Structural engineering of atomic catalysts for electrocatalysis

Abstract

As a burgeoning category of heterogeneous catalysts, atomic catalysts have been extensively researched in the field of electrocatalysis. To satisfy different electrocatalytic reactions, single-atom catalysts (SACs), diatomic catalysts (DACs) and triatomic catalysts (TACs) have been successfully designed and synthesized, in which microenvironment structure regulation is the core to achieve high-efficiency catalytic activity and selectivity. In this review, the effect of the geometric and electronic structure of metal active centers on catalytic performance is systematically introduced, including substrates, central metal atoms, and the coordination environment. Then theoretical understanding of atomic catalysts for electrocatalysis is innovatively discussed, including synergistic effects, defect coupled spin state change and crystal field distortion spin state change. In addition, we propose the challenges to optimize atomic catalysts for electrocatalysis applications, including controlled synthesis, increasing the density of active sites, enhancing intrinsic activity, and improving the stability. Moreover, the structure–function relationships of atomic catalysts in the CO₂ reduction reaction, nitrogen reduction reaction, oxygen reduction reaction, hydrogen evolution reaction, and oxygen evolution reaction are highlighted. To facilitate the development of high-performance atomic catalysts, several technical challenges and research orientations are put forward.

---

1. Introduction

More than 80% of modern chemical industrial processes involve catalytic reactions, which have greatly promoted the development of modern industry.¹ Therein, the wide application of electrocatalysis technology reduces the industrial dependence on fossil fuels and alleviates the environmental crisis, for instance, CO₂ can be converted into chemicals with high added value through the electrocatalytic CO₂ reduction reaction (CO₂RR).^2,3 Fuel cells and water splitting are new energy conversion technologies, in which no carbon-containing compounds are released and the core reactions are the oxygen reduction reaction (ORR), oxygen evolution reaction (OER) and hydrogen evolution reaction (HER).^4–7 Ammonia synthesis by the Haber–Bosch process requires high temperature and high pressure, which can be achieved by the electrocatalytic nitrogen reduction reaction (NRR) at normal temperature and normal pressure. Moreover, the electrocatalytic NRR can be used to treat wastewater and reduce nitrogen pollution.⁸ However, the application of electrocatalysis to industry not only requires high catalytic activity and specific selectivity, but also needs to ensure economic benefit and high catalytic stability. The commonly used electrocatalysts are Pt, Au, Ru, Ir and other precious metals, but their small reserves and high price hinder the large-scale commercial applications.^3,9 The development of efficient and economical transition metal-based catalysts is expected to promote the development of industrial electrocatalysis.

By reducing the size of metal particles of metal-based catalysts from bulk to nanoscale, and further to the single-atom level, atomic dispersion of metal species is achieved. In 1998, Basset et al. modified Sn atoms around Pt atoms by surface organometallic chemistry (SOMC), showing nearly 100% selectivity for isobutane dehydrogenation.¹⁰ In addition, they used the Si–O bond on the surface of silicon oxide to fix single Zr atoms, which can be used for low-temperature hydrogenolysis of alkanes.¹¹ In 2011, the concept of “single atom catalysis” was first proposed by Zhang et al.¹² Single-atom catalysts (SACs) are a new catalytic family. The uniformly dispersed isolated metal atoms on the substrate have well-defined active centers and 100% atom utilization.¹³

As a new family of catalysts, isolated transition metal atoms in SACs anchored to specific substrates act as active sites. In addition to the central metal, the surrounding coordination environment affects the electronic structure and geometric configuration of single metal sites, and the catalytic activity of SACs is related to the strong force between metal centers and the substrate. Therefore, the electrocatalytic activity can be effectively improved by selecting appropriate central metal atoms and adjusting the local microenvironment of SACs.^14,15 Diatomic catalysts (DACs) have isolated pairs of metal atoms as the active sites, in which pairs of the same metal are called homonuclear DACs, while pairs of different metals are called heteronuclear DACs. In catalytic reactions, the metal pairs synergistically participate in the reactions. DACs are an extension of SACs, which not only have the advantages of SACs, but also introduce another degree of freedom for adsorption regulation of reaction intermediates, providing more possibilities for adsorption. At the same time, the interaction between bimetal sites can effectively regulate the activation energy and adsorption energy of reaction intermediates.¹⁶ By further adjusting the number of central metal atoms, isolated active sites with three or four metal atoms called triatomic catalysts (TACs) or tetratomic catalysts can be obtained. The main advantage of such catalysts is the synergistic effect between multi-nuclear metals, providing a variety of reaction intermediates and transition states in catalytic reactions, searching for the optimal reaction path and the fastest reaction rate, which are more prominent for some complex multi-electron transfer processes.^17,18 At present, there are relatively few reports about the application of multi-atomic catalysts in electrocatalysis, although the research on SACs and DACs has achieved great progress. The study of catalysts at an atomic level opens a door to explore the relationship between activity and structure.

With the deepening of research, the electrocatalytic activity of atomic catalysts can be improved by adjusting the coordination structure of metal sites. Müllen et al. synthesized a P and N double-doped Fe single-atom catalyst (P/Fe–N–C) by introducing P atoms into the second coordination layer of Fe–N₄, which showed OER/ORR bifunctional activity. Theoretical calculations showed that P doping can adjust the *OOH/*O adsorption energy at the Fe–N₄ site and accelerate the reactions.¹⁹ Liu et al. used a controlled “precursor preselection” strategy to synthesize a class of Cu and Co DACs with different coordination shells.²⁰ When Cu and Co metals bond to S, they have lower valence states, which are conducive to the ORR. However, when S is in the second shell and is coordinated with N, the metal atoms show an intermediate equivalent state, which is conducive to the OER. Additionally, when S is doped on the base surface it is far away from the metal, which is in a high valence state and is favourable for the HER.

In addition to SACs and DACs, the synthesis and characterization of TACs can be experimentally realized. However, TACs are still in the development stage, since the more central the metal atoms, the greater the probability of metal agglomeration, and the greater the challenge of synthesis. TACs have great potential to catalyse the conversion of CO₂ to C₂ and C₂₊ products due to multi-metal site synergistic catalysis. Using theoretical calculations, Dou et al. demonstrated that spatially confined three-centered metal atoms can simultaneously fix a few molecules of carbon dioxide and have multiple reaction paths, which are spatially and electronically conducive to the coupling of C–C and C₂–C to generate C₂ and C₃ products with high added-value.²¹ On M₃@NG, the reaction steps of CO₂ → *COOH → *CO → *CHO → *CH are similar to those on copper catalysts, and the speed control step is *CO → *CHO. Then *CH can be converted into *CH₂ or *CH–CO, and the C–C coupling process of the latter can be carried out on the middle spatially confined trinuclear metal to form C₂ or continue coupling to produce C₃. This provides a new idea for designing novel electrocatalysts for CO₂ reduction to C₂₊ products. Zhang et al. reported a trinuclear nickel catalyst (TNC-Ni), where the interaction among the three nickel sites plays a key role in charge accumulation and O–O bond formation during water oxidation.²²

About 297 million tons of nitrogen are annually converted by biological nitrogen fixation, in which dimeric enzyme complexes (FeMo, FeV, or FeFe active sites) play a catalytic role, effectively reducing the reaction energy barrier. However, the slow kinetics limits high yields, and the stability and recyclability of catalysts are challenging.^23–26 Inspired by natural nitrogen fixation, Ma et al. proposed a heteronuclear FeV@C₂N diatomic electrocatalyst, which effectively inhibited competitive HER and had good NRR activity and selectivity.²⁷ Dey et al. reported the synthesis of a trinuclear Ni complex (Ni₃S₈) using PhOH as a proton source, which selectively reduces N₂ to N₂H₄.²⁸

In this review, we summarize the recent progress of atomic catalysts in the electrocatalysis of the HER, OER, ORR, CO₂RR and NRR, and a main development timeline of SACs, DACs and TACs is shown in Fig. 1. The regulation strategies of metal active sites are discussed, including reasonable design of substrates, regulating the electronic structure of central metal atoms and the coordination environment. Furthermore, the influencing factors of atomic catalysts for electrocatalysis are theoretically analysed and the strategy of optimizing the atomic catalyst is proposed. Finally, we put forward the current challenges and opportunities for future development of atomic catalysts to guide the reasonable design of multi-atomic catalysts in the future.

Fig. 1 History of catalysts with different nuclear numbers for electrocatalysis. Pt₁/FeO_x.¹² Pt–MoS₂. Reproduced with permission.²⁹ Copyright 2015, The Royal Society of Chemistry. FeTPyP-Co. Reproduced with permission.³⁰ Copyright 2016, American Chemical Society. Ru₃/CN. Reproduced with permission.³¹ Copyright 2017, American Chemical Society. Co–N₅/HNPCSs. Reproduced with permission.³² Copyright 2018, American Chemical Society. Ni/Fe–N–C. Reproduced with permission.³³ Copyright 2019, Wiley-VCH. Se/Te-doped C catalysts. Reproduced with permission.³⁴ Copyright 2020, Wiley-VCH. TNC-Cu. Reproduced with permission.³⁵ Copyright 2021, American Chemical Society. RuCoN_x/C. Reproduced with permission.³⁶ Copyright 2022, American Chemical Society. Homonuclear and heteronuclear DACs. Reproduced with permission.³⁷ Copyright 2023, American Chemical Society. CoMn–N/C. Reproduced with permission.³⁸ Copyright 2023, Wiley-VCH GmbH.

2. Reasonable control of the geometric and electronic structure of metal active sites

Geometric configuration and electronic structure are crucial for metal active sites towards advanced electrocatalysis.³⁹ The geometric configuration reflects the coordination environment between the active metal and the surrounding atoms, and the electronic structure of the active site is closely related to the coordination environment. When the distance between two metal atoms is less than 1.0 nm, the two metal atoms will interact and affect the distribution of electrons.⁴⁰ The selection and modification of the substrate also play an important role in the geometric configuration and electronic structure of central metal atoms, and the selection of special position on the substrate is beneficial to the formation of a unique electronic structure. The active center is composed of metal and coordination atoms, and the interaction between metal and metal/non-metal will cause charge rearrangement, and therefore, reasonable selection of the number and type of metal is very important to design catalysts with different electrocatalytic activities.⁴¹ In this section, we first introduce the rational design of atomic catalysts with different numbers of metal nuclei through defect engineering and spatial limiting, and focus on how to regulate the central metal atoms and the coordination environment.

2.1. Reasonable design of substrates

2.1.1. Defect engineering. Defect engineering can be adopted to regulate the electronic structure and surface morphology of catalysts by controlling the type, concentration and distribution of defect sites, so that the catalysts have unique physical and chemical properties to improve the catalytic activity. The defects in the substrate carbon materials mainly include intrinsic defects and externally induced defects. The intrinsic defects generally include pore, edge, vacancy and topological defects. The defects induced by external factors, such as N, P, S, B and F nonmetallic heteroatom doping and Fe, Co, Ni, Cu and Zn metal atom doping, are caused by the destruction of the original sp² carbon lattice of the carbon material by the bonding of carbon atoms with non-carbon atoms.^42–44 Yao et al. prepared a highly stable atomically dispersed Ni catalyst using defective graphene (A-Ni@DG) for the electrocatalytic OER and HER.⁴⁵ X-ray absorption spectroscopy (XAS) characterization and density functional theory (DFT) calculations showed that different defects within graphene can induce different local electron state densities of Ni species (Fig. 2).

Fig. 2 (a–c) High angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) images of A-Ni@DG. (d–f) Illustrations of three different types of catalytic active sites corresponding to a single Ni atom supported on (d) perfect hexagons, (e) D5775, and (f) di-vacancy. (g and h) Energy profiles of three configurations (d–f) for the HER (g) and OER (h). Reproduced with permission.⁴⁵ Copyright 2018, Elsevier.

Xu et al. reported a carbon-supported defect-anchored Pt SAC.⁴⁶ The ORR on the Pt SAC is a four-electron reaction with a high half-wave potential (0.835 V vs. RHE). Theoretical calculations showed that the Pt–C₄ site with two carbon vacancies is the active center of the ORR. Wang et al. constructed defect sites on TiO₂ nanosheets to stabilize single-atom Au sites, which can improve catalytic activity by reducing the energy barrier and competitive adsorption at isolated Au atomic sites.⁴⁷ Chen et al. prepared a catalyst with a copper atomic pair (Cu⁰_1–Cu₁^x+) that was stabilized by Te surface defects on Pd₁₀Te₃ alloy nanowires. High FE (FE_CO = 92%) was shown for the CO₂RR, and the HER was almost completely suppressed.⁴⁸ Theoretical calculations showed that the Cu⁰_1–Cu₁^x+ site can promote CO₂ activation by the “diatomic activation bimolecular mechanism”. Yao et al. reported a locally distributed atomic Pt–Co nitrocarbon-based catalyst (A-CoPt-NC) with 267 times higher mass ORR activity than commercial Pt/C.⁴⁹ DFT calculations showed that the high ORR activity is due to the synergistic interaction between Pt–Co atom pairs and the surrounding defective coordination environment, and the electron distribution of the Pt–Co active center is asymmetrical.

The introduction of nonmetallic heteroatoms such as P or S into the carbon carrier will cause the change of charge density of active sites due to the mismatch between the heteroatoms and the outer orbitals of carbon, thus regulating the electrocatalytic performance.^50,51 Deng et al. synthesized an N,P codoped FeNi-NPC catalyst with diatomic Fe–Ni sites, where P does not coordinate with the diatomic Fe–Ni site.⁵² By adjusting the microenvironment and designing the structure, the catalytic OER/ORR activities of FeNi-NPC were effectively improved. Zou et al. synthesized a S-doped Mo single atom/carbon multichannel fiber catalyst (MoSA/CMF-S) using template assisted electrospinning.⁵³ The catalyst catalyzed the NRR to generate NH₃ with a yield of 46.6 μg h⁻¹ mg_cat⁻¹ and FE of 28.9% at −0.2 V. Theoretical calculations showed that the catalytic active site of MoSA/CMF-S is Mo₁N₁C₂–S4, where S does not coordinate with a Mo atom. Bader analysis showed that S doping makes the charge of the Mo atom rearrange. By comparing the charge of Mo in Mo₁N₁C₂–S4 and Mo₁N₁C₂, the charge of the Mo atom in the former increases more than that in the latter, and a charge accumulates between S and Mo in the Mo₁N₁C₂–S4. In addition, HER kinetics was slow at the Mo₁N₁C₂–S4, indicating that S doping effectively inhibited the HER. The adsorption energy of N₂ on the side-on (–0.87 eV) and end-on (–0.82 eV) orientations of a single Mo atom at the active site was calculated. The adsorption energy of N₂ at the Mo₁N₁C₂–S4 site was −0.88 eV, which is stronger than that at Mo₁N₁C₂ (−0.79 eV). In addition, the speed control step at the Mo₁N₁C₂–S4 (0.58 eV) is easier to overcome than that at Mo₁N₁C₂ (0.81 eV) (Fig. 3). In short, the introduction of S rearranges the electrons of single-atom Mo, which is conducive to the adsorption of N₂ and effectively inhibits HER activity.

Fig. 3 (a) Bader charge variations of the Mo atom in different S-doped samples. (b) Optimized geometries and contour plots of the charge density differences. (c) Gibbs free energy diagram of enzymatic NRR. Reproduced with permission.⁵³ Copyright 2022, Elsevier B.V.

2.1.2. Spatial confinement. Atomic confinement can be divided into chamber confinement and grid confinement. Chamber confinement is generally achieved by confining isolated transition metal precursors in a stable structure, such as the encapsulation of cobalt acetyl acetone in a chamber of imidazole molecular sieve framework, which requires the size of metal precursors and the cavity.⁵⁴ For grid confinement, metal atoms are stabilized by forming strong coordination and covalent bonds, such as metal porphyrins and metal phthalocyanines in covalent organic frameworks (COFs). It is needed to ensure that the small molecules have symmetry in order to obtain a periodic framework.^55,56 By separating and encapsulating suitable mononuclear metal precursors in porous materials such as zeolites, metal organic frameworks (MOFs), and COFs, the finite dispersion of atoms is achieved.^57–60

Li et al. used cage encapsulation and pyrolysis strategies to obtain a single-atom iron catalyst with good ORR activity, exhibiting a half-wave potential of 0.9 V in 0.1 M KOH and good stability.⁶¹ DFT calculations showed that the good ORR activity is attributable to the formation of single-atom Fe, which coordinates with four N atoms and is axially adsorbed with O₂ (Fig. 4a). Li et al. packaged a W precursor into a MOF and pyrolyzed it to obtain a W single-atom catalyst with a local structure of W₁N₁C₃.⁶² The catalyst showed good HER activity with an overpotential of 85 mV at 10 mA cm⁻² (η₁₀) in 0.1 M KOH. Huang et al. prepared an ultra-stable bimetallic phthalocyanine COF (Fig. 4b).⁶³ The overlap stacking mode of metal phthalocyanine units is conducive to electron transfer and improves the FE of CO₂ conversion to CO (FE = 97%). Niu et al. used encapsulation and pyrolysis strategies to construct a series of diatomic catalysts, including homo-nuclear (Fe₂, Co₂, Ni₂, Cu₂, Mn₂ and Pd₂) and hetero-nuclear (Fe–Cu, Fe–Ni, Cu–Mn and Cu–Co) DACs.³⁷ After pyrolysis, the molecular skeleton of macrocyclic complexes can be preserved (Fig. 4c). Taking Fe–Cu DAC as an example, it showed good ORR activity (E_1/2 = 0.793 V) and stability. Besides the above two atomic confinement methods, Zhang et al. designed a click-limiting strategy to disperse cobalt-coordinated porphyrin units onto the substrate to obtain a Co–N–C electrocatalyst, which showed excellent OER/ORR activities (ΔE = 0.79 V) (Fig. 4d–f).⁶⁴

Fig. 4 (a) Schematic illustration of the formation of Fe-ISAs/CN. (b) Top views of a graphical representation of the eclipsed stacking of CuPcF₈-CoNPc-COF. (c) Schematic illustration of the synthesis of DACs via a macrocyclic precursor-mediated encapsulation–pyrolysis process. (d–f) Different confinement strategies for fabricating M–N–C SACs. (a) Reproduced with permission.⁶¹ Copyright 2017, Wiley-VCH. (b) Reproduced with permission.⁶³ Copyright 2021, American Chemical Society. (c) Reproduced with permission.³⁷ Copyright 2023, American Chemical Society. (d–f) Reproduced with permission.⁶⁴ Copyright 2022, American Association for the Advancement of Science.

2.2. Regulating the electronic structure of central metal atoms

The electronic structure of central metal atoms plays an important role in optimizing the physical and chemical properties of catalysts. The central metal atom determines the energy level of the d orbital, the number of d electrons and the electron configuration. The distance between metal centers and the combination of metal centers are important factors affecting chemical reactions.^65–67 Single, double and triple metal atoms with isolation are obtained by adjusting the distance of metal centers, namely SACs, DACs and TACs, respectively (Fig. 5). The charge distribution between the active sites is optimized by adjusting metal combination, thereby changing the electronic properties of the metal atoms. The electronic structure of the central metal atom determines the atomic charge distribution and Fermi level, thus influencing the adsorption/desorption energy of reaction intermediates.⁶⁸ In this section, we will discuss how to adjust the electronic structure of the central metal atoms of SACs, DACs and TACs.

Fig. 5 (a–c) Typical structures of SACs, DACs and TACs.

2.2.1. Regulating the electronic structure of SACs. The development of SACs has opened a new door for heterogeneous catalysis. SACs have single isolated active sites, high metal utilization and unique electronic structure. The optimized electronic structure of the active center and the surrounding unsaturated coordination environment can increase electrocatalytic performance. SACs have become an ideal system for the study of catalytic mechanisms by combining theory and experiment, further guiding the design of new catalysts.

Hydrogen production by electrolysis of water is a green and sustainable method, which is expected to replace fossil fuels in the future. Precious metal platinum-, palladium-, and iridium-based catalysts play an important role in water electrolysis. However, Pt, Pd and Ir have small reserves and high costs, which are obstacles in large-scale electrocatalytic hydrogen production.⁶⁹ At present, ideal electrocatalysts can be developed by two strategies, one is to prepare SACs with fewer precious metals, and the other is to develop non-precious metal catalysts.^70,71 In the HER, there are three basic steps: Tafel, Heyrovsky, and Volmer, where the Volmer step is the fastest and Tafel is the rate-determining step. The formation and desorption processes of the reaction intermediate H* affect the reaction mechanism and reaction rate. According to Sabatier's principle, when the hydrogen adsorption free energy G_H* is close to 0, the HER is the most favourable to occur. Under alkaline conditions, the activity of the HER is also affected by the kinetic energy barrier of water splitting and the energy barrier of hydroxyl desorption.⁷²

For the HER, Pt-based catalysts are the most active electrocatalysts. Lee et al. loaded Pt onto low-oxide tungsten (Pt SA/WO_3−x), which greatly improved the utilization of Pt, and demonstrated that Pt SA/WO_3−x was significantly superior to WO_3−x loaded with Pt nanoparticles (Pt NP/WO_3−x).⁷³ At an overpotential of 50 mV, the mass activity of Pt SA/WO_3−x was 10.7 times that of Pt NP/WO_3−x. In practical applications, pH has a great effect on HER activity. Single-atom Pt catalysts show good performance under different pH values. Tan et al. modified Pt atoms into nanoporous Co_0.85Se (Pt/np-Co_0.85Se), and the catalyst showed excellent HER activity under neutral conditions, with an initial potential close to 0 and a Tafel slope of 35 mV dec⁻¹.⁷⁴ Operando XAS and theoretical calculations showed that Pt in Pt/np-Co_0.85Se can not only regulate the surface state of the active center, but also decrease the ΔG of the HER, optimize the hydrogen adsorption/desorption energy, and promote thermodynamics and kinetics.

Compared with acidic conditions, the HER rate and apparent activity of catalysts is slower under alkaline conditions, and thus it is relatively difficult to design atomic catalysts that can catalyse the HER under alkaline conditions. The π-electron hybridization at the edge of the carrier graphene is used to influence the electronic state of the Pt 5d orbital. Amal et al. used edge-rich vertical graphene to anchor Pt, and Pt coordinated with the graphene edge showed excellent HER activity with a turnover frequency (TOF) up to 22.6 s⁻¹ at η = 150 mV.⁷⁵

The anchoring of single-atom platinum to a suitable carrier can facilitate the HER due to the optimization of metal–carrier interaction and water adsorption. Jiang et al. anchored single-atom Pt to α-MoC_1−x@C (Pt_SA/α-MoC_1−x@C), which showed high activity and stability for the pH-universal HER.⁷⁶ The η₁₀ of Pt_SA/α-MoC_1−x@C under alkaline and acidic conditions is 21 mV and 12 mV, respectively. X-ray photoelectron spectrometry (XPS) and XAS analysis indicated that Pt in Pt_SA/α-MoC_1−x@C is present in a low state, which is caused by electron migration from Mo to Pt. Experimental results showed that platinum atoms are monodispersed in the α-MoC_1−x@C lattice. DFT analysis showed that the ΔG of hydrogen adsorption decreases significantly (0.60 eV → 0.21 eV) with the construction of the Pt unit point under acidic conditions, which is favourable for hydrogen adsorption/desorption. Under neutral and alkaline conditions, α-MoC_1−x-Pt is beneficial to water adsorption, promotes the capture of reactants, and accelerates the HER. In addition, the hydroxide adsorption free energy is reduced on α-MoC_1−x–Pt, and the interaction between α-MoC_1−x and Pt atoms optimizes the adsorption/desorption energy of reaction intermediates. Thus, Pt_SA/α-MoC_1−x@C showed pH-universal HER activity (Fig. 6).

Fig. 6 (a) Pt 4f XPS spectra. (b) Pt L₃-edge XANES spectra. (c) Fourier transform X-ray absorption fine structure (FT-EXAFS) curves. (d) Mo K-edge X-ray absorption near edge structure (XANES) spectra. (e) FT-EXAFS curves of the Mo K-edge. (f) Differential charge density for Pt₁/α-MoC_1−x. (g) Optimized adsorption geometries. (h) Gibbs free energy diagram. (i) Calculated adsorption free energies. Reproduced with permission.⁷⁶ Copyright 2021, Wiley-VCH GmbH, Weinheim.

In addition to the precious metal Pt, transition metal cobalt and nickel-based catalysts are used to catalyse the HER in acidic environments.^77–80 Studies have shown that transition metal carbides have a low energy barrier in hydrogen adsorption, and the density of states near the Fermi level is similar to that of precious metals. Nickel carbide (NiC) shows stability in a wide pH range. By introducing N into NiC, the electronic structure of Ni can be adjusted to improve the HER activity. In addition, the conductivity of the catalyst can be improved by reducing its size. Sun et al. prepared hybrid Ni–C–N nanosheets with thickness smaller than 2 nm.⁸¹ Ni–C–N is chemically stable and has metal conductivity, which showed good HER activity in 0.5 M H₂SO₄, 1.0 M KOH and 1.0 M PBS (pH = 7). Yu et al. prepared nitrogen-doped porous nanofibers with Fe–N₅ sites (Fe SA/PNCNFs-0.1) by a wet impregnation method.⁷⁹ The catalyst showed a high catalytic HER activity with an η₁₀ of 44.3 mV and Tafel slope of 45.4 mV dec⁻¹. However, there are still some issues in transition metal catalysts catalysing the HER at full pH, because the process will cause the change of proton concentration. To make water splitting more energy-saving, the ideal state should show excellent activity at any pH.

The OER is a complex four-electron process whose reactivity determines the efficiency of hydrogen production by water electrolysis.⁸² At present, the most advanced OER electrocatalysts are IrO₂ and RuO₂, but they have poor stability in alkaline media.⁸³ SACs have attracted much attention because of their unique electronic structure.⁸⁴ By anchoring precious metals to supports, SACs can be prepared to reduce the use of precious metals while ensuring high OER activity. Huang et al. prepared single-atom Rh-doped CuO nanowires on a copper foam (Rh SAC-CuO NAs/CF) using a cation exchange strategy, which exhibited an η₁₀ of 357 mV for the OER.⁸⁵ Theoretical calculations showed that the introduction of Rh atoms into the CuO lattice can effectively reduce the reaction energy barrier and accelerate the kinetics. Gu et al. synthesized a single-atom iridium-doped nickel phosphide catalyst (Ir_SA-Ni₂P) using a top-down phosphating strategy, which showed excellent OER activity with an η₁₀ of 149 mV.⁸⁶ Simulated results of Ir_SA-Ni₁₂P₅ and Ir_SA-Ni₂P nanospheres showed that iridium atoms replace the outermost nickel and occupy the Ni sites in the Ni_xP_y (Ni₂P and Ni₁₂P₅) matrix (Fig. 7).

Fig. 7 (a and d) HAADF-STEM images and corresponding FFT (inset). (b and e) Experimental and simulated HAADF-STEM images. (c and f) Intensity profile of the yellow line in (b and e) with the matrix Ni column and the decorated Ir atom. (g) Surface Pourbaix diagram of a single Ir atom doped Ni₂P (0001) surface (Ir_SA-Ni₂P (0001)). (h) Projected crystal orbital Hamilton populations (COHP) between chemisorbed *O and active metal centers on Ni₂P (0001) and Ir_SA-Ni₂P (0001). (i) The optimized OER intermediates on Ir_SA-Ni₂P (0001). (j) Free energy diagrams. Reproduced with permission.⁸⁶ Copyright 2021, American Chemical Society.

In addition, the abundant and inexpensive transition metals Mn, Fe, Co and Ni are often used as active centers of SACs for the OER.⁸⁷ Kibria et al. synthesized a cobalt-anchoring nitrogen-rich carbon single-atom catalyst (Co SAC) with a Co loading content of 10.6 wt% using a macromolecular assisted strategy.⁸⁸ The Co SAC had an OER η₁₀ of 351 mV under alkaline conditions and a mass activity of 2209 mA at 1.65 V, and showed tolerance for up to 300 h. According to operando XAS analysis, the pyridine nitrogen in a cobalt phthalocyanine tetramer (CoPc) with melem (CoMM) was suitable for the OER and had good reproducibility. Close Co–Co atomic spacing (∼3.85 A) and e_g packing can increase OER activity. In addition, theoretical calculations showed that electron-deficient Co–O coordination intermediates are easily formed to accelerate the reaction kinetics, and the Co–N₄ coordination structure can effectively reduce the reaction energy barrier and increase the OER activity (Fig. 8). According to the Sabatier principle, a good catalyst should optimize the adsorption energy of reaction intermediates. The OER involves adsorption and desorption of three reaction intermediates (*OH, *O and *OOH). Therefore, for a catalyst with high OER activity, the binding strength of the three intermediates to the active site should be moderate. Therefore, regulating the electronic structure and coordination environment of the active sites plays an important role in improving OER activity.^89–91

Fig. 8 (a) Schematics of operando XAS analysis of CoMM. (b) Chemical state change of CoMM during the OER. (c) OER free energy profile of pyridinic-nitrogen–cobalt and pyrrolic-nitrogen–cobalt single-atom models. (d) Projected density of states (PDOS) on the Co atom for both pyridinic-nitrogen–cobalt and pyrrolic-nitrogen–cobalt single-atom models. (e) Top view and (f) side view of *OH. (g) Top view and (h) side view of *O. Reproduced with permission.⁸⁸ Copyright 2023, American Chemical Society.

The ORR plays an important role in energy storage and chemical production, such as in fuel cells, metal air batteries and hydrogen peroxide synthesis. However, these devices are faced with the problems of low ORR efficiency and slow dynamics.^92–96 According to the volcanic curve (Fig. 9a), metals with slightly lower oxygen binding energy than platinum have higher oxygen reduction rate, and transition metals Fe, Co, Ni and Cu exhibit weaker binding energy of reaction intermediates. Anchoring them in the form of single atoms into conductive supports can promote the inherent ORR activity.^97,98 Zhang et al. synthesized an iron-doped mesoporous nitrogen-doped carbon material (Fe SAC-MIL101-T).⁹⁹ When the pyrolysis temperature was 1000 °C, the ORR activity was the best, and the half-wave potential was 0.94 V. The excellent ORR activity was attributed to the high dispersion of FeN_x active centers and abundant pore structure, which was conducive to mass transfer and improved the utilization rate of FeN_x active sites.

Fig. 9 (a) Trends in oxygen reduction activity plotted as a function of the oxygen binding energy. (b) The normalized K-edge XANES spectra and (c) FT-EXAFS spectra. (d) FT-EXAFS fitting curve of Co–N–C in R space and (e) wavelet transform (WT) for the k³-weighted EXAFS signals. (f) ΔG diagram. (g) Kinetic energy barriers. (h) ΔG diagram of the two-electron transfer ORR pathway. (i) PDOS in Co–N₅ near the Fermi level (E_f). (j) PDOS in Co–N₄ near E_f. Reproduced with permission.⁹⁷ (a) Copyright 2004, American Chemical Society. (b–j) Reproduced with permission.¹⁰⁵ Copyright 2021, The Royal Society of Chemistry.

The strong binding of Co-based SACs with oxygen can promote a two-electron process in the ORR, which is conducive to the generation of H₂O₂. Co-based SACs with Co–N₄ structure like cobalt phthalocyanine and cobalt porphyrin are conducive to the 4e⁻ transfer process. However, when the cobalt is coordinated with five nitrogen atoms (Co–N₅), it shows high selectivity of H₂O₂.^100–104 Shao et al. used a coordination engineering strategy to prepare a Co single-atom catalyst with Co–N₅ sites (Co SAC), which had a good selectivity for H₂O₂ in the ORR.¹⁰⁵ The Co SAC demonstrated the two-electron process of the ORR in seawater with fast kinetics and chloride ion tolerance. The FE of H₂O₂ in 0.5 M NaCl solution with pH = 7 is as high as 95.6%. Theoretical calculation results showed that the ORR barrier of the two-electron pathway on the Co–N₅ is lower than that of the four-electron transfer process, and the Co–N₅ is more conducive to the activation and reduction of *O₂ than Co–N₄ (Fig. 9b–j).

In the CO₂RR, most metal-based SACs can reduce CO₂ to CO, and Sn- and Bi-based SACs can convert CO₂ to HCOOH, while Cu-based catalysts with moderate adsorption strength of CO₂ can reduce CO₂ to C₁, C₂ and C₂₊ products.^106–113 Cu is the only metal reported to be able to convert CO₂ into C₂ and C₂₊ products with high added value.^14,114 Chen et al. studied a series of TM-phthalocyanine (TM-Pcs, TM = Fe, Co, Ni, Cu, and Zn) crystal molecular catalysts with TM-N₄ active sites for the CO₂RR and revealed that the binding strength of the metal to reaction intermediates determines the activity of the CO₂RR.¹¹⁵

Shen et al. synthesized a series of M–N–C SACs (M = Fe, CO, Ni and Cu) by a leaching method.¹¹⁶ The FE of CO on Ni–N–C reached up to 97%, and the activity of the CO₂RR was affected by the d-orbital electron configuration of the central metal ion. As shown in Fig. 10a–h, during the activation process, CO₂ interacts with the active site, and electrons are transferred from CO₂ to the active center metal, and then fed back to the π orbital of CO₂. Ni–N–C has higher electron density around Ni than other M–N–C because the outermost d orbital of Ni²⁺ is empty, which is conducive to the transfer of electrons from the C atom in the CO₂ molecule to the outermost orbital of the Ni atom and promotes the progress of the CO₂RR. Theoretical calculations showed that the rate-determining step of Ni–N–C is the formation of the *COOH intermediate, and the energy barrier of the rate-determining step on Fe–N–C, Co–N–C and Cu–N–C in the CO₂RR is greater than that in the HER, which is not conducive to the progress of the CO₂RR (Fig. 10i and j). Crystal field theory showed that the outermost space orbital of Ni²⁺ in the Ni–N–C is beneficial to the transfer of electrons from CO₂ to the active site Ni, effectively activating CO₂ molecules adsorbed on the surface.

Fig. 10 Schematics of d-orbital electronic configurations of central metals in (a) Fe–N–C, (b) Co–N–C, (c) Ni–N–C, and (d) Cu–N–C and their electronic interaction with CO₂ molecules. Side-view differential charge density maps of (e) Fe–N–C, (f) Co–N–C, (g) Ni–N–C, and (h) Cu–N–C. The blue and red regions represent net electron accumulation and depletion, respectively. Reaction pathways and free energy diagrams of (i) the CO₂RR and (j) HER over M–N–C SACs at 0 V versus RHE. Reproduced with permission.¹¹⁶ Copyright 2023, American Chemical Society.

Cu-based catalysts can induce CO* dimerization in the CO₂RR to obtain C₂ and C₂₊ products. The activity and selectivity can be improved by modifying the surface/interface of catalysts, such as adjusting the morphology, ligand modification, vacancy fabrication and doping. Han et al. applied a site-specific underpotential electrochemical deposition (UPD) strategy to treat single-atom Cu-anchoring Ag₂S/Ag nanowires.¹¹⁷ Cu/Ag(S) was synthesized through single atom Cu occupying the defect position after sulfur removal. The FE of C₂₊ on Cu/Ag(S) is significantly higher than that on the Ag(S) catalyst without single-atom Cu, and Cu/Ag(S) showed the maximum FE (17%) at 1.4 V. Theoretical calculations showed that the interaction between Cu and Ag on the surface defects causes the d-band center to shift down, reduces the energy barrier of *CO formation (∼0.5 eV), promotes the adsorption of reaction intermediates, and improves the selectivity of C₂₊.

Ammonia (NH₃) is an important energy carrier and an important raw material for the synthesis of agricultural fertilizers. At present, NH₃ is industrially produced by the Haber–Bosch method, which needs high temperature and pressure. The conditions are harsh and the energy input is not in line with the trend of green and low-carbon development. The electrocatalytic NRR has attracted much attention because of its low energy consumption and mild reaction conditions. However, the NRR is a six-electron-proton transfer process with complex intermediates and slow kinetics. In addition, side reactions can occur through a two-electron pathway to diazene (N₂H₂) and four-electron pathway to hydrazine (N₂H₄). Therefore, in recent years, the development of SACs with high activity and selectivity for the electrocatalytic NRR has been on the rise.^118,119

Since the active metal center of the natural nitrogenase is Mo, it has been inferred that single-atom Mo catalysts might be efficient for the NRR. Xin et al. synthesized a single-atom Mo-anchoring nitrogen-doped porous carbon material (SA-Mo/NPC), showing an FE_NH3 of 14.6 ± 1.6%.¹²⁰ NH₃ yields reached up to 34.0 ± 3.6 μg h⁻¹ mg_cat.⁻¹, and SA-Mo/NPC was well tolerated in the NRR (Fig. 11a–g). Zhang et al. simulated a monatomic Mo-anchoring C₉N₄ catalyst (Mo@C₉N₄) for the NRR.¹²¹ The results showed that the Mo atom is the active center of the NRR, and N₂ can be efficiently converted to NH₃ on the surface of Mo@C₉N₄. By comparing the remote and alternate pathways, the energy barrier (0.40 V) of the rate control step of *N₂ → *N₂H in the remote pathway is lower than that (0.70 V) of the rate control step of *NNH → *NHNH in the alternate pathway. Through the analysis of the bond length and charge changes in each step of the NRR, it was concluded that single-atom Mo serves as both the reservoir and the electron transfer transmitter between the intermediate *N_xH_y and the carrier C₉N₄ in the NRR process (Fig. 11h–k).

Fig. 11 (a) Illustration of SA-Mo/NPC. (b) Mo EDS mapping. (c) HAADF-STEM image. (d) LSV curves. (e) NH₃ yield rate and FE at each given potential. (f) NH₃ yield rates with different Mo loadings. (g) Chronoamperometric curve and FE stability. (h) Mo–N bond length during the NRR process. (i) Charge variation of a single Mo atom during the NRR process. Charge variation of the three moieties (C₉N₄, single Mo atom and the adsorbed N_xH_y species) along (j) distal and (k) alternating pathways. (a–g) Reproduced with permission.¹²⁰ Copyright 2019, Wiley-VCH. (h–k) Reproduced with permission.¹²¹ Copyright 2020, Elsevier B.V.

2.2.2. Regulating the electronic structure of DACs. A DAC refers to a catalyst with metal dimers on the substrate. The interaction between two adjacent active metal sites redistributes electrons, which is conducive to improving the catalytic performance, and catalysing complex reactions involving multiple reactants, reaction intermediates and products. According to the kind of active metal center, DACs are divided into homonuclear DACs and heteronuclear DACs.
2.2.2.1. Homonuclear DACs. Homonuclear DACs can be obtained by increasing metal loadings. Compared with SACs, the active sites of DACs increase due to the increase in the metal loadings.¹²² In addition, the interaction between homonuclear metals can regulate the electronic structure of active centers, which is conducive to reducing the reaction energy barrier. Zhou et al. constructed a diatomic Ag/graphene with AgN₃–AgN₃ structure (Ag₂–G), which showed good activity in catalysing the conversion of CO₂ to CO, with an FE of up to 93.4%.¹²³ Theoretical calculations showed that the AgN₃–AgN₃ active site has a lower energy barrier of *COOH formation than single-site Ag₁–G. Wang et al. reported a diatomic Pd catalyst (Pd₂ DAC) for the CO₂RR, exhibiting an FE_CO of 98.2%.¹²⁴ DFT calculations unveiled that the enhanced activity of the CO₂RR is due to the electron transfer between Pd and Pd, optimizing the adsorption strength of the reaction intermediate CO*. Quan et al. synthesized a nitrogen-coordinated Fe DAC (Fe₂NPC) using Fe₂(CO)₉ as the Fe source, exhibiting an FE_CO of 96.0%.¹²⁵ XAS characterization showed that the Fe₂ dimer was coordinated with six N atoms (Fe₂N₆) (Fig. 12a–c). Theoretical calculations showed that CO₂ → *CO₂ is an exothermic process (−0.03 eV) at Fe₂N₆, while an energy barrier of +0.59 eV needs to be overcome at Fe–N₄. In addition, another CO₂ can be adsorbed at the Fe₂N₆ site to form a simultaneous adsorption of *CO and *CO₂, which is subsequently converted into a *CO and *CO configuration, and the desorption of CO requires only 0.20 eV. The maximum energy barrier of CO₂ to CO at the Fe₂N₆ site is 0.90 eV (Fig. 12d). The desorption temperature of CO₂ at the Fe₂N₆ site is also higher than that of Fe–N₄ (Fig. 12e). The synergistic effect of the Fe₂ dimer is conducive to the adsorption of CO₂ and the desorption of CO.

Fig. 12 (a) FT-EXAFS spectra at the Fe K-edge. (b) Corresponding EXAFS R-space fitting curves and (c) WT of Fe₂NPC. (d) Calculated free energy. (e) CO₂-TPD curves. Reproduced with permission.¹²⁵ Copyright 2022, American Chemical Society.

2.2.2.2. Heteronuclear DACs. The synergistic effect of bimetallic active sites can break the linear proportional relationship of single atomic sites, which is conducive to catalysing complex multi-electron reactions and improving the stability.¹²⁶ The electron coupling of adjacent metal sites can optimize the adsorption energy and adsorption mode of reaction intermediates, reduce the energy barrier and accelerate the reaction kinetics. Due to the unique electronic structure of DACs, great progress has been made in electrocatalytic reactions such as the HER, OER, ORR, CO₂RR and NRR.^9,127

The development of high-performance DACs is expected to replace platinum for the HER. Fan et al. prepared a W₁Mo₁-NG DAC with W–O–Mo–O–C structure by combining a hydrothermal method and chemical vapor deposition (CVD) technology.¹²⁸ The W and Mo diatomic structure is the active center of the HER and has high stability. Theoretical calculations showed that the electron delocalization of the W–O–Mo–O–C configuration provided the best adsorption strength for H and accelerated the kinetics of the HER. Hu et al. used atomic layer deposition to anchor Pt to Ni and N co-doped carbon (PtNi-NC), exhibiting an η₁₀ of 30 mV for the HER in an acidic medium.¹²⁹ XAS characterization indicated the presence of a Pt–Ni dimer in the PtNi-NC. DFT calculations showed that Pt–Ni is beneficial to regulate adsorption energy and increase catalytic activity by regulating the local electronic structure and charge. Liu et al. studied cobalt-based DACs (Co–M DACs) for the HER and found that the activity of DACs was significantly better than that of a single-atom Co catalyst.¹³⁰ Theoretical calculations showed that OH is more easily adsorbed on the CoFe DAC, promoting hydrogen evolution under alkaline conditions.

The ORR and OER are the core reactions of various energy conversion devices, and their main challenge is the complex four-electron transfer process with a high overpotential. Fu et al. used a plasma engineering method to immobilize an Fe–Co dimer onto mesoporous nitrogen-doped carbon nanotubes (Fe, Co SAs-PNCF) with metal loading of up to 9.8 wt%.¹³¹ Fe, Co SAs-PNCF showed good ORR activity (E_onset = 1.04 V and E_1/2 = 0.93 V). XAS analysis revealed that the active center of the Fe, Co SAs-PNCF catalyst is the N₃–Fe–Co–N₃ structure. In situ XAS and Raman spectroscopy were used to study the changes of oxidation states of Fe and Co and the bond length of the N₃–Fe–Co–N₃ structure during the ORR. The Raman peak intensities of CoOOH/FeOOH and Fe(OH)₂/Co(OH)₂ changed with increasing bias, and oxygen intermediates on the catalyst surface increased continuously, indicating that Fe/Co species are the active sites (Fig. 13a–c). In situ XAS results showed that the oxidation states of Fe and Co become smaller and Fe/Co–N bonds become longer during the ORR process at lower voltages (Fig. 13d–f), indicating that the N₃–Fe–Co–N₃ structure is the active center of the ORR. Theoretical calculations were further used to reveal the mechanism of the ORR. The limiting potential (U_L) of Fe, Co SAs–V (0.89 V) is higher than that of Fe, Co SAs (0.81 V) and NCF (0.278 V). The spin orbit of Fe, Co SAs-V is closer to the Fermi level, and the transfer of electrons between metal and oxygen atoms facilitates electron accumulation and accelerates the ORR process (Fig. 13g–i).

Fig. 13 Potential-dependent in situ Raman spectra during the ORR process in (a) alkaline and (b) acidic solutions. (c) Normalized Raman intensities of the stretching mode of OOH at 730 cm⁻¹. (d) In situ XAS spectra recorded at the Fe K-edge. (e) Fitted average oxidation states of Fe from XANES spectra. (f) Lengths of the Fe–N bond obtained from the FT-EXAFS spectra. (g) Optimized structures of OH*, O*, and OOH* intermediates on the Fe, Co SAs-PNCF surfaces. (h) Free energy diagram. (i) Electronic density of states (DOS). Reproduced with permission.¹³¹ Copyright 2021, Elsevier Ltd.

Hu et al. synthesized a cobalt–iron diatomic catalyst using in situ electrochemistry by introducing iron onto a single-atom Co catalyst.¹³² The dimerized Co–Fe structure is the active site of the OER. Lou et al. synthesized a nitrogen-doped carbon material (a-NiCo/NC) anchored by Ni–Co sites using a multi-step template strategy, in which Ni and Co are dispersed on hollow prisms of the carbon.¹³³ DFT calculations showed that the strong interaction between Ni and Co sites optimizes the electronic structure, reduces the reaction energy barrier, and increases the OER activity. Wu et al. obtained a diatomic catalyst (FeCo-NC) by calcining an Fe and Co co-doped zeolite imidazole skeleton (ZIF-8).¹³⁴ FeCo-NC showed good bifunctional catalytic activity, exhibiting E_1/2 of 0.877 V in the ORR and E_j=10 of 1.579 V in the OER. The increase in the activity of the ORR and OER can be attributed to the adjustable electronic structure of FeCo–N₆. In addition, the fractional micropore–mesopore structure of the catalyst can increase the specific surface area and facilitate mass transfer.

Compared with SACs, DACs provide a variety of adsorption sites for the CO₂RR, and the synergistic effect between diatomic sites can improve the selectivity of CO₂ conversion to CO and regulate the electronic structure to promote the subsequent reaction. The CO selectivity on the diatomic Ni–Fe catalyst synthesized by Zhao et al. was as high as 98% at −0.7 V, and the electron interaction between Ni and Fe is beneficial to COOH* formation and CO desorption.³³ Wu et al. synthesized a diatomic catalyst with Co–Cu heterodiatomic pairs, which has an FE of 99.1% for CO generation and FE_CO of more than 95% at high current densities.¹³⁵ The introduction of the second metal can effectively regulate the electronic structure and promote the formation of intermediate *COOH. He et al. prepared a nitrogen-doped carbon matrix catalyst with Fe–Cu diatomic sites, and the CO FE in the CO₂RR was as high as 99.2%.¹³⁶ The synergistic effect of Fe–Cu accelerates the charge transfer, and effectively regulates the position of the d-band center, reducing the reaction energy barrier. He et al. prepared a Ni/Cu–N–C catalyst with Ni–Cu double sites, in which an N₄Ni/CuN₄ site structure was formed (Fig. 14a–c).¹³⁷ The catalyst had an FE_CO of 99.2% at −0.79 V. Theoretical calculation results showed that the adjustment of electronic structure promoted by the two-site interaction can effectively narrow the band gap, promote charge transfer, and reduce the reaction energy barrier, which is conducive to the generation of CO (Fig. 14d–f).

Fig. 14 (a) Ni K-edge FT-EXAFS curves. (b) Ni K-edge EXAFS fitting curves of Ni/Cu–N–C in the R space. (c) HOMO and LUMO spin electron densities of the proposed Ni/Cu–N–C and Ni–N–C models. (d) Optimized catalytic models and reaction pathways. (e) Free energy profiles of the CO₂RR. (f) Difference in limiting potentials for the CO₂RR and HER. Reproduced with permission.¹³⁷ Copyright 2021, American Chemical Society.

At present, C₁ products are mainly reported on DACs in the catalytic CO₂RR, and further in-depth products have been rarely reported. The mechanism of multi-electron reaction to produce methane and methanol on DACs for the CO₂RR was explored, demonstrating the way of C–C coupling to obtain C₂ products. Lu et al. constructed Fe/Co SACs and found that CO was the only product in the CO₂RR.¹³⁸ With the help of CO intermediates, the limiting potential on the FeCo-NC DAC to reduce CO₂ to CH₃OH and CH₄ is as low as −0.64 V. Under applied voltage, FeCo-NC is more likely to generate CH₃OH and CH₄, since the two-site center facilitates spin polarization and electron transfer. *CO tends to adsorb the bridge position of the dual-sites on either side, and C–C coupling is not easy to occur since greater excitation energy is required. In addition, the increase of CO concentration at the dual-sites will cause poisoning.¹³⁹

2.2.3. Regulating the electronic structure of TACs. DACs are an important improvement over SACs, and on this basis, TACs can be fabricated by continuing to adjust the number of central metal atoms, providing more space for the regulation of electronic structure.¹⁴⁰ Compared with SACs and DACs, TACs with suitable coordination structure have the potential to catalyse complex reactions with multiple electron and proton coupling processes. However, due to the difficulty of controllable synthesis and the complexity of electronic structure, the research on TACs is still in the initial stage. The electronic structure can be adjusted by optimizing the metal configuration of active sites. In addition, the structural symmetry of active sites also significantly affects electron distribution. Wu et al. designed a trinuclear catalyst with a manganese atom and two cobalt atoms coordinated with eight nitrogen atoms (Co₂MnN₈/C).³⁸ Co₂MnN₈/C can efficiently catalyse the ORR with a half-wave potential of 0.912 V. DFT calculations showed that Co₂MnN₈–2OH formed by the combination of Co₂MnN₈ and OH was the active site of the ORR, which led to the rearrangement of d electrons and optimized the adsorption energy of the reaction intermediates. Li et al. prepared atomically dispersed Ru₃ on a nitrogen-doped porous carbon catalyst by cage separation and pyrolysis.³¹ The cage of ZIF-8 can just wrap Ru₃(CO)₁₂ molecules, and uniform nitrogen-fixed Ru₃ clusters were synthesized by annealing (Fig. 15). HAADF-STEM and XAS results showed that Ru₃ clusters have a triangular structure, and Ru–N coordination and Ru–Ru metallic bonds were detected, indicating that the nitrogen coordination effectively stabilized Ru₃ clusters. Liu et al. prepared a trinuclear copper-based cluster (Inz-Cu₃) catalyst to research the effect of structural symmetry on the selectivity of CO₂RR products by changing the distance and angle between adjacent Cu active sites.¹⁴¹ The experimental results showed that at the current density of −320 mA cm⁻², the Faraday efficiency of CO₂ conversion to C₂ is 42.20% under acidic conditions, and under alkaline conditions, the Faraday efficiency of CO₂ conversion to C₂ is 66.79%, with FE_C2H4 = 35.27% and FE_C2H5OH = 31.52%. Liao et al. synthesized a stable and inexpensive metal-azo salt trinuclear copper cluster catalyst (Cutrz) that efficiently catalysed CO₂ to ethylene and C₂₊ with Faraday efficiencies of up to 50% and 80%, respectively.¹⁴² Theoretical calculations showed that three CO intermediates can be closely adsorbed on the same side of Cu clusters, accelerating the C–C coupling probability. Yagi et al. used graphene oxide and polynuclear copper precursors to synthesize an electrocatalyst with trinuclear copper sites through a rapid heating strategy, which can effectively catalyse the ORR at pH = 7, exhibiting an initial potential of 0.82 V and good stability.¹⁴³

Fig. 15 (a) HAADF-STEM image of Ru₃/CN. (b) Corresponding intensity maps (left), models (right) and (c) intensity profiles obtained in areas labeled 1, 2, and 3 in (a). (d) Ru K-edge XANES spectra. (e) EXAFS Fourier transformed (FT) k²-weighted χ(k) function spectra. (f) Wavelet transforms for the k²-weighted Ru K-edge EXAFS signals for the high-coordination shells in Ru₃/CN. The colors in the contour figures indicate the moduli of the Morlet wavelet transform. (g) Comparison of the q-space magnitudes for FEFF-calculated k²-weighted EXAFS paths. (h) Corresponding EXAFS R-space fitting curve for Ru₃/CN. (i) Schematic model of Ru₃/CN. Reproduced with permission.³¹ Copyright 2017, American Chemical Society.

To study the synergistic effect between three nuclear metal atoms, Wang et al. systematically simulated the catalytic performance of single atoms, dual atoms and triple atoms in the electrocatalytic NRR by theoretical calculations.¹⁴⁴ The results showed that the catalytic activity of the TAC was generally superior to that of the SAC and DAC since the reduction of adsorption energy of the NH₂ intermediate effectively promotes the rate control step of *NH₂ hydrogenation. Due to the relatively low loading capacity of single/double/triple site catalysts, there are few active sites which affect the catalytic activity. Jiang et al. used density functional theory to construct a triatomic Fe₃-GDY/Gra catalyst with a theoretical mass loading of up to 35.8 wt%, showing excellent activity in the theoretical NRR at a potential of −0.37 V.¹⁴⁵ High loading provides more active sites to activate N₂ more easily. Zhou et al. theoretically demonstrated that a trinuclear catalyst anchored on nitrogen-doped carbon can effectively convert CO₂ into C₂ and C₂₊ products in the CO₂RR.²¹ The trinuclear active sites can adsorb multiple CO₂ at the same time, which is conducive to C–C coupling.

2.3. Regulating the coordination environment of central metal sites

Electrocatalytic active sites are formed by bonding between central metal atoms and the surrounding coordination atoms. Therefore, adjusting the bonding mode can tune the electronic structure of central metal atoms, including the valence and spin states and the electron cloud density near the Fermi level. The target active structure is obtained by adjusting the number, type and configuration of coordination atoms, which is conducive to understanding the relationship between coordination configuration and catalytic performance.^146,147

2.3.1. Adjusting the number of coordination atoms. M–N_x coordination structures have attracted much attention due to their high electrocatalytic activity and durability. For SACs, isolated single metals have high surface free energy, and easily migrate and reunite. This process can be effectively prevented through the interaction between metal and N atoms.^148–150 The usual ways to synthesize M–N_x coordination structures are directly annealing metal- and nitrogen-containing precursors in an inert nitrogen atmosphere. By regulating the number of N atoms in the M–N_x (x = 1–5) coordination structure, the electronic structure of the active site can be effectively regulated, thus promoting the catalytic effect on the HER, OER, ORR, CO₂RR and NRR.^39,114,151

Among them, the M–N₄ structure is the most studied active site, which is like that of metal porphyrins and metal phthalocyanines, but the chemical environment after loading on the carbon carrier is obviously different from that of metal porphyrins/metal phthalocyanines. M–N₄ sites showed good activity for many electrocatalytic reactions. Guo et al. synthesized single-atom Co and Ni catalysts through three steps of impregnation–carbonation–acidification.¹⁵² The single-atom Co catalyst showed excellent catalytic activity and stability for the OER/ORR, and the reversible oxygen overpotential was 0.75 V. It was assembled as a cathode into a zinc–air battery with a capacity of 530.17 mA h g_zn⁻¹. Theoretical calculation results showed that the OH* hydrogenation barrier increases when there are clusters in the catalyst, which is not conducive to the promotion of ORR/OER activity.

Xie et al. synthesized a Ni–N₄–C catalyst using a topochemical transformation strategy, which effectively avoided the agglomeration of Ni metal.¹⁵³ The FE of converting CO₂ to CO was up to 99%. Theoretical calculations showed that the Ni–N₄ site effectively reduces the generation energy of COOH* and accelerates charge transfer to promote the conversion of CO₂ to CO (Fig. 16a–e). The distance between different M–N_x coordination structures affects the selectivity of the CO₂RR. Zheng et al. adjusted the Cu–N_x structure by adjusting pyrolysis temperature and Cu doping amount.¹⁵⁴ Experimental and theoretical results showed that Cu–N_x can promote the conversion of CO₂ to CH₄ when the concentration of Cu is less than 2.4%_mol, while when the concentration is increased to 4.9%_mol, CO₂ can be transformed into C₂H₄. Two CO intermediates can be coupled to produce C₂H₄ at adjacent Cu–N₂ sites (Fig. 16f).

Fig. 16 (a) Schematic illustration of the topochemical transformation strategy. (b) FT of the Ni K-edge EXAFS oscillations of Ni-doped g-C₃N₄. (c) FT of the Ni K-edge EXAFS oscillations of Ni–N₄–C. (d) Calculated free energy diagram. (e) Difference in limiting potentials for CO₂ reduction and H₂ evolution. (f) Schematic of the procedure to synthesize Cu–N–C–T catalysts. (a–e) Reproduced with permission.¹⁵³ Copyright 2017, American Chemical Society. (f) Reproduced with permission.¹⁵⁴ Copyright 2020, American Chemical Society.

Fu et al. used in situ XAS to reveal the relationship between the ORR and the configuration of active sites.¹⁵⁵ They studied the dynamic changes of the Mn–N₄ configuration during the ORR process. Under applied voltage, Mn–N₄ was transformed into Mn–N₃C, and then further transformed into the Mn–N₂C₂ configuration. The valence state of Mn increased from +3.0 to +3.8 and then decreased to +3.2. Therefore, Mn–N₃C and Mn–N₂C₂ configurations really regulated the adsorption energy barrier and adsorption mode of the reaction intermediates in the ORR process. Wang et al. synthesized nitrogen-doped graphene with highly dispersed FeN₅ active sites, which achieved 97.0% FE in converting CO₂ to CO.¹⁵⁶ Theoretical calculations showed that five N coordination depleted the electron density of Fe 3d, effectively promoting the formation of CO. Jiang et al. prepared single-atom Ni–N_x–C with different N coordination numbers by a post-synthetic metal substitution method.¹⁵⁷ The FE on the Ni–N₃–C catalyst was 95.6%, which is significantly higher than that of the Ni–N₄–C catalyst. DFT calculations showed that the Ni–N₃ site was conducive to COOH* production and accelerated CO production. Wang et al. reasonably designed a single-atom W catalyst with a W₁N₁C₃ structure, showing an η₁₀ of 85 mV in 0.1 M KOH solution for the HER.⁶²

In DACs, the M₁M₂-N₆ coordination structure is usually formed. Zhao et al. anchored diatomic Ni–Fe sites on carbon nitride, which showed an FE_CO of up to 90% and CO₂RR stability.¹⁵⁸ Theoretical calculations revealed that the NiN₃–FeN₃ active site reduced the formation energy of the COOH* intermediate and the desorption energy of CO. In addition, other M₁M₂–N_x coordination structures also play an important role in the CO₂RR. Lu et al. anchored diatomic Ni₂–N₃ sites onto carbon nanotubes (CNTs), on which the energy barrier of the *COOH intermediate could be more significantly reduced than on Ni–N₃, Ni–N₄ and Ni₂–N₄ sites.¹⁵⁹ The energy barrier of the rate-limiting step of *CO desorption is 0.76 eV, which is conducive to the conversion of CO₂ to CO. Zhong et al. reported three coordination modes of Ni₂–N₇, Ni₂–N₅C₂ and Ni₂–N₃C₄, among which Ni₂–N₃C₄ showed excellent activity in the CO₂RR.¹⁶⁰ DFT calculations showed that a suitable coordination environment regulated the electronic structure of Ni₂ and reduced the energy barrier of COOH*.

The Fe–N_x structure has been considered to be an effective active site for the ORR.^161–163 However, a two-electron transfer process will inevitably occur, generating a small amount of H₂O₂, which attacks N ligands, resulting in the dissolution of Fe and inactivation. To inhibit the two-electron ORR process, it is needed to regulate the Gibbs free energy of *OOH to prevent the spontaneous generation of H₂O₂.^164,165 By introducing a second metal, the electronic structure of Fe–N₄ can be effectively adjusted to optimize the adsorption free energy of reaction intermediates.¹⁶⁶ Cao et al. synthesized a diatomic catalyst (Fe–Mn–N–C) with FeN₄–MnN₃ sites.¹⁶⁷ There is a defect around FeN₄–MnN₃ and the Fe–Mn bond has a strain effect, leading to a significant change of the space charge density around Fe and Mn. The rate-limiting step of the ORR on FeN₄ is OH* → H₂O, while it changes to O* → OH* after the introduction of MnN₃. The intermediate OOH* is adsorbed at the Fe–Mn bridging position, which is conducive to O–O dissociation, and hydrogen peroxide generation by 2e⁻ can be completely inhibited. The FeN₄–MnN₃ configuration decreased the d-band center of FeN₄ compared with FeN₄, and weakened the binding strength, thus reducing the adsorption energy of oxygen-containing intermediates (Fig. 17).

Fig. 17 (a) FeN₄–MnN₃-d1 structure (d1 referring to the number of defect being 1). (b) Side view of the FeN₄–MnN₃ structure. (c) Charge density difference diagrams of FeN₄–MnN₃-d1. (d) Oxygen-free energy diagrams of FeN₄ and FeN₄–MnN₃-d1. (e) Illustrations of the ORR process on the FeN₄–MnN₃-d1 structure. (f) The d-band changes of Fe in FeN₄ and FeN₄–MnN₃-d1. (g) PDOS of Fe in the d_xz, d_yz and d_z² orbitals and the bonding and anti-bonding state information obtained by COHP. Reproduced with permission.¹⁶⁷ Copyright 2020, Elsevier B.V.

2.3.2. Regulating the type of coordination atoms. In addition to the N ligand, S, P, B, O and other nonmetallic elements can also be used to regulate the local electronic structure of metal active sites. S has a good affinity for transition metal elements, and can coordinate with the central metal atom to form the M–N_xS_x structure, so that the local charge of the active center is rearranged. Zhu et al. designed an S and N co-doped single-atom Fe catalyst with excellent OER/ORR bifunctional activity.¹⁶⁸ S doping modulates the charge and spin distribution around Fe–N_x and improves the catalytic activity. Li et al. synthesized an N and S co-coordinated Cu single-atom (S-Cu-ISA/SNC) with a hierarchically porous structure.¹⁶⁹ XAS results showed Cu–N (1.55 Å) and Cu–S (1.81 Å), and the FT-EXAFS indicated that the central metal Cu coordinated with three N and one S (Fig. 18a). The S-Cu-ISA/SNC catalyst with an asymmetric Cu–S₁N₃ structure can efficiently catalyse the ORR with a half-wave potential of 0.918 V. Subsequently, to study the change of Cu structure in the ORR, in situ XAS tests were conducted. The valence of Cu changed from +2 to +1, and the length of the Cu–N bond was shortened from 1.98 Å to 1.94 Å (0.90 V) and 1.93 Å (0.75 V), while the length of the Cu–S bond was almost unchanged (Fig. 18b–g). According to the volcano-type plot (Fig. 18h), Cu–N₄ and Cu–S₁N₃ are located at the right side, indicating that during the ORR process, the intermediate adsorption at the Cu sites is weak (Fig. 18i).

Fig. 18 (a) Schematic atomic interface model of S-Cu-ISA/SNC. (b) Schematic of the in situ electrochemical cell set-up. (c) Cu K-edge XANES spectra of S-Cu-ISA/SNC at various potentials during the ORR. (d) Differential Δμ XANES spectra. (e) Current density as a function of potential for S-Cu-ISA/SNC (left) and the average oxidation number of Cu species in S-Cu-ISA/SNC as a function of potential (right). (f) k³-weighted FT-EXAFS. (g) The proposed ORR mechanism for the S-Cu-ISA/SNC. (h) η_ORR as a function of ΔG_O* on different Cu-centered moieties. (i) Free-energy diagram. Reproduced with permission.¹⁶⁹ Copyright 2020, Nature.

P and N belong to the same group V, but the electronegativity of P is weaker than that of N, and therefore, the adsorption activity of reaction intermediates can be regulated by introducing P in the common tetragonal configuration. Wang et al. synthesized a single-atom Co catalyst (Co-SA/P-in situ) with a Co₁–P₁N₃ coordination structure, showing an η₁₀ of 98 mV for the HER.¹⁷⁰ The bond lengths of Co–P and Co–N were 2.18 Å and 1.95 Å, respectively. In situ XAS results showed that the valence state of Co increased from +1.53 to +1.78 during the HER, and the bond lengths of Co–P and Co–N became longer, indicating that Co₁–P₁N₃ was the active center. DFT calculations showed that P and N doping promoted the charge transfer, and effectively improved the catalytic activity of the HER. Chen et al. found that a N- and P-codoped single-atom Fe catalyst (Fe–N/P–C) can efficiently catalyze the CO₂RR, and the FE_CO can reach 98%.¹⁷¹ DFT calculations showed that the introduction of P reduces the oxidation state of Fe, optimizes the adsorption energy of the *CO intermediate, and improves the catalytic activity.

B is an electron-deficient atom, and when B and N co-coordinate with the metal, B shares electrons with the electron-rich N to regulate the electronic structure around the metal. Xin et al. synthesized a single-atom Ni catalyst (Ni–B/N–C) with a Ni–B₁N₃ coordination structure, exhibiting a half-wave potential of 0.87 V for the ORR.¹⁷⁴ Theoretical calculation results show that the introduction of B can adjust the local coordination environment of Ni–B₁N₃, reduce the adsorption energy barrier of reaction intermediates, and accelerate the ORR process.

The electronegativity of O is greater than that of N, and O and N can form quad-coordination with the metal in the plane and quintuple coordination outside the plane. Compared to other non-metallic coordination atoms, no additional oxygen source is introduced during the synthesis process due to the presence of oxygen in most carbon precursors. Chen et al. successfully introduced single-atom Mn–N₃O₁ into a 3D graphene framework (Mn/C–NO), optimized the position of the Mn d-band center by regulating the geometry of local coordination, making it conducive to the adsorption and desorption of reaction intermediates, and accelerating the kinetics of the ORR.¹⁷⁵ Chen et al. used a sacrificial template method to controllably anchor Ni–N₄–O sites on nitrogen-doped porous graphite carbon (NiSA-N-PGC).¹⁷² The highest FE of NiSA-N-PGC in the catalytic conversion of CO₂ to CO was 97.2%. In situ infrared spectroscopy was used to study the reaction intermediates adsorbed on the catalyst surface, demonstrating that *COOH and *CO were the main intermediates in the CO₂RR process. With the increase of potential, the peak strength of *CO increased, while the peak strength of *COOH decreased, indicating that *COOH → *CO accelerated, and the activity of the CO₂RR increased. Theoretical calculations revealed the promotion effect of the axial Ni–O coordination on the CO₂RR. In the CO₂RR, the conversion step of CO₂ to CO is CO₂ → *COOH → *CO → CO. The oxidation state of Ni in Ni–N₄–O is +1.12 e, which is higher than that in Ni–N₄ due to the transfer of electrons from Ni to the axial O. Therefore, the Ni 3d orbitals in the former exhibit higher DOS near the Fermi level than the latter. The coordination of axial O with high electronegativity enhances the electron delocalization of the active center metal Ni, reduces the reaction energy barrier in the CO₂RR, and facilitates the adsorption and desorption of reaction intermediates (Fig. 19a–d).

Fig. 19 (a) Potential-dependent operando attenuated total reflection surface enhanced infrared absorption spectroscopy (ATR-SEIRAS) spectra of NiSA(0.3)-N-PGC. (b) Proposed reaction steps of electrocatalytic reduction of CO₂ to CO over the Ni–N₄–O site. (c) PDOS calculated using DFT for Ni 3d. (d) Reaction paths and free energy diagrams. (e) Operando XANES spectra recorded at the Ni K-edge of Ni DSC. (f) Least-squares curve-fitting analysis of operando EXAFS spectra at the Ni K-edge. (g) Comparison between the Ni K-edge XANES experimental spectra (solid lines) and the theoretical spectra. (h) Projected crystal orbital Hamilton population (pCOHP) between the metal center and the COOH* intermediate. (i) Electron density difference plot of the *COOH intermediate adsorption structure and the Bader charge analysis. (a–d) Reproduced with permission.¹⁷² Copyright 2022, American Chemical Society. (e–i) Reproduced with permission.¹⁷⁶ Copyright 2021, American Chemical Society.

In DACs, the electron effect between two adjacent metal atoms can provide a variety of absorbable sites and accelerate electrocatalytic reactions. Hou et al. synthesized an atomically dispersed indium–nickel diatomic catalyst (InNi DS/NC), in which an oxygen atom is axially bridged between In and Ni.¹⁷³ XAS characterization showed that the coordination configuration was O–In–N₆–Ni, and the FE of CO generation was up to 96.7%. In situ characterizations and theoretical calculations showed that the interaction between the In–N₆–Ni site and bridging oxygen reduced the energy barrier of the *COOH intermediate and inhibited the competitive HER. In addition, the electrolyte can provide an oxygen atom axially between the two core metals and affect the activity and efficiency of the catalyst. Yao et al. synthesized a binuclear Ni cluster precursor and then anchored it to a nitrogen-doped carbon material to obtain Ni₂/NC with binuclear Ni₂–N₆ sites.¹⁷⁶ The FE of catalytic conversion of CO₂ to CO was greater than 94%. Operando XAS analysis and theoretical calculation results showed that the Ni₂–N₆ in the electrolyte first adsorbed an oxygen-containing group to form an O–Ni₂–N₆ structure during the initial reaction, and then the CO₂ adsorbed on the O–Ni₂–N₆ structure was rapidly activated, and then the CO* intermediate was obtained through two proton coupling processes. Finally, CO* was desorbed on the O–Ni₂–N₆ structure. The occupancy rate of Ni–COOH* antibonding states at O–Ni₂–N₆ is lower than that at Ni–N₄, indicating stronger adsorption of intermediates at the former. In addition, the positive charge (0.23 e) at O–Ni₂–N₆ was higher than that at Ni–N₄ (0.18 e), indicating that the O–Ni₂–N₆ site was more conducive to the activation of CO₂ conversion to COOH* (Fig. 19e–i).

3. Theoretical understanding of atomic catalysts for electrocatalysis

Electrocatalytic reactions usually take place at the interface or surface of catalysts, and structural variations affect the adsorption and the desorption. On the one hand, well-designed engineering of the geometry or electronic structure produces synergistic effects that can enhance catalytic activity, stability and selectivity. On the other hand, the changes in the electronic structure affect the charge density and the degree of d-band filling in the spin state, which modulates the binding model, leading to a decrease in the reaction energy and an increase in the catalytic activity.^68,177,178

3.1. Synergistic effect

Modulating the coordination environment and electron distribution of atomic catalysts is an effective but still challenging strategy to improve the electrocatalytic performance.¹⁷⁹ Many works have shown the activity of DACs and TACs is higher than that of SACs due to a synergistic effect.⁶⁸ The synergistic effect between polynuclear metal atoms can regulate the local electronic structure, promote the transfer of electrons, optimize the adsorption and desorption properties, and reduce the overall reaction barrier, thus improving the electrocatalytic performance.⁴¹ Theoretical calculations showed that the synergistic action of neighboring metals optimizes the d-band center positions of the metal centers and balances the free energies of the oxygen intermediates, thus increasing the oxygen electrocatalytic activity. Bu et al. reported a homogeneously conceived FeCo-DACs/NC catalyst, in which the synergistic interaction between Fe and Co exerted a significant effect on the reaction barrier.¹⁸⁰ The tiny band gap of FeCo-DACs/NC increased the electronic conductivity and accelerated the electron transport. Meanwhile, the synergistic interaction between Fe and Co made the metal d-band center of FeCo-DACs/NC lower than the Fermi energy level and the energy level position closer to the 2p orbital of O than that of the corresponding SACs (Fig. 20a and b). Lee et al. reported a more efficient Fe₂–N–C electrocatalyst than Fe₁–N–C. The diatomic Fe orbital coupling led to the reduction of orbital energy levels and electron delocalization, which facilitated *CO desorption. The orbital coupling between the dual-site Fe atoms reduced the energy gap between the antibonding and bonding states in *CO adsorption (Fig. 20c–e).¹⁸¹ Li et al. reported the construction of a bimetallic catalyst by introducing single-atom Co into the second ligand shell layer of the Fe center.¹⁸² The addition of the Co atom to the d-band center in FeCo-NC (−0.98 eV) brought it closer to the Fermi energy level than Fe-NC (−1.61 eV), indicating that fewer electrons populate the antibonding orbitals of the active center, thus enhancing the phase hundred interaction between the active and adsorbed substances, and facilitating the subsequent substrate reaction (Fig. 20f).

Fig. 20 (a) Proposed architecture of the Fe–Co metal pair in FeCo-DACs/NC. (b) DOS patterns of FeCo₁, Co-SACs/NC, and Fe-SACs/NC. (c) Bare surface with a range from −0.01e Å⁻³(blue) to +0.04e Å⁻³ (red) and (d) *CO adsorption with an isosurface level of 0.002e Å⁻³. (e) PDOS of Fe–N₄ and N₃–Fe–N–Co. (f) Schematic illustration of orbital interaction between Fe-3d (d_z² and d_xz/d_yz) and adsorbed CO (5σ and 2π*) for Fe₁–N₄–C and Fe₂–N₆–C-o. (a and b) Reproduced with permission.¹⁸⁰ Copyright 2022, Wiley-VCH GmbH. (c and e) Reproduced with permission.¹⁸¹ Copyright 2022, American Chemical Society. (f) Reproduced with permission.¹⁸² Copyright 2023, Wiley-VCH GmbH.

3.2. Defect coupled spin state change

Defects are prevalent in heterogeneous materials, and can be used as active sites to fabricate specific catalysts or to promote catalytic performance. Defect engineering is an effective way to modulate the surface physical/chemical properties of catalytic materials, and the presence of defects will significantly change the electronic structure and chemical properties, leading to the formation of new physicochemical properties or strong synergistic effects, and thus optimizing the catalytic performance.¹⁸³ The change of defect-coupled spin states has gradually attracted more and more attention. Yang et al. reported that the charge redistribution of the attached exfoliated monolayer of iron phthalocyanine (FePc) could be effectively regulated using a defective graphene substrate with 585 defects.¹⁸⁴ The transfer of electrons from the 585 defects to FePc formed an electron-rich region on the Fe atoms, and the high-density electrons further elevated the d-band center of the Fe atoms. Obviously, this adjustment of the electronic structure of Fe atoms favoured the adsorption and reaction of O₂ molecules (Fig. 21a–c). Nitrogen-doped carbon-coupled FeNi₃ intermetallic compounds (FeNi₃@NC) with excellent OER and ORR properties were reported by Mu et al.¹⁸⁵ The defects in FeNi₃@NC, including impurities, dopants, vacancies, and interstitial defects, can significantly change the chemical properties and electronic structure, promoting the formation of new physicochemical properties or strong synergistic effects. Mai et al. utilized the coupling effect of edge defects of coordinated bismuth (Bi) and sulfur to improve the selectivity of the CO₂RR and inhibit the competing HER.¹⁸⁶ DFT calculations showed that sulfur tends to bind to the Bi edge defects, reduces the ligand-unsaturated Bi sites (*H adsorption sites), and modulates the charge state of the neighboring Bi sites to improve the *OCHO adsorption (Fig. 21d and e). Wang et al. utilized a seeded decarboxylate-induced defect strategy to improve the intrinsic activity of the Co SAC by increasing the defect density.⁴³ DFT calculations confirmed that in the defect-rich Co SAC, the defect structure formed near the Co–N₄ site likely influences the charge redistribution, which can greatly improve the intrinsic catalytic performance and reduce the free energy of OH adsorption on the Co–N₄ site, thus facilitating the electrochemical reaction.

Fig. 21 (a) The top views of optimized FePc/DG-585 (FePc/G or FePc/NG) based hybrid interfaces. (b) The side views of the 3D charge density difference plot for the interfaces between the DG-585 (G or NG) sheet and FePc layer. (c) The schematic of a probable ORR electrocatalytic mechanism of FePc/DG-585 for the ORR. (d) The correlation between the adsorption energy (ΔG_*OCHO) and p-band center of active Bi atoms. (e) The formation energies. (a–c) Reproduced with permission.¹⁸⁴ Copyright 2020, Elsevier B.V. (d and e) Reproduced with permission.¹⁸⁶ Copyright 2023, Wiley-VCH GmbH.

3.3. Crystal field distortion spin state change

Although some previous discussions have demonstrated the important role of the coordination field of M–N–C catalysts in improving the catalytic performance, the mechanism is still unclear. Breaking the D_4h symmetry of the FeN₄ active center provides a new way to improve the activity of Fe–N–C catalysts.¹⁸⁷ Zhang et al. achieved the modulation of the electronic spin state of the Fe active center in SAC-Fe–N–C and the modulation of the electronic spin state of the FeN₄ active center from the low-spin state of FeN₅ to the high-spin states of FeN₄ and FeN₃, based on ligand-field theory by converting the defect-rich pyrrole-type N-coordinated FeN_x sites, thus allowing the electrons to easily penetrate the antibonding π-orbitals of oxygen.¹⁸⁸ Fe–N₄–HS has a three-dimensional electronic structure of t_2g and e_g, which promotes the kinetics of the ORR significantly. The asymmetric arrangement of the d-orbital electrons of Fe in the spin channel indicates spin polarization. Utilizing the activity-enhancing effect of the high spin state of Fe(III), the designed Fe–N₄-HS catalyst exhibited higher ORR activity than commercial Pt/C catalysts (Fig. 22a). Ling et al. showed the first crystal-field deformation of the FeN₄ portion achieved by the introduction of sulfur–oxygen-polar functional groups into Fe–N–C catalysts, and revealed that the disruption of FeN₄'s D_4h symmetry leads to the rearrangement of Fe 3d electrons and increases the spin moment of the Fe center.¹⁸⁷ The effective spin-state manipulation optimized the adsorption energy of ORR intermediates, which significantly improved the intrinsic ORR activity of the Fe–N–C catalysts (Fig. 22b–d). Ruan et al. remodelled the metal-centered coordination of the Fe centers by altering the crystal field of the Fe–N portion through the introduction of sacrificial bonds.¹⁶² Selective breaking of the sacrificial Fe–S bond resulted in electron-withdrawing sulfur oxides in the Fe center. The reduced electron localization around the Fe center gave the catalyst excellent ORR activity and remarkable stability, which facilitated the desorption of ORR intermediates.

Fig. 22 (a) Schematic illustration of bond formation between the Fe–N–C and the adsorbates (left), and schematic illustration between the d-band centers and adsorption–desorption abilities of ORR intermediates (right). (b–d) Schematic diagrams of breaking FeN₄ square-planar coordination by introducing polar XO₂ (X = S, Se, and Te) groups. (a) Reproduced with permission.¹⁸⁸ Copyright 2022, Elsevier Ltd. (b–d) Reproduced with permission.¹⁸⁷ Copyright 2022, Wiley-VCH GmbH.

3.4. Stress-induced spin state change

Changes in catalyst geometry induce strain effects that affect the spin polarization of active centers and the changes of charge density. Zhang et al. succeeded in introducing strain effects on the active C–C bond near the edge graphite nitrogen, which increased the proper spin polarization and charge density of the carbon active center, and kinetically favoured the adsorption of O₂ and the activation of O-containing intermediates.¹⁸⁹ The spin-polarization strategy allows the construction of highly curved carbon-based multinuclear metal electrocatalysts with improved ORR performance by introducing strain effects. Dong et al. designed magical concave Pt–Zn nanocubes with high-index faceted platinum skins, which produced a strain effect different from that of the ordinary cubes.¹⁹⁰ The length of the Pt–Pt bond was significantly shortened, which reduced oxygen-containing intermediates' adsorption energy on the platinum surface. Jiang et al. proposed a spin-state transition strategy modulated by axial Fe–O–Ti ligands to enhance the ORR activity of Fe centers.¹⁹¹ Theoretical calculations showed that the Fe–O–Ti ligand in FeN₃O–O–Ti induces a low to medium spin-state transition to optimize the adsorption of O₂ on the FeN₃O. The magnetic moment of Fe increased from 1.51 μ_B (FeN₃O) to 3.52 μ_B (FeN₃O–O–Ti). The low to moderate spin transition provides better E_g filling for Fe, which favors O₂ affinity. FeN₃O–O–Ti has suitable O₂ adsorption energy (1.84 eV) and ICOHP value (1.88 eV), meaning a moderate binding strength of O₂. It was therefore inferred that FeN₃O–O–Ti may have convincing ORR activity.

4. Challenges for optimizing atomic catalysts

4.1. Controlled synthesis

Accurate and controllable synthesis techniques play an important role in the fabrication of atomic catalysts. The active sites determine the catalytic activity and selectivity. Uniform active sites are conducive to the study of the relationship between structure and performance. At present, the common synthesis strategies for atomic catalysts include high temperature annealing and encapsulation.^{13,66,192,193} Random dispersion is suitable for low metal loading, while high metal loading might lead to metal agglomeration since metal atoms migrate at high temperatures. The encapsulation strategy can satisfy the need to obtain DACs with definite structure, but the size of the selected precursor must match the carrier cage size. Using the steric hindrance of the carrier to inhibit atomic migration, Qu et al. constructed a diatomic catalyst with Fe₂, Cu₂ and Ir₂ sites by using the interface encapsulation strategy.¹⁹⁴ To prevent thermal migration during synthesis, the interface encapsulation carrier ZIF-8 with polydopamine was used, and the bimetallic precursor was anchored into the hollow nanocages (Fig. 23a–d). Bu et al. adopted a “pre-constrained metal twins” strategy to encapsulate an FeCo dimer into ZIF-8 in situ for the creation of adjacent Fe–N₄ and Co–N₄ geometric configurations (Fig. 23e).¹⁸⁰ Theoretical calculations showed that the interaction between Fe and Co optimizes the d-band center of the catalyst, optimizes the adsorption energy of oxygen intermediates, and improves the catalytic activity of the ORR/OER.

Fig. 23 (a) HAADF-STEM image. (b) Intensity profiles highlighted in yellow in (a). (c) Fe K edge FT EXAFS spectra. (d) FT EXAFS fitting spectrum at R space. (e) Schematic illustration of the traditional strategy and “pre-constrained metal twins” strategy for fabricating DACs. (f) Schematic diagram of the NP-to-SA-to-DA atomization and sintering strategy. (a–d) Reproduced with permission.¹⁹⁴ Copyright 2022, Wiley-VC. (e) Reproduced with permission.¹⁸⁰ Copyright 2022, Wiley-VC. (f) Reproduced with permission.¹⁹⁵ Copyright 2023, Nature.

The major challenge of controlled synthesis is that the dynamic evolution of metal atoms and ligands during the synthesis is still not clear. To date, there have been some reports revealing this complex process. Huang et al. designed an atomization and sintering strategy to transfer Co nanoparticles to single-atom CoN₄ and then to diatomic Co₂N₅.¹⁹⁵ In this process, the nitrogen source played the role of stripping and fixing the metal atoms, and the Co₂N₅ sites were acquired by thermal migration spontaneous sintering (Fig. 23f). Wu et al. used in situ XAS to investigate the structural changes of the CoN₄ site during synthesis.¹⁹⁶In situ XAS was used to monitor the conversion of inactive Co–OH/O to the active CoN₄ site. The conversion could not be completed below 700 °C, and the best activity was obtained at 900 °C. However, due to the limitation of advanced characterization techniques, it is still very difficult to dynamically study the whole process of the formation of active sites.

4.2. Increasing the density of active sites

For typical synthesis of SACs, it is usual to reduce the metal load on a high specific surface substrate to gain atomic dispersion and prevent aggregation, but due to the low metal load, there are few metal active sites. The number of catalytic active sites can be effectively increased by increasing the metal loading, but the probability of metal agglomeration increases gradually with the increase of metal loading. At present, the commonly used methods to increase metal loading and reduce agglomeration are to stabilize single metal atoms by introducing heteroatoms and using covalent coordination between the metal and nonmetal. Choi et al. loaded 5 wt% Pt in atomically dispersed form on a high sulfur doped zeolite template carbon by a wet impregnation method.¹⁹⁷ The rich sulfur doping and 3D carbon structure provided a stable loading environment.

Compared to SACs, the density of active sites in DACs and TACs can be enhanced. For instance, the loading capacity of iron in the monatomic iron catalyst (Fe–N₄–C) and the positive catalyst of diatomic iron (Fe₂–N₆–C-o) synthesized by Han et al. was 0.155 and 0.313 wt%, respectively.¹⁸¹ More than 80% of the iron was identified as double metal sites by calculation, and the mean distance of Fe–Fe was 2.37 ± 0.31 Å. The Fe and Ni content in diatomic catalyst NiFe-DASC synthesized by Chen et al. using the pyrolysis strategy was 3.24 and 4.05 wt% respectively, which is twice the metal load of the Fe/Ni-SAC synthesized by the same method.¹⁹⁸ The HAADF-STEM images showed that the Ni and Fe atoms in the NiFe DASC are evenly dispersed, with a distance of about 0.24 nm between the Ni and Fe (Fig. 24a–g). Using the large number of periodic N–H functional groups in polymeric carbon nitride (PCN) as a metal coordination site, the metal can replace the H atom and can be fixed by N, avoiding agglomeration. Lu et al. used stepwise ion exchange and ligand removal strategies to synthesize low-coordination bimetallic atoms, where the content of Cu was up to 18.3 wt%.¹⁹⁹

Fig. 24 (a) Schematic illustration of NiFe-DASC synthesis. (b) Dark-field scanning TEM and AFM (inset) images. (c) TEM image and (d) EDS elemental mapping of NiFe-DASC. (e) HAADF-STEM image of NiFe-DASC. (f) Intensity profiles of the three sites in (e). (g) Electronic energy loss spectroscopy (EELS) of NiFe-DASC. Reproduced with permission.¹⁹⁸ Copyright 2021, Nature.

4.3. Enhancing intrinsic activity

To improve the inherent electrocatalytic activity of catalysts, the central metal atom and the coordination environment can be adjusted. Selecting an appropriate central metal atom as the active center can significantly improve the electrocatalytic activity and effectively adjust the product selectivity. For example, Fe-pyridine/pyrrole-N₄ and Mn-pyrrole-N₄ sites can catalyse the ORR, while the Co-pyrrole-N₄ site can catalyse the OER efficiently.²⁰⁰ Cu is the only metal species that can convert CO₂ to C₂₊ in the CO₂RR.²⁰¹ The coordination environment includes a coordination atom, base atom and axial ligand. Unsaturated central metal atoms interact with ligands to regulate the metallic energy band, spin state, and charge redistribution to enhance the inherent activity. The heteroatoms on the base plane can regulate the electronic structure of the active site through long-range delocalization. In addition, the interaction of the active site with reactants and reaction intermediates during electrocatalysis can also effectively regulate its activity. Therefore, the inherent activity of the central metal atoms can be changed by adjusting the coordination environment. Zhang et al. synthesized a series of (Mn, Fe, Co, Ni, Cu or Zn) framework porphyrin (POF) materials, and found that the ORR activity of Co-POF was superior to that of Fe-POF.²⁰² Lu et al. constructed an Fe–N₂S₂ active site, and found that the activity of the CO₂RR was significantly optimized compared to that on an Fe–N₄ porphyrin.²⁰³ On the Fe–N₄ porphyrin, the reduction to HCOOH is favourable at −0.70 V, while on the Fe–N₂S₂ porphyrin, CO₂ can be converted to HCOOH and CH₃OH at −0.38 V and −0.40 V, respectively (Fig. 25). The introduction of para-S atoms elongates the bond length of Fe–N, causing electrons to gather in the Fe–S bond, and the interaction between Fe and S is enhanced. Moreover, additional Fe orbitals in the Fe–N₂S₂ porphyrin can effectively regulate the adsorption/desorption energy of reaction intermediates and accelerate the CO₂RR.

Fig. 25 (a) The schematic of elementary steps of the CO₂RR. (b) CO₂ and H binding energies. (c) Volcano plot. (d) Bader charge difference of C and O atoms of the CO molecule and Fe atom before and after CO binding. (e) Charge density difference of CO binding. Reproduced with permission.²⁰³ Copyright 2021, Wiley-VCH GmbH.

In addition, adjacent active sites can regulate each other to promote the inherent electrocatalytic activity. Yu et al. demonstrated that the catalytic ORR activity was enhanced when the distance between the Fe–N₄ active sites was in the subnanometer level. When the distance between the two sites was less than 1.2 nm, the strong interaction between them altered the electronic structure and improved the inherent ORR activity. The activity is optimal when the distance is about 0.7 nm.²⁰⁴ Qiao et al. demonstrated theoretically and experimentally that when the distance between Ni–N₄ and Cu–N₄ is about 5.3 Å, the electronic structure of the active site can be effectively regulated to improve the selectivity and activity of the CO₂RR.²⁰⁵

4.4. Improving the stability

For atomic active sites, due to the high surface energy of metal atoms and the unsaturated coordination environment, the isolated metal sites will partially agglomerate under working conditions, resulting in a decrease in catalytic activity. The stability of active sites can be improved by enhancing the interaction between isolated metals and ligands and carriers. The electronegativity of the B atom is less than that of the N atom, which can effectively enhance the electron density of the Ni–N₄ center and strengthen the binding ability of the Ni site with reaction intermediates. Xin et al. synthesized a SAC (Ni–B/N–C) with Ni–B₁N₃ sites, which had a higher E_1/2 (0.87 V) than Ni–N–C (0.76 V) in the ORR, and was more stable than most Fe–N–C and Co–N–C catalysts.¹⁷⁴ In addition, based on the degradation mechanism of Fe–N–C SACs, some strategies have been proposed, such as the introduction of a co-catalyst and free radical scavenger in the active sites to effectively reduce the degradation. Sun et al. anchored CeO₂ NPs around Fe–N₄ active sites by an organic gas doping method. CeO₂ NPs were used to eliminate the ˙OH and radical generated at the Fe–N₄ site (Fig. 26).²⁰⁶ When the Fe-NC/Sca^ad-CeO₂ catalyst was assembled into a fuel cell, the attenuation of peak power density was reduced from 69% to 28%.

Fig. 26 Schematic illustration of the radical elimination behaviour. Reproduced with permission.²⁰⁶ Copyright 2023, Wiley-VCH GmbH.

In addition, the active sites can be reconstructed under applied voltage, and the structure is restored upon removal of applied voltage, which is also considered to be a stable catalyst. Fontecave et al. reported nitrogen-doped graphene with CuN₄ sites (Cu–N–C), which reduced CO₂ to ethanol with an FE of 55%.²⁰⁷ Under the applied voltage, the isolated CuN₄ transformed into copper clusters, and Cu–N–C recovered after the voltage was removed. Operando XAS showed that the transformed copper clusters were the active structure. During electrolysis, the reconstruction of CuN₄ in Cu–N–C was reversible.

5. Conclusions and outlook

In the past decade, SACs have made great progress in the field of electrocatalysis due to their unique geometry and electronic structures, which are expected to replace commercial precious metal-based catalysts. DACs with the interaction between metal dimers are also emerging and show excellent electrocatalytic activity, which can solve the limitation caused by SACs to a certain extent. To better suit multi-electron complex reactions, TACs have been developed, which provide more possibilities for the adsorption of intermediates, and are conducive to better study the relationship between the structure and electrocatalytic activity. In this review, we summarize the progress of SACs, DACs and TACs in electrocatalysis systematically. By using defect engineering and an atomic limiting strategy to design precursors, the number of central metal atoms and the coordination environment can be regulated for specific electrocatalytic reactions. Up to now, SACs and DACs designed by electronic regulation have made good progress in electrocatalysis. As a rising star, TACs are in the initial stage of development. In general, electrocatalysts with different nuclear numbers and uniform active sites have broad application prospects, but some problems still need to be solved.

(i) Controlled synthesis is still one of the biggest obstacles for atomic catalysts. At present, most of the atomic catalysts are synthesized by pyrolysis at high temperatures. The randomness of the coordination environment of the obtained catalysts is relatively large, and metal migration and agglomeration will inevitably occur. In addition, for DACs and TACs, active sites may possess multiple structures. By encapsulating precursors into MOFs, COFs and other materials with pore structure, DACs and TACs can be synthesized through high-temperature pyrolysis. However, this method needs to ensure that the size of the precursor and the pores is compatible, which brings certain obstacles to the accurate control of the coordination environment. To solve the issue of migration and agglomeration of metal atoms during high temperature pyrolysis, Joule thermal synthesis technology with ultra-fast heating and cooling (millisecond level) processes has been developed. However, in the current synthesis technology, single-atom sites are unavoidably formed in DACs and TACs, resulting in the reduction of the target effective active sites. In addition, the atomic distance and configuration between active sites cannot be guaranteed to be consistent. The ratio of the target effective active sites to other structural sites plays a crucial role in determining catalytic activity. Therefore, it is pivotal to develop new and operable synthesis techniques to achieve precise regulation of the coordination environment and to better explain the relationship between structure and mechanism.

(ii) Atomic catalysts synthesized by controllable technology provide a good platform for studying the structure–activity relationships and reaction mechanisms. In catalytic processes, reaction intermediates adsorbed at active sites are dynamically evolved. In addition, the structures of some catalysts change under applied voltage, and therefore it is necessary to in situ detect the changes of active sites and reaction intermediates in catalytic reactions. At present, some in situ techniques have been developed, such as in situ XAS, IR, Raman, and TEM, but there are still issues in time accuracy and atomic resolution. For example, it is difficult to distinguish coordination atoms with similar atomic numbers and similar bonds (e.g. Co–N and Co–O). Therefore, advanced characterization techniques are needed to deeply understand the dynamic evolution of active sites during electrolysis.

(iii) At present, DFT calculations have been successfully applied to guide the selection of atomic catalysts. However, screening the coordination environment of metal center atoms is limited by computational cost and time, and traditional DFT calculations are not suitable for such a huge working system. With the development of computers, the first-principles high-throughput screening method will guide the experimental design of more excellent catalysts. In addition, the high-throughput method can also be used to screen the structures of reaction intermediates in catalytic reactions, and the combination with in situ characterizations will further promote the study of dynamic mechanisms. However, these calculations are carried out in an ideal environment, and the external conditions and internal environment in the electrocatalysis process are very complicated, which makes experimental results deviate from theoretical simulation to a certain extent, and a relatively large model is required to fully simulate the electrocatalytic environment. Therefore, new methods need to be developed in a timely manner.

(iv) Industrial grade catalysts need to have high activity and stability. Although atomic catalysts have made good progress, they are still in the basic research stage. At present, theoretical calculations of high-throughput screening can provide guidance for high-performance catalysts. Catalysts synthesized in the laboratory are in the gram scale, whereas the kilogram scale is commercially required. Therefore, catalyst amplification synthesis is also a key for commercial applications, since the mass-synthesized catalysts need to maintain high activity and stability. Moreover, the test of catalyst stability in the laboratory usually takes several hundred hours, but it falls far short of commercial requirements, since it'll take at least a few months. Therefore, more synthetic methods should be developed to synthesize catalysts that can effectively stabilize the active sites in corrosive electrolytes and reduce economic losses caused by frequent catalyst replacement.

In short, although atomic catalysts have made good progress in the field of electrocatalysis, there are still some issues to be addressed. However, we believe that with the development of new synthesis methods, characterization techniques and computing theories, atomic catalysts will play a significant role in the field of energy storage and conversion in the future.

Author contributions

All of the authors contributed to the literature search, writing and editing of this review.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 22075099) and the Natural Science Foundation of Jilin Province (No. 20220101051JC).

Notes and references

1. A. Wang, J. Li and T. Zhang, Nat. Rev. Chem, 2018, 2, 65–81.
2. L. Xu, P. Trogadas and M.-O. Coppens, Adv. Energy Mater., 2023, 13, 2302974.
3. L. Xiao, L. Qi, J. Sun, A. Husile, S. Zhang, Z. Wang and J. Guan, Nano Energy, 2024, 120, 109155.
4. Y.-X. Zhao, J.-H. Wen, P. Li, P.-F. Zhang, S.-N. Wang, D.-C. Li, J.-M. Dou, Y.-W. Li, H.-Y. Ma and L. Xu, Angew. Chem., Int. Ed., 2023, 62, e202216950.
5. J. Chang, Q. Zhang, J. Yu, W. Jing, S. Wang, G. Yin, G. I. N. Waterhouse and S. Lu, Adv. Sci., 2023, 10, 2301656.
6. J. Quílez-Bermejo, S. García-Dalí, A. Daouli, A. Zitolo, R. L. S. Canevesi, M. Emo, M. T. Izquierdo, M. Badawi, A. Celzard and V. Fierro, Adv. Funct. Mater., 2023, 33, 2300405.
7. T. Sun, W. Zang, J. Sun, C. Li, J. Fan, E. Liu and J. Wang, Adv. Funct. Mater., 2023, 33, 2301526.
8. S. Chen, Y. Gao, W. Wang, O. V. Prezhdo and L. Xu, ACS Nano, 2023, 17, 1522–1532.
9. Y. Gao, B. Liu and D. Wang, Adv. Mater., 2023, 35, 2209654.
10. F. Humblot, J. P. Candy, F. Le Peltier, B. Didillon and J. M. Basset, J. Catal., 1998, 179, 459–468.
11. J. Corker, F. Lefebvre, C. Lécuyer, V. Dufaud, F. Quignard, A. Choplin, J. Evans and J.-M. Basset, Science, 1996, 271, 966–969.
12. 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.
13. C. Zhu, S. Fu, Q. Shi, D. Du and Y. Lin, Angew. Chem., Int. Ed., 2017, 56, 13944–13960.
14. T. Tang, Z. Wang and J. Guan, Adv. Funct. Mater., 2022, 32, 2111504.
15. X. Bai, J. Han, X. Niu and J. Guan, Nano Res., 2023, 16, 10796–10802.
16. T. Tang, Z. Wang and J. Guan, Coord. Chem. Rev., 2023, 492, 215288.
17. L. Sun, V. Reddu and X. Wang, Chem. Soc. Rev., 2022, 51, 8923–8956.
18. B. Zhang, Y. Chen, J. Wang, H. Pan and W. Sun, Adv. Funct. Mater., 2022, 32, 2202227.
19. Y. Zhou, R. Lu, X. Tao, Z. Qiu, G. Chen, J. Yang, Y. Zhao, X. Feng and K. Müllen, J. Am. Chem. Soc., 2023, 145, 3647–3655.
20. Y. Chen, J. Mao, H. Zhou, L. Xing, S. Qiao, J. Yuan, B. Mei, Z. Wei, S. Zhao, Y. Tang and C. Liu, Adv. Funct. Mater., 2023, 2311664.
21. W. Pei, S. Zhou, J. Zhao, X. Xu, Y. Du and S. X. Dou, Nano Energy, 2020, 76, 105049.
22. Q.-F. Chen, Y. Xiao, R.-Z. Liao and M.-T. Zhang, CCS Chem., 2022, 5, 245–256.
23. D. E. Canfield, A. N. Glazer and P. G. Falkowski, Science, 2010, 330, 192–196.
24. R. L. Robson, R. R. Eady, T. H. Richardson, R. W. Miller, M. Hawkins and J. R. Postgate, Nature, 1986, 322, 388–390.
25. B. K. Burgess and D. J. Lowe, Chem. Rev., 1996, 96, 2983–3012.
26. J. Rittle and J. C. Peters, J. Am. Chem. Soc., 2016, 138, 4243–4248.
27. Z. Wei, J. He, Y. Yang, Z. Xia, Y. Feng and J. Ma, J. Energy Chem., 2021, 53, 303–308.
28. P. Saha, S. Amanullah and A. Dey, J. Am. Chem. Soc., 2020, 142, 17312–17317.
29. J. Deng, H. Li, J. Xiao, Y. Tu, D. Deng, H. Yang, H. Tian, J. Li, P. Ren and X. Bao, Energy Environ. Sci., 2015, 8, 1594–1601.
30. B. Wurster, D. Grumelli, D. Hötger, R. Gutzler and K. Kern, J. Am. Chem. Soc., 2016, 138, 3623–3626.
31. S. Ji, Y. Chen, Q. Fu, Y. Chen, J. Dong, W. Chen, Z. Li, Y. Wang, L. Gu, W. He, C. Chen, Q. Peng, Y. Huang, X. Duan, D. Wang, C. Draxl and Y. Li, J. Am. Chem. Soc., 2017, 139, 9795–9798.
32. Y. Pan, R. Lin, Y. Chen, S. Liu, W. Zhu, X. Cao, W. Chen, K. Wu, W.-C. Cheong, Y. Wang, L. Zheng, J. Luo, Y. Lin, Y. Liu, C. Liu, J. Li, Q. Lu, X. Chen, D. Wang, Q. Peng, C. Chen and Y. Li, J. Am. Chem. Soc., 2018, 140, 4218–4221.
33. W. Ren, X. Tan, W. Yang, C. Jia, S. Xu, K. Wang, S. C. Smith and C. Zhao, Angew. Chem., Int. Ed., 2019, 58, 6972–6976.
34. Y. Yang, L. Zhang, Z. Hu, Y. Zheng, C. Tang, P. Chen, R. Wang, K. Qiu, J. Mao, T. Ling and S.-Z. Qiao, Angew. Chem., Int. Ed., 2020, 59, 4525–4531.
35. Q.-F. Chen, Z.-Y. Cheng, R.-Z. Liao and M.-T. Zhang, J. Am. Chem. Soc., 2021, 143, 19761–19768.
36. M. Liu, H. Chun, T.-C. Yang, S. J. Hong, C.-M. Yang, B. Han and L. Y. S. Lee, ACS Nano, 2022, 16, 10657–10666.
37. Y.-X. Zhang, S. Zhang, H. Huang, X. Liu, B. Li, Y. Lee, X. Wang, Y. Bai, M. Sun, Y. Wu, S. Gong, X. Liu, Z. Zhuang, T. Tan and Z. Niu, J. Am. Chem. Soc., 2023, 145, 4819–4827.
38. X. Yan, D. Liu, P. Guo, Y. He, X. Wang, Z. Li, H. Pan, D. Sun, F. Fang and R. Wu, Adv. Mater., 2023, 35, 2210975.
39. X. Xie, Z. Zhai, L. Peng, J. Zhang, L. Shang and T. Zhang, Sci. Bull., 2023, 68, 2862–2875.
40. Y. Chen, J. Lin, B. Jia, X. Wang, S. Jiang and T. Ma, Adv. Mater., 2022, 34, 2201796.
41. L. Li, K. Yuan and Y. Chen, Acc. Mater. Res., 2022, 3, 584–596.
42. Y. Zhang, L. Guo, L. Tao, Y. Lu and S. Wang, Small Methods, 2019, 3, 1800406.
43. S. Yuan, J. Zhang, L. Hu, J. Li, S. Li, Y. Gao, Q. Zhang, L. Gu, W. Yang, X. Feng and B. Wang, Angew. Chem., Int. Ed., 2021, 60, 21685–21690.
44. R. Jiang, Z. Qiao, H. Xu and D. Cao, Chin. J. Catal., 2023, 48, 224–234.
45. L. Zhang, Y. Jia, G. Gao, X. Yan, N. Chen, J. Chen, M. T. Soo, B. Wood, D. Yang, A. Du and X. Yao, Chem, 2018, 4, 285–297.
46. J. Liu, M. Jiao, B. Mei, Y. Tong, Y. Li, M. Ruan, P. Song, G. Sun, L. Jiang, Y. Wang, Z. Jiang, L. Gu, Z. Zhou and W. Xu, Angew Chem. Int. Ed. Engl., 2019, 58, 1163–1167.
47. J. Wan, W. Chen, C. Jia, L. Zheng, J. Dong, X. Zheng, Y. Wang, W. Yan, C. Chen, Q. Peng, D. Wang and Y. Li, Adv. Mater., 2018, 30, 1705369.
48. J. Jiao, R. Lin, S. Liu, W.-C. Cheong, C. Zhang, Z. Chen, Y. Pan, J. Tang, K. Wu, S.-F. Hung, H. M. Chen, L. Zheng, Q. Lu, X. Yang, B. Xu, H. Xiao, J. Li, D. Wang, Q. Peng, C. Chen and Y. Li, Nat. Chem., 2019, 11, 222–228.
49. L. Zhang, J. M. T. A. Fischer, Y. Jia, X. Yan, W. Xu, X. Wang, J. Chen, D. Yang, H. Liu, L. Zhuang, M. Hankel, D. J. Searles, K. Huang, S. Feng, C. L. Brown and X. Yao, J. Am. Chem. Soc., 2018, 140, 10757–10763.
50. J. Liang, Y. Jiao, M. Jaroniec and S. Z. Qiao, Angew. Chem., Int. Ed., 2012, 51, 11496–11500.
51. X. Bai, J. Han, S. Chen, X. Niu and J. Guan, Chin. J. Catal., 2023, 54, 212–219.
52. Z. Wang, C. Li, Y. Liu, Y. Wu, S. Zhang and C. Deng, J. Energy Chem., 2023, 83, 264–274.
53. L. Li, W. Yu, W. Gong, H. Wang, C.-L. Chiang, Y. Lin, J. Zhao, L. Zhang, J.-M. Lee and G. Zou, Appl. Catal., B, 2023, 321, 122038.
54. J. Han and J. Guan, Chin. J. Catal., 2023, 49, 1–4.
55. Q. Sun, C. Jia, Y. Zhao and C. Zhao, Chin. J. Catal., 2022, 43, 1547–1597.
56. J. Li, L. Zhang, K. Doyle-Davis, R. Li and X. Sun, Carbon Energy, 2020, 2, 488–520.
57. J. Han, X. Meng, L. Lu, J. Bian, Z. Li and C. Sun, Adv. Funct. Mater., 2019, 29, 1808872.
58. V. Hasija, S. Patial, P. Raizada, A. Aslam Parwaz Khan, A. M. Asiri, Q. Van Le, V.-H. Nguyen and P. Singh, Coord. Chem. Rev., 2022, 452, 214298.
59. N. Han, Y. Wang, L. Ma, J. Wen, J. Li, H. Zheng, K. Nie, X. Wang, F. Zhao, Y. Li, J. Fan, J. Zhong, T. Wu, D. J. Miller, J. Lu, S.-T. Lee and Y. Li, Chem, 2017, 3, 652–664.
60. L. Xiao, Z. Wang and J. Guan, Adv. Funct. Mater., 2023, 2310195.
61. Y. Chen, S. Ji, Y. Wang, J. Dong, W. Chen, Z. Li, R. Shen, L. Zheng, Z. Zhuang, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2017, 56, 6937–6941.
62. W. Chen, J. Pei, C.-T. He, J. Wan, H. Ren, Y. Wang, J. Dong, K. Wu, W.-C. Cheong, J. Mao, X. Zheng, W. Yan, Z. Zhuang, C. Chen, Q. Peng, D. Wang and Y. Li, Adv. Mater., 2018, 30, 1800396.
63. Y. Yue, P. Cai, K. Xu, H. Li, H. Chen, H.-C. Zhou and N. Huang, J. Am. Chem. Soc., 2021, 143, 18052–18060.
64. C.-X. Zhao, J.-N. Liu, J. Wang, C. Wang, X. Guo, X.-Y. Li, X. Chen, L. Song, B.-Q. Li and Q. Zhang, Sci. Adv., 2022, 8, eabn5091.
65. W.-H. Li, J. Yang and D. Wang, Angew. Chem., Int. Ed., 2022, 61, e202213318.
66. R. Li and D. Wang, Adv. Energy Mater., 2022, 12, 2103564.
67. H. Zhang, X. Jin, J.-M. Lee and X. Wang, ACS Nano, 2022, 16, 17572–17592.
68. P. Zhu, X. Xiong, D. Wang and Y. Li, Adv. Energy Mater., 2023, 13, 2300884.
69. L. Xiao, Z. Wang and J. Guan, Chem. Sci., 2023, 14, 12850–12868.
70. T. Tang, Z. Wang and J. Guan, Chin. J. Catal., 2022, 43, 636–678.
71. T. Tang, S. Li, J. Sun, Z. Wang and J. Guan, Nano Res., 2022, 15, 8714–8750.
72. Y. Yang, Y. Yu, J. Li, Q. Chen, Y. Du, P. Rao, R. Li, C. Jia, Z. Kang, P. Deng, Y. Shen and X. Tian, Nano-Micro Lett., 2021, 13, 160.
73. J. Park, S. Lee, H.-E. Kim, A. Cho, S. Kim, Y. Ye, J. W. Han, H. Lee, J. H. Jang and J. Lee, Angew. Chem., Int. Ed., 2019, 58, 16038–16042.
74. K. Jiang, B. Liu, M. Luo, S. Ning, M. Peng, Y. Zhao, Y.-R. Lu, T.-S. Chan, F. M. F. de Groot and Y. Tan, Nat. Commun., 2019, 10, 1743.
75. C. Tsounis, B. Subhash, P. V. Kumar, N. M. Bedford, Y. Zhao, J. Shenoy, Z. Ma, D. Zhang, C. Y. Toe, S. Cheong, R. D. Tilley, X. Lu, L. Dai, Z. Han and R. Amal, Adv. Funct. Mater., 2022, 32, 2203067.
76. W. Wang, Y. Wu, Y. Lin, J. Yao, X. Wu, C. Wu, X. Zuo, Q. Yang, B. Ge, L. Yang, G. Li, S. Chou, W. Li and Y. Jiang, Adv. Funct. Mater., 2022, 32, 2108464.
77. H. Yan, C. Tian, L. Wang, A. Wu, M. Meng, L. Zhao and H. Fu, Angew. Chem., Int. Ed., 2015, 54, 6325–6329.
78. Q. Jin, N. Liu, C. Dai, R. Xu, B. Wu, G. Yu, B. Chen and Y. Du, Adv. Energy Mater., 2020, 10, 2000291.
79. M. Li, J. Yu, Q. Liu, J. Liu, R. Chen, J. Zhu, R. Li and J. Wang, ACS Sustain. Chem. Eng., 2022, 10, 13505–13513.
80. J. Wang, X. Ge, L. Shao, J. Zhang, D. Peng, G. Zou, H. Hou, W. Deng, S. Xu, X. Ji and W. Zhang, Mater. Today Energy, 2020, 17, 100436.
81. J. Yin, Q. Fan, Y. Li, F. Cheng, P. Zhou, P. Xi and S. Sun, J. Am. Chem. Soc., 2016, 138, 14546–14549.
82. J. Han and J. Guan, Nano Res., 2023, 16, 1913–1966.
83. S. Chen, T. Zhang, J. Han, H. Qi, S. Jiao, C. Hou and J. Guan, Nano Res. Energy, 2024, 3, e9120106.
84. T. Tang, Z. Duan, D. Baimanov, X. Bai, X. Liu, L. Wang, Z. Wang and J. Guan, Nano Res., 2023, 16, 2218–2223.
85. H. Xu, T. Liu, S. Bai, L. Li, Y. Zhu, J. Wang, S. Yang, Y. Li, Q. Shao and X. Huang, Nano Lett., 2020, 20, 5482–5489.
86. Q. Wang, Z. Zhang, C. Cai, M. Wang, Z. L. Zhao, M. Li, X. Huang, S. Han, H. Zhou, Z. Feng, L. Li, J. Li, H. Xu, J. S. Francisco and M. Gu, J. Am. Chem. Soc., 2021, 143, 13605–13615.
87. X. Bai, Y. Wang, J. Han, X. Niu and J. Guan, Appl. Catal., B, 2023, 337, 122966.
88. P. Kumar, K. Kannimuthu, A. S. Zeraati, S. Roy, X. Wang, X. Wang, S. Samanta, K. A. Miller, M. Molina, D. Trivedi, J. Abed, M. A. Campos Mata, H. Al-Mahayni, J. Baltrusaitis, G. Shimizu, Y. A. Wu, A. Seifitokaldani, E. H. Sargent, P. M. Ajayan, J. Hu and M. G. Kibria, J. Am. Chem. Soc., 2023, 145, 8052–8063.
89. Q. Shi, C. Zhu, D. Du and Y. Lin, Chem. Soc. Rev., 2019, 48, 3181–3192.
90. Y. Lee, J. Suntivich, K. J. May, E. E. Perry and Y. Shao-Horn, J. Phys. Chem. Lett., 2012, 3, 399–404.
91. W. T. Hong, M. Risch, K. A. Stoerzinger, A. Grimaud, J. Suntivich and Y. Shao-Horn, Energy Environ. Sci., 2015, 8, 1404–1427.
92. C. Zhang, Y.-C. Wang, B. An, R. Huang, C. Wang, Z. Zhou and W. Lin, Adv. Mater., 2017, 29, 1604556.
93. Q. Liu, X. Liu, L. Zheng and J. Shui, Angew. Chem., Int. Ed., 2018, 57, 1204–1208.
94. M. Luo, Z. Zhao, Y. Zhang, Y. Sun, Y. Xing, F. Lv, Y. Yang, X. Zhang, S. Hwang, Y. Qin, J.-Y. Ma, F. Lin, D. Su, G. Lu and S. Guo, Nature, 2019, 574, 81–85.
95. X. Han, X. Ling, Y. Wang, T. Ma, C. Zhong, W. Hu and Y. Deng, Angew. Chem., Int. Ed., 2019, 58, 5359–5364.
96. C. Xia, Y. Xia, P. Zhu, L. Fan and H. Wang, Science, 2019, 366, 226–231.
97. J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard and H. Jónsson, J. Phys. Chem. B, 2004, 108, 17886–17892.
98. C.-X. Zhao, B.-Q. Li, J.-N. Liu and Q. Zhang, Angew. Chem., Int. Ed., 2021, 60, 4448–4463.
99. X. Xie, L. Peng, H. Yang, G. I. N. Waterhouse, L. Shang and T. Zhang, Adv. Mater., 2021, 33, 2101038.
100. Y. Sun, L. Silvioli, N. R. Sahraie, W. Ju, J. Li, A. Zitolo, S. Li, A. Bagger, L. Arnarson, X. Wang, T. Moeller, D. Bernsmeier, J. Rossmeisl, F. Jaouen and P. Strasser, J. Am. Chem. Soc., 2019, 141, 12372–12381.
101. M. Kunitski, N. Eicke, P. Huber, J. Köhler, S. Zeller, J. Voigtsberger, N. Schlott, K. Henrichs, H. Sann, F. Trinter, L. P. H. Schmidt, A. Kalinin, M. S. Schöffler, T. Jahnke, M. Lein and R. Dörner, Nat. Commun., 2019, 10, 1.
102. C. Wang, D. Wang, S. Liu, P. Jiang, Z. Lin, P. Xu, K. Yang, J. Lu, H. Tong, L. Hu, W. Zhang and Q. Chen, J. Catal., 2020, 389, 150–156.
103. J. Zhang, Y. Zhao, C. Chen, Y.-C. Huang, C.-L. Dong, C.-J. Chen, R.-S. Liu, C. Wang, K. Yan, Y. Li and G. Wang, J. Am. Chem. Soc., 2019, 141, 20118–20126.
104. Y. Wang, R. Shi, L. Shang, G. I. N. Waterhouse, J. Zhao, Q. Zhang, L. Gu and T. Zhang, Angew. Chem., Int. Ed., 2020, 59, 13057–13062.
105. Q. Zhao, Y. Wang, W.-H. Lai, F. Xiao, Y. Lyu, C. Liao and M. Shao, Energy Environ. Sci., 2021, 14, 5444–5456.
106. Y. Zhao, J. Liang, C. Wang, J. Ma and G. G. Wallace, Adv. Energy Mater., 2018, 8, 1702524.
107. X. Zu, X. Li, W. Liu, Y. Sun, J. Xu, T. Yao, W. Yan, S. Gao, C. Wang, S. Wei and Y. Xie, Adv. Mater., 2019, 31, 1808135.
108. Z. Jiang, T. Wang, J. Pei, H. Shang, D. Zhou, H. Li, J. Dong, Y. Wang, R. Cao, Z. Zhuang, W. Chen, D. Wang, J. Zhang and Y. Li, Energy Environ. Sci., 2020, 13, 2856–2863.
109. S. Kusama, T. Saito, H. Hashiba, A. Sakai and S. Yotsuhashi, ACS Catal., 2017, 7, 8382–8385.
110. 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.
111. T. Wang, L. Xu, Z. Chen, L. Guo, Y. Zhang, R. Li and T. Peng, Appl. Catal., B, 2021, 291, 120128.
112. Z. Weng, Y. Wu, M. Wang, J. Jiang, K. Yang, S. Huo, X.-F. Wang, Q. Ma, G. W. Brudvig, V. S. Batista, Y. Liang, Z. Feng and H. Wang, Nat. Commun., 2018, 9, 415.
113. C. Zhang, S. Yang, J. Wu, M. Liu, S. Yazdi, M. Ren, J. Sha, J. Zhong, K. Nie, A. S. Jalilov, Z. Li, H. Li, B. I. Yakobson, Q. Wu, E. Ringe, H. Xu, P. M. Ajayan and J. M. Tour, Adv. Energy Mater., 2018, 8, 1703487.
114. T. Tang, Z. Wang and J. Guan, Acta Phys.-Chim. Sin., 2022, 39, 2208033.
115. Q. Chang, Y. Liu, J.-H. Lee, D. Ologunagba, S. Hwang, Z. Xie, S. Kattel, J. H. Lee and J. G. Chen, J. Am. Chem. Soc., 2022, 144, 16131–16138.
116. J. Wang, Y.-C. Huang, Y. Wang, H. Deng, Y. Shi, D. Wei, M. Li, C.-L. Dong, H. Jin, S. S. Mao and S. Shen, ACS Catal., 2023, 13, 2374–2385.
117. Z. Ma, T. Wan, D. Zhang, J. A. Yuwono, C. Tsounis, J. Jiang, Y.-H. Chou, X. Lu, P. V. Kumar, Y. H. Ng, D. Chu, C. Y. Toe, Z. Han and R. Amal, ACS Nano, 2023, 17, 2387–2398.
118. H. Li, F. Pan, C. Qin, T. Wang and K.-J. Chen, Adv. Energy Mater., 2023, 13, 2301378.
119. M. Humayun, M. Israr, A. Khan and M. Bououdina, Nano Energy, 2023, 113, 108570.
120. L. Han, X. Liu, J. Chen, R. Lin, H. Liu, F. Lü, S. Bak, Z. Liang, S. Zhao, E. Stavitski, J. Luo, R. R. Adzic and H. L. Xin, Angew. Chem., Int. Ed., 2019, 58, 2321–2325.
121. Z. Xue, X. Zhang, J. Qin and R. Liu, J. Energy Chem., 2021, 57, 443–450.
122. J. Han and J. Guan, Chin. J. Catal., 2023, 47, 1–31.
123. Y. Li, C. Chen, R. Cao, Z. Pan, H. He and K. Zhou, Appl. Catal., B, 2020, 268, 118747.
124. N. Zhang, X. Zhang, Y. Kang, C. Ye, R. Jin, H. Yan, R. Lin, J. Yang, Q. Xu, Y. Wang, Q. Zhang, L. Gu, L. Liu, W. Song, J. Liu, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2021, 60, 13388–13393.
125. X. Zhao, K. Zhao, Y. Liu, Y. Su, S. Chen, H. Yu and X. Quan, ACS Catal., 2022, 12, 11412–11420.
126. T. Tang, Y. Wang, J. Han, Q. Zhang, X. Bai, X. Niu, Z. Wang and J. Guan, Chin. J. Catal., 2023, 46, 48–55.
127. Y. Ying, X. Luo, J. Qiao and H. Huang, Adv. Funct. Mater., 2021, 31, 2007423.
128. Y. Yang, Y. Qian, H. Li, Z. Zhang, Y. Mu, D. Do, B. Zhou, J. Dong, W. Yan, Y. Qin, L. Fang, R. Feng, J. Zhou, P. Zhang, J. Dong, G. Yu, Y. Liu, X. Zhang and X. Fan, Sci. Adv., 2020, 6, eaba6586.
129. Y. Da, Z. Tian, R. Jiang, Y. Liu, X. Lian, S. Xi, Y. Shi, Y. Wang, H. Lu, B. Cui, J. Zhang, X. Han, W. Chen and W. Hu, Sci. China Mater., 2023, 66, 1389–1397.
130. M. Jiao, Z. Chen, N. Wang and L. Liu, Appl. Catal., B, 2023, 324, 122244.
131. M. Jiang, F. Wang, F. Yang, H. He, J. Yang, W. Zhang, J. Luo, J. Zhang and C. Fu, Nano Energy, 2022, 93, 106793.
132. L. Bai, C.-S. Hsu, D. T. L. Alexander, H. M. Chen and X. Hu, J. Am. Chem. Soc., 2019, 141, 14190–14199.
133. Z. Pei, X. F. Lu, H. Zhang, Y. Li, D. Luan and X. W. Lou, Angew. Chem., Int. Ed., 2022, 61, e202207537.
134. Y. He, X. Yang, Y. Li, L. Liu, S. Guo, C. Shu, F. Liu, Y. Liu, Q. Tan and G. Wu, ACS Catal., 2022, 12, 1216–1227.
135. J.-d. Yi, X. Gao, H. Zhou, W. Chen and Y. Wu, Angew. Chem., Int. Ed., 2022, 61, e202212329.
136. M. Feng, X. Wu, H. Cheng, Z. Fan, X. Li, F. Cui, S. Fan, Y. Dai, G. Lei and G. He, J. Mater. Chem. A, 2021, 9, 23817–23827.
137. H. Cheng, X. Wu, M. Feng, X. Li, G. Lei, Z. Fan, D. Pan, F. Cui and G. He, ACS Catal., 2021, 11, 12673–12681.
138. S. Cao, S. Zhou, H. Chen, S. Wei, S. Liu, X. Lin, X. Chen, Z. Wang, W. Guo and X. Lu, Energy Environ. Mater., 2021, 6, e12287.
139. W. Yang, Z. Jia, B. Zhou, L. Chen, X. Ding, L. Jiao, H. Zheng, Z. Gao, Q. Wang and H. Li, ACS Catal., 2023, 13, 9695–9705.
140. D. Fang, P. Sun, S. Huang, Y. Shang, X. Li, D. Yan, Y. V. Lim, C.-Y. Su, B.-J. Su, J.-Y. Juang and H. Y. Yang, ACS Mater. Lett., 2022, 4, 1–10.
141. R. Wang, L.-Z. Dong, J.-W. Shi, M. Zhang, S.-L. Li, Y.-Q. Lan and J. Liu, ACS Catal., 2024, 14, 741–750.
142. D.-S. Huang, H.-L. Zhu, Z.-H. Zhao, J.-R. Huang, P.-Q. Liao and X.-M. Chen, ACS Catal., 2022, 12, 8444–8450.
143. M. Kato, M. Muto, N. Matsubara, Y. Uemura, Y. Wakisaka, T. Yoneuchi, D. Matsumura, T. Ishihara, T. Tokushima, S.-i. Noro, S. Takakusagi, K. Asakura and I. Yagi, ACS Appl. Energy Mater., 2018, 1, 2358–2364.
144. W. J. Li, Z. X. Lou, H. Y. Yuan, H. G. Yang and H. F. Wang, J. Mater. Chem. A, 2022, 10, 16106–16114.
145. Z. W. Chen, L. X. Chen, M. Jiang, D. Chen, Z. L. Wang, X. Yao, C. V. Singh and Q. Jiang, J. Mater. Chem. A, 2020, 8, 15086–15093.
146. X. Fu, X. Zhao, T.-B. Lu, M. Yuan and M. Wang, Angew. Chem., Int. Ed., 2023, 62, e202219242.
147. Y. Zhang, J. Yang, R. Ge, J. Zhang, J. M. Cairney, Y. Li, M. Zhu, S. Li and W. Li, Coord. Chem. Rev., 2022, 461, 214493.
148. G. Bae, S. Han, H.-S. Oh and C. H. Choi, Angew. Chem., Int. Ed., 2023, 62, e202219227.
149. Y. Wang, D. Wang and Y. Li, Adv. Mater., 2021, 33, 2008151.
150. P. Zhang, K. Chen, J. Li, M. Wang, M. Li, Y. Liu and Y. Pan, Adv. Mater., 2023, 35, 2303243.
151. Y. Zhu, J. Sokolowski, X. Song, Y. He, Y. Mei and G. Wu, Adv. Energy Mater., 2020, 10, 1902844.
152. D. Ji, L. Fan, L. Li, S. Peng, D. Yu, J. Song, S. Ramakrishna and S. Guo, Adv. Mater., 2019, 31, 1808267.
153. X. Li, W. Bi, M. Chen, Y. Sun, H. Ju, W. Yan, J. Zhu, X. Wu, W. Chu, C. Wu and Y. Xie, J. Am. Chem. Soc., 2017, 139, 14889–14892.
154. 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.
155. M. Tong, F. Sun, G. Xing, C. Tian, L. Wang and H. Fu, Angew. Chem., Int. Ed., 2023, 62, e202314933.
156. H. Zhang, J. Li, S. Xi, Y. Du, X. Hai, J. Wang, H. Xu, G. Wu, J. Zhang, J. Lu and J. Wang, Angew. Chem., Int. Ed., 2019, 58, 14871–14876.
157. Y. Zhang, L. Jiao, W. Yang, C. Xie and H.-L. Jiang, Angew. Chem., Int. Ed., 2021, 60, 7607–7611.
158. W. Ren, X. Tan, W. Yang, C. Jia, S. Xu, K. Wang, S. C. Smith and C. Zhao, Angew. Chem., Int. Ed., 2019, 58, 6972–6976.
159. X.-M. Liang, H.-J. Wang, C. Zhang, D.-C. Zhong and T.-B. Lu, Appl. Catal., B, 2023, 322, 122073.
160. Y.-N. Gong, C.-Y. Cao, W.-J. Shi, J.-H. Zhang, J.-H. Deng, T.-B. Lu and D.-C. Zhong, Angew. Chem., Int. Ed., 2022, 61, e202215187.
161. X. Yang, Y. Wang, G. Zhang, L. Du, L. Yang, M. Markiewicz, J.-y. Choi, R. Chenitz and S. Sun, Appl. Catal., B, 2020, 264, 118523.
162. L. Yu, Y. Li and Y. Ruan, Angew. Chem., Int. Ed., 2021, 60, 25296–25301.
163. S. H. Lee, J. Kim, D. Y. Chung, J. M. Yoo, H. S. Lee, M. J. Kim, B. S. Mun, S. G. Kwon, Y.-E. Sung and T. Hyeon, J. Am. Chem. Soc., 2019, 141, 2035–2045.
164. B. Y. Xia, Y. Yan, N. Li, H. B. Wu, X. W. Lou and X. Wang, Nat. Energy, 2016, 1, 15006.
165. M. Ferrandon, X. Wang, A. J. Kropf, D. J. Myers, G. Wu, C. M. Johnston and P. Zelenay, Electrochim. Acta, 2013, 110, 282–291.
166. G. Yang, J. Zhu, P. Yuan, Y. Hu, G. Qu, B.-A. Lu, X. Xue, H. Yin, W. Cheng, J. Cheng, W. Xu, J. Li, J. Hu, S. Mu and J.-N. Zhang, Nat. Commun., 2021, 12, 1734.
167. S. Huang, Z. Qiao, P. Sun, K. Qiao, K. Pei, L. Yang, H. Xu, S. Wang, Y. Huang, Y. Yan and D. Cao, Appl. Catal., B, 2022, 317, 121770.
168. J. Zhang, M. Zhang, Y. Zeng, J. Chen, L. Qiu, H. Zhou, C. Sun, Y. Yu, C. Zhu and Z. Zhu, Small, 2019, 15, 1900307.
169. H. Shang, X. Zhou, J. Dong, A. Li, X. Zhao, Q. Liu, Y. Lin, J. Pei, Z. Li, Z. Jiang, D. Zhou, L. Zheng, Y. Wang, J. Zhou, Z. Yang, R. Cao, R. Sarangi, T. Sun, X. Yang, X. Zheng, W. Yan, Z. Zhuang, J. Li, W. Chen, D. Wang, J. Zhang and Y. Li, Nat. Commun., 2020, 11, 3049.
170. J. Wan, Z. Zhao, H. Shang, B. Peng, W. Chen, J. Pei, L. Zheng, J. Dong, R. Cao, R. Sarangi, Z. Jiang, D. Zhou, Z. Zhuang, J. Zhang, D. Wang and Y. Li, J. Am. Chem. Soc., 2020, 142, 8431–8439.
171. K. Li, S. Zhang, X. Zhang, S. Liu, H. Jiang, T. Jiang, C. Shen, Y. Yu and W. Chen, Nano Lett., 2022, 22, 1557–1565.
172. M. Huang, B. Deng, X. Zhao, Z. Zhang, F. Li, K. Li, Z. Cui, L. Kong, J. Lu, F. Dong, L. Zhang and P. Chen, ACS Nano, 2022, 16, 2110–2119.
173. Z. Fan, R. Luo, Y. Zhang, B. Zhang, P. Zhai, Y. Zhang, C. Wang, J. Gao, W. Zhou, L. Sun and J. Hou, Angew. Chem., Int. Ed., 2023, 62, e202216326.
174. F. Wang, R. Zhang, Y. Zhang, Y. Li, J. Zhang, W. Yuan, H. Liu, F. Wang and H. L. Xin, Adv. Funct. Mater., 2023, 33, 2213863.
175. Y. Yang, K. Mao, S. Gao, H. Huang, G. Xia, Z. Lin, P. Jiang, C. Wang, H. Wang and Q. Chen, Adv. Mater., 2018, 30, 1801732.
176. T. Ding, X. Liu, Z. Tao, T. Liu, T. Chen, W. Zhang, X. Shen, D. Liu, S. Wang, B. Pang, D. Wu, L. Cao, L. Wang, T. Liu, Y. Li, H. Sheng, M. Zhu and T. Yao, J. Am. Chem. Soc., 2021, 143, 11317–11324.
177. X. Li, S. Mitchell, Y. Fang, J. Li, J. Perez-Ramirez and J. Lu, Nat. Rev. Chem, 2023, 7, 754–767.
178. D.-C. Zhong, Y.-N. Gong, C. Zhang and T.-B. Lu, Chem. Soc. Rev., 2023, 52, 3170–3214.
179. Z. Pei, H. Zhang, Y. Guo, D. Luan, X. Gu and X. W. Lou, Adv. Mater., 2023, 2306047.
180. M. Liu, N. Li, S. Cao, X. Wang, X. Lu, L. Kong, Y. Xu and X.-H. Bu, Adv. Mater., 2022, 34, 2107421.
181. Y. Wang, B. J. Park, V. K. Paidi, R. Huang, Y. Lee, K.-J. Noh, K.-S. Lee and J. W. Han, ACS Energy Lett., 2022, 7, 640–649.
182. X. Yi, H. Yang, X. Yang, X. Li, C. Yan, J. Zhang, L. Chen, J. Dong, J. Qin, G. Zhang, J. Wang, W. Li, Z. Zhou, G. Wu and X. Li, Adv. Funct. Mater., 2023, 2309728.
183. B. Zhang, J. Hou, Y. Wu, S. Cao, Z. Li, X. Nie, Z. Gao and L. Sun, Adv. Energy Mater., 2019, 9, 1803693.
184. X. Yu, S. Lai, S. Xin, S. Chen, X. Zhang, X. She, T. Zhan, X. Zhao and D. Yang, Appl. Catal., B, 2021, 280, 119437.
185. D. Chen, J. Zhu, X. Mu, R. Cheng, W. Li, S. Liu, Z. Pu, C. Lin and S. Mu, Appl. Catal., B, 2020, 268, 118729.
186. L. Lv, R. Lu, J. Zhu, R. Yu, W. Zhang, E. Cui, X. Chen, Y. Dai, L. Cui, J. Li, L. Zhou, W. Chen, Z. Wang and L. Mai, Angew. Chem., Int. Ed., 2023, 62, e202303117.
187. R. Wang, L. Zhang, J. Shan, Y. Yang, J.-F. Lee, T.-Y. Chen, J. Mao, Y. Zhao, L. Yang, Z. Hu and T. Ling, Adv. Sci., 2022, 9, 2203917.
188. D. Xue, P. Yuan, S. Jiang, Y. Wei, Y. Zhou, C.-L. Dong, W. Yan, S. Mu and J.-N. Zhang, Nano Energy, 2023, 105, 108020.
189. Y. Wang, W. Cheng, P. Yuan, G. Yang, S. Mu, J. Liang, H. Xia, K. Guo, M. Liu, S. Zhao, G. Qu, B.-A. Lu, Y. Hu, J. Hu and J.-N. Zhang, Adv. Sci., 2021, 8, 2102915.
190. M. Liu, B.-A. Lu, G. Yang, P. Yuan, H. Xia, Y. Wang, K. Guo, S. Zhao, J. Liu, Y. Yu, W. Yan, C.-L. Dong, J.-N. Zhang and S. Mu, Adv. Sci., 2022, 9, 2200147.
191. Y. Liu, X. Liu, Z. Lv, R. Liu, L. Li, J. Wang, W. Yang, X. Jiang, X. Feng and B. Wang, Angew. Chem., Int. Ed., 2022, 61, e202117617.
192. K. Maiti, S. Maiti, M. T. Curnan, H. J. Kim and J. W. Han, Adv. Energy Mater., 2021, 11, 2101670.
193. A. Pedersen, J. Barrio, A. Li, R. Jervis, D. J. L. Brett, M. M. Titirici and I. E. L. Stephens, Adv. Energy Mater., 2022, 12, 2102715.
194. K. Leng, J. Zhang, Y. Wang, D. Li, L. Bai, J. Shi, X. Li, L. Zheng, J. Bai and Y. Qu, Adv. Funct. Mater., 2022, 32, 2205637.
195. X. Wang, L. Xu, C. Li, C. Zhang, H. Yao, R. Xu, P. Cui, X. Zheng, M. Gu, J. Lee, H. Jiang and M. Huang, Nat. Commun., 2023, 14, 7210.
196. Y. He, Q. Shi, W. Shan, X. Li, A. J. Kropf, E. C. Wegener, J. Wright, S. Karakalos, D. Su, D. A. Cullen, G. Wang, D. J. Myers and G. Wu, Angew. Chem., Int. Ed., 2021, 60, 9516–9526.
197. C. H. Choi, M. Kim, H. C. Kwon, S. J. Cho, S. Yun, H.-T. Kim, K. J. J. Mayrhofer, H. Kim and M. Choi, Nat. Commun., 2016, 7, 10922.
198. Z. Zeng, L. Y. Gan, H. Bin Yang, X. Su, J. Gao, W. Liu, H. Matsumoto, J. Gong, J. Zhang, W. Cai, Z. Zhang, Y. Yan, B. Liu and P. Chen, Nat. Commun., 2021, 12, 4088.
199. X. Hai, Y. Zheng, Q. Yu, N. Guo, S. Xi, X. Zhao, S. Mitchell, X. Luo, V. Tulus, M. Wang, X. Sheng, L. Ren, X. Long, J. Li, P. He, H. Lin, Y. Cui, X. Peng, J. Shi, J. Wu, C. Zhang, R. Zou, G. Guillén-Gosálbez, J. Pérez-Ramírez, M. J. Koh, Y. Zhu, J. Li and J. Lu, Nature, 2023, 622, 754–760.
200. H. Xu, D. Cheng, D. Cao and X. C. Zeng, Nat. Catal., 2018, 1, 339–348.
201. S. H. Lee, J. C. Lin, M. Farmand, A. T. Landers, J. T. Feaster, J. E. Avilés Acosta, J. W. Beeman, Y. Ye, J. Yano, A. Mehta, R. C. Davis, T. F. Jaramillo, C. Hahn and W. S. Drisdell, J. Am. Chem. Soc., 2021, 143, 588–592.
202. C.-X. Zhao, B.-Q. Li, J.-N. Liu, J.-Q. Huang and Q. Zhang, Chin. Chem. Lett., 2019, 30, 911–914.
203. S. Cao, S. Wei, X. Wei, S. Zhou, H. Chen, Y. Hu, Z. Wang, S. Liu, W. Guo and X. Lu, Small, 2021, 17, 2100949.
204. Z. Jin, P. Li, Y. Meng, Z. Fang, D. Xiao and G. Yu, Nat. Catal., 2021, 4, 615–622.
205. D. Yao, C. Tang, X. Zhi, B. Johannessen, A. Slattery, S. Chern and S.-Z. Qiao, Adv. Mater., 2022, 35, 2209386.
206. X. Cheng, X. Jiang, S. Yin, L. Ji, Y. Yan, G. Li, R. Huang, C. Wang, H. Liao, Y. Jiang and S. Sun, Angew. Chem., Int. Ed., 2023, 62, e202306166.
207. D. Karapinar, N. T. Huan, N. Ranjbar Sahraie, J. Li, D. Wakerley, N. Touati, S. Zanna, D. Taverna, L. H. Galvão Tizei, A. Zitolo, F. Jaouen, V. Mougel and M. Fontecave, Angew. Chem., Int. Ed., 2019, 58, 15098–15103.

---