TYPE Original Research PUBLISHED 20 March 2023 DOI 10.3389/fchem.2023.1144189 Theoretical study of the catalytic OPEN ACCESS performance of Fe and Cu EDITED BY Nino Russo, University of Calabria, Italy single-atom catalysts supported REVIEWED BY Xiaoli Yang, on Mo2C toward the reverse Qingdao University, China Rajib Kumar Singha, Poornaprajna Institute of Scientific water–gas shift reaction Research (PPISR), India *CORRESPONDENCE Wenjuan Zhang 1†, Anna Vidal-López 1† and Aleix Comas-Vives 1,2* Aleix Comas-Vives, 1 aleix.comas@uab.cat, Department of Chemistry, Universitat Autònoma de Barcelona, Catalonia, Spain, 2Institute of Materials aleix.comas@tuwien.ac.at Chemistry, Technische Universität Wien, Vienna, Austria † These authors have contributed equally to this work and share first authorship SPECIALTY SECTION This article was submitted to Theoretical The reverse water–gas shift (RWGS) is an attractive process using CO2 as a and Computational Chemistry, chemical feedstock. Single-atom catalysts (SACs) exhibit high catalytic activity a section of the journal in several reactions, maximizing the metal use and enabling easier tuning by Frontiers in Chemistry rational design than heterogeneous catalysts based on metal nanoparticles. In this RECEIVED 13 January 2023 study, we evaluate, using DFT calculations, the RWGS mechanism catalyzed by ACCEPTED 27 February 2023 PUBLISHED 20 March 2023 SACs based on Cu and Fe supported on Mo2C, which is also an active RWGS catalyst on its own. While Cu/Mo2C showed more feasible energy barriers toward CITATION Zhang W, Vidal-López A and CO formation, Fe/Mo2C presented lower energy barriers for H2O formation. Comas-Vives A (2023), Theoretical study Overall, the study showcases the difference in reactivity between both metals, of the catalytic performance of Fe and Cu evaluating the impact of oxygen coverage and suggesting Fe/Mo2C as a potentially single-atom catalysts supported on Mo2C toward the reverse water–gas active RWGS catalyst based on theoretical calculations. shift reaction. Front. Chem. 11:1144189. KEYWORDS doi: 10.3389/fchem.2023.1144189 reverse water–gas shift (RWGS) reaction, single-atom catalysis (SAC), Cu/Mo2C, Fe/Mo2C, COPYRIGHT DFT calculations © 2023 Zhang, Vidal-López and Comas- Vives. This is an open-access article distributed under the terms of the 1 Introduction Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is As populations and living standards increase, so does our consumption of fossil fuels, permitted, provided the original author(s) coal, oil-derived combustibles, and natural gas. These energy sources eventually transform and the copyright owner(s) are credited and that the original publication in this their carbon content into carbon dioxide (CO2), a significant greenhouse gas contributing to journal is cited, in accordance with global warming and climate change (Karl and Trenberth, 2003; Olah et al., 2011; Lim, 2015; accepted academic practice. No use, Rodriguez et al., 2015). Consequently, capturing CO2 and converting it into fuels and distribution or reproduction is permitted which does not comply with these terms. commodity chemicals have attracted considerable attention to mitigate their adverse environmental effects on Earth (Zhang et al., 2019). The reverse water–gas shift (RWGS) reaction (Eq. (1)) is a promising CO2 utilization and capture technology because its product can be used directly as feedstock in the Fischer–Tropsch (FT) process, MeOH synthesis processes, and other syngas processes (Guharoy et al., 2019; Jing et al., 2019; Zhang et al., 2019). CO2 + H2 ↔ CO + H2 O ΔH298 K  41.2 kJ mol−1 . (1) Due to the importance of RWGS from both points of view, considerable attention is being paid to improving the reaction kinetic fundamental and practical aspects and designing more efficient RWGS catalysts (Wang et al., 2011; Guharoy et al., 2019). Metal-based catalysts for the RWGS reaction are based on supported particles (Wang Frontiers in Chemistry 01 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 et al., 2011). Nevertheless, an emerging class of catalysts enabling dry reforming of methane, another CO2 conversion process (Kurlov optimal metal utilization is single-atom catalysts (SACs) (Qiao et al., et al., 2020). However, the effect of changing the oxygen coverage on 2011; Lin et al., 2013; Wei et al., 2014; Yang et al., 2015; Li et al., 2016; the catalytic activity was not evaluated in depth. In our previous Wang et al., 2016; Lu et al., 2018; Mondelli et al., 2018; Li et al., work, we found that Cu SACs on Mo2C are highly active catalysts 2020a; Li et al., 2020b; Kaiser et al., 2020; Gao et al., 2021; Xiong toward the CO2 hydrogenation to methanol, showing higher et al., 2021; Zhu et al., 2021). SACs are based on an isolated metal catalytic activity than that of unsupported Cu and Cu/ZnO atom anchored on a solid support. Several studies have shown that catalysts. We found that a Cu SAC supported on Mo2C and SACs can exhibit superior catalytic performance in thermocatalytic surrounded by O has a high cationic character in agreement with processes, such as selective hydrogenation (Wei et al., 2014; Wang the experiment (Zhou et al., 2021). We proposed feasible reaction et al., 2016), CO oxidation (Qiao et al., 2011; Lu et al., 2018), CO2 mechanisms for the CO2 hydrogenation and the RWGS reaction. conversion (Li et al., 2020a; Zhu et al., 2021), and water gas-shift In the present article, we address, using theory, that is, DFT (WGS) and RWGS reactions, C–C coupling, and electrocatalytic and calculations, the study of the RWGS reaction catalyzed by SACs photocatalytic processes (Wang et al., 2016; Mondelli et al., 2018; supported on TMC (Mo2C) with different surface O coverages, Kaiser et al., 2020) (Lin et al., 2013; Yang et al., 2015; Li et al., 2020b), particularly Cu- and Fe-based SACs. We focus on the CO2 activation with high activity, selectivity, metal atom utilization, and stability (Li and the H2O formation, which involve the adsorption of reactants, et al., 2016; Gao et al., 2021; Xiong et al., 2021). For example, Lin direct CO2 dissociation through the redox mechanism (CO2* → et al. (2013) synthesized Ir/FeOx SAC, having exceptionally high CO* + O*), H2 dissociation (H2* → H* + H*), and water formation activity for WGS, where the Ir center greatly enhanced the (2H* + O* → H2O*) (Alonso et al., 2021). reducibility of the FeOx support by generating oxygen vacancies, leading to the excellent catalytic performance. Currently, the development of SACs is a highly active research field (Zhang 2 Computational details et al., 2018). Other metal-based catalysts are also promising candidates for the RWGS reaction (Kim et al., 2015; Juneau We studied Fe’s and Cu’s catalytic performance supported on et al., 2020). Nevertheless, they have drawbacks, that is, their Mo2CTx with different oxygen coverages, which we denote as Fe/ poor natural abundance and high cost. Other alternative Mo2C and Cu/Mo2C, respectively. Spin-polarized density functional materials have also been considered as possible catalysts (Kim theory was used for the energetics as implemented in the Vienna Ab et al., 2015). In this context, MXene materials, a family of two- initio Simulation Package (VASP) (Kresse and Hafner, 1993; Kresse dimensional (2D) carbides, nitrides, and carbonitrides with the and Furthmüller, 1996a; Kresse and Furthmüller, 1996b). We used general formula of Mn+1XnTx (where M is an early transition the BEEF-vdW (Wellendorff et al., 2012) as the exchange- metal; n = 1, 2, and 3; X is C; and/or N and T are surface –O–, correlation functional and projected-augmented wave (PAW)- –OH, and/or –F groups), are currently emerging in thermocatalytic based pseudopotentials for all calculations. A plane-wave basis set applications as catalysts or supports with reactive metal–support with the kinetic energy cutoff of 500 eV was employed to expand the interactions (Li et al., 2018a; Li et al., 2018b; Diao et al., 2018; Zhao wave functions. We set the convergence criteria for minima et al., 2019; Kurlov et al., 2020). As a member of MXene materials, calculations to have a lower force than 0.02 eV/A. A vacuum transition metal carbides (TMCs) have attracted particular attention layer of 10 Å, which is perpendicular to the surface of Fe/Mo2C (Reddy et al., 2019; Lin et al., 2021) as they are cheap, potentially and Cu/Mo2C, was added to avoid spurious interactions between selective, and efficient catalysts. periodic images. For gas-phase calculations of molecules, we TMCs have similar properties as precious metals (Zhang et al., employed a cubic supercell of 15 Å × 15 Å × 15 Å. We included 2020; Morales-Salvador et al., 2021), being active in many reactions, dipole corrections along the z-direction due to the asymmetry of the such as CO hydrogenation, water–gas shift (WGS), hydrogen M/Mo2C surface with co-adsorbed oxygen atoms. We used nudged- evolution reaction (HER), oxygen reduction reaction (ORR), elastic band (NEB) methods to locate the transition states until the methanol oxidation reaction, and methane reforming (Chen atomic forces were less than 0.05 eV/Å. Finally, we constructed the et al., 2016; Wang et al., 2020). As a key member of TMCs, energy profile for the RWGS for all evaluated systems referencing all Mo2C is particularly interesting for CO2 conversion because of minima and transition states against the sum of energies of the given its low cost, dual functionality for H2 dissociation, and C=O evaluated catalyst and initial reactants (CO2 and H2) as the origin of bond scission capability (Porosoff et al., 2014). Many studies energies. have shown that Mo2C is highly active in activating CO2 in various processes, especially for RWGS reactions. Therefore, combining enriched SACs with Mo2C as support is an appealing 3 Results and discussion way to balance catalytic activity, selectivity, and stability effectively (Wang et al., 2022). Theoretical calculations can provide detailed We selected our former model for the Cu/Mo2CTx system insights into the energetics of the catalytic processes (Geiger and (Zhou et al., 2021), hereafter Cu/Mo2C, to evaluate their activity López, 2022). The catalytic cycle, energy barriers catalytic sites, and toward the RWGS reaction for different oxygen coverages (O obtained structure–reactivity relationships of each elementary step ML); 0, 0.33, 0.67, and 0.78, and performed an analogous study can be calculated using DFT-based methods with a good for the hypothetical Fe/Mo2C one. Figure 1 shows the structures compromise between accuracy and computational cost. Mo2C of Fe/Mo2C and Cu/Mo2C. CO2 and H2 adsorption minima close has also been used as a catalyst as the oxygen coverage was a key in energy were considered as initial structures. From their most aspect determining the catalytic activity of the material toward the stable configurations, other minimums and transition states were Frontiers in Chemistry 02 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 We split the RWGS reaction into two key steps, namely, CO2 activation (CO*+O*) and water formation (H2O*). Concerning CO2 activation, hydrogen-assisted routes via formate (HCOO*) and carbonyl (COOH*) are an alternative to direct CO2 activation. Nevertheless, forming HCOO* and COOH* species for both evaluated catalysts is more demanding than just directly splitting CO2 (see Figure 3). Thus, assessing the subsequent C–O bond cleavage of HCOO* and COOH* is not needed to conclude that the redox pathway by direct activation of CO2 is preferred over the hydrogen-assisted routes. We will now describe in detail the CO2 activation step. First, the CO2 molecule adsorbs on the metal atom (M = Cu, Fe)/Mo2C interface, forming a δ–CO2* intermediate. Subsequent CO2 pre-activation is exoenergetic or slightly endoenergetic, depending on the oxygen coverage. In the resulting structure, the carbon atom and one oxygen atom from CO2 carbon bind FIGURE 1 directly to the metal center, while the second oxygen of CO2 coordinates to Optimized structures for M (Cu and Fe)/Mo2C catalytic systems a Mo atom. From the δ–CO2* structure, CO2 can split via TS1 into CO* and four adsorption sites, namely, top (T), bridge (B), and hollow sites and O* in an endoenergetic step for all the cases. This transition state allows at X atom (Hx) or metal (Hm). activating CO2 and cleaving one of the C–O bonds. Oxygen bonds on the highly oxophilic Mo-hollow sites, while CO* remains coordinated to the metal center (M = Cu/Fe). The next step we evaluated is the desorption of localized. For each adsorbate (CO2, CO, O, and H), four high- the CO molecule to the gas phase. This step is endoenergetic for all cases. symmetry sites were explored (Figure 1), namely, top (T), bridge Given the high temperature of the RWGS reaction (200°C–500 °C for (B), and two types of threefold hollow sites, either with an X atom maximum conversion of CO2 ranging from 10% to 50%) (Porosoff et al., (Hx) or a metal (Hm) atom beneath. The adsorption energy for 2016), both the CO2 cleavage and the CO* desorption seem feasible at both each adsorbate (CO2, CO, O, and H) on each site for both systems the kinetic and thermodynamic levels. is provided in Supplementary Table S1 of the Supporting The second part of the mechanism corresponds to the H2O Information. molecule formation. This process is endoenergetic in all cases. This pathway starts with the adsorption of the H2 molecule on the surface, which is exoenergetic in all cases. Next, the H–H bond cleaves (TS2), 3.1 Description of the reverse water–gas giving rise to a proton (H+) and a hydride ion (H−). The latter transition shift reaction mechanism state can be understood as a heterolytic TS, producing formally a metal hydride (M–H) and the proton bonded to the cleaved O*. The resulting We studied the RWGS reaction catalyzed by the M(Cu and formal metal hydride remains at the interface (H*–M/Mo2C) and the Fe)/Mo2C system, which we previously evaluated for a 0.67 O ML hydroxy group on a Mo-hollow site (HO*–Mo). The subsequent coverage as a side reaction of the CO2 hydrogenation to methanol migration of the H* to the OH* group and the O–H bond reaction for the Cu/Mo2C system (Zhou et al., 2021; Geiger and formation to produce the H2O* molecule has a high energy barrier López, 2022). In the present work, we systematically assess the (TS3). This transition state is the most energy demanding along the oxygen coverage’s effect on the energetics of the RWGS reaction energy profile for all the evaluated oxygen coverages. After forming for Cu and Fe SACs supported on Mo2C. Thus, we first evaluated H2O*, its desorption is endoenergetic for all systems. the clean M/Mo2C system, that is, without oxygen being adsorbed, and 0.33, 0.67, and 0.78 O ML systems. We optimized the system’s minima and transition states to 3.2 Fe/Mo2C system evaluate the RWGS mechanism for both catalysts at several oxygen coverages to assess the latter’s effect and compare the The energy profiles of the RWGS catalyzed by Fe/Mo2C with intrinsic activity of Fe and Cu on the RWGS activity (Figure 2). 0 ML, 0.33 ML, 0.67 ML, and 0.78 ML surface oxygen coverages are FIGURE 2 Top and side views of the structure of M/Mo2C catalyst (M = Cu and Fe) with (A) 0 ML, (B) 0.33 ML, (C) 0.67 ML, and (D) 0.78 ML surface oxygen coverages. Gray, cyan, red, and bronze spheres indicate carbon (C), molybdenum (Mo), oxygen (O), and metal (Cu and Fe) atoms, respectively. Frontiers in Chemistry 03 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 FIGURE 3 Energy profiles for the Cu/Mo2C (A) and Fe/Mo2C (B) systems comparing the direct CO2 cleavage (CO + O), that is, the redox pathway and the formation of formate (HCOO) and carbonyl (COOH) intermediates. The energy barriers are reported within each profile (black indicates the lowest pathway for each system). The optimized structures for the transition state (TS) and final state (FS) structures are shown next to the corresponding energy profiles. Energies are referenced against the sum of the initial reactants’ energy in kJ mol–1 (Erel). shown in Figure 4. Table 1 summarizes the energy barriers for all Supplementary Table S2 reports the energy difference involved in evaluated steps. both Fe displacements). The 0 ML coverage has the most favorable The adsorption of CO2 is exoenergetic for all coverages, by 176.6, adsorption energy, confirming a more significant interaction 131.2, 91.5, and 88.5 kJ mol–1 going from lower to higher oxygen between CO2 with Mo sites and the iron center. Among all the coverages. All adsorbed CO2 molecules have a bent structure, in coverages evaluated, 0.67 ML has a lower CO release energy, but which one oxygen is bound to the Fe center, while the other one is 0.78 ML allows a better rate of CO formation due to its affordable bound to a Mo-top site, and the carbon atom is bonded to the Fe and reaction barrier and moderate releasing energy. the two Mo-top sites, as shown in the initial state (IS) on Figure 5. The H2O formation starts via H2 adsorption, which is exoenergetic The oxygen coverage effect can explain the energy differences in in all cases. The adsorption energies are 63.2, 56.4, 9.8, and 37.3 kJ mol–1 CO2 pre-activation. On the one hand, lower oxygen coverage means for oxygen coverages equal to 0, 0.33, 0.67, and 0.78 ML, respectively. less repulsion between the adsorbed surface species and the catalytic Again, the 0 O ML system has the most favorable adsorption energy, in system. Thus, higher CO2 adsorption energies are obtained for the which the location of the H2 molecule coordinates to Fe but is closer to 0 and 0.33 O ML systems shown in Table 1. Overall, the δ–CO2* the co-adsorbed oxygen coming from the CO2 activation than the other intermediate is more stable when decreasing the oxygen coverage. oxygen coverages just above the Fe, and the co-adsorbed oxygen favors On the other hand, the resulting bent CO2 angles of the resulting the adsorption. It means that the more in the middle it is, the better the intermediate (∠O–C–O) are 122° (0 ML), 132.3° (0.33 ML), 136.1° energy absorption. The energy barriers for the subsequent heterolytic (0.67 ML), and 136.4° (0.78 ML), confirming the relationship H2 cleavage (TS2, Figure 6) are equal to 73.2, 53.4, 64.7, and between the bending angle and the energy gain upon adsorbing 69.8 kJ mol–1 for 0, 0.33, 0.67, and 0.78 ML coverages, respectively. CO2 on the catalytic surface. The H2 cleavage forms an OH* species adsorbed on a Mo-hollow site The subsequent CO2 cleavage step (TS1) has energy barriers and a metal hydride intermediate (H*–M + OH*–Mo, Figure 6). equal to 68.7, 73.6, 31.5, and 27.4 kJ mol–1, from lower to higher Finally, H2O forms by reaction of OH* and H* with energy oxygen coverages. These energy barriers are related to the δ–CO2 barriers equal to 147.7, 100.4, 53.9, and 66.6 kJ mol−1 for oxygen pre-activation and stability. The energy barriers for CO2 cleavage coverages of 0, 0.33, 0.67, and 0.78 ML, respectively. TS3 geometries slightly decrease when increasing the ∠O–C–O angle and decreasing differ only in the proximity of the H2O* formed to the Mo-hollow the energy stability of the δ–CO2* intermediate. Overall, reaction and in the migration step of the H* atom from the interface to the energy differences are exoenergetic, so all catalysts are favorable for OH* group (TS3, Figure 6). The position of the OH* group, the the formation of CO, as shown in Table 1. The desorption of the migration site of the H* atom, and the bond lengths Fe–H and resulting CO* species on Fe (FS; Figure 5) is endoenergetic in all H–OH on the TS3 differ depending on the oxygen coverage, as is cases, that is, by 148.7, 166.9, 93.8, and 120.7 kJ mol–1 (see Figure 4). summarized in Table 2. For the first two coverages (0 and 0.33 ML), the energy for the CO* For the clean Fe/Mo2C system, the H* atom comes from the desorption includes slight Fe movement (Supporting Information interface, while the OH* group has more available adsorption sites as Frontiers in Chemistry 04 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 FIGURE 4 Energy profile of the RWGS reaction on Fe/Mo2C with different O ML coverages; black, blue, cyan, and green lines indicate surface oxygen coverages of 0, 0.33, 0.67, and 0.78, respectively. The intermediate and transition state structures and their energies are shown on the right for the 0.67 O ML case. Energies are referenced against the sum of the initial reactants’ energy in kJ mol–1 (Erel). there is no oxygen around. At 0.33 O ML, the H* atom goes to Mo- other coverages, in a very exothermic adsorption step of hollow because it is available in the absence of further oxygen 121.9 kJ mol–1, as mentioned earlier (CO2*–Cu, Figure 8). In this coverage. Finally, for 0.67 and 0.78 ML, there are one or no structure, CO2 bends the most, with an ∠O–C–O angle equal to longer any adsorption sites available, and consequently, the H* 121.4°. For the rest coverages (0.33, 0.67, and 0.78 ML), the carbon atom moves to the metal atom instead of the Fe/Mo2C interface and one oxygen atom of CO2 are bonded to the Cu center, and the before forming H2O*. These newly obtained minima, in which H* is oxygen of CO2 is connected to the top Mo site (shown as IS in solely bonded to the metal center, are shown in Figure 7. In these Figure 9). final cases, the OH* group remains bonded to the Mo-hollow The energy barriers for the subsequent CO2 splitting (TS1) are adsorption site. Structures of initial state (IS), transition state 105.8, 32.4, 2.4, and 4.7 kJ mol–1. As found for the Fe/Mo2C system (TS), intermediate, and final state (FS) are shown in Figure 6. (vide supra), a more stable δ–CO2 intermediate implies a high energy The H2O* structure at 0.67 and 0.78 O ML, that is, with the high barrier for CO2 cleavage; that is, a high oxygen coverage favors CO2 oxygen coverage, is more stable due to the formation of two hydrogen activation. Overall, the reaction energies are exoenergetic, so all catalysts bonds between H2O* and a co-adsorbed oxygen atom (FS, Figure 6). favorably form CO, as shown in Table 3. Once CO* is obtained, the CO* The desorption of H2O is endoenergetic for all systems by 66.5, 76.7, desorption is endoenergetic in all the cases, arising from the strong bond 92.3, and 93.1 kJ mol–1, from lower to higher oxygen coverages. High between CO* and the Cu atom, as we can see in the optimized oxygen coverages, likely present under reaction conditions, provide minimum CO*–Cu + O*–Mo shown in Figure 8 and as FS in the most feasible energy barriers, and therefore, high catalytic activity Figure 9. The 0 ML coverage has the most significant adsorption is expected. In contrast, for low oxygen coverages, the strong energy and the highest reaction barrier for CO2 activation, adsorption of the intermediates increases the key energy barriers suggesting a more substantial interaction between the CO2 and the for the RWGS reaction, suggesting a lower catalytic activity. The most catalyst increases the energy barrier. In contrast, 0.78 ML has the active system along the evaluated series is the Fe/Mo2C surface with a highest CO desorption energy (83.2 kJ mol-1), but all oxygen 0.67 O ML coverage as it presents the lowest energy barrier, with the coverages present CO desorption values within 70.9–83.2 kJ mol–1. highest energy barrier being the H2O formation step, amounting to When the oxygen coverage is equal to 0.67 O ML, the lowest energy 53.9 kJ mol-1. barrier toward CO* + O* is obtained: 2.4 mol–1. After CO desorbs, the adsorption of H2 is exoenergetic by 3.7, 7.9, 13.9, and 10.8 kJ mol–1, from lower to higher oxygen coverages. Next, H2 splits in a heterolytic 3.3 Cu/Mo2C system way. The hydride ion (H−) remains on the Cu/Mo2C interface, while the proton (H+) bonds to the O* atom arising from the CO2 cleavage, Figure 8 shows the complete energy profile of the RWGS forming an OH* group bonded to Mo, as shown in Figure 10. The catalyzed by the Cu/Mo2C with 0 ML, 0.33 ML, 0.67 ML, and reaction barriers for H2 splitting (heterolytic TS2, Figure 10) are 83.2, 0.78 ML oxygen coverages. The energy barriers of each step are 69.3, 73.1, and 68.8 kJ mol–1 for 0 ML, 0.33 ML, 0.67 ML, and 0.78 ML, summarized in Table 3. respectively. We can observe that when the H2 adsorption is higher, the The CO2 adsorption is exoenergetic by 121.9, 25.9, and energy barrier for the H–H bond cleavage decreases (Table 3). 7.6 kJ mol–1 for 0 , 0.33 , and 0.78 ML coverages, respectively, Finally, the energy barriers to forming H2O (TS3, Figure 10) are while for 0.67 ML, it is slightly endothermic, which is about 141.4, 144.1, 100.7, and 80.2 kJ mol–1 from lower to higher oxygen 5.1 kJ mol–1. On the Cu/Mo2C catalyst, a low oxygen coverage coverages. This step has the highest energy barriers for all evaluated decreases the repulsion of the adsorbed CO2 and therefore results oxygen coverages. The related transition states correspond to the in a more favorable CO2 adsorption energy. For Cu/Mo2C 0 ML, formation of the second O–H bond of water by the hydrogen CO2 binds mainly on Mo rather than on Cu in comparison to the transfer of the H* atom at the Cu/Mo2C interface to the OH* Frontiers in Chemistry 05 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 TABLE 1 Summary of energy values (adsorption, release, and overall reaction) H2O desorption steps are all endoenergetic by 66.8, 78.1, 65.2, and energy barriers (transition states) for each step on RWGS with the Fe/Mo2C and 89.1 kJ mol–1 from lower to higher oxygen coverages. These system in kJ mol−1. resulting products with the water molecule adsorbed are more stable Step Energy values Oxygen coverage (O ML) for high oxygen coverages (0.67 and 0.78 O ML) than for the lower (kJ mol−1) ones (0 and 0.33 O ML) due to the formation of hydrogen bonds 0 0.33 0.67 0.78 between water and the co-adsorbed oxygen atom (FS, Figure 10). CO δ–CO2 adsorption −176.6 −131.2 −91.5 −88.5 The Cu/Mo2C system shows that the more favorable the reaction energy, the lower the energy barrier for forming H2O. ∠O–C–O (°) 122.0 132.3 136.1 136.4 Overall, among all the RWGS catalyzed by Cu/Mo2C, the Reaction energy (CO) −21.9 −63.1 −61.5 −70.1 systems with high oxygen coverages have the lowest energy CO desorption 148.7 166.9 93.8 120.7 barriers for CO2 activation and H2O formation (0.67 and 0.78 O ML) compared to the systems with low oxygen H2O H2 adsorption −62.9 −56.1 −9.8 −37.1 coverages (0 and 0.33 O ML). Overall, the system with the Reaction energy 27.2 49.8 7.3 17.7 lowest energy barriers is the Cu/Mo 2C 0.78 O ML one, in (OH*+H*-Fe) which the highest energy barrier is 80.2 kJ mol-1, Reaction energy (H2O) 87.2 24.9 37.3 32.2 corresponding to the water formation step. H2O desorption 66.5 76.7 92.3 93.1 Step Energy barriers (kJ mol−1) Oxygen coverage (O ML) 3.4 Comparison of the RWGS catalytic 0 0.33 0.67 0.78 activity of Cu/Mo2C vs Fe/Mo2C CO CO2 splitting (TS1) 68.7 73.6 31.1 27.4 The discussion will be divided into two parts: one for 0 and H2O H2 splitting (TS2) 73.2 53.6 64.7 69.8 0.33 O ML coverages and the other for 0.67 and 0.78 O ML coverages, respectively. We first describe the results for the CO H2O formation (TS3) 147.7 100.4 53.9 66.6 formation with the 0/0.33 O ML systems. The adsorption of CO2 releases energy in all cases. When increasing the amount of co- adsorbed oxygen, the (∠O–C–O) angle and the adsorption energy group adsorbed to the Mo-hollow site. The geometries of TS3 are decrease; that is, it is less negative—from −121.9 to −25.9 kJ mol–1 similar for all coverages. The main differences observed are the for Cu/Mo2C and −176.6 to −131.2 kJ mol–1 for Fe/Mo2C. These position of the OH* group, the migration site of the H* atom, and values indicate that CO2 interaction is significantly stronger on Fe the distance length when migrating from the Cu/Mo2C interface to than on Cu on clean surfaces. All reaction energies are exoenergetic OH*. Table 4 summarizes the related position/migration and the for cleaving CO2 to CO* and O*. The energy barriers of CO2 Cu–H and O–H bond lengths for TS3. In this case, the bond length splitting for Cu are 105.8 and 32.4 kJ mol–1 for 0 and 0.33 O ML, of the H* atom to Cu of the interface allows a higher reactivity of this respectively, whereas, for Fe, they are equal to 68.7 and 73.6 kJ mol–1 hydrogen atom, resulting in a lower energy barrier. Structures of for 0 and 0.33 O ML, respectively. The Cu system presents an initial state (IS), transition state (TS), intermediate, and final state essential difference between both coverages since the 0.33 O ML (FS) are shown in Figure 10. coverage has a much lower energy barrier than the 0 ML one: FIGURE 5 Top and side views of the initial state (IS), transition state (TS), and final state (FS) for the CO2 cleavage catalyzed by the Fe/Mo2C of 0.78 ML system (see Supplementary Figures S1–S3 for the structures of other intermediates and TS on 0, 0.33, and 0.67 O* ML coverages). Frontiers in Chemistry 06 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 FIGURE 6 Top and side views of structures of initial state (IS), intermediate, transition state (TS), and final state (FS) on Fe/Mo2C of 0.78 O ML coverage for H2 splitting and the formation of H2O (see Supplementary Figures S4-6 for the structures of other intermediates and TS on 0, 0.33, and 0.67 O* ML coverages). TABLE 2 Summary of OH* position, H* migration’s site, and H–Fe and H–OH bond lengths for the H* atom migration step from the Fe/Mo2C interface to OH* involved in the H2O* formation step (TS3) of the RWGS catalyzed by the Fe/Mo2C system in Å. Surf.—Cover. (ML) OH* position H* migration site H–Fe (Å) H–OH (Å) 0 Mo-top Interface 1.85 1.37 0.33 Mo-top Mo-hollow 3.08 1.38 0.67 Mo-hollow Fe-top 1.95 1.29 0.78 Mo-hollow Fe-top 1.75 1.40 71.6 kJ mol-1). Here, a remarkable difference between both systems is that the energy required to desorb CO is much higher for Fe/Mo2C than for Cu/Mo2C, regardless of oxygen coverage. This difference means the CO binding energy is much stronger on Fe/ Mo2C than on Cu/Mo2C. However, at the high temperature of RWGS, desorption is favored entropically, so it should be feasible for both catalysts. Concerning H2 adsorption, it is more favorable on the Fe/Mo2C catalyst than on the Cu/Mo2C one. The Fe system has a maximum energy release of 62.9 kJ mol–1 per 0 ML. The splitting of H2 to OH*+H*–M is endothermic in Fe, while for Cu, it is exothermic. The reaction energy absorbed for Fe or released for Cu energy increases with coverage; 0.33 ML exhibits better OH* FIGURE 7 formation for both metals due to the lower energy barrier. The Top and side views of new minima were obtained where the H* atom remains at the top of the metal center involved in Fe/Mo2C second energy barrier for obtaining H2O, both Fe/Mo2C and Cu/ coverage of (A) 0.67 and (B) 0.78 ML for H2O* formation. Mo2C, shows the same trend, with the highest energy barriers and endothermic processes. Finally, H2O desorption requires similar adsorption energy values on both metal systems and coverages. Concerning the catalytic performance of Fe/Mo2C and Cu/Mo2C 105.8 vs 32 kJ mol-1. In contrast, the Fe system only shows a with surface oxygen coverages of 0.67 ML and 0.78 M O ML, the difference of 5 kJ mol-1 between 0 and 0.33 O ML. The Cu/Mo2C adsorption of CO2 is more favored for Fe/Mo2C than for Cu/Mo2C. The at 0.33 O ML is the most active system toward cleaving CO2. The energy released is less negative upon increasing oxygen content and the reaction energy of this step becomes more negative and, therefore, ∠O–C–O angle, which is consistent with the behavior from 0ML to more favorable upon increasing the surface oxygen coverage for 0.3 ML oxygen coverages. However, Cu/Mo2C has a slight endothermic both Fe/Mo2C and Cu/Mo2C. Nevertheless, the variation is more reaction at 0.67 ML instead of 0.78 ML, which is exothermic, and the significant for Cu/Mo2C, indicating that the presence of surface energy released is less than that of Fe/Mo2C with the same oxygen oxygen atoms substantially affects the Cu/Mo2C system more than coverage. Again, the overall reaction energy differences are exothermic, the Fe/Mo2C one. Finally, CO desorption is endothermic for both so the CO formation is thermodynamically favorable for all catalysts. Fe/Mo2C (148.6 and 167 kJ mol-1) and Cu/Mo2C (70.9 and The energy barriers for the CO2 splitting at 0.67 and 0.78 O ML are the Frontiers in Chemistry 07 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 FIGURE 8 Energy profile of RWGS reaction on Cu/Mo2C with different O ML coverage; black, red, green, and orange lines indicate surface oxygen coverages of 0, 0.33, 0.67, and 0.78, respectively. The intermediate and transition state structures and their energies are shown on the right for the 0.67 O ML case. Energies are referenced against the sum of the initial reactants’ energy in kJ mol–1 (Erel). TABLE 3 Summary of energy values (adsorption, release, and overall reaction) The energy barriers for H2O formation for the oxygen coverages equal and energy barriers (transition states) for each step on RWGS with Cu/Mo2C to 0.67 and 0.78 ML are lower for the Fe/Mo2C than for the Cu/Mo2C system in kJ mol−1. ones, ranging within 53.9–66.6 kJ mol-1 and 80.2–100.7 kJ mol-1, Step Energy values Oxygen coverage (O ML) respectively. Overall, Fe/Mo2C, with 0.67 O ML, has the lowest (kJ mol−1) energy barrier for forming water: 53.9 kJ mol-1. Finally, H2O 0 0.33 0.67 0.78 desorption requires higher adsorption energy values on Fe/Mo2C, CO δ–CO2 adsorption −121.9 −25.9 5.1 −7.6 indicating a strong interaction between H2O and iron instead of copper. In summary, for CO2 and H2 adsorption, Fe/Mo2C with 0 ML ∠O–C–O (°) 121.4 133.8 136 139 coverage is the most energy-releasing system upon the adsorption of Reaction energy (CO) −9.1 −83.2 −103.4 −110.9 the reactants. For CO formation, Cu/Mo2C 0.67 ML has the lowest CO desorption 70.9 71.6 79.9 83.2 energy barrier (TS1, 2.4 kJ mol–1), and it is among those with the lowest reaction energies (−103.4 kJ mol–1). Therefore, it is the system H2O H2 adsorption −3.7 −7.9 −13.9 −10.8 leading most easily to CO. The CO desorption values are higher for Fe Reaction energy −12.8 −28.6 28.7 36.6 catalysts than for Cu ones, indicating a higher interaction of CO on Fe (OH*+H*–Cu) than on Cu/Mo2C. Concerning H2 splitting, Fe/Mo2C 0.33 ML Reaction energy (H2O) 77.8 63.8 16.4 −11.4 presents the lowest energy barrier (TS2, 53.6 kJ mol–1) among all systems. However, the energy barriers for both systems and coverages H2O desorption 66.8 78.1 65.1 89.1 do not differ much, being all within 53.6–83.2 kJ mol-1. Fe/Mo2C Step Energy barriers (kJ mol−1) Oxygen coverage (O ML) 0.67 O ML is the system presenting the lowest energy barrier for forming H2O (TS3: 53.9 kJ mol–1), followed by the Fe/Mo2C 0.78 O 0 0.33 0.67 0.78 ML system (66.6 kJ mol-1). Conversely, the energy barriers for this CO CO2 splitting (TS1) 105.8 32.4 2.4 4.7 transition state (TS3) for 0 and 0.33 O ML for Fe (147.7 and H2O H2 splitting (TS2) 83.2 69.3 73.1 68.8 100.4 kJ mol-1, respectively) and Cu (141.1 and 144.1 kJ mol–1, respectively) present high energy barriers. According to the overall H2O formation (TS3) 141.4 144.1 100.7 80.2 analysis, the coverage of 0.67 O ML is the most effective one in catalyzing the formation of CO and the formation of H2O. While Cu more favorably forms CO and Fe H2O, the best Fe/Mo2C system lowest in both systems and are like each other, although they are lower (0.67 O ML) presents the lowest energy barriers. for the Cu/Mo2C systems (2.4–4.7 kJ mol-1) than for the Fe/Mo2C ones (27.4–31.1 kJ mol-1). For CO formation, Fe/Mo2C and Cu/Mo2C show the same energy trend as oxygen increases, with the highest reaction 4 Conclusion energy released for the oxygen coverage equal to 0.78 ML. Finally, higher energy was required in the desorption of CO 0.78 ML, with both Systematic DFT calculations were performed on Cu/Mo2C metals indicating better interaction with CO*. Concerning the and Fe/Mo2C catalysts to explore the effect of metal and different subsequent H2 adsorption, the Fe/Mo2C 0.78 O ML system releases oxygen coverages through the RWGS reaction. The study more energy than the Cu one. Fe and Cu systems have negligible energy indicates that both catalysts can pre-activate and cleave CO2, differences in the energy barriers for H2 splitting. The reaction energies heterolytically split H2, and form water by reacting with two to form OH become more endoenergetic as oxygen coverage increases. adsorbed hydrogen atoms, formally as a proton (H+) and a Frontiers in Chemistry 08 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 FIGURE 9 Top and side views of the initial state (IS), transition state (TS), and final state (FS) catalyzed by the Cu/Mo2C of 0.78 ML system for the CO2 cleavage (see Supplementary Figures S7-9 for the structures of other intermediates and TS on 0, 0.33, and 0.67 O* ML coverages). FIGURE 10 Top and side views of structures of the initial state (IS), intermediate, transition state (TS), and final state (FS) on Cu/Mo2C of 0.78 O ML coverage for H2 splitting and the formation of H2O (see Supplementary Figures S10-12 for the structures of other intermediates and TS on 0, 0.33, and 0.67 O* ML coverages). TABLE 4 Summary of OH* position, H* migration’s site, and H–Cu and H–OH bond lengths for the H* atom migration step from the Cu/Mo2C interface to OH* involved in the H2O* formation step (TS3) of the RWGS catalyzed by the Cu/Mo2C system in Å. Surf.—Cover. (ML) OH* position H* migration site H–Cu (Å) H–OH (Å) 0 Mo-bridge Interface 2.67 1.26 0.33 Mo-bridge Interface 1.99 1.30 0.67 Mo-hollow Interface 1.76 1.38 0.78 Mo-hollow Interface 1.75 1.36 hydride ion (H−). On the one hand, Cu/Mo2C showed more RWGS reactivity. On the other hand, H2 activation has similar feasible energy barriers for CO formation, while Fe/Mo2C trends for both metals and all oxygen coverages. In conclusion, presented more feasible energy barriers for H2O formation. the calculated energy barriers and reaction energies suggest that Overall, the most active Fe/Mo2C system, having oxygen the Fe/Mo2C 0.67 O ML catalyst has the potential for being a coverage equal to 0.67 O ML, presents lower energy barriers highly active RWGS catalyst, likely arising from the highly than the Cu/Mo2C system, suggesting that it is more active than oxophilic and positive character of Fe, in which the high the latter. The presence of O in the catalysts may explain the oxygen coverage balances the catalytic activity in agreement previously mentioned favorable trends for high coverages for with the Sabatier principle. Frontiers in Chemistry 09 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 Data availability statement Conflict of interest The raw data supporting the conclusions of this article will be The authors declare that the research was conducted in the made available by the authors, without undue reservation. absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Author contributions Publisher’s note AC-V conceived the project. WZ, AV-L, and AC-V planned the research. WZ and AV-L performed the computational calculations, All claims expressed in this article are solely those of the authors and AC-V supervised them. WZ and AV-L wrote the first draft of and do not necessarily represent those of their affiliated the manuscript. All authors contributed to the manuscript, and read organizations, or those of the publisher, the editors, and the and approved the submitted version. reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Funding The authors thank the Spanish “Ministerio de Ciencia e Supplementary material Innovación” for funding the “I + D Generación del Conocimiento” project (PID 2021-128416NB-I00) and The Supplementary Material for this article can be found online acknowledge the grant to AV-L (PRE 2019-089647). WZ thanks at: https://www.frontiersin.org/articles/10.3389/fchem.2023.1144189/ the support of the China Scholarship Council (CSC). full#supplementary-material References Alonso, G., López, E., Huarte-Larrañaga, F., Sayós, R., Prats, H., and Gamallo, P. Kurlov, A., Deeva, E. B., Abdala, P. M., Lebedev, D., Tsoukalou, A., Comas-Vives, A., (2021). Zeolite-encapsulated single-atom catalysts for efficient CO2 conversion. et al. (2020). Exploiting two-dimensional morphology of molybdenum oxycarbide to J. CO2 Util. 54, 101777. doi:10.1016/j.jcou.2021.101777 enable efficient catalytic dry reforming of methane. Nat. Commun. 11 (1), 4920. doi:10. 1038/s41467-020-18721-0 Chen, Y., Choi, S., and Thompson, L. T. (2016). Low temperature CO2 hydrogenation to alcohols and hydrocarbons over Mo2C supported metal catalysts. J. Catal. 343, Li, X., Bi, W., Zhang, L., Tao, S., Chu, W., Zhang, Q., et al. (2016). Single-atom Pt as 147–156. doi:10.1016/j.jcat.2016.01.016 Co-catalyst for enhanced photocatalytic H 2 evolution. Adv. Mat. 28 (12), 2427–2431. doi:10.1002/adma.201505281 Diao, J., Hu, M., Lian, Z., Li, Z., Zhang, H., Huang, F., et al. (2018). Ti3C2Tx MXene catalyzed ethylbenzene dehydrogenation: Active sites and mechanism exploration from Li, Z., Cui, Y., Wu, Z., Milligan, C., Zhou, L., Mitchell, G., et al. (2018). Reactive both experimental and theoretical aspects. ACS Catal. 8 (11), 10051–10057. doi:10. metal–support interactions at moderate temperature in two-dimensional niobium- 1021/acscatal.8b02002 carbide-supported platinum catalysts. Nat. Catal. 1 (5), 349–355. doi:10.1038/s41929- 018-0067-8 Gao, Z., Meng, Y., Shen, H., Xie, B., Ni, Z., and Xia, S. (2021). Theoretical study on low temperature reverse water gas shift (RWGS) mechanism on monatomic transition metal M Li, Z., Yu, L., Milligan, C., Ma, T., Zhou, L., Cui, Y., et al. Two-dimensional transition doped C2N catalyst (M=Cu, Co, Fe). Mol. Catal. 516, 111992. doi:10.1016/j.mcat.2021.111992 metal carbides as supports for tuning the Chemistry of catalytic nanoparticles. Nat. Commun. 2018, 9 (1), 5258.doi:10.1038/s41467-018-07502-5 Geiger, J., and López, N. (2022). Coupling metal and support redox terms in single- atom catalysts. J. Phys. Chem. C 126 (32), 13698–13704. doi:10.1021/acs.jpcc.2c03710 Li, M., Wang, H., Luo, W., Sherrell, P. C., Chen, J., and Yang, J. (2020). Heterogeneous single-atom catalysts for electrochemical CO 2 reduction reaction. Adv. Mat. 32 (34), Guharoy, U., Ramirez Reina, T., Gu, S., and Cai, Q. (2019). Mechanistic insights into 2001848. doi:10.1002/adma.202001848 selective CO2 conversion via RWGS on transition metal phosphides: A DFT study. J. Phys. Chem. C 123 (37), 22918–22931. doi:10.1021/acs.jpcc.9b04122 Li, L., Chang, X., Lin, X., Zhao, Z.-J., and Gong, J. (2020). Theoretical insights into single-atom catalysts. Chem. Soc. Rev. 49 (22), 8156–8178. doi:10.1039/d0cs00795a Jing, H., Li, Q., Wang, J., Liu, D., and Wu, K. (2019). Theoretical study of the reverse water gas shift reaction on copper modified β-Mo2C(001) surfaces. J. Phys. Chem. C 123 Lim, X. (2015). How to make the most of carbon dioxide. Nature 526 (7575), 628–630. (2), 1235–1251. doi:10.1021/acs.jpcc.8b09884 doi:10.1038/526628a Juneau, M., Vonglis, M., Hartvigsen, J., Frost, L., Bayerl, D., Dixit, M., et al. (2020). Assessing Lin, J., Wang, A., Qiao, B., Liu, X., Yang, X., Wang, X., et al. (2013). Remarkable the viability of K-Mo 2 C for reverse water–gas shift scale-up: Molecular to laboratory to pilot performance of Ir 1/FeO x single-atom catalyst in water gas shift reaction. J. Am. Chem. scale. Energy Environ. Sci. 13 (8), 2524–2539. doi:10.1039/d0ee01457e Soc. 135, 15314–15317. doi:10.1021/ja408574m Kaiser, S. K., Chen, Z., Faust Akl, D., Mitchell, S., and Pérez-Ramírez, J. (2020). Single- Lin, Z., Denny, S. R., and Chen, J. G. (2021). Transition metal carbides and nitrides as atom catalysts across the periodic table. Chem. Rev. 120 (21), 11703–11809. doi:10.1021/ catalysts for thermochemical reactions. J. Catal. 404, 929–942. doi:10.1016/j.jcat.2021. acs.chemrev.0c00576 06.022 Karl, T. R., and Trenberth, K. E. (2003). Modern global climate change. Science 302 Lu, Y., Wang, J., Yu, L., Kovarik, L., Zhang, X., Hoffman, A. S., et al. (2018). (5651), 1719–1723. doi:10.1126/science.1090228 Identification of the active complex for CO oxidation over single-atom Ir-on- MgAl2O4 catalysts. Nat. Catal. 2 (2), 149–156. doi:10.1038/s41929-018- Kim, D. H., Han, S. W., Yoon, H. S., and Kim, Y. D. (2015). Reverse water gas shift 0192-4 reaction catalyzed by Fe nanoparticles with high catalytic activity and stability. J. Ind. Eng. Chem. 23, 67–71. doi:10.1016/j.jiec.2014.07.043 Mondelli, C., Puértolas, B., Ackermann, M., Chen, Z., and Pérez-Ramírez, J. (2018). Enhanced base-free formic acid production from CO 2 on Pd/g-C 3 N 4 by tuning of the Kresse, G., and Furthmüller, J. (1996). Efficient iterative schemes for ab initio total- carrier defects. ChemSusChem 11 (17), 2859–2869. doi:10.1002/cssc.201801362 energy calculations using a plane-wave basis set. Phys. Rev. B 54 (16), 11169–11186. doi:10.1103/physrevb.54.11169 Morales-Salvador, R., Gouveia, J. D., Morales-García, Á., Viñes, F., Gomes, J. R. B., and Illas, F. (2021). Carbon capture and usage by MXenes. ACS Catal. 11 (17), Kresse, G., and Furthmüller, J. (1996). Efficiency of ab-initio total energy calculations 11248–11255. doi:10.1021/acscatal.1c02663 for metals and semiconductors using a plane-wave basis set. Comput. Mat. Sci. 6 (1), 15–50. doi:10.1016/0927-0256(96)00008-0 Olah, G. A., Prakash, G. K. S., and Goeppert, A. (2011). Anthropogenic chemical carbon cycle for a sustainable future. J. Am. Chem. Soc. 133 (33), 12881–12898. doi:10. Kresse, G., and Hafner, J. (1993). Ab initio molecular dynamics for liquid metals. Phys. 1021/ja202642y Rev. B 47 (1), 558–561. doi:10.1103/physrevb.47.558 Frontiers in Chemistry 10 frontiersin.org Zhang et al. 10.3389/fchem.2023.1144189 Porosoff, M. D., Yan, B., and Chen, J. G. (2016). Catalytic reduction of CO 2 by H 2 for Wei, H., Liu, X., Wang, A., Zhang, L., Qiao, B., Yang, X., et al. (2014). FeOx-supported synthesis of CO, methanol and hydrocarbons: Challenges and opportunities. Energy platinum single-atom and pseudo-single-atom catalysts for chemoselective hydrogenation Environ. Sci. 9 (1), 62–73. doi:10.1039/c5ee02657a of functionalized nitroarenes. Nat. Commun. 5 (1), 5634. doi:10.1038/ncomms6634 Porosoff, M. D., Yang, X., Boscoboinik, J. A., and Chen, J. G. (2014). Wellendorff, J., Lundgaard, K. T., Møgelhøj, A., Petzold, V., Landis, D. D., Nørskov, Molybdenum carbide as alternative catalysts to precious metals for highly J. K., et al. (2012). Density functionals for surface science: Exchange-correlation model selective reduction of CO 2 to CO. Angew. Chem. 126 (26), 6823–6827. doi:10. development with bayesian error estimation. Phys. Rev. B 85 (23), 235149. doi:10.1103/ 1021/anie.201404109 physrevb.85.235149 Qiao, B., Wang, A., Yang, X., Allard, L. F., Jiang, Z., Cui, Y., et al. (2011). Single-atom Xiong, H., Datye, A. K., and Wang, Y. (2021). Thermally stable single-atom catalysis of CO oxidation using Pt1/FeOx. Nat. Chem. 3 (8), 634–641. doi:10.1038/ heterogeneous catalysts. Adv. Mat. 33 (50), 2004319. doi:10.1002/adma.202004319 nchem.1095 Yang, M., Liu, J., Lee, S., Zugic, B., Huang, J., Allard, L. F., et al. (2015). A common Reddy, K. P., Dama, S., Mhamane, N. B., Ghosalya, M. K., Raja, T., Satyanarayana, C. single-site Pt(II)–O(OH) x – species stabilized by sodium on “active” and “inert” V., et al. (2019). Molybdenum carbide catalyst for the reduction of CO 2 to CO: Surface supports catalyzes the water-gas shift reaction. J. Am. Chem. Soc. 137 (10), 3470–3473. science aspects by NAPPES and catalysis studies. Dalton Trans. 48 (32), 12199–12209. doi:10.1021/ja513292k doi:10.1039/c9dt01774g Zhang, H., Liu, G., Shi, L., and Ye, J. (2018). Single-atom catalysts: Emerging Rodriguez, J. A., Liu, P., Stacchiola, D. J., Senanayake, S. D., White, M. G., and Chen, multifunctional materials in heterogeneous catalysis. Adv. Energy Mat. 8 (1), J. G. (2015). Hydrogenation of CO2 to methanol: Importance of metal–oxide and 1701343. doi:10.1002/aenm.201701343 metal–carbide interfaces in the activation of CO2. ACS Catal. 5 (11), 6696–6706. doi:10. Zhang, Q., Pastor-Pérez, L., Jin, W., Gu, S., and Reina, T. R. (2019). Understanding the 1021/acscatal.5b01755 promoter effect of Cu and Cs over highly effective β-Mo2C catalysts for the reverse Wang, W., Wang, S., Ma, X., and Gong, J. (2011). Recent advances in catalytic water-gas shift reaction. Appl. Catal. B Environ. 244, 889–898. doi:10.1016/j.apcatb. hydrogenation of carbon dioxide. Chem. Soc. Rev. 40 (7), 3703. doi:10.1039/ 2018.12.023 c1cs15008a Zhang, Q., Pastor-Pérez, L., Gu, S., and Ramirez Reina, T. (2020). Transition metal Wang, L., Zhang, W., Wang, S., Gao, Z., Luo, Z., Wang, X., et al. (2016). Atomic-level carbides (TMCs) catalysts for gas phase CO2 upgrading reactions: A comprehensive insights in optimizing reaction paths for hydroformylation reaction over Rh/CoO overview. Catalysts 10 (9), 955. doi:10.3390/catal10090955 single-atom catalyst. Nat. Commun. 7 (1), 14036. doi:10.1038/ncomms14036 Zhao, D., Chen, Z., Yang, W., Liu, S., Zhang, X., Yu, Y., et al. (2019). MXene (Ti3C2) Wang, F., Li, T., Shi, Y., and Jiao, H. (2020). Molybdenum carbide supported vacancy-confined single-atom catalyst for efficient functionalization of CO2. J. Am. metal catalysts (M n/Mo x C; M = Co, Ni, Cu, Pd, Pt) – metal and surface Chem. Soc. 141 (9), 4086–4093. doi:10.1021/jacs.8b13579 dependent structure and stability. Catal. Sci. Technol. 10 (9), 3029–3046. doi:10. Zhou, H., Chen, Z., López, A. V., López, E. D., Lam, E., Tsoukalou, A., et al. (2021). 1039/d0cy00504e Engineering the Cu/Mo2CTx (MXene) interface to drive CO2 hydrogenation to Wang, X., Zhang, Y., Wu, J., Zhang, Z., Liao, Q., Kang, Z., et al. (2022). Single-atom methanol. Nat. Catal. 4 (10), 860–871. doi:10.1038/s41929-021-00684-0 engineering to ignite 2D transition metal dichalcogenide based catalysis: Zhu, Y., Yang, X., Peng, C., Priest, C., Mei, Y., and Wu, G. (2021). Carbon-supported Fundamentals, progress, and beyond. Chem. Rev. 122 (1), 1273–1348. doi:10. single metal site catalysts for electrochemical CO 2 reduction to CO and beyond. Small 1021/acs.chemrev.1c00505 17 (16), 2005148. doi:10.1002/smll.202005148 Frontiers in Chemistry 11 frontiersin.org