixture of two species with identical signal pattern in a 8.3 : 1 molar ratio (Figure S6 in the SI), which could not be separated by ordinary purification techniques. We hypothesize that these two species are originated due to the rigidity imposed by the Ph‐bis‐mim ligand. The orientation of the phenyl ring relative to the Br−Mn−CO axis generates two configurational isomers. As can be observed in Scheme 1, in one of the isomers the bromide ligand is oriented close to the ortho‐phenylene bridge of the bis‐NHC ligand (syn‐ ^Me Ph−Mn−Br), while the bromide ligand is pointing to the opposite direction than the aromatic bridging group in anti‐ ^Me Ph−Mn−Br.

Scheme 1.

Synthesis of ^Me Ph−Mn−Br.

To corroborate this hypothesis, density functional theory computational studies were undertaken (see the SI for details). anti‐ ^Me Ph−Mn−Br was computed to be more stable than syn‐ ^Me Ph−Mn−Br by 1.3 kcal/mol. Heating a C₆D₆ solution of ^Me Ph−Mn−Br at different temperatures (Figure S15 in the SI) did not change the relative ratio of the isomers, indicating, that most likely these two species are not in equilibrium. The less stable nature of syn‐ ^Me Ph−Mn−Br can be justified due to the larger steric hindrance produced between the rigid ortho‐phenylene bridge and the bulky bromide ligand. This situation is minimized in anti‐ ^Me Ph−Mn−Br, which locates the bromide ligand opposite to the bridging group of the bis‐NHC ligand.

A similar behaviour was reported for ^Me CH₂−Mn−Br and ^Me C₂H₄−Mn−Br. ^[48] In the case of ^Me CH₂−Mn−Br the syn isomer is more stable than the anti isomer, thus suggesting that the nature of the bridging group has an influence on the relative stability of the geometrical isomers for this family of complexes. A more complex situation was observed for ^Me C₂H₄−Mn−Br due to the flexible nature of C₂H₄‐bis‐mim, which can give rise to chiral manganese complexes.

The infrared spectrum of a dichloromethane solution of ^Me Ph−Mn−Br shows three intense C−O stretching bands (Table 1, Figure S19 in the SI), which appear at quasi‐identical frequencies as those shown for ^Me CH₂−Mn−Br (2009, 1926 and 1886 cm⁻¹, Figure S16 in the SI), thus suggesting that the two bis‐NHC ligands exert similar electrondonating strength. These ν(CO) values contrast with those shown by the Mn(I) complex with the ethylene‐connected bis‐NHC, ^Me C₂H₄−Mn−Br (2005, 1919 and 1881 cm⁻¹, Figure S17 in the SI), which turns out to be the most electrondonating ligand manganese complex of the family.

Table 1.

Characterization data of anti‐ CH₂−Mn−Br ^[34] anti‐ C₂H₄−Mn−Br and anti‐ Ph−Mn−Br.

Parameter	anti‐ Me CH2−Mn−Br	anti‐ Me C2H4−Mn−Br	anti‐ Me Ph−Mn−Br
Br−Mn−COax (°)	178.96(7)	172.8(2)	169.0(2)
Bite angle (°)	85.11(8)	92.51(17)	86.3(2)
α angle (°)[a]	32.8	37.8	51.9
ν CO (cm−1)[b]	2009 1926 1886	2005 1919 1881	2008 1925 1888

[a] Defined as the average angle between the plane containing the azole ring and the equatorial plane of the complex (which contains the two NHC ligands). [b] IR spectra measured in dichloromethane as solvent.

The molecular structure of anti‐ ^Me Ph−Mn−Br was unambiguously determined by means of single crystal X‐ray diffraction studies (Figure 3a). The complex contains three carbonyl ligands in a fac disposition, with Ph‐bis‐mim and a bromide ligand completing the pseudo‐octahedral coordination sphere about the manganese centre. The bromide ligand is occupying the axial position of the complex, and it is oriented opposite to the phenylene bridge. The coordination of the bis‐NHC ligand to the metal establishes a seven‐membered ring metallacycle with a chair conformation. The angle between the equatorial plane of the complex (containing the two carbene donors) and the phenylene ring (“chair angle”)[ ⁴⁹ , ⁵⁰ ] is 96.7°.

Figure 3.

Structural representation of anti‐ ^Me Ph−Mn−Br (a) and anti‐ ^Me C₂H₄−Mn−Br (b) with ellipsoids drawn at the 50 % level. Hydrogen atoms, dichloromethane and syn‐ ^Me C₂H₄−Mn−Br are omitted for clarity.

Slow diffusion of hexane into a saturated solution of ^Me C₂H₄−Mn−Br in dichloromethane allowed the formation of single crystals suitable for X‐ray diffraction studies. Two conformers of ^Me C₂H₄−Mn−Br were observed in the asymmetric unit cell differing in the orientation of the bromide ligand with respect to the ethylene bridge: anti‐ ^Me C₂H₄−Mn−Br (Figure 3b) and syn‐ ^Me C₂H₄−Mn−Br (Figure S57 in the SI, reported previously). ^[48] The presence of anti‐ ^Me C₂H₄−Mn−Br allows to establish comparisons between the metrical parameters of anti‐ ^Me Ph−Mn−Br and ^Me CH₂−Mn−Br (Table 1), for which the X‐ray structure of both geometrical isomers was previously reported.[ ³⁴ , ³⁶ ]

The Br−Mn−CO_ax angle of anti‐ ^Me CH₂−Mn−Br (178.96(7)°, Table 1) is close to the ideal 180° for the octahedral geometry. This angle is 6 degrees smaller for anti‐ ^Me C₂H₄−Mn−Br (172.8(2)°) and 10 degrees smaller for anti‐ ^Me Ph−Mn−Br (169.0(2)°), as expected for the higher steric congestion produced by the ortho‐phenylene linker in Ph‐bis‐mim, compared to the ethylene and methylene groups in C₂H₄‐bis‐mim and CH₂‐bis‐mim, respectively. Interestingly, Ph‐bis‐mim and CH₂‐bis‐mim display similar bite angles in anti‐ ^Me CH₂−Mn−Br (85.11(8)°) and anti‐ ^Me Ph−Mn−Br (86.3(2)°), both being smaller than the bite angle exhibited by C₂H₄‐bis‐mim in anti‐ ^Me C₂H₄−Mn−Br (92.51(17)°). The average angle between the plane containing the azole ring and the equatorial plane of the complex, the α angle, ^[51] differs by almost 20° between anti‐ ^Me CH₂−Mn−Br (32.8°) and anti‐ ^Me Ph−Mn−Br (51.9°), as a consequence of the different size of the metallacycle of both complexes (Table 1). The six‐membered metallacycle displayed by CH₂‐bis‐mim forces the two NHC donors to lie in a more coplanar situation with respect to the equatorial plane of anti‐ ^Me CH₂−Mn−Br, while a larger α angle is needed to accommodate the rigid seven‐membered ring metallacycle defined by Ph‐bis‐mim in anti‐ ^Me Ph−Mn−Br. C₂H₄‐bis‐mim displays an intermediate value of the α angle in anti‐ ^Me C₂H₄−Mn−Br (37.8°) with respect to anti‐ ^Me CH₂−Mn−Br and anti‐ ^Me Ph−Mn−Br, which can be explained by its more flexible nature.

Manganese electrocatalysts typically promote selective CO₂ electroreduction to CO. Cationic tetracarbonyl species of the type [Mn(bis‐NHC)(CO)₄]⁺ are expected to be involved in the last step of the mechanism, enabling CO loss from the coordination sphere of the catalyst. Therefore, such important intermediates were targeted next.

Treatment of ^Me C₂H₄−Mn−Br or ^Me Ph−Mn−Br in dichloromethane under CO atmosphere, in the presence of NaPF₆, allowed the isolation of the corresponding tetracarbonyl complexes, ^Me C₂H₄−Mn−CO⁺ and ^Me Ph−Mn−CO⁺ , in 20 % and 46 %, respectively (Scheme 2). The pattern of signals displayed in the ¹H NMR spectra of both complexes is consistent with C_S symmetric structures (Figure S10 and S13 in the SI).

Scheme 2.

Synthesis of ^Me C₂H₄−Mn−CO⁺ and ^Me Ph−Mn−CO⁺ .

Cyclic voltammetry experiments were performed with the family of manganese complexes in acetonitrile in order to evaluate their electrochemical properties. Cyclic voltammograms (CVs) of ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br under Ar atmosphere revealed a single reduction wave for both complexes (Figure 4), attributed to a two electron process coupled with bromide dissociation, as indicated by the cathodic shift observed on the reduction peak upon increasing the scan rate of the experiment (Figure S25 and S26).[ ³⁴ , ³⁵ ] We observed that ^Me C₂H₄−Mn−Br displays a reduction feature at 40 mV less negative potential than ^Me Ph−Mn−Br (E_pc( ^Me C₂H₄−Mn−Br)=−2.24 V, E_pc( ^Me Ph−Mn−Br)=−2.28 V, at 100 mV/s, all the potentials reported in this work are referenced vs the Fc⁺/Fc couple), suggesting that the manganese centre in ^Me C₂H₄−Mn−Br is less electronrich than in ^Me Ph−Mn−Br. This observation contrasts with the IR spectroscopy data acquired in dichloromethane, which seems to point to the inverse trend (vide supra). We hypothesized that when acetonitrile is utilized as solvent, bromide substitution by acetonitrile occurs, establishing an equilibrium with the corresponding cationic acetonitrile species and this modifies the observed reduction potential. To investigate this hypothesis, the IR spectra was acquired in acetonitrile, revealing the presence of three strong CO stretches (ν _co=2020, 1928 and 1917 cm⁻¹ for ^Me C₂H₄−Mn−Br, and ν _co=2025, 1937 and 1928 cm⁻¹ for ^Me Ph−Mn−Br, Figure S18 and S20 in the SI, respectively). These bands appear at significantly higher frequencies than those observed for the IR spectra of the same complexes in methylene chloride (Table 1), thus suggesting that the bromide ligand may dissociate in acetonitrile solution yielding a cationic [Mn(bis‐NHC)(CO)₃(S)]⁺ species (with S=acetonitrile). The observed frequencies lie at similar values as those reported for ^Me CH₂−Mn−MeCN⁺ (ν _co=2025, 1935 and 1925 cm⁻¹). ^[36] The lability of the bromide ligand in (bis‐NHC)Mn(I) tri‐carbonyl complexes in acetonitrile solutions has been reported previously.[ ³³ , ³⁴ , ³⁵ ] These results indicate that the assignment of the electronrichness of this type of complexes based on cyclic voltammetry must be performed carefully, as bromide substitution by acetonitrile is a facile process which can shift the reduction potential.

Figure 4.

CV of ^Me C₂H₄−Mn−Br (blue) and ^Me Ph−Mn−Br (red) under Ar atmosphere. Conditions: Ar atmosphere, MeCN, [Mn]=1 mM, [TBAPF₆]=100 mM, 3 mm glassy carbon disc working electrode, Pt wire counter electrode, Ag wire pseudo‐reference electrode, 100 mV/s.

Interestingly, two oxidation events were observed on the return sweep for both complexes (Figure 4). In this case, the potential of the oxidation features matched the donor ability of the ligands (−2.17 V and −1.4 V for ^Me C₂H₄−Mn−Br and −2.10 V and −1.37 V for ^Me Ph−Mn−Br). This behavior was previously observed for the related methylene bridged bis‐NHC manganese complexes ( ^Me CH₂−Mn−Br ^[34] and ^Mes CH₂−Mn−Br, ^[35] Figure 1). The first feature was attributed to a one electron oxidation of the doubly reduced species, yielding a neutral intermediate with a radical character. This species forms a dimer that is oxidized by one electron around −1.4 V. Supportive of similar processes occurring on ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br is the different current of the oxidation features. The oxidation events around −2.15 V have almost identical current intensity for both types of complexes, suggesting a similar one electron oxidation process (Figure 4). However, the oxidation current around −1.4 V is more intense for the complex based on C₂H₄‐bis‐mim, suggesting that the formation of the dimer is more favored for ^Me C₂H₄−Mn⁰ than for ^Me Ph−Mn⁰ . This can be justified by the more flexible nature of C₂H₄‐bis‐mim, which enables conformation adaptability, thus facilitating dimer formation. Moreover, the steric bulk introduced by the rigid ortho‐phenylene bridge of Ph‐bis‐mim, and the larger α‐angle observed in ^Me Ph−Mn−Br (Table 1), produce a higher steric congestion around the apical positions of ^Me Ph−Mn⁰ , limiting the generation of dimeric species, and thereby reducing the current intensity of the second oxidation event.

When the cyclic voltammetry experiments with ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br were performed with the addition of 5 % H₂O under Ar atmosphere, an anodic shift on the peak potential was observed (Figure S28 and S29 in the SI). This behavior is attributed to the protonation of the low valent Mn complexes, with the concomitant generation of the corresponding metal‐hydride.

When performed under CO₂ atmosphere, the CVs of ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br showed a large current enhancement, strongly suggestive of an electrocatalytic process occurring at the surface of the electrode (Figure 5). This observation is in agreement with the behavior previously reported for bis‐NHC or CNC based Mn electrocatalysts, which is not observed on related pyridyl‐based Mn electrocatalyst, unless proton sources are added. When the CV was acquired under CO₂ atmosphere and with addition of 1 % H₂O, the catalytic current of ^Me Ph−Mn−Br and ^Me C₂H₄−Mn−Br increased (Figure 5). Interestingly, addition of larger amounts of H₂O produced a decrease on the intensity of catalytic wave of both complexes (Figures S31 and S32 in the SI). Similar behavior was observed previously for ^Me C₂H₄−Mn−Br ^[34] and it was rationalized by the formation of a metal‐hydride intermediate at high water concentrations, which in the presence of CO₂ generates formate. The Mn‐formato complex needs large negative potentials to release formate, thereby limiting the catalytic current at mild potentials. Considering the similar CV behavior at high H₂O concentrations (Figures S31 and S32 in the SI), we propose that ^Me Ph−Mn−Br and ^Me C₂H₄−Mn−Br follow the same reactivity than ^Me CH₂−Mn−Br. ^[36]

Figure 5.

CV of ^Me C₂H₄−Mn−Br (a) and ^Me Ph−Mn−Br (b) under Ar (red), CO₂ (orange), CO₂ with 1 % added H₂O (black) and Ar with 5 % added H₂O (blue). Conditions: Ar or CO₂ atmosphere, MeCN or MeCN+1 % H₂O or MeCN+5 % H₂O, [Mn]=1 mM, [TBAPF₆]=100 mM, 3 mm glassy carbon disc working electrode, Pt wire counter electrode, Ag wire pseudo‐reference electrode, 100 mV/s.

Controlled potential electrolysis experiments (CPE) at an applied potential of −2.65 V were undertaken to establish the selectivity and stability of the catalysts. In the absence and in the presence of H₂O (1 %) ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br selectively yielded CO as the only product with Faradaic efficiencies higher than 92 % (Table S1 in the SI). In agreement with previous reports on bis‐NHC manganese electrocatalysts, addition of H₂O produces a decay of charge during electrolysis, indicating that the long‐term stability of the catalysts is influenced by the presence of protons (Figures S36 and S38 in the SI). In the absence of H₂O, the current remains relatively stable for more than two hours (Figure S35 and S37 in the SI).

Once the product of the electrocatalytic reactions was established, we decided to explore fast‐scan regimes on cyclic voltammetry to determine the rate constant for catalysis and the overpotential. “S‐shaped” catalytic waves were observed for ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br, from 30 V/s under CO₂ in the absence of any acid, and from 65 V/s under CO₂ in the presence of 1 % added H₂O (Figures S39, S43, S47 and S51 in the SI). This type of catalytic response is indicative of pure kinetic conditions (KS kinetic zone), in which no substrate consumption occurs at the surface of the electrode.[ ⁵² , ⁵³ ] Therefore, the catalytic current plateaus and becomes independent of the scan rate of the experiment. Under these conditions, the rate constant for the rate‐determining chemical step of the reaction (k _obs) can be determined utilizing Eq. 1, which relates the ratio of catalytic current under pure kinetic conditions (i _c) and the current in the absence of catalysis (i _p) to k _obs, the number of electrons in the CO₂ to CO reduction (n _c=2), the number of electrons involved in the reduction process in the absence of substrate (n _p=2) and the scan rate of the experiment in the absence of catalysis (υ).[ ⁵² , ⁵⁴ ]

$${\displaystyle \frac{i_c}{i_p}=2.24\frac{n_c}{n_p}\sqrt{\frac{RT}{n_{p}F}}\sqrt{\frac{1}{\upsilon }}\sqrt{k_{obs}}}$$ (1)

Table 2 shows the k _obs values of the bromide complexes utilized in our study, together with previously reported values for ^Me CH₂−Mn−Br, which are displayed for comparison. ^[34]

Table 2.

Catalytic metrics of Mn CH₂‐bis‐mim, ^[34] C₂H₄‐bis‐mim and Ph‐bis‐mim complexes.

	k obs(CO2) (s−1)[a]	η(CO2) (V)	k obs(CO2+1 % H2O) (s−1)[a]	η(CO2+1 % H2O) (V)
Me CH2−Mn−Br	2.1 ⋅ 103	1.35	3.2 ⋅ 105	1.15
Me C2H4−Mn−Br	(5.07±2.29) ⋅ 104	1.20	(1.39±0.25) ⋅ 105	0.96
Me Ph−Mn−Br	(4.21±1.26) ⋅ 104	1.23	(2.41±0.92) ⋅ 105	0.97

[a] Values obtained from the average of at least two independent experiments. Uncertainty established from the standard deviation of the experiments.

The complexes ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br enable catalysis at rates that compare well with ^Me CH₂−Mn−Br, but their performances are highly dependent on the conditions used. The observed kinetic constants for ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br are about one order of magnitude larger than that for ^Me CH₂−Mn−Br under CO₂ in the absence of added protons, while in the presence of 1 % H₂O they are slightly smaller. These results indicate that both, ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br are among the fastest catalysts for CO₂ to CO electroreduction. It is worth mentioning that the k _obs values found for the tetra‐carbonyl complexes, ^Me C₂H₄−Mn−CO⁺ and ^Me Ph−Mn−CO⁺ , were quasi‐identical to the brominated pre‐catalysts (see Table S1 of the SI), ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br, which strongly suggests that such complexes are intermediates in the catalytic cycle.

From the S‐shaped CVs, E_cat/2 values can be derived, ^[55] enabling to determine the overpotential of the catalysts (η) required to reach half of the maximum rate constant, as the difference between the standard potential for CO₂ reduction (E⁰ _CO2/CO, see Section 4 of the SI for details) and E_cat/2. All catalysts show high overpotentials due to the strong donor ability of the bis‐NHC ligands (Table 2). Interestingly, ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br display lower overpotential than ^Me CH₂−Mn−Br. The difference is more pronounced if the experiments are performed under wet conditions than in the absence of protons. These results indicate that subtle differences in the first coordination sphere of the manganese bis‐NHC catalysts influence and reduce its overpotential in CO₂ electroreduction.

The catalytic Tafel plot for the three manganese bis‐NHC catalysts can be constructed (Figure 6) considering the data displayed in Table 2 (see Section 4 of the SI for details) allowing a visual comparison of the relationship between TOF and η under different conditions.

Figure 6.

Catalytic Tafel plot of ^Me C₂H₄−Mn−Br, ^Me Ph−Mn−Br and ^Me CH₂−Mn−Br under dry conditions (a) and in the presence of 1 % H₂O (b).

Catalysts ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br need an overpotential of about 950 mV to reach a rate constant of 1 s⁻¹ under dry conditions (Figure 6a). The overpotential needed for ^Me CH₂−Mn−Br to reach this rate is 200 mV higher. A similar difference can be observed in the presence of 1 % H₂O (Figure 6b).

Under a mechanistic point of view, we propose, based on previous reports, a CO₂ reductive disproportionation to provide CO and carbonate under water‐free conditions (Scheme 3).[ ³⁴ , ³⁵ ] The bromide precatalysts, which are in equilibrium with the solvato complex, undergo a two‐electron reduction, giving rise to a negatively charged five‐coordinate intermediate, which generates a metallocarboxylate species upon CO₂ binding. The highly nucleophilic nature of this intermediate enables a second CO₂ binding, releasing carbonate and generating a tetracarbonyl intermediate, which can start new catalytic cycle by means of a two‐electron reduction and CO loss.

Scheme 3.

Proposed mechanism for CO₂ electroreduction in the absence of protons.

In the presence of H₂O, the catalysts are proposed to follow a “protonation first” pathway.[ ³⁶ , ⁵⁶ , ⁵⁷ ] Key for this mechanism to be operative is the strong donor ability of the bis‐NHC ligands, which renders a metallocarboxylic acid (after metallocarboxylate protonation) nucleophilic enough to undergo protonation without the need of further reduction events. Both mechanisms have as important intermediates tetracarbonyl compounds, being the species generated before electrochemically triggered CO loss. These intermediates have been isolated in this work and proved to have identical catalytic metrics than the bromide precatalysts.

Considering both mechanisms and the structure of the manganese catalysts, the doubly reduced intermediates bearing Ph‐bis‐mim or C₂H₄‐bis‐mim ( ^Me C₂H₄−Mn⁻ and ^Me Ph−MN⁻ ) can exist as two different isomers, one with the coordination vacancy close to the linker of the bis‐NHC ligand, and the other with the coordination site pointing to the opposite direction (syn and anti isomers, similar to the bromide complexes). Therefore, the steric hindrance of the bridging groups of Ph‐bis‐mim and C₂H₄‐bis‐mim create two different isomers, whose binding properties to CO₂ may differ considerably. The same is true for the different intermediates involved in the catalytic cycle. This situation is less pronounced in CH₂‐bis‐mim derived species, for which the difference in properties and reactivity of syn and anti isomers is expected to be less important due to the lower steric bulk of the methylene bridge. We hypothesize that the lower overpotential observed arises from this structural divergence between isomers of catalytic intermediates, which can give rise to lower energy transition states. Further studies are necessary to shed light into this point, and they are the focus of current efforts in our laboratory.

Conclusions

In this work we have studied the effect of increasing the size of the metallacycle defined by bis‐NHC ligands, from six to seven members, on manganese electrocatalysts for CO₂ reduction. The rigid ortho‐phenylene bridge enabled complexes with similar electronrichness and bite angle than the previously reported ^Me CH₂−Mn−Br, despite the different structural features of Ph‐bis‐mim and CH₂‐bis‐mim. In contrast, the flexible ethylene linker of C₂H₄‐bis‐mim provided more electronrich complexes, which proved to be less stable under aerobic conditions. The complexes bearing bis‐carbene ligands defining a seven‐membered ring metallacycle ( ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br) displayed high selectivity for CO generation and high catalytic activities, comparable to ^Me CH₂−Mn−Br in the presence of 1 % H₂O and one order of magnitude faster in the absence of protons. Although all the manganese catalysts work at large overpotentials due to the strong donor ability of bis‐NHC ligands, ^Me C₂H₄−Mn−Br and ^Me Ph−Mn−Br allowed catalysis at less negative potentials (by 120–190 mV). We hypothesize that this can be due to the asymmetry imposed by the bridging groups of Ph‐bis‐mim or C₂H₄‐bis‐mim on the axial axes of the complexes, which enables syn and anti isomers, whose properties and reactivity during the catalytic cycle may be significantly different from those derived from CH₂‐bis‐mim, due to the higher steric hindrance imposed by ethylene and ortho‐phenylene linkers when compared to the methylene bridge.

We believe that the results reported in this work will contribute to the design and development of highly active and selective electrocatalysts which work at mild overpotentials. We envision that the lower overpotential, enabled by the electrocatalysts containing a larger bridge between the NHC fragments, can be pushed even further by introducing specific functional groups as linkers, which can lead to secondary coordination sphere effects on the performance of the catalysts.

Experimental Section

General Considerations

All operations were carried out by using standard Schlenk techniques under nitrogen atmosphere unless otherwise stated. ^[58] MeCN and CH₂Cl₂ were dried and degassed with a solvent purification system (SPS M BRAUN).

CD₂Cl₂, C₆D₆ and DMSO‐d₆ were purchased from Cambridge Isotope Laboratories Inc. and used as received. ^Me CH₂−Mn−Br, ^{[47] Me} C₂H₄−Mn−Br ^[47] and [Ph‐bis‐mim‐H₂][PF₆]₂ ^[59] were synthesized according to reported procedures.

All other materials were commercially available and used as received, unless otherwise noted. NMR spectra were recorded on a Bruker 300 MHz spectrometer at 25 °C. Variable Temperature NMR (VT NMR) experiments were recorded on a Varian 500 MHz spectrometer. Chemical shifts are reported with respect to residual organic solvents. ^[60]

Infrared spectra (FTIR) were recorded on a Bruker Equinox 55 spectrometer with a spectral window of 4000–400 cm⁻¹. Electrospray mass spectra (ESI‐MS) were recorded on a Micromass Quatro LC instrument; nitrogen was employed as drying and nebulizing gas. Exact mass analyses were recorded by using a Q‐TOF Premier mass spectrometer with an electrospray source (Waters, Manchester, UK) operating at a resolution of about 16000FWHM.

X‐Ray Crystal Data

Deposition Numbers 2352256 (for ^Me C₂H₄−Mn−Br) and 2352253 (for anti‐ ^Me Ph−Mn−Br) contain the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe (www.ccdc.cam.ac.uk/structures) access Structures service.

Synthesis and Characterization

[Ph‐bis‐mim‐H₂][Br]₂ . This compound was synthesized following a modified reported procedure. ^[61] [Ph‐bis‐mim‐H₂][PF₆]₂ (2.00 g, 3.77 mmol) and tetrabutylammonium bromide (3.50 g, 10.86 mmol) were suspended in acetone (150 mL) under aerobic conditions. After stirring at room temperature for 16 h the suspension was filtrated and the solid washed with acetone (3×25 mL). The colourless product was dried in vacuo and the absence of starting material was confirmed by ¹⁹F NMR spectroscopy (1.49 g, 98 % yield). ¹H NMR (300 MHz, DMSO‐d₆): δ 9.53 (s, 2 H, N−CH−N), 7.93 (m, 4 H, CH _Ph), 7.87 (t, 2 H, ³ J _H‐H=1.8 Hz, CH _Im), 7.77 (t, 2 H, ³ J _H‐H=1.8 Hz, CH _Im), 3.93 (s, 6 H, CH ₃) ppm. ¹³C {¹H} NMR (75 MHz, DMSO‐d₆): δ 138.6 (N−CH−N), 132.1 (CH_Ph), 129.7 (C _q), 128.2 (CH_Ph), 124.3 (CH_Im), 122.9 (CH_Im), 36.3 (CH₃) ppm. HRMS (ESI, m/z): 293.1301 [M−2Br−H]⁺ (calculated: 293.1297).

^Me Ph−Mn−Br. [Ph‐bis‐mim‐H₂][Br]₂ (435 mg, 1.09 mmol) and Ag₂O (306 mg, 1.32 mmol) were suspended in CH₂Cl₂ (40 mL) under exclusion of light. After stirring at room temperature for 24 h, MnBr(CO)₅ (300 mg, 1.09 mmol) was added, and the mixture was heated at 50 °C for 16 h. At room temperature, the yellow‐brown suspension was filtrated over a pad of celite under aerobic conditions, washed with 3×20 mL CH₂Cl₂, and the solvent was evaporated under reduced pressure. The yellow solid was dissolved in 20 mL of CH₂Cl₂ and filtrated over a 0.2 μm Nylon‐filter. Addition of 150 mL of pentane enabled the precipitation of the desired product as a yellow powder, which was collected by filtration, washed with 3×20 mL pentane and dried in vacuo (362 mg, 73 % yield). Suitable single crystals for X‐ray diffraction experiments were obtained by slow diffusion of pentane into a saturated solution of the product in CH₂Cl₂ at −30 °C. Two geometrical isomers are observed by NMR‐spectroscopy in a 8.3 : 1 ratio (anti : syn). ¹H NMR (300 MHz, C₆D₆), anti‐ ^Me Ph−Mn−Br: δ 6.87–6.81 (m, 2 H, CH _Ph), 6.63–6.56 (m, 2 H, CH _Ph), 6.22 (d, 2 H, ³ J _H–H=2.1 Hz, CH _Im), 5.89 (d, 2 H, ³ J _H‐H=2.0 Hz, CH _Im), 4.15 (s, 6 H, CH ₃) ppm. syn‐ ^Me Ph−Mn−Br: δ 6.99 (m, 2 H, CH _Ph), 6.63–6.56 (m, 2 H, CH _Ph), 6.30 (d, 2 H, ³ J _H‐H=2.0 Hz, CH _Im), 5.80 (d, 2 H, ³ J _H‐H=2.0 Hz, CH _Im), 3.48 (s, 6 H, CH ₃) ppm. ¹³C {¹H} NMR (75 MHz, C₆D₆), anti‐ ^Me Ph−Mn−Br: δ 199.2 (C _Carbene), 136.1 (C _q), 128.7 (CH_Ph), 128.1 (CH_Ph), 125.1 (CH_Im), 121.8 (CH_Im), 41.7 (CH₃) ppm. The signals corresponding to syn‐ ^Me Ph−Mn−Br as well as C _CO were not observed. HRMS (ESI, m/z): 293.0601 [M−Br−3CO]⁺ (calculated: 293.0599).

^Me Ph−Mn−CO⁺ . A Teflon‐sealed side arm flask was charged with ^Me Ph−Mn−Br (40 mg, 87.5 μmol), NaPF₆ (22 mg, 131.0 μmol) and CH₂Cl₂ (15 mL). CO was bubbled through the suspension for 15 minutes and the flask was sealed under a CO atmosphere. The reaction mixture was stirred 16 h at room temperature and the suspension was filtrated under aerobic conditions. The solvent was evaporated under reduced pressure. The residue was redissolved in 5 mL CH₂Cl₂ and 15 mL pentane were added, enabling precipitation of the desired product as a white powder which was collected by filtration, washed with 15 mL pentane and dried in vacuo (22 mg, 46 % yield). ¹H NMR (300 MHz, CD₂Cl₂): δ 7.75–7.70 (m, 2 H, CH _Ph), 7.56–7.51 (m, 2 H, CH _Ph), 7.46 (d, 2 H, ³ J _H‐H=2.0 Hz, CH _Im), 7.33 (d, 2 H, ³ J _H‐H=2.1 Hz, CH _Im), 4.02 (s, 6 H, CH ₃) ppm. ¹³C {¹H} NMR (75 MHz, CD₂Cl₂): δ 135.2 (C _q), 131.4 (CH_Ph), 128.3 (CH_Ph), 127.3 (CH_Im), 125.2 (CH_Im), 39.9 (CH₃) ppm. The signals corresponding to C _CO and C _Carbene were not observed. HRMS (ESI, m/z): 293.0603 [M−4CO]⁺ (calculated: 293.0599).

Synthesis of ^Me C₂H₄−Mn−CO⁺ . A Teflon‐sealed side arm flask was charged with ^Me C₂H₄−Mn−Br (40 mg, 87.5 μmol), NaPF₆ (25 mg, 148.9 μmol) and 15 mL CH₂Cl₂. CO was bubbled through the suspension for 15 minutes and the flask was sealed under a CO atmosphere. The reaction mixture was heated to 40 °C for 48 h and the resulting suspension was filtrated under aerobic conditions after cooling to room temperature. Addition of 25 mL of pentane enabled the precipitation of the desired product as a white solid, which was collected by filtration, washed with 10 mL pentane and dried in vacuo (10 mg, 20 % yield). ¹H NMR (300 MHz, CD₂Cl₂): δ 7.24 (bs, 2 H, CH _Im), 7.18 (bs, 2 H, CH _Im), 4.69 (bs, 4 H, CH ₂), 3.97 (s, 6 H, CH ₃) ppm. ¹³C {¹H} NMR (75 MHz, CD₂Cl₂): δ 126.5 (CH_Im), 126.2 (CH_Im), 50.2 (CH₂), 40.3 (CH ₃) ppm. The signals corresponding to C _CO and C _Carbene were not observed. HRMS (ESI, m/z): 245.0605 [M−4CO]⁺ (calculated: 245.0599).

Computational Methods

The DFT calculations have been performed with the Gaussian09 software package ^[62] using the B3LYP density functional. ^[63] Geometry optimizations and subsequent frequency calculations have been performed at the B3LYP/6‐311+g** level of theory. The effect of the solvent (acetonitrile) and London interactions are considered through the SMD model ^[64] and Grimme‐D3 dispersion correction, ^[65] respectively. The conversion factor used from Hartrees to kcal/mol was 627.5.

Supporting Information

NMR spectra, electrochemical methods, crystallographic methods, and computational data can be found in the Supporting Information. The authors have cited additional references within the Supporting Information.[ ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ ]

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Richter M. L., Peris E., Gonell S., ChemSusChem 2024, 17, e202401007. 10.1002/cssc.202401007

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.