Electrochemical Characterization and CO2 Reduction Reaction of a Family of Pyridazine-Bridged Dinuclear Mn(I) Carbonyl Complexes

Three recently synthesized neutral dinuclear carbonyl manganese complexes with the pyridazine bridging ligand, of general formula [Mn2(μ-ER)2(CO)6(μ-pydz)] (pydz = pyridazine; E = O or S; R = methyl or phenyl), have been investigated by cyclic voltammetry in dimethylformamide and acetonitrile both under an inert argon atmosphere and in the presence of carbon dioxide. This family of Mn(I) compounds behaves interestingly at negative potentials in the presence of CO2. Based on this behavior, which is herein discussed, a rather efficient catalytic mechanism for the CO2 reduction reaction toward the generation of CO has been hypothesized.


Introduction
Owing to the fact that the applied potential to reduce CO 2 into its radical anion CO 2 •− has a rather negative value (E • (CO 2 /CO 2 •− )= −1.97 V versus SHE (standard hydrogen electrode)) in DMF [1], the catalysts used for the electrochemical CO 2 reduction reaction (CO 2 RR) have been widely researched amongst molecular metal-based systems, as transition metals generally have multiple accessible redox states that facilitate multielectron chemistry. The vast majority of these catalysts employ a single metal center, [2][3][4][5][6][7][8] usually rhenium or manganese. Albeit structurally similar, Re-and Mn-based catalysts show different electrochemical behavior. While rhenium complexes have been studied for longer and are historically thought to be the most efficient in CO 2 RR, Chardon-Noblat, Deronzier, and co-workers [9] evidenced that Mn complexes (where the Mn center is typically in the Mn(I) oxidation state) can act as good CO 2 reduction catalysts as well. Manganese-based complexes have become the focus of intense investigation in recent years as pre-catalysts for CO 2 reduction, in part due to their high product selectivity for CO formation, but also because they are based on a cheaper and more earth-abundant metal, compared to their thoroughly investigated Re-based counterparts.
Mechanistic studies of rhenium-type catalysts have provided evidence for a concentration-dependent formation of dinuclear intermediates during catalysis [10][11][12][13][14][15]. However, only few examples have been related to catalysts with a rigid bridging ligand that links two rhenium or two manganese active sites in close proximity with a predictable intermetallic distance and orientation [16][17][18] in order to clearly highlight the contribution of the dinuclear pathway [19,20]. Indeed, cofacial dinuclear Re complexes, with a rigid backbone structure that prevents Re-Re bonding (leading to the deactivation of carbon dioxide reduction catalysis), showed a beneficial interaction between the two reaction sites, clearly outperforming non-cofacial or mononuclear complexes at a double concentration. In this framework, a new family of dinuclear tricarbonyl rhenium (I) complexes containing the 1,2-diazine bridging ligand and halide anions as ancillary ligands, has been reported, which is able to catalyze CO 2 reduction with higher efficiency than that reported for the benchmark mononuclear complex Re(CO) 3 Cl-(bpy), confirming that the use of dinuclear complexes could be very promising in this field [21,22]. Although efficiencies are lower than those reported by Jurss and co-workers for similar dinuclear Re complexes (TOF = 15 s −1 for our complex vs. 35 s −1 for the complex reported by Jurss) [19] our work highlights a crucial role of the catalyst structure, in terms of the distance of metal centers and capability to facilitate CO 2 insertion.
On the basis of these results, we recently synthesized a new similar family of dinuclear manganese complexes containing 1,2-diazine as the bridging ligand and halides or chalcogenide as the ancillary ligands (structural determination and a thorough characterization of all the compounds investigated in this work is reported elsewhere) [28], which has the general formula [Mn 2 (µ-ER) 2 (CO) 6 (µ-pydz)] (pydz = pyridazine; E = O or S; R = methyl or phenyl). In addition to the rhenium counterpart, only two examples of dinuclear Mn complexes with two "Mn(CO) 3 " moieties connected by only one aromatic nitrogen bridging ligand have been reported to show higher performances [16,18]. It is worth noting that dinuclear manganese complexes (as those reported in this study)-containing not only the aromatic bridging ligand but also two anionic bridging ligands-have been investigated for the efficient CO release [29][30][31] and as electrocatalysts for proton reduction [32][33][34], but no examples have been reported for CO 2 reduction reaction.
In this work, for the first time, we studied the electrochemical behavior of this family of dinuclear Mn compounds (Scheme 1) in aprotic solvents and in the presence of CO 2 to investigate its reduction reaction catalyzed by manganese-based complexes and to explore the beneficial effects on the catalytic performances of such a dinuclear geometry inspired by similar works concerning rhenium [21,22].

Results and Discussion
The electrochemical investigation of compounds Mn-Ome, Mn-SMe, and M reported in Figure 1, was performed in DMF, which was chosen over acetonitrile in order to avoid the higher coordinating properties of ACN molecules in the me

Results and Discussion
The electrochemical investigation of compounds Mn-OMe, Mn-SMe, and Mn-SPh, reported in Figure 1, was performed in DMF, which was chosen over acetonitrile (ACN) in order to avoid the higher coordinating properties of ACN molecules in the metal center. Electrolyte solutions of about 1 mM of each of the three studied complexes were prepared using dry DMF, and voltammetric measurements were taken in an inert atmosphere by Ar-saturating the solution. During the characterization procedures, the sample solutions were kept in a dark environment for the whole time, as Mn complexes are liable to photodegradation. All the systems were electrochemically characterized in a wide potential window by using a glassy carbon disk under the inert Ar atmosphere (Figure 1), before the introduction of CO 2 ; the potentials of the most significant processes are reported in Table 1.
Molecules 2023, 28,1138 4 of 14 making the prevention of some partial decomposition impossible. The photostability of the complexes in different solvents has been investigated and fully reported elsewhere [28]. Figure 1c reports the electrochemical behavior of species Mn-SPh, for which the reversibility of the first peak was mostly lost. This is in agreement with the weaker donating power of the phenylsulfide anion with respect to the methylsulfide or methoxo anions. This led to the weakening of the Mn-S bond, resulting in a possible opening of the bridge bond, formed by the ancillary ligands, and thus a partial dissociation of the complex. As a consequence, the two-electron voltammetric wave beyond −2.0 V was split into two distinct one-electron processes.   Herein, the voltammetric behavior of an Ar-saturated solution of the species Mn-OMe, Mn-SMe, and Mn-SPh was investigated in the negative potential range, as shown in Figure 1. For Mn-SMe and Mn-OMe (Figure 1a,b, respectively), the potential scan at negative values presented a completely reversible reduction with a half-wave potential E 1/2 I at about −1.0 V (vs SCE) and centered on the diazine ligand (vide infra). The reversibility of this peak clearly indicates the good stability of these complexes towards processes involving ligand dissociation and the further formation of dimers containing Mn-Mn bonds, usually observed following the reduction of mononuclear tricarbonyl Mn complexes [9,35,36].
The first reduction was followed by a second voltammetric wave with a two-electron peak at about −2.3 V, which occurred just at the edge of the potential window allowed by the solvent. This last voltammetric peak was completely irreversible, and only a careful inspection of the voltammetric features highlighted the presence of two subsequent overlapping processes with potentials very closely spaced (labeled as II and III in Figure 1a). Moreover, on the backward potential scan, three small reoxidation processes were spawned with peak potentials reported in Table 1, thus suggesting a reaction mechanism of the molecule, following up the second reduction voltammetric peak (vide infra for a description of the mechanism). In the case of Mn-OMe, two other smaller reduction peaks were observed at −0.9 V and −2.2 V, respectively (indicated with asterisks in Figure 1b), that are related to the reduction of some decomposition products. In fact, although all the CV experiments were carefully carried out in dark conditions, the Mn-OMe derivative was highly photosensitive, even upon exposure to indirect sunlight, making the prevention of some partial decomposition impossible. The photostability of the complexes in different solvents has been investigated and fully reported elsewhere [28]. Figure 1c reports the electrochemical behavior of species Mn-SPh, for which the reversibility of the first peak was mostly lost. This is in agreement with the weaker donating power of the phenylsulfide anion with respect to the methylsulfide or methoxo anions. This led to the weakening of the Mn-S bond, resulting in a possible opening of the bridge bond, formed by the ancillary ligands, and thus a partial dissociation of the complex. As a consequence, the two-electron voltammetric wave beyond −2.0 V was split into two distinct one-electron processes.
To shed light on the behavior seen in the Ar-saturated DMF electrolyte solution, simulations were performed to understand a possible undergoing mechanism. In order to obtain a set of reliable and acceptable thermodynamic and kinetic parameters for the hypothesized mechanism (vide infra), the fit of the experimental voltammetric curves was obtained within a two-order-of-magnitude range of scan rates (between 0.4 and 10 V/s). Data for the Mn-SMe sample are reported below in Figure 2.
To shed light on the behavior seen in the Ar-saturated DMF electrolyte solution, simulations were performed to understand a possible undergoing mechanism. In order to obtain a set of reliable and acceptable thermodynamic and kinetic parameters for the hypothesized mechanism (vide infra), the fit of the experimental voltammetric curves was obtained within a two-order-of-magnitude range of scan rates (between 0.4 and 10 V/s). Data for the Mn-SMe sample are reported below in Figure 2. On the basis of the electrochemical and chemical behavior of the various species, the electrochemical mechanism that reasonably fits the experimental data is sketched below in Scheme 2, and the electrochemical and chemical simulation parameters are reported in Table S1 in the Supplementary Materials. In particular, it shows a reversible one-electron process followed by two irreversible closely spaced one-electron reductions. This second process led to three different oxidation peaks on the reverse anodic scan back to the initial potential. On the basis of the electrochemical and chemical behavior of the various species, the electrochemical mechanism that reasonably fits the experimental data is sketched below in Scheme 2, and the electrochemical and chemical simulation parameters are reported in Table S1 in the Supplementary Materials. In particular, it shows a reversible one-electron process followed by two irreversible closely spaced one-electron reductions. This second process led to three different oxidation peaks on the reverse anodic scan back to the initial potential. The first reversible process can be reasonably attributed to the reduction of the aromatic diazine ring, not inducing any bond breaking. The irreversibility of the second voltammetric process (comprising two close one-electron transfers) indicates that the doubly and triply reduced molecule became unstable and that a follow-up chemical reaction has occurred. This can be associated with the double reduction of the two metal centers which most likely has a chemical impact involving the partial breaking of sul-  The first reversible process can be reasonably attributed to the reduction of the aromatic diazine ring, not inducing any bond breaking. The irreversibility of the second voltammetric process (comprising two close one-electron transfers) indicates that the doubly and triply reduced molecule became unstable and that a follow-up chemical reaction has occurred. This can be associated with the double reduction of the two metal centers which most likely has a chemical impact involving the partial breaking of sulfur-Mn bonds between the ancillary bridging ligands and the Mn nuclei.
In this respect, some preliminary molecular orbital DFT calculations on the Mn-SMe species were carried out, confirming that the first reduction is mostly centered on the diazine bridging ligand, while the subsequent two electron injections are mainly metalcentered. Moreover, while the molecular structure optimization at the M06/3-21G* level of the theory for the first reduced species Mn-SMe did not show any significant Mn-S bond elongation (only 0.02 Å), when the second electron was added, the optimization of the two-electron reduced the molecular structure, evidencing a rather large increase (0.22 Å) in one of the two Mn-SR bridging bonds for both the ancillary ligands. Such a 0.22 Å value is an undoubtedly a clear indication and theoretically supports the "opening" of the ancillary bridges (see Figure S1 in the Supplementary Materials).
From the new formed isomeric structure (right-bottom structure in Scheme 2), the reoxidation to the original valence state was possible and the processes were again irreversible because the electron transfers were affected by chemical reactions. In fact, this last chemical process leads to re-establish the bonds involving the bridging sulfurs, thus revealing a square-scheme mechanism, as shown in Scheme 2. Thus, such a reaction mechanism can be involved in the interaction with CO 2 and its reduction (vide infra). It should be noted that voltammetric curves recorded at lower scan rates (<0.4 V/s) show some degree of electrode passivation.

Electrocatalysis
After the preliminary voltammetric investigation in an inert atmosphere, the solution was saturated with CO 2 . The presence of CO 2 led to a remarkable change in the voltammetric pattern with an increase in the current of the two-electron peak occurring at a potential more negative than −2.0 V for all four systems, thus indicating that the highly reduced transient species, electrogenerated at those potentials, are affected in some way by the presence of carbon dioxide. Such an increase in the current wave overlapped with the two-electron peak (i.e., processes II + III) overwhelming it and revealing what looks like a solvent discharge, as shown in Figure 3 for all the Mn complexes. Moreover, it is worth noting the complete absence of the sequence of the three reoxidation processes on the reverse anodic scan, thus suggesting that the new process completely overturns the previously detected one in argon (Figure 3).
A very similar voltammetric behavior was also observed in an acetonitrile solution. Figure 4 reports the CV curves of all the complexes in ACN in the presence of CO 2 , revealing a strong current increase at the second reduction peak potential, as observed in the DMF solution. some way by the presence of carbon dioxide. Such an increase in the current wave ove lapped with the two-electron peak (i.e., processes II + III) overwhelming it and revealin what looks like a solvent discharge, as shown in Figure 3 for all the Mn complexe Moreover, it is worth noting the complete absence of the sequence of the three reoxid    For the sake of comparison, a blank experiment without any of the manganese complexes in solution revealed that CO 2 reduction was observed on a glassy carbon electrode only starting at −2.6 V, with a shift of~0.4-0.5 V towards a more negative potential.
Measurements showing a clear and efficient CO 2 RR, in particular for the dinuclear carbonyl Mn species Mn-SMe and Mn-SPh, allowed the turnover frequency (TOF) to be calculated for such systems. By comparing the relevant peak current in the presence of the substrate CO 2 (i cat ) and in its absence (i p ), it was possible to calculate the TOF for a homogeneous catalyst (in the case of a reversible electron transfer reaction followed by a fast catalytic reaction) using the following equations: F RT where y is equal to 1, assuming pseudo-first-order kinetics, since the concentration of substrate Q (i.e., CO 2 ) is large in comparison to the concentration of the catalyst [37]. Moreover, n cat is the number of electrons required for the catalytic reaction (n cat = 2 for the reduction of CO 2 to CO), F is the Faraday's constant, A is the surface area of the electrode, [cat] is the catalyst concentration, D is the diffusion constant of the catalytically active species, k cat is the rate constant of the catalytic reaction, and [Q] is the substrate concentration. Equation (2) describes the peak current of a reversible electron transfer not affected by any follow-up reaction [38]. R is the universal gas constant, T is temperature, n p is the number of electrons in the reversible non-catalytic reaction, and υ is the scan rate. Dividing Equation (1) by Equation (2) In this equation, factor A is canceled out because the same electrode was used for the experiments under CO 2 and Ar. The diffusion coefficient D is also canceled out because we assumed that the value of the catalytically active species does not change significantly under CO 2 or Ar. By using experimental data from ACN solutions of Mn-SMe and Mn-SPh (Figure 4), the TOF values were calculated through Equation (4).
Using the current peaks of the first voltammetric cycle (to exclude any eventual electrode passivation), the TOF values were 131 s −1 for Mn-SMe and 10 s −1 for Mn-SPh. Data from Mn-SMe and Mn-SPh show that they behaved much better, with a value above 7.2 s −1 , as reported in previous works on manganese-based systems [39]. Scans conducted on Mn-OMe provide evidence that this Mn complex is inactive toward the CO2RR.
The notable large difference of activity between the O-based and S-based bridging moieties for the two systems Mn-OMe and Mn-SMe, respectively, is proof of the influence of the ancillary ligand on the catalytic performances, as reported for the analogous rhenium complexes [22]. In Mn-SMe, both Mn-S bond cleavages take place almost simultaneously, without the decomposition of the complex and/or the formation of an Mn-Mn direct bond, allowing both the metal centers to interact with CO 2 , as shown in Scheme 3. It is unlikely that two CO 2 molecules have enough room to interact independently with both nuclei, but, as previously reported in the literature, a dinuclear reaction site can outperform two independent mononuclear ones, as demonstrated in Re-based systems [19]. Scheme 3. It is unlikely that two CO2 molecules have enough room to interact independently with both nuclei, but, as previously reported in the literature, a dinuclear reaction site can outperform two independent mononuclear ones, as demonstrated in Re-based systems [19]. However, in Mn-SPh, the two Mn-S bond cleavages can initiate a partial decomposition of the complex, due to the lower donating power of the phenyl derivative affording the splitting of the second voltammetric peak in two processes. Interestingly, the CO 2 RR peak takes place clearly when the species with one metal center, formed upon reduction, interacts with the substrate. On the other hand, the strong interaction of the OMe − anion with the metal center, due to the hard nature of the ligand and its higher electronegativity, with respect to the SMe − anion, showed to partially hamper the coordination of CO 2 , thus decreasing the efficiency.

Materials and Experimental Procedures
Tetrabutylammonium hexafluoro-phosphate (TBAH) (99.0%) purchased from Sigma-Aldrich was used as electrolyte, and N,N-dimethylformamide (DMF) (99.5%) and acetonitrile (ACN) (99.5%) purchased from Sigma-Aldrich were used as solvents. Electrochemical investigations linked to cyclic voltammetry (CV) were carried out using a BioLogic SP-300 instrument and a custom-made electrochemical cell. A 1 mm diameter GC disk was used as the working electrode; the working electrode was polished with a 0.3 µm aluminium oxide (Buehler) slurry in distilled water on a felt pad. A Pt spiral wire acted as the counter electrode and an Ag wire was used as a pseudo-reference electrode that was checked to have a stable potential with a negligible drift within the time of the experiment comprising several potential sweeps in the electrolyte/solvent used. The potential scale was corrected vs the saturated calomel electrode (SCE) using the redox couple ferrocene/ferrocenium as an internal reference. The solvent, containing both DMF and ACN, was purified and dried by distillation and stored over activated 4 Å molecular sieves in a Schlenk flask. It was transferred through a cannula system into a custom-designed electrochemical cell containing the supporting electrolyte and the species under examination, immediately before performing the experiment [22]. The argon and carbon dioxide were initially bubbled through the electrolyte solution and as a blanket of gas at a pressure of 1 bar while measurements were performed.
All the E 1 2 potentials have been directly obtained from cyclic voltammetric curves as averages of the cathodic and anodic peak potentials and by digital simulation in the case of non-Nernstian or overlapping processes. Digital simulations of the cyclic voltammetric curves were carried out by Antigona [40] or DigiSim 3.0 (Bioanalytical Systems, Inc. West Lafayette. IN, USA), utilizing a best-fitting procedure of the experimental curves recorded at different scan rates spanning over at least two orders of magnitude [41,42].
The chalcogenide-bridged dinuclear Mn carbonyl complexes herein investigated (see Scheme 1) were obtained through the one-pot synthesis "orthogonal bonding approach" [28], which was previously used for the synthesis of analogue rhenium dinuclear complexes [27] and different dinuclear manganese complexes [30]. The oxidative addition of E 2 R 2 ligands (E = S and R = methyl or phenyl) to Mn 2 (CO) 10 and the further addition of the pyridazine bridging ligand affords complexes labeled as Mn-SMe and Mn-SPh in good yields (about 60%). For the synthesis of the methoxo derivative Mn-OMe, methanol was used as the source of -OMe anions to obtain the corresponding managanese complex Mn-OMe. The structural determination of products and the thorough characterization of all compounds investigated in this work have been reported elsewhere [28].

Conclusions
We described and electrochemically characterized three manganese complexes with different bridging ancillary chalcogenide ligands as promising homogeneous CO 2 RR catalysts. As revealed by the analogous rhenium complexes, the catalysis efficiency, quantified using various methods outlined in the literature, confirms the strong correlation between the molecular structure and catalytic activity in CO 2 reduction. The maximum TOF was observed for complex Mn-SMe, affording an impressive CO 2 RR activity with TOF values up of 131 s −1 that is much higher than those reported by previous works on manganese-based systems, reaching a maximum of 7.2 s −1 . The three-order-of-magnitude difference between the two very similar systems also reveals interesting insight into the mechanism that takes place at the reaction site, showing the effect of the ancillary ligand on catalytic activity. We hypothesized a mechanism based on electrochemical findings, indicating that the reaction site for CO 2 RR is formed when the bond with the ancillary bridging ligand is cleaved so that the CO 2 molecule can access the metal center. In particular, this feature is in agreement with the lower electronegativity of the sulfur with respect to oxygen, together with the higher donor power of methyl with respect to phenyl, which increases the electronic density of the Mn centers and improves the stability of the reduced species, thus increasing the catalytic activity.