Catalyst design for the electrochemical reduction of carbon dioxide: from copper nanoparticles to copper single atoms

Cu nanoparticles

Cu NPs refer to the catalysts with size of Cu particles in the nanometer range. The size, morphology, chemical environment, and crystal plane of Cu NPs have a significant influence on the products and performance of CO₂RR. When the size of Cu particles decreases to the nanometer level^[45,46], the catalyst exhibits higher surface activity and richer surface defects than the Cu particles with larger sizes^[47-49]. The surface properties and structures of Cu NPs can be regulated by appropriate synthesis methods^[50]. Thus, the selectivity and activity of the catalytic reaction can be adjusted^[51]. Modifying the surface of Cu NPs with hydrophobic materials, such as Nafion^[52,53], polytetrafluoroethylene (PTFE)^[54], and alkyl thiol^[55], can inhibit HER and promote CO₂ mass transfer^[56,57], thereby improving the efficiency of CO₂RR and enhancing C₂₊ selectivity. Cu NPs are attractive catalysts for CO₂RR to produce valuable chemicals^[58]. Next, we will summarize some Cu NP catalysts with different morphologies, sizes, and chemical environments to regulate the catalytic effect of CO₂RR and the corresponding selective catalytic mechanism.

Reske et al. synthesized six spherical Cu NP catalysts of different average sizes (2-15 nm) and coordination numbers (CNs) by stirring CuCl₂ loaded micelles and changing the molecular weight of polyvinylpyrrolidone (PVP) heads or metal salt/PVP ratios [Figure 3A-F]^[59]. As shown in Figure 3G and H, the proportion of atoms with low CNs (Cu-CNs < 8) significantly increased in Cu NPs with particle sizes less than 6 nm. During CO₂RR, the catalytic activity increased but tended to promote HER, mainly producing H₂ and CO. Manthiram et al. prepared Cu NPs with a diameter of 7.0 ± 0.4 nm terminated with tetradecyl phosphonate by reducing cupric acetate Cu(Ac)₂ in trioctylamine^[60]. The Faraday efficiency (FE) of CH₄ is 76% at -1.35 V vs. RHE. The current density of CH₄ was four times that of Cu foil electrodes. They found that more isolated NPs exposed more catalytic active sites to form CH₄. However, when the Cu NPs aggregated, CH₄ would be lost in the products. Jung et al. prepared a 20 nm Cu₂O NP/C with cubic morphology. Cu NPs were grown on the carbon carrier using cysteine molecules^[61]. Under negative potential, the 20 nm cubic Cu₂O crystal particles disintegrated into 2-4 nm particles, and FEs of C₂H₄ (FE of C₂H₄) and C₂₊ (FE C₂₊) of the shattered Cu-based NP/C catalyst reached 57.3% and 74%, respectively. Yang et al. reduced Cu(Ac)₂ precursors with tetradecyl phosphonic acid at high temperatures to prepare Cu NPs, and controlled the size of Cu NPs by adjusting the ratio^[62]. The authors established that the selectivity of C₂₊ increased with the proportion of metallic Cu nanograins. The 7 nm Cu NP with a unity fraction of Cu nanograins exhibited six times higher FE C₂₊ than the 18 nm Cu NP ensemble with one-third of Cu nanograins [Figure 3I and J]. As shown in Figure 3K, Zhang et al. adsorbed a mixture of Cu phthalocyanine (CuPc) and carbon black (CB) onto a glassy carbon wafer and electroreduced CuPc to Cu NPs (20 nm) under CO₂RR conditions^[63]. The size of Cu NPs was influenced by the reaction time, reduction potential, oxidation degree of the carbon support, and the loading amount of CuPc. Furthermore, due to the abundance of grain boundaries in large NPs, the selectivity of C₂₊ products also increased from 0% to 40% during the slow growth of Cu NPs (2 nm to 20 nm). The highest FE C₂₊ was 70%, and the current density for C₂₊ was 800 mA cm^-2.

Figure 3. (A-F) Tapping-mode AFM images of micellar Cu NPs; Particle size dependence of (G) the composition of gaseous reaction products (balance is CO₂) during catalytic CO₂ electroreduction over Cu NPs; (H) the faradic selectivities of reaction products during the CO₂ electroreduction on Cu NPs. (Reproduced with permission^[59]. Copyright 2014, American Chemical Society); (I) Structure-activity correlation of relative fraction of active Cu nanograins and C₂₊ FE of Cu NP ensembles with three different NP sizes; (J) FE for all CO₂RR products of three different Cu NP sizes (7, 10 and 18 nm) at -0.8 V. (Reproduced with permission^[62]. Copyright 2023, The Author(s), under exclusive license to Springer Nature Limited); (K) Structure-activity correlation of relative fraction of active Cu nanograins and C₂₊ FE of Cu NP ensembles with three different NP sizes.(Reproduced with permission^[63]. Copyright 2023, American Chemical Society). NPs: Nanoparticles; CO₂RR: Carbon dioxide reduction reaction; FE: Faraday efficiency; AFM: Atomic force microscopy.

Cu NPs with various morphologies and crystal faces have different effects on CO₂RR. Zi et al. found that Cu nanoneedles can induce ultra-high local potassium concentration (4.22 M)^[64]. High concentrations of potassium can promote the C-C coupling, achieving efficient CO₂ reduction in 3-M KCl electrolytes of pH = 1. FE C₂₊ reached up to 90.69% ± 2.15% at 1,400 mA cm^-2. Fu et al. prepared a series of Cu₂O nanocrystals (NCs) with diverse crystal faces and morphologies^[65]. Among them, the o-Cu₂O NCs with high-index facets reached the highest FE C₂₊ (48.3%). They investigated the structural alterations of Cu₂O NCs after stability tests and discovered that abundant crystal defects and high-index facets were observed, which may be the active sites. As shown in Figure 4A, Luo et al. modified the surface of Cu₂O particles using reducing agents^[66], and found that the star-shaped Cu₂O NPs (F-Cu₂O) with exposed (322) facets achieved the highest C₂H₄ selectivity (FE = 74.1%). They found that the Cu₂O (332) surface can significantly reduce the free energy during the coupling process of *CHO intermediates and promote the production of C₂H₄. As shown in Figure 4B, they further modified the surface with ZIF-8. The ZIF-8 shell provided abundant pores, substance exchange channels and appropriate hydrophobic microenvironments, which facilitated the desorption of *CO intermediates. Periasamy et al. prepared stable Cu₂O/polypyrrole (Cu₂O/Ppy) particles on linen-textured (LT) paper^[67]. The Cu₂O/Ppy with octahedral and micro flower shapes had exposed (111) facets, (311) and (211) facets, exhibiting high selectivity towards CH₃OH products. At -0.85 V vs. RHE, the FE of methanol (FE of CH₃OH) was 93% ± 1.2%. Wu et al. prepared a highly stable mixed Cu₂O-Cu nanocube catalyst [Cu₂O (CO) ] by thermal reduction of CuO nanocubes under CO atmosphere^[68]. The material exhibited a high-density CuO/Cu interface and abundant Cu (100) crystal planes. At an industrial current density of 500 mA cm^-2, the FE of C₂₊ products was 77.4% (FE of C₂H₄ was 56.6%). Recently, as shown in Figure 4C-E, Geng et al. prepared hydrophobic porous Cu₂O spheres (pore sizes 140/240/340 nm) using polystyrene (PS) spheres as pore guides by wet chemical methods^[69]. The hydrophobic porous morphology of the catalyst facilitated the transportation and capture of CO₂ and promoted CO₂ activation and the formation of *OCCOH intermediates, thereby promoting C-C coupling. The P-Cu₂O-240 sample had the highest CO₂ mass transfer efficiency. As shown in Figure 4E, at -1.0 A cm^-2, FE C₂ was 75.3% ± 3.1%.

Figure 4. (A) Synthesis of Cu₂O NPs by the reductant-controlling method and Cu₂O@ZIF-8 composites; (B) HAADF-STEM image and elemental mappings. (Reproduced with permission^[66]. Copyright 2022, Wiley-VCH GmbH); (C) Scheme, SEM, and enlarged HAADF-STEM images of P-Cu₂O-140, P-Cu₂O-240, and P-Cu₂O-340; (D) Water contact angle (CA) of P-Cu₂O-140, P-Cu₂O-240, P-Cu₂O-340 (from top to bottom); (E) FEs for C₂₊ products under different applied current densities over porous Cu₂O samples. (Reproduced with permission^[69]. Copyright 2024, American Chemical Society). NPs: Nanoparticles; HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; FE: Faraday efficiency.

The grain boundaries of Cu NPs affect the selectivity of CO₂RR. Frese et al. first discovered an increase in CH₄ production on Cu (100), Cu (110), and Cu (111)^[70]. Hori et al. found that Cu (100) and Cu (111) tended to generate C₂H₄ and CH₄, respectively, while Cu (110) preferably produced acetate and acetaldehyde. They also reported that CO is a key intermediate in CO₂RR^[71]. Schouten et al. observed that the Cu (100) surface tended to form C₂H₄ at a relatively low overpotential, while the Cu (111) surface preferentially generated CH₄ and only a small amount of C₂H₄^[72]. Gao et al. found that the crystal facets exposed by Cu₂O NPs greatly affected the selectivity of CO₂-C₂H₄ [Figure 5A-F]^[73]. Cu₂O NPs with both (111) and (100) crystal faces demonstrated a stronger selectivity for CO₂-C₂H₄ with an FE of C₂H₄ of 59% compared to Cu₂O NPs with only one crystal face. Wu et al. further proved this point^[74]. They found that the Cu (100)/Cu (111) interface had a good localized electronic structure, which enhanced CO adsorption and C-C coupling, and its performance was better than that of the Cu (100) and Cu (111). Ma et al. reconstructed Cu NPs on vertical graphene [plasma-enhanced chemical vapor deposition (PECVD)]^[75]. They constructed incompatible sites and abundant oxygen vacancies on Cu active sites through reduction-oxidation-reduction (ROR), enhancing CO₂RR activity. This catalyst had abundant grain boundaries, producing *COOH-derived products (CO) at high oxidation potentials (+1.2 V vs. RHE), and *OCHO-derived products (HCOOH) at low oxidation potentials (+0.8 V vs. RHE). Recently, Chen et al. utilized tannic acid (TA) molecules in situ to regulate and reconstruct Cu-based materials and prepared a Cu electrocatalyst (TA-Cu)^[76]. Cu (111) NPs and partially oxidized CuO_x were uniformly distributed on TA-Cu nanorods [Figure 5G]. The hydroxyl groups in TA could stabilize the key intermediate *COH through hydrogen bonding, thereby promoting the generation of C₂H₄. At -0.7 V vs. RHE, the process reached an FE of C₂H₄ of 63.6%, a stability of ten h, and a current density exceeding 497.2 mA cm^-2[Figure 5H and I].

Figure 5. SEM images of (A, B) c-Cu₂O NPs; (C, D) o-Cu₂O NPs; and (E, F) t-Cu₂O NPs. (Reproduced with permission^[73]. Copyright 2020, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim); (G) The preparation of TA-Cu via electrochemical reconstruction of CuTA; (H) FEs of TA-Cu obtained at different applied potentials; (I) Stability for CO₂ electrolysis with the TA-Cu. (Reproduced with permission^[76]. Copyright 2023, Wiley-VCH GmbH). NPs: Nanoparticles; TA: Tannic acid ; FE: Faraday efficiency; SEM: Scanning electron microscope.

Carbon intermediates can be effectively limited by forming nano-sized confined spaces in Cu-based materials, thus effectively promoting the formation of C₂₊. For example, Liu et al. prepared a porous Cu nanosphere catalyst (P-Cu)^[77], which can enrich the intermediate *CO in the pore structure. As shown in Figure 6A-C, Zhang et al. found a volcano curve relationship between the selectivity of C₂H₅OH and the nanocavity size of porous CuO in the range of 0 to 20 nm^[78]. The increase in *OH coverage associated with the nanocavity size-dependent confinement effect was believed to account for the remarkable CH₃CH₂OH selectivity. Increased *OH coverage may facilitate the hydrogenation of *CHCOH to *CHCHOH (CH₃CH₂OH pathway). As shown in Figure 6D and E, p-CuO-(12.5 nm) has an FE of CH₃CH₂OH of 44.1% ± 1.0% at a high C₂H₅OH partial current density of 501.0 ± 15.0 mA cm^-2, while FE C₂₊ was 90.6% ± 3.4%. Liu et al. increased the coverage of local CO by introducing a Cu hollow multi-shell structure (HoMSs) and utilizing the nano-constrained effect^[79]. The presence of Cu in the cavity stabilized *CO and promoted dimerization. As the shell layer increased, the selectivity and activity of C₂₊ products significantly improved. The Cu HoMSs with three shells exhibited the largest FE C₂₊ in neutral electrolyte of 77.0% ± 0.3%, and the current density was 513.7 ± 0.7 mA cm^-2. As shown in Figure 6F-H, based on mature CO₂-CO electroreduction technology, Yang et al. reported porous Cu₂O catalysts with nanocavities^[80]. They found that fragmented and solid Cu₂O catalysts mainly produced CO, while the FE of C₂₊ production on porous Cu₂O catalysts was much higher, indicating that the porous structure could promote C-C bonding. At -0.61 Vvs. RHE, the FE C₂₊ was 75.2% ± 2.7%, and the partial current density was 267 ± 13 mA cm^-2. They found that pore cavities could restrict the in-situ generation of carbon intermediates effectively, which bound with Cu sites [Figure 6I]. In addition, this restricted intermediate promoted C-C coupling within the reactive nanocavity.

Figure 6. (A) TEM images of p-CuO-(12.5 nm); (B) HAADF-STEM images of p-CuO-(12.5 nm); (C) High-resolution HAADF-STEM images of p-CuO-(12.5 nm); (D) FE_ethanol and (E) FE C₂₊ over Cu NPs, p-CuO-(7.0 nm), p-CuO-(12.5 nm), and p-CuO-(19.4 nm) catalysts. (Reproduced with permission^[78]. Copyright 2023, PNAS); Multihollow Cu₂O catalyst imaged by SEM (F), TEM (G) and HRTEM (H); The inset in (G) shows the corresponding SAED pattern; (I) Schematic of carbon intermediates that are confined in the nanocavities. White: H; gray: C; red: O; violet: Cu. (Reproduced with permission^[80]. Copyright 2020, American Chemical Society). HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; NPs: Nanoparticles; FE: Faraday efficiency; TEM: Transmission electron microscope; HAADF-STEM: High-angle annular dark-field scanning transmission electron microscope; HRTEM: High-resolution transmission electron microscope; SAED: Selected-area electron diffraction.

Modifying the Cu NPs with different materials can enhance C-C coupling, thereby improving the production of multiple carbon products. For example, Zhao et al. reported an in-situ polymerization strategy of encapsulating hydrophobic polymers onto Cu NPs to synthesize fluorinated polymer-functionalized Cu NPs (Cu poly-1/2)^[58]. Increasing the hydrophobicity of the catalyst surface can suppress HER, which is beneficial for reducing the thickness of the CO₂ gas diffusion layer and promoting CO₂ mass transfer^[56,57]. Compared with bare Cu (FE C₂₊ = 61.59%), the use of Cu poly-1/2 increased the FE C₂₊ to 71.08% at -3.98 V vs. RHE. Pellessier et al. covered commercial Cu NPs with porous PTFE nanocoats^[54]. The physical adsorption of PTFE on Cu NPs generated a large interface surface area, which could improve the binding energy of CO intermediates on Cu and enhance C-C coupling. The FE C₂₊ was 78% within the current density range of 400-500 mA cm^-2. Using ammonia and Cu(NO₃)₂ as raw materials, Zhou et al. prepared AN-Cu(OH)₂ by a one-step wet chemical method. Then, the acidic ionomer Nafion was used to regulate the surface hydrophobicity and local alkalinity of the catalyst [AN-Cu(OH)@Nafion]^[52]. The FE of C₂H₄ of optimal Nafion-activated AN-Cu(OH)@Nafion was 44%, and HER was effectively inhibited. Recently, Su et al. used Nafion to modify Cu NP catalyst (Cu@Nafion)^[53]. Nafion can inhibit HER and enhance CO₂ mass transfer, which is beneficial for C-C coupling. At -1.2V vs. RHE, FE C₂₊ was about 73.5%, and HER decreased from 40.6% to 16.8%. As shown in Figure 7A and B, Chen et al. developed a Cu polyamine hybrid catalyst by co-electroplating polyamines onto the surface of Cu electrodes (Cu-Pi, i =1-5)^[81]. The amine functional group causes higher pH values and has higher intermediate stability, as well as the ability to adsorb more CO, thereby improving the selectivity of C₂H₄. The Cu-P1 catalyst reached the highest FE of C₂H₄ of 72%. As shown in Figure 7C and D, Lin et al. modified alkyl thiols with different alkyl chain lengths to regulate the interfacial wettability^[55]. The FE C₂₊ of the Cu-12C catalyst was the highest, reaching 86.1%. At 400 mA cm^-2, the catalyst achieved an FE C₂₊ of 80.3% [Figure 7E and F], which is one of the best performances at this current density. Increasing the interfacial hydrophobicity of the catalyst effectively prevented the absorption of H₂O, facilitated CO₂ transportation, increased *CO coverage, and thus enhanced C₂₊ selectivity^[82].

Figure 7. (A) Schematic illustration of P1 and Cu co-electroplating on the GDL; (B) FE for all products on Cu-P1. [Reproduced with permission^[81] .Copyright 2020, The Author(s), under exclusive license to Springer Nature Limited]; (C) HRTEM image of Cu-12C, revealing a 2-3 nm continuous and conformal alkanethiol layer; (D) Illustration shows Cu catalysts via hydrophobic treatment by alkanethiol; (E) FE and partial current densities of C₂₊ on Cu-12C under different current densities; (F) The variety of product selectivity with different interfacial wettability. [Reproduced with permission^[55].Copyright 2023, The Author(s)]. FE: Faraday efficiency; GDL: Gas diffusion layer; HRTEM: High-resolution transmission electron microscope.

The above results indicate that enhancing the C-C coupling process in CO₂RR is significant for the formation of C₂₊ products, and the efficiency depends on the coverage of *CO intermediates on the catalyst surface. For smaller-sized Cu-based NPs, adjusting the chemical environment of catalytically active sites can adjust the adsorption capacity for *CO, while for slightly larger catalyst particles, the adsorption capacity of *CO can be adjusted through nano-confined-space strategies. Enhancing the concentration of *CO can effectively promote the formation of C₂₊ products. Modifying the surface of Cu NPs with different materials will influence the selectivity of CO₂RR products. When using hydrophobic materials, HER can be suppressed, thereby enhancing the CO₂ reduction efficiency. When modified with hydrophilic salt materials, C₁ products can be promoted, thereby reducing the generation of C₂₊. The synthesis of nanocavity structures using nanoconfinement effects can enrich and stabilize CO, greatly increasing the concentration of CO on Cu NPs and promoting the generation of C₂₊ products. Adjusting the Cu surface through different molecules is beneficial for reducing the energy barrier of C-C coupling and stabilizing intermediates, thereby improving CO₂RR performance. However, the stability of Cu NPs is generally poor, and using appropriate carriers to anchor Cu NPs and developing new synthesis strategies are feasible methods to increase stability.

Cu-based single atoms

In Cu-based SACs, Cu atoms exist in the form of single atoms and can effectively participate in the reaction process. SACs have maximum atomic utilization efficiency and high catalytic activity^[83], and can promote the catalytic reaction^[84]. Compared with other Cu-based catalysts, Cu SACs have a higher utilization rate of active sites^[85]. These SACs can maximize the utilization of Cu resources through reasonable design and preparation methods^[86]. Due to their monodispersed catalytic sites, some Cu SACs cannot overcome the energy barrier of C-C coupling, inhibiting such reactions and thereby the formation of C₂₊ products and improving the selectivity of C₁ products such as CH₄, CO, CH₃OH, and HCOOH^[87,88]. The key to improving the performance of SACs is to regulate the interaction between catalytically active sites and reaction intermediates^[89]. The CO₂RR performance of Cu SACs can be improved by precisely adjusting the coordination environment and electronic structure of the central metal^[90,91]. Most single-atom Cu is loaded onto C-based or NC materials through coordination with C or N. A strong correlation exists between catalytic activity and the coordination environment (such as CN and coordinating atoms) of Cu metal centers^[92]. For example, Cu-C and Cu-N can promote the generation of CH₄ and CO, respectively, and Cu-N₂, Cu-N₃ and Cu-N₄ also form different catalytic effects. Metal and nitrogen-doped carbon (MNC) catalyst can activate and reduce CO₂ to various products. The separation of electrons and geometric sites of a single metal atom will make the poor adsorption of H, thus eliminating the Tafel-type reaction and inhibiting the HER effectively^[93]. In this section, we will summarize in detail the impact of the coordination environment of Cu SACs on their performance in CO₂RR.

When Cu SACs coordinate with C in graphene to form Cu-C bonds, the activation sites of CO₂ molecules are changed to promote the formation of *OCHO intermediates, thereby promoting the production of CH₄^[94-98]. In the process of CO₂RR, the Cu-C bond formed by loading Cu atoms on carbon materials is usually conducive to the formation of *OCHO, which is converted to CH₄. Shi et al. immobilized Cu single atoms on graphdiyne (GDY) and constructed Cu-C bonds for the first time^[99], which not only stabilized the single Cu atoms but also enhanced the electronic control effect. They found that the Cu-C bond could induce the formation of *OCHO. Furthermore, the catalyst had a stronger binding ability to CO₂ than H₂O, thus suppressing HER. The maximum FE of CH₄ was 81% at a stable electroreduction of ten h. Zhao et al. prepared Cu SACs with Cu-C₂ coordination structure by accurately anchoring single Cu atoms onto hydrogenated graphdiyne (HGDY), using the special structure of 1,3,5-triethynylbenzene through the Sonokashira reaction Cu single atoms (Cu-SAs)/HGDY^[98]. They confirmed that the low-coordination Cu-C₂ provided the Cu single-atom center with more positive charges and promoted the formation of H• and improved the selectivity of CH₄. At -1.1 Vvs. RHE, FE of CH₄ was 44%, and the CH₄ current density was 230.7 mA cm^-2. Importantly, the catalyst achieved a turnover frequency (TOF) of 2,756 h^-1.

When Cu coordinates with N in graphite carbon nitride (g-C₃N₄), it tends to produce CH₄^[100], and it can stabilize Cu single atoms^[99]. Li et al. incorporated single-atom Cu into the nitrogen cavity of the host g-C₃N₄^[100]. The catalyst exhibited an FE of CH₄ of 49.04% and a maximum CH₄/C₂H₄ ratio of 35.03. The Cu-N-C structure enhances the adsorption of *COOH and promotes the desorption of *CO, which is beneficial for the production of CO. Chen et al. anchored single-atom Cu onto a carbon nitride structure (Cu-N-C)^[101], which enhanced the adsorption of *COOH and promoted the desorption of *CO. A high FE of CO of 98% with a local current density of 131.3 mA cm^-2 was achieved at -0.67 V vs. RHE. Cai et al. prepared a carbon dot-supported SAC (Cu-CDs) with a unique CuN₂O₂ site [Figure 8A-E]^[87]. Since the generation of hydrocarbons requires to transfer more electrons and thereby a higher energy barrier, the product of CO₂RR is usually CO. Therefore, the authors used partial carbonization to adjust the electronic structure of the Cu atom, thus reducing the total endothermic energy of *OCHO. Due to the introduction of oxygen ligands, FE of CH₄ increased to 78% [Figure 8F and G]. Furthermore, they believed that pyridinic N was the anchor site for CO₂ capture, due to the oxidation of Cu²⁺ in Cu-CD which had more pyridinic N than CD. Atomic-dispersed Cu-N₃-C catalysts also exhibit excellent selectivity for CO. Chen et al. used biomass guanosine to directly synthesize a two-dimensional (2D) atom-dispersed Cu-N₃-C catalyst (CuG-1000)^[102]. The electrons were transferred from Cu coordinated with N to *COOH, thus regulating the adsorption and desorption of key intermediates and improving the selectivity of CO. At -0.65 V vs. RHE, the FE of CO of CuG-1000 reached 99% with a current density of 6.53 mA cm^-2. Recently, as shown in Figure 8H-K, Purbia et al. synthesized Cu-SAC-N-doped carbon quantum dots (Cu SAC-N-CQDs) catalysts using CuCl₂ as the Cu source and dopamine hydrochloride as the carrier through a hydrothermal reaction at 160 °C in the presence of citric acid for six h^[103]. Figure 8H-K displays the structural schematic diagram and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of the Cu SAC N-CQD. The selectivity of CH₃CH₂OH is highly related to the coordination of Cu SACs, and changes with the variation of Cu loading. During the CO₂ reduction progress, the in-situ generated Cu is the active site, which is beneficial for the production of CH₃CH₂OH. This catalyst had a unique Cu-N-C site and achieved an FE of 70% for CH₃CH₂OH at -0.2 V vs. RHE [Figure 8I], as well as a stability exceeding 50 h. Pan et al. designed C-supported Cu catalysts with atomic-dispersed N, OH-Cu₃ sites Cu-N/interconnected mesoporous carbon fiber (IPCF)^[104]. Atomically dispersed and N-coordinated Cu moieties were loaded onto a mesoporous N-doped carbon fiber carrier. The N, OH-Cu₃ sites had moderate *CO adsorption affinity and low barrier for *CO hydrogenation. Cu-N/IPCF could enhance nano confinement to prevent CO from escaping mesochannels, which greatly increased the residence time of *CO, and continuously reduced to CH₄ on Cu sites. At 300 mA cm^-2, this catalyst showed a high FE of CH₄ of 74.2%.

Figure 8. (A) Scheme of the low-temperature calcining procedure for Cu-CD catalysts; (B) Large-field of view and (C) magnified view of TEM images; (D) Relatively large-field of view and e typical view of HAADF-STEM images of distributed single Cu atoms in carbon dots. Yellow circles in (E) indicate typical single Cu atoms; (F) FE and current density of Cu-CDs; (G) Stability test of Cu-CDs and CuPc at their highest ECR FE potentials. [Reproduced with permission^[87]. Copyright 2021, The Author(s)]; (H) Structural schematic diagram of Cu SACs N CQDs electrocatalyst; (I) Faraday efficiency at different potential of Cu-SACs N CQDs; (J, K) HAADF-STEM image of all the Cu species on the carbon-dots Cu-N-CQDs (scale bar: 2nm). (Reproduced with permission^[103]. Copyright 2024, Elsevier B.V.). HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; FE: Faraday efficiency; SACs: Single-atom catalysts; TEM: Transmission electron microscope; CQDs: Carbon quantum dots.

The coordination of Cu single atoms with N to form Cu-N_x sites results in various CO₂RR products during electroreduction. Dong et al. synthesized locally planar, symmetrically fractured planar-symmetry-broken CuN₃ (PSB-CuN₃) SACs^[92]. The catalyst exhibited an FE of HCOOH of 94.3% at -0.73 V vs. RHE and could run stably in a flow cell for 100 h. The active centers of formate products were highly concentrated in the CuN₃ region. In the CO₂RR process, CuN₄C₄ sites tended to produce *COO-, while CuN₃C₃ and CuN₂C₂ tended to produce *CHO. Compared with highly symmetrical CuN₄C₄, the ΔG (0.23 eV) of the HCOOH on CuN₃C₃ was much lower than that of CO (0.68 eV), CH₄/CH₃OH (0.78 eV), and HER (1.01 eV). Xia et al. prepared ultra-high density Cu single atoms on thin-walled N-doped carbon nanotubes (TWN)^[105], named TWN-Cu_13.35-600-SACs, with a Cu content of 13.35 wt%. The researchers established that the adjacent Cu-N₃ site was the active site. The FE of CH₃CH₂OH improved with the increase of the density of Cu-N₃ sites. As shown in Figure 9A-C, in H-cell, the highest FE of CH₃CH₂OH was about 81.9%. In addition, the stability of the catalyst exceeded 25 h. Xu et al. synthesized a Cu SAC (Cu-N₄-NG) based on N-doped graphene (NG) using a two-step pyrolysis method^[106]. Cu atoms coordinated with four adjacent N atoms. They found that Cu-N₄ portion was beneficial for the CO₂ activation step. The CO₂RR on Cu-N₄-NG was less thermodynamically hindered, effectively limiting HER. Cu-N₄-NG achieved an FE of CO of 80.6% at -1.0 V vs. RHE. Cheng et al. prepared Cu-N₄-C/1100 containing Cu-N₄^[107]. They adjusted the coordination environment and electronic structure of Cu through pyrolysis. The edge-hosted Cu-N₄ was the key active site for generation CO, which strongly interacted with *COOH and facilitated the desorption of *CO. Compared with Cu-N₃-C/800, which contained Cu-N₃ sites, Cu-N₄-C/1100 exhibited excellent performance for CO production. At -0.9 V vs. RHE, the catalyst achieved an FE of CO of 98%. Karapinar et al. prepared Cu-N-C SACs using CuCl₂ as a Cu source through simple pyrolysis. Cu atoms were coordinated by four N atoms to form CuN₄ sites dispersed in an N-doped conductive carbon matrix^[108]. During electrolysis, the separated sites instantaneously transformed into Cu NPs, which may be the active species. At -1.2 V vs. RHE, the FE of CH₃CH₂OH was 55%. As shown in Figure 9D and E, Zhao et al. encapsulated single-atom Cu on N-doped porous carbon (Cu-SA/NPC)^[109]. They found that the reaction mainly occurred at the coordination sites between Cu and four pyrrole-N atoms (Cu-pyrrolic-N₄). The coordination of Cu with four pyrrole-N atoms reduced the free energies for CO₂ activation. Cu-SA/NPC could reduce CO₂ to CH₃COOH, CH₃CH₂OH, and CH₃COCH₃ products, and CH₃COCH₃ was the major product (FE = 36.7%). Generally speaking, Cu SACs with Cu-N₄ sites (CN = 4) have the best catalytic effect on CO₂RR and exhibit superior stability. Recently, Roy et al. used CuCl₂ solution as the Cu source to form Cu-N₂ sites by loading Cu single atoms onto a carbon nitride (CN) matrix through the metal ion exchange method [Figure 9F-H]^[110]. This catalyst had high-density (1.5 at%) single-atom sites. The cooperative Cu-Cu sites generated on the 9N pores in Cu-PTI synergistically enhanced the activity of the CO₂ to CH₄. As shown in Figure 9I and J, at -0.84 V vs. RHE, the FE of CH₄ was 68%.

Figure 9. (A) Preparation and characterization of the catalysts; (B) HAADF-STEM images of TWN-Cu_13.35-600-SACs; (C) FE and the product distribution at different potentials. (Reproduced with permission^[105]. Copyright 2023, American Chemical Society); (D) TEM images of Cu-SA/NPC. (E) FE of CO₂RR products on Cu-SA/NPC. [Reproduced with permission^[109]. Copyright 2020, The Author(s)]; (F) Synthetic scheme of Na-PHI, Li-PTI; (G) HRTEM images of Na-PHI; (H) HRTEM images of Li-PTI; (I) FE of CH₄ comparison of Cu-PHI/PTI; (J) Partial current density of CH₄ for the catalysts. (Reproduced with permission^[110]. Copyright 2024, Wiley-VCH GmbH). HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; FE: Faraday efficiency; SACs: Single-atom catalysts; TWN: Thin-walled N-doped carbon nanotubes; Cu-SA/NPC: Single-atom Cu on N-doped porous carbon; CO₂RR: Carbon dioxide reduction reaction; TEM: Transmission electron microscope; PHI: Poly (heptazine imide); PTI: Poly (triazine imide); HRTEM: High-resolution transmission electron microscope.

Furthermore, Cu coordinated by heteroatoms can promote the formation of C₂₊ products. Wu et al. broke the coordination symmetry of Cu sites to form Cu-S₂N₁ sites in atomic precision in Cu₆ clusters [Cu₆(MBD)₆, MBD = 2-mercaptobenzimidazole]^[111], which was conducive to the generation of *COOH instead of *OCHO, thereby inducing the formation of more valuable hydrocarbons. At -1.4 V vs. RHE, the catalyst exhibited an FE of 65.5% for hydrocarbons, where FE of CH₄ and FE of C₂H₄ accounted for 42.5% and 23%, respectively. Recently, Lv et al. anchored Cu single atoms onto F, O, N-co-doped carbon composites (FONCDs) via precipitation reaction, and the prepared CuFONC catalyst had a stable CuN₂O₁ configuration and a high density of Cu SAs^[112]. They established that the addition of F and O resulted in excellent stability of CuN₂O₁ during CO₂RR. The addition of F adjusted the 3d orbitals charge of Cu atom in CuFONC. The Cu coordinated with N/O atoms, and the Cu-N/Cu-O ratio was 2:1. The Cu-N site in CuFONC was found to be an adsorption site for *CO. At -1.3 V vs. RHE, FE C₂ was 80.5%.

The catalytic activity of Cu SACs is connected with the CN and degree of distortion of the Cu metal. Cu-N and Cu-C are effective active sites for CO₂RR. Cu-C and Cu-N sites generally generate hydrocarbons and oxygen-containing compounds, respectively. Researchers have found that a single Cu site usually produces C₁ products, and adjacent Cu can promote C-C coupling. It is difficult to precisely control the active sites, as they usually generate C₁ products, making it challenging to obtain C₂₊ products. Therefore, the key is to design the active site reasonably. Cu SACs with high CO₂RR performance can be designed by regulating the CN of Cu and N (Cu-N₄ being optimal), adjusting the loading density of Cu single atoms, and doping other heteroatoms to coordinate with Cu to saturate the Cu coordination sites.

CuAu

Ouyang et al. found that on Cu alloy catalysts such as CuAu and CuAg, the surface coupling hydrogenation activity was enhanced and indirectly inhibited solvent hydrogenation of intermediates, resulting in higher CH₃CH₂OH selectivity than Cu (100)^[124]. Morales-Guio et al. deposited Au NPs by physical vapor deposition on top of polycrystalline Cu foil (Au/Cu)^[125]. Au/Cu efficiently reduced CO₂ to CH₃CH₂OH. They observed that CO₂ was reduced to CO on Au NPs, which was then enriched and further reduced to alcohols on nearby Cu. Shen et al. synthesized AuCu alloy embedded in Cu submicron conical arrays (AuCu/Cu SCA)^[114]. Cu atoms were replaced by larger Au atoms, causing lattice expansion and forming AuCu alloys. Altering the content of Au affected CO₂RR activity and the selectivity of C₂H₅OH/C₂H₄ products. The CO generated at Au sites coupled with *CH₂ intermediates at Cu sites to form *COCH₂, which further generated CH₃CH₂OH. In addition, the catalyst could maintain electrocatalytic activity for 24 h at high current densities. Wei et al. prepared an Au-doped Cu nanowire (Cu Au NWs)^[126]. They modified a small amount of Au NPs on the surface of Cu NWs, using the homonuclear method. The addition of Au NPs resulted in a rougher surface, exposing more active sites. In addition, the *CO intermediates generated on Au NPs aggregate on Cu NWs, promoting the adsorption of *CO. The interface electrons transferred from Cu to Au induced electron-deficient Cu, which facilitates the adsorption of *CO. At -1.25 V vs. RHE, the FE C₂₊ increased from 39.7% of Cu NWs without Au to 65.3%. Zheng et al. successfully synthesized an Au-Cu Janus nanostructure catalyst (Au-Cu Janus NSs) using a seed growth strategy^[127]. Cu atoms were deposited on one side of concave cubic Au seeds. The Au-Cu domain boundaries in the spatial separated Janus nanostructures facilitated synergistic catalysis. At -0.75 V vs. RHE, the FE C₂₊ was 67%.

CuAg

Ma et al. prepared Ag₆₅-Cu₃₅ Janus nanostructure (JNS-100) catalysts with (100) facets [Figure 10A-C]^[128]. Compared to Cu, Ag₆₅-Cu₃₅ JNS-100 had higher electron abundance. This catalyst exhibited the highest selectivity for C₂H₄ and C₂₊ products with FE values of 54% and 72%, respectively [Figure 10D-F]. As mentioned earlier, Cu (100) had a lower energy barrier for C-C coupling^[49]. As shown in Figure 10G, Ag₆₅-Cu₃₅ JNS-100 exposed Cu (100), Janus nanostructure induced tandem catalysis and electron transfer between Ag and Cu. Wei et al. improved the FE C₂₊ in CO₂RR by concatenating nano electrocatalysts^[129]. As shown in Figure 10H, they designed a three-dimensional (3D) catalyst electrode with Ag NPs, which were deposited at the bottom of a Cu nanoneedle array (Cu needle-Ag) to generate the intermediate product CO. The dense Cu needles would prevent CO from flowing out of the Ag NPs, causing further dimerization of CO on the Cu nanoneedles to form C₂₊ products. Through this nanostructure design, FE C₂₊ was 70% in a flow cell. Choi et al. synthesized a tight atomic Cu-Ag interface on the Cu NW surface^[130]. The tight atomic Cu-Ag interface promoted the synergistic effect of CO-Ag* and H-Cu*, greatly improving the selectivity of CH₄. The maximum FE of CH₄ at -1.17 V vs. RHE was 72%. Hoang et al. electrodeposited 3,5-diamino-1,2,4-triazole inhibitors on thin CuAg alloy films^[123]. The CO₂RR performance of the alloy wire with an Ag content of 6% was the best, achieving an FE of nearly 60% and 25% for C₂H₄ and C₂H₅OH, respectively, at low potentials of -0.7 V vs. RHE. CuAg wire (6%) exhibits Ag-Cu bonds. Also, the Ag-Ag distance in the CuAg film is slightly reduced, indicating that Ag atoms were at least partially alloyed with smaller Cu atoms. As shown in Figure 10I, Li et al. synthesized Ag,S-Cu₂O/Cu catalysts through an in-situ dual doping strategy^[131]. This catalyst had a typical 3D porous architecture which resulted in more active sites. The Cu atoms near heteroatoms in Cu₂O/Cu served as efficient active centers for CH₃OH production. S^- could regulate the electronic structure and morphology, while Ag⁺ inhibited HER. FE of CH₃OH reached 67.4%. Du et al. cascaded Ag NPs and AgCu single-atom alloy (SAA) (AgCu SANP)^[132]. The Ag NPs selectively reduced CO₂ to CO, followed by the formation of multiple carbon products on AgCu SAA. Owing to the asymmetric binding of Cu atom to the adjacent Ag atom, the incorporation of the single Ag atom enhanced the adsorption energy of CO on the Cu sites. At -0.65 V vs. RHE, FE C₂₊ reached 94% ± 4% under a working current density of about 720 mA cm^-2. Qi et al. synthesized CuAg alloy catalysts (CO₂-10-Cu₉₄Ag₆) by the co-electrodeposition method^[133]. Under CO₂-supersaturated conditions, the co-deposition of Ag and Cu forms a CuAg alloy, with the Cu (100) facets being preferentially exposed. The dispersed Ag atoms on Cu strengthened C-O bond of C₃ chain. The FE 2-propanol in the H-cell with 1.0 mol L^-1 CsHCO₃ was 56.7%. They observed that C₃ products can only be detected when the electrolyte cation is Cs⁺ and CO₂ is supersaturated. This was attributed to the larger radius of Cs⁺ that can interact with two *CO. Recently, Li et al. synthesized Cu₃N-Ag NCs by doping Ag into Cu₃N nanocubes (NCs)^[134]. The Ag sites generated CO, while the Cu sites promoted C-C coupling to *COCHO, which enhanced the production of C₂H₄. As shown in Figure 10J, the cascade Ag-Cu dual sites in Cu₃N-Ag NCs could promote C-C coupling and generate C₂H₄.

Figure 10. (A) HAADF-STEM image of Ag/Cu interface in Ag₆₅-Cu₃₅ JNS-100; (B) HAADF-STEM images of Ag₆₅-Cu₃₅ JNS-100; (C) EDS elemental mappings of Ag₆₅-Cu₃₅ JNS-100; (D) FE of major CO₂-reduction products obtained on Ag₆₅-Cu₃₅ JNS-100; (E) Comparison of C₂H₄ FE between Ag₆₅-Cu₃₅ JNS-100 and Cu NCs at different potentials; (F) Comparison of C₂₊/C₁ product ratios between Ag₆₅-Cu₃₅ JNS-100, Ag + Cu mixture and Cu NCs; (G) Schematic illustration of a plausible CO₂RR mechanism on Ag₆₅-Cu₃₅ JNS-100. (Reproduced with permission^[128]. Copyright 2022, Wiley-VCH GmbH); (H) Schematic illustration of the synthesis of Cu needle-Ag catalyst. (Reproduced with permission^[129]. Copyright 2023, Advanced Functional Materials published by Wiley-VCH GmbH); (I) Schematic diagram of the in situ dual-doping process for preparing the x, y-Cu₂O/Cu catalysts. [Reproduced with permission^[131]. Copyright 2022, The Author(s)]; (J) Introduction of Ag into Cu₃N with weaker binding strength with reaction intermediates of CO₂RR could increase local CO coverage and facilitate *CO protonation to *CHO. (Reproduced with permission^[134] . Copyright 2024, American Chemical Society). HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; JNS: Janus nanostructure; FE: Faraday efficiency; NCs: Nanocrystals; CO₂RR: Carbon dioxide reduction reaction; EDS: Energy dispersive spectrometer.

CuZn

In 2017, Kattel et al. established that the ZnO/Cu (111) interface sites were the real active sites of CuZn catalysts that could reduce CO₂ to CH₃OH^[135]. The synergistic effect of Cu and ZnO facilitated the synthesis of CH₃OH. Zhu et al. prepared Cu/ZnO-CeO₂ catalysts by the flame spray pyrolysis method^[136], and the catalysts exhibited better CH₃OH selectivity (70%) than Cu/ZnO and Cu/CeO₂. In 2022, Amann et al. found that the hydrogenation of CO₂ preferentially occurred on ZnO^[137]. Wan et al. found that phase-separated bimetallic Cu-Zn catalysts had a lower energy barrier to form the *COOH than the core-shell one^[138]. The FE of CO was 94% and stability exceeded 15 h. Zhen et al. found that the synergistic effect between Cu and Zn can result in effective CO adsorption. Both balanced Cu-Zn sites and Zn-rich Cu-Zn sites promoted C-C coupling^[139]. Recently, as shown in Figure 11A-D, Zhang et al. developed CuZn alloy/CuZn aluminate oxide (Al₂O₄) composite electrocatalyst (Cu₉Zn₁/Cu_0.8Zn_0.2Al₂O₄)^[140]. The Al₂O₄ triggered a robust electron interaction among Cu, Zn, and Al, leading to the generation of numerous highly reactive interfaces characterized by exceptional activity. Supported by interface effects, the optimized catalyst achieved an FE C₂₊ of 88.5% and a current density of up to 400 mA cm^-2 [Figure 11E-J].

Figure 11. (A) Schematic illustration for the preparation of the CuZn/CuZnAl₂O₄ catalyst; (B) TEM image; (C) HR-TEM image; (D) HAADF-STEM image; (E) LSV curves toward CO₂RR on CuZn/CuZnAl₂O₄; Product distributions and corresponding FE produced by CuZn/CuZnAl₂O₄ (F); Cu/CuAl₂O₄ (G); and CuZn (H); (I) C₂₊ partial current density; (J) Stability test of CuZn/CuZnAl₂O₄. (Reproduced with permission^[140]. Copyright 2023, ELSEVIER B.V. and Science Press). HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; CO₂RR: Carbon dioxide reduction reaction; TEM: Transmission electron microscope; HR-TEM: High-resolution transmission electron microscope; LSV: Linear sweep voltammetry.

Others

Liu et al. loaded Ni SAC onto Cu catalyst using the electrostatic self-assembly method^[141]. This catalyst promoted the dimerization of CO, reaching an FE of C₂H₄ of approximately 62% at -1.4 V vs. RHE. At 500 mA cm^-2, the catalyst could sustain stability for 14 h. Recently, Song et al. prepared a double-layer hollow spherical nanosphere catalyst (Ni-SA@Cu-NP)^[142]. The inner layer of Ni-SA@Cu-NP was composed of dispersed Ni atoms, and the outer carbon layer consisted of Cu NPs. CO was generated on the Ni site in the inner layer, accumulated in the cavity, and then overflew to the outer Cu site for CO dimerization, achieving FE C₂₊ of 74.4%. The addition of In often reduced CO₂ to HCOOH. Wei et al. prepared a series of bimetallic In_xCu_y NP electrocatalysts^[143]. They adjusted the growth direction of the crystal face by changing the molar ratio of In/Cu. Among the catalysts, In_1.5Cu_0.5 NPs had an FE of HCOOH of 90% at -1.2 V vs. RHE. Wang et al. prepared an In₂O₃/Cu catalyst with 3D succulent plant morphology^[144]. The surface tensile strain and interfacial electronic interaction reduced the energy barrier for the formation of *HCOO. At -1.4 V vs. RHE, FE of HCOOH was 87.5%. Li et al. prepared a metal Cu/Pd-1% catalyst^[145]. The Pd element was uniformly distributed on the Cu NPs. The doping of Pd revealed the d-band center of Cu toward the Fermi level. The higher d-band center location favors the adsorption of intermediates. This catalyst enhanced the affinity for *CO, exhibiting an FE C₂₊ of 66.2%. Zheng et al. introduced a single-atom Pb alloy Cu catalyst (Pb₁Cu) [Figure 12A and B]^[146], which could completely convert CO₂ to HCOOH with an FE of nearly 96%. The activation of Cu sites on Pb₁Cu catalyst directed CO₂RR towards the HCOO* instead of the COOH*. As shown in Figure 12C, the highest FE of HCOOH was 95.7% at -0.72 V vs. RHE, and an FE exceeding 80% could still be maintained even when the partial current density was up to -1200 mA cm^-2. Yang et al. prepared rod-like bimetallic CuBi₇₅ catalysts^[122]. Bi, C, and O were uniformly dispersed on CuBi₇₅ catalysts. At -0.77 V vs. RHE, FE of HCOOH was 100%. As shown in Figure 12D, Cao et al. introduced single Bi atoms on Cu NPs to form BiCu-SAA^[147]. The single-atom Bi on the Cu (111) surface resulted in significant charge redistribution. Local electrons can be transferred to the antibonding orbitals of CO₂ molecules, thereby activating and further reducing CO₂. As shown in Figure 12E and F, the FE C₂₊ of the BiCu-SAA catalyst was 73.4 %, and the catalyst remained stable at the current density of 400 mA cm^-2.

Figure 12. (A) Schematic illustration of CO₂ conversion into HCOOH over a Pb₁Cu SAA; (B) HAADF-STEM image and (C) FEs of all reduction products and j-V curves of Pb₁Cu catalyst. [Reproduced with permission^[146]. Copyright 2021, The Author(s), under exclusive license to Springer Nature Limited]; (D) Schematic diagram of BiCu-SAA (E) The FEs of CO₂RR products on the catalysts with different Bi content at an applied potential of -1.10 V; (F) FE of C₂₊ products on the BiCu-SAA and control Cu-Nano catalysts at different applied potentials. (Reproduced with permission^[147]. Copyright 2023, Wiley-VCH GmbH). HAADF-STEM: High-angle annular dark-field scanning transmission electron microscopy; CO₂RR: Carbon dioxide reduction reaction; FE: Faraday efficiency; SAA: Single-atom alloy.

Hu et al. synthesized La(OH)₃/Cu by modifying the Cu catalyst with La(OH)₃^[148]. The modification of La(OH)₃ caused Cu to be in an electron-deficient state, which favored *CO adsorption and *CO-*COH coupling, thus increasing C₂₊ selectivity. The main component was C₂H₄. This catalyst achieved an FE C₂₊ of 71.2%, which was 1.2 times higher than that of pure Cu, and exhibited a current density of up to 1,000 mA cm^-2. Li et al. doped the p-block metal atom Ga into Cu and prepared a CuGa catalyst^[149]. At -1.07 V vs. RHE, the CuGa catalyst exhibited an excellent FE C₂₊ of 81.5% at a current density of 0.9 A cm^-2. When the current density reached 1100 mA cm^-2, the catalyst still maintained a high FE C₂₊ of 76.9%. It was attributed to the p-d hybrid interaction between Ga and Cu, which redistributed the electronic structure and improved the bonding strength of intermediates. This finesse can also be expanded to other p-block metals, such as Al and Ge. Xie et al. synthesized a bimetallic Mg-Cu catalyst^[150]. The modification with Mg not only enhanced the activation of CO₂ but also stabilized the Cu sites. Compared to Cu catalysts, Mg-Cu catalysts had abundant Cu⁺ sites and stronger adsorption capacity for CO₂ and CO. The FE C₂₊ of the catalyst reached 80% (mainly C₂H₄), and the current density reached 1 A cm^-2.

Generally speaking, introducing other metals as additives can effectively improve CO₂ reduction efficiency. The metals used for doping and the levels of doping significantly influence both the products and the efficiency. For instance, introducing hydrophilic metals may result in the preferential formation of oxygen-containing compounds. The interaction between two metals usually improves CO₂ reduction efficiency and leads to high selectivity for C₂₊ products. When Cu forms an alloy with other metals, the diversity in electronegativity can lead to charge transfer, resulting in a change in the d-band center of active sites. This electronic effect can regulate the adsorption of key intermediates. The addition of various metals often leads to different products. The Cu-Au catalysts are often beneficial to the production of CH₃CH₂OH. The Cu-Ag catalysts usually facilitate the generation of oxygen-containing C₂₊ products. The Cu-Zn catalysts are favorable to the production of C₂H₄. The addition of In, Bi and Sn usually leads to HCOOH. At the same time, doping with other metals can also achieve high current densities, some of which can reach industrial-grade current densities, and there is hope for commercialization. However, most of the doped metals are precious metals, and their use is not very economical. Furthermore, strong interactions between alloy components can lead to better catalytic performance of alloy catalysts than single-component NPs^[151,152].

C-based materials

C-based materials are commonly used to disperse SACs due to their advantages such as low cost, good stability, and good conductivity^[156]. The properties of the substrate can change the electronic state of the active sites, thereby altering the reaction pathway and catalytic mechanism. To prepare commercial carbon-supported Cu SACs, Xu et al. mixed bulk Cu with molten Li and subsequently added CB as the support (Vulcan XC-72), followed by leaching LiOH by thoroughly mixing with water, in which process Cu atoms transferred to the carbon surface^[153]. On this support, Cu single atoms can be highly dispersed. Cu was still atomically dispersed after prolonged electrocatalysis. At -0.7 V vs. RHE, FE of CH₃CH₂OH achieved 91%. The stability exceeded 16 h. Recently, Pan et al. developed a dual continuous mesoporous carbon carrier (IPCF) to support the synthesis of Cu-N/IPCF catalysts from Cu single atoms^[104]. They designed carbon supported Cu catalysts by regulating the atomic scale structure of Cu active sites and designing the mesoscale structure of carbon supports. The catalyst was derived from block copolymers and had interconnected mesopores. Its unique long-range channel provided a microenvironment lacking H₂O, prolonging the transport pathway of CO intermediates. And this catalyst had N, OH-Cu₃ sites, which effectively inhibited HER and achieved an FE of CH₄ of 74.2% at 300 mA cm^-2. As shown in Figure 13A-C, Yang et al. prepared a through-hole carbon nanofiber catalyst with a large and uniformly distributed Cu single-atom modification CuSAs/through-hole carbon nanofibers (TCNFs)^[157]. Due to the self-supporting and through-hole structure of the catalyst, a large number of Cu SAs were fully exposed on the surface, which was conducive to capturing CO₂. At -0.9 V vs. RHE, FE of CH₃OH was 44%, and the CH₃OH partial current density was -93 mA cm^-2.

Figure 13. (A) Pyridine Cu-N₄ structure; (B) LSV curves and (C) FE of all products at CuSAs/TCNFs. (Reproduced with permission^[157]. Copyright 2019, American Chemical Society); (D) TEM image and (E) corresponding enlargement of the sites 1-4 and intensity profiles of Cu_-1/hNCNC; (F) FE, current density (j), and product distribution for Cu_-1/hNCNC at different polarization potentials; (G) Ethanol FE and j as a function of time for Cu_-1/hNCNC during a chronoamperometric test at -0.30 V. (Reproduced with permission^[161]. Copyright 2024, American Chemical Society). FE: Faraday efficiency; LSV: Linear sweep voltammetry; CuSAs/TCNFs: Cu Single Atoms/through-hole carbon nanofibers; TEM: Transmission electron microscope; hNCNCs: NC nanocages.

NC materials

NC materials are ideal carriers due to their tunable pore structure, intense metal-heteroatom interactions, and stability^[104,158]. Compared with C-based materials, NC materials contain N species, a higher N content, and a uniform N arrangement, which can provide abundant and more accurate coordination sites for single atoms^[159]. Feng et al. used B- and N-doped graphene materials (N-doped GDY and B-doped GDY) as supports to anchor individual Cu atoms^[160]. They found that Cu@doped GDY can spontaneously capture CO₂. Cu atoms had strong interactions with adjacent C, B, or N atoms. The electronegativity of coordinated elements is a crucial factor in improving catalytic performance. They discovered that the catalytic performance of Cu@N-doped GDY was better. Recently, as shown in Figure 13D-G, Xu et al. prepared Cu_-1/hNCNC by vapor deposition on multi-stage NC nanocages (hNCNCs) using commercial Cu₂O powder as a Cu source^[161]. This catalyst had CuOCu-N₄-oxygen-bridged binuclear Cu sites on hNCNCs. Cu_-1/hNCNC exhibited the most advanced low overpotential for C₂H₅OH at 0.19 V vs. RHE, with FE of CH₃CH₂OH reaching 56.3% at -0.30 V vs. RHE, and the Cu active site was very steady during the CO₂RR process.

Cu NPs usually employ C-based materials and oxides as supports, and the interaction between appropriate supports and Cu NPs can improve the selectivity of the product^[156]. For example, the g-C₃N₄ supports can not only elevate the d-orbital of Cu to the Fermi level, but also serve as an additional active center for CO₂RR, enhancing the selectivity of CH₄ and C₂H₄^[162]. As shown in Figure 14A and B, Li et al. synthesized CuNCN by the precipitation method, using CuCl as a Cu source and adding NH₃•H₂O and cyanamide (H₂NCN), and then prepared Cu-based/C_xN_y catalysts by pyrolysis of CuNCN^[163]. The state of Cu species on the g-C₃N₄ carrier was regulated by the pyrolysis temperature. Among the prepared catalysts, CuNCN-300 obtained by pyrolysis of CuNCN at 300 °C had the highest FE of C₂H₄ of 48.5% and a current density of 500 mA cm^-2. The good dispersibility of Cu₃N in CuNCN-300 promoted the selectivity of C₂H₄, while the tris-s-triazine structure (g-C₃N₄ fragment) in C_xN_y enhanced the reaction of *CO - *CHO on the Cu surface. Li et al. prepared a p-NG-Cu-7 catalyst by loading 7-nm Cu NPs on pyridine-rich N-graphene (p-NG)^[164]. As shown in Figure 14C, due to the rich Lewis base sites of p-NG, the formation of *COOH was promoted, which was then enriched on Cu NPs and converted into CO and CH_xO with subsequent C-C coupling. Compared with the catalytic performance of GO-Cu-7 catalyst deposited on conventional graphene (FE of C₂H₄ = 7.5%), the p-NG-Cu-7 catalyst had an FE of C₂H₄ of 19%, and the selectivity of C₂H₄ in hydrocarbons reached 79%. The p-NG-Cu-7 catalyst could also produce C₂H₅OH with 63% FE.

Figure 14. (A) The schematic illustration of synthesis of CuNCN and preparation of Cu-based/C_xN_y by CuNCN pyrolysis; (B) HRTEM of CuNCN-300. (Reproduced with permission^[163]. Copyright 2022, Elsevier B.V. All rights reserved); (C) Product selectivity of the hydrocarbons generated from the p-NG-Cu-7 catalyzed reduction at -0.9 V. (Reproduced with permission^[164]. Copyright 2016, Elsevier Ltd). NCN: Elements N and C; HRTEM: High-resolution transmission electron microscope.

Oxides

The synergistic effect between metal oxide carriers and Cu can modulate the electronic structure of the catalyst. Loading Cu on oxide (ZnO/Al₂O₃) can promote the reduction of CO₂ to CH₃OH^[165]. The Lewis acid sites in metal oxides (such as Al₂O₃ and Cr₂O₃) have been proven to activate CO₂ molecules and promote CO₂ methanation^[166-169]. As shown in Figure 15 A, Chen et al. anchored Cu SAs to ultra-thin porous Al₂O₃ rich in Lewis acid centers (Cu/Al₂O₃ SAC)^[170]. When Cu single atoms were supported by ultra-thin porous Al₂O₃, the Cu/Al₂O₃ SAC achieved an FE of C₂H₄ of 62%. As shown in Figure 15B and C, Cu atoms on Al₂O₃ were in a higher oxidation state (electrons transferred from Cu atoms to the carrier). In addition, the Cu atoms on Al₂O₃ exhibited higher d-band centers, indicating an improved electron transfer ability. The electron-accepting properties of Al₂O₃ were beneficial for stabilizing methanation intermediates and reducing the energy barrier. TiO₂ could stabilize Cu NPs and provide more active sites. As shown in Figure 15D and E, Yuan et al. prepared CuO/TiO₂ catalysts by the hydrothermal method using CuO and TiO₂ and then reduced them in situ to Cu/TiO₂ catalysts^[171]. TiO₂, as a semiconductor material, can serve as an oxidation-reduction electron carrier and assist in CO₂ adsorption^[172-174]. It can stabilize the CO₂RR intermediates and reduce overpotential. The CuO/TiO₂ significantly increased the adsorption capacity of CO₂. Firstly, a large amount of CO₂ was adsorbed on TiO₂. Then, the adsorbed CO₂ obtained an electron from the Cu/TiO₂ and was converted to CO₂^-, which could be dimerized to *C₂O₂^-. As shown in Figure 15F and G, the Cu/TiO₂ catalyst effectively reduced CO₂ to multiple oxygen-containing carbon compounds such as CH₃CH₂OH, CH₃COCH₃, and CH₃CH₂CH₂OH. The maximum total FE was 47.4%.

Figure 15. (A) Schematic diagram of CO₂RR on Cu/Al₂O₃ SAC; (B) Electronic structure and (C) projected densities of states (pDOS) of d-orbitals with an aligned Fermi level of Cu/Al₂O₃ SAC and Cu/Cr₂O₃ SAC. Color code: Cu, brick red; Al, purple; Cr, gray; O, red. (Reproduced with permission^[170]. Copyright 2021, American Chemical Society); (D) TEM images of CuO/TiO₂-5 catalyst in low magnification; (E) Schematic diagram of CO₂RR on CuO/TiO₂-5 catalyst; (F) FEs for different products over various CuO/TiO₂ catalysts at -0.85 V vs. RHE in CO₂-saturated 0.5 M KHCO₃ aqueous solution; (G) FEs for different products over CuO/TiO₂-5 catalyst at various potentials. (Reproduced with permission^[171]. Copyright 2018, MDPI, Basel, Switzerland). CO₂RR: Carbon dioxide reduction reaction; SAC: Single-atom catalysts; FE: Faraday efficiency; RHE: Reversible hydrogen electrodes; TEM: Transmission electron microscope.

Others

Abdinejad et al. synthesized Cu-Pd/MXene catalysts using MXene-based (Ti₃C₂T_x) materials as carriers^[175]. They paired 2D MXene with bimetallic Cu-Pd. Compared with Cu-Pd, the Cu-Pd/MXene had a larger active surface area and electron transfer rate. This stemmed from MXene improving electron transfer and having a larger electrochemically active surface area (EASA). Due to the unique multi-layer composition of MXene, the catalytic surface area and conductivity increased. At -0.5 V vs. RHE, FE of HCOOH reached 93%, and overall battery energy efficiency (EE) reached 47%. MOFs are considered as the ideal catalyst supports. The Cu-N coordination bond formed by Cu(II) and N-heterocyclic ligands has moderate strength, located between Cu-O coordination and Cu porphyrin bond^[176]. Thus, Cu-N coordination is more stable. Therefore, Chen et al. designed a MOF (2Bn-Cu@UiO-67) of encapsulated N-heterocyclic carbene (NHC) ligand linked to Cu SAC^[177], which exhibited an FE of CH₄ of 81%. Due to the interaction between Cu and NHCs, the electron occupancy rate on the d orbital was much lower than that of Cu foil. The σ donation from NHC enriched the electron density of Cu single atoms, promoting the adsorption of CHO*. The high porosity promoted the spread of CO₂ molecules towards 2Bn-Cu, remarkably improving the catalytic efficiency.

The above results indicate that the appropriate carrier can improve the performance and selectivity of the catalyst, and some supports can enhance catalyst stability.

The following Table 2 lists the catalytic performance data of different types of Cu catalysts.