Structure–activity relationship of Cu-based catalysts for the highly efficient CO2 electrochemical reduction reaction

Electrocatalytic carbon dioxide reduction (CO₂RR) is a relatively feasible method to reduce the atmospheric concentration of CO₂. Although a series of metal-based catalysts have gained interest for CO₂RR, understanding the structure–activity relationship for Cu-based catalysts remains a great challenge. Herein, three Cu-based catalysts with different sizes and compositions (Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs) were designed to explore this relationship by density functional theory (DFT). The calculation results show a higher degree of CO₂ molecule activation on CuNi₃@CNTs compared to that on Cu@CNTs and Cu₄@CNTs. The methane (CH₄) molecule is produced on both Cu@CNTs and CuNi₃@CNTs, while carbon monoxide (CO) is synthesized on Cu₄@CNTs. The Cu@CNTs showed higher activity for CH₄ production with a low overpotential value of 0.36 V compared to CuNi₃@CNTs (0.60 V), with *CHO formation considered the potential-determining step (PDS). The overpotential value was only 0.02 V for *CO formation on the Cu₄@CNTs, and *COOH formation was the PDS. The limiting potential difference analysis with the hydrogen evolution reaction (HER) indicated that the Cu@CNTs exhibited the highest selectivity of CH₄ among the three catalysts. Therefore, the sizes and compositions of Cu-based catalysts greatly influence CO₂RR activity and selectivity. This study provides an innovative insight into the theoretical explanation of the origin of the size and composition effects to inform the design of highly efficient electrocatalysts.

1 Introduction

With the rapid development of the global society, the increasing emissions of carbon dioxide (CO₂) from the burning of fossil fuels have led to serious environmental and energy crises (Jiao, 2019; Cao et al., 2021; Wang et al., 2022). The electrocatalytic CO₂ reduction reaction (CO₂RR) provides a viable alternative to convert CO₂ into value-added chemical products such as carbon monoxide (CO), methane (CH₄), methanol (CH₃OH), ethylene (CH₂CH₂), and ethanol (CH₃CH₂OH) (Xue et al., 2020; Zhang, 2022; Zhou et al., 2022). Therefore, this is a promising approach to alleviate the global demand for fossil fuels and greenhouse effects. In addition, the advantages of CO₂RR include 1) the feasibility of controlling the process by adjusting the potential and related conditions; 2) electricity consumption from sustainable sources such as the Sun and wind; 3) electrochemical reaction systems that are compact, modular, and easily scaled-up (Gong et al., 2019; Jiao et al., 2019; Bai et al., 2020; Han et al., 2020; He et al., 2020). However, there remain several challenges to the practical application of CO₂RR, such as slow reaction kinetics, poor efficiency, and low product selectivity due to the strong C=O double bond (806 kJ.mol^-1) in CO₂ molecules (Alvarez et al., 2017). Therefore, highly active electrocatalysts must be designed to improve the sluggish kinetics and energy conversion efficiency.

A recent study demonstrated the high activity and selectivity of transition metal catalysts in the conversion of CO₂ to CO, CH₄, or CH₃OH. Among these, Cu-based catalysts have been widely applied to CO₂RR due to their excellent conductivity and availability (Li et al., 2017). The catalyst size and composition structure are usually regulated to tune the catalytic activity and selectivity of metal nanoparticles. Varying particle size has size effects in CO₂RR. Studies have reported on the CO₂RR trends over spherical Cu nanoparticles ranging in size from 2 to 15 nm (Shen et al., 2021). A decrease in Cu nanoparticle size leads to a significant increase in low coordination sites, resulting in strong CO₂ adsorption and hydrogenation. A series of different sizes of Ni-based catalysts (Ni@CNT, Ni₄@CNT, and Ni (110)) were also designed to study their size effects in CO₂RR. Density functional theory (DFT) calculations revealed that the Ni@CNT catalyst is much more conducive to the conversion of CO₂ to CO, while CH₄ molecules were produced on Ni@CNT and Ni (100) with increasing Ni size (Cao et al., 2020a). Therefore, the catalyst size may affect CO₂ activation and final products.

The alloy effect of the metal catalyst is another key factor in catalytic reactions. The electronic structure of metal catalysts can be controlled by alloying with other metal atoms, which is an obvious characteristic of Cu-based catalysts (Weng et al., 2018; Meng et al., 2022). A Cu–Ni alloy catalyst embedded in the N-doped carbon framework (CuNi/NC) synthesized for CO₂RR showed excellent activity and selectivity for electrocatalytic CO₂RR to CO, with a high Faradaic efficiency of 99.7% and a potential of −0.6 V (Zhang, 2019; Zhong et al., 2020). Additionally, a non-noble metal CuSn alloy with a different Cu–Sn composition demonstrated a completely different distribution of products for CO₂RR, including CO and formate (Li et al., 2020a). Tuning the surface composition of Cu-based catalysts via alloying with other metals could directly modify the absorbed state of the intermediate species on their active sites and determine the final product during the CO₂RR. Therefore, the introduction of other metals is an effective strategy to adjust the electronic structure and improve CO₂RR activity and selectivity.

This study designed different sizes and compositions of Cu-based catalysts (Cu@CNT, Cu₄@CNT, and CuNi₃@CNT) to explore the relationship between the structure of catalysts and activity for CO₂RR based on DFT calculations. CO₂ molecule adsorption and activation are essential for the following hydrogenation steps. Therefore, different CO₂ adsorption configurations were optimized to explore the nature of CO₂ activation on the three catalysts by analyzing their geometric and electronic properties. The mechanisms of CO₂RR on different Cu-based catalysts were also assessed, and a plot was constructed to compare the selectivity of CO₂RR among Cu@CNT, Cu₄@CNT, and CuNi₃@CNT.

2 Computational models and methods

All calculations were performed using the Vienna Ab initio Simulation Package (VASP) (Kresse and Hafner, 1993; Kresse and Furthmiiller, 1996), including the Perdew–Burke–Ernzerhof (PBE) functions in the generalized gradient approximation (GGA) to describe electron exchange and correction effect (Blochl, 1994; Perdew et al., 1996; Kresse and Joubert, 1999). The cutoff energy for the plane wave basis was set to 520.0 eV. The Brillouin zone was sampled at 4 × 4 × 1 k points for all calculation models using the Monkhorst–Pack grid. The convergence criteria for energy for geometry optimization were set to 1.0 × 10⁻⁵ eV. The vacuum thickness was 15.0 Å to prevent interlayer interactions. van der Waals correction was implemented using the Grimme’s DFT-D3 method (Krieg, 2010).

The adsorption energy ($E_{\mathrm{a}\mathrm{d}\mathrm{s}}$) of CO₂ is calculated using the following equation:

$E_{\mathrm{a}\mathrm{d}\mathrm{s}}=E_{\mathrm{c}\mathrm{a}\mathrm{t}+\mathrm{m}\mathrm{o}\mathrm{l}}–E_{\mathrm{c}\mathrm{a}\mathrm{t}}–E_{\mathrm{m}\mathrm{o}\mathrm{l}},(1)$

where $E_{\mathrm{c}\mathrm{a}\mathrm{t}+\mathrm{m}\mathrm{o}\mathrm{l}}$ is the total energy of CO₂ molecule or intermediate adsorption on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs; $E_{\mathrm{c}\mathrm{a}\mathrm{t}}$ is the optimized energy of three catalysts; and $E_{\mathrm{m}\mathrm{o}\mathrm{l}}$ is the energy of the adsorbed molecule.

The binding energy ($E_{\mathrm{b}}$) of a single Cu atom, Cu₄, and CuNi₃ cluster anchored on N-doped CNT is defined as follows:

$E_{\mathrm{b}}=E_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}–E_{\mathrm{N}-\mathrm{C}\mathrm{N}\mathrm{T}}–E_{\mathrm{m}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l}},(2)$

where $E_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the total energy of Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs; $E_{\mathrm{N}-\mathrm{C}\mathrm{N}\mathrm{T}}$ denotes the optimized energy of N-doped CNT support; and $E_{\mathrm{m}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the energy of a single Cu atom or Cu₄ or the Cu₃Ni cluster.

The computational hydrogen electrode (CHE) mode was proposed by Nørskov et al (Nadolska et al., 2021), in which the changes in Gibbs free energies of CO₂RR are calculated for every elementary step. The reference electrode is the reversible hydrogen electrode (RHE). Therefore, the chemical potential (μ) of the proton–electron pair is defined as follows:

${\mathrm{\mu }}_{{\mathrm{H}}^{+}}+{\mu }_{{\mathrm{e}}^{-}}\to {\mathrm{\mu }}_{{\mathrm{H}}_2}\left(\mathrm{g}\right),(3)$

The change in Gibbs free energy ($∆G$) for each elementary step is determined by the following equation:

$∆G=∆E+∆E_{\mathrm{Z}\mathrm{P}\mathrm{E}}–T∆S+∆G_U,(4)$

where $∆E_{\mathrm{D}\mathrm{F}\mathrm{T}}$ is the change of electron energy calculated by the DFT calculation, $∆E_{\mathrm{Z}\mathrm{P}\mathrm{E}}$ is the change of zero-point correction energy, $∆S$ is the change of the entropy gradient, and T is the temperature.

$∆G_U=–neU,(5)$

where e is the fundamental charge transfer, n is the proton–electron logarithm transferred, and U is the electrode applied potential relative to the RHE.

The overpotential (η) of CO₂RR is defined as follows:

$\eta =U_{\mathrm{e}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{b}\mathrm{r}\mathrm{i}\mathrm{u}\mathrm{m}}–U_{\mathrm{l}\mathrm{i}\mathrm{m}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}},(6)$

where $U_{\mathrm{e}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{b}\mathrm{r}\mathrm{i}\mathrm{u}\mathrm{m}}$ is the equilibrium potential of CO₂RR. The limiting potential $U_{\mathrm{l}\mathrm{i}\mathrm{m}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}}$ is calculated as U_limiting = –ΔG/e, where ΔG_limiting represents the free energy of the potential-limiting step.

Carbon nanotubes (CNTs) are promising substitutes that have been widely applied in electrocatalysts for their high electrical conductivity, thermal stability, and abundant catalytic active sites (Grimme et al., 2010; Xiong et al., 2019; Bulmer et al., 2021; Cai et al., 2021). As an N-doped CNT (Figure 1A) may further increase interaction between CNT and the active center (Park et al., 2020; Nguyen and Shim, 2021), a single Cu atom, Cu₄ cluster, and CuNi₃ cluster were doped on the N-doped CNTs through the carbon vacancy defect to explore the effects on the CO₂RR sizes and composition (Figures 1B–D). As shown in Figure 1B, the single Cu atom is connected to two N atoms and one C atom to form a pyramid-like structure with E_b values of −3.21 eV. The bond lengths of Cu–N and Cu–C are 2.11 Å and 1.80 Å, respectively. When the Cu₄ cluster is doped on the N-doped carbon nanotubes, the three Cu atoms at the bottom interact with two N atoms and two C atoms (Figure 1C), resulting in a strong interaction with E_b values of −4.87 eV. The Cu–N, Cu–C, and Cu–Cu bond lengths are 2.05 Å, 2.03 Å, and 2.42 Å, respectively. In addition, CuNi₃@CNTs are used to study the compositional effect of Cu-based catalysts for CO₂RR by replacing three Cu atoms in Cu₄@CNTs with three Ni atoms connected to the adjacent N atoms and C atoms on the bottom of the tetrahedron, respectively (Figure 1D). The binding energy of CuNi₃@CNTs is −5.12 eV. The large binding energies indicate that the three different Cu-based catalysts all possess high thermodynamic stability, ensuring their long-term use.

FIGURE 1

FIGURE 1. Optimized structures of (A) nitrogen-doped carbon nanotubes, (B) single Cu atom (Cu@CNTs), (C) Cu₄ cluster (Cu₄@CNTs), and (D) CuNi₃ (CuNi₃@CNTs) supported on the nitrogen-doped carbon nanotubes.

3 Results and discussion

3.1 CO₂ activation on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs

CO₂ adsorption and activation are crucial steps for subsequent hydrogenation in CO₂RR (Jiao et al., 2019). To investigate the behaviors of CO₂ adsorption, we explored the geometric and electronic structures of the adsorbed CO₂ molecules. As shown in Figure 2A, the CO₂ molecule is adsorbed on the single Cu atom site with an E_ads value of −0.50 eV. The C=O bond length in CO₂ is also slightly elongated from 1.18 to 1.25 Å. With increasing Cu active center size (from a single Cu atom to the Cu₄ cluster), the E_ads value of CO₂ decreases to −0.19 eV. The CO₂ prefers to adsorb on the edge site of the Cu₄ cluster (Figure 2B) compared to the top site, with an E_ads value of 0.15 eV. The adsorption site and configuration of CO₂ on the CuNi₃@CNTs (Figure 2C) are the same as that on the Cu₄@CNTs, although the E_ads value increases to −0.82 eV. The C=O bond length also increases to 1.26 Å. The top adsorption site of CuNi₃@CNTs has an adsorption energy of 0.35 eV (Supplementary Figure S1). Thus, the top adsorption site was used for CO₂ molecule adsorption and activation on both the Cu₄@CNTs and CuNi₃@CNTs. The bond angles of O−C−O in the CO₂ molecule changed from 180.00° to 146.74°, 140.77°, and 139.23° on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs, respectively (Table 1). Thus, Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs all significantly activate CO₂, and the size and composition effect greatly influence the degree of CO₂ activation, which may further affect the activity and selectivity of CO₂ hydrogenation.

FIGURE 2

FIGURE 2. (A–C) Optimized adsorption configurations of CO₂ on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs. (D–F) Calculated charge density difference (CDD) and (G–I) the electronic local function diagram (ELF) of CO₂ adsorption. The charge depletion and accumulation are depicted in blue and yellow, respectively. The equivalent face value is 0.005 e/Å^–3.

TABLE 1

TABLE 1. Geometric and electronic structure parameters of adsorbed *CO₂ on the three Cu-based catalysts.

To further investigate the activation mechanism of the CO₂ molecules, the Bader charge, charge density difference (CDD), electronic local function diagram (ELF), projected density of state (PDOS), and integrated crystal orbital Hamilton population (ICOHP) with CO₂ adsorption were investigated. As shown in Figures 2D–F, significant charge transfers occur between C=O double bonds and metal Cu and Ni atoms connected to CO₂. In detail, the electron densities of the O atom, and between the C, O atoms and Cu, Ni atoms increase, while the electron density of the C=O double bond decreases significantly. In addition, the ELF map is used to further verify the bonding properties of C=O double bonds (Peterson et al., 2010; Reske et al., 2014; Wang et al., 2022). The results of the ELF calculation (Figures 2G–I) showed that the C=O double bond is weak, consistent with the results of the CDD analysis. The local electron densities of the Cu-C and Cu-O bonds range from 0.5 to 1.0 in the whole ELF map, which further verifies the metal coordination bond formation between Cu and Ni atoms and between the C and O atoms. The adsorbed *CO₂ molecule gains 0.44|e|, 0.67|e|, and 0.55|e| from the Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs, respectively, based on the Bader charge analysis, resulting in the bending of the bond angles of the adsorbed CO₂ molecules.

The PDOS of CO₂ adsorption are analyzed to further reveal interactions between Cu atoms and CO₂ (Figures 3A–C). The Cu-3d and Ni-3d orbitals play key roles in CO₂ adsorption and activation. The C-2p, O-2p, and Cu-3d orbitals have significant peak overlaps near the Fermi level, indicating strong interactions between C, O, and Cu atoms during CO₂ adsorption on Cu@CNTs and Cu₄@CNTs (Figures 3A,B). With the introduction of the Ni atom, the Cu-3d and Ni-3d orbitals show hybridization, illustrating the strong interaction between the Cu and Ni atoms. Meanwhile, the broadly overlapped regions by C-2p, O-2p, Cu-3d, and Ni-3d hybrids also appear near-Fermi level, with strong peak intensities (Figure 3C). Based on the CDD and PDOS results, the mechanism of CO₂ activation can be summarized as follows: anti-bonding orbitals of CO₂ accept 3d-electrons from the Cu-3d and Ni-3d orbitals; meanwhile, the π electrons from the CO₂ bonding orbitals revert to empty Cu-3d and Ni-3d orbitals. The activation strength of the C=O double bond of the adsorbed CO₂ molecule is quantitatively studied by calculating the ICOHP values, in which lower values indicate stronger bond strength (Li et al., 2021; Song et al., 2021). As seen in Figures 3D–F, the C=O double bond anti-bonding states move down relative to the Fermi level after CO₂ adsorption. Therefore, the strength of the C=O double bond significantly decreases after CO₂ adsorption. The ICOHP values of the C=O double bond are −4.47, −4.43, and −3.82 eV on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs, respectively. The weaker the C=O double bond, the more efficient the CO₂ activation on the catalyst surface. This result demonstrates the higher degree of CO₂ activation on the CuNi₃@CNTs compared to that of the Cu@CNTs and Cu₄@CNTs, consistent with the results of Bader charge, CDD, and adsorption energy calculations.

FIGURE 3

FIGURE 3. (A–C) Projected density of state (PDOS) plots. (D–F) Projected crystal orbital Hamilton population (pCOHP) of CO₂ adsorption on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs. The dotted line represents the Fermi energy level.

3.2 Mechanisms of CO₂RR on Cu-based catalysts

CO₂RR is a complex reaction involving multi-electron (two, six, or eight-electron) reaction pathways (Jiwanti and Einaga, 2019; Niu et al., 2021; Yang et al., 2022). The overall reaction equation can be expressed as follows:

${\mathrm{C}\mathrm{O}}_2+*+{2\mathrm{H}}^{+}+{2\mathrm{e}}^{–}\to \mathrm{C}\mathrm{O}+{\mathrm{H}}_2\mathrm{O}+*(7)$

${\mathrm{C}\mathrm{O}}_2+*+{6\mathrm{H}}^{+}+{6\mathrm{e}}^{–}\to {\mathrm{C}\mathrm{H}}_3\mathrm{O}\mathrm{H}+{\mathrm{H}}_2\mathrm{O}+*(8)$

${\mathrm{C}\mathrm{O}}_2+*+{8\mathrm{H}}^{+}+{8\mathrm{e}}^{–}\to {\mathrm{C}\mathrm{H}}_4+{2\mathrm{H}}_2\mathrm{O}+*(9)$

Scheme 1 shows the different reaction paths for CO₂RR to C1 products (CO, CH₄, and CH₃OH). According to the aforementioned description, the first step is CO₂ adsorption. Then, an electron from the catalyst surface attacks the adsorbed *CO₂ molecules to form the radical CO₂⁻, resulting in the bending of the angle of the adsorbed *CO₂. Subsequently, the radical CO₂⁻ intermediate combines with the H atom to form the key *COOH intermediate. The *COOH intermediate then is further hydrogenated to form *CO and H₂O. In Path 1, the adsorbed *CO diffuses to the gas phase. In the other paths, the adsorbed *CO species is further protonated to form *COH or *CHO via the formation of C-H or O-H bonds. Thus, CH₄ and CH₃OH products can be obtained by further hydrogenation of the key intermediates *COH or *CHO. In Path 2, CH₃OH is synthesized starting from the *COH intermediate. The proton/electron pair (H⁺/e⁻) continuously attaches the C atom of *COH to form the *CHOH, *CH₂OH, and *CH₃OH intermediates. *CH₃OH then desorbs to form the CH₃OH product. In contrast, CH₄ can be formed from *CHO or *COH via paths 3 and 4, respectively. In Path 3, the O atom of *COH is hydrogenated to form H₂O and *C. In the next four hydrogenation steps, *CH, *CH₂, *CH₃, and CH₄ are continuously generated. Moreover, the *CHO intermediate can eventually be protonated to produce CH₄ by the *CH₂O, *CH₃O, O* + CH₄, *OH, and *H₂O species (Path 4). In addition, the C₂ products also can be synthesized on the Cu-based catalysts. However, when two CO₂ molecules are co-adsorbed on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs, only one CO₂ molecule can be adsorbed on the single-atom Cu, Cu₄ cluster, and CuNi₃ cluster; the other CO₂ molecule diffuses to the gas phase (Supplementary Figure S2). Therefore, this study explores only four paths for CO₂RR to C1 products.

SCHEME 1

SCHEME 1. Proposed mechanisms of CO₂RR to C₁ products. *represents the catalytic surface.

3.3 Reaction mechanism for CO₂RR to CH₄ products on Cu@CNTs

To illustrate the catalytic activity and selectivity of CO₂RR on Cu@CNTs, the free energy barrier diagram and corresponding intermediate configurations are shown in Figures 4, 5, respectively. Figure 4 shows that the absorbed CO₂ molecule is hydrogenated to form stable *COOH with a change of the free energy value (ΔG) of 0.47 eV. In the second hydrogenation step, *COOH can react with the (H⁺/e⁻) pair to form *CO and H₂O, with an exothermic energy release of –0.27 eV. The adsorbed *CO is further protonated to form *CHO (ΔG = 0.61 eV) and *COH (ΔG = 1.49 eV) intermediates, or may be directly desorbed with a ΔG value of 0.91 eV. Obviously, the formation of *CHO is more favorable than that of *CHO and *CO desorption. When the *COH species is formed, the C atom of *COH accepts three H atoms to form *CH₃OH with downhill free energies of −0.47, −1.18, and −0.08 eV. *COH is also directly dissociated into *C species (ΔG = 0.60 eV), and *CH₄ is formed by four hydrogenation steps of the *C intermediate. The formation of *CH and *CH₂ is slightly thermodynamically favorable, with ΔG values of −1.22 and −1.83 eV, respectively. However, the formation of *CH₃ has an endothermic energy of 0.78 eV. It is difficult to form the CH₄ molecule in Path 3 due to the high hydrogenation barrier of the formation of *COH species for the whole reaction. In Path 4, *CHO intermediate accepts an (H⁺/e⁻) pair to form *CH₂O with a ΔG value of −1.15 eV. Then, the hydrogenation of *CH₂O to *CH₃O is uphill by 0.26 eV. *CH₃O is further hydrogenated to form *O and a CH₄ molecule with a ΔG value of −0.75 eV. The leaving *O atom is continuously hydrogenated to form H₂O. Thus, the conversion of CO₂ to CH₄ following Path 4 with an overpotential value of 0.37 V is easier than that of CO and CH₃OH in all four paths, and the potential-determining step (PDS) is the *CHO formation on the Cu@CNTs. Wang et al. prepared three Cu-based catalysts for CO₂RR (Weng et al., 2018). Among them, the single-atom Cu phthalocyanine catalyst exhibited the highest activity to produce CH₄ with a Faradaic efficiency of 66% and a partial current density of 13 mA cm⁻² at a potential of −1.06 V versus RHE, findings that are consistent with the calculation results of CO₂RR to CH₄ on the Cu-based single-atom catalysts in the present study.

FIGURE 4

FIGURE 4. Free energy diagram of CO₂RR on Cu@CNTs.

FIGURE 5

FIGURE 5. Optimized structures of intermediates for CO₂RR on Cu@CNT surfaces.

3.4 Reaction mechanism for CO₂RR to CO products on Cu₄@CNTs

To explore the different size effects on CO₂RR activity on Cu-based catalysts, the free energy profiles and optimized geometric structures for CO₂RR on the Cu₄@CNTs are shown in Figures 6, 7, respectively. The Cu₄ clusters embedded in the CNTs catalyst provided multiple active sites for the activation of CO₂ molecules compared to Cu@CNTs and further promoted hydrogenation. The adsorbed *CO₂ accepts two (H⁺/e⁻) pairs to successively convert the key intermediates *COOH and *CO with an entirely endothermic process of 0.54 eV and 0.16 eV, respectively. CO desorption is a thermodynamically favorable process with a low energy of −0.01 eV compared to *CO hydrogenation to *COH (ΔG = 1.32 eV) and *CHO (ΔG = 0.62 eV). In the subsequent hydrogenation steps, the (H⁺/e⁻) pair consecutively attacks the C atom of the *CO intermediates to form *CHO, *CH₂O, *CH₃O, and *O+CH₄ species, with PDS, the process of *O+ CH₄ intermediate formation, with ΔG values of 1.32 eV in Path 4. After *COH formation, *COH hydrogen dissociates to form *C with an uphill energy of 0.72 eV. Then, the *C species is further hydrogenated to form *CH, *CH₂, *CH₃, and *CH₄ (Path 3). The corresponding ΔG values are −0.21, −0.74, −1.11, and −0.44 eV, respectively. In Path 2, *COH is hydrogenated to form *CHOH, *CH₂OH, and *CH₃OH species, with ΔG values of −0.17, −0.43, and −0.86 eV. Comparison of the ΔG values of PDS in all four paths shows that the reactions are more likely to follow Path 1 and CO is produced on the Cu₄@CNTs with an extremely low overpotential value of 0.02 V. More importantly, the desorption energy of CO is greatly reduced with increasing Cu size, which is conducive to the production of CO and avoids poisoning the active sites on the catalyst surface. In addition, the reported atom-pair catalyst with stable Cu₁⁰–Cu₁^x+ pair structures also shows high selectivity and activity for CO₂RR to CO (Li et al., 2020b). Therefore, our calculation results of CO₂RR on cluster catalysts are consistent with experimental results reported previously. These results demonstrate that CO₂RR reduction product changes from CH₄ to CO when the size of Cu–based catalysts changes from a single atom to a Cu₄ cluster. Therefore, designing multiple active sites to change the final reduction products may be a universal strategy for catalyst development.

FIGURE 6

FIGURE 6. ree energy diagram of CO₂RR on Cu₄@CNTs.

FIGURE 7

FIGURE 7. Optimized structure of intermediates for CO₂RR on Cu₄@CNT surfaces.

3.5 Reaction mechanism for CO₂RR to CH₄ products on CuNi₃@CNTs

To explore the composition effect on the mechanism of CO₂RR, the synthesis paths of CO₂ into CH₄, CO, and CH₃OH products were determined on CuNi₃@CNTs. The free energy and geometric structure diagrams are shown in Figures 8, 9, respectively. The *COOH intermediate is obtained in the first step of hydrogenation of the adsorbed *CO₂ molecule, with a downhill energy of −0.44 eV on the CuNi₃@CNTs catalyst, which differs from that on Cu@CNTs and Cu₄@CNTs. According to the aforementioned charge transfer analysis, the adsorbed *CO₂ molecule obtains more electrons from CuNi₃@CNTs compared to Cu@CNTs and Cu₄@CNTs, and the electrons of Ni-3d orbitals are partly transferred to the π anti-bonding orbitals of CO₂. These strong electron transfers lead to the spontaneous formation of a *COOH intermediate. The production of *CO occurs through *COOH hydrogenation to *CO and H₂O molecules. However, *CO desorption requires a large desorption energy of 1.57 eV (Path 1). Thus, *CO is further hydrogenated to *COH (ΔG = 1.83 eV) or *CHO (ΔG = 0.84 eV). In the subsequent hydrogenation steps, *CHO is hydrogenated in two steps to form *CH₂O and *CH₃O in Path 4. Then, the (H⁺/e⁻) pair attacks the C atom in *CH₃O, resulting in CH₄ formation. After release of the CH₄ molecule, one *O atom remains above the bridge site of the Cu–Ni atoms (Figure 9) and is subsequently further hydrogenated into H₂O molecules and released. The reactions are more likely to follow Path 2 to form CH₃OH than Path 3 to form CH₄ based on a comparison of the ΔG values of producing possible species (*CHOH, *CH₂OH; *C, *CH₂, and *CH₃). As discussed previously, CO₂RR is more likely to follow Path 4 to form CH₄ on CuNi₃@CNTs, with an overpotential value of 0.60 V and *COH formation being the PDS. For the Cu–Ni alloy catalysts, the major products of CO₂RR change from ethylene (C₂H₄) and C₂H₅OH to CH₄ with increasing Ni content (Park et al., 2021). These selectivity challenges are attributed to the increased adsorption energies of *CO on the Ni surface based on the results of pressure-dependent experiments and DFT calculations. We observed similar findings in our calculations. The E_ads values of *CO on the CuNi₃@CNTs are much higher than those on the Cu@CNTs and Cu₄@CNTs. Thus, the introduction of Ni atoms in the CuNi₃@CNTs changes the selectivity of CO₂RR to produce CH₄ compared to that on the Cu₄@CNTs, which produce CO.

FIGURE 8

FIGURE 8. Free energy diagram of CO₂RR on CuNi₃@CNTs.

FIGURE 9

FIGURE 9. Optimized structure of intermediates for CO₂RR on the CuNi₃@CNT surfaces.

3.6 Competitive relationship of CO₂RR vs. hydrogen evolution reaction (HER)

The hydrogen evolution reaction (HER) is usually considered a competitive reaction during CO₂RR. Therefore, the selectivity and activity of CO₂RR are tightly related to HER performance (Xue et al., 2020; Zhao and Liu, 2020). Figure 10A shows the free energies of HER for Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNs. HER occurs on three catalysts with ΔG values of −1.12, 0.24, and 0.57 eV, respectively. The difference in limiting potentials for CO₂RR and HER on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs were also calculated to further understand the selectivity of the catalyst for CH₄ or CO formation. The optimized configurations of *H adsorption are shown in Supplementary Figure S3. The higher the U_L (CO₂)–U_L (H₂) difference, the higher is the selectivity for CO₂RR over HER (Shen et al., 2021; Wei et al., 2021). As shown in Figure 10B, the selectivity toward CO₂RR decreases in the order of Cu@CNTs (CH₄ product) > CuNi₃@CNTs (CH₄ product) > Cu₄@CNTs (CO product). The Cu@CNTs have only one adsorption site for the intermediates, while the CuNi₃@CNTs and Cu₄@CNTs have multiple adsorption sites (top, edge, and hollow sites). Thus, smaller-sized Cu@CNTs show the highest CH₄ selectivity among the three catalyst types. However, the large size of the bulk Ni (110) facet shows the highest selectivity for CH₄ compared to Ni@CNTs and Ni₄@CNTs among the Ni-based catalysts (Cao, 2020b). In addition, the selectivity of CH₄ on the CuNi₃@CNTs is higher than that for CO on the Cu₄@ CNTs. Therefore, the sizes and composition of Cu-based catalysts have a significant role in CO₂RR selectivity.

FIGURE 10

FIGURE 10. (A) Free energy diagram of the HER on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs. (B) Differences in limiting potentials for CO₂RR and HER on Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs.

4 Conclusion

In summary, the size and composition effect for the activity and selectivity of CO₂RR were explored using DFT. Three types of Cu-based catalysts (Cu@CNTs, Cu₄@CNTs, and CuNi₃@CNTs) were designed to achieve highly efficient CO₂ reduction. The CuNi₃@CNTs show the largest CO₂ adsorption energy (−0.82 eV) due to CO₂ molecule adsorption on the edge site between the Ni and Cu atoms. This promotes higher charge transfer to CO₂ molecules and weakens the C=O double bond of CO₂. With a size increase from single-atom Cu to the Cu₄ cluster, the CO₂RR product changes from CH₄ to CO. More importantly, Cu₄@CNTs show an extremely low overpotential of 0.02 V for CO formation, while Cu@CNTs exhibit the highest selectivity for CH₄ formation among the three catalysts. When the Ni atoms are introduced in the Cu₄ cluster, the CO₂RR selectivity is improved compared to HER. The activity and selectivity of the Cu-based catalysts for the CO₂RR depended strongly on the size and composition. Our results provide new insight into understanding the size and alloy effect of CO₂RR on Cu-based catalysts.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding authors.

Author contributions

All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.

Funding

This work is financially supported by the National Natural Science Foundation of China (grant no. 22108093), the National Key Research and Development Program of China (grant no. 2018YFB1502900), the Natural Science Foundation of Zhejiang Province (grant nos. LY19B060006 and LQ20E030016), and the Innovation Jiaxxing Elite Leadership Plan and Technology Development Project of Jiaxing University (70518040).

Acknowledgments

The authors also acknowledge support from the G60 STI Valley Industry & Innovation Institute, Jiaxing University, and appreciate the help provided by J. H. Chen

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors, and the 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.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2023.1141453/full#supplementary-material

References

Alvarez, A., Borges, M., Corral-Perez, J. J., Olcina, J. G., Urakawa, A., Cornu, D., et al. (2017). CO₂ activation over catalytic surfaces. Chemphyschem 18, 3135–3141. doi:10.1002/cphc.201700782

PubMed Abstract | CrossRef Full Text | Google Scholar

Bai, X., Li, Q., Shi, L., Niu, X., Ling, C., and Wang, J. (2020). Hybrid Cu(0) and Cu(x) as atomic interfaces promote high-selectivity conversion of CO₂ to C₂H₅OH at low potential. Small 16, e1901981. doi:10.1002/smll.201901981

PubMed Abstract | CrossRef Full Text | Google Scholar

Blochl, P. E. (1994). Projector augmented-wave method. Phys. Rev. B 50, 17953–17979. doi:10.1103/physrevb.50.17953

PubMed Abstract | CrossRef Full Text | Google Scholar

Bulmer, J. S., Kaniyoor, A., and Elliott, J. A. (2021). A meta-analysis of conductive and strong carbon nanotube materials. Adv. Mater 33, e2008432. doi:10.1002/adma.202008432

PubMed Abstract | CrossRef Full Text | Google Scholar

Cai, Y., Fu, J., Zhou, Y., Chang, Y. C., Min, Q., Zhu, J. J., et al. (2021). Insights on forming N,O-coordinated Cu single-atom catalysts for electrochemical reduction CO₂ to methane. Nat. Commun. 12, 586. doi:10.1038/s41467-020-20769-x

PubMed Abstract | CrossRef Full Text | Google Scholar

Cao, C., Zhao, C., Fang, Q., ZhongWang, J., Zhuang, G., Deng, S., et al. (2020b). Hydrogen peroxide electrochemical synthesis on hybrid double-atoms (Pd-Cu) doped N vacancy g-C₃N₄: A novel design strategy for electrocatalysts screening. J. Mat. Chem. A 8, 2672–2683. doi:10.1039/C9TA12468C

CrossRef Full Text | Google Scholar

Cao, C., Ma, D. D., Gu, J. F., Xie, X., Zeng, G., Li, X., et al. (2020a). Metal-organic layers leading to atomically thin bismuthene for efficient carbon dioxide electroreduction to liquid fuel. Angew. Chem. Int. Ed. Engl. 59, 15014–15020. doi:10.1002/anie.202005577

PubMed Abstract | CrossRef Full Text | Google Scholar

Cao, Y., Zhao, J., Zhong, X., Yao, Z., Zhuang, G., Wang, J., et al. (2021). Building highly active hybrid double–atom sites in C₂N for enhanced electrocatalytic hydrogen peroxide synthesis. Green Energy Environ. 6, 846–857. doi:10.1016/j.gee.2020.12.006

CrossRef Full Text | Google Scholar

Gong, Q., Ding, P., Xu, M., Zhou, W., Li, Y., Deng, J., et al. (2019). Structural defects on converted bismuth oxide nanotubes enable highly active electrocatalysis of carbon dioxide reduction. Nat. Commun. 10, 2807. doi:10.1038/s41467-019-10819-4

PubMed Abstract | CrossRef Full Text | Google Scholar

Grimme, S., Antony, J., Ehrlich, S., and Krieg, H. (2010). A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 132, 154104. doi:10.1063/1.3382344

PubMed Abstract | CrossRef Full Text | Google Scholar

Han, L., Song, S., Liu, M., Wu, Q., Luo, J., Xin, H. L., et al. (2020). Stable and efficient single-atom Zn catalyst for CO₂ reduction to CH₄. J. Am. Chem. Soc. 142, 12563–12567. doi:10.1021/jacs.9b12111

PubMed Abstract | CrossRef Full Text | Google Scholar

He, T., Zhang, L., Kour, G., and Du, A. (2020). Electrochemical reduction of carbon dioxide on precise number of Fe atoms anchored graphdiyne. J. CO₂ Util. 37, 272–277. doi:10.1016/j.jcou.2019.12.025

CrossRef Full Text | Google Scholar

Jiao, J. Q., Lin, R., Liu, S. J., Cheong, W. C., Li, Y. D., Chen, Z., et al. (2019). Copper atom-pair catalyst anchored on alloy nanowires for selective and efficient electrochemical reduction of CO₂. Nat. Chem. 11, 222–228. doi:10.1038/s41557-018-0201-x

PubMed Abstract | CrossRef Full Text | Google Scholar

Jiwanti, P. K., and Einaga, Y. (2019). Electrochemical reduction of CO₂ using palladium modified boron-doped diamond electrodes: Enhancing the production of CO. Phys. Chem. Chem. Phys. 21, 15297–15301. doi:10.1039/C9CP01409H

PubMed Abstract | CrossRef Full Text | Google Scholar

Kresse, G., and Furthmiiller, J. (1996). Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mat. Sci. 6, 15–50. doi:10.1016/0927-0256(96)00008-0

CrossRef Full Text | Google Scholar

Kresse, G., and Hafner, J. (1993). Ab initio molecular dynamics for liquid metals. Phys. Rev. B 47, 558–561. doi:10.1103/PhysRevB.47.558

PubMed Abstract | CrossRef Full Text | Google Scholar

Kresse, G., and Joubert, D. (1999). From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758–1775. doi:10.1103/PhysRevB.59.1758

CrossRef Full Text | Google Scholar

Li, Q., Ouyang, Y., Lu, S., BaiZhang, X. Y., Shi, L., Ling, C., et al. (2020a). Perspective on theoretical methods and modeling relating to electro-catalysis processes. Chem. Commun. 56, 9937–9949. doi:10.1039/D0CC02998J

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, Q., Wang, H., Luo, W., Sherrell, P. C., Chen, J., and Yang, J. (2020b). Heterogeneous single-atom catalysts for electrochemical CO₂ reduction reaction. Adv. Mater 32, e2001848. doi:10.1002/adma.202001848

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, S., Nagarajan, A. V., Alfonso, D. R., Sun, M., Kauffman, D. R., Mpourmpakis, G., et al. (2021). Boosting CO₂ electrochemical reduction with atomically precise surface modification on gold nanoclusters. Angew. Chem. Int. Ed. Engl. 60, 6351–6356. doi:10.1002/anie.202016129

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, X., Bi, W., Chen, M., Sun, Y., Ju, H., Yan, W., et al. (2017). Exclusive Ni-N₄ sites realize near-unity CO selectivity for electrochemical CO₂ reduction. J. Am. Chem. Soc. 139, 14889–14892. doi:10.1021/jacs.7b09074

PubMed Abstract | CrossRef Full Text | Google Scholar

Meng, Y., Xu, Z., Shen, Z., Xia, Q., Cao, Y., Wang, Y., et al. (2022). Understanding the water molecule effect in metalfree B-based electrocatalysts for electrochemical CO₂ reduction. J. Mat. Chem. A 10, 6508–6522. doi:10.1039/D1TA10127G

CrossRef Full Text | Google Scholar

Nadolska, M., Przesniak-Welenc, M., Lapinski, M., and Sadowska, K. (2021). Synthesis of phosphonated carbon nanotubes: New insight into carbon nanotubes functionalization. Materials 14, 2726. doi:10.3390/ma14112726

PubMed Abstract | CrossRef Full Text | Google Scholar

Nguyen, A. T. N., and Shim, J. H. (2021). All carbon hybrid N-doped carbon dots/carbon nanotube structures as an efficient catalyst for the oxygen reduction reaction. RSC Adv. 11, 12520–12530. doi:10.1039/D1RA01197A

PubMed Abstract | CrossRef Full Text | Google Scholar

Niu, H., Zhang, Z., Wang, X., Wan, X., Kuai, C., and Guo, Y. (2021). A feasible strategy for identifying single-atom catalysts toward electrochemical NO-to-NH₃ conversion. Small 17, e2102396. doi:10.1002/smll.202102396

PubMed Abstract | CrossRef Full Text | Google Scholar

Park, B. J., Kim, J., and Park, H. S. (2020). Bifunctional electrocatalysts based on hierarchical graphene/iron hybrid architectures branched by N-doped CNT. J. Alloys Compd. 846, 156244. doi:10.1016/j.jallcom.2020.156244

CrossRef Full Text | Google Scholar

Park, J., Cho, M., Rhee, Y. M., and Jung, Y. (2021). Theoretical study on the degree of CO₂ activation in CO₂-coordinated Ni(0) complexes. ACS Omega 6, 7646–7654. doi:10.1021/acsomega.0c06257

PubMed Abstract | CrossRef Full Text | Google Scholar

Perdew, J. P., Burke, K., and Ernzerhof, M. (1996). Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868. doi:10.1103/PhysRevLett.77.3865

PubMed Abstract | CrossRef Full Text | Google Scholar

Peterson, A. A., Pedersen, F. A., Studt, F., Rossmeisl, J., and Nørskov, J. K. (2010). How copper catalyzes the electroreduction of carbon dioxide into hydrocarbon fuels. Energy Environ. Sci. 3, 1311–1315. doi:10.1039/c0ee00071j

CrossRef Full Text | Google Scholar

Reske, R., Mistry, H., Behafarid, F., Cuenya, B. R., and Strasser, P. (2014). Particle size effects in the catalytic electroreduction of CO₂ on Cu nanoparticles. J. Am. Chem. Soc. 136, 6978–6986. doi:10.1021/ja500328k

PubMed Abstract | CrossRef Full Text | Google Scholar

Shen, X., Liu, X., Wang, S., Chen, T., Zhang, W., Cao, L., et al. (2021). Synergistic modulation at atomically dispersed Fe/Au interface for selective CO₂ electroreduction. Nano Lett. 21, 686–692. doi:10.1021/acs.nanolett.0c04291

PubMed Abstract | CrossRef Full Text | Google Scholar

Song, H., Tan, Y. C., Kim, B., Ringe, S., and Oh, J. (2021). Tunable product selectivity in electrochemical CO₂ reduction on well-mixed Ni−Cu alloys. ACS Appl. Mat. Interfaces. 13, 55272–55280. doi:10.1021/acsami.1c19224

CrossRef Full Text | Google Scholar

Wang, F., Meng, Y., Chen, X., Zhang, L., Li, G., Shen, Z., et al. (2022). Effect of nickel-based electrocatalyst size on electrochemical carbon dioxide reduction: A density functional theory study. J. Colloid Interface Sci. 615, 587–596. doi:10.1016/j.jcis.2022.02.032

PubMed Abstract | CrossRef Full Text | Google Scholar

Wei, N., Tian, Y., Liao, Y., Komatsu, N., Kauppinen, E. I., Lyuleeva-Husemann, A., et al. (2021). Colors of single-wall carbon nanotubes. Adv. Mater 33, e2006395. doi:10.1002/adma.202006395

PubMed Abstract | CrossRef Full Text | Google Scholar

Weng, Z., Wu, Y. S., Wang, M. Y., Jiang, J. B., Wang, H. L., Huo, S., et al. (2018). Active sites of copper-complex catalytic materials for electrochemical carbon dioxide reduction. Nat. Commun. 9, 415. doi:10.1038/s41467-018-02819-7

PubMed Abstract | CrossRef Full Text | Google Scholar

Xiong, W., Yang, J., Shuai, L., Qiu, M., Li, X., Hou, Y., et al. (2019). CuSn alloy nanoparticles on nitrogen-doped graphene for electrocatalytic CO₂ reduction. ChemElectroChem 6, 5951–5957. doi:10.1002/celc.201901381

CrossRef Full Text | Google Scholar

Xue, D., Xia, H., Yan, W., Zhang, J., and Mu, S. (2020). Defect engineering on carbon-based catalysts for electrocatalytic CO₂ reduction. Nanomicro Lett. 13, 5. doi:10.1007/s40820-020-00538-7

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, T., Tian, N., Liu, X. H., and Zhang, X. (2022). CuNi alloy nanoparticles embedded in N-doped carbon framework for electrocatalytic reduction of CO₂ to CO. J. Alloys Compd. 904, 164042. doi:10.1016/j.jallcom.2022.164042

CrossRef Full Text | Google Scholar

Zhang, E., Wang, T., D., Zhang, J., Li, Y., Chen, W., Li, A., et al. (2019). Bismuth single atoms resulting from transformation of metal-organic frameworks and their use as electrocatalysts for CO₂ reduction. J. Am. Chem. Soc. 141, 16569–16573. doi:10.1021/jacs.9b08259

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, Y., Yang, R., Li, H., and Zeng, Z. (2022). Boosting electrocatalytic reduction of CO₂ to HCOOH on Ni single atom anchored WTe₂ monolayer. Small 18, e2203759. doi:10.1002/smll.202203759

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhao, X., and Liu, Y. (2020). Unveiling the active structure of single nickel atom catalysis: Critical roles of charge capacity and hydrogen bonding. J. Am. Chem. Soc. 142, 5773–5777. doi:10.1021/jacs.9b13872

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhong, M., Tran, K., Min, Y., Wang, C., Wang, Z., Ulissi, Z., et al. (2020). Accelerated discovery of CO₂ electrocatalysts using active machine learning. Nature 581, 178–183. doi:10.1038/s41586-020-2242-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhou, J., Wang, X., Ge, K., Yang, Z., Xia, L., Guo, C., et al. (2022). Core-shell structured nanocomposites formed by silicon coated carbon nanotubes with anti-oxidation and electromagnetic wave absorption. J. Colloid Interface Sci. 607, 881–889. doi:10.1016/j.jcis.2021.09.022

PubMed Abstract | CrossRef Full Text | Google Scholar