Chemical and electrochemical hydrogenation of CO₂ to hydrocarbons on Cu single crystal surfaces: insights into the mechanism and selectivity from DFT calculations

Abstract

Atomic level mechanistic insights into the chemical and electrochemical reduction of CO₂ on the Cu(111) and Cu(100) surfaces are presented based on DFT-based thermodynamic and kinetic calculations. On Cu(111), CO_ads is firstly formed by the dissociative hydrogenation of CO₂, and CHO_ads and CH₂O_ads are the key intermediates towards the chemical and electrochemical reduction of CO₂ into methanol and CH₄. Despite being thermodynamically or kinetically favoured, it is likely that CH₂OH_ads instead of CH₃O_ads is the intermediate for methanol and CH₄ formation. Based on the activation barriers, CH₂OH_ads intermediate either forms CH₃OH by direct hydrogenation or forms CH_2ads by hydrogenative dissociation, which may be a parallel path in the CO₂ reduction mechanism on the Cu(111) surface. Finally, the CH₂ intermediate leads to formation of the hydrocarbons. On Cu(100), CO₂ reduction takes a different pathway in the early stages; CO is formed through the direct dissociation of CO₂ rather than hydrogenative dissociation as on the Cu(111) surface due to stronger bonding of CO. Further reduction of CO also undergoes a different pathway, in which CO dimerization is more easy to achieve, whereas CO hydrogenation is difficult on the Cu(100) surface. This explains why C₂H₄ is formed more favorably on the Cu(100) surface and CH₄ is predominantly produced on the Cu(111) surface under both chemical and electrochemical conditions. In addition, DFT results showed for the first time that the electrochemical reduction would be expected to be highly favored at potentials of interest by CO₂ reduction compared with chemical reduction and that the carbon dioxide anion radical (˙CO₂⁻) is involved in the initial stage of CO₂ electroreduction. Simultaneously, the results also explain partly why CH₃OH is formed in gas phase chemistry and only CH₄ is observed in electrochemistry on copper surfaces. By analyzing the chemical and electrochemical reduction paths, important mechanistic information is deduced on the Cu single crystal surfaces. The study of CO₂ reduction mechanisms on copper will lead to a deeper understanding of the reaction chemistry and could eventually lead to the design of more efficient and selective catalysts.

---

1. Introduction

Converting CO₂ into organic fuels and/or chemicals would positively impact the global carbon balance and offer a novel solution to the dilemma of growing energy demands and global warming we face nowadays.^1,2 Among various CO₂ fixation methods, reduction with hydrogen gas^3,4 or electricity^1,5–9 may represent the most promising pathways since either of them can be generated on large scales by sustainable energies such as solar, hydro and wind; moreover, they provide ways of storing these intermittently available sustainable energies. The two reduction methods similarly involve the successive hydrogenation of CO₂ molecules, leading to a variety of products such as formic acid, methanol and hydrocarbons. In comparison, the electrochemical reduction of CO₂ offers several advantages such as room temperature and ambient pressure operation, tunable reaction rate and selectivity by electrode potential, and the capability of producing hydrocarbons such as methane (CH₄) and ethylene (C₂H₄),^1,5–13 which are particularly desired due to their high energy densities and widespread use in the current energy and industrial infrastructures. Among electrocatalyst materials explored to date, metallic Cu is the most capable of converting CO₂ into hydrocarbons with high Faraday efficiency.^{5,7,8,10–19} Therefore, the electrochemical reduction of CO₂ on Cu-based materials has received growing interest in recent years.

Although CO₂ can be converted into CH₄ and C₂H₄ with high Faraday efficiency on Cu electrodes, considerable overpotentials are required for the reaction to occur, with onset potentials being around −0.7 V and −0.8 V (vs. RHE), respectively, for the formation of C₂H₄ and CH₄.^5–7,20 At less negative potentials, formate and CO are the main products. Studies on single crystal surfaces showed that more C₂H₄ was formed on the Cu(100) surface, whereas the Cu(111) surface produced more CH₄.^7,21,22 Understanding the atomic-level origin of the product distribution and the surface structure effect in CO₂ electroreduction are crucial for electrocatalyst optimization. This requires a detailed microscopic view on the reaction pathways and kinetics. It has been found that the electroreduction of CO exhibits very similar potential and surface structure dependence of the CH₄ and C₂H₄ formation, while experiments starting with HCOOH showed no detectable products,^20,23 which indicates that CO is one of the intermediates involved in CO₂ electroreduction to these hydrocarbons and CO is formed in a separate pathway from formate formation.^{8,13,21,24,25} However, the reaction pathways and intermediates for CO formation remain unclear.

Based on the reaction free energies calculated using density functional theory (DFT) for various possible elementary steps, Nørskov and coworkers^26,27 recently proposed that the formation of CH₄ and C₂H₄ from CO₂ on various fcc Cu facets similarly involves the hydrogenative reduction of CO to an adsorbed formyl (CHO*) as the rate-determining step (rds), which is followed by successive hydrogenative reduction steps to form adsorbed formaldehyde (CH₂O*) and methoxy species (CH₃O).²⁶ However, experimental results obtained recently by Koper and coworkers²⁸ and earlier by Hori et al.²⁵ seem to suggest that the reaction pathways of CH₄ and C₂H₄ formation are separated at an early stage of CO reduction. Results from a more recent experimental study of CO electroreduction on single-crystal copper electrodes by Koper and coworkers²⁹ further implied that there are two separate pathways for C₂H₄ formation: one that shares an intermediate with the pathway to CH₄ and one that occurs mainly on Cu(100) and probably involves the formation of a CO dimer as the key intermediate.⁸ The latest DFT calculation study by Nie et al.³⁰ suggested an alternative reaction path for both CH₄ and C₂H₄ production through a hydroxymethylidyne (COH) intermediate from CO electroreduction. These controversial results and propositions have left both the rate-determining and selectivity-determining steps for the formation of hydrocarbon products from CO₂ and CO electroreduction on Cu still uncertain. In addition, several other questions in the reaction mechanism remain, such as why methanol, i.e. the dominant product on copper catalysts in chemical reduction, is nearly absent in the electrochemical reduction; whether methylene (:CH₂) proposed in earlier studies⁸ is involved in the hydrocarbon formation; and why C₂H₄ formation is favored over CH₄ at lower overpotentials.

In an attempt to ascertain the dominant reaction pathways for the formation of various products in CO₂ reduction, we present a systematic DFT calculation on the reaction free energies and activation barriers of various possible elementary steps on the two most common types of Cu surface, namely, the closed-packed Cu(111) surface and the open Cu(100) surface. Both chemical reduction by gaseous hydrogen and electrochemical reduction through proton-electron pairing are considered and their similarities and differences are discussed. We calculated the free energies of the reaction and the activation barriers for the chemical reduction, according to which the optimized pathways for chemical reduction are deduced. With the theoretical hydrogen electrode model, the effect of electrode potentials on the steps in these pathways is discussed. Based on the calculated reaction free energy, the pathways with water in the double layer as the hydrogenation agent in the electrochemical environment are discussed. A mechanism for the electrochemical reduction is proposed. In chemical reduction, the adsorbed hydrogen atoms (H*) should be the main reducing and hydrogenation agent, while in electrochemical reduction, the hydrogenative reduction can be either accomplished by H* or by protons from the double layer plus electrons from electrode. Thermodynamically, the two ways are not distinguished, but they could be different kinetically as the transition states in the two cases could be totally different.

2. Computational method and modeling

Calculations were performed in the framework of DFT using the generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof (PBE)³¹ and employing ultrasoft pseudopotentials³² for nuclei and core electrons. Calculations of reaction free energies were performed using periodic super-cells with the Cu electrodes modeled by four-layer slabs with a 3 × 3 surface. A vacuum space of 16 Å was placed above the slabs and adsorption is allowed on only one of the two surfaces exposed. The Kohn–Sham orbitals were expanded in a plane-wave basis set with a kinetic energy cutoff of 26 Ry and a charge density cutoff of 260 Ry. The Fermi-surface effects were treated by the smearing technique of Methfessel and Paxton using a smearing parameter³³ of 0.02 Ry. Calculations were carried out with spin-polarization, which is essential to properly represent the electronic structure of adsorbed CO₂. The PWSCF codes contained in the Quantum ESPRESSO distribution³⁴ were used to implement all calculations, while figures of the chemical structures were produced with the XCRYSDEN^35–37 graphics package. BZ integrations were performed using a (3 × 3 × 1) uniformly shifted k-mesh for a (3 × 3) supercell.

The calculated equilibrium lattice constant for Cu was 3.66 Å, which agreed well with theoretical and experimental values (3.66 and 3.62 Å, respectively).^38,39 During the calculations, the structure of the bottom two layers were fixed at the theoretical bulk positions, whereas the top two layers and the adsorbates were allowed to relax and all the other structural parameters were optimized so as to minimize the total energy of the system. Structural optimization was performed until the Cartesian force components acting on each atom were brought below 10⁻³ Ry Bohr⁻¹ and the total energy converged to within 10⁻⁵ Ry.

The climbing-image nudged elastic band (CI-NEB) method was used to determine the minimum energy paths (MEPs) for all the elementary steps.^40,41 The transition state of the optimized reaction coordinate was approximated by the image of highest energy. The transition state images from the CI-NEB calculations were optimized using the quasi-Newton method, which minimizes the forces to find the saddle point. Geometry optimization was performed for each intermediate point in MEPs, in which a three-layer Cu slab with a 2 × 3 surface unit cell was used. Considering the high cost of the CI-NEB calculation, the bottom two layers of metal atoms were fixed, while the top layer of metal atoms and all other non-metal atoms were allowed to relax.

The adsorption energy (E_ad) of adsorbate “A” was calculated according to E_ad(A) = E(slab-A) − E(slab) − E(A), where E(slab-A), E(slab) and E(A) refer to the total energy of a slab with an adsorbed “A”, the total energy of a slab, and the total energy of the free “A”, respectively. The co-adsorption energy between adsorbates “A” and “B” was calculated according to E_ad(A, B) = E(slab-A, B) − E(slab) − E(A) − E(B), in which E(slab-A, B) refers to the calculated total energy of the slab with co-adsorbed “A” and “B”. The difference value of “E_ad(A, B) − (E_ad(A) + E_ad(B))” should represent the interaction between “A” and “B” on the surface.

3. Results and discussion

3.1 Chemical reduction of CO₂

Table 1 lists the possible reaction steps in the chemical reduction of CO₂ by H₂ and the DFT-calculated reaction free energies for these steps on the Cu(111) and Cu(100) surfaces. The preferred adsorption configurations for all the reactants, intermediates and products listed in Table 1 on the two surfaces are given in Tables S1–S6 and Fig. S1–S3 in the ESI.† For the sake of simplicity, we directly use an H₂ molecule as the reactant in the reaction free energy calculations for all the hydrogenation steps. In reality, the hydrogenation agent in the elementary reaction steps of chemical reduction should be the adsorbed hydrogen atoms (H*) formed through H₂ dissociation. Therefore, the hydrogenation steps listed in Table 1 are not the elementary reaction steps. Although they are not the elementary steps, the calculated reaction free energies in Table 1 should still represent the general thermodynamic preference of various hydrogenation steps.

Table 1 Possible reduction steps in the chemical reduction of CO₂ and their reaction free energies on the Cu(111) and Cu(100) surfaces calculated with DFT^a

Possible reaction steps	Reaction free energies, ΔGreac^b (kJ mol^−1)
	Cu(111)	Cu(100)
a The asterisk (*) indicates that the species is adsorbed on the surface.b In the calculations, the entropies obtained from the literature of Nørskov and coworkers^26,27 are considered for gaseous molecules, including CO2, H2, CH3OH, CH4 and C2H4, whereas the entropies of the adsorbed species are ignored. The zero point energies (ZPE) for all species, which are taken from the literature of Nørskov and coworkers,^26,27 are included in the calculations. For steps involving the co-adsorption of two species, the co-adsorption structures were optimized and used in calculations. For instance, for a reaction step of (A + B)* → (C + D)*, the reaction free energy is calculated according to “E(slab-C, D) − E(slab-A, B) + ZPE(C*) + ZPE(D*) − ZPE(A*) + ZPE(B*)”.
1/2H2(g) → H*	−45.67	−41.34
(1a) CO2(g) + H* → HCOO*	−28.47	−50.66
(1b) CO2(g) + H* → COOH*	32.91	36.35
(1c) CO2(g) → CO* + O*	91.47	22.57
(1d) COOH* → CO* + OH*	−36.22	−49.87
(2a) CO* + 1/2H2(g) → COH*	96.92	63.59
(2b) CO* + 1/2H2(g) → CHO*	60.04	41.41
(2c) CO* → OCCO*	132.55	75.59
(2d) CO* → C* + O*	263.66	154.73
(2e) COH* → CHO*	−36.88	−22.18
(3a) CHO* + 1/2H2(g) → CH2O*	−40.49	−63.45
(3b) COH* + 1/2H2(g) → CHOH*	−25.13	2.30
(3c) COH* → C* + OH*	78.26	5.38
(3d) CHO* → CH* + O*	67.59	27.82
(4a) CH2O* + 1/2H2(g) → CH2OH*	−29.99	−1.64
(4b) CH2O* + 1/2H2(g) → CH3O*	−127.37	−99.55
(4c) CH2O* + 1/2H2(g) → CH2* + OH*	−42.45	−47.05
(4d) CH2O* → CH2* + O*	41.87	10.11
(4e) CHOH* + 1/2H2(g) → CH2OH*	−82.22	−89.57
(4f) CHOH* → CH2O*	−52.23	−87.93
(4g) CHOH* → CH* + OH*	−22.44	−110.11
(5a) CH2OH* + 1/2H2(g) → CH3OH(l)	−112.67	−96.53
(5b) CH2OH* → CH2* + OH*	−12.46	−45.41
(5c) CH3O* + 1/2H2(g) → CH3OH(l)	−15.29	1.38
(5d) CH2OH* → CH3O*	−97.38	−97.91
(5e) CH2OH* + 1/2H2(g) → CH3* + OH*	−118.31	−126.97
(5f) CH3O* + 1/2H2(g) → CH3* + OH*	−20.93	−29.07
(6a) CH2* + 1/2H2(g) → CH3*	−105.85	−61.37
(6b) CH3* + 1/2H2(g) → CH4(g)	−99.41	−78.61
(6c) CH2* → 1/2C2H4*	−42.70	−92.59

---

According to calculated adsorption free energies (Table S1†), CO₂ cannot form stable chemical adsorption states on the Cu(111) and Cu(100) surfaces. Therefore, gaseous CO₂ is used in the free energy calculations. The physisorption state of CH₄ and chemisorption state of C₂H₄ are obtained on the Cu(111) surface. The corresponding adsorbed states of CH₄ and C₂H₄ are also obtained on the Cu(100) surface. Moreover, the physisorption state of CH₃OH is also obtained on the Cu(111) surface. Although the chemical reduction in industry should be operated practically at elevated temperatures, the reaction free energies in Table 1 were calculated for room temperature (298 K) so that the potential effect can be easily determined.

Although the chemical hydrogenation reactions of CO₂ to CH₃OH, CH₄ and C₂H₄, namely, CO₂(g) + 3H₂(g) → CH₃OH + H₂O, CO₂(g) + 4H₂(g) → CH₄ + 2H₂O and CO₂(g) + 3H₂(g) → 1/2C₂H₄ + 2H₂O, have rather negative overall reaction free energies of −109.45 kJ mol⁻¹, −232.43 kJ mol⁻¹ and −180.46 kJ mol⁻¹, respectively, some of the elementary reaction steps, especially the reaction steps of CO, have rather positive reaction free energies. This might be the reason why these reactions are practically difficult.

The DFT-calculated activation energies for H₂ dissociation are 0.52 eV and 0.45 eV on the Cu(111) and Cu(100) surfaces, respectively, which are relatively small compared to activation barriers for other steps in CO₂ reduction as shown later. Therefore, H₂ dissociation should also play a relatively minor role in the kinetics of the chemical reduction of CO₂ by H₂.

In the following, we will first deduce the most possible pathways for the chemical reduction of CO₂ on the two types of Cu surfaces according to the calculated reaction free energies shown in Table 1 and the activation barriers determined from the CI-NEB method. The mechanisms of electrochemical reduction will then be discussed by analyzing the effect of electrode potentials on the steps in these pathways on the basis of the theoretical hydrogen electrode model and by considering water in the double layer as the hydrogenation agent in the electrochemical environment.

3.1.1 The first steps of CO₂ chemical reduction – formation of CO. According to the calculated reaction free energies in Table 1, the direct dissociation of CO₂ into adsorbed CO* and O* is the most endergonic process among various possible initial steps for CO₂ chemical reduction on the Cu(111) surface, while the hydrogenation of CO₂ to an adsorbed formate (HCOO*) is the most exergonic process. In comparison, the hydrogenation to carboxyl (COOH) and the hydrogenative dissociation to adsorbed CO and OH are both mildly exergonic.

Fig. 1 displays the calculated MEPs for these possible initial steps on the Cu(111) surface. In these calculations, the initial state contains a physisorbed CO₂ molecule and a chemisorbed H* at the fcc site. The calculated activation energies for these steps follow the same trend as the reaction free energies, i.e., the direct dissociation of CO₂ exhibits the highest activation barrier (∼1.54 eV), while the formation of formate by the direct hydrogenation of CO₂ requires the lowest activation energy (∼0.86 eV). The activation barrier for the formation of carboxyl (COOH) is ca. 1.18 eV. The dissociative hydrogenation of CO₂ to adsorbed CO and OH is not an elementary step. It involves COOH as an intermediate, whose dissociation requires activation energy of only ca. 0.2 eV. Thus, both the reaction free energies and activation energies from DFT calculations suggest that formate is the most preferred initial product in the chemical reduction of CO₂ on the Cu(111) surface. The formation of CO from CO₂ is most probably through the COOH intermediate, which is formed with an activation barrier slightly higher than the formate formation. Experiments starting with formate showed no detectable products, and CO product was suppressed at a less negative potential when starting with CO₂.^20,23

Fig. 1 Minimum energy paths for the three possible initial steps in the chemical reduction of CO₂ on the Cu(111) surface. Oxygen atoms are red, hydrogen atoms are white, carbon atoms are gray, and copper atoms are blue.

Similar to the Cu(111) surface, the formation of formate is a relatively strong exergonic process (∼92.00 kJ mol⁻¹). Compared to that on the Cu(111) surface, however, the formation of COOH becomes much less exergonic, while the direct dissociation of the CO₂ becomes much less endergonic on the Cu(100) surface. If considering that hydrogenation may occur through adsorbed H atoms, the reaction free energy for the formation of COOH could be more endergonic than the dissociation of CO₂. As will be shown in later MEP calculations, the direct dissociation of CO₂ on the Cu(100) surface requires a lower activation energy than its hydrogenation to COOH.

As shown in Fig. 2, CO formation through the direct dissociation of CO₂ requires considerably lower activation energy (1.29 eV) than that through the hydrogenative dissociation (1.79 eV) on the Cu(100) surface and is different from that on the Cu(111) surface. Therefore, the direct dissociation of CO₂ would be the preferred pathway for CO formation on the Cu(100) surface.

Fig. 2 Minimum energy path of CO formation on the Cu(100) surface by the direct dissociation and hydrogenation of CO₂. Oxygen atoms are red, hydrogen atoms are white, carbon atoms are gray, and copper atoms are blue.

3.1.2 Reaction pathways of CO. There are three possible reactions for the CO formed, namely, direct dissociation to adsorbed C* and O*, hydrogenation to form an adsorbed formyl (CHO*), and hydrogenation to form an adsorbed hydroxymethylidyne (COH*). Both the calculated reaction free energies (Table 1) and the MEPs (Fig. 3a) suggest that these reaction paths are endergonic, and the direct dissociation is particularly energetically demanding (2.75 eV and 3.96 eV for reaction and activation free energies, respectively). Although the two hydrogenation reactions do not significantly differ in reaction free energies (ca. 1.0 eV and 0.67 eV), the formation of CHO* requires an activation barrier nearly 1.5 eV lower than the formation of COH* (∼1.0 eV vs. ∼2.6 eV). On this basis, we may conclude that the most potential reaction path for CO is its hydrogenation to CHO*. By carefully inspecting the intermediate images in the calculated MEPs for the two hydrogenation reaction paths, one finds that the COH* formation actually involves CHO* as the intermediate.

Fig. 3 Minimum energy path of CO dissociation and hydrogenation to form C, COH and CHO (a) on the Cu(111) surface and (b) on the Cu(100) surface. Oxygen atoms are in red, hydrogen atoms are in white, carbon atoms are in gray, and copper atoms are in blue.

The calculated MEPs for the steps on the Cu(100) surface following CO formation suggest they are very similar to that on the Cu(111) surface. The activation energy barriers for the direct dissociation of CO and hydrogenation to form COH* and CHO* are 3.01, 2.43, and 2.36 eV, respectively, as shown in Fig. 3b. These results indicate that CO hydrogenation requires higher activation energy barriers on the Cu(100) surface, which is more easy to achieve on the Cu(111) surface. By carefully inspecting the intermediate images in the calculated MEPs for the two hydrogenation reaction paths on the Cu(100) surface, one finds that CHO* formation involves COH* and the dissociative adsorption state C* + OH* as the intermediates.

In fact, the formation of a formyl intermediate has been recognized in other reactions involving CO hydrogenation, for examples, alcohol production from carbon monoxide and molecular hydrogen on the Rh(111) surface,⁴² Rh/TiO₂ (ref. 43) and Rh/SiO₂ catalysts,^44,45 and formaldehyde and methanol synthesis from CO and H₂ on the Ni(111) surface.⁴⁶

3.1.3 Reaction steps following CO hydrogenation. Among the three possible reaction paths of CHO*, hydrogenation to the adsorbed formaldehyde (CH₂O*) is thermodynamically the most exergonic process (−0.42 eV reaction free energy) and requires the lowest activation energy (ca. 0.7 eV). The other processes, including the hydrogenation to CHOH* and the dissociation to CH* + O* and/or C* + OH*, require activation energies near to or above 2.0 eV, as shown in Fig. 4, although the hydrogenation reaction to CHOH* is only mildly endergonic (0.12 eV reaction free energy). Therefore, CH₂O* should be a key intermediate after CHO* formation for CO₂ chemical reduction on the Cu(111) surface.

Fig. 4 Minimum energy paths on the Cu(111) surface of CHO* dissociation and hydrogenation into CH, C, CHOH and CH₂O. Oxygen atoms are red, hydrogen atoms are white, carbon atoms are gray, and copper atoms are blue.

In a previous study on the mechanism of ethanol synthesis from syngas on the Rh(111) surface, Liu et al.⁴² showed that the CHO* intermediate formed through CO hydrogenation can be more easily hydrogenated to form CH₂O* than the dissociation. Earlier studies^13,23 have also shown that CH₄ can be synthesized on Cu with formaldehyde (CH₂O) as the starting material.

CH₂O may undergo direct dissociation to methylene (:CH₂) and O*, direct hydrogenation to form CH₃O* and/or CH₂OH, and/or hydrogenated dissociation to methylene (:CH₂) and OH. Calculated reaction free energies for the formation of CH₃O* and/or CH₂OH* are −127.37 kJ mol⁻¹ and −29.99 kJ mol⁻¹, respectively, while it is −49.08 kJ mol⁻¹ for :CH₂ formation. This seems to suggest that the CH₃O* pathway is preferable on the Cu(111) surface. As shown by the MEP calculation results in Fig. 5, the formation of CH₃O* and the hydrogenated dissociation of CH₂O* to :CH₂ actually proceed via a CH₂OH* intermediate, which has a lower activation energy than the CH₂O* direct dissociation. Therefore, although the reaction free energy of CH₂OH* formation is less exergonic than CH₃O*, the preferred reaction pathway of CH₂O* should be hydrogenation to form CH₂OH* rather than CH₃O*.

Fig. 5 Minimum energy paths of the direct hydrogenation, the hydrogenative dissociation, and the direct dissociation of CH₂O on the Cu(111) surface. Oxygen atoms are in red, hydrogen atoms are in white, carbon atoms are in gray, and copper atoms are in blue.

It can be seen (Fig. 5) that the formation of CH₂ and OH via CH₂O* hydrogenative dissociation is not an elemental step. The hydrogenation of CH₂O* first produces a CH₂OH* intermediate with an activation barrier of about 0.48 eV; the transformation of CH₂OH* to CH₂ and OH has an activation barrier of about 1.04 eV (Fig. 5). As shown by the calculated MEPs in Fig. 6, the activation barriers for the direct hydrogenation of CH₂OH* to form CH₃OH and the hydrogenative dissociation of CH₂OH* to form CH₂ are ∼0.68 eV and 0.66 eV, respectively. Thus, the CH₂OH* intermediate either forms CH₃OH by direct hydrogenation or forms CH₂ by hydrogenative dissociation. Based on the activation barriers, the simultaneous occurrence of both the paths can be concluded, which may be a parallel path in CO₂ reduction mechanism on the Cu(111) surface. In addition, the hydrogenative reduction of CH₂OH* to CH₃OH and the hydrogenative dissociation to form CH₂ and H₂O are also easier than the abovementioned CH₂OH direct dissociation. The presented results explain the experimental results in which CH₃OH was observed in CO₂ chemical reduction on copper surfaces.

Fig. 6 Minimum energy paths of the hydrogenative dissociation of CH₂OH to CH₂ and H₂O and the direct hydrogenation to CH₃OH, respectively, on the Cu(111) surface. Oxygen atoms are in red, hydrogen atoms are in white, carbon atoms are in gray, and copper atoms are in blue.

The formation of CH₄ through CH₂ hydrogenation is drastically exergonic on the Cu(111) surface. In comparison, C₂H₄ formation is less exergonic. This explains the experimental result that CH₄ is more likely formed on the Cu(111) surface. Fig. 7a shows the MEPs for the formation of CH₄ and C₂H₄ from CH₂. The activation barriers for the two hydrogenation steps in CH₄ formation are 0.63 eV and 1.03 eV. For the association of two CH₂ to form C₂H₄, the activation barrier is 0.21 eV, which is less than that of CH₂ hydrogenation to CH₃. This seemingly indicates that CH₂ dimerization is kinetically more favorable. However, as shown by the calculated MEPs in Fig. 6, CH₃ and OH intermediates are observed during the course of CH₂OH hydrogenative dissociation to CH₂ and H₂O, and the reaction free energy of forming CH₃ and OH intermediates is more negative than that of forming CH₂ and H₂O; an activation barrier of ∼0.53 eV is required from CH₃ and OH to CH₂ and H₂O, namely, the CH₃ intermediate may be easier to form on the Cu(111) surface. Considering that the surface coverage of the CH₂ should be very low, the dimerization reaction would thus still have a much lower reaction rate than the hydrogenation reaction on the Cu(111) surface.

Fig. 7 Minimum energy paths of CH₂ successive hydrogenation into hydrocarbons CH₃ and CH₄ and dimerization into C₂H₄ (a) on the Cu(111) surface and (b) on the Cu(100) surface.

On the Cu(100) surface, however, we find that the CH₂ dimerization into C₂H₄ is a non-activated process, as shown in Fig. 7b. The paths for CH₃ and CH₄ formation have activation energy barriers of 0.34 and 0.55 eV, respectively. The dimerization of CH₂ to C₂H₄ has a more negative reaction free energy than that of CH₂ hydrogenation to CH₃. Thus, C₂H₄ formation should be more easy on the Cu(100) surface than CH₄ formation.

The results from Koper and coworkers²⁹ suggested that there are two separate pathways for the formation of C₂H₄, one that shares an intermediate with the pathway to CH₄, as they observed on the Cu(111) surface and below −0.8 V on the Cu(100) surface at pH 7, and a second pathway that occurs only on the Cu(100) surface. For this second pathway, they suggest that the formation of a CO dimer is the key intermediate in the formation of C₂H₄. As shown in Fig. 8, the activation energy barrier for the formation of CO dimer on the Cu(111) and Cu(100) surfaces is 1.59 and 1.26 eV, respectively. The results indicate that CO dimerization has a higher activation energy barrier than CO hydrogenation to form CHO* on the Cu(111) surface, whereas on the Cu(100) surface, CO dimerization requires a lower activation energy barrier than CO hydrogenation to form CHO*. Therefore, CO hydrogenation is more easy on the Cu(111) surface, whereas CO dimerization into (CO)₂ is more easy on the Cu(100) surface. Such a surface dimer could explain the unique selectivity for C₂H₄ (for detailed arguments, see ref. 7) and is in agreement with the suggestion of Gattrell et al.¹¹ who proposed that this CO dimer would be more stable on the Cu(100) surface. In addition, our results may somewhat agree with Koper and coworkers.

Fig. 8 Minimum energy paths of CO dimerization into OCCO* on the Cu(111) and Cu(100) surfaces.

Based on the abovementioned reaction free energy and MEP calculations, the optimal reaction paths for CO₂ reduction to hydrocarbons on the Cu(111) surface can be summarized in Table 2. CO is first formed by the dissociative hydrogenation of CO₂, which is the most feasible reaction path for CO is its hydrogenation to a CHO* intermediate. The key intermediate CH₂O can be more easily formed through further CHO hydrogenation, and the preferred reaction pathway of CH₂O should be the hydrogenation to form CH₂OH, which is based on the activation energy barriers. The CH₂OH intermediate either forms CH₃OH by direct hydrogenation or forms CH₂ by hydrogenative dissociation, which may be a parallel path in the CO₂ reduction mechanism on the Cu(111) surface; finally, the CH₂ intermediate leads to formation of the hydrocarbons. From a kinetic viewpoint, the relatively slow steps on the Cu(111) surface include CO₂(g) + H* → (CO + OH)* and (CO + H)* → CHO*. The reason may be that the weak Cu–CO₂ interaction is an obstacle to CO₂ dissociative hydrogenation and therefore slows down the overall conversion. The differences in the reaction mechanism on the Cu(111) and Cu(100) surfaces are that the CO formation and further reduction undergoes a different pathway. Simultaneously, the preference for the formation of CH₄ and C₂H₄ is also different; the formation of C₂H₄ is easier on the Cu(100) surface, and the experimental results are explained to some degree on the selectivity of the Cu single crystal surface. The C₂H₄ formation mechanisms via CO dimerization will be studied systematically in our future work.

Table 2 The minimum energy paths for the chemical reduction of CO₂ to hydrocarbons and the activation energy barriers (E_act) in each elementary step on the Cu(111) surface

Reaction paths	Eact (eV)
a The values in parentheses are the activation energy barriers of the corresponding processes on the Cu(100) surface.
CO2(g) + * → (CO + O)*	1.54 (1.29)^a
CO2(g) + H* → (CO + OH)*	1.18 (1.79)
(CO + H)* → CHO*	1.06
(CHO + H)* → CH2O*	0.72
(CH2O + H)* → (CH2 + OH)*	1.12
(CH2O + H)* → CH2OH*	0.95
(CH2O + H)* → CH3O*	1.86
(CH2OH + H)* → CH3OH(l)	0.68
(CH2OH + H)* → CH2* + H2O(l)	0.66
(CH2 + H)* → CH3*	0.63 (0.34)
(CH3 + H)* → CH4(g)	1.03 (0.56)
2CH2* → C2H4*	0.21 (∼0)

---

3.2 Electrochemical reduction of CO₂

One of the major differences between the electrochemical reduction and the chemical reduction is that the H₂ molecules will be replaced by the (H⁺ + e) pair. The overall reaction for the electrochemical reduction of CO₂ to methanol and hydrocarbons can be expressed as follows,

CO₂(g) + 8(H⁺ + e⁻) → CH₄ + 2H₂O (1)

CO₂(g) + 6(H⁺ + e⁻) → 1/2C₂H₄ + 2H₂O (2)

CO₂(g) + 6(H⁺ + e⁻) → CH₃OH + H₂O (3)

Since the chemical potential (μ) of the electron changes with the electrode potential (E), the reaction free energies will vary with E. If the potential of the reversible hydrogen electrode (RHE) is used as the zero potential, the relationship would be μ(H⁺ + e⁻) = 1/2μ(H₂) − eE.⁴⁷ This relationship provides an elegant way to calculate the potential-dependent free energies of the reaction steps in the electrochemical reduction by avoiding the explicit treatment of solvated protons.

In electrochemical reduction, the reaction steps involving hydrogenation will be enhanced thermodynamically as the potential becomes negative. In addition, for reaction steps which produce O* or OH*, the reaction may be promoted since these oxygenated species can be easily removed from the surface at negative potentials. Experimental results^9,16,20 have shown that when the electrode potential is −0.50 V (vs. RHE), formate is the main product; when the electrode potential is −0.67 V (vs. RHE), CO is the main product; when the electrode potential is more negative, i.e. −0.90 V (vs. RHE), hydrocarbons begin to form. We chose these three electrode potentials to calculate the reaction free energies for various steps and the results are given in Table 3.

Table 3 Possible reduction steps in the electrochemical reduction of CO₂ to hydrocarbon and their reaction free energies on the Cu surfaces calculated with DFT

Possible reduction steps	Reaction free energies, ΔGreac (kJ mol^−1)
	Cu(111)			Cu(100)
	−0.50 V	−0.67	−0.90	−0.50 V	−0.67	−0.90
(a) CO2(g) + H^+ + e^− → HCOO*	−98.06	−114.47	−136.66	−119.84	−136.25	−158.44
(b) CO2(g) + H^+ + e^− → COOH*	−16.29	−32.70	−54.89	−32.83	−49.24	−71.43
(c) COOH* + H^+ + e^− → CO* + H2O(l)	−84.07	−101.48	−123.67	−73.12	−89.53	−111.72
(d) HCOO* + H^+ + e^− → CO* + H2O(l)	−3.30	−19.71	−41.90	13.89	−2.52	−24.71
(e) CO* + H^+ + e^− → CHO*	11.79	−4.62	−26.81	−6.84	−23.25	−45.44
(f) CO* + H^+ + e^− → COH*	48.67	32.26	10.07	15.34	−1.07	−23.26
(g) CHO* + H^+ + e^− → C* + H2O(l)	84.63	68.22	46.03	−25.48	−41.89	−64.08
(h) CHO* + H^+ + e^− → CH* + OH*	−58.95	−75.36	−97.55	−133.88	−150.29	−172.48
(i) CH* + OH* + H^+ + e^− → CH* + H2O(l)	−48.84	−65.25	−87.44	−23.25	−39.66	−61.85
(j) CHO* + H^+ + e^− → CH2O*	−88.74	−105.15	−127.34	−111.70	−128.11	−150.30
(k) CHO* + H^+ + e^− → CHOH*	−36.50	−52.91	−75.10	−23.77	−40.18	−62.37
(l) CH2O* + H^+ + e^− → CH3O*	−175.62	−192.03	−214.22	−147.80	−164.21	−186.40
(m) CH2O* + H^+ + e^− → CH2OH*	−78.24	−94.65	−116.84	−49.89	−66.30	−88.49
(n) CH2OH* + H^+ + e^− → CH3OH(l)	−160.92	−177.33	−199.52	−144.78	−161.19	−183.38
(o) CH3O* + H^+ + e^− → CH3OH(l)	−63.54	−79.95	−102.14	−46.87	−63.28	−85.47
(p) CH2O* + H^+ + e^− → CH2* + OH*	−90.71	−107.12	−129.31	−95.30	−111.71	−133.90
(q) CH2* + OH* + H^+ + e^− → CH2* + H2O(l)	−48.84	−65.25	−87.44	−23.25	−39.66	−61.85
(r) CH2* + H^+ + e^− → CH3*	−154.10	−170.51	−192.70	−129.82	−146.23	−168.42
(s) CH3* + H^+ + e^− → CH4(g)	−147.66	−164.07	−186.26	−146.09	−162.50	−184.69
(t) CH2* → 1/2C2H4*	−85.40	−85.40	−85.40	−185.18	−185.18	−185.18

---

The free energies, ΔG_reac, on the Cu(111) surface for the overall reactions (1) are −618.43 kJ mol⁻¹, −749.67 kJ mol⁻¹, and −927.23 kJ mol⁻¹; the ΔG_reac for the overall reaction (2) are −469.96 kJ mol⁻¹, −568.39 kJ mol⁻¹, and −701.56 kJ mol⁻¹; and the ΔG_reac for the overall reaction (3) are −389.95 kJ mol⁻¹, −497.38 kJ mol⁻¹, and −630.55 kJ mol⁻¹ when the electrode potential is −0.50 V, −0.67 V, and −0.90 V (vs. RHE), respectively, which are all highly exergonic. The ΔG_reac in the electrochemical environment is more negative than that in chemical reduction environment; therefore, these reactions may be more favorable in an electrochemical environment.

According to the calculated electrochemical reaction free energies, the formation of formate and COOH will both be promoted at negative electrode potentials. In the electrochemical condition, the hydrogenated dissociation of COOH to CO would become a favorable step as the potential goes more negative than −0.67 V. The subsequent protonation of CO also gradually becomes highly exergonic. Therefore, the hydrocarbons would become the favored products in the electrochemical reduction, especially at more negative potentials. These results are in accordance with the experimental results.^9,16,20

Under the electrochemical condition, CHO* could undergo a hydrogenated dissociation to CH as the OH and O can be easily removed from the surface. As the potential is more negative than −0.67 V, the formation of CH could become more favored than the formation of CH₂O*. However, as shown in Fig. 4, the activation energy for CHO hydrogenation to form CH₂O* is ca. 0.7 eV and the dissociation to CH* + O* requires activation energies above 2.0 eV. CH₂O* could undergo a hydrogenated dissociation to CH₂* + H₂O or be hydrogenated directly to CH₃O* or CH₂OH*. As the potential is more negative than −0.90 V, the formation of CH₂* + H₂O and CH₃O* are strong exergonic processes, and the free energies, ΔG_reac, are approximately equal, indicating that CH₂* + H₂O and CH₃O* may be formed simultaneously by further the reduction of CH₂O*. However, as shown in Fig. 6, CH₃O* formation requires a higher activation energy. Thus, the formation of CH₂* + H₂O could become more favored. It is worth considering why CH₃OH is formed in gas phase chemistry⁴⁸ and CH₄ is formed in electrochemistry^16,17 on copper surfaces. In the current study, we considered that adsorbed CH₃O* or CH₂OH* may form CH₃OH via a proton transfer reaction. However, as our calculations showed that the free energies, ΔG_reac, of CH₂OH* formation are more positive than the CH₂O* hydrogenated dissociation to CH₂* + H₂O, and CH₃O* formation requires a higher activation energy, this would be in favor of CH₄ formation, in agreement with the electrochemical experimental results in which only CH₄, and not CH₃OH, was observed. Accordingly, we speculate that the optimum electrochemical reduction paths of CO₂ on the Cu(111) surfaces are CO₂(g) + 2(H⁺ + e⁻) → CO* + H₂O(l), CO* + (H⁺ + e⁻) → CHO*, CHO* + (H⁺ + e⁻) → CH₂O*, CH₂O* + 2(H⁺ + e⁻) → CH₂* + H₂O(l), CH₂* + (H⁺ + e⁻) → CH₃*, and CH₃* + (H⁺ + e⁻) → CH₄(g), in which CO* + (H⁺ + e⁻) → CHO* may be the potential-limiting step. A similar conclusion has been obtained by Nørskov and coworkers.²⁶

In the calculations conducted by Nørskov et al.,^26,27 direct hydrogenation was considered as the major reaction of CH₂O*. The key intermediate for the formation of hydrocarbons in their reaction mechanism is therefore CH₃O*. By comparing the calculated reaction free energies and activation energies for the two pathways, we find that the hydrogenated dissociation pathway is actually more favorable. Koper et al.^28,29 reported their studies of the reduction of CO₂ on two basal planes, Cu(111) and Cu(100), using online electrochemical mass spectrometry (OLEMS) to investigate the path to CH₄ and C₂H₄. This tip-based sampling technique allows the formation of volatile reaction intermediates and products to be followed while the potential at the electrode is changed. One can measure the reduction of the various species online while changing the potential and therefore follow the formation and consumption of intermediates during the reaction. Their experimental results show that it is very likely that CHO_ads is the key intermediate toward the breaking of the C–O bond and therefore the formation of CH₂ and CH₃ species and CH₄. CH₃O cannot be reduced to CH₄ on the copper electrodes. The experimental observations appear to be in conflict with the DFT results from Nørskov et al., which suggest that CHO_ads is the intermediate in the formation of methane, but the energetically favored route to CH₄ is through adsorbed CH₃O species.

In addition, it has been proposed in earlier studies⁴⁹ that the electrocatalytic reduction of CO₂ on Cu electrodes involves an initial stage forming the carbon dioxide anion radical, CO₂˙⁻, which is not observed on the Cu(111) surface in our previous calculations since the chemisorption state of the CO₂ molecule is not obtained on the Cu(111) surface during the geometry optimization and minimum energy path analysis. Although the chemisorption state of CO₂ molecule is obtained on the Cu(100) surface, the adsorption energies of the chemisorption states of the CO₂ molecule are positive, which indicates that the chemisorption state is a metastable state. Indeed, in the electrochemical reduction of CO₂, formate and COOH species are usually formed through anion radical ˙CO₂⁻ protonation. Due to a large reorganizational energy between the linear molecule and bent radical anion, the outer-sphere single electron reduction of CO₂ to carbon dioxide anion radical should be difficult from both kinetics and thermodynamics.¹ However, when the electrode potential is more negative, the electron level of the electrode could be aligned with the LOMO of CO₂ and electron transfer to the CO₂ molecule is possible. If a proton transfer occurs simultaneously, the products could be more stable. This phenomenon can be verified by the geometry optimization of the CO₂ molecule in the presence of a hydronium ion. Accordingly, in this section, we used the hydronium ion H₃O⁺⋯(H₂O)₃ as a model of solvation in an acid solution in order to model the first proton-coupled electron transfer step during CO₂ reduction at the Cu/electrolyte interface and evaluate the role of the proton, for which the rationalization of the model has been validated in our previous work.⁵⁰ Fig. 9 gives the geometry structural optimization plot of the protonation of CO₂ in the presence of hydronium ion H₃O⁺⋯(H₂O)₃. Optimized results indicate that the anion radical ˙CO₂⁻ is formed initially and then the hydrated proton, H₉O₄⁺, preferentially combined with O atom in the chemisorbed CO₂ to form a COOH species (Fig. 9a and b), rather than a formate species, when the CO₂ molecule adsorbed on the Cu(111) and Cu(100) surfaces for the protonation process, even if the hydrated proton H₉O₄⁺ was closer to the C atom in the initial structure. In other words, formate species formation may require a higher activation energy than COOH under the electrochemical environment. Based on the geometry optimization, it can be concluded that the anion radical ˙CO₂⁻ is formed primarily under electrochemical conditions and then successive proton and electron transfer occurs, eventually leading to the formation of hydrocarbons. Moreover, the experimental result⁵¹ also indicated that the adsorption of alkali metal on the Cu surfaces promoted the formation of the carbon dioxide anion radical ˙CO₂⁻. Thus, the reaction in experiment conditions may be also promoted by alkali metal halides.

Fig. 9 Geometry optimization plot of the protonation of CO₂ in the presence of hydrated proton H₉O₄⁺: (a) COOH adsorbed on the Cu(111) surface and (b) COOH adsorbed on the Cu(100) surface.

Simultaneously, the abovementioned DFT calculations have also shown that there are two possibilities for hydrocarbon formation: one is chemical reduction through direct hydrogenation via adsorbed hydrogen, which occurs by a surface chemistry reaction, for example:

(CH₂ + H)* → CH₃*, (CH₃ + H)* → CH₄(g)

The other possibility to produce hydrocarbons is by direct electrochemical reaction:

CH₂* + H⁺ + e⁻ → CH₃*, CH₃* + H⁺ + e⁻ → CH₄(g)

The two routes are compared in Fig. 10 for the Cu(111) surface at −0.90 V (vs. RHE). As can be seen, in the direct electrochemical route, the intermediate CH₂ formation is exergonic. CH₂ formation can be seen to be ‘uphill’ in energy, and the free energies of CH₂ hydrogenation to form CH₄ are more positive in the hydrogenation route than in the direct electrochemical route. Thus, the direct electrochemical route would be expected to be highly favored at potentials of interest to CO₂ reduction. Simultaneously, under electrochemical conditions, the free energy of CH₄ formation via two protonation steps of CH₂ is more negative on the Cu(111) surface than on the Cu(100) surface (as shown in Fig. 10), and the free energy of C₂H₄ formation is more negative than that of CH₂ protonation to form CH₃ on the Cu(100) surface (as shown in Table 3). Thus, DFT study indicated that CH₄ is favorably formed on the Cu(111) surface and C₂H₄ is preferably formed on the Cu(100) surface under electrochemical conditions, explaining the experimental results.

Fig. 10 Comparison of the production of CH₄(g) from CH₂* via the direct electrochemical route and the combination of adsorbed CH₂* and adsorbed H* on the Cu(111) surface, and effect of crystal face on selectivity of product.

The possible optimized reaction paths of CO₂ reduction to form hydrocarbons on Cu single crystal electrode surfaces and the degree of difficulty of the reaction have been revealed, and a systematic understanding of the CO₂ reduction mechanism has also been achieved using density functional theory (DFT) method in the first principle on the geometry structural optimization, the reaction energy calculations and the minimum energy paths analysis. On this basis, the effect of combining the activation barriers of rate-determining steps and reaction pathways in C₂H₄ production via CO dimerization on the Cu(100) surface will be studied in the future. The research results have great significance for understanding in depth the selectivity of CO₂ electrochemical reduction processes, the sensitivity of surface structure and the size effect of catalysts. Simultaneously, the research results can provide a scientific basis for designing the catalysts for CO₂ electrochemical reduction.

4. Conclusions

In this study, a complete thermodynamic and kinetic description of CO₂ reduction on two Cu single crystal surfaces is carried out for the first time based on a systematic DFT exploration. New insight has been provided on the chemical and electrochemical reduction of CO₂ into CH₄ at the atomic level on the Cu surfaces.

In agreement with the present experiment and theoretical studies, our DFT calculated results indicated that CHO_ads and CH₂OH_ads are the key intermediates toward the chemical and electrochemical reduction of CO₂ into hydrocarbons. CO dimerization is more easy to achieve; CO hydrogenation is difficult to occur on the Cu(100) surface, explaining the unique selectivity for C₂H₄. However, our DFT calculated results indicated CH₃O cannot be reduced to CH₄ on copper electrodes; the main reaction of the CHO_ads is the hydrogenated dissociation to :CH₂, eventually leading to the formation of CH₃ species and CH₄, which are consistent with the present experimental results but conflict the present theoretical results. In addition, DFT calculated results showed for the first time that the electrochemical reduction would be expected to be highly favored at potentials of interest to CO₂ reduction compared with the chemical reduction, and the carbon dioxide anion radical (˙CO₂⁻) is involved in the initial stage of CO₂ electroreduction.

By analyzing the chemical reduction pathways, we deduced important mechanistic information. The possible optimal pathway on the Cu(111) surface is CO₂(g) + H* → COOH* → CO* + OH*, CO* + H* → CHO*, CHO* + H* → CH₂O*, CH₂O* + H* → CH₂OH* → CH₂* + OH*, CH₂* + H* → CH₃*, CH₃* + H* → CH₄(g). CO is first formed by the dissociative hydrogenation of CO₂. The most possible reaction path for CO is its hydrogenation to a CHO* intermediate. The key intermediate CH₂O can be more easily formed through further CHO hydrogenation; the preferred reaction pathway of the CH₂O should be hydrogenation to form CH₂OH. The CH₂OH intermediate either forms CH₃OH by direct hydrogenation or CH₂ by hydrogenative dissociation based on the activation barriers, which may be a parallel path in the CO₂ reduction mechanism on the Cu(111) surface; finally, the CH₂ intermediate leads to the formation of the hydrocarbons. From a kinetic viewpoint, the relatively slow steps on the Cu(111) surface include CO₂(g) + H* → (CO + OH)* and (CO + H)* → CHO*. The reason may be that the weak Cu–CO₂ interaction is an obstacle to CO₂ dissociative hydrogenation and therefore slows down the overall conversion. The differences in the reaction mechanism on the Cu(111) and Cu(100) surfaces are that the CO formation and further reduction undergoes a different pathway. Simultaneously, the preference in the formation of CH₄ and C₂H₄ is also different, the formation of C₂H₄ is easier on the Cu(100) surface, and the experimental results are explained to some degree based on the selectivity of the Cu single crystal surface.

The free energies for various steps in the electrochemical reduction of CO₂ were calculated under different electrode potentials. The results indicated that formate and CO are mainly formed when the potential is more positive than −0.50 V (vs. RHE). The protonated dissociation of CO₂ to form CO and the subsequent protonation of CO become increasingly exergonic as the potential goes negative, so that hydrocarbon gradually becomes the favored product during the electrochemical reduction. Under electrochemical conditions, the possible optimum reaction path on the Cu(111) surface is CO₂(g) + 2(H⁺ + e⁻) → CO* + H₂O(l), CO* + (H⁺ + e⁻) → CHO*, CHO* + (H⁺ + e⁻) → CH₂O*, CH₂O* + 2(H⁺ + e⁻) → CH₂* + H₂O(l), CH₂* + (H⁺ + e⁻) → CH₃*, CH₃* + (H⁺ + e⁻) → CH₄(g), in which CO* + (H⁺ + e⁻) → CHO* may be the potential-limiting step. Simultaneously, the calculated results also partly explained why CH₃OH is formed in gas phase chemistry and only CH₄ is observed in electrochemistry on copper surfaces.

Acknowledgements

This study is financially supported by the National Natural Science Foundation of China (Grant No. 21303048), the Hunan Provincial Natural Science Foundation of China (Grant No. 13JJ4101), the Construct Program of the Key Discipline in Hunan Province (Applied Chemistry) and the Doctoral Start-up Fund of Hunan University of Arts and Science.

References

1. F. Hasegawa, S. H. Yokoyama and K. Imou, Bioresour. Technol., 2010, 101, S109–S111.
2. M. Abu-zahra, L. H. Schneiders, J. P. M. Niederer, P. H. Feron and G. F. Versteeg, Int. J. Greenhouse Gas Control, 2007, 1, 37–46.
3. B. P. Sullivan, Electrochemical and Electrocatalytic Reduction of Carbon Dioxide, Elsevier, Amsterdam, 1993.
4. C. M. Sanchez-Sanchez, V. Montiel, D. A. Tryk, A. Aldaz and A. Fujishima, Pure Appl. Chem., 2001, 73, 1917–1927.
5. M. M. Halmann, Greenhouse Gas Carbon Dioxide Mitigation, Lewis Publishers, Washington, DC, 1999.
6. T. Inui, M. Anp, K. Izui, S. Yanagida and T. Yamaguchi, Advances in Chemical Conversions for Mitigating Carbon Dioxide, Elsevier, Amsterdam, 1998.
7. Y. Hori, K. Kikuchi, A. Murata and S. Suzuki, Chem. Lett., 1986, 15, 897–898.
8. Y. Hori, A. Murata, R. Takahashi and S. Suzuki, J. Am. Chem. Soc., 1987, 109, 5022–5023.
9. J. J. Kim, D. P. Summers and K. W. Frese Jr, J. Electroanal. Chem., 1988, 245, 223–244.
10. Y. Hori, Electrochemical CO₂ reduction on metal electrodes, in Modern Aspects of Electrochemistry, Springer, New York, 2008.
11. M. Gattrell, N. Gupta and A. Co, J. Electroanal. Chem., 2006, 594, 1–19.
12. W. M. Ayers, Carbon Dioxide Chemistry: Environmental Issues, The Royal Society of Chemistry, U. K., 1994.
13. K. Ito, S. H. Ikeda and H. Noda, Solar World Congress Proceedings Biennial Congress International Solar Energy Society, Pergamon, Oxford, 1992.
14. R. L. Cook, R. C. MacDuff and A. F. Sammells, J. Electrochem. Soc., 1989, 136, 1982–1984.
15. D. W. DeWulf, T. Jin and A. J. Bard, J. Electrochem. Soc., 1989, 136, 1686–1691.
16. Y. Hori, A. Murata and R. Takahashi, J. Chem. Soc., Faraday Trans. 1, 1989, 2309–2326.
17. Y. Hori, H. Wakebe, T. Tsukamoto and O. Koga, Electrochim. Acta, 1994, 39, 1833–1839.
18. H. De Jesús-Cardona, C. del Moral and C. R. Cabrera, J. Electroanal. Chem., 2001, 513, 45–51.
19. C. Delacourt, P. L. Ridgway, J. B. Kerr and J. Newman, J. Electrochem. Soc., 2008, 155, B42–B49.
20. H. Noda, S. Ikeda, Y. Oda and K. Ito, Chem. Lett., 1989, 18, 289–292.
21. J. Krause, D. Borgmann and G. Wedler, Surf. Sci., 1996, 347, 1–10.
22. I. A. Bonicke, W. Kirstein and F. Thieme, Surf. Sci., 1994, 307−309, 177–181.
23. Y. Hori, A. Murata, T. Tsukamoto, H. Wakebe, O. Koga and H. Yamazaki, Electrochim. Acta, 1994, 39, 2495–2500.
24. J. Pritchard, Surf. Sci., 1979, 79, 231–244.
25. Y. Hori, O. Koga, H. Yamazaki and T. Matsuo, Electrochim. Acta, 1995, 40, 2617–2622.
26. A. A. Peterson, F. Abild-Pedersen, F. Studt, J. Rossmeisl and J. K. Nørskov, Energy Environ. Sci., 2010, 3, 1311–1315.
27. W. J. Durand, A. A. Peterson, F. Studt, A. A. Peterson and J. K. Nørskov, Surf. Sci., 2011, 605, 1354–1359.
28. K. J. P. Schouten, Y. Kwon, C. J. M. van der Ham, Z. Qin and M. T. M. Koper, Chem. Sci., 2011, 2, 1902–1909.
29. K. J. P. Schouten, Z. Qin, E. P. Gallent and M. T. M. Koper, J. Am. Chem. Soc., 2012, 134, 9864–9867.
30. X. Nie, M. R. Esopi, M. J. Janik and A. Asthagiri, Angew. Chem. Int. Ed., 2013, 52, 2459–2462.
31. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
32. D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 7892–7895.
33. M. Methfessel and A. T. Paxton, Phys. Rev. B: Condens. Matter Mater. Phys., 1989, 40, 3616–3621.
34. S. Baroni, A. Dal Corso, S. de Gironcoli and P. Giannozzi, PWSCF and PHONON: Plane-Wave Pseudo-Potential Codes, 2001, http://www.pwscf.org.
35. A. Kokalj, J. Mol. Graphics Modell., 1999, 17, 176–179.
36. A. Kokalj and M. Causà, Scientific Visualization in Computational Quantum Chemistry, In Proceedings of High Performance Graphics Systems and Applications European Workshop, CINECA-Interuniversity Consortium, Bologna, Italy, 2000.
37. A. Kokalj and M. Causà, XCrySDen: X-Window CRYstalline Structures and DENsities, 2001, http://www-k3.ijs.si/kokalj/xc/XCrySDen.html.
38. J. Greeley, A. A. Gokhale, J. Kreuser, J. A. Dumesic, H. Topsøe, N. Y. Topsøe and M. Mavrikakis, J. Catal., 2003, 213, 63–72.
39. CRC Handbook of Chemistry and Physics, CRC Press/Taylor and Francis, Boca Raton, FL, 91st edn, (Internet Version 2011), 2011.
40. G. Henkelman and H. Jonsson, J. Chem. Phys., 2000, 113, 9978–9985.
41. G. Henkelman, B. P. Uberuaga and H. Jonsson, J. Chem. Phys., 2000, 113, 9901–9904.
42. Y. M. Choi and P. Liu, J. Am. Chem. Soc., 2009, 131, 13054–13061.
43. A. Deluzarche, J. P. Hindermann, R. Kieffer, R. Breault and A. Kiennemann, J. Phys. Chem., 1984, 88, 4993–4995.
44. A. Kiennemann, R. Breault, J. P. Hindermann and M. Laurin, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 2119–2128.
45. C. Diagne, H. Idriss, J. P. Hindermann and A. Kiennemann, Appl. Catal., 1989, 51, 165–180.
46. I. N. Remediakis, F. Abild-Pedersen and J. K. Nørskov, J. Phys. Chem. B, 2004, 108, 14535–14540.
47. J. K. Nørskov, J. Rossmeisl, A. Logadottir and L. Lindqvist, J. Phys. Chem. B, 2004, 108, 17886–17892.
48. J. B. Hansen and P. E. Højlund Nielsen, Methanol synthesis, in Handbook of Heterogeneous Catalysis, ed. G. Ertl, H. Knözinger, F. Schüth and J. Weitkamp, 2008, vol. 13, pp. 2920–2949.
49. M. Jitaru, D. A. Lowy, M. Toma, B. C. Toma and L. Oniciu, J. Appl. Electrochem., 1997, 27, 875–889.
50. L. H. Ou, F. Yang, Y. W. Liu and S. L. Chen, J. Phys. Chem. C, 2009, 113, 20657–20665.
51. J. Onsgaard, L. Bech, C. Svensgaard, P. J. Godowski and S. V. Hoffmann, Prog. Surf. Sci., 2001, 67, 205–216.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra09294a

---