Unlocking the potential of p-block single-atom anchored on the MXene electrocatalyst surface for efficient CO₂ reduction

Abstract

The electrochemical reduction of CO₂ into value-added products has emerged as a promising approach for mitigating CO₂ emissions. In this study, 23 p-block single-atom (PSA) anchored onto the Mo₂CO₂ catalyst for CO₂ reduction have been systematically investigated using density functional theory at the atomic level. Based on the binding energy and cohesive energy, 9 PSA prefer to anchor onto the Mo₂CO₂ hollow carbon site. Side-on and end-on modes are preferred for CO₂ adsorption on PSA anchored Mo₂CO₂ (PSA@Mo₂CO₂). Projected density of states (PDOS) analysis indicates that PSA@Mo₂CO₂ exhibits a metallic-like electronic structure. The Bader charge analysis and charge density difference show unique behavior for Sn@Mo₂CO₂, with a lower Gibbs free energy change for the potential-determining step, CO₂ to *OCHO (0.58 eV). Sn@Mo₂CO₂ is located on top of the volcano plot of limiting potential versus adsorption energy. Furthermore, Sn@Mo₂CO₂ exhibits the highest selectivity for CO₂ reduction to HCOOH and suppresses the competing hydrogen evolution reaction. PDOS analysis of the *OCHO intermediate reveals that the oxygen and Sn p orbitals show moderate overlap. Ab initio molecular dynamics simulations indicate that Sn@Mo₂CO₂ is stable at 300 K. This work provides an orbital-based strategy for catalyst design to enable selective CO₂ reduction to HCOOH.

---

1. Introduction

In this technological and industrial era, energy consumption is increasing rapidly and continuously. Currently, the primary energy sources depend on fossil fuels. As a result, an enormous amount of CO₂ is continually emitted into the atmosphere, which significantly contributes to global warming and has an adverse impact on the climate.¹ The Intergovernmental Panel on Climate Change (IPCC) has predicted that a 1.5 °C temperature rise or more is expected to occur within the next 20 years. CO₂ is inert and thermally stable due to its linear geometry and C=O bonds.² To address the above problems, identifying a suitable method for converting CO₂ into value-added products is essential. Among all approaches, the electrochemical reduction method using a sustainable catalyst offers several advantages, including operating at room temperature, using recyclable electrolytes, and integrating with renewable energy sources. Additionally, a more straightforward setup can be easily scaled to an industrial level.³ Metal oxides, transition metal dichalcogenides, graphenes, metal alloys, h-BN, metal–organic frameworks, and noble metal-based catalysts have been explored as potential electrocatalysts for CO₂ reduction.^4–10 However, all these processes still present numerous challenges, including high activation energies, sluggish kinetics, high overpotential and poor selectivity.^11,12

Recently, single-atom catalysts (SACs) have garnered great interest due to their nearly 100% atom utilization efficiency, unique electronic properties, and low coordination of single atoms anchored on support materials.^11,13 The SAC-based catalysts have shown better activity and performance than their bulk counterparts in various reactions, such as the oxygen evolution reaction (OER), oxygen reduction reaction (ORR),¹⁴ nitrogen reduction reaction (NRR),¹⁵ methane activation,¹⁶ hydrogen evolution reaction (HER),¹⁷ and CO₂ reduction reaction (CRR).¹⁸ Since the single atoms are highly unstable, suitable support materials are required to stabilize them and enhance their activity. Considering this, metal oxides,¹⁹ graphenes,⁶ h-BN,²⁰ transition metal dichalcogenides,²¹ phosphorenes,²² and MXenes²³ are explored aspromising support materials.

Among all supports, MXene is currently the most investigated 2D nanomaterial with fascinating properties and applications and the M_n+1X_nT_x (n = 1 to 3) general formula, where M is an early transition metal, X represents C and/or N, and T_x represents the surface termination group, such as F, Cl, S, O, and OH.^24,25 MXene is a suitable support nanomaterial for single atoms due to its outstanding electrical conductivity, hydrophilicity, large surface area, and stability. The recently discovered MXene support is attracting attention for its exceptional properties.^26–28 The strong interaction of a single atom with MXene results in modified electronic states and an optimal active site for the adsorption of intermediates for better catalytic activity.^29–31 While considering suitable single atoms, researchers have mainly focused on transition-metal based atoms for a wide range of reactions due to their electron-accepting or donating properties toward adsorbates, arising from their partially filled d-orbitals.^23,25,32 In contrast, the p-block elements have filled d-orbitals and partially occupied p-orbitals, making them alternative candidates for a single-atom based active site for the desired reactions.³¹ However, few studies have been performed on p-block single-atom (PSA) for the HER and CRR. Zhang et al.³³ investigated the behavior of 14 non-metal doped Mo₂CO₂ catalysts toward the HER. In this work, surface oxygen gets activated by doping a non-metal in Mo₂CO₂. The calculated Gibbs free energies for hydrogen reveal that Cl (−0.06 eV), Br (−0.01 eV), and I (0.04 eV) doped Mo₂CO₂ show outstanding electrocatalytic activity towards the HER. Ren et al.²¹ designed a p-block (In, Bi, Pb, Sn, and Sb) metal-based single atom anchored on a MoS₂ support for CO₂ reduction. Out of all PSA, only Sb and Bi show an excellent catalytic activity towards formic acid formation, with the limiting potential (U_L) value for Bi and Sb being −0.46 and −0.31 V, respectively. Chen et al.³⁴ considered single and double atoms based on metals from the d and p-block elements on SnS₂, and it was found that the Sn-based single-atom catalyst outperforms all the considered catalysts in the formation of formic acid. Guo et al.³¹ constructed a SnN₄ single-atom active site with different secondary coordination sphere atoms, such as P, S, B, and N, for modulating the activity of the Sn single atom. Among all, Sn–N₃S₁ was found to improve intermediate adsorption and activity. Liu et al.³⁵ reported that, among all PSA-doped Cu₂O, Ga–Cu₂O shows better catalytic performance for the CRR because Ga makes a favourable interaction with the *OCHO intermediate for the formation of HCOOH. Hence, Cu₂O is also considered a suitable support for PSA and has been explored for the CRR.

SnS₂, MoS₂, and Cu₂O are semiconductors,^36–38 whereas the MXene-based support, Mo₂CO₂, is metallic in nature.³⁹ For a better electrocatalyst, metallic behavior is highly inevitable. The Mo₂CO₂ support is metallic, low-cost, and easy to synthesize, making it a viable catalyst in experiments.^39,40 Among MXene-based supports, Mo₂CO₂ has broad applications, including in energy storage, electrocatalysis, and thermoelectric materials.^39–42 Hence, based on its properties, Mo₂CO₂ is considered a suitable support for PSA. Based on the literature, the Mo₂CO₂ MXene-based support for the PSA catalyst has not been explored in greater detail for the CRR. This research aims to explore the potential feasibility of a PSA on a Mo₂CO₂ support for the electrocatalytic CRR to improve the stability, activity, and selectivity. In this work, we have systematically studied and analyzed the PSA on the Mo₂CO₂ catalyst for the CRR by using the density functional theory (DFT) approach.

2. Catalyst model and computational details

The Mo₂CO₂ is modeled by a 3 × 3 × 1 supercell of the Mo₂C (space group-P6₃/mmc) unit cell and is surface-functionalized with oxygen atoms due to their stability. Mo₂CO₂ has a total number of 45 atoms (18 Mo, 9 C, and 18 O).^43,44 Cheng et al.⁴² noted in their study that 3 × 3 × 1 and 4 × 4 × 1 supercells gave similar adsorption energy values for the adsorbate, which indicates that adsorption of the intermediate is not much affected by a change in the supercell, and a 10 Å vacuum (Table S1) is added in the z-direction to prevent interactions between periodic images.^23,39,45 This results in the lattice vector for Mo₂CO₂ 3 × 3 × 1 with dimensions a = 9.20 Å, b = 9.20 Å, c = 15.00 Å, α = 90°, β = 90°, and γ = 120°. The bottom three layers were frozen during Mo₂CO₂ optimization. All DFT simulations were performed using the Quantum Espresso-7.1 software package.⁴⁶ The calculation of the exchange-correlation energy was performed using the GGA-PBE method, where GGA is a generalized gradient approximation employed with the parameters provided by Perdew, Burke, and Ernzerhof, with spin polarization.^47,48 A 36 Ry for cut-off energy, 3 × 3 × 1 k-point for Brillouin zone, and DFT-D3 for van der Waals interactions were employed during the calculations.⁴⁹ The optimization of the structural geometry was allowed to proceed to the force and energy, which attained values of 2.0 × 10⁻² Ry/a.u. and 1.0 × 10⁻⁵ a.u., respectively. Image visualization was performed using the Vesta software.⁵⁰ To understand the stability of the single atom, Mo₂CO₂ binding energy calculations were performed using the following formula.⁵¹

E_b = E_PSA-Mo2CO2 − E_PSA − E_Mo2CO2 (1)

where E_b, E_PSA-Mo2CO2, E_PSA, and E_Mo2CO2 represent the binding energy of a p-block single atom, the total energy of a p-block single atom anchored on Mo₂CO₂, the p-block single atom electronic energy, and the electronic energy of bare Mo₂CO_2, respectively. The Gibbs free energy was calculated for all intermediates present in the electrochemical reduction reaction process using the computational hydrogen electrode (CHE) method.⁵²

ΔG = ΔE_DFT + ΔZPE − TΔS (2)

ΔG, ΔE_DFT, ΔZPE, T, and ΔS refer to the change in Gibbs free energy, DFT electronic energy for the reaction, zero point energy change, temperature, and entropy change, respectively. The entropy and zero-point energy for gaseous molecules and intermediates formed during the reaction are considered, as reported in the literature.^18,53,54 Entropy contributed by gaseous molecules such as CO₂, H₂, and H₂O is considered from the NIST database (Table S2), while the entropy contributed by the solid phase intermediate is neglected.^55,56

The adsorption energy is calculated by using the following equation.

E_ads = E_total − E_adsorbate − E_catalyst (3)

E_ads is the adsorption energy, E_total is the total energy of the system (adsorbate + catalyst), E_adsorbate is the electronic energy of the adsorbate, and E_catalyst is the electronic energy of the catalyst.

The cohesive energy of the system is calculated using the following formula.

The n_Mo, n_C, n_O, and n_PSA represent the number of molybdenum, carbon, oxygen and p-block single atoms, respectively, present in the catalyst and E_Mo, E_C, E_O, and E_PSA, and E_catalyst denote the energy of molybdenum, carbon, p-block single atom, and the catalyst total energy, respectively. The n₁ + n₂ + n₃ + n₄ term represents the total number of atoms.

Limiting potential is calculated by the formula given below.

U_L = −ΔG_max/e (5)

ΔG_max is the maximum Gibbs free energy barrier of the potential determining step (PDS), and e represents the charge of an electron.⁵⁷

3. Result and discussion

3.1 Structure and stability of PSA@Mo₂CO₂

Mo₂C exhibits a hexagonal structure and is formed by a three-layer Mo–C–Mo, where the upper and lower layers are made of Mo, and the middle layer is carbon. A previous study demonstrated that Mo₂C can be synthesized experimentally, and its surface can be functionalized with O, F, or OH.⁵⁸ However, oxygen surface functionalization leads to more stability than other surface functionalizations.^44,59 The calculated Mo–Mo and Mo–C bond lengths were 3.07 Å and 2.07 Å, respectively, which are similar to those reported in previous computational and experimental studies.^58,60 This work initially investigates the PSA stability in various coordination environments on the Mo₂CO₂ two-dimensional nanomaterial support at the atomic level. 23 PSA were anchored onto bare Mo₂CO₂, except for the radioactive one, to avoid environmental problems and considering safety measures. The PSA stability on three different sites (hollow-C (S1), hollow-Mo (S2), and top-O (S3)) is considered on Mo₂CO₂ with different coordination environments (Fig. 1(a)). The PSA anchored on the hollow-C site of Mo₂CO₂, coordinating with the oxygen and molybdenum atoms present on the surface of Mo₂CO₂. For the hollow Mo site, PSA is anchored on the Mo site only, whereas the top-O site PSA is anchored only with the oxygen atom. All these sites were considered to find the most preferred site for PSA on Mo₂CO₂. For being a better catalyst, it should be thermodynamically stable. The thermodynamic stability of PSA@Mo₂CO₂ is explained by the binding energy, which indicates the strength of the interaction between PSA and Mo₂CO₂. The calculated binding energy reveals that PSA@Mo₂CO₂ exhibits a negative binding energy, suggesting that it is thermodynamically stable and that the catalyst can be synthesized experimentally.^23,61 All hollow-C (S1), hollow-Mo (S2), and top-O (S3) site optimized structures of PSA@Mo₂CO₂ are shown in Fig. S3–S5. Among all adsorption sites, the binding energy of the majority of PSA on the hollow-C (S1) site is lower than that on the hollow-Mo (S2) and top-O (S3) sites, as shown in Fig. 1(b), indicating that out of all sites, most of the PSA is thermally stable on the hollow-C site (S1). This is due to the weaker repulsion from the surrounding atoms at the hollow-C site,²³ which is the main reason to consider it for further studies. Out of 23 PSA, silicon shows the lowest binding energy (−7.63 eV), indicating that silicon is the most stable single atom on the Mo₂CO₂ surface and binds strongly to the oxygen site.^62,63 Chen et al.⁶⁴ reported that amorphous SiO₂ with an excess oxygen defect exhibits a Si–O bond length of approximately 1.90 Å, similar to the reported value for Si@Mo₂CO₂. The average bond length of Si anchored on the oxygen surface of Mo₂CO₂ is lower than that of the previously published value of 2.79 Å, demonstrating that Si is most stable on the oxygen surface of Mo₂CO₂ due to a shorter bond length compared to the previous study of Si anchored on the Mo of Mo₂CO₂.³⁹ Silica (β-cristobalite) shows covalent bonding, with a bond length for Si–O of approximately 1.61 Å.^65,66 However, for Si@Mo₂CO₂, the Si–O bond length (1.90 Å) shows a weak covalent bond character as the bond length elongates. Halogens (F, Cl, Br, and I) are unstable at the carbon top position and shift near molybdenum, as halogens tend to coordinate better with metals to minimize repulsion from the surface termination group.⁶⁷ Therefore, they are not considered for further studies. Furthermore, to gain a better understanding of the bonding interaction between PSA and Mo₂CO₂, the calculated average bond lengths for PSA@Mo₂CO₂ are presented in Table S3 of the SI. The atomic radius of PSA generally increases going downward the group from B to Tl, C to Pb, N to Bi, and O to Te. However, when the PSA is anchored on Mo₂CO₂, the average bond length between the PSA and Mo₂CO₂ neighbouring atoms shows variation, as the PSA is anchored on either the oxygen or molybdenum site of Mo₂CO₂. In the case of the 14^th group elements, the average bond length from C to Si decreases from 1.99 Å to 1.90 Å and that from Si to Pb increases from 1.90 Å to 2.35 Å because carbon is anchored on the Mo site and Si, Ge, Sn, and Pb are anchored on the oxygen site of Mo₂CO₂, and the atomic radius of Mo is greater than that of O, which leads to an increase in the average bond length. Next, in the case of the 15^th group elements from As to Bi, the average bond length increases down the group from 1.88 Å to 2.57 Å, except for N (1.95 Å) and P (2.19 Å). N is anchored on the Mo site of Mo₂CO₂ and P is anchored on both Mo and O sites of Mo₂CO₂, which shows a different nature of an increase in bond length due to the increase in atomic radius of Mo compared to O. For the last 16^th group element, the average bond length follows an increasing trend from O (2.00 Å) to Se (2.58 Å) except Te (2.57 Å), Because O, S, and Se are anchored on Mo and Te is anchored on oxygen and Mo, this leads to the difference in the average bond length.

Fig. 1 (a) Side and top views of the Mo₂CO₂ optimized structure. (b) Binding energy of a PSA anchored on Mo₂CO₂ on three different sites: hollow-C (S1), hollow-Mo (S2), and top-O (S3). (c) Comparison between the binding energy versus cohesive energy of PSA@Mo₂CO₂.

Furthermore, the stability and clustering ability of PSA on the Mo₂CO₂ surface are evaluated by comparing the binding and cohesive energies. As shown in Fig. 1(c), the calculated cohesive energy is higher than the binding energy for B, N, F, S, Cl, Ga, As, Se, Br, Sb, Te, I, Tl, and Bi; this reveals that PSA is unstable on the Mo₂CO₂ surface. However, in the case of PSA (C, O, Al, Si, P, Ge, In, Sn, and Pb), the binding energy of C (−6.68 eV), O (−6.54 eV), Al (−7.28 eV), Si (−7.62 eV), P (−7.30 eV), Ge (−6.81 eV), In (−6.36 eV), Sn (−6.82 eV), and Pb (−6.69 eV) is lower than the cohesive energy of C (−6.31 eV), O (−6.31 eV), Al (−6.33 eV), Si (−6.33 eV), P (−6.33 eV), Ge (−6.31 eV), In (−6.31 eV), Sn (−6.32 eV), and Pb (−6.31 eV), which indicates that these single atoms can be stable and uniformly anchored on the Mo₂CO₂ surface.²¹ Therefore, only C, O, Al, Si, P, Ge, In, Sn, and Pb have been considered for further study.

3.2 CO₂ activation

The CO₂ molecule has C–O1 and C–O2 bond lengths of 1.17 Å and a bond angle of 180°. In the CRR process, the first step is the adsorption and activation of CO₂. Initially, CO₂ adsorbs onto the catalyst, with two possible adsorption modes: end-on and side-on site adsorption.⁶⁸

In end-on adsorption mode (Fig. S6 and Table S4), O1 of CO₂ adsorbs through a monodentate binding on PSA@Mo₂CO₂ (C, O, Al, Si, P, Ge, In, Pb, and Sn) and allowing energy minimization. This leads to an increase in the bond length of C–O1 from 1.17 Å to a range of 1.17–1.37 Å. Additionally, the bond angle varies from 180° to a range between 127.21° and 179.92°. These results suggest that CO₂ is being activated on the PSA@Mo₂CO₂ (PSA = Al, Si, Ge, In, Sn, and Pb) catalyst surface. Liu et al.⁶⁹ reported that the C–O1 and C–O2 bond lengths of adsorbed CO₂ on the Al site of γ-Al₂O₃(100) are found to be 1.18 Å and 1.17 Å, indicating that CO₂ activation is more favorable on Al@Mo₂CO₂ compared to γ-Al₂O₃(100). Additionally, CO₂ on Sn@Mo₂CO₂ shows a similar bond length and angle as observed on the SnO₂(110) catalyst reported by wang et al.⁷⁰ In the case of side-on adsorption (Fig. S7 and Table S5), CO₂ tends to adsorb in a bidentate mode through the carbon and the O1 atom on PSA@Mo₂CO₂ (PSA = C, O, Al, Si, P, Ge, In, Sn, and Pb), allowing geometry optimization. The resulting geometries show that the CO₂ molecule has varied structural features, with bond lengths ranging from 1.17 Å to 1.42 Å and bond angles ranging from 129.32° to 179.92°. Among all, based on CO₂ structural features, it is found that Si effectively activates CO_2, consistent with the results of Mao et al.⁷¹ The adsorption of CO₂ on Si@Mo₂CO₂ yields a weak adsorption energy (−0.03 eV). The energy of CO₂ adsorption comes under the physisorption range. However, after CO₂ adsorption, CO₂ bond lengths (C–O1 = 1.42 Å and C–O2 = 1.19 Å) and the angle (129.32°) changed significantly from a linear structure with bond lengths (C–O1 = 1.17 Å and C–O2 = 1.17 Å) and the bond angle (180°). The significant change in the structural features of CO₂ is supported by the Bader charge and charge density difference. The Bader charge analysis shows that after the adsorption of CO₂ on Si@Mo₂CO₂ (C = +0.96e, O1 = −1.10e, and O2 = −1.02e), there is a charge transfer between CO₂ and Si@Mo₂CO₂, as shown in Table S6, compared to a neutral linear CO₂ (C = +2.06e, O1 = −1.03e, and O2 = −1.03e) molecule. This is also confirmed by the charge density difference of CO₂ adsorbed on Si@Mo₂CO₂ (Fig. S8). The charge density difference for CO₂ adsorbed on Si@Mo₂CO₂ shows charge depletion on C and O2 and accumulation on O1. The structural change, Bader charge (Table S6), and charge density difference (Fig. S8) indicate that the adsorption of CO₂ on Si@Mo₂CO₂ is chemisorption. This observation aligns well with the previous study, which showed that weakly adsorbed bent CO₂ on reconstructed Cu₂O (111) undergoes chemisorption due to significant charge transfer.⁷²

For C, O, P, and Ge-based PSA@Mo₂CO₂, no significant structural features of CO₂ are observed when it is in the side-on and end-on configurations. Therefore, these PSA@Mo₂CO₂ configurations were not considered for further analysis. Whereas on Al, Si, In, Sn, and Pb, after optimization, side-on and end-on show bond length (C–O) and angle (O–C–O) variations. Based on the bond distance between PSA and O1 of CO₂ and Bader charge analysis (Table S6), Al@Mo₂CO₂ and Si@Mo₂CO₂ show chemisorption among all PSA, as shown in Tables S5 and S6.

3.3 Electronic structure analysis for Mo₂CO₂ and PSA@Mo₂CO₂

The projected density of states (PDOS) analysis was performed to elucidate the electronic properties of Mo₂CO₂ and PSA@Mo₂CO₂ and to explain the orbital interaction between PSA and the oxygen surface of Mo₂CO₂.¹⁴ For Mo₂CO₂, as shown in Fig. 2(a), below the Fermi level, the Mo d-orbital electronic state intensity was lower, and the O valence p-orbital electronic state intensity was higher. In contrast, Mo contributes more to the conduction band above the Fermi level. At the same time, oxygen exhibits a low-intensity electronic state in the conduction band, indicating electron transfer from Mo to O.

Fig. 2 PDOS for (a) Mo₂CO₂ and (b–f) PSA@Mo₂CO₂ (PSA = Al, Si, In, Sn, and Pb), with the Fermi level indicated by a black dashed line. Below the Fermi level is the valence band, and above the Fermi level is the conduction band on the x-axis.

Additionally, the Mo-d and O-p valence bands overlap effectively, indicating a strong interaction between the Mo-d and O-p valence orbitals. Therefore, Mo₂CO₂ is stable with an O-terminated surface.¹⁷ The PDOS plot in Fig. 2(a) indicates that Mo₂CO₂ exhibits a metallic-like nature, characterized by a zero-band gap, as the molybdenum and oxygen valence band peaks align with the Fermi level.³⁹ However, the introduction of the p-block element shifts the valence band near the Fermi level. In Fig. 2(b), the Al valence p-orbital peak passes through the Fermi value (E–E_F = −0.30 eV), resulting in Al@Mo₂CO₂ exhibiting metallic-like behavior, and the Al valence p-orbital overlaps with the O valence p-orbital, showing p–p orbital hybridization, indicating that Al strongly binds and is stable on the oxygen surface of the Mo₂CO₂ support. The Al p-orbital electronic state on the conduction band side shows transfer of an electron from Al to Mo₂CO₂, which is confirmed by the Bader charge of Al (+1.79e) and the charge density difference of Al@Mo₂CO₂, as shown in Fig. 3(e). Al@Mo₂CO₂ with the most positive Bader charge exhibited the most stable CO₂ adsorption with an adsorption energy of −0.97 eV. As shown in Fig. 2(c), the Si-p and O-p valence orbitals overlap near the Fermi level and show strong p–p hybridization and bonding between Si and O in Mo₂CO₂. The Si-p orbital electronic state in the conduction band (E–E_F = 2.60 eV) side shows transfer of an electron from Si to Mo₂CO₂, which is confirmed by the Bader charge of Si (+1.65e) and charge density difference, as shown in Fig. 3(f). Si@Mo₂CO₂ exhibits a metallic-like character due to a non-zero electronic state at the Fermi level. Furthermore, the charge density difference shows charge transfer between Si and CO₂, as shown in the SI (Fig. S8). The carbon in CO₂ is electrophilic in nature, accepts an electron from Si, and donates an electron from nucleophilic oxygen to Si and adsorbs by side-on mode with weak adsorption (E_ads = −0.03 eV). In shows a low-intensity p-orbital near the Fermi level compared to the Si-p orbital (Fig. 2(d)) and overlaps with the O p orbital, indicating the lower stability of In on the oxygen of Mo₂CO₂ compared to Si. In@Mo₂CO₂ exhibits metallic-like behavior because the valence band crosses the Fermi level. Fig. 2(e and f) shows that Sn-p and Pb-p contribute near the Fermi level. The Pb-p orbital electronic state on the conduction-band side shows an electronic state near the Fermi level, compared to the Sn-p orbital. This indicates less electron transfer from Pb than from Sn to the Mo₂CO₂ surface. This is confirmed by the Bader charge analysis of Sn (+1.40e) and Pb (+1.37e). The Sn p-orbital shows moderate electron transfer to CO₂ with moderate adsorption energy (−0.15 eV) compared to Pb (−0.23 eV). The valence band p-orbital peak intensity for PSA (Al, Si, In, Sn, and Pb) near the Fermi level side is in the order of Si > Al > Sn > Pb > In. This leads to greater charge transfer from Si and Al after CO₂ adsorption, as confirmed by the charge-density difference shown in the SI (Fig. S8). Fig. S9(a–e) represents the PDOS for PSA@Mo₂CO₂. The electronic state present on the valence band side (E–E_F = 0 to −8 eV) below the Fermi level for the Si-p orbital is more populated compared to the Al, In, Sn, and Pb p-orbital electronic states. The In p-orbital electronic state intensity is very low, whereas Al and Pb have p-orbital electronic state intensities higher than In. Sn shows optimal electronic state intensity on the valence-band side. This facilitates the *OCHO intermediate adsorption as evident from the adsorption energy (−2.78 eV), which is more favorable for the electrochemical CRR. Si@Mo₂CO₂ (−4.35 eV) has the lowest adsorption energy and strongly adsorbs *OCHO. Pb@Mo₂CO₂ (−2.33 eV) has a higher adsorption energy for *OCHO compared to Sn@Mo₂CO₂ (−2.78 eV). Al@Mo₂CO₂ (−1.81 eV) shows poorer adsorption of the intermediate than PSA@Mo₂CO₂ (PSA = Si, Sn, and Pb). The addition of PSA broadens the band, increases charge transfer, and facilitates intermediate adsorption for the electrocatalytic CRR.

Fig. 3 Optimized geometry of (a) Al@Mo₂CO₂, (b) Si@Mo₂CO₂, (c) Sn@Mo₂CO_2, and (d) Pb@Mo₂CO₂. Side views of charge density difference of (e) Al@Mo₂CO₂, (f) Si@Mo₂CO₂, (g) Sn@Mo₂CO₂ and (h) Pb@Mo₂CO_2, and the top view of (i) Al@Mo₂CO₂, (j) Si@Mo₂CO₂, (k) Sn@Mo₂CO₂, and (l) Pb@Mo₂CO₂. The charge density difference for the *OCHO intermediate adsorbed on (m) Al@Mo₂CO₂, (n) Si@Mo₂CO₂, (o) Sn@Mo₂CO_2, and (p) Pb@Mo₂CO₂; yellow and cyan indicate charge accumulation and depletion, respectively, with an isosurface value of 0.005 e Å⁻³. The blue arrow shows the transfer of charge from PSA to the *OCHO intermediate (except Sn). In the case of Sn, the blue arrow shows charge transfer from Mo₂CO₂ to Sn, and the red arrow displays Sn to *OCHO charge transfer.

3.4 Bader charge and charge density difference analysis for PSA@Mo₂CO₂

The Bader charge analysis was performed to deepen the understanding of charge transfer in Mo₂CO₂ and PSA@Mo₂CO₂. The charge is analyzed before and after PSA anchoring on Mo₂CO₂. Fig. S10(a) represents a uniformly distributed charge on bare Mo₂CO₂, where O and Mo have −1.01e and +1.71e charges. Fig. S10(b–e) shows that after anchoring PSA (Al, Si, In, Sn, and Pb) on Mo₂CO₂, charge transfer occurs from PSA to the neighboring Mo and O of Mo₂CO₂. Fig. S10(b) displays that after Al is anchored on Mo₂CO₂, the Bader charge on Mo (+1.56e, +1.49e, +1.57e) and O (−1.33e, −1.32e, −1.25e) increases. In Si@Mo₂CO₂, as shown in Fig. S10(c), the charge is transferred from Si to Mo (+1.60e, +1.60e, +1.56e) and O (−1.34e, −1.35e, −1.09e). In the case of In@Mo₂CO₂ (Fig. S10(d)), charge transfer occurs from In to Mo (+1.64e, +1.59e, +1.65e) and O (−1.05e, −1.13e, −1.13e). In Sn@Mo₂CO₂ (Fig. S10(e)), charge transfer occurs from Sn to Mo (+1.58e, +1.53e, +1.58e) and O (−1.08e, −1.19e, −1.19e). For Pb@Mo₂CO₂ (Fig. S10(f)), charge transfer occurs from Pb to Mo (+1.59e, +1.53e, +1.58e) and O (−1.09e, −1.16e, −1.16e). Out of all PSA@Mo₂CO₂ (PSA = Al, Si, In, Sn, and Pb), Sn (+1.40e) has an optimal charge. Further, the Bader charge is calculated after CO₂ adsorption on PSA@Mo₂CO₂. The Bader charges on Al, Si, In, Sn, and Pb were +2.43e, +3.00e, +0.86e, +1.41e, and +1.40e, respectively. Bader charge analysis shows that the charge transfer is negligible for In. Therefore, it is considered less active for the CRR and is excluded from further analysis. In the case of Al@Mo₂CO₂ and Si@Mo₂CO₂, the Bader charge significantly changes after CO₂ adsorption, which shows changes in bond angle and bond length at 175.39° (C–O1 = 1.20 Å and C–O2 = 1.15 Å) and 129.32° (C–O1 = 1.42 Å and 1.19 Å), respectively. Whereas, in the case of Sn@Mo₂CO₂ and Pb@Mo₂CO₂, very little charge transfer occurs when CO₂ is adsorbed, which shows negligible variation in the bond angle and bond length at 179.45° (C–O1 = 1.18 Å and C–O2 = 1.17 Å) and 179.82° (C–O1 = 1.18 Å and C–O2 = 1.17 Å) respectively (Table S5).

As seen in Fig. 3(a–d), the optimized structure of PSA@Mo₂CO₂ (where PSA represents Al, Si, Sn, and Pb) is considered for further study. Fig. 3(e–h) displays the side view of the charge density difference (CDD), and Fig. 3(i–l) shows the top view of PSA@Mo₂CO₂. The CDD analysis is utilized to understand the distribution of charge density on PSA@Mo₂CO₂. The positively charged region exhibits a depletion of charge density, indicated by the cyan-colored region, and the negatively charged region shows charge accumulation; electrons are concentrated in this region, represented by the yellow color. Al@Mo₂CO₂, Si@Mo₂CO₂, Sn@Mo₂CO_2, and Pb@Mo₂CO₂ show charge transfer from Al, Si, Sn, and Pb to Mo₂CO₂.^23,32,73 Most of the electrons are gained by oxygen, which is shown by yellow colour, and cyan color on Mo displays the charge transfer from Mo to O; among Al, Si, Sn, and Pb, more charge is accumulated on Si, followed by Pb, Sn and Al, which indicate Si can donate and accept electrons, and Sn and Pb have lower number of electrons. The Bader charges on Al@Mo₂CO₂ (+1.79e), Si@Mo₂CO₂ (+1.65e), Sn@Mo₂CO₂ (+1.40e), and Pb@Mo₂CO₂ (+1.37e) align with the CDD study, which shows that Al shows charge depletion due to maximum electron transfer. Si shows charge accumulation and charge depletion due to partial charge transfer from Si to Mo₂CO₂. In contrast, Sn shows greater charge depletion and less charge accumulation than Pb, as shown in Fig. 3(k–l) (top view). This is because most of the electron transfer occurs from Sn and Pb to Mo₂CO₂. This is due to the metallic properties of Al, Sn, and Pb, as metals can donate electrons more readily than Si, which is a metalloid.

Fig. 3(m–p) displays the charge transfer between the *OCHO intermediate, PSA, and the Mo₂CO₂ surface. Fig. 3(m–n) shows higher charge transfer from Al and Si to *OCHO, whereas Pb shows lower charge transfer (Fig. 3(p)). The cyan color present on Al, Si, Sn, and Pb indicates charge depletion, with the charge transferred from PSA to *OCHO. The yellow color present on the *OCHO intermediate shows charge gain. Among all PSAs, Sn shows distinct charge transfer, as shown in Fig. 3(o). Charge is transferred from the Mo₂CO₂ surface to Sn and, further, Sn to the *OCHO intermediate, as shown in Fig. S11, due to which Sn shows distinct Bader charge transfer after adsorption of *OCHO on Sn@Mo₂CO₂. The overall Bader charge on Sn is −0.01e after *OCHO intermediate adsorption on Sn@Mo₂CO₂.

3.5 Reaction mechanism for the CRR and HER

Electrochemical CRR is challenging and complicated; therefore, an effective catalyst is required. The reduction of CO₂ to CO and HCOOH on a catalyst surface is a two-step, two-electron pathway that proceeds via a twice-proton-coupled electron–transfer reaction, as shown in Fig. 4, which is widely proposed as the path and accepted as the most likely one. The first step in CO₂ reduction involves the acceptance of a proton and an electron, leading to two possible competitive intermediates, *OCHO and *COOH. This indicates that there are two pathways: either the carbon or the oxygen site of CO₂ can receive a proton and an electron. Further, proton and electron transfer occur, and the *OCHO and *COOH intermediates are converted into *HCOOH and *CO, respectively. In the last step, *HCOOH and *CO desorb from the catalyst surface, forming HCOOH and CO products.⁷⁴ In contrast, the selectivity toward CH₄ and CH₃OH depends on the adsorption of the *CO intermediate. If the adsorption of *CO is weaker, it will terminate at the CO product. As shown in Fig. 5(a–f), the Gibbs free energy for the PSA@Mo₂CO₂ *CO intermediate is poor. Therefore, CO₂ does not reduce to CH₃OH and CH₄.⁷⁵ Furthermore, the p-block metals In, Sn, and Pb are known for electrochemically reducing CO₂ to HCOOH, with faradaic efficiencies exceeding 90%. CH₃OH and CH₄ follow six- and eight-electron pathways, respectively, whereas HCOOH follows a two-electron pathway.⁷⁶ Therefore, CO₂ reduction is not further considered for CH₄ and CH₃OH.

Fig. 4 Schematic of the reaction mechanism for the electrochemical CRR to CO and HCOOH products and the competing HER on the catalyst surface. For the HCOOH product, there are two possible binding sites for the intermediate (*OCHO), one oxygen site or two oxygen sites.

Fig. 5 Gibbs free energy diagram for CO₂ + 2H⁺+2e⁻ to CO and HCOOH products for (a) Mo₂CO₂ and (b–e) PSA@Mo₂CO₂(PSA = Al, Si, Sn, and Pb) and (f) HER Gibbs free energy for Mo₂CO₂ and PSA@Mo₂CO₂ (PSA = Al, Si, Sn, and Pb), where * represents the catalyst surface in all Gibbs free energy diagrams.

The energy-minimized structures of all intermediates involved in the CRR on the surfaces of the Mo₂CO₂ and PSA@Mo₂CO₂ (Al, Si, Sn, and Pb) catalysts are displayed in Fig. S12. The Gibbs free energy diagram for the CRR on Mo₂CO₂ and PSA@Mo₂CO₂ surfaces is shown in Fig. 5. For the bare Mo₂CO₂ surface (Fig. 5(a)), the Gibbs free energy for the first step, CO₂ into *OCHO and *COOH intermediates, is 1.60 eV and 1.05 eV, respectively. The calculated values suggest that the *OCHO and *COOH intermediates are present as PDS with the maximum Gibbs free energy change (Tables S8 and S9).⁶¹ Among the intermediates, the *COOH intermediate formation is more favorable with a lower Gibbs free energy change in the first step compared to *OCHO. Therefore, the subsequent formation of CO is more favorable than that of HCOOH. On Al@Mo₂CO₂ (Fig. 5(b)), the calculated Gibbs free energy change value suggests that *COOH (3.53 eV) and *OCHO (1.56 eV) are the PDS for CO and HCOOH (Tables S8 and S9), respectively. This result shows that *OCHO intermediate formation is more favorable compared to *COOH due to the lower Gibbs free energy change. Therefore, HCOOH product formation is preferred on Al@Mo₂CO₂, consistent with the earlier report.⁷⁷ Furthermore, the Gibbs free energy plot for CO₂ reduction on the Si@Mo₂CO₂ surface (Fig. 5(c)) reveals that the conversion of *COOH to *CO and *OCHO to *HCOOH is found to be PDS, with a maximum Gibbs free energy change of 1.69 eV and 2.34 eV, respectively (Tables S8 and S9). This indicates that Si@Mo₂CO₂ promotes the formation of CO rather than HCOOH. The Sn@Mo₂CO₂ (Fig. 5(d)) catalyst surface shows that the * + CO₂ to *COOH and *OCHO conversion steps are found to be the PDS with Gibbs free energy changes of 1.95 eV and 0.58 eV, respectively (Table S8 and S9). This indicates that the HCOOH product is preferred over the CO product. In the case of the Pb@Mo₂CO₂ surface (Fig. 5(e)), the PDS for the *+CO₂ to *OCHO and *COOH shows a Gibbs free energy change of 1.03 eV and 2.32 eV, respectively (Tables S8 and S9). This indicates that it favors the formation of HCOOH, as its Gibbs free energy change is lower than that for the formation of CO. Out of all the catalysts, the Sn@Mo₂CO₂ surface exhibited the lowest Gibbs free energy change for HCOOH formation from CO_2.

The reason for the Gibbs free energy difference between the *COOH and *OCHO intermediate on PSA@Mo₂CO₂ (PSA = Al, Si, Sn, and Pb) is elucidated using Bader charge analysis. This deepens the understanding of the CO and HCOOH reaction pathway and PSA@Mo₂CO₂ catalytic activity. The Gibbs free energy for the *OCHO intermediate on PSA@Mo₂CO₂, (PSA = Al (1.56 eV), Si (−0.99 eV), Sn (0.58 eV), and Pb (1.03 eV)) is found to be lower as compared to that of the Gibbs free energy for the *COOH intermediate on PSA@Mo₂CO₂, (PSA = Al (3.53 eV), Si (−0.05 eV), Sn (1.95 eV), and Pb (2.32 eV)). The Bader charges present on the PSA of PSA@Mo₂CO₂ (PSA = Al (+1.79e), Si (+1.65e), Sn (+1.40e) and Pb (+1.37e)) before intermediate adsorption are shown in Table S7. After *COOH and *OCHO adsorption on PSA@Mo₂CO₂, the Bader charge present on the PSA of PSA@Mo₂CO₂ increases (except for Sn), as shown in Table S7. Meanwhile, Sn has a decreased Bader charge for *COOH and *OCHO. Interestingly, the *OCHO charge transfer is higher than that of *COOH. The *OCHO intermediate adsorption is preferred on the positively charged region, as the oxygen of *OCHO behaves as a nucleophile.³⁵ This charge transfer makes the *OCHO interact more strongly with PSA@Mo₂CO₂, thereby lowering the Gibbs free energy.

Whereas in the case of Sn@Mo₂CO₂, Sn receives a charge from the Mo₂CO₂ support and transfers an electron to the intermediate *OCHO, as shown in the SI (Fig. S11(a–e)). This demonstrates that Sn behaves as a good mediator in modulating the charge transfer between the intermediate and Mo₂CO₂.³⁴ Furthermore, the potential role of Mo₂CO₂ in charge transfer is explored in detail through Bader charge analysis and charge density difference (Fig. S11(a–e)). First, the Mo₂CO₂ surface Bader charge is calculated, which shows −1.01e present on every oxygen atom and +1.71e present on every molybdenum atom. After that, a single-atom Sn is anchored on Mo₂CO₂ with three coordinated oxygen atoms. The charges present on neighboring Mo and O atoms underwent a significant change. The Bader charge analysis shows a reduction in the positive charge of Mo (+1.58e, +1.58e, and +1.53e) and an increase in the negative charge of O (−1.19e, −1.19e and −1.08e) that indicates charge transfer from Sn (+1.40e) to Mo₂CO₂. This is confirmed by charge density difference: Sn@Mo₂CO₂ shows charge depletion on Sn (Fig. S11(d)), indicated by the cyan color, and the yellow color on the oxygen surface indicates charge gain.

At last, the *OCHO intermediate adsorbed on the Sn@Mo₂CO₂ configuration is shown in Fig. S11(c). After the adsorption of *OCHO on Sn@Mo₂CO₂, the Sn coordination environment changed from three to one on Mo₂CO₂. The positive charge present on the neighboring Mo (+1.59e, +1.56e, +1.71e) increases and the negative charge present on O (−1.00e, −1.05e, −1.08e) decreases further, indicating that the overall charge decreases on the surface of Mo₂CO₂. This shows that the charge transfer occurs from Mo₂CO₂ to Sn. The charge present on Sn (+1.39e) shows a slight decrease in positive charge and tries to maintain its own charge by accepting electrons from the Mo₂CO₂ support and transferring to the *OCHO intermediate. This is confirmed by the charge density difference (Fig. S11). After intermediate *OCHO adsorption on Sn@Mo₂CO₂ (Fig. S11(e)), the charge on the oxygen surface decreases compared to Sn@Mo₂CO₂ before intermediate adsorption, shown in yellow color. While the charge on Sn increases, shown by the yellow color, indicating charge gain by Sn and Sn transferring charge to the *OCHO intermediate. This distinct behavior of Sn as a mediator between *OCHO and Mo₂CO₂ facilitates the formation of HCOOH with a lower Gibbs free energy change compared to other systems, such as Al@Mo₂CO₂, Si@Mo₂CO₂, and Pb@Mo₂CO₂.

During the CRR, the HER also occurs in parallel and competes with the CRR process. Fig. 5(f) shows the calculated HER Gibbs free energy profile for Mo₂CO₂ and PSA@Mo₂CO₂ (PSA = Al, Si, Sn, and Pb) catalyst surfaces. According to the Sabatier principle, the interactions between the catalyst and the intermediate should be within an optimal range, neither too low nor too high,⁷⁸ to improve catalytic activity. The Gibbs free energy changes for H*, *OCHO, and *COOH (ΔG_H*, ΔG_*OCHO, and ΔG_*COOH) intermediates are shown in Fig. S13 to provide deeper insights into the CRR versus the HER. For an efficient electrochemical CO₂ reduction reaction, an electrocatalyst should suppress the hydrogen evolution reaction and accelerate the formation of intermediates towards HCOOH and CO. Therefore, the Gibbs free energy for H*, *OCHO, and *COOH intermediates are computed and compared for Mo₂CO₂, Al@Mo₂CO₂, Si@Mo₂CO₂, Sn@MO₂CO₂ and Pb@Mo₂CO₂. Mo₂CO₂ (−0.16 eV), Al@Mo₂CO₂ (−1.06 eV), and Si@Mo₂CO₂ (−1.26 eV) have lower ΔG for H* compared to *COOH (Mo₂CO₂ (1.05 eV), Al@Mo₂CO₂ (3.53 eV), and Si@Mo₂CO₂ (−0.05 eV)) and *OCHO (Mo₂CO₂ (1.60 eV), Al@Mo₂CO₂ (1.56 eV), and Si@Mo₂CO₂ (−0.99 eV)). Therefore, H* adsorption will inhibit *COOH and *OCHO formation. Whereas in the case of Sn@Mo₂CO₂ (0.58 eV) and Pb@Mo₂CO₂ (1.03 eV), for *OCHO, the Gibbs free energy is lower compared to *COOH (Sn@Mo₂CO₂ (1.95 eV) and Pb@Mo₂CO₂ (2.32 eV)) and H* (Sn@Mo₂CO₂ (2.18 eV) and Pb@Mo₂CO₂ (1.38 eV)). This shows Sn@Mo₂CO₂ and Pb@Mo₂CO₂ facilitate *OCHO formation. The Sn@Mo₂CO₂ has a lower ΔG for *OCHO than Pb@Mo₂CO₂, showing that Sn@Mo₂CO₂ facilitates the conversion of CO₂ to HCOOH. U_L was considered to analyze the electrocatalyst (PSA@Mo₂CO₂) activity.⁷⁹ U_L can be defined as the negative maximum difference between the Gibbs free energy of the reaction step (−ΔG_max) divided by e (electron charge). It is used as an electrocatalyst, demonstrating superior performance among all PSA@Mo₂CO₂ electrocatalysts, as shown in Fig. 6(a). The catalysts exhibit a lower negative value of U_L, which is considered beneficial because less energy is required to overcome the PDS.⁵⁷ For Al@Mo₂CO₂, Sn@Mo₂CO₂, and Pb@Mo₂CO₂, the formic acid product pathway is more favorable due to their low negative U_L values of −1.56, −0.58, and −1.03 V, respectively, compared to CO product formation, as shown in Fig. 6(a), Tables S10 and S11. However, for Mo₂CO₂ and Si@Mo₂CO₂, the CO formation pathway is more favorable due to the lower U_L values of −1.05 and −1.69 V, respectively, compared to HCOOH.

Fig. 6 (a) Limiting potential for the CRR (U^CRR_L) for HCOOH and CO production on Mo₂CO₂ and PSA@Mo₂CO₂. (b–f) Bond length between the oxygen of the intermediate (*OCHO) and the catalyst (Mo₂CO₂ and PSA@Mo₂CO₂). (g) Volcano plot for PSA@Mo₂CO₂ and Mo₂CO₂ using the adsorption energy of *OCHO (E_ads) and U_L. (h) Selectivity of the CRR toward CO and HCOOH products. PSA@Mo₂CO₂ (PSA = Al, Si, Sn, and Pb) is used to denote the corresponding systems.

3.6 Structure–activity relationship

A suitable structure–activity descriptor is developed to predict the catalytic activity. Here, bond length is considered as a descriptor, and an optimal bond length is required between PSA and *OCHO intermediate to form HCOOH effectively. Fig. 6(b–f) shows how the key intermediate *OCHO adsorbs on Mo₂CO₂ and PSA@Mo₂CO₂. On Mo₂CO₂ and Si@Mo₂CO₂, the *OCHO intermediate strongly adsorbs through a monodentate oxygen site with shorter bond lengths of 2.06 Å and 1.64 Å, respectively. In the bidentate adsorption mode of *OCHO on Al@Mo₂CO₂, Sn@Mo₂CO₂, and Pb@Mo₂CO₂, the average bond lengths between the O atom and the PSA (Al, Sn, and Pb) are 1.93 Å (Al), 2.26 Å (Sn), and 2.46 Å (Pb). This bidentate adsorption mode resulted in an elongated bond length, causing weaker adsorption. The bond length for *OCHO adsorbed on Sn@Mo₂CO₂ is shorter than that on Pb@Mo₂CO₂ and longer than that of Al@Mo₂CO₂, Si@Mo₂CO₂, and Mo₂CO₂. To determine the optimal distance for HCOOH formation, a volcano plot has been constructed. The volcano plot displays a structure–activity relationship between the limiting potential and bond length (Fig. S14). As seen in Fig. S14, Sn@Mo₂CO₂ is positioned on top of the volcano with the lowest limiting potential and moderate bond length as compared to those of the other catalysts, suggesting that Sn@Mo₂CO₂ favors the formation of HCOOH.

In addition to that, a structure–activity relationship is developed between the limiting potential and the adsorption energy of the *OCHO intermediate. Here, adsorption energy is considered as a descriptor, and an optimal adsorption energy is required for HCOOH formation. Feaster et al.⁷⁴ plotted a volcano curve using partial current density and binding energy for the elements Ni, Cu, Zn, Sn, Ag, Pt, and Au. Among all, the Sn shows better activity towards CO₂ to HCOOH, as Sn is at the top of the volcano, indicating that the binding is not too high or too low for the *OCHO intermediate. For Al@Mo₂CO₂, Pb@Mo₂CO₂, and Mo₂CO₂ in Fig. 6(g), it is shown that the *OCHO intermediate adsorption is weaker, as it may not favor the formation of HCOOH effectively due to the weaker interaction with the catalyst. Whereas Si@Mo₂CO₂ demonstrates strong adsorption of the *OCHO intermediate, making it difficult to reduce further to HCOOH. Among Al@Mo₂CO₂, Si@Mo₂CO₂, Sn@Mo₂CO₂, Pb@Mo₂CO₂, and Mo₂CO₂, Sn@Mo₂CO₂ demonstrates the optimal adsorption energy for the *OCHO intermediate, which indicates Sn@Mo₂CO₂ favors CO₂ reduction to HCOOH. The tilted *OCHO intermediate on Sn@Mo₂CO₂ leads to optimal adsorption of the *OCHO intermediate on Sn@Mo₂CO₂ with the lowest limiting potential and is on top of the volcano plot, as shown in Fig. 6(g).

3.7 Selectivity for the CRR to CO/HCOOH

In general, competitive HER hinders the CRR at the catalyst surface. If H* adsorbs on the catalyst, then the CRR will be inhibited. For the effective CRR on catalysts, the HER must be avoided at the catalyst surface. The H* adsorption energy is compared with the adsorption energy of *CO₂ on Mo₂CO₂, Al@Mo₂CO₂, Si@Mo₂CO₂, Sn@Mo₂CO₂, and Pb@Mo₂CO₂ surfaces. It reveals that H* adsorption energy on Mo₂CO₂ (−0.42 eV), Al@Mo₂CO₂ (−1.32 eV), and Si@Mo₂CO₂ (−1.52 eV) is lower compared to the *CO₂ adsorption energy on Mo₂CO₂ (−0.41 eV), Al@Mo₂CO₂ (−0.97 eV), and Si@Mo₂CO₂ (−0.03 eV). From the above results, Mo₂CO₂, Al@Mo₂CO_2, and Si@Mo₂CO₂ will not favor the CRR. However, the H* adsorption energy for Sn@Mo₂CO₂ (1.92 eV) and Pb@Mo₂CO₂ (1.12 eV) is higher compared to the CO₂ adsorption energy for Sn@Mo₂CO₂ (−0.15 eV) and Pb@Mo₂CO₂ (−0.23 eV). Hence, Sn@Mo₂CO₂ and Pb@Mo₂CO₂ favor the CRR.⁵³

The selectivity between the CRR and HER was evaluated to determine the best catalyst among Mo₂CO₂ and PSA@Mo₂CO₂ (PSA = Al, Si, Sn, and Pb).⁷⁹ The difference between the U_L of the CRR and HER defines the selectivity (U^CRR_L–U^HER_L) on the PSA catalyst. A more positive U_L difference (U^CRR_L–U^HER_L) shows that the higher valued one yields the more pronounced selectivity towards the CRR.⁵⁷ Fig. 6(h) shows the selectivity towards HCOOH on the catalyst surface found in the order of Sn@Mo₂CO₂ > Pb@Mo₂CO₂ > Mo₂CO₂ > Al@Mo₂CO₂ > Si@Mo₂CO₂. This result suggests that Sn@Mo₂CO₂ exhibits the best activity and selectivity for the conversion of the CRR to HCOOH.

3.8 Orbital interaction analysis

The PDOS analysis was employed to investigate orbital interactions between the *OCHO intermediate and PSA@Mo₂CO₂ (where PSA represents Al, Si, Sn, and Pb) to understand the bonding between PSA (Al, Si, Sn, and Pb) and *OCHO. The presence of two oxygen and one carbon atom shows more and less intense electronic states, respectively. As shown in Fig. 7(a), the Al valence p-orbital electronic state, present far away from the Fermi level, shows lower electron density. The ineffective overlap between Al-p and O-p leads to weaker *OCHO intermediate adsorption on Al@Mo₂CO₂. The Bader charge and CDD show that Al@Mo₂CO₂ has the most positive Bader charge and leads to a weaker adsorption intermediate *OCHO (Fig. 7(b)). On the other hand, Fig. 7(c) shows stronger p–p hybridization between Si-p and O-p on the valence-band side. Hence, the bond length between Si–O is the shortest (1.64 Å), leading to strong bonding between Si and O.

Fig. 7 (a–d) Orbital interaction between PSA (Al, Si, Sn, and Pb) and the *OCHO intermediate. The black short dashed line at 0 eV represents the Fermi level.

In addition, CDD (Fig. 3) shows a higher charge transfer from Si to O of *OCHO. The overlap between the Sn p-orbital and the O p-orbital hybridized near the Fermi level (E–E_F = −1.70 eV) shows p–p hybridization and moderate bonding. Whereas, in the case of Pb (Fig. 7(d)), the peak of the valence p-orbital is far from the Fermi level at −1.80 eV, and the intensity is also lower compared to the Sn p-orbital, which leads to weaker adsorption of the *OCHO intermediate. The bond length of Sn–O (2.26 Å) is lower compared to Pb–O (2.43 Å). This aligns with p–p hybridization and leads to moderate adsorption of the *OCHO intermediate on Sn@Mo₂CO₂ and weaker adsorption on Pb@Mo₂CO₂. The best active Sn@Mo₂CO₂ catalyst stability is evaluated by using ab initio molecular dynamics simulations. The total energy obtained from the AIMD simulations is shown in Fig. S2 and exhibits small fluctuations without causing a rise in the catalyst's energy under dynamical conditions at 300 K for 10 ps. This shows that Sn@Mo₂CO₂ remains stable at ambient conditions.

4. Conclusion

In this work, 23 PSA@Mo₂CO₂ and bare Mo₂CO₂ catalysts for the electrochemical CRR were investigated using density functional theory. The binding and cohesive energy results suggest that, among 23 catalysts, 9 PSA@Mo₂CO₂ catalysts are the most stable. Side-on and end-on adsorption of CO₂ on PSA@Mo₂CO₂ predominantly results in linear geometries after energy minimization. Based on CO₂ adsorption and activation on Mo₂CO_2, PSA (Al, Si, In, Sn, and Pb) were screened, and PDOS analysis shows they exhibit metallic-like behavior. The Gibbs free energy diagram shows that the formation of *OCHO on PSA@Mo₂CO₂ is the potential-determining step, except for Si. The Gibbs free energy change for the potential-determining step is found to be in the order of Sn@Mo₂CO₂ < Pb@Mo₂CO₂ < Al@Mo₂CO₂ < Mo₂CO₂. Remarkably, Sn@Mo₂CO₂ is the most efficient catalyst with a low U_L value of −0.58 V, due to its optimal interaction with *OCHO, and is also located at the top of the volcano plot. Interestingly, the Bader charge and CDD analyses show that Sn@Mo₂CO₂ exhibits a unique charge of −0.01e on Sn, indicating that a higher charge is transferred from Mo₂CO₂ to Sn and then from Sn to *OCHO to maintain the neutral state. Hence, Sn acts as a mediator between Mo₂CO₂ and *OCHO, facilitating effective charge transfer. Additionally, Sn@Mo₂CO₂ shows the best selectivity toward the CRR and inhibits the HER. For better performance, the optimal p–p orbital overlap is essential. Furthermore, Sn@Mo₂CO₂ is thermally stable and may serve as an efficient catalyst for the formation of HCOOH from CO₂. This study provides orbital-based approaches for understanding catalytic behavior. Further, fine-tuning of the electronic properties of Mo₂CO₂ will help effectively convert CO₂ into value-added products.

Author contributions

Anshul Gupta: conceptualization, methodology, software, data curation, investigation, formal analysis, writing original draft. Shanmugam Ramasamy: supervision, validation, resources, funding acquisition, review, and editing.

Conflicts of interest

The authors declare no financial or personal relationship with other people or organizations that could influence or bias this work.

Data availability

The supporting data are provided in the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6na00397d.

Acknowledgements

AG is thankful to the VIT management for the fellowship and CO₂RGTC for the computing facilities. RS acknowledges VIT for the SEED grant support with reference number RGEMS/SG20220030. We thank the HPC facility for computational resources under DST-FIST grant number SR/FST/MS-II/2023/139(C).

References

1. S. C. Peter, ACS Energy Lett., 2018, 3, 1557–1561,  DOI:10.1021/acsenergylett.8b00878.
2. Y. Zhou, H. Cao, Z. An, M. Li, Y. Huo, J. Jiang, J. Xie and M. He, Appl. Surf. Sci., 2023, 616, 156492,  DOI:10.1016/j.apsusc.2023.156492.
3. A. S. Agarwal, Y. Zhai, D. Hill and N. Sridhar, ChemSusChem, 2011, 4, 1301–1310.
4. Q. Cheng, M. Huang, L. Xiao, S. Mou, X. Zhao, Y. Xie, G. Jiang, X. Jiang and F. Dong, ACS Catal., 2023, 13, 4021–4029.
5. Y. Linghu, T. Tong, C. Li and C. Wu, Appl. Surf. Sci., 2022, 590, 153001,  DOI:10.1016/j.apsusc.2022.153001.
6. S. Emken, G. Di Liberto and G. Pacchioni, Electrochim. Acta, 2024, 499, 144714,  DOI:10.1016/j.electacta.2024.144714.
7. L. Liu, H. Akhoundzadeh, M. Li and H. Huang, Small Methods, 2023, 7, 2300482,  DOI:10.1002/smtd.202300482.
8. D. N. Sredojević, I. Vukoje, Đ. Trpkov and E. N. Brothers, Phys. Chem. Chem. Phys., 2024, 26, 8356–8365.
9. G. Xing, S. Liu and J. yao Liu, Int. J. Hydrogen Energy, 2023, 48, 3486–3494.
10. Y. Shao, Q. Yuan and J. Zhou, Small, 2023, 19, 2303446,  DOI:10.1002/smll.202303446.
11. M. Li, H. Wang, W. Luo, P. C. Sherrell, J. Chen and J. Yang, Adv. Mater., 2020, 32, 2001848,  DOI:10.1002/adma.202001848.
12. A. Vasileff, C. Xu, Y. Jiao, Y. Zheng and S. Z. Qiao, Chem, 2018, 4, 1809–1831,  DOI:10.1016/j.chempr.2018.05.001.
13. Y. Cai, J. Fu, Y. Zhou, Y. C. Chang, Q. Min, J. J. Zhu, Y. Lin and W. Zhu, Nat. Commun., 2021, 12, 586,  DOI:10.1038/s41467-020-20769-x.
14. Z. Wang, L. Sun, W. Li, J. Qin, Y. Li and Y. Wang, Mater. Today Commun., 2025, 48, 113505,  DOI:10.1016/j.mtcomm.2025.113505.
15. S. Wang, L. Li, K. San Hui, F. Bin, W. Zhou, X. Fan, E. Zalnezhad, J. Li and K. N. Hui, Adv. Eng. Mater., 2021, 23, 2100405,  DOI:10.1002/adem.202100405.
16. P. Komen, S. Suthirakun, A. Plucksacholatarn, S. Kuboon, K. Faungnawakij and A. Junkaew, J. Colloid Interface Sci., 2025, 679, 1026–1035.
17. Z. Sun, R. Li, Q. Xi, F. Xie, X. Jian, X. Gao, H. Li, Z. Yu, J. Liu, X. Zhang, Y. Wang, Y. Wang, X. Yue and C. Fan, Phys. Chem. Chem. Phys., 2023, 25, 13728–13740,  10.1039/d3cp00779k.
18. Y. Yang, J. Li, C. Zhang, Z. Yang, P. Sun, S. Liu and Q. Cao, J. Phys. Chem. C, 2022, 126, 4338–4346.
19. P. Sikam, K. Takahashi, T. Roongcharoen, T. Jitwatanasirikul, C. Chitpakdee, K. Faungnawakij and S. Namuangruk, Appl. Surf. Sci., 2021, 550, 149380,  DOI:10.1016/j.apsusc.2021.149380.
20. Z. Lu, P. Lv, Z. Yang, S. Li, D. Ma and R. Wu, Phys. Chem. Chem. Phys., 2017, 19, 16795–16805.
21. H. Da Ren, Z. L. Lang, H. Q. Tan, Y. H. Wang and Y. G. Li, J. Phys. Chem. C, 2024, 128, 6702–6710.
22. H. P. Zhang, R. Zhang, C. Sun, Y. Jiao and Y. Zhang, Nanoscale, 2021, 13, 20541–20549.
23. A. Athawale, B. M. Abraham, M. V. Jyothirmai and J. K. Singh, J. Phys. Chem. C, 2023, 127, 24542–24551.
24. Y. Gogotsi and B. Anasori, ACS Nano, 2019, 13, 8491–8494,  DOI:10.1021/acsnano.9b06394.
25. S. Baskaran and J. Jung, Appl. Surf. Sci., 2022, 592, 153339,  DOI:10.1016/j.apsusc.2022.153339.
26. B. Mohanty, B. K. Jena and S. Basu, ACS Omega, 2020, 5, 1287–1295.
27. B. R. Anne, J. Kundu, M. K. Kabiraz, J. Kim, D. Cho and S. Il Choi, Adv. Func. Mater., 2023, 33, 2306100,  DOI:10.1002/adfm.202306100.
28. X. Zhan, C. Si, J. Zhou and Z. Sun, Nanoscale Horiz., 2020, 5, 235,  10.1039/c9nh00571d.
29. X. Sun, L. Sun, G. Li, Y. Tuo, C. Ye, J. Yang, J. Low, X. Yu, J. H. Bitter, Y. Lei, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2022, 61, e202207677,  DOI:10.1002/anie.202207677.
30. J. Yu, J. Li, C. Y. Xu, Q. Li, Q. Liu, J. Liu, R. Chen, J. Zhu and J. Wang, Nano Energy, 2022, 98, 107266,  DOI:10.1016/j.nanoen.2022.107266.
31. J. Guo, F. Yu, Y. You, J. Zhan and L. H. Zhang, Appl. Catal. B, 2024, 355, 124160,  DOI:10.1016/j.apcatb.2024.124160.
32. Y. Li, D. Gao, C. Tang, Z. Guo, N. Miao, B. Sa, J. Zhou and Z. Sun, J. Colloid Interface Sci., 2024, 658, 114–126.
33. C. Zhang, W. Chu, C. Chen and W. Sun, Catal. Today, 2024, 432, 114649.
34. G. Chen, M. Buraschi, R. Al-Heidous, S. Bonakala, F. El-Mellouhi and C. S. Cucinotta, J. Phys. Chem. C, 2024, 128, 15861–15872,  DOI:10.1021/acs.jpcc.4c02283.
35. H. Liu, D. Liu, Z. Yu, H. Bai and H. Pan, J. Colloid Interface Sci., 2025, 683, 170–177.
36. R. K. Mishra, G. J. Choi, H. J. Choi, J. Singh, S. H. Lee and J. S. Gwag, J. Alloys Compd., 2022, 921, 166018,  DOI:10.1016/j.jallcom.2022.166018.
37. Y. Jiang, S. Li, F. Zhang, W. Zheng, L. Zhao and Q. Feng, Catal. Commun., 2021, 156, 106325,  DOI:10.1016/j.catcom.2021.106325.
38. X. Yang, J. Cheng, X. Yang, Y. Xu, W. Sun and J. Zhou, Chem. Eng. J., 2022, 431, 134171,  DOI:10.1016/j.cej.2021.134171.
39. Q. Liu, H. Cheng, X. Wang and P. Qian, Phys. Chem. Chem. Phys., 2023, 25, 5056–5065.
40. R. Meshkian, L. Å. Näslund, J. Halim, J. Lu, M. W. Barsoum and J. Rosen, Scr. Mater., 2015, 108, 147–150.
41. M. Khazaei, M. Arai, T. Sasaki, M. Estili and Y. Sakka, Phys. Chem. Chem. Phys., 2014, 16, 7841–7849.
42. C. Cheng, X. Zhang, M. Wang, S. Wang and Z. Yang, Phys. Chem. Chem. Phys., 2018, 20, 3504–3513.
43. K. Kotmool, T. Kaewmaraya, T. Hussain, R. Ahuja, W. Luo and T. Bovornratanaraks, Phys. Chem. Chem. Phys., 2022, 24, 17862–17869.
44. J. Gan, F. Li, Y. Tang and Q. Tang, ChemSusChem, 2020, 13, 6005–6015.
45. R. Morales-Salvador, J. D. Gouveia, Á. Morales-García, F. Viñes, J. R. B. Gomes and F. Illas, ACS Catal., 2021, 11, 11248–11255.
46. P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. de Gironcoli, S. Fabris, G. Fratesi, R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A. P. Seitsonen, A. Smogunov, P. Umari and R. M. Wentzcovitch, J. Phys.: Condens. Matter, 2009, 21, 395502.
47. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
48. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1998, 80, 891.
49. S. Grimme, J. Comput. Chem., 2006, 27, 1787–1799.
50. K. Momma and F. Izumi, J. Appl. Crystallogr., 2011, 44, 1272–1276.
51. H. Shen, T. Wang, H. Jiang, P. Zhao, Z. Chen, Y. Feng, Y. Cao, Y. Guo, Q. Zhang and H. Zhang, Appl. Catal. B, 2023, 339, 123140.
52. J. K. Nørskov, T. Bligaard, A. Logadottir, J. R. Kitchin, J. G. Chen, S. Pandelov and U. Stimming, J. Electrochem. Soc., 2005, 152, J23.
53. S. Wang, L. Li, J. Li, C. Yuan, Y. Kang, K. S. Hui, J. Zhang, F. Bin, X. Fan, F. Chen and K. N. Hui, J. Phys. Chem. C, 2021, 125, 7155–7165.
54. S. Zhou, W. Pei, J. Zhao and A. Du, Nanoscale, 2019, 11, 7734–7743.
55. NIST Chemistry WebBook, https://webbook.nist.gov/chemistry/, accessed 20 October 2025.
56. J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard and H. Jónsson, J. Phys. Chem. B, 2004, 108, 17886–17892.
57. S. Cao, Y. Hu, C. Yang, J. Li, H. Chen, S. Wei, S. Liu, Z. Wang, D. Sun and X. Lu, J. CO₂ Util., 2022, 62, 102074,  DOI:10.1016/j.jcou.2022.102074.
58. H. Zhou, Z. Chen, E. Kountoupi, A. Tsoukalou, P. M. Abdala, P. Florian, A. Fedorov and C. R. Müller, Nat. Commun., 2021, 12, 5510,  DOI:10.1038/s41467-021-25784-0.
59. B. Huang, N. Li, W. J. Ong and N. Zhou, J. Mater. Chem. A, 2019, 7, 27620–27631.
60. S. A. Khan, G. Rehman, I. Ahmad, M. Maqbool, C. Franchini and B. Amin, Chem. Phys. Lett., 2019, 731, 136614,  DOI:10.1016/j.cplett.2019.136614.
61. Y. W. Sun and J. Y. Liu, Phys. Chem. Chem. Phys., 2023, 25, 4773–4779.
62. B. H. Lee, J. Park, A. Kumar, S. Choi, D. H. Kim and S. Y. Lee, Mater. Today Commun., 2022, 33, 104809,  DOI:10.1016/j.mtcomm.2022.104809.
63. C. Wu, X. Li, J. Lu, Z. Ye, J. Zhang, T. Zhou, R. Sun, L. Chen, B. Lu and X. Pan, Appl. Phys. Lett., 2013, 103, 082109,  DOI:10.1063/1.4818728.
64. Z. Chen, J. W. Wang, Y. Song and X. Zuo, AIP Adv., 2017, 7, 105118,  DOI:10.1063/1.4998280.
65. C. J. Lombard, C. G. C. E. van Sittert, J. N. Mugo, C. Perry and D. J. Willock, Phys. Chem. Chem. Phys., 2023, 25, 6121–6130.
66. Z. Wei, Y. Li, P. Li, L. Pan, W. Li, X. Hu and Y. Gu, RSC Adv., 2024, 14, 6262–6269.
67. J. Zhao, N. Ma, T. Wang, N. Li, Y. Wang and J. Fan, J. Mater. Chem. A, 2022, 10, 21611–21621.
68. S. Cao, Y. Liu, Y. Hu, J. Li, C. Yang, Z. Chen, Z. Wang, S. Wei, S. Liu and X. Lu, J. Colloid Interface Sci., 2023, 642, 273–282.
69. S. Liu, Z. Zhou, J. Chen, Y. Fu and C. Cai, Appl. Surf. Sci., 2023, 611, 155645,  DOI:10.1016/j.apsusc.2022.155645.
70. X. Wang, H. Qin, Y. Chen and J. Hu, J. Phys. Chem. C, 2014, 118, 28548–28561.
71. X. Mao, G. Kour, L. Zhang, T. He, S. Wang, C. Yan, Z. Zhu and A. Du, Catal. Sci. Technol., 2019, 9, 6800–6807.
72. L. I. Bendavid and E. A. Carter, J. Phys. Chem. C, 2013, 117, 26048–26059.
73. F. Li and Q. Tang, Nanoscale, 2019, 11, 18769–18778.
74. J. T. Feaster, C. Shi, E. R. Cave, T. Hatsukade, D. N. Abram, K. P. Kuhl, C. Hahn, J. K. Nørskov and T. F. Jaramillo, ACS Catal., 2017, 7, 4822–4827.
75. J. Quílez-Bermejo and J. Herranz, J. Mater. Chem. A, 2025, 13, 41555,  10.1039/d5ta03854e.
76. Z. Yang, F. E. Oropeza and K. H. L. Zhang, APL Mater., 2020, 8, 060901,  DOI:10.1063/5.0004194.
77. S. Y. Wu, T. C. Chuang and H. T. Chen, Appl. Surf. Sci., 2025, 681, 161523,  DOI:10.1016/j.apsusc.2024.161523.
78. A. J. Medford, A. Vojvodic, J. S. Hummelshøj, J. Voss, F. Abild-Pedersen, F. Studt, T. Bligaard, A. Nilsson and J. K. Nørskov, J. Catal., 2015, 328, 36–42.
79. S. H. Zhou, W. Wei, X. Cai, D. D. Ma, S. M. Wang, X. Li and Q. L. Zhu, Adv. Funct. Mater., 2024, 34, 2311422,  DOI:10.1002/adfm.202311422.

---