Insight into the Confined Space Between Copper Nanoparticles for the Electrochemical CO₂ Reduction to CO

Abstract

The electrochemical carbon dioxide reduction reaction (CO₂RR) offers a promising route to mitigate excessive CO₂ emissions while enabling the production of value-added chemicals. However, achieving high catalytic selectivity and activity toward specific products remains a critical challenge. Here, we engineer a confined interfacial environment formed between adjacent copper nanoparticles and systematically investigate its impact on CO₂RR performance toward CO production. Our theoretical calculations reveal that the confined space effectively stabilizes the *COOH intermediate, a key species governing the CO₂-to-CO conversion pathway. In contrast, this geometric confinement exerts a negligible influence on the adsorption energetics of *H, which is associated with the competing hydrogen evolution reaction (HER). As a consequence, the catalyst exhibits a markedly reduced onset potential for CO₂RR, accompanied by enhanced selectivity and catalytic activity toward CO formation. These findings highlight the critical role of nanoscale confinement in modulating reaction energetics and provide a viable strategy for the rational design of highly efficient and selective catalysts for CO₂RR.

1. Introduction

The electrochemical carbon dioxide reduction reaction (CO₂RR) has attracted considerable attention as a viable strategy to mitigate the rising concentration of atmospheric CO₂ [1,2]. By converting renewable electricity into chemical energy, CO₂RR enables the production of value-added products, including CO, methane, ethanol, and other hydrocarbons or oxygenates [3]. Among these products, CO is of particular interest because it serves as a key feedstock for the Fischer–Tropsch process, through which it can be further converted into fuels and industrial chemicals. Consequently, the selective reduction of CO₂ to CO has been extensively investigated [4].

Copper is widely recognized as the only metal catalyst capable of producing a broad spectrum of hydrocarbons during the CO₂RR, which has stimulated extensive research interest [5,6,7,8]. However, bulk copper catalysts typically suffer from limited selectivity and insufficient activity toward specific target products, largely due to the competitive hydrogen evolution reaction (HER) [8,9,10]. Therefore, the development of advanced copper-based catalysts with enhanced selectivity and catalytic efficiency for the conversion of CO₂ into desired products remains highly desirable [10,11].

A variety of strategies have been explored to enhance the catalytic performance of copper-based systems, including modulation of surface morphology, nanostructure, facet engineering, and heteroatom doping [12,13,14,15,16,17,18,19]. In addition, external reaction conditions, such as electrolyte pH, cation identity, and applied pressure, have been shown to play important roles in governing CO₂RR performance [20,21,22,23,24,25,26]. More recently, constructing confined catalytic environments has emerged as an effective approach to improve electrochemical activity [27,28,29,30]. For example, metallic Sn quantum sheets confined within few-layer graphene exhibit markedly enhanced CO₂RR activity compared with Sn nanoparticles or bulk counterparts [31]. Beyond CO₂RR, confinement effects have also been demonstrated in other electrochemical processes, including hydrogen evolution at Ni surfaces encapsulated by graphene overlayers, oxygen evolution within three-dimensional RuO₂ nanochannels, and oxygen reduction at interparticle-confined interfaces between adjacent Pt nanoparticles [28,29,32].

Recent experimental advances in nanoparticle self-assembly, porous Cu architectures, MOF-derived catalysts, and interparticle interface engineering have demonstrated that confined Cu nanogaps on the sub-nanometer to nanometer scale can be experimentally achieved [33,34,35,36]. These confined environments may generate localized electronic coupling and intermediate stabilization effects that differ substantially from those of isolated Cu surfaces. Motivated by these findings, we investigate the catalytic behavior of confined interfacial regions formed between adjacent facets of copper nanoparticles for CO₂RR toward CO production. Unlike confinement environments created by porous hosts, graphene encapsulation, or nanochannels, the present work focuses specifically on the nanoscale interparticle confined interface formed between adjacent Cu nanoparticle surfaces. Such confined metallic interfaces may induce direct electronic coupling and localized adsorption-energy modulation within the gap region, thereby generating confinement effects fundamentally different from conventional pore-confined or encapsulated catalytic systems.

The confined environment is modeled as a nanoscale gap (d_Cu) between two copper surfaces separated by a defined distance (Scheme 1). Different separated distances and exposed facets of copper nanoparticles were taken into consideration. Focusing on the CO₂-to-CO pathway, the adsorption characteristics of key intermediates, *COOH and *CO, within the confined region are examined in detail. For comparison, the adsorption free energy of *H associated with the competing hydrogen evolution reaction (HER) is also evaluated. The results reveal that spatial confinement exerts a more pronounced effect on the adsorption free energy of *COOH than on *CO or *H, which can be attributed to the larger molecular size and stronger geometric sensitivity of the *COOH intermediate. Consequently, the onset potential for CO₂RR to CO is significantly modulated by the presence of the confined environment. These insights provide a mechanistic basis for understanding confinement effects in CO₂RR and offer guidance for the rational design of copper-based catalysts with enhanced selectivity and activity.

2. Results and Discussion

It should be noted that the present confined Cu slab models are intentionally simplified representations of realistic interparticle nanospaces. Under practical electrochemical conditions, confined regions between Cu nanoparticles may involve structural heterogeneity, surface reconstruction, defects, oxidation states, grain boundaries, and strain effects. However, the purpose of the present work is to isolate the intrinsic influence of interfacial confinement distance on the adsorption energetics of key reaction intermediates. Therefore, an idealized and well-defined slab-gap framework was adopted to minimize additional coupled structural variables and to establish clearer structure–activity relationships associated with nanoscale confinement.

All possible adsorption configurations of *COOH, *CO, and *H within the confined systems, i.e., c-sCu(100), c-sCu(110), and c-sCu(111), were systematically evaluated, and only the most stable structures are discussed here (Figure 1 and Table S1). As the interfacial distance (d_Cu) decreases from 11.0 to 5.0 Å, the adsorption geometries and preferred binding sites of *CO and *H remain essentially unchanged across all surfaces. In contrast, *COOH exhibits a distinct structural response to spatial confinement. Although the adsorption site of *COOH is preserved, notable reorientation occurs when d_Cu falls below 6.0 Å. Specifically, the O and OH moieties rotate around the C atom, as illustrated in Figure 1 and Figure S1. This structural flexibility can be attributed to the larger molecular size and steric sensitivity of *COOH compared to *CO and *H, making it more susceptible to geometric constraints imposed by the confined environment.

To further quantify the effect of spatial confinement, the adsorption free energies of *COOH, *CO, and *H were systematically evaluated as a function of interfacial distance (d_Cu) for c-sCu(100), c-sCu(110), and c-sCu(111) (Figure 2). The results clearly indicate that ΔG_ad-*COOH is significantly more sensitive to variations in d_Cu than ΔG_ad-*CO and ΔG_ad-*H. As d_Cu decreases from 20.0 to 6.0 Å, ΔG_ad-*COOH decreases by 0.33, 0.18, and 0.34 eV for c-sCu(100), c-sCu(110), and c-sCu(111), respectively. In contrast, the corresponding changes in ΔG_ad-*CO are limited to 0.11, 0.02, and 0.06 eV, while ΔG_ad-*H shows negligible variation (−0.02, −0.02, and 0.00 eV, respectively). These results are consistent with the structural evolution of *COOH discussed above, confirming that confinement effects predominantly influence this intermediate. When d_Cu is increased to 11.0 Å, the adsorption free energies of all intermediates become comparable to those on the corresponding open Cu(100), Cu(110), and Cu(111) surfaces, indicating that the confinement effect is effectively diminished at larger interfacial separations.

To further evaluate the thermodynamic consequences of confinement-induced adsorption modulation, he reaction free-energy profiles for CO₂RR toward CO formation and HER toward H₂ evolution were constructed based on the calculated ΔG_ad-*COOH, ΔG_ad-*CO, and ΔG_ad-*H values (Figure 3). For CO₂RR, the formation of *COOH is identified as the potential-determining step (PDS) across all confined systems, including c-sCu(100), c-sCu(110), and c-sCu(111), as it corresponds to the largest uphill free-energy change along the reaction pathway. In contrast, the PDS for HER exhibits surface-dependent behavior. On c-sCu(100), the initial proton adsorption step (* + H⁺ + e⁻ → *H) is the rate-limiting process. However, for c-sCu(110) and c-sCu(111), the subsequent hydrogen evolution step (*H → * + 1/2 H₂ (g)) becomes the PDS, indicating distinct kinetic limitations on different surface configurations.

To render all elementary steps in the CO₂RR and HER pathways thermodynamically downhill, an external bias potential must be applied. The minimum potential required to achieve this condition is defined as the limiting (or onset) potential. Accordingly, the limiting potentials for CO₂RR and HER are estimated as U_CO2RR = −|ΔG_ad-*COOH|/|e| and U_HER = −|ΔG_ad-*H|/|e|, respectively. As shown in Figure 4, U_CO2RR exhibits a pronounced dependence on the interfacial distance (d_Cu), becoming more negative with increasing d_Cu across all confined systems. In contrast, U_HER remains nearly constant, indicating a negligible sensitivity to spatial confinement. For instance, on c-sCu(100), U_CO2RR shifts from −0.48 V at d_Cu = 11.0 Å to −0.18 V at d_Cu = 6.0 Å, whereas U_HER changes by only ~0.02 V over the same range. More importantly, the difference between the two limiting potentials (U_HER − U_CO2RR), which serves as a descriptor of reaction selectivity, decreases substantially with decreasing d_Cu. Specifically, this value is reduced from 0.45 to 0.13 V for c-sCu(100), from 0.20 to 0.00 V for c-sCu(110), and from 0.55 to 0.22 V for c-sCu(111) as d_Cu decreases from 11.0 to 6.0 Å. The significant narrowing of U_HER − U_CO2RR indicates that CO₂RR becomes thermodynamically more competitive relative to HER under stronger confinement conditions. This behavior indicates that decreasing the interfacial distance preferentially stabilizes key CO₂RR intermediates compared with *H adsorption, thereby favoring thermodynamically preferred CO formation over H₂ evolution. These results further highlight the important role of nanoscale confinement in regulating the thermodynamic selectivity tendency of Cu-based catalysts toward CO₂ reduction.

To further elucidate the role of spatial confinement in modulating CO₂RR and HER, the electronic properties of the confined systems were systematically analyzed. The electron density distributions of *COOH, *CO, and *H on c-sCu(100), c-sCu(110), and c-sCu(111) are presented in Figure 5. A clear trend is observed: *COOH exhibits interaction with the opposing (upper) surface at d_Cu ≤ 7 Å, whereas *CO shows similar interaction only at d_Cu ≤ 6 Å. In contrast, *H remains largely unaffected by the presence of the upper surface even at d_Cu = 6 Å. This behavior indicates that spatial confinement induces significantly stronger electronic perturbations for *COOH than for *CO and *H. Such differential sensitivity can be attributed to the enhanced electronic coupling between *COOH and the adjacent surface under confinement, which plays a dominant role in modulating its adsorption energetics. This interpretation is further supported by charge density difference analysis (Figure S10), which reveals pronounced electron redistribution associated with *COOH adsorption. Consequently, the progressively strengthened interfacial electronic interaction under decreasing d_Cu directly contributes to the substantial variation in ΔG_ad-*COOH and the corresponding shift in the onset potential for CO₂RR.

To further quantify the confinement-induced electronic interaction between the Cu surfaces and adsorbed intermediates, Bader charge analysis was performed for representative weak-confinement (d_Cu = 20 Å) and strong-confinement (d_Cu = 6 Å) configurations. As summarized in Table S2, the electron transfer from the Cu surfaces to *COOH increases from 0.44 |e| to 0.46 |e| as dCu decreases from 20 Å to 6 Å, while *CO exhibits a similar increase from 0.48 |e| to 0.51 |e|. In contrast, the electron gains of *H slightly decrease from 0.25 |e| to 0.23 |e| under stronger confinement conditions. These results quantitatively confirm that nanoscale confinement promotes electron accumulation on key CO₂RR intermediates while slightly weakening the electronic interaction associated with HER intermediates, consistent with the observed thermodynamic trends in adsorption free energies.

The above results demonstrate that spatial confinement can effectively modulate the interaction between key reaction intermediates and copper surfaces, thereby tuning the onset potential of CO₂RR. Such confinement-induced effects highlight the potential of engineering nanoscale interfacial environments to enhance catalytic performance. In particular, electrocatalysts featuring well-defined confined geometries are expected to exhibit improved selectivity and activity toward desired products. Experimentally, such configurations may be realized through controlled nanoparticle dispersion or by encapsulating metal nanoparticles within porous hosts possessing tailored pore dimensions, such as zeolites or metal–organic frameworks [30,37,38].

3. Materials and Methods

All calculations were performed using spin-polarized DFT (ISPIN = 2) with the Perdew–Burke–Ernzerhof (PBE) functional implemented in the Vienna ab initio simulation package (VASP) [39,40]. The ion-electron interaction was described by the projector augment wave (PAW) method, and the van der Waals interaction was also considered with the Grimme D3 method [41,42]. The computational model consists of two copper slabs separated by a tunable interfacial distance of d_Cu (Scheme 1). Three low-index copper surfaces, (100), (110), and (111), were selected to construct the confined configurations [15,41,43]. Specifically, (4 × 4) supercells with three atomic layers were used for Cu(100) and Cu(111), while a (3 × 4) supercell with five layers was employed for Cu(110), ensuring adequate representation of surface structures. To systematically evaluate the effect of spatial confinement in the adsorption of intermediates during the CO₂RR and HER, d_Cu was varied from 5.5 to 11.0 Å in increments of 0.5 Å. The open Cu surface is simulated by setting d_Cu to 20.0 Å, where interfacial interactions become negligible. All calculations were performed using a plane-wave basis set with a kinetic energy cutoff of 450 eV. Structural optimizations were conducted until the residual forces on each atom were below 0.02 eV Å⁻¹. Brillouin zone sampling was carried out using Monkhorst–Pack grids of 11 × 11 × 11 for bulk calculations and 3 × 3 × 1 for slab models. The confined systems constructed from Cu(100), Cu(110), and Cu(111) surfaces are denoted as c-sCu(100), c-sCu(110), and c-sCu(111), respectively.

The reaction pathway for CO₂ reduction to CO is considered as follows:

* + CO₂(g) + H⁺ + e⁻ → *COOH

*COOH + H⁺ + e⁻ → *CO + H₂O(l)

*CO → * + CO(g)

where * and *X denote the surface and an adsorbed species on the surface.

For comparison, the hydrogen evolution reaction proceeds via:

* + H⁺ + e⁻ → *H,

*H → * + 1/2 H₂(g)

Accordingly, *COOH, *CO, and *H are identified as the key intermediates governing the CO₂RR and HER processes. Based on the definition of the standard hydrogen electrode (SHE), the free energy of the proton–electron pair is referenced to hydrogen gas, such that G(H⁺ + e⁻) = 1/2 G(H₂) at U = 0 V (vs. SHE) and pH = 0. At arbitrary electrode potentials and pH values, this relationship can be expressed as: [8]

G(H⁺ + e⁻) = 1/2 G(H₂(g)) − |e|U − 0.0592·pH

Within the framework of the computational hydrogen electrode (CHE) model [8,43], the adsorption free energies of key intermediates are defined as:

ΔG_ad-*COOH = G(*COOH) − G(*) − G(CO₂(g)) − 1/2 G(H₂(g)) + |e|U + 0.0592·pH

ΔG_ad-*CO = G(*CO) + G(H₂O(l)) − G(*COOH) − 1/2 G(H₂(g)) + |e|U + 0.0592·pH

ΔG_ad-*H = G(*H) − G(*) − 1/2 G(H₂(g)) + |e|U + 0.0592·pH

Here, G(*) denotes the energy of the clean surface, while G(CO₂), G(H₂), G(H₂O), G(*COOH), G(*CO), and G(*H) correspond to the computed energies of gas-phase molecules and adsorbed intermediates, respectively. These adsorption-free energies are dependent on both the electrode potential (U) and the electrolyte pH, thereby influencing the overall free-energy profiles of CO₂RR and HER. Although the CHE model employed here explicitly includes corrections associated with the applied potential and electrolyte pH, these contributions largely cancel when evaluating the relative free-energy changes and limiting potentials of elementary proton-coupled electron transfer steps. Therefore, to focus specifically on the intrinsic effect of nanoscale confinement, all calculations were consistently performed under the reference condition of U = 0 V and pH = 0.

It should also be noted that realistic electrochemical interfaces are considerably more complex than the simplified framework adopted in the present study. Factors such as explicit solvent effects, electric double-layer structures, local ion accumulation, interfacial water configurations, and field-induced polarization may further influence the stabilization of reaction intermediates within confined nanospaces. However, rigorous treatment of these effects typically requires substantially more sophisticated methodologies, including explicit solvation models, constant-potential simulations, or ab initio molecular dynamics. Accordingly, the present work is intended to provide a fundamental thermodynamic understanding of confinement-regulated adsorption energetics on Cu surfaces, while the incorporation of more realistic electrochemical environments will be the subject of future investigations.

4. Conclusions

To elucidate the influence of confined space on the catalytic behaviors of copper nanoparticles during the CO₂RR and HER, we have considered different sizes of confined space reflected by the separated distance d_Cu, different exposed copper surfaces and adsorption free energies of the key intermediates. The calculation results show that a smaller confined space makes the adsorbed *COOH more stable, but has little effect on the adsorption free energy of *H. A confined space would change the onset potential for the CO₂RR, but does not change that for the HER, which would make the CO₂RR more thermodynamically competitive in the electrochemical reaction. Overall, the present work reveals that interparticle nanoscale confinement can serve as an effective thermodynamic modulation mechanism for regulating adsorption energetics and the relative CO₂RR/HER competitiveness on Cu surfaces. The confinement-induced stabilization of key CO₂RR intermediates originates from localized electronic redistribution and interfacial coupling within the confined metallic gap region. These findings provide theoretical insights into how confined metallic interfaces may be utilized to manipulate reaction thermodynamics in electrocatalytic CO₂ reduction systems. More broadly, the present study highlights the importance of nanoscale interfacial engineering in designing confinement-regulated catalytic environments for future Cu-based CO₂RR electrocatalysts.