Capture and Reaction of CO2 and H2 Catalyzed by a Complex of Coronene: A Computational Study.

An organometallic complex of coronene and chromium was designed and used as a catalyst in a simulated process in which a CO 2 molecule is captured, activated, and then reacts with a hydrogen molecule (H 2 ) to yield formic acid (HCOOH). The thermodynamic and kinetic results from this process are analyzed, as well as the nature of the orbitals in which the process is carried out.


Introduction
The presence of CO 2 in the earth's atmosphere has existed since the archaic era, rst resulting from volcanic activity and subsequently because of the presence of living beings, although for many years, the phenomenon of photosynthesis helped maintain an equilibrium in terms of the presence of this substance on the planet [1]. However, at present, the increase in human population and the boom in fossil combustible compounds have caused a serious disequilibrium concerning the quantity of this chemical agent, giving place to the greenhouse phenomenon [2]. The amount of CO 2 in the atmosphere has changed from 325 ppm in 1980 to 410 ppm in 2019 [3], but was only 280 ppm at the beginning of the XIX century [4] and the emission per year derived from human activities is of 37 GTon of CO 2 [5]. If this tendency continues, the temperature of the atmosphere will rise by 2°C over the next 50 years; therefore, it is mandatory to reduce the current level by 45% in the near future, to avoid a very taxing atmospheric situation [6].
A possible strategy for curtailing CO 2 accumulation is to capture it and in the best case scenario transform it into useful products [7]. This set of procedures has been termed Carbon Capture Utilization and Storage (CCUS) by several organizations that are interested in research with the potential to develop methods, processes and chemical agents (catalysts, adsorbents, etc.), aimed at capturing and transforming CO 2 . In this context, there are several groups throughout the world working towards particular goals in this matter [8][9][10][11][12][13]; indeed the subsequent transformation of carbon dioxide into simple organic substances makes the capture of this gas an attractive procedure because of its intrinsic economic bene t. It is therefore possible to predict an amount with the production of chemicals such as formaldehyde, formic acid, methanol, etc. [14,15].
The proposition is the molecule shown in Fig. 1: The molecular complex being studied represents a large analogue of bis(benzene)chromium [18], in which the coronene plays the role of the original benzene decks and is complemented by three chrome atoms, put in a position which maximizes Clar's theorem aromaticity [19]. This peculiar sandwich structure for the Cr 3 -Cor 2 complex (Compound 1) was proposed by Duncan et al. [20] to explain the photofragmentation processes of Cr m (Cor) n + clusters produced by laser vaporization. In that study, the authors pointed out the di culty involved in fragmenting the 3:2 complex, possibly indicating its high stability. However, no one has successfully attempted to isolate complexes like these and determine their properties. In addition, few studies on chromium-coronene sandwich complexes have been published, the most relevant being one carried out by Philpott et al. [21], who reported a C 2V symmetry structure with a chromium atom located on the outer edge of one of the perimeter rings. Therefore, it would be the rst time that a D 3h symmetry structure has been reported for the Cr 3 -Cor 2 complex. However, giving a detailed description of the complex is not the purpose of this work, but to take advantage of the stability of this system to promote a reaction between CO 2 and H 2 .
The ability of L n -MT metal complexes to catalyze CO 2 transformation reactions, by means of different mechanisms, is well known [2,7,22]. In this case, the electron-rich region within the complex, or around the Clar sites where the metal atoms are located, can be used as a catalytic activity zone.
The aim of this work is to design a molecule of this kind and subsequently carry out DFT calculations to model the CO 2 hydrogenation reaction through the Cr 3 -Cor 2 complex. The interaction and how this reaction may proceed to produce HCOOH are analyzed. The results are discussed, as well as the main thermodynamic and kinetic data of the process. Propositions about the molecular orbital role are also studied.

Methods
Results and structures reported in this work were mainly obtained from DFT calculations of electronic structure, for which we used the functional meta-GGA of Truhlar M06 [23], as recommended for applications in organometallic chemistry and for non-covalent interactions. All calculations were performed with the 6-31G* base set and Gaussian 16 computational software [24]. Structures were optimized without symmetry constraints and then con rmed as either steady states (GS) or transition states (TS), by evaluating vibrational frequencies using the same level of theory in all cases. Besides this, zero point energy (ZPE) has been added to correct the electronic energies of all molecules and is included in the binding energies (B. E.). Similarly, we evaluated Grimme's correction (GD3) [25] in order to consider the energy resulting from long-range dispersive interactions. are often compared [29]. To model the hydrogenation reaction, several methods were used: a simple optimization was performed and later, a QST2 calculation was carried out, which involves the Cr 3 -Cor 2 complex and the reagents of interest (CO 2 and H 2 ), to nd a TS leading to the formation of HCOOH.
Finally, these results were validated by means of an IRC calculation, using the identi ed TS.

Structure and Aromaticity
The main feature of the trimetallic sandwich structure is the existence of three equivalent sites, referred to here as Clar sites, where the chromium atoms are located. (See Fig. 1) This arrangement maximizes the separation between metal atoms (4.32 Å) and minimizes the disruption of the π bond system of the coronene molecule [20]. The η 6 -coordination at the center of the ring produces changes in all the bonds around the Clar site, but in general we can identify three types of rings (a, b and c), clearly differentiated by their structural and aromatic properties, as discussed below. The average distance from the metal atom to the complexing ring (Cr-R) is 1.59 Å and for the separation between coronene units, the average C-C' distance between ligands is 3.21 Å. In addition, the six Cr-C bonds have an average length of 2.14 Å, therefore, there is marked geometric similarity between the Clar sites and the structure presented by bis(benzene)chromium [30]. (See Table 1) It is thus possible to consider that the Clar sites of the Cr 3 -Cor 2 complex are analogous to the Cr-Bz 2 fragment. One way to validate this is by checking the presence of aromatic sextets in the alternated (type a) rings of the complex. Table 1 Representative distances in the Cr 3 -Cor 2 complex. For comparison, the reported values of bis(benzene)chromium [30] (theoretical and experimental).  (Table 2). In the coronene molecule, evidently the inner ring is less aromatic than the rings on the periphery, as expected [32]; in another context, three different values are observed for the Cr 3 -Cor 2 complex, with the type b outer rings being less aromatic than the Clar sites and with the central ring being rather anti-aromatic, according to the values from the NICS (ZZ)1 index [28]. Therefore, the presence of the aromatic sextets, not only explains the general stability of the complex as a result of the local aromaticity of the Clar sites, but is also important for understanding the reactivity of these sites.

Electronic and bonding properties
Initiating with the representation of the coronene molecule, according to the AdNDP model, [31] we can also describe the binding scheme in the Cr 3 -Cor 2 complex. The image produced using the method developed by Zubarev, et al. [31,32] (see Fig. 2a) is obtained by distributing the 24 π electrons of the coronene into the following groups: 6 π bonds in the periphery (12e), which are maintained in the complex; 3 π bonds in the central ring (6e) reposition towards the Clar sites; Finally, the 3 delocalized π bonds on the entire surface of the coronene (24C-6e) can also be relocated to the Clar sites, meaning that there are now 6 π-electrons provided by each of the a-type rings (12e contributed by the fragment sandwich). Likewise, there is a total count of 18e at each Clar site, just as occurs in the Cr-Bz 2 . Thus, the exposed model makes it possible to explain the structural resemblance of the Clar site to bis(benzene)chromium, also demonstrating the importance of the C-C bonds that join the central ring to the external 18C fragment, [33,34] which help to stabilize the system from the redistribution of electronic density in charge transfer processes.
The union of the chromium atom with the sandwich fragment, which represents the coronene dimer, is produced by the overlap of the AO's of the metal with the MO's of the π system. The same bonding scheme of Cr-Bz 2 [30,35] can be applied in this case, as we can identify similar characteristics, such as the shape of the dz 2 atomic orbital in the HOMO of the complex, (see Fig. 3). From the ordering of the energy levels, it is apparent that the Cr 3 -Cor 2 complex can donate a pair of electrons from the HOMO, through the circular lobe of the dz 2 orbital of Cr (Fig. 3b), at any of the three Clar sites. In contrast, LUMO orbitals ( Fig. 3a) extend throughout the coronene molecule, connecting the metal atoms with both ligands, enabling a large part of the electron density donated by a donor to be delocalized and distributed to the C's, thus stabilizing the complex and maintaining its planarity. In summary, there is a donation from the MO's π to the AO's of the unoccupied metals (M←L); similarly, a back-donation takes place from the occupied orbitals of the M to the MO's π* (LUMO) of the coronene (M→L), the latter having greater relevance, as indicated by the values for NBO charges.

Energies and interactions in the Cr 3 -Cor 2 complex
Page 6/17 The separation that exists between the ligands (3.2 Å is smaller than that observed in the coronene sandwich dimer (3.6 Å) [36]. The interaction between ligands would tend to be repulsive. However, at this precise distance, the overlap and the orbital interactions between the metal and the MO's of the π system are maximized, making it possible to compensate for the repulsion of the Assessing interaction between the Cr 3 -Cor 2 complex and a CO 2 molecule represented a step prior to evaluating the reaction of interest. There are various modes of M-CO 2 coordination, which can include either a metal center or multiple metallic atoms, and these have been described in the eld of organometallic chemistry [22b]. For the Cr 3 -Cor 2 complex, several stationary states were identi ed with the adsorbed CO 2 molecule, nding that interactions involving more than one metal center usually occur inside the complex, whereas in the case outside the complex, the CO 2 only interacts with a Clar site.
These stable structures represent metal carbon dioxide complexes (L n M-CO 2 ), which are important catalysts in CO 2 conversion reactions, promoting both stoichiometric reactions, with the formation of C − C or C − X bonds, as well as reducing or dissociating CO 2 . [7,22] The most stable L n M-CO 2 complexes are precisely those which undergo an interaction with a single metal atom, as conserving the other two Clar sites only produces distortions at the edge of the said a-type site, as if the sandwich bites the host tail. (Fig. 4) Inside the complex, the CO 2 molecule promotes the opening of the structure, due to the repulsion of the Cor 2 fragment; while on the outside, some Cr − C haptic bonds are still preserved at the Clar site, and these hold the ligands together.
The CO 2 adsorption energy on the part of the complex is in all cases positive, due to the loss of stability in the system. The structures shown in Fig. 4 are the most probable, with a binding energy (B. E.) of + 2.84 eV and + 3.62 eV (Compound 2 and 3, respectively). The interaction of CO 2 with the complex can be described as the η 2 -coordination mode or the η 1 -mode. (See Fig. 5a) In the rst case, a d-type OA of chromium overlaps with the antibonding π* orbital of CO 2 ; whereas in the η 1 -mode, the π* orbital overlaps with OA dz 2 . In both cases, the result is an electron density transfer from the metallic center towards the LUMO of the CO 2 , when the union occurs by means of the double bond or the carbon atom and one of the oxygens; this charge transfer (CT) tends to bend the CO 2 molecule and elongate the various bonds. (See data in Fig. 4).
In another context, the Cr 3 -Cor 2 complex with an energy gap of 2.64 eV (which manifests semiconductor behavior), being able to donate and accept electron density represents something that is necessary to achieve simultaneous acid-based activation through η 2 -coordination, as this mode of interaction is also related to the double bond scheme of the DCD model. [7,22] 3.4 Hydrogenation of CO 2 to formic acid The possibility of carrying out the CO 2 hydrogenation reaction through the Cr 3 -Cor 2 complex was evaluated, for which a direct interaction of the molecules involved (CO 2 and H 2 ) in the vicinity of a Clar site was tested. (Fig. 6a) The formation of stable L n M-CO 2 complexes has previously been veri ed; therefore, the idea is to take advantage of the fact that the M − CO 2 interaction activates the CO 2 molecule [11] and subsequently when approaching an H 2 molecule (about 1.1 Å of CO 2 ) hydrogenation occurs and yields formic acid (HCOOH). (Fig. 6b) The calculation presented here indicated that changes in energy minimization lead to the formation of the product of interest by an optimization process that nally achieves the release of HCOOH from the Cr 3 -Cor 2 complex. Because the reaction takes place outside the complex, it is easy for the desorption of the nal product to take place, as the binding energy M − CO 2 is positive and presents a bond polarity M δ+ −C δ+ . Likewise, the regeneration of the Cr 3 -Cor 2 complex to its original structure after the reaction, which acts as a catalyst per se, contributes to promoting the release of HCOOH.
For this reaction to occur, a relevant aspect involves the orientation of the reagents CO 2 and H 2 . Both types of coordination can favor a reaction with H 2 : the oxygen atoms concentrate the charge donated by the metal (which causes the activation of CO 2 ), while the carbon presents a partial positive charge.
Furthermore, because the Clar site tends to concentrate electron density, this environment contributes to causing the distortion and stabilization of the M − CO 2 fragment. Therefore, when the interaction with H 2 is promoted, the H − H bond is broken and hydrogen is added twice over; rstly into one of the oxygens and then in the carbon atom. In order for this process to occur in a concerted manner, the H 2 molecule must be close to the intermolecular plane of carbon dioxide.
In another test of the study process, reagents were placed in the vicinity of the Clar site in a favorable con guration as described above, and a QST2 calculation was carried out by placing the product of interest next to the Cr 3 -Cor 2 complex. After nding a suitable transition state (TS), in which both reactants are in an activated complex form, this structure was used to perform an IRC (Intrinsic Reaction Coordinate) type calculation, in order to ensure that the TS connects reactants and products. IRC results show the complex with its trimetallic sandwich structure and HCOOH as the nal state (Fig. 7). As a result of the dual addition of hydrogen on both sides of the C = O bond, the cis form of HCOOH is obtained, which has a rotation barrier of 5.3 kcal/mol, when it reaches its most stable form: the HCOOH-trans isomer. The thermodynamic data of interest are shown in the lower gure, resulting in an activation barrier of about 54.8 kcal/mol.

Conclusions
A complex formed out of two coronene and three chromium atoms was computationally designed to work as a reactor in a capture and reaction process, which involves the con ictive CO 2 molecule. The CO 2 hydrogenation reaction through the Cr 3 -Cor 2 complex was studied theoretically. The system is capable of capturing a CO 2 molecule and activating it to subsequently react with H 2 and produce formic acid. The molecular fragment associated with the Clar site concentrates and transfers electronic density to CO 2 , in addition to providing stability to the system. The results showed that the study reaction occurs in this way and after releasing the product of interest, the Cr 3 -Cor 2 complex is regenerated to its initial structure.
Declarations Figure 1 Two views of the study system, the Cr 3 -Cor 2 complex.