Strategies for the proton-coupled multi-electron reduction of CO₂ on single-atom catalysts

Abstract

Catalytic reduction of CO₂ to high value-added chemicals is an important approach for tackling the rising CO₂ concentration in the atmosphere. Recently, a range of heterogeneous and potentially low-cost single-atom catalysts (SACs) have emerged as promising candidates for the reduction of CO₂. However, in comparison to conventional metal nanoparticle catalysts, SACs have long faced challenges in reactions involving multiple reactants and multiple reaction steps due to the limitation of isolated metal sites. This review presents the most recent research advances on the development of single-atom catalysis for deep reduction of CO₂. Based on the approaches proposed for proton-coupled multi-electron transfer, detailed introductions and summaries were classified into three categories: 1) strengthen the metal–support interaction to achieve a synergistic catalysis; 2) rational design and regulation of the coordination environment of isolated metal atoms; 3) development of SACs with multi-atom active sites. Finally, the main challenges and future research directions in the field of SACs for CO₂ reduction are proposed.

---

1. Introduction

With the escalating consumption of fossil fuels, the rapid global industrial revolution has resulted in a significant surge in greenhouse gas emissions, causing irreversible damage to the Earth's ecological environment. Since the 1990s, numerous nations have commenced collaboration in tackling the issue of global climate change. In 1997, the Kyoto Protocol achieved a momentous milestone by constructing an international institutional framework to address climate change, effectively restraining emissions of greenhouse gases.^1–3

Utilizing CO₂ as a sustainable C1 feedstock to produce valuable chemical products via catalytic reduction is an important approach for CO₂ mitigation. Despite its potential advantages, there are still technological difficulties in its practical application. The intrinsic thermodynamic inertness of CO₂ (with a standard Gibbs free energy of −394.38 kJ mol⁻¹) presents the most fundamental challenge.⁴ CO₂ is the ultimate byproduct of fossil fuel combustion, and it possesses the lowest energy density among all carbon-containing species. Transition of CO₂ requires significant energy input to overcome the energy barrier of the first C=O bond dissociation. As a result, the activity and selectivity of catalysts for CO₂ reduction are relatively low.^5,6

Presently, the techniques for CO₂ conversion can be divided into three categories: electrocatalysis, photocatalysis and thermocatalysis.^7,8 The electrocatalytic reduction of CO₂ is mainly driven by sustainable energy sources such as solar and wind power. Nevertheless, the electrochemical CO₂ reduction competes with hydrogen evolution reaction in aqueous solution.^9,10 The sluggish dynamics of CO₂ activation and the limited selectivity of this technology hinder its practical applications. Photocatalysis technology offers a solution to alleviate the emission of CO₂ by harnessing solar energy as a driving force.¹¹ However, regulating the dynamics of photoinduced charge transfer is still a significant scientific issue. Additionally, quantum efficiency is also a significant factor influencing the efficiency of a photocatalyst.¹² Thermocatalysis converts CO₂ into more valuable products through heat energy. Thermocatalysis often grapples with the challenges of high temperature-induced side reactions and catalyst sintering. However, it possesses the capacity to convert large amounts of CO₂ concurrently, rendering it more amenable to large-scale application.^13–15

In the past decades, there has been considerable focus on the single-atom catalysts (SACs). SACs offer unique distinctiveness, primarily due to their unsaturated binding sites, the highest atomic utilization efficacy, and the explicitly defined active sites.^16,17 In 1979, Yates et al. utilized nuclear magnetic resonance (NMR) spectroscopy and ¹³C relaxation time distribution to discover that Rh species can be stably distributed in atomic form on the surface of Al₂O₃.¹⁸ Subsequently, Heiz et al.¹⁹ employed the soft-landing approach to deposit Pd atoms onto the surface of MgO. They found that the single-atom Pd catalyst can drive acetylene cyclo-trimerization at 300 K. In 2011, Zhang et al.¹⁷ reported a Pt₁/Fe₂O₃ SAC prepared with the wet chemistry method. They demonstrated that this catalyst has excellent catalytic performance in the catalytic oxidation of CO. Subsequently, research into SACs has entered an intense development stage and a significant body of literature in the field of CO₂ conversion has been produced.

Li et al.²⁰ first reported the use of Ni SACs for the electrocatalytic reduction of CO₂ to CO. The single-atom Ni–N–C catalysts, synthesized via a ZIF-8-assisted approach, demonstrated superior performance compared to Ni nanoparticles. Recent density functional theory (DFT) studies revealed that using graphitic carbon nitride (g-C₃N₄) as support, the single-atom Pd sites can reduce CO₂ into HCOOH, whereas the single-atom Pt sites favour the production of CH₄. Carbon nanotube (CNT) supported Ni- and Ru-based SACs were also reported for CO₂ hydrogenation. Both catalysts were capable of CO₂ hydrogenation at 210 °C, with Ni SACs exhibiting higher selectivity for CO, while Ru SACs exhibiting higher selectivity for CH₄.

Despite the significant success achieved by SACs, they face significant challenges in catalyzing reactions that involve multiple reactants.²¹ In a SAC, only the single isolated atom acts as the active site for the adsorption of reactant. Nevertheless, in most catalytic reactions, co-adsorption of multiple reactants is involved on the catalyst surface. In addition, single atoms often have a positive charge and are in a state of electron deficiency, resulting in weaker electron transfer capabilities when compared to clusters and particles. Consequently, SACs tend to convert CO₂ into CO, as the electronic state limitations of the active center result in limited product diversity. This issue is illustrated by the difficulty in achieving six-electron transfer and methane or methanol production during CO₂ hydrogenation with SACs.^22,23

This review summarizes the latest breakthroughs in single-atom catalysis, with the aim of achieving deep reduction of CO₂via electrocatalysis, thermocatalysis, and photocatalysis. The ways to regulate the activity of CO₂ reduction on SACs can be summarized as follows: 1) strengthen the metal–support interaction to achieve synergistic catalysis; 2) rational design and regulation of the coordination environment of isolated metal atoms; 3) development of SACs with multi-atom active sites. Furthermore, the reasons for the change in catalytic activity, from single atoms up to clusters, are also explored in this article.

2. Overview of catalytic CO₂ reduction paths

2.1 Mechanism of thermocatalytic CO₂ hydrogenation

Methanol is an important C1 product in thermal catalytic CO₂ hydrogenation, while it is often accompanied by the co-production of several other byproducts (CO/CH₄/HCOOH). This paragraph will analyse the mechanism of C1 product generation from the perspective of the methanol production pathway. Numerous reaction pathways have been proposed for the methanol synthesis through CO₂ hydrogenation, owing to variations in the catalysts system and reaction conditions. The common reaction pathways are the formate pathway, the reverse water gas shift (RWGS) pathway, and the trans-COOH pathway (Fig. 1).²⁴ In the formate pathway, CO₂ is hydrogenated to form formate (HCOO*), formyl (H₂COO*), formaldehyde (H₂CO*), methoxy (H₃CO*), and eventually methanol (CH₃OH). The formation of HCOO* and H₂COO* in the formate pathway is the rate-controlling step.^25,26 In the RWGS pathway, CO₂ undergoes hydrogenation to generate a *COOH intermediate, which subsequently dissociates into CO. The CO is then further hydrogenated to produce formyl (HCO*), H₂CO*, H₃CO*, and CH₃OH.^27–29 The trans-COOH pathway generates a variety of intermediates, including carboxyl groups (COOH*), dihydroxy methylene (COHOH*) and hydroxymethyl (COH*). Further hydrogenation leads to the formation of hydroxyformyl (HCOH*) and hydroxymethyl (H₂COH*) intermediates, ultimately resulting in the production of CH₃OH.³⁰

Fig. 1 Possible reaction pathways of CO₂ hydrogenation to CO, CH₃OH, and CH₄. *(X) indicates adsorbed species. Reproduced from ref. 24 with permission from American Chemical Society, copyright 2017.

The direct hydrogenation of CO₂ to produce C2 products is a complex process involving multiple steps of bond breaking and carbon coupling, generating various C1 products such as CH₃OH, CO, and CH₄. There remains controversy over the reaction mechanism of CO₂ hydrogenation due to the presence of different concentrations of surface species. Arakawa et al. proposed a mechanism involving CO as an intermediate, combining the RWGS and synthesis gas-to-ethanol mechanisms. This involves CO generation, CO and H₂ dissociation promoting carbon chain growth to form C_xH_y, followed by undissociated CO insertion (CO insertion into alkyl or alkyl insertion into CO), and finally hydrogenation to produce high-carbon alcohols. C–O dissociation-insertion and C–C coupling occur at two active sites and require the distance between the active sites to be close enough to increase the possibility of the reaction occurring.^31,32

2.2 Mechanism of electrocatalytic CO₂ reduction

CO₂ electrocatalytic reduction is a multistep proton–electron coupled transfer reaction involving the transfer of 2e⁻, 4e⁻, 6e⁻, and 8e⁻. As the CO₂ electroreduction reaction occurs at the interface between a solid-phase electrocatalyst electrode and a CO₂ saturated electrolyte, the heterogeneous reaction at this interface mainly involves the following three steps: (1) CO₂ chemisorption on the surface of the electrocatalyst; (2) electron transfer or proton–electron coupled transfer causing the cleavage of the C=O bond and/or the formation of C–H bond; (3) the product desorbing from the surface of the electrocatalyst and diffusing into the electrolyte (Fig. 2).³³

Fig. 2 Overview of reaction pathways for CO₂RR towards different products. Black spheres, carbon; red spheres, oxygen; white spheres, hydrogen; blue spheres, (metal) catalyst. The arrows indicate whether proton, electron or CPETs take place. Reproduced from ref. 33 with permission from Springer Nature, copyright 2019.

Two pathways are involved in the formation of C1 products. The first pathway is known as the formate pathway, which occurs when hydrogen reacts with the C-site of CO₂ to form an *OCHO. This intermediate is further hydrogenated to produce HCOOH. The second pathway is referred to as the carboxyl pathway, which occurs when hydrogen reacts with an O-site of CO₂ to form the *COOH intermediate. This intermediate is then hydrogenated to form *CO. CO is a crucial intermediate that can desorb from the surface of the catalyst. Moreover, it can be reduced to *CHO or *COH and proceed through various proton–electron transfers to generate CH₄ and CH₃OH ultimately. Along with the other pathway, HCOOH can also be formed.

In the process of generating C2 products, CO is a crucial intermediate produced by the reduction of CO₂. The process of C2 product generation involves two primary paths of C–C coupling. In the first path, two CO molecules undergo dimerization, giving rise to *OC–CO, which then couples. In the second path, CO undergoes hydrogenation to produce *CHO/*COH, which subsequently couples with another CO molecule to form *OC–CHO/*OC–COH. Crucial intermediates, such as *OCCO, *OCCHO, and *OCCOH, undergo multiple hydrogenation steps to generate C2 products like C₂H₄ and C₂H₅OH.³³

The generation of C3 products, including propanol (C₃H₇OH), starts with the CH₃CHO* intermediate. *CO insertion triggers the formation of the intermediate CH₃COCHO*, which then undergoes four H⁺/e⁻ transfers, leading to the intermediate CH₃CH₂CHO*. Further two H⁺/e⁻ transfer steps eventually yield C₃H₇OH.³⁴

2.3 Mechanism of photocatalytic CO₂ reduction

The CO₂–H₂O reaction on a semiconductor surface results in several pathways, such as the formaldehyde, carbene, and glyoxal pathways, leading to the production of high-value-added products. Activation of CO₂ in the formaldehyde pathway (Fig. 3a) is achieved by combining the active site of the catalyst with one O atom from CO₂. The transfer of an electron to CO₂ results in the formation of a ·CO₂⁻ radical, followed by a proton addition, resulting in a ·COOH radical. The coupling of a proton and an electron in the ·COOH ultimately gives rise to the production of formic acid. Finally, the formic acid accepts two protons, producing formaldehyde and H₂O. CH₃OH and CH₄ can also be synthesized via this pathway if adequate electrons and protons are present, with CH₄ formation occurring through the ·CH₃ radical intermediate.³⁵

Fig. 3 Schematic illustration for the (a) formaldehyde pathway, (b) carbene pathway and (c) glyoxal pathway for CO₂ reduction. The blue and red dashes highlight the starting CO₂˙⁻ species and key intermediates involved in the pathways. Reproduced from ref. 35 with permission from Royal Society of Chemistry, copyright 2020.

The carbene pathway initially consumes two electrons to generate CO, which serves as an intermediate for subsequent reactions to yield either CH₃OH or CH₄ (Fig. 3b). The adsorption strength between CO and the catalyst determines whether CO quickly desorbs from the surface or accepts electrons and protons to produce subsequent products. Subsequently, the ·CH₃ radical undergoes two reaction pathways: one involves its reaction with OH⁻ to form CH₃OH, while the other involves its combination with a proton to produce CH₄.^36,37

The glyoxal pathway for CO₂ transformation involves both of its O atoms coordinating with the active site in a bidentate manner (Fig. 3c). Subsequently, ·CO₂⁻ radical interacts with H⁺ to form formate. Coupling of formate with another H⁺ generates formic acid. The formyl radical (·HCO) dimerizes to produce glyoxal, a precursor to C₂ and C3 products. Like the reaction pathway mentioned above, the combination of ·CH₃ radical with a proton generates CH₄, with CO acting as a byproduct of the reaction. The appearance of other reaction intermediates in the catalytic system may also cause polymerization of ·CH₃ into C2 products.³⁸

3. Transforming CO₂ into deep reduction products over SACs

3.1 SACs for CO₂ hydrogenation by thermocatalysis

Heterogeneous catalytic hydrogenation is one of the most viable industrial pathways for CO₂ conversion. CO₂ exhibits high inertness, which requires the catalytic hydrogenation to occur at elevated temperatures. High-pressure reaction conditions are necessary for producing high-valued products; however, the associated high energy consumption and cost pose significant challenges to the large-scale production.³⁹ Harsh reaction conditions also present a heightened challenge in preserving the stability of SACs. The study of thermal catalytic CO₂ hydrogenation endeavours to enhance catalyst durability, diminish energy consumption, and generate value-added products.⁴⁰

3.1.1 Regulation of the support. Single-atom catalyzed CO₂ hydrogenation typically occurs under harsh conditions (high temperatures or pressures). Supports need to have a strong anchoring effect on the metal atoms. Metal oxides are often used as the support to enhance the catalyst stability. In addition, the difference in support properties can lead to significant differences in the modulation of the electronic structure of active centers. Ja Hun Kwak et al.⁴¹ investigated the performance of Pd SACs supported on both carbon and Al₂O₃ in the CO₂ methanation. Interestingly, the study revealed that Pd/Al₂O₃ showcased superior activity compared to the carbon-supported counterpart. DFT calculations indicated that the Al₂O₃ support regulates the electronic structure of Pd atoms, enabling them to efficiently participate in the adsorption and conversion of intermediate species.

Fan et al.⁴² employed a vacuum filtration technique to immobilize Ru atoms onto a porous hexagonal boron nitride substrate (Table 1). Under the conditions of 350 °C and 1 MPa, a remarkable CO₂ conversion of 28.7% was attained, with CH₄ selectivity as high as 93.5%. The high specific surface area and abundant surface defects of boron nitride facilitate a strong anchoring effect for Ru atoms, thereby ensuring atomic dispersion as well as high catalyst stability. The Ru K-edge X-ray absorption near edge structure (XANES) reveals the formation of a coordinated structure between the Ru and N atoms, leading to a substantial decrease in the oxidation state of Ru. According to the DFT results (Fig. 4), low-valence Ru is identified as the primary active center in CO₂ methanation. Camila Rivera-Cárcamo et al.⁴³ found that Ru single atoms on carbon carriers were difficult to maintain atomic dispersion, while those on TiO₂ were more stable. The obtained SAC exhibited remarkable differences in the CO₂ methanation reaction. X-ray photoelectron spectroscopy (XPS) revealed that the carbon-supported Ru atoms manifest an electron-deficient state, in contrast to those supported by TiO₂. The Ru species, abundant in electrons, foster CH₄ formation more favorably.

Table 1 Comparison of the catalytic performance of different SACs in CO₂ hydrogenation

Catalyst	Temp. (°C)	P (MPa)	CO2 Con. (%)	CH4 Sel. (%)	HCOOH (TON)	CH3OH Sel. (%)	C2H5OH (TOF/h^−1)	Ref.
Ru/pBN-0.58%F	350	1	28.7	93.5	—	—	—	42
RuSA/TiO2	210	0.61	15.6	97.3	—	—	—	43
Rh1/ZrO2	440	1	62	99	—	—	—	48
Ir/AP-POP	120	6	—	—	25 135	—	—	44
Ru@MCM-3	90	3.5	—	—	>2000	—	—	49
Ir1/In2O3	200	6	—	—	—	—	481	45
Cu/ZrO2	180	3	∼2.7	—	—	100	—	51
7.5% Pt/MoS2	150	3.2	—	—	—	81.3	—	52

---

Fig. 4 Reaction free-energy diagram for CO₂ hydrogenation at 350 °C (623 K). Top view of RuO and RuO₂ at the BN vacancy. Reproduced from ref. 42 with permission from American Chemical Society, copyright 2019.

The liquid-phase hydrogenation reaction of CO₂ occurs in organic solvents, wherein the metal sites are susceptible to loss. Substrate materials such as nitrogen-rich organic polymers can effectively stabilize metal atoms. Shao et al.⁴⁴ fabricated an amino pyridine-functionalized porous organic polymer for the construction of Ir SAC. (Fig. 5a), which demonstrated excellent activity for the hydrogenation of CO₂ into formate under liquid-phase conditions. XPS analysis indicated that the Ir atoms are primarily in a trivalent oxidation state, which is attributed to the coordination of the aminopyridine with Ir atom. DFT calculations demonstrated that this unique structure facilitates hydrogen dissociation, thereby promoting formate production.

Fig. 5 (a) A schematic illustration of the synthesis of Ir/AP-POP SAC. Reproduced from ref. 44 with permission from Elsevier, copyright 2019. (b) In situ FTIR spectra of CO₂ hydrogenation with Ir₁–In₂O₃ catalyst bubbled in aqueous solution. (c) Ir–In₂O₃ samples with different Ir loadings after exposure to CO₂ at 200 °C. Reproduced from ref. 45 with permission from American Chemical Society, copyright 2020.

Lately, the controlled growth of the carbon chain via CO₂ hydrogenation to yield high-value C2 products has attracted considerable interest. However, the pursuit of attaining remarkable selectivity for ethanol by single-atom catalysis continues to pose a challenge. Ye et al.⁴⁵ engineered a bifunctional Ir₁/In₂O₃ SAC by anchoring single-atom Ir onto an In₂O₃ support. By integrating two active catalytic centres, this Ir SAC proved effective for the hydrogenation of CO₂ in the liquid-phase, presenting high selectivity for ethanol (>99%) and a high initial turn-over frequency (481 h⁻¹). Detailed characterization studies revealed that isolated Ir atoms couple with neighboring oxygen vacancies to form Lewis acid–base pairs, activating CO₂ and producing a carbonyl (CO*) intermediate species adsorbed on the Ir atom. This CO* then couples with the methoxy to form a C–C bond (Fig. 5b and c). This efficacious bifunctional SAC introduces a novel strategy for designing catalysts for single-atom catalysis, through the synergistic utilization of single-atom sites and the roles of substrates.

3.1.2 Adjustment the coordination environment. Compared to nanoparticle catalysts, changes of the coordination environment have a more significant impact on SACs. Such modifications could improve stability, regulate electronic structure, and modify the adsorption strength of intermediate species, leading to an improvement of catalytic performance. Chen et al.⁴⁶ deployed MIL to synthesize Pt single-atom and particle catalysts. The Pt SAC exhibited superior activity in the liquid-phase hydrogenation of CO₂ to CH₃OH. In situ XPS and Fourier transform infrared spectroscopy (FTIR) measurements revealed distinct reaction pathways for the two catalysts, where HCOO and COOH served as intermediates for the Pt SAC and nanoparticle catalyst, respectively. EXAFS and DFT calculations indicated a stronger metal–ligand synergy in the SAC, which reduces reaction activation energy and enhances CH₃OH selectivity.

Büchel et al.⁴⁷ used the double-nozzle flame pyrolysis method to prepare Rh SACs supported on Al₂O₃. When different promoters were added, the catalyst activity significantly varied. Potassium-promoted catalysts primarily produced CO, while barium-promoted catalysts could convert CO₂ into CH₄. The CO-FTIR results indicated that promoters had a significant impact on the valence state of Rh. The presence of barium resulted in the Rh catalyst exhibiting a valence state closer to zero. The valence state distribution of Rh atoms was the main cause of the selectivity difference. The research conducted by Li et al.⁴⁸ further demonstrated that tailoring both the electronic properties of metal atoms and carriers through alkali ion doping can serve as an effective strategy for tuning the selectivity of CO₂ hydrogenation. They found that by doping Na⁺ into Rh₁/ZrO₂, even at high temperatures of up to 440 °C and a high CO₂ conversion of 43%, the reaction pathway can be switched from methanation to RWGS. Mechanistic studies indicated that Na⁺ enhances the adsorption of CO₂ and COOH* on the ZrO₂ carrier through Na–O interactions, thereby inducing the formation of CO. Additionally, Na⁺ endows the Rh single atom with more electron deficiency, which greatly suppresses the activation of H₂ on the catalyst surface and promotes the desorption of CO. Wang et al.⁴⁹ employed molecular sieve synthesis technology to encapsulate Ru atoms within the inner walls of MCM-41. The resultant structure imparted the essential structural stability required for the dispersion of Ru at the atomic level (Fig. 6a). The catalyst facilitated the conversion of CO₂ into formate during the liquid-phase hydrogenation process. DFT calculations unveiled that the highly dispersed Ru atoms served as the active sites for H₂ dissociation (Fig. 6b).

Fig. 6 (a) Scheme of the synthesis of Ru@MCM catalyst. (b) Catalytic cycle and respective Gibbs free energies (kJ mol⁻¹) for the proposed mechanism for Ru@MCM-3. DFT calculations performed with PBE0/TZP//PBE0/SVP at 298 K. Reproduced from ref. 49 with permission from Elsevier, copyright 2021.

3.1.3 Optimization the interatomic distances. SACs often face challenges in producing C2+ products due to highly dispersed active centres. By modulating the distance between the dual atom sites, one can modulate the electronic states of the active metal centre, thereby altering the adsorption behavior of intermediate species and facilitating C–C coupling. Millet et al.⁵⁰ found that, on the MgO surface, atomic dispersed Ni sites exhibited high activity in catalyzing CO₂ into CO. As the nickel loading increased, the preference for CH₄ and CH₃OH consistently increased, corresponding to the diminishing distance between the nickel atoms. DFT calculations unveiled that those variations in the nickel concentration resulted in distinctive reaction mechanisms during the hydrogenation of CO₂. Hence, fine-tuning the interatomic distances presents a promising avenue for single-atom catalysis in producing more valuable compounds from CO₂.

Zhao et al.⁵¹ reported a copper-based catalyst featuring isolated copper sites for the conversion of CO₂ into methanol. Detailed characterization studies revealed that the catalyst, comprised of single-atom Cu–Zr units with Cu₁–O₃ units, exclusively facilitates the synthesis of methanol at approximately 180 °C. In contrast, the presence of small copper clusters or nanoparticles leads to the formation of the by-product, CO (Fig. 7a and b). The exceptional activity of these isolated copper sites, coupled with the discernible structural patterns identified in the study, broadens the prospects of single-atom catalysis in thermal catalytic CO₂ hydrogenation. Li et al.⁵² reported that the synergistic interaction between adjacent Pt monomers on MoS₂ markedly enhances the catalytic activity for CO₂ hydrogenation, compared with isolated single atoms. Adjacent Pt monomers were achieved by increasing the Pt mass loading, while maintaining the dispersion of Pt at the atomic level. As shown in Pt L₃-edge XANES profiles, with the increase of Pt loading, the white line intensity for Pt/MoS₂ decreased which indicated the difference of electron density. Mechanistic studies indicated that the synergies between adjacent Pt single-atom sites can effectively reduce the reaction barrier and follows a reaction pathway distinct from that of isolated Pt SAC (Fig. 7c). Isolated Pt sites tend to catalytically convert CO₂ directly into methanol, whereas adjacent Pt SAC hydrogenates CO₂ stepwise into formic acid and methanol (Fig. 7d and e).

Fig. 7 (a) The catalytic activity of CAZ-x and CAZ-15–H. (b) Gibbs free-energy profile of CO₂ hydrogenation to CH₃OH/CO. Reproduced from ref. 51 with permission from Springer Nature, copyright 2022. (c) Histogram of the contents of isolated Pt monomers, neighbouring Pt monomers, and patches of Pt monomers for atomically dispersed Pt/MoS₂ with different Pt loadings. (d) In situ DRIFT spectra of 0.2% Pt/MoS₂ and 7.5% Pt/MoS₂. The spectra were obtained after exposure of the samples to the mixed gas (CO₂ : H₂ = 1 : 3, 1 bar) at 150 °C for 0.5 h. (e) The Arrhenius plots of 0.2% Pt/MoS2 and 7.5% Pt/MoS₂. Reproduced from ref. 52 with permission from Springer Nature, copyright 2018.

3.2 SACs for CO₂ reduction by electrocatalysis

Electrocatalytic CO₂ reduction typically occurs in inorganic salt solutions, accompanied by hydrogen evolution reaction.^53,54 During the process of electrocatalysis, the electrochemical workstation acts as an electron pump, providing the requisite electrons. CO₂ is adsorbed onto a cathode electrode coated with a catalyst, forming adsorbed products following the dissociation of CO₂ and the transfer of protons and electrons. The difficulty of obtaining electrons during the CO₂ reduction process, as well as the ability of inorganic salt to provide carbon, result in different CO₂ reduction products.^55,56 Moreover, a significant advantage of CO₂ reduction by electrocatalysis lies in their ability to take place under mild reaction conditions. Currently, the catalytic reduction of CO₂ using SACs has garnered considerable attention. Important progress has been made in understanding the dynamic evolution of catalyst structures during the reaction process, reaction mechanisms, and the development of novel SACs.

3.2.1 Regulation of the support. Due to the requirement of conductivity in electrocatalytic processes, electrocatalysts are typically composed of metal oxides and carbon materials with good conductivity. However, traditional carbon supports have poor ability to anchor metal atoms, leading to metal aggregation during the reaction. Xu et al.⁵⁷ studied carbon-supported Cu SACs and found that during the CO₂ reduction process, single-atom copper tends to aggregate and form particles, as evidenced by in situ EXAFS results. Therefore, dynamic evolution of SACs is prone to occur during the reaction, highlighting the importance of selecting suitable supports to maintain their dispersion.

Genovese et al.⁵⁸ successfully achieved the high-selectivity conversion of CO₂ to acetic acid under low-potential conditions using Fe SAC supported on carbon–nitrogen. HAADF-STEM images (Fig. 8a) confirmed the isolated Fe single-atom sites in the catalyst. Temperature-programmed XPS (Fig. 8b) measurements confirmed that the Fh–FeOOH phase transitions from an Fe(III) phase to a mixed Fe(III)/Fe(II) phase. Combining with the results from EXAFS spectra, it can be seen that CO₂ primarily adsorbs onto the Fe(II) sites and bonds with the pyridine N. N dopants on carbon have a dual function: they coordinate with CO₂-related species and stabilize these Fe(II) species, hindering their further reduction. The excellent C–C coupling performance is due to the synergistic effect between the surface nitrogen species of the support and Fe–OOH.

Fig. 8 (a) HAADF STEM micrographs of sample Fe/N–C (scale bar 5 nm). (b) Normalized Fe K edge spectra of Fe/N–C at −0.5 V in 0.05 M KHCO₃ (black line); Fe/N–C at OCP in 0.05 M KHCO₃ component of the fit (red line); Fe foil component of the fit (magenta line); envelope of Fe/N–C at −0.5 V with Fe(III) component (Fe/N–C at OCP) and Fe⁰ component (Fe foil) in fit 1 (green line). Reproduced from ref. 58 with permission from Springer Nature, copyright 2018.

Metal oxides frequently contain oxygen vacancies, which are a key factor in improving their catalytic performance.⁵⁹ Oxygen vacancies typically contain extra electrons, thereby facilitating the adsorption and activation of CO₂ reactants. Consequently, the overpotential of the reaction decreases while its activity increases. Chen et al.⁶⁰ utilized a porous monolayer of Al₂O₃ as a substrate for the deposition of individual copper atoms, enabling the highly selective electrochemical reduction of CO₂ to CH₄ (Fig. 9a). Bader charge analysis indicated that copper atoms on the Al₂O₃ surface had higher oxidation states. The Lewis acid sites on the surface of the Al₂O₃ facilitated the transfer of electrons between the copper atoms and the support. DFT calculations (Fig. 9b) suggested that Cu atoms with a low electron density favor bonding with CO₂, producing electron-rich intermediates that facilitate the production of CH₄. Hung et al.⁶¹ synthesized a copper-loaded Fe SAC for the electrochemical conversion of CO₂ to CH₄. In situ X-ray absorption spectroscopy (XAS) results indicated that the Fe single atoms remain highly dispersed during the reaction process. DFT calculations showed that the Fe single atoms efficiently stabilize CO species while promoting the formation of *COH intermediate species. The formation of *COH leads to a decreased energy barrier during the conversion of CO₂ to CH₄. On the other hand, Wang et al.⁶² designed a CeO₂ supported Cu SAC to reduce CO₂ to CH₄. The strong metal–support interaction between CeO₂ and Cu ensures a high dispersion of Cu on the CeO₂ surface. Theoretical simulations revealed the presence of three oxygen vacancies surrounding the Cu atoms. A high concentration of surface defects aids in promoting the adsorption of CO₂ and carbon-containing intermediates, thereby enhancing the efficiency of CO₂ conversion into CH₄.

Fig. 9 (a) CO₂RR performance of Cu/C–Al₂O₃ SAC. (b) Calculated free-energy diagrams for the CO₂RR over Cu/Al₂O₃ SAC and (c) Cu/Cr₂O₃ SAC. Reproduced from ref. 60 with permission from American Chemical Society, copyright 2021.

3.2.2 Adjustment the coordination environment. The coordination structure of SACs primarily depends on the type of central atoms, as well as the ligands in their surroundings and the adjacent active sites. The difference in electronegativity between the metal and adjacent atoms leads to charge transfer and structural reconstruction.⁶³ As a result, the d-band center of the metal atom deviates from the Fermi level, influencing the adsorption properties of the metal atom on the intermediate. Therefore, the rational construction of the coordination environment for SACs can effectively enhance their electrocatalytic activity for CO₂ electroreduction.

Zhao et al.⁶⁴ designed a carbon–nitrogen coated Cu SAC for the electrochemical reduction of CO₂ to multi-carbon products. The reaction achieved a selectivity of 36.7% for acetone. The interaction between Cu and pyrrole nitrogen is elucidated through EXAFS and XPS measurements. DFT calculations (Fig. 10a) confirmed that the Cu atoms, which coordinated with four pyrrole nitrogen, serve as the active sites, and can decrease the free energy required for CO₂ activation and carbon–carbon coupling. Meanwhile, Yang et al.⁶⁵ developed a through-hole carbon nanofiber-loaded Cu SAC, a flexible material (CuSAc/TCNFs) (Fig. 10b). This catalyst can significantly promote the electrocatalytic reduction of CO₂ to methanol. DFT calculations show (Fig. 10c) that Cu atoms coordinated with different nitrogen species exhibit different adsorption abilities to intermediate species. Specifically, Cu atoms coordinated with four pyridine nitrogen have strong adsorption capabilities to the *CO intermediate species, allowing CO₂ to be transformed into CH₃OH instead of desorbing to produce gaseous CO.

Fig. 10 (a) Free energy diagrams calculated at a potential of −0.36 V for CO₂ reduction to CH₃COCH₃ on Cu–pyridinic-N₄ and Cu–pyrrolic-N₄ sites of Cu–SA/NPC. Reproduced from ref. 64 with permission from Springer Nature, copyright 2020. (b) The schematic and methanol faradaic efficiencies of CuSAs/TCNFs. (c) Free energy diagram of CO₂ to CO on pyridine N, Ni–N₄, and Cu–N₄ structure. Reproduced from ref. 65 with permission from American Chemical Society, copyright 2019. (d) Structural comparison between CoPc and CoPc–NH₂. (e) Product selectivity (FE) and total current density for a 12 h electrolysis of CO₂ reduction catalysed by CoPc–NH₂/CNT at −1.00 V versus RHE. Reproduced from ref. 67 with permission from Springer Nature, copyright 2019.

In addition to catalytic activity, regulating the N-coordination environment around the metal centres can also have a significant impact on the stability of the SACs.⁶⁶ Wu et al.⁶⁷ prepared a single-atom Co catalyst by absorbing cobalt phthalocyanine onto carbon nanotubes (CoPc/CNT), which demonstrated remarkable efficacy in catalyzing the reduction of CO₂ into CH₃OH (Fig. 10d). However, CoPc is prone to reduction during the reaction, leading to poor stability of the catalyst. Intriguingly, the introduction of an amino functional group on the benzene ring can significantly enhance the stability of the catalyst. Amino, acting as an electron-donating group, diminishes the electron transfer from Co to the benzene ring, effectively maintaining the stability of Co during the reaction process. (Fig. 10e) This inference can be substantiated by the UV-vis spectra of CoPc–NH₂/CNT before and after the reaction, as the amino-functionalized CoPc catalyst shows no significant structural changes under the reaction conditions.

3.2.3 Optimization the interatomic distances. The interaction between adjacent sites is a crucial factor influencing catalyst activity. Unlike the hybridization of p–d orbitals between metal and non-metal atoms, the d–d orbital hybridization between neighbouring metal sites results in significant electron asymmetry.⁶⁸ This distinction provides opportunities for the design of novel SAC systems. Karapinar et al. prepared a Cu–N–C catalyst with a CuN₄ structure via a pyrolysis method.⁶⁹In situ EXAFS measurements revealed that isolated Cu atoms undergo migration during the reaction, leading to the formation of multiple Cu centres that enable carbon–carbon coupling, thereby shifting the product distribution from CO to ethanol and ethylene (Fig. 11). Thus, rational control of interatomic distances proves to be an effective approach for C2 product production.

Fig. 11 N-doped carbon containing isolated Cu atoms has highly selective for CO₂ electroreduction to ethanol. Reproduced from ref. 69 with permission from John Wiley and Sons, copyright 2019.

Guan et al.⁷⁰ employed a nitrogen coordination strategy to load Cu single atoms on carbon–nitrogen materials and investigated their performance in electrocatalytic CO₂ reduction. Interestingly, at low Cu loading (2.1 mol%), Cu primarily existed in the form of single atoms, coordinated with four nitrogen atoms, which predominantly catalyzed the reduction of CO₂ to CH₄. At high Cu loading (4.9 mol%), the active sites transformed into adjacent Cu–N₂ structures. DFT calculations confirmed that the adjacent Cu–N₂ structures exhibited stronger adsorption towards CO species, promoting the formation of OC–COH intermediates. This elucidates the high selectivity observed in the catalytic reduction of CO₂ to C₂H₄ by adjacent Cu–N₂ centres. Furthermore, constructing bimetallic SACs is another effective approach to enhance catalytic activity. Apart from regulating the distance between metal centres by adjusting the metal loadings, the interaction between bimetallic species also plays a crucial role in determining catalyst performance.

Xie et al.⁷¹ reported an efficacious atomic pair electrocatalyst on a carbon nanosheet array electrode with a hierarchical integration, demonstrating a synergistic effect in prompting the activity and selectivity of the CO₂ reduction towards formate formation. The atomic pair catalyst comprises adjacent Ni and Sn, each surrounded by four nitrogen atoms (N₄–Ni–Sn–N₄). The synthesized NiSn atomic pair catalyst manifests exceptional activity for the CO₂ reduction towards formate with a turnover frequency of 4752 h⁻¹, outperforming previously reported SACs. Both experimental data and DFT calculations corroborated the electron redistribution of Sn influenced by the neighbouring Ni, which diminishes the energy barrier of the formation of *OCHO intermediate and renders this crucial step thermodynamically spontaneous. This synergistic catalysis offers a paradigm for the rational design and fabrication of atomic pair electrocatalysts with enhanced efficacy.

3.3 SACs for CO₂ reduction by photocatalysis

Photo-driven CO₂ conversion involves the reduction of CO₂ and the oxidation reactions of H₂O. The reaction process involves the transfer of electrons and holes and happens in three steps: (1) when the incident light energy (hv) is greater than or equal to the forbidden band width (E_g), electrons in the semiconductor catalyst valence band (VB) absorb energy from photons and are excited to the conduction band (CB), while holes are left in the VB simultaneously. (2) Photo-generated electrons in the CB and the holes in the VB are then transferred from the bulk to the surface of the catalyst. (3) The electrons and holes that have moved to the surface react with molecules adsorbed on the catalyst surface, such as CO₂ and H₂O.^72,73 Semiconductor materials (such as TiO₂, ZnS) or heterojunction structures composed of two materials are commonly used as traditional photocatalysts. Traditional photocatalytic materials are prone to decreased stability due to light corrosion. Metal–organic frameworks (MOFs) materials are increasingly used in photocatalysis due to their high specific surface area and porosity, which promote charge and mass transfer. Metal sites in MOFs materials are typically in a coordination unsaturated state, which have a stronger adsorption capacity for CO₂.^74,75 Currently, the main research directions are in the development of efficient MOFs-based photocatalysts and the study of the mechanism of the reaction process.

3.3.1 Regulation of the support. MOFs exhibit semiconductor properties, manifested in the fact that both the metal nodes and organic ligands can be excited and participate in the reduction of CO₂. The common ligand in MOFs is an aromatic carboxylate salt with high electron density, which can absorb photons upon excitation and generate photocarriers. MOFs have the following unique advantages as photocatalysts for light-driven CO₂ conversion: (1) the diversity of nodes and organic ligands in MOFs and their tuneable porosity can enable the construction of a variety of MOF-based catalysts containing rich catalytic sites; (2) the high porosity of MOFs facilitates the fast diffusion of reactants and products; (3) the three-dimensional crystal structure of MOFs provides a perfect platform for studying reaction processes; (4) the porous nature of MOFs endows them with excellent CO₂ adsorption capabilities, thereby increasing the concentration of CO₂ around active sites and promoting the light-driven conversion of CO₂.

Sun et al.⁷⁶ prepared a MOF-253 material and used it as a carrier to prepare Ru-based photocatalysts. Upon loading of Ru–carbonyl complex (MOF-253-Ru(CO)₂Cl₂), MOF-253 exhibits remarkable photocatalytic activity for CO₂ reduction under visible light irradiation. It is worth noting that the performance can be further enhanced by the simultaneous loading of a photosensitizer. Detailed characterization reveals that MOF-253 not only acts as a solid ligand for constructing a supported photocatalyst, but also serves as a platform for constructing composite photocatalytic systems, facilitating charge transfer between the photo-sensors and the surface-anchored photocatalysts.

Ye et al.⁷⁷ synthesized a novel catalyst using a MOF as the matrix, wherein coordinatively unsaturated single Co atoms were incorporated. This newly developed catalyst exhibits remarkable capability in capturing and photo catalytically reducing CO₂ under visible-light irradiation. Detailed mechanistic studies revealed that the presence of single Co atoms in the MOF greatly enhances the electron–hole separation efficiency in the porphyrin units, allowing the photo-generated excitons to selectively migrate from the porphyrin to the single Co atom center, thereby providing a long-lived electron supply to the adsorbed CO₂ molecules. As a result, the Co-containing porphyrin-MOF exhibits a significantly enhanced CO₂ photoreduction rate compared to the unmodified MOF, with a 3.13-fold increase in CO produce rate and a 5.93-fold increase in CH₄ generation rate.

In addition to MOFs, the covalent organic framework (COF) material, a category of crystalline porous materials formed through the covalent bonding of organic building blocks, has garnered increasing interest in the field of photocatalytic CO₂ conversion. Kou et al.⁷⁸ integrated the advantages of micropore structure and single atom catalysis by introducing MoN₂ single-atom sites into COF, realizing the first application of COF-based catalyst in reducing CO₂ to C_xH_y under visible light, with the C₂H₄ selectivity up to 42.92%. The experimental and theoretical findings elucidated that the integration of a single Mo atom within the COF enhances the adsorption and activation of CO, thereby promoting its subsequent hydrogenation reaction. Concurrently, the energy barriers for the production of C₂H₄ is lowered. This research provides a novel concept for the photochemical reduction of CO₂ to C2 derivatives utilizing metal single-atom modified COFs.

3.3.2 Adjustment the coordination environment. g-C₃N₄ is a semiconductor polymer composed of C and N elements that shows promise as a photocatalytic material due to its narrow bandgap and response to visible light. Theoretically, the energy band positions of VB and CB are about −1.1 eV and 1.6 eV (vs. NHE), making it an ideal photocatalyst. However, g-C₃N₄ suffers from issues such as serious charge carrier recombination, limited light absorption performance, and insufficient photocatalytic activity for practical applications. To enhance its photocatalytic activity, it is typically necessary to modify g-C₃N₄ by incorporating metal sites. As a 2D material with a layered stacking structure and π-conjugated system, g-C₃N₄ has many metal attachment sites within its layers, making it suitable for the preparation of SACs with high loading. Furthermore, aside from its capacity to stabilize metal atoms, nitrogen possesses the capability to form M–N_x, which can modify the coordination environment of the metal. This, in turn, leads to an enhancement in the adsorption of CO₂.

Sharma et al.⁷⁹ reported a simple synthesis strategy toward single-atom Ru catalyst entrapped in porous C₃N₄ (RuSA–C₃N₄). Remarkably, by employing this RuSA–C₃N₄ composite under ambient conditions, an efficient photocatalytic conversion of CO₂ into CH₃OH was achieved, boasting an impressively high yield of 1500 μmol g⁻¹. XPS measurements disclosed the formation of two distinct Ru moieties: Ru⁴⁺ and Ru²⁺, without any metallic Ru species (Fig. 12a). XANES and Raman analysis confirmed the bonding of Ru to N/C sites (Fig. 12b). These sites serve as a bridge, facilitating the electron transfer and increasing the charge density on Ru, thereby diminishing the barrier for photocarrier transfer. Additionally, the introduction of Ru single atoms extended the lifetime of charge carriers, with the average lifetime on RuSA–C₃N₄ being 7.6 ns compared to 7.1 ns onC₃N₄. This, in turn, facilitates the heightened photocatalytic efficacy of RuSA–C₃N₄ in the aqueous reduction of CO₂ to methanol.

Fig. 12 (a) Ru 3d high-resolution XPS spectra of RuSA–mC₃N₄. (b) Ru K edge XANES of RuSA–mC₃N₄ (black), RuO₂ standard (blue), RuCl₃ standard (pink) and Ru foil (red). Reproduced from ref. 79 with permission from John Wiley and Sons, copyright 2021. (c) Calculated free-energy diagram for CO₂ reduction to CO and CH₄ on the Er₁/CN-NT catalyst, as well as the adsorption configurations of key intermediates. The Er, C, H, O, and N atoms are given in yellow, brown, white, red, and blue, respectively. Reproduced from ref. 80 with permission from John Wiley and Sons, copyright 2020.

In addition, introducing rare-earth atoms into carbon–nitrogen materials can also improve its photocatalytic performance. Ji et al.⁸⁰ developed a Er₁–CN-NT catalyst by employing the atom-confinement and coordination (ACC) strategy, where Er single atoms are distributed uniformly on CN nanotubes. The catalyst exhibited superior activity in the photoreduction of CO₂ to CH₄ in pure water solutions. In situ XANES spectra of Er₁/CN-NT revealed that Er was oxidized to a higher valence state by CO₂, confirming that Er played a crucial role in the adsorption and conversion of CO₂. DFT calculations (Fig. 12c) indicated that CO₂ through the –COOH pathway generated CO and CH₄.

Zhao et al.⁸¹ incorporated Dy single atoms into the heterojunction of CdS and C₃N₄, which resulted in a six-fold increase in catalytic activity for the photoreduction of CO₂ to CH₄ (Fig. 13a). The rise in activity is primarily attributed to the function of Dy as a charge transfer bridge, which elevates the separation efficiency of electrons and holes. HAADF-STEM measurements revealed that Dy is dispersed on the catalyst surface in a monatomic form. Furthermore, XPS results for C 1s, N 1s, and Dy 4d indicated that the introduction of Dy leads to the disappearance of COOR and C–N=C species, while a new peak appears at 155.6 eV in the Dy 4d spectrum, suggesting coordination between Dy³⁺ and C and N (Fig. 13b). DFT calculations showed that the introduction of Dy reduces the bond length of Cd–N at the heterojunction interface, indicating a significant promoting effect of Dy single atom on the formation of the heterojunction structure.

Fig. 13 (a) The proposed mechanism diagram of pCO2RR. (b) Dy 4d XPS spectra of CdS. Reproduced from ref. 81 with permission from John Wiley and Sons, copyright 2021.

3.3.3 Optimization the interatomic distances. The synergistic effect primarily arises from the hybridization and coupling of electronic orbitals between metal atoms, as well as the specific spatial arrangement between two metal sites. When two isolated single atoms on the catalyst surface are in proximity, they interact strongly with each other, leading to the formation of a highly active bimetallic center that greatly enhances catalytic performance. This phenomenon is commonly referred to as the geometric effect of bimetallic active sites. The presence of neighbouring metal atoms can modify the adsorption state and energy of reaction intermediates, facilitating the activation of chemical bonds and subsequently influencing the reaction barrier.

Using UiO-66–NH₂ as support, Wang et al.⁸² developed a Cu SAC (Cu–SAs/UiO-66–NH₂) through a photoreduction method, which significantly promotes the photocatalytic reduction of CO₂. Importantly, the developed Cu–SAs/UiO-66–NH₂ facilitated the solar-driven CO₂ conversion to methanol and ethanol, with producing rates of 5.33 and 4.22 μmol h⁻¹ g⁻¹, respectively. These yields markedly surpass those of UiO-66–NH₂ and Cu nanoparticle counterpart. Theoretical calculations indicated that introducing Cu single atoms on UiO-66–NH₂ reduces the reaction barrier for the conversion of CO₂ into CHO* and CO* intermediates, thereby demonstrating excellent selectivity for methanol and ethanol.

Ma et al.⁸³ utilized a thermolysis-induced gasification approach to prepare a Co/g-C₃N₄ SAC with a significant metal loading of up to 24.6 wt%. The obtained Co SAC exhibited remarkable activity for the photo-reduction of CO₂ to methanol and dimethyl ether (Fig. 14a). HAADF-STEM and EDX mapping images unveiled the atomic dispersion of Co, and no Co–Co bonds were detected by the EXAFS measurements. The XPS results show that the electron density of carbon and nitrogen significantly increases with the increase in Co loading, effectively promoting the capture of photocarriers. Additionally, the generated Co–N₂C structure alters the band structure of g-C₃N₄. The valence band maximum (VBM) and conduction band minimum (CBM) of Co/g-C₃N₄ moved up 0.99 eV and 0.78 eV as compared to those of g-C₃N₄, respectively. Interestingly, this substantial change in the band gap can only be achieved when the density of single atom sites is sufficiently high. The formation of high-density Co–N₂C sites facilitates the transfer of photocarriers and simultaneously suppresses the recombination of electrons and holes.

Fig. 14 (a) H₂, CO, CH₄, C₂H₄, C₂H₆, C₃H₆, CH₃OCH₃ products at 4 h over g-C₃N₄, Co/g-C₃N₄-0.1 SAC, Co/g-C₃N₄-0.2 SAC, Co/g-C₃N₄-0.25, and CoO_x/g-C₃N₄-0.2. Reproduced from ref. 83 with permission from Elsevier, copyright 2022. (b) Photocatalytic activity of the different Cu loading catalysts. (c) In situ DRIFTS spectra of 3%Cu/WO₃. Reproduced from ref. 84 with permission from Elsevier, copyright 2023.

In addition, appropriate atomic density is vital for the generation of C2 products. In a study conducted by Zeng et al.,⁸⁴ it was found that two-dimensional hexagonal tungsten oxide (h′-H_xWO₃) with an appropriate loading of Cu atoms can promote the photocatalytic reduction of CO₂ to acetic acid (Fig. 14b). HAADF-STEM and EDX measurements confirmed the atomic dispersion of Cu. This inference is further supported by the linear adsorption of CO on Cu at 2120 cm⁻¹ (Fig. 14c). XPS results indicated that the oxidation state of Cu is +1, while electron transfer from W to Cu results in an increase in the oxidation state of W from +5 to +6. Compared to h′-H_xWO₃, 3%Cu/WO₃ with appropriate atomic spacing exhibited stronger adsorption capabilities for intermediate species (HCO₃⁻, COOH, CO), promoting the carbon–carbon coupling between COOH and CO to produce acetic acid. In addition to its impact on the key step of carbon–carbon coupling, the loading of Cu single atoms also plays an important role in the adsorption of CO₂ on the catalyst surface. CO₂ temperature-programmed desorption (TPD) results showed that three CO₂ adsorption configurations, including HCO₃⁻, bidentate carbonate (b-CO₃²⁻), and monodentate carbonate (m-CO₃²⁻), can form on h′-H_xWO₃. The introduction of an appropriate amount of Cu single atoms enhanced the binding abilities of all three adsorbed species, which explains the promoting effect of Cu single atoms on the catalytic performance.

4. Conclusions and perspectives

This article provides an overview of the pathways for achieving multi-electron transfer for CO₂ reduction by single-atom catalysis. In addition, detailed discussions are presented on the strategies for tuning the geometry and electronic structure of single-atom metal sites, as well as the relationship between local coordination structures and the catalytic activity. This study presents new insights into the atomic-level design of highly active SACs for CO₂ reduction. To accelerate research on efficient SACs for CO₂ reduction, we have identified the critical challenges that necessitate attention in this field.

1) The long-term stability of SACs

SACs have emerged as a promising avenue for CO₂ reduction owing to its precise coordination structure, complete atomic efficiency, and high catalytic activity. However, catalyst deactivation poses a major hurdle for its industrial implementation. At low metal loading, catalytic activity falls short of production requirements, while increasing loading leads to aggregation of active metal sites, even under mild electrochemical and photocatalytic conditions as discussed above. Therefore, to achieve stable production by single atom catalysis, stability under reaction conditions needs to be addressed. Stability can be improved by selecting suitable metal precursors, such as metal phthalocyanines, with N anchoring ability and steric hindrance effect that inhibits metal atom aggregation. Additionally, engineering metal oxide carriers with increased oxygen vacancy concentration or introducing N and S atoms for carbon carriers can enhance anchoring ability. Finally, confinement of metal sites using MOFs, molecular sieves, and hierarchical zeolites, combined with structural matching between carriers and metal precursors, can greatly improve catalyst stability.

2) In situ characterization techniques for single-atom catalysis

During the reaction process catalysts are typically not static. Regardless of the cause, such as adsorption and desorption of reaction molecules or induction of energy from the reaction environment, metal atoms on the catalyst's surface are subject to disturbance. For instance, copper phthalocyanine exists in a dispersed single-atom state before and after electrochemical CO₂ reduction, but in situ testing reveals the truth of Cu atom aggregation. This highlights the need for characterization of SACs, particularly under catalytic conditions. Although electron microscopy, synchrotron-based X-ray absorption spectroscopy, infrared spectroscopy, Raman spectroscopy, and XPS are currently employed techniques for observing catalysts during reaction, time and spatial resolution limitations prevent obtaining specific details such as oxidation state or complex coordination environments. Given that reactions occur at the molecular and even atomic level, featuring timescales of nanoseconds or even picoseconds, current methods prove insufficient for detecting.

3) Exploring multi-atomic catalysts

In traditional SACs, each metal center exists independently as individual active sites. This reduces the catalytic ability when multiple reactants and multiple-step reactions are involved. To overcome this limitation, dual-atom catalysts were developed by introducing metal atoms at adjacent sites. These catalysts possess two neighbouring metal sites that can work synergistically during complex reactions, while still maintaining the advantages of SACs. Furthermore, dual-atom catalysts can alter intermediate adsorption configurations to lower the elementary reaction energy barrier. In the presence of two adjacent active sites, the adsorption of intermediate species typically undergoes significant changes. Dual-atom catalysts can also facilitate the conversion of CO₂ to C2 products. Typically, SACs during the CO₂ hydrogenation process mainly produce C1 products because of the high dispersion of active sites. Dual-atom catalysts, however, can close the gap between active metals to allow for C–C coupling.

4) Novel synthetic methodologies for SACs

In recent years, numerous methodologies have been developed for the synthesis of SACs. However, the pursuit of novel synthetic strategies to achieve precise control over the coordination structure of active sites remains a pivotal task. For instance, although various methods have been reported for the preparation of carbon-based SACs, selective modulation of the active site structure through dopants such as P and N species still proves to be exceedingly challenging. Furthermore, due to the diversity of active sites in heterogeneous SACs, identifying true active sites and synthesizing SACs with identical structures pose significant challenges. Theoretically, this task can be solved via the efforts of combining advanced synthetic methodologies, characterization techniques, and DFT calculations. However, all the above tools currently present several shortcomings. For example, the current characterization techniques are even difficult to determine the atomic composition of the active site in SACs, let alone determine the number of coordinating atoms. As the design of SACs based on detailed synthetic understanding remains a long-term objective, it is essential to develop novel strategies that can circumvent these limitations for achieving visible breakthroughs in single-atom catalysis.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors gratefully acknowledge the CAS Project for Young Scientists in Basic Research (YSBR-022), the Strategic Priority Research Program of the Chinese Academy of Sciences (No. XDB36030200), the National Natural Science Foundation of China (21925803, U19A2015, 22008136).

References

1. M. Ronda-Lloret, G. Rothenberg and N. R. Shiju, ChemSusChem, 2019, 12, 3896–3914.
2. W. W. Ingwersen, A. S. Garmestani, M. A. Gonzalez and J. J. Templeton, Clean Technol. Environ. Policy, 2013, 16, 719–730.
3. A. V. Eliseev, M. M. Arzhanov, P. F. Demchenko and I. I. Mokhov, Izv., Acad. Sci., USSR, Atmos. Oceanic Phys., 2009, 45, 271–283.
4. Q. Liu, X. Yang, L. Li, S. Miao, Y. Li, Y. Li, X. Wang, Y. Huang and T. Zhang, Nat. Commun., 2017, 8, 1407.
5. W. Zhou, K. Cheng, J. Kang, C. Zhou, V. Subramanian, Q. Zhang and Y. Wang, Chem. Soc. Rev., 2019, 48, 3193–3228.
6. Q. Liu, J. Ma and C. Chen, Chin. J. Catal., 2022, 43, 898–912.
7. M. Aresta, A. Dibenedetto and A. Angelini, J. CO2 Util., 2013, 3–4, 65–73.
8. D. H. Nam, P. De Luna, A. Rosas-Hernandez, A. Thevenon, F. Li, T. Agapie, J. C. Peters, O. Shekhah, M. Eddaoudi and E. H. Sargent, Nat. Mater., 2020, 19, 266–276.
9. J. Zhang, G. Zeng, L. Chen, W. Lai, Y. Yuan, Y. Lu, C. Ma, W. Zhang and H. Huang, Nano Res., 2022, 15, 4014–4022.
10. A. Alvarez, M. Borges, J. J. Corral-Perez, J. G. Olcina, L. Hu, D. Cornu, R. Huang, D. Stoian and A. Urakawa, ChemPhysChem, 2017, 18, 3135–3141.
11. M. U. Khan, L. Wang, Z. Liu, Z. Gao, S. Wang, H. Li, W. Zhang, M. Wang, Z. Wang, C. Ma and J. Zeng, Angew. Chem., Int. Ed., 2016, 55, 9548–9552.
12. L. Zeng, Z. Wang, Y. Wang, J. Wang, Y. Guo, H. Hu, X. He, C. Wang and W. Lin, J. Am. Chem. Soc., 2019, 142, 75–79.
13. K. Chen, J. Yu, B. Liu, C. Si, H. Ban, W. Cai, C. Li, Z. Li and K. Fujimoto, J. Catal., 2019, 372, 163–173.
14. Y. Li, S. H. Chan and Q. Sun, Nanoscale, 2015, 7, 8663–8683.
15. J. Peng, S. Wang, H.-J. Yang, B. Ban, Z. Wei, L. Wang and B. Lei, Fuel, 2018, 224, 481–488.
16. J. Wang, Q. Hao, H. Zhong, K. Li and X. Zhang, Nano Res., 2022, 15, 5816–5823.
17. B. Qiao, A. Wang, X. Yang, L. F. Allard, Z. Jiang, Y. Cui, J. Liu, J. Li and T. Zhang, Nat. Chem., 2011, 3, 634–641.
18. J. T. Yates, S. D. Worley, T. M. Duncan and R. W. Vaughan, J. Chem. Phys., 1979, 70, 1225–1230.
19. A. S. S. Abbet, U. Heiz, W.-D. Schneider, A. M. Ferrari, G. Pacchioni and N. Rösch, J. Am. Chem. Soc., 2000, 122, 3453–3457.
20. C. Zhao, X. Dai, T. Yao, W. Chen, X. Wang, J. Wang, J. Yang, S. Wei, Y. Wu and Y. Li, J. Am. Chem. Soc., 2017, 139, 8078–8081.
21. L. Wang, D. Wang and Y. Li, Carbon Energy, 2022, 4, 1021–1079.
22. X. Duan, J. Xu, Z. Wei, J. Ma, S. Guo, S. Wang, H. Liu and S. Dou, Adv. Mater., 2017, 29, 1701784.
23. T. Zheng, K. Jiang and H. Wang, Adv. Mater., 2018, 30, 1802066.
24. S. Kattel, P. Liu and J. G. Chen, J. Am. Chem. Soc., 2017, 139, 9739–9754.
25. B. Yan, B. Zhao, S. Kattel, Q. Wu, S. Yao, D. Su and J. G. Chen, J. Catal., 2019, 374, 60–71.
26. Y. Kim, T. S. B. Trung, S. Yang, S. Kim and H. Lee, ACS Catal., 2016, 6, 1037–1044.
27. P. Panagiotopoulou and X. E. Verykios, J. Phys. Chem. C, 2017, 121, 5058–5068.
28. L. Liu, F. Fan, M. Bai, F. Xue, X. Ma, Z. Jiang and T. Fang, Mol. Catal., 2019, 466, 26–36.
29. C. Liu and P. Liu, ACS Catal., 2015, 5, 1004–1012.
30. Y. F. Zhao, Y. Yang, C. Mims, C. H. F. Peden, J. Li and D. H. Mei, J. Catal., 2011, 281, 199–211.
31. H. Kusama, K. Okabe, K. Sayama and H. Arakawa, Appl. Organomet. Chem., 2000, 14, 836–840.
32. D. Xu, Y. Wang, M. Ding, X. Hong, G. Liu and S. C. E. Tsang, Chem, 2021, 7, 849–881.
33. Y. Y. Birdja, E. Pérez-Gallent, M. C. Figueiredo, A. J. Göttle, F. Calle-Vallejo and M. T. M. Koper, Nat. Energy, 2019, 4, 732–745.
34. M. Zhou, W. Kong, M. Xue, H. Li, M. A. Khan, B. Liu, F. Lu and X. Zeng, Catal. Sci. Technol., 2023, 2044–4761.
35. T. Kong, Y. Jiang and Y. Xiong, Chem. Soc. Rev., 2020, 49, 6579–6591.
36. Z. Sun, H. Wang, Z. Wu and L. Wang, Catal. Today, 2018, 300, 160–172.
37. X. Chang, T. Wang and J. Gong, Energy Environ. Sci., 2016, 9, 2177–2196.
38. S. N. Habisreutinger, L. Schmidt-Mende and J. K. Stolarczyk, Angew. Chem., Int. Ed., 2013, 52, 7372–7408.
39. N. Cheng, L. Zhang, K. Doyle-Davis and X. Sun, Electrochem. Energy Rev., 2019, 2, 539–573.
40. S. Wang, L. Wang, D. Wang and Y. Li, Energy Environ. Sci., 2023, 16, 2759–2803.
41. J. H. Kwak, L. Kovarik and J. Szanyi, ACS Catal., 2013, 3, 2094–2100.
42. M. Fan, J. D. Jimenez, S. N. Shirodkar, J. Wu, S. Chen, L. Song, M. M. Royko, J. Zhang, H. Guo, J. Cui, K. Zuo, W. Wang, C. Zhang, F. Yuan, R. Vajtai, J. Qian, J. Yang, B. I. Yakobson, J. M. Tour, J. Lauterbach, D. Sun and P. M. Ajayan, ACS Catal., 2019, 9, 10077–10086.
43. C. Rivera-Cárcamo, C. Scarfiello, A. B. García, Y. Tison, H. Martinez, W. Baaziz, O. Ersen, C. Le Berre and P. Serp, Adv. Mater. Interfaces, 2020, 8, 2001777.
44. X. Shao, X. Yang, J. Xu, S. Liu, S. Miao, X. Liu, X. Su, H. Duan, Y. Huang and T. Zhang, Chem, 2019, 5, 693–705.
45. X. Ye, C. Yang, X. Pan, J. Ma, Y. Zhang, Y. Ren, X. Liu, L. Li and Y. Huang, J. Am. Chem. Soc., 2020, 142, 19001–19005.
46. Y. Chen, H. Li, W. Zhao, W. Zhang, J. Li, W. Li, X. Zheng, W. Yan, W. Zhang, J. Zhu, R. Si and J. Zeng, Nat. Commun., 2019, 10, 1885.
47. R. Büchel, A. Baiker and S. E. Pratsinis, Appl. Catal., A, 2014, 477, 93–101.
48. S. Li, Y. Xu, H. Wang, B. Teng, Q. Liu, Q. Li, L. Xu, X. Liu and J. Lu, Angew. Chem., 2023, 135, e202218167.
49. Q. Wang, S. Santos, C. A. Urbina-Blanco, W. Y. Hernández, M. Impéror-Clerc, E. I. Vovk, M. Marinova, O. Ersen, W. Baaziz, O. V. Safonova, A. Y. Khodakov, M. Saeys and V. V. Ordomsky, Appl. Catal., B, 2021, 290, 120036.
50. M.-M. Millet, G. Algara-Siller, S. Wrabetz, A. Mazheika, F. Girgsdies, D. Teschner, F. Seitz, A. Tarasov, S. V. Levchenko, R. Schlögl and E. Frei, J. Am. Chem. Soc., 2019, 141, 2451–2461.
51. H. Zhao, R. Yu, S. Ma, K. Xu, Y. Chen, K. Jiang, Y. Fang, C. Zhu, X. Liu, Y. Tang, L. Wu, Y. Wu, Q. Jiang, P. He, Z. Liu and L. Tan, Nat. Catal., 2022, 5, 818–831.
52. H. Li, L. Wang, Y. Dai, Z. Pu, Z. Lao, Y. Chen, M. Wang, X. Zheng, J. Zhu, W. Zhang, R. Si, C. Ma and J. Zeng, Nat. Nanotechnol., 2018, 13, 411–417.
53. T. Tang, Z. Wang and J. Guan, Adv. Funct. Mater., 2022, 32, 2111504.
54. J. Yang, W. Li, D. Wang and Y. Li, Adv. Mater., 2020, 32, 2003300.
55. X. Wang, X. Sang, C. L. Dong, S. Yao, L. Shuai, J. Lu, B. Yang, Z. Li, L. Lei, M. Qiu, L. Dai and Y. Hou, Angew. Chem., 2021, 60, 11959–11965.
56. K. Jiang, S. Siahrostami, T. Zheng, Y. Hu, S. Hwang, E. Stavitski, Y. Peng, J. Dynes, M. Gangisetty, D. Su, K. Attenkofer and H. Wang, Energy Environ. Sci., 2018, 11, 893–903.
57. H. Xu, D. Rebollar, H. He, L. Chong, Y. Liu, C. Liu, C.-J. Sun, T. Li, J. V. Muntean, R. E. Winans, D.-J. Liu and T. Xu, Nat. Energy, 2020, 5, 623–632.
58. C. Genovese, M. E. Schuster, E. K. Gibson, D. Gianolio, V. Posligua, R. Grau-Crespo, G. Cibin, P. P. Wells, D. Garai, V. Solokha, S. K. Calderon, J. J. Velasco-Velez, C. Ampelli, S. Perathoner, G. Held, G. Centi and R. Arrigo, Nat. Commun., 2018, 9, 935.
59. Z. Geng, X. Kong, W. Chen, H. Su, Y. Liu, F. Cai, G. Wang and J. Zeng, Angew. Chem., Int. Ed., 2018, 57, 6054–6059.
60. S. Chen, B. Wang, J. Zhu, L. Wang, H. Ou, Z. Zhang, X. Liang, L. Zheng, L. Zhou, Y. Q. Su, D. Wang and Y. Li, Nano Lett., 2021, 21, 7325–7331.
61. S. F. Hung, A. Xu, X. Wang, F. Li, S. H. Hsu, Y. Li, J. Wicks, E. G. Cervantes, A. S. Rasouli, Y. C. Li, M. Luo, D. H. Nam, N. Wang, T. Peng, Y. Yan, G. Lee and E. H. Sargent, Nat. Commun., 2022, 13, 819.
62. Y. Wang, Z. Chen, P. Han, Y. Du, Z. Gu, X. Xu and G. Zheng, ACS Catal., 2018, 8, 7113–7119.
63. J. Zhang, C. Zheng, M. Zhang, Y. Qiu, Q. Xu, W.-C. Cheong, W. Chen, L. Zheng, L. Gu, Z. Hu, D. Wang and Y. Li, Nano Res., 2020, 13, 3082–3087.
64. K. Zhao, X. Nie, H. Wang, S. Chen, X. Quan, H. Yu, W. Choi, G. Zhang, B. Kim and J. G. Chen, Nat. Commun., 2020, 11, 2455.
65. H. Yang, Y. Wu, G. Li, Q. Lin, Q. Hu, Q. Zhang, J. Liu and C. He, J. Am. Chem. Soc., 2019, 141, 12717–12723.
66. J. Gu, C. S. Hsu, L. Bai, H. M. Chen and X. Hu, Science, 2019, 364, 1091–1094.
67. Y. Wu, Z. Jiang, X. Lu, Y. Liang and H. Wang, Nature, 2019, 575, 639–642.
68. X. Liu, Y. Jiao, Y. Zheng, K. Davey and S.-Z. Qiao, J. Mater. Chem. A, 2019, 7, 3648–3654.
69. D. Karapinar, N. T. Huan, N. R. Sahraie, J. Li, D. Wakerley, N. Touati, S. Zanna, D. Taverna, L. H. G. Tizei, A. Zitolo, F. Jaouen, V. Mougel and M. Fontecave, Angew. Chem., Int. Ed., 2019, 58, 15098–15103.
70. A. Guan, Z. Chen, Y. Quan, C. Peng, Z. Wang, T.-K. Sham, C. Yang, Y. Ji, L. Qian, X. Xu and G. Zheng, ACS Energy Lett., 2020, 5, 1044–1053.
71. W. Xie, H. Li, G. Cui, J. Li, Y. Song, S. Li, X. Zhang, J. Y. Lee, M. Shao and M. Wei, Angew. Chem., 2021, 60, 7382–7388.
72. M. Tawalbeh, R. M. N. Javed, A. Al-Othman and F. Almomani, Energy Convers. Manage., 2023, 279, 116755.
73. Y. Wang, X. Cui, J. Zhang, J. Qiao, H. Huang, J. Shi and G. Wang, Prog. Mater. Sci., 2022, 128, 100964.
74. H. Chen, K. Shen, Q. Mao, J. Chen and Y. Li, ACS Catal., 2018, 8, 1417–1426.
75. X. Fang, Q. Shang, Y. Wang, L. Jiao, T. Yao, Y. Li, Q. Zhang, Y. Luo and H. L. Jiang, Adv. Mater., 2018, 30, 1705112.
76. D. Sun, Y. Gao, J. Fu, X. Zeng, Z. Chen and Z. Li, Chem. Commun., 2015, 51, 2645–2648.
77. H. Zhang, J. Wei, J. Dong, G. Liu, L. Shi, P. An, G. Zhao, J. Kong, X. Wang, X. Meng, J. Zhang and J. Ye, Angew. Chem., Int. Ed., 2016, 55, 14310–14314.
78. M. Kou, W. Liu, Y. Wang, J. Huang, Y. Chen, Y. Zhou, Y. Chen, M. Ma, K. Lei, H. Xie, P. K. Wong and L. Ye, Appl. Catal., B, 2021, 291, 120146.
79. P. Sharma, S. Kumar, O. Tomanec, M. Petr, J. Z. Chen, J. T. Miller, R. S. Varma, M. B. Gawande and R. Zboril, Small, 2021, 17, e2006478.
80. S. Ji, Y. Qu, T. Wang, Y. Chen, G. Wang, X. Li, J. Dong, Q. Chen, W. Zhang, Z. Zhang, S. Liang, R. Yu, Y. Wang, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2020, 59, 10651–10657.
81. Y. Zhao, Z. Han, G. Gao, W. Zhang, Y. Qu, H. Zhu, P. Zhu and G. Wang, Adv. Funct. Mater., 2021, 31, 2104976.
82. G. Wang, C. T. He, R. Huang, J. Mao, D. Wang and Y. Li, J. Am. Chem. Soc., 2020, 142, 19339–19345.
83. M. Ma, Z. Huang, D. E. Doronkin, W. Fa, Z. Rao, Y. Zou, R. Wang, Y. Zhong, Y. Cao, R. Zhang and Y. Zhou, Appl. Catal., B, 2022, 300, 120695.
84. D. Zeng, H. Wang, X. Zhu, H. Cao, Y. Zhou, W. Wang, L. Zhang and W. Wang, Chem. Eng. J., 2023, 451, 138801.

---