Unraveling the Effect of Aromatic Groups in Mn(I)NNN Pincer Complexes on Carbon Dioxide Activation Using Density Functional Study

Carbon dioxide utilization is necessary to reduce carbon footprint and also to synthesize value-added chemicals. The transition metal pincer complexes are attractive catalysts for the hydrogenation of carbon dioxide to formic acid. There is a need to understand the factors affecting the catalytic performance of these pincer complexes through a structure–activity relationship study using computational methods. It is a well-established fact that aromatic functionalities offer stability and selectivity to transition metal catalysts. However, their impact on the performance of the catalysts is lesser known in the case of metal pincer complexes. Hence, it is necessary to investigate the catalytic performance of Mn(I)NNN pincer complexes with variably activated aromatic functionalities. In this context, 15 catalysts are designed by placing different types of aromatic rings at the pincer carbons and two terminal nitrogen of Mn(I)NNN pincer complexes. A benzene moiety, placed at C2–C3 carbons of Mn(I)NNN pincer complex with identical aromatic groups at the terminal nitrogen, is found to be most efficient toward CO2 hydrogenation than the rest of the catalysts. On the other hand, when N,N-dimethyl aniline is placed at C2–C3 carbons of Mn(I)NNN pincer complexes, then the catalytic performance is significantly decreased. Thus, the present study unravels the impact of aromatic groups in Mn(I)NNN pincer complexes toward the catalytic hydrogenation of carbon dioxide.


INTRODUCTION
The burning of fossil fuels continuously increases CO 2 concentration in the atmosphere, leading to a substantial and negative impact on the world climate. About 33,890.8 million tons of CO 2 was released into the atmosphere in 2018, and the global CO 2 concentration in the atmosphere reached 407.65 (BP Statistical Review of World Energy, 2019; Global Monitoring Division, 2019) Nearly ∼35 GT of CO 2 is being added to the atmosphere per year, and there is a considerable gap between the amount of CO 2 produced and utilized. Therefore, carbon dioxide utilization becomes a necessity to save the world from global warming. CO 2 utilization would not only help to remove CO 2 from the atmosphere but also helps to get alternate fuels and to reduce dependence on petrochemicals and, thus, restricts CO 2 expulsion in the atmosphere (Centi and Perathoner, 2009;Balaraman et al., 2011). Apart from forming fuels from carbon dioxide, many synthetically useful chemicals are being synthesized from carbon dioxide (Aresta, 2010). The conversion of formic acid from carbon dioxide has manyfold benefits to the environment and economy. Formic acid is considered to be potential chemical hydrogen storage material because of its stability, nontoxicity, and easy accessibility (Dörthe et al., 2016;Jörg et al., 2017;Miriam et al., 2019;Ganesh, 2014;Kassem et al., 2020). The global formic acid market is growing at a CAGR of 1.3%, and it is expected to reach US$828.1 million by 2025 due to its wide applications in agriculture, leather, textile, rubber, chemical and pharmaceuticals, and industries (https://dataintelo.com/report/formic-acid-market). Therefore, developing an energy-efficient and environmentally benign protocol to get formic acid from CO 2 becomes significant.
In carbon dioxide activation and formation of formic acid or formate derivatives, the use of rhodium, ruthenium, and iridium pincer complexes have made remarkable progress in recent years (Wen, et al., 2021;Wang et al., 2013;Huff andSanford, 2011, Wesselbaum et al., 2012;Kothandaraman et al., 2016;Wesselbaum et al., 2015;Campos et al., 2014). However, the high price and limited availability of these metals restrict their wide applications to an industrial scale. Therefore, the development of new catalytic technologies based on earth-abundant metals such as Fe, Mn, Ni, and Co is of considerable interest (Figure 1) (Zell and Milstein, 2015;Furstner, 2016;Langer et al., 2011;Zhang et al., 2015;Ge et al., 2016;Choi and Lee., 2020;Curley et al., 2018). The transition metal pincer complexes of these earth-abundant metals are of particular importance due to their thermal stability, cost efficiency, and flexibility for modification. The flexibility in the modification could be useful to fine-tune structural and electronic properties of the metal pincer complexes to make them more reactive as well as more selective (Peris and Crabtree, 2018). In this context, Mn pincer complexes in catalytic carbon dioxide hydrogenation have seen much progress over a much shorter time (Bertini et al., 2017;Garbe et al., 2017;Kar et al., 2017). The PNP, PCP PNN pincer complexes have been commonly used to develop transition metal pincer complexes (Kumar et al., 2019;Bernskoetter and Hazari, 2017;Irina et al., 2016;Bertini et al., 2016;Jan et al., 2020). It is necessary to understand the donor-acceptor strength of the ligand during the rational ligand designing for new catalyst development. In all the cases, the metal-ligand interaction will depend highly on the choice of the transition metal, oxidation state, coordinating sites of the ligands, and the substituents on ligands. Steric bulk is also an extremely important aspect not only for enhancing the stability of the complexes but also for providing stereoselectivity (Garbe et al., 2019;Wen et al., 2021;Jessica et al., 2017). The NNN pincer ligands are of particular importance due to their accessibility, scalability, stability, and affordability. In the recent review, Crabtree mentioned that there is a necessity to explore transition metal NNN pincer complexes and to understand their catalytic performance (Peris and Crabtree, 2018). Herein, 15 Mn(I)NNN pincer complexes are designed to understand steric and electronic factor ligands on the catalytic efficacy toward carbon dioxide hydrogenation ( Figure 2). However, it would be difficult to understand the efficacy of these 15 Mn(I)NNN complexes and also to throw light on the impact of aromatic and heterocyclic rings present in the NNN pincer ligands by using experimental methods. Therefore, the use of computational methods to assess a large number of complexes by evaluating the mechanistic pathway and energetics of the reaction is highly desirable. GRAPHICAL ABSTRACT | TOC for paper.

COMPUTATIONAL DETAILS
The Gaussian-16 program package was used to perform the computational calculations (Frisch et al., 2016). The meta-GGA hybrid M06 functional with dispersion corrections without the imposition of any symmetry constraints were used to optimize reactants and intermediates (Zhao and Truhlar, 2008a). M06 is the accurate, economical functional for transition metals than B3LYP functional (Hehre et al., 1972). The SDD basis set was used for manganese, and Pople's 6-31G(d,p) basis set was used for all the main group elements (Hariharan and Pople, 1973;Hay and Wadt, 1985). All complexes were treated as neutral species to compute optimized energies. The ground states of intermediates and transition states were confirmed as singlets through the comparison with optimized high-spin analogs. All transition states exhibited a single imaginary frequency, corresponding to the eigenvector along the reaction path. Frequency analysis of all the stationary points was performed at the same level of theory to confirm stationary points as minima or first-order saddle points along with the reaction coordinate (Liu et al., 2012;Jain et al., 2018;Zhao and Truhlar, 2008b). Intrinsic reaction coordinates (IRCs) were carried out on the transition states to endorse that such structures were indeed connecting the two minima, reactant, and product (Fukui, 1970;Fukui, 1981). All the calculations presented in this work were performed in the presence of water (dielectric constant 78.39) using the integral equation formalism variant-like solvation model based on density (SMD) .
NBO analysis was carried out using the NBO 3.1 suite as implemented in Gaussian-16 (Foster and Weinhold, 1980;Glendening et al., 1987;Carpenter et al., 1988;Reed et al., 1988). The second-order perturbative estimation of donor-acceptor stabilization energy (E s ) was calculated using the following equation, where q i is the donor orbital occupancy number, and F ij is the offdiagonal element of the Fock matrix in the NBO basis. Δε ji ε j − ε i is the orbital energy difference between the acceptor (j) and donor (i) NBO. The solvent-corrected relative Gibbs free energies (ΔG) for the transition states and intermediate as well as overall activation Gibbs free energies (ΔG ⧧ ) for the catalytic cycle was calculated at 1 atm pressure and 298.15 K temperature. Visualization of all optimized structures was performed using the CYLview software (CYLview, 2020), and imaging of all NBO structures and orbitals was performed using the Chemcraft Visualization software (Chemcraft, 2021).

RESULT AND DISCUSSION
Hydrogenation of Carbon Dioxide to Formic Acid Catalyzed by Group 1 Catalysts, 1a-1e Catalysts 1a-1e are designed to explore the impact of donating and withdrawing groups on the catalytic performance in the 3 kcal/mol) by forming an Mn-OCHO bond. This is the most stable state TDI (Turn-over Determining Intermediate), and hence, it is also a rate-determining state of the reaction. The IN3R again isomerizes to IN3 with an energy barrier of ∼15.0 kcal/mol, and then, the protonation takes place. The hydrogen molecule adds to IN3 to provide the IN4 comprising formic acid. The relative Gibbs free energy of the IN4 is minimum for 1c (−4.5 kcal/mol) and maximum for 1e (1.4 kcal/mol) ( Figure 5).
Interestingly, we found that in Group 1, the catalyst with a shorter Mn-N (1a, Mn-N 2.13 Å) distance performs better than the catalyst with a longer Mn-N bond (1e,  Mn-N 2.20 Å) distance ( Figure 6). However, there is no significant change in the Mn-N bond distance during the catalytic reaction. The presence of a strong H-bond (HCOO-H 2 1.64 Å) in 1c than in 1e (HCOO-H 2 2.43 Å) lowers the relative Gibbs free energy in 1c (−4.5 kcal/mol) than 1e (1.36 kcal/mol) in IN4. However, TS[4-5] of 1c is found to have higher Gibbs free energy than 1b, due to strong H 2 polarization and strong bond-forming interactions in 1b (1.00 Å) than in 1c (0.98 Å). A similar dihydrogen polarization is found to be efficient in the M-H 2 σ-complex of 1c (0.81 Å) than in 1e (0.78 Å). The IN4 is further converted into IN5, the regenerated catalyst and formic acid complex through the TS[4-5] with the relative Gibbs free energy of 3.3-6.0 kcal/mol for all the catalysts. However, the activation energy barrier for this step is the lowest for 1a (3.3 kcal/mol) and the highest for 1c (6.0 kcal/mol). In this entire mechanism, the dihydrogen dissociation and catalyst regeneration steps are more energy demanding by 1-3 kcal/mol than the hydride transfer transition state for the catalyst complexes 1c, 1d, and 1e. On the other hand, the catalysts complexes 1a and 1b show facile protonation via dihydrogen polarization than the hydride transfer from the Mn center to carbon dioxide. In overall CO 2 to formic acid formation, catalysts 1a and 1d have almost the same catalytic efficiency as the rest of the catalysts with ΔG ⧧ of 23.9 and 24.6 kcal/mol (Table 1), whereas catalyst complex 1c is most sluggish toward carbon dioxide hydrogenation. This indicates that unsubstituted carbon centers of Mn(I)NNN with unsubstituted nitrogen pincer ligands of Group 1 show better catalytic performance than the one with electronwithdrawing groups. The donating groups and moderate withdrawing groups at terminal nitrogen make 1b and 1c catalysts less reactive than the Mn(I)NNN complex 1a and

Hydrogenation of carbon dioxide to formic acid catalyzed by Group 2 catalysts 2a-2e
The unsubstituted pincer ring carbons provided exciting results; therefore, we decided to investigate the effect of aromatic substituents of the pincer carbons (C2 and C3) on the catalytic performance (Curley et al., 2018;Kumar et al., 2019;Luca et al., 2020) (Figure 2). Herein, C2 and C3 carbons of catalysts 1a-1e are used to attach the unsubstituted aromatic ring to form the new pincer complexes 2a-2e (Figure 7). After the addition of carbon dioxide to these pincer complexes, IN2 is formed. The relative Gibbs free energy of the IN2 is considered as 0.0 kcal/mol to calculate the relative Gibbs free energies of the transition state and the intermediate states ( Figure 8) (Figure 9). The    Frontiers in Chemistry | www.frontiersin.org November 2021 | Volume 9 | Article 778718 with the relative Gibbs free energy of 2.0 to 5.0 kcal/mol. The IN5 contains the product, formic acid, and the regenerated catalyst complex. The IN5 further expels formic acid and provides regenerated catalysts, 2a-2e. The ΔG for the entire catalytic reaction is found to be minimum for catalysts 2b (22.1 kcal/ mol) and 2e (22.8 kcal/mol) and maximum for catalysts 2c (27.2 kcal/mol) and 2d (27.7 kcal/mol) ( Table 2). This indicates that in Group 2, catalysts show better catalytic performance when both the terminal nitrogen are attached to either electron-donating groups, N,N-dimethyl aniline, or strong electron-withdrawing groups, pyrimidine. On the other hand, the catalytic performance slows down when these two terminal nitrogens carry mixed electron-withdrawing and donation groups 2a and 2c. However, catalyst 2a with an unsubstituted terminal nitrogen shows average performance toward carbon dioxide hydrogenation.
Hydrogenation of carbon dioxide to formic acid catalyzed by Group 3 catalysts, 3a-3e The unsubstituted Group 1, as well as benzene, substituted Group 2 manganese pincer complexes provided a deeper insight into their catalytic performances toward carbon dioxide hydrogenation reaction. Therefore, we also investigated the effect of the activated aromatic ring by placing N,N-dimethyl aniline at the C2-C3 of the pincer ligands of the Mn(I)NNN complex in Group 3 (Figures 2 and 10) . Complexes 3a-3e are designed to understand the effect of Mn(I)NNN complexes on carbon dioxide hydrogenation. After the addition of carbon dioxide to these pincer complexes, the IN2 is formed. The relative Gibbs free energy of the IN4 is considered as 0.0 kcal/mol to calculate the relative Gibbs free energies of the transition and the intermediate states ( Figure 11).     [4-5] with a maximum relative Gibbs free energy for catalyst 3c (9.9 kcal/mol) and minimum for catalyst 1a, 1d, and 1e (∼2.0 kcal/ mol). The entire CO 2 hydrogenation is found to be thermodynamically favorable for all the catalysts except for catalyst 3c. This indicates that electron-donating and weakwithdrawing groups at an Mn center make dihydrogen polarization and catalyst generation sluggish. On the other hand, unsubstituted terminal nitrogen and terminal nitrogen with strong withdrawing groups facilitate the dihydrogen polarization and carbon dioxide to the formic acid formation (Table 3).

NATURAL BOND ORBITAL
Natural bond orbital (NBO) analysis is performed to gain a mechanistic insight into the carbon dioxide hydrogenation reaction mechanism (Chemcraft, 2021 formation based on the second-order perturbative interaction energy. At the transition state TS[2-3], prominent bondforming interactions are observed for the C-H(a) bond. (Figure 13).
Here, the TS[2-3] of 1e has stronger bond-forming interaction energy (45.01 kcal/mol) and lower relative Gibbs free energy(3.0 kcal/mol), whereas the TS[2-3] of 1a shows weaker bond-forming interaction energy of (13.53 kcal/mol) and higher relative Gibbs free energy (4.8 kcal/mol). A similar trend is observed when NBO analysis is performed for the IN4 of all the catalysts (Figure 14).

CONCLUSION
The structure-activity relationship study of computationally modeled Mn(I)NNN pincer complexes emphasize that the position and nature of the aromatic rings attached to the pincer ligands affect the catalytic performance to a considerable amount for carbon dioxide hydrogenation reaction. The Group 2 Mn(I)NNN pincer complexes with benzene substituent at C2-C3 and identical substituents at both the terminal nitrogen are superior to all the catalysts from the three groups, whereas Group 1 catalysts without any aromatic substituents at C2-C3 show moderate catalytic Frontiers in Chemistry | www.frontiersin.org November 2021 | Volume 9 | Article 778718 performance, and Group 3 catalysts with N,N-dimethyl aniline at C2-C3 are sluggish toward carbon dioxide hydrogenation.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
SP and VA conceived the project. SVP carried out the computational calculations and analysis of data. All authors were involved in the manuscript writing and revision. All authors have made substantial, direct and intellectual contributions to the work, and approved it for publication.