Theoretical insight into effect of doping of transition metal M (M = Ni, Pd and Pt) on CO₂ reduction pathways on Cu(111) and understanding of origin of electrocatalytic activity

Abstract

The effect of the doped transition metal M (M = Ni, Pd and Pt) on CO₂ reduction pathways and the origin of the electrocatalytic activity are investigated systematically by focusing on the CH₄ and CH₃OH formation pathways based on DFT calculations associated with the computational hydrogen electrode model. Our studies show that the doping of Ni, Pd and Pt can promote CO₂ reduction into hydrocarbons and influence the selectivity of reduction pathways, in which the doping of Pt may be able to lead to the strongest catalytic activity. The adsorption behavior between reaction intermediates and surfaces is crucial and the interactions of intermediates with the catalysts should be moderate in order to efficiently catalyze CO₂ reduction into CH₄ and CH₃OH, and avoid OH surface poisoning. The enhanced electrocatalytic activity of transition metal-doped Cu(111) surfaces may be owing to decreased overpotential and moderate electronic interactions between Cu and the doped transition metals. The doped Ni, Pd and Pt atoms can considerably decrease the overpotential and remove surface OH poisoning, in which the doped Pt can simultaneously reduce overpotential for CO formation and further reduction, and most easily remove OH, thus suggesting the best electrocatalytic activity. The moderate electron interaction between Cu and Pt and moderate upshift of the d-band center of Pt also explain why the Pt-doped Cu(111) surface has the best electrocatalytic activity for CO₂ reduction. Two possible descriptors can be proposed in order to scale the electrocatalytic activity of Cu-based electrocatalysts for CO₂ reduction, in which an ideal Cu-based electrocatalyst should be able to reduce barriers for CO formation and further reduction, and should have moderate electron interactions between Cu and the doped transition metals, and a moderate upshift of d-band center of the doped transition metals. In these ways, CO₂ reduction pathways can be facilitated and the yield of hydrocarbons CH₄ and CH₃OH can be enhanced.

---

1. Introduction

Excessive consumption of fossil fuels has gradually led to the increasing anthropogenic emission of CO₂ and depletion of finite natural resources.¹ To mitigate over-production of CO₂ and prepare for a fluctuating supply of fossil fuels, efforts are being made to convert CO₂ into reusable hydrocarbons.² Among various approaches for conversion, electrochemical reduction of CO₂ is considered a potentially clean and promising technique to produce various value-added hydrocarbons at only the cost of a sustainable supply of electrical energy.^3,4 However, the selectivity and faradic efficiency of the CO₂ electroreduction process are dependent on many factors, such as the electrode materials, the surface structures of the electrode and the type of electrolytes,^5–10 in which the major obstacle preventing the efficient reduction of CO₂ is the lack of electrode materials that can readily couple with electric energy to achieve rapid and selective cleavage of C–O bonds in CO₂ and formation of new bonds in the products. Towards this goal, various electrode materials and promoters have been screened experimentally and analyzed computationally to optimize their activity and selectivity for CO₂ reduction,^10–19 such as transition metals,¹⁵ ionic liquids,¹⁶ organometallic complexes^17,18 and doped graphene.¹⁹ Unfortunately, present electrocatalysts for CO₂ reduction suffer from high overpotentials, low current densities, low selectivity or poor durability.²⁰ Among these electrocatalysts explored to date, Cu has received considerable attention because it has been demonstrated experimentally as a valid low-cost electrocatalyst for CO₂ reduction^21,22 and is the only pure elemental catalyst that produces hydrocarbons at significant faradic efficiency.²⁰ Reduction products include formate, CH₄, C₂H₄ and other higher hydrocarbons.^23–26 Furthermore, theoretical study of CO₂ reduction on transition metals Cu, Pt, Rh, Pd, Ni, Au and Ag by Nørskov and co-authors²⁷ revealed the “volcano” type of the activity diagram, in which Cu has also the best catalytic activity and is located at the top of the diagram among these transition metals. Thus, experimental and theoretical studies all showed that Cu is the best known metal electrocatalyst for CO₂ reduction. However, in order to achieve high faradic efficiency, significant overpotentials are required on Cu electrodes. For example, the theoretical reversible potentials for CH₄ and C₂H₄ are 0.17 V and 0.08 V (vs. RHE), respectively, but experimentally it was observed that a very large potential of about −1.0 V is applied for CO₂ reduction into these hydrocarbons.²⁸ Simultaneously, the H₂ evolution reaction (HER) at sufficiently negative potential can reduce the faradic efficiency for CO₂ reduction by consuming protons and electrons. Thus, to make the CO₂ reduction more active and selective, properly designing new electrocatalysts is extremely urgent under the current situation.

In surface-catalyzed reactions, the surface selectivity, activity and stability can be dramatically improved through the doping of transition metals to the surface of a host, and some novel properties that are not present on the parent metal surfaces are often exhibited.^29–32 Thus, bimetallic Cu-based alloys with Cu-rich composition experimentally have been extensively used to improve CO₂ electroreduction owing to the high overpotential and low current density of the pure Cu surface, such as Cu–Ni, Cu–Zn, Cu–Cd, Cu–Sn, Cu–Pb, Cu–Au, Cu–Ag, Cu–Pd and Cu–Pt.^33–40 Compared to pure Cu, doping of other transition metals into Cu catalysts can modify the activation barrier for different steps, thereby leading to a reduction of overpotential and a major change of faradic efficiency. For example, Cu–Au alloy has higher faradic efficiency for CO₂ electroreduction than that of pure Cu,³⁷ and the experimental onset potential on the Cu–Au alloy was positively shifted, indicating that the overpotential of CO₂ reduction can be reduced through the doping of Au.³⁸ The Ni-doped Cu surface also displayed experimentally a superior catalytic activity with respect to the CH₃OH synthesis from a mixture of CO, CO₂ and H₂ in comparison with pure Cu.^39,40 Such an improvement was ascribed to the capability of Ni to promote CH₃OH production by activating CO₂ and stabilizing the intermediates owing to the higher oxygen affinity of Ni.⁴¹ Recently, a Pd–Cu catalyst for CO₂ electroreduction was investigated.⁴² A sharp increase of the reduction current and positive shift of potential were observed on the Pd–Cu electrode compared with those for pure Cu, indicating that the doping of Pd could effectively suppress the HER and enhance the CO₂ electroreduction activity. Actually, in more previous studies, both Pd single-crystal⁴³ and oxide-supported Pd catalysts^44,45 have also been shown to be catalytically active toward CH₃OH synthesis from CO₂ reduction. Given that Pt is also a metal with high chemical stability and oxygen affinity, the doping of Pt could be a promising approach to improve CO₂ electroreduction. Thus, a Cu–Pt alloy with high Cu concentration was developed by Xiong and co-authors. The greatly improved chemical stability and superior electrocatalytic activity towards CO₂ reduction were exhibited owing to the presence of Pt.³⁶ Moreover, the alloying of Cu with Pt can lower the decomposition activation energy of formate by up to 13%.⁴⁶ In the recent theoretical studies on CO₂ reduction, density functional theory (DFT) was employed by Liu and co-authors to investigate the CH₃OH synthesis reaction from CO₂ reduction on transition metals-doped Cu(111) surfaces.⁴⁷ The overall CH₃OH yield was increased by the doped Ni, Pd and Pt compared with pure Cu, suggesting that the doping of Ni, Pd and Pt is able to promote the CH₃OH production of the Cu(111) surface. Using DFT calculations associated with the standard hydrogen electrode model,^48,49 Hirunsit and co-authors performed a systematic thermodynamic investigation for CO₂ electroreduction into CH₄ and CH₃OH on Cu-based electrocatalysts with a Cu-rich composition of Cu₃X (X is Ag, Au, Co, Ni, Pd, Pt, Rh and Ir). The investigation exhibited that a considerably different electrocatalytic activity was produced and hydrocarbon selectivity was changed compared with those on the pure Cu surface, in which on the mostly Cu-based alloy catalysts CH₄ is more energetically favorable to be yielded than CH₃OH, and CH₃OH was found to be more favorable than CH₄ production on Cu₃Pd and Cu₃Pt surfaces. On most surfaces the potential-limiting step is the CO protonation with the exception of on Cu₃Au and Cu₃Co surfaces. Most recently, the activation of CO₂ on transition metal TM (TM = Fe, Co, Ni, Ru, Rh, Pd, Ag, Os, Ir, Pt and Au)-doped Cu(111) and Cu(100) surfaces was investigated with a dopant coverage of 1/9 ML by Qiu and co-authors using first-principle DFT calculations combined with a slab model.^50,51 The studies predicted that Co, Ru and Os may be potential dopants to enhance the chemisorption of CO₂ on both TM-doped Cu surfaces. However, previous experimental and theoretical studies only showed that the doped transition metals could reduce the thermodynamic overpotential of CO₂ reduction and activate CO₂ molecules. The exact CO₂ electroreduction mechanisms and origin of electrocatalytic activation on Cu-based electrocatalysts still remain unclear, especially kinetic analysis of the elementary reaction steps and the relationship between the kinetic barriers and adsorption ability of the reaction intermediates. Essentially, the doping of transition metals can change the CO₂ reduction pathways and reduce the activation barriers, thereby leading to the reduced overpotential and enhanced catalytic activity. A general understanding of the adsorption behavior, reaction mechanism and atomic-level origin of the electrocatalytic activity of Cu-based catalyst bimetallic surfaces will help us discover more efficient catalysts for a given reaction.

Based on the above analysis, Ni, Pd and Pt atoms as potential dopants may be able to enhance CO₂ reduction into hydrocarbons on the Cu surface. Furthermore, Cu, Ni, Pd and Pt have the same face-centered cubic structure, which gives them a crystallographic match. Thus, the activation barriers of key steps can be maneuvered through the miscible bimetallic combinations of Cu with Ni, Pd and Pt, thereby resulting in improved surface catalytic activity and selectivity. Focusing on the CH₄ and CH₃OH formation pathways, the effect of the doped transition metal M (M = Ni, Pd and Pt) on the CO₂ reduction pathways and the origin of the electrocatalytic activity are investigated systematically based on DFT calculations associated with the computational hydrogen electrode model. The (111) facet is chosen owing to its high selectivity for CH₄ production. The major goal of this work is to examine the effects of the doped Ni, Pd and Pt atoms on the CO₂ reduction pathways, surface electrocatalytic activity and selectivity in comparison with pure Cu catalyst. Some key factors that influence surface electrocatalytic activity, such as OH surface poisoning, the limiting potential, potential-limiting step and electronic interactions between Cu and Ni, Pd and Pt, are also demonstrated. The development of more superior electrocatalysts requires an essential understanding of these effects.

2. Computational method and modeling

All DFT calculations were implemented through the PWSCF codes included in the Quantum ESPRESSO distribution,⁵² while figures of the chemical structures were produced by the XCRYSDEN graphical package.^53–55 The exchange–correction interactions were treated by using the generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof (PBE).⁵⁶ The nuclei and core electrons were described by ultrasoft pseudopotentials.⁵⁷ The transition metal M (M = Ni, Pd and Pt)-doped Cu electrodes modeled by three-layer (111) slabs with a 2 × 3 supercell were used to perform the calculations of the all stable geometry structure of various adsorption and co-adsorption intermediates, which has been shown to be thick enough to well describe the CO₂ reduction. The calculated equilibrium lattice constant of Cu was 3.66 Å, which was in good agreement with experimental and theoretical values (3.62 and 3.66 Å, respectively).^58,59 During the geometry optimization, the bottom two atomic layers in the transition metal M (M = Ni, Pd and Pt)-doped Cu(111) slabs were kept fixed, whereas the top one layer and the adsorbates were relaxed to geometry configurations with the lowest energy. A vacuum region of 16 Å was used along the z-direction to avoid any periodic interactions. For the 2 × 3 supercell, the surface Brillouin zone was sampled by using a 4 × 3 × 1 uniformly shifted k-mesh. The Kohn–Sham orbitals were expanded in a plane-wave basis set, and the plane-wave cutoff energy was optimized at 26 Ry. The smearing technique of Methfessel and Paxton was used to treat the Fermi-surface effects with a value of smearing parameter of 0.02 Ry.⁶⁰ The convergence criteria for the total energy and Cartesian force components acting on each atom were set to within 10⁻⁵ Ry and below 10⁻³ Ry Bohr⁻¹, respectively. Additionally, it has been shown that dopants Pd and Pt prefer to stay in the surface of Cu(111), while Ni is in favor of the bulk.⁶¹ However, based on the experiments conducted by Chorkendorff and coauthors,^39,40 the active adsorbents are able to pull the Ni atom out to the surface under hydrocarbon synthesis conditions, such as CO. Thus, our present study did not include the segregation of Ni into the bulk. The doping range of dopants on the Cu-based electrocatalyst surfaces from ca. 1/9 to 1/4 has been studied in the previous experimental and theoretical studies,^{35,36,47,50,51,62} and the surface with dopant coverage of ca. 1/6 ML has superior electrocatalytic activity towards CO₂ reduction. Thus, only one surface Cu atom was substituted by a dopant (Ni, Pd and Pt) atom in each periodic supercell in the present study, corresponding to a dopant coverage of 1/6 ML. Selecting dopant coverage of 1/6 ML is also in order to compare with our previous studies on a pure 2 × 3 Cu(111) surface.⁶³

The minimum energy paths (MEPs) for each step of the CO₂ reduction into hydrocarbons were determined by the climbing-image nudged elastic band (CI-NEB) method.^64,65 The image of highest energy approximated the transition state of the optimized reaction coordinate, and the transition state images from the CI-NEB calculations were optimized by the quasi-Newton method, which minimizes the forces to find the saddle point. For each intermediate point in the MEPs, geometry optimization was also performed.

3. Results and discussion

3.1 The optimal reaction pathways of CO₂ reduction

The calculated results of CO₂ reduction into CH₄ and CH₃OH on the pure Cu(111) surface were reported in our previous work⁶³ and are summarized in the ESI† again to complete the systematic comparison. Various possible pathways obtained by minimum energy pathway (MEP) analysis during the course of CO₂ reduction on transition metal M (M = Ni, Pd and Pt)-doped Cu(111) are also demonstrated in the ESI.† Based on the MEP analyses of the elementary reaction steps, it can be concluded that CO, CHO and CH₂O are key reaction intermediates during the course of CO₂ reduction into hydrocarbons on Ni-, Pd- and Pt-doped Cu(111) surfaces, in which CO is firstly formed through CO₂ dissociative hydrogenation, and the COOH intermediate is involved during the course of CO formation. On the Ni-doped Cu(111) surface, formation of intermediates CH₃O, CH₂OH and CH₂ through further CH₂O hydrogenation and hydrogenative dissociation are all favorable reaction pathways, which may be parallel pathways in the CO₂ reduction mechanism based on the MEP analyses. CH₃OH formation more easily occurs through CH₂OH hydrogenation in CH₃O and CH₂OH further reduction on the Ni-doped Cu(111) surface. Simultaneously, CH₂ that is produced through CH₂O hydrogenative dissociation also leads to CH₄ formation. Notably, CH₂OH hydrogenation into CH₃OH formation and CH₂ serial hydrogenation into CH₄ require very low activation barriers on the Ni-doped Cu(111) surface, which can be all overcome by the thermoactive process at ambient temperature. Thus, we conclude that CH₄ and CH₃OH may be able to be formed during CO₂ reduction simultaneously on the Ni-doped Cu(111) surface. The MEP analyses on the Pd-doped Cu(111) surface show that CH₂O direct hydrogenation into CH₃O is the most favorable pathway among all possibilities of further CH₂O reduction. The most preferred reaction pathway among further CH₃O reduction is CH₃O hydrogenative dissociation into CH₃ on the Pd-doped Cu(111) surface. Thus, CH₄ can be formed through CH₃ hydrogenation. On the Pt-doped Cu(111) surface, direct hydrogenation of CH₂O to form CH₃O and CH₂OH intermediates are the most favorable pathways. Thus, CH₃O and CH₂OH may be reaction intermediates in CO₂ reduction on the Pt-doped Cu(111) surface. The MEP analyses show that CH₃ and CH₃OH can be formed easily through CH₃O hydrogenative dissociation and CH₂OH hydrogenation, respectively. Simultaneously, a CH₂ intermediate can be also formed through CH₂OH hydrogenative dissociation. Intermediates CH₂ serial hydrogenation and CH₃ hydrogenation possibly result in the formation of the final CH₄ product. Thus, we can conclude that CH₄ and CH₃OH in CO₂ reduction on the Pt-doped Cu(111) surface are also possible products. Based on thermodynamic and kinetic analyses, (CO + H)* → CHO* is the slowest step on these three surfaces. The optimal reaction pathways of CO₂ reduction on the pure and transition metal M (M = Ni, Pd and Pt)-doped Cu(111) surfaces are summarized in Fig. 1.

Fig. 1 The optimal reaction pathways of CO₂ reduction into CH₄ and CH₃OH on: (a) pure Cu(111); (b) Ni-doped Cu(111); (c) Pd-doped Cu(111); and (d) Pt-doped Cu(111).

3.2 Comparison of CO₂ reduction pathways among pure and Ni-, Pd- and Pt-doped Cu(111)

The reaction free energies and activation barriers for the optimal CO₂ reduction pathways into CH₄ and CH₃OH production are listed in Tables 1 and 2 on the pure and transition metal Ni-, Pd- and Pt-doped Cu(111) surfaces.

Table 1 The optimal CO₂ reduction pathways into hydrocarbons CH₄ and CH₃OH and the reaction free energies (E_reac) and activation barriers (E_act) for each elementary step on the pure and Ni-doped Cu(111) surfaces^a

Reaction paths	Pure Cu(111)		Ni-doped Cu(111)
	ΔGreac (eV)	Eact (eV)	ΔGreac (eV)	Eact (eV)
a The asterisk (*) indicates that the species is adsorbed on the surface. The zero point energies (ZPE) for all species, which are obtained by the present DFT calculations, are included in the reaction free energy. For instance, for the reaction step of (CO + H)* → CHO*, the reaction free energy is calculated according to “E(CHO*) − E(CO + H)* + ZPE(CHO*) − ZPE(CO*) − ZPE(H*)”.
CO2(g) + H* → (CO + OH)*	0.27	1.18	−0.23	0.40
(CO + H)* → CHO*	0.90	1.06	1.03	1.08
(CHO + H)* → CH2O*	−0.20	0.72	−0.04	0.28
(CH2O + H)* → (CH2 + OH)*	0.01	1.12	0.06	0.24
(CH2O + H)* → CH2OH*	0.13	0.95	−0.60	0.27
(CH2O + H)* → CH3O*	−1.05	1.86	−0.08	0.24
(CH2OH + H)* → CH2* + H2O (l)	−0.19	0.66	−0.13	0.58
(CH2OH + H)* → CH3OH (l)	−0.95	0.68	−0.61	0.16
(CH2 + H)* → CH3*	−0.83	0.63	−0.55	0.30
(CH3 + H)* → CH4*	−0.80	1.03	−0.59	0.21

---

Table 2 The optimal CO₂ reduction pathways into hydrocarbons CH₄ and CH₃OH and the reaction energies (E_reac) and activation barriers (E_act) for each elementary step on Pd- and Pt-doped Cu(111) surfaces

Reaction paths	Pd-doped Cu(111)		Pt-doped Cu(111)
	ΔGreac (eV)	Eact (eV)	Ereac (eV)	ΔGreac (eV)
CO2(g) + H* → (CO + OH)*	0.31	0.83	0.07	0.65
(CO + H)* → CHO*	0.51	0.87	0.41	0.84
(CHO + H)* → CH2O*	−0.25	0.10	−0.02	0.35
(CH2O + H)* → (CH2 + OH)*	0.24	0.52	0.26	0.82
(CH2O + H)* → CH2OH*	−0.12	0.68	−0.35	0.59
(CH2O + H)* → CH3O*	−0.80	0.06	−0.62	0.17
(CH2OH + H)* → CH2* + H2O (l)	−0.01	0.33	0.10	0.60
(CH2OH + H)* → CH3OH (l)	−0.86	0.30	−0.53	0.34
(CH3O + H)* → CH3OH (l)	−0.10	1.86	−0.17	1.36
(CH3O + H)* → (CH3 + OH)*	0.01	0.29	−0.12	0.63
(CH2 + H)* → CH3*	−0.83	0.35	−0.95	0.33
(CH3 + H)* → CH4*	−1.01	0.23	−0.82	0.70

---

Based on the data in Tables 1 and 2, potential energy diagrams for formation of CH₄ and CH₃OH through CO₂ reduction on the pure and transition metal Ni-, Pd- and Pt-doped Cu(111) surfaces are given in Fig. 2–5, respectively. We observed that the activation barriers of CO formation through initial CO₂ reduction are considerably decreased by the doping of Ni, Pd and Pt, which is the rate-determining step of CO₂ reduction on the pure Cu(111) surface (ca. 1.18 eV). Thus, the doping of Ni, Pd and Pt can activate CO₂ and enhance the surface activity of Cu catalysts for CO₂ reduction. Therefore, CO formation requires significantly lower activation barriers on Ni and Pt-doped Cu(111) surfaces (ca. 0.40 and 0.65 eV), whereas an activation barrier of 0.83 V is required after doping of Pd (see Fig. S1, S19 and S16†), which may be attributed to chemisorbed CO₂ molecule being observed after doping of Ni and Pt by binding C atom at Ni and Pt sites, whereas only physisorbed CO₂ was observed on the Pd-doped Cu(111) surface. For further CO reduction, CO hydrogenation into CHO was shown to be the preferred reaction pathway on the Ni-, Pd- and Pt-doped Cu(111) surfaces, as obtained on the pure Cu(111) surface.⁶³ Our previous study and the work from Mei and co-authors^63,66 have shown that CO prefers to desorb from Cu rather than to undergo further hydrogenation into CHO, which is in favor of dissociation back to CO and H. For example, CO hydrogenation to form CHO needs to overcome an activation barrier of 1.06 eV on the pure Cu(111) surface, whereas the inverse process only requires an activation barrier of 0.16 eV. The corresponding pathway has nearly equal activation barriers on the Ni-doped Cu(111) surface and the pure Cu(111) surface (1.06 eV vs. 1.08 eV), and a very low activation barrier for CHO dissociation back to CO was required (ca. 0.10 eV). Both steps may hinder CH₄ and CH₃OH formation in CO₂ reduction on the pure and Ni-doped Cu(111) surfaces. In fact, a large amount of CO was observed experimentally on Cu for CH₃OH synthesis.⁶⁷ Based on our present calculations, the doping of Pd and Pt helps to improve the corresponding hydrogenation process. The activation barriers are 0.87 and 0.84 eV for CO hydrogenation after the doping of Pd and Pt, which are 0.19 and 0.22 eV lower than the corresponding barrier on the pure Cu(111) surface, respectively. Moreover, the activation barriers for CHO dissociation back to CO are increased on Pd- and Pt-doped Cu(111) surfaces (see Fig. S11 and S18†), which are 0.10 and 0.27 eV higher than the corresponding barriers on the pure Cu(111) surface, respectively. Improvement of further CO reduction may be attributed to the stronger CO adsorption. The adsorption energy of CO is 0.13 and 0.48 eV stronger after doping of Pd and Pt than that on the pure Cu(111) surface, respectively, i.e. the CO intermediate can be stabilized. Our present studies also show that the CO molecule can be stabilized by the doping of Ni. However, CO adsorption is stronger than that on the pure Cu(111) surface by 0.80 eV after doping of Ni, which is significantly higher than that on the Pd- and Pt-doped Cu(111) surfaces, thereby leading to a relatively poor improvement for further CO reduction into CHO. Although the doped Ni atom cannot significantly improve the further CO reduction into CHO, it makes CO formation the most facile to occur. According to our calculations, CO prefers to adsorb on the top sites of Ni, Pd and Pt. In comparison, the energetically most favorable site for CO adsorption is the fcc-hollow site on the pure Cu(111) surface.⁶³

Fig. 2 Potential energy diagrams for the CH₄ and CH₃OH formation through CO₂ reduction on the pure Cu(111) surface. TS stands for transition state.

Fig. 3 Potential energy diagrams for CH₄ and CH₃OH formation through CO₂ reduction on the transition metal Ni-doped Cu(111) surface.

Fig. 4 Potential energy diagrams for CH₄ and CH₃OH formation through CO₂ reduction on the transition metal Pd-doped Cu(111) surface.

Fig. 5 Potential energy diagrams for CH₄ and CH₃OH formation through CO₂ reduction on the transition metal Pt-doped Cu(111) surface.

Subsequent further CHO reduction into the CH₂O intermediate is the most favorable pathway among various possibilities on the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces (see Fig. S4, S12 and S19†). The activation barriers of 0.28, 0.10 and 0.35 eV are required for this pathway, respectively, which are significantly lower than that on the pure Cu(111) surface (ca. 0.72 eV). Simultaneously, it can be found that the CHOH intermediate may be formed on the Pt-doped Cu(111) surface since CH₂O and CHOH formation through CHO hydrogenation have nearly equal activation barriers (ca. 0.35 and 0.40 eV). We also observed that the further reduction of CHO is energetically comparable to CHO dissociation back to CO after the doping of Ni, Pd and Pt. Thus, although the doping of Ni has a poor improvement for further CO reduction, it considerably improves the further CHO reduction into key intermediate CH₂O. The significantly lower activation barriers after the doping of Ni, Pd and Pt may be owing to the stronger CHO adsorption. CHO prefers to bind at Ni, Pd and Pt sites through a C atom after the doping of Ni, Pd and Pt, which is 0.67, 0.29 and 0.68 eV more stable than that on the pure Cu(111) surface, respectively. Thus, the CHO intermediate is able to be stabilized and further CHO reduction can be promoted by the doping of Ni, Pd and Pt atoms.

On the Ni-doped Cu(111) surface, CH₂O hydrogenation into CH₃O and CH₂OH, and hydrogenative dissociation into CH₂ may be parallel pathways since they have very low and almost equal activation barriers. The barriers for the corresponding pathways are 0.24, 0.27 and 0.24 eV, respectively. Only CH₃O formation is the most preferred pathway in further CH₂O reduction on the Pd-doped Cu(111) surface with a very low activation barrier of 0.06 eV, which is a non-activated process, as shown in Fig. S13.† On the Pt-doped Cu(111) surface, CH₃O and CH₂OH formations through CH₂O hydrogenation are the most favorable pathways among four possibilities owing to the relatively lower activation barriers (0.21 and 0.59 eV), as shown in Fig. S20.† However, considerably higher activation barriers are required on the pure Cu(111) surface for further CH₂O reduction compared with that on Ni-, Pd- and Pt-doped Cu(111), and the most favorable reduction pathway is hydrogenation of CH₂O into CH₂OH with an activation barrier of 0.95 eV. The significantly lower barriers may be owing to the doping of Ni, Pd and Pt stabilizing the CH₂O intermediate, which adsorbs on the Ni-, Pd- and Pt-doped Cu(111) surfaces through the C atom directly interacting with Ni, Pd and Pt and O atoms interacting with Cu and transition metal M (M = Ni, Pd and Pt). In comparison, the adsorption of CH₂O intermediate on the Ni-, Pd- and Pt-doped Cu(111) surfaces is 0.55, 0.19 and 0.31 eV more stable than that on pure Cu(111), respectively. On the Pt-doped Cu(111) surface, CH₂ and CH₂OH may be able to be formed through CHOH hydrogenative dissociation and direct hydrogenation owing to almost identical activation barriers (ca. 0.53 and 0.54 eV). However, CHOH dissociation back into CHO only requires a very low activation barrier of ca. 0.04 eV, as shown in Fig. S4 and S19,† which is a non-activated process. Thus, although CHOH can be formed on the Pt-doped Cu(111) surface, its further reduction may not be able to occur.

Further reduction of CH₂OH intermediate into CH₃OH is more favorable to occur with an activation barrier as low as 0.16 eV on the Ni-doped Cu(111) surface, which is significantly more facile in contrast with CH₃O further reduction, as shown in Fig. S6 and S7.† Thus, CH₃OH can be formed easily by CH₂OH hydrogenation after the doping of Ni atom. CH₃O hydrogenative dissociation into the CH₃ intermediate is the most preferred pathway in further CH₃O reduction on the Pd- and Pt-doped Cu(111) surfaces with relatively lower activation barriers of 0.29 and 0.63 eV, respectively. On the Pt-doped Cu(111) surface, CH₃OH and CH₂ can be also formed easily through hydrogenation and hydrogenative dissociation in further reduction of CH₂OH with relatively lower activation barriers of 0.34 and 0.60 eV, respectively. Thus, the doping of Pt may lead to simultaneous formation of CH₂, CH₃ and CH₃OH, as shown in Fig. S22 and S23.† Our previous studies also suggested that CH₃OH and CH₂ were formed on the pure Cu(111) surface through further reduction of CH₂OH with activation barriers of 0.68 and 0.66 eV, respectively.⁶³ However, the doping of Ni and Pt leads to a significant decrease of the activation barriers for CH₃OH formation. Thus, further CH₂OH reduction can be improved through the doped Ni and Pt atoms, whereas the doping of Pd promotes CH₃O formation and further reduction, leading to enhanced CH₄ and CH₃OH yields. Further reduction of CH₃O was not considered on the pure Cu(111) surface owing to the significantly higher activation barrier of CH₃O formation. The significant promotion effect on further CH₂OH reduction may be owing to the stronger CH₂OH adsorption after doping of Ni and Pt atoms. The binding of CH₂OH is 0.39 and 0.51 eV more stable than that on the pure Cu(111) surface, and it prefers to bind at the top site of the Ni and Pt atoms by C atom on the Ni- and Pt-doped Cu(111) surfaces. Compared with that on the pure Cu(111) surface, CH₃O on the Ni-doped Cu(111) surface is 0.05 eV more stable. However, CH₃O is 0.09 and 0.25 eV less stable after the doping of Pd and Pt than that on the pure Cu(111) surface, which may explain why the CH₃ intermediate can be formed through hydrogenative dissociation of CH₃O after doping of Pd and Pt.

Based on the above analyses, CH₄ production possibly occurs through serial CH₂ hydrogenation on the Ni- and Pt-doped Cu(111) surfaces or direct CH₃ hydrogenation on the Pd- and Pt-doped Cu(111) surfaces. The activation barriers for CH₂ hydrogenation into CH₃ and CH₃ hydrogenation into CH₄ are 0.30 and 0.21 eV, 0.35 and 0.23 eV, 0.33 and 0.70 eV on Ni, Pd- and Pt-doped Cu(111) surfaces (see Fig. S8, S15 and S24†), respectively, which are notable lower than that for the corresponding pathways on the pure Cu(111) surface (0.63 eV and 1.03 eV, respectively). Based on our present calculations, the doping of Ni and Pt helps to stabilize CH₂ and CH₃ intermediates and enhance the corresponding hydrogenation processes, in which CH₂ and CH₃ are 0.43 and 0.22 eV, 0.18 and 0.23 eV more stable after doping of Ni and Pt, respectively. Although CH₃ is 0.07 eV less stable on the Pd-doped Cu(111) surface, the doping of Pd still can improve further CH₃ reduction into CH₄. This may be attributed to different electronic interactions among Ni and Pd, Ni and Pt, which will be confirmed by subsequent electronic structure analysis. Additionally, OH can be formed during the course of CO₂ reduction into CH₄ and CH₃OH on the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces, which may lead to poisoning of surface active sites. Thus, the OH removal to form H₂O is an important step for CO₂ reduction. The activation barriers for OH removal are 1.09, 0.68, 0.56 and 0.21 eV on pure and Ni-, Pd- and Pt-doped Cu(111) surfaces, respectively, as shown in Fig. 6. We observe that the doping of Ni, Pd and Pt atoms significantly decreases the activation barriers of OH removal, which is 0.41, 0.53 and 0.88 eV lower than the corresponding barrier on pure Cu(111), respectively. Therefore, the reduced value of the activation barrier after the doping of Ni atom is lower than that on the Pd- and Pt-doped Cu(111) surfaces. The significant improvement of OH removal on the Pd- and Pt-doped Cu(111) may be owing to the weaker OH adsorption, in which the adsorption of OH is 0.16 and 0.26 eV less stable than that on the pure Cu(111) surface, respectively, whereas the doping of Ni makes OH adsorption 0.08 eV stronger. Thus, the doping of metals with less affinity toward OH, such as Pd and Pt, improves the OH removal step and proves to be more efficient than pure Cu and high OH affinity metal Ni.

Fig. 6 The minimum energy pathways of OH removal to form H₂O on pure and Ni-, Pd- and Pt-doped Cu(111) surfaces.

On this basis, the effect of the doped Ni, Pd and Pt atoms on the selectivity of the reaction pathways and reduction production can be revealed. The doped Ni, Pd and Pt atoms changed the rate-determining step of CO₂ reduction, which is (CO₂ + H)* → (CO + OH)* on the pure Cu(111) surface, whereas it is changed into (CO + H)* → CHO* after the doping of Ni, Pd and Pt. The activation barriers of rate-determining steps are also reduced, which are 0.10, 0.31 and 0.34 eV lower than that of the corresponding process on the pure Cu(111) surface, respectively. Simultaneously, we observed that the activation barriers in overall optimal pathways for CO₂ reduction into CH₄ and CH₃OH are 1.88, 1.12, 1.18 and 0.90 eV on the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces, respectively, as shown in Fig. 2–5, in which the barriers are considerably decreased compared with that on pure Cu(111) and the doping of Pt may be able to lead to the strongest catalytic activity owing to the most reduced value. Therefore, higher CH₄ and CH₃OH yields can be expected on the Ni-, Pd- and Pt-doped Cu(111) surfaces. The higher reactivity of Ni-, Pd- and Pt-doped Cu(111) surfaces is also consistent with the previous experimental observations.^36,39–42 The adsorption behavior of intermediates may be able to determine CO₂ reduction activity based on the above discussion. The most stable adsorption sites and adsorption energies for the possible intermediates involved in CO₂ reduction into CH₄ and CH₃OH on the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces are given in Table 3. It is clear that the doping of Ni, Pd and Pt atoms can stabilize the relatively weaker adsorbed C- and O-containing intermediates on the pure Cu(111) surface, such as CO, CHO and CH₂O, where C and O atoms directly interact with the doped Ni, Pd and Pt atoms, thereby resulting in easier formation and reduction of CO, CHO and CH₂O in spite of the doped Ni atom not notably improving further CO reduction into CHO owing to excessively strongly adsorbed CO. However, the stronger adsorbed intermediates, such as CH₃O and OH, are destabilized by the doped Pd and Pt atoms in comparison with pure Cu(111), which make CH₃O formation and further reduction able to occur on the Pd- and Pt-doped Cu(111) surfaces and OH removal be easier. Owing to the high carbon and oxygen affinity of Ni, all intermediates can be stabilized by the doped Ni atom in CO₂ reduction and the Ni-doped Cu(111) surface has the strongest adsorption ability among the Ni-, Pd- and Pt-doped Cu(111) surfaces, whereas the doping of Pd and Pt leads to the weakest and moderate adsorption of most intermediates, respectively, which may explain why Pt-doped Cu(111) surfaces have the best catalytic activity for CO₂ reduction. Thus, it is concluded that moderate adsorbed intermediates on the Cu-based electrocatalyst surfaces will favor CO₂ reduction.

Table 3 The most stable adsorption sites and adsorption energies (E_ads) for the possible intermediates involved in CO₂ reduction into CH₄ and CH₃OH on pure and Ni-, Pd- and Pt-doped Cu(111) surfaces

Species	Cu(111)		Ni-doped Cu(111)		Pd-doped Cu(111)		Pt-doped Cu(111)
	Sites	Eads (eV)	Sites	Eads (eV)	Sites	Eads (eV)	Sites	Eads (eV)
COOH	Bridge	−1.56	Bridge	−1.95	Bridge	−1.69	Bridge	−2.03
CO	fcc	−0.80	Top	−1.59	Top	−0.93	Top	−1.28
CHO	fcc	−1.25	Top	−1.93	Top	−1.54	Top	−1.93
CH2O	fcc	−0.05	fcc	−0.60	fcc	−0.24	fcc	−0.36
CHOH	fcc	−2.50	Top	−3.02	fcc	−2.60	Top	−2.95
CH2OH	fcc	−0.90	fcc	−1.29	fcc	−1.11	fcc	−1.41
CH3O	fcc	−2.28	fcc	−2.33	fcc	−2.19	hcp	−2.03
CH2	fcc	−2.94	fcc	−3.37	fcc	−2.90	fcc	−3.12
CH3	fcc	−1.36	fcc	−1.58	Top	−1.29	Top	−1.59
OH	fcc	−3.16	fcc	−3.24	hcp	−3.00	hcp	−2.90

---

3.3 Origin of catalytic activity

3.3.1 The limiting potential and potential-limiting step. The Ni-, Pd- and Pt-doped Cu(111) surfaces may be able to decrease the overpotentials of CO₂ electroreduction, which can be linked directly to the proton and electron transfer to adsorbed intermediates being strongly bonded to the surface and determines the catalytic activity of the electrocatalysts. Thus, potential dependent reaction free energies and limiting potentials are obtained in order to ascertain the origin of the overpotential and electrocatalytic activity using the computational hydrogen electrode (CHE) model proposed by Nørskov and co-authors.⁶⁸ The mechanistic details are elucidated in the ESI.† A thermodynamic equilibrium potential of 0.17 V vs. RHE for hydrocarbons formation is used to calculate the reaction free energies if we assume CO₂ reduction to be in equilibrium and neglect ohmic losses. This model could provide an elegant way to calculate the potential-dependent reaction free energies for each step under electrochemical conditions by avoiding the explicit treatment of solvated protons. The potential where the reaction free energy is zero is referred to as the limiting potential, which serves as an estimate of the onset potential for each intermediate; the most endothermic reaction step in the reaction pathways is referred to as the potential-limiting step.

The calculated limiting potentials (E) and reaction free energies (ΔG_reac) at 0.17 V vs. RHE in CO₂ electroreduction to CH₄ and CH₃OH on pure and Ni-, Pd-, and Pt-doped Cu(111) surfaces are listed in Tables 4 and 5. We found that the potential-limiting step is CHO formation (CO* + H⁺ + e⁻ → CHO*) with the most positive reaction free energy at 0.17 V vs. RHE on the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces, requiring limiting potentials of −0.60, −0.72, −0.44 and −0.40 V vs. RHE by the CHE model, respectively. It can be observed that the limiting potentials are positively shifted after the doping of Pd and Pt atoms, whereas it is more negative after the doping of Ni atom, implying that the doped Pd and Pt atoms can decrease the overpotential of this pathway. The strongest CO adsorption on the Ni-doped Cu(111) surface may induce the protonation to be potentially limited, resulting in a relatively high overpotential for CHO formation. The formation of CHO is also the rate-determining step in CO₂ reduction on the Ni-, Pd- and Pt-doped Cu(111) surfaces based on the above kinetic analyses, and its activation barriers are lower after the doping of Pd and Pt atoms, and slightly higher after the doping of Ni atom, compared with that on the pure Cu(111) surface. Although the doped Ni atom leads to more negative limiting potential (−0.72 V) and more positive reaction free energy (0.89 eV) for CHO formation, it can notably shift the limiting potential and reaction free energy to more positive (1.11 V) and more negative values (−0.94 eV) for CO formation (CO₂* + H⁺ + e⁻ → (CO + OH)*), respectively, in which the formation of CO has a slightly negative limiting potential of −0.07 V and a positive reaction free energy of 0.24 eV on the pure Cu(111) surface. The above kinetic studies also show that the activation barrier of CO formation is significantly decreased by the doping of Ni atom. Thus, the decreased overpotential for CO formation can be concluded after the doping of Ni. As shown in Table 5, the doping of Pt atom also can significantly shift the limiting potential of CO formation to a more positive value and make the reaction free energy more negative, thus leading to decrease of the overpotential. However, the limiting potential and reaction free energy of CO formation after the doping of Pd are almost unchanged compared with that on the pure Cu(111) surface. These may be owing to chemisorbed CO₂ being observed on Ni- and Pt-doped Cu(111) surfaces and only physisorbed CO₂ is formed on the Pd-doped Cu(111) surface. Simultaneously, the limiting potentials of OH removal are also shifted positively after the doping of Ni, Pd and Pt, which are 0.06, 0.24 and 0.36 V more positive than that on the pure Cu(111) surface, respectively, leading to the reduction of the overpotential. Therefore, the Pt-doped Cu(111) surface with the weakest OH adsorption has the most positive limiting potential and the most negative reaction free energy, which is in agreement with the above kinetic analyses, suggesting that the doped metals with less affinity toward OH could decrease surface OH poisoning and enhance the catalytic activity. Thus, the best electrocatalytic activity of the Pt-doped Cu(111) surface for CO₂ reduction may be able to be attributed to the simultaneous reduction of overpotential for CO formation and further reduction, and the easiest OH removal.

Table 4 The limiting potentials (E) and reaction free energies (ΔG_reac) at 0.17 V vs. RHE (thermodynamic equilibrium potential) in CO₂ electroreduction to CH₄ and CH₃OH on the pure and Ni-doped Cu(111) surfaces

Possible reduction steps	Pure Cu(111)		Ni-doped Cu(111)
	E (V)	ΔGreac (eV)	E (V)	ΔGreac (eV)
(a) CO2* + H^+ + e^− → (CO + OH)*	−0.07	0.24	1.11	−0.94
(b) CO* + H^+ + e^− → CHO*	−0.60	0.77	−0.72	0.89
(c) CHO* + H^+ + e^− → CH2O*	0.45	−0.28	0.32	−0.15
(d) CH2O* + H^+ + e^− → (CH2 + OH)*	0.38	−0.21	0.10	0.07
(e) CH2O* + H^+ + e^− → CH3O*	1.26	−1.09	0.76	−0.59
(f) CH2O* + H^+ + e^− → CH2OH*	0.23	−0.06	0.07	0.10
(g) CH2OH* + H^+ + e^− → CH3OH (l)	1.19	−1.02	0.76	−0.59
(h) CH2OH* + H^+ + e^− → CH2* + H2O (l)	0.43	−0.26	0.40	−0.23
(i) CH3O* + H^+ + e^− → CH3OH (l)	0.17	0.00	0.07	0.10
(j) CH3O* + H^+ + e^− → (CH3 + OH)*	0.24	−0.07	0.22	−0.05
(k) CH2* + H^+ + e^− → CH3*	1.05	−0.88	0.84	−0.67
(m) CH3* + H^+ + e^− → CH4*	1.07	−0.90	0.86	−0.69
(n) OH* + H^+ + e^− → H2O (l)	0.12	0.05	0.18	−0.01

---

Table 5 The limiting potentials (E) and reaction free energies (ΔG_reac) at 0.17 V vs. RHE (thermodynamic equilibrium potential) in CO₂ electroreduction to CH₄ and CH₃OH on the Pd- and Pt-doped Cu(111) surfaces

Possible reduction steps	Pd-doped Cu(111)		Pt-doped Cu(111)
	E (V)	ΔGreac (eV)	E (V)	ΔGreac (eV)
(a) CO2* + H^+ + e^− → (CO + OH)*	−0.11	0.28	0.82	−0.65
(b) CO* + H^+ + e^− → CHO*	−0.44	0.61	−0.40	0.57
(c) CHO* + H^+ + e^− → CH2O*	0.35	−0.18	0.08	0.09
(d) CH2O* + H^+ + e^− → (CH2 + OH)*	−0.08	0.25	−0.16	0.33
(e) CH2O* + H^+ + e^− → CH3O*	0.97	−0.80	0.69	−0.52
(f) CH2O* + H^+ + e^− → CH2OH*	0.25	−0.08	0.43	−0.26
(g) CH2OH* + H^+ + e^− → CH3OH (l)	0.93	−0.76	0.62	−0.45
(h) CH2OH* + H^+ + e^− → CH2* + H2O (l)	0.11	0.06	0.00	0.17
(i) CH3O* + H^+ + e^− → CH3OH (l)	0.21	−0.04	0.36	−0.19
(j) CH3O* + H^+ + e^− → (CH3 + OH)*	0.01	0.16	0.32	−0.15
(k) CH2* + H^+ + e^− → CH3*	1.03	−0.86	1.11	−0.94
(m) CH3* + H^+ + e^− → CH4*	1.14	−0.97	0.84	−0.67
(n) OH* + H^+ + e^− → H2O (l)	0.36	−0.19	0.48	−0.31

---

3.3.2 The analysis of electronic structure. The adsorption behavior and reduction mechanism may be able to be determined by electronic interactions between the Cu(111) surface and the doped transition metals M (M = Ni, Pd and Pt). The analyses of electronic structures facilitate the understanding the origin of catalytic activity of the transition metal Ni-, Pd- and Pt-doped Cu(111) surfaces toward CO₂ reduction. Thus, in order to determine the origin of the catalytic activity and understand the electron transfer between the transition metal M (M = Ni, Pd and Pt) and the Cu(111) surface, we performed local density of states (LDOS) analysis of pure and Ni-, Pd- and Pt-doped Cu(111) surfaces in the present study since it can describe the number of electron at Fermi energy level that are available to be occupied. A high LDOS at the Fermi energy level means that there are many electrons available for occupation. Fig. 7 and 8 give the LDOS of the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces. For comparison, the LDOS of pure Ni, Pd and Pt are also included. The results show that the doping of Ni, Pd and Pt leads to an obvious change of the s state of Cu at the Fermi energy level compared with that of the pure Cu(111) surface, in which s orbitals of Cu in Ni- and Pt-doped Cu(111) have relatively lower LDOS, as shown in Fig. 7(a), implying that more electrons in the s states of Cu are transferred after the doping of Ni and Pt. The LDOS of the d orbital of Cu in the pure and Ni-, Pd- and Pt-doped Cu(111) are almost unchanged, as shown in Fig. 7(b), suggesting that electron transfer may not occur in the d state of Cu. Thus, the observed significantly lower LDOS of the s states of Cu in Ni- and Pt-doped Cu(111) means that more electrons are transferred, thereby leading to stronger adsorption of reaction intermediates in CO₂ reduction on Ni- and Pt-doped Cu(111) surfaces than that on the Pd-doped Cu(111) surface (Table 3).

Fig. 7 Local density of states of pure and transition metals Ni-, Pd- and Pt-doped Cu(111) surfaces: (a) s orbital of Cu and (b) d orbital of Cu.

Fig. 8 Local density of states of pure and transition metals Ni-, Pd- and Pt-doped Cu(111) surfaces: (a) s orbitals of Ni, Pd and Pt; and (b) d orbitals of Ni, Pd and Pt.

In the mean time, we also observed that the s and d states of Ni, Pd and Pt in Ni-, Pd- and Pt-doped Cu(111) are changed significantly near the Fermi energy level compared with those of pure Ni, Pd and Pt, as shown in Fig. 8(a) and (b). The considerably higher LDOS near the Fermi energy level shows that electrons in the s state of Cu are transferred into the s and d states of Ni, Pd and Pt. The highest LDOS in the d states of Ni in Ni-doped Cu(111) may imply that the most electron transfer and strongest interactions occur between Cu and Ni. Thus, the present studies explain why the reaction intermediates have the strongest adsorption on the Ni-doped Cu(111) surface. The relatively weaker interaction between Cu and transition metals Pd and Pt may result in weaker adsorption of intermediates.

The d-band center of surface atoms, ε_d, is a key parameter that influences surface adsorption characteristics,^69,70 and its upshift and downshift can be used to judge the catalytic activity of the electrocatalysts.^71–73 The shift of the d-band center into a lower energy level (downshift) corresponds to the weaker adsorption of reaction intermediates, whereas the shift into a higher energy level (upshift) corresponds to the stronger adsorption. Thus, the d-band centers of surface atoms were calculated in pure and Ni-, Pd- and Pt-doped Cu(111), which is the first moment of the projected d-band density of states on the surface atoms referenced to the Fermi energy level. The corresponding values for surface atoms are listed in Table 6. It was found that there is a slight and considerably upshift of the d-band center of surface Cu and Ni atoms (ca. 0.02 and 0.74 eV) in Ni-doped Cu(111) compared with that of the pure metal surfaces, respectively, whereas that of surface Cu atoms is not changed after the doping of Pd and Pt atoms. The relatively lower upshift of the d-band center of the surface Pt atom in Pt-doped Cu(111) is also observed (ca. 0.20 eV). However, the d-band center of the surface Pd atom is downshifted significantly in Pd-doped Cu(111) (ca. −0.34 eV) compared with that of pure Pd. Thus, we conclude that the doping of Ni and Pd can lead to the strongest and weakest adsorption of reaction intermediates on the Ni and Pd-doped Cu(111) surfaces, respectively. Furthermore, the significant upshift of the d-band center of the surface Ni atom and the strongest interaction between CO and the Ni-doped Cu(111) surface may also be able to explain why further CO reduction has the most positive activation barrier and the most negative limiting potential on the Ni-doped Cu(111) surface among the Ni-, Pd- and Pt-doped Cu(111) surfaces, whereas the significant downshift of the d-band center of the surface Pd atom in Pd-doped Cu(111) explains why only physisorbed CO₂ is observed, which even leads to slightly more negative limiting potential for CO formation in comparison with that on pure Cu(111). Accordingly, we infer that the dopant Pt with moderate upshift of the d-band center is the most capable of enhancing the electrocatalytic activity of Cu catalysts for CO₂ reduction, which confirms the above thermodynamic and kinetic studies.

Table 6 The DFT calculated d-band center, ε_d of surface atoms in pure and Ni-, Pd- and Pt-doped Cu(111)

Surface atoms		εd (eV)	Δεd^a (eV)
a Δεd represents difference of the d-band center, εd of surface atoms between surface atoms in Ni-, Pd- and Pt-doped Cu(111) and pure surface atoms.
Pure Cu		−2.52	—
Pure Ni		−1.79	—
Pure Pd		−2.02	—
Pure Pt		−2.69	—
Ni-doped Cu(111)	Cu	−2.50	0.02
	Ni	−1.05	0.74
Pd-doped Cu(111)	Cu	−2.52	0.00
	Pd	−2.36	−0.34
Pt-doped Cu(111)	Cu	−2.52	0.00
	Pt	−2.49	0.20

---

Based on our present calculations, CO formation and further reduction are key reaction steps during the course of CO₂ reduction into hydrocarbons. CO₂ reduction on the Cu catalyst can be promoted and higher hydrocarbons yield may be able to be expected by the doping of transition metals. On the basis of the above discussion, two possible descriptors can be proposed in order to scale the electrocatalytic activity of Cu-based alloy catalysts for CO₂ reduction. One is activation barriers of CO formation and further reduction. The transition metal-doped Cu-based alloy electrocatalysts with lower activation barriers for CO formation and further reduction will exhibit better catalytic activity for CO₂ reduction. For example, the barriers of CO formation are decreased by the doped Ni, Pd and Pt atoms, and those of further CO reduction on the Pd- and Pt-doped Cu(111) surfaces are also reduced. The chemisorbed CO₂ on Cu-based electrocatalyst surfaces will favor the formation of the key intermediate CO, i.e. the doping of Ni and Pt lead to the formation of chemisorbed CO₂ and make CO formation easier. The moderately adsorbed CO on Cu-based electrocatalyst surfaces can avoid CO desorption and poisoning on surface active sites, thereby being in favor of further CO reduction, i.e. the Ni-doped Cu(111) surface has excessively strong CO adsorption, resulting in the highest barrier and the most negative limiting potential for further CO reduction among pure, Ni-, Pd- and Pt-doped Cu(111) surfaces. Another is the electronic structure of the transition metals-doped Cu-based electrocatalysts. The moderate interactions between Cu and transition metals and the moderate upshift of the d-band center of the doped transition metals in transition metal-doped Cu(111) will favor CO₂ reduction, which can lead to moderate adsorption of intermediates. For example, the Ni-doped Cu(111) surface with the strongest interaction between Cu and Ni and the most upshift of the d-band center of the Ni atom, and Pd-doped Cu(111) surface with the downshift of the d-band center of Pd atom have relatively poorer electrocatalytic activity for CO₂ reduction than the Pt-doped Cu(111) surface. Thus, an ideal Cu-based alloy electrocatalyst toward CO₂ reduction should be able reduce activation barriers for CO formation and further reduction, and should have moderate electron interactions between Cu and the doped transition metals, and a moderate upshift of the d-band center of the doped transition metals. In these ways, CO₂ reduction pathways can be facilitated and the yield of hydrocarbons CH₄ and CH₃OH can be enhanced.

4. Conclusions

In our present paper, the effect of the doped transition metal M (M = Ni, Pd and Pt) on CO₂ reduction pathways and the origin of the electrocatalytic activity are investigated systematically by focusing on the CH₄ and CH₃OH formation pathways based on DFT calculations of the key adsorbates associated with the computational hydrogen electrode model. Our studies show that the doping of Ni, Pd and Pt can promote CO₂ reduction and influence the selectivity of the reduction pathways and products, in which the doping of Pt may be able to lead to the strongest catalytic activity. The adsorption behavior of intermediates may be able to determine the CO₂ reduction activity. The destabilization of the stronger adsorbed intermediates to CH₃O by the doped Pd and Pt atoms make CH₃O formation and further reduction occur, thus the selectivity of reduction pathways is revealed. Owing to high carbon and oxygen affinity of Ni, all intermediates can be stabilized by the doped Ni atom. The Ni-doped Cu(111) surface has the strongest adsorption ability among the Ni-, Pd- and Pt-doped Cu(111) surfaces, whereas the doping of Pd and Pt leads to the weakest and moderate adsorption of most intermediates, respectively, which may explain why Pt-doped Cu(111) surfaces have the best catalytic activity for CO₂ reduction. Thus, it is concluded that moderate adsorbed intermediates on the Cu-based electrocatalyst surfaces will favor CO₂ reduction.

The enhanced electrocatalytic activity of the transition metal-doped Cu(111) surfaces may be owing to the decrease of the overpotential and electronic interactions between Cu and the transition metals. Further reduction of CO into CHO is the potential-limiting step on the pure and Ni-, Pd- and Pt-doped Cu(111) surfaces. The doping of Pd and Pt atoms can reduce the overpotential of CHO formation, whereas the doped Ni makes it slightly higher in comparison with pure Cu(111). The overpotential of CO formation is notably decreased after the doping of Ni, which is the rate-determining step on pure Cu(111). The significantly reduced overpotential for CO formation is also observed on the Pt-doped Cu(111). However, a slightly increased overpotential is obtained on the Pd-doped Cu(111) for CO formation. Simultaneously, the Pt-doped Cu(111) surface has the lowest overpotential for OH removal, suggesting that the doped metals with less affinity toward OH could enhance the catalytic activity. Thus, the doped Pt can simultaneously reduce the overpotential for CO formation and further reduction and most easily remove OH. The electronic structures analyses show that there is a moderate interaction between Cu and Pt and a moderate upshift of the d-band center of Pt. These analyses confirm why the Pt-doped Cu(111) surface has the best electrocatalytic activity for CO₂ reduction.

Two possible descriptors can be proposed in order to scale the electrocatalytic activity of Cu-based electrocatalysts for CO₂ reduction. One is activation barriers of CO formation and further reduction. The transition metal-doped Cu-based electrocatalysts with lower barriers will exhibit better catalytic activity for CO₂ reduction. The chemisorbed CO₂ on the Cu-based electrocatalyst will favor the formation of the key intermediate CO. The moderate adsorbed CO on the Cu-based electrocatalyst can avoid CO desorption and poisoning on surface active sites, thereby being in favor of further reduction of CO. Another is the electronic structure of the Cu-based electrocatalysts. The moderate interactions between Cu and the doped transition metals and the moderate upshift of the d-band center of the doped transition metals will be in favor of CO₂ reduction. Thus, an ideal Cu-based electrocatalyst should be able to reduce barriers for CO formation and further reduction, and should have moderate interactions between Cu and the doped transition metals, and a moderate upshift of d-band center of the doped transition metals. In these ways, CO₂ reduction pathways can be facilitated and the yield of hydrocarbons CH₄ and CH₃OH can be enhanced.

Acknowledgements

This work is financially supported by the National Natural Science Foundation of China (Grant No. 21303048), the Outstanding Youth Foundation of the Education Department of Hunan Province (Grant No. 16B178), the Hunan Provincial Natural Science Foundation of China (Grant No. 13JJ4101), the Construct Program of the Key Discipline in Hunan Province (Applied Chemistry), the Doctoral Start-up Fund of Hunan University of Arts and Science, the Hunan Provincial College Students Inquiry Learning and Innovative Pilot Projects (Grant No. 201510549007) and the Innovation Project of Hunan University of Arts and Science (Grant No. YB1509 and YB1612).

References

1. A. Goeppert, M. Czaun, R. B. May, G. K. S. Prakash, G. A. Olah and S. R. Narayanan, Carbon Dioxide Capture from the Air Using a Polyamine Based Regenerable Solid Adsorbent, J. Am. Chem. Soc., 2011, 133, 20164–20167.
2. C. Graves, S. D. Ebbesen, M. Mogensen and K. S. Lackner, Sustainable Hydrocarbon Fuels by Recycling CO₂ and H₂O with Renewable or Nuclear Energy, Renewable Sustainable Energy Rev., 2011, 15, 1–23.
3. D. T. Whipple and P. J. A. Kenis, Prospects of CO₂ Utilization via Direct Heterogeneous Electrochemical Reduction, J. Phys. Chem. Lett., 2011, 1, 3451–3458.
4. W. L. Zhu, R. Michalsky, O. Metin, H. F. Lv, S. J. Guo, C. J. Wright, X. L. Sun, A. A. Peterson and S. H. Sun, Monodisperse Au Nanoparticles for Selective Electrocatalytic Reduction of CO₂ to CO, J. Am. Chem. Soc., 2013, 135, 16833–16836.
5. K. P. Kuhl, E. R. Cave, D. N. Abram and T. F. Jaramillo, New Insights into the Electrochemical Reduction of Carbon Dioxide on Metallic Copper Surfaces, Energy Environ. Sci., 2012, 5, 7050–7059.
6. K. J. P. Schouten, E. Pérez Gallent and M. T. M. Koper, Structure Sensitivity of the Electrochemical Reduction of Carbon Monoxide on Copper Single Crystals, ACS Catal., 2013, 3, 1292–1295.
7. M. Le, M. Ren, Z. Zhang, P. T. Sprunger, R. L. Kurtz and J. C. Flake, Electrochemical Reduction of CO₂ to CH₃OH at Copper Oxide Surfaces, J. Electrochem. Soc., 2011, 158, E45–E49.
8. X. Nie, G. L. Griffin, M. J. Janik and A. Asthagiri, Surface Phases of Cu₂O(111) under CO₂ Electrochemical Reduction Conditions, Catal. Commun., 2014, 52, 88–91.
9. C. W. Li and M. W. Kanan, CO₂ Reduction at Low Overpotential on Cu Electrodes Resulting from the Reduction of Thick Cu₂O Films, J. Am. Chem. Soc., 2012, 134, 7231–7234.
10. B. Innocent, D. Liaigre, D. Pasquier, F. Ropital, J. M. Legerand and K. B. Kokoh, Electro-Reduction of Carbon Dioxide to Formate on Lead Electrode in Aqueous Medium, J. Appl. Electrochem., 2009, 39, 227–232.
11. Y. Chen and M. W. Kanan, Tin Oxide Dependence of the CO₂ Reduction Efficiency on Tin Electrodes and Enhanced Activity for Tin/Tin Oxide Thin-Film Catalysts, J. Am. Chem. Soc., 2012, 134, 1986–1989.
12. M. Jitaru, D. A. Lowy, M. Toma, B. C. Toma and L. Oniciu, Electrochemical Reduction of Carbon Dioxide on Flat Metallic Cathodes, J. Appl. Electrochem., 1997, 27, 875–889.
13. H. Noda, S. Ikeda, Y. Oda, K. Imai, M. Maeda and K. Ito, Electrochemical Reduction of Carbon Dioxide at Various Metal Electrodes in Aqueous Potassium Hydrogen Carbonate Solution, Bull. Chem. Soc. Jpn., 1990, 63, 2459–2462.
14. A. A. Peterson and J. K. Nørskov, Activity Descriptors for CO₂ Electroreduction to Methane on Transition-Metal Catalysts, J. Phys. Chem. Lett., 2012, 3, 251–258.
15. K. P. Kuhl, T. Hatsukade, E. R. Cave, D. N. Abram, J. Kibsgaard and T. F. Jaramillo, Electrocatalytic Conversion of Carbon Dioxide to Methane and Methanol on Transition Metal Surfaces, J. Am. Chem. Soc., 2014, 136, 14107–14113.
16. B. A. Rosen, A. Salehi-Khojin, M. R. Thorson, W. Zhu, D. T. Whipple, P. J. A. Kenis and R. I. Masel, Ionic Liquid-Mediated Selective Conversion of CO₂ to CO at Low Overpotentials, Science, 2011, 334, 643–644.
17. J. F. Hull, Y. Himeda, W. H. Wang, B. Hashiguchi, R. Periana, D. J. Szalda, J. T. Muckerman and E. Fujita, Reversible Hydrogen Storage Using CO₂ and a Proton-Switchable Iridium Catalyst in Aqueous Media under Mild Temperatures and Pressures, Nat. Chem., 2012, 4, 383–388.
18. R. Angamuthu, P. Byers, M. Lutz, A. L. Spek and E. Bouwman, Electrocatalytic CO₂ Conversion to Oxalate by a Copper Complex, Science, 2010, 327, 313–315.
19. V. Tripkovic, M. Vanin, M. Karamad, M. E. Björketun, K. W. Jacobsen, K. S. Thygesen and J. Rossmeisl, Electrochemical CO₂ and CO Reduction on Metal-Functionalized Porphyrin-Like Graphene, J. Phys. Chem. C, 2013, 117, 9187–9195.
20. Y. Hori, Electrochemical CO₂ Reduction on Metal Electrodes, in Modern Aspects of Electrochemistry, ed. G. Constantinos, C. Vayenas, R. E. White and M. E. Gamboa-Aldeco, Springer, New York, 2008, vol. 42, pp. 89–189.
21. N. Yang, F. Gao and C. E. Nebel, Diamond Decorated with Copper Nanoparticles for Electrochemical; Reduction of Carbon Dioxide, Anal. Chem., 2013, 85, 5764–5769.
22. W. Tang, A. A. Peterson, A. S. Varela, Z. P. Jovanov, L. Bech, W. J. Durand, S. Dahl, J. K. Nørskov and I. Chorkendorff, The Importance of Surface Morphology in Controlling the Selectivity of Polycrystalline Copper for CO₂ Electroreduction, Phys. Chem. Chem. Phys., 2012, 14, 76–81.
23. Y. Hori, K. Kikuchi, A. Murata and S. Suzuki, Production of Methane and Ethylene in Electrochemical Reduction of Carbon Dioxide at Copper Electrode in Aqueous Hydrogen Carbonate Solution, Chem. Lett., 1986, 15, 897–898.
24. Y. Hori, A. Murata, R. Takahashi and S. Suzuki, Electroreduction of CO to CH₄ and C₂H₄ at a Copper Electrode in Aqueous Solutions at Ambient Temperature and Pressure, J. Am. Chem. Soc., 1987, 109, 5022–5023.
25. R. L. Cook, R. C. MacDuff and A. F. Sammells, Evidence for Formaldehyde, Formic Acid, and Acetaldehyde as Possible Intermediates during Electrochemical Carbon Dioxide Reduction at Copper, J. Electrochem. Soc., 1989, 136, 1982–1984.
26. Y. Hori, A. Murata and R. Takahashi, Formation of Hydrocarbons in the Electrochemical Reduction of Carbon Dioxide at a Copper Electrode in Aqueous Solution, J. Chem. Soc., Faraday Trans. 1, 1989, 85, 2309–2326.
27. A. A. Peterson and J. K. Nørskov, Activity Descriptors for CO₂ Electroreduction to Methane on Transition-Metal Catalysts, J. Phys. Chem. Lett., 2012, 3, 251–258.
28. K. P. Kuhl, E. Cave, D. N. Abram and T. F. Jaramillo, New Insights into the Electrochemical Reduction of Carbon Dioxide on Metallic Copper Surfaces, Energy Environ. Sci., 2012, 5, 7050–7059.
29. J. G. Chen, C. A. Menning and M. B. Zellner, Monolayer Bimetallic Surfaces: Experimental and Theoretical Studies of Trends in Electronic and Chemical Properties, Surf. Sci. Rep., 2008, 63, 201–254.
30. W. Yu, M. D. Porosoff and J. G. Chen, Review of Pt-Based Bimetallic Catalysis: From model Surfaces to Supported Catalysts, Chem. Rev., 2012, 112, 5780–5817.
31. D. A. Hansgen, D. G. Vlachos and J. G. Chen, Using First Principles to Predict Bimetallic Catalysts for the Ammonia Decomposition Reaction, Nat. Chem., 2010, 2, 484–489.
32. J. K. Nørskov, T. Bligaard, J. Rossmeisl and C. H. Christensen, Towards the Computational Design of Solid Catalysts, Nat. Chem., 2009, 1, 37–46.
33. N. D. Subramanian, G. Balaji, C. S. S. R. Kumar and J. J. Spivey, Development of Cobalt–Copper Nanoparticles as Catalysts for Higher Alcohol Synthesis from Syngas, Catal. Today, 2009, 147, 100–106.
34. N. Schumacher, K. Andersson, L. C. Grabow, M. Mavrikakis, J. Nerlov and I. Chorkendorff, Interaction of Carbon Dioxide with Cu Overlayers on Pt(111), Surf. Sci., 2008, 602, 702–711.
35. M. Watanabe, M. Shibata, A. Kato, M. Azuma and T. Sakata, Design of Alloy Electrocatalysts for CO₂ Reduction III. The Selective and Reversible Reduction of CO₂ on Cu Alloy Electrodes, J. Electrochem. Soc., 1991, 138, 3382–3389.
36. X. Y. Zhao, B. B. Luo, R. Long, C. M. Wang and Y. J. Xiong, Composition-Dependent Activity of Cu-Pt Alloy Nanocubes for Electrocatalytic CO₂ Reduction, J. Mater. Chem. A, 2015, 3, 4134–4138.
37. J. Christophe, T. Doneux and C. Buess-Herman, Electroreduction of Carbon Dioxide on Copper-Based Electrodes: Activity of Copper Single Crystals and Copper–Gold Alloys, Electrocatalysis, 2012, 3, 139–146.
38. Z. Xu, E. Lai, Y. Shao-Horn and K. Hamad-Schifferli, Compositional Dependence of the Stability of AuCu Alloy Nanoparticles, Chem. Commun., 2012, 48, 5626–5628.
39. J. Nerlov and I. Chorkendorff, Methanol Synthesis from CO₂, CO, and H₂ over Cu(100) and Ni/Cu(100), J. Catal., 1999, 181, 271–279.
40. J. Nerlov, S. Sckerl, J. Wambach and I. Chorkendorff, Methanol Synthesis from CO₂, CO and H₂ over Cu(100) and Cu(100) Modified by Ni and Co, Appl. Catal., A, 2000, 191, 97–109.
41. E. Vesselli, L. D. Rogatis, X. Ding, A. Baraldi, L. Savio, L. Vattuone, M. Rocca, P. Fornasiero, M. Peressi, A. Baldereschi, R. Rosei and G. Comelli, Carbon Dioxide Hydrogenation on Ni(110), J. Am. Chem. Soc., 2008, 130, 11417–11422.
42. X. Liu, L. S. Zhu, H. Wang, G. Y. He and Z. Y. Biao, Catalysis Performance Comparison for Electrochemical Reduction of CO₂ on Pd–Cu/Graphene Catalyst, RSC Adv., 2016, 6, 38380–38387.
43. P. J. Berlowitz and D. W. Goodman, Kinetics of Methanol and Methane Synthesis over Pd/SiO₂ and Pd/La₂O₃, J. Catal., 1987, 108, 364–368.
44. M. L. Poutsma, L. F. Elek, P. A. Ibarbia, A. P. Risch and J. A. Rabo, Selective Formation of Methanol from Synthesis Gas over Palladium Catalysts, J. Catal., 1978, 52, 157–168.
45. Y. A. Ryndin, R. F. Hicks, A. T. Bell and Y. I. Yermakov, Effects of Metal-Support Interactions on the Synthesis of Methanol over Palladium, J. Catal., 1981, 70, 287–297.
46. J. P. Reilly, D. O'Connell and C. J. Barnes, Modification of Formate Stability by Alloying: the Cu(100)-c(2 × 2)-Pt System, J. Phys.: Condens. Matter, 1999, 11, 8417–8430.
47. Y. X. Yang, M. G. White and P. Liu, Theoretical Study of Methanol Synthesis from CO₂ Hydrogenation on Metal-Doped Cu(111) Surfaces, J. Phys. Chem. C, 2012, 116, 248–256.
48. P. Hirunsit, Electroreduction of Carbon Dioxide to Methane on Copper, Copper–Silver, and Copper–Gold Catalysts: A DFT Study, J. Phys. Chem. C, 2013, 117, 8262–8268.
49. P. Hirunsit, W. Soodsawang and J. Limtrakul, CO₂ Electrochemical Reduction to Methane and Methanol on Copper-Based Alloys: Theoretical Insight, J. Phys. Chem. C, 2015, 119, 8238–8249.
50. M. Qiu, Z. X. Fang, Y. Li, J. Zhu, X. Huang, K. N. Ding and W. K. Chen, First-Principles Investigation of the Activation of CO₂ Molecule on TM/Cu (TM = Fe, Co and Ni) Surface Alloys, Appl. Surf. Sci., 2015, 353, 902–912.
51. M. Qiu, Y. Liu, J. Wu, Y. Li, X. Huang, W. K. Chen and Y. F. Zhang, Theoretical Investigations of the Activation of CO₂ on the Transition Metal-Doped Cu(100) and Cu(111) Surfaces, Chin. J. Struct. Chem., 2016, 35, 669–678.
52. S. Baroni, A. Dal Corso, S. de Gironcoli and P. Giannozzi, PWSCF and PHONON: Plane-Wave Pseudo-Potential Codes, 2001, http://www.pwscf.org.
53. A. Kokalj, XCrySDen—A New Program for Displaying Crystalline Structures and Electron Densities, J. Mol. Graphics Modell., 1999, 17, 176–179.
54. 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.
55. A. Kokalj and M. Causà, XCrySDen: (X-Window) CRYstalline Structures and DENsities, 2001, http://www-k3.ijs.si/kokalj/xc/XCrySDen.html.
56. J. P. Perdew, K. Burke and M. Ernzerhof, Generalized Gradient Approximation Made Simple, Phys. Rev. Lett., 1996, 77, 3865–3868.
57. D. Vanderbilt, Soft Self-Consistent Pseudopotentials in a Generalized Eigenvalue Formalism, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 7892–7895.
58. J. Greeley, A. A. Gokhale, J. Kreuser, J. A. Dumesic, H. Topsøe, N. Y. Topsøe and M. Mavrikakis, CO Vibrational Frequencies on Methanol Synthesis Catalysts: A DFT study, J. Catal., 2003, 213, 63–72.
59. W. M. Haynes, CRC Handbook of Chemistry and Physics, CRC Press/Taylor and Francis, Boca Raton, FL, 93rd edn, Internet Version 2012, 2012.
60. M. Methfessel and A. T. Paxton, High-Precision Sampling for Brillouin-zone Integration in Metals, Phys. Rev. B: Condens. Matter Mater. Phys., 1989, 40, 3616–3621.
61. B. Hammer and J. K. Nørskov, Theoretical Surface Science and Catalysis-Calculations and Concepts, Adv. Catal., 2000, 45, 71–129.
62. H. A. Hansen, C. Shi, A. C. Lausche, A. A. Peterson and J. K. Nørskov, Bifunctional Alloys for the Electroreduction of CO₂ and CO, Phys. Chem. Chem. Phys., 2016, 18, 9194–9201.
63. L. H. Ou, Chemical and Electrochemical Hydrogenation of CO₂ to Hydrocarbons on Cu Single Crystal Surfaces: Insights into the Mechanism and Selectivity from DFT Calculations, RSC Adv., 2015, 5, 57361–57371.
64. G. Henkelman and H. Jonsson, Improved Tangent Estimate in the Nudged Elastic Band Method for Finding Minimum Energy Paths and Saddle Points, J. Chem. Phys., 2000, 113, 9978–9985.
65. G. Henkelman, B. P. Uberuaga and H. Jonsson, A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths, J. Chem. Phys., 2000, 113, 9901–9904.
66. Y. F. Zhao, Y. Yang, C. Mims, C. H. F. Peden, J. Li and D. Mei, Insight into Methanol Synthesis from CO₂ Hydrogenation on Cu(111): Complex Reaction Network and the Effects of H₂O, J. Catal., 2011, 281, 199–211.
67. Y. X. Yang, J. Evans, J. A. Rodriguez, M. G. White and P. Liu, Fundamental Studies of Methanol Synthesis from CO₂ Hydrogenation on Cu(111), Cu Clusters, and Cu/ZnO(0001), Phys. Chem. Chem. Phys., 2010, 12, 9909–9917.
68. J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard and H. Jónsson, Origin of Overpotential for Oxygen Reduction at a Fuel-Cell Cathode, J. Phys. Chem. B, 2004, 108, 17886–17892.
69. B. Hammer and J. K. Nørskov, Electronic Factors Determining the Reactivity of Metal Surfaces, Surf. Sci., 1995, 343, 211–220.
70. B. Hammer and J. K. Nørskov, Why Gold is the Noblest of All the Metals, Nature, 1995, 376, 238–240.
71. J. Greeley, J. K. Nørskov and M. Mavrikakis, Electronic Structure and Catalysis on Metal Surfaces, Annu. Rev. Phys. Chem., 2002, 53, 319–348.
72. B. Hammer, Y. Morikawa and J. K. Nørskov, CO Chemisorption at Metal Surfaces and Overlayers, Phys. Rev. Lett., 1996, 76, 2141–2144.
73. M. Gajdoš, A. Eichler and J. Hafner, CO Adsorption on Close-Packed Transition and Noble Metal Surfaces: Trends from ab initio Calculations, J. Phys.: Condens. Matter, 2004, 16, 1141–1164.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra28815d

---